Localization of HArep-Containing Genes on the Chromosome of Porphyromonas gingivalis W83

Abstract

We have mapped a group of virulence genes of Porphyromonas gingivalis to a single large fragment of the genome. These genes (rgpA, kgp, and hagA) all contain a consensus repeat sequence (HArep). rgpA and kgp encode cysteine proteases with Arg-X and Lys-X specificity, respectively, and hagA encodes a hemagglutinin. Genomic DNA fragments separated by pulse-field gel electrophoresis were blotted and probed in order to localize the genes to a 0.25-Mb NheI fragment of the P. gingivalis W83 genome. Further hybridization analyses with single- and double-restriction digestion allowed us to generate a physical map of the fragment and determine the precise locations of the protease and hemagglutinin genes. In addition, we found an insertion-like sequence, IS195, near the ends of the 0.25-Mb NheI fragment. A similarly sized fragment carrying HArep sequences was also demonstrated in the P. gingivalis W12 and W50 genomes.

Porphyromonas gingivalis is an anaerobic, asaccharolytic bacterium that is recognized as an important etiologic agent in adult periodontitis (7, 18, 34). Virulence factors of P. gingivalis identified in the mouse abscess animal model include the cysteine proteases Arg- and Lys-gingipain (10, 20). An allelic-exchange mutant of P. gingivalis W83 deficient in arginine-specific cysteine protease activity displayed reduced virulence in this animal model (10). Similar results have been obtained with naturally occurring and allelic-exchange lysine-specific protease mutants of P. gingivalis W83 (20). The specific role of such proteases in virulence has not been elucidated, but they might contribute to the ability of the bacteria to colonize the oral cavity by the exposure of cryptic sites and binding to an extracellular matrix, the evasion of host defense mechanisms through the hydrolysis of immunoglobulin and complement proteins, and the alteration of neutrophil antimicrobial activity by degradation of bactericidal proteins and acquisition of essential nutrients (1, 14, 15, 17, 24, 32, 35, 36).

While the cysteine protease activity with arginine specificity originates from two different genes, rgpA and rgpB (24, 31), the lysine-specific cysteine protease activity is derived from a single gene, kgp (formerly called prtP) (20). Structure-function studies of the cysteine proteases have revealed that rgpA and kgp encode domains specifying different functions (protease activity and hemagglutinin/adhesin activity) (Fig. 1) (2, 4, 26–29). Nucleotide and protein sequence analyses revealed that the sequence encoding the hemagglutinin/adhesin domain shows a high degree of homology to the hemagglutinin gene hagA (Fig. 1). A nucleotide sequence of approximately 1.3 kb found in both the rgpA and kgp genes is also present in tandemly repeated multiple copies in the hagA gene (Fig. 1) (12, 16). This sequence is known as the HArep consensus (12). Both in vitro and in vivo studies have provided evidence that the protein sequence (HArep) encoded by this region might have a role in virulence properties. Curtis et al. (6) found that a hemagglutination-inhibiting monoclonal antibody, 1A1, recognized an amino acid sequence within the HArep. Another monoclonal antibody, 61BG1.3, recognized a similar sequence present within the HArep, and this immunoglobulin inhibited hemagglutination and prevented P. gingivalis recolonization in periodontal patients for up to 9 months (5). In addition, the HArep contains the hemoglobin receptor domain (25). Protoheme is an essential nutrient for P. gingivalis and is probably derived from erythrocytes present in the organism’s natural niche. Thus, it appears critical that P. gingivalis be able to attach to both erythrocytes and hemoglobin in order to survive in its host. Furthermore, the HArep consensus is flanked by 138 bp of conserved repeated nucleotide sequence (CRS) encoding 46 amino acids (Fig. 1). This amino acid sequence contains motifs implicated in binding to fibronectin, collagen, and laminin (37). Other studies have demonstrated (30) that arginine- and lysine-specific cysteine proteases have the ability to bind to fibrinogen, fibronectin, and laminin. Southern blot analyses have demonstrated multiple DNA fragments bearing the CRS among P. gingivalis strains (4, 10). In order to evaluate the role and regulation of genes containing the CRS, we characterized all genes containing the CRS in P. gingivalis W83. By cloning DNA fragments containing the repeated sequence we demonstrated that the CRS was present along with HArep in the same genes: hagA (three HArep copies in P. gingivalis W83), rgpA, and kgp (Fig. 1). We have used this information to locate and position HArep-containing genes on the W83 genome.

FIG. 1.

(Left) Genetic maps of P. gingivalis cysteine proteases with Arg-X and Lys-X specificity and hemagglutinins/adhesins. rgpA, encoding protein composed of signal peptide (“s”), prepropeptide (“propt.”), a protease domain (“protease”), and a hemagglutinin domain(s) (“hemagglutinin”), is depicted at the top. Numbers below the map refer to the sizes, in kilobases. The structure of polypeptide encoded by kgp resembles that of RgpA. Shadings and patterns are meant to convey nucleotide sequence similarity. CRSs are indicated by small black boxes and roman numerals. The insertion-like sequence IS195 and a truncated copy of IS1126 (ΔIS1126) flanking kgp are shown. The broken line indicates sequences with some similarity. Four tandem HArep repeats present within hagA are indicated below the map of the gene. A single copy of the repeat is shown in rgpA and kgp. Sequences used as probes in this study are depicted by outward-facing arrowheads and lowercase letters: a, 0.6-kb PCR amplicon obtained with primers complementary to the rgpA sequence (forward, 5′ ATCGTTGCGCATCTATACGG 3′ [positions 1488 to 1508], and reverse, 5′ TCACCATTGTCTGCGGATGG 3′ [positions 2111 to 2091]); b, PCR amplicon obtained with primers complementary to the rgpA portion of the gene carrying the HArep (forward, 5′ CATGAGTATTGCGTGGAAGT 3′ [positions 3623 to 3643], and reverse, 5′ CTTCGTACCGTCACGATACAC 3′ [positions 4940 to 4919]); c, PCR amplicon constructed with primers complementary to the internal portion of IS195 (20); d, 0.45-kb HindIII fragment of the kgp gene encoding the prepropeptide domain (830 to 1,283 bp). (Right) Effects of mutations in rgpA, kgp, and hagA on P. gingivalis (arrows represent reduction in indicated function). hemaggl., hemagglutinin.

P. gingivalis W83 (from H. A. Schenkein, Virginia Commonwealth University), W12 (from A. Progulske-Fox, University of Florida), and W50 (from M. Curtis, Medical Research Council, St. Bartholomew’s, and the Royal London School of Medicine and Dentistry, London, England) were used in this work. P. gingivalis strains were grown in brain heart infusion broth (Difco Laboratories, Detroit, Mich.) supplemented with hemin (5 μg/ml), vitamin K₃ (0.5 μg/ml), and cysteine (1%).

The DNA probes used to localize HArep, rgpA, kgp, and IS195 are illustrated in Fig. 1. Probes specific for rgpB and ragB were constructed by restriction enzyme digestion of clones containing these genes that were obtained from M. Curtis. The tla-specific probe was constructed by PCR amplification of a 1-kb fragment with primers complementary to the TonB-linked portion of the gene (forward, 5′ GTTGTAGAAGCAGGAATCGG 3′ [positions 626 to 645]; reverse, 5′ CGTTGCTACAGTATAGTCGC 3′ [positions 1658 to 1639]). The recA-specific probe was an amplicon constructed with primers complementary to 5′ GCCGATCAGATACTAAACGG 3′ (forward; positions 730 to 749) and 5′ GATAATATCCAGCTCTACACC 3′ (reverse; positions 1649 to 1629) of the recA gene.

DNA for pulse-field gel electrophoresis by contour-clamped homogeneous electric field (CHEF) technology was prepared as previously described (20). Agarose blocks were incubated for 2 h in 0.5× Tris-base-EDTA buffer, followed by 4 h of incubation in 1× restriction enzyme buffer. Half of each agarose plug was used for each reaction. The digestion was carried out in a 100-μl reaction mixture containing restriction enzymes. The following amounts of enzymes were used alone or in combination as needed: 50 U of SfiI, 20 U of AvrII, 100 U of XbaI, 15 U of SpeI, and 25 U of NheI. The digestion was carried out overnight at 37°C (except SfiI digestion, which was carried out at 50°C). This overnight digestion was followed by a second 2 h of digestion with half of the amount of restriction enzyme that had been added for overnight digestion. A pulse-field gel electrophoresis instrument with hexagonal electrodes, the Chef Mapper XA system (Bio-Rad Laboratories, Richmond, Calif.), was used to separate large DNA fragments. Tris-base-EDTA buffer (0.5×) was used. Ultimately, electrophoresis was performed with one fourth of the original agarose block in a 1% pulse-field certified agarose gel at 14°C. λ DNA ladders (λ Ladder PFG Marker and MidRange PFG Marker II; New England Biolabs, Beverly, Mass.) were used as DNA size markers. Following electrophoresis, the gel was stained with ethidium bromide in electrophoresis buffer for 15 min and destained for 30 min.

The genomic library of P. gingivalis W83 used in this work was constructed by H. Fletcher (9).

For Southern blotting, electrophoretically separated DNA was transferred to positively charged nylon membranes (Boehringer Mannheim Corp., Indianapolis, Ind.), and hybridization was performed with a probe labeled by nick translation with [α-³²P]dCTP (kit from Promega Corp., Madison, Wis.) or with positively charged peroxidase (ECL kit; Amersham Corp., Arlington Heights, Ill). Autoradiography was done with intensifying screens on X-Omat LS films (Eastman Kodak Company, Rochester, N.Y.).

Estimation of chromosome size.

Previously, we estimated that the genome size of P. gingivalis W83 was 2.2 Mb, based on the summation of AvrII restriction fragments (20). The DNA base composition of P. gingivalis ranges from 46 to 48 mol% G+C (33). Based on this information, we surveyed appropriate rare-cutting restriction enzymes in order to identify others that yielded quantitatively desirable DNA fragments. Several appropriate restriction enzymes were found in addition to AvrII: SfiI, NheI, XbaI, and SpeI. These data indicate that other restriction enzymes cutting N′CTAGN sequences are potential candidates for mapping purposes. Among them are SrfI and SseI. The restriction enzyme survey data have also rendered information on the size of the P. gingivalis W83 genome. Based on summations of the various fragment sizes (assessed by their migration relative to coelectrophoresed marker fragments) produced by AvrII and NheI, we estimated the size of the P. gingivalis W83 genome to be approximately 2.2 Mb (Fig. 2; Table 1). Similarly, we have determined the genome sizes of P. gingivalis W12 (2.2 Mb) and W50 (2.2 Mb) (Fig. 2A and B; Table 1). Restriction enzymes SfiI, XbaI, and SpeI rendered multiple fragments that might exist as doublets and were indistinguishable from each other. Thus, no attempt to approximate the genome size from these digests was made.

FIG. 2.

CHEF gel electrophoresis of AvrII-digested (A) and NheI-digested (B) DNA of P. gingivalis W83, W12, and W50. Cleaved total DNA was subjected to CHEF gel electrophoresis as described in the text. Lanes in panel A: 1, W83; 2, W12; 3, W50; 4, molecular size marker. Lanes in panel B: 1, W83; 2, W12; 3, W50; 4, molecular size marker. (C) Southern blot analysis of NheI-cleaved DNA of P. gingivalis W83 (lane 1), W12 (lane 2), and W50 (lane 3) with the HArep domain as a probe (Fig. 1). The size of the hybridizing component is given at the right.

TABLE 1.

Sizes of P. gingivalis W83, W12, and W50 DNA fragments by AvrII and NheI restriction endonucleases

Fragment no.	Fragment size (kb)
	AvrII			NheI
	W83	W12	W50	W83	W12	W50
1	320	500	340	580	500	480
2	291	320	291	340	320	340
3	225	235	250	291	291	300
4	225	225	230	250a	291	291
5	200	200	200	195	250a	250a
6	170	160	170	145	145	145
7	150	130	150	125	115	110
8	130	120	130	110	100	110
9	125	100	115	70	70	70
10	97	60	97	40	40	40
11	60	52	60	30	30	30
12	52	50	52	25	25	25
13	50	37	50
14	37	30	37
15	30		30
16	25		25
Total	2,187	2,219	2,227	2,201	2,177	2,191

^aSize of HArep-hybridizing fragment. 

Although the genome sizes of P. gingivalis W83, W12, and W50 were similar, their fragment patterns rendered by AvrII and NheI differed (Fig. 2A and B; Table 1), reflecting genotypic heterogeneity in these strains.

Location of rgpA, kgp, and hagA.

Southern blot analysis of NheI-digested chromosomal DNA of P. gingivalis W83 with the HArep consensus of the hemagglutinin domain found in hagA, rgpA, and kgp (probe b [Fig. 1]) as a probe revealed one hybridizing fragment 0.25 Mb in size (Fig. 2C). The NheI restriction fragment patterns for strains W83, W12, and W50 were different (Fig. 2B, lanes 1 to 3), but all of the strains contained a 0.25-Mb fragment that hybridized with the probe. This suggests the conservation of a large region within the genomes of these strains. Differences among these three strains were also noted in the AvrII digests. These results are consistent with studies from other laboratories that have established that clonal diversity is characteristic of P. gingivalis isolates (21, 22).

We constructed a restriction map of the 0.25-Mb NheI fragment and determined the locations of rgpA, kgp, and hagA. We screened a lambda genomic library of P. gingivalis W83 and found three clones containing CRSs: lambda 10 (hagA), lambda 7 (kgp), and lambda 4 (rgpA). These clones were analyzed for the presence of AvrII, NheI, XbaI, SfiI, and SpeI restriction sites. The analysis revealed that the lambda 4 clone, carrying the rgpA gene, contained an SfiI site approximately 10 kb upstream from the gene (Fig. 3). This information also allowed us to determine the size of the SfiI fragment located upstream from the SfiI fragment carrying rgpA. The Southern blot analysis of SfiI-cleaved chromosomal DNA with the 3.5-kb fragment of DNA adjacent to the SfiI fragment carrying rgpA revealed a hybridizing band of approximately 85 kb (Fig. 3). In addition, the probe hybridized to 45 kb of SfiI-NheI-digested DNA, indicating that the NheI restriction site was present approximately 55 kb upstream from the rgpA gene (Fig. 3). The lambda 10 clone, containing the hagA gene, also carried a 3.5-kb 5′ region of kgp gene sequence (Fig. 3). By PCR and DNA sequencing analyses we determined that these genes were approximately 3 kb apart (results not shown). Furthermore, the lambda 10 clone carried an NheI restriction site at the beginning of the hagA gene, indicating that the gene is located at the end of the 0.25-Mb NheI fragment (Fig. 3). By combining the results of the lambda 4 and lambda 10 analyses, we were able to position rgpA, kgp, and hagA on the 0.25-Mb NheI fragment (Fig. 3).

FIG. 3.

Map of the 0.25-Mb NheI genomic fragment. (Top left) SfiI restriction map of lambda 4. (Top right) Mapping of kgp and hagA on the 0.25-Mb NheI fragment. Sequences of both hagA and kgp are present on the lambda 10 clone. They are separated by a 3-kb fragment and located in the vicinity of the NheI site. The orientation of these genes is shown by arrows. (Bottom) Partial restriction enzyme map of the 0.25-Mb NheI fragment. rgpA, kgp, and hagA were localized on the 0.25-Mb fragment (black boxes). The orientation of these genes is depicted by arrows above the genes. Parentheses indicate approximate locations of restriction sites or the IS195 element. The brackets indicate the presence of the insertion element in the V2543 variant of P. gingivalis W83. The numbers below the map refer to the sizes, in kilobases.

Location of other genes.

Additional Southern blot analyses with single and double digests of chromosomal DNA revealed the presence of two copies of the insertion-like sequence IS195 near the termini of the 0.25-Mb NheI fragment, allowing us to further refine our map (Fig. 3). Furthermore, we analyzed the NheI-digested genomic DNA of P. gingivalis W83 for the presence of other genes: ragB, encoding a 55-kDa immunodominant antigen (13); rgpB, encoding Arg-X-specific protease (23); recA, encoding RecA protein (8); and tla, encoding TonB-linked protein (3). However, none of these genes hybridized to the 0.25-Mb fragment.

Discussion.

The physical mapping of the rgpA, kgp, and hagA genes revealed an interesting genetic architecture. First, all of these genes are considered to be involved in the virulence of soft tissue infections based on direct analysis of mutants in animal infection models (10, 20) or on inferences derived from in vitro analyses (6, 19, 24) or from antibody effects on microbial behavior in vivo (5). Second, all three of these genes occur on a conserved 0.25-Mb NheI genomic fragment that represents about 10% of the genome size of P. gingivalis W83. It is possible that this fragment differs qualitatively in the strains we examined, though the discovery of any differences depends on more complete mapping and DNA sequence information from multiple stains. Even so, the structural paradigm that our map presents will be useful in extrapolating genetic organizational relationships among isolates of P. gingivalis. The genomic nucleotide sequence of strain W83 is presently being determined (6a). However, genomic extrapolations may prove difficult given the genetic heterogeneity of clinical isolates of P. gingivalis (21, 22, 38). The virulence gene-bearing fragment characterized in our work may provide a useful landmark for rapid genetic comparison of clinical isolates and for further genetic exploration of virulence determinants. Thus, we believe mapping studies such as those reported here will complement genomic sequence analyses specifically related to the investigation of genetic determinants involved in colonization and virulence. Finally, the occurrence of IS-like elements near the termini of the 0.25-Mb NheI fragment suggests fragment mobility and the possibility that the basis of this region is a pathogenicity island (11). These regions of virulence genes are found in virulent isolates but not in avirulent variants of the same bacterial species. Considerable work is needed to establish that this fragment is part of a pathogenicity island in P. gingivalis, but its conservation in multiple strains, its carriage of multiple virulence genes, and the association of IS elements with its termini provide the basis for formulating a hypothesis.