Purification and characterization of bacteriocins-like inhibitory substances from food isolated Enterococcus faecalis OS13 with activity against nosocomial enterococci

Nosocomial infections caused by enterococci are an ongoing global threat. Thus, finding therapeutic agents for the treatment of such infections are crucial. Some Enterococcus faecalis strains are able to produce antimicrobial peptides called bacteriocins. We analyzed 65 E. faecalis isolates from 43 food samples and 22 clinical samples in Egypt for 17 common bacteriocin-encoding genes of Enterococcus spp. These genes were absent in 11 isolates that showed antimicrobial activity putatively due to bacteriocins (three from food, including isolate OS13, and eight from clinical isolates). The food-isolated E. faecalis OS13 produced bacteriocin-like inhibitory substances (BLIS) named enterocin OS13, which comprised two peptides (enterocin OS13α OS13β) that inhibited the growth of antibiotic-resistant nosocomial E. faecalis and E. faecium isolates. The molecular weights of enterocin OS13α and OS13β were determined as 8079 Da and 7859 Da, respectively, and both were heat-labile. Enterocin OS13α was sensitive to proteinase K, while enterocin OS13β was resistant. Characterization of E. faecalis OS13 isolate revealed that it belonged to sequence type 116. It was non-hemolytic, bile salt hydrolase-negative, gelatinase-positive, and sensitive to ampicillin, penicillin, vancomycin, erythromycin, kanamycin, and gentamicin. In conclusion, BLIS as enterocin OS13α and OS13β represent antimicrobial agents with activities against antibiotic-resistant enterococcal isolates.


Materials and methods
Bacterial strains and culture conditions. The bacterial strains used in this study and their culture conditions are presented in Table 1. All strains were maintained in glycerol stock 40% (v/v) and stored at − 80 ºC.

Isolation and identification of E. faecalis. Different food samples
(milk products; mozzarella cheese and buttermilk-pastries; commercial cakes and bread) and clinical samples (urine and stool) were collected in sterile plastic cups from Beni-Suef City, Egypt, and refrigerated for 24 h prior to processing. Urine sample precipitates (collected after centrifugation at 5000×g for 10 min) and 1 g of each of the food and stool samples were inoculated into tubes containing 5 mL of GM17 broth (Oxoid, Hampshire, UK) and incubated aerobically at 37 °C for 24 h.
Enterococci were isolated by selective culture on bile esculin agar (Oxoid), where they grow in the presence of 4% (w/v) bile and hydrolyze esculin to esculetin that reacts with Fe 3+ to form a dark brown to black precipitate on the agar plate. The resultant black colonies were presumptively identified as Enterococcus sp. using the l-pyrrolidonyl-β-naphthylamide (PYR) test and then identified at species level using the API 50  www.nature.com/scientificreports/ Hilden, Germany). Quantification of the PCR products was determined by optical density measurement using a NanoDrop ND1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA). The PCR product was sequenced and analyzed using an ABI Prism BigDye Terminator v3.1 Ready Reaction Cycle Sequencing Kit on Table 1. Inhibition spectrum of BLIS from E. faecalis isolates using a deferred inhibition test method. a ATCC, "American Type Culture Collection". b SDA, "Sabouraud Dextrose Agar"; BHI, "Brain Heart Infusion"; MRS, "deMan, Rogosa and Sharpe"; GM17, "Glucose-M17". c + , clear inhibition zones of 12 mm or more were recorded as a positive antibacterial activity. -, no inhibition.

Indicator strain a Description
Growth medium b  Tested indicator  strain  BHI  37  -----------Candida albicans  ATCC 60193   Tested indicator  strain  SDA  30  ---------- Pediocin sensitive  strain  GM17  37  -----------Pseudomonas  aeruginosa ATCC  9027   Tested indicator  strain  BHI  37  -----------Salmonella  enterica ATCC  35664   Tested indicator  strain  BHI  37  ---------- Screening for antimicrobial activity production. The antimicrobial activity was detected using a deferred inhibition test 27 . Briefly, the colonies were overlaid with 5 mL of soft agar containing 70 µl of the indicator strains which were grown to late exponential phase (Table 1). A confirmatory test was established by adding aliquots of the pH-adjusted supernatant (50 µl) into 0.5-cm wells, made in the inoculated agar plates with indicator strain. Then, the plates were incubated overnight and checked for the presence of growth inhibition zones. Quantitative estimation of BLIS in the cell-free culture supernatant was performed using a microtiter plate assay 28 , where 100× dilutions of the sensitive indicator strains were exposed to twofold serial dilutions of bacteriocin in growth media. The plates were incubated overnight, and the growth was measured spectrophotometrically at 620 nm using a microtiter plate reader (iEMS Reader; Labsystems, Helsinki, Finland). One arbitrary activity unit (AU) was defined as the reciprocal of the highest dilution of bacteriocin, causing 50% growth inhibition of the treated indicator strain by measuring growth turbidity inhibition at 620 nm in the microtiter plate reader compared to the untreated control 29 . The bacteriocin activity was calculated according to the following equation; Bacteriocin activity (AU/mL) = (d × 1/m), where "d"; is the used dilution factor and "m" is the highest dilution causing 50% growth inhibition.
Determination of virulence traits (hemolytic activity, proteolytic gelatinase activity, and bile salt hydrolase activity). Hemolytic activity, being alpha, beta, or gamma, was determined by streaking putative bacteriocinogenic E. faecalis isolates onto a fresh plate of Columbia blood agar (Oxoid) containing 5% (v/v) defibrinated horse blood and incubating at 37 °C for 24 h. The formation of a clear zone around the colonies indicated beta complete hemolytic activity. In contrast, the formation of greenish discoloration surrounding the colonies indicated alpha partial hemolytic activity, and the absence of any hemolysis was referred to as gamma or no hemolysis. Proteolytic gelatinase activity was detected by inoculating E. faecalis isolates into tubes containing gelatin medium composed of 0.5% (w/v) peptone, 2% (w/v) gelatin, 0.3% (w/v) beef extract, 0.5% (w/v) NaCl and final pH was adjusted to 7.2. All tubes, including the control (uninoculated), were incubated at 37 °C for 24 h. Gelatin liquefaction observed after keeping the tubes at 4 °C for 3 h, indicated positive gelatinase activity. Putative bacteriocinogenic E. faecalis isolates were cultured on both control MRS agar and MRS agar containing 0.5% (w/v) thioglycolate and 0.5% (w/v) sodium glychodeoxycholate (bile salt) to investigate bile salt hydrolase activity. The formation of white precipitate around the colonies indicated bile salt hydrolase activity.
Antibiotic susceptibility testing and minimum inhibitory concentration (MIC) determination. In order to perform antibiotic susceptibility typing of the 11 putative bacteriocinogenic E. faecalis isolates to check whether they are relevant to each other or from different clones, and to overview their biosafety for further applications, the Clinical and Laboratory Standards Institute (CLSI) broth microdilution method in microtiter plates 30 was used to determine the antibiotic resistance phenotype and MIC of these isolates against 13 antibiotics with different concentration ranges (Table S2), using Mueller-Hinton broth (Oxoid). The microtiter plates were incubated at 37 °C for 20 h before visually determining the MIC, which is the lowest drug concentration that causes complete inhibition of microorganism growth.

Multilocus sequence typing (MLST).
An MLST scheme based on seven housekeeping genes was performed to investigate the epidemiology and population structure of the 11 putative bacteriocinogenic E. faecalis isolates 31 . The PCR primers are listed in Table S1. The allelic profile or sequence type (ST) was determined, according to the E. faecalis MLST database (http:// efaec alis. mlst. net/) 31 . The ST distribution among all putative bacteriocinogenic E. faecalis isolates, included in the MLST database, was evaluated using eBurst v3 32 .
Purification of BLIS. The BLIS was purified as previously described 33 . Briefly, 2 L of E. faecalis OS13 cultivated in MRS broth at 30 °C for 18 h was used for further purification steps. After centrifugation at 14,500×g for 10 min at 4 °C (Sorvall RC-6; Kendro Laboratory Products, Asheville, NC, USA) and filtration (500 mL Steritop 0.22 µm filter system; Millipore, Dublin, Ireland), the supernatant was subjected to ammonium sulfate precipitation 50% (w/v) and shaken vigorously at 4 °C for 30 min. The protein pellet was harvested by centrifugation at 14,500×g for 10 min at 4 °C and then redissolved in 40 mL of distilled water to obtain a 50× concentrated crude BLIS solution. Whenever necessary, the pH was adjusted to 3.5 using 0.1 M HCl.

Results
Isolation and identification of E. faecalis. The 65 isolates (43 from food and 22 from clinical samples) that were collected from various food and healthcare units were positive for esculin hydrolysis on bile esculin agar and for l-pyroglutamic acid beta-naphthylamide hydrolysis on PYR discs (results not shown). They were identified at the species level as E. faecalis using an API 50 CHL carbohydrate fermentation kit and were positive for glycerol, ribose, galactose, glucose, fructose, mannose, mannitol, sorbitol, N-acetyl glucosamine, amygdalin, arbutin, esculin, salicin, cellobiose, maltose, saccharose, trehalose, melezitose, and tagatose. Furthermore, 16S rDNA sequence analysis confirmed their identity as E. faecalis. The obtained sequences of the bacteriocinogenic isolates were submitted to the NCBI GenBank database under specific accession numbers (Table S3). The partial 16S rRNA gene sequences were aligned to closely related species retrieved from the NCBI GenBank database and assembled in MEGA7 software to construct a phylogenetic tree (Fig. 1).
Screening for bacteriocin genes and spectrum of antimicrobial activity. Interestingly, when the putative bacteriocinogenic E. faecalis isolates were screened by PCR for genes corresponding to any of 17 known bacteriocin genes, none of these genes were found in 11 of the isolates (three from food samples and eight from clinical samples). Of them, the food-isolated E. faecalis OS13 produced the most potent antimicrobial activity and was chosen for further study and purification of its biologically active antimicrobial substance. Some Grampositive bacteria, such as L. sakei, L. plantarum, E. faecalis, and E. faecium were sensitive to the antimicrobial activity of the OS13 isolate. Still, other Gram-positive bacteria, such as Listeria, Bacillus, and Staphylococcus, and all tested Gram-negative bacteria were resistant. The most potent antimicrobial activity was shown against L. sakei LMG 2313, which was then used as the primary indicator microorganism during purification.
Characterization of the isolates producing antimicrobial activities at biosafety and epidemiology levels. All selected 11 isolates with antimicrobial activities demonstrated bile salt hydrolase activity except strains OS11, 13, 16, and 64. Furthermore, all strains were non-hemolytic and gelatinase-positive. They were sensitive to ampicillin and vancomycin (Table S2) but resistant to clindamycin, fusidic acid, apramycin, streptomycin, tetracycline, and polymyxin B (Table S2, Fig. S1). Other antibiotics sensitivities and resistance patterns are listed (Table S2). Clustering based on the resistance pattern of these isolates is presented in Fig. S1. Finally, the multilocus sequence typing (MLST) of seven housekeeping genes 31 revealed that these isolates belonged to ST 6 and ST116 (Table S3). ST116, representing E. faecalis OS13, did not belong to clonal complex (CC) 6 or CC9, which are commonly spread in the hospital environment (Fig. S2).  Table 2). The activity (AU/mL) was determined by microtiter plate assay using L. sakei LMG 2313 as the indicator strain. The final step of reverse-phase chromatography using a Sephasil Peptide C8 column allowed the separation of two peaks with antimicrobial activity, eluted at approximately 39.5% (v/v) and 40.5% (v/v) isopropanol, respectively (Fig. 2).

Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and mass spectrometry.
The partially purified BLIS obtained by cation exchange chromatography were analyzed by one-dimen-  Table 2. Purification of enterocin OS13. a (AU/mL), Arbitrary unit per mL.  (Fig. 3A). The molecular weights of the purified agents were further determined by MALDI-time of flight mass spectrometry (TOF/MS) analysis. The two BLIS were designated enterocin OS13α and OS13β and showed a molecular weight of 8079.1 Da (Fig. 3B) and 7858.6 Da (Fig. 3C), respectively. A peptide mass similarity search using ExPASy revealed that there were no molecular weight similarities with any other AMPs except holotricin-2 (7858 Da) 36 .

Purification step Vol (mL) Activity per mL (AU/mL) a Total activity (AU) Recovery (%)
Effect of heat and proteinase K enzyme. When the crude BLIS with an initial activity of 10,240 AU/ mL was heated to 100 °C, the activity was reduced to 640 AU/mL (6.25%). Also, the antimicrobial activity was decreased to 1280 AU/mL (12.5%) after proteinase K treatment. Repeating the same treatment with the concentrated purified enterocin OS13α with initial activity (1,024,000 AU/mL) and enterocin OS13β with initial activity (128,000 AU/mL) confirmed the sensitivity of both peptides to heat treatment, and their activities were entirely abolished. Proteinase K treatment reduced the activity of enterocin OS13α to 512,000 AU/mL (50%), but enterocin OS13β remained unaffected with an activity of 128,000 AU/mL.

Discussion
Enterococcus faecalis are widely spread in the environment and are commonly isolated from foodstuffs and clinical samples 37 . E. faecalis and other LAB have attracted more attention over the last few years due to the production of many useful AMPs or BLIS, with several safe applications in the food industry and medical treatments 38,39 . The identification of antimicrobial agents to treat antibiotic-resistant nosocomial enterococci is crucial because the treatment of such infections is becoming increasingly problematic. In addition, enterococci gained greater medical and public attention since the mid-1980s when vancomycin resistance was first present in clinical isolates (VRE) 40 . BLIS with the potential to inhibit the growth of such bacteria could remedy infections without affecting endogenous beneficial microbiota or risking dissemination of R-factors to non-enterococcal pathogens (e.g.,

vancomycin-intermediate Staphylococcus aureus).
Our study searched for suitable candidate BLIS in 65 E. faecalis isolates from Egypt of food and clinical origin. The 11 unknown putative bacteriocinogenic E. faecalis isolates were characterized epidemiologically by MLST and tested for antibiotic resistance, virulence traits, and antimicrobial activities. Our study showed that, E. faecalis OS13 produced highly active (> 10 6 AU/mL) BLIS designated enterocin OS13α and OS13β that had antimicrobial activity against some Gram-positive bacteria, such as L. sakei, L. plantarum, E. faecalis, and E. faecium, but not against Gram-negative bacteria. Generally, the bacteriocins produced from Gram-positive bacteria are active against closely related species and not directly active against Gram-negative bacteria. However, some other bacteriocins from LAB showed antagonistic activities against both Gram-positive and Gram-negative bacteria, like enterocin E760 41 , enterocin E50-52 42 , and salivaricin SMXD51 43 . Owing to the narrow-spectrum capabilities of enterocin OS13α and OS13β against target pathogens, they are favored when compared with conventional broad-spectrum antibiotics that can interfere with the host microbiota, affecting their crucial beneficial roles 22 . Especially during infancy and early childhood, exposure to broad-spectrum antibiotics is often unfavorable, as  www.nature.com/scientificreports/ the early microbiota lacks diversity and stability, rendering it distinctly vulnerable to disturbance, in addition to the importance of intestinal microbiota in early development and education of the immune system 23 .
Noticeably, bacteriocins from isolate OS13 were active against multi-resistant nosocomial enterococci as Enterococcus faecalis V583, which was the first VRE Enterococcus isolated in the United States, and it was also resistant to high levels of aminoglycosides, macrolides, lincosamides, and streptogramin B 44 . Therefore, BLIS could represent a non-conventional alternative of interest to combat pathogens like enterococci, recognizing that the presence of natural resistance prevents a variety of modern drugs widely used, such as cephalosporins and quinolones, from working effectively against enterococcal infections 45 . Most VREs are also resistant to antibiotics as ampicillin, vancomycin, and aminoglycosides, widely used in the hospital environment. Therefore, only a few therapeutic options are commercially available to treat multidrug-resistant VRE infections, including linezolid, tedizolid, tigecycline, and daptomycin 46,47 . The increased use of these last-source antibiotics and the ability of enterococci to readily acquire new resistance determinants are currently raising the need to find alternative therapeutics and explore the effectiveness of natural antimicrobials such as bacteriocins.
In this study, purification of two BLIS, OS13α and OS13β, from strain OS13 and based on their chromatographic properties exposed that both BLIS were cationic and hydrophobic, which is in agreement with most bacteriocins purified from LAB [48][49][50][51] . At each purification level, the antimicrobial activity decreased, especially after reverse-phase chromatography, which was observed in other studies 34,52 . SDS-PAGE did not accurately determine the molecular weight of both BLIS and was unable to differentiate between them 53,54 .
Although both enterocin OS13α and OS13β were heat-labile, and their activity was wholly abolished after heating at 100 ºC for 30 min, their activity together as a crude BLIS in the supernatant was decreased to 6.25% following heat treatment. There are previous reports of many BLIS becoming less heat-stable during subsequent purification steps 55,56 . The larger size of enterocin OS13α and OS13β being 8079.1 Da and 7858.6 Da, respectively, could interpret the reasons behind their heat sensitivity 39,57,58 . Although the crude BLIS with an initial activity of 10,240 AU/mL was decreased to be 1,280 AU/mL (12.5%) after proteinase K treatment but the purified enterocin OS13α was sensitive to proteinase K (50% reduction in its activity), which strongly suggests its proteinaceous nature 52,59 . Also, enterocin OS13β was resistant to proteinase K, and this is probably not due to the absence of digestion sites but more likely due to a tightly folded structure, making it inaccessible to enzymatic digestion 60,61 . The variation in sensitivity of crude BLIS compared to purified ones to proteinase K could indicate that the OS13 isolate might also produce other BLIS besides OS13α and OS13β that is/are more sensitive to proteinase K but was/were not isolated and purified.
Generally, Enterococcus spp. found in food are presumed less virulent than clinical isolates 62,63 . Regarding the virulence traits of strain OS13, neither hemolytic activity nor bile salt hydrolase activity was demonstrated, while gelatinase activity was observed. Cytolysin (also called hemolysin) is a lantibiotic, but it differs from most lantibiotics because it is both antibacterial and hemolytic 64 . A higher incidence of cytolysin genes has been shown in clinical isolates (33%) compared with food isolates (6%) 65 . In our study, PCR analysis was used to confirm the absence of the cytolysin gene cluster (cyl). Although E. faecalis can be an opportunistic pathogen that possesses bile salt hydrolase activity 66 , such activity might be a desirable trait for a probiotic bacterium since it could maximize its ability to survive in the hostile environment of the gastrointestinal tract 67 . The gelatinase enzyme is involved in the virulence of Enterococcus sp. and hydrolyzes gelatin, casein, hemoglobin, and other bioactive peptides 65 . The primary role of gelatinase in enterococcal pathogenesis is to provide nutrients to the bacteria by degrading host tissue. However, it also appears to play a role in biofilm formation 68 .
The food chain is considered the primary transmission route of antibiotic-resistant bacteria between populations. Enterococcus sp. behave as opportunistic pathogens in nosocomial infections and spread antibiotic resistance through the food chain with potential risks to humans 69 . In biosafety evaluations of enterococci, their resistance to glycopeptides, such as vancomycin, should be considered 70 . In our study, the pathogenic potential of the food isolated E. faecalis OS13, and the spectrum of resistance to various antibiotics was determined. OS13 was sensitive to ampicillin, penicillin G, vancomycin, erythromycin, kanamycin, and gentamicin. It was resistant to chloramphenicol, streptomycin, tetracycline, clindamycin, fusidic acid, and polymyxin B. These results are in agreement with previous publications 71,72 , where all tested E. faecalis strains were sensitive to ampicillin, penicillin, and vancomycin. Resistance to chloramphenicol, tetracycline, erythromycin, gentamicin, and kanamycin is becoming common among enterococci from different sources due to the abuse of antibiotics in clinics, animal and fish farming, and agriculture. Thus, enterococcal antibiotic resistance is not exclusive to the clinical field and is also prevalent in the food industry 71 .
Although MLST is a tool for global epidemiological studies, the possible role of plasmids in transferring virulence between different strains should be considered 73 . MLST analysis of E. faecalis OS13 revealed that it belongs to ST116 and CC116. E. faecalis with ST116 has been reported as isolated twice from human clinical samples; vancomycin-resistant E. faecalis from a blood sample in Denmark and E. faecalis DQ287 from a catheter in Cuba 74,75 . To our knowledge, ours is the first report of the isolation of E. faecalis with ST116 from non-human samples.
The ability of CC116 to spread in hospitals and cause infection is unknown. However the most common clonal complexes that spread in the hospital environment are CC2, CC6, and CC9 31,76 .

Conclusion
The food-isolated E. faecalis OS13 produced BLIS named enterocin OS13α (8079 Da) and OS13β (7859 Da), with activities against VRE and MDRE. MLST of E. faecalis OS13 revealed its belongings to ST116, being the first from non-human samples. E. faecalis OS13 was non-hemolytic, bile salt hydrolase-negative, and sensitive to ampicillin, penicillin, vancomycin, erythromycin, kanamycin, and gentamicin, suggesting several safe applications in the food industry and medical treatments. Although our BLIS showed sensitivity to heat treatments, they could be Scientific Reports | (2021) 11:3795 | https://doi.org/10.1038/s41598-021-83357-z www.nature.com/scientificreports/ used for preservation of raw or minimally processed foods stored at low temperatures besides their likely role in the treatment of clinical infections with VRE and MDRE. Future works are still required to figure out the amino acid sequences, structures, encoding genes, and mode of actions of such BLIS.