Prevalence of Smqnr and plasmid-mediated quinolone resistance determinants in clinical isolates of Stenotrophomonas maltophilia from Japan: novel variants of Smqnr.

Stenotrophomonas maltophilia is an important pathogen in healthcare-associated infections. S. maltophilia may contain Smqnr, a quinolone resistance gene encoding the pentapeptide repeat protein, which confers low-level quinolone resistance upon expression in a heterologous host. We investigated the prevalence of Smqnr and plasmid-mediated quinolone resistance (PMQR) determinants in S. maltophilia isolates from Japan. A total of 181 consecutive and nonduplicate clinical isolates of S. maltophilia were collected from four areas of Japan. The antimicrobial susceptibility profiles for these strains were determined. PCR was conducted for Smqnr and PMQR genes, including qnrA, qnrB, qnrC, qnrS, aac(6')-Ib and qepA. PCR products for Smqnr and aac(6')-Ib were sequenced. For the S. maltophilia isolates containing Smqnr, pulsed-field gel electrophoresis (PFGE) was performed using XbaI. Resistance rates to ceftazidime, levofloxacin, trimethoprim-sulfamethoxazole, chloramphenicol and minocycline were 67.4%, 6.1%, 17.7%, 8.8% and 0%, respectively. The minimum inhibitory concentration required to inhibit the growth of 50% and 90% of organisms were 0.5 and 2 mg/L for moxifloxacin but 1 and 4 mg/L for levofloxacin, respectively. Smqnr was detected in 104 of the 181 S. maltophilia isolates (57.5%), and the most frequent was Smqnr6, followed by Smqnr8 and Smqnr11. Eleven novel variants from Smqnr48 to Smqnr58 were detected. The 24 Smqnr-containing S. maltophilia isolates were typed by PFGE and divided into 21 unique types. Nine S. maltophilia isolates (5.0%) carried aac(6')-Ib-cr. No qnr or qepA genes were detected. This study describes a high prevalence of Smqnr and novel variants of Smqnr among S. maltophilia from Japan. Continuous antimicrobial surveillance and further molecular epidemiological studies on quinolone resistance in S. maltophilia are needed.

Introduction

In recent years, hospital infections and outbreaks due to glucose nonfermenting Gram-negative bacilli such as the multidrug-resistant Pseudomonas aeruginosa and Acinetobacter baumannii have become a major concern worldwide. In hospital environments, Stenotrophomonas maltophilia has been isolated from tap water, nebulizers, ventilator circuits, contaminated disinfectants and other locations, and it is an important opportunistic pathogen causing healthcare-associated infections [1]. Infections due to S. maltophilia are increasing as a result of developments in advanced medical treatments such as organ transplantation, chemotherapy, immunosuppressant drugs, increased numbers of compromised hosts and the increased use of broad-spectrum antibiotics. S. maltophilia causes various infections, including respiratory tract infections (pneumonia and acute exacerbations of chronic obstructive pulmonary disease); bacteremia; and infections of urinary tract, skin and soft tissues, bones and joints and eye [2]. Nosocomial infections have high mortality, which has been reported to range 12.5% to 41% due to bacteremia, 40% to 50% due to pneumonia and 39% due to endocarditis [1,3]. S. maltophilia infections are difficult to treat because the bacterium may exhibit high-level intrinsic resistance to many antimicrobial agents. Trimethoprim–sulfamethoxazole, ticarcillin/clavulanate, and quinolones are used for treatment of S. maltophilia infections, but resistance to these antibiotics has been increasingly described [3]. Few studies on the antimicrobial susceptibility of S. maltophilia have been reported in Japan.

Quinolone resistance is mainly caused by chromosomal mutations of the quinolone resistance–determining regions in DNA gyrase and DNA topoisomerase IV, as well as porin alterations or overexpression of efflux pump systems, resulting in decreased quinolone accumulation [4-6]. However, a plasmid-mediated horizontally transferable quinolone resistance gene encoding a pentapeptide repeat protein (qnr) was discovered in 1998 [7]. Since then, other plasmid-mediated quinolone resistance (PMQR) determinants, including aac(6′)-Ib-cr, encoding a variant aminoglycoside acetyltransferase, and qepA, encoding an efflux pump, have been reported worldwide in clinical isolates of Enterobacteriaceae [5,6]. Studies of PMQR determinants in Enterobacteriaceae have been increasing, though PMQR determinants in S. maltophilia remain unclear.

The main mechanism of high-level quinolone resistance in S. maltophilia is thought to involve overexpression of efflux pumps, in particular SmeDEF, and low permeability of the outer membrane [8]. Quinolone resistance in S. maltophilia can also occur as a result of mutations of DNA gyrase and topoisomerase genes [2]. Smqnr have been recently identified in S. maltophilia as a novel chromosomal quinolone resistance gene encoding the pentapeptide repeat protein SmQnr. Similar to PMQR determinants in Enterobacteriaceae, Smqnr confers low-level quinolone resistance in Escherichia coli and may transmit quinolone resistance genes to other bacteria [9-11]. There are few studies on the frequency of Smqnr and PMQR genes in S. maltophilia. In this study, we investigated the antimicrobial susceptibility of clinical isolates of S. maltophilia and the prevalence of Smqnr and PMQR determinants in S. maltophilia isolates from Japan.

Materials and Methods

Bacterial isolates

A total of 181 consecutive and nonduplicate clinical isolates of S. maltophilia were collected between November 2009 and March 2010 from four areas of Japan, including 43 isolates from Hokkaido, 49 from Tokyo, 43 from Osaka and 46 from Fukuoka. S. maltophilia were identified using the Microscan Walkaway System (Siemens Healthcare Diagnostics, Tokyo, Japan) and the Vitek-2 System (Sysmex-bioMérieux, Tokyo, Japan). Isolates included 148 from sputum (81.8%), 13 from nasal and throat swabs (7.2%), eight from urine (4.4%) and 12 from other specimens (6.6%). There were no isolates obtained from blood cultures in this study. The isolates included 160 (88.4%) from hospitalized patients and 21 (11.6%) from community patients. Institutional review board approval was not required for this study according to the research guidelines of the Japanese government (http://www.niph.go.jp/wadai/ekigakurinri/guidelines.pdf).

Antimicrobial susceptibility testing

Minimum inhibitory concentrations (MICs) of nine antibiotics, including piperacillin (range, 0.06–128 mg/L), ceftazidime (0.06–128 mg/L), imipenem (0.06–128 mg/L), minocycline (0.06–128 mg/L), amikacin (0.06–128 mg/L), levofloxacin (0.06–128 mg/L), moxifloxacin (0.06–128 mg/L), trimethoprim–sulfamethoxazole (1/19–8/152 mg/L) and chloramphenicol (2–128 mg/L) were determined by the broth microdilution method. Clinical and Laboratory Standards Institute (CLSI) interpretive criteria were used for each antibacterial agent [12]. Quality control was performed using E. coli American Type Culture Collection (ATCC) 25922 and E. coli ATCC 35218.

PCR and sequencing of Smqnr and PMQR genes

The primers used to amplify an 811 bp fragment containing the entire Smqnr coding sequence were as follows: 5′-ACACAGAACGGCTGGACTGC-3′ and 5′-TTCAACGACGTGGAGCTGT-3′ [9,11]. Screening of S. maltophilia isolates for aac(6′)-Ib and PMQR genes, including qnrA, qnrB, qnrC, qnrS and qepA, was conducted by multiplex PCR amplification, as described previously [13]. PCR products for Smqnr and aac(6′)-Ib were sequenced using ABI BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) on an ABI 3730xl Analyzer (Applied Biosystems). Nucleotide and amino acid sequences were analysed with BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi), Expasy translate tool (http://web.expasy.org/translate/) and CLUSTALW2.1 (http://www.ddbj.nig.ac.jp/searches-j.html). If one or more amino acid differed compared to the previously reported Smqnr sequences in GenBank and/or published articles, Smqnr alleles were considered to be novel and were named numerically according to qnr nomenclature [14].

Pulsed-field gel electrophoresis

For the 24 S. maltophilia isolates containing Smqnr variants which were detected from four Japanese areas, pulsed-field gel electrophoresis (PFGE) was performed using XbaI (Takara Bio Inc., Otsu, Japan). Electrophoresis was performed on a 1% PFGE agarose gel with a CHEF-DR III system (Bio-Rad Laboratories, Richmond, CA, USA). Electrophoretic patterns were analysed with GelCompar II version 3.0 (Applied Maths, Kortrijik, Belgium). Similarity between two tracks was calculated using the coefficient of Jaccard and band positions. Cluster analysis and dendrogram generation were performed by the unweighted pair group method using arithmetic averages. Isolates with similarity >85% were considered to be within a cluster [15].

Statistical analysis

Statistical significance was evaluated by two-sided chi-square test or Fisher’s test, according to the sample size. A p value of <0.05 was considered significant.

Results

Antimicrobial susceptibility of S. maltophilia isolates

MICs and susceptibility profile of S. maltophilia for the tested antimicrobial agents are shown in Table 1. Resistance rates to ceftazidime, levofloxacin, trimethoprim–sulfamethoxazole and chloramphenicol were 67.4%, 6.1%, 17.7% and 8.8%, respectively. All isolates were susceptible to minocycline. The MIC required to inhibit the growth of 50% of organisms (MIC50) of piperacillin, imipenem and amikacin was 128 mg/L, >128 mg/L and 128 mg/L, respectively. The MIC50 and MIC required to inhibit the growth of 90% of organisms (MIC90) of moxifloxacin (0.5 and 2 mg/L, respectively) were lower than those of levofloxacin (1 and 4 mg/L, respectively). Fig. 1 shows a comparison of susceptibility to levofloxacin and moxifloxacin among S. maltophilia isolates with or without Smqnr genes. Eight (7.7%) and three (2.9%) of the 104 Smqnr-positive isolates were resistant to levofloxacin and moxifloxacin using the resistance breakpoint of MICs ≥8 mg/L, while three (3.9%) and three (3.9%) of the 77 Smqnr-negative isolates were resistant to levofloxacin and moxifloxacin, respectively (p >0.05) for both levofloxacin and moxifloxacin between Smqnr-positive and -negative isolates). However, 41 (39.4%) and 20 (19.2%) of the 104 Smqnr-positive isolates were resistant to levofloxacin and moxifloxacin using the resistance breakpoint of MICs ≥2 mg/L, while 17 (22.1%) and seven (9.1%) of the 77 Smqnr-negative isolates were resistant to levofloxacin and moxifloxacin, respectively (p 0.0134 for levofloxacin and p >0.05 for moxifloxacin between Smqnr-positive and -negative isolates).

Table 1

Minimum inhibitory concentrations (MICs) and susceptibility profiles of Stenotrophomonas maltophilia for the tested antimicrobial agents

Antibiotic	MIC range (mg/L)	MIC50 (mg/L)	MIC90 (mg/L)	Resistance rate (%)
Piperacillin	0.06–128	128	>128
Ceftazidime	0.06–128	64	128	67.4
Imipenem	0.06–128	>128	>128
Minocycline	0.06–128	0.25	1	0
Amikacin	0.06–128	128	>128
Levofloxacin	0.06–128	1	4	6.1
Moxifloxacin	0.06–128	0.5	2
Trimethoprim–sulfamethoxazole	1/19–8/152	1/19	8/152	17.7
Chloramphenicol	2–128	8	16	8.8

MIC50, minimum inhibitory concentration required to inhibit the growth of 50% of organisms; MIC90, minimum inhibitory concentration required to inhibit the growth of 90% of organisms.

Fig. 1

Comparison of levofloxacin (A) and moxifloxacin (B) susceptibility among Stenotrophomonas maltophilia isolates with or without Smqnr genes. Susceptible, ≤2 mg/L; intermediate, 4 mg/L; resistant, ≥8 mg/L.

Prevalence of Smqnr and PMQR genes in S. maltophilia

The distribution of Smqnr and PMQR genes for S. maltophilia is summarized in Table 2. Smqnr was detected in Japan at high frequencies in clinical isolates of S. maltophilia (57.5%; 104/181 isolates). The Smqnr frequency by region was 69.8% (30/43) in Hokkaido, 53.1% (26/49) in Tokyo, 46.5% (20/43) in Osaka and 60.9% (28/46) in Fukuoka. The most frequent was Smqnr6 (27.9%, 29/104), followed by Smqnr8 (19.2%, 20/104) and Smqnr11 (13.5%, 14/104). PMQR determinants were detected in nine S. maltophilia isolates (5.0%). No S. maltophilia isolates carried qnrA, qnrB, qnrC, qnrS or qepA. Nine (5.0%) were positive for aac(6′)-Ib, and all of them possessed the -cr variant.

Table 2

Distribution of Smqnr and plasmid-mediated quinolone resistance (PMQR) genes in 181 clinical isolates of Stenotrophomonas maltophilia

Gene	n	%
Smqnr
Smqnr2	5	2.8
Smqnr4	2	1.1
Smqnr6	29	16.0
Smqnr8	20	11.0
Smqnr9	3	1.7
Smqnr11	14	7.7
Smqnr18	1	0.6
Smqnr35	4	2.2
Smqnr37	5	2.8
Smqnr41	4	2.2
Smqnr48	1	0.6
Smqnr49	2	1.1
Smqnr50	1	0.6
Smqnr51	1	0.6
Smqnr52	6	3.3
Smqnr53	1	0.6
Smqnr54	1	0.6
Smqnr55	1	0.6
Smqnr56	1	0.6
Smqnr57	1	0.6
Smqnr58	1	0.6
Total	104	57.5
PMQR genes
qnrA	0	0
qnrB	0	0
qnrC	0	0
qnrS	0	0
qepA	0	0
aac(6′)-Ib-cr	9	5.0
Total	9	5.0

PFGE typing of Smqnr-containing S. maltophilia isolates

Dendrogram and PFGE of XbaI-digested genomic DNA from 24 S. maltophilia with Smqnr are shown in Fig. 2, including 11 isolates with Smqnr6 and 13 isolates with Smqnr11. The 24 Smqnr-containing S. maltophilia isolates were divided into 21 unique PFGE types. Of the 11 Smqnr6-containing S. maltophilia isolates, five isolates were novel, while two isolates from Osaka and Fukuoka, Tokyo and Osaka, or Fukuoka only clustered together (SM1, SM5 or SM6). All 13 isolates possessing Smqnr11 were unique.

Fig. 2

Dendrogram and pulsed-field gel electrophoresis of XbaI-digested genomic DNA from 24 Smqnr-containing Stenotrophomonas maltophilia isolates. A, Hokkaido; D, Tokyo; F, Osaka; I, Fukuoka; M, marker.

Novel variants of Smqnr and comparison with previous variants

In this study, 11 novel variants of Smqnr were discovered, and the following nucleotide accession numbers were assigned in the EMBL/GenBank/DDBJ database: Smqnr48,AB852568; Smqnr49,AB852569; Smqnr50,AB852570; Smqnr51,AB852571; Smqnr52,AB852572; Smqnr53,AB852573; Smqnr54,AB905279; Smqnr55,AB905280; Smqnr56,AB905281; Smqnr57,AB905282; and Smqnr58,AB905283. The novel variants were compared to those reported previously: Smqnr1–Smqnr11 (accession numbers AB430839–AB430843, AB430845–AB430850); Smqnr12–Smqnr18 (accession numbers FJ596752–FJ596758); Smqnr19–Smqnr27 (accession numbers EU681371–EU681373, EU681375, EU681378, EU681381–EU681382, EU681384–EU681385); Smqnr28–Smqnr47 (accession numbers HQ874463–HQ874464, HQ896258–HQ896275). A phylogenetic tree showing the relationship from Smqnr1 to Smqnr58 is shown in Fig. 3. Differences in amino acid sequences between novel variants in this study and previous variants are as follows: SmQnr48 (55S→T, 194T→A of SmQnr36 ), SmQnr49 (18D→G of SmQnr30), SmQnr50 (119A→V of SmQnr35), SmQnr51 (20T→H of SmQnr9), SmQnr52 (72R→H, 203V→M of SmQnr11), SmQnr53 (67K→E, 146W→G of SmQnr37), SmQnr54 (119A→E, 169V→I of SmQnr11), SmQnr55 (82S→N, 115G→W, 194A→T of SmQnr38), SmQnr56 (6S→F, 57T→A, 82S→N of SmQnr39), SmQnr57 (105S→N, 169V→I of SmQnr11), and SmQnr58 (6S→F, 57T→A, 82S→N, 205Q→X of SmQnr39).

Fig. 3

Phylogenetic analysis of the relationship of SmQnr1–58. The phylogram was provided by the neighbor-joining method using ClustalW2.1.

Discussion

Global infectious disease surveillance indicates that resistance rates for trimethoprim–sulfamethoxazole, ticarcillin–clavulanic acid, ciprofloxacin, levofloxacin and minocycline in S. maltophilia isolates are less than 4.7%, 16.1%, 40%, 6.5% and 5%, respectively [16]. However, large differences in the susceptibility of S. maltophilia were seen, depending on the country and/or region. In our study, resistance to trimethoprim–sulfamethoxazole was approximately 18% in Japan, while it has been reported to be 8% in the Asia-Pacific region, 10% in Europe, 15% in Turkey, 25% in Taiwan and 27% in Spain [3]. The resistance rate to levofloxacin in this study was 6.1%, which was the same as seen in a previous study in China [17]. Although the percentage of strains resistant to quinolones ranged from 13–96% for ciprofloxacin, 3–11% for levofloxacin, 0–46% for gatifloxacin, 6% for moxifloxacin, 56% for norfloxacin and 5% for ofloxacin [3], the levofloxacin resistance rate in this study showed a higher tendency in a MIC scale than to moxifloxacin. Furthermore, our resistance rate using a MIC of ≥2 mg/L for levofloxacin among Smqnr-positive isolates was significantly higher than that in Smqnr-negative isolates; however, there was no significant difference in levofloxacin resistance rate using a MIC of ≥8 mg/L (the same as the CLSI criteria) between isolates with or without Smqnr, although the number of sample strains tested in this study was small. Little association was seen between Smqnr presence and high-level quinolone resistance in S. maltophilia, though Smqnr can provide low-level quinolone resistance, as described in previous studies [9-11,17]. The role of Smqnr remains unclear, but the reduced susceptibility may be associated with the selection of mutants having higher resistance, as described in PMQR determinants [7].

In this study, we identified 21 different PFGE types among the 24 S. maltophilia Smqnr-containing isolates in Japan. These results suggest that S. maltophilia Smqnr-containing isolates in Japan are not a single clone but rather are multiple clones. Superinfection with S. maltophilia is common, with a reported frequency of 20–40% [3,18], and S. maltophilia may be a reservoir for the dissemination of quinolone resistance to other bacteria. Thus, it is necessary to monitor Smqnr genes in Enterobacteriaceae and other organisms as well as in S. maltophilia.

Smqnr, a chromosome-carried quinolone resistance gene in S. maltophilia, was first reported by Shimizu and colleagues in 2008 [9]. There have been several experimental and epidemiologic studies regarding the mechanism and distribution of Smqnr in clinical isolates of S. maltophilia[9-11,17,19,20]. SmQnr1–11 proteins with 2–6 amino acid alterations were initially identified in Japan [9], and subsequent SmQnr variants were described as follows: SmQnr12–18 in UK [11], SmQnr19–27 in Spain [10], SmQnr28–47 in China [17] and SmQnr48–58 in Japan (this study). All 11 novel Smqnr variants in our study were derived from Smqnr genes reported previously from Japan and China. Our study suggests a high prevalence of Smqnr in S. maltophilia from Japan (Table 2), and there were regional differences of 47–70% in Smqnr frequency. The most common Smqnr gene in our study was Smqnr6, followed by Smqnr8 and Smqnr11, while Smqnr8 and Smqnr35 were most common in China. Further molecular epidemiological studies investigating the prevalence of Smqnr variants in other countries are needed.

Unlike Enterobacteriaceae, PMQR determinants in S. maltophilia have not been fully clarified. A previous study in China showed that PMQR determinants, including qnrA, qnrB, qnrC, qnrD and qnrS, were not detected in S. maltophilia[17]. In this investigation, we report the detection of aac(6′)-Ib-cr in S. maltophilia, but no qnr or qepA genes were seen. The aac(6′)-Ib-cr gene, which confers resistance to both quinolones and aminoglycosides, has often been detected in CTX-M-producing Enterobacteriaceae [6,21] and is a concern for facilitating further multidrug resistance in S. maltophilia.

In conclusion, our study revealed a high prevalence of Smqnr among S. maltophilia from Japan and described novel variants of Smqnr. We also investigated the prevalence of PMQR determinants and demonstrated the presence of aac(6′)-Ib-cr genes in clinical isolates of S. maltophilia. Continuous antimicrobial surveillance for S. maltophilia and further investigations into quinolone resistance mechanisms in S. maltophilia are required.

Conflict of interest

None declared.