Enterococcus species: insights into antimicrobial resistance and whole-genome features of isolates recovered from livestock and raw meat in Ghana

Introduction Enterococcus spp. have gradually evolved from commensals to causing life-threatening hospital-acquired infections globally due to their inherent antimicrobial resistance ability and virulence potential. Enterococcus spp. recovered from livestock and raw meat samples were characterized using antimicrobial susceptibility testing and whole-genome sequencing. Materials and methods Isolates were confirmed using the MALDI-ToF mass spectrometer, and antimicrobial susceptibility was determined using the Kirby-Bauer disk diffusion method. Whole genome sequencing was performed on isolates resistant to two or more antibiotics. Bioinformatics analysis was performed to determine sequence types, resistance and virulence gene content and evolutionary relationships between isolates from meat and livestock samples, and other enterococci genomes curated by PATRIC. eBURST analysis was used to assign genomes to clonal complexes. Results Enterococcus spp. were predominantly E. faecalis (96/236; 41%) and E. faecium (89/236; 38%). Overall, isolates showed resistance to erythromycin (78/236; 33%), tetracycline (71/236; 30%), ciprofloxacin (20/236; 8%), chloramphenicol (12/236; 5%), linezolid (7/236; 3%), ampicillin (4/236; 2%) and vancomycin (1/236, 0.4%). Resistance to two or more antimicrobial agents was detected among 17% (n = 40) Enterococcus spp. Resistance genes for streptogramins [lsa(A), lsa(E), msr(C)], aminoglycosides [aac(6′)-Ii, aph(3′)-III, ant(6)-Ia, aac(6′)-aph(2″), str], amphenicol [cat], macrolides [erm(B), erm(T), msr(C)], tetracyclines [tet(M), tet(L), tet(S)] and lincosamides [lsa(A), lsa(E), lnu(B)] were detected among the isolates. Genes for biofilm formation, adhesins, sex pheromones, cytolysins, hyaluronidase, oxidative stress resistance, quorum-sensing and anti-phagocytic activity were also identified. Potential plasmids with replicon sequences (rep1, rep2, repUS43, repUS47, rep9a, rep9b) and other mobile genetic elements (Tn917, cn_5536_ISEnfa1, Tn6009, ISEnfa1, ISEfa10) were detected. Clinically relevant E. faecium ST32 and ST416 clones were identified in meat samples. Conclusion The occurrence of antimicrobial-resistant Enterococcus spp. in livestock and raw meat samples, carrying multiple resistance and virulence genes, including known clones associated with hospital-acquired infections, underscores the critical need for employing robust tools like whole genome sequencing. Such tools provide detailed data essential for ongoing surveillance efforts aimed at addressing the challenge of antimicrobial resistance with a focus on one health.


Introduction
Enterococcus spp.exist as commensals in the gut of warm-blooded animals and humans but possess virulence genes, and are recognized as opportunistic pathogens that can cause a variety of hospitalacquired infections, such as urinary tract and intra-abdominal infections, bacteremia, and endocarditis (Iweriebor et al., 2015;Ngbede et al., 2017).E. faecium and E. faecalis are commonly associated with opportunistic infections, with other species being rarely pathogenic (Beukers et al., 2017).Enterococci have notable resistance to adverse environmental conditions, hence the capacity to colonize different ecological niches and spread within the food chain through contaminated animals and foods (Klibi et al., 2013).The inherent high antimicrobial resistance (AMR) ability of Enterococcus spp., and their ability to genetically acquire and transmit antimicrobial drug resistant determinants among themselves and other bacteria in the environment, presents a significant challenge for therapeutic measures (Iweriebor et al., 2015).
One common phenomenon that increases AMR among bacterial species including Enterococci is the use of antibiotics in animal production, which subsequently creates ideal conditions for evolution and selection of resistant strains (Iweriebor et al., 2015;Ngbede et al., 2017).Several studies have documented the widespread use of antibiotics in animal husbandry in Ghana, with majority of farmers using antibiotics for the purposes of treatment or prophylaxis, and at a lesser extent for growth stimulation (Donkor et al., 2012;Osei Sekyere, 2014;Boamah et al., 2016).Although Donkor et al. (2012) reports that there is no significant difference in the frequency of antimicrobial usage among different animals (cattle, goat, sheep, pig, poultry), Osei Sekyere (2014) indicates that the frequency of antimicrobial usage is influenced by the size or financial status of farms and mixed farming practices.In an effort to reduce costs and maintain profits, farmers in Ghana disregard the recommended waiting period after administering antibiotics to food animals by selling animal products like eggs and meat for human consumption (Boamah et al., 2016;Nkansa, 2020).This causes persistence of antibiotic residues in animal products, thereby exposing microorganisms in animals and humans to sub-inhibitory concentrations of antibiotic, leading to the emergence and spread of antibiotic-resistant bacterial strains through food or other environmental pathways (Iweriebor et al., 2015;Boamah et al., 2016;McEwen and Collignon, 2018;Collignon and McEwen, 2019).Unavailability of regulations regarding the use of antimicrobials in animals, limited involvement of veterinarians in the administration of antibiotics, and the unavailability of optimal storage conditions for antibiotics are among factors identified to be influencing the extensive and inappropriate use of antimicrobials in animal husbandry in the country (Donkor et al., 2012;Osei Sekyere, 2014;Boamah et al., 2016;Nkansa, 2020).
Enterococci are one of two indicator bacteria used to study the extent of AMR in populations owing to their proven resistance and virulence transmissibility as demonstrated in surveillance programs implemented in developed countries (Bager, 2000; European Food Safety Authority, European Centre for Disease Prevention and Control, 2013).Although the multidrug resistance potential of Enterococcus spp.has been demonstrated in other studies across Africa (Klibi et al., 2013;Iweriebor et al., 2015;Ngbede et al., 2017;Katakweba et al., 2018), data on Enterococci of human, animal and environmental origin in Ghana is scarce.Reports (Obiri-Danso et al., 2003, 2005;Ayum and Gifty, 2010;Labi et al., 2015;Quansah et al., 2018Quansah et al., , 2020;;Akrong et al., 2019;Akita et al., 2021) of isolated Enterococcus spp.from human clinical sources and environmental sources in Ghana have consistently lacked genomic characterization.Given that there are clear reports of inappropriate use of antibiotics on farms in Ghana, the limited application of genomic characterization obscures understanding of the genetic basis of ensuing antibiotic resistance, including the evolution and transmission of resistance genes, and the dynamics of resistance in bacterial populations.The application of whole genome sequencing (WGS) offers extensive genomic information and is valuable for characterizing bacterial isolates and conducting genomic surveillance of AMR in bacteria.This study employed WGS and antimicrobial susceptibility testing to examine Enterococcus spp.recovered from archived raw meat and livestock samples collected in Ghana.

Study site and isolates
Samples used in the study were obtained from three livestock farms and three slaughterhouses in southern Ghana independent of each other, from 2018 to 2019 and archived at the Bacteriology Department, Noguchi Memorial Institute for Medical Research.Livestock samples consisted of fecal swabs collected from cattle, goats, pigs, poultry and sheep.Meat samples were collected from the thigh, brisket and flank/mid-loin of carcasses at slaughterhouses using sterile swabs and stored in brain heart infusion broth.Carcasses sampled included beef, chevon and mutton.One milliliter of each archived broth sample was pre-enriched with 9 mL of trypticase soy broth and incubated overnight at 37°C.A loopful of pre-enriched samples was then cultured on Bile Esculin Azide agar (Merck, Germany) and incubated for 18-24 h at 37°C.Presumptive Enterococcus spp.were seen as small transparent colonies with black halos.Presumptive Enterococcus spp.were confirmed using MALDI-TOF mass spectrometer (Bruker, Billerica, MA, USA) following subculture on nutrient agar (Oxoid, Basingstoke, Hants, United Kingdom).

Whole-genome sequencing and analysis
Genomic DNA was extracted from Enterococcus spp.resistant to two or more antimicrobial agents using QIAamp ® DNA mini kit   (QIAGEN Inc.GmbH, Holden, Germany), with reference to the manufacturer's protocol.Libraries of DNA were prepared using the Illumina ® DNA Prep (M) Tagmentation kit (Illumina Inc.San Diego, CA 92122, United States) followed by multiplexing and 300 bp paired-end sequencing on the Illumina Miseq platform (Illumina, San Diego, CA, United States).Fastq files generated were assessed for quality using Fastqc (v0.11.9) and trimmed with Trimmomatic (v.0.39) at a minimum quality threshold of ≥ Q20 (Andrews, 2010;Bolger et al., 2014).Reads that met the quality threshold (Q20 or higher) after trimming, were subjected to de-novo assembly using Unicycler (v0.4.8).De-novo assembly was performed using the following parameters: minimum contig size of 200 bp, number of contigs <400, and minimum coverage of 20x.All assembled sequence data were submitted to GenBank and assigned accession numbers under Bio project PRJNA851427.
1 http://www.genomicepidemiology.org/ Using the CSI phylogeny tool, evolutionary relationships between E. faecalis and E. faecium isolates from this study and other African Enterococci genomes (curated by Bacterial and Viral Bioinformatics Resource Center), 2 were analyzed and inferred from single nucleotide polymorphisms (SNPs), by comparing assembled sequences to the reference genomes E. faecium SRR24 (NZ_CP038996.1)and E. faecalis V583 (226185.9).Close genetic relatedness was determined based on a pairwise SNP difference below 10 between the genomes (Weterings et al., 2015).The maximum-likelihood phylogenetic tree generated was annotated in Interactive Tree of Life (iTOL; Letunic and Bork, 2021).The accession numbers of downloaded genomes have been included in the Supplementary materials.

Statistical analysis
Statistical analysis was performed to assess the difference in Enterococcus spp.isolation from the respective samples and to also assess the correlation between resistance phenotypes and their respective genotype using the chi square test of independence and logistic regression.A significance level of α = 0.05 or lower was employed to determine statistical significance.Analyses were carried out using Stata version 16.

Genomic characteristics of Enterococcus species
A summary of the genome characteristics of Enterococcus spp.sequenced is presented in Table 3.All assembled genomes passed QC with Q scores >20, minimum coverage of 20, number of contigs <400 and minimum contig size of 200 bp.The assembled genomes ranged in size from 2.5-3.1 Mb, with average GC content of 37.6.

Phylogenetic analysis
The pairwise distance matrix of SNPs showed close genetic relatedness between majority of the E. faecalis isolates from livestock in this study (Figure 1).Clustering of E. faecalis isolates was mostly observed between isolates from the same farm (Figure 2).E. faecalis isolates recovered in this study also showed close genetic relatedness with the South African livestock genomes irrespective of the livestock type (Figures 1, 2).Two isolates from sheep (76F and 78 M) in this study were found clustering together with South African human clinical isolates, with a pairwise SNP difference below 10 (Figures 1,  2).Similar patterns in the distribution of plasmid replicons, AMR and virulence genes were seen among E. faecalis genomes occurring in the same clade (Figure 2).
E. faecium isolates from livestock sources exhibited greater genetic diversity compared to those from South Africa (Figure 3).E. faecium isolates from meat and livestock were mostly genetically divergent (Figure 4).However, one E. faecium isolate from meat (isolate 70) showed close genetic relatedness to an isolate from sheep (isolate 83) in the livestock group, with a SNP difference below 10 (Figures 3, 4).

Discussion
This study revealed a substantial diversity among the isolates collected from both meat and livestock samples.While E. faecalis and E. faecium were the most prevalent species detected among the samples, no statistically significant differences were noted in the distribution of Enterococcus spp.This lack of differentiation among the Enterococcus spp.recovered from the various sources, suggests that these isolates may not exhibit strong niche-specific adaptions based on the particular source or animal host, contrary to reports elsewhere (Zaheer et al., 2020).The variation in the distribution of Enterococcus spp.among livestock across different geographic regions can be attributed to differences in dietary patterns, which impact the composition of gut commensal bacteria (Klibi et al., 2013;Ngbede et al., 2017).Similar to our findings, other studies (Golob et al., 2019;Habib et al., 2022) have also reported consistent contamination of retail meat by Enterococcus spp.and this has been linked to fecal contamination.Furthermore, the presence of antimicrobial resistance genes in these Enterococci isolates, including those related to clinically important antimicrobials like aminoglycosides, streptogramin A, macrolides, and tetracycline, suggests a narrowing of available treatment options (beta-lactam monotherapy or beta-lactam in combination with an aminoglycoside or glycopeptide) for opportunistic Enterococcus infections including those caused by vancomycin resistant strains (Arias et al., 2010;O'Driscoll and Crank, 2015).Among the isolates resistant to two or more antimicrobial agents, resistance to erythromycin, tetracycline, ciprofloxacin and chloramphenicol was common.The incidence of erythromycin and tetracycline resistance in Enterococcus spp.has been reported in other studies in Africa, and has been associated with the extensive use and misuse of these antimicrobials in animal production (Iweriebor et al., 2015;Ngbede et al., 2017;Molechan et al., 2019).The increased levels of resistance recorded against tetracycline could be attributed to the use of tetracycline as recorded in the livestock farms sampled for this study.Nonetheless, erythromycin resistance might also be due to cross resistance associated with another macrolide, tylosin documented to be used among livestock in Ghana (Jackson et al., 2004;Akansale et al., 2019).Resistance to erythromycin among Enterococcus spp. was found to be mediated by erythromycin ribosomal methylation (erm) genes which encode for the modification of ribosomal target through methylation and msr(C) genes which mediate the extrusion of the antibiotic (Munita and Arias, 2016).Moreover, the detection of tet(L), tet(S) and tet(M) genes indicate that tetracycline resistance was due to ribosomal target protection and efflux-mediated mechanisms (Munita and Arias, 2016).Chloramphenicol resistance was mediated by the chloramphenicol acetyltransferase (cat) gene responsible for inactivation of chloramphenicol antibiotic, while ciprofloxacin   resistance was encoded by gyrA and parC mutations in the subunits of the DNA gyrase and topoisomerase IV enzymes (Munita and Arias, 2016).Furthermore, the presence of ClpL gene suggests exposure of these Enterococcus spp. to heat and other adverse environmental conditions.These environmental stress factors have been demonstrated to select for resistant bacteria and influence susceptibility to a range of antibiotics (Huang et al., 2021).The finding of low resistance to ampicillin, linezolid and vancomycin is in line with other reports (Mannu et al., 2003;Bortolaia et al., 2016).Although linezolid resistance is considered uncommon due to its limited use, findings show that the use of other antimicrobials such as macrolides may select for resistance against linezolid (Tyson et al., 2018).
All three livestock farms sampled in this study used antibiotics including tetracycline, penicillin and colistin, similar to other studies in the country where antibiotic use has been described as inappropriate and unregulated (Donkor et al., 2012;Osei Sekyere, 2014;Boamah et al., 2016;Nkansa, 2020).Enterococci notably acquire AMR genes through plasmids and transposons, chromosomal exchange or mutation (Iweriebor et al., 2015;Torres et al., 2018).Mobile genetic elements including integrative conjugative elements, plasmid replicons, insertion sequences and transposons were identified among the isolates confirming the potential of these isolates to acquire and transfer AMR determinants between likespecies and other pathogens.Co-occurrence between plasmid replicons (rep1,rep2,repUS43,repUS47,rep9a,rep9b) and AMR genes on the same contig was observed in 70% (28/40) of the isolates.Tetracycline resistance genes [tet(M),tet(L)] and macrolide resistance gene [erm(B)] more frequently co-existed with repUS43, rep9a and rep9b plasmid replicons.The association between tet(M) and tet(L) and repUS43 and rep9b was also reported in a similar study (Fatoba et al., 2022).Aminoglycoside resistance genes [aph(3′)-III, ant(6)-Ia] tended to be more frequently associated with rep1 plasmid, while another aminoglycoside resistance gene (str) often occurred on the same contig with rep7a plasmid replicon.Other mobile genetic elements commonly associated with AMR genes encoding resistance to tetracycline, macrolides and aminoglycosides included transposons (Tn917, cn_5536_ISEnfa1), integrative conjugative element (Tn6009) and insertion sequences (ISEnfa1, ISEfa10).Macrolide resistance gene [erm(B)] was frequently associated with Tn917 consistent with reports elsewhere (Torres et al., 2018).The abundance of plasmid replicon genes and other mobile genetic elements detected in this study confirms the plasticity of the genomes of Enterococcus spp.and supports the potential for horizontal dissemination of the AMR genes (Fatoba et al., 2022).The dissemination of these genes is intensified by clonal expansion which results when horizontally transferred AMR genes become chromosomally integrated (Waddington et al., 2022).MSLT analysis revealed diverse clones including novel sequence types for both E. faecium and E. faecalis isolates.Novel sequence types for both E. faecalis and E. faecium harbored genes for virulence and resistance with demonstrated phenotypic resistance.The most common African clone, E. faecalis ST16, (Jolley et al., 2018) was found among the isolates (n = 2) in this study.The other clones (ST4, ST16, ST32, ST300) detected, have also been reported in Nigeria and South Africa from similar sources (Jolley et al., 2018).Clonal relationship was observed in four E. faecium isolates [ST1442 (n = 2), ST2269 (n = 2)] from meat and livestock samples.E. faecium ST32 and ST416 recovered from meat clustered with sequence types belonging to the hospital-adapted CC17, known to cause hospital-acquired and clinical infections worldwide (Leavis et al., 2006).Majority of isolates belonging to CC17 are characterized by resistance to quinolones and ampicillin, and the presence of enterococcal surface protein (Leavis et al., 2006;Top et al., 2008).E. faecium isolates (ST32 and ST416) showed resistance to ciprofloxacin but were both susceptible to ampicillin phenotypically, although mutations in the pbp5 gene was detected for both isolates.The pathogenicity of bacterial species requires not only antibiotic resistance but also the possession of specific virulence factors.The adherence of bacterial cells to host tissues is a critical stage in the development of infection or forming biofilms (Stępień-Pyśniak et al., 2019).In this study, both acm and efaAfm virulence genes were detected in all E. faecium isolates.These genes have also been documented in clinically-derived E. faecium isolates in the United States and Malaysia (Nallapareddy et al., 2008;Soheili et al., 2014).Acm gene, known for encoding an adhesin responsible for E. faecium attachment to collagen, plays a significant role in the competitiveness of this species within clinical environments and is crucial for survival, colonization and infection (Nallapareddy et al., 2008).While the precise function of efaAfm is yet to be confirmed, it is hypothesized to be involved in cell wall adherence (Soheili et al., 2014;Stępień-Pyśniak et al., 2019).Consistent with other reports (Jahansepas et al., 2018;Farman et al., 2019;Trościańczyk et al., 2022), a broader spectrum of virulence genes was identified in E. faecalis than E. faecium, suggesting that E. faecalis is more virulent than E. faecium.Genes mediating adherence to collagen (ace), biotic and abiotic surfaces (efaAfs) and expression of pili on cell surface (ebpA, ebpB, ebpC, SrtA) were all identified in E. faecalis isolates, facilitating cell adhesion and biofilm formation.Genes encoding sex pheromones (cad,camE,cCF10,cOB1) identified in all E. faecalis isolates are noted for promoting biofilm formation and regulation by inducing conjugation between enterococcal cells and mediating the transfer of pheromone-responsive plasmids, which may contain virulence genes (Stępień-Pyśniak et al., 2019;Yoon et al., 2020;Trościańczyk et al., 2022).Furthermore, genes encoding cytolysin (cylB, cylL, cylM), gelatinase (gelE), and hyaluronidase (hylA, hylB), which are known for their ability to damage cells were detected in E. faecalis isolates (Golińska et al., 2013;Stępień-Pyśniak et al., 2019;Lang et al., 2020).Although studies suggest that hyaluronidase encoded by hyl is specific for E. faecium it has been reported in E. faecalis isolates in other studies (Golob et al., 2019;Kiruthiga et al., 2020).ElrA and tpx genes, which are associated with evading host's immune defenses were also detected, possibly indicating an evolved mechanism in these E. faecalis isolates for immune system evasion.Similar virulence factors have been isolated from E. faecalis isolates from food, animal, and clinical sources in Africa, Europe and Asia (Iweriebor et al., 2015(Iweriebor et al., , 2016;;Beshiru et al., 2017;Jahansepas et al., 2018;Farman et al., 2019;Stępień-Pyśniak et al., 2019;Yoon et al., 2020).The ST16, ST4 and ST480 E. faecalis identified in this study, Population structure of Enterococcus faecalis isolates.Phylogeny of 24 study isolates and 38 isolates from South Africa downloaded from BV-BRC.The genomes were compared using E. faecalis V583 strain (226185.9)as reference.SNP-based maximum likelihood tree was constructed using CSI Phylogeny and visualized in iTOL.This phylogenetic tree shows the evolutionary relationship between E. faecalis isolates in this study and other E. faecalis genomes from South Africa.Clustering between E. faecalis isolates from livestock was mostly observed among isolates in this study and other genomes from South Africa.For each isolate in this study, the source, farm (F)/slaughter house (SH), the sequence types, AMR, virulence genes and plasmids distribution are shown.have also been reported in cases of clinical infection in other geographical regions (Watanabe et al., 2009;Olsen et al., 2012;Poulsen et al., 2012;Farman et al., 2019) suggesting the widespread nature of these clones.Similar to findings in this study, E. faecalis ST480 identified in Saudi Arabia carried virulence genes encoding sex pheromones [cOB1], biofillm formation-associated pili genes [ebpA, ebpB, SrtA] and adhesins [ace, efaAfs], and were resistant to erythromycin, ciprofloxacin and tetracycline (Farman et al., 2019).

FIGURE 1
FIGURE 1 Heatmap and dendrogram showing pairwise SNP distance matrix and relationship among E. faecalis (n = 24) isolates in this study and other South African E. faecalis (n = 38) genomes downloaded from BV-BRC, labeled with the suffix "sa."Isolates in this study included: Cattle = 2; Goat = 5; Meat = 1; Pig = 1; Poultry = 5 and Sheep = 10.The diagram shows the genetic variation between the isolates in the study.The pairwise distance matrix of SNPs showed close genetic relatedness between majority of the E. faecalis isolates from livestock in this study, with SNP difference below 10.Complete matrix data with pairwise distances recorded has been included in Supplementary material.Heatmap and dendrogram were created with gplots (v.3.1.3)and dendextend (v.3.1.3)packages, respectively, in R.

FIGURE 3
FIGURE 3Heatmap and dendrogram showing pairwise SNP distance matrix and relationship among E. faecium (n = 15) isolates in this study and other E. faecium (n = 16) genomes downloaded from BV-BRC, labeled with the suffix "sa" and "et".Isolates in this study included: Cattle = 2; Goat = 2; Meat = 5; Pig = 3 and Sheep =3.The diagram shows the genetic variation between the isolates in the study.The pairwise distance matrix of SNPs showed significant genetic diversity between E. faecium isolates, with SNP difference greater than 10.Complete matrix data with pairwise distances recorded has been included in Supplementary material.Heatmap and dendrogram were created with gplots (v.3.1.3)and dendextend (v.3.1.3)packages, respectively, in R.
Enterococcus species

TABLE 1
Distribution of Enterococcus spp.recovered.
Other Enterococcus spp.isolated included: E

TABLE 2
Antimicrobial resistance profile of Enterococcus spp.recovered from livestock and raw meat samples.

TABLE 3
Genome characteristics of sequenced Enterococcus spp.
narrower spectrum of virulence genes was detected in E. faecium and E. lactis genomes; acm and efaAfm all encode adhesins.

TABLE 4
Characteristics of Enterococcus spp.resistant to ≥2 antimicrobial agents.

TABLE 5
Isolates with MGEs co-occurring with AMR genes on the same contig.