Targeting Enterococci with Antimicrobial Activity against Clostridium perfringens from Poultry

Necrotic enteritis (NE), caused by Clostridium perfringens, is an emerging issue in poultry farming. New approaches, other than antibiotics, are necessary to prevent NE development and the emergence of multidrug-resistant bacteria. Enterococci are commensal microorganisms that can produce enterocins, antimicrobial peptides with activities against pathogens, and could be excellent candidates for protective cultures. This study aimed to screen and characterize Enterococcus strains of poultry origin for their inhibitory activity against C. perfringens. In total, 251 Enterococcus strains of poultry origin plus five bacteriocin-producing (BP+) E. durans strains of other origins were screened for antimicrobial activity against the indicator C. perfringens X2967 strain using the “spot on the lawn” method. We detected thirty-two BP+ strains (eleven Enterococcus faecium, nine E. gallinarum, eight E. faecalis, three E. durans, and one E. casseliflavus). We further studied the antimicrobial activity of the supernatants of these 32 BP+ strains using agar well diffusion and microtitration against a collection of 20 C. perfringens strains. Twelve BP+ enterococci that were found to exhibit antimicrobial activity against C. perfringens were characterized using whole genome sequencing. Among these, E. faecium X2893 and X2906 were the most promising candidates for further studies as protective cultures for poultry farming. Both strains belong to the sequence type ST722, harbor the genes encoding for enterocin A and enterocin B, do not possess acquired resistance genes, do not carry plasmids, and present the acm gene, which is implicated in host colonization. Further research is needed to determine the utility of these strains as protective cultures.


Introduction
Antibiotic resistance is a serious public health concern that compromises the treatment of infections in humans and animals and is associated with the unnecessary prescription and/or misuse of antibiotics. Besides their clinical use in humans, antibiotics are also used in veterinary and animal farming. Antibiotics have also been extensively used as growth promoters in food-producing animals; however, even though this practice has been banned in Europe since 2006 [1] and also in several other countries, it is still allowed in some others [2]. This contributes to the increase and spread of antibiotic resistance, not only among pathogenic bacteria but also among commensal bacteria of the intestinal tract of humans and animals, which can lead to contamination via feces. Therefore, resistant bacteria can reach humans via the food chain and water or by contact with animals. For this reason, the World Health Organization (WHO) proposed to address this issue from a "One Health" perspective, establishing new alternatives to the use of antibiotics in livestock and agriculture [3].
Sixty enterococcus strains were isolated from poultry meat samples collected from local markets in La Rioja, Spain. These strains were identified using MALDI-TOF mass spectrometry as E. faecium (n = 33), E. faecalis (n = 19), E. gallinarum (n = 5), E. casseliflavus (n = 1), E. durans (n = 1), and E. avium (n = 1). These isolates were combined with another 191 Enterococcus, previously obtained from poultry (in Spain and Tunisia), and with five bacteriocin-producing (BP+) Enterococcus from other origins, to develop the entire collection of 256 Enterococcus used to detect and characterize the BP+ isolates.

Effects of the Supernatants of BP+ Enterococci on C. perfringens Isolates
The supernatants of the 32 BP+ enterococci were tested against a collection of 20 C. perfringens isolates of poultry origin. The antimicrobial activity was detected in 18 concentrated supernatants against at least one of the C. perfringens strains. Nevertheless, antimicrobial activity was observed in six of the heated supernatants (HS) and non-heated supernatants (NHS) (Figure 2, Table 2), corresponding to four E. faecium and two E. durans isolates. In general, the inhibitory activities of the HS and NHS were similar; both inhibited the growth of 2-8 strains of the 20 C. perfringens tested. The concentrated supernatants showed a broad spectrum of inhibition against 2-20 C. perfringens isolates ( Table 2). The remaining 14 supernatants, either HS, NHS, or concentrated supernatants, did not show any inhibitory activity.
Antibiotics 2022, 11, x FOR PEER REVIEW

Effects of the Supernatants of BP+ Enterococci on C. perfringens Isolates
The supernatants of the 32 BP+ enterococci were tested against a collection perfringens isolates of poultry origin. The antimicrobial activity was detected in 18 trated supernatants against at least one of the C. perfringens strains. Nevertheless crobial activity was observed in six of the heated supernatants (HS) and non-he pernatants (NHS) (Figure 2, Table 2), corresponding to four E. faecium and two isolates. In general, the inhibitory activities of the HS and NHS were similar; bo ited the growth of 2-8 strains of the 20 C. perfringens tested. The concentrated supe showed a broad spectrum of inhibition against 2-20 C. perfringens isolates (Tabl remaining 14 supernatants, either HS, NHS, or concentrated supernatants, did n any inhibitory activity.  Poultry Poultry E. faecalis --1   Supernatant activity could only be quantified for the E. faecalis X3198 and E. faecium X3179 strains (16 AU/mL).

Phenotypic and Genotypic Characterization of the Selected BP+ Enterococci
For a complete genome analysis, 12 BP+ enterococci were selected based on their antimicrobial activity detected using the previously described methods. Five E. faecium and two E. faecalis of poultry origin were selected, as well as five E. durans of milk and camel milk origin, chosen as the BP+ controls.

Bacteriocinome
Structural genes encoding for bacteriocins were detected in 12 BP+ strains ( Table 3). The structural genes for enterocins P and Enterocin L50 A/B were detected in all five E. durans isolates, and the genes for bac 32 were also observed in three of them. Genes encoding enterocin A and enterocin B were detected in all the E. faecium strains; two of these strains carried the genes encoding enterocin NKR-5-3-A/D/Z. Moreover, the genes encoding enterocin SE-K4 and staphylococcin C55a/b were identified in two E. faecalis strains. Five of the twelve selected BP+ enterococci (41.7%) were susceptible to the nine antibiotics tested, all of them from the species E. durans. The remaining strains were resistant to at least one of the antibiotics tested. The most frequent resistance was against ciprofloxacin (58.3%), followed by tetracycline (25.0%), erythromycin (25.0%), penicillin (16.7%), chloramphenicol (8.3%), high-level streptomycin (8.3%), and high-level gentamicin (8.3%). In addition, all the isolates showed susceptibility to vancomycin and linezolid.
Genes encoding antibiotic resistance were detected in all 12 BP+ strains (Table 4), although only five (three E. faecium and two E. faecalis isolates) had genes for acquiredtype resistance. The mutations associated with resistance phenotypes for beta-lactams (pbp5) and fluoroquinolones (gyrA and parC) were detected only in E. faecium isolates (Supplementary Material).  (6 ) • Virulome Among the 12 BP+ enterococcal strains, virulence genes were detected in E. faecium and E. faecalis but not in E. durans (Table 5).  The replicon plasmids identified in the selected enterococci are listed in Table 6. All of the E. durans strains carried RepA_N, Inc18, and Rep3 or Rep1 plasmidic replicons. Both E. faecalis strains carried the type Rep trans. Moreover, most of the faecium strains carried at least three different types of plasmidic replicons.

Genetic Lineages
Multi-locus sequence typing (MLST) of the two E. faecalis and five E. faecium strains yielded the following results: (a) the two E. faecalis strains were typed as ST397; (b) the five E. faecium strains showed four different sequence types, with two isolates typed as ST722, one isolate typed as ST784, and the remaining two with an unknown ST (Table 5, Figure 3). The reference strain, ATCC 29212, was also included.

Screening for BP+ Enterococci
A total of 32 of the 256 enterococci tested (12.84%) showed antimicrobial activity against the C. perfringens X2967 strain, as determined using the "spot on the lawn" method; however, among these, only 18 supernatants of the BP+ strains were active against the collection of 20 C. perfringens isolates used as indicators. The inhibitory activities of these supernatants were attributed to the Enterococcus-derived enterocins [14]. The absence of inhibitory activity in the supernatants obtained from the strains showing inhibition using the spot-on-the-lawn method may be explained by the fact that bacteriocins sometimes remain attached to the cell wall and are not released in the supernatant. Furthermore, the production of bacteriocins is commonly mediated by quorum sensing [15]; hence, we detected 14 strains as BP+ via the spot-on-the-lawn method (in which the producer and the indicator strains are confronted) but without activity in their supernatants (the extract produced without previous exposure to the indicator bacteria) [16].

Phenotypic and Genotypic Characteristics of the BP+ Enterococci
According to their antimicrobial activity, 12 BP+ enterococci were selected for further characterization.

Screening for BP+ Enterococci
A total of 32 of the 256 enterococci tested (12.84%) showed antimicrobial activity against the C. perfringens X2967 strain, as determined using the "spot on the lawn" method; however, among these, only 18 supernatants of the BP+ strains were active against the collection of 20 C. perfringens isolates used as indicators. The inhibitory activities of these supernatants were attributed to the Enterococcus-derived enterocins [14]. The absence of inhibitory activity in the supernatants obtained from the strains showing inhibition using the spot-on-the-lawn method may be explained by the fact that bacteriocins sometimes remain attached to the cell wall and are not released in the supernatant. Furthermore, the production of bacteriocins is commonly mediated by quorum sensing [15]; hence, we detected 14 strains as BP+ via the spot-on-the-lawn method (in which the producer and the indicator strains are confronted) but without activity in their supernatants (the extract produced without previous exposure to the indicator bacteria) [16].

Phenotypic and Genotypic Characteristics of the BP+ Enterococci
According to their antimicrobial activity, 12 BP+ enterococci were selected for further characterization.

Bacteriocinome
The structural genes for enterocins P and Enterocin L50A/B were detected in all five E. durans isolates. Enterocin P (entP) was first detected in an E. faecium strain isolated from a dry-fermented sausage [17], showing activity against gram-positive pathogenic bacteria such as C. perfringens, L. monocytogenes, and S. aureus. Enterocin P is chromosomally encoded [18,19]; however, other studies have detected entP genes in the plasmid location [20]. Enterocin P and L50A/B have been detected in different enterococcal species [21]. This study is the first study to detect Enterocin P in E. durans.
Enterocin bac 32 was identified in three of our five E. durans strains. This peptide was firstly detected in a vancomycin-resistant clinical E. faecium VRE200 strain, exhibiting activity against Enterococcus spp [24]. Although this bacteriocin has not been extensively studied, it seems to be identical to enterocin IT [25].
The strain E. durans 61A has been previously described, and durancin 61A and enterocins L50A and L50B were identified using mass spectrometry [26,27]. However, the genetic determinants for these bacteriocins were not detected in strain 61A using whole genome sequencing (WGS) in our study; instead, enterocin P was detected. Duracin 61A is not in the anti-SMASH and BAGEL4 databases (we used data from the NCBI and NCBI plus UniProt, respectively), whose genetic determinants have yet to be described. In contrast, enterocin P might not have been detected in other studies, as it is a temperature-regulated bacteriocin that is synthesized optimally at 37-47 • C [23].
Genes encoding enterocin A and enterocin B were detected in all of our five E. faecium strains, two of which also carried the genes encoding Enterocin NKR-5-3-A/D/Z.
Enterocin A was first identified in 1996 [28] and is produced by several strains of E. faecium-CTC492, T136, and P21-isolated from Spanish sausage; BFE900 from black olives; DPC 1146, WHE 81, and EFM01 from dairy products; and the N5 strain of "nuka", a Japanese rice paste. Enterocin A shows activity against Enterococcus spp., Lactobacillus spp., Pediococcus spp., and L. monocytogenes [10]. However, its activity has not been tested against clostridial species. Enterocin A is usually co-produced with enterocin B, which is produced by E. faecium T136 isolated from Spanish fermented sausages [29]. Enterocin B shows antimicrobial activity against gram-positive bacteria, such as L. monocytogenes, Propionibacterium spp., C. sporogens, and C. tyrobutyricum [29]. When enterocin A and enterocin B are co-produced, they form a heterodimer, and studies have demonstrated its potential anti-bacterial and anti-biofilm activities against S. aureus, Acinetobacter baumannii, L. monocytogenes, and E. coli [30].
The genetic determinants for enterocin NKR-5-3-A/B/C/D/Z were detected in two of our E. faecium strains. These enterocins have been purified and studied previously [31]. NKR-5-3-A (identical to brochocin A) and NKR-5-3-Z are class IIb bacteriocins and exhibit synergistic antimicrobial activity. NKR-3-5-B is a novel circular bacteriocin belonging to class IIc bacteriocins with a broad spectrum of antimicrobial activities against Bacillus spp., Enterococcus spp., and gram-negative bacteria (E. coli and Salmonella). NKR-5-3-C is a class IIa bacteriocin with strong antimicrobial activity against L. monocytogenes. NKR-5-3-D, a class IId bacteriocin, has a weak antimicrobial activity but can be produced even under unfavorable conditions [32,33]. NKR-5-3-A, D, and Z variant genes were detected in the two E. faecium strains. The genetic determinants of enterocins NKR-5-3-A/C/D/Z are closely located in a gene cluster (13 kb long) and include specific bacteriocin biosynthetic genes, such as an ABC transporter gene (enkT), two immunity-related genes (enkIaz and enkIc), a response regulator (enkR), and a histidine protein kinase (enkK). This gene cluster is essential for the biosynthesis and regulation of NKR-5-3 enterocins [34].
Genes encoding enterocin SE-K4 and staphylococcin C55a/b were identified in the two E. faecalis strains in this study. Enterocin SE-K4 was first identified in E. faecalis K-4 isolated from grass silage [35]; it grows at 43-45 • C and exhibits antimicrobial activities against E. faecium, E. faecalis, B. subtilis, C. beijerinckii, and L. monocytogenes. This enterocin has a high degree of homology to bacteriocin 31 and T8/43 [10]. Staphylococcin C55a/b was originally found to be produced by S. aureus C55 [36], consisting of three distinct peptide components termed staphylococcins C55a, C55b, and C55g. Staphylococcins C55a and C55b (lantibiotic components) acted synergistically against S. aureus and M. luteus [36]. It is a plasmid-encoded bacteriocin [37]; thus, the plasmid transfer between the producer, Staphylococcus, and the E. faecalis strains could account for the presence of the genetic determinants of this bacteriocin.

BP+ Enterococcus Resistance Phenotype and Resistome
Five of the twelve BP+ enterococci, all from E. durans isolates, were susceptible to the nine antibiotics tested. The remaining strains showed resistance to at least one of the antibiotics. Generally, the enterococci of poultry origin have more resistance genes than those of other origins (camel and camel milk). The only gene discovered in the E. durans strains of milk origin was aac(6 )-Iih, which is intrinsically present in E durans [38,39]. Antibiotics are commonly used in poultry farming, leading to the development of acquired resistance mechanisms in poultry-derived strains.
The genus Enterococcus is characterized by its intrinsic resistance to several antibiotics and ability to acquire new resistance mechanisms [40]. Enterococci are naturally resistant to semisynthetic penicillins (a reduced susceptibility), aminoglycosides (in low levels), vancomycin (at a low level and only in the species E. gallinarum and E. casseliflavus/E. flavescens, which are carriers of vanC genes), to lincosamides, polymyxins, and streptogramins (the species E. faecalis) [41]. In addition, E. faecium carries some intrinsic genes, such as msrC and aac(6 )-Ii, whereas E. durans harbors the gene aac(6 )-Iih [38,39]. Antibiotic resistance can occur either through the acquisition of genetic elements containing the resistance genes or via DNA mutations (mostly in genes encoding antibiotic targets), which are favored when there is a selective antibiotic pressure [40].
E. faecium strains X2893 and X2906 carry only chromosomal and intrinsic resistance genes (msr(C) and aac(6 )-Ii), which are non-transferable; therefore, these strains are excellent candidates for use as potential protective cultures.
Specific mutations in the pbp5 and gyrA/parC genes are associated with resistance to beta-lactams and fluoroquinolones, respectively [43][44][45]. Different mutations in the pbp5, gyrA, and parC genes have been detected in our strains, although, in most cases, with an unknown resistance phenotype associated.

Virulence of BP+ Enterococci
Different virulence factors are involved in the attachment to host cells and extracellular matrix proteins (AS, Esp, Hyl, and EfaA), macrophage resistance (AS), and cell and tissue damage (Cyl and GelE) [46,47]. Thus, although enterococci are commensal bacteria found in the intestine, they can still cause infections. Therefore, the Food and Drug Administration (FDA) has not yet assigned them to the GRAS category. Genes encoding these virulence factors are located in conjugative plasmids (agg, cyl, or hyl), in the chromosome (gelE or fsr), or in regions of the chromosome called pathogenic islands (esp and cyl) [48,49].
In the 12 enterococcal strains, virulence genes were detected in E. faecium and E. faecalis but not in E. durans. E. faecalis has already been described as more virulent than other species [50]. Fifteen virulence genes were detected in both E. faecalis strains. However, the presence of these genes is not always related to the virulence potential, as they are sometimes silenced and not associated with the phenotype [49]. Both strains carried the gelE gene, which is associated with gelatinase activity, but only strain X3198 was positive for gelatinase activity.
All the E. faecium strains carried the functional collagen adhesin gene, acm, which plays an essential role in colonization by binding to collagen type I, with less affinity to collagen type IV [51]. As these E. faecium strains did not carry other virulence factors, the presence of acm might be positive, as it could facilitate the colonization of this beneficial strain. Nevertheless, as mentioned before, the presence of a virulence gene does not always indicate that it is being expressed [49]. Therefore, further studies must uncover whether acm is, in fact, expressed as a virulence factor.

Plasmidome of the BP+ Enterococci
Ten of the BP+ enterococci harbored at least one plasmid. Interestingly, strains X2893 and X2906 did not present any mobile genetic elements, which, along with the other characteristics, makes them good candidates for potential protective cultures [52].

Enterococcus Sampling and Identification
In total, 251 enterococcal isolates of poultry origin were used in this study: (a) 60 isolates were collected during this study from poultry carcass samples obtained from different supermarkets and butchers in the La Rioja Region (Spain), the isolates recovered in the Slanetz-Bartley agar (OXOID); (b) 166 isolates were previously obtained from poultry carcasses at the slaughterhouses' level in Tunisia; (c) 25 poultry isolates were obtained from the University of La Rioja's collection (Spain). Additionally, 5 BP+ enterococci of other origins (2 isolates from cow milk and 3 from camel milk) were obtained from the University of LAVAL's strain collection (Canada).

Screening for Anti-C. perfringens Activity Using the "Spot on the Lawn" Method
The antimicrobial activity of the 256 Enterococcus isolates against the indicator strain, C. perfringens X2967 (a clinical strain obtained from the Hospital San Pedro, Logroño, Spain), was analyzed using the "spot-on-the-lawn" method [53]. The active isolates were identified as BP+. Briefly, a fresh culture of C. perfringens strain X2967 was suspended in brain-heart infusion broth (BHI) (turbidity 0.5 MacFarland). Subsequently, 10 µL of this indicator microorganism solution was added to tubes containing 5 mL of semi-solid melted tryptic soy broth (TSB) and supplemented with 0.7% agar and 0.3% yeast extract. Finally, the semi-solid TSB medium with the indicator microorganism was poured onto tryptic soy agar plates (TSA). Once the plates were dried, the enterococcal microorganisms were stingseeded, and the plates were incubated at 37 • C for 24 h under strict, anaerobic conditions. Strains that showed inhibitory activity against C. perfringens strain X2967 were tested against other relevant pathogens and multidrug-resistant (MDR) bacteria using the same test. This panel included E. casseliflavus C1232, E. gallinarum C2310, E. faecium C2321, E. faecalis C410, E. durans C1433, E. hirae C1436, MSSA C411, MRSA C1570, M. luteus C157, L. monocytogenes C137, S. suis C2058, E. coli C408, S. enterica C660, Y. enterocolitica X3080, and P. aeruginosa X3282. A blood agar plate was used for S. suis testing. All strains used as indicator bacteria came from the University of La Rioja's collection.

Screening for Anti-C. perfringens Activity Using the Agar Diffusion Method
NHS and HS extracts were prepared from Enterococcus isolates showing inhibitory activity in the spot-on-the-lawn assay. These supernatants were tested against a collection of 20 C. perfringens isolates using the previously described agar diffusion method [54], with nisin as a positive control. The C. perfringens isolates were collected from the NE of poultry origin (University of Laval, Quebec, QC, Canada).
To prepare the NHS, enterococci were inoculated in 10 mL of TSB in sterile tubes and were incubated overnight at 37 • C. Then, the culture medium was centrifuged at 5000× g rpm for 5 min and filtrated using 0.20 µm filters. Next, a fraction of this supernatant was heated at 100 • C for 15 min and used as the HS. For the concentrated supernatants, the culture cell media were concentrated 10 times using a Speed Vac (Thermo Scientific Savant, Asheville, NC, United States) after centrifugation.
For the agar well-diffusion method, C. perfringens was cultured in a reinforced clostridial medium (RCM) (Himedia, Kennett Square, PA, USA) supplemented with 10% agar. The plates were incubated overnight at 37 ºC under strict, anaerobic conditions.

Anti-C. perfringens Activity Determination Using Microtitration Assay
A microtitration assay was performed to determine the total activity (AU/mL) of the active supernatant of BP+ enterococci against the C. perfringens ATCC 13124 strain, as described previously [55,56]. The BHI was used as the growth medium for C. perfringens and was added to the wells, with a final bacterial concentration of~10 5 CFU/well. The microplate was incubated for 24 h at 37 • C under strict, anaerobic conditions. After incubation, the optical density was measured at 595 nm using a microplate reader (Infinite M200, Tecan, Männedorf, Switzerland) to determine the number of wells in which inhibition occurred.
The following formula was used to calculate the total arbitrary activity: where 2 is the dilution factor, n is the number of inhibition wells, 1000 is the factor for reporting the result per mL, and 125 is the volume of the solution tested in microliters.

Characterization of BP+ Enterococci
Twelve BP+ enterococci were chosen for further characterization based on their antimicrobial activity against C. perfringens strains.

Gelatinase Activity and Hemolysis
The gelatinase activity and hemolytic capacity of BP+ enterococci strains were determined as reported previously [58], using TSA supplemented with 3% skim milk and blood agar, respectively.

Conclusions
Among the 12 enterococci that showed inhibitory activity against C. perfringens, the strains E. faecium X2893 and X2906 seem to be the most promising candidates for use as protective cultures in poultry farming. Both strains belong to the sequence type ST722 and harbor enterocin A and Enterocin B genetic determinants. These strains also do not have acquired resistance genes, do not carry plasmids, and only carry the acm gene, which is implicated in host colonization and might be a desirable feature for protective strains. Both are gelatinase-negative and gamma-hemolytic.
The strains derived from other origins (milk and camel milk) and belonging to the species E. durans might be also good candidates as protective cultures, as they do not harbor any virulence factors or resistance genes, and they produce bacteriocins. However, these strains carry more than one plasmid and have not been isolated from poultry.
Concluding, E. faecium X2893 and X2906 showed potential to be considered in further studies as protective cultures in poultry farming, a promising alternative to antibiotic use in this sector.