Plant Cell Reports (1995) 14:471-476 Plant Cell Reports �9 Springer-Verlag 1995

Efficient genetic transformation of red raspberry, Rubus ideaus L.

Helena Mathews, W. Wagoner, C. Cohen, J. Kellogg, and R. Bestwick

Agritope Inc., 8505 S.W. Creekside, P1. Beaverton, OR 97005, USA

Received 24 August 1994/Revised version received 7 October 1994 - Communicated by I. K. Vasil

Abstract. We have developed an efficient transformation system for red raspberry (Rubus ideaus L.) using Agrobacterium mediated gene transfer. Using this system we have successfully introduced a gene that encodes an enzyme, S-adenosylmethionine hydrolase (SAMase), in raspberry cultivars Meeker (MK), Chilliwack (CH) and Canby (CY). Leaf and petiole explants were inoculated with disarmed Agrobacterium tumefaciens strain EHA 105 carrying either of two binary vectors, pAG1452 or pAG1552, encoding gene sequences for SAMase under the control of the wound and fruit specific tomato E4 promoter. Primary shoot regenerants on selection medium were chimeral containing both transformed and non- transformed cells. Non-chimeral transgenic clones were developed by iterative culture of petiole, node and leaf explants, on selection medium, from successive generations of shoots derived from the primary regenerants. Percent recovery of transformants was higher with the selection marker gene hygromycin phosphotransferase (hpt), than with neomycin phosphotransferase (nptll). Transformation frequencies of 49.6%, 0.9% and 8.1% were obtained in cultivars Meeker, Chilliwack and Canby respectively from petiole explants using hygromycin selection. Genomic integration of transgenes was confirmed by Southern hybridization. Transgenic plants from a total of 218 independent transformation events (161 MK, 4 CH, 53 CY) have been successfully established in s0il.

Key words: genetic transformation, S- adenosylmethionine hydrolase gene, red raspberry

Abbreviations: ACCO: amincocyclopropane-1- carboxylic acid oxidase; AS: acetosyringone; BA: 6- benzylaminopurine; CH: cultivar Chiltiwack; CY: cultivar Canby; cv: cultivar; hpt: hygromycin phosphotransferase; IBA: indolebutyric acid; MK: cultivar Meeker; npt H: neomycin phosphotransferase; SAMase: S- adenosylmethionine hydrolase; TDZ: Thidi~uron (N- phenyl-N'-l,2,3-thidi~ol-5-ylurea).

Introduction The potential for cultivar improvement through traditional breeding methods is limited in Rubus due to the heterozygous nature of the species and its severe inbreeding depression. It can take 20 to 30 years to breed a Rubus cultivar with a characteristic transferred from unimproved germplasm (Jennings and McNicol 1991). Gene transfer technology allows introduction of new traits in proven cultivars without disrupting their otherwise desirable genetic constitutions. However, the recalcitrant nature and poor transformation rate of woody species have placed them far behind the herbaceous group in the application of genc liansfer methods (Schuennan and Dandekar 1993). Though there have been several reports of plant regeneration in Rubus species (Cousineau and Donnelly 1991; Fiola et al. 1990; McNicol and Graham 1990; De Novoa 1992; De Novoa and Conner 1992; Vysotski and Upadyshev 1992), transgenic Rubus is still a rarity. De Novoa and Conner (1991) did an extensive study on the host specificity of Agrobacterium to different Rubus genotypes. The first report on transgenic red raspberry plants using a disarmed Agrobacterium strain came from the Scottish Crop Research Institute (Graham and McNicol 1990, Graham et al. 1990) using marker genes uidA and nptll. They reported that the nptll gene was not suitable for Rubus species since they!: could not obtain regenerants in the presence of the selection agent, kanamycin. However, Hassan et al. (1993) obtained transgenic blackberry (Rubus sp.) plants using kanamycin selection which is perhaps the first report of transgenic Rubus with definite proof of transformation. They reported transgenic blackberry, xesistant to chlorsulfuron herbicide, at a frequency of 0.25%. Although there have been attempts to produce insect resistant plants using the cowpea trypsin inhibitor (CpTi) gene (McNicol and Graham 1989), there has not been any published report on the integration of a functional gene in red raspberry. Here we report the incorporation of a gene of possible economic importance, S-adenosylmethionine hydrolase (SAMase) into red raspberry cultivars Meeker, Chilliwack and Canby. S-adenosylmethionine (SAM) is the metabolic precursor of 1-aminocyclopropane-l-carboxylic acid (ACC), the proximal precursor to ethylene. SAMase lowers the

Correspondence to: H. Mathews

472

available substrate for ACC synthase and, therefore, reduces a plant's ability to produce ethylene (Good et al. 1994). Because SAM is a required cofactor in many methyl donor reactions and serves as a substrate for polyamine biosynthesis, constitutive expression of SAMase has a potentially adverse effect on plant metabolism. Regulated expression of SAMase to those tissues that are producing ethylene effectively eliminates this caveat and was used to create low ethylene tomatoes with increased post harvest shelf life (Good et al. 1994). We expect a similar strategy for the control of ethylene biosynthesis may prove useful in raspberries. The present paper focusses on development of an efficient transformation protocol that resulted in generating transgenic plants of three different floricane cultivars of raspberry.

Materials and Methods

Explant s o u r c e : Leaves and petioles were excised from proliferating shoot cultures maintained on modified MS (Murashige and Skoog 1962) medium supplemented with 1 rag/1 BA and gelled with 0.2% phytagel. Petioles were cut into 4-6 mm segments. Leaf blades 4-5 mm in length were used as whole explants or they were cut into transverse halves when blades were >6 mm in size. Leaf explants were cultured with the adaxial surface in contact with the medium. Bacterial strain and binary vectors Agrobacterium tumefaciens strain EHA105 contained the disarmed supervimlent plasmid pTiBo542 in the C58 chromosomal background (Hood et ai.1993). The plasmid vectors containing the SAMase gene were constructed using the backbone of pGPTV binary vectors where the marker genes had been divergently organized for efficient expression and could be easily removed or replaced (Becker et al. 1992). The binary vector, pAG1452 had the hpt gene for resistance to the antibiotic hygromycin under the nos promoter located near the left border and the SAMase gene driven by the wound and fruit specific tomato E4 promoter (Cordes et al. 1989) located near the right border (Fig. 1). The binary vector pAG1552 differed from pAG1452 by having the nptH marker gene in place of hpt gene (Fig. 1).

ku :~:

- - - -~pAno~SAMase [ E4~,~ I pr, o, I hpt IpAg;'J-I b" pAG-1452 BR ~ ........ ~ eL

~pAr~s~SAMase I E4p~o U Pnos I not/I IPAg71 - ~ pAG-1552 BR ~ ~ BL

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

kb I l I I I I l I I 1

Fig. 1. Restriction maps of the Agrobacterium tumefac&ns binary vectors used in the present study. The Eco RI and Hind llI sites shown in pAG1452 are present at the same location in pAGI552. The Agrobacterium Ti right and left borders are abbreviated B R and BL, respectively. All the remaining genetic elements are explained or referenced in the Material and Methods.

Co.cultivation of explants with Agrobacterium A freshly grown single colony of Agrobacterium was inoculated into 30 ml of MG/L (Garfinkel and Nester 1980) liquid medium supplemented with 50 mM acetosyringone, pH 5.6, and grown on a shaker at 200 rpm for overnight (16-18 hrs). The bacterial suspension had an average

count of 0.5-0.6 x 109 cclls/ml at the start time of co-cultivation with plant tissues. The petioles and leaf explants soon after excision were soaked in Agrobacterium suspension. After 30-60 minutes, these were transferred to flasks containing liquid MS medium and kept on shaker at 100 rpm. The density of explants was about 100-120 segments in about 30 ml of medium in 125 ml flasks. After 2 days of co-cultivation, the explants were rinsed with liquid medium, blotted and plated on regeneration medium with selection. Tissue culture media and growth conditions The shoot regeneration medium consisted of modified MS medium supplemented with 0.1 rag/1 IBA, 0.1-l.0 rag/1 TDZ and 10 mg/l silver nitrate. For cultivar Meeker, 3% sucrose was replaced by 3% D-glucose as the carbon source, qhe pH of the medium was adjusted to 5.8 before

golfing with 0.2-0.25% phytagel. Depending on the plasrnid strain the regeneration medium contained antibiotics carbenicil]in (500 mg]l) and geneticin (3/5-25 rag/l) or hygromycin (10-30 rag/l) for screening o f transformed shoots. The level of antibiotics chosen for each cuhivar and stage of explant was based on the in vitro responses of control explants to different selection agents (data not shown). The culture medium was

autoclaved at 1200C at 1.1 kg.cm 2 except for the antibiotics and silver

nitrate which were filter sterilized. Cultures were kept at 250C with 16- hr photoperiod provided by white fluorescent light at an average intensity

-2 -1 of 50 ]~rnol m s . Observations were recorded every 3-4 weeks and cultures were transferred to fresh medium of the same composition with appropriate changes in the level of antibiotics. Soon after co-cultivation the explants received the lowest level of selection mentioned above. As the explants underwent proliferation in culture, the selection level was gradually increased based on the rate of dedifferentiated tissue on the explant. The selection level was elevated to a maximum of 25 rng/l of geneticin or 30 mg/1 hygrornycin in the case of explants with prolific callus growth. The putatively transformed shoot regenerants were isolated and cultured on shoot proliferation medium with selection at 15- 20 mg/1 of geneticin or hygromycin. Rooting medium for transgenic shoots contained haft strength MS medium supplemented with 3% sucrose (or D-glucose for cv. Meeker), 0.05 mg/l IBA, 300 rng/1 earbeniciUin and 10 rag/1 hygrornycin or geneticin. Leaf, petiole and nodal explants were cultured in petri plates with 40 ml of regeneration/proliferation medium. There were 20 petiole segments and -10 leaf segments per plate. Shoot explants (6-9) for multiplication/rooting were cultured on phytatrays with 120 ml of the proliferation or roofing medium. T r e a t m e n t of primary shoot regenerants and recovery o f transgenic clones Shoot regenerants were isolated and cultured on shoot proliferation medium containing 15-20 mg]l geneticin or hygromycin. Leaves, petioles and nodal segments were isolated from shoots which withstood selection, and cultured on regeneration or proliferation medium with 20 mg/1 geneticin or hygromyicn. This iterative process of reculture of excised tissues from regenerants was continued until no part of the shoots necrosed or bleached under selection pressure. Shoots were considered as fully transformed only after they passed the above criteria. Such shoots were multiplied on proliferation medium for generating clonal plants of different transformation events. Tissue samples of such plants were used for molecular confirmation of transformation events. Transformation efficiency Frequency of transformation was defined as the number of explants which gave antibiotic resistant callus and/or shoot regenerants relative to the total number of co-cultivated explants. Transformation frequency was expressed in percentage points. Establishment of transgenic shoots~plants in the green house Individual shoots from profusely proliferating shoots on selection medium were isolated for root induction or were transferred directly to soil. For direct transfer the shoot base was dusted with rooting mix (Hormex powder #3) before placing in potting mix. Phytatrays with rooted plants were kept in the greenhouse with loosened lids for 2-4 days, followed b y transfer to soil after rinsing off the media with water. Southern hybridization Genomic DNA was isolated from leaf tissue of transgenic and non- transformed control plants following the method of Dellaporta et aL (1983). The DNA was digested with either Eco RI or Eco RI and HindHI, electrophoresed and transferred to a nylon membrane (Oncor,

Gaithersburg, MD), hybridized with 32p radiolabeled probes and autoradiographed using standard procedures (Sambrook et al. 1989). The hybridization probes used were a 550 bp fragment containing the SAMase gene or a 1050 bp fragment containing the genes for SAMase and a putative raspberry aminocyclopropane-l-carboxylic acid oxidase (ACCO) gene (manuscript in preparation) . Transgene copy number was estimated by comparing the ACCO and SAMase band intensities using a high-resolution flat bed scanner and the NIH-Image analysis program on an Apple Macintosh computer.

R e s u l t s

Response of Agrobacterium co-cultivated explants on selection medium cv. M e e k e r : M e e k e r was transformed with the binary vector pAG1452 in the disarmed A . t u m e f a c i e n s strain EHAI05. After co-cultivation petioles and leaf explants were cultured on regeneration medium with 10 mg/1 hygromycin. In 3-4 weeks dedifferentiation of cut edges was observed. At the end of the second transfer period (6 weeks), about 24.0% of the leaf and 33.3% of the petiole explants (data not shown) underwent shoot regeneration.

473

The number of petiole explants undergoing shoot regeneration (Fig. 2A) increased with subsequent transfers. The petiole explants gave 49.6% shoot regeneration at the end of 4 months of culture on selection medium (Table 1). The shoot regeneration frequency of leaf explants decreased to 15.9% at the end of 4 months since many of the early regenerants underwent complete necrosis after transfer to proliferation medium with hygromycin. Frequency of transformed callus from both petiole and leaf explants was significantly higher than that of shoot regeneration and was capable of profuse growth in the presence of 30 mg/l hygromycin. Control leaf and petiole explants did not withstand even 10 mg/l hygromycin and were completely necrosed by the second culture period. The percent recovery of transformed shoots was higher with petiole than with leaf explants.

T A B L E 1

, of Transformation obtained , cvs. Meeker , Chil

Cultivar IExpt i Plasmid 113 I Explant I # lnitial Explants Tra.Events in soil

Meeker RT18 pAG1452 Petiole 244 74.1 49.6 88

pAG1452 Leaf 675 31.3 15.9 73

Chillliwack i RTI8 pAG1452 Petiole* 214 37.0 0.9 2

pAG 1452 Leaf* 430 14.4 0.7 2

Canby RT24 pAG 1552 I Petiole 467 91.4 2.6 I0

pAG1552 Leaf 314 0.3 0.3 1

Canby RT25 pAG1552 Petiole 190 95.8 3.7 4

pAGI552 Leaf* 223 9.4 0 0

Canby RT28 pA(i 1552 Petiole 308 66.9 5.2 5

pAGI552 Leaf 322 43.5 3.1 8

Canby RT27 pAG1452 Petiole 308 51.6 8.1 19

pAGI452 Leaf 288 30.6 2.4 6

Note: Culture period t20 tO 125 days * -50% explants lost due to contamination.

cv. Chi l l iwack:As shown in table 1, cv. CH treated with A. tumefaciens strain EHA 105 containing the binary vector pAG1452, gave a transformation frequency of 0.9 and 0.7% from petiole and leaf explants, respectively. Initially, 25% of the petiole and 9.8% leaf explants (data not shown) underwent shoot regeneration during 4-6 weeks culture on regeneration medium with selection at 10 mg/1 hygromycin. But most of these initial shoot regenrants completely necrosed on excision and culture on proliferation medium with selection. Unlike cv. Meeker, explants of

Chil l iwack did not undergo shoot regeneration on subsequent transfers. In addition, many of the leaf explants were lost due to hypersensitivity to Agrobacterium during the co-cultivation periodl cv.Canby:Cultivar Canby was transformed with both binary vectors, pAG1552 and pAG1452. Petiole and leaf explants after co-cultivation with pAG1552 were cultured on selection medium with 5 and 10 mg/l geneticin, respectively. The explants co-cultivated with pAG1452 were cultured on regeneration medium with 10 and 20 rag/1 hygromycin for petiole and leaf explants, respectively. In 3-4 weeks the cut edges of explants showed dedifferentiation along with spontaneous shoot regeneration in some of the explants. Most of the shoot differentiation took place over a period of 4 months on selection medium. The responses of four independent experiments in cv. Canby, three with plasmid pAG1552 and one with pAG1452, are summarised in Table 1. As in the other two cultivars, the number of explants which gave transformed callus was significantly higher than the number of explants which gave rise to

transformed shoots. The transformed callus grew uninhibited at 25 mg/l geneticin or 30 mg/1 hygromycin, depending on the transgene nptll or hpt while growth of control non-a'ansformed tissues was completely inhibited at much lower concentrations (5-10 mg/1). Frequency of shoot regeneration was higher with explants selected on medium with hygromycin while the frequency of transformed callus was higher with explants selected on medium with geneticin. Petiole explants gave higher rates of transformation, both in terms of callus and shoot regeneration, irrespective of selection agent. Treatment of primary shoot regenerants and recovery of transgenie clones During the periodic transfer of leaf and petiole explants to fresh medium, the differentiated shoots, about 10-15 mm in size, were excised and individually cultured on shoot p r o l i f e r a t i o n m e d i u m with 15-20 mg/l geneticin/hygromycin depending on the strain pAG1452 or pAG1552 used in the experiment. Mainly four types of responses were observed: (a) complete necrosis of the shoot; (b) arrested shoot development with pale green leaves; (c) vigorous growth of single shoot ;(d) multiple shoot proliferation. Among these, the first two categories were considered as non-transformed escapes or with very few transformed cells and were discarded. Vigorously growing single and multiple shoots were considered as putative transformants and used for further analysis to check whether the whole shoot was uniformly transformed. Leaves, petioles and nodal explants of the primary regenerants were excised and cultured on regeneration/proliferation medium with 20 mg/1 of the selection agent geneticin/hygromycin. The majority of the explants underwent callusing and shoot regeneration, or bud growth in the case of nodal segments, while some of the explants necrosed and responded just like control, non-transformed tissues on selection. When the presence of such non-transformed regions were identified in the putative transformants, the process of explant isolation and reculture on selection medium was repeated with the secondary shoot regenerants. Shoots were considered as fully transformed only when isolated tissue explants were capable of growth on selection medium. Transgenic shoots were then cultured on proliferation medium with selection for generating clonal plants of each event. Southern analysis DNA isolated from transgenic plants were subjected to Southern blot analysis to confirm the presence of the transgenes and to characterize the structure of the integrated DNA. DNA was digested with Eco RI alone or in conjunction with Hind III. Eco RI cleaves the region between the borders of the binary vector once and therefore produces junction fragments that can be recognized by the SAMase probe. Junction fragments can provide information on the transgene copy number and the number of independent integration events. The EcoRI/Hind III double digest produces a single fragment internal to the T- DNA borders that can be detected using a SAMase hybridization probe. Figure 3A and 3B show Southerns of transgenic cvs MK with pAG-1452, and CY with pAG- 1552, respectively. In general, a variety of integration patterns are detected including single, double, and triple insertions and multiple insertions at a single site. Fig. 3A, lanes 3 to 13 show the junction fragments from eleven transgenic events in cv. MK.

474

Fig. 2A. Petiole explants of red raspberry cv. Meeker with callus and shoot regenerants on regeneration medium with 30 mg/1 hygromycin

Fig. 2B. Transgenic plants of cv. Meeker established in greenhouse

Fig. 3. Southern blots of DNA isolated from non-transformed cvs. Meeker (A; lanes 2 and 14) and Canby (B; lane 7) compared to DNA from independent transgenic lines in the same cultivars. DNAs were digested with either Eco RI alone (A; lanes 2-13, B; lanes 1-6) or, in matching sequence, with both Eco RI and Hind III (A; lanes 14-25, B; lanes 7-13). The blots were probed with a 32P-labeled probe consisting of either SAMase and a putative raspberry ACC oxidase gene (A) or with SAMase alone (B). The blot shown in B was also probed with the SAMase/ACC- oxidase probe which resulted in the expected 1.4 kb band appearing in each DNA track (not shown). Molecular weight markers in kb are shown along the left margins. Exposures were for 8 hours.

Most are single integration events with two double (lanes 5 and 13) and one triple (lane 7) events. Lane 8 contains several bands including one that is smaller than the predicted 4.2kb Eco RI to left border fragment size of pAG-1452 (see Fig. 1), indicating integration of partial fragments. The appearence of faint signals in lanes 9 and 10 may be indicative of plants that are chimeral for the transgene, despite the tissue culture efforts to select nonchimeral plants. Interestingly both these events gave profuse multiplication on selection medium like other events. In Fig. 3B, the Eco RI digested DNA produced bands that seem to fall into two intensity categories suggesting a rather complex integration structure. However, all of the integration events contained an intact 2.0 kb Eco RI to Hind II! fragment indicating an accurate T-DNA structure (Fig. 3B, lanes 7-13). One line had multiple integrations ( see Fig. 3B, lanes 6 and 13) estimated at greater than 20 copies of the SAMase gene. Estimates of the SAMase gene copy number were also made using a hybridization probe consisting of both the SAMase gene and a putative raspberry ethylene forming enzyme gene or ACC oxidase (ACCO) which has one allelle (data not shown). The ACCO probe produces a band intensity equivalent to two copies of the gene. It also serves to confirm that the DNA has been completely digested by Eco RI and provides a measure of the relative amounts of DNA in each lane. Because the ACCO and SAMase genes are on a single plasmid from which the double hybridization probe is made, the intensity differences between the Eco RI/Hind III SAMase bands and the ACCO bands can be compared and were used to quantitate the SAMase gene copy number. In Figure 3A there is good agreement between the number of integrations predicted by the junction fragment analysis and the intensity of the 2.0 kb Eco RI/Hind III SAMase bands (lanes 14-25). Lanes 3, 4, 6 and 11 indicated single integration events which was confirmed in lanes 15, 16, 18, and 23 respectively. The double integration events seen in lanes 5 and 13 and the triple integration event in lane 7 were also confirmed by proportional band intensities in lanes 17, 25 and 19 respectively. Rooting and transplantation Individual shoots isolated from multiple shoot complex were cultured on rooting medium with selection at 10 mg/1 geneticin or hygromycin. All the rooted shoots were successfully transplanted to soil. When shoots were directly transplanted to soil from proliferation medium, the survival rate was 60-70%. Occasional abnormalities among transgenics One of the events in cv. Canby (RT24 expt) gave shoots which were stunted in growth and bristle type in appearence. In cv. Meeker an interesting kind of shoot differentiation was observed occasionally on leaves still attached to the shoots growing on proliferation medium. Shoot buds capable of further growth arose from all over the adaxial surface of leaves. This phenomenon was not observed on the control non-transformed proliferation cultures while stunted and bristle type shoots were occasionally observed among the control regenerants in non-selection medium. However only the normal looking transgenic plants were transferred to soil. Figure 2B shows transgenic plants established in greenhouse. Transgenic raspberry plants from a total of 218 independent transformation events (161 MK, 4 CH, 53

475

CY) have been established in soil for further evaluation of growth parameters and expression of the introduced traits. These plants were indistinguishable from control non- tissue cultured plants.

Discussion In our studies, petioles proved to be more efficient explants than leaves in all the cultivars independent of whether the selection agent was hygromycin or geneticin (see table 1). Similar explant specificity was also reported in alfalfa transformation (Du et al. 1994) where more consistent results were obtained with petiole rather than leaf and stern explants. We found hpt to be more efficient than nptll for recovering transformants. Our experiments on the evaluation of different selection markers had shown that hygromycin and geneticin gave clear cut-off points compared to kanamycin in arresting the growth of control raspberry tissue (data not shown). Nevertheless, we obtained stable transgenic plants, at a frequency of 2.6-5.2% from petiole explants using the selection marker nptH gene. This is contrary to Graham et al's report (1990) that nptH gene is unsuitable for selection in Rubus and transformants had to be screened based on GUS assay alone. It is possible that geneticin is a more effective antibiotic for Rubus than kanamycin, although both antibiotics are aminoglycoside compounds that are detoxified by the neomycin phosphotransferase enzyme. In lettuce transformed and non-transformed seeds were easily identified using 10-15 mg/1 geneticin while progeny assay was not possible even with 250 mg/l kanamycin in the germination medium (Michelmore et al. 1987). In addition to the types of selection agents used, the very selection scheme in terms of gradual increase in concentrations depending on tissue responses was an important element for our success in transformation.

It is very important to produce non-chimeral transgenics in plants like raspberry where the genetic constitution is maintained by vegetative means and recovery of pure transformants through seed segregation is not practical. Since shoot meristems normally arise from more than one cell and not necessarily from cells of clonal origin (Poethig 1989), the formation of chimeral plants may be more of a rule than an exception in transformation experiments, specially where plant regeneration occurs via organogenesis. This is well illustrated in recent reports on tobacco (Oono et al. 1993, Schmulling and Schell 1993), flax (Dong and McHughen 1993a and 1993b) and melon (Dong and McHughen I99 I). It is critical to establish that the antibiotic resistant plants contained the introduced transgenes. Tissues tested at random from any of the transformation events were positive on Southern analysis both for npt/hpt (data not shown) as well as for SAMase gene. Most of the transgenic plants we investigated have intact single or double integration events. Of interest is the rather complex pattern of integration structures seen in Figure 3B. The presence of bands with different intensities is difficult to reconcile based on the available evidence. The simple explanation that the Eco RI digest was incomplete is not supported by the blot obtained with the ACCO]SAMase probe (not shown) which showed a complete digestion of the ACCO gene to a 1.4kb fragment. It is possible that the Eco RI sites in the vicinity of the transgenes were resistant to cleavage but this

476

suggestion is not corroborated by the Eco RI/Hind III digested DNA which produces a single 2.0 kb SAMase band, as expected of a complete digest. It is also interesting that these multiple integration patterns appear to be specific for the Canby cultivar and the npt II selectable marker gene.

Acknowledgements We thank Stan Gelvin for providing us with the Agrobacterium tumefaciens strain EHA105 and Valerie Dewey, Ethel Lupulesa, Debbie Schuster, Don Wanek and Richard Harding for their excellent technical contributions. We are also grateful for the excellent assistance of Jana Dieter and Katherine Lawrence in the preparation of manuscript and figures.

References Becket D, Kemper E, Schell J, Masterson R. (1992) Plant Mol Biol

20:1195-1197. Cordes S, Deikman J, Margossian LJ, Fischer RL. (1989) The Plant Cell

1:1025-1034. Cousineau JC, Donnelly DJ. (1991) Plant Cell, Tissue and Organ Culture

27:249-255. De Novoa COY. (1992) New Zealand Natural Sciences 19:79-86. De Novoa COY, Conner AJ. (1991) J Genet and Breed 45:359-368. De Novoa COY, Conner AJ. (1992) New Zealand Journal of Crop and

Horticultural Science 20:471-476. Dellaporta SL, Wood J, Hicks JB. (1983) Plant Mol Biol Rep 1:19-21. Dong JZ, McHughen A. (1991) Plant Cell Reports 10:555-560. Dong JZ, McHughen A. (1993a) Plant Science 88:61-71. Dong JZ, McHughen A. (1993b) Plant Science 91:139-148. Du S, Erickson L, Bowley S. (1994) Plant Cell Reports 13:330-334 Fiola JA, Hassan MA, Swartz HJ, Bors RH, McNicols R. (1990) Plant

Cell Tissue and Organ Culture 20:223-228. Gaffinkel DJ, Nester EW. (1980) J Bacteriol 144:732-743. Good X, Kellogg JA, Wagoner WA, Langhoff D, Matsumura W, Ferro

AJ, Bestwick RK. (1994) Plant Molecutar Biology in press: Graham J, McNicol RJ. (1990) Acta Horticulturae 280:517-522. Graham J, McNicol RJ, Kumar A. (1990) Plant Cell Tissue and Organ

Culture 20:35-39. Hassan MA, Swartz HJ, Inamine G, Mullineaux P. (1993) Plant Cell

Tissue and Organ Culture 33:9-17. Hood EE, Gelvin SB, Melchers I_S, Hoekema A. (1993) Transgenic

Research 2:208-218. ffennings DL, McNicol RJ. (1991) Plant Breeding Abstracts 61:755-758. McNicol RJ, Graham J. (1989) Acta Horticulturae 262:41-46. McNicol RJ, Graham J. (1990) Plant Cell Tissue and Organ Culture

21:45-50. Michelmore R, Marsh E, Seely S, Landry B. (1987) Plant Cell Reports

6:439-442 Murashige T, Skoog F. (1962) Physiol Plant 15:473-497. Oono Y, Suzuki T, Toki S, Uchimiya H. (1993) Plant Cell Physiol

34(5):745-752. Poethig S. (1989) Trends Genet. 5: 273-277. Sambrook if, Fritsch EF, Maniatis T. Molecular cloning - a laboratory

manual. 2nd ed. Cold Spring Harbour Laboratory Press: NewYork, 1989.

Schmulling T, Schell J. (1993) Plant Mol Bid 21:705-708. Schuerman P, Dandekar AM. (1993) Scientia Horticulturae 55:101-124. Vysotskii VA, Upadyshev MT. (1992) Soviet Plant Physiology 39:375-

380.