Whole-genome long-read sequencing to unveil Enterococcus antimicrobial resistance in dairy cattle farms exposed a widespread occurrence of Enterococcus lactis

ABSTRACT Enterococcus faecalis (Efs) and Enterococcus faecium (Efm) are major causes of multiresistant healthcare-associated or nosocomial infections. Efm has been traditionally divided into clades A (healthcare associated) and B (community associated) but clade B has been recently reassigned to Enterococcus lactis (Elc). However, identification techniques do not routinely differentiate Elc from Efm. As part of a longitudinal study to investigate the antimicrobial resistance of Enterococcus in dairy cattle, isolates initially identified as Efm were confirmed as Elc after Oxford-Nanopore long-fragment whole-genome sequencing and genome comparisons. An Efm-specific PCR assay was developed and used to identify isolates recovered from animal feces on five farms, resulting in 44 Efs, 23 Efm, and 59 Elc. Resistance, determined by broth microdilution, was more frequent in Efs than in Efm and Elc but all isolates were susceptible to ampicillin, daptomycin, teicoplanin, tigecycline, and vancomycin. Genome sequencing analysis of 32 isolates identified 23 antimicrobial resistance genes (ARGs, mostly plasmid-located) and 2 single nucleotide polymorphisms associated with resistance to 10 antimicrobial classes, showing high concordance with phenotypic resistance. Notably, linezolid resistance in Efm was encoded by the optrA gene, located in plasmids downstream of the fexA gene. Although most Elc lacked virulence factors and genetic determinants of resistance, one isolate carried a plasmid with eight ARGs. This study showed that Elc is more prevalent than Efm in dairy cattle but carries fewer ARGs and virulence genes. However, Elc can carry multi-drug-resistant plasmids like those harbored by Efm and could act as a donor of ARGs for other pathogenic enterococcal species. IMPORTANCE Enterococcus species identification is crucial due to differences in pathogenicity and antibiotic resistance profiles. The failure of traditional methods or whole-genome sequencing-based taxonomic classifiers to distinguish Enterococcus lactis (Elc) from Enterococcus faecium (Efm) results in a biased interpretation of Efm epidemiology. The Efm species-specific real-time PCR assay developed here will help to properly identify Efm (only the formerly known clade A) in future studies. Here, we showed that Elc is prevalent in dairy cattle, and although this species carries fewer genetic determinants of resistance (GDRs) than Enterococcus faecalis (Efs) and Efm, it can carry multi-drug-resistant (MDR) plasmids and could act as a donor of resistance genes for other pathogenic enterococcal species. Although all isolates (Efs, Efm, and Elc) were susceptible to critically or highly important antibiotics like daptomycin, teicoplanin, tigecycline, and vancomycin, the presence of GDRs in MDR-plasmids is a concern since antimicrobials commonly used in livestock could co-select and confer resistance to critically important antimicrobials not used in food-producing animals.

Efs and 82 Efm), that is, one isolate for each Enterococcus species per sampling time and age group, were selected for MIC determination.
In all, 32 isolates (17 Efs and 15 Efm) were then selected for WGS based on the Enterococcus species, the phenotypic AMR profile, and the age group.ONT sequencing provided a median of 20,239 reads per sample [interquartile range (IQR) = 15,393-34,315] at a median of 405.4 Mb per sample (IQR = 356.7-654.2Mb), corresponding to 143× median of coverage value (IQR = 125-226×).The N50 median value was 25,712 bases per sample (IQR = 23,638-29,753).Assembly achieved circularized chromosomes in 29 of the 32 sequenced isolates.In addition, 38 plasmid contigs were identified from 26 isolates, all of which were successfully assembled into complete circular plasmids.Information regarding sequencing output and assembly statistics is provided in Table S1.
Taxonomic identification results with Kraken2 were in total agreement with RT-PCR1 results.However, the pangenome analysis revealed clear pattern differences between Efs and Efm, and separated the 15 Efm isolates into two subclusters, in both the dendrogram based on the presence or absence of accessory genes (Fig. 1A) and the phylogenetic tree based on core genome alignment (Fig. 1B).Both subclusters shared 13.8% of core genes (1,011/7,314, total size: 110,229 bases).To delve into these findings, genome-scale phylogenetic analysis of the sequenced genomes and the genomes of other Enterococ cus spp.was performed at the Type (Strain) Genome Server (TYGS) (Fig. S1).This analysis confirmed the identity of the 17 Efs.However, the 15 isolates identified by RTi-PCR1 as Efm were separated into two clusters: one that included six of our isolates along with other Efm isolates belonging to clade A2, and another cluster that included the remaining nine isolates along with clade B Efm isolates, Elc BT159 (DSM23655; type strain) and E. xinjiangensis (JCM30200; heterotypic synonym of Elc).The intergenomic comparison between these nine isolates and Elc BT159 resulted in a mean DNA:DNA hybridization d 4 value of 88.8 (range: 83.9-90.3),confirming their classification as Elc (by contrast, E. faecium NBRC100486 d 4 = 59.6 [range: 58.2-59.5]).G + C content of both Elc BT159 and E. faecium NBRC100486 was 38.1%.Based on this phylogenomic analysis, the nine isolates were more closely related to Elc than to any other Enterococcus species.

Design and performance of a real-time PCR assay for the specific identifica tion of E. faecium
Since RTi-PCR1 did not distinguish Efm from Elc, a new real-time PCR assay (RTi-PCR2) was designed to amplify the gluP gene.RTi-PCR2 produced an amplification signal for Efm (former Efm clade A) but not for Elc (formerly known as Efm clade B) or Efs, as shown when testing two control strains (kindly provided by Hospital Ramón y Cajal, Madrid) and the field isolates included in this study (6 Efm and 9 Elc).
When the 82 isolates identified as Efm by RTi-PCR1 were tested using RTi-PCR2, only 23 were confirmed as Efm, the remaining 59 being regarded as Elc (Fig. 2; Table S2).The isolation frequency of each of the enterococci species did not differ between age groups.Overall occurrence of enterococci (any species) was higher in calves compared to heifers (OR adj = 9.90 (1.80-54.44),P = 0.008).
MDR, defined as resistance to at least three antimicrobial classes, was mainly found in Efs (38.6% of isolates) recovered from all the farms (Fig. S2B), with 55.6% of the Efs isolates from calves, 27.3% from heifers and 26.7% from lactating cows being MDR.The only phenotypic MDR patterns present in all age groups were CHL-ERY-LZD-TET and CHL-ERY-TET.In addition, the profile CHL-ERY-TET was present in four of the five farms (Fig. 2).Farm F5 showed the highest percentage of MDR Efs isolates (83.3%) and farm F3 the lowest (12.5%) despite harboring the only isolate resistant to five antimicrobial classes.

Genomic composition of the isolates
The 32 genomes were assigned to 18 MLST types (STs), that is, 17 Efs belonged to 10 STs, six Efm to four STs, and nine Elc to four STs (Fig. 3).The most prevalent MLST type in Efs was ST21 (n = 5), two STs were detected twice (ST32 and ST771) in Efm, and ST296 predominated in Elc (n = 5).Efm and Elc did not share STs, and when assigned to a clonal complex, Efm isolates belonged to CC17 and Elc isolates to CC94 (Fig. 3).The phylogenetic tree based on the core genome single nucleotide polymorphisms (SNPs) was consistent with the MLST results (Fig. 1B and C).
A total of 79 virulence genes (71 chromosomally encoded, seven only present in plasmids, and one in both) were identified.The percentage of identity shared was above 90% for most genes from enterococci (n = 55) and ranged between 64.1% and 80.1% for genes from other Gram (+) bacteria (n = 24) (Table S3).VFs were more abundant in Efs, with isolate ENT0115 carrying the largest number of VFs, including the aggregation substance (AS) and the cytolysin operon, which were not found in the other isolates.Although Efm and Elc shared most of their VFs, seven of the nine Elc carried the nidogenbinding LPXTG surface adhesin gene sgrA, which was not identified neither in Efm nor Efs.

Detection of genes and chromosomal point mutations associated with antimicrobial resistance
In all, 25 different GDRs (23 ARGs and two chromosomal point mutations) that encode resistance to antimicrobials belonging to 10 different classes were identified, resulting in 18 different genotypic resistance profiles (Fig. 3).Efs genomes carried the largest number of GDRs (n = 23), 10 of them also present in Efm and Elc.Genes intrinsically present in the chromosome of all isolates were lsa(A) (resistance to lincosamides and streptogramin A) in Efs, and aac(6′)-Ii (resistance to tobramycin, kanamycin, and amikacin) and msrC (low-level resistance to ERY and quinupristin) in both Efm and Elc.Only one Elc isolate (ENT0020) carried acquired ARGs (nine ARGs in a plasmid), and all of them were also detected in Efm.The two SNP mutations were only detected in two Efs and conferred resistance to fluoroquinolones: one in the gyrA gene (S83Y) and another in the parC gene (S80I).In addition, PointFinder identified several mutations in the Penicillin-Binding Protein 5 (PBP5) sequences of six Efm and one Elc (ENT0020) corresponding to intermediate PBP5-S/R hybrid types (Table S4), but none presented the consensus PBP5-R allele.
ARGs encoding aminoglycoside resistance were the most abundant and diverse (Fig. 3).The only isolate resistant to GEN, an MDR Efs isolated from lactating cows in F3 (ENT0072), carried the aac(6′)-aph(2″) gene in the chromosome.Two genes encoding streptomycin (STR) resistance were detected, that is, str in Efs (n = 3 isolates) and ant( 6)-Ia in the three species (n = 16).The aph(3′)-III gene was also found in the three species (n = 13) as part of the ant( 6)-Ia-sat4-aph(3′)-III aminoglycoside-streptothricin resistance gene cluster.The spectinomycin-resistant gene ant( 9)-Ia was present in two Efs isolates (in plasmids in both and with a second copy in the chromosome in one) and it was always present in the same plasmid or chromosomal region as the erm(A) gene (ERY).All isolates with MIC ERY ≥128 mg/L, that is, two Efm, one Elc, and 10 Efs, harbored the erm(B) gene, generally located in a plasmid.However, four other ERY-resistant Elc (MIC ERY = 8 mg/L, three isolates and MIC ERY = 16 mg/L, one isolate) did not carry the erm(A) or erm(B) genes.
Determinants of CHL resistance were mostly restricted to Efs and included catA8 (cat pC223 ) (n = 4) and catA7 (cat pC221 ) (n = 9); catA8 (cat pC223 ) was present in plasmids in two (duplicate) isolates and the chromosome in another two unrelated isolates, whereas catA7 (cat pC221 ) was always located in plasmids along with several other ARGs.The only Efm isolate (ENT0092) carrying a CHL resistance gene, catA7 (cat pC221 ), did not show phenotypic resistance to CHL (MIC CHL = 32 mg/L, one dilution step below the ECOFF).Four Efs isolates were resistant to both CHL and LZD and besides catA7 (cat pC221 ), they all carried the fexA 688 nt upstream of the optrA gene in other plasmids.
The lsa(E) gene, associated with resistance to lincosamides and streptogramin A antibiotics, was detected in the chromosome of one Efs and in plasmids of five Efm and one Elc.Genes that encode the nucleotidyl transferase and confer resistance to lincosamides were also detected, that is, lnu(B) in Efm (n = 5) and Elc (n = 1), and lnu(G) in Efs (n = 2).Trimethoprim resistance-associated genes were only detected in Efs, dfrD (n = 3) and dfrG (n = 3).Finally, all TET-resistant isolates harbored the tet(M) gene (n = 22), 13 of them along with the tet(L) gene.The tet(M) gene was generally chromosomally encoded in Efs and plasmid-encoded in Efm and Elc, whereas tet(L) was mostly located in plasmids in all species (Fig. 3).

Within-farm transmission dynamics of resistant enterococci
In some instances, duplicate isolates (same ST, GDRs, and VFs) were found during several samplings infecting animals of different age groups within a farm (Fig. 3).This was the case of two ST19 Efs isolated from calves and heifers during two consecutive samplings in F5 and two Efs isolates of ST100 recovered from calves and lactating cows in F1.In Efm, isolates of ST32 (recovered at samplings S03 and S05 from calves in F1) shared their phenotypic and genotypic AMR profiles and carried identical VFs.On the other hand, ST21, the most abundant MLST type in Efs, was identified in F2 (one isolate) and F4 (four isolates) and all presented a different content of ARGs.Similarly, the two Efs isolates belonging to ST570, also from F1 and F4, were different in their gene content and antimicrobial susceptibility.In Efm, the two ST771 isolates (F2 and F5) carried a different repertoire of plasmid-encoded ARGs.In Elc, isolates with the same genomic profile were isolated from different farms.Thus, ST361, isolated from lactating cows in F1 and F4, was identical in all the examined features, and ST296, found in all age groups and two farms (F4, n = 3 and F1, n = 2), shared their GDR content (aac(6')-Ii and msrC genes) even though four isolates showed MICs for ERY above the ECOFF (MIC ERY = 8-16 mg/L but lacking an ERY-associated GDR).

DISCUSSION
This study was aimed at describing the within-farm dynamics and AMR profiles of Efm and Efs isolated from animals of different age groups (calves, heifers, and lactating cows) in dairy cattle farms.However, WGS and pangenome analysis of a selection of isolates showed that several isolates initially identified as Efm were in fact Elc, an Enterococcus species first isolated from Italian raw milk cheese (3).Like several recent reports (12,13), our results confirm the proposed reclassification of Efm clade B as Elc (6).Consequently, the former division of Efm between commensal-and healthcare-asso ciated lineages would now be the result of two different species.Upon analysis of the genome sequences, we verified the presence of the ddlA gene in the genomes of both Efm and Elc, confirming the amplification of both species with RTi-PCR1.Considering that most identification techniques available did not differentiate Efm clade A from Elc, we designed a new real-time PCR assay for the specific identification of Efm (only the formerly known clade A).Protocols for the specific PCR identification of Elc have already been described (14,15) but here we developed a TaqMan real-time PCR assay that uses a specific hydrolysis probe for the detection of Efm.The primers and probe designed here can be used in a duplex format with the mltF-targeting primers and probe (16) to simultaneously detect Efs and Efm (data not shown).When used to re-analyze all the isolates initially identified as Efm with RTi-PCR1, the study herein showed a widespread occurrence of Elc in dairy cattle.
Resistance to antibiotics considered critically or highly important to human medicine has been described to be low in enterococci recovered from cattle (10,17).However, resistant enterococci can easily exchange GDRs with other enterococci and Gram-pos itive bacterial species and contribute to the spread of resistance (18).In this study, resistance was more frequent in Efs than in Efm and Elc but all isolates were suscep tible to ampicillin, vancomycin, teicoplanin, tigecycline, and daptomycin.Enterococci produce chromosomally encoded lowaffinity penicillin-binding proteins (PBPs) that bind weakly to β-lactam antibiotics resulting in reduced susceptibility to penicillin.The most important mechanism for high-level ampicillin resistance in Efm is specific amino acid changes in the PBP5 sequence (19).Here, even though six Efm and one Elc showed intermediate PBP5-S/R hybrid types, none contained the PBP5 profile associated with ampicillin resistance.This agrees with previous studies that showed that strains associated with animals (primarily of subclade A2) displaying ampicillin MICs in the range 0.5-128 mg/L, mainly harbored hybrid-like PBP5 (PBP5-S/R) alleles (20).In fact, beside the allelic type, the ampicillin resistance phenotype is related to the PBP5 expression levels which are likely associated with differences in the upstream region of pbp5 (21).Acquired (enhanced) resistance to ampicillin is rare in Efs but common among clinical Efm isolates.The prevalence of AMP resistance in Efm of animal origin varies depending on the region and host, and in cattle is lower than in poultry or swine (10,22).Similarly, vancomycin resistance in enterococci isolated from food-producing animals is more common in Efm than Efs, and isolation of vancomycin-resistant Efm is more frequent in poultry and pigs compared to cattle (22).Here, vancomycin-resistant isolates were not recovered even when selective isolation in vancomycin-containing media was performed.
LZD, the first oxazolidinone antibiotic, is a critically important antimicrobial used to treat vancomycin-resistant enterococci.Resistance is associated with ribosomal mutations or the acquisition of the cfr, poxtA or optrA genes (23).Here, Efs of two different MLST types (ST19 in F5 and ST21 in F4) that shared the same MDR phenotypic profile (CHL-ERY-LZD-TET) showed LZD resistance.They carried the oxazolidinone-pheni col transferable resistance gene optrA that codes for an ABC transporter and confers a decreased susceptibility to oxazolidinones and phenicols in plasmids of different sizes but always co-localized with the fexA gene.Co-carriage of additional ARGs in the same plasmid is of particular relevance for the dissemination of optrA-carrying plasmids even when LZD is not used (not approved for use in food-producing animals) (23).The use of florfenicol in farm animals may select for florfenicolresistant bacteria that also carry the optrA gene (24).Although LZD resistance is still rare in humans, Efs harboring transferable optrA and poxtA genes is spreading in Spain (25).Similarly, the increasing levels of resistance in enterococci recovered from food-producing animals and food of animal origin reported in different countries, mainly associated with optrA gene carriage (22,23,26), is a concern.Resistance to CHL was widespread in Efs, and all resistant isolates sequenced carried variants of the cat gene that code for chloramphenicol acetyltransferase enzymes.As previously reported (22), the catA7 (cat pC221 ) gene variant was more prevalent, whereas the catA8 (cat pC223 ) gene was more sporadically detected.Interestingly, the four LZD-resistant isolates that carried the optrA and fexA genes also carried catA7 (cat pC221 ) but in a different plasmid.
Resistance to ERY (MIC ERY >4 mg/L) was moderate in Efm and similarly high in Efs and Elc.However, MIC values among ERY-resistant isolates were higher in Efs and Efm compared to Elc, and only isolates with MIC ERY >128 mg/L carried the erm(B) gene.Conversely, most ERY-resistant Elc isolates showed MIC ERY values just above the ECOFF and those sequenced lacked any ERY-associated GDR.The erm(B) gene, that confers resistance to macrolide-lincosamide-streptogramin B antibiotics, is the most common genetic determinant of ERY resistance and is widespread in human and animal isolates (22).Q/D, a mixture of streptogramin A (dalfopristin) and B (quinupristin), is effective against Efm but not against Efs (27) that intrinsically carry the lsa(A) gene (28).In Efm, resistance against streptogramin A-type antibiotics can be mediated by the PLS A resistance gene lsa(E) (29) or via enzymatic inactivation by the acetyltransferases VatD and VatE (30).Resistance to streptogramins B is either encoded by the erm(B) gene or the vgbA gene, which encodes a staphylococcal-type lactonase (31).Here, vat and vgb genes were not detected but the presence of the erm(B) and lsa(E) genes (1 Efm and 1 Elc) was associated with higher MIC values for Q/D (MIC Q/D = 8 mg/L).The presence in isolates from cattle of plasmids carrying both Isa(E) and erm(B) genes is a concern since Q/D is a valuable alternative to vancomycin for the treatment of multi-drug-resistant Efm infections.
Enterococci have intrinsic low-level resistance to aminoglycosides due to their low cell wall permeability or inactivation by naturally occurring enterococcal enzymes (32).For example, all Efm and Elc isolates tested in this study carried a chromosomally encoded 6′-acetyltransferase enzyme (AAC(6′)-Ii) capable of modifying aminoglycosides like tobramycin, kanamycin, and amikacin (33,34).Therefore, only two aminoglycosides (GEN and STR) are not readily affected by intrinsic enzymes produced by enterococci and are reliably used in clinical practice (for synergism with β-lactams) (2).In the present study, a high level of resistance to GEN (MIC GEN >1,024) was observed in three Efs isolates from the same farm (F3, recovered from calves, heifers, and lactating cows).Of them, the only isolate subjected to WGS carried the bifunctional gene aac(6′)-aph(2″), the most important aminoglycoside-modifying enzyme encoding gene since it confers resistance to most of the clinically available aminoglycosides (e.g., GEN, tobramycin, amikacin, kanamycin) but not STR (32,35).On the other hand, the ant( 6)-Ia gene (mostly within the aminoglycoside-streptothricin cluster ant( 6)-Ia-sat4-aph(3′)-III) was widespread among the sequenced isolates indicating a high rate of STR-resistance.Other acquired aminoglycoside resistance genes detected in our panel of isolates included ant(9)-Ia, aph(3′)-III, str, sat4, and spw.
TET resistance was low in Elc but high in Efm and very high in Efs.The higher prevalence of TET resistance in Efs was associated with the widespread presence of tet(M), mainly located in the chromosome.Conversely, in Efm and Elc, tet(M) was mostly encoded in plasmids along with tet(L).All MDR isolates were resistant to TET, an antimicrobial widely used in livestock that might co-select enterococci resistant to other antimicrobial agents.Genes associated with trimethoprim resistance (dfrD and dfrG) were occasionally present in Efs; however, since enterococci can utilize exogenous sources of folate inhibitors trimethoprim is ineffective in vivo (36).Although fluoroquinolones are commonly used in dairy cattle, here, resistance to CIP was low and only found in two Efs isolates, in both cases associated with mutations in the gyrA and parC genes, the most common mechanism of resistance (37).
Horizontal gene transfer and the remarkable genome plasticity of enterococci play a key role in their evolution and survival, and the acquisition and dissemination of GDRs.The most frequently reported mechanism for foreign DNA acquisition in enterococci is conjugation via conjugative plasmids and conjugative transposons (or integrative conjugative elements) (2).Here, most isolates carried the majority of acquired resistant genes in plasmids, with up to 10 ARGs in a single plasmid.Efs isolate ENT0072 was the exception since it carried on the chromosome 13 different ARGs and SNP mutations in the gyrA and parC genes.
Isolates with identical MLST type, ARG content, and plasmids were occasionally obtained from consecutive samplings within the same farm suggesting that the same strain persisted in the farm over time infecting animals of different age groups.This was the case with the LZD-resistant isolates, which were detected in different samplings and age groups in two farms.Exceptionally, within the same farm isolates of the same ST carried different ARG, mainly in plasmids but also in the chromosome (e.g., Efs ST21 in F4 isolated over 10 months in calves and lactating cows), indicating the frequent occurrence of gene exchange.By contrast, isolates from different farms were in general distinct and even when genetically related (same MLST type), they generally carried different ARGs in different plasmids.Also noteworthy was the presence of multiple copies of certain ARGs in several isolates, either in the chromosome, in a plasmid, or both.Notably, one Efm isolate carried two copies of tet(M) in the chromosome and another in a plasmid.Conjugative elements such as the Tn916/1545 and Tn5385 transposon families, commonly associated with tet(M) (38), could have mediated a recombination event.This redundancy of determinants of resistance has been described in settings under high selective pressure.
Although Efs and Efm are opportunistic pathogens, Elc has probiotic characteristics.The pathogenic potential of Enterococcus spp. is based on their adhesion capacity, their ability to evade the immune system, and their potential to form biofilms.VFs were more widespread in Efs and similarly distributed among Efm and Elc isolates.The isolate with the largest number of VFs (n = 46) was an Efs.Interestingly, it was the only one that carried the cytolysin operon (Cyl), that encodes a secreted two-peptide lytic protein that damages host cells and promotes infection (39), and the aggregation substance AS, a pheromone-induced surface protein that facilitates plasmid conjugation and potentiates the pathogenic effect of Cyl (40).Regarding Elc, they all had a MIC AMP ≤2 mg/L and lacked the virulence genes IS16, hyl, and esp, thus meeting the criteria set by the European Food Safety Authority (EFSA) to consider a strain safe (41).However, one Elc (ENT0020) carried eight ARGs in a plasmid, which poses a risk of transmitting resistance to other bacteria.Besides, seven Elc isolates carried the sgrA gene that codes for a nidogen-binding LPXTG surface adhesin.SgrA binds to extracellular matrix proteins and is involved in biofilm formation (42).The two Elc that did not carry sgrA (ENT0020 and ENT0245) were the most divergent isolates within the Elc cluster in the pangenome analysis.
In conclusion, the identification of enterococci at the species level has clinical relevance due to differences in pathogenicity and antibiotic resistance profiles.Traditional methods did not distinguish Elc from Efm.Even commonly used taxonomic classifiers like Kraken2 failed to correctly assign the corresponding species due to the misidentification of the reference strains included in the database.Therefore, the interpretation of data on Efm epidemiology and characterization from previous studies biased by Elc misidentification should be reconsidered.The Efm speciesspecific real-time PCR assay developed here will help to properly identify Efm (only the formerly known clade A) in future studies.Here, we showed that Elc is prevalent in cattle.Whole-genome characterization showed that GDRs and VFs were more widespread in Efs than in Efm, and GDRs were more abundant in both species compared to Elc.Although Elc most probably poses a lower human health hazard than Efm or Efs, some isolates can carry MDR plasmids similar to those harbored by Efm and could act as a donor of ARGs for other pathogenic enterococcal species.Although all isolates were susceptible to critically or highly important antibiotics like daptomycin, teicoplanin, tigecycline, and vancomycin, as well as ampicillin, resistance to LZD (Efs) and Q/D (Efm and Elc) was detected in a few isolates.Co-carriage of several ARGs in MDR-plasmids is a concern since antimicrobials commonly used in livestock could co-select and confer resistance to critically important antimicrobials not used in food-producing animals.Further studies are in progress to deeply characterize the mobilome and ARG synteny of the chromo somes sequenced in this study.

Sampling design
Five dairy cattle farms in the Basque Country were selected to study the within-farm diversity and dynamics of Enterococcus spp.and their resistance profiles.The Basque Country is a 7,234 km 2 region located in northern Spain, where 18,126 dairy cattle are currently managed under an intensive production system according to the 2020 census (https://en.eustat.eus).The commercial farms enrolled in the study (designated F1-F5) were representative of the style of farming in the region.Samplings, initially planned on a monthly basis for 1 year, extended between 2019 and 2020 due to the COVID-19 pandemic and consisted of 7-10 samplings in four of the farms (F1-F4) and three in another (F5).At each sampling time, rectal fecal samples were collected from five apparently healthy animals from each of three age groups (1-to 5-month-old calves, 5-to 22-month-old heifers, and lactating cows) and analyzed in a single 25 g pool per age group (5 g per animal).At five sampling visits in F2 and two in F4, heifers were not available for sampling.Thus, a total of 535 rectal fecal samples were collected and analyzed in 107 pools.Samplings were carried out by veterinary practitioners in strict accordance with Spanish ethical guidelines and animal welfare regulations (RD 53/2013) as part of their routine veterinary practice, and therefore, ethical review and approval of the Ethics Committee for Animal Experimentation was not required.Informed oral consent was obtained from the farmers at the time of sampling.

Enterococcus selective isolation
Feces (a total of 25 g) were diluted 1:10 in Brain Heart Infusion (BHI, Difco Laboratories, Detroit, MI, USA) broth supplemented with sterile NaCl 6.5% and incubated at 37 ± 1°C for 18 to 24 h.Subsequently, 20 µL of the incubated broth was subcultured onto m-Enterococcus Agar (Difco Laboratories) for the selective isolation of Enterococcus spp.Simultaneously, for the selective isolation of vancomycin-resistant enterococci (VRE), 10 µL of the initial broth was subcultured on BHI broth supplemented with 2 mg/L vancomycin (Sigma-Aldrich, Saint Louis, MO, USA).After incubation (37 ± 1°C for 18 to 24 h), 20 µL was subcultured onto m-Enterococcus plates with vancomycin (6 mg/mL) and incubated at 37 ± 1°C for 18 to 24 h.Ten colonies per plate were selected based on colony morphology and subcultured onto Blood Agar (Columbia agar + 5% sheep blood, bioMérieux, Marcy-l'Etoile, France) for further identification of Efs and Efm by the simultaneous real-time PCR amplification of the mltF gene of Efs and the ddlA gene present in Efm (RTi-PCR1), as described elsewhere (16).

Design of a real-time PCR assay for the specific identification of E. faecium
A real-time PCR (RTi-PCR2) amplification assay was developed for the amplification of the gluP (rhomboid protease) gene for the specific detection of Efm.The gluP gene was selected based on the pangenome analyses performed by Belloso Daza et al. (14); this gene is present in Efm and Elc but only shares a 90%-92% homology.Aligned sequences of the gluP gene of 16 Efm and 34 Elc were compared for primer and probe design, and an in silico specificity analysis was then performed for each primer and probe by submitting their nucleotide sequences against the GenBank databases using BlastN.Thus, a pair of primers (Efm-gluP-F: 5′-ACATAACCCAGCGATCCAG-3′; Efm-gluP-R: 5′-CC AATTAGCCCACCGACAT-3′) and a probe (FAM─5′-CCCAAACAT/ZEN/CTACAGAGGTATCCA GAAGAC-3′─Iowa Black FQ; Integrated DNA Technologies, Coralville, Iowa, USA) were designed to specifically amplify a 120 bp fragment of the gluP gene of Efm but not Elc.A QuantStudio 5 real-time PCR system (ThermoFisher Scientific, Waltham, MA, USA) was used to perform the RTi-PCR using 96-well (0.2 mL) consumables.Reactions were performed in a volume of 15 µL and included 1× Premix Ex Taq (TaKaRa Bio, Mountain View, CA, USA), 0.5× ROX Reference Dye II (TaKaRa Bio), 0.3 µM of each primer, and 0.2 µM of probe.Cycling consisted of a denaturation step of 30 s at 95°C, followed by 35 amplification cycles of 95°C for 15 s and 60°C for 1 min.Negative (sterile water) and positive controls were included in each real-time PCR amplification assay.Threshold values (Ct) below 32 cycles were considered positive and those greater than 32 cycles were considered uncertain.

Antimicrobial susceptibility test determination by broth microdilution
MIC values were determined for 12 antimicrobial agents belonging to 10 classes using Sensititre EUVENC antimicrobial susceptibility test (AST) Plates (ThermoFisher Scientific), following the recommendations in Commission Implementing Decision 2020/1729/EU regarding antimicrobials and dilution ranges.The results were inter preted using epidemiological cutoff (ECOFF) values defined by EUCAST (European Committee on Antimicrobial Susceptibility Testing; http://www.eucast.org) to discrim inate microorganisms with and without acquired resistance mechanisms (non-wildtype resistant and wild-type susceptible, respectively).Efs are intrinsically resistant to pleuromutilins, lincosamides, and streptogramins A (the so-called PLS A phenotype) due to the expression of the lsa(A) gene ( 28) and, therefore, MIC results for Q/D were not interpreted.For Efm and Elc, a MIC of >1 mg/L was considered as a reference following EFSA recommendations (43) but only those with MIC Q/D >4 mg/L were interpreted as resistant.

Whole-genome sequencing, genome assembly, and bioinformatics analysis
Genomic DNA was extracted with NZY Microbial gDNA Isolation kit (NZYTech, Lisbon, Portugal) and it was quantified and quality-assessed using a NanoDrop 1000 spectropho tometer (ThermoFisher Scientific) and a Qubit 2.0 Fluorometer (ThermoFisher Scientific).WGS was carried out using Oxford Nanopore Technology (ONT, Oxford, UK) for long-read sequencing.Library preparation was performed following the genomic DNA ligation protocol (SQK-LSK109), and sequencing was carried out on a MinION MK1C device (ONT) using FLO-MIN106D (R9.4.1) and FLO-MIN111 (R10.3)flow cells (ONT).The output files generated by ONT sequencing were base-called in high accuracy mode and quality filtered using Guppy v.5.0.13 (Qscore >8).Sequencing adapters were removed using Porechop v.0.2.4 with default parameters (44) and Filtlong v.0.2.0 (https://github.com/rrwick/Filtlong) was employed to filter sequences based on their size and quality.Reads with a size smaller than 1,000 base pairs (bp) were discarded, and a subset comprising the top 90% of reads, as determined by their quality score (Q), was selected.In sam ples where the sequence data were larger than 1,000 Mbp, lower quality reads were progressively removed until obtaining 1,000 Mbp of sequences.
Virulence genes were filtered at 85% coverage and 60% identity, and the pattern of presence/absence of these genes was used to generate a dendrogram.The hierarchical clustering analysis for the dendrogram was performed with the unweighted pair-group method with arithmetic mean (UPGMA) based on the Jaccard distance matrix, using the function hclust (v.3.6.1) of the R statistical package v.3.6.3.The presence of IS16 was investigated by screening the genomes against a custom database containing the target sequence downloaded from ISFinder (51) using ABRIcate.
Isolates from the same source (farm) that exhibited the same MLST type, VFs, and GDR patterns were considered as duplicate isolates; isolates from different farms sharing these features were considered the same strain.When distinguishable by any of these features they were referred to as different strains.

Pangenome analysis
Chromosomic pangenome analysis was conducted to investigate the genetic diversity and evolutionary relationships among the Enterococcus isolates in the European Galaxy server (https://usegalaxy.eu/).Genomes annotation was performed using Prokka v.1.14.6 (52), and the pangenome was calculated using Roary v.3.13.0 (53), employing a minimum blastp identity threshold of 95% for gene family definition.Genes from the Roary output were categorized as core genes (present in 99%-100% of the isolates), soft-core genes (95%-99%), shell genes (15%-95%), and cloud genes (0%-15%).In addition, a matrix was generated to illustrate the gene presence and absence patterns across the different species, along with the dendrogram that illustrates the relationship among isolates based on the accessory genes.
For the classification and identification at the species level of our isolates, genomebased taxonomy was performed at TYGS (Type Genome Server), a server that infers genome-scale phylogenies and state-of-the-art estimates for species and subspecies boundaries from userdefined and automatically determined closest type genome sequences through pairwise genome comparisons (54).To infer the phylogenetic relationships among the Enterococcus isolates, parsnp v1.7.4 (55) along with the implemented RaxML v.8.2.12 (56) was used to construct phylogenetic trees based on core genome SNPs, with default parameters and specifying -r! parameter to randomly select the reference from the set of genomes analyzed.The resulting trees were visualized and edited with iTOL v.6.8 (https://itol.embl.de).

Statistical analysis
Multivariate logistic regressions were performed to statistically examine the differences among various variables in relation to age groups and/or Enterococcus species.These variables encompassed (i) the overall distribution of enterococci, (ii) the distribution of each specific Enterococcus species, (iii) the occurrence of phenotypical AMR to each of the studied antimicrobial agents, (iv) pan-susceptibility, and (v) MDR, defined as resistance against three or more antimicrobial classes.Odds ratios (OR), along with their corresponding 95% confidence intervals (95% CI), were employed as a measure of association between positivity and the explanatory variables (age groups and Enterococ cus species), adjusted for the variable of farms.Differences were considered statistically significant if P < 0.05. the recipient of a fellowship within the 2020 call of grants for the training of scientific and technologist personnel in the scientifictechnological and business environment of the Basque agrifisheries and food sector funded by the Department of Economic Development, Sustainability, and Environment of the Basque Government.The funders had no role in the study design, data collection, and interpretation, or the decision to submit the work for publication.

FIG 1
FIG 1 Pangenome analyses of Enterococcus genomes.(A) Distribution of the 15,766 genes that make up the pangenome across the 32 isolates.The dendrogram represents the hierarchical clustering of the genomes based on the distribution of their accessory genes (presence/absence).The cluster framed in red corresponds to isolates identified as E. faecalis and in green as E. faecium by RT-PCR1.(B) Phylogenetic tree based on the core-genome single nucleotide polymorphisms (SNPs) of E. faecium and E. lactis (size = 110,229 bases).(C) Phylogenetic tree based on the core-genome SNPs of E. faecalis (size = 50,251 bases).The phylogenetic trees (B and C) were constructed using Parsnp and RaxML, and corresponding metadata such as age group, farm, sampling time, and MLST results (ST, Sequence Type; CC, Clonal Complex) are indicated for each isolate.

FIG 3
FIG 3 Heatmap of the distribution of the AMR genes detected by WGS.The presence and location of the GDRs are indicated as per the legend.
A.H. conceived and coordinated the study.M.M. and B.O. performed the laboratory analyses.J.L.L. and M.O.carried out the bioinformatic data analyses.M.M. and M.O.performed the statistical analyses.A.H., M.M., and M.O.interpreted the data and wrote the manuscript.B.O. and J.L.L. contributed to the manuscript revision.All authors read and approved the final version of the manuscript.