Plant Cell Reports

Plant Cell Reports (1990) 8:512-516

© Springer-Verlag1990

Improved efficiency of the walnut somatic embryo gene transfer system Gale H. MeGranahan, Charles A. Leslie, Sandra L. Uratsu, and Abhaya M. Dandekar Department of Pomology, University of California, Davis, CA 95616, USA Received August 21, 1989/Revised version received November 22, 1989 - Communicated by G. C. Phillips

SUNMARY. An Agrobacterium-mediated gene transfer system which relies on repetitive embryogenesis to regenerate transgenic walnut plants has been made mere efficient by using a more virulent strain of Agrobacteriumandvectors containing genes for both kanamycin resistance and beta-glucuronidase (GUS) activity to facilitate early screening and selection. Two plasmids (pCGNTO01 and pCGN7314) introduced individually into the disarmed Agrobacterium host strain EHAI01 were used as inoculum. Embryos maintained on medium containing 100 mg/l kanamycin after co-cultivation produced more transformed secondary embryos than embryos maintained on kanamycin-freemedium. Of the 186 GUS-positive secondary embryo lines identified, 70% were regenerated from 3 out of 16 primary embryos inoculated with EHA101/pCGN7314 and grown on kanamycin- containing medium, 28% from 4 out of 17 primary embryos inoculated with EHAI01/ pCGN7001 and grown on kanamycin medium, and 2% from one out of 13 primary embryos inoculated with EHAIOI/pCGN7001 but not exposed to kanamycin. Because kanamycin inhibits but does not completely block new embryo formation in controls, identification of transformants formerly required repetitive selection on kanamycin for several months. Introduction of the GUS marker gene allowed positive identification of transformant secondary embryos as early as 5-6 weeks after inoculation. DNA analysis of a representative subset of lines (n=13) derived from secondary embryos confirmed transformation and provided evidence for multiple insertion events in single inoculated primary embryos.

INTRODUCTION

An Agrobacterium-mediated gene transfer system that targets embryogenic cells of walnut somatic embryos has been previously reported (McGranahan e t a[., 1988). One attribute of this system is that it allows regeneration and multiplication of transformants from species, including walnuts and many other woody perennials, that can be induced to form repetitively embryogenic somatic embryos (Tulecke, 1987) but that have never been regenerated from adventitious buds, a common procedure for Agrobacterium-mediated gene insertion in annual crop species (Horsch et al., 1985, McCormick et al., 1986). Another attribute is the potential for elimination of chimeric transformants if secondary embryos originate directly from single cells as in walnut (Polito et al., 1989). Although an inoculated set of embryos (E0 Offprint requests to." G. I-I. McGranahan

generation) could give r i s e to chimeric secondary embryos (E 1 generation) depending on their developmental stage at the time of inoculation, the next embryo generation (E 2) which o r i g i n a t e s from s i n g l e celts of the E1 enToryos should be free of chimeras. Thus s e l e c t i o n in the E2 or l a t e r generation is a means to screen out chimeric transformants. Avoidance of chimeras is e s s e n t i a l in ctonatty propagated woody species with long generation i n t e r v a l s l i k e walnut because the eventual goal is to insert useful genes i n t o commercial, c t o n a t t y propagated cultivars without Loss of genetic integrity. The major drawback of the system has been the Low efficiency of transformation and the time and labor involved in identifying transformants. Because kanamycin inhibits but does not completely block new embryo formation in untransformed embryos, identification of transformants previously required repetitive selection on kanamycin-containing media for several months. In this paper we describe methods to improve transformation efficiency by applying early selection pressure for transformed cells and by including a second marker gene that is easily and reliably detected in transgenic somatic embryos. In addition, we have used a strain of Agrobacterium that has been shown to behypervirulent (Hood et al., 1986). We used a bacterial 9ene encoding the enzyme beta-glucuronidase (GUS) (Jefferson et al., 1987) in addition to the gene encoding resistance to the antibiotic kanamycin in a binary vector system in Agrobacterium strain EHAI01 (Hood et al., 1986). Our specific goals were to determine if embryo exposure to kanamycin shortly after inoculation with the Agrobacterium enhanced the frequency of transgenic embryos recovered and if the GUS 9ene product was expressed and reliably detected in transformed walnut tissue.

NATERIALS AND METHODS

The two binary vectors pCGN7001 and pCGN7314 used in these experiments were provided by L. Comai (Catgene Inc., Davis). The vector pCGNTO01 is a 29 kb ptasmid of which almost 18.8 kb is derived from the large Sat I to Bgl !I fragment of pVCKI02 (Knauf and Nester, 1982), the remainder being the T-DNA region. At e i t h e r end of the T-DNA region a r e DNA fragments containing the Left and r i g h t border sequences (bp 625-1617 and

513 13774-15208 from pTiA6; Barker et at., 1983). Between the border sequences are the following: a chimeric gene encoding bacterial APH(3')II (kanamycin resistance) containing the CaMV 35S promotor (bp 7144-7569 of CaMV; Gardener et at., 1981) and the polyadenylation region of transcript 7 (tr 7, bp 2920-2396 of pTiA6; Barker et at., 1983); a chimeric GUS gene encoding beta-glucuronidase activity (coding region of uid A gene of Escheria coli, Jefferson, 1987) containing the mannopine synthase promotor and the 3' polyadenylation region of pTiA6 (bp 20128-20857 and 18474-19239; Barker et at., 1983); and a 2.4 kb fragment of DNA from plasmid pPHIJI (Bam HI to Hind Ill) which includes a gene encoding gentamycin resistance to allow for selection of the plasmid in bacteria. The second vector pCGN7314, is about 18 kb in length. The plasmid backbone contains two origins of replication, one (CoIEI) from pBR322 and the other from pRiA4, and the gene for gentamycin resistance described earlier. The T-DNA region has left and right border sequences from pTiA6 (bp 625-2122 and 13990-14273; Barker e t a L., 1983) and the following: a chimeric gene encoding bacterial APH(3')II containing the mannopine synthase promotor and 3' polyadenylation sequence from pTiA6 (bp 20128-20857 and 18474-19239; Barker et at., 1983); a chimeric GUS gene encoding beta-glucuronidase activity containing an enhanced CaMV 35S promoter (bp 7445-6492 of CaMV with an enhancer region, bp 7343-7070, inserted at sequence 7343 to create a promoter with duplicated enhancer regions) and the 3' polyadenylation of pTiA6 described above. Plasmids pCGN7001 and pCGN7314 were each mated (Ditta e t a_L., 1980) separately into the disarmed A. tumefaciens strain EHAI01 (Hood et at., 1986a, 1986b) creating a binary vector system (i.e., EHAIOI/pCGNTO01 and EHAI01/ pCGN7314) capable of gene transfer into plants. The construction of EHAI01 which contains a disarmed derivative of the Ti plasmid pTiBo542 has been previously described (Hood et at., 1986a). EHAI01 by itself without the T-DNA plasmids was used as a control in the plant inoculations. The Agrobacterium strains EHAI01, EHAIOI/pCGN7001 and EHA101/pCGN7314 were grown at 26-28°C in 523 medium (Rodriguez and Tait, 1983) overnight, centrifuged (500~ x g, 10 min.) and resuspended to a density of 2.5 x 10~ cells/ml in DKW medium (Driver and Kuniyuki, 1984; McGranahan et a t . , 1987) containing 100 #M acetosyringone. These resuspended celts were used as inocu Lum. Each Agrobacterium strain (EHAI01 control,EHA101/pCGN7001, EHAI01/ pCGN7314) was used to inoculate approximately 50 small (2-5 mm) white intact embryos of a repetitively embryogenic culture line designated "SU-2". This line was derived from the immature cotyledon of an open-pollinated seed of the walnut cultivar SunLand and had been maintained in culture for over three years using standard methods (Tulecke and McGranahan, 1985; McGranahan e t a L., 1987). Embryos were soaked 10-15 min in the inoculum, blotted lightly on sterile filter paper and plated on solid basal DKW medium (Driver and Kuniyuki, 1984; McGranahan et at., 1987) containing 100 /~M acetosyringone. After 48 h of co-cultivation, embryos were rinsed in basal DKW containing 500 mg/l cefotaxime and transferred to plates of the same medium solidified with Gelrite (Merck). After 24 h half of the embryos in each treatment were transferred to solid DKW ~ i u m containing both cefotaxime (500 mg/l) and kanamyc~n (100 mg/l), the other half remaining on medium without kanamycin, resulting in 6 different treatments of 25 embryos each. Embryos in each treatment were maintained in the dark at room temperature and transferred to their appropriate fresh

mediumevery 7-10 d throughout the experiment. Embryos selected f o r shoot c u l t u r e s were placed on the same media under cool white f l u o r e s c e n t l i g h t s (55-110 #E'm-2"sec-1) f o r germination and then micropropagated using standard techniques (McGranahan et a l . 1987). Eight of the o r i g i n a l embryos (E generation) in each 0 treatment were s a c r i f i c e d f o r e v a l u a t i o n of GUS expression a f t e r one and 2 1/2 wk. The remaining E0 embryos were maintained and secondary embryos (E 1) were removed from them f o r e v a l u a t i o n at one to two wk i n t e r v a l s from 5 1/2 to 16 wk from i n o c u l a t i o n . Expression of the GUS gene was evaluated by an X-Gluc (5-bromo-4-chloro-3 i n d o l y l glucuronide) histochemical assay based on the method of Jefferson (1987). X-Gluc is a substrate which produces a l o c a l i z e d blue p r e c i p i t a t e in c e l l s expressing the GUS gene. X-Gluc substrate s o l u t i o n was prepared by d i s s o l v i n g X-Gluc (Clontech) to a 0.3% v / v s o l u t i o n in dimethylformamide. This was diluted to I mM X-Gluc with 100 mM sodium phosphate buffer (pH 7.0) containing 0.006% Triton-X 100 and 0.5 n~l K+Fe Cyanide. E embryos were tested for expression of GUS by I. cuttlng a small plece from each embryo and immersing it in X-Gluc solution in multi-well plates at room temperature or 37C. E0 embroys were tested intact. Embryos or pieces were observed for blue color at intervals beginning 10 min after immersion for up to 24 h. If the piece developed a distinct blue color the embryo from which it was cut was multiplied on the appropriate medium as a subcLone for retesting, DNA assays and production of transgenic plants. Stem segments from germinated embryos were assayed for GUS activity using the same protocol as for embryos; leaf tissue for GUS assays was obtained from shoots etiofated in the dark for one week. To confirm absence of bacterial contamination a subset of GUS-positive embryos was cultured in 523 liquid medium (Rodriquez and Tait, 1983) on a rotary shaker for one wk and observed daily for the presence of bacteria. Embryo DNA was isolated and analyzed by Southern blot analysis as previously described (Dandekar et at., 1988, McGranahan et at., 1988). The Southern blots were hybridized sequentially with two probes, one for detecting the right border and the other for detecting the chimeric kanamycin resistance gene. The probe used to detect the kanamycin resistance region has been previously described (Dandekar e taL., 1988, McGranahan et at., 1988). To detect the right border sequences a 567 bp Eco RI fragment from pCGN1532 (K. McBride, Calgene Inc., unpublished) containingbp13990-14273of pTiA6 (Barker et al.,1983) was used as a probe. Apart from the 283 bp border region this fragment contains 284 bp of DNA from pUC18 (bp 396 to 680). Processing of T-DNA occurs at position 14062 (Atbright e t a_~[.,1987) leaving only a 72 bp (13990-14602, Barker et al.,1983) homology region with the right border of T-DNA from pCGNTO01. The right border of T-DNA from pCGN7314 has 356 bp region of homology upon processing, that would include in addition to the 72 bp border region described above, a 284 bp region from pUC18. DNA fragments used as probes were obtained by removing the appropriate DNA fragments from low malt agarose gets. The DNA fragments were separated from the agarose by phenol extraction followed by alcohol precipitation as described by Ausubel et at., 1988. Purified DNA fragments were then labelled using the random primed reaction (Feinberg and Vogelstein, 1983 and Feinberg and Vogelstein, 1984) using a Boehringer Mannheim kit. Hybridization, washing and autoradiography were as

514 described earlier (Dandekar et at., 1988, McGranahan e t a L., 1988).

RESULTS AND CONCLUSION

Eoembryos tested in X-Gtuc one wk after inoculation had abundant bacteria on their surfaces which interfered with the detection of transformed cells. Thecefotaxime rinse before plating was designed to reduce this problem but was unsuccessful. E0 embryos tested at 2 I/2 wk stilt had some bacterial c~ntamination and we concluded that the visualization of transformation on the E0 embryos was not feasible if the bacteria usod also express betagtucuronidase activity. GUS-positive E I embryos were easily detected within 4 h of exposure to the X-Gtuc substrate. The distinct blue color first could be observed on cut surfaces or areas damaged in handling (Fig. I). With longer periods of exposure, the blue color was evident at the cotyledon tips where the tissue is thinnest. In most cases after 8 h the entire embryo piece was distinctly blue (Fig. I). The blue from the transformed plant tissue could be distinguished from that caused by the bacteria if present. The blue pigment resulting from beta-gtucuronidase activity in the bacteria was diffuse on the embryo surface, tended to be a lighter shade and became apparent after a longer exposure. When bacteria were present, the X-Gtuc solution itself became cloudy and blue over the 24 h period; the solution surrounding clean embryos remained clear. Testing of E I embryos began at 5 I/2 wk after inoculation and continued at I to 2 wk intervals through week 16 (Table I). Several of the E0 en~ryos failed to survive through 5 wk, particularly in the kanamycin treated control. In that treatment the 3 embryos that did survive produced no E I embryos. Over the the 10 wk testing pariod 186 E I embryos expressed betagtucuronidase a c t i v i t y and were maintained as d i s t i n c t tines. Of these, 130 were derived from 3 E0 embryos inoculated with EHAlO1/pCGN7314and exposed to k~namycin, 53 from 4 EN embryos inoculated with EHAIO1/pCGN7001 and exposed to Eanamycin, and 3 from one Eoembryo inoculated with EHAI01/ pCGN7001 but not exposed to kanamycin. No embryos inoculated with the control EHAI01 produced GUS positive embryos. Kanamycinexposure enhanced transgenic embryo production; 21.2% (7/33) of the exposed embryos produced transgenic embryos, whereas only 3.4% (1/29) of the unexposed embryos did. Of the GUS positive E1 embryos produced, almost 100%(183/186) were derived from kanamycin treated embryos. The percent of GUS positive E1 embryos emerging from transformant producing E0 embryos increased with time in the presence of kanamycin (Table 2). From the GUS a c t i v i t y data we were able to i d e n t i f y those embryos that were transformed but i t was not clear whether the GUS-positive E I embryos derived from a single E0 embryo were the result of multiplication at a single site of insertion and therefore products of one transformation event or whether they resulted from several to many independent transformation events on the surface of the E0 embryo. In order to distinguish between these two possibilities DNA was isolated from 5 E I embryo lines derived from different E0 embryos (Fig. 2) as well as 8 E I embryo lines derived from a single E0 embryo (Fig. 3). The DNA was digested with the restriction endonuctease Eco R! which cuts near the right border within the T-DNA and into the flanking plant DNA, yielding unique fragments corresponding to each integration event. This enzyme also yields internal T-DNA fragments of ~4.5kb in

pCGN7001 and ~0.9 kb in pCGN7314 containing the coding region of the APH(3')II enzyme. Southern blots of the DNA were f i r s t hybridizod with a right border (RB) probe (Fig. 2A and 3A). After autoradiography the blots were washed free of the RB probe andreh~/oridized with an APH(3')II probe (Figs. 2B and 3B). When hybridized with the RB probe each of the lines originating from d i f f e r e n t E0 en~ryos displayed a unique pattern of fragments confirming separate transformation events. The patterns indicated that from one (GUS 20, 255, and 55; Fig. 2A) to three (GUS 78; Fig. 2A) copies of T-DNA had been incorporated in the embryos. The fragment about 3 kb tong that appeared in a l l lanes was due to hybridization of an endogenous sequence in walnut DNA and therefore was not counted. The differences in intensity of hybridization observed in the RB hybridizations in Fig. 2A were due to differences in the region of homology of the RB probe with respect to T-DNA from each of the vectors as described in the methods section. The RB probe had a greater homology (356 bp, see Methods) to T-DNA from pCGN7314 and therefore gave a m e r e intense hybridization (Fig. 2A, lanes d and e; GUS 255 and 55) as compared to the corresponding pCGN7001 where the region of homology was only 72 bp (Fig. 2A lanes a,b and c; GUS 78,8, and 20). In contrast, the APH(3')II probe hybridized to the two different T-DNA fragments in an identical manner with the result that the single copy insertions from pCGN7314 (Fig. 2B lanes d arid e; GUS 255 and 55) are identical in intensity to the single copy insertion from pCGNTO01 (Fig. 2B Lane c; GUS 20). The single copy insertions could be distinguishedbytheir lower intensity of hybridization when compared with two copies (Fig. 2B lane b; GUS 8) and 3 copies (Fig. 2B lane a; GUS 78). The blots in Fig. 3A and B represent the DWA of E I lines obtained from a single E0 embryo inoculated with EHAI01/ pCGN7001. At least 4 of the 8 lines testod displayed unique patterns (Fig. 3A lanes b, c, e and h), 2 had simiLiar patterns (Fig. 3A lanes g and h) and for three there was no detectable hybridization (Fig. 3A lanes a, d and f). The unique patterns confirmed independent insertion events from which E I embryos were derived (Fig. 3A). Variation in band intensity (Fig. 3B) again correspanded to the copy number of T-DNA molecules inserted per genome. The san~oles GUS6, GUS16 and GUS33 that did not display detectable hybridization with the right border probe, displayed GUS activity and have a single T-DNA as judged byhybridization with the APH(3')II probe (Fig. 3B lanes a, d and f). A passible explanation for the lack of hybridization in the three samples could be due to further toss of the 72 homology region with the probe (see methods) caused by processing of T-DNA during its insertion and incorporation into walnut genomic DNA. The san~ole GUS12 possibly containod two insertions and these ran very close together around 2 kb (Fig 3A lane b). This could explain the more intense hybridization observed with the APH(3')II probe (Fig 3A, lane b). The band above themajor hybridization band in lane b was due to p a r t i a l digestion. T h e s e results indicated that individual E1 embryos generated from the surface of a single inoculated E0 embryo may come from d i f f e r e n t transformodcetts and~epresent separate transformation events. As the GUS-pasitive secondary c~nbryo lines multiplied they were retested for GUS activity and examinod for the presence of contaminating bacteria. Of the 33 lines retested (representing 7 E0 embryos) 32 were found to express GUS activity. The negative Line derived from E0 embryo ID # I was reproducing slowly and the original pasitive reading may have been due to

515

I

[

E036

o

o

o

~D (D

0

(D (D

e

b

C

d

f

g

e

h

Eo56 -

-

]

co

co

co

g gggg

co

~

co

(o

b

c

d

e

f

cj

h

gg~= a

kb

kb

9.56.74.3-

25.7 9.56.74.5-

2.~2.0-

2520-

0.6-

0.6-

25.7 -

RIGHT-BORDER PROBE

APH (5')11

PROBE

Fig. 3 Southern blots of DNA from8 transformed embryo lines all derived from inoculation of a single embryo with EHAIOI/pCGNTO01 showing results of hybridization with right border probe (A) and hybridization with APH(3')!! probe (B).

Fig. I GUS expression in walnut embryos transformed with EHAIOI/pCGNTO01 after 2 (a), 6 (b) and 24 (c) hours and untransformed embryo after 24 hours (d) in X-Gluc solution (bar = Imm).

LO

0 CO

D

D

LO D

LO GO i~ CO

D

(.9

kb

J

2

~2,,,

a

b

c

d

o

e

kb

25.7-

25.7 -

9.56.74-.3-

9.56.74.5-

2,3-

2.52.0-

2.0-

a

GO CO ~ CO

0 OJ (;3 ~ (.9

0

0

b

c

LO ~ OJ 09 ~ ~

CO D CD

d

e

Fig. 4 GUS expression in shoot from germinated embryo transformed with EHAIOI/pCGN7001 after 24 hour in XGtuc solution (right) and untransformed shoot treated similarly (teft)(bar = Icm).

TABLE 1. Frequency of GUS-positive E1 embryos over a 10 week p e r i o d r e l a t i v e t o the number of E0 walnut somatic embryos i n o c u l a t e d and m u l t i p l i e d .

0.6-

0.6-

Vector RIGHT BORDER PROBE

Kanamycin medium

E embryos E embryos GUSpositive 0 ~reduced inoculated E I embryos (no.) (no.) (no.) (%)

APH(5')II- PROBE

Fig. 2 Southern blots of DNA from 5 transformed embryo lines derived from different original embryos showing results of hybridization with right border probe (A) and hybridization with APH(3')I] probe (B). Lines EO11/GUS78 , EO26/GUS8 and En36/GUS20 (lanes a-c) resulted fFom inoculations with EHAIOI/pCGN7001, and EO43/GUS255 and EO47/GUS55 (lanes d and e) from EHA101/pCGN7314.

pCGNT314

+

16 16

403 299

130 0

32 0

pCGN7001

+

17 13

331 416

53 3

16 I

EHAI01

+

3 15

0 654

0 0

0 0

516 TABLE 2. Proportion of GUS-positive E1 embryos from E0 walnut somatic embryos that produced transformants according to time from inoculation. E0 Embryo ID I 43 47 9 11 32 36 26

Vector pCGN7314 pCGN7314 pCGN7314 pCGNTO01 pCGN7001 pCGN7001 pCGN7001 pCGN7001

T 0 T A L S (%)

Kanamycin medium + + + + + + +

6 7 I/4 = 0/3 0/20 --1/40 --......... --0/25 0/13 --3/14 13/16 0/I I/9

5/92 5

14/53 26

Weeks from inoculation TOTALS 9 10 12 13 14 16 n . . . . . . . . . . . . . . . . . . I/7 0/22 0/17 3/5 3/11 4/5 4/5 14/85 2/38 32/78 11/16 7/18 22/31 40/48 115/269 I/I . . . . . . . . . . . . I/I --1/29 --0/6 ...... 1/60 1/17 0/11 ...... 0/4 0/I 1/46 7/16 11/22 12/43 --4/6 0/4 50/121 2/11 0/10 0/I 0/4 --0/22 3/58

12/104 11

45/168 27

26/65 40

10/39 26

30/46 65

% 14 16 43 100 02 02 41 05

44/80 186/647 55 29

a Number of GUS-positive E I en~oryos/Number of E I embryos harvested

our testing the only positive sector of a chimeric E I embryo. The tests for bacterial contamination were all negative, i.e. we were unable to detect viable bacteria after 7-8 days of culture. Shoots have been regenerated from 7 secondary embryo lines obtained from two E 0 embryos. Leaves and stem segments of all shoots tested have expressed theactivity. GUS expression in the leaves is less striking than in the embryos loecause of the green pigmentation but in chlorophyll deficient leaves the pigment is easily detected (Fig. 4). As with the embryos the blue color first became apparent at the cut surfaces but the entire leaf surface colored after 24 h, or more, of exposure. In summary, early kanamycin exposure enhanced transgenic embryo recovery and the GUS marker served as an efficient, reliable and easily detected marker for transformation. Screening for GUS expression in the E I embryo generation reduced the time and work involved in maintaining large embryo populations for screening in later generations as was recommended in our previous paper (McGranahan e t at., 1988). Screening E I embryos should be followed by selection in later generations as chimeric E I embryos do occur. Although it appears from our small sample (n=8) that most transgenic E I embryos represent unique transformation events, DNA analysis of a larger population of E I en~oryos from a single E0 is needed to confirm this. This gene transfer system is suitable for for species which are susceptible to infection by Agrobacterium tumefaciens and are repetitively enforyogenic from single cells near the embryo surface. It canbe used to transform commercially important cultivars only if the embryogenic cultures are genetically stable and are derived from maternal tissue of the selected cultivars.

Acknowledgments. This work was done while the senior author was employed by USDA/ARS. The authors thank Luca Comai arid Tom Gradziel for critical review of this manuscript; Don Edwards for photographs; John Preece for manuscript preparation; and the Walnut Marketing Board for partial financial support. The use of product names or vendors does not imply endorsement by the USDA.

LITERATURE CITED

Albright LM, Yanofsky MF, Leroux B, Ma D, Nester EW (1987) J Bacteriol 169:1046-1055 Ausubel FM, Brent R, Kingston RE, Moore DD, Seid~n JG, Smith JA, Struhl K (1987) Current Protocols in Molecular Biology. Vol I. Greene Publishing Associ ates and Wiley-lnterscience, New York Barker RF, Idler KB, ThonNoson DV, Kemp JD (1983) Plant Mot Biol 2:335-350 Dandekar AM, Martin LA, and McGranahan GH (1988) J Amer Soc Hort Sci 113:945-949 Ditta G, Stanfield G, Corbin D, Helsinki DR (1980) Proc Nat[ Acad Sci USA 77:7347-7351 Driver JA, Kuniyuki AH (1984) Hortscience 19(4):507-509 Feinberg AP, Vogelstein B (1983) Anal Biochem 132:6-13 Feinberg AP, Vogetstein B (1984) Anal Biochem 137:266267 Gardner RC, Howarth AJ, Hahn P, Brown-Luedi M, Shepard RJ, Messing J (1981) Nucl Acid Res 9:2871-28&B Hood EH, Chilton WS, Chillon M-D, Fraley RT (1986) J Bacteriol 168:1283-1290 Hood EH, Helmer GL, Fratey RT, Chilton M-D (1986) J Bacteriol 168:1291-1301 Horsch RB, Fry JE, Hoffman NL, Eichholtz D, Rogers SG, Fraley RT (1985) Science 227:1229-1231 Jefferson RA (1987) Plant Mol Bio Rep 5:387-405 Jefferson RA, Kavanagh TA, Bevan MW (1987) EMBO J 6:3901-3907 Knauf VC, Nester EW (1982) Plasmid 8:45-54 McCormick S, Niedermeyer J, Fry J, Barnason A, Horsch R, Fraley R (1986) Plant Celt Reports 5:81-84 McGranahan GH, Driver JA, Tulecke W (1987) In: Bonga JM, Durzan DJ (eds) Cell and tissue culture in forestry, Vol 3. Martinus Nijhoff, Boston, pp 261-271 McGranahan GH, Leslie CA, Uratsu SL, Martin LA, Dandekar AM (1988) Bio/Technology 6:800-804 Polito VS, McGranahan GH, Pinney K, Leslie CA (1989) Plant Cell Reports 8:219-221 Rodriguez RL, Tait RC (1983) Recon~inant DNA Techniques. Benjamin/Cummings, Menlo Park, CA Tulecke, W (1987) In: Bonga JM, Durzan DJ (eds) Cell and Tissue Culture in Forestry. Vol 2. Martinus Nijhoff, Boston, pp 61-91 Tulecke W, McGranahan GH (1985) Plant Science 40:53-67