Competence without a Competence Pheromone in a Natural Isolate of Streptococcus infantis

Abstract

Many streptococcal species belonging to the mitis and anginosus phylogenetic groups are known to be naturally competent for genetic transformation. Induction of the competent state in these bacteria is regulated by a quorum-sensing mechanism consisting of a secreted peptide pheromone encoded by comC and a two-component regulatory system encoded by comDE. Here we report that a natural isolate of a mitis group streptococcus (Atu-4) is competent for genetic transformation even though it has lost the gene encoding the competence pheromone. In contrast to other strains, induction of competence in Atu-4 is not regulated by cell density, since highly diluted cultures of this strain are still competent. Interestingly, competence in the Atu-4 strain is lost if the gene encoding the response regulator ComE is disrupted, demonstrating that this component of the quorum-sensing apparatus is still needed for competence development. These results indicate that mutations in ComD or ComE have resulted in a gain-of-function phenotype that allows competence without a competence pheromone. A highly similar strain lacking comC was isolated independently from another individual, suggesting that strains with this phenotype are able to survive in nature in competition with wild-type strains.

In streptococci belonging to the mitis and anginosus phylogenetic groups, competence for genetic transformation (natural competence) is a transient state that enables bacterial cells to take up and incorporate extracellular DNA into their genomes by homologous recombination. Natural competence in these bacteria is regulated by a quorum-sensing mechanism that is controlled by the competence-stimulating peptide (CSP), the product of the comC gene (15). In addition to CSP, the quorum-sensing mechanism consists of a CSP secretion apparatus (ComAB) (14), and a two-component regulatory system (ComDE) (28). Transcription of the comAB and comCDE operons is regulated by factors in the external milieu that are still poorly defined (1, 7). When the growth conditions are right, ComABCDE are produced at a low basal level, resulting in a slow accumulation of mature CSP in the growth medium. CSP will induce the competent state when the pheromone reaches an external concentration of 1 to 10 ng/ml by triggering the signal transduction pathway consisting of the CSP receptor histidine kinase ComD (16) and its cognate response regulator ComE. By analogy with other two component regulatory systems, binding of CSP will lead to receptor dimerization, followed by autophosphorylation in trans between the monomers. In its phosphorylated state, ComD acts as a specific phosphodonor for ComE, its cognate cytoplasmic response regulator (36). Upon phosphorylation, ComE activates transcription from promoters containing a characteristic imperfect direct repeat motif (35). The presence of this motif in the promoters of the comAB and comCDE operons amplifies the response to the pheromone, resulting in higher amounts of CSP in the environment and, presumably, an increased level of phosporylated ComE inside the cells (18). Autoinduction increases the amount of pheromone in the medium 30-fold or more and presumably serves to coordinate competence induction throughout the bacterial population. A closely related imperfect direct repeat motif has also been found in the promoter region of a recently characterized gene encoding an alternative sigma-factor (ComX), indicating that transcription of comX is ComE dependent (24). When expressed, ComX will replace the primary sigma factor from the RNA polymerase holoenzyme and activate transcription of genes containing a cinbox (TACGAATA) in their promoter regions. These so called late genes, which together constitute a competence regulon, encode the proteins that carry out DNA uptake, processing, and recombination.

Homologs of the late genes have been found in all streptococcal species examined thus far, but DNA uptake by natural competence has never been detected in many of these species (3, 8, 18). Interestingly, species that have been demonstrated to be naturally competent, such as the members of the mitis and anginosus phylogenetic groups (19), possess the ComABCDE quorum-sensing mechanism in addition to the late genes (17). These findings suggest that natural competence is widespread among streptococci but is difficult to detect under laboratory conditions except in species in which competence is regulated by the ComABCDE cell density monitoring mechanism. Sequencing of the genes encoding CSP (comC) from different species and strains belonging to the mitis phylogenetic group have revealed that a large number of CSPs with different primary structures are produced (17, 37). The largest CSP diversity is found among members of the species Streptococcus mitis (unpublished results). Numerous studies have shown that induction of the competent state is inhibited by any mutation, change in growth condition, or biochemical treatment that prevent CSP accumulation in the medium, demonstrating that competence development cannot take place without the CSP pheromone (34, 33, 4, 28, 18). Recently, however, some laboratory made gain-of-function mutations in comD and comE have been described that make the mutant strains partially or completely CSP independent (7, 23, 26). We describe in the present study the first natural isolate (Atu-4) that possess CSP-independent competence.

MATERIALS AND METHODS

Bacterial strains, mutants, and culture media.

The Atu-4 strain was isolated from the tongue of a healthy person in 1999. The SK-348 strain was isolated from a human oral cavity in the USA It was thoroughly characterized phenotypically by using methods and principles described previously (21) and by partial sequencing of its 16S rRNA gene. Streptomycin (Sigma)-resistant (Sm^r) mutants of Atu-4 and SK348 were isolated by plating 200 μl of a 50×-concentrated overnight culture on Todd-Hewitt (TH) broth (Difco) agar plates containing 1 mg of streptomycin/ml. The plates were incubated anaerobically at 37°C for 36 h before colonies were picked and inoculated in TH broth containing 100 μg of streptomycin/ml. All bacterial strains used in the present study were grown in TH broth. When transformation assays were carried out, the TH broth was supplemented with 100 μg of streptomycin/ml. Rapid ID32 Strep tests were conducted as described by the manufacturer (bioMérieux).

DNA purification, PCR, and DNA sequencing.

Isolation of streptococcal genomic DNA used for transformation studies, Southern blots, and as a template in PCRs was carried out by using Qiagen Genomic-Tip 100. The procedure was essentially as described by the manufacturer, with some modifications in the lysis step: 40 to 80 ml of an overnight culture was harvested at an optical density at 550 nm (OD₅₅₀) of 0.5, and the pellet was resuspended in 3.5 ml of lysis buffer B1 (Qiagen) containing 1.4 mg of RNase A (Sigma), 2.3 mg of lysozyme (Sigma), 0.5 mg of protease (Qiagen), and 115 U of mutanolysin (Sigma)/ml. After the cells were lysed at 37°C for 30 to 60 min, the remaining steps of the DNA purification protocol were carried out as described for the standard procedure supplied by Qiagen.

PCR amplification of DNA fragments corresponding to the comCDE and comDE operons of the SK348 and Atu-4 strains was performed as previously described (16) with the primers tArg2 (5′-CATAGCTCAGCTGGATAGAGCATTCGCCTTC-3′) and tGlu (5′-GGCGGTGTCTTAACCCCTTGACCAACGGACC-3′). The resulting PCR fragments were sequenced in a stepwise manner on a Perkin-Elmer/ABI Prism 377 sequencer by using the BigDye terminator cycle sequencing kit (Applied Biosystems). In the first step, the original amplification primers, tArg2 and tGlu, were used as sequencing primers. Sequence information obtained in this step was then used to synthesize new primers, etc., until sequencing of both strands of the fragments had been completed. To determine 16S rRNA sequences, a DNA fragment corresponding to the16S rRNA gene was first amplified by PCR. The PCR was performed with primers 1F (5′-GAGTTTGATCCTGG-CTCAG-3′) and 6R (5′-AGAAAGGAGGTGATCCAGCC-3′) with genomic DNA as a template. Sequencing of the fragment was carried out as described above with the following sequencing primers: 1F, 6R, and 3R (5′-CCCGTCAATTCATTTGAGTT-3′), 4R (5′-GACGGGCGGTGTGTA-3′), and 9R (5′-CGTATTACCGCGGCT-3′).

Southern blots.

Genomic DNA (5 to 10 μg) from strain Atu-4 was digested with three different restriction enzymes (ScaI, NdeI, and HindIII), loaded onto a 0.8% agarose gel, and run for 3 h at 4 to 6 V/cm. Prior to blotting, the gel was treated with depurination solution (0.25 M HCl) for 20 min, denaturation solution (0.5 M NaOH, 1.5 M NaCl) for 30 min, and neutralization buffer (0.5 M Tris-Cl [pH 7.4], 1.5 M NaCl) for 30 min. DNA was transferred from the agarose gel to a positively charged nylon membrane (Biodyne B; Pall) by capillary blotting for 12 to 16 h in 10× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and fixed to the membrane by baking at 80°C for 30 min. Prehybridization was performed at 68°C for 30 to 60 min in a buffer containing 5× SSC, 0.1% sodium dodecyl sulfate (SDS), 100 μg of nonhomologous DNA/ml, and 1× Denhardt solution. A digoxigenin (DIG)-labeled DNA probe was added directly to the prehybridization solution and incubated overnight at 68°C in a roller bottle. The 1,245-bp DNA probe was made by PCR amplification with the primers Atu5 (5′-CTACTAATCTTCTTTGCAATGA-ATGCC-3′) and Atu6 (5′-CCAATAATCCTTAACTTATGAGATACTCC-3′) and with chromosomal DNA from strain Atu-4 as a template. The probe was labeled by using the DIG High Prime Labeling and Detection Starter Kit I (Boehringer Mannheim) according to the manufacturer's instructions. After hybridization, the membrane was washed twice with gentle agitation for 5 min (each time) in 2× SSC-0.1% SDS at room temperature and then twice for 15 min (each time) in 0.1× SSC-0.1% SDS at 68°C. DIG-labeled nucleic acids were detected by an enzyme-linked immunoassay with a highly specific anti-DIG-alkaline phosphatase antibody conjugate and the color substrates nitroblue tetrazolium and BCIP (5-bromo-4-chloro-3-indolylphosphate).

Transformation assays.

Overnight cultures of Atu-4 and SK348 were diluted in preheated TH broth (37°C) to an OD₅₅₀ of 0.05 and then further diluted 10,000-fold in preheated TH broth to a final volume of 100 ml and placed in a water bath at 37°C for 30 min to restore vigorous growth. At this point (0 h in Table 1) 0.5-ml samples were removed, and 5 μg of DNA/ml from Sm^r mutants of Atu-4 and SK348 was added to wild-type Atu-4 and SK348 cultures, respectively. After the addition of Sm^r DNA, the 0.5-ml samples were incubated in a water bath at 37°C for 2 h before serial dilutions were made and plated on agar plates with or without streptomycin (100 μg/ml). These steps were repeated every hour for a period of 10 h (i.e., 10 times). After overnight incubation at 37°C the total number of CFU/ml and the number of CFU/ml growing on agar plates containing streptomycin were determined for each time point.

TABLE 1.

Competence for natural transformation in strains Atu-4 and SK348 at different population densities

Strain and time (h) after dilutiona	Total CFU/ml	Smr CFU/ml	Transformation effiency (%)
Atu-4
0	1,000	88	8.8
1	2,512	228	9
2	6,309	1,700	27
3	19,958	6,000	30
4	63,096	17,833	28
5	251,119	62,000	34.7
6	1.26 × 106	680,000	54
7	6.3 × 106	1.46 × 106	23.2
8	4.5 × 107	240,000	4.7
9	3.98 × 108	0	0
10	6.3 × 109	0	0
SK348
0	794	0	0
1	1,995	0	0
2	5,012	0	0
3	15,849	0	0
4	50,119	0	0
5	251,119	200	0.08
6	1.26 × 106	6,500	0.52
7	7.94 × 106	3.1 × 106	39
8	6.3 × 107	1 × 106	2.54
9	6.3 × 108	199,000	0.03
10	1 × 1010	0	0

^aSee Materials and Methods.

Transformation experiments with trypsin were conducted essentially as described above. After the initial dilutions and restoration of growth at 37°C for 30 min, cultures of the Atu-4 and SK348 strains were each split in two, and trypsin (Sigma) was added to one of the parallels to a final concentration of 0.02 mg/ml. Samples (0.5 ml) from all cultures were collected every 60 min, and 5 μg of Sm^r DNA/ml was immediately added to each of them. In addition, the OD₅₅₀ of each culture was determined every time a sample was collected. After incubation at 37°C for 2 h with the Sm^r DNA, serial dilutions of each 0.5-ml sample were plated on TH agar plates containing 100 μg of streptomycin/ml.

To determine whether Atu-4E, a comE mutant of strain Atu-4 (see below), is still competent for natural transformation, an overnight culture of this mutant was diluted to an OD₅₅₀ of 0.05 and then incubated at 37°C for 30 min to allow reconstitution of growth. Then, 0.5 ml of the bacterial culture was transferred to an Eppendorf tube, and 5 μg of Sm^r DNA/ml was added. After incubation at 37°C for 2 h, serial dilutions of the bacteria were plated on TH agar containing 100 μg of streptomycin/ml. The Atu-4 strain was run in parallel as a positive control. All of the transformation assays described above were repeated at least three times to verify results and reproducibility.

Construction of a comE knockout mutant.

PCR was performed with the primers atu4.kpn (5′-ATTATGGTACCGCACAGT TTATTCGTCGCCACAATCCC-3′) and atu4.bam (5′-ATTATGGATCCATCATGGTAGGGAATTTTCAAATCATTTCCC-3′) containing a KpnI and a BamHI site at their 5′ ends, respectively, and chromosomal DNA from strain Atu-4 as a template. The 257-bp PCR product, an internal fragment of the comE gene, was cleaved with KpnI and BamHI and ligated into the corresponding sites in the pEVP3 polylinker (6). The resulting plasmid, pST1, was transformed into Escherichia coli DH5α (Gibco-BRL) for plasmid propagation and finally into the Atu-4 strain by natural transformation. The plasmid pST1, which contains a chloramphenicol marker, does not replicate in streptococci. Therefore, only bacteria with the plasmid integrated into their genomes will grow in the presence of chloramphenicol. Such mutants (comE::pST1) were selected for by plating Atu-4 cells, transformed with purified pST1 plasmid on TH agar plates containing 2 μg of chloramphenicol/ml. After overnight incubation at 37°C, several colonies were observed growing on the plates, one of which was picked for further analysis. The presence of the insert in this mutant was confirmed by PCR with combinations of primers flanking the pST1 polylinker and primers complementary to sequences upstream and downstream of the comE gene. Analyses of the resulting PCR products by agarose gel electrophoresis showed that the pST1 plasmid had been intergrated into the comE gene as expected.

Computer-aided analyses.

Database searches were performed by using the programs BLASTN and BLASTP (2) on the NCBI server (http://www.ncbi.nlm.nih.gov/blast/). DNA sequence alignments were performed by using CLUSTAL W (32) with the BLOSUM62 (12) comparison matrix and with the following parameters: open gap cost, 10; and gap extension cost, 0.2. Dot plot analysis was performed by using the OMIGA 2.0 package (Oxford Molecular).

Nucleotide sequence accession number.

The nucleotide sequences have been deposited in the GenBank database under accession no. AF498313 and AF498314.

RESULTS

Identification and characterization of a naturally competent streptococcal isolate lacking the comC gene.

An ongoing project in our laboratory has been to determine the sequence of a large number of comC alleles to get an overview of the diversity of competence pheromones produced by different isolates of streptococci. The first step in this process is to purify chromosomal DNA from the isolate to be examined and then use this DNA as a template for PCR to amplify a fragment corresponding to the comCDE operon. One of the natural isolates examined (Atu-4) gave rise to a PCR product that was ca. 150 bp shorter than the 2.7-kb comCDE fragment obtained with other strains. Sequencing of this PCR fragment revealed that the comC gene is missing in the Atu-4 strain, whereas the comDE genes and the upstream promoter region are intact (Fig. 1). To our surprise, transformation assays carried out with the Atu-4 strain, with chromosomal DNA from an Atu-4 mutant selected for streptomycin resistance, gave a high number of transformants. This finding suggests that the Atu-4 strain has a functional comC gene located elsewhere on its chromosome or that this strain somehow circumvents the ComC dependent quorum sensing mechanism essential for competence development in other streptococci.

FIG. 1.

Organization of the comCDE competence regulation operon in strain SK348 (A) and the corresponding operon in strain Atu-4 (B). Both operons are flanked by genes encoding Arg-tRNA and Glu-tRNA. The nucleotide sequences shown in detail below the arrows representing open reading frames in panels A and B represent the binding site of ComE. The direct repeat motif is underlined. The amino acid sequence of the CSP precursor (ComC) is shown above the diagram of open reading frames in panel A. The amino acid sequence of the mature secreted competence pheromone (CSP) is underlined.

To further characterize the Atu-4 strain, we carried out a Rapid ID32 Strep test and determined the sequence of its 16S rRNA. The Rapid ID32 Strep test identified the strain as S. oralis (results not shown), whereas BLAST searches with the Atu-4 16S rRNA sequence revealed the highest homology (99.1% identity) to the corresponding sequence from S. infantis (20). Phenotypic characteristics of the latter recently described species are not included in the Rapid ID32 database. Furthermore, no phenotypic traits allowing differentiation of S. infantis from other mitis group streptococci have been identified (20). The most likely identity of the Atu-4 strain is therefore S. infantis. We also searched in gene banks and among our own unpublished sequences to identify the histidine kinase and response regulator most similar to ComD and ComE from Atu-4. The closest homologs were from a streptococcal isolate termed SK348. The most likely identity of this strain is also S. infantis, since its 16S rRNA sequence shows highest homology to the corresponding sequence from the type strain of S. infantis (98%). The SK348 strain has an intact comCDE operon and will autoinduce to competence in a cell density-dependent manner when grown in liquid medium. An alignment of the 250-amino-acid ComE response regulators from SK348 and Atu-4 showed that they differ only in three positions as follows: at position 51, leucine (SK348) versus valine (Atu-4); at position 77, histidine (SK348) versus arginine (Atu-4); and at position 130, asparagine (SK348) versus serine (Atu-4). Since all substitutions are situated in the receiver domain of the ComE response regulators, they should recognize the same or very similar direct repeat motifs. A comparison of the ComE binding sites in the promoter regions of the comCDE and comDE operons (Fig. 1) shows that they are practically identical. An alignment of the ComD receptors of the two strains showed that they are 80% identical (Fig. 2) and that the majority of the amino acid substitutions are located among the 80 to 100 N-terminal residues. This difference in sequence homology is also evident at the nucleic acid level. The dot plot in Fig. 3 shows that the first 280 bp of the comD genes from SK314 and Atu-4 are much less conserved than the rest of these genes. This mosaic-like structure strongly suggests that an interspecies recombinational exchange has taken place, presumably by natural transformation. It is possible that this event also caused the loss of comC from the genome of the Atu-4 strain.

FIG. 2.

Sequence alignment of the ComD histidine kinase receptors from Atu-4 and SK348. Amino acid substitutions are shown in a black background.

FIG. 3.

Comparison of the nucleic acid sequences of the comD genes from the Atu-4 and SK348 strains. Each comD gene consists of 1,320 bp. The comparison is visualized graphically as a dot plot.

S. infantis strains SK348 and Atu-4 share many properties. They are phylogenetically quite closely related, they have the same morphology (chains consisting of about 50 bacterial cells), and their ComDE proteins are highly homologous. For this reason and since the SK348 strain behaves normally in natural transformation, we decided to use SK348 as a reference in the further characterization of Atu-4.

Regulation of competence development in the Atu-4 and SK 348 strains.

In vigorously growing pneumococcal cultures, autoinduction of competence has been shown to take place when the culture reaches a critical cell density (34). Various laboratories have reported critical cell densities varying between 2 × 10⁶ and 2 × 10⁸ cells/ml (5). The critical cell density for autoinduction is reproducible from one experiment to another but varies somewhat with medium composition, bacterial strain, and other unknown variables. To determine whether competence development in the Atu-4 strain is cell density dependent, transformation experiments with this strain were carried out at different population densities. The SK348 strain was run in parallel as a positive control. The results of these experiments, given in Table 1, show that the Atu-4 and SK348 strains behave very differently. A high transformation efficiency (9%) was obtained with the Atu-4 strain even in very dilute cultures containing as little as 1,000 CFU/ml. As the Atu-4 and SK348 strains grow in long chains containing ca. 50 cells each, 1,000 CFU/ml corresponds to about 5 × 10⁴ cells/ml. A population density of ca. 7.9 × 10⁶ CFU/ml (i.e., 4 × 10⁸ cells/ml) is needed to get a similar level of transformation efficiency with the SK348 strain, even though a low number of transformants are observed at a population density of 1.3 × 10⁷ cells/ml (Table 1). Competence development in the SK348 strain follows the normal pattern where the ability to take up naked DNA from the environment is induced at a cell density of 1.3 × 10⁷ to 4 × 10⁸ cells/ml and is then turned off again a few hours later. In contrast, the competent state in the Atu-4 strain seems to be constitutive and not induced by cell density. The transformation efficiency is more or less at the same level between 5 × 10⁴ and 2.3 × 10⁹ cells/ml. Above this cell density, competence in the Atu-4 strain is shut off. The unusually high transformation efficiency obtained with Atu-4 and SK348 compared to S. pneumoniae is probably due to the presence of long chains, which interferes with the calculation of transformation efficiency.

Is the Atu-4 strain competent without a competence pheromone?

It is possible that an additional copy of the comDE operon, containing an intact comC gene, exists at a separate locus in the Atu-4 genome. To examine this possibility, Southern blots were conducted with a 1,245-bp DNA probe covering the C-terminal half of ComD and most of ComE. Genomic DNA from stain Atu-4 was digested with three different restriction enzymes, followed by separation on agarose gel and transfer to a nylon membrane. The results did not indicate that additional comDE genes are present in the Atu-4 genome, since only one band was detected on the nylon filter for each restriction enzyme used (data not shown).

In contrast to streptococci from the mitis and anginosus groups, the comC gene of S. mutans is not cotranscribed with the comDE genes but is situated on a separate operon (25). To exclude the possibility that a comC gene encoding a functional competence pheromone (CSP) could be located elsewhere on the Atu-4 genome, we decided to investigate whether competence development in this strain is inhibited by the presence of trypsin. Trypsin cleaves a polypeptide chain behind the basic amino acids arginine and lysine. Since CSPs of streptococci from the mitis and anginosus phylogenetic groups contain an essential arginine residue at position number three from the N-terminal end and additional important arginine residues at the C-terminal end, these pheromones are very susceptible to inactivation by trypsin (34, 22). This susceptibility to trypsin is demonstrated in the results presented in Fig. 4, which show that competence development in a growing culture of SK348 is completely abolished by the presence of 0.02 mg of the protease/ml in the medium. In contrast, transformation of the Atu-4 strain is not significantly affected by the presence of trypsin, strongly indicating that competence development in this strain is not induced by CSP.

FIG. 4.

Effect of trypsin on competence development in strains Atu-4 and SK348. When no trypsin was added to cultures of Atu-4 (•) and SK348 (▴), the number of Sm^r CFU obtained varied with growth phase as expected. However, when trypsin was added to the SK348 culture (▵) (see Materials and Methods) competence was completely abolished. In contrast, competence development in the Atu-4 strain is unaffected by trypsin (○).

Competence in strain Atu-4 depends on a functional comE gene.

Since the competent state in Atu-4 most likely is independent of CSP and seems to be constitutive at levels below ca. 2.3 × 10 ⁹ cells/ml, we hypothesized that a gain-of-function mutation had taken place in one of the key regulatory proteins of natural competence. The most obvious candidates for such a mutation are ComD, ComE, or ComX. To determine whether the signal transduction pathway containing these regulatory proteins is still needed for competence induction in the Atu-4 strain, we decided to disrupt the comE gene by using insertion-duplication mutagenesis. Transformation experiments performed with this mutant (comE::pST1) clearly showed that competence was completely abolished (data not shown). This experiment demonstrated that a functional comE gene product is needed for competence in the Atu-4 strain and that the signal transduction pathway downstream of ComE is intact. It also follows from this experiment that the unusual phenotype of the Atu-4 strain must be due to a gain-of-function mutation in ComD or ComE.

DISCUSSION

In naturally competent streptococci belonging to the mitis group a quorum-sensing mechanism controlled by CSP regulates competence development. In all likelihood, the naturally competent Atu-4 strain described here lacks a CSP, and thus competence induction in this strain cannot be regulated by cell density. In accordance with this finding, competence in the Atu-4 strain seems to be a constitutive property, except at high cell densities where the competent state is shut off. How can competence in the Atu-4 strain be turned on without the interaction between CSP and its histidine kinase receptor ComD? Previous work on natural competence in Bacillus subtilis and S. pneumoniae suggests that deletions or point mutations introduced into the pheromone receptor histidine kinase or its cognate response regulator can result in uncoupling of competence development from cell density control. In B. subtilis, competence development is partially regulated by the peptide pheromone ComX, the histidine kinase ComP, and the response regulator ComA, which monitor the density of the bacterial population in essentially the same way as ComCDE in S. pneumoniae (27). When deletions are introduced in the N-terminal part of the membrane domain of ComP, it becomes constitutively active and independent of its ligand ComX (29). In S. pneumoniae laboratory-made point mutations in ComD or ComE yielded mutant strains with similar phenotypes (7, 23, 26). Judging from these findings and the from fact that the Atu-4 ComE knockout mutant is completely noncompetent, we consider it likely that the constitutive phenotype seen in Atu-4 is caused by mutations in ComD and/or ComE. As mentioned above, only three amino acid substitutions separate the ComE proteins of the Atu-4 and SK348 strains. Two of these substitutions—leucine-51 (SK348)→valine (Atu-4) and asparagine-130 (SK348)→serine (Atu-4)—are probably not responsible for the constitutive phenotype of Atu-4, since valine and serine are commonly found in these positions in ComE alleles from closely related streptococci. In contrast, searches in sequence databases did not identify any ComE alleles that have an arginine in position 77, suggesting that the histidine (SK348)→arginine (Atu-4) substitution is unique and therefore might be responsible for the constitutive phenotype of the Atu-4 strain. The comparison of the ComD alleles from Atu-4 and SK348 depicted in Fig. 2 shows that the two proteins differ in 90 of 439 amino acid positions. Most of these substitutions (i.e., 77 of them) are located in the membrane domain, which is often referred to as the input domain. Previous studies of the CSP receptor of S. gordonii have shown that the first 80 to 100 amino acids at the N-terminal end of ComD determines pheromone specificity and consequently must be important for CSP binding. The alignment in Fig. 2 shows that the highest density of amino acid substitutions is located among the 100 N-terminal residues. We therefore suspect that changes in this important region somehow mimic ligand binding and switch ComD to a constant on-mode, causing the constitutive phenotype observed in the Atu-4 strain. Further research is needed, however, to clarify these matters and to identify the exact mutations involved.

Even though induction of competence is no longer regulated by cell density, the competence shutoff mechanism is intact in the Atu-4 strain. In S. pneumoniae, the most studied species of naturally transformable streptococci, the competent state lasts for only 40 to 60 min after induction (15). Apparently, pneumococci are a special case; as in other strains and species of streptococci from the mitis group, e.g., SK348, the competent state can last for several hours (10). The mechanism that shuts down the competent state in streptococci is not understood, but Lee and Morrison (24) obtained data suggesting that one of the genes upregulated by the alternative sigma factor ComX, i.e., one of the late genes, effects shutdown of competence by some kind of feedback mechanism. In contrast to pneumococci, the Atu-4 strain is continuously competent for a period of at least 8 h. If the feedback mechanism proposed by Lee and Morrison effects shutdown in the Atu-4 strain, it must act extremely slowly. To us it seems more plausible that competence in the Atu-4 strain is shut down by an independent mechanism that senses some signal in the environmental that will trigger shutdown of the competent state. This signal could be population density monitored by a system different from ComABCDE or by monitoring the concentration of particular nutrients or metabolites in the growth medium.

The Atu-4 strain was isolated in 1999 from the tongue of a young healthy adult. After isolation and characterization of the Atu-4 strain, a nearly identical strain, lacking comC and having the same competence phenotype as Atu-4, was discovered. This strain, designated E20, was isolated from a child with bacteremia in 1995. Partial sequencing of the comDE operon of this strain revealed that, except for a few base substitutions, the sequence was identical to the corresponding sequence from Atu-4. These findings indicate that the constitutive competence phenotype observed in these strains is not strongly selected against in nature. In most naturally transformable bacteria, the competent state is transient. The exceptions are Neisseria gonorrhoeae and N. meningitidis, which both are unusual in maintaining competence throughout the growth cycle (31). However, in contrast to streptococci, which will take up DNA from any source, the two Neisseria species show preferential uptake of homologous DNA. Their DNA uptake apparatus recognizes a 10-bp uptake signal sequence distributed throughout the genomes of the Neisseria spp. and will not efficiently take up DNA lacking these conserved sequence motifs (11, 30). It is widely believed that indiscriminate uptake of foreign DNA by a bacterium could have negative consequences. It could, for instance, increase the risk of recombination events resulting in gene disruption or genes encoding nonfunctional gene products. As described above, a quorum-sensing mechanism controlled by a strain-specific peptide pheromone (CSP) regulates competence development in streptococci. Most likely, the purpose of CSP specificity and quorum sensing in this context is to monitor the concentration of potential gene donors in the immediate neighborhood. Thus, the CSP-controlled cell density monitoring mechanism regulating competence development in streptococci most likely serves the same function as DNA uptake signal sequences in Neisseria spp., namely, to increase the likelihood that the DNA taken up during competence comes from closely related bacteria. It was therefore unexpected to find that strains with CSP-independent constitutive DNA uptake could survive in competition with CSP-dependent “wild-type” strains in nature. It is possible that these strains have evolved compensatory mechanisms, for example, a more efficient mismatch repair system (13), to reduce the potentially harmful effects of taking up large amounts of foreign DNA. On the other hand, from a nutritional point of view it could be an advantage for the Atu-4 and E20 strains to take up all available DNA in the environment. It could support bacterial growth by being used as a source of nucleotides and maybe even as a source of carbon and energy (9). In sum, our findings suggest that any negative effects of indiscriminate DNA uptake must be more than compensated for by the advantages of being able to take up naked DNA from the environment. Otherwise, natural competence would not have been maintained in the Atu-4 and E20 strains.