Phenotypic and genotypic characterization of antimicrobial resistances reveals the effect of the production chain in reducing resistant lactic acid bacteria in an artisanal raw ewe milk PDO cheese

Food Research International 187 (2024) 114308 Available online 18 April 2024 0963-9969/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Phenotypic and genotypic characterization of antimicrobial resistances reveals the effect of the production chain in reducing resistant lactic acid bacteria in an artisanal raw ewe milk PDO cheese Gorka Santamarina-García a , c , d , * , Gustavo Amores a , c , d , Diego Llamazares a , Igor Hern´ andez a , c , d , Luis Javier R. Barron b , d , Mailo Virto a , c , d a Lactiker Research Group, Department of Biochemistry and Molecular Biology, Faculty of Pharmacy, University of the Basque Country (UPV/EHU), Paseo de la Universidad 7, 01006 Vitoria-Gasteiz, Spain b Lactiker Research Group, Department of Pharmacy and Food Sciences, Faculty of Pharmacy, University of the Basque Country (UPV/EHU), Paseo de la Universidad 7, 01006 Vitoria-Gasteiz, Spain c Bioaraba Health Research Institute-Prevention, Promotion and Health Care, 01009 Vitoria-Gasteiz, Spain d Joint Research Laboratory on Environmental Antibiotic Resistance, Department of Biochemistry and Molecular Biology, Faculty of Pharmacy, University of the Basque Country (UPV/EHU), Paseo de la Universidad 7, 01006 Vitoria-Gasteiz, Spain ARTICLE INFO Keywords: Sheep Raw ewe milk cheese Antimicrobial resistance Antibiotic resistance Antimicrobial susceptibility testing Broth microdilution method High-throughput quantitative PCR Resistance genes Mobile genetic elements ABSTRACT Antimicrobial resistance (AMR) is a significant public health threat, with the food production chain, and, specifically, fermented products, as a potential vehicle for dissemination. However, information about dairy products, especially raw ewe milk cheeses, is limited. The present study analysed, for the first time, the occurrence of AMRs related to lactic acid bacteria (LAB) along a raw ewe milk cheese production chain for the most common antimicrobial agents used on farms (dihydrostreptomycin, benzylpenicillin, amoxicillin and polymyxin B). More than 200 LAB isolates were obtained and identified by Sanger sequencing (V1-V3 16S rRNA regions); these isolates included 8 LAB genera and 21 species. Significant differences in LAB composition were observed throughout the production chain (P ≤0.001), with Enterococcus (e.g., E. hirae and E. faecalis) and Bacillus (e.g., B. thuringiensis and B. cereus) predominating in ovine faeces and raw ewe milk, respectively, along with Lactococcus (L. lactis) in whey and fresh cheeses, while Lactobacillus and Lacticaseibacillus species (e.g., Lactobacillus sp. and L. paracasei) prevailed in ripened cheeses. Phenotypically, by broth microdilution, Lactococcus, Enterococcus and Bacillus species presented the greatest resistance rates (on average, 78.2 %, 56.8 % and 53.4 %, respectively), specifically against polymyxin B, and were more susceptible to dihydrostreptomycin. Conversely, Lacticaseibacillus and Lactobacillus were more susceptible to all antimicrobials tested (31.4 % and 39.1 %, respectively). Thus, resistance patterns and multidrug resistance were reduced along the production chain (P ≤0.05). Genotypically, through HT-qPCR, 31 antimicrobial resistance genes (ARGs) and 6 mobile genetic elements (MGEs) were detected, predominating Str, StrB and aadA-01, related to aminoglycoside resistance, and the transposons tnpA02 and tnpA-01. In general, a significant reduction in ARGs and MGEs abundances was also observed throughout the production chain (P ≤0.001). The current findings indicate that LAB dynamics throughout the raw ewe milk cheese production chain facilitated a reduction in AMRs, which has not been reported to date. 1. Introduction Antibiotics are chemical compounds that attack essential bacterial physiology and biochemistry to cause cell death or growth cessation (Lade & Kim, 2021). For decades, antibiotics have been overused, both in human medicine and in animal production (Sobierajski et al., 2022), including the currently forbidden use of subtherapeutic doses as growth promoters (Patel et al., 2020). As a result, bacterial communities have been exposed to antibiotics and have developed the ability to withstand or resist the action of one or more antimicrobial agents, which is called * Corresponding author at: Lactiker Research Group, Department of Biochemistry and Molecular Biology, Faculty of Pharmacy, University of the Basque Country (UPV/EHU), Paseo de la Universidad 7, 01006 Vitoria-Gasteiz, Spain. E-mail address: [email protected] (G. Santamarina-García). Contents lists available at ScienceDirect Food Research International journal homepage: www.elsevier.com/locate/foodres https://doi.org/10.1016/j.foodres.2024.114308 Received 17 January 2024; Received in revised form 27 March 2024; Accepted 16 April 2024 Food Research International 187 (2024) 114308 2 antimicrobial resistance (AMR) (Konopka et al., 2022; Virto et al., 2022). Bacteria can be intrinsically resistant to certain antimicrobial groups or agents, mediated by chromosomal genes and linked to physiological or anatomical characteristics. Nonetheless, acquired resistance also occurs due to horizontal transmission between bacteria by means of mobile genetic elements (MGEs), which can carry one or more resistance genes (Iskandar et al., 2022; Nunziata et al., 2022), or due to generational genetic transmission by point mutations in genes that give rise to resistance or increased expression of resistance mechanisms (vertical transmission) (Iskandar et al., 2022; Wall et al., 2016). The antibiotics utilized in human medicine belong to the same pharmacological classes as those used in veterinary medicine (Devirgiliis et al., 2011); consequently, acquired resistance to certain antimicrobial agents is widespread to the point that effective treatment of certain fatal infections is already compromised (Virto et al., 2022). In fact, the proliferation of antimicrobial-resistant (AR) microorganisms has become one of the most important threats to human health (Wang et al., 2022) and is classified as one of the top 10 threats to global public health (WHO, 2022). It causes approximately 700,000 deaths worldwide per year and is projected to increase to 10 million each year by 2050 (IACG, 2019). Thus, AMR is of utmost importance and is included within the sustainable development goals (SDGs) set by the United Nations. Specifically, AMR affects SDG 3 on good health and well-being since it hinders the ability to control infectious diseases, increasing morbidity and mortality and resulting health care costs (United Nations, 2015). The food and food production chain is classified as a possible vehicle for the dissemination of AR bacteria and genes (Caniça et al., 2019); and, specifically, fermented products are considered notable reservoirs (Wang et al., 2006; Yasir et al., 2022). In this regard, several studies have been developed recently (Zhao et al., 2022), for instance, on raw beef, sheep and lamb meat (S¸anlıbaba, 2022) and dry-fermented sausages (Fraqueza, 2015). However, information about dairy products, especially raw milk cheeses, is limited, with most studies focused on raw cow milk cheeses (Dos Santos et al., 2022; Rola et al., 2016) and scarce information on raw sheep milk cheeses (Gaglio et al., 2016; Slyvka et al., 2022). Milk is an ideal growth medium for microorganisms due to its high nutrient content (Fusco et al., 2020). Consequently, the microbiota of raw ewe milk is diverse and is primarily composed of lactic acid bacteria (LAB), psychotropic bacteria and pathogens (Biçer et al., 2021; Santamarina-García et al., 2022a). Nonetheless, the cheese-making and ripening processes have a clear impact on bacterial communities, with a general predominance of LAB (Cardinali et al., 2021; SantamarinaGarcía et al., 2022a). Several studies have highlighted the presence of resistant bacteria in raw ewe milk and cheese, including pathogenic Escherichia coli and Staphylococcus aureus (Imre et al., 2022; Karahutov´ a & Bujˇ n´ akov´ a, 2023; Výrostkov´ a et al., 2020, 2021). Nonetheless, despite the predominance of LAB (Quigley et al., 2013; Santamarina-García et al., 2022a), there has been limited research on AMRs in LAB from raw ewe milk and derivate cheeses (Výrostkov´ a et al., 2020, 2021). In particular, species of the genus Enterococcus, such as E. faecium and E. faecalis, known as important opportunistic pathogens in nosocomial infections (Conde-Est´ evez et al., 2011), have been described as the most remarkable AR LAB (Výrostkov´ a et al., 2021). Addressing AMRs in LAB is essential since they can serve as potential reservoirs for the transfer of resistance genes to other bacteria, including pathogenic bacteria (Caniça et al., 2019). Several studies have reported the preference of consumers for raw milk cheeses (Colonna et al., 2011; Meunier-Goddik & Waite-Cusic, 2019), based on their richer and more intense aromatic profiles than pasteurized milk cheeses (Barron et al., 2007; O’Sullivan & Cotter, 2017). Given the pressing need to minimize the development and dissemination of AR LAB to safeguard public health (Výrostkov´ a et al., 2021), the present study is focused on Idiazabal protected designation of origin (PDO) cheese. It is a semihard or hard cheese from the Basque Country (southwestern Europe) produced with raw milk from the Latxa and/or Carranzana autochthonous sheep breeds, and it has a minimum mandatory ripening period of 60 days (Official Journal of the European Communities, 1996). Thus, this study aimed to characterize the prevalence of AMRs in LAB from ovine faeces, raw ewe milk, whey, fresh cheeses and 2-month-old ripened cheeses by means of phenotypic and genotypic approaches. Moreover, the potential differences among producers producing the same kind of raw ewe milk cheese were also analysed. To our knowledge, no study has comprehensively analysed the prevalence of AMRs along the production chain of a raw ewe milk cheese. 2. Methods 2.1. Area of study To evaluate the prevalence of AMRs in LAB along the production chain of artisanal raw ewe milk cheeses, this study was carried out within the European PDO Idiazabal cheese. This particular cheese was selected as a case study because its production is primarily carried out by small-scale artisanal dairies that oversee the entire process, from herd management to cheese-making. Idiazabal cheese is a semihard or hard cheese made from the raw milk of the autochthonous Latxa and/or Carranzana sheep breeds and has a mandatory minimum ripening time of 2 months. Herd management and milk production for cheese-making occur in the Basque Country, covering an area of 17,213.06 km 2 in southwestern Europe (43◦27 ′ −41◦54 ′ N and 1◦5 ′ −3◦37 ′ W). This region corresponds to the natural habitat of the sheep breeds (Official Journal of the European Communities, 1996). Herd management involves the use of indoor forage from October to March and semiextensive or extensive grazing from March to October (Aldalur et al., 2019). Milk collection and cheese production mainly occur between January and June, following the traditional seasonal approach dictated by the biological rhythms of the sheep (Boletín Oficial del Estado, 1993). 2.2. Sampling For sampling, four producers attached to the PDO Idiazabal cheese were chosen and identified as A, B, C, and D. Each producer came from one of the distinct geographical production areas (Alava, Biscay, Gipuzkoa, and Navarre). All the producers adhered to similar flock management and cheese-making practices in accordance with the specifications outlined by the Idiazabal PDO regulatory board (Boletín Oficial del Estado, 1993). The flocks consisted of approximately 350–400 Latxa breed sheep, following the management practices mentioned earlier. Milking was conducted automatically, and the milk was promptly refrigerated (3–4 ◦C) until cheese-making. For the cheesemaking process, the milk was warmed to 25 ◦C, and the commercial mesophilic lyophilized starter culture Choozit MM 100 LYO 50 DCU (a mixture of Lactococcus lactis subsp. lactis, Lactococcus lactis subsp. cremoris, and Lactococcus lactis subsp. lactis biovar. diacetylactis, DuPont NHIB Ib´ erica S.L., Barcelona, Spain) was added. Coagulation occurred at 28–32 ◦C for 20–45 min using artisanal rennet and/or the commercial NATUREN® 195 Premium (Chr. Hansen Holding A/S, Hørsholm, Denmark). The resulting curds were cut into 5–10 mm diameter grains and heated to 36–38 ◦C. Cheeses were then moulded, pressed and salted in saturated brine, and subsequently ripened in chambers maintained at 80–95 % relative humidity and 8–14 ◦C for 2 months. Thus, ovine faeces, raw ewe milk, whey, fresh cheeses (1-day-old), and 2-month-old ripened cheese samples were obtained from each producer. Samples were collected aseptically in quadruplicate, with each set of samples corresponding to the same batch. The sampling was conducted by the producers, eliminating the need for approval from the Ethics Committee for Animal Experimentation. Verbal consent was obtained from dairies during samples collection. Samples were collected from healthy flocks, excluding animals that underwent antibiotic treatment. Samples were transported under refrigerated conditions (3 ±1 ◦C) for analysis. G. Santamarina-García et al. Food Research International 187 (2024) 114308 3 2.3. Reagents and materials The peptone water was supplied by Panreac Química (Barcelona, Spain). De Man, Rogosa and Sharpe (MRS) agar, MRS broth medium, sodium citrate and sodium chloride were purchased from Scharlab (Barcelona, Spain). Tryptic soy broth (TSB) was obtained from Condalab (Madrid, Spain). Glycerol was obtained from Honeywell Fluka (Madrid, Spain). Amoxicillin was supplied by Sigma-Aldrich (Madrid, Spain). Dihydrostreptomycin and polymyxin B were purchased from Glentham Life Sciences (Corsham, United Kingdom). Benzylpenicillin was supplied by Tokyo Chemical Industry Co. (Tokyo, Japan). Mag-Bind Bacterial DNA 96 Kit was purchased from Omega Bio-Tek, Inc. (Norcross, United States). KAPA HiFi HotStart ReadyMix Kit was obtained from Roche Molecular Systems, Inc. (Branchburg, United States). CleanNGS and CleanDTR kits were obtained from CleanNA (Waddinxveen, The Netherlands). DNA 5 K Reagent Kit was obtained from PerkinElmer, Inc. (Waltham, United States). BigDye Terminator v3.1 Cycle Sequencing Kit and exonuclease I were purchased from Thermo Scientific (Waltham, United States). QIAamp® PowerFecal® Pro DNA Kit and QIAGEN® Multiplex PCR Kit were purchased from Qiagen (Valencia, United States). Petri dishes and 96-well plates were obtained from Deltalab (Barcelona, Spain). The Master Mix SsoFastTM EvaGreen® Supermix Kit with Low ROX was purchased from Bio-Rad Laboratories (Hercules, United States). 2.4. Phenotypic characterization of AMRs 2.4.1. LAB isolation and enumeration For the faeces and cheese samples, 10 g was diluted in duplicate 1:10 in peptone water and homogenized for 30 s three times in a stomacher (Masticator Basic 400, IUL Instruments, K¨ onigswinter, Germany). Serial dilutions were made in peptone water and plated on MRS agar media. For the raw ewe milk and whey samples, 100 µL was taken directly and serially diluted. The plates were incubated at 37 ±1 ◦C for 48 h. Three presumptive LAB isolates were randomly selected per sample based on colony morphology diversity. Then, the isolates were preenriched in TSB and subcultured onto MRS agar to ensure purity. All the cultures were stored at −80 ◦C in 20 % (v/v) glycerol. 2.4.2. Identification of LAB isolates by Sanger sequencing 2.4.2.1. DNA extraction. Isolates were preenriched in TSB and subcultured onto MRS agar prior to DNA extraction to ensure purity and viability. Bacterial DNA was extracted using the Mag-Bind Bacterial DNA 96 Kit following the manufacturer’s instructions for agar cultures, but the elution volume was reduced to 60 µL to improve the DNA yield. The quantity and quality of the DNA obtained were verified with a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Massachusetts, USA), in which the absorbance was measured at a wavelength of 260 nm and the 260/280 and 260/230 ratios were analysed. DNA extraction and subsequent Sanger sequencing were conducted in the Sequencing and Genotyping Unit of the Genomic Facility/SGIker (supported by UPV/EHU, MICINN, GV/EJ, FSE) of the University of the Basque Country. 2.4.2.2. Sanger sequencing. The V1–V3 regions of the 16S rRNA gene were amplified via PCR with the KAPA HiFi HotStart ReadyMix Kit using the forward primer 16S–V1–8F: 5 ′ - AGAGTTTGSTCCTGGCTCAG-3 ′ and the reverse primer 16S–V3–534R: 5 ′ - ATTACCGCGGCTGCTGG −3 ′ . The PCR products were purified by means of the CleanNGS Kit following the manufacturer’s instructions. Amplicon quantification was performed using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific) and a LabChip GX Touch Nucleic Acid Analyser (PerkinElmer) with a DNA 5 K Reagent Kit. Sequencing of the purified product was performed using the BigDye Terminator v3.1 Cycle Sequencing Kit following the manufacturer’s protocol. The sequencing product was purified using a magnetic bead-based CleanDTR kit, and Sanger sequencing was performed on the SeqStudio platform (Thermo Fisher). 2.4.2.3. Bioinformatic analysis. Quality filtering and trimming of the raw reads were performed using SeqStudio Reporter software (Thermo Scientific). The sequences were visualized and edited by means of the BioEdit Sequence Alignment Editor software 7.2.5 (Hall et al., 2011). The resulting sequences were approximately 500 bp in length. Taxonomic classification was performed against the Nucleotide Basic Local Alignment Search Tool (NBLAST) 2.14.0+(Zhang et al., 2000), with default parameters and taking into account the e-value, score, query cover and percentage of identification as quality indicators. 2.4.3. Antimicrobial susceptibility testing (AST) via the broth microdilution method The minimum inhibitory concentration (MIC) of the LAB isolates was evaluated by the broth microdilution method for the most widely used antimicrobial agents on farms (namely, amoxicillin, dihydrostreptomycin, benzylpenicillin and polymyxin B) according to the updated International Organization for Standardization and International Dairy Federation Standards (ISO/IDF, 2010) and European Food Safety Authority guidance (Rychen et al., 2018), with minor modifications. Briefly, a 96-well plate was inoculated with MRS broth medium supplemented with serial (1:2) concentrations of antibiotics (amoxicillin: 0.0313–16 µg/mL; dihydrostreptomycin: 1–512 µg/mL; benzylpenicillin: 0.0313–32 µg/mL; and polymyxin B: 2–1024 µg/mL). The inocula of each isolate were prepared in saline solution (0.85 %, m/v) by picking up single colonies from previously subcultured isolates on MRS agar to obtain an optical density equivalent to 0.5 on the MacFarland scale. The inoculum was subsequently diluted 1:10 in antibiotic-free MRS broth, and 50 µL of the diluted suspension was added to each well and incubated at 37 ±1 ◦C for 48 h. The inoculum in a well with MRS broth without antibiotics was used as a positive control, and an inoculum-free well was used as a negative control. The antimicrobial susceptibility or resistance was interpreted using the available microbiological cut-off values defined by the European Food Safety Authority (EFSA) Panel on Additives and Products or Substances used in Animal Feed (FEEDAP) (Rychen et al., 2018) and employing the epidemiological cut-off values (ECOFFs) proposed by the European Committee for Antimicrobial Susceptibility Testing (EUCAST; https://www.eucast. org). 2.5. Genotypic characterization of AMRs 2.5.1. DNA extraction To analyse the presence of antimicrobial resistance genes (ARGs), DNA was extracted as previously described (Santamarina-García et al., 2022a), with some modifications. Briefly, for the faecal and cheese samples, 10 g was suspended in 90 mL of 2 % (w/v) sterile sodium citrate (pH 8.0) and homogenized six times (20 s ON and 10 s OFF) in a stomacher (Masticator Basic 400; IUL Instruments, K¨ onigswinter, Germany). The resulting suspension was centrifuged at 6500 ×g for 8 min at 4 ◦C, after which the fat-containing supernatant was discarded. The obtained pellet was washed with 50 mL of sodium citrate and centrifuged at 6500 ×g for 8 min at 4 ◦C. The pellet was resuspended in 800 µL of sodium citrate and centrifuged three times at 6500 ×g for 8 min at 4 ◦C. The DNA was extracted with a QIAamp® PowerFecal® Pro DNA Kit according to the manufacturer’s protocol, but a double DNA elution step was carried out with 25 μ L of C6 solution to improve DNA yields. To extract DNA from the milk and whey samples, 10 mL was processed as described above, but without the need for homogenization in the stomacher. The DNA was stored at −80 ◦C until analysis. G. Santamarina-García et al. Food Research International 187 (2024) 114308 4 2.5.2. High-throughput quantitative PCR (HT-qPCR) The detection of ARGs was performed by means of HT-qPCR in a nanofluidic qPCR BioMarkTM HD system using 96.96 Dynamic Array Integrated Fluidic Circuits (IFCs) (Fluidigm Corporation), as previously described (Jauregi et al., 2021). A total of 48 primer sets were used (Supplementary Table 1) to target the ARGs conferring resistance against the most commonly used antimicrobial agents on farms (12 ARGs encoding resistance to dihydrostreptomycin, 24 ARGs for benzylpenicillin and amoxicillin, 2 ARGs for polymyxin B and 2 multidrug ARGs conferring resistance to more than one of the aforementioned antimicrobial agents), MGE genes (5 genes encoding transposases and 2 genes encoding integrases) and the 16S rRNA gene as a reference gene. These genes were selected considering the CARD database for LAB (Alcock et al., 2023). The primers used for qPCR were previously validated (Gorecki et al., 2022; Hu et al., 2016). DNA samples were preamplified using the QIAGEN® Multiplex PCR Kit and a primer pool (final concentration for each primer pair =50 nM), following the amplification program (at 95 ◦C for 15 min and 14 PCR cycles at 95 ◦C for 15 s, 60 ◦C for 4 min and a final extension step at 4 ◦C). Then, the samples were treated with exonuclease I (at 37 ◦C for 30 min for digestion, 80 ◦C for 15 min for inactivation of exonuclease I and kept at 4 ◦C). Subsequently, 1:10 dilutions of specific target amplification reactions were loaded onto the Dynamic Array IFCs following the Fluidigm’s Fast Gene Expression Analysis—EvaGreen® Protocol (Fluidigm Corporation). For amplification, the Master Mix SsoFastTM EvaGreen® Supermix Kit with Low ROX was used, with a final concentration of primers of 500 nM, both forward and reverse. The program consisted of 1 min of denaturation at 95 ◦C, followed by 30 cycles of 95 ◦C for 5 s and 60 ◦C for 20 s, a melting curve at 60 ◦C for 3 s and a ramp rate of 1 ◦C/3 s up to 95 ◦C. Four replicates were included for each sample. Analyses were conducted at the Gene Expression Unit of The Genomics Facility/ SGIker (supported by UPV/EHU, MICINN, GV/EJ, FSE) of the University of the Basque Country. 2.5.3. Bioinformatic analysis Raw data were processed with Fluidigm Real-Time PCR Analysis Software (v.3.1.3, Fluidigm Corporation), with linear baseline correction and manual threshold settings. A cycle threshold (CT) value of 30 was chosen because the highest CT value obtained in this study was 29.0. The detection of an ARG or MGE gene was considered positive when 3 out of the 4 technical replicates for each sample were above the detection limit. The relative abundances of the ARGs were calculated on the basis of the comparative CT method (Jauregi et al., 2021), normalized to the abundance of the 16S rRNA control gene and expressed as the fold change (FC). ΔCT(per replicate) = CT(target gene) − CT(16S rRNA gene) ΔΔCT(per sample) = ΔCT FC =2−ΔΔCT 2.6. Statistical analysis IBM SPSS statistical package version 26.0 (IBM SPSS, Inc., Chicago, IL, USA, 2019) was used for data preparation and analysis. Plot generation was performed in RStudio version 2023.03.1 and R version 4.3.0 (R Core Team, Vienna, Austria, 2023) with the “ggplot2” package (https ://github.com/tidyverse/ggplot2) and in Microsoft Office Professional Plus 2016 Excel® version 16.0.5413 (Microsoft, Albuquerque, United States). Kruskal–Wallis one-way analysis of variance (ANOVA) with Bonferroni correction was performed with the SPSS package to determine the significance (P ≤0.05) of the effects of producer and production chain (sample type) factors on the bacterial counts and abundances and phenotypic and genotypic results. Permutational multivariate analysis of variance (PERMANOVA) was carried out in R with the “vegan” package (https://github.com/vegandevs/vegan) to analyse the overall effect of producer and production chain factors. The data were log transformed when necessary and subjected to unit variance (UV) scaling, and a heatmap with hierarchical clustering analysis (HCA) was generated with the “pheatmap” package (https://github. com/raivokolde/pheatmap) to analyse the clustering of the phenotypic and genotypic results. Clustering of samples according to phenotypic and genotypic results was also performed by means of a dendrogram in R with the “factoextra” package (https://github.com/ka ssambara/factoextra). Trends in the bacterial counts and abundances and phenotypic and genotypic results according to producer and production chain factors were explored by means of principal component analysis (PCA), applied to log-transformed, when necessary, and UVscaled data and performed in SIMCA software version 17.0.2.34594 (Umetrics AB, Umeå, Sweden). The number of principal components (PCs) was determined by the eigenvalues (greater than 1.0) and crossvalidation. Similarly, orthogonal partial least squares-discriminant analysis (OPLS-DA) was performed with SIMCA software to analyse whether the samples differed according to the producer and production chain factors. Variable influence on projection (VIP) values and loading weights were used to analyse the importance of each parameter in the model. 3. Results 3.1. LAB prevalence and distribution throughout the cheese production chain Fig. 1A shows the prevalence of LAB throughout the Idiazabal cheese production chain. Overall, large differences were found according to the sample type (P ≤0.001). Specifically, a mean LAB prevalence of 6.45 ± 0.451 log CFU/g was observed in the faeces. In the raw ewe milk, the LAB count was 3.73 ±0.0659 log CFU/mL, which subsequently increased to 5.65 ±0.0623 log CFU/mL in the whey and to 7.99 ±0.172 log CFU/g in the fresh cheeses. However, during ripening, the LAB count slightly decreased to 7.75 ±0.202 log CFU/g, although the difference was not significant (Fig. 1A). Moreover, among producers, significant differences were also observed for the whey (P ≤0.01) and fresh cheese samples (P ≤0.05), with producer A clearly differentiated from the rest due to the lower values. Using multivariate analysis, PERMANOVA confirmed the differences among sample types (P ≤0.001) and, to a lesser extent, among producers (P ≤0.05). To identify the LAB communities, 203 isolates were obtained from the raw ewe milk Idiazabal cheese production chain. As expected, all the isolates belonged to the phylum Firmicutes and class Bacilli (Fig. 1B). Two orders were identified, predominantly Lactobacillales (69.0 %) and, to a lesser extent, Bacillales (31.0 %). All the isolates of the Bacillales order belonged to the Bacillaceae family and the Bacillus genus, identifying 4 different species, B. cereus (8.37 %), B. thuringiensis (6.40 %), B. paramycoides (2.96 %), and B. anthracis (0.99 %), in addition to other unidentified species (Bacillus sp., 12.3 %). The isolates of the order Lactobacillales belonged mainly to the Enterococcaceae (37.4 %) and Lactobacillaceae (22.2 %) families and, to a lesser extent, to the Streptococcaceae (9.36 %). All the Enterococcaceae isolates corresponded to the genus Enterococcus, identifying different species, such as E. hirae (19.2 %) and E. faecalis (13.3 %), and to a lesser extent, E. faecium (2.46 %), E. mundtii (1.48 %), E. avium (0.49 %) and E. durans (0.49 %). The Lactobacillaceae isolates belonged to the genus Lactobacillus, without being able to identify species (Lactobacillus sp., 7.88 %); Lacticaseibacillus (7.88 %), with the species L. paracasei (6.40 %) and L. casei (1.48 %); Levilactobacillus (3.94 %), namely, L. brevis; and Lactiplantibacillus (2.46 %), identified as L. plantarum (0.49 %) and L. plantarum subsp. plantarum (1.97 %). The Streptococcaceae isolates belonged to the genera Lactococcus (8.87 %), identifying L. lactis (4.43 %) and L. lactis subsp. lactis (3.45 %), in addition to unidentified strains (0.99 %), and Streptococcus, for which species could not be identified (Streptococcus sp., G. Santamarina-García et al. Food Research International 187 (2024) 114308 5 0.49 %). Within all the samples, E. hirae, E. faecalis, Bacillus sp., B. cereus and Lactobacillus sp. were some of the most important species throughout the production chain of the Idiazabal cheese. PERMANOVA confirmed the difference in LAB composition among the collected samples throughout the production chain (P ≤0.001) (Fig. 1B). E. hirae clearly predominated in the faeces (62.5 %), followed by B. thuringiensis (12.5 %) and other unidentified species (Bacillus sp.) (6.25 %). In raw ewe milk, instead, E. faecalis (36.4 %) predominated, followed by B. thuringiensis (13.6 %), Bacillus sp. (9.09 %) and E. hirae (9.09 %). During cheese-making, Bacillus species, such as B. cereus (15.8 %) or Bacillus sp. (13.2 %), dominated the whey; together with Enterococcus, such as E. hirae (10.5 %) or E. faecalis (7.89 %); and Lactococcus, L. lactis (7.89 %) and L. lactis subsp. lactis (7.89 %), or L. brevis (7.89 %). In fresh cheeses, a similar trend was maintained, with a predominance of unidentified Bacillus species (22.0 %), along with Enterococcus species, such as E. hirae (14.6 %) and E. faecalis (14.6 %), and also L. lactis (12.2 %). However, after ripening, Lactobacillus species predominated (20.8 %), followed by Lacticaseibacillus, specifically L. paracasei (16.7 %). In general, the abundance of other species, such as E. hirae (10.4 %) and Bacillus sp. (10.4 %), decreased during ripening. The greatest differences among sample types along the production chain were mainly observed for E. faecalis, L. paracasei, L. lactis and Lactobacillus sp. (P ≤0.05). PERMANOVA corroborated the lack of differentiation among producers (P >0.05). 3.2. Phenotypic profile of antimicrobial resistance Subsequently, antimicrobial susceptibility was tested by the broth microdilution method for more than 200 LAB isolates. The distributions of MICs are shown in Table 1. Clear differences were observed in the AMR phenotypes among the LAB communities (P ≤0.05) (Table 1 and Fig. 2A), which was confirmed by an OPLS-DA model (Supplementary Fig. 1). Overall, Lactococcus and Streptococcus species had the greatest resistance rates (on average, 78.2 % and 75.0 % of resistance of all isolates to all antibiotics, respectively), followed by Levilactobacillus, Enterococcus and Bacillus (65.6 %, 56.8 % and 53.4 %, respectively). Lactiplantibacillus, Lacticaseibacillus and Lactobacillus species, instead, were the most susceptible bacteria (31.3 %, 31.4 % and 39.1 %, respectively). In more detail, clear differences were observed in the AMR phenotypes among LAB species from the same genera and families (P ≤0.05). Within the Bacillaceae and Bacillus genus, B. anthracis species clearly differed from the other species because of their low resistance (12.5 %), with the remaining species, B. cereus, B. paramycoides, B. thuringiensis and Bacillus sp., exhibiting greater resistance (69.1 %, 58.3 %, 71.2 % and 56.0 %, respectively). For the Enterococcaceae and Enterococcus isolates, differences were also detected among the species, with E. durans and E. faecium being the most resistant (100 % and 80.0 %, respectively); E. hirae, E. mundtii and E. faecalis presenting greater susceptibility (57.1 %, 58.3 % and 45.4 %, respectively); and E. avium isolates being sensitive to all the antibiotics tested. On the other hand, most Lactobacillaceae genera and species showed similar low resistance rates, including unidentified Lactobacillus species (39.1 %), Lacticaseibacillus species, namely, L. paracasei and L. casei (21.2 % and 41.7 %, respectively), and Lactiplantibacillus species, specifically L. plantarum and L. plantarum subsp. plantarum (25.0 % and 37.5 %, respectively). The Levilactobacillus genus and the L. brevis species were unique exceptions for their higher levels of resistance. Finally, Streptococcaceae isolates also showed low differences among genera and species. All Lactococcus species, including L. lactis, L. lactis subsp. lactis and other unidentified species (72.2 %, 75.0 % and 88.0 % on average, respectively), exhibited high resistance, similar to unidentified Streptococcus species (75.0 % on average). Thus, the HCA and dendrogram divided the LAB communities into two clusters (Fig. 2A and B). E. avium, L. plantarum, L. plantarum subsp. plantarum and Fig. 1. Mean counts (log CFU/g or mL) and relative abundance (%) of lactic acid bacteria throughout the Idiazabal cheese production chain (faeces, raw milk, whey, fresh cheese and ripened cheese samples). The different lowercase letters for each type of sample indicate statistically significant differences. G. Santamarina-García et al. Food Research International 187 (2024) 114308 6 B. anthracis were the most differentiated LAB, as they were the most sensitive bacteria to all the antibiotics tested (on average, 18.8 % of all the isolates were resistant to all the antibiotics) (cluster 1). On the other hand, the remaining LAB species exhibited higher resistance rates (cluster 2). L. casei, L. paracasei, E. faecalis and Lactobacillus sp. were closely related, as they showed higher but still lower resistance rates (on average, 36.8 %) (cluster 2.3). E. durans, Lactococcus sp., E. mundtii and Streptococcus sp. stood out (clusters 2.1 and 2.2), showing the highest resistance rates to all the antimicrobial agents tested (93.8 %). The resistance rate against dihydrostreptomycin was the main reason for the Table 1 Distribution of minimum inhibitory concentration (MIC) values for the 202 isolates obtained throughout the production chain (ovine faeces, raw ewe milk, whey, fresh cheese and 60-day-old ripened cheese) of raw ewe milk Idiazabal cheese. 1 The range of dilutions tested for each antimicrobial agent is indicated in white. The vertical lines indicate the epidemiological cut-off (ECOFF) values according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) or European Food Safety Authority (EFSA). MICs lower than the lowest concentration tested are indicated in the closest concentration of the grey range. 2 n.d. =not detected. 3 MIC 50 (µg/kg) =MIC requeried for the inhibition of the growth of the 50% of the isolates. 4 MIC 90 (µg/kg) =MIC requeried for the inhibition of the growth of the 90% of the isolates. G. Santamarina-García et al. Food Research International 187 (2024) 114308 7 difference between these two clusters (2.1 and 2.2). The remaining bacterial species belonging to the Bacillus, Enterococcus, Lactococcus or Levilactobacillus genera exhibited similar high-intermediate resistance (66.7 % and 67.2 %, respectively) (clusters 2.4 and 2.5). The MIC 50 and MIC 90 , defined as the MIC required for the inhibition of the growth of 50 % and 90 %, respectively, of the isolates confirmed the high resistance of some of the bacterial species (Table 1). However, some trends were observed along the production chain (Supplementary Table 2). B. cereus, E. hirae, E. mundtti, L. lactis subsp. lactis and L. brevis maintained similar MIC 50 and MIC 90 values throughout the production chain, while for B. paramycoides, E. faecium, L. paracasei and Lactobacillus sp. increased, indicating a greater prevalence of resistant bacteria throughout the production chain. Finally, the MIC 50 and MIC 90 of B. thuringiensis, Bacillus sp., E. faecalis, L. casei, L. plantarum subsp. plantarum and L. lactis decreased, which indicated an increase in the abundance of sensible bacteria during the cheese production process. In general, 24.1 % (49/203) of the LAB isolates were susceptible to all the antimicrobial agents tested (Fig. 2C). Thus, 75.9 % (154/203) were resistant to at least one of the antibiotics tested, with resistance to 3 or 4 antimicrobial agents being the most common (46/203, 22.7 % and 50/203, 24.6 %, respectively) (Fig. 2C). However, differences were observed throughout the cheese production chain and were mainly related to differences in LAB composition (P ≤0.001) (Fig. 2C). In faeces, LAB isolates resistant to 3 antimicrobial agents predominated (25.0 %), followed by those resistant to 4 antimicrobial agents (21.9 %), since the predominant E. hirae was mainly resistant to 2 or 3 antimicrobial agents. Other minor species, such as L. brevis or Streptococcus sp., were principally resistant to 3 or 4 individual compounds. In raw ewe milk, resistance to 4 antimicrobial agents predominated (34.1 %), followed by resistance to 3 antimicrobial agents (20.5 %), since the predominant species, E. faecalis, B. thuringiensis, and E. hirae, and most minor species were mainly resistant to 4 and, to a lesser extent, to 3 antimicrobial agents. In whey, a similar trend compared to milk was observed, although resistance to 3 antimicrobial agents predominated (31.6 %), followed by resistance to 4 antimicrobial agents (28.9 %). In this case, the predominant B. cereus was equally resistant to 3 or 4 antimicrobial agents, while Bacillus sp. was mainly resistant to 4 antimicrobial agents, E. hirae was resistant to 3 antimicrobial agents, and most minor species were resistant to 3 or 4 antimicrobial agents. In fresh cheeses, resistance to 4 antimicrobial agents was dominant (26.8 %), followed by resistance to 1 antimicrobial agent (22.0 %). The predominant Bacillus sp. species were equally resistant to 3 or 4 antimicrobial agents, while other dominant species such as E. hirae being mainly resistant to 1 antimicrobial, E. faecalis to 2 antimicrobial agents and L. lactis to 4 antimicrobial agents. A large proportion of the minor species were also resistant to 1 or 4 antimicrobial agents. Finally, in the ripened cheeses, resistance to 2 or 3 antimicrobial agents predominated (25.0 % in both cases), since the predominant Lactobacillus sp. and L. paracasei were resistant to 2 antimicrobial agents, and, to a lesser extent, other minor important species (B. cereus or Bacillus sp.) were resistant to 3 antimicrobial agents. Overall, the proportion of LAB resistant to 1 or 2 antimicrobial agents increased throughout the production chain, while the proportion of strains resistant to 3 or 4 antimicrobial agents decreased. No differences in terms of abundance were Fig. 2. HCA heatmap (A), dendogram clustering (B), box plot representations (C, D and E) based on antimicrobial susceptibility testing (AST) results according to the bacterial species (A and B), number of antimicrobials (C) and sample type along production chain (D and E). Abbreviations: DHS: dihydrostreptomycin; PB: polymyxin B; PG: benzylpenicillin; AMX: amoxicillin. G. Santamarina-García et al. Food Research International 187 (2024) 114308 8 observed for the susceptible bacteria throughout the production chain (P >0.05), although the bacteria differed taxonomically. In faeces, raw ewe milk and whey, susceptible LAB belonged, mainly, to E. hirae and/or E. faecalis species, while in fresh cheese, they were, mainly, unidentified Bacillus species; and in ripened cheeses, they corresponded to L. paracasei; and, to a lesser extent, to L. plantarum subsp. plantarum and Lactobacillus sp. Regarding individual compounds, the resistance against polymyxin B was the most common (67.0 %), followed by benzylpenicillin (54.7 %) and amoxicillin (51.2 %), and being clearly more sensitive to dihydrostreptomycin (37.4 %) (Table 1, Fig. 2D-E). The predominance of resistance to polymyxin B was observed in all the samples except for ripened cheeses, where a higher prevalence of LAB isolates resistant to benzylpenicillin was observed. Resistance to benzylpenicillin was primarily observed in ripened cheeses, while amoxicillin resistance was mainly detected in faeces and whey, and dihydrostreptomycin resistance was mainly detected in raw ewe milk and fresh cheeses (Fig. 2D-E). This differentiation along the production chain was related to the LAB composition of each sample type (P ≤0.05) (Table 1). E. durans or Lactococcus sp., followed by E. faecium, were the most resistant to dihydrostreptomycin (Fig. 2A), and the most sensitive species were B. anthracis, E. avium, E. mundtii, L. plantarum or Streptococcus sp. In the case of benzylpenicillin, L. plantarum subsp. plantarum, E. durans, Lactococcus sp. and Streptococcus sp. were the most resistant, while B. anthracis or E. avium were the most sensitive. E. durans, Lactococcus sp., L. lactis subsp. lactis and Streptococcus sp. were the most resistant to polymyxin B, while E. avium and L. plantarum subsp. plantarum were the most susceptible. Finally, E. durans, E. mundtii or Streptococcus sp. were the most resistant to amoxicillin, and B. anthracis, E. avium or L. plantarum were the most sensitive. Regarding the AMR profiles of LAB (Fig. 3A-B), which results from all possible combinations of all antibiotics tested, resistance to all antibiotics predominated (24.6 %), followed by polymyxin B-benzylpenicillinamoxicillin (16.7 %), polymyxin B (6.90 %) and dihydrostreptomycinpolymyxin B (5.42 %). These patterns were mainly observed for E. hirae and, to a lesser extent, for Bacillus sp., Bacillus cereus and Fig. 3. Box plot representations based on the resistance patterns according to the sample type along production chain and bacterial species (A and B, respectively), and box plot representation (C), HCA heatmap (D) and dendogram clustering (E) based on the resistance against the antimicrobial classes tested according to the sample type along production chain (C) and bacterial species (D and E). Abbreviations: DHS: dihydrostreptomycin; PB: polymyxin B; PG: benzylpenicillin; AMX: amoxicillin. G. Santamarina-García et al. Food Research International 187 (2024) 114308 9 E. faecalis. Resistance patterns were significantly related to the LAB communities (P ≤0.01), consequently leading to differentiation in the cheese production chain. Resistance to all antibiotics was mainly observed for E. hirae and Bacillus sp., followed by L. lactis and E. faecalis, while the pattern of resistance to polymyxin B-benzylpenicillin-amoxicillin was most common for E. hirae, B. cereus and Bacillus sp. (Table 1, Supplementary Fig. 2). Resistance to polymyxin B was related to E. hirae, E. faecalis and Lactobacillus sp., and dihydrostreptomycin-polymyxin B combination was observed in E. hirae, E. faecalis and L. paracasei (Table 1, Supplementary Fig. 2). Thus, considering LAB communities throughout the production chain, resistance to polymyxin B-benzylpenicillin-amoxicillin (25.0 %) and to all antibiotics predominated (21.9 %) in faeces. In raw ewe milk, a similar trend was observed, although resistance to all antimicrobials was notably greater (34.1 %) than that to polymyxin B-benzylpenicillin-amoxicillin (9.09 %), which was maintained in whey (28.9 % and 23.7 %, respectively) and fresh cheeses (26.8 % and 7.32 %, respectively), albeit at different proportions. In the ripened cheeses, the resistance to all the antimicrobial agents was clearly lower (12.5 %), and the resistance to polymyxin B-benzylpenicillin and benzylpenicillin was also notable (12.5 % and 10.4 %, respectively). Notably, no isolate resistant to the combination of dihydrostreptomycin-amoxicillin or dihydrostreptomycin-benzylpeni cillin-amoxicillin was found in any type of sample. In relation to multidrug resistance (MDR) (Fig. 3C-E), which is defined as resistance to 3 or more classes of antibiotics, it was observed in 30.5 % of the LAB isolates. The MDR differed throughout the production chain (P ≤0.05). Specifically, the proportion of bacteria resistant to 1 class of antimicrobial did not differ throughout the production chain (12.9–29.0 %) (P >0.05). However, the number of isolates resistant to 2 classes differed significantly according to the sample type (P ≤0.05), with the highest prevalence found in ripened cheeses (32.8 %) and faeces (21.3 %) and the lowest in raw ewe milk and fresh cheeses (13.1 % in both cases) (Fig. 3C). The prevalence of multirresistant bacteria also differed throughout the production chain (P ≤0.01), with the highest rate observed in raw ewe milk (32.3 %), which decreased throughout cheese-making and ripening processes (12.9 %) (Fig. 3C). These dynamics were related to the LAB communities (P ≤0.05) (Supplementary Fig. 3), since MDR was mainly observed in E. hirae (17.2 %), E. faecalis (15.6 %), B. thuringiensis (15.6 %) and Bacillus sp. (15.6 %). Resistance to 2 classes was detected mainly in E. hirae (23.7 %), B. cereus (15.3 %) and Bacillus sp. isolates (13.6 %), and resistance to 1 class was more common in Lactobacillus sp. (19.4 %) and E. hirae (16.1 %). Moreover, the HCA and dendrogram divided LAB species into 4 clusters according to the predominant phenotype within each species. Thus, the MDR phenotype was predominant within E. durans, Lactococcus sp., E. faecium, L. lactis, B. thuringiensis and B. paramycoides. 3.3. Genotypic profile of antimicrobial resistance Regarding the ARGs and MGEs, 37 out of the 47 genes studied were detected (Fig. 4A, Supplementary Table 3). Among all the samples, the predominant ARGs were Str (average relative abundance of 387), followed by StrB (39.3) and aadA-01 (19.3), while among the MGEs, tnpA02 and tnpA-01 predominated (71.3 and 26.5, respectively). The ARGs aph, aph6ia, blaZ, blaTEM, blaGES, blaCTX-M−03, blaOKP and pbp2x and the MGEs intI1 and tnpA-03 were not detected. In general, aminoglycoside ARGs presented greater abundances than β-lactam, polymyxin and multidrug ARGs. Among the antimicrobial agents, the Str gene exhibited the highest relative abundance within aminoglycosides, followed by StrB and aadA-01, while aacA/aphD and aadA5-01 presented the lowest abundances. For β-lactams, bla-ACC-1, pbp and blaCMY2-01 dominated, while ampC, blaIMP-01 and pbp5 were minor ARGs. For polymyxins, mcr-2 presented the greatest abundance compared to mcr-1. Among the multidrug ARGs, two genes were also detected, namely, tolC01 and mexD, the first presenting the greatest abundances. Among the MGEs, transposons presented the greatest abundance, especially tnpA-02 and tnpA-01, while tnpA-07 was the least abundant. A similar abundance was observed for integrons, namely, intI and IS613. Overall, predominant transposons presented higher abundances than predominant aminoglycoside ARGs, with the exception of Str, while lower abundances were observed for integrons, similar to β-lactams, polymyxins and multidrug ARGs. PERMANOVA indicated that there were no differences among producers in terms of the detected ARGs or MGEs (P >0.05). Nonetheless, Fig. 4. HCA heatmap (A) and Venn diagram (B) showing the distribution of ARGs and MGEs along cheese production chain (faeces, raw milk, whey, fresh cheeses and ripened cheeses). G. Santamarina-García et al. Food Research International 187 (2024) 114308 16 P´ erez-Elortondo, F. J., Ald´ amiz Echobarria, P., Albisu, M., & Barcina, Y. (1998). Indigenous lactic acid bacteria in Idiazabal ewes’ milk cheese. International Dairy Journal, 8, 725–732. https://doi.org/10.1016/S0958-6946(98)00113-7 P´ erez-Elortondo, F. J., Albisu, M., & Barcina, Y. (1999). Physicochemical properties and secondary microflora variability in the manufacture and ripening of Idiazabal cheese. Lait, 79, 281–290. P´ erez Elortondo, F. J., Albisu, M., & Barcina, Y. (1999). Brining time effect on physicochemical and microbiological parameters in Idiazabal cheese. International Journal of Food Microbiology, 49, 139–149. https://doi.org/10.1016/S0168-1605 (99)00068-9 Possas, A., Bonilla-Luque, O. M., & Valero, A. (2021). From cheese-making to consumption: Exploring the microbial safety of cheeses through predictive microbiology models. Foods, 10, Article 355. https://doi.org/10.3390/ foods10020355 Quigley, L., O’Sullivan, O., Stanton, C., Beresford, T. P., Ross, R. P., Fitzgerald, G. F., & Cotter, P. D. (2013). The complex microbiota of raw milk. FEMS Microbiology Reviews, 37, 664–698. https://doi.org/10.1111/1574-6976.12030 Rajput, K., Dubey, R. C., & Kumar, A. (2022). Probiotic potential and immunomodulatory properties in Enterococcus faecium GMB24 and Enterococcus hirae SMB16 isolated from goat and sheep milk. Archives of Microbiology, 204, 619. https://doi.org/ 10.1007/s00203-022-03217-w Rocha, R., Couto, N., Pinto, R. P., Vaz-Velho, M., Fernandes, P., & Santos, J. (2023). Microbiological characterization of protected designation of origin serra da estrela cheese. Foods, 12, Article 2008. doi: 10.3390/FOODS12102008. Rola, J. G., Czubkowska, A., Korpysa-Dzirba, W., & Osek, J. (2016). Occurrence of Staphylococcus aureus on farms with small scale production of raw milk cheeses in poland. Toxins, 8, Article 62. https://doi.org/10.3390/TOXINS8030062 Rosander, A., Connolly, E., & Roos, S. (2008). Removal of antibiotic resistance genecarrying plasmids from Lactobacillus reuteri ATCC 55730 and characterization of the resulting daughter strain, L. reuteri DSM 17938. Applied and Environmental Microbiology, 74, 6032–6040. https://doi.org/10.1128/AEM.00991-08 Rüegsegger, F., Ruf, J., Tschuor, A., Sigrist, Y., Rosskopf, M., & H¨ assig, M. (2014). Antimicrobial susceptibility of mastitis pathogens of dairy cows in Switzerland. Schweizer Archiv Fur Tierheilkunde, 156, 483–488. https://doi.org/10.1024/00367281/a000635 Ruiz-Moyano, S., Gonçalves dos Santos, M. T. P., Galv´ an, A. I., Merch´ an, A. V., Gonz´ alez, E., de C´ ordoba, M. G., & Benito, M. J. (2019). Screening of autochthonous lactic acid bacteria strains from artisanal soft cheese: Probiotic characteristics and prebiotic metabolism. LWT, 114, Article 108388. https://doi.org/10.1016/j. lwt.2019.108388 Rychen, G., Aquilina, G., Azimonti, G., Bampidis, V., Bastos, M. L., Bories, G., Chesson, A., Cocconcelli, P. S., Flachowsky, G., Gropp, J., Kolar, B., Kouba, M., L´ opez-Alonso, M., Puente, S. L., Mantovani, A., Mayo, B., Ramos, F., Saarela, M., Villa, R. E., Wallace, R. J., Wester, P., Glandorf, B., Herman, L., K¨ arenlampi, S., Aguilera, J., Anguita, M., Brozzi, R., & Galobart, J. (2018). Guidance on the characterisation of microorganisms used as feed additives or as production organisms. EFSA Journal, 16, Article 5206. S¸anlıbaba, P. (2022). Prevalence, antibiotic resistance, and enterotoxin production of Staphylococcus aureus isolated from retail raw beef, sheep, and lamb meat in Turkey. International Journal of Food Microbiology, 361, Article 109461. https://doi.org/ 10.1016/J.IJFOODMICRO.2021.109461 Santamarina-García, G., Amores, G., Hern´ andez, I., Mor´ an, L., Barr´ on, L. J. R., & Virto, M. (2023). Relationship between the dynamics of volatile aroma compounds and microbial succession during the ripening of raw ewe milk-derived Idiazabal cheese. Current Research in Food Science, 6, Article 100425. https://doi.org/10.1016/ J.CRFS.2022.100425 Santamarina-García, G., Amores, G., L´ opez de Armentia, E., Hern´ andez, I., & Virto, M. (2022). Relationship between the dynamics of gross composition, free fatty acids and biogenic amines, and microbial shifts during the ripening of raw ewe milk-derived idiazabal cheese. Article 3224 Animals, 12. https://doi.org/10.3390/ani12223224. Santamarina-García, G., Fresno, J. M., Virto, M., Amores, G., & Aranceta, J. (2020). Cheese microbiota and its functional importance. Revista Espanola de Nutricion Comunitaria-Spanish Journal of Community Nutrition, 26, 248–256. Santamarina-García, G., Hern´ andez, I., Amores, G., & Virto, M. (2022). Characterization of microbial shifts during the production and ripening of raw ewe milk-derived idiazabal cheese by high-throughput sequencing. Biology, 11, 769. https://doi.org/ 10.3390/BIOLOGY11050769 Sharma, P., Tomar, S. K., Goswami, P., Sangwan, V., & Singh, R. (2014). Antibiotic resistance among commercially available probiotics. Food Research International, 57, 176–195. https://doi.org/10.1016/j.foodres.2014.01.025 Shi, H., Zhou, M., Zhang, Z., Hu, Y., Song, S., Hui, R., Wang, L., Li, G., & Yao, L. (2023). Molecular epidemiology, drug resistance, and virulence gene analysis of Streptococcus agalactiae isolates from dairy goats in backyard farms in China. Frontiers in Cellular and Infection Microbiology, 12, Article 1049167. https://doi.org/ 10.3389/fcimb.2022.1049167 Slyvka, I., Tsisaryk, O., Musii, L., Kushnir, I., Koziorowski, M., & Koziorowska, A. (2022). Identification and Investigation of properties of strains Enterococcus spp. Isolated from artisanal Carpathian cheese. Biocatalysis and Agricultural Biotechnology, 39, Article 102259. https://doi.org/10.1016/J.BCAB.2021.102259 Sobierajski, T., Mazi´ nska, B., Chajęcka-Wierzchowska, W., ´ Smiałek, M., & Hryniewicz, W. (2022). Antimicrobial and antibiotic resistance from the perspective of polish veterinary students: an inter-university study. Antibiotics, 11, Article 115. https://doi.org/10.3390/ANTIBIOTICS11010115 Sozzi, T., & Smiley, M. B. (1980). Antibiotic resistances of yogurt starter cultures. Streptococcus thermophilus and Lactobacillus bulgaricus. Applied and Environmental Microbiology, 40, 862–865. https://doi.org/10.1128/aem.40.5.862-865.1980 ˇ Stˇ sepetova, J., Taelma, H., Smidt, I., Hütt, P., Lapp, E., Aot¨ aht, E., & M¨ andar, R. (2017). Assessment of phenotypic and genotypic antibiotic susceptibility of vaginal Lactobacillus sp. Journal of Applied Microbiology, 123, 524–534. https://doi.org/ 10.1111/JAM.13497 Sun, G., Zhang, H., Wei, Q., Zhao, C., Yang, X., Wu, X., Xia, T., Liu, G., Zhang, L., Gao, Y., Sha, W., & Li, Y. (2019). Comparative analyses of fecal microbiota in European mouflon (Ovis orientalis musimon) and blue sheep (Pseudois nayaur) living at low or high altitudes. Frontiers in Microbiology, 10, Article 1735. doi: 10.3389/ fmicb.2019.01735. Tosi, L., Berruti, G., Danielsen, M., Wind, A., Huys, G., & Morelli, L. (2007). Susceptibility of Streptococcus thermophilus to antibiotics. Antonie van Leeuwenhoek, 92, 21–28. https://doi.org/10.1007/s10482-006-9130-6 Tsigkrimani, M., Panagiotarea, K., Paramithiotis, S., Bosnea, L., Pappa, E., Drosinos, E. H., Skandamis, P. N., & Mataragas, M. (2022). Microbial Ecology of Sheep Milk, Artisanal Feta, and Kefalograviera Cheeses. Part II: Technological, Safety, and Probiotic Attributes of Lactic Acid Bacteria Isolates. Foods, 11, Article 459. doi: 10.3390/foods11030459. United Nations (UN), 2015. The 2030 Agenda for Sustainable Development. Retrieved from https://sdgs.un.org/publications/transforming-our-world-2030-agenda-sustainabledevelopment-17981. Accesed June 25, 2023. Vidal, C. A., & Collins-Thompson, D. L. (1987). Resistance and sensitivity of meat lactic acid bacteria to antibiotics. Journal of Food Protection, 50, 737–740. https://doi.org/ 10.4315/0362-028X-50.9.737 Virto, M., Santamarina-García, G., Amores, G., & Hern´ andez, I. (2022). Antibiotics in dairy production: Where is the problem? Dairy, 3, 541–564. https://doi.org/ 10.3390/dairy3030039 Výrostkov´ a, J., Regecov´ a, I., Dudrikov´ a, E., Marcinˇ c´ ak, S., Vargov´ a, M., Kov´ aˇ cov´ a, M., & Mal’ov´ a, J. (2021). Antimicrobial resistance of Enterococcus sp. Isolated from sheep and goat cheeses. Foods, 10, Article 1844. doi: 10.3390/foods10081844. Výrostkov´ a, J., Regecov´ a, I., Kov´ aˇ cov´ a, M., Marcinˇ c´ ak, S., Mal’ov´ a, J., & Dudrikov´ a, E. (2020). Antimicrobial resistance of Lactobacillus johnsonii and Lactobacillus zeae in raw milk. Processes, 8, Article 1627. doi: 10.3390/PR8121627. Wall, B.A., Mateus, A., Marshall, L., Pfeiffer, D., Lubroth, J., Ormel, H.J., Otto, P., Patriarchi, A., 2016. Drivers, dynamics and epidemiology of antimicrobial resistance in animal production. Retrieved from https://www.fao.org/documents/card/en? details=d5f6d40d-ef08-4fcc-866b-5e5a92a12dbf/. Accessed June 8, 2023. Wang, H. H., Manuzon, M., Lehman, M., Wan, K., Luo, H., Wittum, T. E., Yousef, A., & Bakaletz, L. O. (2006). Food commensal microbes as a potentially important avenue in transmitting antibiotic resistance genes. FEMS Microbiology Letters, 254, 226–231. https://doi.org/10.1111/J.1574-6968.2005.00030.X Wang, J., Li, M., Wang, J., Liu, M., Yang, K. U. N., Zhang, J. I. E., Fan, M., & Wei, X. (2018). Antibiotic resistance of coagulase-negative staphylococci and lactic acid bacteria isolated from naturally fermented chinese cured beef. Journal of Food Protection, 81, 2054–2063. https://doi.org/10.4315/0362-028X.JFP-18-195 Wang, Y., Xu, Y., Guo, X., Wang, L., Zeng, J., Qiu, H., Tan, Y., Chen, D., Zhao, H., & Gu, Y. (2022). Enhanced antimicrobial activity through the combination of antimicrobial photodynamic therapy and low-frequency ultrasonic irradiation. Advanced Drug Delivery Reviews, 183, Article 114168. https://doi.org/10.1016/J. ADDR.2022.114168 Waseem, H., Jameel, S., Ali, J., Ur Rehman, H. S., Tauseef, I., Farooq, U., Jamal, A., & Ali, M. I. (2019). Contributions and challenges of high throughput qPCR for determining antimicrobial resistance in the environment: A critical review. Molecules, 24, Article 163. https://doi.org/10.3390/molecules24010163 Wei, N., Lu, J., Dong, Y., & Li, S. (2022). Profiles of microbial community and antibiotic resistome in wild tick species. MSystems, 7, Article e00037–22. https://doi.org/ 10.1128/msystems.00037-22 Winand, R., Bogaerts, B., Hoffman, S., Lefevre, L., Delvoye, M., Van Braekel, J., Fu, Q., Roosens, N. H. C., De Keersmaecker, S. C. J., & Vanneste, K. (2019). Targeting the 16S rRNA gene for bacterial identification in complex mixed samples: comparative evaluation of second (illumina) and third (oxford nanopore technologies) generation sequencing technologies. International Journal of Molecular Sciences, 21, Article 298. https://doi.org/10.3390/IJMS21010298 Wolfe, B. E. (2023). Are fermented foods an overlooked reservoir of antimicrobial resistance? Current Opinion in Food Science, 51, Article 101018. https://doi.org/ 10.1016/J.COFS.2023.101018 World Health Organization (WHO), 2022. Antimicrobial resistance. Retrieved from https ://www.who.int/news-room/fact-sheets/detail/antimicrobial-resistance. Accessed February 18, 2023. Yang, L. Y., Zhou, S. Y. D., Lin, C. S., Huang, X. R., Neilson, R., & Yang, X. R. (2022). Effects of biofertilizer on soil microbial diversity and antibiotic resistance genes. Science of the Total Environment, 820, Article 153170. https://doi.org/10.1016/j. scitotenv.2022.153170 Yap, M., Ercolini, D., ´ Alvarez-Ord´ o˜ nez, A., O’Toole, P. W., O’Sullivan, O., & Cotter, P. D. (2022). Next-generation food research: use of meta-omic approaches for characterizing microbial communities along the food chain. Annual Review of Food Science and Technology, 13, 361–384. https://doi.org/10.1146/ANNUREV-FOOD052720-010751 Yasir, M., Al-Zahrani, I. A., Bibi, F., Abd El Ghany, M., & Azhar, E. I. (2022). New insights of bacterial communities in fermented vegetables from shotgun metagenomics and identification of antibiotic resistance genes and probiotic bacteria. Food Research International, 157, Article 111190. https://doi.org/10.1016/J. FOODRES.2022.111190 Zhai, Z., Cui, C., Li, X., Yan, J., Sun, E., Wang, C., Guo, H., & Hao, Y. (2023). Prevalence, antimicrobial susceptibility, and antibiotic resistance gene transfer of Bacillus strains isolated from pasteurized milk. Journal of Dairy Science, 106, 75–83. https://doi.org/ 10.3168/jds.2022-22199 G. Santamarina-García et al. Food Research International 187 (2024) 114308 17 Zhang, Z., Schwartz, S., Wagner, L., & Miller, W. (2000). A greedy algorithm for aligning DNA sequences. Journal of Computational Biology, 7, 203–214. https://doi.org/ 10.1089/10665270050081478 Zhao, J., Liu, R., Sun, Y., Yang, X., & Yao, J. (2022). Tracing enterococci persistence along a pork production chain from feed to food in China. Animal Nutrition, 9, 223–232. https://doi.org/10.1016/J.ANINU.2022.01.005 Zheng, B., Feng, C., Xu, H., Yu, X., Guo, L., Jiang, X., & Song, X. (2019). Detection and characterization of ESBL-producing Escherichia coli expressing mcr-1 from dairy cows in China. Journal of Antimicrobial Chemotherapy, 74, 321–325. https://doi.org/ 10.1093/jac/dky446 Zheng, J., Wittouck, S., Salvetti, E., Franz, C. M. A. P., Harris, H. M. B., Mattarelli, P., O’toole, P. W., Pot, B., Vandamme, P., Walter, J., Watanabe, K., Wuyts, S., Felis, G. E., G¨ anzle, M. G., & Lebeer, S. (2020). A taxonomic note on the genus Lactobacillus: Description of 23 novel genera, emended description of the genus Lactobacillus beijerinck 1901, and union of Lactobacillaceae and Leuconostocaceae. International Journal of Systematic and Evolutionary Microbiology, 70, 2782–2858. https://doi.org/10.1099/ijsem.0.004107 Zhou, J. S., Pillidge, C. J., Gopal, P. K., & Gill, H. S. (2005). Antibiotic susceptibility profiles of new probiotic Lactobacillus and Bifidobacterium strains. International Journal of Food Microbiology, 98, 211–217. https://doi.org/10.1016/j. ijfoodmicro.2004.05.011 G. Santamarina-García et al.