Variation between pathogenic serovars within Salmonella pathogenicity islands

Abstract

Although four of the five Salmonella pathogenicity islands (SPIs) have been characterized in detail for Salmonella enterica serovar Typhimurium, and the fifth has been characterized for Salmonella enterica serovar Dublin, there have been limited studies to examine them in detail in a range of pathogenic serovars of S. enterica. The aim of this study was to examine these regions, shown to be crucial in virulence, in pathogenic serovars to identify any major deletions or insertions that may explain variation in virulence and provide further understanding of the elements involved in the evolution of these regions. Multiple strains of each of the 13 serovars were compared by Southern blot hybridization using a series of probes that together encompassed the full length of all five SPIs. With the exception of serovar Typhimurium, all strains of the same serovar were identical in all five SPIs. Those serovars that differed from serovar Typhimurium in SPI-1 to SPI-4 and from serovar Dublin in SPI-5 were examined in more detail in the variant regions by PCR, and restriction endonuclease digestion and/or DNA sequencing. While most variation in hybridization patterns was attributable to loss or gain of single restriction endonuclease cleavage sites, three regions, in SPI-1, SPI-3, and SPI-5, had differences due to major insertions or deletions. In SPI-1 the avrA gene was replaced by a 200-base fragment in three serovars, as reported previously. In SPI-5, two serovars had acquired an insertion with similarity to the pagJ and pagK genes between pipC and pipD. In SPI-3 the genes sugR and rhuM were deleted in most serovars and in some were replaced by sequences that were very similar to either the Escherichia coli fimbrial operon, flanked by two distinct insertion sequence elements, or to the E. coli retron phage PhiR73. The distribution of these differences suggests that there have been a number of relatively recent horizontal transfers of genes into S. enterica and that in some cases the same event has occurred in multiple lineages of S. enterica. Thus, it seems that insertion sequences and retron phages are likely to be involved in continuing evolution of the pathogenicity islands of pathogenic Salmonella serovars.

FIG. 1.

Variation between Salmonella serovars within SPI-3 as detected using Southern hybridization and chromosomal amplification analysis. (A) SPI-3 genes and their direction of transcription (modified from reference 3). (B) Extent of hybridization probe 3 and PCRs (shaded blocks) used for detailed analysis. (C) EcoRV and HindIII restriction maps for serovar Typhimurium SL1344. The numbers indicate the sizes of the HindIII and EcoRV fragments detected (in kilobases). +, presence of HindIII and EcoRV fragment of the same size as that seen in serovar Typhimurium SL1344. ∗, serovar Typhimurium strains 9806584 (PT9), C5, LT2, and SL1344.

FIG. 2.

Variant SPI-3 regions in serovars Derby, Ratchaburi, Virchow, Infantis, Bovismorbificans, and Zanzibar compared to serovar Typhimurium. The regions encoding selC and rmbA are shown as arrows indicating the direction of transcription. Narrower boxes indicate the sequences shared by some of the serovars, with shading used to distinguish the different regions. The large insertions in serovars Typhimurium, Derby, and Ratchaburi and the genes they encode are shown as double-headed arrows. The insertions are shown at 1/10 the scale of the other regions.

FIG. 3.

Insertion at left-hand end of SPI-3 in serovars Ratchaburi and Derby compared to that in serovar Typhimurium, with selC at the left side of each schematic. Open reading frames are shown as arrows indicating the direction of transcription. Lines in tpase2 of serovar Derby indicate interruptions to the reading frame by a nonsense mutation and a deletion creating a frame shift.

FIG. 4.

(A) The 431-bp insertion in serovars Derby and Ohio, between pipC and pipB of SPI-5. (B) Predicted amino acid sequence of this insertion compared to PagJ and PagK, showing 24% identity and 42% similarity. Asterisks and dots indicate identical and similar amino acid residues, respectively.

FIG. 5.

The phylogenetic tree for the pathogenicity islands of 13 Salmonella serovars inferred using restriction site data and Restml of PHYLIP 3.6. Major genetic events identified in the pathogenicity islands (insertions and deletions compared to serovar Typhimurium) are indicated as follows: (I), deletion of avrA and replacement with a 0.2-kb sequence; (IIIa), deletion within sugR; (IIIb), deletion of sugR and rhuM and replacement with a 0.7-kb sequence; (IIIc), deletion of sugR and rhuM and replacement with a 3-kb sequence; (IIId), deletion of sugR and rhuM and replacement with a 10-kb sequence; and (V), insertion of a 0.5-kb sequence within SPI-5. The number next to each branch indicates the confidence that can be placed on that branch as determined by bootstrap resampling of the restriction site data 100 times, inferring the best tree for each set of resampled data and then inferring a consensus tree using Consensus (PHYLIP 3.6). Each number indicates the percentage of the trees inferred from the resampled data that contained that branch.