PGI2 Is a Novel SGI1-Relative Multidrug-Resistant Genomic Island Characterized in Proteus mirabilis

ABSTRACT

A novel 61,578-bp genomic island named Proteus genomic island 2 (PGI2) was characterized in Proteus mirabilis of swine origin in China. The 23.85-kb backbone of PGI2 is related to those of Salmonella genomic island 1 and Acinetobacter genomic island 1. The multidrug resistance (MDR) region of PGI2 is a complex class 1 integron containing 14 different resistance genes. PGI2 was conjugally mobilized in trans to Escherichia coli in the presence of a conjugative IncC helper plasmid.

TEXT

Genomic islands (GIs), such as pathogenicity islands and antimicrobial resistance islands, are approximately 10- to 200-kb discrete DNA regions in the bacterial genome acquired through horizontal gene transfer (1, 2). GIs enhance genome plasticity by the acquisition of virulence genes, antimicrobial resistance genes, and other genes to promote the adaptive evolution of bacteria (3). Salmonella genomic island 1 (SGI1) is a mobilizable genetic element mediating multidrug resistance (MDR) initially found in Salmonella enterica serovar Typhimurium DT104 (4) and also identified in Proteus mirabilis (5) and Morganella morganii (6). SGI1 consists of a 27.4-kb backbone and an MDR region that is a complex class 1 integron named In104 (7). In recent years, three SGI1-relative elements, SGI2 (8), Proteus genomic island 1 (PGI1) (9, 10), and Acinetobacter genomic island 1 (AGI1) (11), were reported in S. enterica, P. mirabilis, and Acinetobacter baumannii, respectively. All four GIs incorporate into chromosomes at a specific location (3′ end of the trmE gene) and carry diverse antimicrobial resistance genes, indicating they may have the same origin and have evolved into different branches. In the present study, we characterized another novel SGI1-relative element, Proteus genomic island 2 (PGI2), in P. mirabilis of swine origin in China.

P. mirabilis strain BC11-24 was isolated from the liver of a diarrheic piglet on a swine farm in November 2016 in Sichuan province. Antimicrobial susceptibility was determined by the disk diffusion method according to the CLSI guidelines (12–14). Strain BC11-24 was resistant to ampicillin, kanamycin, streptomycin, spectinomycin, apramycin, trimethoprim, trimethoprim-sulfamethoxazole, tetracycline, nalidixic acid, ciprofloxacin, norfloxacin, chloramphenicol, and florfenicol. The presence of SGI1/PGI1 was determined by PCR amplification and sequencing targeting the SGI1/PGI1 integrase gene (15) (see Table S1 in the supplemental material). A BLAST analysis showed that the integrase gene in strain BC11-24 had nucleotide identities of 96 and 82% to int_PGI1 and int_SGI1, respectively, suggesting that strain BC11-24 harbored a novel SGI1-relative element.

The whole genome of BC11-24 was sequenced using a PacBio RS II sequencing instrument (100-fold average read depth) and an Illumina HiSeq platform (400-bp paired-end reads with approximately 200-fold average coverage). The chromosome was assembled into one scaffold using software SMRT portal v.3.2.0. The genome sequence predicts a single mutation in both GyrA (S83I) and ParC (S80I), explaining the nalidixic acid, ciprofloxacin and norfloxacin resistance. Fifteen antimicrobial resistance genes were identified by ResFinder 2.1 (http://www.genomicepidemiology.org/). A sequence analysis indicated that 14 resistance genes, except for tetracycline resistance gene tet(J), were carried by an SGI1-relative element that was designated PGI2. A BLAST analysis showed that PGI2 was 61,578 bp in size incorporated into the 3′ end of trmE (Fig. 1a). PGI2 contains 68 open reading frames (ORFs) composed of a 23.85-kb backbone (Fig. 1a) and a 37.73-kb MDR region (Fig. 1b).

FIG 1.

Structure of PGI2. (a) PGI2 backbone compared with the SGI1 and AGI1 backbones. Structures are drawn to scale from GenBank accession numbers MG201402 (PGI2), KM234279 (SGI1), and KP054476 (AGI1). Genes and ORFs are shown as arrows, which indicate their orientations of transcription. The locations of integrons in PGI2, SGI1, and AGI1are indicated. attL and attR represent the attachment sites of the islands shown as vertical bars. Shared regions with significant nucleotide identities are indicated by shading with the identities shown. (b) Structure of the complex class 1 integron in PGI2. Structures are drawn to scale from GenBank accession numbers MG201402 (PGI2) and KY924928 (pWJ1). The different regions corresponding to Tn4352, Tn21, and integrons are indicated on the horizontal line. Shared regions with >99.9% identity are indicated by shading. Antimicrobial resistance genes are in red and transposase or integrase genes are in blue. IRi and IRt represent the inverted repeats defining the left and right ends of class 1 integron, respectively.

The backbone of PGI2 that consists of 27 ORFs is related to those of SGI1 and AGI1 (Fig. 1a). A partial backbone of PGI2, which includes rep, traN, traG, and traH, is almost identical (98%) to those of SGI1 and AGI1 (Fig. 1a). The amino acid sequence of Int_PGI2 has 97, 89, and 68% identities with those of Int_PGI1, Int_SGI1, and Int_AGI1, respectively. The right-hand portion of the PGI2 backbone contains three ORFs (PGI2-066 to PGI2-68). The nucleotide sequence of PGI2-066 is 97% identical to that of S044 of SGI1. It is very interesting that another two ORFs, PGI2-067 encoding a universal stress protein family protein (290 amino acids [aa]) and PGI2-068 encoding a putative sulfate permease (508 aa), show 90% and 93% nucleotide identities to orf52 and orf51, respectively, of integrative and conjugative element R391 (16).

The MDR region of PGI2 (corresponding to bases 21,610 to 59,340 in GenBank accession number MG201402) is a complex class 1 integron located adjacent to PGI2_024. It includes two class 1 integrons and many transposons and insertion sequences elements (Fig. 1b). The first class 1 integron harbors five gene cassettes, dfrA16 (trimethoprim), bla_PSE-1 (ampicillin), aadA2 (streptomycin and spectinomycin), cmlA1 (chloramphenicol), and aadA1 (streptomycin and spectinomycin). On the right side of this integron is the IS26-blms (bleomycin) region and IS26-mediated translocatable unit Tn4352, suggesting that IS26 might promote the accumulation of transposons and resistance genes (17). The 20.85-kb region (corresponding to bases 33,307 to 54,155 in GenBank accession number MG201402) containing aacC4 (apramycin, netilmicin and tobramycin), hph (hygromycin), sul2 (sulfonamides), and floR (chloramphenicol and florfenicol) resistance genes is identical to the corresponding region of IncHI2 plasmid pWJ1 (18) with only three base pairs changes, indicating they might have a common origin. The In4 family class 1 integron is found adjacent to the tnp module of Tn21, containing aadA2 and lnuF (lincosamide) gene cassettes that were also observed in SGI1-W (19). The MDR region of PGI2 is not surrounded by 5-bp duplications, which might result from the combination and rearrangement of two class 1 integrons located in different sites of the PGI2 backbone. The lack of a resolvase-encoding gene in PGI2 compared with SGI1, PGI1, and AGI1 might support this hypothesis.

SGI1-relative GIs (SGI1, SGI2, and PGI1) could shape the free circular form and transfer to Enterobacteriaceae bacteria with the help of conjugative IncA/C plasmids (20–23). The circular extrachromosomal form could be detected in strain BC11-24 after two rounds of PCR amplification by using the primers listed in Table S1. To determine the transferability of PGI2, a cefotaxime-resistant IncC plasmid pPm14C18 from Escherichia coli J53-pPm14C18 (24) was conjugated into BC11-24 using Shigella and Salmonella agar plates containing 4 mg/liter cefotaxime and 64 mg/liter tetracycline to select P. mirabilis transconjugants. The conjugation experiment was performed by using P. mirabilis BC11-24 carrying pPm14C18 as the donor strain and sodium azide-resistant E. coli J53 as the recipient strain with selection on eosin-methylene blue agar plates containing 300 mg/liter sodium azide and 100 mg/liter apramycin. The transconjugant was further determined by the mobilization of PGI2 and its location in E. coli with primers listed in Table S1. The results showed PGI2 could be successfully transferred to E. coli and incorporated into the 3′ end of trmE, which is identical to what was found for SGI1 and PGI1 (20, 23). PGI2 was shown to be mobilized by an IncC plasmid, presumably because it retains the tra genes traN, traG, and traH (25).

In conclusion, we characterized a novel MDR SGI1-relative GI, PGI2, in P. mirabilis. PGI2 was a mobilizable GI that could be transferred to E. coli with the help of the IncC plasmid. More SGI1-relative GIs might be discovered with more whole-genome sequencing, which will help to reveal the evolutionary history of SGI1-relative GIs, including the evolution of backbones and the acquisition process of antimicrobial resistance genes.

Accession number(s).

The complete nucleotide sequences of PGI2 and the genome of P. mirabilis strain BC11-24 characterized in this study were submitted to GenBank and assigned accession numbers MG201402 and CP026571, respectively.