Molecular Characterization of Multidrug-Resistant Pseudomonas aeruginosa Isolates in Hospitals in Myanmar

The emergence of multidrug-resistant (MDR) Pseudomonas aeruginosa has become a serious worldwide medical problem. This study was designed to clarify the genetic and epidemiological properties of MDR P. aeruginosa strains isolated from hospitals in Myanmar.

ABSTRACT

The emergence of multidrug-resistant (MDR) Pseudomonas aeruginosa has become a serious worldwide medical problem. This study was designed to clarify the genetic and epidemiological properties of MDR P. aeruginosa strains isolated from hospitals in Myanmar. Forty-five MDR P. aeruginosa isolates obtained from different patients in seven hospitals in Myanmar were screened using the broth microdilution method. The whole genomes of the MDR isolates were sequenced using a MiSeq platform (Illumina). Phylogenetic trees were constructed from single nucleotide polymorphism concatemers. Multilocus sequence types were deduced, and drug resistance genes were identified. Of the 45 isolates, 38 harbored genes encoding carbapenemases, including DIM-1, IMP-1, NDM-1, VIM-2, and VIM-5, and 9 isolates had genes encoding 16S rRNA methylases, including RmtB, RmtD3, RmtE, and RmtF2. Most MDR P. aeruginosa strains isolated in Myanmar belonged to sequence type 1047 (ST1047). This is the first molecular epidemiological analysis of MDR P. aeruginosa clinical isolates in Myanmar. These findings strongly suggest that P. aeruginosa ST1047 strains harboring carbapenemases, including DIM-, IMP-, NDM-, and VIM-type metallo-β-lactamases, have been spreading throughout medical settings in Myanmar.

INTRODUCTION

The emergence of multidrug-resistant (MDR) Pseudomonas aeruginosa has become a serious worldwide medical problem (1, 2). Most MDR P. aeruginosa isolates encode carbapenem and aminoglycoside resistance factors, including metallo-β-lactamases (MBLs), aminoglycoside-modifying enzymes, and 16S rRNA methylases (1, 2). In particular, MBLs and 16S rRNA methylases are associated with high resistance to carbapenems and aminoglycosides, respectively.

MBLs confer resistance to all β-lactams, except the monobactams, and are characterized by their efficient hydrolysis of carbapenems (3). P. aeruginosa isolates producing IMP- or VIM-type MBLs have been detected in various Asian countries (4). For example, isolates producing IMP-type MBLs have been observed in China, Japan, South Korea, Malaysia, Singapore, Thailand, and Vietnam, and isolates producing VIM-type MBLs have been identified in China, India, Indonesia, South Korea, Malaysia, Saudi Arabia, and Taiwan (5, 6). In addition, a P. aeruginosa isolate producing an NDM-type MBL has been reported in India (7). In non-Asian countries, IMP-type MBL-producing P. aeruginosa clinical isolates were detected in Australia, Belgium, Brazil, Canada, the Czech Republic, Denmark, France, Germany, Portugal, and Slovakia, whereas VIM-type MBL producers were detected in Argentina, Austria, Belgium, Bulgaria, Chile, Columbia, Croatia, Egypt, France, Germany, Greece, Hungary, Italy, Kenya, Mexico, Poland, Portugal, Spain, Sweden, Tunisia, the United Kingdom, the United States, and Venezuela (8).

Acquired 16S rRNA methylase genes responsible for extremely high levels of resistance to various aminoglycosides are widely distributed among Gram-negative bacteria, including Enterobacteriaceae and glucose-nonfermentative bacteria (9). To date, 10 kinds of 16S rRNA methylases, including ArmA, RmtA, RmtB, RmtC, RmtD, RmtE, RmtF, RmtG, RmtH, and NpmA, have been found in clinical isolates (10–13).

This is the first molecular epidemiological analysis of MDR P. aeruginosa clinical isolates in Myanmar. Our present study, therefore, would be useful to understand the dynamics of MDR P. aeruginosa strains in Asian countries.

RESULTS

Drug susceptibility and drug resistance factors.

All 45 MDR isolates were sensitive to colistin (Table 1), and all 45 isolates were resistant to amikacin and ciprofloxacin. Thirty-five isolates (77.8%) were resistant to aztreonam, 39 isolates (86.7%) were resistant to cefepime, 44 isolates (97.8%) were resistant to ceftazidime, 40 isolates (88.9%) were resistant to imipenem, and 44 isolates (97.8%) were resistant to meropenem.

TABLE 1.

MIC₉₀ and MIC₅₀ values and percent antimicrobial resistance of MDR P. aeruginosa clinical isolates^a

Antimicrobial agent	Breakpoint for resistance (μg/ml)b	% resistance	MIC data (μg/ml)
			Range	MIC50	MIC90
Amikacin	≥64	100	64 to >1,024	256	>1,024
Aztreonam	≥32	77.8	8 to >1,024	32	>1,024
Cefepime	≥32	86.7	8 to >1,024	512	>1,024
Ceftazidime	≥32	97.8	4 to >1,024	>1,024	>1,024
Ciprofloxacin	≥4	100	4 to >1,024	64	128
Colistin	≥8	0	0.063 to 1	0.5	0.5
Imipenem	≥8	88.9	1 to 512	128	512
Meropenem	≥8	97.8	2 to 1,024	128	512

^an = 45.

^bBreakpoints for antimicrobial resistance were determined according to the guidelines of the Clinical and Laboratory Standards Institute.

Of the 45 MDR P. aeruginosa isolates, 38 (84.4%) encoded metallo-β-lactamase genes, including DIM-1, IMP-1, NDM-1, VIM-1, VIM-2, and VIM-5 (Table 2). Of these, 12 harbored genes encoding both DIM-1 and NDM-1, and one harbored genes encoding both IMP-1 and NDM-1. None of these isolates harbored other carbapenemase-encoding genes.

TABLE 2.

Characterization of multidrug-resistant P. aeruginosa isolates^a

MLST	No. of isolates belonging to same ST	Hospital(s)	No. of strains harboring the gene/total no. of strains for:		QRDR (no. of strains with the amino acid substitution/total no. of strains) forb:
			Carbapenemase-encoding gene(s)	Aminoglycoside resistance gene(s)c	GyrA	ParC
ST233	5	A, B, G	blaVIM-2	aacA7, aacA5, aadA2 (2/5)	S83I	S80L
ST235	4	A, C	blaIMP-1 (2/4), blaNDM-1 (1/4)	aacA1 (1/4), aacA4 (2/4), aacC5 (1/4), aadA1 (3/4), aadA6, aadB	S83I	S80L (2/4)
ST273	4	A, B, D	blaNDM-1 (3/4)	rmtB (1/4), rmtE, aacA4 (2/4), aadA1, aadB	S83I, D87H	S80L
ST314	1	A	blaNDM-1	aacA4	S83I	S80L
ST316	1	C		rmtD3, aacA4	S83I
ST357	3	D, F	blaVIM-2 (1/3), blaVIM-5 (2/3)	aacA4 (1/3), aacA7 (1/3), aacC5 (1/3), aadA1 (1/3)	S83I (2/3), S83T (1/3)	S80L (2/3)
ST446	1	A	blaVIM-2	aacA7, aacC5	S83I	S80L
ST983	2	C		rmtF2, aacA4	S83I, D87N	S80L
ST1047	23	A, B, G	blaDIM-1 (12/23), blaNDM-1 (15/23), blaVIM-2 (8/23)	rmtB4 (2/23), aacA4 (6/23), aadA1 (6/23)	S83I (22/23), S83L (1/23), D87N (1/23)	S80L
ST1121	1	A		aadA1	S83I	S80L

^an = 45.

^bQRDR, quinolone resistance-determining region.

^cAminoglycoside resistance genes, including those encoding aminoglycoside-acetyltransferase and aminoglycoside-adenylyltransferase.

Nine (20.0%) of the 45 isolates harbored genes encoding 16S rRNA methylase, including RmtB4, RmtD3, RmtE, and RmtF2 (Table 2). One isolate harbored genes encoding both RmtB and RmtE. In addition, the isolates harbored genes encoding various aminoglycoside-modifying enzymes, including AacA1, AacA4, AacA7, AacC5, AadA1, AadA2, AadA6, and AadB.

All 45 isolates were found to have point mutations in the quinolone resistance-determining regions of gyrA and/or parC, containing one to three amino acid substitutions. Of the 45 isolates, 35 (77.8%) had the amino acid substitutions Ser83Ile in GyrA and Ser80Leu in ParC, four (8.9%) had the substitutions Ser83Ile and Asp87His in GyrA and Ser80Leu in ParC, two (4.4%) had amino acid substitutions Ser83Ile and Asp87Asn in GyrA and Ser80Leu in ParC, two (4.4%) had the substitution Ser83Ile in GyrA, one (2.2%) had the substitutions Ser83Leu and Asp87Glu in GyrA and Ser80Leu in ParC, and one (2.2%) had the substitution Ser83Thr in GyrA (Table 2).

MLST and phylogenetic analysis.

Of the 45 MDR P. aeruginosa isolates in Myanmar, 23 (51.1%) belonged to sequence type 1047 (ST1047) (allelic profile 18, 8, 5, 5, 1, 6, 4). In addition, five (11.1%) isolates belonged to ST233, four (8.9%) isolates belonged to ST235, four (8.9%) isolates belonged to ST273, three (6.7%) isolates belonged to ST357, two (4.4%) isolates belonged to ST983, and one (2.2%) isolate each belonged to ST314, ST316, ST446, and ST1121 (Table 2). The phylogenetic tree revealed four clades designated clades I, II, III, and IV. Clade I consisted of the isolates belonging to ST1047, clade II of the isolates belonging to ST316, ST357, ST446, and ST1121, clade III of the isolates belonging to ST235, and clade IV of the isolates belonging to ST233, ST273, ST314, and ST549 (reference strain,P. aeruginosa PAO1) and ST983. The 23 strains of P. aeruginosa ST1047 (clade I) were isolated from hospitals A, B, and G and found to have spread in a clonal manner (Fig. 1).

FIG 1.

Molecular phylogenetic tree of the 45 MDR P. aeruginosa strains isolated in medical settings in Myanmar. A maximum likelihood phylogenetic tree was constructed from the 45 MDR isolates.

Genomic environments surrounding carbapenemase-encoding genes.

Sequences derived from each contig after assembling the raw read data showed that seven types of genetic structures surrounded the carbapenemase-encoding genes, including bla_DIM-1, bla_IMP-1, bla_NDM-1, bla_VIM-2, and bla_VIM-5 (Fig. 2). The environment surrounding bla_DIM-1 was intI1-bla_DIM-1-orf1-trpR-tnpA (type A in Fig. 2). The structure was a unique gene cassette array, although the structure, intI1-bla_IMP-1-aacA4, was frequently identified in IMP-type MBL-producing P. aeruginosa strains in Asian countries, including Japan and South Korea (1). The genomic environment surrounding bla_NDM-1 was IS91-bla_NDM-1-IS91-mrsB-mrsA-orf2-corA-orf3-tnpA (type C in Fig. 2). The structure IS91-bla_NDM-1-IS91 was identical to that in a chromosome of P. aeruginosa N15-01092, which was identified in 2015 in Canada (GenBank accession no. CP012901). The genomic environment surrounding bla_IMP-1, intI1-bla_IMP-1-aacA4-IS1595, was unique (type B in Fig. 2). The genomic environment surrounding bla_VIM-2 consisted of three types of genetic structures (types D, E, and F in Fig. 2). The structure D, intI1-aacA7-bla_VIM-2-dfrB-aacC5-tniR-tniQ-tniB-tniA, was identical to that of the chromosome of P. aeruginosa K34-7 identified in 2006 in Norway (GenBank accession no. CP029707) (14) and PA83 in 2013 in Germany (GenBank accession no. CP017293). The structure E, intI1-orf4-bla_VIM-2-tniC-tniQ-tniB-tniA, was identical to that in the plasmid of Pseudomonas putida PPV2-2 (GenBank accession no. GQ227991) identified in 2005 in Spain (15). The structure F, intI1-qnrVC1-aacA4-bla_VIM-2-tniC-tniQ-tniB-tniA, was a unique gene cassette array, although part of this structure, aacA4-bla_VIM-2-tniC-tniQ-tniB-tniA, was identical to that in the transposon Tn5090 of P. aeruginosa R22 in China (GenBank accession no. AM993098). The genetic environment surrounding bla_VIM-5 was intI1-bla_VIM-5-aadB-dfrA1-orf5-qacEΔ1-sulI. The structure, intI1-bla_VIM-5-aadB, was identical to that in the plasmid of VIM-1-producing P. aeruginosa strain PAMBL-1 (GenBank accession no. GQ422829) in 2007 in Spain (16).

FIG 2.

Structures of the genomic environments surrounding carbapenemase-encoding genes, including bla_DIM-1 (A), bla_IMP-1 (B), bla_NDM-1 (C), bla_VIM-2 (D to F), and bla_VIM-5 (G). orf1, haloacid dehalogenase type II-encoding gene; orf2, glutathione S-transferase-encoding gene; orf3, hypothetical protein-encoding gene; orf4, putative AAC(6′)-encoding gene; orf5, hypothetical protein-encoding gene.

Pulsed-field gel electrophoresis (PFGE) and Southern hybridization analyses revealed that 8 of 11 isolates harbored plasmids ranging in the sizes from 48 to 145 kbp (see Table S1 in the supplemental material). None of these isolates contained a plasmid harboring bla_DIM-1, bla_IMP-1, bla_NDM-1, bla_VIM-2, and/or bla_VIM-5 (data not shown), indicating that these genes were located on the chromosomes.

DISCUSSION

To our knowledge, this is the first molecular epidemiological analysis of MDR P. aeruginosa clinical isolates in Myanmar. Of the MDR P. aeruginosa isolates tested in Myanmar, 84.4% possessed genes encoding carbapenemases, including DIM-1, IMP-1, NDM-1, VIM-2, and VIM-5, and 20% of the isolates possessed genes encoding 16S rRNA methylases, including RmtB, RmtD3, RmtE, and RmtF2. These genes likely contributed to the high resistance of P. aeruginosa isolates in Myanmar to carbapenems and aminoglycosides.

Another key finding of this study was that the majority (51.1%) of MDR P. aeruginosa clinical isolates in Myanmar belonged to ST1047. To date, only one P. aeruginosa isolate belonging to ST1047 had been registered in the MLST website. This isolate, identified in Norway in 2010, was found to produce VIM-type MBLs. Recently, VIM-2-producing P. aeruginosa ST1047 isolates were obtained from dogs in South Korea with pyoderma and otitis (17). P. aeruginosa high-risk clones, including ST233, ST235, and ST357, were also identified in Myanmar but from only 26.7% of the 45 isolates. Further studies are needed to determine whether MDR P. aeruginosa isolates belonging to ST1047 are emerging and spreading in countries neighboring Myanmar, including Bangladesh, China, Laos, India, and Thailand. These results indicate that MDR P. aeruginosa strains may evolve in unique environments in Myanmar, such as the climate and the antibiotic usage, because ST1047 strains, which had completely different allelic profiles from those of ST235 strains, were spreading in medical settings in Myanmar.

P. aeruginosa ST1047 strains may have the ability to integrate various drug resistance factors, including carbapenemases, 16S rRNA methylases, and aminoglycoside-modifying enzymes, suggesting that P. aeruginosa ST1047 may become a high-risk clone in Southeast Asia. Most P. aeruginosa isolates obtained to date from medical settings in Southeast Asian countries belong to the high-risk clone ST235, but some isolates belonging to the high-risk clones ST111 and/or ST175 have been obtained from medical settings in East and South Asian countries, including China, India, Japan, and South Korea (1).

Our study strongly suggests that P. aeruginosa ST1047 has been spreading throughout medical settings in Myanmar. It is necessary to survey MDR P. aeruginosa isolates in medical settings in Myanmar.

MATERIALS AND METHODS

Bacterial isolates.

Forty-five isolates of MDR P. aeruginosa, defined as strains showing resistance to carbapenem (MIC, ≥8 μg/ml), amikacin (MIC, ≥64 μg/ml), and fluoroquinolones (MIC, ≥4 μg/ml) (18), were obtained between December 2015 and May 2017 from patients treated at seven hospitals in Myanmar (26 isolates from hospital A, eight from hospital B, four from hospital C, three from hospital D, two from hospital E, and one each from hospitals F and G). Bacteria were identified using the Vitek 2 system (bioMérieux, Marcy l’Etoile, France), with identities confirmed by sequencing of 16S rRNA (19). Of the 45 isolates, 18 isolates were from wounds, 14 isolates were from urine, six isolates were from sputum, three isolates were from pus, and one isolate each was from blood, bone pus, an endotracheal tube, and a nasopharyngeal swab. MICs were determined using the broth microdilution method (18).

Whole-genome sequencing.

Genomic DNAs of the 45 isolates were extracted using DNeasy blood and tissue kits (Qiagen, Tokyo, Japan) and sequenced using the MiSeq platform (Illumina, San Diego, CA) with the Nextera XT DNA library prep kit and MiSeq reagent kit version 3 (600 cycle; Illumina). More than 31-fold coverage was achieved for each isolate. Raw reads of each isolate were assembled using CLC Genomics Workbench version 8.0.2, and drug resistance genes were identified using ResFinder 3.0 (https://cge.cbs.dtu.dk//services/ResFinder/). The criteria for ResFinder 3.0 were defined as select threshold for identification (ID) of 90% and select minimum length of 80%. Fluoroquinolone resistance has been associated with mutations in the quinolone resistance-determining region, which includes the gyrA and parC genes that encode DNA gyrase and topoisomerase IV, respectively (20). Multilocus sequence typing (MLST) was deduced, as described by the protocols of the PubMLST (http://pubmlst.org/paeruginosa/) databases.

Phylogenetic analysis based on SNPs.

To identify single nucleotide polymorphisms (SNPs) among the 45 whole genomes, all reads of each isolate were aligned against the P. aeruginosa PAO1 sequence (GenBank accession no. AE004091) using CLC Genomics Workbench version 8.0.2 (CLC bio, Tokyo, Japan). SNP concatenated sequences were aligned by MAFFT (http://mafft.cbrc.jp/alignment/server/). Models and parameters used for the phylogenetic analyses were computed using j-Model Test-2.1.4. A maximum likelihood phylogenetic tree was constructed from SNP alignment with PhyML 3.0 (21).

Pulsed-field gel electrophoresis and Southern hybridization.

Plasmids from each ST strain (11 isolates), which harbored the bla_DIM-1, bla_IMP-1, bla_NDM-1, bla_VIM-2, and/or bla_VIM-5 genes, were extracted (Table S1) and separated by pulsed-field gel electrophoresis (22). Southern hybridization was performed using probes of each of the above-mentioned genes. Signal detections were carried out using DIG High Prime DNA labeling and detection starter kit II (Roche Applied Science, Indianapolis, IN, USA).

Data availability.

The whole-genome sequences of all 45 isolates were deposited at GenBank as accession number DRA007442 (experiment accession numbers DRX143958 to DRX144002 and run numbers DRR153275 to DRR153319).