Bacterial Interactomes: Interacting Protein Partners Share Similar Function and Are Validated in Independent Assays More Frequently Than Previously Reported
2016; Elsevier BV; Volume: 15; Issue: 5 Linguagem: Inglês
10.1074/mcp.m115.054692
ISSN1535-9484
AutoresMaxim Shatsky, Simon Allen, B. Gold, Nancy L. Liu, Thomas R. Juba, Sonia A. Reveco, Dwayne A. Elias, Ramadevi Prathapam, Jennifer He, Wenhong Yang, Evelin D. Szakal, Haichuan Liu, Mary E. Singer, Jil T. Geller, Bonita R. Lam, Avneesh K. Saini, Valentine V. Trotter, Steven C. Hall, Susan J. Fisher, Steven E. Brenner, Swapnil R. Chhabra, Terry C. Hazen, Judy D. Wall, H. Ewa Witkowska, Mark D. Biggin, John‐Marc Chandonia, Gareth Butland,
Tópico(s)Biotin and Related Studies
ResumoNumerous affinity purification-mass spectrometry (AP-MS) and yeast two-hybrid screens have each defined thousands of pairwise protein-protein interactions (PPIs), most of which are between functionally unrelated proteins. The accuracy of these networks, however, is under debate. Here, we present an AP-MS survey of the bacterium Desulfovibrio vulgaris together with a critical reanalysis of nine published bacterial yeast two-hybrid and AP-MS screens. We have identified 459 high confidence PPIs from D. vulgaris and 391 from Escherichia coli. Compared with the nine published interactomes, our two networks are smaller, are much less highly connected, and have significantly lower false discovery rates. In addition, our interactomes are much more enriched in protein pairs that are encoded in the same operon, have similar functions, and are reproducibly detected in other physical interaction assays than the pairs reported in prior studies. Our work establishes more stringent benchmarks for the properties of protein interactomes and suggests that bona fide PPIs much more frequently involve protein partners that are annotated with similar functions or that can be validated in independent assays than earlier studies suggested. Numerous affinity purification-mass spectrometry (AP-MS) and yeast two-hybrid screens have each defined thousands of pairwise protein-protein interactions (PPIs), most of which are between functionally unrelated proteins. The accuracy of these networks, however, is under debate. Here, we present an AP-MS survey of the bacterium Desulfovibrio vulgaris together with a critical reanalysis of nine published bacterial yeast two-hybrid and AP-MS screens. We have identified 459 high confidence PPIs from D. vulgaris and 391 from Escherichia coli. Compared with the nine published interactomes, our two networks are smaller, are much less highly connected, and have significantly lower false discovery rates. In addition, our interactomes are much more enriched in protein pairs that are encoded in the same operon, have similar functions, and are reproducibly detected in other physical interaction assays than the pairs reported in prior studies. Our work establishes more stringent benchmarks for the properties of protein interactomes and suggests that bona fide PPIs much more frequently involve protein partners that are annotated with similar functions or that can be validated in independent assays than earlier studies suggested. Proteins often function by interacting with partner proteins to form complexes, which range from heterodimers to large macromolecular assemblies (1Kristensen A.R. Foster L.J. High throughput strategies for probing the different organizational levels of protein interaction networks.Mol. Biosyst. 2013; 9: 2201-2212Crossref PubMed Scopus (10) Google Scholar, 2Vidal M. Cusick M.E. Barabási A.L. Interactome networks and human disease.Cell. 2011; 144: 986-998Abstract Full Text Full Text PDF PubMed Scopus (1183) Google Scholar). If we can accurately learn the heteromeric interactions that each protein makes, it will greatly aid the modeling of all aspects of cellular biochemistry and physiology. Over the last 15 years, protein-protein "interactomes" have been characterized on a genome-wide scale in bacteria and eukaryotes by yeast 2-hybrid (Y2H) 1The abbreviations used are:Y2Hyeast two-hybridAP-MSaffinity purification-mass spectrometryBLASTBasic Local Alignment Search ToolCYcytoplasmic proteinFDRfalse discovery rateNSAFnormalized spectral abundance factorOMouter membrane proteinPEperiplasmic proteinPPIprotein-protein interactionTIGRThe Institute of Genome ResearchFNfalse negative. and affinity purification-mass spectrometry (AP-MS) screens (1Kristensen A.R. Foster L.J. High throughput strategies for probing the different organizational levels of protein interaction networks.Mol. Biosyst. 2013; 9: 2201-2212Crossref PubMed Scopus (10) Google Scholar, 3Häuser R. Ceol A. Rajagopala S.V. Mosca R. Siszler G. Wermke N. Sikorski P. Schwarz F. Schick M. Wuchty S. Aloy P. Uetz P. A second-generation protein-protein interaction network of Helicobacter pylori.Mol. Cell. Proteomics. 2014; 13: 1318-1329Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar, 4Rajagopala S.V. Sikorski P. Kumar A. Mosca R. Vlasblom J. Arnold R. Franca-Koh J. Pakala S.B. Phanse S. Ceol A. Häuser R. Siszler G. Wuchty S. Emili A. Babu M. et al.The binary protein-protein interaction landscape of Escherichia coli.Nat. Biotechnol. 2014; 32: 285-290Crossref PubMed Scopus (148) Google Scholar, 5Yu H. Braun P. Yildirim M.A. Lemmens I. Venkatesan K. Sahalie J. Hirozane-Kishikawa T. Gebreab F. Li N. Simonis N. Hao T. Rual J.F. Dricot A. Vazquez A. Murray R.R. et al.High quality binary protein interaction map of the yeast interactome network.Science. 2008; 322: 104-110Crossref PubMed Scopus (1109) Google Scholar, 6Marchadier E. Carballido-López R. Brinster S. Fabret C. Mervelet P. Bessières P. Noirot-Gros M.F. Fromion V. Noirot P. An expanded protein-protein interaction network in Bacillus subtilis reveals a group of hubs: exploration by an integrative approach.Proteomics. 2011; 11: 2981-2991Crossref PubMed Scopus (44) Google Scholar, 7Parrish J.R. Yu J. Liu G. Hines J.A. Chan J.E. Mangiola B.A. Zhang H. Pacifico S. Fotouhi F. DiRita V.J. Ideker T. Andrews P. Finley Jr., R.L. A proteome-wide protein interaction map for Campylobacter jejuni.Genome Biol. 2007; 8: R130Crossref PubMed Scopus (195) Google Scholar, 8Sato S. Shimoda Y. Muraki A. Kohara M. Nakamura Y. Tabata S. A large-scale protein-protein interaction analysis in Synechocystis sp. PCC6803.DNA Res. 2007; 14: 207-216Crossref PubMed Scopus (145) Google Scholar, 9Titz B. Rajagopala S.V. Goll J. Häuser R. McKevitt M.T. Palzkill T. Uetz P. The binary protein interactome of Treponema pallidum–the syphilis spirochete.PLoS ONE. 2008; 3: e2292Crossref PubMed Scopus (83) Google Scholar, 10Rolland T. Tașan M. Charloteaux B. Pevzner S.J. Zhong Q. Sahni N. Yi S. Lemmens I. Fontanillo C. Mosca R. Kamburov A. Ghiassian S.D. Yang X. Ghamsari L. Balcha D. et al.A proteome-scale map of the human interactome network.Cell. 2014; 159: 1212-1226Abstract Full Text Full Text PDF PubMed Scopus (890) Google Scholar, 11Murali T. Pacifico S. Yu J. Guest S. Roberts 3rd, G.G. Finley Jr., R.L. DroID 2011: a comprehensive, integrated resource for protein, transcription factor, RNA and gene interactions for Drosophila.Nucleic Acids Res. 2011; 39: D736-D743Crossref PubMed Scopus (141) Google Scholar, 12Simonis N. Rual J.F. Carvunis A.R. Tasan M. Lemmens I. Hirozane-Kishikawa T. Hao T. Sahalie J.M. Venkatesan K. Gebreab F. Cevik S. Klitgord N. Fan C. Braun P. Li N. et al.Empirically controlled mapping of the Caenorhabditis elegans protein-protein interactome network.Nat. Methods. 2009; 6: 47-54Crossref PubMed Scopus (216) Google Scholar, 13Butland G. Peregrín-Alvarez J.M. Li J. Yang W. Yang X. Canadien V. Starostine A. Richards D. Beattie B. Krogan N. Davey M. Parkinson J. Greenblatt J. Emili A. Interaction network containing conserved and essential protein complexes in Escherichia coli.Nature. 2005; 433: 531-537Crossref PubMed Scopus (937) Google Scholar, 14Gavin A.C. Aloy P. Grandi P. Krause R. Boesche M. Marzioch M. Rau C. Jensen L.J. Bastuck S. Dümpelfeld B. Edelmann A. Heurtier M.A. Hoffman V. Hoefert C. Klein K. et al.Proteome survey reveals modularity of the yeast cell machinery.Nature. 2006; 440: 631-636Crossref PubMed Scopus (2116) Google Scholar, 15Krogan N.J. Cagney G. Yu H. Zhong G. Guo X. Ignatchenko A. Li J. Pu S. Datta N. Tikuisis A.P. Punna T. Peregrín-Alvarez J.M. Shales M. Zhang X. Davey M. et al.Global landscape of protein complexes in the yeast Saccharomyces cerevisiae.Nature. 2006; 440: 637-643Crossref PubMed Scopus (2333) Google Scholar, 16Kühner S. van Noort V. Betts M.J. Leo-Macias A. Batisse C. Rode M. Yamada T. Maier T. Bader S. Beltran-Alvarez P. Castaño-Diez D. Chen W.H. Devos D. Güell M. Norambuena T. et al.Proteome organization in a genome-reduced bacterium.Science. 2009; 326: 1235-1240Crossref PubMed Scopus (377) Google Scholar, 17Hu P. Janga S.C. Babu M. Díaz-Mejía J.J. Butland G. Yang W. Pogoutse O. Guo X. Phanse S. Wong P. Chandran S. Christopoulos C. Nazarians-Armavil A. Nasseri N.K. Musso G. et al.Global functional atlas of Escherichia coli encompassing previously uncharacterized proteins.PLos Biol. 2009; 7: e96Crossref PubMed Scopus (262) Google Scholar, 18Arifuzzaman M. Maeda M. Itoh A. Nishikata K. Takita C. Saito R. Ara T. Nakahigashi K. Huang H.C. Hirai A. Tsuzuki K. Nakamura S. Altaf-Ul-Amin M. Oshima T. Baba T. et al.Large-scale identification of protein-protein interaction of Escherichia coli K-12.Genome Res. 2006; 16: 686-691Crossref PubMed Scopus (335) Google Scholar). Y2H screens detect binary physical interactions between pairs of proteins expressed in a non-native host, whereas AP-MS detects proteins that co-purify with a tagged protein expressed in the native organism. The Y2H method can detect transient, lower affinity interactions of as little as micromolar affinity (19Estojak J. Brent R. Golemis E.A. Correlation of two-hybrid affinity data with in vitro measurements.Mol. Cell. Biol. 1995; 15: 5820-5829Crossref PubMed Scopus (447) Google Scholar), whereas AP-MS is better suited to identify stable protein complexes (5Yu H. Braun P. Yildirim M.A. Lemmens I. Venkatesan K. Sahalie J. Hirozane-Kishikawa T. Gebreab F. Li N. Simonis N. Hao T. Rual J.F. Dricot A. Vazquez A. Murray R.R. et al.High quality binary protein interaction map of the yeast interactome network.Science. 2008; 322: 104-110Crossref PubMed Scopus (1109) Google Scholar). The resulting networks generally comprise thousands of pairwise interactions between proteins in which hub proteins are highly connected to functionally diverse arrays of other proteins (1Kristensen A.R. Foster L.J. High throughput strategies for probing the different organizational levels of protein interaction networks.Mol. Biosyst. 2013; 9: 2201-2212Crossref PubMed Scopus (10) Google Scholar, 20Seebacher J. Gavin A.C. SnapShot: protein-protein interaction networks.Cell. 2011; 144 (1000, 1000 e1001 doi:10.1016/j.cell.2011.02.025)Abstract Full Text PDF PubMed Scopus (52) Google Scholar), with the total interactome being estimated to contain ∼10,000 protein pairs in Escherichia coli (4Rajagopala S.V. Sikorski P. Kumar A. Mosca R. Vlasblom J. Arnold R. Franca-Koh J. Pakala S.B. Phanse S. Ceol A. Häuser R. Siszler G. Wuchty S. Emili A. Babu M. et al.The binary protein-protein interaction landscape of Escherichia coli.Nat. Biotechnol. 2014; 32: 285-290Crossref PubMed Scopus (148) Google Scholar) to ∼130,000 in humans (21Venkatesan K. Rual J.F. Vazquez A. Stelzl U. Lemmens I. Hirozane-Kishikawa T. Hao T. Zenkner M. Xin X. Goh K.I. Yildirim M.A. Simonis N. Heinzmann K. Gebreab F. Sahalie J.M. et al.An empirical framework for binary interactome mapping.Nat. Methods. 2009; 6: 83-90Crossref PubMed Scopus (651) Google Scholar). yeast two-hybrid affinity purification-mass spectrometry Basic Local Alignment Search Tool cytoplasmic protein false discovery rate normalized spectral abundance factor outer membrane protein periplasmic protein protein-protein interaction The Institute of Genome Research false negative. As part of a large interdisciplinary project (enigma.lbl.gov), we are conducting detailed system-wide analyses of the model sulfate-reducing bacterium Desulfovibrio vulgaris, a Deltaproteobacteria and obligate anaerobe (22Zhou J. He Q. Hemme C.L. Mukhopadhyay A. Hillesland K. Zhou A. He Z. Van Nostrand J.D. Hazen T.C. Stahl D.A. Wall J.D. Arkin A.P. How sulphate-reducing microorganisms cope with stress: lessons from systems biology.Nat. Rev. Microbiol. 2011; 9: 452-466Crossref PubMed Scopus (132) Google Scholar). D. vulgaris has been extensively characterized by functional genomic studies of its response to environmentally relevant conditions (23He Q. He Z. Joyner D.C. Joachimiak M. Price M.N. Yang Z.K. Yen H.C. Hemme C.L. Chen W. Fields M.M. Stahl D.A. Keasling J.D. Keller M. Arkin A.P. Hazen T.C. et al.Impact of elevated nitrate on sulfate-reducing bacteria: a comparative study of Desulfovibrio vulgaris.ISME J. 2010; 4: 1386-1397Crossref PubMed Scopus (52) Google Scholar, 24Mukhopadhyay A. Redding A.M. Joachimiak M.P. Arkin A.P. Borglin S.E. Dehal P.S. Chakraborty R. Geller J.T. Hazen T.C. He Q. Joyner D.C. Martin V.J. Wall J.D. Yang Z.K. Zhou J. Keasling J.D. Cell-wide responses to low-oxygen exposure in Desulfovibrio vulgaris Hildenborough.J. Bacteriol. 2007; 189: 5996-6010Crossref PubMed Scopus (85) Google Scholar), but relatively few protein complexes have been analyzed to date (25Han B.G. Dong M. Liu H. Camp L. Geller J. Singer M. Hazen T.C. Choi M. Witkowska H.E. Ball D.A. Typke D. Downing K.H. Shatsky M. Brenner S.E. Chandonia J.M. et al.Survey of large protein complexes in D. vulgaris reveals great structural diversity.Proc. Natl. Acad. Sci. U.S.A. 2009; 106: 16580-16585Crossref PubMed Scopus (25) Google Scholar, 26Oliveira T.F. Vonrhein C. Matias P.M. Venceslau S.S. Pereira I.A. Archer M. The crystal structure of Desulfovibrio vulgaris dissimilatory sulfite reductase bound to DsrC provides novel insights into the mechanism of sulfate respiration.J. Biol. Chem. 2008; 283: 34141-34149Abstract Full Text Full Text PDF PubMed Scopus (127) Google Scholar, 27Walian P.J. Allen S. Shatsky M. Zeng L. Szakal E.D. Liu H. Hall S.C. Fisher S.J. Lam B.R. Singer M.E. Geller J.T. Brenner S.E. Chandonia J.M. Hazen T.C. Witkowska H.E. et al.High throughput isolation and characterization of untagged membrane protein complexes: outer membrane complexes of Desulfovibrio vulgaris.J. Proteome Res. 2012; 11: 5720-5735Crossref PubMed Scopus (16) Google Scholar). Therefore, we have performed a global AP-MS screen to characterize its interactome and have also critically reexamined nine published AP-MS and Y2H screens. We have developed a rigorous data analysis strategy that has identified 459 high confidence protein-protein interactions (PPIs) for D. vulgaris and 391 PPIs from an existing AP-MS dataset for E. coli, many of which are supported by low throughput data from the literature. Importantly, compared with the protein-protein networks proposed previously, our two interactomes are smaller, less interconnected. and more strongly enriched in protein partners that share similar function or whose interactions have been validated in independent high throughput assays. We also show that the ∼3% of PPIs from the earlier Y2H and AP-MS screens that were reciprocally confirmed as both bait-prey and prey-bait pairs, and thus are more confidently detected, share very similar characteristics with our two high confidence interactomes. The remaining ∼97% of protein pairs from the earlier screens, in contrast, do not. Our work provides more stringent criteria for assessing the quality of protein interactomes and suggests that the number of bona fide interactions from the earlier screens that are supportable by independent evidence is limited to hundreds and not the thousands claimed. D. vulgaris Hildenborough wild-type ATCC29579 was genetically engineered to encode locus-specific affinity purification (AP)-tagged fusion proteins using electroporation of non-replicating "suicide constructs" (28Chhabra S.R. Butland G. Elias D.A. Chandonia J.M. Fok O.Y. Juba T.R. Gorur A. Allen S. Leung C.M. Keller K.L. Reveco S. Zane G.M. Semkiw E. Prathapam R. Gold B. et al.Generalized schemes for high throughput manipulation of the Desulfovibrio vulgaris genome.Appl. Environ. Microbiol. 2011; 77: 7595-7604Crossref PubMed Scopus (10) Google Scholar). Of the 3525 predicted D. vulgaris protein-coding genes, we attempted to create tagged strains for a priority list of 2086 genes. These genes were selected based on several criteria, including detection of the proteins they encode in fractionated cell-free extracts by MS, 2M. Dong, H. Liu, J. JIn, H.E. Witkowska and M.D. Biggin., unpublished observations. expected complexes based on E. coli interologs, and functional interest, such as energy generation. We constructed plasmids for generating chromosomal AP-tagged alleles for 1963 of the priority genes, 1681 of which were successfully integrated into the D. vulgaris chromosome. From this set, 1498 strains expressing an AP-tagged fusion protein were verified by Western blot, of which 1415 were constructed using Sequence and Ligation Independent Cloning, 77 using Gatewayn and 6 using recombineering procedures, supplemental Dataset S1. The primary AP tag utilized was Strep-TEV-FLAG (1231 strains); however, Strep-TEV-FLAG-His6 (237 strains) and Sequential Peptide Affinity tag (30 strains) (29Zeghouf M. Li J. Butland G. Borkowska A. Canadien V. Richards D. Beattie B. Emili A. Greenblatt J.F. Sequential peptide affinity (SPA) system for the identification of mammalian and bacterial protein complexes.J. Proteome Res. 2004; 3: 463-468Crossref PubMed Scopus (152) Google Scholar) were also used. A non-redundant total of 1401 unique genes are represented as AP-tagged alleles in the 1498 strains constructed. All affinity purifications were performed as described previously (28Chhabra S.R. Butland G. Elias D.A. Chandonia J.M. Fok O.Y. Juba T.R. Gorur A. Allen S. Leung C.M. Keller K.L. Reveco S. Zane G.M. Semkiw E. Prathapam R. Gold B. et al.Generalized schemes for high throughput manipulation of the Desulfovibrio vulgaris genome.Appl. Environ. Microbiol. 2011; 77: 7595-7604Crossref PubMed Scopus (10) Google Scholar). In all cases, Strep-TEV-FLAG-His6 strains were treated exactly as Strep-TEV-FLAG strains for the purposes of affinity purification of protein complexes. The majority of AP samples were analyzed by parallel gel-free and gel-based workflows. In a gel-free approach, AP-isolated proteins were digested with trypsin utilizing a 96-well PVDF membrane-based protocol and analyzed by LC MS/MS using either LTQ XL linear ion trap mass spectrometer (Thermo Scientific, Fremont, CA) or LTQ Velos Orbitrap mass spectrometer (Thermo Scientific), essentially as described by Chhabra et al. and Roan et al., respectively (28Chhabra S.R. Butland G. Elias D.A. Chandonia J.M. Fok O.Y. Juba T.R. Gorur A. Allen S. Leung C.M. Keller K.L. Reveco S. Zane G.M. Semkiw E. Prathapam R. Gold B. et al.Generalized schemes for high throughput manipulation of the Desulfovibrio vulgaris genome.Appl. Environ. Microbiol. 2011; 77: 7595-7604Crossref PubMed Scopus (10) Google Scholar, 30Roan N.R. Chu S. Liu H. Neidleman J. Witkowska H.E. Greene W.C. Interaction of fibronectin with semen amyloids synergistically enhances HIV infection.J. Infect. Dis. 2014; 210: 1062-1066Crossref PubMed Scopus (9) Google Scholar). Five sample sets, however, used a QSTAR XL mass spectrometer (AB Sciex, Framingham, MA) as described by Chiu et al. (31Chiu Y.L. Witkowska H.E. Hall S.C. Santiago M. Soros V.B. Esnault C. Heidmann T. Greene W.C. High molecular-mass APOBEC3G complexes restrict Alu retrotransposition.Proc. Natl. Acad. Sci. U.S.A. 2006; 103: 15588-15593Crossref PubMed Scopus (211) Google Scholar). In all cases, an additional wash-run (5-μl injection of 50% isopropyl alcohol to clean the trap cartridge; 30-min gradient over analytical column, including two 5-min ramps from 3% acetonitrile to 97% acetonitrile) and a protein standards run consisting of bovine 6-protein mix (Michrom Bioresources, Auburn, CA) (10-fmol injection; 15-min gradient 3% acetonitrile to 40% acetonitrile) were incorporated between AP samples to minimize carry-over of D. vulgaris proteins. The final 6-protein mix standard-run was used to assess carry-over between samples (i.e. represents the "background-run" described below). In the gel-based workflow, proteins were fractionated by SDS-PAGE (12%) and bands visualized by silver staining (13Butland G. Peregrín-Alvarez J.M. Li J. Yang W. Yang X. Canadien V. Starostine A. Richards D. Beattie B. Krogan N. Davey M. Parkinson J. Greenblatt J. Emili A. Interaction network containing conserved and essential protein complexes in Escherichia coli.Nature. 2005; 433: 531-537Crossref PubMed Scopus (937) Google Scholar). Selected bands were excised from the gel; proteins were in-gel digested with trypsin using a ProGest robot (Genomics Solutions, Ann Arbor, MI), and proteolytic peptides were analyzed by LC MS/MS using LTQ XL linear ion trap mass spectrometer (Thermo Scientific), as described in Walian et al. (27Walian P.J. Allen S. Shatsky M. Zeng L. Szakal E.D. Liu H. Hall S.C. Fisher S.J. Lam B.R. Singer M.E. Geller J.T. Brenner S.E. Chandonia J.M. Hazen T.C. Witkowska H.E. et al.High throughput isolation and characterization of untagged membrane protein complexes: outer membrane complexes of Desulfovibrio vulgaris.J. Proteome Res. 2012; 11: 5720-5735Crossref PubMed Scopus (16) Google Scholar). For the in-gel workflow, additional wash and standard runs were introduced between samples that resulted from processing of gel slices from a single lane of the SDS-polyacrylamide gel, i.e. between different AP purifications. Wash runs consisted of a blank injection followed by a 30-min gradient containing two 5-min ramps from 2% acetonitrile to 97% acetonitrile. Protein standard runs consisted of a 25-fmol injection of bovine 6-protein mix via a 10-ml metered injection followed by a 14-min gradient from 2 to 50% acetonitrile. The 6-protein mix standard runs were used to assess carryover between samples (i.e. represents the background-run described below). Peak lists were extracted from .raw files using the Mascot Distiller 2.3.2.0 software (Matrix Science, London, United Kingdom). Data were searched with an in-house Mascot version 2.2.04 search engine (Matrix Science) against a custom protein database containing all potential protein products generated via 6-frame translation of the D. vulgaris genome supplemented by frequently observed contaminants and concatenated with the decoy database generated by reversing all D. vulgaris protein sequences (102,572 sequences; 9,848,210 residues) (32Elias J.E. Gygi S.P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry.Nat. Methods. 2007; 4: 207-214Crossref PubMed Scopus (2827) Google Scholar). Search was limited to doubly and triply charged precursors. The following search parameters were utilized for most searches: precursor mass tolerance of 0.8 Da for the LTQ XL-generated and 3 ppm for the LTQ Velos Orbitrap-generated data, respectively; fragment mass tolerance of 0.8 Da for both instruments; tryptic digestion allowing for cleavages before Pro; 1 missed cleavage; fixed modification, Cys-carbamidomethyl; variable modifications, deamidation (Asn and Gln), Met-sulfoxide, and Pyro-Glu (N-terminal Gln). A limited number of searches were performed with a precursor mass tolerance of 1.5 Da and 50 ppm for LTQ XL-generated and LTQ Velos-generated data, respectively. Precursor and product ion mass tolerances for analysis of the QSTAR-generated data were 100 ppm and 0.15 Da, respectively. Significance threshold was set to a p value of ≤0.05. Protein acceptance required the presence of at least one distinct peptide with expectation value of ≤0.05. >90.5% of peptide identifications met the "bold red" Mascot match quality criteria, i.e. (i) peptide assignment to a protein with the highest score (rank) within the potential homologs with overlapping sequences, and (ii) a top scoring match for the spectrum. <0.5% of peptide identifications had rank two scores. 97% of these, however, were the only identification supporting a protein and as such were filtered out at a later step, as described below. The remaining 25 cases supported a protein identified by at least one peptide that met the bold red criteria and were thus retained. All peptide matches with expectation value of ≤0.05 were used for spectral counting (33Liu H. Sadygov R.G. Yates 3rd., J.R. A model for random sampling and estimation of relative protein abundance in shotgun proteomics.Anal. Chem. 2004; 76: 4193-4201Crossref PubMed Scopus (2066) Google Scholar, 34Lundgren D.H. Hwang S.I. Wu L. Han D.K. Role of spectral counting in quantitative proteomics.Expert Rev. Proteomics. 2010; 7: 39-53Crossref PubMed Scopus (313) Google Scholar). High abundance or "sticky" proteins were observed in some cases in subsequent unrelated protein samples even after extensive washing of the LC column between samples (background-run). Proteins identified based on the presence of these peptides were designated "carry-over" and removed from subsequent analysis if the Mascot score for the protein in the sample in question was lower than its Mascot score from the immediately preceding background-run. This automatic procedure was augmented in 21 cases by manual removal of a single protein that appeared to be a contaminant from other samples processed the same day. In addition, peptides that cannot be unambiguously mapped to a single D. vulgaris protein were removed from the analysis and not used to assign the proteins' identification. Some ambiguous peptides were retained for spectral counting, however, but only if the identified protein was also supported by at least one unambiguous peptide. Peptide level mass spectrometry data for the resulting partially filtered dataset are provided in supplemental Dataset S2. For our final high confidence interactome, we additionally filtered out low signal proteins and overly abundant proteins by removing prey proteins identified with a single-peptide hit from the results of a given purification: ribosomal proteins and protein chaperones (DnaK, DVU0811; GrpE, DVU0812; GroEL, DVU1976; and GroES, DVU1977); and the following top nine frequent fliers: PpaC (DVU1636); Mrp (DVU2109); GroEL; DVU2405; ApsA (DVU0847); Sat (DVU1295); Pyc (DVU1834); DnaK; and Tuf (DVU2920) (supplemental Dataset S3). The 31 instances of a bait being detected by a single peptide were retained at this stage, however (see supplemental Dataset S2). Skyline software (35MacLean B. Tomazela D.M. Shulman N. Chambers M. Finney G.L. Frewen B. Kern R. Tabb D.L. Liebler D.C. MacCoss M.J. Skyline: an open source document editor for creating and analyzing targeted proteomics experiments.Bioinformatics. 2010; 26: 966-968Crossref PubMed Scopus (2963) Google Scholar) was used to generate a spectral library for these baits, and the spectra have been deposited at Panoromaweb. After this series of filtering steps, 53,506 protein pairs remained for a Matrix model (supplemental Dataset S4) and 5177 for a Spoke model (20Seebacher J. Gavin A.C. SnapShot: protein-protein interaction networks.Cell. 2011; 144 (1000, 1000 e1001 doi:10.1016/j.cell.2011.02.025)Abstract Full Text PDF PubMed Scopus (52) Google Scholar). To compare protein-protein networks from various species as well as to project the EcoCyc reference set onto species other than E. coli, we mapped homologs between all nine studied species using bi-directional best Basic Local Alignment Search Tool (BLAST) searches (36Altschul S.F. Gish W. Miller W. Myers E.W. Lipman D.J. Basic local alignment search tool.J. Mol. Biol. 1990; 215: 403-410Crossref PubMed Scopus (70338) Google Scholar). All predicted protein sequences encoded by one genome were queried against a database of protein sequences encoded by another genome using BLASTP 2.2.9 with default options, and then the search direction was switched. Pairs in which each protein was the most significant hit for a query from the other genome and for which both E-values were at least as significant as 10−4 were mapped to each other. The supplemental Table S1 lists the number of mapped homologs between all nine species. A pair of PPIs (a and b) and (a′ and b′) from two different species is called an interolog if a is th emapped homolog of a′ and b is the mapped homolog of b′. Computational analysis was performed using curated gold standard sets of interacting and non-interacting pairs of proteins (supplemental Dataset S5). Because of the lack of truly known interacting and non-interacting proteins in D. vulgaris, our gold standard sets should be considered as imperfect gold standards. 38 of the gold standard positive set are pairs of D. vulgaris proteins that have been shown to interact in stable complexes in low throughput experiments in this organism. The remaining 500 gold-positives were E. coli interologs, i.e. D. vulgaris proteins mapped to homologous E. coli proteins (as described above), of either PPIs from EcoCyc version 12.0 (supplemental Dataset S6) (37Karp P.D. Riley M. Saier M. Paulsen I.T. Collado-Vides J. Paley S.M. Pellegrini-Toole A. Bonavides C. Gama-Castro S. The EcoCyc Database.Nucleic Acids Res. 2002; 30: 56-58Crossref PubMed Scopus (340) Google Scholar) or reciprocally confirmed PPIs from recent AP-MS experiments in E. coli (17Hu P. Janga S.C. Babu M. Díaz-Mejía J.J. Butland G. Yang W. Pogoutse O. Guo X. Phanse S. Wong P. Chandran S. Christopoulos C. Nazarians-Armavil A. Nasseri N.K. Musso G. et al.Global functional atlas of Escherichia coli encompassing previously uncharacterized proteins.PLos Biol. 2009; 7: e96Crossref PubMed Scopus (262) Google Scholar). This dataset was curated to account for expected differences between E. coli and D. vulgaris complexes (e.g. a classical RNA degradosome complex configuration was not expected to be found in D. vulgaris due to the truncation of a scaffold protein (38Chhabra S.R. Joachimiak M.P. Petzold C.J. Zane G.M. Price M.N. Reveco S.A. Fok V. Johanson A.R. Batth T.S. Singer M. Chandonia J.M. Joyner D. Hazen T.C. Arkin A.P. Wall J.D. et al.Towards a rigorous network of protein-protein interactions of the model sulfate reducer Desulfovibrio vulgaris Hildenborough.PLoS One. 2011; 6: e21470Crossref PubMed Scopus (10) Google Scholar)). We also excluded all interactions with ribosomal proteins, as this complex is atypical due to the RNA component as well as highly abundant. T
Referência(s)