Field:

The invention is in the field of determining or quantifying an activity of a proteolytic enzyme of Clostridium difficile in a sample and concerns substrates, to be used therefore and an assay process and an assay kit and device therefore. The invention additionally is in the field of a pharmaceutical composition for treatment of an individual suffering from infection with C. difficile.

Clostridium difficile is an anaerobic, Gram-positive spore-forming bacterium. Human intake of spores occurs through the faecal-oral route. When conditions become favourable (after leaving the stomach) in the small intestine, the C.

difficile spores germinate, liberating the vegetative forms. These vegetative cells adhere to the mucosal epithelium, multiply and produce toxins responsible for diarrhoeal symptoms (Rupnik et al., 2009. Nat Rev Microbiol 7: 526-536). Once germinated, the vegetative cells in the colon encounter an unreceptive environment. Besides competition with the normal flora, immune responses, gastric fluids (Muniz et al., 2012. Front Immunol 3: 310.) and specialized antimicrobial peptides, which all act against the developing infection, another hurdle is the physical barrier formed by a layer of glycoproteins (mucins) covering the underlying epithelial cells. In order for the bacteria to reach this epithelial adhesion surface and to exploit host tissues for nutrients, it must first penetrate this protective layer. The best studied virulence factors in C. difficile are the toxins, TcdA and TcdB (Kuehne et al., 2010. Nature 467: 711-713; Lyras et al., 2009. Nature 458: 1176-1179). These toxins cause destruction of the intestinal barrier by disrupting the epithelial actin cytoskeleton. The subsequent increased permeability of the intestinal epithelium is speculated to lead to increased exudation of fluids, including nutritional substances. Damaging the intestinal mucosa causes the symptoms of C. difficile infection, including

pseudomembranous colitis.

Many enteric pathogens express factors that reduce competition, allow evasion of host immune responses and promote adhesion and/or invasion of tissues. These virulence factors, which communicate with the environment are located in the cell membrane/wall (controlling adhesion and protection) or are secreted

(modifying the surrounding environment). In C. difficile, few examples are described that illustrate how the bacteria circumvent host defense mechanisms and efficiently adhere to epithelial cells. In response to attack by antimicrobial peptides, C. difficile expresses a set of genes that change the surface charge, thereby diminishing the interaction of cationic antimicrobial peptides on the bacterial surface (McBride and Sonenshein, 2011. Microbiology 157: 1457-1465). C. difficile cell membrane and cell wall proteins interact with the host and are most likely involved in adhesion and colonization. Footholds on the host cell surface proteins include extracellular matrix components fibronectin, laminin, collagen and fibrinogen (Lin et al., 2011. J Biol Chem 286: 3957-3969; Hennequin et al., 2003. Microbiology 149: 2779-2787.; Sebaihia et al., 2006. Nat Genet 38: 779-786; TuUi et al., 2013. Cell Microbiol) aU implicated in C. difficile adhesion. In contrast to cell associated proteins, bacterial secretory proteins such as the toxins are released into the surrounding milieu where they freely interact with host cell to exert their function. It is very likely that other factors in C. difficile are involved in concealing the bacteria from the host's defence or direct interaction with host tissues, but little is known about extracellular C. difficile factors that contribute to manipulation of their surroundings (Sebaihia et al., 2006. Nat Genet 38: 779-786; Vedantam et al., 2012. Gut Microbes 3: 121-134).

Present methods for monitoring infections C. difficile infections (CDI) include enzyme immune-detection, nucleic acid amplification-based detection, and animal-assisted detection (Bomers et al., 2012. BMJ 345:e7396). Enzyme immune-detection has a low sensitivity and a relatively long turnaround time. Nucleic acid amplification-based test such as, for example, BD GeneOhm Cdiff assay (San Diego, CA) and the Prodesse ProGastro CD assay (W aukesha, WI) have an acceptable sensitivity and a turnaround time of about 4 hours. However, these tests are expensive and require specialized equipment and expertise.

A fast, sensitive method for monitoring CDI is essential for patient management and infection control. The present invention provides a peptide or peptide analog comprising the amino acid sequence A/PjA/P-V/P/I/C/T-P or N-A/PjA/P- V/P/I/C/T-P, wherein the arrow denotes the cleavage site. Said peptide is a substrate of a C. difficile-specific protease, CD2830. This protease is highly specific and important for virulence of C. difficile. Said peptide comprising the amino acid sequence A/P jA/P-V/P/I/C/T-P or N-A/PjA/P-V/P/I/C/T-P, provides a highly sensitive, fast and direct detection method that can be performed on culture medium, colonies and a sample, preferably feces, of infected individuals, without purification of the enzyme.

The term peptide, as used herein, refers to a short chain of between 5 and at most 100 amino acid monomers, preferably of between 6 and at most 50 amino acid monomers, more preferred between about 7 and at most 30 amino acid monomers, more preferred at most 25 amino acid monomers, more preferred at most 20 amino acid monomers such as, for example, eight amino acid monomers, nine amino acid monomers, ten amino acid monomers, eleven amino acid monomers, twelve amino acid monomers, thirteen amino acid monomers, fourteen amino acid monomers, fifteen amino acid monomers, sixteen amino acid monomers, seventeen amino acid monomers, eighteen amino acid monomers, nineteen amino acid monomers, twenty amino acid monomers, twenty one amino acid monomers or twenty two amino acid monomers.

The term peptide includes a peptide in which one or more of the amino acid monomers have been modified, for example by acetylation, amidation and/or glycosylation. The term peptide also includes a peptide that is directly or indirectly attached to a solid surface, for example a surface of a multiwell plate or a magnetic bead. The term peptide analog includes peptide analogues or peptidomimetics which are or which comprise small protein-like chains such as peptoids and 6-peptides designed to mimic a peptide. The altered chemical structure is preferably designed to adjust one or more properties such as, for example, stability, of a peptide.

The epidemiology of CDI has changed dramatically during the last years.

Infection rates have increased markedly in most countries with detailed surveillance data. According to a recent survey (see Freeman et al., 2010. Clin Microbiol Rev 23: 529-549), ribotype 001 is the most common isolated toxinogenic isolate in clinical infections (13%), followed by ribotype 014 (9%). Ribotypes 002, 012, 017, 020, and 027 were each found in 6% of toxinogenic isolates, whereas ribotype 078 was found in 3% of toxinogenic isolates. Patients infected with ribotype 027 (sometimes referred to as BI/NAPl/027) are likely to have suffered from a more severe disease and to have been treated with metronidazole or vancomycin, compared with patients infected by another PCR ribotype. The C. difficile-specific protease, CD2830, was found to be present in all toxinogenic ribotypes.

Said peptide preferably comprises the amino acid sequence V/L/I/P-V/L/I- N- A/P jA/P-V/P/I/C/T-P-P, more preferred V/L/I- N-A/P |A/P-V/P/I/-P-P, more preferred V/L/I- N-Pj P-V-P -P.

A preferred peptide according to the invention comprises the amino acid sequence AjA-V-P, AjA-P-P, AjA-I-P, AjA-C-P, AjA-T-P, PjA-V-P, PjA-P-P, PjA-I-P, PjA- C-P, P jA-T-P, AjP-V-P, AjP-P-P, AjP-I-P, AjP-C-P, AjP-T-P, P jP-V-P, P jP-P-P, PjP-I-P, PjP-C-P, PjP-T-P, N-AjA-V-P, N-AjA-P-P, N-AjA-I-P, N-AjA-C-P, N- A|A-T-P, N-P|A-V-P, N-P|A-P-P, N-P|A-I-P, N-P|A-C-P, N-P|A-T-P, N-A|P-V-P, N-A|P-P-P, N-A|P-I-P, N-A|P-C-P, N-A|P-T-P, N-P|P-V-P, N-P|P-P-P, N-P|P-I- P, N-P|P-C-P, or N-P|P-T-P. A most preferred peptide or analog according to the invention comprises the amino acid sequence

PjP-V-P or PjP-P-P, such as, preferably, N-P jP-P-P-D and/or N- PjP-P-P.

Methods for detecting proteolytic activity of C. difficile-specific protease, CD2830, are known in the art and include methods for detecting the enzyme itself, methods for detecting products of peptide hydrolysis or methods for detecting residual substrate peptide. The methods available for detection and assay of proteolytic activity vary in their simplicity, rapidity, range of detection and sensitivity. Said methods comprise qualitative assays such as protein agar plate assay, radial diffusion and thin layer enzyme assay and, preferred, quantitative assays which provide a measure of the proteolytic activity of the enzyme. The commonly used methods employ natural or synthetic substrates using techniques such as enzyme-linked immunosorbent assay-based assays (ELISA),

spectrophotometry, fluorimetry, and radiometry.

In enzyme-linked immunosorbent assay (ELISA)-based assays, C. difficile- specific protease CD2830 is detected by an antibody that is directed against said protease. A preferred ELISA is a double-antibody-sandwich ELISA, or an inhibition ELISA (Clements et al. 1990. Appl Environ Microbiol 56: 1188-1190) which can detect small quantities of said protease. A peptide or analog according to the invention is preferably labeled, for example with a chromogenic group, a (chemo)luminescent group, a radiolabel and/or, most preferred, a fluorescent group. A labeled peptide is preferably used in

spectrophotometry, fluorimetry, and radiometry. Said label is preferably present at a terminus of the peptide. Said terminus is either the amino-terminus (N- terminus) or the carboxy-terminus (C-terminus). The term "terminus" indicates that the label preferably is present on one or more of the first five amino acid monomers from the N-terminus, and/or on one or more of the last five amino acid monomers at the C-terminus. Said label is preferably present at the N-terminal amino acid monomer, and/or at the C-terminal acid monomer. The skilled person will understand that said label can be indirectly coupled to the N- or C-terminus, for example through a linker that is attached to N-terminus and/or C-terminus. Said linker preferably is an amino acid residue, for example a glutamic acid residue at the N-terminus, and/or an aspartic acid or a cysteine residue at the C- terminus.

A preferred chromogenic group is 2'-azino-bis-(3-ethylbenzothiazoline-6-sulfonic acid (ABTS), o-phenylenediamine (OPD), 3,3',5,5'-tetramethylbenzidine (TMB), p- nitroanihde, paranitrophenol and/or 5-bromo-4-chloro-3-hydroxyindole. A preferred (chemo)luminescent group is a dioxetane derivative such as 1,2- dioxetanedione (C204), 3,3,4,4-tetramethyl-l,2-dioxetane and 3,3,4-trimethyl- 1,2 -dioxetane, 3-(4-Methoxyspiro[l,2-dioxetane-3,2'-tricyclo[3,3,l, 13,7]decan]-4- yl- )-l-aniline and luminol. A preferred radiolabel is 35S and/or 14C. The labeled proteolytic products are preferably separated from the unhydrolyzed substrate, for example by employing magnetic beads.

Said group preferably comprises one or more fluorescent group. Fluorescent groups are known and have been described, for example, in U.S. Pat. No.

7,256,012 and U.S. Pat. No. 7,410,769. Some non-limiting examples of fluorescent groups include coumarin derivatives such as 7-amino-4-methylcoumarin (AMC), 7-acetoxy-4-methylcoumarin (7-AC-4-MC) and 7-hydroxycoumarin, fluorescein, tetramethylrhodamine, rhodamine B, lissamine, rhodamine X, Texas Red, cyanine dyes, Dabcyl, BODIPY dyes, alexa dyes, QSY 7 and QSY 9 dyes, and other fluorescent dyes commonly available from, for example, Invitrogen Corp (Carlsbad, Calif.). Other dyes known to those skilled in the art may also be used.

A preferred fluorescence-based protease assay is simple, inexpensive and sensitive. Said protease assay preferably comprises a soluble fluorescein isothiocyanate (FITC)-labeled peptide, or a soluble Alexa 488-labeled peptide. A peptide or analog according to the invention preferably comprises a quencher group, preferably a fluorescent group and a quencher group, whereby said fluorescent group is present at a first terminus and said quencher group is present at a second terminus of the peptide. The presence of a fluorescent group and a quencher group allows detection of CD2830 activity by a fluorescent peptide energy transfer assay. In this assay, said quencher group is in close proximity to the fluorescent group while the peptide remains intact. Little, if any, light is emitted when the fluorescent group is excited, and a low or no signal is measured. When the peptide is cleaved by the enzyme, the quencher is separated from the fluorescent group, which now emits light when excited, and a signal can be measured. The intensity of the signal is proportional to the amount of peptide that is cleaved, which in turn is proportional to the amount of active enzyme, as the peptide is in excess. Examples of fluorescent groups (donor)/quencher group (acceptor) pairs include: fluorescein/tetramethylrhodamine,

IAEDANS/fluorescein, EDANS/Dabcyl, fluorescein/fluorescein, 5 - carboxyfluorescein (5-FAM)/QXL 520, BODIPY FL/BODIPYFL, FITC/graphene oxide, EDANS/Alexa 488, Fluorescein/QSY 7 and QSY 9 dyes, and further combinations thereof. A preferred peptide or analog according to the invention comprises N-[4-(4- dimethylamino)phenylazo]benzoic acid (DABCYL) at a first terminus, and 5 - [(2 - aminoethyl)amino]naphthalene - 1 - sulfonic acid (EDANS) at a second terminus, or comprises Alexa 488 (Alexa Fluor® 488 carboxylic acid, succinimidyl ester) at a first terminus and EDANS at a second terminus, or comprises FITC at a first terminus, and graphene oxide at a second terminus. Said graphene oxide is preferably covalently attached to said peptide.

The term "at a first terminus" refers to the N-terminus or C-terminus of the peptide. The term "at a second terminus" refers to the N-terminus or C-terminus of the peptide that differs from the first terminus. The invention further provides an aqueous composition comprising the peptide or analog according to the invention. Said composition preferably comprises means for maintaining the pH of the aqueous composition at a pH of between 2 and 10, preferably between 3 and 9, preferably between 4 and 8, preferably between 5 and 7, preferably around 6.

Said composition preferably additionally comprises an inorganic salt, preferably NaCl and/or ZnC12. A preferred composition according to the invention preferably comprises a protease inhibitor. Said protease inhibitor is added to the composition to suppress the activity of other proteases that may be present in the sample. Said other protease includes a protease such as, for example, a serine protease, a threonine protease, a cysteine protease, an aspartate protease, and/or a glutamic protease. Said other protease may be trypsin, chemotrypsin, cathepsin, papain, kallikrein, plasmin, thrombin, pepsin, elastase, factor Xa, and/or subtilisin.

A preferred protease inhibitor is selected from aprotinin, 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (AEBSF), 4- amidinophenylmethanesulfonyl fluoride hydrochloride (APMSF),

phenylmethylsulfonyl fluoride (PMSF), bestatin, chymostatin E-64 (N-[N-(L-3- trans-carboxirane-2-carbonyl)-L-leucyl]-agmatine), leupeptin, trypsin inhibitor, pepstatin and/or phosphoramidon. A most preferred protease inhibitor is a protease inhibitor cocktail, preferably a cocktail without EDTA, most preferred the protease inhibitor cocktail cOmplete, EDTA-free (Roche).

The invention further provides a kit of parts or a device comprising the peptide or analog according to the invention or the composition of the invention. Said kit or device provides a highly specific and highly sensitive tool for monitoring C.

difficile in a sample of an individual. The term "individual", as used herein, includes birds and mammals such as a pet, for example dog and cat; ungulates including pig, horse and ruminants such as sheep, cow; and goat; fowl, including chicken, duck, goose, and turkey, and primates, including human. A preferred individual is a chicken, pig or human.

Said kit or device preferably comprises a holder comprising the peptide according to the invention, preferably in an aqueous composition comprising means for maintaining the pH of the aqueous composition at a pH of between 2 and 10. Said kit or device preferably comprises one or more separate holders such as a receptacle, for example a multiwell plate, and a holder comprising a protease inhibitor. Said protease inhibitor may be added to the holder comprising the peptide prior to the start of an assay.

The term "sample" includes meat, such as beef, veal, pork, and poultry, of the birds and mammals such as a pet, for example dog and cat; ungulates including pig, horse and ruminants such as sheep, cow; and goat; fowl, including chicken, duck, goose, and turkey. The term "sample" also includes a tissue sample, an amount of bodily fluid such a blood or urine and, preferably, stool from the birds and mammals such as a pet, for example dog and cat; ungulates including pig, horse and ruminants such as sheep, cow; and goat; fowl, including chicken, duck, goose, and turkey; and primates, including human.

The invention further provides a method for identifying an agent that is capable of modulating protease activity of CD2830, comprising (a) providing said agent to a testing system comprising CD2830 protease; (b) contacting said testing system with a peptide of any one of claims 1-6; and (c) detecting reduced or increased cleavage of the peptide, compared to a testing system to which the agent was not provided. Said testing system comprising CD2830 protease further comprises means for determining the activity of the CD2830 protease, preferably a peptide according to the invention, preferably a labeled peptide. Said agent preferably inhibits the protease activity of CD2830 protease. The invention also provides an inhibitor of the CD2830 protease that is identified by the methods and means provided by this invention. Said inhibitor preferably is used in a method for the treatment of an individual suffering from infection with C. difficile.

Said agent preferably is a chemical compound. Protease-targeted compound libraries are available, for example from Otava Ltd, Canada, TimTec LLC

(Newark, DE), and Enzo Life Sciences (Farmingdale, NY). Said libraries can be screened using the methods and means provided by this invention for identifying an inhibitor of the CD2830 protease. The invention further provides an agent that is identified by the methods of the present invention, for use in a method of treatment an individual suffering from infection with C. difficile, whereby the agent inhibits the protease activity of CD2830.

The invention further provides an in vitro method for determining the presence of Clostridium difficile in a sample, comprising incubating the sample with the peptide of the invention, the composition of invention, or the kit or device of the invention. Said sample, preferably a stool sample, is preferably directly incubated with the peptide of the invention, the composition of invention, or the kit or device of the invention, without prior purification of the protease.

The C. difficile- specific protease, CD2830, was found to be thermostabile and remains active after incubation of the sample for several minutes at a

temperature of more than 55 °C. To reduce or eliminate the activity of other protease during the determination of the presence of Clostridium difficile in a sample, the sample is preferably heated prior to determining the presence of Clostridium difficile in the sample, preferably heated for 1-10 minutes, preferably for 2-5 minutes, at a temperature between 55 and 80 °C, more preferred at a temperature of about 65 °C. The inventors have established that a mutant strain of Clostridium difficile, in which the coding sequence for the C. difficile-specific protease, CD2830, had been functionally rendered inactive by a so called "knock out" mutation, is less virulent compared to the parent strain of C. difficile that does not comprise the mutation. It was additionally found that the knock out strain adhered better to a substrate. Without being bound by theory, the improved adherence might interfere with the colonisation of the gut in infected individuals. A peptide that inhibits the activity of the C. difficile-specific protease, CD2830, might therefore be used to treat infections with C. difficile. Said inhibitory peptide preferably is a modified peptide of the invention that is not hydrolizable by the protease

CD2830.

The invention therefore also provides a pharmaceutical composition comprising protease CD2830, or an activator of this protease. The presence of active protease CD2830 in the gut is thought to prevent adherens of C. difficile, resulting in an effective flush from the gut. Methods for providing a delayed release formulation of CD2830, or of an activator of this protease, for example by providing a tablet comprising the active protease or activator with an enteric coating, are known in the art.

The invention therefore also provides a pharmaceutical composition comprising a peptide according to the invention. Said peptide preferably inhibits the protease activity of CD2830. A preferred inhibitor peptide comprises a non-cleavable Pro- Pro mimetic such as, for example, a structure in which the proline rings have been linked together by an additional bridge. Such structures have been described,, for example, in Reuter et al., (2011). Chem. Eur. J. 17, 12037-12044. A further preferred non-cleavable Pro-Pro mimetic comprises a vinylidene bridge between the PI and Pl'prolines to restrict cleavage (Hack et al., 2013. Angew Chem Int Ed Engl 52: 9539-9543).

Said inhibitor peptide is preferably modified to adjust some properties of the peptide such as, preferably, the stability of the peptide. Said modification comprises, for example, replacing the peptide bond at the cleavage site with an alkene dipeptide isostere. Said inhibitor preferably mimics the transition state during hydrolysis of a peptide by CD2830. Said inhibitor peptide preferably is a phosphinic peptide, in which one or more peptide bonds (CO-NH) are replaced by a phosphinic acid moiety, for example (P02-CH2) (Dive et al., 2004. CMLS 61: 2010-2019). Said phosphinic acid moiety preferably replaces the peptide bond that is cleaved by CD2830. Further preferred phosphinic acid moieties are (PO2- NH) and (P02-0). A further preferred inhibitor peptide comprises a sulfurous acid moiety, in which the peptide bond that is cleaved by CD2830 is replaced by a sulfurous acid moiety preferably (S02-CH2), (S02-NH), or (S02-0). Inhibition of protease activity by these peptides can be easily monitored in, for example, a FRET fluorogenic assay. Methods to produce such peptides are known in the art.

It will be clear to the skilled person that only the (CO) part of a peptide bond is replaced by a phosphinic acid moiety (PO2) or a sulfurous acid moiety (SO2) in case the P l-prime amino acid of said amino acid bond is a proline.

A further preferred inhibitor peptide comprises either the unprimed amino acid sequence V/L/I- N-A/P, preferably V/L/I/P-V/L/I- N-A/P with a C-terminal chelating group, or the primed amino acid sequence A/P-V/P/I/C/T-P, preferably A/P-V-P, A/P-V/P/I/C/T-P-P or P-V-P-P, with an N-terminal chelating group. Said chelating group is preferably selected from the group consisting of a thiolate group, a carboxylate group, a phosphinyl group and, preferably, a hydroxamate group. Said inhibitor peptide is preferably for use in a method of treatment of an individual suffering from infection with C. difficile, whereby the inhibitor peptide inhibits the protease activity of CD2830. Methods to produce such peptides are known in the art. For example, methyl, ethyl or N-hydroxy-succinimide ester precursors of a desired peptide are prepared by using classical peptide synthesis methodology. These precursors are reacted with excess hydroxylamine in either ethanol or Ν,Ν-dimethylformamide, generating a hydroxamic acid derivative of the peptide according to the invention. Said unprimed amino acid sequence is preferably selected from V-N-A, V-N-P, L- N-A, L-N-P, I-N-A, I-N-P, G-N-L, V-V-N-A, V-V-N-P, V-L-N-A, V-L-N-P, V-I-N-A, V-I-N-P, L-V-N-A, L-V-N-P, L-L-N-A, L-L-N-P, L-I-N-A, L-I-N-P, I-V-N-A, I-V-N- P, I-L-N-A, I-L-N-P, I-I-N-A, I-I-N-P, P-V-N-A, P-V-N-P, P-L-N-A, P-L-N-P, P-I- N-A, V-G-N-L and P-I-N-P.

Yet a further preferred inhibitor peptide comprises the amino acid sequence DTIVINP or DTIVGNL, with a C-terminal chelating group selected from a thiolate group, a carboxylate group, a phosphinyl group and, preferably, a hydroxamate group.

Said primed amino acid sequence is preferably selected from A-V-P, P-V-P, A-P- P, P-P-P, A-I-P, P-I-P, A-C-P, P-C-P, A-T-P, P-T-P, A-V-P-P, P-V-P-P, A-P-P-P, P- P-P-P, A-I-P -P, P-I-P-P, A-C-P-P, P-C-P-P, A-T-P -P, and P-T-P-P.

The invention further provides an inhibitor peptide according to the invention, preferably a hydroxamic acid derivative, for use in a method of treatment of an individual suffering from infection with Clostridium difficile. The invention further provides a method of treatment of an individual suffering from infection with Clostridium difficile, the method comprising administering a pharmaceutical composition comprising an inhibitor peptide according to the invention, preferably a hydroxamic acid derivative according to the invention.

Figure legends

Figure 1. Clostridium difficile CD2830 is a metalloprotease with a fold similar to the Anthrax Lethal Factor catalytic domain

(A) Amino acid sequence alignment of C. difficile CD2830 with Anthrax lethal factor (ALF) according to the Phyre2 protein structure prediction. Identical amino acids are shaded in dark grey and similar residues shaded in gray. Zinc coordinating residues are highlighted in green. Δ (triangle) point to ALF amino- acids involved in peptide substrate recognition. HEXXH metal binding site is indicated.

(B) 3D stereo ribbon model of Anthrax Lethal Factor (left) and CD2830 predicted structure (right). ALF colouring and domain annotation according to Pannifer et al. 2001 (PDB: lj7n). The MAPKK-2 target peptide is shown as a red ball-and- stick model. The Anthrax Lethal Factor consists of four domains (I, II, III, IV). The CD2830 only consists of a proteolytic domain with similar fold to domain IV of ALF.

Figure 2. Proteolytic assays reveal that human HSP906 is a substrate of CD2830

(A) Caco-2 whole cell lysates analyzed by SDS PAGE followed by Coomassie staining.

Lane 1, Caco-2 lysate incubated at 4 °C; Lane 2, Caco-2 lysate incubated at 37 °C; Lane 3, Caco-2 lysate plus rCD2830 at 37°C . Arrow points at cleavage product.

(B) Recombinant HSP90ct and HSP906 analyzed by SDS-PAGE followed by Coomassie staining.

Left: HSP906 incubated with rCD2830 for different time periods; Right: HSP90ct after 20hr incubation with (+) or without (-) rCD2830. MW, molecular weight in kDalton.

(C) Alignment of the C-terminal amino-acid sequences of human HSP906 and HSP90a. Arrow indicates the CD2830 cleavage site.

(D) MALDI-ToF MS spectrum from a synthetic peptide containing the HSP906 sequence containing the CD2830 cleavage site after 20 hr. incubation with rCD2830. Figure 3. Characterisation of the preferred CD2830 protease-cleavage motif and identification of C. difficile LPXTG-motif containing proteins as putative CD2830 substrates

(A) A synthetic peptide library was constructed where all 6 positions surrounding the CD2830 scissile bond were permutated to the 20 standard amino acids within a core synthetic peptide (KAAEEPNAAVPDEIK), resulting in a total set of 120 peptides. Peptides were individually incubated 16hr. with rCD2830 and measured by MALDI-ToF MS to determine whether efficient cleavage had occurred. The resulting CD2830 cleavage motif based on this peptide library screen is shown.

(B) Identification of multiple CD2830 cleavage sites (indicated by arrows) in the C-terminal region of the C. difficile adhesion proteins CD2831 and CD3246. Both putative substrates contain an LPXTG-motif (PPXTG and SPXTG, respectively) and show considerable sequence alignment (shaded areas) around the CD2830 cleavage sites.

(C) CD2830 cleavage motif based on the alignment of in total 13 CD2830 cleavage sites within CD2831 and CD3246. Figure 4. Clostridium difficile adhesin CD2831 is efficiently cleaved by the CD2830 protease

(A) Schematic representation of the C. difficile adhesin CD2831 showing the putative collagen binding domains, collagen stalk and transmembrane domain. A his-tagged recombinant CD2831 (rCD2831) protein corresponding to aa 732-947 (arrow) was produced, containing all the CD2830 cleavage sites.

(B) SDS-PAGE analysis of rCD2831 treated with rCD2830 for different time periods.

(C) Mass spectrometric analysis (Q-ToF-MS) of the small peptide cleavage products formed from rCD2831 after incubation with rCD2830 for 30 minutes. Figure 5. Detection of endogenous CD2830 protease activity in the secretome of C. difficile

A synthetic peptide containing one of the CD2830 cleavage sites in CD2831 (KDTIVINPPVPPSEK) was incubated with the medium collected from cultures of C. difficile WT (WT secretome) and CD2830 knockout cells (CT::CD2830 secretome). After incubation for 16 hrs, samples were analysed by MALDI-ToF MS to determine substrate peptide cleavage. As a positive control, the substrate peptide was incubated with recombinant CD2830 (+ recombinant CD2830) and as a negative control the peptide was incubated alone (- CD2830).

Figure 6. CD2830 cleaves CD2831 from live/intact C. difficile cells

(A) C. difficile cells from begin stationary phase were collected by centrifugation and after washing incubated with recombinant CD2830 for "cell surface shaving". Samples were then measured by LC-ion trap MS analysis with specific emphasis on the detection of cleaved CD2831 peptides. Upper graph: elution profile and MS/MS identification of the CD2831 peptide PAPPNTDEPIVNP resulting from cleavage of recombinant CD2831 by recombinant CD2830. Middle graph: Elution profile from WT cells treated with recombinant CD2830, showing the absence of the CD2831 peptide PAPPNTDEPIVNP. Lower graph: Elution profile from

CT:CD2830 cells treated with recombinant CD2830, showing the presence of the CD2831 peptide PAPPNTDEPIVNP as demonstrated by the similarity in MS/MS spectrum and elution time as shown for the peptide from recombinant CD2831 (upper graph).

(B) C. difficile WT and CT::CD2830 cells were cultured in minimal medium till the beginning of the stationary phase. Cell culture medium (secretome) was then analysed by LC-ion trap MS for the presence of the CD2831 peptide

PAPPNTDEPIVNP (as in A) resulting from cleavage by endogenous CD2830. The CD2831 peptide was only observed in the WT secretome. Figure 7. C. difficile CD2830 protease knockout strain shows higher adherence to a collagen matrix in vitro

Y-axis presents the relative number of hve, adhering bacteria per 24 well (2cm2) cell, CT::CD2830/WT, either on confluent monolayer of Caco-2 cells or uniformly coated with Collagen type I. * indicates the statistical difference, of four independent experiments, P<0.05.

Figure 8. C. difficile CD2830 protease knockout strain shows attenuated virulence in vivo

Groups of 8 hamsters were infected with C. difficile 630Aerm (WT, ·) or the

CD2830 knockout (CT::CD2830,.i). Time from inoculation to endpoint (sacrifice) is indicated. * indicates the statistical difference, P<0.05.

Figure 9. Optimal reaction conditions for CD2830 activity.

MALDI-ToF MS spectrum from synthetic peptide containing the HSP906 sequence covering amino acids 691-714, after 16 hrs incubation

with 0.1 ug rCD2830.

(A) Small peptide cleavage products after incubation with rCD2830 at

temperatures between 30 and 70°C showed active enzyme.

(B) Cleavage by rCD2830 is sensitive to EDTA, as no cleavage products were observed in the presence of 10 mM EDTA.

(C) Cleavage by rCD830 is active between pH 6-9. At pH 5 or lower no cleavage product were observed. Figure 10. Growth curve, OD600 versus time (hours), of C. difficile strain 630 AErm (WT) and CT::CD2830 shows no differences in growth rate.

Cells were grown in trypton-yeast (TY) based broth.

Figure 11. FRET (Fluorescence Energy Transfer assay) for in vitro enzymatic activity of CD2830.

A: Titration curve of recombinant CD2830 protein.

B: WT or CT::CD2830 (CD2830 knock-out) C. difficile. Figure 12. Influence of amino acid changes within a core Dabcyl-EVNP | PVPD- Edans FRET peptide at the six positions surrounding the CD2830 cleavage site (P3, P2, PI I PI', Ρ2', P3') on the efficiency of CD2830-mediated cleavage. Figure 13. FRET assay for inhibition of in vitro CD2830 cleavage of fluorogenic peptide Dabcyl-EVNPPVPD -Edans by various hydroxamate peptides.

Examples

Example 1 EXPERIMENTAL PROCEDURES General

Bioinformatics analysis

Predictions of signal sequences were carried out using the SignalP 4.1 Server, (http://www.cbs.dtu.dk/services/SignalP). Prediction of cell wall binding motifs, anchors and subcellular prediction were performed at

http : //www . ncbi . nlm . nih . go v/Structur e/cdd/wrp sb . cgi , Lip oP server ,

http://www.cbs.dtu.dk/services/LipoP/ and PSORTB ( www.psort.org/psortb/ ). Predictions were performed using standard settings. All sequence alignments were performed by use of Clustal Omega Multiple Sequence Alignment, available from EMBL-EBI European bioinformatics institute,

http://www.ebi.ac.uk/Tools/msa/clustalo/. All predictions were performed using standard settings. Structural models of CD2830 were generated by the automated I-TASSER

(threading, assembly and refinement) simulation method,

http://zhanglab.ccmb.med.umich.edu/I-TASSER/ or the Phyre2 protein fold recognition server, www.sbg.bio.ic.ac.uk/phyre2/. All predictions were done using the standard parameters.

Clostridium difficile strains and growth conditions

Clostridium difficile strains were grown anaerobically in a microaerobic cabinet (Don Whitley DG 250) at 37°C in pre-reduced 3% Bacto Tryptose (Difco), 2% Yeast extract (Difco) and 0.1% thioglycolate (pH 7.4) medium (TY) or Brain Heart Infusion broth (Oxoid) supplemented with 0.5% yeast extract and 0.01% L- cysteine (Sigma) (BHIS) (Bakker et al., 2012. PLoS. One. 7: e43247) or minimal medium broth (Cartman and Minton, 2010. Appl. Environ. Microbiol. 76: 1103- 1109). When required, the broths were supplemented with appropriated antibiotics. Mid-logarithmic growth phase pre-cultures (OD600 0.4-0.8) were used to inoculate pre-reduced TY broth to a starting OD600 of 0.05 (± 0.01).

Optical density readings and samples for total toxin quantification were taken hourly in the exponential growth phase and at 12 and 24 hours post inoculation.

Proteomic analysis of the Clostridium difficile secretome

For analysis of the exoproteome, 12 ml minimal medium from C. difficile strain 630 grown to early stationary phase was collected. Intact bacterial cells were removed by centrifugation 10 min. at 7000xg. The resulting supernatant was filtered first through a 0.45 μιη filter (Whatman, FP 30/0.45), followed by a 0.2 μιη filter (Whatman FP 30/0,2). The filterered supernatant was subsequently concentrated to 0.5 mL using a Amicon Ultra- 15 Centrifugal Filter Unit NMWL 3.000 (Millipore). 20uL of this concentrated sample was run on a SDS-page gel, Coomassie stained and resulting bands analysed. In-gel digestion and LC-ion trap MS/MS analysis was performed as described previously (Zauner et al., 2012. J. Proteome. Res. 11: 5804-5814). Data was searched against the Clostridium difficile strain 630 database using the Mascot search algorithm (Mascot 2.2, Matrix Science). An MS tolerance of 0.6 Da (with # 13C=1) and a MS/MS tolerance of 0.5 Da were used. Trypsin was designated as the enzyme and up to one missed cleavage site was allowed. Carbamidomethylcysteine was selected as a fixed modification and oxidation of methionine as a variable modification. Only significant protein hits with at least two unique peptides with a score above 30 were selected.

PCR Detection to determine the genetic prevalence of CD2830

The PCR reactions to determine the genetic prevalence of the CD2830 gene were performed with GoTaq polymerase, according to the instructions of the

manufacturer with the forward primer AGGGATGGGAAGGTACTGGA and the reverse primer GTTTGTGGACAAGCTGATTTTAACT. RESULTS

In order to identify proteins secreted by C. difficile strain 630 (Ribotype 012), we analysed the culture medium at early stationary growth phase, using SDS-PAGE followed by in-gel tryptic digestion and reversed phase LC-ion trap MS/MS. This approach identified 124 proteins in total. It is well known that such a proteomic subfraction often contains not truly secreted proteins, but for example also major cytoplasmic proteins (Boetzkes et al., 2012. Arch Microbiol 194: 675-687). Indeed, SignalP analysis (Petersen et al., 2011. Nat Methods 8: 785-786) showed that only 39 (31%) of the identified proteins contain a signal export sequence. A signal sequence directs proteins to the outside of the cell but for release into the extracellular environment, cleavage of the signal sequence is required.

Furthermore, cell wall binding motifs or a lipid anchor can retain the secreted proteins at the surface of the bacteria. We therefore analysed the 39 signal sequence containing proteins for signal cleavage sites, lipid anchors and cell wall binding motifs using SignalP, LipoP, PSORTB Server and NCBI CDD (see table 1). Surprisingly, only a few proteins seemed to be truly cleaved and released (i.e. missing membrane anchors) and the appearance of the other proteins in the cell culture medium is probably related to cell surface shedding. One of the predicted genuinely secreted proteins, the uncharacterized protein CD2830 (Q183R7), seems highly expressed based on the number of identified unique peptides (12) and overall sequence coverage (50%). In addition, the MS/MS data of the secreted CD2830 protein indicated that the signal sequence had been cleaved off at the predicted position 26 (at sequence AHA | DSTT, data not shown). This data prompted us to analyse CD2830 in more detail.

Comparing the CD2830 protein sequence with the NCBI non redundant protein sequences database using BlastP, resulted in identical top hits with other C. difficile strains. The first annotated protein resembling our query was Anthrax toxin A moiety lethal factor from Bacillus with 42% amino-acid identity. Because a structure of a protein is more conserved in evolution than its amino acid sequence, we validated the homology of CD2830 using the Phyre2 three- dimensional structure prediction server (Kelley and Sternberg, 2009. Nat Protoc 4: 363-371). This analysis showed with 100% confidence, over 97% alignment coverage and 18% amino-acid identity (Figure 1A), similarity to the catalytic domain of Anthrax toxin lethal factor (ALF), residues 575-771 (Pannifer et al., 2001. Nature 414: 229-233). Among the identical residues is the well conserved HExxH motif, in which the two histidine residues coordinate the zinc ion in the active site, common to zinc metalloproteases (Figure 1A).

Three observations can be made with respect to the CD2830 fold in comparison to ALF. First, the structure of ALF comprises four structural domains (I, II, III and IV) including an N-terminal domain that binds the membrane-translocating component (domain I, Figure IB). Structural similarity with CD2830 only extends to the C-terminal domain containing the metalloprotease domain

(domain IV, Figure IB). Secondly, ALF is a metalloprotease that cleaves MAP kinase kinase enzymes and the structure of ALF in complex with this target peptide has been solved (Pannifer et al., 2001. Nature 414: 229-233; Turk et al., 2004. Nat Struct Mol Biol 11: 60-66). ALF amino-acids making substrate specific contacts are assigned but none of these contacts are identical to CD2830 (Figure 1A, arrows). Identical residues between CD2830 and ALF mostly correspond to internal, structural fold determining residues. Finally, ALF contacts with the MAPKK peptide are also present in domain III that forms a recognition groove together with the metalloprotease catalytic domain. This domain III is absent in CD2830. In summary, this indicates that CD2830 is a secreted functionally active zinc metalloprotease, with a fold similar to ALF, but with a different mechanism of action (lacks membrane translocating domain) and different targets.

To determine the distribution of the cd2830 gene among the genomes of

Clostridium difficile, we evaluated the genomic DNA of 30 strains, encompassing the 6 major lineages (Knetsch et al., 2012. Infect Genet Evol 12: 1577-1585), by PCR using specific primers for CD2830. We found that in all strains tested the cd2830 gene is present (data not shown), emphasizing the importance of this gene throughout the species. Table 1

Example 2

EXPERIMENTAL PROCEDURES

Chemicals

Collagen I from Life technologies (cat. no. A1048301 ), Mucin from Sigma-Aldrich (cat.no. M3895), IgA from VWR international (cat.no. 401098), Fibrinogen from (EMD Millipore). Recombinant Hsp906 (cat. no. SPR- 102A) and recombinant HSP90a (cat. no. SPR- 101A) were from Sanbio. Sequencing Grade Modified Trypsin (Promega). Peptides were synthesized at the LUMC-facility as described previously (Hiemstra et al., 1997. Proc Natl Acad Sci USA 94: 10313-10318).

Bacterial expression constructs

To construct a hislO-tagged CD2830 expression plasmid, the CD2830 sequence was amplified by PCR from C. difficile strain 630 genomic DNA, using specific primers, AGGGAATCATATGGATAGTACTACTATACAACAAAATAAAGACAC (forward) and TATTGGATCCCTATTTAGCTAAATTTTGCAAAAAGC (reverse). The PCR products were digested with Ndel and BamHI and ligated into vector pET16b (Novagen) similarly digested with Ndel and BamHI. This resulted in the construction of CD2830 expression vectors containing 10 consecutive histidines at its N-terminus replacing the signal sequence.

To construct His-tagged CD2831, a similar approach was used, using primers, AGTTCCATATGAAGCAAGGTTATGCTTTTGAAGC (forward) and

TCTGGCTCGAGCTATGTAGTACTATCCCCTGTTTTTGG (reverse), which were Ndel and Xhol cloned in pET16b, resulting in an expression construct containing aa 732 to 947 of CD2831.

Purification of CD2830 and CD2831

The bacterial expression constructs were transformed into the Escherichia coli DE3 C43 strain (Lucigen, USA ) and a single colony cultivated on a rotary shaker (200 rpm) in 100 ml Luria Broth at 37°C until an OD600 of 0.5, after which the cultures were induced with 1 M isopropylthio-6-galactoside for 5 hours, using ampicillin (50pg/mL) as a selection marker. The E. coli cells were collected by centrifugation at 6000 x g. We followed the protocol for the preparation of cleared E. coli lysates under native conditions from QIAGEN as described in the fifth edition of the QIAexpressionist with the following modifications: the lysates were sonicated 3x10 seconds at 22 kHz, the pH of the lysis buffer, wash and elution buffer were set at 7.4. The E. coli lysates (50 mM sodium phosphate buffer, pH 7.4, 5 mM 2-mercaptoethanol, 0.1% NP40, 300 mM NaCl) containing histidine- tagged proteins were loaded on a 1 mL HisTrap HP column (GE Healthcare). The column was washed with 20 mL wash buffer (50 mM sodium phosphate buffer pH 7.4, 300 mM NaCl, 5 mM 2-mercaptoethanol, 5% glycerol, 20 mM Imidazol). The His-tagged proteins eluted at 150 mM imidazole while using a 25 mL linear gradient ranging from 20 to 250 mM imidazole.

Proteolytic assays with CD2830

CD2830 proteolytic assays were performed at 37°C in phosphate buffered saline (PBS) [pH 7.4], 0.5 mM ZnC12, unless otherwise stated, for a duration of 16 hours. Reactions were stopped by the addition of Laemmli loading buffer (62.5 mM Tris-HCl, pH 6.8, 25% glycerol, 2% SDS, 5% 6-mercaptoethanol or 350 mM dithiothreitol and 0.01% bromophenol blue) and by heating the samples at 95°C. Proteolytic assays with the single targets were performed with 0.5 μg of rCD2830 and 5 μg of Fibrinogen, 5 μg of Mucin, 5 μg of IgA, 3 μg of HSP90 a and 3 μg of HSP90 6. For the proteolytic assay with a Caco-2 lysate, the lysate was prepared as described in the manual of Pierce IP lysis buffer (87787) from Thermo scientific. 10 pg lysate was incubated with 1 μg of rCD2830. Protease activity was visualized by SDS-PAGE electrophoresis of the samples followed by Coomassie staining.

The peptide cleavage assays were performed with 20 pmol peptides and 0.1 μg of rCD2830. After overnight incubation at 37 °C, samples were analysed with MALDI-ToF-MS (Ultraflex II, Bruker Daltonics) using dihydroxybenzoic acid as a matrix. Identification of CD2830 cleavage products with direct infusion Q-ToF MS

For direct identification of the CD2830 cleavage products from HSP90P and recombinant CD2831, proteolytic digestions were performed as described above but samples were subsequently desalted on a reverse phase cartridge (C-18 Oasis HLB lcc 30 mg, Waters). Peptides were eluted using 50% acetonitrile, 0.1% formic acid and then directly infused to the Q-TOF-MS (maXis, Bruker Daltonics) with a syringe pump (flow rate 30 μ/h). For mass spectrometric analysis, ions were generated using a Captive Spray (Bruker Daltonics) at a spray voltage of 1.4 kV. The temperature of the heated capillary was set to 180°C. Data were recorded for 5 minutes (mass range m/z 300 to 1400).

LC-ion trap MS/MS analysis

Peptides were injected onto a trap column equilibrated with 0.1% formic acid

(Acclaim PepMap lOO ΙΟΟμιη x 2cm, nano Viper C18, 5μιη, 100 A, Thermo

Scientific) and separated on a analytical reverse phase column (Acclaim PepMap

RSLC 75 μιη x 50 cm, nanoViper C18, 2μιη, 100A, Thermo Scientific) coupled to an ion trap mass spectrometer (amaZon, Bruker Daltonics) at the flow rate of 300 nl/min. The liquid chromatography was performed with piece-linear gradient (0- 10 min at 2% B, 10-25 min to 5% B, 25-165 min to 25% B, 165-175 min to 30% B, 175-190 min to 35% B, 190-195 min to 100% B, 195-210 min at 100% B, were B is 95% acetonitrile/0.1% formic acid). For mass spectrometric analysis, ions were generated using a Captive Spray (Bruker Daltonics) at a spray voltage of 1.4 kV. The temperature of the heated capillary was set to 180°C. Identification of the CD2830 cleavage site motif

The combination of short synthetic peptides and MALDI-ToF MS analysis, allowed a fast screening of the optimal reaction conditions for CD2830 activity as well as cleavage site specificity. Varying assay conditions such as pH,

temperature and the presence of a metal ion chelating agent (Figure 9) showed a robust enzyme activity between 30 and 70°C, pH 6-9 having sensitivity for EDTA. Based on these results, we decided to perform subsequent cleavage reactions at pH 7.4, 37°C, 0.5 mM Zn2+. For the analysis of try tic digests, eluting peptides were analyzed using the data dependent MS/MS mode over a m/z 300-1400 range. The 10 most abundant ions in an MS spectrum were selected for MS/MS analysis by collision-induced dissociation using helium as the collision gas.

For the analysis of CD2830 cleavage products, a targeted mass spectrometric approach was applied. For this purpose, MS/MS data acquisition was

continuously performed on a single or a few m/z values corresponding to the expected peptide(s) during the whole LC-MS run.

RESULTS

To investigate the predicted metalloprotease activity of CD2830, we cloned the cd2830 gene into a bacterial expression vector. The histidine tag containing recombinant CD2830 protein (rCD2830) was overexpressed in E.coli and purified using a nickel affinity column. Although CD2830 lacks a structure known to be associated with membrane translocation, we tested toxicity of purified rCD2830 on monolayers of human epithelial colorectal Caco-2 cells, but did not observe any adverse effect on cell viability, tested at various concentrations of rCD2830 (l-100 g/mL) and incubation times (1 -24 hrs.).

We then focused on the identification of substrates of CD2830. Because we had a priori no indication as to whether CD2830 would target a C. difficile or host protein, we considered both options plausible. First of all, we tested whether known extracellular substrates of other bacterial proteases, Mucin, Fibrinogen, IgA and collagen (Mistry and Stockley, 2011. Virulence 2: 103-110; Ashida et al., 2012. Nat Chem Biol 8: 36-45; Ohbayashi et al., 2011. Microbiology 157: 786-792; Govindarajan et al., 2012. PLoS One 7: e32418), could also be substrates for recombinant CD2830. However, when we performed incubations under

proteolytic assay conditions comparable to those described for ALF (Kim et al., 2004. Biochem Biophys Res Commun 313: 217-222), we did not observe any

CD2830-mediated cleavage of any of these substrates (data not shown). As a next step we incubated total C. difficile and Caco-2 cell lysates with and without rCD2830 and analysed these samples by SDS-PAGE. No difference between the rCD2830 treated and untreated C. difficile samples was apparent (data not shown). However, when the Caco-2 lysate was incubated with rCD2830, a clear cleavage product of approximately 85 kDa was noticeable (illustrated with the arrow in Figure 2 A). As a negative control we incubated the lysate at 4°C to ensure that the appearance of this band was not due to proteolytic activity in the Caco-2 lysate. We subsequently analysed this 85kDa cleavage product using reversed phase LC-ion trap MS/MS after in-gel tryptic digestion. Using this approach, several proteins were identified but, based on peptide counts, the major protein was Heat Shock Protein 906 (data not shown).

To confirm that HSP906 was cleaved by rCD2830, we performed a proteolytic assay with purified HSP906 and also included the HSP90a isoform, which has 93% sequence identity to HSP906. As can be seen in figure 2B (left panel), purified HSP906 is cleaved by rCD2830, which correlates with the proteolytic assay with the Caco-2 lysate. Surprisingly, purified HSP90a was not cleaved after incubation with rCD2830 (Figure 2B, right panel).

The cleavage product observed with HSP906, is approximately 3-5 kDa smaller compared to the intact protein (Figure 2B, left panel). This product therefore must be derived from a cleavage event near either the N- or C-terminus of the protein. In order to determine the precise cleavage site, we employed two strategies. First, we compared the LC-MS/MS analysis of the in-gel tryptic digests of the rCD2830-cleaved and the uncleaved HSP906. Secondly, we captured the putative small peptide fragment after rCD2830 cleavage and analysed this by mass spectrometry. Both approaches demonstrated that HSP906 is cleaved by rCD2830 between alanine 701 and 702 (AEEPNA | AVPDE).

Because both homologous alanines are also present in the HSP90a sequence (Figure 2C) which is not susceptible to CD2830 cleavage, this indicates that surrounding amino acid residues are important for CD2830 specificity. To determine whether CD2830 is also capable of cleaving small peptides, we generated a synthetic peptide (KAAEEPNAAVPDEIK) based on the identified cleavage site of HSP906 and incubated this with rCD2830. Using MALDI-ToF MS analysis, we indeed observed cleavage of this peptide and the products at m/z 771.41 [M+H]+ (AVPDEIK) and 829.40 [M+H]+ (KAAEEPNA) confirmed the

CD2830 cleavage between the two alanine residues (Figure 2D). The other visible peaks within the spectrum correspond to the sodium and potassium adducts of the same cleavage products. When we assayed an HSP90a peptide containing the homologous sequence surrounding the cleavage site, DDTSAAVTEEM (see Figure 2C), no cleavage was observed (data not shown), again demonstrating the specificity of CD2830 for the C-terminal region of HSP906.

We next determined the substrate cleavage specificity by permutating the six positions flanking the cleavage site (P, N, A | A, V, P) within the peptide described above (KAAEEPNAAVPDEIK) to each of the 20 possible amino acids. Each peptide was synthesized and individually incubated for 16hrs with rCD2830 and then measured by MALDI-ToF MS to determine whether cleavage had occurred (data not shown). Amino acid residues in a substrate undergoing cleavage are commonly designated P3, P2, P I | PI', Ρ2', P3'. In brief (Figure 3A), the screening of the 120 synthetic peptides demonstrated that the most stringent position is P3', i.e. only peptides with a proline at this position are cleaved. The two positions flanking the scissile bond, A | A, only tolerate replacement by a proline. The P2' position can, in addition to valine, also accommodate a proline or an isoleucine. Less stringent requirements were observed for positions P3 and P2, where most amino acids are tolerated although cleavage efficiencies did vary (data not shown). The absolute requirement of a proline at position P3' explains why HSP90a (threonine at P3', see Figure 2C) was not cleaved.

Example 3

CD2830 substrate cleavage sites are found in C. difficile LPXTG adhesion molecules To find other potential targets of CD2830, we performed a ScanProsite search (Castro et al., 2006. Nucleic Acids Res 34: W362-W365) of the human as well as the C. difficile proteome using the above described cleavage motif ([P/A] T

[P/A][V/P/I] P). As expected for a small motif, we found ~600 potential human extracellular target proteins. Several of these cellular proteins may play a role in defence against bacterial infections (not shown).

In contrast to the human proteome, within the C. difficile strain 630 proteome we found only four potential signal sequence containing targets: CD0515 (D-alanyl- D-alanine carboxypeptidase), CD2831 (Adhesin), CD3043 (a putative

transglutaminase) and CD3246 (surface protein, putative collagen binding).

Remarkably, CD2831, located next to CD2830 in the genome, and CD3246 contain six and seven consecutive cleavage sites, respectively (Figure 3B). The region containing the cleavage sites seems conserved between CD2831 and CD3246 (shaded areas in Figure 3B). Moreover, CD2831 and CD3246 have features common to LPXTG cell wall anchored proteins and the protease cleavage sites are directly adjacent to this peptidoglycan anchor motif. The multitude of target sites suggests an effective cleavage of these two cell surface proteins and based on the 13 potential cleavage sites in these two proteins a consensus motif was constructed (Figure 3C). Apart from the obvious similarity between the cleavage motif deduced from the peptide library screen (Figure 3A) a striking overall proline rich motif becomes apparent and preferences for a proline a P4' and asparagine at P2 seem plausible. To confirm cleavage of CD2831 (adhesion protein, Figure 4A) by CD2830

(protease), we produced part of CD2831 as a recombinant protein and tested it for cleavage. As shown in Figure 4B, within 5 minutes incubation with rCD2830, all recombinant CD2831 was cleaved. Therefore, this cleavage is much more efficient than observed for HSP906 (Figure 2B). From the gel it was not obvious to deduce which of the six putative cleavage sites are cleaved. We therefore also analysed the CD2831 cleavage products by Q-ToF MS. This showed that indeed each predicted CD2830 cleavage site within recombinant CD2831 was cleaved (Figure 4C).

Example 4

EXPERIMENTAL PROCEDURES

Generation of Clostridium difficile CD2830 knockout strain

We generated an isogenic CD2830 mutant by insertional inactivation of the CD2830 gene in the wild type strain 630AErm using ClosTron technology (Heap et al., 2010. Methods Mol Biol 646: 165-182; Heap et al., 2007. J Microbiol Methods 70: 452-464). Briefly, the Perutka algorithm on the ClosTron website (http://www.clostron.com) was used to design primers for retargeting the Group II intron (Sigma; Targetron). The retargeted intron was cloned using the restriction enzymes BsrGI and Hindlll into plasmids pMTL007C-E2 and transformed to E.coli CA434 and transferred by conjugation into the wild type strain 630AErm (Purdy et al., 2002. Mol Microbiol 46: 439-452; Hussain et al., 2005. J Med Microbiol 54: 137-141). The selection of C. difficile transconjugants was done by subculturing on pre-reduced BHIS agar supplemented with thiamphenicol (Sigma; 10pg/mL) and C. difficile selective supplement (Oxoid). This was followed by several rounds of subculturing on pre-reduced BHIS agar supplemented with lincomycin (Sigma; 20 pg/mL) and C. difficile selective supplement to promote integration of the GroupII intron into the CD2830 gene. Chromosomal DNA isolated from the transconjugants using a QIAamp blood kit (Qiagen) was used in conventional PCRs and sequence runs to confirm the disruption of CD2830 and the nucleotide position of the insertion in the CD2830 gene.

The genotype of the disruption was confirmed with conventional PCRs using the primer oDB0029 [5'-ATGAGACCAAGTAAAAAATT] and the EBS universal primer (5'-CGAAATTAGAAACTTGCGTTCAGTAAAC) and with primer pairs oDB0029 and oDB0030 [CTATTTAGCTAAATTTTGC A] , flanking the ClosTron insertion site. Sequence analysis confirmed that the disruption was at the expected site in CD2830 gene (data not shown). In addition, Southern blot analysis using an ermB specific probe clearly confirmed a specific single insertion of the Group II intron in the genome (data not shown).

RESULTS

Clearly, the CD2830 protease is very active and specific in vitro. To elucidate its potency and role in vivo, we generated a knockout of CD2830 (CT::CD2830) in Clostridium difficile strain 630AErm using the Clostron system (Heap et al., 2010. Methods Mol Biol 646: 165-182). For phenotypic characterisation, we first determined the growth rates of the WT and CT::CD2830 strain in TY medium and observed no differences demonstrating that the CT::CD2830 strain has no growth defect in vitro (Figure 10). We then sought to test whether we could measure endogenous CD2830 activity in the medium of growing cells and whether such activity was only observed in WT cells compared to CT::CD2830 cells. We therefore grew WT and CT::CD2830 C. difficile in minimal medium (Cartman and Minton, 2010. Appl Environ Microbiol 76: 1103-1109) to the beginning of the stationary phase and collected the medium after high speed centrifugation (20 min. 20.000xg) and filtration (0.2μιη). To this conditioned medium, we added a synthetic peptide containing a CD2830 cleavage site identified in CD2831 (KDTIVINPPVPPSEK). As a positive control we used fresh medium to which we added rCD2830 protein. As shown in Figure 5 both the recombinant protein as well as activity in the WT secretome cleaves the peptide. The knockout CT::CD2830 medium, in contrast, showed no peptide cleavage activity. This shows that the CD2830 protease is functionally secreted and is the only activity within the C. difficile secretome that can cleave a CD2831 peptide containing the CD2830 target site.

Example 5

EXPERIMENTAL PROCEDURES

lxlOE9 Clostridium difficile cells were collected by centrifugation 10 min. at 7000xg and washed 3 times in PBS containing complete, EDTA-free Protease

Inhibitor Cocktail (Roche). Cells were subsequently resuspended in 0.5 mL of the same buffer containing 0.5 mM ZnC12. 1 μg of rCD2830 was added and the suspension incubated 1 hr. at 37°C. After this cells were spun 20 min. 20.000xg. The resulting supernatant was collected and filtered through a 30 kDa cut-off filter (Microcon, Ultracell YM-30, Millipore). Samples were subsequently desalted on a reverse phase cartridge (C- 18 Oasis HLB lcc 30 mg, Waters) and peptides were eluted stepwise using 20 and 50% acetonitrile in 0.1% formic acid, respectively. Prior to LC-MS/MS analysis, acetonitrile was evaporated in a vacuum concentrator. Similarly, 10 ml of conditioned minimal growth medium was desalted and eluted on a reverse phase cartridge for analysis of peptides in the medium.

RESULTS

Having established that CD2830 is actively secreted and one of its targets is CD2831, a cell surface located adhesion molecule anchored to the cell wall through a LPXTG motif, we investigated whether rCD2830 protease can actively remove this adhesion molecule from the surface of C. difficile cells. For this purpose, we performed "cell surface shaving" experiments using C. difficile cells and rCD2830, in analogy to proteomic approaches which have been developed to study cell surface proteomes in general (Ythier et al., 2012. Mol Cell Proteomics 11: 1123- 1139.; Solis and Cordwell, 2011. Proteomics 11: 3169-3189). Following incubation of the C. difficile cells with rCD2830 for 1 hr, cells were pelleted by centrifugation, the resulting supernatant collected and peptides purified on a C- 18-cartridge. As a positive control, recombinant CD2831 was incubated with rCD2830 protease and resulting peptides similarly purified. The peptides were analysed by reversed phase LC-ion trap MS/MS. During these analyses, we focused our attention to one of the expected CD2831 product peptides

(PAPPNTDEPIVNP, Figure 4C) by continuously performing MS/MS analysis at m/z 680.8 [M+2H]2+, corresponding to this peptide. We first analysed the recombinant CD2831 sample treated with rCD2830 protease. The rCD2831 product peptide PAPPNTDEPIVNP eluted around 100 min as shown by the summed extracted ion chromatogram of fragment ions at m/z 1032.5, 1131.5 and 1245.6, characteristic for this peptide (Figure 6A, upper part). Surprisingly, we did not observe this peptide when we analysed the WT C. difficile cells treated with rCD2830 protease (Figure 6A, middle part) while a clear signal was apparent in the sample from the CT::CD2830 cells treated likewise (Figure 6A, lower part, left). The MS/MS spectrum confirmed the identity of the CD2831- peptide that was "shaved" from the CT::CD2830 cells (Figure 6a, lower part, right). Overall, this data shows that rCD2830 can cleave CD2831 from the cell surface of C. difficile cells. Moreover, the fact that we only observed this when we used CT::CD2830 cells, implies that the endogenous CD2830 from WT cells (Figure 5) is responsible for cleaving CD2831 during culturing of the cells, resulting in the loss of cell surface expression of CD2831 and concomitant inability to identify it in the "cell shaving" experiments. Obviously, this suggests that CD2831 product peptides, resulting from CD2830 cleavage, would be present in the medium of cultured WT cells but not in that of CT::CD2830 cells. To test this, we grew cells in minimal medium and collected the medium after O/N culturing and analysed it as described above. Indeed, we observed the CD2831 product peptide PAPPNTDEPIVNP in the culture medium of WT cells but not in that of CT::CD2830 cells (Figure 6B). Of note, CD2831 was also observed on the gel from our secretome analysis (see Table I).

Example 6

EXPERIMENTAL PROCEDURES

Caco-2 cells were grown in RPMI (GE healthcare), 10% fetal calf serum, Penicilin and Streptavidin in a humidified 5% CO2 / 95% air atmosphere at 37°C in a 75 mL culture flask until the cells were nearly confluent.

Adhesion of C. difficile to Caco-2 cells (24 wells cell culture plates, Greiner, cat.no. 662160) and collagen coated plates (24 wells, Life technologies cat.no.

A11428) was performed essentially as described previously (Cerquetti et al.,

2002. FEMS Immunol Med Microbiol 32: 211-218) with few modifications.

Collagen coated plates were pre-incubated with PBS, 0.05% Tween-20 for 2hrs in a microaerobic cabinet (Don Whitley DG 250) at 37°C. Confluent monolayers of Caco-2 cells were washed twice in PBS before transfer to the cabinet. To each well lxlOE6 bacteria were added in 200 μΐ pre-reduced PBS. After 3 hrs incubations, wells were washed 5 times in PBS. Adherent bacteria were released in 100 μΐ lxTrypsin (PAA laboratories, Cat. No. Ll l- 001) before counting cells.

RESULTS

We have shown above that the endogenously secreted CD2830 protease cleaves the cell surface adhesion molecule CD2831 from WT cells during culturing. On the contrary, the CT::CD2830 cells retain surface adhesion molecules. Conserved domain analysis (Figure 4A) as well as structural fold predictions suggested that CD2831 contains a 'Collagen Hug' motif. We therefore speculated that the

CT::CD2830 cells would show enhanced adhesion, particularly to collagen. Hence, we assayed adhesion of C. difficile WT and CT::CD2830 cells to Collagen (type I) coated plates. Adhesion of CT::CD2830 cells to collagen was indeed ~3 fold higher than the wild type strain (Figure 7). We also tested adhesion of C. difficile WT and CT::CD2830 cells to a monolayer of Caco-2 epithelial cells as described previously (Cerquetti et al., 2002. FEMS Immunol Med Microbiol 32: 211-218). We observed no difference in adhesion of WT and CT::CD2830 to Caco-2 cells (Figure 7). Possibly, adhesion to Caco-2 cells involves additional C. difficile adhesion molecules (Fagan et al., 2011. J Med Microbiol 60: 1225-1228) which are not substrates for CD2830. In addition, the extracellular matrix protein(s) bound by CD2831, and possibly CD3246, may not be highly expressed on Caco-2 cells.

Example 7

EXPERIMENTAL PROCEDURES

Hamster infection model

All hamster infection studies were carried out in accordance with UK Home Office requirements and all experiments approved by the University of

Nottingham ethics committee.

A block design with final group sizes of 7 animals was used for this study. Female Golden Syrian hamsters of between 100-120 g were obtained from Charles River UK. Hamsters were housed singly in individually ventilated cages. Each hamster was given clindamycin (30mg/kg) via the oral route five days prior to being infected orally with 10,000 spores of either strain 630Aerm or CT::CD2830.

Hamsters were monitored for signs of infection and euthanised when a predetermined end point was reached. Hamsters were monitored 4-5 times per day following infection and were assessed for several parameters including presence and severity of wet tail, weight loss, level of activity, piloerection, sunken eyes, hunched posture and response to stimulus. In order to quantify changes in the condition of the animal a scoring system based of the severity of changes observed (ranging from 0-3 for each parameter) was used, with the animals being euthanised when a pre-determined cumulative value was reached. Faecal pellets were collected daily and upon euthanasia a caecum sample was taken from each hamster. Samples were homogenised and plated to determine the presence of C. difficile, PCR was performed to determine the genotype of each strain recovered from the hamsters. Samples were heat treated (55 °C for 30 minutes) and plated on to fructose agar (C. difficile agar base. Oxoid) supplemented with cycloserine (Oxiod), cefoxitin (Oxoid), taurocholate (Sigma) and amphotericin (Sigma) to select for C. difficile. The following primers were used to determine the genotypes of the recovered strains:

oDB-0029 5'-ATGAGACCAAGTAAAAAATT and

oDB-0030 5'-CTATTAGCTAAATTTTGCA.

RESULTS

Having characterized CD2830 in vitro, we tested the virulence of the mutant strain in a hamster infection model (Kuehne et al., 2010. Nature 467: 711-713). Hamsters were individually challenged with 10.000 spores of a C. difficile strain,

5 days after an oral dose of clindamycin. Both WT and CT::CD2830 cells were administered to eight hamsters in total. All hamsters were colonized by C.

difficile and all but one developed symptoms of C. difficile infection.

The mean time of infection to end-point (death) for WT was ~63 hrs. In contrast, the mean time of the CT::CD2830 strain was 129 hrs (Figure 8), demonstrating an attenuated virulence for this mutant strain (p<0.05). Example 8

EXPERIMENTAL PROCEDURES

AH incubations were performed at 37°C in ΙΟΟμΙ _^ phosphate buffer saline pH 7.4 (Braun Melsungen), 0.5 mM ZnC12 and 50μΜ fluorogenic peptide Dabcyl- EVNPPVPD-Edans in a μΟΙΛϊΑΕν plate, black 96 well (Greiner Bio-one). Various amounts of recombinant CD2830 protein (indicated) were added to initiate reaction (time = 0, x-axis). Also, growth medium from WT cells and CD2830 knockout cells was incubated with the fluorogenic peptide. Progressive cleavage (y-axis) was monitored by measuring fluorescence at 485 nm in a Mithras LB 940 Multimode Microplate Reader (Berthold Technologies) after excitation at 350 nm.

RESULTS

See figures 11A and 11B. A FRET (Fluorescence Energy Transfer) assay for in vitro enzymatic activity of CD2830. A dose-dependent response curve was obtained with recombinant CD2830 protein (Figure 11A). Cleavage was also obtained with endogenous CD2830 protease from WT cells secreted in the growth medium, while no cleavage was observed with growth medium derived from CD2830 knockout cells (CD2830KO) (Figure 11B), demonstrating that CD2830 is the only activity secreted by C. difficile cells that can cleave the peptide.

Example 9

Materials and methods

See Example 2

Results

Within a core Dabcyl-EVNP | PVPD-Edans FRET peptide, amino acids were permutated to other amino acids as indicated. Initial cleaving velocity of all peptides was individually measured and compared to the peptide Dabcyl- EVNP I PVPD-Edans, which was set at 100%. The initial cleaving velocity was measured as long as substrate cleavage (determined by the increase in fluorescence) was linear. In the upper right corner of each inset shown in Figure 12, the permutated position is indicated and the X-axis shows which amino acid replaced the original amino acid within the core FRET peptide. Based on the results shown in Figure 12, we conclude that EVNPPPPD is the optimal cleavage site for CD2830. To increase the sensitivity of our assay, we synthesized this peptide where we replaced the fluorescent Edans group for Alexa 488: Dabcyl-EVNPPPPD-Alexa488. Setting for detection are Excitation filter 485nm; Emission filter 535 nm. With this peptide we are able to identify 0.1 recombinant CD2830.

Example 10

A hydroxamic acid-based metalloprotease inhibitor was designed, which is predicted to block the active site of CD2830 by binding to the catalytic Zn2+ ion through a chelating hydroxamate. The use of a chelating hydroxamate for inhibitors of metalloproteases has been described in Jialiang Hu et al., 2007. Nature Reviews Drug Discovery 6, 480-498. The inhibitor comprises the CD2830 cleavage half site plus hydroxamate i.e. DTIVINP-hydroxamate for P4-P3-P2-P1-NHOH. In addition three mutated hydroxamate peptides were designed as controls for the specificity of the inhibitor. The half site is thought to target the inhibitor peptide to the active site where the hydroxamate will interact with the zinc ion (/displace H20) and block the active site.

The inhibitors used are:

1 DTIVINP-hydroxamate

2 DTIVGGP-hydroxamate

3 DTIVGGL-hydroxamate

4 DTIVGNL-hydroxamate As is shown in Figure 13, the DTIVINP-hydroxamate, corresponding to the CD2830 recognition half-site showed inhibition, whilst control peptides

DTIVGGP-hydroxamate and DTIVGGL-hydroxamate did not. However, the DTIVGNL-hydroxamate, containing a Leucine at the PI position showed an even more efficient inhibition of the cleavage reaction. We speculate that the Leucine at the PI position might reflect a post-cleavage conformation. In addition, comparison of DTIVGGL-hydroxamate (no inhibition) and DTIVGNL- hydroxamate (efficient inhibition) emphasizes the importance of Asn at P2. Selectivity of inhibition by hydroxamates DTIVLNP and DTIVGNL was confirmed by showing a lack of inhibition on human matrix metalloproteinase-2 (gelatinase A) enzyme activity (data not shown).