Occurrence and Transfer by Conjugation of Linezolid-Resistance Among Non-Enterococcus faecalis and Enterococcus faecium in Intensive Pig Farms

Abstract

Enterococcus spp. are opportunistic and nosocomial pathogens. Intensive pig farms have been recently described as important hotspots for antibiotic resistance and reservoirs of potentially pathogenic enterococci, including other species than the most known E. faecalis and E. faecium. Here, we identified Linezolid-resistant non-E. faecalis and E. faecium (NFF) Enterococcus strains isolated from different production stages (suckling piglets, weaning pigs, and fatteners) across six intensive pig farms. The transferability of the linezolid-resistance determinants was assessed by bacterial conjugation and strains were also characterized for biofilm production, hemolytic and gelatinase activity. Among 64 identified NFF Enterococcus strains, 27 were resistant to at least three different antibiotic classes and 8/27 specifically to Linezolid. E. gallinarum and E. casseliflavus both transferred their Linezolid resistance determinants to the main pathogenic species E. faecium. Remarkably, this is the first report of the optrA gene transfer from E. casseliflavus to E. faecium by conjugation, which can greatly contribute to the spread of antibiotic resistance genes among pathogenic enterococcal species. The “weaning pigs” stage exhibited a significantly higher number of antibiotic-resistant enterococci than the “fatteners”. These findings highlight the importance of monitoring pig farms as hotspots for the spread of antibiotic-resistant enterococci, especially in the early stages of production. Furthermore, they underscore the significant role of NFF Enterococcus species as carriers of antibiotic resistance genes, even to last-resort antibiotics, which may be transferable to the major enterococcal species.

1. Introduction

Enterococci are Gram-positive bacteria, originally classified under the Streptococcus genus and only in 1984 recognized as a proper genus [1]. They belong to lactic acid bacteria and are commensals in the human gastrointestinal tract, but they can also be pathogenic, particularly as leading agents of healthcare-associated infections (HAIs) [2]. Indeed, enterococci are involved in cases of septicemia, endocarditis and urinary tract infections, often insensitive to antibiotic treatment and even showing recurrence. E. faecalis is the most common Enterococcus species detected within the human gut, followed by E. faecium, E. casseliflavus, E. durans, and E. gallinarum, present with different abundances [3], and together with E. faecium constitutes the main species involved in HAIs [2]. Enterococci are also able to adapt to different environmental conditions (i.e., wide ranges of temperature [10–45 °C] and pH [4.9 to 9.5]) and can be found in soil, water and in the gut of animals, where, similar to humans, they exhibit a dual nature, being involved in bacterial diseases as well. This is true not only for E. faecalis and E. faecium but also for other species such as E. durans and E. hirae [4].

Furthermore, enterococci are intrinsically resistant to several antibiotics and, due to their genetic plasticity, are prone to acquire antibiotic resistance genes (ARGs) through horizontal gene transfer [5], encoded by mobile genetic elements, making them “infamous for their resistance to many antibiotics” [6]. Currently, resistance to Ampicillin, high doses of aminoglycosides, Quinupristin-Dalfopristin, and Vancomycin are among the most dangerous phenotypes for the clinical management of infections. This, within the context of the spread of antibiotic-resistant bacterial infections, is contributing to an increase in patient morbidity and mortality [7]. Additionally, particularly important in the treatment of enterococcal infections is Linezolid (LNZ), the only antibiotic approved by the FDA for treating infections caused by Vancomycin-resistant enterococci [5]. LNZ, introduced in clinical use in 2000, is mainly used to treat Gram-positive bacterial infections and is often considered as a last-resort antibiotic [8]. Although LNZ is again effective against Enterococcus, with the incidence of LNZ resistance in this bacterium being <1%, several outbreaks of LNZ resistant enterococci have occurred globally [9]. Some mutations in the central loop of the V domain of 23S rDNA are considered the most common cause of chromosome-encoded resistance; however, a more worrisome hazard is represented by transferable genes, often encoded by plasmids and/or transposons, which can be shared between different bacterial clones [10,11]. To further complicate this scenario, mobile genetic elements carrying antibiotic resistance determinants have been shown to be selected and transferred even under the pressure of low heavy-metal concentrations within the environment [12]. This underscores the importance of monitoring the dissemination of this antibiotic resistance not only in clinical settings and in E. faecalis and E. faecium, but also in other environments such as farms and across other species of enterococci.

Animal farms are significant hotspots of antibiotic-resistant bacteria and ARGs [13,14,15]. In particular, swine farms, especially considering the global increase in pork production [16], deserve closer attention in this regard. Florfenicol use in intensive farming, regulated in Italy by Legislative Decree n. 218/2023 and in line with EU Regulation n. 2019/6, could be associated with the selection of linezolid-resistant enterococci due to co-resistance mechanisms mediated by mobile genetic elements. Genes such as optrA and cfr can confer resistance to both antibiotics when located on the same plasmid or transposon. Moreover, optrA alone can confer cross-resistance to both antibiotic classes [17,18,19]. This poses a serious health risk, since antibiotic resistant (AR) enterococci can be responsible for food contamination, especially during animal slaughtering and meat manipulation, and for the spread of ARGs to human pathogenic species, thus further contributing to the antibiotic resistance development. In our previous study, we investigated the presence and distribution of class 1 integrons, which are considered a proxy for the dissemination of ARGs in the environment [20], as well as of the distribution of enterococci, focusing on E. faecalis and E. faecium across different pig growth stages of six intensive pig farms. We found that the “Suckling piglets” production stage was particularly important for the potential dissemination of the selected targets [21].

However, other Enterococcus species, potentially displaying antibiotic resistance [22,23], were not considered in that study. While they pose a minor threat to human health, different reports have highlighted the isolation of LNZ resistant non-E. faecalis and E. faecium (NFF) strains from the environment [24,25]. Therefore, there is a significant risk of resistance spreading to major pathogenic species via horizontal gene transfer, mediated by the consumption of contaminated meat. Among the NFF Enterococcus spp., E. casseliflavus and E. gallinarum are again the most significant, often involved in bloodstream, intra-abdominal or catheter-associated infections [26]. Their presence is also documented in environmental reservoirs, since they are included in the gut microflora even of livestock animals [27]. The human colonization easily occurs, directly in veterinarians or farm workers and indirectly in the community, by the consumption of contaminated food.

In this study, we aimed to isolate, identify and characterize the antibiotic resistance profile and virulence of NFF Enterococcus species isolated from different pig production stages at the previously investigated farms. Additionally, we focused on the LNZ resistant enterococci, analyzing their potential to transfer LNZ resistance genes via conjugation to the clinically relevant species E. faecium.

2. Materials and Methods

2.1. Bacterial Strains, Culture Media and Antibiotics

The enterococcal strains included in this study were previously isolated [21], from six traditional closed-cycle fattening pig farms in Piedmont (North-Western Italy). The farms were sampled three times per year in 2017 and 2018, recovering stool samples from animals, divided in three production stages, i.e., suckling piglets (15–21 days old), weaning pigs (50–60 days old) and fatteners (200–250 days old). A total of 171 enterococcal isolates were obtained. After identification of the main pathogenic species E. faecalis and E. faecium by VITEK MS (BioMérieux Italia, Florence, Italy), the isolates considered NFF were selected.

The reference strains E. faecalis ATCC 29212 and E. faecium 64/3 belong to the Microbiology strain collection of the Department of Biomolecular Sciences, University of Urbino. All the strains were grown in Brain Hearth Infusion (BHI) broth or agar plates and conserved in 25% glycerol stocks in the same medium at −80 °C.

All the media were purchased from Oxoid (Thermofisher, Scientific, Basingstoke, UK), while all antibiotics from Sigma-Aldrich (Milwaukee, WI, USA).

2.2. Molecular Identification

The strains were previously identified by VITEK MS (Biomerieux Italia, Florence, Italy) [21] and then by molecular assays. The molecular identification of the isolates was performed by amplifying species-specific genes, as reported in Supplementary Table S1 [28,29,30], along with the corresponding amplification protocols. DNA extraction was performed by an in-house developed protocol: bacterial cultures in BHI supplemented with glycine 0.4% were incubated at 37 °C for 1 h in a first lysis buffer (10 mM Tris-HCl pH 8, 100 mM NaCl, 1 mM EDTA pH 8, 20% sucrose and lysozyme 2.5 mg/mL) and then for a further hour at 60 °C in a second lysis buffer (50 mM KCl, 10 mM Tris pH 8.3, 0.1 mg/mL gelatin, 0.45%, Nonidet P-40, 0.45% Tween 20, 5 μg/mL Proteinase K). After a final heating process at 95 °C for 10 min, DNA samples were stored at −20 °C before being processed. The reactions were carried out using 1X buffer, 10 µM of each primer, and 1 U of yourSial Taq (S.i.a.l. Group, Roma, Italy). The PCR products were visualized through gel electrophoresis on a 1% agarose gel (GellyPhor, Euroclone, Milan, Italy). The obtained results were further confirmed by biochemical assays, i.e., by MALDI-TOF.

2.3. Antibiotic Resistance Phenotype

The strains were primarily screened by plating on BHI agar plates supplemented with LNZ, Vancomycin, Erythromycin, Tetracycline, Streptomycin and Gentamicin. Antibiotics were added at the concentration corresponding to the resistance breakpoints for each drug, whose values are indicated in the CLSI guidelines [31]. The strains were inoculated at an optical density (OD₆₀₀) 0.1 and growth on the plates was monitored for 24 h. The isolates who exhibited growth on BHI plates were further characterized by determining the specific antibiotic Minimum Inhibitory Concentrations (MICs) using the broth microdilution method in cation-adjusted Mueller-Hinton II (MHII) broth. MIC values were interpreted based on CLSI clinical breakpoints, with incubation of MIC plates for 20 h at 37 °C. E. faecalis ATCC 29212 was used as quality control strain.

2.4. Antibiotic Resistance Genotype

The presence of selected genes responsible for resistance to LNZ, Erythromycin, Tetracycline, Streptomycin and Gentamicin was investigated using specific PCR assays for all the resistant strains, performed as described above, using primer pairs and amplification protocols, reported in Supplementary Table S2, targeting the most common and transferable genes [32,33,34,35,36,37,38,39,40,41].

2.5. Virulence Phenotype

2.5.1. Hemolysis Test

The hemolysis test was performed using a protocol similar to the CAMP test [42], with modifications. Bacterial strains were first inoculated into 5 mL of Luria–Bertani broth and incubated overnight at 37 °C. The cultures were then diluted in sterile saline solution to an OD₆₀₀ of 0.1, and a 10 µL aliquot was spotted onto Luria–Bertani agar plates supplemented with 5% (v/v) sheep blood (Sigma). Each strain was tested in three technical replicates. The inoculated plates were incubated at 37 °C for 24–48 h under aerobic conditions. Hemolytic activity was evaluated by observing the presence of clear zones (beta-hemolysis), greenish discoloration (alpha-hemolysis), or no visible change (gamma-hemolysis) surrounding the spots.

2.5.2. Gelatinase Production

The production of gelatinase was assessed by the gelatin hydrolysis test. Bacterial strains were inoculated into 5 mL of Nutrient Gelatin medium at an OD₆₀₀ of 0.1 and incubated at 37 °C for 12–15 h. After incubation, the tubes were transferred to 4 °C for 1 h to evaluate gelatin hydrolysis, indicated by the liquefaction of the medium.

2.5.3. Biofilm Plate Assay

The biofilm plate assay was carried out following the protocol described by Baldassarri et al. [43]. Briefly, biofilms were developed in 96 well plates, inoculated with bacterial suspensions in BHI broth at an OD₆₀₀ of 0.1 and incubated at 37 °C for 24 h. Each well was rinsed three times with sterile saline, then biofilms were stained with 1% crystal violet for 15 min, rinsed again with saline and let dry for 1 h. Crystal violet was resuspended in 96% ethanol for 20 min, then its absorbance quantified at 570 nm. Each enterococcal strain was tested at least in technical triplicate in two independent experiments.

2.6. Molecular Typing for Linezolid Resistant Strains

2.6.1. Molecular Typing by Random Amplified Polymorphic DNA PCR (RAPD-PCR) for E. gallinarum

The E. gallinarum molecular typing was performed by Random Amplified Polymorphic DNA PCR (RAPD-PCR), as described above using the primer pair F′ TGCTCTGCCC R′GTAGACCCGT and the following amplification protocol: 1′ at 95 °C, 30″ at 95 °C, 30″ at 34 °C, 4′ at 72 °C, 1′ at 72 °C, number of cycles 40. The results were visualized by gel electrophoresis and compared between the different tested isolates.

2.6.2. Molecular Typing by Pulsed Field Gel Electrophoresis (PFGE) for E. casseliflavus

The E. casseliflavus strains were typing using the Pulsed Field Gel Electrophoresis (PFGE) protocol previously described by Vignaroli et al. [44].

2.7. Filter Mating Experiments

Conjugation experiments were performed by filter mating as previously described [45] using E. faecium 64/3 as recipient. Transconjugants were selected on BHI agar plates containing Fusidic acid, Rifampicin and Florfenicol (all at 10 µg/mL). Conjugation frequency was expressed as number of transconjugants per recipient cell. All transconjugants were confirmed as E. faecium by species-specific PCR, targeting the ddl gene, using a previously described primer pair (F-TAGAGACATTGAATATGCC; R-CTAACATCGTGTAAGCT [28]). Moreover, the gene transfer was confirmed both phenotypically, by LNZ MIC determination, and genotypically, by amplifying the gene of interest as described above. Only the isolates presenting the resistance genes were regarded as true transconjugants. Each conjugation was verified in two independent experiments.

2.8. Statistical Analysis

The statistical analyses were made in the R environment v 4.1.2 [46] with the aim to investigate if the differences in phenotypic and genotypic antibiotic resistance were related to farming category and/or Enterococcus species. For both phenotypic and genotypic resistance data, we created a presence/absence matrix, where we inserted, for each isolate, 0 = “not resistant” or 1 = “resistant”. First, we assessed if the single phenotypic and genotypic resistances were a function of the interaction between the farming category and Enterococcus species (1), using a generalized linear model (GLM) with a binomial error distribution. If the interaction was not significant, the model was repeated excluding it from the analysis (2). Moreover, we calculated the total resistances for each isolate as the sum of the single resistances and tested them by negative binomial generalized linear model (NB-GLM) (3,4).

(1)

glm (pheno- or genotypic resistance ~ Category × Species, family = binomial)

(2)

glm (pheno- or genotypic resistance ~ Category + Species, family = binomial)

(3)

nb-glm (sum of pheno- or genotypic resistances ~ Category × Species)

(4)

nb-glm (sum of pheno- or genotypic resistances ~ Category + Species)

Starting from the presence/absence data of phenotypic and genotypic resistance, we built a matrix of distances among samples based on Jaccard’s dissimilarity index. We used the matrixes to plot sample composition in a tree, after hierarchical cluster analysis, and applied to them a PERMANOVA to evaluate if sample composition was a function of farming category and Enterococcus species. The pheno- and genotypic dissimilarity matrixes were correlated by Mantel test (Pearson’s correlation analysis) and considered correlated for r > 0.6 and p < 0.05.

3. Results

3.1. Identification

The 64 strains isolated from swine fecal samples were previously identified as NFF Enterococcus by mass spectrometry and subsequently confirmed by PCR assays as E. casseliflavus (22), E. hirae (18), E. gallinarum (16), E. avium (7) and E. durans (1).

3.2. Antibiotic Resistance Phenotype and Genotype

Out of 64 isolated strains, 50 showed a resistance at pheno- and genotypic level (Figure 1 and Figure 2, and

Table 1. Antibiotic resistance profile of the non-Enterococcus faecalis and Enterococcus faecium isolates.

Strains ID	MIC μg/mL					Genotype
	Linezolid	Tetracycline	Gentamicin	Streptomycin	Erythromycin
E. casseliflavus 2 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; ant; ermB
E. casseliflavus 10 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	ermB
E. casseliflavus 55 *	S (<4)	R (>128)	R (>500)	S (<1000)	R (>8)	tetM; tetL; aac(6′)-aph(2″); ermB
E. casseliflavus 61 *	R (8)	R (>128)	S (≤500)	S (<1000)	R (>8)	cfr(D); optrA; tetM; ermB
E. casseliflavus 62 *	R (8)	R (>128)	R (>500)	S (<1000)	R (>8)	cfr(D); optrA, aac(6′)-aph(2″); ermB
E. casseliflavus 63	S (<4)	S (<8)	S (≤500)	S (<1000)	R (>8)	ermB
E. casseliflavus 65	R (8)	S (<8)	S (≤500)	S (<1000)	R (>8)	cfr(D); optrA; ermB
E. casseliflavus 75	S (<4)	R (>128)	S (≤500)	S (<1000)	S (<1–4)	tetM
E. casseliflavus 77	S (<4)	R (16)	S (≤500)	S (<1000)	R (>8)	ermB
E. hirae 14	S (<4)	R (>128)	S (≤500)	R (>1000)	S (<1–4)	tetM; tetL; tetK
E. hirae 15 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 16 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 17 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 18 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 19 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 21	S (<4)	R (128)	S (≤500)	S (<1000)	S (<1–4)	tetM; tetK
E. hirae 22 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 23 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 33 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; tetK; ermB
E. hirae 34	S (<4)	R (128)	S (≤500)	S (<1000)	S (<1–4)	tetM; tetK
E. hirae 36 *	S (<4)	R (128)	S (≤500)	R (>1000)	R (>8)	tetM; tetK; ant; ermB
E. hirae 37 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; ermB
E. hirae 38 *	S (<4)	R (>128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; ermB
E. hirae 42	S (<4)	R (>128)	S (≤500)	S (<1000)	S (<1–4)	tetM; tetL
E. hirae 43	S (<4)	R (>128)	S (≤500)	R (>1000)	S (<1–4)	tetM; tetL; ant
E. hirae 44	S (<4)	R (>128)	S (≤500)	R (>1000)	S (<1–4)	tetM; tetL; ant
E. hirae 73	S (<4)	R (128)	S (≤500)	S (<1000)	S (<1–4)	tetM
E. gallinarum 1	S (<4)	S (<8)	R (>500)	S (<1000)	S (<1–4)	aac(6′)-aph(2″)
E. gallinarum 4	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	ermB
E. gallinarum 13 *	S (<4)	R (>128)	R (>500)	R (>1000)	R (>8)	tetM; aac(6′)-aph(2″)
E. gallinarum 29	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetM; ermB
E. gallinarum 35	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetM; tetL; ermB
E. gallinarum 46 *	R (8)	R (>128)	S (≤500)	S (<1000)	R (>8)	cfr(D); optrA, tetM; ermB
E. gallinarum 47 *	R (8)	R (>128)	R (>500)	S (<1000)	R (>8)	cfr(D); optrA, tetM; aac(6′)-aph(2″); ermB
E. gallinarum 48 *	R (8)	R (>128)	R (>500)	S (<1000)	R (>8)	cfr(D); optrA, tetM; aac(6′)-aph(2″); ermB
E. gallinarum 49 *	R (8)	R (>128)	R (>500)	S (<1000)	R (>8)	cfr(D); optrA, tetM; aac(6′)-aph(2″); ermB
E. gallinarum 50 *	R (8)	R (>128)	R (>500)	S (<1000)	R (>8)	cfr(D); optrA, tetM; aac(6′)-aph(2″); ermB
E. gallinarum 54 *	S (<4)	R (>128)	R (>500)	R (>1000)	R (>8)	tetM; aac(6′)-aph(2″)
E. gallinarum 56	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetM; tetL; ermB
E. gallinarum 57	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetM; tetL; ermB
E. gallinarum 58	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetL; ermB
E. gallinarum 59	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetM; ermB
E. avium 12	S (<4)	R (128)	S (≤500)	S (<1000)	R (>8)	tetM; tetO
E. avium 20 *	S (<4)	R (128)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; ermB
E. avium 40 *	S (<4)	R (64)	R (>500)	R (>1000)	R (>8)	tetM; ermB
E. avium 51	S (<4)	R (64)	S (≤500)	S (<1000)	R (>8)	tetM
E. avium 67 *	S (<4)	R (>128)	S (≤500)	S (<1000)	R (>8)	tetM; tetL; ermB
E. avium 74 *	S (<4)	R (64)	S (≤500)	R (>1000)	R (>8)	tetM; tetL; ermB
E. avium 76	S (<4)	R (64)	S (≤500)	S (<1000)	R (>8)	tetM
E. durans 11	S (<4)	R (128)	S (≤500)	S (<1000)	S (<1–4)	tetO

MIC resistance breakpoints (CLSI) were as follows: Linezolid, resistant at 4 μg/mL; Tetracycline, resistant at >16 μg/mL; Gentamicin, resistant at >500 μg/mL; Streptomycin, resistant at >1000 μg/mL; Erythromycin, resistant at >8 μg/mL. The strains with * are Multidrug resistant isolates. S = susceptible, I = intermediate, R = resistant.

). We found 28 resistant isolates in Suckling Piglets category (3 E. avium, 2 E. casseliflavus, 1 E. durans, 5 E. gallinarum, 17 E. hirae), 16 resistant isolates in Weaning Pigs (1 E. avium, 5 E. casseliflavus, 10 E. gallinarum) and 6 resistant isolates in Fatteners (3 E. avium, 2 E. casseliflavus, 1 E. hirae) (Figure 1).

At phenotypic level, the most frequent resistance was the one against Tetracycline (47/50 isolates), followed by the resistances against Erythromycin (40/50 isolates), Gentamicin (22/50 isolates), and LNZ (8/50 isolates) (Figure 2A). Moreover, 27 strains were Multidrug resistant (MDR) (4 E. avium, 5 E. casseliflavus, 7 E. gallinarum, 11 E. hirae), showing resistance against, at least, three different classes of antibiotics (Figure 2A). No resistance against Vancomycin was determined. At genotypic level, tetM was the most frequently detected ARG (41/50 isolates), followed by ermB (35/50 isolates) and tetL (23/50 isolates) (Figure 2B). The vanA, vanB, poxtA, ermA, mef genes were not found. At phenotypic level, only the resistance against Erythromycin showed a significantly different response, resulting significantly more frequent in Weaning Pigs than in Fatteners (GLM pairwise comparison: p = 0.0138). No differences among samples were observed for the other resistances, neither in terms of farming category nor of species. The sum of the single phenotypic resistances per sample significantly varied according to production stage and species: it was significantly higher in Weaning Pigs than in Fatteners (NB-GLM pairwise comparison: p = 0.0182) and significantly lower in E. casseliflavus than in E. avium (GLM pairwise comparison: p = 0.0030) and E. hirae (GLM pairwise comparison: p = 0.0139). At genotypic level, ermB had a significantly higher occurrence in Weaning Pigs than in Fatteners (NB-GLM pairwise comparison: p = 0.0414), whereas tetL and tetM genes were significantly less frequent in E. casseliflavus than in E. hirae (GLM pairwise comparison: p = 0.0067) and E. gallinarum (GLM pairwise comparison: p = 0.0448), respectively. The sum of the single genotypic resistances per sample significantly varied according to production stage and species and considering the interaction between the two factors (NB-GLM: p ≤ 0.0315). In particular, it was lower in E. casseliflavus than in E. hirae in Suckling Piglets (GLM pairwise comparison: p = 0.0067) and in E. casseliflavus than in E. avium in Fatteners (GLM pairwise comparison: p = 0.0029), respectively. At phenotypic level, in terms of sample composition, 12.6 and 9.7% of variance were explained by species and production stage, respectively (PERMANOVA: p ≤ 0.027).

Indeed, in the tree depicting sample composition, isolates only limitedly clustered based on these factors (Supplementary Figure S1A). In case of genotypic resistance data, the species explained 20.8% of variation in sample composition (PERMANOVA: p = 0.001) while the category accounted for a lesser extent (5.2%, PERMANOVA: p = 0.074). The tree, obtained after hierarchical linkage analysis, showed that, although no net clusters were formed, isolates mainly grouped following the species (Supplementary Figure S1B). Moreover, the sample composition at phenotypic level was not correlated with the one found at genotypic level (MANTEL TEST: r = 0.578, p = 0.001).

3.3. Virulence Phenotype

Virulence of the eight LNZ-resistant strains was assessed through biofilm production, hemolysis and gelatinase activity. No strain showed hemolytic activity nor resulted positive to the gelatinase test. Only two strains, E. casseliflavus 61 and E. gallinarum 49, which were also MDR, were strong biofilm producers, while the remaining ones resulted weak producers.

3.4. Genetic Diversity Among the Linezolid Resistant Strains

The LNZ-resistant NFF Enterococcus typing was performed using both PFGE and the RAPD-PCR method, for E. casseliflavus and E. gallinarum, respectively, to assess the genetic diversity among the LNZ-resistant strains. The results revealed that the three strains of E. casseliflavus were grouped into two distinct clones (Figure 3B). Similarly, the five strains of E. gallinarum were classified into two separate clones (Figure 3A).

3.5. Conjugation Experiments

The conjugation experiments, performed by filter mating, demonstrated the transfer of genetic material from selected LNZ-resistant donor strains of E. casseliflavus and E. gallinarum to the recipient strain E. faecium 64/3. The results of conjugation are reported in

Table 2. Conjugation experiments between E. casseliflavus and E. gallinarum (donors) and the recipient E. faecium 64/3.

Donors Strains	Donors CFU/mL	Recipient CFU/mL	Transconjugant CFU/mL	Transfer Frequency *
E. casseliflavus 61	5.3 × 108	4 × 1010	8.30 × 105	2.08 × 10−5 ± 9.69 × 10−6
E. casseliflavus 62	3.4 × 108	1.80 × 1010	5.60 × 104	3.11 × 10−6 ± 4.17 × 10−7
E. gallinarum 46	7.3 × 108	1.3 × 1010	1.3 × 105	1 × 10−5 ± 2.58 × 10−6
E. gallinarum 49	7.5 × 108	2 × 1010	1.15 × 104	5.75 × 10−7 ± 1.06 × 10−8

* evaluated on the base of the transconjugant/recipient frequency.

.

The results confirm the successful transfer of genetic material between the tested strains, though with varying efficiencies, highlighting differences in the conjugation potential among the donor strains. Transconjugants were confirmed by evaluating resistance phenotypes and genotypes. All transconjugants exhibited resistance to LNZ with MIC values of 8 µg/mL. The presence of resistance genes (cfr(D) and optrA) varied among transconjugants: in the mating between E. faecium 64/3 and E. casseliflavus 61, 10 out of 10 transconjugants carried the optrA gene, while 1 of 10 also harbored the cfr(D) gene. In the mating between E. faecium 64/3 and E. casseliflavus 62, all 10 transconjugants possessed the optrA gene, and 3 of 10 carried the cfr(D) gene. All transconjugants derived from E. gallinarum 46 and E. gallinarum 49 displayed the optrA gene, but the cfr(D) gene was undetected (

Table 3. Phenotype and genotype of linezolid resistance in the transconjugants.

Transconjugants	Strek	Linezolid MIC μg/ml	ddl E. faecium	cfr(D)	optrA
E. casseliflavus 61 x E. faecium 64/3	10/10	8	10/10	+ (1/10)	+ (10/10)
E. casseliflavus 62 x E. faecium 64/3	10/10	8	10/10	+ (3/10)	+ (10/10)
E. gallinarum 46 x E. faecium 64/3	10/10	8	10/10	-	+ (10/10)
E. gallinarum 49 x E. faecium 64/3	10/10	8	10/10	-	+ (10/10)

).

4. Discussion

NFF Enterococcus spp. are often neglected species, since not considered as relevant pathogens to mankind. However, they are often found in animal hosts and can easily develop and spread ARGs, even through the food chain. Indeed, these species represent a hazard for humans: the zoonotic transmission of AR isolates, selected upon antibiotic use in farms, and the subsequent colonization of the human intestinal environment, could lead to the spread of resistance determinants and even to modifications of the human normal microflora [47]. Such a risk has already been described, especially in farm workers, who are exposed to the colonization by environmental bacteria by both direct contact with the animals and by the consumption of meat products and vegetables produced in the farm environment. The direct consequence is an enrichment of ARGs in the human gut resistome, most likely mediated by the gene transfer from the environmental isolates, which is more evident according to the time spent in the farms themselves [48].

The percentage of NFF Enterococcus isolated strains from the sampled intensive pig farms was 37% (with 107 out of 171 isolated Enterococcus strains identified as either E. faecalis or E. faecium by mass spectrometry, according to Di Cesare et al. [21]). This percentage is lower than the 48–67% range obtained in previous studies from pig farms [49,50]. This discrepancy could be attributed to the source of samples, as we only analyzed pig fecal samples, whereas the previous studies investigated also feed or environmental samples from pig farms, which may have led to a higher frequency of NFF Enterococcus strain isolation. This choice was purposely made to specifically investigate the health risk connected with the consumption of contaminated meat products, mostly the raw or undercooked ones, hence with the contamination of livestock animals by NFF Enterococcus spp. The involvement of putative enterococcal reservoirs (i.e., water, environment and feeding) is currently under investigation.

Noteworthy, considering only NFF Enterococcus strains, a percentage close to 80% of the isolates resulted antibiotic resistant with a significant number of MDR (42%). This clearly highlights the potential contribution of the NFF Enterococcus species isolated from pig farms as reservoirs of antibiotic resistance. We focused our attention on transferable genes, mostly encoded by mobile genetic elements to assess the risk of vehiculating resistance determinants to the main species E. faecalis and E. faecium, well-known human pathogens. Significantly, no Vancomycin-resistant strain was recovered from the preliminary screening. This may indicate that the banning of Avoparcin, which is known to select for Vancomycin-coresistance, from veterinary use in 1997 has produced the desired effect, i.e., the limitation of the Vancomycin resistant phenotype to the clinical setting and the depletion of environmental reservoirs. Both E. casseliflavus and E. gallinarum present Vancomycin resistance determinants (i.e., vanC1 and vanC2) encoded by the chromosome; however, these are known for conferring low-level resistance, which is not sufficient to produce a resistant phenotype, according to the values recommended by the CLSI guidelines [31]. The antibiotic surveillance is generally focused on “true-resistant” phenotypes, such as those conferred by the vanA or vanB genes. The “Weaning Pigs” resulted the pig production stage that warrants more attention regarding the potential contribution of NFF Enterococcus strains in disseminating antibiotic resistance. This might be due to the change and introduction of animal feeds, putatively vehiculating environmental enterococcal species, other than E. faecalis and E. faecium, exhibiting antibiotic resistance [51]. Consistently, in another study a comparable pig production stage named “Weaners” was identified as a category of interest for the spread of ARGs [13]. This claims for more research focused on the role of the different stages of production of pigs for the dissemination of ARGs.

Furthermore, a total of eight strains (close to 11%) of NFF Enterococcus strains were resistant to LNZ, specifically 5 E. gallinarum and 3 E. casseliflavus; both species have previously been reported as LNZ resistant in food producing pigs [52]. We purposely focused the molecular typing on this strain subset, considering the importance of the spreading LNZ resistant clones. Both RAPD-PCR and PFGE were performed for the typing of E. gallinarum and E. casseliflavus, respectively. We acknowledge that whole genome sequencing would provide more comprehensive and descriptive information about the enterococcal epidemiology. However, we have adopted these screening techniques to identify the main enterococcal clones present within our collection and to avoid redundances in the strain characterization. Interestingly, the genotype of LNZ resistance of these strains was identical, consisting of the co-occurrence of two transferable resistance genes, cfr(D) and optrA. The cfr gene was first detected in an E. faecalis strain isolated from a farm animal and carried by the plasmid pEF-01 [53] and only in 2014, for the first time, E. gallinarum and E. casseliflavus isolated from swine farms were reported as positive for this gene [54]. The optrA gene was first detected in an E. faecium strain isolated from human in 2005 and has since been widely disseminated among enterococci isolated from human and then from animal origin samples [55]. The optrA determinant is typically detected in E.faecalis and E. faecium, only sporadically in E. gallinarum [51], and more recently in E. casselliflavus isolated from pigs [56,57]. Our results clearly demonstrate that also the NFF Enterococcus species are important carriers of ARGs particularly interesting at clinical level. Moreover, the two genetically distinct strains of both isolated species were capable of transferring the optrA gene by conjugation. On the other hand, only E. casselliflavus strains were able to transfer the cfr(D) gene to E. faecium, aligning with the previous studies documenting the horizontal transfer of these genes among Enterococcus spp. strains [58,59,60,61,62]. The localization of the two genes in the two NFF Enterococcus spp., as well as the features of the involved genetic elements are currently being evaluated and will represent the topic for future works; in the current paper, our main aim was to prove the potentiality of the LNZ resistance gene transfer from NFF Enterococcus spp. to a pathogenic species, such as E. faecium. Although this limits the information about the specific mechanisms of the spreading of LNZ resistance, it should be underlined that such gene transfer is worrisome, since NFF Enterococcus spp. are often neglected in the monitoring of antibiotic resistance. A more detailed analysis (i.e., by whole genome sequencing, S1-PFGE and blotting techniques) is in progress on the enterococcal strains presenting the higher tendency to spread resistance determinants and will provide valuable insights into this phenomenon in future papers. We chose to select the E. faecium transconjugants in the presence of Florfenicol to reliably mimic the antibiotic pressure the strains are exposed to within the farming environment [17]. The efficient transfer of both genes underlines two worrisome concepts: (a) resistance to clinic-restricted drugs can be developed by exposure to other antibiotics, which can share transferable resistance determinants and cause co-resistance [19], thus demanding a high control and update of the procedures associated with livestock health and handling; (b) the NFF Enterococcus spp. may represent an overlooked hazard, not only for their involvement in infections, but even and more importantly for the spread of resistance determinants. Indeed, to be best of our knowledge, this is the first study reporting the optrA gene transfer by conjugation from E. casseliflavus to a pathogenic enterococcal species, which had been previously obtained only by transformation [62]. This is an outstanding result, since enterococci are well-known to share genetic information, including ARGs, by conjugation and the possible transfer of resistance determinants, even to last resort drugs, to the main pathogenic species should not be neglected in the surveillance of antibiotic resistance in the environment. This further highlights the remarkable risk associated with the presence of NFF Enterococcus species in intensive pig farms.

Interestingly, two strains that successfully transferred LNZ resistance genes via horizontal gene transfer in vitro were also MDR and strong biofilm producers. This finding might suggest that, despite being less frequent causes of infection compared to E. faecalis and E. faecium, E. casseliflavus and E. gallinarum could still pose a pathogenic risk to humans, as reported in the literature [63]. The isolation rate of both these enterococcal species from bloodstream infections has increased in the latest years. Moreover, biofilm production facilitates bacterial adhesion to surfaces and persistence to disinfectants, thus posing a health risk for food contamination during the processing. Within the biofilm matrix, bacterial cells are stratified and in direct communication, which favor the quorum sensing signaling and, subsequently, the gene transfer by conjugation, especially for both virulence and antibiotic resistance genes [64,65]. Biofilm formation should thus be considered when assessing the hazard of enterococcal isolates.

5. Conclusions

Overall these results highlight that NFF Enterococcus species pose a direct threat to human health due to their resistance to last-resort antibiotics and virulence, as well as an indirect threat through their potential to transfer LNZ resistance genes via horizontal gene transfer. Interestingly, the optrA gene transfer by conjugation from an NFF Enterococcus spp. to the pathogenic species E. faecium was pointed out for the first time, further stressing the need for considering even these less frequent enterococcal species in the spread of resistance to last resort drugs, such as LNZ. Furthermore, the different stages of pig farming for food production have a quantitatively different impact on the presence of MDR and virulent NFF Enterococcus species. This reinforces the need for further research focused on this aspect and the related farming conditions to mitigate the human health risks associated with potential infections by these pathogens through the food chain, especially due to the consumption of raw or undercooked meat. A special concern should be posed about animal feed: indeed, the change from the maternal nutrition to the use of fodder could represent a source of contamination with antibiotic resistant NFF enterococci, as suggested by the higher number of resistant isolates recovered from weaning pigs. Surveillance practices should take into consideration even testing this material to avoid the diffusion of antibiotic resistance into livestock.