Proteomic analysis of cell surface proteins from Clostridium difficile

REGULAR ARTICLE

Proteomic analysis of cell surface proteins from

Clostridium difficile

Anne Wright1, Robin Wait2, Shajna Begum2, Ben Crossett1, Judit Nagy1,Katherine Brown1 and Neil Fairweather1

1 Centre for Molecular Microbiology and Infection, Department of Biological Sciences,Imperial College London, UK

2 Kennedy Institute of Rheumatology Division, Faculty of Medicine, Imperial CollegeLondon, UK

Clostridium difficile is a bacterium that causes disease of the large intestine, particularly aftertreatment with antibiotics. The bacterium produces two toxins (A and B) that are responsible forthe pathology of the disease. In addition, a number of bacterial virulence factors associated withadhesion to the gut have previously been identified, including the cell wall protein Cwp66, thehigh-molecular weight surface layer protein (HMW-SLP) and the flagella. As the genomesequence predicts many other cell wall associated proteins, we have investigated the diversity ofproteins in cell wall extracts, with the aim of identifying further virulence factors. We have used anumber of methods to remove the proteins associated with the cell wall of C. difficile. Two of theresulting extracts, obtained using low pH glycine treatment and lysozyme digestion of the cellwall, have been analysed in detail by two-dimensional electrophoresis and mass spectrometry.One hundred and nineteen spots, comprising 49 different proteins, have been identified. Thetwo surface layer proteins (SLPs) are the most abundant proteins, and we have also found com-ponents of the flagellum. Interestingly, we have also determined that a number of paralogs of theHMW-SLP are expressed, and these could represent targets for further investigation as virulencefactors.

Received: September 1, 2004Revised: October 18, 2004

Accepted: December 1, 2004

Keywords:

Bacterial pathogenesis / Cell wall proteins / Clostridium difficile / Surface proteins

Proteomics 2005, 5, 2443–2452 2443

1 Introduction

Clostridium difficile is a gram-positive anaerobic bacteriumthat commonly causes enteric disease in humans particular-ly after treatment with antibiotics, which disrupt the gut flora

allowing C. difficile to colonise and multiply [1]. Two toxins(TcdA and TcdB) are produced that cause severe tissue dam-age and result in the manifestation of the disease. As well assevere diarrhoea and abdominal pain, infection with C. diffi-cile can lead to further complications including pseudo-membraneous colitis, inflammation and ulceration of thelining of the intestinal wall [2, 3]. Commonly acquired inhospitals, C. difficile can result in extra costs of £ 4000 perpatient due to the extra time spent in hospital, as a result ofthe treatment and tests required [4], amounting to an esti-mated $ 1.1 billion per year in the US alone [5].

Adhesion is an early, critical step in colonisation and thesubsequent disease process. To date, a number of virulencefactors involved in the adherence of C. difficile to the gut wallhave been identified and characterised, but the full mecha-

Correspondence: Dr. Neil Fairweather, Flowers Building, Centrefor Molecular Microbiology and Infection, Department of Biolog-ical Sciences, Imperial College London, Exhibition Road, LondonSW7 2AZ, UKE-mail: [email protected]: 144-20-7594-3069

Abbreviations: HMW, high-molecular weight; LMW, low molecu-lar weight; SLP, surface layer protein

2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.de

DOI 10.1002/pmic.200401179

2444 A. Wright et al. Proteomics 2005, 5, 2443–2452

Figure 1. Arrangement of genesaround slpA in the C. difficilechromosome. The region of thegenome of C. difficile 630 har-bouring a series of ORFs thatinclude slpA and its paralogs(ORFs 1–12) is shown. Paralogsshow amino acid homology tothe C-terminal domain (HMW –cell wall binding domain) of

SlpA and to the autolysin CwlB from B. subtilis, and are all potentially secreted. Black shading indicates regions of homology to the cellwall binding domain, and the non-homologous region is shown in white. Dark grey indicates unrelated ORFs. SlpA is post-translationallycleaved to release the signal sequence and two mature SLPs.

nism by which the bacteria are attached to the mucosa andbegin an infection remains to be elucidated. One adhesin isthe flagellum and its components [6]. C. difficile is known tohave peritrichious flagella, and the adherence of flagellatedstrains to mouse cecum is ten-fold higher compared to non-flagellated strains. In addition, flagellin, the flagellar tipprotein, and crude flagella can adhere to axenic mouse tis-sue [6].

Another adhesin is the high-molecular weight surfacelayer protein (HMW-SLP) [7]. C. difficile is surrounded by aparacrystalline array of protein, external to the cell wall,which forms the surface (S)-layer. The S-layer of C. difficile iscomposed of two proteins, derived from a single gene, slpA(Fig. 1) [7, 8]. The translated gene product undergoes tworounds of post-translational cleavage, firstly to remove thesignal sequence following secretion, and then internally atposition 355 to release the two mature SLPs of 45 kDa(HMW-SLP) and 36 kDa low-molecular weight surface layerprotein (LMW-SLP). Recombinant HMW-SLP is able toadhere to human and mouse intestinal tissue whilst anti-bodies against it are able to block adherence of C. difficile tocultured cells [9]. Interestingly, the HMW-SLP has severalparalogs throughout the genome, with 11 of these clusteredaround slpA (Fig. 1) and the others dispersed around the ge-nome [10]. Each paralog contains an HMW-SLP-like domain,which is thought to function in cell wall binding and whichhas homology to the amidase CwlB from Bacillus subtilis [11].The majority of paralogs found in C. difficile also contain asignal sequence and a second variable domain, which variesbetween paralogs but is conserved between strains. One ofthe paralogs, known as Cwp66 (Fig. 1), has been shown toplay a role in adherence [12].

To provide an insight into the pathways that C. difficileuses to secrete proteins, particularly during toxin production,the extra cellular proteome of C. difficile strain VPI 10463 wasrecently analysed [13]. In addition to the toxins TcdA andTcdB, a small number of proteins were found to be exportedinto the media including SlpA, SLP paralog 2, TolC (a chan-

nel forming protein) and XkdX (a holin-like protein). How-ever, this study was restricted largely to the secreted fractionand did not examine the cell wall fraction in detail.

The discovery of a number of adhesion factors suggeststhat there may be a whole consortium of proteins involved inthe attachment of C. difficile to the gut wall. Certain SLPparalogs, which are all potentially secreted, may have impor-tant functions in the attachment of the bacterium to hostcells and its subsequent survival. The genome sequence ofC. difficile has recently been completed (www.sanger.ac.uk/Projects/C_difficile) and the annotation project is ongoing.This has allowed us to begin a proteomic analysis of C. diffi-cile, in order to identify potential virulence factors, in partic-ular those proteins that are present in the cell wall fractionsof the bacterium.

In this study, we describe a number of methods used toextract proteins from the surface of the bacterium and eval-uate the number and abundance of proteins removed using2-DE and MS. This will facilitate further studies examiningthe adhesins of C. difficile to establish how this pathogen isable to colonise a host and cause disease.

2 Materials and methods

2.1 Bacterial strains and culture conditions

C. difficile 630 was grown in brain-heart infusion broth(Oxoid, Basingstoke, UK) at 377C in an anaerobic chamber inan atmosphere of 10% CO2/10% H2/80% N2 for 24 h.

2.2 Protein extraction

Standard SLP extraction was essentially as described in [14].C. difficile cells from a 24 h culture were harvested by cen-trifugation (4000 rpm for 15 min), washed once in PBS andresuspended in 0.04 volumes of 0.2 M glycine pH 2.2. Afterincubation for 30 min at room temperature, the suspension


Proteomics 2005, 5, 2443–2452 Microbiology 2445

was centrifuged at 13 000 rpm for 10 min to remove bacteriaand the supernatant neutralised with 2 M Tris. To comparemethods of extraction, glycine was replaced by one of thefollowing solutions: (i) 10 or 70 mM Na2EDTA in PBS; (ii) 1or 5 M LiCl; (iii) 8 M urea, 50 mM Tris-Cl pH 8.0, 1 mM PMSF.A method based on digestion of the peptidoglycan layerusing lysozyme to release all proteins external to the cell wallwas employed, based on the method of Jonquieres [15]. Bac-terial cell pellets were resuspended in digestion buffer (Tris-sucrose buffer with 1 mg?mL21 lysozyme, 60 mg?mL21

mutanolysin, 50 mg?mL21 lysostaphin, 250 mg?mL21 RNA-seA, and 2 mM 4-(2-aminoethyl) benzenesulfonyl fluoridehydrochloride) (all from Sigma, St. Louis, MO, USA) andincubated at 377C with gentle rotating agitation for 1 h. Theresulting ‘protoplasts’ were pelleted by centrifugation at10 000 rpm for 10 min and the supernatant containing theSLPs and cell wall associated proteins were retained. Allextracts were dialysed against 10 mM Tris-HCl pH 7.4 at 47Cfor 24 h and filtered.

2.3 Two-dimensional electrophoresis

2-DE was performed in triplicate experiments using 11 cmpH 3–11 non-linear IPG strips. Fifty microgram proteinsamples were cup loaded onto strips rehydrated with buffercontaining 8 M urea, 2% w/v CHAPS and 0.05% w/v bromo-phenol blue. IEF was performed using an IPGPhor (GEHealthcare, Slough, UK) for a total of 17 kVh. Strips wereequilibrated in 50 mM Tris-HCl pH 8.8, 6 M urea, 30% v/vglycerol, 2% SDS and 10 mg?mL21 DTT and embedded inagarose on the top of a 4–12% gradient acrylamide gel. SDS-PAGE was carried out using a Criterion sytem (Bio-Rad,Hercules, CA, USA) at 150 V for 40 min. Proteins werevisualised by staining with colloidal CBB or silver accordingto established protocols [16, 17].

2.4 Mass spectrometry

In-gel digestion with trypsin was performed according topublished methods [17–19] modified for use with a roboticdigestion system (Genomic Solutions, Huntingdon, UK)[20]. Digestion proceeded for 8 h at 377C and products wererecovered by sequential extractions with 25 mM ammoniumhydrogen carbonate, 5% formic acid and ACN. The pooledextracts were lyophilised and redissolved in 0.1% formic acidfor MS.

Tandem electrospray mass spectra were recorded using aQ-Tof hybrid quadrupole/orthogonal acceleration TOF spec-trometer (Micromass, Manchester, UK) interfaced to aMicromass CapLC capillary chromatograph. Samples weredissolved in 0.1% aqueous formic acid and 6 mL injected ontoa Pepmap C18 column (300 mm 6 0.5 cm; LC Packings),and washed for 3 min with 0.1% aqueous formic acid (withthe stream select valve diverting the column effluent towaste). The flow rate was then reduced to 1 mL?min21, thestream select valve was switched to the data acquisition

position and the peptides were eluted into the mass spec-trometer with an ACN/0.1% formic acid gradient (5–70%ACN over 20 min). The capillary voltage was set to 3500 V,and data dependent MS/MS acquisitions were performed onprecursors with charge states of 2, 3 or 4 over a survey massrange of 540–1000. Trypsin autolysis products and keratin-derived precursor ions were automatically excluded. Thecollision voltage was varied between 18 and 45 V dependingon the charge and mass of the precursor. Product ion spectrawere charge-state de-encrypted and deisotoped with a max-imum entropy algorithm (MaxEnt 3, Micromass).

2.5 Database searching

Proteins were identified by the correlation of uninterpretedtandem mass spectra to entries in Swiss-Prot/TrEMBL and tothe translated C. difficile genome, using ProteinLynx GlobalServer (Version 1.1, Micromass). One missed cleavage perpeptide was allowed, and an initial mass tolerance of 50 ppmwas used in all searches [21].

3 Results and discussion

A number of methods exist in the literature for the removalof SLPs from the surface of a bacterial cell [7, 22, 23]. In orderto identify other proteins associated with the cell wall ofC. difficile, we have compared a number of different methodsthat remove one or both of the SLP proteins. Figure 2 shows1-D SDS-PAGE analysis of the proteins removed by thetreatment of the whole bacteria with low pH glycine, lyso-zyme, EDTA, LiCl or urea. These treatments remove proteinsfrom the cell surface by a variety of different mechanisms.The two SLPs are recovered in roughly equimolar amounts

Figure 2. Extraction of S-layer and associated proteins fromC. difficile 630 and analysis by SDS-PAGE. S-layers were extract-ed from an overnight culture of C. difficile 630 using: lane 2,0.2 M glycine pH 2.2; lane 3, lysozyme treatment; lane 4, 10 mM

EDTA; lane 5, 70 mM EDTA; lane 6, 1 M LiCl; lane 7, 5 M LiCl andlane 8, 8 M urea. Molecular weights in kDa are shown in lane 1.Arrows indicate the high (HMW) and low (LMW) molecularweight SLPs.



Figure 3. 2-DE of SLPs andassociated proteins extracted byglycine treatment. Proteins wereextracted from C. difficile 630grown at 377C in BHI by treat-ment with glycine pH 2.2 and50 mg resolved on 2-DE. Fiftyspots were identified and arenumbered G1–G50 (see Table 1).

by both glycine and 8 M urea treatments. The lysozyme,EDTA and LiCl treatments appear to remove predominantlythe LMW-SLP, but the lysozyme and 1 M LiCl treatments alsoremove other proteins, presumably those attached to the cellwall. These results are consistent with the hypothesis thatthe LMW-SLP protein is localised on the external surface ofthe bacterium, whereas the HMW-SLP is a cell wall bindingprotein intimately attached to the underlying peptidoglycan,and is therefore not removed by the milder treatments. Twoextracts were chosen for further study by 2-DE. Low pH gly-cine treatment removes both SLPs in roughly equimolarproportions, with a small number of contaminating proteins.Lysozyme digestion was also chosen, as we believed that thiswould give the best chance of removing all proteins exteriorto the cell wall in a methodical way, giving a complete pictureof this portion of the proteome.

Analysis of the glycine extracted proteins by 2-DErevealed over 50 spots in this fraction (Fig. 3). There was aconcentration of spots in the low pI range (approximately 4–5) with few spots in the higher pI range. The identities of50 spots, comprising 12 proteins, were determined againstthe translated but as yet unannotated genome and are shownin Table 1. The quality of these identifications can be judgedby viewing the protein probe score and percentage coverage.Most identifications were based on more than one peptidematch. The majority of spots were found to be derived fromthe 45 kDa HMW-SLP (the putative cell wall binding do-main) and the 36 kDa LMW-SLP (both products of the slpAgene), consistent with the results from the 1-D analysis. Aspredicted, several other proteins were identified as paralogsof the HMW-SLP. Paralogs 2, 6 and 12 were detected in thisextract, and are encoded in a cluster of genes surroundingslpA (Fig. 1). The remaining paralogs are not found in clus-ters, but are dispersed throughout the genome. Of these, wedetected paralogs 15, 24 and 25 in this extract. In contrast to

SlpA, paralogs 2, 6, 12, 24 and 25 were all visualised as spotsof molecular mass close to that predicted for each matureprotein after the removal of the signal peptide, indicatingthat a second internal cleavage had not taken place. Interest-ingly, paralog 15, which has a predicted molecular mass of167 kDa, was visualised as a series of spots of molecularmass between 47 and 85 kDa. Paralog 15 contains a cell wallbinding domain linked to a series of 9 peptide repeats each of120 residues. The 47 kDa fragment corresponds to the cellwall binding domain, and the remaining higher molecularmass spots correspond to a varying numbers of peptiderepeats, without the cell wall binding domain. This protein istherefore subjected to processing at various positions result-ing in the display of the peptide repeats in the cell wall. Twocomponents of the flagellum were identified in this extract;the flagellin subunit (FliC or FlaA) [24] and FlgE, the flagellarhook protein. Two other proteins were also seen, a transglu-taminase-like predicted protease and NifU, a molecular scaf-fold for Fe-S clusters.

The lysozyme cell wall extract was found to be composedof over 100 spots (Fig. 4). This extract was generated bydigesting the peptidoglycan of the cell wall to release theassociated proteins. This treatment released approximatelytwice as many proteins as the glycine treatment, reflectingthe variety of proteins associated with the cell wall. Again,the majority of spots were concentrated in the low pI range,but with a number of small basic proteins, mainly ribonu-clease and lysozyme, which were used in the extractionmethod. A total of 69 protein spots, comprising 45 proteins,were identified by comparison with the translated genome,and are shown in Table 2. The two SLPs are predominant inthis extract, although the HMW protein is less abundantthan the LMW protein. A number of paralogs were alsofound, namely 2, 4, 15 and 25. Most proteins found are pre-dicted to contain a signal peptide and are therefore expected



Figure 4. 2-DE of proteins extracted by lysozyme treatment. Proteins were extracted from C. difficile 630 grown at377C in BHI by treatment with lysozyme, mutanolysin, lysostaphin and RNAseA and 50 mg resolved on 2-DE. Sixty-nine spots were identified and are numbered L1–L69 (see Table 2).

Table 1. C. difficile proteins identified in extracts prepared by glycine treatment

Spot GenomeIDa)

Description ProteinProbescoreb)

% cover-age

Number ofpeptides(residues)

Theoreticalmolecularmassc)

TheoreticalpI

Signalsequenced)

G1 CD2793 LMW SLP 50.8 21.2 3 (73) 34228 4.9 HG2 CD2793

CD0514HMW SLPSLP paralog 15

85.992.7

24.313.7

4 (82)4 (69)

39504167580

4.64.7

HH

G3 CD2793 LMW SLP 39.3 13.6 2 (47) 34228 4.9 HG4 CD0514 SLP paralog 15 53.6 11.6 2 (35) 167580 4.7 HG5 CD0514

CD2793SLP paralog 15HMW SLP

111.440.5

13.420.3

4 (57)3 (76)

16758039504

4.74.6

HH

G6 CD2791 SLP paralog 2 140.8 14.0 6 (87) 66427 6.4 HG7 CD2791 SLP paralog 2 59.6 7.1 3 (44) 66427 6.4 HG8 CD2791 SLP paralog 2 199.6 20.4 8 (126) 66427 6.4 HG9 CD2784 SLP paralog 6 43.3 6.2 2 (42) 73024 7.2 HG10 CD2784

CD2793SLP paralog 6HMW SLP

18.633.6

4.312.8

2 (29)2 (48)

7302439504

7.24.6

HH

G11 CD2793 HMW SLP 66.8 16.8 3 (63) 39504 4.6 HG12 CD2793 HMW SLP 244.5 51.6 10 (193) 39504 4.6 HG13 CD2793 HMW SLP 325.0 55.9 14 (212) 39504 4.6 HG14 CD2793 HMW SLP 230.3 48.1 10 (180) 39504 4.6 HG15 CD2793 HMW SLP 309.4 41.7 11 (150) 39504 4.6 HG16 CD2793 HMW SLP 142.8 24.3 6 (91) 39504 4.6 HG17 CD2793 LMW SLP 906 65.1 11 (209) 34228 4.9 HG18 CD2793 SlpA 378.9 35.3 15 (255) 76115 4.8 HG19 CD2793 LMW SLP 157.8 29.9 8 (103) 34228 4.9 HG20 CD2793 LMW SLP 2983.6 52.8 12 (182) 34228 4.9 HG21 CD0239

CD2793FlaA, flagellinLMW SLP

138.3424.2

13.162.0

4 (38)14 (214)

3075534228

8.64.9

HH

G22 CD0239CD2793

FlaA, flagellinLMW SLP

76.2232.4

18.628.7

3 (54)9 (99)

3075534228

8.64.9

xH



Table 1. Continued

Spot GenomeIDa)


% cover-age



TheoreticalpI

Signalsequenced)

G23 CD2793CD0239

LMW SLPFlaA, flagellin

68.259.4

18.022.8

3 (62)3 (66)

3422830755

4.98.6

Hx

G24 CD0239CD2793

FlaA, flagellinHMW SLP

132.4240.1

15.832.6

3 (49)10 (121)

3075539504

8.64.6

xH

G25 CD2793 HMW SLP 352.9 34.2 11 (116) 39504 4.6 HG26 CD2793

CD0239LMW SLPFlaA, flagellin

89.9193.9

15.3633.8

4 (53)7 (98)

3422830755

4.98.6

Hx

G27 CD0844 SLP paralog 25 40.6 14.4 3 (45) 33776 5.2 HG28 CD0239

CD2793CD0844

FlaA, flagellinLMW SLPSLP paralog 25

799.461.070.1

28.621.514.4

5 (83)4 (74)3 (45)

307553422833776

8.64.95.2

xHH

G29 CD2793CD0239

LMW SLPFlaA, flagellin

3571.7105.7

55.78.5

12 (192)5 (61)

3422830755

4.98.6

Hx

G30 CD0239 FlaA, flagellin 86.5 26.6 4 (52) 30755 8.6 xG31 CD2793

CD0239LMW SLPFlaA, flagellin

232.8117.4

34.523.1

8 (119)4 (67)

3422830755

4.98.6

Hx

G32 CD0255 FlgE, flagellar hook protein 62.2 10.1 2 (33) 34571 5.2 xG33 CD2793 HMW SLP 225.0 38.2 9 (143) 39504 4.6 HG34 CD2793 LMW SLP 75.8 17.4 3 (60) 34228 4.9 HG35 CD2793 HMW SLP 232.7 33.2 8 (124) 39504 4.6 HG36 CD2793 HMW SLP 75.8 8.3 3 (60) 39504 4.6 HG37 CD2793 HMW SLP 107.2 25.1 6 (94) 39504 4.6 HG38 CD1280 NifU, nitrogen fixation/Fe-S

cluster formation33.8 8.90 1 (13) 15853 4.5 x

G39 CD2793 HMW SLP 109.0 19.3 4 (72) 39504 4.6 HG40 CD2793 HMW SLP 53.7 7.0 2 (26) 39504 4.6 HG41 CD2793 HMW SLP 66.7 12.8 3 (48) 39504 4.6 HG42 CD2793 HMW SLP 35.1 12.3 2 (46) 39504 4.6 HG43 CD2793 LMW SLP 870.5 21.7 3 (75) 34228 4.9 HG44 CD2794 SLP paralog 12 59.3 12.1 2 (64) 57586 8.8 HG45 CD0514 SLP paralog 15 22.0 10.8 1 (22) 167580 4.7 HG46 CD2793 LMW SLP 177.3 7.5 1 (26) 34228 4.9 HG47 CD0514 SLP paralog 15 129.5 21.2 5 (68) 167580 4.7 HG58 CD2793 LMW SLP 914.4 27.5 4 (95) 34228 4.9 HG49 CD1156 Transglutaminase-like

predicted protease55.0 5.0 2 (18) 39277 9.4 H

G50 CD2193 SLP paralog 24 33.3 4.4 1 (13) 31501 7.5 H

a) Genome ID corresponds to the gene identification by The Sanger Institute (www.sanger.ac.uk/C_difficile)b) Protein Probe score determined by Protein Lynx Global Server (Micromass)c) Predicted molecular mass (Da) based on sequences containing signal peptide (where applicable) except for the case of the HMW and

LMW SLPsd) Predicted by Signal P. H indicates presence, x indicates absence of signal peptide.

to be either in the cell wall or associated with cytoplasmicmembrane. In addition, a number of proteins expected to belocalised in the cytoplasm were detected, a possible indicatorof cell lysis during the extraction procedure.

The use of low pH glycine releases large quantities of theSLPs from the cell surface, together with smaller amounts ofother proteins, the majority of which are paralogs of theHMW-SLP. In contrast, digestion of the peptidoglycan cellwall with lysozyme releases a greater number of proteins

including the two SLPs, certain parologs and other proteinsof diverse function. The differences between the extracts arefurther illustrated by the detection of different SLP paralogsin each extract. For example, paralog 4 is only seen in thelysozyme extract whereas paralogs 6, 12 and 24 are only seenin the glycine extract. This may reflect differences in the waythese paralogs are attached to the underlying cell wall. Anumber of proteins found in the lysozyme extract do nothave signal peptides, and were possibly derived from the



Table 2. C. difficile proteins identified in extracts prepared by lysozyme treatment

Spot GenomeIDa)


% cover-age



TheoreticalpI

Signalsequenced)

L1 CD0514 SLP paralog 15 71.1 19.0 3 (50) 167580 4.7 HL2 CD0514 SLP paralog 15 23.7 10.8 1 (22) 167580 4.7 HL3 CD0514 SLP paralog 15 56.9 9.5 2 (32) 167580 4.7 HL4 CD2966 AdhE, alcohol dehydrogenase 96.4 9.1 5 (80) 96674 5.8 xL5 CD2966 AdhE, alcohol dehydrogenase 45.9 4.0 2 (35) 96674 5.8 xL6 CD2966

CD3394AdhE, alcohol dehydrogenasePykF, pyruvate kinase

75.352.3

9.34.8

4 (80)2 (28)

9667463092

5.85.0

xx

L7 CD2791CD3091

SLP paralog 2TreA/C trehalose-6-phosphate

hydrolase

142.0146.7

14.111.7

6 (87)6 (65)

6642766104

6.46.3

Hx

L8 CD3394 PykF pyruvate kinase 65.3 5.0 2 (29) 63092 5.0 xL9 CD2755 PtsI, phosphoenolpyruvate

phosphotransferase52.0 5.3 2 (30) 63132 4.8 x

L10 CD2672 AppA, ABC transport systemoligopeptide binding protein

173.0 15.8 8 (84) 59706 5.6 H


118.1 14.1 7 (75) 59706 5.6 H

CD3394 PykF, pyruvate kinase 59.3 4.4 2 (26) 63092 5.0 xL12 CD2793

CD0395CD3173CD1059

LMW SLPPutative coA transferasePgk, Phosphoglycerate kinaseAcetyl coA acetyl transferase

156.592.580.161.4

14.818.813.515.4

3 (53)5 (74)3 (54)3 (63)

34228442164304843171

4.95.25.15.4

Hxxx


66.0 8.1 3 (43) 59706 5.6 H

CD0855 Putative transport systempeptide-binding protein

19.5 10.7 2 (56) 58496 5.4 H

L14 CD3285 Pgi, glucose 6-phosphate iso-merase

53.7 14.9 4 (57) 50519 5.6 x

CD3170 Eno, enolase 44.2 4.2 2 (19) 48149 4.6 xL15 CD3173 Pgk, phosphoglycerate kinase 47.1 10.0 2 (40) 43048 5.1 xL16 CD0395 Putative coA transferase 302.6 49.1 13 (196) 44216 5.2 xL17 CD3285 Pgi, glucose 6-phosphate iso-

merase84.3 17.3 4 (78) 50519 5.6 x

L18 CD0395CD1279

Putative coA transferaseNifS cysteine desulfurase

85.731.2

14.311.4

3 (57)2 (45)

4421644387

5.26.3

xx

L19 CD3174 Glyceraldehyde-3-phosphatedehydrogenase

167.7 21.2 6 (71) 36015 5.9 x

L20 CD3174

CD1594

Glyceraldehyde-3-phosphatedehydrogenase

CysK cysteine synthase

282.1237.3

29.937.1

9 (100)8 (112)

3601532647

5.95.4

xx

L21 CD0672 Nif3 related protein 22.7 8.3 2 (24) 33099 5.4 xL22 CD0672 Nif3 related protein 79.8 17.0 4 (49) 33099 5.4 xL23 CD1594 CysK cysteine synthase 178.2 36.8 8 (105) 32647 5.4 xL24 CD0401

CD2139

EtfA1, electron transport flavo-protein

Elongation factor Ts

115.072.9

18.318.2

5 (63)3 (55)

3712933122

4.65.2

xx

L25 CD3664CD0395

Putative aminotransferasePutative coA transferase

110.4169.6

20.117.3

5 (80)4 (69)

4477944216

5.25.2

xx

L26 CD3173CD2793CD3664

Pgk, phosphoglycerate kinaseLMW SLPPutative aminotransferase

105.640.174.2

17.84.99.0

4 (71)2 (35)3 (36)

430483422844779

5.14.95.2

xHx

L27 CD2793 LMW SLP 240.3 20.3 9 (153) 34228 4.9 HL28 CD3173

CD2793Pgk, phosphoglycerate kinaseLMW SLP

120.7228.7

19.518.8

5 (78)10 (135)

4304834228

5.14.9

xH



Table 2. Continued

Spot GenomeIDa)


% cover-age



TheoreticalpI

Signalsequenced)

L29 CD0394CD2793

Lactate dehydrogenaseLMW SLP

196.356.4

27.43.6

7 (90)2 (23)

3644934228

5.14.9

xH

L30 CD2793 LMW SLP 313.4 27.5 13 (199) 34228 4.9 HL31 CD2793 LMW SLP 64.3 3.6 2 (23) 34228 4.9 HL32 CD2793 LMW SLP 382.0 26.2 13 (187) 34228 4.9 HL33 CD0239

CD0873FlaA, FlagellinABC transporter substrate

binding protein

130.148.2

37.211.2

6 (31)2 (38)

3075536049

8.65.4

xH

L34 CD0844CD2793

SLP paralog 25LMW SLP

55.7122.8

13.16.8

3 (41)4 (49)

3377634228

5.24.9

HH

L35 CD2755 PtsI, phosphoenolpyruvatephosphotransferase

73.6 10.53 3 (60) 63132 4.8 x

CD3225 Dihydrodipicolinate synthetase 56.4 11.2 3 (32) 31577 4.8 xL36 CD0795 Conserved hypothetical 50.0 8.5 2 (21) 26930 4.6 xL37 CD3174 Glyceraldehyde-3-phosphate

dehydrogenase56.2 8.1 2 (27) 36015 5.9 x

L38 CD0239CD3172

FlaA, flagellinTpiA, triospehosphate iso-

merase

71.168.3

14.118.6

2 (41)3 (46)

3075527174

8.65.0

xx

L39 CD2755 PtsI, phosphoenolpyruvatephosphotransferase

73.3 9.7 3 (55) 63132 4.8 x

L40 CD2240 Dihydrodipicolinate synthetase 58.8 16.6 3 (50) 34291 5.3 xL41 CD2461 DnaK, chaperone protein 129.5 21.2 5 (68) 66449 4.7 xL42 CD3172 TpiA, triosephosphate iso-

merase193.1 36.8 7 (80) 27174 5.0 x

L43 CD0401 EtfA1, electron transport flavo-protein

102.4 12.5 4 (43) 37129 5.1 x

L44 CD0401 EtfA1, electron transport flavo-protein

76.2 6.7 3 (23) 37129 5.1 x

L45 CD0239 FlaA, flagellin 267.2 9.3 2 (28) 30755 8.6 xL46 CD1125 Nitroreductase 92.3 21.2 4 (62) 32711 5.3 xL47 CD3205 Nitroreductase 35.6 9.7 2 (19) 22344 7.4 xL48 CD0058 Elongation factor Tu 71.2 5.5 2 (22) 44008 4.6 xL49 CD2793 HMW SLP 90.9 20.3 4 (76) 39504 4.8 HL50 CD0401 EtfA1, electron transport flavo-

protein68.2 15.6 4 (44) 37129 5.1 x

L51 CD2461 DnaK, chaperone protein 33.7 5.7 2 (35) 66449 4.7 xL52 CD1634 TerD, telurium resistance

protein56.2 19.8 3 (38) 20742 4.5 x

L53 CD1280 NifU, nitrogen fixation/Fe-Scluster formation

87.8 26.0 3 (38) 15853 4.5 x

L54 CD2513 YjgF, putative translationinitiation factor

55.6 11.1 2(14) 13582 4.9 x

L55 CD1355 CspA, putative cold shockprotein

22.7 25.8 1 (17) 7105 4.7 x

L56 CD2793 HMW SLP 70.8 13.9 3 (52) 39504 4.6 HL57 CD2793 HMW SLP 66.8 16.8 3 (63) 39504 4.6 HL58 CD1355 CspA, putative cold shock

protein62.9 48.5 3 (32) 7105 4.7 x

L59 CD1690 Thioredoxin 31.0 20.9 2 (22) 11708 5.1 xL60 CD2756 HPr, putative phosphocarrier 113.9 60.5 4 (51) 9046 5.2 xL61 CD3663 RpsF, 30S ribosomal protein 125.2 66.3 6 (61) 10502 7.3 xL62 CD2137 Ribosome recycling factor 91.7 33.0 5 (61) 21090 7.5 xL63 Ribonuclease From sample preparation 489.0 24.8 2 (35) 16442 9.0 –L64 Ribonuclease From sample preparation 1469 56.7 5 (85) 16442 9.0 –



Table 2. Continued

Spot GenomeIDa)


% cover-age



TheoreticalpI

Signalsequenced)

L65 Lysozyme From sample preparation 100.0 12.9 2 (19) 16220 10.7 –L66 CD2787

LysozymeSLP paralog 4From sample preparation

27.8635.5

4.736.7

1 (17)5 (54)

3857716220

6.710.7

H–

L67 Lysozyme From sample preparation 606.9 32.0 4 (47) 16220 9.5 –L68 Ribonuclease From sample preparation 2196.2 87.1 6 (101) 16442 9.0 –L69 Lysozyme From sample preparation 148.0 30.6 3 (45) 16220 10.7 –

a) Genome ID corresponds to the gene identification by The Sanger Institute (www.sanger.ac.uk/C_difficile)b) Protein Probe score determined by Protein Lynx Global Server (Micromass)c) Predicted molecular mass (Da) based on sequences containing signal peptide (where applicable) except for the case of the HMW and

LMW SLPsd) Predicted by Signal P. H indicates presence, x indicates absence of signal peptide

cytoplasm following partial cell lysis. This is in contrast tothe glycine extract that contains predominantly proteins withsignal peptides or those exported from the cell by a dedicatedmechanism.

Twelve paralogs of the SLP proteins are localised in oneregion of the genome [7, 8] (Fig. 1), and RT-PCR analysis hasshown that paralogs 1–6 are all expressed [7]. In addition,transcript mapping [25] has indicated that paralogs 2 and 3(Cwp66) are cotranscribed and that paralogs 12 and SlpAhave a common transcript. This is consistent with the para-log 12 being detected in our glycine extracts. However, inneither of our protein preparations are paralogs 3 or 5 visua-lised, whereas paralog 2 is highly expressed (Figs. 3, 4).Paralog 3, the Cwp66 adhesin, was shown by immunoblot-ting to be present in membrane and cell wall extracts of heatshocked C. difficile [12]. This difference may be due to theincreased sensitivity of immunoblotting, incompatibility ofthe protein with the 2-DE procedure or the lack of expressionunder the conditions we used. Paralog 4 was identified in thelysozyme extract, although the protein probe score and per-cent coverage were low. However mRNA for this protein, alsotermed the Cwp84 protease, is transcribed at the early expo-nential phase of growth [25], increasing our confidence inthe identity of this spot. Our analysis has only shown 8 of30 paralogs of the SLP proteins to be expressed under theconditions we used. It is tempting to speculate that many ofthe remaining paralogs are either not expressed at all,expressed only at very low levels or are expressed only underparticular culture conditions. Many bacterial proteinsinvolved in virulence are expressed upon stress, e.g. heat orcold shock or in response to environmental stimulation. Inthe lysozyme extract, a putative cold shock protein (CD1335)was found, together with the heat shock protein DnaK andseveral ABC transporter proteins.

4 Concluding remarks

Our proteome analysis of the cell wall of C. difficile hasrevealed that the composition of extracts depends greatly onthe methods used for their isolation. While the SLP proteinsare the most abundant in all the extracts, many other pro-teins are present as minor species. Several proteins in theglycine and lysozyme extracts were identified as paralogs ofthe SLPs, consistent with the hypothesis that these are cellwall associated proteins. This work provides a basis for fur-ther analysis of the expression of these proteins under arange of conditions, including those that mimic conditionsfound in the gut, e.g. temperature or antibiotic stress.

This work was supported by a BBSRC studentship to A. W.,and B. C. was supported by a grant from Dstl. The authors wouldlike to thank the Sanger Institute for releasing the genomesequence prior to publication and the Wellcome Trust for fundingJ. N. and the proteomic facilities at the CMMI, Imperial College.R. W. thanks the Arthritis Research Campaign and the MRC forfinancial support.

5 References

[1] Kyne, L., Farrell, R. J., Kelly, C. P., Gastroenterol. Clin. NorthAm. 2001, 30, 753–777, ix-x.

[2] Borriello, S., J. Antimicrob. Chemother. 1998, 41, 13–19.

[3] Poxton, I. R., McCoubrey, J., Blair, G., Clin. Microbiol. Infect.2001, 7, 421–427.

[4] Wilcox, M. H., Cunniffe, J. G., Trundle, C., Redpath, C., J.Hosp. Infect. 1996, 34, 23–30.



[5] Kyne, L., Hamel, M. B., Polavaram, R., Kelly, C. P., Clin. Infect.Dis. 2002, 34, 346–353.

[6] Tasteyre, A., Barc, M. C., Collignon, A., Boureau, H., Karja-lainen, T., Infect. Immun. 2001, 69, 7937–7940.

[7] Calabi, E., Ward, S., Wren, B., Paxton, T. et al., Mol. Micro-biol. 2001, 40, 1187–1199.

[8] Karjalainen, T., Waligora-Dupriet, A. J., Cerquetti, M., Spi-gaglia, P. et al., Infect. Immun. 2001, 69, 3442–3446.

[9] Calabi, E., Calabi, F., Phillips, A. D., Fairweather, N. F., Infect.Immun. 2002, 70, 5770–5778.

[10] Calabi, E., Fairweather, N., J. Bacteriol. 2002, 184, 3886–3897.

[11] Kuroda, A., Sekiguchi, J., J. Bacteriol. 1991, 173, 7304–7312.

[12] Waligora, A. J., Hennequin, C., Mullany, P., Bourlioux, P. etal., Infect. Immun. 2001, 69, 2144–2153.

[13] Mukherjee, K., Karlsson, S., Burman, L. G., Akerlund, T.,Microbiology 2002, 148, 2245–2253.

[14] McCoy, E. C., Doyle, D., Burda, K., Corbeil, L. B., Winter, A. J.,Infect. Immun. 1975, 11, 517–525.

[15] Jonquieres, R., Bierne, H., Fiedler, F., Gounon, P., Cossart, P.,Mol. Microbiol. 1999, 34, 902–914.

[16] Neuhoff, V., Arold, N., Taube, D., Ehrhardt, W., Electropho-resis 1988, 9, 255–262.

[17] Shevchenko, A., Wilm, M., Vorm, O., Mann, M., Anal. Chem.1996, 68, 850–858.

[18] Wilm, M., Neubauer, G., Mann, M., Anal. Chem. 1996, 68,527–533.

[19] Jeno, P., Mini, T., Moes, S., Hintermann, E., Horst, M., Anal.Biochem. 1995, 224, 75–82.

[20] Wait, R., Gianazza, E., Eberini, I., Sironi, L. et al., Electropho-resis 2001, 22, 3043–3052.

[21] Peirce, M. J., Wait, R., Begum, S., Saklatvala, J., Cope, A. P.,Mol. Cell. Proteomics 2004, 3, 56–65.

[22] Cerquetti, M., Molinari, A., Sebastianelli, A., Diociaiuti, M. etal., Microbiol. Pathogenesis 2000, 28, 363–372.

[23] Sleytr, U. B., Beveridge, T. J., Trends Microbiol. 1999, 7, 253–260.

[24] Tasteyre, A., Barc, M. C., Karjalainen, T., Dodson, P. et al.,Microbiology 2000, 146, 957–966.

[25] Savariau-Lacomme, M. P., Lebarbier, C., Karjalainen, T., Col-lignon, A., Janoir, C., J. Bacteriol. 2003, 185, 4461–4470.


Documents

Proteomic analysis of cell surface proteins from Clostridium difficile