Comparison between Some Phenotypic and Genotypic Methods for Assessment of Antimicrobial Resistance Trend of Bovine Mastitis Staphylococcus aureus Isolates from Bulgaria

Simple Summary The aim of the study was to assess the resistance of bovine mastitis Staphylococcus aureus isolates to antimicrobials by comparison of laboratory methods and statistical analysis. For this, 546 mastitis milk samples at 14 farms from 9 districts in Bulgaria were examined. A total of 92 Staphylococcus aureus strains were isolated and identified. The results showed high resistance to sulfadimethoxine (87%), followed by resistance to penicillin (33.7%), erythromycin (13%), streptomycin (8.7%), tetracycline (6.5%) and gentamicin (1.1%). The comparison of the methods demonstrated more than 90% agreement for 9 tested antimicrobial drugs, hence proving reliability of results from monitoring of resistance. Considering the detected discrepancies for some of isolates, an integral evaluation through standard and modern molecular methods for Staphylococcus aureus is recommended. The results from this study add to the microbiology laboratory experience and strategies for mastitis prevention and control programmes. Abstract The aim of this study was to assess the resistance of bovine mastitis S. aureus isolates from farms in Bulgaria to different classes of chemotherapeutic drugs by comparison of some phenotypic and genotypic methods by means of Cohen’s kappa statistics. The study comprised 546 milk samples from subclinical and clinical mastitis at 14 farms from 9 districts in the country. A total of 92 Staphylococcus aureus strains were isolated from tested samples and identified by nuc PCR. The results demonstrated high levels of resistance to sulfadimethoxine (87%), followed by resistance to penicillin (33.7%), erythromycin (13%), streptomycin (8.7%), tetracycline (6.5%) and gentamicin (1.1%). The comparison of both phenotypic tests with respect to 9 antimicrobials revealed strong agreement with kappa coefficient 0.836. An almost complete agreement was evidenced between phenotypic resistance to penicillin and blaZ gene presence, to methicillin with mecA gene, to tetracycline with tet genes, but the agreement between erythromycin resistance and erm genes presence was moderate. This study was the first to demonstrate discrepancy between the behaviour to cefoxitin in the disk diffusion test and oxacillin in the MIC test for an isolate shown to carry the mecA gene in the subsequent genetic analysis. Considering the detected discrepancies for some of isolates, an integral evaluation through phenotypic and molecular methods for monitoring of antimicrobial resistance of Staphylococcus aureus is recommended.

A total of 546 milk samples from cows with subclinical and clinical mastitis were selected in 2019-2021 after screening for increased cell elements content with Mastitis test NK (Biovet, Ivanovice na Nané, Czech Republic) following manufacturer's instructions. In addition to the increased somatic cell content in the tested quarters, inclusion criteria for sampling subclinical mastitis milk were both the absence of concurrent disease and dry period in the animal. Exclusion criterion for clinical mastitis sampling was the pre-treatment of the animal with either systemic or intramammary chemotherapeutic drug. From positive milk quarters, individual samples were collected for bacteriological examination in line with approved practices [25]. In the lab, 10 µL from each milk sample were routinely inoculated on tryptic soy agar (TSA, Sigma-Aldrich, Darmstadt, Germany) supplemented with 5% defibrinated sheep blood (TSBA), MacConkey agar (Himedia, Mumbai, India) for detection of Gram-negative bacterial microflora and mannitol salt phenol-red agar (Sigma-Aldrich) as a selective medium for presumptive identification of S. aureus. Plates were incubated aerobically for 24-48 h at 35 • C. Identification of Staphylococcus spp. was based on the Vet. Sci. 2022, 9,401 3 of 14 absence of growth on McConkey agar, growth on blood agar, mannitol salt phenol-red agar and colonial characteristics of the isolates on the blood agar-size, morphology, presence of haemolysis. Fermentation of mannitol on the selective medium was also considered. After obtaining pure culture, Gram staining, catalase/oxidase tests and coagulase production in tube coagulase test with rabbit plasma (Hi Media, Mumbai, India) were carried out.

PCR Identification of S. aureus Isolates
DNA from presumptive S. aureus isolates was isolated with QIAmp DNA mini kit (Quiagen, Hilden, Germany) according to the protocol for Gram-positive bacteria. For cell wall lysis, the bacterial depot was treated with buffer containing 20 mg/mL lysozyme, 20 mM Tris HCl with pH 8.0, 2 mM EDTA and 1.2% Triton. DNA concentration was determined by UV-Vis spectrophotometry (Cary 60, Agilent Technologies, Santa Clara, CA, USA). Extracted DNA was stored at −20 • C until analysis. Isolates were identified through PCR amplification of species-specific sequences from the nuc gene of S. aureus with primers generating 359 bp amplicons [26]. The sequence of primers was as follows: au-F3 TCGCTTGCTATGATTGTGG, au-nucR GCCAATGTTCTACCATAGC. The reaction volume of 20 µL contained 0.25 µM of each primer, 1x PCR buffer, 1.5 mM MgCl 2 , 0.2 mM dNTPs, 1 U/rxn Taq polymerase (Canvax, Córdoba, Spain), 1 µL DNA. PCR temperature details are presented in Table S1. PCR products were run in 2% agarose gel (Sigma-Aldrich, Burlington, MA, USA) stained with 10 ng/mL ethidium bromide (Sigma-Aldrich, Merck KGaA, Saint Louis, MO, USA).
Minimum inhibitory concentrations (MIC) were determined on Sensititre mastitis plate format CMV1AMAF with two-fold dilutions of selected antimicrobial drugs presented in Table S2. Plates were inoculated with 50 µL standardised inoculum prepared according to the former method and diluted in cation adjusted Mueller-Hinton broth to achieve a final concentration of approximately 5 × 10 5 cfu/mL. After sealing with adhesive foil provided by the manufacturer, plates were aerobically incubated at 35 • C for 24 h. After the incubation, results were read manually, considering first positive growth controls (3 wells) for each of isolates. MICs were interpreted as the lowest antimicrobial drug concentration inhibiting visible growth. The reference strains S. aureus ATCC 29213 and Escherichia coli ATCC 25922 were used as controls for each batch of plates. The interpretation of results from both phenotypic methods was carried out according to CLSI criteria [27] 2.4. Detection of Genes Encoding Resistance to Antibiotics S. aureus isolates resistant and sensitive to beta-lactams, tetracyclines and erythromycin were analysed for presence of mecA, blaZ, tetM, tetK, ermB and ermC genes by means of conventional PCR. The sequences of primers expected PCR product size and their origin are listed in [19,28] Both singleplex and multiplex PCR reactions were performed. In the first variant, the PCR protocol described for nuc gene detection was followed. A multiresistant methicillin-resistant S. aureus DSM 29134 provided by the German Collection of Microorganisms and Cell Cultures GmbH served as positive control.
The multiplex reaction was performed with two primer pairs-blaZ/ermC and blaZ/tetK. The optimum hybridisation temperature of primers was selected through a gradient PCR within the range 49.9-56.1 • C. The reaction included 35 cycles with the following concentrations of reagents for the first primer pair: 0.5 µM of each primer, 1x PCR σyφep, 2 mM MgCl 2 , 0.3 mM dNTPs, 1.5 U/rxn Taq polymerase (Canvax, Spain), 4 µL DNA and molecular biology grade water (Sigma-Aldrich, USA) up to 20 µL. For the second primer pair, reagent concentrations were: 0.5 µM of each primer, 1x PCR σyφep, 1.5 mM MgCl 2 , 0.2 mM dNTPs, 1 U/rxn Taq polymerase, 4 µL DNA and water up to 20 µL Temperature conditions and exposures comprised single denaturation at 95 • C for 5 min, denaturation at 95 • C for 30 ceK, annealing at 55 • C for the first reaction mix and 52 • C for the second reaction mix for 35 s, elongation at 72 • C for 60 s and final amplification at 72 • C for 7 min.

Statistical Analysis
The agreement of results from the behaviour of isolates against chemotherapeutic drugs by the disk diffusion method and MIC method were compared by Cohen's kappa statistic and interpreted according to McHugh [29]. The Mueller-Hinton broth dilution micromethod was used as gold standard for comparisons [30]. The sensitivity and specificity of PCR tests for detection of genes encoding resistance to antibiotics and the corresponding behaviour at the phenotypic level were determined by formulas described by Martin [31]. Cohen's kappa statistic was used to determine the agreement between phenotypic resistance and presence of genes of resistance to beta-lactams, erythromycin and tetracyclines. Statistically significant differences in the resistance against the tested antibacterial drugs in the investigated farms were assessed by ANOVA test, followed by Bonferoni posthoc test. Statistycal evaluation of the results was performed with a specialized software (STATISTICA for Windows 10.0, StatSoft, Inc., Tulsa, Oklahoma, USA).

Primary Identification of S. aureus Isolates
A total of 92 presumptive S. aureus strains were isolated from tested 546 milk samples from cows with subclinical and clinical mastitis. The results from primary identification tests of the staphylococcal isolates are presented in Table S4. With the exception of 3 isolates, all demonstrated haemolytic activity on blood agar with sheep red blood cells: 63/68.5%double haemolysis and 26/28.3%-beta haemolysis. Gram-stained microscopic preparations demonstrated Gram-positive cocci arranged in grape-like clusters. Tests for production of catalase and oxidase were positive and negative, respectively. The pigmentation of colonies, mannitol fermentation on mannitol-salt agar and coagulase production of isolates were variable. A major part of isolates produced colonies with yellowish tint -69/75%, the rest were not pigmented. Seven isolates gave a mild positive mannitol fermentation reaction (7.6%), and another 8 were negative in the free coagulase tube test (8.7%). The strains originated from 14 farms located in 9 administrative districts and 4 regions in Bulgaria (Table S5).

PCR Identification of S. aureus Isolates
Presumptive S. aureus isolates were precisely identified at the species level by conventional PCR. The results demonstrated 359 bp amplification products, specific for the nuc gene of S. aureus in all tested isolates ( Figure 1). Vet. Sci. 2022, 9, x FOR PEER REVIEW 5 of 14 (8.7%). The strains originated from 14 farms located in 9 administrative districts and 4 regions in Bulgaria (Table S5).

PCR Identification of S. aureus Isolates
Presumptive S. aureus isolates were precisely identified at the species level by conventional PCR. The results demonstrated 359 bp amplification products, specific for the nuc gene of S. aureus in all tested isolates ( Figure 1).

Antimicrobial Susceptibility Testing
In this case, 32 (34.8%) of the 92 S. aureus isolates tested by the disk diffusion method demonstrated resistance against selected antimicrobial drugs. Resistant isolates originated from 9 farms in 4 regions of the country. At the farm level, the resistance of S. aureus to tested antimicrobials varied from 0% tо 75%. In 5/14 (35.7 %) of farms, resistance to none of used chemotherapeutics was detected. These farms were located in Pazardzhik. Stara Zagora, Sliven, Burgas and Kardzhali districts, situated in the South central and Southeastern regions of Bulgaria (Тable S5).

Antimicrobial Susceptibility Testing
In this case, 32 (34.8%) of the 92 S. aureus isolates tested by the disk diffusion method demonstrated resistance against selected antimicrobial drugs. Resistant isolates originated from 9 farms in 4 regions of the country. At the farm level, the resistance of S. aureus to tested antimicrobials varied from 0% to 75%. In 5/14 (35.7 %) of farms, resistance to none of used chemotherapeutics was detected. These farms were located in Pazardzhik. Stara Zagora, Sliven, Burgas and Kardzhali districts, situated in the South central and Southeastern regions of Bulgaria (Table S5).

PCR Identification of S. aureus Isolates
Presumptive S. aureus isolates were precisely identified at the species level by conventional PCR. The results demonstrated 359 bp amplification products, specific for the nuc gene of S. aureus in all tested isolates ( Figure 1).

Antimicrobial Susceptibility Testing
In this case, 32 (34.8%) of the 92 S. aureus isolates tested by the disk diffusion method demonstrated resistance against selected antimicrobial drugs. Resistant isolates originated from 9 farms in 4 regions of the country. At the farm level, the resistance of S. aureus to tested antimicrobials varied from 0% tо 75%. In 5/14 (35.7 %) of farms, resistance to none of used chemotherapeutics was detected. These farms were located in Pazardzhik. Stara Zagora, Sliven, Burgas and Kardzhali districts, situated in the South central and Southeastern regions of Bulgaria (Тable S5).
Data about minimum inhibitory concentrations of the 10 chemotherapeutic drugs in the mastitis pathogens plate for S. aureus are presented in Table 1.
In most isolates determined as sensitive in the disk diffusion test, MIC values of chemotherapeutics in the plate were within the reference ranges. There were isolates, whose growth was not inhibited by the highest concentration of an antimicrobial in the plate. For example, no inhibition was detected from penicillin, ampicillin, erythromycin, tetracycline and sulfadimethoxine in 6.5%, 5.4%, 8.7%, 6.5% and 87% of tested isolates. For them, respective MIC 90 values were 2 µg/mL, 4 µg/mL, 1 µg/mL, 1 µg/mL and >256 µg/mL. Table 2 presents the data from the general comparison between the results from the disk diffusion and MIC tests for 9 chemotherapeutic drugs against bovine mastitis S. aureus isolates. The agreement between both tests was 92.39% with kappa coefficient 0.836 [95% confidence interval (CI) 0.719-0.952; standard error (SE) of kappa-0.059].
The results for cefoxitin/oxacillin were the same as for ampicillin, and those for ceftiofur-the same as for penicillin. The comparison of tests with respect to erythromycin is shown in Table S9. The results demonstrated 96.74% agreement with kappa coefficient 0.83(95% CI 0.662-1.000; SE of kappa 0.090). For tetracycline (Table S10), the tests agreed completely: 100%; kappa coefficient 1.000 (95% CI 1.000-1.000; SE 0.000). The results of cephalotin, penicillin/novobiocin and pirlimycin corresponded to those for tetracycline.

Minimum Inhibitory Concentration
The results for cefoxitin/oxacillin were the same as for ampicillin, and those for ceftiofur-the same as for penicillin. The comparison of tests with respect to erythromycin is shown in Table S9. The results demonstrated 96.74% agreement with kappa coefficient 0.83(95% CI 0.662-1.000; SE of kappa 0.090).

Minimum Inhibitory Concentration
The results for cefoxitin/oxacillin were the same as for ampicillin, and those for ceftiofur-the same as for penicillin. The comparison of tests with respect to erythromycin is shown in Table S9. The results demonstrated 96.74% agreement with kappa coefficient 0.83(95% CI 0.662-1.000; SE of kappa 0.090).

Detection of Resistance Genes
The singleplex PCR assays demonstrated amplification products with size 310 bp, 517 bp (Figure 3), 190 bp and 169 bp (Figure 4), corresponding to, mecA, blaZ genes and ermC and tetK genes, respectively, in 2% agarose gel ((Sigma-Aldrich, USA) stained with 10 ng/mL ethidium bromide (Sigma-Aldrich, USA).  of PCR assay for detection of resistance to penicillin were 100% and 96.8%, respectively. Among the 12 isolates resistant to erythromycin, the erm gene was found out in 7, furthermore, it was detected in 2 of susceptible isolates; therefore, the sensitivity of PCR for detected of resistance to erythromycin was 58.3%, and specificity-97.5%. With regard to resistance to methicillin and tetracycline, both the sensitivity and specificity of the reaction were 100%.
The results from the subsequent multiplex PCR analysis corresponded entirely to those of reactions with single primer pairs ( Figure 5). Out of the 29 isolates resistant to penicillin, all possessed the blaZ gene. BlaZ gene was not detected in 61 out of all 63 penicillin-sensitive isolates. The sensitivity and specificity of PCR assay for detection of resistance to penicillin were 100% and 96.8%, respectively. Among the 12 isolates resistant to erythromycin, the erm gene was found out in 7, furthermore, it was detected in 2 of susceptible isolates; therefore, the sensitivity of PCR for detected of resistance to erythromycin was 58.3%, and specificity-97.5%. With regard to resistance to methicillin and tetracycline, both the sensitivity and specificity of the reaction were 100%.
The results from the subsequent multiplex PCR analysis corresponded entirely to those of reactions with single primer pairs ( Figure 5). The agreement between phenotypic resistance to penicillin and the presence of the blaZ gene was 97.83% with kappa coefficient 0.951 (95% CI 0.883-1.000; SE of kappa 0.035). The agreement between phenotypic and genotypic tests for erythromycin was 92.39% with kappa 0.625 (95% CI 0.371-0.878; SE of kappa 0.129); for methicillin and tetracyclinekappa 1.000; 100% agreement.

Discussion
The first step for characterisation of bacterial strains is isolation of pure culture and precise identification by means of contemporary methods with high sensitivity and specificity. The performed primary identification of bovine mastitis S. aureus isolates in this study showed variable results from key tests, e.g., presence of haemolysins, pigment production, mannitol fermentation, rabbit plasma coagulation, which was in line with previous observations of ours [32,33]. In a number of small-scale microbiological labs, the identification of S. aureus at the species level is based on the free coagulase tube test. It is not considered a definitive test, as coagulase-negative S. aureus strains are also isolated, probably due to mutations in the coa gene-an important epidemiological marker for detection of different S. aureus variants [34]. Thus, such S. aureus isolates may be erroneously classified as coagulase-negative staphylococci (CNS), which usually cause self-limiting udder infections [35]. The wrong diagnosis can influence the sufficiency of measures for mastitis control. On the other hand, coagulase-positive staphylococci other than S. aureus (S. intermedius, S. pseudintermedius, S. shleiferi subsp. coagulans) may be also involved in the etiology of mastitis [32,36,37]. In this study, all presumptive S. aureus isolates after basic biochemical tests were identified with conventional PCR based on the nuc gene, encoding thermonuclease, including isolates with atypical behaviour. The selected gene is strongly conservative and may be successfully used for precise identification of S. aureus, as confirmed by other research reports [19,26,38,39]. The agreement between phenotypic resistance to penicillin and the presence of the blaZ gene was 97.83% with kappa coefficient 0.951 (95% CI 0.883-1.000; SE of kappa 0.035). The agreement between phenotypic and genotypic tests for erythromycin was 92.39% with kappa 0.625 (95% CI 0.371-0.878; SE of kappa 0.129); for methicillin and tetracycline-kappa 1.000; 100% agreement.

Discussion
The first step for characterisation of bacterial strains is isolation of pure culture and precise identification by means of contemporary methods with high sensitivity and specificity. The performed primary identification of bovine mastitis S. aureus isolates in this study showed variable results from key tests, e.g., presence of haemolysins, pigment production, mannitol fermentation, rabbit plasma coagulation, which was in line with previous observations of ours [32,33]. In a number of small-scale microbiological labs, the identification of S. aureus at the species level is based on the free coagulase tube test. It is not considered a definitive test, as coagulase-negative S. aureus strains are also isolated, probably due to mutations in the coa gene-an important epidemiological marker for detection of different S. aureus variants [34]. Thus, such S. aureus isolates may be erroneously classified as coagulase-negative staphylococci (CNS), which usually cause self-limiting udder infections [35]. The wrong diagnosis can influence the sufficiency of measures for mastitis control. On the other hand, coagulase-positive staphylococci other than S. aureus (S. intermedius, S. pseudintermedius, S. shleiferi subsp. coagulans) may be also involved in the etiology of mastitis [32,36,37]. In this study, all presumptive S. aureus isolates after basic biochemical tests were identified with conventional PCR based on the nuc gene, encoding thermonuclease, including isolates with atypical behaviour. The selected gene is strongly conservative and may be successfully used for precise identification of S. aureus, as confirmed by other research reports [19,26,38,39].
The testing of antimicrobial susceptibility of S. aureus isolates as well as of other causative agents is essential element of mastitis control strategies [40]. Selected antimicrobials comprised drugs used for treatment of mastitis in our country as well as preparations non-registered for use: pirlimycin or ceftiofur, erythromycin under the form of intramammary infusion, as well as penicillin/novobiocin to investigate the trends in antimicrobial resistance of this bacterial species, isolated from mastitic milk in Bulgaria. In this case, 14 substances were tested in the first method, MICs were determined for 10 antimicro-bials, and comparisons of S. aureus behaviour in both phenotypic tests were made for 9 antimicrobial drugs included in the Sensititre mastitis plate format. The results determined 32 isolates (34.8%) as resistant to at least one of antimicrobial drugs by the disk diffusion test, whereas the MIC test identified 35 isolates (38%); therefore, a strong agreement between the two tests was found out with kappa value of 0.836. The resistance rates of isolates varied among the farms, districts and regions from 0% to 75%, which required a differentiated approach in programmes for mastitis prevention and control. For penicillin and ampicillin, both tests showed an almost perfect agreement with kappa coefficient 0.975. Disagreements referred to two different isolates. One isolate was interpreted as resistant by the disk diffusion test: 20 mm inhibition zone (IZ), but sensitive in the MIC test (0.12 µg/mL), and another isolate-sensitive to ampicillin by the disk diffusion test (39 mm IZ), but resistant by MIC (8 µg/mL). It should be noted that tests were performed under strictly controlled conditions to eliminate factors that would lead to imprecision in detection of results [27]; also, tests were repeated in independent experiments when inconsistent results for respective isolates were obtained. The paradoxical findings can be most probably due to the Eagle phenomenon-the ability of microorganisms to survive and replicate at concentrations higher than their optimum minimum bactericidal concentrationeffect described for S. aureus against penicillin and other beta-lactams [41]. Discrepancies between the disk diffusion test and MIC determined by the E-test were reported by another study [42] having tested the sensitivity of S. aureus isolates from mammary glands of cows and sheep, the authors detected error rates of 3.3% for penicillin and 3.8% for ampicillin. Disagreement was demonstrated by Khalili et al. [43] in testing S. aureus and Enterococcus spp. susceptibility to beta-lactams, glycopeptide antibiotics and clindamycin. Before being reported as sensitive to penicillin, S. aureus strains should be tested for production of beta-lactamase [27]. That is why this study tested isolates with phenotypic resistance and those sensitive to penicillin for the presence of the blaZ gene. The results showed a high sensitivity and specificity of the PCR method for blaZ gene detection compared to phenotypic tests with almost perfect agreement (kappa coefficient 0.951), in line with data from other studies [19,44]. The established 33.7% resistance to penicillin after the genetic analysis (n = 31/92 blaZ positive isolates) was relatively low compared to reported rates from Estonia (61.4%), England (36 to 46%), Korea (38.6 to 78.8%), some farms in China (>60%); northern Ethiopia (>90%), southeastern Brazil (60.7%), Comarca Lagunera region of Mexico (97%) etc. [45][46][47][48][49][50][51]. Lower rates of resistance to penicillin were reported from Chinese dairy farms in Shandong, Jiangsu and Guangdong provinces-24.2% [12], farms in East Poland-23.6% [52], Switzerland-14% [53] and others. In our country, Nikolova et al. [54] have tested the susceptibility of Staphylococcus spp. isolates from two dairy cattle farms in northeastern Bulgaria against different antimicrobials. The authors identified 8 coagulase-positive staphylococci, which were 100% sensitive to tested antimicrobial drugs at farm 1 whereas at farm 2-18.2% were resistant to amoxicillin: a rate comparable to our results.
A surprising finding of this study was the detection of a methicillin-resistant S. aureus isolate from clinical bovine mastitis, which was initially clearly defined as sensitive to cefoxitin by the disk diffusion test on the basis of the 22 mm IZ. A zone of the same size was observed also with the MRSA strain DSM 29134, used as reference strain in the PCR assay, which was also atypical. Subsequent analyses demonstrated resistance to oxacillin with MIC > 4 µg/mL and presence of the mecA gene in both strains. To our best knowledge, this is the first report for isolation of MRSA from a cow with mastitis in Bulgaria, confirmed to carry the mecA gene. Additionally, this clinical isolate showed an unique resistance profile: to penicillin (ampicillin), ceftiofur, streptomycin, gentamicin and tetracycline by the disk diffusion test and susceptibility to amoxillin/clavulanic acid. It was determined as sensitive to erythromycin by both phenotypic tests, but according to the PCR assay, carried the ermC gene. Further detailed studies of this isolate at the molecular level are necessary to establish its SCCmec type as well as full genome sequencing for better understanding of antimicrobial resistance mechanisms. Oxacillin-sensitive mecA positive S. aureus are reported in the literature [55,56], yet the present study is the first to report disagreement between the behaviour to cefoxitin in the disk diffusion test (used as surrogate in oxacillin resistance testing) and that to oxacillin in the MIC test.
With respect to erythromycin, the agreement of results of both phenotypic tests was strong (kappa coefficient 0.839), yet the agreement of phenotypic test and the presence of erm genes was only moderate (kappa coefficient 0.625). From all 12 detected erythromycinresistant isolates (9 in the disk diffusion and MIC tests; another 3 only in the MIC test), seven strains carried the ermC gene whereas the ermB gene was absent in either resistant or sensitive isolates. The observed sensitivity of the PCR method for detection of genes encoding resistance to erythromycin-58.3% was much lower than the 100% sensitivity reported by Gao et al. [19] for detection of ermC/ermB genes. This was probably due to the presence of ermA gene that was not tested in our isolates and is a limitation of this study. Nevertheless, the prevalence of erm genes in different research reports is highly variable. For example, Ghanbari et al. [23] observed the highest prevalence of ermC followed by ermB and ermA genes among tested human S. aureus isolates; in addition, they did not detect any erm gene in two strains with inducible MLS B phenotype. Another study [57] reported 60% prevalence of the ermA gene among erythromycin-resistant S. aureus isolates. The interpretation of resistance to macrolides should be careful and integral as the different resistance mechanisms produce a different phenotypic profile [58]. At the farm level, the resistance to erythromycin varied from 0 to 46.2%, taking into account that in Bulgaria, this antibiotic is not applied as intramammary infusion. The common use of tylosin in veterinary medical practice may have resulted in horizontal transfer of genetic elements conferring cross resistance in S. aureus to this antimicrobial drug class [59]. In addition, resistance was confirmed in both subclinical and clinical mastitis isolates.
The results for the behaviour of isolates to tetracycline agreed almost perfectly between the two phenotypic tests, as well as with the presence of tet genes, among which only the tetK gene was found out. The resistance rate to tetracycline in all tested isolates was low (6.5%) compared to other reports [49,50,60]. In the different farms however, resistance rates varied from 0 to 50% which was not surprising at the background of the broad use of this antibiotic in cattle farming, including for treatment of mastitis. Rates from 0 to 30.8% in farms were obtained also with respect to streptomycin (in the disk diffusion test), used in combination with penicillin as intramammary infusion-an inevitable effect of the selective pressure of antimicrobial drugs on bacteria [18].
A perfect agreement on susceptibility of S. aureus with kappa value 1.000 was obtained from the two phenotypic tests for cephalothin, penicillin/novobiocin and pirlimycin. With regard to ceftiofur, included in Sensititre mastitis plate format, disagreement referred only to one isolate exhibiting intermediate sensitivity in the disk diffusion test and therefore classified as resistant, but then showing susceptibility with MIC = 2 µg/mL. Against sulfadimethoxine, the isolates from the present study showed high resistance rates: 87% were not inhibited at concentrations of 256 µg/mL, conversely to the behaviour of S. aureus strains in the study of Freu et al. [50] reporting 100% susceptibility. In this study, the sensitivity to amoxicillin/clavulanic acid, fluoroquinolones (enrofloxacin) and cotrimoxazole, tested by the disk diffusion method was 100%, contrary to other research data [39,49,61].

Conclusions
Bovine mastitis S. aureus isolates in the present study demonstrated preserved sensitivity to most of tested antimicrobial drug classes, high resistance to sulfadimethoxine and moderately high resistance to penicillinase-sensitive penicillin (33.7%). The comparative analysis between the disk diffusion test used in smaller veterinary diagnostic laboratories and MIC determination by the serial dilution micromethod applied in labs handling a greater number of samples showed more than 90% agreement for 9 tested antimicrobial drugs, hence proving reliability of results from phenotypic monitoring of resistance. By reason of observed discrepancies between the behaviour of some isolates in the phenotypic tests, e.g., against penicillin, ampicillin, cefoxitin/oxacillin and erythromycin, integral evaluation of resistance patterns with inclusion of molecular biological methods is advised. This approach resulted in detection of a cefoxitin-sensitive but oxacillin-resistant mecApositive S. aureus isolate, as well as of few isolates sensitive in the disk diffusion test but resistant in the MIC test and vice versa. We believe that the results from this study will contribute to the experience of microbiological diagnostic labs and strategies implemented in mastitis prevention and control programmes.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/vetsci9080401/s1, Table S1: Nuc PCR temperature conditions; Table S2: Classes of antimicrobial drugs included in the mastitis pathogens testing plate (Trek Diagnostic Systems Ltd., UK), their concentrations and susceptible MIC breakpoints according to CLSI [27]; Table S3: Sequence of primers, targeted genes, expected size of PCR products and their source; Table S4: Primary identification of bovine mastitis Staphylococcus spp. Isolates; Table S5: Distribution of isolates resistant in the Bauer-Kirby disk diffusion test by regions, districts and farms; Table S6: Resistance patterns of bovine mastitis S. aureus isolates resistant in the Bauer-Kirby disk diffusion test by regions, districts and farms; Figure S1: Some resistance patterns depicted by the disk diffusion test; Table S7: Comparison between disk diffusion and MIC methods for testing resistance to penicillin; Table S8: Comparison between disk diffusion and MIC methods for testing resistance to ampicillin; Table S9: Comparison between disk diffusion and MIC methods for testing resistance to erythromycin; Table S10. Comparison between disk diffusion and MIC methods for testing resistance to tetracycline.

Data Availability Statement:
The data presented in this study can be found in the manuscript.