-
PDF
- Split View
-
Views
-
Cite
Cite
Sophie Octavia, Weizhen Xu, Oon Tek Ng, Kalisvar Marimuthu, Indumathi Venkatachalam, Bernadette Cheng, Raymond T P Lin, Jeanette W P Teo, Identification of AbaR4 Acinetobacter baumannii resistance island in clinical isolates of blaOXA-23-positive Proteus mirabilis, Journal of Antimicrobial Chemotherapy, Volume 75, Issue 3, March 2020, Pages 521–525, https://doi.org/10.1093/jac/dkz472
- Share Icon Share
Abstract
blaOXA-23 is a class D carbapenemase-encoding gene typical of the Acinetobacter genus. However, its occurrence in the Enterobacteriaceae is uncommon. Here we provide the genome characterization of blaOXA-23-positive Proteus mirabilis.
In Singapore, a national surveillance of carbapenem non-susceptible clinical Enterobacteriaceae has enabled the collection of OXA-23 bearing isolates. Three clinical P. mirabilis were whole-genome sequenced using Oxford Nanopore MinION and Illumina platforms. The sequence accuracy of MinION long-read contigs was enhanced by polishing with Illumina-derived short-read data.
In two P. mirabilis genomes, blaOXA-23 was detected as two copies, present on the chromosome and on a 60018 bp plasmid. blaOXA-23 was associated with the classic Acinetobacter composite transposon Tn2006, bounded by two copies of ISAba1 bracketing the carbapenemase gene. The Tn2006 itself was embedded within an Acinetobacter baumannii AbaR4 resistance island. In the chromosome, the AbaR4 was found integrated into the comM gene, which is also the preferred ‘hotspot’ in A. baumannii. In the plasmid, AbaR4 integrated into a putative colicin gene.
Our description of an A. baumannii AbaR4 encoding blaOXA-23 in P. mirabilis is to our knowledge the first description of an Acinetobacter resistance island in Proteus and suggests that P. mirabilis may be a reservoir for this class D carbapenemase gene.
Introduction
blaOXA-23 is a prominent carbapenem-hydrolysing class D β-lactamase (CHDL) and is one of the carbapenemase gene families amongst the five known acquired gene clusters of CHDLs that feature in Acinetobacter populations.1 Transposons Tn2006, Tn2007, Tn2008 and Tn2009 are known to be associated with blaOXA-23 carriage.2–5
In the Enterobacteriaceae, CHDLs are typically encoded by a genetically dissimilar and unrelated set of genes, the blaOXA-48-like genes.6 In contrast, Acinetobacter CHDLs such as blaOXA-51,7,blaOXA-588,9 and blaOXA-2310–12 have been infrequently detected in the Enterobacteriaceae.
The presence of blaOXA-23 in Enterobacteriaceae is exceptional and has only been described in two species, Proteus mirabilis and Escherichia coli. In the 1990s, chromosomally encoded blaOXA-23 was first described in P. mirabilis11 and again more recently in a single clinical isolate12 and as a clonal cluster of community-acquired infections in France.13 In 2014, we described the first detection of blaOXA-23 in E. coli.10 The blaOXA-23 was flanked by IS1 and did not resemble Acinetobacter blaOXA-23-associated transposons. blaOXA-23-positive E. coli were subsequently observed in 14 clinical isolates of E. coli in India.14 In the latter two reports, the blaOXA-23 was plasmid located.
Since our initial characterization of blaOXA-23-positive E. coli, continued national surveillance on non-carbapenem-susceptible Enterobacteriaceae isolates has detected several more such blaOXA-23-positive E. coli and P. mirabilis. Here, we focus on the genomic characterization of blaOXA-23-positive P. mirabilis isolates, in particular, the genetic environment of blaOXA-23.
Materials and methods
Sample collection
A total of 25 non-duplicate blaOXA-23-positive isolates were collected over 2014–17. These isolates came from retrospectively collected, phenotypically carbapenem-resistant Enterobacteriaceae (CRE) submitted to the reference National Public Health Laboratory for mandatory CRE surveillance. Of the 25 blaOXA-23-positive isolates, 16 were clinical isolates and 9 were from screening specimens. Screening specimens comprised rectal swabs and stool samples. Rectal swabs were collected from: (i) high-risk patients on admission (defined as prior hospitalization in the preceding 1 year); (ii) patients being transferred to high-risk units (e.g. ICU and haematology and oncology wards); and (iii) epidemiologically linked contacts of carbapenemase-producing-Enterobacteriaceae-positive patients. Four of the blaOXA-23-positive isolates were P. mirabilis and the remainder were E. coli. The blaOXA-23-positive isolates made up 1.7% (25/1468) of all carbapenemase genotypes screened at the reference laboratory for the duration of the study period.
Phenotypic susceptibility testing
Antibiotic susceptibility testing and organism identification for clinical cultures were conducted at participating institutions. CREs were defined as Enterobacteriaceae isolates with MICs of >1 mg/L for imipenem or meropenem using automated testing systems (VITEK® 2 instrument). PCR screening for blaOXA-23 was performed as previously described.10
Plate conjugation assays
Solid media conjugation assays were performed to assess the transferability of blaOXA-23 from the P. mirabilis clinical isolates to an azide-resistant recipient, E. coli J53. Selection for transconjugants was made on LB agar containing 100 mg/L of ampicillin and 50 mg/L of sodium azide.
WGS and analysis
Three P. mirabilis underwent long-read sequencing using the MinION R9.4.1 flow cell (Oxford Nanopore Technologies). The long reads were combined with Illumina MiSeq paired-end reads of 300 bp for hybrid assembly using Unicycler v0.3.1.15 Prokka16 was used to annotate the contigs. ABRicate (https://github.com/tseemann/abricate) was used to screen the presence of antibiotic resistance genes. IS elements were identified using ISfinder (https://isfinder.biotoul.fr/blast.php). Complete genomes were obtained when only one contig per chromosome/plasmid was obtained and this contig could be circularized with overlapping reads of >100 bp at both ends. The plasmid incompatibility type was identified using PlasmidFinder and pMLST at https://cge.cbs.dtu.dk/services/. The alignment and visualization of plasmids were performed with BRIG v0.95.17
GenBank accession numbers
Assembled genomes of PM1301, PM1157 and PM1224 have been deposited in GenBank under the accession numbers CP044135, CP044136 and CP044134, respectively. Raw reads from both Nanopore and MiSeq are available in the Short Reads Archive (BioProject number PRJNA540912).
Core genome MLST (cgMLST)
The clonal relationship between the P. mirabilis genomes was assessed by comparing the genomes of PM1157, PM1224 and PM1301 with 132 other genomes downloaded from the NCBI genomes database. Briefly, a database containing annotated genes present in the genome of the reference strain P. mirabilis HI432018 was built using the pyMLST tool (https://github.com/bvalot/pyMLST). A multi-fasta alignment file generated from the script was then used to build a maximum likelihood tree and the pairwise difference between the genomes was calculated in MEGA X.19
Results and discussion
All three clinical P. mirabilis isolates were isolated in the same year, from urine specimens of different patients from the same hospital (Table 1). The meropenem MICs of the isolates were borderline susceptible (Table 1). This phenotype of relative susceptibility has been repeatedly observed.10,12 Therefore, blaOXA-23-positive Enterobacteriaceae were likely to be under-detected when only phenotypic screening was employed.
Isolate . | Date of isolation . | Specimen source . | Replicon . | Additional acquired resistance determinants . | MIC (mg/L) . | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
IPM . | MEM . | CAZ . | CTX . | FEP . | ATM . | TZP . | AMK . | GEN . | TOB . | CIP . | LVX . | CST . | |||||
PM1157 | 5 May 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa | 8 | 8 | <1 | <2 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1224 | 4 June 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa, adA5, strA, strB, mph(A), sul1, sul2, dfrA17 | 8 | 2 | <1 | 4 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1301 | 4 July 2014 | urine, clinical | no replicon detected | blaCTX-M-3, blaSHV-12, blaTEM-1, aac(3)-IId, aadA1, aph(3′)-Ia, strA, strB, sul2, dfrA1 | 8 | 4 | <1 | 16 | 8 | <2 | 32/4 | <4 | >8 | 2 | 2 | 4 | >4 |
Isolate . | Date of isolation . | Specimen source . | Replicon . | Additional acquired resistance determinants . | MIC (mg/L) . | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
IPM . | MEM . | CAZ . | CTX . | FEP . | ATM . | TZP . | AMK . | GEN . | TOB . | CIP . | LVX . | CST . | |||||
PM1157 | 5 May 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa | 8 | 8 | <1 | <2 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1224 | 4 June 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa, adA5, strA, strB, mph(A), sul1, sul2, dfrA17 | 8 | 2 | <1 | 4 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1301 | 4 July 2014 | urine, clinical | no replicon detected | blaCTX-M-3, blaSHV-12, blaTEM-1, aac(3)-IId, aadA1, aph(3′)-Ia, strA, strB, sul2, dfrA1 | 8 | 4 | <1 | 16 | 8 | <2 | 32/4 | <4 | >8 | 2 | 2 | 4 | >4 |
MICs were determined using standard microbroth dilution methods.29
IPM, imipenem; MEM, meropenem; CAZ, ceftazidime; CTX, cefotaxime; FEP, cefepime; ATM, aztreonam; TZP, piperacillin/tazobactam constant 4; AMK, amikacin; GEN, gentamicin; TOB, tobramycin; CIP, ciprofloxacin; LVX, levofloxacin; CST, colistin.
Isolate . | Date of isolation . | Specimen source . | Replicon . | Additional acquired resistance determinants . | MIC (mg/L) . | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
IPM . | MEM . | CAZ . | CTX . | FEP . | ATM . | TZP . | AMK . | GEN . | TOB . | CIP . | LVX . | CST . | |||||
PM1157 | 5 May 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa | 8 | 8 | <1 | <2 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1224 | 4 June 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa, adA5, strA, strB, mph(A), sul1, sul2, dfrA17 | 8 | 2 | <1 | 4 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1301 | 4 July 2014 | urine, clinical | no replicon detected | blaCTX-M-3, blaSHV-12, blaTEM-1, aac(3)-IId, aadA1, aph(3′)-Ia, strA, strB, sul2, dfrA1 | 8 | 4 | <1 | 16 | 8 | <2 | 32/4 | <4 | >8 | 2 | 2 | 4 | >4 |
Isolate . | Date of isolation . | Specimen source . | Replicon . | Additional acquired resistance determinants . | MIC (mg/L) . | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
IPM . | MEM . | CAZ . | CTX . | FEP . | ATM . | TZP . | AMK . | GEN . | TOB . | CIP . | LVX . | CST . | |||||
PM1157 | 5 May 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa | 8 | 8 | <1 | <2 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1224 | 4 June 2014 | urine, clinical | no replicon detected | blaTEM-2, aac(3)-IIa, adA5, strA, strB, mph(A), sul1, sul2, dfrA17 | 8 | 2 | <1 | 4 | 4 | <2 | >64/4 | <4 | >8 | 8 | >2 | 4 | >4 |
PM1301 | 4 July 2014 | urine, clinical | no replicon detected | blaCTX-M-3, blaSHV-12, blaTEM-1, aac(3)-IId, aadA1, aph(3′)-Ia, strA, strB, sul2, dfrA1 | 8 | 4 | <1 | 16 | 8 | <2 | 32/4 | <4 | >8 | 2 | 2 | 4 | >4 |
MICs were determined using standard microbroth dilution methods.29
IPM, imipenem; MEM, meropenem; CAZ, ceftazidime; CTX, cefotaxime; FEP, cefepime; ATM, aztreonam; TZP, piperacillin/tazobactam constant 4; AMK, amikacin; GEN, gentamicin; TOB, tobramycin; CIP, ciprofloxacin; LVX, levofloxacin; CST, colistin.
The relatedness between isolates was determined by cgMLST.13,P. mirabilis isolates PM1157 and PM1224 were likely to be clonal in nature with only two SNPs identified, although the epidemiological link between these two isolates was not found. In contrast, PM1301 had 1859 differences when compared with PM1157 and PM1224. For comparison, these isolates had >1800 nucleotide differences when compared with the French cluster of blaOXA-23-positive isolates13 and not deemed to be phylogenetically related.
Within P. mirabilis genomes PM1157 and PM1224, two copies of blaOXA-23 were detected, one on the chromosome and the other on the 60018 bp plasmid (GenBank accession number MK941846) (Figure 1). In the remaining isolate, PM1301, the plasmid was not observed and blaOXA-23 was present only on the chromosome. Specifically, the AbaR4 had 100% nucleotide identity to the AbaR4 (GenBank accession number JN107991.2) that was first described in a carbapenem-resistant A. baumannii strain, D36, isolated in Australia.20 Diverse genetic contexts are not unusual for AbaR4, appearing in the chromosome or within plasmids in A. baumannii populations.1,21 The AbaR4 has only one resistance gene, blaOXA-23, which itself is carried within Tn2006 bounded by two copies of the insertion sequence ISAba120 (Figure 1a). The backbone of AbaR4 is Tn6022, comprising a set of transposition genes, tniC–tniA–tniB–tniD–tniE, that help catalyse transposition, uspA (universal stress protein encoding gene) and sup (sulphate permease).20 Other than in Australia, AbaR4-type detection amongst carbapenem-resistant A. baumannii has also been reported in South Korea,22 Taiwan23 and Europe. To our knowledge, this is the first description of an Acinetobacter AbaR4-D36-type resistance island described in P. mirabilis.
(a) Genetic comparison of the chromosomal genetic environment of blaOXA-23 in P. mirabilis (PM1224, PM1301, PM1157) with GenBank reference A. baumannii strain D3620 and P. mirabilis strain ESBL4969.12,blaOXA-23 is represented by a red arrow and the remaining ORF genes are colour coded by functional category: dark blue, ISs; sky blue, plasmid transfer, replication, establishment, stability, maintenance; grey, hypothetical protein. Integration of AbaR4 into the comM gene is depicted by green arrows. The 5 bp target duplication site for AbaR4 is indicated by pink vertical lines. The extents of AbaR4 and Tn2006 are also indicated. GenBank accession numbers are cited in brackets. aKU302354 was found to have an identical genetic context to the blaOXA-23-positive P. mirabilis cluster identified in France (GenBank accession number SLUF00000000).13 (b) BRIG alignment of the P. mirabilis 60018 bp plasmid. The P. mirabilis blaOXA-23-bearing plasmid is represented in green. The plasmid backbone has 99% nucleotide identity to P. rettgeri strain 16pre36-2 (GenBank accession number KX832926.1), indicated in blue. The A. baumannii resistance island AbaR4 (JN107991.2) is in purple and has integrated in the putative colicin gene of the P. mirabilis plasmid, indicated as the yellow ORF. The left and right inverted repeats of AbaR4 (IRL and IRR, respectively) are indicated by red perpendicular lines. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.
In the genomes of PM1157 and PM1224, a completely closed plasmid of 60018 bp was obtained. The plasmid appears to be chimeric. The plasmid backbone was 99% identical to the 43191 bp plasmid p16Pre36-2 from Providencia rettgeri strain 16pre3624 (GenBank accession number KX832926.1) with the integration of AbaR4 into a putative colicin gene (Figure 1b). Other than blaOXA-23, two other resistance genes, aac(3)-IIa and blaTEM-2, were identified. No replicons were detected in the 60018 bp plasmid. Not all schemes can fully classify plasmids; Shintani et al.25 found that the proportion of Enterobacteriaceae plasmids that could be replicon typed was about 75%. In P. rettgeri p16Pre36-2, an Inc group was not previously assigned.24 AbaR4 in the P. mirabilis chromosome was flanked by a 5 bp target site duplication of 5′-GCGGT-3′. Likewise, inverted repeats of AbaR4 were also present in the plasmid copy but flanked by a different sequence of 5 bp direct repeats of 5′-GAAAAT-3′. The integration sites of AbaR4 differ for the chromosomal and plasmid locations. The chromosomal copy of AbaR4 was found inserted into the comM gene (encoding ATPase), but the integration sites had no homology to that found in A. baumannii D36. The ComM of P. mirabilis and A. baumannii have only 48.9% amino acid identity. In the plasmid, AbaR4 integrated into a putative colicin gene (GenBank accession number CAQ34879.1). In Acinetobacter, the comM gene is the preferential target site for AbaR integration, although other genomic sites, such as pho, acoA and uup, have also been observed.21 Additionally, plasmid integration of AbaR4 has been described in A. baumannii.26
The earliest report of blaOXA-23-positive P. mirabilis was in 2002, where the carbapenemase gene was found to be chromosomally encoded.9 However, there was no information regarding the genetic environment. In 2016, a clinical P. mirabilis blaOXA-23 isolate was described.12 The blaOXA-23 was not plasmid borne. The OXA-23 gene was located on a 28 kb contig and harboured by Tn2008 (GenBank accession number KU302354) (Figure 1a). Within the same contig, a variety of insertion elements (ISAba125, ISAba15 and ISCR2) as well as different resistance genes (sul2, floR, strB and strA) were detected. This was the same genetic configuration as that found for 19 blaOXA-23-positive P. mirabilis isolates that had formed a cluster.13 Overall, the blaOXA-23 genetic context of previously described isolates was very different from those found in our isolates (Figure 1).
We did not manage to obtain E. coli J53 transconjugants in the P. mirabilis conjugation assays even though components of the conjugative machinery, including type IV coupling and secretion systems (vir region),27 were detected (Figure 1b). It is possible that the conjugation has extremely low efficiency (below our detection limit) or requires specific induction conditions.28 This low transmissibility of blaOXA-23 could account for the infrequent detection and limited spread of OXA-23 P. mirabilis observed in our CRE populations.
Interspecies transfer of genomic islands and resistance genes between Acinetobacter and Enterobacteriaceae have been documented.23,24 AbaRs are non-self-transmissible elements21 and likely would have utilized plasmids as vehicles for mobility into the Proteus host genome.
Acknowledgements
We wish to thank the members of the Carbapenemase-Producing Enterobacteriaceae in Singapore (CaPES) Study Group: Michelle Ang, Benjamin Cherng, Deepak Rama Narayana, Douglas Chan Su Gin, De Partha Pratim, Hsu Li Yang, Indumathi Venkatachalam, Jeanette Teo, Kalisvar Marimuthu, Koh Tse Hsien, Nancy Tee, Nares Smitasin, Ng Oon Tek, Ooi Say Tat, Raymond Fong, Raymond Lin Tzer Pin, Surinder Kaur Pada, Tan Thean Yen and Thoon Koh Cheng.
Funding
This study was supported by: the National Medical Research Council (NMRC) Clinician-Scientist Individual Research Grant (NMRC/CIRG/1463/2016); Singapore Ministry of Education Academic Research Fund Tier 2 grant ‘New Delhi Metallo-β-Lactamase: A global multi-centre, whole-genome study’ (MOE2015-T2-2–096); NMRC Collaborative Grant ‘Collaborative Solutions Targeting Antimicrobial Resistance Threats in Health Systems’ (CoSTAR-HS) (NMRC CGAug16C005); and NMRC Clinician Scientist Award (NMRC/CSA-INV/0002/2016). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Transparency declarations
None to declare.