Susceptibility of Methicillin-Resistant and -Susceptible Staphylococcus aureus Isolates of Various Clonal Lineages from Germany to Eight Biocides

ABSTRACT

Few studies have been conducted on the susceptibility of bacteria to biocides. A total of 182 methicillin-resistant and -susceptible Staphylococcus aureus isolates collected from healthy or diseased humans and animals in Germany were included in the present study. Sixty-three isolates of animal origin and 119 human isolates were tested for their MICs to eight biocides or heavy metals by the broth microdilution method. The MIC₅₀ and MIC₉₀ values of human and animal isolates were equal or differed by not more than 1 dilution step, and statistical analysis revealed that differences between MICs of human and animal isolates were not significant. However, when taking into account the multilocus sequence type (MLST), a strong tendency (P = 0.054) to higher MICs of silver nitrate was detected for clonal complex 398 (CC398) isolates from humans compared to those from animals. Furthermore, a comparison of MIC values from isolates belonging to different clonal lineages revealed that important human lineages such as CC22 and CC5 exhibited significantly (P < 0.05) higher MICs for the biocides chlorhexidine, benzethonium chloride, and acriflavine than the main animal lineage sequence type 398 (ST398). Isolates with elevated MIC values were tested for the presence of biocide and heavy metal tolerance-mediating genes by PCR assays, and the following genes were detected: mepA (n [no. of isolates containing the gene] = 44), lmrS (n = 36), norA (n = 35), sepA (n = 22), mco (n = 5), czrC (n = 3), smr (n = 2), copA (n = 1), qacA and/or -B (n = 1), qacG (n = 2), and qacJ (n = 1). However, only for some compounds was a correlation between the presence of a biocide tolerance gene and the level of MIC values detected.

IMPORTANCE Biocides play an essential role in controlling the growth of microorganisms and the dissemination of nosocomial pathogens. In this study, we determined the susceptibility of methicillin-resistant and -susceptible S. aureus isolates from humans and animals to various biocides and heavy metal ions and analyzed differences in susceptibilities between important clonal lineages. In addition, the presence of biocide or heavy metal tolerance-mediating genes was investigated. We demonstrated that important human lineages such as CC22 and CC5 had significantly higher MIC values for chlorhexidine, benzethonium chloride, and acriflavine than the main farm animal lineage, ST398. In addition, it was shown that for some combinations of biocides and tolerance genes, significantly higher MICs were detected for carriers. These findings provide new insights into S. aureus biocide and heavy metal tolerance.

INTRODUCTION

Staphylococcus aureus is a frequent colonizer of the nasal vestibules of humans and also of a large number of animal species (1, 2). Beyond asymptomatic carriage, S. aureus (in particular, methicillin-resistant S. aureus [MRSA]) regarded as one of the most important human nosocomial pathogens worldwide and are able to cause clinical conditions ranging from mild skin infections to life-threatening invasive infections (3). In animals, there are a variety of diseases caused by S. aureus, such as mastitis, botryomycosis, and urinary tract infections (4). Even though farm animals are often carriers of MRSA, they are only rarely infected. The frequent occurrence of MRSA, mainly of clonal complex 398 (CC398), in livestock and occasionally in humans exposed to livestock is a serious public health concern since transmission between animals and humans cannot be excluded (1, 4, 5).

To control the growth of microorganisms and the dissemination of nosocomial pathogens, biocides play an essential role and are used extensively for many topical and hard-surface applications (6, 7). The quaternary ammonium compounds (QACs) acriflavine, alkyldiaminoethyl glycine hydrochloride, benzalkonium chloride, and benzethonium chloride are membrane-active agents targeting the cytoplasmic membrane of bacteria. They are used as antiseptics and disinfectants for many clinical purposes, but also for hard-surface cleaning and deodorization (7, 8). Chlorhexidine, a bisbiguanide, is one of the most widely used active ingredients in antiseptic products and has long been used for its disinfectant and preservative properties. Of note, the biocide is included in decolonization strategies to decrease MRSA rates in hospitals (9). Chlorhexidine damages the outer cell layers and attacks the cytoplasmic membrane of bacteria (7). It has long been known that heavy metal ions such as Ag⁺, Co²⁺, and Zn²⁺ develop antimicrobial activity, and some compounds have been used for centuries (7, 10). Silver compounds are in use especially for the treatment of burns, chronic wounds, and eye infections, while nano-silver is used as a coating substance, e.g., for medical devices, food contact materials, or cosmetic products (11). In contrast, copper sulfate, zinc oxide, and much less frequently zinc chloride are used in livestock production as supplements to animal feed to reduce microbial growth (12). As a mode of action, an interaction of heavy metals such as silver salts with specific groups in enzymes and proteins or direct damage to bacterial membranes is assumed (7, 13).

While resistance to antibiotics is common, tolerance to biocides is believed to occur more rarely due to the multiplicity of targets within the bacterial cell (14). However, there are reports of MRSA with decreased susceptibility to various biocides or heavy metals, including benzalkonium chloride, chlorhexidine, and zinc chloride (15, 16). Tolerance to biocides is mainly a result from alterations in the cell envelope, enhanced efflux pump activity, or, notably in staphylococci, the acquisition of plasmid-mediated genes (7, 14).

Therefore the aims of the present study were (i) to determine the susceptibility of methicillin-resistant and -susceptible S. aureus isolates to various biocides and heavy metal ions, (ii) to analyze differences in susceptibilities between important clonal lineages, and (iii) to identify genetic determinants involved in biocide or heavy metal tolerance.

RESULTS AND DISCUSSION

Results from biocide and heavy metal susceptibility testing.

The results from susceptibility testing are shown in Table 1. The MIC values of the isolates ranged between ≤2 and 32 μg/ml acriflavine, ≤0.06 and 0.25 μg/ml alkyldiaminoethyl glycine hydrochloride, ≤2 and 64 μg/ml benzalkonium chloride, ≤1 and 4 μg/ml benzethonium chloride, 0.12 and 4 μg/ml chlorhexidine, 512 and 8,192 μg/ml copper sulfate, 2 and 16 μg/ml silver nitrate, and 32 and >8,192 μg/ml zinc chloride. Unimodal distributions of MICs were detected for most biocides and heavy metal ions, with the exception of benzalkonium chloride and zinc chloride, for which a bimodal (or multimodal) distribution was found. A bimodal distribution is considered to be indicative for the presence of an antibiotic-resistant (or less-susceptible) subpopulation of bacteria, typically due to mutational or acquired mechanisms of resistance (17). With regard to biocides, inconsistent findings have been reported. For most combinations of bacteria and biocides, unimodal distributions of MIC values were obtained. Even in the presence of genes known to mediate tolerance to biocides or heavy metals (e.g., qac genes in S. aureus), a unimodal distribution pattern of MICs was detected (17, 18). However, bimodal distributions were detected for specific combinations, such as S. aureus and triclosan susceptibility, Enterobacter and chlorhexidine susceptibility, or triclosan susceptibility of Escherichia coli and Enterobacter (17, 19).

TABLE 1.

Distribution of Staphylococcus aureus MIC values for biocides and heavy metal ions^a

^aACR, acriflavine; ADH, alkyldiaminoethyl glycine hydrochloride; BKC, benzalkonium chloride; BEN, benzethonium chloride; CHX, chlorhexidine; COP, copper sulfate; SIL, silver nitrate; ZKC, zinc chloride. Asterisks indicate the number of isolates exhibiting a MIC that is greater than the highest concentration tested or less than or equal to the lowest concentration tested. The white and light-colored areas represent the tested ranges of biocides.

In the present study, the MIC₅₀ and MIC₉₀ values of almost all biocides or heavy metals were equal or differed by not more than 1 or 2 dilution steps. Only slight differences were detected between MIC₅₀ and MIC₉₀ values of human and animal isolates (Table 1). Slightly higher MIC₉₀ values of animal isolates were found for benzalkonium chloride, benzethonium chloride, and copper sulfate, whereas isolates collected from humans exhibited higher MIC₅₀ values of silver nitrate and higher MIC₉₀ values of acriflavine (Table 1). In a study recently published by Morrissey et al. (17), S. aureus epidemiological cutoff values (ECOFFs) were set for the biocides chlorhexidine (8 μg/ml) and benzalkonium chloride (16 μg/ml) to differentiate a wild-type from a non-wild-type MIC phenotype. According to these ECOFFs, none of the isolates of the present study was considered non-wild type for chlorhexidine, with MIC₅₀ and MIC₉₀ values lower than those in the comparative study. In contrast, three isolates from the present strain collection from Germany were classified as benzalkonium chloride non-wild type, and MIC_50/90 values were in a comparable range (17). However, it has to be taken into account that ECOFFs were only derived from human S. aureus isolates and did not include MICs of isolates from animals (17). Furthermore, analysis of 1,602 clinical S. aureus isolates by Furi and coworkers did not uncover a clear indication for chlorhexidine and benzalkonium chloride ECOFF values (18). For the other biocides and heavy metals tested, a classification of isolates as wild type or non-wild type was hampered by the lack of proposed ECOFF values. However, ECOFFs estimated by visual inspection of MIC distributions indicated values of ≥64 μg/ml for acriflavine, ≥0.5 μg/ml for alkyldiaminoethyl glycine hydrochloride, ≥8 μg/ml for benzalkonium chloride, ≥8 μg/ml for benzethonium chloride, ≥6 μg/ml for chlorhexidine, ≥4,096 for copper sulfate, ≥32 μg/ml for silver nitrate, and ≥512 μg/ml for zinc chloride for this strain collection.

Comparison of results from the present study to those of further studies is problematic, since different concentrations of biocides and heavy metals were tested and/or the methodologies used for MIC determinations were not identical (15, 20, 21). However, in a study from the United States, a broth dilution method according to CLSI guidelines was also used to determine chlorhexidine MICs of 829 MRSA isolates from nursing home residents. The values were 4-fold higher than those of the German isolates (22). Results from a study on 95 MRSA and 164 MSSA isolates from Malaysia revealed higher MIC values for benzethonium chloride, benzalkonium chloride (range of 3.9 to 15.6 μg/ml for both) and chlorhexidine (range of 10.3 to 20.7 μg/ml) than isolates of the present study (23). Similar ranges of MIC values (4 to 16 μg/ml) were detected in a study by Randall and coworkers, who investigated the susceptibility to silver nitrate of staphylococci collected between 1997 and 2010 in hospitals throughout Europe and Canada (13). Another study described chlorhexidine MIC values from 219 methicillin-susceptible S. aureus (MSSA) and 82 MRSA isolates that were collected in three African countries. There was a close similarity between the observed MIC range and the MICs of 90% of the isolates compared to the values found in the present study (24). In another study from Denmark, MICs of 43 S. aureus isolates obtained from farm animals were determined (12). The ranges of MICs for benzalkonium chloride and chlorhexidine were in good accordance with our results from animal isolates. However, MIC values for copper sulfate (lower) and zinc chloride (higher) differed slightly from MICs obtained in the present study.

Comparison of MIC values between isolates from animals and humans and between different clonal lineages.

To detect differences in the susceptibilities of isolates from humans and animals and between isolates belonging to different clonal lineages, statistical analysis was performed. Although slight differences in MIC_50/90 values from human and animal isolates were observed for some of the biocides (Table 1), statistical analysis revealed that MICs of human and animal isolates did not differ significantly (P > 0.05) (Fig. 1). Taking into account the MLST genotype, the results showed a strong tendency (P = 0.054) to higher MIC values of silver nitrate for human CC398 isolates compared to animal CC398 isolates. Since silver salts are in use for some clinical purposes, one might speculate that the selection pressure imposed by the use of silver led to higher MIC values of human isolates. For the other genotypes (CC5, CC9, and CC22), statistical analysis was not carried out due to the uneven distribution of human and animal isolates within the clonal lineages.

FIG 1.

Comparison of mean MIC values of silver nitrate (SIL), chlorhexidine (CHX), alkyldiaminoethyl glycine hydrochloride (ADH), benzethonium chloride (BEN), benzalkonium chloride (BKC), and acriflavine (ACR) (a) and zinc chloride (ZKC) and copper sulfate (COP) (b) between S. aureus isolates originating from humans and animals.

Analysis of differences in MIC values between clonal lineages (independent of the origin of isolates from humans or animals) revealed that isolates belonging to CC22 and CC5 (clonal lineages that are considered typically human and among the most frequently occurring lineages in Germany) exhibited significantly higher MICs of chlorhexidine (P = 0.0017 and P = 0.0002, respectively), benzethonium chloride (P < 0.0001 and P = 0.002), and acriflavine (P = 0.0002 and P = 0.0024) than isolates of sequence type 398 (ST398) (Fig. 2). The biocides chlorhexidine and benzethonium chloride are frequently used in hospitals, e.g., for MRSA decolonization and as active ingredients in household products (25, 26). Isolates of the lineages CC22 and CC5 also had significantly higher MICs of chlorhexidine (P = 0.0078 and P = 0.0032, respectively) than ST9 isolates. In contrast, isolates of ST9 showed significantly higher MIC values of benzalkonium chloride than isolates of the lineages CC22 (P < 0.0001), CC5 (P = 0.0002), and ST398 (P < 0.0001), and they exhibited higher MIC values of benzethonium chloride than CC5 (P = 0.0196) and ST398 (P < 0.0001) isolates. Furthermore, isolates of CC22 exhibited significantly (P = 0.005) lower MIC values of silver nitrate than isolates of CC5. For the other heavy metals (zinc chloride and copper sulfate) and biocides tested, no significant differences were detected between sequence types. The heavy metals copper and zinc are currently in use as feed additives in livestock farming (27). Zinc, mostly in form of zinc oxide, can be administered to piglets for 2 to 3 weeks postweaning as an antibiotic alternative to prevent colibacillosis, while copper sulfate is used to suppress bacterial action in the gut of farm animals with the objective of maximizing feed utilization (27, 28). Therefore, divergences in the susceptibilities between human and animal isolates or between clonal lineages to zinc chloride might have been expected, but did not come true for this strain collection. Statistically significant differences could also not be seen between isolates from diseased or colonized hosts. In addition, it should be noted that for some sequence types (e.g., ST9), only a limited number of isolates was available. Therefore, the results should be confirmed by including a greater number of isolates of this lineage.

FIG 2.

Comparison of mean MIC values of silver nitrate (SIL), chlorhexidine (CHX), alkyldiaminoethyl glycine hydrochloride (ADH), benzethonium chloride (BEN), benzalkonium chloride (BKC), and acriflavine (ACR) (a) and zinc chloride (ZKC) and copper sulfate (COP) (b) between S. aureus isolates belonging to clonal complexes CC22, CC398, CC5, and CC9. Statistically significant differences (P < 0.05) are indicated by asterisks.

PCR screening for the presence of biocide and heavy metal tolerance genes and localization on plasmids.

Isolates with MICs of ≥4 μg/ml benzalkonium chloride and/or 4 μg/ml benzethonium chloride (see Table S1 in the supplemental material) were screened for the presence of the genes qacA and/or qacB (here named “qacA/B” since the PCR assay detects both genes), qacG, qacH, qacJ, smr, norA, mepA, sepA, and lmrS, which are known to be involved in tolerance to quaternary ammonium compounds. A total of 10 isolates exhibited elevated MICs for both biocides, while 8 isolates from humans showed decreased susceptibility only to benzalkonium chloride (2 isolates) or benzethonium chloride (6 isolates). PCR analysis demonstrated that all isolates harbored mepA, and all except one were positive for the gene lmrS, while norA was detected in 13 isolates, sepA in 4 isolates, and smr in 2 isolates. Single isolates were positive for the qacA/B, qacG, and qacJ resistance genes (Table 2; Table S1). The gene qacH was not detected in our strain collection. Subsequent sequence analysis of the qacA/B PCR amplicon confirmed the presence of qacA in our strain collection. In contrast to the chromosomally located genes norA, lmrS, mepA, and sepA, the gene smr was found to be located on plasmids with estimated sizes of >25 kb in two MRSA isolates (ST225 and CC22, respectively). A plasmid localization of the genes qacA, qacB, and smr was previously reported for the majority of Staphylococcus species isolates (29–31). Nevertheless, the qacA/B gene was detected in the chromosome of a human ST45 MRSA isolate in this strain collection, whereas the genes qacG and qacJ were located on small plasmids of approximately 1.5 to 2.0 kb in the CC80 and ST398 strains. With regard to the chromosomally located genes norA and sepA, frequent occurrence in staphylococci (>90%) has recently been reported (32). In another study of clinical MRSA isolates from the United Kingdom, only 36.7% of isolates were positive for norA (33). In this study, the gene norA was present in 13 out of 18 isolates with elevated MICs. However, whether point mutations in the primer binding sites caused negative results in PCR assays for the chromosome-mediated determinants, as recently assumed by Liu and coworkers (32), remains to be clarified, just like the expression levels of the efflux genes.

TABLE 2.

PCR analysis of isolates and presence of biocide or heavy metal tolerance-mediating genes

Biocide or heavy metal	MIC (μg/ml)	No. of isolates investigateda	No. of isolates positive for geneb
Quaternary ammonium compounds			lmrS	mepA	norA	sepA	qacA/B	qacG	qacH	qacJ	smr
Benzalkonium chloride	≥4	12	11	12	9	3		2			1
Benzethonium chloride	≥4	16	16	16	11	4	1	2		1	2
Acriflavine	≥16	27	21	27	21	19					2
Bisbiguanides			lmrS	mepA	norA	sepA	qacA/B	qacG	qacH	qacJ	smr
Chlorhexidine	≥2	9	8	9	5	3
Heavy metal ions			copA	mco	czrC
Copper sulfate	≥4,096	7	1	5	NT
Zinc chloride	>8,192	3	NT	NT	3

^aSome isolates were included in PCR screening for more than one biocide or heavy metal ion and may therefore be listed repeatedly in this table.

^bNT, isolates were not tested for the tolerance-mediating genes listed above.

Of the 27 isolates with MIC values of ≥16 μg/ml acriflavine, all were positive for mepA, 21 were positive for norA, 21 for lmrS, 19 for sepA, and 2 for smr (Table S1). Only 6 of these isolates (including both smr-positive isolates) showed elevated MICs of benzalkonium chloride or benzethonium chloride, even though the norA, sepA, smr, and qac genes are known to be involved in tolerance to all quaternary ammonium compounds tested (34–36).

In eight out of nine isolates with a MIC of ≥2 μg/ml chlorhexidine, the gene lmrS was detected. Of these, five isolates additionally carried norA and two isolates sepA, and a single isolate was positive for sepA and norA, while neither qacA/B, qacG, qacH, qacJ, nor smr was present in isolates with elevated MICs. However, a correlation between the presence of tolerance-mediating genes and the level of MIC values could not been detected. This is in good accordance with a study by Skovgaard et al., who did not detect an association between the presence of qac genes and altered chlorhexidine susceptibility (37).

Among seven isolates with MICs of ≥4,096 μg/ml copper sulfate, five were positive for the gene mco encoding a multicopper oxidase. One of them carried a plasmid-located copper resistance gene, copA, in addition to mco; however, this isolate did not exhibit a MIC value of copper sulfate higher than those of the remaining isolates tested. For all but one mco-harboring S. aureus isolate and the copA-carrying isolate, a plasmid localization of the genes was detected, with plasmids ranging in size between 15 and 25 kb.

PCR amplicons for the zinc (and cadmium) resistance gene czrC were obtained in all three isolates exhibiting MICs of >8,192 μg/ml zinc chloride (Table 2). In these isolates (two MRSA ST398 isolates and the single MRSA ST9 isolate of avian origin), the gene was found to be located in the chromosome. As previously reported by Cavaco and coworkers, the gene czrC is linked to the staphylococcal chromosome cassette mec (SCCmec) element in MRSA CC398 isolates. Thus, the use of zinc chloride in animal production may lead to coselection of methicillin resistance in S. aureus (16, 28).

For comparison, 15 isolates with MICs of biocides far below the above-mentioned values were subjected to PCR analysis. The gene qacG was detected in a single isolate, while qacA/B, qacJ, and qacH were not present in the selected isolates. However, all isolates harbored the genes mepA and sepA, 13 isolates were positive for lmrS and 8 for copA, and 7 isolates each carried mco and norA.

Comparison of MIC values from isolates with and without tolerance genes.

To investigate differences in MIC values between isolates with and without tolerance genes, statistical analysis was performed. However, significant differences were neither detected for zinc chloride and isolates carrying or not carrying czrC (P = 0.22), nor for copper sulfate and isolates carrying the genes copA (P = 0.35) and mco (P = 0.20). For the bisbiguanide chlorhexidine, isolates carrying sepA, norA, and lmrS showed no significant differences in MICs compared to isolates negative for the genes. In contrast, significantly higher MICs were detected for the quaternary ammonium compounds benzethonium chloride and acriflavine and isolates carrying the genes lmrS (P = 0.03) and norA (P = 0.03), respectively, while there was no difference (P > 0.05) in MICs obtained from isolates positive or negative for the gene sepA (see Table S2 in the supplemental material). Furthermore, the presence of the genes lmrS, norA, and sepA did not result in significantly higher MICs of benzalkonium chloride or alkyldiaminoethyl glycine hydrochloride (P > 0.05). For the other combinations of biocides and genes, a statistical analysis was precluded due to the low number of isolates positive for the genes or due to the presence of the gene (mepA) in all isolates investigated. Hence, these findings show that the presence of biocide or heavy metal tolerance-mediating genes is not always associated with elevated MICs of the compound. Only for some combinations of biocides and tolerance genes were significantly higher mean MICs detected for carriers (Table S2), indicating that the role of these genes and their interaction remain to be elucidated by including a larger number of isolates.

Conclusion.

In our strain collection, slight but statistically significant differences in the susceptibility to biocides and heavy metal ions have been detected between S. aureus isolates belonging to different major clonal lineages in Germany. However, further investigations are needed to elucidate the role of biocide tolerance-mediating genes and their association with elevated or lowered MIC values of the compounds.

MATERIALS AND METHODS

Bacterial isolates.

A total of 182 S. aureus isolates from humans (n = 119), animals (n = 59), and the environment of a veterinary practice (n = 4) were included in the study (Table 3). The isolates were part of the strain collections of the National Reference Centre for Staphylococci and Enterococci, Robert-Koch Institute, Wernigerode, Germany, and the Institute for Food Quality and Food Safety, University of Veterinary Medicine Hannover, Foundation, Hannover, Germany. The collection comprised 166 methicillin-resistant S. aureus (MRSA) and 16 methicillin-susceptible S. aureus (MSSA) isolates belonging to important human and animal clonal lineages in Germany. The MSSA isolates were included for comparison. The human isolates were collected in the years 2010 through 2011 from diseased and healthy people in Germany and included the following clonal lineages: CC5, CC8, CC22, CC30, CC45, CC80, CC121, ST1, ST7, ST9, ST12, ST15, ST25, ST34, ST59, ST88, ST97, ST101, ST152, ST225, ST398, and ST772 (Table 3). Animal isolates were collected over the past 8 years from diseased and healthy animals and from veterinary clinical settings in Germany and belonged to the clonal lineages ST1, ST9, ST20, ST50, ST97, ST133, ST398, ST425, ST504, ST1380, ST1643, and CC30 (Table 3). The isolates were stored at −80°C as cryopreserved cultures and were grown at 35 ± 1°C overnight on Columbia sheep blood agar (Oxoid, Wesel, Germany) prior to susceptibility testing.

TABLE 3.

Clonal lineages and origins of the MRSA and MSSA isolates used in this study

MLST type or clonal complex (n)	spa type(s) (no. of isolates)	Origin(s)	Disease/colonization
ST398 (67)	t011 (33), t034 (23), t571 (3), t108 (1), t109 (1), t1451 (2), t1184 (1), t1197 (1), t6867 (2)	29 poultry, 25 human, 5 pet animal, 1 dairy cow, 3 animal,a 4 veterinary practice interior	Poultry (29 colonized), human (12 diseased, 8 colonized, 5 NDb), pet animal (5 diseased), dairy cow (1 diseased), animal (3 ND), veterinary practice (4)
ST225, CC5 (22)	t003 (15), t002 (2), t067 (1), t214 (1), t311 (1), t481 (1), t7333 (1)	22 human	Human (7 diseased, 3 colonized, 12 ND)
CC22 (19)	t005 (3), t022 (1), t025 (1), t032 (10), t1214 (1), t310 (1), t608 (1), t9062 (1)	19 human	Human (9 diseased, 4 colonized, 6 ND)
CC8 (9)	t008 (6), t197 (1), t334 (1), t068 (1)	9 human	Human (4 diseased, 3 colonized, 2 ND)
ST09 (8)	t1430 (7), t587 (1)	7 poultry, 1 human	Poultry (7 colonized), human (1 diseased)
ST15 (6)	t084 (5), t094 (1)	6 human	Human (5 diseased, 1 ND)
CC45 (5)	t015 (1), t303 (1), t004 (1), t1574 (1), t095 (1)	5 human	Human (5 diseased)
ST133 (4)	t1181 (2), t6384 (1), t1403 (1)	3 wild boar, 1 dairy cow	Wild boar (3 colonized), dairy cow (1 diseased)
CC30 (4)	t018 (1), t019 (2), t1347 (1)	3 human, 1 dairy cow	Human (1 diseased, 1 colonized, 1 ND), dairy cow (1 diseased)
ST97 (4)	t267 (1), t224 (1), t359 (1), t521 (1)	3 human, 1 dairy cow	Human (3 ND), dairy cow (1 diseased)
ST7 (4)	t091 (3), t1867 (1)	4 human	Human (3 colonized, 1 ND)
CC80 (3)	t044 (2), t203 (1)	3 human	Human (3 diseased)
CC121 (3)	t159 (1), t435 (1), t8660 (1)	3 human	Human (1 diseased, 2 ND)
ST59 (3)	t163 (1), t437 (1), t216 (1)	3 human	Human (2 diseased, 1 colonized)
ST1 (3)	t127 (2), t922 (1)	2 human, 1 dairy cow	Human (2 diseased), dairy cow (1 diseased)
ST34 (3)	t089 (1), t166 (1), t2080 (1)	3 human	Human (1 diseased, 1 colonized, 1 ND)
ST101 (2)	t056 (1), t150 (1)	2 human	Human (2 diseased)
ST152 (2)	t595 (1), t355 (1)	2 human	Human (1 colonized, 1 ND)
ST425 (2)	t6368 (1), t6782 (1)	2 wild boar	Wild boar (2 colonized)
ST88 (1)	t1814 (1)	1 human	Human (1 diseased)
ST772 (1)	t657 (1)	1 human	Human (1 ND)
ST1643 (1)	t6385 (1)	1 wild boar	Wild boar (1 colonized)
ST12 (1)	t160 (1)	1 human	Human (1 diseased)
ST25 (1)	t078 (1)	1 human	Human (1 diseased)
ST504 (1)	t529 (1)	1 dairy cow	Dairy cow (1 diseased)
ST20 (1)	t1023 (1)	1 dairy cow	Dairy cow (1 diseased)
ST50 (1)	t518 (1)	1 dairy cow	Dairy cow (1 diseased)
ST1380 (1)	t2873 (1)	1 dairy cow	Dairy cow (1 diseased)

^aAnimal species not documented.

^bND, disease or colonization status not documented.

Susceptibility testing of biocides and heavy metal ions.

Determinations of MICs were performed in single tests in a broth microdilution assay using customized microtiter plates (Merlin Diagnostics, Bornheim-Hersel, Germany). The wells of the microtiter plates were coated with increasing concentrations of vacuum-dried biocides in 2-fold serial dilutions, including the following substances and test ranges: acriflavine, 2 to 256 μg/ml; alkyldiaminoethyl glycine hydrochloride, 0.0625 to 32 μg/ml; benzalkonium chloride, 2 to 256 μg/ml; benzethonium chloride, 1 to 256 μg/ml; chlorhexidine, 0.0625 to 128 μg/ml; copper sulfate, 32 to 8,192 μg/ml; silver nitrate, 0.5 to 64 μg/ml; and zinc chloride, 4 to 8,192 μg/ml. For MIC determinations, inoculum density, growth medium, and incubation times followed the recommendations given in the Clinical and Laboratory Standards Institute (CLSI) documents M07-A9 and VET01-A4 (38). In brief, a single S. aureus colony was inoculated in sterile sodium chloride solution and adjusted to a 0.5 McFarland standard. A volume of 50 μl of this suspension was diluted in 10 ml cation-supplemented Mueller-Hinton broth II (BD Diagnostics, Heidelberg, Germany) to achieve a concentration of approximately 5 × 10⁵ S. aureus CFU/ml. Of this suspension, 50 μl was inoculated into each biocide-containing well and the microtiter plates were incubated at 35 ± 1°C. After 18 ± 2 h of incubation, the microtiter plates were analyzed visually and the lowest concentration preventing visible growth of bacteria was defined as the MIC. For quality control, microtiter plates contained the antimicrobial agent ciprofloxacin (test range, 0.015625 to 4 μg/ml) in addition to the panel of biocides. The quality control strain S. aureus ATCC 29213 was used each test day to test if MICs of ciprofloxacin were in the acceptable range. In addition, the Salmonella enterica serovar Enteritidis strains 7112 (susceptible) and 84482 (with elevated MICs of chlorhexidine, benzalkonium chloride, and acriflavine) were also tested each test day as quality controls. For both strains, MICs for the biocides chlorhexidine, benzalkonium chloride, and acriflavine were previously determined by the broth microdilution and agar dilution methods, and test ranges were compared with previous results (39). Elevated MICs of copper sulfate (≥4,096 μg/ml) and zinc chloride (≥256 μg/ml), for which endpoints of growth were difficult to interpret by broth microdilution due to the color or turbidity of the solved antimicrobial substances in the wells, were confirmed by the broth macrodilution method as described previously (39).

DNA preparation, PCR amplification, and sequence analysis.

S. aureus strains exhibiting elevated MIC values of biocides (isolates with MIC values at the right edge of the distribution; exact MIC values are presented in Results and Discussion) were investigated by PCR amplification for the presence of genes known to confer decreased susceptibility to biocides or heavy metals. For this, total genomic DNA was extracted by using the spin column protocol of the DNeasy blood and tissue kit (Qiagen, Hilden, Germany). PCR amplifications included screening for the presence of the copper resistance genes copA and mco, the zinc and cadmium tolerance-mediating gene czrC, and genes encoding multidrug transporters of the major facilitator superfamily (MFS) (lmrS, norA, and qacA/B), the small multidrug resistance (SMR) family (qacG, qacH, qacJ, and smr), the resistance nodulation cell division family (sepA), and the multidrug and toxic compound extrusion (MATE) family (mepA). All primers used to amplify internal parts of the genes are shown in Table 4, along with the expected sizes of the amplicons and the annealing temperatures. The standard PCR was performed in a volume of 25 μl and included the use of Taq DNA polymerase (Thermo Fisher Scientific, Darmstadt, Germany) under the following conditions: an initial cycle of 95°C for 1 min, followed by 30 cycles of 1 min at 95°C, 1 min at the annealing temperature as specified (Table 4), and 1 min at 72°C, with a final extension step of 72°C for 5 min. The specificity of all PCR amplicons was confirmed by sequencing (Eurofins Genomics MWG, Ebersberg, Germany). For the purpose of comparison, a smaller set of 15 isolates exhibiting lower MICs to all biocides or heavy metals (left edge of distributions) was also included in PCR screening of tolerance-mediating genes.

TABLE 4.

Primer sequences and PCR conditions

Primer	Sequence (5′→3′)	Amplicon size (bp)	Annealing temp (°C)	Source or reference
sepA-f	TGTACTTTCTGGTGCGAT	176	52	This study
sepA-r	GCTCGACTGCAAATATGA
mepA-f	TGCTATCTCTCTAACACTGCCA	657	56	This study
mepA-r	GCGAAGTTTCCATAATGTGC
norA-f	GATTGGTGGATTTATGGCAG	509	56	This study
norA-r	TTGTAATGGCTGGTCGTATC
mco-f	AAATGGCTCCAATGCTCG	418	50	This study
mco-r	ACGGGTGCTTCATACCACTC
lmrS-f	TGATGTCAATGGTTGGACC	462	58	This study
lmrS-r	AATGCGATGGCGATGTAG
smr-f	ATAAGTACTGAAGTTATTGGAAGT	285	50	40
smr-r	TTCCGAAAATGTTTAACGAAACTA
qacA/B-f	ATTCCATTGAGTGCCTTTGC	198	52	41
qacA/B-r	TGGCCCTTTCTTTAGGGTTT
qacH-f	CAAGTTGGGCAGGTTTAGGA	121	52	41
qacH-r	TGTGATGATCCGAATGTGTTT
qacG-fw	CAACAGAAATAATCGGAACT	275	48	42
qacG-rv	TACATTTAAGAGCACTACA
qacJ-fw	CTTATATTTAGTAATAGCG	306	48	42
qacJ-rv	GATCCAAAAACGTTAAGA
czrC-f	TAGCCACGATCATAGTCATG	655	56	43
czrC-r	ATCCTTGTTTTCCTTAGTGACTT
copA-f	CATGCTTTAGGCTTGGCAAT	662	55	43
copA-r	TCTTCTGGCATGAGTTGTGC

Southern blotting.

Southern blot hybridization studies were performed with plasmid DNA and EcoRI- and BglII-digested genomic DNA of all isolates positive for one of the PCR-screened biocide or heavy metal tolerance genes. The gene probes consisted of the PCR-amplified internal fragments of the genes and were nonradioactively labeled by using the PCR digoxigenin (DIG) probe synthesis kit according to the manufacturer′s recommendations (Roche Diagnostics, Mannheim, Germany).

Data evaluation and statistical analysis.

The primary aims of this study were the exploration of MICs and the investigation of differences in MIC values of MRSA and MSSA isolates for eight different biocides. Differences in MICs depending on the origin of isolates (animal or human), their clonal lineage (four lineages analyzed: ST398, CC22, CC5, and ST09), and the presence of specific resistance genes and their origins from diseased or colonized persons/animals were investigated. For this, the MICs were initially log transformed to base 2. Two-way analysis of variance (ANOVA) models were used to analyze the effects of origin and clonal lineage as well as of origin and disease or colonization status of the host. The effects of the presence or absence of specific tolerance-mediating the genes were analyzed with the one-way ANOVA model. Due to the exploratory nature of the experiments, no multiple adjustments were performed and comparison-wise P values were reported. Statistical analyses were carried out using SAS software, version 9.3 (SAS Institute, Cary, NC).