Streptococcus rattus strain BHT is a species representative and strong bacteriocin producer. Here we report that S. rattus BHT produces two quite different types of bacteriocin activity, named BHT-A and BHT-B. The two bacteriocins were purified and analysed for activity and by MALDI-TOF mass spectrophotometry. BHT-A was found to be a variant of the two-component lantibiotic, Smb. BHT-B is a non-modified 5195 Da peptide with some similarity to the tryptophan-rich Staphylococcus aureus bacteriocin, aureocin A53. Six S. rattus and two S. mutans strains were found to contain both the BHT-A and BHT-B genetic loci.
The bacteriocins are proteinaceous antibacterials commonly produced by all types of bacteria. In recent years the lactic acid bacteria have been especially studied for their bacteriocinogenicity and a number of practical applications ranging from food preservation to probiotics have been touted for some inhibitory strains [1,2]. The mutans streptococci, renowned for their etiological role in dental caries, are prolific producers of bacteriocins, collectively referred to as mutacins . Currently characterised mutacins include members of bacteriocin classes I (lantibiotics) and II (small, heat-stable non-lanthionine-containing peptides) .
Caufield and associates characterised the lantibiotic mutacins I, II and III and the di-peptide non-lantibiotic mutacin IV [5–9]. In this laboratory, bacteriocin-producing streptococci are initially categorized by assessing their inhibitory activity against nine indicator bacteria using bacteriocin P (production) typing . Strains inhibiting all nine indicators are given the code designation P-type 777. Crooks et al. found that 3% of freshly isolated S. mutans displayed P-type 777 activity . Balakrishnan classified 16 P-type 777 S. mutans strains into four groups (A–D) largely on the basis of their cross-immunity and activity spectrum profiles . The mutacins produced by his groups A and B equate to Caufield's mutacin types I and II, respectively.
Streptococcus rattus strain BHT, a member of the mutans streptococci, also has a 777 P-type, but its extended inhibitory profile differs from that of S. mutans groups A–D . S. rattus strain BHT was isolated by Zinner from a human source and has been widely studied as a species representative . Like other mutans streptococci S. rattus principally colonises the tooth surface, but its small numbers in plaque make it unlikely to be a major contributor to dental caries. The pioneering studies of Kelstrup and Gibbons first showed strain BHT to be bacteriocin positive and interestingly that it had an inhibitory spectrum identical to that of the rat isolate, S. rattus strain FA1 . A number of research groups have investigated the bacteriocin activity of strain BHT, but none has reported purification of the active inhibitory agent(s) [3,16–18]. Delisle showed that high concentrations of yeast extract (4%) promoted bacteriocin production by both strain BHT and S. mutans GS5 in broth cultures . Several groups also tried to establish a link to plasmid carriage by generating bacteriocin-negative mutants but no direct proof of plasmid linkage was obtained [19,20].
In the present study we report that strain BHT (like several other strains of S. rattus) has two distinct bacteriocin loci, one encoding a two-component class I lantibiotic variant of the recently described Smb from S. mutans GS5 and the other encoding a tryptophan-rich peptide having some of the characteristics of class II bacteriocins.
Materials and methods
Bacterial strains and media
The bacterial strains were from the following sources: A.H. Rogers, University of Adelaide, Australia (S. rattus BHT, S. sobrinus OIHI); M.C. Lavoie, Universite Laval, Canada (S. rattus LG1, IB, 67–3 and GF71) and S. Hamada, The National Institute of Health, Tokyo, Japan (S. rattus FAI, S. mutans OMZ175, S. sobrinus B13) and G. Colman, Central Public Health Laboratory, London, UK (S. mutans strain NCTC10449). S. mutans strain K34–1 was isolated from a plaque specimen obtained from an adult human. The nine standard indicator strains (I1–I9) used in this laboratory for discrimination between streptococcal bacteriocins have been described previously . Indicator strains I1 (Micrococcus luteus), I6 (Lactococcus lactis), I7 (S. pyogenes) and I9 (S. equisimilis) were also used in comparative assays of bacteriocin activity.
Stock cultures were stored in skim milk at −70 °C. Strains in regular use were maintained as plate cultures and sub-cultured every two weeks on Columbia Agar Base (Difco) supplemented with 5% whole human blood and 0.1% calcium carbonate (BaCa). Inhibitory activity was assessed by deferred antagonism, either on BaCa or trypticase soy yeast extract calcium carbonate agar (TSYCa; trypticase soy broth (Baltimore Biological Laboratories, Becton Dickinson and Company, USA) +2% yeast extract (Difco Laboratories, Detroit, MI, USA) +1.5% Davis agar (Davis Gelatine Ltd, Christchurch, New Zealand) +0.3% calcium carbonate. Soft TSYCa agar (containing only 0.7% Davis agar) was used for bacteriocin purification.
Assay of bacteriocin activity
Inhibitory activity was determined by end-point titration using a surface spot method (SSM). All titrations were performed in duplicate. Briefly, 20 μl samples of serial twofold saline dilutions of all preparations were spotted onto the surface of BaCa and allowed to dry. The agar surface was then sterilized by exposure to chloroform vapours for 30 min followed by airing. The surface was then swabbed evenly from a Todd Hewitt broth (THB) (Difco) culture of the indicator strain. Following incubation, the titre of bacteriocin in arbitrary units (AU) per ml was taken to be the reciprocal of the highest dilution to show definite inhibitory activity. All incubations of cultures were at 37 °C for 18 h in air plus 5% CO2.
The procedures for bacteriocin isolation and purification were identical to those previously used in this laboratory for purification of mutacin N . Successive 0.5 ml fractions from FPLC (Pharmacia) reversed phase chromatography were assessed against standard indicators I1 and I9. Two distinct sets of fractions containing inhibitory activity eluted at ca. 25% and 41% AcN and these were pooled respectively as partially purified bacteriocin BHT-A (PBHT-A) and BHT-B (PBHT-B).
PBHT-A was lyophilised, re-dissolved in 4 ml of 20 mM sodium acetate buffer (pH 4) and fractionated by use of FPLC cation exchange chromatography (MonoS HR 5/5 cation exchange column, Pharmacia). Absorbance was monitored at 214 nm. Active fractions were pooled, lyophilised and re-dissolved in 12 ml buffer A (0.1% trifluoroacetic acid). Four millilitres of this material was fractionated by FPLC reversed phase chromatography and this was called BHT-Aa. 2.5 ml of this was desalted by FPLC using a Superose 12 gel filtration column (Pharmacia) equilibrated in 50 mM ammonium acetate buffer (pH 6) plus 30% AcN. Five hundred microlitres of this material was fractionated by FPLC reversed phase chromatography as described above and this was called BHT-Ab.
PBHT-B material was lyophilised and redissolved in 3 ml buffer A. This preparation was fractionated by HPLC reversed phase chromatography (C18, 10 μm, 250 × 46 mm (Allteck) using the buffers described for FPLC reversed phase chromatography above. Absorbance was monitored at 214 nm. The active fraction was called BHT-B.
Analysis of the purified peptides
Both BHT-Aα and BHT-Aβ were analysed for molecular mass and BHT-Ab was used for N-terminal amino acid sequencing. BHT-B was similarly analysed for molecular mass and for N-terminal amino acid sequencing. The molecular mass was determined using MALDI-TOF mass spectrometry (Lasermat 2000, Finnigan Mat, Hemel Hempstead, England). N-terminal amino acid sequencing (Edman degradation) was done using a pulsed-liquid phase protein sequencer (Procise 492, Perkin–Elmer/Applied Biosystem, Foster City, CA, USA).
Activity of purified BHT-A and BHT-B
PBHT-A and PBHT-B were assayed for activity against indicator strains I1, I6, I7 and I9 (see above), S. mutans strains 10449 and OMZ175, S. sobrinus strains B13 and OIHI and S. rattus strain BHT using SSM. PBHT-A and PBHT-B were tested for pH stability by mixing 100 μl with 100 μl of 20 mM sodium acetate buffer (pH 2.0, 4.0 or 6.0) or 20 mM Tris buffer (pH 8.0, 10.0 or 12.0), incubating at room temperature for 24 h and determining residual activity by SSM. The heat stabilities of PBHT-A and PBHT-B were tested by heating in a water bath at 95 °C for 30 min and determining residual activity by SSM.
Identification of the BHT-B gene locus
Degenerate primers (Bht-B FP 5′-GGIACHAARGCWGTTCAATGGGC-3′ and Bht-B RP 5′-YAAWGCYAAAATWCGWCCCCACAT-3′) to the amino acid sequence obtained for BHT-B were used for inverse PCR using the Expand Long Template PCR system (Roche). The PCR program used had an extension time of 6 min, an annealing temperature of 55 °C and 35 cycles. Primer concentrations used were 2.4 and 6.4 μM for Bht-B FP and Bht-B RP, respectively. The product was gel purified, digested with Sau3AI and cloned into pBluescript SK II (+) digested with BamHI. Both direct and inverse PCR were then used to sequence the genetic locus surrounding the BHT-B gene. Primers used to screen for the BHT-B gene were BHT-B US 5′-AGAAAGCACCTCCGCCGCCACT-3′ and BHT-B DS 5′-AGTCTTACCAGCACCATTTTCACCTA-3′.
Identification of a two-component lantibiotic operon in strain BHT
The primers lanM Fwd (5′-TTGCWAGWYWTGCWCATGG-3′) and lanM Rev (5′-CCTAATGAACCRTRRYAYCA-3′) were used to amplify a region of the M gene from lantibiotic operons. The program used for the PCR had an annealing temperature of 40 °C, a 1 min extension time and 35 cycles. Resultant PCR products were then cloned and sequenced. Inverse PCR and primer walking were then used to sequence the entire genetic locus surrounding the identified lanM genes. Primers used to screen for the BHT-B gene were BHT-A US 5′-GGGTAGCTGGGTGGGACAAAGGAGAAG-3′ and BHT-A DS 5′-TCCTACTAACCACATTGACTGACTGAT-3′.
GenBank Accession Nos. for sequences described: BHT-A genetic locus-DQ145752, BHT-B genetic locus-DQ145753.
Results and discussion
Purification and characterisation of mutacins BHT-A and BHT-B
Analysis of active, purified BHT-A using MALDI-TOF mass spectrophotometry identified two peptides of 2802 and 3375 Da (Fig. 1(a)). Further attempts at separation of these two peptides by reversed phase chromatography resulted in loss of the 2802 Da peak. Some weak bacteriocin activity against indicator I9 was detected in fractions containing the purified 3375 Da peptide. However, since isolation of both pure peptides was not achieved no analysis of their possible synergistic activity could be performed. N-terminal sequencing was carried out on the purified 3375 Da peptide and this was found to start isoleucine glycine (Fig. 1(c)).
MALDI-TOF mass spectrophotometry of purified BHT-B identified a peak of 5195 Da corresponding to inhibitory activity (Fig. 1(b)). Edman sequencing of this peptide, although yielding a clean call for all (except residue 21) of the first 28 amino acids (Fig. 1(c)), was notable for its extremely low yield, a finding consistent with the N-terminus of the peptide being at least partially blocked. Since the peptide sequence for BHT-B starts with a methionine, it could mean that the biologically active peptide is completely unmodified resulting in partial blockage of sequencing due to the presence of formyl-methionine at the N-terminus of the peptide.
The two purified mutacins were analysed for stability and for their inhibitory activity against various indicators (Fig. 1(d)). Both BHT mutacins were stable when heated to 95 °C for 30 min and tolerated 24 h exposure to pH in the range 2–12. Mutacin BHT-A was active to varying degrees against all of the indicators tested and especially S. pyogenes strain I7 and S. equisimilis strain I9, while mutacin BHT-B was principally active against the non-mutans Streptococcus indicators tested. Testing of mixed preparations of mutacins BHT-A and BHT-B showed that there was no apparent synergistic activity between these inhibitors (data not shown). The combined activity spectra of the two different bacteriocins account for all known inhibitory activity of S. rattus BHT.
Identification and analysis of the BHT-B gene and locus
The N-terminal sequence of mutacin BHT-B was used to design degenerate primers for inverse PCR. This gave a single, weak band of approximately 4.5 kbp. Sequencing of Sau3AI sub-clones generated from this band identified a putative ABC-type transporter and oligopeptidase genes. Specific primers designed to these sequences were used for both direct and inverse PCR, which led to the identification of a small open reading frame (ORF), which started with the 28 amino acid sequence obtained for the BHT-B peptide. The sequence of the mutacin BHT-B ORF encodes a 44 amino acid peptide with a predicted mass of 5164 Da (Fig. 2(a)), 31 Da smaller than the mass obtained by MALDI-TOF mass spectrophotometry. The initiation methionine is still present in the active peptide and the formylation of this to give N-formylmethionine accounts very well for this size discrepancy. Mutacin BHT-B is apparently unmodified and amphipathic in nature, with 23 hydrophobic residues and five tryptophans (Fig. 2(a)). Similarity searching identified some similarity to aureocin A53, an amphipathic, tryptophan-rich, membrane-active peptide isolated from Staphylococcus aureus strain A53 (GenBank Accession No.: AAN71834) [22,23]. Unusually, as was shown for aureocin A53, mutacin BHT-B is expressed and exported with no leader sequence or signal peptide. Another gene with similarity to aureocin A53 has also been identified from Corynebacterium jeikeium (GenBank: YP_220784) . An alignment of these three genes is presented in Fig. 2(b). Significant similarity occurs through the N-terminal half of the peptides while the C-terminus of BHT-B shows little similarity to the same regions of the other two peptide sequences. These three peptides have a very similar predicted secondary structure and we anticipate that mutacin BHT-B acts through a similar membrane disruptive mechanism to that of aureocin A53 . Both aureocin A53 and the C. jeikeium AucA gene are harboured on plasmids. We found no evidence of a plasmid in S. rattus BHT using standard and pulsed-field gel electrophoresis (data not shown).
The locus surrounding the mutacin bht-b gene was sequenced and is illustrated in Fig. 2(c). Upstream of the bht-b gene is an ORF with similarity to the receptor component of the mutacin I receptor/kinase genes and an operon encoding proteins predicted to be involved in sugar metabolism. Downstream of the bht-b gene are two ORFs with motifs common to ABC transporter proteins and three ORFs with no similarity to any database sequences or known functional motifs. These genes may encode proteins involved in either the transport of or immunity to mutacin BHT-B. The oligopeptidase identified from the initial inverse PCR product is located downstream of the second putative ABC transporter protein.
Identification of a two-component lantibiotic operon in strain BHT
Two putative lanM genes were identified in S. rattus BHT using degenerate primers. Direct and inverse PCR were used to sequence the loci surrounding these genes. They were found to be part of the same operon encoding a two-component lantibiotic (Fig. 3(a)). Overall the operon was very similar to the recently published smb two-component lantibiotic-encoding operon from S. mutans strain GS5 , with the two operons sharing 95% identity at a DNA level. Analysis of the structural gene sequences (Fig. 3(b)) and alignment with those from other two-component lantibiotic structural genes (Fig. 3(c)) allowed prediction of probable cleavage sites. The predicted sizes of 3451 and 2893 Da strongly correlated with the two sizes identified in the purified BHT-A preparation (Fig. 1(a)) when accounting for the dehydration of some serine and/or threonine residues involved in lantibiotic formation. A point mutation in the BHT-Aα gene has resulted in a C-terminus differing from that of the smb-α peptide (Fig. 3(b)), which could result in a slightly altered inhibitory spectrum. The difficulty in obtaining quality sequence from the purified BHT-A preparations is likely due to the presence of serine and threonine residues, which are often dehydrated in lantibiotic peptides, resulting in blocked N-terminal sequencing.
At a structural level the BHT-Aα peptide appears to retain the amino acids required for the more globular, mersacidin-like, ring structure also found in other two-component lantibiotics, such as lacticin 3147 α, at its C-terminus, while the BHT-Aβ peptide more closely resembles the lacticin 3147 β peptide, which has a more linear structure . These structural features are critical to the synergistic mode of action of the two lacticin 3147 peptides. The lacticin 3147 α peptide, like the BHT-Aα peptide, has inhibitory activity by itself. This, given the structural similarity to mersacidin, is thought to be due to its ability to bind lipid-2 . Morgan et al. demonstrated that the synergistic activity of the lacticin 3147 α and β peptides required the addition of first the α peptide, then the β peptide . The proposed mode of action is that the α peptide binds to lipid-2, then recruits the β peptide, which forms pores in the membrane of the target organism. We predict that the BHT-A two-component lantibiotic functions through a similar mechanism.
Distribution of the loci encoding the bacteriocins produced by strain BHT
PCR testing of five other S. rattus strains (67-3, FA1, GF71, IB and LG-1) showed that all had both the mutacin bht-a and bht-b structural genes. In addition, S. mutans strains GS5 and K34–1 (a local P-type 777 isolate selected because of its apparent cross-immunity with strain BHT) yielded PCR products for both loci. The sequences of the S. mutans K34–1 structural genes were identical to smb and bht-b. Following loss of production from the S. mutans GS5 smb operon, strain GS5 still showed residual bacteriocin activity which was attributed to mutacin IV . BHT-B could also account for some of the residual inhibitory activity found in the smb mutants derived from S. mutans GS5 . The two genes encoding mutacin IV (nlmA and nlmB) are also present in S. mutans strain K34-1, indicating that this strain is very similar to GS5. Our finding that strains carrying the smb-like two-component lantibiotic genes all carry the BHT-B gene indicates that the genetic loci encoding these two quite different classes of bacteriocins are probably closely linked.
This work was supported in part by the New Zealand Dental Research Foundation and the Health Research Council of New Zealand.