Extensive bacteriocin gene shuffling in the Streptococcus bovis/Streptococcus equinus complex reveals gallocin D with activity against vancomycin resistant enterococci

Streptococcus gallolyticus LL009 produces gallocin D, a narrow spectrum two component bacteriocin with potent activity against vancomycin-resistant enterococci. Gallocin D is distinct from gallocin A, a separate two component bacteriocin produced by S. gallolyticus. Although the gene clusters encoding gallocin A and gallocin D have a high degree of gene synteny, the structural genes are highly variable and appear to have undergone gene shuffling with other streptococcal species. Gallocin D was analysed in laboratory-based experiments. The mature peptides are 3,343 ± 1 Da and 3,019 ± 1 Da and could be readily synthesized and display activity against a vancomycin resistant Enterococcus strain EC300 with a MIC value of 1.56 µM. Importantly, these bacteriocins could contribute to the ability of S. gallolyticus to colonize the colon where they have been associated with colorectal cancer.

With the rise of antibiotic resistant pathogens and the decreasing number of novel antibiotics, the search for alternative antimicrobials is of increasing importance 1 . Bacteriocins are potential antimicrobial candidates and consist of different classes of ribosomally-synthesized antimicrobial peptides which are either narrow or broad spectrum 2 . Narrow spectrum bacteriocins are of particular interest as targeted therapeutics since they could be expected to have minimal impact on resident microbiota 3,4 . Bacteriocin-producing bacteria have been isolated from a range of sources including food, skin, and the gastrointestinal tracts of both animals and humans 5 . Among the functions attributed to bacteriocins are competition, quorum sensing and host signalling 6 . They are classified into multiple types; class I are lantibiotics such as nisin which are subject to post-translational modification, and class II that are unmodified or cyclic peptides. The class II bacteriocins are divided into several subgroups 7 .
Class IIb are two-peptide bacteriocins where the two components are required for maximal activity. The structural genes encoding them are usually adjacently located, together with a gene encoding an immunity protein that protects the cell from being killed by its own bacteriocin 8 . Class IIb bacteriocin operons usually also contain an ABC transporter and an accessory protein. Both peptides are synthesised as pre-peptides with a leader sequence at the N terminal that is cleaved during export at a GG motif to produce the extracellular mature active peptide. This cleavage is performed by the ABC transporter or a peptidase which recognises the leader sequence, and transports the peptide across the cell membrane. Two-component bacteriocins require both peptides for optimal activity and both peptides interact with one another at the same target site to form one antibacterial unit. The mode of action of these bacteriocins involves the binding of the peptides to a target in the cell membrane, leading to pore formation causing leakage and cell death.

Results
Identification of the bacteriocin produced by Streptococcus gallolyticus LL009. Streptococcus gallolyticus LL009 (Sgg LL009) was isolated from raw goat milk sourced in New Zealand; milk samples were stored at − 20 °C until processing. The strain was initially isolated by streaking 10 µl of milk onto Streptococcus thermophilus selective agar that was incubated at 42 °C for 24 h. To test for bacteriocin production, individual isolates were streaked onto BHI agar and incubated overnight at 37 °C, aerobically. When overlaid with Lb. delbrueckii ssp bulgaricus LMG 6,901, zones of inhibition were observed around the colonies of Sgg LL009 (Fig. 1). www.nature.com/scientificreports/ No antimicrobial activity was observed when a cell-free supernatant was tested by well diffusion assay, but a zone of clearing was observed when cell-containing broth was used in this assay. Inhibitory activity was restored to the cell free supernatant when Tween80 was added. We designated this bacteriocin activity as gallocin D, pending further analysis. Sgg LL009 was tested for the production of capsular exopolysaccharide (EPS), a feature of some S. gallolyticus strains. Colonies were positive for the loop touch test and were white on ruthenium red supplemented medium, indicating a ropy type EPS 26,27 . On sucrose supplemented medium, a large amount of mucous type EPS was produced.
Sgg LL009 was not completely resistant to any antibiotic tested (Table 1). It is non-proteolytic in that no zones of clearing were observed surrounding colonies on 10% (w/v) reconstituted skim milk (RSM). A green colouration was observed surrounding cells on blood agar for hemolysis testing, indicative of alpha hemolysis. Alpha hemolysis is defined as bruising of the red blood cells and not true lysis.
In silico analysis of Streptococcus gallolyticus LL009. Following draft genome sequencing of the gallocin D producer Sgg LL009, in silico analysis was performed using BAGEL4 and antiSMASH to identify bacteriocin-associated genes 28,29 . A single bacteriocin operon was identified in Sgg LL009 that consists almost entirely of genes also found in the gallocin A operon of strain S. gallolyticus DSM16831 ( Fig. 2A). The predicted function of each gene is shown in Table 2. However, the bacteriocin structural genes (gllD1 and gllD2) and that encoding the immunity protein (gllDI) do not have homologs in the gallocin A producing S. gallolyticus. These genes are homologous to genes found in the infantaricin ABCDEFG bacteriocin cluster from S. infantarius LP90 (Fig. 2B). The predicted structural genes in the LL009 operon are variants of infantaricin A, where gllD1 shares 98% amino acid identity with infA1, while gllD2 shares 90% amino acid identity with infA2. The predicted molecular weight of gllD2, 3,019.54 Da, is consistent with the 3,021.29 Da mass found by MALDI TOF mass spectrometry analysis (Fig. 1). This mass could not be matched to any known bacteriocin in the bactibase database.
Nine strains of S. gallolyticus and S. infantarius with available genomic data (including Sgg LL009) were analysed for the presence of bacteriocin operons ( Table 3). The ABC transporters are highly conserved in both species, with 97% amino acid identity (Fig. 3). Infantaricin A is encoded by the infantaricin ABCDEFG bacteriocin operon in S. infantarius LP90. Twelve structural bacteriocin genes can be identified in the operon, ten of which make up five two-peptide bacteriocins 11 . There is significant variation in the structural gene arsenal of each strain. Each previously sequenced strain of S. gallolyticus contained the gallocin A structural genes, gllA1and gllA2, while all S. infantarius strains other than CJ18 contained a homolog of infantaricin C. The infantaricin C bacteriocin genes, infC1 and infC2, in S. infantarius LP90 have 66% and 78% identity with gllA1 and gllA2, respectively. One strain of S. gallolyticus ssp macedonicus has gllA2 next to a gene with 60% identity to gllA1. Sgg LL009 is the only S. gallolyticus strain that lacks a homolog for gallocin A, and instead harbours a variant of infantaricin A, here termed gallocin D.

Synthesis and spectrum of inhibition of infantaricin A and gallocin D.
The gallocin D1 prepeptide encoded by gllD1 is 51 amino acids and the D2 pre-peptide encoded by gllD2 is 52 amino acids in length, while the active peptides are predicted to be 30 and 29 amino acids, respectively. We synthesised both D1 and D2 peptides for further characterisation, the mass of the synthesized D2 peptide matched the observed mass of 3,021 Da. The D2 peptide displays no activity in a well diffusion assay against Lb. delbrueckii ssp bulgaricus, while www.nature.com/scientificreports/ the D1 peptide has solo activity, which is enhanced when in a 1:1 combination with D2. When tested against pathogenic bacteria, both peptides are required for activity. The infantaricin A peptides were also synthesized and it was found that the peptides cross-complemented each other, A1/D2 or D1/A2 (Fig. 4). A number of indicator organisms were tested to determine the spectrum of activity, using both the overlay method on a plate and a well diffusion assay using the synthesized peptide ( Table 4). The bacteriocin was active against a narrow range of indicator organisms that included the clinically important pathogens S. pneumoniae and vancomycin resistant enterococci (VRE) strains but was inactive against other unrelated pathogenic and commensal bacteria. The minimum inhibitory concentration (MIC) of the gallocin D peptide was assessed at 1.56 µM against VRE EC300.
Streptococcus gallolyticus DPC6501 was isolated from a porcine jejunum in a previous study and is reported to produce a bacteriocin 30 . This strain was tested for sensitivity to Sgg LL009 and the synthesized gallocin D peptides. The overlay assay of Sgg LL009 showed zones of inhibition against S. gallolyticus DPC6501 and a zone of inhibition from 15 µM gallocin D peptides was observed in a well diffusion assay (Fig. 5).
Resistance development in VRE EC300 against gallocin D. The microtitre plates from the MIC determination experiment were plated in an attempt to identify growth of VRE EC300. Following 24 h growth, VRE EC300 cells from the control and wells containing 100 µM gallocin D, the highest concentration included, were spread plated on BHI, and then sub-cultured into fresh BHI. At 24 h, the sub-cultured broths were serially diluted and plated. The bacteriocin treated cells showed no growth at any dilution, while the untreated cells reached 4.9 × 10 8 cfu ml −1 . At this concentration and inoculum, no colonies were found following treatment with gallocin D.
For the time-kill assay actively growing VRE EC300 at 10 8 cfu ml −1 was treated with gallocin D at 15.6 µM, 10 × the MIC value. Within the first 4 h, VRE EC300 was undetectable in all treated broths, while the untreated control remained at 10 8 cfu ml −1 . At 6 h, VRE EC300 was detected in treated broths while at 24 h the treated broths were lower than the untreated control at 10 7 cfu ml −1 (Fig. 6).

Discussion
The bacteriocin operons within the S. gallolyticus subspecies gallolyticus of the SBSEC were found to be generally highly conserved, with the exception of the newly isolated Sgg LL009 which lacks the gallocin A structural genes and has different structural genes and a putative immunity gene at this locus. The structural genes in Sgg LL009 are variants of genes in an infantaricin A producing S. infantarius LP90 (Sii LP90), and its associated immunity gene. Infantaricin A is encoded by an operon with seven predicted bacteriocin structural genes, of which infC shows sequence homology to the gallocin A structural genes. Sgg LL009 is the only genome analysed found to lack genes for gallocin A/infantaricin C, and contain a variant of previously identified infantaricin A. In a BLAST www.nature.com/scientificreports/ search limited to the SBSEC, no gallocin D hits were found outside of the S. gallolyticus or S. infantarius species and only one potential gallocin A homolog was identified in S. macedonicus. Gallocin A is a two-peptide bacteriocin which has been reported to give S. gallolyticus a competitive advantage in conditions found in in the gut of patients with CRC. This bacteriocin is absent from closely related species, but we identified possible homologs in S. infantarius strains, and gllA2 was found in S. gallolyticus ssp macedonicus together with a gene with 60% identity to gllA1. High genome plasticity has been reported in the SBSEC and S. gallolyticus is reported to have retained the largest genome and highest functional capacity. Two human isolates of S. infantarius, NCTC 13,760 and ATCC BAA-102 possess the infantaricin ABCDEFG bacteriocin operon while the dairy isolate Sii CJ18 does not have the complete operon. Sgg LL009 and Sii LP90 were isolated from goat and water buffalo raw milk samples, respectively. In previous comparative genomics studies, high numbers of IS elements were found in these species. The organisation of the operon and high sequence identity between Table 2. In silico analysis of the predicted function of each gene involved in bacteriocin production.    www.nature.com/scientificreports/ genes suggests that transfer of bacteriocin structural genes occurred between these strains due to adaptive pressure in the microbe-rich environment of the gut 21 . Sgg LL009 has unique structural genes which are produced within the gallocin D operon. The association between S. gallolyticus and CRC is not fully understood 31 , but it is thought that the conditions in the colon when tumours are present provide a suitable niche for S. gallolyticus if competitors can be controlled with gallocin A. Gallocin A and gallocin D share similar characteristics and target organisms. Gallocin A production is enhanced in the presence of secondary bile acids, a known risk factor of CRC. The cell free supernatant of S. gallolyticus UCN34 shows no activity in a well diffusion assay in the absence of a detergent or secondary bile acids, a feature that is also observed for Sgg LL009 producing gallocin D 10 . S. gallolyticus mutants lacking gllA1 and gllA2 do not have the same colonisation advantage in tumour-bearing mice, leading to the conclusion that S. gallolyticus is not a causative factor of CRC but does promote its acceleration if pre-malignant tumours are present and the strain colonises. Both gallocin A and gallocin D target enterococci, with enhanced bacteriocin production in CRC conditions. This suggests that Sgg LL009 could well retain the colonisation advantage seen in S. gallolyticus UCN34 in CRC conditions, despite their distinct amino acid sequences. Table 4. Indicator organisms with growth conditions. Inhibition is represented by relative activities + 0.5-2 mm zone, ++ 2-4 mm zone, +++ > 4 mm zone + , and no inhibition −.  The operons in the various S. gallolyticus strains were strikingly similar with the exception of the bacteriocin structural genes, suggesting that these bacteriocins would be subject to similar regulation and would be produced under similar conditions. Whether this strain would have a protective effect in the colon is unclear. It is likely, due to their similar target organisms and production characteristics, that this strain of S. gallolyticus producing gallocin D would occupy the same niche as S. gallolyticus producing gallocin A. Further studies are required to assess if this gallocin D producing S. gallolyticus is able to colonise and accelerate cancer development.

Indicator organism Growth medium Incubation conditions Inhibition
This work confirms horizontal gene transfer of bacteriocin structural genes between members of the SBSEC, which have been shown to have high genome plasticity 18 . Sgg LL009 and Sii CJ18 are both dairy isolates and both show the most "unusual" operons; the structural and immunity genes of Sgg LL009, and Sii CJ18 has the lowest number of bacteriocin structural genes of all the Sii strains. This hints at a role for these bacteriocins, gallocin D and infantaricin A, in streptococcal strains colonising ruminant animals. Further identification of bacteriocin producers from the SBSEC could lead to further evidence of shuffling of bacteriocin structural genes.
A further strain of bacteriocin producing S. gallolyticus from our culture collection, S. gallolyticus DPC6501, is sensitive to gallocin D, which strongly indicates that the genomic annotation of the immunity protein is correct. This is the only predicted immunity protein present in Sgg LL009 and absent in related S. gallolyticus strains and also the only immunity protein shared between Sgg LL009 and the infantaricin A producer. Sgg LL009 is not inhibited by Sgg DPC6501, this strain has not been sequenced, but is known to be a bacteriocin producer from previous work 30 . We hypothesise that Sgg LL009 is not susceptible to gallocin A, but that synthesized gallocin D could provide an alternative strategy for the control of S. gallolyticus infections.
The application of gallocin D is not limited to its potential role in controlling the growth of other S. gallolyticus strains in CRC. It is a narrow spectrum bacteriocin with potent activity against VRE, opportunistic pathogens that are particularly relevant in hospital settings, for patients who are immunocompromised and those under antibiotic treatment for endocarditis 32 . VRE can infect the urinary tract, surgical wounds or the bloodstream and are spread by direct contact. Many patients who develop VRE infections have underlying illnesses and due to antibiotic resistance this infection can lead to serious problems or fatalities 33 . The kill curve of gallocin D shows it can reduce the numbers of VRE from 10 8 cfu ml −1 to undetectable levels in two hours, though the cells regrow after 8 h. Importantly, when added to growing cells at 10 5 cfu ml −1 , which is regarded as a clinical infection, the VRE EC300 did not recover and no resistant colonies were found. This suggests that the mode of action of this bacteriocin is related to cell contact; if the bacteriocin is present in the right ratio to cells the infection can be cleared. The MIC and resistance testing results for gallocin D show that it completely kills VRE EC300 when present at 10 5 cfu ml −1 , and gallocin D is present at the MIC of 1.56 µM.

conclusions
A combination of laboratory and in silico analyses led to the discovery and characterisation of gallocin D, a class IIb bacteriocin with activity against VRE, S. pneumoniae, and a related strain of S. gallolyticus. Gallocin D is a variant of infantaricin A produced by closely-related S. infantarius species. The operon is similar to those found in other S. gallolyticus genomes, with the exception of the structural and immunity genes. This work highlights the shuffling of bacteriocin structural genes within these closely related species. Gallocin D could be applied for treatment of VRE infections, and potentially for the control of other species of S. gallolyticus.

Materials and methods
Bacterial strains culture conditions. S. gallolyticus LL009 was isolated from raw goat milk produced in New Zealand on S. thermophilus agar at 42 °C. S. gallolyticus LL009 was routinely cultured under aerobic conditions at 37 °C, in brain heart infusion (BHI) medium (Oxoid Ltd., Basingstoke, Hampshire, United Kingdom), all other reagents were sourced from Sigma-Aldrich (Wicklow, Ireland) unless otherwise stated. Indicator strains used and their incubation conditions are listed in Table 3. prepared for sequencing using a Nextera XT kit (Illumina) for library preparation. DNA was quantified using a Qubit 2.0 fluorometer. Sequencing was carried out using an Illumina MiSeq platform with paired-end 2 × 300 base pair reads at the Teagasc Sequencing Centre, Teagasc Food Research Centre Moorepark. Assembly was performed de novo using SPADES and automatically annotating using GAMOLA 34 . The whole genome sequence was run through Bagel4 and antiSMASH to search for bacteriocin genes. The genome was also compared to other available S. gallolyticus and S. infantarius genomes from NCBI listed in Table 4.
Antimicrobial activity assays. Sgg LL009 was grown on BHI agar and incubated overnight at 37 °C. Sgg LL009 was assayed against various indicator organisms, listed in Table 4 along with incubation conditions. Plates were overlaid with MRS agar (7.5 g L −1 agar) seeded with L. bulgaricus LMG6901 and incubated overnight at 37 °C, anaerobically. Activity was defined by a clearing in the overlay medium.
Overnight cultures were centrifuged and filtered to obtain cell free supernatant (CFS). MRS agar was seeded with an overnight culture of L. bulgaricus LMG6901 and wells were made. Both broth and CFS were added to the wells in 50 µl volumes. Plates were incubated overnight at 37 °C. Triton X-100 was added to CFS at 1% concentrations. All assays were performed in triplicate. The well diffusion method was repeated using synthesized gallocin D, using peptides alone and in combination (1:1).

Minimum inhibitory concentration. Minimum inhibitory concentration (MIC) determinations of gal-
locin D (D1D2) were carried out in triplicate in microtitre plates as previously described 35 . Briefly, VRE EC300 was grown overnight in BHI broth at 37 °C and subcultured at 0.5% into fresh broth. The strain was grown to OD600 of 0.5 and diluted to a final concentration of 10 5 cfu ml −1 in a final volume of 100 µl. Synthesized gallocin D peptides in a 1:1 ratio were made up to 100 µM concentration and serially diluted to a concentration of 0.98 µM in sterile water. The peptide solutions were added at 100 µl volumes to the VRE broth. The plate was incubated at 37 °C for 16 h, MIC was determined as the lowest concentration causing visible inhibition of growth.
Resistance testing. In order to assess resistance development by VRE EC300, the MIC protocol was repeated. The microtitre wells containing 100 µM gallocin D and control wells were plated directly and cultured in fresh BHI broth. At 24 h, the sub-cultured broths were serially diluted and plated. The whole microtitre plate was reincubated for a further 48 h and observed for cloudiness indicative of growth.
Kill curve. VRE EC300 was grown overnight in BHI and adjusted in fresh BHI to a final concentration of 10 8 cfu ml −1 . Synthesized gallocin D was added at concentration of 10 × MIC (15.6 µM) to triplicate testing broths, and sterile water added to control VRE broths. Samples were taken at multiple time points until 8 h and again at 24 h, serial diluted in MRD and plated on BHI medium. Plates were incubated for 24 h at 37 °C before colonies were enumerated.
Colony mass spectrometry. Colony MALDI-TOF MS (Axima TOF 2 MALDI-TOF mass spectrometer, Shimadzu Biotech) was used to determine the molecular mass of the peptides present on the surface of colonies as follows: cells were first mixed in 70% (v/v) 2-propanol/0.1% TFA (IPA) and vortex mixed, the sample was separated by centrifugation at 14,000 r.p.m and the supernatant was subsequently used for analysis. A MALDI target plate was precoated with CHCA matrix solution, 0.5 µL of the supernatant from the cell extract was then placed on the target and a final layer of matrix solution was added. Positive-ion linear mode was used to identify the peptide masses on an Axima TOF 2 MALDI TOF mass spectrometer (Shimadzu Biotech, Manchester, UK). The masses detected were then compared to those of known bacteriocins.
Growth curve. A single colony of VRE was inoculated into BHI medium and incubated at 37 °C for 16 h, and subcultured at 1% into fresh medium. Samples were taken at 0, 2, 4, 5, 6, 7, 8 and24 hours for OD 600 and plating. OD was read in duplicate, 100 µl was serially diluted in MRD to 10 -8 and plated on BHI agar. Plates were incubated for 24 h and enumerated. Experiments were completed in triplicate, each with technical duplicates. Antibiotic resistance. The MIC value of sixteen antibiotics was assessed using the VetMIC Lact-1 and Lact-2 MIC determination plates (National Veterinary Institute, Sweden). The antibiotics tested were ampicillin, penicillin, vancomycin, erythromycin, virginiamycin, tetracycline, clindamycin, chloramphenicol, kanamycin, gentamycin, streptomycin, neomycin, linezolid, rifampicin, ciprofloxacin, and trimethoprim. Briefly, colonies were resuspended in MRD at a concentration of ~ 1 × 10 8 cfu ml −1 and transferred into ISO-MRS broth for a final inoculum of 5 × 10 5 cfu ml −1 . VetMIC plates were inoculated with 100 µl, sealed and incubated for 24 h. Exopolysaccharide screening. Multiple screening methods were used to test for ropy and non-ropy type EPS; ruthenium red agar, the loop touch test and sucrose-supplemented MRS. Ruthenium red was filter sterilised, added to cooling MRS agar at 0.08% and mixed before pouring plates27. For the loop touch test, a sterile loop was touched to a single colony and slowly pulled away. A string between the loop and colony was recorded as a positive result. 10% (w/v) sucrose and 10% (w/v) lactose supplemented MRS plates were autoclaved at 121 °C for 15 min and poured, a mucous phenotype was characterised as a positive result.
Proteolysis. 10% reconstituted skim milk (RSM) was autoclaved at 121 °C for 5 min and combined with a 3% (w/v) agar solution, autoclaved at 121 °C for 15 min, the solutions were allowed to cool to ~ 45 °C and combined 1:1, and poured into petri dishes. Previously grown Sgg LL009 was re-streaked onto the RSM plates and incubated at 37 °C for 48 h, plates were viewed every 24 h. Proteolysis was defined as clear zones surrounding colonies.