OUP user menu

Characterization of enterocin- and salivaricin-producing lactic acid bacteria from the mammalian gastrointestinal tract

Eileen F. O'Shea, Gillian E. Gardiner, Paula M. O'Connor, Susan Mills, R. Paul Ross, Colin Hill
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01427.x 24-34 First published online: 1 February 2009


Bacteriocin production may be a factor contributing to bacterial dominance within complex microbial populations and may therefore be a common trait within the gut microbiota. However, of 278 antimicrobial-producing culturable lactic acid bacteria (LAB) isolated from a range of mammalian intestinal sources in this study, characterization revealed just 23 distinct strains producing bacteriocin-like inhibitory substances and one Streptococcus hyointestinalis strain producing a potentially novel protease-insensitive antimicrobial. Three class II bacteriocins previously isolated from intestinal-derived LAB were identified as enterocin A and two salivaricin P-like bacteriocins. Moreover, this is the first report of intestinal-derived Streptococcus salivarius producing variants of the lantibiotic salivaricin A.

  • gastrointestinal tract
  • lactic acid bacteria
  • bacteriocin
  • enterocin A
  • salivaricin P
  • salivaricin A


The human gastrointestinal tract (GIT) is estimated to contain up to 1014 microbial cells, and over 1000 different species, outnumbering human somatic and germ cells by a factor of 10 (O'Hara & Shanahan, 2006). Bacteriocin production, a feature common among lactic acid bacteria (LAB) and some Gram-negative bacteria, is thought to confer a competitive advantage on the producing strain, perhaps enabling it to influence or to dominate complex microbial populations (Ryan et al., 1996, 2001). This may occur within the intestinal microbial ecosystem and, indeed, several bacteriocins have been isolated from species resident or even predominant within the gut (Itoh et al., 1995; Twomey et al., 2002). Moreover, bacteriocins have demonstrated potential for use in the control of gastrointestinal pathogens in vivo (Corr et al., 2007). Therefore, screening the gut microbial community may prove a successful approach for the isolation of novel bacteriocins, which may find applications as food biopreservatives or as alternatives to antibiotics. Bacteriocin-producing isolates originating from the GIT may also have additional benefits due to their potential for use as probiotics.

Probiotics are defined as ‘live microorganisms, which when administered in adequate amounts, confer a health benefit on the host’ (FAO/WHO, 2001). LAB are known to produce numerous bacteriocins (De Vuyst & Leroy, 2007). LAB also comprise the largest bacterial group commonly used as probiotics and are considered to play a beneficial role within the mammalian gastrointestinal microbial community (Mombelli & Gismondo, 2000). Interestingly, antimicrobial production has been suggested as a probiotic mechanism because it potentially confers the producing organisms with a competitive advantage over other intestinal microbial communities, perhaps aiding probiotic survival within the GIT (Doron & Gorbach, 2006). While the mechanisms underlying the antibacterial activity of probiotics appear to be multifactorial, and include lowering of pH and production of organic acids and antibacterial compounds such as bacteriocins (Servin, 2004), studies in our laboratory have shown bacteriocin production to be particularly important in vivo. In fact, production of the salivaricin P bacteriocin appears to provide an advantage for a probiotic Lactobacillus salivarius strain, in terms of survival over coadministered probiotics in the porcine ileum (Walsh et al., 2008). Also, production of the bacteriocin, mutacin, by a Streptococcus mutans strain, generated for the prevention of dental caries by replacement therapy, provided a selective advantage, allowing the strain to supersede existing S. mutans populations and persistently colonize the oral cavity (Hillman et al., 2000). Furthermore, one of the most compelling pieces of evidence for the role of bacteriocins in mediating probiotic effects is provided by a recent study that demonstrated the anti-listerial effect of a L. salivarius probiotic in mice was attributed to the Abp-118 bacteriocin produced by the strain (Corr et al., 2007).

Hence, bacteriocin production may be a common trait among gut-derived bacteria. It has therefore been suggested that screening for bacteriocin producers within the GIT could be a suitable approach for probiotic selection (O'Connor et al., 2005). The aim of this study was therefore to investigate bacteriocin production among culturable LAB isolated from a variety of mammalian intestinal sources, with a view to identifying strains with potential for probiotic and other applications in biomedicine or as biopreservatives.

Materials and methods

Bacterial strains and culture conditions

All LAB isolated from mammalian intestinal samples were cultivated under anaerobic conditions in MRS broth (Difco Laboratories, Detroit, MI) at 37 °C for 24 h. Anaerobic conditions were maintained through the use of anaerobic jars together with Anaerocult® A gas packs (Merck, Darmstadt, Germany). Indicator strains used for antimicrobial characterization and their respective growth conditions are listed in Table 1.

View this table:
Table 1

Indicator strains used in this study and their growth conditions

Indicator strainCulture mediumTemperature (°C)Conditions
Bacillus cereus NCIMB 700577BHI37Aerobic
Bacillus subtilis DPC3344BHI37Aerobic
Bifidobacterium animalis DSMZ 20104mMRS37Anaerobic
Bifidobacterium bifidum NCIMB 795mMRS37Anaerobic
Bifidobacterium breve NCIMB 8807mMRS37Anaerobic
Bifidobacterium infantis NCFB 2205mMRS37Anaerobic
Bifidobacterium lactis Bb12mMRS37Anaerobic
Bifidobacterium pseudolongum NCIMB 702244mMRS37Anaerobic
Bifidobacterium psychroaerophilum LMG 21775mMRS37Anaerobic
Bifidobacterium scardovi DSMZ 13734mMRS37Anaerobic
Clostridium difficile ATCC 43593RCM37Anaerobic
Enterobacter sakazakii ATCC 12868BHI37Aerobic
Enterococcus durans LMG 10746MRS37Anaerobic
Enterococcus faecalis LMG 7973MRS37Anaerobic
Enterococcus faecium LMG 11423MRS37Anaerobic
Enterococcus hirae LMG 6399MRS37Anaerobic
Enterococcus malodoratus LMG 10747MRS37Anaerobic
Enterococcus mundtii LMG 10748MRS37Anaerobic
Enterococcus saccharolyticus LMG 11427MRS37Anaerobic
Escherichia coli DPC6054BHI37Anaerobic
Lactobacillus acidophilus LMG 9433MRS37Anaerobic
Lactobacillus agilis LMG 9186MRS37Anaerobic
Lactobacillus amylovorus LMG 9496MRS37Anaerobic
Lactobacillus buchneri LMG 6892MRS37Anaerobic
Lactobacillus casei LMG 6904MRS37Anaerobic
Lactobacillus crispatus LMG 9479MRS37Anaerobic
Lactobacillus delbrueckii ssp. bulgaricus LMG 6901MRS37Anaerobic
Lactobacillus delbrueckii ssp. lactis LMG 7942MRS37Anaerobic
Lactobacillus fermentum LMG 6902MRS37Anaerobic
Lactobacillus gallinarum LMG 9435MRS37Anaerobic
Lactobacillus gasseri LMG 9203MRS37Anaerobic
Lactobacillus johnsonii DSM 10533MRS37Anaerobic
Lactobacillus murinus LMG 14189MRS37Anaerobic
Lactobacillus paracasei LMG 7955MRS37Anaerobic
Lactobacillus paracasei NFBC 338MRS37Anaerobic
Lactobacillus pentosus LMG 10755MRS37Anaerobic
Lactobacillus plantarum LMG 6907MRS37Anaerobic
Lactobacillus reuteri NCIMB 11951MRS37Anaerobic
Lactobacillus rhamnosus GGMRS37Anaerobic
Lactobacillus rhamnosus LMG 6400MRS37Anaerobic
Lactobacillus ruminis DSM 20403MRS37Anaerobic
Lactobacillus sakei LMG 13558MRS37Anaerobic
Lactobacillus salivarius DPC6005MRS37Anaerobic
Lactobacillus salivarius LMG 9477MRS37Anaerobic
Lactobacillus salivarius UCC118MRS37Anaerobic
Lactococcus lactis HPLM1730Aerobic
Leuconostoc mesenteroides CNRZ 1091MRS30Aerobic
Listeria innocua DPC3572BHI37Aerobic
Listeria monocytogenes Scott ABHI37Aerobic
Micrococcus luteus DPC6275BHI30Aerobic
Micrococcus luteus LMG3293BHI30Aerobic
Pediococcus pentosaceus LMG11488MRS30Anaerobic
Salmonella enterica serovar Typhimurium DT104BHI37Aerobic
Staphylococcus aureus DPC5246BHI37Aerobic
Streptococcus algalactiae LMG 14694BHI37Aerobic
Streptococcus bovis ATCC 9809BHI37Aerobic
Streptococcus dysgalactia DPC5345BHI37Aerobic
Streptococcus mutans DPC6152BHI37Aerobic
Streptococcus mutans DPC6155BHI37Aerobic
Streptococcus salivarius DPC6382BHI37Aerobic
Streptococcus thermophilus DPC1780BHI37Aerobic
Streptococcus uberis DPC5344BHI37Aerobic

Detection and isolation of antimicrobial-producing LAB

More than 100 mammalian intestinal samples from a range of sources were screened for antimicrobial-producing bacteria. Samples included human faeces from 19 healthy adults ranging in age from 22 to 50 years (obtained from Moorepark Food Research Centre), 32 neonates (2–12 days old), 26 Clostridium difficile (aged 60+) and five Crohn's disease patients, from local hospitals, Cork, Ireland. Animal samples included nine porcine jejunum and caecum digesta and 25 bovine manure and rumen contents, obtained at Moorepark Dairy Production Research Centre, Teagasc, Ireland.

Samples were homogenized in maximum recovery diluent (MRD; Oxoid Ltd, Basingstroke, Hampshire, UK) as 10-fold dilutions, further diluted in MRD and appropriate dilutions spread-plated in duplicate on MRS agar and the plates were incubated anaerobically at 37 °C. After 48 h, the colonies were enumerated and overlaid with 5 mL soft agar seeded with early stationary-phase cultures of two indicator strains, Listeria innocua DPC3572 (1 × 107 CFU mL−1 BHI agar, Oxoid Ltd) and Lactobacillus bulgaricus LMG 6901 (9 × 106 CFU mL−1 MRS agar), and incubated aerobically and anaerobically, respectively, at 37 °C for 24 h. Colonies surrounded by zones of inhibition were then cultured in MRS broth and stocked at −80 °C for further characterization.

Characterization of antimicrobial activity

The presence of antimicrobial activity in the cell-free supernatants (CFS) of the isolates was determined using the agar well diffusion method (Ryan et al., 1996), using Lactococcus lactis HP, Staphylococcus aureus DPC5246, C. difficile ATCC 43593, L. bulgaricus LMG 6901 and L. innocua DPC3572 as indicators (Table 1). In an effort to eliminate inhibition due to acid production, the activity of the antimicrobials was assessed at a neutral pH. The CFS were adjusted to pH 7.0 using 1 M NaOH and assayed in duplicate against each of the indicator strains. To determine the protease sensitivity of the antimicrobials, CFS from each isolate was mixed with an equal volume of 50 mg mL−1 solutions of each of the enzymes, proteinase K, trypsin, α-chymotrypsin and pepsin (Sigma, Poole, Dorset, UK), and assayed for antimicrobial activity, with the respective CFS mixed with sterile water used as the control. The Streptococcus hyointestinalis DPC6484 and Lactobacillus amylovorus DPC6485 and DPC6486 isolates were similarly assayed with four further enzymes: protease XIII (from Aspergillus saitoi, Sigma), protease XIV (from Streptomyces griseus, Sigma), protease VIII (from Bacillus licheniformis, Sigma) and finally catalase (Sigma). The heat stability of each antimicrobial was determined by assaying the antimicrobial activity of CFS following heating to 100 °C for 10 min. A spectrum of inhibition was performed for each of the antimicrobial-producing isolates using neutralized CFS in agar well diffusion assays using 62 indicator strains (listed in Table 1). All plates were examined for growth inhibition following a 24-h incubation period.

Differentiation and identification of antimicrobial-producing isolates

Molecular fingerprinting of the isolates demonstrating antimicrobial activity from neutralized CFS was performed by pulsed-field gel electrophoresis (PFGE) as described previously (Simpson et al., 2002) using ApaI and SmaI restriction endonucleases and a low-range-molecular-weight DNA marker (9.42–194.0 Kb; New England Biolabs, Beverly, MA). One strain representing each PFGE macro-restriction pattern was selected for further characterization.

Representative isolates were identified by 16S rRNA gene sequencing as described previously (Casey et al., 2004), with sequencing performed by Cogenics (Essex, UK) and sequence data analysed using lasergene 6 software (DNAStar Inc., Madison, WI). Comparison of sequence data was performed using the basic local alignment search tool (blast) on the National Centre for Biotechnology information (NCBI) server (http://www.ncbi.nlm.nih.gov) and the ribosomal database project (http://rdp.cme.msu.edu).

Purification and identification of bacteriocins

Purification of antimicrobial peptides was performed by reverse phase (RP)-HPLC. The bacteriocins were purified from an overnight culture of the producing strain grown at 37 °C in MRS-IM-G broth [tryptone (Oxoid) 10.0 g L−1; yeast extract (Oxoid) 5.0 g L−1; Tween 80 1.0 g L−1; di-potassium hydrogen phosphate 2.6 g L−1; sodium acetate 3H2O (Sigma) 5.0 g L−1; di-ammonium hydrogen citrate (Merck) 2.0 g L−1; MgSO4·7H2O 0.2 g L−1; MnSO4·4H2O (Sigma) 0.05 g L−1; Glucose (BDH) 20.0 g L−1]. Cells were harvested by centrifugation and resuspended in 70% propan-2-ol, adjusted to pH 2.0 using HCl. Following 3.0 h stirring at 4 °C, the cell debris was removed by centrifugation and the propan-2-ol removed from the supernatant. The resulting sample was applied to a 5-g (20 mL volume) C18 Bond Elute column (Phenomenex, Cheshire, UK) pre-equilibrated with methanol and water. After washing with 40% ethanol, the bacteriocin was eluted with 70% propan-2-ol (pH 2.0). After removal of the propan-2-ol from the preparation, a 2-mL aliquot was applied to a Phenomenex C12 RP-HPLC column (Jupiter 4u proteo 90 Å, 250 × 10.0 mm, 4 μm, Phenomenex) previously equilibrated with 30% propan-2-ol, 0.1% trifluoroacetic acid (TFA). The column was subsequently developed in a gradient of 30% propan-2-ol containing 0.1% TFA to 60% propan-2-ol containing 0.1% TFA from 4 to 40 min at a flow rate of 1.5 mL min−1. Bacteriocin activity was monitored throughout the purification procedure by well diffusion assay. MS was performed on bioactive fractions as described previously (Cotter et al., 2006).

Detection of salA variants was achieved by PCR using template DNA isolated from the Streptococcus salivarius isolates and salA specific primers SalAUS and SalADS as described previously (Wescombe et al., 2006). Similarly, salivaricin P and enterocin A production was detected in L. salivarius and Enterococcus faecium isolates, using specific primers (118αF and 118imR; TH8 and TH10, respectively), also described previously (Aymerich et al., 1996; Barrett et al., 2007). Template DNA from Streptococcus bovis and Streptococcus gallolyticus isolates was used for PCR amplification using primers specific for bovicin 255 and bovicin HJ50, as described previously (Cookson et al., 2004; Xiao et al., 2004). Template DNA from L. amylovorus isolates was used for PCR amplification using amylovorin L-specific primers, also described previously (De Vuyst et al., 2004). Sequencing was performed by Cogenics and sequence data analysed using lasergene 6 software. Comparison of sequence data was performed using blast.

Results and discussion

Detection and isolation of antimicrobial-producing LAB

Over 40 000 LAB colonies from a variety of intestinal sources were screened and antimicrobial activity against Listeria spp. and/or Lactobacillus spp. was detected from a total of 278 colonies, as indicated by zones of clearing in deferred antagonism assays. In an effort to select for anti-listerial activity, a nonpathogenic Listeria strain (L. innocua DPC3572) was chosen as an initial indicator. Listeria monocytogenes is a Gram-positive food-borne pathogen responsible for listeriosis, an invasive disease usually occurring in immunocompromised individuals. Although the incidence of listeriosis is relatively low, it has a high mortality rate of 30–40%. The gut flora has demonstrated a protective role against Listeria infection (Gahan & Hill, 2005). A protective effect has also been demonstrated by the administration of a probiotic strain, L. salivarius UCC118, which produces the anti-listerial bacteriocin Abp118, directly mediating protection against Listeria infection in a mouse model (Corr et al., 2007). It is therefore of interest to isolate potential probiotic strains with anti-listerial activity offering a protective effect against listeriosis. Lactobacillus bulgaricus LMG 6901 was also chosen for the initial screening as it was previously shown to be an antimicrobial-sensitive acid-resistant strain (Casey et al., 2004).

Antimicrobial activity was confirmed in the neutralized CFS of just 84 isolates, demonstrating the possible presence of bacteriocins. The low frequency of isolation of bacteriocin-producing LAB may be explained by limitations of culture-based screening procedures. For example, the genetic determinants responsible for bacteriocin production are usually tightly regulated; bacteriocin production therefore often goes undetected by culture-based screening approaches under conditions where the responsible operons are switched off (Kleerebezem, 2004). Also, bacteriocins usually have a narrow spectrum of inhibition and may not be detected from such an extensive microbial community using only two indicator strains for screening, as in this study (De Vuyst et al., 2004). Alternative molecular and genomic approaches have demonstrated the potential to overcome such limitations of culture-based screening techniques with the construction of web-based genome mining tools and metagenomic screening approaches (Williamson et al., 2005; de Jong et al., 2006).

Differentiation and identification of antimicrobial-producing isolates

Genotyping of the 84 antimicrobial-producing isolates revealed 24 individual strains (Fig. 1) indicating a high incidence (71%) of repeated isolation of antimicrobial-producing strains, particularly within individual intestinal samples or from animals in close proximity (data not shown), supporting the competitive advantage theory. A combination of 16 s rRNA gene sequence data and protease sensitivity assays revealed that, of the 24 distinct strains producing active antimicrobial compounds, 23 belonging to nine different species produce bacteriocin-like inhibitory substances (BLIS) (Enterococcus faecalis, E. faecium, L. amylovorus, Lactobacillus gasseri, Lactobacillus reuteri, L. salivarius, S. bovis, S. gallolyticus and S. salivarius), while one porcine caecal isolate, S. hyointestinalis, produces a potentially novel protease-insensitive antimicrobial (Table 2).

Figure 1

PFGE macrorestriction patterns for the restriction enzyme SmaI. This gel illustrates 17 of the 24 individual molecular fingerprints generated when 81 isolates were cleaved with the restriction enzyme SmaI. Each individual pattern and hence strain is represented by a letter subject to identification by 16S rRNA gene sequencing (Table 2).

View this table:
Table 2

Identification of antimicrobial-producing LAB isolated from the mammalian gut and some relevant characteristics

SourcePFGEIsolate no.Radius of zone of inhibition from MRS supernatant pH 7.0 (mm)Protease sensitivity
(SmaI)(Identified by 16s rRNA gene sequence)L. bulgaricusL. lactis HPL. innocua(50 mg mL−1)
Neonate faecesAStreptococcus salivarius DPC648773Proteinase K
Streptococcus salivarius DPC648121Proteinase K
BLactobacillus salivarius DPC64885Proteinase K, α-chymotrypsin, pepsin, trypsin
Enterococcus faecium DPC648226Proteinase K
Enterococcus faecalis DPC64832.5Proteinase K
Patient with Crohn's disease faecesELactobacillus gasseri DPC64893Proteinase K, trypsin, pepsin
C. difficile positive patients faecesLactobacillus gasseri DPC648051Proteinase K
Healthy adults faecesFStreptococcus salivarius DPC64904.5α-Chymotrypsin, pepsin
Bovine faecesJStreptococcus bovis DPC64912.5Proteinase K, α-chymotrypsin, pepsin
KStreptococcus bovis DPC64926Proteinase K, α-chymotrypsin
LStreptococcus bovis DPC649351Proteinase K, α-chymotrypsin
MStreptococcus bovis DPC649451Proteinase K, α-chymotrypsin, pepsin
NStreptococcus bovis DPC649552Proteinase K, α-chymotrypsin
OStreptococcus bovis DPC649642Trypsin
Bovine rumenPStreptococcus bovis DPC649731Trypsin
QStreptococcus bovis DPC649862.5Proteinase K, α-chymotrypsin
ManureCLactobacillus amylovorus DPC64993Proteinase K
DStreptococcus bovis DPC65005.52α-Chymotrypsin, proteinase K
Porcine jejunumGStreptococcus galloyticus DPC65014Proteinase K, α-chymotrypsin, pepsin
HLactobacillus salivarius DPC65025Proteinase K, α-chymotrypsin, trypsin
ILactobacillus reuteri DPC65033Proteinase K, trypsin
Porcine caecumStreptococcus hyointestinalis DPC648474Not sensitive
Lactobacillus amylovorus DPC64854Protease XIII, protease XIV
Lactobacillus amylovorus DPC64864Protease XIII, protease XIV
  • * Mean of triplicate assays.

  • Proteinase K was the only enzyme used to determine protease sensitivity of these isolates.

  • Isolates producing heat-sensitive antimicrobials, all others are tolerant to 100°C for 10 min.

  • –, no activity.

Identification and characterization of antimicrobial compounds

The antimicrobial substances produced by the intestinal isolates were assayed for antagonistic activity against a variety of indicator bacteria typical of the mammalian intestine (listed in Table 1). The protease-insensitive antimicrobial compound produced by the S. hyointestinalis DPC6484 porcine isolate exhibited the broadest spectrum of activity, inhibiting 27 of the 62 indicator strains tested (Table 3). These included seven strains of Bifidobacterium, 15 Lactobacillus strains and at least one strain from each of the following genera: Enterococcus, Streptococcus, Pediococcus, Leuconostoc and Lactococcus. Interestingly, of all the isolates assayed, the porcine caecal isolates, S. hyointestinalis DPC6484 and L. amylovorus DPC6485 and DPC6846, had the broadest and narrowest inhibition spectra, respectively. Such potent activity has not been documented previously for S. hyointestinalis, and further analysis is currently underway to determine the nature of the antimicrobial compound produced by this isolate. A previously characterized amylovorin L-producing isolate L. amylovorus LMG 9434, also of porcine intestinal origin, similarly displayed a narrow inhibition spectrum, inhibiting the same single indicator strain Lactobacillus delbrueckii ssp. bulgaricus LMG 6901 (De Vuyst et al., 2004). Each of the L. amylovorus isolates was immune to the BLIS produced by the other L. amylovorus isolates in this study, as determined by cross-sensitivity assays (data not shown). However, none inhibited the indicator strain L. amylovorus LMG 9496 (Table 3). This strain was also previously shown to produce a BLIS and was found to be antagonistic to the three L. amylovorus isolates of this study (data not shown) as well as the amylovorin L-producing isolate L. amylovorus LMG 9434 (De Vuyst et al., 2004). This suggests that the three L. amylovorus isolates in this study are producing similar or even identical BLIS. Interestingly, differences were evident in the inhibition spectra of L. amylovorus DPC6485 and DPC6846 isolates and the L. amylovorus DPC6499 isolate; this is common among bacteriocins from this species, including those with identical structural genes (De Vuyst et al., 2004).

View this table:

The BLIS produced by L. gasseri DPC6480, L. salivarius DPC6502 and S. gallolyticus DPC6501 also exhibited relatively wide spectra of activity, inhibiting 26, 22 and 21 indicator strains, respectively (Table 3). However, overall antagonistic activity was limited to closely related LAB and bifidobacteria. This suggests a possible role for these antimicrobials in creating a niche for the producing strains within the gut microbial ecosystem, but may not be beneficial when considering potential perturbation of beneficial intestinal LAB populations. Interestingly, anti-listerial activity was exclusive to two neonatal Enterococcus isolates.

Although PFGE results indicated the presence of nine distinct S. bovis strains (Fig. 1), similarities were evident in the spectra of inhibition and the protease sensitivity of the BLIS produced by these isolates (Tables 2 and 3). This suggests the possibility that different S. bovis strains are producing similar or even identical antimicrobial peptides. This was also indicated by cross-sensitivity assays, as each was immune to the antimicrobial peptides produced by all other S. bovis strains (data not shown). A recent study reported the extensive occurrence of horizontal gene transfer between Lactobacillus species within the GIT, potentially influencing up to 40% of core genome genes (Nicolas et al., 2007). This may explain the production of similar antimicrobial peptides by genetically distinct strains in close proximity. Indeed, horizontal gene and plasmid transfer have been suggested as possible mechanisms responsible for the production of similar bacteriocins by genetically distinct S. mutans and L. salivarius strains, respectively (Balakrishnan et al., 2002; Barrett et al., 2007). Isolation of a single bacteriocin-producing species from bovine origin again supports the theory of dominance within complex ecosystems as a consequence of antimicrobial production.

A strain from each intestinal source was selected for further investigation based on the frequency of occurrence or the associated inhibitory activity. These included S. salivarius DPC6481 and DPC6490 isolated from neonates and healthy adults, respectively, the anti-listerial neonatal isolate E. faecium DPC6482, L. salivarius DPC6488 and DPC6502 of human and porcine origin, respectively, L. amylovorus DPC6485 and DPC6486 porcine caecal isolates and manure isolate DPC6499 as well as S. gallolyticus DPC6501 and S. bovis DPC6498, the most frequently isolated strains of porcine and bovine origin, respectively.

MS analysis of the purified antimicrobial peptides produced by S. salivarius DPC6481 and DPC6490 revealed peptides with molecular masses of 2366 Da and 2329 Da, respectively (Fig. 2). These correspond to the masses of SalA2 and SalA5, respectively, variants of salivaricin A, a previously characterized lantibiotic produced by S. salivarius 20P3 (Wescombe et al., 2006). Similarly, a peptide of 4829 Da was purified from the anti-listerial isolate, E. faecium DPC6482, a mass that corresponds to that of the bacteriocin enterocin A (Aymerich et al., 1996).

Figure 2

(a) RP-HPLC profile of the antimicrobial peptide produced by Streptococcus salivarius DPC6481. (b) Matrix-assisted laser desorption/ionization time-of-flight MS data of SalA2 produced by S. salivarius DPC6481. (c) PCR amplification of S. salivarius DPC6481 and DPC6490 template DNA with salA-specific primers. (d) Peptide sequence comparison of SalA and variants, produced by S. salivarius DPC6481 and DPC6490. Amino acid differences are underlined.

PCR, which involved exploiting primers specific for the structural genes of previously characterized bacteriocins, verified the presence of the structural genes for the candidate bacteriocin peptides as suggested by MS analysis. In the case of the S. salivarius isolates, a product of 338 bp was amplified using primers specific for salA, the structural gene of salivaricin A (Fig. 2). Sequencing revealed 100% homology to the respective variants of the salA structural gene that encode SalA2 and SalA5 (Fig. 2), in agreement with the MS data. Salivaricin A and its variants have previously been detected within predominant members of the oral microbial community of humans, S. salivarius and Streptococcus pyogenes (Wescombe et al., 2006). The screening strategy used here, however, selected a species that is not commonly detected within the human intestinal microbial community, S. salivarius, due to the production of the salivaricin A variants. To our knowledge, this is the first time that production of variants of salivaricin A has been reported in intestinal-derived bacteria. Furthermore, a salivaricin A2 producer currently used as an oral probiotic in New Zealand is touted to be effective for the treatment and prevention of streptococcal pharyngitis and halitosis via lantibiotic-mediated influence on the oral microbiota (Tagg, 2004; Burton et al., 2006). It was also recently demonstrated that ingestion of SalA-producing streptococci enhanced the SalA-like inhibitory activity of the commensal oral microbiota, thus improving protection against S. pyogenes infection (Dierksen et al., 2007). Likewise, SalA-producing bacteria from intestinal origin may beneficially influence the gastrointestinal microbiota.

PCR amplification of template DNA from E. faecium DPC6482 similarly confirmed the presence of a structural gene for enterocin A, whose expression was verified by MS detection of the bacteriocin. Likewise, PCR amplification of the template DNA from L. salivarius DPC6488 and DPC6502 isolates revealed the production of two-component class II salivaricin P-like bacteriocins, a common feature of intestinal L. salivarius isolates (Barrett et al., 2007). PCR was used to establish whether the L. amylovorus isolates possess structural genes for the bacteriocin amylovorin L and whether the S. gallolyticus and S. bovis strains isolated in this study possess structural genes for the bacteriocins bovicin HJ50 or bovicin 255, the latter being a particularly common trait among rumen isolates of these species (Cookson et al., 2004). However, their respective structural genes could not be identified by PCR. This suggests that the bacteriocins produced by the S. bovis and S. gallolyticus isolates may be novel. The structural genes for amylovorin L were previously only weakly amplified from L. amylovorus LMG 9434, also of porcine intestinal origin, after extensive PCR; this could not be achieved for the L. amylovorus strains isolated in this study (De Vuyst et al., 2004).

Overall, culture-based antimicrobial screening approaches are limited in their requirement for optimal conditions for growth of producing strains and expression of antimicrobials. Such approaches, therefore, most likely allow for detection of only a fraction of potential bacteriocin-producing microorganisms within complex microbial ecosystems such as the GIT. A genuine reflection of the frequency of antimicrobial-producing LAB present within the mammalian gut microbial community may therefore require more sensitive genomic approaches. Nevertheless, a diverse range of antimicrobial-producing LAB were found among the culturable mammalian gut flora in the present study, albeit detection of antimicrobial production was less frequent than anticipated and bioactivity was limited to Gram-positive targets. Also, predominant species and repeated isolation of the same strains were evident within individual intestinal sources, which may be a consequence of their associated antimicrobial activity or perhaps a bias towards their selective isolation using the specific approach adopted. However, our screening system was sufficient for the isolation of intestinal species for which bacteriocin production is known to be a common feature (e.g. L. salivarius), but more significantly also selected for S. salivarius, a species less frequently detected among the human intestinal microbial community. Findings showing that antagonistic activity was limited to closely related species suggest the possibility that antimicrobial (and in particular bacteriocin) production is primarily a mechanism for survival and prevalence of the producing strain within the intestine rather than a means of pathogen inhibition. This concept perhaps requires further substantiation, but may support the hypothesis that antimicrobial production could aid in the establishment of probiotic bacteria within the GIT, providing evidence that it could be considered a probiotic trait. However, not all the bacteriocin-producing gut-derived strains isolated in this study are necessarily suitable for use as probiotics and each will require rigorous assessment (safety and otherwise) for use in probiotic or other applications.


The authors would like to thank Christine Beecher for technical assistance. E.F.O'S. is in receipt of a Teagasc Walsh Fellowship. This work was funded by the Food Institutional Research Measure of the Department of Agriculture, Fisheries and Food and the Science Foundation of Ireland through a Centre for Science, Engineering and Technology (SCI-CSET) award to the Alimentary Pharmabiotic Centre.


  • Editor: Rustam Aminov


View Abstract