A novel streptococcal surface protease promotes virulence, resistance to opsonophagocytosis, and cleavage of human fibrinogen

Group B streptococcus (GBS) is an important human pathogen. In this study, we sought to identify mechanisms that may protect GBS from host defenses in addition to its capsular polysaccharide. A gene encoding a cell-surface–associated protein (cspA) was characterized from a highly virulent type III GBS isolate, COH1. Its sequence indicated that it is a subtilisin-like extracellular serine protease homologous to streptococcal C5a peptidases and caseinases of lactic acid bacteria. The wild-type strain cleaved the α chain of human fibrinogen, whereas a cspA mutant, TOH121, was unable to cleave fibrinogen. We observed aggregated material when COH1 was incubated with fibrinogen but not when the mutant strain was treated similarly. This suggested that the product(s) of fibrinogen cleavage have strong adhesive properties and may be similar to fibrin. The cspA gene was present among representative clinical isolates from all nine capsular serotypes, as revealed by Southern blotting. A cspA^– mutant was ten times less virulent in a neonatal rat sepsis model of GBS infections, as measured by LD₅₀ analysis. In addition, the cspA^– mutant was significantly more sensitive than the wild-type strain to opsonophagocytic killing by human neutrophils in vitro. Taken together, the results suggest that cleavage of fibrinogen by CspA may increase the lethality of GBS infection, potentially by protecting the bacterium from opsonophagocytic killing.

Group B streptococcus (GBS) is a major cause of serious neonatal bacterial infections (1) and is the most common cause of sepsis and pneumonia in newborns. GBS is also a significant cause of postpartum endometritis. In addition, recent data implicate GBS as an increasingly important cause of invasive infections in adults, especially among the immunocompromised (2).

Bacterial pathogens have evolved a diverse array of defenses to combat the innate immune system, which is an important first line of defense against bacterial infections in the nonimmune host. Complement cells, as well as phagocytic cells such as polymorphonuclear leukocytes (PMNs) and macrophages, play major roles in innate immunity. The antiphagocytic mechanisms of a number of Gram-negative bacteria such as Yersiniae and Pseudomonas aeruginosa (3) are well characterized. However, much remains to be understood about how Gram-positive pathogens evade phagocytic mechanisms. Streptococcus pneumoniae strains produce polysaccharide capsules (4) that allow the bacterium to resist complement deposition in the absence of type-specific capsule antibodies. In addition, S. pneumoniae strains produce a C3 binding protein that may function in the evasion of complement-mediated host defense (5). Enterococcus faecalis, a major cause of hospital-acquired infections, has recently been reported to synthesize a capsular polysaccharide that provides resistance to opsonophagocytic killing (6). Strains of group A streptococcus (GAS) use numerous mechanisms to evade this immune pathway, including the M protein (7) and the hyaluronic acid capsule (8).

The GBS capsule is the most well-defined virulence factor of GBS. The capsule protects GBS from opsonization by C3 through inhibition of the alternative complement pathway in the absence of type-specific capsule antibodies (9). However, GBS may express additional factors that allow it to resist opsonophagocytosis. A recent report indicated that the surface-localized streptococcal β protein binds human complement factor H and that the GBS–factor H complex retains its ability to downregulate complement activation (10).

In this study, we identify cspA, a novel gene encoding a surface-localized, serine protease–like protein that promotes GBS survival through evasion of opsonophagocytosis. CspA shows homology to a family of proteases that include C5a proteases of pathogenic streptococci (11) as well as caseinases expressed by nonpathogenic Gram-positive cocci (12). Surprisingly, we observed that CspA does not have enzymatic activity against C5a in vitro and the presence of the cspA gene was not required for casein degradation. However, the cspA gene was required for GBS cleavage of human fibrinogen, which indicates that CspA is active as a protease. Mutants that failed to express cspA displayed a significantly decreased GBS virulence in a neonatal rat model of infection and displayed increased sensitivity to opsonophagocytosis. Our findings provide evidence that CspA is a novel, surface-localized protease that plays an important role in GBS pathogenesis as an antiphagocytic surface factor.

It is likely that many factors contribute to the virulence of GBS (29). In order to identify novel virulence factors, we performed a screen to identify transposon mutants with reduced virulence in GBS. In the process of screening Tn916 ΔE mutants of the highly virulent type III GBS isolate, COH1, we identified a gene with homology to cell-surface–associated proteases. The open reading frame, which we designated cspA (Figure 1), encodes a 1,571–amino acid protein and displays homology to several extracellular serine proteases from the subtilase family. The greatest similarity of CspA (51.2% identity and 58.3% similarity) was to PrtS, which is an extracellular caseinase produced by S. thermophilus (12). The next highest similarity was to a putative extracellular protease (39% identity and 55% similarity) from the GAS genome sequence (34). The third highest similarity (45% similarity and 36% identity) was to ScpA (35) and ScpB (28), which participate in the proteolytic inactivation of the chemotactic factor C5a in GAS and GBS, respectively. In addition, CspA was similar to several caseinases of lactic acid bacteria (27% identity and 36% similarity) encoded by the prtB, prtH, and prtP genes (36).

The sequence of CspA indicates that it shares the functional and structural domains of the cell envelope–associated protease (CEP) family (36). The CEPs are a subfamily of subtilisin-like serine proteases that are found in a wide variety of bacteria. An excellent comprehensive review of the properties of the CEP protein family may be found in Siezen (36). The motifs required for the catalytic function of this family of serine proteases are present in CspA at residues 168–179 (aspartic acid motif, VAIISDGLNTNH), 237–248 (histidine motif, HGMHVTSIATA) and 565–575 (serine motif, GTSMASPHVAG); this suggests that CspA functions as a protease.

A general characteristic of caseinases is that they are synthesized as pre-pro enzymes and, after translocation across the cell membrane, are activated by autocatalytic cleavage of a pro-peptide sequence (36). A putative pre-pro domain, spanning residues 1–143, was identified in CspA (36), implying that CspA may also be synthesized as a pre-pro enzyme and may similarly undergo autocatalytic maturation.

Inactivation of cspA and scpB. To facilitate the functional analysis of CspA, allelic replacement mutagenesis of cspA was performed. A plasmid bearing a cspA::erm allele was created by deleting a portion of cspA and replacing it with an erythromycin resistance gene (see Methods). This construct was used to replace the wild-type cspA allele of COH1 (a highly virulent, type III GBS clinical isolate), as detailed in the Methods. A single clone, TOH121, was selected for further study, and the presence of the cspA::erm mutation on the chromosome of this strain was verified by Southern blotting. To provide controls for evaluating GBS C5a protease activity (see below), a mutant bearing a kanamycin resistance element insertion in the GBS C5a peptidase gene, scpB, was also constructed in the COH1 genetic background and was designated TOH97. A double mutant bearing both the cspA::erm and the scpB::kan mutations was constructed and designated TOH144. Southern blotting was also used to confirm the presence of the desired mutation for the two scpB::kan mutants (data not shown).

The cspA gene is transcribed as a monocistronic operon. We characterized the transcriptional organization of the region proximal to cspA. Examination of the DNA region next to cspA revealed open reading frames located both 5′ and 3′ of cspA (Figure 1). A gene designated sblA (for streptococcus group B lipoprotein) is located 5′ of cspA and is oriented in the opposite transcriptional direction as cspA. The predicted product shares 33% identity with AzlC from Deinococcus radiodurans (37); the precise biological function of the AzlC protein is currently unknown (38). Adjacent to cspA and oriented in the same direction are two putative genes (designated sbrA and sbsA) that encode products with homology to the response regulator and sensor proteins of two-component regulatory systems (39). A potential promoter sequence is located in the 233 bp of noncoding sequence between cspA and sbrA, suggesting that sbrA and sbsA are transcribed from a promoter that is distinct from that of cspA.

To characterize the transcriptional organization of the cspA region, RNA was isolated from THB-grown COH1 at different growth stages (early-log, mid-log, and stationary phase) and Northern analysis was performed using a probe internal to the cspA coding region (Figure 2a). A discrete band corresponding to a transcript size of 4.7 kb was present in cells at different stages of growth (Figure 2a; lanes 1–3). A control for probe specificity was performed with strain TOH121; no hybridizing material was seen for this strain, as expected. Maximal expression was observed in early-log phase (OD₆₀₀ = 0.3) (Figure 2a, lane 1). Since the coding region of cspA is approximately 4.7 kb, these observations suggest that cspA is expressed on a monocistronic transcript.

Figure 2

Northern blot analysis of cspA expression in COH1 and TOH121. Standard Northern blots were performed using equivalent amounts of RNA (5 μg) from the indicated strains. (a) Blot developed using cspA probe. Lane 1, COH1 at OD₆₀₀ = 0.3 in THB; lane 2, COH1 at OD = 0.6; lane 3, COH1 at OD = 1.7; lane 4, TOH121 at OD = 0.3; lane 5, TOH121 at OD = 1.7. (b) Blot developed using sbrA probe. Lane 1, COH1 at OD = 0.6; lane 2, TOH121 at OD = 0.6. Migration of RNA size standards is indicated (in kb) on the right.

To confirm our hypothesis that the cspA transcript is monocistronic, we performed Northern blot analysis on a gene located 3′ to cspA. It was important to verify that the phenotypes (see below) of the cspA mutant, TOH121, were not due to a polar effect of the cspA mutation onto the adjacent sbrA and sbsA genes (Figure 1). RNA was isolated from COH1 and TOH121 in mid-log phase and was hybridized to a probe internal to the coding region of sbrA (see Figure 1). The sbrA-hybridizing material ran as a smear at molecular weights that corresponded to less than 2.4 kb. Similar amounts of hybridizing material were observed in both COH1 and TOH121 (Figure 2b). The cspA and sbrA probes hybridized to distinct transcripts, suggesting that cspA is transcribed independently from the adjacent genes and that cspA is transcribed as a monocistron.

CspA is a surface-associated, cell-wall–anchored protein. In addition to predicting that CspA functions as a protease, the sequence of CspA suggests that it is secreted and subsequently anchored to the cell surface. A 35-residue signal peptide within the N-terminal end of CspA was identified using SIGSEQ (40) (Figure 1). A classic C-terminal cell-wall attachment site sequence (LPKTG) characteristic of Gram-positive surface-associated proteins (41) was located at amino acids 1536–1540 (Figure 1). To investigate the hypothesis that CspA is a surface-attached protein, we determined the subcellular localization of CspA by Western blot analysis. Antibody was raised to a GST–CspA fusion protein expressed from E. coli (see Methods), and the antibody was used to test different cellular fractions of GBS for the presence of CspA. Periplasmic extracts (see Methods) of E. coli strain TOH50 (bearing plasmid pTH5; see above) reacted strongly with the antibody raised to GST–CspA. Plasmid pTH5 contains a mutation in the CspA coding region (in comparison with the cspA gene of the wild-type isogenic strain, COH1) that prematurely terminates translation at Leu-1121; this accounts for the lower molecular mass observed for the TOH50 extracts in comparison with wild-type CspA. Proteins from culture supernatants of COH1 did not react with the antibody, even when concentrated 10-fold (data not shown). In contrast, Western blots of mutanolysin-extracted surface proteins from both COH1 and TOH97 (cspA⁺, scpB^–) revealed two protein bands of molecular masses 142 and 80 kDa (Figure 3). The migration of COH1-derived CspA on SDS-PAGE was anomalous, since it corresponded to a lower molecular mass (142 kDa) than predicted by sequence analysis for mature CspA (153 kDa). No cross-reactive bands were seen at the 142-kDa position for the cspA mutants (TOH121 and TOH144), confirming that the antibody does react with wild-type CspA from GBS. The results of the sequence analysis of CspA taken together with the Western blot data indicate the CspA is a surface-localized protein.

Figure 3

Western blot analysis of CspA expression in E. coli and GBS strains using anti-CspA sera. Lane 1 shows periplasmic extract from E. coli DH5α containing pBS (negative control); Lane 2 shows periplasmic extract from E. coli DH5α containing pTH5 (XbaI fragment bearing cspA in pBSKS–; Stratagene). Note that that the lower migration of CspA in periplasmic extracts is due to a mutation in pTH5 that prematurely terminates translation (see Methods). Lanes 3–6 show mutanolysin-extracted GBS surface proteins from COH1 (lane 3), TOH121 (cspA^–, lane 4), TOH97 (scpB^–, lane 5), and TOH144 (cspA^–, scpB^–, lane 6). Migration of molecular mass markers (in kDa) is indicated on the left.

CspA does not function as a C5a protease. Given the strong similarity of CspA to ScpB, the C5a peptidase of GBS, we hypothesized that CspA may also be active as a C5a protease. We measured C5a-ase activity (Figure 4) with a functional assay described previously (11). Briefly, GBS was preincubated with recombinant human C5a, purified human PMNs were added, and C5a-stimulated PMN adherence to gelatin-coated plastic was measured. COH1 (wild-type, scpB⁺) served as a positive control, and TOH97 (scpB^–) and another scpB^– strain, GW (32), served as negative controls. The effect of the cspA mutation was measured both in the presence and absence of the GBS C5a protease (ScpB) by comparing TOH121 (scpB⁺, cspA^–) and TOH144 (scpB^–, cspA^–) with the controls. Preincubation of C5a with COH1 or TOH121 abolished C5a activity, and the C5a protease activities of the strains were similar (Figure 4). In contrast, C5a protease activity of strains GW, TOH97, and TOH144 were low and not significantly different from each other. Thus, inactivation of CspA did not appear to have a detectable effect on C5a activity, as evaluated in this functional assay. C5a protease activity only correlated with the presence of functional ScpB. The above results, taken together, are consistent with the conclusion that CspA does not function as a C5a protease.

Figure 4

Functional assay for C5a protease activity. Shown is the percent adhesion by human PMNs to gelatin-coated tissue-culture wells after the indicated GBS strains were incubated with recombinant human C5a (28). As controls, buffer alone (buffer) or 100 ng/ml of untreated C5a (C5a) were incubated without bacteria before exposure to PMNs. GBS strains that were tested were COH1 (positive control, cspA⁺, scpB⁺), GW (32) (negative control, a type III GBS strain lacking C5a-ase activity), TOH97 (negative control, cspA⁺, scpB^–), TOH121 (cspA^–, scpB⁺), and TOH144 (cspA^–, scpB^–).

Evaluation of GBS phenotypic traits. We evaluated the expression of several GBS phenotypic traits, some of which are known GBS virulence determinants. The cspA mutant expressed α-hemolysin, CAMP factor, hippuricase, and hyaluronidase similarly to wild-type GBS (data not shown). These strains also expressed equivalent amounts of type III capsule as measured by competitive ELISA (72.7 ± 15.8 μg/mg for COH1 and 82.9 ± 26.3 μg/mg dry weight for TOH121). Invasion of A549 epithelial cell monolayers was 1.2% ± 0.3% for COH1 and 1.3% ± 0.2% for TOH121, expressed as a percentage of the number of total input bacteria invading the A549 cells.

Many of the proteases that are homologous to CspA are important for bacterial growth. For example, caseinases from lactic acid bacteria participate in the degradation of extracellular casein before utilization of the casein peptides as a nutritional source. Because GBS is auxotrophic for multiple amino acids, it must rely on exogenous amino acids. Thus, we compared the growth characteristics of the cpsA mutant to the wild-type strain in different growth media. Growth of the strains was comparable in all media that were tested, including RPMI plus 5% Casamino acids and THB (data not shown). Growth in human plasma was compared, and the strains exhibited doubling times of 0.6 hours and 0.52 hours for COH1 and TOH121, respectively. This suggests that CspA does not play a role in nutritional scavenging, at least under the experimental conditions tested.

Caseinase activity of GBS. Because of the strong similarity of CspA to PrtS, which functions as a caseinase, we tested whole cells of COH1 and TOH121 for casein degradation. Whole GBS cells were incubated with casein, the mixture was centrifuged, SDS-PAGE was performed on the supernatant, and the amount of intact casein was measured by quantitation of Coomassie-stained gels. Very little casein degradation was observed under the experimental conditions employed, and experiments with the two strains yielded similar results (Figure 5).

Figure 5

Caseinase activity assay. GBS strain COH1 was grown to stationary phase in THB, washed in PBS, concentrated 20-fold, and resuspended in PBS. A saturated solution of protease assay–grade casein (Sigma-Aldrich) was prepared, and 0.2 ml of bacteria was mixed with 0.3 ml of casein. A mock reaction with PBS (no bacteria) was also prepared. The mixtures were incubated for 24 hours at 37ºC, and an aliquot was removed for SDS-PAGE. Molecular-weight markers are depicted on the left. Lane 1 shows the PBS control, lane 2, COH1 after 0 hours of incubation, and lane 3, COH1 after 24 hours of incubation. Migration of molecular mass markers (in kDa) is indicated on the left.

The cspA gene is required for cleavage of human fibrinogen. We hypothesized that CspA, as a putative surface-localized protein, proteolyses a host factor. Therefore, to test the ability of CspA to function as a protease, we compared the ability of the cspA mutant and the wild-type strain to degrade a variety of host proteins. We tested purified human fibronectin, purified complement component C3, and purified human fibrinogen. To evaluate degradation, whole bacteria and/or mutanolysin-extracted surface proteins of COH1 and TOH121 were incubated with the test substrate, and degradation was assessed by evaluation of Coomassie-stained SDS-PAGE gels or by Western blotting. No difference in the cleavage of the test substrates was observed for all substrates except fibrinogen when the mutant and wild-type strains were compared (data not shown).

Human fibrinogen was cleaved by the wild-type strain but was not cleaved by the cspA mutant TOH121. Fibrinogen is a dimer of nonidentical subunits — α, β, and γ — that are covalently linked together by disulfide bonds (42). For fibrinogen isolated from fresh human plasma, we found that SDS-PAGE resolves the α subunit of fibrinogen into a doublet. We excised both putative α bands from an SDS-PAGE gel and confirmed their identity by MALDI-TOF MS. Peptides matching the sequence of the α fragment of human fibrinogen were obtained from both putative α bands; 28% of the fibrinogen sequence was covered by the tryptic peptides that were obtained. The lower band of the α doublet is believed to be an in vivo cleavage product in which the C-terminal 27 amino acids are removed (43). When the wild-type and CspA mutants were compared for fibrinogen cleavage, SDS-PAGE revealed that COH1 cleaved the minor species (corresponding to the lower band of the doublet) of the α subunit, whereas the mutant did not cleave fibrinogen under the conditions used (Figure 6). We also noticed that when the mutant and wild-type were incubated with fibrinogen overnight, the wild-type strain formed macroscopic aggregates that were much more prominent in comparison with the mutant strain. These aggregates were not observable in controls in which the wild-type strain was incubated in PBS alone.

Figure 6

Fibrinogen degradation assay. GBS strains COH1 (wild type) and TOH121 (cspA^–) were grown to stationary phase in THB. Cells were washed once in an equivalent volume of PBS and concentrated 20-fold. Fibrinogen was then added to a concentration of 0.61 μg/μl. The suspension was incubated with slow rotation overnight (16 hours) at 37ºC, bacteria were removed by centrifugation, and the supernatant, containing 3 μg of total fibrinogen, was analyzed by SDS-PAGE as follows: lane 1 shows COH1 after 0 hours of incubation; lane 2, TOH121 after 0 hours incubation; lane 3, COH1 after 16 hours of incubation; and lane 4, TOH121 after 16 hours of incubation. The arrow at the left of the figure denotes the migration position of the minor species of the α fragment of fibrinogen that is proteolyzed by CspA. Migration of molecular mass markers (in kDa) is indicated on the right.

CspA is essential for virulence in a neonatal rat model. Extracellular proteases have been implicated in virulence both in Gram-negative and Gram-positive bacteria. To determine if CspA contributes to virulence, we compared the LD₅₀ of TOH121 (cspA^–) to the isogenic wild type strain, COH1. A series of 10-fold dilutions of each strain were introduced by intraperitoneal injection into neonatal rats (24–48 hours old). Five separate lethality experiments were performed. The mean LD₅₀ values were 2.9 × 10³ and 2.9 × 10⁴ CFU per animal for COH1 and TOH121, respectively (P ≤ 0.0431 by the Wilcoxon matched-pair signed-ranks test). These results suggest that mutation of cspA significantly impairs the virulence of GBS in the neonatal sepsis model.

CspA promotes evasion of opsonophagocytosis. We hypothesized that CspA may allow GBS to avoid innate immune clearance in the nonimmune host, perhaps by a novel mechanism, since capsular polysaccharide expression was not affected by the cspA mutation (see above). Opsonophagocytosis is an important mechanism for bacterial clearance, and neutrophils play a major role in the elimination of GBS from the bloodstream (44). We hypothesized that the attenuated virulence of TOH121 may be a consequence of increased susceptibility to phagocytic clearance as compared with COH1. Fresh PMNs were isolated from human donors, and pooled human serum was used as a source of complement and was preabsorbed with COH1 to remove antibodies directed against the bacteria. The assay was repeated four times, and the results of a representative experiment are shown in Figure 5. COH 1-13 (16), which is an unencapsulated mutant of COH1 known to be very susceptible to PMN killing, was included as a control. The growth index (GI) of each strain was calculated as the output CFU per milliliter divided by the input CFU per milliliter. COH1 growth during the assay corresponded to more than one doubling (GI = 2.2) (Figure 7). In contrast, the negative control strain, COH1-13, was markedly killed by human PMNs during the 1-hour incubation (GI = 0.04). TOH121 exhibited a sensitivity that was intermediate between COH1 and COH1-13 (GI = 0.81). All three strains grew in the presence of heat-inactivated sera and in the absence of PMNs, yielding GIs between 2 and 4 (Figure 7 and data not shown). We also averaged the results of the four experiments and expressed the results as a ratio of the GI of the mutant to the GI of the wild type, which was 0.46 (P < 0.001 by Student’s t test). These findings suggest that CspA promotes resistance to nonimmune opsonization and killing by PMNs, though not to the same degree as observed for the capsular polysaccharide (16).

Figure 7

Opsonophagocytosis of GBS strains by PMNs. GBS strains COH1, TOH121, and the unencapsulated mutant COH 1-13 were compared for resistance to opsonophagocytosis by PMNs. Bacteria (1 × 10⁶) and human PMNs (3 × 10⁶) were incubated in 10% human sera as a complement source (preabsorbed with COH1 to remove GBS antibody) for 1 hour at 37°C. The GI was calculated as the output CFU per milliliter divided by the input CFU per milliliter. One representative experiment of four is depicted. Controls were performed for the test strain, TOH121, with heat-inactivated serum (HIS) and without PMNs.

The cspA gene is widely distributed among the GBS serotypes. Since CspA contributes to virulence, we investigated the prevalence of cspA in the other GBS serotypes. Southern blots were performed using GBS genomic DNA isolated from representative strains from each of the nine serotypes. A single DNA fragment from 16 of 18 strains hybridized to the cspA probe (data not shown), and DNA from at least one representative of each serotype hybridized to the probe. Therefore, the cspA gene is prevalent among all the serotypes examined.

It is clear that bacterial proteases have diverse functions in pathogenesis. Proteases can benefit the invading microbe by liberating nutrients from the host (45, 46) or can promote disease through direct destruction of host tissues (47). Several proteases activate host zymogens, such as plasminogen (48), kininogens, matrix metalloproteases, and proenzymes of the clotting system (47). Some of the proteases inactivate host-defense proteins, such as immunoglobulins, components of the complement system, and antimicrobial peptides (48–50), whereas others are required for intracellular survival in macrophages (49) and for adherence as well as uptake into epithelial cells (50).

In this study, we identified a GBS gene encoding a protein that shows strong similarity to CEPs from other Gram-positive bacteria and contributes to GBS virulence. CEPs are a branch of a larger family of proteins known as subtilases, which use a serine active site (51). The overall similarity of CspA to CEP proteins as well as the conservation of residues associated with the catalytic activity of serine proteases (aspartic acid, histidine, and serine motifs) suggest that CspA functions as a protease. For a comprehensive review of the domain structure of CEPs, see Siezen (36). The presence of a putative signal peptide and a cell-wall anchor (LPXTG) motif indicates that CspA is a cell-surface–associated protein, and Western blots of cell-surface–associated proteins confirmed the surface localization of CspA. Although CspA shares many of the features of the CEPs, including a putative pre-pro domain and anchor domain, CspA diverges from members of the CEP family in other aspects of its sequence. Two domains that are present in other CEPs, the “H” domain, thought to function as a cell-wall spacer, and the “W” domain, thought to function as a cell-wall–spanning domain, are absent in CspA. These domains are replaced with a novel sequence that is not found in the other CEPs; no homologous sequences in the existing protein database were found (36). The significance of these observations is not known at present but suggest that CspA has a distinct structure from the known CEPs in the vicinity of the cell-wall–proximal region, which may also influence localization of CspA.

CspA appears to be a novel member of the CEP family of subtilase-like proteases on the basis of its functional properties also. Most of the CEP family members that have been studied are from dairy lactic acid bacteria and function in the acquisition of nutrients through degradation of casein (36). In fact, the CEP family member that CspA shows the highest overall similarity to is PrtS from S. thermophilus (12), which catalyzes casein degradation. However, the inactivation of CspA did not affect the ability of GBS to degrade casein in vitro; in fact, GBS did not exhibit substantial caseinase activity under the conditions that we tested. We also considered that CspA may function to degrade other proteins for the purpose of nutrient acquisition. However, we did not detect a difference in growth rates in human plasma when the wild-type strain and the cspA mutant were compared in vitro (doubling times of 0.6 hours and 0.52 hours for COH1 and TOH121, respectively). The findings suggest that CspA does not perform a function homologous to the caseinases (36) and does not play an essential role in nutrient acquisition, at least under the conditions that we tested.

The finding that whole cells of TOH121 (cspA^–) cannot cleave fibrinogen strongly suggests that cspA encodes a protein with active proteolytic activity. Although it is a formal possibility that the cspA gene is simply necessary for the function of a distinct protease that actually cleaves fibrinogen, the simplest hypothesis to explain our data is that cspA encodes the fibrinogen-degrading protease. It is intriguing that CspA degrades the lower α band of fibrinogen, which corresponds to fibrinogen with the C-terminal 29 amino acids removed (43). Many proteases, including plasmin, will cleave the C terminus of fibrinogen (43), but it appears that CspA acts on the form of fibrinogen that has already been cleaved at the C terminus of the α chain. The significance of this observation is not currently known; however, CspA may proteolyse fibrinogen at a position distinct from the positions acted on by other bacterial proteases.

Another bacterial protease that degrades fibrinogen is the SpeB protease of GAS. SpeB also cleaves fibrinogen, but in a manner that is rather different from what we observed for CspA activity. SpeB degrades the α band of fibrinogen, which leads to the nearly complete disappearance of the α band on SDS-PAGE gels (52). In contrast, CspA does not degrade the major species of the α chain, and its cleavage of the minor α species yields bands that differ in size from the major band by less than 10 kDa. These results suggest that CspA may cleave fibrinogen more specifically than SpeB; additionally, SpeB is a cysteine protease, whereas CspA is a serine protease.

The connection between the attenuated opsonophagocytosis, the reduced virulence of TOH121, and the fibrinolytic activity of CspA is not yet clear. However, the sensitivity of TOH121 to opsonophagocytic killing may reflect the inability of TOH121 to cleave fibrinogen. GBS may be exploiting the adhesive properties of a fibrin-like substance by cleaving host fibrinogen in a manner similar to thrombin. Fibrin is formed from fibrinogen and is the key structural element of blood clots as well as a component of the extracellular matrix. In blood, the proteolysis of the N termini of the α and β chains of fibrinogen leads to the polymerization of cleaved fibrinogen, forming fibrin, which, as the structural component of blood clots, is highly adhesive. If CspA cleaves the α chain of fibrinogen at the N terminus, it could potentially expose the charged regions present on fibrinogen that are responsible for polymerization of fibrinogen and may promote the aggregation of GBS or the coating of GBS with fibrin. We have observed that when the wild-type strain (COH1) is incubated with fibrinogen, aggregates are formed and can be observed as macroscopic particles, whereas in the TOH121 (cspA^–) mutant strain, aggregation is much less prominent. These aggregates may represent GBS complexed with a fibrinogen-derived cleavage product.

One explanation for the reduced opsonophagocytosis of TOH121 is that wild-type GBS (cspA⁺) is coated with a fibrin-like substance and this substance reduces the access of opsonins or neutrophils to GBS. The other possibility is that the cellular aggregation is induced by fibrinogen cleavage; aggregated GBS may be protected from host-defense mechanisms. A third possibility is that the reduced opsonophagocytosis of GBS is not a consequence of CspA-cleaved fibrinogen. For example, it is possible that CspA proteolyses a protein distinct from fibrinogen and that degradation of this protein is responsible for the attenuated opsonophagocytosis of TOH121. Such a protein could even be bacterial in nature; the SpeB protease of GAS is known to cleave bacterial surface proteins (52). We tested for degradation of the GBS αC protein and found that the cspA gene was not required for cleavage of this surface-localized protein (data not shown). Further studies will be needed in order to link the defect in virulence to the attenuated opsonophagocytosis and fibrinogen cleavage.

We attempted to complement the cspA mutant in order to verify that that decreased opsonophagocytosis, virulence, and defect in fibrinogen cleavage were linked. However, we were unable to subclone cspA to a plasmid; attempts at cloning on both high- and low-copy plasmids failed. We are currently attempting to replace the erm^r cassette of TOH121 with the wild-type gene in order to demonstrate the causality of the cspA::erm^r mutation.

Gram-positive bacteria are important human pathogens; however, much remains to be understood concerning how Gram-positive bacteria evade opsonophagocytosis. To our knowledge, this study represents the first report of a GBS protease that functions in the evasion of opsonophagocytosis. In summary, we have identified a novel GBS protease with extensive homology to the C5a-ases of GAS and GBS as well as to other members of the subtilase family of serine proteases. All GBS serotypes that were examined bore the cspA gene. Mutation of cspA attenuated GBS virulence in a neonatal rat sepsis model and decreased resistance to phagocytosis. The protease is necessary for the cleavage of human fibrinogen. Understanding the contribution of this novel serine protease–like gene to the pathogenesis of GBS disease will broaden our understanding of how this significant pathogen causes severe infections of the newborn infant.

We thank Donald Chaffin, Jason Short, Clementien Vermont, and Peter Vickerman for their technical assistance with this project as well as Lakshmi Rajagopal for the accompanying scanning micrograph. We also thank Roland Siezen and Glen Tamura for their advice. This work was supported by NIH grants AI30068 (C. E. Rubens) and RO1AI 40918 (J.F. Bohnsack) and by National Research Service Award 5 F32 AI10426-03 (D. W. Shelver).

Theresa O. Harris and Daniel Shelver contributed equally to this work.

Conflict of interest: The authors have declared that no conflict of interest exists.

Nonstandard abbreviations used: group B streptococcus (GBS); cell-surface–associated protein (CspA); polymorphonuclear leukocytes (PMNs); group A streptococcus (GAS); Todd-Hewitt broth (THB); matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS); cell envelope–associated protease (CEP).

1. Baker, C.J., and Edwards, M.S. 1995. Group B Streptococcal infections. In Infectious diseases of the fetus and newborn infant. J.S. Remington and J.O. Klein, editors. W. B. Saunders. Philadelphia, Pennsylvania, USA. 980–1054.
2. Farley, MM, et al. A population-based assessment of invasive disease due to group B streptococcus in nonpregnant adults. N. Engl. J. Med. 1993. 328:1807-1811.
   View this article via: PubMed CrossRef Google Scholar
3. Ernst, J. Bacterial inhibition of phagocytosis. Cell. Microbiol. 2000. 2:379-386.
   View this article via: PubMed CrossRef Google Scholar
4. Dillard, JP, Vandeersea, MW, Yother, J. Characterization of the cassette containing genes for type 3 capsular polysaccharide biosynthesis in Streptococcus pneumoniae. J. Exp. Med. 1995. 181:973-983.
   View this article via: PubMed CrossRef Google Scholar
5. Cheng, Q, Finkel, D, Hostetter, MK. Novel purification scheme and functions for a C3-binding protein from Streptococcus pneumoniae. Biochemistry. 2000. 39:5450-5457.
   View this article via: PubMed CrossRef Google Scholar
6. Hancock, LE, Gilmore, MS. The capsular polysaccharide of Enterococcus faecalis and its relationship to other polysaccharides in the cell wall. Proc. Natl. Acad. Sci. USA. 2002. 99:1574-1579.
   View this article via: PubMed CrossRef Google Scholar
7. Ashbaugh, C, Warren, HB, Carey, VJ, Wessels, MR. Molecular analysis of the role of the group A streptococcal cystein protease, hyaluronic acid capsule, and M protein in a murine model of human invasive soft-tissue infection. J. Clin. Invest. 1998. 102:550-560.
   View this article via: JCI PubMed CrossRef Google Scholar
8. Wessels, MR, Moses, AE, Goldberg, JB, DiCesare, TJ. Hyaluronic acid capsule is a virulence factor for mucoid group A streptococci. Proc. Natl. Acad. Sci. USA. 1991. 88:8317-8321.
   View this article via: PubMed CrossRef Google Scholar
9. Rubens, CE, Wessels, MR, Heggen, LM, Kasper, DL. Transposon mutagenesis of type III group B Streptococcus: correlation of capsule expression with virulence. Proc. Natl. Acad. Sci. USA. 1987. 84:7208-7212.
   View this article via: PubMed CrossRef Google Scholar
10. Areschoug, T, Stalhammar-Carlemalm, M, Karlsson, I, Lindahl, G. Streptococcal β protein has separate binding sites for human factor H and IgA-Fc. J. Biol. Chem. 2002. 277:12642-12648.
    View this article via: PubMed CrossRef Google Scholar
11. Bohnsack, JF, et al. Purification of the proteinase from Group B Streptococci that inactivates human C5a. Biochim. Biophys. Acta. 1991. 1079:222-228.
    View this article via: PubMed Google Scholar
12. Fernandez-Espla, MD, Garault, P, Monnet, V, Rul, F. Streptococcus thermophilus cell wall-anchored proteinase: release, purification, and biochemical and genetic characterization. Appl. Environ. Microbiol. 2000. 66:4772-4778.
    View this article via: PubMed CrossRef Google Scholar
13. Martin, TR, Rubens, CE, Wilson, CB. Lung antibacterial defense mechanisms in infant and adult rats: implications for the pathogenesis of Group B Streptococcal infections in the neonatal lung. J. Infect. Dis. 1988. 157:91-100.
    View this article via: PubMed Google Scholar
14. Tamura, GS, Rubens, CE. Group B streptococci adhere to a variant of fibronectin attached to a solid phase. Mol. Microbiol. 1995. 15:581-589.
    View this article via: PubMed CrossRef Google Scholar
15. Chmouryguina, I, Suvorov, A, Ferrieri, P, Cleary, PP. Conservation of the C5a peptidase genes in Group A and B Streptococci. Infect. Immun. 1996. 64:2387-2390.
    View this article via: PubMed Google Scholar
16. Rubens, CE, Heggen, LM, Haft, RF, Wessels, MR. Identification of cpsD, a gene essential for type III capsule expression in Group B Streptococci. Mol. Microbiol. 1993. 8:843-855.
    View this article via: PubMed CrossRef Google Scholar
17. Sambrook, J.E., Fritsch, F., and Maniatis, T. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press. Cold Spring Harbor, New York, USA. Pp. 1659.
18. Yim, HH, Rubens, CE. Use of a dental amalgamator to extract RNA from the gram-positive bacterium Streptococcus agalactiae. Biotechniques. 1997. 23:229-231.
    View this article via: PubMed Google Scholar
19. Maguin, E, Duwat, P, Hege, T, Ehrlich, D, Gruss, A. New thermosensitive plasmid for gram-positive bacteria. J. Bacteriol. 1992. 174:5633-5638.
    View this article via: PubMed Google Scholar
20. Rubens, CE, Heggen, LM. Tn916 delta E: a Tn916 transposon derivative expressing erythromycin resistance. Plasmid. 1988. 20:137-142.
    View this article via: PubMed CrossRef Google Scholar
21. Framson, PE, Nittayajarn, A, Merry, J, Youngman, P, Rubens, CE. New genetic techniques for Group B Streptococci: high-efficiency transformation, maintenance of temperature-sensitive pWV01 plasmids, and mutagenesis with Tn917. Appl. Environ. Microbiol. 1997. 63:3539-3547.
    View this article via: PubMed Google Scholar
22. Yim, HH, Rubens, CE. Site-specific homologous recombination mutagenesis in group B streptococci. Methods. Cell. Sci. 1998. 20:13-20.
    View this article via: CrossRef Google Scholar
23. Chaffin, DO, Beres, SB, Yim, HH, Rubens, CE. The serotype of type Ia and III Group B Streptococci is determined by the polymerase gene within the polycistronic capsule operon. J. Bacteriol. 2000. 182:4466-4477.
    View this article via: PubMed CrossRef Google Scholar
24. Nizet, V, et al. Group B streptococcal beta-hemolysin expression is associated with injury of lung epithelial cells. Infect. Immun. 1996. 64:3818-3826.
    View this article via: PubMed Google Scholar
25. Richman, PG, Baer, H. A convenient plate assay for the quantitation of hyaluronidase in Hymenoptera venoms. Anal. Biochem. 1980. 109:376-381.
    View this article via: PubMed CrossRef Google Scholar
26. Ferrieri, P, Wannamaker, LW, Nelson, J. Localization and characterization of the hippuricase activity of group B streptococci. Infect. Immun. 1973. 7:747-752.
    View this article via: PubMed Google Scholar
27. Winram, SB, Tamura, GS, Rubens, CE. In vitro systems for investigating group B streptococcal host cell and extracellular matrix interactions. Methods. Cell. Sci. 1998. 20:191-201.
    View this article via: CrossRef Google Scholar
28. Bohnsack, JF, Mollison, KW, Buko, AM, Ashworth, JC, Hill, HR. Group B streptococci inactivate complement component C5a by enzymic cleavage at the C-terminus. Biochem. J. 1991. 273:635-640.
    View this article via: PubMed Google Scholar
29. Jones, AL, Knoll, KM, Rubens, CE. Identification of Streptococcus agalactiae virulence genes in the neonatal rat sepsis model using signature-tagged mutagenesis. Mol. Microbiol. 2000. 37:1444-1455.
    View this article via: PubMed CrossRef Google Scholar
30. Harlow, E., and Lane, D. 1988. Antibodies: a laboratory manual. Cold Spring Harbor Laboratory Press. Cold Spring Harbor, New York, USA. 68 pp.
31. Tanaka, T, Weisblum, B. Construction of a colicin E1-R factor composite plasmid in vitro: means for amplification of deoxyribonucleic acid. J. Bacteriol. 1975. 121:354-362.
    View this article via: PubMed Google Scholar
32. Bohnsack, JF, et al. Genetic polymorphisms of group B streptococcus scpB alter functional activity of a cell-associated peptidase that inactivates C5a. Infect. Immun. 2000. 68:5018-5025.
    View this article via: PubMed CrossRef Google Scholar
33. Baltimore, RS, Kasper, DL, Baker, CJ, Goroff, DK. Antigenic specificity of opsonophagocytic antibodies in rabbit anti- sera to Group B Streptococci. J. Immunol. 1977. 118:673-678.
    View this article via: PubMed Google Scholar
34. Ferretti, JJ, et al. Complete genome sequence of an M1 strain of Streptococcus pyogenes. Proc. Natl. Acad. Sci. USA. 2001. 98:4658-4663.
    View this article via: PubMed CrossRef Google Scholar
35. Chen, CC, Cleary, PP. Cloning and expression of the streptococcal C5a peptidase gene in Escherichia coli: linkage to the type 12 M protein gene. Infect. Immun. 1989. 57:1740-1745.
    View this article via: PubMed Google Scholar
36. Siezen, RJ. Multi-domain, cell-envelope proteinases of lactic acid bacteria. Antonie Van Leeuwenhoek. 1999. 76:139-155.
    View this article via: PubMed CrossRef Google Scholar
37. White, O, et al. Genome sequence of the radioresistant bacterium Deinococcus radiodurans R1. Science. 1999. 286:1571-1577.
    View this article via: PubMed CrossRef Google Scholar
38. Belitsky, BR, Gustafsson, MC, Sonenshein, AL, Von Wachenfeldt, C. An lrp-like gene of Bacillus subtilis involved in branched-chain amino acid transport. J. Bacteriol. 1997. 179:5448-5457.
    View this article via: PubMed Google Scholar
39. Stock, JB, Ninfa, AJ, Stock, AM. Protein phosphorylation and regulation of adaptive responses in bacteria. Microbiol. Rev. 1989. 53:450-490.
    View this article via: PubMed Google Scholar
40. von Heijne, G. A new method for predicting signal sequence cleavage sites. Nucleic. Acid. Res. 1986. 14:4683-4690.
    View this article via: PubMed CrossRef Google Scholar
41. Navarre, WW, Schneewind, O. Proteolytic cleavage and cell wall anchoring at the LPXTG motif of surface proteins in gram-positive bacteria. Mol. Microbiol. 1994. 14:115-121.
    View this article via: PubMed CrossRef Google Scholar
42. Doolittle, RF. Fibrinogen and fibrin. Annu. Rev. Biochem. 1984. 53:195-224.
    View this article via: PubMed Google Scholar
43. Cottrell, BA, Doolittle, RF. The amino acid sequence of a 27-residue peptide released from the alpha-chain carboxy-terminus during the plasmic digestion of human fibrinogen. Biochem. Biophys. Res. Commun. 1976. 71:754-761.
    View this article via: PubMed Google Scholar
44. Nizet, V., Ferrieri, P., and Rubens, C.E. 2000. Molecular pathogenesis of group B streptococcal disease in newborns. In Streptococcal infections: clinical aspects, microbiology, and molecular pathogenesis. Oxford University Press. New York, New York, USA. pp.180–221.
45. Lantz, MS. Are bacterial proteases important virulence factors? J. Periodontal Res. 1997. 32:126-132.
    View this article via: PubMed CrossRef Google Scholar
46. Projan, S.J., and Novick, R.P. 1997. The molecular basis of pathogenicity. Churchill Livingstone. New York, New York, USA. pp. 55–81.
47. Maeda, H, Yamamoto, T. Pathogenic mechanisms induced by microbial proteases in microbial infections. Biol. Chem. Hoppe. Seyler. 1996. 377:217-226.
    View this article via: PubMed Google Scholar
48. Lähteenmäkia, K, Pentti, K, Korhonen, TK. Bacterial plasminogen activators and receptors. FEMS Microbiol. Rev. 2001. 25:531-555.
    View this article via: PubMed Google Scholar
49. Gaillot, O, Pellegrini, E, Bregenholt, S, Nair, S, Berche, P. The ClpP serine protease is essential for the intracellular parasitism and virulence of Listeria monocytogenes. Mol. Microbiol. 2000. 35:1286-1294.
    View this article via: PubMed CrossRef Google Scholar
50. St. Geme (III), JW, Cutter, D. Evidence that surface fibrils expressed by Haemophilus influenzae type b promote attachment to human epithelial cells. Mol. Microbiol. 1995. 15:77-85.
    View this article via: PubMed CrossRef Google Scholar
51. Siezen, RJ, Leunissen, JA. Subtilases: the superfamily of subtilisin-like serine proteases. Protein. Sci. 1997. 6:501-523.
    View this article via: PubMed Google Scholar
52. Matsuka, Y, Pillai, S, Gubba, S, Musser, JM, Olmsted, SB. Fibrinogen cleavage by the Streptococcus pyogenes extracellular cysteine protease and generation of antibodies that inhibit proteolytic activity. Infect. Immun. 1999. 67:4326-4333.
    View this article via: PubMed Google Scholar
53. D’Souza, SE, Ginsberg, MH, Plow, EF. Arginyl-glycyl-aspartic acid (RGD): a cell adhesion motif. Trends. Biochem. Sci. 1991. 16:246-250.
    View this article via: PubMed CrossRef Google Scholar
54. Beckmann, C, Waggoner, JD, Harris, TO, Tamura, GS, Rubens, CE. Identification of novel adhesins from Group B Streptococci by use of phage display reveals that C5a peptidase mediates fibronectin binding. Infect. Immun. 2002. 70:2869-2876.
    View this article via: PubMed CrossRef Google Scholar