Using resistance genes from a crossable donor to obtain cultivars resistant to diseases and the use of such cultivars in production appears an economically and environmentally advantageous approach. In apple, introgression of resistance genes by classical breeding results in new cultivars, while introducing cisgenes by biotechnological methods maintains the original cultivar characteristics. Recently, plants of the popular apple ‘Gala’ were genetically modified by inserting the apple scab resistance gene Rvi6 (formerly HcrVf2) under control of its own regulatory sequences. This gene is derived from the scab-resistant apple ‘Florina’ (originally from the wild apple accession Malus floribunda 821). The vector used for genetic modification allowed a postselection marker gene elimination to achieve cisgenesis. In this work, three cisgenic lines were analysed to assess copy number, integration site, expression level and resistance to apple scab. For two of these lines, a single insertion was observed and, despite a very low expression of 0.07- and 0.002-fold compared with the natural expression of ‘Florina’, this was sufficient to induce plant reaction and reduce fungal growth by 80% compared with the scab-susceptible ‘Gala’. Similar results for resistance and expression analysis were obtained also for the third line, although it was impossible to determine the copy number and TDNA integration site–such molecular characterization is requested by the (EC) Regulation No. 1829/2003, but may become unnecessary if cisgenic crops become exempt from GMO regulation.
Cisgenesis is a promising tool to rapidly deploy disease resistance in apple. By adding only the gene(s) of interest, and therefore the corresponding phenotypic trait(s) from a crossable donor to an existing cultivar, it circumvents most of the problems encountered during traditional resistance breeding (Jacobsen and Schouten, 2008). Addition of a single trait by genetic engineering to an apple plant maintains all the characteristics of the original cultivar (Borejsza-Wysocka et al., 2010).
Decreased susceptibility of the popular cultivar ‘Gala’ to the fungus V. inaequalis, which causes apple scab, is a highly desirable trait, as it would reduce fungicide input during the production of fruits of this cultivar. This, in turn, would be advantageous for the environment, the consumer and the producer (Gessler, 2011). Cisgenic scab-resistant apple lines may cause less concerns compared with a transgenic product, for example allergenicity caused by the product of the inserted gene, as the trait is already present in traditionally bred scab-resistant cultivars (e.g. ‘Florina’, ‘Santana’) and avoids the generation of a completely new cultivar with unpredictable agronomic qualities. So far, the only scab resistance gene to be cloned is HcrVf2 (newly Rvi6), originally derived from Malus floribunda 821 and was shown to incite scab resistance in the susceptible apple ‘Gala’ under the control of the 35S promoter (Belfanti et al., 2004) and under its own native promoter (Szankowski et al., 2009). However, the plants used in these studies still contained the selection marker gene nptII and foreign regulatory sequences. Joshi et al. (2011) also generated transgenic lines with HcrVf1 and HcrVf2 (newly Rvi6) using the vector pMF1 and their own regulatory sequences.
Cisgenic lines of apple ‘Gala’ carrying the Rvi6 scab resistance gene have been reported recently (Vanblaere et al., 2011). The endogenous apple scab resistance gene Rvi6 with its native regulatory sequences was inserted into the genome of the susceptible apple ‘Gala’. Cisgenic plant lines were generated through the Agrobacterium-mediated transformation method developed by Schaart et al. (2004), which consists of a two-step process: transgenic lines containing the Rvi6 gene and an excisable cassette carrying the marker fusion gene codA/nptII and the RecLBD gene were first generated. Then, a chemical activation of the recombination led to excision of the cassette resulting in a truly cisgenic genotype. During this process, Vanblaere et al. (2011) reported the regeneration of ten transgenic lines, eight of them amplifying–in addition to the target gene Rvi6 and plant selection gene codA/nptII–sequences corresponding to the bacterial selection gene nptIII present on the backbone of the pMF1-vector. Two transgenic lines without backbone (T11.1 and T12.1) and one (T7.1) with backbone were selected for detailed analysis, together with the cisgenic lines (C7.1.49, C11.1.53, C12.1.49) derived from the aforementioned transgenic lines. Following recombination, no amplification of the nptIII gene was obtained from the cisgenic lines, suggesting that the corresponding vector backbone sequences were excised during the recombination step resulting in cisgenic line C7.1.49. Copy number assessment by Southern blot hybridization showed a single integration of the marker gene codA/nptII used for selection in each of the three transgenic selected mother lines (Vanblaere et al., 2011).
Cisgenesis in apple as reported here has been achieved by an Agrobacterium-mediated transformation technique that randomly inserts the cisgene of interest into the genome of the target cultivar by T-DNA transfer to the host cell. Due to the process of the integration, genomic rearrangements and the interruption of important functional genes or pathways can occur, which could alter the phenotype, quality and traits of the plant (Rosati et al., 2008). The random integration of the T-DNA into different loci in the genome can also be responsible for variation in the transcriptional activity of the inserted gene, thereby altering its phenotypic expression (Day et al., 2000). Identification of genomic sequences flanking the integration site allows detailed investigation of the transition sequences. Moreover, the availability of the ‘Golden Delicious’ genome (Velasco et al., 2010) coupled with EST and gene prediction information may allow estimation of where the integration took place in the genome.
The legal regulation of GM plants in Europe (e.g. (EC) 1829/2003) requires molecular characterization, including the amount of vector and/or donor nucleic acid remaining in the modified organism, sequence data for the flanking regions, level of expression of the specific protein resulting from the genetic modification and information on the organization of the inserted genetic material at the insertion site and should specifically aim to identify whether the genetic modification(s) raise(s) any issues regarding the potential for producing new toxins or allergens (EFSA 2011).
In this study, we report the examination of the insertion site of the T-DNA (based on the ‘Golden Delicious’ genome, which is a parent of ‘Gala’) and the T-DNA/plant genomic DNA junction of these lines. Southern blot hybridization with a probe designed for the cisgene (Rvi6) was used to determine copy number insertions of the T-DNA in the mother transgenic lines and in the derived cisgenic lines. A quantitative real-time (qRT)-PCR analysis was performed to assess the gene expression in the trans- and cisgenic plants compared with ‘Florina’ (a traditionally bred cultivar carrying the Rvi6 gene). Moreover, scab resistance was evaluated visually as well as by qPCR. The generated data serves as further characterization of the cisgenic lines, as requested for biosafety assessment based on European GMO regulations.
Complete backbone integration in six of 10 transgenic ‘Gala’ lines
Backbone integration was investigated in the lines by PCR. Amplification of one or more of the four amplicons distributed along the pMF1 backbone sequence used for transformation indicated genomic integration of the corresponding sequences. Six of the eight transgenic lines (T7.1, T7.2, T7.3, T7.4, T8.1 and T11.2; Figure 1 and Table 2), which contain backbone sequences as mentioned by Vanblaere et al. (2011), were found to amplify all four amplicons. This indicates that most of the sequences of the pMF1 vector backbone were integrated in these lines either as a single sequence or fragmented in several parts (truncated copies). The other two transgenic lines (T8.2, T8.3) produced only two of the four amplicons, indicating only partial integration of the vector backbone sequences (Table 2). Transgenic lines T11.1 and T12.1 did not amplify any of the amplicons on the backbone, leading to the assumption that only the T-DNA was inserted. In the putative cisgenic lines–‘C'7.1.17, ‘C'7.1.63, ‘C'8.2.7, ‘C'8.3.4, ‘C'8.3.6–at least one of the four PCR products indicating the presence of the vector backbone were amplified following the recombination step, indicating that these lines are not truly cisgenic (and therefore marked as ‘C’), as they contain foreign genes. Cisgenic lines C7.1.3, C7.1.49, C11.1.53 and C12.1.49 were found to be free from backbone integration as no backbone amplicon was detected (data not shown).
Determination of T-DNA border junctions and the genomic point of T-DNA integration in lines T7.1, T11.1 and T12.1
The product of the inverse PCR (iPCR) of line T7.1 was sequenced, and a 379 bp sequence flanking the insertion site was obtained. This sequence matched with the genome of ‘Golden Delicious’ chromosome 4 (contig MDC017356.126) and indicated integration into an exon of the putative transcript MDP0000250224, containing a cyclin-like F-box domain. Primers T7.1_a and T7.1_c (Table 1) were designed for this sequence and amplified a 200 bp amplicon in ‘Gala’.
Table 1. Primer sequences
Primer Sequence (5′–3′)
Similarly, a 343 bp sequence flanking the insertion site was obtained for line T11.1, and BLAST analysis of this sequence against the ‘Golden Delicious’ genome revealed a match to a sequence on chromosome 12 (contig MDC021013.168). In this line, the T-DNA is inserted in a putative ORF (MD00G455290) to which the gene ontology term ‘microtubule cytoskeleton organization’ has being assigned. The 343 bp sequence obtained from untransformed ‘Gala’ using primers T11.1_b/T11.1_d revealed that two different alleles were amplified. For line T12.1, two different sequences of 188 bp and 924 bp flanking the insertion site were obtained by iPCR and matched Golden Delicious genomic sequences on chromosomes 11 (MDC004866.187) and 14 (MDC018357.590). A third sequence of 193 bp was found in line 12.1 flanking the recombination site 1 (RS1) of the transformation vector and corresponded to a Tnos, a terminator sequence present on the pMF1 vector. Primers T12.1_a/T12.1_d developed on both regions flanking the insertion site failed to amplify an amplicon from ‘Gala’ DNA, thus indicating that the two corresponding sequences are noncontiguous in both ‘Golden Delicious’ and ‘Gala’–the insertion site of the T-DNA in plant line T12.1 could therefore not be determined.
PCR was then attempted on the transgenic and cisgenic lines, using the HcrVf2term2 primer in combination with the primers developed on the corresponding sequence flanking the left border (T7.1_c, T11.1_d and T12.1_d, respectively). In both transgenic T7.1 and cisgenic C7.1.49, an amplicon of approximately 1.2 kbp was amplified. An amplicon of about 500 bp was obtained from line C11.1.53, while no amplification was observed in line T11.1 and in untransformed ‘Gala’.
Sequencing of the PCR product amplified in line T7.1 with primers pmf_IPCR_9/T7.1_a showed that the right border and the PacI restriction site used for cloning were not integrated. The integrated T-DNA starts at the promoter of the Rvi6 gene, runs over the marker fusion gene cassette, LB, the whole pMF1 backbone, RB, Rvi6 gene, Tnos terminator sequence and is interrupted at approximately 1/5 of the end of the RecLBD gene. The partial integration of RecLBD is demonstrated by successful amplification using primers Rec1/T7.1_c but not with primer pair Rec2/T7.1_c (Figure 1, Figure 2 and Table 1). In the cisgenic line C7.1.49, following recombination the left border (LB) is also missing. Sequencing of the PCR products amplified in line T11.1 and C11.1.53 with primers pmf_IPCR_9/T11.1_b showed no integration of the right border and the PacI restriction site as well as the first five nucleotides of the Rvi6 (Figure 1 and Table 1). The left border is also absent, and the recombination site (RS) next to the left border is truncated. A minimal integration of foreign DNA (72 bp), resulting from the vector cloning site and the recombination site, was found in cisgenic plant line C11.1.53 (Figure 3).
Copy number of Rvi6 estimated by Southern hybridization
Lines T7.1 (~16 000 bp) and C11.1.53 (~9000 bp) had one additional band for Rvi6 compared with ‘Gala’ (Figure 4). Lines T11.1 (~16000 bp and ~9000 bp), T12.1 (~10500 bp and ~3900 bp) and C12.1.49 (~3900 bp and ~2700 bp) possessed two additional bands for Rvi6. For the cisgenic line C7.1.49, there was no specific band for Rvi6 visible.
Expression level of Rvi6 in the cisgenic lines compared with ‘Florina’
The expression level of Rvi6 in all trans- and cisgenic lines, as measured by quantitative real-time PCR, was below the expression level of the scab-resistant apple ‘Florina’. Rvi6 expression level in lines T7.1 and C7.1.49 was 14-fold lower compared with the control, while in lines T11.1 and C11.1.53, Rvi6 is 300 and 450 times less expressed, respectively. Rvi6 showed sixfold lower expression in line T12.1, whereas in C12.1.49, it was 200 times less expressed than in ‘Florina’. Very low Rvi6 expression could be observed for ‘Gala’, corresponding to 0.000749 times the expression in ‘Florina’ and indicating a nonspecific amplification as ‘Gala’ lacks Rvi6. To display the different expression levels on the same graph, a log-scale transformation was used (Figure 5).
Apple scab resistance evaluation
Susceptible control ‘Gala’ showed abundant sporulation without chlorosis (class 4 symptoms) 21 days after artificial inoculation, whereas the resistant control ‘Florina’ showed no symptoms at all (class 0 symptoms). All three investigated cisgenic lines showed a strong foliar deformation, chlorosis and in some cases also sporulating lesions (class 2-3b symptoms). Results of visual scoring are shown in Table 3.
Pathogen leaf colonization by quantitative PCR
Pathogen/host DNA ratio was calculated following qPCR of the DNA extracted from whole leaves. The CCcorr values corresponding to a colonization coefficient expressed as the ratio between pathogen and host DNA evinced from the infected leaves of the investigated cisgenic lines are shown in Figure 6. ‘Florina’, the resistant control, showed an average CCcorr value of 0.007 (N = 2), while the susceptible control ‘Gala’ reached 0.133 (± 0.079, N = 7) 21 days after inoculation. All cisgenic lines had lower values: C7.1.49 showed a CCcorr value of 0.022 (± 0.020, N = 8), C11.1.53 a value of 0.037 (± 0.029, N = 7) and C12.1.49 a CCcorr value of 0.028 (± 0.013, N = 4).
The European Commission has stated in its latest report that cisgenic crops are becoming more and more accepted in Europe (Gaskell et al., 2010). In the European Union, cisgenics underlie the GMO regulations. We report here molecular investigations on three cisgenic lines of the apple ‘Gala’. The creation of these lines has been described recently (Vanblaere et al., 2011).
Cisgenic line C7.1.49
In this cisgenic line, a single insertion of the T-DNA occurred in chromosome 4 of ‘Gala’, putatively interrupting a predicted exon of a cyclin-like F-box domain. The single insertion was already supported by Vanblaere et al. (2011) using a nptII-specific probe and was confirmed by hybridization with the Rvi6 probe of a 17 kbp band in the mother line T7.1. No extra band is observed in C7.1.49. However, following recombination and subsequent loss of 7.3 kbp of the excisable cassette, the resulting band in the derived cisgenic line C7.1.49 would overlay the 10 kbp band observed in ‘Gala’. PCR primers designed for the flanking sequences amplified a 200 bp band in ‘Gala’, confirming the proximity of the two sequences separated by T-DNA insertion in transgenic and cisgenic lines T7.1 and C7.1.49, respectively. PCR using HcrVf2term and T7.1_c primers suggests that spontaneous recombination already took place before induction by dexamethasone treatment, as both lines amplify a fragment of the size corresponding to the expected one following removal of the excisable cassette (data not shown), while the amplicon expected prior excision could not be amplified probably due to its size of 8.4 kbp. Line C7.1.49 lacks both right and left border sequences. The line still contains the AscI restriction site used for cloning, the RS site and part of the RecLBD gene and its Tnos terminator. This leads to the presence of approximately 1 kbp of nonfunctional DNA sequences belonging to the vector. Although foreign DNA is present, C7.1.49 still falls under the definition of ‘cisgenic’ from Schouten et al. (2006). Therefore, we conclude that T7.1 has two copies of Rvi6, but the final product–the cisgenic line C7.1.49–contains only one copy of Rvi6 (Figure 2). The expression level of Rvi6 in T7.1 and C7.1.49 was sevenfold lower than in ‘Florina’. No difference in expression level was observed between T7.1 and C7.1.49, despite the extra copy of Rvi6 in T7.1. Joshi et al. (2011) also found no variation in the expression level between one or two copies of the gene, concluding that there is no correlation between the copy number and gene expression level, whereas the material they analysed consisted of independently transformed lines with integration in different genomic regions. Despite the low expression level of Rvi6, line C.7.1.49 was found to exhibit a defence reaction varying between class 2 and 3b of Chevalier's classes and the strong foliar deformation that seldom resulted in abundant sporulation (Table 3). Therefore, we conclude that cisgenic line C7.1.49 shows a level of resistance comparable with other Vf genotypes (Tartarini et al., 2000), even if it does not attain the complete resistance found in ‘Florina’.
In the transgenic mother line T7.1, it was deduced that due to the read-through, the complete vector pMF1 containing Rvi6, plus a second partial fragment of about 800 bp of the T-DNA, was integrated leading to the presence of three recombination sites in line T7.1 (Figure 2). An extra recombination site can result in the regeneration of three different plant genotypes after recombination, one for each possible pair of recombination sites. We did not detect any line that still had the RecLBD gene and codA-nptII after activation of the recombinase due to the use of the 5-fluorocytosine in selective regeneration media following dexamethasone treatment (Vanblaere et al., 2011). It was possible to detect lines corresponding to the other two possible genotypes (Figure 2, following recombination type 1 and 2). Type-2 recombination removed the second copy of Rvi6 and the whole backbone, leaving only one copy of the Rvi6 gene, one recombination site and the Tnos terminator sequence with part of the recombination gene (Figure 2). The presence of both types has been proved through PCR in lines C7.1.49 (Type 2), C7.1.3 g (Type 2), ‘C'7.1.17a (Type 1) and ‘C'7.1.63 (Type 1) (Vanblaere et al., 2011). Type-2 recombination regenerates a cisgenic plant, whereas type 1 regenerates a plant that remains transgenic, as it still contains the whole backbone vector DNA with its transgenes.
Cisgenic line C11.1.53
In cisgenic line C11.1.53, as well as in transgenic line T11.1, no amplification of vector backbone sequences was observed (Table 2), and it was therefore assumed that the inserted DNA corresponded to the expected one from the right border to the left border. A single insertion was found for line C11.1.53 as corroborated by Southern blot analysis, while results obtained for line T11.1 indicated two copies of Rvi6 (Figure 4). However, the lower band (~9 kbp) corresponds to the band in C11.1.53 and is ~7300 bp smaller than the higher band (~16 kbp) in T11.1. The difference in size between both bands corresponds to the size of the excisable cassette that is deleted during recombination. The lower band in T11.1 most likely results from spontaneous recombination, leading to a chimera composed of cisgenic and transgenic cells. Genomic sequences flanking the right and left border matched the same contig of ‘Golden Delicious’, which is located on linkage group 12. Primers developed for these sequences amplified two 343 bp amplicons in ‘Gala’ corresponding to two alleles. Spontaneous recombination could not be shown to occur by PCR amplification of the fragment encompassing the excisable cassette (primers HcrVf2term2 and T11.1_d). In fact, only line C11.1.53 amplified a fragment of about 0.5 kbp, while no amplification was observed in line T11.1. Very low expression levels were found in T11.1 and C11.1.53 compared with the natural control plant ‘Florina’. A 100- and 500-fold lower expression level of the Rvi6 gene was observed in the transgenic, respectively, cisgenic lines. Despite the low expression of the Rvi6 gene, a sufficient level of resistance, comparable with the one of the cisgenic lines C7.1.49, could be achieved (Table 3 and Figure 6). Possible explanations for the low expression levels are the position on the chromosome where the T-DNA insert is integrated, position-dependent spreading of methylation or condensed chromatin structure from surrounding genomic DNA, or a genomic scanning mechanism that identifies and methylates ‘invading’ DNA (Kooter et al., 1999). However, the aforementioned possible explanations for low expression levels do not explain why there is a significant difference in expression level following recombination between the lines T12.1 and C12.1.49 but not between T7.1 and T7.1.49 or between T11.1 and C11.1.53. Further research will be necessary to address this problem.
Table 2. Results of the PCR amplification of four amplicons scattered on the pMF1 vector's backbone. The position of the amplicons can be evinced from Figure 1
Primer pairs amplifying the vector's backbone
Table 3. Distribution of V. inaequalis inoculated apple leaves into the different scab symptom classes after Chevalier et al. (1991). The three youngest leaves per shoot have been scored by three independent researchers at day 21 after inoculation (three observations/leaf/shoot). Gala is the susceptible genotype used for transformation, while Florina is a conventionally bred scab-resistant cultivar carrying the Rvi6 gene
Number of observations
Scab Symptom classes after Chevalier et al. (1991)
Cisgenic line C12.1.49
Southern blot analysis of the line T12.1 with a probe derived from nptII (Vanblaere et al., 2011) showed a single integration, whereas Southern blot analysis with an Rvi6 specific probe revealed two bands. In T12.1, we found a band (~9500 bp) that corresponds to the band (~2500 bp) found in C12.1.49 after a normal recombination, followed by the excision of the cassette of 7300 bp. The second band present in T12.1 and C12.1.49 has the same size, ~3800 bp and is assumed to be part of a second copy of the Rvi6 gene (Figure 4). In this second integration, a truncated T-DNA was integrated only partially without the presence of the excisable cassette because Southern blot analysis with an nptII probe showed only one copy integration of the gene. We assume that two independent integrations took place on possibly different chromosomes/linkage groups based on the localization in the ‘Golden Delicious’ genome. In line T12.1, no backbone sequence amplification was observed. However, the amplification of the flanking regions led to different sequences, located by comparing with the ‘Golden Delicious’ genome, on two different chromosomes (11 and 14). No amplification of the insertion site was achieved, and it is therefore impossible to elucidate either the structure of the integrated DNA(s) or its position in the genome. Gene expression level analysis in transgenic line T12.1 had only a twofold lower expression level compared with the expression level of ‘Florina’, but in C12.1.49, there was a 100-fold lower level of gene expression compared with ‘Florina’ corresponding to a 50-fold lower level of expression than in T12.1 (Figure 5). However, this low expression was sufficient to enhance resistance to apple scab, as shown by the results of visual scoring and qPCR (Figure 6 and Table 3).
Assessment of the ability of pMF1:Rvi6 vector to generate cisgenic lines
This system, used to produce cisgenic plants in apple, was successful in generating cisgenic plant lines. Backbone integration tests indicated a very frequent (80%) read-through over the left border (Table 2). The phenomenon of ‘leaky borders’ has been reported in the past (Kohli et al., 2010; Joshi et al., 2011), but no explanation was found as to why the T-DNA transfer stopped at other sequences than the left border (LB) region. Optimization of the procedure through reduction in the backbone transfer by resolving the ‘leaky border’ problem could be achieved through the use of more effective borders (Kuraya et al., 2004; Rommens et al., 2004). The recombination system used to excise the cassette containing the nptII-codA and RecLBD recombinase was found to be not completely repressed in the absence of dexamethasone. This is supported by the observation of spontaneous recombination events in lines T7.1 and T11.1 leading to chimeras of transgenic and cisgenic genotypes, while for line T12.1, chimerity could not be tested. However, in this case, the spontaneous recombination did not hamper the generation of transgenic and cisgenic lines, due to the use of the positive and negative selection system (Vanblaere et al., 2011). Recombination also led to a reduction in the expression level of the cisgene in two of three lines (C11.1.53 and C12.1.49), whereas for the latter, no data about T-DNA integration extent were obtained. The introgression of Rvi6 led to increased resistance to apple scab, even though no resistance level similar to ‘Florina’ was observed. On one hand, the observed resistance reaction of the classes 2-3b is also found in Vf genotypes with other genetic backgrounds and may be sufficient to incite resistance under field conditions (Tartarini et al., 2000). On the other hand, however, in all cisgenic lines, the expression of Rvi6 was much lower than in the resistant control. This overall low expression level in the three trans- and cisgenic lines could be linked with the native promoter length of 242 bp used. Szankowski et al. (2009) proved that a promoter length of 779 bp has a 100-fold increased expression level compared with a promoter length of 115 bp. The latter promoter induced similar expression to the one observed in ‘Florina’, while it could not induce complete scab resistance. Also, the use of the native terminator could be a limiting factor in overall gene expression level. Szankowski et al. (2009) used a Tnos terminator, which could have a stronger influence on gene transcription termination than the native terminator. Also, the length of the native terminator might influence a more efficient gene transcription termination. Joshi et al. (2011) used the native terminator with a length of more than 2 kbp, whereas our terminator length was a mere 220 bp. However, despite the low expression, partial resistance was observed in our lines: the results of qPCR quantification of the colonization coefficient (CCcorr, Figure 6) indicated that the results obtained 21 dpi for ‘Gala’ and ‘Florina’ were similar to the ones reported by Gusberti et al. (2012). The cisgenic lines C7.1.49, C11.1.53 and C12.1.49 showed CCcorr values similar to those of ‘Milwa’, a moderately resistant cultivar (Gusberti et al., 2012). The present work represents the complete molecular characterization of two of three cisgenic lines reported by Vanblaere et al. (2011). However, such molecular characterization, as requested by the (EC) Regulation No. 1829/2003, may become unnecessary if cisgenic crops are made exempt from GMO regulations as suggested by Schouten et al. (2006).
Transgenic apple lines T7.1, T7.2, T7.3, T7.4, T8.1, T8.2, T8.3, T11.1, T11.2 and T12.1 were used (Vanblaere et al., 2011). Cisgenic apple lines were regenerated following recombination of the transgenic lines, and, for naming, the prefix T (transgenic) was replaced by C (cisgenic), and the number of each regenerated cisgenic line was added to the name (e.g. C7.1.49 is the 49th regenerated cisgenic line derived from the transgenic mother line T7.1). Cisgenic lines C7.1.49, C7.1.3, C7.1.17, C7.1.63, C11.1.53 and C12.1.49 were previously reported by Vanblaere et al. (2011).
Genomic plant DNA of trans- and cisgenic ‘Gala’ lines was extracted with the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. To detect and quantify backbone integrations, several primers were designed to amplify different parts of the backbone of vector pMF1 containing the Rvi6 gene used for transformation (Vanblaere et al., 2011). The primer sequences are listed in Table 1, and their relative position on the linearized vector is shown in Figure 1. Sequences of primers pmf_bb3 and pmf_bb4 have been published previously (Vanblaere et al., 2011). For PCR amplification, the Qiagen multiplex kit (Qiagen, Hilden, Germany) was used following the manufacturer's protocol, with each primer pair applied singularly. To evaluate the possibility of a read-through over the whole backbone (and therefore a second copy of the T-DNA), a primer pair was developed to span the plasmid sequence before and after the right border (RB).
For each of the three transgenic lines T7.1, T11.1 and T12.1, genotype inverse PCR (iPCR) was used to isolate genomic regions of ‘Gala’ flanking the RB, as well as the LB of the T-DNA insert in separate reactions, combining for each reaction the suitable restriction endonuclease and a corresponding pair of primers. The method of Triglia et al. (1988) was used with following modifications: 1 μg of DNA was digested in a total volume of 20 μL. For the self-ligation reaction, 100 ng (2 μL) of the digestion mixture was mixed with 10 μL 10X ligation buffer, 10 μL ATP (10 mm), and 4 μL T4 DNA ligase (3U) (New England Biolabs Inc., Beverly, MA) in a total volume of 100 μL and incubated overnight at 16 °C. The ligation reaction was deactivated by heating at 65 °C for 10 min. Five μL was taken from the ligation mixture and added directly to the PCR reaction. PCR reactions consisted of 1X Qiagen Multiplex PCR Master Mix, 0.5X Q-solution and 0.2 μm of each primer in a total volume of 50 μL. PCR reactions were performed applying the manufacturer's protocol conditions (95 °C for 15 min, followed by 35 cycles at 94 °C for 30 sec, 55 °C for 90 sec, 72 °C for 3 min, and a final extension at 72 °C for 10 min). The amplicons were analysed by agarose gel electrophoresis (1%), and a nested PCR was performed in a second PCR reaction. The reaction mixture and conditions were the same as for the first PCR reaction, with 5 μL of the PCR product used as the template. The amplicons were analysed by agarose gel electrophoresis (1%); the bands were cut out of the gel and sequenced (primer sequences are listed in Table 1) using the BigDye terminator kit 3.1 (Applied Biosystems, Foster City, CA). The obtained sequences, after trimming of the T-DNA sequences, were subjected to BLAST analysis against the whole genome sequence of ‘Golden Delicious’ (Velasco et al., 2010) in order to estimate the site of insertion in the Malus genome. Primers (T7.1_a, T7.1_c, T11.1_b, T11.1_d, T12.1_a, T12.1_d) were designed on the genomic DNA regions and used to amplify the corresponding region in untransformed ‘Gala’ to identify single nucleotide polymorphisms (SNPs) between the two alleles of ‘Gala’. In a second step, the primers were used for amplification on DNA of the transgenic ‘Gala’ lines T7.1, T11.1 and T12.1. As the inserted T-DNA (11 kbp) was very large for a normal PCR amplification, only amplicons from the allele in which the integration did not occur should be obtained from these lines. The developed primers were also used separately in combination with primers on the T-DNA (HcrVf2termrev2, pmf_IPCR_9; Table 1) to amplify the transition region between T-DNA and genomic DNA at the right and left borders, respectively. The amplicons were sequenced to verify the extent of the T-DNA integration. Finally, the developed primers were used to find out whether spontaneous recombination events had occurred, amplifying the DNA of the transgenic mother lines T7.1, T11.1 and T12.1. Spontaneous recombination is assumed to occur if the obtained amplicon is smaller in size than the expected T-DNA insert (11 kbp) and is shorter by the length of the excisable cassette (about 7.3 kbp shorter, Vanblaere et al., 2011). In the corresponding cisgenic line, the 11 kbp band should be absent.
Southern blot hybridization
To confirm the number of integration events obtained by hybridization of the Southern blots using the nptII probe (Vanblaere et al., 2011), a second hybridization using a PCR probe on Rvi6 was performed (primers RT1for and RT2rev (Vanblaere et al., 2011)). Transgenic and cisgenic lines were analysed by Southern blotting as described by Vanblaere et al. (2011). ‘Gala’ was added to distinguish between the Rvi6 homologues and the actual Rvi6 gene. pMF1 with Rvi6 was added as a positive control. Genomic DNA was digested by XhoI, run on an agarose gel, blotted onto a nylon membrane and hybridized using digoxygenin-labelled probes of Rvi6.
Quantitative Real-Time RT–PCR analysis
Quantitative real-time RT–PCR on cDNA was used to assess the relative level of expression of the Rvi6 gene in transgenic and cisgenic lines T7.1, C7.1.49, T11.1, C11.1.53, T12.1 and C12.1.49. As negative and positive control plants, ‘Gala’ and ‘Florina’ were used, respectively. ‘Gala’ does not contain the Rvi6 gene, while ‘Florina’ is a classically bred Vf-cultivar resistant to apple scab carrying the Rvi6 gene. The Rvi6 mRNA expression measured for ‘Florina’ was used as reference for normalizing the expression measured for the other lines. Total RNA was isolated using PureLink™ Plant RNA Reagent (Invitrogen, Paisley, Scotland) and converted into cDNA using the M-MLV reverse transcriptase, RNase H Minus Kit with oligo(dT)18 primers (Promega, Madison, WI, USA) according to the protocol described by Vanblaere et al. (2011). The cDNA was afterwards diluted 1:9. The PCR reaction mix consisted of 6 μL of cDNA, 2X SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) and 60 nm of each primer in a total volume of 20 μL. All reactions were performed according to the procedure outlined in the manufacturer's instructions (95 °C for 20 sec, followed by 40 cycles at 95 °C for 3 sec, 62 °C for 30 sec.) with the Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Ct values were obtained using specific primers for Rvi6 (formerly HcrVf2, primers HcrVf2-2273For and HcrVf2-2331Rev; Table 1) and for the reference gene ubiquitin-conjugating enzyme (MDP0000223660; primers UBCforward and UBCreverse, Pagliarani et al., 2013). All reactions were performed in biological and technical triplicates. The average of the three technical replicates was taken over each sample, while the biological replicates were analysed independently from one another. The mathematical model of Pfaffl (2001) was used to calculate the relative expression ratio. The results, including standard deviation, are presented in Figure 2 as an average of the three biological repeats.
Inoculation Tests for Apple Scab Resistance
Scab resistance was tested by inoculating grafted greenhouse plants of each cisgenic line with V. inaequalis. As susceptible controls, grafted plants of ‘Gala’ from an untransformed in vitro culture were used. As resistant controls, grafted plants of ‘Florina’ (not from in vitro culture) were used. At least three plants with one shoot per line were inoculated for each plant line used. Active growing shoots with more than three leaves were inoculated by spraying with a hand-operated mist blower until a fine water film was visible and the inoculated shoots were kept in a plastic tent to maintain the water film for 48 h in the dark at 15 °C. V. inaequalis conidia suspension was obtained from leaves with sporulating scab lesions of susceptible seedlings of the Agroscope ACW breeding programme representing a mixture of genotypes, but not containing a Vf-virulent genotype. Conidia concentration was adjusted to 1.5 × 105 using tap water. After 48 h, the plants were moved into a greenhouse cabin with day/night temperature of 20°/15°, 8-h night period, 70% rH. Symptoms were classified by three independent observers after 21 days in the greenhouse according to Chevalier et al. (1991) with class 0 no visible symptoms, class 2 yellow chlorotic flecks, whose centre could be slightly necrotic, class 3a ‘necrotic lesions and some chlorotic lesions with occasional very slight sporulation’, class 3b ‘clearly sporulating chlorotic and necrotic lesions’ and class 4 abundant sporulation and no necroses. Symptoms were evaluated on the three youngest leaves per shoot (most susceptible). Estimation of fungal colonization was assessed using the qPCR method developed by Gusberti et al. (2012). DNA was extracted from infected leaves (second youngest leaf at the time of inoculation, harvested at the time of visual screening), and the DNA was subjected to qPCR applying two different TaqMan probes targeting host DNA (Elongation factor 1) and pathogen (ATP binding cassette 2). The ratio between pathogen and host DNA (colonization coefficient, CCcorr) was calculated.
The authors wish to acknowledge the financial support by the Swiss National Science Foundation NRP59 and COST Action 864. Contributions of Michele Gusberti, Roberta Paris, Gabriela Ziltener, Ines Hiller, Jan Schaart and Henk Schouten are gratefully acknowledged.