cytokinin biosynthesis

Proc. Nati. Acad. Sci. USA
Vol. 81, pp. 5994-5998, October 1984
Biochemistry
T-DNA of Agrobacterium tumefaciens encodes an enzyme of
cytokinin biosynthesis
(crown gall/tumor-inducing plasmid/tmr gene/plant tumorigenesis)
D. E. AKIYOSHI*, H. KLEEt, R. M. AMASINOt, E. W. NESTERt, AND M. P. GORDON*
*Department of Biochemistry, SJ-70, and tDepartment of Microbiology and Immunology, SC-42, University of Washington, Seattle, WA 98195
Communicated by Edmond H. Fischer, June 27, 1984
ABSTRACT Phytohormone overproduction in crown gall
tumors is due to the expression of several T-DNA genes. The
data strongly suggest that the Imr gene (transcript 4) is responsible
for cytokinin overproduction by encoding dimethylallylpyrophosphate:AMP
dimethylallyltransferase (DMA transferase),
an enzyme directly involved in cytokinin biosynthesis.
Cell-free extracts of Escherichia coli strains containing the tmr
gene from pTiA6NC had DMA transferase activity. No activity
was present in the control strain containing only the plasmid
vector. The cytokinins synthesized were isopentenyladenine,
isopentenyladenosine, and isopentenyladenosine 5'-monophosphate.
DMA transferase activity was also detected in cloned
crown gall tumors incited by Agrobacterium tumefaciens wildtype
A6NC and a tms mutant. Enzymatic activity in cell-free
extracts of E. coli and tumors could be abolished by transposon
insertion within the tmr gene.
Infection of a wound site by Agrobacterium tumefaciens
strains containing Ti (tumor-inducing) plasmids results in the
neoplastic plant disease known as crown gall (1-5). In the
plant genome, the integrated region (T-DNA) of the Ti plasmid
expresses two phenotypic traits characteristic of transformed
cells: phytohormone autonomy and opine synthesis.
The T-DNA in tumors incited by octopine-type Ti plasmids
encodes eight transcripts (6). Five of these map within three
loci (tmr, tms, tml), which affect tumor morphology (7-11).
Transposon inactivation of the tms locus (transcripts 1 and 2)
or the tmr locus (transcript 4) results in tumors that proliferate
shoots or roots, respectively. Inactivation of the tml
locus (transcripts 6a and 6b) results in larger tumors on
certain plant species. The region of T-DNA containing these
loci is shown to be highly conserved in octopine and nopaline
plasmids (12-14). The nucleotide sequences of the tmr (15,
16) and tms (15, 17-19) genes have been recently determined
and have been shown to be highly conserved. The synthesis
of several proteins from the T-DNA region has been demonstrated
in Escherichia coli minicells and in coupled
transcription/translation systems of E. coli and A. tumefaciens
(20). Three of these proteins had molecular weights
identical to the predicted sizes of the tmr and tms gene
products, as determined from the nucleotide sequence.
Previous studies have also shown that the tms and tmr loci
are involved in auxin and cytokinin metabolism, respectively
(9-11, 21, 22). Tumors incited by tms mutants have decreased
endogenous auxin levels (21) and are auxin dependent
in culture (23, 24). Recently, Schroder et al. (25) and
Inze et al. (26) independently demonstrated that transcript 2
(tms) codes for indole-3-acetamide hydrolase, an enzyme
involved in the auxin biosynthetic pathway.
Several lines of evidence suggest a role of the tmr locus in
cytokinin metabolism. Tumors incited by tmr mutants have
significantly decreased cytokinin levels compared to wild-
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked "advertisement"
in accordance with 18 U.S.C. §1734 solely to indicate this fact.
5994
type tumors (21), and some tmr tumors have been shown to
require cytokinin for growth in culture (26).
The first step in cytokinin biosynthesis is the attachment of
the dimethylallyl side chain to AMP:
N
Ni
OH
HO4+OH2 ',o..
OH Kw)I
OH OH
AMP
DMA TRANSFERASE
?2H 5t H3
+ HO-

Biochemistry: Akiyoshi et al.
393 349 149 169
tma tmr (ml Ocs
5(1.20) 2(1.75) 4(1.25) 3(1.45)
7(0.73) 1(2.7) 6ab(0.92)
BamHI 19H II8 11291
2
EcoRi 2 A 7 111 24 |
pDQ64
PD0 14
pNW34D-7 -I _
FIG. 1. Locations of T-DNA loci and transcripts in the T-DNA
of pTiA6NC. The four identified genetic loci (tmr, tms, tml, ocs)
and the locations of the eight transcripts are shown (6, 9, 11, 34).
The polarity of each transcript is indicated by the direction of the
arrow and the transcript size (in kilobases) is shown. Transposon
(TnS) insertions used in this study are indicated at the top of the
figure. Solid bars below the map represent the plasmid clones containing
T-DNA fragments.
same direction as 5' -+ 3' polarity of the tmr gene, and
orientation II is defined as the opposite orientation. Digestion
of pUC18-EcoRI 7 (orientation I) with Kpn I followed
by religation produced a 4.2-kilobase EcoRI/Kpn I subclone
containing only transcripts 1 and 4 (pDQ14; Fig. 1). A 6.1kilobase
EcoRI/Sal I subclone (pDQ64; Fig. 1), containing
transcripts 4, 6a, and 6b, was generated by Sal I digestion
and religation to pUC18-EcoRI 7 (orientation II).
The plasmid pRA120 was constructed in the following
manner: The 5' end of the tmr gene was isolated as an Ava
II/BamHI fragment, which extends from the second codon
of the tmr gene to the BamHI 29/19 junction. The Ava II
overhang was filled in to produce a blunt end and the fragment
was ligated into pUC18 cut with Sma I/BamHI. This
generated an in-frame translational fusion between the a
peptide of 83-galactosidase and the 5' end of the tmr gene. The
tmr gene was then reconstituted by cutting this plasmid with
BamHI/Sal I, and a restriction fragment consisting ofBamHI
19 to the Sal I site of a TnS insertion in transcript 6a was
inserted. Thus, pRA120 contains the entire tmr gene linked
to the first 10 codons of the a peptide of f8-galactosidase. The
resulting protein is expected to have 10 additional amino
acids (Thr-Met-Ile-Thr-Asn-Ser-Ser-Ser-Val-Pro) at the
amino terminus of the tmr gene.
Construction of pRR290-EcoRI (pNW34D-7-1) and
pRK290-EcoRI 7::TnS (pNW34D-7-1::TnS) plasmids have
been described (34). Tn5 insertions 149 and 169 map within
the tmr locus, and insertion 393 maps within the tms locus.
Insertion 349 is a wild-type TnS insertion (Fig. 1) (11).
The cloned crown gall tumor lines (Nicotiana tabacum cv.
Xanthi nc or White Burley) were incited by A. tumefaciens
strains containing either pTiA6NC or pTiA6NC::Tn5 (34).
The tumor lines were grown under continuous light on solid
Murashige-Skoog medium (35) without phytohormones.
Preparation of Cell-Free Extracts. Bacterial cells (2 g, wet
weight) were harvested by centrifugation (4100 x g, 5 min,
40C), resuspended in 3-5 ml of buffer A [40 mM Hepes, pH
7.3/10 mM MgCl2/20 mM KCl/1 mM EDTA/0.5 mM
dithiothreitol/0.1 mM phenylmethylsulfonyl fluoride/20%
(vol/vol) glycerol] and sonicated 6 times (30 sec each) (Sonifier
Cell Disruptor, Heat System/Ultrasonics, Plainview,
NY). The cell lysate was diluted to 20 ml with buffer A,
centrifuged (12,000 x g, 30 min, 40C) to remove cell debris,
and solid ammonium sulfate (5.6 g/ml; 85% saturation) was
added over a 20-min period. After 90 min, the precipitate was
collected by centrifugation (12,000 x g, 60 minm 4C), dissolved
in 7 ml of buffer A, and dialyzed against buffer A
(1000 ml, 2 changes, 14 hr, 40C). The dialysate was clarified
by centrifugation (12,000 x g, 20 min, 40C), diluted with 100
IX6
Proc. Natl. Acad. Sci. USA 81 (1984) 5995
ml of buffer A containing 50 mM Hepes (1:1, vol/vol) and
applied to a small DEAE-cellulose column. The protein
fraction containing DMA transferase activity was eluted with
buffer A containing 0.3 M KCl and stored at - 70'C for
assay.
Cell-free extracts of tumor tissue (3 weeks old; 10 g, fresh
weight) were prepared as described (27) except XAD-4 was
omitted and the samples were dialyzed as described above
for the bacterial samples.
DMA Transferase Assay. Enzyme preparations from either
bacterial cells (equivalent to 0.5 g, wet weight, of cells) or
tumor tissue (equivalent to 2 g, fresh weight, of tissue) were
assayed in duplicate. The assay reaction (8 ml) contained
0.12 M Hepes, pH 7.3/0.16 M KCI/15 mM MgCl2/7 mM
KF/0.5 mM EDTA/0.25 mM dithiothreitol/0.05 mM phenylmethylsulfonyl
fluoride/10% (vol/vol) glycerol. The reaction
was started by the addition of 0.43 mM dimethylallylpyrophosphate
synthesized by the method of Cornforth
and Popjak (36) and [2-3H]AMP (74-148 kBq; 443
GBq/mmol; Amersham). After a 90-min incubation at 220C,
cytokinins were adsorbed onto a small octadecylsilica column,
eluted with methanol, and dried in vacuo.
Some samples were treated with alkaline phosphatase
prior to HPLC as follows: alkaline phosphatase (bovine
intestinal; 12 units; Sigma) in 1 ml of 50 mM Tris HCl (pH
9.2) and 20 mM MgCl2 was added to the dried samples and
incubated 15-18 hr at 370C. Cytokinins were adsorbed onto
a small octadecylsilica column, eluted with methanol, and
dried.
HPLC. Samples were subjected to HPLC as described
(27), with minor modifications. To resolve the radiolabeled
products of the cell-free extracts, samples were fractionated
on an octadecylsilica column (25 cm x 4.6 mm; Ultrasphere
ODS 5 um; Altex, Berkeley, CA) using the following gradient
program (flow rate, 1 ml/min): 12%-18% CH3CN over
10 min (curve 3, Waters Automated Gradient Programmer),
18%-23% over 12 min (curve 3), held at 23% for 6 min,
23%-30o over 2 min (curve 6), and then held at 30% for 2
min. The samples were applied in 100 ttl of CH3CN/0.1 M
triethylammonium acetate (60:40, vol/vol), fractions were
collected (0.5 min), and radioactivity was determined.
To estimate DMA transferase activity in bacterial and
tumor extracts, samples were dissolved in 80 ,ul of methanol
and chromatographed under isocratic conditions (0.1 M triethylammonium
acetate, pH 3.4/CH3CN) (79:21, vol/vol).
RESULTS
Identification of Cytokinins. Addition of dimethylallylpyrophosphate
and [3H]AMP to cell-free extracts of E.
coli containing either the plasmid vector or the tmr gene
resulted in the incorporation of radioactivity into cytokinins,
which were subsequently identified by HPLC. The strain
containing only the pUC18 plasmid had a single peak of
radioactivity at a retention time (3.5-4.5 min) corresponding
to the mobility of polar compounds including AMP, adenine,
and adenosine (Fig. 2 A and B). However, strains containing
the tmr gene had three additional radioactive species, which
were not found in the control strain without the tmr gene.
The retention times of these species corresponded to mobilities
of authentic standards of isopentenyladenosine 5'monophosphate
(15-16 min), isopentenyladenine (25.5-27
min), and isopentenyladenosine (27.5-28 min). A typical
radioactivity profile for strains containing the tmr gene is
shown in Fig. 2D.
Over 50% of the radioactivity incorporated into cytokinins
was present as isopentenyladenine. If the samples were
treated with alkaline phosphatase prior to HPLC, radioactivity
in the isopentenyladenosine 5'-monophosphate fraction
disappeared, but radioactivity in the isopentenyladenosine

5996 Biochemistry: Akiyoshi et al.
0
x
E
0
0
15 20
Time, min
FIG. 2. Radioactivity profile of E. coli cell-free extracts on
HPLC. Cytokinins were adsorbed onto octadecylsilica, eluted, and
fractionated on HPLC. One set of extracts (A and C) were treated
with alkaline phosphatase prior to HPLC. (A and B) E. coli control
strain containing plasmid vector; (C and D) E. coli strains containing
tmr gene. Retention times of cytokinin standards are indicated
by horizontal bars (-) in D: Ade, adenine; Ado, adenosine; AMP,
adenosine 5'-monophosphate; t-Z, trans-zeatin; t-ZR, transribosylzeatin;
c-ZR, cis-ribosylzeatin; iPA-P, isopentenyladenosine
5'-monophosphate; iP, isopentenyladenine; iPA, isopentenyladenosine.
region increased (Fig. 2C). The radioactivity in the isopentenyladenine
and isopentenyladenosine fractions can be precipitated
by isopentenyladenosine-specific antibodies (data
not shown). No radioactivity was detected in fractions corresponding
to the retention times of trans-zeatin, transribosylzeatin,
cis-zeatin, or cis-ribosylzeatin (Fig. 2 C and D).
Proc. Natl. Acad. Sci. USA 81 (1984)
Table 1. DMA transferase activity in E. coli strains and crown
gall tumors
Radioactivity in
iP/iPA region,*
cpm
NDt
10,100
Plasmid
pUC18
pRA120
pUC18-EcoRI 7
(orientation I)
pDQ14
pUC18-EcoRI 7
(orientation II)
pDQ64
pRK290
pRK290-tmr-149::TnS
pRK290-tmr-169::TnS
pRK290-EcoRI::TnS
(349)
Crown gall tumor lines
A6S/2t
tmr-149
tmr-169
tms-393
7,730
14,900
4,190
2,750
20
80
2,400
10,820
8,950
ND
450
31,830
The assay mixture contained cell-free extract from E. coli (0.5 g,
fresh weight) or crown gall tumor tissue (2 g, fresh weight), 0.43 mM
dimethylallylpyrophosphate and 74-148 kBq [3H]AMP. Incubation
was for 90 min at 220C.
*13 cpm = 1 fmol isopentenyladenine (iP)/isopentenyladenosine
(iPA) synthesized.
tND, not detected.
tN. tabacum cv. White Burley. Other tumor lines were incited on
N. tabacum cv. Xanthi nc.
DMA Transferase Activity in Bacterial Ceil-Free Extracts.
Having established the identity of the cytokinins made by
cell-free extracts of E. coli, DMA transferase activity was
estimated in several strains containing the tmr gene. Samples
were treated with alkaline phosphatase prior to HPLC and
were chromatographed under conditions whereby isopentenyladenine
and isopentenyladenosine had identical retention
times (11-12 min). The EcoRI fragment 7 from pTiA6NC,
an octopine-type plasmid, was cloned into pUC18 in both
orientations. DMA transferase activity was detected in both
strains containing this fragment, although activity was higher
in orientation I (Table 1). The EcoRI fragment 7 contains, in
addition to tmr, transcripts 1 (tms) and 6a and 6b (tml). Two
additional constructs were made, which removed either the
tms or tml transcripts. This eliminated the possibility that
one of these transcripts was responsible for DMA transferase
activity. Strains containing the EcoRI/Kpn I subclone
(pDQ14), which contains tmr and transcript 1 but not transcripts
6a and 6b, retained DMA transferase activity. Enzymatic
activity was also retained in strains containing the
EcoRI/Sal I subclone (pDQ64), which contains tmr and
transcript 1.
To demonstrate unambiguously that DMA transferase activity
was correlated with the tmr gene, a translational fusion
was constructed (pRA120) fusing the tmr structural gene to
the ,B-galactosidase a peptide encoded by pUC18. This fusion
deletes the initiation codon from the amino terminus of
the tmr protein and replaces it with the first 10 codons of pgalactosidase.
Cell-free extracts of E. coli containing pRA120
had DMA transferase activity equivalent to the strain containing
pUC18-EcoRI 7 (orientation I, Table 1).
The availability of E. coli strains containing Tn5 insertions
in the tmr gene made it possible to extend the correlation
between DMA transferase activity and the tmr gene. HB101
containing only pRK290 had no detectable DMA transferase
activity (Table 1). However, HB101 containing the EcoRI 7
fragment with a TnS insertion in a region that had no detect-

Biochemistry: Akiyoshi et al.
able effect on tumor morphology (Fig. 1), had activity comparable
to that observed for BNN103 strains containing the
tmr gene. Two tmr mutations, 149 and 169, encoded no
activity and 22% activity, respectively. Additional mapping
studies of insertion 169 have determined its location to be 25
+ 5 bases downstream from the 3' end of the open reading
frame.
DMA Transferase Activity in Crown Gall Tumor Extracts.
Finally, DMA transferase activity in cell-free extracts of
cloned tumor lines incited by tmr mutants was compared to
activity in wild-type tumors. Consistent with the level of
activity measured in the bacterial extracts, tumors incited by
tmr-149::TnS had no detectable DMA transferase activity,
while tumors incited by tmr-169::TnS had o5% of the activity
measured in wild-type tumors (Table 1). A tms-transcript
2 mutant Qms-393::TnS) had a 3-fold higher DMA
transferase activity than wild-type tumors.
DISCUSSION
Transposon mutagenesis defines three genetic loci (tmr, tms,
tml) in the octopine T-DNA that affect tumor morphology
(7-11). On tobacco stems, tms mutants and mr mutants
incite tumors characterized by shoot or root proliferation,
respectively. Tumors incited by tml mutants are unorganized,
like parental wild-type tumors, but larger than wildtype
tumors are observed on Kalanchoe leaves. The morphology
of these tumors results from a change in the hormonal
balance due to expression of the T-DNA and follows
the pattern originally described by Skoog and Miller for
normal tobacco tissue (37). Specifically, tmr tumors have
low cytokinin/auxin ratios, which favor root formation, and
tms mutants have high cytokinin/auxin ratios, which promote
shoot production (21). Tumors incited by tml mutants
have ratios similar to the wild-type unorganized tumors.
In this paper, we have presented strong evidence suggesting
that the tmr gene encodes the cytokinin biosynthetic
enzyme, DMA transferase, which attaches the dimethylallyl
side chain to AMP to form isopentenyladenosine 5'monophosphate.
Acquisition of DMA transferase activity in
E. coli was strictly correlated with the presence of the tmr
gene, which was expressed either from its own promoter or
from the pUC18 lac promoter. In one strain, enzymatic
activity was abolished by TnS insertion into the tmr gene. In
another tmr strain containing a TnS insertion between the 3'
end of the open reading frame and the polyadenylylation
signal, DMA transferase activity was decreased by almost
80%. No activity was detected in the strain containing only
pRK290. We could not detect DMA transferase activity in A.
tumefaciens containing pTiA6NC or pTiA6NC: :Tn5; although,
we could detect a similar activity in the nopaline
strain C58 (data not shown). This suggests that either the
octopine tmr gene is not expressed in A. tumefaciens or it is
expressed at levels too low to detect enzymatic activity.
In the present study, DMA transferase was detected in a
cloned octopine A6S/2 tumor line and in a tms-2 tumor line,
but activity was 1/20th that in tmr tumor lines. DMA transferase
activity has been previously reported in several cloned
octopine crown gall tumors (27) and in cytokinin-independent
tobacco tissue (28), but there is controversy as to its
presence in nontransformed tobacco tissue (27, 29).
The presence of the tmr gene in crown gall tumors is
correlated with overproduction of the cytokinins, trans-zeatin
and trans-ribosylzeatin. The tmr-directed synthesis of
DMA transferase provides an explanation for this overproduction.
In nontransformed tissue, the rate-limiting step
in cytokinin biosynthesis appears to be the synthesis of
isopentenyladenosine. In crown gall tumors, the production
of isopentenyladenosine is no longer rate-limiting, presumably
because of the presence of the tmr gene. However, in
Proc. Natl. Acad. Sci. USA 81 (1984) 5997
the tumors, increased isopentenyladenosine synthesis is not
reflected by an increase in the endogenous isopentenyladenosine
pool; instead, the newly synthesized isopentenyladenosine
is rapidly converted to trans-zeatin and transribosylzeatin.
Recently, transcript 2 of the tms locus has been shown to
code for indole-3-acetamide hydrolase, an enzyme that hydrolyzes
indole-3-acetamide to indole-3-acetic acid (25, 26).
The addition of a-naphthalene acetamide can promote tumor
formation by tms-1 mutants (26), suggesting that gene 1 may
encode an enzyme involved in the synthesis of the acetamide
derivative in the auxin biosynthetic pathway. In addition,
sequence homology has been observed between tms-1 and
the adenine binding region of Pseudomonas fluorescens phydroxybenzoate
hydroxylase (18).
Thus, there is accumulating evidence indicating that the T-
DNA codes for genes directly involved in auxin and cytokinin
biosynthesis. This would provide a molecular explanation
for the ability of crown gall tumors to grow in culture
without the presence of phytohormones.
We thank J. Peschon for valuable technical assistance and Dr.
R. 0. Morris for the iPA antibody. The work was supported by
Standard Oil Company (Indiana) and grants from the American
Cancer Society (NP-336A) and National Institutes of Health (ROI
CA12015-12). D.E.A. was supported by a postdoctoral traineeship
from the National Institutes of Health (T32 GM07187). H.K. was
supported by a postdoctoral fellowship from Standard Oil Company
(Indiana). R.M.A. was supported by postdoctoral fellowships from
the Damon Runyon-Walter Winchell Foundation and National Institutes
of Health.
1. Bevan, M. W. & Chilton, M.-D. (1982) Annu. Rev. Genet. 16,
357-384.
2. Kahl, G. & Schell, J., eds. (1982) Molecular Biology of Plant
Tumors, (Academic, New York), 615 pp.
3. Depicker, A., Van Montagu, M. & Schell, J. (1983) in Genetic
Engineering of Plants: An Agricultural Perspective, eds.
Kosuge, T., Meredith, C. P. & Hollaender, A. (Plenum, New
York), pp. 143-176.
4. Hooykaas, P. J. J. & Schilperoort, R. A. (1984) Adv. Genet.
22, 210-283.
5. Nester, E. W. & Kosuge, T. (1981) Annu. Rev. Microbiol. 35,
531-565.
6. Willmitzer, L., Simons, G. & Schell, J. (1982) EMBO J. 1,
139-146.
7. Ream, L. W., Gordon, M. P. & Nester, E. W. (1983) Proc.
Natl. Acad. Sci. USA 80, 1660-1664.
8. Holsters, M., Silva, B., Van Vliet, F., Genetello, C., De Block,
M., Dhaese, P., Depicker, A., Inze, D., Engler, G., Villarroel,
R., Van Montagu, M. & Schell, J. (1980) Plasmid 3, 212-230.
9. Ooms, G., Hooykaas, G., Moolenaar, G. & Schilperoort, R. A.
(1981) Gene 14, 33-50.
10. Leemans, J., Deblaere, R., Willmitzer, L., De Greve, H.,
Hernalsteens, J. P., Van Montagu, M. & Schell, J. (1982)
EMBO J. 1, 147-152.
11. Garfinkel, D. J., Simpson, R. B., Ream, L. W., White, F. F.,
Gordon, M. P. & Nester, E. W. (1981) Cell 27, 143-153.
12. Depicker, A., Van Montagu, M. & Schell, J. (1978) Nature
(London) 275, 150-152.
13. Chilton, M.-D., Drummond, M. H., Merlo, D. J. & Sciaky, D.
(1978) Nature (London) 275, 147-149.
14. Engler, G., Depicker, A., Maenhaut, R., Villarroel-Mandiola,
R., Van Montagu, M. & Schell, J. (1981) J. Mol. Biol. 152,
183-208.
15. Barker, R. F., Idler, K. B., Thompson, D. V. & Kemp, J. D.
(1983) Plant Mol. Biol. 2, 335-350.
16. Lichtenstein, C., Klee, H., Montoya, A., Garfinkel, D., Fuller,
S., Flores, C., Nester, E. & Gordon, M. (1984) J. Mol. Appl.
Genet., in press.
17. Heidekamp, F., Dirkse, W. G., Hille, J. & van Ormondt, H.
(1983) Nucleic Acids Res. 11, 6211-6223.

5998 Biochemistry: Akiyoshi et al.
18. Klee, H., Montoya, A., Hordyski, F., Lichtenstein, C., Garfinkel,
D., Fuller, S., Flores, C., Peschon, J., Nester, E. &
Gordon, M. (1984) Proc. Natl. Acad. Sci. USA 81, 1728-1732.
19. Sciaky, D. & Thomashow, M. F. (1984) Nucleic Acids Res. 12,
1447-1461.
20. Schroder, G., Klipp, W., Hillebrand, A., Ehring, R., Koncz,
C. & Schroder, J. (1983) EMBO J. 2, 403-409.
21. Akiyoshi, D., Morris, R., Hinz, R., Mischke, B., Kosuge, T.,
Garfinkel, D., Gordon, M. & Nester, E. (1983) Proc. Natl.
Acad. Sci. USA 80, 407-411.
22. Joos, H., Inze, D., Caplan, A., Sormann, M., Van Montagu,
M. & Schell, J. (1983) Cell 32, 1057-1067.
23. Binns, A. N., Sciaky, D. & Wood, H. N. (1982) Cell 31,
605-612.
24. Binns, A. (1983) Planta 158, 272-279.
25. Schroder, G., Waffenschmidt, S., Weiler, E. W. & Schroder,
J. (1983) Eur. J. Biochem. 138, 387-391.
26. Inze, D., Follin, A., Van Lijsebettens, M., Simoens, C., Genetello,
C., Van Montagu, M. & Schell, J. (1983) Mol. Gen.
Genet. 194, 265-274.
Proc. Nati. Acad. Sci. USA 81 (1984)
27. Akiyoshi, D. E. (1984) Dissertation (Oreon State University,
Corvallis, OR).
28. Chen, C.-m. & Melitz, D. K. (1979) FE.S Lett. 107, 15-20.
29. Nishinari, N. & Syono, K. (1980) Z. Pflanzenphysiol. 99,
383-392.
30. Young, R. A. & Davis, R. W. (1983) Proc. Nail. Acad. Sci.
USA 80, 1194-1198.
31. Boyer, H. W. & Roulland-Dussoix, D. (1969) J. Mol. Biol. 41,
459-472.
32. Miller, J. H. (1972) Experiments in Molecular Genetics (Cold
Spring Harbor Laboratory, Cold Spring Harbor, NY) p. 21.
33. Narrander, J., Kempe, T. & Messing, J. (1983) Gene 26,
101-106.
34. Grfinkel, D. J. & Nester, E. W.- (190) J. Bacteriol. 144,
732-743.
35. Murashige, T. &-Skoog, F. (1962) Physiol. Plant. 15, 473-497.
36. Cornforth, R. H. & Popjdk, G. (1969) in Methods in Enzymology,
ed. Clayton, R. B. (Academic, NY) Vol. 15, pp. 385-387.
37. Skoog, F. & Miller, C. 0. (1957) Symp. Soc. Exp. Biol. 11, pp.
118-131.