Cotyledons'

Plant Physiol. (1982) 70, 1162-1168
0032-0889/82/70/1 162/07/$00.50/0
Fluorescence Immunohistochemical Localization of Malate
Dehydrogenase Isoenzymes in Watermelon Cotyledons'
A DEVELOPMENTAL STUDY OF GLYOXYSOMES AND MITOCHONDRIA
Received for publication January 19, 1982 and in revised form June 5, 1982
CHRISTOF SAUTTER AND BERTOLD HOCK
Department of Botany, Faculty of Agriculture and Horticulture, Technical University of Munich, D-8050
Freising 12, West Germany
ABSTRACT
Monospecific antibodies to glyoxysomal, mitochondrial, and cytosolic I
malate dehydrogenase were used for the fluorescence immunohistochemical
localization of these isoenzymes in dark-grown watermelon (Citrlus
vulgaris Schrad.) cotyledons. It was demonstrated that, with cell organelies
isolated by sucrose density gradient centrifugation, antibodies to glyoxysomal
malate dehydrogenase were specific markers for glyoxysomes, and
similarly, antibodies to mitochondrial malate dehydrogenase were markers
for mitochondria. The time course of the glyoxysomal malate dehydrogenase
appearance and decline was not synchronous for the individual tissues
and differed completely from that of the mitochondria. The cytosolic malate
dehydrogenase I was confined to restricted regions of the lower epidermis.
The activity which was definitively localized outside the cell organeiles
decreased during the first days of germination.
The MDH2 isoenzymes in cotyledons of dark-grown fatty seedlings
are involved in different metabolic functions, ie. the glyoxylate
pathway, the tricarboxylic acid cycle, and the malate/aspartate
shuttle (1 1). The individual forms exhibit different molecular
properties (19, 20), and they can be easily separated by PAGE,
isoelectric focusing, or other techniques (16). The association of
the isoenzymes with different cellular compartments has been
established by electrophoretic analyses of cell organelles which
were purified by sucrose density gradient centrifugation (8).
The individual isoenzymes which are numbered consecutively
from the anode to the cathode are distributed in the following
way: MDH V is located in the glyoxysomes; MDH III in the
mitochondria; whereas MDH I, II, and IV are cytosolic isoenzymes.
Up to now, an in situ localization of the MDH isoenzymes in
plant tissues has not been reported. Clearly, histochemical stains
which assay enzyme activities would not achieve this purpose
since they do not discriminate between the different isoenzymes.
The availability of antibodies directed against individual isoenzymes
which do not cross-react with the other isoenzymes (17, 18)
' Supported by the Deutsche Forschungsgemeinschaft (Grant Ho 383/
19). 2 Abbreviations: gMDH, glyoxysomal malate dehydrogenase; mMDH,
mitochondrial malate dehydrogenase; cMDH, cytoplasmic malate dehydrogenase;
PAGE, polyacrylamide gel electrophoresis; FP, 0.25% (w/v)
formaldehyde (freshly prepared from paraformaldehyde in 0.5 M K-phosphate
(pH 7.0); FITC, fluoresceine isothiocyanate.
overcame this difficulty.
This report presents information on the intracellular distribution
ofgMDH, mMDH, and cMDH I in the cotyledons of watermelons
during seed germination by the aid of immunofluorescence microscopy.
For this purpose, indirect immunolabeling was used
with monospecific antibodies against the isoenzymes and FITCcoupled
goat-anti-rabbit immunoglobulin G's. The data confirm
the hypothesis of the strict and precise compartmentation of the
MDH isoenzymes and provide new information on their tissuespecific
distribution.
MATERIALS AND METHODS
Plant Material. Watermelon seed (Citrullus vulgaris Schrad.,
var. Stone Mountain, harvest 1978) were obtained from Vaughan's
Seed Company (Ovid, MI). They were germinated at 300C in the
dark under sterile conditions on 0.8% agar as described before (7).
When indicated, 2-d-old seedlings were exposed to continuous
white light (36 tuE/m2. s) at 25°C.
Preparation of Monospecific Anti-MDH Antisera. Anti-gMDH
and anti-mMDH antisera were produced according to the methods
of Walk and Hock (17, 18). For the production of anti-cMDH I
antiserum, the isoenzyme was purified according to Kaiser (10),
involving DEAE-cellulose (Serva) chromatography; ammonium
sulfate fractionation; followed by chromatography on the Pharmacia
gels Sephadex G-25, Sephacryl S-200, CM-Sephadex C-50,
5'-AMP-Sepharose 4B, QAE-Sephadex A-50, Blue Sepharose CL-
6B, and isoelectric focusing. The immunization schedule was the
same as with gMDH and mMDH as antigens. All antisera were
fractionated by ammonium sulfate precipitation (12) in order to
recover the immunoglobulin G (IgG) fractions.
Separation of Cell Organelles. Glyoxysomes and mitochondria
were purified by sucrose density gradient centrifugation of a crude
particulate fraction (crude 10,000g pellet after a 10-min centrifugation,
corresponding to 30 cotyledons from 3-d-old dark-grown
seedlings) as described before (8).
Immunofluorescence Localization of MDH Isoenzymes. The
tissue processing followed in general the procedures of Baumgartner
et al. (1) and Tokuyasu and Singer (15). One mm thick crosssections
were handcut from cotyledons and fixed with FP under
slight evacuation at 20°C for 1 h. The samples were carried
through a series of increasing sucrose concentrations (0.25, 0.5, 1.0
M) in FP, each step for 30 min. The sections were mounted on
copper rods and frozen in melting nitrogen.
Glyoxysomal and mitochondrial fractions were fixed in the
fractionation medium with final concentrations of 0.25% (w/v)
formaldehyde and 25 mm K-phosphate (pH 7.0) for 15 min at
4°C. Before centrifugation (5 min at 20,000g), the volume of the
fractions was doubled by the addition of FP. The pellets were
1162

embedded in 2% (w/v) agar, and frozen in melting nitrogen.
The frozen samples (tissue slices or organelle pellets, respectively)
were cut with a Leitz 'Grundschlittenmikrotom' into sections
of 40 pam thickness using a knife angle of 20 and a temperature
of -30°C. The sections were collected at the surface of a
large droplet containing 1% (w/v) gelatin, 0.3% (w/v) agarose,
and 10 mm K-phosphate (pH 7.0). The antibody binding was
carried out at 20°C for 10 min following precisely the procedure
of Tokuyasu and Singer (15) using 5 pl IgG from antiserum or
control serum, respectively, per droplet of buffered gelatin and
agarose solution. To remove unbound antibodies, the sections
were washed three times with 10 mM glycine in 0.8% NaCl (w/v)
and 10 mm K-phosphate (pH 7.0), and then carried through three
droplets of the same buffer without glycine. The subsequent
incubation with fluorescent-labeled secondary antibodies (goat
anti-rabbit FITC-labeled IgG obtained from Behringwerke AG,
Marburg) was similar. Afterwards, the sections were mounted in
FP with the labeled side pointing upwards, and covered with a
cover slip.
The sections were examined in a Zeiss Photomicroscope II
equipped with epifluorescence (filter combination 450-490, FT
510, LP 520). Micrographs were taken on Ilford PanF, Ilford HP5,
and Agfachrome 50 L.
Histochemical Localization of MDH Activity. Two-d cotyledons
were fixed and cryosectioned as described above. For the
histochemical localization of MDH activity, the slices were incubated
for 45 min at 37°C in a freshly prepared medium, described
by Hanker et al. (6) except for 10 mM L-malate substituting for
lactate. In the controls, malate was omitted. The sections were
mounted in 0.5 M K-phosphate (pH 7.0).
Electron Microscopy. The preparation of organelle fractions for
electron microscopy was carried out according to Sautter et al.
(13). The sections were counterstained with uranylacetate and
lead citrate and examined with a Zeiss EM 10.
RESULTS
The use of monospecific antibodies as cytochemical markers for
the intracellular localization of glyoxysomes and mitochondria
was demonstrated with anti-gMDH and anti-mMDH immunoglobulin
fractions, respectively. To exclude any artificial binding
of the antibodies, the assays were first carried out with isolated
glyoxysomes and mitochondria. For this purpose, the organelles
from 3-d cotyledons were separated by sucrose density gradient
centrifugation and identified by measuring the isocitrate lyase and
fumerase activities as markers for glyoxysomes and mitochondria,
respectively. The gradients were loaded to the limits of their
capacity by an equivalent of 30 pairs of cotyledons per gradient
and yielded glyoxysomes at a density of 1.24 g/cmn and mitochondria
at a density of 1.19 g/cm3. The two organelle fractions
were prefixed, embedded in agar and frozen. After sectioning, the
primary immunoreaction was carried out with anti-MDH antibodies
as indicated, followed by treatment with secondary FITClabeled
antibodies.
Figure 1 shows the immunofluorescence of the glyoxysomal (A,
C) and the mitochondrial fractions (B, D) challenged with antigMDH
(A, B) and anti-mMDH antibodies (C, D) respectively.
The fluorescence patterns correspond to the electron microscopic
analysis (E, F). A high level of specific staining was observed
when glyoxysomes were treated with anti-gMDH antibodies and
mitochondria with anti-mMDH antibodies. In addition to the
strong fluorescence of the particles, there was in both cases a
distinct background fluorescence which was due to the leakage of
the isoenzymes out of the organelles into the medium during
preparation. When the antisera were substituted by control serum
obtained prior to the immunization, a much weaker background
without any particulate staining was observed. In spite of the
absence of isocitrate lyase activity in the mitochondrial fraction,
IMMUNOHISTOCHEMICAL MDH LOCALIZATION
1163
there was a slight cross-contamination of the mitochondrial fraction
by glyoxysomes (Fig. 1B) which was typical for the high
gradient loads. The reverse case shown in Figure IC is not
representative; moreover, the weak fluorescence was due to an
unspecific background staining, resulting from the loss of FITC
marker from the secondary antibody. These data prove the suitability
of the antibody labels for the immunohistochemical detection
of cell organelles.
Figure 2 shows the distribution of glyoxysomes in frozen sections
obtained from dark-grown cotyledons of different age. On
the left, representative sections of the palisade parenchyma are
shown, while on the right, corresponding sections of the storage
parenchyma are seen. In l-d cotyledons (Fig. 2, A and B), occasionally
a few glyoxysomes could be detected, usually close to the
vascular bundles. The main appearance of the organelles did not
occur before day 2 (Fig. 2, C and D). The first places where the
glyoxysomes could be seen were in the surroundings of the vascular
bundles and in the lower epidermis and in a few neighboring
layers of the future spongy parenchyma which at this time serves
as a storage parenchyma. The diameters of the globular organelles
were between 1 and 2 ,um. On day 3, a distinctive progression of
organelle production was observed. In the spongy parenchyma
(Fig. 2F), the glyoxysomes reached their maximal number and
their most intensive fluorescence, whereas in the palisade layers,
the maximum was achieved 1 d later (Fig. 2G). At this time, a
dramatic decline in glyoxysomal number and fluorescence had
already taken place in the spongy parenchyma (Fig. 2H). For a
comparison of the cell types, fluorescence micrographs of the
sections are shown in Figure 2, G and H, with the corresponding
bright field micrographs in Figure 2, 1 and K. With control serum,
shown here for 4-d cotyledons, no fluorescence could be detected
(Fig. 4E). From these data, a characteristic developmental pattern
for glyoxysomes is. inferred with an increase from almost zero
levels to a maximum in organelle numbers and fluorescence
intensities followed by a fast decline. The analysis of a large
number of slices has shown that the different cotyledonary tissues
exhibit shifted time courses, beginning with the lower epidermis,
followed by the spongy parenchyma, especially in the neighborhood
of vascular bundles, and later by the palisade parenchyma.
The differences in organelle appearance and number at different
developmental stages were not due to an artificial covering of
available antibody binding sites, e.g. by changing concentrations
of cell constituents such as fat, etc. This possibility was ruled out
by the immunocytochemical labeling of mitochondria with antimMDH
antibodies which exclusively bind mMDH. Here, an
entirely different labeling pattern was seen. Figure 3 shows the
time course of fluorescence labeling in the storage parenchyma of
1- to 4-d cotyledons. By direct microscopic examination of fluorescent
organelles which permits a quick evaluation of several
focusing planes, the different appearance of mitochondria with
their smaller and often curled forms became evident. Most importantly,
there was already a significant and specific labeling in l-d
cotyledons (Fig. 3A) which increased to a high level at day 2 (Fig.
3B). Three- and 4-d cotyledons exhibited a further increase which,
however, was much smaller than the increase in the glyoxysomal
number during the same period (Figs. 3, C and D). The other
cotyledonary tissues (not shown) yielded comparable patterns of
mitochondrial development. These experiments have shown that
the immunohistochemical localization of organelles is feasible
even in early stages of fatty cotyledons. The organelle numbers
correspond roughly to the enzyme activities ofgMDH and mMDH
extracted from cotyledons of different developmental stages (17,
18) Ṫhe availability of monospecific antibodies against one cytosoic
form of MDH, cMDH I, which is representative for the cotyledons
and the embryo axis, raises the intriguing problem of the tissuespecific
localization of this isoenzyme. This isoenzyme was con-

1164 SAUTTER AND HOCK
Plant Physiol. Vol. 70, 1982
E~~~~~~.
A..
F
A
F.
.z
I.:-
K
1M: '.
:R K.:.
SFVz wr.
~~~~~Af. ....
FIG. 1. Immunofluorescent labeling of glyoxysomes (A, C) and mitochondria (B, D) by rabbit anti-gMDH antibodies (A, B) and anti-mMDH
antibodies (C, D), respectively, followed by FITC-labeled goat anti-rabbit antibodies, after sucrose density gradient centrifugation of a crude organelle
fraction (lO,OOOg pellet from watermelon cotyledons). Electronmicrographs from glyoxysomal (E) and mitochondrial (F) fractions are shown for control
purposes.
fined to the lower epidermis (Figs. 4, A, C, and D): it was entirely
lacking in the upper epidermis (Fig. 4B). The most intense fluorescence
was observed in l-d cotyledons (Fig. 4A), and it decreased
as germination progressed (Fig. 4, C and D). During the later
stages, the cytosolic localization of the isoenzyme became evident.
The various organelles were observed as dark spots. The fluorescence
label was never uniformly distributed within the tissue but
was confined to small groups of cells within the lower epidermis.
When the tissue of 2-d cotyledons was histochemically stained for
general MDH activity, the most intensive concentration of the dye
was again observed within small groups of the lower epidermis
which contained the differentiating guard mother cells, followed

IMMUNOHISTOCHEMICAL MDH LOCALIZATION
1165
'" A'
FIG. 2. Fluorescent immunohistochemical localization of glyoxysomes in frozen sections of dark-grown watermelon cotyledons by anti-gMDH
antibodies followed by FITC-labeled secondary antibodies, 1 d (A, B), 2 d (C, D), 3 d (E, F), and 4 d (G, H) after germination. Brightfield micrographs
of sections G and H are shown in plates I and K. Left column, palisade parenchyma; right column, spongy parenchyma. Bar, 10 ,tm.

1166 SAUTTER AND HOCK
Plant Physiol. Vol. 70, 1982
FIG. 3. Fluorescent immunohistochemical localization of mitochondria in frozen sections from spongy parenchyma of dark-grown watermelon
cotyledons by anti-mMDH antibodies followed by FITC-labeled secondary antibodies, 1 d (A), 2 d (B), 3 d (C), and 4 d (D) after germination. Bar,
IoOLm.
by the neighboring layers of the storage parenchyma and the
surroundings of the vascular bundles (Fig. 4F). It is tempting to
envisage a functional connection between the future stomatal
apparatus, which are not completely differentiated in these early
stages, and the lower epidermal groups stained for cMDH activity.
DISCUSSION
Immunocytochemistry is the method of choice for the intracellular
localization of isoenzymes when the enzyme reaction does
not permit discrimination. The use of gMDH and mMDH as
markers for glyoxysomes and mitochondria, respectively, was
demonstrated by immunofluorescent labeling of organelles previously
purified by sucrose density gradient centrifugation and
identified by isocitrate lyase or fumarase activity as well as electron
microscopy. By this technique, the cross-contamination of the two
organelle fractions was checked; it confirmed the high purity of
the glyoxysomal fraction in contrast to the mitochondrial fraction,
which contained some glyoxysomes in the case of high organelle
loads. Considering the sensitivity of the serological tests, it is likely
that antibodies to glyoxysomal membranes will detect determinants
of glyoxysomal origin -in mitochondrial fractions which were
previously judged as pure on the basis of marker enzymes (9).
This type of experiment, therefore, does not provide clues to the
question of common determinants in the glyoxysomal and the
outer mitochondrial membrane. For this purpose, ultrastructural
studies combined with immunochemical investigations are required.
The immunofluorescent labeling of gMDH and mMDH in cell
organelle fractions and tissue sections yielded a considerable
concentration of the dye in the organelles. These must have
remained relatively intact, since particles cut prior to fixation only
contribute to the background staining (C. Sautter, unpublished).
The penetration of primary and secondary antibodies through the
organelle envelopes without a prior leakage of the isoenzymes into
the surrounding areas is not a contradiction. The treatment with
low concentrations of formaldehyde fixed the original isoenzyme
location without causing a destruction of the antibody binding
sites. The subsequent passage through increasing sucrose concentrations
provided the necessary cryoprotection during deep freezing
and sectioning and also changed the permeability of the
organelle membranes during thawing of the cut sections. This has
been noticed before and was confirmed by electron microscopy

FIG. 4. Fluorescent immunohistochemical localization of cMDH I in frozen sections from the lower (A, C, D) and the upper epidermis (B) of darkgrown
watermelon cotyledons by anti-cMDH I antibodies followed by FITC-labeled secondary antibodies, 1 d (A, B), 3 d (C), and 4 d (D) after
germination. E, 4-d spongy parenchyma treated with control serum instead of antiserum. F, 2-d cotyledon, histochemically stained for MDH activity.
Bar, 10 ,tm (A-E); F, 100 ttm.
1167

1168 SAUTTER AND HOCK
Plant Physiol. Vol. 70, 1982
(C. Sautter, unpublished). The globular shape of the glyoxysomes
deviated from their original conditions in vivo where the organelles
appear as irregularly lobed and invaginated particles squeezed
into the spaces, especially those between oleosomes (14). It was
the infiltration by sucrose which allowed easy access to the matrix
of the glyoxysomes and caused this conspicuous change. Spherical
forms were also seen with other histochemical stains, e.g. malate
synthase (2).
Antibody labeling of the organelles within the tissues was
limited to cut cells. Again, the cryocutting of prefixed tissue
sections guaranteed the preservation of the original antibody
binding sites. The choice of thick cryosections with a diameter of
40 ,um and the use of epifluorescence equipment resulted from the
need to view large areas. The additional advantages of this procedure
lie in a uniform staining over the entire surface and in the
short time required. The results were available in less than 4 h
after chopping the cotyledons.
It was verified by different controls that the developmental
course of fluorescence labeling of gMDH in the cotyledonary
tissues reflected the true pattern of glyoxysomal development. The
lack of fluorescent labeling in early stages of seed germination was
not due to the inability of the anti-gMDH antibodies to reach
their binding sites, since the infiltration and labeling of the sections
with anti-mMDH did not present any problem. The labeling
patterns were in accordance with the results of serological isoenzyme
determination in crude extracts (17, 18). Moreover, in hypocotyls
which were known to lack gMDH, no glyoxysomes could
be detected by immunohistochemical techniques (not shown). On
the other hand, fluorescent labeling of cMDH I was restricted to
areas outside the organelles, which was to be expected in the case
of a cytosolic isoenzyme.
Evidence for the different temporal and spatial pattern of
glyoxysomal development in contrast to mitochondrial is based
on the analysis of a large number of slices, of which only a few
have been presented in this paper. This difference reflects the
functional separation of fat degradation and respiration. The
sequential appearance of glyoxysomes in the different cotyledonary
tissues is closely correlated to the fat degradation which is
known to be nonsimultaneous in the different parts of the cotyledons
(5). The increase in mitochondria during germination as
detected by fluorescent labeling ofmMDH appeared to be almost
simultaneous in the different tissues. This is in contrast to the
findings of Flinn and Smith (3) who analyzed in pea cotyledons
the distribution pattern of Cyt oxidase and succinate dehydrogenase
by enzyme specific staining. This difference might be due to
a contrast in fat and starch storing seedlings.
The immunochemical demonstration of cMDH I provides the
first tissue-specific localization of a cytosolic MDH. The choice of
this isoenzyme was governed by its occurrence in the embryo axis
and the cotyledons, whereas the cMDH II and IV are restricted to
the cotyledons (8). The association ofcMDH I with distinct groups
of epidermal cells, probably meristemoids giving rise to the stomata,
raises the question of the metabolic role of this isoenzyme.
Biochemical studies in our laboratory (4) have shown that in
contrast to cMDH I, the organelle-bound MDH isoenzymes are
synthesized as higher mol wt precursors which are processed
during the importation into their organelles. This event is probably
due to a posttranslational transport mechanism. It would be most
important to study the intracellular route of the isoenzymes from
the site of synthesis to their final destination. The immunochemical
localization at the electron microscopic level provides the
basis for these efforts, and significant contributions are to be
expected using this technique.
Acknowledgment--The competent technical assistance of Mrs. Danuta Weber
is gratefully acknowledged. We thank Dr. R. Youngman (Technical University of
Munich) for his comments on the manuscript.
LITERATURE CITED
1. BAUMGARTNER B, KT TOKUYASU, MJ CHRUSPEELS 1980 Immunocytochemical
localization of reserve protein in the endoplasmic reticulum of developing bean
(Phaseolus vulgaris) cotyledons. Planta 150: 419-425
2. BuRKE JJ, RN TRELEASE 1975 Cytochemical demonstration of malate synthase
and glyoxylate oxidase in microbodies of cucumber cotyledons. Plant Physiol
56: 710-717
3. FLINN AM, DL SMITH 1967 The localization of enzymes in the cotyledons of
Pisum arvense L. during germination. Planta 75: 10-22
4. GiETL C, B HOCK 1982 Organelle-bound malate dehydrogenase isoenzymes are
synthesized as higher molecular weight precursors. Plant Physiol 70: 484-487
5. HACKER M, H STOHR 1966 Der Abbau von Speicherfett in den Kotyledonen von
Sinapis alba L. unter dem Einfluss des Phytochroms. Planta 68: 215-225
6. HANKER JS, CJ KuSYK, FE BLOOM, AGE PEARSE 1973 The demonstration of
dehydrogenases and monoamine oxidase by the formation of osmium at the
sites of Hatchett's brown. Histochemie 33: 205-230
7. HOCK B 1969 Die Hemmung der Isocitratlyase bei Wassermelonenkeimlingen
durch Weisslicht. Planta 85: 340-350
8. HOCK B 1973 Kompartimentierung und Eigenschaften der Malatdehydrogenase-
Isoenzyme aus Wassermelonenkeimblattern. Planta 112: 137-148
9. HOCK B 1974 Antikorper gegen Glyoxysomenmembranen. Planta 115: 271-280
10. KAISER S 1978 Reinigung der cytoplasmatischen Malatdehydrogenase I aus
Wassermelonenkeimlingen. Master thesis. Ruhr University, Bochum, West
Germany
11. METTLER IJ, H BEEVERS 1980 Oxidation of NADH in glyoxysomes by a malateaspartate
shuttle. Plant Physiol 66: 555-560
12. NOVOTNY A 1969 Basic Exercises in Immunochemistry. Springer-Verlag, New
York, pp 3-5
13. SAUTrER C, HC BARTSCHERER, B HOCK 1981 Separation of plant cell organelles
by zonal centrifugation in reorienting density gradients. Anal Biochem 113:
179-184
14. SCHOPFER P, 0 BAJRACHARYA, R BERGFELD, H FALK 1976 Phytochrome-mediated
transformation of glyoxysomes into peroxisomes in the cotyledons of
mustard (Sinapis alba L.) seedlings. Planta 133: 73-80
15. TOKUYASU KT, SJ SINGER 1976 Improved procedures for immunoferritin labeling
of ultrathin frozen sections. J Cell Biol 71: 894-896
16. WALK RA, B HOCK 1976 Separation of malate dehydrogenase isoenzymes by
affinity chromatography on 5'-AMP-sepharose. Eur J Biochem 71: 25-32
17. WALK RA, B HOCK 1976 Mitochondrial malate dehydrogenase of watermelon
cotyledons: time course and mode of enzyme activity changes during germination.
Planta 129: 27-32
18. WALK RA, B HOCK 1977 Glyoxysomal malate dehydrogenase of watermelon
cotyledons: de novo synthesis on cytoplasmic ribosomes. Planta 124: 277-285
19. WALK RA, B HOCK 1977 Glyoxysomal and mitochondrial malate dehydrogenase
of watermelon (Citrullus vulgaris) cotyledons. II. Kinetic properties of the
purified isoenzymes. Planta 136: 221-228
20. WALK RA, S MICHAELI, B HOCK 1977 Glyoxysomal and mitochondrial malate
dehydrogenase of watermelon (Citrullus vulgaris) cotyledons. I. Molecular
properties of the purified isoenzymes. Planta 136: 211-220