Mol Gen Genet (1992) 233 : 53-64
© Springer-Verlag 1992
Petunia plants escape from negative selection
against a transgene by silencing the foreign DNA via methylation Suzy Renckens 1, Henri De Greve 1, Marc Van Montagu ~'2, and Jean-Pierre Hernalsteens 1 1 Laboratorium voor GenetischeVirologic, Vrije Universiteit Brussel, Paardenstraat 65, B-1640 St-Genesius-Rode, Belgium 2 Laboratorium voor Genetica, RijksuniversiteitGent, K.L. Ledeganckstraat35, B-9000 Gent, Belgium Received September 1, 1991
Summary. Transgenic Petunia hybrida clones harbouring the T-DNA gene 2 of Agrobacterium tumefaeiens were used to test a strategy for the trapping of plant transposable elements. In the Petunia line used, floral variegation is due to the presence of the non-autonomous transposable element dTphl at the Anl locus. The gene 2 product converts the auxin precursor indole-3-acetamide and its analogue 1-naphthalene acetamide into the active auxins indole-3-acetic acid and 1-naphthalene acetic acid. Plant cells that express gene 2 can use a low concentration of the precursors as auxins and become sensitive to the toxicity of high concentrations of these compounds. By selecting protoplast-derived microcalli or seedlings able to grow on medium with high precursor concentrations, variant plants were obtained in which gene 2 was no longer expressed. Southern analysis, using gene 2-specific probes, revealed that in one variant the T-DNA was deleted. For 30 other variants no alteration in gene 2 structure was observed, indicating that transposable element insertion was not responsible for the inactivation of gene 2. Analysis with restriction enzymes allowing discrimination between methylated or non-methylated DNA sequences showed that the inactivated gene 2 sequences were methylated. Addition of the in vivo methylation inhibitor 5-azacytidine to the medium led to reactivation of gene 2 expression in some of the variants. These observations demonstrated that reversible DNA methylation was the main cause of silencing of gene 2 in this system. Key words: DNA methylation - Petunia hybrida - Transgenie plants - Gene inactivation- Agrobacterium tumefaciens
Introduction The T - D N A onc genes of the Ti plasmid are responsible for the hormone-independent growth of plant tumours
Correspondence to . S. Renckens
following infection by Agrobacterium tumefaeiens (reviewed in Inz~ et al. 1987; Weiler and Schr6der 1987). These onc genes have regulatory sequences similar to those of typical eukaryotic genes. They influence the differentiation of higher plants by encoding the biosynthesis of auxins and cytokinins, thereby changing the endogenous hormonal balance. The protein encoded by gene 2 (also iaaH or tins2) is the enzyme indole-3-acetamide hydrolase, which converts indole-3-acetamide (IAM) into the active auxin indole-3-acetic acid (IAA) (Schr6der et al. 1984; Thomashow and Reeves 1984). Similarly, the synthetic analogue 1-naphthalene acetamide (NAM) is converted into 1-naphthalene acetic acid (NAA) (Inz6 et al. 1984). Several pathways have been proposed for the synthesis of the naturally occurring auxin IAA (for review see Marumo 1986). According to Sembdner et al. (1980) IAM is not an intermediate in auxin biosynthesis in plants. Normal plant cells do not contain other compounds that can be converted into biologically active auxin by the gene 2 product (Follin et al. 1985; Inz6 et al. 1987). Therefore, gene 2 itself does not interfere with endogenous auxin metabolism in plants. Transgenic plant expressing gene 2, as the only T-DNA derived gene, are able to use low concentrations of NAM or IAM as the sole auxin supplement. On the other hand, increased auxin concentrations are known to be toxic to plant cells (Caboche 1980; Muller et al. 1983). Consequently, in association with high concentrations of IAM or NAM, gene 2 can be used as a negative selection marker at the level of the plant or protoplast (Budar et al. 1986; Depicker et al. 1988; Karlin-Neumann et al. 1991). Together with the gene 2 probe, this provides a means to study gene inactivation, mutagenesis and T-DNA stability in higher plants. The original aim of this work was to examine whether the gene 2 selection system could provide us with a general mechanism to trap transposable elements. Transposable elements are interesting tools for the cloning of plant genes (reviewed by Chandlee 1990). A Petunia hybrida line harbouring an active transposable element (Doodeman et al. 1984) was used as a model system. In
54 Petunia, instability at several loci involved in flower anthocyanic colouration has been described. Spontaneous instability at the Anl locus has been genetically proven to be due to a two-element system (Wijsman 1986). The element at Anl transposes at high frequency (Doodeman et al. 1984) to other loci on each of the seven Petunia chromosomes (Gerats et al. 1989). Recently, the nonautonomous transposable element dTphl, causing instability at the Anl locus of P. hybrida W138, has been characterized (Gerats et al. 1990). Use of the gene 2 system also allows selection at the cellular level (Depicker et al. 1988). Additionally, the Agrobacterium transfer system allows the introduction of a single gene copy and consequently does not require the use of haploid cells. Transposable elements may be active in tissue culture (Planckaert and Walbot 1989; James and Stadler 1989) and previously silent controlling elements can even be activated in vitro (Peschke et al. 1987). In the present paper we describe the application of negative selection on Petunia protoplasts and germinating seeds expressing gene 2. Analysis of the resulting variant clones revealed no transposable element insertion. We observed that gene 2 was efficiently shut off by DNA methylation. Deletion of gene 2 was also detected. Materials and methods
DNA techniques. Restriction enzyme digestion, ligation, transformation and plasmid preparation were accomplished by standard methods (Maniatis et al. 1982). Vector construction. The T - D N A vector pGV974 was constructed by introducing the AsuII fragment of pGV0153 (De Vos et al. 1981), containing the T - D N A gene 2 of A. tumefaciens Ach5 (Willmitzer et al. 1982), into the HpaI site of the binary vector pGV941 (Deblaere et al. 1987). The binary vector pGV974 was mobilized from Eseherichia coli GJ23 (Van Haute et al. 1983) to the non-oncogenic A. tumefaeiens strain C58C1Rif R (pGV2260) (Deblaere et al. 1985). Hybridization probes. The AsuII fragment of pGV0153 containing gene 2 was cloned in the AccI site of pUC8 (Vieira and Messing 1982) resulting in the plasmid pGV975. The 2.6 kb PstI-BamHI fragment of pGV975 (probe A) harbours the coding sequence, as well as the 5' and 3' regulatory sequences of gene 2 (Fig. 1). The 1051 bp EeoRI fragment of pGV975 (probe B) contains the first 970 bp of the coding sequence of gene 2. The 690 bp EcoRI fragment of pGV975 (probe C) spans the 3' end of the coding sequence and the 3' regulatory sequences of gene 2. The 1800 bp HindIII-BamHI fragment of plasmid pKC7 (Rao and Rogers 1979) was used to detect nptII sequences (probe D). The DNA fragments used as hybridization probes were extracted from agarose gels by electroelution (Allington et al. 1978) and radiolabelled using an Amersham Multiprime DNA labelling kit (RPN. 1601). Clonin9 and sequencin9. After protection of the EcoRI sites with EcoRI methylase and addition of EcoRI lin-
kers, HindIII DNA fragments hybridizing to probe B were cloned into the EcoRI site of the vector lambda gtl0 (Murray et al. 1977). For selection of recombinant phages, the host strain C600Hfl was used (Protoclone Lambda gtl0 System, Promega Biotec). The dideoxy method (Sanger et al. 1977) was used for DNA sequence analysis. Plant material. The P. hybrida line W138, harbouring an active transposable element (Doodeman et al. 1984) was provided by Dr. A. Gerats. To improve its tissue culture and regeneration properties, the F1 hybrid, F1E, of the Petunia lines W138 and Mitchell (Mitchell et al. 1980) was used. Plant media. Medium 277 is composed of the minerals of the medium of Murashige and Skoog (1962), 30 g/1 sucrose, the vitamins of B5 medium (Gamborg et al. 1968) and 7 g/1 agar. Medium 15 is medium 277 supplemented with 1 mg/1 6-benzylaminopurine (BAP) and 0.1 mg/1 NAA. Medium 236 is half-strength hormone-free Murashige and Skoog (1962) medium solidified with 7 g/1 agar. Medium 195 is B5 medium (Gamborg et al. 1968) with 0.4 M sucrose, supplemented with 0.5% Cellulase R10 and 0.2% Macerozyme R10. The pH is adjusted to 5.6 before filter sterilization. Medium 19 is K3 medium (Nagy and Maliga 1976) without growth regulators. Medium 313 is medium 19 supplemented with 20 mg/1 sodium pyruvate, 40 mg/1 malic acid, 40 mg/1 citric acid, 40 mg/1 fumaric acid, 0.56 mg/1 BAP, 0.93 rag/1 NAA and 0.11 mg/1 2,4-dichlorophenoxyacetic acid (2,4-D). The pH is adjusted to 5.7 before filter sterilization. All cultures were grown in 9 cm plastic petri dishes. Plant transformation. P. hybrida F1E was transformed by the Agrobacterium strain C58C1Rif R (pGV2260) (pGV974) using the leaf disc method (Horsch et al. 1985). Infected leaf discs were incubated on medium 15, supplemented with 100mg/1 kanamycin sulphate and 500 mg/1 cefotaximum (Claforan, Hoechst, Frankfurt, FRG). The resulting shoots were either propagated in vitro on medium 236, or transferred to soil in the greenhouse and propagated by cuttings in Jiffy-7 peat discs. Test for expression of introduced 9enes. The expression of the chimeric nosmptII gene was investigated by testing the ability of the selected clones to root on medium 236 supplemented with kanamycin sulphate (50 rag/l), or to produce callus and shoots from leaf explants on medium 15 containing geneticin (50 and/or 100 mg/1). Expression was confirmed by means of an enzymatic assay (McDonnell et al. 1987). The expression of gene 2 was tested by incubating leaf fragments on medium 277 supplemented with 1 mg/1 BAP and 1 mg/1 NAM. Clones expressing gene 2 develop callus and roots on this medium, while leaves of untransformed plants produce shoots (Budar et al. 1986). Alternatively, shoots were tested on medium 236 supplemented with 1 mg/1 NAM. After 14 days, plants expressing gene 2 develop hairy white callus or only a few thick abnormal roots on this medium, while those not expressing gene 2 produce a normal root sys-
55 tern. This rooting test for gene 2 expression is fast and reproducible. It allows detection of intermediate levels of expression.
medium, whereas seedlings not expressing gene 2 developed normal roots. To test nptII expression, seeds were sown on medium 236 containing 20 mg/1 geneticin.
Plant DNA isolation and hybridization conditions. Total plant D N A was prepared from in vitro material as described by Dellaporta et al. (1983). The digested D N A (5 gg per lane) was subjected to electrophoresis in 0.8% agarose gels, transferred (Southern 1975) to nylon membranes (Hybond-N; Amersham) and hybridized with radiolabelled probes as described (Membrane transfer and detection methods, Amersham).
Selection of seedlings. To select for inactivation of gene 2, surface-sterilized seeds were sown on medium 236 supplemented with 1 mg/1 N A M and 10 mg/1 geneticin. Plants that developed normally on the selection medium were tested again for their ability to root on medium 236 containing 1 mg/1 N A M and to produce callus on medium 15 containing 25 mg/1 geneticin.
5-Azacytidine treatment. Reactivation of gene 2 expression in the presence of a methylation inhibitor was tested by incubating seeds or cuttings on medium 236 containing 1 rag/1 N A M and 30gM 5-azacytidine (5-AZA). Medium 236 containing only 5-AZA (30 gM) was used as a control.
Selection for inactivation of 9ene 2 in mesophyll protoplasts. Axenic plants were propagated from cuttings in jars containing medium 236 under a 16 h light (2000 lx)/ 8 h dark cycle at 24 ° C. Mesophyll protoplasts were isolated by cutting 10 to 15 fully expanded young leaves into thin strips (1-2 mm) followed by overnight incubation in 10 ml of enzyme mixture (medium 195). After careful mixing, the suspension was filtered through a 100 gm sieve and centrifuged at 100 9 for 12 rain. The floating protoplasts were resuspended in medium 19, centrifuged again at 100 g for 12 rain and cultured at a density of 105 protoplasts/ml in medium 313. After 1 week, 25 000 microcolonies were plated onto solid medium 313, without NAA, and supplemented with 5 to 200 mg/1 N A M or IAM. Surviving calli were transferred to medium 15 for regeneration. Calli not producing shoots after two transfers on this medium were transferred to B5 medium (Gamborg et al. 1968) containing 1 mg/1 of zeatin as the only growth regulator. Elongated shoots formed roots after transfer onto medium 236 and regenerated into normal plants.
Introduction of 9ene 2 and analysis of transgenic Petunia plants The Petunia hybrid line F1E (W138x Mitchell) was transformed using the Agrobacterium strain C58C1Rit ~ (pGV2260)(pGV974). The plasmid pGV974 contains, in its T-region, both gene 2 and a chimeric 5'nos-nptII-3'ocs gene encoding neomycin phosphotransferase II, that confers kanamycin and geneticin resistance to plant cells (Fig. 1). All 51 independent plants, obtained by incubation of Agrobacterium-infected leaf discs on regeneration medium containing kanamycin, were morphologically normal. The expression of the chimeric nos-nptII gene in these plants was confirmed by testing the ability to root on hormone-free medium containing kanamycin or to produce callus from leaf fragments on medium containing geneticin. Expression of gene 2 was tested by incubating leaf fragments on culture medium, supplemented with BAP and with N A M as only auxin source. After
Segregation of transgenes. Seeds from transgenic Petunia plants were surface-sterilized by treatment with a 10% solution of commercial bleach (15°Chl) containing 0.05% Triton X 100 for 30 rain and rinsed with sterile water. After drying, the seeds were germinated on medium 236 containing 1 mg/1 NAM. Seedlings expressing gene 2 were unable to root on NAM-containing
EeoRI HindIII BstNI/EcoRII HpaII/MspI
>73o > 951
I 421 I
Fig. 1. T-DNA of pGV974. Sizes of restriction fragments and minimal sizes of border fragments are indicated in bp. RB, LB, right and left T-DNA borders; 5'-GENE 2-3', gene 2 with original 5'
and 3' regulatory sequences; 5'NOS-NPTII-3'OCS, chimeric nptII gene encoding neomycin phosphotransferase. Regions of the probes that are homologous to the T-DNA are shaded
56 normal root system (Fig. 2B). Thirty-three plants expressed both the nptII gene and gene 2 (Table 1). In five plants only gene 2 expression was detected, whereas four plants expressed only the nptII gene. Nine plants showed no transgene expression. The number of T - D N A inserts in the transgenic plants was determined by Southern blotting. HindIII and/or EeoRI digests of the plant D N A were hybridized with a probe homologous to gene 2, including its regulatory sequences (probe A, Fig. 1). Twelve transformed plants harboured a single copy of gene 2. Eleven of these plants expressed the nptII gene (Table 1). This nptII expression was confirmed by enzymatic assay. Leaf discs of ten "single-copy" plants expressed gene 2 (Table 1). Transformed plants expressing only neomycin phophotransferase activity all contained at least one copy of gene 2. Apparently gene 2 expression was either absent, or below the phenotypically detectable level. Similarly, out of nine plants not showing expression of either of the introduced genes, only six contained no gene 2 copy.
Segregation of gene 2 in the progeny of transformed plants
Fig. 2 A, B. Phenotypic expression of gene 2 in leaf explants and shoots of transgenic plants. A On medium containing 1-napthalene acetamide (NAM), explants from wild-type F1E plants (left) produce shoots, whereas the transgenic plant expressing gene 2 (right) develops callus and roots. B Shoots of untransformed F1E plants (left) develop a normal root system on medium containing NAM; transformed shoots expressing gene 2 (right) cannot root normally on this medium as a result of the conversion of the precursor into an active auxin
Seeds, from self-fertilization of the transformed singlecopy plants expressing gene 2, were germinated on hormone-free medium, supplemented with NAM. Seedlings expressing gene 2 were unable to grow on this medium, but formed a white callus from the radicle (Budar et al. 1986) owing to the conversion of N A M into a toxic quantity of NAA. Untransformed seedlings developed normally on NAM-containing medium. The segregation values indicated that the transmission mode of the gene into the progeny was not in all cases that expected for a dominant Mendelian marker (Table 2).
Table 1. Expression of transgenes in plants obtained from the transformation experiment
Selection for inactivation of gene 2 on mesophyll protoplasts
Expression of transgenes gene 2 nptII
Total number assayed
Number of gene 2 copies 0 1 >1
ND 3 1 1
+ or - , expression or no expression of the transgene, respectively. ND, number of gene 2 copies was not determined
3 weeks, shoot development started at the edges of explants from untransformed control plants due to the cytokinin present in the medium. In contrast, transformed leaf fragments expressing gene 2 developed calli and roots, indicating conversion of N A M into an active auxin (Fig. 2A). On hormone-free medium, supplemented with N A M , cuttings of these transformed plants were unable to root, while untransformed plants developed a
Mesophyll protoplasts were prepared from the untransformed control plant F1E and the transformed plants 7a, 14a, 32a, 61a (one gene 2 copy) and 9a and 13a (three gene 2 copies) expressing both transgenes. Plants with three T - D N A copies were included in the experiment as controls. The probability that all three copies would be inactivated by transposable element insertion is negligible. The expression of gene 2 was tested on freshly isolated protoplasts by culture in modified medium 313, in which the auxins N A A and 2.4-D were replaced by the precursors N A M (0.5 mg/1) or IAM (0.5 mg/1). Wild-type Petunia protoplasts did not divide in the absence of an active auxin; protoplasts of the transgenic plants expressing gene 2, on the contrary, developed into callus with an efficiency comparable to that on medium 313. Addition to the medium of N A A concentrations above 10 rag/1 was severely toxic for non-transformed F 1E cell clones. Subsequently, selection against gene 2 expression was performed. One-week-old protoplast-derived cell clones
57 Table 2. Expression of transgenes in the progeny of single gene 2 copy plants Plant
7a 14a 18a 23a* 32a 37a 44a* 50a 55a 60a 61a 63a
nptII gene nptlI +
Gene 2 +
76 140 118 3 201 145 57 124 39 ND 51 ND
33 36 165 51 71 50 20 35 131 ND 83 ND
1.6 1.9 167.4 138.9 0.18 0.04 0.04 0.76 245.7 ND 97.5 ND