Characterization of coagulase-negative staphylococci and macrococci isolated from cheese in Germany

Cheese, especially ripened varieties, harbor a very complex and heterogeneous microbiota. In addition to the desired microorganisms (starter cultures) added during cheese production, potentially harmful bacteria may also enter the production chain. Regarding the latter, the focus of this study was on coagulase-negative staphylococci (CNS) and Macrococcus caseolyticus . Both are known to harbor a variety of genes coding for antibiotic resistance, including mec A, mec B, mec C, and mec D. Coagulase-negative staphylococci or macrococci carrying such genes or other virulence factors should not be present in cheese. Cheese samples (101 in total) were collected from retail sources. Coagulase-negative staphylococci and M. caseolyticus were isolated utilizing selective agars, and species were identified by phe-notypical tests and partial sequencing of the sod A gene. The results allowed identification of 53 CNS strains and 19 M. caseolyticus strains. Among the CNS, 11 isolates of Staphylococcus saprophyticus and one Staphylococcus epidermidis isolate were obtained. Both species are potential human pathogens and may thus adversely affect the safety of these food products. Screening for antimicrobial resistance was performed by application of disc diffusion tests, a gradient strip-test, and 14 different PCR tests. Evidence for methicillin resistance (by either positive disc diffusion assay for cefoxitin or by mec PCR) was found in CNS isolates and M. caseolyticus (9 isolates each). Regarding other virulence factors, no genetic determinants for coagulase or the most common staphylococcal enterotoxins sea , seb , sec , sed , and see were detected in any of the CNS or M. caseolyticus isolates by PCR testing. In conclusion, the presence of facultatively pathogenic CNS and carriers of genes for antibiotic resistance in both groups of microorganisms, especially mec genes, and the respective food safety issues need further evaluation and surveillance.


INTRODUCTION
Cheese, especially the ripened varieties, harbor a very complex and heterogeneous microbiota.In addition to the desired microorganisms, such as the starter cultures added during cheese production, which stem from the brining bath or the ripening process (Bockelmann and Hoppe-Seyler, 2001;Mounier et al., 2006;Montel et al., 2014;Stavropoulou et al., 2018), detrimental bacteria may also enter the processing chain (Stavropoulou et al., 2018;Hammer et al., 2019).This study focused on CNS and Macrococcus caseolyticus, which can be opportunistic pathogens if they carry or express genes for virulence factors and antibiotic resistance.
Regarding the CNS, the presence of Staphylococcus saprophyticus in cheese as a potential human pathogen needs attention (Hammer et al., 2019).Further, while M. caseolyticus is not currently considered a human pathogen, it appears to be an emerging pathogen in veterinary medicine (Schwendener et al., 2017;Li et al., 2018;MacFayden et al., 2018;Schwendener et al., 2019;Ramos et al., 2021).
Coagulase-negative staphylococci and M. caseolyticus are known to harbor a variety of genes encoding antibiotic resistance, including mecA, mecB, mecC, and mecD (Baba et al., 2009; International Working Group on the Classification of Staphylococcal Cassette Chromosome Elements, 2009;Fontes et al., 2013;Chajęcka-Wierzchowska et al., 2015;Schwendener, et al., 2017;Schwendener, et al., 2019;Becker, 2021;da Silva Abreu et al., 2021).In addition, it has been shown that bacteria are able or at least have the potential to exchange genes with another species of the same genus or even a species of another genus (Alonso et al., 2001;Resch et al., 2008;Becker et al., 2018;Schwendener and Perreten, 2018).Therefore, when CNS or macrococci (as well as other bacteria) carrying genes coding for antibiotic resistance are present in cheese or any other foods, they can potentially cause harm to the consumer.
As shown in the literature cited in an earlier study (Hammer et al., 2019), CNS are capable of producing staphylococcal enterotoxins.In terms of M. caseolyticus, however, 10 strains from Turkish fermented sausage (Çetin and Tuncer, 2016) and 1 from camel milk (Njage et al., 2013) were previously shown to be incapable of enterotoxin production owing to a lack of respective genes.
The objective of this study was to characterize CNS and M. caseolyticus strains isolated from cheeses sold in Germany in terms of virulence factors and potential antibiotic resistance.

MATERIALS AND METHODS
No animals were used in this study, and ethical approval for the use of animals was thus deemed unnecessary.

Cheese Samples and Isolation of Staphylococci and Macrococci
Cheese samples (101 in total) were obtained from retail sources, focusing on producers (15 in total) for which brine bath samples were examined in a previous study (Hammer et al., 2019) and staphylococci and macrococci were found to be abundant.From each sample, 10 g (ratio rind/curd 1:10) was homogenized in double buffered peptone water [peptone water 20.0 g/L (wt/vol) (Oxoid), 3.5 g/L (wt/vol) Na 2 HPO 4 , 1.5 g/L (wt/vol) KH 2 PO 4 , (both Merck)] prewarmed to 37°C.A 0.1-mL aliquot as well as 4 × 0.25 mL aliquots were spread-plated onto either rabbit plasma fibrinogen Baird-Parker agar (start of the study) or Brilliance MRSA 2 agar (after change of laboratory practice owing to availability of agar) (both Oxoid) and incubated for 24 h at 37°C.Presumptive staphylococcal and macrococci colonies (one of each morphological type from each plate) were streaked onto blood agar [Columbia blood agar base supplemented with 5% (vol/vol) defibrinated sheep blood (both Oxoid)] and incubated for 24 h at 37°C.

Species Identification
Presumptive macrococci or CNS strains from cheese were isolated on blood agar and were identified as described earlier by phenotypical and genotypical testing (Hammer et al., 2019).Briefly, Gram staining, catalase test, hemolysis, coagulase test (PCR), API 32 ID STAPH (bioMérieux), and partial sequencing of the sodA gene (429 bp) were used to identify the isolates as CNS or micrococci, using described previously methods (Poyart et al., 2001;Hammer et al., 2019).The pro-posal for reassignment of Staphylococcus fleurettii and Staphylococcus sciuri to a new genus Mammaliicoccus (Madhaiyan et al., 2020) was not considered in this study.

Random Amplified Polymorphic DNA PCR
Random amplified polymorphic DNA (RAPD)-PCR was done to detect clonal relationships and thus avoid multiple isolation of strains with identical fingerprints for further study.Furthermore, RAPD-PCR fingerprinting was performed to compare strains obtained from cheese in this study to isolates obtained from brine baths used to salt the cheeses from the same producers in an earlier study (Hammer et al., 2019).The RAPD-PCR amplifications were performed with oligonucleotide primers OLP11 (5′-ACGATGAGCC-3′), OLP13 (5′-ACCGCCTGCT-3′), OPR-1 (5′-TGC-GGGTCCT-3′), and OPR 16 (5′-CTCTGCGCGT-3′), as described by Williams et al. (1990).The PCR amplifications were carried out with 20 ng of DNA [extracted as described in Hammer et al. (2019)] in a total volume of 15 µL containing 0.3 µmol of primers using Eppendorf Hot Master Mix.After thermal cycling (95°C for 3 min; 35 cycles of 95°C for 30 s, 40°C for 30 s, and 72°C for 120 s; 72°C for 10 min; 4°C), DNA fragments were separated by electrophoresis in 1.5% agarose gel containing 0.1 µg/mL ethidium bromide.Fingerprint patterns were then visualized and documented under UV light.

Antibiotic Resistance Testing
Screening for antibiotic resistance of strains was performed by application of the disc diffusion assay as described earlier (Hammer et al., 2019).Antibiotic compounds and loads were as follows: cefoxitin (30 µg), enrofloxacin (5 µg), erythromycin (15 µg), gentamicin (10 µg), penicillin G (10 U), and tetracycline (30 µg) (Oxoid).For evaluation, breakpoints provided by the Clinical and Laboratory Standards Institute (2014) or the European Committee on Antimicrobial Susceptibility Testing (2020) were used as far as available (neither organization provided breakpoints for all isolated bacterial species).
To estimate the MIC for cefoxitin, a gradient striptest (Etest, bioMérieux) was applied according to the manufacturer's instructions.Staphylococcus aureus ATCC 29213 with an expected MIC of 1 to 4 µg/mL was used as a cefoxitin-resistant positive control.For the specific lot of test strips used in this study, the actual MIC for this control strain was 3 µg/mL, which was subsequently used as the breakpoint value indicating resistance to this compound.
Although strains belonged to the same species in some cases, they did not show any obvious clonal relationship on the basis of RAPD-PCR fingerprint patterns (data not shown).
As reported and discussed previously (Poyart et al., 2001;Hammer et al., 2019), the discriminatory power of sodA sequencing is superior to phenotypical testing utilizing the API 32ID STAPH miniaturized test kit.None of the M. caseolyticus strains could be identified by the API test, and only 1 of 16 strains of S. equorum was correctly identified by this test.Mismatches in identification when comparing the API 32ID miniaturized test kit results with sodA sequencing-based identification (numbers in parentheses) also occurred for S. carnosus (2), S. fleurettii (2), and S. sciuri (1); whereas, 100% congruent results were obtained for S. saprophyticus (n = 17) and S. xylosus (n = 11).
The high number of S. saprophyticus isolates still needs further consideration.As reported earlier (Hammer et al., 2019), S. saprophyticus ssp.saprophyticus is a biosafety-level 2 microorganism and should not be present in cheese.Discrimination of the 2 proposed subspecies, where subspecies S. saprophyticus ssp.bovis is classified as a biosafety level 1 microorganism, remains difficult.As already reported, phenotypic tests for nitrate reductase, pyrrolidonyl arylamidase, and d-ribose fermentation as distinguishing characteristics between the subspecies according to Hájek et al. (1996) and the respective UK standard (Public Health England, 2014) did not perform as well as sodA sequencing and rpoB sequencing (Hammer et al., 2019).The proposal of Madhaiyan et al. (2020) for unification of both subspecies is thus supported by our observations as well.Staphylococcus epidermidis is also classified as a biosafety-level 2 microorganism.Though well known as a colonizer of human skin, its significance for nosocomial infections in terms of biofilm formation in catheters or on implants is still emerging (Otto, 2012;Oliveira et al., 2018).In addition, its capability of forming biofilms is also a cause for hygienic problems in the food industry (Zou and Liu, 2020).

Antimicrobial Resistance Testing
All CNS and M. caseolyticus isolates were tested for antibiotic resistance by applying disc diffusion assays and PCR tests (Table 2).For strains showing evidence of methicillin resistance, the Etest was also applied to determine the MIC values for cefoxitin.In total, 14 PCR tests were applied for testing for the presence of genes encoding antibiotic resistance.None of the following genes were detected in isolates of blaDHA,blaFox,ermA,ermB,mecC,tetM,and vanB.Of the CNS strains, 42 showed a positive PCR result for tetK, and 8 were positive for tetL.The tetK and tetL genes encode efflux pumps responsible for pumping tetracycline out of the bacterial cell and have also been determined to occur in CNS from foods of animal origin in previous studies (Resch et al., 2008;Chajęcka-Wierzchowska et al., 2015).However, only the S. xylosus (n = 11) strains showed resistance against tetracycline in the disk diffusion assay.In contrast, 14 CNS isolates were phenotypically resistant to penicillin G, but only 1 of these carried the blaTEM and 2 the blaZ gene.Coagulase-negative staphylococci showing evidence for methicillin resistance (being positive for mecA, mecB, or mecD or in the cefoxitin disk diffusion assay) are listed in Table 3.Interestingly, both S. fleurettii strains positive for mecA and 1 isolate each of S. saprophyticus and S. xylosus positive for mecD, were negative in the disk diffusion assay for cefoxitin.The MIC values were 2.0 and 0.5 µg/mL for S. fleurettii, 2.0 µg/mL for S. saprophyticus, and 3.0 µg/mL for S. xylosus; these values were below or directly at the breakpoint of 3.0 µg/mL.Carriage of the mecA gene by S. fleuretti is in accordance with phylogenetic work identifying this species as having an ancient origin and being the origin of the spread of the gene to at least other staphylococcal species (Tsubakishita et al., 2010;Rolo et al., 2017).
Screening for methicillin resistance by use of the cefoxitin disc diffusion assay showed positive results for 5 CNS (none of these carried a mec gene) and 7 strains of M. caseolyticus (6 of these carried the mecD gene).In addition, all of these M. caseolyticus isolates showed resistance to penicillin G in the disc diffusion assay.Genetic determinants for antibiotic resistance in M. caseolyticus (number in parentheses) were detected for blaTEM (3), mecD (7), tetK (12), and tetL (5).Although 12 strains carried genes for tetK, with 5 of these also carrying genes for tetL, only one exhibited phenotypic resistance, which was similar to the results for CNS.Several strains showed resistance to multiple antibiotics (Table 4).One isolate carried genes for mecD, tetK, and tetL and phenotypically showed resistance to cefoxitin, penicillin G, and tetracycline.Three isolates showed a positive disc diffusion assay result for cefoxitin, with a MIC for cefoxitin of 16.0 µg/ mL, and the mecD gene was consistently detected.For 3 other isolates, results for cefoxitin disc diffusion assay and presence of mecD gene were consistent, the MIC for these, however, were varying: 8.0, 4.0, and 1.5 µg/ mL, respectively.One strain showed a MIC of 8.0 and a positive cefoxitin test but no mecD gene, and another had a MIC of 2.0, a negative cefoxitin test, and the presence of the gene.To overcome these inconsistencies, it could be helpful to define an own reference strain of M. caseolyticus for determination of breakpoints for MIC and disc diffusion assay.Interestingly, no mecB gene was detected in the isolates, although Becker et al. (2018) postulated that M. caseolyticus was the ancient source of this gene.According to Schwendener et al. (2017), frequent detection of mecD genes can be expected, which the results of this study confirmed.These authors identified 2 genomic resistance islands for mecD in M. caseolyticus, both containing an integrase gene.The potential for broad host dissemination   (Schwendener and Perreten, 2018).Accordingly, the Macrococcus group needs continuous surveillance owing to its emergence as a potential carrier and vector of mec genes (Becker, 2021) on the one hand, and its application as starter culture (Ramos et al., 2021) on the other.

Genetic Determinants for Enterotoxin Production
Genetic determinants for the most common staphylococcal enterotoxins sea, seb, sec, sed, and see were not detected in any CNS or M. caseolyticus isolates.As discussed earlier (Hammer et al., 2019), Rodrigues et al. (2017) found an absence of the above-mentioned staphylococcal enterotoxin genes in several CNS species, whereas the presence of such genes was reported by other authors (Vernozy-Rozand et al., 1996;Oliveira et al., 2010;Casaes Nunes et al., 2016).For M. caseolyticus, only strains isolated from meat products (n = 10, from Turkish sausages) and 1 sample of camel milk have been tested for the presence of staphylococcal enterotoxin genes, showing no evidence to date for staphylococcal enterotoxin production capability by this species (Njage et al., 2013;Çetin and Tuncer, 2016).

CONCLUSIONS
Cheese microbiota should not contain organisms that have pathogenic potential or carry antibiotic resistance.The genes encoding common staphylococcal enterotoxins were not present in the CNS group of bacteria and the M. caseolyticus species investigated in this study.Among the CNS, a high number of S. saprophyticus was isolated, which needs further consideration in terms of this microorganism's potential as a human pathogen.Genes coding for antibiotic resistance, including mecA and mecD, were detected in several isolates of the CNS group, while mecD was shown to be present in M. caseolyticus.These findings imply a potential food safety issue because a potential for broad-host dissemination of the mecD gene exists with M. caseolyticus.
Klempt et al.: STAPHYLOCOCCI AND MACROCOCCI IN CHEESETable 2. Summary of results for detection of genes for antibiotic resistance and phenotypic tests for CNS and Macrococcus caseolyticus isolates from cheese in Germany 1 assay: FOX = cefoxitin; PEN = G; TET = tetracycline.Genes are in italics.
Klempt et al.: STAPHYLOCOCCI AND MACROCOCCI IN CHEESE

Table 1 .
, the dominant species in spontaneously fermented cheeses were S. equorum, S. saprophyticus, and S. xylosus.The Klempt et al.: STAPHYLOCOCCI AND MACROCOCCI IN CHEESE Origin, number, and summarized resistance pattern of isolates of CNS and Macrococcus caseolyticus 1

Table 3 .
Isolates of CNS from cheese in Germany showing multiple phenotypic and genotypic indications for antibiotic resistance and evidence for methicillin resistance (positive tests for either FOX, mecA, mecB, or mecD) 1 2 of the resistance gene was shown in this connection

Table 4 .
Macrococcus caseolyticus isolates from cheese in Germany showing multiple phenotypic and genotypic indications for antibiotic resistance 1 Disc diffusion assays: FOX = cefoxitin; PEN = penicillin G; TET = tetracycline.MIC for FOX determined with Etest.Genes are in italics.ND = not done.Etest was done only for strains showing evidence for methicillin resistance. 3