Uptake of transforming DNA in Gram-positive bacteria: a view from Streptococcus pneumoniae

Authors

  • Mathieu Bergé,

    Corresponding author
    1. Laboratoire de Microbiologie et Génétique Moléculaire, UMR 5100 CNRS-Université Paul, Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex, France.
    • *For correspondence. E-mail claverys@ibcg.biotoul.fr; Tel. (+33) 5 61 33 59 11; Fax (+33) 5 61 33 58 86

      These authors contributed equally to this study.

    Search for more papers by this author
  • Miriam Moscoso,

    Corresponding author
    1. Laboratoire de Microbiologie et Génétique Moléculaire, UMR 5100 CNRS-Université Paul, Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex, France.
    • *For correspondence. E-mail claverys@ibcg.biotoul.fr; Tel. (+33) 5 61 33 59 11; Fax (+33) 5 61 33 58 86

      These authors contributed equally to this study.

    Search for more papers by this author
  • Marc Prudhomme,

    1. Laboratoire de Microbiologie et Génétique Moléculaire, UMR 5100 CNRS-Université Paul, Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex, France.
    Search for more papers by this author
  • Bernard Martin,

    1. Laboratoire de Microbiologie et Génétique Moléculaire, UMR 5100 CNRS-Université Paul, Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex, France.
    Search for more papers by this author
  • Jean-Pierre Claverys

    Corresponding author
    1. Laboratoire de Microbiologie et Génétique Moléculaire, UMR 5100 CNRS-Université Paul, Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex, France.
    • *For correspondence. E-mail claverys@ibcg.biotoul.fr; Tel. (+33) 5 61 33 59 11; Fax (+33) 5 61 33 58 86

      These authors contributed equally to this study.

    Search for more papers by this author

Summary

In a working model for the uptake of transforming DNA based on evidence taken from both Bacillus subtilis and Streptococcus pneumoniae, the ComG proteins are proposed to form a structure that provides access for DNA to the ComEA receptor through the peptidoglycan. DNA would then be delivered to the ComEC–ComFA transport complex. A DNA strand would be degraded by a nuclease, while its complement is pulled into the cell by ComFA through an aqueous pore formed by ComEC. The nuclease is known in S. pneumoniae only as EndA. We have examined the processing (i.e. binding, degradation and internalization) of DNA in S. pneumoniae strains lacking candidate uptake proteins. Mutants were generated by transposon insertion in endA, comEA/C, comFA/C, comGA and dprA. Processing of DNA was abolished only in a comGA mutant. As significant binding was measured in comEA mutants, we suggest the existence of two stages in binding: surface attachment (abolished in a comGA mutant) required for and preceding deep binding (by ComEA). Abolition of degradation in comGA and comEA mutants indicated that, despite its membrane location, EndA cannot access donor DNA by itself. We propose that ComEA is required to deliver DNA to EndA. DNA was still bound and degraded in comEC and comFA mutants. We conclude that recruitment of EndA can occur in the absence of ComEC or ComFA and that EndA is active even when the single strands it produces are not pulled into the cell. Finally, inactivation of dprA had no effect on the internalization of DNA, indicating that DprA is required at a later stage in transformation.

Introduction

Natural genetic transformation, as first observed in Streptococcus pneumoniae (Griffith, 1928), is widespread among bacterial species (Lorenz and Wackernagel, 1994). Genetic transformation requires the development of competence, a genetically programmed transient state permitting uptake of DNA. Competence development is controlled in S. pneumoniae by a peptide signal transduction pathway including the ComC-derived competence-stimulating peptide (CSP), the ComD membrane-bound histidine kinase and the ComE response regulator, which are encoded by the comCDE operon (Håvarstein et al., 1995; Pestova et al., 1996). ComE controls the expression of comX, a gene encoding an alternative sigma factor specifically required for the synthesis of competence proteins including those involved in DNA uptake and processing (Lee and Morrison, 1999). The existence of this regulatory cascade accounts for the distribution of competence (com) genes into two classes (Alloing et al., 1998; Peterson et al., 2000; Rimini et al., 2000), early (i.e. ComE dependent) and late (i.e. ComX dependent) com genes (reviewed by Claverys and Håvarstein, 2002).

DNA uptake has been characterized to various extents in both Gram-positive, particularly Bacillus subtilis and S. pneumoniae, and Gram-negative bacteria, such as Haemophilus influenzae and Neisseria gonorrhoeae (reviewed by Dubnau, 1999). The first step in the process is the binding of double-stranded DNA, which requires the presence of a species-specific uptake sequence only in the latter two bacteria. Evidence for import of single-stranded fragments (Lacks, 1962; Piechowska and Fox, 1971; Dubnau and Cirigliano, 1972; Méjean and Claverys, 1984) and for linear uptake (Strauss, 1965; Gabor and Hotchkiss, 1966) was obtained in S. pneumoniae and B. subtilis. In both species, release of acid-soluble degradation products in the medium is concomitant with uptake (Dubnau and Cirigliano, 1972; Lacks and Greenberg, 1973; Morrison and Guild, 1973; Lacks et al., 1974; Méjean and Claverys, 1993). A 3′–5′ polarity of DNA transport was observed in S. pneumoniae (Méjean and Claverys, 1988) and H. influenzae (Barany et al., 1983). In the former species, entry of the transported strand and degradation of a strand equivalent occur at similar rates, ≈ 100 nucleotides s−1 at 30°C, and proceed with opposite polarities (Méjean and Claverys, 1993), consistent with a model in which degradation of one strand and internalization of its complement are coupled. The nuclease responsible for degradation in S. pneumoniae is the product of the endA gene (Lacks et al., 1975; Puyet et al., 1990). EndA is required for transport, but not for binding of donor DNA to the cell surface (Lacks et al., 1975). It is a constitutively expressed, membrane-located protein (Lacks and Neuberger, 1975), which is likely to be recruited by the DNA uptake complex that assembles in competent cells (Rosenthal and Lacks, 1980).

Most of the information regarding proteins in the DNA uptake complex comes from studies in B. subtilis (Dubnau, 1999). In a working model for DNA uptake based on evidence taken from both S. pneumoniae and B. subtilis, the ComG proteins are proposed to form (after processing by the ComC peptidase in B. subtilis) a structure that provides access of DNA to its receptor, ComEA, through the peptidoglycan (Dubnau, 1999). ComEA-dependent binding of donor DNA would be rapidly followed by cleavage by a nuclease, NucA in B. subtilis (Provvedi et al., 2001), and the newly formed DNA terminus would then be delivered to a transport complex, containing ComEC and ComFA. The non-transported strand would be degraded by a nuclease, EndA in S. pneumoniae, possibly providing energy for ComFA-dependent driving of the complementary DNA into the cell through an aqueous pore formed by ComEC (Dubnau, 1999).

Candidate S. pneumoniae orthologues of the B. subtilis DNA uptake proteins were identified among late com genes in several independent studies (Campbell et al., 1998; Claverys and Martin, 1998; Pestova and Morrison, 1998; Lee et al., 1999; Peterson et al., 2000; Rimini et al., 2000; Bartilson et al., 2001) and were given different names. [For clarity, in agreement with D. Morrison, we adopt here a simplified nomenclature referring to most of these genes on the basis of B. subtilis names, as comEA/C (cilE in Campbell et al., 1998; celA/B in Pestova and Morrison, 1998), comFA/C (Claverys and Martin, 1998; cfl in Lee et al., 1999) and comGA–F (cilD in Campbell et al., 1998; cglA–EPestova and Morrison, 1998).] Insertion–duplication or simple insertion mutants constructed at the orthologue of Bsu-comC (cilC in Campbell et al., 1998; cclA in Lee et al., 1999), at comGA (Campbell et al., 1998; Pestova and Morrison, 1998) and at comEC (Campbell et al., 1998; Pestova and Morrison, 1998) loci were reported to be severely affected (more than 100-fold reduction) or completely deficient for transformation. In another study, comFA mutants were reported to yield <1% transformants (Lee et al., 1999). However, in no case have the steps in transformation affected in the mutants been characterized. In addition, although the nuclease responsible for degradation of the non-internalized strand has been identified only in S. pneumoniae, the genetic requirements for its recruitment at the entry pore have never been investigated. This prompted us to generate mutants of candidate uptake genes.

We included a dprA mutant in our study, as inactivation of DprA was reported to lead to chromosomal transformation deficiency in H. influenzae (Karudapuram et al., 1995). In H. influenzae, the dprABC operon is induced at competence (Karudapuram and Barcak, 1997). The S. pneumoniae orthologue of dprA was identified among genes upregulated at competence as cilB (Campbell et al., 1998) or dal (Lee et al., 1999). It is expressed as a late com gene (Peterson et al., 2000; Rimini et al., 2000), and its inactivation was also reported to abolish transformation (Campbell et al., 1998; Lee et al., 1999). Inactivation of dprA in H. influenzae was not concluded to affect DNA uptake, but the processing of donor DNA (hence the name, dpr for DNA processing; Tomb et al., 1989; Karudapuram et al., 1995). This conclusion was based on the observation that, after contact with dprA competent cells, donor DNA became resistant to externally added DNase I. However, although resistance to DNase in Gram-positive bacteria can be unambiguously attributed to bona fide uptake across the cell membrane, in the Gram-negative H. influenzae, protection can result from entry into the periplasmic compartment or into specialized structures (called transformasomes by Kahn et al., 1983). Therefore, the question of whether donor DNA really enters the cytosolic compartment of dprA mutant cells remained totally open, leading to the suggestion that DprA may be needed to transport DNA across the inner membrane (Dubnau, 1999).

Here, we report on the characterization of candidate uptake and dprA mutants of S. pneumoniae with respect to degradation, binding and internalization of donor DNA, after normalization of competence levels in the various mutants using com::luc transcriptional reporters. Our observations give an insight into the genetic requirement for the recruitment of EndA by the DNA transport machinery and led us to propose refinements of the working model of DNA uptake in Gram-positive bacteria (Dubnau, 1999).

Results

Comparative analysis of donor DNA processing and of transformation frequency in wild-type and endA mutant cells

In a preliminary experiment to validate the binding, degradation and uptake assays (Experimental procedures), a wild-type strain was used in parallel with an endA mutant. Insertional inactivation of endA was first achieved through transfer of the endA::pJDC9 construct from strain 1142 (Puyet et al., 1990) to generate strain R1033 (Table 1). Surprisingly, residual transformation in this strain (0.5–3% of wild type; data not shown) appeared to be 20- to 125-fold higher than that reported for endA mutants (Puyet et al., 1990). This observation raised the question of a residual activity of EndA in the mutant generated by integration of the non-replicative pJDC9 plasmid derivative and prompted us to generate endA mutants through transposon insertion (in vitro mariner mutagenesis; see Experimental procedures) (Fig. 1). In addition, as competence for DNA uptake can easily vary 10-fold among different cultures even when grown similarly, a com::luc transcriptional reporter was introduced into all strains used in this study to permit normalization of competence (Experimental procedures). As shown in Fig. 2, measurement of transformation frequency in parallel with luciferase activity in strain R825 (wild type, but comC::luc, comC+; Table 1) under widely varying levels of com-petence demonstrated a direct correlation between luciferase activity and transformation level. The use of mariner insertions confirmed that endA mutants exhibit a residual transforming activity close to 1% of wild type (Table 2, lines 1 and 2, last column). As endA strains used by Lacks (1970) frequently carried an additional mutation inactivating exoA, residual transforming activity was also measured in exoA and endA–exoA mutants. Although exoA mutants displayed slightly reduced colony-forming ability (Table 2, column 2), there was no hint that the exoA mutation could affect transforming activity (Table 2, lines 3 and 4, last column). It is possible that the lower residual transforming activity of endA mutants reported previously (Puyet et al., 1990) resulted from the lack of normalization for competence. Alternatively, the difference could be accounted for by the presence of some other unidentified mutation(s).

Table 1. S. pneumoniae strains, plasmids and primers used in this study.
StrainsGenotype and descriptionSource or reference
1142TM6 trt1 hex4 endA::pJDC9; Eryr Puyet et al. (1990)
1374R6, but end1 exoA::cat; CmrS. A. Lacks
R800R6 derivative Martin et al. (1985)
R243R800, but comA::ermAMA Martin et al. (2000)
R304R800, but nov1, rif23, str41; Novr; Rifr; Smr Mortier-Barrière et al. (1998)
R825R800, but comC::luc (comC+) (by transformation with plasmid pR414); EryrThis study
R895R800, but ssbB::luc (ssbB+); Cmr Chastanet et al. (2001)
R1033R800, but endA::pJDC9 (by transformation with 1142 chromosomal DNA); EryrThis study
R1046R825, but exoA::cat (by transformation with 1374 chromosomal DNA); Eryr; CmrThis study
R1047R895, but comA::ermAM (by transformation with R243 chromosomal DNA) Eryr; CmrThis study
R1048R1051, but exoA::cat (from 1374); Eryr; Kanr; CmrThis study
R1049R1047, but endA::kan6C mariner insertion; Eryr; Cmr; KanrThis study
R1051R825, but endA::kan6C mariner insertion; Eryr; KanrThis study
R1059R1047, but dprA::spc21C mariner insertion; Eryr; Cmr; SpcrThis study
R1061R1047, but comEC::spc4A mariner insertion; Eryr; Cmr; SpcrThis study
R1062R1047, but comGA::spc3C mariner insertion; Eryr; Cmr; SpcrThis study
R1063R1047, but comFA::spc9C mariner insertion; Eryr; Cmr; SpcrThis study
R1146R1047, but comEA::spc3A mariner insertion; Eryr; Cmr; SpcrThis study
Plasmids  
p5.00ColE1 derivative carrying the firefly luciferase gene; 10 162 bp; Eryr; TcrM. Stieger
pR410ColE1 derivative, ApR, KanR; carries a 1337-bp-long mini-transposon containing the
inverted repeats (IRs) of the Himar1 transposon and ≈ 100 bp of Himar1 transposon
sequences flanking the kan gene
Sung et al. (2001)
pR412ColE1 derivative, ApR, SpcR; carries a 1145-bp-long mini-transposon containing the
inverted repeats (IRs) of the Himar1 transposon and ≈ 100 bp of Himar1 transposon
sequences flanking the spc gene
Martin et al. (2000)
pR413Same as pR412, but reverse orientation of the spc gene with respect to the IRs of the
Himar1 transposon
This study
pR414p5.00 derivative carrying a 389 bp comC targeting fragment; insertion–duplication in
S. pneumoniae generates a comC′::luc (comC+) fusion; Eryr
 
pR424ColE1 (pEVP3) derivative carrying a 246 bp ssbB targeting fragment and the firefly luciferase
gene; insertion–duplication in S. pneumoniae generates an ssbB′::luc (ssbB+) fusion; Cmr
Chastanet et al. (2001)
PrimersSequence; gene; position within the deposited sequence indicated;
GenBank/EMBL accession number
  • a. 

    Small letters in the oligonucleotide sequence indicate nucleotide change or extension to introduce convenient restriction sites (BamHI and HindIII underlined in BM37, BM40 and MP120 respectively).

  • C and A indicate, respectively, co-transcribed and reverse orientation of the inserted mini-transposon with respect to the targeted ORF.

BM37aggaattcggaTCCTAAACTCCAAAGTTTCTGCGTC; upstream from the ami operon; AE005672
BM40aaaggcctggatcCTTCAAAGCTACAAACTGTTCC; comC; 720–699; U33315
BM77ACAAAACTTGCTGGTTTAGATTGGA; upstream from comGA; AE005672; see Fig. 1
BM78TCAAGTAGGTCTGCACAAAGATTCC; downstream of comGA; AE005672; see Fig. 1
BM112CCTGCccGGGCTGAACCAGTCATCCCAG; downstream of the ami operon; AE005672
ComFA3AAATTCAATTTCATAACACC; AE005672; see Fig. 1
ComFA4ACGATGTTGTTAAGTTTAAC; AE005672; see Fig. 1
ComEC1GAGCGTTTATTGCTCAGACCTGTAA; AE005672; see Fig. 1
ComEC2GTCAATGTTTCCTGATGGGG; AE005672; see Fig. 1
MMN13TCGCCATTCCCAAGCCAAGA, downstream of endA; AE005672; see Fig. 1
MMN14TAGAAGAGATGCGCCGCATGGG; upstream from epuA; AE005672; see Fig. 1
MP120agccgcgaagcttCTAGTTCTTGTTGAACAAA; orfL (upstream from comC); 332–350; U33315
MP127CCGGGGACTTATCAGCCAACC; mariner cassette universal primer; internal to terminal IRs; outward orientation
MP128TACTAGCGACGCCATCTATGTG; mariner cassette universal primer; adjacent to IRL; outward orientation
MP133TTGGAGTCTATGGACTGTCTAT; upstream from dprA; AE005672; see Fig. 1
MP134GCACCAATGATACCTGCTATCC; downstream of dprA; AE005672; see Fig. 1
Figure 1.

Distribution of mariner insertions at loci encoding putative components of the DNA uptake machinery of S. pneumoniae. PCR fragments used for in vitro mariner mutagenesis (Experimental procedures) were, respectively, 1840- (endA), 3364- (comEA/C), 4003- (comFA/C), 1825- (comGA/B) and 3928-bp-long (dprA). ORFs corresponding to genes encoding proteins potentially required for DNA uptake are shown in grey. The location of oligonucleotide primers used for PCR is indicated by vertical arrows. Filled and open flags indicate, respectively, co-transcribed and reverse orientation of the kan or spc gene with respect to the targeted ORF. Digits below the maps correspond to insertions for which orientation of the mariner cassette was not determined. Underlined digits identify insertions with sequenced junctions used in this study (Table 1; data not shown). endA::kan6C, endA::kan15A, comEA::spc3A, comEA::spc6C, comEC::spc4A, comFA::spc9C, comFA::spc1C, comGA::spc3C, dprA::spc22A, dprA::spc21C and dprA::kan184A insertions occurred, respectively, after positions 340, 280, 221, 281, 560, 34, 880, 273, 502, 253 and 302 with respect to the start of the corresponding gene.

Figure 2.

Monitoring of competence using a com::luc (com+) transcriptional fusion. Spontaneous competence of strain R825 (Table 1) was assayed in C+Y medium by withdrawing aliquots at 15 min intervals and incubating for 20 min at 30°C with R304 chromosomal DNA (1 μg ml−1). SmR transformants were determined by plating as indicated in Experimental procedures. Luciferase activity in the culture was monitored in parallel, at 2 min intervals. The data presented are from two independent experiments.

Table 2. Transforming activity of wild type, endA and exoA mutants of S. pneumoniae.
  RLU × 10−2 ml−1SmR transformants
Straincfu × 10−7 ml−1−CSP+CSP×10−3 ml−1%% of wild typea
  • a. 

    Measurement of competence through expression of the comC::luc transcriptional fusion (RLU; third and fourth column) allowed normalization of competence levels (Experimental procedures), before making the ratio with the wild type.

R825 (wt)16.4 91530 08016 80010.2100
R1051 (endA)21.8119219 451110 0.05 0.8
R1048 (endAexoA) 7.7 86619.47530 0.04 0.6
R1046 (exoA) 8.1 87223 3652 200 2.7 34.0

The amount of donor DNA degraded was similar to the amount taken up in wild-type cells (Fig. 3A; see also Fig. 3B–D; data not shown). In agreement with previous reports, the exoA mutant behaved similarly to wild type with respect to processing of donor DNA (Fig. 3A; data not shown). On the other hand, endA::kan competent cells showed no detectable degradation of donor DNA to oligonucleotides and very limited uptake (the average of the three experiments shown is ≈ 5% of the value in the wild type). The observation of much higher DNA binding values in endA mutant strains (21- to 36.5-fold increase) compared with endA+ strains (Fig. 3A; see also Fig. 3D; data not shown) was consistent with previous reports (Lacks et al., 1974). The low binding value observed for the wild type is caused by the strict definition of binding adopted here, as in pre-vious studies (Lacks et al., 1974). Binding defined as labelled DNA co-sedimenting with cells through centrifugation and solubilized after DNase treatment (Experimental procedures) is only transient in wild-type cells, as bound donor DNA is rapidly processed, being degraded or internalized at a rate of ≈ 100 nucleo-tides s−1 (Méjean and Claverys, 1993).

Figure 3.

Characterization of degradation, binding and internalization of donor DNA in wild-type and entry mutant strains. Each value (Experimental procedures) was corrected by deducting counts measured in control non-competent cells and normalized for competence. Measurement of luciferase activity (Experimental procedures) resulting from induction of the comC::luc transcriptional fusion was used to monitor competence (Fig. 2). Counts measured in control cells (indicated as –CSP) are shown only for the wild-type strain in (B) and (C). Ordinate values represent units of 3496, 558, 874 and 342 c.p.m. (for 0.5 ml of culture) in (A)–(D) respectively. Note that the strains shown in a given section were analysed in the same experiment. The strains analysed were: in (A), wild-type R825, wt; R1046, exoA; R1048, exoA endA; R1051, endA. In (B), wild-type R1047, wt; R1062, comGA; R1059, dprA. In (C) wild-type R1047, wt; R1061, comEC; R1063, comFA. In (D), wild-type R1047, wt; R1146, comEA; R1062, comGA; R1049, endA. All strains used in the series of experiments shown in (B)–(D) were comA derivatives to abolish spontaneous competence development, which ensured the lowest possible background for control non-competent cells. Single data points are shown but, for each insertion mutant, at least triplicate independent experiments were carried out, using at least two different genetic backgrounds (comA+ and comA; data not shown).

Interestingly, the amount of material bound by endA mutant cells represents 52% (endAexoA strain; Fig. 3A), 64% (endA strain; Fig. 3D) and 90% (endA strain; Fig. 3A) of total donor DNA processed (i.e. degraded plus internalized) by wild-type cells. Thus, the extra binding is only apparent and is readily explained by the failure of the mutant to process bound DNA further, which therefore tends to accumulate at the surface of competent endA cells.

Inactivation of putative DNA transport genes through mariner mutagenesis

In vitro mariner mutagenesis of polymerase chain reaction (PCR) fragments overlapping the comEA/C, comFA/C, comGA/B and dprA loci (Fig. 1) was carried out as described in Experimental procedures. At least two randomly chosen insertion mutants were analysed at every locus. comEA, comEC, comFA, comGA and dprA insertion mutants revealed that chromosomal transformation dropped by at least six orders of magnitude (in fact below the limit of detection of transformants; data not shown), confirming that the corresponding proteins are absolutely required for genetic transformation.

Characterization of donor DNA processing in comGA and dprA mariner mutants

Inactivation of comGA abolished the binding and any further processing of donor DNA (Fig. 3B and D; data not shown). Inasmuch as the mariner cassette in the insertion mutant used was in the co-transcribed orientation and therefore should not abolish expression of the downstream gene, comGB, as judged from insertions at other loci (Martin et al., 2000), it is likely that the mutant pheno-type resulted from inactivation of comGA. This is consistent with observations in B. subtilis (Dubnau, 1999) and the proposal that ComGA is required for assembling a structure that traverses the peptidoglycan (Provvedi et al., 2001). The complete lack of donor DNA degradation in the comGA mutant strongly suggests that, despite its membrane location (Lacks and Neuberger, 1975), EndA is unable to access donor DNA by itself. On the other hand, dprA mutant cells were indistinguishable from wild type with respect to processing of donor DNA (Fig. 3B; data not shown). Analysis of two other dprA mutants (dprA::spc22A and dprA::kan184A; Fig. 1) gave similar results (data not shown).

Characterization of donor DNA processing in comEC and comFA mariner mutants

Inactivation of comEC or comFA completely abolished internalization of donor DNA (Fig. 3C; data not shown), a result similar to that in B. subtilis (Dubnau, 1999). However, comEC mutant cells exhibited near wild-type degradation of donor DNA and a binding value increased twofold compared with wild type (Fig. 3C). comFA mutant cells exhibited binding values close to that of wild type and significant degradation of donor DNA (40% of wild type; Fig. 3C). Similar results were obtained with another comFA mutant (comFA::spc1C; Fig. 1; data not shown). This situation is in contrast to that in B. subtilis, for which it was reported that comEC and comFA mutations not only prevented transport but also completely abolished donor DNA degradation (Provvedi et al., 2001). Our data suggest that, in S. pneumoniae, EndA requires neither ComEC nor ComFA for recruitment by the DNA uptake complex and functioning.

Characterization of donor DNA processing in comEA mariner mutants

Finally, DNA processing was analysed in comEA mutant cells, in parallel with wild type, and endA and comGA mutant cells as controls (Fig. 3D; data not shown). Inactivation of comEA completely abolished degradation and internalization of donor DNA, as expected. However, surprisingly, comEA mutant cells exhibited a 2.8-fold higher binding value compared with wild type. Note that essentially no binding was observed in the comGA control run in parallel (Fig. 3D; see also Fig. 3B). Analysis of another comEA insertion mutant (comEA::spc6C; Fig. 1) gave similar results (data not shown). Although the binding value for comEA mutant cells is lower than that for endA mutant cells (14-fold increase compared with wild type; Fig. 3D), it is significant (23 500 c.p.m. bound by ≈ 5 × 107 competent cells, well above a background value of 2600 c.p.m. for non-competent control cells). Material bound to endA and comEA competent cells represents, respectively, 64% and 12% of total donor DNA processed (i.e. degraded plus internalized) by wild-type cells in the same experiment (Fig. 3D). The value for comEA mutant cells suggests that they can still bind donor DNA, although much less efficiently than endA mutant cells.

Discussion

DprA is not required for DNA uptake but at a later stage in transformation

Among the mariner insertion mutants analysed in this study, only dprA exhibited normal processing of donor DNA (Fig. 3B). These data rule out the hypothesis that DprA is needed to transport DNA across the cytoplasmic membrane (Dubnau, 1999). Together with the observation that no transformant could be recovered in dprA mutant strains (Campbell et al., 1998; Lee et al., 1999; this study), they strongly suggest that entered single strands cannot be processed into viable transformants in the absence DprA. Work is in progress to determine the fate of donor DNA in mutant cells with the aim of elucidating the role of DprA in transformation.

Binding of donor DNA in S. pneumoniae

In B. subtilis, the ComG proteins are proposed to form a structure that provides access for DNA to its receptor, ComEA, through the peptidoglycan (Dubnau, 1999). The only mariner insertion leading to complete inhibition of DNA binding in S. pneumoniae is that in comGA (Fig. 3B and D). Although abolition of binding in comGA mutants does not imply that the ComGs are involved in surface binding of DNA, this result clearly indicates that ComEA is unable by itself to access donor DNA. In addition, ComEA mutant cells exhibited a 2.8-fold increased binding value compared with wild type (Fig. 3D). Collectively, these data suggest that, in S. pneumoniae, ComEA is not the only protein able to retain DNA at the cell surface.

ComGA, together with other ComGs, could play a role in the primary binding of donor DNA. Although biochemical evidence that B. subtilis ComGs bind double-stranded DNA could not be obtained in vitro (Provvedi et al., 2001), as the assay was conducted with purified His-tagged preparations, it remains possible that a complex structure and/or interactions with the peptidoglycan are necessary for DNA binding by the ComGs.

Our data with comEA mutant cells are not inconsistent with the observations of Inamine and Dubnau (1995) in B. subtilis. These authors, defining bound material as total DNA associated with cells after washing (i.e. including internalized DNA), observed abolition of binding only for a comEA mutant (Δ2) lacking 22 residues from the extreme carboxyl-terminus of ComEA, whereas two other comEA mutants exhibited only a limited reduction in binding (twofold for Δ1 that carried a deletion removing residues 102–140; and fourfold for a Tn917-lacZ insertion mutant). It should be pointed out that no ComEA signal could be detected with anti-ComEA antiserum for the Δ2 deletion and the Tn917 insertion mutant (Inamine and Dubnau, 1995). That the latter could nevertheless achieve ≈ 25% as much irreversible binding as the wild type is not consistent with ComEA being the surface receptor of DNA in B. subtilis and, indeed, Inamine and Dubnau (1995) concluded that ComEA was absolutely required for uptake but not for binding. On the other hand, several lines of evidence, including the use of a membrane vesicle DNA-binding assay system, clearly showed that B. subtilis ComEA binds double-stranded DNA through its C-terminal domain (Provvedi and Dubnau, 1999). The His-tagged ComEA orthologue of Neisseria gonorrhoeae was also recently shown to exhibit DNA-binding activity (Chen and Gotschlich, 2001). The lack of detecTable sequence specificity in binding (whereas binding of DNA in this species is sequence specific) suggests that ComEA is not the cell surface receptor of this species, a result not unexpected because of the presence of the outer membrane in this Gram-negative bacterium.

To reconcile the various observations, we suggest the existence of two stages in binding, surface attachment and deep binding. Surface attachment would depend on the ComGs and would be a prerequisite for deep binding, as suggested by the observation that binding was abolished in a comGA mutant. ComEA, which displays DNA-binding activity in the absence of wall (Provvedi et al., 2001), would be responsible for deep binding. In the absence of ComEA, DNA would be less tightly bound to the surface, thus accounting for the lower binding value in comEA compared with endA mutant cells (Fig. 3D). Alternatively, the engagement of donor DNA in a ComG-dependent transpeptidoglycan structure could account for the retention of some donor DNA at the surface of comEA competent cells.

Genetic requirement for the recruitment of EndA by the DNA uptake machinery

The complete lack of degradation in the comGA mutant (Fig. 3B and D) indicates that, although EndA is a constitutively expressed protein with a membrane location (Lacks and Neuberger, 1975), it cannot access donor DNA by itself. Obviously, abolition of primary binding in the comGA mutant would prevent EndA-mediated degradation of donor DNA. Alternatively, ComGA together with other ComGs could be required not (or not only) for granting access of donor DNA through the cell wall but for the formation of a structure essential for the recruitment of EndA at the entry pore.

The complete absence of degradation observed in the comEA mutant cannot be explained solely by a defect in the binding of donor DNA, as significant binding was still observed in ComEA mutant cells (Fig. 3D). Interestingly, a binding value similar to that in ComEA was observed in the comFA mutant and was accompanied by significant degradation (≈ 40% of wild type; Fig. 3C). We take these observations as evidence that ComEA is not only involved in the binding of DNA, but is also essential to present donor DNA to EndA, rather than directly to the transport pore as suggested by Dubnau (1999). A flexible sequence in ComEA, which was suggested to allow a bending important to deliver DNA to the uptake apparatus (Dubnau, 1999), could be instrumental in this presentation. EndA would then degrade one strand, allowing transport of its complement into the cytosol.

Relationship between degradation and internalization of donor single strands

In contrast to the situation in B. subtilis in which degradation of donor DNA could not be detected in a comEC mutant (Provvedi et al., 2001), near-wild-type degradation levels were observed in an S. pneumoniae comEC mutant (Fig. 3C). It has been proposed that ComEC forms part or all of an aqueous pore essential for internalization of donor DNA (Dubnau, 1999). Although our data say nothing concerning the exact role of ComEC, they suggest that a structure at least partially functional is maintained in its absence. This structure allows the binding of DNA, the recruitment of EndA and the degradation of donor DNA. ComGB, a protein predicted to contain transmembrane segments (Dubnau, 1999), could play such a structural role, in addition to its postulated role in transmembrane transport of the ComGs.

Significant donor DNA degradation occurred in S. pneumoniae comFA mutants (Fig. 3C), also in contrast to the situation in B. subtilis (Provvedi et al., 2001). These data indicate that single strands produced by EndA need not be pulled into the cytosol by ComFA for degradation of donor DNA to continue. Although it is possible that degradation is somewhat reduced in the absence of ComFA (≈ 40% of wild type; Fig. 3C), our observations provide additional evidence that, in S. pneumoniae, DNA degradation can be uncoupled from transmembrane transport. The previous finding that carefully prepared single-strand donor DNA retained appreciable transforming activity (Miao and Guild, 1970; Barany and Boeke, 1983) and the behaviour of several transformation-defective mutants similar to that shown here for comFA and comEC mutants (Morrison et al., 1983) also provided support for this idea.

The uncoupling of DNA degradation and internalization of single strands in comEC and comFA mutants raises the question of the fate of the single strands produced. It would interesting to determine whether these single strands are also degraded or could be reisolated intact from the cell surface.

How could the inactivation of comEC and comFA have a different impact on DNA degradation in B. subtilis and S. pneumoniae? As no homologue of endA was detected in the genome of B. subtilis (Dubnau, 1999), a different nuclease could be recruited in this species. This nuclease may differ from EndA in its requirement for functioning at the entry pore. Alternatively, it is possible that, in B. subtilis, there is no recruited nuclease, the nucleolytic activity necessary for uptake being intrinsic to the uptake machinery. This putative intrinsic activity may require the presence of ComEC and/or ComFA. Interestingly, the ComEC protein belongs to the metallo-β-lactamase fold superfamily and, as such, contains conserved catalytic residues that could carry out some as yet unidentified hydrolytic reaction (Aravind, 1999). It is tempting to speculate that this hydrolytic reaction is a nucleolytic one. Such an intrinsic nucleolytic activity could also account for residual entry (Fig. 3A and D; data not shown) and residual transforming activity (Table 2) exhibited by S. pneumoniae endA mutants.

Experimental procedures

Bacterial strains, growth conditions and transformation

Escherichia coli K-12 strain DH5α (Life Technologies) was used for cloning experiments. E. coli strains were grown in LB medium (Sambrook et al., 1989) and transformed by electroporation with selection on LB plates supplemented with erythromycin (Ery; 200 μg ml−1) or chloramphenicol (Cm; 10 μg ml−1). S. pneumoniae strains used in this work are listed in Table 1. S. pneumoniae strains were grown in Todd–Hewitt medium (supplemented with 5 g l−1 yeast extract) and in CAT medium (Porter and Guild, 1976), or in C + Y medium (Alloing et al., 1998) under ‘acid growth’ conditions (i.e. with an initial pH value of 6.9).

Transformation of S. pneumoniae with R304 chromosomal or radiolabelled DNA (see below) was performed as described previously (Martin et al., 2000), using precom-petent cells treated at 37°C for 10 min with synthetic CSP1 (100 ng ml−1) to induce competence. Transformants were selected by plating on CAT agar supplemented with 4% horse blood, followed by challenge with a 10 ml overlay containing the appropriate antibiotic after phenotypic expression for 120 min at 37°C. Antibiotic concentrations used for the selection of transformants were: Cm, 2.5–4.5 μg ml−1; Ery, 0.05– 0.2 μg ml−1; kanamycin (Kan), 250 μg ml−1; spectinomycin (Spc), 100 μg ml−1; streptomycin (Sm), 200 μg ml−1.

Measurement of competence

Monitoring of competence was achieved using transcriptional fusions between the early (comC) or the late (ssbB) com genes and the Photinus pyralis luc gene, which encodes the firefly luciferase (Stieger et al., 1999). Transformation of S. pneumoniae cells with plasmid pR414 (see below) or pR424 (Table 1), with selection for Ery or Cm resistance, leads to integration of the corresponding plasmid at the comC or ssbB locus respectively. The resulting strains harbour a comC′::luc (comC+) or a ssbB′::luc (ssbB+) transcriptional fusion. For the detection of luciferase activity, 20 μl of CSP-activated or control cells was added to 100 μl of luciferin solution (0.2 mM in 10 mM phosphate buffer, pH 7.2), and luminescence (expressed in relative luminescence units, RLU) was measured for 30 s in a Berthold Lumat LB 9501 luminometer. A direct correlation between luciferase activity and transformation level was observed under widely varying levels of competence (Fig. 2). Luciferase activity was therefore used in this study as an indicator of competence of entry mutants.

Plasmids, oligonucleotides and plasmid constructions

Plasmids and oligonucleotides used in this study are listed in Table 1. The luc gene was placed under the control of the comC promoter in plasmid pR414 as follows. A DNA fragment overlapping the 5′ extremity of the comC gene was amplified from S. pneumoniae R800 chromosomal DNA by the PCR technique using oligonucleotides BM40-MP120 and digested with BamHI–HindIII to generate a 389 bp fragment. This fragment was cloned into a 9004-bp-long BamHI–HindIII fragment from plasmid p5.00 (generously provided by M. Stieger), which confers Ery resistance and carries the luc gene, thus generating plasmid pR414 (Table 1).

In vitro mariner mutagenesis

Mutagenesis of PCR fragments was performed as described previously (Martin et al., 2000). Briefly, plasmids used as sources for the 1337 bp kan (Kanr) and the 1145 bp spc (Spcr) mariner mini-transposons were pR410 (kan; Table 1) for mutagenesis of the endA fragment, pR412 (spc; Table 1) for the comEA/C, comFA/C and comGA/B PCR fragments and pR413 (spc; Table 1) for the dprA fragment. Plasmid DNA (≈ 1 μg) was incubated with PCR fragments (≈ 1 μg) in the presence of purified Himar1 transposase (Lampe et al., 1996), in a total volume of 40 μl leading to random insertion of the mini-transposon within the fragments. Gaps in transposition products were repaired as described previously (Akerley et al., 1998), and the resulting in vitro-generated transposon insertion library was used to transform S. pneumoniae. Location and orientation of the mariner cassette was determined through PCRs using primers MP127 or MP128 (Table 1) in combination with either of the two primers used to generate each PCR fragment. In some cases, cassette–chromosome junctions were sequenced (Fig. 1) using the ThermoSequenase cycle sequencing kit (USB) and primer MP128.

Labelling of donor DNA

Radiolabelled donor DNA was prepared by PCR using Sigma's AccuTaq LA DNA polymerase and S. pneumoniae R800 chromosomal DNA. PCRs (50 μl total volume) were carried out under the conditions recommended by the manu-facturer with the BM37–BM112 primer pair (Table 1) in the presence of dNTPs (0.2 mM each) and [α32P]-dATP (Amersham; specific activity 111 × 106 kBq mmol−1; 1850 kBq) to generate a 10397-bp-long fragment. Amplification proceeded for 30 cycles as follows: 1 min at 98°C, 30 s at 94°C, 1 min at 62°C, 20 min at 68°C, followed by a 10 min extension cycle.

Measurement of degradation, binding and internalization of donor DNA

Radioactively labelled donor DNA (2 μl; ≈ 50 ng; ≈ 4 × 105 c.p.m.) was added to 0.5 ml of competent cells for 15 min at 30°C. Cells were chilled on ice and centrifuged for 1 min at 10 600 g. For the measurement of degradation, the supernatant fraction was precipitated with 10% TCA for 20–30 min on ice and centrifuged for 10 min at 10 600 g at 4°C. An aliquot of 100 μl of the TCA soluble fraction was mixed with scintillation fluid (EcoLite; ICN), and radioactivity was counted in a Beckman LS 3801.

For the measurement of DNA binding, the pellet from the first centrifugation above was carefully washed once by layering 1 ml of cold CAT medium over the cells without resuspending them (Lacks et al., 1974) and centrifuging for 1 min at 10 600 g. Cells were resuspended in 100 μl of CAT (containing 10 mM MgCl2) and treated with DNase I (50 μg ml−1) for 5 min at 30°C. After centrifugation for 1 min at 10 600 g, 100 μl of supernatant was mixed with scintillation fluid, and radioactivity was counted.

For the measurement of DNA uptake, the pellet from the last centrifugation above (i.e. cells containing radioactive donor DNA rendered resistant to externally added DNase I because of internalization) was resuspended in 100 μl of SEDS (0.15 M NaCl, 0.015 M EDTA, 0.02% SDS and 0.01% sodium deoxycholate); cells were then incubated for 5 min at 37°C (to allow for autolysis), and total radioactivity was counted after mixing with scintillation fluid.

Acknowledgements

We thank S. Lacks for the gift of strains 1142 and 1374, Chantal Granadel for expert technical assistance, and The Institute for Genomic Research for generously providing access to the complete Streptococcus pneumoniae type 4 genome sequence before publication. Mathieu Bergé was the recipient of a PhD thesis fellowship from the Ministère de la Recherche, and Miriam Moscoso was the recipient of a Marie Curie Individual Fellowship (QLK2-CT-1999-51509). This research was financed in part by the European Union (grant QLRK 2000-00543).

Ancillary