Transgenic Research10: 105–112, 2001.© 2001Kluwer Academic Publishers. Printed in the Netherlands.

105

Genetic transformation of major wine grape cultivars of Vitis vinifera L.

P. Iocco1, T. Franks1,2,3 & M.R. Thomas1,∗1CSIRO Plant Industry, Hartley Grove, Urrbrae, SA 5064, Australia2Cooperative Research Centre for Viticulture, Plant Research Centre, Hartley Grove, Urrbrae, SA 5064, Australia3Present address: Department of Horticulture, Viticulture and Oenology, The University of Adelaide, PMB1 GlenOsmond, SA 5064, Australia

Received 18 May 2000; revised 21 August 2000; accepted 5 September 2000

Key words: Agrobacterium tumefaciens, grapevine, somatic embryogenesis, transgene expression, transgenicplants,Vitis vinifera

Abstract

We have developed anAgrobacterium-mediated transformation system for a number of important grapevine cul-tivars used in wine production. Transgenic plants were obtained for the seven cultivars: Cabernet Sauvignon,Shiraz, Chardonnay, Riesling, Sauvignon Blanc, Chenin Blanc and Muscat Gordo Blanco. Embryogenic calluswas initiated from anther filaments and genotypic differences were observed for initiation and subsequent prolif-eration with Chardonnay responding most favourably to culture conditions. The transformation system allowedthe recovery of germinating transgenic embryos 10–12 weeks afterAgrobacteriuminoculation and plants within18 weeks. Examination of the expression patterns of the green fluorescent protein gene under the control of theCAMV35S promoter in leaf tissue of transgenic plants showed that for up to 35% of plants the pattern was notuniform. The successful transformation of a genetically diverse group of wine grape cultivars indicates that thetransformation system may have general application to an even wider range ofVitis viniferacultivars.

Abbreviations:2,4-D – 2,4-dichlorophenoxyacetic acid; BAP – 6-benzylaminopurine; TDZ –N-(1,2,3-thiadiazol-5-yl)-N ′-phenylurea (Thidiazuron); CPPU –N-(2-chloro-4-pyridyl)N ′-Phenylurea; NOA – ß-naphthoxyaceticacid; NAA – 1-naphthaleneacetic acid; MES – 4-morpholineethanesulphonic acid; X-gluc – 5-bromo-4-chloro-3-indoyl ß-D-glucuronide

Introduction

Grapevine cultivars which have a reputation forproducing premium quality wine include CabernetSauvignon, Shiraz (Syrah), Chardonnay, SauvignonBlanc and Riesling. These cultivars and others whichare used for quality wine production are establishedgenotypes ofVitis vinifera which have been vegetat-ively propagated for hundreds of years and in somecases for over a thousand years (Viala & Vermo-rel, 1901–1910; Galet, 1990). Clonal propagationhas preserved cultivar identity but has also restrictedthe improvement of these genotypes to strategies that

∗ Author for correspondence:E-mail: mark.thomas@pi.csiro.au

do not involve conventional breeding. The strategyof clonal selection has been used to obtain cloneswithin a cultivar which exhibit phenotypic or perform-ance differences. However, this is a slow process andclonal differences may arise not only by somatic muta-tions but may also be due to epigenetic changes anddisease (e.g. viral) load. A transgenic approach to ge-netic improvement of traditional wine grape cultivarswould allow modification of traits such as disease andpest resistance, product quality, and production ef-ficiency without altering the essential characteristicsof the cultivar. This is extremely important withinthe wine industry where the cultivar name is usedwidely for wine differentiation and product labelling.Fungal resistance to powdery and downy mildew are

106

important traits lacking in allV. vinifera cultivarsgrown for premium quality wines. Introduction ofresistance genes from other species by genetic trans-formation could reduce both management costs andthe requirement for chemical control measures.

Successful application of gene technology tograpevines requires an efficient transformation systemwhich can be applied to a wide range of cultivars.Regeneration of transgenicV. vinifera plants afterAgrobacteriuminoculation has only been previouslyreported for the table grape and multipurpose cultivarsKoshu Sanjaku, Superior Seedless and Sultana (Na-kano et al., 1994; Perl et al., 1996; Scorza et al., 1996;Franks et al., 1998) and the single wine grape cultivarChardonnay (Mauro et al., 1995).

Here we report the regeneration of transgenicplants for a number of grapevine cultivars of signific-ance in the world wine industry. The described trans-formation system appears to be suitable for applicationto otherV. viniferacultivars.

Materials and methods

Initiation and maintenance of somatic embryogeniccultures

Somatic embryogenic cultures were initiated forthe nine cultivars Chardonnay, Shiraz, CabernetSauvignon, Semillon, Sauvignon Blanc, MuscatGordo Blanco, Chenin Blanc, Riesling and Pinot Noirfrom anthers of immature field grown flowers as de-scribed by Franks et al. (1998). In the first year,anther cultures for Chardonnay, Cabernet Sauvignon,Chenin Blanc, Pinot Noir, Riesling, Sauvignon Blanc,Semillon and Shiraz were initiated on PIV (Frankset al., 1998), HT and JK initiation media. Anther cul-tures for Muscat Gordo Blanco were plated on PIV.In the second year, anther cultures for Chardonnay,Shiraz, Cabernet Sauvignon and Semillon were platedon PIV, HT and C1P (Torregrosa, 1998) initiationmedia. The phytohormones used were; PIV (4.5µM2,4-D, 8.9µM BAP), HT (10µM 2,4-D, 5µM TDZ),JK (2.5µM 2,4-D, 5µM CPPU, 2.5µM NOA), C1P

(5µM 2,4-D, 1µM BAP). The basal medium of PIV,HT, and JK was NN (Nitsch & Nitsch, 1969), 6%sucrose, pH 5.7, and 0.3% Phytagel (Sigma ChemicalCo. USA). The medium C1P was solidified with 0.5%Phytagel. Embryogenic callus was maintained on C1P

medium with a subculture every 4 weeks. Tissue wastransferred onto GS1CA medium (Franks et al., 1998)for 1 month prior to transformation.

Constructs andAgrobacterium tumefaciensstrains

The binary construct pNTG+ (Franks et al., 1998)containing the chimericgusA reporter gene and p35S-gus-INT (Vancanneyt et al., 1990) containing thegusAreporter gene interrupted by a plant intron sequencewere used for ß-glucuronidase (GUS) expression stu-dies. The binary plasmid pBINm-gfp5-ER (Haseloffet al., 1997) contains the green fluorescent protein(GFP) gene targeted to the endoplasmic reticulumand was used forgfp expression studies. The re-porter genes in pNTG+, p35S-gus-INT and pBINm-gfp5-ER were both controlled by the CAMV35Spromoter. Binary vectors contained thenptII cod-ing region, under the control of thenos promoter,conferring kanamycin resistance for selection of trans-formed plant cells.Agrobacterium tumefaciensstrainEHA101 (Hood et al., 1986) was transformed with thebinary plasmid pNTG+ and Agrobacterium tumefa-ciensstrain EHA105 was transformed with the bin-ary vectors pBINm-gfp5-ER and p35S-gus-INT byelectroporation (Mattanovich et al., 1989).

Transformations

Embryogenic callus of Cabernet Sauvignon, Semillon,Chenin Blanc, Pinot Noir, Chardonnay and SauvignonBlanc were co-cultivated with EHA105/pBINm-gfp5-ER. Muscat Gordo Blanco, Shiraz and Riesling wereco-cultivated with EHA101/pNTG+ and Chardon-nay with EHA105/p35S-gus-INT. A single colonyof Agrobacteriumwas incubated overnight at 28◦Cshaking in 50 ml of modified MG/L medium(Garfinkel & Nester, 1980) (5 g/l, mannitol, 1 g/lL-glutamate, 5 g/l tryptone, 2.5 g/l yeast extract, 2.5 ml/lFe-EDTA solution (7.44g/l Na2EDTA.2H2O and1.86 g/l FeSO4.7H2O), 5 g/l NaCl, 150 mg/l KH2PO4,100 mg/l MgSO4.7H2O) with biotin (20 ng/ml) and50µg/ml spectinomycin for pNTG+ or 100µg/mlkanamycin for pBINm-gfp5-ER and p35S-gus-INT.The culture was centrifuged at 2600× g for 5 min andthe pellet resuspended in induction medium (2 mMsodium phosphate, 40 mM MES pH 5.6, 0.5% gluc-ose, AB salts (Chilton et al., 1974) and 100µMacetosyringone) and incubated for a further 2 h. Theculture was centrifuged as above and the pellet re-suspended in liquid culture medium (GS1CA withouthormones) to an OD550 of 0.3. Embryogenic callus(1–2 g) collected from GS1CA medium was sub-merged in theAgrobacteriumsuspension (50 ml) for7 min with gentle shaking and then co-cultivated fortwo days in the dark at 22◦C on GS1CA medium.

107

The Agrobacteriumwas then removed by rinsing thecallus in 30 ml of liquid culture medium containing1 mg/ml timentin (SmithKline Beecham Pty Ltd). Thecallus was collected from the liquid culture mediumby filtering through a 100µ 3 M Nylon Net Filter(Millipore). The callus was then placed on GS1CAmedium plus 1 mg/ml timentin. Culture conditionswere at 28◦C in the dark. After 3 weeks from in-oculation the callus was subcultured onto the samemedium plus 100µg/ml kanamycin to allow for cal-lus proliferation under selection. After 5 weeks frominoculation the callus was placed on a hormone freeGS1CA medium plus kanamycin and timentin to se-lect for the germination of transformed embryos. Se-lection efficiency at embryo germination was determ-ined by GUS assays of root tips or by visualisation ofGFP. Germinated embryos were cut at the hypocotyland transferred to the light onto shooting medium(SM: major elements (160 mg/l, NH4NO3, 2325 mg/lKNO3, 340 mg/l KH2PO4, 440 mg/l CaCl2.2H2O and370 mg/l MgSO4.7H2O), minor elements (Murashige& Skoog, 1962), Fe-EDTA; (Dalton et al., 1983),B5 vitamins; (Gamborg et al., 1968), 15 g/l sucrose,1% agar, pH 5.7) containing a cytokinin (see Res-ults section) for shoot growth. Once shoots weredeveloped they were placed in MAGENTATM GA7-3Vessels (Life Technologies) containing rooting me-dium (SM with 0.5µM NAA). Rooted plantlets werepotted into soil and placed in a glasshouse after ac-climatisation. The glasshouse was set to a 30◦C, 16 hday, achieved by supplementary lighting in winter, andnight temperature of 25◦C.

Southern analysis, GUS and GFP visualisation

DNA extraction from plants and Southern analysiswas performed as previously described (Franks et al.,1998). Genomic DNA was digested withEcoRV re-striction enzyme and probed with thenptII codingregion. ForgusA expression analysis, root tips ofgerminated embryos were placed in X-gluc assay buf-fer (Franks et al., 1998) and left to stain overnightat 37◦C. Leaves fromin vitro grown plantlets werevacuum infiltrated with X-gluc assay buffer for 1 h be-fore a 37◦C overnight incubation. After staining, rootand leaf tissue were cleared in 70% ethanol. Trans-formed callus, embryos and leaves from transgenicplants were visualised for GFP fluorescence using aLeica MZ12 dissecting microscope fitted with a 50 Wmercury lamp and ‘GFP plus’ filter combination.

Results

Initiation and maintenance of embryogenic culture

Embryogenic callus was successfully initiated andmaintained for each of the cultivars. Embryogeniccallus initiation and growth from immature anther fil-aments was recorded for Chardonnay, Shiraz, Caber-net Sauvignon and Semillon on PIV, HT and C1initiation media (Table 1). For these genotypes HTmedium appeared superior for initiation. However,when the frequency of embryogenic callus initiationwas calculated after two subcultures on C1P mediumHT medium was best for the initiation of Chardon-nay embryogenic callus with PIV medium best forShiraz, Cabernet Sauvignon and Semillon (Table 1).Chardonnay showed the highest frequency of 9.38%with Shiraz, Cabernet Sauvignon and Semillon havingcomparatively low frequencies below 0.6% (Table 1).Not all initial embryogenic callus continued growthafter subculture. C1P was chosen as the proliferationmedium based on the observation of its superiorityover the other initiation media and GS1CA mediumfor maintaining embryogenic callus over 2 years.Difficulties experienced with the other media included

Table 1. Embryogenic callus growth from immature antherfilaments of various wine grape cultivars

Cultivar Initiation Callus Proliferating

medium initiation on callus after

initiation subculture (%)∗∗medium (%)∗

Chardonnay HT 17.55 9.38

PIV 12.66 3.49

C1P 4.80 4.80

Shiraz HT 36.94 0.00

PIV 21.81 0.59

C1P 1.49 0.00

Cabernet HT 25.73 0.05

Sauvignon PIV 22.49 0.14

C1P 6.48 0.00

Semillon HT 44.91 0.00

PIV 20.62 0.11

C1P 5.04 0.00

Minimum number of anthers per medium was 360.∗Percent of anthers forming initial embryogenic callus on ini-tiation medium.∗∗Percent of embryogenic calli after two transfers onto C1P

medium per total number of initial anthers.

108

slow growth and gradual depletion of proembryostructures as a result of embryo differentiation. Em-bryogenic callus from all cultivars proliferated on C1P

medium where they remained at an early stage ofproembryo development described as Type I callus(Franks et al., 1998). Proliferation appeared to oc-cur by secondary embryogenesis of proembryo-likestructures.

Transformation and regeneration efficiency oftransgenic plants

For transformation studies embryogenic callus wastransferred from C1P medium to GS1CA medium forone month prior toAgrobacteriuminoculation. Wash-ing the overnight culture ofAgrobacteriumin liquidculture medium and adjusting the OD550 to 0.3 greatlyreduced the necrotic effects ofAgrobacteriumco-cultivation previously observed forV. vinifera (Perlet al., 1996) thereby avoiding the use of anti-oxidants.Efficiency of selection was determined at the stageof embryo germination when roots appeared eitherby gfp expression or GUS assays of root tips. Trans-genic embryos began germinating as early as 10 weeksafter inoculation. A typical transformation experi-ment for Shiraz with kanamycin selection producedan overall selection efficiency of 62% with a high of94% over a 14 week period (Figure 1). In all trans-formation experiments selection efficiency was low inembryos that germinated before 12 weeks from in-oculation (Figure 1). At around 12 weeks onwardsselection efficiency increased and remained high un-til the experiment was terminated at the end of 32

Figure 1. Efficiency of selection of transgenic Shiraz embryos atvarious weeks afterAgrobacteriumco-cultivation. Root tips of em-bryos which germinated on selective media were assayed for GUSto determine percentage of transformants.

weeks. The first shoots emerged from germinatingembryos 2 weeks after transfer to cytokinin medium.BAP at a concentration of 10µM promoted shootingin Shiraz, Chardonnay, Chenin Blanc, and Rieslingas was previously found for Sultana (Franks et al.,1998). Cabernet Sauvignon and Sauvignon Blanc,however, responded better to a lower BAP concentra-tion of 2µM and Muscat Gordo Blanco required 5µMZeatin Riboside for shoot development. The earliesttransformed plantlets were potted into soil 18 weeksafter inoculation. From 1.3 g of Shiraz embryogeniccallus inoculated withAgrobacterium, 161 transgenicembryos were identified and 45 transgenic plantletsobtained. The frequency of transgenic plantlet regen-eration from these Shiraz GUS positive embryos was28%. Transgenic plantlet regeneration from embryosof other cultivars ranged from 10–33% (Table 2). Oneof the main limitations for obtaining transgenic plantswas inefficient shoot formation.

The standard transformation and selection protocoldid not include kanamycin when embryos were trans-ferred to shooting medium (see Materials and meth-ods). However, kanamycin (100µg/ml) was includedat the shooting and rooting stage in some experimentsto determine its usefulness. There was no differencein greening and shoot formation between transgenicand non-transgenic embryos when they were placedon shooting medium with kanamycin. In contrast,when kanamycin was included in rooting mediumonly transgenic shoots were capable of forming roots.This represents an effective screen for identifyingtransgenic plantlets prior to potting out.

Transgenic plants were obtained for Chardonnay,Shiraz, Riesling, Cabernet Sauvignon, Muscat GordoBlanco, Chenin Blanc and Sauvignon Blanc (Table 2).GFP fluorescence was present in Semillon and PinotNoir proliferating callus indicating that stable trans-formation had occurred but no embryos developed tothe germination stage.

Transgenic plants

The transgenic status of plants was confirmed bySouthern analysis (Figure 2) and by visualisation ofreporter gene expression in leaves (Figure 3). Thenumber of T-DNA copies inserted into the genomewas determined by digesting genomic DNA with theEcoRV restriction enzyme and probing with thenptIIcoding region. For all constructs, oneEcoRV restric-tion site occurs within the T-DNA and the other siteoccurs elsewhere in the plant genome. Transgene copy

109

Table 2. Transformed red and white wine cultivars ofVitis vinifera

Cultivar Construct Transgenic Transgenic plant Independent

embryogenic callus regeneration∗ transformation events∗∗

Chardonnay pBINm-gfp5-ER Yes 2/2 2/2

Chardonnay p35S-gus-INT Yes 5/50 4/4

Cabernet Sauvignon pBINm-gfp5-ER Yes 31/136 4/5

Sauvignon Blanc pBINm-gfp5-ER Yes 5/23

Chenin Blanc pBINm-gfp5-ER Yes 19/57

Muscat Gordo Blanco pNTG+ Yes 1/9

Shiraz pNTG+ Yes 45/161 10/13

Riesling pNTG+ Yes 2/19 2/2

Semillon pBINm-gfp5-ER Yes 0

Pinot Noir pBINm-gfp5-ER Yes 0

∗Number of plants regenerated over the number of GFP and GUS expressing embryos placed onto shooting medium.For all cultivars except Muscat Gordo Blanco and Riesling the numbers of embryos plated for plant regenerated wasrestricted. Thus the potential number of transgenic plants which could have been recovered in each experiment forthese cultivars is much higher than indicated.∗∗Number of independent transformation events over the number of plantlets as determined by Southern analysis.

Figure 2. Southern blots of plant genomic DNA digested withEcoRV and probed with a fragment of thenptII gene to reveal T-DNA copynumber. Shiraz transformed with pNTG+ (lanes 2, 3), Riesling transformed with pNTG+ (lanes 5, 6), Muscat Gordo Blanco transformed withpNTG (lane 8), Cabernet Sauvignon transformed with pBINm-gfp5-ER (lanes 10, 11), Chardonnay transformed with pBINm-gfp5-ER (lanes13, 14). Lanes 1, 4, 7, 9 and 12 represent non-transformed plants of the respective cultivars.

number as estimated by Southern hybridisation rangedfrom 1 to 6 copies (Figure 2). Plantlets regeneratedfrom independent transformation events were distin-guishable by their differing banding patterns on South-ern blots and depending on the cultivar this rangedfrom 77–100% (Table 2). Fifty seven transgenic plantswere assessed for the pattern ofgfp expression inleaves. In some of the transgenic plantsgfp expres-sion was not uniform (Figure 3). While the majorityof plants had a uniform expression pattern the per-centage of plants exhibiting variegatedgfp expressionwas 35% with 9% lackinggfp expression in leaf vas-cular tissue and 26% also showing patchy expressionin the leaf lamina. Thegfp gene was driven by theCAMV35S promoter.

All cultivars regenerated after transformation hada lobed leaf phenotype previously described for trans-genic Sultana (Franks et al., 1998) confirming thatregeneration by somatic embryogenesis induces a ju-venile phenotype similar to that seen for seedlingsgerminated from seed.

Discussion

All nine V. vinifera cultivars tested were capable offorming embryogenic callus from anther filaments andthis embryogenic callus could be maintained by sub-culture. However, differences were observed betweencultivars for the efficiency of embryogenic callus ini-

110

Figure 3. Expression ofgfp from the CAMV35S promoter in transgenic plants. Uniform expression in leaf tissue of Cabernet Sauvignon (a)and Chardonnay (c) plants. Variegated expression patterns in leaf tissue of Cabernet Sauvignon (b) and Chardonnay (d) plants. The outlines ofindividual cells (arrow) are clearly seen with GFP fluorescence restricted to the thin layer of cytoplasm between the cell wall and vacuole. Barrepresents 5µm.

tiation and subsequent proliferation indicating geno-typic differences. Table 1 shows that the percentageof embryogenic callus capable of growth after sub-culture is significantly lower (54–0.5%) than initialembryogenic callus formed on anther filaments. Thissuggests that an accurate frequency value for suc-cessful grapevine embryogenic initiation should bedetermined after 1–2 subcultures instead of at initi-ation. Genotypic differences were also observed atthe embryo shooting stage with Muscat Gordo Blancopreferring the cytokinin Zeatin Riboside to BAP andCabernet Sauvignon and Sauvignon Blanc requiringa lower concentration of BAP. Differences in regen-eration response of different grapevine species havealso been observed when cultured on solid or liquidmedium (Mozsar & Viczian, 1996). It is likely thatthese differences are due to the highly heterozygousgenotypes of cultivars ofV. vinifera(Thomas & Scott,1993; Thomas et al., 1993). While it is apparent thatcultivar genotypic differences should be taken intoaccount at all culture steps from initiation to regener-ation of transgenic grapevine plants the tissue cultureprotocol described in this paper appears to have ap-

plication to a wide variety ofV. viniferacultivars. Thetransformation procedure used was based on our pre-vious work with Sultana (Franks et al., 1998). Theinclusion of an initialAgrobacteriumpre-treatmentstep to prevent browning and tissue necrosis afterAgrobacteriumco-cultivation was successful in redu-cing the frequent and significant tissue loss previouslyexperienced.

Results from different cultivars (see Figure 1 forShiraz) suggests that it takes between 10–12 weeksfrom Agrobacteriuminoculation for the first trans-genic somatic embryos to develop and mature to agermination stage. After this time transgenic embryosrepresented a consistently high percentage (61–100%)of germinating embryos. Non transgenic embryos de-veloping before 10–12 weeks could be attributed to thepresence of some advanced multicellular embryos atthe time ofAgrobacteriuminoculation. These untrans-formed or partly transformed multicellular embryosmay be resilient to the effects of kanamycin. Embryosthat germinate after 10–12 weeks are more likely to beof single cell origin at the time of inoculation and giverise to fully transgenic germinated embryos. It appears

111

that kanamycin selection is more effective when ap-plied to grapevine somatic embryos at an early stageof development but even then it is not completelyeffective in preventing nontransformed embryos fromgerminating. In our experience the most efficient se-lection of transgenic embryos has occurred when closeattention has been paid to the development stage of theembryogenic callus used forAgrobacteriuminocula-tion by choosing callus lacking later stage embryos.Kanamycin selection applied at the culture step of rootinduction on shoots was found to be very effectivewith only transgenic shoots being capable of form-ing roots. This additional kanamycin selection stepis an improvement on that originally used for Sul-tana (Franks et al., 1998) and removes the need toPCR test plantlets for a transgene prior to transfer tosoil.

Stable transformation and transgenic plant regener-ation ofV. viniferacultivars Shiraz, Riesling, CabernetSauvignon, Sauvignon Blanc, Chenin Blanc and Mus-cat Gordo Blanco have been demonstrated for the firsttime. Cultivars ofV. viniferaare vegetatively propag-ated and the mode of inheritance (and hence the num-ber of insertion sites) of a transgene is inconsequentialas long as the desired phenotype is achieved and main-tained. Transformation of Pinot Noir and Semilloncallus was successful but no plants were regenerated.Embryogenic callus quality at the time of transforma-tion and the development of mature somatic embryosafter transformation appeared to be the primary causesof successful or unsuccessful transgenic plant regen-eration between the different cultivars. Relatively highnumbers of transgenic embryos could be obtainedfrom a small amount of embryogenic callus over a32 week period (see Results). The high frequency ofplantlets regenerated from independent transformationevents (Table 2) indicates efficient transformation attheAgrobacteriuminoculation stage. The limiting stepin obtaining transgenic plants was the developmentof shoots from embryos. Further research is requiredat this step in the procedure to improve the overallefficiency of the system.

The gfp reporter gene, with its documented ad-vantages overgusA (Haseloff et al., 1997), was usedto examine gene expression in live tissue of trans-genic plants. The nonuniform transgene expressionpatterns observed were in some cases similar to pat-terns found previously with transgenic Sultana GUSpositive plants (Franks et al., 1998) withgfp ex-pression absent in leaf vascular tissue and in othercases quite different (Figure 3). BothgusA and gfp

reporter genes were driven by the CAMV35S pro-moter. In some cases of variegatedgfp expression inthe leaf lamina the expression appeared to be asso-ciated with groups of cells (see Figure 3b) while inother instances this was not apparent (see Figure 3d).Since the plants were regenerated from embryos withuniform gfp expression patterns it is unlikely thatthe variegated patterns are due to plant regenerationfrom partially transformed chimeric embryos. In ad-dition the patterns are not typical for a chimera. Thecause of these abnormalgfp expression patterns havenot been investigated although cytosine and adeninemethylation of transgenes introduced into grapevinehas previously been shown (Franks et al., 1998) andmay be associated with silencing of gene expres-sion. Alternatively the phenotype may be indicative ofan absence of gene expression in some cells due toposition-effect variegation (Martin & Whitelaw, 1996)or other gene silencing mechanisms including post-transcriptional mediated gene silencing (Stam et al.,1997). The observation that 35% ofgfp expressinggrapevine plants had an aberrant expression patternshould be taken into account for studies seeking toproduce an improved transgenic plant for commercialproduction.

This study has described the development of atransformation system which can be applied to grapev-ine cultivars of importance to the wine industry andrepresents a significant step towards the successful ap-plication of gene technology to the improvement ofestablished premium wine cultivars.

Acknowledgements

We thank Don Mackenzie for technical assistance.This work was supported in part by the Grape andWine Research and Development Corporation and theCooperative Research Centre for Viticulture.

References

Chilton MD, Currier TC, Farrand SK, Bendich AJ, Gordon MP andNester EW (1974)Agrobacterium tumefaciensDNA and PS8bacteriophage DNA not detected in crown gall tumors.Proc NatlAcad Sci USA71: 3672–3676.

Dalton CC, Iqbal K and Turner DA (1983) Iron phosphate precipita-tion in Murashige and Skoog media.Physiologia Plantarum57:472–476.

Franks T, He DG and Thomas MR (1998) Regeneration of trans-genicVitis vinifera L. Sultana plants: genotypic and phenotypicanalysis.Mol Breeding4: 321–333.

112

Galet P (1990) Cépages et vignobles de France. Tome II.L’ampélographie francaise. Vol. Charles Dehan, Montpellier.

Gamborg OL, Miller RA and Ojima K (1968) Nutrient requirementsof suspension cultures of soybean root cells.Experimental CellRes50: 1515–1518.

Garfinkel DJ and Nester EW (1980)Agrobacterium tumefaciensmutants affected in crown gall tumorigenesis and octopine cata-bolism.J Bacteriol144: 732–743.

Haseloff J, Siemering KR, Prasher DC and Hodge S (1997) Removalof a cryptic intron and subcellular localization of green fluores-cent protein are required to mark transgenicArabidopsisplantsbrightly. Proc Natl Acad Sci USA94: 2122–2127.

Hood EE, Helmer GL, Fraley RT and Chilton MD (1986) Thehypervirulence ofAgrobacterium tumefaciensA281 is encodedin a region of pTiBo542 outside of T-DNA.J Bacteriol 168:1291–1301.

Martin DI and Whitelaw E (1996) The vagaries of variegatingtransgenes.BioEssays18: 919–923.

Mattanovich D, Rüker F, da Câmara Machado A, Laimer M, RegnerF, Steinkellner H, Himmler G and Katinger H (1989) Efficienttransformation ofAgrobacteriumspp. by electroporation.NuclAcids Res17: 6747.

Mauro MC, Toutain S, Walter B, Pinck L, Otten L, CoutosthevenotP, Deloire A and Barbier P (1995) High efficiency regeneration ofgrapevine plants transformed with the GFLV coat protein gene.Plant Sci112: 97–106.

Mozsar J and Viczian O (1996) Genotype effect on somatic em-bryogenesis and plant regeneration ofVitis spp.Vitis 35: 155–157.

Murashige T and Skoog F (1962) A revised medium for rapidgrowth and bioassays with tobacco tissue cultures.PhysiologiaPlantarum15: 473–497.

Nakano M, Hoshino Y and Mii M (1994) Regeneration of trans-genic plants of grapevine (Vitis vinifera L.) via Agrobacteriumrhizogenes-mediated transformation of embryogenic calli.J Ex-perimental Bot45: 649–656.

Nitsch JP and Nitsch C (1969) Haploid plants from pollen grains.Science163: 85–87.

Perl A, Lotan O, Abu-Abied M and Holland D (1996) Estab-lishment of anAgrobacterium-mediated transformation systemfor (Vitis vinifera L.) – the role of antioxidants during grape-Agrobacteriuminteractions.Nat Biotechnol14: 624–628.

Scorza R, Cordts JM, Gray DJ, Gonsalves D, Emershad RL andRamming DW (1996) Producing transgenic ‘Thompson Seed-less’ grape (Vitis viniferaL.) plants.J Am Soc Horticult Sci121:616–619.

Stam M, Mol JNM and Kooter JM (1997) The silence of genes intransgenic plants.Ann Bot79: 3–12.

Thomas M and Scott NN (1993) Microsatellite repeats in grapevinereveal DNA polymorphisms when analysed as sequence-taggedsites (STSs).Theor Appl Genet86: 985–990.

Thomas MR, Matsumoto S, Cain P and Scott NS (1993) RepetitiveDNA of grapevine: classes present and sequences suitable forcultivar identification.Theor Appl Genet86: 173–180.

Torregrosa L (1998) A simple and efficient method to obtain stableembryogenic cultures from anthers ofVitis vinifera L. Vitis 37:91–92.

Vancanneyt G, Schmidt R, O’Connor-Sanchez A, Willmitzer Land Rocha-Sosa M (1990) Construction of an intron-containingmarker gene: Splicing of the intron in transgenic plants and itsuse in monitoring early events inAgrobacterium-mediated planttransformation.Mol Gen Genet220: 245–250.

Viala P and Vermorel V (1901–1910) Ampélographie. Vol. 1–7Masson, Paris.