Plant Cell Reports

Plant Cell Reports (1996) 15:549-554

9

Verlag 1996

Agrobacterium-mediated transformation of apple (Malus x domestica Borkh.): an assessment of factors affecting regeneration of transgenic plants An De Bondt t' 2, Kristel Eggermont ~, Iris Penninckx t, Inge Goderis ~, and Willem E Broekaert F.A. Janssens Laboratorium voor Genetica, Katholieke Universiteit Leuven, Willem de Croylaan 42, B-3001 Heverlee, Belgium

2 N.V. Jo Nicolai & Co, Gorsemdorp 51, B-3803 Saint-Truiden, Belgium Received 15 March 1995/Revised version received 22 June 1995 - Communicated by A. M. Boudet

ABSTRACT

INTRODUCTION

We have previously developed a protocol for efficient gene transfer and regeneration of transgenic calli following cocultivation of apple (cv. Jonagold) explmlts with Agrobaeterium tumefaeiens (De Bondt et al. 1994, Plant Cell Reports 13: 587-593). Now we report on the optimization of postcultivation conditions for efficient and reproducible regeneration of transgenic shoots from the apple cultivar Jonagold. Factors which were found to be essential for efficient shoot regeneration.were the use of gelrite as a gelling agent and the use of the cytokinin-mianicing thidiazuron in the selective postcultivation medium, hnproved transformation efficiencies were obtained by combining the hormones thidiazuron aud zeatin and by using leaf explants from in vitro grown shoots not older than 4 weeks after multiplication. Attempts to use phosphinofltrich~ acetyl transferase as a selectable marker were not successfill. Using selection on kanamycin under optimal postcultivation conditions, about 2% of file leaf explants developed transgenic shoots or shoot clusters. The presence and expression of the transferred genes was verified by [3glucuronidase assays and Southern analysis. The transfonnation procedure has also been succesfully applied to several other apple cultivars.

Conventional breeding of woody species is a low efficiency process becanse of their long life cycle and the high heterozygcsity of the parental lines. Transforuaation tectmiques are aimed at the introduction of additional genes corresponding to beneficial properties into the otherwise mlaltered genome of a cultivar, and may therefore speed up the breeding process. Many transformation procedures of herbaceous plants are based on explants derived from seed or seedling orgmls such as wltole embryos, hypocotyls or cotyledons. However, as passage through a sexual stage results in a drastic reshuffling of the genome and hence in an alteration of cultivar properties, it is essential that transformation procedures of woody species are based on organ explants such as leaves or stems which are derived from vegetatively propagated plants. Apple leaf explants have previously been shoml to regenerate shoots at efficiencies close to 100% (James et al. 1988; Fasolo et al. 1989; Sriskandarajah et al. 1990; Welander mid Maheswaran 1992). Based on the high regenerative capacity of leaf explmlts, Agrobacterium tumefaciens mediated transformation protocols have been worked out for the apple cultivars Greensleeves (James et al. 1989) and Golden Delicious (Sriskandarajah et al. 1994) mid for the apple rootstock M26 (Norelli et al. 1994; Maheswaran et al. 1992). Relatively few studies have addressed in depth the question which factors affect transformation efficiencies. James et al. (1993) have studied the influence of acetosyringone, an inducer of virulence gene expression in A. tumefaeiens, and of the osmoprotectant betaine phosphate on the efficiency of gene transfer to apple (cv. Greensleeves) explants. They found that addition of these compounds to the bacterial suspension 6 hours prior to inoculation significantly e~fllanced gene transfer efficiency. We have previously shown that transformation of apple (cv. Jonagold) leaves with A. mmefaeiens strain EHA101 (pEHA101) yielded five times more transgenic calli relative to two other disarmed strains tested (De Bondt et al. 1994). Moreover, it was observed that the hormonal composition and the nature of the carbon soarce of the cocultivation medium drastically affected gene transfer efficiency (De Bondt et al. 1994). Maheswaran et al. (1992) have reported that no

Abbreviations: BAP, benzylaminopurine; CTAB, hexadecyltrimethylan~noniumbromide; Na2EDTA, ethylenediamine-tetra-acetate di-sodium salt; FeNaEDTA, ethylenediamfile-tetra-acetate ferric-sodium salt; GA3, gibberellic acid 3; GusA, [3-glucuronidase; gusA, ~glucuronidase gene of Escheriehia toll; IAA, indole acetic acid; IBA, indole butyric acid; 2iP, N6-2-isopentenyl adenine; NAA, naphthalene acetic acid; nptlI, neolnycinphosphotransferase 1~ gene; bar, phosphinothricin acetyl transferase gene; PCR, pobanerase chain reaction; PPT, phosphinothricin; STS, silver thiosulphate; T-r)NA, transferred DNA, TDZ, thidiazttron; XGluc, 5-bromo-4-chloro-3-indolyl ~-D-glucuronide; Zea, transZeatin Correspondence to: W. E Broekaert

550 regeneration o f transgenic shoots o f rootstock M26 occurred on selective m e d i u m (kanmnycin 100 mg/1) gellified with agar, w h e r e a s trmlsfonnation efficiency on gelrite-containing m e d i a r e a c h e d 2-4%. hi the p r e s e n t study, w e report on the succesful regeneration o f transgenic shoots from l e a f explants o f several COlnlnercial apple cultivars. The influence o f several components in the

on sterile filter paper and transferred to a postcultivation medium, AS 188 (unless stated otherwise), supplemented with 100 mg/l k,'mamycin for EHA101 (pFAJ3003) or 10 mg/l phosphinothricin (PPT) for EHA105 (pFAJ3027), 200 rag/1 triacillin and 200 mg/l cefotaxime ,and solidified with 2.5 g/1 gelrite. Postcultivation of tbe leaves was done in the dark at 23~ After 4 weeks, the leaves were transferred to the same postcultivation medium but with reduced concentrations of Agrobacterium killing antibiotics (100 mg/l triacillin, 100 mg/1 r This transfer was repeated every 4 weeks until appearance of buds.

postcultivation m e d i u m on the regeneration efficiency o f transgenic shoots from the cultivar Jonagold was assessed.

=

A

..... I..,o.,

MATERIALS AND METHODS

Bacterial strains and vectors. The supervirulent succinamopine Agrobacter/um tumefaciens strain EHA101 (pEHA101) (Hood et al. 1986) provided with the binmTr expression vectors pFAJ3003 and pFAJ3027 (Fig. 1) were used. pFAJ3003 was constructed as follows. In a first step a 2.4 kb AT~aI fi'agment from pCGN1547 (McBride and Summerfelt 1990) containing the nptlI coding region flanked by the promoter and temfinator elements of the lnammpine syutbase Onas) gene was cloned into the SalI site of vector pDE3. Vector pDE3 is a pUC19 derivative containing in between the ,gamHI and XbaI sites of the polycloning site tbe nopaline synthase trios) promoter, the gusA coding region and the nos terminator. The gusA and nptll cassettes of the resulting vector were cut out as a 5 kb HindlII-BamHI fragment and cloned into the HindlII-Bglll sites of the binary vector pGSC1701A2. pGSC 1701 A2 is a derivative of pGSC 1700 (Co,'nelissen and Vandewiele 1989) from which the pBR322 derived ampicilline resistance marker has been deleted. The resulting vector is called pFAJ3003 and is schematically ,'epresented in Fig. 1A. Vectors pDE3 and pGSC1701A2 were kindly provided by Dr. J. Bottennan of Plant Genetic Systems N.V. (Ghent, Belgium). pFAJ3027 (Fig. 1B) was constructed by cloning a 2.5 kb HindIII-SaoI liagment tiom p35S-GUSint (Vancmmeyt et aI. 1990) into the HindIII-SacI sites of the binary vector pGPTV-BAR (Becker et al. 1992). pFAJ3003 was introduced into A. tum@eiens strain EHA101 (pEHA101) and vector pFAJ3027 into A.. tumefaeiens strain EHA105 (pEHA105) (Hood et al. 1993) by electroporation. A. mm@eiens strain EHA105 (pEHA105) is a kanamycin sensitive derivative of EHA101 (pEHA101). Plant tissue culture and translbnnation. Apple shoots were established by meristem-tip culture and subsequent micropropagation as previously described by Druart (1988). Explants used for transformation were the youngest four fully expanded leaves of 5-week-old (unless stated otherwise) micropropagated shoots. The leaves were wounded by making three incisions perpendicular to the midrib taking care not to cut through the leaf edges (Druart.1990). To infect the leaves with ,4. tum@eiens, a bacterial suspension was made up as follows. A. tumefiwiens was grown in YEB medium (10 g/l bacto peptone, 5 g/l NaCI and 10 g/1 yeast ex-h'act) supplemented witb the appropriate antibiotics (300 mg/l streptomycin and 100 rag/1 spectinomycin for strain EHAI01 and 50 rag/1 kanamycin tb," strain EHA105) until an ()D600 of 0.5 to 0.7 (midlog phase) was reached. The bacteria were spun down by centrifugation (4000xg, 10 minutes) and resuspended in an equal volmne of liquid AS123 medium supplemented with 10 mM MgSO 4. Medinm AS123 contained the macro elements of Murashige and Skoog (1962), micro elements as reported by Druart (1980), 40 mg/l FeNaEDTA, 100 rag/1 myo-inositol, 2 mg/l thiamine-HCl, I mg/I pyridoxine-HCl, I rag/1 nicotinic acid, 2 rag/1 glycine, 250 mg/l NZamine (Sigma) and 2% sucrose. The leaves were shaken gently in this suspension for about 1 minute and then blotted dry on sterile filter paper. They were transferred to AS188 medium solidified with 2.5 g/1 gehite, with tbeir adaxial side in contact with the culture nledium and eocultivated during 4 days at 25~ in the dark. Medium AS 188 is AS 123 supplemented with 4 plant growth reguhttors: 0.25 mg/l indole-butyrie acid (IBA), 5 mg/l thidiazuron (TDZ), 2 m~t N6-[2-isopentenyl]adenine (2iP) and 0.1 rag/1 gibberellic acid 3 (GA,3). After cocultivation, the leaves were washed tbr t minute in liquid AS123 medium containing 400 mg/l triacillin, blotted dry

k .... I

LB

lOOO bp

RB

LB 1000 bp

Figure 1: Schematic representation of the T-DNA of the binary vectors pFAJ3003 (A) and pFAJ3027 (B). The region ex-temal to the T-I-)NA border sequence is identical to that of pGSC1701A2. Restriction sites marked with an asterisk are unique. Abbreviations: gusA~ coding region of the [3-glucuronidase gene; bar, coding region of the phospbinotlniein acetyl transferase gene; gusA-intron, intron containing coding region of the [~glucuronidase gene; LB, left border of T-DNA; nptII, coding region of the neomycin phosphotransferase gene; Penh35S, promoter of 35S RNA of cauliflower mosaic virus with duplicated enhancer region; Pmas, promoter of the mammpine synthase gene; Pnos, promote," of the nopaline synthase gene; RB, right border ofT-DNA; Tg7, termination sigual of T-DNA gene 7; Tmas, termination sigual of the manopine synthase gene; Thus, termination signal of the nopaline synthase gene. The appe,'uing buds and the callus, from which they originated, were excised from the explant and transferred to AS248, a shoot elongalion medium, solidified with 5 g/l agar and supplemented with 50 mg/I kanamycin fbr EHA101 or 5 mg/l PPT tbr EHA105o 100 mg/I triacillin mad 100 mg/l cefotaxime and incubated under the light from fluorescent tubes at 45 laE/m2s with a 16 hours plmtope,iod. Medium AS248 contained the mineral nutrients ,as described by Druart (1988), 40 mg/I FeNaEDTA, 100 mg/l myo-inositol, 1 rag/1 thiamine-HCl, 0.5 mg/I pyridoxine-HC1, 0.5 rag/1 nicotinic acid, 2 rag/1 glycine, 1 mg/l BAP, 0.1 mg/I IBA, 1% galactose and 2% sucrose. TransJ:bnnation efficiency is expressed as the percentage of explants regenerating kanamyein resistant shoots or clusters ofkanamyein resistant shoots. Usually, only one shoot or one shoot cluster regenerated per leaf explant. Statistical analysis was performed following Fisher's Exact Test. Siguifieant differences between treatments were stated with 95% co~ffidence and are marked in the figures with different characters. Two to tlnee months after appearance of the shoots a leaf was excised and tested for GusA activity following a histochemical staining procedure previously described (De Bondt et al. 1994). Between 70 and 100% of the surviving shoots showed detectable GnsA activity. Transgenic shoots were rooted as follows. The stem bases of excised shoots were put in AR18, a medium consisting of 0.9 g/l KNO> 0.6 g/l Ca(NOs)2.4H20, 0.18 g/1 MgSO4.7H20, 0.135 g/l KH2PO4 as macro elements, half strength micro elements as reported by Druart (1980), 40 rag/1 FeNaEDTA, 100 mg/l myo-inositol, 1 mg/l thiamine-HCl, 2 mg/1, IBA 2% sucrose and 5 g/1 agar. After incubation during 4 days at 23~ in the dark, the shoots were transferred to AR19 under white light from fluorescent tubes at 45 ~.tE/m2swith a 16 hours photoperiod. AR19 is AR18 without IBA, with sucrose at 3% and with the macro elements as described by Druart (1980). After tln'ee weeks, rooted shoots were removed fi'om the culture medium and chilled in a humid sterile culture vessel during 7 days at 4~C. Atter this cold treatment, the shoots were transferred to a 1:3 mixture of vemficulite ,and soil in a sealed

551 micropropagator at 100% relative humidity and placed in a growth chamber (20~ daytime temperature, 15~ nighttime temperature) with light at 100 ~E/m~s during a 16 hours photoperiod. After four weeks, the relative humidity was graduaUy reduced to 70% over a 5 days period, whereafter the lid of the micropropagator was taken off DNA preparation and Southern analysis. Genomic DNA was isolated from kanamycin resistant shoots, 5 weeks after multiplication, according to the method of Aldrich and Cullis (1993) with two additional steps. After CTAB extraction, proteinase K was added to the extract to a final concentration of 100 gg/ml and the extract was incubated tbr 30 minutes at 56~ After subsequent extractions with chlorofonn:octanol (24:1), ethanol was added to the aqueous phase to a final concentration of 10% (v/v). After a 30 mintltes incubation on ice, the precipitate was removed by centrifugation (10 minutes at 2000xg) and the supematant fimher treated as described by Aldrich and Cullis (1993). Aliquots of 10 /.$g DNA in a total volume of 500 Ixl were digested either with EcoRI or with HindlII, fi'actionated on a 0.8% agarose gel and transfelTed to a positively charged nylon membrane (Boeltringer). The membrane was hybridised with a digoxigenin labeled nptlI probe (0.7 kb) made by PCR using pFAJ3003 as template DNA and the primers described in HamiU et al. (1991). After washing the membrane, bound probe was visualized following the immunochemihuniniscent detection protocol of Boelu'inger.X-ray films were exposed for 30 minutes.

RESULTS AND DISCUSSION Vector construction. The T-DNA structure of pFAJ3003, a multipurpose vector developed for trmlsfonnation of heterozygous woody crops, is shown in Fig. 1A. Between the TDNA borders, this vector contains the nptII coding region driven by a mammpine synthase promoter as a selectable marker and the gusA coding region fnsed to a nopaline synthase promoter as a histochemical marker. The presence of a histochemical marker gene in transtbnnation vectors tbr heterozygous woody plant species is convenient since chimaerism can not be circumvented through a sexual reproduction step (as this would lead to alteration of cultivar properties) and hence must be checked tbr by histochemical staining of different parts of the trm~sgenic plant. The nopaline synthase promoter, driving the gusA histochemical marker gene, does not yield detectable gusA expression in A. tumefaciens (unpublished results). This eliminates the problem of false positive marker gene detection in transgenic plant tissue due to the presence of A. tumeJ'aciens. The connnonly used strong consfitntive cauliflower mosaic virus 35S RNA (CaMV35S) promoter is purposedly absent I?om both marker genes. This promoter can be used to drive an additional gene of interest which can be inserted into the uniqne HindlII site between the right border and the marker genes. Since this promoter is absent from the marker genes, introduction of a CaMV35S driven gene of interest can not cause recombination events between repeated DNA segments. In addition, pFAJ3003 does not contain a bacterial ampicillin resistmlce marker gene, which allows penicillin derivatives (e.g. carbenicillin, triacillin) to be used during postcultivation for killing A. utmefaeiens. Figure 1B shows the T-DNA structure of pFAJ3027. This vector was constructed to test the phosphinothricin acetyl transt~rase gene (bar) as an alternative for nptlI as selectable

marker. Besides a nopaline synthase promoter driven bat" gene, the vector contains a CaMV35S driven, intron containing gusA gene (gusA-intron) as a histochemical marker, gusA-intron was developed by Vancamleyt et al. (1990) as a histochemical marker which can not be expressed in A. tumefaciens. Unlike pFAJ3003, which has a streptomycin/ spectinomycin adenyl transferase gene as a bacterial selection marker, pFAJ3027 has a bacterial nptg selection marker which makes this vector inappropriate for introduction into the kanamycin resistmlt A. tumefaciens strain EHAI01 (pEHAI01). For this reason, pFAJ3027 was introduced in the kanamycin sensitive derivative ofEHA101 (pEHA101), EHA105 (pEHA105). blfluence of the ~elling agent and the hormonal composition of the postcultivation medimn. In a first experiment, we have compared the capacity of Agro-infected apple cv. Jonagold leaves to regenerate transgenic shoots on 5 different postcultivation media using 400 leaf explants per medium. All media contained 0.25 rag/1 IBA as the auxin and 5 my/1 of either Zeatm0 BAP, 2iP or TDZ as the cytokinin and were all solidified with gelrite (2.5 g/l). A fifth medium contained IBA (0.25 rag/l) and TDZ (5 rag/l) as in one of the former combinations, but was solidified with agar (5 g/l). As presented in table 1, the postcultivation medium containing TDZ and geh'ite resulted in the highest transformation efficiency (1.2%), whereas only either a zero or a very low transformation efficiency was reached on the other media. The results of this orientating experiment are consistent with those of Maheswarm~ et al. (1992) who were able to regenerate transgenic shoots fi'om apple rootstock M26 leaf explants on media solidified with gelrite (1-4% efficiency) but not on agar containing media. In contrast, Sriskandarajah et al. (1994) succeeded in the recovery of transgenic Golden Delicious shoots with an efficiency of 1-2 % on a medium solidified with agar. Our finding that TDZ is required tbr et'ficient regeneration of transgenic shoots is consistent with that of Sriskandarajah et al. (1994). Many authors have shown that the cytokinin mimicking compound thidiazuron (TDZ) is more effective than true cytokinins for indncing shoot organogenesis in woody species (Van Nieuwkerk et al. 1985; Escalettes and Dosba 1993; Hassan et al. 1993; Huetteman and Preece 1993; Perales and Schieder 1993). hi a second experiment, we tested whether combinations of TDZ and true cytokinins could elfllance the transformation efficiency. All media compared in table 1 contained 5 mg/l TDZ in combination with a second cytokinin (either zeatin, BAP or 2iP) at either 2 or 5 rag/1. Addition of zeatin or 2iP to the postcultivation media resulted in transformation effciencies between 0.8 and 1.5% which are significantly higher relative to media containing BAP (between 0 and 0.6%). Moreover, the effect on transformation efficiency of three different a u x i n s at a concentration of 0.25 rag/1 (IBA, NAA, IAA) was compared (table 1). However, no significant differences in transtbnnation efficiencies were obtained by varying the auxin type. As can be seen in table 1, the addition of the gibberellin GA 3 at 0.1 my/1 to the postcultivation medium did not increase transformation efficiency. All media described hereafter contain the grow~th

552

Table 1: Influence of the hormonal compositionof the postcultivation medium on the percentage of apple (cv. Jonagold) explants regenerating kan,'unycin resistant shoots E x P

Gelling Agent

Hormonal Composition

(g/I)

Transformation efficiency

TDZ (rag/I)

Cytokinin (rag/I)

Auxin (mg/I)

GA 3 (mg/I)

# explants tested

#Km a shoots

% Km R shoots 1 a

1

gelrite (2.5)

0

Zea (5)

IBA (0.25)

0

400

0

0.0

1

gelrite ( 2 . 5 )

0

BAP (5)

IBA (0.25)

0

40O

1

0.3

1

gelrite (2.5)

0

2iP (5)

IBA (0.25)

0

400

0

0.0

8 a

1

gelrite (2.5)

5

0

IBA (0.25)

0

400

5

1.3

1

agar (5)

5

0

IBA (0.25)

0

400

0

0.0

2

gelrite (2.5)

5

Zea (2)

IBA (0.25)

0

600

8

1.3

2

gelrite (2.5)

5

Zea (5)

IBA (0.25)

0

600

9

1.5

b a

2

gelrite (2.5)

5

BAP (2)

IBA (0.25)

0

605

2

0.3

2

gelrite (2.5)

5

BAP (5)

IBA (0.25)

0

600

0

0,0

2

gelrite (2.5)

5

2iP (2)

IBA (0.25)

0

595

5

0,8

2

gelrite (2.5)

5

2iP (5)

IBA (0.25)

O

600

6

1.0

2

gelrite (2.5)

5

2iP (5}

NAA (0.25)

0

600

6

1.0

2

gelrite (2.5)

5

2iP (5)

IAA (0.25)

0

600

3

0.5

3

gelrite (2.5)

5

2iP (5)

IBA (0.25)

0

390

9

2.3

3

gelrite (2.5)

5

2iP (5)

IBA (0.25)

0.1

280

5

1.8

a

b b a a a a

a

IData differing significantly within one experiment at the 9 5 % confidence level are indicated with different superscript characters.

Influence of the vitamin composition of the postcultivation medium. According to Patat-Ochatt et al. (1988), the vitmnin components in the medimn for regeneration of apple shoots from protoplast-derived calli are at least as important as the hormonal composition. We therefore compared two vitamin treatments: our reference vitmnin composition consisting of 5 vitamfils (see M&M) and a complex mixture of 10 vitamins as described by Patat-Ochatt et al. 0988). On the medium containing 5 vitamins, 5 shoots regenerated out of 280 apple cv. Jonagold explants (1.8% transformation efficiency), whereas on the medium with the 10 vitamins, 6 shoots regenerated out of 420 explants (1.4% transformation efficiency). These transformation efficiencies are not significantly diftbrent at the 95% comqdence level.

surviving Agobacteria. On the other hmld, 11o explant overgrowth was observed on media containing either 400 my/1 triacillin or a combination of 200 my/1 cetbtaxime and 200 mg/1 triacillin. The media with the combination of cefotaxime mid triacillin resulted in a transfonnation efficiency of 2.2% (3 shoots out of 115 explants) whereas the media with triacillin alone yielded only 0.8% (1 shoots out of 125 explants) regenerating explants. At relatively low concentrations (200 rag/l), cefotaxime has previously been found to have a stimulatory effect on regeneration from leaf explants of apple and pear (James et al. 1990; Maheswaran et al. 1992 and Predieri et al. 1989). It is concluded that cefotaxime allows efficient shoot regeneration but can not be used for adequate control ofA. tumefaciens strain EHA101. This is in agreement with Maheswarml et al. (i 992), who have previously reported that the growth of A. tumefaeiens EHA101 is not efficiently controlled by cefotaxime alone.

hlfluence of the A. tumefaciens killing antibiotics in the postcultivation medimn. Two different antibiotics, cefotaxime (400 rag/l) and triacillin (400 rag/l), as well as a combination of both antibiotics at 200 rag/1 each, were compared tbr their effect on transtbnnation efficiency and tbr their capacity to control explant overgrowth by A. tumefaciens. On selective postcultivation medimn containing 400 rag/1 cefotaxime, 5% of the explants (5 shoots out of 100 explants) regenerated kanamycin resistant buds. Soon after transfer of these buds to the light on elongation medimn, all buds were overgrown by

hlfluence of the addition of malate in the postcultivation medium. Nichol et al. (1991) studied the inflnence of several organic acids on the regeneration ofMedicago sativa L. shoots from callus tissue, hi their experiments and in our preliminary comparative experiments, malate was one of the most stimulatory organic acids for regeneration. We have therefore assessed the effect of the addition of 0.45 g/l malate in the postcultivation medium on transformation efficiency. On the postcultivation mediuna containing 0.45 g/l malate 37 shoots regenerated out of 1590 explants (2.3 % transformation

regulators TDZ (5 mg/1), 2iP (2 rag/l) and IBA (0.25 mg/1) (tmless stated otherwise).

553 efficiency), which is sig::ificantly higher (at the 95 % confidence level) than the regeneration efficiency o1: the same postcultivation medium lacking malate (31 shoots out of 2190 explants or 1.4 % transformation efficiency). Influence of the explant age. In our previous study, we showed that apple leaves (cv. Jonagold) are most competent for regeneration and most susceptible to Agrobaeterium mediated gene transfer when isolated fi-om invit~.v grown shoots 20 to 40 days after multiplication (De B0ndt et al. 1994). The influence of the explant age on the regeneration of kanamycin resistant shoots from apple cv. Jonagold leaves is depicted in Fig. 2. About 2% of the leaves collected from4-week-old shoots regenerated kanamycin resistant shoots. The transformation efficiency decreases drastically with increasing shoot age. Leaves from shoots yotmger than 4 weeks are difficult to manipulate and axe more susceptible to A. mmefaeiens overgrowth during cocultivation (unpublished results). Therefore, the optimal age of the leaves used in our transformation system appears to be 4 weeks.

explants transfom:ed with EHAI01 (pFAJ3003) it: a parallel experiment regenerated 3 kanamycin resistant shoots. Transformation of other genotypes. We applied our standard transformation protocol developed for Jonagold to other apple cultivars, namely Braeb:ma, Elstar, Fuji, Gala, Golden Delicious, Jonagold King, Jonagored and Merlijn, as well as one pear cultivar, namely Conf6rence (Fig. 3). This protocol consists of cocultivation of wounded leaves with EHAI01 (PEHA101/pFAJ3003) during 4 days at 25~ on AS188 (see M&M) followed by a postcultivation on the same medium supplemented with kanamycin (100 rag/l), cefotaxime (200 rag/l) and triacillin (200 rag/l). The transformation efficiency appears to depend strongly on the genotype. Significant differences are marked in Fig. 3. Jonagold King, Jonagored and Jonagold show comparable transformation efficiencies (1-2%) which could be expected since Jonagold King and Jonagored are fruit color nmtants of Jonagold. Golden Delicious and Merlijn yielded somewhat lower transtbnnation efficiencies (0.6 m:d 0.4%, respectively). The transfommtion efficiency detem:ined here for Golden Delicious is lower than that reported by Sriskandarajah et al. (1994) tbr the stone

3./

2,5."

4 a

3,5

2 "/ 0 0 e-

1.6.

i

/

1 ./

oO

bc

1 .S

~,h

ab

0

....

"

C

0.5 . / O, / 27 d a y s

33 days

39 days

45 days

ii

r

Figure 2: Influenceof explant age on the percentage of apple cv. Jonagold explants regenerating kanamycin resistant shoots. The explants were isolated from shoots 27, 33, 39, or 45 days aRer multiplication. The tests were performed on 1200, 1100, 1200 and 500 leaves respectively. Data differing significantly at the 95% confidence level are marked with ditl~rent characters. Selection on phosphinothricin containing media. To find out whether the herbicide phosphinothricin is a better alternative to the mltibiotic kanamycin for the selection of trm:sgenic apple cv. Jonagold shoots, experiments were set up whereby A. tumefaciens strain EHA105 (pEHA105/pFAJ3027)was used instead of EHA 101 (pEHA 10 I/pFAJ3003). Vector pFAJ3027 contains a selectable marker cassette based on the phosphinothricin detoxification gene bar'. The postcultivation medium used in this experilnent was identical to the standard postcultivation medium except that NZamine was omitted and that kanamycin was replaced by 10 rag/1 phosphinothricin (PPT). NZamine was omitted because Dekeyser et al. (1989) have previously found that the uptake of phosphinottucicin can be antagonized by amino acids in the medium. Of the 400 explants transformed with EHA105 (pFAJ3027), none regenerated PPT :'esistant shoots. On the other hand, 400

eq

--;

-S

b

b ~

~ =.a

b :~.

-5

Figure 3: Influence of genotype on the percentage of apple explants regenerating kanamycin resistant shoots: Golden Delicious {(3. Deli.), Jonagold King (J. King), Jonagored (J. red.), Jonagold (Jona.), Elstar, Gala, Braebum (Breb.), Merlijn and Fuji. Each genotype was tested on 170, 300, 280, 250, 210, 280, 200, 440 and 200 explants respecfively Data differing significantlyat the 95% comfidencelevel are marked with different characters.. cultivar (1-2%), which suggests that their transfonnation protocol is better suited for Golden Delicious. The cultivar Elstar outperfom:s the other genotypes with almost 4% of the explants regenerating shoots. Braebum, Fuji and Gala did not regenerate shoots on selective medium. Tllere is a striking correlation between these data on the relative transtbrmation efficiencies of different apple cultivars and those obtained in onr previous study based on the number of GusA expressing calli tbnned per leaf explant on selective medium (De Bond: et al. 1994). h: both comparative experiments, Elstar pertbnns best and Gala, Braebum m:d Fuji appear highly recalcitrant. There is, however, no correlation at all between shoot transformation efficiency and the number of zones showing transient GusA expression as measured innnediately after cocultivation (De Bondt et al. 1994). 111 the latter experiment,

554 Elstar was only weakly and Braebum highly competent to A. tumefaeiens mediated DNA transfer. Hence it must be concluded that transient expression data have little or no predictive value for determining the overall efficiency of transformation, as, was also pointed out by McCabe and Christou (1993).

GusA expression assay and I)NA blot analysis. To assess for chimaerism of the transgenic shoots obtained by our transformation procedure, regenerated shoots were micropropagated and a whole shoot was stained histochemically for GusA activity. O f 8 independent transgenic apple (cultivar Jonagold) clones, all shoots analysed in this way were tmifonnely stained but showed different levels of expression (results not shorn1). The stem of the shoot did not show any staining except at the cut surface, indicating that the GusA substrate (X-Gluc) does not penetrate the intact stem tissue. Two independent transgenic clones and one untransfomted clone of apple cultivar Jonagold were also analysed by DNA blotting using a digoxigenin labelled nptlI probe. As can be seen in Fig. 4, the probe hybridized with high molecular weight tmdigested genomic DNA and with fragments longer than 5.6 kb of HindlII digested DNA. The size of 5.6 kb corresponds to the distance between the right border of the T-DNA and the unique HindlII restriction site in pFAJ3()03.

Figure 4: Southern blots oftmcut (lanesl, 4 and 5), EcoRI digested (l,'mes 2 and 6) and HindlIl digested DNA (lanes 3 and 7) of two independent transgenic clones (clone #1 :lanes 1 --~ 3; clone #2: lane 4 and lanes 5 --~ 7) probed with digoxigenin labeled nptlI probe. The DNA in lane 4 is isolated from an untransfonned shoot. Hybridisation bands t?om HindlII digested DNA larger than 5.6 kb are indicative of insertion in plant DNA. The clone represented ill lanes 1-3 (clone #I) of Fig. has most probably two separated insertions of the T-DNA, whereas the clone represented in lanes 5-7 (clone #2) has probably three sites of insertion with at least one in the tbnn of tandem repeats as suggested by the intense 6.(1 kb braid. On EeoRI digested DNA, the nptll probe hybridizes to a 1.1 kb fragment corresponding to the DNA fragment situated between the two EcoRI sites surrounding the nptlI gene in vector pFAJ3003. It is concluded from these data that at least in the two clones mmlyzed, the T-DNA is integrated into tile apple genome.

Transgenic Jonagold clones were rooted and acclimated to growth chamber conditions. The plants acclimated equally well as untransformed in vitro grown apple cultivar Jonagold shoots (results not shown).

ACKNOWLEDGEMENTS

This work was supported in part by the 'Vlaams Aktieprogralmna Bioteelmologie' of the Flemish Government (Project VLAB 014)' and by the N.V. Jo Nicolai & Co. W.F. Broekaert is Senior Research Associate of the Belgian 'Nationaal Fonds voor Wetenschappelijk Onderzoek'. The authors wish to thank Dr. J. Bottennan (Plant Genetic Systems N.V., Gent, Belgium) for providing the plant transformation vectors pDE3 and pGSC 1701A2, Pros J. Keulemans mad J. Nicolai for stimulating support. REFERENCES

Aldrich J; Cullis CA (1993) Plant Mol Biol Rep 11:128-141 Becker D, Kemper E, Schell J, Masterson R (1992) Plant Mol Biol 20: 1195-1197 Cornelissen M, Vandewiele M (1989) Nucleic Acids Res 17:19-29 De Bondt A, Eggermont K, Druart P, De Vii M, Goderis I, Vanderleyden J, Broekaert W (1994) Plant Cell Reports 13:587-593 Dekeyser R, Claes B, Marichal M, Van Montagu M, Caplan A (1989) Plant Physio190:217-223 Druart P (1980) Scientia Hortic 12:339-342 Druart P (1988) Acta HoiSt227:369-380 Druart P (1990) Aeta Hort 280:117-124 Escalettes V, Dosba F (1993) Plant Sci 90:201-209 Fasolo F, Zinunerman RH, Fordham I (1989) Plant Cell Tiss Org Cult 16:75-87 Hanlill JD, Rounsley S, Spencer A, Todd G, Rhodes MJC (1991) Plant Cell Reports 10:221-224 Hassan MA, Swartz HJ, Inamine G, MuIlineaux P (1993) Plant Cell Tiss Org Cult 33:9-17 Hood EE, Helmer GL, Frayley RT, Chilton M-D (1986) J Bacteriol 168:1297-1301 Hood EE, Gelvin SB, Melchers LS, Hoekema A (1993) Transgenic Research 2:208-218 Huetteman CA, Preece JE (1993) Plant Cell Tiss Org Cult 33:105-119 James DJ, Passey AJ, Barbara D (1990) In: Genetic engineering of crop plants, 239-248, (eds.) Lyeett GW, Grierson D, Butterworths-London James DJ, Passey AJ, Rugini E (1988) J Plant Physiol 132:148-154 James DJ, Passey AJ, Barbara DJ, Bevan MW (1989) Plant Cell Reports 7:658-661 James DJ, Uratsu S, Cheng J, Negri P, Viss P, Dandekar AM (1993) Plant Cell Reports 12:559-563 Maheswaran G, Welander M, Hutchinson J F, Graham M W, Richards D (1992) J Plant Physiol 139:560-568 McBride KE, Summerfelt KR (1990) Plant Mol Biol 14:269-276 McCabe D, Christou P (1993) Plant Cell Tiss Org Cult 33:227-236 Murashige T, Skoog F (1962) Physiol Plant 15:473-497 Nichol JW ,Slade D, Viss P, Stuart DA (1991) Plant Sci 79:181-192 Norelli JL, Aldwinckle HS, Dest~fano-Beltran L, Jaynes JM (1994) Euphytica 77:123-128 Patat-Ochatt E, Ochatt SJ, Power JB (1988) J Plant Physiol 133:460-465 Perales EH and Schieder O (1993) Plant Cell Tiss Org Cult 34:71-76 Predieri 8, Fasolo F, Passey ALl,Ridout MS, James DJ (1989) J Hort Sci 64:553-559 Sriskandarajala S, Skirvin RM, Abu-Qaoud H, Korban SS (1990) J Hort Sci 65:113-121 Sriskandarajab S, Goodwin PB, Speirs J (1994) Plant Cell Tiss Org Cult 36:317-329 Vancanneyt G, Schmidt R, O'Connor-Sanchez A, Willnfitzer L, RochaSosa M (1990) Mol Gen Genet 220:245-250 Van Nieuwkerk JP, Zinunerman RH, Fordham I (1985) HortSci 20:523 Welander M, M,'daeswaran G (1992) Plant Physiol 140:223-228