Evaluation of Staphylococcus aureus virulence factors using a silkworm model


Correspondence: Chikara Kaito, Laboratory of Microbiology, Graduate School of Pharmaceutical Sciences, The University of Tokyo, 3-1, 7-Chome, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Tel.: +81 3 5841 4825; fax: +81 3 5684 2973; e-mail: kaito@mol.f.u-tokyo.ac.jp


Previous studies have indicated that the silkworm model is useful for identifying virulence genes of Staphylococcus aureus, a human pathogenic bacterium. Here we examined the scope of S. aureus virulence factors that can be evaluated using the silkworm model. Gene-disrupted mutants of the agr locus, arlS gene and saeS gene, which regulate the expression of cell surface adhesins and hemolysins, exhibited attenuated virulence in silkworms. Mutants of the hla gene encoding α-hemolysin, the hlb gene encoding β-hemolysin, and the psmα and psmβ operons encoding cytolysins, however, showed virulence in silkworms indistinguishable from that of the parent strain. Thus, these S. aureus cytolysins are not required for virulence in silkworms. In contrast, the gene-disrupted mutants of clfB, fnbB and sdrC, which encode cell-wall-anchored proteins, attenuated S. aureus virulence in silkworms. In addition, the mutant of the srtA gene encoding sortase A, which anchors cell-wall proteins, showed attenuated virulence in silkworms. These findings suggest that the silkworm model can be used to evaluate S. aureus cell-wall proteins and regulatory proteins as virulence factors.


The infectious process of pathogenic bacteria in host animals involves adherence to host cell surfaces, destruction of host cells and dissemination into other tissues. In these processes, bacterial factors interact with host factors. In host animals, innate immunity, which is independent of antibody function, plays a major role at the earliest stage of infection in discriminating pathogens. The innate immune system is highly conserved among vertebrates and invertebrates (Garsin et al., 2001; Sifri et al., 2003; Begun et al., 2005; Garcia-Lara et al., 2005). For example, Toll receptors recognize pathogens in both humans and Drosophila (Hoffmann, 1995). Therefore, invertebrate model animals such as Drosophila melanogaster and Caenorhabditis elegans have been used to study the interaction between host and human pathogens to gain knowledge of the events applicable in mammals (Tan et al., 1999; Needham et al., 2004; Garcia-Lara et al., 2005).

We previously reported that Staphylococcus aureus, a human pathogen, was lethal to silkworms, which are larvae of the moth Bombyx mori, a lepidopteran species that has been cultured by humans for over 5000 years for silk production (Kaito et al., 2002; Hamamoto et al., 2004). The large size of the silkworm allows for injection of quantitative amounts of samples into the hemolymph using syringes, a marked advantage over small invertebrate animals, including D. melanogaster and C. elegans (Kaito & Sekimizu, 2007; Kurokawa et al., 2007; Fujiyuki et al., 2010). The silkworm can be maintained at 37 °C, the temperature at which most pathogenic bacteria against humans show high virulence (Kaito et al., 2011a). Use of the silkworm model enabled us to identify S. aureus novel virulence genes, cvfA, cvfB, cvfC and sarZ, from hypothetical genes that are conserved among bacteria (Kaito et al., 2005, 2006; Matsumoto et al., 2007, 2010; Nagata et al., 2008; Ikuo et al., 2010). These genes contribute to virulence in mice and regulate the expression of hemolysins. Injection of α-hemolysin and β-hemolysin from S. aureus into silkworm hemolymph is lethal to silkworms (Hossain et al., 2006; Usui et al., 2009). α-Hemolysin and β-hemolysin contribute to S. aureus virulence in mammals (O'Callaghan et al., 1997; Bubeck Wardenburg et al., 2007). Phenol-soluble modulins (PSMs) have recently been identified as cytolysins against erythrocytes and neutrophils (Wang et al., 2007). Expression of these hemolysins is positively regulated by the agr locus (Novick, 2003; Queck et al., 2008). Our previous studies of the RN4220 strain transformed with the intact agr locus indicated that the agr locus contributes to S. aureus virulence in silkworms (Kaito et al., 2005). These findings suggest that S. aureus possesses virulence factors that are not only specific for humans but also applicable to other invertebrates, and that the silkworm model is effective for the functional analysis of S. aureus virulence factors. The overall scope of S. aureus virulence factors that can be evaluated using the silkworm model, however, remains to be elucidated.

In the present study, we evaluated virulence factors of S. aureus that have been characterized in mammalian infection models. Using mammalian systems, S. aureus genes encoding hemolysins and adhesins were identified to be involved in the infectious process (see Table 3 below). We constructed disruption mutants for these genes and compared their virulence in silkworms with that of the parent strain. We previously evaluated the virulence of the S. aureus RN4220 strain in silkworms (Kaito et al., 2005). The strain is constructed by mutagen treatment and contains previously unidentified mutations in the genome (Traber & Novick, 2006; Nair et al., 2011). For example, a point mutation in the agr locus, which positively regulates the expression of exotoxins, was discovered in the genome of RN4220, and results in decreased hemolysin production (Traber & Novick, 2006). Here, we used NCTC8325-4 as the parent strain.

Materials and methods

Bacterial strains and growth conditions

Escherichia coli JM109 was used as the host for plasmids. Escherichia coli strains transformed with plasmids were cultured at 37 °C in Luria–Bertani broth containing 25 μg mL−1 chloramphenicol or 50 μg mL−1 kanamycin. Staphylococcus aureus strains were aerobically cultured in tryptic soy broth at 37 °C, and 12.5 μg mL−1 chloramphenicol, 50 μg mL−1 kanamycin or 1 μg mL−1 tetracycline was added to maintain the chromosomally integrated plasmids. Bacterial strains and plasmids used in this study are listed in Table 1.

Table 1. List of the bacterial strains and plasmids used
Strain or plasmidGenotypes or characteristicsSource or reference
  1. Cm, chloramphenicol; Km, kanamycin; Tet, tetracycline.

S. aureus
RN4220NCTC8325-4, restriction mutantPeng et al. (1988)
NCTC8325-4NCTC8325 cured of φ11, φ12 and φ13Novick (1967)
M1007NCTC8325-4 hla::pT1007; CmrThis study
M1811NCTC8325-4 hlb::pT1811; CmrThis study
M1811-1NCTC8325-4 hlb::pT1811R; KmrThis study
M1007-1NCTC8325-4 hla::pT1007, hlb::pT1811R; Cmr, KmrThis study
M0406-7NCTC8325-4 psmα::ermAM; ErmrThis study
M1056-7NCTC8325-4 psmβ::aph; KmrThis study
M1844NCTC8325-4 agr::tetM (transduction from RN6911); TetrMatsumoto et al. (2007)
M1248NCTC8325-4 arlS::pT1248; CmrThis study
M0661NCTC8325-4 saeS::pT0661; CmrThis study
M2316NCTC8325-4 srtA::pT2316; CmrThis study
M0107NCTC8325-4 spa::pT0107; CmrThis study
M0742NCTC8325-4 clfA::pT0742; CmrThis study
M2423NCTC8325-4 clfB::pT2423; CmrThis study
M2291NCTC8325-4 fnbA::pT2291; CmrThis study
M2290NCTC8325-4 fnbB::pT2290; CmrThis study
M1964NCTC8325-4 fmtB::pT1964; CmrThis study
M1577NCTC8325-4 sasC::pT1577; CmrThis study
M0977NCTC8325-4 isdA::pT0977; CmrThis study
M2285NCTC8325-4 sasG::pT2285; CmrThis study
M0022NCTC8325-4 sasH::pT0022; CmrThis study
M1552NCTC8325-4 sasI::pT1552; CmrThis study
M0519NCTC8325-4 sdrC::pT0519; CmrThis study
M0520NCTC8325-4 sdrD::pT0520; CmrThis study
E. coli
JM109General purpose host strain for cloningTakara Bio
pCK20S. aureus integration vector for gene inactivation; CmrIchihashi et al. (2003)
pSF151S. aureus integration vector for gene inactivation; KmrTao et al. (1992)

DNA manipulation

Transformation of S. aureus with plasmids was performed by electroporation (Schenk & Laddaga, 1992). Phage transduction was performed using phage 80α (Novick, 1991). Transformation of E. coli, extraction of plasmid DNA, PCR and Southern blot hybridization were performed according to Sambrook & Russell (2001). Genomic DNA from S. aureus cells was extracted using a QIAamp DNA Blood Kit (Qiagen) after digestion of cell-wall components with lysostaphin.

Gene disruption of S. aureus

The S. aureus gene was disrupted by the integration of a suicide vector into a chromosome by single cross-over homologous recombination (Kaito et al., 2005). The internal region of the target ORF near the translation initiation site was amplified by PCR using primer pairs (Table 2) and NCTC8325-4 genomic DNA as template. The amplified DNA fragment was cloned into the multi-cloning site of pCK20 or pSF151, resulting in targeting vectors. Staphylococcus aureus RN4220 was transformed with the targeting vector, resulting in the gene-disrupted mutant of RN4220. The gene disruption was transferred to strain NCTC8325-4 using phage 80α, resulting in the gene-disrupted mutant of NCTC8325-4. The gene disruption was confirmed by Southern blot hybridization analyses (Supporting Information, Fig. S1). To delete the psmα and psmβ operons, the deletions in strain RN4220 (Kaito et al., 2011b) were transferred to NCTC8325-4 using phage 80α, resulting in M0406-7 and M1056-7.

Table 2. Primers used in this study
PrimerTargetSequence (5′–3′)

Infection experiment using silkworms

Silkworms were raised from fertilized eggs at 27 °C in an air incubator (MIR-554; Sanyo Electric Co., Tokyo, Japan) (Kaito et al., 2002, 2005). The fertilized eggs were purchased from Ehime Sansyu Co. (Ehime, Japan). Hatched larvae were fed an artificial diet (Silkmate 2S; Nosan Corp., Kanagawa, Japan). Fifth instar larvae were fed an antibiotic-free artificial diet (Katakura Industries Co., Ltd, Tokyo, Japan) for 1 day and injected with serial dilutions of S. aureus overnight cultures using a 1-mL syringe equipped with a 27G needle and maintained at 27 °C in a safety cabinet (Airtech Japan). Silkworms that did not move when picked up with a platinum loop at 24 h after the injection were confirmed dead.


Staphylococcus aureushla- and hlb-encoding hemolysins do not contribute to virulence in silkworms

We injected silkworms with twofold serial dilutions of overnight culture of NCTC8325-4 and monitored the survival of silkworms at 24 h after the injection. The lethal dose (50%; LD50) of NCTC8325-4 was 1 × 107 CFU (Table 4), identical to that of strain RN4220 (Kaito et al., 2005), indicating no difference between these two strains in their silkworm killing abilities.

We used strain NCTC8325-4 to examine whether the hla gene encoding α-hemolysin, the hlb gene encoding β-hemolysin, the psmα operon encoding PSMα1, α2, α3 and α4, the psmβ operon encoding PSMβ1 and β2, and the agr locus contribute to S. aureus virulence in silkworms. The LD50 values of the hla-disrupted mutant, hlb-disrupted mutant, hla/hlb double-disrupted mutant, psmα-deleted mutant and psmβ-deleted mutant were similar to those of the parent strain (Table 4). Thus, hla, hlb, psmα and psmβ encoding hemolysins do not contribute to S. aureus virulence in silkworms. In contrast, the LD50 of the agr mutant was 2.5-fold higher than that of the parent strain (Table 4). This confirms previous findings that the agr locus contributes to S. aureus virulence in silkworms, and suggests that the agr locus functions in silkworms via hla-, hlb-, psmα- and psmβ-independent pathways.

Two-component virulence regulatory systems of S. aureus, arlRS and saeRS, contribute to virulence in silkworms

Staphylococcus aureus possesses 16 two-component regulatory systems (Cheung et al., 2004). Among them, arlRS and saeRS broadly regulate the expression of virulence genes (Fournier et al., 2001; Liang et al., 2005, 2006). The arlRS-deleted mutant exhibited attenuated virulence in a mouse systemic infection model (Benton et al., 2004). The saeRS-deleted mutant showed attenuated virulence in a mouse pyelonephritis infection model (Liang et al., 2006). We examined whether the arlS and saeS genes of S. aureus contribute to virulence against silkworms. The LD50 values of the arlS- and saeS-disrupted mutants were 2.7- and 1.8-fold higher than that of the parent strain, respectively (Table 4). This indicates that arlS and saeS contribute to virulence of S. aureus against silkworms.

Sortase A of S. aureus is required for virulence in silkworms

Cell-wall-anchored proteins of S. aureus are reported to contribute to virulence by facilitating bacterial attachment to host tissues or escape from immune systems (Foster & Hook, 1998). Sortase A is required for anchoring of various proteins to the cell wall (Mazmanian et al., 1999). A gene-disrupted mutant of srtA encoding sortase A had attenuated virulence in mouse infection models (Table 3) (Jonsson et al., 2002, 2003; Weiss et al., 2004). We tested whether the srtA-disrupted mutant showed decreased virulence in silkworms. The LD50 of the srtA-disrupted mutant was 3.1-fold higher than that of the parent strain (Table 4). This suggests that the anchoring of cell-wall proteins by sortase A is required for S. aureus virulence in silkworms.

Table 3. Staphylococcus aureus virulence genes examined in this study
CategoryGeneProductSA no.Vertebrate modelInvertebrate model
  1. SA no., the identity number for S. aureus ORFs in the N315 genome database (Kuroda et al., 2001). The ‘Model’ column refers to the gene-disrupted mutant of S. aureus with attenuated virulence in the model. ND, not determined in previous studies. NA, the gene-disrupted mutant did not show attenuated virulence in the model. Manduca sexta is a lepidopteran insect species.


Mouse pneumonia (Bubeck Wardenburg et al., 2007)

Rabbit corneal infection (O'Callaghan et al., 1997)

C. elegans (Sifri et al., 2003; Bae et al., 2004)
hlbβ-HemolysinSA1811Rabbit corneal infection (O'Callaghan et al., 1997)ND









Not registered

Mouse systemic infection (Wang et al., 2007)

Mouse skin infection (Wang et al., 2007)






Not registeredNA in mouse systemic infection (Wang et al., 2007)ND

AgrA, AgrB,





Mouse pneumonia (Heyer et al., 2002)

Rabbit corneal infection

(O'Callaghan et al., 1997)

Silkworm (Kaito et al., 2005)

C. elegans (Sifri et al., 2003)

Manduca sexta (Fleming et al., 2006)

NA in Drosophila (Needham et al., 2004)

arlRSArlR, ArlS




Mouse systemic infection (Benton et al., 2004)NA in Drosophila (Needham et al., 2004)
saeRSSaeR, SaeS



Mouse pyelonephritis infection (Liang et al., 2006)

C. elegans (Bae et al., 2004)

NA in Drosophila (Needham et al., 2004)

SortasesrtASortase ASA2316

Mouse acute lethal infection (Weiss et al., 2004)

Mouse arthritis (Jonsson et al., 2002, 2003; Weiss et al., 2004)

Mouse kidney infection (Weiss et al., 2004)

Mouse renal abscess (Cheng et al., 2009)

Rat endocarditis (Weiss et al., 2004)

NA in C. elegans (Bae et al., 2004)
Surface proteinspaProtein ASA0107

Mouse arthritis (Palmqvist et al., 2002)

Mouse renal abscess (Cheng et al., 2009)

clfAClumping factor ASA0742Mouse renal abscess (Cheng et al., 2009)ND
clfBClumping factor BSA2423

Mouse renal abscess (Cheng et al., 2009)

Human nasal colonization (Wertheim et al., 2008)

fnbAFibronectin binding protein ASA2291Mouse arthritis (Palmqvist et al., 2005)ND
fnbBFibronectin binding protein BSA2290Mouse arthritis (Palmqvist et al., 2005)ND
isdAIsdASA0977Mouse renal abscess (Cheng et al., 2009)ND
sasHSasHSA0022Mouse renal abscess (Thammavongsa et al., 2009)ND
sdrDSdrDSA0520Mouse renal abscess (Cheng et al., 2009)ND
Table 4. Killing activity of virulence gene-disrupted mutants in the silkworm model
StrainGenotypeLD50 (×10CFU)LD50 ratioP-valueAttenuation
  1. Overnight cultures of Staphylococcus aureus gene-disrupted mutants were twofold serially diluted and injected into the silkworm hemolymph (= 5), and survival was determined at 24 h after the injection (Table S1). The CFU values that cause 50% of the silkworms to die (LD50) were determined from survival curves (sigmoid dose–response curves; PRISM software package, Graphpad software). The data shown are means ± SD from at least three independent measurements of LD50. The LD50 ratio is the LD50 of the mutant/LD50 of the parent strain. Student's t-test P-value was calculated between the gene-disrupted mutant and parent strain. NS, not significant (> 0.05). ‘Attenuation +’, LD50 ≥ 1.5 and  0.05, which we judged as the mutant having attenuated virulence against silkworms.

NCTC8325-4Parent10 ± 41.0  
M1007hla8 ± 10.8NS 
M1811hlb12 ± 31.2NS 
M1007-1hla/hlb11 ± 31.1NS 
M0406-7psmα7 ± 20.7NS 
M1056-7psmβ14 ± 81.4NS 
M1844agr25 ± 82.50.01+
M1248arlS27 ± 152.7< 0.01+
M0661saeS18 ± 61.8< 0.01+
M2316srtA31 ± 103.1< 0.01+
M0107spa8 ± 30.8NS 
M0742clfA16 ± 81.6NS 
M2423clfB15 ± 81.50.03+
M2291fnbA8 ± 10.8NS 
M2290fnbB19 ± 21.9< 0.01+
M1964fmtB10 ± 21.0NS 
M1577sasC10 ± 31.0NS 
M0977isdA12 ± 61.2NS 
M2285sasG13 ± 11.3NS 
M0022sasH10 ± 51.0NS 
M1552sasI15 ± 41.5NS 
M0519sdrC31 ± 93.1< 0.01+
M0520sdrD15 ± 81.5NS 

Cell-wall-anchored proteins of S. aureus contribute to virulence in silkworms

Twenty proteins are known to be anchored to the cell wall by sortase A in S. aureus (Roche et al., 2003). Among them, we selected 13 proteins – protein A, clumping factor A and B, fibronectin binding protein A and B, FmtB, SasC, IsdA, SasG, SasH, SasI, SdrC and SdrD – and tested whether these proteins are required for the virulence of S. aureus against silkworms. All of the spa-, clfA-, fnbA-, fmtB-, sasC, isdA-, sasG-, sasH-, sasI- and sdrD-disrupted mutants showed virulence in silkworms similar to that of the parent strain (Table 4). In contrast, the LD50 values of the clfB-, fnbB- and sdrC-disrupted mutants were significantly higher than that of the parent strain (Table 4). These findings indicate that ClfB, FnbB and SdrC contribute to the virulence of S. aureus in silkworms. The sdrC-disrupted mutant had severely attenuated virulence in silkworms, indicating that SdrC plays a prominent role in infection by S. aureus in silkworms.


Little contribution of hemolysins to virulence of S. aureus against silkworms

Our previous studies indicated that injection of α-hemolysin and β-hemolysin was lethal to silkworms (Hossain et al., 2006). The findings of the present study revealed that genes encoding α- and β-hemolysin were not necessary for S. aureus to kill silkworms. In the S. aureus infectious processes in silkworms, levels of α- and β-hemolysin expression might be too low to kill silkworms.

Roles of virulence regulators of agr, arlS and saeS for S. aureus infectious processes in silkworms

The findings of this and our previous study revealed that the agr locus, which positively regulates the expression of genes encoding hemolysins, contributes to the virulence of S. aureus in silkworms. The agr system also senses cell density and broadly regulates the expression of virulence factors (Novick, 2003). The finding that disruption of genes encoding α-hemolysin, β-hemolysin and PSM peptides did not affect virulence of S. aureus in silkworms led us to hypothesize that factors other than α-hemolysin, β-hemolysin and PSMs, which that are regulated by the agr locus, contribute to S. aureus virulence.

Here, we revealed that arlS and saeS, encoding sensor proteins of the two-component systems, are required for S. aureus virulence in silkworms. The expression of arlRS is activated by high osmolarity or quinolone, an inhibitor of DNA gyrase (Fournier & Klier, 2004). The expression of saeRS is activated by hydrogen peroxide or α-defensin, an antimicrobial peptide (Kuroda et al., 2007; Geiger et al., 2008; Palazzolo-Ballance et al., 2008). These findings suggest that S. aureus requires ArlRS and SaeRS to adapt similarly to the stress induced by silkworm innate immunity.

Roles of cell-wall-anchored proteins in S. aureus infectious processes in silkworms

The present study revealed that protein A is not required for S. aureus virulence in silkworms. Protein A contributes to the virulence of S. aureus by interacting with immunoglobulin in mammalian blood (Palmqvist et al., 2002). The lack of the requirement for spa in S. aureus infection of silkworms is presumably due to the absence of immunoglobulin in invertebrates, including silkworms.

We demonstrated that cell-wall-anchored proteins, ClfB, FnbB and SdrC, contributed to the virulence of S. aureus in silkworms. To our knowledge, this is the first report that cell-wall-anchored proteins contribute to the virulence of S. aureus in an invertebrate model animal. ClfB binds cytokeratins of mammalian epithelial cells and the interaction is required for S. aureus colonization onto nasal epithelial cells (Wertheim et al., 2008); FnbB binds mammalian fibronectin and contributes to the virulence of S. aureus (Palmqvist et al., 2005); and SdrC is required for adherence of S. aureus to mammalian epithelial cells (Barbu et al., 2008; Corrigan et al., 2009). Therefore, ClfB, FnbB and SdrC are presumably required for adherence of S. aureus to silkworm tissues by binding silkworm proteins that are homologous to the mammalian target proteins.

Comparison of silkworm model with other invertebrate models

Invertebrate animal models of S. aureus infection include C. elegans, D. melanogaster and Manduca sexta, in addition to silkworms (Sifri et al., 2003; Needham et al., 2004; Fleming et al., 2006). In the C. elegans model, bacteria were eaten by worms and the number of surviving worms was counted (Sifri et al., 2003). In the D. melanogaster model, bacteria were injected into adult flies by injuring animals with tungsten needles that were dipped in a solution containing bacteria, and the number of surviving flies was counted (Needham et al., 2004). In the M. sexta model, bacteria were injected into larvae by using microsyringes (Fleming et al., 2006). In the C. elegans model, the agr locus, saeRS and hla genes of S. aureus are required to kill worms, although srtA is not (Table 3) (Sifri et al., 2003; Bae et al., 2004). In the D. melanogaster model, the agr locus, saeRS and arlRS of S. aureus were not required for killing flies (Table 3) (Needham et al., 2004). In the M. sexta model, the agr locus of S. aureus is involved in killing larvae (Table 3) (Fleming et al., 2006). Our present study revealed that agr, saeRS, arlRS and srtA of S. aureus were required for killing silkworms, whereas hla was not required. The different results between these animal models may be due to different sensitivities of animals against exotoxins, different adhesive characteristics of cell surfaces to bacterial cells, and different experimental conditions, such as temperatures and infection routes.


The findings of the present study revealed that genes encoding hemolysins of S. aureus are not required for killing silkworms, whereas some genes encoding cell-wall proteins and regulatory proteins are required. Therefore, the silkworm model is applicable for evaluating cell-wall proteins and its regulators in S. aureus. Whether the gene-disrupted mutants that attenuated killing ability against silkworms show characteristic clinical presentation in silkworms compared with the wild-type strain should be investigated in future studies to understand the roles of virulence factors in S. aureus infection. Silkworms have a smaller genome and fewer genes than mammals. The size of the silkworm is also larger than that of other invertebrate model animals, supplying an adequate bulk of biomaterials for biochemical studies. These advantages of the silkworm model will contribute to promote an understanding of basic virulence systems of S. aureus and other pathogens.


We thank Timothy J. Foster and Richard P. Novick for the S. aureus strains. This work was supported by Grants-in-Aid for Scientific Research, and in part by the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (NIBIO) and Genome Pharmaceuticals Institute.