Choline-Binding Protein D (CbpD) in Streptococcus pneumoniae Is Essential for Competence-Induced Cell Lysis

Streptococcus pneumoniae is an important human pathogen thatis able to take up naked DNA from the environment by a quorum-sensing-regulatedprocess called natural genetic transformation. This propertyenables members of this bacterial species to efficiently acquirenew properties that may increase their ability to survive andmultiply in the human host. We have previously reported thatinduction of the competent state in a liquid culture of Streptococcuspneumoniae triggers lysis of a subfraction of the bacterialpopulation resulting in release of DNA. We have also proposedthat such competence-induced DNA release is an integral partof natural genetic transformation that has evolved to increasethe efficiency of gene transfer between pneumococci. In thepresent work, we have further elucidated the mechanism behindcompetence-induced cell lysis by identifying a putative mureinhydrolase, choline-binding protein D (CbpD), as a key componentof this process. By using real-time PCR to estimate the amountof extracellular DNA in competent relative to noncompetent cultures,we were able to show that competence-induced cell lysis andDNA release are strongly attenuated in a cbpD mutant. Ectopicexpression of CbpD in the presence or absence of other competenceproteins revealed that CbpD is essentially unable to cause celllysis on its own but depends on at least one additional proteinexpressed during competence.

INTRODUCTION

Top

Introduction

Streptococcus pneumoniae and related streptococci have the abilityto take up exogenous DNA for potential genomic transformation.Mapping has shown that this intrinsic property is not constitutivebut regulated by the competence-stimulating peptide (CSP) andits cognate signal transduction pathway, ComDEX (13, 14, 17,25). ComD, which is a transmembrane histidine kinase, monitorsthe concentration of CSP outside the cytoplasmic membrane. Atthe critical concentration of CSP, which is affected by thegrowth conditions used, the competent state is turned on (7).

Recent mapping of CSP-responsive genes in Streptococcus pneumoniaeby use of DNA microarrays has shown that eight early genes and14 late genes are required for transformation (26). The sevenproteins encoded by the essential early genes include the CSPsecretion apparatus (ComAB), the precursor of CSP (ComC), thetwo-component regulatory system ComDE, the alternative sigmafactor ComX (encoded by two identical genes, comX1 and comX2),and the newly discovered positive regulator of competence termedComW (18). The 14 essential proteins encoded by the late genesare members of the ComX regulon and include DNA-processing proteinsplus the components of the DNA uptake machinery. Interestingly,the 22 genes that are indispensable for transformation constituteonly about 18% of the CSP-inducible genes in Streptococcus pneumonia(26). This finding immediately raises the question of whetherCSP could regulate other biological functions in addition tonatural transformation.

Several studies have reported that ComD mutants are attenuatedwith respect to pathogenicity, indicating that at least oneof the CSP-responsive genes contributes to virulence (12, 16,19). It is also possible that some of the CSP-responsive genesare involved in other aspects of horizontal gene transfer thanuptake and integration of incoming DNA. We have previously shownthat a subfraction of the cell population lyse when a liquidculture of pneumococci is induced to competence by additionof CSP (35, 36). The biological significance of this phenomenonis not clear, but an attractive hypothesis is that some cellslyse to release DNA that can be taken up by the others. Themolecular mechanism behind the observed cell lysis remains tobe elucidated, but previous studies strongly indicate that choline-bindingproteins are involved (35). Choline-binding proteins containa number of short choline-binding repeats which anchor themnoncovalently to teichoic and lipoteichoic acid in the pneumococcalcell wall (9, 33). Here we show that CbpD, a CSP-inducible choline-bindingprotein and putative murein hydrolase, plays a key role in competence-inducedcell lysis.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains and growth conditions. The strains used in this study were derived from S. pneumoniaestrain Rx (see Table 1). Bacterial strains were grown in CATmedium or on CAT agar plates containing (per liter) 167 mmolof K₂HPO₄, 5 mg of choline chloride, 5 g of tryptone, 10 g ofenzymatic casein hydrolysate, 1 g of yeast extract, and 5 gof NaCl (22). After sterilization, glucose was added to a concentrationof 0.2%.

View this table:
[in this window]
[in a new window]
TABLE 1. S. pneumoniae strains used in this study

Construction of S. pneumoniae mutants. To investigate the expression profile of cbpD during competenceinduction, the S. pneumoniae L1 mutant strain was constructed(Table 1). First, a 400-bp internal fragment of cbpD was amplifiedfrom the CP1415 strain using the primers cbpDrapFor (5'-GAGAGAGCATGCGGAAGGTTATCGTGTAGATAATACACCG-3')and cbpDrapBack (5'-CTCTCTGGATCCGGCAGTATAACTCAACCACTTATAGCCG-3').Then the fragment was digested with BamHI and SphI and ligatedinto the corresponding restriction sites of the nonreplicativepEVP3 vector (6), resulting in a transcriptional fusion betweenthe cbpD gene and the pEVP3 lacZ. The resulting vector was subsequentlyused to transform the S. pneumoniae EK100 strain (Table 1) bynatural transformation. A single transformant, termed L1, wasisolated from an agar plate containing 2.5 µg/ml chloramphenicol.Correct integration of the recombinant vector was confirmedby employing PCR with flanking primers followed by DNA sequencingof the PCR product.

Strain EK4166 was used to construct the cbpD deletion mutantEK4168 (Table 1). A 953-bp DNA fragment corresponding to theregion immediately upstream of cbpD was amplified using theprimers cbpD.5 (5'-ATTAGACGTCCTGACATCTTCAAGTAATAACTTGTTCCT-CTTGC-3')and cbpD.1 (5'-ATTAGCTAGCGCTGCAGAAGGCTTGGTGGAGTCAC-3') and ligatedinto pCR2.1-TOPO (Invitrogen) according to the manufacturer'srecommendations. The fragment was excised from the pCR2.1TOPOvector with AatII and NheI and ligated into multiple cloningsite I of the nonreplicative vector pFW13 (27). Next, a 766-bpfragment corresponding to the region downstream of cbpD wasamplified using the primers cbpD.3 (5'-ATTACCATGGGGCTGGTACCAGGTCTCTGGTGAGTG-3')and cbpD.4 (5'-ATTACATATGGCCACTCTCAAGGTCGCCCAGCTATGG-3') andcloned into a pCR2.1TOPO vector. The fragment was excised frompCR2.1TOPO with NcoI and SpeI and cloned into multiple cloningsite II of the pFW13 vector harboring the 953-bp fragment. Usingthis construct as a template and the oligonucleotides cbpD.1and cbpD.4 as primers, a 3,300-bp fragment consisting of theKan^r gene of pFW13 flanked by the two cloned fragments describedabove was amplified. This linear DNA fragment was then usedto transform EK4166 by natural transformation. A transformantgrowing on agar plates containing 100 µg/ml kanamycinwas isolated, and correct insertion of the Kan^r cassette wasverified by PCR and DNA sequencing.

cbpD deletion mutant L3, used in the quantitative PCR experiments,was made by transforming strain CP1415 with the 3,300-bp PCRfragment described above. The L5 mutant was constructed in orderto compare the amount of DNA release mediated by CbpD when thisprotein was expressed together with or separate from the othercompetence proteins. First, an 842-bp DNA fragment encompassingthe inducible p1 promoter of the qsrAB operon plus the ${approx}$ 200-bpBOX element immediately downstream of the p1 promoter (15) wasamplified by PCR using the primers Box1 (5'-ATATAGGATCCTATACTCAATGAAAATCAAAGAGC-3')and 4144bgal2 (5'-ATTAAATGCATCTCTGATGGACGTAATTTATGGCTAG-3').This fragment was digested with BamHI and NsiI and ligated intothe corresponding sites of the pEVP3 vector, giving rise tothe vector pEVP3qsrI.

To obtain a p1 promoter that would respond equally well to inductionby CSP-1 and BIP-1, we introduced two point mutations in theright repeat of this promoter. This was done as described byKnutsen et al. (15) for the construction of the OE4145 mutantstrain. Then, a unique SpeI restriction site was introducedin pEVP3qsrI close to the 3' end of the cloned 842-bp fragment,using the primers mutSpeIF (5' CCCCAGCTTGAATTGATACACTAGTGCTTTTATATAGGGAAAAGGTGG-3')and mutSpeIB (5'-CCACCTTTTCCCTATATAAAAGCACTAGTGTATCAATTCAAGCTGGGG-3')according to the Quick Change PCR mutagenesis method (Stratagene).Next, the cbpD gene was amplified from genomic CP1415 DNA, usingthe primers cbpDBDraIII (5'CTCTCTCACAGCGTGCTATACTCGTTCTCCATCGTTTACTC-3')and cbpDFspeI (5'-AGAGAGACTAGTCGCTTTATATTATTTTTTAAGGAGGAAGAATG-3').This fragment was digested with DraIII and SpeI and ligatedinto the corresponding sites of the mutagenized pVP3qsrI vectordescribed above. The resulting vector, pVP3qsrII, was transformedinto the L3 strain by natural transformation. Correct chromosomalintegration was verified by DNA sequencing.

Assay of ß-galactosidase activity. Competence-induced expression from the cbpD promoter was monitoredusing the L1 mutant strain. An overnight culture of L1 was dilutedin CAT medium to an OD₅₅₀ of 0.05. The culture was grown at37°C. At selected optical densities, 1-ml samples were collectedand incubated at 30°C for 10 min before addition of 250ng/ml CSP. The tubes were subsequently incubated at 30°Cfor 30 min and cells were lysed by addition of 0.1% (vol/vol)Triton X-100. Incubation was carried out at 30°C becausecompetence-induced cell lysis is less effective at 37°C(36). Parallel samples containing uninduced cells were collectedas described above; 200 µl cell lysate was added to 250µl 5x Z buffer [5 mM MgCl₂, 250 mM ß-mercaptoethanol,50 mM KCl, 0.3 M Na₂HPO₄ · 7H₂O, 0.2 M NaH₂PO₄ ·H₂O, 4 mg/ml o-nitrophenyl-ß-D-galactopyranoside (ONPG),pH 7.0] and 760 µl CAT medium, and subsequently incubatedat 30°C for 1 h. The ß-galactosidase reactionwas stopped by addition of 500 µl 1 M Na₂CO₃, and hydrolysisof ONPG was recorded in a spectrophotometer at 420 nm. Enzymeactivity was calculated according to the method of Miller (20).ß-Galactosidase release assays were performed as describedby Steinmoen et al. (35), except that competence induction wasperformed at 30°C.

Real-time PCR assays. Overnight cultures were diluted to an OD₅₅₀ of 0.05 in CAT mediumand grown at 37°C to an OD₅₅₀ of 0.3. Then a second dilutionto an OD₅₅₀ of 0.1 was performed. When the cultures reachedan OD₅₅₀ of 0.3, 0.5 µg/ml of salmon sperm DNA and 15pg/ml of the plasmid pLNO2HM1 (24) harboring a variant of thehm1 gene (4) were added. The cultures were next transferredto 30°C and treated with either 250 ng/ml CSP-1, 250 ng/mlBIP-1, or kept untreated. After 30 min of incubation, 1.5-mlsamples were collected in microtubes and centrifuged at 4°Cfor 2 min at 15,000 x g. Then the supernatants were sterilefiltered (pore size of 0.2 µm) into cooled cryotubes andstored at –80°C. Real-time PCR was performed on thesupernatant fractions in triplicate reactions, using specificoligonucleotides for the pneumolysin gene (ply) and the hm1gene (Table 2), and the TaqMan universal PCR master mix accordingto the manufacturer's instruction (User Bulletin 2, AppliedBiosystems, Foster City, USA). Real-time PCR was performed withthe ABI Prism 7700 sequence detection system (Applied Biosystems,Foster City, CA). Standard curves were obtained for both plyand hm1 to verify that the amplification efficiency was similarin order to apply the 2^–Ct method (Applied Biosystems)for the analysis. Plasmid DNA (hm1) was used as an internalcontrol for data normalization. The amount of released DNA wasexpressed as the n-fold difference relative to the untreatedculture, and the final value represents the mean of three independentexperiments.

View this table:
[in this window]
[in a new window]
TABLE 2. Primers and probes used in TaqMan real-time PCR analysis^a

RESULTS

Top

Results

CbpD is highly expressed during competence. As described above, we have previously reported that an unidentifiedcholine-binding protein must be important for competence-inducedcell lysis (35, 36). Since CbpD is a competence-specific choline-bindingprotein that is dispensable for transformation, we consideredit a likely candidate for this hitherto-unidentified componentof the lysis mechanism. Previous studies using microarray analysishave shown that transcription of the cbpD gene increases uponcompetence induction and that cbpD belongs to the late competencegenes (26, 32). To verify this, and to determine the basal levelof transcription in noncompetent cells, we inserted a lacZ reportergene downstream of the cbpD promoter. Our results show thattranscription of the cbpD gene is very low or possibly absentin noncompetent cells, but is highly induced during competence(Fig. 1). The level of ß-galactosidase expressionis somewhat reduced at an OD₅₅₀ of 0.5 compared an OD₅₅₀ of0.4. This is in accordance with the fact that an increasingfraction of the pneumococcal population becomes refractory toCSP at high cell density (13). Interestingly, we and othershave found that the efficiency of transformation is not reducedin a cbpD mutant (26) (unpublished results). Thus, even thoughCbpD clearly is a competence-specific protein, it is not requiredfor DNA binding, processing, uptake, or recombination.

View larger version (9K):
[in this window]
[in a new window]
FIG. 1. cbpD promoter activity in competent and noncompetent cells. The L1 mutant was collected at various cell densities (OD₅₅₀ of 0.1 to 0.5) and induced to competence by addition of 250 ng/ml of synthetic CSP-1. After 30 min at 30°C, cultures were lysed and assayed for ß-galactosidase activity. Corresponding uninduced samples were run in parallel. Solid squares, cell lysates from cells induced with CSP-1; open squares, cell lysates from uninduced cells. The data presented are representative of results from three independent experiments. Strain L1 is comA egb cbpD::pEVP3.

Deletion of cbpD strongly reduces competence-induced cell lysis and DNA release. To directly test our hypothesis that CbpD is important for competence-inducedcell lysis, we deleted the cbpD gene from the EK4166 strainand compared the resulting ${Delta}$ cbpD mutant (EK4168) to the parentalstrain. Experimentally, cell lysis was detected as release ofcytoplasmic ß-galactosidase to the growth medium.The EK4166 strain is negative for endogenous ß-galactosidaseactivity but harbors a chromosomally located transcriptionalfusion between the hirL gene (GenBank accession no.AAL00764)and the Escherichia coli lacZ reporter gene (35). Since thehirL gene is highly expressed throughout the exponential growthphase, EK4166 and EK4168 cells have a high cytoplasmic contentof ß-galactosidase during this period. In addition,the two strains lack a functional CSP transporter (ComA) andcan for this reason develop the competent state only if syntheticCSP is added to the culture.

To compare the properties of the ${Delta}$ cbpD mutant (EK4168) and itsparental strain (EK4166), we induced competence at cell densitiesranging from an OD₅₅₀ of 0.1 to 0.5 and measured the amountof ß-galactosidase released to the growth medium after30 min. The result of a representative experiment is presentedin Fig. 2. It shows that almost no ß-galactosidaseactivity was detected in sterile filtered culture supernatantsfrom the ${Delta}$ cbpD mutant, demonstrating that deletion of the cbpDgene has a dramatic effect on competence-induced cell lysis.

View larger version (10K):
[in this window]
[in a new window]
FIG. 2. Comparison of competence-induced release of cytoplasmic ß-galactosidase from wild-type S. pneumoniae cells and a mutant lacking a functional cbpD gene. Bacterial cultures of strains EK4166 and EK4168 were collected at different cell densities (OD₅₅₀ of 0.1 to 0.5) and induced to competence by addition of 250 ng/ml of synthetic CSP-1. After 30 min at 30°C, culture supernatants were harvested by centrifugation followed by sterile filtration. The supernatants were subsequently assayed for ß-galactosidase activity. Solid squares, supernatants from EK4166 cells; open squares, supernatants from EK4168 cells. The experiment has been repeated several times with similar results. Genotypes of strains: EK4166, comA egb hirL::pEVP3; and EK4168, comA egb hirL::pEVP3 cbpD.

Competence-induced cell lysis is believed to increase the efficiencyof horizontal gene exchange by providing extracellular DNA thatcan be taken up by competent recipient cells. We therefore consideredit important to develop an assay that would allow us to directlymeasure DNA release from lysing cells in a competent cultureof pneumococci. The obvious choice was to employ real-time PCR,a technique that makes it feasible to accurately compare theamount of extracellular DNA in competent and noncompetent cultures.

The primers and probes used for real-time PCR were directedagainst the pneumolysin gene of S. pneumoniae. Our results showthat the content of extracellular DNA in a competent cultureof strain CP1415 (comA) is more than a hundredfold higher thanin a noncompetent culture of the same strain (Fig. 3). Sampleswere withdrawn from the cultures 30 min after addition of CSPin order to make the measurement within the period of competence.Transformation reaches a maximum 20 to 25 min after the additionof CSP and then gradually decays during the following 20 to30 min (13). In a corresponding experiment, carried out withthe ${Delta}$ cbpD mutant strain termed L3, the results revealed onlya fivefold increase in extracellular DNA content in competentcultures relative to their noncompetent counterparts (Fig. 3).This drastic reduction in CSP-induced release of DNA is in goodagreement with the data presented in Fig. 2. Together the resultsobtained with real-time PCR and the ß-galactosidaserelease assay clearly demonstrate that CbpD is a key componentof the lysis mechanism.

View larger version (17K):
[in this window]
[in a new window]
FIG. 3. Quantitative comparison of competence-induced DNA release in wild-type and ${Delta}$ cbpD cells. When the bacterial cultures reached an OD₅₅₀ of 0.3, 15 pg/ml control DNA (pLNO2HM1) and 0.5 µg/ml salmon sperm DNA were added. The cultures were immediately split in two, and one half of the culture was induced to competence by addition of 250 ng/ml of synthetic CSP while the other was kept untreated. After 30 min of incubation at 30°C, supernatants were collected by centrifugation and sterile filtered. Subsequently, the collected supernatants were used in a real-time PCR assay with primers and probes directed against the pneumolysin gene or the hm1 gene. The amount of released DNA in untreated (noncompetent) cultures was set to 1. Data are plotted as the means of three independent experiments with error bars representing the standard deviations. Genotypes of strains: CP1415, comA; and L3, comA cbpD.

Ectopic expression of CbpD. As removal of CbpD almost abolished competence-induced celllysis, we considered it of interest to determine if ectopicexpression of CbpD in noncompetent cells would provoke celllysis at a level comparable to that observed for competent CP1415cells (see Fig. 3). In other words, is CbpD the only gene productexpressed during competence that is required for cell lysis,or are additional competence specific gene products needed aswell?

To answer this question, we inserted the cbpD gene into thechromosome behind the inducible promoter of the qsrAB operon.The qsrAB genes (accession no. AAL00365 and AAL00366) encodean ABC-transporter of unknown function. Recently, we showedthat transcription from this promoter is induced by both CSP-1and BIP-1 (15). CSP-1 is the competence-stimulating peptideproduced by pneumococci possessing the comC-1 allele, such asthe S. pneumoniae strain Rx used in this study (28). The BIP-1peptide pheromone shares no sequence homology with CSP-1 andis not able to induce competence in S. pneumoniae strain Rx.It is part of a homologous but separate quorum-sensing system(BlpABCSRH) that regulates production of bacteriocins in strainRx (8, 30). The inducible direct repeat promoter upstream qsrAbinds both BlpR and ComE and can therefore be activated by additionof CSP-1 or BIP-1 to the growth medium.

To obtain the same level of CbpD expression with both peptidepheromones, two point mutations were introduced in the rightrepeat as previously described for the OE4145 mutant constructedby Knutsen et al. (15). The resulting strain (L5) was then usedto compare the amount of DNA release mediated by CbpD when thisprotein was expressed together with or separate from the othercompetence proteins. Our results clearly show that other proteinsexpressed during competence are needed in addition to CbpD toobtain DNA release comparable to that observed for the CP1415strain (Fig. 4). Only a very small, but probably significant,increase in DNA release was detected when CbpD was expressedalone. In sum, the data presented in Fig. 4 show that even thoughCbpD is a crucial component of the lysis mechanism, at leastone additional competence protein must be involved.

View larger version (15K):
[in this window]
[in a new window]
FIG. 4. Ectopic expression of CbpD. When cultures of the L3 and L5 mutants reached an OD₅₅₀ of 0.3, 15 pg/ml control DNA (pLNO2HM1) and 0.5 µg/ml salmon sperm DNA were added. The cultures were immediately split in three and induced with 250 ng/ml CSP-1, 250 ng/ml BIP-1, or kept untreated. After 30 min incubation at 30°C, supernatants were collected by centrifugation followed by sterile filtration. Subsequently, the collected supernatants were used in a real-time PCR assay with primers and probes directed against the pneumolysin gene or the hm1 gene. The amount of released DNA in untreated (noncompetent) cultures was set to 1. Data are plotted as the means of three independent experiments with error bars representing the standard deviations. Genotypes of strains: L3, comA cbpD; and L5, comA cbpD qsrA::cbpD.

LytA and LytC. We have previously reported that significantly less ß-galactosidaseis released from a lytA mutant during competence developmentthan from the wild-type positive control (36). The differencewas two- to fourfold depending on the density of the cell culturewhen competence was induced. In the present work, we used real-timePCR to directly compare the amount of pneumococcal DNA presentin the medium of competent and noncompetent cultures of a lytAmutant (H3). At an OD₅₅₀ of 0.3, the culture was split in two,and 250 ng/ml of CSP was added to one of them. After 30 minincubation at 37°C, samples were withdrawn, briefly centrifuged,sterile filtered, and stored at –80°C for subsequentreal-time PCR analysis. Our results show that there is about50-fold more DNA present in the growth medium of the competentH3 culture than in the noncompetent H3 culture run in parallel(Fig. 5). Thus, competence-induced cell lysis is reduced abouttwofold in a lytA mutant compared to the wild-type CP1415 strain(Fig. 3). In contrast, a lytC mutant (H5) analyzed in the sameway as H3 was not affected with respect to DNA release (Fig.5). The difference in extracellular DNA concentration betweencompetent and noncompetent cultures of the lytC mutant was foundto be more than a hundredfold, i.e., the same as for the wild-typeCP1415 strain (Fig. 3).

View larger version (21K):
[in this window]
[in a new window]
FIG. 5. Competence-induced DNA release in ${Delta}$ lytA and ${Delta}$ lytC mutants. Cultures of the H3 and H5 mutants were incubated at 37°C until they reached an OD₅₅₀ of 0.3. Then 15 pg/ml control DNA (pLNO2HM1) and 0.5 µg/ml salmon sperm DNA were added. The cultures were immediately split in two, transferred to 30°C, and induced with 250 ng/ml CSP or kept untreated. After 30 min incubation at 30°C, supernatants were collected by centrifugation and sterile filtered. Subsequently, the collected supernatants were used in a real-time PCR assay with primers and probes directed against the pneumolysin gene or the hm1 gene. The amount of released DNA in untreated (noncompetent) cultures was set to 1. Data are plotted as the means of three independent experiments with error bars representing the standard deviations. Genotypes of strains: H3, comA lytA::pEVP3 Nov^r; and H5, comA egb lytC::pEVP3.

DISCUSSION

Top

Discussion

In the present work, we have obtained strong evidence that CbpDis a key component of the competence-induced cell-lysis mechanism.Computer-aided analysis of the 448-amino-acid sequence of CbpDshows that it is composed of three different domains plus asignal sequence at the N-terminal end. Downstream of the signalsequence is a CHAP (cysteine, histidine-dependent amidohydrolase/peptidase)domain consisting of about 150 amino acids followed by two SH3domains and a choline-binding domain with three or four choline-bindingrepeats (1, 3, 31). The function of choline-binding domainshas been thoroughly studied. They are known to anchor proteinspossessing such domains to the cell wall of Streptococcus pneumoniaeby binding noncovalently to phosphorylcholine moieties linkedto teichoic and lipoteichoic acid (9, 33). Consequently, CbpDmust be considered an extracellular protein that is noncovalentlyattached to the pneumococcal cell wall. This is in agreementwith the results of Gosink et al. (11), who used specific antibodiesto detect CbpD in choline eluates from pneumococcal cells.

The CHAP family of enzymes consists predominantly of bacterialand phage amidases (1, 3, 31). Phylogenetic analysis has shownthat CbpD belongs to the AcmB/LytN subfamily of these enzymes(1). Members of this subfamily are mostly found among gram-positivebacteria and function as cell wall hydrolases. Consequently,it is reasonable to assume that the CHAP domain of CbpD functionsas a murein hydrolase as well. Work is in progress at our laboratoryto verify this experimentally. The function of bacterial SH3domains is not known. They are however often associated withCHAP domains and other cell wall-degrading enzymes. It is thereforetempting to speculate that SH3 domains are involved in attachingthese enzymes to their substrate by binding directly eitherto murein or to other cell wall components. Indeed, it has beenshown that the SH3 domain of lysostaphin targets this bacteriolyticenzyme to the cell wall of Staphylococcus aureus (2). Thus,available data on CbpD and its CHAP, SH3, and choline-bindingdomains indicate that CbpD is a cell wall-anchored murein hydrolasethat functions as a competence-specific lysin.

Interestingly, our results strongly indicate that overexpressionof CbpD alone is not sufficient to obtain effective cell lysisand that expression of additional competence-induced genes isrequired (Fig. 4). Transcriptome analyses have shown that CSPinduces the expression of a large number of proteins apparentlydispensable for DNA binding, uptake, and recombination. It istherefore likely that one or more of these proteins are partof the lysis mechanism. We have previously found that disruptionof the gene encoding the autolysin LytA reduce competence-inducedcell lysis two- to fourfold (36).

In the present paper, we used real-time PCR two quantify releaseof DNA in competence-induced and noncompetent cultures of pneumococcilacking a functional lytA gene. With this method, a twofoldreduction of DNA release was observed in the lytA mutant. Theseresults suggest that LytA acts together with CbpD to degradethe cell wall of cells undergoing lysis during competence. Asonly a fraction (5 to 20%) of the cells in a competent batchculture will lyse within the peak of competence, i.e., duringthe 30 min following addition of CSP (35), a mechanism mustexist that protects the majority of the cells in the populationor makes only a subfraction of the cells susceptible to lysis.This mechanism is not understood but may involve additionalproteins required to control the activity of CbpD and LytA.

Other important questions regarding the mechanism of lysis remainunanswered as well. Initially,, we hypothesized that competence-inducedrelease of DNA takes place by autolysis, in a process wheresome altruistic cells in the population commit suicide to provideDNA to the others. Recently, however, we discovered that heterolysis(the lysis of one bacterium brought about by another) takesplace in a competent S. pneumoniae culture and that heterolysisprobably requires cell-to-cell contact (36). This finding doesnot necessarily prove that our original hypothesis is wrong.It is possible that heterolysis is a secondary effect of autolysis,caused by murein hydrolases released from cells undergoing autolysis.At present, it is therefore not clear whether autolysis or heterolysisis the principal cell lysis mechanism operating in a competentpneumococcal cell culture. Hopefully, future elucidation ofthe exact function of CbpD will help solve this question andlead to a better understanding of the biological significanceof competence-induced cell lysis.

In a previous study, we used a ß-galactosidase releaseassay to show that disruption of the lytC gene has no effecton competence-induced cell lysis (36). The lytC gene encodesa choline-binding autolytic lysozyme that does not belong tothe CSP-responsive competence genes (10). In the present study,we employed real-time PCR to measure CSP-induced DNA releasefrom the lytC mutant. Again we were not able to detect any differencebetween the lytC mutant (H5) and the wild-type strain CP1415(Fig. 3 and Fig. 5). In contrast to our results, Moscoso andClaverys (23) found that inactivation of the lytC gene affectsDNA release to the same extent as inactivation of lytA. Theyalso found that CSP-induced DNA release is almost abolishedin a lytA lytC double mutant. Our results, on the other hand,strongly indicate that CbpD is a key component of the lysismechanism and that this putative murein hydrolase acts togetherwith LytA to degrade the pneumococcal cell wall.

At present, there is no obvious explanation for these conflictingresults. However, as pointed out by Moscoso and Claverys, itis possible that the observed discrepancy results from the useof different strains and media. Another possibility is thatthe conflicting results are due to different experimental designs.We have measured release of DNA or ß-galactosidasein samples collected 30 min after addition of CSP, i.e., withinthe period of competence. In contrast, Moscoso and Claverysmeasured DNA release at multiple time points for at least 3h post-competence induction, i.e., for several hours after thecompetent state had been shut off. They found that DNA releasein cultures that had gone through a period of competence continuedto increase throughout the exponential and into the stationaryphase. No such release was seen in parallel cultures that hadremained noncompetent during the experiment.

In our opinion, it is conceivable that different mechanismsare behind early and late DNA release. Early DNA release takesplace during the period of competence and is, according to ourresults, strongly dependent on CbpD. Cells that lyse and disintegrateduring this early period will shed much of their enzyme contentinto the environment, presumably including the autolysins LytAand LytC. Reminiscent of a chain reaction, it is possible thatLytA and LytC released in this way will attack and lyse newcells, leading to release of more autolysins, and so forth.This scenario is not pure speculation, as we have observed thata wild-type strain, which always undergoes autolysis in latestationary phase due to the activity of LytA, will provoke lysisof a lytA mutant strain when the two strains are grown in amixed culture (unpublished observation). This experiment clearlydemonstrates that pneumococcal cells that lyse and fall apartrelease components (presumably autolysins) that attack and lyseother cells in the same culture. Consequently, it is possiblethat late DNA release is mainly caused by autolysins that arecontinuously released by disintegrating cells. This would explainwhy early DNA release requires CbpD and not LytC, whereas lateDNA release depends mainly on the joint action of the two autolysinsLytA and LytC.

ACKNOWLEDGMENTS

This work was supported by grants from the Research Councilof Norway.

FOOTNOTES

* Corresponding author. Mailing address: Department of Chemistry, Biotechnology, and Food Science, Biotechnology Building, Norwegian University of Life Sciences, P.O. Box 5003, N-1432 Ås, Norway. Phone: 47-64965883. Fax: 47-64947720. E-mail: sigve.havarstein{at}umb.no.

REFERENCES

Top