Global Survey of Human T Leukemic Cells by Integrating Proteomics and Transcriptomics Profiling

• DOI: 10.1074/mcp.m700017-mcp200
• Publication Date: 2007-05-23T17:30:35Z
• Authors: Linfeng Wu; Sun-Il Hwang; Karim Rezaul; Long J. Lu; Viveka Mayya; Mark Gerstein; Jimmy K. Eng; Deborah H. Lundgren; David K. Han
• Keywords: Profiling (computer programming)

A global protein survey is needed to gain systems-level insights into mammalian cell signaling and information flow. Human Jurkat T leukemic cells are one of the most important model systems for study, but no comprehensive proteomics has been carried out in this type. In present study we combined subcellular fractionation, multiple enrichment methods, replicate tandem mass spectrometry analyses determine expression pattern a single The proteome dataset was evaluated by comparison with genome-wide mRNA same total 5381 proteins were identified high confidence. Rigorous RNA afforded removal false positive identifications redundant entries rescued scoring peptide, resulting final identification 6471 unique gene products among which 98% corresponding transcripts detected probability. Using hierarchical clustering patterns five fractions (cytosol, light membrane, heavy mitochondria, nuclei), primary localization 2241 assigned confidence including 792 previously uncharacterized proteins. This landscape can serve as useful platform understanding organelle composition cellular functions human cells. An goal functional genomics globally profile biological systems. Many studies have utilized cDNA or oligonucleotide microarrays measure level, deducing (1Pan Q. Shai O. Misquitta C. Zhang W. Saltzman A.L. Mohammad N. Babak T. Siu H. Hughes T.R. Morris Q.D. Frey B.J. Blencowe Revealing regulatory features alternative splicing using quantitative microarray platform.Mol. Cell. 2004; 16: 929-941Abstract Full Text PDF PubMed Scopus (258) Google Scholar, 2Su A.I. Wiltshire Batalov S. Lapp Ching K.A. Block D. J. Soden R. Hayakawa M. Kreiman G. Cooke M.P. Walker J.R. Hogenesch J.B. atlas mouse protein-encoding transcriptomes.Proc. Natl. Acad. Sci. U. A. 101: 6062-6067Crossref (2851) Scholar). However, despite obvious dependence synthesis on mRNA, many reported that more than half non-coding (3Okazaki Y. Furuno Kasukawa Adachi Bono Kondo Nikaido I. Osato Saito Suzuki et al.Analysis transcriptome based annotation 60,770 full-length cDNAs.Nature. 2002; 420: 563-573Crossref (1394) 4Carninci P. Katayama Gough Frith M.C. Maeda Oyama Ravasi Lenhard B. Wells al.The transcriptional genome.Science. 2005; 309: 1559-1563Crossref (2804) 5Cheng Kapranov Drenkow Dike Brubaker Patel Long Stern Tammana Helt Sementchenko V. Piccolboni Bekiranov Bailey D.K. Ganesh Ghosh Bell Gerhard D.S. Gingeras Transcriptional maps 10 chromosomes at 5-nucleotide resolution.Science. 308: 1149-1154Crossref (975) addition, measurements show only moderate even poor correlation between level due different translational efficiency post-translational turnover (6Gygi S.P. Rochon Franza B.R. Aebersold Correlation abundance yeast.Mol. Biol. 1999; 19: 1720-1730Crossref (3187) 7Chen Gharib T.G. Huang C.C. Taylor J.M. Misek D.E. Kardia S.L. Giordano T.J. Iannettoni M.D. Orringer M.B. Hanash S.M. Beer D.G. Discordant lung adenocarcinomas.Mol. Proteomics. 1: 304-313Abstract (784) 8Kislinger Cox Kannan Chung Hu Ignatchenko Scott M.S. Gramolini A.O. Hallett M.T. Rossant Emili Global organ mouse: proteomic transcriptomic profiling.Cell. 2006; 125: 173-186Abstract (399) Furthermore cannot be accurately predicted expression. Therefore, ultimately need explained completed draft sequences genome (9Lander E.S. Linton L.M. Birren Nusbaum Zody Baldwin Devon K. Dewar Doyle FitzHugh al.Initial sequencing analysis genome.Nature. 2001; 409: 860-921Crossref (17795) 10Venter J.C. Adams Myers E.W. Li P.W. Mural R.J. Sutton G.G. Smith H.O. Yandell Evans C.A. Holt R.A. sequence 291: 1304-1351Crossref (10623) Scholar) several other organisms (11Kunst F. Ogasawara Moszer Albertini A.M. Alloni Azevedo Bertero M.G. Bessieres Bolotin Borchert complete gram-positive bacterium Bacillus subtilis.Nature. 1997; 390: 249-256Crossref (3126) 12Adams Celniker S.E. Gocayne J.D. Amanatides P.G. Scherer Hoskins Galle R.F. Drosophila melanogaster.Science. 2000; 287: 2185-2195Crossref (4824) made large scale feasible (13Peng Schwartz Elias J.E. Thoreen Cheng Marsischky Roelofs Finley Gygi approach ubiquitination.Nat. Biotechnol. 2003; 21: 921-926Crossref (1310) 14Ptacek Devgan Michaud Zhu X. Fasolo Guo Jona Breitkreutz Sopko McCartney R.R. Schmidt Rachidi Lee S.J. Mah A.S. Meng L. Stark M.J. D.F. De Virgilio Tyers Andrews Gerstein Schweitzer Predki P.F. Snyder phosphorylation yeast.Nature. 438: 679-684Crossref (819) huge diversity dynamic range expressed proteins, especially cells, all remained greatest challenges (15Wu Han Overcoming problem spectrometry-based shotgun proteomics.Expert Rev. 3: 611-619Crossref (76) Although proteome-scale tissues, organelles reported, type date. performed line combining transcriptomics profiling. popular studying signal transduction because key advances field receptor (16Abraham R.T. Weiss development T-cell signalling paradigm.Nat. Immunol. 4: 301-308Crossref (386) Moreover also used phenomena such apoptosis engulfment (17Arur Uche U.E. Rezaul Fong Scranton Cowan A.E. Mohler Annexin I an endogenous ligand mediates apoptotic engulfment.Dev. 587-598Abstract (364) 18Wang Song J.H. Hao Role death mitochondrial pathways conventional chemotherapy drug induction apoptosis.Cell. Signal. 18: 1528-1535Crossref (55) in-depth characterizations function transduction. contrast recent (8Kislinger 19Foster L.J. de Hoog C.L. Xie Mootha V.K. Mann map 187-199Abstract (470) Scholar), our type, making it suitable network flow within A3 from American Type Culture Collection (Manassas, VA) study. grown maximal density 0.5–0.8 × 106 cells/ml then collected centrifugation 400 g min 4 °C. pellets washed twice ice-cold PBS. To obtain whole lysates, resuspended lysis buffer (50 mm HEPES, pH 7.5, 100 NaCl, 1 EDTA, 1% Tween 20, mixture protease inhibitors (Roche Diagnosis GmbH)) ice 30 centrifuged 12,000 20 supernatant (designated lysate) stored −80 °C later analysis. Subcelluar fractionation described below. One volume pellet incubated 5 volumes hypotonic (buffer A: KCl, 1.5 MgCl2, EGTA, DTT, inhibitors) 2 mixed equal supplemented 0.5 m sucrose 0.25 isolation buffer. After incubation ice, homogenized glass Dounce homogenizer until ∼50% became trypan blue-positive. homogenates 650 postnuclear supernatants fraction preparation. further ∼95% nuclear isolated rinsed once. gently 2.5 ml 1.28 sucrose, layered over 2.3 60,000 90 purified A, min, subsequent experiments. isolate cytosol, mitochondria (20Rezaul Wu Mayya Hwang S.I. systematic characterization leukemia cells.Mol. 169-181Abstract (100) plasma membrane fraction, (∼2 109) 15 times. °C, removed. 100,000 h 10–15 times, 6.5 38% sucrose), 200,000 h. membranes phase diluted buffer, designated membrane. lipid raft PBS-washed (25 MES, 6.5, 150 0.1% Triton X-100, sodium vanadate, inhibitors). Cells 20–30 times 80% without X-100. lysates placed bottom 14 89-mm clear centrifuge tube (Beckman), overlaid 30% 3.5 5% 18 (the machine set "no brake" condition). low raft. details culture approaches shown supplemental materials. Proteins digested trypsin analyzed minor modifications outlined Digested linear ion trap spectrometer (Finnigan LTQ, Thermo Finnigan, San Jose, CA). Samples loaded onto 10-cm 100-μm capillary C18 reversed-phase column microautosampler (Famos, Dionex, Sunnyvale, CA) followed LC-MS/MS LTQ. For stable isotope-free peptide samples, each full MS scan MS/MS scans intense peaks spectrum exclusion enabled. m/z either 300–1700 400–1700 range. isotope-labeled mainly come phosphoprotein enrichment, peak All raw files converted .dat Xcalibur software (version 1.4 SR1) mzXML conversion dat2xml (Institute Systems Biology). Peak lists generated automatically smoothing deisotoping, charge states (21Eng McCormack Yates III, correlate spectral data peptides amino acid database.J. Am. Soc. Mass Spectrom. 1994; 5: 976-989Crossref (5439) minimum count 1000. searched against local copy non-redundant database (56,709 entries, November 30, 2004 release version) NCI, National Institutes Health, Advanced Biomedical Computing Center SEQUEST algorithm (SEQUEST-PVM version 27 (revision 0)) parameters follows: filtering thresholds off; tolerance 1.0 Da precursor ions fragment ions; tryptic constraint allowing missed cleavage; oxidization (+16 Da) methionine. If contained acids, modification allowed. detailed description labeled acids experiment Supplemental Table S1. search results processed INTERACT program (22Han Eng Zhou Quantitative profiling differentiation-induced microsomal isotope-coded affinity tags spectrometry.Nat. 946-951Crossref (828) filtered following criteria: Xcorr cutoff values 1.9, 2.2, 3.7 1+, 2+, 3+ peptides, respectively; ΔCn value 0.1; partially excluded. least two distinct experimental estimate rate, datasets forward reversed concatenated (23Peng Licklider Evaluation multidimensional chromatography coupled (LC/LC-MS/MS) large-scale analysis: yeast proteome.J. Proteome Res. 2: 43-50Crossref (1383) address redundancy issue list above compute probability PeptideProphet 1.0) (24Keller Nesvizhskii Kolker E. Empirical statistical accuracy search.Anal. Chem. 74: 5383-5392Crossref (3897) ProteinProphet 2.0) softwares (25Nesvizhskii Keller identifying spectrometry.Anal. 75: 4646-4658Crossref (3633) combine group indicate whether group. searching parameter 0.0 include transcript compared dataset. Sentrix Human-6 Expression BeadChip (Illumina, Diego, contains 50-mer gene-specific probes >46,000 variants There average 30× per array. Total log guanidine thiocyanate method. solutions materials RNase-free if necessary. (1 108) PBS, lysed GTC (4 thiocyanate, acetate, 2-mercapotaethanol, 7.0), 20-gauge syringe needle lysate top 3 CsCl solution (5.7 CsCl, acetate). Then sample 27,000 rpm SW 41 rotor. removed carefully, dissolved 200 μl H2O. phenol:chloroform:isoamyl alcohol (25:24:1, v/v/v) precipitate acetate ethanol. lyophilized speed vacuum Poly(A)-enriched Absolute Purification kit (Stratagene). qualities poly(A)-enriched examined electrophoresis formaldehyde, agarose gel Northern blot hybridization (Supplemental Fig. S1). amplified TotalPrep amplification (Ambion, Foster City, hybridization, labeling, scanning chips according Illumina protocol. extracted, normalized, BeadStudio software. semiquantitation calculated normalizing counts relative fraction. normalized profiles hierarchically clustered uncentered centroid linkage Cluster 3.0 (26Eisen Spellman P.T. Brown P.O. Botstein display patterns.Proc. 1998; 95: 14863-14868Crossref (13227) visualized Java TreeView (27Saldanha A.J. Treeview—extensible visualization data.Bioinformatics. 20: 3246-3248Crossref (2343) begin identify separated 33 μg one-dimensional cut lane slices. slices trypsinized, extracted LTQ spectrometer. process repeated amount (33 μg) sample. Each 2200–2700 (Fig. 1A overlap replicates about 82% 1.8% standard deviation. Differences attributed complexity random sampling during acquisition LC-MS/MS. When cumulative number analyzed, found 3620 replicates. We observed coverage enhanced 40% when sequenced 1B). first reached 94% analyses. accumulation curve slope 0 after nine repeats, suggesting enough achieve saturation. saturate analyses, did not next attempted detect refractory aimed increase reducing well established validated methods 28Hwang Lundgren D.H. J.K. Systematic apoptosis: differential extraction isotope labeling.Mol. 1131-1145Abstract (57) First, subfractionated seven fractions, cytosolic, mitochondrial, nuclear, raft, fractions. Western assessments confirmed appropriate partitioning organellar markers across these providing basic confirmation purity S2). Second, enrich phosphorylated blotting S3). Third, enriched glycosylated lectin wheat germ agglutinin, preferentially binds N-acetylglucosamine (GlcNAc), terminal GlcNAc structures, sialic acid. S4). Fourth, tried masked subsets depletion. Because binding rates particular medium, hypothesized complex might shift binding, investigate hypothesis, CNBr-activated Sepharose beads, covalently bind free amine groups. Uncoupled time points. shifted depletion S5). Next repeatedly proteomes subfractions LC-MS/MS, 1707 runs performed. diagram strategy 2A. stringent filtering, 9611 Further selection apparent contaminants introduced handling (e.g. keratins) led ∼1000 ∼3600 2B, I, rate lower 0.7% entire (Table I). These indicated criteria filter spectra stringent, very few identifications.Table ISummary dataExperimental fractionProteinsUnique peptidesSpectraPeptide rate%Whole lysates3,62026,865256,3750.43Cytosol2,01213,61872,8930.64Heavy membrane1,5999,07334,6710.39Light membrane2,15413,61758,0010.42Mitochondrion1,1545,97121,0370.46Nucleus1,75010,33844,6730.37Raft1,1125,46220,6450.28Plasma membrane1,5298,23949,3810.28Glycoprotein9366,17523,6680.08Phosphoprotein1,0334,96421,6930.50Depletion3,03520,156195,1670.51Total5,38143,693798,204 Open table new tab essential closely related splice variants, homologs, paralogs, orthologs, database), leading overestimation number. issue, computed vast majority (80%) unambiguously, representative distinguished (peptide ≥0.9) unambiguous peptide. remaining could unambiguously mapped protein. cases, reliable claim truly simply bioinformatics prediction. Therefore straight method inadequate completely issues, indicating support data. independent Both prepared commercialized transcript-specific probe Duplicate triplicate arrays parallel respectively. Genes detection p less 0.05 regarded identifications. 15,592 15,