QnrS1 structure–activity relationships

Abstract

Objectives

Loop B is important for low-level quinolone resistance conferred by Qnr proteins. The role of individual amino acids within QnrS1 loop B in quinolone resistance and gyrase protection was assessed.

Methods

qnrS1 and 11 qnrS1 alleles with site-directed Ala mutations in loop B were expressed in Escherichia coli BL21(DE3) and proteins were purified by affinity chromatography. Ciprofloxacin MICs were determined with and without IPTG. Gyrase DNA supercoiling was measured with and without ciprofloxacin IC₅₀ and with various concentrations of QnrS1 proteins.

Results

Wild-type QnrS1 and QnrS1 with Asn-110→Ala and Arg-111→Ala substitutions increased the ciprofloxacin MIC 12-fold in BL21(DE3), although QnrS1 with Gln-107→Ala replacement increased it 2-fold more than wild-type did. However, QnrS1 with Ala substitutions at His-106, Val-108, Ser-109, Met-112, Tyr-113, Phe-114, Cys-115 and Ser-116 increased ciprofloxacin MIC 1.4- to 8-fold less than wild-type QnrS1. Induction by 10–1000 μM IPTG increased ciprofloxacin MICs for all mutants, reaching values similar to those for wild-type. Purified wild-type and mutated proteins differed in protection of gyrase from ciprofloxacin action. Wild-type QnrS1 produced complete protection of gyrase supercoiling from ciprofloxacin (1.8 μM) action at 0.05 nM and half protection at 0.5 pM, whereas QnrS1 with Ala replacements that conferred the least increase in ciprofloxacin MICs also required the highest QnrS1 concentrations for protection.

Conclusions

Key individual residues in QnrS1 loop B affect ciprofloxacin resistance and gyrase protection from ciprofloxacin action, supporting the concept that loop B is key for interaction with gyrase necessary for quinolone resistance.

Introduction

Since the discovery of the first plasmid-encoded qnr from a resistant clinical isolate of Klebsiella pneumoniae,¹ a large number of qnr homologues have been found on both chromosomes and plasmids in Gram-negative and Gram-positive bacteria.^2,3 The qnr genes confer low-level resistance to quinolones, their gene products interact with DNA gyrase and topoisomerase IV to protect against quinolone toxicity and they are frequently associated with other resistance determinants on the same plasmid, thus structurally linking quinolone resistance to multidrug resistance.^2,4

DNA gyrase and topoisomerase IV are essential type II topoisomerases that utilize energy of ATP hydrolysis to alter DNA conformation.^5,6 Gyrase is the only Escherichia coli topoisomerase that can introduce negative supercoils into closed circular DNA.⁶ Quinolones act by binding to specific sites in the enzyme–DNA complexes in a manner that traps the complex and blocks DNA replication.⁷ Mutations in select domains of gyrase and topoisomerase IV confer resistance to quinolones and reduce drug affinity.²

The Qnr proteins were discovered to bind to and block quinolone inhibition of DNA gyrase and topoisomerase IV.^8–10 They belong to the pentapeptide repeat protein (PRP) family,² which has >500 members in the prokaryotic and eukaryotic kingdoms.^3,11 These proteins contain tandem 5 amino acid repeats with a consensus sequence of [S,T,A,V][D,N][L,F]-[S,T,R][G].^3,12–18 The biological functions of the vast majority of PRP family members are unknown.¹¹

The pentapeptide repeat units of nine representative Qnr proteins have been determined via a structure-based alignment approach and the structures of AhQnr of Aeromonas hydrophila,^18,19 EfsQnr of Enterococcus faecalis and plasmid-encoded QnrB1 have been determined directly.^15,17 The structure of the Qnr homologue MtMfpA, which is associated with reduced fluoroquinolone susceptibility in mycobacteria, has also been determined.¹⁴ Each of these four proteins folds as a right-handed quadrilateral β-helix that is characteristic of PRPs.^{11,12,15,17,18} Nevertheless, despite the overall similarity in the β-helical fold, there are differences among Qnr proteins from Gram-negative bacteria compared with MtMfpA and EfsQnr from Gram-positives. The folding of pentapeptides into the β-helical turns is interrupted by two non-canonical PRP sequences that produce outward-projecting loops. Loop A and loop B (the larger loop) extend out from the surface of the β-helical fold and are present (or predicted to be present) in all Qnr proteins found in Gram-negative bacteria, but absent in two Qnr proteins found in Gram-positives, MtMfpA and EfsQnr.^14,15,18,19 Loop B has been shown to be important for quinolone resistance in AhQnr and QnrB1 and has been proposed to be conformationally flexible.^17,18 The present work evaluated the role of loop B in QnrS1 and individual amino acids within it in quinolone resistance and gyrase protection to extend understanding of the structure–activity relationships of loop B of Qnr proteins.

Materials and methods

Plasmid construction of pET-28a:QnrS1

A DNA fragment encompassing the qnrS1 gene was amplified by PCR from plasmid pMG306 (GenBank accession no. DQ485529.1) using the primers listed in Table 1.²⁰ The resulting DNA fragment was double-digested with SacI and XhoI (New England Biolabs, Ipswich, MA, USA) and cloned into expression vector pET-28a (Novagen, USA).

Table 1.

Primers used in this study

	Primers
	Primers for amplification of QnrS1
	5′-GGGAATTCGAGCTCATGGAAACCTACAATCATACATATC-3′
	5′-CGCCTCGAGTTAGTCAGGATAAACAACAATACC-3′
Amino acid changea	Primers for site-directed mutagenesis
His-106→Ala	5′-CAACTTTTCCCGAACAAACTTTGCCGCCCAAGTGAGTAATCGTATGTACTTTT-3′
	5′-AAAAGTACATACGATTACTCACTTGGGCGGCAAAGTTTGTTCGGGAAAAGTTG-3′
Gln-107→Ala	5′-CTTTTCCCGAACAAACTTTGCCCATGCCGTGAGTAATCGTATGTACTTTTGC-3′
	5′-GCAAAAGTACATACGATTACTCACGGCATGGGCAAAGTTTGTTCGGGAAAAG-3′
Val-108→Ala	5′-CGAACAAACTTTGCCCATCAAGCCAGTAATCGTATGTACTTTTGCT-3′
	5′-AGCAAAAGTACATACGATTACTGGCTTGATGGGCAAAGTTTGTTCG-3′
Ser-109→Ala	5′-GAACAAACTTTGCCCATCAAGTGGCCAATCGTATGTACTTTTGCTCAGC-3′
	5′-GCTGAGCAAAAGTACATACGATTGGCCACTTGATGGGCAAAGTTTGTTC-3′
Asn-110→Ala	5′-CGAACAAACTTTGCCCATCAAGTGAGTGCCCGTATGTACTTTTGCTCAGC-3′
	5′-GCTGAGCAAAAGTACATACGGGCACTCACTTGATGGGCAAAGTTTGTTCG-3′
Arg-111→Ala	5′-CCGAACAAACTTTGCCCATCAAGTGAGTAATGCCATGTACTTTTGCTCAGC-3′
	5′-GCTGAGCAAAAGTACATGGCATTACTCACTTGATGGGCAAAGTTTGTTCGG-3′
Met-112→Ala	5′-CTTTGCCCATCAAGTGAGTAATCGTGCCTACTTTTGCTCAGCATTTATTTCTG-3′
	5′-CAGAAATAAATGCTGAGCAAAAGTAGGCACGATTACTCACTTGATGGGCAAAG-3′
Tyr-113→Ala	5′-CCCATCAAGTGAGTAATCGTATGGCCTTTTGCTCAGCATTTATTTCTG-3′
	5′-CAGAAATAAATGCTGAGCAAAAGGCCATACGATTACTCACTTGATGGG-3′
Phe-114→Ala	5′-CCCATCAAGTGAGTAATCGTATGTACGCCTGCTCAGCATTTATTTCTGGATGTA-3′
	5′-TACATCCAGAAATAAATGCTGAGCAGGCGTACATACGATTACTCACTTGATGGG-3′
Cys-115→Ala	5′-GCCCATCAAGTGAGTAATCGTATGTACTTTGCCTCAGCATTTATTTCTG-3′
	5′-CAGAAATAAATGCTGAGGCAAAGTACATACGATTACTCACTTGATGGGC-3′
Ser-116→Ala	5′-GCCCATCAAGTGAGTAATCGTATGTACTTTTGCGCCGCATTTATTTCTG-3′
	5′-CGGGTAGTTCACTCATTAGCATACATGAAACGGCGGCGTAAATAAAGAC 3′

Primers have been included in the following order: forward primer, reverse primer.

^aAmino acid replacement resulting from the use of the corresponding pair of mutagenic oligonucleotide primers.

Production of QnrS1 mutants

Plasmid pET-28a:QnrS1 was transformed into E. coli BL21(DE3) (New England Biolabs) and plated on Luria–Bertani (LB) agar (Becton Dickinson, USA) containing 100 mg/L kanamycin to maintain the plasmid. The 11 QnrS1 proteins mutated in loop B were produced using the QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies, USA), plasmid pET-28a:QnrS1 and the primers listed in Table 1. Each amino acid from positions 106 through 116 in the original amino acid sequence of QnrS1 (Figure 1) was substituted by alanine. Alanine scanning is frequently used to determine the role of specific residues in the stability or function of a given protein. The alanine side chain is small and non-reactive and thereby allows for assessment of the roles of other amino acid side chains with different size and chemical characteristics.²¹ The native alanine residue 105 in QnrS1 loop B was not modified (Figure 1).¹⁹ The 11 mutated plasmids were transformed into E. coli BL21(DE3) and their DNA sequences were confirmed at the Tufts University core facilities.

Figure 1.

Comparison of plasmid-encoded homologues of QnrS1 loop B amino acid sequence. Amino acid residues that are coincident with those of QnrS1 loop B are underlined. GenBank accession number for proteins are available in the qnr numbering and sequence web page.³⁶

MIC determinations

The strain BL21(DE3) was chosen together with expression vector pET-28a to assess the changes in the ciprofloxacin MIC conferred by cloned qnrS1 with or without mutations in qnrS1 loop B codons, because this system allows controlled IPTG-regulated expression of the cloned gene.^22,23 Ciprofloxacin MICs were determined for E. coli BL21(DE3), E. coli BL21(DE3) pET-28a:QnrS1 and the 11 mutants in QnrS1 loop B with at least three independent experiments by the broth microdilution method following CLSI guidelines.²⁴ Ciprofloxacin MICs were also determined in the presence of 10, 100 and 1000 μM IPTG (bioMérieux, France) for induction of expression. Etest for ciprofloxacin MIC determinations could not be used since the point at which E. coli BL21(DE3) growth intersected the MIC reading scale was below the ciprofloxacin minimum concentration in the Etest strip (bioMérieux, France), 0.002 mg/L.

Protein production and purification

Overnight cultures of BL21(DE3) strains with pET-28a:QnrS1 or the 11 mutants with single amino acid substitutions were inoculated in 4 L flasks containing 650 mL of LB broth (Becton Dickinson, USA) with kanamycin (100 mg/L) and grown at 37°C with vigorous shaking until the mid-exponential phase of growth. QnrS1 protein expression was induced by the addition of 1.0 mM IPTG and incubation continued for 6 h, as described previously.⁹ The cells were harvested by centrifugation, washed with cold buffer A [50 mM Tris, pH 8.0, 200 mM (NH₄)₂SO₄, 10% glycerol and 20 mM imidazole] and the cell pellet was resuspended in buffer A supplemented with 1 mg/mL lysozyme and 0.5% Triton X-100 and kept at 4°C overnight, followed by sonication and addition of 0.1 mg/mL DNase. After centrifugation and filtration of the lysed samples, the resulting supernatant was applied to a 5 mL HiTrap Chelating HP column (GE Healthcare, USA) that had been previously equilibrated with buffer A. The column was washed with a series of imidazole concentrations (30, 40, 50, 100, 200 and 300 mM) in buffer A containing EDTA-free protease inhibitor cocktail tablets (Roche, USA). The collected eluate was immediately aliquotted and frozen at −80°C. Purified protein extracts were dialysed against buffer B containing 20 mM Tris, pH 8.0, 10% glycerol and 50 mM arginine to improve the solubility of the protein as previously described.^17,25 Protein extracts were concentrated using Amicon Ultra-4 tubes (Fisher Scientific, USA) and the protein concentrations were determined with a NanoDrop 1000 Spectrophotometer (Thermo Scientific, USA).¹³ The yield of purified protein was expressed in mg of protein/g cell paste that was calculated for each purified QnrS protein taking into account the final total soluble protein after nickel-iminodiacetic acid (Ni-IDA) elution and the final OD₆₀₀ units of 2.6 L of culture in each case.

Analysis of final purification products

The presence of the polyhistidine-tagged fusion proteins in the purified samples was confirmed by western blot using the SuperSignal West HisProbe Kit (Thermo Scientific, USA). To determine the purity of the final products, aliquots were separated in NuPAGE 12% Bis-Tris Mini Gels (Novex, USA) and stained with a Coomassie R-250 formulation (Thermo Scientific, USA) and with silver stain (GE Healthcare, USA).

Gyrase assays

DNA supercoiling assays were performed using E. coli DNA gyrase holoenzyme (TopoGEN Inc., USA). Reaction mixtures contained 2 U (9.2 nM) gyrase, with a unit defined as the amount of gyrase that catalysed >90% conversion of 0.5 μg of relaxed pHOT-1 DNA (TopoGEN Inc.) to supercoiled plasmid in 1 h at 37°C in 40 μL of gyrase buffer (TopoGEN Inc.; containing 35 mM Tris-HCl, pH 7.5, 24 mM KCl, 4 mM MgCl₂, 2 mM dithiothreitol, 1.8 mM spermidine, 1 mM ATP, 6.5% glycerol and 0.1 mg/mL BSA). Reactions were terminated by the addition of 40 μL of chloroform/iso-amyl alcohol (24/1).^17,26 The reaction mixtures were subjected to agarose gel electrophoresis, stained with ethidium bromide and visualized under ultraviolet light. The intensity of the most highly supercoiled DNA band was determined using Image Lab software (Bio-Rad Laboratories Inc., USA).

The IC₅₀ of ciprofloxacin for E. coli gyrase supercoiling activity was determined as the drug concentration causing a 2-fold reduction in the intensity of the most highly supercoiled DNA band.²⁶ Likewise, the activities of wild-type and mutated QnrS1 proteins were titrated using a series of protein concentrations in the presence of a partially inhibitory concentration of ciprofloxacin with protection based on the extent of restoration of the intensity of the most highly supercoiled DNA band. The half protective concentration (PC₅₀) and the concentration providing complete protection (PC₁₀₀) were defined as the concentrations of QnrS1 protein that relieved 50% and 100%, respectively, of ciprofloxacin inhibition of gyrase activity. At least three independent assays with consistent results were made for each set of concentrations with each of the 12 QnrS1 proteins studied. Only standard deviations <10% for density measurements for each assayed concentration and QnrS1 protein were accepted. PC₅₀ values were expressed as means since the range of concentrations for wild-type and mutated QnrS1 proteins that relieved 50% of ciprofloxacin inhibition of gyrase activity was narrow, whereas there were wide ranges of concentrations that relieved 100% inhibition for the studied Qnr proteins. In addition, the possible inhibition of gyrase DNA supercoiling activity by wild-type and mutated QnrS1 proteins was also assessed in the presence and absence of ciprofloxacin IC₅₀.

Results

Effects of single amino acid substitutions in loop B of QnrS1 on ciprofloxacin MICs in whole cells

E. coli BL21(DE3) containing pET-28a:QnrS1 in the absence of IPTG induction showed a 12-fold increased resistance to ciprofloxacin relative to BL21(DE3) strain or BL21(DE3) containing the pET-28a vector plasmid alone (Table 2). Induction of expression of QnrS1 by IPTG further increased resistance 2- to 4-fold. In the absence of IPTG induction, all but three loop B alanine mutants of QnrS1 expressed resistance to ciprofloxacin 1.4- to 8-fold less than that of wild-type QnrS1, with the greatest reductions in the mutants of the most highly conserved amino acids Phe-114, Cys-115 and Ser-116 (Figure 1 and Table 2). Notably, with increasing IPTG induction, all loop B mutants produced increases in the ciprofloxacin MIC to wild-type or near-wild-type levels (Table 2).

Table 2.

Effect of expression of wild-type qnrS1 and mutated qnrS1 genes encoding single alanine substitutions in loop B on susceptibility to ciprofloxacin of E. coli BL21(DE3)

	Mean ciprofloxacin MICs (mg/L) and standard deviationsa at the following concentrations of IPTG (μM):
	0	10	100	1000
BL21(DE3)b
no plasmid vector	0.0013 ± 0.0006	0.001	0.001	0.001
vector alone	0.0013 ± 0.0006	0.001	0.001	0.001
QnrS1 (wild-type)c	0.016	0.053 ± 0.018	0.064	0.032
BL21(DE3) and mutated qnrS1d
His-106→Ala	0.011 ± 0.005	0.016	0.027 ± 0.009	0.027 ± 0.009
Gln-107→Ala	0.032	0.032	0.064	0.032
Val-108→Ala	0.004	0.016	0.027 ± 0.009	0.016
Ser-109→Ala	0.006 ± 0.0023	0.016	0.032	0.032
Asn-110→Ala	0.016	0.029 ± 0.007	0.032	0.032
Arg-111→Ala	0.016	0.027 ± 0.009	0.053 ± 0.018	0.032
Met-112→Ala	0.006 ± 0.0023	0.016	0.032	0.032
Tyr-113→Ala	0.005 ± 0.0023	0.032	0.053 ± 0.018	0.032
Phe-114→Ala	0.0035 ± 0.001	0.011 ± 0.005	0.032	0.024 ± 0.009
Cys-115→Ala	0.002	0.016	0.027 ± 0.009	0.032
Ser-116→Ala	0.0035 ± 0.001	0.032	0.064	0.032

^aStandard deviation is included when it is >0.

^bE. coli strain BL21(DE3) with and without pET-28a expression plasmid vector.

^cBL21(DE3) expressing wild-type qnrS1 cloned into pET-28a.

^dE. coli strain BL21(DE3) expressing mutated qnrS1 encoding single alanine substitutions in loop B.

Wild-type QnrS1 protein activity in gyrase protection assays

Wild-type and QnrS1 proteins with amino acid changes in loop B were expressed with an N-terminal His₆ tag. The yield of total soluble protein after Ni-IDA elution for the 12 different purified proteins ranged from 7.7 to 15.5 mg/g cell paste and after concentration ∼5 mg/mL. All the purified extracts containing QnrS1 proteins separated by SDS–PAGE showed a band that migrated to ∼25 kDa, consistent with the expected size of QnrS1, and was positive for the His₆ tag by western blot. Based on relative band intensity, the QnrS1 proteins constituted 90% of total purified protein.

DNA gyrase catalysed 95% ± 4.5% conversion of 0.5 μg of relaxed substrate pHOT-1 DNA to supercoiled form, whereas the presence of ciprofloxacin 1.8 μM (0.6 mg/L) resulted in 48% ± 3% of the supercoiled form.

Purified wild-type QnrS1 protein at 0.05 nM fully protected gyrase supercoiling activity from 1.8 μM ciprofloxacin and 100-fold lower concentrations of QnrS1 produced half protection (Table 3). At a higher concentration of ciprofloxacin (6 μM), the same concentration of QnrS1 (0.05 nM) produced only partial protection, indicating that the interactions among QnrS1, gyrase and ciprofloxacin are concentration dependent. Gyrase was inhibited by QnrS1 at concentrations ≥6 μM, but this effect was not seen at concentrations <2 μM (data not shown).

Table 3.

QnrS1 protection of DNA gyrase from ciprofloxacin inhibition

QnrS1 protein	PC50 (nM)a	PC100 (nM)b
Wild-type	5 × 10−4	0.05–250
His-106→Ala	5 × 10−4	0.5–50
Gln-107→Ala	10−5	0.05–250
Val-108→Ala	5	5–50c
Ser-109→Ala	0.05	5–50
Asn-110→Ala	5 × 10−4	0.05–250
Arg-111→Ala	10−3	0.2–250
Met-112→Ala	0.05	5–50c
Tyr-113→Ala	0.5	25–250
Phe-114→Ala	0.5	5–50c
Cys-115→Ala	5	5–50c
Ser-116→Ala	0.5	50–100c

^aPC₅₀, mean values of concentrations of QnrS1 proteins that half relieved gyrase supercoiling inhibition by 1.8 μM ciprofloxacin.

^bPC₁₀₀, range of concentrations of QnrS1 proteins that completely relieved gyrase supercoiling inhibition by 1.8 μM ciprofloxacin (with the exception of c–g).

^cConcentrations of QnrS1 mutant proteins that relieved 75%–85% of gyrase supercoiling inhibition by 1.8 μM ciprofloxacin.

Effects of amino acid substitutions in loop B of QnrS1 on protection of gyrase from quinolone action

Similar to the effects of various QnrS1 alanine replacements on quinolone resistance in whole cells, QnrS1 proteins with Asn-110→Ala and Arg-111→Ala or Gln-107→Ala, which conferred the same or a 2-fold increase in ciprofloxacin MICs relative to wild-type QnrS1 (Table 2), had both PC₅₀ and PC₁₀₀ values similar to or lower (Gln-107→Ala) than those of wild-type QnrS1 (Table 3). These three mutated QnrS1 proteins, like wild-type QnrS1, also maintained their gyrase protective activity at concentrations up to 2 μM without any evident inhibitory effect, while inhibition of gyrase was seen at concentrations ≥6 μM (data not shown). QnrS1 protein with His-106→Ala change, which resulted in a ciprofloxacin MIC only 1.4-fold lower than wild-type QnrS1 (Table 2), exhibited the same PC₅₀ value as wild-type (Table 3). In contrast, large increases in PC₅₀ and PC₁₀₀ values were found for the remaining seven mutated QnrS1 proteins, which increased the fluoroquinolone MIC 2.7- to 8-fold less than wild-type QnrS1 did. Furthermore, five QnrS1 proteins with Ala replacements in loop B did not reach a PC₁₀₀ value and only relieved ciprofloxacin inhibition of gyrase supercoiling by a maximum of 75%–85% (Table 3). In the case of QnrS1 proteins with alanine substitutions in the most highly conserved residues Phe-114, Cys-115 and Ser-116 as well as in Val-108 and Tyr-113, PC₅₀ values were ≥1000-fold higher than that of wild-type (Table 3 and Figure 2). Likewise, QnrS1 proteins with Ser-109→Ala and Met-112→Ala changes exhibited 100-fold increases in PC₅₀ values compared with those of wild-type QnrS1 (Table 3 and Figure 2c and e). In addition, QnrS1 proteins with replacements in Val-108, Ser-109, Met-112, Tyr-113, Phe-114, Cys-115 and Ser-116 at concentrations ≥0.3–0.5 μM (Figure 2) did not protect gyrase supercoiling activity against quinolone inhibition. Likewise, concentrations ≥1.2 μM of the above seven mutated QnrS1 proteins and QnrS1 His-106→Ala protein produced complete inhibition of gyrase supercoiling activity (data not shown).

Figure 2.

DNA gyrase protection assays with QnrS1 purified proteins. Reactions of 25 μL were analysed by agarose gel electrophoresis. Reactions mixtures (40 μL) contained 0.5 μg of relaxed pHOT-1 DNA (lanes 1–12), 2 U of purified E. coli DNA gyrase (lanes 2–12) and 1.8 μM (0.6 mg/L) ciprofloxacin (lanes 3–12). The concentrations of the 12 different QnrS1 purified proteins are expressed in nM above each lane; the numbers in parentheses refer to the power of 10 that multiplies the value. (a) QnrS1 wild-type (lanes 4–7) and QnrS1 Q107A (lanes 8–12). (b) QnrS1 H106A (lanes 4–7) and QnrS1 V108A (lanes 8–12). (c) QnrS1 S109A (lanes 4–7) and QnrS1 N110A (lanes 8–11). (d) QnrS1 R111A (lanes 4–7) and QnrS1 Y113A (lanes 8–12). (e) QnrS1 M112A (lanes 4–8). (f) QnrS1 F114A (lanes 4–7) and QnrS1 C115A (lanes 8–11). (g) QnrS1 wild-type (lanes 4–7) and QnrS1 S116A (lanes 8–12). nc, l and sc indicate positions of the nicked circular, linear and supercoiled forms of pHOT-1, respectively.

Discussion

The amino acids in loop B of QnrB1 and AhQnr and those in the same loop B domain of 18 other Qnr proteins have substantial conservation of sequence (Figure 1),^{17–20,27–35} suggesting that they are important for Qnr function.^17–19 In fact, the amino acid compositions of the B loops of all identified variants of QnrB are identical,^3,34,36 as are the compositions of the B loops of QnrS proteins, with the exception of QnrS2 and QnrS6, which have Asn-106 instead of His-106 (Figure 1).^{3,27,29,34,36} Likewise, both residues Phe-114 and Cys-115 (following QnrS1 numbering) are conserved not only in the B loops of QnrA1, QnrB1, QnrS1, QnrC, QnrD and QnrVC but also in the alignment of Qnr homologues from 57 different bacterial species.³ In contrast, the composition of loop A is more variable and its deletion results in only a partial loss of activity.^17–19 QnrS1, however, has not been directly evaluated for the contribution of loop B to resistance nor has the contribution of all individual amino acids throughout loop B been studied. Thus, we assessed the contributions of individual amino acids in loop B of QnrS1 to its ability to confer resistance to ciprofloxacin and to interact with DNA gyrase.

It is notable that E. coli strain BL21(DE3) is hypersusceptible to ciprofloxacin. This hypersusceptibility is likely due to the absence of the Lon protease that results in high levels of the cell division inhibitor SulA (which is a Lon substrate), after its induction by ciprofloxacin, as has been described for another E. coli Δlon strain and levofloxacin.³⁷ Nevertheless, the presence of wild-type qnrS1 cloned into pET-28a with or without induction conferred 12- to 64-fold increases in the ciprofloxacin MIC for BL21(DE), as previously described for other cloned qnr genes in other strain backgrounds.³⁸ Likewise, all qnrS1 genes encoding amino acid changes in loop B resulted in increases in the ciprofloxacin MIC, albeit to different levels, indicating that the ability of the mutated QnrS1 proteins to protect gyrase from quinolone action was not completely eliminated, but reduced with some mutations in loop B (Figure 1). Thus, the findings of the present study indicate that none of the individual amino acids alone is essential for the ability of qnrS1 to confer resistance, despite the requirement for the presence of the loop B itself as previously described.^17,18 Nevertheless, the fact that some mutations in QnrS1 loop B conferred lower ciprofloxacin resistance could indicate a lesser affinity or altered positioning of certain mutated QnrS1 proteins on the target gyrase. Consistent with our findings, a previous study of mutagenesis of qnrA1, qnrB1 and qnrS1 also found that single amino acid substitutions by tyrosine or aspartic acid at positions 114–116 of QnrA1, 114–116 of QnrS1 or 111–113 of QnrB1 conferred lower ciprofloxacin MIC increases than those by their respective wild-type proteins.³⁹ This report, however, did not assess mutants in other amino acids in loop B or the effect of induced expression on resistance.³⁹ Our findings also extend those previously reported³⁹ and indicate that the individual position and properties of the native residues of loop B contribute to its role in quinolone resistance; thus, substitutions in Val-108 and Ser-109 but also those localized in any of the last five C-terminal amino acids of loop B (Figure 1) resulted in a reduction of conferred ciprofloxacin resistance and gyrase protection.

Assays to measure gyrase protection of ciprofloxacin inhibition by the purified wild-type QnrS1 protein showed partial protection at sub-picomolar concentrations of QnrS1, as previously also described for QnrB1 using a higher ciprofloxacin concentration (6 μM).⁴ The PC₅₀ value found for QnrB1 (0.5 nM), however, was higher than that found for QnrS1 in the present study (5 × 10⁻⁴ nM). Thus, higher concentrations of ciprofloxacin require higher Qnr concentrations for DNA gyrase protection as previously described.^8,17

Reductions in the gyrase protection of different QnrS1 proteins with alanine substitutions correlated in general with the relative magnitude of their reduced ability to confer resistance. Notably, those two mutated QnrS1 proteins (Asn-110→Ala and Arg-111→Ala) with PC₅₀ values that were the same or 2-fold higher than that for wild-type QnrS1 conferred the same ciprofloxacin MIC and, in turn, QnrS1 with Gln-107→Ala, which had relative to wild-type QnrS1 a 50-fold lower PC₅₀ value, conferred a ciprofloxacin MIC 2-fold higher than for wild-type. Indeed, Gln-107→Ala represents the first example of a Qnr loop mutant with documented increased gyrase protection. Although a previous report described a 4-fold increase in the ciprofloxacin MIC associated with the substitution Asp-185→Tyr in QnrS1, it was in a residue located outside loops A and B.^16,19,39 For those mutated QnrS1 proteins with the smallest increases in conferred resistance (Val-108→Ala, Ser-109, Met-112, Tyr-113, Phe-114, Cys-115→Ala and Ser-116), PC₅₀ values increased hundreds- to thousands-fold with respect to wild-type QnrS1 PC₅₀. Likewise, for QnrS1 proteins with alanine substitutions in Val-108, Met-112, Phe-114, Cys-115 and Ser-116, 100% protection potency was not found at any protein concentration, whereas minimum PC₁₀₀ values for Ser-109→Ala and Tyr-113→Ala exhibited large increases. These findings suggest that for some proteins with loop B mutations there may be additional factors beyond those inside the intact bacterial cells that contribute to their reduced activity in the gyrase protection assay. For instance, QnrB1 proteins with mutations Met-102→Arg and Cys-112→Arg were less soluble than wild-type at high concentrations 10–15 mg/mL.¹⁷

Furthermore, the inhibition of DNA gyrase supercoiling activity by some Qnr protein concentrations, which had been previously described for QnrB4 and QnrB1,^4,13 was also seen for both wild-type QnrS1 and mutated proteins, suggesting that there is a limit in the conferred protection by QnrS1 proteins on DNA gyrase against ciprofloxacin and at higher concentrations inhibition appears in vitro assays. This finding could explain the observation that 100 μM IPTG resulted in the best induction of ciprofloxacin MIC increases for BL21(DE3) expressing wild-type QnrS1 and QnrS1 loop B mutants, which did not increase further with 1000 μM IPTG. Indeed, maximum protection potency was reached for the studied QnrS1 proteins at concentrations ranging from 0.05 to 250 nM and higher protein concentrations did not improve the protection potency. Finally, the fact that the ciprofloxacin MIC for BL21(DE3) was increased by the expression of all 11 mutated proteins, including those with lower gyrase protection activity than wild-type QnrS1, suggests that levels of QnrS1 proteins inside the bacterial cell are maintained near those at which the protective potencies of the proteins with loop B mutations are effective. Definitive understanding of the mechanism by which the B loop of Qnr proteins contributes to their gyrase protective functions, however, will likely require X-ray crystallographic analysis of the complex between Qnr, gyrase and DNA.

Funding

This work was supported by grants from the US Public Health Service, National Institutes of Health (R01 AI057576 to D. C. H. and G. A. J.) and the Spanish Ministry of Education (Ministerial Order EDU/3378/2010, 21st December and extension, PR2010-0522 to M.M.T.P.-ULPGC).

Transparency declarations

None to declare.