| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Microbiology, School of Dentistry, Aichi-Gakuin University, 1-100 Kusumoto-cho, Chikusa-ku, Nagoya, Aichi 464-8650, Japan
2 Oral and Maxillofacial Surgery II, School of Dentistry, Aichi-Gakuin University, 1-100 Kusumoto-cho, Chikusa-ku, Nagoya, Aichi 464-8650, Japan
Correspondence
Yukitaka Murakami
yukimu{at}dpc.agu.ac.jp
Received 14 March 2007
Accepted 24 July 2007
ragA mutant reduced RagB expression considerably and the ragB mutant produced degraded RagA. Growth of the three mutants in a nutrient-rich medium and synthetic media containing digested protein as a unique nutrient source was similar to that of the parental strain; however, both the ragA and ragAB mutants exhibited very slow growth in a synthetic medium containing undigested, native protein, and the two mutants tended to lose their viability during experiments, although gingipain (protease) activities were unchanged in the mutants. A mouse model showed that the ragB mutant had reduced virulence. Cell-surface labelling with biotin and dextran revealed that both RagA and RagB localized on the outermost cell surface. A cross-linking experiment using wild-type P. gingivalis showed that RagA and RagB were closely associated with each other. Furthermore, co-immunoprecipitation confirmed that RagA and RagB formed a protein–protein complex. These results suggest that physically associated RagA and RagB may stabilize themselves on the cell surface and function as active transporters of large degradation products of protein and in part as a virulence factor.
Abbreviations: CAT, chloramphenicol acetyltransferase; CBB, Coomassie brilliant blue; DAB, 3,3'-diaminobenzidine.
Present address: Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720-3202, USA. 
The GenBank/EMBL/DDBJ accession number for the P. gingivalis ATCC 33277T ragA, ragB and flanking region sequence is AB205195.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The outer membrane of the Gram-negative bacterium directly interacts with other bacteria, host cells and their environment. These interactions must be closely related to growth, colonization, biofilm formation, accomplishment of infection and development of diseases. We have established an isolation method for the outer membrane and identified major outer-membrane proteins designated Pgms 1–7 from P. gingivalis strain ATCC 33277T (hereafter referred to as 33277) (Murakami et al., 2002, 2004). Among the major outer-membrane proteins, Pgm1 and Pgm4 were identified as RagA and RagB, respectively, and their peptide fingerprinting patterns and N-terminal or internal amino acid sequences were compared with those of P. gingivalis strain W83, whose complete genome sequence has been published (Nelson et al., 2003). We have also identified major outer-membrane proteins from W83 (Imai et al., 2005). RagA is observed as a 110 kDa protein by SDS-PAGE in both strains, although the mobility of RagB is different, appearing as 47 and 55 kDa proteins in 33277 and W83, respectively.
RagA and RagB were originally identified as immunodominant surface antigens recognized by sera from periodontitis patients (Curtis et al., 1991). It has been reported that ragB, which is located adjacent to and downstream of ragA, is co-transcribed with ragA as a polycistronic message, and that expression of ragAB is influenced by temperature (Hanley et al., 1999; Bonass et al., 2000). RagA has homology to TonB-linked outer-membrane receptors, which are involved in the recognition and active transport of specific external ligands by a wide range of Gram-negative species (Curtis et al., 1999). RagB is predicted to be a lipoprotein from the primary amino acid sequence and is more diverse than RagA among the species, with only short regions of sequence conservation (Hall et al., 2005; Imai et al., 2005). Based on circumstantial evidence, Hanley et al. (1999) claimed that RagA and RagB might form a functionally linked complex on the outer surface of P. gingivalis, involved in a TonB-dependent, active process.
Both 33277 and W83 are widely used for P. gingivalis studies. 33277 has both long and short fimbriae, encoded by fimA and mfa1, respectively (Yoshimura et al., 1984; Hamada et al., 1996), whereas W83 lacks both of these, although it possesses apparently intact fimA (PG2132, as annotated in the P. gingivalis W83 genome database) and mfa1 (PG0178) disrupted with an insertion element (ISpg4). As whole-genome sequencing of 33277 has not yet been carried out, we wanted to analyse the rag locus from 33277.
In this study, we determined the DNA sequence of the ragAB region from 33277 and compared it with that from W83. Then we constructed three mutants from W83 by deleting the ORFs of ragA and/or ragB, and examined the physiological and pathological functions of RagA and RagB. We also present experimental evidence that RagA and RagB, localized on the cell surface, are physically and functionally associated with each other.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
(Gardner et al., 1996; Nelson et al., 2003). All P. gingivalis strains were grown at 37 °C under anaerobic conditions [10 % CO2, 10 % H2 and 80 % N2] on Brucella HK agar (Kyokuto Pharmaceutical) supplemented with 5 % laked rabbit blood, 2.5 µg haemin ml–1, 5 µg menadione ml–1 and 0.1 mg dithiothreitol ml–1 (BHK agar), and in trypticase soy broth (Becton, Dickinson) supplemented with 2.5 mg yeast extract ml–1, 2.5 µg haemin ml–1, 5 µg menadione ml–1 and 0.1 mg dithiothreitol ml–1 (sTSB). For growth experiments, we also used Dulbecco's modified Eagle's medium (DMEM) supplemented with 1 % (w/v) tryptone (Becton, Dickinson), 1 % (w/v) neopeptone (Becton, Dickinson), 1 % (w/v) BSA (fraction V; Miles), or 1 % (w/v) BSA digested with 0.25 % trypsin (Invitrogen) for 10 h at 37 °C. Escherichia coli strains were cultivated in Luria–Bertani medium supplemented with the necessary antibiotics. Bacterial growth was monitored by measuring the OD660.
|
DNA sequencing. A purified PCR product and plasmid DNA were usually used as templates for the DNA cycle sequencing with a BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems). The products of the DNA cycle sequencing reaction were purified and analysed using an ABI PRISM 3100-Avant Genetic Analyzer (Applied Biosystems).
ragA, ragB and their flanking regions in the 33277 genome were sequenced. We previously examined the N-terminal and several internal amino acid sequences of RagA and RagB of 33277 (Murakami et al., 2002, 2004). Based on these amino acid sequences, primers were designed and the internal region between ragA and ragB was amplified by PCR and sequenced. Next, the outside sequence of the region was determined by inverse PCR (Ochman et al., 1988). Briefly, the 33277 genome was digested with PshBI and then ligated to generate a pool of monomeric DNA circles. Using the DNA circles as templates, the outside region was amplified by PCR, and the amplified fragment was subjected to DNA sequencing.
Construction of deletion mutants from W83.
Deletion mutants were constructed by the PCR-based overlap-extension method (Horton et al., 1993) essentially as described previously (Nagano et al., 2005). The primers and their annealing sites are shown in Table 2 and Fig. 2(a, b), respectively. The cat gene, encoding chloramphenicol acetyltransferase (CAT), was amplified with primers AGU01 and AGU02 from pKD260. For construction of the ragA-deletion cassette, the flanking sequence upstream of ragA was amplified with primers AGU51 and AGU52, which have homology to the 5' end of the cat fragment. The flanking sequence downstream of ragA was amplified with AGU56 and AGU55, which have homology to the 3' end of the cat fragment. The cat, ragA-upstream and ragA-downsteam fragments were used as templates for overlap-extension PCR to generate a deletion cassette in which ragA was replaced by cat. As the inserted cat gene contained only its reading frame without a promoter/terminator, its transcription was dependent on the rag promoter, and a downstream gene such as ragB was expected to be transcribed by the read-through mechanism (Nagano et al., 2005). The deletion cassettes of ragB and both ragA and ragB were generated by similar procedures. Each deletion cassette was ligated into pCR-Blunt II-TOPO (Invitrogen) and the resulting recombinant plasmids were transformed into E. coli TOP10 (Invitrogen). A plasmid construct carrying each deletion cassette was sequenced before electroporation to rule out unintended base changes. For electroporation of P. gingivalis, the plasmid constructs were linearized by digestion with the endonucleases SpeI and XbaI, and introduced into electrocompetent cells of W83. After 6 h of anaerobic incubation in sTSB, the pulsed cells were plated on BHK agar supplemented with 10 µg chloramphenicol ml–1 and the plates were incubated anaerobically for 7 days.
|
|
ragB) was transformed with the shuttle vector pTCB bearing a DNA fragment that linked the promoter region of rag or fimR with the ragB ORF from W83 or 33277. To generate restriction sites at the 5' and 3' ends of the promoters by PCR, the primers ragPNotF and ragPBamR were used to generate a 482 bp fragment of the rag promoter, and fimRPSpeF and fimRPBamR to generate a 460 bp fragment of the fimR promoter (Nishikawa & Yoshimura, 2001) (Table 2
). The fimR and rag promoters amplified and digested with appropriate restriction enzymes (SpeI/BamHI and NotI/BamHI, respectively) were ligated into pTCB digested with the same enzymes to create pTCBex and pTCBex2, respectively (Table 1). The ragB ORF of W83 (ragBW83) was amplified with primers W83RagBBamF and W83RagBHindR from the W83 genome. The ragB ORF of 33277 (ragB33277) was amplified with primers 33277RagBBamF and 33277RagBHindR from the 33277 genome. Amplified ragBW83 or ragB33277 was digested with BamHI and HindIII and inserted into pTCBex or pTCBex2 digested with the same enzyme.
The recombinant plasmid vector was introduced into E. coli S17-1 by electroporation. We confirmed that the DNA sequence of the insert had no unintentional base changes after extraction of the plasmid from the transformants. The purified plasmid was then introduced into ragB via conjugation under aerobic mating conditions at 37 °C for 12 h and the mated bacteria were spread on BHK agar containing 100 µg gentamicin ml–1 and 1 µg tetracycline ml–1 (Nishikawa & Yoshimura, 2001). After the plates had been incubated anaerobically for about 1 week at 37 °C, the plasmid was extracted again from black-pigmented transformants and verified by PCR and restriction enzyme digestion analysis.
Cell fractionation, SDS-PAGE and Western blotting. Preparation of the bacterial whole-cell lysate and envelope fraction, SDS-PAGE and Western blotting analyses were performed essentially as described previously (Murakami et al., 2002; Nagano et al., 2005). For co-immunoprecipitation, whole-cell lysate was prepared using BugBuster HT protein extraction reagent according to the manufacturer's instructions (Novagen). The gels were stained with Coomassie brilliant blue R-250 (CBB). As primary antibodies in Western blotting, we used antigen-specific antisera against RagA and RagB obtained from a rabbit immunized with purified RagA protein from 33277 (Murakami et al., 2004) and recombinant RagB constructed on the basis of the W83 sequence (Imai et al., 2005), respectively.
Animal experiments. P. gingivalis strains were grown in sTSB for 24 h until the early stationary phase, harvested and washed twice with 10 mM PBS (pH 7.4). The bacterial concentration (c.f.u. ml–1) was adjusted with PBS by reference to a standard curve of concentration against OD660. The numbers of viable cells prepared were counted and confirmed by spreading the bacterial suspension on BHK agar for every experiment. Six-week-old BALB/c mice were challenged with a dorsal subcutaneous injection of 0.2 ml of bacterial suspension. General health, weight and the appearance and location of lesions were assessed daily. To determine bacteraemia, spleens were excised under aseptic conditions, placed in sTSB, minced and suspended gently by pipetting. Different dilutions of these suspensions were spread on BHK agar and cultured anaerobically for 1 week. Black colonies, which are characteristic of P. gingivalis, were counted. Only a few colonies without black pigmentation were observed. The black colonies randomly selected were confirmed as P. gingivalis by microscopic examinations such as Gram staining. Statistical analyses were performed using Student's t-test. Values that were significantly different from control values (P <0.05) are indicated by asterisks in the figures. These animal experiments were performed in accordance with the Guidelines for Animal Experiments at the School of Dentistry, Aichi-Gakuin University.
Antimicrobial susceptibility. The antibiotics used were ampicillin (Wako Pure Chemical Industries), cefotaxime, cefuroxime, cephalexin, ceftriaxone, chloramphenicol, erythromycin, tetracycline, minocycline and norfloxacin (Sigma-Aldrich). MICs were evaluated by an agar double-dilution assay, as recommended by the NCCLS (1997). Briefly, serial dilutions of the corresponding antibiotics were added to BHK agar. After the bacteria had been grown to the early stationary phase in sTSB, 2 µl bacterial culture was spotted on the antibiotic-containing agar. After 2 days of anaerobic incubation, the susceptibility breakpoints were determined.
Electron microscopy. P. gingivalis cells were grown to the early stationary phase in sTSB and directly spotted onto carbon-coated grids. After negative staining with 2 % (w/v) uranyl acetate, the cells were observed using a JEM-1210 electron microscope (JEOL).
Cell-surface labelling. P. gingivalis cells suspended in 10 mM PBS (pH 8.0) were labelled with sulfo-NHS-LC-biotin (Pierce) at 4 °C for 2 h. The reaction was stopped by the addition of 0.1 M glycine. Dextran labelling was carried out by the method of Kamio & Nikaido (1977). Briefly, Dextran T-10 (Pharmacia) activated with CNBr was added to P. gingivalis cells suspended in 0.1 M sodium carbonate buffer (pH 8.5). Coupling was carried out at 20 °C for 1 h and terminated by the addition of 1 M ethanolamine/HCl (pH 7.4). After labelling, envelope fractions were subjected to SDS-PAGE and blotted onto nitrocellulose membranes. Biotinylation was detected using peroxidase-conjugated streptavidin using 3,3'-diaminobenzidine (DAB). Dextran-labelled proteins were detected by Western blotting using anti-RagA and anti-RagB antibodies developed with a Pro-Q Western blot stain kit (Invitrogen).
Chemical cross-linking. Washed P. gingivalis cells were suspended in 0.1 M triethanolamine/HCl buffer (pH 8.5). The cross-linker disuccinimidyl suberate (11 Å length; Sigma-Aldrich) dissolved in DMSO was added to a final concentration of 50 mM. After incubation of the reaction mixture at 4 °C for 2 h, the cross-linking reaction was stopped by the addition of 0.1 M Tris/HCl buffer (pH 8.0). The envelope fraction was prepared from the cells and analysed by SDS-PAGE and Western blotting using anti-RagA and anti-RagB antibodies developed with DAB.
Co-immunoprecipitation. Co-immunoprecipitation was performed using a ProFound Co-Immunoprecipitation kit (Pierce) according to the manufacturer's instructions. Briefly, an anti-RagA or anti-RagB antibody was coupled to AminoLink Plus Gel. The slurry of antibody-coupled beads was incubated with whole-cell lysate from W83 with gentle rocking overnight at 4 °C. Beads were washed three times with Dulbecco's PBS. The bound proteins were eluted from the beads with ImmunoPure IgG Elution Buffer, immediately neutralized with 1 M Tris/HCl (pH 9.5) and analysed by SDS-PAGE.
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
. For RagA, a sequence identity of 70 % and a sequence similarity of 81 % were found between the two strains, and the N- and C-terminal domains were almost identical. The regions boxed in RagA in Fig. 1 are conserved in TonB-linked proteins (Hanley et al., 1999). These regions were completely or highly conserved between the strains. In contrast, RagB in W83 and 33277 showed only 48 % identity and 63 % similarity over the whole region. When the nucleotide sequences of W83 and 33277 were also compared, ragA showed a higher sequence identity (72 %) than ragB (59 %). The ragB locus including ragA was found only in 17 of the 132 P. gingivalis strains by Southern blot analysis using a DNA probe (436 bp) derived from P. gingivalis strain W50 and was present mostly in virulent strains, but not in 381 and 33277 (Hanley et al., 1999; Frandsen et al., 2001). However, we have reported that 33277 expresses both RagA and RagB in the outer membrane, based on peptide mapping and internal sequence analysis (Murakami et al., 2002, 2004) and Western immunoblotting (Imai et al., 2005). In this study, we verified by sequencing that the genes ragA and ragB were present in the 33277 chromosome. More recently, Hall et al. (2005) reported sequence diversity and antigenic variation at the rag locus, which was classified into four alleles. The rag locus was in the same genomic location; however, there was some polymorphism. W83 and 33277 represented rag-1 and rag-4, respectively, in the report. The ragAB nucleotide sequence from 33277 in our study (GenBank accession no. AB205195) was identical to that derived by Hall et al. (2005) (GenBank accession no. AY842852). The degree of similarity between RagB variants (43–56 % identity at the protein level and 53–62 % identity at the nucleotide level) would not have been detected by Southern blotting under high-stringency conditions (Frandsen et al., 2001).
|
. To confirm the replacement of target ORFs by cat, the cloning sites were amplified from flanking regions by PCR using each mutant chromosome as a template with primers of AGU59 and AGU60 designed to anneal outside the cloning sites as shown in Fig. 2(b). Amplified PCR products were then digested with selected restriction enzymes to verify the genotypes of the putative mutants. As shown in Fig. 2(c), gel electrophoresis of digested DNA yielded the predicted fragments. Thus three mutants of W83 with deletion of ragA (ragA), ragB (ragB) or both (ragAB) were obtained.
Complementation of RagB
Complementation analysis was performed with shuttle plasmid pT-COW derivatives incorporating the ragA and fimR promoters, presumed to be strong and weak promoters, respectively (Gardner et al., 1996; Nishikawa & Yoshimura, 2001). Although we tried to construct a series of ragB complemented with ragB, namely ragB from W83 (ragBW83) or ragB from 33277 (ragB33277) under the rag promoter and ragBW83 or ragB33277 under the fimR promoter, we obtained only two strains: ragBW83 downstream of the rag promoter (ragB+ragBW83) and ragB33277 downstream of the fimR promoter (ragB+ragB33277) (Table 1). Although further research is required to determine the reason for this, various explanations include the possibility that: (i) RagB33277 may not be compatible with RagBW83 because of their low homology (Hall et al., 2005); and (ii) full expression of RagB33277 via the rag (strong and authentic) promoter in ragB in the presence of RagAW83 may cause cell viability to deteriorate, resulting in failure to establish the complementation strain carrying pTCBex2 : : ragB33277. Also, insufficient expression of RagBW83 via the fimR (weak) promoter may make cells unstable in the presence of sufficient RagAW83 in ragB as shown in the following section. We did not attempt to complement ragA into ragA because of the lower viability during experiments with the mutant (see section on physical and morphological characteristics later).
SDS-PAGE and Western blot analyses of envelope fractions
Envelope fractions denatured in SDS with 2-mercaptoethanol were subjected to SDS-PAGE and visualized with CBB (Fig. 3a). The 110 kDa band of RagA was clearly observed at high intensity in 33277 and W83, and at low intensity in ragB, whereas such bands were not observed in ragA or ragAB. Complemented ragB+ragBw83 was able to recover the 110 kDa band (RagA) with higher intensity than ragB, whereas RagA was recovered in ragB+ragB33277 at a low level similar to ragB. RagB bands of high intensity were clearly observed at 47 and 55 kDa in 33277 and W83, respectively, and with low intensity in ragB+ragBw83; however, corresponding bands were not observed in the other strains, ragA, ragB, ragAB and ragB+ragB33277. RagA and RagB could be detected on SDS-PAGE gels even in whole-cell lysates, although their contents were higher in the envelope fraction, indicating that the proteins in the mutants localized on the membrane as in the parent.
|
) and RagB (Fig. 3c). The anti-RagA antibody detected a strong band of 110 kDa in 33277 and W83 and several weaker bands of lower molecular masses, which appeared to be degradation products. In ragB, a new strong 60 kDa band was detected, although a weak signal was seen at the 110 kDa position. Such bands were not observed in ragA or ragAB. ragB+ragBw83 was able to recover a strong 110 kDa band together with several lower-molecular-mass bands, including a faint 60 kDa band, although overall signals appeared to be slightly weaker than those of W83. In contrast, in ragB+ragB33277, the signal intensity of RagA did not increase; the 110 kDa band and several other bands were not recovered to the level of ragB+ragBw83. The anti-RagB antibody detected weak and strong signals in 33277 (47 kDa) and W83 (55 kDa), respectively. As antiserum against the RagB antigen from W83 was used, the signal for RagB from 33277 was weak, indicating an antigenic difference between the two strains (note the two equivalent CBB-stained RagB bands of lanes 1 and 2 in Fig. 3a compared with the two clearly different immunostained bands in the same lanes in Fig. 3c). The migration position of RagB following SDS-PAGE was also clearly different between the two strains. The precise reason is unclear; however, we speculate that the degree of lipid modification might be different in lipoprotein RagB, as the calculated molecular mass and isoelectric point were almost the same in the two strains. When the protein motif was analysed by the Motif Scan program (http://myhits.isb-sib.ch/cgi-bin/motif_scan), nine possible N-myristoylation sites were detected in 33277, whereas there were four sites in W83. RagB was not detected in ragB and ragAB, whereas a faint band was detected in ragA. The marked decrease in RagB may not be a polar effect in ragA as the inserted cat gene was supposed to be a non-polar cassette. Instead, the decrease in RagB may have been caused by a deficiency in RagA due to instability of RagB alone in the outer membrane as discussed later.
As production of RagA was decreased by deletion of RagB as described above, both RagA and RagB may exist as mutually associated structures, and their interaction is probably important in their stabilization in cells. Expression of RagA and RagB in ragB+ragBw83 was almost fully recovered and the degradation product of RagA (60 kDa protein) disappeared. In contrast, recovery of RagA and RagB in ragB+ragB33277 was marginal, although the 60 kDa protein derived from RagA decreased and RagB33277 was not detected, partially due to the use of an antiserum with less specificity for RagB33277. This also suggested that RagB33277 under the fimR promoter was not expressed sufficiently and might not interact with RagAW83 adequately because of their incompatibility.
Physical and morphological characteristics
ragA and ragAB tended to lose their viability during experiments. We therefore investigated whether they became more oxygen sensitive than the parental strain W83. However, we did not observe any difference in oxygen sensitivity between them (data not shown). These mutants probably had decreased resistance to physical stresses such as centrifugal force, mixing and pipetting. It is possible that the complete disappearance of RagA existing on the outer membrane in large amounts might destabilize the membrane and make it fragile. Such fragility was not observed in ragB, presumably due to the presence of RagA to some degree. We did not observe any morphological changes in the three mutants when they were compared with W83 by negative staining in electron microscopy (data not shown).
Growth in nutrient-rich and synthetic media
An early stationary-phase culture in sTSB, a nutrient-rich medium, was inoculated at a 1 : 20 ratio into several media. In sTSB (Fig. 4a), the three mutants ragA, ragB and ragAB showed growth rates similar to the parental strain W83 and reached the stationary phase within 24 h. The three mutants also showed growth curves similar to the parental strain in DMEM supplemented with 1 % tryptone (Fig. 4b) and 1 % neopeptone (data not shown). In DMEM supplemented with 1 % BSA, the parental strain reached the stationary phase within 36 h, slightly later than in sTSB, and ragB reached it even later, whereas ragA and ragAB grew considerably slower and reached the same level as the parental strain at about 72 h (Fig. 4c). However, in DMEM supplemented with BSA pre-digested with trypsin, ragA and ragAB showed growth curves similar to the parental strain and ragB (Fig. 4d). It has been reported that P. gingivalis utilizes proteins such as serum albumin as nutrition by digestion with gingipains (Grenier et al., 2001). The amounts and activities of gingipains were almost the same among the strains used here (data not shown). Therefore, the mutants, especially ragA and ragAB, deficient in RagA and RagB, might not uptake larger peptides and thus would need more time to grow until smaller peptides become available via further digestion of BSA with the many proteases in this organism. Digestion of BSA by trypsin or tryptone, a pancreatic digest of casein, could diminish such growth lags, resulting in no apparent difference in growth curves. RagAB could play a role in the active transport of macromolecules such as larger peptides that are not passively permeable to the outer membrane in this organism. This is the first experimental evidence that this assumption is likely, although RagAB has been suggested to play a role in the uptake of macromolecules such as polysaccharides or glycoproteins based on sequence comparison with Bacteroides thetaiotaomicron SusCD (Curtis et al., 1999; Hall et al., 2005). However, further work is required to elucidate the exact function of RagAB.
|
Bacterial cells of W83 (2x109 or 1x109 c.f.u.) or ragB (4x109 or 2x109 c.f.u.) were injected into five mice each. ragA and ragAB were not used in animal experiments because they tended to lose their viability during the experimental procedure, as described above. One day after injection, ruffled hair was observed in every mouse. Although body weight fell from 2 to 4 days after injection, it then recovered and began to increase again. All of the animals injected survived for the experimental period of 2 weeks because the cell numbers used were much lower than those used by others (Kumagai et al., 2000). Abdominal ulcers were observed within 2 days in all mice injected with 2x109 c.f.u. W83 and in two mice injected with 1x109 c.f.u W83. Ulcer formation was also observed in all mice within 2 days when 4x109 c.f.u. ragB was injected; however, ulcer size was clearly smaller than that induced by 2x109 c.f.u. W83. Only small ulcers were formed in three mice after injection of 2x109 c.f.u. ragB. These results suggested that ragB might be less virulent than the parental W83.
To examine the difference in virulence between the strains quantitatively, bacterial cell numbers in the spleen were also evaluated. As a pilot study, 1x109 c.f.u. W83 was injected and bacterial cells in the spleen were counted in three mice every day from days 1 to 7. P. gingivalis was detected in all three mice at day 1 and in two of the three mice at day 2. At days 3 to 5, a small number of P. gingivalis cells were detected in a mouse with a large lesion. However, at days 6 and 7, P. gingivalis was not detected in any mouse. P. gingivalis was not detected in mice injected only with PBS as a control. Therefore, we decided to inject 1x109 c.f.u. W83, ragB or ragB+ragBW83 to count bacterial cell numbers in the spleen at days 1 and 2. At days 1 and 2, 104–107 c.f.u. W83 per spleen was detected in all four mice. In contrast, much lower cell numbers were detected in two mice and one mouse out of four at days 1 and 2, respectively, and almost no bacteria were detected in five other mice in the case of ragB (Fig. 5). Unexpectedly, bacteria were not detected when the complementation strain ragB+ragBW83 was used. We do not know the exact reasons for this, but there are at least two possible explanations. First, although RagB almost recovered to the normal level, RagA was not fully stabilized, as shown in Fig. 3. Secondly, the strain, bearing a recombinant plasmid vector with a large size, might become weak in vivo.
|
Antimicrobial susceptibility and other characteristics
We also examined whether deletion of RagA and/or RagB influenced MICs to various antibiotics. The MIC of chloramphenicol for the wild-type was 4 µg ml–1, whilst for the mutants it was 64 µg ml–1, presumably due to the strong CAT expression. Strong CAT activity was detected only in mutants where cat had been introduced (data not shown). However, no significant difference in the MICs of other antibiotics was observed between the wild-type and mutants. Moreover, the various antibiotics used were effective against the P. gingivalis strains, showing considerably low MICs. It has been reported that P. gingivalis is susceptible to most antibiotics, including ß-lactams, partially due to the absence of ß-lactamase (Andrés et al., 1998; Kleinfelder et al., 1999; Ikeda & Yoshimura, 2002). Thus we concluded that RagA and RagB might not be involved in antimicrobial susceptibility.
In addition, we tested the effects of deletion of RagA and/or RagB on haemagglutinating activity as one of the representative characteristics of P. gingivalis; however, no difference from the parental strain was observed, implying that they were unrelated to this characteristic (data not shown).
Cell-surface labelling
To examine localization of RagA and RagB, we used different labelling molecules, namely biotin and dextran. Both RagA and RagB were strongly detected by a streptavidin-linked reagent after biotinylation (Fig. 6b), based on comparison with patterns after amido black staining (Fig. 6a). Sulfo-NHS-LC-biotin has been demonstrated predominantly to label cell-surface proteins of Gram-negative bacterial cells (Bradburne et al., 1993). However, as this reagent is relatively small (Mr 557) and hydrophilic, it may penetrate the outer membrane through diffusion pores such as porins, and outer-membrane proteins might be labelled from the periplasmic space. Therefore, we used macromolecular dextran as another labelling reagent that could not pass through the outer-membrane pores. After CNBr-activated dextran labelling, the intensities of RagA and RagB bands decreased, whereas thick smears appeared at the top of the separation gel (Fig. 7a), indicating that RagA and RagB were shifted up by conjugation with dextran. Western blotting using anti-RagA and anti-RagB antibodies confirmed that large quantities of both RagA and RagB existed in the high-molecular-mass smears after labelling (Fig. 7b, c). Consequently, these results suggested that both RagA and RagB were exposed on the cell surface.
|
|
). Western blotting using specific antibodies confirmed that most RagB migrated to much higher positions as high-molecular-mass bands, although RagA behaved in a slightly different manner (Fig. 8b, c). These results suggested that RagA and RagB were in close proximity and physically interacted with each other.
|
). Essentially, no other bands were stained. This strongly suggested that RagA and RagB have a physical molecular interaction that permits them to be chemically cross-linked as described above.
|
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. E. Zeituni, R. Jotwani, J. Carrion, and C. W. Cutler Targeting of DC-SIGN on Human Dendritic Cells by Minor Fimbriated Porphyromonas gingivalis Strains Elicits a Distinct Effector T Cell Response J. Immunol., November 1, 2009; 183(9): 5694 - 5704. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Hasegawa, J. Iwami, K. Sato, Y. Park, K. Nishikawa, T. Atsumi, K. Moriguchi, Y. Murakami, R. J. Lamont, H. Nakamura, et al. Anchoring and length regulation of Porphyromonas gingivalis Mfa1 fimbriae by the downstream gene product Mfa2 Microbiology, October 1, 2009; 155(10): 3333 - 3347. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Mazumdar, E. S. Snitkin, S. Amar, and D. Segre Metabolic Network Model of a Human Oral Pathogen J. Bacteriol., January 1, 2009; 191(1): 74 - 90. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
, P. gingivalis W83 (parental strain); 
, 
, 
, 
