We purchased 10 probiotic products available for dogs and cats from retailers selling veterinary products in Hungary. The isolation of strains indicated on the products was carried out by the Department of Microbiology and Infectious Diseases at the University of Veterinary Medicine Budapest. The isolation of Enterococcus faecium strains was successful in all nine of the products listing it (9/9). Among the other strains listed on the products, we were able to isolate Lactobacillus plantarum in one case (1/4), Pediococcus pentasaceus in two cases (2/2), and Pediococcus acidilactici in one case (1/2). Additionally, certain products indicated the presence of Pediococcus strains (one strain) and Lactococcus strains (eight strains) in deactivated form. The species identity of the strains was determined using MALDI-TOF mass spectrometry (Bruker, Mannheim, Germany). The isolates were used for minimal inhibitory concentration (MIC) testing. The properties of each product are summarized in Table 1.

The classification of strains into categories is overseen by three main international organizations. Based on their standards, probiotic strains found in products are provided with unique identifiers. The CECT (Spanish Type Culture Collection) is a Spanish strain collection accredited with ISO 9001 standards (24). The NCIMB (National Collection of Industrial, Food and Marine Bacteria) is a privately-owned strain collection located in the United Kingdom (25). The DSM (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH) is a German strain collection (26).

Strain solutions were prepared from the active ingredients in accordance with the recommendations of the Clinical Laboratory Standards Institute (CLSI), utilizing materials from Merck KGaA, Darmstadt, Germany (27). For the experiments, strain solutions were prepared at a concentration of 1,024 μg/mL, with adjustments made for the purity specified by the manufacturer of each active ingredient. The experiments were conducted within a 2-fold dilution range, spanning from 128 to 0.125 μg/mL.

The phenotypic expression of resistance was examined by determining the minimum inhibitory concentration (MIC) values of each bacterial strain, which was conducted using the CLSI methodology (27). The breakpoints were determined based on the guidelines of both the CLSI and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) (28). The selection of antimicrobial agents was made based on international literature, aiming to cover the most frequently used antimicrobial groups.

The bacterial strains were stored at −80°C and inoculated into 3 ml of Mueller-Hinton broth (MHB) the day before the experiment, followed by incubation at 37°C for 18–24 h. The experiments were conducted using a 96-well microtiter plate (VWR International, LLC., Debrecen, Hungary). The first column of the working plates, filled with 90 μl of MHB, received a 4-fold dilution of the strain solutions, establishing an initial concentration of 128 μg/mL, from which a 2-fold dilution series was prepared, except in the last two columns of 96-well microtiter plate. The penultimate column served as a positive control (containing only bacteria), while the final column served as a negative control (containing only the drug). Bacterial suspension adjusted to 0.5 McFarland turbidity was inoculated onto the working plates up to the positive control column using a nephelometer (CheBio fejleszto Kft., Budapest, Hungary) (27). Evaluation took place after 24 h of incubation at 37°C using the SWIN automatic MIC reader (CheBio fejleszto Kft., Budapest, Hungary) and the VIZION system (CheBio fejleszto Kft., Budapest, Hungary). The reference isolate used was Escherichia coli (ATCC 25922).

Probiotic samples were prepared by using 0.1 g of probiotic powder dissolved in 1 ml PBS (1:10 dilution). Nucleic acid was extracted from the mixture using the Quick-DNA Fecal/Soil Microbe Miniprep Kit (Zymo Research, CA, USA) according to the manufacturer's instruction. Samples were disrupted by using a TissueLyzer LT Bead Mill (Qiagen, Germany). The concentration of purified DNA was measured with Qubit 2.0 equipment using the Qubit dsDNA BR Assay Kit (Thermo Scientific, Waltham, MA, USA).

Nucleotide sequences were determined by next-generation sequencing on an Illumina^® NextSeq 500 sequencer (Illumina, San Diego, CA, USA) following the reference guide provided by Illumina. Illumina^® Nextera XT DNA Library Preparation Kit (Illumina, San Diego, CA, USA) and Nextera XT Index Kit v2 Set A (Illumina, San Diego, CA, USA) were used to prepare Illumina specific libraries. DNA samples were diluted to 0.2 ng/μL in nuclease-free water (Promega, Madison, WI, USA) in a final volume of 2.5 μL. Reaction components were used at a reduced volume. For the tagmentation reaction, 5 μL Tagment DNA buffer with 2.5 μL AmpliconTagment Mix were used. During tagmentation, the samples were incubated at 55°C for 6 min, using the GeneAmp PCR System 9700 (Applied Biosystems/Thermo Fisher Scientific, Foster City, CA, USA). The samples were then allowed to cool to 10°C before the addition of 2.5 μL of the Neutralize Tagment buffer. Neutralization was performed for 5 min at room temperature. A total of 7.5 μL of the Nextera PCR Master Mix and 2.5 μL each of the i5 and i7 index primers were added to the tagmented DNA samples. The index primers were attached to library DNA via 12 PCR cycles (each cycle consisted of the following steps: 95°C for 10 s, 55°C for 30 s, followed by 72°C for 30 s). Following the PCR cycles, the samples were held at 72°C for 5 min and then at 10°C. Libraries were purified using Gel/PCR DNA Fragments Extraction Kit (Geneaid Biotech Ltd., Taipei, Taiwan). The concentration of the purified libraries was measured with Qubit 2.0 equipment using Qubit dsDNA HS Assay Kit (Thermo Fischer Scientific, Waltham, MA, USA). Library DNAs were pooled and denatured. The denatured library pool was loaded onto a NextSeq 500/550 Mid Output flowcell at a final concentration of 1.5 pM. Sequencing was performed using an Illumina^® NextSeq 500 sequencer (Illumina, San Diego, CA, USA).

During the bioinformatic data processing, quality control of the raw sequences was performed using FastQC v0.11.9 (29), followed by the removal of low-quality sections using TrimGalore v0.6.6 (30). The reads were assembled into longer sequences (contigs) using MEGAHIT v1.2.9 (31). From these contigs, all possible open reading frames (ORFs) were determined using Prodigal v2.6.3 (32). Protein sequences were derived from these ORFs based on their nucleotide sequences. Subsequently, the protein sequences were compared to antimicrobial resistance gene (ARG) sequences in The Comprehensive Antibiotic Resistance Database (CARD) using the Resistance Gene Identifier (RGI) v5.1.0 software (33). Only results reaching the specified threshold (95%) in the CARD database were retained.

The potential mobility of identified resistance genes was assessed using the MobileElementFinder v1.0.3 program (34) which predicted genes occurring as mobile genetic elements (MGEs) on contigs. Only those ARGs found within the longest composite transposon distance specific to the species in the database within the contig were considered as potentially mobile. Additionally, plasmid encoding was examined using the PlasFlow v1.1 software, and the presence of phage genomes on contigs was determined using the VirSorter v2.2.2 (35) software.

Table 2 summarizes the results of susceptibility testing of Enterococcus faecium strains isolated from the examined products. All the products contained the same strain (NCIMB10415) except for the J-product (DSM7134). Similar MIC values were observed for all strains. All the tested strains were sensitive to penicillin, amoxicillin, amoxicillin-clavulanate, oxytetracycline, doxycycline, clindamycin, tylosin, and vancomycin. For gentamicin (100%; MIC >32 μg/mL) and potentiated sulfonamides (sulfamethoxazole and trimethoprim in a 19:1 ratio; 100%; MIC 16 μg/mL), all strains were resistant. Regarding gatifloxacin, a fourth-generation fluoroquinolone, six strains were susceptible, while three were resistant (66.7%; MIC >2 μg/mL).

Table 3 summarizes the antibiotic susceptibility profiles of Lactobacillus and Pediococcus species. Among the Lactobacillus strains, one Lactobacillus plantarum strain was isolated, which showed sensitivity to penicillin, amoxicillin-clavulanate, gentamicin, oxytetracycline, doxycycline, clindamycin, tylosin, and vancomycin. It was resistant to amoxicillin (>1 μg/mL MIC), the combination of sulphonamides (>16 μg/mL MIC), and gatifloxacin (>2 μg/mL MIC). From the Pediococcus species, two Pediococcus pentasaceus and one Pediococcus acidilactici strains were isolated, all of which were susceptible only to clindamycin (< 8 μg/mL MIC), showing resistance to all other tested antibiotics.

Table 4 summarizes the results obtained from next-generation sequencing, including the isolated resistance genes from the products, their taxonomic origin, indicating the ARG family for each gene, the resistance mechanism determined by the respective gene, and the antibiotic groups against which resistance is conferred.

A total of 19 types of ARGs were identified, out of which 11 were found on plasmids, and no genes were identified on phages. Of particular concern is that in two cases, a plasmid-contained gene also acted as a mobile genetic element (MGE). One of these was the APH(3′)-Ia gene, found in the I-product. The activation of this resistance gene through enzymatic means (phosphotransferase) can inactivate aminoglycoside antibiotics (36). The other gene was the tetS gene, found in the A-product. This gene is capable of reducing sensitivity to tetracycline antibiotics through target protection (ribosomal mosaic), thereby preventing the binding of the antibiotic (37).

During the determination of MIC values, a correlation can be observed between resistant values and the identified ARGs in the Enterococcus faecium strains of various products. The expression of the AAC(6′)-Ii gene enzymatically (acetyl-transferase) leads to resistance to aminoglycoside antibiotics. This gene was found in all tested products and could be one of the genes responsible for resistance (≥32 μg/mL) observed in all strains. It's typically chromosomally located, but in one case (F-product), it was found on a plasmid. The eatAv gene confers resistance to lincosamides and pleuromutilins through target protection (ABC-F type). This gene was present in all Enterococcus faecium strains from the tested products. The efmA gene determines an MFS-type efflux pump for macrolides and fluoroquinolones, potentially contributing to resistance to gatifloxacin (2 μg/mL MIC) observed in one case (J-product). The rsmA gene determines an RND-type multidrug efflux pump, which can confer resistance to phenicols, fluoroquinolones, and diaminopyrimidines. This gene was also found in all Enterococcus faecium strains and may explain the observed resistance (>128 μg/mL MIC) to potentiated sulfonamides, gatifloxacin (2 μg/mL MIC) in the H-product, I-product, and J-product cases, as well as potential florfenicol resistance (8 μg/mL MIC).

All phenotypic resistances were successfully attributed to identified ARGs. Of particular concern is the high proportion of genes found on plasmids (57.9%), with two instances where these were also MGEs. Therefore, based on our findings, it would be necessary to conduct such examinations, similar to those required for food-producing animals, before introducing probiotic preparations intended for companion animals into circulation. In product A, containing strains of Lactobacillus plantarum (NCIMB41638), Lactobacillus fermentum (NCIMB41636), and Lactobacillus rhamnosus (NCIMB41640), no ARGs were found. Therefore, the phenotypic resistance observed in this case cannot be genetically supported by detected genes.

The B-product contained Enterococcus faecium (NCIMB10415), from which we identified 15 ARGs, out of which seven were located on plasmids and one (tetS) gene was on an MGE. Four genes were specifically derived from Enterococcus species, while the rest originated from Lactococcus, Klebsiella, Pseudomonas, Hafnia, Acinetobacter, and Streptococcus species. Similarly, in product C and product H containing Enterococcus faecium (NCIMB10415), we identified 3 ARGs, all of which were chromosomal genes originating from Enterococcus spp. The D-product, E-product, and J-product containing Lactobacillus plantarum (DSM12837), Pediococcus acidilactici (DSM16243), and Enterococcus faecium (NCIMB10415) species, respectively, all showed the presence of the same four ARGs. These genes were chromosomally located, and all originated from Enterococcus spp.

The F-product contained strains of Pediococcus pentasaceus (DSM1283U), Lactobacillus brevis (DSM12835), Lactobacillus bucherii (DSM12856), Lactobacillus plantarum (DSM12836), Lactobacillus rhamnosus (NCIMB30121), Lactobacillus acidophilus (CECTU529), and Enterococcus faecium (NCIMB10415), with Pediococcus and Lactobacillus species present in deactivated form in the product. Four resistance genes matched those found in the E-product. However, in the case of the F-product, the AAC(6′)-Ii gene was located on a plasmid, while the rest were chromosomal. All identified genes were attributed to Enterococcus spp. The composition of the G-product was identical to that of the F-product, but only two chromosomally located ARGs originating from Enterococcus spp. could be identified in the tablets.

The I-product, which contained the Enterococcus faecium (NCIMB10415) strain, contained the most identified ARGs, totaling 14. Of these, eight were found on plasmids, and among these, the APH(3′)-Ia gene was also an MGE. Five of the genes were originally of Enterococcus spp. origin, while the rest may have been acquired from other species through horizontal gene transfer, including Staphylococcus, Klebsiella, Lactobacillales, Enterobacteriaceae, Enterobacterales, and Streptococcus species.