Staphylococcus aureus encodes a secreted von Willebrand factor-binding protein (vWbp) of 482 amino acids. The N-terminal part of this protein is homologous to staphylocoagulase and therefore we investigated whether vWbp has coagulating activity. Recombinant vWbp was shown to coagulate human and porcine plasma efficiently, but was less active against plasma from other species. The coagulation efficiency was concentration dependent, and could be inhibited by specific antibodies against vWbp. Furthermore, the species-specific coagulation by vWbp depended on the interaction with prothrombin. This interaction also resulted in specific cleavage of vWbp, releasing the C-terminal part from the coagulating domain.
Staphylococcus aureus is an important human pathogen responsible for a multitude of diseases ranging from trivial to life threatening. S. aureus encodes a broad spectrum of putative virulence factors consisting of many exoenzymes and toxins, and also cell wall-bound and secreted proteins that interact with host proteins in plasma and in the extracellular matrix.
Staphylocoagulase, or coagulase, is secreted by most strains of S. aureus, and coagulates blood and plasma from various animals. This enzyme activator, and its mode of action, has been extensively studied, and it is known to form an equimolar reactive complex with prothrombin [1,2]. The crystal structures of coagulase bound to human alpha-thrombin and prethrombin-2 have recently been solved . Coagulase does not cleave prothrombin into thrombin, which is the normal physiological activation mechanism. Instead, coagulase acts as a cofactor, which induces a conformational change in prothrombin, resulting in an active complex that can convert fibrinogen into fibrin. Thrombin is the key effector enzyme in the blood coagulation cascade, cleaving fibrinogen into fibrinopeptides and fibrin, which results in fibrin clotting.
The importance of coagulase in staphylococcal infections is not yet clear. Virulence studies with an isogenic S. aureus mutant lacking the coagulase gene have been performed in different animal models. In murine models of subcutaneous and intramammary infection, respectively, no diminished virulence was detected at 24 or 36 h post-infection . In a different mouse model, employing the same two strains, a possible role for coagulase in the development of blood-borne staphylococcal pneumonia was suggested, more specifically in promoting bacterial proliferation during later stages of the infection .
A secreted von Willebrand factor-binding protein (vWbp) from S. aureus was recently identified . This vWbp shows sequence homology to coagulase [6,7]. Here we demonstrate that the homology is also reflected in a conserved function. Thus, besides being a protein with von Willebrand factor-binding ability, vWbp is a coagulase.
Materials and methods
Proteins, plasma and reagents
Bovine fibrinogen, human prothrombin and plasma from chicken, mouse, pig and rat were from Sigma–Aldrich. Plasma from cow, goat, horse, rabbit and sheep were from the National Veterinary Institute, Sweden. Human plasma from different individuals was from Uppsala University Hospital, Sweden. Restriction enzymes and DNA modifying enzymes were from MBI Fermentas or Amersham Biosciences. Oligonucleotides were from Invitrogen.
Bacterial strains and growth conditions
Escherichia coli strain ER2566 (New England Biolabs) was grown in Luria–Bertani (LB)-broth or on LB-plates (LB with 1.5% agar) supplemented with 50 μg ml−1 ampicillin (LB-amp). S. aureus strains Newman and 8325-4 were grown in Tryptone Soya Broth (TSB, Oxoid) or on TSB-plates (TSB with 1.5% agar).
Recombinant proteins and antibodies
The amino acids (aa) corresponding to the full-length vWbp (aa 1–482) and coagulase (aa 1–611) were expressed using the IMPACT T7 expression system (New England Biolabs) according to the manufacturer's instructions. DNA was amplified with two DNA polymerase (Roche Diagnostics) using the primers vWbpFd (5′-AATATACCATGGTGGTTTCTGGGGAGAAGAAT-3′), vWbpRe (5′-TTTGCCATTATATACTTTATTGAT-3′), CoaFd (5′-AATTATCCATGGCGATAGTAACAAAGGATT-3′), CoaRe (5′-TTTTGTTACTCTAGGCCCAT-3′) and S. aureus strain Newman and 8325-4 DNA, respectively, as templates. The PCR products were digested with NcoI, treated with T4 polynucleotide kinase and ligated into the pTYB4 vector digested with NcoI and SmaI. After DNA-sequence confirmation, proteins were expressed in E. coli strain ER2566. The introduction of an extra N-terminal Ala-residue in coagulase was required due to cloning constraints (A-coagulase).
Four C-terminally truncated versions of vWbp were expressed and purified as above, comprising aa 1–405, aa 1–300, aa 1–250 and aa 1–200, respectively. For cloning, primer vWbpFd was used in combination with the different reverse primers: vWbp405Re (5′-TCTACCAGGTAAAGCAGACGT-3′), vWbp300Re (5′-TGTTTTCTTAATTTTTTGATTATCC-3′), vWbp250Re (5′-TGATTCATCACTTTTTGCTGCT-3′) and vWbp200Re (5′-TAAGTCTTCTTTTTTATTTTCTAAC-3′). To express aa 124–392 of vWbp, primers vWbp124Fd (5′-TTAATACCATGGCTAACCCTGAATTGAAAGACTT-3′) and vWbp392Re (5′-ATTATTATGCGTGTGATTTGAA-3′) were used.
Antibodies against vWbp and A-coagulase were generated in chickens by Agrisera, Vännäs, Sweden and purified on the respective protein immobilised on HiTrap columns (Amersham Biosciences).
Bactident Coagulase, a tube test for coagulation based on lyophilised rabbit plasma with EDTA, was purchased from MERCK Diagnostica. The tube coagulation assay was performed in small borosilicate glass tubes by mixing 190 μl plasma with 10 μl of recombinant protein at different concentrations. The fibrinogen-binding part of protein Fbe from Staphylococcus epidermidis expressed in E. coli was used as a negative control . The tubes were incubated at 37 °C, and the level of coagulation was observed by tilting the tubes. The test was regarded as positive if the tube content formed a coherent clot.
In the plate coagulation assay, adapted from Hwang et al. , 5 μl of recombinant protein in concentrations ranging from 40 to 2.5 μg ml−1, was added to small wells punched out in agarose plates (0.9%) containing 0.4% PEG 8000, 3 mg ml−1 bovine fibrinogen and 1% plasma (PFP-plates). Coagulation zones were measured after incubation at 37 °C overnight.
Inhibition of coagulation was carried out in 1.5 ml spectrophotometer cells. After mixing 0.5 ml of human plasma with 0.5 ml vWbp protein (in 0.9% NaCl) to a final concentration of 1 μg ml−1, coagulation was monitored at 405 nm. Affinity-purified antibodies against vWbp or, as a control, against A-coagulase, were added in different concentrations.
SDS–PAGE, mass spectrometry and N-terminal sequencing
Protein samples were prepared for gel electrophoresis by mixing equal volumes of protein solution and 2× sample buffer (1× sample buffer: 62.5 mM Tris–HCl pH 6.8, 10% glycerol, 2% SDS, 5%β-mercaptoethanol and 0.01% bromophenol blue). After boiling, the samples were analysed by SDS–PAGE using the PhastSystem (Amersham Biosciences) with PhastGel Gradient 8–25% gels. Protein bands were excised from a Coomassie Brilliant Blue stained SDS–PAGE gel. Peptide analysis was performed by H. Larsson (Department of Plant Biology, SLU), with electrospray ionisation mass spectrometry on a Q-Tof mass spectrometer and the Masslynx software (Micromass, Manchester, UK). N-terminal sequencing (Edman-degradation) of vWbp and A-coagulase was performed by M. Ståhlberg, PAC, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Stockholm, Sweden.
Results and discussion
Recombinant vWbp coagulates plasma
Part of the von Willebrand factor binding protein (vWbp) of S. aureus shows homology to coagulase (Fig. 1). To determine if vWbp had coagulating activity, vWbp and coagulase were expressed using the IMPACT system, where the C-terminal purification tag is removed by intein self-cleavage. During the course of this study, the importance of the N-terminal amino acids in coagulase was demonstrated by Friedrich et al. . Failure to remove the initiating Met, or removal of the following amino acid, resulted in proteins with lower activities, whereas removal of two amino acids almost completely abolished the coagulation ability. Therefore, the N-terminal amino acids of our recombinant proteins were determined, and in vWbp found to be identical to the N-terminal amino acids of the deduced native protein. Recombinant coagulase was obtained in two different sizes, of which the larger is identical to the native coagulase except for an additional N-terminal Ala-residue, while the smaller variant lacked the first 115 aa. Thus, the initiating Met was removed from both vWbp and coagulase after protein translation, and the additional Ala-residue in coagulase (A-coagulase) was due to cloning constraints. Therefore, the coagulating activity of A-coagulase may not reflect the full activity of coagulase.
The coagulating ability of vWbp and A-coagulase was assayed using the Bactident Coagulase test based on rabbit plasma. Both vWbp and A-coagulase showed coagulating activity, although roughly ten times more vWbp than A-coagulase was required to give the same level of coagulation (data not shown). Nevertheless, S. aureus produces two distinct proteins that cause plasma to coagulate.
Species specificity of vWbp
It is known that plasmas from different species differ markedly in their sensitivity to coagulase . When performing the tube coagulation assay, plasma from many different animal species coagulated in the presence of vWbp (10 μg ml−1), but with varying efficiency. Most interestingly, vWbp was a very efficient coagulator of human and porcine plasma, which solidified into a clot within a minute. Human plasma from different individuals coagulated with the same efficiency (data not shown). Plasma from goat and sheep coagulated in 10 min, equine plasma in 1 h, and bovine plasma in 6 h, while plasma from mouse, rabbit, rat and chicken did not coagulate within 18 h. These results suggested the existence of some species-specific determinant responsible for the differences in coagulation efficiency.
To study this, 1% human plasma was added to plasma from chicken, cow, horse, mouse and rabbit. This small addition of human plasma markedly reduced the vWbp-induced coagulation time to 10 min, except for chicken plasma, which did not coagulate within 2 h. Thus, some component in human plasma in combination with vWbp enhanced the coagulation rate for plasma from other species.
Concentration-dependent coagulation time
To investigate the coagulating ability of vWbp and A-coagulase further, we used different concentrations of both proteins to coagulate human and rabbit plasma, respectively (Table 1). The coagulation time depended on the concentrations of the recombinant proteins, and a higher concentration led to a shorter coagulation time over the range tested.
The table shows the time needed for the plasma to coagulate into a solid clot. N, no coagulation after 24 h. One representative experiment is shown.
Another method of determining coagulating activity of coagulase, is the plate coagulation assay . Here, the conversion of fibrinogen into a network of fibrin is seen as a milky precipitate in an agarose plate. Different amounts of vWbp were added to wells cut in PFP-plates containing either human, rabbit or horse plasma. After incubation, coagulation zones of different sizes were observed, depending of the concentration of vWbp (Fig. 2). This shows that fibrinogen is converted into fibrin by the action of vWbp in a concentration-dependent manner. As with the above results, vWbp caused the most pronounced coagulation when the plate contained human plasma, exhibiting coagulation zones of approximately twice the diameter of those in plates with horse or rabbit plasma (data not shown).
Inhibition of coagulation with specific antibodies against vWbp
By mixing human plasma and vWbp in a spectrophotometer cell, the vWbp-dependent coagulation of plasma could be followed by the increase in optical density over time (Fig. 3). The coagulation caused by vWbp could be inhibited in a dose-dependent manner by addition of antibodies against vWbp, but not by addition of antibodies against A-coagulase. Vice versa, coagulation induced by A-coagulase could be inhibited by anti-A-coagulase antibodies, but not by anti-vWbp antibodies (data not shown). These results further strengthen the observation that vWbp is a coagulase. Although vWbp and coagulase are homologous, the antibodies showed very little cross-reactivity in a Western blot (data not shown). Taken together, antibodies specifically developed against both vWbp and coagulase, are probably required to fully inactivate the coagulating ability of S. aureus.
The coagulation domain is located in the N-terminal part of vWbp
To determine the borders of the coagulating domain, we expressed truncated variants of the vWbp protein. These were assayed for their ability to coagulate human plasma (Fig. 4). The first 250 amino acids were found to be sufficient to induce coagulation, while the N-terminally truncated vWbp-protein did not have any apparent coagulating activity. These data are in agreement with the recent findings by Freidrich et al. , who showed the mechanism of prothrombin activation by coagulase and also proposed a domain organisation of vWbp based on homology to coagulase. The D1 domain of coagulase (aa 1–150) activates prothrombin by inserting the N-terminal amino acids into the activation pocket of prothrombin, while the D2 domain (aa 150–261) is required for docking with prothrombin. The corresponding D1 domain in vWbp is located between aa 1–131 and the D2 domain between aa 131–261.
vWbp interacts with prothrombin
In the presence of vWbp, some factor in human plasma enhanced coagulation of plasma from other species. It is also known that coagulase interacts with prothrombin. Therefore, we tested if prothrombin was the component that interacted with vWbp in a species-specific manner. The tube coagulation assay described above was performed with vWbp and the addition of human prothrombin, both at the final concentration of 10 μg ml−1. Plasma from chicken, cow, goat, horse, mouse, rabbit and sheep was used in this experiment. Plasma from all species, except chicken, now coagulated within 1 min. Addition of human prothrombin alone had no effect. Thus, the level of prothrombin activation determines the species-specific coagulation by vWbp. In a control experiment, bovine prothrombin was added to plasma from the species above, and as expected this did not affect the coagulating efficiency of vWbp (data not shown). Since S. aureus is not restricted to human infections, it would be interesting to study isolates from different species and investigate if the coagulating specificity is shifted towards the corresponding prothrombin.
Prothrombin is a conserved protein, but species-specific differences still exist, e.g. human and mouse prothrombin show 81% identity. It is generally accepted that pathogens and their host(s) are under a co-evolutionary pressure. This can lead to the emergence of virulence factors, the specificity of which reflects the host range of the bacteria. It was recently shown that streptokinase and staphylokinase showed a species-specific activation of plasminogen . The staphylokinase-activation of plasminogen dissolves blood clots and can be considered as the opposite to the clotting induced by vWbp. It is interesting that both these S. aureus proteins show a species-specificity producing the most pronounced activity with the corresponding human target molecule. This is especially interesting in view of the recent finding that expression of staphylokinase was inversely correlated to virulence of S. aureus in humans . Thus, isolates of S. aureus from nasal carriers, or non-lethal bacteraemia, had a higher staphylokinase-expression than S. aureus isolated from patients with lethal bacteraemia. It is tempting to speculate that staphylokinase-expression counteracts the coagulating effect of vWbp and coagulase.
vWbp and coagulase are cleaved in the presence of human prothrombin
On a SDS–PAGE gel, vWbp is mainly seen as a 66 kDa protein, although some degradation products are visible. Using antibodies against vWbp we confirmed that the minor proteins in the preparation are also derived from vWbp (data not shown). When mixed with human prothrombin, full-length vWbp is cleaved into two protein fragments, the smaller of ∼10 kDa (Fig. 5). Some full-length vWbp was still visible due to incomplete cleavage. The amount of prothrombin (∼5 ng) was too low to be detected. In all preparations of A-coagulase, two proteins of slightly different sizes are visible; the larger had the expected N-terminal sequence, but the smaller lacked the first 115 aa. When prothrombin and recombinant coagulase were mixed, both forms of coagulase were cleaved, resulting in a small protein fragment (∼15 kDa), and two larger fragments. This suggests that both forms of coagulase are cleaved at the same site. The cleavage of coagulase is in agreement with the observation of Kawabata et al. , who reported that coagulase is autodigested when mixed with human prothrombin. The molecular masses of the small fragments derived from vWbp and A-coagulase were determined using electrospray ionisation mass spectrometry. For A-coagulase, the measured molecular mass (14735 Da) correspond to cleavage of coagulase between aa 474 and 475 (P473R474–F475) to give a C-terminal fragment consisting of 136 aa with a calculated mass of 14724 Da. In vWbp, two cleavage sites were identified, one located between aa 405 and 406 (G404R405–K406) and one after Lys406, resulting in C-terminal fragments of 77 aa and 76 aa. The measured masses of these fragments are 9065 and 8935 Da, which is close to the calculated molecular masses of 9058 and 8930 Da. The identified cleavage sites are in agreement with the reported specificity of thrombin [13,14].
The observed cleavage of both A-coagulase and vWbp in the presence of human prothrombin is intriguing. In the case of coagulase, the fibrinogen-binding repeats are released, and for vWbp, the 77/76 C-terminal amino acids are released. This part of vWbp is very similar (68% identity) to a small (173 aa) hypothetical extra cellular protein (accession number BAB56976.1). It should be noted that the identified cleavage site in coagulase is found only in the first Fg-binding repeat in the publicly available genome sequences . Similarly, in the three variants of vWbp deduced from microbial genomes, the prothrombin cleavage site is conserved. It is possible that these cleavage sites exist by pure coincidence, but it appears more likely that cleavage of vWbp and coagulase is involved in regulation of coagulating activity, or in localisation of this activity. Alternatively, the released protein domains could have a function distinct from coagulation.
Most likely, S. aureus benefits from “short-circuiting” the host coagulation system, or proteins like vWbp and coagulase are unlikely to have evolved. It is interesting to note that both proteins have additional binding capacities, which are not required the coagulation. However, it seems likely that these functions are linked to coagulation, perhaps acting as a homing device, to direct coagulation to different locations in the vicinity of the bacteria.
The Swedish Foundation for Strategic Research (I&V Research Program, grant 24/98) and Biostapro AB supported this work.