Mutations in Ribosomal Protein RplA or Treatment with Ribosomal Acting Antibiotics Activates Production of Aminoglycoside Efflux Pump SmeYZ in Stenotrophomonas maltophilia

TEXT

Aminoglycoside resistance in the important opportunistic pathogen Stenotrophomonas maltophilia is multifactorial. Reduced amikacin and tobramycin susceptibility is conferred by a chromosomal aac(6′)-Iz gene, which is present in approximately 50% of clinical isolates (1, 2). The chromosomal aph(3′)-IIc gene is more widespread among S. maltophilia clinical isolates and is responsible for reduced kanamycin and neomycin susceptibility (3). A wide variety of other genes have been shown to play minor roles in intrinsic aminoglycoside susceptibility (4–9), but by far, the most significant inducible resistance mechanisms is the SmeYZ efflux system, where SmeZ is an RND-type efflux pump (10). Following mutation, smeYZ can be overexpressed, which leads to hyperresistance to all aminoglycosides, and, most importantly, smeYZ-overexpressing mutants have been seen in the clinic (10). Disruption of smeYZ also affects S. maltophilia virulence (11), suggesting a more pleotropic role in physiology, as is common for RND-type efflux systems (12).

It has been shown that smeYZ expression is controlled by a two-component regulatory system encoded immediately upstream, named SmeSyRy (13). One S. maltophilia aminoglycoside hyperresistant mutant shown to overexpress smeYZ is K M5 (10, 14). This mutant is a derivative of the clinical S. maltophilia isolate K279a (15, 16) and was selected for its ability to grow on amikacin at 4 times the MIC (14). In an attempt to identify the mutation responsible for smeYZ overexpression in K M5, we resurrected it from the freezer and first confirmed using CLSI disc susceptibility testing (17) that it expresses a “resistance profile 3” phenotype, as previously defined (10, 14): particularly, reduced zone diameters for aminoglycosides, quinolones, and tetracyclines (Table 1). Both SmeY and SmeZ were found to be approximately 8-fold upregulated in K M5 relative to expression in the parental strain, K279a, according to liquid chromatography tandem mass spectrometry (LC-MS/MS) envelope proteomics (Fig. 1), which was carried out as described previously (18). Whole-genome sequencing (WGS) was performed and analyzed as described before (19) and showed that smeSy and smeRy are wild type in K M5. Instead, there is only one difference from K279a, confirmed by PCR sequencing using the primers rplA F, 5′-GCGAAGGAACCGGATCTGA-3′, and rplA R, 5′-CGCCTGCGGTCTTTGAC-3′. The single point mutation in K M5 is predicted to cause a Gly67Asp change in the largest protein from the 50S ribosomal subunit, L1, encoded by the rplA gene. The previously described clinical isolate 9189, which also has the resistance profile 3 phenotype (Table 1) and overexpresses smeYZ (10), was found by PCR sequencing to also carry differences in rplA relative to that in K279a: predicted to cause a Phe22Leu change in RplA and a frameshift mutation at codon 212 caused by the insertion of a single adenine base. To confirm a role for the observed rplA mutation in reduced aminoglycoside susceptibility in K M5, wild-type rplA was amplified alongside the overlapping rplK by PCR using the primers rplAK F, 5′- AAAGCGGCCGCATCCAGCTGTAGAGTCGAGC-3′, and rplAK R, 5′- AAAGCGGCCGCCTGCGGTCTTTGACGGCTAC-3′, cut with NotI (introduced site underlined) and ligated into the broad host range vector pBBR1MCS (20), and the recombinant or empty vector was used to transform K M5 to chloramphenicol resistance (30 μg/ml). Using MIC and disc susceptibility testing, we confirmed that carriage of wild-type rplA by K M5 increased amikacin and gentamicin susceptibility relative to that of the plasmid-only control, though not to wild-type levels (Table 2, Fig. 2A). The other markers of resistance profile 3, reduced susceptibility to fluoroquinolones and tetracyclines, were not reversed by complementation of the rplA mutation in K M5, and we have previously shown that this part of the phenotype is due to the overexpression of a different pump operon, smeIJK (10). Surprisingly, in fact, fluoroquinolone and tetracycline susceptibility was reduced in K M5(pBBR1MCS::rplA) compared to that for the plasmid-only control (Fig. 2A). It is known that there is an inverse correlation between SmeYZ production and SmeDEF production in S. maltophilia (21), and indeed, proteomics confirmed that, as well as SmeY production (Fig. 2B) being reduced in K M5(pBBR1MCS::rplA) relative to that for the plasmid-only control (though not to wild-type levels, as seen for aminoglycoside MICs [Table 2]), production of SmeD, SmeE, and SmeF increased 17.3-fold, 11.2-fold, and 17.3-fold, respectively (P < 0.001, n = 3 for all). SmeDEF is a known tripartite efflux pump for fluoroquinolones and tetracyclines in S. maltophilia (22), explaining our findings (Fig. 2A).

• Open in new tab
• Download powerpoint

FIG 1

Production of SmeYZ in K M5. Protein abundance data for whole-envelope proteomics were normalized to the average 30S and 50S ribosomal protein abundance for each. Abundance of SmeY and SmeZ (UniProt B2FQ54 and B2FQ55) are reported as means ± standard errors of the means (SEMs), n = 3. Differences between K279a and K M5 were statistically significant (P < 0.05).

• Open in new tab
• Download powerpoint

FIG 2

Effect of complementing K M5 with rplA. (A) Disc susceptibility testing for antimicrobials against K M5 carrying pBBR1MCS alone or pBBR1MCS::rplA. Data are means ± SEMs, n = 3. LEV, levofloxacin, 5 μg disc; CIP, ciprofloxacin, 5 μg disc; MH, minocycline, 30 μg disc; TE, tetracycline, 30 μg disc; CN, gentamicin, 10 μg disc; AK, amikacin, 10 μg disc; SXT, sulfamethoxazole-trimethoprim, 25 μg disc. (B) Abundances of SmeY normalized to average ribosomal protein abundances based on proteomics analysis is reported as means ± SEMs, n = 3. The differences between K279a and K M5 and between K M5(pBBR1MCS) and K M5(pBBR1MCS::rplA) are statistically significant (P < 0.05).

In Pseudomonas aeruginosa, MexXY is considered the most important efflux pump involved in aminoglycoside resistance (23). It has been stated that at least two P. aeruginosa clinical isolates that hyperexpress mexXY have truncations in rplA, though the data were not presented and were reported as “unpublished” (24). Mutations in other ribosomal subunit genes have more conclusively been shown to cause mexXY overexpression (25). Given the similarities between MexXY and SmeYZ, this provides precedence for our experimental observation that rplA disruption is the cause of smeYZ overexpression in S. maltophilia.

Expression of mexXY in P. aeruginosa is inducible in response to ribosomal acting antibiotics (25, 26), and since ribosomal protein damage by mutation is associated with SmeYZ overproduction in S. maltophilia (Fig. 1), we next tested the inducibility of smeZ expression by the ribosomal acting antibiotics gentamicin and minocycline. Expression of smeZ was assessed by reverse transcriptase quantitative PCR (RT-qPCR) using RNA prepared from a nutrient broth (NB) culture of K279a exposed to gentamicin (32 μg/ml) or minocycline (0.5 μg/ml) versus untreated control. The method used was as described previously (18) with the primers smeZ RT F, 5- TGTCCAGCGTCAAGCACC-3, and smeZ RT R, 5- GCCGACCAGCATCAGGAAG-3. Abundance of smeZ was normalized to rrnB abundance (as an RNA loading control) using the primers rrnB F, 5- GACCTTGCGCGATTGAATG-3, and rrnB R, 5- CGGATCGTCGCCTTGGT-3. This experiment confirmed that, as predicted, the normalized expression of smeZ in the gentamicin- or minocycline-treated K279a cultures was significantly more (approximately 8-fold and 3-fold, respectively, P < 0.05, n=3 biological replicates, each with 4 technical replicates) than in the control (Fig. 3).

• Open in new tab
• Download powerpoint

FIG 3

RT-qPCR analysis of the effect of ribosomal acting antibiotics on smeZ expression in K279a. K279a cultures were diluted 1:100 from an overnight culture and each grown in NB for 3 h. The experimental group was grown in the presence of gentamicin (CN) (32 μg/ml) or minocycline (MH) (0.5 μg/ml). RNA was purified and the abundance of smeZ in each RNA preparation was assayed using RT-qPCR calculated using the threshold cycle (2^ΔΔCT) method (18) using the rrnB gene as an internal control for RNA abundance. Values for the smeZ/rrnB ratio from treated cells were normalized to those of the untreated control (NB). *, P < 0.05 versus control. There were three biological and, for each, four technical replicates.

Based on our findings, we therefore conclude that RplA damage in S. maltophilia constitutively mimics the effects of treatment with ribosomal acting antibiotics and constitutively activates SmeYZ production, raising aminoglycoside MICs. Our finding that rplA mutations exist in SmeYZ-overproducing clinical isolates confirms that such a mutation does not impair fitness or virulence so much that the mutants cannot survive, colonize, and cause infections in humans. Indeed, since reduced SmeYZ production reduces S. maltophilia virulence, at least in vitro and in a mouse model of infection (11), it may be that the advantage of rplA mutation in S. maltophilia is greater than simply raising aminoglycoside MICs. Given that S. maltophilia is frequently a colonizer of the lungs of cystic fibrosis patients, for which the aminoglycoside tobramycin is a regular therapy (27), the potential for selecting mutants with elevated SmeYZ production seems high.