One of the surprises of our whole genome analysis

and comparison of the 14 ATCC serovars showed the mba genes to be part of a large complex gene superfamily comprising 183 UPA and UUR genes and 22 subfamilies (Figure  5). There were a limited number of unique variable domains as shown in Table  5. We found that all UUR serovars and UPA1 and 6 had more than one tandem repeating unit type in their mba locus. Although some Epacadostat cost of the TRUs in the loci have not yet been observed to be attached to the conserved domain of the mba, they are surrounded by inverted repeats that contain a putative recombinase recognition site. This suggested that these TRUs were involved with the mba and contributed to surface ACP-196 concentration antigen variation. We consider genes without tandem repeats that are in the mba locus and have the putative recombination recognition site to be part of the MBA superfamily. The UPA serovars had a simpler MBA phase variation

systems than the UUR serovars: the UPA conserved domain was surrounded by inverted single base pair repeats, containing the 25 base pair putative recombinase recognition site (Figures  6 and 7). The inverted repeats and a site-specific recombinase were potentially involved in inverting the orientation of the transcriptional promoter and conserved domain in order for expression to occur with one or the other TRU. A list of all genes encoding potential recombinases or transposases is provided in the Additional file 5: 19UU_Recombinases.xls. In most serovars a recombinase or a transposase is located in close

proximity to the mba locus. Selleck ABT-737 Experimental evidence is needed to determine which recombinase is responsible for the rearrangement of the locus. It is interesting to note that one TRU was short and had a high copy number (18 nt – UPA1, 12 nt – UPA6, repeated >30X) and the other one was long and had a low copy number (327 nt -UPA1, 336 nt – UPA6, repeated <5X). Rearrangements of the mba locus were evident in the smaller contigs of unfinished serovar genomes (Figures  6 and 7). UPA1 genome sequencing FER data clearly shows a sub-population in which the conserved domain of the mba is attached to the alternative TRU ([GenBank: NZ_ABES01000008] -gcontig_1106430400161, [GenBank: NZ_ABES01000003] – gcontig_106430400170; Figure 6 & Table  5) and another subpopulation in which another gene is present between the two TRUs ([GenBank: NZ_ABES01000002] – gcontig_1106430400172). The high repeat number of the mba TRUs, and the existence of a subpopulation in the culture being sequenced that has a rearrangement of the mba locus, represent an ambiguity for the assembly software, resulting in the generation of smaller alternative contigs that cannot be assembled into the chromosome. The alternative 327 nt mba TRU of UPA1 is on a 1399 nt long contig [GenBank: NZ_ABES01000008] that contains only this gene, and it ends truncating the 327 nt TRU at only 2.