Identification of Multiple Bacteriocins in Enterococcus spp. Using an Enterococcus-Specific Bacteriocin PCR Array

Twenty-two bacteriocin-producing Enterococcus isolates obtained from food and animal sources, and demonstrating activity against Listeria monocytogenes, were screened for bacteriocin-related genes using a bacteriocin PCR array based on known enterococcal bacteriocin gene sequences in the NCBI GenBank database. The 22 bacteriocin-positive (Bac+) enterococci included En. durans (1), En. faecalis (4), En. faecium (12), En. hirae (3), and En. thailandicus (2). Enterocin A (entA), enterocins mr10A and mr10B (mr10AB), and bacteriocin T8 (bacA) were the most commonly found structural genes in order of decreasing prevalence. Forty-five bacteriocin genes were identified within the 22 Bac+ isolates, each containing at least one of the screened structural genes. Of the 22 Bac+ isolates, 15 possessed two bacteriocin genes, seven isolates contained three different bacteriocins, and three isolates contained as many as four different bacteriocin genes. These results may explain the high degree of bactericidal activity observed with various Bac+ Enterococcus spp. Antimicrobial activity against wild-type L. monocytogenes and a bacteriocin-resistant variant demonstrated bacteriocins having different modes-of-action. Mixtures of bacteriocins, especially those with different modes-of-action and having activity against foodborne pathogens, such as L. monocytogenes, may play a promising role in the preservation of food.


Introduction
Enterococcus is a genus of Gram-positive, catalase-negative, facultative anaerobic cocci. Its members were removed from the genus Streptococcus in the 1980s because of DNA hybridization data showing that they were not closely related to streptococci and were moved into their own genus. The genus Enterococcus is now comprised of more than 20 species of which En. faecalis and En. faecium are the most common species and are readily found in the feces of mammals. Fecal enterococci have been classified as "fecal indicator" bacteria in much the same way as fecal coliforms, i.e., as an indicators of unsanitary manufacture or processing of foods. Because of their association with animals and fecal distribution, enterococci are found worldwide in a large variety of fresh and prepared foods, vegetables, hard cheeses, meats and prepared meat products [1][2][3]. In France, Enterococcus spp. have been found in 44% of pasteurized-milk cheeses and as high as 92% of raw-milk cheeses [4]. They are present in numerous so-called "artisanal" food products that many are considered safe for consumption even though they are not generally regarded as safe (GRAS) by USDA regulatory approval. Although they have been involved in human enterococcal disease, they are considered opportunistic infections and account for a considerable amount of hospital-acquired nosocomial infections in the US [5]. Certain strains, however, are considered non-pathogenic and have been used as dietary adjuncts for reducing antibiotic-associated diarrhea in children [6,7]. In a comparative genomic study of probiotic En. faecium strain T-110, a non-pathogenic strain (NRRL B-2354), and three pathogenic strains, the non-pathogenic/probiotic strain was found to be lacking a number of virulence-related genes that were found in the pathogenic strains [8]. In an effort to define the desirable characteristics for Enterococcusbased probiotics, researchers have been examining genes in isolates from healthy individuals vs. those obtained as pathogenic strains from clinical isolates. Pathogenic strains of Enterococcus spp. exhibit a number of virulence factors including gelatinase, adhesion to collagen, aggregation substance, endocarditis antigen, and β-hemolytic substances which make them less likely candidates for use as food starter cultures [9].
Enterococcus spp. are also known to produce bacteriocins that have demonstrated antimicrobial activity against various foodborne pathogens [10,11]. Bacteriocins are ribosomally synthesized peptides produced by bacteria that are capable of killing other bacteria by forming pores in target membranes [12][13][14]. In lieu of issues with live cultures which may be consumed, the application of cell-free culture supernatants may be far less objectionable and find greater acceptance as food preservatives. This study examines the use of an Enterococcus-based PCR bacteriocin array to determine the presence of bacteriocin genes in 22 Bac+ Enterococcus strains isolated from food and animal sources having activity against Listeria monocytogenes.

Bacterial Strains, Storage, and Growth Conditions
This study included 22 isolates of Enterococcus spp. of which twelve were En. faecium, four were En. faecalis, three were En. hirae, two were En. thailandicus, and one was En. durans. Frozen master cultures of enterococcal strains were stored in milk-based freezing media at −80 °C, and 100 μL of master culture was added to 10 mL of de Man, Rogosa, and Sharpe (MRS) broth at 30 °C for 12-16 h prior to use. Although MRS broth/agar would not usually be the initial choice for Enterococcus spp., they were initially isolated on MRS agar and grew well when cultured in MRS broth, as they were suspected of being traditional lactic acid bacteria until their eventual identification by 16S rRNA PCR and sequence analysis [15].

Activity against Listeria Monocytogenes
All 22 strains of Enterococcus spp. were previously isolated from retail foods and animal sources as bacteriocin-producing enterococci as determined by overlay of 'sandwiched' colonies from plated samples ( Figure 1) with indicator lawns containing L. monocytogenes 39-2 as described previously [15,16]. Subsequent analysis of specific strains was observed using culture spot assay against the wild type L. monocytogenes 39-2 (R0 strain) as well as a variant of L. monocytogenes 39-2 R0 that was isolated by selection against the bacteriocin produced by Lactobacillus curvatus FS47 (curvaticin FS47) and deemed "R1". Inhibitory activity against L. monocytogenes 39-2 R0 (wild type) and R1 (curvaticin FS47 R ) was by Bac+ culture spot assay against Listeria indicator lawns.

Primer Creation
Bacteriocin genes were found using the online database Bactibase (Institut Supé rieur des Sciences Biologiques Appliqué es de Tunis, Tunis, Tunisia) [17] by searching for individual gene sequences produced by members of the genus, Enterococcus. Duplicates and highly homologous gene sequences were condensed into single primer sets. Gene selections were made using bacteriocin structural genes and neighboring immunity proteins to increase the size of the amplicon that would allow exclusion of homologous sequences. Adjacent ABC transporters and other similar features were excluded due to high homology between genes. Primers were designed from gene sequences using the online Primer3 software (Whitehead Institute for Biomedical Research, Cambridge, MA, USA) [18]. Once created, primers were analyzed against each bacteriocin gene using the MEGA 5.2 software (The Biodesign Institute, Tempe, AZ, USA) [19] to ensure cross-amplification would not occur. Primers were then ordered from Integrated DNA Technologies (IDT, Coralville, IA, USA).

PCR Detection of Enterococcal Bacteriocin Genes
The presence of bacteriocin related genes was determined by using 16 sets of primers in a PCR bacteriocin array as done previously [20]. Initially, total bacterial DNA was isolated using the BAX ® protease lysis method (DuPont Qualicon, Wilmington, DE, USA). Overnight culture (5 μL) was lysed in 200 μL protease mixture at 55 °C for 60 min followed by a deactivation step at 95 °C for 10 min. PCR amplification was performed on a DNA Engine Opticon 2 (MJ Research/Bio-Rad, Hercules, CA, USA) in 25 μL reaction mixtures. Each 25 μL reaction contained 12.5 μL iTaq™ Universal SYBR ® Green Supermix (Bio-Rad), 3.5 μL of each primer (final concentration of 60 nM for each primer), and 5 μL of the cell lysate diluted 1:5 with sterile, molecular grade water (DNAse-/RNAse-free). The cycling program was preceded by an initial denaturation at 95 °C for 15 min. The specific cycling parameters consisted of 40 cycles of the following: denaturation at 95 °C for 15 s, annealing at 60 °C for 60 s, and elongation at 72 °C for 60 s followed by a plate read. PCR products were verified by melting curve analysis (50 °C to 90 °C with a read every 0.2 °C and hold for 0.02 s) and electrophoresis on a 2% (mg/mL) agarose (FMC Corporation, Philadelphia, PA, USA gel at 80 V for 1 h, using a 100 bp ladder for size verification and viewed using the ChemiDoc™ XRS System UV transilluminator (Bio-Rad). PCR products were purified using the GenCatch™ PCR Cleanup Kit (Epoch Life Science, Missouri City, TX, USA) and submitted for sequencing to the Oklahoma State University Recombinant DNA and Protein Core Facility.

Primer Design
Approximately 37 Enterococcus bacteriocin DNA sequences found in the Bactibase database were examined for designing primers. Alignment of the sequences in the database revealed that genes with identical sequences were often denoted as different bacteriocins with different accession numbers (Table 1). Upon grouping of highly homologous sequences, the DNA sequences were condensed to 16 unique enterococcal bacteriocin sequences listed in Table 1 with the names of highly homologous sequences cited under the heading "Homologous Genes". Primary names for the primer pairs created with homologous genes were chosen subjectively.

PCR Detection of Enterococcal Bacteriocin Genes by Real-Time PCR
The 22 Enterococcus isolates used in this study were pre-screened for the ability to inhibit L. monocytogenes ( Figure 1) and we assumed they contained at least one bacteriocin gene [15]. The Enterococcus bacteriocin PCR array was obtained from previously sequenced bacteriocin genes from enterococci (as noted above), weeding out duplicates of those showing near-identical homology in order to reduce the number of primer sets in the "primer array" (Table 1). During analysis of our 22 isolates with all 16 sets of primers (i.e., the "enterococcal PCR bacteriocin array"), the gene for the bacteriocin, enterocin A (entA), occurred most frequently among our isolates (77.3%, in 17 of 22 isolates) and was found among members of four of the five species examined in this study (Table 2). Bacteriocin mr10AB was the next most frequently detected bacteriocin among our isolates (63.6%). Values for homology relatedness for the partial sequences of various Enterococcus bacteriocins are presented in Table 2. Sequences obtained from our isolates using bacteriocin mr10AB primers were more highly homologous to enterococcin L50A and L50B sequences in the GenBank database than to other enterococcal bacteriocins because they are highly homologous (>95%) and were considered interchangeable in this study. Other enterococcal bacteriocin genes found within our isolates that showed high homology to known enterococcal bacteriocins include bacA (in seven of 22 isolates, 31.8%), enxAB (18.2%), entP (18.2%), entB (9.1%), and munA (9.1%). Many of the strains demonstrated the presence of more than one bacteriocin structural gene based on real-time PCR reactions (Table 2; Figure 2A). Sequence and phylogenetic tree analysis with the genes from which their primers were derived demonstrated four bacteriocin structural genes in En. faecium 326F ( Figure 2B). Structural genes for the Enterococcus bacteriocins avicin A (avcA), columbicin A (colA), durancin Q (duqQ), enterocin 96 (ent96), enterocin C (entC), enterocin SE-K4 (entSE-K4), enterocins W α and β (enwAB), enterocin Q (entqA), and mundticin KS (munA) were not detected among our isolated strains. A maximum likelihood tree created from alignment of sequenced amplicons (using MEGA 5.2 software) show the distribution and similarity of bacteriocins among the Enterococcus isolates we examined (Figure 3).

Discussion
Enterococcus spp. are common inhabitants of mammalian intestinal flora and are often found among other lactic acid bacteria in the environment and isolated from foods. Similarly, our Bac+ enterococcal isolates represent a wide distribution of food and environmental samples such as fecal matter (323F, 326F, 341FA, Pop4), meat processing (JCP-9/B-5/M-2), dairy (Milk 5/12), rumen fluid (323RL1) and herbs (THYME 2/3). The 22 strains of Enterococcus spp. examined in this study had inhibitory activity against L. monocytogenes which was not attributed to bacteriophage, acid inhibition, or hydrogen peroxide and was eliminated by treatment with protease, complying with the classic definition of bacteriocins [15]. Enterococci have been well documented for the production of bacteriocins which may give them an environmental advantage against susceptible bacteria [21]. Bacteriocin-producing strains of Enterococcus have been isolated from fermented foods [22], the intestinal flora of mammals including humans [23], and retail raw and processed foods [16].
Although enterococci are part of the natural flora of the mammalian gastrointestinal tract, some strains produce virulence factors, are listed as BSL-2 strains, and can also be opportunistic causes of infection [22]. The presence of enterococci as part of the flora isolated from artisanal food products made from raw ingredients has been the subject of debate on whether they should be considered as food starter cultures [24] or allowed for use as surrogate organisms (that mimic pathogen survival) to define process conditions in food manufacturing facilities [25]. Some strains of Enterococcus have also been characterized for use as probiotics in commercial products, citing the absence of aggregation substance, collagen binding protein, hemolysins, resistance to reactive oxygen, and vancomycin resistance that could be troublesome in opportunistic infections [26]. Although there are successful applications of enterococcal probiotic cultures listed in the literature, FAO/WHO is still reluctant to give universal approval of Enterococcus as probiotic cultures based on potential acquisition and dissemination of antibiotic resistance determinants, notably vancomycin resistance [27].
The ability to produce multiple bacteriocins is a common feature among bacteriocinogenic lactic acid bacteria [21]. Liu et al. [9,28] found two bands of activity when they attempted to purify the bacteriocin produced by En. faecium LM-2, and further DNA sequencing identified both enterocin P and enterocin L-50 related structural genes in strain LM-2. Perez et al. [29] used electrospray ionization-liquid chromatography and mass spectrometry (ESI-LC/MS) to identify and monitor production profiles of four bacteriocins in culture supernatants of En. faecium NKR-5-3 that was isolated from fermented fish. Rehaiem et al. [22] used nine primer sets in PCR assays against En. faecium MMRA and found genes encoding four different enterocins. In this study, an Enterococcus bacteriocin real-time PCR array facilitated the identification of bacteriocin genes in all 22 Bac+ strains tested. The 16 primer sets used in this study (derived from 37 enterococcal bacteriocins) were employed in a bacteriocin PCR array that was successful in identifying not only individual bacteriocin structural genes in all 22 strains, but multiple bacteriocin genes in 15 of the 22 strains tested. DNA sequencing and sequence analysis confirmed that many of the multiple PCR reactions that occurred with any individual strain were due to different bacteriocins within each of the strains (Figure 2) whereby seven strains possessed one bacteriocin gene, five strains possessed two bacteriocins, seven strains possessed three bacteriocins, and three strains demonstrated the presence of as many as four different bacteriocins ( Table 2; Figures 2 and 3). The enterococcal bacteriocin PCR array defined herein is a quick and easy screen in future studies of bacteriocinogenic Enterococcus spp.
The presence of multiple bacteriocin genes suggests a potent antimicrobial mixture, however, it does not necessarily mean that all of them are produced in concert. Perez et al. [29] indicated that some of the multiple bacteriocins in En. faecium NKR-5-3 were not expressed simultaneously and that some regulated the expression of others. Although Liu et al. [9] used lengthy traditional biochemical protein purification techniques to identify multiple bacteriocins, and Perez et al. [29] used more sophisticated mass spectrometry, the identification of multiple enterococcal bacteriocin structural genes in this study was readily determined in a few hours by real-time PCR using an array of primers covering all known enterococcal bacteriocins (Table 2, Figure 4). Enterocin A was the most commonly found bacteriocin structural gene from the 22 Enterococcus isolates. Enterocin A shows sequence homology to class IIa bacteriocins such as the pediocin-like bacteriocins which are characterized for having antilisterial activity [30]. A high frequency of "pediocin-like" bacteriocins with antilisterial activity was expected among the Bac+ strains examined in this study since selection was biased on the use of L. monocytogenes as an indicator strain [31,32]. It is not clear if the spectrum of antilisterial activity demonstrated by these Enterococcus spp. is representative of enterococcal bacteriocins in general, as the distribution of Bac+ isolates was biased by the Listeria indicator used for their selection.
The ability to acquire resistance to LAB bacteriocins has been observed within L. monocytogenes. Naghmouchi et al. [33] observed cross-resistance in L. monocytogenes that was linked to changes in antibiotic sensitivity and cell wall palmitic acid content. Bacteriocin-resistant (Bac R ) variants of L. monocytogenes were also used to identify same/different "modes-of-action" by various bacteriocins [34]. Bacteriocins that remained inhibitory to the Bac R variant were considered to have a different mode-of-action than those which lost activity due to the resistance mutation [34]. Similarly, En. faecium Milk5, JCP B-5, FS97-2, FS56-1, 326F, and En. thailandicus RP-1 were all inhibitory against L. monocytogenes R0 (wild-type) but only JCP B-5, FS56-1, and RP-1 retained activity against L. monocytogenes R1 (Bac R ; derived from R0) ( Figure 4; Table 2). One reasonable explanation is that the elimination of activity by multiple bacteriocins was due to a common mode-of-action, or site-of-action, that was modified by the formation of the Bac R mutation and the activity demonstrated by the remaining Bac+ strains is possibly due their action at a different site, or by a different mechanism ( Figure 4B). Partial sequence analysis shows that En. thailandicus RP-1 possesses the structural gene for entA, En. faecium JCP B-5 possesses that for mr10AB, bacA, and entP, and En. faecium FS56 possesses entA and mr10AB, and all three strains are inhibitory to L. monocytogenes R1 (Bac R ). However, both En. faecium Milk5 and 326F possess most of those genes and others (entA, mr10AB, enxAB, and bac3), yet lost activity against L. monocytogenes R1 (Table 2, Figure 4B). It is not clear what is involved to explain these differences in activity against L. monocytogenes R0 and R1. Different modes-of-action demonstrated by various bacteriocins have previously been shown to result in enhanced bacteriocin activity when mixed and this could be valuable information to use in further studies on the use of bacteriocins as antimicrobial preservatives in food [15,34].
The presence of Enterococcus spp. in numerous RTE food products, the commercial use of Enterococcus as probiotics, and the ability to produce inhibitors of pathogenic bacteria makes enterococcal bacteriocins potential candidates as natural food preservatives. Many of the concerns of Enterococcus as a probiotic is with the use of the live culture itself. However, the use of bacteriocin-containing cell-free culture supernatants as food preservatives may be less disconcerting than the use of viable bacteria. Further studies involving enterococcal bacteriocins, and the prospects of combining bacteriocins having different mode/site-of-action, as food preservatives may provide significant strides in the growing field of natural antimicrobials.

Conclusions
Enterococcus spp. carry many bacteriocin-related genes that provide antimicrobial activity capable of inhibiting competitors in a hypothetical ecological role, but may also be useful in activity against spoilage and pathogenic organisms in select food applications. Many strains encode more than one bacteriocin, although it remains to be seen which ones might be actively expressing bacteriocin proteins. The bacteriocin PCR array allowed quick and easy identification of the presence of bacteriocin-encoding genes and provided coverage for all known Enterococcus bacteriocins. Not only did the 22 strains have activity against L. monocytogenes, but the use of bacteriocin-resistant variants of L. monocytogenes allowed the identification of bacteriocins having different modes-of-action which may be useful in designing mixtures of bacteriocins that have synergistic activity against target organisms.

Author Contributions
Chris Henning is the graduate student who performed the research and wrote the initial draft and had input on experimental design; Dhiraj Gautam is a post-doc who assisted with several experiments and discussions on experimental design; Peter Muriana is the research advisor of Chris Henning, wrote the supporting grant (and grant PI) under which the work was performed, and edited the final version of the submitted paper.