INTRODUCTION

The present invention relates to a recombinant fibrinogen binding protein clumping factor A (CIfA) variant or fragment thereof wherein the ability to bind fibrinogen compared to the non-mutated protein or fragment thereof is altered. The present invention also relates to improved microbial antigen vaccines, pharmaceutical compositions, immunogenic compositions and antibodies comprising the recombinant fibrinogen binding protein clumping factor A (CIfA) variant or fragment thereof and their use in the treatment of microbial infections, particularly those of bacterial origin, including Staphylococcal origin.

Multiple drug resistance (MDR) is an increasing problem amongst gram positive bacteria, particularly in hospitals. The widespread use of antibiotics and other agents to treat bacterial infections has led to the rapid development of bacteria resistant to the agents and many bacteria have multiple drug resistance. Thus, there is now a need to provided improved therapies for dealing with such drug resistant infections.

Staphylococci are Gram-positive bacteria of spherical shape, usually arranged in grape-like irregular clusters. Some are members of the normal flora of the skin and mucous membranes of humans, others cause suppuration, abscess formation, a variety of pyogenic infections, and even fatal septicaemia. Pathogenic staphylococci often hemolyze blood, coagulate plasma, and produce a variety of extracellular enzymes and toxins.

The genus Staphylococcus has at least 30 species. The three main species of clinical importance are Staphylococcus aureus, Staphylococcus epidermidis, and Staphylococcus saprophytics. Staphylococcus aureus is coagulase-positive, which differentiates it from the other species. S. aureus is a major pathogen for humans. Almost every person has some type of S. aureus infection during a lifetime, ranging in severity from food poisoning or minor skin infections to severe life-threatening infections. The coagulase-negative staphylococci are normal human flora which sometimes cause infection, often associated with implanted devices, especially in very young, old and immunocompromised patients. Approximately 75% of the infections caused by coagulase-negative staphylococci are due to S. epidermidis. Infections due to Staphylococcus warneri, Staphylococcus hominis, and other species are less common.

S. saprophytics is a relatively common cause of urinary tract infections in young women. Staphylococci produce catalase, which differentiates them from the streptococci. S. lugdunensis is also relevant in a clinical and is present in approximately 5 to 10% of cases of infective endocarditis.

S. aureus colonization of the articular cartilage, of which collagen is a major component, within the joint space appears to be an important factor contributing to the development of septic arthritis. Hematogenously acquired bacterial arthritis remains a serious medical problem. This rapidly progressive and highly destructive joint disease is difficult to eradicate. Typically, less than 50% of the infected patients fail to recover without serious joint damage. S. aureus is the predominant pathogen isolated from adult patients with hematogenous and secondary osteomyelitis.

In hospitalized patients, Staphylococcus bacteria such as S. aureus are a major cause of infection. Initial localized infections of wounds or indwelling medical devices can lead to more serious invasive infections such as septicaemia, osteomyelitis, mastitis and endocarditis. In infections associated with medical devices, plastic and metal surfaces become coated with host plasma and matrix proteins such as fibrinogen and fibronectin shortly after implantation. This ability of S. aureus and other staphylococcal bacteria to adhere to these proteins is essential to the initiation of infection. Vascular grafts, intravenous catheters, artificial heart valves, and cardiac assist devices are thrombogenic and prone to bacterial colonization. Of the staphylococcal bacteria, S. aureus is generally the most damaging pathogen of such infections.

A significant increase in S. aureus isolates that exhibit resistance to most of the antibiotics currently available to treat infections has been observed in hospitals throughout the world. The development of penicillin to combat S. aureus was a major advance in infection control and treatment. Unfortunately, penicillin-resistant organisms quickly emerged and the need for new antibiotics was paramount. With the introduction of every new antibiotic, S. aureus has been able to counter with β-lactamases, altered penicillin-binding proteins, and mutated cell membrane proteins allowing the bacterium to persist. Consequently, methicillin-resistant S. aureus (MRSA) and multidrug resistant organisms have emerged and established major footholds in hospitals and nursing homes around the world (Chambers, H. F., Clin Microbiol Rev, 1 :173, 1988; and Mulligan, M.E., et al., Am J Med, 94:313, 1993). Today, almost half of the staphylococcal strains causing nosocomial infections are resistant to all antibiotics except vancomycin, and it appears to be only a matter of time before vancomycin will become ineffective as well. Thus, there remains a very strong and rapidly growing need for therapeutics to treat infections from staphylococci such as S. aureus which are effective against antibiotic resistant strains of the bacteria. In gram positive pathogens, such as Staphylococci, Streptococci and Enterococci, proteins, called adhesins, mediate such infections, for example by promoting colonization, attachment to blood clots and traumatized tissue. These specific microbial surface adhesins are termed MSCRAMMs (microbial surface components recognizing adhesive matrix molecules) (Patti, J., θt al., Ann Rev Microbiol, 48:585-617, 1994; Patti, J. and Hook, M., Cur Opin Cell Biol., 6:752- 758, 1994). MSCRAMMs specifically recognize and bind to extracellular matrix (ECM) components, such as fibronectin, fibrinogen, collagen, and elastin. These MSCRAMMs are found in many gram positive pathogens and their amino acid sequences are related, they have similar modular design and common binding domain organization. MSCRAMMs on the bacterial cell surface and ligands within the host tissue interact in a lock and key fashion resulting in the adherence of bacteria to the host. Adhesion is often required for bacterial survival and helps bacteria evade host defence mechanisms and antibiotic challenges. Once the bacteria have successfully adhered and colonized host tissues, their physiology is dramatically altered and damaging components such as toxins and enzymes are secreted. Moreover, the adherent bacteria often produce a biofilm and quickly become resistant to the killing effect of most antibiotics.

A bacterium can express MSCRAMMs that recognize a variety of matrix proteins. Ligand- binding sites in MSCRAMMs appear to be defined by relatively short contiguous stretches of amino acid sequences (motifs). Because a similar motif can be found in several different species of bacteria, it appears as though these functional motifs are subjected to interspecies transfer (Patti and Hook, Cur Opin Cell Biol, 6:752-758, 1994). In addition, a single MSCRAMM can sometimes bind several ECM ligands. MSCRAMMs can mediate infection by binding to proteins including Fibrinogen (Fg) and/or Fibronectin (Fn) etc. Fibrinogen and Fibronectin are proteins found in blood plasma and play key roles in hemostasis and coagulation.

Fibrinogen is composed of six polypeptide chains, two Aa, two Bβ and two γ-chains. The C- terminal part of the γ-chain is biologically important and interacts with the platelet integrin during platelet adherence and aggregation. It is this region which is also targeted by Staphylococcus aureus resulting in Fibrinogen-dependant cell clumping and tissue adherence.

Staphylococcus aureus has several surface expressed proteins which stimulate platelet activation and aggregation. The Staphylococcus aureus MSCRAMM proteins include but are not limited to the following:

Fibrinogen binding protein clumping factor A (CIfA);

Fibrinogen binding protein clumping factor B (CIfB);

Fibronectin-fibrinogen binding protein A (FnBPA);

Fibronectin-fibrinogen binding protein B (FnBPB); and

S. aureus surface proteins SasA, SasG, SasK etc.

Table 1 below outlines a selection of various Staphylococcus aureus cell wall-anchored surface proteins.

TABLE 1

Surface protein aa ^a Ligand(s) ⁶ Motif Sortase ^d

Protein A (Spa) 508 Immunoglobulin, von Willebrand LPETG A

Factor, TNFR"

Fibronectin binding protein A 1 ,018 Fibronectin, fibrinogen, elastin LPETG A (FnbpA)

Fibronectin binding protein B 914 Fibronectin, fibrinogen, elastin LPETG A (FnbpB)

Clumping factor A (CIfA) 933 Fibrinogen.complement factor I LPDTG A

Clumping factor B (CIfB) 913 Fibrinogen, cytokeratin 10 LPETG A

Collagen adhesion (Cna) 1 ,183 Collagen LPKTG A

SdrC 947 Unknown LPETG A

SdrD 1 ,315 Unknown LPETG A

SdrE 1 ,166 Unknown LPETG A

PIs 1 ,637 Unknown LPDTG A

SasA 2,261 Unknown LPDTG A

SasB 937 Unknown LPDTG A

SasC 2,186 Unknown LPNTG A

SasD 241 Unknown LPAAG A SasE/lsdA 354 Heme' LPKTG A

SasF 637 Unknown LPKAG A

SasG/Aap 1 ,117 Unknown ⁹ LPKTG A

SasH 308 Unknown LPKTG A

Sasl/HarA/lsdH 895 Haptoglobin LPKTG A

SasJ/lsdB 645 Hemoglobin, heme LPQTG A

SasK 211 Unknown LPKTG A

IsdC 227 Heme NPQTN B a aa, protein length in amino acids.

⁶ Molecular component(s) recognized and bound by protein.

⁰ Consensus motif recognized by sortase and present in C-terminal cell wall sorting signal.

d Sortase for which cell wall surface protein is substrate.

⁸ TNFR, tumor necrosis factor receptor

'also binds to proteins in desquamated epithelial cell. Promotes resistance to bactericidal lipids and lactoferrin

⁹ also binds to desquamated nasal epithelial cells. Involved in biofilm formation.

Clumping factor A (CIfA) was the first Fibrinogen γ-chain-binding S. aureus adhesin identified. Clumping factor A (ClfA)is a surface located protein of Staphylococcus aureus. CIfA is an important virulence factor of S. aureus. It contributes to the pathogenesis of septic arthritis and endocarditis. CIfA contains a 520 amino acid N-terminal A domain (the Fibrinogen Binding Region), which comprises three separately folded subdomains N1 , N2 and N3. The A domain is followed by a serine-aspartate dipeptide repeat region and a cell wall- and membrane-spanning region, which contains the LPDTG-motif for sortase-promoted anchoring to the cell wall. CIfA is present in practically all S. aureus strains (Peacock SJ, Moore CE, Justice A, Kantzanou M, Story L, Mackie K, O ^' Neill G, Day NPJ (2002) Virulent combinations of adhesin and toxin genes in natural populations of Staphylococcus aureus. Infect lmmun 70:4987-4996). It binds to the C- terminus of the γ-chain of fibrinogen, and is thereby able to induce clumping of bacteria in fibrinogen solution (McDevitt D, Nanavaty T, House-Pompeo K, Bell E, Turner N, McEntire L, Foster T, Hook M (1997) Characterization of the interaction between the Staphylococcus aureus clumping factor (CIfA) and fibrinogen. Eur J Biochem 247:416-424 and McDevitt D, Francois P, Vaudaux P, Foster TJ (1994) Molecular characterization of the clumping factor (fibrinogen receptor) of Staphylococcus aureus. MoI Microbiol 11 :237-248). 3D Structural analysis of CIfA and the related fibrinogen-binding proteins SdrG and CIfB has revealed that the ligand-binding A domain in all these related proteins are all composed of three subdomains N1 , N2 and N3, with residues 221-559 corresponding to Regions N2-N3 being the smallest truncate that retains the ability to bind fibrinogen. It has been found that amino acid residues 532 to 538 correspond to the latching peptide region of CIfA. Each subdomain comprises nine /3-strands that form a novel IgG-type fold. The fibrinogen /-chain peptide- binding site in these proteins is located in a hydrophobic groove at the junction between N2 and N3. It has been found that there is significant structural similarity between the 3d structure of these proteins, this is due to one or more of related amino acid sequence, similar modular design and common binding domain organization.

Expression of CIfA on S. aureus hampers phagocytosis by both macrophages and neutrophils (Palmqvist N, Patti JM, Tarkowski A, Josefsson E (2004) Expression of staphylococcal clumping factor A impedes macrophage phagocytosis. Microb Infect 6:188-195 and Higgins J, Loughman A, van Kessel KPM, van Strijp JAG, Foster TJ (2006) Clumping factor A of Staphylococcus aureus inhibits phagocytosis by human polymorphonuclear leukocytes. FEMS Microbiol Lett 258:290-296). In neutrophils this is due to both a fibrinogen-dependent mechanism and to a fibhnogen-independent mechanism. In contrast, platelets are activated by bacteria expressing CIfA through its interaction with GPIIb/llla leading to aggregation. This is most efficiently executed when fibrinogen is present, but there is also a fibrinogen-independent pathway for platelet activation (Loughman A, Fitzgerald JR, Brennan MP, Higgins J, Downer R, Cox D, Foster TJ (2005) Roles of fibrinogen, immunoglobulin and complement in platelet activation promoted by Staphylococcus aureus clumping factor A. MoI Microbiol 57:804-818 and O ^' Brien L, Kerrigan SW, Kaw G., Hogan M., Penades J., Litt D., Fitzgerald DJ. , Foster T.J. & Cox D. (2002) Multiple mechanisms for the activation of human platelet aggregation by Staphylococcus aureus: roles for the clumping factors CIfA and CIfB, the serine-aspartate repeat protein SdrE and protein A. MoI Microbiol 44, 1033-1044).

CIfA is a virulence factor for induction of septic arthritis in mice (Josefsson E., Hartford O., O'Brien L, Patti JM, Foster T (2001) Protection against experimental Staphylococcus aureus arthritis by vaccination with clumping factor A, a novel virulence determinant. J Infect Dis

184:1572-1580). In addition, elimination of CIfA together with another fibrinogen binding protein

CIfB protected against systemic inflammation at the early stages of infection (Palmqvist N,

Foster T, Fitzgerald R, Josefsson E, Tarkowski A (2005) Fibronectin-binding proteins and fibrinogen-binding clumping factors play distinct roles in staphylococcal arthritis and systemic inf lammation. J lnf Dis 191 :791 -798).

The Staphylococcus aureus fibrinogen binding protein CIfA has been isolated and characterized and is the subject of, for example, US Patent Nos. 6, 008, 341 and 6,177,084. CIfA and CIfB have an identical structural (3D) organization and approximately 27% amino acid identity. FnBPA has an approximately 25% amino acid identity to CIfA.

At present there are no MSCRAMM based vaccines approved and on the market. Veronate®, a donor-selected staphylococcal human immune globulin intravenous (IGIV) targeting CIfA and SdrG, performed poorly in phase III clinical trials and was withdrawn from trials. It is currently being re-evaluated to determine whether it is a viable treatment for Staphylococcal infections.

Thus, in view of the prevalence of multiple drug resistance in gram positive bacteria and the lack of successful therapies and vaccines for these multi-drug resistant bacteria, alternative therapies which can deal with such bacterial infections without using antibiotics will be of significant value.

Furthermore, an improvement in efficacy over any known treatments or vaccines will be of particular importance, especially in a clinical setting.

Thus, the present invention is directed to providing an improved recombinant fibrinogen binding protein clumping factor A (CIfA), ideally for use in the therapy of bacterial infections.

STATEMENT OF THE INVENTION

According to a first aspect of the invention, there is provided a recombinant fibrinogen binding protein clumping factor A (CIfA) or fragment thereof comprising at least part of the fibrinogen binding region, wherein amino acid residue D ₃₂₁ or an amino acid adjacent to D ₃₂₁ is mutated and the recombinant protein has reduced ability or lacks the ability to non-covalently bind fibrinogen compared to the non-mutated protein or fragment thereof.

According to a second aspect of the invention, there is provided a recombinant fibrinogen binding protein clumping factor A (CIfA) or fragment thereof according to the invention for use in therapy. According to a third aspect of the invention, there is provided a recombinant fibrinogen binding protein, or fragment thereof, of the invention for use in the treatment or prophylaxis of microbial infections, including sepsis, septic arthritis and/or endocarditis, preferably caused by Staphylocci. According to a fourth aspect of the invention, there is provided the use of the recombinant protein, or fragment thereof, of the invention in the manufacture of a medicament for the treatment or prophylaxis of a microbial infection, preferably caused by Staphylocci.

According to a fifth aspect of the invention, there is provided a method of inducing an immune response in a patient comprising administering to the patient a recombinant protein, or fragment thereof, of the invention.

According to a sixth aspect of the invention, there is provided a method of treating a patient having a microbial infection comprising administering a recombinant protein or fragment thereof, or vaccine comprising the recombinant protein or fragment thereof of the invention to a patient in need thereof.

According to a seventh aspect of the invention, there is provided a nucleic acid construct, fusion protein, expression vector or host cell expressing the recombinant protein, or fragment thereof, of the invention.

According to a eighth aspect of the invention, there is provided a vaccine comprising the recombinant protein, or fragment thereof, of the invention. According to a ninth aspect of the invention, there is provided an antibody raised against the recombinant protein, or fragment thereof, of the invention preferably in the form of a hyperimmune serum.

According to a tenth aspect of the invention, there is provided an immunogenic pharmaceutical composition comprising the recombinant protein, or fragment thereof, of the invention, and a pharmaceutically acceptable adjuvant.

According to an eleventh aspect of the invention, there is provided a process for making an immunogenic composition comprising the step of adding the pharmaceutically acceptable excipient to the recombinant protein, or fragment thereof, of the invention. DETAILED DESCRIPTION

In this specification, the terms "adhesin", "MSCRAMM" and "cell-wall anchored proteins" will be understood to be interchangeable and cover all microbial derived ligand binding proteins. Ideally, these proteins bind fibrinogen, heme or haemoglobin, haptoglobin-haemoglobin, haemin, collagen and other such ligands. The term "MSCRAMM-like" proteins are intended to cover proteins or adhesins which have related amino acid sequences, similar modular design and/or common/similar binding domain organization to such MSCRAMM proteins, such as lsd proteins. Ideally, the MSCRAMM-like proteins have similar binding domain organization/modular design. Additionally, the MSCRAMM-like proteins may have at least 50%, preferably 60%, preferably 75%, more preferably 85%, even more preferably 95%, still more preferably 99% or more amino acid sequence identity with the MSCRAMM proteins. Any reference herein to MSCRAMM will be understood to be interchangeable with recombinant fibrinogen binding protein clumping factor A (CIfA).

It will also be understood that any of the percentage identities or homologies referred to in the specification are determined using available conventional methods over the entire/whole length of the sequence. The term "micro-organism", "microbe", "microbial" or the like includes but is not limited to organisms including bacteria, fungi, viruses, yeasts and/or moulds.

The term "immunologically effective amount" covers those amounts which are capable of stimulating a B cell and/or T cell response.

It will be understood that the improved recombinant fibrinogen binding protein clumping factor A or fragment thereof of the invention has reduced ability or lacks the ability to non-covalently bind fibrinogen compared to the wild type or non-mutated protein. It will also be understood that the recombinant fibrinogen binding protein clumping factor A fragment has substantially the same immunogenic activity as the full length protein. In this manner the fragment is capable of eliciting an immune response similar to that of the full length protein. It will also be understood that recombinant fibrinogen binding protein clumping factor A or fragment thereof is mutated or altered i.e. undergoes a nucleotide or amino acid substitution, insertion, deletion or addition at the residue of interest or adjacent to the residue of interest. The term "adjacent" is intended to cover an amino acid residue immediately adjacent or at a distance of 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acids either side (N-or C-terminal) to the amino acid of interest.

According to a first aspect of the invention, the MSCRAMM protein of the invention is a recombinant fibrinogen binding protein clumping factor A (CIfA) or fragment thereof comprising at least part of the fibrinogen binding region, wherein amino acid residue D _32I is substituted with tyrosine (D ₃₂ iY) and the recombinant protein has reduced ability or lacks the ability to non- covalently bind fibrinogen, for use in therapy.

Fibrinogen binding can be readily measured using conventional techniques used in the .Examples. We have found that the recombinant fibrinogen binding protein clumping factor A (CIfA) variant or fragment thereof of the invention has an altered ability to bind fibrinogen compared to a non-mutated or wild-type protein or fragment thereof.

It has been established that amino acid residues 221 to 559, covering the N2 and N3 regions, of CIfA play an important part in the binding to fibrinogen and are the minimal fibrinogen binding region. We have unexpectedly found that mutation of amino acid residues in this region results in an expressed protein which can be recognized by the host immune defences but lacks fibrinogen binding and hence, reduces the associated virulence. This region (the 339 amino acid fibrinogen binding domain) of CIfA has a specific 3D structure, a so-called DE-variant IgG fold, and is the minimum Fg-binding truncate which if altered (via substitution or deletion etc) can provide an improved therapy.

The alteration to result in the loss of fibrinogen binding activity may take place by substitution, addition or insertion or deletion at either the nucleotide or amino acid level. Ideally, the substitution negatively affects the 3D structure (e.g. of the so-called DE-variant IgG fold) of the protein or fragment so it can no longer bind fibrinogen. Ideally, the nucleotide or amino acid substitution reduces the non-covalent interaction with fibrinogen, preferably by preventing ligand binding to the hydrophobic pocket separating N2 and N3 of Region A of the fibrinogen binding protein. Alternatively, the latching peptide region corresponding to amino acids 532 to 538 may be altered by substitution or deleted to prevent ligand binding. Additionally, a truncate/fragment lacking the latching peptide region and optionally the remainder of the C- terminal protein residues, i.e. lacking amino acid residues 532 to 559, may be used. The fibrinogen /-chain peptide-binding site is located in a hydrophobic groove at the junction between N2 and N3 of CIfA. Thus, the substitutions or deletions mentioned above are designed to alter the MSCRAMM protein-ligand interaction and prevent the non-covalent binding of CIfA to fibrinogen.

Thus, we have found that by altering the recombinant fibrinogen binding protein clumping factor A (CIfA) or fragment thereof in this manner, it is possible to provide a ligand binding protein without the ability to bind its ligand, which stimulates a greater immune response upon immunization than the wild type protein. This reduces systemic inflammation, thereby decreasing microbial virulence. Consequently, this altered ligand binding CIfA or ClfA-like protein which lacks the ability to bind its ligand can be advantageously used in the treatment of microbial infections. Additionally, we have surprisingly found that the alteration of the specific amino acid residue D ₃₂ i, particularly the substitution with tyrosine, provides a more stable protein compared to other protein variants. As shown in EΞxample 4 the CIfA D _32I YP ₃₃₆ SY ₃₃₈ A (triple mutant) protein is easier to purify than CIfA P ₃₃₆ SY ₃₃₈ A without the D ₃₂ i mutation. These findings present a new and valuable vaccine/immunization therapeutic against bacterial infections which provides better results when compared to a vaccine or immunization therapeutic derived from the wild type protein.

CIfA is a 993 amino acid protein, comprising a 520 amino acid fibrinogen binding domain (from amino acids 40 to 559). This fibrinogen binding domain is the N Terminal A domain comprising subregions N1 , N2 and N3. It will be understood that the entire fibrinogen region spanning N1 to N3 from amino acid 40 to amino acid 559 may be used in the invention. Alternatively, a truncate of the N1 to N3 region may be used, e.g. 221 to 559 (the minimal fibrinogen binding region), 221 to 531 (the minimal fibrinogen region without the latching peptide and following residues) etc. Ideally, subregions N2 and N3, the minimal fibrinogen binding region, may be used which correspond to amino acid residues 221 to 559. Alternatively, a fragment of these subregions may be used.

It is established that the first step in binding of an MSCRAMM to its ligand involves a non- covalent interaction via the DLL model. These are the primary non-covalent MSCRAMM interactions with the ligand. The final stages in MSCRAMM-ligand binding involve covalent interactions. The DLL model was elucidated from the 3D structure of SdrG in complex with its ligand. CIfA has now been shown to act by a minor variation of the DLL mechanism (Ganech et al (2008) "A structural model of the Staphylococcus aureus Clfa-fibrinogen interaction opens new avenues for the design of anti-staphylococcal therapeutics". PIoS Pathog 4(11); e1000226). The DLL model specifically relates to the non-covalent interactions involved in ligand binding.

We have found that the recombinant protein, or fragment thereof of the invention has reduced or lacks the ability to non-covalently bind its host ligand due to altered dock, lock and latching. One or more of the dock, lock or latching steps may be altered. Thus, we postulate that, according to the invention, that the non-covalent binding that takes place during binding, via Dock, Lock and Latching (DLL), of the MSCRAMM or MSCRAMM-like protein (i.e. CIfA) to its ligand may be reduced or prevented.

In relation to MSCRAMMs ClfA/ClfB in particular, it has been found that the minimal ligand binding domain comprises Region A subregions N1 to N3, specifically subregions N2 and N3 which comprise a variant Dev-lgG Ig fold. The variant Dev-lgG Ig fold is new variant of the immunoglobulin motif also called the DE-variant. It is postulated that a hydrophobic pocket formed between the two DEv-IgG domains of ClfA/B is the ligand-binding site for the fibrinogen γ-chain. Essentially, the ligand binds to the hydrophobic groove separating N2 and N3. Specifically, during ligand binding the unfolded peptide component of the ligand inserts into the groove located between the N2 and N3 subdomains. The latching peptide at the C-terminus of subdomain N3 undergoes a conformational change and inserts between two beta strands in subdomain N2, thus, locking the ligand in place. Indeed, mutagenic substitution of residues Tyr256, Pro336, Tyr338 and Lys389 in the clumping factor, which are proposed to contact the terminal residues ⁴⁰⁸ AGDV ⁴¹¹ of the fibrinogen y -chain, resulted in proteins with no or markedly reduced affinity for fibrinogen. It will be understood that the complete ligand binding protein, the ligand binding domain, the minimal ligand binding domain or a fragment thereof may be used. The use of truncated proteins of the ligand binding protein such as the ligand binding domain, the minimal ligand binding domain, or the use of fragments thereof is advantageous for ease of manufacture and overcoming other problems such as unwanted cleavage of the protein. For example, the latching peptide, present in the minimal ligand binding domain, may be deleted/removed or altered. For example, the latching peptide in CIfA corresponds to Region A amino acids 532 to 538 and in CIfB to Region A amino acids 530-540 (Walsh et al (2004) JBC 279(49): 50691- 50699). These residues may be altered, substituted or removed/deleted in order to prevent the ligand binding to the MSCRAMM via DLL. In this way the DLL "latching" of the MSCRAMM to its ligand is prevented. This "latching" occurs by way of a non-covalant interaction. In one embodiment, the latching peptide is removed in its entirety along with the remaining Region A C-terminal amino acid residues. According to another embodiment, the latching peptide region only is removed. According to yet another embodiment, the latching peptide region undergoes amino acid substitution to result in the reduction or prevention of ligand binding/latching.

Furthermore, such alterations in the ligand binding domain may take place at the amino acid level, by amino acid substitution or deletion, using either the full length protein, ligand binding domain, minimal ligand binding domain or fragment thereof. It will be understood that proteins or fragments thereof with sufficiently high homology to the ligand binding protein may also be used. High homology as defined herein occurs when at least 50%, preferably 60%, preferably 70%, preferably 80%, more preferably 90%, even more preferably 95%, still more preferably 95% to 99%, still more preferably 99% or more of the nucleotides or match over the entire length of the DNA sequence or when used in connection with amino acid sequences when the amino acid sequences are not identical but produce a protein having the same functionality and activity. It will be understood that these comments about high homology may also relate to the 3D structure of the protein, i.e. modular binding domain organization.

It will be understood that whilst these teachings relate to CIfA in particular, they are equally applicable ClfA-like proteins, which have similar modular binding domain organization and bind ligands in similar ways.

It will be understood that the recombinant fibrinogen binding protein may optionally further comprise one or more nucleotide or amino acid residue substitutions, point mutations, insertions, deletions or additions in the fibrinogen binding region.

According to a further embodiment of this aspect of the invention, the recombinant fibrinogen binding protein CIfA may comprise further amino acid substitutions wherein amino acid residue P ₃₃₆ and/or Y ₃₃₈ of the fibrinogen binding Region A is substituted with either serine or alanine to result in (CIfAP ₃₃₆ S Y ₃₃₈ A or 1"CIfAP ₃₃₆ A Y ₃₃₈ S. Thθ choice of residues was based on the X-ray crystal structure of CIfA and the observation that individual changes to the proline or the tyrosine reduced binding affinity. Surprisingly, we found that this mutant CIfA protein (rClfA D ₃₂₁ Y P ₃₃₆ S Y ₃₃₈ A and rClf A P ₃₃₆ A Y ₃₃₈ S) stimulated an immune response and can be used in the generation of a much more effective vaccine or antibody therapy. This substitution may take place in the full length fibrinogen binding protein, the fibrinogen binding region, the minimal fibrinogen binding region, or a fragment thereof.

Advantageously, the recombinant fibrinogen binding protein clumping factor A (CIfA), or a fragment thereof comprising at least part of the fibrinogen binding region, comprises the amino acid substitutions wherein residue D ₃₂₁ is substituted with tyrosine, P ₃₃₆ and/or Y ₃₃₈ are substituted with either serine and/or alanine.

The following sequences outlined in the table below may be used in accordance with the invention.

¹ Additional N residues (N-terminal extension (6 x His tag and additional residues) comprise 6 His residues, followed by GIy and Ser. Additional C terminal residues comprise Lys followed by Leu (other additional N and C terminal residues may be used - depending on the primer used or N/C terminal tags required)

² Additional N residues (6 x His tag and additional residues) comprise 6 His residues, followed by GIy and Ser. Additional C terminal residues comprise Arg followed by Ser (other additional N and C terminal residues may be used - depending on the primer used or N/C terminal tags required))

³ without the latching peptide corresponding to aa residues 532 to 538 and remainder A Region C-terminal residues i.e. lacking amino acid residues 532 to 559.

Ideally, the recombinant Staphylococcal fibrinogen binding protein comprises the amino acid sequence according to any of SEQ ID Nos. 1 to 3 wherein residue D ₃₂₁ is substituted with tyrosine (D ₃ 21Y) and optionally P ₃₃₆ and/or Y ₃₃₈ are substituted with either serine and/or alanine, or a fragment thereof.

Alternatively, the fragment of the recombinant Staphylococcal fibrinogen binding protein comprises the amino acid sequence according to any of SEQ ID No. 4 to SEQ ID No. 11. SEQ ID NOs 4 and 5 correspond to the CIfA A domain N1 , N2, N3 only, rClfA D ₃₂₁ Y P ₃₃₆ S Y ₃₃₈ A and rClfA P ₃₃₆ A Y ₃₃₈ S respectively as outlined in the table above.

It is also postulated, based on the substitutions in the latch which were made in SdrG, that substitutions in the latch that are defective in the conformational change or beta strand complementation willl also be defective in ligand binding. Thus, ideally, the substitutions are in amino acid residues 532 to 538 which correspond to the latching peptide and affect the ability of the peptide to undergo conformational change, or bind the ligand or both. Alternatively, the alteration may comprise removing the amino acid residues 532 to 538 (delta latch peptide) altogether, to give similar results. Additionally, a C-terminal truncation mutant lacking amino acid residues 532 to 559 (including the latching peptide residues) will also effect binding to the ligand. However, it will also be contemplated that other amino acid residues could be substituted other than those specifically recited above. For example, GIu 526, VaI 527, Tyr 256 and Lys 389 may be substituted to alter the fibrinogen binding properties of the protein or fragment thereof. Thus, any substitution which reduces binding ability may be contemplated. Ideally, such substitutions or deletions effect the hydrophobic pocket and associated mechanism for binding the ligand in the hydrophobic trench such as homologues Val527 in CIfA and N526 in CIfB. In CIfB, Q235 and N526 have been studied to shown to reduce binding. A similar study was done with FnBPA where N304 and F306 were shown to be important for Fg binding. Thus, mutations in these amino acid residues will affect ligand binding. Thus, according to a still further embodiment of this aspect of the invention, the recombinant fibrinogen binding protein may further comprise the amino acid substitution wherein amino acid residues Ala254, Tyr256, Pro336, Tyr338, Ne387, Lys389, Glu526 and/or Val527 are substituted with either Ala or Ser.

Alternatively, the alteration may be in the form of a deletion, comprising the fibrinogen binding region without the latching peptide sequence (amino acids 532 to 538), to result in a recombinant fibrinogen binding protein without the ability to non-covalently bind fibrinogen. In this embodiment, amino acid residues 221 to 531 of Region A of CIfA are used, which lack the latching peptide and following C-terminal residues. Alternatively, an amino acid substitution in the latching peptide amino acids 532 to 538 which prevents the DLL of the fibrinogen may be contemplated.

It is understood that all proteins in the Clf-Sdr family binds ligands by the DLL model. By modelling the 3D structure, it is possible to predict the latching peptide and make a truncate that lacks it, either in the full length (N1 to N3) or the minimal ligand binding truncate N2-N3, or a fragment thereof.

We found that these substitution rClfA proteins (whether deletion mutants, substitutions or truncates) reduced virulence and disease outcome, and surprisingly induced less systemic inflammation that the wild type protein.

Thus, immunization with these mutant proteins is expected, based on the proteins tested, to enhance the level of antibodies which recognized both the mutant and wild type protein and to provide for a greater immune response than the wild type protein. It will be understood that the recombinant Staphylococcal fibrinogen binding protein, or fragment thereof, of the invention may be used in therapy, specifically in the treatment of microbial infections, preferably Staphylococci infections such as in the treatment of sepsis, septic arthritis and/or endocarditis or other similar conditions or disease states. The fibrinogen binding region of the protein is altered so that it no longer binds fibrinogen. As stated above, the alteration may take place at the nucleotide or amino acid level. It will be understood that proteins or fragments thereof with sufficiently high homology to the fibrinogen binding protein may also be used. High homology as defined herein occurs when at least 50%, preferably 60%, preferably 70%, preferably 80%, more preferably 90%, even more preferably 95%, still more preferably 95% to 99%, still more preferably 99% of the nucleotides match over the entire length of the DNA sequence or when used in connection with amino acid sequences when the amino acid sequences are not identical but produce a protein having the same functionality and activity. It will be understood that these comments about high homology may also relate to the 3D structure of the protein.

It will be understood that the complete fibrinogen binding protein, the fibrinogen binding region, the minimal fibrinogen binding region, or a fragment thereof may be used. The use of truncated proteins or fragments thereof is advantageous for ease of manufacture and overcoming other problems such as unwanted cleavage of the protein.

Such fragments should ideally comprise at least part of the fibrinogen binding region of the MSCRAMM. The advantages of using a truncated protein or fragment thereof of the, comprising for example one or more subdomains of the ligand-fibrinogen binding region only, relate to the ability to purify the protein at high yields without degradation.

The CIfA protein fibrinogen binding region, otherwise referred to as the A Region, comprises 3 subregions, N1 , N2 and N3. Thus, the immunogenic fragment may comprise subregions N1 , N2 and/or N3 of the CIfA A Region or a fragment thereof. Thus, for example, in relation to CIfA, the fragment may comprise one or more of subdomains of Region A, N1 , N2 or N3. Ideally, N2 and N3 may be used as this truncate is less likely to undergo proteolysis (a protease cleavage site has been reported between N1 and N2 in CIfA and CIfB) and can be expressed at higher levels in E. coli. N2 and N3 are the minimal fibrinogen binding region of CIf proteins.

We have unexpectedly found that this altered fibrinogen binding protein, truncate or fragment thereof, without the ability to bind fibrinogen stimulates a greater immune response upon immunization than the wild type protein which binds to fibrinogen in the normal manner. Advantageously, this altered fibrinogen binding protein does not provoke systemic inflammation when expressed by S. aureus, thus, microbial virulence is decreased. Consequently, this altered protein which lacks the ability to bind fibrinogen can be advantageously used in the treatment of microbial infections. We have also found contrary to expectations that the protection effect of the altered fibrinogen binding protein is greater than the wild type protein. We have found that a pharmaceutical composition or vaccine comprising such an altered recombinant protein is more effective than a pharmaceutical composition or vaccine comprising the same recombinant protein in an unaltered (wild type) form, such as CIfA, CIfB, SdrG etc.

Thus, these findings present a new and valuable vaccine/immunization therapeutic against bacterial infections which provides better results when compared to the wild type protein when also used as a vaccine/immunization therapeutic. It will be understood that the altered protein of the invention, may be used in the generation of antibodies, including monoclonal, polyclonal, chimeric, humanized antibodies or fragments thereof, for use in the treatment of such microbial infections. Compositions may then be provided which include such antibodies, such as a hyperimmune serum, and these compositions may be used in the treatment of patients infected with Staphylococcus infections.

Thus, the proteins or active fragments thereof may be used to inhibit the binding of Staphylococci to the extra-cellular matrix (ECM) and to prevent/treat Staphylococci infections in a patient. Furthermore, the proteins or active fragments thereof, and antibodies to the proteins are useful in the treatment of infections from Staphylococcal infections, for the development of vaccines for active or passive vaccination, and when administered as a pharmaceutical composition to a wound or a medical device, both the proteins and antibodies are useful as blocking agents to prevent microbial infection. For example, these proteins or fragments thereof may be used in active vaccines, and the antibodies to these proteins in passive vaccines.

These vaccines and products described herein present a significant improvement over the prior art, which teaches the general use of MSCRAMMs to impart immunization, but does not teach the unexpected and improved vaccines or products described herein. The preparation of proteins, DNA and antibodies are well known in the art and will not be described in detail herein. Conventional techniques are ideally used in the generation of these molecules. The invention will also be understood to cover nucleic acid constructs containing the nucleic acid or amino acid sequence of interest, recombinant host cells containing such nucleic acid constructs to express the protein of interest, and immunogenic compositions.

For administration, the protein composition may be dispersed in a sterile, isotonic saline solution or other pharmaceutically acceptable adjuvant. It will be understood that the vaccine may be a DNA or protein vaccine.

Immunization may take place by the injection of DNA, protein or antibodies. Alternatively, an attenuated live organism that includes and expresses the DNA may be administered. The amount of DNA, protein or antibodies that may be administered will depend on several mitigating factors, including dependence on the promoter strength, protein expression and immunogenicity of the expressed gene. These may be altered for each new application to obtain the desired immunologically effective amount required. According to another embodiment of this invention, there is provided a method of inducing an immune response in an individual and/or treating a patient having a microbial infection, comprising administering to the individual a recombinant Staphylococcal fibrinogen binding protein, or fragment thereof comprising at least the fibrinogen binding region, without the ability to bind fibrinogen.

According to further preferred embodiment of the invention, there is provided a vaccine comprising a recombinant Staphylococcal fibrinogen binding protein, or fragment thereof comprising at least part of the fibrinogen binding region, without the ability to bind fibrinogen. According to a still further preferred embodiment of the invention, there is provided an antibody raised against a recombinant Staphylococcal fibrinogen binding protein, or fragment thereof comprising at least part of the fibrinogen binding region, without the ability to bind fibrinogen, preferably in the form of a hyperimmune serum. According to a yet further preferred embodiment of the invention, there is provided an immunogenic pharmaceutical composition comprising a recombinant Staphylococcal fibrinogen binding protein, or fragment thereof comprising at least part of the fibrinogen binding region, without the ability to bind fibrinogen and a pharmaceutically acceptable adjuvant.

Ideally, the recombinant Staphylococcal fibrinogen binding protein or fragment thereof is derived from S. aureus, S. epidermidis and/or S. lugdunensis.

In the specification, the terms "comprise, comprises, comprised and comprising" or any variation thereof and the terms "include, includes, included and including" "consist, consists, consisted and consisting" or any variation thereof are considered to be totally interchangeable and they should all be afforded the widest possible interpretation.

The invention is not limited to the embodiment hereinbefore described, but may be varied in both construction and detail within the scope of the claims.

The present invention will now be described with reference to the following non-limiting figures and examples. Figures 1 to 15 show the results of Example 1.

Figure 1 shows the severity of arthritis (A), measured as arthritic index, and weight loss (B) in mice inoculated with S. aureus strain Newman, and clfAPYl, clfAPYU, and elf A null mutants. 3.2 x 10 ⁶ - 6.0 x 10 ⁶ cf u of S. aureus strains were inoculated. Data are presented as medians

(squares or center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Data from three experiments are pooled. Λ^ _e w _m an = 27 - 30, Λ/ _C/MPYI = 30, N _c n _APγn = 10, and Nα, _A = 16 -

20.

Figure 2 shows the bacterial growth in kidneys in mice 7-8 days after inoculation with 3.2 x

10 ⁶ - 6.0 x 10 ⁶ cfu of S. aureus strain Newman, and clfAPYi, clfAPΥW, and elf A null mutants. Data are presented as cfu per kidney pair. Where no growth was detectable, the count was put to highest possible count according to what dilution was used. Data from three experiments are pooled. Λtawman = 26, NdMPYi = 30, Nci _f A _PY n = 10, and N _M = 15.

Figure 3 shows the survival of mice after inoculation with 5.2, 5.1 or 3.3 x 10 ⁷ cfu of

S.aureus strain Newman, clfAPYi mutant or elf A null mutant, respectively. N= 10 per group from start.

Figure 4 shows the survival of mice after inoculation with 9.4, 7.9, 10.7 or 9.8 x 10 ⁶ cfu of

S.aureus strain LS-1 , and clfAPYi, clfAPYU or elf A null mutants, respectively. N = 15 per group from start. Figure 5 shows the survival of mice immunized with BSA, recombinant CIfA or recombinant CIfAPY (i.e. CIfAPYI recombinant protein A domain) and inoculated with 2.3 x 10 ⁷ cfu of S. aureus Newman. N= 15 per group from start.

Figure 6 shows the frequency of arthritic mice inoculated with 3.2 x 10 ⁶ - 6.0 x 10 ⁶ cfu of S. aureus strain Newman wild-type, and elf APW, clfAPYW, and clfA null mutants. Data from three experiments are pooled. /Newman = 27 - 30, /V _c/MPY ι = 30, N _mPΥΑ = 10, and NM = 16 - 20.

Figure 7 shows the severity of arthritis measured as arthritic index in mice inoculated with 5.2, 5.1 or 3.3 x 10 ⁷ cfu of S.aureus strain Newman wild-type, elf APW mutant or elf A null mutant, respectively. Data are presented as medians (squares), interquartile ranges (boxes), and 80% central ranges (whiskers). Λ^ _e wman = 0-10, Λ/ _C/ MPYI = 9-10, and Λ/ _c/M = 0-10.

Figure 8 shows the weight loss in mice inoculated with 5.2, 5.1 or 3.3 x 10 ⁷ cfu of S.aureus strain Newman wild-type, clfAPW mutant or clfA null mutant, respectively. Data are presented as medians (center line), interquartile ranges (boxes), and 80% central ranges (whiskers). Mj _θw m _a n = 0-10, /V _dMPY i = 9-10, and N _cHA = 0-10.

Figure 9 shows the severity of arthritis measured as arthritic index in mice immunized with

BSA, recombinant CIfA or recombinant CIfAPY (i.e. CIfAPYI recombinant protein A domain) and inoculated with 4.0 x 10 ⁶ cfu of S. aureus Newman. Data are presented as medians (squares), interquartile ranges (boxes), and 80% central ranges (whiskers). Λfe _SA = 14, Λ/ _C/ MPY = 14, and Λ/ _c/M = 15 per group from start.

Figure 10 gives the nucleotide and amino acid sequence of wild-type CIfA A domain protein

(rClfA), domains N123 only, with the residues highlighted which are altered in the following examples, D ₃₂₁ , P ₃₃₆ and Y ₃₃₈ to give rise to rClfAPYI and P ₃₃₆ and Y ₃₃₈ rClfAPYII (SEQ ID No.3).

It is this recombinant protein A domain which was used in vaccination in the following examples.

Figure 11 shows an illustrative representation of the structure of FnBPA, CIfB, CIfA and SdrG proteins. Region A is the fibrinogen binding region, S is the signal sequence, W is the cell wall spanning domain, M is the membrane anchor including the LPXTG motif, + represent positively charged residues and R is the repeat region. In CIfA Region A comprises N123 (not shown). The BCD region of FnBPA (and the shorter CD region of FnBPB - not shown) binds fibronectin.

Figure 12 shows the specific antibody responses to recombinant ClfAPY40-559 in serum samples of mice immunized with bovine serum albumin (BSA), recombinant ClfA40-559 (rClfA), or recombinant ClfAPY40-559 (rClfAPY), 9 days after the second booster immunization, which was one day before infection with 2.3 x 10 ⁷ cfu/mouse of S. aureus strain Newman wildtype for induction of sepsis. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Λ/ _BSA = 13-15, /V^ _A = 15, and N^ _fAPY = 15. Figure 13 shows the specific antibody responses to recombinant ClfA40-559 in serum samples of mice immunized with bovine serum albumin (BSA), recombinant CIf A40-559 (rClfA), or recombinant ClfAPY40-559 (rClfAPY), 9 days after the second booster immunization, which was one day before infection with 2.3 x 10 ⁷ cfu/mouse of S. aureus strain Newman wildtype for induction of sepsis. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Λfes _A = 13-15, /V^HA = 15, and N^ _mPY = 15.

Figure 14 shows the specific antibody responses to recombinant ClfAPY40-559 in serum samples of mice immunized with bovine serum albumin (BSA), recombinant ClfA40-559 (rClfA), or recombinant ClfAPY40-559 (rClfAPY), 9 days after the second booster immunization, which was one day before infection with 4.0 x 10 ⁶ cfu/mouse of S. aureus strain Newman wildtype for induction of septic arthritis. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Λ/ _BSA = 14-15, Λ/rcitA = 15, and N^APY = 15.

Figure 15 shows the specific antibody responses to recombinant ClfA40-559 in serum samples of mice immunized with bovine serum albumin (BSA), recombinant ClfA40-559 (rClfA), or recombinant ClfAPY40-559 (rClfAPY), 9 days after the second booster immunization, which was one day before infection with 4.0 x 10 ⁶ cfu/mouse of S. aureus strain Newman wildtype for induction of septic arthritis. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Λ/ _B S _A = 14-15, Λtci _fA = 15, and Λ/ICH _A PY = 15.

Figure 16 of Example 2 shows the specific antibody responses to recombinant ClfAPY221-559 in serum samples of mice immunized with bovine serum albumin (BSA), recombinant ClfA221-559 (rClfA221-559), or recombinant ClfAPY221-559 (rClfAPY221-559), 9 days after the second booster immunization. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Λ/ _B S _A = 15, Λ/,αfA22i-559 = 14-15, and /V _rC ifAPY22i-559 = 14-15.

Figure 17 of Example 2 shows the specific antibody responses to recombinant ClfA221-

559 in serum samples of mice immunized with bovine serum albumin (BSA), recombinant ClfA221 -559 (rClfA221 -559), or recombinant ClfAPY221-559 ( rClf AP Y221-559), 9 days after the second booster immunization. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). Λ/BSA = 15, = 14-15, and Λ/fC»APY221-559 = 14-15.

Figure 18 of Example 3 shows the specific antibody responses to recombinant ClfA221-531 in serum samples of mice immunized with recombinant ClfAPY221-531 (rClfAPY221-531 ), 9 days after the second booster immunization. Data are presented as medians (center lines), interquartile ranges (boxes), and 80% central ranges (whiskers). /Vrcif _A22 i- ₅₃ i ⁼ 14-15.

Figure 19 gives the nucleotide and amino acid sequence of CIfA ₄₀ - _SS g D321Y/P336S/Y338A (CIfAPYI).

Figure 20 gives the nucleotide and amino acid sequence of CIfA ₄₀ - _SS g D321Y/P336S/Y338 A including N- and C-terminal extension (CIfAPYI).

Figure 21 shows the results of Comparative Example 1.

Examples

Example 1

rClf A A region truncates comprising N1 , N2 and N3 (amino acids 40-559)

MATERIAL AND METHODS

Full details of the numeric references in brackets given in the Examples are provided at the end of this section. Mice

NMRI mice were obtained from Scanbur BK (Sollentuna, Sweden) and were maintained in the animal facility of the Department of Rheumatology, University of Gδteborg, Sweden. Gόteborg animal experiment ethical board approved the experiments. They were housed up to 10 animals per cage with a 12 h light-dark cycle, and were fed standard laboratory chow and water ad libitum. The animals were 6 to 16 weeks old at the start of the experiments.

Bacterial strains

For infection of animals the S. aureus wildtype strains Newman (14) and LS-1 (11) and constructed derivatives thereof were used. The elf A D ₃₂₁ YP ₃₃₆ SY ₃₃₈ A (clfAPYl) and elf A Pa _S eAY ₃₃₈ S (clfAPYW) derivatives were constructed in strain Newman and transduced to strain LS-1 (see below). The deletion mutants Newman clfA2::Tn917 mutant DU5876 (3) and LS-1 clfA2::Tn917 mutant (J. R. Fitzgerald et al., unpublished) were also used. Bacteria were grown on blood agar plates for 48 h, harvested, and kept frozen at -2O ⁰ C in PBS containing 5% (wt/vol) BSA (Sigma Chemicals) and 10% (vol/vol) dimethyl sulfoxide. Before injection into animals, the bacterial suspensions were thawed, washed in PBS, and adjusted to appropriate cell concentrations. The number of viable bacteria was measured in conjunction with each challenge by cultivation on blood agar plates and counting colonies.

Construction of elf APYl and elf APYIl mutations in S. aureus Newman and LS-1

In this experiment, a full length CIfA A region truncate, comprising N1 , N2 and N3, corresponding to amino acids 40 to 559, was used. In the following description and figures :

- CIfA may also be referred to as rClfA 40-559 (SEQ ID NO 3);

- CIfA D _32I YP ₃₃₆ SY ₃₃₈ A may also be referred to as clfAPYI, rclfAPY or rclfAPYI (i.e clfAPYI 40-559) (SEQ ID NO 4) (triple mutant); and

- CIfA P ₃₃₆ AY ₃₃₈ S may also be referred to as clfAPYII, rclfAPYII (i.e. clfAPYII 40-559) (SEQ ID NO 5) (double mutant).

A 1.02 kb Psfl-SamHI fragment of pCF77 PY (Loughman et al., 2005) containing the mutations P ₃₃₆ S and Y ₃₃₈ A in clfA was cloned into pBluescriptll SK- (Stratagene). This plasmid was linearised with HindiW and ligated to /-//πcΛII-cut pTSβrmC (J. Higgins, unpublished) to generate plasmid pARM, which is a temperature sensitive E. coli-S. aureus shuttle vector containing the P ₃₃₆ S and Y ₃₃₈ A substitutions. We generated a double mutant, in which the order of the substitutions was reversed, yielding P ₃₃₆ A and Y ₃₃₈ S. To generate this a plasmid pJH2, analogous to pARM but containing the P ₃₃₆ A and Y ₃₃₈ S subsitutions, was generated. Overlap primer PCR was used with the same flanking primers used to make pCF77 PY (6), and a different pair of overlapping mutagenic primers: F3: GCAACTTTGACCATGGCCGCTTCTATTGACCCTGAAAATG and

R3: CATTTTCAGGGTCAATAGAAGCGGCCATGGTCAAAGTTGC

(mutations in bold and underlined) to generate pCF77 PYII. The 1.02 kb Pstl-Hindlll fragment of this plasmid was used as described above to generate pJH2, a temperature sensitive E colf-S. aureus shuttle vector containing the P ₃₃₆ A and Y ₃₃₈ S substitutions.

Both pARM and pJH2 were transferred to RN4220 (15) by electroporation and subsequently transduced using phage 85 (16) to S. aureus Newman (14) and LS-1 (11). In these strains the plasmids were induced to insert into the chromosome and then excise, leaving the mutations in the chromosome of a proportion of transformants, generating Newman clfAPYl, Newman elf APYW, LS-1 clfAPYl and LS-1 clfAPYW. Transformants were screened for loss of the plasmid and a loss of fibrinogen-binding activity. Integrity of the clfA gene was verified by Southern hybridisation using a clfA probe (data not shown). Expression of an immunoreactive protein (CIfAPY) was verified by Western immunoblotting using anti-ClfA region A polyclonal rabbit antiserum (data not shown). The mutations were verified by PCR across the KprA-BamH\ fragments from genomic DNA and commercial sequencing of the products. The about 700 bases of the elf A gene of strain LS- 1 that were sequenced were identical to the corresponding bases in the Newman elf A gene of strain Newman. Production of Recombinant CIfA and CIfAPY

His-tagged recombinant CIfA region A, domains N123 (amino acids 40-559), was produced from pCF40 as previously described (17), with an additional polishing step through an anion-exchange column. Plasmid pCF77 PY (6) was used as template to clone c/MPYI domains N123 into pQE30 to generate pCF40PY. Using this plasmid, recombinant CIfAPY was also produced by nickel affinity chromatography and anion exchange chromatograpy, as was described for rClfA. Eluates were dialysed against two changes of PBS before concentration and freeze-drying.

Septic arthritis and sepsis experiments

In experiments 1-3 all the mice (n=10 per group) were infected with strain Newman to trigger arthritis. In experiments 4 and 5, the mice were infected with strain Newman and LS-1 , respectively, to induce sepsis (n=10 per group).

-Experiment 1 Mice were infected by intravenous injection with 3.5 x 10 ⁶ cfu/mouse of S.aureus strain Newman or with 4.3 x 10 ⁶ cfu/mouse of Newman c/MPYI mutant, both in 200 μl PBS. Clinical arthritis and weight change was followed until day 7. Mice were sacrificed day 8, kidney growth of bacteria were assessed and serum IL-6 and total IgG levels were measured. Synovitis and bone destruction was studied histologically on the joints of fore and hind legs.

Experiment 2 Mice were infected with 5.0 x 10 ⁶ cfu, 6.0 x 10 ⁶ cfu or 4.3 x 10 ⁶ cfu of S.aureus strain Newman, clfAPY\ mutant or Newman clfA::Erm ^R (clfA null mutant), respectively. Clinical arthritis and weight change was followed until day 7. Mice were sacrificed day 7, kidney growth of bacteria were assessed and serum IL-6 and total IgG levels were measured. Synovitis and bone destruction was studied histologically on the joints of fore and hind legs. Experiment 3 Mice were infected with 4.7 x 10 ⁶ cfu, 3.2 x 10 ⁶ cfu, 3.9 x 10 ⁶ cfu or 4.8 x 10 ⁶ cfu of S.aureus strain Newman, clfAPY\ mutant, Newman elf APYW mutant or Newman elf A null mutant, respectively. Clinical arthritis and weight change was followed until day 7. Mice were sacrificed day 8 and kidney growth of bacteria were assessed. The outcome of the experiments 1-3 were very similar, so data were pooled and presented together.

In Experiment 4 mice were injected intravenously with 5.2 x 10 ⁷ cfu, 5.1 x 10 ⁷ cfu or 3.3 x 10 ⁷ cfu of S.aureus strain Newman, clfAPYl mutant or elf A null mutant, respectively. Mortality, weight change and clinical arthritis were followed until day 10.

In Experiment 5 mice were infected with 9.4 x 10 ⁶ cfu, 7.9 x 10 ⁶ cfu, 10.7 x 10 ⁶ cfu or 9.8 x 10 ⁶ cfu of S.aureus strain LS-1 , LS-1 clfAPYl mutant, LS-1 clfAPYW mutant, or LS-1 elf A null mutant, respectively. Mortality, clinical arthritis and weight change was followed until day 16.

Intra-articular injection of bacteria

One knee joint per mouse was injected with 2.4 x 10 ⁴ cfu, 2.4 x 10 ⁴ cfu, or 3.4 x 10 ⁴ cfu of strain Newman wildtype, clfAPYl mutant or elf A knockout mutant, respectively, in 20 μl PBS. N = 10 per group. Mice were sacrificed 3 days later, and the knee joints were collected for histopathological examination.

Vaccination with wild-type and mutant recombinant CIfA

Purified rClfA40-559, rClfAPY40-559 (i.e. rClfAPYI) or BSA were dissolved in physiologic saline and emulsified 1:1 in Freund ^' s complete adjuvant (Difco Laboratories). Two hundred μl of the emulsion containing 30 μg (= 0.53 nmol) of protein was injected subcutaneously (s.c.) on day 0. First booster immunization with 30 μg of protein in physiologic saline in incomplete Freund ^' s adjuvant was performed on day 11. Second booster immunization was done day 21. On day 30 the mice were bled and sera were frozen for later analysis of antibody responses. On day 31 , 14-15 mice per group were infected by i.v. injection of 4.0 x 10 ⁶ cfu/mouse for induction of septic arthritis, or by 2.3 x 10 ⁷ cfu/mouse for induction of sepsis. Clinical arthritis, weight change and mortality were followed for 11 and 15 days, respectively. Bacterial growth in kidneys was assessed in the septic arthritis experiment. Clinical evaluation of infected mice

The clinical evaluation was performed in a blinded manner. Each limb was inspected visually. The inspection yielded a score of 0 to 3 (0, no swelling and erythema; 1 , mild swelling and/or erythema; 2, moderate swelling and/or erythema; 3 marked swelling and/or erythema). The arthritic index was constructed by adding the scores from all four limbs of an animal. The overall condition of each mouse was also examined by assessing signs of systemic inflammation, i.e., weight decrease, reduced alertness, and ruffled coat. In cases of severe systemic infection, when a mouse was judged too ill to survive another 24 h, it was killed by cervical dislocation and considered dead due to sepsis. Histological examination

Histological examination of joints was performed using a modification (8) of a previously described method (18).

Bacterioloqic examination of infected kidneys

Kidneys were aseptically dissected, kept on ice, homogenised, serially diluted in PBS and spread on blood agar plates. After 24 h of incubation in 37°C the number of cfu per kidney pair was determined.

Measurement of serum IqG

Levels in serum of total IgG were measured by the radial immunodiffusion technique (19). Goat- Anti-Mouse-lgG and mouse IgG standard were purchased from Southern Biotech, Birmingham, AL.

Specific antibodies - ELISA

Serum samples from immunized mice were obtained 9 days after the second booster immunization. The serum specific antibody response against rClfA and rClfAPY was measured by ELISA. Microplates (96-well; Nunc) were coated with 5 μg/ml of recombinant protein in PBS. Blocking agent, serum samples, biotinylated antibodies, and ExtrAvidin-proxidase were all diluted in PBS. The assay was run according to a previous description (8). All serum samples were diluted 1 :20000, and antibody response was monitored as absorbance at 405 nm.

In a second run, to get a more accurate measure of the specific antibody responses in the different immunization groups, the responses were determined at several serum dilutions. Thus, all serum samples were diluted 1 :5000, 1 :20000, 1 :80000 and 1 :320000, and antibody response was monitored as absorbance at 405 nm.

IL-6 analysis - Serum IL-6 was detected by a method previously described (20).

Statistical analysis

Statistical evaluation was done by using the Mann-Whitney U test.. P<0.05 was considered to be significant. Data are reported as medians, interquartile ranges, and 80% central ranges, unless otherwise mentioned.

RESULTS

Exchange of amino acids necessary for CIfA binding to fibrinogen hampers development of septic arthritis and sepsis

Amino acids (D321 , P336 and/or Y338) that are known to be required for fibrinogen binding by CIfA were altered by allelic exchange to create mutants of strains Newman and LS1 that expressed a non-fibrinogen-binding CIfA protein on the cell surface. The level of expression and integrity of the protein was measured by Western blotting which established that there was good expression of the mutant proteins on the bacterial surface and expressed protein was the right size. The ability of Newman wild-type and Newman elf A D _32I Y, P ₃₃₆ S Y ₃₃₈ A (clfAPY\) to provoke septic arthritis was investigated. Septic arthritis was induced by intravenous inoculation of 3.5 x 10 ⁶ to 5.0 x 10 ⁶ colony-forming units (cfu) and 3.2 x 10 ⁶ to 6.0 x 10 ⁶ cfu of Newman wild-type and the clfAPYl mutant, respectively. The development of arthritis was studied clinically for 7 days. The clfAPY\ mutant provoked significantly less severe arthritis than the wild-type strain over the entire experimental period (P > 0.001 , Fig. 1 A). The frequency of arthritis was lower for Newman clfAPYl at most time points (Fig. 6).

Unexpectedly, it appears that the new amino acid composition in the ClfAPYl molecule fits for interaction with a host anti-bacterial defence. To check for this possibility, a new construct was made where different amino acids were substituted for P336 and Y338 (clfA P ₃₃₆ A Y ₃₃₈ S: clfAPYW). Mice that were inoculated with 3.9 x 10 ⁶ cfu of Newman clfAPYll developed arthritis to the same low extent as the clfAPYl mutant (Fig. 1 A), and with a similar frequency (Fig. 6). This outcome suggests strongly that the loss of fibrinogen binding is responsible for the reduced level of arthritis.

It is possible that CIfA is involved in the development of arthritis by mechanisms that do not involve fibrinogen binding. To test this a CIfA deletion mutant lacking the CIfA protein was compared to mutants expressing the modified non-fibrinogen binding CIfA protein. However, mice that were infected with 4.3 x 10 ⁶ to 4.8 x 10 ⁶ cfu of elf A null mutant developed arthritis in a manner not different from the clfAPYl and elf APYU mutant infected mice (Fig. 1 A). The frequency of arthritis was also indistinguishable (Fig. 6).

Infected joints were also investigated histologically. The synovitis in Newman c/MPYI-infected mice was significantly milder than in wild-type infected mice in both experiment 1 and 2 (P = 0.02 and 0.001 , respectively). Bone destruction, a major cause of sequels in human septic arthritis, was almost absent in the Newman c/MPYI-infected samples (EΞxperiment 2, P = 0.001). The synovitis and bone destruction induced by the Newman clfA null mutant were also less pronounced compared to mice infected with Newman wild-type (P = 0.003 and 0.006, respectively), but somewhat more severe than in the Newman clfAPYi group, although not significantly so.

Next, the metabolic consequences of the clfA mutations for the infectious process were analysed. Mice infected with the Newman wild-type strain lost up to about 30% of their body weight during the experimental period. Mice that were infected with the fibrinogen binding-deficient mutants Newman clfAPYi and Newman elf APYW lost hardly any weight at all (P > 0.0001 versus wild-type). In contrast, the Newman clfA null mutant had an intermediate effect on the weight loss, causing significantly less than the wild-type strain, but significantly more than the clfAPYi and elf APYW mutant strains (P≤ 0.02 in most cases, Fig. 1 B). The serum levels of IL-6, a measure of systemic inflammatory response, were analyzed at day 7- 8 of infection. The pattern of IL-6 expression was similar to weight changes. Newman wild-type evoked high levels of serum IL-6 (4.8 (2.8, 5.7) ng/ml), the Newman clfAPYi mutant evoked considerably lower IL-6 (0.2 (0.07, 2.4) ng/ml, P < 0.0001) while the Newman clfA null mutant gave rise to an intermediate response (2.5 (1.3, 3.2) ng/ml) with significant differences to both the wild-type and clfAPY\ mutant group (P = 0.009 and P = 0.008, respectively) (median, interquartile range).

The growth of bacteria in kidneys was significantly greater in Newman wild-type-infected mice, compared to both of the Newman elf APY mutants and the Newman clfA null mutant (P < 0.0001 , P= 0.011 , and P= 0.005, respectively; Figure 2). The Newman clfA null mutant-infected mice had significantly more bacterial growth in kidneys than Newman c/WPYI-infected mice (P = 0.0005, Fig. 2).

Total IgG in sera was measured in mice on day 7-8 of infection. There was a significantly lower increase of IgG levels in both the Newman clfAPYi- and Newman clfA null mutant-infected groups as compared to mice infected with the wild-type strain (3.1 (1.2, 4.9); 2.3 (1.0, 2.6); and 6.4 (5.0, 11.0), respectively (median, interquartile range); P≤ 0.0003). There were no significant differences between the two mutant groups. The mortality was 17% in the Newman wild type-infected mice, 0% in the Newman elf APW and clfAPYW mutant groups and 30% in the Newman clfA null mutant group. There were significant differences in mortality between the wild-type and the clfAPYl groups, and between the clfAPY\ and elf A null mutant groups (P < 0.05 and P< 0.01 , respectively). It appears that direct and indirect signs of systemic inflammation are lower in mice infected with S.aureus expressing CIfA that is deficient in fibrinogen binding. Unexpectedly, the strain which lacked CIfA expression altogether induced more systemic inflammation than a CIfAPY mutant- expressing strain. Sepsis was induced in mice by increasing the inoculation dose of S. aureus. Mice were infected with 5.2 x 10 ⁷ cf u of Newman wild type, 5.1 x 10 ⁷ cfu of the Newman clfAPY\ mutant and 3.3 x 10 ⁷ cfu of the Newman clfA null mutant. Within 5 days all wild-type infected mice were dead, but only one C/Λ4PYI mutant mouse out of ten were dead after 10 days of infection (P < 0.0001 , Fig. 3). Mice infected with the clfA null mutant also survived a significantly shorter time than the clfAPY\ mutant- infected mice (P < 0.0001 , Fig. 3). In this experiment the mice challenged with the clfA null mutant developed significantly more arthritis than the clfAPYl mutant group, while at the same time they lost significantly more weight (Fig. 7 and 8). Thus, by analogy with the measures of systemic inflammation in the septic arthritis experiments, the survival of the mice is prolonged if the CIfA molecule is expressed, as long as it lacks fibrinogen binding properties.

Injection of bacteria into joints

To test if the inflammatory reaction in the joint is dependent on fibrinogen binding, Newman wild- type, Newman clfAPY\ or Newman clfA null were injected directly into a knee joint of mice, thereby by-passing the systemic compartment. Synovitis, including polymorphonuclear infiltration of the joint cavity, and bone destruction was studied by histology 3 days later. The mice received 2.4 x 10 ⁴ cfu of wild-type, 2.4 x 10 ⁴ cfu of the clfA null mutant, or 3.4 x 10 ⁴ cfu of clfAPY\ mutant in one knee. The synovitis and the polymorphonuclear infiltration histologic index in the joint cavity was 0.25 (0, 3.0) for knees infected with wild-type, 2.38 (0.25, 3.0) for the clfA null mutant and 0.25 (0, 0.25) for the clfAPY\ mutant (median, interquartile range). The histologic index for destruction of bone was 0 (0, 1.0) for wild-type, 1.0 (0, 1.0) for the clfA null mutant, and 0 (0, 0) for the elf APYl mutant (median, interquartile range; P = 0.01 between the clfAPYl mutant and the elf A null mutant). Since the clfAPYl mutant evoked very little synovitis and destruction, despite the fact that 42% more of that strain was given to mice than the other strains, it is concluded that CIfA- promoted fibrinogen binding is needed for the maximal inflammatory response within the joint. Again, the absence of CIfA expression enhanced inflammation compared to the fibrinogen binding deficient CIfA mutant.

PY mutation in strain LS-1

To determine if the ability of CIfA to bind fibrinogen affects virulence of other strains of S.aureus, the clfAPYl, clfAPYU and elf A null mutations were transduced to the TSST-1 expressing S. aureus strain LS-1. Mice were challenged with 9.4 x 10 ⁶ cfu of LS-1 wild-type, 7.9 x 10 ⁶ cfu of LS-1 clfAPY\ , 10.7 x 10 ⁶ cfu of LS-1 elf APYW, or 9.4 x 10 ⁶ cfu of the LS-1 elf A null mutant. Sepsis was studied by following the survival rate. After 16 days only 40% of mice challenged with the wild-type strain were alive while 90% of the mice challenged with the clfAPYl mutant and elf A null mutant groups and 80% mice infected with the ctf APYIl mutant were alive (Fig. 4). The clfAPYl mutants and the elf A null mutant of LS-1 were significantly less virulent (P= 0.014, P = 0.05 and P= 0.03, respectively).

Immunization with recombinant CIfA proteins

The effect of vaccination with recombinant wild-type CIfA A domain protein (rClfA) and mutant ClfAPYl protein (rClfAPY) was studied in both the septic arthritis model and the sepsis model. Mice were sensitized and then boosted twice with control protein BSA, rClfA, or rClfAPY, and subsequently infected with 4.0 x 10 ⁶ cfu of S. aureus strain Newman to induce septic arthritis, or with 2.3 x 10 ⁷ cfu of strain Newman to induce sepsis. Immunization with rClfAPY (i.e. ClfAPYl recombinant protein A domain) protected significantly against septic death as compared to control mice (P = 0.01 , Fig. 5) while rClf A immunization did not achieve significant protection. One day before bacterial infection there was a much higher specific serum antibody response to both rClfAPY and rClfA in mice immunized with rClfAPY (A ₄₀₅ = 0.39 (0.33, 0.56) and 0.71 (0.52, 0.81)) as compared to mice immunized with rClfA (A ₄₀₅ = 0.13 (0.07, 0.17) and 0.15 (0.10, 0.24), P < 0.0001 in both comparisons (median, interquartile range)). Control immunized animals had only background levels (A _{405 nm} = 0 and 0.01 (0, 0.01) (median, interquartile range)). The immunized mice which were to be infected with the lower, arthritic bacterial dose had similar antibody responses to rClfA and rClfAPY as the mice in which sepsis were induced (data not shown). Immunization with both rClfA and rClfAPY protected against the development of arthritis, although the protection was not significant (Fig. 9). During day 5 to 9 after infection the weight loss was significantly reduced in the rClf APY and rClfA immunized mice, as compared to the control mice (data not shown). A trend to diminished bacterial growth in kidneys of mice immunized with rClfAPY or rClfA at day 11 after infection (BSA: 38 (3, 436); rClfAPY: 7 (2, 17); rClfA: 10 (7, 54) x 10 ⁷ cfu / kidney pair) was observed.

To get a more accurate measure of the specific antibody responses in the different immunization groups, the responses were determined at several serum dilutions (the second run). Data shows that there were very likely higher titers of specific antibodies in sera from rClfAPY immunized mice to both the rClfAPY and rClfA wildtype antigens, in both the mice which were to be infected with the septic and the arthritic bacterial dose, respectively, than in sera from rClf A wildtype immunized mice, since there wassignificantly higher antibody responses measured as absorbance in mice immunized with rClfAPY at each serum dilution in all comparisons (P<0.0001 to P=0.008, Figure

12-15). BSA immunization evoked only a background antibody response.

CONCLUSION

The results strongly suggest that the CIfA - fibrinogen interaction is crucial for the bacterial virulence and disease outcome. The ability of CIfA to bind fibrinogen was associated with enhanced virulence in terms of the ability to cause septic death. In both staphylococcal strains tested, a clfAPY mutant induced less septic death than the wild-type. Also, the severity of arthritis was strongly reduced in mice infected with the non-fibrinogen binding clfAPY mutant. A likely mechanism for the promotion of virulence by the fibrinogen - bacterial cell surface interaction is inhibition of neutrophil phagocytosis (5). Neutrophils are crucial for the host defence in the early phase of S. aureus infection (13). Without neutrophils, bacterial growth is strongly augmented in blood and kidneys, and the frequency of arthritis and mortality increases. Fibrinogen mediated inhibition of neutrophil phagocytosis by CIfA could explain at least in part the more pronounced virulence of wildtype S. aureus compared to the clfAPY mutants. Binding of fibrinogen to CIfA could decrease opsonophagocytosis by neutrophils by reducing opsonin deposition or access to opsonins by neutrophil receptors. Alternatively bound fibrinogen might block the binding of an unknown protective host factor to S. aureus. Another option is that the fibrinogen - CIfA interaction promotes bacterial passage from blood vessel into the tissue or promotes colonization in tissues. Unexpectedly, our data also show the CIfA null mutant was more virulent than the clfAPY mutant strains. Possibly the CIfA protein has functions in vivo other than interacting with fibrinogen. This interaction is clearly disadvantageous for the host as shown in this study. Other functions of CIfA are presently not well mapped but non-fibrinogen dependent platelet aggregation exerted by CIfA might result in trapping of big amounts of S. aureus in circulation with subsequent elimination of the bacterial-platelet complexes through the reticuloendothelial system. Such platelet aggregation mediated elimination of staphylococci would readily occur in the wild-type and clfAPY mutated strains but not in the clfA knockout. Whereas in the wild-type strain the fibrinogen interaction would overshadow the other events, in the clfAPY mutants such bacterial elimination might be highly beneficial to the host.

The clfA knockout mutant protected against septic death to the same degree as the clfAPY mutation in S. aureus strain LS-1 , but protected less, if at all, in strain Newman. The overall impact of CIfA expression on bacterial virulence could differ between different S. aureus strains depending on the level of expression and the presence of other virulence factors.

The issue whether the clfAPY mutant displays equal or lower virulence once in the joint cavity is of certain importance having in mind that in inflamed synovial fluid fibrinogen and fibrin are abundant. Our data suggest that the clfAPY mutant is less destructive for cartilage and bone.

The protective effect of recombinant CIfA A domain non-fibrinogen binding D _32I P ₃₃₆ Y ₃₃₈ mutant was greater than for wildtype rClfA. Immunization with ClfAPY very likely induced a better immune response since higher specific antibody responses were evoked against both the immunogen and the wildtype CIfA protein. More importantly, it induced a greater protective immune response against septic death than wildtype CIfA.

In conclusion, our results show that rClfAPY is a better vaccine candidate than wild type recombinant CIfA. We hypothesize that binding of fibrinogen by wild-type CffA protein during the immunization phase decreases antigen presentation due to hiding of important epitopes on the CIfA molecule and hence impairs specific antibody production. Example 2

rClfA A region truncate comprising N2 and N3 (rClfA 221-559)

MATERIALS & METHODS:

The protocols outlined in Example 1 were followed in this example which utilized

- rClfA 221-559 (i.e. CIfA A region truncate comprising N2 and N3 corresponding to amino acids 220-559)

- rCifAPYI221-559; and

- BSA.

There were 15 female NMRI mice per group who were 8 weeks old at start of experiments. In this Example, the constructs used for immunization were CIfA wild type/native N2N3 truncate, CIfA N2N3 truncate with mutation PYI as defined in Example 1. BSA was used as the control.

Vaccination with wild-type and mutant recombinant CIfA

The mice were immunized with rClfA 221-559, rClfAPYI 221-559 or BSA in accordance with the protocol of Example 1. Purified rClfA221-559, rClfAPYI221-559 (i.e. CIfAPYI recombinant protein A subdomains N2 and N3) or BSA were dissolved in PBS and emulsified 1 :1 in Freund ^' s complete adjuvant. Two hundred μl of the emulsion containing 30 μg (= 0.79 nmol) of protein was injected s.c. on day 0. First booster immunization with 30 μg of protein in physiologic saline in incomplete Freund ^' s adjuvant was performed on day 12. Second booster immunization was done day 22. On day 31 the mice were bled and sera were frozen for later analysis of antibody responses.

Specific antibodies - ELISA

Serum samples from immunized mice were obtained 9 days after the second booster immunization. The serum specific antibody response against rClfA221-559 and rClfAPYI221-559 was measured by ELISA. Microplates (96-well; Nunc) were coated with 5 μg/ml of recombinant protein in PBS. Blocking agent, serum samples, biotinylated antibodies, and ExtrAvidin-proxidase were all diluted in PBS. The assay was run according to a previous description (8). All serum samples were diluted 1 :5000, 1 :20000, 1 :80000 and 1 :320000, and antibody response was monitored as absorbance at 405 nm. RESULTS:

SPECIFIC ANTIBODY RESPONSE:

The antibody response was measured by absorbance in an ELISA-assay, as per Example 1 , with four different serum dilutions. The data obtained was very similar to the data in the Example 1.

It was found that rClfAPYI221-559 immunization very likely gave rise to higher titers of specific antibodies to both native rClfA221-559 and rClfAPYI221-559, as compared to native rClfA221-599 immunization, since there were significantly higher antibody responses measured as absorbance in mice immunized with rClfAPYI221-559 at each serum dilution in all comparisons but one (P=0.001 to 0.025, see Figures 16 and 17). BSA immunization evoked only background levels of antibody response.

CONCLUSION

We found that immunization with a rClfAPYI221-559 protein gave rise to significantly higher antibody responses to both the immunogen and the wildtype CIfA protein, than immunization with the native protein.

Based on these findings, we conclude that DPY-immunization, regardless of whether the DPY protein comprises amino acids 40 to 550 as in Example 1 or amino acids 221 to 559 as in Example 2, induces a better immune response than immunization with native CIfA of the corresponding size.

Example 3

CIfA A region truncate (δ/delta latch truncate)

MATERIALS & METHODS:

The protocols outlined in Example 1 were followed in this example which utilized the following construct: - rClfA 221 -531 (i.e. rClf A A region truncate comprising N2 and N3 amino acids 220-

559 but without the latching peptide amino acids 532-538 and the subsequent proline-rich residues.

There were 15 female NMRI mice in the group who were 8 weeks old at start of experiment. In this Example, the above construct was used for immunization. The mice were immunized with the above truncate in accordance with the protocol of Example 1.

Vaccination with wild-type and mutant recombinant CIfA

Purified rClfA221-531 was dissolved in PBS and emulsified 1 :1 in Freund ^' s complete adjuvant. Two hundred μl of the emulsion containing 0.79 nmol of protein was injected s.c. on day 0. First booster immunization with 0.79 nmol of protein in physiologic saline in incomplete Freund ^' s adjuvant was performed on day 12. Second booster immunization was done day 22. On day 31 the mice were bled and sera were frozen for later analysis of antibody responses. Specific antibodies - ELiSA

Serum samples from immunized mice were obtained 9 days after the second booster immunization. The serum levels of specific antibodies was measured by ELISA. Microplates (96- well; Nunc) were coated with 4.6 μg/ml of rClfA221-531 protein which is equimolar to 5μg/ml of rClfA221-559 and rClfAPYI221-559 from Examples 1 and 2. Blocking agent, serum samples, biotinylated antibodies, and ExtrAvidin-proxidase were all diluted in PBS. The assay was run according to a previous description (8). All serum samples were diluted 1 :5000, 1 :20000, 1 :80000 and 1 :320000, and antibody response was monitored as absorbance at 405 nm.

RESULTS:

The antibody response was measured by absorbance in an ELISA-assay, as per Example 1. It was found that rClfA221-531 immunization gave rise to an immune response, measured as a specific antibody response (Figure 18).

CONCLUSION:

We found that the fragment rClfA221-531 also works well as an immunogen, since the antigen evokes a specific antibody response.

EXAMPLE 4 - Protein Purification CIfA proteins tested & Protocol

• Wildtype CIfA full length

• CIfA D ₃₂₁ YP ₃₃₆ SY ₃₃₈ A (triple mutant) (i.e clfAPYI 40-559) (SEQ ID NO 4)

• ClfAPYI without the D ₃₂₁ Y substitution (i.e. CIfAP ₃₃₆ SY ₃₃₈ A) Recombinant his-tagged proteins outlined above were purified by immobilised nickel chelate affinity chromatography. Expression was induced by the addition of 1 mM IPTG to exponentially growing cells. After 3 h induction, cells were harvested by centrifugation and resuspended in 30 ml binding buffer (0.5 M NaCI, 20 mM Tris-HCI, 2OmM imidazole, pH7.9) containing protease inhibitors (Complete EDTA-free, Roche). Cells were lysed in a French pressure cell and the lysate was cleared by high-speed centrifugation followed by filtration. A HiT rap Chelating HP column (GE Healthcare) was charged with 150 mM Ni ²⁺ and equilibrated with binding buffer. The cleared lysate was applied to the column and the column was then washed with binding buffer. Bound protein was eluted with a continuous linear gradient of imidazole (5 - 100 mM) in 0.5 M NaCI, 20 mM Tris-HCI (pH 7.9).

Samples were analysed by SDS-PAGE (shown in Figure 21 A and 21 B) and fractions containing protein of the correct molecular weight were pooled and dialysed against PBS for 16 h at 4°C. Following dialysis the proteins were analysed by SDS-PAGE to assess integrity (Figure 21C). Results

The results are shown in Figure 21 where

A. Purification of CIfA wt (1) and triple mutant (2). The proteins are pure and no breakdown is evident.

B. Purification of CIfAPYI without the D ₃₂₁ Y substitution. Protein product breakdown is evident.

C. Following overnight dialysis at 4 ³ C, 10 μl of triple mutant (1) or CIfAPYI without the D _32I Y substitution (2) (10 μM). This figure shows that the triple mutant (1) does not break down.

Conclusion

By observing the breakdown of the proteins during purification, we found that the CIfA D ₃₂₁ YP ₃₃₆ SY ₃₃₈ A (triple mutant) provides a more stable protein than the CIfAPYI without the D ₃₂₁ Y substitution (i.e. CIfAP ₃₃₆ SY ₃₃₈ A). Thus, these results show that the CIfA D ₃₂₁ YP ₃₃₆ SY ₃₃₈ A (triple mutant) protein is easier to purify when the D ₃₂₁ mutation is present. Thus, the presence of the D ₃₂₁ mutation leads to a protein with increased stability.

REFERENCES

1. Peacock SJ, Moore CE, Justice A, Kantzanou M, Story L, Mackie K, O ^' Neill G, Day NPJ

(2002) Virulent combinations of adhesin and toxin genes in natural populations of Staphylococcus aureus. Infect lmmun 70:4987-4996. 2. McDevitt D, Nanavaty T, House-Pompeo K, Bell E, Turner N, McEntire L, Foster T, Hook M (1997) Characterization of the interaction between the Staphylococcus aureus clumping factor (CIfA) and fibrinogen. Eur J Biochem 247:416-424.

3. McDevitt D, Francois P, Vaudaux P, Foster TJ (1994) Molecular characterization of the clumping factor (fibrinogen receptor) of Staphylococcus aureus. MoI Microbiol 11 :237-

248.

4. Palmqvist N, Patti JM, Tarkowski A, Josefsson E (2004) Expression of staphylococcal clumping factor A impedes macrophage phagocytosis. Microb Infect 6:188-195.

5. Higgins J, Loughman A, van Kessel KPM, van Strijp JAG, Foster TJ (2006) Clumping factor A of Staphylococcus aureus inhibits phagocytosis by human polymorphonuclear leukocytes. FEMS Microbiol Lett 258:290-296.

6. Loughman A, Fitzgerald JR, Brennan MP, Higgins J, Downer R, Cox D, Foster TJ (2005) Roles of fibrinogen, immunoglobulin and complement in platelet activation promoted by Staphylococcus aureus clumping factor A. MoI Microbiol 57 :804-818.

7. O ^' Brien L, Kerrigan SW, Kaw G., Hogan M., Penades J., Litt D., Fitzgerald D.J., Foster T.J. & Cox D. (2002) Multiple mechanisms for the activation of human platelet aggregation by Staphylococcus aureus: roles for the clumping factors CIfA and CIfB, the serine-aspartate repeat protein SdrE and protein A. MoI Microbiol 44, 1033-1044.

8. Josefsson E., Hartford O., O'Brien L, Patti JM, Foster T (2001) Protection against experimental Staphylococcus aureus arthritis by vaccination with clumping factor A, a novel virulence determinant. J Infect Dis 184:1572-1580.

9. Palmqvist N, Foster T, Fitzgerald R, Josefsson E, Tarkowski A (2005) Fibronectin- binding proteins and fibrinogen-binding clumping factors play distinct roles in staphylococcal arthritis and systemic inflammation. J lnf Dis 191 :791 -798.

10. Deivanayagam CCS, Wann ER, Chen W, Carson M, Rajashankar KR, Hook M, Narayana SVL (2002) A novel variant of the immunoglobulin fold in surface adhesins of Staphylococcus aureus: crystal structure of the fibrinogen-binding MSCRAMM, clumping factor A. The EMBO Journal 21 :6660-6672.

11. Bremell T, Langβ S, Yacoub A, Ryden C, Tarkowski A (1991) Experimental

Staphylococcus aureus arthritis in mice. Infect lmmun 59:2615-2623.

12. Sakiniene E, Bremell T, Tarkowski A (1996) Addition of corticosteroids to antibiotic treatment ameliorates the course of experimental Staphylococcus aureus arthritis. Arthritis Rheumatism 39:1596-1605.

13. Verdrengh M, Tarkowski A (1997) Role of neutrophils in experimental septicemia and septic arthritis induced by Staphylococcus aureus. Infect lmmun 65:2517-2521.

14. Duthie ES, Lorenz LL (1952) Staphylococcal coagulase: mode of action and antigenicity. J Gen Microbiol 6:95-107.

15. Kreiswirth BN, Lόfdahl S, Betley MJ, O'Reilly M, Schlievert PM, Bergdoll MS, Novick RP (1983) The toxic shock syndrome exotoxin structural gene is not detectably transmitted by a prophage. Nature 305:709-712.

16. Foster TJ (1998) in Methods in Microbiology Vol. 27: Bacterial Pathogenesis, eds

Williams P, Ketley J, Salmond G (Academic Press, London), pp 433-454.

17. O'Connell DP, Nanavaty T, McDevitt D, Gurusiddappa S, Hook M, Foster TJ (1998) The fibrinogen-binding MSCRAMM (clumping factor) of Staphylococcus aureus has a Ca ²⁺⁺ - dependent inhibitory site. J Biol Chem 273:6821-6829.

18. Sakiniene E, Bremell T, Tarkowski A (1999) Complement depletion aggravates Staphylococcus aureus septicaemia and septic arthritis. Clin Exp Immunol 115:95-102.

19. Mancini G, Carbonara AO, Heremans JF (1965) Immunochemical quantitation of antigens by single radial immunodiffusion, lmmunochemistry 2:235-254.

20. Bremell T, Abdelnour A, Tarkowski A (1992) Histopathological and serological progression of experimental Staphylococcus aureus arthritis. Infect lmmun 60:2976-

2985.