Genomic islands 1 and 2 carry multiple antibiotic resistance genes in Pseudomonas aeruginosa ST235, ST253, ST111 and ST175 and are globally dispersed

Sir,

Studies of MDR or XDR Pseudomonas aeruginosa often describe an inability to transfer antimicrobial resistance loci into recipient cells, suggesting that plasmids do not play a prominent role in dissemination of MDR loci in this species.^1–6 While two recent studies have highlighted the important role of genomic islands (GIs) in the carriage and transfer of multiple antimicrobial resistance in P. aeruginosa,^7,8 much of the hypothesis was based on earlier studies.^4,7,9,10 Here we present a pilot bioinformatics analysis (strategy detailed in Figure S1, available as Supplementary data at JAC Online) on 22 complete and 252 draft P. aeruginosa (Taxa ID: 287, resistance profiles unknown) genome sequences in the NCBI-Microbial-BLAST database (on 11 April 2016) as evidence of the presence of GI1, GI2 and associated transposons on other globally dispersed clonal lineages.

Excluding our Australian isolates in Table S1, 11% (31/274) of P. aeruginosa genomes in the database carry GI1 while 14% (38/274) carry GI2 or variants of them. It is notable that none of the non-Australian genomes contains both GI1 and GI2. The 31 P. aeruginosa strains that contained GI1 were from the USA, Spain, France, Germany, Japan, Mexico, Argentina and Israel and belonged to different P. aeruginosa STs, including ST235 (19 strains), ST253 (5 strains), ST348 (3 strains), ST179 (2 strains) and ST463 (1 strain). GI2 was identified in 38 P. aeruginosa strains from various European countries, including Germany, Romania, France, Spain, the Netherlands, Belgium, Croatia, Italy, Portugal and Greece as well as the USA, Columbia and India. While most belonged to ST111 (19 isolates), ST175 (10 isolates) or ST235 (7 isolates), one strain of P. aeruginosa typed as ST395 and another as ST823. P. aeruginosa ST235 appears to play a significant role in the carriage of GI1, although some ST235 strains harbour GI2. GI2 associates mostly with P. aeruginosa ST111 and ST175 (Table S1). We also sequenced four additional ST235 isolates collected between 2007 and 2011 from various sources in Sydney (GenBank accession numbers LVEC01000000, LVED01000000, LVEF01000000 and LWGS01000000) and found versions of GI1 and GI2 that were identical to the islands described in strain RNS_PA1 from 2006.

Given that GI1 or GI2 were identified in 69/274 (25%) of the P. aeruginosa genome sequences deposited in the database, we examined the frequency of carriage of transposons Tn6060, Tn6162, Tn6163 and Tn6249 within the selected cohort (Table S1, part C). Nineteen of the 274 (7%) genomes carried a Tn6060-family transposon, and in 74% (14/19) of the genomes it was inserted at an identical location in GI1. Of these 14 strains, 6 belonged to ST235, 5 were ST253 and the remaining 3 strains were ST179, ST463 and an unknown ST, from different countries. Irrespective of ST, 11 of the 14 strains carrying GI1 also contained the cassette array (aadA6-gcuD) found in Tn6162. Strains NCGM1900 and NCGM1984 from Japan had an extra cassette (aacA7) inserted in the array, while the cassette array in strain U2504 was different.

Of 274 genomes in the database 30 (11%) harboured a Tn6163 backbone. Fourteen of these (47%), with ST111, ST175 and ST235, were in GI2 (Table S1, part C). In the remaining 16 strains, Tn6163 was not integrated into GI2 and was found in strains with various STs. None of the P. aeruginosa strains with GI2 carried the Ambler class A carbapenemase gene bla_GES-5 that is seen in XDR strains from Sydney.⁷ ST111 strains from different European countries, all of which carried a Tn6163-like transposon in GI2, contained an aacA4-blaP1b-aadA2 cassette array encoding resistance to gentamicin/tobramycin (aacA4), carbenicillin (blaP1b) and streptomycin/spectinomycin (aadA2). Thus, it is evident that while GI1 and GI2 are hotspots for the insertion of Tn6060-family transposons and Tn6163, respectively, these transposons can also integrate at alternative sites in the P. aeruginosa genome.

While these data are consistent with ST235 being a globally dispersed clone, the acquisition (or loss) of the GIs and the transposons the GIs harbour are likely to influence their resistome. Since all of the P. aeruginosa strains in our study carry one or more class 1 integrons in their genomes, opportunities exist to gain or lose resistance gene cassettes or evolve complex antibiotic resistance loci via homologous recombination, as seen in Tn6060 and Tn6249.

Data presented in the current study provide a snapshot of a global scenario that implicates GI1 and GI2 in the mobilization of MDR loci not only within ST235 but in other globally dominant clones of P. aeruginosa, including ST111 and ST175. All ST235 strains from Sydney that carried both GI1 and GI2 clustered within ST235 strains that contain either GI1 or GI2 (Figure S2), suggesting the ST235 clonal lineage dominant in Sydney is distinct and may have arisen either by transfer of GI1 into a strain of P. aeruginosa that carried GI2 or by the phage-mediated transfer of GI2 into a strain that carried GI1, followed by clonal expansion.

Acknowledgements

We wish to thank colleagues at the Microbiology Laboratory of Royal North Shore Hospital (RNSH) in Sydney for identifying the bacteria and making this project possible.

Funding

This work was funded by the Australian Centre for Genomic Epidemiological Microbiology (AUSGEM) partnership and by a University of Technology Sydney (UTS) small grant to P. R. C.

Transparency declarations

None to declare.

Author contributions

P. R. C. initiated the study, designed and performed BLAST searches and compiled the tables and figures (with assistance from M. J. S.). S. P. D. and P. R. C. wrote the manuscript.

Supplementary data

Figures S1 and S2 and Table S1 are available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/).