Genetic snapshots of the Rhizobiumspecies NGR234 genome

Abstract

Background: In nitrate-poor soils, many leguminous plants form nitrogen-fixing symbioses with members of the bacterial family Rhizobiaceae. We selected Rhizobium sp. NGR234 for its exceptionally broad host range, which includes more than 112 genera of legumes. Unlike the genome of Bradyrhizobium japonicum, which is composed of a single 8.7 Mb chromosome, that of NGR234 is partitioned into three replicons: a chromosome of about 3.5 Mb, a megaplasmid of more than 2 Mb (pNGR234b) and pNGR234a, a 536,165 bp plasmid that carries most of the genes required for symbioses with legumes. Symbiotic loci represent only a small portion of all the genes coded by rhizobial genomes, however. To rapidly characterize the two largest replicons of NGR234, the genome of strain ANU265 (a derivative strain cured of pNGR234a) was analyzed by shotgun sequencing. Results: Homology searches of public databases with 2,275 random sequences of strain ANU265 resulted in the identification of 1,130 putative protein-coding sequences, of which 922 (41%) could be classified into functional groups. In contrast to the 18% of insertion-like sequences (ISs) found on the symbiotic plasmid pNGR234a, only 2.2% of the shotgun sequences represent known ISs, suggesting that pNGR234a is enriched in such elements. Hybridization data also indicate that the density of known transposable elements is higher in pNGR234b (the megaplasmid) than on the chromosome. Rhizobium-specific intergenic mosaic elements (RIMEs) were found in 35 shotgun sequences, 6 of which carry RIME2 repeats previously thought to be present only in Rhizobium meliloti. As non-overlapping shotgun sequences together represent approximately 10% of ANU265 genome, the chromosome and megaplasmid may carry a total of over 200 RIMEs. Conclusions: 'Skimming' the genome of Rhizobium sp. NGR234 sheds new light on the fine structure and evolution of its replicons, as well as on the integration of symbiotic functions in the genome of a soil bacterium. Although most putative coding sequences could be distributed into functional classes similar to those in Bacillus subtilis, functions related to transposable elements were more abundant in NGR234. In contrast to ISs that accumulated in pNGR234a and pNGR234b, the hundreds of RIME elements seem mostly attributes of the chromosome.