Enzymatic Properties of Dipeptidyl Aminopeptidase IV Produced by the Periodontal Pathogen Porphyromonas gingivalis and Its Participation in Virulence

The bacterial strains used in this study are described in Table 1.P. gingivalis was grown anaerobically (80% N₂, 10% CO₂, 10% H₂) in brain heart infusion (BHI) broth (Difco, Detroit, Mich.) or on BHI agar plates supplemented with 5 μg of hemin (Sigma-Aldrich, St. Louis, Mo.) per ml and 0.5 μg of menadione (Sigma-Aldrich) per ml. For animal experiments,P. gingivalis was grown as previously described (12). Escherichia coli was grown in Luria-Bertani (LB) broth or on LB agar plates. SY327λpir(24), λpir lysogen of SY327, was the host for transformation with plasmids containing the R6K replicon, pGP704 (24), and its derivatives. SM10λpir(24) was used for conjugal transfer of pGP704 derivatives. BL21 (36) was the host for pTD-T7 (8) and its derivatives, and DH1 (36) was the host for transformation with other plasmids. Antibiotics were used at the following concentrations for E. coli: ampicillin sodium salt (Sigma-Aldrich), 50 μg/ml; kanamycin sulfate (Sigma-Aldrich), 50 μg/ml; erythromycin (Sigma-Aldrich), 300 μg/ml; and gentamicin sulfate (Sigma-Aldrich), 100 μg/ml. For P. gingivalis, 10 μg of erythromycin per ml was added as required.

DPPIV activity was measured with 0.5 mM glycylprolylp-nitroanilide (Gly-Pro-pNA) (Peptide Institute, Inc., Osaka, Japan) as a substrate in 20 mM potassium phosphate buffer (pH 7.5) at 25°C, and released p-nitroaniline was spectrophotometrically monitored at 405 nm. P. gingivalisW83 was harvested at the stationary phase. Cells were washed with 10 mM potassium phosphate–1 mM EDTA buffer (pH 7.5), resuspended in the same buffer, frozen-thawed, sonicated with an Ultrasonic Generator (Nihonseiki, Tokyo, Japan), and centrifuged at 70,000 × g for 1 h. The resulting supernatant was fractionated on a DEAE-Sephacel column (Amersham Pharmacia, Little Chalfont, United Kingdom) that had been equilibrated with 10 mM potassium phosphate–1 mM EDTA buffer (pH 7.5). Retained proteins were eluted with a linear gradient of 50 to 120 mM NaCl in 10 mM potassium phosphate–1 mM EDTA buffer (pH 7.5). Pooled active fractions were concentrated with an Ultrafree C3-LGC centrifugal filter unit (Millipore, Bedford, Mass.) and applied to a Sephacryl S-300 HR column (Amersham Pharmacia) that had been equilibrated with 20 mM potassium phosphate–100 mM NaCl–1 mM EDTA buffer (pH 7.5), followed by elution with the same buffer. Ammonium sulfate was added to the pooled active fractions at a final concentration of 1.5 M (35% saturation), and the fractions were then applied to a butyl-Sepharose column (Amersham Pharmacia) that had been equilibrated with 10 mM potassium phosphate–1 mM EDTA–1.5 M ammonium sulfate buffer (pH 7.5), followed by elution with a linear gradient of 1.5 to 0 M ammonium sulfate in 10 mM potassium phosphate–1 mM EDTA buffer (pH 7.5). Combined active fractions were finally applied to a hydroxyapatite column (Bio-Rad, Hercules, Calif.) that had been equilibrated with 150 mM potassium phosphate buffer (pH 7.5), and retained proteins were eluted with a linear gradient of 150 to 300 mM phosphate buffer (pH 7.5). The purity of proteins was confirmed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) (20). The protein concentration was determined with bovine serum albumin as a standard by use of the BCA (bicinchoninic acid) Protein Assay Kit (Pierce, Rockford, Ill.).

DNA was extracted from agarose gels with the Geneclean II Kit (Bio 101, Inc., Vista, Calif.). Transformation ofE. coli with plasmids was performed by electroporation with a Gene Pulser II (Bio-Rad). Preparation of DNA probes, colony hybridization, and Southern hybridization were carried out with the DIG (digoxigenin) DNA Labeling and Detection Kit (Boehringer GmbH, Mannheim, Germany). A series of deletion mutants for DNA sequencing were prepared with the Kilo-Sequence Deletion Kit (TaKaRa, Kusatsu, Japan). Sequencing was performed with the AutoRead Sequencing Kit (Amersham Pharmacia), followed by analysis with an A.L.F.DNA Sequencer (Amersham Pharmacia). Other methods were as described previously (36).

When we began and completed sequencing the gene coding for DPPIV, no sequence information was available for DPPIV from any strains ofP. gingivalis, including 381 (18) and W83. Therefore, we used amino acid sequences of the purified enzyme to obtain information for cloning. The N-terminal amino acid sequence could not be determined after several attempts. Instead, internal amino acid sequences were determined as follows. The purified enzyme was digested with lysyl endopeptidase (Wako, Osako, Japan) at a 200:1 (mol/mol) ratio at 25°C for 20 h, and the resulting peptides were isolated by reverse-phase high-pressure liquid chromatography (HPLC) (ODS-120T; Tosoh, Tokyo, Japan) with a linear gradient of 0 to 80% acetonitrile in 0.1% trifluoroacetic acid at a flow rate of 0.7 ml/min. The sequences of the nine fragments were automatically determined with a PPSQ-10 gas-phase peptide sequencer system (Shimadzu, Kyoto, Japan). Since two of the nine peptides exhibited high homology with other eukaryotic DPPIVs in a search of the GenBank/EMBL/DDBJ database, primers for PCR were designed based on the sequences of the peptide fragments as follows: 5′-CCGGATCCGA (C/T)GG(A/C/G/T)(A/C/T)G(A/C/G/T)ATGGT(A/C/G/T)GC-3′, containing a BamHI site (underlined), and 5′-GGCTGCAGTC(G/A)TA(G/A)AA(A/C/G/T)C(T/G)CCA(G/A)TC(A/C/G/T)GC-3′, containing a PstI site (underlined). PCR was performed withP. gingivalis W83 chromosomal DNA as the template. The resulting 1.5-kbp PCR fragment was used as a probe to screen a plasmid library made by partially digesting P. gingivalischromosomal DNA with MboI and ligating the DNA to pUC119 digested with BamHI by colony hybridization. Three overlapping DNA fragments derived from a single gene, as demonstrated by restriction endonuclease analysis, were isolated. Southern blotting revealed that one copy of the gene exists on the chromosome (data not shown). We combined the sequence data for the three clones and obtained a complete sequence.

A DNA fragment containing the dpp gene was generated by PCR with W83 chromosomal DNA as the template and with the following primers: ML1, 5′-GGGTCGACCATCGTAACCATGTGTGCC-3′, with aSalI site (underlined) beginning at base 33 from the translation initiation codon, and OPR3, 5′-CCGCATGCCTGTATTAAAGATTGTCG-3′, with anSphI site (underlined) ending 5 bp downstream from the stop codon. The resulting PCR fragment was digested with SalI andSphI and ligated to SalI- andSphI-digested pTD-T7 (8) to yield pYKP403. To construct the plasmid for overproducing wild-type DPPIV, pYKP406, “loop-out” mutagenesis was performed by the Kunkel method (19) with primer DP4MQ-2 (5′-TTTGTTCCCCGTCTGCATAGCTGTTTCCTG-3′) according to the instruction manual for Mutan-K (TaKaRa). BL21 was transformed with pYKP406 to create strain DPPRWT. pYKP407 for producing mutant DPPIV protein S593A, pYKP408 for producing D668A, and pYKP409 for producing H700A were created by the Kunkel method (19) with primers DPP4S (5′-GCCGCCATAGGCCCACCCCCA-3′), DPP4D (5′-AACATTGTCGGCTGCCGATCC-3′), and DPP4H-2 (5′-CCCGTATATACTAGCGTTCTTGTCCAT-3′), respectively. BL21 was transformed with these plasmids to produce strains DPPRSA, DPPRDA, and DPPRHA, respectively (Table 1).

A 1/20 portion of overnight cultures of E. coli overproducers was added to LB broth without glucose and incubated aerobically at 30°C for 3 h. Isopropyl-β-d-thiogalactopyranoside (IPTG) was added to each culture at a final concentration of 0.1 mM, and the cultures were incubated aerobically at 30°C for 4 h. Cells were harvested, resuspended in 10 mM Tris-HCl buffer (pH 7.5), frozen-thawed, sonicated with an Ultrasonic Generator, and centrifuged at 15,000 × g for 30 min. The resulting supernatant was diluted threefold and subjected to chromatography on a hydroxyapatite column equilibrated with 0.2 M Tris-HCl buffer (pH 7.5). After the column was washed with the same buffer, retained proteins were eluted with a linear gradient of 0.2 to 0.6 M Tris-HCl buffer (pH 7.5). The protein concentration and kinetic constants of the purified enzyme were determined.

Antiserum against P. gingivalis DPPIV was prepared by immunizing a rabbit with the wild-type DPPIV sample purified from strain DPPRWT as an antigen according to a previously described method (36). Western blotting was performed as described previously (36).

Binding of P. gingivalisDPPIV or human DPPIV to ADA was investigated by use of modifications of the procedure described by Iwaki-Egawa et al. (16). Briefly, samples of P. gingivalis DPPIV or human DPPIV (1 μg) were incubated with calf intestine mucosa ADA (1 μg) (Sigma-Aldrich) for 30 min at 37°C in 10 mM Tris-HCl buffer. For electrophoresis under nondenaturing conditions, human DPPIV reaction mixtures were loaded onto a 3 to 8% Tris-acetate gel (NOVEX, San Diego, Calif.) andP. gingivalis DPPIV reaction mixtures were loaded onto a 50 mM Tris-HCl nondenaturing gel (pH 8.8) or onto an isoelectric focusing gel (pH 3 to 10) (NOVEX). Proteins on gels were detected by Coomassie brilliant blue (CBB) staining and by Western blotting with anti-ADA antibody and anti-P. gingivalis DPPIV antibody. ADA affinity column chromatography was performed as previously described (10), except for the detergents.

dpp null mutants were constructed as depicted in Fig.1. The EcoRI fragment of pUC4K (42) containing the kanamycin resistance gene was inserted into the EcoRI site of pVA2198 (12) to yield pYKP001. The PstI fragment containing theermF-ermAM cassette, which confers erythromycin resistance (Em^r) to both P. gingivalis and E. coli (12), was ligated to the PstI site of suicide vector pGP704 (24). The resulting plasmid, pYKP002, was digested with KpnI to eliminate one of the twoSphI sites and self-ligated to generate pYKP009. pYKP002 and pYKP009 possess several cloning sites, including SphI andSacI, and are suicide vectors transferable to P. gingivalis and selectable for the Em^r phenotype when inserted into P. gingivalis chromosomal DNA. A derivative of pYKP009, pYKP200, which possesses a fused fragment ALAR of the flanking regions with the reading frame of dpp deleted (Fig. 1A), was created to construct null mutants.

The fragment containing the flanking regions was generated by PCR withP. gingivalis W83 chromosomal DNA as the template. The AL fragment (463 bp), located upstream of dpp (positions −511 to −49 from the translation initiation codon), was generated with primer AL1, 5′-GGGCATGCTATCCACAACTATGAAAGAG-3′, with an SphI site (underlined), and primer AL2, 5′-CCCAGCTGCAATGCACGATCCATCTCTC-3′, with aPvuII site (underlined). The AR fragment (422 bp), located downstream of the dpp gene, including 135 bp of the 3′ coding region, was amplified with primer AR1, 5′-GGCAGCTGACAGAGGCACTGGTTCAGGC-3′, containing aPvuII site (underlined), and primer AR2, 5′-CCGAGCTCATCACCACCGATCATGAAGG-3′, containing aSacI site (underlined). The AL and AR fragments were digested with SphI/PvuII andSacI/PvuII, respectively. PCR was performed with a mixture of both the AL and the AR fragments as templates and with primers AL1 and AR2. The resulting ALAR fragment was digested withSphI and SacI and ligated with SphI- and SacI-digested pYKP009 to yield pYKP200 (Fig. 1A).

SM10λpir was transformed with pYKP200, and the plasmid was then transferred to P. gingivalis W83 by selection with erythromycin and counterselection with gentamicin as described elsewhere (28), with some modifications. Em^rtransconjugants of P. gingivalis W83 thus formed were purified and grown successively in BHI broth to obtain erythromycin-sensitive (Em^s) colonies, which were then purified and examined by PCR, Southern hybridization, and Western blotting to confirm that the fragment consisting of the vector and thedpp gene had been lost. Two null mutants, 4351 and 4361, were thus isolated.

The dpp gene insertion mutant, EM23, was constructed as shown in Fig. 2. The EcoRV fragment containing the dpp gene was cloned from W83 chromosomal DNA by colony hybridization with pBR322 as the vector to create pYKP300. The PstI fragment of pYKP001 mentioned above (Fig. 1A) and containing the ermF-ermAM cassette was ligated with pYKP300 digested with EcoT22I, resulting in pYKP301. pYKP301 was linearized by EcoRV digestion, and theEcoRV fragment containing both the dpp gene and the ermF-ermAM cassette was introduced into P. gingivalis W83 by electroporation as previously described (12). Erythromycin was used as a selection marker instead of clindamycin. Except for spontaneous mutants, Em^r colonies should have arisen only when the ermF-ermAM cassette was inserted into the dpp gene on the chromosome as a result of a double-crossover event, since the introduced DNA cannot replicate in W83 (Fig. 2B). Em^r transformants thus obtained were purified and confirmed to have no DPPIV activity. One of the mutants was designated EM23.

W83 and DPPIV-deficient mutants were grown until the early stationary phase. The growth phase was judged by measuring the optical densities of cultures. Cells were harvested and washed with phosphate-buffered saline, and W83 and the mutants were adjusted to the same cell concentrations in phosphate-buffered saline. Numbers of viable cells were counted by spreading the cultures on BHI agar plates, and we confirmed the absence of contamination. BALB/c mice (11 weeks) were challenged with a dorsal subcutaneous injection of 0.2 ml of bacterial suspension. General health, weight, and presence and location of lesions were assessed daily. Statistical significance of differences in lesion formation and lethality caused by injection of W83 and mutant strains was examined by Fisher's exact probability test.

The DDBJ accession no. assigned to the dpp gene of P. gingivalis W83 isAB008194 .

More than 90% of the DPPIV activity was found in the supernatant (hereafter referred to as the soluble fraction) after sonication and centrifugation. No DPPIV activity was found in the culture fluids. The enzyme was purified to about 1,000-fold from the soluble fraction and reached homogeneity (Fig.3A). Nineteen micrograms of the purified enzyme was obtained from 1.5 liters of bacterial culture.

The purified enzyme exhibited a 78-kDa single band in SDS-PAGE. When the sample was chromatographed on an HPLC gel filtration column (TSK gel G3000SWXL; Tosoh) under nondenaturing conditions, the retention time was at a position corresponding to a molecular mass of 140 kDa (data not shown). The K_m andV_max values were estimated to be 0.28 mM and 18 μmol/min/mg, respectively, with Gly-Pro-pNA as a substrate. The range of optimum pH of the enzyme was 7.5 to 8.5, with the highest activity occurring at pH 8.0. After incubation at 50°C for 30 min, 70% of the enzyme activity remained. DPPIV activity was not affected by 5 mM dithiothreitol.

By the methods described in Materials and Methods, we isolated and sequenced the gene coding for DPPIV, designated dpp, which was composed of a coding region of 2,169 nucleotides. The open reading frame encodes a protein of 723 amino acids with a calculated molecular mass of 81.8 kDa. This value is in good agreement with that estimated by SDS-PAGE (78 kDa). The deduced amino acid sequence contains sequences of the nine peptides derived from purified DPPIV and mentioned in Materials and Methods.

The amino acid sequence exhibits high homology with those of other DPPIVs—about 30% identical to human DPPIV (7, 25, 38) and mouse DPPIV (22) and 43% identical to C. meningosepticum DPPIV (17). Serine proteases and serine esterases share a Gly-X-Ser-X-Gly consensus motif found around the active serine (5). This motif occurs in the C-terminal catalytic domain of other DPPIVs (21) with the consensus sequence Gly-Trp-Ser-Tyr-Gly. Three residues, Ser624, Asp702, and His734, are thought to form a catalytic triad in mouse DPPIV (9). The consensus motif around the active Ser and the three amino acids corresponding to the catalytic triad are found in P. gingivalis DPPIV from alignment data. The putative catalytic triad in P. gingivalis DPPIV is Ser593, Asp668, and His700. On the other hand, no significant homology with other bacterial peptidases, such as Lactococcus (23, 30) andLactobacillus (41) X-Pro dipeptidyl aminopeptidases, was found, except for a few residues around the active Ser and the residues of the catalytic triad, even though the enzymatic activities of the peptidases are the same as those of DPPIV.

The amino acid sequence deduced from the dpp gene contains a typical signal sequence in its N terminus. The signal sequence consists of three domains, N, H, and C (34). Lys and Arg residues are found in the N domain, the H domain is composed predominantly of hydrophobic residues, and one Gly residue is found in the H domain. When the (−3, −1) rule of von Heijne (44) is applied to the N-terminal sequence of P. gingivalis DPPIV, the enzyme is predicted to be cleaved between residue 19 (Ala) and residue 20 (Gln) upon secretion, since residue 19 Ala (−1) and residue 17 Ala (−3) are in good agreement with the (−3, −1) rule. We constructed an E. coli strain expressing P. gingivalis DPPIV protein whose second amino acid after the initiation Met is Gln (residue 20 of the enzyme). The dpp gene was located between the codon for the initiation Met and the transcription termination signal of plasmid pTD-T7 (8), so that the gene was under the control of the T7 promoter, resulting in pYKP403. Loop-out mutagenesis was carried out to ligate the codons for the initiation Met and for residue 20 Gln to produce pYKP406. Strain DPPRWT, BL21 carrying pYKP406, was cultured, and the expression of DPPIV protein was induced by IPTG. High-level expression was observed by CBB staining (Fig. 3B) and Western blotting (data not shown). The recombinant protein was purified from the overproducer to near homogeneity (Fig. 3B). From 1 liter of bacterial culture, 7.6 mg of recombinant protein was obtained. The N-terminal amino acid sequence of the protein was Met followed by Gln, indicating that the initiation Met was not removed in E. coli. TheK_m and V_max values of the recombinant protein were estimated to be 0.21 mM and 23 μmol/min/mg, respectively. The values were almost equal to those of DPPIV purified from P. gingivalis, indicating that DPPIV isolated fromP. gingivalis and recombinant DPPIV have almost the same enzymatic properties.

It was reported that human DPPIV cleaved human CC chemokines (β-chemokines) at the C terminus of the Pro residue on the penultimate position of the N terminus (3, 40). Cleavage of two synthetic peptides, Ser-Pro-Tyr-Ser-Ser-Asp-Thr-Thr (corresponding to the N-terminal amino acids of human RANTES) and Gln-Pro-Asp-Ala-Ile-Asn-Ala-Pro (corresponding to the N-terminal amino acids of human MCP1), by recombinant DPPIV was examined. The recombinant enzyme cleaved the peptides in the same fashion as human DPPIV did, as demonstrated by reverse-phase HPLC with a TSK gel ODS-80Ts column (Tosoh) (data not shown).

As reported previously (16), the present study reconfirmed the ability of human DPPIV to bind to ADA. In contrast, we did not detect the binding of P. gingivalis DPPIV to ADA. After incubation of recombinant DPPIV and ADA, the reaction mixture was loaded onto a 50 mM Tris-HCl nondenaturing gel (pH 8.8). Under these conditions, protein bands of DPPIV and ADA were detected not only by CBB staining but also by Western blotting with anti-P. gingivalis DPPIV antibody and anti-ADA antibody, respectively. However, the band detected by anti-P. gingivalis DPPIV antibody was not stained by anti-ADA antibody. The same result was obtained when the reaction mixture was loaded onto an isoelectric focusing gel (pH 3 to 10). DPPIV purified from wild-type P. gingivalis W83 showed the same results (data not shown). When DPPIV from P. gingivalis was tested on an ADA affinity column, binding of the enzyme to the column was not observed.

When P. gingivalis DPPIV was aligned with other DPPIVs, Ser593, Asp668, and His700 were suggested to participate in DPPIV activity. To explore the involvement of these three amino acid residues in catalytic function, we performed in vitro site-directed mutagenesis of P. gingivalis DPPIV. Each of the residues was substituted with Ala to create S593A, D668A, and H700A mutant proteins. The levels of expression of the mutant proteins in strains DPPRSA, DPPRDA, and DPPRHA were almost the same as that of the wild-type protein in strain DPPRWT. The specific activity of wild-type DPPIV was estimated to be 15 μmol/min/mg, while those of the mutant proteins were all less than 2 nmol/min/mg, with 0.4 mM Gly-Pro-pNA as a substrate. Reduction of DPPIV activity of the mutant proteins to less than 1/7,500 wild-type activity suggests that the three amino acid residues are directly involved in the catalytic reaction and that Ser593 is likely to be the catalytic Ser.

To investigate the importance of DPPIV in the virulence of P. gingivalis, we constructed DPPIV-deficient mutants of P. gingivalis W83 as described in Materials and Methods. Two null mutants, 4351 and 4361, were isolated. DPPIV activity was detected in neither of them. The chromosomal DNAs of the two strains were analyzed by PCR with the primer sets primer AL1-primer AL2 and primer AL1-primer AR2 (Fig. 1A). A 463-bp fragment was amplified from both W83 and 4351 chromosomal DNAs with the primer AL1-primer AL2 set. A 2,970-bp fragment and an 885-bp fragment were developed from W83 and 4351 chromosomal DNAs, respectively, with the primer AL1-primer AR2 set.

For Southern hybridization analysis, chromosomal DNAs of W83 and 4351 were digested with PvuII or KpnI. With the AR fragment (Fig. 1A) as a probe, 2.7-kbp PvuII and 6.6-kbpKpnI fragments from W83 DNA hybridized, while 2.4-kbpPvuII and 4.5-kbp KpnI fragments from 4351 DNA hybridized (Fig. 1B). For Western blotting, crude extracts prepared from W83 and 4351 were loaded onto an SDS-polyacrylamide gel, and proteins on the gel were transferred to a membrane. The anti-P. gingivalis DPPIV antibody stained a 78-kDa band from W83 but no band from 4351. Strain 4361 showed the same results in PCR, Southern blotting, and Western blotting as strain 4351. Taken together, these results indicate that 4351 and 4361 are null mutants with most of thedpp gene deleted. Mutant 4351 was chosen for further analysis.

One of the insertion mutants derived from W83, EM23 (Fig. 2), was further analyzed by Southern blotting. Chromosomal DNAs of W83 and EM23 were digested with KpnI or HindIII. With theermF-ermAM fragment (Fig. 2) as a probe, no hybridized band was seen with W83 DNA; on the contrary, a 7.2-kbp KpnI fragment and 7.6- and 2.5-kbp HindIII fragments from EM23 DNA hybridized (Fig. 2B). With the ALAR fragment (Fig. 1) as a probe, a 6.6-kbp KpnI fragment and an 8.0-kbpHindIII fragment from W83 DNA hybridized, while 7.2- and 1.5-kbp KpnI fragments and 7.6- and 2.5-kbpHindIII fragments from EM23 DNA hybridized (Fig. 2B).

The growth rates of the mutants, 4351, 4361, and EM23, were similar to that of parental strain W83 in BHI broth, indicating that DPPIV is not essential for bacterial growth in rich medium.

Total cell numbers, as indicated by optical densities of both wild-type W83 and the mutant strains (4351 and EM23), were proportional to viable cell numbers at both the logarithmic and the early stationary phases. In contrast, after the cells reached the stationary phase, the numbers were no longer proportional, presumably because some cells were not viable at this phase. Therefore, we harvested bacterial cells when they were determined by optical density measurements to have reached the early stationary phase. W83 or mutant (4351 or EM23) cells were injected subcutaneously into the dorsal surface of mice. There was no significant decrease in viable cell numbers due to exposure to the atmosphere during the period of injection. Mice displayed ruffled hair within 24 h and began to die 3 days after injection. The numbers of animals dead 6 days after injection of 7 × 10⁹ to 9 × 10⁹ bacterial cells were 13 of 17 with W83, 5 of 13 with 4351, and 1 of 4 with EM23 (Fig.4 and Table2). The difference in lethality between W83 and the mutant strains was significant, as determined by Fisher's probability test (P <0.05). There was no significant difference in lethality between the two mutants. All animals that had survived the first 6 days after injection remained alive during the entire observation period of 3 weeks.

Mice challenged with W83 or mutant strains displayed the following macroscopic findings. (i) Mice exhibited no severe lesions at the sites of injection on the dorsal surfaces, except for redness. (ii) Mice developed abscesses on the abdomens 2 days after injection. Within 1 day of abscess appearance, the abscesses ruptured, and drainage and hemorrhage occurred, resulting in ulcer formation. Sixteen of the 17 W83-challenged animals displayed widespread abscesses or ulcerative lesions covering the abdomens. Among the mutant-challenged animals, 6 of 13 injected with 4351 and 4 of 4 injected with EM23 developed similar extensive lesions on the abdomens. The frequencies of extensive lesions caused by W83 and by mutant strains were significantly different (P < 0.05) (Table 2). (iii) Seven mice injected with 4351 survived throughout the observation period with only localized lesions or with no apparent pathological changes (Table 2). In contrast, one mouse injected with W83 exhibited no abscess on the abdomen but developed severe expansion of the abdomen (Table 2), leading to death in 3 days. Autopsy of this animal revealed that an intestinal protrusion had penetrated through a hole in the peritoneum to reach the subcutaneous region. (iv) Two W83-injected mice and one 4351-challenged animal developed secondary extensive lesions in the thoracic regions within 4 days. (v) Finally, among the surviving animals, one mouse challenged with W83 and three mice challenged with 4351, small ulcerative lesions scattered over several places on the bodies, such as the axillary regions and/or articulations, were found in addition to abscesses on the abdomens.

An increase in body weight was observed in 4351-injected mice at 6 days after injection, while it took longer (more than 2 weeks) for the body weight of W83-injected mice to increase in experiment 2 (Table 2 and Fig. 5). Similar results were obtained in other experiments shown in Table 2. In concurrence with weight gain, surviving mice showed scarring of ulcerative lesions. Lesions caused by W83 healed much more slowly than those caused by the mutants. The findings indicated that mice challenged with the mutants exhibited faster recovery from the infection than those injected with W83.