Molecular analysis of the rfb (O antigen) gene cluster of Salmonella enterica serovar anatum (group E1)

Abstract

Salmonella enterica is highly polymorphic for the 0 antigen, which is synthesized under the control of rfb genes. To understand the genetic basis of this polymorphism, comparative studies of the 0 antigen genetics of S. enterica groups A, B, C1, C2, D, and E1 have been undertaken in this laboratory (Verma et al., 1988; Verma and Reeves, 1989; Jiang et al., 1991; Liu et al., 1991; Lee et al., 1992; Brown et al., 1992). The work on groups A, B, D and C2 has shown that there is substantial similarity in the rfb region, with only limited regions of low similarity and non homology. This thesis describes the cloning and sequence of rfb region from S. enterica strain M32 (serovar anatum, group E1), and compare it to the rfl) region of strain LT2 (serovar Typhimurium, group B) as a representative of the related groups B, A and D. Some of the work on other groups was done at the same time as the study of group B.

The rfb gene cluster of strain M32 was cloned as a series of overlapping clones, and compared to that of strain LT2 by Southern hybridization and restriction mapping. The data revealed that the M32 17‘?) gene cluster has regions which are essentially identical to that of strain LT2, and which flank a central 6.5kb region of low or non similarity. It was also found that a 1.8kb region in the right end of M32 rfb gene cluster is not present in strain LT2. The regions unique to strain M32 were sequenced. Analysis of the sequences revealed six complete open reading frames in the central region and one in the right end region. All the open reading frames are transcribed in the same direction.

Comparison of the rfb gene cluster of strain M32 to that of strain LT2 revealed that genes for biosynthetic pathways common to both strains are conserved and have very similar sequences, and theses genes are TDP-rhamnose and GDP—mannose biosynthetic pathway genes, and the galactose and rhamnose transferase genes. In contrast, the five genes for CDP-abequose synthesis, present in strain LT2, are absent in strain M32. A region containing 07772.8, 0rf14.1 and the mannose transferase gene of strain LT2 is replaced in strain M32 by the DNA containing 0rf7.9, 0779.6 and 0rf10.8 which do not show any detectable nucleotide sequence similarity with the three LT2 genes. 0217.9 encodes a protein with 12 predicted transmembrane segments, and the distribution of these 12 transmembrane segments is similar to that of the 12 segments of the protein encoded by 0rf12.8 of strain LT2. The function of 0rf7.9 is unclear, but it may play a role in 0 antigen translocation across the inner membrane. 0rf10.8 encodes the mannose transferase (Liu et al., 1992), and the function of 0179.6 is unknown. 0rf17.4 which is located in the right end region unique to strain M32 encodes a protein which shows significant similarity to the protein encoded by rfc of group B S. enterica, and it could be the rfc gene of strain M32.

The rfb gene clusters of strains M32 and LT2 differ in their central regions but are conserved in flanking regions, with a gradient in similarity level between the central and flanking regions. It was concluded that much of these two gene clusters is of common ancestry, and the substantial differences in the central regions result from one or more recombination events that replaced one set of genes by another unrelated set.

Like the rfb gene cluster of group B, the rfb gene cluster of strain M32 is also low in G+C content. It was further concluded that the group E1 rfl) gene cluster, like that of group B, arose in a low G+C content species and was acquired by S. enterica by lateral transfer recently.

The location and functional analysis of rbeT is also presented in this thesis.