Quantitative Profile of Five Murine Core Proteomes Using Label-free Functional Proteomics

Quantitative Profile of Five Murine CoreProteomes Using Label-free FunctionalProteomics*□S

Pedro R. Cutillas‡§¶� and Bart Vanhaesebroeck‡¶**

Analysis of primary animal and human tissues is key inbiological and biomedical research. Comparative pro-teomics analysis of primary biological material wouldbenefit from uncomplicated experimental work flows ca-pable of evaluating an unlimited number of samples. Inthis report we describe the application of label-free pro-teomics to the quantitative analysis of five mouse coreproteomes. We developed a computer program and nor-malization procedures that allow exploitation of the quan-titative data inherent in LC-MS/MS experiments for rela-tive and absolute quantification of proteins in complexmixtures. Important features of this approach include (i)its ability to compare an unlimited number of samples, (ii)its applicability to primary tissues and cultured cells, (iii)its straightforward work flow without chemical reactionsteps, and (iv) its usefulness not only for relative quanti-fication but also for estimation of absolute protein abun-dance. We applied this approach to quantitatively charac-terize the most abundant proteins in murine brain, heart,kidney, liver, and lung. We matched 8,800 MS/MS peptidespectra to 1,500 proteins and generated 44,000 independ-ent data points to profile the �1,000 most abundant pro-teins in mouse tissues. This dataset provides a quantita-tive profile of the fundamental proteome of a mouse,identifies the major similarities and differences betweenorgan-specific proteomes, and serves as a paradigm ofhow label-free quantitative MS can be used to character-ize the phenotype of mammalian primary tissues at themolecular level. Molecular & Cellular Proteomics 6:1560–1573, 2007.

Experiments on immortalized cell lines have resulted in thegeneration of a vast amount of information on the biologicaland biochemical processes that govern the function of cul-tured cells. However, discerning the mechanisms by whichgenes control mammalian physiology in vivo may only beachieved by investigations that involve the use of animal

models of which the laboratory mouse (Mus musculus) offersmany advantages (1). There is a wealth of resources related tothe molecular biology of mouse cells that have benefited fromgenomics, transcriptomics, and, more recently, proteomicsprojects aimed at profiling the molecular composition of mu-rine tissues (2–4). The generation of gene-targeted mice isparticularly useful in advancing our understanding of howgenes control fundamental processes of mammalian physiol-ogy (e.g. Refs. 5–9).

Once created, finding the phenotype of a gene-targetedstrain of mice is not a trivial task (1, 10). Several years areneeded to fully characterize phenotypic alterations, and sub-tle phenotypes often go unnoticed. Robust and high through-put methods for profiling the proteomes of primary tissues ina quantitative fashion may expedite the search for phenotypicchanges in gene-targeted and other animal models. Insightsgained by powerful methods for the molecular characteriza-tion of primary tissues could also direct classical physiologicalexperiments, reduce the number of experimental animals, andmore comprehensively exploit the scientific knowledge thatcan be gained from animal models.

Thus, phenotypes could in principle be characterized sys-tematically by comparing the proteomes of primary cells usingunbiased proteomics approaches based on MS. Several an-alytical strategies exists that use MS for relative quantificationof proteins and proteomes. However, current methods forquantitative proteomics have shortcomings for the analysis ofprimary tissues. First, metabolic labeling, the ideal approachto compare proteomes, is difficult to apply to mammalianorganisms. Therefore, although powerful to quantify proteinsfrom immortalized cell lines (11), stable isotope labeling byamino acids in cell culture (SILAC) and other metabolic label-ing approaches for quantitative MS cannot be used to easilyquantify proteins from primary tissues. Second, an ideal ap-proach for quantitative proteomics should not rely on chem-ical derivatization strategies if the method is to be useful forthe comparison of an unlimited number of samples and toprovide statistically sound results. Thus although powerful inother contexts, strategies that rely on chemical labeling withcompounds enriched in isotopes (e.g. isobaric tags for relativeand absolute quantification (iTRAQ) and ICAT approaches (12,13)), heavy water (14), or fluorescence labels (e.g. the two-dimensional DIGE approach (15)) are not ideal for the analysisof gene-targeted mice (or for clinical studies). This shortcom-

From the ‡Cell Signalling Group, Ludwig Institute for Cancer Re-search, 91 Riding House Street, London W1W 7BS, United Kingdom,**Department of Biochemistry and Molecular Biology, University Col-lege London, Gower Street, London WC1E 6BT, United Kingdom, and§Proteomics Unit, Ludwig Institute for Cancer Research, CruciformBuilding, Gower Street, London WC1E 6BT, United Kingdom

Received, January 30, 2007, and in revised form, May 24, 2007Published, MCP Papers in Press, June 12, 2007, DOI 10.1074/

mcp.M700037-MCP200

Research

© 2007 by The American Society for Biochemistry and Molecular Biology, Inc.1560 Molecular & Cellular Proteomics 6.9This paper is available on line at http://www.mcponline.org

ing is also encountered when using metabolic labeling ap-proaches. Third, the method should provide sufficientthroughput, precision, and dynamic range, and the ideal tech-nique for proteome profiling should therefore consider thetrade-off between speed and depth of analysis. Approachesfor extensive characterization of proteomes, which rely oncomplex cellular fractionation, are useful for characterizingthe molecular architecture of cells (4, 16), but unfortunatelythese strategies do not offer the throughput required for com-parison of proteomes.

We and others have shown that the data in LC-MS andLC-MS/MS experiments have inherent quantitative informa-tion such that it is possible to use this type of data to assessprotein amounts in cells and tissues (17–19). Although label-free methods for quantitative proteomics based on spectralcounts have been described (4, 20, 21), their level of precisiononly permits the use of spectral counts as an approximateindication of protein abundance (17). In contrast, quantitativeMS methods based on the determination of peptide ion in-tensities may offer levels of precision close to those obtainedby isotopic labeling strategies (17, 19). The aim of the currentstudy was to evaluate the performance of a label-free quan-titative proteomics approach (which we designed taking intoaccount the considerations listed above) for the analysis ofprimary tissues. We report on the creation of a computerprogram and on the development of standardization proce-dures for downstream data processing that can be used toautomate the quantitative analysis of label-free LC-MS/MSdata such that this approach can be used for large scalecomparative analysis of any protein mixture, including thosein mammalian primary tissues. After assessing the perform-ance of the created protocols, we used these tools in aproof-of-principle experiment aimed at obtaining a low reso-lution map of the major proteins in the mouse. We found that,in addition to cost and simplicity, an advantage of label-freemethods is that they may be useful for providing relative andabsolute values of protein amounts simultaneously. Our re-sults describe five core proteomes of a mouse in quantitativeterms, provide new insights into the major similarities anddifferences between the protein compositions of the mainmurine organs, and serve as an example of how this approachmay be used to characterize mammalian organisms pheno-typically at the molecular level.

EXPERIMENTAL PROCEDURES

Tissue Extraction

Cell culture reagents were from Invitrogen. The WEHI-231 B cellline was cultured in RPMI 1640 medium supplemented with 10% fetalbovine serum, 1% penicillin/streptomycin, and 0.05 mM �-mercapto-ethanol at 37 °C in 5% CO2. Mouse tissues were obtained from amouse of the C57BL/6 genetic background that was killed by cervicaldislocation. Tissues were excised and processed without freezingsteps.

Cells were lysed in Triton X-100 lysis buffer (150 mM NaCl, 1% (w/v)Triton X-100, 1 mM EDTA, 50 mM Tris�HCl, pH 7.4) supplemented with

protein and phosphatase inhibitors. Fresh primary tissues were ho-mogenized in Triton X-100 lysis buffer using a micropestle (Eppen-dorf). Protein concentrations in cell lysates and organ homogenateswere determined using the Bradford assay.

Immunoblotting

Proteins for Western blotting were separated by 10% SDS-PAGEand transferred to PVDF membranes. Membranes were blocked with5% skimmed milk in TBS-T buffer (20 mM Tris�HCl, pH 8.0, 150 mM

NaCl, 0.1% Tween) and then probed with monoclonal antibodiesagainst actin or tubulin (both from Sigma) followed by incubationswith IRDye 800CW goat anti-mouse secondary antibody (LI-CORBiosciences, Cambridge, UK). Fluorescent immunoblot signals werequantified using an Odyssey imaging system (LI-COR Biosciences)and reported as -fold over the mean intensities of the samples to becompared.

Preparation of Samples for MS

Proteins were prepared for MS essentially as described previously(19). Briefly cell lysates or tissue homogenates were separated by10% SDS-PAGE and visualized by colloidal Coomassie Blue staining.Gels were scanned in a Bio-Rad G-800 densitometer. The OD of eachof the sections to be excised for downstream MS analysis wasdetermined and expressed as a percentage of the OD of the total lane(this step was important for the estimation of absolute proteinamounts; see below). Proteins in these gel sections were extracted byin-gel digestion using standard procedures with the exception that 2pmol of a standard protein (fetuin or lysozyme) was added to each gelpiece prior to in-gel digestion. These proteins served as internalstandards to normalize intensity readings of endogenous proteins.

Mass Spectrometry

Protein-derived peptides were analyzed by LC-MS/MS in a Q-Tof-1 mass spectrometer (Micromass, Manchester, UK) connectedon line with an Ultimate nanoflow HPLC system (LC-Packings/Dionex,Amsterdam, The Netherlands). Settings and conditions for this setuphave been described (19, 22). Gradient elution was applied from 5%B to 40% B in 60 min followed by a ramp to 90% B over 5 min.Solvent B was 80% (v/v) acetonitrile, 0.1% (v/v) formic acid, and thebalance solvent A was 0.1% (v/v) formic acid. To ensure that chro-matographic peak shapes obtained from MS traces were compatiblewith quantification and that information such as peak heights werenot lost during MS/MS analyses, we performed data-dependent ac-quisition (DDA)1 experiments with settings that switched from MS/MSto MS after just 1 s of MS/MS acquisition.

Data Analysis

Identification of Proteins by Database Searching—Deisotopedpeak lists, obtained from MS/MS raw data using Distiller Version 1(Matrix Science, London, UK), were used by Mascot Version 2.1.03(Matrix Science) to interrogate the National Center for BiotechnologyInformation non-redundant (NCBInr) July 6, 2005 protein database(containing 2,543,645 sequences) restricted to mammalian entries(391,497 sequences) or the mouse International Protein Index (IPI)August 27, 2006 database (containing 51,559 sequences). We usedIntegra (Matrix Science) for the automation of these databasesearches. This is a laboratory information management system that

1 The abbreviations used are: DDA, data-dependent acquisition;CV, coefficient of variation; GO, gene ontology; XIC, extracted ionchromatogram; Pescal, Peak Statistic Calculator; tR, retention time.

Quantitative Profile of a Mouse Core Proteome

Molecular & Cellular Proteomics 6.9 1561

also allows for effective management of proteomics data. Settingswere as follows: mass accuracy window for parent ion, 100 ppm;mass accuracy window for fragment ions, 150 millimass units; fixedmodification, carbamidomethylation of cysteines; variable modifica-tions, oxidation of methionine and acetylation of N termini. We ac-cepted the returned protein identifications when Mascot scores wereabove the statistically significant threshold (p � 0.05) and at least twopeptides matched the identified protein. Results from databasesearches were imported into Excel files.

Extraction of Quantitative Information from LC-MS/MS Data—Massspectral peaks in LC-MS/MS runs were smoothed and centroidedprior to quantitative analysis. We wrote a program in Visual Basic thatwe termed Pescal (Peak Statistic Calculator) and incorporated it intoan Excel macro. This program uses m/z and retention time (tR) valuesfor each identified peptide ion to generate extracted ion chromato-grams (XICs). The generated XICs consist of arrays of time/intensitypairs centered at the tR and m/z values entered in Excel cells withuser-defined windows, which for our experiments we set up to 3 minand 100 ppm, respectively. We determined experimentally that thesetime and mass accuracy windows were sufficiently narrow to ensurethat potential co-eluting isobaric ions were not confounding the quan-titative analysis (see “Results and Discussion”). The width of tR andm/z windows may be modified for other instruments that offer differ-ent levels of tR reproducibility and other mass accuracies. It should benoted that the narrower these windows are the less probability forconfounding the results as a consequence of co-eluting isobariccompounds. The use of narrow windows also has the effect of re-ducing background signals. Algorithms in Pescal then calculate peakheights and areas in the generated arrays and return these values tothe original Excel file for further analysis.

Normalization Procedures—Normalized ion intensities for eachpeptide X (NIX) were obtained by subtracting from the experimentalpeptide ion intensities (EIX; peak areas or heights) the intensities of thesame ions in blank LC-MS/MS runs (BIX) and by dividing this figure bythe peptide intensities of the internal standard (EIstd; see above)(Equation 1). The relative quantity of a peptide (RQPEPT) was calcu-lated relative to the mean of peptide normalized ion intensities of thispeptide across the samples to be compared (Equation 2; n is theindex for the number of samples to be compared).

NIx �EIx � BIx

EIstd(Eq. 1)

RQPEPT �NIx

1n

� �x�1

n

NIx

(Eq. 2)

Intensity values of peptides matching each protein were averagedto give a relative quantity value for each identified protein (RQPROT;Equation 3; m is the index for the number of peptides identified perprotein), which is thus reported as -fold expression relative to themean expression.

RQPROT �1m

� �PEPT�1

m

RQPEPT (Eq. 3)

For the estimation of variation, the standard deviation (S.D.) of themean (i.e. RQPROT) was calculated taking N as the number of peptidesquantified per protein (Excel was used for these statistical analyses).The percentage coefficient of variance (CV%) was calculated bydividing the S.D. by RQPROT times 100.

For absolute quantification of proteins (AQPROT), the intensities of

peptides matching a given protein were added up. This aggregateintensity value for a given protein was then divided by the total ioncurrent (TIC; which for the purpose of this manuscript was defined asthe sum of the intensity values of all the identified peptides for allproteins in an LC-MS/MS experiment) and multiplied by the percent-age of OD that this gel section (ODBAND) had relative to the OD of thetotal gel lane (ODLANE; Equation 4).

AQPROT �

�x�1

m

NIx

TIC�ODBAND

ODLANE� 100 (Eq. 4)

Hierarchical cluster analysis of quantitative LC-MS/MS data wasperformed using Cluster and Treeview (23) by complete linkage clus-tering using Pearson correlation (uncentered) similarity metric for bothgenes and arrays. Data were log10-converted prior to cluster analysis.

Gene Ontology (GO) Analyses—GO annotations were retrievedusing PIGOK (Protein Interrogation of Gene Ontology and KEGG data-bases) (24) and the IPI accession number of the identified protein entry.

RESULTS AND DISCUSSION

Numerous independent studies have shown that label-freeapproaches that use the inherent quantitative information inLC-MS/MS data are suitable for quantitative proteomics (17–19, 25). However, to be practical, this strategy requires infor-matics tools to automate the extraction of quantitative datafor all the identified peptides in LC-MS experiments (18). Wetherefore wrote a script (which we named Pescal) to automatethe generation of XICs and the calculation of their statistics(area under the curve and intensity at maximum peak height;see “Experimental Procedures” for more information). An im-portant feature of Pescal is that it calculates peak areas andheights of generated XICs for peptide ions across sampleseven if these peptides are not always selected for MS/MS inDDA experiments with the only requirement being that thepeptide is detected at least once across the samples to becompared. This is an important feature because the set ofpeptides selected for MS/MS may vary across samples due toundersampling caused by the limited duty cycle of commer-cially available mass spectrometers.

Fig. 1 shows a scheme of the strategy we used to deriveprotein quantitative information. Note that XICs are generatedfrom all the peptides identified from all data files to be com-pared irrespective of whether these peptides are present in allthe samples. In cases where a peptide is not present in asample, the returned intensity value (peak height or areas) iszero or very close to zero (background). Supplemental Tables1 and 2 provide examples on the application of this analyticalstrategy. Pescal does not allow for manual correction of peakintegration but permits calculation of peak intensities for thou-sands of peptide ions in a relatively short time (�5 peptideions/s), making it a useful tool for comprehensive quantifica-tion of proteomes.

Although it is normally believed that chromatographic peakareas are best suited as the intensity readout for the quanti-tative analysis of LC-MS and HPLC data, peak heights are


1562 Molecular & Cellular Proteomics 6.9

sometimes preferred as the intensity value (26) especiallywhen peaks are not well resolved as is often the case in theanalysis of trace compounds by HPLC. Our initial experimentsindicated that the quantitative data generated with our

method was more precise when using peak heights ratherthan peak areas as the intensity readout (data not shown).This may be because of intrinsic integration errors associatedwith the calculation of peak areas of poorly resolved chro-matographic peaks. This problem is analogous to that en-countered in the analysis of trace compounds by HPLC (26),and it may be argued that quantification of peptides in com-plex mixtures by LC-MS presents a related analytical chal-lenge. Furthermore chromatographic peak heights are moreaccurately calculated than peak areas when retention timesshift across samples. Therefore, in the experiments describedbelow, we used chromatographic peak heights as ion inten-sity readouts for correlation with protein abundance.

Validation of Chromatographic Peak Heights, as Obtainedby Pescal, as the Intensity Readout for Relative Protein Quan-tification—We aimed at assessing the performance of thisapproach for protein quantification. To investigate whetherthe chosen retention time and mass accuracy windows weresufficiently narrow to ensure accurate calculations of proteinamounts, we first performed experiments as in Ref. 19 inwhich serial dilutions of albumin (BSA) spiked in a constantamount of cell lysate from the murine WEHI-231 B cell linewere separated by SDS-PAGE (Fig. 2A), and the proteinspresent in a gel section around 65 kDa were analyzed byLC-MS/MS and Pescal. Similar experiments were done on aconstant amount of BSA spiked into a dilution series of WEHI-231 lysate before SDS-PAGE. Fig. 2B shows that, althoughproteins derived from WEHI-231 cells (and exogenouslyadded trypsin) showed the same level of expression in repli-cate analyses, BSA signals correlated linearly with theamounts spiked in the WEHI-231 cell lysates (with the dy-namic range being linear from 100 ng to at least 1000 ng ongel and 25 ng on column; this is close to the detection limit ofthe LC-MS/MS system used). Similarly the amounts of pro-teins in WEHI-231 cell lysates correlated linearly with theamounts of lysate loaded onto SDS-PAGE gels (Fig. 2C). Incontrast, constant amounts of albumin and trypsin, addedprior to and after electrophoretic separation, respectively,showed equal ion intensities in these samples (Fig. 2C).

Analysis of the variation indicated that this approach allowsfor protein quantification with reasonably good CV values(15–16% on average) with 70–80% of the data points show-ing CV values below 20% (Fig. 2, D and E). We have previ-ously reported that by careful manual integration of peaks inXICs it is possible to quantify proteins with mean CV values of�12% (19). We considered that the larger variation observedin our automated analysis (�16%) with Pescal (Fig. 2, D andE) is acceptable and that the loss of �4 percentage points inprecision may be the trade-off for being able to obtain quantityvalues for thousands of peptide ions in a time frame compatiblewith large scale analyses. Nevertheless this variation is stillrelatively low and adequate given that these data indicate thatautomated quantification of LC-MS/MS data using Pescal pro-vides levels of precision close to those obtained using strategies

FIG. 1. Scheme for label-free quantification of proteins used inthis study. Samples to be compared (there could be an unlimitednumber of samples, and they could be proteolytic digests from ac-rylamide gel pieces, HPLC fractions, or unfractionated biomaterial)are analyzed by LC-MS/MS, thus generating LC-MS data files. Peaklists in these data files are extracted and merged to perform a singledatabase search (in our case Mascot). Identified peptides are filteredso that peptide ions detected more than once across samples arelisted as a single entry. These lists contain accurate m/z and tRinformation, which is used by the Pescal software to generate XICsfrom each LC-MS data file. Pescal also calculates the peak areas andheights of these XICs and exports these values into Excel whereseveral normalization procedures are performed (described in Equa-tions 1–3 and in Supplemental Tables 1 and 2) to derive proteinquantification data. Examples and step by step descriptions on howto apply the approach are presented in Supplemental Tables 1 and 2.



that rely on chemical labeling in our previous work (22) and tothose reported for metabolic labeling (27).

Taken together, these data demonstrate that normalizedpeak heights, as returned by Pescal, are a true representationof protein abundance because these ion intensity readoutsare directly proportional to protein abundance (Fig. 2C) andcan be obtained in a reproducible manner (Fig. 2B) and withreasonably good precision (Fig. 2, D and E). We thereforeconcluded that these mass accuracy and retention time set-tings (100 ppm and 3 min, respectively) provide an adequatelevel of precision for protein quantification. However, the the-ory predicts that narrowing these windows would be advan-tageous because this would completely eliminate the possi-bility of artifacts due to the co-elution of peptides with theclose m/z values. In this respect, using the new generation ofhigh mass accuracy mass spectrometers and nanoflow HPLC

systems capable of increased retention time reproducibilitywould be desirable in principle, albeit perhaps not completelynecessary, as the data presented here indicate.

Validation of Chromatographic Peak Heights, as Obtainedby Pescal, as the Intensity Readout for Estimation of AbsoluteProtein Amounts—Although the information provided by rel-ative protein quantification is often sufficient to derive conclu-sions from certain biological experiments, the ability to obtainabsolute quantification of proteins adds value to proteomicsexperiments. Thus absolute quantification is important toclassify proteins within a sample according to abundance.This information may be particularly important for understand-ing the relative contribution of different isozymes to theirbiochemical pathway, a type of information that cannot beaccessed by performing relative quantification experiments asin “classical” proteomics studies (28).

FIG. 2. Validation of Pescal for rela-tive quantification of proteins by LC-MS/MS. A, serial dilutions of BSA orWEHI-231 B cell lysate were mixed withfixed amounts of cell lysate or BSA, re-spectively, separated by electrophoresisfollowed by excision of the bands cen-tered at �65 kDa (boxed in the gel im-age) for LC-MS/MS analysis. B and Cshow the results obtained from the BSAand cell lysate dilution experiments, re-spectively, and illustrate the linearity andreproducibility of quantitative data ob-tained by Pescal analysis of LC-MS/MSdata for endogenous B cell proteins andexogenously added BSA and trypsin. Dand E show the CV distributions for theexperiments shown in B and C, respec-tively. Error bars in A and B correspondto the S.D. of the mean expression of allthe peptides matching the named pro-tein. CV values in D and E were calcu-lated by dividing the S.D. by the meanexpression for each protein times 100.



Targeted approaches for absolute quantification that useisotopically labeled peptides have been described (29–31),but this approach requires synthesis of an internal standardpeptide for each of the proteins to be quantified. For largescale absolute quantification, it has been suggested thatspectral counts (the number of peptides selected for MS/MSin DDA experiments) roughly correlate with protein abun-dance (21), providing an approximate estimation of absoluteprotein amounts. However, the accuracy of quantitative ap-proaches based on spectral counts is very limited, and such

approaches may only provide a semiquantitative estimation ofabsolute protein levels within a sample (17).

We tested the idea that large scale absolute quantificationof proteins may be better achieved by adding up the ionintensities of all peptides derived from the same protein, thuscombining the principle behind spectral counts with the addedquantitative information stored in the ion intensities for each ofthese peptide ions. Testing this hypothesis was made possiblebecause our approach makes it possible to obtain intensityvalues for all the peptides selected for MS/MS in LC-MS/MS

FIG. 3. Validation of Pescal for ap-proximate absolute quantification ofproteins by LC-MS/MS. A, graph show-ing that total ion currents were directlyproportional to the amount of cell lysate(in protein weight) loaded in gels. B, theproportional amounts of BSA (in weight)were plotted against the proportionalamount of ion intensities. The resultshowed a strong correlation betweenthese two parameters. C, the sum of ionintensities of endogenous proteins in Bcell lysates (from the experiment in Fig.1C) produced responses that were linear(left panel) when plotted against theamount of protein loaded in gels withcorrelation coefficients (R2) approximat-ing unity in all instances (right panel).

TABLE IPrecision of absolute quantification based on addition of peptide ion intensities

The intensities of peptides matching a particular protein were added up and expressed normalized to total protein in the gel section. Theresults of five replicate experiments analyzing the expression of one gel band containing a total cell lysate fraction are shown. Raw data areshown in Supplemental Fig. 2. n, number of peptides used for quantification per protein.

GI No. Protein Name n

Amount of protein in gel section (ng/�gof total protein) in replicate experiment

number Mean S.D. CV%

1 2 3 4 5

31543113 Lymphocyte cytosolic protein 1 29 7.6 10.1 13.1 13.9 14.4 11.8 2.9 24.5984636 65-kDa macrophage protein 26 6.7 9.0 11.8 12.3 12.7 10.5 2.5 24.211066098 Transketolase 16 6.2 9.9 10.9 12.3 12.1 10.3 2.5 24.252783203 Carboxykinase 7 1.3 1.9 2.4 2.4 2.5 2.1 0.5 23.5226021 Growth-regulated nuclear 68 6 1.2 1.4 1.8 1.9 1.8 1.6 0.3 19.356385 SWAP-70 17 1.1 1.0 1.0 3.9 10.1 3.4 3.9 114.013242328 Lymphocyte cytosolic protein 1 5 1.1 1.3 1.6 1.7 1.7 1.5 0.3 18.1293689 65-kDa macrophage protein 6 1.1 1.6 1.7 2.0 2.1 1.7 0.4 25.016359229 Transketolase 8 0.9 1.2 1.4 1.5 1.5 1.3 0.3 19.816741161 Carboxykinase 6 0.6 0.8 0.8 1.0 1.1 0.8 0.2 23.419882225 Growth-regulated nuclear 68 4 0.5 0.7 0.9 1.0 1.0 0.8 0.2 24.42137619 Lymphocyte cytosolic protein 1 2 0.3 0.4 0.5 0.5 0.6 0.5 0.1 24.02323410 65-kDa macrophage protein 3 0.2 0.3 0.2 0.3 0.3 0.3 0.0 12.0



experiments. As shown in Fig. 3A, the added intensities of allpeptide ions of a gel section centered at 65 kDa were directlyproportional to the amount of cell lysate loaded in gels. In linewith this observation, increasing amounts of BSA, when nor-malized to the total amount of all proteins in the sample, pro-duced normalized ion intensities with good linearity (Fig. 3B).We also observed that the signals of added peptide ion inten-

sities derived from proteins identified in the experiment shownin Fig. 2C were linear (average correlation coefficient, R2 � 0.99)with respect to the amount of cell lysate analyzed (Fig. 3C).

We further assessed the precision of quantification by per-forming replicate experiments (summarized in Table I). Thedata, obtained by comparing five replicate experiments as inFig. 2C, demonstrated good reproducibility of quantification

FIG. 4. Strategy for the identificationand quantification of proteins ex-tracted from mouse primary tissuesand qualitative analysis of results. A,strategy for the identification and quan-tification of proteins from brain, heart,kidney, and lung murine tissues usingMascot for protein identification andPescal for their quantification. B, distri-bution of number of peptides (left) andstatistical scores (right) of returned pro-tein hits. C, molecular weight of proteinsidentified in the 11 gel fractions analyzed(plotted values represent means � S.E.of the mean). D, total ion counts per gelsection. E, F, and G, distribution of pep-tides and proteins identified per gel sec-tion. H, overview of results tabulatingsome other statistics of the data ob-tained from these experiments.



when total ion intensities for a given protein were normalizedto the sum of total intensities of the LC-MS/MS run. Takentogether, these data indicate that the sum of the peptideintensities for a given protein correlates with the proportionalamount of this protein in the protein mixture, and thereforethis information can be used to estimate protein amounts inabsolute units with reasonable precision. Thus we proposethat the proportion of ion signal for a particular protein relativeto the total ion intensity of an LC-MS/MS experiment may betranslated to the proportional weight of this protein in the totalprotein mixture.

Although we are aware that the precision of this absolutequantification approach may not be as high as that affordedby isotope dilution MS, it permits cataloguing hundreds ofproteins within a sample according to abundance with ac-ceptable confidence. This is thus another useful strategy formaximizing the amount of information obtained from LC-MS/MS data.

Quantitative Profile of the Most Abundant Proteins in MouseOrgans—Having found that our protocols can be used toquantify proteins in complex mixtures, we designed experi-ments aimed at profiling the most prominent mouse proteins.

FIG. 5. A quantitative proteomicsmap of the 1000 most abundant mem-bers of five murine proteomes. The rel-ative expression levels of proteins in dif-ferent organs were clustered using toolsfor the analysis of gene expression data(23). Black, average expression; green,expression below average; red, expres-sion above average. The figure displaysthe IPI number and description as pro-vided by the IPI database.



Fig. 4A shows a scheme of the approach used for theseexperiments. Proteins from brain, heart, kidney, liver, andlung homogenates were separated by SDS-PAGE, andlanes were cut into 11 sections. Proteins in these sectionswere digested with trypsin and analyzed by LC-MS/MS.Analyses were performed in duplicate. Mascot was used forsearches against the mouse IPI database, and returnedpeptide hits were used by Pescal to derive peak intensityvalues for each of the identified peptide sequences. Weobtained good data for 1,487 protein entries as evidencedby Mascot scores and the number of unique peptidesmatched per protein (Fig. 4B). Indeed �75% of proteinidentifications matched to more than two peptide se-quences, and �80% of these had Mascot scores �50.Another indication of the quality of the data was obtainedfrom the molecular weight distribution of the identified pro-teins (Fig. 4C), which generally agreed with the predicted

values; the exception was in the proteins found in fraction 1probably because of precipitation occurring at the interfacebetween the stacking and separating SDS-PAGE gels.Some other statistics of the data obtained with these ex-periments are shown in Fig. 4, D–H. No obvious bias wasobserved in the number of peptides of proteins identifiedper fraction (Fig. 4, E, F, and G), and the total ion counts perfraction (Fig. 4D) was consistent with Coomassie staining(Fig. 4A). After deleting duplicates (i.e. proteins that werepresent in more than one organ), we obtained a protein listconsisting of 942 entries that matched 8,842 peptides. Ionintensity values for these peptides were obtained for the fivedifferent organs analyzed, leading to a total number of44,210 independent quantitative data points and the quan-tification of 4,710 proteins (942 proteins per organ).

Relative Quantification of Proteins in Mouse Organs—Ourdata allowed comparison of the expression levels of the major

FIG. 6. Validation of the protein ex-pression data obtained from Pescaland LC-MS/MS analysis of mamma-lian primary tissues. A, expression pat-terns of selected marker proteins forspecific organs showing expected levelsof enrichment. B, comparison of thequantitative information obtained usingWestern blot (WB) and Pescal analysis ofLC-MS/MS data for actin-1 and �-tubu-lin. Western blots were carried out intriplicate; ODs were measured by fluo-rescence and expressed as mean � S.D.Insets show representative Western blotimages. C, correlation of the quantitativeinformation obtained from the datashown in B. Error bars correspond to theS.D. of the mean of all the peptidesmatching the named protein.



proteins present in distinct mouse tissues. We clustered thedata using approaches originally designed for the analysis ofcDNA microarray data (Fig. 5), allowing easy visualization ofthe protein patterns of these five mouse organs. The list of allthe identified proteins together with their levels of expressionis provided as Supplemental Tables 4 and 5.

These quantitative data were first validated by comparisonof the expression levels of proteins known to be enriched inspecific organs. For example, Fig. 6A shows that synaptotag-min-1 and excitatory amino acid transporter 1, two proteinsinvolved in neurotransmission (32, 33), were found to be moreabundant in brain. Tissue-specific isoforms of fatty acid-bind-ing proteins were enriched in the respective organs (the heartisoform in heart and the liver isoform in liver; Fig. 6A). Similarlyand as expected, megalin (the multifunctional protein receptorresponsible for reabsorbing polypeptides from the glomerularfiltrate) and meprin (a glomerular structural protein) were ex-

pressed at high levels in the kidney. Also in line with expecta-tions was the high expression of myoglobin in heart and alcoholdehydrogenase-1 in liver and kidney (Fig. 6A) consistent withthe known expression pattern of this metabolic enzyme (34, 35).Fig. 6A also shows enriched expression of the lung isoform ofcarbonyl reductase and endothelial filamin A in the lung; thelatter is consistent with the abundance of endothelial cells in thistissue. Other proteins were detected with similar levels of ex-pression across different tissues. As an example, �-hemoglobinand HSP-90� produced similar ion intensities across the differ-ent tissue samples (Fig. 6A) consistent with a similar degree oferythrocyte presence in these organs and the chaperone HSP-90� being a ubiquitous housekeeping protein (36).

We also validated the data by comparing the relative ex-pression levels of �-tubulin and actin-1 by Western blot usingspecific antibodies and by Pescal analysis of LC-MS/MSdata. Fig. 6B shows that expression patterns of �-tubulin and

FIG. 7. Overview of absolute quanti-fication values of protein expressionin different murine organs. A, plot ofabsolute protein levels of the majormouse proteins in five different organsand their average. Each point representsa protein entry. Error bars correspond tothe S.D. of the mean of all the peptidesmatching the named protein. B, absoluteprotein amounts of the most abundantproteins in brain, heart, kidney, lung, andliver. 3-P, 3-phosphate.



actin-1 were similar when analyzed by these two differentmethods, a finding that was confirmed by the high degree ofcorrelation between the two sets of quantitative data (Fig. 6C).Further analysis of replicates indicated good reproducibility ofthe quantitative data (Supplemental Fig. 1 and SupplementalTable 3).

Absolute Quantification of Mouse Organ Proteins—As dis-cussed above, our approach can also provide an estimationof the absolute abundance of proteins in complex mixtures.Thus we obtained estimates of absolute quantities of all iden-tified proteins as a percentage of total ion intensities, which,as discussed above, we propose may be translated to per-centage of total weight (or mg of protein/100 mg of total tissueprotein). Fig. 7A shows that the quantitative readings spannedat least 3 orders of magnitude, whereas Fig. 7B indicates thatmetabolic enzymes (e.g. �-enolase) and structural proteins(e.g. actin-1) are among the most abundant proteins in theseorgans. Also highly present are proteins such as albumin dueto the presence of blood in the organs. Nevertheless some ofthese abundant proteins are also known to have organ-spe-cific roles. Thus, Na�/K�-ATPases subunits were found torepresent a high percentage of the brain and kidney pro-teomes (Fig. 7B) when compared with those of heart, lung,and liver. This is consistent with the high requirement of

neurons and renal epithelial cells for the electrical and con-centration gradients that are generated by the Na�/K�-ATPase. These gradients are used by nerve cells for thetransmission of information in the form of postsynaptic po-tentials and by renal epithelial cells for the reabsorption ofsolutes in the glomerular filtrate. It is also interesting to notethat enzymes involved in the generation of ATP (ATP synthase� and �) constitute a larger percentage of heart and kidneyproteomes than in other organs (Fig. 7B). This finding is con-sistent with the high energy requirements of these organs tomaintain constant heart muscle contraction and for the largerate of vacuolar ATPase-dependent endocytosis taking placein renal epithelial cells (37).

Gene Ontology Analysis Identifies the Most Prevalent Cel-lular Functions—We next performed a bioinformatics analysisof our dataset to provide insight into the most common cel-lular functions in the different organs analyzed (Fig. 8). For thiswe used a program termed PIGOK (24) to query GO annota-tions on the IPI protein database (38). About 40 and 60% of allthe identified proteins returned GO annotations on biologicaland molecular function process, respectively. We then com-pared the total number of proteins matching a particular GOannotation with the total absolute amounts of proteins (cal-culated by LC-MS and Pescal) within each GO description.

FIG. 8. Gene ontology analysis iden-tifies the most prevalent cellular bio-logical processes and molecular func-tions in five different primary tissues.A, absolute protein amounts of proteinsmatching a particular molecular functionor biological process were added andplotted in the graphs. B, graphs showaggregate protein amounts in absoluteunits of quantity for selected geneontologies.



We noted that the number of proteins matching a particularGO description did not always correlate with the combinedion intensities (Fig. 8). As an example, Fig. 8A shows that thenumber of proteins matching structural, carrier, transporter,regulator, and motor ontologies was lower than their com-bined intensities. In other words, the number of genes match-ing a particular GO group may not always reflect the abun-dance of this activity or process in cells and argues for the useof absolute protein amounts for refined functional analysis ofprotein expression data.

Metabolism was the most enriched biological process in alltissues with about 16% of the total ion intensity correspond-ing to metabolic proteins (Fig. 8A). Other abundant biologicalprocesses in all tissues were transport (14%), cell death (9%),and development (7%) (Fig. 8A).

As for their molecular function, proteins with oxidoreduc-tase, hydrolase, and transferase activities produced the larg-est combined intensity values in all tissues (Fig. 8A), indicatingthat these are the three most abundant biochemical functionsin primary murine cells. Overall these prevalent molecularfunctions were about 2 orders of magnitude more abundant(as judged by their combined intensities) than proteins match-ing signal transduction and antioxidant ontologies and 1 orderof magnitude more abundant than receptors, regulatory pro-teins, and proteins with kinase activity. These values may bean underestimate because only about 60% of all proteinsidentified had GO annotations. However, we do not expectthe ratios to change significantly after all entries have beenannotated.

The differences in molecular function and biological proc-ess (as determined by GO analysis) between the differentorgans were not pronounced. Fig. 8B shows that proteinsinvolved in metabolic functions produced a similar proportionof added ion intensities in the different organs. Neverthelesswe observed several interesting differences in the functionalclassification of the proteomes analyzed. For example, pro-teins with ligase and transferase ontologies produced moreabundant ion signals in liver than in brain, heart, and lung (Fig.8B). Ligase and transferase activities are involved in biosyn-thetic pathways, and their enrichment in the liver may reflecta high rate of anabolic metabolism, a known liver function, inthis organ (39, 40).

Interestingly proteins with signal transduction and proteintransport activity were 3–5-fold more abundant in brain than inthe other organs, whereas other functions were diluted in thisorgan, including proteins with electron transport activity(needed for the generation of metabolic energy in mitochondria)whose added intensities were �3.5-fold less abundant in brainthan in heart. This finding is consistent with the �3-fold highernumber of mitochondria in rodent heart than in brain (41, 42).

Comparison between Relative and Absolute Protein Quan-tification—Our dataset also allowed us to directly comparethe merits and utilities of relative and absolute methods forprotein quantification. As an example, we assessed relative

and absolute expression of the isoforms of the enolase en-zyme in different organs. In agreement with published data(43), the expression of �-enolase was found to be enriched inbrain (Fig. 9A, top panel); expression of �-enolase and �-eno-lase was more equally distributed in the different tissues.However, absolute quantification (Fig. 9A, bottom panel) re-vealed that, although �-enolase expression was more pro-nounced in brain relative to other organs, the most abundantenolase isozyme (in absolute terms) was the � isoform in allorgans analyzed, including the brain.

The expression patterns of the 17 distinct small GTPasesthat we identified provide another example of the extra infor-mation provided by absolute protein quantification. Relativequantification revealed that most Rab isoforms were �2 timesmore abundant in brain than in other organs (Fig. 9B, upper

FIG. 9. Comparison of data obtained by relative and absolutequantification of proteins. A, relative (top panel) and absolute (bot-tom panel) expression of enolase isozymes. B, analyses of relativeand absolute expression of GTPases are shown in the top and bottompanels, respectively.



panel, graphs 1–12). In contrast, R-Ras expression seems tobe diluted in brain and more abundant in the lung (Fig. 9B,graph 17). These data, however, did not allow us to assesswhich of the identified small GTPases are more abundantwithin a defined tissue.

Absolute quantitation analysis revealed that the most abun-dant Rab isoforms in brain are Rab-1A, Rab-3A, and Rab-3D(Fig. 9B, bottom panel), which are approximately 1 order ofmagnitude more abundant in this tissue than other membersof the Rab family of small GTPases such as Rab-11B andRab-15. It is also interesting to note that although relativequantification suggests that R-Ras may be more abundant inlung than in any other organ (Fig. 9B, upper panel, graph 17),assessment of absolute amounts indicates that R-Ras is ex-pressed at low amounts in all tissues analyzed, including thelung (Fig. 9B, bottom panel).

Conclusion—Using a first generation Q-TOF mass spec-trometer, the approach described here allowed the quantifi-cation the �1,000 most abundant proteins in different murineorgans, thus demonstrating the power of this technique forquantitative proteomics. This approach can be used for pro-filing the proteomes of cell lines and primary tissues andcomparing an unlimited number of samples is facilitated bythe fact that it is not restricted to the number of availableisotopic labels. These are two important requirements forphenotypic analysis of primary clinical samples and cells andtissues from model organisms. The bioinformatics tools andstandardization procedures described here (in combinationwith new generation mass spectrometers) will provide us withthe opportunity for maximizing the biochemical informationthat may be obtained from our gene targeting efforts (5,44–46), which are ultimately aimed at understanding themechanisms of signal transduction in physiologically relevantsystems.

Acknowledgments—We thank M. D. Waterfield for allowing accessto analytical instrumentation, M. Graupera for animal dissection, R.Jacobs for advice using Integra, and John Timms for feedback on themanuscript.

* Work in the laboratory of B. V. was supported by the LudwigInstitute for Cancer Research. This work was also supported byadditional funds from the Association for International Cancer Re-search (to P. R. C. and B. V.). The costs of publication of this articlewere defrayed in part by the payment of page charges. This articlemust therefore be hereby marked “advertisement” in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

□S The on-line version of this article (available at http://www.mcponline.org) contains supplemental material.

¶ Present address: Analytical Signalling Laboratory, Centre for CellSignalling, Inst. of Cancer, Barts and the London Medical School, 3rdfl. John Vane Science Bldg., Charterhouse Square, London EC1M6BQ, UK.

� To whom correspondence should be addressed: Cell Signalling inCancer Laboratory, Ludwig Inst. for Cancer Research, 91 RidingHouse St., London W1W 7BS, UK. Tel.: 44-020-7882-8264; E-mail:[email protected].

REFERENCES

1. Battey, J., Jordan, E., Cox, D., and Dove, W. (1999) An action plan formouse genomics. Nat. Genet. 21, 73–75

2. Mouse Genome Sequencing Consortium (2002) Initial sequencing andcomparative analysis of the mouse genome. Nature 420, 520–562

3. The FANTOM Consortium and the RIKEN Genome Exploration ResearchGroup Phase I & II Team (2002) Analysis of the mouse transcriptomebased on functional annotation of 60,770 full-length cDNAs. Nature 420,563–573

4. Foster, L. J., de Hoog, C. L., Zhang, Y., Zhang, Y., Xie, X., Mootha, V. K.,and Mann, M. (2006) A mammalian organelle map by protein correlationprofiling. Cell 125, 187–199

5. Foukas, L. C., Claret, M., Pearce, W., Okkenhaug, K., Meek, S., Peskett, E.,Sancho, S., Smith, A. J., Withers, D. J., and Vanhaesebroeck, B. (2006)Critical role for the p110� phosphoinositide-3-OH kinase in growth andmetabolic regulation. Nature 441, 366–370

6. Kadowaki, T. (2000) Insights into insulin resistance and type 2 diabetesfrom knockout mouse models. J. Clin. Investig. 106, 459–465

7. Vidal-Puig, A. J., Grujic, D., Zhang, C. Y., Hagen, T., Boss, O., Ido, Y.,Szczepanik, A., Wade, J., Mootha, V., Cortright, R., Muoio, D. M., andLowell, B. B. (2000) Energy metabolism in uncoupling protein 3 geneknockout mice. J. Biol. Chem. 275, 16258–16266

8. Dobkin, C., Rabe, A., Dumas, R., El Idrissi, A., Haubenstock, H., and Brown,W. T. (2000) Fmr1 knockout mouse has a distinctive strain-specificlearning impairment. Neuroscience 100, 423–429

9. Silva, A. J., Frankland, P. W., Marowitz, Z., Friedman, E., Laszlo, G. S.,Cioffi, D., Jacks, T., and Bourtchuladze, R. (1997) A mouse model for thelearning and memory deficits associated with neurofibromatosis type I.Nat. Genet. 15, 281–284

10. Rao, S., and Verkman, A. S. (2000) Analysis of organ physiology in trans-genic mice. Am. J. Physiol. 279, C1–C18

11. Ong, S. E., Blagoev, B., Kratchmarova, I., Kristensen, D. B., Steen, H.,Pandey, A., and Mann, M. (2002) Stable isotope labeling by amino acidsin cell culture, SILAC, as a simple and accurate approach to expressionproteomics. Mol. Cell. Proteomics 1, 376–386

12. Gygi, S. P., Rist, B., Gerber, S. A., Turecek, F., Gelb, M. H., and Aebersold,R. (1999) Quantitative analysis of complex protein mixtures using iso-tope-coded affinity tags. Nat. Biotechnol. 17, 994–999

13. Ross, P. L., Huang, Y. N., Marchese, J. N., Williamson, B., Parker, K.,Hattan, S., Khainovski, N., Pillai, S., Dey, S., Daniels, S., Purkayastha, S.,Juhasz, P., Martin, S., Bartlet-Jones, M., He, F., Jacobson, A., andPappin, D. J. (2004) Multiplexed protein quantitation in Saccharomycescerevisiae using amine-reactive isobaric tagging reagents. Mol. Cell.Proteomics 3, 1154–1169

14. Bonenfant, D., Schmelzle, T., Jacinto, E., Crespo, J. L., Mini, T., Hall, M. N.,and Jenoe, P. (2003) Quantitation of changes in protein phosphorylation:a simple method based on stable isotope labeling and mass spectrom-etry. Proc. Natl. Acad. Sci. U. S. A. 100, 880–885

15. Gharbi, S., Gaffney, P., Yang, A., Zvelebil, M. J., Cramer, R., Waterfield,M. D., and Timms, J. F. (2002) Evaluation of two-dimensional differentialgel electrophoresis for proteomic expression analysis of a model breastcancer cell system. Mol. Cell. Proteomics 1, 91–98

16. Kislinger, T., Cox, B., Kannan, A., Chung, C., Hu, P., Ignatchenko, A., Scott,M. S., Gramolini, A. O., Morris, Q., Hallett, M. T., Rossant, J., Hughes,T. R., Frey, B., and Emili, A. (2006) Global survey of organ and organelleprotein expression in mouse: combined proteomic and transcriptomicprofiling. Cell 125, 173–186

17. Old, W. M., Meyer-Arendt, K., Aveline-Wolf, L., Pierce, K. G., Mendoza, A.,Sevinsky, J. R., Resing, K. A., and Ahn, N. G. (2005) Comparison oflabel-free methods for quantifying human proteins by shotgun proteom-ics. Mol. Cell. Proteomics 4, 1487–1502

18. Jaffe, J. D., Mani, D. R., Leptos, K. C., Church, G. M., Gillette, M. A., andCarr, S. A. (2006) PEPPeR, a platform for experimental proteomic patternrecognition. Mol. Cell. Proteomics 5, 1927–1941

19. Cutillas, P. R., Geering, B., Waterfield, M. D., and Vanhaesebroeck, B.(2005) Quantification of gel-separated proteins and their phosphorylationsites by LC-MS using unlabeled internal standards: analysis of phospho-protein dynamics in a B cell lymphoma cell line. Mol. Cell. Proteomics 4,1038–1051

20. Mallick, P., Schirle, M., Chen, S. S., Flory, M. R., Lee, H., Martin, D., Ranish,J., Raught, B., Schmitt, R., Werner, T., Kuster, B., and Aebersold, R.



(2007) Computational prediction of proteotypic peptides for quantitativeproteomics. Nat. Biotechnol. 25, 125–131

21. Ishihama, Y., Oda, Y., Tabata, T., Sato, T., Nagasu, T., Rappsilber, J., andMann, M. (2005) Exponentially modified protein abundance index (em-PAI) for estimation of absolute protein amount in proteomics by thenumber of sequenced peptides per protein. Mol. Cell. Proteomics 4,1265–1272

22. Cutillas, P. R., Chalkley, R. J., Hansen, K. C., Cramer, R., Norden, A. G.,Waterfield, M. D., Burlingame, A. L., and Unwin, R. J. (2004) The urinaryproteome in Fanconi syndrome implies specificity in the reabsorption ofproteins by renal proximal tubule cells. Am. J. Physiol. 287, F353–F364

23. Eisen, M. B., Spellman, P. T., Brown, P. O., and Botstein, D. (1998) Clusteranalysis and display of genome-wide expression patterns. Proc. Natl.Acad. Sci. U. S. A. 95, 14863–14868

24. Jacob, R. J., and Cramer, R. (2006) PIGOK: linking protein identity to geneontology and function. J. Proteome Res. 5, 3429–3432

25. Silva, J. C., Denny, R., Dorschel, C., Gorenstein, M. V., Li, G. Z., Richard-son, K., Wall, D., and Geromanos, S. J. (2006) Simultaneous qualitativeand quantitative analysis of the Escherichia coli proteome: a sweet tale.Mol. Cell. Proteomics 5, 589–607

26. Snyder, L. R., Kirkland, J. J., and Glajch, J. L. (1997) Practical HPLCMethod Development, 2nd Ed., pp. 666–680, John Wiley and Sons, Inc.,New York

27. Olsen, J. V., Blagoev, B., Gnad, F., Macek, B., Kumar, C., Mortensen, P.,and Mann, M. (2006) Global, in vivo, and site-specific phosphorylationdynamics in signaling networks. Cell 127, 635–648

28. Geering, B., Cutillas, P. R., and Vanhaesebroeck, B. (2007) Regulation ofPI3K: is there a role for free subunits. Biochem. Soc. Trans. 35, 199–203

29. Anderson, L., and Hunter, C. L. (2006) Quantitative mass spectrometricmultiple reaction monitoring assays for major plasma proteins. Mol. Cell.Proteomics 5, 573–588

30. Steen, H., Jebanathirajah, J. A., Springer, M., and Kirschner, M. W. (2005)Stable isotope-free relative and absolute quantitation of protein phos-phorylation stoichiometry by MS. Proc. Natl. Acad. Sci. U. S. A. 102,3948–3953

31. Cutillas, P. R., Khwaja, A., Graupera, M., Pearce, W., Gharbi, S., Waterfield,M., and Vanhaesebroeck, B. (2006) Ultrasensitive and absolute quanti-fication of the phosphoinositide 3-kinase/Akt signal transduction path-way by mass spectrometry. Proc. Natl. Acad. Sci. U. S. A. 103,8959–8964

32. Andrews, N. W., and Chakrabarti, S. (2005) There’s more to life thanneurotransmission: the regulation of exocytosis by synaptotagmin VII.Trends Cell Biol. 15, 626–631

33. Yoshihara, M., and Montana, E. S. (2004) The synaptotagmins: calcium

sensors for vesicular trafficking. Neuroscientist 10, 566–57434. Balak, K. J., Keith, R. H., and Felder, M. R. (1982) Genetic and develop-

mental regulation of mouse liver alcohol dehydrogenase. J. Biol. Chem.257, 15000–15007

35. Buhler, R., Pestalozzi, D., Hess, M., and von Wartburg, J. P. (1983) Immu-nohistochemical localization of alcohol dehydrogenase in human kidney,endocrine organs and brain. Pharmacol. Biochem. Behav. 18, Suppl. 1,55–59

36. Burrows, F., Zhang, H., and Kamal, A. (2004) Hsp90 activation and cellcycle regulation. Cell Cycle 3, 1530–1536

37. Christensen, E. I., and Birn, H. (2002) Megalin and cubilin: multifunctionalendocytic receptors. Nat. Rev. Mol. Cell. Biol. 3, 256–266

38. Camon, E., Magrane, M., Barrell, D., Binns, D., Fleischmann, W., Kersey, P.,Mulder, N., Oinn, T., Maslen, J., Cox, A., and Apweiler, R. (2003) TheGene Ontology Annotation (GOA) project: implementation of GO inSWISS-PROT, TrEMBL, and InterPro. Genome Res. 13, 662–672

39. Mason, T. M. (1998) The role of factors that regulate the synthesis andsecretion of very-low-density lipoprotein by hepatocytes. Crit. Rev. Clin.Lab. Sci. 35, 461–487

40. Russell, D. W. (2003) The enzymes, regulation, and genetics of bile acidsynthesis. Annu. Rev. Biochem. 72, 137–174

41. Clark, J. B., and Nicklas, W. J. (1970) The metabolism of rat brain mito-chondria. Preparation and characterization. J. Biol. Chem. 245,4724–4731

42. LaNoue, K., Nicklas, W. J., and Williamson, J. R. (1970) Control of citric acidcycle activity in rat heart mitochondria. J. Biol. Chem. 245, 102–111

43. Sakimura, K., Kushiya, E., Takahashi, Y., and Suzuki, Y. (1987) The struc-ture and expression of neuron-specific enolase gene. Gene (Amst.) 60,103–113

44. Ali, K., Bilancio, A., Thomas, M., Pearce, W., Gilfillan, A. M., Tkaczyk, C.,Kuehn, N., Gray, A., Giddings, J., Peskett, E., Fox, R., Bruce, I., Walker,C., Sawyer, C., Okkenhaug, K., Finan, P., and Vanhaesebroeck, B. (2004)Essential role for the p110delta phosphoinositide 3-kinase in the allergicresponse. Nature 431, 1007–1011

45. Bilancio, A., Okkenhaug, K., Camps, M., Emery, J. L., Ruckle, T., Rommel,C., and Vanhaesebroeck, B. (2006) Key role of the p110delta isoform ofPI3K in B-cell antigen and IL-4 receptor signaling: comparative analysisof genetic and pharmacologic interference with p110delta function in Bcells. Blood 107, 642–650

46. Okkenhaug, K., Bilancio, A., Farjot, G., Priddle, H., Sancho, S., Peskett, E.,Pearce, W., Meek, S. E., Salpekar, A., Waterfield, M. D., Smith, A. J., andVanhaesebroeck, B. (2002) Impaired B and T cell antigen receptor sig-naling in p110� PI 3-kinase mutant mice. Science 297, 1031–1034



Documents

Quantitative Profile of Five Murine Core Proteomes Using Label-free Functional Proteomics