Identification of Surface Proteins of Helicobacter pylori by Selective Biotinylation, Affinity Purification, and Two-dimensional Gel Electrophoresis

Abstract

Helicobacter pylori is a widespread human pathogen that can cause gastric ulcers and cancer. To identify surface proteins that may play a role in pathogen-host interactions and represent potential targets for the control of this infection, we selectively biotinylated intact H. pylori with the hydrophilic reagent sulfosuccinimidyl-6-(biotinamido)-hexanoate and purified the labeled proteins by membrane isolation, solubilization, and affinity chromatography. After separation of 82 biotinylated proteins on two-dimensional gels, 18 were identified with comparison to proteome data and peptide mass fingerprinting. Among the identified proteins, 9 have previously been shown to be surface-exposed, 7 are associated with virulence, and 11 are highly immunogenic in infected patients. In conclusion, this generally applicable combined proteome approach facilitates the rapid identification of promising targets for the control of H. pylori and might be applicable to numerous other human pathogens although larger biotinylation reagents might be required in some cases to prevent permeation of porin channels in the outer membrane. Helicobacter pylori is a widespread human pathogen that can cause gastric ulcers and cancer. To identify surface proteins that may play a role in pathogen-host interactions and represent potential targets for the control of this infection, we selectively biotinylated intact H. pylori with the hydrophilic reagent sulfosuccinimidyl-6-(biotinamido)-hexanoate and purified the labeled proteins by membrane isolation, solubilization, and affinity chromatography. After separation of 82 biotinylated proteins on two-dimensional gels, 18 were identified with comparison to proteome data and peptide mass fingerprinting. Among the identified proteins, 9 have previously been shown to be surface-exposed, 7 are associated with virulence, and 11 are highly immunogenic in infected patients. In conclusion, this generally applicable combined proteome approach facilitates the rapid identification of promising targets for the control of H. pylori and might be applicable to numerous other human pathogens although larger biotinylation reagents might be required in some cases to prevent permeation of porin channels in the outer membrane. cytotoxin-associated genes sulfosuccinimidyl-6-(biotinamido)hexanoate matrix-assisted laser desorption ionization-mass spectrometry isoelectric focusing hypothetical outer membrane protein 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid Helicobacter pylori is a micro-aerophilic, Gram-negative bacterium that colonizes in the stomach of ∼50% of the world human population (1Walker M.M. Crabtree J.E. Ann. N. Y. Acad. Sci. 1998; 859: 96-111Crossref PubMed Scopus (60) Google Scholar, 2Telford J.L. Covacci A. Rappuoli R. Chiara P. Curr. Opin. Immunol. 1997; 9: 498-503Crossref PubMed Scopus (71) Google Scholar). Infection results in chronic inflammation of the gastric mucosa, which in most cases remains asymptomatic. However, in 10% of those infected severe pathological consequences develop including gastric and duodenal ulcer, atrophic gastritis, adenocarcinoma, or mucosa-associated lymphoid tissue (MALT) lymphoma. The surface of H. pylori provides an important interface for pathogen-host interactions including flagella-driven motility in the stomach mucus layer, local buffering of stomach acid by urease, adhesion to gastric epithelial cells mediated by several adhesins including BabA (3Ilver D. Arnqvist A. Ogren J. Frick I.M. Kersulyte D. Incecik E.T. Berg D.E. Covacci A. Engstrand L. Boren T. Science. 1998; 279: 373-377Crossref PubMed Scopus (991) Google Scholar), AlpA/AlpB (4Odenbreit S. Till M. Hofreuter D. Faller G. Haas R. Mol. Microbiol. 1999; 31: 1537-1548Crossref PubMed Scopus (207) Google Scholar), HopZ (5Peck B. Ortkamp M. Diehl K.D. Hundt E. Knapp B. Nucleic Acids Res. 1999; 27: 3325-3333Crossref PubMed Scopus (150) Google Scholar), translocation of CagA into host cells via the type IV secretion apparatus encoded in the Cag1(cytotoxin-associated genes) pathogenicity island (6Segal E.D. Cha J., Lo, J. Falkow S. Tompkins L.S. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 14559-14564Crossref PubMed Scopus (668) Google Scholar, 7Backert S. Ziska E. Brinkmann V. Zimny-Arndt U. Fauconnier A. Jungblut P.R. Naumann M. Meyer T.F. Cell. Microbiol. 2000; 2: 155-164Crossref PubMed Scopus (369) Google Scholar, 8Odenbreit S. Puls J. Sedlmaier B. Gerland E. Fischer W. Haas R. Science. 2000; 287: 1497-1500Crossref PubMed Scopus (1070) Google Scholar, 9Asahi M. Azuma T. Ito S. Ito Y. Suto H. Nagai Y. Tsubokawa M. Tohyama Y. Maeda S. Omata M. Suzuki T. Sasakawa C. J. Exp. Med. 2000; 191: 593-602Crossref PubMed Scopus (424) Google Scholar, 10Stein M. Rappuoli R. Covacci A. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 1263-1268Crossref PubMed Scopus (519) Google Scholar), and adherence of neutrophils to endothelial cells induced by the neutrophil-activating protein (NAP) (11Evans D.J., Jr. Evans D.G. Takemura T. Nakano H. Lampert H.C. Graham D.Y. Granger D.N. Kvietys P.R. Infect. Immun. 1995; 63: 2213-2220Crossref PubMed Google Scholar). Envelope proteins are particularly accessible to host immune responses and to drug therapy and thus could be used to control this important human pathogen. Because of its outstanding importance, several approaches have been developed to identify and characterize additional H. pylori envelope proteins. Global theoretical predictions of outer membrane proteins based on the two completely sequenced genomes and physico-chemical properties and sequence similarities to proteins with known surface localization have yielded 64 putative candidates (12Alm R.A. Bina J. Andrews B.M. Doig P. Hancock R.E. Trust T.J. Infect. Immun. 2000; 68: 4155-4168Crossref PubMed Scopus (260) Google Scholar). Several of these have been experimentally verified, but it is unclear if all of the predicted surface proteins are expressed. Moreover, several proposed surface proteins including urease, Hsp60, Hsp70, flagellar sheath protein, and catalase do not share common properties of surface proteins and are therefore neglected by theoretical predictions. Surface proteins including outer membrane proteins can be experimentally identified by selective solubilization and sucrose gradient centrifugation for enrichment of outer membrane fractions. InH. pylori, these techniques yielded enriched outer membrane fractions, and many putative surface proteins have been found. However, these preparations are heavily contaminated by inner membrane components (13Nilsson C.L. Larsson T. Gustafsson E. Karlsson K.A. Davidsson P. Anal. Chem. 2000; 72: 2148-2153Crossref PubMed Scopus (94) Google Scholar, 14Exner M.M. Doig P. Trust T.J. Hancock R.E. Infect. Immun. 1995; 63: 1567-1572Crossref PubMed Google Scholar), so that independent evidence, such as specific antibody binding to intact bacteria, is necessary to verify true surface proteins (15Doig P. Trust T.J. Infect. Immun. 1994; 62: 4526-4533Crossref PubMed Google Scholar). Five outer membrane proteins with the typical β-barrel structure were identified by their anomalous temperature-dependent migration in polyacrylamide gels (14Exner M.M. Doig P. Trust T.J. Hancock R.E. Infect. Immun. 1995; 63: 1567-1572Crossref PubMed Google Scholar), but surface proteins with different structures escaped this method. Selective labeling with reactive reagents is an alternative approach to identify surface proteins. In H. pylori urease could be labeled with 125I demonstrating its surface localization (16Dunn B.E. Perez-Perez G.I. Blaser M.J. Infect. Immun. 1989; 57: 1825-1833Crossref PubMed Google Scholar). Labeling with a reactive sugar ligand confirmed the specific binding of the Lewisb antigen to the surface-exposedH. pylori adhesin BabA (3Ilver D. Arnqvist A. Ogren J. Frick I.M. Kersulyte D. Incecik E.T. Berg D.E. Covacci A. Engstrand L. Boren T. Science. 1998; 279: 373-377Crossref PubMed Scopus (991) Google Scholar). The hydrophilic reagent S-NHS-LC-biotin (sulfosuccinimidyl-6-(biotinamido)hexanoate) allowed us to selectively biotinylate surface proteins of Escherichia coli although to a lesser extent periplasmic proteins were also labeled (17Bradburne J.A. Godfrey P. Choi J.H. Mathis J.N. Appl. Environ. Microbiol. 1993; 59: 663-668Crossref PubMed Google Scholar). Here, we confirmed the periplasmic labeling for the well characterized MalE protein of E. coli and showed that lysis followed by membrane purification removes all of the labeled MalE protein. We then applied this technique to identify surface proteins of H. pylori. Biotinylated membrane proteins of H. pyloriwere further enriched by avidin-affinity chromatography, separated by two-dimensional gel electrophoresis, and identified by peptide mass fingerprinting. This generally applicable approach led to the rapid experimental identification of both new and previously known H. pylori envelope and putative surface proteins. E. coli HB2151 was grown in M9 minimal medium supplemented with 0.2% maltose, to induce the mal operon (18Ehrmann M. Boos W. J. Bacteriol. 1987; 169: 3539-3545Crossref PubMed Google Scholar), 0.1% hydrolysated casein, and 1 μm biotin to block active uptake of the biotinylation reagent. The cells were harvested at an OD600 nm of 0.6–0.7, corresponding to the mid-exponential growth phase. H. pylori strain 26695 (19Eaton K.A. Morgan D.R. Krakowka S. Infect. Immun. 1989; 57: 1119-1125Crossref PubMed Google Scholar) was cultured at 37 °C in a micro-aerobic atmosphere (5% O2, 85% N2, and 10% CO2) on serum-agar plates (20Odenbreit S. Wieland B. Haas R. J. Bacteriol. 1996; 178: 6960-6967Crossref PubMed Google Scholar) for 3 days, and grown for 1 additional day on fresh plates. The bacteria were harvested and suspended in ice-cold phosphate-buffered saline, 1 mmCaCl2, 0.5 mm MgCl2, 1.6 mm biotin at an OD600 nm of 2.5–3.5 (equivalent to 1–2 × 109 colony-forming units). The bacteria were surface-labeled by incubation with 200 μm(final concentration) S-NHS-LC-biotin (Pierce) for 30 min on ice. The reaction was stopped by adding two volumes of TNKCM (50 mmTris, pH 7.4, 100 mm NaCl, 27 mm KCl, 1 mm CaCl2, 0.5 mmMgCl2). After 10 min of incubation at room temperature, the bacteria were sedimented by centrifugation at 3,500 ×g for 10 min and washed three times with TNKCM. The viability of the bacteria before and after labeling was determined by plating on serum-agar and by flow cytometry as well as fluorescence microscopy using a membrane-permeable (Syto9) and a membrane-impermeable (propidium iodide) fluorophore according to the manufacturer's instructions (LIVE/DEAD kit, Molecular Probes) except that an altered ratio in millimolar of 3:27 of the dyes Syto9/propidium iodide was used. To isolate biotinylated membrane proteins, labeled bacteria were resuspended in 50 mm Tris-HCl, pH 7.4, 1 mmMgCl2 with protease 1 μm 1 μm mm and by a at After of intact bacteria by two at × and °C for 10 were at × and °C for 30 min and then and resuspended in (50 mm KCl, 10 mm protease pH were to a protein of with and for 1 at °C with were by for 1 at × and The was purified by affinity on binding according to the instructions of the Molecular with In of membrane proteins were in 100 mm NaCl, pH and with 1 of in mm mm NaCl, pH 0.2% After 30 min of incubation at room temperature, the was washed times with of The biotinylated proteins were by the times for min at 37 °C with were and by and on were according to Biotinylated proteins were on the using (Pierce) and according to the manufacturer's The protein MalE was with a antibody and a gel and protein identification by were as previously P.R. D. Haas G. Zimny-Arndt U. P. S. A. Meyer T.F. Mol. Microbiol. 2000; PubMed Scopus Google Scholar). In protein were for 30 min at in mm the of protein a × 30 two-dimensional gel was used. the identification of proteins or were applied to the of the were identified by peptide mass after of The peptide was with a acid in and were applied to the of a mass mass were with the using all H. pylori proteins in the data a mass of 100 for the peptide two of acid at the and of by were in these Surface proteins may important pathogen-host interactions in H. pylori To identify surface proteins, we a that been developed for E. coli (17Bradburne J.A. Godfrey P. Choi J.H. Mathis J.N. Appl. Environ. Microbiol. 1993; 59: 663-668Crossref PubMed Google Scholar). In this bacteria are with the hydrophilic biotinylation reagent S-NHS-LC-biotin that a of its this with of surface proteins. However, some of the reagent the channels of E. coli outer membrane in labeling of periplasmic proteins, which the localization of previously proteins. To these we labeled E. coli and intact separated the by two-dimensional gel electrophoresis, and biotin by not all proteins of the labeled were biotinylated that most E. coli proteins accessible In labeling of intact bacteria in a of biotinylated putative surface proteins. However, which is a well characterized periplasmic protein and can be with a was also heavily We to the labeling by the reagent with the but this not prevent biotinylation of We then to by using a larger biotinylation reagent. periplasmic labeling to be the of permeation of the reagent porin of the reagent might prevent to the in We the by via its to The biotinylation a mass of which the of porin channels in E. H. J. Bacteriol. PubMed Google and H. J. PubMed Scopus Google T. J. Chem. PubMed Google Scholar). this the biotinylation of MalE in comparison with other protein but the MalE labeling was not not the reagent was to we its the to were we to the periplasmic proteins by lysis followed by membrane After this membrane MalE protein that the proteins were the biotinylated proteins most represent true surface proteins We then applied this to H. pylori. In to E. this is to in M. D. B.E. Infect. Immun. 1996; PubMed Google Scholar, D.R. PubMed Scopus Google Scholar), which could prevent To if such lysis the viability of the bacteria was determined before and after labeling by fluorescence and flow using a of membrane-permeable and dyes (LIVE/DEAD that 10% of the bacteria were incubation and Moreover, of the bacteria a membrane after labeling as determined by The of bacteria might have some but most of this be components are not to the labeled proteins. In H. pylori, mass of data not as shown by two-dimensional electrophoresis, and H. were all proteins that could be by with were also that all proteins accessible Among the proteins that not were the previously identified and P.R. D. Haas G. Zimny-Arndt U. P. S. A. Meyer T.F. Mol. Microbiol. 2000; PubMed Scopus Google Scholar). In and the urease were in but in the with other This is Hsp60, and urease many these are not accessible to the biotinylation reagent. In to the labeling of proteins, labeled intact H. pylori that were biotinylated and in three independent Several as of and such might represent different of a protein this into protein were isoelectric in the comparison of labeled intact cells labeled and a of biotinylation as for a hydrophilic reagent and intact In the reagent biotin labeled several known proteins, permeation of the on with labeled periplasmic proteins in E. coli that can be by membrane we enriched putative biotinylated proteins by lysis and membrane in the of Helicobacter this not in a in the two-dimensional of biotinylated proteins, and all were that if periplasmic proteins been could be that H. pylori strain 26695 may porin channels which the biotinylation reagent can to the are required to this To further the biotinylated proteins, were with and applied to a and washed and with of these purified labeled proteins corresponding to different as by most of which T. C. A. Infect. Immun. 1997; PubMed Google Scholar, C. A. Mol. Microbiol. 1999; 31: PubMed Scopus Google Scholar, B.E. M.M. T. Immun. 1997; PubMed Google Scholar, K.A. C.L. Morgan D.R. Krakowka S. Infect. Immun. 59: PubMed Google Scholar, P. C. Haas R. M. J. E. 1994; PubMed Scopus Google of were biotinylated The were of proteins as by their Moreover, additional corresponding to different in the but were not by to their with biotinylated H. pylori, the binding of purified proteins were but and to of and in these additional escaped in of In all but of biotinylated proteins H. pylori could also be in the enriched protein To identify the biotinylated proteins, we the binding with the H. pylori strain 26695 in data P.R. D. Haas G. Zimny-Arndt U. P. S. A. Meyer T.F. Mol. Microbiol. 2000; PubMed Scopus Google Scholar). The of 18 biotinylated with previously proteins To these the most biotinylated were the and identified by peptide mass fingerprinting. of the were with the which confirmed the of the of proteins, peptide that theoretical of the corresponding protein This allowed and sequence was but all mass with and 10 and additional proteins with sequence the identification based on comparison to the proteome the of this proteins of H. of in peptide mass identified by comparison with the proteome or in other by infected according to combined data three M. T. Berg D.E. Immunol. 1998; PubMed Google Scholar, B. A. R. R. W. D. Infect. Immun. 2000; 68: PubMed Scopus Google Scholar, G. G. S. Zimny-Arndt U. S. Meyer T.F. T. Jungblut P.R. 2: PubMed Scopus Google Scholar, H. B. V. J. Bacteriol. 1989; PubMed Google not in a new of in peptide mass identified by comparison with the proteome or in other by infected according to combined data three M. T. Berg D.E. Immunol. 1998; PubMed Google Scholar, B. A. R. R. W. D. Infect. Immun. 2000; 68: PubMed Scopus Google Scholar, G. G. S. Zimny-Arndt U. S. Meyer T.F. T. Jungblut P.R. 2: PubMed Scopus Google Scholar, H. B. V. J. Bacteriol. 1989; PubMed Google Scholar). not proteins of H. pylori as in the two-dimensional In most labeled protein as which could represent the protein with that the This was confirmed for three of a biotinylated corresponding to catalase not a of with the biotinylation reagent in a of and The of were and in the of of which is with this a of biotinylation many were not Moreover, of were enriched the avidin-affinity chromatography, which further for highly biotinylated Surface proteins of H. pylori important pathogen-host interactions that are for and of this pathogen. To pylori surface proteins, several approaches have been used the a proteome we combined a selective biotinylation of with membrane followed by affinity purification of proteins, two-dimensional electrophoresis, and peptide mass fingerprinting. for this is the of that are to the The that except for two predicted proteins all or several of these to be on the of the corresponding protein all protein pylori can be labeled using the hydrophilic biotinylation reagent S-NHS-LC-biotin that with In in intact bacteria a of the proteome was that the the of most proteins for the highly hydrophilic biotinylation reagent. the inner membrane is for the (17Bradburne J.A. Godfrey P. Choi J.H. Mathis J.N. Appl. Environ. Microbiol. 1993; 59: 663-668Crossref PubMed Google Scholar), which we confirmed in this have shown that in E. the outer membrane is for in labeling of periplasmic proteins. membrane channels that the of hydrophilic with to T. J. Chem. PubMed Google Scholar, R. J. PubMed Scopus Google Scholar, G. N. J. H. P. H. C. Proc. Natl. Acad. Sci. U. S. A. PubMed Scopus Google Scholar), which is well the mass of S-NHS-LC-biotin are the cause of of the biotinylation reagent the could its for surface proteins. We a reagent with a mass of which showed but not in a well E. coli This that larger reagents might be highly specific but of the of we could not this we that the reaction with proteins could be completely by lysis with a membrane applied to H. pylori, this approach showed that a biotinylated were membrane that the outer membrane of H. pylori may be for S-NHS-LC-biotin with the outer membrane of E. have for the H. pylori (14Exner M.M. Doig P. Trust T.J. Hancock R.E. Infect. Immun. 1995; 63: 1567-1572Crossref PubMed Google Scholar), P. M.M. Hancock R.E. Trust T.J. J. Bacteriol. 1995; PubMed Google Scholar), and B. Ortkamp M. U. M. Hundt E. Knapp B. Infect. PubMed Scopus Google Scholar) coli with a of In Helicobacter might have a larger but this porin is not in B. Ortkamp M. U. M. Hundt E. Knapp B. Infect. PubMed Scopus Google Scholar). The of porin channels that permeation of the S-NHS-LC-biotin in H. pylori is also by the that the of identified biotinylated proteins are to be true H. pylori surface proteins based on with independent which that the labeling was highly selective for surface proteins. In 82 putative H. pylori protein were by selective labeling and membrane isolation, and 18 of these were is that H. some additional surface proteins that escaped proteins were labeled in that their localization could not be using this although their surface localization previously been B.E. J. Med. 1998; Google Scholar, M. Infect. Immun. 1998; PubMed Google Scholar, B.E. M.M. T. Immun. 1997; PubMed Google Scholar). Among the 18 identified proteins, several have been predicted to be in H. pylori based on results independent that urease and a of the flagellar sheath protein, are on the surface M. D. B.E. Infect. Immun. 1996; PubMed Google Scholar, B.E. J. Med. 1998; Google Scholar, V. 31: PubMed Scopus Google Scholar, Evans D.G. 2000; PubMed Scopus Google Scholar). the and have been to be into the medium P. Infect. Immun. 1998; PubMed Google Scholar, D. S. M. Zimny-Arndt U. N. Meyer T.F. Jungblut P.R. Infect. Immun. PubMed Scopus Google Scholar). The surface-exposed, proteins and previously characterized proteins is with the of of proteins at the H. pylori M. D. B.E. Infect. Immun. 1996; PubMed Google Scholar, D.R. PubMed Scopus Google Scholar, R. J. PubMed Scopus Google Scholar). is to an outer membrane protein E. coli J.E. R.A. M. Trust T.J. Hancock R.E. 2000; PubMed Scopus Google Scholar, A. C. J. C. B. J. PubMed Scopus Google Scholar), is to an outer membrane protein of 1993; PubMed Scopus Google Scholar). In the of is to be in the based on data coli J. B. G. E. J. Chem. 1997; PubMed Scopus Google Scholar). However, a of this protein might be other periplasmic proteins that have been in this and to be pylori In the that surface proteins, although independent be for the several candidates that have not been previously of 18 identified proteins have been predicted to be surface proteins, and of the hypothetical outer membrane proteins have been found. The strain is known to BabA several of the (3Ilver D. Arnqvist A. Ogren J. Frick I.M. Kersulyte D. Incecik E.T. Berg D.E. Covacci A. Engstrand L. Boren T. Science. 1998; 279: 373-377Crossref PubMed Scopus (991) Google Scholar, B. Ortkamp M. Diehl K.D. Hundt E. Knapp B. Nucleic Acids Res. 1999; 27: 3325-3333Crossref PubMed Scopus (150) Google Scholar). been to be in this strain M. T. Berg D.E. Immunol. 1998; PubMed Google Scholar) and might be the labeled protein Several surface proteins of H. pylori important This is also the for some of the 18 proteins that were identified in this of have been previously as Moreover, the flagellar sheath protein is of that are also for is a of the pathogenicity island that is known to responses to H. pylori. is for the of the which is a of the type IV secretion A. J.L. G. J. Rappuoli R. Science. 1999; PubMed Scopus Google W. Puls J. R. B. S. Haas R. Mol. Microbiol. PubMed Scopus Google Scholar). The secretion in cells J. L. M. G. N. W. J. Med. Microbiol. Scholar). other human of catalase J. Med. Microbiol. PubMed Scopus Google Scholar, S. J. V. J. Bacteriol. PubMed Scopus Google Scholar, Jr. J.L. Infect. Immun. 2000; 68: PubMed Scopus Google Scholar) and the protease and M.J. Mol. Microbiol. 1997; PubMed Scopus Google Scholar) are important for all fresh human of H. pylori which is for of the in N. Meyer T.F. Mol. Microbiol. 2000; PubMed Scopus Google Scholar). a potential role for in H. pylori is be to characterize in the additional surface proteins with known in other Surface proteins of H. pylori are particularly to the host immune and therefore might represent 11 of 18 identified proteins are the previously that are by a of infected M. T. Berg D.E. Immunol. 1998; PubMed Google Scholar, B. A. R. R. W. D. Infect. Immun. 2000; 68: PubMed Scopus Google Scholar, G. G. S. Zimny-Arndt U. S. Meyer T.F. T. Jungblut P.R. 2: PubMed Scopus Google Scholar). This that identification enriched for highly immunogenic of 18 surface proteins of This of proteins therefore a to antigen candidates for of 18 identified proteins have previously been shown to induce a H. pylori in the We have shown that and are also highly in the Helicobacter R. T. and D. are In conclusion, a rapid proteome approach been developed to identify surface proteins of H. pylori that are promising targets for the control of this important human pathogen. larger biotinylation reagents this approach be generally applicable to characterize the surface of pathogens to identify new potential proteins for drug therapy and We (Pierce) for on the of a larger biotinylation reagent. We for the H. sequence for proteins.