ABSTRACT
The dlt operon of gram-positive bacteria comprises four genes (dltA, dltB, dltC, and dltD) that catalyze the incorporation of d-alanine residues into the lipoteichoic acids (LTAs). In this work, we characterized thedlt operon of Streptococcus agalactiae, which, in addition to the dltA to dltD genes, included two regulatory genes, designated dltR and dltS, located upstream of dltA. The dltR gene encodes a 224-amino-acid putative response regulator belonging to the OmpR family of regulatory proteins. The dltS gene codes for a 395-amino-acid putative histidine kinase thought to be involved in the sensing of environmental signals. The dlt operon ofS. agalactiae is mainly transcribed from the PdltR promoter, which directs synthesis of a 6.5-kb transcript encompassing dltR, dltS, dltA, dltB, dltC, and dltD, and from a weaker promoter, PdltA, which is located in the 3′ extremity ofdltS. We demonstrate that PdltR, but not PdlA, is activated by DltR in the presence of DltS in d-Ala-deficient LTA mutants resulting from insertional inactivation of the dltA gene, which encodes the cytoplasmic d-alanine-d-alanyl carrier ligase DltA. Expression of the dlt operon does not require DltR and DltS, since the basal activity of PdltR is high, being 20-fold that of the constitutive promoter PaphA-3 which directs synthesis of the kanamycin resistance gene aphA-3 in various gram-positive bacteria. We hypothesize that the role of DltR and DltS in the control of expression of the dlt operon is to maintain the level of d-Ala esters in LTAs at a constant and appropriate value whatever the environmental conditions. The DltA− mutant displayed the ability to form clumps in standing culture and exhibited an increased susceptibility to the cationic antimicrobial polypeptide colistin.
The cell wall of gram-positive bacteria contains two types of anionic polymers: (i) the teichoic acids, which are covalently linked to the peptidoglycan; and (ii) the lipoteichoic acids (LTAs), which are polyphosphoglycerol substituted with a d-Ala ester or a glycosyl residue and are anchored in the membrane by their glycolipid moiety (13,15). Incorporation of d-Ala residues into the LTAs necessitates the activity of four gene products (DltA to DltD) that are encoded by the dlt operon (see Fig. 1A). Inactivation of genes within this operon in various gram-positive bacilli and cocci results in the complete absence of d-Ala ester from LTAs, and these d-Ala-deficient LTA mutants were found to exhibit a variety of phenotypic changes that could be attributed to the resulting charge modification of their cell surface. In particular, due to their increased electronegativity, these mutants are thought to bind cationic compounds more efficiently. Consistently,d-Ala-deficient LTA mutants of Staphylococcus aureus and Staphylococcus xylosus are more susceptible to cationic antimicrobial peptides than the wild-type strains are (27). Moreover, S. aureus cells lackingd-Ala esters in LTAs are more susceptible to vancomycin and possess reduced autolytic activity (27). InLactobacillus casei, inactivation of the dltDgene resulted in an enhanced microbial activity of cationic detergents (11). The enhancement of endogenous and β-lactam-induced cell lysis observed with d-Ala-deficient LTA mutants ofBacillus subtilis was postulated to occur through increased binding of cationic autolysins to negatively charged LTAs by electrostatic interactions (40). In Streptococcus gordonii, it was hypothesized that d-alanyl LTAs may provide binding sites for a 100-kDa cell surface protein and a scaffold for the proper presentation of this adhesin to mediate intrageneric coaggregation (7). Accordingly, it has been recently suggested that d-Ala substitution of LTAs could modulate the rate of posttranslocational folding of exported proteins inB. subtilis by maintaining a high concentration of cations (Ca2+, Fe3+) at the membrane-wall interface (21). Insertional inactivation of genes within thedlt operon was also associated with unexpected phenotypes, such as sensitivity to UV radiation, as reported in the case ofLactococcus lactis (12), or the inability to accumulate intracellular polysaccharides and to adapt to acid stress, as shown in the case of Streptococcus mutans (2,35).
Lancefield's group B streptococcus (GBS), also referred to asStreptococcus agalactiae, is one of the leading causes of invasive infections (septicemia, meningitis, and pneumonia) in neonates (34). Newborns are colonized intrapartum by aspiration of contaminated amniotic fluid, and the lung is a likely portal entry for GBS into the bloodstream, since these bacteria can adhere to and invade alveolar epithelial (31) and endothelial cells (18). The physiopathology of GBS infections implies that this bacterium can rapidly adapt to various growth conditions, including pH, osmolarity, and temperature variations (37). Bacterial responses to environmental stimuli are often mediated by two-component regulatory proteins, which comprise a histidine-protein kinase as a sensor and a response regulator (19). Signal transduction occurs through autophosphorylation of the sensor, which then transfers a phosphoryl group to the regulator. However, several histidine-protein kinases also act as phosphatases and catalyze the dephosphorylation of the associated response regulator. Regulation may therefore take place by modulating either the kinase or the phosphatase activities of the sensor. Response regulators share a conserved amino-terminal domain that is approximately 120 amino acids (aa) long and that contains two regions of strong amino acid sequence identity. Based on this property, it is possible to characterize response regulators in a wide range of bacteria by using degenerate oligonucleotides in PCRs (41). We have developed a similar strategy to identify response regulators in S. agalactiae(data not shown). Among the four putative regulatory gene fragments cloned and sequenced from this bacterium, one has retained our attention because it belongs to a hitherto novel two-component regulatory system that was located immediately upstream from thedlt operon responsible for the formation ofd-alanyl esters of LTAs. We report in this study a genetic and functional analysis of the dlt operon of S. agalactiae.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions.The main characteristics of the relevant bacterial strains and plasmids used in this study are listed in Table 1. All strains were grown at 37°C in brain heart infusion (BHI) broth or agar plates (Difco Laboratories). Unless otherwise specified, antibiotics were used at the following concentrations: forEscherichia coli, ampicillin, 100 μg/ml; erythromycin, 150 μg/ml; kanamycin, 50 μg/ml; spectinomycin, 60 μg/ml; and streptomycin, 50 μg/ml; for S. agalactiae, chloramphenicol, 5 μg/ml; erythromycin, 5 μg/ml; kanamycin, 1,000 μg/ml; spectinomycin, 250 μg/ml; streptomycin, 500 μg/ml; and nalidixic acid, 50 μg/ml.
Bacterial strains and plasmids used in this study
Construction of plasmids.A 527-bpEcoRI-BamHI DNA fragment containing the PdltR promoter was inserted into the multiple cloning site of pTCV-erm to give pTCV-ermΩPdltR. The pair of oligonucleotides IntB and IntP was used to amplify a promoterlessint gene cassette from the DNA of the conjugative transposon Tn1545. After digestion with BamHI andPstI, this amplicon was cloned in the proper orientation downstream from pTCV-ermΩPdltR to give pTCV-int (Table 1). This low-copy vector plasmid directs synthesis of the transposon-encoded integrase Int-Tn inS. agalactiae and can be readily lost following subculture at 41°C in the absence of antibiotic selective pressure.
Construction of bacterial strains.To construct S. agalactiae strains NEM1636, NEM1637, and NEM1693, we inserted in the same direction of transcription the promoterless and terminatorless kanamycin resistance cassette aphA-3 within DNA segments internal to dltA, dltR, and dltS, respectively. This was done by ligation after digestion with the appropriate enzymes, of the amplicons LA21-LA22, KanK-KanB, and LA23-LA24 for NEM1636 construction; LA17-LA18, KanK-KanB, and LA19-LA20 for NEM1637 construction; and LA28-LA29, KanK-KanB, and LA30-LA31 for NEM1693 construction. The corresponding EcoRI-PstI fragments were cloned into pG+host5, and the resulting recombinant vectors were introduced by electroporation into NEM316. The double crossover events leading to the expected gene replacements were screened and obtained as described previously (1). In NEM1637, NEM1639, NEM1693, and NEM1786, the promoterless resistance cassettes used to inactivatedltR or dltS are transcribed from PdltR, and the insertional inactivation strategy described above was used because we initially showed in this study that this promoter was active in the absence of the regulatory proteins DltR and DltS.
The double mutant NEM1639 was constructed similarly to that of NEM1637 by insertional inactivation of the dltR gene of NEM1636 with the chloramphenicol resistance cassettecatpIP501 . To construct the double mutant NEM1786, the dltS gene of NEM1636 was insertionally inactivated with the streptomycin resistance cassetteaad6pJH1 . The catpIP501 and aad6pJH1 resistance cassettes were also designed devoid of promoter and terminator sequences to avoid transcriptional polar effects. Southern analysis of restriction enzyme-digested DNA revealed that all mutant strains were devoid of sequences related to pG+host5 and that insertion of the resistance cassette(s) occurred at the expected location (data not shown).
For complementation analysis, we used the following strategy. The pairs of oligonucleotides LA25-LA26 and LA11-LA4 were used to amplify thedltA gene associated with the upstream promoter PdltA and the dltR gene associated with the upstream promoter PdltR. The corresponding amplicons, following digestion with the appropriate enzymes (BamHI plus PstI and EcoRI plus PstI, respectively) were inserted into pAT113/Sp to give pAT113/SpΩdltA and pAT113/SpΩdltR. These vectors were conjugatively transferred from HB101/pRK24 toS. agalactiae NEM1636/pTCV-int and NEM1637/pTCV-int to restore the DltA and DltR activities, respectively, in these mutant strains. In both cases, the plasmid insertion sites were characterized by inverted PCR in three integrants harboring a single copy of the integrative vector inserted at different loci. This was done by using ligated Sau3A-digested chromosomal DNA as a template in PCRs carried out with the primer pairs attRin-attRout and attLin-attLout to characterize the right and left chromosome-plasmid junction fragments, respectively (attL and attR were previously arbitrarily defined) (39). Sequence analysis of the six insertion sites revealed that in neither case was the integrative vector inserted within an open reading frame (data not shown). The complemented strains NEM1638 and NEM1687 were chosen for further studies because no transcript running through the corresponding vector integration site in NEM316 was detected by Northern blot analysis (data not shown).
Genetic techniques.Recombinant plasmid DNAs were introduced by transformation into Escherichia coli (33). Electrocompetent cells of S. agalactiae were prepared as described previously (9). IncP mobilizable shuttle vectors (pTCV-lac, pTCV-int, pAT113/Sp, and their derivatives) were transferred by conjugation from the mobilizing donor strain E. coli HB101/pRK24 to S. agalactiaerecipients (29).
DNA manipulations. E. coli DH5α was used as a host for plasmid constructions. Plasmid DNA from E. coli(33) and total DNA from S. agalactiae(30) were extracted as described above. PCRs were carried out in a final volume of 50 μl containing 50 ng of DNA, 0.1 μM each primer, 200 μM each deoxynucleoside triphosphate, and 2 U of AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster City, Calif.) in a 1× amplification buffer (15 mM Tris-HCl [pH 8.0], 50 mM KCl, 2.5 mM MgCl2). The PCR mixture was submitted to a denaturation step (10 min at 95°C), followed by 30 cycles of amplification (90 s of annealing at 55°C, 90 s of elongation at 72°C, and 60 s of denaturation at 95°C). DNA was sequenced with an ABI 310 automated DNA sequencer by using the ABI PRISM dye terminator cycle sequencing kit (Applied Biosystems).
RNA preparations, Northern blot, RT-PCR, and primer extension analysis.Total RNAs were extracted as described previously (28) from mid-exponential (optical density at 600 nm [OD600] = 0.6)-phase cultures of S. agalactiaegrown in BHI broth at 37°C without agitation. For Northern blot analysis, 40 μg of RNA was separated through a 1.3% formaldehyde–agarose gel (33) and transferred to a Hybond-N+ membrane (Amersham, Uppsala, Sweden). The filters were baked for 2 h at 80°C in an oven. Prehybridization and hybridization were performed under stringent conditions as described previously (33). The DNA probes used (S1 [LA3-LA4], S2 [LA5-LA6], S3 [LA7-LA8], and S4 [LA9-LA10]) were PCR fragments obtained from NEM316 genomic DNA by using the primers indicated in brackets. DNA fragments were labeled with [α-32P]dCTP by using an Amersham nick translation kit.
RT-PCRs were carried out by using the Superscript One-Step reverse transcription-PCR (RT-PCR) system (Gibco BRL). Reactions were carried out in a final volume of 50 μl containing 0.5 μg of RNAs, 0.2 μM each primer, 1 μl of RT-Taq mix, and 25 μl of 2× reaction mix. Primer annealing and RT (30 min at 50°C) were followed by 35 cycles of amplification (30 s of annealing at 55°C, 60 s of elongation at 72°C, and 15 s of denaturation at 94°C).
For primer extension analysis, 50,000 cpm of γ-32P-labeled primer O33, complementary to the spoVG-lacZ gene of pTCV-lac (Table 1), was mixed with 10 μg of RNA in 1× hybridization buffer (33). The mixture was heated for 10 min at 85°C and incubated overnight at 30°C. Primer extension was performed for 90 min at 42°C with 50 U of avian myeloblastosis virus reverse transcriptase (Boehringer, Mannheim, Germany). The products were separated on an 8% polyacrylamide sequencing gel next to a DNA sequence reaction (Sequenase sequencing kit; Amersham) obtained by using the same oligonucleotide as primer.
Construction of spoVG-lacZ transcriptional fusions and β-galactosidase assay.The amplified fragments tested for promoter activity were digested with EcoRI plusBamHI and cloned into pTCV-lac. All inserts were sequenced as described previously (29) to verify that no misincorporation of nucleotides occurred during the PCR assay. S. agalactiae cells containing lacZ fusions were cultivated in BHI broth at 37°C without agitation. Cells were collected in the mid-exponential phase, and β-galactosidase activities were assayed as described previously (22), except that the cells were permeabilized by treatment with 0.5% toluene and 4.5% ethanol. The β-galactosidase specific activities, determined in three experiments, are expressed as [103 × (OD420 of the reaction mixture −1.75 OD550 of the reaction mixture)]/[time of the reaction in min × OD600 of the quantity of cells used in the assay].
Scanning electron microscopy.Aliquots (20 to 50 μl) of overnight cultures of S. agalactiae cells grown without agitation in BHI medium at 37°C were gently spread on microscope glass slides. The cells were fixed with a 3% glutaraldehyde solution in 0.1 M cacodylate buffer (pH 7.3). Preparations were then coated with gold palladium after critical point drying. Examination was performed with a JEOL 840A at the Centre Inter-Universitaire de Microscopie Electronique (University of Paris 7, Paris, France).
Biochemical analysis of LTAs.LTAs were extracted and purified under conditions that preserve the native substitution withd-Ala and were analyzed for d-Ala ester content as previously described (32).
Western blot analysis of surface proteins.Surface proteins of S. agalactiae were extracted by gentle sodium dodecyl sulfate (SDS) treatment (0.1% SDS, 5 min, 20°C) as described previously (36). Under these conditions, the viability ofS. agalactiae was barely altered (data not shown). Following electrophoresis under denaturing conditions, the proteins were transferred onto a Nylon membrane and revealed as described previously (17) by using a rabbit serum (diluted 1:1,000 in phosphate-buffered saline) containing polyclonal antibodies raised against formaldehyde-treated S. agalactiae NEM316.
Oligonucleotides.The sequences (5′ to 3′) of the oligonucleotides used in this study were as follows: LA1, GGRIRTIYTIRTIGTIGAIGA; LA2, AAIGGYTTIRYIAIRTARTCIIIIGYICC; LA3, CTGCAAAATCACAGACTTATG; LA4, AACTGCAGCAATAAGGATTGACATGGTC; LA5, GTTGGAGAAGAGAGTCAG; LA6, TCATGTATCATTTGTACCATC; LA7, CTTGAAGTTGATATTGTCC; LA8, CTAATTCAATACGATAGCCG; LA9, GTTAGAGCTTTCAGTGATCC; LA10, GCATCTGTCTTCTCAAATACCC; LA11, CGGAATTCGCCAACGTAAACACGGATTC; LA12, GGGGACAATAAAGCCTGAATAGCC; LA13, ATTGAATTCGTTTAGTGACTTAGG; LA14, GAAGAATTCCTTAATTTGTCTAGACTTGATG; LA15, CGGAATTCTGGATTGTTGGAGAAGAGAGTCAG; LA16, GGGGATCCATCTCCTAGIIITCATT; LA17, GGGGATCCATAACCTCTTTGG; LA18, CGGGAATTCGTAGACAAATAGCCCC; LA19, GGGGTACCTCATGTGAATCTGTGTTAAGCC; LA20, GGGATCCTTCAGACAATTCAGAATATAACC; LA21, TTCTGCAGTCATTTCTGTCTCTACAGGTCC; LA22, CGGAATTCGGTGGTCAAGAATATG; LA23, CGGGTACCGCGCCATATCAGCAAATGATGG; LA24, CGGGATCCGATGGGGAAGAGTTAACTGTC; LA25, AACTGCAGATGGCATCATGTAATCC; LA26, ATGGATCCTGGATTGTTGGAGAAGAGAGTCAG; LA27, TACTGCAGATTCCCATAAACATCAAGTGAGG; LA28, TAGGAATTCCAGGCGATGAACCG; LA29, GAGTGGTACCTGTCTCTACAGGTCCTAC; LA30, TCTGGATCCTGTTGTTTCGGAGACTAAGCG; LA31, GCTCTGCAGCTCCCCTTATGGCGTTCCACG; LA32, TTCTTATTGGAAATATCTTTATAGCG; LA33, ATGAGACTTCTTGTAGTTGAGG; LA34, CAGGTTTGGACTTCGACACC; LA35, TTGAAAGGGTCACAACGACATTTC; LA36, AACACTCATTTGATATCTAG; LA37, GGACGTCCTGTATATTTTGCCC; SmK, TCGGTACCGAAGAAGATGTAATAATATAG; SmB, TTGGATCCCTGTAATCACTGTTCCCGCCT; KanK, GGGGTACCTTTAAATACTGTAG; KanB, TCTGGATCCTAAAACAATTCATCC; CatK, GGGGTACCAGAGGATTATTCCTCC; CatB, TCGGATCCGTGTATAAAATTAAATTCAC; IntB, CTGGATCCATAAAGGAAAGGAGC; IntP, TTCTGCAGTACTACTAAGCAACAAGAC; attRin, GGGATATATCAACGGTGG; attRout, GATAAGTCCAGTTTTTATGCGG; attLin, CCTTCTCGTTCGGAGGAAATCC; attLout, TTCTGACAGCTAAGACATGAGG; O33, CGTCAGTAACTTCCACAGTAGTTCACCACC.
RESULTS
Sequence analysis of the dlt operon of S. agalactiae NEM316.The molecular cloning of the 300-bp amplification product obtained by using the degenerate primers LA1 and LA2 enabled the sequencing of four segments encoding the NH2-terminal phosphoaccepting receiver domains of putative response regulators in S. agalactiae NEM316. Sequences of the region located downstream from these DNA fragments internal to putative response regulator genes were characterized by an inverse PCR strategy (data not shown). This analysis revealed that one of these regulator genes was associated with a gene encoding a putative protein kinase located immediately upstream of an open reading frame related to the dltA genes of various gram-positive bacilli and cocci. This gene encodes ad-alanine-d-alanyl carrier ligase involved in the esterification of the membrane-associated LTA. A sequencing strategy that combined inverse PCR and regular PCR (data not shown) was used to determine the sequence of a 6,937-bp-long DNA segment that contains the dlt operon (Fig.1A). Structural analysis of this DNA fragment revealed that it contained six open reading frames (dltR, dltS, dltA, dltB, dltC, and dltD) that have the same polarity of transcription and that encode putative peptides sharing significant homology with cognate proteins. However,S. agalactiae is the only gram-positive bacterium characterized so far that possesses a dlt operon, including a two-component regulatory system.
Characterization of the dlt operon ofS. agalactiae NEM316. (A) Comparison of the dltoperon of NEM316 with those of various gram-positive bacteria. The accession numbers or URL sites used to collect these sequences were as follows: B. subtilis,genomeweb.pasteur.fr/GenoList/SubtiList/; Listeria monocytogenes, AJ012255; Lactobacillus rhamnosus,U43894; S. aureus, D86240; S. xylosus, AF032440;Enterococcus faecalis, www.tigr.org/tdb/mdb/mdb.html;S. mutans, AF051356; Streptococcus pneumoniae,genome.microbio.uab.edu/strep/strep.asp; Streptococcus pyogenes, www.genome.ou.edu/strep.html; and S. agalactiae, AJ291784. The dltR and dltSgenes encode the regulatory and sensor proteins of a two-component regulatory system, respectively; dltA, dltC, dltB, anddltD are genes encoding thed-alanine-d-alanyl carrier ligase DltA, thed-alanyl carrier protein DltC, and the transmembrane proteins DltB and DltD, respectively. The length, location, and designation of the probes used in Northern blot analysis are indicated by heavy black lines (S1, S2, S3, and S4). The major transcripts initiated at PdltR and PdltA, which enable the expression of thedlt operon, are depicted by horizontal bold arrowheads. The locations of the pairs of primers used in RT-PCR analysis are shown below the transcriptional start site. The stem-loop structure represents the transcriptional terminator. (B) Northern blot analysis of the dlt operon of NEM316. Total RNAs were extracted from mid-exponential-phase cultures, subjected to denaturing agarose gel electrophoresis, and transferred onto a nylon membrane. Equal amounts of RNAs (40 μg) were applied per lane. The locations of the probes used (S1, S2, S3, and S4) are shown in panel A. (C) RT-PCR analysis of the dlt operon of NEM316. DNAs and RNAs extracted from mid-exponential-phase cultures were analyzed by PCR and RT-PCR, respectively, by using the pairs of primers LA32-LA19 (lane 1), LA33-LA19 (lane 2), LA34-LA8 (lane 3), LA35-LA36 (lane 4), and LA35-LA37 (lane 5).
dltR encodes a 224-aa-long polypeptide, DltR, that exhibits significant sequence homology with various response regulators (data not shown). In particular, its NH2-terminal moiety contains the 3 aa residues (D8, D51, and K100) that are invariably found in the corresponding region of response regulators from gram-positive bacteria (24). An inspection of the primary structure of DltR failed to reveal a canonical helix-turn-helix DNA-binding motif in the COOH-terminal domain of this protein. However, this region contains the 3 aa residues (R198, G207, and Y208) that are invariant in the COOH-terminal DNA-binding domains of response regulators belonging to the OmpR family (R209, G229, and Y230, according to the OmpR numbering) (23). This suggests that DltR is a member of the OmpR family of regulatory proteins.
The start codon of dltS overlaps the stop codon ofdltR by 2 bp, and this gene codes for a 395-aa-long putative histidine kinase designated DltS. This protein contains in its COOH half the four invariant residues (H180, N292, G353, and L354) found in the corresponding region of histidine protein kinases (24). Based on these sequence homologies, the histidyl residue at position 180 could correspond to the site of autophosphorylation. DltS contains membrane-spanning segments in its NH2-terminal half, which suggests that it is a membrane-associated protein likely involved in the sensing of environmental signals.
The dlt operon, which is responsible for the formation ofd-alanyl esters of LTA, is located 145 bp downstream ofdltS. The structure of this operon, which comprises four genes designated dltA, dltB, dltC, and dltD, is very similar to those found in other gram-positive bacteria (Fig. 1A). On the basis of sequence similarities, we assume that dltAcodes for a 511-aa cytoplasmicd-alanine-d-alanyl carrier protein ligase (designated DltA or Dcl) that catalyzes the d-alanylation of the 79-aa d-alanyl carrier protein DltC (or Dcp) encoded by dltC; dltB codes for a 421-aa transmembrane protein thought to be involved in the efflux of activated d-alanine to the site of acylation; and the 423-aa dltD gene product is a secreted protein that may serve to recognize nonalanylated acceptor LTA (25, 26). The gene dltB overlaps the stop codon of dltA by 4 bp, and dltD overlaps the stop codon of dltC by 8 bp, whereas dltB anddltC are separated by 14 bp. This genetic organization may result in translational coupling to provide coordinated synthesis of the corresponding proteins, as suggested in the case of thedlt operon of Lactobacillus casei(10). However, it is worth noting that both thedltB (GAGG) and dltD (GGAG) genes are preceded by a putative ribosome binding site, which suggests that their translation could be independent from that of the upstream gene. In fact, our complementation analysis indicates that translation of dltBis partly coupled to that of the upstream dltA gene.
Transcriptional analysis of the dlt operon of S. agalactiae NEM316.Total RNAs extracted from cells collected in mid-exponential phase were probed with various32P-labeled DNA fragments (S1, S2, S3, and S4) spanning thedlt operon (Fig. 1B). A large transcript of approximately 6.5 kb was detected with all four probes, whereas a transcript of 4.4 kb was detected with the S2, S3, and S4 probes only. The additional minor transcripts (5.5 and 2 kb) detected in this study were considered as corresponding to partially degraded or processed mRNAs. Sequence analysis revealed the presence of a palindromic sequence forming a possible stem-loop transcriptional terminator (Δ G = −13.8 kcal/mol) immediately downstream from dltD (Fig.1A). We interpret these data as indicating that the 6.5-kb transcript was initiated at a promoter PdltR located upstream of dltR and that the 4.4-kb transcript was initiated at a promoter PdltA located upstream of dltA. Both transcripts likely ended at the palindrome located downstream of dltD, since they were not detected when DNA fragments located downstream of this structure were used as probes (data not shown). To characterize PdltRand PdltA, the promoter activities of various amplicons cloned in the proper orientation upstream of the promoterlessspoVG-lacZ gene of pTCV-lac were studied in NEM316 cells collected in mid-exponential phase.
DNA fragments thought to contain PdltR were amplified by using the primers LA11 plus LA12and LA13 plus LA12 (Fig.2A). The amount of β-galactosidase synthesized from the LA11-LA12 amplicon (201 Miller units) was twice that obtained with the LA13-LA12 amplicon (Table2) (data not shown). The pTCV-lac derivative carrying the LA11-LA12 amplicon was used as a template to map the transcriptional start point (TSP) initiated at PdltR. A single TSP located immediately upstream of the dltR gene was located in this DNA fragment in NEM316, NEM1636, and NEM1637 (Fig. 3A) (data not shown). Sequence analysis of PdltR revealed a canonical −10 sequence preceded by a poorly conserved −35 sequence that is part of a short palindrome (Fig. 2A). The lower transcriptional activity of the LA13-LA12 amplicon, as compared to that obtained with LA11-LA12, may suggest that PdltR regulatory sequences are located in the 5′ region immediately upstream of LA13.
Nucleotide sequence of the DNA segments containing PdltR (A) and PdltA (B). The transcriptional start points are written in uppercase and marked with arrows. The positions of the presumed −35 and −10 sequences (underlined characters) and ribosome binding sites (RBS) are indicated above the sequences. Sequences forming putative palindromes are indicated by convergent horizontal arrows. The nucleotide sequences of the oligonucleotides used to amplify the promoter regions are boxed. Three inosine residues, marked with stars, were included in LA16 to remove an EcoRI site. The NH2-terminal sequence of DltR and the COOH-terminal sequence of DltS are written in the single-letter code below the nucleotide sequence in panels A and B, respectively. The numbering refers to the first base of each nucleotide sequence as presented.
Activities of PdltR and PdltA in various gram-positive bacteria
Primer extension analysis of PdltR-lacZ (A) and PdltA-lacZ (B) transcripts. RNAs were extracted from NEM316/pTCV-lac (lane 1), NEM316/pTCV-lacΩPdltR (LA11-LA12 amplicon) (lane 2), and NEM316/pTCV-lacΩPdltA (LA15-LA17 amplicon) (lane 3). Asterisks indicate the major transcriptional start sites. The sequences of the cloned amplicons are shown in Fig. 2.
The LA14-LA17 and LA15-LA16 amplicons were also cloned into pTCV-lac to characterize PdltA. These overlapping amplicons enabled the characterization of transcriptional start site(s) located in the 670-bp DNA segment upstream ofdltA, since the primer LA16 was designed to hybridize with the ribosome binding site of this gene (Fig. 2B). A similar low level of β-galactosidase activity (34 Miller units) was observed with both DNA fragments, which suggested that PdltA was located between LA15 and LA17 primers (Table 2) (data not shown). The pTCV-lac derivative carrying the LA15-LA17 amplicon was used as a template to map the TSP initiated at PdltA. A single TSP located in the 3′ extremity of dltS was detected in this DNA fragment (Fig. 3B). Sequence analysis of PdltArevealed a nearly canonical −10 sequence preceded by a poorly conserved −35 sequence, which is included in a short imperfect palindrome (Fig. 2B). Transcription initiated at PdltA should direct synthesis of the 4.4-kb mRNA detected by Northern blot analysis (Fig. 1B). Surprisingly, this mRNA was barely detectable with the S2 probe, whereas a strong signal was obtained with the S3 and S4 probes. This might indicate that the 5′ end of this mRNA is rapidly degraded or processed. This could alternatively indicate that a promoter stronger than PdltA is located in the 5′ extremity of dltA. The latter possibility is unlikely, since the 5′-half moiety of dltA did not possess any detectable promoter activity. Moreover, in this case, such a promoter would not direct synthesis of the entire dltA to -D operon.
Transcriptional initiation at PdltR was further characterized by RT-PCR in RNAs extracted from cells collected in mid-exponential phase. This was done by using the reverse primer LA19 and either of the forward primers LA32 and LA33, which hybridize immediately upstream (LA32) or downstream (LA33) from the transcriptional start site initiated at PdltR. An efficient amplification of the expected DNA fragment was obtained with the pair of primers LA33-LA19, whereas a barely detectable signal was obtained with the pair LA32-LA19 (Fig. 1C, lanes 1 and 2). Transcriptional termination at the palindrome located downstream ofdltD was studied similarly by using the forward primer LA35 and either of reverse primers LA36 and LA37, which hybridize upstream or downstream from the putative transcriptional terminator. As shown in Fig. 1C (lanes 4 and 5), amplification of the expected DNA fragment was only obtained with the primer pair LA35-LA37. Taken together, these results confirm the location of the transcriptional start site initiated at PdltR and demonstrate that the palindrome located downstream of dltB is an efficient transcriptional terminator. In these experiments, the primer pair LA32-LA19 was used as a positive control for RT-PCR analysis and the primability of each primer pairs used was assessed in PCR experiments (Fig. 1C).
d-Alanine ester contents in LTAs of S. agalactiae NEM316 and derivatives.The sequence of thedlt operon of NEM316 was used to inactivate thedltA (NEM1636) and dltR genes (NEM1637) by inserting, by double-crossover, a promoterless kanamycin resistance cassette within each gene. The dltA gene was reinserted into the chromosome of NEM1636 to give NEM1638 (Table 1). We determined that 20.8% of the glycerophosphate residues of the LTAs of the wild-type strain NEM316 were substituted with d-Ala ester. Insertional inactivation of dltA in S. agalactiaeNEM1636 caused complete absence of d-Ala ester from LTAs, whereas d-alanyl incorporation was partially restored into the LTAs of the complemented strain NEM1638 (12.8% ofd-Ala ester substitution). In this strain, the functionaldltA gene is transcribed from the PdltA, which is sixfold weaker than PdltR, and from a strong promoter located in the integrative vector that directs synthesis of the plasmid-borne spectinomycin resistance gene. We therefore concluded that the partial complementation observed in NEM1638 likely reflects the fact that translation of dltB is no longer coupled to that of the upstream dltA gene. Insertional inactivation ofdltR in NEM1637 (22.2% of d-Ala ester substitution) did not modify the d-Ala ester content of the LTAs compared to that of the wild-type strain.
Regulation of the promoters of the dlt operon ofS. agalactiae NEM316.The activities of PdltR and PdltAdetermined during the growth of NEM316 by measuring their abilities to direct β-galactosidase synthesis, were maximal and constant during the exponential phase of growth and diminished thereafter when the cells entered the stationary phase (Fig.4). The activities of these promoters were also assayed in heterologous gram-positive hosts such asEnterococcus faecalis and Listeria monocytogenesand compared to that of the constitutive promoter PaphA-3, which directs synthesis of the kanamycin resistance gene aphA-3 in various gram-positive bacteria (29). In S. agalactiae, the efficiencies of PdltR are 6- and 19-fold those of PdltA and PaphA-3, respectively (Table 2). In E. faecalis and L. monocytogenes, the activity of PdltR is similar to that of PaphA-3. Interestingly, PdltA was not active in these bacterial species, which suggests that its activity might require host factors that are specific for S. agalactiae.
Measurement of the β-galactosidase activities directed by PdltR and PdltA during the growth of S. agalactiae NEM316. Bacteria were grown in BHI broth containing erythromycin at 37°C. Samples were removed at the times indicated and assayed for turbidity (⧫) and β-galactosidase activities (◊, PdltR; □, PdltA). Experiments were conducted three times. The curves represent results from a representative experiment.
The β-galactosidase synthesis directed by PdltR and PdltA was also assayed in NEM1636 (DltA−), NEM1638 (DltA−/DltA+), NEM1637 (DltR−), NEM1639 (DltA− DltR−), NEM1687 (DltR−/DltR+), NEM1693 (DltS−), and NEM1786 (DltA− DltS−), to measure their activities in different genetic backgrounds (Table3). The activity of PdltR was found to be higher in NEM1636, NEM1638, and NEM1687 than in the wild-type strain NEM316 (2.5-, 1.5-, and 2-fold increases, respectively). In contrast, a twofold decrease in β-galactosidase activity was observed in NEM1637, NEM1639, NEM1693, and NEM1786 when compared to NEM316 (Table 3). The PdltA promoter was found to direct a similar level of β-galactosidase synthesis in NEM316 and NEM1637, but a threefold increase was observed in NEM1636. This promoter has a similar activity in NEM1638 and NEM1639, which is twofold that measured in NEM316 (Table 3). Taken together, these results indicate that the activities of PdltR and PdltA increase as the d-Ala ester content of the LTAs decreases. They also indicate that DltR positively regulates PdltR, but not PdltA, and that this regulation requires DltS. The activity of PdltR is significantly higher in NEM1687 than in NEM316. This is likely due to the fact that in NEM1687, the functional dltR gene associated with its natural promoter is inserted downstream from the plasmid-borne spectinomycin resistance gene.
Activities of PdltR and PdltA in S. agalactiae NEM316 and derivatives
Characterization of S. agalactiae NEM316 and derivatives.Comparison of the growth curves of NEM316 (wild-type strain), NEM1636 (DltA−), and NEM1637 (DltR−) when cultivated in BHI broth at 37°C did not reveal any significant differences (data not shown). However, we observed that in standing cultures, the mutants NEM1636 and NEM1639 (DltA−DltR−) formed clumps, whereas NEM316, NEM1637, and the complemented strain NEM1638 (DltA−/DltA+) did not. These clumps were visible by optic and scanning microscopy (Fig.5 shows part of this analysis). We thus concluded that the formation of clumps was related to the absence ofd-Ala ester in the LTAs of NEM1636 and NEM1639. We also observed that a significant number of NEM1636 cells possessed an aberrant morphology, being either poorly separated or multiseptated (data not shown). Western blot analysis of the surface proteins extracted from S. agalactiae NEM316 and derivatives revealed that proteins of approximately 73, 33, and 28 kDa were detected in the mutants NEM1636 and NEM1639, but not in NEM316 and NEM1637 (Fig.6). Interestingly, the 73-kDa protein was barely detectable in the extracts originating from the complemented strain NEM1638, which suggests that its presence depends on thed-Ala ester content of the LTAs. The activity of various antibiotics inhibiting cell wall synthesis (penicillin, vancomycin) or interfering with the cell membrane (colistin) was tested against NEM316 and derivative strains. The MICs of penicillin (0.047 μg/ml) and vancomycin (0.75 μg/ml) were identical with all strains tested, whereas that of the cationic peptide colistin was significantly higher with the wild-type strain NEM316 (>512 μg/ml) than those obtained with NEM1636 (32 μg/ml) and NEM1639 (32 μg/ml). The MIC of colistin seems to reflect the d-Ala ester content of the LTAs, since that of the complemented strain NEM1638 (128 μg/ml) was intermediate between those of NEM316 and the DltA− strains NEM1636 and NEM1639 (32 μg/ml).
Examination by scanning electron microscopy of S. agalactiae NEM316, NEM1636, and NEM1637 at two different magnifications. Note that NEM1636 forms large aggregates that are absent in NEM316 and NEM1637. The scale of each panel is indicated in the bottom right corner. WT, wild type.
Western blot analysis of SDS extracts of S. agalactiae NEM316 and derivatives. The proteins were electrophoresed under denaturing conditions on a 7.5% acrylamide gel and transferred onto a nylon membrane. The blot was developed with a rabbit serum (diluted 1:1,000) raised against NEM316. Lane 1, NEM316 (wild-type strain); lane 2, NEM1636 (DltA−); lane 3, NEM1637 (DltR−); lane 4, NEM1639 (DltR−DltA−); lane 5, NEM1638 (DltA−/DltA+). The three proteins (73, 33, and 28 kDa) detected in the DltA− mutants NEM1636 and NEM1639 are indicated by arrowheads. PM, protein size markers.
DISCUSSION
The dlt operon of gram-positive bacteria comprises four genes (dltA, dltB, dltC, and dltD) that catalyze the incorporation of d-alanine residues into the LTAs. In this work, we characterized the dlt operon ofStreptococcus agalactiae NEM316, which, in addition to thedltA to dltD genes, included two regulatory genes, designated dltR and dltS, located upstream of dltA. Twenty-one percent of the glycerophosphate residues of the LTAs of NEM316 were substituted with d-Ala ester, and insertional inactivation of the dltA gene encoding the cytoplasmic d-alanine-d-alanyl carrier protein ligase results in the complete absence of d-Ala esters from LTAs. The main phenotypic alterations caused by d-Ala ester deprivation were an increased susceptibility to the cationic antimicrobial polypeptide colistin, as already shown in staphylococci (27), and the ability of the corresponding mutant to form clumps in standing culture. We showed that additional cell surface proteins were extracted from this DltA− mutant, which might suggest that they preferentially bind to negatively charged LTAs. However, it remains to be demonstrated whether either of these proteins participates in intrageneric coaggregations by acting as an adhesin. These findings constitute the second report of a mutation in thedlt operon leading to altered adherence properties, but in the previous example, the consequences of a similar mutation inS. gordonii were opposite, since it resulted in the concomitant loss of an adhesin and of the intrageneric aggregative properties (7).
Although numerous dlt operons from gram-positive bacilli and cocci have been characterized, little is known about the regulation of their expression. In B. subtilis, this operon is controlled by a ςD-dependent promoter, and, consequently, its expression is activated during the late logarithmic growth and turned off before the transition phase by the concerted activities of the regulatory proteins SpoOA and AbrB (26). In S. mutans, maximal expression of the dlt operon occurs during the mid-log phase of growth when the medium contains carbohydrates internalized via the phosphoenolpyruvate phosphotransferase system (PTS), whereas it is constitutively expressed during all stages of growth in medium containing non-PTS sugars (35). However, the molecular basis of this regulation has not yet been characterized. The dlt operon of S. agalactiae included two regulatory genes that are located upstream of dltA and that encode the two-component regulatory system DltR-DltS. This operon is mainly transcribed from the PdltR promoter, which directs synthesis of a 6.5-kb transcript encompassing dltR, dltS, dltA, dltB, dltC, and dltD. The basal activity of PdltR(i.e., that measured in the DltR− mutant) increased fourfold in the DltA− mutant, and this upregulation depended on the presence of both DltR and DltS, since it was not observed in the DltR− DltA− and DltS− DltA− double mutants (Table 3). We thus concluded that PdltR is regulated by DltR in the presence of DltS and is activated in d-Ala-deficient LTA mutants. Since dltS encodes a putative membrane-associated histidine kinase, it is conceivable that the DltR-DltS regulatory system responds to an external signal related to the absence ofd-Ala esters of LTAs. The presence of d-Ala in the growth medium does not abolish the activation of PdltR in the DltA− mutant (data not shown), and efforts are currently being made to characterize the stimuli required for DltR-DltS activation. The basal activity of PdltR measured in the DltR− mutant is high, being 20-fold that of the constitutive promoter PaphA-3, which directs synthesis of the kanamycin resistance gene aphA-3 in various gram-positive bacteria (29), and this promoter was found to be active in heterologous gram-positive hosts such as E. faecalis andL. monocytogenes, which are devoid of sequences highly homologous to dltR and dltS (data not shown). This observation is consistent with the fact that inactivation ofdltR did not modify the d-Ala ester content of the LTAs compared to that of the wild-type strain. Thus, DltR and DltS are not required for, but modulate the expression of the dltoperon of S. agalactiae.
It is likely, at least in E. faecalis, that the amount ofd-Ala incorporated into the LTAs is limited by the availability of this amino acid in the cytoplasm (20). During the infectious process, S. agalactiae is exposed to hostile growth conditions, including nutrient starvation, and in this situation, the d-Ala ester content of its LTAs may decrease in response to the decreased availability of this precursor in the bacteria. It is conceivable that under such growth conditions. DltR and DltS may trigger the expression of the dlt operon to increase the formation of the d-Ala ester of LTAs. Unfortunately, this bacterium is not auxotrophic for alanine, which makes this hypothesis difficult to test in vitro. On the other hand,d-Ala ester residues of LTAs are susceptible to spontaneous base-catalyzed hydrolysis even at pH 7 (6, 14), and the existence of an enzyme-catalyzed hydrolysis has also been suggested (13). Thus, the role of DltR and DltS in the control of expression of the dlt operon could be to maintain the level of d-Ala esters in LTAs at a constant and appropriate value whatever the environmental conditions. The dlt operon is also transcribed from the PdltA promoter, which is located in the 3′ extremity of dltS and directs synthesis of a 4.4-kb mRNA. The activity of this promoter in the wild-type strain NEM316 is sixfold lower than that of PdltR and increases in the DltA− (threefold) and DltR−DltA− (twofold) mutants, which suggests that it is not regulated by DltR. Consistently, PdltR and PdltA do not share significant sequence homology. Unlike PdltR, PdltA is not active in E. faecalisand L. monocytogenes, which suggests a host-specific regulation.
In conclusion, we have shown that the dlt operon ofS. agalactiae is transcribed from two promoters that are upregulated when the amount of d-Ala incorporated into the LTAs decreases. This might indicate that the formation of thed-Ala ester of LTAs is essential for the lifestyle of this extracellular pathogen. Further support for this hypothesis comes from the observation that both the DltA− and DltR−mutants exhibit decreased virulence in a murine model (data not shown).
ACKNOWLEDGEMENTS
We thank S. Naı̈r and T. Msadek for critical reading of the manuscript, P. Berche for interest in this work and material support, and the “Centre Inter-Universitaire de Microscopie Electronique” for scanning electron microscopy.
This work was supported by the Institut National de la Santé et de la Recherche Médicale, the Pasteur Institute (PTR17), and the University of Paris V.
FOOTNOTES
- Received 16 April 2001.
- Accepted 8 August 2001.
- Copyright © 2001 American Society for Microbiology
