Results

Seven distinct tomato (Lycopersicon esculentum) unigenes (version: build 3) in the SOL Genomics Network (SGN) database (http://www.sgn.cornell.edu/) encoding putative glutamate oxalaoacetate transaminases were obtained by searching for sequences annotated synonymously as aspartate aminotransferases. Nine pairs of PCR primers (Table 1) were designed in regions of high sequence confidence, using the SGN intron finder (http://www.sgn.cornell.edu/cgi-bin/tools/intron_detection/find_introns.pl) to predict intron locations and where possible design primers that amplified across one or two introns. Where available, corresponding L. pennellii ESTs were also used to guide primer design and the choice of restriction enzyme for conversion of PCR products into cleaved amplified polymorphic sequence (CAPS) markers. In the absence of corresponding L. pennellii sequences, a panel of restriction enzymes with sites in the amplified L. esculentum sequence were tested to identify restriction site polymorphisms. After amplification, seven primer pairs produced clear single bands and two primer pairs, GOT-C and GOT-H, amplified two fragments each (Table 1). The smaller fragments from these two primer pairs were polymorphic in size between M82 and LA716 and therefore, could be employed as sequence characterised amplified region (SCAR) markers. The remaining nine bands were non-polymorphic between the two parental lines, but digestion with restriction enzymes revealed cleavage polymorphisms in all of them, thus generating nine CAPS markers.

Table 1 Markers developed from SGN sequences related to glutamate oxaloacetate transaminases
Full size table

The nine CAPS markers and two SCAR markers were bin-mapped to eight different chromosomal locations (Table 1) using a set of 50 introgression lines (ILs), each containing an introgressed segment of the L. pennellii LA716 genome in a L. esculentum cv. M82 background (Eshed and Zamir 1994). GOT-F, GOT-G and GOT-H were developed from different parts of unigene sequence U213889 and therefore, amplified the same gene on chromosome 8. The GOT-C and GOT-H primer pairs each produced two bands of different size and in each case the derived markers did not map to the same regions, suggesting the existence of sequence duplications in different regions. This was further complicated following additional database analysis using U216620, from which GOT-C was developed. The SP6 end sequence of BAC clone SL_MboI0121F21 (http://sgn.cornell.edu/maps/physical/clone_read_info.pl?chrid = 209409) was found to contain an exon identical to part of U216620. An SSR marker was identified and developed from insert sequence at the T7 end of this clone (http://sgn.cornell.edu/maps/physical/clone_read_info.pl?chrid=194433), and subsequently mapped to the L. pennellii introgression carried by IL7-4. This was a different location from where the GOT-C CAPS or SCAR markers mapped, suggesting that three copies of GOT-C may exist in the tomato genome. The markers mapping to the introgressions in IL4-2/4-3, IL7-2/7-4, IL7-4 and IL8-1/8-2 corresponded well to the map locations of the classical Got-1, Got-2, Got-3 and Got-4 isozyme markers.

Discussion

In tomato, many isozymes have been exploited as biochemical markers to construct genetic maps and assist selection in breeding programs. For example, acid phosphatase (Aps-1) proved to be a powerful and reliable screening tool for nematode resistance conferred by the Mi gene (Gunther et al. 1988). Focusing on GOT, we have shown that isozyme markers can be developed into simple, reliable DNA markers through mapping sequences annotated with the same enzymatic function as the isozyme marker. The positions of several GOT markers correlate well with the positions of the known Got genes on chromosomes 4, 7 and 8. Four Got loci have been placed on three chromosomes in the classical genetic map of tomato; Got-1 on chromosome 4, Got-2 and Got-3 on chromosome 7, and Got-4 on chromosome 8 (Tanksley et al. 1992). GOT-E based on U216856 mapped to chromosome 4 in the same region as Got-1 suggesting that U216856 probably corresponds to Got-1. Similarly, GOT-B based on U215965 mapped to chromosome 7 in the same region as Got-2, suggesting that U215965 probably corresponds to Got-2. The Got-2 isozyme marker is linked tightly to I-3 (Bournival et al. 1989), a gene for resistance to Fusarium wilt, and has been used for the marker-assisted breeding of I-3. In a separate study, we have mapped GOT-B 0.15 cM north of I-3 (data not shown), strengthening the conclusion that U215965 corresponds to Got-2. The GOT-B marker, therefore, provides a useful alternative to Got-2 isozyme analysis for the marker-assisted selection of lines carrying I-3 in tomato breeding. The GOT-C SSR marker, derived from a BAC containing sequences identical to U216620, mapped to the same region of chromosome 7 as Got-3, suggesting that U216620 probably corresponds to Got-3. GOT-A based on U214922, and GOT-F, GOT-G and GOT-H-CAPS based on U213889, all mapped to chromosome 8 in the same region as Got-4, but the position of Got-4 relative to the IL has been approximated from a different mapping population, so it was not possible to determine from the available positional data whether U214922 or U213889 more likely corresponded to Got-4. However, Got-4 isozyme activity has been localised to the chloroplast (Moore and Sink 1988) and U214922 encodes a protein with a chloroplast transit peptide, whereas U213889 encodes a protein with a peroxisomal targeting signal (Table 2). This suggests that GOT-A (U214922) probably corresponds to Got-4.

Table 2 Correspondence between glutamate oxaloacetate transaminase (Got) genes in tomato and aspartate aminotransferase (Asp) genes in Arabidopsis
Full size table

Reciprocal blast searches indicate that tomato U216856 (Got-1), U215965 (Got-2) and U214922 (Got-4) are probably orthologues of Arabidopsis At2g22250, Asp1 (At2g30970) and Asp5(At4g31990), respectively, (Tables 2 and 3). At2g22250 has not hitherto been clearly identified as an Asp gene, although Liepman and Olsen (2004) note that it encodes a protein with a high degree of sequence similarity to an aspartate aminotransferase from Thermus thermophilus. Our data suggest that At2g22250 does indeed encode an aspartate aminotransferase. Similarly, tomato unigenes U216620 (Got-3) and U213889 are highly homologous to Asp2 (At5g19550), Asp3 (At5g11520) and Asp4 (At1g62800) suggesting additional orthologous relationships (Table 3). The proteins encoded by U216620 and U213889 show highest matches to the proteins encoded by Asp2 and Asp3, respectively, in reciprocal blast searches and neither matches the protein encoded by Asp4 as well as they do those encoded by Asp2 and Asp3 (Table 3). Moreover, the proteins encoded by U213889 and Asp3 have 44 and 46 amino acid extensions at their amino termini compared to those encoded by U216620 and Asp2, respectively. These amino terminal extensions are homologous to one another and contain the type 2 peroxisomal targeting signal R(IL)xxxxxHL (Reumann 2004). This suggests that U216620 and U213889 are probably orthologues of Asp2 and Asp3, respectively, (Tables 2 and 3). These correspondences to functional aspartate aminotransferase genes in Arabidopsis suggest that the tomato genome harbours at least two additional Got loci corresponding to Asp3 and Asp4. However, the tomato genome may harbour up to four additional Got loci corresponding to the GOT-C (U216620) and GOT-H (U213889) SCAR and CAPS markers. It is possible that one of the GOT-C (U216620) or GOT-H (U213889) markers corresponds to Asp4. However, it remains uncertain whether the GOT-D and GOT-I markers correspond to Got genes or genes encoding aminotransferases with different substrate specificity. Nevertheless, development of these markers has enabled functional genomic comparisons to be made between tomato and Arabidopsis with functional implications about homologous sequences in both genomes. One interesting outcome of this comparison is the observation that the tomato Got-1 and Arabidopsis At2g22250 genes appear to be unrelated to the other tomato Got or Arabidopsis Asp genes (Table 3). Rather than suggesting convergent evolution, this observation might instead indicate that the Got isozyme assay does not discriminate between differences in aminotransferase function between Got-1 and Got-4, which are both located in the chloroplast, and/or that Got-1 is a gene of endosymbiotic origin.

Table 3 Amino acid identity between proteins encoded by glutamate oxaloacetate transaminase (Got) genes in tomato and aspartate aminotransferase (Asp) genes in Arabidopsis
Full size table

The development of PCR-based markers from multigene families, exemplified here by the development of 12 new PCR-based markers from the tomato GOT gene family, has the potential to complement PCR-based markers based on unique gene sequences, because many genes are members of multigene families e.g., 65% of the genes in Arabidopsis are members of multigene families (Arabidopsis Genome Initiative 2000), and unique gene sequences may therefore, be distributed sparsely in some regions of the plant genome. The conversion of old isozyme markers into new PCR-based markers will also enable better integration of classical genetic mapping data based on isozyme and morphological markers into recent genetic maps based on molecular markers. Moreover, markers developed from multigene families have the potential to assist genome-sequencing projects by anchoring BAC ends to individual chromosomes, which might otherwise be problematic for BAC ends containing members of multigene families e.g., the Got-3 SSR marker anchors BAC clone SL_MboI0121F21 to tomato chromosome 7, which may be useful for the tomato chromosome 7 sequencing project.

The Got isozymes have been useful genetic markers in tomato because of the variation in electrophoretic mobility they display between tomato and its wild relatives. Although not necessarily amplifying the coding subregions responsible for these polymorphisms, the intron-spanning SCAR and CAPS markers that we have developed are also likely to detect variation between related tomato species and may be useful in introgression breeding in tomato e.g., in the introduction of new alleles of the I-3 gene for Fusarium wilt resistance that have novel resistance specificities. Given that SSR markers are often used to reveal polymorphisms between tomato varieties, the GOT-C SSR marker may also be useful for intra-specific breeding in addition to introgression breeding. As with any marker, the GOT markers reported here will need to be adapted to the specific objectives of the tomato breeder. The GOT markers serve to anchor their associated unigene sequences to specific subregions of the tomato genome and these in turn can be used to underpin the development of new markers tailor-made for marker-assisted breeding of traits of interest in these subregions.