Tomato yellow leaf curl disease, a devastating disease of Solanum lycopersicum (tomato), is caused by a complex of begomoviruses generally referred to as Tomato yellow leaf curl virus (TYLCV). Almost all breeding for TYLCV resistance has been based on the introgression of the Ty-1 resistance locus derived from Solanum chilense LA1969. Knowledge about the exact location of Ty-1 on tomato chromosome 6 will help in understanding the genomic organization of the Ty-1 locus. In this study, we analyze the chromosomal rearrangement and recombination behavior of the chromosomal region where Ty-1 is introgressed. Nineteen markers on tomato chromosome 6 were used in F2 populations obtained from two commercial hybrids, and showed the presence of a large introgression in both. Fluorescence in situ hybridization (FISH) analysis revealed two chromosomal rearrangements between S. lycopersicum and S. chilense LA1969 in the Ty-1 introgression. Furthermore, a large-scale recombinant screening in the two F2 populations was performed, and 30 recombinants in the Ty-1 introgression were identified. All recombination events were located on the long arm beyond the inversions, showing that recombination in the inverted region was absent. Disease tests on progenies of informative recombinants with TYLCV mapped Ty-1 to the long arm between markers MSc05732-4 and MSc05732-14, an interval overlapping with the reported Ty-3 region, which led to the indication that Ty-1 and Ty-3 may be allelic. With this study we prove that FISH can be used as a diagnostic tool to aid in the accurate mapping of genes that were introgressed from wild species into cultivated tomato.
Tomato yellow leaf curl disease (TYLCD) has been one of the most devastating diseases in Solanum lycopersicum (tomato) in the last few decades (Moriones and Navas-Castillo, 2000). TYLCD is a viral disease caused by tomato yellow leaf curl viruses (TYLCVs), which all belong to the Begomovirus genus within the Geminiviridae (Fauquet, 2005). Whereas most begomoviruses (geminiviruses) contain a bipartite circular, single-stranded (ss) DNA genome, TYLCV only contains one circular ssDNA of about 2.7 kb. Its genome has six partially overlapping open reading frames (ORFs) that are bidirectionally organized and separated by an intergenic region (IR) of approximately 200 nucleotides (Gronenborn, 2007). Geminiviruses easily recombine during mixed infections, which not only leads to new variants and diversification within the TYLCV cluster, but also makes its taxonomic classification more and more complex (Monci et al., 2002; García-Andrés et al., 2007, 2009). Meanwhile, at least 11 TYLCV species have been reported, and a standardized system has been set up to assist in creating a transparent and useful nomenclature for Geminiviridae, including newly identified species and strains (Fauquet et al., 2008).
Tomato yello leaf curl virus (TYLCV) is widespread in warm and (sub) tropical regions worldwide, and is a limiting factor for tomato production (Cohen and Lapidot, 2007). The disease is still spreading, with recent outbreaks reported in California and Hawaii, USA, and in China (Rojas et al., 2007; Zhang et al., 2009; Melzer et al., 2010). TYLCV infections lead to stunting, yellowing, leaf curling and flower abortion. When plants are infected at a young stage crop losses of up to 100% may occur (Varma and Malathi, 2003). The virus has a large host range, including many economically important crops like tomato, Nicotiana tabacum (tobacco), Capsicum annuum (pepper) and Solanum tuberosum (potato)(Polston and Anderson, 1997), and is transmitted by the sweetpotato whitefly Bemisia tabaci.
Disease management of TYLCV, aimed at controlling the whitefly insect vector, is expensive and labor intensive, and management includes insecticide applications sometimes combined with physical barriers like polyethylene sheets or large plants like Sorghum bicolor (sorghum; Hilje et al., 2001; Palumbo et al., 2001). More recently, whitefly insecticide resistance has been reported (Horowitz et al., 2007), stressing the need for alternative management strategies, such as breeding TYLCV-resistant tomato cultivars. Whereas domesticated tomatoes (S. lycopersicum) are susceptible to TYLCV, high levels of resistance were found in several wild tomato species, including Solanum pimpinellifolium, Solanum peruvianum, Solanum chilense, Solanum habrochaites and Solanum cheesmaniae (Ji et al., 2007b). Some of these have been used for intensive genetic studies, which have so far led to the mapping of five TYLCV resistance genes (Table 1).
Table 1. Mapped TYLCV resistance loci identified from wild Solanum species
aThe source of the Ty-5 gene was the tomato breeding line TY172, which is derived from four different accessions of Solanum peruvianum.
Currently, five loci (Ty-1–Ty-5) for TYLCV resistance are available for commercial breeding (Ji et al., 2007b; Anbinder et al., 2009). The Ty-1 locus from S. chilense LA1969 was the first mapped TYLCV resistance locus (Zamir et al., 1994), and was shown to be linked with the Ty-3 locus on chromosome 6 (Ji et al., 2007a). In many (commercial) breeding programs worldwide, Ty-1 has been introgressed into cultivated tomatoes, and these cultivars are for sale on the market (Ji et al., 2007b). However, the Ty-1 introgression in these cultivars is generally accompanied by undesired horticultural traits (such as autonecrosis, http://www.faqs.org/patents/app/20100212048), a phenomenon that is known as linkage drag. Ty-1 was first mapped to the pericentromere of tomato chromosome 6 (Zamir et al., 1994), but follow-up studies with newly developed molecular markers tightly linked to Ty-1 presented contradictory results on the genetic position of the Ty-1 locus. In one study, Ty-1 was linked to the REX-1 locus in the Mi-1 gene cluster, suggesting that Ty-1 is located on the short arm of chromosome 6 (Milo, 2001). In another study, Pérez de Castro et al. (2007) reported linkage of Ty-1 to marker CT21, which is located below the centromere on the long arm. So far, the exact position of Ty-1 has not been determined and nor has the underlying gene been elucidated (Pérez de Castro et al., 2007).
Mapping of genes in the pericentromere is very inaccurate because of the suppression of recombination. One example is the mapping of Mi-1, a tomato gene conferring resistance to three different pathogens, which is located in the pericentromere in the short arm of tomato chromosome 6. Suppression of recombination was reported in the F2 populations derived from interspecific crosses between S. lycopersicum and S. peruvianum, which were used for cloning the Mi-1 gene (Kaloshian et al., 1998). Considering the report of linkage between Ty-1 and the Mi-1 gene (Milo, 2001), the failure of efforts to fine-map Ty-1 and to reduce the introgression size is probably caused by the suppression of recombination in this region. Although causes for the recombination suppression are not known, the location of a target gene in pericentromere heterochromatic regions and/or chromosomal rearrangement(s) between cultivated and wild tomatoes may play a role (Tang et al., 2008; Szinay et al., 2010).
Recently, fluorescence in situ hybridization (FISH) has been shown to facilitate genetic mapping by the visualization of physical locations of bacterial artificial chromosomes (BACs) on pachytene chromosomes (Szinay, 2010). Moreover, cross-species FISH has been successfully applied to detect chromosomal rearrangements between Solanum species (Iovene et al., 2008; Lou et al., 2010.; Szinay et al., 2010; Tang et al., 2008). In this study, we applied BAC-FISH and large-scale recombinant screening to analyze the chromosomal structure and recombination behavior in the chromosomal region where Ty-1 is introgressed. Markers that cover the large S. chilense Ty-1 introgressions in tomato hybrids have been developed and used for large-scale recombinant screenings on F2 populations derived from two Ty-1 hybrids. Our results demonstrate that Ty-1 is located on the long arm of tomato chromosome 6 near the Ty-3 locus. The failure of previous efforts to precisely locate Ty-1 was caused by low marker coverage in combination with severe recombination suppression in the previously reported Ty-1 region, which is the result of chromosomal inversions between S. lycopersicum and S. chilense LA1969.
Large introgression fragments are present in cultivars carrying Ty-1
Two F2 populations (P1 and P2), each with an introgression from S. chilense LA1969 and segregating for TYLCV resistance, were used to more precisely map the Ty-1 locus. The presence of the Ty-1 introgression was verified in P1 by challenging a small set of F2 plants (n = 45), along with the susceptible control, Moneymaker (MM), with viruliferous whiteflies. Three weeks after infestation, TYLCV symptoms, i.e. yellowing and curling of the leaves, were clearly visible on MM plants (Figure 1). From P1, 15 F2 plants showed symptoms and were scored as susceptible (S), whereas 30 plants remained symptomless and were scored as resistant (R) (Figure 1). Five markers linked to Ty-1, i.e. REX-1, Aps-1, TG97, TG231 and JB-1 (Table S1; Pérez de Castro et al., 2007), were applied to this set of plants. All markers showed a homozygous or heterozygous S. chilense genotype for the resistant plants and an S. lycopersicum genotype for the susceptible plants. The same markers were also applied in P2, and all showed similar polymorphisms among F2 plants, indicating the presence of a Ty-1-carrying S. chilense introgression in each population.
In order to determine the size of the introgression in this material, molecular markers were designed from the tomato genome sequence information. Eighteen BACs physically mapped to the Ty-1 region on chromosome 6 (Figure 2; Peters et al., 2009; Tang et al., 2008) were targeted for marker development, and cleaved amplified polymorphic sequences (CAPS) markers were successfully developed from nine of these (Table S1). Tomato genome scaffold sequences were used to design additional markers corresponding to gaps between the BAC contigs (Figure 2; Table S1).
Based on polymorphisms between two DNA pools that were made containing DNA of either 10 R or 10 S F2 plants, the introgression in P1 spans the region between BAC H304P16 and BAC M005H10. The introgression thus covers a part of the short arm, the centromere and a part of the long arm of chromosome 6 (Figure 2). Markers derived from these two BACs were applied in P2, and the same marker polymorphisms were observed, suggesting a similar-sized introgression in both populations. As both F2 populations are derived from commercial F1 hybrids, our results clearly demonstrate that a large chromosomal fragment from S. chilense carrying Ty-1 was introgressed into cultivated tomatoes. This region spans, according to the latest tomato sequence release (WGS2.31), approximately 30 million base pairs.
Suppression of recombination in the S. chilense introgression region
To more precisely map Ty-1, an initial screening was performed by genotyping approximately 3000 plants from P1 with markers M-H304P16 and M-M005H10, which flank the chromosomal region where Ty-1 and Ty-3 are located, according to previous studies (Zamir et al., 1994; Ji et al., 2007a; Pérez de Castro et al., 2007). A total of 26 recombinants between these two markers were identified and further genotyped with additional markers in the region (Table 2). Results revealed that all recombination events occurred downstream of marker MSc09883-6 (Figure 2). Surprisingly, no recombinants were found between markers M-H304P16 and MSc09883-6, an interval corresponding to more than 60% of the physical distance between markers M-H304P16 and M-M005H10, indicating the severe suppression of recombination in this region.
Table 2. Genotypes of CAPS markers from recombinants between M304P16-2 and M005H10 identified in an F2 population 1, and used in disease tests
To test whether this suppression was population specific, two other populations (P2 and P3) were used for recombinant screening. In P2, where Ty-1 is present, only four recombination events between M-H304P16 and M-M005H10 were identified by screening 1600 F2 plants. Again, no recombination events occurred between markers M-H304P16 and MSc09883-6, confirming the suppression of recombination within this region in populations derived from S. chilense LA1969.
In the Ty-3 population (P3), approximately 6000 plants were screened with markers Mi23 and M-M005H10, which yielded 150 recombinants. Fifty-four of these recombination events were identified between markers Mi23 and TG97, but only six occurred within the nearly 17-Mb interval between M-H304P16 and C2_At5g61510. No recombination events were observed within the nearly 6-Mb region between markers C2_At5g61510 and TG97. Clearly recombination is also severely suppressed in this region in this Ty-3-carrying population.
Chromosomal rearrangements in the S. chilense introgression
In order to check whether the suppression of recombination is caused by chromosomal rearrangements between homologous chromosomes in the Ty-1 introgression, we performed FISH experiments using five selected BACs (Figure 3) that had been used in previous BAC-FISH experiments (Szinay et al., 2008; Tang et al., 2008). Cot100 repeat blocking was applied in order to guarantee locus-specific signals. Nine F2 plants of P1 were used, of which three were homozygous for S. chilense alleles in the Ty-1 introgression (‘b’ plants), three were homozygous for S. lycopersicum alleles (‘a’ plants) and three were heterozygous (‘ht’ plants). On the pachytene chromosome of the ‘a’ plants (Figure 3a,d), the BACs hybridized to the expected locations as in cv. MM and cv. Heinz 1706: two on the short arm and three on the long arm (Figure 3). On the ‘b’ plants (Figure 3b,e–h), we could not obtain an image where all five BACs gave signals on the same chromosome. We therefore compared BAC locations by combining different representative images obtained from the same plant (Figure 3e–h). As each BAC is present on at least two images, relative positions of all the five BACs are determined. For BAC H176D13 it was difficult to obtain a clear signal in ‘b’ plants, possibly because of the highly condensed heterochromatin of ‘b’ plants. BAC H242H19, which localized on the short arm above the centromere in ‘a’ plants, showed a signal on the long arm below the centromere in ‘b’ plants. BAC H309K01 and H003K02 were located on the long arm with an inverted order between ‘a’ and ‘b’ plants. BAC H308F14, most distal in the long arm pericentromere heterochromatin, was syntenic between ‘a’ and ‘b’ plants. On the ‘ht’ plants (Figure 3c,i), BAC H176D13 and H308F14 each gave a locus-specific signal. BAC H242H19 gave two signals: one on the short arm and one on the long arm. For the other two BACs, the positions of which were inverted between ‘a’ and ‘b’ plants, multiple signals appeared on the ‘ht’ plants, indicating that paring between homologous chromosomes was interrupted. The results altogether suggested the occurrence of two chromosomal inversions between S. chilense LA1969 (the donor species of the Ty-1 locus) and S. lycopersicum: i.e. one involving the centromere, as shown by BAC H242H19, and the other one on the long arm pericentromere heterochromatin, as shown by BACs H309K01 and H003K02 (Figure 3). Both inversions localized to the chromosomal region where suppression of recombination was observed (Figure 2; Table 2). The failure in hybridizing BAC H176D13 on ‘b’ plants is probably the result of inefficient Cot100 blocking because of the high degree of (mostly gypsy-type) retrotransposable elements in the heterochromatin of the short arm (Figure 3e–h), or the result of undiscovered complex chromosome rearrangements in this region.
To verify the FISH results obtained in the F2 plants, we attempted to paint the BACs on S. chilense LA1969 (donor of Ty-1) and on S. chilense LA2779 (donor of Ty-3). These experiments failed because in almost 2 years we could not obtain young flower buds in the right stage for FISH. Therefore, we painted the same BACs onto TY52 and Su09E941-164-1, which carry Ty-1 and Ty-3 introgressions, respectively. Based on our marker data (Table 3), TY52 harbors a large introgression that spans the inverted regions, and FISH data confirmed in this line that BAC H242H19 was on the long arm and that BACs H309K01 and H003K02 were inverted compared with ‘a’ plants. Su09E941-164-1 has a small introgression, likely below the inversions (Table 3), and the BAC order was collinear between MM and Su09E941-164-1 in the FISH experiments.
Table 3. Marker genotypes of a fixed Ty-1 and a fixed Ty-3 line
Ty-1 maps near the Ty-3 locus
To further localize the Ty-1 locus, disease assays were performed on selfed progenies of the recombinants of P1. In total, 18 informative F3 families were challenged by agroinoculation with an infectious TYLCV clone. Approximately 4 weeks after agroinoculation, MM plants showed clear TYLCV symptoms. One F3 family homozygous for the S. chilense Ty-1 introgression was used as a resistant control. All plants of the resistant control were symptomless, and plants of recombinant F3 families were unambiguously scored as either resistant or susceptible. Although autonecrosis is known for some Ty-1 cultivars, the resistant plants in our test showed an MM-like morphological phenotype. Previous reports locate Ty-1 to the pericentromere, above the Ty-3 locus, towards the centromere. According to this position, the markers M-H304P16 and MSc09983-6 (Figure 4; Table 2) should flank the resistance gene, and F3 families of the recombinants Z-G9, M-A7, U-F6, S-F7, K-D1 and R-C2 should breed true for resistance, whereas families of W-G5, J-D10, L-D5, R-G10 and Z-D8 should all be susceptible. However, only Z-G9 bred true for resistance, and all other families segregated for resistance, showing that Ty-1 is actually distal of marker MSc05732-4 and located in the region where Ty-3 is mapped (Figure 4). This result was supported by the analysis of the V-A9 and K-A5 families that, according to the reported Ty-1 position, should have segregated for resistance, but in fact bred true for susceptibility. Moreover, segregation in the U-A2 and M-G6 families places Ty-1 above marker MSc05732-14 (Table 2). Results from all other tested recombinant families support the location of Ty-1 between marker MSc05732-4 and MSc05732-14 (Table 2), a marker interval that partly overlaps with the mapped marker interval of Ty-3.
Although five TYLCV resistance loci are available, introgressions of Ty-1 from S. chilense LA1969 have so far been a major focus in breeding programs (Zamir et al., 1994; Ji et al., 2007a; Pérez de Castro et al., 2007; Vidavski, 2007). A lack of knowledge about the exact location of Ty-1 on tomato chromosome 6 has hindered efforts to reduce the size of the Ty-1 introgression. Consequently, Ty-1 introgressions are often accompanied by linked detrimental traits. In this study, FISH and large-scale recombinant screening was applied to analyze the chromosomal structure and recombination behavior in the chromosomal region where Ty-1 is introgressed. Our results show that the precise mapping of Ty-1 has been hampered by chromosomal inversions between S. lycopersicum and S. chilense LA1969, and that Ty-1 is actually located on the long arm of tomato chromosome 6, near the chromosomal region where Ty-3 is mapped. Moreover, our study demonstrates that FISH is not only a very helpful tool to reveal the occurrence of chromosomal rearrangements but also a diagnostic technique if one wants to correctly map introgressed genes from wild species.
Ty-1 and Ty-3 are likely to be located in an overlapping chromosomal region
The results presented here map Ty-1 between markers MSc05732-4 and MSc05732-14, a region of approximately 600 kb. This interval was surprising, as it is nearly 5 Mb below the reported map position. Zamir et al. (1994) originally mapped Ty-1 to an approximately 40-cM introgression [spanning the restriction fragment length polymorphism (RFLP) markers TG297, TG97 and TG119, as well as the Mi locus for S. peruvianum-derived nematode resistance]. These authors also identified a resistant recombinant inbred line (RIL) with an LA1969 introgression spanning the upper portion of this introgression to TG97, but lacking the lower region represented by TG119. They assumed that the tolerance of some S. peruvianum accessions to TYLCV might be based on Ty-1, so different nematode-resistant lines containing Mi introgressions from TG297 to TG97 were examined for their response to the virus, but all were susceptible. Based on these findings, it was concluded that Ty-1 was linked to TG97 and located below the marker. This conclusion, however, did not consider that TYLCV resistance in S. peruvianum accessions might be conferred by a locus other than Ty-1; this is confirmed by the recent mapping of Ty-5 from S. peruvianum to chromosome 4 (Ji et al., 2007b). Furthermore, mapping Ty-1 near TG97 did not consider the possibility that the introgression in the RIL might extend well below TG97. Better marker coverage would have shown this, as we demonstrate in the present study by revealing that TY52 – developed from the original Ty-1 mapping work – has an introgression extending into the 600-kb genic region. It appears that subsequent approaches to more precisely map Ty-1 were based on the originally reported position. These attempts were unsuccessful in correctly locating the gene because they targeted recombination in a nearby but non-genic region.
The resistance gene Ty-3, derived from either S. chilense LA2779 or LA1932 accessions, was also recently mapped to the long arm of chromosome 6, between markers cLEG-31-P16 and T1079 (Ji et al., 2007a). However, as with the original Ty-1 work, defining the location of Ty-3 within this interval did not take into account how far the introgression might extend above cLEG-31-P16 or below T1079, into regions where there was previously no marker coverage. In fact, preliminary data suggests that Ty-3 is located between T0774 and cLEG-31-P16 (Hutton et al., 2010). The introgression present in Su09E941-164-1 is shown to extend approximately 300 kb above cLEG-31-P16, beginning below MSc05732-4 (Table 3). Together, these data indicate that Ty-1 and Ty-3 are likely to be located in an overlapping chromosomal region and may be allelic (Figure 4). Efforts are underway to fine-map both genes to determine if this is indeed the case.
BAC-FISH can be used as a diagnostic tool in introgression breeding
Our initial aim was to fine-map the Ty-1 locus on tomato chromosome 6, but screening a large number of plants for recombinants revealed a strong suppression of recombination in the S. chilense LA1969 introgressed region. This was shown to be the case in two independent commercial Ty-1 hybrids. With the use of FISH, we uncovered that this suppression was caused by two inversions between S. chilense LA1969 and S. lycopersicum in the pericentromere heterochromatic regions of both the short and long arms of chromosome 6 (Figure 3). Because of these inversions, chromosomal pairing during meiosis has been interrupted. The presence of these inversions has caused recombination suppression, which contributed to the failure to accurately map the Ty-1 locus. In short, because recombination was suppressed in this region, the association of resistance with markers on the short arm (such as REX-1) (Milo, 2001) was no less accurate than the mapping of Ty-1 near markers on the long arm (such as CT21) (Pérez de Castro et al., 2007). Furthermore, the large introgression present in Ty-1 hybrids shows that inverted chromosomal regions are introgressed as a whole. It is not yet clear whether the same inversion is present in other S. chilense accessions, although the suppression of recombination in this region in P3 indicates a likely inversion in S. chilense LA2779. Chromosomal rearrangements between two related species have previously been described: e.g. in the Mi-1 cluster between S. lycopersicum and S. peruvianum; in the sun locus between S. lycopersicum and S. pimpinellifolium; and in the short arm of chromosome 6 between S. lycopersicum and S. tuberosum (Seah et al., 2004; Van der Knaap et al., 2004; Tang et al., 2008). Recently, Szinay et al. (2010) described a comprehensive study on a large series of chromosomal rearrangements among Solanum species, which altogether suggests that rearrangements as described here are not unique.
For many traits the gene pool of S. lycopersicum is quite narrow, which has forced breeders to use related wild species in breeding programs. This so-called ‘introgression breeding’ allows access to the variation present in numerous Solanum accessions. Our results suggest that this type of breeding can be hampered by chromosomal rearrangements between related species, and that these rearrangements can thus influence the success of introgression breeding when interspecific crosses are used. Here in, we present the concept on how to apply BAC-FISH as a diagnostic tool to investigate chromosomal rearrangements in genetic mapping and introgression breeding, which will have a potential impact on introgression breeding in many crops. Furthermore, the occurrence of chromosomal rearrangements stresses the importance of a physical map to order scaffolds if related Solanum species are being sequenced (current efforts of large sequencing consortia).
Impact of the presented results on plant breeding
Disease resistance genes in the tomato genome tend to be clustered as opposed to randomly distributed (Yang and Francis, 2007), and repulsion linkages can be problematic for tomato breeders who want to combine resistance genes that are closely linked. Doing this becomes extremely difficult when introgressions from wild species overlap and there is recombination suppression, as has been shown in this study. The general genomic region where Ty-1 and Ty-3 reside has the greatest number of reported disease resistance genes in the entire genome. Besides the mentioned resistance to root knot nematode are genes for resistance to Ralstonia solanacearum (bacterial wilt), Clavibacter michaginensis (bacterial canker), Oidium lycopersicum (powdery mildew) and Cladosporium fulvum (leaf mold). It will be useful to fine map these resistance genes and develop lines with minimal introgression sizes to facilitate combining the genes in cis, and to eliminate any associated linkage drag. Once combined, the close linkages will be beneficial for keeping the genes together.
The present work shows that Ty-1 and Ty-3 are very close and are perhaps allelic. This finding has important breeding implications. It reduces the likelihood that the genes can be pyramided homozygously, but instead points to the making of hybrids heterozygous for both genes/alleles by combining parents homozygous for Ty-1 with parents homozygous for Ty-3. The Ty-1 locus has shown dominance for TYLCV resistance, but has been ineffective against some TYLCV strains (Scott, 2007) and against bipartite begomoviruses (Mejia et al., 2005). The Ty-3 locus has generally shown less dominance, but a wider range of resistance against TYLCV strains and bipartite begomoviruses. Hybrids with the heterozygous combination of both genes/alleles may prove to be effective and durable against a wide array of such viruses. Vidavski (2007)showed that combining different begomovirus resistance genes can have unanticipated synergistic effects, and this combination should be tested in this regard.
Two commercial hybrids derived from differing parental lines were used to produce two F2 populations (P1 and P2) in which Ty-1 was segregating. For both F2 populations multiple F1 plants were selfed to get enough F2 seeds. These two hybrids were provided by breeding companies within the cooperative framework of the Centre for BioSystems Genomics (CBSG). The Ty-1 locus was introgressed from S. chilense LA1969 in the genetic background of cultivated tomato S. lycopersicium. These two F2 populations were used for recombinant screenings, and selected recombinants were selfed to produce F3 families for further testing with TYLCV. As a susceptible control, S. lycopersicum cv. Moneymaker was included.
Another population (P3) was developed in a Ty-3 fine-mapping effort at the University of Florida, Gainesville, FL, USA. F2 plants segregating for a large Ty-3 introgression from S. chilense accession LA2779 (spanning approximately 20 cM from the Mi locus to M-M005H10) were screened for recombination events within this interval. Su09E941-164-1 is an RIL developed from this population; resistance to TYLCV in this line has been confirmed over multiple seasons, and is conferred by an approximately 4-cM Ty-3 introgression that does not overlap the Ty-1 region reported previously by Zamir et al. (1994) (S.F. Hutton, unpublished data). TY52 is an introgression line that contains Ty-1 introgression from S. chilense LA1969 (Michelson et al., 1994).
All markers are PCR-based CAPS markers taken from publicly available data (Table S1). Tomato BACs in the region of Ty-1 and Ty-3 were selected according to a physical map of tomato chromosome 6 (Tang et al., 2008; Peters et al., 2009) (Figure 2), and sequences were downloaded from the SOL website (http://solgenomics.net). For each BAC, multiple random primer pairs, which would result in products of approximately 800 bp, were designed using PrimerBLAST from NCBI (http://www.ncbi.nlm.nih.gov/tools/primer-blast). As the parental lines of the F2 populations could not be obtained because of confidentiality, resistant and susceptible F2 plants were selected. Two DNA pools were made containing DNA of either 10 resistant (R) or 10 susceptible (S) F2 plants. These pools (R-pool and S-pool) were used for marker development. In case PCR products obtained from both pools showed no length polymorphism, several restriction enzymes were used to search for polymorphisms. BAC-derived markers are named after the BAC name, e.g. M-H304P16 means marker (M) generated from the sequence of BAC H304P16 (Table S1). Scaffolds that covered the gaps between the BAC contigs were also selected and used for marker development using the same strategy as described.
For P1 and P2, leaf samples were collected from 2-week-old seedlings and DNA was extracted according to the NaOH extraction method, as described by Wang et al. (1993), with some slight modifications. In brief, small leaf samples were ground for 5 min in 20 μl of 0.5 m NaOH using a tissue striker (Kisan Biotech™, http://www.kisanbiotech.com). After the addition of 20 μl of 100 mm Tris (pH 7.5), 5 μl was added to 200 μl of 100 mm Tris (pH 7.5) to prepare (crude) genomic DNA template for PCR. For P3, tissue was collected according to the methods described above and frozen at −80°C. Frozen leaf samples were then ground for 2 min using a high-throughput homogenizer (Talboys; Troemner, http://www.troemner.com), and DNA was extracted according to the method described by Fulton et al. (1995). PCR amplification was performed according to standard protocols in either an Applied Biosystems GeneAmp 2700 system (Applied Biosystems, http://www.appliedbiosystems.com) or an Eppendorf Mastercycler® pro (Eppendorf, http://www.eppendorf.com).
TYLCV inoculation and disease evaluation
For P1, 2-week-old seedlings were put in a cage in the presence of whiteflies carrying TYLCV-Alm. After 4 days, whiteflies were killed by adding imidocloprid (Admire; Bayer CropScience, http://www.bayercropscience.com) to the soil. After another 2 weeks, plants were scored as R (no symptoms) or S (showing curling and yellowing of the leaves).
To test plants of recombinant families selected from P1, agroinoculation was used. An infectious TYLCV clone (pTYCz40a, kindly provided by Dr. Eduardo Rodríguez Bejarano, Universidad de Malaga) was transformed to Agrobacterium tumefaciens strain LBA4404 and used to agroinoculate tomato seedlings. To this end, A. tumefaciens containing the TYLCV clone was grown overnight at 28°C in 3 ml LB medium (10 g l−1 trypton, 5 g l−1 yeast, 4 g l−1 NaCl, 1 g l−1 KCl, 3 g l−1 MgSO4.2H2O). From the overnight culture 600 μl was transferred to 3 ml of induction medium [10.5 g l−1 K2HPO44, 4.5 g l−1 KH2PO4, 1 g l−1 (NH4)2SO4, 0.5 g l−1 sodium citrate.2H2O, 1 mm MgSO4.7H2O, 0.2% (w/v) glucose, 0.5% (v/v) glycerol; after autoclaving, 50 μm acetosyringone and 10 mm 2-(N-morpholine)-ethanesulphonic acid (MES), pH 5.6, was added] and grown overnight at 28°C. Bacteria were pelleted by centrifugation for 10 min at 2000 g and resuspended in MS medium (supplemented with 150 μm acetosyringone and 10 mm MES) at an OD600 of 0.5. The first pair of true leaves of 3-week-old seedlings were agroinoculated with the TYLCV construct by pressure inoculation with a syringe.
Plants were kept under glasshouse conditions at a temperature of 23°C and relative humidity of 60% during a 16-h day/8-h night regime. Four weeks after agroinfiltration, plants were evaluated for TYLCV symptoms by using the disease severity index as described by (Lapidot and Friedmann, 2002).
Fluorescence in situ hybridization
Slide preparation: Young flower buds were collected in the glasshouse and fixed overnight in fresh Carnoy solution (1:3, acetic acid:ethanol), then transferred to 70% ethanol for storage at 4°C. Slides were prepared according to the method described by Szinay et al. (2008), without post-fixation with 1% formaldehyde.
BAC and COT 100 isolation and labelling: BAC DNA was isolated using the High Pure Plasmid Isolation Kit (cat. no. 11754785001; Roche, http://www.roche.com). Isolated DNA was labeled by nick translation according to the manufacturer’s protocol. Probes were labeled either directly with Cy3-dUTP (Amersham, http://www.amershambiosciences.com), Cy3.5-dCTP (Amersham) and diethylaminocoumarin-5-dUTP (DEAC) (Perkin Elmer, http://www.perkinelmer.com) or indirectly with biotin or digoxigenin. Cot 100 isolation, necessary to block repeats present in most of the BACs, was performed according to the method described by Szinay et al. (2008).
Procedure, microscopy and data analysis: FISH experiments were performed according to Szinay et al. (2008), with some slight modifications. Probes were hybridized for 48 h at 37°C (Rens et al., 2006). For stringent washing, 64% formamide was used three times for 5 min at 42°C (Schwarzacher and Heslop-Harrison, 2000). Biotin-labeled probes were amplified using streptavidin Cy5 and biotinylated anti-streptavidin. Digoxigenin-labeled probes were amplified using anti-dig fluorescein isothiocyanate (FITC) and anti-sheep FITC. Microscopy and data analysis were performed according to Szinay et al. (2008).
Much of this project was carried out within the research programme of the Centre of BioSystems Genomics (CBSG), which is part of the Netherlands Genomics Initiative/Netherlands Organization for Scientific Research. Another part of this research was supported by a USDA CREES-NRI (now AFRI) grant. The infectious TYLCV clone was kindly provided by Professor Eduardo Rodríguez Bejarano (Universidad de Málaga). We thank Dr. I. Dick Peters for his help with the whitefly transmission and Dick Lohuis for his help with agroinfiltrations. We thank Tingting Xiao, Xiaoxin Liu and Ada Lucia Angulo Fernandez for their help with marker development and recombinant screenings. Special thanks are also given to Dolly Cummings and Jose Diaz for help with DNA extractions and recombinant screenings.