|Home | About | Journals | Submit | Contact Us | Français|
In streptococcal species, the key step of competence development is the transcriptional induction of comX, which encodes the alternative sigma factor σX, which positively regulates genes necessary for DNA transformation. In Streptococcus species belonging to the mitis and mutans groups, induction of comX relies on the activation of a three-component system consisting of a secreted pheromone, a histidine kinase, and a response regulator. In Streptococcus thermophilus, a species belonging to the salivarius group, the oligopeptide transporter Ami is essential for comX expression under competence-inducing conditions. This suggests a different regulation pathway of competence based on the production and reimportation of a signal peptide. The objective of our work was to identify the main actors involved in the early steps of comX induction in S. thermophilus LMD-9. Using a transcriptomic approach, four highly induced early competence operons were identified. Among them, we found a Rgg-like regulator (Ster_0316) associated with a nonannotated gene encoding a 24-amino-acid hydrophobic peptide (Shp0316). Through genetic deletions, we showed that these two genes are essential for comX induction. Moreover, addition to the medium of synthetic peptides derived from the C-terminal part of Shp0316 restored comX induction and transformation of a Shp0316-deficient strain. These peptides also induced competence in S. thermophilus and Streptococcus salivarius strains that are poorly transformable or not transformable. Altogether, our results show that Ster_0316 and Shp0316, renamed ComRS, are the two members of a novel quorum-sensing system responsible for comX induction in species from the salivarius group, which differs from the classical phosphorelay three-component system identified previously in streptococci.
Among pathogenic streptococci, some species display the ability to naturally acquire exogenous DNA by entering a physiological state known as “competence for transformation.” This mechanism relies on the synthesis of machinery for DNA uptake and recombination. Competence is a virulence determinant (35) and is potentially involved in a number of functions: increasing genome plasticity through the acquisition of new genes (9), DNA repair (46), and/or fulfilling nutritional requirements by supplying a carbon, nitrogen, phosphorus, and energy source (17).
In the model species Streptococcus pneumoniae and Streptococcus mutans, transformation is tightly regulated by cell density, but also by various environmental stresses. In the literature, competence is increasingly viewed as a general stress response because it is activated by the presence of reactive oxygen species, pH or temperature changes, antibiotics, or mutagens (1, 8, 41). In S. pneumoniae and S. mutans, development of the competence state involves two steps: early and late. The early step involves 5 genes, comABCDE (orthologous to blpABCHR in S. mutans) (38), encoding a quorum-sensing system responsible for competence activation. The precursor of the inducer peptide of competence, encoded by comC (blpC), is matured and secreted in the extracellular medium through an ABC-type transporter encoded by comAB (blpAB) (27, 38). The expression of comC increases with cell density and with the presence of stress signals in the medium (41, 46). When its extracellular concentration reaches a critical threshold, the mature peptide activates ComD (BlpH), a transmembrane histidine kinase, which in turn stimulates the phosphorylation of the response regulator ComE (BlpR). The phosphorylated regulator positively regulates the expression of the early genes comABCDE (positive feedback loop), as well as comX, encoding the alternative sigma factor σX, specifically required for activation of the late step of competence. The σX factor transiently associates with the RNA polymerase core, which can then bind to the promoter region of the late genes (27). The late essential genes are required for the biosynthesis of the DNA uptake machinery, protection of single-stranded DNA, and integration of the new genetic material by homologous recombination (27, 42). The σX factor also induces the expression of genes involved in other functions, such as adaptation to stress conditions (41, 42).
Streptococcus thermophilus, a member of the salivarius group, is of major economic importance, since it is widely used for the manufacture of yoghurt and cheese. In silico analysis of the genomes of three strains of S. thermophilus (LMG18311, CNRZ1066, and LMD-9) has revealed the presence of a comX gene and of all the late competence genes essential for natural competence (22). However, it seems to lack orthologues of the quorum-sensing system that controls comX expression in S. mutans and S. pneumoniae, since the blpABCHR genes of S. thermophilus were shown to control bacteriocin production in the species (18, 19). In 2006, Blomqvist et al. (4) showed the functionality of natural transformation in S. thermophilus LMG18311 at a remarkably high frequency (10−2 to 10−3). The strategy was based on the overexpression of comX under the control of a BlpC-inducible promoter from S. thermophilus (4). Recently, Gardan et al. (20) showed that competence development in S. thermophilus can spontaneously turn on during growth in chemically defined medium (CDM). However, competence efficiency under those conditions differs between strains, since LMD-9, LMG18311, and CNRZ1066 were shown to be highly (106 transformants/ml), poorly (102 transformants/ml), and not transformable, respectively. In strain LMD-9, the oligopeptide ABC transporter Ami was shown to be absolutely required for the transcriptional induction of comX and activation of natural transformation under those conditions (20). Furthermore, it was shown that the two oligopeptide-binding proteins AmiA1 and AmiA3 are differentially involved in the triggering of competence, since their inactivation results in 99% and 50% decreases in the transformation rate, respectively (20). In S. pneumoniae, mutations in the orthologous oligopeptide binding proteins were also shown to modulate competence development, but the underlying regulatory mechanisms remained unclear (2). The Ami system of Gram-positive bacteria actively imports oligopeptides present in the extracellular medium and is known to have both nutritive and signaling functions (39). Since CDM is a peptide-free medium, Gardan et al. (20) hypothesized that S. thermophilus LMD-9 synthesizes and secretes a specific competence-stimulating peptide, which is then sensed and imported by the Ami system. Once internalized, this pheromone would then interact with a specific cytoplasmic regulator, leading to the transcriptional induction of comX (20). The pheromone and the transcriptional regulator responsible for comX expression and natural transformation under CDM conditions are still unknown.
The aim of the present study was to identify the genetic determinants involved in the early steps of competence development of S. thermophilus LMD-9 under CDM-inducing conditions. In order to reveal early competence genes, we compared the transcriptomes of Ami− and ComX− strains, both unable to switch on late competence genes. This transcriptomic approach led to the identification of an Rgg-like transcriptional regulator and its cognate pheromone, both essential for the transcriptional control of comX. Moreover, we showed that this pheromone can stimulate or activate competence development in S. thermophilus and Streptococcus salivarius strains that are poorly or not transformable.
The bacterial strains and plasmids used in the present study are listed in Table Table1.1. Plasmids derived from pG+host9 (37) were constructed in Escherichia coli EC1000 (32). E. coli was grown in LB medium with shaking at 37°C (49). S. thermophilus and S. salivarius were grown anaerobically (BBL GasPak Systems; Becton Dickinson, Franklin Lakes, NJ) at 37°C in M17 broth (Difco Laboratories Inc., Detroit, MI) or in CDM, as described by Letort and Juillard (34). Both media contain 1% (wt/vol) lactose. When required, erythromycin (250 μg/ml for E. coli; 2.5 μg/ml for S. thermophilus and S. salivarius) or chloramphenicol (20 μg/ml for E. coli; 5 μg/ml for S. thermophilus) was added to the media. Solid agar plates were prepared by adding 2% (wt/vol) agar to the medium.
Growth, luciferase activity, and transformation experiments for S. thermophilus were monitored in the Synergy HT multimode reader (BioTek, Winooski, VT). The reader automatically measures the optical density at 600 nm (OD600) and luminescence at 595 nm (expressed in relative light units [RLU]) of culture samples at 10-min intervals.
Before the experiment was started, small volumes (300 μl) of culture samples were transferred to the wells of a prewarmed (37°C) sterile covered white microplate with a transparent bottom (Greiner, Alphen a/d Rijn, the Netherlands). The luciferase activity catalyzed by LuxAB requires the presence of nonanal as a substrate. In all experiments, we supplied nonanal (Acros Organics, Geel, Belgium) in a volatile form to the cultures by placing 50 μl of a solution containing 1% nonanal diluted in mineral oil (Sigma) in the spaces between the wells of a covered microplate as described by Bachmann et al. (3). The microplate was next transferred to the prewarmed automatic reader (37°C) (time zero [t0]) and incubated at 37°C for 5 h.
An overnight culture of S. thermophilus grown in M17 at 37°C was washed twice (5,000 × g; 9 min; room temperature) in 1 volume of CDM and resuspended in 1 volume of CDM. The washed culture was then 30-fold diluted in CDM. Small volumes (300 μl) of the culture sample were next transferred into the wells of a sterile microplate (Greiner). In one sample, 1 μg of plasmid (pGIUD0855ery or pG+host9) or 500 ng overlap PCR fragments was added at the beginning of the experiment (t0). The plate was next incubated at 37°C in the prewarmed Synergy HT multimode reader (see above). In the supplementation experiments, different concentrations of synthetic forms of Shp0316 (NH2-COOH) peptides (purity > 95%) supplied by Peptide 2.0 (Chantilly, VA) were added to the 300-μl culture samples after 1 h 30 min of growth at 37°C (t1.5). After 5 h (t5), samples (100 μl of serial dilutions in M17 broth) containing DNA, or not (negative control), were spread on M17 plates containing erythromycin in the case of pGIUD0855ery and pG+host9 or chloramphenicol in the case of overlap PCR fragments. The transformation rate was calculated after 30 h of incubation at 37°C or 29°C (only for pG+host9) as the number of antibiotic-resistant CFU per ml divided by the total number of viable CFU per ml. After the transformation experiments, the integration of the antibiotic resistance cassette at the right location or acquisition of pG+host9 was checked by PCR. For pGIUD0855ery and pG+host9, the primer pairs Chstu0855A-Chstu0855B and pGhost1-pGhost2, respectively, were used. The primer pairs used for overlap PCR fragment are listed in Table S1 in the supplemental material.
General molecular biology techniques were performed according to the instructions given by Sambrook et al. (49). Electrotransformation of E. coli was performed as described by Dower et al. (14). Electrocompetent S. thermophilus cells were prepared as previously described (4). After transformation with 1 μg of pG+host9 derivative plasmids, the cells were immediately resuspended in 1 ml M17 and incubated anaerobically for 6 h at 29°C. S. thermophilus chromosomal DNA was prepared as described by Ferain et al. (15). PCRs were performed with Fhusion high-fidelity DNA polymerase (Finnzymes, Espoo, Finland) in a GeneAmp PCR system 2400 (Applied Biosystems, Foster City, CA). The primers used in this study were purchased from Eurogentec (Seraing, Belgium) and are listed in Table S1 in the supplemental material.
Plasmid pGIUD0855ery was constructed to assess and compare the natural transformation rates of S. thermophilus strains. It allowed the replacement of the stu0855 open reading frame (ORF) of strain LMG18311 (ster_0891 and str0855 in the case of strains LMD-9 and CNRZ1066, respectively) by an erythromycin resistance cassette. This plasmid was constructed as follows. In the first step, an overlapping PCR was performed to create a DNA fragment containing the erythromycin expression cassette (Pery-ery), flanked by the 1-kb upstream (UP) and downstream (DN) regions of stu0855. The Pery-ery cassette and the UP and DN fragments were separately amplified using pUC18ery and S. thermophilus chromosomal DNA, respectively, as templates (the primer pairs are listed in Table S1 in the supplemental material). The 3′ end of UP and the 5′ end of DN fragments, respectively, were complementary to the 5′ and 3′ ends of Pery-ery. The three fragments were then mixed in equimolar concentrations and joined together by PCR with the external primers. In the second step, the overlap fragment obtained was digested with SphI/KpnI and cloned in the similarly digested pUC18 vector.
The reporter strains CB001 and CB002 were constructed by replacing part of the blp locus or comX ORF, respectively, in strain LMD-9 with the transcriptional fusions PcomX-luxAB. Replacements of blp bacteriocin operons (18) and comX were obtained by double homologous recombination of the pG+host9 derivative plasmids pGICB001 and pGICB002 after two steps of temperature shift, as previously described (37). The recombinant strains were confirmed by PCR with primers located upstream and downstream of the recombination regions. The pGICB001 and pGICB002 plasmids were constructed from the pJIM4900 vector, which is a pG+host9 derivative plasmid containing the Photorabdus luminescens luxAB genes (a generous gift from E. Guédon). Plasmid pGICB001 was used to create a blpD-blpX::PcomX-luxAB reporter strain and was obtained in two steps. First, 1-kb fragments corresponding to the UP region of blpD and the DN region of blpX were PCR amplified with specific primer pairs (see Table S1 in the supplemental material); digested by PshA1/SpeI and SalI/PvuII, respectively; and cloned upstream and downstream of the luxAB genes of plasmid pJIM4900, respectively, using the same restriction enzymes. Second, to create a PcomX-luxAB transcriptional fusion, the fragment containing the comX expression signals was amplified with the primer pair DPX1-DPX2, digested by SpeI/EcoRI, and cloned in transcriptional fusion with luxA in the plasmid obtained in the first step. Plasmid pGICB002 was used to create a comX::PcomX-luxAB reporter strain and was obtained as follows: 1-kb fragments corresponding to the UP and DN regions of comX were PCR amplified with specific primer pairs (see Table S1 in the supplemental material); digested by PshA1/SpeI and SalI/PvuII, respectively; and cloned upstream and downstream, respectively, of the luxAB genes of plasmid pJIM4900 (the same restriction enzymes were used). After two steps of homologous recombination of plasmids pGICB001 and pGICB002, the strains that had integrated the PcomX-luxAB fusion, CB001 and CB002, respectively, were confirmed by PCR.
Strain CB003 was constructed by deleting the comX ORF in strain LMD-9 by double-crossover events of plasmid pGICB003, a pG+host9 derivative (37). Plasmid pGICB003 was constructed as follows. Plasmid pGICB002 was digested with SalI and partially digested by EcoRI to remove the luxAB genes. The linearized plasmid was then filled in with the Klenow fragment and ligated.
The LMD-9 derivative strain LF115 (amiA1-amiF::P32-cat) and CB001 derivative strains LF116 (amiA1-amiF::P32-cat), LF117 (ster_0316::P32-cat), LF118 (shp0316::P32-cat), LF119 (IRshp0316::P32-cat) (IRshp0316 is the inverted repeat located upstream of shp0316), and LF120 (amiA3::P32-cat) were constructed by replacing the sequence between the start and stop codons of the target gene or a specific region with the chloramphenicol expression cassette P32-cat (double homologous recombination). DNA fragments containing P32-cat flanked by the UP and DN regions of the target gene or region were made in vitro by overlapping PCR. For this purpose, the 3′ ends of UP fragments and the 5′ ends of DN fragments are complementary to the 5′ and 3′ ends of P32-cat, respectively. The strategy used was the following. In the first step, the UP, DN, and P32-cat fragments were PCR amplified separately. The P32-cat cassette (1.3 kb) was amplified with primers Uplox66 and DNlox71 using plasmid pNZ5319 as a template (30). The UP and DN regions (1.2 kb) were PCR amplified using the LMD-9 chromosome as a template with specific primers (UPA-UpB and DNA-DNB primers, respectively, which are listed in Table S1 in the supplemental material). In the second step, the UP, DN, and P32-cat fragments were mixed in equimolar concentrations and joined together by PCR using the specific UpA and DNB external primers. The resulting fragments (500 ng) were further used to transform naturally competent cells of strain LMD-9 or CB001, as described above. Transformants were selected on M17 plates containing chloramphenicol and checked by PCR using the specific primers ChA and ChB listed in Table S1 in the supplemental material.
The experimental procedures for RNA extraction and microarray experiments were detailed in Text S1 in the supplemental material.
In S. pneumoniae and S. mutans, the comABCDE genes responsible for comX induction are regulated by a ComX-independent positive feedback loop, which is initiated when a critical extracellular concentration of the competence-stimulating peptide is reached and sensed (27). In S. thermophilus, we postulated that a similar autoamplification mechanism governs the regulation of the early competence genes and that activation depends on the reimportation of a secreted pheromone via the Ami system (20). Based on this hypothesis, our leading strategy to identify the early competence genes was to compare the transcriptomes of Ami− and ComX− strains, since early genes should be induced only in the latter. To determine the appropriate time to extract RNAs, we monitored the kinetics of comX induction in different S. thermophilus LMD-9 backgrounds, i.e., strains that contain or do not contain a functional ComX or Ami system. For this purpose, three comX reporter strains were constructed. Strain CB001 (ComX+ Ami+) was obtained by replacing the blp bacteriocin operons (18, 19) with a transcriptional fusion between the intergenic region upstream of comX, called PcomX, and the luciferase genes luxAB of Photorabdus luminescens. In strain CB002 (ComX− Ami+), luxAB genes are under the control of the native PcomX promoter and replace the comX ORF (comX::luxAB). To construct strain LF116 (ComX+ AmiA1-AmiF−), the amiA1-amiF region of strain CB001 (ComX+ Ami+) was replaced by the chloramphenicol resistance cassette P32-cat. Prior to any experiment, we assessed the competence efficiency of the reporter strains. As expected, the transformation rate of strain CB001 (ComX+) was not affected compared to that of the wild-type LMD-9 strain, while the deletion of comX in strain CB002 or amiA1-amiF in strain LF116 abolished competence (Table (Table2)2) (20). Growth and luciferase activities were monitored under CDM growth conditions at 37°C in an automatic multimode reader. The results are presented in Fig. Fig.1A.1A. By monitoring the number of RLU per OD600 unit (RLU/OD600) of strain CB001 (ComX+ Ami+), we found that PcomX activity started to increase in the early exponential phase and reached a maximum (RLU/OD600 = 15,000) before mid-log phase, after 50 min of induction. The activity of PcomX remained stable for 10 min and then constantly decreased to reach an RLU/OD600 of 380 in stationary phase. These kinetics are in accordance with those of the appearance of transformed cells in CDM (20). The kinetics of PcomX activity in strain CB002 (ComX− Ami+) followed a similar pattern, except that the maximum number of RLU per OD unit obtained was 3-fold higher and the activity remained stable for a longer time than with strain CB001 (Fig. (Fig.1A).1A). As expected, the luciferase activity of strain LF116 remained low (maximum RLU/OD600 = 250) throughout growth, confirming the role of the Ami system in comX transcription. Interestingly, when we compared the growth curves of the transformable strain CB001 to those of the competence-deficient strains CB002 and LF116, we observed a ComX-dependent decrease in the growth rate and the final growth yield. Deletion of the blp bacteriocin operons in strain CB001 was not responsible for this phenotype, since the growth of the wild-type LMD-9 strain was similar to that of CB001 (Fig. (Fig.1B).1B). This shows that competence development of LMD-9 under CDM conditions has a deleterious effect on growth.
For transcriptomic study purposes, ComX− (CB003; LMD-9 ΔcomX) and AmiA1-AmiF− (LF115, LMD-9 amiA1-amiF::P32-cat) strains harboring a native blp locus were constructed. These competence-defective strains displayed growth curves in CDM similar to those of their Blp− isogenic strains, with an improved growth rate and yield compared to the wild-type strain (Fig. (Fig.1B).1B). Total RNA was extracted from mid-log-phase growing cells (CDM conditions), where the RLU/OD600 of strain CB002 was found to be maximal. Labeled RNAs were then cohybridized to Agilent LMG18311 microarrays as described in Materials and Methods. As expected, the ami probes were highly induced in the ComX− strain, except for amiA1, for which specific probes were not present in the LMG18311 microarrays (see Table S2 in the supplemental material). In contrast, no comX-specific probes were repressed, confirming that comX is no longer induced in the AmiA1-AmiF− strain grown in CDM (data not shown).
To identify genes involved in the activation of comX expression, we focused our attention on genes that were induced in the ComX− strain compared to the AmiA1-AmiF− strain. We considered relevant only those genes that were represented by more than 50% induced probes with a mean absolute fold change (FC) of at least 10. Induced genes can be separated into highly (10 < mean FC < 14) and very highly (19 < mean FC < 100) regulated genes. Interestingly, the latter are organized into 4 loci that are related to bacteriocin production (see Table S2 in the supplemental material). The first locus (ster_0317-ster_0319) encodes a truncated ABC transporter specific for 2-Gly peptides. Loci 2 (blpABC) and 3 (orf4-orf6) belong to the blp bacteriocin cluster and, respectively, encode the induction factor and the secretion apparatus of thermophilin 9 and potential immunity proteins (18, 19). Finally, genes of locus 4 (ster_1720-ster_1718) include homologues of the mutacin I biosynthetic genes. Of the three genes belonging to the group of less induced genes (blpG, ster_0935, and ster_1924), only blpG is related to bacteriocin production, since it was shown to be involved in thermophilin 9 interspecies activity (19). Coprogrammed production of bacteriocins and competence development was previously reported in several transformable streptococcal species and was proposed to play an important role in the acquisition of substrate DNA for transformation (7, 21, 29, 41).
To provide further evidence that the four identified loci specifically include early competence genes, we aligned their upstream regions with the intergenic region upstream of comX. This allowed us to highlight a conserved organization of their promoter sequences (60 nucleotides [nt]; 43% identity). Importantly, they all contain a highly conserved motif (20 nt; 80% identity) consisting of two 9-bp inverted repeats with the following consensus sequence: 5′-TAGTGACATNTATGTCACTA-3′, called the early competence box (ECom box). This motif is followed by a conserved T tract and a putative −10 box, which is located 21 bp downstream of the ECom box in all promoters (Fig. (Fig.2A).2A). The ECom box could serve as a binding site for a positive transcriptional regulator specific to the early competence step. This motif is not present upstream of the three less induced genes, strongly suggesting that they are differentially regulated and probably not specific to early competence in S. thermophilus.
Since we hypothesized that an autoamplification mechanism governs comX induction, we searched for genes coding for precursors of signaling peptides and transcriptional regulators among the four early competence operons.
None of the induced gene products displayed obvious regulator features, but two transcriptional regulators were encoded in the direct vicinity of early competence genes: BlpR and Ster_0316. They belong to the LytR and Rgg families of regulators, respectively. Strain LMD-9 ΔblpR (LF101) from our mutant collection (18) displayed a competence efficiency similar to that of the wild-type strain (data not shown). To assess the role of ster_0316 in early competence induction, the ster_0316 ORF of the PcomX-luxAB reporter strain CB001 was replaced by the P32-cat cassette (LF117) (Table (Table11 and Fig. Fig.2B).2B). Compared to the control strain CB001 cultivated in CDM, the low luciferase activity (maximum RLU/OD600 = 160) of strain LF117, together with its improved growth rate and yield, unequivocally showed the key role of ster_0316 in the activation of comX transcription (Fig. (Fig.3A).3A). As expected from PcomX activity, competence experiments revealed that strain LF117 was no longer transformable (Table (Table2),2), supporting its role in natural transformation.
Since most Rgg regulator genes are located close to their associated pheromone-encoding genes (25, 26), a closer examination of the ster_0317 early competence locus was performed. We identified a 75-bp ORF that was not annotated in the public database. It encodes a 24-amino-acid hydrophobic peptide and was therefore called Shp0316 (Shp for small hydrophobic peptide) (Fig. (Fig.2B).2B). The putative signaling function of Shp0316 was tested by monitoring the PcomX activity and transformation rate of the Shp0316− strain (LF118) (Table (Table11 and Fig. Fig.2B),2B), which was constructed using the same strategy described above for strain LF117. The results in Fig. Fig.3B3B showed that the deletion of shp0316 impaired comX induction in CDM. Indeed, a maximum RLU/OD600 of 230 was measured in LF118, which is dramatically reduced compared to the control strain CB001 and similar to the Ster_0316− and AmiA1-AmiF− reporter strains. In addition, no transformant could be obtained for strain LF118 in natural competence experiments (Table (Table2).2). These results demonstrate that Shp0316 is essential for early competence development and acts at the comX transcriptional level in S. thermophilus.
The essential role of shp0316 in comX induction and the presence of an ECom box in its promoter sequence suggest that its transcription depends on a positive feedback loop. In this context, the Rgg regulator Ster_0316 could play a role in this regulation pathway either directly, by binding to the ECom box of the Pshp0316 promoter sequence, or indirectly by stimulating the transcription of a second regulator. To assess the importance of the putative Ster_0316 binding site for comX induction, we replaced the corresponding ECom box and T tract (IRshp0316) with the P32-cat cassette in the reporter strain CB001 (LF119) (Table (Table11 and Fig. Fig.2B).2B). In this construct, the putative −10 box of Pshp0316 and shp0316 translation signals were maintained, but shp0316 transcription was under the control of the upstream constitutive P32 promoter. The consequences of this promoter exchange were studied by comparing the PcomX activities of the reporter strains CB001 and LF119 (IRshp0316::P32-cat) grown in CDM (Fig. (Fig.3C)3C) and by performing competence experiments. Interestingly, the growth rates of the CB001 and LF119 strains were similar, but the kinetics of PcomX induction dramatically changed. Expression was no longer induced during growth, but rather, constitutively increased from the beginning of growth until the cells entered the stationary phase. The maximum RLU/OD600 measured was lower than that of the control strain CB001 but high enough to develop normal competence (Fig. (Fig.3C3C and Table Table2).2). This clearly shows that the ECom box of Pshp0316 plays a critical role in the kinetics of comX induction, supporting the hypothesis that Shp0316 is part of an autoamplification loop governing the early competence state.
Altogether, our results strongly suggest that Ster_0316 and Shp0316 are part of an Rgg regulator/signaling peptide system controlling early competence induction in S. thermophilus LMD-9.
To fulfill its putative signaling function, peptide Shp0316 must be secreted and probably matured in order to be sensed by S. thermophilus cells. For example, the peptides internalized by oligopeptide transporters and involved in the regulation of conjugation in Enterococcus faecalis (33) and virulence in Bacillus cereus (52) are synthesized as longer precursor peptides in the cytoplasm. Then, they undergo several maturation steps leading to the release of the active pheromone in the extracellular medium (5, 43). Shp316 (MKTLKIFVLFSLLIAILPYFAGCL) displays all the characteristics of a signal sequence from a lipoprotein: a size between 20 and 25 amino acids, a positively charged N-terminal end, a central hydrophobic core, and a conserved −2, −1 cleavage site followed by a +1 Cys residue (lipobox) (24) (Fig. (Fig.4A).4A). Interestingly, all mature conjugation peptides of E. faecalis identified to date correspond to the C-terminal part of lipoprotein signal sequences (10).
To test the hypothesis that Shp0316 is matured and released in the medium in order to act as a competence pheromone, experiments with supplementation of the competence defect of the Shp0316− strain (LF118) were carried out. Based on the sizes and sequences of known conjugation-inducing peptides (10) and of the virulence-signaling peptide PapR (43), a heptapeptide corresponding to the C-terminal part of Shp0316 was synthesized (Shp031618-24) (Fig. 4A and B) and tested for its ability to induce PcomX activity and natural transformation. Different concentrations of Shp031618-24 (0 nM, 10 nM, 100 nM, 500 nM, 1 μM, and 2.5 μM) were added to the early log growth phase of LF118 cells grown under CDM conditions at 37°C. Luciferase activity was next monitored for 5 h before competence experiments were performed. The results presented in Fig. Fig.4C4C show that the luminescence driven by PcomX-luxAB from strain LF118 (Shp0316−) displayed a dose response to the amount of Shp031618-24, which is typical of pheromone-regulated systems. Indeed, Shp031618-24 concentrations lower than 100 nM had no inducing effect on PcomX-luxAB activity, while the response was almost linear between 100 nM and 1 μM. At concentrations higher than 1 μM, the luciferase activity of LF118 reached a plateau, indicating saturation of Shp031618-24 induction. In each case, the maximum RLU/OD600 was measured before the mid-log growth phase. The measured transformation rates and PcomX-luxAB activity of strain LF118 displayed a similar dose response to the Shp031618-24 concentration (Fig. (Fig.4D).4D). Transformation rates in the presence of 500 nM or higher concentrations of Shp031618-24 were even higher than those of the control strain CB001 grown in the absence of inducing peptide (Table (Table22 and Fig. Fig.4D).4D). Altogether, our results demonstrate that Shp0316 is the precursor of the competence-stimulating peptide in S. thermophilus LMD-9.
Longer Shp0316 derivative forms were synthesized in order to test the importance of the N-terminal residues in the signaling activity of Shp0316 (peptides Shp031617-24, Shp031616-24, and Shp031615-24) (Fig. (Fig.4B).4B). The maximum peptide length tested was fixed at 10 amino acids, since importation by the Ami/Opp systems of Gram-positive species was shown to be optimal in this peptide size range (13). Since lipoprotein signal sequences are cleaved upstream of the +1 Cys residue, we also assessed the role of the two last C-terminal amino acids by synthesizing peptides Shp031618-23, Shp031618-22, and Shp031618-20 (Fig. (Fig.4B).4B). The inducing potential of 1 μM of each peptide was evaluated in strain LF118 (Shp0316−) as described above. The transformation rates obtained (Fig. (Fig.4B)4B) show that (i) the N-terminal residues of Shp0316, or at least amino acids 15 to 17, are not critical for the signaling activity of Shp0316, while (ii) the C-terminal Leu residue is specifically involved in this function. The latter conclusion was further confirmed, since the removal of Leu24 from the active Shp0316 derivative forms displaying a longer N-terminal part abolished their inducing properties (peptides Shp031617-23, Shp031616-23, and Shp031615-23) (Fig. (Fig.4B4B).
To provide evidence that the Ami system and the Ster_0316 regulator are obligate intermediates between Shp0316 signaling and comX induction during early competence, we performed induction experiments with the nontransformable AmiA1-AmiF− and Ster_0316− reporter strains and the poorly transformable AmiA3− reporter strain (LF120) (Table (Table1);1); the latter displayed a 100-fold reduced transformation rate compared to the control strain CB001 (Table (Table2),2), as reported previously by Gardan et al. (20). Peptide Shp031618-24 was added to CDM at a final concentration of 1 μM. As expected, addition of Shp031618-24 could not restore the PcomX-luxAB activity and competence deficiency of AmiA1-AmiF− and Ster_0316− strains (maximum RLU/OD600, 210 and 250, respectively) (Table (Table2),2), which strongly supports their respective hypothetical roles in Shp0316 importation and comX transcription. In contrast, the competence level of strain AmiA3− was partially restored in the presence of 1 μM Shp031618-24, since the maximum RLU/OD600 and the transformation rate increased 1.3-fold and 4-fold, respectively (Table (Table2).2). In this strain, a high extracellular Shp031618-24 concentration could eventually facilitate its recognition by AmiA1, which was also shown to play a role in competence induction (20).
The ability of S. thermophilus to naturally turn on high competence levels under CDM growth conditions seems restricted to strain LMD-9 (reference 20 and L. Fontaine, unpublished results). Since all of the key early competence genes identified to date are in the genome of the poorly transformable LMG18311 and CNRZ1066 strains, we assumed that differences in competence efficiency between S. thermophilus strains could be due to different levels of Shp0316 production, maturation, and/or recognition. To test this hypothesis, we performed competence experiments with strains LMG18311 and CNRZ1066 induced with 1 μM Shp031618-24. The results presented in Table Table22 show that the extracellular addition of the heptapeptide restores the competence defect of these strains to levels similar to those of LMD-9 grown in CDM.
From our work and recent results of Gardan et al. (20), it is clear that evolution has selected different regulation mechanisms to govern competence in the salivarius and mitis/mutans groups of streptococci. In this context, analysis of the recently available draft genome sequence of S. salivarius SK126 (GenBank accession no. NZ_ACLO00000000) allowed the identification of a locus with a strong sequence identity to the ster_0316-shp0316 genes of S. thermophilus. The orthologue of Shp0316 from SK126 displays 79% identity to Shp0316 from LMD-9 but with only a single amino acid difference (A21T) in the C-terminal heptapeptide (Fig. (Fig.4A).4A). To provide evidence that S. salivarius has a similar competence regulation pathway, we performed transformation experiments with S. salivarius ATTC 25975, JIM8777 (12), and JIM8780 (12). S. salivarius was grown at 37°C in CDM containing 1 μg of pG+host9 plasmid DNA. Induction was performed with 1 μM Shp031618-24 from LMD-9, as described for S. thermophilus. The S. salivarius strains were poorly transformable or not transformable in the absence of the inducing peptide, but addition of Shp031618-24 activated competence development of the three strains with a transformation rate for strain JIM8777 similar to that of S. thermophilus CNRZ1066 (Table (Table22).
The presence of late competence genes in all streptococcal genomes sequenced to date strongly suggests that natural DNA transformation has been positively selected through evolution of these microorganisms, whatever their niche. However, the regulatory mechanisms governing comX induction in early competence development seem to have evolved independently in species from the S. mitis/mutans and salivarius groups. These systems mainly differ in the mechanisms of competence pheromone recognition and information transfer for comX induction. Indeed, the results obtained in our study strongly suggest the following model for early competence induction in S. thermophilus and S. salivarius (Fig. (Fig.5):5): under CDM growth conditions, peptide Shp0316 (renamed ComS for competence signal) is produced, matured, secreted, and mainly sensed by the AmiA3 substrate-binding protein when a critical extracellular concentration is reached. The mature Shp0316 form, ComS*, would then be imported by the Ami transporter and interact with the cytoplasmic Rgg regulator Ster_0316 (renamed ComR for competence regulator), resulting in its activation. Subsequent binding to the ECom box could finally promote comX transcription and the activation of the ComS-mediated positive feedback loop.
This model of competence development in S. thermophilus has similarities to the peptide-mediated circuits governing conjugation in Enterococcus faecalis (33, 51), virulence in the B. cereus group (52), and Phr signaling in Bacillus subtilis (45). Precursors of the conjugation pheromone, PapR (43), and the competence-stimulating factor, CSF (31), that, respectively, regulate these systems undergo several maturation steps. For example, the release of active conjugation octapeptides from signal sequences of lipoproteins requires two sequential processing steps, potentially catalyzed by the signal peptidase II Lsp (10) and the pheromone-specific peptidase Eep (5). In the case of cCF10, a still unknown additional peptidase is required to remove the C-terminal residue, yielding a heptapeptide pheromone (5). The composition of ComS and the presence of a putative lipobox initially suggest that it could be exported through the type II Sec-dependent secretion pathway, as suggested for lipoprotein precursors of conjugation-signaling peptides. However, the signal peptidase II does not seem to be involved in ComS maturation, since we showed that the C-terminal Leu residue is essential for ComS-inducing properties. However, the Eep homologue of S. thermophilus and S. salivarius may be responsible for the release of the C-terminal active ComS from the precursor, which is predicted to form one transmembrane segment (from amino acids 4 to 18) with a surface-exposed C terminus (TMHMM prediction [http://www.cbs.dtu.dk/services/TMHMM]). Additionally, we could not exclude the possibility that ComS is further matured by peptidases after its internalization. Whatever the actors involved in the maturation step(s), the active form of ComS shows some flexibility in its N terminus, similar to PapR (43). Indeed, the addition of 3 amino acids to the N terminus of ComS18-24 did not affect its inducing properties.
ComR belongs to the Rgg family of pleiotropic transcriptional activators. This family is well represented among enterococci, lactococci, and streptococci, where they regulate various functions, such as production of secreted proteins (RopB of Streptococcus pyogenes) (6), production of mutacin II (MutR of S. mutans) (47) and lactocin S (LasX of Lactococcus lactis) (48), production of acid (GadR of L. lactis) (50), and oxygen tolerance (RggC of S. thermophilus) (16). The Rgg family is characterized by a well-conserved N-terminal XRE helix-turn-helix (HTH) DNA binding motif and a more variable C-terminal domain (36). ComR displays the highest level of identity with its orthologue in S. salivarius (93% identity) and with SMU.3181c of S. mutans (42% identity) and a lower level of identity (ranging from 34% to 28%) with regulators present in other streptococcal genomes. Prediction of the secondary and tertiary structures of ComR identified the quorum-sensing regulators PlcR of Bacillus thuringiensis and PrgX of E. faecalis as the best homologues (LOMETS predictions [http://zhang.bioinformatics.ku.edu/LOMETS]). The members of this superfamily of regulators are characterized by a C-terminal regulatory domain composed of 11 α-helices involved in dimerization and specific pheromone interaction. Binding of their cognate signaling peptide induces a conformational change in the regulatory domain, which rearranges the DNA-binding domain (11, 51). The C-terminal domains of the PrgX, PlcR, MutR, and Rap proteins are believed to derive from a common ancestor domain to which additional domains, like HTH or phosphatase domains, have been added, yielding proteins with new regulatory functions (11). Since PlcR is known to bind to inverted repeats similar to those of ComR (30% identity) (44), it is likely that ComS-activated ComR directly binds to the in silico-identified ECom box. The activator-versus-repressor function of ComR is supported by three observations: (i) PcomX is no longer induced in the ComR− strain, (ii) ECom boxes are localized upstream of putative −10 boxes, and (iii) most regulators of the Rgg family are activators (6, 44, 47, 50), except PrgX (51) and LasX, with the latter displaying both repressor and activator functions (48).
Remarkably, we observed that competence development in S. thermophilus LMD-9 has a clear deleterious effect on growth under CDM conditions. In all competence-deficient LMD-9 derivative strains, both the growth rate and yield were improved. This effect was not observed in the poorly transformable strain LMG18311, where comX deletion had no impact on growth (data not shown). The negative effect of a high level of competence could be the consequence of the energy cost associated with induction of late competence genes. Alternatively, it is possible that a subpopulation of cells undergoes lysis when competence is induced, as reported for S. pneumoniae (53) and S. mutans (41). In this context, most early competence genes identified in our transcriptomic study are related to bacteriocin production, a process shown to be responsible for competence-induced lysis in streptococci (41). Interestingly, most S. thermophilus strains are unable to spontaneously turn on competence under CDM conditions while keeping the ability to respond to the addition of ComS18-24 or to develop competence in coculture experiments with LMD-9 (unpublished results). This could indicate “cheating” behavior that is generally observed when the benefits of a particular function are balanced by energy costs that are unfavorable for the population fitness (e.g., bacteriocin production and immunity and protease+ and protease− strains of L. lactis) (23, 28). An alternative explanation could be that the environmental conditions that trigger ComS production in most S. thermophilus strains are different from those in strain LMD-9.
Multilocus sequence typing (MLST) studies performed on S. thermophilus, S. salivarius, and Streptococcus vestibularis revealed that these species have recently diverged from a common ancestor (12). This is further supported by our cross-induction experiments, which showed that ComS-mediated communication is still possible between S. salivarius and S. thermophilus. An interesting perspective to gain further insight into the mechanisms involved in speciation within the salivarius group would be to investigate the diversity of the comRS locus in order to evaluate the presence or absence of distinct competence pherotypes. Finally, the identification of ComRS as a key quorum-sensing system activating competence instead of a classical phosphorelay three-component quorum-sensing system raised the question of the recruitment through evolution of similar regulation systems to control competence in other streptococci and related species.
This research was carried out with financial support from Danisco and FNRS. C.B. holds a doctoral fellowship from FRIA. L.F. is a postdoctoral researcher at FNRS. P.H. is a research associate at FNRS.
We are grateful to E. Maguin for providing the pG+host9 vector, E. Guédon for providing plasmid pJIM4900, and P. Renault for providing S. salivarius JIM8777 and JIM8780. We are grateful to E. J. Smit for the generous gift of custom Agilent microarrays of S. thermophilus LMG18311. We warmly thank P. Goffin for critically reading the manuscript.
Published ahead of print on 18 December 2009.
†Supplemental material for this article may be found at http://jb.asm.org/.