Silkworms are killed by injection of pathogenic bacteria, such as Staphylococcus aureus and Streptococcus pyogenes, into the haemolymph. Gene disruption mutants of S. aureus whose open reading frames were previously uncharacterized and that are conserved among bacteria were examined for their virulence in silkworms. Of these 100 genes, three genes named cvfA, cvfB, and cvfC were required for full virulence of S. aureus in silkworms. Haemolysin production was decreased in these mutants. The cvfA and cvfC mutants also had attenuated virulence in mice. S. pyogenes cvfA-disrupted mutants produced less exotoxin and had attenuated virulence in both silkworms and mice. These results indicate that the silkworm-infection model is useful for identifying bacterial virulence genes.
Pathogenic bacterial infection of hosts consists of the following stages; attachment to and colonization of host surfaces, invasion of the epithelial cell layer, and systemic dissemination of the microbes. Animal models are important for the identification and evaluation of bacterial genes that are required for virulence and to study infectious stages at a molecular level. Invertebrate animals such as Caenorhabditis elegans (C. elegans) and Drosophila melanogaster have been used for this purpose for both economic and ethical considerations and have contributed to our understanding of the molecular basis of host–pathogen interactions (Aballay and Ausubel, 2002; Ewbank, 2002; Hoffmann, 2003).
The silkworm (Bombyx mori) is an easily bred invertebrate animal used for basic studies because of its importance in sericulture. We reported a silkworm-infection model, in which silkworms are killed by injection of microorganisms that are virulent in humans, such as Staphylococcus aureus (S. aureus), Pseudomonas aeruginosa, Vibrio cholerae, Stenotrophomonas maltophilia, pathogenic strains of Escherichia coli, or true fungi Candida albicans, but not by non-pathogenic strains of E. coli or Saccharomyces cerevisiae. The lethality of the infected silkworms requires bacterial growth, as antibiotics have therapeutic effects against the infection (Kaito et al., 2002; Hamamoto et al., 2004). The silkworm is large enough to inject sample solutions into the haemolymph, which is difficult in C. elegans and Drosophila larvae. Therefore, it might be possible to quantitatively evaluate the virulence of various bacteria or true fungi and to identify their virulence genes in the silkworm infection model.
In the present study, 3 of 100 previously uncharacterized genes that are conserved among pathogenic bacteria were required for virulence of S. aureus in the silkworm pathogenic bacteria infection model. Two of these three genes were also required for virulence in mice. In addition, the gene named cvfA was required for virulence of Streptococcus pyogenes (S. pyogenes) in both silkworms and mice. These results suggest the silkworm-pathogenic bacteria infection model is useful for identifying bacterial virulence genes relevant to pathogenesis in mammals.
Identification of virulence genes of S. aureus using the silkworm model
To establish the silkworm model for identification of the virulence gene, we first tested the effect of mutation in a known virulence gene, agr, which is required for the expression of variety of toxins, including haemolysin, protease, and DNase as described below. Overnight culture of the agr-disrupted mutant and its wild-type were diluted 20-fold with saline, injected into the haemolymph of fifth-instar larvae, and survival was monitored (Fig. 1A and Fig. 3). Half of the silkworm larvae injected with the parent strain CK501 died within approximately 30 h after injection, whereas 60 h was required for 50% of the silkworm larvae injected with agr-disrupted mutants to die (Fig. 3). The result suggests that the silkworm S. aureus infection model is applicable for identification of virulence genes.
While searching for novel virulence genes, we hypothesized that some virulence genes are conserved among various pathogenic bacteria. Our previous reports on the S. aureus genome project revealed that there are 589 genes whose functions are not known and that are conserved among other organisms (Kuroda et al., 2001). Therefore, we constructed gene-disrupted mutants of the S. aureus RN4220 strain for 100 of these genes (supplemental data online, Table S1) and tested their virulence in a silkworm infection model. The RN4220 strain is defective for restriction enzyme and is highly competent for transformation with plasmid DNA and thus useful for construction of an integrational deletion mutant. Overnight cultures of each mutant (approximately 2 × 109 to 4 × 109 CFU ml−1) were diluted 20-fold and 10-fold with saline, 0.05 ml of them were injected into the haemolymph of 10 fifth-instar larvae, and survival was monitored. Silkworm larvae injected with the S. aureus parent strain RN4220 died within approximately 30 h after the injection, whereas those injected with the three mutants survived for over 72 h after the injection (Fig. 1B). We named the genes disrupted in these mutants cvfA, cvfB and cvfC (conserved virulence factor A, B and C respectively). The cvfA, cvfB and cvfC genes correspond to SA1129, SA1223 and SA1262, respectively, in the S. aureus N315 genome database on GIB (the Genome Information Broker program) from the DNA Data Bank of Japan. Northern blot analysis revealed that the sizes of the cvfA, cvfB and cvfC gene transcripts were 1.9, 1.1 and 2.3 kb, respectively (data not shown). The lengths of the open reading frames (ORFs) in the cvfA, cvfB and cvfC genes were 1557, 900 and 1122 bp respectively. Genome sequence information of neighbouring genes suggested that the cvfA and cvfB genes are transcribed as monocistrons, whereas the cvfC gene forms an operon with three other genes. To further confirm their contribution to silkworm virulence, complementation experiments and phage transduction analysis using phage 80 alpha were performed. The attenuated killing ability of the cvfA mutant was complemented by a plasmid harbouring the intact S. aureus cvfA gene (Fig. 1C). The cvfA disruption mutation was cotransduced with the attenuated killing ability, supporting the notion that the cvfA mutation is responsible for the attenuated virulence in silkworms (data not shown). Although decreased virulence of cvfB and cvfC mutants was not complemented by a plasmid harbouring each gene, phage transduction experiments of the cvfB and cvfC mutations suggested that each mutation was responsible for its attenuated virulence in the silkworm (data not shown). In addition, integration of the same suicide vector into the cvfB and cvfC region without disrupting each gene did not decrease the virulence in silkworms (data not shown). The results do not support the possibility that the decreased virulence of cvfB and cvfC mutants is because of the polar effect of the integrated vector and suggests that each mutation is responsible for its attenuated virulence. The failure of complementation for the cvfB and cvfC mutants was also observed as a decreased production of haemolysin in vitro(Fig. 2A), which we assumed was because of the multicopy effects of the genes. These results suggested that the cvfA gene, the cvfB gene, and the cvfC operon of S. aureus are required for full virulence in silkworms.
The doubling times of the cvfA and cvfB deletion mutants in Luria–Bertani broth at 27°C, the temperature at which silkworms were fed, were 1.3 and 1.4 times longer than that of the parent strain RN4220 (data not shown). The cvfC mutant had a doubling time similar to the parent strain, RN4220, at this temperature. As in the case of attenuated virulence in silkworms, the slow-growth phenotype of the cvfA mutant was complemented by a plasmid harbouring the intact cvfA gene and was cotransduced with the cvfA disruption mutation in phage transduction analyses; whereas that of the cvfB mutant was not complemented, but was cotransduced with the cvfB-disruption mutation in phage transduction analyses. To determine whether the decreased growth rate of S. aureus mutants is the reason for the attenuated virulence in the silkworm, we re-examined the killing abilities of four randomly selected mutants with one disrupted gene (SA1274, SA1661, SA1662 and SA1877 in the N315 genome database), whose doubling time was 1.5 times longer than that of the parent. Each mutant had equivocal silkworm killing ability compared with the parent strain (data not shown). Therefore, the attenuated virulence of the cvfA and cvfB mutants cannot be explained by a decreased growth rate.
We next examined whether these mutants with attenuated virulence in silkworms also had attenuated virulence in mice. Mice were intraperitoneally injected with different cell numbers of the mutants and their parent strain, RN4220, and mouse survival after 3 day was determined. The dose required to kill 50% of the mice (LD50) of the cvfA or cvfC mutant was 3.5 times higher than that of RN4220 (Table 1). The results suggest that the cvfA gene and the cvfC operon contribute to S. aureus virulence in mice.
Table 1. Attenuated killing ability in mice of the cvfA, cvfB, and cvfC mutants of S. aureus and the cvfA mutant of S. pyogenes.
LD50 (cell number)
Relative attenuation (LD50 for mutant/LD50 for wild-type)
CD-1 mice (n = 5/group) were intraperitoneally injected with various doses of each strain of S. aureus and S. pyogenes, and the LD50 was determined using the method of Reed and Muench (1938).
3.4 × 108
1.2 × 109
5.4 × 108
1.2 × 109
1.3 × 108
9.3 × 108
The cvfA, cvfB and cvfC genes are required for production of haemolysin
Because certain bacteria kill C. elegans via toxin-mediated mechanisms and S. aureus secretes various extracellular toxins (Aballay and Ausubel, 2002; Ewbank, 2002), we examined whether the cvfA, cvfB and cvfC mutants produced decreased amounts of exotoxins. Overall activities of haemolysin, protease and DNase were detected by using substrates for each exotoxin. There was decreased secretion of haemolysin in the cvfA, cvfB and cvfC mutants and DNase and protease in the cvfA mutant (Fig. 2A). The in vitro phenotype of the cvfA mutant was complemented by introducing a plasmid harbouring the intact cvfA gene, although those of the cvfB or cvfC mutants were only partially complemented by a plasmid harbouring the cvfB or cvfC gene respectively (Fig. 2A). Each of the phage transductants of the cvfA, cvfB, or cvfC mutants decreased the secretion of haemolysin to a similar level as the original mutant (data not shown). Thus, the cvfA, cvfB and cvfC gene mutations were responsible for the decreased exotoxin production. Because the four slow growing mutants described above secreted equivocal amounts of haemolysin compared with the parent strain (data not shown), the decreased haemolysin production of the cvfA and cvfB mutants is not because of slow growth.
Expression of exotoxin genes in S. aureus, including genes encoding haemolysin, protease, and DNase, is positively regulated at the transcriptional level by a global virulence regulatory locus, agr, which encodes two divergent transcripts, RNAII and RNAIII (Peng et al., 1988). Production of RNAIII is activated by RNAII, and RNAIII upregulates gene expression encoding for various exotoxins and downregulates gene expression for cell wall-associated proteins including Protein A, which is encoded by the spa gene (Novick et al., 1993). We used Northern blot analysis to examine whether disruption of the cvfA gene would affect the synthesis of RNAII. Because the RNAII expression level in the RN4220 strain is extremely low (Peng et al., 1988), we transformed the RN4220 strain with the wild agr operon and used this strain (CK501) as the parent in the following study. Disruption of the cvfA gene of CK501 decreased the amount of RNAII (Fig. 2B). Furthermore, RNAIII expression was greatly reduced in both the cvfA mutant and the agr null mutant. Consistent with this notion, expression of the spa gene, which is suppressed by RNAIII as described above, was repressed in CK501, whereas the spa gene was highly expressed in both the cvfA mutant and the agr null mutant (Fig. 2B). Changes in the amount of RNAII, RNAIII and spa transcripts induced by the cvfA mutation were restored to the level of the CK501 strain by a plasmid harbouring the intact cvfA gene (Fig. 2B). These results suggest that the cvfA gene activated transcription of the agr operon, resulting in the expression of the exotoxin genes.
To evaluate the contribution of the agr inactivation to decreased virulence in the cvfA-disrupted mutant, the silkworm virulence of the agr null mutant was compared with that of the cvfA-disrupted mutant. The agr null mutant reduced killing ability in silkworms compared with the parent CK501, but was more virulent than the cvfA mutant (Fig. 3). In addition, the agr and cvfA double mutant attenuated virulence compared with the agr null mutant (Fig. 3). These quantitative assessments of the virulence in the silkworm infection model suggest that the cvfA gene is required for the full S. aureus pathogenicity, not only via transcriptional activation of the agr locus but also via other role(s) that remain to be elucidated.
The cvfA gene of S. pyogenes contributes to virulence in both silkworms and mice
Table 2 shows that homologues of the S. aureus cvfA, cvfB and cvfC genes are present in various pathogenic bacteria. Therefore, we examined whether the function of the cvfA gene was conserved in other bacterial species. A disruption mutant of the cvfA gene in a clinically isolated strain of S. pyogenes (SSI-9, M type 1) was constructed and its virulence was examined in silkworms and mice. When 9 × 108 cells of wild-type S. pyogenes were injected into silkworms, 50% of the larvae were dead 30 h after the injection, whereas more than 90 h were required for 50% of the larvae injected with the S. pyogenes cvfA mutant to die (Fig. 4). This attenuated killing ability in silkworms was complemented by a plasmid harbouring the intact S. pyogenes cvfA gene (Fig. 4). In the mouse infection experiment, the LD50 value of the wild-type SSI-9 was 1.3 × 108 CFU, whereas that of the cvfA mutant was 9.3 × 108 CFU (Table 1). Therefore, the S. pyogenes cvfA gene contributes to virulence in both silkworms and mice. Further, in vitro analysis indicated that the secretion of haemolysin, DNase, streptokinase and protease was decreased in the cvfA mutant, and were complemented by a plasmid harbouring the intact S. pyogenes cvfA gene (Table 3). The results indicate that the cvfA gene contributes to the production of exotoxins in S. pyogenes.
Table 2. Conservation of the cvfA, cvfB, and cvfC genes among bacteria.
‘Yes’ indicates the presence of homologous proteins to S. aureus CvfA, CvfB, or CvfC having more than 20% amino acid identities and E-values lower than e-10. Only the species that had at least one of the three genes are noted.
Table 3. Decrease in the amount of secreted exotoxins in the cvfA mutant of S. pyogenes.
1+ to 3+ indicates relative strength of signals on skim milk agar plate.
Activities of haemolysin, DNase, streptokinase, and protease in supernatants of overnight bacterial culture of S. pyogenes SSI-9 or M1633 harbouring either a plasmid vector (pYC60) or a plasmid containing the intact cvfA gene (pYC601). For haemolysin, DNase, and protease, overall activity was assessed without regard to a specific protein. Means ± standard deviation from three independent experiments are presented.
Requirement of KH and HD domains for S. aureus CvfA protein
To predict the molecular functions of CvfA protein, motifs in the CvfA protein were searched using the Motif search service (GenomeNet service by Kyoto University Bioinformatics Center) and the SOSUI program on the Internet. It was determined that there was a transmembrane domain at the N-terminal region, and KH and HD domains in the internal region; the latter two have RNA binding activity (Siomi et al., 1994) and metal-dependent phosphohydrolase activity (Aravind and Koonin, 1998), respectively (Fig. 5A). To examine the requirement of the KH and HD domains for S. aureus CvfA protein function, site-directed mutagenesis was performed at the amino acid residues reported to be important for the RNA binding activity of the KH domain or the phosphohydrolase activity of the HD domain (Siomi et al., 1994; Turko et al., 1999). Plasmids that harbour the point-mutated cvfA gene were transformed to the S. aureus cvfA deletion mutant and their complementation activity for the secretion of haemolysin and DNase, and killing ability against silkworms were examined. Strains harbouring the point mutated cvfA gene with G229A, I232N, or D423A had decreased levels of haemolysin and DNase compared with a strain harbouring the intact cvfA gene (Table 4). Consistent with this, these strains also exhibited reduced virulence in the silkworm compared with a strain harbouring the intact cvfA gene (Fig. 5B). These results suggest that the KH and HD domains are required for the full virulence of CvfA protein.
Table 4. Decreased complementation abilities for haemolysin and DNase production of KH or HD domain mutants of S. aureus CvfA protein.
Haemolysin (HU ml−1)
DNase (AU ml−1)
Activities of haemolysin and DNase in supernatants of overnight bacterial culture of S. aureus M1129 (cvfA-disrupted) strains that each harbour plasmids containing the intact (pC1129) or mutated cvfA gene (pG229A, pI232N, pD423A). pND50 is an empty vector. For haemolysin and DNase, overall activity was assessed without regard to a specific protein. Means ± standard deviation of three independent experiments are presented.
Novel virulence genes of S. aureus, cvfA, cvfB, and cvfC were identified using the silkworm infection model. In these three mutants, the amount of secreted haemolysins was decreased, and the cvfA and cvfC mutants also had attenuated virulence in mice. Moreover, a cvfA mutant of S. pyogenes had attenuated virulence in both silkworms and mice. Therefore, the silkworm-infection model might be useful for identifying bacterial virulence factors against mammals.
Staphylococcus aureus null mutant for the agr locus, which regulates a variety of virulence genes and is involved in mouse pathogenesis (Abdelnour et al., 1993; Heyer et al., 2002), attenuated the virulence in silkworms (Fig. 3). The results suggest that there is a common pathogenic mechanism between silkworms and mammals. In addition, a virulence system other than the agr pathway might contribute to virulence in silkworms because cvfA disruption reduced virulence in silkworms of the S. aureus agr null mutant (Fig. 3). The alternative pathway and its contribution to the pathogenesis in mammals are critical and are currently under investigation.
Functional analyses of CvfA protein in two distinct bacterial species, S. aureus and S. pyogenes, demonstrated that CvfA protein might function in a signal transduction system to induce expression of the virulence genes in these bacteria. Because CvfA protein is widely conserved among pathogenic bacteria, including Bacillus anthracis, Listeria monocytogenes, Helicobacter pylori, Treponema pallidum, etc. (Table 2), it is possible that CvfA protein is involved in an important pathway for pathogenic bacteria to exert virulence in host animals, such as sensing microenvironments and then transducing the signals for expression of the virulence genes. This is consistent with recent findings that the S. aureus agr locus, whose expression requires the cvfA gene (Fig. 2), is induced by host factors and facilitates quorum sensing-dependent virulence (Rothfork et al., 2003). To evaluate the molecular functions of CvfA protein, it will be necessary to focus on potential interactions with previously described transcriptional activators for agr and on three highly conserved domains of CvfA protein; a transmembrane domain, a KH domain and an HD domain. The KH and HD domains were suggested to be important for the function of CvfA protein (Fig. 5). The identification of target molecules of KH and HD domains might facilitate studies to examine the cvfA-dependent alternative pathway described above.
Studies of the cvfB gene demonstrate that not all the genes whose disruption has a significant effect on the virulence in silkworms has a noticeable role in mammals. The expression and/or action of the gene might be specific to the host, as reported for genes determined using C. elegans as the host (Ewbank, 2002). In contrast to the cvfA and cvfB genes, the cvfC gene is conserved only among closely related Gram-positive bacteria, and thus functional analysis of the cvfC gene might reveal the physiology that specifically associates with these bacteria.
About 80% of the genome sequence of silkworm has been determined by Japanese research groups (Mita et al., 2004), and forward and reverse genetic methods for silkworm have been established. These genetic resources enable genetic analyses of host animals in the silkworm-pathogenic bacteria infection model, and thereby this model system might contribute to basic research to reveal the dynamic interactions between bacteria and host factors.
Bacterial strains, plasmids and culture conditions
The JM109 strain of E. coli was used as a host for pMutinT3, pND50, pSF151, and their derivatives. E. coli strains transformed with the plasmids were cultured in Luria–Bertani broth containing 100 µg ml−1 ampicillin, 12.5 µg ml−1 chloramphenicol, or 50 µg ml−1 kanamycin. S. aureus strains were aerobically cultured in tryptic soy broth at 37°C unless indicated, and 10 µg ml−1 erythromycin or 12.5 µg ml−1 chloramphenicol were added to the medium if required. S. pyogenes strains were cultured in Todd-Hewitt broth supplemented with 0.4% yeast extract at 37°C in an atmosphere containing 5% CO2, and 500 µg ml−1 kanamycin or 12.5 µg ml−1 chloramphenicol was added to medium if required. Details on bacterial strains and plasmids used in this study are shown in Table 5.
Table 5. A list of bacterial strains and plasmids used
Strain or plasmid
Genotypes or characteristics
Source or reference
Provided by Dr Keiichi Hiramatsu (Juntendo University, Tokyo, Japan).
The expression of RNAII of the agr locus is low.
Provided by Dr Richard P. Novick (New York University School of Medicine, NY).
Provided by Dr Teiko Murai (Toho University, Tokyo, Japan).
Provided by Dr Naotake Ogasawara (Nara Institute of Science and Technology, Nara, Japan).
Em, erythromycin; Cm, chloramphenicol; Km, kanamycin; Tet, tetracycline.
pND50 with S. aureus cvfA harboring G229A mutation
pND50 with S. aureus cvfA harboring I232N mutation
pND50 with S. aureus cvfA harboring D423A mutation
Transformation by plasmid DNA of S. aureus was performed by electroporation (Inoue et al., 2001). Phage transduction was performed using phage 80α(Novick, 1991). Preparation of genomic DNA from S. pyogenes and transformation of plasmid DNA into S. pyogenes were performed according to the methods previously reported (Caparon and Scott, 1991). Transformation of E. coli, extraction of plasmid DNA from E. coli, polymerase chain reaction (PCR), and Southern blot analyses were performed as previously reported (Inoue et al., 2001). pYC60 was constructed by insertion of the Streptococcal origin (HindIII fragment) from pYT1140 (Terao et al., 2001) into pCK20 at the HindIII site (Ichihashi et al., 2003). To construct the plasmid for complementation experiments, the cvfA, cvfB and cvfC genes of the S. aureus RN4220 strain and the cvfA gene of the S. pyogenes SSI-9 strain were amplified by PCR using the oligonucleotide primers FccvfA and RccvfA, FccvfB and RccvfB, FccvfC and RccvfC, or FccvfAsp and RccvfAsp (supplemental data online, TableS2). The amplified DNA fragments were inserted into pND50 and pYC60, resulting in pC1129, pC1223, pC1262 and pYC601 respectively (Table 5).
One hundred genes were selected from the S. aureus genomic genes whose function was unknown, that had an ORF size of over 800 bp, and that were conserved among other bacteria based on previous studies (Kuroda et al., 2001) (Table S1). DNA fragments containing an approximately 600-bp internal region of each ORF near the translation initiation site (referred to as +1) were amplified by PCR using RN4220 genomic DNA as a template, and inserted into pMutinT3. The resulting plasmids were introduced into the S. aureus strain RN4220 and colonies resistant to erythromycin were isolated on Luria–Bertani agar plates containing 10 µg ml−1 erythromycin. In the cvfA gene (SA1129 in the N315 genome database), the internal region (+84 to +710) was amplified by PCR using oligonucleotide primers FtcvfA and RtcvfA (TableS2). The amplified 645-bp fragment was digested and inserted into pMutinT3 at the HindIII and BamHI site, resulting in pT1129. A strain resistant to erythromycin was obtained by transformation of the RN4220 strain with pT1129, resulting in the M1129 strain (Table 5). In the cvfB and cvfC genes (SA1223 and SA1262, respectively, in the N315 genome database), the internal regions (cvfB gene, +83 to +706 and cvfC gene, +16 to +702) were amplified by PCR using the primers listed in TableS2, and used for targeting. To insert the suicide vector near the cvfB or cvfC genes without disrupting the corresponding gene, alternative primers were used for PCR amplification of Fd1223 and Rd1223 for the cvfB locus and Fd1262 and Rd1262 for the cvfC locus (TableS2). Disruption of the cvfA, cvfB, and cvfC gene and integration of their corresponding targeting vector to the desired chromosomal locus were confirmed by Southern blot analyses using the PCR-amplified region of each gene used for targeting and pMutinT3 as probes.
In the cvfA gene (SPY1633 in the SF370 genome database) of S. pyogenes, the internal region (+10 to +646) was amplified by PCR using SSI-9 genomic DNA as a template and oligonucleotide primers FtcvfAsp and RtcvfAsp (TableS2), and was inserted into pSF151, resulting in pCKT1633. A strain resistant to kanamycin was obtained by introducing pCKT1633 to the SSI-9 strain of S. pyogenes, resulting in the M1633 strain (Table 5). Disruption of the cvfA gene on chromosomal DNA was confirmed by Southern blot analysis.
Toxin production analysis
Secretion of haemolysin by S. aureus was examined on tryptic soy broth agar containing 5% sheep blood. An aliquot (2 µl) of overnight bacterial culture was spotted onto the plates and incubated overnight at 37°C. To examine the production of haemolysin by S. aureus or S. pyogenes in liquid medium, supernatants of overnight cultures (centrifuged at 15 krpm, 2 min) were incubated with sheep or rabbit red blood cells, respectively, at 37°C for 1 h (Vandenesch et al., 1991; Ruiz et al., 1998). Reaction mixtures were centrifuged (1 krpm, 5 min) and the increase in OD405 of the supernatant was determined. The activity was expressed by haemolytic units ml−1 (HU ml−1) corresponding to the reciprocal of the dilution of supernatant that yielded 50% lysis of erythrocytes.
To examine DNase activity secreted from S. aureus, overnight cultures (2 µl) were spotted onto DNA agar plates (Eiken Chemical) and incubated overnight. DNA was visualized by precipitation with 1.5 N HCl. Production of DNase secreted from S. aureus or S. pyogenes in liquid medium was examined according to the previously reported method (Cuatrecasas et al., 1967). In brief, salmon sperm DNA (500 µg) was incubated with supernatants (2 µl) of centrifuged-overnight cultures for 10 min at 37°C. The reaction cocktails were mixed with the same volume of 10% trichloroacetic acid, and centrifuged at 15 krpm for 10 min. The OD260 of the supernatants was measured.
To determine the activity of proteases secreted from S. aureus and S. pyogenes, overnight cultures (2 µl) were spotted onto tryptic soy broth agar plates containing 3.3% skim milk and incubated overnight (Pailin et al., 2001). The size of the cleared zone appearing around the growing bacteria was measured.
To determine the streptokinase activity of S. pyogenes, 50 µl of the 10-fold diluted supernatant of overnight culture was incubated with 6 µg of human plasminogen at 37°C for 30 min. The mixture was reacted with a synthetic substrate for plasmin (D-Val-Leu-Lys-p-nitroanilide dihydrochloride, Sigma Aldrich). The OD405 was determined with a microplate reader and the amount of digested substrate was calculated (Kreikemeyer et al., 2001).
Construction of the agr-expressing strain
The intact agr region (5457 bp) was amplified by PCR using primers Fagr and Ragr (TableS2), from genomic DNA of the NCTC8325 strain as a template. The amplified DNA fragment was inserted into pSF151 and the resulting plasmid was named ‘pIagr’. The RN4220 strain was transformed with pIagr and a kanamycin-resistant strain (CK501) was obtained. The integration of the intact agr operon into chromosomal DNA was confirmed by Southern blot analysis with probes of the agr region (5457 bp) and pSF151.
Northern blot analysis
Staphylococcus aureus cells were treated with 200 µg ml−1 of lysostaphin for 5 min and total RNA was extracted using an RNeasy Protect Bacteria Kit (Qiagen). RNA 2.5 µg was electrophoresed in a 1.2% agarose gel containing 6.6 M formaldehyde and transferred onto a nitrocellulose membrane (GeneScreen Plus, Perkin Elmer Life Sciences). DNA fragments of the agrA, RNAIII, and spa genes were amplified by PCR using the primers listed in TableS2, labelled with [α-32P]dCTP by random priming, and used as probes. Hybridization was performed at 42°C.
Site-directed mutagenesis was performed according to the method of Weiner et al. (1993). The cvfA gene was cloned into pUC119, resulting in pUC119-cvfA. Using the primer pairs for mutagenesis listed in TableS2 (such as FG229A and RG229A), PCR amplification was performed using pUC119-cvfA as a template. The amplified DNA fragment was self-ligated and transformed into JM109. The plasmids were extracted and sequenced to confirm the desired cvfA point mutation. The mutated cvfA gene was excised from the plasmid and subcloned into pND50, resulting in pG229A, pI232N, and pD423A.
Bacterial infection experiment with silkworms
Fertilized silkworm eggs (Hu·Yo × Tukuba·Ne), were obtained from Ehime Sansyu. Hatched larvae were fed Silkmate 2S (Nosan Corporation) at 27°C. The fourth molted larvae were gathered and hatched fifth-instar larva were fed with an antibiotic-free artificial food, Silkmate (Katakura Industries), for 1 day. The larvae were injected with 0.05 ml of bacterial suspensions into the haemolymph through the dorsal surface using a 27-gauge needle (Hamamoto et al., 2004). Overnight cultures of S. aureus were diluted 10- and 20-fold with saline, whereas overnight cultures of S. pyogenes were concentrated 50-fold. If the numbers of cells were not comparable, the assay were re-examined. The injected larvae were maintained without food in a safety cabinet (BHC-1303IIA; Airtech Japan) at 27°C with 50% humidity and survival was monitored for 5 day after the injection. To make the assay more reproducible, food was not supplied during infection. The challenge dose (1 × 107 CFU) of S. aureus RN4220 used in this study (1/20 cultures) caused death of half of the larvae at 30 h post infection. The lower dose, 1/50 cultures, caused death of the silkworms approximately 80 h post infection (data not shown). A lower challenge dose makes the assay more sensitive and effective to identify the genes with a smaller contribution to virulence. The plasmid stability for pND50 in S. aureus or pYC60 in S. pyogenes in silkworms at 27 h post infection was over 90% or 100% respectively. Statistical analyses of the survival curves were performed using one-sided rank log tests (the PRISM software package, Graphpad software). P-values of less than 0.05 were considered to be statistically significant.
Lethality test of mice
CD-1 mice (5 week-old, female, 20–25 g) were obtained from Charles River Laboratories. Overnight cultures of S. aureus were suspended in 5% hog gastric mucin and aliquots (0.5 ml) were injected intraperitoneally. Survival of the mice was examined for 3 day (Mazmanian et al., 2000). For S. pyogenes, overnight cultures were suspended in PBS. Mice were injected intraperitoneally with the suspensions (0.5 ml) and survival was determined after 6 day. LD50 was determined using the method of Reed and Muench (1938).
We thank Drs R. P. Novick, N. Ogasawara, T. Ohta, T. Murai, K. Hiramatsu, T. Ito and K. Murakami for kindly providing bacterial strains, plasmids, and phages. This study was supported in part by research grants from the Japan Society for the Promotion of Science (JSPS). C. K. was the recipient of a predoctoral fellowship from JSPS.