Microbiological Quality and Safety of Fresh Turkey Meat at Retail Level, Including the Presence of ESBL-Producing Enterobacteriaceae and Methicillin-Resistant S. aureus

The aim of this work was to study the microbiological safety and quality of marketed fresh turkey meat, with special emphasis on methicillin-resistant S. aureus, ESBL-producing E. coli, and K. pneumoniae. A total of 51 fresh turkey meat samples were collected at retail level in Spain. Mesophile, Pseudomonas spp., enterococci, Enterobacteriaceae, and staphylococci counts were 5.10 ± 1.36, 3.17 ± 0.87, 2.03 ± 0.58, 3.18 ± 1.00, and 2.52 ± 0.96 log CFU/g, respectively. Neither Campylobacter spp. nor Clostridium perfringens was detected in any sample. ESBL-producing K. pneumoniae and E. coli were detected in 22 (43.14%), and three (5.88%) samples, respectively, all of which were multi-resistant. Resistance to antimicrobials of category A (monobactams, and glycilcyclines) and category B (cephalosporins of third or fourth generation, polymixins, and quinolones), according to the European Medicine Agency classification, was found among the Enterobacteriaceae isolates. S. aureus and methicillin-resistant S. aureus were detected in nine (17.65%) and four samples (7.84%), respectively. Resistance to antimicrobials of category A (mupirocin, linezolid, rifampicin, and vancomycin) and category B (cephalosporins of third- or fourth generation) was found among S. aureus, coagulase-negative staphylococci, and M. caseolyticus isolates.


Introduction
Consumption of turkey meat has increased in recent years due to its characteristics of low cost, high protein content, and low fat content (1.21%) (lower than the fat content of chicken) [1]. However, turkey meat has been involved in outbreaks of Salmonella, Staphylococcus aureus, Campylobacter spp., Clostridium perfringens, and Listeria monocytogenes [2].
A total of 51 fresh turkey meat samples were purchased in Logroño (La Rioja, Spain) from 10 different retailers that were representative of a variety of trade models. The samples were collected between January 2020 and January 2021. The number of samples of each commercial brand was determined according to the place-of-purchase data [29]. All the samples were produced in Spain. Fourteen samples were collected in two different hypermarkets (HA and HB), 35 were collected in seven different supermarkets (SA, SB, SC, SD, SE, SF, and SG) and two were collected in traditional shops (TAs).
The 51 meat samples were evaluated by the ultra-performance liquid chromatography quadrupole time of flight (UPLC-QTOF) method to detect antibiotic residues, as indicated in an earlier study [26]. Doxycycline was found in one sample at levels of 6.6 µg/kg. Antibiotic residues were not detected in the other 50 samples [30].
For the initial microbiological analysis, 10 g of turkey meat were aseptically taken and homogenized using a masticator blender (IUL Instruments, Barcelona, Spain) for 2 min with 90 mL of sterile peptone water (0.1% w/v) (Oxoid, Basingstoke, Hampshire, UK). Decimal dilutions were carried out using the same diluent. The next microbiological analyses were then carried out for Mesophiles, Pseudomonas spp., enterococci, Enterobacteriaceae, staphylococci, Campylobacter spp., and Clostridium perfringens. Mesophile counts were determined on Plate Count agar (Scharlau, Barcelona, Spain) after incubation for 48 h at 30 • C. The enumeration of Pseudomonas spp. was conducted in a chromogenic agar for Pseudomonas To determine the presence of Campylobacter spp., 10 g of turkey meat were homogenized for 2 min in 90 mL of Bolton broth (Oxoid) and incubated at 42 • C for 1 day in a microaerobic atmosphere, followed by streaking on Agar Brilliance CampyCount agar incubated at 42 • C for 2 days under microaerobic conditions.
In addition, a screening was performed to determine methicillin-resistant S. aureus and ESBL-and carbapenemase-producing Enterobacteriaceae. Two grams of turkey meat were placed in flasks containing 50.0 mL of Brain Heart Infusion (BHI) broth (Oxoid) and incubated at 37 • C for 24 h. For the screening of methicillin-resistant S. aureus (MRSA), after incubation, the samples were plated with the streak-plate method in chromID MRSA agar (BioMérieux, Lyon, France) and incubated at 37 • C for 24 h. Presumptive MRSA colonies were selected for further analysis. For the screening of ESBL-and carbapenemaseproducing Enterobacteriaceae, after incubation, the samples were plated with the streak-plate method in chromID ESBL and chromID CARBA SMART agar (BioMérieux) and incubated at 37 • C for 24 h. Presumptive Escherichia coli and Klebsiella pneumoniae were selected, according to the manufacturer's instructions, for further analysis.

Isolation and Identification
From each turkey meat sample and culture media five colonies of the highest dilution that generated growth were randomly selected and isolated. The morphology of suspected colonies was taken into consideration when specific media were used. Isolates were purified in Tryptone Soy agar (Scharlau) and Brain Heart Infusion broth (Scharlau). The purified isolates were maintained at −80 • C. Bacterial identification was carried out by a MALDI-TOF biotyper (Bruker, Daltonik, Bremen, Germany).

Phenotypic Confirmation of ESBL Producers
Further analyses were carried out with isolates from chromID ESBL identified by MADI-TOF as K. pneumoniae and E. coli. Phenotypic confirmation of these ESBL producers was performed using the disc-diffusion method according to the Clinical Laboratory Standards Institute's guidelines [31]. Isolates from other media identified as K. pneumoniae and E. coli were also analyzed.

Phenotypic Confirmation of Methicillin Resistance of S. aureus
The methicillin resistance of S. aureus was confirmed in accordance with the Clinical Laboratory Standards Institute's guidelines [31] by a diffusion-agar assay using cefoxitin (30 µg).

Statistical Analysis
The microbial counts were changed to log CFU/g. Analysis of variance techniques using Duncan's multiple range test was carried out to separate averages and evaluate the three factors that were investigated: microbial group, retailer, and month. The level of significance was determined at p < 0.05. All the tests were conducted with SPSS version 26 software (IBM SPSS Statistics).

Results
Mesophile counts were 5.10 ± 1.36 log CFU/g, with counts in the range 2.3-7.23 log CFU/g. Only two samples showed levels above 7 log CFU/g. No significant differences (p > 0.05) in mesophile counts were observed between samples from hypermarkets and those from supermarkets. Nevertheless, significantly lower counts (p < 0.05) of mesophiles were observed in samples from traditional shops than in those from supermarkets and hypermarkets. No significant differences (p > 0.05) in mesophile counts were found between samples from the two hypermarkets. Significantly lower counts (p < 0.05) of mesophiles were found in supermarket SD than in the other six supermarkets analyzed.
The bacteria identified from the Plate Count agar were mainly one rifampicin (RD, 5 µg) and actic acid bacteria (37.66%), followed by Brochotrix thermosphacta (22.94%) ( Table 1). Pseudomonas spp., Enterobacteriaceae, Micrococcaceae, and enterococci were isolated to a lesser extent (9.09%, 8.23%, 7.79%, and 1.30%, respectively) ( Table 1). In addition, Chryseobacterium spp., Acinetobacter spp., Brevundimonas diminuta, Stenotrophomonas rhizophila, Wautersiella falsenii, Psychrobacter pulomonis, Microbacterium spp., Rhodococcus erythropolis, and Bacillus endophyticus were isolated (Table 1). P. fragi was the predominant Pseudomnas spp. isolated from Plate Count agar (47.62%) ( Table 1). The meat sample in which doxycycline was detected showed mesophile counts of 5.15 ± 0.01 log CFU/g, being the species identified as Brochotrix thermosphacta Rhodococcus erythropolis, Microbacterium liquefaciens, and Microbacterium maritypicum. R. erythropolis and M. maritypicum were not identified in any other sample, while M. liquefaciens was isolated in two other samples. No significant differ-ences (p > 0.05) in mesophile counts were found between the doxycycline-positive sample and those that were negative. The doxycycline levels detected in the positive sample were below the maximum residue limits (MRLs) of antimicrobials in meat, as established by Regulation 37/2010 (100 µg/kg) [33]. Pseudomonas spp. counts below 1 log CFU/g were observed in 15 samples (29.41%). The other 36 samples (70.59%) showed counts between 2.00 log CFU/g and 5.02 log CFU/g, with an average number of 3.17 ± 0.87 log CFU/g. No significant differences (p > 0.05) in pseudomonas counts were found between samples from hypermarkets and those from supermarkets. Nevertheless, significantly lower counts (p < 0.05) of pseudomonas were observed in samples from traditional shops than in those from supermarkets and hypermarkets. No significant differences (p > 0.05) in pseudomonas counts were observed between samples taken in the two hypermarkets. Significantly lower counts (p < 0.05) of pseudomonas were found in supermarket SD than in the other six supermarkets analyzed.
Pseudomonas spp. distribution is shown in Table 2. P. libanensis (31%) and P. extremorientalis (14%) were the prevailing species, followed by P. fluorescens (12%). The meat sample in which doxycycline was detected showed Pseudomonas counts of 2.24 ± 0.24 log CFU/g, being the only species isolated, P. rhodesiae. Table 2. Percentage and number of Pseudomonas spp. isolated from chromogenic agar for Pseudomonas in turkey samples.

Pseudomonas synxantha 3 3
Pseudomonas marginalis 2 2 Pseudomonas cedrina 2 2 Pseudomonas trialis 2 2 Pseudomonas kilorensis 1 1 Pseudomonas proteolítica 1 1 Pseudomonas koreensis 1 1 Total Pseudomonas spp. 100 100 Enterococci counts below 1 log CFU/g were found in 12 samples (23.53%). The other 39 samples (76.47%) displayed counts between 1.30 log CFU/g and 3.28 log CFU/g, with an average number of 2.03 ± 0.59 log CFU/g. No significant differences (p > 0.05) in enterococci were observed between samples from hypermarkets and those from supermarkets. However, significantly lower counts (p < 0.05) of enterococci were found in samples from traditional shops than in those from supermarkets and hypermarkets. No significant differences (p > 0.05) in enterococci counts were found between samples taken in the two hypermarkets. The Enterococcus spp. distribution is shown in Table 3. E faecium was the prevailing enterococci (38.10%), followed by E. faecalis (23.81%) and E. gallinarum (16.67%). In addition, Streptococcus gallolyticus was isolated in 12 samples (23.53% of the samples analyzed). Enterobacteriaceae counts below 1 log CFU/g were found in 13 samples (25.49%). The other 38 samples (74.51%) showed counts between 1.60 and 4.99, with an average number of 3.18 ± 1.00. No significant differences (p > 0.05) in Enterobacteriaceae counts were observed between samples from hypermarkets and those from supermarkets. However, significantly lower counts (p < 0.05) of Enterobacteriaceae were observed in samples from. traditional shops than in those from supermarkets. No significant differences (p > 0.05) in Enterobacteriaceae counts were found between samples taken in the two hypermarkets. Significantly lower counts (p < 0.05) of staphylococci were found in supermarkets SD, SE, SF, and SG than in supermarkets SA, SB, and SC. Table 4 shows the species distribution. Serratia liquefaciens was the dominant specie (16.42%), followed by Hafnia alvei (14.18%) and Escherichia coli (14.18%). In addition, Klebsiella pneumoniae, Moellerella wisconcensis, and Yersinia enterocolitica were isolated. The meat sample in which doxycycline was detected showed Enterobacteriaceae counts of 2.69 ± 0.09 log CFU/g, being E. coli (40%) and K. pneumoniae (60%) the only species isolated. K. pneumoniae was not isolated from Mac-Conkey agar in any other sample. Twenty-three of the 51 turkey samples were positive in chromID ESBL (45.1%). ESBLproducing K. pneumoniae and E. coli were detected in three and 23 samples, respectively. ESBL-producing E. coli were confirmed phenotypically in 22 of 23 samples, while all ESBL-producing K. pneumoniae were confirmed. Both ESBL-producing K. pneumoniae and ESBL-producing E. coli were isolated from the meat sample in which doxycycline was detected. The K. pneumoniae isolates obtained from MacConkey agar in the doxycyclinepositive sample were the ESBL-producing phenotype, while the two isolates of K. oxytoca obtained from doxycycline-negative samples were ESBL-negative. However, none of the E. coli isolates obtained from MacConkey agar showed the ESBL phenotype, although some of the isolates were obtained from samples that were positive in chromID ESBL. Carbapenemase-producing Enterobacteriaceae were not recovered from the chromID CARBA SMART medium.
The antimicrobial resistance phenotype of E. coli isolates is displayed in Figure 1. All 23 E. coli isolates from chromID ESBL were multi-resistant (i.e., resistant to three or more antibiotic classes), with the highest rates of resistances to ampicillin (100%); piperacillin, ceftriaxone, and aztreonam (91.30%); cefpodoxime, gatifloxacin, and tetracycline (82.61); streptomycin (78.26%); cetftazidime (73.91%); and enrofloxacin (69.57%). For+ antimicrobial classes, the highest resistance corresponded to penicillins, cephalosporins, and monobactams. In addition, resistance to colistin was found (8.69%). Of the 14 E. coli isolates from MacConkey agar, 71.43% were multi-resistant. The highest resistance rates were observed against streptomycin (92.86%); ampicillin (78.57%); and piperacillin, tetracycline, and doxycycline (64.29%). None of the isolates showed susceptibility to all of the 36 tested antibiotics. Table 5 shows the antimicrobial resistance phenotype of multi-resistant E. coli isolated from turkey meat. Multi-resistant strains were isolated from samples obtained in supermarkets and hypermarkets. The highest number of multi-resistant E. coli was obtained in hypermarket HB (six isolates). However, no resistant E. coli strain was isolated from a traditional shop (TA).  Of the 14 E. coli isolates from MacConkey agar, 71.43% were multi-resistant. The highest resistance rates were observed against streptomycin (92.86%); ampicillin (78.57%); and piperacillin, tetracycline, and doxycycline (64.29%). None of the isolates showed susceptibility to all of the 36 tested antibiotics. Table 5 shows the antimicrobial resistance phenotype of multi-resistant E. coli isolated from turkey meat. Multi-resistant strains were isolated from samples obtained in supermarkets and hypermarkets. The highest number of multi-resistant E. coli was obtained in hypermarket HB (six isolates). However, no resistant E. coli strain was isolated from a traditional shop (TA).

Species (Number of Isolates) Antibiotic Resistance Phenotype 1 (Number of Isolates) Retailer 6
Klebsiella oxytoca ( Staphylococci counts were below 1 log CFU/g in 14 samples (27.45%). Counts ranged between 1.30 log CFU/g and 4.81 log CFU/g with an average number of 2.52 ± 0.96 log CFU/g. No significant differences (p > 0.05) in staphylococci counts were observed between samples from hypermarkets and those from supermarkets or a traditional shop. No significant differences (p > 0.05) in staphylococci counts were observed between samples taken in the two hypermarkets evaluated. Significantly lower counts (p < 0.05) of staphylococci were observed in supermarkets SD, SE, SF, and SG than in supermarkets SA, SB, and SC. Table 7 shows the Staphylococcus spp. distribution, with S. saprophyticus (31.45%) and S. equorum (13.7% being the dominant species. S. aureus was detected in nine samples (17.65%), being the fourth most often staphylococci isolated (8.1%). Methicillin-resistant S. aureus was detected in four samples (7.84%). Macrococcus caseolyticus was also identified (12.1%) ( Table 6). The meat sample in which doxycycline was detected showed staphylococci counts of 1.3 ± 0.00 log CFU/g, being the only species identified, S. warneri.

Species (Number of Isolates) Antibiotic Resistance Phenotype 1 (Number of Isolates) Retailer 3 (Number of Isolates)
Macrococcus caseolyticus (8) susceptible to all antibiotics tested (4) Table 9 shows the antimicrobial resistance phenotype of methicillin-sensitive and methicillin-resistant S. aureus isolates from turkey meat. Most of the S. aureus isolates (88.89%) and all the methicillin-resistant isolates showed a multi-resistant phenotype. All the S. aureus isolates showed resistance to tetracycline, penicillin, and benzilpenicillin. Resistance to enrofloxacin was observed in 66.67% of the S. aureus isolates. Resistance to amikacin, chloramphenicol, kanamycin, mupirocin, tobramycin, ceftaroline, gentamycin, quinupristin-dalfopristin, rifampicin, and fusidic acid was only observed in 25-50% of the methicillin-resistant isolates, while resistance to clindamycine, erythromycin and tylosine was observe in 75% of these isolates. All the S. aureus isolates were susceptible to linezolid, vancomycin, nitrofurantoin, trimethoprim-sulfamethoxazole, and trimethoprim. Multi-resistant S. aureus were isolated from samples from hypermarkets HA and HB and supermarkets SD, SE, SF, and SG. Table 9. Antimicrobial resistance phenotype of S. aureus from turkey samples.

Specie (Number of Isolates) Antibiotic Resistance Phenotype 1 (Number of Isolates) Retailer 4 (Number of Isolates)
Staphylococcus equorum (7) susceptible to all antibiotics tested (1)  Neither Campylobacter spp. nor Clostridium perfringens was detected in any sample.

Discussion
We observed mesophile counts of 5.10 ± 1.36 log CFU/g in turkey meat. Jaber et al. [17] found higher counts in turkey meat from Moroccco (6.44 log CFU/g). Lower counts were reported by Augustyńska-Prejsnar et al. [34]. (4.25 ± 0.07 log CFU/g). It should be noted that poultry spoilage occurs when mesophile counts reach 8-9 log CFU/g [35], populations that were not reached in the present study. The bacterial load on poultry meat is influenced by the physiological conditions of animals at slaughter, as well as by processing, distribution, and storage circumstances [3].
Among lactic acid bacteria, Lactobacillus spp., Leuconostoc spp., and Carnobacterium spp. are linked with the spoilage of fresh meat [37]. Other authors have also found C. maltaromaticum and C. divergens in fresh meat, with C. divergens being the dominant species, as in the present work [38]. Carnobacterium spp. have been associated with the spoilage of chicken meat [6,39]. We observed that Carnobacterium spp. represented 50.6% of lactic acid bacteria, followed by Lactobacillus spp. (34.5%) and Leuconostoc spp. (8.05%). In addition, Vihavainen et al. [39] reported that C. maltaromaticum and C. divergens were the dominant bacteria in chicken. It should be noted that Lactobacillus spp. has been isolated from broiler feathers and skin, while Leuconostoc spp. and Carnobacterium spp. have been isolated from the plant-processing environment [39].
As in the present study, Raouterella spp. has previously been found in raw turkey and chicken meat, although the earlier study found a different species, Raouterella ornithinolytica, instead of Raoultella planticola [30,34].
Acinetobacter spp. have also been isolated from chicken carcasses; their presence is related to cross-contamination during processing [41]. A. johnsonii, A. lwoffii, and A. guillouiae have been detected in chicken [42]. Chryseobacterium spp. has also been isolated from chicken [12,43]. Psychrobacter spp. was previously reported in chicken meat [16,40,43]. Brevundimonas diminuta was isolated from pork meat [44]. The isolation of B. diminuta may be of concern, as this bacterium is considered an emerging pathogen and an important multidrug-resistant microorganism [45]. In addition, Sterophonomas spp. and Waurtersiella spp. have been isolated from fresh meat [40,43].
As in the present work, Höll et al. [6] isolated Microbacterium spp. and Rhodococcus spp. from chicken meat. In addition, Bacillus spp. has been isolated from fresh meat [40]. In the present work, M. maritypicum and R. erythropolis were only isolated from the sample in which doxycycline was detected. These bacteria have been reported for their antimicrobial resistance [46,47]. Our results suggest that the presence of doxycycline may influence meat microbiota. It should be noted that tetracyclines are usually administered intramuscularly to food-producing animals, having an extended mean residence time in muscles, and consequently there is an extended withdrawal period for these antibiotics [48]. A study found that antimicrobial levels in muscles decreased as the withdrawal period moved forward [48]. Therefore, although the amounts of doxycycline detected in the positive meat sample were low (below the MRLs), large amounts would be present in early stages and could affect animal microbiota, which could be a source of contamination of meat.
Higher Pseudomonas spp. counts have been identified by Augustyńska-Prejsnar et al. [34] in turkey meat (4.29 ± 0.05 log CFU/g) compared to the counts observed in the present research (3.17 ± 0.87 log CFU/g). Pseudomonas spp. are important spoilage bacteria. Some species, such as P. fluorescens, P. fragi, P. lundensis, and P. putid, are often found in spoiled meat [40]. Some authors have reported that P. putida was the most common Pseudomonas spp. isolated from turkey meat, but this species was not isolated in the present work [34]. It is worth noting that P. putida has often been isolated from spoiled meat [40]. On the other hand, in the present work the dominant flora was Carnobacterium spp. rather than Pseumdomonas spp. Similarly, Pseudomonas spp. has been isolated from chicken by other authors [13,49]. Kačániová et al. [49] also isolated P. brenneri, P. proteolytica, and P. fluorescens from chicken meat. Oakley et al. [13] also reported the presence of the following Pseudomonas spp. in chicken: P. libanensis, P. extremorientalis, P. antarctica, P. veronii, P. synxantha, P. marginalis, P. cedrina, P. koreensis, P. brenneri, and P. trivialis. The presence of P. orientalis has also been found in meat by other authors [50]. We also isolated other Pseudomonas spp., including P. rhodesiae, P. azotoformans, and P. kilorensis. A total of 16 different species of Pseudomonas were identified in the present study. Kačániová et al. [49] isolated nine different Pseudomonas spp. from chicken meat. It is worth noting that the main contamination source of Pseudomonas spp. is the processing environment [11].
Enterococci counts below 1 log CFU/g were observed in 12 samples (23.53%). The other 39 samples (76.47%) showed counts between 1.30 log CFU/g and 3.28 log CFU/g. Other authors have reported that enterococci counts are usually present in raw meat at levels between 2-4 log CFU/g [51]. The dominant enterococci found in the present work was E. faecium. However, other authors reported that the predominant enterococci in turkey is E faecalis [52]. Moreover, Aslam et al. [53] did not isolate either E. faecium or E. hirae from turkey meat. Enterococci are often contaminants of poultry meat [20]. Turkey meat may become contaminated with E. faecium and E. faecalis at slaughter. As enterococci are commensals in the gut of poultry, the contamination of carcasses by fecal bacteria can occur if hygienic standards are low [20].
We isolated S. gallolyticus, a non-enterococcal group D streptococci, from 12 samples (23.53%) [54]. This bacterium has been previously reported in turkey feces [55]. As far as we know, there are no previous works on its presence in turkey meat. The isolation of S. gallolyticus may be of concern because this bacterium is an opportunistic pathogen in humans and can cause bacteremia, meningitis, and endocarditis [56]. In addition, the presence of this species has been linked to colon cancer in humans [52].
The presence of Enterobacteriaceae in fresh meat is of particular relevance, since some species are pathogens for humans [57]. In addition, these bacteria have a high deteriorating potential [57]. Higher Enterobacteriaceae counts in turkey meat have been reported by Augustyńska-Prejsnar et al. [34] (3.96 ± 0.03 log CFU/g, compared to 3.18 ± 1.00 log CFU/g in the present study). Augustyńska-Prejsnar et al. [34] observed that the most frequently Enterobacteriaceae isolated in raw turkey was Enterobacter cloacae, followed by Hafnia alvei and Pantoea agglomerans. In contrast, we observed that Serratia liquefaciens was the dominant species, followed by Hafnia alvei and Escherichia coli. Our findings that Serratia spp. is the dominant Enterobacteriaceae agrees with others studies in chicken and turkey meat [6,58]. As in the present work, Höll et al. [6] pointed out that the dominant Enterobacteriaceae in chicken meat were Serratia spp. However, they reported that the largest part of the genus Serratia was represented by S. proteomaculans rather than S. liquefaciens, as was observed in the present work. In addition, S. fonticola has been isolated from poultry [59]. Other authors have also found Kluyvera intermedia, Klebsiella pneumoniae, and K. oxytoca in fresh turkey meat [25,34]. Rahnella aqualis has also been isolated from chicken meat [40]. This species has been linked to the spoilage of pork meat [60]. Buttiauxella warmboldiae and B gavininae have also been isolated from chicken meat [49]. The presence of B. agrestis in fresh meat has also been reported by other authors [40]. Buttiauxella spp. has been associated with meat spoilage [57]. Enterobacter spp. has been found in chicken meat by other authors [49]. Moellerella wisconsensis may cause infections in humans [61]. This bacterium has been isolated from wild birds [62] but its presence in turkey meat has not been previously reported. We only isolated E. coli (40%) and K. pneumoniae (60%) from MacConkey agar in the meat sample in which doxycycline was detected. K. pneumoniae was not isolated from MacConkey agar in any other sample. Our results suggest that the presence of doxycycline could influence the Enterobacteriaceae species, which is dominated by E. coli and K. pneumoniae. As mentioned above, although the amounts of doxycycline detected in the positive meat sample were low (below the MRLs), large amounts would be present in early stages and could affect the animal microbiota, which could be a source of contamination of meat. Further studies are needed to confirm these findings, as there have been a limited number of samples with antibiotic residues.
As in the present work, other researchers have observed a high prevalence of E. coli in turkey meat [17,24]. We observed that 45.4% of the turkey samples showed positive results in chromID ESBL, a lower percentage than that reported by Díaz-Jiménez et al. [25] (84%). The use of the abovementioned medium allows detecting ESBL producers when they are present at low concentrations-particularly E. coli, which is one of the most common ESBL producers [23]. This fact can explain that none of the E. coli isolates obtained from MacConkey agar showed the ESBL phenotype, although some of the isolates were obtained from samples that were positive in chromID ESBL.
We isolated both ESBL-producing K. pneumoniae and ESBL-producing E. coli from the doxycycline positive sample. It should be noted that the K. pneumoniae isolates obtained from MacConkey agar in that positive sample were also the ESBL-producing phenotype. These findings suggest that doxycycline may promote the presence of ESBL-producing K. pneumoniae. In fact, some studies indicate that the use of tetracyclines requires attention, due to the development of the antimicrobial resistance of K. pneumoniae and E. coli [63]. Like Díaz-Jiménez et al. [25], we did not isolate any carbapenemase-producing Enterobacteriaceae from chromID CARBA SMART.
We observed that E. coli isolates from turkey meat showed higher resistance rates than those reported by Díaz-Jiménez et al. [25] for ampicillin (100% vs. 90.2%), while lower rates were found for trimethoprim-sulfamethoxzoale (17.39% vs. 53.7%) and ciprofloxacin (52.17% vs. 53.7%). Higher rates of resistance than those found in the present work have been reported for E. coli isolates from poultry meat for nalidixic acid (60.7% vs. 43.48% in the present work) and gentamicin (19% vs. 0% in the present work), while lower rates of resistance were found for doxycycline (29.8% vs. 65.22% in the present work) [25].
We found high resistance rates to aztreonam (91.3%) in E. coli isolates recovered from chromID ESBL. This finding is relevant, as aztreonam is categorized as "Category A: antimicrobial to avoid" in animals [64]. In addition, we observed high resistance rates to fluoroquinolones and cephalosporins of the third or fourth generation. Further, resistance to colistin was observed (8.69%). It should be pointed out that fluoroquinolones, cephalosporins of third or fourth generation, and colistin have been categorized as "Category B: antimicrobials to restrict" in animals [64].
We observed high resistance rates to aztreonam (87.3%) in K. pneumoniae isolates. In addition, resistance to tigecycline was observed (25%). Both aztreonam and tigecycline are categorized as "Category A: antimicrobial to avoid" in animals [64]. Moreover, we found high resistance rates to fluoroquinolones and cephalosporins of the third or fourth generation. In addition, resistance to colistin was observed (50%). As mentioned above, fluoroquinolones, cephalosporins of the third or fourth generation and colistin have been categorized as "Category B: antimicrobials to restrict" in animals [64]. K. pneumoniae is an opportunistic pathogen that is capable of persisting in various reservoirs, including These findings are of concern, as coagulase-negative staphylococci could be a reservoir of clinically relevant resistant genes that could be transferred to S. aureus isolates [83].
Like Mezher et al. [84], we did not isolate any Campylobacter spp. in turkey meat. However, other studies have shown a high prevalence of Campylobacter spp. in turkey meat [85]. Narvaez et al. [86] found Campylobacter spp. in 14.2% of the turkey samples.
Clostridium perfringens was not detected in the present work; few works deal with the detection of this pathogen in poultry meat, indicating populations of 1.0-1.2 log CFU/g [87].
In total, 35 different genera were identified in the present work, a higher number than that identified by Kačániová et al. [40] in chicken meat (15 genera). In addition, we detected some species that are considered as opportunistic pathogens and others that are recognized foodborne pathogens.

Conclusions
This study emphasized that turkey meat microbiota can be a source of both recognized foodborne pathogens and opportunistic or emerging pathogens. Moreover, turkey meat can be a source of K. pneumoniae, E. coli, S. aureus, coagulase negative staphylococci, and M. caseolyticus resistance to critical antibiotics, according to European Medicine Agency (EMA) criteria.
The presence of multi-resistant bacteria in turkey meat is of particular concern, and special measures should be taken within the framework of the One Health approach. Funding: Sixty-five percent of this project was co-financed by the European Regional Development Fund (ERDF) through the Interreg V-A Spain-France-Andorra program (POCTEFA, Programa INTERREG V-A España-Francia-Andorra) (2014-2020) (EFA (España-Francia-Andorra)) (152/16). POCTEFA aims to reinforce the economic and social integration of the French-Spanish-Andorran border. Its support is focused on developing economic, social, and environmental cross-border activities through joint strategies favouring sustainable territorial development. A.M.L. acknowledges and extends her thanks to the University of La Rioja and Rioja Government for her predoctoral fellowship (UR-CAR-2019).