For purposes of studying intron structures and predicting consensus splice motifs, a total of 102 legume species were used to isolate introns across the family. Of 196 gene-targeted PCR primer pairs, we successfully amplified 118 intron-containing genes (60.2%) and obtained a total of 1,870 introns with an average size of 143 nucleotides. Species-based compilation of 5’- and 3’-splicing motifs showed lineage-specific conservation in each splicing motif. Compilation of the entire intron set permitted prediction of the consensus sequences of splicing signal motifs in legumes, AYGWGTABABGH and TVNC/TAGGHTV for the 5’- and 3’-splicing motifs, respectively. Interestingly, these consensus motifs are very similar to the corresponding splicing signals of two model systems, Arabidopsis and rice. This result is suggestive of conservation of pre-mRNA splicing mechanisms in higher plants. Multiple alignments of CALTL introns demonstrated that the region from the branch point to 3’ splice site was relatively more conserved than the region from5’ splice site to the branch point. Phylogenetic analysis demonstrated that each of three splicing motifs, 5’-splice sites, 3’-splice sits, and branch site, was relevant to evolutionary divergence of species and phylogenetically informative, suggesting that splice signal sequences would be useful as a potential tool for the molecular phylogenetic analysis.

A core genetic map of the legume Medicago truncatula has been established by analyzing the segregation of 288 sequence-characterized genetic markers in an F2 population composed of 93 individuals. These molecular markers correspond to 141 ESTs, 80 BAC end sequence-tags, and 67 resistance gene analogs, covering 513 cM. In the case of EST-based markers we used an intron-targeted marker strategy, with primers designed to anneal in conserved exon regions and amplify across intron regions. Polymorphisms were significantly more frequent in intron vs exon regions, thus providing an efficient mechanism to map transcribed genes. Genetic and cytogenetic analysis produced eight well-resolved linkage groups, which have been previously correlated with eight chromosomes by means of FISH with mapped BAC clones. We anticipated that mapping of conserved coding regions would have utility for comparative mapping among legumes; thus 60 of the EST-based primer pairs were designed to amplify orthologous sequences across a range of legume species. As an initial test of this strategy, we used primers designed against M. truncatula exon sequences to rapidly map genes in Medicago sativa. The resulting comparative map, which includes 68 bridging markers, indicates that the two Medicago genomes are highly similar, and establishes the basis for a “Medicago” composite map.

Cross-species translation of genomic information may play a pivotal role in applying biological knowledge gained from one species to other genomes. Abiotic stress-responsive genes in Arabidopsis have been translated to a legume model system, Medicago truncatula. A total of 1,370 Arabidopsis genes were identified by searching TAIR database, expression profiling data and literatures. For purposes of cross-genome identification of orthologous genes, tBlastX or BlastP were employed between these two model systems. Candidate genes potentially associated with abiotic stress responses were classified into 18 functional criteria and corresponding genomic locations were analyzed by Circos program. To do this, user-friendly bioinformatic analysis platform was established. In order to discover abiotic stress-associated genes, gene network and/or interactome analyses were conducted using a combination of AraNet web-based platform and CytoScape program. As a result, we could identify 240 key genes that appeared to play an important role within central gene networks. We anticipate that these genes may impact molecular breeding programs by developing them into genetic markers and discovering trait-associated nucleotide variations.