Potential integration sites of the Salmonella genomic island 1 in Proteus mirabilis and other bacteria

Abstract

Sir, We read with great interest the article of Ahmed et al. which described the occurrence of a variant Salmonella genomic island 1 (SGI1) in a clinical isolate of Proteus mirabilis. SGI1 was described initially in epidemic multidrug-resistant (MDR) Salmonella enterica serovar Typhimurium (S. Typhimurium) phage type DT104 strains. The 43 kb SGI1 is located between the thdF and int2 genes of the chromosome of S. Typhimurium and contains an antibiotic resistance gene cluster conferring resistance to ampicillin, chloramphenicol/florfenicol, streptomycin/spectinomycin, sulphonamides and tetracyclines. The int2 gene is part of a retron sequence which has been reported only in S. Typhimurium. In other S. enterica serovars, SGI1 is located between the thdF and yidY genes. The antibiotic resistance gene cluster is located near the 30 end of SGI1 and constitutes a complex class 1 integron called In104. Variant SGI1 antibiotic resistance gene clusters have been described in a wide variety of S. enterica serovars and have been named SGI1-A to SGI1-L. – 5 These gene clusters were likely generated after chromosomal recombinational events or by antibiotic resistance gene cassette replacement at one of the attI1 sites. Recently, SGI1 has been shown to be mobilized using the R55 conjugative plasmid and thus transferred by conjugation to Salmonella or Escherichia coli recipient strains. The site specific integration in Salmonella or E. coli occurs at the last 18 bp of the 30 end of the thdF gene named attB. The attP site, which is almost identical to attB, is present on the mobile circular form of SGI1. Thus integration or excision of SGI1 into the chromosome occurs through homologous recombination between the attB and attP sites mediated by the integrase int gene which is the first gene found at the 50 end of SGI1. SGI1 is thus considered to be an integrative mobilizable element. Interestingly, for the first time, Ahmed et al. identified SGI1 in a bacterium other than Salmonella. The antibiotic resistance gene cluster identified corresponded to that of the SGI1-L complex integron reported previously in S. Newport where the aadA2 gene cassette is replaced by a dfrA15 trimethoprim resistance gene cassette at the first attI1 site of this complex integron. However, although a circular extrachromosomal form of SGI1 was detected in P. mirabilis, the authors failed to identify the integration site in the P. mirabilis chromosome. The authors used the primers U7-L12 (targeting the Salmonella thdF gene) and LJ-R1 (targeting the SGI1 int gene) to confirm by PCR the location of SGI1 downstream of thdF in the Salmonella chromosome. Since this PCR failed to produce an amplicon in P. mirabilis 18306, the authors concluded that the integration site of SGI1 into the chromosome of P. mirabilis 18306 may be different from that of S. enterica. Using the Blastn algorithm (http://www.ncbi.nlm.nih.gov/) we searched for a thdF gene homologue in the P. mirabilis strain HI4320 sequenced genome (http://www.sanger.ac.uk/ cgi-bin/blast/submitblast/p_mirabilis) and identified one gene that is 70% identical in nucleotide sequence to that of the S. Typhimurium thdF gene. The last 18 bp of the putative P. mirabilis thdF gene had a nucleotide sequence that was 100% identical to that of the SGI1 attP site and thus likely constitutes the attB site in P. mirabilis. In S. enterica serovars or E. coli, the attB sites usually differ in 1–2 bp from the SGI1 attP site. Analysis of the potential U7-L12 (20-mer primer) binding site in the thdF gene of the P. mirabilis strain HI4320 sequenced genome revealed seven nucleotide differences which may explain why the PCR failed to generate an amplicon. The identification of SGI1 in a P. mirabilis clinical isolate is of great interest as the spread of the multidrug resistance phenotype in conjunction with the potential for enhanced virulence has significant clinical implications. Since the SGI1 may also spread to other bacterial species, we searched for potential attB target sites of SGI1 in bacteria other than Salmonella and E. coli, using several databases and Blastn. The bacteria found with a potential attB SGI1 target site differing in no more than three nucleotides from the SGI1 attP site are listed in Table 1. The fact that potential attB sites were identified in diverse bacteria such as Shigella spp., Vibrio spp., Pseudomonas spp., Brucella spp., Legionella pneumophila and Klebsiella pneumoniae raises the potential for SGI1 to emerge in other human pathogens. Of particular interest in the evolution of SGI1 is that in some Shewanella genome sequencing projects, parts of SGI1 are found downstream of the respective thdF gene, also named trmE in the following database (http://genome.jgi-psf.org/ mic_home.html). In conclusion, the horizontal transfer of SGI1 may play an important role in the dissemination of multidrug resistance and potential virulence not only among S. enterica serovars but also among other bacterial genera of clinical importance.