Coordinate Hyperproduction of SmeZ and SmeJK Efflux Pumps Extends Drug Resistance in Stenotrophomonas maltophilia

Abstract

A Stenotrophomonas maltophilia mutant that coordinately hyper-expresses three resistance nodulation division-type efflux pump genes, smeZ, smeJ, and smeK, has been identified. SmeZ is responsible for elevating aminoglycoside MICs; SmeJ and SmeK are jointly responsible for elevating tetracycline, minocycline, and ciprofloxacin MICs and conferring levofloxacin resistance. One clinical isolate with this same phenotype was identified from a sample of six, and the isolate also coordinately hyper-expresses smeZ and smeJK, confirming the clinical relevance of our findings.

TEXT

Stenotrophomonas maltophilia is an emerging nosocomial pathogen that is intrinsically resistant to many antimicrobial drugs and causes respiratory tract and bloodstream infections in patients who are immunocompromised or severely debilitated. In addition, it can cause wound, burn, and other soft tissue infections and ocular infections, and it is an increasingly common colonizer of the lungs of cystic fibrosis patients (1, 2). Intrinsic β-lactam resistance (including to carbapenems) in S. maltophilia is due to two chromosomally encoded β-lactamase enzymes, L1 and L2, which are produced at elevated levels (induced) during β-lactam challenge through an AmpR-mediated mechanism (3, 4). Intrinsic aminoglycoside resistance involves the modifying enzymes AAC(6′)-Iz (5, 6) and APH(3′)-IIc (7) and a resistance nodulation division (RND)-type efflux pump, SmeZ (8). SmeJ and SmeK, which are encoded in an operon and appear to form a single RND pump, also have a role in intrinsic drug resistance (8), as does SmeE (9).

As well as having a role in intrinsic resistance when expressed at wild-type levels, SmeE, when overproduced, causes fluoroquinolone resistance (10). The fluoroquinolone levofloxacin is one of the few drugs that are generally efficacious against S. maltophilia, so fluoroquinolone resistance mechanisms are particularly important to characterize (11). To this end, CLSI agar dilution MICs (12) of a cross-section of antimicrobials were determined against S. maltophilia K279a, the prototype S. maltophilia genome sequence strain, and a previously selected mutant derivative, named K M5, which has reduced fluoroquinolone susceptibility but does not overexpress smeE (13). β-Lactam MICs against K279a and K M5 where identical. Blocking β-lactamase production, which might otherwise mask an ancillary β-lactam resistance mechanism, by disruption of ampR using the method described previously (4) reduced β-lactam MICs to identical levels in K279a and K M5 (Table 1). Therefore, the resistance mechanism activated in K M5 does not affect β-lactam susceptibility. Elevated MICs against K M5 were seen for all aminoglycosides and tetracyclines tested and for ciprofloxacin (Table 1). Importantly, K M5 gained resistance to levofloxacin, according to CLSI breakpoints (14).

Table 1.

MICs of antimicrobials against S. maltophilia clinical isolate K279a and mutant derivatives^a

Isolate or mutant	MIC (μg/ml) of:
	TET	MIN	SXT	CIP	LVX	AMK	GEN	TZP	CAZ	ATM	IPM
K279a	16	0.25 (S)	1 (S)	2	2 (S)	64	16	256	8 (S)	256	256
K M5	64	1 (S)	1 (S)	32	8 (R)	512	128	256	8 (S)	256	256
K279a ampRFS	16	0.25 (S)	1 (S)	2	2 (S)	64	16	32	0.25 (S)	16	0.5
K M5 ampRFS	64	1 (S)	1 (S)	32	8 (R)	512	128	32	0.25 (S)	16	0.5
K M5 smeZFS	64	1 (S)	1 (S)	32	8 (R)	32	1	256	8 (S)	256	256
K M5 smeJFS	16	0.125 (S)	1 (S)	0.5	2 (S)	64	16	256	8 (S)	256	256
K M5 smeKFS	16	0.125 (S)	1 (S)	0.5	2 (S)	64	16	256	8 (S)	256	256
K M5 smeZJKFS	16	0.125 (S)	1 (S)	0.5	2 (S)	32	1	256	8 (S)	256	256

^aAbbreviations: TET, tetracycline; MIN, minocycline; SXT, trimethoprim-sulfamethoxazole (the concentration given is for trimethoprim; the sulfamethoxazole concentration was 19 times higher); CIP, ciprofloxacin; LVX, levofloxacin; AMK, amikacin; GEN, gentamicin; TZP, piperacillin-tazobactam (the concentration given is for piperacillin; the tazobactam concentration was 4 μg/ml); CAZ, ceftazidime; ATM, aztreonam; IPM, imipenem. Where a resistance breakpoint has been set for S. maltophilia (14), a designation for susceptible (S) or resistant (R) is given. The FS gene superscript indicates that the gene carries a frameshift mutation.

In order to find out if any of the nine RND-type efflux pump genes previously identified from the S. maltophilia K279a genome sequence (8) is overexpressed in K M5 relative to expression of a gyrA control gene, one-step reverse transcription-PCR (RT-PCR) was performed as described previously (13) using the primers in Table 2. Both smeZ and the smeJK operon were markedly overexpressed in K M5 compared with levels in the parent isolate (Fig. 1A), but all of the other efflux pump genes tested were identically expressed in the two strains (not shown). To test whether their overexpression is responsible for extended drug resistance in K M5, smeZ, smeJ, and smeK were disrupted, separately or all together, by introducing frameshift mutations as previously described (8). Disruption of smeZ significantly lowered aminoglycoside MICs against K M5 but not non-aminoglycoside MICs (Table 1). Disruption of either smeJ or smeK reversed the acquisition of levofloxacin resistance in K M5 and reduced the MICs of tetracyclines and ciprofloxacin, but the effect on aminoglycoside MICs was limited (Table 1). To get full reversion of the extended resistance phenotype acquired by K M5, it was necessary to disrupt smeZ, smeJ, and smeK at the same time (Table 1), confirming that coordinate overproduction of SmeZ and SmeJK is required to acquire this phenotype. There is no putative regulatory gene close to the smeJK operon, but close to smeZ is a putative two-component system (TCS) operon (8). Sequencing of this TCS operon in K M5 following PCR amplification did not reveal any mutation, so more work needs to be done to test the role of this putative TCS in coordinate control of smeZ and smeJK expression.

Table 2.

RT-PCR primers used in this study

Primera	Sequence (5′ to 3′)
smeB+	TGGAAGCGCGCATCGATGAC
smeB−	GCCATGCGGCCGTACTTCTGC
smeE+	GCATCAACGTGGCCA
smeE−	TCAGCATGTTCACCGAGA
smeH+	GTGGTGGCGCTGCAGTCCGATAC
smeH−	CGCGGAAGTTCTGCAGGAAG
smeJK+	GACCTCGCAGACGCAGTCG
smeJK−	CAGGTAATCGCGCAGGGTC
smeN+	CCGAAGAAGCGCGTCCG
smeN−	GATCACCGCCAGCACCAGG
smeP+	CGTTGACCGGCAAGGGCTTC
smeP−	CTGCGGCTGGTCGGACAC
smeW+	GACACCCTGTACATGCGCAAG
smeW−	CAACGGGATGATCGAGGCAC
smeZ+	GATTCCGGCTTCCTGATGCTG
smeZ−	CGCGGTGTCGAACGGAATG
gyrA+	AACTCAACGCGCACAGCACAAGCC
gyrA−	CCGATTCCTTTTCGTCGTAGTTGGG

^a+, forward primer; −, reverse primer.

Fig 1.

RT-PCR for smeZ and smeJK RND efflux pump gene expression in S. maltophilia clinical isolates and mutant derivatives. RNA was extracted from isolate K279a and K M5 (A) or from six clinical isolates (B to D) and used for RT-PCR as set out previously (13) to quantify the amount of smeZ (B, C) and smeJK operon (D) mRNA expressed in each strain using the primers in Table 2. The RT-PCR was run for the standard 30 cycles (B) or for an additional 5 cycles (C) to confirm that isolate 2597 expresses very little smeZ. RT-PCR products were separated using 1%, wt/vol, agarose gels, stained with ethidium bromide, and visualized using UV transillumination on a Bio-Rad Gel Doc system (Hemel, Hempstead, United Kingdom). The images are photographs of these stained gels and are representative of three separate experiments, each using a different preparation of RNA.

S. maltophilia is a very heterogeneous species (3), and as such there are likely to be many confounding factors when trying to identify the function, in MIC terms, of a particular resistance mechanism in the species as a whole from studying the mechanism in a single isolate. However, MICs of antimicrobials against six clinical S. maltophilia isolates were correlated with their smeZ, smeJ, and smeK expression profiles. The isolates, originally from the SENTRY antimicrobial surveillance program (15), were chosen at random from a larger group of isolates that are phylogenetically related to K279a (3) specifically to reduce background variability. Only two of the six isolates are levofloxacin resistant (Table 3). Levofloxacin-resistant isolate 49-6147 expresses a phenotype known as resistance profile 1: high MICs of fluoroquinolones and tetracyclines but not aminoglycosides (Table 3). It is known to hyper-express the smeDEF efflux pump operon, which is the probable cause of levofloxacin resistance (13), and does not hyper-express smeJK or smeZ (Fig. 1B and D). Isolate 49-6147 is also ceftazidime resistant, and the MIC for it is very high (Table 3); the isolate is known to hyper-produce L1 and L2 β-lactamases (3), which is the probable cause. The other levofloxacin-resistant isolate, 9189, expresses a phenotype known as resistance profile 3, which is almost identical to that of K M5: very high aminoglycoside, fluoroquinolone, and tetracycline MICs (Table 3). Also like K M5, isolate 9189 overexpresses smeJK and smeZ (Fig. 1B and D) but not smeE (13).

Table 3.

MICs of antimicrobials against S. maltophilia clinical isolates

Isolate	Country of origin	MIC (μg/ml) ofa:											Comment
		TET	MIN	SXT	CIP	LVX	AMK	GEN	TZP	CAZ	APM	IPM
49-6147	Venezuela	64	8 (I)	2 (S)	32	16 (R)	64	32	>1,028	>256 (R)	>512	>512	Resistance profile 1b + L1/L2 hyper-producerc
9189	Chile	128	4 (S)	>32 (R)	8	8 (R)	512	512	512	4 (S)	256	256	Resistance profile 3b + sul1 expresserd
2597	Italy	16	0.25 (S)	>32 (R)	1	1 (S)	64	1	256	32 (R)	256	256	sul1 expresserd
489I	USA	128	16 (R)	>32 (R)	1	2 (S)	256	128	64	8 (S)	512	256	Resistance profile 2b + sul1 expresserd
1696	USA	8	0.25 (S)	1 (S)	2	2 (S)	64	16	>1,028	>256 (R)	>512	>512	Resistance profile 4b
4225	Belgium	16	0.25 (S)	>32 (R)	1	1 (S)	64	32	256	16 (I)	>512	256	sul1 expresserd
K279a	UK	16	0.25 (S)	1 (S)	2	2 (S)	64	16	256	8 (S)	256	256	Wild type

^aAbbreviations: TET, tetracycline; MIN, minocycline; SXT, trimethoprim-sulfamethoxazole (the concentration given is for trimethoprim; the sulfamethoxazole concentration was 19 times higher); CIP, ciprofloxacin; LVX, levofloxacin; AMK, amikacin; GEN, gentamicin; TZP, piperacillin-tazobactam (the concentration given is for piperacillin; the tazobactam concentration was 4 μg/ml); CAZ, ceftazidime; ATM, aztreonam; IPM, imipenem; S, sensitive; I, intermediate resistant; R, resistant (14).

^bAs defined in reference 13.

^cAs shown in reference 3.

^dAs shown in reference 15.

None of the levofloxacin-susceptible clinical isolates overexpress smeJK, smeZ (Fig. 1B and D), or smeE (13). One of these isolates, 489I, expresses a phenotype known as resistance profile 2, namely, elevated MICs of aminoglycosides and tetracyclines but not fluoroquinolones (Table 3), for which the mechanism is unknown. Isolate 1697 is interesting because it is ceftazidime resistant with a very high MIC (Table 3); however, because this isolate is known not to hyper-produce either the L1 or the L2 β-lactamase (3), its phenotype is characterized as resistance profile 4 (13), and again, the mechanism is not known. Isolate 2597 was noticeably more susceptible to aminoglycosides than the others (Table 3). RT-PCR revealed that this isolate had very little detectable expression of smeZ, even when large numbers of RT-PCR cycles were used (Fig. 1B andC). The effect of smeZ disruption in K279a is to reduce aminoglycoside MICs (8), so this lack of smeZ expression is likely to be the reason for the reduced MICs of aminoglycosides against isolate 9189.

In conclusion, we have defined the mechanism conferring extended-drug-resistance profile 3 in S. maltophilia as coordinate hyper-production of SmeZ and SmeJK. Resistance profile 3 is seen in clinical isolates and has clear clinical relevance, particularly when one considers that it can be selected by the use of aminoglycosides (K M5 was selected using a concentration of amikacin above the MIC [13]) but results in cells that are much less susceptible to fluoroquinolones. This may well have implications for therapeutic choice where S. maltophilia infection is suspected. It may also have implications for patients who have cystic fibrosis, since S. maltophilia is an increasingly common colonizer of their lungs (16) and they are regularly prescribed aminoglycosides to treat pulmonary exacerbations, often followed up with a fluoroquinolone (17).