analysis of the listeria cell wall proteome by two-dimensional nanoliquid chromatography coupled to...

REGULAR ARTICLE

Analysis of the Listeria cell wall proteome by

two-dimensional nanoliquid chromatography

coupled to mass spectrometry

Enrique Calvo1, M. Graciela Pucciarelli2, Hélène Bierne4, Pascale Cossart4,Juan Pablo Albar3 and Francisco García-del Portillo2

1 Unidad de Proteómica, Fundación Centro Nacional de Investigaciones Cardiovasculares (CNIC),Tres Cantos, Madrid, Spain

2 Departamento de Biotecnología Microbiana, Centro Nacional de Biotecnología-Consejo Superiorde Investigaciones Científicas (CSIC), Campus de Cantoblanco, Madrid, Spain

3 Servicio de Proteómica, Centro Nacional de Biotecnología-Consejo Superior de InvestigacionesCientíficas (CSIC), Campus de Cantoblanco, Madrid, Spain

4 Unité des Interactions Bactéries-Cellules, Institut Pasteur, INSERM U406, Paris, France

Genome analyses have revealed that the Gram-positive bacterial species Listeria monocytogenesand L. innocua contain a large number of genes encoding surface proteins predicted to be cova-lently bound to the cell wall (41 and 34, respectively). The function of most of these proteins isunknown and they have not even been identified biochemically. Here, we report the first char-acterization of the Listeria cell wall proteome using a nonelectrophoretic approach. The materialanalyzed consisted of a peptide mixture obtained from a cell wall extract insoluble in boiling 4%SDS. This extract, containing peptidoglycan (intrinsically resistant to proteases) and stronglyassociated proteins, was digested with trypsin in a solution with 0.01% SDS, used to favor proteindigestion throughout the peptidoglycan. The resulting complex peptide mixture was fractionatedand analyzed by two-dimensional nanoliquid chromatography coupled to ion-trap mass spec-trometry. A total of 30 protein species were unequivocally identified in cell wall extracts of thegenome strains L. monocytogenes EGD-e (19 proteins) and L. innocua CLIP11262 (11 proteins).Among them, 20 proteins bearing an LPXTG motif recognized for covalent anchoring to thepeptidoglycan were identified. Other proteins detected included peptidoglycan-lytic enzymes, apenicillin-binding protein, and proteins bearing an NXZTN motif recently proposed to directprotein anchoring to the peptidoglycan. The marked sensitivity of the method makes it highlyattractive in the post-genome era for defining the cell wall proteome in any bacterial species. Thisinformation will be useful to study novel protein-peptidoglycan associations and to rapidly iden-tify new targets in the surface of important bacterial pathogens.

Received: April 26, 2004Revised: June 21, 2004

Accepted: June 24, 2004

Keywords:

Listeria / LPXTG motif / Proteome / Two-dimensional nanoliquid chromatography-tandem mass spectrometry

Proteomics 2005, 5, 433–443 433

1 Introduction

The bacterial cell wall plays a crucial role in ensuring cellviability by maintaining the high internal osmotic pressure.It also provides the cell with a specific shape. The cell wall ofGram-positive bacteria is composed by a peptidoglycan mac-romolecule and different types of molecules attached to it,including teichoic acids, teichuronic acids, polyphosphates,and carbohydrates [1]. A specific subset of surface proteinsare also known to be located in the cell wall, establishing

Correspondence: Dr. Francisco García-del Portillo, Departamentode Biotecnología Microbiana, Centro Nacional de Biotecnología-Consejo Superior de Investigaciones Científicas (CSIC), Campusde Cantoblanco, Madrid, SpainE-mail: [email protected]: 134-91-5854506

Abbreviations: 2DnLC-MS/MS, two-dimensional nanoliquidchromatography coupled to ion-trap tandem mass spectrometry;Inl, internalin; Q-TOF-MS, quadrupole-time of flight-mass spec-trometry; SrtA, sortase

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

DOI 10.1002/pmic.200400936

434 E. Calvo et al. Proteomics 2005, 5, 433–443

interactions with either the peptidoglycan or accessorymolecules as teichoic acids [2]. The most remarkable exam-ple is probably that of the proteins covalently linked to thepeptidoglycan. These proteins are processed and anchoredvia a transpeptidation reaction carried out by specificenzymes named sortases [2, 3].

Cell wall-associated proteins decorate the surface ofGram-positive bacteria and play functions in essential pro-cesses, such as sensing of the external environment, nutrienttransport, adhesion to biological and abiotic surfaces, survi-val within infected host cells, peptidoglycan metabolism, andcontrol of cell division [2]. These essential traits have led to anincreasing interest for identifying new cell wall-associatedproteins, especially in Gram-positive bacterial pathogens [2].However, the biochemical characterization of cell wall-asso-ciated proteins requires laborious procedures, involving iso-lation of cell envelopes and solubilization of proteins fromthe peptidoglycan lattice. Protein solubilization demands theemployment of hydrolytic enzymes, such as lysozyme,lysostaphin, phage murein hydrolases, amidases, or fungimuramidases, which cleave specific bonds in the pepti-doglycan [2]. Only upon this treatment, proteins can beresolved and analyzed by gel electrophoresis. On the otherhand, due to the apparent scarcity shown for some of theseproteins [2], their identification and characterization requiresin most cases the availability of specific antisera.

In the genome era, a large number of new genes havebeen annotated as encoding putative envelope proteins dis-playing association to the cell wall. This feature is especiallyrelevant in Gram-positive bacteria, such as Listeria, Staphylo-coccus, and Streptococcus [4–9]. In the case of Listeria, a com-parative genome analysis revealed that L. monocytogenes andL. innocua contain a large repertoire of surface proteins pre-dicted to be covalently bound to the peptidoglycan [8]. Listeriaspecies are capable for adapting to diverse and extreme envi-ronmental conditions of pH, temperature, and osmolarity[10]. These traits may explain why these bacteria encode somany surface proteins exposed to the environment. Thecontribution of many of these proteins to the physiology ofListeria remains at present undefined. However, in the fewcases known, mostly in the pathogen L. monocytogenes, it hasbeen shown that cell surface proteins direct processes suchas host cell invasion and cell-cell spreading promoted byremodelling of host-actin [7, 10–12].

Among the 2853 and 3052 open reading frames identifiedin the L. monocytogenes and L. innocua genomes, respectively,more than 600 genes are annotated in the ‘cell envelope andcellular processes’ functional category [8]. More than 330genes of this category are predicted to encode transportbinding/lipoproteins whereas about 60 genes (63/57 inL. monocytogenes/L. innocua respectively) encode putative sur-face proteins. Remarkably, only one protein included in thislatter subcategory of “surface proteins” has an ortholog in theclose relative bacterium Bacillus subtilis. This specific group ofsurface proteins includes: (i) the protein family sharing anLPXTG motif close to the C-terminal end recognized by sor-

tases [3, 7]; (ii) proteins containing GW modules that promoteprotein-lipoteichoic acid interactions [7]; (iii) proteins con-taining an hydrophobic C-terminal end [7]; and (iv) proteinsbelonging to the P60 family, with peptidoglycan-lytic activity[7]. Listeria species have by far the highest number of LPXTGproteins among all Gram-positive organisms whose genomesequence is known [13, 14]. L. monocytogenes and L. innocuacontain 41 and 34 genes, respectively, encoding LPXTG pro-teins [8]. Despite this large set of LPXTG proteins, only onemember of the family, internalin A (InlA), has been character-ized at a biochemical and functional level [11]. InlA mediatesbacterial entry in epithelial cells by interacting with the eukar-yotic receptor E-cadherin [11].

A first step to gain insights into the function of membersof the Listeria LPXTG protein family is to obtain evidence oftheir presence in the cell wall. This information is crucial toaddress functional studies. To date, only three ListeriaLPXTG proteins, Lmo0130, Lmo0880, and Lmo2714, havebeen identified by a ‘gel-less’ proteomic approach based onquadrupole-time of flight-mass spectrometry (Q-TOF-MS)[15]. Noteworthy, InlA, revealed in cell wall extracts by West-ern assays, was not identified in this study [15]. This resultsuggested that LPXTG proteins might be relative minor pro-teins.

The aim of this study was to use state-of-the art prote-omic technology [16] for identifying new Listeria LPXTGproteins in cell wall extracts. A nonelectrophoretic proteomicapproach based on two-dimensional nanoliquid chromatog-raphy coupled to ion-trap mass spectrometry (2DnLC-MS/MS) was used. To optimize the method, we exploited twoproperties of the peptidoglycan chemistry: (i) the peptidogly-can remains insoluble upon extensive boiling in SDS-con-taining solutions, and (ii) it is intrinsically resistant to pro-teases as trypsin. Thus, a complex peptide mixture repre-sentative of the cell wall proteome was rapidly obtained upontrypsin digestion of the SDS-insoluble envelope material(enriched in peptidoglycan and strongly associated proteins).The 2DnLC-MS/MS analysis of this complex peptide mixtureallowed us to unequivocally identify a total of 30 proteins incell wall extracts of L. monocytogenes and L. innocua. Amongthese proteins, 20 members of the LPXTG protein familywere identified, including InlA.

2 Materials and methods

2.1 Preparation of peptidoglycan-enriched material

from Listeria

The bacterial strains L. monocytogenes EGD-e and L. innocuaCLIP11262, with genomes completely sequenced [8], weregrown overnight at 377C in 50 mL brain-heart-infusion (BHI)growth medium (Difco, Detroit, MI, USA). Bacteria werespun down by centrifugation (10 000 3 g, 10 min, 47C), sus-pended in 10 mL phosphate-buffered saline (PBS), pH 7.4,containing a cocktail of protease inhibitors (Roche, Indiana-


Proteomics 2005, 5, 433–443 Microbiology 435

polis, IN, USA) and DNAse (100 mg?mL21). Bacteria werelysed in three passes through a French press. Unbroken cellswere removed by centrifugation (10 000 3 g, 10 min, 47C)and cell envelopes obtained by centrifugation of supernatantat a higher speed (18 000 3 g, 20 min, 47C). The pellet con-taining envelopes was resuspended in 1.5 mL PBS (pH 7.4)and gently mixed with 1.5 mL boiling 8% SDS. Boiling con-ditions were maintained for 4 h and the material furtherincubated overnight at 807C. The SDS-insoluble material,enriched in peptidoglycan and strongly associated proteins,was collected by centrifugation at high speed (200 000 3 g,20 min, 307C) and washed four times with 2.5 mL distilledwater. The amount of peptidoglycan was critical for the effi-cacy of the washing steps. Thus, as assessed by 2DnLC-MS/MS, peptidoglycan purified from a larger volume of bacterialculture (250 mL) and washed in equal volumes of distilledwater retained some cytosolic proteins nonexpected to asso-ciate strongly to peptidoglycan (data not shown). The washedmaterial was finally resuspended in 500 mL 50 mM ammo-nium bicarbonate/0.01% SDS. Addition of this low amountof SDS was considered to favor trypsin digestion within thepeptidoglycan lattice. Modified trypsin (sequencing grade;Promega, Madison, WI, USA) was added at a final amount of2 mg per sample. Digestion was performed under shakingconditions overnight at 377C. Peptides were separated fromthe undigested macromolecular peptidoglycan by cen-trifugation (200 000 3 g, 20 min, 307C). The supernatantcontaining the peptide mixture was lyophilized and kept at–20ºC.

2.2 SDS removal in lyophilized material containing

the peptide mixture

The lyophilized peptide mixture was redissolved in 100 mL Asolution (A = 0.5% acetic acid in water) and quantified.Finally, the solution was cleaned up with cartridges thatspecifically trap the remaining SDS (Michrom BioResources,Auburn, CA, USA). Prior to use, cartridges were washed with90% ACN and 0.1% HCl and then treated with ACN toremove HCl for regeneration. Finally, cartridges were equili-brated in 30% ACN and 1% acetic acid in water.

2.3 2DnLC-MS/MS analysis

The tryptic peptide mixtures (4 mg each) were injected onto astrong-cation-exchange microprecolumn (500 mm ID 3

15 mm BioX-SCX) (LC Packings, Amsterdam, The Nether-lands) with a flow rate of 30 mL/min as a first-dimensionalseparation. Peptides were eluted from the column as frac-tions by injecting three salt steps of increasing concentrationof ammonium acetate (10, 100, and 2000 mM). Each of thethree fractions together with the nonretained fraction was online injected onto a C-18 reversed-phase microcolumn(300 mm ID 3 5 mm PepMap) (LC Packings) to removesalts, and the peptides were analyzed in a continuous aceto-nitrile gradient consisting of 0–50% B in 45 min, 50–90% B

in 1 min (B = 95% ACN, 0.5% acetic acid in water) on a C-18reversed-phase self-packing nanocolumn (100 mm ID 3

15 cm Discovery BIO Wide pore) (Supelco, Bellefonte, PA,USA). A flow rate of ca. 300 nL/min was used to elute pep-tides from the reversed-phase nanocolumn to a PicoTip

emitter nanospray needle (New Objective, Woburn, MA,USA) for real-time ionization and peptide fragmentation onan Esquire HCT ion-trap (Bruker-Daltonics, Bremen, Ger-many) mass spectrometer. Every 1 s, the instrument cycledthrough acquisition of a full-scan mass spectrum and oneMS/MS spectrum. A 4 Da window (precursor m/z 6 2), anMS/MS fragmentation amplitude of 0.8 V and a dynamicexclusion time of 0.3 min were used for peptide fragmenta-tion. 2DnLC was automatically performed on an advancedmicrocolumn switching device (Switchos, LC Packings) cou-pled to an autosampler (Famos; LC Packings) and a nano-gradient generator (Ultimate nano-HPLC; LC Packings). Thesoftware Hystar 2.3 was used to control the whole analyticalprocess.

2.4 Database analysis

MS/MS spectra were batch-processed by using DataAnalysis5.1 SR1 and MS BioTools 2.0 software packages and searchedagainst the L. monocytogenes and L. innocua protein databasesusing Mascot software (Matrix Science, London, UK).

3 Results and discussion

3.1 Optimization of a gel-independent proteomic

approach to analyze the Listeria cell wall

proteome

Our first objective was to evaluate the efficacy of 2DnLC-MS/MS for identifying bacterial proteins strongly associated tothe cell wall. A material highly enriched in Listeria pepti-doglycan and strongly associated proteins was obtained byextensive boiling of cell envelopes in 4% SDS [15]. A key stepin the protocol was the omission of the digestion step in-volving peptidoglycan-hydrolytic enzymes. Instead, the SDS-insoluble material was digested with trypsin. To facilitateaccess of the protease to proteins buried into the peptidogly-can lattice, a low amount of SDS (0.01%) was added to thedigestion buffer. This treatment yielded a peptide mixturethat was easily separated from the macromolecular pepti-doglycan by centrifugation. Such tryptic-peptide mixture wasfractionated and analyzed by 2DnLC coupled to electrosprayionization and tandem mass spectrometry (ESI-MS/MS).The 2DnLC was performed for the first-dimensional separa-tion by means of a strong-cation-exchange column, applyinga nonlinear salt gradient. For the second dimension, a C-18reversed-phase column was used with an ACN gradient. Thenonbound material and the peptides eluted at each salt stepwere submitted to fragmentation on an ion-trap mass spec-trometer. The resulting MS/MS data were matched against



protein databases for identification. Figure 1A displays chro-matograms corresponding to the different salt steps of thenonlinear gradient. The capacity of this method to revealeven minor composition differences is depicted in Fig. 1B,which shows chromatographic differences when severalpeptides corresponding to proteins present only in one of thetwo samples were monitored in the 10 mM and 100 mM

ammonium acetate fractions. This was the case for ions 2and 4 (Fig. 1B top and bottom, respectively), which areascribed to sequences from proteins present exclusively inthe peptidoglycan of L. monocytogenes (internalin-H andinternalin-A, respectively). Likewise, chromatographic dif-ferences were also noted when monitoring signals fromortholog proteins present in both Listeria samples but show-ing slightly different amino acid composition. This was thecase for peptide 1*, which was ascribed to a L. innocua pro-tein (Lin2289) with ortholog in L. monocytogenes (Lmo2185)(Fig. 1B, top). This peptide 1*, TQISGALQDVK, is specific ofLin2289 and absent in the sequence of Lmo2185, which isTTLAGTLQDVK for the same amino acid-relative positions.Signals derived from peptides present on both samplesshowed identical retention times and ion intensities (ions 3and 3*, Fig. 1B, bottom). The amino acid sequences of ionswere identified by analyzing the corresponding MS/MSspectra (Fig. 2A–D). The comprehensive analysis of this largeset of data was performed in less than 24 h per sample. Atotal of 30 protein species were unambiguously identified inthe peptidoglycan purified from the L. monocytogenes EGD-eand L. innocua CLIP11262 genome strains (19 and 11 pro-teins respectively) (Table 1). With respect to reproducibility,more than 95% of the protein species were detected in threeanalyses corresponding to peptidoglycan preparationsobtained from three independent experiments (data notshown).

3.2 Envelope proteins identified in the Listeria cell

wall extract

All the 30 proteins species identified by 2DnLC-MS/MS (19and 11 in L. monocytogenes and L. innocua, respectively)belong to the “cell envelope and cellular processes” func-tional category, as defined in the functional classificationcodes of the Listeria genome database (http://genolist.pas-teur.fr/ListiList/help/function-codes.html) (Table 1). Withinthis category, the identified proteins are included in eitherthe “cell wall” or “cell surface proteins” subcategories (Table1). Three main protein groups were found among the 30protein species encompassing the cell wall proteome. Thefirst group formed by members of the LPXTG protein family(see below). The highest percentage of protein species iden-tified were in this group: 13 out of 19 in L. monocytogenes and7 out of 11 in L. innocua (Table 1). The second group wasformed by proteins predicted to have enzymatic activityrelated to peptidoglycan metabolism: the penicillin-bindingprotein 2A (PbpA) and the hydrolytic enzymes P60 (alsoknown as Iap, invasion-associated protein), P45 (Slp) [17]

and an N-acetyl-muramidase (MurA/Nam) [18, 19]. Thefunction of P45, P60 and MurA/Nam have been related topeptidoglycan hydrolysis required for cell separation anddivision [17, 19]. Noteworthy, P60 and MurA/Nam containtwo and four copies respectively of the cell wall-anchoringrepeat motif LysM [7, 18, 19]. These two autolytic enzymeshave been recently reported as substrates of an alternativesecretion machinery dependent on a Listeria membrane pro-tein named SecA2 [18]. The third group of proteins identifiedincluded protein species bearing a NXZTN motif close to theC-terminal end: Lmo2185 (SvpA), Lmo2186 (SvpB), andtheir L. innocua orthologs Lin2289 and Lin2290. For the spe-cific case of Lmo2185 (SvpA), as many as 20 distinct peptidesof the protein were identified (data not shown), suggestingthat this protein might be abundant in the Listeria pepti-doglycan. The NXZTN motif has recently been proposed as asequence recognized by the second sortase of Listeria, sor-tase-B (SrtB), to anchor this specific type of surface proteinsto the peptidoglycan [20]. Although the exact mechanism ofassociation of Lmo2185 (SvpA) to the peptidoglycan is stillundefined, it is known that it remains associated to pepti-doglycan withstanding boiling in 4% SDS and that it is a SrtBsubstrate [20].

Importantly, no envelope protein was identified in thecell wall proteome belonging to the remaining subcategoriesincluded in the “cell envelope and cellular processes” func-tional category: transport/binding proteins and lipoproteins;protein sensors; proteins involved in membrane bioener-getics, mobility or chemotaxis; proteins of secretionmachineries, soluble internalins, and proteins involved intransformation and competence. These results confirm thatthe stringent purification method used in our study is suffi-cient to prevent unspecific retention of envelope proteins tothe peptidoglycan lattice along the cell fractionation proce-dure. The cell wall proteome profile obtained by 2DnLC-MS/MS is fully consistent with our previous studies showing thatthe peptidoglycan-enriched material insoluble in boiling 4%SDS contains a very limited subset of envelope proteins.Only surface proteins displaying strong interaction with thepeptidoglycan, such as those covalently bound, are retainedin this material. Thus, InlA, a LPXTG protein, is detected byWestern analysis in purified peptidoglycan material whereasInlB, loosely associated to the cell wall via its interaction withlipoteichoic acids [21], is not [20].

3.3 Identification of new Listeria LPXTG proteins by

2DnLC-MS/MS

The increase in sensitivity provided by the 2DnLC-MS/MSproteomic approach was remarkable in the case of theLPXTG protein family (Table 1). Internalin A (InlA), which,as previously mentioned, is easily detected in peptidoglycanextracts by Western analysis [15], was however not identifiedby a nonelectrophoretic proteomic approach using Q-TOF-MS [15]. In fact, only three L. monocytogenes LPXTG proteins(Lmo0130, Lmo0880, and Lmo2714) were identified by



Figure 1. Data attained by 2DnLC. (A) MS and MS/MS base peak chromatogram for each salt step from L. mono-cytogenes EGD-e and L. innocua CLIP11262 samples: NBF (nonbound fraction), 10 mM, 100 mM and 2 M ammo-nium acetate. (B, top) MS extracted ion chromatogram (EIC) corresponding to the 10 mM ammonium acetatefraction for the ions at m/z 579.8 and m/z 656.8 (ions 1* and 2, respectively). Ion 1* is ascribed to a sequence fromprotein Lin2289. The corresponding peptide in the L. monocytogenes ortholog (Lmo2185) has a different aminoacid sequence and therefore fails to show in the chromatogram. (B, bottom) MS EIC corresponding to the 100 mM

ammonium acetate fraction for the ions at m/z 703.8 and m/z 806.4 (ions 3/3* and 4, respectively). Ions 3 and 3* areascribed to a sequence from protein Lmo2185 and its corresponding ortholog Lin2289, respectively. Both retentiontime and intensity are nearly identical for 3 and 3*. Ion 2 (B, top) and ion 4 (B, bottom) are ascribed to sequencesfrom proteins that are present in the L. monocytogenes preparation exclusively (InlH and InlA, respectively).Arrows indicates the absence of signals for specific monitored ions.



Figure 2. Fragmentation spectra. MS/MS spectra displaying themain fragmentation series (b- and y-series) from doubly chargedparent ions from the following sequences: (A) TQISGALQDVK;(B) ITELELSGNPLK; (C) MPANDITLYAQFTK; and (D) YPVKDG-TANTDVK.

Q-TOF-MS in the genome strain EGD-e, reaching a max-imum of three peptides detected per protein species [15]. Incontrast, our 2DnLC-MS/MS analysis resulted in the identifi-cation of 13 LPXTG proteins, including InlA, in the sameL. monocytogenes EGD-e strain (Tables 1, 2). A total of 7 LPXTGproteins were identified in L. innocua. To our knowledge, thisis the first study reporting the biochemical identification ofLPXTG proteins in this nonpathogenic Listeria species. Forsome of the LPXTG proteins present in the peptidoglycanmaterial, such as Lmo0880, as many as 17 peptides wereidentified (Table 2). In the specific cases of InlA and InlH, 14and 12 distinct peptides respectively, were unequivocallyidentified (Table 2). With the only exception of the LPXTGproteins Lmo1413 and Lin2281, which were identified basedin a single peptide, the rest 18 LPXTG proteins were identifiedwith a minimum of 3–4 peptides per protein (Table 2).

It is noteworthy that, as opposed to the predicted aminoacid sequence of their precursor forms, no tryptic peptidelocated between the LPXTG motif and the C-end was detect-ed in any of the 20 LPXTG proteins identified (Table 2). Insome cases, such as Lmo0842 and Lin0141, a peptide wasidentified just before the cleavage site proposed for Sortase-A(SrtA) between the T-G residues of the LPXTG motif [2].Thus, for Lmo0842 the peptide identified was TYEEAKPLPK(residues 2000–2009) while the motif is LPKTG (residues2007–2011). In the case of Lin0141, the peptide identifiedwas TATYQQAIPR (residues 1951–1960) and its anchoringmotif is IPRTG (residues 1958–1962). These results arguefor an exclusive presence in our cell wall sample of the fullyprocessed form of the LPXTG protein, i.e., the one anchoredcovalently to the peptidoglycan. Table 2 also reflects that insome cases the identification of high-molecular-weightLPXTG proteins was generally based in a relatively lowernumber of peptides when compared to medium-to-low-mo-lecular-weight LPXTG proteins. These differences mayaccount for: (i) a higher relative abundance of medium-to-low-molecular-weight LPXTG proteins; (ii) an uniquearrangement of the protein(s) into the peptidoglycan latticewhich could hamper the endopeptidase (trypsin) activity; and(iii) the sequence-dependent capability of specific peptidesfor ionizing and be analyzed by 2DnLC-MS/MS.

In some instances, the identification of a specific LPXTGprotein in one Listeria species was not accompanied by thatof its ortholog protein in the other species. That was the caseof Lmo0327 and Lmo0610 identified in L. monocytogenes andLin0514 identified in L. innocua (Table 1). These differencesmay arise from a distinct relative abundance of a concreteprotein species in each Listeria species in the growth condi-tion selected for this study (BHI medium, overnight at 377C).Alternatively, small changes in primary sequence betweenortholog proteins may compromise detection by 2DnLC-MS/MS. Future work, focused on the extraction of peptidoglycanmaterial in a panel of growth conditions, may provide moreinsights on whether these ortholog protein pairs are sub-jected to distinct regulation by both Listeria species in specificenvironments.



Table 1. Surface proteins of L. monocytogenes and L. innocua identified by 2DnLC-MS/MS from purified peptidoglycan

Functionalcategory a)

L. monocyto-genes b)

L. innocua b) Putative function Remarks

1.1 Cell wall Lmo0582 Lin0591 P60, invasion-associated protein (Iap),peptidoglycan hydrolase

Both orthologs identified

Lmo2505 Lin2648 P45, peptidoglycan-lytic enzyme (Slp) Both orthologs identifiedLmo1892 similar to penicillin-binding protein 2A

(PbpA)Ortholog (Lin2006) nonidentified in

L. innocuaLmo2691 Similar to autolysin, N-acetyl-muramidase

(MurA/Nam)Ortholog (Lin2838) nonidentified in

L. innocua

1.8 Cell surfaceproteins

Lmo0130 Lin0177 LPXTG protein, similar to 5’-nucleotidase Both orthologs identified

Lmo0160 Lin0203 LPXTG protein, unknown function Both orthologs identifiedLmo0262 – LPXTG protein, InlG, unknown function Absent in L. innocuaLmo0263 – LPXTG protein, InlH, unknown function Absent in L. innocuaLmo0327 LPXTG protein, unknown function Ortholog (Lin0352) nonidentified in

L. innocuaLmo0433 (InlA) – LPXTG protein, InlA, invasion

mammalian cellsAbsent in L. innocua

Lmo0610 LPXTG protein, unknown function Ortholog (Lin0619) nonidentified inL. innocua

Lmo0842 Lin0141 LPXTG protein, unknown function Both orthologs identifiedLmo0880 Lin0879 LPXTG protein, unknown function Both orthologs identifiedLmo1413 – LPXTG protein, unknown function Absent in L. innocuaLmo1666 – LPXTG protein, unknown function Absent in L. innocuaLmo2085 – LPXTG protein, unknown function Absent in L. innocuaLmo2714 Lin2862 LPXTG protein, unknown function Both orthologs identified

Lin0514 LPXTG protein, unknown function Ortholog (Lmo0514) nonidentifiedin L. monocytogenes

– Lin2281 LPXTG protein, unknown function Absent in L. monocytogenesLmo2185 Lin2289 NXZTN protein, SvpA Both orthologs identifiedLmo2186 Lin2290 NXZTN protein, SvpB Both orthologs identified

Total 19 11

a) Functional categories were assigned according to the “ListiList functional classification codes” agreed by the Listeria Genome Euro-pean Consortium (http://genolist.pasteur.fr/ListiList/)

b) Lmo and Lin are the prefix codes used for L. monocytogenes and L. innocua proteins, respectively. In brackets, alternate protein des-ignation

Finally, it is worth to recall that among the 13 LPXTGproteins identified in L. monocytogenes, six of them (InlA,InlG, InlH, Lmo1413, Lmo1666, and Lmo2085) have noortholog in L. innocua. Whereas the internalins InlA, InlG,and InlH have been implicated in virulence [7, 22], nothing isknown about the function of the other three L. mono-cytogenes-specific LPXTG proteins identified in the pepti-doglycan.

4 Concluding remarks

To our knowledge, this study provides the first comprehen-sive proteomic analysis dedicated exclusively to analyze bac-terial proteins present in a subcellular fraction containinghighly pure peptidoglycan material. Standard protocols in-volving preparation of bacterial protein extracts for proteom-

ic analysis, either by electrophoretic or nonelectrophoreticapproaches [23, 24], include centrifugation steps to clearthe cell lysate. However, unless peptidoglycan-degradativeenzymes are used, these steps imply the removal of insolublematerial as peptidoglycan and proteins strongly associated toit. In this sense, we consider our approach novel since itaddresses the analysis of a subproteome that might be par-tially or completely discarded during preparation of standardprotein extracts. This lack of information may be especiallyrelevant for proteins bound covalently to the peptidoglycan,as those of Gram-positive bacteria bearing an LPXTG motif[2]. This conclusion is also further sustained when compar-ing the Listeria cell wall proteome with the recent report ofRamnath et al. [25], in which 2-D gel electrophoresis wasused to define a partial proteome of L. monocytogenes strainEGD-e. A total of 33 protein species were identified in mem-brane, cytosol, and total protein extracts [25]. Neither LPXTG



Table 2. Peptides identified by 2DnLC-MS/MS corresponding to LPXTG proteins of L. monocytogenes and L. innocua

Bacterial species LPXTGprotein

Proteinsize (aa)

Peptides identified by LC-MS/MS(amino acid positions)

Residue number ofLPXTG-motif a)

L. monocytogenes(EGDe)

Lmo0130 784 ILDGVSTNK (147–155)SVTPNADITAVTEDAK (349–364)VVINDFLFGGGDGFSAFK (526–543)TLTGTTLPGATVSVQK (615–630)FSVDVTSLNLK (653–663)

755–759 (LPTTG)

Lmo0160 571 EGDTMEFVLPPELK (83–96)QNIQNAVYEDFIGPK (214–228)VDLGNLTDSVK (269–279)GQLTGDNFIK (298–307)YDTYQLIETK (379–388)APQGYVLDASPVK (389–401)FTIDDTHQSLFVSK (402–415)LTVADLLPGEYQFVETK (464–480)ISTEALNVTVTK (496–507)

541–545 (LPKTG)

Lmo0262(InlG)

490 ENQISDASPLVNMTDLTVLHLEK (173–195)TGGTEWDFATSK (395–406)MPTSDITLYAR (407–417)

456–460 (LPKTS)

Lmo0263(InlH)

548 DNQITDLTPLK (101–117)ITELELSGNPLK (122–133)NVSAIAGLQSIK (134–145)ADDNKISDISPLASLPNLIEVHLK (215–238)ISDISPLASLPNLIEVHLK (220–238)NTTVPFSGTVTQPLTEAYTAVFDVDGK (328–354)QTSVTVGANELIK (355–367)QTSVTVGANELIKEPTAPTK (355–374)EGYTFTGWYDAK (382–393)VTYQSLLEEPVAPTK (430–444)VTYQSLLEEPVAPTKDGYTFTGWYDAK (430–456)WDFATGK (462–468)

515–519 (LPTAG)

Lmo0327 1348 GVFGASDETVNYVYK (1058–1072)TTAAVIPGYTLVAIPK (1095–1110)AAVIPGYTLEK (1179–1189)ANDYQLTSTFK (1211–1221)DQQGNEIALPTVDAK (1222–1236)TYHIHEAYTTK (1236–1246)LIPGYSLVAAPK (1250–1261)

1313–1317 (LPKTG)

Lmo0433(InlA)

800 PLANLTTLER (202–211)VSDISVLAK (219–227)LFFYNNK (366–372)ITQLGLNDQAWTNAPVNYK (407–425)ANVSIPNTVK (426–435)GTTTFSGTVTQPLK (482–495)EVEAGNLLTEPAKPVK (512–527)EGHTFVGWFDAQTGGTK (528–544)WNFSTDK (544–550)TGGDKWDFATSK (610–622)STTQAVDYQGLLK (648–660)STTQAVDYQGLLKEPK (648–663)WDFATDK (685–691)MPANDITLYAQFTK (692–705)

767–771 (LPTTG)

Lmo0842 2044 ITLTVKAK (1633–1640)TINLTVHPR (1780–1788)TYEEAKPLPK (2000–2009)

2007–2011 (LPKTG)



Table 2. Continued


Proteinsize (aa)



Lmo0880 462 VAYDFVITQPVASGETMTLTIPDQLK (58–83)NIQTDQVNPIAFPVK (140–154)NIQTDQVNPIAFPVKNTTQTVTPYISK (140–175)NTTQTVTPYISK (155–166)VIATNIQPIALDADR (236–250)ATLSGDNLDAVSR (283–295)NATVNDYGSGGQGTGTPPAPPVK (296–318)EEPPFIPAEK (319–328)TVETDFGPLEIVK (334–346)TVETDFGPLEIVKDSEQNGK (334–353)VKDGDTLPGVANK (360–372)DGDTLPGVANK (362–372)FDVSVAEIK (373–381)FDVSVAEIKDWNNLTSDTLQAGQK (373–396)DWNNLTSDTLQAGQK (382–396)LQLTIEK (397–403)ITVPPVQK (409–416)

433–437 (LPHTG)

Lmo1413 439 LGETYTTSPK (161–170) 405–409 (LPKTG)

Lmo1666 1711 NEFNPLDNTFWAK (260–272)TEAGAVIK (873–880)TEQAFYTDIK (1419–1428)AATSDNSPISSDFSK (1429–1443)TGNYEVLLR (1449–1457)INVLVQDTIAPVIK (1471–1484)GTPMTEQQLLAK (1495–1506)TVTEFLQDIHATTDDGSK (1578–1595)ITTDFDPNMLK (1596–1606)YTIHLNAVDADGNK (1611–1624)

1682–1686 (IPALG)

Lmo2085 562 ESATTNNIGSYLFTDVLPGDYQVK (249–272)FSLPNNDFIFSK (273–284)TGIASVNVPNLK (300–311)KFTITYGDTNPVK (393–405)GLANAVFDVK (429–439)SIDGTTLKK (439–447)GYALAENLQPGTYVITEVTAPPGYEK (454–479)VTIPFNPQK (486–494)IMVPLKPTPTK (505–515)

532–536 (LPQTG)

Lmo2714 315 TTDTTENIITEEK (42–54)TEPAEPTKDTTVTPPQK (55–71)VQTTIDITDSQEVYSYEQNSK (74–94)DFTATLTDYTDTK (99–111)LSTAQTGIQEVTLLGDNEDGK (123–143)LSTAQTGIQEVTLLGDNEDGKTTAVK (123–148)TAQLTITDMNFDLETK (157–172)QNEAVVTAPDK (182–192)QNEAVVTAPDKVK (182–194)DGYFYAEWATTPK (204–216)

283–287 (LPSTG)

L. innocua(CLIP11262)

Lin0141 1993 LSDVTVIGR (1809–1817)YDGVTSVSR (1832–1840)VAITPSENSNSDVKPGK (1932–1948)TATYQQAIPR (1951–1960)

1958–1962 (IPRTG)

Lin0177 785 ILDGVSTNK (147–155)QDVAGEAANMMTK (268–280)

756–760 (LPTTG)



Table 2. Continued


Proteinsize (aa)



TTGDFVSPPDAK (331–342)SVTPNADITAVTEDAK (349–364)VASVTTEDGTPLKADQK (507–523)SQAEIDKETEDAAIK (584–598)

Lin0203 586 VNLGNLTDSVK (269–279)YDTYQLIETK (379–388)SNLITDNAGK (454–463)IVSSDEQPTTLPK (546–558)

556–560 (LPKTG)

Lin0514 611 LTYLYAQNSMK (156–166)ALAAFGQNTGR (208–218)VVDLTTPGK (465–473)ANNTTLTTNENVR (555–567)

578–582 (LPKTG)

Lin0879 463 ATLSGDNLDAVSR (295–307)ITVPTPQK (410–417)LQLTIEK (398–404)

434–438 (LPHTG)

Lin2281 1622 GNVILTKK (1105–1112) 1589–1593 (LPSTG)

Lin2862 320 NTITEDPSKTPEPPK (48–62)IALKDFTDTLTTHTGTK (104–120)DFTDTLTTHTGTK (108–120)ADKDGNFYAQWATVPK (210–225)

288–292 (LPATG)

a) In brackets, specific sequence of each protein matching the LPXTG motif

proteins nor peptidoglycan lytic enzymes were identified inany of these three fractions. These differences highlight theunique features of the cell wall proteome described in ourstudy.

The application of the 2DnLC-MS/MS nonelectro-phoretic proteomic approach has allowed to decipher thetype of proteins that co-purify with peptidoglycan uponextensive boiling in 4% SDS. Due to the peculiar biochemicalproperties of this material, we obviated the usage of pepti-doglycan-degradative enzymes and, instead, performed an‘in situ’ trypsin digestion that led to the rapid collection of apeptide mixture representative of the cell wall proteome.This approach led to the unequivocal identification of 30protein species, 19 in L. monocytogenes EGD-e and 11 inL. innocua CLIP11262, in a single growth condition (BHImedium, stationary phase, 377C). Taking into account boththe speed of the procedure and its high sensitivity (only 4 mgof peptide mixture were required to obtain this bulk of data),we are currently pursuing the analysis of the Listeria cell wallproteome in diverse environmental conditions. Extendingthe cell wall proteome information to other growth condi-tions will lead to a more rational generation of knockoutmutants in surface proteins to be used in functional studies.The demonstration that besides InlA, L. monocytogenes syn-thesizes at detectable levels at least 12 additional LPXTGproteins, including the internalins InlG and InlH, certainlyopens new avenues of research. It will also be of great inter-est to pursue studies addressing the role of the rest of LPXTG

proteins identified, all of them of unknown function. On theother hand, the analysis of the cell wall proteome in mutantsdefective in sortases, like the two identified in Listeria (SrtA,SrtB), can also provide rapidly very valuable information onthe substrate specificity existing during the process of pro-tein anchoring to the peptidoglycan.

In summary, our work corroborates recent studiesshowing that the 2DnLC-MS/MS proteomic technology is arapid, sensitive, and economically affordable method toidentify bacterial proteins in complex mixtures [23, 24]. Inaddition, here we demonstrate using Listeria as a modelorganism that the technique is valuable for identifying froma proteomic perspective bacterial proteins strongly associatedto the cell wall. Considering that these proteins are exposedin the bacterial surface, and that the few cases in which theirfunction is known it is related to pathogenicity, this prote-omic approach becomes highly attractive for application toother Gram-positive pathogens causing important humandiseases. Such bacterial pathogens include Streptococci,Staphylococci, and Clostridi, as well as bioterrorism agentssuch as Bacillus anthracis. The rapid analysis of many bacte-rial cell wall proteomes in a near future will facilitate theidentification of new surface-exposed targets that cancertainly improve drug and vaccine design.

This work was supported by grants QLG2-CT-1999–00932from the European Union and BIO2000–3202-CE from theSpanish Ministry of Science and Technology. M. G. Pucciarelli is



supported by a Researcher Post-doctoral Contract from the ‘Con-sejería de Educación of Comunidad de Madrid’, Spain (ref. 02/0335/2002). H. Bierne is in the staff of the Institut National dela Recherche Agronomique. We are indebted to E. Camafeita forcritical reading of the manuscript.

5 References

[1] Hancock, I. C., Biochem. Soc. Trans. 1997, 25, 183–187.

[2] Navarre, W. W., Schneewind, O., Microbiol. Mol. Biol. Rev.1999, 63, 174–229.

[3] Mazmanian, S. K., Ton-That, H., Schneewind, O., Mol.Microbiol. 2001, 40, 1049–1057.

[4] Rigden, D. J., Galperin, M. Y., Jedrzejas, M. J., Crit. Rev. Bio-chem. Mol. Biol. 2003, 38, 143–168.

[5] Roche, F. M., Massey, R., Peacock, S. J., Day, N. P. et al.,Microbiology 2003, 149, 643–654.

[6] Pallen, M. J., Chaudhuri, R. R., Henderson, I. R., Curr. Opin.Microbiol. 2003, 6, 519–527.

[7] Cabanes, D., Dehoux, P., Dussurget, O., Frangeul, L., Cossart,P., Trends Microbiol. 2002, 10, 238–245.

[8] Glaser, P., Frangeul, L., Buchrieser, C., Rusniok, C. et al., Sci-ence 2001, 294, 849–852.

[9] Doumith, M., Cazalet, C., Simoes, N., Frangeul, L. et al.,Infect. Immun. 2004, 72, 1072–1083.

[10] Vazquez-Boland, J. A., Kuhn, M., Berche, P., Chakraborty, T.et al., Clin. Microbiol. Rev. 2001, 14, 584–640.

[11] Cossart, P., Pizarro-Cerda, J., Lecuit, M., Trends Cell Biol.2003, 13, 23–31.

[12] Portnoy, D. A., Auerbuch, V., Glomski, I. J., J. Cell Biol. 2002,158, 409–414.

[13] Buchrieser, C., Rusniok, C., Kunst, F., Cossart, P., Glaser, P.,FEMS Immunol. Med. Microbiol. 2003, 35, 207–213.

[14] Pallen, M. J., Lam, A. C., Antonio, M., Dunbar, K., TrendsMicrobiol. 2001, 9, 97–102.

[15] Bierne, H., Mazmanian, S. K., Trost, M., Pucciarelli, M. G. etal., Mol. Microbiol. 2002, 43, 869–881.

[16] Patterson, S. D., Aebersold, R. H., Nat. Genet. 2003, 33Suppl., 311–323.

[17] Schubert, K., Bichlmaier, A. M., Mager, E., Wolff, K. et al.,Arch. Microbiol. 2000, 173, 21–28.

[18] Lenz, L. L., Mohammadi, S., Geissler, A., Portnoy, D. A., Proc.Natl. Acad. Sci. USA 2003, 100, 12432–12437.

[19] Carroll, S. A., Hain, T., Technow, U., Darji., A. et al., J. Bac-teriol. 2003, 185, 6801–6808.

[20] Bierne, H., Garandeau, C., Pucciarelli, M. G., Sabet, C. et al.,J. Bacteriol. 2004, 186, 1972–1982.

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

[22] Raffelsbauer, D., Bubert, A., Engelbrecht, F., Scheinpflug, J.et al., Mol. Gen. Genet. 1998, 260, 144–158.

[23] Kolker, E., Purvine, S., Galperin, M. Y., Stolyar, S. et al., J.Bacteriol. 2003, 185, 4593–4602.

[24] Corbin, R. W., Paliy, O., Yang, F., Shabanowitz, J. et al., Proc.Natl. Acad. Sci. USA 2003, 100, 9232–9237.

[25] Ramnath, M., Rechinger, K. B., Jansch, L., Hastings, J. W. etal., Appl. Environ. Microbiol. 2003, 69, 3368–3376.


analysis of the listeria cell wall proteome by two-dimensional nanoliquid chromatography coupled to...

Documents