April - June 2007 - Kasetsart University

April - June 2007 

Volume 41 Number 2 

http://www.rdi.ku.ac.th

KASETSART JOURNAL 

NATURAL SCIENCE 

(http://www.rdi.ku.ac.th) 

Kasetsart Journal (Natural Science) is a peer-reviewed journal of 

Kasetsart University (www.ku.ac.th) which publishes original research articles on 

natural sciences and other topics dealing with current knowledge and advances in 

technology. The Kasetsart Journal (Natural Science) is issued four times per year. Articles 

from researchers worldwide are welcomed. 

EDITORS 

Editor-in-Chief 

Wanchai Chanprasert 

Associate Professor, Seed Science and Technology, 

Faculty of Agriculture, Kasetsart University, Thailand 

e-mail: agrwcc@ku.ac.th 

Associate Editor 

Amara Thongpan 

Associate Professor, Cell and Molecular Biology, 

Faculty of Science, Kasetsart University, Thailand 

EDITORIAL BOARD 

Roberto Bonoan 

Researcher, Department Manager, Soil Science, 

National Tobacco Administration, Department of 

Agricualture, Philippines 

Korchoke Chantawarangul 

Assistant Professor, Civil Engineering, Faculty of 

Engineering, Kasetsart University, Thailand 

Parnjit Damrongkulkamjorn 

Assistant Professor, Electrical Engineering, Faculty of 

Engineering, Kasetsart University, Thailand 

Praparat Hormchan 

Associate Professor, Entomology, Faculty of 

Agriculture, Kasetsart University, Thailand 

Kanapol Jutamanee 

Associate Professor, Plant Physiology, Faculty of 

Science, Kasetsart University, Thailand 

Onanong Naivikul 

Professor, Cereal Chemistry and Technology, Faculty of 

Agro-Industry, Kasetsart University, Thailand 

Eiji Nawata 

Associate Professor, Tropical Agriculture, Kyoto 

University, Japan 

Saran Petpiroon 

Associate Professor, Marine Ecology and Pollution, 

Faculty of Fisheries, Kasetsart University, Thailand 

Witoon Prinyawiwatkul 

Professor, Food Science and Technology, Louisiana State 

University and LSU Agricultural Center, USA 

Hathairat Rimkeeree 

Assistant Professor, Product Development, Faculty of 

Agro-Industry, Kasetsart University, Thailand 

U Ravi Sangakkara 

Professor, Crop Ecology, Faculty of Agriculture, 

University of Peradeniya, Sri Lanka 

Uthaiwan Sangwanit 

Lecturer, Forest Resources, Faculty of Forestry, 

Kasetsart University, Thailand 

Thongchai Suwonsichon 

Associate Professor, Food Science, Faculty of Agro- 

Industry, Kasetsart University, Thailand 

Aree Thankijjanukij 

Associate Library Director, Agricultural Information 

Management, Kasetsart University 

Satoru Tsuchikawa 

Assistant Professor, Near Infrared Spectroscopy and 

Agricultural Science, Graduate School of 

Bioagricultural Sciences, Nagoya University, Japan 

Sirichai Wongnarkpet 

Assistant Professor, Veterinary Epidemiology, Faculty 

of Veterinary Medicine, Kasetsart University, Thailand 

Kenji Yamane 

Associate Professor, Floriculture and Horticulture, 

Faculty of Agriculture, Utsunomiya University, Japan 

MANAGING EDITORS 

Orawan Wongwanich 

Senior Researcher, Kasetsart University Research and 

Development Institute, Kasetsart University, Thailand 

Somporn Maneeprasopsuk 

Researcher, Kasetsart University Research and 

Development Institute, Kasetsart University, Thailand 

EDITORIAL ADVISORY BOARD 

Gerald T. Baker 

Professor, Entomology, Mississippi State University, USA 

A. Bruce Bishop 

Professor, Civil and Environmental Engineering, Utah State 

University, USA 

Samakkee Boonyawat 

Associate Professor, Forest Resource Management, Faculty of 

Forestry, Kasetsart University, Thailand 

Delores Chambers 

Assistant Professor, Food Science, College of Human Ecology, 

Kansas State University, USA 

Edgar Chambers 

Professor, Food Science, College of Human Ecology, Kansas State 

University, USA 

Angsumarn Chandrapatya 

Professor, Entomology, Faculty of Agriculture, Kasetsart University, 

Thailand 

Mauricio A. Elzo 

Professor, Animal Breeding and Genetics, University of Florida, 

USA 

John Hampton 

Professor, Seed Science and Technology, Lincoln University, 

New Zealand 

Parichat Hongsprabhas 

Assistant Professor, Food Science, Faculty of Agro-Industry, 


Sathaporn Jittapalapong 

Associate Professor, Veterinary Parasitology, Faculty of Veterinary 

Medicine, Kasetsart University, Thailand 

Helen H. Keenan 

Professor, Environmental Science, University of Strathclyde, 

Scotland 

Saichol Ketsa 

Professor, Postharvest Technology, Faculty of Agriculture, 


Jumras Limtrakul 

Professor, Physical Chemistry, Faculty of Science, Kasetsart 

University, Thailand 

Chitochi Miki 

Professor, Structural Engineering, Tokyo Institute of Technology, 

Japan 

Larry Miller 

Professor, Agricultural Education College of Food, Agriculture and 

Environment, Ohio State University, USA 

Tadashi Miyata 

Professor, Entomology, Nagoya University, Japan 

Punpiti Piamsa-nga 

Assistant Professor, Computer Engineering, Faculty of Engineering, 


Wiroj Rujopakarn 

Professor, Transportation Engineering, Faculty of Engineering, 


Ed Sarobol 

Associate Professor, Crop Physiology, Faculty of Agriculture, 


Narongrit Sombatsompop 

Professor, Polymer Processing, School of Energy and Materials, 

King Mongkut’s University of Technology Thonburi, Thailand 

Peerasak Srinives 

Professor, Plant Breeding, , Faculty of Agriculture at Kamphaeng 

Saen, Kasetsart University, Thailand 

Rungsit Suwanmankha 

Professor, Weed Science, Faculty of Agriculture, Kasetsart 

University, Thailand 

Chanvit Vajrabukka 

Professor, Animal Science, Physiology, Animal Behavior, Faculty 

of Agriculture, Kasetsart University, Thailand 

PUBLISHER 

The Kasetsart Journal (Natural Science) is published by Kasetsart University Research 

and Development Institute (KURDI), Kasetsart University, Bangkok, Thailand 

EDITORIAL OFFICE 

Kasetsart University Research and Development Institute (KURDI), Kasetsart 

University, Chatuchak, Bangkok 10900, Thailand 

TEL 66 (2) 5795548; FAX 66 (2) 5611474 

e-mail: kj_rdi@ku.ac.th

KASETSART JOURNAL 

NATURAL SCIENCE 

The publication of Kasetsart University 

VOLUME 41 April - June 2007 NUMBER 2 

Changing in TSS, TA and Sugar Contents and Sucrose Synthase Activity in Ethephon-Treated 

‘Pattavia’ Pineapple Fruit 

...................... Ngarmnij Chuenboonngarm, Niran Juntawong, Arunee Engkagul, 

.............................................................. Wallop Arirob and Surin Peyachoknakul 205 

Phylogenetic Analysis of Thai Amomum (Alpinioideae: Zingiberaceae) Using AFLP Markers 

............. Wittaya Kaewsri, Yingyong Paisooksantivatana, Uamporn Veesommai, 

............................................................. Wichan Eiadthong and Srunya Vajrodaya 213 

Prediction of Sweet Corn Seeds Field Emergence under Wet Soil Condition 

............................................ Vichai Wongvarodom and Wikanate Rangsikansong 227 

Modifying Controlled Deterioration for Evaluating Field Weathering Resistance of Soybean 

.............................................Ye Changrong, Prapa Sripichitt, Sunanta Juntakool, 

....................................................................Vipa Hongtrakul and Arom Sripichitt 232 

Composite Line Method for the Development of Early Generation Hybrids 

of Maize (Zea mays L.) 

................. Nguyen Phuong, Krisda Samphantharak and Vatcharee Lertmongkol 242 

Anther Culture of BC 1F 1 (KDML105//IRBB5/KDML105) Hybrid to Produce Bacterial Blight 

Resistance Doubled Haploid Rice 

................................. Supanyika Sengsai, Surin Peyachoknagul, Prapa Sripichitt, 

.......................................................... Amara Thongpan and Pradit Pongtongkam 251 

Novel PCR Primers for Specific Detection of Xanthomonas citri subsp. citri the Causal Agent 

of Bacterial Citrus Canker 

Udomsak Lertsuchatavanich, Ampaiwan Paradornuwat, Junlapark Chunwongse, 

....................................................... Norman W. Schaad and Niphone Thaveechai 262 

Soil-to-Plant Transfer of Radiocaesium in Thailand 

................................................Thitika Thammavech and Teerasak Veerapaspong 274 

Beta-carotene, Mimosine and Quality of Leucaena Silage Kept at Different Duration 

............................ Wanna Angthong, Boonlom Cheva-Isarakul, Somkid Promma 

................................................................................ and Boonserm Cheva-Isarkul 282 

Effects of Natural Mineral Soils on Body Weight and Liver Minerals of Black Head Somali 

Sheep in Ethiopia 

.................................... Sisay Tilahun, Pravee Vijchulata, Pornsri Chairatanayuth 

.............................................................................. and Suwapong Swasdiphanich 288 

Protoplast Isolation and Culture of Aquatic Plant Cryptocoryne wendtii De Wit 

.................. Kanchanaree Pongchawee, Uthairat Na-Nakorn, Siranut Lamseejan, 

.......................................................... Supawadee Poompuang and Salak Phansiri 300

Anti HSV-1 Activity of Spirulina platensis Polysaccharide 

.... Nattayaporn Chirasuwan, Ratana Chaiklahan, Marasri Ruengjitchatchawalya 

.......................................................... Boosya Bunnag and Morakot Tanticharoen 311 

Taura Syndrome Virus Disease in Farm-Reared Penaeus monodon in Thailand 

.................................................................... Chalor Limsuwan and Niti Chuchird 319 

Optimization of Docosahexaenoic Acid (DHA) Production and Improvement of Astaxanthin 

Content in a Mutant Schizochytrium limacinum Isolated from Mangrove Forest in Thailand 

.................... Wassana Chatdumrong, Wichien Yongmanitchai, Savitree Limtong 

...................................................................... and Wanchai Worawattanamateekul 324 

Cloning, Expression, Purification and Biological Activities of Recombinant Mouse Interleukin-2 

in E. coli M15 

........... Sanchai Chantajorn, Ratchanee Hongprayoon and Thaweesak Songserm 335 

Production and Partial Characterization of Chitosanases from a Newly Isolated Bacillus cereus 

..... Sutee Wangtueai, Wanchai Worawattanamateekul, Mathana Sangjindavong, 

.............................................. Nuanphan Naranong and Sarote Sirisansaneeyakul 346 

Application of Pectin Coating in the Production of Vitamin Fortified Rice 

..................................................... Lalita Chatiyanont and Phaisan Wuttijumnong 356 

The effects of starter cultures on biogenic amine and free amino acid contents 

in Nham during Fermentation 

......... Sasithorn Limsuwan, Wonnop Visessanguan and Jirasak Kongkiattikajorn 

Product Development System in Pattern Construction System, Standard Body Measurement 

and Suitable Fitting Allowance for Thai Ladies Brand in Fashion Industry 

................................................................................. Foengfurad Mungtavesinsuk 373 

A Nonlinear Optimization Problem for Determining Safety Stocks in a Two-Stage 

Manufacturing System 

.............................................................................................Parthana Parthanadee 380 

Design and Implementation of a Framework for .NET-based Utility Computing Infrastructure 

............................................. Thanapol Rojanapanpat* and Putchong Uthayopas 394

Kasetsart J. (Nat. Sci.) 41 : 205 - 212 (2007) 

Changing in TSS, TA and Sugar Contents and Sucrose Synthase 

Activity in Ethephon-Treated ‘Pattavia’ Pineapple Fruit 

Ngarmnij Chuenboonngarm 1 , Niran Juntawong 2 *, Arunee Engkagul 3 , 

Wallop Arirob 2 and Surin Peyachoknakul 4 

ABSTRACT 

Exogenous ethylene increases endogenous ethylene which plays a crucial role on ripening in 

climacteric fruits. Although pineapple is a non-climacteric fruit, ethylene released from ethephon is 

effectively used to hasten the harvesting period. Effects from the use of a high concentration of ethephon 

on eating quality, fruit size and the reduction in harvesting period have been reported. In this paper, the 

effect of a low concentration of ethephon on pineapple fruit quality and sucrose synthase (SuSy) activity 

was investigated. Field experiment was arranged in split plot design. In the main plot, two levels of 

ethephon concentrations, i.e. 0 and 500 mg/l, were used by spraying at 110 days after forcing (DAF) 

fruits. The sub plot was harvesting time, i.e. 5 times of one-week intervals from 124 to 152 DAF. We 

found that the total soluble solid (TSS) was significantly increased in most of harvesting-treated fruits 

while the titratable acid (TA) was significantly increased at 131 DAF of harvesting-treated fruits. Only 

at 131 DAF harvesting time, the glucose content and SuSy activity of ethephon-treated fruits were 

significantly reduced and return to the control level afterward. However, ethephon had no effect on the 

fructose and sucrose contents at all harvesting times. In conclusion, fruit quality with shortening of 

harvesting time could be improved by applying 500 mg/l ethephon at 110 DAF since TSS content which 

is one of the parameter predicting eating quality of pineapple was increased without decreasing fruit 

quality. 

Key words: ‘Pattavia’ pineapple, ethephon, total soluble solid (TSS), titratable acidity (TA), sucrose 

synthase 

INTRODUCTION 

Ethephon is one of the most effective 

inflorescence forcing agents in pineapple [Ananas 

comosus L. (Merr.)] that is widely used presently 

(Bartholomew et al., 2003). Its function is to 

stimulate the respiration rate of fruit while 

chlorophyll remains in shell (Dull et al., 1967). 

Moreover, it accelerates the ripening process and 

concentrates the harvest peak (Chalermglin, 1979; 

Smith, 1991). In other non-climacteric fruit such 

as pepper, exogenous ethylene promotes and 

increases a cellulase activity (Ferrarese et al., 

1995). 

1 Bioscience Interdisciplinary Graduate Program, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

2 Department of Botany, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

3 Department of Biochemistry, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

4 Department of Genetics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: fscinrj@ku.ac.th 

Received date : 19/06/06 Accepted date : 06/10/06

206 

To achieve the high fruit quality, high 

total soluble solid (TSS) at the range of 12-14% 

and relatively low titratable acidity (TA) of citric 

acid at the range of 0.4-0.6% in pineapple flesh 

are recommended for pineapple production in 

Thailand (Thongtham, 1983). Though TSS and 

TA are eating quality prediction parameters, TSS 

is the only parameter suitable as a year-round index 

(Bartolome et al., 1995). Bartolome et al. (1996) 

found that TSS in pineapples was positively 

correlated with total sugars. Beside reflecting fruit 

quality, TA also indicates the sourness. In 

pineapples, TA is reported as citric acid, not malic 

acid. It varies primarily with fruit developmental 

stages but does not relatively respond to short-term 

environmental changes, while the malic acid varies 

with environmental changes especially the light 

(Singleton and Gortner, 1965). 

Many factors including ethephon have 

affected pineapple fruit quality (Bartholomew et 

al., 2003). An application time and the quantity 

of ethephon have influences on the quality of fruit. 

Too early application causes the reduction in size 

and weight of crown and fruit, whereas low TSS 

and high TA contents are also found (Audinay, 

1970; Chalermglin, 1979). TSS is highly 

correlated with test-panel eating quality (Smith, 

1988) and with total sugars (Bartoleme et al., 

1996). In pineapple fruits, fructose, sucrose, and 

glucose play important roles in flavor 

characteristics and are major sugars which vary 

according to the stage of fruit development. 

Sucrose content is lowest in the flesh during the 

early stage of fruit growth but rapidly increases at 

6 weeks before harvest and becomes predominant 

in mature fruit. In the early stage, glucose is 

slightly higher than fructose and remains relatively 

constant through development while fructose 

slightly increases at 2 weeks before harvest (Chen 

and Paull, 2000). The changes in total sugar 

contents are affected by the developmental stage 

of fruits, climates, and varieties (Bartoleme et al., 

1996), nevertheless the change of each sugar 

Kasetsart J. (Nat. Sci.) 41(2) 

content in ethephon-treated pineapple fruits has 

not been reported. 

In a sink organ, sugar accumulation is 

related to the presence of sucrose metabolizing 

enzymes. One of them is sucrose synthase (SuSy) 

(Taiz and Zeiger, 1998) which reversibly converts 

sucrose and UDP to fructose and UDP-glucose. 

SuSy is important in cell metabolism not only in 

sink strength (Nguyen-Quoc and Foyer, 2001) but 

also in cell wall synthesis (Nakai et al., 1999; Ruan 

et al., 2003), and starch synthesis (D / Aoust et al., 

1999). Furthermore, it accumulates sucrose in 

edible tissue of satsuma mandarin fruit (Komatsu 

et al., 2002) and saves ATP in glycolysis pathway 

(Huber and Azakawa, 1986). Chen and Paull 

(2000) reported that in pineapple fruits SuSy 

activity was higher at young stage, lower at 6 

weeks before harvest, and then constant till 

harvesting time. The change of SuSy activity in 

ethephon treated fruit has also not been reported. 

The objective of this work is to answer the question 

if ethephon could increase TSS, TA, sugar content 

and SuSy activity in pineapple fruit. 

MATERIALS AND METHODS 

Plant and fruit materials 

Field-grown ‘Pattavia’ pineapple 

[Ananas comosus L. (Merr.) cv. smooth cayenne] 

planted at Sam Praya district, Petchburi Province, 

Thailand, were used. Forcing of pineapple 

inflorescence was done in the evening of 

November 18, 2002, by spraying 50 ml of 250 

mg/l ethephon (a.i. 48% w/v) including 3% (w/v) 

urea on shoot. The experimental design used in 

this study was split plot design. Main plot was 

ethephon concentration of 0 and 500 mg/l by 

spraying 50 ml volume per fruit at the age of 110 

days after forcing (DAF). Pineapple fruit at this 

age is pointed-eyes stage 3 according to the Dole 

Company, Thailand, which is the last stage of 

pointed-eyes pineapple (immature) and thereafter 

the eyes will become flatted. Sub-plot was

harvesting time which started from 124 DAF until 

152 DAF. Three replications, 8 fruits each, were 

analyzed. 

Fruit samples were brought to laboratory 

and cut transversely into 3 sections after the size 

and weight of crowns and fruits were measured. 

Only the flesh of the middle section was used in 

this study. A half of the flesh was crushed and the 

juice was then used for determination of TSS and 

TA. The other half, sliced into small pieces, was 

used for the determination of the sugar content 

and sucrose synthase activity. These sliced fleshes 

of 8 fruits were pooled together as one of three 

replications at each harvesting time. The tissues 

were then frozen immediately in liquid nitrogen 

and stored at -80°C until use. 

Soluble sugar content 

TSS was determined from extracted juice 

using hand sugar refractometer. Soluble sugars in 

the form of sucrose, fructose and glucose were 

extracted following the method of Chen and Paull 

(2000). After extraction, the solution was filtered 

through a 0.45 mm filter, and 20 ml was injected 

and analyzed with HPLC by using a Waters 2690 

Separation Model instrumented with a Waters 410 

Differential Refractometer detector, employing a 

Sugar-PAK I (Waters Associates, Milford, USA) 

column of stainless steel (300 mm length × 6.5 

mm internal diameters). The eluting buffer was 

0.1 mM calcium EDTA and the flow rate was 0.5 

ml/min. Experiments were performed at 90°C. 

Soluble sugars were quantified by comparing the 

peak areas with external sucrose, glucose and 

fructose standard solutions (Sigma Co., Ltd.). 

Titratable acidity 

TA was analyzed from extracted juice 

after the determination of TSS contents and 

reported as citric acid according to AOAC (1990). 

Sucrose synthase determination 

Sucrose synthase (SuSy) in frozen flesh 

tissue was extracted as described by Chen and 

Kasetsart J. (Nat. Sci.) 41(2) 207 

Paull (2000). The extracted solution was desalted 

by Hitrap ® Desalting column (Amersham 

Biosciences) and 50 µl of desalted mixture was 

used to determine the enzymatic activity in 

synthesis direction according to the method of 

Hubbard et al. (1989), as modified by Chen and 

Paull (2000). 

Statistical analysis 

All data were analyzed the variance 

(ANOVA) using statistical analysis software of 

IRRISTAT version 93-3. 

RESULTS AND DISCUSSION 

The last harvesting time in this study 

(152 DAF) was planned to coincide with 

commercial harvesting time. The commercial 

harvesting index for cannery fruit industry is 

apparent when fruits reach full-size and the shell 

color at the basal portion starts to change. The 

effects of ethephon and harvesting time on fruit 

quality, sugar content and SuSy activity after 

treating at 110 DAF are shown in Table 1. 

Ethephon concentration did not reduce the size and 

weight of the crowns and fruits. The crowns and 

fruits continued to develop after the treatment and 

the crowns reached a full-size one week (138 DAF) 

before the fruits did (145 DAF). Maximum growth 

of the crowns indicated that the fruits were nearly 

ready for harvest (Paull and Reyes, 1996). The 

concentration of ethephon plays a significant role 

in increasing the mean of TSS contents (11.02° 

Brix) when compared with the mean of untreated 

fruits (8.90°Brix). The mean of TA and sugar 

contents including SuSy activity did not change, 

compared with untreated fruits. The harvesting 

time at 145 DAF provided the highest TA, TSS 

and sucrose contents of 0.62% citric acid, 12.16° 

Brix and 54.12 g/kg FW, respectively (P

208 


Table 1 Effects of ethephon concentrations and harvesting times on fruit quality, sugar content and sucrose synthase activity after treated at 110 days 

after forcing (DAF). 

Crown Fruit Flesh 

Width Length Weight Width Length Weight TA1/ TSS2/ Sucrose Glucose Fructose SuSy activity 

_____(cm)_____ (g) _____(cm)_____ (g) (%citric (°Brix) ________(g/kg FW)________ (mmole/h/ 

acid) g FW) 

Ethephon 

concentration 

0 mg/l 12.6 12.6 140.8 11.0 14.0 882.4 0.54 8.90b 31.07 14.78 11.61 2.374 

500 mg/l 11.6 11.1 127.1 11.3 14.1 922.9 0.59 11.02a 43.20 13.42 11.40 1.717 

Harvesting time 

124 DAF 11.4b 9.8b 124.8 11.5 14.2 846.2b 0.42c 8.08b 20.90d 14.46 10.86 2.297 

131 DAF 11.3b 10.4b 114.4 10.7 13.4 760.4b 0.52b 8.78b 23.46cd 13.96 10.43 1.984 

138 DAF 13.5a 13.4a 139.4 10.8 13.8 849.8b 0.58ab 8.48b 37.08bc 15.22 12.02 1.964 

145 DAF 12.2ab 13.0a 143.3 11.8 15.0 1149.8a 0.62a 12.16a 54.12a 14.50 13.08 2.127 

152 DAF 12.1ab 12.7a 147.9 11.1 13.8 906.7b 0.67a 12.30a 50.13ab 12.34 11.12 1.864 

Ethephon 

concentration ns ns ns ns ns ns ns * ns ns ns ns 

Harvesting time * ** ns ns ns ** ** ** ** ns ns ns 

Ethephon 

concentration X ns ns ns ns ns ns * * ns * ns * 

Harvesting time 

Mean followed by the same letter within the same column are not significantly different at the 5% level according to LSD. Symbols * and ** indicate significance at the 0.05 and 0.01 levels 

analyzed by DMRT, ns indicates no significant. 

1/ TA = Tritratable acidity 

2/ TSS = Total soluble solid

y Chen and Paull (2000). Figure 1 also showed 

that sucrose content was low in immature fruit and 

the highest content was achieved at 145 DAF while 

glucose and fructose contents were relatively 

constant during fruit growth as reported by Chen 

and Paull (2000). 

The interaction of ethephon 

concentration with harvesting time significantly 

affected TA, TSS and glucose contents at P

210 

were higher than that of the untreated fruits (Figure 

2B). The exogenous ethylene which was 

suggested to increase the lipoxygenase activity by 

Yu et al. (2003) might change the permeability of 

the membrane and cause the increase of TSS in 

these mature fruits. From the results on high TSS 

(13.53°Brix) and TA (0.6% citric acid) contents 

measured at 145 DAF, the treated fruits which are 

in the range of high eating-quality fruit 

(Bartholomew et al., 2003) could be harvested one 

week earlier. Chalermglin (1979) also reported 

that after applying 1,500 mg/l of ethephon at 112 

DAF, the treated fruits could be harvested 11 days 

TA (%citric acid) 

Glucose content (g/kg FW) 

0.9 

0.8 

0.7 

0.6 

0.5 

0.4 

0.3 

0.2 

0.1 

25 

20 

15 

10 

5 

0 

0 

a 

a a 

a 

b 

a 

a 

b 

a 

(A) 

a 

(C) 

a 

124 131 138 145 152 DAF 

a 

124 131 138 145 152 DAF 

a 

a 


a 

a a a 

a 

a 

earlier than those of the control. However, TA was 

found to be inereased in treated fruits while fruit 

size was reduced and TSS was unchanged. This 

study indicates that the application of 500 mg/l 

ethephon to 110 DAF fruits hastened the 

harvesting time without reducing fruit quality. 

Figure 2 also showed SuSy activities 

which were affected by a significant interaction 

between ethephon concentration and harvesting 

time. When harvested at 131 DAF, the SuSy 

activity of the treated fruits was significantly lower 

than that of the untreated fruits. Chen and Paull 

(2000) suggested that the low SuSy activity in 

Figure 2 Changes in tritratable acidity (TA) (A), total soluble solid (TSS) (B) and glucose contents (C) 

and sucrose synthase activity (D) in pineapple fruits flesh at various harvesting times after 

treated with 500 mg/l ethephon ( ) and without ethephon ( ) at 110 days after forcing 

(DAF). Error bars represent standard error of the means of three replications. Bars with the 

same letter assigned are not significantly different at 0.05 probability level. 

TSS (°Brix) 

SuSy Activity (µmole/h/g FV) 

18 

16 

14 

12 

10 

8 

6 

4 

2 

0 

4 

3.5 

3 

2.5 

2 

1.5 

1 

0.5 

0 

a 

b b 

a 

a 

a a 

124 131 138 145 152 DAF 

a 

a 

b 

(B) 

a 

124 131 138 145 152 DAF 

(D) 

a 

b 

a 

a 

a a 

b 

a 

a

pineapple fruit allowed the accumulation of 

sucrose. However, we found that the low SuSy 

activity in harvested fruits treated at 131 did not 

enhance the sucrose accumulation (no significant 

interaction of sucrose was found, Table 1). 

Therefore, the SuSy activity was not related to the 

accumulation of sucrose in pineapples which is in 

contrast to the activity in non-climacteric, satsuma 

mandarin fruits (Komatsu et al., 2002). The 

decrease of SuSy activity of harvested fruits treated 

at 131 DAF might be resulted from the increase 

in respiration rate which increases the amount of 

ATP in cells. Therefore, SuSy activity which 

involves in energy-saving pathway of glycolysis 

(Huber and Akazawa, 1986) should be decreased. 

SuSy is an important enzyme for synthesizing 

UDP-glucose, the cellulose precursor (Nakai et al., 

1999). Thus, exogenous ethylene enhances a 

cellulase activity (Ferrarese et al., 1995) which 

leads to high production of UDP-glucose that may 

act as a negative feedback to the SuSy activity. 

The exact mechanisms of the SuSy activity as well 

as the effect of ethylene on SuSy activity have still 

not been well-defined. 

CONCLUSION 

We conclude that the ethephon at the rate 

of 500 mg/l spraying at 110 DAF could increase 

TSS in pineapple fruit, but not TA, sugar contents 

and SuSy activity, and the treated fruits could be 

harvested at 145 DAF without the decrease of fruit 

size and weight. 

ACKNOWLEDGEMENTS 

The work was partially supported by 

Thesis and Dissertation Support Fund, Graduate 

School, Kasetsart University. Special thank to 

Assoc. Prof. Dr. Napavarn Noparatnaraporn for 

her suggestion in preparation of this manuscript. 


LITERATURE CITED 

AOAC. 1990. Official Methods of Analysis. Vol. 

II, 15 th ed., Association of Official Analytical 

Chemists, Inc., Arlington, Virginia. 

Audinay, A. 1970. Trial on the artificial control 

of ripening of pineapples with ethrel. Fruits 

25: 757. 

Bartholomew D.P., E. Malezieux, G.M. Sanewski 

and E. Sinclair. 2003. Inflorescence and fruit 

development and yield, pp. 167-202. In D.P. 

Bartholomew, R.E. Paull and K.G. Rohrbach 

(eds.). The Pineapple: Botany, Production 

and Uses. CABI Publishing, U.K. 

Bartolome A.P., P. Ruperez and C. Fuster. 1995. 

Pineapple fruit: morphological characteristic 

chemical composition and sensory analysis of 

Red Spanish and Smooth Cayenne cultivars. 

Food Chemistry 53: 75-79. 

Bartolome A.P., P. Ruperez and C. Fuster. 1996. 

Changes in soluble sugars of two pineapple 

fruit cultivars during frozen storage. Food 

Chemistry 56: 163-166. 

Chalermglin P. 1979. Effect of ethephon on 

maturation and some fruit characteristics 

of pineapple. MS dissertation. Kasetsart 

University, Bangkok. 

Chen, C.-C. and R.E. Paull. 2000. Sugar 

metabolism and pineapple flesh translucency. 

J. Amer. Soc. Hort. Sci. 125: 558-562. 

D / Aoust, M.-A., S. Yelle and B. Nguyen-Quoc. 

1999. Antisense inhibition of tomato fruit 

sucrose synthase decreases fruit setting and 

the sucrose unloading capacity of young fruit. 

Plant Cell 11: 2407-2418. 

Dull, G.G., R.E. Young and J.B. Biale. 1967. 

Respiratory patterns in fruit of pineapple, 

Ananas comosus detached at different stages 

of development. Physiol. Plant. 20: 1059- 

1065. 

Ferrarese, L., L. Trainotti, P. Moretto, de L.P. 

Polverino, N. Rascio and G. Casaeoro. 1995. 

Differential ethylene-inducible expression of

212 

cellulase in pepper plants. Plant Mol. Biol. 

29: 735-747. 

Holcroft, D.M. and A.A. Kader. 1999. Controlled 

atmosphere-induced changes in pH and 

organic acid metabolism may affect color of 

stored strawberry fruit. Postharvest Biol. 

Technol. 17: 19-32. 

Hubbard, N.L., S.C. Huber and D.M. Pharr. 1989. 

Sucrose phosphate synthase and acid invertase 

as determinants of sucrose concentration in 

developing muskmelon (Cucumis melo L.) 

fruits. Plant Physiol. 91: 1527-1534. 

Huber, S.C. and T. Akazawa. 1986. A novel 

sucrose synthase pathway for sucrose 

degradation in culture sycamore cells. Plant 

Physiol. 81: 1008-1013. 

Komatsu, A., T. Moriguchi, K. Koyame, M. Omura 

and T. Akihama. 2002. Analysis of sucrose 

synthase genes in citrus suggests different 

roles and phylogenetic relationships. J. Exp. 

Bot. 53: 61-71. 

Nakai, T., N. Tonouchi, T. Konishi, Y. Kojima, T. 

Tsuchida, F. Yoshinaga, F. Sakai and T. 

Hayashi. 1999. Enhancement of cellulose 

production by expression of sucrose synthase 

in Acetobacter xylinum. PNAS 96: 14-18. 

Nguyen-Quoc, B. and C.H. Foyer. 2001. A role 

for ‘futile cycles’ involving invertase and 

sucrose synthase in sucrose metabolism of 

tomato fruit. J. Exp. Bot. 52: 881-889. 

Ruan, Y.-L., D.J. Llewellyn and R.T. Furbank. 

2003. Suppression of sucrose synthase gene 


expression represses cotton fiber cell 

initiation, elongation, and seed development. 

Plant Cell 15: 952-964. 

Singleton, V.L. and W.A. Gortner. 1965. Chemical 

and physical development of pineapple fruit 

II. Carbohydrate and acid constituents. J. 

Food Sci. 30: 19-23. 

Smith, L.G. 1988. Indices of physiological 

maturity and eating quality in Smooth 

Cayenne pineapples. 2. Indices of eating 

quality. Queensl. J. Agric. Animal Sci. 45: 

219-228. 

. 1991. Effects of ethephon on ripening 

and quality of freshmarket pineapples. 

Austral. J. Exp. Agric. 31: 123-127. 

Taiz, L. and E. Zeiger. 1998. Plant Physiology, 

2 nd ed., Sinaver Associates, Inc., Publishers, 

Sunderland. 792 p. 

Thongtham, J. 1983. Pineapple and Pineapple 

Industry in Thailand. Kasetsart University 

Press, Bangkok. 

Ulrich, R. 1970. Organic acids, pp. 89-118. In 

A.C. Hulme (ed.). The Biochemistry of 

Fruits and Their Products, Vol 1. Academic 

Press, London. 

Yu, Z., C. Kun-Song, C. Qing-Jun, Z. Shang-Long 

and R. Yi-Ping. 2003. Effects of 

acetylsalicylic acid (ASA) and ethylene 

treatments on ripening and softening of 

postharvest kiwifruit. Acta Bot. Sinica 45: 

1447-1452.


Phylogenetic Analysis of Thai Amomum (Alpinioideae: Zingiberaceae) 

Using AFLP Markers 

Wittaya Kaewsri 1 *, Yingyong Paisooksantivatana 1 , Uamporn Veesommai 1 , 

Wichan Eiadthong 2 and Srunya Vajrodaya 3 

ABSTRACT 

The AFLP technique was used to assess the genetic relationships among 45 zingiberaceous 

plants including 40 collections of Amomum and 5 outgroup taxa: Alpinia, Etlingera 1, Etlingera 2, 

Elettaria and Geostachys. Cluster analysis using unweighted pair group method with arithmetic mean 

(UPGMA), based on AFLP data from 122 polymorphic bands generated with five primer combinations, 

was performed. The grouping of accessions of most species corresponded with their fruit morphological 

characteristics and were found to be consistent with previous studies. The species of Thai Amomum 

were classified into 3 groups based on AFLP markers: A. aculeatum group, A. biflorum group, and A. 

dealbatum group. The genetic relationships among genus Amomum and other genera in the tribe 

Alpinioideae are still incompletely understood. 

Key words: phylogenetic, Amomum, AFLP, Thailand 

INTRODUCTION 

Amomum Roxb. is one of the largest 

genera in the ginger family (Zingiberaceae) with 

about 150-180 species. As currently recognized, 

Amomum occurs from the Himalayas through 

Southeast Asia, Northern Australia and extends 

into the central Pacific and is widely distributed 

in Southeast Asia (Kiew, 1982; Smith, 1985). 

Sirirugsa (2001) estimated about 15-20 species to 

be found in Thailand. Plants of Amomum are 

generally evergreen herbs inhabiting wet forests 

in light gaps and at forest margins (Sakai and 

Nagamasu, 1998). Many species are used as 

medicine, spice, condiment and vegetable. Even 

though the plants from this genus have been long 

utilized, the identification is still confusing because 

of the absence of a comprehensive work on the 

genus and the much confused taxonomic problems. 

These bring about many changes in their 

taxonomic status. 

Four species of Amomum were first 

recognized by Linnaeus (1753) including: A. 

cardamomum, A. zingiber, A. zerumbet and A. 

grana-paradisi. These species have since been 

transferred to Elettaria Maton, Zingiber Boehm 

and Aframomum K. Schum. by Burtt and Smith 

(1972). Baker (1892), classified Amomum into 5 

sections; Geanthus, Achasma, Hornstedtia, 

Euamomum and Cenolophon based on 

1 Department of Horticulture, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand. 

2 Department of Forest Biology, Faculty of Forestry, Kasetsart University, Bangkok 10900, Thailand. 

3 Department of Botany, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: wittayakaewsri@yahoo.com 


214 

morphological characteristics of spike, labellum 

and anther crest. Schumann (1904) used the 

characteristics of anther crest and classified 

Amomum into 2 sections and 4 series. Section 

Geanthus was divided into 2 series, series 

Oliganthae and Polyanthae, distinguished by the 

absence of an anther crest. Section Euamomum 

was comprised of series Lobulatae and Integrae, 

characterized by an anther crest. Gagnepain (1906) 

separated Amomum into 3 groups based on the 

characteristics of floral morphology such as anther 

crest and lateral staminode. Loesener (1930) 

classified Amomum into 2 main groups using 

anther crest, Lobulatae and Integrae. 

Xia et al. (2004) investigated the 

phylogenetic status of Amomum using ITS and 

matK DNA sequence data. They indicated that 

Amomum as currently defined is polyphyletic with 

three major groups of species (A. villosum Group, 

A. tsao-ko Group and A. maximum Group) that do 

not correspond with any previously recognized 

sectional classification of the genus. They also 

mentioned that some morphological characters 

such as anther crest and fruit type could be useful 

for classification. 

The AFLP technique has been used to 

study genetic diversity and phylogenetic 

relationships in a wide range of plant species; 

Lubberstedt et al. (2000) studied relationships 

among early European maize inbreds, Garcia- 

Mass et al. (2000) used AFLP marker for 

measuring genetic diversity in melon, Abdalla et 

al. (2001) used AFLP marker for estimating 

genetic relationships across a wide range of 

taxonomic levels and for analyzing the 

evolutionary and historical development of cotton 

cultivars at the genomic level, Larson et al. (2001) 

studied AFLP variation in agamospermous and 

dioecious bluegrasses of western North America, 

Mizumoto et al. (2003) used AFLP for studying 

the diversity of nuclear and chloroplast genome 

in wild einkorn wheat (Triticum urartu). 

Because the relationships within genus 


Amomum and other genera in tribe Alpinioideae 

are still incompletely understood, a more detailed 

analysis using other molecular techniques is 

necessary. Knowledge of the genetic relationships 

among them is essential to the classification of 

the genus. This study was intended to determine 

genetic relationships among species of the 

Amomum genus occurring in Thailand using AFLP 

markers. 


Plant materials 

Forty accessions of Amomum and 5 

accessions of outgroup taxa: Alpinia, Elettaria, 

Etlingera 1, Etlingera 2 and Geostachys were used 

in this study (Table 1). All plant materials were 

grown and kept at Department of Horticulture, 

Kasetsart University, Bangkok, Thailand. 

DNA isolation and AFLP analysis 

Total genomic DNAs were extracted 

from 100 mg fresh young leaves using Qiagen 

DNeasy ® Plant Mini kit (Qiagen GmbH, Hilden, 

Germany). 

AFLP analysis was performed following 

the method of Vos et al. (1995) with minor 

modifications. From each sample, 2 templates 

were prepared by digesting 20-50 ng DNA with 

the restriction enzyme combination EcoRI-MseI 

and by ligating the corresponding oligonucleotide 

adaptors in a total volume of 10 µl. Preselective 

PCR amplification with primers corresponding to 

adaptor core sequences (E+A and M+C) was 

performed in a 10 µl reaction containing 3 µl of 

AFLP template. PCR contained 10X PCR buffer, 

0.5 µmol/L of each primer, 1 µmol/L of each dNTP, 

and 1 U Taq DNA polymerase (Fermentas, 

Lithuania) and was performed using a Biosystems 

Mod. Gene Amp ® PCR system 9700 (Biosystems, 

Montgomeryville, PA). PCR conditions consisted 

of 1 cycle of 5 min at 50°C, 1 cycle of 3 min at 

94°C, 24 cycles of 30 s at 94°C, 24 cycles of 1


Table 1 List of Thai Amomum accessions and outgroup taxa used in AFLP study. 

Accessions Species Collected number Collected places(provinces) 

1 A. aculeatum Roxb. Kaewsri-02 Kanchanaburi 

2 A. biflorum Jack Kaewsri-52 Chanthaburi 

3 A. dealbatum Roxb. Kaewsri-110 Chiang Mai 

4 A. koenigii 1 Kaewsri-03 Kanchanaburi 

5 A. koenigii 2 Kaewsri-29 Nakhon Nayok 

6 A. micranthum Ridl. Kaewsri-63 Chanthaburi 

7 A. repoense Gagnep. Kaewsri-64 Chanthaburi 

8 A. rivale1* Kaewsri-04 Kanchanaburi 

9 A. rivale2* Kaewsri-23 Kanchanaburi 

10 A. cf. rivale Kaewsri-33 Kanchanaburi 

11 A. siamense Craib Kaewsri-14 Tak 

12 A. testaceum 1 Kaewsri-15 Tak (Cultivated) 




16 Amomum cf. testaceum Kaewsri-86 Chumphon 

17 A. uliginosum1 Kaewsri-30 Nakhon Nayok 

18 A. uliginosum2 Kaewsri-92 Tak 

19 A. uliginosum3 Kaewsri-32 Trat 

20 A. cf. villosum1 Kaewsri-12 Tak (Cultivated) 

21 A. cf. villosum2 Kaewsri-13 Tak 

22 Amomum sp.1 Kaewsri-01 Kanchanaburi 


24 Amomum sp.3 Kaewsri-19 Prachuap Khiri Khan 



27 Amomum sp.6a Kaewsri-113 Chiang Mai 

28 Amomum sp.6b Kaewsri-88 Tak 

29 Amomum sp.7 Kaewsri-27 Uthai Thani 

30 Amomum sp.8 Kaewsri-35 Ranong 


32 Amomum sp.10 Kaewsri-50 Sakon Nakhon 

33 Amomum sp.11 Kaewsri-68 Chumphon 

34 Amomum sp.12 Kaewsri-70 Chumphon 


36 Amomum sp.14 Kaewsri-94 Tak 

37 Amomum sp.15 Kaewsri-108 Chiang Mai 

38 Amomum sp.16 Kaewsri-111 Chiang Mai 

39 Amomum sp.17a Kaewsri-134 Nan 

40 Amomum sp.17b Kaewsri-138 Nan 

41 Alpinia nigra - Cultivated at KU 

42 Elettaria cardamomum - Tak 

43 Etlingera littoralis - Kanchanaburi 

44 Etlingera pavieana - Chanthaburi 

45 Geostachys sp. - Nakhon Nayok 

The number 1, 2, 3 or 4 of each species = Amomum’s specimens that were collected from different places.

216 

min at 56°C, and 24 cycles of 1 min at 72°C, 

followed by an extension of 5 min at 72°C. 

Amplification products were diluted in 100 µl 

deionized H 2O and 2 µl were used for selective 

amplification in a total volume of 10 µl containing 

1 µmol/L of 10X PCR Buffer, 5 µmol/L of Especific 

primer extended by 3 selective 

neucleotides (Table 2), 5 µmol/L of M-specific 

primer extended by 3 selective nucleotides (Table 

2), 1 U of Taq DNA polymerase (Fermentas, 

Lithuania) and 1 µmol/L of each dNTPs. PCR was 

performed using a touchdown protocol with initial 

denaturation of a cycle of 30 s at 94°C, 30 s at 65° 

C (decreasing the temperature by 1°C after each 

cycle until 57°C) and 1 min at 72°C, followed by 

30 cycles of 30 s at 94°C, 30 s at 56°C and 1 min 

at 72°C with a final extension of 4 min at 72°C. 

Following amplification, 10 µl of formamide 

loading dye was added to the PCR products. The 

products were electrophoresed on 8% nondenaturing 

polyacrylamide gel. The bands were 

visualized using silver stain. 

Data analysis 

Each accession was scored (1) for 

presence and (0) for absence of each polymorphic 

band. AFLP bands within accessions were scored 

as missing if they were poorly resolved on the gel 

or if the template DNA did not amplify well. 

Similarity coefficient was calculated on the basis 

of Dice similarity coefficients (Dice, 1945) and is 


written as 

Cjk = 2a/(2a+b+c) 

In which Cjk is similarity coefficient, a 

is number of AFLP markers present in both j and 

k accessions, b is number of AFLP markers present 

only in j accessions and c is number of AFLP 

markers present only in k accessions. The 

similarity matrix was subjected to cluster analysis 

by the unweighted pair-group method with 

arithmetic mean (UPGMA) and a dendrogram was 

created using the NTSYS-pc version 2.01d 

program (Rohlf, 1997). 

RESULTS 

Five informative AFLP primer 

combinations generated a total of 364 reproducible 

amplification fragments across all species of 

Amomum, among which 122 bands were 

polymorphic (Table 2). The number of amplified 

AFLP bands per primer pair varied from 66 to 81 

with an average of 72.8 bands. The average 

number of polymorphic bands detected was 24.4 

per primer combination. The fragment sizes were 

determined by comparing each one with the 

standard DNA ladder, ranging from about 140 to 

726 base pairs (bp). Two primer combinations (E- 

AGG, M-CAA (Figure 2) and E-ACC, M-CAA) 

produced 30 polymorphic bands, a relatively 

higher numbers of polymorphisms compared to 

the other primers used in this study. 

Table 2 AFLP primer pairs and their number of amplified and polymorphic bands for phylogenetic 

study of Thai Amomum. 

Primer combinations Amplified bands No. of polymorphic bands 

(EcoRI+3/MseI+3) 

E-AGG, M-CAA 81 30 

E-ACC, M-CTA 73 17 

E-ACC, M-CAA 66 30 

E-AGC, M-CTC 74 25 

E-AGG, M-CTC 70 20 

Total 364 122 

Mean 72.8 24.4

Cluster analysis 

Cluster analysis using UPGMA 

(unweighted pair group method with arithmetic 

mean) was performed to examine genetic 

relationships among Thai Amomum species. A 

dendrogram was produced from the UPGMA 

cluster analysis of genetic similarity (GS) matrix 

for 45 accessions, 40 accessions of Amomum 

species and 5 accessions of out taxa, based on 

AFLP markers varied from 43% to 88% with a 

total average genetic similarity of 74.5% (Table 

3). Two main clusters (A and B) were separated 

at 57% GS. The A cluster was separated into 2 

groups: C and D, at 58% genetic similarity. The D 

group is subdivided into 2 subgroups (I and II) at 

59% GS while the B Cluster generated 2 groups 

(E and F) at 59% GS (Figure 3). 

The A cluster is characterized by spiny 

fruit (rarely smooth fruit). The C group contains 

Amomum koenigii 1, A. koenigii 2, A.uliginosum 

1, Amomum sp.9, Etlingera littoralis, A. 

aculeatum, Amomum sp.12 and Geostachys sp. 

while D group is divided into two subgroups (I 

and II). Subgroup (I) consists of A. testaceum 1, 

A. testaceum 2, A. testaceum 3 and Amomum cf. 

testaceum. Subgroup (II) consists of A. testaceum 

4, A. cf. villosum2, Amomum sp.5, Amomum sp.7, 

A. uliginosum2, A. cf. villosum1, A. rivale1, A. 

micranthum, A. rivale2, A. cf.rivale, Amomum sp. 

8, Amomum sp.4, A. biflorum and Amomum sp.13. 

The B cluster is characterized by smooth, ridged 

or wing fruit (rarely spiny fruit). This cluster 

contains E and F groups. The E group consists 

Amomum sp.16, Amomum sp.3, Amomum sp.2, 

Amomum sp.17a, A. siamense, Amomum sp.6b, 

Amomum sp.6a, Elettaria cardamomum, Amomum 

sp.17b, A.dealbatum, Amomum sp.15, Amomum 

sp.10 and Alpinia nigra. The F group contains 

Amomum sp.1, Amomum sp.14, A. repoense, and 

Etlingera pavieana. Regarding the out group taxa; 

Alpinia nigra, Elettaria cardamomum and 

Etlingera pavieana were inserted in B group while 

Etlingera littoralis and Geostachys sp. were placed 

in A group. 


DISCUSSION 

In this study, 40 accessions of Thai 

Amomum species were fingerprinted including 5 

outgroup taxa. One hundred twenty two 

polymorphic AFLP markers were produced from 

five primer combinations. UPGMA cluster 

analysis (Rohlf, 1997) with genetic similarity of 

57% separated Amomum into 2 main clusters: A 

consists of C and D groups and B consists of E 

and F groups (Figure 1). 

Regarding the C group, A. koenigii 1 and 

A.koenigii 2 were collected from Kanchanaburi 

and Nakhon Nayok provinces, respectively. It is 

clear that both collections are closely related 

(74%), even though the peduncular lengths vary 

greatly. The plants from Nakhon Nayok have a 

much shorter peduncle than those found in 

Kanchanaburi. The variation in phenotype could 

be caused by differences in their respective 

habitats. The placement of this species is similar 

to morphological analysis that placed it in spiny 

fruit group. This result is confirmation of the 

paraphyletic relationship between A.koenigii and 

the spiny fruit species (A. uliginosum and A. 

aculeatum). A.aculeatum and Amomum sp.12 are 

placed together at 90% GS. These closely related 

species are similar in leafy stem but differ in 

peduncular length, colour and size of labellum. 

From the results, the species Amomum sp.12 

should be established as a new variety. However, 

this is difficult to decide from only a single plant. 

More collections are needed to solve this problem. 

D group is divided into two subgroups (I 

and II). Subgroup D (I) consists of A. testaceum 

1, A. testaceum 2, A. testaceum 3 and Amomum 

cf. testaceum. Regarding A. testaceum species 

complex, the dendrogram suggests that this species 

can be separated into at least three varieties; 

especially A. testaceum 4 which was isolated from 

the group. The placement of A. testaceum is rather 

close to the spiny fruit species (Amomum sp.1 and 

Amomum sp.14). This result does agree with Xia 

et al. (2004) whose work was based on ITS and

218 

MatK genes. They placed A.testaceum among the 

spiny fruit species of A.villosum group. A possible 

explanation for this was a paraphyletic origin of 

A. testaceum complex. Although its morphological 

characteristics are different, its genotype is close 


to spiny fruit species. The Amomum cf. testaceum 

that was collected from Chumphon is also placed 

in this group. Its leafy stem is similar to A. 

testaceum but dif fers in its hairiness on the lower 

surface of leaves. 

Figure 1 Some species of Amomum used in AFLP study. 

A. A. aculeatum Roxb. B. A. biflorum Jack C. A. dealbatum Roxb. 

D. A.koenigii Gmelin. E. A. repoense Pierre ex Gagnep. F. A. rivale Ridl. 

G. A. testaceum Ridl. H. A. uliginosum K?nig ex Retz. I. A. siamense Craib


Figure 2 AFLP fingerprint of Thai Amomum species and out-groups using E-AGG, M-CAA primer pair. 1. A. koenigii1 , 2. A. koenigii 2, 3. Amomum 

sp.16, 4. A. testaceum1, 5. A. testaceum 2, 6. A. testaceum 3, 7. A. testaceum 4, 8. Amomum sp.1, 9. A. aculeatum Roxb., 10. Amomum sp.12, 

11. A. rivale1, 12. A. rivale2, 13. A.cf. villosum1, 14. A. cf. villosum2, 15. Amomum sp.4, 16. Amomum sp.5, 17. Amomum sp.7, 18. A. 

uliginosum1, 19. A. uliginosum2, 20. A. uliginosum3, 21. A Amomum cf. rivale., 22. Amomum sp.17a, 23. Amomum sp.8, 24. Amomum sp.10, 

25. A. biflorum Jack , 26. A. micranthum Ridl., 27. Amomum sp.11, 28. Amomum sp.13, 29. none use, 30. none use, 31. A. siamense Craib,32. 

.Amomum sp.3, 33. Amomum sp.2, 34. Amomum sp.6b, 35. A. uliginosum4, 36. Amomum sp.17b, 37. A. repoense Gagnep., 38. Amomum sp.6a, 

39. none use, 40. Amomum sp.14, 41. Amomum sp.15, 42. A. dealbatum Roxb., 43. Elettaria cardamomum, 44. Etlingera littoralis, 45. 

Etlingera pavieana, 46. Alpinia nigra, 47. Geostachys sp., 48.Amomum cf. testaceum and M=φXHinfI

220 

Table 3 Dice’s coefficient of similarity matrix from AFLP fingerprints of 40 accessions of Amomom and 5 outgroup taxa. 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 

1 A. koenigii 1 1.00 

2 A. koenigii 2 0.71 1.00 

3 Amomum sp. 16 0.58 0.57 1.00 

4 A. testaceum 1 0.56 0.54 0.45 1.00 

5 A. testaceum 2 0.56 0.51 0.47 0.78 1.00 

6 A. testaceum 3 0.57 0.57 0.51 0.78 0.88 1.00 

7 A. testaceum 4 0.50 0.55 0.47 0.59 0.62 0.64 1.00 


8 Amomum sp. 1 0.64 0.60 0.60 0.50 0.53 0.57 0.60 1.00 

9 A. aculeatum 0.59 0.65 0.60 0.66 0.63 0.64 0.56 0.61 1.00 

10 Amomum sp. 12 0.58 0.63 0.57 0.64 0.62 0.60 0.60 0.64 0.90 1.00 

11 A. rivale 1 0.63 0.63 0.61 0.65 0.67 0.68 0.61 0.68 0.65 0.61 1.00 

12 A. rivale 2 0.57 0.59 0.65 0.57 0.55 0.63 0.65 0.67 0.60 0.59 0.76 1.00 

13 A. cf. villosum 1 0.54 0.56 0.51 0.57 0.52 0.52 0.64 0.52 0.61 0.57 0.65 0.60 1.00 

14 A. cf. villosum 2 0.50 0.53 0.43 0.53 0.53 0.53 0.73 0.53 0.58 0.59 0.56 0.64 0.75 1.00 

15 Amomum sp. 4 0.61 0.63 0.58 0.68 0.68 0.71 0.64 0.60 0.67 0.64 0.80 0.74 0.71 0.62 1.00 

16 Amomum sp. 5 0.57 0.60 0.53 0.60 0.58 0.60 0.65 0.57 0.66 0.65 0.64 0.67 0.77 0.83 0.68 1.00 

17 Amomum sp. 7 0.53 0.55 0.52 0.62 0.68 0.64 0.69 0.59 0.60 0.66 0.67 0.67 0.65 0.70 0.75 0.71 1.00 

18 A. uliginosum 1 0.67 0.83 0.61 0.57 0.56 0.59 0.53 0.60 0.68 0.64 0.63 0.60 0.60 0.51 0.63 0.64 0.55 1.00 

19 A. uliginosum 2 0.53 0.57 0.47 0.65 0.68 0.65 0.69 0.57 0.62 0.64 0.64 0.67 0.71 0.74 0.77 0.74 0.94 0.58 1.00 

20 A. uliginosum 3 0.53 0.52 0.57 0.57 0.56 0.54 0.63 0.62 0.56 0.58 0.67 0.64 0.67 0.64 0.69 0.65 0.75 0.58 0.74 1.00 

21 Amomum cf.rivale 0.55 0.57 0.69 0.59 0.57 0.62 0.63 0.64 0.62 0.58 0.72 0.85 0.62 0.64 0.78 0.68 0.66 0.61 0.67 0.67 1.00 

22 A.momum sp.17a 0.53 0.52 0.58 0.53 0.51 0.48 0.53 0.56 0.60 0.60 0.63 0.56 0.60 0.48 0.60 0.53 0.60 0.50 0.57 0.58 0.60 1.00

Table 3 (Continued) 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 

23 Amomum sp. 8 0.55 0.57 0.66 0.57 0.57 0.59 0.61 0.62 0.59 0.55 0.77 0.85 0.65 0.62 0.77 0.67 0.63 0.61 0.66 0.69 0.92 0.60 1.00 

24 Amomum sp. 10 0.55 0.50 0.63 0.54 0.53 0.48 0.57 0.59 0.57 0.58 0.60 0.57 0.60 0.53 0.60 0.51 0.53 0.58 0.58 0.61 0.63 0.61 0.66 

25 A. biflorum Jack 0.58 0.58 0.61 0.65 0.65 0.68 0.61 0.65 0.60 0.61 0.71 0.71 0.59 0.56 0.69 0.64 0.67 0.64 0.67 0.67 0.71 0.55 0.71 

26 A. micranthum Ridl. 0.53 0.54 0.62 0.61 0.60 0.63 0.71 0.63 0.63 0.62 0.73 0.75 0.63 0.63 0.64 0.61 0.64 0.57 0.64 0.67 0.76 0.56 0.74 

27 Amomum sp. 11 0.50 0.54 0.51 0.64 0.69 0.64 0.71 0.60 0.55 0.57 0.62 0.67 0.61 0.67 0.68 0.71 0.81 0.59 0.81 0.73 0.70 0.53 0.65 

28 Amomum sp. 13 0.54 0.67 0.65 0.67 0.61 0.63 0.60 0.60 0.66 0.64 0.73 0.75 0.60 0.60 0.71 0.61 0.68 0.64 0.68 0.70 0.78 0.56 0.78 

29 A. siamense Craib 0.54 0.62 0.57 0.58 0.60 0.57 0.54 0.55 0.60 0.64 0.70 0.57 0.58 0.55 0.62 0.58 0.59 0.64 0.59 0.64 0.60 0.60 0.65 


30 Amomum sp. 3 0.52 0.53 0.61 0.53 0.51 0.53 0.55 0.64 0.56 0.52 0.63 0.51 0.51 0.48 0.60 0.47 0.57 0.55 0.55 0.61 0.53 0.52 0.57 

31 Amomum sp. 2 0.50 0.53 0.66 0.51 0.51 0.50 0.57 0.67 0.51 0.49 0.60 0.56 0.48 0.47 0.57 0.43 0.53 0.53 0.50 0.58 0.57 0.50 0.55 

32 Amomum sp. 6b 0.58 0.60 0.58 0.54 0.53 0.51 0.53 0.60 0.57 0.58 0.67 0.60 0.57 0.60 0.63 0.60 0.60 0.57 0.58 0.64 0.63 0.64 0.64 

33 A. uliginosum 4 0.59 0.67 0.65 0.58 0.58 0.58 0.60 0.57 0.63 0.62 0.65 0.64 0.60 0.58 0.67 0.64 0.62 0.74 0.62 0.62 0.68 0.57 0.67 

34 Amomum sp.17b 0.57 0.56 0.60 0.57 0.63 0.58 0.62 0.69 0.63 0.60 0.68 0.64 0.63 0.52 0.65 0.60 0.60 0.62 0.62 0.59 0.64 0.64 0.67 

35 A. repoense Gagnep. 0.53 0.52 0.58 0.50 0.54 0.54 0.55 0.70 0.56 0.52 0.58 0.59 0.57 0.54 0.57 0.51 0.52 0.52 0.55 0.63 0.63 0.64 0.63 

36 Amomum sp. 6a 0.59 0.54 0.60 0.50 0.47 0.47 0.45 0.55 0.60 0.57 0.59 0.53 0.50 0.43 0.53 0.49 0.56 0.51 0.54 0.51 0.54 0.60 0.54 

37 Amomum sp. 14 0.53 0.53 0.57 0.55 0.57 0.53 0.64 0.67 0.57 0.60 0.65 0.66 0.57 0.61 0.57 0.61 0.59 0.56 0.57 0.60 0.64 0.48 0.62 

38 Amomum sp. 15 0.57 0.54 0.56 0.57 0.58 0.55 0.56 0.57 0.61 0.57 0.70 0.66 0.66 0.55 0.65 0.58 0.54 0.54 0.59 0.62 0.67 0.68 0.70 

39 A. dealbatum Roxb. 0.61 0.64 0.57 0.56 0.57 0.62 0.52 0.70 0.60 0.53 0.63 0.60 0.64 0.54 0.67 0.57 0.55 0.64 0.60 0.60 0.66 0.60 0.67 

40 Elettaria cardamomum 0.65 0.56 0.60 0.44 0.52 0.53 0.51 0.74 0.57 0.62 0.62 0.58 0.53 0.50 0.51 0.50 0.56 0.57 0.54 0.59 0.53 0.62 0.56 

41 Etlingera littoralis 0.69 0.63 0.64 0.56 0.56 0.59 0.52 0.60 0.59 0.60 0.60 0.62 0.60 0.57 0.60 0.64 0.58 0.71 0.58 0.60 0.63 0.57 0.64 

42 Etlingera pavieana 0.58 0.58 0.57 0.54 0.57 0.54 0.57 0.71 0.64 0.63 0.67 0.54 0.51 0.47 0.61 0.50 0.58 0.55 0.58 0.61 0.55 0.61 0.53 

43 Alpinia nigra 0.52 0.55 0.55 0.53 0.53 0.50 0.60 0.56 0.56 0.50 0.61 0.50 0.62 0.56 0.61 0.56 0.55 0.52 0.60 0.58 0.55 0.49 0.58 

44 Geostachys sp. 0.57 0.60 0.59 0.55 0.57 0.53 0.60 0.64 0.69 0.65 0.64 0.57 0.61 0.58 0.65 0.61 0.59 0.67 0.62 0.65 0.65 0.60 0.62 

45 Amomum cf. testaceum 0.53 0.49 0.61 0.70 0.70 0.68 0.58 0.57 0.60 0.60 0.63 0.67 0.47 0.53 0.63 0.56 0.53 0.49 0.53 0.55 0.67 0.57 0.66

222 

Table 3 (Continued) 

24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 

24 Amomum sp. 10 1.00 

25 A. biflorum Jack 0.60 1.00 

26 A. micranthum Ridl. 0.62 0.67 1.00 

27 Amomum sp. 11 0.57 0.73 0.63 1.00 

28 Amomum sp. 13 0.64 0.74 0.75 0.64 1.00 

29 A. siamense Craib 0.65 0.57 0.58 0.57 0.64 1.00 

30 Amomum sp. 3 0.69 0.58 0.57 0.51 0.64 0.60 1.00 


31 Amomum sp. 2 0.66 0.58 0.57 0.53 0.60 0.57 0.88 1.00 

32 Amomum sp. 6b 0.69 0.55 0.60 0.53 0.65 0.82 0.63 0.58 1.00 

33 A. uliginosum 4 0.59 0.65 0.66 0.57 0.74 0.64 0.60 0.57 0.62 1.00 

34 Amomum sp. 17b 0.70 0.70 0.58 0.67 0.63 0.63 0.62 0.64 0.65 0.52 1.00 

35 A. repoense Gagnep. 0.71 0.61 0.68 0.57 0.60 0.54 0.61 0.58 0.60 0.54 0.62 1.00 

36 Amomum sp. 6a 0.62 0.56 0.57 0.52 0.60 0.67 0.57 0.57 0.71 0.52 0.66 0.56 1.00 

37 Amomum sp. 14 0.54 0.70 0.66 0.61 0.64 0.52 0.60 0.64 0.47 0.61 0.60 0.59 0.43 1.00 

38 Amomum sp. 15 0.67 0.65 0.67 0.57 0.64 0.60 0.54 0.53 0.65 0.58 0.72 0.64 0.69 0.55 1.00 

39 A. dealbatum Roxb. 0.64 0.63 0.54 0.56 0.59 0.59 0.61 0.63 0.64 0.59 0.74 0.66 0.60 0.57 0.70 1.00 

40 Elettaria cardamomum 0.60 0.57 0.63 0.49 0.53 0.66 0.59 0.59 0.68 0.55 0.61 0.67 0.69 0.53 0.61 0.59 1.00 

41 Etlingera littoralis 0.55 0.67 0.54 0.57 0.59 0.62 0.55 0.57 0.60 0.71 0.62 0.58 0.53 0.60 0.53 0.69 0.60 1.00 

42 Etlingera pavieana 0.64 0.63 0.59 0.56 0.59 0.60 0.63 0.63 0.57 0.54 0.67 0.69 0.59 0.64 0.60 0.63 0.57 0.52 1.00 

43 Alpinia nigra 0.67 0.55 0.56 0.54 0.59 0.56 0.61 0.57 0.58 0.54 0.54 0.61 0.54 0.54 0.53 0.60 0.51 0.49 0.64 1.00 

44 Geostachys sp. 0.70 0.64 0.64 0.57 0.64 0.67 0.65 0.62 0.70 0.63 0.71 0.62 0.57 0.60 0.64 0.62 0.64 0.54 0.64 0.57 1.00 

45 Amomum cf. testaceum 0.60 0.64 0.65 0.62 0.65 0.51 0.53 0.60 0.55 0.54 0.59 0.63 0.53 0.54 0.65 0.57 0.53 0.55 0.57 0.44 0.54 1.00


Figure 3 Dendrogram depicting the genetic relationship of 45 accessions of Amomum based on AFLP 

fingerprint, using similarity coefficient by DICE, clustering with UPGMA.

224 

Subgroup D (II) consists of A. testaceum 

4, A. cf. villosum, Amomum sp.5, Amomum sp.7, 

A. uliginosum 2, Amomum sp.11, A. uliginosum 3, 

A. villosum 1, A. rivale 1, A. micranthum, A. rivale 

2, A. cf. rivale, Amomum sp.8, Amomum sp.4 and 

Amomum sp.13. All members have spiny fruit and 

leafy stem less than 1.50 m tall. Regarding 

uliginosum 2 and 3 which were collected from Tak 

province, they were separated from A. uliginosum 

1 and 4 (from Nakhon Nayok and Ranong 

provinces, respectively). Their morphological 

characteristics differ from the ones in C group in 

its shorter leafy stem and smaller inflorescence. 

A possible explanation for this is that their 

morphological characteristics were the result of 

long time adaptation in the surrounding habitats 

which resulted in two ecotypes of A. uliginosum. 

B cluster consists of E and F groups. It 

is characterized by smooth, ridged or winged fruit 

(rarely spiny fruit). 

E and F groups include Amomum sp.16, 

Amomum sp.3, Amomum sp.2, Amomum sp.17a, 

A. siamense, Amomum sp.6b, Amomum sp.6a, 

Amomum sp. 17b, A. dealbatum, Amomum sp. 15, 

Amomum sp. 10, Elettaria cardamomum, Alpinia 

nigra, Amomum sp. 1, Amomum sp.14, A. repoense 

and Etlingera pavieana. 

The dendrogram suggests the placement 

of smooth fruit (Amomum sp.16) between spiny 

and winged fruit. Similar to the result of Amomum 

sp.1 and 10 both of which are spiny fruit but were 

placed among winged fruit species. Amomum 

sp.17a and 17b from Nan province are similar in 

their morphology but were placed in different 

clusters. More study is needed to properly identify 

the position of these species. A.siamense with fruit 

longitudinally ridged is also placed in this group. 

This species should be closely related to winged 

fruit species. Although the cluster is not completely 

separated from the others, all winged fruit species 

are clearly placed. Therefore, the results have the 

tendency to be consistent with the A.maximum 

group of Xia et al. (2004). 


The outgroup taxa (Alpinia, Elettaria, 

Etlingera and Geostachys) are placed among 

Amomum species. The result indicates a closer 

relationship among them and the spiny fruit species 

of Amomum. This result is similar to Xia et al. 

(2004) who found that Etlingera littoralis was 

placed in the clade of A.villosum group. The results 

then confirmed that Etlingera is related to the 

genus Amomum. Furthermore, some species of 

Alpinia, Elettaria and Geostachys are also closely 

related to the genus Amomum. 

Twenty-six representives of Thai 

Amomum can be classified into 3 groups by using 

AFLP evidence: A. aculeatum, A. biflorum and A. 

dealbatum groups. 

The A. aculeatum group consists of 4 

species: A. koenigii, A. uliginosum, A. aculeatum 

and Amomum sp. 12. Species in this group have 

smooth and spotted or spiny fruit, anther crest 3 

lobes, leafy stem stout and usually more than 1.5 

m tall. 

The A. biflorum group contains 10 

species: A. testaceum, Amomum cf. villosum, 


Amomum sp.8, A.rivale, A. micranthum, Amomum 

sp.11 and Amomum sp.13. All members of this 

group are defined by smooth or spiny fruit. Most 

species of this group are spiny fruit. In the case of 

smooth fruit, its fruit shape is usually globular and 

fruit colour is white or pale brown. The leafy stem 

is usually slender and shorter than 1.5 m. 

The A. dealbatum group contains 12 

species of Amomum: A. dealbatum, A. repoense, 

A. siamense, Amomum sp.1, Amomum sp.2, 


Amomum sp.14, Amomum sp.15, Amomum sp.16 

and Amomum sp.17. The species in this group are 

characterized by winged, ridged or smooth fruit 

(rarely spiny fruit and 3 lobes) and entire, round 

or truncate anther crest.

CONCLUSION 

AFLP markers classified Thai Amomum 

species into three groups (A. aculeatum group, A. 

biflorum group, and A. dealbatum group) which 

correspond to the fruit and leafy stem 

characteristics. 


The authors are thankful to the curators 

of Bangkok Herbarium (BK) and Forest 

Herbarium (BKF) for their kind permission and 

suggestion during this study. Also, this work was 

supported by the TRF/BIOTEC Special Program 

for Biodiversity Research and Training grant 

T_14009. 


Abdalla, A.M, O.U.K. Reddy and K.M. El-Zik. 

2001. Genetic diversity and relationships of 

diploid and tetraploid cottons revealed using 

AFLP. Theor. Appl. Genet. 102: 222-229. 

Baker, J.G. 1892. Scitamineae, pp. 198-264. In 

J. D. Hooker. Flora of British India vol.VI. 

L. Reeve & Co., London. 

Burtt, B.L. and R.M. Smith. 1972. Key species in 

the taxonomic history of Zingiberaceae. Note 

RBG. Edinb. 31: 177-227. 

Dice, L.R. 1945. Measures of the amount of 

ecological association between species. 

Ecology 26: 297-302. 

Gagnepain, F. 1906. Du Muséum. Bulletin de La 

Societé Botanique de France. 53: 136-145. 

Garcia-Mass, J., M. Oliver and H. Gomez- 

Paniagua. 2000. Comparing AFLP, RAPD and 

RFLP markers for measuring genetic diversity 

in melon. Theor. Appl. Genet. 101: 860-864. 

Kiew, K. Y. 1982. The genus Elettariopsis 

(Zingiberaceae) in Malaya. Notes RBG 

Edinb. 42: 295-314. 

Larson, S.R., B.L. Waldron, S.B. Monsen, L.St. 


John, A.J. Palazzo, C.L. McCracken and R.D. 

Harrison. 2001. AFLP Variation in 

agamospermous and dioecious bluegrasses of 

Western North America. Crop Sci. 41: 1300- 

1305. 

Linnaeus, C. 1753. Monandria. In Species 

Plantarum-A facsimile of the first edition. 

London, Bernard Quaritch Ltd. 560 p. 

Loesener, T. 1930. Zingiberaceae (Amomum), 

pp.599-602. In A. Engler and K. Prantl, ed. 

Die Naturlichen Pflanzenfamilien, Leipzig, 

E. Haberland. 

Lubberstedt, T., A. E. Melchinger, C. DuBle, M. 

Vuylsteke and M. Kuiper. 2000. Relationships 

among early European maize inbreds: IV. 

Genetic diversity revealed with AFLP markers 

and comparison with RFLP, RAPD, and 

pedigree data. Crop Sci. 40: 783-791. 

Mizumoto, K., S. Hirosawa, C. Nakamura and S. 

Takumi. 2003. Nuclear and chloroplast 

genome genetic diversity in the wild einkorn 

wheat, Triticum urartu, revealed by AFLP and 

SSLP analyses. Hereditas 137: 208-214. 

Rohlf, F.J. 1997. NTSYS-pc 2.01d: Numerical 

taxonomy and multivariate analysis 

system, version 2.01. Exeter Software, 

Setauket, NY. 

Sakai, S. and H. Nagamasu. 1998. Systematic 

studies of Bornean Zingiberaceae I. Amomum 

in Lambir Hills, Sarawak. Edinb. J. Bot. 

55(1): 45-64. 

Schumann, K. 1904. Zingiberaceae, pp. 1-458. 

In A. Engler, ed. Das Pflanzenreich, Heft 20 

(IV, 46) Leipzig. 

Sirirugsa, P. 2001. Zingiberaceae of Thailand. 

Pp. 63-77. In V. Baimai and R. Kumhom. 

BRT Research Reports 2001. Biodiversity 

Research and Training Program. Jirawat 

Express Co.,Ltd., Bangkok. 

Smith, R.M. 1985. A review of Bornean 

Zingiberaceae:1(Alpineae). Notes RBG 

Edinb. 42: 295-314. 

Vos, P., R. Hoger, M.Bleeker, M.Reijans, T. van

226 

de Lee, M. Hornes, A. Frijters, J. Pot, J. 

Peleman, M. Kuiper and M. Zabeau. 1995. 

AFLP: a new technique for DNA 

fingerprinting. Nucleic Acids Research 

23(21): 4407-4414. 


Xia, Y.M., W.J. Kress and L.M. Prince. 2004. 

Phylogenetic analyses of Amomum 

(Alpinioideae: Zingiberaceae) using ITS and 

matK DNA sequence data. Systematic 

Botany 29(2): 334-344.


Prediction of Sweet Corn Seeds Field Emergence under 

Wet Soil Condition 

Vichai Wongvarodom* and Wikanate Rangsikansong 

ABSTRACT 

Field emergence prediction of sweet corn seeds under wet soil conditions was studied using 

three different quality seeds of Hawaiian Sugar Super Sweet and Super Sweet Argo Extra MT varieties. 

The seeds were subjected to germinate in sand at room temperature. Germination was evaluated for 7 

days after planting (DAP). Flooded germination was done by planting the seeds in 1000 g of clay soil in 

a plastic basket, flooding at 1 cm above soil level for 5 hours and evaluating at 7 DAP. Field emergence 

was studied under daily watering, three times a day, to simulate wet planting soil condition. Field 

emergence evaluation was performed at 7 DAP. Results showed that the sweet corn seeds with 79.50- 

91.00% germination of Hawaiian Sugar Super Sweet and Super Sweet Argo Extra MT varieties had 

field emergence of 61-79% under wet soil condition. The lower quality seed had the field emergence of 

lower than 60%. Sand germination and flooded germination in clay soil did not correspond to field 

emergence under wet soil condition. The field emergence under wet soil condition of sweet corn seeds 

could be predicted by the polynomial equations which gave better results than sand germination test. 

Key words: field emergence, wet soil condition, sand germination, flooded germination, sweet corn 

seeds 

INTRODUCTION 

Germination test is an analytical 

procedure to evaluate seed viability under 

standardized (favorable) laboratory conditions 

(ISTA, 1999; AOSA, 2001). The percentage of 

germination reflects the planting value of a seed 

lot (Liu et al., 1999). However, it was frequently 

found that standard germination did not correspond 

to the field performance under stress planting 

conditions (Delouche and Baskin, 1973; Vieira et 

al., 1999). Sand germination test in room 

temperature for corn seeds has been used widely 

in many of the developing countries, due to simple 

and low cost test, and using less specific 

equipment. Most importantly, result of the test has 

to be highly accurate with standard germination 

(Dungpatra, 1986). Many crop production areas, 

including Thailand, are faced with heavy raining 

in the planting season which resulted in wet soil 

condition in the planting field (Martin et al., 1988; 

Jittham, 2002). Some vigor tests have been 

develop for more accurate prediction for field 

emergence under the stressful planting condition 

in corn (Sawatdikarn, 2002), sweet corn (Jittham, 

2002) and cucumber (Werakul, 2003) in the humid 

Division of Agricultural Technology, Department of Industry and Technology, Faculty of Science and Technology, Prince of 

Songkla University, Muang, Pattani 94000, Thailand 

* Corresponding author, e-mail: seedinter@yahoo.com 


228 

tropics. The purpose of this study was to 

investigate the relationship between sand 

germination and field emergence and flooded 

germination in predicting potential to field 

emergence under wet planting soil condition. 


Two varieties, namely Hawaiian Sugar 

Super Sweet and Super Sweet Argo Extra MT 

commercial corn seeds obtained from Songkhla 

Field Crop Research Center and a seed company 

were used as high quality seeds with germination 

of 86.50-91.00%. The seed samples having 

different germination percentage (74.50-79.50 and 

53.50-66.00%) after accelerated aging at 42°C 

(AOSA, 2002) for 48 and 96 hours were used as 

medium and low quality seeds, respectively. All 

tests were done with four replications. 

Sand germination 

Fifty seeds per replication were subjected 

to germinate in 1,000 g sand in plastic basket at 

room temperature and were watered daily. First 

and final counts were done at 4 and 7 days after 

planting, respectively (AOSA, 2001). Normal 

seedlings were averaged as the germination 

percentage. 

Flooded germination test 

Fifty seeds per replication were subjected 

to germinate in 1,000 g of clay soil in plastic basket 

at room temperature. The planting baskets were 

placed in plastic trays and were flooded at 1 cm 

above soil level for 5 hours. After the end of 

flooding duration, the water was drained (the soil 

moisture content was still near to saturation after 

drainage) and the seeds were placed for further 

germinating. The germination percentage was 

evaluated 7 days after planting (Jittham, 2002). 

Field emergence under wet soil condition 

Fifty seeds per replication were planted 


at a depth of 2.5 cm in clay soil of the experimental 

field of Division of Agricultural Technology, 

Prince of Songkla University, Pattani. Irrigation 

was given daily for three times a day, morning, 

noon, and evening, to simulate wet planting field 

condition. Also in the each irrigation time, the soil 

was watered till wet. The normal seedlings were 

counted at 4 and 7 days after planting, respectively. 

Field emergence percentage was calculated using 

the same procedure as described in AOSA (2001). 

Analysis of variances for a completely 

randomized design among sand germination, field 

emergence, and flooded germination was 

performed. The statistical significance of 

differences among means was tested by Duncan , s 

Multiple Range Test (DMRT). The relation 

between sand germination and field emergence 

were plotted as well as mathematical equations 

for predicting the field emergence which are also 

presented as polynomial. 


Comparison of the sweet corn seed 

germination among sand, flooded condition, and 

wet field planting condition was undertaken (Table 

1). The sweet corn seeds with 79.50-91.00% 

germination of Hawaiian Sugar Super Sweet and 

Super Sweet Argo Extra M.T. varieties had field 

emergence of 61-79%. The lower quality seeds 

had the field emergence of lower than 60%. 

Seeds of Hawaiian Sugar Super Sweet 

and Super Sweet Argo Extra MT varieties in sand 

test showed significant higher germination 

percentages than field emergence under wet 

planting field condition. The seeds germinated in 

soil in baskets under the flooded condition gave 

lower germination percentage than those in both 

sand test and under the field condition. This is not 

in agreement with the earlier report by Jittham 

(2002) that the flooded germination gave the same 

germination percentage as field emergence in rainy 

season planting. This is probably due to the

difference of soil texture used in the flooded 

germination test causing different results. The high 

amount of water holding after drainage in clay soil 

used in this study might cause more reduction of 

oxygen diffusion and become more compact 

during the germination period comparing to sandy 


loam soil, and as a consequence sweet corn seed 

germinability dramatically reduced. 

Sand germination showed a significant 

correlation with field emergence both in Hawaiian 

Sugar Super Sweet corn (r=0.694*) (Figure 1) and 

Super Sweet Argo Extra MT corn (r=0.919**) 

Table 1 Germination of Hawaiian Sugar Super Sweet and Super Sweet Argo Extra MT corn seeds 

with three quality classes tested in sand, flooded condition and under wet field condition. 

Test methods and Germination (%) 

field conditions High Medium Low 

Hawaiian Sugar Super Sweet 

Sand 91.00 A 74.50 A 66.00 A 

Flooded condition 13.00 C 11.00 C 7.50 C 

Wet field condition 60.50 B 55.00 B 46.50 B 

F-test ** ** ** 

C.V.(%) 12.73 12.31 14.05 

Super Sweet Argo Extra MT 

Sand 86.50 A 79.50 A 53.50 A 

Flooded condition 48.00 C 27.00 B 29.00 C 

Wet field condition 79.00 B 77.00 A 41.50 B 

F-test ** ** ** 

C.V.(%) 6.47 12.75 17.26 

** = significant at P

230 

(Figure 2). This is in agreement with the previous 

report by Kulik and Schoen (1982) that emergence 

of sweet corn seeds in sand bench was highly 

correlated with field emergence. However, sand 

germination could not predict seedling emergence 

of sweet corn under wet soil condition or in rainy 

planting season as data shown in Table 1. The 

Field emergence under wet soil 

condition (%) 

100 

80 

60 

40 

20 

0 


y = -0.0231x 2 + 4.2883x - 119.05 

R 2 = 0.8443 

0 20 40 60 80 100 

Sand germination (%) 

Figure 2 Relation between sand germination and field emergence under wet soil condition of Super 

Sweet Argo Extra MT corn seeds, r=0.919**. 

Germination difference (%) 

3 

2.5 

2 

1.5 

1 

0.5 

0 

-0.5 

-1 

-1.5 

HSSS SAEMT 

results of this study showed that percentage of field 

emergence calculated using the mathematical 

equation, polynomial, is very closely to field 

emergence (Figure 3). The data suggested that in 

sweet corn, the calculated field emergence is 

superior to sand germination and flooded 

germination tests in predicting field emergence 

Figure 3 Germination differences between the predicted field emergence and field emergence under 

wet soil condition of different quality seeds of Hawaiian Sugar Super Sweet (HSSS) and 

Super Argo Extra MT (SAEMT) corn. 

High 

Medium 

Low

under wet soil condition planting. 

Additional evaluations of other varieties 

and lots as well as hybrid varieties are needed to 

confirm the present results and to investigate more 

optimum mathematical equation which could be 

widely used in most sweet corn. Future study 

should also be conducted to relate field emergence 

results to drought planting condition. 

CONCLUSION 

1. Sweet corn seeds with 79.50-91.00% 

germination of Hawaiian Sugar Super Sweet and 

Super Sweet Argo Extra MT varieties had field 

emergence of 61-79% under wet soil condition. 

The lower quality seed had the field emergence of 

lower than 60%. 

2. Sand germination and flooded 

germination in clay soil did not correspond to field 

emergence under wet soil condition. 

3. The field emergence under wet soil 

condition of sweet corn seeds could be predicted 

by the polynomial equations obtained from 

relationship between sand germination and the 

field emergence which gave better results than 

sand germination test. 


Thanks are due to the Faculty of Science 

and Technology, Prince of Songkla University, for 

financing this research. We wish to thank the 

Songkhla Field Crop Research Center for 

supplying the seeds. 


Association of Official Seed Analysts. 2001. Rules 

for Testing Seeds. The Association of Official 

Seed Analysts. 126 p. 


Association of Official Seed Analysts. 2002. Seed 

Vigor Testing Handbook. Contribution 

No.32 to the Handbook on Seed Testing. 105 

p. 

Delouche, J. C. and C. C. Baskin. 1973. 

Accelerated aging techniques for predicting 

the relative storability of seed lots. Seed Sci. 

& Technol. 1: 427-452. 

Dungpatra, J. 1986. Seed Testing and Analysis. 

Agri Book Group. Bangkok. 194 p. (In Thai). 

International Seed Testing Association. 1999. 

International Rules for Seed Testing. Seed Sci. 

& Technol. 27: Supplement. 

Jittham, O. 2002. Germination test under water 

stress conditions for sweet corn seed vigor 

evaluation. MS. Thesis. Prince of Songkla 

University. Songkhla. 

Kulik, M.M. and J.F. Schoen. 1982. Germination, 

vigor and field emergence of sweet corn seeds 

infected by Fusarium moniliforme. Seed Sci. 

& Technol. 10: 595-604. 

Liu, H., L. O. Copeland, O. Schabenberger and 

D. Jamieson. 1999. Variability of germination 

tests of corn and soybeans. J. Seed Technol. 

21: 25-33. 

Martin, B.A., O.S. Smith and M.O. Neil. 1988. 

Relationships between laboratory germination 

tests and field emergence of maize inbreds. 

Crop Sci. 28: 801-805. 

Sawatdikarn, S. 2002. Germination test of corn 

seed under water stress conditions. MS. 

Thesis. Prince of Songkla University. 

Songkhla. 

Vieira, R. D., J. A. Paiva-Aguero, D. Perecin and 

S. R. M. Bittencourt. 1999. Correlation of 

electrical conductivity and other vigor tests 

with field emergence of soybean seedlings. 

Seed Sci. & Technol. 27: 67-75. 

Werakul, S. 2003. Germination test under water 

stress conditions for cucumber seed vigor 

evaluation. MS. Thesis. Prince of Songkla 

University. Songkhla.


Modifying Controlled Deterioration for Evaluating Field 

Weathering Resistance of Soybean 

Ye Changrong 1,2 , Prapa Sripichitt 2 *, Sunanta Juntakool 2 , 

Vipa Hongtrakul 3 and Arom Sripichitt 4 

ABSTRACT 

To develop practical methods for testing field weathering resistance of soybean varieties, pods 

and seeds from CM60 (susceptible) and GC10981 (resistant) were tested by seven treatments. Among 

the treatments, modified incubator weathering (yellow pods were incubated at 30°C under 90-100% 

relative humidity for 7 days) and the controlled deterioration (dry seeds were soaked in distilled water 

for 60 minutes and then incubated at 41°C under 90-100% relative humidity for 3 days) showed widerange 

differences in seed germination and viability between CM60 and GC10981. These two treatments 

were then tested on 11 soybean varieties comparing with a field weathering treatment. The germination 

of seeds treated by controlled deterioration was highly correlated to the germination of seeds subjected 

to field weathering treatment (r=0.964**, n=11). The viability of seeds submitted to both incubator 

weathering and controlled deterioration were also correlated to the viability of seeds exposed to field 

weathering (r=0.697* and 0.716*, n=11). The modified incubator weathering and controlled deterioration 

methods were further used to evaluate the field weathering resistance of 139 F 2 progenies derived from 

the cross CM60/GC10981. There was a significant correlation between the incubator weathering and 

the controlled deterioration by considering the germination and viability of seeds (germination r=0.331**, 

viability r=0.425**, n=139). Both the modified incubator weathering and controlled deterioration were 

efficient for evaluating the field weathering resistance of soybean varieties. Particularly, controlled 

deterioration method was found to be a useful way for evaluating the field weathering resistance of 

soybean seeds. 

Key words: Glycine max (L.) Merr., field weathering resistance, incubator weathering, controlled 

deterioration 

INTRODUCTION 

Soybean [Glycine max (L.) Merrill] is 

one of the world’s leading sources of vegetable 

oil and plant protein. As the world demand for 

vegetable oil and protein meal continues to 

increase, soybean production has spread rapidly 

from the temperate zone into the hot and humid 

1 Present address: School of Land and Food Science, The University of Queensland, Brisbane, Queeensland, Australia. 

2 Department of Agronomy, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand. 


4 Department of Plant Production Technology, Faculty of Agricultural Technology, King Mongkut’s Institute of Technology 

Ladkrabang, Bangkok 10520, Thailand. 

* Corresponding author: e-mail: agrprs@ku.ac.th 


tropics. Following the expansion, weather 

conditions have become the major factor affecting 

the soybean seed quality and production in tropical 

and subtropical regions (Tekrony et al., 1980). 

Weather conditions (mainly high temperature and 

relative humidity) during the post-maturation and 

pre-harvest period increase the difficulty in 

producing soybean seed with high quality in the 

tropics. This obstacle in producing good quality 

seed is also the most important factor that limits a 

distribution of soybean production in the tropics. 

Soybean seed attains its highest vigor, 

viability and potential quality at physiological 

maturity (maximum seed dry weight). However, 

due to high moisture content at physiological 

maturity (about 55%), the seed cannot be harvested 

commercially at this stage and must remain on the 

plant through a desiccation period till the seed 

reaches a harvestable moisture level. This period 

may vary from a few days to over 3 weeks 

depending on the environmental conditions in the 

field. The seeds deteriorate rapidly during this 

period (Delouche, 1980). Deterioration of seed 

vigor, as well as viability, due to high temperature 

and high relative humidity between the stages of 

seed physiological maturity and harvesting is 

referred to as field weathering (Tekrony et al., 

1980). Improving the field weathering resistance 

of new varieties is an important objective for 

soybean breeding programs in the tropics. 

To evaluate the field weathering 

resistance of soybean varieties, various methods 

have been developed (Kueneman, 1982; Dassou 

and Kueneman, 1984; Horlings et al., 1994). 

Among these methods, delayed harvest and 

incubator weathering have been widely used for 

field weathering evaluation. However, it is difficult 

to unify the maturation time of different varieties 

to make them suffer the same weather damage by 

delayed harvest (Kueneman, 1982). Dassou and 

Kueneman (1984) compared three weathering 

treatments and concluded that the incubator 

weathering treatment (incubated at 30°C under 


90-95% relative humidity for 10 days) minimized 

intraplant variability and environmental effects 

among genotypes with different maturity. The 

incubator weathering method promised a practical 

screening procedure for identification of resistance 

to field weathering of soybean seeds and has been 

widely used since then. However, the incubator 

conditions are conducive to the rapid growth of 

pathogens which is likely to encourage 

deterioration (Balducchi and McGee, 1987). 

Horlings et al. (1994) modified the treatment to 

incubate the pods at 27°C under 90-100% relative 

humidity for 4 days and indicated that the modified 

incubator treatment had the most detrimental effect 

on seed quality. But this treatment is probably too 

gentle since the temperature in tropical soybean 

fields is normally hotter and the duration is longer. 

The field weathering resistance of 

soybean is usually evaluated by the germination 

and vigor of the weathered seeds. Controlled 

deterioration method has been widely used to 

evaluate seed vigor and viability of seeds 

(Matthews, 1980; Powell and Matthews, 1981), 

but the relationship between field weathering and 

controlled deterioration has not been studied. The 

purpose of this study is to evaluate the possibility 

and efficiency of using controlled deterioration 

treatment for evaluating the field weathering 

resistance of soybean seeds. 



Twelve soybean varieties/lines, namely 

Chiangmai 60 (CM60), Yodson, TGX814-26D, 

Kalitor, 9520-21, 9519-1, Jakapan-1, Lee, CM 

9501-3-17, MK-35, SSR 8502-14-1 and GC10981, 

and 139 F 2 progenies from the cross CM60 / 

GC10981 were used in this study. Soybean CM60 

and GC10981 were employed as susceptible and 

resistant control for field weathering resistance, 

respectively (Kaowanant, 2003).

234 

Methods 

Experiment A 

Soybean CM60 and GC10981 were 

grown in a field at the National Corn and Sorghum 

Research Center, Nakhon Ratchasima Province. 

Water, fertilizer, pesticide and fungicide were 

applied when necessary. The yellow pods at 

physiological maturity were harvested for field 

weathering test. The field weathering resistance 

was evaluated by seven treatments as follows. 

1. Incubator weathering: Fresh yellow 

pods (36 pods for each treatment) were placed 

upright in the cells of a grid to avoid pod contact, 

and then sealed in a plastic box with 1 cm water 

under the grid to ensure a high relative humidity 

(90-100%) during the incubation. The boxes with 

pods inside were incubated by three treatments as 

follows: 

• 30°C for 10 days 



After the incubation, the pods were dried, 

threshed and the seeds were used for germinating 

evaluation. 

2. Accelerated ageing test and 

controlled deterioration: Fresh yellow pods were 

dried and threshed. The seeds (50 seeds for each 

treatment) were subjected to the following four 

treatments: 

• Seeds were put in a wire-mesh tray. 

The trays were then sealed in a plastic box with 1 

cm water under the trays to ensure a high relative 

humidity (90-100%) during the incubation. The 

boxes with seeds inside were incubated at 41°C 

for 3 days (standard AA test). 

• Seeds were soaked in distilled water 

for 15 minutes, and then put in wire-mesh tray 

and sealed in a plastic box with 1 cm water under 

the trays. The boxes with seeds inside were 

incubated at 41°C for 7 days. 












After the treatment, the seeds from each 

treatment and 50 non-treatment seeds (control) 

were germinated between wet papers at 25°C for 

5 days. The normal seedlings, abnormal seedlings, 

fresh ungerminated seeds, hard seeds and dead 

seeds were counted (AOSA, 2000). The field 

weathering resistance of the variety was evaluated 

by germination (percentage of normal seedlings 

and hard seeds) and viability (percentage of normal 

seedlings, abnormal seedlings, fresh ungerminated 

and hard seeds) of the treated seeds. 

Experiment B 

Soybean seeds of CM60, Yodson, 

TGX814-26D, Kalitor, 9520-21, 9519-1, Jakapan- 

1, Lee, CM 9501-3-17, MK-35, and SSR 8502- 

14-1 were grown in a greenhouse at the 

Department of Agronomy, Kasetsart University, 

Bangkok. Water, fertilizer, pesticide and fungicide 

were applied when necessary. At physiological 

mature stage, the yellow pods were subjected to 

the following treatments: 

1. Control (no treatment): The yellow 

pods were harvested, dried and threshed, and then 

50 seeds of each variety/line were germinated and 

investigated as described in experiment A. 

2. Field weathering: The yellow pods 

were left on the plant (green and brown pods were 

cut out) for 2 weeks with water spraying twice a 

day. Then the pods were harvested, dried and 

threshed. Fifty seeds of each variety/line were 

germinated and investigated as described in 

experiment A. 

3. Incubator weathering: The yellow 

pods (36 pods for each variety/line) were harvested 

and placed upright in the cells of a grid, and then

sealed in a plastic box with 1 cm water under the 

grid. The boxes with pods inside were incubated 

at 30°C for 7 days. After the treatment, the pods 

were dried and threshed, and fifty seeds of each 

variety/line were germinated and investigated as 

described in experiment A. 

4. Controlled deterioration: The 

yellow pods were harvested, dried and threshed. 

After measuring the moisture content (MC, wet 

weight basis) of the seeds, fifty seeds from each 

variety/line were weighed and soaked in distilled 

water for 60 minutes, and then the seeds were 

quickly dried by tissue paper, weighed again and 

put in a wire-mesh tray. The trays with seeds inside 

were sealed in a plastic box with 1 cm water under 

the trays to ensure a high relative humidity (90- 

100%) during the incubation. The boxes were then 

incubated at 41°C for 3 days. After the treatment, 

the seeds were weighed, germinated and 

investigated as described in experiment A. 

Experiment C 

Soybean seeds of CM60, GC10981 and 

their F 2 progenies were grown in a field at the 

National Corn and Sorghum Research Center, 

Nakorn Rachasima Province. Water, fertilizer, 

pesticide and fungicide were applied when 

necessary. The pods at physiological maturity were 

harvested from each plant for field weathering 

evaluation by incubator weathering and controlled 

deterioration test: 

1. Incubator weathering: Fresh yellow 

pods (18 pods for each progeny) were placed 

upright in the cells of a grid, and then sealed in a 

plastic box with 1 cm water under the grid. The 

boxes with pods inside were incubated at 30°C 

for 7 days. After the treatment, the pods were dried 

and threshed, and then the seeds were germinated 

and investigated as described in experiment A. 

2. Controlled deterioration: The 

yellow pods were harvested, dried and threshed. 

Twenty-five seeds from each progeny were soaked 

in distilled water for 60 minutes, and then the seeds 


were put in a wire-mesh tray. The trays with seeds 

inside were sealed in a plastic box with 1 cm water 

under the trays. The boxes were then incubated at 

41°C for 3 days. After the treatment, the seeds 

were germinated and investigated as described in 

experiment A. 

Data analysis 

The frequency distribution, paired t-test 

and Pearson correlation test of germination and 

viability data were carried out following the 

procedure of Komez and Komez (1984). 

RESULTS 

The efficiency of different treatments 

To determine the optimum treatment for 

evaluating the field weathering resistance of 

soybean varieties, the seed quality of CM60 and 

GC10981 was evaluated by seven treatments. The 

germination and viability of soybean seeds of 

CM60 and GC10981 after being subjected to seven 

treatments (experiment A) are shown in Table 1. 

For incubator weathering, after the pods were 

incubated, the germination of CM60 and GC10981 

seeds were 0-9.2% and 0-42.9%, and the viability 

of CM60 and GC10981 seeds were 19.5-41.5% 

and 27.3-72.5%. The most serious seed 

deterioration was caused by incubating the pods 

at 30°C for 10 days due to a serious pathogen 

infection. The seeds of both varieties/lines had lost 

their germinability (0% germination). Decrease in 

both treating temperature and time could increase 

the germination and viability of the treated seeds. 

After incubating the pods at 30°C for 7 days, the 

difference between CM60 and GC10981 in 

germination (33.7%) and viability (31.0%) were 

greater than the other two treatments. Therefore, 

this treatment was considered to be more efficient 

to distinguish the seed weathering of CM60 and 

GC10981 than the other two treatments. 

For the standard accelerated aging (AA) 

test, the germination of CM60 and GC10981 were

236 

decreased from 92% to 60% and from 94% to 84%, 

respectively, but the viability was only decreased 

slightly from 100% to 92% for CM60 comparing 

to the control. For controlled deterioration 

treatments, soaking the seeds in distilled water 

prior to incubation further decreased the 

germination and viability of the treated seeds. The 

lowest germination and viability were caused by 

soaking the seeds in distilled water for 15 minutes 

and incubating at 41°C for 7 days. Soaking the 

seeds in distilled water for 60 minutes and then 

incubating at 41°C for 3 days showed the widest 

difference in seed germination (38.0%) and 

viability (17.0%) between CM60 and GC10981. 

Thus this treatment was considered to be an 

optimum one to distinguish the seed weathering 

of CM60 and GC10981. 

If ignore the fault treatment (incubator 

weathering at 30°C for 10 days), by comparing 

the means of the other 6 treatments (paired t-test), 

there was significant difference between CM60 

and GC10981 (t 5=5.825, p=0.002 for germination 

and t 5=4.083, p=0.01 for viability). It was clear 

that the field weathering resistance of CM60 and 

GC10981 was significantly different. 

Relationship among field weathering, incubator 

weathering and controlled deterioration 


Table 1 The germination and viability of the soybean seeds of CM60 and GC10981 after being subjected 

to various treatments. 

Treatment 1 Germination (%) 2 Viability (%) 

CM GC Dif. CM GC Dif. 

IW: 30°C, 10 days 0.0 0.0 0.0 19.5 27.3 7.8 

IW: 35°C, 7 days 6.6 15.2 8.6 35.6 46.8 11.2 

IW: 30°C, 7 days 9.2 42.9 33.7 41.5 72.5 31.0 

CD: Water 15 min. + 41°C, 7 days 8.0 32.0 24.0 52.0 60.0 8.0 



AA: 41°C, 3 days 60.0 84.0 24.0 92.0 100.0 8.0 

Control (no treatment) 92.0 94.0 2.0 100.0 100.0 0.0 

1 IW= incubator weathering, CD= controlled deterioration, AA= accelerated ageing 

2 CM= CM60, GC= GC10981, Dif.=difference= GC-CM. 

To confirm the possibility of using the 

modified controlled deterioration method for 


soybean seed, field weathering (delayed harvest 

with water spraying) along with incubator 

weathering and controlled deterioration were 

carried out on 11 soybean varieties/lines. The seed 

characters and the changes in moisture content 

during controlled deterioration treatments are 

shown in Table 2. The seed moisture content of 

all the varieties/lines increased (3.12 to 31.1%) 

after soaking in distilled water for 60 minutes. The 

moisture content continued to increase during 

incubation. After soaking and incubation, the seed 

moisture content increased from 15.72 to 31.25% 

depending on the variety/line. The final moisture 

content of all the treated seeds varied from 23.45 

to 37.95%. Different varieties absorbed water and 

moisture at different speeds. The seed moisture 

content after soaking was highly correlated to the 

final moisture content (r=0.951**, n=11) and the 

final moisture content increase (r= 0.950**, n=11). 

The increase in moisture content after soaking was 

also correlated to the final moisture content 

increase (r=0.952**, n=11). The treated seeds that 

absorbed water faster also showed the higher 

increases in moisture content finally. There was a 

significant negative correlation between the

moisture content increment after soaking and the 

moisture content increase after incubation (r= 

-0.958**, n=11). The treated seeds that absorbed 

more water during soaking absorbed less moisture 

during incubation. In contrast, the treated seeds 

that absorbed less water during soaking absorbed 

more moisture during incubation. 


The germination, viability and 

hardseedness of 11 soybean varieties/lines after 

being subjected to 3 different treatments (field 

weathering, incubator weathering and controlled 

deterioration) are shown in Table 3. The 

germination and viability of seeds decreased after 

being subjected to all the three treatments 

Table 2 The seed characters and changes in moisture content during controlled deterioration treatment 

of 11 soybean varieties/lines. 

Variety/line Seed 100 seeds MC* before MC after MC after 

color weight (g) soaking (%) soaking (%) incubation (%) 

(increase) (increase) 

CM60 Yellow 15.11 7.04 30.33 (23.29) 31.63 (24.59) 

Yodson Black 12.13 7.20 15.47 ( 8.27) 26.76 (19.57) 

TGX-814-26D Yellow 9.47 7.16 22.29 (15.13) 30.83 (23.67) 

Kalitor Black 7.47 7.73 10.85 ( 3.12) 23.45 (15.72) 

9520-21 Yellow 15.74 7.02 32.51 (25.49) 34.09 (27.07) 

9519-1 Yellow 11.73 6.70 37.80 (31.10) 37.95 (31.25) 

Jakapan-1 Yellow 11.84 7.05 15.95 (8.91) 25.78 (18.73) 

Lee Yellow 8.62 7.21 33.64 (26.43) 35.20 (27.99) 

CM9501-3-17 Yellow 8.44 7.18 11.87 ( 4.69) 28.09 (20.91) 

MK35 Yellow 10.25 7.40 21.66 (14.26) 29.94 (22.54) 

SSR8502-14-1 Black 13.28 7.13 13.90 ( 6.77) 24.84 (17.71) 

* MC = moisture content 

Table 3 The germination, viability and hardseedness of 11 soybean varieties/lines after being subjected 

to 3 different treatments. 

Variety/line Germination (%)* Viability (%) Hard seeds (%) 

CK FW IW CD CK FW IW CD CK FW IW CD 

CM60 86 44 30 50 90 70 46 56 4 12 0 0 

Yodson 96 66 56 82 98 86 78 86 6 10 6 4 

TGX-814-26D 90 62 82 76 96 78 98 88 0 0 0 0 

Kalitor 98 76 68 96 100 90 94 100 24 52 12 34 

9520-21 96 58 80 72 98 86 96 90 0 0 0 0 

9519-1 96 54 56 64 98 84 82 68 0 0 0 0 

Jakapan-1 98 56 58 70 98 76 64 80 22 26 0 8 

Lee 90 56 68 66 92 80 94 86 0 0 0 0 

CM9501-3-17 94 72 64 90 96 88 78 92 14 38 0 8 

MK35 92 50 42 68 96 74 62 82 0 0 0 0 

SSR8502-14-1 90 62 78 84 100 86 94 96 4 14 2 4 

* CK = check (no treatment), FW = field weathering, IW = incubator weathering, CD = controlled deterioration.

238 

comparing to the control. There was a significant 

correlation between the germination of seeds 

treated by field weathering and by controlled 

deterioration (r=0.964**, n=11), as well as the 

viability of seeds subjected to these two treatments 

(r=0.716*, n=11). The correlation between the 

germination of seeds treated by field weathering 

and by incubator weathering was not obvious. 

However, there was a significant correlation 

between the viability of seeds treated by field 

weathering and by incubator weathering 

(r=0.697*, n=11). To a certain extent, the 

controlled deterioration treatment was more 

efficient for indicating the field weathering 

resistance of soybean than the incubator 

weathering treatment. The correlation between the 

germination of seeds treated by incubator 

weathering and by controlled deterioration was not 

obvious. There was only a correlation between the 

viability of seeds treated by incubator weathering 

and by controlled deterioration (r=0.739**, n=11). 

All the tested varieties/lines showed higher 

germination and viability than CM60. Among 

these varieties/lines, Kalitor, a variety with black 

seed coat and high percentage of hardseedness, 

showed high germination and viability in every 

treatment. It is a useful resource for future breeding 

programs of soybean field weathering resistance. 

Application of incubator weathering and 

controlled deterioration on F 2 progenies 

The field weathering resistance of 139 

F 2 plants was investigated by incubator weathering 

and controlled deterioration treatment. For 

incubator weathering treatment, the seed 

germination of the F 2 progenies ranged from 21.3 

to 81.6%, whereas those of their parents, CM60 

and GC10981, were 34.7% and 75%, respectively. 

The viability of the F 2 progenies varied from 47.8 

to 95.6%, whereas those of CM60 and GC10981 

were 59.7% and 93.4%, respectively. For 

controlled deterioration test, the germination of 

the F 2 progenies extended from 20 to 82%, 


whereas those of CM60 and GC10981 were 32% 

and 72%, respectively. The viability of the F 2 

progenies ranged from 44 to 90%, whereas those 

of CM60 and GC10981 were 54% and 94%, 

respectively. The distribution of the germination 

and viability of the F 2 progenies are shown in 

Figure 1. The germination and viability of the 

seeds under both treatments showed normal 

distribution (skewness < ± 0.5, kurtosis < ± 0.1). 

There was a significant correlation between the 

incubator weathering and the controlled 

deterioration as assessed by germination 

(r=0.331**, n=139) and viability (r=0.425**, 

n=139). Both treatments could be used for 


soybean seeds. 

DISCUSSION 

Since the field weathering occurs under 

the hot and humid conditions in the field after the 

seeds are physiologically mature, the most 

common procedure for evaluating seed resistance 

to field weathering is to leave the plants in the 

field beyond the normal harvest period and then 

assess the quality of the seed by visual score, 

examining seed-borne fungi, seed vigor, or use a 

combination of these assessment methods. This 

delayed harvest technique for evaluating the field 

weathering has several limitations, for example, 

genotypes matured at different times are subjected 

to different environmental weathering stresses and 

different periods of weathering. It is difficult to 

apply the same environmental stress conditions to 

cultivars of different maturities by delayed harvest. 

In an attempt to overcome the limitations of 

delayed harvest, Kueneman (1982) developed 

spreader row and overhead irrigation techniques 

to accelerate weathering based on the delayed 

harvest method and found the cultivar differences 

were highly significant. Artificial seed weathering 

methods, such as incubator weathering, can 

minimize the effects of variable pod maturity. In

Frequency 

Frequency 

30 

20 

10 

0 

20.0 

40 

30 

20 

10 

0 

50.0 

30.0 

40.0 

IW germination 

CM60 GC10981 

CM60 

60.0 

50.0 

60.0 

Germination (%) 

IW viability 

70.0 

Viability (%) 

80.0 

70.0 


90.0 

80.0 

GC10981 

0 

20.0 

Germination (%) 

Figure 1 The distribution of seed germination and viability of the F 2 progenies after being subjected 

to incubator weathering (IW) and controlled deterioration (CD) test. The skewness and 

kurtosis for each distribution are as follows: IW germination (0.079, -0.523), CD germination 

(-0.319, -0.742), IW viability (0.199, 0.656), CD viability (-0.395, 0.721). 

experiment A of this study, three incubator 

weathering treatments were carried out to identify 

the difference between susceptible variety CM60 

and resistant variety GC10981. After the fresh 

yellow pods were incubated at 30°C and 90-100% 

relative humidity for 10 days, a serious pathogen 

infection occurred, and some seeds even 

germinated during the incubation. The remaining 

seeds had lost their germinating ability (0% 

germination) and showed a very low viability 

(19.5% and 27.3%). Increasing the temperature 

and shortening the incubating time (35°C, 7 days) 

Frequency 

Frequency 

30 

20 

10 

40 

30 

20 

10 

0 

40.0 

CM60 

30.0 

CM60 

50.0 

CD germination 

40.0 

60.0 

50.0 

CD viability 

70.0 

Viability (%) 

60.0 

GC10981 

70.0 

GC10981 

80.0 

80.0 

90.0 

reduced the pathogen infection and germination 

during incubation, but the treatment still caused 

serious damage to the seeds. This treatment had 

been successfully used to identify the field 

weathering difference between CM60 and 

GC10981 by Kaowanant (2003). However, the 

constant temperature at 35°C practically does not 

occur in the soybean field. By reducing the 

temperature (30°C, 7 days), the pathogen growth 

and seed germination during incubation were 

controlled. The results showed obvious differences 

in germination and viability between CM60 and

240 

GC10981 which could be used to identify the field 

weathering resistance of these varieties. 

On the other hand, since the ability of 

seed coat to absorb moisture from the environment 

is a decisive factor in field weathering, the faster 

the seed absorbs moisture from the environment, 

the more serious the weathering that occurs. Thus, 

the controlled deterioration method developed by 

Matthews (1980) was modified to evaluate the 

seed weathering. The modified treatments 

emphasized the relationship between seed 

moisture absorbing speed and seed weathering. 

The original controlled deterioration method was 

modified into three combinations of water soaking 

time and incubating time to compare with the 

standard accelerated aging test. Soaking the seeds 

in distilled water for 60 minutes and incubating at 

41°C under 90-100% relative humidity for 3 days 

showed a wide-ranging difference in germination 

and viability between CM60 and GC10981. Since 

the difference in field weathering resistance of 

these two soybean varieties had been stated by 

Kaowanant (2003), the treatments which showed 

more difference between these varieties should be 

more efficient for distinguishing the field 

weathering resistance of soybeans. Thus, this 

treatment was considered to be efficient for testing 

field weathering resistance of soybean. 

The efficiency of the modified incubator 

weathering and controlled deterioration were 

further confirmed on 11 soybean varieties/lines in 

experiment B. Highly significant correlation was 

found between the germination of seeds treated 

by field weathering and by controlled deterioration 

(r=0.964**, n=11), as well as the viability of seeds 

subjected to these two treatments (r=0.716*, 

n=11). It is possible to use this controlled 

deterioration method to predict the field 

weathering resistance of soybean varieties. There 

was highly significant negative correlation 

between the water absorbing speed and the final 

germination of the treated seeds (r=-0.785**, 

n=11), it was confirmed that the seeds absorbing 


water faster would suffer more serious 

deterioration during the incubation resulting in a 

lower percentage of germination. The incubator 

weathering and controlled deterioration methods 

were further applied to evaluate the field 

weathering resistance of F 2 progenies derived from 

the cross CM60/GC10981. The germination and 

viability of seeds subjected to both treatments were 

continuous with a normal distribution. There was 

highly significant correlation between incubator 

weathering and controlled deterioration by 

considering germination and viability of the seeds 

(germination r=0.331**, viability r=0.425**, 

n=139). Both incubator weathering and controlled 

deterioration may be used to determine the field 

weathering resistance of soybean varieties. 

However, it is difficult to treat a great number of 

pods at the same time in incubator weathering test 

due to the laboratory limitations, especially in 

large-scale breeding programs. Controlled 

deterioration method makes it possible to harvest 

the pods at physiological maturity, dry to a similar 

moisture content level, and then store for testing. 

This will be very beneficial to large-scale 

screening in breeding programs. 

CONCLUSION 

The modified controlled deterioration 

(soaking seeds in distilled water for 60 minutes 

and incubating at 41°C under 90-100% relative 

humidity for 3 days) was confirmed to be useful 

for evaluating field weathering resistance of 

soybean seeds based on the hypothesis of a 

correlation between the water absorbing speed and 

field weathering resistance of seeds, especially for 

large-scale soybean breeding programs that focus 

on seed quality. 


This study was partly supported by the 

Graduate Research Scholarship from the Graduate

School, Kasetsart University. The authors also 

would like to express thanks Mr. Adrian Hillman 

for editing the English grammar of this paper and 

to Miss Sirikwan Sawatsitung, Miss Peeraya 

Thanarog, Miss Phan Thi Thanh, and Miss 

Supaporn Dechkrong for their helps in the 

experimental field. 


Association of Official Seed Analysts. 2000. Rules 

for testing seeds. Proc. Assoc. Off. Seed Anal. 

60(2): 1-39. 

Balducchi, A.J. and D.C. McGee. 1987. 

Environmental factors influencing infection 

of soybean seed by Phomopsis and Diaporthe 

species during seed maturation. Plant Disease 

71: 209-212. 

Dassou, S. and E.A. Kueneman. 1984. Screening 

methodology for resistance to field weathering 

in soybean seed. Crop Sci. 24: 774-779. 

Delouche, J.C. 1980. Environmental effects on 

seed development and seed quality. 

HortScience 15(6): 775-780. 

Horlings, G.P., E.E. Gamble and S. 

Shanmugasundaram. 1994. Weathering of 

soybean in the tropics as affected by seed 

characteristics and reproductive development. 

Trop. Agric. (Trinidad): 71(2): 110-115. 


Kaowanant, R. 2003. Varietal Differences of 

Soybean in Quality and Physical 

Characteristics of Seeds in Resistance to 

Field Weathering. M.S. thesis. King 

Mongkut’s Institute of Technology 

Ladkrabang. Bangkok. 

Komez, K.A. and A.R. Komez. 1984. Statistical 

Procedures for Agriucltural Research. 2 nd 

ed. International Rice Research Institute. 

John Willey & Sons, Inc., New York. 

Kueneman, E.A. 1982. Genetic differences in 

soybean seed quality screening methods for 

cultivar improvement, pp 31-41. In J.B. 

Sinclair and J.A. Jackobs (eds.). Soybean 

Seed Quality and Stand Establishment. 

International Agriculture Publications. 

University of Illinois, Urbanna-Champaign, 

IL. 

Matthews, S. 1980. Controlled deterioration: A 

new vigor test for crop seeds, pp 647-660. In 

P.D. Hebblethwaite (ed.). Seed Production 

Butterworths, London. 

Powell, A.A. and S. Matthews. 1981. Evaluation 

of controlled deterioration, a new vigor test 

for crop seeds. Seed Sci. and Tech. 9: 633- 

640. 

Tekrony, D.M., D.B. Egli and A.D. Phillips. 1980. 

Effect of field weathering on the viability and 

vigor of soybean seed. Agron. J. 72: 749-755.


Composite Line Method for the Development of Early Generation 

Hybrids of Maize (Zea mays L.) 

Nguyen Phuong, Krisda Samphantharak* and Vatcharee Lertmongkol 

ABSTRACT 

Six commercial single crosses were used for the improvement of composite and inbred lines. 

Modified S 1-full sib selection was applied to improve the three sister line composite. Lines were visually 

selected under low-competition environment in honeycomb arrangement with equilateral triangular 

side of 0.866 m. Testcross as well as diallel cross were applied to identify high combining lines. All 

yield trials were conducted in randomized completed block design with 4 replications, 1 row plot of 5 m 

long and 0.75 × 0.25 m plant spacing. Standard cultural practices were regulated and irrigation was 

applied as needed. 

Statistically, there was no clear advantage of yield between composite and inbred lines in early 

generation testcrosses. Besides, the diallel sets of both groups of lines gave similar results. However, 

the top hybrids of overall trials came from composite crosses even though it was not significant. In 

addition, composite lines were superior to S 3 lines in yield, earliness and plant height. Modified S 1-full 

sib selection is a flexible breeding method but its merit for the construction of early generation hybrids 

must be thoroughly investigated even though the positive results were observed. 

Key words: maize breeding, testcross, honeycomb, composite line 

INTRODUCTION 

Development of single cross hybrid of 

maize is the ultimate goal of most of maize 

breeding programs. However, finding stable high 

yield inbred lines to ensure the high level of 

economic return for commercial hybrid seed 

production is the main obstacle of small and new 

emerged single cross development programs. 

Combined line selection and testing for combining 

ability is time and space consuming processes. 

Instead of five or six generations of selfing usually 

practiced in the development of inbred lines, 

composite-sibbing lines from individual of S 1 

progenies have been proposed (Kinman, 1952). 

The method fixed the composite-sibbed lines since 

the first selfing and therefore improvement in the 

combining ability or other characteristics of 

composite-sibbing lines can not be made after 

several generations of mass sibbing unless 

effective selection is practiced. In other way, line 

selection from cross between closely related 

parents has been proved to be an effective method 

for inbred line development (Rasmusson and 

Phillips, 1977; Troyer, 1999). Selection for high 

and low yield lines effectively separated lines into 

high and low combining ability groups but yield 

of lines within group cannot be used as criterion 

Department of Agronomy, Faculty of Agriculture, Kasetsart university, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: agrkrs@ku.ac.th 


for combining ability of lines (Lamkey and 

Hallauer, 1986). In addition, for effective 

differentiation of lines, Fasoula and Fasoula (1997) 

proposed line selection under nil-competition 

environment in honeycomb designs. In order to 

improve yield and combining ability of population, 

Landi and Frascaroli (1993) applied full-sib 

selection in F 2 population of single cross. The 

method proved to be very effective for several 

cycles of selection. However, the previous study 

of Genter (1976) which applied the same method 

suggested that using S 1 instead of S 0 to form fullsibs 

was more effective to identify high yielding 

full-sibs as well as in improvement of population 

per se. This finding agreed well with suggestion 

of Lonnquist (1950) that testing for combining 

ability after one generation of selfing is desirable 

when the composite sib-breeding method is used. 

The above finding suggested that 

alternate selfing and full sibbing among few 

closely related lines under low-competition 

environment should lead to uniform, high yield 

and high combining composite lines as high level 

of homozygosity is approached and provide a 

chance for continuous improvement of composite 

lines in the successive cycles. 

The present study therefore aim to 

formulate the effective breeding method for the 

development of composite lines and evaluate its 

merit as compared to the conventional line 

selection with early generation testing for 

combining ability. The modified S 1-full sib 

selection within related lines is proposed. 


Six commercial single cross hybrids 

comprised Monsanto 949, Monsanto 919, Pioneer 

A33, Pioneer 3012, Pacific 984 and Syngenta NK 

48 were planted in normal plant spacing (0.75 × 

0.25m) and selfed to obtain S 1 ears. Nine S 1 ears 

within each family were randomly grouped in to 

3 ear sets, 3 sets per family and therefore resulted 


in 18 sets of 3 S 1 and 54 individual S 1 lines. They 

were separately ear-rowed in honeycomb 

arrangement (HC) with equilateral triangular side 

of 0.866m. Three best S 1 plants within each set 

were intercrossed (full sibbing) to form 18 intraset 

diallel crosses which will be refered to as full 

sib sets while 3 best S 1 plants from each family 

were also selfed to obtain 18 S 2 lines. 

Consequently, they were ear-rowed in 

HC, the 18 S 2 plants were selfed as well as 

testcrossed to the inbred tester, KRi 208 to obtain 

18 S 3 lines and 18 testcrosses, S 2 × KRi 208 

hybrids. The best S 2 lines by visual selection, one 

from each family, were also intercrossed to form 

15 diallel crosses of 6 S 2. In the meantime, the 

best 3 F 1 plants from each full sib set were crossed 

in all possible combinations to form 18 composite 

lines and they will be referred to as composite line 

cycle-1 (C#1). The method is essentially similar 

to S 1 and full-sib selection of which it will be 

referred to as modified S 1-full sib selection for 

composite line development. Afterward, C#1 were 

testcrossed to KRi 208. As a result, 18 C#1 

testcrosses were obtained. In addition, the best C#1 

by visual selection, one from each family, were 

intercrossed to form 15 diallel crosses of 6 C#1. 

Yield trials of 18 S 3 lines, 18 C#1, 18 

testcrosses of S 2 × KRi 208, 18 testcrosses of C#1 

× KRi 208, 15 diallel crosses of 6 C#1 and 15 

diallel crosses of 6 S 2 lines were conducted in 

separate trials in adjacent areas in randomized 

completed block design with 4 replications, 1 row 

plot of 5m long and 0.75 × 0.25m plant spacing. 

Five original hybrids were included as common 

checks in all hybrid yield trials. Pacific 984 was 

excluded and replaced by Suwan 4452 because 

the former was dropped out from the market and 

there was no seed available. 

All experiments were conducted from 

September 2004 to March 2006 at National Corn 

and Sorghum Research Center, Suwan Farm, in 

Nakhon Ratchasima province (14 0 30’N, 101 0 30’ 

E, and 356m asl.), Thailand under standard cultural

244 

practices. Basal fertilizers were applied at planting 

time at the rate of 75 kg ha -1 of N and 100 kg ha -1 

of P 2O 5. Top-dressing was done at the 6 to 8 leaf 

stages with the rate of 75 kg N ha -1 . Pre-emergence 

herbicides, Atrazine and Alachlor were used by 

mixing at the rate 1.5 and 1 kg a.i. per ha, 

respectively. Thinning was done at 14 days after 

sowing. Irrigation was applied when necessary. 


Mean grain yields and other agronomic 

traits of top-10 S 3 lines are presented in Table 1. 

All selected lines were statistically not different 

except line 406-3 and only line 401-6 showed 

significant difference over the inbred check, KRi 

208. However, the KRi 208 had higher level of 

homozygosity and therefore, if further inbreeding 

was applied, all lines were expected to be similar 

in yield level. There was no clear evidence for the 

advantage or disadvantage of other agronomic 

traits among the top ten lines but line 403-5 


demonstrated a superior shelling percentage over 

other lines including the checks. 

In comparison S 3 with C#1 lines, the C#1 

lines were consistently superior in the 

characteristics used as measures of vigor; grain 

yield, earliness of anthesis and silking, plant and 

ear height. They were earlier, taller and had higher 

yield regardless of germplasm sources. Moreover, 

better distribution of germplasm sources of top- 

10 C#1 was evident. All six germplasm sources 

were present in the top-10 C#1 while in the top- 

10 S 3 lines, visual selection leaned toward 

Monsanto 949 and Pacific 984. The results 

indicated that C#1 was more stable by outcrossing. 

On the other hand, inbred lines from each 

germplasm source should have different level of 

inbreeding depression and thus selection for 

performance per se was biased toward the less 

inbreeding depression germplasm. In this case, 

Pioneer 3012 was lost from the top-10 S 3 lines. 

The present results agreed well with 

report presented by Kinman (1952) of which 

Table 1 Grain yields at 15 percent moisture and other agronomic traits of top 10 S 3 lines and KRi 208 

at Suwan Farm, Thailand in November 2005 (dry season). 

S3 lines Source of Grain Days to Days to Moisture Plant Ear Shelling 

germplasms Yield Anthesis Silking Content Height Height (%) 

(ton/ha) (days) (days) (%) (cm) (cm) 

401-6 Pac.984 4.21 a 68.7 a-d 69.0 bcd 21.9 a-d 131.5 54.2 bc 76.6 a-d 

402-6 Mon.949 3.77 ab 68.3 a-e 67.3 edf 25.6 a 140.2 60.3 abc 73.6 bcd 

404-4 Pio.A33 3.71 ab 67.7 b-f 67.3 edf 23.7 abc 140.3 63.5 abc 73.9 bcd 

402-8 Mon.949 3.63 ab 66.7 ef 67.0 ef 24.9 ab 134.3 60.7 abc 75.6 bcd 

401-9 Pac.984 3.40 ab 70.0 a 68.8 b-e 23.4 a-d 131.3 54.3 bc 79.1 ab 

402-7 Mon.949 3.38 ab 66.3 f 66.3 f 23.9 abc 138 52.3 c 75.7 bcd 

405-4 Syn. 48 3.36 ab 67.0 def 67.7 c-f 22.8 a-d 134.7 60.5 abc 71.8 cd 

403-5 Mon.919 3.21 abc 68.0 b-f 68.0 b-f 20.2 cd 126 52.3 c 82.2 a 

401-7 Pac.984 3.08 a-d 70.0 a 69.0 bcd 22.0 a-d 131.5 58.7 abc 78.0 ab 

406-3 Pio.3012 2.89bcd 70.0 a 68.8 b-e 24.5 ab 151.5 68.2 ab 74.6bcd 

KRi 208 Pio.3012/3013 3.01bcd 69.0 abc 69.7 ab 25.5 a 120.3 50.2 c 73.5 bcd 

Mean 3.48 68.2 68.0 23.4 132.8 56.7 76.0 

F-value1/ ** ** ** ** ns ** ** 

CV(%) 19.568 1.338 1.353 8.489 10.561 13.461 3.973 

1/ ns: non significant, * : significant, ** : highly significant

selective mass sibbing within individual S 1 

progenies was used. In Kinman’s words, the 

population is closed at the time of first sibbing, it 

should not be expected that improvement in the 

combining ability or other characteristics of 

composite-sibbed lines will be made even after 

several generations of mass sibbing unless 

effective selection is practiced. Unlike Kinman’s 

method, the modified S 1-full sib selection 

employed in the present study provided a more 

flexible approach. Selection for S 1 performance 

per se alternate with diallel cross of individual of 

3 selected S 1 lines (full sibbing) should improve 

general combining ability as well as specific 

combing ability of S 1 lines from successive cycles. 

In the meantime, the newly emerged individual 

S 1 as well as full sib of each cycle can be fixed by 

mass sibbing method and used in early generation 

hybrid combinations while the successive cycles 

of composite sets move slowly toward higher level 

of homozygosity and hence more uniform lines 

and hybrids in later stages. 

Lamkey and Hallauer (1986) found that 


inbred line performance per se can be used as a 

criterion to differentiate combining ability between 

high and low yield inbreds. However, yield per se 

within high or low yielding groups cannot be used 

to predict line performance in hybrid 

combinations. Yielding ability of line per se in 

Table 1 and their testcross performance in Table 3 

clearly supported the above finding. Since all 18 

inbred lines came from the top-3 high yield lines 

of each original hybrid therefore they should be 

considered high yield lines. However, their 

yielding ability did not represent the combining 

ability of lines in the testcross combinations with 

the inbred tester (KRi 208), line 403-4 which was 

excluded from the top-10 lines gave the highest 

yield in the testcrosses while the top yield line, 

401-6 ranked 9 th in testcrosses. Besides, only two 

Pioneer lines, 406-1 and 404-4 were present in the 

top-10 testcrosses. This is not unexpected because 

the tester line, KRi 208 derived from Pioneer 3012/ 

Pioneer 3013. Therefore, genetic background of 

tester played an important role in the combinations 

with tested lines. However, 406-1/KRi 208 is 

Table 2 Grain yields at 15 percent moisture and other agronomic traits of composite lines of cycle 1 st 

at Suwan Farm, Thailand in November 2005 (dry season). 

Compos Source of Grain Yield Days to Days to Moisture Plant Ear Shelling 

_ite lines germplasms (ton/ha) Anthesis Silking Content Height Height (%) 

(days) (days) (%) (cm) (cm) 

Set 4 Mon.949 6.13 a 67.3 abc 67.3 bc 25.5 a 168.7 82 77.2 

Set 5 Mon.949 5.53 ab 65.0 d 67.0 bc 25.6 a 165.6 71.2 79.2 

Set 18 Mon.919 5.26 abc 63.0 e 66.0 c 21.4 cd 157.4 63.9 80.8 

Set 10 Syn. 48 4.91 bc 66.0 cd 66.7 bc 21.9 bcd 173.3 80.3 79.6 

Set 11 Syn. 48 4.87 bcd 66.0 cd 67.0 bc 22.9 bcd 161.2 67.7 72.9 

Set 2 Pac.984 4.81 bcd 68.3 ab 67.7 bc 23.8 abc 171.8 75.2 79.4 

Set 8 Pio.A33 4.66 bcd 66.7 a-d 67.0 bc 22.8 bcd 165.8 80.8 80.2 

Set 14 Pio.3012 4.64 bcd 68.0 abc 67.7 bc 22.0 bcd 158.1 69.2 80.2 

Set 3 Pac.984 4.63 bcd 68.7 a 68.3 ab 23.8 abc 159.3 65.5 79.7 

Set 7 Pio.A33 4.58 bcd 66.3bcd 67.0 bc 22.6 bcd 150.3 67.7 79.5 

Mean 5.00 66.5 67.2 23.2 163.2 72.4 78.9 

F-value1/ * ** * ** ns ns ns 

CV(%) 14.076 1.73 1.575 5.815 7.111 15.454 3.299 


246 

essentially a backcross to sister line and ranked 

6 th in the top-10 testcrosses indicated a strong 

additive effect in this hybrid combination. Since 

different testers gave different performance with 

the same group of lines (Castellanos et al., 1998), 

all high yield lines should be tested for their hybrid 

combinations directly to their counterpart parental 

lines to identify the best hybrid combination. 

Statistically, all top-10 testcrosses 

yielded as high as the top-4 checks but somewhat 

better than Monsanto 949 and Monsanto 919. 

However, 403-4/KRi 208 gave an outstanding 

feature of yield and earliness even though it was 

taller and lower in shelling percentage than the 

average. 


Top S 2 lines, one from each of six 

original hybrids were intercrossed and the top-10 

interfamily hybrids are presented in Table 4. As 

expected, the average of top-10 S 2 interfamily 

diallel hybrids was lower than that of top-10 S 2 

testcrosses because both parental lines of S 2interfamily 

hybrids were more heterogeneous than 

the tester line, KRi 208 in S 2 testcrosses. Therefore, 

the specific combining ability of lines were more 

pronounced. Seven out of 10 S 2-interfamily 

hybrids were involved with Pioneer 404-6 and 

Pioneer 406-1 and 6 out of 10 were crosses 

between Pioneer and Monsanto lines. Evidently, 

both germplasm sources complimented each other 

of which they showed a good heterotic pattern. 

Table 3 Grain yields at 15 percent moisture and other agronomic traits of top 10 testcrosses between 

selected S 2 × KRi 208 and original hybrids conducted at Suwan Farm, Thailand in November 

2005 (dry season). 

Lines × Source of Grain Yield Days to Days to Moisture Plant Ear Shelling 

KRi 2082/ germplasms (ton/ha) Anthesis Silking Content Height Height (%) 


403-4 Mon.919 8.96 a 61.3 h 61.7 j 22.6 def 165.8 b-f 81.7 a-f 75.9 h-k 

405-5 Syn. 48 8.82 ab 62.3 e-h 64.0 d-i 23.0 c-f 153.3 fgh 70.3 fg 77.1 f-i 

402-6 Mon.949 8.40 a-d 62.0 fgh 62.7 hij 24.9 abc 161.0 d-h 80.3 a-f 75.3 k 

405-4 Syn. 48 8.14 a-e 61.7 gh 62.3 hij 23.2 c-f 158.0 d-h 79.0 b-g 77.5 efg 

405-6 Syn. 48 8.08 a-f 63.3 b-g 64.3 b-g 24.3 bcd 157.6 d-h 82.7 a-f 75.6 ijk 

406-1 Pio.3012 8.07 a-f 63.0 c-h 63.7 e-i 24.1 bcd 164.2 b-f 81.7 a-f 78.3 def 

402-7 Mon.949 8.00 a-f 61.3 h 62.0 ij 24.0 bcd 150.0 gh 73.7 d-g 77.0 f-i 

404-4 Pio.A33 7.89 a-f 65.0 ab 65.3 c-f 23.6 b-f 159.8 d-h 76.8 b-g 75.4 jk 

401-6 Pac.984 7.70 a-g 62.7 e-h 63.3 f-j 23.4 c-f 156.3 e-h 72.7 d-g 79.5 cd 

403-5 Mon.919 7.66 a-g 62.0 fgh 62.7 hij 21.9 ef 162.1 c-h 89.8 a-g 81.1 bc 

Check Pio.A33 8.47 abc 64.7 bc 66.0 bcd 21.9 ef 182.0 a 91.9 a 79.8 bc 

Check SW 4452 8.24 a-e 66.7 a 68.7 a 25.0 abc 177.3 ab 84.2 a-e 76.3 g-k 

Check Pio.3012 8.08 a-f 66.7 a 68.0 ab 22.4 def 165.2 b-f 89.1 ab 77.2 fgh 

Check Syn. 48 7.86 a-f 64.0 b-e 66.0 bcd 22.7 def 171.3 a-d 74.4 dfg 77.9 ef 

Check Mon.949 7.57 b-g 62.7 e-h 64.3 d-g 26.7 a 178.3 ab 81.6 a-f 77.9 ef 

Check Mon.919 

Mean of top 

7.32 c-g 63.3 b-g 65.0 c-g 23.9 b-e 176.1 abc 88.1 abc 82.9 a 

10 topcrosses 8.17 62.5 63.2 23.5 158.8 78.9 77.3 

F-value1/ ** ** ** ** ns * ** 

CV(%) 11.006 1.836 1.985 5.198 5.296 9.685 1.179 

1/ ns: non significant, * : significant, ** : highly significant 

2/ Pedigree of KRi 208 is Pio.3012/Pio.3013

Although most of S 2-interfamily hybrids were 

significantly not different from the checks, 404- 

6/402-6 (Pioneer A33/Monsanto 949) gave 

outstanding features for yielding ability, earliness, 

plant and ear height while retained good shelling 

percentage. Therefore, beside the conventional 

testcross program, diallel cross between the top 

high yield lines is necessary for thorough use of 

germplasms and identification of new unique 

hybrid combination. 

The numbers of original germplasm 

sources involved in top-10 S 2 and C#1 testcrosses 

in Table 3 and 5 were almost the same; 4:5 

(Monsanto), 3:3 (Syngenta), 2:1 (Pioneer) and 1:1 

(Pacific) indicated that they responded similaly to 


the same tester, even though each S 2 line derived 

from visual selection within each composite set. 

However, average yield of S 2 testcrosses was 

higher than that of C#1 testcrosses but top testcross 

yields of both groups as well as the best check 

were more or less the same. 

The average yield of S 2 diallel crosses 

in Table 4 and that of C#1 diallel crosses in Table 

6 were almost the same but with the higher trend 

toward the C#1 lines. Evidently, general 

combining ability of S 2 and C#1 were somewhat 

the same even though the C#1 were more 

heterogeneous. Surprisingly, the top-2 hybrids of 

C#1 gave higher yield over other hybrids and 

checks tested in the present studies eventhough 

Table 4 Grain yields at 15 percent moisture and other agronomic traits of interfamily diallel hybrids 

of selected S2 lines and original hybrids at Suwan Farm, Thailand in November 2005 (dry 

season). 

S2 × S2 Source of Grain Days to Days to Moisture Plant Ear Shelling 

germplasms Yield Anthesis Silking Content Height Height (%) 

(ton/ha) (days) (days) (%) (cm) (cm) 

404-6×402-6 Pio.A33 × Mon.949 8.86 a 62.3 d 63.3 g 25.1 ab 171.9 de 80.5 d-g 76.1 gh 

404-6×403-6 Pio.A33 × Mon.919 7.82 abc 62.7 cd 64.3 efg 23.1 b-f 170.3 de 81.7 c-g 77.2 efg 

406-1×402-6 Pio.3012 × Syn.48 7.61 bcd 66.3 a 67.3 abc 24.0 b-e 202.3 a 106.9 a 74.8 h 

406-1×405-5 Pio.3012 × Syn.48 7.53 b-e 64.0 bcd 65.3 def 24.3 bcd 197.5 ab 96.7 ab 77.4 efg 

403-6×401-9 Mon.919 × Pac.984 7.51 b-e 62.3 d 63.7 fg 21.8 efg 169.2 de 77.0 fg 80.6 b 

406-1×404-6 Pio.3012 × Pio.A33 7.47 b-e 65.7 ab 66.7 bcd 23.8 b-f 194.7 abc 97.2 ab 78.2 c-f 

404-6×401-9 Pio.A33 × Pac.984 7.11 cde 64.0 bcd 66.3 bcd 23.3 b-f 175.4 de 84.1 c-f 80.0 bc 

406-1×403-6 Pio.3012 × Mon.919 6.59 de 63.7 bcd 65.0 d-g 22.1 def 183.5 bcd 93.3 bc 77.8 efg 

405-5×402-6 Syn.48 × Mon.949 6.55 de 63.3 cd 65.3 def 24.6 abc 178.3 de 79.9 efg 75.0 h 

401-9×402-6 Pac.984 × Mon.949 6.32 e 63.0 cd 64.3 efg 23.4 b-f 180.9 cde 77.3 efg 78.2 c-f 

Check Pio.A33 8.47 ab 64.7 abc 66.0 cde 21.9 ef 182.0 bcd 91.9 bcd 79.8 bcd 

Check SW 4452 8.26 abc 66.7 a 68.7 a 25.0 ab 177.3 de 84.2 c-f 76.3 fgh 

Check Pio.3012 8.08 abc 66.7 a 68.0 ab 22.4 def 165.2 e 89.1 b-e 77.2 efg 

Check Syn. 48 7.86 abc 64.0 bcd 66.0 cde 22.7 c-e 171.3 de 74.4 g 77.9 d-g 

Check Mon.949 7.57 bcd 62.7 cd 64.3 efg 26.7 a 178.3 de 81.6 c-g 77.9 d-g 

Check Mon.919 

Mean of top 10 

7.32 b-e 63.3 cd 65.0 d-g 23.9 b-e 176.1 de 88.1 b-f 82.9 a 

interfamily cross 7.34 63.7 65.2 23.6 182.4 87.5 77.5 

F-value1/ ** ** ** ** ** ** ** 

CV(%) 10.322 1.9 1.839 5.706 5.327 8.393 1.494 


248 

they were statistically not different. 

Evidences from previous studies (Genter, 

1976; Landi and Frascaroli, 1993; Rasmusson and 

Phillips, 1997 and Troyer, 1999) showed that 

selections in a very narrow base populations were 

very effective for the improvement of the 

populations as well as inbred lines per se. The 

method for composite line improvement used in 

the present studies is very similar to that suggested 

by Genter (1976) for population improvement but 

only 3 S 1 lines were used to form new population 

of each cycle, aiming to get uniform, high yield 

and high combining ability composite lines for 

better hybrid combinations. The method is simply 

a modification of S 1 and full-sib selection and 


therefore it will be referred to as modified S 1-full 

sib selection method. Data presented in this study 

did not show any clear advantage of line selection 

over the composite line method. More advanced 

cycles of S 1-full sib selection are underway to 

prove the merit of the method as compared to the 

conventional line selection by pedigree method. 

The composite-sibbed lines as proposed 

by Kinman (1952) is clearly had an advantage over 

line selection method when time and space are 

involved. Composite-sibbed lines are ready for 

final testing without five or six generations of 

selfing usually practice in the development of 

inbred lines. In the modified S 1-full sib selection, 

composite-sibbed lines can be derived from 

Table 5 Grain yields at 15 percent moisture and other agronomic traits of top 10 testcrosses between 

composite lines of cycle 1 × KRi 208 and original hybrids at Suwan Farm, Thailand in 

November 2005 (dry season). 

Set Source of Grain Yield Days to Days to Moisture Plant Ear Shelling 

numbers germplasms (ton/ha) Anthesis Silking Content Height Height (%) 


set 4 Mon.949 8.74 a 62.7 fg 63.3 ghi 24.9 abc 172.5 a-d 84.4 76.1 def 

set 3 Pac.984 8.31 abc 65.7 a-d 67.0 a-d 23.8 b-g 165.5 c-g 84.9 78.9 bc 

set 11 Syn. 48 7.80 a-d 64.0 c-f 65. d-h 22.9 b-h 163.9 d-g 78.5 74.6 fg 

set 5 Mon.949 7.65 a-d 61.7 g 62.7 i 23.8 b-g 168.8 b-g 80.5 76.5 c-f 

set 12 Syn. 48 7.44 b-d 63.3 efg 64.7 e-h 22.9 b-g 160.7 efg 77.2 76.3 def 

set 17 Mon.919 7.37 b-f 63.0 efg 65. d-h 21.1 h 161.9 d-g 81.9 76.2 def 

set 10 Syn. 48 7.34 b-f 62.7 fg 63.0 hi 22.2 fgh 160.5 fg 81.2 77.1 cde 

set 15 Pio.3012 7.24 c-g 66.0 abc 67.0 a-d 24.0 b-g 189.9 g 82.3 77.2 cde 

set 16 Mon.919 7.20 c-g 63.3 efg 64.7 e-h 21.7 gh 161.9 d-g 83.7 75.3 efg 

set 18 Mon.919 7.20 c-g 63.7 efg 65.3 c-g 23.9 b-g 169.3 b-g 85.4 77.3 cde 

Check Pio.A33 8.47 ab 64.7 b-f 66.0 b-f 21.9 gh 182.0 a 91.9 79.8 b 

check SW 4452 8.26 abc 66.7 ab 68.7 a 25.0 ab 177.3 ab 84.2 76.3 def 

Check Pio.3013 8.08 a-d 66.7 ab 68.0 ab 22.4 e-g 165.2 d-g 89.1 77.2 cde 

Check Syn. 48 7.86 a-d 64.0 c-f 66.0 b-f 22.7 c-g 171.3 a-e 74.4 77.9 bcd 

Check Mon.949 7.57 a-d 62.7 fg 64.3 ghi 26.7 a 178.3 ab 81.6 77.9 bcd 

Check Mon.919 


7.32 b-f 63.3 efg 65. d-h 23.9 b-g 176.1 abc 88.1 82.9 a 

topcrosses 7.63 63.6 64.8 23.1 167.5 82.0 76.6 

F-value1/ ** ** ** ** ** ns ** 

CV(%) 10.175 1.923 1.921 5.964 3.873 7.455 1.93 


composite sets as used in this study or using the 

individual S 1 and full-sib of each successive cycle. 

In addition, S 1 lines may be selfed for one or two 

additional generations in order to eliminate the 

undesirable alleles and several desirable sister lines 

may then be composited to establish the 

composite-sibbed lines. 

CONCLUSION 

Line selection combined with early 

generation testing for combining ability is an 

effective method. It gave higher average yield of 

top-10 S 2 testcrosses over the composite 


testcrosses. However, statistically, there was no 

clear advantage of yield between both groups of 

lines in early generation testcrosses. Besides, the 

selected S 2 and composite lines showed similar 

results in diallel cross sets. Visual selection under 

low-competition environment proved to be a very 

effective method to identify good combining and 

relatively high yield lines. However, testcross and 

diallel cross should be applied for thorough test 

of combining ability of lines. 

Composite lines had clear advantages 

over S 3 lines in yield, earliness and plant height. 

The modified S 1-full sib selection for the 

improvement of composite lines is a flexible 

Table 6 Grain yields at 15 percent moisture and other agronomic traits of interfamily diallel hybrids of 

composite lines (cycle 1) and original hybrids at Suwan Farm, Thailand in November 2005 

(dry season). 

Composite2/ Source of germplasms Grain Days to Days to Moisture Plant Ear Shelling 

× Yield Anthesis Silking Content Height Height (%) 

composite (ton/ha) (days) (days) (%) (cm) (cm) 

2 × 4 Pac.984 × Mon.949 9.33 a 65.0 c-h 66.3 b-e 22.9 b-e 190.1 95.3 abc 78.8 b-e 

4 × 7 Mon.949 × Pio.A33 9.18 ab 63.7 g-j 64.6 def 24.4 bc 181.4 92.9 a-d 78.5 b-e 

2 × 15 Pac.984 × Pio.3012 7.98 c-f 65.7 a-f 66.3 b-e 19.3 f 182.9 91.9 b-e 78.9 b-e 

4 x 11 Mon.949 × Syn. 48 7.98 c-f 63.0 ij 64.0 f 23.0 b-e 184.9 89.3 b-f 77.5 de 

7 × 17 Pio.A33 × Mon.919 7.62 c-h 63.7 g-j 65.0 def 21.4 ef 177.6 95.7 ab 77.9 cde 

11 × 15 Syn. 48 × Pio.3012 7.20 d-i 65.3 b-g 66.3 b-e 22.8 c-e 187.5 100.5 a 77.6 de 

2 × 17 Pac.984 × Mon.919 7.06 e-i 66.0 a-e 66.7 a-d 20.5 ef 189.5 86.3 e-g 77.3 de 

7 × 15 Pio.A33 × Pio.3012 6.97 e-i 66.7 abc 68.3 ab 21.4 ef 169.7 87.1 c-g 77.8 de 

4 × 15 Mon.949 × Pio.3012 6.86 f-i 64.3 e-j 65.0 def 23.1 b-e 186.5 92.4 a-e 77.8 de 

15 × 17 Pio.3012 × Mon.919 6.75 ghi 63.3 i-j 65.0 def 21.0 ef 175.6 89.1 b-f 81.2 abc 

Check Pioneer A33 8.47 abc 64.7 d-i 66.0 d-f 21.9 def 182 91.9 b-e 79.8 a-d 

Check Suwan 4452 8.26 a-d 66.7 abc 68.7 a 25.0 ab 177.3 84.2 efg 76.3 e 

Check Pioneer 3012 8.08 b-e 66.7 abc 68.0 abc 22.4 c-f 165.2 89.1 b-g 77.2 de 

Check Syngenta NK 48 7.86 c-g 64.0 f-j 66.0 d-f 22.7 c-f 171.3 74.4 h 77.9 cde 

Check Monsanto 949 7.57 c-h 62.7 j 64.3 ef 26.7 a 178.3 81.6 fgh 77.9 cde 

Check Monsanto 919 


7.32 c-i 63.3 hij 65.0 def 23.9 bcd 176.1 88.1 b-g 82.9 a 

interfamily cross 7.69 64.7 65.8 22.0 182.6 92.1 78.3 

F-value1/ ** ** ** ** ns ** * 

CV(%) 9.449 1.825 1.941 5.827 7.565 5.779 2.557 

1/ ns: non significant, * : significant, ** : highly significant 

2/ Crosses between two sets of composite lines.

250 

method which can be applied to improve the 

composite as well as inbred lines. However, further 

investigation is required to prove its merit for the 

construction of early generation hybrids as well 

as for the improvement of inbred lines. 

ACKNOWLEDGMENT 

We are grateful to Higher Education 

Project of Nong Lam University, Vietnam for 

financial support and to the staff of National Corn 

and Sorghum Research Center, Nakhon 

Ratchasima, Thailand for their kind helps during 

the time we did experiments. 


Castellanos, J.S., A.R. Hallauer and H.S. Cordova. 

1998. Relative performance of testers to 

identify elite lines of corn (Zea mays L.). 

Maydica 43: 217-226. 

Fasoula, D.A. and V.A.Fasoula. 1997. Competitive 

ability and plant breeding. Plant Breeding 

Review 14: 89-138. 


Genter, C.F. 1976. Recurrent selection for yield 

in the F 2 of maize single cross. Crop Sci. 16: 

350-352. 

Kinman M.L. 1952. Composite-sibbing versus 

selfing in development of corn inbred lines. 

Agron. J. 44: 209-241. 

Lamkey, K.R. and A.R. Hallauer. 1986. 

Performance of high x high, high x low and 

low x low crosses of lines from the BSSS 

maize synthetic. Crop Sci. 26: 1114-1118. 

Landi, P. and E. Frascaroli. 1993. Responses to 

four cycles of full-sib family recurrent 

selection in an F 2 maize population. Maydica 

38:31-37. 

Lonnquist, J.H. 1950. The effect of selection for 

combining ability within segregating lines of 

corn. Agron. J. 42: 503-508. 

Rasmusson, D. C. and R. L. Phillips. 1997. Review 

and interpretation: plant breeding progress and 

genetic diversity from de novo variation and 

elevated epistasis. Crop Sci. 37: 303-310. 

Troyer, A.F. 1999. Review and interpretation: 

background of U.S. hybrid corn. Crop Sci. 

39: 601-626.


Anther Culture of BC 1F 1 (KDML105//IRBB5/KDML105) Hybrid to 

Produce Bacterial Blight Resistance Doubled Haploid Rice 

Supanyika Sengsai 1 , Surin Peyachoknagul 1 , Prapa Sripichitt 2 , 

Amara Thongpan 1 and Pradit Pongtongkam 1 * 

ABSTRACT 

Maltose was found to be a better carbon source for callus induction in BC 1F 1 (KDML 105// 

IRBB5/KDML105) anther culture compared with sucrose. Statistical analysis, however, showed that 

increasing maltose or sucrose concentrations had no differential promotive effects on callus formation. 

One-step plantlet formation was found when maltose and NAA were supplemented together in the 

induction media. Adding 2 mg/l 2,4-D to the medium further increased the percentage of callusing 

anthers from 5.57% to 10.19%. However, the highest percentage of green plant regeneration was obtained 

(1.29%) from calli induced on N 6 medium without 2,4-D and subsequently cultured on regeneration 

medium containing MS supplemented with 2 mg/l BAP, 0.2 mg/l NAA, 300 mg/l casein hydrolysate, 

15% coconut water, and 30 g/l sucrose. AFLP analysis of all six anther-derived plants showed 57.3% 

to 67.12% recurrent parental alleles. After planting, seeds were detected in two out of six anther 

culture-derived plants indicating the occurrence of spontaneous chromosome doubling in these plants. 

Unfortunately, none of these six plants contained bacterial blight resistant gene (xa5) as detected by 

specific PCR-based RG556 marker and pathogen inoculation. 

Key words: KDML 05, anther culture, maltose, 2,4-D, AFLP, RG556, bacterial blight 

INTRODUCTION 

The production of haploid plants and 

doubled haploid plants from anther culture offers 

a rapid achievement of homozygous lines for early 

release of new crop varieties. Many desirable traits 

such as high grain weight, disease resistance, dwarf 

plant type and abiotic stress tolerance were 

introgressed into rice breeding population by 

culturing of anthers. Unfortunately, low 

percentages of both callus induction and plant 

regeneration are the principal constraints in 

establishing successful anther culture in some rice 

varieties especially in indica rice since these 

critical culturing responses are genotype 

dependent (Roy and Mandal, 2005). Consequently, 

the effective culture medium used for some rice 

varieties may not be appropriate for others, and 

the composition of culture media should be 

carefully selected when the anthers of particular 

rice variety was subjected to culture. 

Sucrose is generally added in rice anther 

culture media to serve as the standard carbon 

source and the osmotic regulator. However, many 



* Corresponding author, e-mail: fscipdp@ku.ac.th 


252 

reports revealed that the percentages of callus 

induction and plant regeneration could be 

increased by using maltose instead of sucrose 

(Lentini et al., 1995). In addition, the type of auxin 

in anther culture medium has been proposed to 

regulate the formation of rice callus. It was found 

that culturing on 2,4-D supplemented-media 

stimulated callus induction and cell proliferation 

in rice anther culture whereas having NAA resulted 

in direct androgenesis (Ball et al., 1993). 

This work aimed at investigating the 

effects of maltose, sucrose and 2,4-D on the anther 

culture response of the BC 1F 1 hybrid of two 

recalcitrant genotypes, Khao Dawk Mali 105 

(KDML105) a well-known aromatic rice variety 

and IRBB5 a bacterial blight resistant rice variety 

containing xa5 resistant gene. In addition, the 

regeneration ability of the anther calli was also 

observed. After anther culture-derived plants (ACderived 

plants) were obtained, the Amplified 

Fragment Length Polymorphism (AFLP) was used 

to assess the contribution of the two parental 

genomes in these plants, and subsequent detection 

of xa5 resistant gene for bacterial blight resistance 

was done using PCR-based marker. 


BC 1F 1 seeds culturing and panicles collection 

BC 1F 1 seeds (KDML105//IRBB5/ 

KDML105) were obtained from the crossing 

between KDML 105 (a bacterial blight susceptible 

variety) and IRBB5 (the donor parent containing 

xa5 bacterial blight resistant gene). The BC 1F 1 

seeds were cultured on a modified MS medium 

(Murashige and Skoog, 1962) supplemented with 

2-3 mg/l BAP, 1 g/l yeast extract, 15% coconut 

water, 30 g/l sucrose and 0.8% agar. The BC 1F 1 

plants having xa5 gene were selected by PCRbased 

RG556 marker (Huang et al., 1997). Healthy 

tillers from the selected BC 1F 1 plants were 

separated and grown in pots. The panicles were 

collected from the primary, secondary and also 


tertiary tillers of plants when the microspores in 

anther were at the mid-to-late uninucleate stage 

as seen from the distance between the auricles of 

the two last leaves that reached 6-12 cm. 

Anther culture 

The panicles covered with flag-leaf 

sheaths were wrapped in a moist-soft paper, sealed 

in a plastic bag, and kept in the dark for 8-10 days 

at 12°C. After this cold-pretreatment, the panicles 

were surface-sterilized by spraying with 70% 

ethanol and the flag-leaf sheaths were removed. 

Then the anthers were cut off and cultured on 

callus induction media containing N 6 salts and 

vitamins (Chu, 1978), supplemented with 2 mg/l 

NAA, 1 mg/l kinetin, 500 mg/l casein hydrolysate, 

0.7% agar and also different concentrations of 

maltose or sucrose (40 , 50 , 60 g/l) at the adjusted 

pH of 5.8. The cultures were maintained under 

alternate 16/8 h light/dark at 25 ± 2°C for 45-50 

days. The numbers of anther calli formation were 

recorded and the percentages of calli induction 

were calculated. The experiment was set as 2×3 

factorial in completely randomized design with 

three replications. 

An addition of 2 mg/l of 2,4-D to callus 

induction medium of the same formular described 

above having 50 g/l maltose was also performed 

to determine the effect of 2,4-D on callus 

induction. 

Callus differentiation 

Anther calli of 1-2 mm diameter were 

randomly collected and transferred to the 

regeneration media consisted of MS salts and 

vitamins, supplemented with different 

concentrations of kinetin (1, 2, 3 mg/l), 300 mg/l 

casein hydrolysate, 1 g/l L-proline, 15% coconut 

water, 30 g/l sucrose and 2.5 g/l phytagel. An 

addition of regeneration medium having 2 mg/l 

BAP and 0.2mg/l NAA to replace kinetin and Lproline 

was set. The cultures were kept under 

alternate 16/8 h light/dark at 25 ± 2°C for 15-20

days. The numbers of calli producing complete 

plantlets were recorded and the percentage were 

calculated. 

AFLP analysis 

To assess the contribution of two parental 

genomes in AC-derived plants, AFLP was 

performed as described by Vos et al. (1995) with 

some modification. Fifteen combinations of primer 

(synthesized by KU Vector), set E primer (EcoRI 

end) and set M primer (MseI end), were used. Only 

clear AFLP bands were scored as present or absent. 

AFLP fingerprints of KDML105, IRBB5 and also 

AC-derived plants were analyzed. 

Detection of xa5 resistant gene for bacterial 

blight resistance by PCR-based marker 

The DNA of the resistant donor parent 

(IRBB5), recurrent susceptible parent 

(KDML105), and AC-derived plants were 

extracted from the leaves using the method 

described by Agrawal et al. (1992) and subjected 

to PCR amplification using synthesized primers 

(KU Vector). The RG556 primer linked to the xa5 

resistant gene was used to detect the presence of 

resistant gene. The sequence of RG556 F is 5′ 

TAGC TGCTGCCGTGCTGTGC 3′ while RG556 


R is 5′ AATATTTCAGTGTGCATCTC 3′ (Huang 

et al., 1997). PCR products were digested with 

Hpy CH4 IV restriction enzyme to detect the 

polymorphic DNA bands from bacterial blight 

resistant and susceptible plants (Sanchez et al., 

2000). 


Anther culture 

Ten days after incubation on callus 

induction media, approximately 90% of anthers 

turned brown (data not shown). This result, 

however, was not surprising because Guzman and 

Zapata-Arias (2000) also reported this changing 

of anther colors which is possibly due to the 

transition of gametophytic phase to sporophytic 

phase during androgenesis. In addition, our results 

showed that anther calli asynchronously emerged 

through the split lobes of these browning anthers 

after 45-50 days of culturing (Figure 1A). Most 

of the responding anthers produced multiple calli 

which ultimately became yellowish in color having 

both campact and friable callus types (Figure 1B). 

The calli were all in satisfactory condition. This is 

the first report on anther culture of BC 1F 1 seeds 

(KDML105//IRBB5/KDML105). 

Figure 1 Calli formation: (A) calli emerged through the split lobes of anther (arrow), (B) the anther 

calli (arrow), after 45-50 days of culturing on callus induction media containing maltose.

254 

The effects of maltose and 2,4-D on callus 

induction 

Using different concentrations of matose 

and sucrose in the culture medium, it was found 

that the percentages of calli induction ranging from 

4.61% to 6.11% in maltose but only 3.33% to 

3.43% in sucrose (Table 1). Callus formation was 

significantly affected by the type of sugar (P=0.01) 

but not significantly affected by the concentration 

of sugar itself. The highest percentage of callus 

induction was obtained by culturing BC 1F 1 anthers 

on N 6 medium supplemented with 2 mg/l NAA, 1 

mg/l kinetin, 500 mg/l casein hydrolysate, 60 g/l 

maltose and 7 g/l agar. Maltose, therefore, seemed 

to be a preferred carbon source for prolific callus 


formation in rice anther culture as also reported 

by Lentini et al. (1995). The beneficial effect of 

maltose has been ascribed to its slow degradation 

which results in stabilization of medium 

osmolarity (Kuhlmann and Foroughi-Wehr, 1989). 

In contrast, sucrose is rapidly hydrolysed to 

glucose and fructose, thereby, the osmolarity of 

the medium became double causing the negative 

effect on callus formation (Xie et al., 1995). 

Although callus induction rate gradually 

increased as the concentration of maltose or 

sucrose increased (Table 1), statistical analysis 

showed that rising of sugar concentration (40, 50, 

60 g/l) had no differential promotive effects on 

callus formation. This result did not agree with 

Table 1 Effects of maltose, sucrose and 2,4-D on anther callus formation of BC1F1 (KDML105// 

IRBB5/KDML105) hybrid. 

Types of sugar Concentration Number of Number of Percentage of callus 

in callus induction of sugar cultured callusing induction (%) 

medium (g/l) anthers anther 

Sucrose 40 1,140 38 3.33 

50 1,808 62 3.43 

60 848 29 3.42 

Maltose 40 4,252 196 4.61 

50 9,979 557 5.57 

60 7,695 470 6.11 

50a 4,378 446 10.19 

Analysis of variance: ANOVA table 

Source Df SS MS F(cal) F(table) 

value value 

5% 1% 

Treatment 5 22.16 

Factor A 1 18.68 18.68 14.83** 4.75 9.33 

Factor B 2 1.69 0.98 0.78 ns 3.88 6.93 

A×B 2 1.52 0.76 0.60 ns 3.88 6.93 

Error 12 15.12 1.26 

Total 17 37.28 

** indicated highly significance at 1%, ns indicated no significance 

50a referred to 50 g/l maltose supplemented medium + 2 mg/l 2,4-D 

Factor A was types of sugar, viz. maltose and sucrose. 

Factor B was sugar concentration, viz. 40, 50 and 60 g/l.

the report of Ching (1982) showing the increase 

of sugar concentration (30, 60, 90 g/l) to promote 

the higher percentage of callusing anthers as well 

as plantlet formation. The contradictory results 

may be due to the narrow range of sugar 

concentrations used which could not cause 

distinctive effects on callus induction in this study. 

It is also interesting to find that some of 

yellowish compact calli grown in 50 g/l or 60 g/l 

maltose containing media could differentiate to 

complete plantlets. The development of complete 

plantlets while they are culturing on induction 

medium is known as “one-step plantlet formation”. 

This result agreed with other reports showing that 

addition of maltose to NAA containing callus 

induction media promoted the formation of 

complete plantlet in rice anther culture (Zhao et 

al., 1999). Since the formation of plantlets by one 

step did not frequently occur in anther culture of 

KDML105 hybrids (Lertvichai, 1995; Boonintara, 

2004), three complete green plantlets (2.63%, data 

not shown) obtained from this experiment is 

considered a positive and satisfactory result. It 

should be noted here that genotype of anther donor 

plants, the type and concentration of sugar as well 

as the type of auxin have some effect on one-step 

plantlet formation (Zhao et al., 1999). 


To further increase the percentage of 

callus formation, 2 mg/l of 2,4-D was added to 50 

g/l maltose supplemented-medium. The results 

showed that the percentage of callus formation was 

increased from 5.57% to 10.19% (Table 1). 

Although the addition of 2,4-D favoured callus 

initiation and proliferation, the organogenesis of 

the calli might be inhibited hence one-step plantlet 

formation was not obtained. This result did not 

agree with those obtained by Datta et al. (1990) 

showing the combination of NAA and 2,4-D 

supplemented in callus induction medium 

promoted green plantlet formation in rice. This 

opposing result is possibly due to the powerful 

influence of genotype of anther donor plants on 

the response of rice anthers to these auxins. 

Callus differentiation 

After culturing on regeneration media, 

differentiation of the calli was observed at 15-20 

days. The percentages of calli development are 

shown in Table 2. The results indicated that the 

highest percentage of green plantlet formation 

(1.29%) was obtained when the calli grown on 

induction media without 2,4-D were transferred 

onto regeneration medium containing MS salts and 

vitamins supplemented with 2 mg/l BAP, 0.2 

Table 2 Development of calli on regeneration media. 

Callus Regeneration Number of Browning Proliferated Plantlet formation (%) 

induction media cultured calli calli Green Albino 

media anthers (%) (%) plantlets plantlets 

without SR1 140 14.29 74.29 0.71 4.29 

2,4-D SR2 123 18.71 65.85 0.81 1.63 

SR3 180 10.00 58.33 0.00 13.89 

MR1 77 9.09 12.89 1.29 12.99 

with SR1 15 13.33 80.00 0.00 0.00 

2 mg/ SR2 82 2.44 96.34 0.00 12.12 

2,4-D SR3 93 0.00 92.47 0.00 4.30 

MR1 92 4.34 92.39 1.09 0.00 

SR1, SR2, SR3 = MS + different concentration of kinetin (1 mg/l, 2 mg/l, 3 mg/l) +300 mg/l casein hydrolysate + 1 g/l L-proline 

+ 15% coconut water + 30 g/l sucrose + 2.5% phytagel 

MR1 = The same formular as SRs media but having 2 mg/l BAP and 0.2 mg/l NAA to replace kinetin and L-proline.

256 

mg/l NAA, 300 mg/l casein hydrolysate, 15% 

coconut water, 30 g/l sucrose and 2.5% phytagel 

(MR1 medium). It was also found that the 

percentage of green plantlet formation of calli 

cultured on the media without BAP and NAA was 

increased when the concentration of kinetin 

increased. By two steps culturing (callus induction 

and subsequent regeneration of calli), the total of 

four green plantlets were obtained (Figure 2A,B). 

However, one of these plantlets died during 

subculturing leaving only three plantlets for further 

investigation. 

It is interesting to find that regeneration 

medium highly supported callus proliferation 

(80%-96.34%, Table 2) of those previously grown 

in the medium containing 2 mg/ 2,4-D but the 

complete plantlets formation was better formed 

in the media without 2,4-D. These results implied 

that regeneration response of calli was possibly 

affected by the interaction between the 

composition, particularly the types of auxin, of 

induction media and regeneration media. 

Although the percentages of callus 

induction (3.33-6.11%) and green plantlet 

formation (0.00-1.29%) of BC 1F 1 in this study 

were lower than those of Lemont/KDML 105 

hybrid (2.22%-44.46% and 0.00%-20% 


respectively) as reported by Lertvichai (1995), they 

were comparable to those of KDML 105/Chainat1 

hybrid (Boonintara, 2004). The low outcome of 

callus formation and recovering plants from anther 

culture of indica rice are known to be genotypic 

dependent (Khanna and Raina, 1998). 

A large percentage of albinos (more than 

90%, data not shown) obtained in this study is 

considered unsatisfactory but similar to other 

reports of albinos ranging from 5% to 100% in 

rice anther culture, especially in indica rice 

(Bhojwani et al., 2001). Several factors, including 

pre-treatment, culture medium and culturing steps 

considerably affected the frequency of albinism. 

However, high sugar concentration might be 

another cause of albino plant formation as seen in 

the increase percentage of albino plantlet of 

japonica rice (Tainan 5) in the increased sugar 

culture (Chen, 1978). 

AFLP based background analysis in anther 

culture-derived plants 

AFLP analysis was performed on six ACderived 

plants (three plants obtained by one-step 

plantlet formation and the others from two-step 

plantlet formation) to assess the contribution of 

two parental genomes in these plants. Out of 373 

Figure 2 Development of calli: (A) green spot producing callus (arrow), (B) green plantlet regeneration, 

after 15-20 days of culturing on regeneration media containing MS supplemented with 2 

mg/l BAP, 0.2 mg/l NAA, 300 mg/l casein hydrolysate, 15% coconut water, and 30g/l sucrose.

clearly amplified bands generated by 15 primer 

combinations, 73 bands showed polymorphisms 

between the parents (Figure 4) of which 50 bands 

were specific for KDML105 and 23 specific bands 

for IRBB5 (data not shown). By assuming random 

distribution of AFLP markers in rice genome, it 

was found that the percentage of recurrent parental 

alleles (KDML105) recovered in AC-derived 

plants population was found ranging from 57.3% 

to 67.12% (Table 3). This result is somewhat 

narrower than the anticipated distribution of 

chromosomes containing recurrent parental alleles 

(50%-100%) in plantlets obtained from anther 

culture of BC 1F 1 plants, which is possibly due to 

the fact that AFLP primers used in the present study 

did not cover the whole genome. Furthermore, 

only six AC-derived plants were obtained and 

could not, thereby, represent the broad distribution 

of the overall AFLP alleles in the population. In 

addition, Guiderdoni (1991) reported that 

androgenesis of microspores containing more 

genetic make-up of recurrent parent may be 

masked by gametic selection. This selection 

resulted in the segregation distortion of alleles and 

preventing AC-derived plants from being truly 


BC 1F 1 gametic array as also shown in the anther 

culture of japonica/indica and indica/indica rice 

hybrids. 

After planting six healthy AC-derived 

plants, seeds were obtained in two out of these six 

plants (33.33%, data not shown) indicating the 

occurrence of spontaneous chromosome doubling 

which resulted in two homozygous lines from 

anther culture of BC 1F 1 (KDML105//IRBB5/ 

KDML105). This result was not surprising since 

the mechanisms to double chromosome can occur 

at various stages in vitro, including callus 

formation, callus re-differentiation and 

embryogenesis in rice anther culture, and even in 

tillers (Bishnoi et al., 2000). Although the 

percentage of chromosome doubling obtained 

from this study (33.33%) was lower than the anther 

culture of KDML105/RD23 (86.1%) reported by 

Pakdeechanuan (1997), it was higher than that of 

KDML105/Lemont (25%) (Lertvichai, 1995) and 

comparable to those of KDML105/Chainat1 

(33.33%) (Boonintara, 2004). Genotypic 

dependence was suspected to be the main cause 

affecting the frequency of chromosome doubling 

in rice anther culture (Sopory et.al. 1996). 

Table 3 The percentage of recurrent parental alleles (KDML105) recovered in AC-derived plants 

population as detected by AFLP using 15 primer combinations. 

Individual AC- Number of specific Number of specific Percentage of 

derived plants bands presented bands presented recurrent parental 

only in KDML105 only in IRBB5 alleles (KDML105) 

recovered in 

AC-derived plants 

(%) 

1 49 24 67.12 

2* 47 26 64.38 

3 42 31 57.53 

4* 47 26 64.38 

5 45 28 61.64 

6 43 30 58.90 

Average percentage of recurrent parental alleles recovered in 

AC-derived plants 

62.33 

* indicated spontaneous double haploids

258 

Detection of xa5 resistant gene for bacterial 

blight resistance by PCR -based marker 

Since the anther donor plants (BC 1F 1) 

were confirmed to be heterozygous for the xa5 

resistant gene (data not shown), AC-derived plants 

from BC 1F 1 were also tested for the xa5 gene using 

PCR-based marker RG556. The PCR product gave 

monomorphic amplification products of 1,600 bp 

(Figure 3A). However, after digesting with Hpy 

CH4 IV, polymorphism of DNA bands between 

resistant and susceptible plants was detected. Two 

bands of 1,000 bp and 300 bp (doublet) were found 

in IRBB5 (bacterial blight resistant variety) while 

non-digested DNA band (1,600 bp) was shown in 


susceptible variety of KDML105 (Figure 3B). 

Genotyping by PCR-based method revealed none 

of these six AC-derived plants contained the xa5 

gene for bacterial blight resistance (Figure 3B). 

This result was confirmed by pathogen inoculation 

test on 45 days old plants grown from healthy seeds 

of two homozygous lines (obtained by 

spontaneous chromosome doubling as described 

in AFLP analysis section) which showed bacterial 

blight susceptability (data not shown). 

Sanchez et al. (2000) reported that the 

distance between RG556 marker and the xa5 gene 

was 0.8 cM, then the loss of the xa5 gene in six 

AC-derived plants in the present study possibly 

Figure 3 PCR analysis of the bacterial blight susceptible variety (KDML105), resistant variety (IRBB5) 

and AC-derived plants (the samples number 1, 2, 3, 4, 5, 6): (A) Monomorphic bands amplified 

with primer RG556, (B) PCR products digested with Hpy CH4 IV. M= 1 kb plus DNA marker.

caused by either the recombination of RG556 

marker and the xa5 resistant gene or the 

segregation of gene during gametogenesis. 

CONCLUSION 

Two homozygous lines were rapidly 

achieved by anther culture of BC 1F 1 hybrid 

(KDML105//IRBB5/KDML105) and subsequent 

spontaneous chromosome doubling. In the present 

study maltose has proven to be a preferred carbon 


source compared to sucrose as seen from the 

significant effect on callus formation. Furthermore, 

one-step plantlet formation was promoted when 

callus induction medium supplemented with 

combination of maltose and NAA was used. 

Although the percentages of callusing anther and 

callus proliferation were increased by adding 

2,4-D to NAA containing induction medium, the 

percentages of organs formation as well as 

complete plantlet formation were very low, 

moreover one-step plantlet formation did not 

Figure 4 AFLP fingerprint of the bacterial blight susceptible variety (KDML105), resistant variety 

(IRBB5) and AC-derived plants (the samples number 1, 2, 3, 4, 5, 6) generated by different 

primer combinations: (A) E-AGG / M-CTC primer (B) E-AAG / M-CAA primer (C) E- 

AAG / M-CAG primer, M = 25 bp DNA size marker (Life Technologies); arrows indicate 

polymorphic bands specific for KDML105.

260 

occur. By two-steps culturing, the highest 

percentage of green plant regeneration was 

obtained (1.29%) from calli induced on N 6 

medium without 2,4-D and subsequently cultured 

on regeneration medium containing MS 

supplemented with 2 mg/l BAP, 0.2 mg/l NAA, 

300 mg/l casein hydrolysate, 15% coconut water, 

and 30g/l sucrose. The contribution of recurrent 

parental genome in AC-derived plants was 

revealed by AFLP analysis. Even though these ACderived 

plants did not contain a bacterial blight 

resistant gene (xa5) when screened by PCR-based 

RG556 marker, other desirable traits such as dwarf 

plant type, photoperiod insensitive response and 

aroma, characteristic of the parents could be 

obtained from them. 


This research was financially supported 

by Kasetsart University Research and 

Development Institute (KURDI) and Thesis and 

Dissertation Support Fund, Graduate School, 

Kasetsart University. The authors also would like 

to thank Dr. Kanchana Klakhaeng, Patumtani Rice 

Research Center for supplying rice seeds, and Dr. 

Nongrat Nilpanit, Division of Plant Pathology and 

Microbiology, Department of Agriculture, for 

providing pathogen inoculation test. 


Agrawal, R.K., R.N. Pandey and V.P. Agrawal. 

1992. Isolation of DNA from Choerospondias 

asillaris leaves. Biotech. Biodiv. 

Lett. 2: 19-24. 

Ball, S.T., H.P. Zhou and C.F. Konzak. 1993. 

Influence of 2,4-D, IAA and duration of callus 

induction in anther culture of spring wheat. 

Plant Sci. 90: 195-200. 

Bhojwani, S., H. Pande and A. Raina. 2001. 

Factors affecting androgenesis in indica rice. 

Available Source : http://www.bibd.un.- 


giessen.de/gdoe. September 21, 2001. 

Bishnoi, U.S., R.K. Jain, K.R. Gupta, V.K. 

Chowdhury and J.B. Chowdhury. 2000. High 

frequency androgenesis in indica x basmati 

rice hybrids using liquid culture media. Plant 

Cell Tiss. Org. Cult. 61: 153-159. 

Boonintara, J. 2004. Anther culture of F 1 hybrid 

and F 2 progenies of aromatic rice (Oryza 

sativa L.) for the production of double 

haploid with photoperiod insensitive 

character. MS. Thesis, Kasetsart University, 

Bangkok. 

Chen, C.C. 1978. Effect of sucrose concentration 

on plant production in anther culture of 

rice.Crop Sci.18: 905-906. 

Chu, C.C. 1978. The N 6 medium and its 

application to anther culture of cereal crops, 

pp. 43-50. In Proceeding of Symposium on 

Plant Tissue Culture. Science Press, Beijing. 

Datta, S.K., K. Datta and I. Potrykus. 1990. 

Embryogenesis and plant formation from 

microspores of both ‘indica’ and ‘japonica’ 

rice (Oryza sativa). Plant Sci. 67: 83-88. 

Guiderdoni, E., E. Galinato, J. Luistro and G. 

Vergara. 1991. Anther culture of tropical 

japonica x indica hybrids of rice (Oryza sativa 

L.). Euphytica 62: 219-224. 

Guzman, M. and F.J. Zapata-Arias. 2000. 

Increasing anther culture efficiency in rice 

(Oryza sativa L.) using anthers from ratooned 

plants. Plant Sci. 151: 107-114. 

Huang, N., E.R. Angles, J. Domingo, G. 

Magpantay, S. Singh, G. Zhang, K. 

Kumaravadiel, J. Bennett and G.S. Khush. 

1997. Pyramiding of bacterial blight resistance 

gene in rice:marker-assisted selection using 

RFLP and PCR. Theor. Appl. Genet. 95: 313- 

320. 

Khanna, H.K. and K. Raina. 1998. Genotype x 

culture medium interaction effects on 

regeneration response of three indica rice 

cultivars. Plant Cell Tiss. Org. Cult. 52: 145- 

153.

Kuhlmann, V. and B. Foroughi-Wehr. 1989. 

Production of doubled haploid lines in 

frequencies sufficient for barley breeding 

programs. Plant Cell Rep. 8: 78-81. 

Lentini, Z., P. Reyes, C.P. Martinez and W.M. 

Roca. 1995. Androgenesis in highly 

recalcitrant rice genotype with maltose and 

silver nitrate. Plant Sci. 110: 127-138. 

Lertvichai, R. 1995. Anther culture of hybrid 

rice (Lemont/Khao Dawk Mali 105). MS. 

Thesis, Kasetsart University, Bangkok. 

Murashige, T. and F. Skoog. 1962. A revised 

medium for rapid growth and bioassays with 

tobacco tissue culture. Physiol. Plant 15: 473- 

492. 

Pakdeechanuan, P. 1997. Anther culture of F(1) 

hybrid rice (KDML 105/RD23) for studying 

for brown planthopper (Nilaparvata lugens) 

resistance. MS. Thesis, Kasetsart University, 

Bangkok. 

Roy,B. and A.B. Mandal. 2005. Anther culture 

response in indica rice and variations in major 

agronomic charaters among the androclones 

of a scented cultivar , Karnal local. African 

J. Biotech. 4(3): 235-240. 


Sanchez, A.C., D.S. Brar, N. Huang, Z. Li and G.S. 

Khush. 2000. Sequence tagged site markerassisted 

selection for three bacterial blight 

resistance gene in rice. Crop Sci. 40: 792-797. 

Sopory, S.K. and M. Munshi. 1996. Anther culture, 

pp. 145-176. In S.M Jain, S.K. Sopory and 

R.E. Veilleux (eds.). In Vitro Haploid 

Production in Higher Plants, vol.1. Kluwer 

Academic Publishers, Netherlands. 

Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. Van 

de Lee, M. Hornes, A. Frijters, J. Pot, J. 

Peleman, M. Kuiper and M. Zabeau. 1995. 

AFLP: a new technique for DNA 

fingerprinting. Nucl. Acids Res. 23: 4407- 

4414. 

Xie, J., M. Gao, Q. Cai, X. Cheng , Y. Shen and Z. 

Liang. 1995. Improved isolated microspore 

culture efficiency in medium with maltose and 

optimized growth regulator combination in 

japonica rice (Oryza sativa L.). Plant Cell 

Tiss. Org. Cult. 42: 245-250. 

Zhao, J., C. Zhou and H.Y. Yang. 1999. In vitro 

development of early proembryos and plant 

regeneration via microculture in Oryza sativa. 

Plant Cell Tiss. Org. Cult. 55: 167-174.


Novel PCR Primers for Specific Detection of Xanthomonas citri 

subsp. citri the Causal Agent of Bacterial Citrus Canker 

Udomsak Lertsuchatavanich 1 , Ampaiwan Paradornuwat 1 , Junlapark Chunwongse 2 , 

Norman W. Schaad 3 and Niphone Thaveechai 1 * 

ABSTRACT 

The new primers were developed for specific detection of Xanthomonas citri subsp. citri (Hasse) 

(Xcc) [syn. X. axonopodis pv. citri (Xac)], the causal agent of Asiatic citrus canker disease. Twenty 

three strains of Xcc and 34 strains of other xanthomonads including X. fuscans subsp. aurantifolii, X. 

alfalfae subsp. citrumelonis, X. campestris pv. campestris, X. campestris pv. glycines, X. citri subsp. 

malvacearum and X. fuscans subsp. fuscans were tested for specificity of new primers by classical PCR. 

The results showed that these 354 F/R primers specifically amplified all of Xcc strains but not other 

xanthomonad strains. The 354-bp PCR fragment was sequenced and its nucleotide sequences were 

compared for similarity with Genbank database. The 354-bp nucleotide sequences were 99.7% similar 

to gene XAC2443 of Xac strain 306 (Accession AE011881). The sensitivity of these specific primers 

for detection of viable cells and total DNA of Xcc were 70 CFU/µl and 50 pg/µl, respectively. Therefore, 

these novel primers can be used as an alternative application for rapid and specific detection of Xcc. 

Key words: Xanthomonas, bacterial citrus canker, detection, polymerase chain reaction 

INTRODUCTION 

Bacterial canker of citrus is a serious 

disease of most citrus species and cultivars in many 

citrus-producing areas worldwide. Five forms of 

the disease have been described, cankers A, B, C, 

D, and E. Canker A or A-strain (Asiatic canker) is 

the most common and most damaging of the citrus 

canker strains (Schubert et al., 2001). It was 

originally found in Asia and is by far the most 

widespread. Recently, information based upon 

DNA sequences comparison or alignment of 16S- 

23S internal transcribed spacers (ITS) regions with 

amplified fragment length polymorphism (AFLP) 

analysis of the five recognized forms of citrus 

canker was demonstrated by Schaad et al. (2005, 

2006). Citrus pathogens were reclassified into 

three pathovars of Xanthomonas campestris (or 

X. axonopodis): pathovars citri for strain “A”, 

aurantifolii for strains “B/C/D” and citrumelo for 

strain “E”, which were revealed as taxon I 

including all “A” strains; taxon II containing all 

“B”, “C”, and “D” strains; and taxon III containing 

all “E” strains. The taxa I, II and III citrus strains 

were reinstated with new names, respectively as 

Xanthomonas citri subsp. citri (Hasse, 1915), 

1 Department of Plant Pathology, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand. 

2 Department of Horticulture, Faculty of Agriculture, Kasetsart University, Kamphaeng Saen Campus, Nakhon Pathom 73140, 

Thailand. 

3 USDA-ARS FDWSRU, 1301 Ditto Avenue, Fort Detrick, MD 21702-5023, USA. 

* Corresponding author, e-mail: agrnpt@ku.ac.th 


Xanthomonas fuscans subsp. aurantifolii (Gabriel 

et al., 1989), and Xanthomonas alfalfae subsp. 

citrumelonis (Riker et al., 1935). A new strain of 

X. citri subsp. citri, designated A*, was identified 

in southwest Asia and has a restricted natural host 

range to Mexican lime (Verniere et al., 1998). 

Another Aw-strain, which behaves similarly, has 

recently been discovered in Florida. This strain 

has a restricted host range that includes Mexican 

lime and alemow (Citrus macrophylla) (Sun et al., 

2004). The A-strain is the target of international 

quarantine efforts, in which the development of 

rapid and reliable procedures for the diagnosis of 

this pathogen has been a priority. 

The polymerase chain reaction (PCR) is 

a principle for plant disease diagnosis (Henson and 

French, 1993). However, routine application of 

PCR for detection of plant pathogens can result in 

false-positive diagnosis when PCR primers are 

non-specific to the pathogen. Several sets of 

primers have been developed for diagnosis of Xcc. 

Non-specific amplification of Hartung’s primers 

(Hartung et al., 1993) for detection of Xcc were 

reported by Zaccardelli and Mazzucchi (1997). 

Miyoshi et al. (1998) studied similarity of the 

intergenic spacer region between 16S-23S rRNA 

genes among X. citri subsp. citri, X. campestris 

pv. glycines, X. alfalfae subsp. alfalfae (X. c. pv. 

alfalfae), X. c. pv. physalidicola, X. c. pv. pisi, X. 

c. pv. pruni, X. c. pv. cucurbitae and X. c. pv. 

vesicatoria, and designed primers XCF and XCR 

based on the data from this study. Primers XCF- 

XCR were not only for detection of Xcc but also 

for X. c. pv. glycines. Kingsley et al. (2000) 

developed fluorogenic PCR assay for specific 

detection of Xcc A and A*-strain using the 

forward/reverse primers and probes designed from 

unique RAPD fragment to target 126 bp amplicon. 

Mavrodieva et al. (2004) designed primers, VM3 

and VM4, for real-time PCR by selecting to 

amplify the pthA gene family and run experiments 

to compare with Kingsley’s primers, KF and KR. 

The results showed that Xcc (A, A* and Aw) and 


X. fuscans subsp. aurantifolii (B and C) reacted 

and gave expected product sizes with VM3-VM4 

primers. On the other hand, Kingsley’s primers 

gave prominent band with Xcc A and A*-strain 

but the reaction with A w -strain and X. fuscans 

subsp. aurantifolii (B and C) were inconsistent and 

also gave more primer-dimer products when 

compared with VM3-VM4 primers. The purpose 

of this study was to design specific PCR primers 

of Xcc from genomic DNA especially gene 

XAC2443 (Accession AE011881) in order to 

apply them for detection of this international 

quarantine bacterial pathogen of citrus. 


Bacterial strains 

To obtain original local strains to be used 

in this study, canker lesions on lime (Citrus 

aurantifolia), mandarin (C. reticulata), sweet 

orange (C. sinensis) and leach lime (C. hystrix) 

were collected from leaves, twigs, and fruit from 

each major citrus growing area in Thailand. The 

corky-like raised surface lesions, surrounded by a 

yellow halo were washed in running water for 1-2 

minutes, sprayed with 70% ethyl alcohol, and airdried. 

Each lesion was removed from the leaf and 

cut into 4-5 pieces then soaked in 0.85% NaCl for 

20 minutes. A loop of the suspension was streaked 

onto Fieldhouse and Sasser (FS) agar (Schaad et 

al., 2001) and incubated at 30°C. After 3-4 days, 

plates were examined for small green-colored 

starch hydrolyzing colonies typical of Xcc. 

Promising colonies of Xcc were transferred onto 

nutrient agar (NA) (Schaad et al., 2001) twice and 

stored either in sterile distilled water at room 

temperature or on NA slants at 4°C, and in 50% 

glycerol at -80°C. Several bacterial strains from 

Japan and the United States of America were 

included in this experiment (Table 1). 

PCR primers 

The new primer pair namely 354F-354R

264 


Table 1 Geographical origin, host and year of isolation of strains of Xanthomonas species used in this 

study 

Bacterial strain Geographical origin Host Year 

X. citri subsp. citri 

T1 Kamphaeng Phet Thailand Citrus sinensis 2003 

T3 Chiang Mai Thailand Citrus grandis 2003 

T4 Chiang Mai Thailand Citrus reticulata 2003 

T5 Kamphaeng Phet Thailand Citrus aurantifolia 2003 



T10 Chiang Mai Thailand Citrus grandis 2003 

T13 Kamphaeng Phet Thailand Citrus sinensis 2003 

NT14 Kamphaeng Phet Thailand Citrus reticulata 2003 

NT18 Chiang Mai Thailand Citrus aurantifolia 2003 

NT20 Sukhothai Thailand Citrus aurantifolia 2003 

NT22 Chiang Mai Thailand Citrus grandis 2003 

NT25 Kamphaeng Phet Thailand Citrus reticulata 2003 

OCr1.1 Chiang Rai Thailand Citrus reticulata 2002 

OCr1.2 Chiang Rai Thailand Citrus reticulata 2002 

LCp2.1 Chumphon Thailand Citrus aurantifolia 2002 

LCp2.2 Chumphon Thailand Citrus aurantifolia 2002 

SWRb Ratchaburi Thailand Citrus sinensis 2003 

Fp1-2 Chiang Rai Thailand Citrus grandis 2003 

XCC-32 Shimizu Japan Citrus natsudaidai 1998 

XCC-131 Yui Japan Citrus unshiu 1998 

1258 (Hartung, Xc-322) Saudi Arabia Citrus sp. ND 

1270 (Hartung, Xc-328) Saudi Arabia Citrus sp. ND 

X. fuscans subsp. aurantifolii 

1415 (IBSBF 392) Brazil Citrus limon 1981 

1416 (IBSBF 423) Uruguay Citrus limon 1981 

1417 (IBSBF 1583) Argentina Citrus limon 1990 

1418 (IBSBF 380) Brazil Citrus aurantifolia 1981 


X. fuscans subsp. aurantifolii 



1460 ND ND ND 

1461 ND ND ND 

1463 ND ND ND 

X. alfalfae subsp. citrumelonis 

1267 (X-85, J. Miller) Florida Citrus sp. 1985 

1274 (4600, D. Gabriel) Florida Citrus sp. ND


Table 1 (continued) 

Bacterial strain Geographical origin Host Year 

X. citri subsp. malvacearum 

1318 (ATCC 14982) Uganda Gossypium hirsutum ND 

317 Sukhothai Thailand Gossypium hirsutum 1984 

579 ND Thailand Morus sp. 1986 

584 Sukhothai Thailand Gossypium hirsutum 1986 

1034 Nakhon Sawan Thailand Gossypium hirsutum 1990 

1035 Nakhon Sawan Thailand Gossypium hirsutum 1990 

1037 Lop Buri Thailand Gossypium hirsutum 1990 

1051 Loei Thailand Gossypium hirsutum 1990 

1232 Prachin Buri Thailand Gossypium hirsutum 1993 

X. fuscans subsp. fuscans 

1316 (NCPPB 381) Canada Phasolus vulgaris ND 

X. campestris pv. campestris 

657 Phetchaburi Thailand Brassica oleracea 2004 

X. campestris pv. glycines 

NKR21 Nakhon Ratchasima Thailand Glycine max 2001 

CM 60-1 Nakhon Ratchasima Thailand Glycine max 2002 

No.21-1 Chiang Mai Thailand Glycine max 2002 

RE 07 Khon Kaen Thailand Glycine max 2002 

239 Chachoengsao Thailand Glycine max 1983 

241 Phitsanulok Thailand Glycine max 1982 

281 Phitsanulok Thailand Glycine max ND 

X. campestris pv. glycines 

285 Phitsanulok Thailand Glycine max ND 

728 Chiang Rai Thailand Glycine max 1987 

1204 Songkhla Thailand ND 1992 

1324 Songkhla Thailand Vigna radiata 1994 

Abbreviations: IBSBF, Phytobacteria Culture Collection of Instituto Biological, Campinas, Brazil; ATCC, American Type Culture 

Collection, Manassas, VA; NCPPB, National Collection Plant Pathogenic Bacteria, York, England; ND, not determined. 

was designed from Xanthomonas axonopodis pv. 

citri strain 306 at section no. 259 from 469 sections 

of complete genome. This target region resulted 

from subtractive hybridization (Schaad et al., 

unpublished). The target at position 4411 to 5228 

(partial gene XAC2443) was used for designing 

new primers for classical PCR by using 

DNASTAR software (LASERGENE, Version5.1). 

The sequences of the primers were 354F at position 

4675-4693 (5’-GACGGCGCGGCTCAGGATG- 

3’) and 354R at position 5006-5028 (5’- 

CAGCCCAGCCAACTCAGCACCAG-3’). 

Other primer pairs also evaluated in this 

experiment were designed by Kingsley et al. 

(2000), KF (5’-TCCACTGCATCCCACAT CTG- 

3’), and KR (5’-CAGGTGTACTGCGCTC 

TTCTTG-3’); Mavrodieva et al. (2004), VM3 (5’- 

GCATTTGATGACGCCATGAC-3’), and VM4 

(5’-TCCCTGATGCCTGGAG GATA-3’); and 

Hartung et al. (1993), 2 (5’-CACGGGTGCAAAA 

AATCT-3’), and 3 (5’-TGGTGTCGTCGCTT 

GTA T-3’) which are respectively referred to as

266 

Kingley’s, Mavrodieva’s and Hartung’s primers 

in the article. 

PCR reaction 

PCR was carried out in a 25 µl reaction 

that consisted of 1x PCR buffer, 3mM MgCl 2 for 

354 F/R and 2-3 primers and 2mM MgCl 2 for KF- 

KR and VM3-VM4 primers, 0.1 mM dNTPs, 0.6 

unit Taq DNA polymerase, DNA template 1 µl (50 

ng), 0.4 pmole of each primer for 354 F/R, KF- 

KR and VM3-VM4 primers and 1 pmole for 2-3 

primer. 

The PCR profiles were designed for each 

primer as follow: 1.) 94°C for 10 min and 30 cycles 

of 94°C for 30 sec, 60°C for 30 sec, 72°C for 60 

sec and 72°C for 10 min for 354 F/R primers, 2.) 

94°C for 10 min and 30 cycles of 94°C for 30 

sec, 57°C for 30 sec, 72°C for 60 sec and 72°C for 

10 min for KF-KR and VM3-VM4 primers and 

3.) 95°C for 10 min and 35 cycles of 95°C for 70 

sec, 60°C for 70 sec, 72°C for 60 sec and 72°C for 

10 min for 2-3 primers. 

Primers specificity tests 

Strains of Xanthomonas species in Table 

1 were used for specificity assay by comparing 

354 primers with the Kingsley’s, Mavrodieva’s and 

Hartung’s primers. Ten microliters of PCR product 

of each primer was determined by gel 

electrophoresis on agarose gels in 0.5x TBE buffer 

at concentration of 1% for 354 bp-PCR primers 

and 1.5% for Kingsley’s, Mavrodieva’s and 

Hartung’s primers. 

Sensitivity tests 

Genomic DNA of Xcc strain T7 was 

calculated from the absorbance at 260 nm with 

UV-Visible Spectrophotometer (UV-1601, 

SHIMADZU ® ) and adjusted by ten-fold dilution 

with sterile distilled water from 50 ng to 50 fg for 

sensitivity tests. Cell suspension of Xcc strain T7 

at 0.2 OD of wavelength 600 nm which was about 

10 8 CFU/ml was also used for sensitivity tests with 


the ten-fold serial dilutions. 

Cloning and sequencing of target DNA 

fragment 

Taq polymerase-amplified PCR products 

using primer pair 354 F/R were purified and 

recovered with commercial silica spin column 

(Promega ® ). Cloning reactions were according 

to pCR ® 8/GW/TOPO ® TA Cloning ® Kit 

(Invitrogen ® ). Briefly, the mixture was incubated 

at room temperature for 5 min, mixed with One 

Shot ® Mach1 TM -T1 R Chemically Competent E. 

coli and incubated on ice for 5 min. The cells were 

transformed for 30 sec at 42°C without shaking 

and immediately transferred on ice. A 250 µl 

aliquot of S.O.C. medium were added and 

incubated on a rotary shaker for 1 hr at 37°C. 

The transformed cells were centrifuged and 

suspended in new S.O.C. medium and then 50µl 

was spread onto LB agar containing 100 µg/ml 

spectinomycin. 

Recombinant clones were screened by 

PCR amplification with 354 primers, as described 

above. Sequencing of target DNA product was 

commercially provided by BSU (Bioservice Unit) 

using GW1 and GW2 as sequencing primers. The 

nucleotide sequences were analyzed by the Vector 

NTI ® Advance 9.0 software (Invitrogen ® ). 

Southern blot hybridization: The 354bp 

PCR fragment was amplified by using 354 F/R 

primers and used as the target DNA probe. The 

method for recovery the DNA fragment from gel 

was modified from Yue and Orban (2001). The 

DNA fragment was excised from 0.7% agarose 

gel in 0.5x TBE with a razor blade. The gel slice 

was ground with a sterile pestle in a microtube 

and 300 µl of phenol was added. After vigorously 

mixing with a vortex, the suspension was 

centrifuged at 10,000 rpm for 10 min and then 200- 

300 µl of the supernatant was collected and added 

to 0.5 volume of 7.5M ammonium acetate and 2.5 

volume of absolute ethanol. The supernatant was 

centrifuged at 10,000 rpm for 15 min and the pellet

was collected and washed with 70% ethyl alcohol. 

After centrifuging at 10,000 rpm for 10 min, the 

pellet was dried and suspended in 20-30 µl of 

sterile distilled water. 

The purified target fragment from the 

previous experiment was labeled with 

digoxigenin-11-dUTP (Dig-11-dUTP) by using 

10xDIG-11-dUTP mixs. The procedure was as 

follows: the DNA template was diluted to 50 ng 

and prepared for the 50 µl labeling reaction 

containing 1µl of DNA template, 5µl of 10x PCR 

buffer (200mM TrisHCl, 500mM KCl, 20mM 

MgCl 2), 5µl of 10x PCR DIG labeling, 2µl of 

each 20 pmole/µl primer and 1µl of Taq DNA 

Polymerase (5 units/µl). The labeling PCR 

product was separated as described above. The 

labeled DNA probe was stored at -20°C and 

denatured by heating in boiling water for 10 min 

and immediately chilled on ice for 5 min before 

use. 

The agarose gel containing PCR products 

was depurinated in 0.25% HCl for 30 min and 

neutralized in 0.4M NaOH for 15 min, and 

transferred to Highbond N + nylon membrane by 

alkaline 0.4N NaOH. DNAs were fixed under UV 

transilluminator for 2.5 min to crosslink the DNA 

to the membrane, and washed as suggested by its 

manufacturer (Roche ® ). The membrane was 

placed into a hybridization bottle containing 3 ml 

hybridization solution containing 1% blocking 

solution. After incubating for 1 hr at 65°C the 

hybridization solution was replaced with a new 

hybridization solution containing labeled DNA 

probe and incubated at 65°C for additional 18-24 

hr. The membrane was removed and washed on a 

rotary shaker in solution I (2xSSC, 0.1% SDS) at 

65°C for 5 min. The solution was replaced and 

the membrane was washed for additional 15 min. 

Finally the membrane was washed twice 

consecutively with solution II (1xSSC, 0.1%SDS) 

and solution III (0.5xSSC, 0.1%SDS) for 15 min 

each at 65°C. 


After being washed briefly in washing 

buffer and 30 min in 1% blocking buffer, the 

membrane was transferred to anti-digoxigenin 

alkaline phosphatase conjugated (Roche ® ) and 

incubated on a rotary shaker for 45 min at room 

temperature. The membrane was washed two 

times with washing buffer for 15 min before being 

transferred to plastic bag. After adding 500 µl 

CDP-Star solution, the bag was sealed, placed into 

a Kodak ® x-ray cassette and moved to the dark 

room for the detection step. X-ray film was cut to 

the proper size and placed over of the membrane 

and the closed cassette for 30-60 sec. The film 

was transferred to the developer solution until the 

band was visible. After washing briefly in water, 

the film was removed to the fixer solution until 

the background was clear. Finally the film was 

washed briefly in water and dried at room 

temperature before being photographed. 

RESULTS 

Bacterial strains 

X. citri subsp. citri strains of Thailand 

were isolated from different kinds of Citrus spp., 

namely, mandarin, (C. reticulata), lime (C. 

aurantifolia), pummelo (C. grandis) and sweet 

orange (C. sinensis) from major citrus producing 

provinces of Thailand (Table 1). Total X. citri 

subsp. citri strains in this study were 19 strains 

from Thailand, 2 strains from Japan, and 2 strains 

from Saudi Arabia. Other xanthomonads included 

in this study consisted of 10 strains of X. fuscans 

subsp. aurantifolii, 2 strains of X. alfalfae subsp. 

citrumelonis, 1 strain of X. campestris pv. 

campestris, 11 strains of X. campestris pv. glycines, 

9 strains of X. citri subsp. malvacearum and 1 

strain of X. fuscans subsp. fuscans. 

PCR specificity 

The specific 354-bp PCR fragment was 

amplified with 354 F/R primers from all 23 strains 

of Xsc (Table 2 and Figure 1A). No fragment of

268 


Table 2 Comparison of specificity test of 354 F/R, VM3-VM4, KF-KR and 2-3 primers by classical 

PCR. Specific amplification product of each primer pair was determined on agarose gel at 

concentration of 1% for 354 F/R primers and 1.5% for VM3-VM4, KF-KR and 2-3 primers in 

0.5x TBE buffer. 

Xanthomonas species PCR primers 

354F-354R VM3-VM4 KF-KR 2-3 

X. citri subsp. citri (23) y 23 z 23 23 23 

X. fuscans subsp. aurantifolii (10) 0 5 0 0 

X. alfalfae subsp. citrumelonis (2) 0 0 0 0 

X. citri subsp. malvacearum (9) 0 9 0 9 

X. fuscans subsp. fuscans (1) 0 0 0 0 

X. campestris pv. campestris (1) 0 0 0 0 

X. campestris pv. glycines (11) 0 9 0 0 

y Total number of Xanthomonas species in specificity test. 

z Total number of classical PCR positive results of each Xanthomonas species and 0 = negative result. 

Figure 1 A) PCR amplification products of 354 F/R primers on 1% agarose gel 0.5x TBE buffer. B) 

Southern blot hybridization with 354 bp probe of Xanthomonas species. Lane 1) DNA 

marker 1 kb (Biolab ® ); 2-5) X. citri subsp. citri: T7, J131, 1258, 1270; 6-11) X. fuscans 

subsp. aurantifolii: 1415, 1416, 1419, 1420, 1360, 1361; 12-13) X. alfalfae subsp. citrumelonis: 

1267, 1274; 14) X. campestris pv. glycines: NKR 21; 15) X. citri subsp. malvacearum: 1318; 

16) X. fuscans subsp. fuscans: 1316; 17) X. campestris pv. campestris: 657.

expected size was amplified with 354 F/R primers 

from 34 strains of other xanthomonads including 

10 strains of X. fuscans subsp. aurantifolii, 2 

strains of X. alfalfae subsp. citrumelonis, 9 strains 

of X. citri subsp. malvacearum, 1 strain of X. 

fuscans subsp. fuscans, 1 strain of X. campestris 

pv. campestris and 11 strains of X. campestris pv. 

glycines (Table 2). 

Other primer pairs, VM3-VM4, KF-KR 

and 2-3, also amplified expected fragment of PCR 

product from all strains of Xcc (Table 2). 

However, VM3-VM4 primers still provided the 

expected fragment from 5 strains of X. fuscans 

subsp. aurantifolii including 1 strain of B-strain 


and 4 strains of C-strain, 9 strains of X. citri subsp. 

malvacearum and 9 strains of X. campestris pv. 

glycines. The KF-KR primers were not crossreacted 

to other xanthomonads. The 2-3 primers 

also gave expected fragment from 9 strains of X. 

citri subsp. malvacearum. 

PCR sensitivity 

Sensitivity of 354 F/R primers for 

detection of viable cells of Xcc and purified total 

DNA of Xcc strain T7 were 70 cells per µl and 50 

pg per µl (Figure 3), respectively by PCR reaction 

and amplification program were followed as 

previously. 

Xsc(T7) 1 GACGGCGCGGCTCAGGATGCTGCTAAGGGAGCTGGACGCGCGAAAGGTAATCTGGAAGAC 60 

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 

AE011881.1 4675 GACGGCGCGGCTCAGGATGCTGCTAAGGGAGCTGGACGCGCGAAAGGTAATCTGGAAGAC 4734 

Xsc(T7) 61 CAGCTGCGTGTTGCCAACGAGCTACTGCGTGGC * TTGCAAATCCTTGGCATTAGCGACGAA 120 

|||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||| 

AE011881.1 4735 CAGCTGCGTGTTGCCAACGAGCTACTGCGTGGT * TTGCAAATCCTTGGCATTAGCGACGAA 4794 

Xsc(T7) 121 GCCGAAGCGTTGGAGCAGGACCTCACCGGGATCTTAAATGCCTTTTCAAAGTCGATTCTG 180 

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 

AE011881.1 4795 GCCGAAGCGTTGGAGCAGGACCTCACCGGGATCTTAAATGCCTTTTCAAAGTCGATTCTG 4854 

Xsc(T7) 181 CAAAGTGAAAGAGGGATCGCGACTGCTGAGGAGGCTAGACGCGAGCAGGCTCTCAATACG 240 

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 

AE011881.1 4855 CAAAGTGAAAGAGGGATCGCGACTGCTGAGGAGGCTAGACGCGAGCAGGCTCTCAATACG 4914 

Xsc(T7) 241 CTTGTTGCATTTCTAATGAGCTTCGCGAGCCGAAGCGGCGTACGTGATCGACTGAACATC 300 

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 

AE011881.1 4915 CTTGTTGCATTTCTAATGAGCTTCGCGAGCCGAAGCGGCGTACGTGATCGACTGAACATC 4974 

Xsc(T7) 301 TTTACCACCAACTATGACAGGCTAATCGAAGCTGGTGCTGAGTTGGCTGGGCTG 354 

|||||||||||||||||||||||||||||||||||||||||||||||||||||| 

AE011881.1 4975 TTTACCACCAACTATGACAGGCTAATCGAAGCTGGTGCTGAGTTGGCTGGGCTG 5028 

Figure 2 Comparison of nucleotide sequences of PCR product fragment of 354 F/R primers from 

Xanthomonas citri subsp. citri (T7 strain) and Xanthomonas axonopodis pv. citri strain 306 

gene XAC 2443 (Accession AE011881) with BlastN program showed 99.7% similarity 

* non-similar nucleotide

270 

Southern blot hybridization 

The amplified PCR products from all 

strains of Xcc (Table 2) were hybridized with 354bp 

probe but not with other xanthomonads (Figure 

1B). 

Sequencing of target PCR product 

The 354 bp, expected PCR fragment 

from Xcc was amplified by 354 F/R primers. The 

sequences obtained from 354 F/R cloned were 

blast(N) in Genbank database at National Center 

for Biotechnology Information (http://www.ncbi. 

nlm.nih.gov/BLAST/). Searching results showed 

that sequences of 354 bp of expected product were 

99.7% similar to sequence of Xac strain 306 gene 

XAC 2443 (Accession AE011881) (Figure 2). 

DISCUSSION 

Xanthomonas citri subsp.citri (Xcc) is 

the causal agent of citrus bacterial canker disease, 

an important pathogen of Citrus species, and it is 

important to international phytosanitary quarantine 

in many citrus producing countries worldwide 

(OEPP/EPPO, 2005). Other bacterial citrus 

pathogens, X. fuscans subsp. aurantifolii and X. 

alfalfae subsp. citrumelonis are closely related to 

Xcc and X. alfalfae subsp. citrumelo and have been 


easily misidentified as Xcc (Schoulties et al., 

1987). 

Effective control and eradication of citrus 

canker needs a rapid, specific, and sensitive 

detection techniques. The polymerase chain 

Figure 3 PCR amplification products of 354 F/ 

R primers on 1% agarose gel 0.5x TBE 

buffer. Lane 1) marker DNA 1 kb 

(Fermentas ? ), 2-7) chromosomal DNA 

of X. citri subsp. citri at concentration 

from 50 ng to 50 fg per microliter by 

ten-fold dilution. 

Table 3 Sensitivity of classical PCR for detecting viable cells and purified DNA of Xanthomonas citri 

subsp. citri strain T7. 

Dilution a Cell/µl PCR results c DNA concentration b PCR results c 

0.1 OD 600 nm 7.0×10 4 + 50 ng/µl + 

10 -1 7.0×10 3 + 5 ng/µl + 

10 -2 7.0×10 2 + 500 pg/µl + 

10 -3 7.0×10 + 50 pg/µl + 

10 -4 7.0 - 5 pg/µl - 

10 -5 0 - 500 fg/µl - 

50 fg/µl - 

a Cell suspension of Xcc was adjusted to turbidity 0.1 O.D. of wavelength 600nm which gave 7.0×104 cell/µl and ten-fold 

serially diluted to 10-5 . The number of cell per microliter was counted by haemacytometer. 

b DNA of Xcc was adjusted by ten-fold dilution from 50 ng/µl to 50 fg/µl. 

c PCR specific amplification with 354 F/R primers were performed following the reaction mix and amplification program in 

methods. Presence (+) or absence (-) of unique predicted PCR product size after agarose gel electrophoresis.

eaction technique has been used for rapid and 

reliable detection for many plant pathogens 

(Henson and French, 1993). Thus, this PCR 

technique is suitable for routine assay in 

international quarantine which requires a rapid and 

sensitive method for routine assay. Plasmid (pthA 

gene family) and chromosomal DNA of Xcc have 

been used to design specific PCR primers (Hartung 

et al., 1993, Kingsley et al., 2000, Mavrodieva et 

al., 2004) for detection of Xcc. 

In this experiment, new specific primers, 

354 F/R primers, were designed from a fragment 

in chromosomal DNA of Xcc by substractive 

hybridization that translated to conserved 

hypothetical protein (gene XAC2443). The results 

were that 354 primers showed specific DNA 

amplification of all strains of Xcc. These Xcc 

strains were isolated from different hosts and 

geographical areas in Thailand including strains 

from Japan and Saudi Arabia which gave the 

expected 354 bp PCR fragment but not from other 

xanthomonads (Table 2). This is the first report 

of using sequences from a conserved hypothetical 

protein in chromosomal DNA to design specific 

PCR primers for detection of Xcc. The PCR 

primers from conserved hypothetical protein 

region showed more specificity than primers from 

plasmid DNA (VM3-VM4 and 2-3 primers, Table 

2). 

The primers designed from chromosomal 

DNA, KF-KR, had specific amplification with all 

strains of Xcc in this experiment. The primers 

also produced prominent band with Xcc A and A*strain 

but the reactions with A w -strain and X. 

fuscans subsp. aurantifolii (B and C-strain) were 

inconsistent and also gave more primer-dimer 

products (Mavrodieva et al., 2004). At present, 

PCR product fragment of KF-KR primers still 

cannot be identified when searching with Genbank 

database by using BLAST program provided by 

National Center for Biotechnology Information 

(NCBI). 

The primers designed from plasmid 


DNA, 2-3 primers and pthA gene family, VM3- 

VM4 primers, amplified not only all strains of Xcc 

but also other xanthomonad strains. Primers 2-3 

cross-reacted with X. citri subsp. malvacearum and 

primers VM3-VM4 cross-reacted with X. fuscans 

subsp. aurantifolii, X. citri subsp. malvacearum 

and X. campestris pv. glycines (Table 2) because 

these primers were designed for detection of 

avirulence or pathogenicity genes which are 

commonly found in the genus Xanthomonas 

(Gabriel, 1997). 

The plasmid DNA has been reported as 

being easily cured, frequently mutants within the 

internal sequence and not present in all pathogens 

(Miyoshi, 1998). The propose of VM3-VM4 

primers is to develop universal detection of Xsc 

and X. fuscans subsp. aurantifolii. However, 

results in this experiment showed that the primers 

did not completely detect all target strains of 

xanthomonad. They detected one third from X. 

fuscans subsp. aurantifolii (B-strain), all 4 strains 

of X. fuscans subsp. aurantifolii (C-strain) but did 

not detect any of X. fuscans subsp. aurantifolii (Dstrain). 

Nine strains of X. citri subsp. malvacearum 

and X. campestris pv. glycines also reacted with 

VM3-VM4 primers. The pthA, pthB and pthC, 

members of pthA gene family, belong to a family 

of avirulence or pathogenicity genes found in the 

genus Xanthomonas (the avrBs3/pthA gene family; 

Leach and White, 1996; Gabriel, 1997) and these 

may be transferred horizontally on plasmids 

between Xcc and X. fuscans subsp. aurantifolii 

(Brunings and Gabriel, 2003). The results of this 

experiment also confirmed that the avrBs3/pthA 

gene family is distributed in X. citri subsp. 

malvacearum and X. campestris pv. glycines. 

The assay of 354 F/R primers with 

classical PCR had the ability to detect a lower limit 

of about 70 CFU/µl of viable cells and the lower 

limit of detection of 50 pg/µl of purified Xcc total 

DNA. Sensitivity of other primers to detect viable 

cells of Xcc and purified Xcc total DNA were 10 

CFU/µl and 25 pg/µl for 2-3 and 10 CFU/µl and 1

272 

pg/µl for VM3-VM4 primers. The target PCR 

product of 354 primers was located in 

chromosomal DNA of which Xcc carries a single 

copy per cell, lower than plasmid that Xcc carries 

multiple copies per cell (Mavrodieva et al., 2004). 

However, the novel 354 F/R primers gave more 

specific and reliable detection of Xcc than other 

primers. Real-time PCR techniques, which are 

based on hybridization of specific probe 

sequences, are faster and have higher sensitivity 

and specificity than classical PCR (Schaad et al, 

2002). Combination of the new specificity primers 

(354 F/R) with real-time PCR technique will 

improve the efficacy of Xcc detection to be more 

specific, sensitive, accurate, reliable, and faster in 

the future work. 


The authors are thankful for the financial 

support from the Thailand Research Fund under 

the Royal Golden Jubilee Ph.D. Program and 

Kasetsart University Research and Development 

Institute. The authors are also thankful to the kind 

support of bacterial cultures from Ms. Nuttima 

Boonwatana, Plant Pathology Research Group, 

Plant Protection Research and Development 

Office, Department of Agriculture and Dr. Srimek 

Chowpongpang, Department of Plant Pathology, 

Kasetsart University, for his kind suggestions and 

support for cloning and sequencing of Xcc. 


Brunings, A.M. and D.W. Gabriel. 2003. 

Xanthomonas citri: Breaking the surface. Mol. 

Plant Pathol. 4: 141-157. 

Gabriel, D.W. 1997. Targeting of protein signals 

from Xanthomonas to the nucleus. Trends 

Plant Sci. 2: 204-206. 

Gabriel, D.W., M.T. Kingley, J.E. Hunter and T.R. 

Gottwald. 1989. Reinstatement of 

Xanthomonas citri (ex Hasse) and X. phaseoli 


(ex Smith) to species and reclassification of 

all X. campestris pv. citri strains. Int. J. Syst. 

Bacteriol. 39: 14-22. 

Hasse, C.H. 1915. Pseudomonas citri, the cause 

of citrus canker. J. Agric. Res. 4: 97-100. 

Hartung, J.S., J.F. Daniel and P. Pruvost. 1993. 

Detection of Xanthomonas axonopodis 

pv.citri by the polymerase chain reaction 

method. Appl. Environ. Microbiol. 59: 1143- 

1148. 

Henson, J. and R. French. 1993. The polymerase 

chain reaction and plant disease diagnosis. 

Annu. Rev. Phytopathol. 31: 81-109. 

Kingsley, M.T. and L.K. Fritz. 2000. Identification 

of citrus canker pathogen Xanthomonas 

axonopodis pv. citri A by fluorescent PCR 

assays.(Abstr.) Phytopathology 90 (suppl.): 

S42. 

Leach, J.E. and F.F. White. 1996. Bacterial 

virulence genes. Annu. Rev. Phytopathol. 34: 

153-179. 

Mavrodieva, V., L. Levy and D.W. Gabriel. 2004. 

Improved sampling methods for real-time 

polymerase chain reaction diagnosis of citrus 

canker from field samples. Phytopathology 

94: 61-68. 

Miyoshi, T., H. Sawada, Y. Tachibana and I. 

Matsuda. 1998. Detection of Xanthomonas 

campestris pv. citri by PCR using primers 

from the spacer region between the 16s and 

23s rRNA genes. Phytopathol. Soc. Jpn. 64: 

249-254. 

Riker, A.J., F.R. Jones and M.C. Davis. 1935. 

Bacterial leaf spot of alfalfa. J. Agric. Res. 

51, 177-182. 

OEPP/EPPO. 2005. List of A1 pests regulated as 

quarantine pests in the EPPO region. Available 

source: http://www.eppo.org/QUARANTINE/ 

quarantine.htm, May 17, 

Schaad, N.W., J.B. Jones and W. Chun. 2001. 

Laboratory Guide for Identification of 

Plant Pathogenic Bacteria, 3 rd ed. American 

Phytopathological Society, St. Paul, MN 373.

Schaad, N.W., D. Opgenorth and P. Gush. 2002. 

Real-Time Polymerase Chain Reaction for 

One-Hour On-Site Diagnosis of Pierce’s 

Disease of Grape in Early Season 

Asymptomatic Vines. Phytopathology 92: 

721-728. 

Schaad, N.W., E. Postnikova, G.H. Lacy, A. 

Sechler, I. Agarkova, P.E. Stromberg, V.K. 

Stromberg and A.K. Vidaver. 2005. 

Reclassification of xanthomonas species 

pathogenic on citrus. Syst. and Appl. 

Microbiol. 28: 494-518 

Schaad, N.W., E. Postnikova, G.H. Lacy, A. 

Sechler, I. Agarkova, P.E. Stromberg, V.K. 

Stromberg and A.K. Vidaver. 2006. Emended 

classification of xanthomonad pathogens on 

citrus. Syst. and Appl. Microbiol. 29:690- 

695. 

Schoulties, C.L., E.L. Civerolo, J.W. Miller, R.E. 

Stall, C.J. Krass, S.R. Poe and E.P. DuCharme. 

1987. Citrus canker in Florida. Plant Dis. 71: 

388-395. 

Schubert, T.S., S.A. Rizvi, X. Sun, T.R. Gottwald, 

J.H. Graham and W.N. Dixon. 2001. Meeting 

challenge of eradicating citrus canker in 


Florida-again. Plant Dis. 85: 340-356. 

Sun, X., R.E. Stall, J.B. Jones, J. Cubero, T.R. 

Gottwald, J.H. Graham, W.N. Dixon, T.S. 

Schubert, P.H. Chaloux, V.K. Stromberg, G.H. 

Lacy and B.D. Sutton. 2004. Detection and 

characterization of a new strain of citrus 

canker bacteria from Key/Mexican lime and 

alemow in South Florida. Plant Dis. 88: 1179- 

1188. 

Verniere, C., J.S. Hartung, O.P. Pruvost, E.L. 

Civerolo, A.M. Alvarez, P. Maestri and J. 

Luisetti. 1998. Characterization of 

phenotypically distinct strains of 

Xanthomonas axonopodis pv. citri from 

Southwest Asia. Eur. J. Plant Pathol. 104: 

477-487. 

Yue, H.G. and L. Orban. 2001. Rapid isolation of 

DNA from fresh and preserved fish scales for 

polymerase chain reaction. Mar. Biotechnol. 

3: 199-204. 

Zaccardelli, M. and U. Mazzucchi. 1997. 

Shortcomings of PCR and protein 

electrophoregrams for the identification of 

Xanthomonas campestris pv.citri groups. 

Phytopath. Medit. 36: 12-18.


Soil-to-Plant Transfer of Radiocaesium in Thailand 

Thitika Thammavech and Teerasak Veerapaspong* 

ABSTRACT 

Soil-to-plant transfer factors (TF) of radiocaesium-137 were estimated by considering soil 

properties of 51 provinces in Thailand, and by using the model of Absalom. According to our study, the 

Absalom model could estimate average TF values to be 0.0852 ± 0.0475. Compared with average 

measured TF values which was 0.1289 ± 0.0529, it was found that calculated TF values decreased with 

increasing pH, clay contents and exchangeable K + . The corresponding calculated TF values increased 

with increasing organic matter contents and NH 4 + concentrations. Statistical analysis showed that Relative 

Euclidean Difference (RED) was 0.238, reliability index (k) was 0.661 and geometrically intuitive 

reliability index (k g) was 1.97, which confirmed that the Absalom model was reasonably accurate. 

Calculated TF values by the Absalom model were in good agreement with the measured ones. However, 

calculated TF values were found to be significantly different from the measured ones for some provinces 

in Thailand. The parameters used in the Absalom model needed to be modified to suitably match soil 

properties in Thailand. 

Key words: transfer factor, Absalom model, radiocaesium; soil properties 

INTRODUCTION 

Radionuclides produced by nuclear 

explosion and nuclear facilities have the potential 

to be released into the atmosphere. These nuclides 

are part of the fallout which is deposited on the 

ground and reach human bodies via food chain 

(Eisenbud, 1973). Among deposited radionuclides, 

radiocaesium ( 137 Cs, half life is 30 years) is the 

dominant fission product which has a high relative 

mobility in the soil–plant system, long term 

bioavailability, high radiotoxicity, continuing to 

cycle through the soil plant-animal system, and is 

longlived. The plant uptake of deposited 137 Cs 

from soil, commonly expressed as soil to plant 

transfer factor (TF) is widely used while 

Department of Physics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: fscitsv@ku.ac.th 

calculating the radiological humus dose via the 

ingestion pathway. 

Absalom et al. (2001) presented a model 

which predicted the radiocaesium soil to plant 

transfer factor (TF) on the basis of easily measured 

soil characteristics (pH, clay content, organic 

matter content, exchangeable K + + 

and NH4 concentration). In the present work, data of soil 

properties and 137Cs activity concentrations in soil 

and grass of some selected provinces in Thailand 

were collected and were used as input parameters 

to calculate transfer factor (TF) in the Absalom 

model. Finally, the calculated TF values were 

compared with the measured TF values to test 

whether the Absalom model could be applied to 

the soil characteristics in Thailand. 



Model descriptions 

Absalom et al. (1999) presented a semimechanistic 

model, which predicted activity 

concentrations of 137 Cs in plants. The model 

utilized as input soil characteristic parameters 

including clay content and exchangeable K + . In 

2001, Absalom et al. (2001) developed the model 

which accounted for the effect of organic matter 

on 137 Cs adsorption by soil and uptake by plants. 

Therefore, radiocaesium bioavailability is strongly 

influenced by soil properties such as pH, clay 

content, organic matter and exchangeable K + 

(Cremers et al., 1988). This model can be applied 

to mineral and organic soils simultaneously to 

provide a more generally applicable simulation of 

137 Cs dynamics. The model of Absalom et al. 

(2001) assumed that 137 Cs adsorption occurred 

exclusively on both clay and humus surfaces, 

however, fixation only occured on clay, and the 

radiocaesium adsorbed on the organic fraction was 

not subject to fixation. The relationship between 

adsorbed and solution of 137 Cs was described by a 

labile 137 Cs distribution coefficient (k dl, dm 3 kg -1 ) 

which was estimated as a function of clay content 

and exchangeable K + . Plant uptake of 

radiocaesium was described by a concentration 

factor (CF, Bq kg -1 plant/Bq dm -3 soil solution) 

which was related to solution K + concentration 

([m K], moles dm -3 ). 

Data sources 

According to input parameters, the data 

referred to six different regions in Thailand. 

Samples were collected from several provinces in 

the north, northeast, east, west, middle and south 

of Thailand. Each soil sample consisted of 

subsamples collected from an area of 100 m 2 . The 

samples were taken from 0 to 10 cm upper soil 

layer. Specific soil parameters in each province 

were available for comparison with 137 Cs 

concentration in the grass samples. 


Five independent soil properties (pH, 

clay content, organic matter, exchangeable K + and 

+ 

NH4 concentration) and initial 137Cs activity in 

soil were required as the model input parameters 

in the Absalom model assuming certain days after 

a deposition of 137Cs in soil for the prediction of 

TF values in the selected regions. Organic matter 

(OM) content was calculated as OM = organic 

carbon × 1.724 (Nelson and Sommers, 1982). The 

five values (pH, clay content, organic matter, 

exchangeable K + + 

and NH4 concentration) in Table 

1 (LLD, 1988) are used as the input parameters to 

calculate transfer factor of soil-to-plant (here, it 

was grass) in the model. The soil and grass were 

dried and homogenized before being analysed. 

137Cs activities in soil and grass, measured by a 

Hyperpure Germanium gamma-ray detector 

(HPGe), are also shown in Table 1 

(Itthipoonthanakorn). 


Since the Absalom model takes into 

account the time-dependent changes in TF due to 

radiocaesium fixation, the calculations were 

performed assuming 365 days after uniform 

deposition of a certain amount of 137 Cs (Bq m -2 ) 

in soil. The same parameters as in the model were 

used in the calculations. 

Predicted and observed 137 Cs transfer 

factor (TF) values for grass are given in Table 2 

and Figure 1. 

Calculated TF values of 137 Cs from soil 

to grass grown in tropical Thailand are shown in 

Figures 2-6 compared to different functions of soil 

properties. It can be seen from Figures 2-4 that 

the calculated TF values decrease with increasing 

pH, clay content and exchangeable K + . The 

corresponding calculated TF values increase with 

increasing organic matter content and NH 4 + 

concentration, as shown in Figures 5-6.

276 


Table 1 Soil properties and 137 Cs activities in soil and grass of some selected provinces in Thailand. 

Region Province pH Clay Organic Ex-K + [NH + 

4 ] 137Cs activity 137Cs activity 

content matter (cmolc kg-1 ) (×10-5 ) concentration in concentration in 

(%) (%) (mol dm -3 ) soil (Bq kg -1 ) a grass (Bq kg -1 ) a 

North 1.Chaiang Rai 4.3 8.0 0.914 0.10 2.40 1.669 0.108 ± 0.018 

2.Chaiang Mai 5.3 15.6 1.810 0.20 7.70 0.989 ± 0.235 0.086 ± 0.022 

3.Nakhon Sawan 8.2 30.7 2.879 0.20 34.30 0.813 ± 0.182 0.068 ± 0.021 

4.Phayao 5.7 9.5 1.379 0.10 5.00 1.171 ± 0.269 0.066 ± 0.024 

5.Phichit 4.5 45.0 2.689 0.20 18.50 0.685 ± 0.223 0.118 ± 0.061 

6.Phetchabun 5.9 6.0 1.672 0.10 4.00 0.619 ± 0.140 0.060 ± 0.031 

7.Phrae 5.1 12.0 2.069 0.10 6.00 1.033 ± 0.252 0.081 ± 0.028 

8.Uthai Thani 4.8 13.5 3.448 0.10 5.60 1.150 ± 0.304 0.155 ± 0.024 

Central 9.Bangkok 4.2 61.5 0.879 0.60 25.50 1.197 ± 0.480 0.078 ± 0.023 

10.Kanchanaburi 4.7 36.5 0.759 0.10 8.40 0.684 ± 0.158 0.088 ± 0.033 

11.Chai Nat 6.0 19.8 0.345 0.10 6.30 0.676 ± 0.345 0.050 ± 0.009 

12.Nakhon Nayok 5.1 44.9 0.172 0.20 10.60 0.734 ± 0.218 0.097 ± 0.030 

13.Nakhon Pathom 5.0 65.4 4.241 0.50 29.40 0.997 ± 0.327 0.073 ± 0.037 

14.Nonthaburi 7.2 52.3 0.721 0.49 5.31 0.953 ± 0.284 0.141 ± 0.051 

15.Pathum Thani 4.2 46.0 1.569 0.20 20.30 0.898 ± 0.312 0.111 

16.Ratchaburi 4.8 65.1 9.775 0.30 36.50 0.474 0.104 ± 0.019 

17.Lop Buri 7.8 52.0 2.534 0.70 84.00 0.969 ± 0.242 0.172 ± 0.048 

18.Samut Prakan 5.3 74.5 1.827 1.00 28.10 0.519 0.068 ± 0.026 

19.Saraburi 6.6 86.0 1.327 0.30 54.30 0.977 ± 0.244 0.162 ± 0.048 

20.Sing Buri 5.9 44.5 2.155 0.30 27.00 0.630 ± 0.046 0.079 ± 0.043 

21.Ang Thong 5.0 79.6 3.827 0.50 32.20 0.902 ± 0.327 0.127 ± 0.034 

22.Ayuthaya 5.0 65.1 1.207 0.30 25.40 0.955 ± 0.268 0.139 ± 0.064 

North-East 23.Kalasin 6.6 10.0 0.034 0.03 0.60 0.834 ± 0.167 0.097 ± 0.027 

24.Khon Kaen 6.0 5.8 1.379 0.10 5.20 1.456 ± 0.113 0.170 ± 0.032 

25.Chaiyaphum 4.7 8.7 0.241 0.03 2.80 0.643 ± 0.157 0.075 ± 0.024 

26.Nakhon Phanom 5.4 6.1 2.862 0.20 5.30 0.963 ± 0.150 0.106 ± 0.033 

27.Maha Sarakham 5.4 2.5 0.931 0.10 2.90 0.791 ± 0.139 0.099 ± 0.024 

28.Mukdahan 5.0 3.6 3.069 0.10 4.80 0.497 ± 0.158 0.101 ± 0.034 

29.Yasothon 5.2 10.8 0.162 0.42 3.58 0.541 ± 0.170 0.082 ± 0.022 

30.Roi Et 5.3 6.6 0.103 0.03 1.00 0.769 ± 0.176 0.132 ± 0.028 

31.Loei 6.1 6.3 0.345 0.07 1.63 0.705 ± 0.133 0.070 ± 0.026 

32.Si Sa Ket 5.0 17.0 0.914 0.03 3.30 0.329 0.098 ± 0.029 

33.Sakon Nakhon 5.9 11.0 8.068 0.40 23.50 0.906 ± 0.243 0.106 ± 0.024 

34.Surin 4.3 10.7 1.862 0.10 7.40 0.762 ± 0.181 0.131 ± 0.029 

35.Nong Bua Lam Phu 4.1 7.9 0.197 0.19 0.92 1.140 ± 0.211 0.127 ± 0.030 

36.Ubon Ratchathani 4.9 2.0 0.414 0.10 1.58 0.681 ± 0.122 0.070 ± 0.016 

East 37.Chachoengsao 5.5 2.8 0.793 0.10 1.60 0.731 ± 0.216 0.108 ± 0.041 

38.Chon Buri 5.1 6.6 0.707 0.10 1.60 1.659 ± 0.265 0.105 ± 0.028 

39.Prachin Buri 5.8 4.8 0.707 0.05 1.90 0.646 ± 0.184 0.124 ± 0.029 

40.Sa Kaeo 4.6 7.5 0.271 0.53 5.30 0.5989 ± 0.182 0.095 

West 41.Prachuap Khiri Khan 7.3 1.5 1.741 0.10 4.20 0.403 0.109 ± 0.030 

42.Phetchaburi 7.1 4.0 0.155 0.10 1.40 0.994 ± 0.288 0.098 ± 0.026 

South 43.Krabi 4.3 8.6 2.327 0.10 3.40 0.876 ± 0.350 0.122 ± 0.025 

44.Trang 6.0 11.0 2.638 0.10 5.40 0.726 ± 0.252 0.058 ± 0.025 

45.Nakhon Si Thammarat 4.7 19.0 2.276 0.10 4.70 0.811 ± 0.230 0.071 ± 0.055 

46.Narathiwat 4.3 14.2 8.448 0.30 34.30 1.341 ± 0.288 0.204 ± 0.027 

47.Pattani 6.3 8.3 0.586 0.10 1.60 0.742 ± 0.083 0.163 ± 0.035 

48.Phangnga 5.9 7.0 1.879 0.10 2.60 1.210 ± 0.164 0.121 ± 0.026 

49.Phuket 4.6 18.5 3.293 0.10 4.50 1.132 ± 0.285 0.151 ± 0.053 

South 50.Songkhla 4.6 8.0 1.017 0.10 2.20 1.132 ± 0.262 0.070 ± 0.023 

51.Satun 4.8 14.5 4.207 0.30 6.30 1.076 ± 0.301 0.040 ± 0.018 

a (average value ± standard error) 

Source: Land Development Department or LDD (1988)


Table 2 Measured and calculated TF values of some provinces in Thailand. 

Region Province Measured TF value Calculated TF value 

North 1.Chaiang Rai 0.0647 0.0677a 2.Chaiang Mai 0.0871 0.0448b 3.Nakhon Sawan 0.0830 0.0790 

4.Phayao 0.0562 0.0738a 5.Phichit 0.1725 0.1124 

6.Phetchabun 0.0975 0.0837 

7.Phrae 0.0783 0.0918a 8.Uthai Thani 0.1347 0.1053 

9.Bangkok 0.0655 0.0350b 10.Kanchanaburi 0.1289 0.1242 

central 11.Chai Nat 0.0743 0.0640 

12.Nakhon Nayok 0.1318 0.0550b 13.Nakhon Pathom 0.0735 0.0479 

14.Nonthaburi 0.1482 0.0095b 15.Pathum Thani 0.1237 0.1256a 16.Ratchaburi 0.2195 0.1527b 17.Lop Buri 0.1772 0.0329b 18.Samut Prakan 0.1320 0.0160b 19.Saraburi 0.1663 0.1173 

20.Sing Buri 0.1248 0.0613b 21.Ang Thong 0.1410 0.0543b 22.Ayuthaya 0.1459 0.0787b North-East 23.Kalasin 0.1159 0.0792 

24.Khon Kean 0.1166 0.0856 

25.Chaiyaphum 0.1161 0.1876a 26.Nakhon Phanom 0.1101 0.0635b 27.Maha Sarakham 0.1253 0.1093 

28.Mukdahan 0.2032 0.1683 

29.Yasothon 0.1513 0.0185b North-East 30.Roi Et 0.1713 0.1174 

31.Loei 0.0989 0.0661 

32.Si Sa Ket 0.2979 0.2344 

33.Sakon Nakhon 0.1168 0.0772 

34.Surin 0.1713 0.1139 

35.Nong Bua Lam Phu 0.1114 0.0358b 36.Ubon Ratchathani 0.1027 0.1001 

East 37.Chachoengsao 0.1476 0.0915 

38.Chon Buri 0.0633 0.0601 

39.Prachin Buri 0.1927 0.1096b East 40.Sa Kaeo 0.1587 0.0199b West 41.Prachuap Khiri Khan 0.2697 0.2134 

42.Phetchaburi 0.0981 0.0525b

278 


South 43.Krabi 0.1398 0.0876 

44.Trang 0.0806 0.0856a 45.Nakhon Si Thammarat 0.0871 0.0830 

46.Narathiwat 0.1523 0.1506 

47.Pattani 0.2202 0.0473b 48.Phangnga 0.0998 0.0721 

49.Phuket 0.1331 0.0894 

50.Songkhla 0.0611 0.0651a Table 2 (continued). 

Region Province Measured TF value Calculated TF value 

51.Satun 0.0370 0.0370 

a The calculated TF values were found to be overestimating compared to the measured TF values for most provinces in 

Thailand. 

b The calculated TF values were found to be significantly different from the measured TF values for several provinces in 

Thailand. 

measured TF value 

0.35 

0.30 

0.25 

0.20 

0.15 

0.10 

0.05 

0.00 

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 

calculated TF value 

Figure 1 Measured and calculated TF values of 137 Cs for grass in Thailand. The solid line indicates 1:1 

relationship for measured and calculated values. 


0.25 

0.20 

0.15 

0.10 

0.05 

0.00 

0.0 2.0 4.0 6.0 8.0 10.0 

pH 

Figure 2 Calculated TF values are shown as a function of pH. The solid line is a curve fitted to the data 

in the graph.



0.25 

0.20 

0.15 

0.10 

0.05 


0.00 

0.0 20.0 40.0 60.0 80.0 100.0 

% Clay 

Figure 3 Calculated TF values are shown as a function of clay content. The solid line is a curve fitted 

to the data in the graph. 


0.25 

0.20 

0.15 

0.10 

0.05 

0.00 

0.00 0.20 0.40 0.60 0.80 1.00 1.20 

0.25 

0.20 

0.15 

0.10 

0.05 

Ex-K + (cmol c kg -1 ) 

Figure 4 Calculated TF values are shown as a function of exchangeable K + . The solid line is a curve 

fitted to the data in the graph. 

0.00 

0.00 2.00 4.00 6.00 8.00 10.00 12.00 

% OM 

Figure 5 Calculated TF values are shown as a function of organic matter (OM) content. The solid line 

is a curve fitted to the data in the graph.

280 


0.25 

0.20 

0.15 

0.10 

0.05 

Measured TF values of 137 Cs for grass 

were observed to be 0.037 - 0.298 in the north, 

northeast, east, west, middle and south, with an 

average of 0.129 ± 0.053. These values were 

relatively high compared to the corresponding 

values (0.010 - 0.234 in the north, northeast, east, 

west, middle and south, with an average of 0.085 

± 0.048) predicted by the Absalom model. 

These calculated values differed 

significantly from the measured values. This is due 

to the differences in soil, the types of grass and 

the environmental conditions. In addition, soil 

management such as plough, cultivation method 

and fertilization, microbial process, root density, 

soil moisture and 137 Cs uptake may decrease with 

increasing 137 Cs-soil contact time after the 

deposition on soil (Bell et al., 1988; Kirk and 

Staunton, 1989; Noordijk et al., 1992; Ehlken and 

Kirchner, 2002; Rahman and Voigt, 2004). 

Simple statistical analysis (Williams and 

Leggett, 1984) showed that the agreement between 

model and measured values (Relative Euclidean 

Difference, RED) was 0.238, the value of the 

reliability index (k) was 0.661 and the 

geometrically intuitive reliability index (k g) was 

1.97, which confirmed that the Absalom model 

was reasonably accurate. Calculated TF values by 

the Absalom model were in good agreement with 


0.00 

0.00 10.00 20.00 30.00 40.00 50.00 60.00 

+ -5 -3 

[NH4 ] × 10 (mol dm ) 

Figure 6 Calculated TF values are shown as a function of NH 4 + concentration. The solid line is a 

curve fitted to the data in the graph. 

the measured ones. However, calculated TF values 

were found to be significantly different from the 

measured ones for some provinces. As a result, 

the parameters used in the Absalom model needed 

to be suitably modified to the characteristics of 

soils in Thailand. 

CONCLUSION 

In this work, the uptake of deposited 

137 Cs has been predicted based on the soil 

properties, such as pH, clay content, organic matter 

content, exchangeable K + + 

and NH4 concentration 

valid for the tropical environment in Thailand, and 

using the Absalom model. It has been found that 

the calculated TF values differ significantly from 

the measured values for some provinces in 

Thailand, which implies that the soil properties in 

these provinces differ from those used in the 

Absalom model and they need to be measured 

practically in order to validate the model. 

Furthermore the parameters (k3, k4, k5, k6, kfast, kslow, Pslow and CECclay) could be re-evaluated for 

the tropical environment of Thailand. 


The authors would like to express their

sincere gratitude and deep appreciation to Dr. 

Kanokrat Tiyapun at the Bureau of Technical 

Support for Safety Regulation, Office of Atom for 

Peace (OAP), Thailand, for her initiative idea and 

guidance for utilizing the Absalom model, and also 

fruitful discussions. For supporting data of 137 Cs 

activities in soil and grass, they would like to thank 

Mr. Thawatchai Itthipoonthanakorn at the Bureau 

of Technical Support for Safety Regulation, OAP. 


Absalom, J.P., S.D. Young and N.M.J. Crout. 

1995. Radiocaesium fixation dynamics: 

Measurement in six Cumbrian soils. Eur. J. 

Soil Sci. 46: 461-469. 

_____, _____, _____, A.F. Nisbet, R.F.M. 

Woodman, E. Smolders and A.G. Gillett. 

1999. Predicting soil to plant transfer of 

radiocaesuim using soil characteristics. 

Environ. Sci. Technol. 33: 1218-1223. 

_____, _____, _____, A. Sanchez, S.M. Wright, 

E. Smolders, A.F. Nisbet and A.G. Gillett. 

2001. Prediction the transfer of radiocaesium 

from organic soils to plant using soil 

characteristics. J. Environ. Radioact. 52: 31- 

43. 

Bell, J.N.B., M.J. Minski and H.A. Grogan. 1988. 

Plant uptake of radionuclides. J. Soil Use 

Manage. 4 (3): 76-84. 

Cremers, A., A. Elsen and P. DePreter. 1988. 

Quantitative analysis of radiocaesium 

retention in soil. Nature 335: 247-249. 

Ehlken, S. and G. Kirchner. 2002. Environmental 

processes affecting plant root uptake of 


radioactive trace elements and variability of 

transfer data: a review. J. Environ. Radioact. 

58: 97-112. 

Eisenbud, M. 1973. Environmental Radioactivity. 

Academic Press, New York. p. 118-136. 

Itthipoonthanakorn, T., at the Bureau of Technical 

Support for Safety Regulation, Office of Atom 

for Peace (OAP). private communication. 

Kirk, G.L.D. and S. Staunton. 1989. On predicting 

the fate of radioactive caesium in soil beneath 

grassland. J. Soil Sci. 40: 71-84. 

LDD. 1988. Soil group database search. Land 

Development Department. Ministry of 

Agriculture and Cooperatives. Available 

Source: http://www.ldd.go.th/dinthai/, 

October 9, 2005. 

Nelson, D.W. and L.E. Sommers. 1982. Total 

carbon, organic carbon and organic matter, p. 

539-577. In A.L. Page, R.H. Miller and R. 

Keeney, (eds.). Methods of soil analysis. 

Part 2. Chemical and microbiological 

properties. American Society of Agronomy, 

Madison, Wisconsin. 

Noordijk, H., K.E. Bergeijk, J. Lembrechts and 

M. Frissel. 1992. Impact of ageing and 

weather conditions on soil to plant transfer of 

radiocesium and radiostrontium. J. Environ. 

Radioact. 15: 277-286. 

Rahman, M.M. and G. Voigt. 2004. Radiocaesium 

soil to plant transfer in tropical environments. 

Environ. Sci. Technol. 71: 128-138. 

Williams, L.R. and R.W. Leggett. 1984. A measure 

of model reliability. Health Phys. 46 (1): 85- 

95.


Beta-carotene, Mimosine and Quality of Leucaena Silage Kept 

at Different Duration 

Wanna Angthong1 , Boonlom Cheva-Isarakul2 *, Somkid Promma3 and Boonserm Cheva-Isarkul2 ABSTARCT 

Leucaena leucocephala leaves (LL) were ensiled by mixing with 20% rice bran and 20% 

water (fresh LL basis). The material was kept for 21, 51, 81 and 111 days in vacuumed double layer 

plastic bags, each containing 26 kg. Five bags were randomly taken at each interval for quality evaluation 

by organoleptic test as well as by organic acid and chemical analysis. It was found that the ensiling 

period did not have much influence on most of the chemical compositions. All samples of leucaena leaf 

silage (LLS) had pH of 4.4-4.5 and 35.22-35.65% DM (dry matter). The compositions on DM basis 

were 21.49-22.29% CP (crude protein), 7.76-8.22% EE (ether extract), 31.18-33.68% NDF (neutral 

detergent fiber), 2.0-2.9% acetate and 6.9-9.7% lactate (DM basis). DM loss was 10.35-12.32% which 

was in the normal range for good quality silage. The most interesting points were the increment of βcarotene 

after ensiling from 88.50 to 99.92-120.25 mg/kg DM while mimosine content decreased over 

90% (from 1.79 to 0.12-0.16% of DM) which were superior to a drying method. It indicated that LLS 

is a good alternative for preserving LL and for reduction of mimosine. 

Key words: leucaena, β-carotene, mimosine, silage, organic acids 

INTRODUCTION 

Leucaena leucocephala is a legume 

plant, commonly found in Thailand and many 

other tropical countries. The nutritive value is 

comparable to alfalfa. The leaf contains around 

24% CP and 116-161 mg β-carotene/kg DM 

(Lamchoun, 1998). It is widely used in animal 

feed for monogastrics and ruminants as a source 

of CP, vitamins and minerals. In addition, it also 

provides pigment for skin and egg yolk. However, 

there is a limitation of using leucaena leaves (LL) 

as animal feed because of its high mimosine 

content. This toxic substance is a non-protein 

amino acid. The chemical name is β-N-(3hydroxy-4-pyridone)-α-amino 

propionic acid. 

After ingestion it converts to 3-hydroxy-4(1H)pyridone 

(DHP), that can induce goiter (Jone, 

1994). And since the structure of mimosine is 

similar to that of tyrosine, it becomes an antagonist 

to this amino acid and inhibits protein synthesis. 

Therefore, it reduces growth and production 

performance. In addition, it interferes with B 6 

activity which is necessary for cystathionine 

synthetase and cystathionase in converting 

methionine to cystine, thus causes hair loss (Liener, 

1 Department of Livestock Development, Payathai, Bangkok 10400, Thailand. 

2 Department of Animal Science, Faculty of Agriculture, Chiang Mai University, Chiang Mai 50200, Thailand. 

3 Chiang Mai Livestock Research and Breeding Center, Sanpatong, Chiang Mai 50120, Thailand. 

* Corresponding author, e-mail: agibchvs@chiangmai.ac.th 


1989). In monogastrics, the inclusion of LL above 

5-10% of the ration depresses growth rate and 

induces alopecia, cataract, paralysis, infertility, low 

production efficiency and finally death (Norton, 

1994). 

In Thailand, mimosine content in LL is 

1-5% and in commercial LL meal is 1-2% (Panja, 

1983; Pharkwiwat, 1984). Ruminants in South 

East Asia can tolerate higher mimosine level due 

to gram negative rod shape rumen microbes 

“Synergists jonesii” which can degrade mimosine 

and its derivative DHP into non toxic substances 

(Jone, 1994). However, these microbes are not 

found in monogastrics, therefore it is necessary 

to reduce this toxic substance before use. The most 

popular method is drying under the sun. Although 

other methods such as soaking in ferrous sulfate 

or in running water or the supplement of amino 

acids, Fe, Al and Zn are reported to be effective 

(Panja, 1983; Pharkwiwat, 1984; Wee and Wang; 

1987), there are some practical limitations. 

Ensiling may be an appropriate method 

for preservation and toxic reduction because LL 

is surplus in wet season during which drying is 

rather difficult. Hongo et al. (1986) and Sunagawa 

et al. (1989) reported that around 90% of mimosine 

is destroyed after 14-21 days of ensiling. Since 

LL is a legume, it contains low soluble 

carbohydrate and high buffering capacity. 

Therefore silage additives such as rice bran (RB) 

should be added (Thepprapakarn, 2001). 

However, the report on mimosine residue 

and β-carotene content in leucaena leaf silage 

(LLS) is still limited, therefore this study aimed 


to determine the quality and chemical composition 

of LLS during 111 days of ensiling. 


Fresh LL leaves with green petioles are 

chopped into 2-5 cm length and mixed with rice 

bran and water at the ratio of 100:20:20 (w/w). 

Then the mixture was filled in 2 layer plastic bags 

(26 kg each), vacuumed, closed tightly and kept 

for 21, 51, 81 and 111 days. At the end of each 

period, 5 bags were randomly opened and the 

samples were taken for Proximate and Detergent 

analysis (AOAC, 1984; Georing and Van Soest, 

1970). Physical quality of the silage was evaluated 

by organoleptic test (Gross, 1982). The pH was 

determined according to Bal et al. (1997). The 

concentrations of lactic, acetic and butyric acids 

were analysed by distillation (Zimmer, 1966). 

Mimosine and β-carotene content were analysed 

by the modified methods of Hegarty et al. (1964). 


Chemical composition of ingredient and LLS 

Chemical compositions of LL, RB and 

LLS ensiled with RB are shown in table 1. Crude 

protein of fresh LL in this experiment is higher 

than that reported by Göhl (1975) but lower than 

that reported by Cheva-Isarakul (1982) who 

reported 21.0 and 26.0% CP, respectively. It might 

be due to the age of plant, the ratio of leaves, stems 

and pods as well as season and cultivation 

condition. NDF and ADF were similar to those 

Table 1 Chemical compositions (DM basis) of fresh leucaena leaves (LL), rice bran (RB) and leucaena 

silage (LLS). 

DM OM CP EE Ash NDF1/ ADF1/ ADL NFC pH 

(% DM) 

LL 30.49 92.01 23.50 3.08 7.99 39.14 23.70 8.67 26.29 6.03 

RB 91.01 90.72 12.70 10.55 9.28 19.75 9.71 1.52 47.72 - 

LL2/ 37.07 91.89 22.82 7.95 8.11 34.76 16.54 5.01 26.36 5.08 

1/ ash free 

2/ ensiled with 20% rice bran

284 

reported by Halim (1992). 

LLS had lower CP, ADF and ADL but 

higher EE and ash than those of fresh LL. This 

might be due to the inclusion of RB as a silage 

additive. 

Physical and chemical property of LLS 

The quality of LLS, as evaluated by 

organoleptic test, was good to fairly good (Table 

2). Although the precision of this test method may 

not be high due to the experience and the 

sensitivity of test panels, it is practical and popular 

since it needs no equipment. It was found that the 

scent of lactic acid in LLS was milder than that in 

corn silage. The odour of RB was also noticed 

but no smell of fungi or rotting was detected. The 

odour of LLS was similar to that of tea leaf silage 

which is a local product for human consumption 

in Northern Thailand. 

The colour of LLS was darker than that 

of good quality grass silages. It might be owing 

to the loss of Mg in chlorophyll when reacted with 

organic acids and became phaeophytin which has 


brown color (Watson and Nash, 1960). In addition, 

legume leaves have more pigments than grass, 

therefore legume silage had darker colour than 

grass silage. However, the colour intensity also 

depends on other factors such as oxygen amount 

in silo and temperature during ensiling. LLS in 

this study had good texture and had no mold. Only 

small amount of fungi was found at the opening 

point of some bags. It might be due to oxygen 

remaining at this point after closing these bags. 

Organic acids of LLS ensilaged at 

different ages, determined by distillation 

techniques, are shown in Table 3. Although lactic 

acid of the silage kept for 81 days was significantly 

higher than that of the others, no significant 

differentce was found on quality scores because 

light smell of butyric acid was also noticed. All 

samples were considered good grade silage even 

though pH were higher than 3.7-4.2 which is 

generally found in good quality grass silage. It is 

owing to the fact that leucaena is leguminous plant, 

therefore it contains high buffering capacity, thus 

inhibits pH change. However, pH and acid content 

Table 2 Quality of leucaena silage ensiled at different durations. 

Ensiling period (days) 

21 51 81 111 

Odor 11.06ab 10.10a 10.92a 11.90b Color 2.00b 1.68a 1.92b 1.96b Texture 3.40 3.06 3.60 3.82 

Score 16.48b 14.84a 16.48b 17.68b Values in a row with different superscripts differ significantly (p

of all LLS samples in this study were in the normal 

ranges of good quality silage according to the 

report by Watson and Nash (1960); i.e. ≤65% 

moisture, pH < 4.8, lactate 3-14% and butyrate < 

0.2% (DM basis). 

Chemical compositions of LLS at 

different ensiling periods are shown in Table 4. It 

was found that the length of ensiling period had 

no influence on DM loss and most of the chemical 

compositions. The loss of DM of 10.35-12.32% 

found in this experiment was in a normal range. 

McDonald et al. (1991) reported that the 

unavoidable loss of silage due to the action of plant 

enzymes, microbial enzymes, plant respiration and 

ensiling technique were 1-2, 2-4 and 2-5%, 

respectively. 

No significant difference was found on 

CP content between prior and after ensiling with 

the exception of the lower CP content of the group 

kept for 111 days. The unremarkable protein loss 

was due to the good ensiling condition since the 

bags were vacuumed, thus only minute amount of 

oxygen remained in the bags. In addition, moisture 

level of the ensiling material was optimal (63%) 

therefore no excess heat was produced. These 

conditions led to the low dry matter and nutrient 


loss (Watson and Nash, 1960; McDonald et al., 

1991). 

Furthermore, the low CP loss might be 

due to the fact that LL has condensed tannin (4- 

6% DM basis; Balogun et al., 1998). This 

substance is able to inhibit protein degradation by 

microbial and animal enzymes (Albrecht and 

Muck, 1991). Moreover, trypsin inhibitor in RB 

may also inhibit protein degradation. Most (85- 

90%) of this inhibitor was found in embryo. The 

other part of RB (without embryo) had 5-10% 

while polished rice had less than 1% of this 

inhibitor (Juliano, 1985). 

The concentration of ash and that of EE 

were not affected by ensiling period. Even though 

the forms of minerals and the pattern of fatty acids 

may change during ensiling process, their amount 

should not decrease because no seepage was found 

due to low moisture content (

286 

was found on cellulose. These results were similar 

to those reported by Jaurena and Pichard (2001). 

Beta-carotene and mimosine content in 

leucaena silage 

β-carotene of LLS increased significantly 

(Table 5). Even though no clear explanation can 

be given, the result was in agreement with that 

reported by Peterson et al. (1935; cited by Watson 

and Nash, 1960) who found the increment of 

carotene in alfalfa ensiling with acid. However, 

Hellbery (1945; cited by Watson and Nash, 1960) 

reported that 11-75% of carotene was loss by 

oxidation during fermentation. The extent of loss 

depended on oxygen content and temperature in 

the silo. 

Fermentation decreased 91-93% of 

mimosine. The length of ensiling had no effect 

on mimosine loss. The result was in agreement 

with that reported by Hongo et al. (1986) and 

Sunagawa et al. (1989) who found mimosine 

reduction over 90% in LLS either with or without 

additives. The reduction of mimosine by ensiling 

being higher than by sun drying (14.5-51.1% of 

the original samples) was reported by Panja 

(1983), Parkwiwat (1984) and Wee and Wang 

(1987). These results indicated that LLS is an 

interesting alternative for feed preservation. 

CONCLUSION 

The ensiling of LL by mixing with 20% 


Table 5 β-carotene and mimosine content in leucaena mixed with rice bran and ensiled at different 

durations. 

Ensiling period β-carotene Mimosine 

(days) mg/kg DM % increment % DM % lost 

0 88.50a 0.00 1.79b 0.00 

21 99.92ab 12.90 0.13a 92.74 

51 116.29bc 31.40 0.14a 92.18 

81 120.28c 35.91 0.16a 91.06 

111 105.21abc 18.88 0.12a 93.30 

Value in a column with different superscripts differ significantly (p

Melbourne, Parkville, Victoria. 

Goering, H. K. and P. T. Van Soest. 1970. Forage 

fibre analysis (apparatus, reagents, 

procedure and some application). USDA/ 

ARS Agricultural Handbook No. 379, 

Washington, D.C. 

Göhl, B. 1975. Tropical feeds. FAO Feed 

International Centre, Rome. 

Gross, F. 1982. Grundlagen der 

Futterkonservierung. pp. 6-42. In Tierische 

Erzeugung-Grundlagen. Achte uberarbeitete 

Auflage. BLV Verlagsgesellschaft. München. 

Halim, R. A. 1992. Productivity and nutritive value 

of six fodder tree species. pp 54. In 

Proceeding of The Sixth AAAP Animal 

Science Congress, vol. III, Bangkok. 

Hegarty, M. P., R. D. Court and P. M. Thorne. 1964. 

The determination of mimosine and 3,4dihydroxypyridine 

in biological material. 

Aust. J. Agric. Res. 15: 168-179. 

Hongo, F., S. Tawata, Y. Watanabe and S. Shiroma. 

1986. Mimosine degradation as affected by 

ensiling of Leucaena leucocephala de Wit. 

Japanese J. Zootech Sci. 57(3): 223-230. 

Jaurena, G. and G. Pichard. 2001. Contribution of 

storage and structural polysaccharides to the 

fermentation process and nutritive value of 

Lucerne ensiled alone or mixed with cereal 

grains. Anim. Feed Sci. Technol. 92: 159- 

173. 

Jone, R. J. 1994. Management of anti-nutritive 

factors-with special reference to leucaena. pp. 

216-231. In R.C. Gutteridge and H.M. Shelton 

(eds.). Forage, Tree Legumes in Tropical 

Agriculture. CAB International, Wallingford. 

Juliano, B. O. 1985. Rice bran. pp. 647-687. In 

B.O. Juliano (ed.). Rice Chemistry and 

Technology. 2 nd ed. American Association of 

Cereal Chemists Inc., Minnessota. 

Lamchoun, W. 1998. Seasonal variation and 

effect of sources of beta-carotene on plasma 

beta-carotene concentration of dairy cattle. 


M.Sc. Thesis, Chiang Mai University, Chiang 

Mai. 

Liener, I. E. 1989. Antinutritional factors. pp. 339- 

382. In R.H. Matthews (ed.). Legumes 

Chemistry, Technology and Human 

nutrition. Marcel Dekker, Inc., New York. 

McDonald, P., N. Henderson and S. Heron. 1991. 

The Biochemistry of silage. 2 nd ed. 

Chalcombe Publications, UK. 340p. 

Norton, B. W. 1994. The nutritive value of tree 

legumes. pp. 177-191. In R.C. Gutteridge and 

H.M. Shelton (eds.). Forage Tree Legumes 

in Tropical Agriculture. CAB International, 

Wallingford. 

Panja, P. 1983. Determination of mimosine 

content and toxic reduction in Leucaena 

leucocephala leaves. M.Sc. Thesis, Kasetsart 


Pharkwiwat, S. 1984. Study on nutritive values 

and detoxification methods of mimosine 

reduction in Leucaena leucocephala leaves. 

M.Sc. Thesis, Kasetsart University, Bangkok. 

Sunagawa, L., F. Hongo, Y. Kawashima and S. 

Tawatana. 1989. The effect of mimosinereduced 

leucaena feed on sheep. Japanese J. 

Zootech Sci. 60(2): 133-140. 

Thepprapakorn, R. 2001. Research and 

development on method for production and 

using giant leucaena leaves silage for local 

farmer dairy cattle feeding. M.Sc. Thesis, 

Chiang Mai Rajabhat University, Chiang Mai. 

Watson, S. J. and M. J. Nash. 1960. The 

Conservation of Grass and Forage Crops. 

2 nd Ed. Oliver and Boyd Ltd., Edinburgh. 

Wee, K. L. and S. S. Wang. 1987. Effect of postharvest 

treatment on the degradation of 

mimosine in Leucaena leucocephala leaves. 

J. Sci. Food Agric. 39: 195-201. 

Zimmer, E. 1966. Die Neufassung des 

Garfutterschlussels nuch Flieg. Das 

Wirtschaltseigene Futter. 12: 299-203.


Effects of Natural Mineral Soils on Body Weight and Liver Minerals 

of Black Head Somali Sheep in Ethiopia 

Sisay Tilahun1 , Pravee Vijchulata2 *, Pornsri Chairatanayuth2 and Suwapong Swasdiphanich3 ABSTRACT 

The effects of different mineral soils on body weight and liver mineral concentration were 

investigated using 48 Black Head Somali Sheep in Jijiga Somali region, Ethiopia. The soil samples 

collected from 4 different sites were compared with a complete mineral mix and a control non supplement 

treatment. Chemical composition of the soils indicated that they all are alkaline. Arabi, Jair and Hermokale 

soils from different localities had adequate amount of Ca, K and Mg whereas Mn, Fe and Zn were below 

the recommended standard by 76 to 95%, 87 to 97% and 68 to 88%, respectively. The mean daily 

mineral intakes of sheep supplemented with Jair, Hermokale, Arabi and Bole soil were 18.14, 16.51, 

16.02 and 11.86 grams/sheep/day, respectively. No significant differences were observed in mineral 

intake among Jair, Arabi and Hermokale groups. When compared to other treatment the daily weight 

gain (mean 74.79 g), and total weight gain (mean 8.98 kg) were recorded highest (p0.05) in liver minerals of sheep provided 

with different mineral soils. Liver Mg in sheep from Bole treatment group was significantly different 

when compared to those receiving Jair, Hermokale and Arabi soils. In addition, when compared to 

animals fed on different minerals soils, mineral concentration in the liver of sheep fed Bole soil was 

lower (p>0.05) in Mn (5.49 ppm) and Zn (92 ppm) than those from the other groups. 

Keywords: Mineral soils, weight gain, Black Head Somali sheep, liver minerals 

INTRODUCTION 

In the lowland parts of Ethiopia, sheep 

rearing has been hampered over the years primarily 

due to non-availability of good quality and 

adequate feeds. During the dry season when the 

available forage is low in both quantity and quality 

what usually occurs is loss of live weight, low birth 

weight, lower resistance to disease and poor 

fertility. 

In Somali region, sheep usually suffer 

from diseases resulting from shortage of feed and 

mineral deficiencies. Mineral imbalance 

(deficiencies or excesses) in soil and forages were 

1 Somali Pastoral and Agro Pastoral Research Institute P. O. Box 398, Jijiga, Ethiopia. 

2 Department of Animal Science, Faculty of Agriculture, Kasetsart University, Bangkok 10900,Thailand. 


* Corresponding author, e-mail: agrpvv@ku.ac.th 


considered responsible for low productive and 

reproductive performance of grazing ruminants in 

the tropics (McDowell et al., 1997). Mineral 

deficiencies are considered to be one of the 

nutritional constraints to animal productivity. Poor 

body conditions, slow live weight gain, low 

fertility and high mortality are normally observed 

in mineral-deficient animals (McDowell et al., 

1983; Vijchulata, 1995). 

The main sources of mineral for animals 

in the Somali region are salty water, shrub plants 

and natural mineral soils. The soil mineral known 

as Carro is found in vast area of Afder, 

Degehabour, Gode, Jijiga, Liben and Shinile 

zones. It is commonly observed that pastoralists 

in these zones feed natural mineral soils to animals. 

Considerable use is being made of the natural 

mineral licks since they are relatively free and are 

easily obtained as compared to complete mineral 

mixture. Supplementation with multi-nutrient 

blocks and local mineral soils such as Bole and 

Megadua in some parts of Ethiopia may provide 

an adequate or even excess amount of most of the 

essential minerals except phosphorus (Tolera and 

Said, 1994). 

Studies regarding mineral supplementation 

have not been conducted in the region. 

Moreover, attention has not been paid to its effect 

on Black Head Somali (BHS) sheep. The main 

objective of this study was, therefore, to determine 

the mineral composition of these soils, and to 

evaluate the effect of their supplementation in 

comparison to Bole soil and complete mineral 

mixture on body weight and liver mineral 

concentrations of BHS. 


Animals and management 

The study was conducted in Jijiga Somali 

Region from July to October 2004. Forty eight 

males BHS sheep about 12 months of age 

weighing 20–25 kg were randomized by weight 


assigned to six groups of eight sheep each. Prior 

to the commencement of the experiment, the 

animal were kept for 15 days for adaptation and 

to observe their health status. All animals were 

ear-tagged. They were also provided with neck 

strips of six different colours for group 

identification. All the animals were drenched with 

a broad-spectrum antihelmentic and vaccinated 

against Anthrax, Pasteuriolosis and Blackleg 

diseases. 

The natural mineral soil Arabi soil, Jair 

soil and Hermokale soil were collected from Jijiga 

and Shinile districts, Somali region, Bole soil was 

collected from Zeway district, Oromiya region and 

complete mineral lick from Thailand (Phosrich 

Rockie: Phillips International Co. Ltd.). Six 

treatment groups were randomly assigned to 

mineral supplementation. Group I (control) was 

not supplemented. Group II, III, IV, V were 

supplemented with Arabi soil, Jair soil, Hermokale 

soil and Bole soil, respectively. Group VI was 

provided with complete mineral lick. 

Sheep barn was constructed using 

eucalyptus wood with 19 m × 5 m dimension and 

was divided into 48 equal pens (1m × 0.8m) for 

individual feeding of the minerals and to 

accommodate the animals at night. Sheep were 

allowed to graze together in flock on the same 

pasture from 8:00 am to 6:00 pm. Between 12:30 

pm and 1:30 pm and 6:00 pm to 8:00 am the 

animals were driven into their pens where they 

were fed individually with their respective mineral 

supplements. Mineral soils were offered ad libitum 

in the boxes which were fixed at the corner of 

individual sheep pen. Mineral residues were 

weighed on weekly basis and intakes for each 

sheep were calculated. The animals were weighed 

on monthly basis through out the experimental 

period. All the experimental animals were 

provided ad libitum with water in the pens. 

Soil sampling and analysis 

To study the mineral content of different

290 

natural mineral soils, the soil samples were 

collected from different sites of Somali region and 

from Zeway area of East Showa zone, Oromiya 

Regional State. The sites in Somali region were 

selected from Jijiga and Shinile zones based on 

the suggestion of Somali Regional Pastoral and 

Agro Pastoral Research Institute (SoRPARI) and 

the clan leaders. Soil samples were collected from 

the sites (Arabi, Jair, Hermokale and Bole) where 

pastoralists mostly use for mineral supplementation 

to their animals. Individual natural mineral 

soil was collected using hoe to the depth of 30 cm 

from three different spots which fall in the radius 

of 50 to 70 m. Soil samples were mixed together 

and then filled into a plastic collection bag, labeled 

and stored at room temperature. A total of four 

composite soil samples from the study sites were 

taken and analyzed for physical and chemical 

properties at National Soil Laboratory for Research 

Center (NSLRC). Analysis of each sample was run 

in duplicate. 

The pH and electrical conductivity of the 

soils were measured according to the procedure 

of Landon (1984). Ca, Mg, Fe, Mn, Zn, and Cu 

were determined by atomic absorption 

spectrophotometer (Lindsay and Norvell, 1978). 

Sodium and potassium were analysed by flame 

photometer (Black, 1965). Organic carbon was 

determined following the wet digestion method 

of Walkley and Black (1934). Available 

phosphorus was determined following Olsen 

methods (Olsen et al., 1954). Total nitrogen was 

determined using the modified Kjeldahl method 

(Jakson, 1958). 

Liver tissue collection and analysis 

A total of sixteen sheep (two sheep from 

each treatment group) were randomly selected and 

slaughtered to determine liver mineral contents. 

Liver tissue samples of 50 to 100 g were taken 

from the right lobule of liver of individual animal. 

The samples were stored frozen until analysed for 

Ca, Mg, Fe, Mn, Cu, Zn and Co using flame atomic 


absorption spectrophotometer (Perkins-Elmer, 

Model 2380). 


Data were analyzed using the PROC 

ANOVA procedure of SAS (1999). Differences 

among treatment means were evaluated using 

Duncan Multiple Range Test (Cody and Smith, 

1997). The statistical model used was: 

yij = µ+ Ti + εij 

Where, Yij = Response variable (body 

weight gain, mineral intake and mineral in liver 

tissue) 

µ = Overall mean 

Ti = The effect of ith treatment 

(i = 1, 2, 3, 4, 5, 6) 

εij = Residual effect 


Physical properties of the mineral soils 

Most naturally occurring mineral 

deficiencies in livestock are associated with 

specific regions, and they are related to both soil 

mineral concentration and soil characteristics 

(McDowell 1986). Physical properties of mineral 

soils collected from different sites are presented 

in Table 1. pH of the soil samples ranged from 8.0 

for soils from Arabi to 9.5 for soils from Bole sites 

indicating that they were alkaline in nature. The 

pH values soils was higher than the pH values 

(ranged from 7.86 to 8.05) of natural mineral soil 

from Southern low lands of Ethiopia (Kabaija and 

Little, 1987; Fikre, 1990). Colours of the soils 

varied from site to site ranging from dark brown 

to light brownish colours. The soil textures varied 

from silt to loam. The soils in Jair and Hermokale 

areas were predominantly silt loam but Arabi soil 

were silt in texture whereas Bole soils were sandy 

clay. The color and texture of individual natural 

soils are as shown in Figure 1.


Figure 1 Mineral soil collected from different localities 

A) Arabi soil, B) Jair soil, C) Hermokale soil, D) Bole soil, E) Complete mineral lick. 

Table 1 Physical properties of soil licks collected from different locations in Ethiopia. 

Items Jair Hermokale Arabi Bole 

pH 9.3 8.6 8.0 9.5 

Sand,% 11 17 9.0 49 

Silt, % 64 70 84 10 

Clay, % 25 13 7.0 41 

Texture Silt loam Silt loam Silt Sandy clay 

Chemical composition of the mineral soils 

Minerals and certain chemical 

characteristics of mineral soils are shown in Table 

2 and 3. The organic carbon in the soils ranged 

from 0.28 % for soils from Jair to 1.07 % for soils 

from Hermokale area. Bole soils were less in 

nitrogen than soils from any other sites in the 

Somali region. Based on the critical levels set by 

Mtimuni (1982) and McDowell (1983) for tropical 

soils, Arabi, Jair and Hermokale soils are sufficient 

in Na, Ca, K and Mg. Therefore, the soils can be 

used as supplements for certain macro element. 

Arabi, Jair and Hermokale soils had lower K and 

P and higher Ca, Mg, Mn, and Cu, than Bole soil. 

The macro mineral composition of Bole reported 

by Mohammed et al. (1989) and Adugna (1990)

292 

was similar to what was found in the present study. 

However, in contrast to their studies Mn (0.90 

ppm) and Zn (0.50 ppm) were found extremely 

low in the present study. This could be due to the 

difference in soil sampling site. In the present 

study, all of the mineral soils collected from four 

sites in the region could not be used as phosphate 

supplement for sheep as they contain relatively 

low amount of this mineral than the 10 ppm critical 

level suggested by McDowell (1997). This is in 

agreement with the works of (Kabaija and Little, 

1987; Mohamed et al., 1989; Fikre, 1990) who 

found that soils from Central and Eastern parts of 

Ethiopia low in phosphorus. Phosphorus 

deficiency results in reduced growth and feed 

deficiency, decreased appetite, impaired 

reproduction and weak fragile bone (Underwood, 

1981). The Fe, Mn and Zn concentrations in all 

the soil samples were below critical level at 19 

ppm for Fe set by Mtimuni (1982) and at 19 ppm 

and 2 ppm for Mn and Zn by McDowell (1997). 

Comparing to the stipulated critical levels, Fe, Mn 

and Zn were found to deficit by 87 to 96%, 76 to 

95% and 68 to 82%, respectively. 

The fact that the Bole soils in this study 

had lower Mn, Zn and Cu than the suggested 


Table 2 Macro mineral compositions of soils collected from different sites. 

Sites Ca P Mg K Na 

Jair, ppm 39.42 2.06 16.20 5.07 191.00 

Hermokale, ppm 52.20 8.00 6.91 5.26 179.00 

Arabi, ppm 67.86 6.30 8.42 6.31 60.32 

Bole, ppm 4.10 2.54 0.98 3.35 84.73 

Complete mineral, % 8.19 10.01 0.50 - 20.78 

Table 3 Micro mineral composition (ppm), Total Nitrogen and Organic Carbon of soils collected from 

different sites. 

Sites Fe Cu Co Mn Se Zn I TN (%) OC (%) 

Jair 2.42 2.22 - 2.36 - 0.36 - 0.05 0.28 

Hermokale 1.78 1.50 - 4.58 - 0.64 - 0.07 1.07 

Arabi 0.78 0.66 - 2.96 - 0.50 - 0.09 0.98 

Bole 1.00 0.28 - 0.90 - 0.50 - 0.04 0.67 

Complete mineral 3,000.00 - 50.00 2,500.00 100.00 300.00 300.00 2.07 - 

critical values. This was on the contrary to what 

was reported by Kabaija and Little (1987) and 

Fikre (1990) for Southern parts of Ethiopia. This 

may be attributed to the difference in the time and/ 

or the specific area the soil samples were collected. 

The Cu content of the soils from Jair and 

Hermokale was higher than the critical level of 

0.6 ppm (Mtimuni, 1982) but there was a high 

degree of variation among samples. Apart from 

Ethiopia, Cu deficient soils have also been reported 

in several other African countries (Sillanpää, 

1982). The current study is in agreement with the 

result of Faye et al. (1983) who reported that Cu 

and Zn were deficient in forage taken from 

rangelands in Southern part of Ethiopia. Blood et 

al. (1983) also reported that if the amount of copper 

in the diet is inadequate and copper deficiency may 

occur. Composition of the natural mineral soil in 

the area indicated that Na was the dominant 

mineral. According to NRC (1985) recommendation, 

an appropriate mineral supplement should 

contain at least 0.04% to 0.10% Na. The current 

study indicated that Jair soil was found the highest 

in Na concentration compared to other mineral 

soils (Table 2). Although, macro minerals in the 

soil are high compared to the requirement of

mineral supplementation suggested by NRC 

(1985) but it may not be sufficient to fulfill the 

mineral requirement. Moreover, micro minerals 

in the soil were below sufficient levels for mineral 

supplementation. Individual minerals should be 

adjusted to meet a minimum of 50% of the daily 

intake requirement of the sheep while formulating 

mineral supplement. If the imbalance minerals 

were rectified to meet standard mineral mixture 

requirement, the soils would be more beneficial 

to animals. 

Mineral intake and live weight of the sheep 

The mineral intake and live weight 

change of sheep during the experimental 

period are presented in Table 3. The mean daily 

mineral intake of sheep supplemented with 

mineral soils of Jair, Hermokale, Arabi, Bole and 

complete mineral were 18.14, 16.51, 16.02, 11.86 

and 15.00 gram/head, respectively. There was no 

significant difference (p>0.05) in mineral 

intake among the sheep fed with soils from Arabi, 

Jair and Hermokale areas. The lowest mineral 

intake was recorded for sheep fed with Bole soil 

(11.86 g/day) from Zeway area. Bole soil had 

lower Cu (0.28 ppm), Mn (0.80 ppm) and Zn (0.50 

ppm) than the suggested critical values set by 

Mtimuni (1982) and McDowell (1983). When 

compared to all other mineral soils the level of K 

in Bole soil was higher. However, it had lower 

concentration of Na, Ca and Mg. The imbalance 


of minerals of Bole soil may attribute to the lower 

mineral intake by the animals. Khalili, (1993) 

reported that sodium deficiency which was 

evident in central parts of Ethiopia usually 

causes increased soil ingestion among grazing 

livestock. 

The mean daily weight gains and total 

weight gain of sheep fed complete mineral mixture 

were significantly higher (p

294 

Mohammed et al. (1989) who reported that 

weight gains of Arsi sheep improved by an 

average of 19±8 g/day when fed with natural 

mineral lick as free choice. The low body 

weight gain observed might be due to the fact 

that Bole soil had lower Mn, Zn and Cu but 

higher in pH, sand and clay than the other soils 

(Table 1). The imbalance mineral content of the 

soil might result in depressing effect on 

feed intake which in turn affects the body weight 

gain. Allen et al. (1986) observed a decrease 

in weight gain when 1000 mg/kg of Fe were 

included in the diet of sheep. Hodgson (1962) 

reported that because of large quantities in the 

soil, animals are also likely to augment their 

Fe supplies through direct ingestion of soil or from 

soil contaminated herbage Moreover, in 

New Zealand it was reported that the amount of 

soil ingested annually reach 75 kg for sheep 

(Healy, 1978). 

The average monthly intake of different 

mineral soils is illustrated in Figure 2. During the 

first month of the experimental period, there was 

Mineral intake (gram/month) 

700 

650 

600 

550 

500 

450 

400 

350 

300 


a steady increase in mineral intake of sheep in all 

the treatment groups. This was followed by a 

nearly constant intake over the period from second 

to third month. This decline coincided with 

seasonal summer rain and emergence of green 

grasses after which an increase was observed until 

the end of the fourth month. According to 

McDowell (1997), energy and protein supplies 

from emerging forages during the wet season are 

higher, livestock gain weight more rapidly 

resulting in high mineral requirements. 

Mean monthly body weight gain of the 

sheep fed different mineral sources during the 

experimental period is shown in Figure 3. Sheep 

supplemented with the complete mineral mixture 

tended to gain more live weight than others 

through out the experimental period. Sheep 

receiving complete mineral lick were less heavy 

than that of control animals during the 1st and 2nd 

month. However, it was observed that they attained 

2 kg weight more than the control group by the 

4th month. Sheep fed with Bole soil were found 

to loose weight during the 2nd to the 3rd months 

July August September October 

Experimental period (months) 

Jair Hermokale Arabi Bole 

Figure 2 Average monthly intakes of different mineral soils over the four months period

Body weight chane (kg/head) 

32 

30 

28 

26 

24 

22 

20 

of the experimental period followed by gain in the 

later period. This could be due to lower preference 

and hence lower intake of this soil by the animals. 

Animals in control group had slower increase in 

body weight. Sheep in the control group showed 

compensatory gain from second to fourth month 

but grew slower than that of supplemented 

animals. Animals in the remaining treatment 

groups showed a sharp increase in body weight 

after third month of the experiment. This could be 

partly due to the availability of good pasture as 

rain started during the second month of 

experiment. 

Liver mineral concentration 

Mineral concentration in liver of sheep 

fed on different mineral sources is presented in 

Table 5. McDowell (1992) reported that liver is 

the organ that often represents the status of several 

elements in animals. Liver minerals varied from 

0.04 to 0.08 ppm for Ca, 0.04 to 0.05 ppm for Mg, 

160 to 282 ppm for Fe, 5.49 to 11.79 ppm for Mn, 


July August September October 

Experimental period (months) 

Control Jair Hermokale Arabi Bole Complete 

Figure 3 Average monthly body weight change of sheep supplemented with different mineral soils 

and complete mineral. 

113 to 229 ppm for Cu and 92 to 110 ppm for 

Zn. With exception to Cu and Zn, all the liver 

minerals varied significantly (p

296 


Table 5 Mineral concentration (ppm) in the liver of sheep supplemented with different sources of 

minerals (Means ± Standard Deviation). 

Treatment Ca Mg Fe Mn Cu Zn 

Control 0.04±0.00b 0.050±0.01a 160.00±27.57b 8.69±2.53ab 139.80±1.14a 97.85±4.38a Jair 0.05±0.00 b 0.040±0.01 b 244.15±62.30 ab 10.30±1.59 ab 179.62±119.37 a 94.84±1.28 a 

Hermokale. 0.06±0.01 ab 0.040±0.01 b 194.59±37.29 ab 9.43±0.14 ab 228.48±47.49 a 106.52±9.5 a 

Arabi 0.05±0.00 b 0.040±0.01 b 162.42±24.99 b 9.11±1.29 ab 112.90±60.61 a 96.44±5.19 a 

Bole 0.06±0.02 ab 0.050±0.01 a 170.47±27.98 b 5.49±3.21 b 192.80±60.74 a 92.10±8.19 a 

Complete minerals 0.08±0.01 a 0.045±0.01 ab 282.20±7.49 a 11.79±0.43 a 167.01±7.85 a 109.63±8.93 a 

Critical. Level - - < 180 ** 6 * 25-75 * < 84 ** 

* Georgievskii (1982) 

** Mtimuni (1982) 

Means within the same column with different superscripts are significantly different (p0.05) in liver minerals 

of sheep provided with different mineral soils. 

Liver Mg in sheep from Bole treatment group was 

significantly different (p0.05) 

in Mn (5.49 ppm) and Zn (92 ppm) than those 

from the other groups. The low Mn and Zn 

concentration in the liver might be induced by the 

low soil Mn (0.90 ppm) and Zn (0.50 ppm) mineral 

content in Bole soils (Table 3). 

Although no significant difference 

(p>0.05) was observed in liver Fe among mineral 

soil groups, when compared to the suggested Fe 

standard deficiency status at 180 ppm by Mitimuni 

(1982), Fe concentration was lower in the Control 

(160 ppm), Arabi (162 ppm) and Bole (170 ppm) 

treated sheep. The animals fed on Jair soil showed 

higher Fe concentration in the liver than 

Hermokale, Arabi and Bole treatment groups. This 

might be due to the fact that Jair soil had higher 

Fe (2.42 ppm) content than the other mineral soils. 

Similarly, sheep provided with Arabi soil showed 

low Fe concentration in the liver. This agrees with 

the low Fe (0.78 ppm) content in this soil. Besides 

liver Fe, sheep fed on Bole soil was lower in liver 

Mn (5.49 ppm) than the suggested deficiency 

standard at 6 ppm by Georgievskii et al. (1982). 

This corresponds to the low Mn level found in Bole 

soil. 

According to Georgevskii et al. (1982) 

and Mtimuni (1982), the suggested standard 

deficiency levels, for liver Cu and Zn ranges from 

25 to 75 ppm and

CONCLUSION 

AND RECOMMENDATION 

With the exception of Bole soils, the 

present study revealed that supplementation with 

the three remaining mineral soils improved the 

total weight gain over the negative control sheep. 

Moreover, the daily mineral soil intake of sheep 

fed on Jair soil was higher than sheep received 

different mineral sources. Sheep in all soil mineral 

treatments except Bole soil, consume minerals at 

the same level as complete mineral treated group. 

However, the daily weight gain of animals fed 

complete mineral lick was highest when compared 

to all the remaining treatment sheep. Based on liver 

analysis, the present study reveals that all treatment 

group animals do not require additional micro 

minerals such as Mn, Cu and Zn. Hence, 

pastoralists can use natural mineral soils as mineral 

supplement sources to their animals. Though, there 

is a need to correct the deficiencies of certain 

minerals. In order to achieve the desire result in 

sheep production, phosphorus should be adjusted/ 

corrected in mineral supplementation. For 

improved mineral feeding, the provision of salt 

licks together with mineral soils and bone meal 

would provide a convenient and effective means 

of ensuring adequate mineral supplementation. 

This could be beneficial to the pastoralists and in 

return would have a national benefit in having 

sustainable sheep production. Pastoralists should 

be made aware of the possible incidence of mineral 

deficiencies as parts of range land are lacking in a 

number of mineral elements that are essential in 

animal nutrition. Therefore, it is recommended that 

planned mineral surveys must be conducted in 

wide areas of the region in order to detect mineral 

inadequacies for formulating balanced mineral 

mixture to the animals. 


The authors would like to express their 


gratitude to the management of Somali Pastoral 

and Agro Pastoral Research Institute (SoRPARI) 

staff for their endless support during the 

experimental period. We gratefully acknowledge 

EARO/ARTP for funding this study. Moreover, 

we appreciate Assistance Professors Dr. Sakron 

Koonawootrittriron and Dr. Panwadee 

Sopannarath for their valuable assistance in 

statistical analysis. 

LITRETURE CITED 

Adugna, T. 1990. Animal Production and Feed 

resource Constraints in Welayta Sodo and 

the Supplementary value of Desmodium 

intortum, Stlosanthes guianensis and 

Macrotyloma axillae When Fed to Growing 

Sheep Feeding on Basal Diet of Maize 

Stover. M.Sc. Thesis, Agricultural University 

of Norway. 

Allen, V. G., F. P. Horn and J. P. Fontenot. 1986. 

Influence of ingestion of aluminum, acid and 

soil on mineral metabolism of lactating beef 

cows. J. Anim. Sci. 62 :1396–1403. 

Black, C. A. 1965. Methods of Soil Analysis (part 

1 and 2). American Society of Agronomy. 

Inc. Madison, Wisconsin, USA. 1573 p. 

Blood, D. C., O. M. Radostits and J. A. Henderson. 

1983. Veterinary Medicine. Sixth edition: 

1021-1030. 

Cody, R. P. and J. K. Smith. 1997. Applied 

Statistics and the SAS Programming 

Language. New Jersy . 403 p. 

Faye, B., G. Grillet and A. Tessema. 1983. Report 

on survey of trace element status in forages 

and blood of domestic ruminants in Ethiopia. 

National Veterinary Institute Biochemistry 

and Nutrition Laboratory, Debre Zeit. 

Ethiopia. http://www.fao.org/Wairdocs/ 

ILRI/x5491E/x5491e18.htm 

Fikre, A. 1990. The Mineral Composition and 

Effect of Bole Local Natural Lick 

Supplementation on Feed Intake,

298 

Digestibility and Weight Gain of Black 

Head Ogaden sheep. M.Sc. Thesis, Alemaya 

University of Agriculture. Alemaya. 

Giorgievskii, V. I., B. N. Annenkov and V. T. 

Samochin. 1982. Mineral Nutrition of 

Animals. London, Butterworths. 475 p. 

Hadgson, J. F., R. M. Leach and W. H. Allway. 

1962. Micronutrients in soil and plants in 

relation to animal nutrition. J. Agri. Food and 

Chemistry. 10:171-174. 

Helay, W. B., W. J. McCabe and C. F Wilson. 1970. 

Ingested soil as source of microelement for 

grazing animals. N. Z. J. Agirc. Res. 13: 503. 

Hadgson, J. F., R. M. Leach and W. H. Allway. 

1962. Micronutrients in soil and plants in 

relation to animal nutrition. J. Agric. Food 

and Chemistry. 10:171-174. 

Intraraska, R., N. Phoengpong, S. Wongkasemchit, 

M. Hayashi and Y. Ogura. 1998, pp. 86-90. In 

T. Kawashima and P. Vijchulata. (eds.). Status 

and requirements of mineral in ruminants in 

Thailand. Current knowledge and future 

research. Proceedings of the Workshop Held 

at Bangkok. May 9-10, 1998. Bangkok, 

Thailand. 

Jakson, M. L. 1958. Soil Chemical Analysis. 

Prentice Hall, New York, USA.498 p. 

Kabaija, E and D. A. Little. 1987. Composition of 

native mineral supplements in Ethiopia. pp. 

24-28. In Forage Network in Ethiopia. News 

letter No.15, International Livestock Center 

for Africa (IlCA). Addis Ababa, Ethiopia. 

Khalili, M., E. Linndgren and T. Varvikko. 1993. 

A survey of minerals status of soil, feed and 

cattle in the Selale Ethiopia highlands. I. 

Macro elements. Tropic.Anim. Hlth. Prod. 

25: 162-172. 

Landon, J. R. 1984. Booker Tropical Soil Manual: 

A Hand Book for Soil and Agriculture Land 

Evaluation in the Tropics and Sub Tropics. 

Longman Inc., New York. 447 p. 

Lindsay, W. L. and W. A. Norvell. 1978. 

Development of DTPA soil test for zinc, iron, 


manganese and copper. Soil Sci. Soc. Amer. 

J. 42: 421-428. 

McDowell, L. R. 1992. Minerals in Animal and 

Human Nutrition. Academic press. New 

York. 

McDowell, L. R., J. H. Conard, G. L Ellis and J. 

K. Loosli. 1983. Minerals for Grazing 

Ruminants in Tropical Regions. Department 

of Animal Science Center for Tropical 

Agriculture, University of Florida, 

Gainesville, Florida. 87 p. 

McDowell, L. R.1986. Nutrition of Grazing 

Ruminants in Warm Climates. Academic 

Press, Inc, Harcourt Brace Jovanovich, 

Publishers. 443 p. 

McDowell, L. R. 1997. Minerals in Animal and 

Human Nutrition. Academic Press, New 

York. 524 p. 

McDowell, L. R. 1997. Minerals for Grazing 

ruminants in Tropical Regions. Animal 

Sciences Department Center for Tropical 

Agriculture University of Florida. Bulletin 

Third Edition. 83 p. 

Mohammed, Y. K., I. Deissie and M. Mukassa. 

1989. Effect of supplementation with lake soil 

on the growth of Arsi sheep. pp. 139-142. In 

Proceedings of The Third National 

Livestock Improvement Conference. 24-26 

May 1989. Institute of Agricultural Research. 

Addis Ababa, Ethiopia. 

Mtimuni, J. P. 1982. Identification of mineral 

deficiencies in soil, plant and animal 

tissues as constraints to cattle production in 

Malawi. CTA Report 6 IFAS University of 

Florida: Gainesville, Florida. 

NRC, 1975. Nutrient Requirements of Sheep. 

6 th ed. National Academy press. Washington, 

DC, USA. 72 p. 

Olsen, S. R., C. V. Cole, F. S. Watanable, and L. 

A. Dean. 1954. Estimation of available 

phosphorous in soils by extraction with 

sodium bicarbonate. USDA, Circular 39: 1- 

19.

SAS. (Statistical Analysis System). 1999. SAS 

User Guide, SAS Institute Inc., Carry 

Carolina. 584 p. 

Sillanp‰‰, M. 1982. Micronutrients and the 

nutrient status of soil: a global study. FAO 

Soil Bulletin 48. 

Tolera, A. and A. N.Said.1994. Assessment of Feed 

resources in Welayta Sodo: 

Quantity Estimation and Laboratory Evaluation. 

Ethiopian J. Agri. Sci. 14: 

69-87. 

Underwood, E. J. 1981. The Mineral Nutrition 

of Livestock. Commonwealth Agricultural 

Bureax, London, England. 180 p. 


Van Eys, J. E., S. Silitonga S., I. W. Mathius and 

W. L. Johnson. 1985. Research methodology 

for livestock on-farm trials, pp. 313-317. 

Proceedings of a Workshop Held at Aleppo 

Syria, 25-28 March 1985. Ottawa, Ont., 

IDRC. 

Vijchulata, P. 1995. Blood plasma minerals and 

fertility of dairy cows in Central Thailand. 

Asian–Australasian J. Anim. Sci. 8(3): 131- 

137. 

Walkley, A. and C. A. Black. 1934. Estimation of 

soil organic carbon by the chromic acid 

titration method. Soil Sci. 37: 29-38.


Protoplast Isolation and Culture of Aquatic Plant 

Cryptocoryne wendtii De Wit 

Kanchanaree Pongchawee 1 *, Uthairat Na-Nakorn 2 , Siranut Lamseejan 3 , 

Supawadee Poompuang 2 and Salak Phansiri 4 

ABSTRACT 

The optimum conditions for protoplast isolation and culture of Cryptocoryne wendtii 

De Wit were investigated. Protoplasts were successfully isolated from in vitro four-week-old leaves 

using an enzyme mixture comprising 2% Cellulase Onozuka R-10, 0.2% Pectolyase Y-23, 0.5 M mannitol, 

2.5 mM CaCl 2.2H 2O and 5 mM 2 (N-morpholino)-ethanesulfonic acid (MES), pH 5.6. Approximately 

1.04±0.06 × 10 7 protoplasts per gram fresh weight with 90.79±4.80% viability were obtained after 

incubating in enzyme solution for 4 hours in the dark and purified with 16 % sucrose gradient 

centrifugation. Protoplasts were cultured on modified MS medium supplemented with 0.2 mg/l 2,4dicholorophenoxyacetic 

acid (2,4-D), 1 mg/l α-naphthalene acetic acid (NAA), 0.5 mg/l zeatin, 0.15 M 

sucrose and 0.3 M mannitol by agarose-bead with thin layer liquid culture. The protoplasts regenerated 

cell walls within 24 hours. First cell division was observed after culturing for 2-3 days, and microcolonies 

were formed within 4 weeks. Enzyme mixture, osmotic solution, incubation time, age of leaves, 

and sucrose solution concentration were found to influence both yield and viability of protoplasts. Culture 

media, plant growth regulators and method of culture affected protoplast division. 

Key words: aquatic plant, Cryptocoryne wendtii De Wit, protoplasts isolation, protoplasts culture 

INTRODUCTION 

The Cryptocoryne genus is a member of 

Araceae with more than 50 different species. They 

are distributed throughout Southeast Asian coastal 

zones. Some species are commercially cultivated 

as aquarium plants (Mühlberg, 1982). 

Cryptocoryne wendtii De Wit is an important 

species used in the aquarium plant trade (Rajaj 

and Horeman, 1977). It is a medium-sized species 

with thin rhizomes and runners, able to grow 

1 Aquatic plant and Ornamental Fish Research Institute, Bangkok 10900 , Thailand. 

2 Department of Aquaculture, Faculty of Fisheries, Kasetsart University, Bangkok 10900, Thailand. 

3 Gamma Irradiation Service and Nuclear Technology Research Center, Kasetsart University, Bangkok 10900, Thailand. 

4 Scientific Equipment Center, KURDI, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: kanchanp@fisheries.go.th 

emerged or submersed and is propagated by 

runners (Mühlberg, 1982). The aerial leaves are 

oblong with round or heart shaped base, 8 to 10 

cm long by 2 to 3 cm wide and below water. The 

blade are narrower (Allgayer and Teton, 1986). 

In order to increase the value of exports 

and to cope with international market demand, the 

improvement of new aquatic plant varieties for 

desirable traits such as variable leaf color and form 

are the key to success. Related or relevant genera 

of cultivated crops contain a large reservoir of 


genes covering a variety of desirable traits (Liu et 

al., 2005). However, reproductive incompatibility 

generally prevents simple hybridization between 

taxa. Somatic cell fusion enables nuclear and 

cytoplasmic genomes to be combined, fully or 

partially, at the interspecific and intergeneric levels 

to circumvent naturally occurring sexual 

incompatibility barriers (Davey et al., 2005). There 

have been many reports of the transfer of useful 

agronomic traits by protoplast fusion for 

production of triploid (Fu et al., 2003) and 

polyploid (Mizuhiro et al., 2001) plants and 

increasing plant vigour (Cheng et al., 2003). This 

technique may be a possible alternative for the 

genetic improvement of Cryptocoryne. For 

successful protoplast fusion, a reliable procedure 

for protoplast isolation and culture is a prerequisite. 

Up to date, there are a few reports of protoplast 

isolation and culture of aquatic plants such as 

seagrass (Balestri and Cinelli, 2001). 

In this study, the procedures for isolation 

and culture of C. wendtii protoplasts were 

established for the first time. The information 

obtained from this study will greatly benefit further 

genetic improvement of Cryptocoryne. 



Shoot tip explants of C. wendtii were 

surface-sterilized by immersion in 50% (V/V) 


ethanol for 1 min and 1.05 % NaOCl containing 1 

drop of Tween-20 per 100 ml for 12 min, followed 

by rinsing three times with sterile distilled water 

(Kane et al., 1999). Explants were cultured on 

semi-solid MS medium (Murashige and Skoog, 

1962) supplemented with 2 mg/l 6-benzyladenine 

(BA), 0.25 mg/l NAA, 30 g/l sucrose and 1.6 g/l 

gelrite (Sigma, USA). The cultures were incubated 

under 16/8 h light/dark photoperiod at 25°C. 

Plantlets derived from shoot tips were subcultured 

into the same medium every four weeks. Leaves 

of plantlets were used as the explants for protoplast 

isolation. 

Factors affecting the protoplast isolation 

1. Enzyme mixtures 

Five enzyme mixtures (Table 1) were 

examined for the suitable protoplast isolation. The 

tested enzyme mixtures were dissolved in 0.5 M 

mannitol, 2.5 mM CaCl 2.2H 2O and 5 mM 2-Nmorpholino-ethanesulfonic 

acid (MES) pH 5.6. 

One gram of in vitro four-week-old leaves were 

cut transversely into 1-2 mm wide strips in a 

washing solution (0.45 M mannitol, 2.5 mM 

CaCl 2.2H 2O and 5 mM MES, pH 5.6). The sliced 

tissues were plasmolysed by immersion in washing 

solution for 30 minutes. The plasmolysis solution 

was pipetted off, replaced with 5 ml of the filtersterilized 

(Satorius, pore size 0.20 µm) enzyme 

mixtures and incubated in the dark on a gyratory 

shaker (40 rpm) at 25°C for 4 hr. The protoplasts 

Table 1 Components of enzyme mixtures used for protoplast isolation of C. wendtii 

Enzyme Enzyme concentration (% w/v) 

mixtures Cellulase Pectinase 

Cellulase R-10a Cellulase RSa Macerozyme R-10a Pectolyase Y-23b E1 2 - 2 - 

E2 2 - - 0.2 

E3 - 2 2 - 

E4 - 2 - 0.2 

E5 2 2 0.1 

a Yakult, Tokyo. 

b Seishin, Tokyo.

302 

were then gently filtered through a 60 and 40 µm 

nylon mesh to remove undigested tissue and 

debris. The filtrate was centrifuged for 5 min at 

750 rpm. The same process was repeated once 

more. The protoplast pellets were purified by 

floating on 20 % sucrose solution and centrifuged 

at 800 rpm for 10 min. The purified protoplasts 

were further washed twice with washing solution. 

Protoplast yield was estimated by a 

hemocytometer (Gleddie, 1995). Viability of 

protoplasts was assessed using 0.01% (w/v) 

fluorescein diacetate staining (FDA) (Sigma, 

USA) followed by observation with a UV 

fluorescence microscope (Widholm, 1972). 

2. Concentration of osmoticum 

solution 

The best result of experiment 1 was used 

in experiment 2. One gram of four-week-old in 

vitro leaves was incubated in 5 ml of filtersterilized 

enzyme mixture, 2% (w/v) Cellulase 

Onozuka R-10 (Yacult Honsha, Japan), 0.2% (w/ 

v) Pectolyase Y-23 (Kyowa Chemical, Japan) in 

washing solution of four varied mannitol 

concentrations; 0.4, 0.5, 0.6 or 0.7 M. The 

protoplasts were isolated and purified as 

previously described. Protoplast yield and 

viability were determined. 

3. Incubation period 

The best result of experiment 2 was used 

in experiment 3. One gram of four-week-old in 

vitro leaves was incubated in 5 ml of enzyme 

mixture, 2% Cellulase Onozuka R-10, 0.2% 

Pectolyase Y-23, 0.5 M mannitol, 2.5 mM 

CaCl 2.2H 2O and 5 mM MES. The digestion was 

performed for 3, 4, 5 or 6 hr in the dark. The 

protoplasts were then harvested and purified as 

previously described. Protoplast yield and 

viability were determined. 

4. Age of leaves 

One gram of four-, six-, eight- and ten- 


week-old leaves was isolated using enzyme 

mixture, 2% Cellulase Onozuka R-10, 0.2% 

Pectolyase Y-23, 0.5 M mannitol, 2.5 mM 

CaCl 2.2H 2O and 5 mM MES, and incubated in 

the dark on a gyratory shaker (40 rpm) at 25°C for 

4 hr. The protoplasts were then harvested and 

purified as previously described. Protoplast yield 

and viability were determined. 

5. Sucrose concentrations for 

purification 

One gram of four-week-old in vitro 

leaves was incubated in 5 ml of enzyme mixture, 

2% Cellulase Onozuka R-10, 0.2% Pectolyase Y- 

23, 0.5 M mannitol, 2.5 mM CaCl 2.2H 2O and 5 

mM MES. The protoplasts were harvested and 

purified with varying levels of sucrose solution; 

16, 18, 20 and 22 % and centrifuged at 800 rpm 

for 10 min. Protoplast yield and viability were 

determined. 

Factors affecting the protoplast culture 

1. Culture medium 

The purified protoplasts at the density of 

5 × 10 5 protoplasts/ml were cultured in two kinds 

of liquid culture media; MS (Murashige and 

Skoog, 1962) and KM8P (Kao and Michayluk, 

1975) containing 0.2 mg/l 2,4-D, 1 mg/l NAA, 

0.5 mg/l zeatin, 0.15 M sucrose and 0.3 M 

mannitol incubated at 25°C in the dark. The cell 

division was observed periodically with an 

inverted microscope. The plating efficiency (% of 

plated protoplasts which were under cell division) 

and the survival rate of protoplasts were 

determined after 10 days of culture. 

2. Plant growth regulators 

The protoplasts were cultured in liquid 

MS medium containing various combinations of 

growth regulators. Three culture media tested for 

protoplast culture were M1 (1.5 mg/l NAA and 

0.4 mg/l BA), M2 (0.2 mg/l 2,4-D, 1 mg/l NAA 

and 0.5 mg/l zeatin) and M3 (0.2 mg/l 2,4-D, 2

mg/l NAA and 0.5 mg/l zeatin) incubated at 25°C 

in the dark. The plating efficiency and percentage 

of survival were evaluated after 10 days of culture. 

3. Culture method 

Protoplasts were cultured using two 

methods, namely, the liquid thin layer and agarose 

bead methods. For the liquid thin layer method, 

protoplasts in liquid MS medium at the density of 

5 × 10 5 protoplasts/ml were poured into a 6 cm 

Petri dish. For agarose bead method, one volume 

of the protoplast suspension was gently mixed with 

one volume of modified MS medium containing 

0.2 mg/l 2, 4-dichlorophenoxyacetic acid (2,4-D), 

1 mg/l NAA and 0.5 mg/l Zeatin with 1.2 % (w/v) 

agarose (SeaPrep ® , FMC BioProducts, U.S.A.). 

The protoplast suspension was dropped into a 6 

cm Petri dish. The droplets were covered with 3 

ml of modified liquid MS medium and incubated 

at 25°C in the dark for 10 days, dim light for 10 

days, and then in the light for 30 days. Cell wall 

regeneration was observed using 0.01% (w/v) 

calcofluor white staining under a fluorescence 

microscope (Phansiri et al., 1992). The plating 

efficiency and percentage of protoplast survival 

were examined after 10, 30 and 50 days of culture. 


All data were assessed by one-way 

analysis of variance (ANOVA), and the means 

were compared by the Turkey test at 95% interval 

of confidence (*P

304 

(Figure 5B). Their viability was 87.14 % and 82.76 

% for the four- and six-week-old leaves, 

respectively, as determined by FDA staining 

(Figure 5C). Protoplast viability decreased 

Yield (×10 5 prtoplasts/ g FW) 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

a 

a 

d 

c 

a 

E1 E2 E3 

Enzyme mixtures 

E4 E5 

ab 

Yield Viability 

Figure 1 Effect of different enzyme mixtures on 

yield and viability of C. wendtii 

protoplasts. Data represent mean ± 

standard error of three replicates. 


120 

100 

80 

60 

40 

20 

0 

a 

b 

c 

b 

c 

c 

ab 

3 4 5 6 

Incubation time (hours) 


Figure 3 Effect of incubation time on yield and 

viability of C. wendtii protoplasts. 

Data represent mean ± standard error 

of three replicates. 

c 

b 

b 


a 

bc 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

Viability (%) 

Viability (%) 

significantly with the increase in leaf age (Figure 

4). It was also found a remarkable number of 

raphids when using the leaves as a source of 

protoplasts. 


100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

b 

ab 

d 

c 

0.4 0.5 0.6 0.7 

Mannitol concentration (M) 


Figure 2 Effect of mannitol concentrations in the 

enzyme mixture on yield and viability 

of C. wendtii protoplasts. Data 

represent mean ± standard error of 

three replicates. 


120 

100 

80 

60 

40 

20 

0 

c 

b 

c 

ab 

c 

b 

b 

ab 

4 6 8 10 

Leaf age (weeks) 


a 

a 

a 

a 

0 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

Figure 4 Effect of leaf age on yield and viability 

of C. wendtii protoplasts. Data 

represent mean ± standard error of three 

replicates. 

90 

80 

70 

60 

50 

40 

30 

20 

10 

Viability (%) 

Viability (%)

5. Purification by various sucrose 

concentrations 

There was a significant difference 

between the yield of protoplasts centrifuged in the 

four sucrose concentrations tested, but no 

significant difference in the viability (Table 2). 

Purification with 16 % sucrose solution gave the 


Figure 5 Isolation, culture and cell division of Crytocoryne wendtii protoplasts. Four-week-old plantlets 

suitable for the isolation of leaf protoplasts (A), protoplasts after purification with 16 % 

sucrose solution (B), vigorous protoplasts fluoresce a yellow-green color when stained with 

FDA (C), first cell division of protoplast culture in agarose bead after a few days of culture 

(D), second cell division after 10 days of culture (E), small cell colonies after culturing for 30 

days (F). Bar = 20 µm. 

highest yield of 103.62×10 5 protoplasts/g FW with 

the viability of 90.79 %, and without cell debris. 

Factors affecting the protoplast culture 

1. Culture medium 

MS medium was found to be more 

effective than KM8P medium. The first cell

306 

division was found within 2-3 days in MS medium 

supplemented with 0.2 mg/l 2,4-D, 1 mg/l NAA 

and 0.5 mg/l Zeatin, 0.15 M sucrose and 0.3 M 

mannitol. The plating efficiency and survival rate 

at 10 days after culture were 21.27 % and 60.44 

%, respectively (Table 3). In contrast, the 

protoplasts cultured in KM8P medium with the 

same growth regulator as MS medium did not 

divide but turned brown and died after 10 days of 

culture. This indicates that MS medium was 

suitable for culturing mesophyll protoplasts of C. 

wendtii. 

2. Plant growth regulators 

Protoplasts did not divide after being 

cultured in M1 (1.5 mg/l NAA, 0.4 mg/l BA) for 


10 days. The highest plating efficiency (21.27 %) 

and cell survival (57.11 %) were observed in M2 

(0.2 mg/l 2,4-D, 1 mg/l NAA and 0.5 mg/l Zeatin), 

which was statistically similar to that in M3 (0.2 

mg/l 2,4-D, 2 mg/l NAA and 0.5 mg/l Zeatin) 

(Table 4). 

3. Culture method 

The freshly isolated protoplasts cultured 

in liquid and agarose bead culture regenerated cell 

walls within 24 hr. The first division of protoplasts 

was observed in 2-3 days (Figure 5D). After 10, 

30 and 50 days there were no significant 

differences within the plating efficiency and 

survival rate of both culture methods (Table 5, 6). 

The plating efficiency and survival rate decreased 

Table 2 Effect of sucrose concentration on yield and viability of C. wendtii protoplasts. 

Sucrose (%) Yield (×10 5 protoplasts/g FW) Viability (%) 

16 103.62±5.63 c 90.79±4.80 ns 

18 80.38±1.78 b 84.74±3.23 ns 

20 81.04±1.78 b 80.27±4.52 ns 

22 58.79±2.96 a 76.96±1.50 ns 

Data represent mean ± S.E. of three replicates. Means in the same column sharing the same superscript letter are not significantly 

different as determined by Turkey’s test (*P>0.05). 

Table 3 Effect of culture medium on cell division and survival of C. wendtii protoplasts after culturing 

for 10 days. 

Culture media Plating efficiency (%) Survival rate (%) 

MS 21.27 ± 1.32 b 60.44 ± 3.61 b 

K8 0.00 ± 0.00 a 0.00 ± 0.00 a 

Data represent mean ± S.E. of three replicates. Means in the same column not sharing the same superscript letter are significantly 

different as determined by Turkey’s test (*P

as the culture period increased. Some protoplasts 

survived, divided and developed to small colonies 

only in agrarose bead (Figure 5F). However, callus 

was not formed, they turned brown and finally 

died. 

DISCUSSION 

The success in protoplast isolation of C. 

wendtii was influenced by the enzyme mixture, 

osmoticum solution, incubation period, age of 

leaves, and sucrose concentration. The 

combination of enzyme solution has been reported 

to be an important factor on yield and viability of 

protoplasts in many plant species such as Artemisia 

judaica L. and Echinops spinosissimus Turra (Pan 

et al., 2003) and Echinacea augustifolia (Zhu et 

al., 2005). Cellulase Onozuka R-10 was a preferred 

enzyme for leaf protoplast isolation of C. wendtii 

rather than Cellulase RS which had higher 

cellulase activity (Marchant et al., 1997). Cellulase 

Onozuka 10 combined with Pectolylase Y-23 was 

the most efficient for protoplast isolation of C. 

wendtii. Pectolyase Y-23 was efficient for digestion 

of mesophyll protoplast (Nagata and Ishii, 1979; 

Eriksson, 1985) due to Pectolyase Y-23 having 

endo-polygalacturonase activity about 50 times 


stronger than Macerozyme R-10 (Nagata and Ishii, 

1979). 

In isolating protoplasts, the wall pressure 

must be replaced by osmotic pressure in the 

isolation mixture. Mannitol is considered to be 

relatively inert metabolically and infuses slowly 

into the protoplast (Eriksson, 1985). The 

concentration of mannitol in the enzyme solution 

was another important factor affecting C. wendtii 

protoplast release. The yield and viability of 

protoplasts were shown to decrease with the 

increasing of mannitol concentration due to the 

protoplasts being plasmolyzed (Sinha, 2003). The 

prolonged incubation period decreased the yield 

and viability of protoplasts because of the over 

digestion (Zhu et al., 2005). 

The ages of the leaves were also critical 

for the successful protoplast isolation of C. wendtii. 

The younger leaves gave the maximum of both 

viability and yield because less pectic substances 

accumulate in young cell walls than in the old cells 

(Babaoˇglu, 2000), and the cell wall of a rapidly 

expanding leaf is thinner (Marchant et al., 1997). 

There were many calcium oxalate needles found 

when leaves were used as the source of protoplasts. 

These crystals are able to puncture and burst 

protoplasts during isolation (Price and Earle, 1984; 

Table 5 Effect of culture methods on plating efficiency of C. wendtii protoplasts in MS medium. 

Culture method Plating efficiency (%) 

Day 10 Day 30 Day 50 

Liquid 28.61 ± 4.72 ns 18.77 ± 3.50 ns 13.60 ± 1.80 ns 

Agarose bead 25.82 ± 2.46 ns 20.66 ± 4.67 ns 14.76 ± 2.14 ns 


different as determined by Turkey’s test (P>0.05) 

Table 6 Effect of culture methods on survival rate of C. wendtii protoplasts in MS medium. 

Culture method Survival rate (%) 

Day 10 Day 30 Day 50 

Liquid 67.25 ± 5.46 ns 57.47 ± 4.65 ns 33.67 ± 4.06 ns 

Agarose bead 68.12 ± 5.37 ns 54.93 ± 5.65 ns 36.57 ± 3.08 ns 


different as determined by Turkey’s test (P>0.05)

308 

Kunasukdakul and Smitamana, 2003). However, 

all raphids and debris could be successfully 

removed by centrifugation of protoplasts with 16 

% sucrose solution. 

For the protoplasts culture of C. wendtii, 

the culture media, culture method and plant growth 

regulators were important factors affecting plating 

efficiency and survival rate. The protoplasts could 

divide in liquid as well as in agarose bead culture. 

However, microcolonies were formed only in 

agarose bead culture. The agarose bead culture 

methods have been found to be an efficient method 

for cell division and microcolony formation in 

many crop species including Lavatera thuringiaca 

(Vazquez-Tello et al., 1995); Rosa hybrida 

(Marchant et al., 1997) and Cucumis melo ‘Green 

Delica’ (Sutiojono et al., 1998). The enhanced 

protoplast division observed in bead culture was 

due to the dilution of substances having inhibitory 

effects on protoplast division which are secreted 

from the cell to the medium (Mizuhiro et al., 

2001). 

Colony formation was observed after 

culturing protoplasts in MS medium supplemented 

with 0.2 mg/l 2,4-D, 1 mg/l NAA, 0.5 mg/l Zeatin, 

0.3 M mannitol, and 0.15 M sucrose for 30 days. 

However, it did not form a callus but turned brown 

and finally died. It has been reported that the 

protoplasts isolated directly from leaves of 

monocotyledons, except rice, was very difficult 

to culture (Kuehnle and Nan, 1990). It was 

suggested that leaf cells rapidly lose totipotency 

thus preventing cells from dedifferentiating and 

reentering the cell cycle (Krautwig and Lörz, 

1995). Plant regeneration has been found possible 

when callus and cell suspension were used as the 

source of protoplast isolation and culture 

(Kobayashi et al., 1993; Pauk, et al., 1994). 

CONCLUSION 

The procedure for simple and reliable 

isolation and culture of C. wendtii protoplasts has 


been described for the first time. It might lead to 

the improvement of the Cryptocoryne through 

somatic hybridization, somaclonal variation and 

genetic engineering by using the protoplast 

technique. Even though the viable protoplasts of 

C. wendtii could form microcolonies, further 

research is needed to develop the efficient 

procedure for the protoplast regeneration. 

ACKNOWLEDGMENTS 

This research was financially supported 

by the Department of Fisheries, Ministry of 

Agriculture and Co-operatives, Bangkok, 

Thailand. We thank Dr. Sureeya Tantiwiwat, 

Department of Botany, Kasetsart University, Dr. 

Yuphin Khentry and Mr. Adrian Hillman, 

Graduate School, Kasetsart University for editing 

this manuscript and their helpful suggestions. 


Allgayer, R. and J. Teton. 1986. Aquarium 

Plants. Worlds Lock, Ltd., London. 157 p. 

Babaoˇglu, M. 2000. Protoplast isolation in Lupin 

(Lupinus mutabilis Sweet): determination of 

optimum explant sources and isolation 

conditions. Turk. J. Bot. 24: 177-185. 

Balestri, E. and F. Cinelli. 2001. Isolation and cell 

wall regeneration of protoplasts from 

Posidonia oceanic and Cymodocea nodosa. 

Aqua. Bot. 70: 237-242. 

Cheng, Y.J., W.W. Gou and X.X. Deng. 2003. 

Molecular characterization of cytoplasmic and 

nuclear genomes in phenotypically abnormal 

Valencia orange (Citrus sinensis) plus Meiwa 

kumquat (Furtunella crassifolia) intergeneric 

somatic hybrids. Plant Cell Rep. 21: 445-451. 

Davey, M.R., P. Anthony, J.B. Power and K.C. 

Lowe. 2005. Plant protoplasts: status and 

biotechnological perspectives. Biotechnol. 

Advances 23: 131-171. 

Eriksson, T.R. 1985. Protoplast isolation and

culture, pp. 1-20. In L.C. Fowke and F. 

Constabel (eds.). Plant Protoplasts. CRC 

Press, Florida. 

Fu, C.H., W.W. Guo, J.H. Liu, X.X. Deng. 2003. 

Regeneration of Citrus sinensis (+) Clausena 

lansium intergeneric tripliod and tetraploid 

somatic hybrids and their identification by 

molecular markers. In Vitro Cell. Dev. Biol- 

Plant 39: 360-364. 

Gleddie, S.C. 1995. Protoplast isolation and 

culture, pp. 167-180. In O.L. Gamborg and 

G.C. Phillips (eds.). Plant Cell Tissue and 

Organ Culture : Fundamental Methods. 

Springer-Verlag, Berlin Heidelberg. 

Kane, M.E., G.L. Davis, D.B. McConnell and J.A. 

Gargiulo. 1999. In vitro propagation of 

Cryptocoryne wendtii. Aquat. Bot. 63: 197- 

202. 

Kao, K.N. and M.R. Michayluk. 1975. Nutritional 

requirements for growth of Vicia hajastana 

cells and protoplasts at a very low population 

density in liquid media. Planta 126: 105-110. 

Kobayashi, R., T. Kameya and S. Ichihashi. 1993. 

Plant regeneration from protoplasts derived 

from callus of Phalaenopsis spp. Plant Tiss. 

Cult. Lett. 10: 267-270. 

Krautwig, B. and H. Lörz. 1995. Review article 

cereal protoplasts. Plant Sci. 111: 1-10. 

Kuehnle, A.R. and G.L. Nan. 1990. Factors 

influencing the isolation and culture of 

protoplasts from Hawaiian Dendrobium 

cultivars, pp. 259-262. In J. Kernohan, D.G. 

Bonhanm, N. Bonhanm and L. Cobb (eds.). 

Proceedings of the 13 th World Orchid 

Conference. 13 th World Orchid Conference 

Proceeding Trust, Auckland. 

Kunasakdakul, K. and P. Smitamana. 2003. 

Dendrobium Pratum Red protoplast. Thai J. 

Agric. Sci. 36: 1-8. 

Liu, J.H., X. Xu and X.X. Deng. 2005. 

Intergeneric somatic hybridization and its 

appplication to crop genetic improvement. 

Plant Cell Tiss. Org. Cult. 82: 19-44. 


Marchant, R., M.R. Davey and J.B. Power. 1997. 

Isolation and culture of mesophyll protoplasts 

from Rosa hybrida. Plant Cell Tiss. Org. 

Cult. 50: 131-134. 

Mizuhiro, M., K. Ito and M. Mii. 2001. Production 

and characterization of interspecific somatic 

hybrids between Primula malacoides and P. 

obconica. Plant Sci. 161: 489-496. 

Mühlberg, H. 1982. The Complete Guide to 

Water Plants. Sterling Publishing Co., Inc., 

German Democratic Republic. 392 p. 

Murashige, T. and F. Skoog. 1962. A revised 

medium for rapid growth and bioassays with 

tobacco tissue cultures. Physiol. Plant 15: 

473-497. 

Nagata, T. and S. Ishii. 1979. A rapid method for 

isolation of mesophyll protoplast. Can. J. Bot. 

57: 1820-1823. 

Pan, Z.G., C.Z. Liu, S.J. Murch, M. El-Demerdash 

and P.K. Saxena. 2003. Plant regeneration 

from mesophyll protoplasts of the Egyptian 

medicinal plants Artemisia judaica L. and 

Echinops spinosissimus Turra. Plant Sci. 

165: 681-687. 

Pauk, J., Z. Kertész, B. Jenes, L. Purnhauser, O. 

Manninen, S. Pulli, Z. Barabás and D. Dudits. 

1994. Fertile wheat (Triticum aestivum L.) 

regenerants from protoplasts of embryogenic 

suspension culture. Plant Cell Tiss. Org. 

Cult. 38: 1-10. 

Phansiri, S., H. Miyake and E. Maeda. 1992. 

Studies on cell wall regeneration and cell 

division in soybean protoplasts using 

fluorescence and scaning electron microscope. 

Jpn. J. Crop Sci. 61: 487-493. 

Price, G.C. and E.D. Earle. 1984. Source of orchid 

protoplasts for fusion experiments. Amer. 

Orchid Soc. Bull. 53: 1035-1043. 

Rajaj, K. and T.J. Horeman. 1977. Aquarium 

Plant: Their Identification, Cultivation and 

Ecology. T.F.H. Pulb. Inc., West Sylvania. 

448 p. 

Sinha, A., A.C. Wetten and P.D.S. Caligari. 2003.

310 

Optimisation of protoplast production in 

white lupin. Biol. Plant. 47: 21-25. 

Sutiojono, E., H.M. Nonhebel and A.S. 

Kantharajah. 1998. Factor affecting protoplast 

culture of Cucumis melo ‘Green Delica’. Ann. 

Bot. 81: 775-777. 

Vazquez-Tello, A., M. Hidaka and T. Uozumi. 

1995. Somatic embryogenesis and plant 

regeneration from isolated protoplasts of 

Lavatera thuringiaca. Plant Cell Tiss. Org. 

Cult. 40: 169-177. 


Widholm, J.M. 1972. The use of fluorescein 

diacetate and phenosafranine for determining 

viability of cultured plant cells. Stain Techno. 

47: 189-194. 

Xu, Z.H. and H.W. Xue. 1999. Plant regeneration 

from culture protoplasts, pp. 167-180. In W.Y. 

Soh and S.S. Bhojwani (eds.). 

Morphogenesis in Plant Tissue Culture. 

Kluwer Academic Publishing, Dorddrecht. 

Zhu, L., B.Wang, J. Zhou, L. Chen, C. Dai and C. 

Duan. 2005. Protoplast isolation of callus in 

Echinacea angustifolia. Coll. Surf. B: 

Biointerfaces 44: 1-5.


Anti HSV-1 Activity of Spirulina platensis Polysaccharide 

Nattayaporn Chirasuwan 1 *, Ratana Chaiklahan 1 , Marasri Ruengjitchatchawalya 2 

ABSTRACT 

Boosya Bunnag 2 and Morakot Tanticharoen 3 

Aqueous extracts of Spirulina platensis were precipitated by cetyltrimethylammonium bromide 

(CTAB). The hot water extract was found anti Herpes simplex virus type 1 (HSV-1) activity at 50% 

inhibitory concentration (IC 50) values of 21.32 µg/ml. Partial purification by gel filtration of the crude 

extract on Sepharose 6B column gave two fractions, SHP-F1 and SHP-F2, which revealed about 4 and 

2 times higher activity than that of the crude hot water extract, respectively. The crude hot water extract 

was a polysaccharide with rhamnose as the main sugar component. Calcium ion and sulfate groups in 

this polysaccharide had major roles in antiviral activity. However, the crude hot water extract 

polysaccharide contained approximately 42% carbohydrate and 31% protein. Decreasing the amount 

of protein by precipitation with trichloroacetic acid (TCA) resulted in higher purity of the crude hot 

water extract polysaccharide. 

Key words: Spirulina platensis, Herpes simplex virus type 1 (HSV-1), polysaccharide 

INTRODUCTION 

Spirulina platensis is one of the edible 

microalgae that has been used as health food and 

feed for a long time. There is an increased interest 

in components of S. platensis because of their 

potential properties such as anti thrombin activity 

(Hayakawa et al., 1996), lowering cholesterol level 

and blood pressure (Kato et al., 1984; Nakaya et 

al., 1988). Herpes simplex virus type1 is a 

common human pathogen causing infections of 

the orofacial mucosal region (Whitley and 

Roizman, 2001). Over the past decade, the 

incidence and severity of HSV infection have 

increased due to the increase in number of 

immuno-compromised patients produced by 

aggressive chemotherapy treatments, organ 

transplant and human immunodeficiency 

infections. Acyclovir, a synthetic drug which has 

remarkable effect against HSV-1 infection, inhibits 

virus replication by acting on viral DNA synthesis 

(Elion et al., 1977; Schaeffer et al., 1978). 

Acyclovir-resistant HSV infections have emerged 

due to the increase in drug use frequency (Field 

and Biron, 1994). Therefore, many researchers 

have attempted to search for effective and 

1 Pilot Plant Development and Training Institute King Mongkut’s University of Technology Thonburi, Bangkhuntien, 

Bangkok 10150, Thailand. 

2 School of Bioresources and Technology, King Mongkut’s University of Technology Thonburi, Bangkhuntien, 

Bangkok 10150, Thailand. 

3 National Center for Genetic Engineering and Biotechnology, National Science and Technology Development Agency, 

Pathum Thani 12120, Thailand. 

* Corresponding author, e-mail: nattayaporn@pdti.kmutt.ac.th, nattayaporn1@yahoo.com 


312 

inexpensive anti-viral agents from natural sources. 

The inhibitory effects of polysaccharides from 

marine algae on virus replication were first 

reported almost four decades ago. 

Gerber et al. (1958) reported that algal 

polysaccharides exhibited antiviral activity toward 

mumps and influenza B virus. Further, Hayashi et 

al. (1993) reported the anti HSV-1 activity of 

aqueous extracts from S. platensis. Our 

preliminary study revealed that both water soluble 

and non-polar extracts of S. platensis exhibited 

antiviral activity (HSV-1). This study investigated 

the anti HSV-1 activity of polysaccharides (water 

soluble compound) extracted from S .platensis. 

The isolation, partial purification, and composition 

determination of the anti HSV-1 activity extracts 

are described. 


Extraction of polysaccharides 

The lipid component was extracted from 

freeze-dried powder of S. platensis with 

CHCl 3:MeOH (2:1). Then, the residue was 

extracted with distilled H 2O. After fillration, the 

filtrate was precipitated by 1% CTAB 

(cetyltrimethylammonium bromide). After 

centrifuging, the precipitant was washed stepwise 

with saturated sodium acetate in 95% EtOH, 95% 

EtOH, absolute EtOH and diethyl ether, 

respectively. The resulting in cold water extract 

polysaccharide was obtained. For the extraction 

of hot water extract polysaccharide, the same 

method was performed except using boiling H 2O. 

Partial purification of polysaccharide 

The hot water extract polysaccharide was 

dissolved in 0.01 M citrate buffer, pH 7.0 

containing 0.1 M NaCl. The soluble portion was 

applied on to a Sepharose 6B (Pharmacia) column 

(3×30 cm) and eluted with the same citrate buffer. 

Fractions of 5 ml were collected and monitored 

using the phenol-sulfuric method with detection 


by spectrophotometer (Bausch&Lomb, Spectronic 

21) at an absorbance of 485 nm (Hayashi et al., 

1996a). The collected fraction was concentrated 

using an evaporator (below 40°C under reduced 

pressure), dialyzed with deionized water and 

lyophilized. 

Preparation of sugar derivatives for GC 

analysis 

One milligram of sugar was treated with 

1 ml of 20 g/l sodium tetraborohydride and cooled 

down to nearly 0°C. After standing over night, 

amberite IR-120 (H + ) was slowly added until no 

bubble. The solution was filtered through filter 

paper (Whatman #541). After filtration, the 

solution was evaporated under reduced pressure 

to thick syrup. The syrup was repeatedly dissolved 

in methanol and evaporated to remove boric acid. 

The syrup was further treated with 0.5 ml of acetic 

anhydride and 0.5 ml of pyridine at 80°C for 2 h. 

The solution was then immersed in an ice bath 

and 1 ml of methanol was added to the solution. 

The solution was then evaporated to remove 

methyl acetate. Then, 1 ml of heptane was added 

and evaporated to remove the remaining pyridine. 

Dried sample was dissolved in 200 µl of 

dichloromethane and analyzed by GC (Shimadzu, 

17A) using Rtx-2330 capillary column (Blakeney 

et al., 1983). 

Hydrolysis of hot polysaccharide solution 

A 5 mg sample from partial purification 

of polysaccharide and 1 mg of internal standard 

(inositol) were mixed and treated with 2 N H 2SO 4 

at 100°C for 8 h. The hot solution was neutralized 

with barium carbonate to pH 5 and filtered through 

filter paper (Whatman #541). Barium was 

eliminated from the supernatant using Amberite 

IR-120 (H + ) acidic cation-exchange resin. The 

solution was then applied to a Dowex 1-X8 anionexchange 

column and eluted with distilled H 2O. 

The fraction was evaporated and converted to an 

alditol acetate derivatives form and analyzed by

GC. 

Analytical methods Total carbohydrate 

content was estimated by phenol sulfuric acid 

assay (Dubois et al., 1956). Total protein content 

and lipid content were determined according to 

the methods of Lowry et al. (1951) and Folch 

Folch et al. (1957), respectively. Calcium content 

was carried out by Inductive Couple Plasma 

Spectroscopy (ICP, Model JY 124) and 

quantitative analysis of sulfate was performed by 

precipitation with 10% BaCl2 (Burns, 1995). 

Removal of calcium was achieved using 

a cation exchange column on Dowex 50W (X-8, 

H + form) (Hayashi et al., 1993). 

Desulfation pH of the polysaccharide 

solution was adjusted to pH 7.6 with pyridine and 

the pyridinium salt was eliminated with dimethyl 

sulfoxide (containing 10% of MeOH) at 80-100° 

C (Nagasawa et al., 1977). 

Antiviral activity was detected by using 

a colorimetric method modified from Skehan et 

al. (1990). Herpes simplex virus type 1 (HSV-1) 

was maintained in a Vero cell line (kidney 

fibroblasts of an African green monkey), which 

was cultured in Eagle’s minimum essential 

medium (MEM) with the addition of 10% heat 

inactivated fetal bovine serum (FBS) and 

antibiotics. The test samples were put into wells 

of a microtiter plate at final concentrations ranging 

from 20-50 µg/ml. The viral HSV-1 (30 PFU) was 

added into a 96-well microplate, followed by 

plating of Vero cells (1 × 105 cells/ml); the final 

volume was 200 µl. After incubation at 37°C for 

72 h, under 5% of CO2 atmosphere, cells were 

fixed with 50% trichloroacetic acid (TCA) and 

stained with 0.05% sulforhodamine B in 1% acetic 

acid and optical density was measured at 510 nm 

using a microplate reader. Acyclovir was used as 

the reference compound. 

Determination of cytotoxicity assay 

Compounds were tested for their 

cytotoxicity against Vero cells (African green 


monkey kidney fibroblasts in 96-well tissue culture 

plates). One hundred and ninety µl of Vero cell 

suspension containing 1 × 10 5 cells/ml and 10 µl 

of tested compound were added to each well in 

triplicate. Elliptine and 10%DMSO were used as 

positive and negative control, respectively. The 

cells were incubated at 37°C for 72 h in 5%CO 2. 

After incubation, the cytotoxicity was determined 

by the colorimetric method as described by 

Skehan et al. (1990). The cytotoxicity was 

expressed as IC 50, i.e., the concentration of the 

compound which inhibits cell growth by 50%, 

compared with untreated cell. 

RESULTS 

Crude cold and hot water polysaccharides 

were obtained from extraction of dried S. platensis 

by distilled water (room temperature) and boiling 

water, respectively. The extracts were precipitated 

by CTAB solution. The freeze-dried extracts as 

fine creamy powder were shown in Figure 1. The 

yields of the cold and hot water extracts were 1.2 

and 0.3 % (W/W), respectively. The hot water 

extract polysaccharide showed an IC 50 value 

against HSV-1 at 21.32 µg/ml whereas no activity 

was detected in the cold water extract 

polysaccharide. 

Figure 1 Cold water polysaccharide (pale color) 

and hot water polysaccharide (dark 

color).

314 

The crude hot water polysaccharide was 

partially purified by gel-filtration on Sepharose 6B 

column. Two fractions, SHP-F1 and SHP-F2, were 

collected (Figure 2). Attempts to completely 

separate the two fractions by decreasing flow 

rate from 1.2 to 0.8 ml/min was not successful. 

The curde hot water polysaccharide comprised of 

approximately 40% of fraction 1 (SHP-F1) and 

60% of fraction 2 (SHP-F2). The fractions of SHP- 

F1 and SHP-F2 were repeatedly run using the same 

method at a lower flow rate of 0.5 ml/min. 

Absorbance (OD485) 

4.5 

4 

3.5 

3 

2.5 

2 

1.5 

1 

0.5 

SHP-F1 

SHP-F2 

0 

0 10 20 30 40 50 60 70 80 

Fraction number 

Figure 2 Elution profile of hot water 

polysaccharide by Sepharose 6B 

column chromatography. 


4.5 

4 

3.5 

3 

2.5 

2 

1.5 

1 

0.5 

0 

0 10 20 30 40 50 60 70 80 



A 

After the pool fraction of SHP-F1 was 

repeatedly applied on Sepharose 6B column, the 

purified SHP-F1 was obtained (Figure 3A). 

However, SHP-F2 still exhibited 2 peaks of 

absorbances, a small peak and a bigger one, 

designated as SHP-F2/1 and SHP-F2/2, 

respectively (Figure 3B). When the partially 

purified fractions of the hot water polysaccharide 

(SHP-F1 and SHP-F2) were subjected to 

cytotoxicity and anti HSV-1 assays, both fractions 

exerted non-toxicity on the growth of Vero cells 

at the maximum concentrations tested and had 

significantly higher anti HSV-1 activity than the 

crude hot water polysaccharide (about 4 and 2 

times, respectively) (Table 1). 

The analysis of monosaccharide was 

performed by GC. It was found that SHP-F1 

fraction contained only three sugars; rhamnose, 

ribose and arabinose, whereas, the SHP-F2 fraction 

contained rhamnose, ribose, arabinose, glucose, 

mannose, galactose and xylose. Both fractions 

contained rhamnose as the main sugar component 

(Table 2). 

Table 3 showed the comparison of 

proximate analysis of dried cells of Spirulina and 

the crude hot water extract polysaccharide. Results 

showed that dried cells consisted of 21.9% 

carbohydrate, 61.4% protein, 7.2% lipid and 7.2% 


4.5 

4 

3.5 

3 

2.5 

2 

1.5 

SHP- F2/ 2 

1 

0.5 SHP- F2/ 1 

0 

. 

0 10 20 30 40 50 60 70 80 


Figure 3 Elution profile of SHP-F1 (A) and SHP-F2 (B) by Sepharose 6B column chromatography. 

B

ash. The crude hot water polysaccharide contained 

42.5% carbohydrate, 31.0% protein, 12.9% ash 

and trace of calcium and sulfate. 

Table 4 shows the remaining 

carbohydrate and protein precipitated by various 

concentrations of trichloroacetic acid (TCA). 

These data suggested that at the highest 

concentration of TCA (50% TCA), 41% of the 

protein in dry cells was eradicated (a decrease from 

31% to 18%), while the percentage of carbohydrate 

increased about 25% (from 42.5% to 53.1%). After 


the treated crude hot water polysaccharide 

(precipitated with 50% TCA) was tested for anti 

HSV-1 activity, results showed that the activity 

was not significantly different from the untreated 

crude hot water polysaccharide (data not shown). 

To determine the role of calcium ion and 

sulfate groups of the hot water polysaccharide in 

antiviral activity, calcium ion and sulfate groups 

in the polysaccharide were eliminated before 

testing for cytoxicity and anti HSV-1 activity. The 

results indicated that all of the calcium-free 

Table 1 Cytotoxicity and Anti HSV-1 activity of the hot water polysaccharide fractions from S. platensis. 

Fractions Cytotoxicitya Anti HSV-1b (IC50: µg/ml) (IC50: µg/ml) 

SHP-F1 > 50 5.25 

SHP-F2 > 50 9.61 

a maximum concentration of compound for cytotoxicity test was 50 µg/ml 

compound was non-toxic to the growth of Vero cells when IC50 >50 µg/ml 

(if compound was toxic on the growth of Vero cells, the compound will be subjected to the serial dilution 

for determination of IC50 value) 

b % inhibition of HSV-1; < 25%= inactive, 25-35%= weakly active, >35-50% = moderately active, 

> 50%= active (the compound will be subjected to the serial dilution 

for determination of IC 50 value) 

Table 2 Sugar composition of fractions of the hot water extract polysaccharide. 

Fractions % Sugar composition 

Rhamnose Ribose Arabinose Glucose Mannose Galactose Xylose 

SHP-F1 75.6 13.4 11.0 - - - - 

SHP-F2 30.4 27.1 10.0 18.2 7.5 4.5 2.3 

Table 3 Composition of S. platensis powder and crude hot water polysaccharide. 

Composition Dry weight (%) 

Spirulina powder Crude hot water polysaccharide 

Carbohydrate 21.9 ± 0.8 42.5 ± 0.3 

Protein 61.4 ± 1.1 31.0 ± 0.8 

Lipid 7.2 ± 1.3 0 

Calcium -* 0.123 ± 0.0006 

Sulfate -* 1.44 ± 0.03 

Ash 

Mean ± standard deviation (n = 3) 

7.2 ± 0.1 12.9 ± 0.4 

* It was not determined

316 

compound exerted weak anti HSV-1 activity when 

compared with that of the crude hot water 

polysaccharide, whereas, in the absence of sulfate 

groups in polysaccharide, no significant anti 

HSV-1 activity was detected in this compound 

(Table 5). 

DISCUSSION 

This study found that the hot water 

extract polysaccharide exhibited anti HSV-1 

activity, while the cold extract of the 

polysaccharide did not. Previous studies found that 

the majority of potential antiviral algal 

polysaccharides were extracted from tissues by hot 

water, dilute acid or alkali solution (Damonte et 

al., 1994; Hoshino et al., 1998). Crude hot water 

polysaccharide still contained a high level of 

protein which may co-precipitate when CTAB is 


used for polysaccharide precipitation (Tomanee et 

al., 2004). 

Partial purification of the hot water 

polysaccharide by gel-filtration on Sepharose 6B 

column gave 2 fractions (SHP-F1 and SHP-F2), 

both fractions effectively inhibited HSV-1 activity. 

Results reported by Hayashi et al. (1996a) revealed 

3 fractions (SP-H-1, SP-H-2 and SP-H-3) but only 

a SP-H-2 fraction had anti HSV-1 activity. The 

sugars found in SHP-F1 and SHP-F2 fractions in 

this study are almost the same as previously 

reported by Hayashi et al. (1966a) except for 

arabinose which was found in this study instead 

of fructose which was reported by the same 

researchers. 

Calcium ion and sulfate groups in the hot 

water polysaccharide were important for the anti 

HSV-1 activity. This result was supported by 

Hayashi’s study that when the calcium-free 

Table 4 Carbohydrate and protein content of crude hot water polysaccharide which was precipitated 

by trichloroacetic acid (TCA). 

TCA concentration (%) % w/w of crude hot water polysaccharide 

Carbohydrate Protein 

0 42.5 ± 0.3 31.0 ± 0.8 

10 41.8 ± 1.5 28.1 ± 2.1 

20 45.3 ± 1.8 25.4 ± 1.3 

30 49.2 ± 2.3 22.3 ± 1.5 

50 53.1 ± 2.0 18.0 ± 0.8 

Mean ± standard deviation (n = 3) 

Table 5 Anti HSV-1 activity in the crude hot water polysaccharides from S. platensis. 

Sample Cytotoxicity a (IC 50 : µg/ml) Anti HSV-1 b (IC 50 : µg/ml) 

Polysaccharide > 50 21.3 

Polysaccharide (-Ca 2+ ) > 50 38.4 

Polysaccharide (-SO 4 2- ) > 50 Inactive 

a maximum concentration of compound for cytotoxicity test was 50 µg/ml 

compound was non toxic on the growth of Vero cells if IC 50 >50 µg/ml 

(if compound was toxic on the growth of Vero cells, the compound will be subjected to the serial dilution 

for determination of IC 50 value) 

b % inhibition of HSV-1; < 25%= inactive, 25-35%= weakly active, >35-50% = moderately active, 

> 50%= active (the compound will be subjected to the serial dilution 

for determination of IC 50 value)

spirulan (H-SP), and a desulfated compound from 

Ca-SP were subjected to cytotoxicity and antiviral 

assay, both compounds exerted strong toxicity to 

the growth of host cell (HeLa cells) and weakly 

inhibited HSV-1 (Hayashi et al., 1996a). Ca-SP 

was found to inhibit replication of several 

enveloped virus and selectively inhibited the 

penetration of virus into host cell (Hayashi et al., 

1996b). Loya et al. (1998) postulated that the 

negatively charged (e.g., sulfonate vs. sulfate) may 

interact with the positively charged side chains on 

the DNA polymerase and Witvrouw et al. (1994) 

assumed that sulfated polysaccharides disruption 

of ionic interactions between positively charged 

regions of viral surface glycoproteins and cellular 

membrane phospholipids. 

CONCLUSION 

Results from this study demonstrated the 

significant potential of S. platensis polysaccharide 

for activity against HSV-1. The hot water extract 

polysaccharide which contained rhamnose as the 

main sugar component showed anti HSV-1 activity 

at IC 50 21.3 µg/ml. Calcium ion and sulfate groups 

in the polysaccharide had major roles in the anti 

HSV-1 activity. S. platensis, is a possible source 

for new drugs in the treatment of HSV-1 and other 

viral diseases. 


This study was supported by TRF (The 

Thailand Research Fund, RDG4330032). 


Blakeney, A. B., P. J. Harris, R. J. Henry and B. A. 

Stone. 1983. A simple and rapid preparation 

of alditol acetates for monosaccharide 

analysis. Carbohydr. Res. 113: 291-299. 

Burns, B. A. 1995. In Official Methods of Analysis 

of Association of Official Analytical Chemists 


International, Chapter 5, 16 th Ed Volume II, 

Edited by Cunniff, P., AOAC International, 

Virginia, p. 5 

Damonte, E. B., J. Neyts, C. A. Pujol, R. Snoeck, 

G. Andrei, S. Ikeda, M. Witvrouw, D. Reymen, 

H. Haines and M. C. Matulewicz. 1994. 

Antiviral activity of a sulfated polysaccharide 

from the red seaweed Nothogenia fastigiata. 

Biochem Pharmacol. 47: 2187-2192. 

Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. 

Rebers and F. Smith. 1956. Colorimetric 

Method for Determination of Sugars and 

Related Substances. Anal. Chem. 28: 350- 

356. 

Elion, G. B., P. A. Furman, J. A. Fyfe, P. De 

Miranda, L. Beauchamp and H. J. Schaeffer. 

1977. Selectivity of action of an antiherpetic 

agent, 9-guanine. Proc. Natl. Acad. Sci. USA. 

74: 5716-5720. 

Field, A. K. and K. K. Biron. 1994. The end of 

innocence revisited: resistance of herpes 

viruses to antiviral drugs. Clin. Microbiol. 

Rev. 7: 1-13. 

Folch, J., M. Lees and G. H. S. Stanley. 1957. A 

simple method for the isolation and 

purification of total lipids from animal tissues. 

J. Biol. Chem. 226: 497-509. 

Gerber, P., J. D. Dutcher, E. V. Adams and J. H. 

Sherman. 1958. Protective effect of seaweed 

extracts for chicken embryos infected with 

influenza B or mumps virus. Proc. Soc. Exp. 

Biol. Med. 99: 590-593. 

Hayakawa, Y., T. Hayashi, K. Hayashi, K. Ozawa, 

K. Niiya and N. Sakuragawa. 1996. Heparin 

cofactor II-dependent antithrombin activity of 

calcium spirulan. Blood Coagul. Fibrinol. 7: 

554-560. 

Hayashi, K., T. Hayashi and M. Morita. 1993. An 

extract from Spirulina platensis is a selective 

inhibitor of Herpes simplex virus type 1 

penetration into HeLa cells. Phytother. Res. 

7: 76-80. 

Hayashi, T., K. Hayashi, M. Maeda and I. Kojima.

318 

1996a. Calcium Spirulan, an inhibitor of 

enveloped virus replication, from a blue-green 

alga Spirulina platensis. J. Nat. Prod. 59: 

83-87. 

Hayashi, K., T. Hayashi and I. Kojima. 1996b. A 

natural sulfated polysaccharide, Calcium- 

Spirulan, isolated from Spirulina platensis: in 

vitro and ex vivo evaluation of anti-herpes 

simples virus and anti-human 

immunodeficiency virus activities. AIDS Res. 

Hum. Retroviruses. 12: 1463-1471. 

Hoshino, T., T. Hayashi, K. Hayashi, J. Hamada, 

J. B. Lee and U. Sankawa. 1998. An antivirally 

active sulfated polysaccharide from 

Sargassum horneri (TURNER) C. 

AGARDH. Biol. Pharm. Bull. 21: 730-734. 

Kato, T., K. Takemoto, H. Katayama and Y. 

Kuwabara. 1984. Effects of Spirulina 

(Spirulina platensis) on dietary 

hypercholesterolemia in rats. Jap. Nutr. Food 

Assoc. J. 37: 323-332. 

Lowry, O. H., N. J. Rosenbrough, A. L. Farr and 

K. J. Kandall. 1951. Protein measurement with 

the Folin phenol reagent. J. Biol. Chem. 193: 

265-275. 

Loya, S., V. Reshef, E. Mizrachi, C. Silberstein, 

Y. Rachamim, S. Carmeli and A. Hizi. 1998. 

The inhibition of the reverse transcriptase of 

HIV-1 by the natural sulfoglycolipids from 

cyanobacteria: Contribution of different 

moieties to their high potency. J. Nat. Prod. 

61: 891-895. 


Nagasawa, K., Y. Inoue and T. Kamata. 1977. 

Solvolytic desulfation of 

glycosaminoglycuronan sulfates with 

dimethyl sulfoxide containing water or 

methanol. Carbohydr. Res. 58: 47-55. 

Nakaya, N., Y. Honma and Y. Goto. 1988. 

Cholesterol lowering effect of Spirulina. 

Nutr. Rep. Int. 37: 1329-1337. 

Schaeffer, H. J., L. Beauchamp, P. De Miranda, 

G. B. Elion, D. J. Bauer and P. Collins. 1978. 

9-(2-hydroxyethoxymethyl) guanine activity 

against viruses of the herpes group. Nature. 

272: 583-585. 

Skehan, P., R. Storeng, D. Scudiero, A. Monks, J. 

Mcmahon, D. Vistica, J. T. Warren, H. 

Bokesch, S. Kenney and M. R. Boyd. 1990. 

New colorimetric cytotoxic assay for 

anticancer - drug screening. J. Nat. Can. Inst. 

82: 1107-1112. 

Tomanee, P., J. T. Hsu and Y. Ito. 2004. 

Fractionation of protein, RNA, and plasmid 

DNA in centrifugal precipitation 

chromatography using cationic surfactant 

CTAB containing inorganic salts NaCl and 

NH 4Cl. Biotechnol Bioeng. 88: 52-59. 

Whitley, R. J. and B. Roizman. 2001. Herpes 

simplex virus infections. Lancet. 357: 1513- 

1518. 

Witvrouw, M., J. Desmyter and E. De Clercq. 

1994. Antiviral portrait series. 4. Polysulfates 

as inhibitors of HIV and other enveloped 

viruses. Antiviral Chem. Chemother. 5: 345- 

359.


Taura Syndrome Virus Disease in Farm-Reared Penaeus monodon 

in Thailand 

ABSTRACT 

Chalor Limsuwan and Niti Chuchird* 

Taura syndrome virus (TSV) has caused major economic losses to shrimp aquaculture throughout 

the world. TSV has been reported to infect a number of penaeid species as hosts. In this study, we 

reported the natural infection of TSV in farm-reared Penaeus monodon from eastern provinces of Thailand 

between June to September 2004. There were different degrees of disease outbreak severity. In some 

cases large number of shrimp died and caused great losses to farmers. However, in most cases only 

small number of shrimp died and the farmers could control the situation enough to raise the majority to 

marketable size. Diseased shrimp varied in size from aged 40-50 days (4 g) to 20 g. Infected shrimp was 

characterized by black cuticular lesions and loose shell. Histopathological changes in infected shrimp 

showed multifocal to extensive areas of necrosis in the sub-cuticular epithelium, connective tissue and 

adjacent striated muscle. Affected cells often displayed an increased cytoplasmic eosinophilia, nuclear 

pyknotic and karyorrhexis. In situ hybridization tests gave positive results with the tissues of shrimp 

collected from the TSV outbreaks. In addition to TSV infection, most moribund shrimp also had dual 

infections with microsporidians in the hepatopancreas and/or gregarines in the gut. 

Key words: Taura syndrome virus, Penaeus monodon 

INTRODUCTION 

Taura syndrome was first recognized as 

a shrimp disease in farms near the mouth of the 

Taura river, Ecuador, in June 1992 (Jimenez, 1992; 

Rosenbery, 1993; Lightner et al., 1994). The 

infectious agent was named Taura syndrome virus 

or TSV in 1994 (Hasson et al., 1995; Lightner et 

al., 1995). TSV was first isolated from Litopenaeus 

vannamei and characterized as a non-enveloped, 

icosahedral particle, 31-32 nm in diameter, with a 

density of 1.338 g/ml in CsCl. Its genome consists 

of a linear, positive sense ssRNA molecule of 

approximately 10.2 kb and it is classified as a 

Picornavirus (Bonami et al., 1997; Brock et al., 

1997). From nucleotide sequence data, TSV is 

more closely related to the cricket paralysis- like 

viruses (Mari et al., 2002). 

The occurrences of TSV outbreaks in L. 

vannamei include cultured shrimp stocks in 

Hawaii, Peru, Ecuador, Colombia, Panama, Costa 

Rica, Nicaragua, El Salvador, Honduras, 

Guatemala and Mexico (Lightner 1996). The 

outbreaks were reported for P. stylirostris, P. 

setiferus and P. schmitti in Ecuador and Peru 

(Lightner et al., 1995; Brock et al., 1997). In Asia, 

TSV was first reported from Taiwan in 1999 (Tu 

et al., 1999). 

Department of Fishery Biology, Faculty of Fisheries, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: ffisntc@ku.ac.th 


320 

In Thailand, since early 2000, the 

cultivation of black tiger shrimp, Penaeus 

monodon, had suffered slow growth, leading 

shrimp farmers to shift to the cultivation of L. 

vannamei. Most of the nauplii were illegally 

imported from China and Taiwan. Alarmed by the 

possibility of TSV introduction, the Thai 

Department of Fisheries permitted legal 

importation of L. vannamei from March 2002- 

February 2003, if the imported stocks were 

certified free of TSV by RT-PCR testing. However, 

in early 2003, TSV outbreaks occurred in inland 

farm-reared L. vannamei (Limsuwan, 2003; 

Nielsen et al., 2005). Since then more TSV 

outbreaks were reported in L. vannamei in most 

areas of cultivation. Shortly thereafter, in early 

2004, mortalities were observed in P. monodon 

intensive culture ponds. Diseased shrimp were 

PCR-negative for both white spot syndrome virus 

(WSSV) and yellow-head virus (YHV) but 

positive for TSV. This disease was widespread and 

caused heavy mortalities to some farms. 

This paper describes an epizootic of TSV 

including gross signs, histopathology and in situ 

hybridization in intensively reared P. monodon in 

Thailand. 


Penaeus monodon samples were 

collected from TSV-affected farm ponds in the 

eastern provinces of Thailand during June to 

September 2004. The shrimp samples weighing 

of 4-20 g were preserved in Davidson’s fixative 

solution and then transferred to 70% ethanol after 

48 h. All histological materials were prepared 

using standard histological procedures for shrimp 

and stained with haematoxylin and eosin (H&E) 

as described in Bell and Lightner (1988). A 

commercially available in situ hybridization probe 

for TSV (Diagxotics Inc.) was used according to 

the manufacturer’s instructions. The protocols 

have been outlined by Lightner (1996) and Mari 


et al. (1998). 


Taura syndrome virus has been reported 

to infect a number of penaeid species as hosts. 

However, only Litopenaeus vannamei appears to 

be highly susceptible to the disease (Lightner 

1996). Overstreet et al. (1997) and Lightner (1996) 

reported natural TSV infections in P. setiferus and 

experimental infections have been reported in 

P. schmitti, P. aztecus, P. duoraram, P. chinensis, 

P. monodon, and P. japonicus. TSV may be 

transmitted horizontally by co-habitation or 

cannibalism (Lotz et al., 2003). In Thailand, TSV 

was first reported from intensive farm-reared P. 

monodon in June 2004. Moribund shrimp aged 

40-50 days were found in scattered areas around 

the edges of the pond. Although in some farms it 

could be found in younger or older shrimp as well. 

Diseased shrimp was characterized by black 

cuticular lesions and loose shell. Shrimp with these 

black lesions are at some risk of mortality during 

the succeeding molt, but if they survive, lesions 

disappear from the cuticle and shrimp look normal. 

However, affected shrimp did not display signs of 

red body or tail (Figure 1 and 2) which was 

different from the report of Lightner et al. (1995) 

indicated that the expansion of red chromatophores 

in the appendages, especially of the uropods, 

telson, and pleopods of L. vannamei infected with 

TSV. There were different degrees of disease 

outbreak severity in cultured P. monodon. In some 

cases large number of shrimp died quickly and 

caused great losses to farmers. However, in most 

cases only small number of shrimp died and the 

farmers could control the situation enough to raise 

the majority to marketable size and harvest them 

for sale. 

Histopathology of moribund shrimp 

showed multifocal to extensive areas of necrosis 

in the sub-cuticular epithelium, connective tissue 

and adjacent striated muscle (Figure 4). Affected

cells often displayed an increased cytoplasmic 

eosinophilia, nuclear pyknosis and karyorrhexis 

(Figure 5). Some samples also showed necrosis 

in the cells of haematopoietic tissue corresponded 

to those previously described for TSV infections 

(Lightner et al., 1995). In situ hybridization tests 

also gave positive results with the tissues of shrimp 

collected from the TSV outbreaks (Figure 6). In 

addition to TSV infection, most moribund shrimp 

also had infections with microsporidians in the 

hepatopancreas (Figure 7) and gregarines in the 

gut (Figure 8). These protozoans are highly 

pathogenic and frequently cause epizootics in 

Figure 1 Moribund shrimp with TSV during the 

first 2 months of culture with multiple 

melanized cuticular lesions. 

Figure 3 Normal subcuticular epidermal and 

connective tissue (H&E). 


crustacean populations (Overstreet,1973; 

Sindermann, 1990). Sprague and Couch (1971) 

indicated that in addition to microsporidians, 

shrimps in the ponds often harbor cephaline 

gregarines, similar to the results in this report. 

Brock et al. (1997) reported experimental infection 

of P. monodon with TSV and indicated that P. 

monodon was susceptible to TSV but suffered few 

mortalities. To avoid TSV infections or a 

significant outbreak of the disease, farmers must 

have sufficient reservoir ponds available and only 

refill the shrimp ponds or stocking postlarvae into 

the pond with water that has been left to rest for at 

Figure 2 Affected shrimp at harvest with 

multiple black melanized cuticular 

lesions. 

Figure 4 Typical TSV lesion showing area of 

extensive subcuticular epidermal and 

connective tissue necrosis (H&E).

322 

least 15 days (Chuchird and Limsuwan, 2005). It 

will then be less likely that the virus will be alive 

in the water and the farmers will have a greater 

chance of rearing a good harvest of shrimp. 

CONCLUSION 

Gross sign of TSV in P. monodon was 

characterized by black cuticular lesions and loose 

shell. Histologically, sub-cuticular lesions were 

characterized by large numbers of spherical 

Figure 5 Higher magnification of TSV lesion 

with numerous nuclear pyknosis (P) 

and karyorrhexis (K), (H&E). 

Figure 7 Microsporidians (arrows) infection in 

the hepatopancreas of TSV infected 

shrimp (H&E). 


eosinophilic to densely basophilic inclusions and 

gave the tissue a kind of “buck-shot” appearance. 

Most moribund shrimp had dual infections with 

microsporidians in the hepatopancreas and/or 

gregarines in the gut. 


The authors would like to thank the 

National Research Council of Thailand (NRCT) 

for financial support. 

Figure 6 Tissue section of cuticular epithelium 

with positive in situ hybridization 

reaction for TSV (arrows). 

Figure 8 Gregarine (arrow) in the gut of TSV 

infected shrimp (H&E).


Bell, T.A. and D.V. Lightner. 1988. A Handbook 

of Normal Shrimp Histology. World 

Aquaculture Society. 

Bonami, J.R., K.W. Hasson, J. Mari, B.T. Poulos 

and D.V. Lightner. 1997. Taura syndrome of 

marine penaeid shrimp: characterization of the 

viral agent. J. Gen. Virol. 78: 313-319. 

Brock, J.A., R.B. Gose, D.V. Lightner and K.W. 

Hasson. 1997. Recent developments and an 

overview of Taura Syndrome of farmed 

shrimp in the Americas, pp. 267-283. In T. W. 

Flegel and I.H. MacRae, eds. Diseases in 

Asian Aquaculture III. Fish Health Section, 

Asian Fisheries Society, Manila, Philippines. 

Chuchird, N. and C. Limsuwan. 2005. The 

viability of Taura syndrome virus on lowsalinity 

water. Kasetsart J. (Nat. Sci.) 39: 

406-410. 

Hasson, K.W., D.V. Lightner, B.T. Poulos, R.M. 

Redman, B.L. White, J.A. Brock and J.R. 

Bonami. 1995. Taura syndrome in Penaeus 

vannamei : Demonstration of a viral etiology. 

Dis. Aquat. Org. 23: 115-126. 

Jimenez, R. 1992. Syndrome de Taura (Resumen). 

Aqucultura del Ecuador 1: 1-16. 

Lightner, D.V. 1996. A Handbook of Pathology 

and Diagnostic Procedures for Diseases of 

Penaeid Shrimp.World Aquaculture Society. 

Lightner, D.V., R.M. Redman, B.T. Poulos, J.L. 

Mari, J.R. Bonami and M. Shariff. 1994. 

Distinction of HPV-type virus in Penaeus 

chinensis and Macrobrachium rosenbergii 

using a DNA probe. Asian Fisheries Science 

7: 267-272. 

Lightner, D.V., R.M. Redman, K.W. Hasson and 

C.R. Pantoja. 1995. Taura syndrome in 

Penaeus vannamei (Crustacea: Decapoda): 

gross signs, histopathology and ultrastructure. 

Dis. Aquat. Org. 21: 53-59. 

Limsuwan, C. 2003. Diseases of Pacific White 

Shrimp (Litopenaeus vannamei) in Thailand. 


AAHRI Newsletter 12(1): 1-4. 

Lotz, J.M. , A.M.Flowers and V. Breland. 2003. A 

model of Taura syndrome virus (TSV) 

epidemics in Litopenaeus vannamei. J. 

Invertebr. Pathol. 83: 168-176. 

Mari, J., J.R. Bonami and D.V. Lightner. 1998. 

Taura syndrome of penaeid shrimp:cloning of 

viral genome fragments and development of 

specific gene probes. Dis. Aquat. Org. 33: 

11-17. 

Mari, J., B.T. Poulos, D.V. Lightner and J.R. 

Bonami. 2002. Shrimp Taura syndrome virus: 

genomic characterization and similarity with 

members of the genus Cricket paralysis-like 

viruses. J. Gen. Virol. 83: 915–26. 

Nielsen, L., W. Sang-oum , S. Cheevadhanarak 

and T.W. Flegel. 2005. Taura syndrome virus 

(TSV) in Thailand and its relationship to TSV 

in China and the Americas. Dis. Aquat. Org. 

63(2-3): 101-106. 

Overstreet, R.M. 1973. Parasites of some penaeid 

shrimps with emphasis on reared hosts. 

Aquaculture 2: 105-140. 

Overstreet, R.M, D.V. Lightner, K.W. Hasson, S. 

McIIwain and J.M. Lotz. 1997. Susceptibility 

to TSV of some penaeid shrimps native to 

the Gulf of Mexico and Southeastern US. 

J. Invertebr. Pathol. 69: 165-176. 

Rosenbery, B. 1993. World Shrimp Farming 1993. 

Annual Report Shrimp News International. 

52 p. 

Sindermann, C.J. 1990. Principle Diseases of 

Marine Fish and Shellfish, 2 nd ed. Academic 

Press. 

Sprague, V. and J.A. Couch. 1971. An annotated 

list of protozoan parasites, hyper-parasites and 

commensals of decapod Crustacea. J. 

Parasitol. 18: 526-573. 

Tu, C., H. Huang, S. Chuang, J. Hsu, S. Kuo, N. 

Li, T. Hsu, M. Li and S. Lin. 1999. Taura 

syndrome in Pacific white shrimp Penaeus 

vannamei culture in Taiwan. Dis. Aquat. Org. 

38: 159-161.


Optimization of Docosahexaenoic Acid (DHA) Production and 

Improvement of Astaxanthin Content in a Mutant Schizochytrium 

limacinum Isolated from Mangrove Forest in Thailand 

Wassana Chatdumrong 1 , Wichien Yongmanitchai 1 *, Savitree Limtong 1 

ABSTRACT 

and Wanchai Worawattanamateekul 2 

Polyunsaturated fatty acids including DHA are essential dietary fatty acids. At present, fish 

oils are a major source, but an alternative supply is needed because of increasing demand and fish 

dwindling stocks. This need might be satisfied using a thraustochytrids found in mangrove forests of 

Thailand and identified by 18S rDNA sequencing as either Schizochytrium limacinum or Thraustochytrium 

aggregatum. S. limacinum was tested in various culture conditions to find the optimal yield of DHA. 

This culture medium contained 7.5% glucose, 0.5% peptone, 0.5% yeast extract (with either 0.25% 

soybean meal or 1% skimmed milk) and 0.75% sea salt at 20-30°C. The C:N ratio was about 15:1. The 

culture was mutated using NTG and one isolate showed high DHA content and also a red pigment 

identified as astaxanthin by TLC and HPLC. Astaxanthin synthesis peaked on day 6 - 10 of incubation 

in medium containing 2% glucose using shaking flasks at 180 rpm, 25°C, 2 kLux light intensity with a 

18:6 h light:dark periods. Six days of incubation yielded the highest yields of both DHA (224.6 mg/l) 

and astaxanthin (8.9 µg/ml of medium). These results suggested that this microorganism could provide 

a commercial source of this valuable lipid and pigment. 

Key words: astaxanthin, Schizochytrium, DHA, mutation, mangrove forest, Thailand 

INTRODUCTION 

Thraustochytrids such as Schizochytrium 

and Thraustochytrium are aquatic heterotrophic 

microorganisms commonly found in marine and 

estuarine environment (Barr, 1992). The capacity 

of thraustochytrids to accumulate large amounts 

of polyunsaturated fatty acids (PUFAs), especially 

omega-3 fatty acids including docosahexaenoic 

acid (C22:6, DHA), is well recognized (Lewis et 

al., 1999; Huang et al., 2001). They are important 

in preventing and treating pathologies such as 

coronary heart disease, stroke and rheumatoid 

arthritis (Kinsella, 1987), provide protection 

against asthma, dyslexia, depression and some 

forms of cancer (Simopoulos, 1989; Takahata et 

al., 1998). DHA is an essential fatty acid for 

neuronal development (Yongmanitchai and Ward, 

1989). Demand of these fatty acids as a dietary 

supplement has increased and the major supply is 

presently derived from fish oil. But dwindling fish 

stocks and increasing demand has created a need 

1 Department of Microbiology, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

2 Fishery Product Department, Faculty of Fishery, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: fsciwcy@ku.ac.th 


for alternative sources of supply. 

Astaxanthin (3,3’-dihydroxy-β,βcarotene-4,4’dione) 

is a carotenoid found 

especially in marine crustaceans. It is added to 

food products (Vazquez et al., 1997) and use as a 

colorant for cultured fish, poultry (Johnson and 

An, 1991) and shrimp. It also acts as a scavenger 

of free oxygen radicals which damage DNA and 

oxidizes proteins (Schroeder and Johnson, 1993). 

Astaxanthin used as an animal feed is often 

produced commercially by chemical synthesis. 

However, the public have a preference for 

additives coming from natural source (Fang and 

Cheng, 1993) such as algae, fungi and small 

crustaceans. When in the food chain, they lead to 

pigmentation of larger animals including fish 

(especially salmon), lobsters, krill and small 

marine and freshwater organisms (Johnson and 

Lewis, 1979). 

Recently, microbial production of 

astaxanthin pigment has been improved through 

isolated or combined strategies, i.e., mutagenesis 

and media fermentation (Fontana et al., 1996). The 

thraustochytrids, Schizochytrium aggregatum 

(Valadon, 1976) and Thraustochytrium CHN-1 

(Marvelisa et al., 2003), have both been found to 

contain this pigment. This study aims to improve 

both the astaxanthin and DHA production by the 

creation of mutations of Schizochytrium sp. 

BR2.1.2 and also by optimizing the media and 

conditions in small scale cultures and then 

applying to larger vessels. 


Microorganisms 

Wild type strain 

The wild type of thraustochytrid selected 

strain BR2.1.2 was isolated from mangrove forest 

at Bang-rong area, Amphur Thalang in Phuket 

province, Southern Thailand. The isolation was 

carried out in GPY agar medium (Huang et al., 

2001) by baiting technique. After a series of 


streaking on the agar medium, pure culture was 

obtained for this study. 

Identification of microorganism by 

18S rDNA sequencing 

Morphological characteristics of 

thraustochytrid BR2.1.2 resembled 

Schizochytrium. Identification was further 

confirmed by 18S rDNA sequencing. Two primers 

of NS1 and NS8 were used for amplification of 

18S rDNA by PCR technique using a thermal 

cycler (Perkin Elmer GeneAmp PCR system 

2400). The amplification program was carried out 

following the protocol of Mo et al. (2001). Purified 

PCR product of 18S rDNA was analyzed by DNA 

autosequencer with NS1-8 primers set according 

to White et al. (1990) 

Mutagenesis 

Increased expression of astaxanthin was 

sought by mutagenesis of the wild type BR2.1.2 

using N-methyl-N’-nitro-N-nitrosoguanidine 

(NTG) (Fluka Chem, AG) modified from Chaunpit 

(1993). The initial concentration of 1-9×10 6 cells/ 

ml was treated with NTG (0.1 mg/ml) for 20 min 

with shaking. NTG was removed from suspension 

by centrifugation and cells pellet washed by 0.5 

M phosphate buffer pH 7.0 and spreaded on GPY 

agar plate. The treated culture contained 0.05 – 

0.1 % of the initial cells. Red colonies indicating 

accumulation of astaxanthin were collected for 

further assessment of growth, astaxanthin and 

DHA contents. 

Optimization of growth and DHA production 

by wild type BR2.1.2 

Culture conditions for thraustochytrid 

BR2.1.2 were optimized for growth and DHA 

production. The following conditions were applied 

throughout unless otherwise stated. Thirty 

milliliters of GPY medium (Huang et al., 2001) 

composed of 3% glucose, 1% peptone, 0.5% yeast 

extract and 50% of natural sea water was used as

326 

the basal medium and placed in a 125 ml 

Erlenmeyer flask. All experiments were carried 

out in triplicate flasks. Cultivation was initiated 

by addition of 1 ml of inoculum (adjusted cells 

concentration to 1.0 at OD 600 nm). Incubation 

was done on a rotary shaker (Sac Science-ENG 

LTD, Part) at 140 rpm at room temperature for 4 

days. To test different media, the basal media was 

modified in the following manner: 

1. The carbon source replaced by either 

glucose, fructose, sucrose, glucose syrup and 

agricultural products i.e. molasses, and sugar cane 

juice (Sahakarnnamtan Co. Ltd., Chonburi, 

Thailand). 

2. The nitrogen sources replaced by 

peptone, soybean meal, skimmed milk, 

ammonium sulfate, potassium nitrate, sodium 

nitrate, monosodium glutamate (MSG). 

3. Sea salt concentration (salinity) 0- 

200% of sea water. 

The effect of temperature on growth and 

DHA production were also determined by using 

temperature gradient incubator (Model TN-3, 

Toyo, Kagaku Sangyo Co., Ltd., Tokyo, Japan) 

set at 15, 20, 25, 30 and 35°C. 

Optimization of growth and astaxanthin 

production by thraustochytrid mutant 

The mutant was cultivated in a 125 ml 

Erlenmeyer flask containing 30 ml of GYC 

medium (Marvelisa et al., 2003) and kept in an 

incubator shaker at 180 rpm for 10 days at 25°C. 

Light was provided by fluorescent lamps at the 

intensity of 2 kLux with light:dark periods at 16:8 

hrs. Effects of carbon sources such as sugar cane 

juice, molasses and maltose:glucose (1:1, w/w) 

contained the same carbon equivalent as 2% 

glucose were studied. Environmental conditions 

such as light intensity at 0, 5 and 10 kLux and 

temperature (as described above) were also 

determined. 


Analytical procedures 

Growth was determined as the dry weight 

of the cells (drying conditions). 

Lipid was extracted by the modified 

method of Bligh and Dyer (1959), followed by 

transmethylation according to Holub and Skeaff 

(1987). Fatty acid methyl esters were analyzed in 

a gas-liquid chromatography (GC-14B; Shimadzu, 

Tokyo, Japan) equipped with flame ionization 

detector and a split injector at 1:40 ratio using 

capillary column in 30 m length, 0.25 mm internal 

diameter, 0.25 mm. film thickness (AT-WAX, 

Alltech Associates Inc, USA). Fatty acids were 

identified by comparing retention times with 

authentic standards from Sigma by using C-R6A 

Chromatopac Data Integrator (Shimadzu, Japan). 

The astaxanthin content was determined 

by the method modified from Fontana et al. (1996). 

The concentration was quantified by using 

absorbance values at 479 nm calculated with the 

specific absorption coefficient a (1cm,1%) = 1600 as 

proposed by Anderwes et al. (1976). Isomers of 

astaxanthin were identified by thin layer 

chromatography according to Donkin (1976) 

compared with reference standards extracted from 

Haematococcus pluvialis. These determinations 

were confirmed by HPLC (model HP1100, Agilent 

Technology) following the procedure of Marvelisa 

et al. (2003). 


Identification of thraustochytrid BR2.1.2 by 

18S rDNA sequence analysis 

The corrected partial sequence of 18S 

rDNA of thraustochytrid BR2.1.2 was 912 bases 

in length after gaps, inserts and ambiguous 

positions had been removed and was deposited in 

DDBJ as accession number 794133. A 

phylogenetic tree was constructed from an 

alignment of the BR2.1.2 sequence with those 

from related species obtained from GenBank by 

the NJ method (Figure 1). It was clearly seen that

BR2.1.2 formed the same clade with 

Thraustochytrium aggregatum and Schizochytrium 

limacinum but with slight distance. Hence the 

strain BR2.1.2 was finally identified as 

Schizochytrium limacinum. 

Effect of culture conditions on growth and DHA 

production by S. limacinum BR2.1.2 

1. Carbon source 

Among the various carbon sources 

tested, S. limacinum BR2.1.2 exhibited highest 

growth rates in 3% fructose and glucose with 14.3 

and 13.4 g/l of CDW, respectively (Figure 2A). 

DHA yields were 392.5 and 362.1 mg/l with DHA 


contents at 49.1 and 49.7% of TFA, respectively. 

Although, relatively good growth rates were 

obtained in complex carbon sources, i.e. molasses 

(10.5 g/l) and sugar cane juice (11.5 g/l), DHA 

production was low. Sucrose and glucose syrup 

were poorer carbon source for this organism. The 

results coincided with those of Wu et al. (2005) as 

glucose syrup contained mainly oligosaccharides 

that could not support growth for many 

microorganisms. Although glucose was slightly 

inferior compared to fructose, it is considered to 

be the good carbon source, because it was ready 

available and substantially cheaper. 

Figure 1 Phylogenetic tree reconstruction based on 18S rDNA sequence by neighbour-joining (NJ) 

method. The number at each branch shows bootstrap values 1000 replications.

328 

Figure 2B demonstrates the effect of 

glucose concentration. Cell mass depended on 

glucose concentrations and was maximal with 7% 

(cell mass 28.3 g/l). However, the highest DHA 

production was obtained in 5% glucose (732 

mg/l) making up 50.6% of TFA. When glucose 

concentration was increased to 7%, the proportion 

of DHA (641.1 mg/l) was 44.6% of TFA. 

2. Nitrogen source 

Further experiments used 5% glucose 

and 0.5% yeast extract to determine the effect of 

the nitrogen source in Figure 3A. Results revealed 

CDW (g/l); DHA (% of TFA)) 


60 

50 

40 

30 

20 

10 

0 

60 

50 

40 

30 

20 

10 

0 

Fructose Glucose Glucose 

Syrup 


Carbon source 

that among complex nitrogen sources (1% 

peptone) was the best in supporting growth for 

both CDW (20.9 g/l) and DHA (828.2 mg/l). 

Soybean meal and skimmed milk although 

relatively good nitrogen source for CDW but DHA 

production was considerably lower at 441.8 and 

545.9 mg /l, respectively. In the medium 

containing 0.2% MSG, BR2.1.2 grew at 20.3 g/l 

and produced DHA 768.5 mg/l, almost the same 

levels as supported by 1% peptone. This result 

agrees with those using with Thraustochytriun 

aureum ATCC 34304 that grew well in medium 

containing glucose, peptone, yeast extract and 

Molasses Sucrose Sugar Cane 

Juice 

CDW (g/l) DHA (% of TFA) DHA Production (mg/l) 

2 3 4 5 6 7 

Glucose concentration (%) 


Figure 2 Effect of (A) carbon sources and (B) glucose concentration on growth and DHA production 

by S. limacinum BR2.1.2 cultivated at room temperature in shaker 140 rpm. 

(A) 

(B) 

450 

400 

350 

300 

250 

200 

150 

100 

50 

0 

800 

700 

600 

500 

400 

300 

200 

100 

0 

DHA Production (mg/l)) 

DHA Production (mg/l))

supplement with glutamate (Iida et al., 1996). 

Although, soybean meal and skimmed 

milk were slightly inferior compared to peptone 

and probably not suitable as sole nitrogen source, 

they are agricultural products that are less 

expensive and readily available in Thailand. 

Moreover, soybean meal not only provided protein 

but also carbohydrate, fat, mineral and vitamins 

which was likely to support growth and DHA 

production of thraustochytrids (Fan et al., 2002). 

Therefore, they could partially replace peptone 



60 

50 

40 

30 

20 

10 

0 

50 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

Ammonium Potassium 

sulfate nitrate 

P0.5 

P0.5+SBM0.25 


MSG Sodium Peptone Soybean 

nitrate 

meal 

Nitrogen source 


P0.5+SBM0.5 

P1.0+SBM0.25 

P 1.0+SBM0.5 

Nitrogen source 

P1.5+SBM0.25 


which was expensive and economically unsuitable 

for large scale production. Figure 3B shows the 

effect of various peptone and soybean meal 

mixtures on growth and DHA production by S. 

limacinum BR2.1.2. In this experiment, the base 

medium consisted of 5% glucose, 1% skimmed 

milk and 0.2% MSG and 0.5% yeast extract. 

Clearly, treatment with 0.5% peptone and 0.25% 

soybean meal produced highest DHA contents at 

1,170.9 mg/l which was 45.3% of TFA. 

P1.5+SBM0.5 

skimmed 

milk 

Figure 3 Effect of (A) single nitrogen source and (B) combined nitrogen source on growth and DHA 

production by S. limacinum BR2.1.2 cultivated at room temperature in shaker 140 rpm (P = 

peptone; SBM = soybean meal). 

(A) 

(B) 

0 

900 

800 

700 

600 

500 

400 

300 

200 

100 

0 

1400 

1200 

1000 

800 

600 

400 

200 



330 

3. C/N ratio 

Lipid accumulation in oleaginous 

microorganisms can be enhanced by providing 

excess carbon while limiting nitrogen (Ratledge, 

2004). Figure 4 showed that optimum C/N ratio 

at 15:1 was suitable for S. limacinum BR2.1.2 in 

terms of growth and DHA production of 2,416.7 

mg/l. Although, the cell concentration was 

improved (27.6 g/l) the highest biomass of 38.0 

g/l was achieved in medium with C/N ratio of 20:1. 


60 

50 

40 

30 

20 

10 

0 


0:1 10:1 15:1 20:1 25:1 30:1 

C/N ratio 


4. Salinity 

Seawater was used as the source of 

salinity in this study. It should be noted that 

although S. limacinum BR2.1.2 was isolated from 

marine environment, it could grow and produced 

DHA at all levels of salinity (Figure 5). The results 

coincided with Yokochi et al. (1998) who reported 

that S. limacinum SR21 could grow in condition 

at zero salinity or without salt. However, a salinity 

equivalent to 25% of natural sea water appeared 

Figure 4 Effect of C:N ratio on growth and DHA production in S. limacinum BR2.1.2 cultivated at 

room temperature in shaker 140 rpm. 


45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

0 25 50 75 100 150 200 

Salinity (as % of seawater) 


Figure 5 Effect of salinity (as % of seawater) on growth and DHA production in S. limacinum BR2.1.2 

cultivated at room temperature in shaker 140 rpm. 

3000 

2500 

2000 

1500 

1000 

500 

0 

1200 

1000 

800 

600 

400 

200 

0 



optimal for S. limacinum BR2.1.2 for DHA 

production (975.4 mg/l , 41.1% of TFA). At the 

highest salinity (200%), the organism showed 

good growth but DHA production was lowest at 

277.5 mg/l This contrasts to T. aureum which 

fails to grow at zero salinity and also completely 

inhibited at 200% salinity of sea water (Iida et al., 

1996). 

5. Effect of temperature 

In this study cultures were grown in Lshaped 

tubes containing 10 ml medium and 

incubated in a temperature gradient incubator. S. 

limacinum BR2.1.2 grow well and produced fairly 

constant DHA levels at a wide range of 

temperature between 20-30°C. Growth of culture 

varied from 8.7-10.3 g/l, and DHA contents were 

220-236 mg/l (Figure 6). 

Improvement of astaxanthin content by 

mutation 

Although, culture of S. limacinum 

BR2.1.2 in liquid GPY medium was creamy white 

color, it developed orange colonies on agar plate 

after several weeks of incubation. This might be 

explained by an accumulation of carotenoid 

pigments. Preliminary analysis of the pigments 


60 

50 

40 

30 

20 

10 

0 


15 20 25 30 35 

Temperature (C) 


by TLC and HPLC confirmed that it was 

astaxanthin. Hence, it was considered to be 

appropriate to improve the content of this pigment 

in S. limacinum BR2.1.2 by mutation. If successful 

this organism would provide two important 

nutrients, i.e., DHA and astaxanthin making it 

suitable for animal and human consumption 

1. Isolation of S. limacinum BR2.1.2 

mutants 

From an initial S. limacinum BR2.1.2 

concentration of 8.75×10 6 cells/ml, the culture was 

treated with NTG for 20 minutes which yielded a 

0.05% cell survival rate. The treated culture was 

then plated on GYP medium but only one colony 

showed a distinctive red color. After sub-culturing 

for several times the deep red color persisted which 

showed that it was stably expressed. The mutant 

was then used for further investigation. 

The mutant grew rapidly for the first 2 

days with cell concentration of 7.8 g/l. Maximum 

cells mass was obtained on the 6 th days at 10.8 

g/l and declined gradually (Figure 7). Astaxanthin 

contents in cell mass increased corresponding with 

growth and reached highest value at 8.9 µg/ml and 

remained relatively constant towards the end of 

fermentation. This result coincided with Marvelisa 

Figure 6 Effect of temperature on growth and DHA production in S. limacinum BR2.1.2 cultivated in 

L-shaped tubes. 

250 

200 

150 

100 

50 

0 


332 

et al. (2003) who reported that, carotenoid contents 

of Thraustochytrium CHN-1 paralleled the 

biomass and cell growth. The mutant S. limacinum 

BR2.1.2 could produce both DHA and astaxanthin 

at moderate amounts. However, DHA production 

decreased from 224.6 mg/l day 6 to only 29.8 

mg/l at day 10. Hence it seemed that we have to 

sacrifice either DHA or astaxanthin production 

depending on the degree of necessity. 

2. Effect of light intensity on 

astaxanthin accumulation by S. limacinum 

BR2.1.2 mutant 

CDW (g/l); Astaxanthin (ug/ml) 

12 

10 

8 

6 

4 

2 

0 


2 4 6 8 10 

Time (days) 

Under dark condition, the mutant 

accumulated 5.6 µg/ml of astaxanthin at 25°C after 

incubation for 8 days. However, when fluorescent 

light source of 5 kLux was provided, the culture 

produced higher pigment yield of 13.1 µg/ml. 

Further increase of light intensity to 10 kLux had 

adverse effect on astaxanthin production (10.7 

µg/ml) (Figure 8). Therefore, moderate light was 

an important bioinduction for carotenogenesis as 

it was also shown by Phycomyces blaksleeanus 

and several species of Rhodotolula (Goodwin, 

1984). Yamaoka et al. (2004) also demonstrated 

that Thraustochytrium sp. CHN-1 grown under 

Astaxanthin (ug/ml) DHA production (mg/l) CDW (g/l) 

Figure 7 Growth, astaxanthin and DHA production by S. limacinum BR2.1.2 mutant strain in GYC 

broth at 25°C with 2 kLux light intensity and light:dark 16:8 hrs. 

Astaxanthin (ug/ml) 

14 

12 

10 

8 

6 

4 

2 

0 

0 5 10 

Light intensity (kLux) 

Figure 8 Effect of light intensity for astaxanthin production by S. limacinum BR2.1.2 mutant strain in 

GYC broth at 25°C, 180 rpm for 8 days. 

250 

200 

150 

100 

50 

0 

DHA production ((mg/l)

fluorescent lamp at 1.5 kLux developed orange to 

red color. 

CONCLUSIONS 

A thraustochytrid strain BR2.1.2 was 

isolated from mangrove forest in Thailand. The 

strain showed an ability to grow rapidly while 

accumulating large amounts of DHA. 

Identification of the strain based on morphological 

characteristics and 18S rDNA sequence revealed 

that it belonged to Schizochytrium limacinum 

species. Under optimal culture conditions, i.e., 5% 

glucose, combined nitrogen source (0.5% peptone, 

0.2% MSG, 0.25% soybean meal and 1% skimmed 

milk) and C/N ratio at 15:1, the DHA yield was 

2,416.7 mg/l from a cell dry weight of 27.6 g/l. 

Furthermore S. limacinum BR2.1.2 had a unique 

feature of growing in media having a wide range 

of salinity equating to 0-200% seawater. When the 

strain was cultivated in liquid GPY medium the 

culture appeared creamy white color. But on agar 

medium with prolong incubation, color of the 

colony developed into typical orange color of 

carotenoid pigment which was identified as 

astaxanthin. Improvement of S. limacinum 

BR2.1.2 for astaxanthin content by mutation with 

NTG was carried out and resulted in a colony with 

intense red color. This mutant produced 

astaxanthin in liquid medium even without light. 

Optimization of culture conditions in liquid GYC 

medium, particularly high light intensity at 5 kLux 

at 25°C caused the mutant to accumulate the 

pigment at 13.1 µg/ml. 


The authors would like to thanks The 

Graduate School of Kasetsart University, 

Bangkok, Thailand for providing research grant. 

Special thank to Dr. C. Norman Scholfield, Queen 

University, UK, for his English editing was also 

acknowledged. 



Andrews, A.G., H.J. Phaff and M.P. Starr. 1976. 

Carotenoids of Phaffia rhodozyma, a redpigmented 

fermenting yeast. Phytochem. 15: 

1003-1007. 

Bligh, E.G. and W.J. Dyer. 1959. A rapid method 

for total lipid extraction and purification. Can. 

J. Biochem. Physiol. 37: 911-917. 

Barr, D.J.S. 1992. Evolution and kingdom of 

organisms from the perspective of a 

mycologist. Mycologia. 84: 1-11. 

Chaunpit, D., 1993. Mutagenesis in fungi. pp. 

5.24-5.25. In Wattanalai Panbanklet and 

Sroung Udomworabpun (eds.). Biotechnology 

Laboratory Manual I. Thai Society for 

Biotechnology. 

Donkin, P. 1976. Ketocarotenoid biosynthesis by 

Haematococcus pluvialis. Phytochemistry 

15: 711-718. 

Fan, K.W., L.L.P. Vrijmoed. and E.B.G. Jones. 

2002. Physiological studies of subtropical 

mangrove thraustochytrids. Botanical 

Marina 45: 50-57. 

Fang, T.J. and Y.S. Cheng. 1993. Improvement of 

astaxanthin production by Phaffia rhodozyma 

through mutation and optimization of culture 

conditions. J. Ferment. Bioeng. 75: 466-469. 

Fontana, J.D., B. Czeczuga, T.M.B. Bonfim, M.B. 

Chociai, B.H. Oliveira, M.F. Guimaraes and 

M. Baron. 1996. Bioproduction of 

carotenoids: The comparative use of raw 

sugarcane juice and depolymerized bagasse 

by Phaffia rhodozyma. Bioresource 

Technology 58: 121-125. 

Goodwin, T.W. 1984. The Biochemistry of the 

Carotenoids, Vol 1. Plants, 2 nd ed, Chapman 

and Hall, London. 377 p. 

Holub, B.J. and C.M. Skeaff. 1987. Nutritional 

regulation of cellular phosphatidyllinositol. 

Methods in Enzymology 141: 234-244. 

Huang, J., T. Aki., K. Hachida., T. Yokochi., S. 

Kawamoto., S. Shigeta., K. Ono. and O.

334 

Suzuki. 2001. Profile of polyunsaturated fatty 

acids production by Thraustochytrium sp. 

KK17-3. J. Am. Oil. Chem. Soc. 78: 605- 

610. 

Iida, I., T. Nakahara., T. Yokochi., Y. Kamisaka., 

H. Yagi., M. Yamaoka. and O. Suzuki. 1996. 

Improvement of docosahexaenoic acid 

production in culture of Thraustochytrium 

aureum by medium optimization. J. Ferment. 

Bioeng. 81: 76-78. 

Johnson, E. A. and G.H. An. 1991. Astaxanthin 

from microbial sources. Crit. Rev. 

Biotechnol. 11: 297-326. 

Johnson, E.A. and M.J. Lewis. 1979. Astaxanthin 

formation by the yeast Phaffia Rhodozyma. 

J. Gen. Microbiol. 115: 173-183. 

Kinsella, J.E. 1987. Seafoods and Fish Oils in 

Human Health and Disease. New York: 

Marcel Decker. 320 p. 

Lewis, T.E., P.D. Nichols and T.A. McMeekin. 

1999. The Biotechnological Potential of 

Thraustochytrids. Mar. Biotechnol. 1: 580- 

587. 

Mo, C. and B. Rinkevich. 2001. A simple. Reliable 

and fast protocol for Thraustochytrid DNA 

extraction. Mar. Biotechnol. 3: 100-102. 

Marvelisa, L.C., T. Naganuma and Y. Yamaoka. 

2003. Identification by HPLC-MS of 

carotenoids of Thraustochytrium CHN-1 

strain isolated from Seto Inland Sea. Biosci. 

Biotechnol. Biochem. 67: 884-888 

Ratledge, C. 2004. Fatty acid biosynthesis in 

microorganisms being used for Single Cell Oil 

production. Biochimie. 86: 807-815. 

Schroeder, W.A. and E.A. Johnson. 1993. 

Antioxidant role of carotenoids carotenoids 

in Phaffia rhodozyma. J. Gen. Microbiol. 

139: 907-912. 


Simopoulos, A.P. 1989. Summary of NATO 

Advanced Research Workshop on dietary ω3 

and ω6 fatty acids: biological effects and 

nutritional essentiality. J. Nutr. 119: 521-528. 

Takahata, K., K. Monobe, M. Tada and P.C. Weber. 

1998. The benefits and risks of n-3polyunsaturated 

fatty acids. Biosci. 

Biotechnol. Biochem. 62: 2079-2085. 

Valadon, L.R.G. 1976. Carotenoids as additional 

taxonomic characters characters in fungi: A 

review. Trans. Br. Mycol. Soc. 67: 1-15. 

Vazques, M., V. Santos and J.C. Parajo. 1997. 

Effect of the carbon source on the carotenoid 

Profiles of Phaffia rhodozyma strains. J. Ind. 

Microbiol. and Biotech. 19: 263-268. 

White, T.J., T. Bruns., S. Lee., and J. Taylor. 1990. 

Amplification and direct sequencing of fungal 

ribosomal RNA genes for phylogenetics, pp. 

315-322. In M.A. Innis, D.H. Gelfand, J.J. 

Sninsky and T.J. White (eds.). PCR Protocol 

: A Guide to Methods and Applications. 

Academic Press, San Diego. 

Wu, S.T., S.T. Yu, and L.P. Lin. 2005. Effect of 

culture conditions on docosahexaenoic acid 

production by Scizochytrium sp. S31. Proc. 

Biochem. 40: 3103-3108. 

Yamaoka, Y., M.L. Carmona and S. Oota. 2004. 

Growth and carotenoid production of 

Thraustochytrium sp. CHA-1 cultured under 

superbright red and blue light-emitting diodes. 

Biosci. Biotechnol Biochem. 68: 1594-1597. 

Yokochi, T., D. Honda., T. Higashihara. and T. 

Nakahara. 1998. Optimization of 

docosahexaenoic acid production by 

Schizochytrium limacinum SR 21. Appl. 

Microbiol. Biotechnol. 49: 72-76. 

Yongmanitchai, W. and O.P. Ward. 1989. Omega- 

3 fatty acids: Alternative sources of 

production. Proc. Biochem. 24: 117-125.


Cloning, Expression, Purification and Biological Activities 

of Recombinant Mouse Interleukin-2 in E. coli M15 

Sanchai Chantajorn 1 , Ratchanee Hongprayoon 2 * and Thaweesak Songserm 3 

ABSTRACT 

Molecular cloning, sequencing and expression of recombinant mouse interleukin 2 (rmIL-2) 

were described. The interleukin-2 (IL-2) cDNA, 450 base pairs in length with repeating CAG, was 

amplified using specific primers. The IL-2 cDNA showed high homology at 100%, 100%, 91%, 96% 

and 94% with five strains of mice previously reported (GeneBank accession number AY147902.1, 

MMU41494, MMU41504, MMU41505 and MMU41506). The IL-2 gene was cloned into the pDrive 

cloning vector and consequently expressed using pQE30 expression vector which provided high 

expression level of the recombinant protein. The predicted rmIL-2 sequence is 161 amino acids with a 

molecular weight of 19 kDa. The expressed protein was then purified by Ni-NTA column under denaturing 

condition. Analysis of the rmIL-2 by SDS-PAGE demonstrated two bands of 19 and 38 kDa representing 

monomeric and dimeric forms of this protein. The biological activity in stimulating T-cell proliferation 

was also described and the binding signal to the receptor was easily observed by immunofluorescence. 

Key words: mouse interleukin-2, cloning, protein expression, protein purification, immunofluorescence 

INTRODUCTION 

Interleukin-2 (IL-2) is a growthpromoting 

activator for bone marrow-derived T 

lymphocytes (Smith, 1989), and was among the 

first cytokines to be characterized at the molecular 

level. Interleukin-2 was the major autocrine 

growth factor for T lymphocytes, and the quantity 

of IL-2 synthesized by activated CD4 + T cells was 

an important determinant of the magnitude of 

immune response. The action of IL-2 on T-cells 

was mediated by binding to IL-2 receptor proteins. 

This system was perhaps the best understood 

mechanism of all cytokine receptors (Morgan et 

al., 1976). Interleukin-2 exerts its effects on many 

cell types, the most prominent of which is the T 

lymphocyte. Indeed, one of the most rapid 

consequences of T cells activation through its 

antigen receptor is the de novo synthesis of IL-2. 

This was quickly followed by expression of a high 

affinity IL-2 receptor on surface membrane of 

CD4 + T cells, CD8 + T cells, B cells and natural 

killer (NK) cells (Lenardo et al., 1999). 

Interleukin-2 induce cells proliferation via pro- 

1 Centre for Agricultural Biotechnology, Kasetsart University, Kamphaeng Saen Campus, Nakhon Pathom 73140, Thailand. 

2 Department of Plant Pathology, Faculty of Agriculture at Kamphaeng Saen, Kasetsart University, Kamphaeng Saen Campus, 

Nakhon Pathom, 73140, Thailand. 

3 Department of Veterinary Pathology, Faculty of Veterinary, Kasetsart University, Kamphaeng Saen Campus, Nakhon Pathom 

73140, Thailand. 

* Corresponding author, e-mail: agrrat@ku.ac.th 


336 

proliferative signals through the proto-oncogenes 

c-myc and c-fos, in combination with antiapoptotic 

signals through an essentially identical 

receptor (Ma, 2000; Carson et al., 1997; Giri et 

al., 1994; Giri et al., 1995). Interleukin-2 promotes 

production of NK-derived cytokines such as TNF·, 

and granulocyte macrophage colony stimulating 

factor (GMCSF). Furthermore, IL-2 acts 

synergistically to enhance NK cytotoxic activity 

(Khatri, 1998). A number of functions for IL-2 in 

B cells have been identified, mostly pertaining to 

antibody secretion. In IgM expressing B cells, IL- 

2 (in synergy with IL-5) upregulates expression 

of heavy and light chain genes as well as inducing 

de novo synthesis of the immunoglobulin J chain 

gene (Blackman et al., 1986). The latter is required 

for oligomerization of the IgM pentamer, and 

represents a tightly controlled stage in B cells 

activation (Koshland, 1985). As in T cell, IL-2 

increases expression of IL-2Rα in B cells, thus 

enhancing their responsiveness to IL-2 (Gaffen et 

al., 1996). Therefore, IL-2 is one of the key 

cytokines in immunology-base studies. 

Measurement of IL-2 by ELISA method has been 

widely used in clinical investigations and research. 

There are quite a number of commercially 

available IL-2 ELISA kits which are very 

expensive. Production of the ELISA system is 

necessary for our use to investigate the effect of 

plant extract on mouse immune response. The 

objectives of this study were, therefore, to produce 

mouse recombinant IL-2 (rmIL-2) by cloning its 

gene into pDrive cloning vector and express the 

rmIL-2 in pQE30 expression vector to investigate 

its biological activities in vitro for the use in 

immunization. 


Media and reagents 

Complete culture medium was RPMI 

1640 (Hyclone, Utah, USA) supplemented with 2 

mM L-glutamine, 0.05 mM 2-mercaptoethanol 


and 10% fetal calf serum (Hyclone, Utah, USA), 

100 U/ml penicillin and 100 µg/ml streptomycin 

(Sigma-Aldrich, St. Louis, USA). Concanavalin 

A (Con A) (Sigma-Aldrich, St. Louis, USA) was 

used for stimulation of splenocytes. 

Preparation of mouse splenocytes 

A BALB/c male mouse weighing 25-30 

g, 12 weeks of age was purchased from the 

National Laboratory Animal Centre, Mahidol 

University, Nakhon Pathom, Thailand and used 

in the experiments. Mouse spleen was removed 

aseptically, homogenated and cells were washed 

with cold RPMI 1640 medium (HyClone, Utah, 

USA) and resuspended in complete medium. 

Viability and number of splenocytes were 

determined microscopically by staining with 

trypan blue (Gibco, New York, USA). Splenocytes 

(5 × 10 6 cell/ml) were activated with Con A 

(2 µg/ml) in complete medium and transferred into 

24 well, flat-blottom tissue culture plate (Costar, 

New York, USA). Cells were incubated for 36 h 

at 37°C in 5% CO 2 and were harvested for mRNA 

extraction. 

Primers design 

The PCR primers were designed 

specifically to mouse IL-2 gene by the FastPCR 

program. Briefly, primers for the amplification of 

IL-2 mRNA were selected from the conserved 

nucleotide sequences of the five strains of mice 

(GeneBank accession numbers AY147902.1, 

MMU41494, MMU41504, MMU41505, 

MMU41506). Additionally, primer sequences for 

the detection of mouse β-actin gene were taken 

from the literature (Deng et al., 2000) (Table 1). 

RNA extraction and cDNA library synthesis 

Splenocytes were harvested after the 

incubation period (36 h). Total RNA was extracted 

from mitogen-stimulated cells (1 × 10 7 cells) and 

non-stimulated controls according to the methods 

of RNA purification kit (Epicentre, Madison,

Wisconsin, USA). Total extracted RNA was 

applied to 1.2% agarose gels and electrophoresed 

in TBE buffer following standard procedures 

(Sambrook et al., 1989). A cDNA library was 

generated from total RNA using oligo (dT) 15 (DNA 

Technology Laboratory, Kasetsart University, 

Nakhon Pathom, Thailand) as primers. 

SuperScript III reverse transcriptase (Invitrogen, 

California, USA) was used for cDNA synthesis 

following the standard procedures. 

Cloning of the interleukin-2 

The synthesized cDNA library was used 

as a template for the amplification of IL-2 gene 

by two steps RT-PCR. Briefly, PCR master mixture 

consisted of 1X PCR buffer, 1.6 mM MgCl 2, 0.5 

mM dNTP, 0.25 µmol of specific primers (Table 

1) and 1U Platinum Taq DNA polymerase 

(Invitrogen, California, USA). The PCR samples 

were then denatured at 94°C for 2 min and 

continually cycled for 30 times at 94°C for 45 s, 

60°C for 45 s and 72°C for 1 min. For complete 

amplification, an additional extension step at 

72°C for 7 min was included. The PCR products 

were analysed in 1.2% agarose gel electrophoresis 

and visualized by ethidium bromide staining. The 

PCR products were cloned into pDrive cloning 

vector (Qiagen, Valencia, USA) and transformed 


Table 1 Primer sequences used for the amplification of IL-2 and house keeping gene transcripts in 

lymphocytes. 

Gene primer sequences (5′-3′) Direction Nucleotide Length Reference 

position (bp) 

IL-2 GCACCCACTTCAAGCTCCACTTC S 61-83 450 - 

TTATTGAGGGCTTGTTGAGATGATGC AS 485-510 

IL-2 GGATCCGCACCCACTTCAAGC* S 61-75 462 - 

GTCGACTTATTGAGGGCTTGTTGAG** AS 478-510 

β-actin TGTATTCCCCTCCATCGTG S 87-105 491 (Deng et al., 2000) 

GGATCTTCATGAGGTAGTCTGTC AS 555-577 

bp: base pair, IL-2: interleukin-2; β-actin: mouse beta-actin; S: sense strand; AS: antisene strand 

* sense strand primer with restriction site GGATCC (BamHI) 

** antisense strand primer with restriction site GTCGAC (SalI) 

into Escherichia coli (DH5α). The transformants 

were easily observed by blue/white screening and 

the PCR was applied for conformation. 

DNA sequence analysis 

The DNA from positive clones were 

sequenced by the automated DNA sequencer ABI 

377 (GMI, Minnesota, USA). Comparison and 

multiple alignment of BALB/c nucleotide and 

amino acid sequences with those of other mice 

strains were carried out using ClustalW version 

1.83 with additional manual adjustments. 

Expression of rmIL-2 and purification 

IL-2 forward and reverse primers 

including the restriction site were used for the 

amplification of IL-2 gene from the positive 

clones. The PCR products were cloned into a 

pDrive cloning vector and then transformed into 

E. coli (DH5α). The IL-2 gene was amplified and 

digested with BamHI/SalI, the IL-2 gene was 

ligated into the same sites of the expression vector 

pQE30 (Qiagen, Valencia, USA) and then 

transformed into E.coli M15 strain by heat shock 

method (Sambrook et al., 1989). Screeninng of 

the transformants was carried out for ampicillin 

and kanamycin resistance. The positive clones 

were induced by culturing at 37°C for 5 h in 2YT

338 

medium containing 100 µg/ml ampicillin, 25 µg/ 

ml kanamycin and 1 mM isopropyl-1-1-thio-β-Dgalactoside 

(IPTG). Cells were harvested and 

extracted by denaturing condition and then the 

recombinant IL-2 was purified with Ni-NTA resin 

affinity column chromatography according to the 

recombinant protein purification procedures 

(Qiagen, Valencia, USA). The rmIL-2 was allowed 

to refold in native conformation by dialysis in PBS 

and the protein concentration was determined by 

Bradford protein assay (Bradford, 1976). The 

protein purity was determined by SDS-PAGE 

(Laemmli, 1970). 

Western blotting 

Twenty micrograms of the rmIL-2 was 

loaded in a mini-gel apparatus and resolved on a 

12% SDS-PAGE gel and transferred to 

nitrocellulose membrane by electroblotter (Bio- 

Rad, California, USA). The blot was blocked with 

5% skim milk, incubated with rat anti-mouse 

interleukin 2 IgG monoclonal antibody (Serotech, 

North Carolina, USA) (1 µg/ml) for 30 min at 

room temperature. After washing, it was incubated 

with goat anti-rat IgG conjugated with alkaline 

phosphatase at 1:10,000 dilution (Sigma-Aldrich, 

St. Louis, USA). Detection was performed using 

the 5-bromo-4-chloro-3-indolyl phosphate/ 

nitroblue tetrazolium (Zymed, South San Fancisco, 

USA) as substrates. 

Cell proliferation assay 

The rmIL-2 was investigated for the 

ability to stimulate cell proliferation which was 

quantified by the colorimetric assay based on the 

2,3-bis (2-Methoxy-4-nitro-5-sulfophenyl)-5- 

[(phenylamino)-carbonyl]-2H-tetrazolium 

hydroxide (XTT) assay as previously described 

(Scudiero et al., 1988). 

Splenocytes were transferred into 96 well 

microtitre plates (Costar, New York, USA) at a 

density of 1 × 10 5 cells/well for 48 h in complete 

medium. The XTT (Sigma-Aldrich, St. Louis, 


USA) solution was prepared freshly at 1 mg/ml in 

prewarmed balance salt solution without phenol 

red. Then, 5 mM phenazine methosulfate (PMS) 

(Sigma-Aldrich, St. Louis, USA) solution was 

prepared in PBS, stored at 4°C until use and 

protected from light. Culture medium was 

removed from each well, after that a 50 µl of XTT 

solution with 0.025 mM phenazine methosulfate 

was added. After 5 h of incubation, the absorbance 

at 450 nm was determined by a Multiskan EX 

(Labsystems, Finland). 

Receptor binding assay 

A New Zealand white rabbit was first 

immunized with a mixture of rmIL-2 (1 mg/ml) 

and Freund’s Complete adjuvant (Sigma-Aldrich, 

St. Louis, USA) at 1:1 ratio following by three 

injections at weekly intervals with the same 

antigen and Freund’s Incomplete adjuvant (Sigma- 

Aldrich, St. Louis, USA). The antiserum with the 

highest titre was used for immunofluorescent 

detection of IL-2 receptor binding. The splenocytes 

were stimulated with Con A mitogen at 5 µg/ml 

final concentration compared with non-stimulated 

control. Cells were incubated for 6 h at 37°C in 

5% CO 2 and harvested for receptor binding assay. 

They were washed twice with PBS and then 

resuspended in 50 µl of PBS, and fixed with 100 

µl of 4% paraformaldehyde for 30 min in the dark 

at 4°C. The experiment was done on a glass slide. 

Proteins on cell surface were stained with 500 µM 

sulforhodamine B (SRB) (Sigma-Aldrich, St. 

Louis, USA), washed with 1% acetic acid and 

PBS. Cells were then incubated with recombinant 

IL-2 for 1 h at 37°C, washed with PBS and reacted 

with rabbit anti-rmIL-2 polyclonal antibody 

(1:500). After washing step, FITC goat anti-rabbit 

IgG (H+L) conjugate (Zymed, South San 

Fancisco, USA) (1:50) was added and incubated 

for 1 h at 37°C and then washed with PBS. Cells 

were analyzed by reflected light fluorescence 

illuminator BH2-RFL (Olympus, New York, USA) 

within 5 h.

RESULTS 

Cloning and sequencing of mouse IL-2 

The cDNA library of a BALB/c mouse 

was synthesized from total RNA and amplified by 

using the specific primers for IL-2. The cDNA 

synthesis was compared by using β-actin primers 

with two steps RT-PCR. The PCR product of IL-2 

gene showed a band of 450 bp and 491 bp for the 

β-actin gene which was a positive control (Figure 

1). The PCR product was cloned into the pDrive 

cloning vector and its sequence was analysed 

(Figure 2). The IL-2 cDNA sequence consisted of 

eight codons of CAG which differs from the 

previous reports of other mouse strains including 

C3HeB/FeJ (AY147902.1), RF (MMU41494), 

C57BL6/J (MMU41504), CZECHII/Ei 


(MMU41505), and BKL (MMU41506). The other 

mouse strains showed CAG codon with 8, 8, 12, 

21 and 21 contiguous codons, respectively. The 

IL-2 cDNA shows 99%, 99%, 96%, 97% and 96% 

high homology with C3HeB/FeJ (AY147902.1), 

RF (MMU41494), C57BL6/J (MMU41504), 

CZECHII/Ei (MMU41505), and BKL 

(MMU41506). The deduced mouse IL-2 protein 

included 149 amino acid with a predicted 

molecular weight of 17,101 Da. Amino acid 

alignment of mouse IL-2 to those of the other 

strains showed high homology at 100%, 100%, 

91%, 96% and 94% with C3HeB/FeJ 

(AY147902.1), RF (MMU41494), C57BL6/J 

(MMU41504), CZECHII/Ei (MMU41505), and 

BKL (MMU41506), respectively. 

Figure 1 Agarose gel electrophoresis of the amplified IL-2 cDNA products in non-stimulated, mitogenstimulated 

splenocytes and β-actin cDNA for house keeping gene after an incubation period 

of 36 h. A: Positive control (β-actin gene 491 bp); B: Non-stimulated Control and C: Con A 

stimulated cells (IL-2 450 bp). The sizes of PCR products were compared with φX174 DNA 

- Hae III markers (M).

340 

(B) 

(A) 


Figure 2 The 450 base pairs of IL-2 cDNA sequence of a BALB/c mouse (A). Alignment of the predicted 

protein sequences of BALB/c mouse (DQ836354), C3HeB/FeJ mouse (AY147902.1), RF 

mouse (MMU41494), C57BL6/J mouse (MMU41504), CZECHII/Ei mouse (MMU41505), 

and BKL mouse (MMU41506) which were analyzed by GeneDoc (B). The homology of 

amino acid was marked by black stripes.

Expression of rmIL-2 and purification 

The recombinant plasmid containing 

pQE30/mouse IL-2 cDNA was induced with IPTG 

to produce rmIL-2. The bacterial extracts 

containing the rmIL-2 gene could be obtained 

according to the protein pattern from SDS-PAGE. 

The expressed protein tended to aggregate with 

the cell debris as observed in the insoluble fraction 

by SDS-PAGE. However, under denaturing 

condition during the protein extraction, the rmIL- 

2 could be resolved and then purified by Ni-NTA 

column chromatography. The rmIL-2 obtained 

from Ni-NTA resin column was highly purified 

(Figure 3). The bands of fusion proteins on SDS- 

PAGE showed a molecular weigh of 19,000 Da 

and 38,000 Da for monomeric and dimeric forms, 

respectively. In addition, these two bands were 

positively reacted with the IL-2 specific 

monoclonal antibody by Western blotting (Figure 

4). 

Cell proliferation assay 

The purified rmIL-2 was analyzed for 


their biological activities by measuring the XTT 

colorimetric assay. The splenocytes were 

stimulated with serial dilutions of the purified 

rmIL-2 compared with non-stimulated cell culture 

as a negative control. The result showed the 

increasing values of OD450 nm and proliferative 

response after adding rmIL-2 at the concentration 

range of 5-2,560 ng/ml. When 40 ng/ml of the 

rmIL-2 was added, the increasing rate was 

distinctively seen and reached the plateau at the 

concentrations of 640-2,560 ng/ml (Figure 5). 

Receptor binding assay 

Mouse splenocytes were used in this 

study to determine the ligand-receptor binding 

activity. Cell surface membrane was stained red 

with sulforhodamine B (SRB) while the rmIL-2 

bound to the cell receptor showed fluorescence 

green of the FITC conjugate. No or less signal was 

observed in non-stimulated cells. However, the 

splenocytes stimulated by mitogen exhibited 

increasing signal with high expression of IL-2 

receptors on the cell surface (Figure 6). 

Figure 3 Ni-NTA purification of the recombinant mouse IL-2, analyzed by SDS-PAGE, demonstrated 

the two bands 19 kDa monomeric and 38 kDa dimeric forms in the eluate fractions (E1-E3) 

compared with non-induced transformant (C/-), IPTG-induced transformant (C/+), washing 

fractions (W1-W3) and molecular weight markers (M).

342 

DISCUSSION 

The ability to produce and purify large 

quantities of biologically active interleukin-2 has 

been made possible by the use of recombinant 

DNA technology. The mouse cDNA library of 

BALB/c strain was cloned and characterized for 


its activity. The IL-2 cDNA consisted of 450 base 

pairs, repeating CAG and showed high homology 

at 100%, 100%, 91%, 96% and 94% with five 

strains of mice previously reported (GeneBank 

accession number AY147902.1, MMU41494, 

MMU41504, MMU41505 and MMU41506). The 

rmIL-2 has been intensively studied and found that 

Figure 4 Detection of the recombinant mouse IL-2 (rmIL-2) by SDS-PAGE (A) and Western blotting 

(B) by probing with mouse IL-2 specific monoclonal antibody. M; molecular weight markers. 

Figure 5 The effect of various concentrations of the recombinant IL-2 on cell proliferation of the 

lymphocytes determined by XTT colorimetric assay while non-incubated cells with the 

recombinant IL-2 showed the OD450 nm = 0.248 (data not shown).

other strains of mice have different effects on the 

biological activity of IL-2 (Matesanz and Alcina, 

1996). The IL-2 cDNA did not contain the 

hydrophobic leader sequence of a 20 amino acid 

peptide and the expressed rmIL-2 was purified 

from the cells later. According to Robb et al. 

(1981), even though the IL-2 exhibited O-linked 

glycosylation at threonine 3 of N-terminus and the 

E. coli system did not provide the posttranslational 

glycosylation, it did not affect the IL-2 activity 

nor change its activity in standard bioassay. The 

functional significance of glycosylation of IL-2 

was not known but it was likely that it enhances 

solubility in aqueous environments. Thus an Nterminal 

20 amino acid sequence was reported to 

be essential for the interaction with the IL-2 

receptor (Eckenberg et al., 2000). The 

polymorphism of the CAG sequence has been 

reported among different strains of mice including 

C3HeB/FeJ mouse (AY147902.1), RF mouse 

(MMU41494), C57BL6/J mouse (MMU41504), 

CZECHII/Ei mouse (MMU41505), and BKL 


mouse (MMU41506) (GeneBank data base) which 

contained the sequences of 8, 8, 12, 21 and 21 

codons, respectively. Characterization of the rmIL- 

2 by ProtParam program (ExPASy) showed that it 

consisted of 149 amino acids of mature IL-2 

protein and 12 amino acids of protein tag from 

the expression vector. The expressed protein was 

estimated to weigh 18,489 Da, with the isoelectric 

point (pI) at 5.87 with good solubility. However, 

the protein bands observed on the SDS-PAGE 

were found to be 19 and 38 kDa which were 

predicted as a monomeric and dimeric forms of 

the protein. The increased molecular weight from 

the data obtained by program analysis may due to 

the phosphorylation of the rmIL-2 (Adachi et al., 

1997; Brennan et al., 1997; Gesbert et al., 1998; 

Justement, 2001; Cook and Unger, 2002; Michelle 

et al., 2003; Stoker, 2005). The rmIL-2 were 

applied to the cell culture with the following 

addition of XTT to examine its biological activity. 

The result showed that the activity was raised 

according to the increasing concentration of rmIL- 

Figure 6 The binding of the recombinant mouse IL-2 to the IL-2 receptors was analyzed by 

immunofluorescence. Cells stained with sulforhodamine B (SRB) illustrating red color and 

the signal for IL-2 binding showed greenish fluorescence (arrows). (A) Non-stimulated cells 

and (B) mitogen-stimulated cells, after 6 h of incubation.

344 

2 fusion proteins. In the receptor binding assay, 

the rmIL-2 bound to its receptor showing green 

fluorescence on the cell surface. This experiment 

confirmed the rmIL-2 biological activity in binding 

to its receptor and leading to cell proliferation by 

XTT assay. The future plan of our project will be 

the use rmIL-2 as an antigen to raise anti-IL 2 

polyclonal antibody and develop the ELISA 

method for the measurement of mouse IL-2 for 

further investigation. 


The authors would like to acknowledge 

the Centre for Agricultural Biotechnology, 

Kasetsart University. Kamphaeng Saen Campus, 

Nakhon Pathom for the facility and financial 

support throughout this study. 


Adachi, M., M. Ishino, T. Torigoe, Y. Minami, T. 

Matozaki, T. Miyazaki, T. Taniguchi, Y. 

Hinoda and K. Imai. 1997. Interleukin-2 

induces tyrosine phosphorylation of SHP-2 

through IL-2 receptor β chain. Oncogene 14: 

1629–1633. 

Blackman, M.A., M.A. Tigges, M.E. Minie and 

M.E. Koshland. 1986. A model system for 

peptide hormone action in differentiation: 

interleukin-2 induces a B lymphoma to 

transcribe the J. chain. gene. Cell 47: 609-617. 

Bradford, M. 1976. A rapid and sensitivity method 

of measuring microgram quantities of proteins 

utilizing the principle of protein-dye coupling. 

Anal. Biochem. 72: 248-264. 

Brennan, P., J.W. Babbage, B.M. Burgering, B. 

Groner, K. Reif and D.A. Cantrell. 1997. 

Phosphatidylinositol 3-kinase couples the 

interleukin-2 receptor to the cell cycle 

regulator E2F. Immunity 7: 679–689. 

Carson, W.E., T.A. Fehniger, S. Haldar, K. Eckhert, 

M.J. Lindemann and C.F. Lai. 1997. A 


potential role for interleukin-15 in the 

regulation of human natural killer cell 

survival. J. Clin. Invest. 99: 937-943. 

Cook, W.S. and R.H. Unger. 2002. Protein tyrosine 

phosphatase 1B: a potential leptin resistance 

factor of obesity. Dev. Cell. 2(4): 385-387. 

Deng, Y.M., E.T. Bernard and D.R. Willium. 2000. 

Transmission of porcine endogenous 

retroviruses in severe combined 

immunodefficient mice xenotransplanted with 

fetal porcine pancreatic cells. 

Transplantation 70: 1010-1016. 

Eckenberg, R., J.L. Moreau, O. Melnyk and J. 

Theze. 2000. IL-2R beta agonist P1-30 acts 

in synergy with IL-2, IL-4, IL-9, and IL-15: 

biological and molecular effects. J. Immunol. 

165: 4312-4318. 

Gaffen, S.L., S.L. Wang and M.E. Koshland. 1996. 

Expression of the immunoglobulin J chain in 

a murine B lymphoma is driven by autocrine 

production of interleukin-2. Cytokine 8: 513- 

524. 

Gesbert, F., M. Delespine-Carmagnat and J. 

Bertoglio. 1998. Recent advances in the 

understanding of interleukin-2 signal 

transduction. J. Clin. Immunol. 18: 307-320. 

Giri, J.G., M. Ahdieh, J. Eisenmann, K. 

Shanebeck, K. Grabstein and S. Kumaki. 

1994. Utilization of the b and g chains of the 

IL-2 receptor by the novel cytokine IL-15. 

EMBO. J. 13: 2822-2830. 

Giri, J.G., S. Kumaki, M. Ahdieh, D.J. Friend, A. 

Loomis and K. Shanebeck. 1995. 

Identification and cloning of a novel IL-15 

binding protein that is structurally related to 

the a chain of the IL-2 receptor. EMBO. J. 

14: 3654-3663. 

Justement, L.B. 2001. The role of the protein 

tyrosine phosphatase CD45 in regulation of 

B lymphocyte activation. Int. Rev. Immunol. 

20(6): 713-738. 

Khatri, V.P., T.A. Fehniger, R.A. Baiocchi, F. Yu, 

M.H. Shah and D.S. Schiller. 1998. Ultra low

dose interleukin-2 therapy promotes a type 1 

cytokine profile in vivo in patients with AIDS 

and AIDS-associated malignancies. J. Clin. 

Invest. 101: 1373-1378. 

Koshland, M.E. 1985. The coming of age of the 

immunoglobulin J chain. Annu. Rev. 

Immunol. 3: 425-453. 

Laemmli, U.K. 1970. Cleavage of structure 

proteins during the assembly of the head 

bacteriophage T4. Nature 277: 680-685. 

Lenardo, M., F.K.-M. Chan, F. Hornung, H. 

McFarland, R. Siegel and J. Wang. 1999. 

Mature T lymphocyte apoptosisdimmune 

regulation in a dynamic and unpredictable 

environment. Annu. Rev. Immunol. 17: 221- 

253. 

Ma, A. 2000. Pleiotropic functions of IL-15 in 

innate and adaptive immunity. Mod. Aspects 

Immunobiol. 1: 102-104. 

Matesanz, F. and A. Alcina. 1996. Glutamine and 

tetrapeptide repeat variations affect the 

biological activity of different mouse 

interleukin-2 alleles. Eur. J. Immunol. 26(8): 

1675-1682. 

Michelle, M.F., R. Forest and L.M. Kathleen. 

2003. IL-2 activation of a PI3K-dependent 


STAT3 serine phosphorylation pathway in 

primary human T cells. Cell. Signal. 15: 625- 

636. 

Morgan, D.A., F.W. Ruscetti and R. Gallo. 1976. 

Selective in vitro growth of T lymphocytes 

from normal human bone marrows. Science. 

193: 1007-1008. 

Robb, R.J., A. Munck and K.A. Smith. 1981. T 

cell growth factor receptors. Quantitation, 

specificity, and biological relevance. J. Exp. 

Med. 154: 1455-1474. 

Sambrook, J., E.F. Fritsch and T. Maniatis. 1989. 

Molecular Cloning: a Laboratory Manual, 

Cold Spring Harbor Laboratory, Cold Spring 

Harbor, NY. 

Scudiero, D.A., R.H. Shoemaker, K.D. Paull, A. 

Monks, S. Tierney, T.H. Nofziger, M.J. 

Currens, D. Seniff and M.R. Boyd. 1988. 

Evaluation of a soluble tetrazolium/ formazan 

assay for cell growth and drug sensitivity in 

culture using human and other tumor cell lines. 

Cancer Res. 48: 4827-4833. 

Smith, K.A. 1989. The interleukin-2 receptor. 

Annu. Rev. Cell. Biol. 5: 397-425. 

Stoker, A.W. 2005. Protein tyrosine phosphatases 

and signalling. J. Endocrinol. 185(1): 19-33.


Production and Partial Characterization of Chitosanases from a 

Newly Isolated Bacillus cereus 

Sutee Wangtueai 1 , Wanchai Worawattanamateekul 1 *, Mathana Sangjindavong 1 , 

Nuanphan Naranong 2 and Sarote Sirisansaneeyakul 3 

ABSTRACT 

The production of chitosanases by a newly isolated Bacillus cereus TP12.24 was studied both 

in shake flask and fermenter cultures. The M9-chitosan medium was found most suitable with 0.5% 

chitosan as a sole carbon source optimized under aerobic growth conditions at pH 6.0 and 30°C. The 

specific rates of growth, substrate consumption, and enzyme production were improved using controlled 

completely aerobic conditions in 2-l fermenter. While the yield of biomass was considerably increased, 

the enzyme yield was on the contrary decreased. As a result, the volumetric chitosanases productivity 

was 43.55 U/l h, which was 1.2 times that obtained from shake flask culture due to higher specific rates 

of chitosan consumption and chitosanases production. In this work, the crude chitosanases from Bacillus 

cereus TP12.24 showed their optimal pH and temperature at 6.5 and 55°C, while the stabilities to pH 

and temperature were found at 3.0-8.0 and 30-50°C, respectively. The Bacillus cereus chitosanases 

could be used for preparing the chitosano-oligosaccharides under mild temperature. 

Key words: chitosanases, chitosan, Bacillus cereus, optimization, batch culture 

INTRODUCTION 

Chitosan (poly-β-(1→4)-2-amino-2deoxy-D-glucose) 

is a long chain polymer derived 

from chitin by deacetylation (Kumar et al., 2000). 

Mostly, the sources of chitin in Thailand are solid 

wastes derived from the shrimp processing 

industries. Chitosan has been utilized as multipurpose 

products in food, semi-food and non-food 

industries. Whereas the production of chitosanderived 

oligosaccharides shows its potential as 

high value added food product, the enzymatic 

hydrolysis rather than chemical degradation that 

provides an attractive process is obviously limited. 

Chitosanase (EC 3.2.1.132) is exploited for the 

production of chitosano-oligosaccharides. Various 

sources of enzyme could be obtained from soil 

fungi and bacteria, such as Bacillus circulans MH- 

K1 (Yabuki et al., 1988), Bacillus sp. No.7-M 

(Uchida and Ohtakara, 1988), Bacillus 

licheniformis UTK (Uchida et al., 1992), Bacillus 

cereus S1 (Kurakake et al., 2000), Streptomyces 

N-174 (Boucher et al., 1992), Streptomyces sp. 

No.6 (Price and Storck, 1975), Amycolatopsis sp. 

CsO-2 (Okajima et al., 1994), and Burkholderia 

gladioli strain CHB101 (Shimosaka et al., 2000). 

1 Department of Fishery Products, Faculty of Fisheries, Kasetsart University, Bangkok 10900, Thailand. 

2 Department of Applied Biology, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, 

Thailand. 

3 Department of Biotechnology, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: ffiswcw@ku.ac.th 


The chitosano-oligosaccharides are 

water-soluble and possess versatile bioactivities 

such as immunopotentiating, bacteriostatic 

activities which have their advantages in food 

materials, agricultural and medical, and antitumor 

activity (Tominaga and Tsujisaka, 1975; Price and 

Storck, 1975; Suzuki et al., 1984; Papineau et al., 

1991; Somashekar and Joseph, 1996; Jeon and 

Kim, 1998, 2000). The purpose of the present work 

was to optimize the production of chitosanases 

from the newly isolated Bacillus cereus TP12.24 

(Wangtueai et al., 2006). The crude chitosanases 

were also partly characterized for their optimal and 

stability based on pH and temperature. 


Microorganism 

Bacillus cereus TP12.24, a newly 

isolated soil bacterium (Wangtueai et al., 2006) 

was used throughout the experiments. The stock 

culture was maintained on the chitosanasedetection 

agar medium (CDA) (Cheng and Li, 

2000) and freshly transferred every 2 weeks. 

Factors affecting enzyme production in shake 

flask culture 

All experiments were carried out in shake 

flask cultures using 500-ml Erlenmeyer flask 

containing 250-ml M9-chitosan medium at 250rpm 

for 72 h. Samples were taken every 6 h for 

determining total viable cells, dry cell weight, 

residual chitosan and enzyme activity. The culture 

conditions and all analyses have been described 

previously (Wangtueai et al., 2006). 

Effect of pH 

The study on pH optimum for enzyme 

production was carried out by varying the pH 

values of M9-0.5% chitosan medium at 4.0 to 8.0 

at 30°C. 


Effect of chitosan 

The M9-chitosan media containing 0.1, 

0.5, 1.0 and 2.0 % chitosan were used for the 

production of chitosanases under optimized initial 

pH 6.0 at 30°C. 

Effect of temperature 

The temperatures of 30, 40 and 50°C 

were investigated for growth and enzyme 

production under optimized initial pH 6.0 and 

0.5% chitosan concentration. 

Enzyme production in 2-l fermenter 

The 2-l fermenter (EYELA Mini jar 

fermenter, Model M-100, Tokyo Rikakikai Co., 

Ltd.) which contained 1.5-l M9-chitosan medium 

with 0.5% chitosan was used for the production 

of chitosanases from Bacillus cereus TP12.24. The 

fermentation conditions were controlled 

automatically at 30°C, pH 6.0, 1 vvm aeration rate 

and 400 rpm agitation rate for 58.5 h of cultivation 

time. The samples were taken every 6 h for 

determining the total viable cells, dry cell weight, 

residual chitosan and enzyme activity (Wangtueai 

et al., 2006). The fermentation kinetics of bacterial 

growth and chitosanases production were studied 

based on the experimental results. 

Characterization of crude chitosanases 

The crude chitosanases were prepared by 

growing cells in 2-l fermenter under the optimal 

conditions obtained in this work. The enzyme 

supernatant was collected from the culture broth 

after centrifugation at 8,000 rpm, 4°C for 20 min. 

This supernatant as crude chitosanases was used 

for the determination of enzyme optimal and 

stability on the basis of pH and temperature. 

Optimal pH 

The crude chitosanase activity was 

measured at various pH values, using 80% 

deacetylated chitosan as a substrate. The reaction 

mixtures consisting of 1.0 ml of 1% soluble

348 

chitosan and 1.0 ml of the crude enzyme solution 

were incubated for 10 min at 30°C. The extended 

pH ranges of 3.0-7.5 and 8.0-9.0 were monitored 

by 0.05 M citrate phosphate buffer and 0.05 M 

carbonate bicarbonate buffer, respectively. 

Optimal temperature 

The temperatures were varied from 30- 

70°C for optimizing the enzyme activity for 10 

min at the optimal pH 6.5 obtained in this work. 

The reaction mixture prepared was the same as 

mentioned above. 

pH stability 

The prepared crude enzyme was diluted 

5 times with buffers at various pH’s (crude 

enzyme:buffer, 1:4) using 0.05 M citrate phosphate 

buffer for pH 3.0-8.0 and 0.05 M carbonatebicarbonate 

buffer for pH 9.0-11.0. The diluted 

enzyme solutions at these various pH’s were 

incubated at 40°C for 60 min. Then the residual 

activities of chitosanases were determined under 

the specified conditions modified from Shimosaka 

et al. (1995) and Cheng and Li (2000). 

Temperature stability 

The diluted crude enzyme solutions were 

prepared with 0.05 M citrate phosphate buffer pH 

6.5 and incubated at different temperatures varying 

from 30-80°C for 30 min. The residual enzyme 

activities of chitosanases were determined under 

the specified conditions modified from Shimosaka 

et al. (1995) and Cheng and Li (2000). 

Analyses 

Determination of growth 

The total number of viable cells was 

determined by spread plate technique and the dry 

cell weight was calculated from the prepared 

standard curve of dry cell weight and total viable 

cells. 


Determination of chitosan 

The concentration of chitosan in culture 

broths was measured by the procedure described 

by Kobayashi et al. (1988). 

Chitosanase assay 

The 1% soluble chitosan was prepared 

by dissolving one gram of chitosan in 40 ml of 

deionized water and 9 ml of 1.0 M acetic acid. 

The solution was stirred for 2 h and the pH was 

adjusted to 6.0 with 1.0 M sodium acetate. This 

solution was finally made up to 100 ml by adding 

0.05 M acetate buffer pH 6.0. 

Chitosanase activity was analyzed by 

estimating the reducing ends of chitooligosaccharides 

produced from the catalytic 

hydrolysis of chitosan. The assay was performed 

by mixing 1.0 ml of 1 % chitosan solution at pH 

6.0 and 1.0 ml of suitably diluted enzyme. After 

10 min incubation at 30°C, the reaction was 

stopped by boiling the mixture for 3 min. A 1.0 ml 

sample of the reaction mixture was taken for 

determining reducing sugar by the procedure 

described by Miller (1959). One unit chitosanase 

activity was defined as the amount of enzyme 

required to release 1 µmol of detectable reducing 

sugar at 30°C in 1 min. 


Optimizing chitosanases by shake flask culture 

Effect of pH 

Bacillus cereus TP12.24 grown in the 

M9-chitosan medium with varying initial pH 4.0, 

5.0, 6.0, 7.0 and 8.0 at 30°C, gave the highest 

enzyme activities of 336.24 U/l in 24 h, 503.31 U/ 

l in 30 h, 2,040.64 U/l in 54 h, 428.71 U/l in 30 h 

and 567.40 U/l in 48 h, respectively. While the 

maximal dry cell weights were 0.421 g/l at 24 h, 

0.503 g/l at 30 h, 2.125 g/l at 54 h, 1.246 g/l at 30 

h and 1.733 g/l at 30 h, respectively. Mostly, the 

production of chitosanases was associated with the 

bacterial growth, in which the concentrations of

enzyme and cells were maximized by using the 

initial pH of 6.0. The maximal specific growth rate 

obtained was 0.260 h -1 at the initial pH 6.0 (Table 

1). Higher or lower initial pH’s gave less favorable 

specific growth rates. At this optimal initial pH 

6.0, the specific rates of chitosan consumption and 

chitosanases production were 0.091 g/g h and 

31.99 U/g h, respectively. As a result, the yield 

and volumetric productivity of chitosanases were 

247.59 U/g and 35.29 U/l h, respectively (Table 

1). The optimal pH obtained in this work was quite 

similar to the results reported by Yoshihara et al. 

(1990) culturing Pseudomonas sp. at pH 6.3 and 

Tominaka and Tsujisaka (1975) producing Bacillus 

R-4 chitosanases at pH 6.0. Moreover, at higher 

pH 6.5 chitosan was difficult to dissolve and could 

not provide a useful carbon source for the bacterial 

growth. Especially, at initial pH 7.0 and 8.0, 

chitosan appeared in large particle sizes, which 

was barely consumed by the bacterial cells. The 

solubility of commercial chitosan being most 

excellent in diluted organic acids has been also 

reported (Kim et al., 2001). In particular, it is clear 

that the specific rate of chitosan consumption was 

enhanced 2.3-7.0 times higher at pH 4.0 than those 


at elevated pH’s. Nevertheless, the specific growth 

rate maximized at pH 6.0 dictated the production 

yields of both cells and enzymes, so that the better 

substrate consumption could no longer monitor the 

production of enzymes. 

Effect of chitosan 

With 0.1 % chitosan, the highest 

concentrations of cells and enzymes were 0.474 

g/l and 475.70 U/l at 66 and 54 h, respectively. 

The cell and enzyme concentrations were 

increased to 2.125 g/l and 2,040.64 U/l at 54 h, 

respectively, when using 0.5% chitosan as the main 

substrate. No bacterial growth was found at 1.0 

and 2.0% chitosan because high viscosity of the 

culture medium limited oxygen availability for the 

bacterial growth. It was also reported that high 

chitosan concentration can inhibit the bacterial 

growth (No et al., 2001). In this study, the chitosan 

concentration of 1.0 and 2.0% could not be used 

as appropriate substrate concentration for the 

production of chitosanases. Therefore, 0.5% 

chitosan was finally selected for the optimal 

growth and chitosanases production from Bacillus 

cereus TP12.24. The specific growth rate and the 

Table 1 Factors affecting growth and chitosanases production by Bacillus cereus TP12.24 using shake 

flask culture. 

Factors Variables µ Y X/S Y P/S q S q P Q P 

(h -1 ) (g/g) (U/g) (g/g h) (U/g h) (U/l h) 

pH 4.0 0.043 0.112 122.19 0.218 36.13 9.81 

5.0 0.155 0.046 37.74 0.094 45.91 13.06 

6.0 0.260 0.352 247.59 0.091 31.99 35.29 

7.0 0.138 0.395 39.89 0.031 13.46 5.64 

8.0 0.126 0.739 107.27 0.068 25.28 8.91 

Chitosan (%) 0.1 0.111 0.221 222.15 0.253 57.39 7.59 

0.5 0.260 0.352 247.59 0.091 31.99 35.29 

Temperature 30 0.260 0.352 247.59 0.091 31.99 35.29 

(°C) 40 0.175 0.032 168.62 1.388 285.06 23.01 

50 0.208 0.025 167.59 2.428 441.38 22.14 

Note: Specific growth rate (µ) obtained from plotting the graph between log dry cell weight and culture time, the yields (Y X/S, Y P/ 

S) obtained from plotting the graph of dry cell weight or enzyme activity with substrate, and the specific rates (q S, q P) 

calculated at the maximal enzyme production with culture time using average dry cell weight.

350 

volumetric enzyme productivity were 2.3 and 4.6 

times higher, respectively, as compared to 0.1% 

chitosan (Table 1). As discussed above, the more 

chitosan consumption, shown as the higher 

specific rate of chitosan consumption, did not favor 

the production of cells and enzymes even at 

optimal pH 6.0. Here, the limiting substrate at 

0.5% chitosan which maximized the bacterial 

growth played an important role instead, for the 

production of chitosanases. 

Effect of temperature 

The maximal concentrations of dry cell 

weight and chitosanases were 2.125, 0.169 0.154 

g/l and 2,040.64, 1,433.09 and 1,444.13 U/l at 30, 

40 and 50°C in 54 h culture, respectively. The 

bacterial growth was clearly retarded at higher 

temperatures of 40 and 50°C, in which the cell 

concentrations decreased markedly after 18 

and$12 h of culture times, respectively (data not 

shown). Both specific growth rate and the 

volumetric enzyme productivity decreased when 

increasing the growth temperatures beyond 30°C. 

Therefore, growth and enzyme production were 

found optimum at 30°C, as shown in Figure 1. 

pH 

8.0 

7.5 

7.0 

6.5 

6.0 

5.5 

Chitosanase activity (U/l) 

3000 

2500 

2000 

1500 

1000 

500 

0 

Chitosan (g/l) 

6 

4 

2 


Although the specific rates of substrate 

consumption were much higher at elevated 

temperatures (Table 1), these higher temperatures 

inhibited the bacterial growth and resulted in lower 

specific growth rate and the yields of cell and 

enzyme production. In conclusion, the factors that 

maximized the bacterial growth affected the 

production of both cells and enzymes. This 

revealed that chitosanases from the newly isolated 

Bacillus cereus TP12.24 was the growth associated 

enzymes. 

The production of chitosanases in 2-l fermenter 

Optimal conditions obtained from the 

shake flask culture were applied for kinetic study 

of the production of chitosanases in a laboratory 

fermenter, using the M9-chitosan medium 

containing 0.5% chitosan. The conditions were 

controlled at 30°C and pH 6.0 under completely 

aerobic cultivation (1 vvm aeration and 400 rpm 

agitation). Bacillus cereus TP12.24 produced 

highest dry cell weight at 0.904 g/l in 21 h, enzyme 

activity at 1,562.12 U/l in 28.5 h (Figure 2). 

However, the enzyme was harvested at 58.5 h at 

the end of cultivation for studying the properties 

pH 

Dry cell weight 

Total viable cells 

Chitosan 

Enzyme activity 

0 

0 

0 6 12 18 24 30 36 42 48 54 60 66 72 

Time (h) 

Figure 1 The production of chitosanases by Bacillus cereus TP12.24 in shake flask culture controlled 

at 30°C. 

6 

5 

4 

3 

2 

1 

Dry cell weight (g/l) 

8 

6 

4 

2 

0 

(CFU/ml) 

Total viable cells x 10 7

of crude chitosanases. Fortunately, the enzyme 

activity was found stable after its maximal at 

28.5 h. 

The kinetic parameters for growth and 

enzyme production were summarized in Table 2. 

The bacterial growth was promoted noticeably in 

fermenter cultivation, resulting in rapid production 

of chitosanases. As the specific growth rate 

increased, the high yield of cells provided higher 

cell concentration with higher specific rates of 

chitosan consumption and chitosanases 

production. As a result, the volumetric productivity 

of chitosanases was 1.2 times increased under 

aerobic conditions in fermenter. This indicated that 

Total viable cells x 10 7 (CFU/ml) 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

0.0 

Dry cell weight (g/l) 

1.00 

0.80 

0.60 

0.40 

0.20 


0.00 

0 10 20 30 40 50 

oxygen plays a very important role in promoting 

the bacterial growth and the production of 

chitosanases. More or less, any suitable parameters 

for monitoring the supply of oxygen during 

cultivation, such as DO or K La might be a key 

strategic optimization for scaling up the production 

of chitosanases in a large scale fermenter. 

Moreover, when compared to the shake 

flask culture, the lag period of bacterial growth in 

fermenter culture was reduced to 6 h from 18 h 

(Figure 1 and 2). Substrate was also rapidly 

consumed under aerobic condition in fermenter. 

Chitosanases were produced in 18-36 and 6-20 h 

in shake flask and fermenter cultures, respectively. 

Time (h) 

Dry cell weight 

Total viable cell 

Chitosan 

Chitosanase activity 

Figure 2 The production of chitosanases by Bacillus cereus TP12.24 in fermenter culture controlled at 

1 vvm aeration, 400 rpm agitation, pH 6.0 and 30°C. 

Table 2 Fermentation kinetics of Bacillus cereus TP12.24 from shake flask and fermenter cultures. 

Culture µ YX/S YP/S qS qP QP conditions (h-1 ) (g/g) (U/g) (g/g h) (U/g h) (U/l h) 

Flask 0.260 0.352 247.59 0.091 31.99 35.29 

Fermenter 

Note: 

0.304 0.447 181.01 0.682 154.37 43.55 

(1) Flask culture referred to optimized conditions at initial pH 6.0, 0.5% chitosan and 30°C. 

(2) The optimal conditions for fermenter culture were pH 6.0, 30°C, 400 rpm and 1 vvm. 

(3) Calculations were done at maximal chitosanase activity obtained. 

6 

5 

4 

3 

2 

1 

0 

Chitosan (g/l) 

1800 

1600 

1400 

1200 

1000 

800 

600 

400 

200 

Chitosanase activity (U/l)

352 

The enzymes were also increased at stationary 

growth phase to show the non-growth associated 

enzyme production. However, enzyme was quite 

stable in fermenter culture. Oxygen might confirm 

its important role during declining growth phase 

in promoting enzyme stability. Further 

investigation will be conducted on optimizing an 

effect of oxygen for the production of chitosanases. 

The properties of crude chitosanases 

The crude chitosanases after cell 

removal, prepared from the 2-l fermenter 

mentioned earlier were used without any further 

treatment for studying the pH and temperature 

optimum and stability of enzyme. 

Effect of pH 

The optimal pH of crude chitosanases 

was at pH 6.5 (Figure 3). At lower pH 3.0 and 

higher pH 9.0, the relative enzyme activities were 

47.18 and 56.64%, respectively. This optimal pH 

was comparable to Bacillus cereus S1 chitosanases 

(pH 6.0) (Kurakake et al., 2000) and similar to 

Relative activity (%) 


chitosanases from Bacillus circulans MH-K1 

(Yabuki et al., 1988) and Bacillus sp. No. 7-M 

(Uchida and Ohtakara, 1988). This, however, 

differed totally from those produced by Bacillus 

subtilis IMR-NK1 (Chiang et al., 2003) and 

Bacillus megaterium P1 (Pelletier and Sygusch, 

1990) which were optimized at pH 4.0 and 4.5- 

6.5, respectively. As previously reported, the 

optimal pH’s for various chitosanases were in a 

broad range of 4.0-8.0 (Somasheka and Joseph, 

1996) depending on the bacterial strains. 

Bacillus cereus TP12.24 chitosanases 

were found stable at a wide pH range of 3.0-8.0 

retaining more than 70% activity after 

preincubation at 40°C for 60 min. However, at pH 

9.0 and 11.0, the relative activities were decreased 

to 47.14 and 34.29%, respectively. Different 

chitosanases showed different pH stability, such 

as pH 6.0-11.0 for Bacillus cereus S1 chitosanases 

(Kurakake et al., 2000) and pH 5.0-9.0 for Bacillus 

subtilis IMR-NK1 chitosanases after preincubation 

at 25°C for 1 h (Chiang et al., 2003). 

100 pH optimum 

pH stability 

80 

60 

40 

20 

3 4 5 6 7 8 9 10 11 

Figure 3 Optimal pH and pH stability of Bacillus cereus TP12.24 chitosanases. 

pH

Relative activity (%) 

100 

80 

60 

40 

Effect of temperatures 

The activity of crude chitosanases from 

Bacillus cereus TP12.24 was found optimal at 

55°C (Figure 4). At lower or higher temperatures, 

the relative activities were reduced to 86.36 and 

85.86% at 30 and 70°C, respectively. This optimal 

temperature was slightly lower than that of 

Bacillus cereus S1 (60°C) (Kurakake et al., 2000), 

but higher than those of Bacillus subtilis IMR-NK1 

(45°C) (Chiang et al., 2003) and Bacillus 

megaterium P1 (50°C) (Pelletier and Sygusch, 

1990). 

Bacillus cereus TP12.24 chitosanases 

were stable at temperature of 30-50°C showing 

74.26-81.19% activity. However the enzyme 

activity was decreased at temperature higher than 

50°C. Chitosanases from Bacillus cereus S1 were 

ever reported to be stable at temperature higher 

than 60°C at pH 5.0 for 30 min (Kurakake et al., 

2000). Therefore, Bacillus cereus TP12.24 

chitosanases were not the thermostable enzyme. 

The enzyme could be used at moderate 

temperature and neutral pH under the wide pH 

range of stability. 


Temperature optimal 

Temperature stability 

20 

30 40 50 60 70 80 

Temperature (°C) 

Figure 4 Optimal temperature and temperature stability of Bacillus cereus TP12.24 chitosanases. 


The present study was financially 

supported by the National Center for Genetic 

Engineering and Biotechnology (Biotec), Thailand 

and partially supported by DNA Technology 

Laboratory, Kasetsart University Kamphaeng Saen 

Campus in association with the Commission on 

Higher Education, Thailand. The laboratory at the 

department of Biotechnology, Kasetsart University 

Bangkhen Campus was greatly acknowledged for 

providing the fermentation facilities. 


Boucher, I., A. Dupuy, P. Vidal, W. A. Neugebuer 

and R. Brzezinski. 1992. Purification and 

characterization of a chitosanase from 

Streptomyces N174. Appl. Microbiol. 

Biotechnol. 38: 188-193. 

Cheng, C. Y. and Y. K. Li. 2000. An Aspergillus 

chitosanase with potential for large-scale 

preparation of chitosan oligosaccharides. 

Biotechnol. Appl. Biochem. 32: 197-203.

354 

Chiang, C. L., C. T. Chang and H. Y. Sung. 2003. 

Purification and properties of chitosanase 

from a mutant of Bacillus subtilis IMR-NK1. 

Enzyme Microb. Technol. 32: 260-267. 

Jeon, Y. J. and S. K. Kim. 1998. Bioactivities of 

chitosan oligosaccharides and their derivative, 

pp. 328-333. In R.H. Chen and H.C. Chen 

(eds.). Advances in Chitin Science, vol. 3, 

P.R. China. 

Jeon, Y. J. and S. K. Kim. 2000. Continuous 

production of chito-oligosaccharides using a 

dual reactor system. Process Biochem. 35: 

623-632. 

Kim, S. B., B. K. Ham. B. O. Rhee, W. J. Lee and 

D. H. Jo. 2001. Effects of solvents on the 

viscosity of chitosan solution, pp.105-106. In 

T. Uragami, K. Kurita and T. Fukamizo, eds. 

Chitin and Chitosan in Life Science. 

Kodansha Scientific, Tokyo. 

Kobayashi, T., Y. Takiguchi, K. Shimahara and T. 

Sannan. 1988. Distribution of chitosan in 

Absidia strains and some properties of the 

chitosan isolated. Nipon Nogeikagaku 

Kaishi 62: 1463-1469. 

Kumar, G., J. F. Bristow, P. J. Smith and G. F. 

Payne. 2000. Enzymatic gelation of the natural 

polymer chitosan. Polymer. 41: 2157-2168. 

Kurakake, M., S. K. Nakagawa, M. Sugihara and 

T. Komaki. 2000. Properties of chitosanase 

from Bacillus cereus S1. Current Microbiol. 

40: 6-9. 

Miller, G. L. 1959. Use of dinitrosalicylic acid 

reagent for determination of reducing sugar. 

Anal Chem. 31: 426-428. 

No, H. K., N. Y. Park, S. H, Lee and S. P. Meyers. 

2001. Antibacterial activities of chitosan and 

chitosan oligomers with different molecular 

weights on spoilage isolated from tofu, pp. 

270-271. In T. Uragami, K. Kurita and T. 

Fukamizo, eds. Chitin and Chitosan in Life 

Science, Kodansha Scientific, Tokyo. 

Okajima, S. A. Ando, H. Shinoyama and T. Fujii. 

1994. Purification and characterization of an 


extracellular chitosanase produced by 

Amycolatopsis sp. CsO-2. J. Ferment. 

Bioeng. 77: 617-620. 

Papineau, A. M., D. G. Hoover, D. Knorr and D. 

F. Farkas. 1991. Antimicrobial effect of water 

soluble chitosan with high hydrostatic 

pressure. Food Biotecnol. 5: 45-57. 

Pelletier, A. and J. Sygusch. 1990. Purification and 

characterization of three chitosanase activities 

from Bacillus megaterium P1. Appl. Environ. 

Microbiol. 56: 844-848. 

Price, J. S. and R. Storck. 1975. Production, 

purification and characterization of an 

extracellular chitosanase from Streptomyces. 

J. Bacteriol. 124: 1574-1585. 

Shimosaka, M., M. Nagawa, X.Y. Wang, M. 

Kumehara and M. Okazaki. 1995. Production 

of two chitosanase from a chitosanassimilating 

bacterium, Acinetobacter sp. 

strain CHB101. Appl. Environ. Microbiol. 

61: 138-442. 

Shimosaka, M., Y. Fukumori, X. Y. Zhang, N. J. 

He, R. Kodaira and M. Okazaki. 2000. 

Molecular cloning and characterization of a 

chitosanase from the chitosanolytic bacterium 

Burkholderia gladioli strain CHB101. Appl. 

Microbiol. Biotechnol. 54: 354-360. 

Somashekar, D. and R. Joseph. 1996. 

Chitosanases-properties and application: a 

review. Biores. Technol. 55: 35-45 

Suzuki, K., Y. Okawa, K. Hashimoto, S. Suzuki 

and M. Suzuki.1984. Protecting effect of 

chitin and chitosan on experimentally induced 

marine candidiasis. Microb. Immunol. 28: 

903-912. 

Tominaga, Y. and Y. Tsujisaka. 1975. Purification 

and some enzymatic properties of the 

chitosanase from Bacillus R-4 which lyses 

Rhizopus cell walls. Biochim. Biophys. Acta 

410: 145-155. 

Uchida, Y. and A. Ohtakara. 1988. Chitosanase 

from Bacillus species. Method Enzymol. 

161: 501-505.

Uchida, K. Tateishi, O. Shida and K. Kodowaki. 

1992. Purification and properties of 

chitosanases from Bacillus licheniformis UTK 

and their application, pp.282-291. In C. J. 

Brine, P. A. Stanford and J. P. Zikakis, eds. 

Advances in Chitin and Chitosan. Elsevier, 

London. 

Wangtueai, S., W. Worawattanamateekul, 

M. Sangjindavong, N. Naranong and 

S. Sirisansaneeyakul. 2006. Isolation and 

screening of chitosanase producing 

microorganisms. Kasetsart J. (Nat. Sci.) 

40: 944-948. 


Yabuki, M., A. Uchiyama, A. Suzuki, A. Ando and 

T. Fujii.1988. Purification and properties of 

chitosanase from Bacillus circulans MH-K1. 

J. Gen. Appl. Microbiol. 34: 255-270. 

Yoshihara, K., J. Hosokawa, T. Kubo, M. 

Nishiyama, Y. Koba and M. Nishiyama. 1990. 

Isolation and identification of a chitosan 

degrading bacterium belonging to the genus 

Pseudomonas and the chitosanase production 

by isolate. Agric. Biol. Chem. 54: 3341-3343.


Application of Pectin Coating in the Production of 

Vitamin Fortified Rice 

Lalita Chatiyanont* and Phaisan Wuttijumnong 

ABSTRACT 

The quantity of vitamins in rice grain is decreased by milling, washing and cooking process. 

Therefore, the production of vitamin fortified rice using edible coating was investigated. Three types of 

low methoxyl pectin (36% degree of methoxyl, 31% degree of methoxyl with 21% degree of amidation 

and 28% degree of methoxyl with 18% degree of amidation) and control (no pectin coating) were 

studied. The results showed that L* a* b* values and moisture contents of rice premix were not 

significantly different (p > 0.05). Their values were 71.67-73.00, 13.07-14.32, 78.97-80.92 and 8.01- 

8.93%, respectively. Rice premix coated with pectin at 36% degree of methoxyl showed the lowest loss 

of thiamine, riboflavin and niacin during washing. However, pectin coating could not prevent the 

significant loss of thiamine and riboflavin during cooking in excess water (p > 0.05). The suitable ratio 

of rice premix to milled rice was 1:70. The cooked vitamin fortified rice at this ratio had 0.17 mg/100 g 

of thiamine and 27.89 mg/100 g of niacin content. The results of consumer acceptance test using Central 

Location Test (CLT) and Home Use Test (HUT) were similar. It was found that vitamin fortified rice 

was accepted by consumers at 95% (CLT) and 98% (HUT), respectively. 

Key words: low methoxyl pectin, edible coating, vitamin fortified rice 

INTRODUCTION 

Rice is a staple food of Thai population. 

Estimated consumption in 2004 was 10.24 million 

tons (Organization of Agricultural Economics, 

2005). Rice is eaten in 2 forms, brown and white 

rice. But the trend of eating white rice is still 

upward. Causes of nutrient loss especially soluble 

vitamins such as thiamine, riboflavin and niacin 

are milling, washing and cooking process. 

However, rice can be enriched to restore those lost 

in milling, washing and cooking by using edible 

coating. Peil et al. (1981) reported that rice coated 

with combined hydroxypropylmethylcellulose and 

methylcellulose (3:1 ratio) retained 70, 100, 18, 

18 and 21% of vitamin A, iron, niacin, thiamine 

and riboflavin, respectively. Shrestha et al. (2003) 

reported that rice premix coated with low methoxyl 

pectin retained 9% and 31% of folic acid during 

washing and cooking in excess water, respectively. 

The objectives of this study were to study 

the effects of edible coating on qualities of rice 

premix, to determine the suitable ratio of rice 

premix to milled rice in order to attain desired 

enrichment levels in the final product and to 

determine the consumer acceptance of edible 

coated vitamin fortified rice. 

Department of Product Development, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: nan_foodtu@hotmail.com 



1. Materials 

Milled rice (Khao Dauk Mali 105) was 

purchased from Tesco Lotus. Purple Ribbon Pure 

pectin (from yellow apple and citrus peel, Degree 

of methoxyl 36% and pectin content 85-100%) was 

obtained from Nutrition Partnership Limited. 7210 

pectin (from citrus peel, 28% Degree of methoxyl 

with 21% degree of amidation and 63% pectin 

content) and 7220 pectin (from citrus peel, 31% 

Degree of methoxyl with 18% degree of amidation 

and 63% pectin content) were obtained from The 

East Asiatic (Thailand) Public Company Limited. 

Thiamine hydrochloride, riboflavin and 

niacinamide were obtained from DSM Nutritional 

Product Co.,Ltd. 

2. Effects of edible coating on qualities of rice 

premix 

2.1 Preparation of mixed vitamin 

solution 

Mixed vitamin solution was prepared by 

dissolving 95 mg of the thiamine hydrochloride, 

52.6 mg of the riboflavin and 559 mg of the 

niacinamide in 8 ml of the distilled water. 

2.2 Preparation of low methoxyl pectin 

solutions 

Pectin solutions were prepared followed 

Rolin et al. (1998) by dissolving 1% of Purple 

Ribbon Pure pectin, 2% of 7220 and 7210 pectins 

in hot water (60-80°C) in a high-speed mixer. The 

viscosity of these solutions was 31-36 cP. 

2.3 Preparation of rice premix 

Rice grain (100 g) was coated with mixed 

vitamin solution and then with pectin solution 

followed by calcium chloride solution in a tablet 

coating pan (model SPKR, MITSUBISHI). The 

coating of pectin solution followed by calcium 

chloride solution was repeated and finally, the 

coated rice was dried in a dryer at 50 degree celcius 

for 2 hours. 

2.4 Washing and cooking of rice premix 


Washing test was carried out in a 250 ml 

erlenmeyer flask by rinsing 20 g rice premix with 

60 ml distilled water and gently swirling for 

exactly 60 s. In cooking test, 5 g of rice was cooked 

in 125 ml erlenmeyer flask with 100 ml distilled 

water for 30 min in a water bath (97 ± 3 °C) and 

cooled immediately. (Shrestha et al., 2003). 

2.5 Quality measurements 

2.5.1 Structure images analysis 

Unwashed, washed and cooked rice 

premixes were viewed on a Laser scanning 

confocal microscope (model AXIO, ZEISS Laser 

LSM 5 PASCAL). He/Ne laser at 488 nm was used 

as a light source to excite the riboflavin (Gue et 

al.,1999). The images acquired with a 5x, 0.15NA., 

dry objective and 512 × 512 pixel resolution. They 

were individual placed in a glass slide without 

further preparation. 

2.5.2 Color measurement 

The color characteristics (L* a* and b* 

values) of rice premixes were quantitatively 

measured using spectrophotometer (model CM- 

3500d , MINOLTA). L*, a* and b* values indicate 

lightness, red to green and yellow to blue, 

respectively. 

2.5.3 Moisture contents of rice premixes 

were analyzed by using hot air oven (model 

FD115, WTB binder) at 105 ± 1°C until the weight 

was constant (A.O.A.C., 2000). 

2.5.4 Determination of vitamin loss after 

washing and cooking 

Unwashed, washed and cooked rice 

premixes were analyzed for thiamine, riboflavin 

and niacin. Thiamine and riboflavin contents were 

determined by fluorometric method (A.O.A.C., 

2000). Niacin contents was determined by the 

Food Quality Assurance Service Center, Kasetsart 

University. 

3. The suitable ratio of rice premix to milled 

rice to attain desired enrichment level in the 

final product 

3.1 Preparation of fortified rice

358 

The rice premix coated with the low 

methoxyl pectin was obtained from part 2 

according to the highest vitamins retained after 

washing and cooking. The rice premix was blended 

with milled rice with different ratios (1:100, 1:85 

and 1:70) in a cubic mixing tank for 10 min. The 

milled rice was used as control sample. 

3.2 Washing and cooking of fortified 

rice 

Milled and fortified rice were washed the 

same way as in 2.4 using rice to water ratio 1 : 3. 

Cooking was done in automatic rice cooker (model 

SR-D10HN, Panasonic) using rice to water ratio 

of 1:1.25. 

3.3 Quality measurements 

3.3.1 Color measurement 

The color (L*, a* and b*) values of 

cooked fortified rice were measured by 

spectrophotometer. 

3.3.2 Determination of vitamin contents 

Unwashed, washed and cooked rice were 

analyzed for thiamine, riboflavin and niacin. 

Thiamine and riboflavin contents were determined 

by fluorometric method (A.O.A.C., 2000). Niacin 

contents was determined by the Food Quality 

Assurance Service Center, Kasetsart University. 

3.3.3 Sensory evaluation 

The likina scores of cooked rice were 

evaluated by 50 untrained panelists using 9-points 

hedonic scale (1 = dislike extremely to 9 = like 


extremely) for appearance, color, odor, flavor and 

overall liking. 

4. Consumer acceptance test 

Consumer acceptance test was carried 

out using Central Location test (CLT) and Home 

Use Test (HUT). 100 consumers were used in CLT 

at two locations (Kasetsart University cafeterias 

1 and 2). The samples (before and after cooking 

fortified rice) and questionnaires were provided 

for the consumers. For HUT, 100 consumers were 

provided with samples (fortified rice 142 g for 1 

meal) and questionnaires. The 9-points hedonic 

scale was used to score the consumers’ liking. The 

acceptability of fortified rice was also evaluated 

by consumers. 


1. Effect of edible coating on qualities of rice 

premix 

1.1 Physical properties 

The appearances of unwashed, washed 

and cooked rice premix coated with pectin viewed 

by confocal laser scanning microscopy were 

shown in Figure 1. It was found that there were 

cracks in washed rice and the kernel shape seems 

to be lost in cooked rice. This may cause a heavy 

losses of vitamins in rice premix. The color 

characteristics (L* a* and b* values) of rice premix 

Figure 1 Rice premix as viewed in the CLSM at 5X (a) before cooking (b) after washing (c) after 

cooking in excess water and draining.

without coating and coated with pectins were not 

significantly different (p>0.05). Their values were 

between 71.67-73.00, 13.07-14.32 and 78.97- 

80.92 for L*, a* and b* values, respectively. The 

rice premix has yellow color (high b* value) due 

to addition of riboflavin. 

1.2 Chemical properties 

Moisture contents of all rice premix 

samples were not significantly different (p>0.05). 


Their values were between 8.01-8.93%. Table 1-3 

showed the loss of thiamine riboflavin and niacin 

in washed and cooked rice premix without coating 

and coated with pectins in excess water and 

draining. The rice premix coated with pectins 

showed lower vitamin losses after washing than 

those without coating. The higher degree of 

methoxyl pectin showed the lower vitamin losses 

in washed rice premix than lower degree of 

Table 1 Thiamine contents in rice premix, washed and cooked rice in excess water and draining, 

washing and cooking losses. 

Rice premix Thiamine contents (mg/100 g) Washing loss Cooking 

Rice premix Washed rice Cooked rice (%) loss (%) 

premix premix (ns) 

- No pectin coating 

- Coated with pectin 

23.47 ± 0.13 4.92 ± 0.24 1.68 ± 0.20 79.03 ± 1.14 a 92.84 ± 0.88 

36 % degree of methoxyl 



25.84 ± 0.12 11.41 ± 0.02 1.37 ± 0.32 55.86 ± 0.29 c 94.69 ± 1.28 

with 18 % degree of amidation 



19.28 ± 4.50 8.19 ± 0.93 1.46 ± 0.05 56.90 ± 5.23 c 92.15 ± 2.09 

with 21% degree of amidation 24.25 ± 0.27 7.57 ± 0.15 0.84 ± 0.00 68.78 ± 0.99 b 96.67 ± 0.23 

Note: alphabets a-c were different within column mean values were significantly different (p≤0.05) 

ns means values within column were not significantly different (p>0.05) 

Table 2 Riboflavin contents in rice premix, washed and cooked rice in excess water and draining, 

washing and cooking losses. 

Rice premix Riboflavin contents (mg/100 g) Washing loss Cooking 





41.98 ± 2.76 8.19 ± 1.41 2.82 ± 0.20 80.54 ± 2.07 a 93.28 ± 0.50 




38.68 ± 3.46 12.52 ± 1.46 3.01 ± 0.25 67.66 ± 0.90 c 92.17 ± 1.34 



28 %degree of methoxyl 

45.48 ± 0.30 12.34 ± 0.33 3.32 ± 0.01 72.86 ± 0.55 b 92.69 ± 0.08 

with 21 % degree of amidation 38.81 ± 1.79 10.91 ± 0.50 2.90 ± 0.02 71.28 ± 2.60 bc 92.53 ± 0.29 

Note: alphabets a-c were different within column mean values were significantly different (p≤0.05) 

ns means values within column were not significantly different (p>0.05)

360 

methoxyl pectin. This may be due to the fact that 

the presence of calcium ion was not enough to 

strengthen gel (Thakur et al., 1997). 

However, pectin coatings were not good 

enough to prevent leaching of these vitamins from 

rice premix when boiled in excess water. The 

preparation process of rice premix consisted of 

many steps of coating and drying which had an 

effect on rice cracking. During boiling, water can 

easily access into the interior of the cracked grain, 

this causes increasing of hydration and 

subsequently leaching of vitamin into the cooking 

water. (Shrestha et al., 2003) 

Since, the rice premix coated with pectin 

(36% degree of methoxyl) had the lowest vitamin 

loss during washing and cooking, it was selected 


for the next experiment. 

Table 3 Niacin contents in rice premix, washed and cooked rice in excess water and draining, washing 

and cooking losses. 

Rice premix Niacin contents (mg/100 g) Washing loss Cooking 





429.28 ± 11.28 92.46 ± 14.67 22.87 ± 0.83 78.41 ± 3.99 a 94.67 ± 0.06 




473.71 ± 0.68 211.14 ± 2.82 25.87 ± 0.37 55.43 ± 0.54 b 94.54 ± 0.08 




405.07 ± 23.81 182.41 ± 16.54 22.53 ± 0.07 54.77 ± 6.74 b 94.43 ± 0.35 

with 21 % degree of amidation 478.86 ± 30.57 181.14 ± 2.93 24.88 ± 0.25 62.08 ± 3.03 b 94.79 ± 0.28 

Note: value in the same column with different superscripts differ significantly (p≤0.05) 

ns means values within column were not significantly different (p>0.05) 

Table 4 L* a* and b* values of cooked vitamin fortified rice. 

Rice premix to milled rice L* a* b* 

Milled rice 77.49 ± 0.10 a -2.11 ± 0.06 b 8.94 ± 0.22 d 

Fortified rice 

(rice premix to milled rice) 

1 : 100 77.46 ± 0.12 a -2.59 ± 0.06 a 9.93 ± 0.17 c 

1 : 85 77.47 ± 0.10 a -2.63 ± 0.06 a 11.43 ± 0.07 b 

1 : 70 76.64 ± 0.02 b -2.57 ± 0.07 a 11.66 ± 0.19 a 


2. The suitable ratio of rice premix to milled 

rice to attain desired enrichment level in the 

final product 

2.1 Physical and chemical properties 

Cooked fortified rice had light yellow in 

color, due to the leaching out of vitamins from 

surface of rice premix during washing and 

cooking. The a* and b* values increased with 

increasing rice premix to milled rice ratios (Table 

4). 

The amount of vitamins (thiamine, 

riboflavin and niacin) in unwashed, washed and 

cooked milled rice and fortified rice were shown

in Table 5-7. It was found that the amount of 

vitamins in unwashed, washed and cooked rice 

increased with increasing ratios of rice premix to 

milled rice). The ratio of rice premix to milled rice 

at 1:70 met the requirement for thiamine and niacin 

fortification of rice, according to Thai Reference 

Daily Intake (Thai RDI) in which the cooked 

fortified rice should have thiamine and niacin 

contents more than 10% of cooked milled rice. 

But the riboflavin content was less than 10% of 

cooked milled rice. 


2.2 Sensory evaluation 

The liking score for each attribute of 

cooked milled rice and fortified rice were not 

significantly different (p>0.05) (Table 8). This 

indicates that the fortification of rice with vitamins 

by mixing rice premix with milled rice had no 

effects on the panelists preference. 

Therefore, the ratio of rice premix to 

milled rice at 1:70 was selected for study on 

consumer acceptance. 

Table 5 Amount of thiamine in unwashed, washed and cooked rice and fortified rice. 

Rice premix to milled rice Amount of thiamine (mg/100 g) 

Unwashed Washed rice Cooked rice 

Milled rice 0.05 ± 0.01 c 0.04 ± 0.01 c 0.01 ± 0.03 c 



1 : 100 0.28 ± 0.01 b 0.17 ± 0.00 b 0.10 ± 0.00 b 

1 : 85 0.30 ± 0.01 b 0.19 ± 0.01 b 0.12 ± 0.01 b 

1 : 70 0.39 ± 0.21 a 0.25 ± 0.03 a 0.17 ± 0.01 a 


Table 6 Amount of riboflavin in unwashed, washed and cooked rice and fortified rice. 

Rice premix to milled rice Amount of riboflavin (mg/100 g) 

Unwashed Washed rice Cooked rice 

Milled rice 0.04 ± 0.00 b 0.03 ± 0.00 c 0.01 ± 0.00 b 



1 : 100 0.43 ± 0.04 a 0.12 ± 0.01 b 0.04 ± 0.01 a 

1 : 85 0.52 ± 0.04 a 0.18 ± 0.02 a 0.05 ± 0.01 a 

1 : 70 0.55 ± 0.01 a 0.20 ± 0.00 a 0.05 ± 0.00 a 


Table 7 Amount of niacin in unwashed, washed and cooked rice and fortified rice. 

Rice premix to milled rice Amount of niacin (mg/100 g) 

Unwashed Washed rice (ns) Cooked rice 

Milled rice 



27.86 ± 3.72 b 26.17 ± 1.39 14.70 ± 1.50 b 

1 : 100 49.25 ± 0.12 a 30.10 ± 0.80 12.09 ± 4.23 b 

1 : 85 49.77 ± 0.52 a 28.83 ± 0.21 11.34 ± 4.57 b 

1 : 70 48.33 ± 0.97 a 32.12 ± 8.07 27.89 ± 0.44 a 


ns means values within column were not significantly different (p>0.05).

362 

3. Consumer acceptance test 

The results showed that the vitamin 

fortified rice was significantly accepted by the 

consumers at 95% (Central Location Test) and 

98% (Home Use Test). The overall liking scores 

of the vitamin fortified rice before and after 

cooking were 6.9 and 7.5 (CLT) ; 7.4 and 7.8 

(HUT), respectively. 

CONCLUSION 

Rice premix coated with pectin (36% 

degree of methoxyl) had a minimal loss of vitamins 

during washing. But pectin coating could not 

prevent vitamins from cooking loss. The vitamins 

fortified rice at ratio of 1:70 was suitable. It was 

significantly accepted by consumers at 95% and 

98% with overall liking scores 7.5 and 7.8 for 

Central Location Test and Home Use Test, 

respectively. 

For further study, we recommend to 

focus on the protein-based films because they are 

better in mechanical and barrier properties than 

polysaccharide based films. Therefore it might 

protect vitamin loss in washing and cooking 

process. 


This research was kindly supported by 

The Thailand Research Fund (TRF) in major 

Science and Technology, 2005. 


Table 8 Liking scores of cooked rice and fortified rice. 

Attributes Rice premix to milled rice 

Milled rice 1 : 100 1 : 85 1 : 70 

Appearance (ns) 6.7 ± 1.4 6.7 ± 1.3 6.4 ± 1.4 6.7 ± 1.3 

Color (ns) 7.0 ± 1.3 6.7 ± 1.4 6.5 ± 1.3 6.6 ± 1.1 

Odor (ns) 5.9 ± 1.1 6.1 ± 1.5 6.1 ± 1.5 6.2 ± 1.7 

Flavor (ns) 6.1 ± 1.2 6.3 ± 1.2 6.2 ± 1.3 6.5 ± 1.2 

Texture (ns) 6.1 ± 1.5 6.0 ± 1.4 5.9 ± 1.4 6.4 ± 1.4 

Overall liking (ns) 6.3 ± 1.5 6.3 ± 1.4 6.1 ± 1.3 6.6 ± 1.3 

Notes: ns means values within row were not significantly different (p>0.05). 


A.O.A.C. 2000. Official Method of Analysis. 17 th 

ed., The Association of Official Analytical 

Chemists Gaithersburg, Maryland. 

Guo, H.X., J. Heinämäki and J. Yliruusi. 1999. 

Characterization of particle deformation 

during compression measured by confocal 

laser scanning microscopy. International 

Journal of Pharmaceutics 186: 99-108. 

Organization of Agricultural Economics Ministry 

of Agriculture and Cooperative. 2005. Data 

Base of Economic. Organization of 

Agricultural Economics Ministry of 

Agriculture and Cooperative, Bangkok. 

Peil, A., F. Barrett, C. Rha and R. Langer. 1981. 

Retention of micronutrients by polymer 

coatings used to fortify rice. J. Food Sci. 47: 

260-262, 266. 

Rolin, C., B.U. Niclsen and P.E. Glahn. 1998. 

Pectin, pp. 377-431. In S. Dumitriu (ed.). 

Polysaccharide, structural diversity and 

functional versatility, Marcel Dekker, Inc., 

New York. 

Shrestha, A.K., J. Arcot and J.L. Paterson. 2003. 

Edible coating materials-their properties and 

use in the fortification of rice with folic acid. 

Food Reseach International 36: 921-928. 

Thakur, B.R., R.K. Singh and A.K. Handa. 1997. 

Chemistry and Uses of Pectin-A Review. 

Critical Reviews in Food Science and 

Nutrition 37: 47-73.


The Effects of Starter Cultures on Biogenic Amine and 

Free Amino Acid Contents in Nham during Fermentation 

Sasithorn Limsuwan 1 *, Wonnop Visessanguan 2 and Jirasak Kongkiattikajorn 1 

ABSTRACT 

Fermented pork sausage, or Nham, is a Thai-style fermented food which relies mainly on 

adventitious microorganisms, normally found in raw materials. The fermented foods usually contain 

biogenic amines produced by the microorganisms which caused the reaction of amino acids 

decarboxylation. These compounds are associated with toxicological symptoms. The objective of this 

study was to study the influence of two decarboxylase negative starter cultures in the presence of biogenic 

amines and free amino acid contents in Nham. Derivative biogenic amines by dansyl chloride were 

determined by high performance liquid chromatography (HPLC). Cadaverine and tyramine were 

determined during ripening process. The highest concentrations of biogenic amines were found at the 

end of the ripening process in the control sausage with no starter culture. Starter cultures test showed 

that Lactobacillus sakei BCC102 and Debaryomyces hansenii BCC 106 were efficient in reducing the 

amine production since these strains caused a quick pH drop during sausage fermentation. Total free 

amino acids after fermentation process decreased and the high decreases in the contents were glutamine 

and arginine while tyrosine and lysine, precursors for tyramine and cadaverine, respectively, increased 

in all batches. This study suggested that the use of decarboxylase negative lactic acid bacteria as starter 

cultures, which produced a rapid decrease on the pH of the meat mixture, was important factor to be 

considered to reduce the levels of biogenic amines in Nham. 

Key words: starter culture, Nham, fermentation, biogenic amine, amino acid 

INTRODUCTION 

Biogenic amines (BAs) are naturally 

present in many foods and relatively high contents 

of some BAs can be present in fermented foods. 

BAs are organic molecules with low molecular 

weight. These compounds are usually generated 

by microbial decarboxylation of amino acids 

present in foods. The aromatic BAs (histamine, 

tyramine, serotonin, β-phenylethylamine, 

tryptamine) have been reported as vasoactive or 

psychoactive amines and they have been 

associated with food histaminic intoxications, 

food-induced migraines, and severe hypertensive 

crisis due to monoamine oxidase inhibitor (MAOI) 

drug interactions. Moreover, diamines such as 

putrescine and cadaverine could generate 

carcinogenic nitrosamines in the presence of 

nitrites (Scanlan, 1983). Interest in cadaverine, 

putrescine, tyramine and histamine also lies in their 

1 School of Bioresources and Technology, King Mongkut’s University of Technology Thonburi, Bangkok 10140, Thailand. 

2 National Center for Genetic Engineering and Biotechnology, Phathum Thani 12120, Thailand. 

* Corresponding author, e-mail: jirasak.kon@kmutt.ac.th 


364 

potential as spoilage indicators of food. In addition, 

they may have unpleasant odours and it was also 

found that putrescine and cadaverine could inhibit 

the activity of muscle aminopeptidases (Flores et 

al., 1996). 

Nham is a Thai traditional fermented 

pork sausage. Nham fermentation generally takes 

3-5 days. The microorganisms involved in the 

fermentation process can yield much higher BA 

amounts than those found in the corresponding raw 

materials, because some BAs are the result of 

amino acid decarboxylation by microbial enzymes. 

BAs may represent a food-poisoning hazard in 

conjunction with additional promoting factors such 

as MAOI antidepressant drugs, alcohol, other food 

amines, gastrointestinal diseases and genetically 

deficient detoxification systems (Vidal-Carou et 

al., 1990). 

Meat fermentation offers favourable 

conditions for BA formation, since the main 

required factors are present, i.e. there is growth of 

microorganisms over several days, a certain degree 

of proteolysis takes place giving rise to the 

presence of free amino acids as precursors of BA 

and, finally, the existence of an acidic environment 

can favour the amino acid decarboxylase activity 

of microorganisms. It has been reported that 

bacteria could be encouraged to produce 

decarboxylase enzymes in such acidic 

environments as part of their defense mechanisms 

against adverse conditions (Eitenmiller et al., 

1978). Since microbial flora naturally present in 

the raw materials seem to have a strong influence 

on BA formation during ripening, the choice of 

good quality raw materials helps to minimize the 

number of amine-producing bacteria (Halasz et al., 

1994). An important factor suggested for 

preventing amine accumulation is the addition of 

adequate starter cultures to complete the 

fermentation. Starter cultures usually consist of 

one or several strains such as lactic acid bacteria 

(LAB). LAB are being widely used as starter 

cultures in fermented sausages. The absence of BA 


formation of LAB was proposed as a selection 

criterion for starter cultures (Buckenhuskes, 1993). 

Proteolysis during the fermentation of 

meat products is favoured by the denaturation of 

proteins as a consequence of the acidity, 

dehydration and the action of sodium chloride 

(DeKetelaere et al., 1974). During the 

fermentation, production of free amino acids from 

proteolysis might occur. Therefore, the 

determination of free amino acid contents can be 

useful in studying the potential relationship 

between proteolysis and BA formation in 

fermented sausages. 

The objectives of the present study were: 

(1) to study the changes in BA levels during the 

fermentation processes (2) to examine the effect 

on BA formation of L. sakei BCC 102 and D. 

hansenii BCC 106 which are nondecarboxylase 

activity used as starter cultures added naturally 

fermented Nham and (3) to determine the effects 

of starter culture on the formation of free amino 

acid during the fermentation of Nham (4) to 

compare the formation of BAs in control Nham 

(naturally fermented) and in Nham fermented with 

starter microorganisms. 


Microorganisms 

Two starter culture strains (Lactobacillus 

sakei BCC 102 and Debaryomyces hansenii BCC 

106) isolated from Nham were chosen after 

screening for nondecarboxylase activities. Both 

strains were gift from the Culture Collection 

Laboratory, National Center for Genetic 

Engineering and Biotechnology (BIOTEC), 

Patumthani, Thailand. Cultures were stored at - 

80°C in 20% glycerol. One loopful of a stock 

culture was cross-streaked on Man, Rogosa and 

Sharpe (MRS) agar and then incubated at 30°C 

for 48 h. A single colony on MRS agar was grown 

in MRS broth at 30°C for 18-24 h. Cell-free 

supernatants were obtained after centrifugation of

the cultures at 10,200 × g for 15 min at 4°C. The 

starter culture was prepared to obtain an 

approximate cell concentration of 10 7 CFU/ml in 

sterile deionized water. 

Preparation of Nham 

Nham was prepared to make a total 100 

g Nham by mixing 52 g minced pork, 35g cooked 

pork rind, 1.9 g curing salt, 0.2 g sodium 

erythrobate, 0.2g sodium tripolyphosphate, 4.3 g 

minced garlic, 4.3 g minced cooked rice, 2 g chilli, 

0.4 g sucrose, 0.2 g monosodium glutamate, 0.01 

g potassium nitrite and 0.6 g sodium chloride. The 

ingredients were thoroughly mixed and divided 

into three fractions for three different batches, i.e. 

control or naturally fermented without the addition 

of starter cultures (NC) and batches I and II 

processed through a starter-mediated fermentation, 

with a single starter culture of 10 4 CFU/g L. sakei 

BCC 102 (LS batch) and mixed starter culture 

consisting 10 4 CFU/g L. sakei BCC 102 combined 

with 10 6 CFU/g D. hansenii BCC 106 (LSY 

batch), respectively. Approximately 200 g of Nham 

was stuffed into a plastic casing 3 cm diameter 

and sealed tightly prior to incubation at 30°C for 

120 h. Samples were taken every 24 h for chemical 

and microbiological analysis. 

pH Measurement 

pH of Nham samples of 0-5 days of 

fermentation were measured directly using a 

microcomputerized pH meter (Mettler Teledo 320, 

UK; Mettler Toledo, Inlab® 427) by inserting the 

electrode into the centre of Nham and recorded as 

the mean value of three measurements. 

Microbiological analysis 

For microbial analysis, 25 g of Nham was 

aseptically removed from the casing, cut into small 

pieces, placed in sterile Stomacher bag and 

homogenized using a Stomacher (IUL Instrument, 

Spain) with 225 ml of sterile diluent containing 

0.1% peptone. Serial decimal dilutions were 


prepared. LAB were enumerated on MRS agar 

adding with 0.5 % calcium carbonate and 

incubated anaerobically at 30°C for 48 h. 

Biogenic amine determination 

HPLC determinations were performed 

with a LC 10 AD Shimadzu LC using a 20 µl loop. 

Detection was at 254 nm with UV detector. LC 

column C18-Hypersil BDS (200 mm.× 4.6 mm, 5 

µm particle size) was used. Amine standard 

solutions were prepared in water to a final 

concentration of 5 mg/ml for each biogenic amine. 

Tyramine, putrecine, cadaverine, tryptamine, 

phenylethylamine and histamine were used. 

Biogenic amine concentrations in the working 

standard solutions chosen for the calibration curve 

were 0.005, 0.01, 0.05, 0.1, 0.5 and 1 mg/mL. 

These working solutions were made by further 

dilution of the stock solution with water. Internal 

standard solution was prepared by diluting 15 mg 

of 1, 7-diaminoheptane in 5 ml of water. The 

gradient-elution system was methanol as solvent 

A and water as solvent B. The system was 

equilibrated for 5 min before the next analysis. 

The flow rate was 1.5 ml/min. 

Sample preparation and extraction 

Four grams of sample was mixed with 

10 ml of 5% trichloroacetic acid and extracted 

using homogenizer. The homogenate was 

centrifuged at 17,212 × g for 10 min at 4°C, the 

supernatant was collected and the precipitate was 

extracted again with 10 ml of 5% trichloroacetic 

acid. After centrifugation, the supernatant was kept 

at -20°C. 

Derivatization of sample extracts and mixed 

standards 

A 100 µl of 2 N NaOH and 150 of µl 

saturated NaHCO 3 were added to 0.5 ml of the 

extract, mixed with 1 ml of dansyl chloride (10 

mg/ml in acetone) and incubated at 40°C in a water 

bath for 45 min. To remove residual dansyl

366 

chloride, 50 µl of 100% ammonia was added and 

the solution was centrifuged at 500 × g for 30 min 

and the supernatant was filtered through a 0.45 

mm filter. Dansyl derivatives of the calibration 

standards were mixed with the samples as 

previously described (Eerola et al., 1994). 

Free amino acid (FAA)determination 

Free amino acids were determined 

according to the method of Benjakul and 

Morrissey (1997) using an amino acid analyzer 

(Waters 2690 Alliance with 280 nm Fluorescent 

detector). The column was an AccQ-Tag TM C18, 

4 µm. The solvent system consisted of three 

eluents: (A) AccQ Tag Eluent pH 5.02; (B) HPLCgrade 

acetonitrile and (C) Nanopure distilled water. 

The flow rate was set at 1.0 ml/min. Five g of 

Nham blend was mixed with 20 ml of 5% 

trichloroacetic acid, then stomachered at 200 

rpm/min for 8 min and centrifuged at 12,000 × g 

for 15 min. A 100 µl of supernatant was mixed 

with 20 µl of 2.5 mM ABAA and 800 ml nanopure 

pH 

6.5 

6 

5.5 

5 

4.5 

4 

3.5 

3 


distilled water and filtered through 0.45 µm 

(Minisart RC4, Sartorious), then 10 µl aliquot of 

filtrate was transferred into a vial, and 70 µl of 

Waters AccQ Fluor borate buffer was added. A 

20 µl of AccQ Fluor reagent was added and the 

mixture was incubated at 55°C for 10 min before 

HPLC analysis. 

Statistical analyses 

The differences between the results of 

physical, chemical and microbiological analysis 

of Nham fermented by different strains were tested 

using one-way analysis of variance (ANOVA). 


pH determination 

Changes of pH in Nham during ripening 

are shown in Figure 1. The pH of the control 

sample and the batch with L. sakei and D. hansenii 

decreased after 4 h of incubation while that of 

batch with L. sakei decreased after 8 h of 

0 20 40 60 80 100 120 140 

Time (h) 

NC LS LSY 

Figure 1 Changes in pH values during the ripening of Nham from different batches. Nham without 

added cultures (NC), Nham fermented with L. sakei BCC102 (LS), L. sakei BCC102 and D. 

hansenii BCC106 (LSY), for 120 h at 30°C (each data point represents the mean and standard 

deviation of three independent trials).

incubation and reached the final pH values of 4.5, 

4.2, and 4.2 for the control, batch with L. sakei 

and batch with L. sakei and D. hansenii, 

respectively. The initial pH of all batches of Nham 

was pH 6.1, thereafter it decreased rapidly in the 

batch with starter cultures to pH 4.7, at 28 h. The 

pH of the control Nham was decreased to 4.9 at 

28 h of fermentation. After 28 h, the pH in all 

batches gradually decreased throughout of 

incubation and pH values of the control slightly 

decreased less than that of the batches with starter 

cultures. The pH reduction during processing 

probably due to organic acid production by the 

inoculated starter cultures as well as the lactic flora. 

Statistical analysis of pH values recorded 

throughout ripening revealed significant 

differences between treatments at 24 h. 

Microbiological analyses 

Microbial counts increased in both the 

controls and starter added samples during 

fermentation (Figure 2). Initial lactic acid bacteria 

Log total LAB (CFU/g) 

10 

8 

6 

4 

2 

0 


(LAB) counts in the starter added samples were 

higher than in the control samples. LAB counts 

increased during fermentation for both control and 

starter added samples. Initial counts of LAB (7.08 

± 0.24 log, 8.04 ± 0.32 log and 8.11 ± 0.21 log 

CFU/g for control, L. sakei added samples and 

mixed cultures of L. sakei and D. hansenii added 

samples, respectively) increased during 

fermenting, till the microorganism being at 9.48 ± 

0.22 log, 9.45 ± 0.15 log and 9.66 ± 0.19 log CFU/ 

g in the control, L. sakei added samples and mixed 

culture of L. sakei and D. hansenii added samples, 

respectively. This was an increase due to the 

environmental conditions which made gramnegative 

bacterial grow. After 16 h of fermentation, 

no significant differences were observed in total 

LAB counts in all batches. LAB increased during 

the ripening process, reaching maximum levels at 

day five in all types of sausages. 

Changes in microbial counts in Nham 

inoculated with single starter culture of L. sakei 

and mixed starter cultures of L. sakei and D. 

0 20 40 60 80 100 120 140 

Time (h) 

NC LS LS Y 

Figure 2 Changes in the population levels of LAB in Nham without added cultures (NC), Nham 

fermented with L. sakei BCC102 (LS), L. sakei BCC102 and D. hansenii BCC106 (LSY), for 

120 h at 30°C (each data point represents the mean and standard deviation of three independent 

trials).

368 

hansenii were similar to that of naturally fermented 

Nham (Figure 1). This microbial group rapidly 

increased after casing and reached the values of 

about 10 9 CFU/g in all the sausages, even in the 

samples to which starter cultures were not added. 

These high values remained relatively constant 

during ripening. Due to relatively high microbial 

load in Nham raw mix(10 7 CFU/g), inoculation 

of starter cultures at levels of 10 4 and 10 6 CFU/g 

had no significant effects on the LAB counts 

during fermentation. Similar to the results of 

Khieokhachee et al. (1997), initial flora of the 

Nham derived mainly from the raw materials. The 

number of LAB increased drastically to a 

maximum of 10 9 CFU/g within 18 h and remained 

the same until the fermentation was completed at 

24 h for all batches . Thus, fermentation of Nham 

involving successive growth of different groups 

of microorganisms was dominated by LAB. 

Various metabolic products of LAB, such as shortchain 

organic acids, carbon dioxide, hydrogen 

peroxide, diacetyl, and bacteriocin, are known as 

antimicrobial agents (Rowan et al., 1998). 

Accumulation of organic acids also resulted in a 

decrease of pH. Thus, the dominance of LAB is 


likely to contribute to the inhibition of other 

microorganisms. 

The addition of L. sakei and D. hansenii 

as starters might prevent the development of flora 

LAB that were present naturally in the initial 

mixture, and L. sakei might be able to dominate 

the whole ripening period in the batches while the 

control predominated with LAB flora. 

Biogenic amines determination 

Tyramine and cadaverine were present 

exclusively in 24 h fermented samples in typical 

quantitative sequence: cadaverine content was 

more than tyramine content (Table 1), other amines 

were not detectable. Both biogenic amines 

increased after 24 h of ripening until the final 

ripening of 120 h. 

The differences between Nham 

elaborated with and without starter culture were 

observed, and the control Nham had significantly 

higher values of cadaverine and tyramine than the 

Nham inoculated with starter cultures. The L. sakei 

had significantly lower concentrations of biogenic 

amines than L. sakei and D. hansenii. Cadaverine 

and tyramine in fermented sausages were produced 

Table 1 Biogenic amine contents (mg/kg) in Nham. 

Time (h) Control L. sakei L. sakei and D. hansenii 

CAD TYR CAD TYR CAD TYR 

0 ND ND ND ND ND ND 

24 135.47 74.28 174.54 57.28 197.41 67.91 

± 14.85 aA ± 5.71 aB ± 18.22 aC ± 7.96 aD ± 16.34 aF ± 7.14 aG 

48 197.32 95.72 204.69 68.46 201.36 94.85 

± 18.92 bA ± 8.42 bB ± 23.85 bA ± 4.85 bC ± 14.89 aA ± 5.54 bB 

72 216.41 97.61 225.87 75.37 223.52 96.47 

± 21.56 cA ± 7.48 bB ± 19.51 cC ± 6.29 bcD ± 26.25 bAC ± 8.20 bB 

96 218.82 102.32 218.34 77.16 227.28 104.62 

± 24.63 cA ±12.85 bB ± 25.18 cA ± 7.28 cC ± 21.94 bA ± 7.87 bcB 

120 235.95 138.59 224.74 82.23 232.86 107.58 

± 17.44 dA ± 14.74 cB ± 21.82 cC ± 5.45 dD ± 25.17 cA ± 6.41 cE 

Mean values and standard deviations with different letters (a, b, c) in the same column indicate significant differences (P

y lysine- and tyrosine-decarboxylase activities, 

of Enterobacteriaceae, respectively. So, L. sakei 

and D. hansenii in Nham might inhibit the growth 

of Enterobacteriaceae resulting in decrease the 

lysine- and tyrosine-decarboxylase activity and 

biogenic formation (Bover-Cid et al., 2001b). 

The main factors seemed to be a suitable 

starter culture and good quality raw materials 

(Bover-Cid et al., 2001a). However, in the present 

study, the high quality raw materials used were 

not effective in preventing the production of 

cadaverine and tyramine in control Nham (Table 

1), and low contents of these amines were obtained 

only when a starter culture was included in sausage 

formulation. 

In conclusion, to avoid the presence of 

high concentrations of biogenic amines in Nham, 

it was advisable to use a competitive starter culture 

such as L. sakei, a negative-decarboxylase strain, 

which might be predominant throughout the 

process, thus it would prevent the growth of 

bacteria which could produce biogenic amines. 

Low occurrence of biogenic amines in 

raw pork meat: usually tyramine did not exceed a 

few mg kg -1 (tyramine less than 3.5 mg kg -1 as 

observed by Hernandez-Jover et al., 1997) while 

ripened and cured meat showed a general increase 

of amines (Bover-Cid et al., 1999). The choice of 

starters could be useful tool to control and reduce 

the development of some Enterobacteriaceae 

strains. However, the presence of biogenic amines 

in ripened dry uncooked fermented meat was 

fundamentally a consequence of the activity of 

decarboxylase-positive strains of Lactobacillaceae 

and Enterococcaceae. 

Biogenic amines content depended also 

on an equilibrium between the decarboxylating and 

amine oxidizing activity of microflora (Gardini et 

al., 2002). 

Therefore, to obtain Nham with low 

amine concentrations besides the high quality raw 

materials and good manufacturing practices, it is 

necessary to employ highly competitive amino 


acids decarboxylase negative starters cultures and 

the starter culture should be able to compete and 

grow well at the temperature intended for 

processing of the product (Maijala et al., 1995). 

Analysis of FAA 

The main differences in the content of 

total FAA among batches were detected at the end 

of the processing (5 days), where lower quantities 

were detected in all batches. From Table 2, after 5 

days of incubation at 30°C, NC caused decrease 

of 36.8 % in total FAA while Nham with L. sakei 

and the Nham with mixed cultures of L. sakei and 

D. hananii caused decrease of 13.3% and 19.73%, 

respectively, in the concentration of FAA. The total 

FAA of Nham with starter culture was higher than 

that of the control (Table 2). 

This suggested that the starter cultures 

batches might have higher proteolytic activity than 

the non-inoculated control batch and /or 

catabolized free amino acids to be the other 

products such as biogenic amines less than the 

control due to the batches added with starters 

lagged of amino acid decarboxylase. 

Free amino acids precursors of biogenic 

amines were detected by HPLC. During 

fermentation step, the increases of tyrosine and 

lysine which were the precursors of tyramine and 

cadaverine, respectively were obsereved in all 

batches. However, after 5 days of ripening, the 

concentrations of tyrosine and lysine in Nham with 

starter cultures were more than that of the control, 

while the biogenic amines, tyramine and 

cadaverine in the control were more than that of 

the batches with starter cultures. 

Evaluation of FAA during the ripening 

of fermented Nham sausages indicated an increase 

in most FAA over the 0-5 day fermentation period. 

The main changes observed in the 

decrease of free amino acids at the end of 

processing showed a higher decrease proportion 

of glutamic acid and arginine in the control than 

that of the batch with L. sakei and the batch with

370 

L. sakei and D. hansenii after 5 days of ripening. 

The decreases in glutamic acid and arginine 

contents might be due to these amino acids were 

used for the growth of the microorganisms and 

might be metabolized to flavours. The quantities 

of alanine, aspartic acid, glycine, isoleucine, 

leucine, methionine, phenylalanine, tyrosine, 

valine and lysine in the control after 5 days of 

ripening were higher than those before ripening. 

Some of these amino acids in the batches with L. 

sakei and the batches with L. sakei and D. hansenii 

also increased after ripening. 

Some amino acids, especially those 

branched-chain amino acids, have been 

metabolised to generate volatile compounds (Dura 

et al., 2004). The contents of alanine, isoleucine, 

histidine and proline were similar in the three 


batches at the end of processing. Alanine, 

contributors of sweet taste was found in higher 

contents after ripening of fermented sausages. 

Therefore, the balance of these free amino acids 

would affect the sensory characteristics of the 

product (Ordonez et al., 1999). The addition of 

starter culture produced a limited effect on the free 

amino acid generation although the effect was 

different depending on the quantity of 

microorganisms inoculated. Many factors could 

affect the generation of free amino acids such as 

the presence of different substrates, the pH, the 

presence of different microorganisms and their 

evolution during processing. The significant (P < 

0.05) reduction in the concentration of free amino 

acids could be produced by a more intense 

microorganism metabolism than their production 

Table 2 Total and free amino acid contents (mg/100 g) in Nham during fermentation. 

Amino Starter culture 

acid NC, 0 h NC, 120 h L. sakei L. sakei L. sakei and L. sakei and 

0 h 120 h D. hansenii D. hansenii 

0 h 120 h 

Ala 1.59 ± 0.16a 3.13 ± 0.29b 1.98 ± 0.04c 3.59 ± 0.12b 1.96 ± 0.05c 3.15 ± 0.36b Asp 0.28 ± 0.04a 0.49 ± 0.04b 0.27 ± 0.01a 0.81 ± 0.10c 0.25 ± 0.01a 0.76 ± 0.03c Gly 0.84 ± 0.10a 1.47 ± 0.07b 0.95 ± 0.02c 1.87 ± 0.18d 0.93 ± 0.01c 1.69 ± 0.08d Ile 0.14 ± 0.01a 0.51 ± 0.04b 0.16 ± 0.00a 0.59 ± 0.04b 0.16 ± 0.00a 0.56 ± 0.02b Leu 0.22 ± 0.01a 1.35 ± 0.13b 0.23 ± 0.00a 0.11 ± 0.01c 0.26 ± 0.01a 1.70 ± 0.07d Met 0.04 ± 0.02a 0.39 ± 0.03b 0.61 ± 0.01c 0.53 ± 0.03c 0.05 ± 0.00a 0.49 ± 0.02c Phe 0.10 ± 0.01a 0.55 ± 0.04b 0.12 ± 0.01a 0.73 ± 0.04c 0.12 ± 0.01a 0.61 ± 0.03c Tyr ND 0.14 ± 0.01a 0.14 ± 0.01a 0.29 ± 0.03b 0.15 ± 0.01a 0.17 ± 0.01a Val 0.27 ± 0.02a 0.72 ± 0.04b 0.33 ± 0.01a 0.24 ± 0.02a 0.33 ± 0.00a 0.66 ± 0.05b Arg 12.89 ± 1.71a 3.66 ± 0.51b 13.63 ± 0.28c 7.71 ± 0.04c 13.66 ± 0.06d 5.93 ± 0.20e Cys 1.26 ± 0.20a 0.32 ± 0.07b 1.45 ± 0.05c 1.43 ± 0.05c 1.49 ± 0.03c 1.34 ± 0.04a Glu 7.65 ± 0.99a 2.57 ± 0.49b 8.62 ± 0.27c 5.37 ± 0.39d 7.57 ± 0.04a 4.37 ± 0.15e His 0.79 ± 0.11a 0.40 ± 0.05b 0.93 ± 0.02a 0.53 ± 0.02c 0.93 ± 0.01a 0.44 ± 0.05b Ser 0.96 ± 0.16a 0.40 ± 0.12b 1.08 ± 0.02a 0.67 ± 0.09c 1.08 ± 0.02a 0.40 ± 0.09b Lys 0.55 ± 0.07a 1.01 ± 0.05b 0.61± 0.00c 1.48 ± 0.02d 0.62 ± 0.01c 1.34 ± 0.04d Pro 0.28 ± 0.04a 0.22 ± 0.01a 0.29 ± 0.01a 0.33 ± 0.02ab 0.37 ± 0.01b 0.28 ± 0.01a Thr 0.98 ± 0.13a 0.28 ± 0.01b 1.05 ± 0.02a 1.15 ± 0.02c 1.04 ± 0.00a 0.99 ± 0.05a Total 28.84 ± 3.66a 17.59 ± 1.79b 31.83 ± 0.71a 27.43 ± 0.90a 30.97 ± 0.20a 24.86 ± 1.13c Mean values and standard deviations with different letters (a, b, c) in the same column indicate significant differences (P

during the stages of ripening as suggested by 

Hughes et al. (2002) and Ordonez et al. (1999). 

The changes of free amino acid contents 

represented the degradation of protein and the 

conversion of these free amino acids to the other 

compounds such as biogenic amines and flavours 

as well as growth metabolism of the 

microorganisms. 

In conclusion, this study determined the 

effect of the starter cultures on biogenic amine 

formation in fermented Nham sausages. In 

addition to these, amino acid contents were 

analyzed to note changes of amino acids in Nham 

sausages. The starter culture, L. sakei BCC102, 

decreased pH quickly and suppressed the 

accumulations of tyramine. To avoid the formation 

of high concentration of biogenic amine in Nham, 

it is advisable to inoculate starter culture with 

negative-decarboxylate activity such as L. sakei 

BCC102 and use to top-quality raw meat materials 

for the manufactured food products. 

CONCLUSIONS 

The production of biogenic amines is 

dependent of several variables, such as the growth 

of the microorganisms, their proteolytic and 

decarboxylase activities, which interact with each 

other. Furthermore, there is not a univocal rule 

linking these variables with the different metabolic 

mechanisms necessary for the formation of 

biogenic amines. The results indicated that 

inoculation of starter cultures with decarboxylase 

negative should be carried out to initiate 

fermentation process. Inoculation with appropriate 

starter may lead to the decrease of biogenic amine 

as fermentation progressed. This study suggests 

that the use of L. sakei as starter culture was 

effective to reduce the accumulation of biogenic 

amine; cadaverine, during the ripening of 

fermented Nham. 



This study was supported by a grant of 

the National Center for Genetic Engineering and 

Biotechnology (BIOTEC), Thailand. 


Benjakul, S. and M. T. Morrissey. 1997. Protein 

hydrolysates from Pacific whiting solid 

wastes. J. Agric. Food Chem. 45: 3423-3430. 

Bover-Cid, S., S. Schoppen, M. Izquierdo-Pulido 

and M. C. Vidal-Carou. 1999. Relationship 

between biogenic amine contents and the size 

of dry fermented sausages. Meat Sci. 51: 305- 

311. 

Bover-Cid, S., M. Izquierdo-Pulido and M. C. 

Vidal-Carou. 2001a. Effectiveness of a 

Lactobacillus sakei starter culture in the 

reduction of biogenic amine accumulation as 

a function of the raw material quality. J. Food 

Protect. 64: 367-373. 

Bover-Cid, S., M. Izquierdo-Pulido and M. C. 

Vidal-Carou. 2001b. Changes in biogenic 

amine and polyamine contents in slightly 

fermented sausages manufactured with and 

without sugar. Meat Sci. 57: 215-221. 

Buckenhuskes, H. J. 1993. Selection criteria for 

lactic acid bacteria to be used as starter 

cultures for various food commodities. FEMS 

Microbiol. Rev. 12: 253-272. 

DeKetelaere, A., D. I. Demeyer, P. Vanderkerhove 

and I. Verveake. 1974. Stoichiometry of 

carbohydrate fermentation during dry sausage 

ripening. J. Food Sci. 39: 297-300. 

Dura, M. A., M. Flores and F. Toldra. 2004. Effect 

of growth phase and dry-cured sausage 

processing conditions on Debaryomyces spp. 

generation of volatile compounds from 

branched-chain amino acids. Food Chem. 86: 

391-399. 

Eerola, S., R. Hinkkanen, E. Lindfors and T. Hirvi. 

1994. Liquid chromatographic determination

372 

of biogenic amines in dry sausages. J. Assoc. 

Off. Anal. Chem.. Int. 76: 575-577. 

Eitenmiller, R. R., P. E. Koehler and J. O. Reagan. 

1978. Tyramine in fermented sausages: factors 

affecting formation of tyramine and tyrosine 

decarboxylase. J. Food Sci. 43: 689–693. 

Flores, M., M. C. Aristoy and F. Toldra. 1996. 

Biogenic polyamines affect activity of 

aminopeptidase B and alanyl aminopeptidase 

from porcine skeletal muscle. J. Food Sci. 61: 

13-27. 

Gardini, F., M. Martuscelli, M. A. Crudele, A. 

Paparella and G. Suzzi. 2002. Use of 

Staphylococcus xylosus as a starter culture in 

dried sausages: effect on the biogenic amines 

content. Meat Sci. 61: 275-283. 

Halasz, A., A. Barath, L. Simon-Sakardi and W. 

Holzapfel. 1994. Biogenic amines and their 

production by microorganisms in food. 

Trends Food Sci. Technol. 5: 42-49. 

Hernandez-Jover, T., M. Izquierdo-Pulido, M. T. 

Veciana-Nogues, A. Marin_e-Font and M. C. 

Vidal-Carou. 1997. Biogenic amines and 

polyamine contents in meat and meat 

products. J. Agric. Food Chem. 45: 

2098–2102. 

Hughes, M. C., J. P. Kerry, E. K. Arendt, P. M. 

Kenneally, P. L. H. McSweeney and E. E. 

O’Neill. 2002. Characterization of proteolysis 

during the ripening of semi-dry fermented 

sausages. Meat Sci. 62: 205-216. 


Khieokhachee, T., W. Praphailong, C. 

Chowvalitnitithum, S. Kunawasen, S. 

Kumphati, V. Chavasith, S. Bhumiratana and 

R. Valyasevi. 1997. Microbial interaction in 

the fermentation of Thai pork sausage, pp. 

312–318. In Proceedings of the sixth 

ASEAN food conference , November 24–27, 

1997. Singapore. 

Maijala, R., E. Nurmi and A. Fischer. 1995. 

Influence of processing temperature on the 

formation of biogenic amines in dry sausages. 

Meat Sci. 39: 9-22. 

Ordonez, J. A., E. M. Hierro, J. M. Bruna and L. 

de la Hoz. 1999. Changes in the components 

of dry-fermented sausages during ripening. 

Crit. Rev. Food Sci. Nutr. 39: 329-367. 

Rowan, N. J., J. G. Anderson and J. E. Smith. 1998. 

Potential Infective and Toxic 

Microbiological Hazards Associated with 

the Consumption of Fermented Food. In B. 

J. B. Wood (ed.), Microbiology of fermented 

foods. Blackie Academic and Professional, 

London. 

Scanlan, R.A. 1983. Formation and occurrence of 

nitrosamines in foods. Cancer Res. 43 

(Suppl.): 2435–2440. 

Vidal-Carou, M. C., M. Izquierdo-Pulido, M. C. 

Mart ′ yn-Morro and A. Marine-Font. 1990. 

Histamine and tyramine in meat products: 

relationship with meat spoilage. Food Chem. 

37: 239-249.


Product Development System in Pattern Construction System, 

Standard Body Measurement and Suitable Fitting Allowance for 

Thai Ladies Brand in Fashion Industry 

ABSTRACT 

Foengfurad Mungtavesinsuk 

Concept of the brand, theme of the design is the spirit of the collection in the fashion branding. 

But the Pattern Construction System, standard body measurement and suitable fitting allowance are the 

sustainable part for the branding in the market. 

The study found that most brands in Thailand did not correct and less detailed, about body 

measurement, standard sizing with appropriate fitting allowance for pattern construction and the pattern 

construction system. 

The objectives of this research, firstly the author used the Germany Pattern System and 

appropriate German standard body measurements to make pattern construction. The results showed that 

humans with different figures (and it does not matter in which country), the body type selected and the 

size range of body measurement are almost similar. 

Secondary, German Standard fitting allowance was applied to the Thai fashion industry. The 

results showed that the tight fit should be used for the first or second step (of fitting allowance) and the 

blouse item in second or third step of fitting allowance, etc. 

Additionally, the study found that through systemization, productivity increased and the cost 

of the product development was reduced. 

Key words: pattern construction, body measurement, fitting, fashion brand, fashion industry 

INTRODUCTION 

The fashion business is an exciting, 

stimulating, fascinating, ever changing, never the 

same. In fashion, as in everything else, there are 

always ups and downs, stops and starts. The 

movement of fashion is always forward, never 

backward. Its movements depend on the 

environment. From the designer to the consumer, 

everyone involved in the movement of fashion. 

As a business, fashion was once 

considered an art form controlled by designers who 

dictated its content. But fashion has now evolved 

into a science that can be measured and evaluated. 

Modern fashion manufacturing was born during 

the industrial revolution and has matured in the 

age of technology. Without machines, clothing 

could never be mass-produced. Technology has 

revolutionized the way fashion is made. Almost 

all stages of clothing production from design to 

delivery rely to same extent on technology. (Stone, 

1990) 

Department of Textile Science and Technology, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand. 

e-mail: ofrm@ku.ac.th 


374 

A fashion retailer is in the business of 

selling fashion products and not art. The modem 

merchandiser is able to plan to supply very unique 

local demand patterns. For example, people in 

different sized, immigration into local areas can 

often fundamentally change the sizing patterns 

required in a local shop. 

For this reasons, the product 

development process needs to have fitting sample. 

The fitting sample is checked on models that are 

the “base size” (a medium or a size 12 in women’s 

wear). Most fashion retailers have a limit to the 

number of fit sample amendments they will accept, 

the comment being up to three. After this, the style 

is at risk of being cancelled. (Jackson et al., 2001) 

The first pattern is important in the 

process and has to be carefully and methodically 

produced. And pieces need to match in the right 

places. (Shreeve et al., 2004) 

Bangkok Fashion city project guide the 

Thai Fashion Industry into the global market. The 

global market is highly competitive and needs 

really professional and knowledgeable team. 

As the experience in the global fashion 

business , business will suffer, no matter how good 

the design is or the accountants in the back office, 

without the right goods they will be not be able to 

generate enough sales, and ultimately enough 

profit . That means, fashion business needs the 

whole, healthy team and the product development 

is part of it. 

In the Fashion Industry, product 

development covers the material and sampling. It 

needs very strong knowledge about apparel 

technology and management. And for the 

sampling, it needs detail of styles with correct 

information about fitting allowance from designer 

and request standards body measurement from the 

item for the first sample. 

Thailand has limited information about 

the standard body measurement and sizing. The 

fitting allowance for pattern construction system 

is mostly by experience but not methodical 


workers. So, how and what can the fashion brands 

in Thailand fashion industry do? 

In this research, the standard body 

measurement used was from Hohenstein Institue 

Germany; the fitting step and pattern methodology 

used was from University of Applied Science 

Niederrhein Germany; applied to Thai Fashion 

Industry. Through this research, the most 

appropriate body measurement, sizing, fitting 

allowance and suitable pattern system for the Thai 

fashion industry was found especially for ODM 

(Original Design Manufacture) in Ladies Fashion. 


Document and samples data 

Hohenstein Institute made the research 

about body measurement and sizing for the 

Germany and EU people, and set up standards for 

body measurements in different figure groups. This 

basic data is used by us as reference in standard 

body measurement. The fitting step and pattern 

methodology used as reference came from the 

document of University of Applied Science 

Niederrhein Germany. 

The target group for this case study is 

from companies in the fashion industry with local 

ladies in Thai market. All together, 25 companies 

with 28 Ladies outwear brandings joined the 

research as a case study. For full scale period of 

the fashion collection and market feed back, a long 

term study is needed to repeat the process and get 

the correct result. So the 25 companies were 

divided into three groups and three phases, each 

phase running for one year and the process of each 

group was the same. The first phase: 5 companies 

as pilot group, second phase: 10 companies 

repeated the process and the third phase: 10 

companies repeated the process and to get 

confirmation of the results, it was done by the first 

two phases. 

The process design

As the 25 companies were divided into 

three phases for the case study, each group 

followed the process and the methodology to get 

the data and the result as standardized and 

systemized. We analyzed the problems of the 

existing products by fitting samples in the body 

measurement, fitting allowance and pattern 

construction system from the 28 brands and design 

process for this study. 

Firstly, we analyzed the suitable sample 

size and tried the standard size 38 and size19 for 

medium size. 

Secondly, as this is the rainy season in 

Thailand we used the fitting steps 1-5 as reference 

for the fitting allowance in different items of 

product. As the standards: step 1-2 for the tight fit 

or tank top, step 2-3 for the blouse and step 3-5 

for jacket, all the fitting allowance data are with 

percentage by calculation instead of by experience 

data. 

Thirdly, we took the medium size body 

measurement with suitable fitting step to calculate 

and apply it to the pattern construction in German 

methodically system. 

All three processes were transferred to 

the product development and made the samples 

by each brand, then we did the sample fitting to 

analysis the data and the methodology. After first 

sample fitting, we corrected or adjusted the 

reference data necessary, we remade the samples 

and rechecked it again to fix the standards body 

measurement, fitting step and construction 

methodology for each brand. 

As soon as the standard body 

measurement, sizing, fitting step and pattern 

construction methodology is confirmed, the 

product development section of the brand done in 

the collection will be put into market for sale. 

When the feed back from market are good that 

means the system is going in the correct direction 

then the standardization of the product 

development system is fixed and each brand can 

set up the standard basic block of pattern for each 


season following the fashion trend. 


After three years study, we grouped the 

problems from 28 different Thai Ladies Brands as 

found out in this research and divided them into 

three parts to show the result and for discussion 

1. Standard body measurement and 

sizing 

2. Fitting allowance in different steps 

3. Pattern construction system 

Standard body measurement and sizing 

After three years study, we grouped the 

28 different Thai Ladies Brands by market segment 

and items of product, and then found out the 

results: the young generation group is mostly fixed, 

the sample size or medium size in size 38 and the 

older generations mostly are in size 19. This also 

shows the human body development of the 

different generations and the development of the 

social environment. 

Thailand has very limited information 

about body measurement; in cases there are some 

but still not the full scale of body measurement 

for the pattern construction. 

In Germany, research of standard body 

size specification is made every 10 years and 

divided size group as the figure in normal high 

group around 168 cm, short group around 160 cm 

and extra high group around 176 cm. 

(Mungtavesinsuk, 2005) As the grouping 

compared to Thai peoples figure we can use the 

normal and short group to apply in the Thailand 

market. 

Group, in normal high will take size 38 

as standard medium size for sample fitting and in 

short group will take size 19 as standard medium 

size for sample fitting. After the size is selected, 

detail of the body measurement placed in the size 

table will be used for pattern construction. Through 

all three phases as market segment for carrier

376 

women will be in normal group and is fixed with 

the Germany body size, in old generation group 

will be in German size 19 but need shorter in back 

waist length and in young attitude will be in 

German size 18. As soon as the results come up, 

all the sample making will follow the standards 

body measurement and sizing for branding. 

Fitting allowance in different step 

Most in the local brand during product 

development has not fixed the standard fitting 

allowance by percentage but with experience data 

added into the finishing garment measurement. 

Those experience data to make the samples, can 

be good for this sample and this size but not sure 

for next sample. That means the reprocess in 

product development are uncountable and the cost 

of product development is higher. In the study, we 

were gave the fitting allowance in German system 

as reference for the pattern construction to made 

the first sample. After the first fitting, maybe some 

have a bit adjustment but mostly almost fix as 

request. Through the try out, we set up the suitable 

fitting step for each item and each brand to make 

the pattern construction and then get the standard 

basic block for whole collection. It means, during 


the product development, there should be the 

standards medium size and the fitting allowance 

should be fixed in same level for the same item in 

same collection and it depends on market request 

and fashion trend too. 

For Thai local market we have only 

summer item, so the fitting allowance do not need 

the whole range from step 1 to step 7. We just 

need the fitting allowance from step 1 to step 5. 

That means the garment is more fit on bodies. 

Pattern construction system 

In the fashion industry, the first sample 

is very important. But it needs the most correct 

information for pattern construction. As the 

standards body measurement, sizing, and the step 

of fitting allowance are fixed, it should get the best 

fit sample. But why it still has problems in the 

sample fitting? 

Using the German pattern construction 

system, formula is calculated methodically. All 

formula is based on the standard body 

measurement with fitting allowance step to 

calculate in percentage. The pattern construction 

actually needs very strong mathematical back 

ground especially nowadays with the computer. 

Table 1 The German standard sizes 38 as medium size apply into Thai local market in medium size as 

38, 19, or 18. 

Group S M L XL 

Normal – carrier women 36 38 40 42 

Short – old generation 18 19 20 21 

Short – young generation 17 18 19 20 

Note: S = Small size; M = Medium size; L= Large size; XL = Extra large 

Table 2 Fitting allowance for different items such as body suit, blouse, jacket, etc. 

Fitting allowance in step Standard German allowance Thailand and new fashion trend 

(add % in chest) 

1 st step (6%) For body suit For tight fit knit wear 

2 nd step (9%) For tight fit knit wear For tight fit blouse or with elasticity 

3 rd step (12%) For blouse and shirt For lose blouse 

4 th step (15%) For tight fit jacket or suit For suit some with 3 1 / 2 step 

5 th step (18%) For lose fit jacket or suit For Jacket some with 4 th step

All the data comes up with informative system. 

But before computerizing the correct pattern 

construction system is needed otherwise the 

sample making still has problems. 

During research it was found that all 

points which happen frequently is mostly from 

basic pattern construction problems. This means 

the fundamental pattern construction system has 

problems. 

During the study most problems in 

trousers were with wrinkles in the crotch position 

and leg twist. The top items: center front rides up 

and center backs are too loose. Those problems 

are all from the basic pattern construction system, 

Figure 1 The problem of crotch in trousers. 

Figure 2 The problem with leg twist. 


which the pattern for samples are by experienced 

but not systemized. 

Wrinkle in crotch of trousers 

In Figure 1: the sample piece is in “A” 

and the standard one is in “B”. We can see the 

width in crotch position especially in part “A” is 

too narrow as part “B”. That means the proportion 

in crotch position in back rise should be 1/8 of 

Hip circumference of body measurement. 

Leg twist in trousers 

In Figure 2 the standard piece is in “A” 

and the sample piece is in “B”. The middle line

378 

for the trousers must be in the middle, like back 

piece “A” the width of leg must be b1=b2, b3=b4 

and front piece”B”b5=b6, b7=b8, but in back piece 

“C” t1>t2, t3>t4 and front piece “D” t5

as piece “A” or straight as piece “C”. We can see 

the center part in pieces “B” & “D” that the center 

back line is not straight in follow the body figure 

like piece “C” but in straight down like pieces 

“B” & “D”. That means the pieces “B” & “D” are 

wider then pieces “A” & “C”, this cause too loose 

in center back. 

CONCLUSION 

Systemization is the successful key for 

the fashion industry, it does not matter in which 

section, even the product development. As the 

result we found through this research that we need 

to have a standard body measurement, standard 

fitting allowance step and systemization of pattern 

construction. The human figure actually can be 

dividing into different groups and different sizes. 

As soon as the data base is there, that can be 

applied to almost everybody’s figure. Thailand 

does not have the standard body measurement. But 

the competitiveness and the chance would not wait. 

The best solution for the Thai ladies fashion brands 

is to follow the German Pattern construction 

system as base to get the best of the technology 

for product development. Through this research 

we also confirm the system from Germany is 

workable for Thailand too. Everyone who is 

concerned about product development for the 

fashion industry must follow this format to running 

the fashion business. 


Additionally, as soon as the product 

development systems are running smoothly, then 

quick response and cost goes down which is 

happening for the fashion business too, because 

everyone is working in the same direction and 

speaks the same language. The high technology is 

introduced to the fashion industry. With strong 

fundamental knowledge will bring high quality 

through high technology. 

ACKNOWLEDGEMENT 

The author would like to thanks the 

Department of Industry Promotion from Industry 

Ministry to support this research and the 25 

companies cooperated to get the valuable data for 

this research. 


Mungtavesinsuk, F. 2005. Industrial Pattern 

Construction of Lady’s wears in German 

system. 2 nd ed. Kasetsart University. 

Bangkok, Thailand. 115p. 

Shreeve, A. and C. Kelly. 2004. Developing 

communication skills through fashion 

design. IFFTI 2004: 51 – 63 

Stone, E . 1990. Fashion Merchandising. 5 th ed. 

Merchandising management. MACMILLAN 

PRESS LTD. London. 454p.


A Nonlinear Optimization Problem for Determining Safety Stocks 

in a Two-Stage Manufacturing System 

ABSTRACT 

Parthana Parthanadee 

Safety stock is the inventory which is used to buffer against the uncertainties in business 

operations. Managers must decide how much safety stock of each raw material and each finished product 

should be maintained. Determining appropriate safety stock levels is an important decision. Too much 

safety stock would incur extra inventory carrying costs, whereas too less safety stock would increase 

the risk of having product stockouts and lost sales. In this paper, a nonlinear programming problem for 

determining safety stock levels in a two-stage manufacturing system, was presented. Instead of using 

the well-known search algorithms, simple decision rules for determining safety stock levels were derived 

from an analysis of the derivatives of cost functions, with respect to the delivery performances of suppliers 

and prior manufacturing process. Two algorithms based on those decision rules were proposed and 

tested on seventy-five problem instances. The results showed that the proposed algorithms provided, 

within 1 second, the solutions with less than 3% deviations, on average, from the known integer solutions 

or the best lower bounds. The algorithms also performed better than the pattern search algorithm, which 

was the method applied in the previous research. 

Key words: safety stock, inventory, nonlinear programming problem, two-stage manufacturing systems 

INTRODUCTION 

Safety stock or buffer stock is the amount 

of inventory held in a short run to protect against 

demand and supply uncertainties and forecasting 

errors in business operations. When demands are 

underestimated, or supplies are insufficient or 

backordered, product stockouts may occur and 

cause the company some lost sales, especially 

when the degree of product substitutability is high. 

On the other hand, if too many safety stock 

quantities are held, high inventory costs would be 

charged to the company. The two types of costs: 

opportunity costs and inventory costs must be 

traded off to find the appropriate safety stock 

levels. 

The classical approach for determining 

safety stock is to specify a desired service level or 

a stockout probability and use it to identify a safety 

factor, k. If the demand during lead time is assumed 

normally distributed, the safety factor is usually 

set to z and the safety stock is set to z⋅σ L, where z 

denotes the z-score to achieve the desired service 

level and σ L denotes the standard deviation of the 

probability distribution of demand during lead time 

(Vollmann et al., 1997). The other choices of safety 

factor, demand deviation, and safety stock 

calculations can be found in Krupp (1997); Silver 

Program of Agro-Industry Technology Management, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand. 

e-mail: fagiptp@ku.ac.th 


et al. (1998); Zeng (2000); and Talluri et al. (2004). 

Maia and Qassim (1999) derived 

optimum safety stocks for a one-stage 

manufacturing system, in which a finished product 

was produced from a number of raw materials. 

The problem was formulated as a nonlinear 

program (NLP), which minimized the total of 

inventory and opportunity costs. From the analysis, 

Maia and Qassim (1999) found that it was 

economical to either hold every safety stock at its 

maximum level or not hold it at all. A set of 

decision rules for finding optimum safety stocks 

was provided and illustrated through a small 

numerical example. 

Siribanluoewut (2006) extended the 

work by Maia and Qassim (1999) to determine 

safety stocks for a two-stage manufacturing 

system. The problem was solved using three 

optimization heuristics, which were genetic 

algorithm, pattern search algorithm, and the hybrid 

genetic algorithm with pattern search. All the 

optimization heuristics performed efficiently on 

the test problems and the qualities of solutions 

reported were found not statistically different from 

each other. However, the pattern search algorithm 

provided good solutions in significantly shorter 

time than other heuristics did. 

Inderfurth and Minner (1998) formulated 

an optimization problem of determining safety 

stocks in multi-stage manufacturing systems with 

normally distributed demands. The system was 

assumed to be under a periodic review, base-stock 

control policy, in which inventories were reviewed 

every fixed period of time and replenished up to a 

specified level. The safety factor in this study was 

found to be depending on service level, type of 

service level, and coverage time. The service level 

and coverage time for different types of multi-stage 

manufacturing systems were derived to establish 

the optimal policy for determining safety stocks 

in these multi-stage systems. 

In this paper, the problem for determining 

safety stocks in the two-stage manufacturing 

system, as presented in Siribanluoewut (2006), was 


considered. Instead of using the optimization 

heuristics, which required the users to understand 

their mechanisms, a set of simple decision rules 

for finding optimum safety stocks was developed, 

and tested on the number of test instances as shown 

in the following sections. 


Problem description 

A two-stage manufacturing system, as 

presented in Siribanluoewut (2006), was 

considered in this study. In such system, a 

manufacturer ordered m raw materials (RMs) for 

its stage-1 manufacturing process and n raw 

materials for its stage-2 manufacturing process. 

Each raw material was ordered from a single 

supplier. The stage-1 process produced a workin-process 

(WIP) from those m raw materials. The 

WIP and the n other raw materials were then fed 

to stage 2 to produce a final product. Figure 1 

illustrated this two-stage manufacturing system. 

The model formulation of this system was 

modified from that of the one-stage manufacturing 

system by Maia and Qassim (1999). The notations 

used in the formulation were as follows. 

Stage 1 

i index of raw materials in stage 1; i = {1, 

2,…, m} 

p1,i the on-time delivery performance of 

supplier i 

* 

q1, i 

q 1,i 

x1,i k1,i c 1,i 

p s1 

qw xw the quantity of stage-1 raw material i that 

is delivered on time 

the quantity of stage-1 raw material i that 

is ordered 

the safety stock of stage-1 raw material i 

the delivery performance to manufacture 

of stage-1 raw material i 

the unit inventory cost of stage-1 raw 

material i 

the stage-1 manufacturing performance 

the quantity of WIP that is required 

the safety stock of WIP

382 

Figure 1 The two-stage manufacturing system. 

k w 

c w 

the WIP delivery performance to stage- 

2 manufacturing process 

the unit inventory cost of WIP 

Stage 2 

j index of raw materials in stage 2; j = {1, 

2,…, n} 

p2,j the on-time delivery performance of 

supplier j 

* 

q2, j 

q 2,j 

x2,j k2,j the quantity of stage-2 raw material j that 

is delivered on time 

the quantity of stage-2 raw material j that 

is ordered 

the safety stock of stage-2 raw material j 

the delivery performance to manufacture 

of stage-2 raw material j 


c 2,j 

p s2 

the unit inventory cost of stage-2 raw 

material j 

qp the stage-2 manufacturing performance 

the quantity of finished product that is 

required 

xp the safety stock of finished product 

kp The finished product delivery 

cp performance to customer 

the unit inventory cost of finished 

product 

co the unit opportunity cost of finished 

product that is not delivered on time 

As in Maia and Qassim (1999), the ontime 

delivery performance of supplier i and the 

on-time delivery performance of supplier j could 

be calculated from the past data records, using 

Equations (1) and (2), respectively.

p 

1, 

i 

* 


q1, 

i 

= (1) 

q 

1, 

i 

* 

q2, 

i 

p2, 

i = (2) 

q2, 

i 

If the manufacturer held safety stocks for every raw material, the delivery performances to 

manufacture of stage-1 raw material i and stage-2 raw material j could be defined as in Equations (3) 

and (4), respectively. 

k 

1, 

i 

* 

1i 1, 

i 

1, 

i 

q , + x 

= 

q 

x1, 

i 

= p1, 

i + 

q 

∀ i = {, 12K , , m} 

1, 

i 

* 

q2, j + x2, 

j x2, 

j 

k2, 

j = = p2, 

j + 

q2, 

j 

q2, 

j 

∀ j = {, 12K , , n} 

(4) 

The inventory cost of the safety stock of each raw material could be computed from Equations 

(5) or (6) as follows. 

C1, i = c1, ix1, i = c1, iq1, i( k1, i − p1, 

i) 

∀ i = {, 12K , , m} 

(5) 

C2, j = c2, jx2, j = c2, jq2, j( k2, j − p2, 

j) 

∀ j = {, 12K , , n} 

(6) 

The manufacturing performance of stage-1 process, ps1 , was defined as the ratio between ontime 

and planned production, accounting for all delays that may occur, but excluding those caused by 

material stockouts. The ps1 could be found from Equation (7). The WIP delivery performance to stage- 

2 manufacturing process, kw, was given in Equation (8). 

p 

s1 

= 

q 

q 

* 

w 

w 

m 

x 

kw = ps k i + w 

1∏ 

1, 

i= 

1 q 

(8) 

w 

The inventory cost of the WIP safety stock could be calculated from Equation (9). 

m 

Cw = cwxw = cwqw( kw − ps ∏ k i) 

1 1, 

(9) 

i= 

1 

Similarly, the manufacturing performance of stage-2 process, ps2 , the product delivery 

performance, kp, and the product inventory cost could be calculated as follows. 

p 

s2 

= 

q 

q 

* 

p 

p 

x 

kp = ps kw k j + 

2 ∏ 2 

q 

n 

j= 

1 

(3) 

(7) 

(10) 

p 

, 

p 

(11) 

C = c x = c q ( k − p k k ) 

p p p p p p s2w 2, 

j 

j= 

1 

n 

∏ 

(12)

384 


Finally, the opportunity cost, defined as the cost incurring whenever the finished product failed 

to be delivered to the customers on time, was given in Equation (13). 

C = c q ( 1 −k 

) 

(13) 

o o p p 

Mathematical model 

A nonlinear programming (NLP) model, for determining the delivery performances k 1,i, k w, 

k 2,i, and k p was formulated in this section. The objective of this NLP model was to minimize the total of 

the inventory costs charged for holding all the safety stocks and the opportunity costs, subject to the 

bounds on the delivery performances. The model was formulated as follows. 

Subject to 

m 

⎛ m ⎞ 

Min C=coqp( 1 − kp)+ ∑ c1, iq1, 

i( k1, i − p1, i)+ cwpw⎜kw − ps k 

1∏ 

1, 

i⎟ 

i= 

1 

⎝ i= 

1 ⎠ 

n 

⎛ 

n ⎞ 

+ ∑ c2, jq2, 

j( k2, j − p2, j)+ cpqp⎜kp − ps kw∏k 2 2, 

j⎟ 

(14) 

j= 

1 

⎝ 

j= 

1 ⎠ 

p1, i ≤ k1, 

i ≤1 ∀ i = {, 12K , , m} 

(15) 

p2, j ≤ k2, 

j ≤1 ∀ j = {, 12K , , n} 

(16) 

m 

ps ∏ k1, i ≤ kw≤ 

1 

(17) 

1 

i= 

1 

n 

ps kw∏k2, j ≤ kp≤ 

1 

(18) 

2 

j= 

1 

Solution analysis 

It was known that the optimal solution of the NLP is necessarily on the border of the feasible 

region, if the Hessian matrix of the objective function is indefinite, as in this problem (see Marsden and 

Tromba (1981), for example). Therefore, the optimal delivery performances k1,i, kw, k2,j, and kp in the 

presented NLP must be either on their lower bounds or upper bounds. In this paper, the analysis followed 

* 

the method in Maia and Qassim (1999) by defining reference costs, c1, i for the stage-1 raw material 

* * 

i, cw for the WIP, and c2, j for the stage-2 raw material j, as shown in Equations (19) - (21). The upper 

bounds of these reference costs were found from the derivatives of the cost function with respect to the 

delivery performances k1,i, kw, k2,j, and k2,j. c 

c1iq1i ≤ 

m 

q p p 

* 

1, 

i 

, , 

w s1 ∏ 

i2=+ i 1 

1, 

i2 

c 

* 

w 

c q 

≤ w w 

n 

q p p 

∏ 

p s2 2, 

j 

j= 

1 

∀ i = {, 12K , , m} 

(19) 

(20)

c 

≤ 

c2 jq2 j 

n 

q p k p 

* 

2, 

j 

, , 

p s2 w ∏ 

j2= j+ 

1 

2, 

j2 


∀ j = {, 12K , , n} 

(21) 

* * * 

The reference costs, c1, i cw and c2, j were then analyzed against all the unit costs in the model 

to identify when the corresponding delivery performances and safety stocks should be set to their lower 

or upper bounds. If the opportunity cost was high, the manufacturer should hold safety stocks to prevent 

the products shortages. In contrary, it would not be economical to stock the materials, when the inventory 

costs (and hence the reference costs) were costly. The optimal solution of the presented optimization 

model could be derived as follows: 

Stage-1 raw materials: 

⎧ * 

c then k 1 1 and x 

* ⎪ o ≤ c1, i ,i = p ,i 1,i 

= 0 

(i) If c 1,i 

≤ min( cw, cp) 

and ⎨ * 

⎩⎪ c o > c1, i then k 1,i = 1and x1,i 

= q1,i 1− 

p1,i 

* 

(ii) If c 1,i > min( cw, cp) then k 1,i = p1,i 

and x1,i 

= 0 

Work-in-process: 

⎧ 

m 

* 

⎪c 

o ≤ cw then kw = ps 

∏ k 1,i 

and xw 

= 0 

1 

* ⎪ 

i= 1 

(iii) If c w ≤ cp 

and ⎨ 

m 

⎪ * 

⎛ 

⎞ 

⎪ 

c o > cw then k w = 1and xw = qw⎜1− ps 

k 

1∏ 

1,i⎟ 

⎩ 

⎝ i= 1 ⎠ 

m 

* 

(iv) If c w > cp then kw = ps 

∏ k 1,i 

and xw 

= 0 

1 

i= 1 

Stage-2 raw materials: 

* 

⎧c 

then k and x 

* ⎪ o ≤ c2, j 2, j = p2, 

j 2, 

j = 0 

(v) If c 2, 

j ≤ cp 

and ⎨ * 

c o > c , then k = 1and x = q 1− 

p 

⎩ 

⎪ 

( ) 

2 j 2, j 2, j 2, j 2, 

j 

(vi) 

* 

If c 2, j > cp then k 2, j = p2, 

j and x2, 

j = 0 

Finished product: 

(vii) 

n 

c o > cp then kp = ps 

kw∏k j and x p = 

2 2, 0 

j= 1 

⎛ 

n ⎞ 

(viii) c o > cp then k p = 1and xp = qp⎜1− ps kw∏k 

2 2, 

j⎟ 

⎝ 

j= 1 ⎠ 

( )

386 

Proposed algorithms 

Since the exact values of the reference 

costs were not known, they could be initially set 

to their upper bounds in which all other rawmaterial 

delivery performances, besides the one 

corresponding to the considered raw material, (k 1,i 

: ∀i ≠ i’ and k 2,j : ∀j ≠ j’) were set at their lower 

bounds. The estimated reference costs of the raw 

materials in every stage were sorted in a nondecreasing 

order and the values were recalculated 

as in Equations (19) and (21). This solution finding 

algorithm was specified as Algorithm 1. 

From the preliminary testing, it was 

found that when the estimated values of reference 

costs were not much different from each other or 

from the opportunity cost, Algorithm 1 may not 

always provide the optimal solutions. Algorithm 

2 was then proposed. Again, the reference costs 

of the raw materials in every stage were sorted as 

in Algorithm 1. At the initial step, the delivery 

performances and safety stocks of all raw materials 

were set to their lower bounds. The delivery 

performances and safety stocks of WIP and 

finished product were found from the decision 

rules presented in the previous section, 

accordingly. The total cost was calculated and 

recorded. Then, the delivery performance and 

safety stock of each raw material in each stage 

were increased to their upper bounds, one by one, 

corresponding to the non-decreasing order of the 


raw-material reference costs. The delivery 

performances and safety stocks of WIP and 

finished product, including the total costs, were 

recalculated and recorded at every step. Finally, 

the minimum total cost and the best solution were 

identified. 

A numerical example 

In this section, a small example, 

consisting of three raw materials in stage 1 and 

two raw materials in stage 2, was presented. The 

data for this example was given in Table 1. The 

opportunity cost was assumed to be 8.44 baht. 

Algorithm 1: 

The initial reference costs for stage-1 raw 

materials 1, 2 and 3 were found to be 2.1778, 

5.4652 and 9.2014 baht, respectively. Thus, the 

stage-1 raw material order followed the natural 

order. The reference costs for RM 1, RM 2 and 

RM 3 were recalculated and their values became 

2.1778, 5.1868 and 8.1043 baht, respectively. 

Following the proposed decision rules, the safety 

stocks of RM 1 and RM 2 should be set to their 

upper bounds, which were 8 and 10 units, 

respectively. The safety stock for RM 3 and WIP 

were found unnecessary. 

Next, the initial reference costs for stage- 

2 raw materials 4 and 5 were found to be 7.5490 

and 5.5337 baht, respectively. Hence, the 

Table 1 Data for a small example with three raw materials in stage 1 and two raw materials in stage 2. 

Materials q q* p c Initial ref. Ref. cost Algorithm 1 Algorithm 2 

cost (baht) (baht) k x k x 

RM 1 157 149 0.9490 2.86 2.1778 2.1778 1.0000 8 1.0000 8 

RM 2 139 129 0.9281 8.29 5.4652 5.1868 1.0000 10 0.9281 0 

RM 3 244 242 0.9918 7.44 9.2014 8.1043 0.9918 0 0.9918 0 

WIP 232 224 0.9655 9.40 - 16.5918 0.9576 0 0.8887 0 

RM 4 117 107 0.9145 8.88 7.5490 6.9549 1.0000 10 0.9145 0 

RM 5 216 199 0.9213 3.50 5.5337 5.5337 1.0000 17 1.0000 17 

Product 173 156 0.9017 7.20 - - 1.0000 23.61 1.0000 46.21 

Total cost 424.1001 415.0962 

(baht)

algorithm would consider RM 5, prior to RM 4. 

The reference costs of RM 4 and RM 5 were 

recalculated and found to be 6.9549 and 5.5337 

baht. Thus, the safety stocks of RM 4 and RM 5 

were set to their upper bounds, which are 10 and 

17 units, respectively. Finally, the product safety 

stock was computed and set to 23.61 units. The 

corresponding total cost is 424.10 baht. 

Algorithm 2: 

Following the initial reference costs 

found in Algorithm 1, the priority for increasing 

raw-material safety stock levels would be in the 

orders of RM 1 – RM 2 – RM 3 and RM 5 – RM 

4. Twenty-four solutions were evaluated and 


shown in Table 2. From the Table, the sixth 

solution provided the minimum total cost of 415.10 

baht, with the safety stock levels set to 8 units for 

RM 1, 17 units for RM 5, and 46.21 units for the 

finished product. Algorithm 2 provided a superior 

solution to Algorithm 1 for this test instance. 

RESULTS 

To facilitate the implementation, 

Algorithms 1 and 2 were coded in MATLAB® 

6.5. Both algorithms were tested on 75 test 

instances (from 5 test problem sets, each with 15 

instances) in Siribanluoewut (2006). Table 3 

presented structures of the test instances and the 

Table 2 The twenty-four solutions evaluated by Algorithm 2. 

No. Safety Stocks (units) Total cost 

RM 1 RM 2 RM 3 WIP RM 4 RM 5 Product (baht) 

1 0 0 0 0 0 0 62.14 447.42 

2 8 0 0 0 0 0 56.19 427.44 

3 8 10 0 0 0 0 47.13 445.15 

4 8 10 2 0 0 0 46.09 452.54 

5 0 0 0 0 0 17 52.67 438.73 

6 8 0 0 0 0 17 46.21 415.10 

7 8 10 0 0 0 17 36.38 427.23 

8 8 10 2 0 0 17 35.25 433.98 

9 0 0 0 0 10 17 41.43 446.56 

10 8 0 0 0 10 17 34.36 418.58 

11 8 10 0 0 10 17 23.61 424.10 

12 8 10 2 0 10 17 22.38 430.09 

13 0 0 0 36.33 0 0 41.56 640.70 

14 8 0 0 25.82 0 0 41.56 564.82 

15 8 10 0 9.84 0 0 41.56 497.48 

16 8 10 2 8.00 0 0 41.56 495.10 

17 0 0 0 36.33 0 17 30.33 619.36 

18 8 0 0 25.82 0 17 30.33 543.48 

19 8 10 0 9.84 0 17 30.33 476.14 

20 8 10 2 8.00 0 17 30.33 473.76 

21 0 0 0 36.33 10 17 17.00 612.16 

22 8 0 0 25.82 10 17 17.00 536.28 

23 8 10 0 9.84 10 17 17.00 468.94 

24 8 10 2 8.00 10 17 17.00 466.56

388 

average percentage of deviations from the optimal 

NLP total costs, including the solution times, by 

Algorithms 1 and 2. The result showed that 

Algorithm 2 did outperform Algorithm 1. 

As aforementioned, the optimization 

model presented in this paper was an NLP model. 

Therefore, the levels of safety stocks in the final 

solution may be reported as non-integers. This 

safety stock determination problem could also be 

modeled as a mixed integer nonlinear program 

(MINLP) for minimizing the total of the 

opportunity costs and the inventory costs charged 

for holding all the safety stocks, subject to the 

bounds on the safety stock levels. The safety stocks 

x 1,i, x w, x 2,j, and x p, which were required to be 

integers, would be sought from the MINLP, in lieu 

of the delivery performances k 1,i, k w, k 2,j, and k p in 

the NLP. However, the MINLP was a much more 

complex problem. It may not be solved in 

reasonable computation times with regular 

optimization methods, even for small-size problem 

instances. Thus, it was suggested that the safety 

stock levels should be found by applying 

Algorithm 2 and then rounding down the noninteger 

safety stocks to their nearest integers. The 

rounded solutions were compared with true 

optimal integer solutions found from the 

enumeration method, in which all possible integer 

solutions were enumerated and evaluated. 

However, the enumeration method could not be 

implemented on the large problem instances, due 

to their long computation times. Therefore, only 


the true optimal integer solutions of test problem 

sets 1 and 2 could be identified. The qualities of 

these rounded solutions were presented in Tables 

4 and 5. For test problem sets 3, 4 and 5, the 

rounded solutions were compared with the 

corresponding MINLP lower bounds (i.e. the 

optimal NLP solutions) instead. The qualities of 

these solutions were given in Tables 6-8. 

Furthermore, the pattern search algorithm (using 

a complete search, a mesh expansion factor of 1.0 

and a mesh contraction factor of 0.5) was also 

investigated. The details of this algorithm can be 

found in Kolda et al. (2003). The qualities of the 

integer solutions found from the pattern search 

algorithm were also presented in Tables 4-8, for 

comparison purpose. 

From Tables 4 and 5, Algorithm 2 with 

solution rounding provided high-quality results. 

The rounded solutions were 2.10% deviating from 

the known MINLP optimum on average (with a 

maximum deviation of 10.20%) for problem set 

1, and 3.86% deviating from the known MINLP 

optimum on average (with a maximum deviation 

of 20.23%) for problem set 2. Algorithm 2 with 

solution rounding provided the good solutions in 

much shorter times (i.e. less than 1 second) than 

the enumeration method did (i.e. more than 7 

minutes for problem set 1 and more than 35 

minutes for problem set 2, on average). For larger 

test problem sets, the average deviation of the 

Algorithm-2 solutions from the corresponding 

MINLP lower bounds were less than 2.5%, with 

Table 3 Structures of the test instances and the average percentage of deviations from the true optimal 

total costs of Algorithms 1 and 2. 

Set No. of RMs No. of % Deviation from true optimum Average solution time (sec.) 

Stage 1 Stage 2 instances Algorithm 1 Algorithm 2 Algorithm 1 Algorithm 2 

1 3 1 15 0.00% 0.00% 0.0013 0.0047 

2 3 2 15 0.47% 0.00% 0.0013 0.0033 

3 7 2 15 0.24% 0.00% 0.0020 0.0047 

4 12 2 15 0.98% 0.00% 0.0033 0.0073 

5 15 2 15 2.79% 0.00% 0.0013 0.0087 

Average 0.90% 0.00% 0.0019 0.0057


Table 4 The quality of the rounded solutions for test problem set 1. 

No. Enumeration Algorithm 2 Pattern search 

Total Time Total Time % Dev. Total Time % Dev. 

costs (seconds) costs (sec) from costs (sec) from 

(baht) (baht) opt. (baht) opt. 

1 226.2863 557.00 226.2863 0.03 0.00 226.2863 0.656 0.00 

2 223.0773 454.98 245.4778 0.00 10.04 223.3209 0.547 0.11 

3 188.9893 574.17 191.7372 0.00 1.45 190.3275 0.89 0.71 

4 174.6790 233.69 179.7564 0.00 2.91 174.6790 0.453 0.00 

5 422.4267 616.30 434.7873 0.00 2.93 442.0500 0.656 4.65 

6 538.7574 805.59 538.7574 0.00 0.00 538.7574 0.343 0.00 

7 85.4226 753.52 86.9119 0.00 1.74 89.8888 0.484 5.23 

8 194.1343 40.50 196.3468 0.00 1.14 199.1684 0.578 2.59 

9 53.3469 9.66 58.7888 0.00 10.20 58.6819 0.344 10.00 

10 240.7639 16.64 240.7639 0.00 0.00 240.7639 0.391 0.00 

11 396.7343 419.02 398.7735 0.00 0.51 453.2415 0.453 14.24 

12 173.3962 692.55 173.4065 0.00 0.01 176.3493 0.61 1.70 

13 359.3905 51.55 359.3905 0.00 0.00 359.3905 0.562 0.00 

14 225.1412 171.69 226.3883 0.02 0.55 244.0558 0.422 8.40 

15 399.7456 1182.19 399.7456 0.00 0.00 399.7456 0.484 0.00 

Average 438.6033 0.0033 2.10 0.5249 3.18 


No. Enumeration Algorithm 2 Pattern search 

Total Time Total Time % Dev. Total Time % Dev. 

costs (seconds) costs (sec) from costs (sec) from 

(baht) (baht) opt. (baht) opt. 

1 231.96 388.13 231.96 0.05 0.00 231.96 0.70 0.00 

2 231.15 310.00 267.04 0.00 15.52 232.33 0.58 0.51 

3 195.83 843.94 196.89 0.00 0.54 198.17 0.55 1.20 

4 286.68 1640.67 291.76 0.00 1.77 286.68 0.63 0.00 

5 448.69 3548.72 461.05 0.00 2.75 468.31 0.84 4.37 

6 623.79 3839.44 623.79 0.00 0.00 623.79 0.48 0.00 

7 95.46 3736.39 99.51 0.00 4.24 97.66 0.61 2.30 

8 233.65 252.59 235.87 0.00 0.95 245.51 0.53 5.08 

9 63.30 22.25 76.10 0.00 20.23 73.49 0.42 16.11 

10 353.81 161.06 353.81 0.00 0.00 353.81 0.59 0.00 

11 415.36 1184.09 415.36 0.00 0.00 415.36 0.72 0.00 

12 137.41 1197.42 150.03 0.00 9.18 142.33 0.64 3.58 

13 110.35 9400.22 111.73 0.00 1.25 110.35 0.63 0.00 

14 171.36 5738.92 173.87 0.00 1.47 238.33 0.69 39.08 

15 210.21 930.83 210.21 0.02 0.00 210.21 0.64 0.00 

Average 2212.98 0.00 3.86 0.62 4.82

390 



No. LB of Algorithm 2 Pattern search 

Total costs Total costs Time % Dev. Total costs Time % Dev. 

(baht) (baht) (sec) from LB (baht) (sec) from LB 

1 341.4262 349.9014 0.05 2.48 381.6719 0.532 11.79 

2 698.4675 699.5436 0.00 0.15 698.7817 0.641 0.04 

3 306.6030 313.5858 0.00 2.28 327.0063 0.469 6.65 

4 782.9640 789.0657 0.00 0.78 791.1078 0.516 1.04 

5 141.2319 145.3478 0.00 2.91 147.0369 0.625 4.11 

6 248.9744 252.0126 0.00 1.22 250.6189 0.5 0.6 

7 418.7751 418.7751 0.00 0.00 418.7751 0.313 0.00 

8 870.7950 871.3350 0.00 0.06 871.3350 0.516 0.06 

9 188.2021 189.3901 0.00 0.63 189.3901 0.563 0.63 

10 607.2085 611.1003 0.02 0.64 618.0523 0.766 1.79 

11 310.3231 317.2456 0.00 2.23 313.7932 0.578 1.12 

12 767.5441 779.7637 0.00 1.59 775.4064 0.687 1.02 

13 796.4144 811.3657 0.00 1.88 872.1025 0.765 9.50 

14 216.6816 220.6952 0.00 1.85 226.5848 0.563 4.57 

15 329.8299 331.9408 0.00 0.64 348.9675 0.453 5.80 

Average 0.0047 1.29 0.5658 3.25 





1 781.7861 781.7861 0.05 0.00 937.4422 0.875 19.91 

2 116.0982 128.7529 0.00 10.90 118.4608 1.125 2.04 

3 250.6759 253.0172 0.00 0.93 276.9489 1.281 10.48 

4 863.7852 866.8405 0.02 0.35 932.2620 0.844 7.93 

5 416.318 425.6585 0.00 2.24 434.4976 1.282 4.37 

6 769.3648 784.0514 0.00 1.91 775.9314 0.906 0.85 

7 831.0751 834.5863 0.00 0.42 861.3861 0.875 3.65 

8 732.6523 737.9136 0.00 0.72 769.4482 1.157 5.02 

9 171.9773 180.0650 0.02 4.70 175.0647 0.765 1.80 

10 356.2244 366.1220 0.00 2.78 359.5120 1.062 0.92 

11 883.9080 883.9080 0.02 0.00 883.9080 0.39 0.00 

12 197.3155 217.3515 0.00 10.15 206.1805 0.672 4.49 

13 532.1986 533.0392 0.00 0.16 632.9300 0.703 18.93 

14 905.1188 910.7604 0.00 0.62 921.6216 1.125 1.82 

15 805.6426 807.9736 0.00 0.29 845.8867 0.781 5.00 

Average 0.0073 2.41 0.9229 5.81

the maximum deviation of about 10%. The solving 

times were still less than 1 second for all test 

instances. 

The qualities of solutions and the 

computation times from Algorithm 2 and from 

pattern search seemed to be competitive, especially 

for the small-size test problems. The differences 

between the total costs found from Algorithm 2 

and from pattern search were compared using the 

paired t-test and the signed rank test (Montgomery 

and Runger, 2004). The former was tested whether 


or not the average of the differences in total costs 

equaled zero. The latter was a non-parametric 

hypothesis test on the median of the differences 

in total costs. Under the normality assumption of 

data, the paired t-test was more powerful than the 

signed rank test. However, the signed rank test was 

less sensitive to the outliers. Herein, the signed 

rank test was applied because the distributions of 

the total costs showed significant departures from 

normal distributions. The summary of the 

statistical tests was presented in Table 9. 





1 335.2583 341.7339 0.05 1.93 338.9744 1.204 1.11 

2 448.8165 457.5385 0.02 1.94 468.6784 1.078 4.43 

3 241.1376 243.6205 0.00 1.03 243.0759 1.313 0.80 

4 689.6022 692.9453 0.00 0.48 708.7569 1.641 2.78 

5 977.388 980.5293 0.02 0.32 1010.1633 1.391 3.35 

6 326.5053 328.5352 0.00 0.62 328.1805 1.078 0.51 

7 750.4167 751.3261 0.00 0.12 751.2409 1.000 0.11 

8 810.929 813.4720 0.00 0.31 858.0251 1.766 5.81 

9 763.4033 764.7848 0.00 0.18 777.3780 1.250 1.83 

10 1047.4942 1047.4942 0.02 0.00 1047.4942 1.734 0.00 

11 964.0926 972.7866 0.00 0.90 1626.5400 1.532 68.71 

12 600.9621 612.3057 0.00 1.89 685.6725 0.656 14.10 

13 819.0913 820.9975 0.00 0.23 1098.4323 1.265 34.10 

14 966.8934 968.9883 0.00 0.22 1042.7315 1.360 7.84 

15 730.5132 730.5132 0.02 0.00 730.5132 1.172 0.00 

Average 0.0087 0.68 1.2960 9.70 

Table 9 The statistical results from the paired t-test and the signed rank test. 

Problem set Average of Median of p-value 

the difference the difference Paired t-test Signed rank test 

in total costs in total costs 

1 3.9592 0.0000 0.3578 0.7109 

2 1.9540 0.0000 0.7095 < 1.0000 

3 8.6375 1.6891 0.3865 0.1180 

4 27.9770 10.8612 0.0372* 0.0438* 

5 79.2191 12.5932 0.0999 0.0287* 

* indicates a significant difference in total costs found from both methods

392 

From the statistical tests, the qualities of 

solutions found from both methods were not 

significantly different for test problem sets 1, 2 

and 3. However, Algorithm 2 became superior to 

the pattern search for larger problem sets. Notice 

on the test results, the total cost obtained from the 

pattern search could be as poor as 68% deviating 

from the MINLP lower bounds in large problem 

instances, while those from Algorithm 2 would not 

be worse than 20% from the lower bounds. 

Algorithm 2 was hence the most efficient method 

for solving this safety stock determination 

problem, in terms of both solution quality and 

computation time. 

DISCUSSION 

It had been shown in the previous section 

that Algorithm 2, which was based on a basic NLP 

theorem, could provide high quality solutions in 

short computation times for the safety stock level 

determination problem in the considered two-stage 

manufacturing system. The algorithm utilized only 

a set of simple decision rules, in contrast to the 

pattern search heuristic, which required the users 

to comprehend its mechanisms. The decision rules 

for finding optimum safety stocks also matched 

the common managerial logics that when the 

opportunity cost was high, the safety stocks should 

be held to prevent the product deficiency, but they 

should not be stocked when the inventory costs 

were high. Moreover, the search heuristic such as 

pattern search would terminate the search after 

some stopping criteria had been satisfied. 

Therefore, in some cases, it might not thoroughly 

search the solution space for the solutions. 

CONCLUSION 

In this research, two algorithms for 

determining safety stocks in a two-stage 

manufacturing system were proposed by analyzing 

the derivatives of cost function and the cost 


comparisons. The algorithms were found to work 

very efficiently on the test problems. They could 

provide high-quality solutions for every test 

instance in less than 1 second. The deviations from 

the known integer solutions or the lower bounds 

were less than 3% on average. The algorithms also 

outperformed the pattern search algorithm, which 

was presented in the previous research. 

ACKNOWLEDGEMENT 

This project was funded by faculty of 

Agro-Industry, Kasetsart University. The author 

gratefully acknowledges this support. 

LITURATURE CITED 

Inderfurth, K. and S. Minner. 1998. Safety Stocks 

in Multi-Stage Inventory Systems under 

Different Service Measures. Eur. J. Oper. 

Res. 106: 57-73. 

Krupp, J.A.G. 1997. Safety Stock Management. 

Prod. Inventory Manag. J. 38(3): 11-18. 

Kolda, T.G., R.M. Lewis and V. Torczon. 2003. 

Optimization by Direct Search: New 

Perspectives on Some Classical and Modern 

Methods. Siam Rev. 45(3): 385–482 

Maia, L.O.A. and R.Y. Qassim. 1999. Minimum 

Cost Safety Stocks for Frequent Delivery 

Manufacturing. Int. J. Prod. Econ. 62: 233- 

236. 

Marsden, J.E. and A. Tromba. 1981. Vector 

Calculus. 2 nd ed. W.H. Freeman. San 

Francisco. 591 p. 

Montgomery, D.C. and G.C. Runger. 2004. 

Applied Statistics and Probability for 

Engineers, 3 rd ed. John Wiley & Sons. New 

Jersey. 706 p. 

Silver, E.A., D.F. Pyke and R. Peterson. 1998. 

Inventory Management and Production 

Planning and Scheduling. 3 rd ed. John Wiley 

& Sons. New Jersey. 754 p. 

Siribanluoewut, Y. 2006. Determining Safety

Stock Quantities Using Heuristic 

Optimization. M.S. Thesis. Kasetsart 


Talluri, S., K. Cetin and A.J. Gardner. 2004. 

Integrating Demand and Supply Variability 

into Safety Stock Evaluations. Int. J. Phys. 

Distrib. Logist. Manag. 34: 62-69. 


Vollmann, T.E., W.L. Berry and D.C. Whybark. 

1997. Manufacturing Planning and Control 

Systems, 4 th ed. McGraw-Hills. New York. 

836 p. 

Zeng, A.Z. 2000. Efficiency of Using Fill-Rate 

Criterion to Determine Safety Stock: A 

Theoretical Perspective and a Case Study. 

Prod. Inventory Manag. J. 41(2): 41-44.


Design and Implementation of a Framework for .NET-based Utility 

Computing Infrastructure 

Thanapol Rojanapanpat * and Putchong Uthayopas 

ABSTRACT 

Future organizations must handle a very large and complex IT infrastructure that consists of 

very diverge and highly heterogeneous computing systems. Moreover, the future generation applications 

must access services and resources regardless of the geographical location, access methods, and domain 

of authorization. In order to meet these challenging requirements, a very high degree of virtualization 

has to be implemented using a smart middleware. This is a very challenging problem for both theory 

and practice. 

This paper presents a new framework called OpenUCI (Open Utility Computing Infrastructure). 

The OpenUCI project aims to explore the innovative design of scalable and flexible software infrastructure 

that manages large scale heterogeneous distributed system ranging from large Server, PC, and Mobile 

Devices. OpenUCI exploits a well established technology such as Grid, Web services and .NET technology 

to build a virtualized and unify access to resources. Basic services that need to be presented will be 

discussed. The prototype system has been implemented along with the prototype financial engineering 

application. The results are presented along with the discussion of the experiences learned. With OpenUCI, 

users can easily harness computing and storage of large distributed system. 

Key words: utility computing, .NET technology, web services 

INTRODUCTION 

The competition in business causes 

organizations to be ready to handle a large amount 

of demand of users, which need more high 

performance computing system. It is a risk for the 

small and medium organizations to invest in the 

high performance computing system, because they 

have to pay for the system maintenance cost. There 

are two solutions. Firstly they can outsource the 

computing power. The other solution is to create 

the supercomputing system by utilizing the already 

existing personal computers (PC) in their 

company. Building a supercomputing system from 

personal computers or desktop PCs now is not an 

imagination, because the speed and performance 

of PCs has been increasing as well as the speed 

and bandwidth of network. From this advantage, 

it emerges many new computing systems; one of 

them is the utility computing system. 

Utility computing (Eilam et al., 2004) is 

a computing model that involves the use of many 

diverge technology such as grid computing (Foster 

et al., 2002) and autonomic computing (Ganek and 

Corbi, 2003). Utility computing system focuses 

on the creating of virtual computing environment 

High Performance Computing and Networking Center, Faculty of Engineering, Kasetsart University, Bangkok 10900, Thailand, 

* Corresponding author, e-mail: thanapolr@hpcnc.cpe.ku.ac.th, pu@ku.ac.th 


which dynamically and automatically virtualizes, 

provisions and manages resources and services on 

users’ demand. The major benefits of utility 

computing are 

• better utilization – the resources in 

utility computing system can be shared and used 

in efficient ways, 

• more flexibility – utility computing 

system provides flexibility in the creation of the 

dynamic computing environment which can 

automatically increase or decrease the computing 

resources corresponding to users’ demand, and 

• lower total cost of ownership – utility 

computing can provide IT and business process 

outsourcing which help reducing cost of investing 

in resources such as hardware assets, maintenance 

cost, training cost, etc. 

The design and building of utility 

computing infrastructure is still a complex and 

challenging task. Figure 1 shows the concept of 

resources virtualization and resources provisioning 

in a modern IT infrastructure. Utility computing 

system must consist of a way to provide both 

features. Firstly, resources virtualization is a 

feature of system that can make resources 

transparent to application. Since the resources that 

we use for build a utility computing infrastructure 

are PCs, these systems have a high dynamism e.g. 

resources can be available and unavailable from 

time to time. The utility computing system must 

have mechanisms for collecting resources and 

monitoring its status. Furthermore, resource 

virtualization should have mechanisms for 

discovering and accessing resources. Secondly, 

resources provisioning is a feature of a system to 

perform an on-demand resources allocation to 

application. A good utility computing system must 

have mechanism for creating the automatic 

adjustable virtual computing environments which 

consist of hardware resources and utility services 

in order to keep responsiveness when users’ 

demand increases. Moreover, the utility computing 

system must provide friendly-used interfaces to 


user, which may be business manager for 

accessing and managing this virtual computing 

environment. These interfaces can be Windows 

applications, Web applications, command line, and 

API. 

Most utility computing systems are based 

on current distributed system technology. Thierry 

(2006) provides a good survey of platform 

technology that is available. There are three 

commonly used distributed systems and 

technologies. The first one is distributed 

information system that focuses on sharing 

knowledge such as the web. Secondly, a distributed 

storage system for sharing data such as peer-topeer 

files sharing. Finally, distributed computing 

or metacomputing (Smarr and Catlett, 1992) 

frameworks for sharing computing power. The 

systems that are classified in this area and related 

to OpenUCI framework are the followings. 

Business 

Processes/ 

Appications 

Virtual 

Computing 

Environment 

Resources/ 

Services 

Resources Provisioning 

Virtualized Resource 

Business 

Processes/ 

Appications 

Virtual 

Computing 

Environment 

Resources Virtualization 

Resources/ 

Services 

Resources/ 

Services 

Figure 1 The relationship of resource 

viortualization and resource 

provisioning.

396 

Grid computing (Foster et al., 2002) 

focuses on integrating geographically distributed 

resources into a unified system. Grid computing 

provides concept of Virtual Organization (VO) 

which is an integrated resources shared by real 

organizations, and it also has a well-defined 

architecture, services and protocols such as 

resource discovery, job submission, system 

monitoring and accounting, which are good 

patterns for designing and developing the utility 

computing system. The most well-known project 

in this area is Globus (The Globus Alliance, 

2005a). 

Peer-to-Peer (P2P) computing is a class 

of applications that takes advantage of resources 

such as storage, CPU cycles, and content that are 

available on the Internet. There are two major 

categories of P2P system, P2P networking (file 

sharing) and P2P computing (CPU sharing), The 

P2P networking is a communication model in 

which each node (peer) has the same capabilities 

and either node can directly initiate a 

communication session. The P2P computing is a 

processing power sharing rather than a files 

sharing. 

Volunteer computing (Sarmenta, 2001) 

focuses on making computers to be a part of 

metacomputer dynamically when computing 

power is available. The topology of volunteer 

computing is usually similar to the third generation 

of peer-to-peer computing. The peer can be both 

client, who submits jobs to server (super-peer), 

and can be worker who dedicates itself to execute 

jobs. This includes system such as SETI@home 

(Anderson et al., 2002), Bayanihan (Sarmenta et 

al., 2002), and Alchemi (Luther, 2005). 

In this paper, we present a design and 

implementation of a framework called OpenUCI 

(Open Utility Computing Infrastructure) which is 

for constructing the utility computing 

infrastructure from Windows-based personal 

computers, because the most of computers in the 

organization are Windows-based operating system 


and the most of users are familiar to Windows. To 

solve the resource virtualization and resource 

provisioning problems, OpenUCI framework 

provides many services such as resource 

collecting, resource monitoring, resource 

discovering, resource invocation, and etc. In 

addition, we use Microsoft’s .NET technology for 

implementing the OpenUCI system because it 

provides a powerful and comfortable development 

environment and it also provides ASP.NET Web 

service, a standard way for communication 

between systems. So, we can ensure that all 

OpenUCI’s components can work together and can 

communicate to other systems seamlessly. 


1. Hardware and software requirements 

This paper develops and tests a 

framework on Windows-based system. The 

computers used in this development comprise one 

manager node, 32 worker nodes, and one user 

node. All nodes are connected with Fast Ethernet 

switch. The system configuration is shown in 

Figure 2. 

The software for developing and testing 

the framework is as follows: 

• Microsoft Windows Server 2003 

• Microsoft Windows XP Professional 

• .NET framework redistributed 1.1 

and 2.0 

• Microsoft Visual Studio .NET 2003 

and 2005 

Figure 2 The windows cluster.

2. Framework architecture and components 

In this paper, utility service is a function 

provided by any computers. The utility service 

must depand on the Service Oriented Architecture 

(SOA) technology such as .NET web services, and 

Grid services. The example of utility service is 

such web service for calculating risk of trading 

stock (VaR) (Rojanapanpat et al., 2005). The 

resource is an entity shared by a computer and can 

be computing power (CPU), storage, files and 

utility services. 

According to the utility computing 

system development problems mentioned before, 

Resources Virtualization and Resources 

Provisioning, the proposed framework, OpenUCI, 

must be designed to solve these problems. 

To deal with Resources Virtualization 

problem, OpenUCI must have mechanism to 

support the dynamism, heterogeneity, scalability, 

interoperability of resources. The mechanisms are 

such resource collecting for gathering resources 

and track its status, resource discovery used to find 

and select the resources, resource accessing which 

defines a unite way to use and interoperate 

resources and etc. 

In the Resources Provisioning problem, 

OpenUCI must provide mechanisms for creating 

virtual computing environments that can be 

automatically adjustable depending on demand of 

users. Moreover, OpenUCI must provide userfriendly 

interfaces and tools using OpenUCI 

system and accessay resources to users. 

The architecture of the OpenUCI 

framework is shown in Figure 3. There are four 

layers of the OpenUCI framework, i.e. resources, 

.NET platform, core services, and applications. 

2.1 Resources layer 

Resources layer is the layer of shared 

resources distributed on the network. The shared 

resources consist of CPU, storage and utility 

services. 

2.2 .NET platform layer 

.NET platform layer provides a runtime 


environment, .NET framework, which OpenUCI 

system relies on. This layer also provides 

technologies for implementing OpenUCI system, 

and sharing resources. These technologies are 

.NET web services, .NET remoting and 

WSRF.NET. The resources can be shared via these 

technologies. 

2.3 Core layer 

This layer provides a set of necessary 

services for building the utility computing 

infrastructure and supporting the basic functions 

of the application running on the utility computing 

infrastructure. The core services are classified into 

two groups according to our requirements. 

The core services that solve the resources 

virtualization problem consist of resource 

management service, data management service 

and execution management service. 

1. Resources Management Service 

(RMS) is responsible for gathering resources 

distributed on the network and tracking the 

existence and status of resources. Moreover, RMS 

also provides mechanisms for resource discovery, 

resource reservation and etc. 

2. Data Management Service (DMS) is 

responsible for transferring files and sharing files 

JOB 

MANAGEMENT 

RESOURCES 

MANAGEMENT 

.NET WEB 

SERVICES 

APPLICATIONS & TOOLS 

VIRTUAL 

COMPUTER 

MANAGEMENT 

RESOURCES PROVISIONING 

EXECUTION 

MANAGEMENT 

RESOURCES VIRTUALIZATION 

CORE SERVICES 

.NET WEB 

REMOTING 

.NET PLATFORM 

USER 

MANAGEMENT 

DATA 

MANAGEMENT 

WSRF .NET 

CPU STORAGE SERVICE 

RESOURCES 

Figure 3 The OpenUCI architecture.

398 

among computers in the OpenUCI system. 

3. Execution Management Service 

(EMS) is used to start and controls processes. 

Furthermore, EMS also supports the invocation 

of web and grid service jobs. 

The core services that address the 

resources provisioning problem consist of user 

management service, virtual computer 

management service and job management service. 

1. User Management service (UMS) 

handles authentication, authorization, accounting 

and users profiles. 

2. Virtual Computer Management 

Service (VCMS) is used for managing and 

controlling the virtual computing environment 

created by users. 

3. Job Management Service (JMS) is 

used for creating jobs and supporting job 

submission from users. JMS also provides job 

queuing and scheduling mechanisms. 

2.4 Applications and tools layer 

Applications and tools layer is the layer 

of user applications developed for using facilities 

of OpenUCI system. OpenUCI system also 

provides basic command-line tools and web 

application interfaces for login, logout, virtual 

computer creation, resources discovering, job 

submission and etc. 

There are three main components in 

OpenUCI system as shown in Figure 4. 

Users 

Applications 


Manager 

Core Services 

Figure 4 The interaction of manager, worker and user. 

1. Manager is a computer that provides 

core services used for managing shared resources 

and supporting incoming requests of users. 

2. Workers are computers that share its’ 

resources such as computing power, files, storage 

and utility services. There are two worker types in 

the OpenUCI system, dedicated and non-dedicated 

workers. Dedicated workers are always online and 

cannot reject jobs assigned by managers. For nondedicated 

workers, they can be online or offline 

all the time and they will request for a job and 

execute it when they are not busy. 

3. Users are the people who need to 

access resources. They can discover resources, 

create job, submit job, download and upload files 

and any services provided by managers. 


1. Proof of concept application 

Currently, the high performance 

computing is widely needed and not limited to the 

computer research field anymore. The financial 

engineering (FE) is a field that requires the high 

computing power because it has to handle and 

analyze a large amount of data in order to reduce 

or keep turn around time constantly as number of 

users increased. We evaluated the performance of 

OpenUCI system by applying the existing 

financial engineering application named Value-at- 

CPU 

Workers 

Agent 

Storage Services

Risk (VaR) calculation which was implemented 

in .NET web services. The VaR measures the 

maximum loss money which may be occurred in 

portfolio at a given time horizon (time of holding 

portfolio) and at a given level of confidence. The 

formula for calculating VaR has high complexity. 

Then, we will show the general form of formula. 

VaR = –Vp* (µp – Q*σp) 

The V p is the portfolio value, and the µp 

and the σp are the expected return and the standard 

deviation, respectively. The Q is the quantile value 

of %confidence level. For example, the 99% 

confidence level gives ~2.326 quantile value and 

the 95% confidence level gives ~1.645 quantile 

value. 

In this test, we used the VaR calculation 

web service as a utility service of OpenUCI system 

which was installed to all worker machines and 

then we developed VaR client program with 

Microsoft Excel. The VaR Excel program uses the 

OpenUCI API to connect to manager, discover 

VaR web services and then invoke them. 

2. Test configuration 

The topology of test system is shown in 

Figure 2. The software that was installed on each 

machine is shown in Table 1. 

3. Test assumptions 

• Each worker executes only one job 

at a time. Since the test application is a compute 

intensive application, the execution of more than 


one job on each worker will not give a better 

performance due to the overhead of task switching. 

• The input data is already in the 

workers. This can be done by preloading fixed data 

and table to worker prior to the execution. Thus, 

the communication can be minimized which yield 

a better performance for the system. 

4. OpenUCI throughput test 

We evaluated the throughput of 

OpenUCI by submitting jobs to OpenUCI system 

that has 1, 2, 4, 8, 16, and 32 workers, and the run 

times used for testing are changed from 10, 30, 

60, 90, 120, 180, 240, and 300 seconds. Figure 5 

shows the procedure of this testing. 

1. The client application discover URLs 

of web service located on the worker nodes from 

the manager. 

2. The manager runs the resource 

selection algorithm and returns the URLs of the 

chosen worker node to requested client 

application. 

3. The client application uses the 

returned URLs for connecting and invoking web 

service on worker nodes. After that, the client 

application will wait until there are some available 

workers. 

4. The worker node executes the service 

and then it returns a result to client application. 

5. The client program invokes web 

service on an available worker 

Table 1 Hardware and software configuration for testing OpenUCI system. 

Machines Hardware Operating system Software 

1 Manager AMD Athlon 2.0GHz, 512 Windows server 2003 OpenUCI Broker, MS 

MB RAM SQL 2005 for 

OpenUCI database 

32 Workers Intel Celeron 2.53GHz, 512 Windows XP OpenUCI Worker, 

MB RAM Professional MS SQL 2005 

Express for VaR 

database 

1 User Intel Pentium M 1.5GHz, Windows server 2003 VaR client application 

768 MB RAM

400 

VaR Client 

(OpenUCI User) 

Wait for available 

worker 

1) Discovery for VaR web 

service 

2) Return the suitable 

VaR web service URLs 

Figure 5 The throughput test procedure. 

The result of throughput test is shown in 

Table 2 and Figure 6. Figure 7 shows average of 

throughput of OpenUCI system based on the 

different number of workers. 

From these results, it shows that 

OpenUCI system gave a good throughput when 

the number of workers increased and the 

increasing of throughput was nearby the increasing 

of number of workers. For example, the average 

throughput of 32 workers system was ~6.4 jobs/ 

sec and the average throughput of 1 worker system 

was ~0.214 jobs/sec. The throughput was 

increased about 30 times. 


OpenUCI Manager OpenUCI Worker1 OpenUCI Worker n 

3) Invoke VaR web 

service on all workers 

4) Return the result 

5) Continue invoking 

5. OpenUCI speed up test 

In this test, we observed the run time used 

to finish jobs when the number of workers was 

changed from 1, 2, 4, 8, 16, to 32 workers. The 

procedure of the speed up testing was similar to 

the throughput testing, but the speed up test 

changed the number of jobs submitted to system 

and observed the run time instead of fixing the 

run time and observed the number of finished jobs. 

Table 3 and Figure 8 show the run time 

of this testing. Table 4 and Figure 9 show the speed 

up. Table 5 and Figure 10 show the efficiency. 

Table 2 The throughtput of OpenUCI. 

Time 1 Worker 2 Workers 4 Workers 8 Workers 16 Workers 32 Workers 

10 0.20 0.30 0.70 1.40 3.10 6.00 

30 0.23 0.43 0.73 1.63 3.17 6.27 

60 0.22 0.42 0.82 1.57 3.18 6.33 

90 0.21 0.41 0.79 1.64 3.24 6.44 

120 0.22 0.43 0.84 1.67 3.31 6.47 

180 0.21 0.42 0.83 1.63 3.30 6.59 

240 0.21 0.42 0.84 1.67 3.31 6.56 

300 0.21 0.42 0.84 1.65 3.31 6.60

Throughput (job/sec) 

7 

6 

5 

4 

3 

2 

1 

0 


Throughput 

0 30 60 90 120 150 180 210 240 270 300 330 

Time (second) 

Figure 6 The throughput of OpenUCI system. 

Average Throughput (job/sec) 

7.00 

6.00 

5.00 

4.00 

3.00 

2.00 

1.00 

0.00 

0.21 

0.41 

Average Throughput 

0.80 

1.61 

1 2 4 8 16 32 

Number of Workers 

Figure 7 The average throughput of OpenUCI system. 

3.2 4 

1 Worker 

2 Workers 

4 Workers 

8 Workers 

16 Workers 

32 Workers 

6.40

402 


Table 3 The run time of testing (second). 

Worker 100 Jobs 500 Jobs 1000 Jobs 2000 Jobs 3000 Jobs 

1 476.33 2359.17 4726.77 10083.33 14794.66 

2 248.03 1191.59 2400.58 4734.84 7106.32 

4 122.14 596.77 1185.05 2386.19 3566.53 

8 61.82 303.70 609.31 1216.19 1825.01 

16 33.30 157.66 308.28 619.43 923.29 

32 19.88 76.25 151.92 301.97 451.55 

Time (second) 

100000 

10000 

1000 

100 

10 

Figure 8 The run time plot. 

1 

Run time 

100 Jobs 

500 Jobs 

1000 Jobs 

2000 Jobs 

3000 Jobs 

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 

Number of workers 

Table 4 The speed up of testing. 


1 1.00 1.00 1.00 1 1 

2 1.92 1.98 1.97 2.13 2.08 

4 3.70 3.95 3.99 4.23 4.15 

8 7.71 7.77 7.76 8.29 8.11 

16 14.31 14.96 15.33 16.28 16.02 

32 23.97 30.94 31.11 33.39 32.76

Speed up 

40 

35 

30 

25 

20 

15 

10 

5 

0 

Figure 9 The speed up plot. 


Speed up 

100 Jobs 

500 Jobs 

1000 Jobs 

2000 Jobs 

3000 Jobs 

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 


Table 5 The efficiency of testing. 


1 1.00 1.00 1.00 1.00 1.00 

2 0.96 0.99 0.98 1.06 1.04 

4 0.97 0.99 0.99 1.06 1.04 

8 0.96 0.97 0.97 1.04 1.01 

16 0.89 0.94 0.96 1.02 1.00 

32 0.75 0.97 0.97 1.04 1.02 

The speed up (S) of n-workers system is 

defined by the run time of 1-worker system 

(sequential run time, Ts) divided by the run time 

of n-workers system (parallel run time, Tp), and 

the efficiency (E) is defined as the speed up (S) 

divided by number of workers (P). From Figure 9 

and Figure 10, we found that there were three 

interesting characteristic results. 

1. The speed up and efficiency were 

decreased when the number of workers increased, 

for example, 100 jobs testing. This characteristic 

happened because all workers in system are not 

fully utilized. For example, in 32-workers system, 

it had to use 4 iterations to finish 100 jobs 

(32+32+32+4 = 100). So, in the last iteration, there 

were 28 workers free. Assume that 1 job used 1 

second for executeing. The speed up was 25 (Ts/ 

Tp = 100/4 = 25), and the efficiency was 0.78 (S/ 

P = 25/32 = 0.78). If we submitted 128 jobs 

(32+32+32+32) to this system, the speed up and 

efficiency would be 32 (128/4) and 1, respectively. 

2. The speed up and efficiency were 

almost perfect. The perfect speed up was the speed 

up that was equal to number of workers in system. 

The perfect efficiency was the efficiency that is 

equal to 1. Basically, the communication overhead

404 

Effciency 

1.2 

1 

0.8 

0.6 

0.4 

such as input data transfer time makes the speed 

up and efficiency dropped. In this test, we reduced 

the data transfer time by replicating VaR database 

to all workers. So, the efficiency and speed up were 

nearly perfect. 

3. The super speed up and the over 

efficiency. This characteristic happened because 

the overhead time before calling web services of 

client application makes the run time of client 

application increased. The high number of jobs 

made the total overhead time grower. However, 

the total overhead time was reduced by the 

increasing of number of workers. So, at the large 

amount of jobs such as 2000 and 3000 jobs, the 

run times of 2, 4, 8, 16, and 32 workers system 

were decreased more than the number of workers 

in system. 

CONCLUSION 

The demand of using super computing 

system in organizations has been increasing. They 


Effciency 

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 

Figure 10 The efficiency plot. 


100 Jobs 

500 Jobs 

1000 Jobs 

2000 Jobs 

3000 Jobs 

need the system that has more dynamicity and 

flexibility in order to support the various types and 

large amount of demand of customers. Moreover, 

this system must provide an easy and familiar 

mechanism for customers to use the power of 

system. This paper proposed the design and 

implementation of framework used for building 

the computing environment that can achieve these 

requirements. This framework is called OpenUCI 

(Open Utility Computing Infrastructure) which 

works on Microsoft .NET platform. OpenUCI will 

gather resources distributed on the network, and 

automatically adjust and provisioning resources 

to users. The prototype of OpenUCI has already 

been implemented and evaluated with a financial 

engineering application named VaR calculation. 

The result of evaluation showed that OpenUCI can 

give a good performance and high utilization when 

the number of computers and demand of users 

increased 

The prototype version of OpenUCI has 

only a few modules such as resource collecting

and discovery, resource selection and broker 

mechanism. There are still many necessary 

modules that should be implemented, for example, 

web and grid services invoker, job queue manager 

and virtual computer management. The following 

is the list of future work. There are many possible 

works in the future such as integrating the 

executable file launcher implemented in another 

related project to OpenUCI system, implementing 

the job queue management module, implementing 

the web and grid services invoker module, 

implementing the virtual computer management 

service, implementing the user authentication and 

accounting modules, implementing the data 

transfer service, exploring the mechanisms for 

handling fault of machines and jobs and 

investigating a proper workload distribution 

scheme and study using simulation. 

All these works will make OpenUCI 

more useful in the modern computing 

environments. 

LITERATURE SITED 

Albaugh V. and H. Madduri. 2004. The utility 

metering service of the Universal 

Management Infrastructure. IBM Systems 

Journal 43(1): 159-178 

Anderson D., J.Cobb, E. Korpela, M. Lebofsky 

and D. Werthimer. 2002. SETI@home: An 

Experiment in Public-Resource Computing. 

Communications of the ACM 45(11): 56- 

61 

Eilam T., K. Appleby, J. Breh, G. Breiter, H. Daur, 

S.A. Fakhouri, G.D.H. Hunt, T. Lu, S.D. 

Miller, L.B. Mummert, J.A. Pershing and H. 

Wangner. 2004. Using a utility computing 

framework to develop utility systems. IBM 

System Journal 43(1): 97-120 

Foster I., C. Kesselman, J. Nick and S. Tuecke. 

2002. The Physiology of the Grid: An Open 

Grid Services Architecture for Distributed 

Systems Integration. Open Grid Service 

Infrastructure WG. Globus Grid Forum 


Ganek G. and T. A. Corbi. 2003. The dawning of 

the autonomic computing era. IBM System 

Journal 42(1): 5-18 

Humphrey M. and G. Wasson. 2005. Architectural 

Foundations of WSRF.NET., International 

Journal of Web Services Research 2(3): 83- 

97. 

Luther A., R Buyya and S. Venugopal. 2005. 

Alchemi: A .NET-Based Enterprise Grid 

Computing System. Proceedings of the 6th 

International Conference on Internet 

Computing (ICOMP’05), June 27-30, 2005, 

Las Vegas, USA 

Rojanapanpat T., P. Uthayopas, S. Chaisiri, J. 

Pichitlamken, S. Phakhawirotkul and T. 

Vorakosit. 2005. Implementing a Distributed 

High Volume Risk Analysis Software on PC 

Farm using OpenUCI System. The 9th 

National Computer Science and 

Engineering Conference (NCSEC2005), 

October 27-28, 2005, Bangkok, Thailand. 

Sarmenta L. F. G. 2001. Volunteer Computing. 

Ph.D. thesis, Massachusetts Institute of 

Technology. 

Sarmenta L. F. G., S. J. V Chua, P. Echevarria, J. 

M. Mendoza, R. R. Santos and S. Tan. 2002. 

Bayanihan Computing NET: Grid Computing 

with XML Web Services. Workshop on 

Global and Peer-to-Peer Computing at the 

2nd IEEE International Symposium on 

Cluster Computing and the Grid (CCGrid 

’02), May 2002, Berlin, Germany. 

Smarr L. and C. Catlett. 1992. Metacomputing, 

pp. 44-52. Communication of the ACM, 35. 

The Globus Alliance. 2005. Welcome to The 

Globus Toolkit Homepage. The Globus 

Toolkit. Available source: http:// 

www.globus.org/toolkit/, March 14, 2006. 

Thierry P. 2006. CoreGRID: European Research 

Network on Foundations. Software 

Infrastructures and Applications for large 

scale distributed GRID and Peer-to-Peer 

Technologies. Available source: http:// 

www.coregrid.net/, March 14, 2006.

Text and Journal Publication Co., Ltd. 

158/3 Soi Viphawadi Rangsit 5, Viphawadi Rangsit Road, Ladyao, Chatuchak, Bangkok 10900 

tel. 0-2617-8611-3 fax. 0-2617-8616

April - June 2007 - Kasetsart University

Create successful ePaper yourself

Delete template?

Save as template?