Fluoroquinolone Resistance Does Not Impose a Cost on the Fitness of Clostridium difficile In Vitro

Abstract

Point mutations conferring resistance to fluoroquinolones were introduced in the gyr genes of the reference strain Clostridium difficile 630. Only mutants with the substitution Thr-82→Ile in GyrA, which characterizes the hypervirulent epidemic clone III/027/NAP1, were resistant to all fluoroquinolones tested. The absence of a fitness cost in vitro for the most frequent mutations detected in resistant clinical isolates suggests that resistance will be maintained even in the absence of antibiotic pressure.

TEXT

Modifications within the quinolone resistance-determining regions (QRDRs) of the DNA gyrase subunits GyrA and GyrB are the only identified mechanism of resistance to fluoroquinolones in Clostridium difficile (1, 2). The substitution Thr-82→Ile in GyrA characterizes the epidemic clone III/027/NAP1, and it is prevalent in isolates resistant or showing a reduced susceptibility to moxifloxacin (3, 4). The substitution Thr-82→Val has also been detected, although rarely (3, 5). In GyrB substitutions Asp-426→Asn and Asp-426→Val are some of the most common changes found in resistant isolates (2, 6).

The main parameters that are likely to affect the emergence, maintenance, and diffusion of mutations conferring antibiotic resistance are the mutation rate itself and the subsequent burden imposed on the fitness of the bacterium. To this day, nothing is known concerning this burden in C. difficile, despite the strong association between the use of fluoroquinolones and C. difficile infection (CDI).

In this study, fluoroquinolone-resistant mutants were generated by introduction of point mutations into either the gyrA or gyrB QRDR domains of the fluoroquinolone-susceptible reference strain C. difficile 630, using a two-step allele exchange recently developed (7). These mutants were characterized to determine the impact of the introduced mutations in terms of resistance and fitness cost in vitro.

All strains used in this study are listed in Table 1. C. difficile was grown at 35°C under anaerobic conditions on brain heart infusion (BHI [Oxoid]) supplemented with 0.5% yeast extract and 0.1% l-cysteine (BHIS). When appropriate, BHIS plates were supplemented with d-cycloserine (250 mg/liter), cefoxitin (8 mg/liter), and thiamphenicol (15 mg/liter). C. difficile minimal medium (CDMM) supplemented with 5-fluorocytosine (5-FC) was used to select for loss of pMTL-SC7315 (7). Escherichia coli was grown aerobically at 37°C in LB. Chloramphenicol was used at final concentrations of 12.5 and 25 mg/liter in broth and agar, respectively.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Characteristics	Source or reference(s)
Strains
E. coli
α-select competent cells	F− deoR endA1 recA1 relA1 gyrA96 hsdR17(rk− mk+) supE44 thi-1 phoA Δ(lacZYA-argF)U169 ϕ80lacZΔM15 λ−	Bioline, United Kingdom
CA434	HB101 [F− mcrB mrr hsdS20(rB− mB−) recA13 leuB6 ara-14 proA2 lacY1 galK2 xyl-5 mtl-1 rpsL20(Smr) glnV44 λ−] with plasmid R702	16, 17
C. difficile
630 (wild type)	PCR ribotype 012	18
ISST82Ia, ISST82Ib, ISST82Ic	630 gyrA(Thr-82→Ile)	This study
ISST82Va, ISST82Vb, ISST82Vc	630 gyrA(Thr-82→Val)	This study
ISSD426Na, ISSD426Nb, ISSD426Nc	630 gyrB(Asp-426→Asn)	This study
ISSD426Va, ISSD426Vb, ISSD426Vc	630 gyrB(Asp-426→Val)	This study
Plasmid
pMTL-SC7315	E. coli-C. difficile shuttle vector (pCB102 catP ColE1 traJ codA)	7

Genomic DNA was extracted from C. difficile using the NucleoBond AXG columns and NucleoBond buffer set III (Macherey-Nagel). Plasmid DNA isolation from E. coli was carried out using the NucleoSpin plasmid kit (Macherey-Nagel). DNA was purified from agarose gels using the NucleoSpin extract II kit (Macherey-Nagel). All kits were used according to the manufacturer's recommendations.

Allele exchange cassettes were generated by PCR using the primers listed in the Table S1 in the supplemental material and cloned into the PmeI site of pMTL-SC7315. Ligation products were used to transform E. coli α-select competent cells. Plasmid DNA was then extracted, Sanger sequenced, introduced in E. coli CA434, and transferred to C. difficile by conjugation, as described previously (7). Transconjugants were streaked on BHIS plates with 4 mg/liter of moxifloxacin to directly select allele exchange events between the vector and the chromosome. The colonies obtained were grown again on the same medium and incubated for at least 48 h to allow subsequent plasmid loss. Colonies were then patch plated onto BHIS with thiamphenicol and CDMM supplemented with 5-FC (7). Desired mutations in thiamphenicol-susceptible and 5-FC-resistant colonies were confirmed by Sanger sequencing of the gyrA and gyrB genes. Three separate clones of each mutant were isolated, each from independent conjugations.

MICs were determined using the Etest (AB Biodisk) as recommended by the Clinical and Laboratory Standards Institute (8). The breakpoint used for moxifloxacin, levofloxacin, ciprofloxacin and gatifloxacin was 8 mg/liter, in accordance with the guidelines established by the CLSI (8).

Bacterial fitness was estimated by competition assays performed as already described (9). The number of viable cells was determined at the end of every cycle by plating aliquots on BHI agar plates containing 4 mg/liter moxifloxacin or without antibiotic. The competition index (CI) was calculated as the CFU ratio of the resistant and susceptible strains at time t₁ divided by the same ratio at time t₀. The selection coefficient s was calculated as s = ln(CI)/[t × ln(2)]. Fitness of the susceptible strain was set to 1, and the relative fitness of the mutants was determined as 1 + s. Each independent mutant was tested versus the wild-type strain three times, for a total of nine independent assays for each single mutation analyzed. Statistical analyses were performed using the GraphPad software and an unpaired Student's t test.

Mutations introduced in the genome of C. difficile 630 resulted in the amino acid changes Thr-82→Ile or Thr-82→Val in GyrA and Asp-426→Asn or Asp-426→Val in GyrB. Genome sequencing of the mutants ISST82Ia and ISST82Va and the progenitor strain 630 (GATC, Constance, Germany, for ISST82Ia and progenitor 630; DEEPSEQ, the University of Nottingham, Nottingham, United Kingdom, for ISST82Va) validated the efficiency of the mutation strategy. Single nucleotide polymorphism (SNP) analysis was performed using the deposited genome sequence of C. difficile 630 (GenBank accession no. NC_009089) as the reference. Overall genome coverage levels were 324.2-, 283.0-, and 795.0-fold for strains 630, ISST82Ia, and ISST82Va, respectively. All strains contained the same six SNPs compared to the reference sequence. As expected, the C→T mutation introduced at position 6310 was detected in ISST82Ia, while the AC→GT substitutions at positions 6309 to 6310 were detected in ISST82Va, resulting in the Thr-82→Ile and Thr-82→Val substitutions in the encoded GyrA sequence, respectively (data not shown). In addition, a G→T mutation was detected at position 3949724 in ISST82Va only, leading to an Arg-113→Ser substitution at a nonconserved position of a magnesium-transporting ATPase. We anticipated that this substitution would affect neither antimicrobial resistance nor fitness of the bacterium.

MIC values for moxifloxacin, ciprofloxacin, levofloxacin, and gatifloxacin were identical between replicate clones containing the same mutation. All mutants were highly resistant to ciprofloxacin, levofloxacin, and gatifloxacin, except for mutants with the substitution Asp-426→Val in GyrB, which showed reduced susceptibility to the last antibiotic (MIC, 6 mg/liter) (Table 2). While mutants with Thr-82→Ile substitution in GyrA were highly resistant to moxifloxacin, the other mutants showed a MIC of 4 mg/liter, in accordance with the observation that substitution Thr-82→Ile is generally associated with high-level resistance (MIC, ≥32 mg/liter) (6). Notably, mutant ISST82V, which had undergone two mutations in the triplet ACT coding for Thr-82, showed decreased MIC values to moxifloxacin and gatifloxacin compared to mutant ISST82I (Table 2).

TABLE 2.

MICs of C. difficile 630 and the mutants obtained in this study

Strain(s)	MIC (mg/liter) ofa:
	MXF	GAT	LVX	CIP
630 (wild type)	0.5	0.75	3	6
ISST82Ia, ISST82Ib, ISST82Ic	32	≥32	≥32	≥32
ISST82Va, ISST82Vb, ISST82Vc	4	8	≥32	≥32
ISSD426Na, ISSD426Nb, ISSD426Nc	4	8	≥32	≥32
ISSD426Va, ISSD426Vb, ISSD426Vc	4	6	≥32	≥32

^aMXF, moxifloxacin; GAT, gatifloxacin; LVX, levofloxacin; CIP, ciprofloxacin.

Even if Asp-426→Val has been associated with higher MICs to fluoroquinolones than Asp-426→Asn, which is associated with intermediate and high levels of resistance (4, 6), both substitutions conferred to strain 630 mutants high levels of resistance only to levofloxacin and ciprofloxacin (Table 2).

Substitution Thr-82→Ile in GyrA had no detectable cost on the fitness of C. difficile 630, while substitution Thr-82→Val reduced the fitness of the bacterium (Table 3). The fitness cost imposed by mutations occurring in corresponding positions of GyrA in other pathogens is variable, ranging from disadvantage to benefit (10–15). In vitro results obtained in this study suggest that substitution Thr-82→Ile in GyrA can be maintained in the bacterial population even in the absence of antibiotic selective pressure. This is in accordance with the large spread of strains carrying this mutation, in particular the epidemic clone III/027/NAP1. The fact that two nucleotide mutations are required to produce the substitution Thr-82→Val, together with the observation that this substitution reduces both fitness and MICs to fluoroquinolones, may explain its rarity compared to Thr-82→Ile among C. difficile clinical isolates (Tables 2 and 3).

TABLE 3.

Competition assays between wild-type C. difficile strain 630 and fluoroquinolone-resistant mutants

Mutants tested vs 630	sa	Relative fitness/generationb	P valuec
ISST82Ia, ISST82Ib, ISST82Ic	−0.0216	0.9784 ± 0.0757	0.4036
ISST82Va, ISST82Vb, ISST82Vc	−0.0952	0.9018 ± 0.0462	<0.0001
ISSD426Na, ISSD426Nb, ISSD426Nc	0.0877	1.0877 ± 0.0813	0.0052
ISSD426Va, ISSD426Vb, ISSD426Vc	0.0589	1.0590 ± 0.0749	0.0312

^aSelection coefficient.

^bFitness relative to that of the susceptible strain.

^cStatistical significance of difference in fitness.

Interestingly, both substitutions Asp-426→Asn and Asp-426→Val in GyrB conferred a slight advantage to the mutants in vitro, compared to the wild-type strain 630 (Table 3). However, it is interesting to note that, among fluoroquinolone-resistant clinical isolates, mutations in gyrB are less frequent than mutations in gyrA.

To conclude, allele exchange was successfully used to introduce point mutations in the gyrA and gyrB genes of C. difficile 630. The mutation leading to the substitution Thr-82→Ile in GyrA conferred the highest level of resistance to fluoroquinolones. Moreover, the most frequent mutations detected in fluoroquinolone-resistant clinical isolates of C. difficile displayed no fitness cost in vitro. These results suggest that resistance to fluoroquinolones could be maintained in C. difficile populations in the absence of antibiotic pressure, although in vivo studies will be necessary to better characterize the cost of fluoroquinolone resistance in C. difficile.