IDENTIFICATION OF CITRUS SINENSIS BAC CLONES CONTAINING GENES RELEVANT TO FRUIT QUALITY

HarvEST:Citrus (http://harvest.ucr.edu/), an EST database-viewing software developed at the University of California - Riverside, displays 141 libraries and 469,618 ESTs from Citrus and Poncirus. We used the C38 EST assembly (36.980 citrus unigenes) and OligoSpawn software (http://138.23.178.42) to design 31.230 overgo probes (36-bp gene-specific hybridization probes that have been prescreened to mask out all known repeat elements) employed to screen a Citrus sinensis 'Vaniglia' BAC library (19X). For BAC library screening we selected 81 overgo probes associated with unigenes that putatively code for enzymes relevant to fruit quality (flavonol, anthocyanin, carotenoid, chlorophyll, cellulose, starch, ascorbic acid, aromatic amino acid and lignin biosynthesis; sucrose catabolism; glycolysis; oxidative/nonoxidative pentose phosphate pathway; fatty acid biosynthesis and oxidation; Krebs cycle). To verify that the overgos selected were not designed to genomic regions that contained intron/exon junctions we used the software "HarvEST BLAST Search" and the database "JGI Citrus" (http://138.23.178.42/blast/blast.html), and the software "Intron Finder" (http://solgenomics.net/tools/intron_detection/find_introns.pl). A two-dimensional 9 X 9 overgo pooling strategy was used. Hybridization probes were pooled and hybridized in groups of intersecting rows and columns to high-density BAC filters, followed by a deconvolution process that established BAC-probe addresses. Matrix deconvolution was carried out using graphical user-interface derived from HybSweeper (Lazo et al., 2005) and a Perl deconvolution script developed at Clemson University Genomics Institute (USA). BAC addresses were obtained for 75 of the 81 overgo probes initially selected, for a total of 1018 BAC clones, a number consistent with the depth of coverage of the BAC library. All the BACs identified hybridized with just one overgo. The average numbers of hits per overgo was 13.6. Repeat masking was clearly effective, as only 4% of the overgos (3) identified more than 27 BACs in the library and only one identified a disproportionate number of BACs (61). BAC end sequencing was carried out and sequence similarity searches (BLAST) were conducted against draft assemblies of the Citrus clementina genome (http://phytozome.net). The BAC clones corresponding to each probe were mapped within the same scaffold as the target gene, demonstrating that the approach we used was successful in isolating the target genes.