Characteristics of quinolone resistance in Escherichia coli isolated from wildlife in Poland

Abstract

The widespread use of quinolones (QNs) in human and veterinary medicine has resulted in the emergence of QN-resistant bacteria. The most common and widespread type of resistance to FQ is the presence of point mutations in the QRDR region and/or PMQR genes. The aim of the study was to analyze the mechanisms of quinolone resistance among E. coli strains isolated from wild animals in Poland. Out of 42 tested isolates, 39 strains had at least one point mutation in the gyrA gene; however, 76.2% of the isolates had at least three mutations in the GyrA and ParC domains, mainly S83L or D87L in the gyrA gene and S80I in the parC gene. Single isolates had also other mutations: D87Y in the GyrA domain and E84G, A108V, G78C, or S80R in the ParC domain. In 12 strains (28.5%), including three in which no QRDR mutations were confirmed, the presence of single PMQR genes: qnrB, qnrS, and aac(6’)-Ib-cr was also demonstrated. Accumulation of point mutations was associated with high values of MIC of fluoroquinolones and single mutations or the presence of PMQR genes were associated with low-level resistance (MIC from 2 to 8 mg/L). The comparative method (MAS-PCR) confirmed the results of sequencing in most cases (90.5%). The E. coli strains isolated from the wildlife were characterized by the same mechanisms of QNs-resistance as those most often identified among isolates from anthropogenic environments. The obtained results indicate that the level of QNs-resistance is really high and monitoring should be obligatorily extended to the natural environment.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-18971-2.

Introduction

Free-living animals are an integral part of the natural environment, and their health is inextricably linked to human health. The World Health Organization indicates that 60% of emerging infectious diseases are zoonoses, with over 70% originating from free-living animals. Studies conducted over the past 20 years indicate that the microbiota of these animals is a significant reservoir of not only typical pathogens but also microorganisms characterized by resistance to many antibiotics, including the so-called last-resort antibiotics in human medicine¹. Escherichia coli as a species commonly found in the microbiota of the digestive tract of animals and humans is one of the important indicator microorganisms in evaluation of the level of drug susceptibility among microorganisms^2,3. The common occurrence and resistance of E. coli, i.e. a species that is both a natural component of the intestine microbiota and a cause of intestinal and extraintestinal infections, pose a particularly problematic challenge in the prevention and therapy of infections⁴. With the increasing introduction of antimicrobials to the market, resistance among microorganisms is emerging almost simultaneously. This phenomenon is often associated with the simultaneous occurrence of several different drug resistance mechanisms, which significantly increase the minimal inhibitory concentration for a given drug. Resistance mechanisms are mostly encoded on mobile genetic elements, thereby significantly increasing the possibility of their spread. At the same time, in many cases, there is a possibility of transferring several different genes responsible for resistance to drugs belonging to different groups, which obviously determines multidrug resistance⁵.

Quinolones (QNs), which include first-generation quinolones (nalidixic acid) and second-fourth-generation fluoroquinolones (FQs), have been used in the treatment of human and animal infections for over 40 years⁶. The bactericidal activity of these drugs is based on the inhibition of DNA synthesis in the bacterial cell and QNs have no or very little affinity for eukaryotic cell, which limits the side effects of their use⁷.

Due to their broad spectrum of action, low toxicity, and favorable pharmacokinetic properties, QNs are agents against many clinically important pathogens in humans and animals^8,9. However, transmission of QNs-resistant bacteria between animals, humans, and the environment is certainly possible^10–12.

The long-term and excessive use of this group of antimicrobials has contributed to the growth of resistance and has forced the introduction of restrictions on their use. The WHO has placed FQs (including ciprofloxacin) in the “Watch group of antibiotics”, which means that they are recommended as essential first or second choice empirical treatment options for a limited number of specific infectious syndromes only.

This correlates with the EMA report on the sale of FQs and their use in EU countries¹⁰, which account for only 3% of the total sales of antimicrobials in veterinary medicine. Nevertheless, there is a very large discrepancy between individual countries, and Poland is in the forefront of countries with the highest sales of FQs¹⁰. This may translate into an increasing level of resistance to this drug among strains isolated from both humans and animals in our country^13,14. Among the mechanisms of resistance to QNs, point mutations in the QRDR region (Quinolone Resistance-Determining Region) and the presence of PMQR (Plasmid-Mediated Quinolone Resistance) genes are of the greatest importance. Mutations in the QRDR region are usually associated with high values of the minimum inhibitory concentration (MIC) of QNs and the presence of highly resistant strains. On the other hand, the presence of PMQR genes located on plasmids, although it only slightly increases the MIC values, is responsible for rapid spread of these genes between bacterial strains. Moreover, the very frequent occurrence of both point mutations in the QRDR region and the presence of plasmid resistance determinants leads to high resistance to QNs¹⁵.

There are few reports in the literature on the mechanisms of resistance to these antimicrobials in wild animals, especially from Poland^16–18. Therefore it seems important to evaluate and analyze the mechanisms of resistance to QNs in wild animals, where human activity has not yet reached such intensity. We hypothesize that the natural environment with a theoretically low impact of anthropization may be a real reflection of the level of resistance to these antimicrobial agents.

The aim of the study was to analyze the mechanisms of quinolone resistance among E. coli strains isolated from free-living animals in Poland, using a rapid protocol that should be employed for screening the genetic background of resistance to quinolones.

Results

Genetic background of quinolone resistance

In our study, 42 strains of E. coli with a ciprofloxacin-resistant or intermediate phenotype were tested (all of them were resistant to nalidixic acid). Mutations in gyrA or parC genes were detected in 39 (92.9%) of the E. coli isolates. Only in three strains, no mutations were detected in any of the domains. Of all the strains, 32 showed the presence of at least three mutations in the tested genes, including five strains with an MIC for ciprofloxacin above 32 mg/L characterized by the presence of four mutations in both genes tested. Three strains had single mutations in each of the GyrA and ParC domains. However, in the case of four isolates, only one mutation in the gyrA gene was detected.

Twelve of the 42 strains tested, had one PMQR gene, with genes encoding protective proteins (qnrS or qnrB) or the aac(6’)-Ib-cr encoding acetyltransferase gene (Table 1). Noteworthy, these genes occurred mainly in strains with the MIC for ciprofloxacin of 2–8 mg/L. In the case of three strains with the MIC equal to 2 mg/L, no mutations were found in the gyrA and parC genes. However, in another three strains, despite the MIC was equal to 2 mg/L, single mutations were simultaneously noted in one or both genes (Table 1). Interestingly, in addition to a double mutation in both genes, the presence of the aac(6’)-Ib-cr gene was additionally demonstrated in three strains isolated from the martens with MIC > 32 mg/L.

Table 1.

Characterization of quinolone resistance of E. coli strains tested in this study.

No of sample	MIC (mg/L) CIPa	MIC (mg/L) NALb	Mutations in gene after sequencing		MAS-PCR			PMQR genes
			gyrA	parC	gyrA		parC
Red foxes (n = 11)
2/CEF	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
11/CEF	2	64						qnrB
14/CEF	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
15/CEF	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
22/CEF	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
28/CEF	8	> 64	S83L, D87N	S80I	GyrA83		ParC80
32/CEF	32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
33/CEF	8	> 64	S83L	-	GyrA83		-	qnrS
34/CEF	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
35/CEF	32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
40/CEF	2	64						qnrB
American minks (n = 10)
N20/TET	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N19/CHL	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N19/TET	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N13/KAN	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N13/TET	8	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N22/CHL	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N17/TET	8	> 64					ParC80	qnrS
N13/CHL	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
N21/TET	16	> 64	S83L, D87Y	S80I	GyrA83, GyrA87		ParC80
N18/KAN	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
Beech martens (n = 7)
26/CEF	> 32	> 64	S83L, D87N	S80I, E84G	GyrA83, GyrA87		ParC80, ParC84	aac(6’)-Ib-cr
29/CEF	> 32	> 64	S83L, D87N	S80I, E84G	GyrA83, GyrA87		ParC80, ParC84	aac(6’)-Ib-cr
31/CEF	> 32	> 64	S83L, D87N	S80I, E84G	GyrA83, GyrA87		ParC80, ParC84	aac(6’)-Ib-cr
K34/CHL	32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
K108/TET	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
K29/CHL	32	> 64	S83L, D87N	S80I, A108V	GyrA83, GyrA87		ParC80
K29/CEF	2	> 64	S83L		GyrA83			qnrB
Common squirrels (n = 5)
S141/TET	4	> 64	S83L		GyrA83			qnrS
S145/CHL	32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
S136/CHL	8	> 64	S83L	G78C	GyrA83		ParC80
S43/KAN	8	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
S42/CHL	32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
Badgers (n = 4)
B100/TET	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
B8/TET	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
B100/KAN	> 32	> 64	S83L, D87N	S80I, E84G	GyrA83, GyrA87		ParC80, ParC84
B101/KAN	> 32	> 64	S83L, D87N	S80I, E84G	GyrA83, GyrA87		ParC80, ParC84
European roe-deers (n = 2)
D89/KAN	2	> 64	S83L	S80R	GyrA83		ParC80	aac(6’)-Ib-cr
D90/KAN	2	> 64	S83L	S80R	GyrA83		ParC80	aac(6’)-Ib-cr
Red-deer (n = 1)
J175/CHL	8	> 64	S83L		GyrA83			qnrS
Rat (n = 1)
R60/KAN	16	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80
Grapevine snail (n = 1)
8/CEF	> 32	> 64	S83L, D87N	S80I	GyrA83, GyrA87		ParC80

^a ciprofloxacin, ^b nalidixic acid.

The isolates obtained from the red foxes were mostly (72.7% of strains) characterized by the presence of two mutations in the gyrA gene and two mutations in the parC gene with a wide range of MIC for ciprofloxacin (from 8 to ≥ 32 mg/L). In all these strains, the same types of mutations were detected: mutation S83L and D87N in the gyrA gene and mutation S80I in the parC gene. One strain (28/CEF) with a ciprofloxacin MIC of 8 mg/L had a single mutation in the gyrA gene. However, in two isolates with a ciprofloxacin intermediate phenotype (MIC equal to 2 mg/L), no mutations were found in any of the analyzed genes. In these three strains, PMQR genes (qnrS or qnrB) were detected.

Nine out of the ten mink isolates tested showed simultaneously two mutations in the gyrA gene and a single mutation in the parC gene. All mutations in the ParC domain of topoisomerase IV consisted in the substitution of serine for isoleucine (S80I). In the GyrA domain, the S83L and D87N mutations were present simultaneously in eight strains. However, one isolate had the S83L and D87Y mutations (substitution of aspartate for tyrosine), but the MIC range was also wide (from 8 to ≥ 32 mg/L). One strain (N17/TET) with MIC equal to 8 mg/L showed no changes in any of the genes tested but had the qnrS gene (Table 1).

The E. coli strains isolated from the martens exhibited a high ciprofloxacin MIC of 32 mg/L, except for one strain. Only this isolate (K29/CEF) with an intermediate phenotype to ciprofloxacin had a single mutation in the gyrA gene; however, this strain was also characterized by the presence of the qnrB gene. The remaining isolates were characterized by the simultaneous presence of three or four mutations in the studied genes. All the mutations in the GyrA domain of topoisomerase II involved the substitution of leucine for serine (S83L) and the substitution of asparagine for aspartic acid (D87N). In the ParC domain of topoisomerase IV, three strains had simultaneous S80I and E84G (glutamate to glycine) mutations (these strains were also characterized by the presence of PMQR aac(6’)-Ib-cr genes), one strain had S80I and A108V mutations (alanine-108 to valine substitution), and two isolates had a single isoleucine substitution in place of serine (S80I).

Three of the five investigated squirrel isolates showed two mutations in the gyrA gene and a single mutation in the parC gene simultaneously. The S136/CHL strain had single substitutions in both genes: S83L in gyrA and G78C (glycine to cysteine) in parC, respectively. Additionally, a single isolate with a MIC value equal to 4 mg/L had one mutation in the gyrase subunit connected with the occurrence of the qnrS gene.

All the badger isolates showed high MIC values for ciprofloxacin (from 16 to > 32 mg/L). All these isolates had substitutions S83L and D87N in the gyrA gene and S80I in the parC gene. However, two strains showed additional mutations in the ParC domain - E84G. None of these strains had PMQR genes.

Both strains obtained from the roe-deer showed an intermediate profile to ciprofloxacin, and each of them had one S83L mutation in the gyrA gene and one S80R (serine to arginine) mutation in the parC gene as well as the aac(6’)-Ib-cr gene.

Single E. coli strains isolated from the rat and snail showed the presence of three typical mutations: two in the gyrA gene and one in the parC gene with MIC 16 and > 32 mg/L respectively. In contrast, a red deer isolate with MIC equal to 8 mg/L had a single mutation only in the GyrA domain of gyrase and PMQR gene qnrS.

Results of MAS-PCR

Additionally, we conducted a control study to detect the most common mutations in the GyrA domain at codon 83 and 87 and in the ParC domain at codon 80 and 84. The results obtained from the amplicon sequencing and the results from MAS-PCR were confirmed in almost all the tested strains (Figs S1 and S2). However, differences were observed in four isolates. In the case of 28/CEF (isolate from a red fox), the MAS-PCR amplification detected only a single mutation in the gyrA gene, while the sequencing of the amplicons also revealed a mutation in codon 87. In the case of N17/TET (isolate from a mink), although MAS-PCR indicated the presence of a mutation in codon 80 of the parC gene, after the sequencing, no mutation was detected either at the same site or in close proximity. In strain S136/CHL (from a squirrel), the MAS-PCR amplification indicated the presence of mutations in codon 83 of the gyrA gene and codon 80 of the parC gene, while the sequencing confirmed the mutation in the gyrase domain, and the mutation in the parC gene was detected in the adjacent codon No 78. In the strain isolated from a marten (K29/CHL), we found double mutations in the gyrA and parC regions. However, one mutation in the latter region was atypical (A108V); therefore, it was not detected by the MAS-PCR amplification.

Discussion

According to the World Health Organization, drug resistance in microorganisms is currently one of the most important global health problems in the world. Due to the complexity of resistance mechanisms and their multidirectional impact, there is a need to monitor all elements included in the One Health concept, including the environment, especially in the aspect of antimicrobials, which may soon be the drugs of last resort.

Such antibiotics are represented by FQs¹⁹. Research on the level of resistance to FQ in Enterobacterales, especially E. coli, Salmonella spp., and Klebsiella spp., is mainly focused on food-producing animals²⁰ and humans^14,15,20,21. Bacterial resistance to FQs is particularly often observed in intensive poultry and pig production^22–24 and in clinical infections²⁵. However, FQs-resistant E. coli has been discovered in environmental samples as well^26–28. Moreover, the presence of FQ residues in the natural environment or animal environments seems to be common^29,30.

In Poland, problem of FQ- resistance most often concerns poultry farms^13,31. According to EFSA report³² Poland is also a country with the highest combined resistance to ciprofloxacin and cefotaxime (4.9% and 3.4% in broilers and turkeys, respectively) among the 31 countries studied. Therefore, such a high percentage of resistance to QNs translates into the results of our recent studies conducted among free-living animals in Poland^33–36. We showed phenotypic resistance to ciprofloxacin at a level of about 30% in bats and carnivores and at a slightly lower level (about 13%) in the other groups of animals. In contrast to our results, studies conducted earlier by Wasyl et al.¹⁸ showed only the presence of eight phenotypically FQ-resistant E. coli strains, which constitutes only 1, 44% of the pool of the free-living animals tested. A similarly low percentage of FQs-resistance among E. coli strains isolated from free-living animals was shown by Guyomard-Rabenirina et al.³⁷ and Kaspersen et al.³⁸. This discrepancy may be related to two reasons: it is possible that the level of FQs-resistance was underestimated, as in the study conducted by Wasyl et al.¹⁸, where only PMQR genes were tested. The other reason may result from differences in the isolation procedure: in our studies, we cultured the samples directly on media supplemented with antibiotics (i.a. cefotaxime, ), which allowed the selective isolation of multidrug-resistant strains, including those with co-resistance to FQs. Interestingly, many studies have confirmed differences in the level of drug resistance between E. coli strains isolated in the non-selective procedure (directly on MacConkey agar without antimicrobials) and E. coli strains isolated in the selective isolation procedure, most often by selective isolation of ESBL/AmpC-producing strains^23,39.

Such a high level of phenotypic resistance to FQs among E. coli isolated from various species of free-living animals in Poland prompted us to investigate the mechanisms responsible for this phenomenon. In the case of 30 strains (71.4%), for which the MIC value ranged from 8 to > 32 mg/L, there were two most frequently occurring mutations in gyrA (Ser(83)Leu and Asp(87)Asn) and usually a single mutation in the parC gene (Ser(80)Ile). Additionally, in the case of one strain isolated from a mink, there was a mutation in the gyrA gene responsible for the substitution with another amino acid at position 87 (Asp(87)Tyr), which probably did not affect the reduction of the level of resistance to ciprofloxacin (MIC = 16 mg/L). A similar difference in the mutation pattern was shown in another strain isolated from a squirrel, in which a mutation was found in the parC gene at position A108V. This strain, with MIC = 32 mg/L, was generally characterized by double mutations in the QRDR; therefore, it is not clear whether the change in the mutation site affected the resistance to ciprofloxacin. On the other hand, studies conducted by Wang et al.⁴⁰ on clinical E. coli strains with the MIC for ciprofloxacin > 32 mg/L showed that three of the isolates tested had an identical mutation pattern as in the K29/CHL strain. A similar pattern was also shown by other authors⁴¹.

In clinical isolates, more than 90% are serine-substitution mutants. The dominance of the serine mutation is explained by the fact that the substitution in this residue has little effect on the catalytic activity of topoisomerase II and IV^42,43.

Five strains (three isolated from the martens and two isolated from the badgers) that had MIC > 32 mg/L showed an additional mutation in the parC gene, which is also the most common type of mutation in the QRDR region (Glu(84)Gly). This result confirms that, similarly to other studies, double mutations in both genes are correlated with a higher MIC value. This also confirms that the mutation pattern occurring in E. coli of wildlife origin resistant to high concentrations of FQs is mostly the same as among strains isolated from humans and domestic animals over many years^9,17,23.

Free-living animals are generally not subjected to targeted therapy; therefore, other factors can influence the occurrence of mutations. One of the most important determinants is the anthropogenic input into the environment⁴⁴. Numerous studies have shown that even sublethal doses of antibiotics, including FQs, in the environment can generate E. coli mutations, especially in the GyrA region⁶. Moreover, a retrospective analysis of antibiotic residues in various environments showed that the predicted no-effect concentration (PNEC), was significantly exceeded in most of the environmental samples tested (municipal, hospital, and industrial wastewater and surface water) in the case of ciprofloxacin³⁰. The presence of ciprofloxacin in such a high concentration in wastewater can be partly explained by the fact that 15–50% of ciprofloxacin is excreted in urine and feces as an unchanged drug⁴⁵. Also, metabolites of ciprofloxacin may have antibacterial activity similar to or lower than other FQs. In the treatment of animals, enrofloxacin is mainly used, which is usually excreted in urine and feces in the form of an active metabolite - ciprofloxacin (enrofloxacin is de-ethylated)⁴⁶. Since FQs are quite stable, they can concentrate, causing selection pressure on the microbiota of free-living animals.

It is generally accepted that bacterial resistance acquired either by horizontal exchange or chromosomal mutations is associated with increased fitness costs, which is unfavorable for bacteria. However, studies have shown that E. coli bears little or no costs of transferring mutations determining substitutions S83L and D87N in GyrA and S80I in ParC^47,48. Studies conducted by the Nair et al.⁴⁹ have shown that the presence of mutations alleviates the fitness costs of plasmids probably by reducing the level of plasmid replication and by silencing the expression of some high-costs plasmid genes. This causes positive selection of strains with mutations and drug resistance plasmids and ensures stable and longer maintenance of such a drug-resistant population in the environment, compared to the wild-type population⁴⁹. This may also partially explain the fact that strains with resistance to high concentrations of FQs are usually multidrug-resistant, as in our study, because they are better hosts for “storing” multidrug-resistant plasmids^33–36.

Our studies also showed that high resistance to FQs in E. coli strains can be determined simultaneously by at least two mechanisms: mutations in QRDR combined with the presence of single PMQR genes. In our studies, we showed only two strains without any mutations that had a single qnrB gene, and these strains were characterized by reduced sensitivity to ciprofloxacin (MIC = 2 mg/L). In the remaining ten isolates, the PMQR gene accompanied at least one mutation in the GyrA region, which caused an increase in MIC from 2 to 8 mg/L. These results confirmed the regularity associated with the determination of reduced susceptibility to QNs in the case of the occurrence of PMQR genes, also among strains isolated from free-living animals. Moreover, we showed only the presence of genes that are also most common among strains isolated from humans, farm animals, and companion animals, which confirms the dominance of the qnrS, qnrB, and aac(6’)-Ib-cr genes in strains isolated from different sources in the worldwide and in Poland^{14,23,31,50–52}.

The presence of even one PMQR gene in combination with a single QRDR mutation increases the MIC for ciprofloxacin to clinical resistance, which not only excludes the possibility of using FQs but also reduces or completely eliminates the synergistic effect obtained through combined application of these drugs with aminoglycosides or beta-lactams⁵³. Moreover, the presence of a bifunctional variant of aminoglycoside acetyltransferase - AAC(6’)-Ib-cr is also responsible for modification the aminoglycoside molecule, causing bacterial resistance to the most commonly used aminoglycosides (kanamycin, amikacin, and tobramycin). This gene may be transferred along with other FQ resistance genes, and often occurs together with genes encoding beta-lactamases⁵⁴.

In our research, we also used an alternative method (MAS-PCR) to detect the most common point mutations in E, coli. The results of the sequencing and MAS-PCR were not the same only in the case of four isolates However the high efficiency of detection of mutations in the gyrA and parC genes indicates the validity of using the MAS-PCR method for screening in a quick analysis of the genetic mechanisms of fluoroquinolone resistance in E. coli isolated from wildlife animals, especially that these mutations are the most common among environmental E. coli isolates⁵⁵. The limitations of this method include the inability to detect mutations in the GyrB and ParE domains, which are also occasionally noted among E. coli isolates. Moreover, this method does not detect mutations in other codons of the tested genes (as in K29/CHL isolate) and may give erroneous results in identifying mutations in neighboring codons (as in S136/CHL isolate).

Our study also has some limitations. Our analysis was performed on strains with an MIC of at least 2 mg/L, while it is known that the presence of PMQR genes is associated with the MIC value from 0.06 to 1 mg/L for ciprofloxacin and from 8 to 32 mg/L for nalidixic acid in E. coli strains²². Therefore, if we had also examined strains with MICs below 2 mg/L, it could have turned out that the occurrence of PMQR genes is much more common among E. coli strains isolated from free-living animals. Another limitation is the fact that we did not analyze all the mechanisms of FQ resistance discovered so far (mutations in the gyrB and parE genes, the presence of efflux pump, or the reduced permeability of cell barriers), which may, although to a lesser extent, also co-generate insusceptibility to FQs.

Moreover, during the review process, we were asked by the reviewer to perform a statistical analysis of the obtained results. We have been analyzing the comment regarding the lack of statistical analysis for quite a long time. After a comprehensive analysis of the results, we as Authors have concluded that we are not able to conduct a reliable statistical analysis based on the generated results for several reasons: In our studies, we analyzed the occurrence of two different resistance mechanisms (PMQR and mutations in the QRDR region). Since these mechanisms often occurred simultaneously in the same strain, it would be very difficult to confirm with statistical analysis, e.g. the correlation between the MIC value and the occurrence of a specific mechanism or mechanisms, especially in the case of strains in which single or double mutations were found in each of the regions and were accompanied by PMQR genes, especially when the MIC value was above 32 mg/L. Second reason was related to the too narrow range of the analyzed MIC values from 2 to above 32 mg/L does not allow for such an analysis. Perhaps if these values were analyzed in wider range of concentrations, i.e. higher than 32, 64, 128, 246 mg/L or lower than 2 mg/L, it would be possible to conduct a reliable statistical analysis. We also believe that a very large cross-section of host species and a small number within the species from which the strains originate are also not conducive to reliable comparative analyses.

On the other hand, it should be emphasized that despite numerous studies on quinolone resistance E. coli, there is still little information and investigated links in QR-E. coli isolated from free-living animals, which successfully represent the environment in the One Health concept. Despite the lack of targeted therapy in these animals, the high level of quinolone resistance indicates a high, persistent level of contamination of the natural environment which should not be underestimated.

Conclusion

We have certainly confirmed that E. coli strains isolated from animals never subjected to targeted therapy are characterized by the same mechanisms of FQ resistance that occur most often among isolates from anthropogenic environments. Moreover, since mutations in the QRDR region can be detected in a significant percentage by both sequencing and the MAS-PCR method, it seems that the combination of PMQR detection and MAS-PCR is sufficient for general monitoring of FQ resistance in the natural environment and obtaining reliable results. This strategy is justified both by the possibility to reduce the cost of monitoring procedure and by its simplicity and repeatability.

The diversity of FQ resistance mechanisms, the influence of many environmental, genetic, and metabolic factors, and the connections between different genes and physiological processes indicate that the intended effect associated with a decrease in global drug resistance may not necessarily be achieved by limiting the use of only some antibiotics. Moreover, bacteria have evolved a wide variety of mechanisms that allow constant and stable persistence of multi-resistant microorganisms in the environment despite the usually high fitness costs associated with acquisition of resistance.

Materials and methods

Selection of bacterial isolates

The research material included 42 selected ciprofloxacin- and nalidixic acid-resistant E. coli strains isolated from different species of free-living animals in Poland during 2014–2018. The material from vertebrate animals was collected after their death by qualified veterinary staff of the State Veterinary Laboratory in Lublin on the basis of the regulation on combating rabies⁵⁶. The area of ​​collected samples covered the region of south-eastern Poland: 51º 15’ 2.012” N 22º 34’ 12.368” E. In the same area, 50 individuals of Helix pomatia were also collected, in accordance with the date specified by the Regional Directorate for Environmental Protection for obtaining snails for commercial purposes.

The phenotypic resistance of these strains has already been described in other publications^34–36. In the case of six tested strains isolated from foxes and martens (Table 1), the occurrence of genes responsible for PMQR was determined partially in a previous study in ESBL-producing strains, denoted in Table 1 as CEF³⁶, but these genes were not determined in the remaining strains (Table 1). However, the mutations in the QRDR region have not been previously analyzed in any of selected strain. Quinolone-resistant strains were obtained from rectal swabs from the animal species listed in Table 1, according to protocols described previously^34–36. The drug resistance of the isolates was determined using the microdilution method and the minimum inhibitory concentration (MIC) value of at least 2 mg/L for ciprofloxacin, classifying them as intermediate and resistant strains according to the CLSI guidelines. In all the tested strains, the MIC value for nalidixic acid was ≥ 64 mg/L. Detailed MIC values for each isolate are shown in Table 1.

DNA extraction and amplification and sequencing of gyrA and parC genes

DNA for molecular analysis was obtained using a ready-made kit for isolation (Tissue and Bacterial DNA Purification Kit, Eurx, Gdańsk, Poland) from 24-hour bacterial cultures on nutrient agar (BioMaxima, Lublin, Poland) according to the manufacturer’s procedure.

The amplification of the gyrA and parC genes in the QRDR region was performed using PCR Mix Plus (A&A Biotechnology, Gdynia, Poland). Specific primers (Genomed, Warsaw, Poland) and conditions of reactions are presented in Table S1. All the reactions were carried out in the thermal cycler (T100™ Thermal Cycler, Bio Rad, Hercules, California, USA). The amplicons were purified using a ready-made Clean-up kit (A&A Biotechnology, Gdynia, Poland) and sent for sequencing to Genomed S.A. (Warsaw, Poland). The resulting sequences were translated into amino acid sequences of gyrA and parC using NCBI’s BLAST. Next, the amino acid changes were analyzed by comparison with wild-type GyrA (GenBank accession no. NP_416734) and ParC (GenBank accession no. NP_417491) of E. coli K-12 (GenBank accession no. NC_000913).

Multiple allele specific PCR (MAS-PCR)

MAS-PCR was performed to detect the most common mutations located at codons 83 and 87 in the gyrA gene and at codons 80 and 84 in the parC gene in fluoroquinolone-resistant strains. Allele-specific primers were used for detection of wild-type gyrA and parC mutations. UspA gene coamplification was performed as an internal control. The MAS-PCR procedure was performed in accordance with the protocol described by Onseedaeng and Ratthawongjirakul⁵⁷. DNA obtained from E. coli ATCC 25,922 was the positive control, while sterile distilled water was used as the negative control. The amplicons were visualized on a 2% agarose gel containing the SimplySafe dye (Eurx, Gdańsk, Poland).

Amplification of PMQR genes

The analysis of the occurrence of PMQR genes was performed based on six genes: qnrA, qnrB, qnrC, qnrS (encoding gyrase-protection repetitive peptides), qepA (encoding efflux pumps), and aac(6’)-Ib-cr (encoding aminoglycoside and quinolone inactivating acetyl-transferase). Gene amplification was performed using PCR Mix Plus (A&A Biotechnology, Gdynia, Poland). Specific primers (Genomed, Warsaw, Poland) and conditions of reactions are presented in Table S1. The sequence of the applied research procedures is presented in Figure S3.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

MO: Conceptualization, Funding acquisition, Investigation, Methodology, Software, Writing—original draft, Writing—review and editing. AN: Conceptualization, Supervision, Investigation, Methodology, Software, Writing—original draft, Writing—review and editing. PZ: Data curation, Resources, Writing—original draft, Writing—review and editing. AT: Methodology, Software, Writing—original draft, Writing—review and editing. AHS: Investigation, Software, Writing—original draft, Writing—review and editing, ATr: Investigation, Writing—original draft, Writing—review and editing.

Funding

This research was funded by an internal project of the University of Life Sciences in Lublin: “Isolation, determination of the phenotypic and genetic profile of drug resistance, and genotyping of Escherichia coli strains isolated from the red fox (Vulpes vulpes)”. Grant number WKD/MN-1/19 financed by the Ministry of Science and Higher Education.

Data availability

The datasets used and/or analysed during the current study available in the RepOD repository (https://repod.icm.edu.pl/) under link: 10.18150/XYT5OV. Contact person: marcelina.osinska@up.lublin.pl.

Declarations

Competing interests

The authors declare no competing interests.