Measuring gene expression by quantitative proteome analysis

Measuring gene expression by quantitative proteome analysis

396 Measuring gene expression by quantitative proteome analysis Steven P Gygi, Beate Rist and Ruedi Aebersold Proteome analysis is most commonly acco...

231KB Sizes 0 Downloads 21 Views

396

Measuring gene expression by quantitative proteome analysis Steven P Gygi, Beate Rist and Ruedi Aebersold Proteome analysis is most commonly accomplished by the combination of two-dimensional gel electrophoresis for protein separation, visualization, and quantification and mass spectrometry for protein identification. Over the past year, exceptional progress has been made towards developing a new technology base for the precise quantification and identification of proteins in complex mixtures, that is, quantitative proteomics. Addresses Department of Molecular Biotechnology, University of Washington, Box 357730, Seattle, WA 98195-7730, USA Correspondence: Steven P Gygi; e-mail: [email protected]

ability to display, quantify, and identify thousands of proteins in a single gel [4]. Closer examination of the proteins routinely identified by proteome studies, however, suggests that 2DE-MS does not represent a truly global technique. Several reports have analyzed this question and proposed potential answers. Some classes of proteins have long been known to be excluded or underrepresented in 2D gel patterns. These include very acidic or basic proteins, excessively large or small proteins, and membrane proteins. By examining codon bias values (see below) of proteins detected in 2D gels, it has now been shown that the 2DE-MS approach is incapable of measuring lowabundance proteins without pre-gel enrichment [5••].

Current Opinion in Biotechnology 2000, 11:396–401 0958-1669/00/$ — see front matter © 2000 Elsevier Science Ltd. All rights reserved. Abbreviations 2DE two-dimensional gel electrophoresis CBD codon bias distribution HPLC high-performance liquid chromatography ICAT isotope-coded affinity tag LC liquid chromatography MS mass spectrometry RP reversed phase

Introduction A major goal of proteomics is the global and quantitative measurement of the proteins expressed in cells or tissues [1]. The purpose of this review is to emphasize the past, current status, and recent advances in proteome analysis in light of this goal. We will be discussing recent papers presenting both incremental and landmark advances. The emerging field of proteomics has grown out of the mature technology of high-resolution two-dimensional gel electrophoresis (2DE) for protein separation and quantification [2] and increasingly refined technologies for the identification of separated proteins. Today, mass spectrometry (MS) is overwhelmingly utilized as the technology base for protein identification from 2D gels [3] (Figure 1). However, the transition from identifying a single protein from one spot in a 2D gel to systematically cataloging all features present is not trivial and has required many complimentary advances. Today, 2DE and protein MS represent an integrated technology by which several thousand protein species can be separated, detected and quantified in a single operation, and hundreds of the detected proteins can be identified in a highly automated fashion by sequential analysis of the peptide mixtures generated by digestion of individual gel spots.

2DE-MS for global proteome analysis It is commonly assumed that 2DE-MS can serve as the technology base for global proteome analysis based on its

It is thought that the codon bias value for a gene is a measure of protein abundance because highly expressed proteins generally have large codon bias values [6]. There are 61 possible codons that code for 20 amino acids. Codon bias is a measure of the propensity of an organism to selectively utilize certain codons, which result in the incorporation of the same amino acid residue in a growing polypeptide chain. Low-abundance proteins (e.g. transcription factors, protein kinases) generally have low codon bias values (<0.1). Figure 2a shows the codon bias distribution (CBD) for the known yeast genome (6139 genes) [7]. The graph shows that more than one half of yeast genes have codon bias values of <0.2 with more than 2500 genes in the 0.0–0.1 range alone. Figure 2b shows the combined CBD for all proteins detected from the major yeast proteome 2D gel studies (343 proteins) [4,5••,6,7,8•,9•,10] and indicates that the CBD pattern for the proteins from 2D gels is highly biased towards abundant proteins because almost no proteins with codon bias values <0.1 were detected even though more than onehalf of all yeast genes have codon bias values <0.1. Proteins of all abundances including low-abundance proteins can be analyzed, however, if starting loads are increased beyond the capacities of 2D gels [5••] or if total cell lysates are prefractionated prior to separation by 2DE [11]. The introduction of new fluorescent stains for protein detection in 2D gels has also improved the dynamic range of detection and facilitated protein quantification [12]. However, these stains only marginally improve the detectability of low-abundance proteins from unseparated cell lysates. The ramifications from these studies are threefold. First, the number of spots on a 2D gel is clearly not representative of the overall number of genes expressed in the analyzed sample. Second, the analysis of total cell lysates by standard 2DE techniques only measures abundant proteins and important classes of regulatory proteins (transcription factors, protein kinases, etc.) are not detected. Third, sample prefractionation allows the detection of low-abundance proteins, but quantification is probably lost.

Measuring gene expression by quantitative proteome analysis Gygi, Rist and Aebersold

397

Figure 1 (a)

12

Peptides

14 Time (min)

16

Separate peptides (b)

Fragmentation 200 400 600 800 1000 1200

m/z Q1

Q2 Collision cell

Q3

MS/MS spectrum

(C) Computer sequence database searching

200

400

600

800 1000 1200

m/z

Peptide identification Protein identification

Theoretical

200

400

600

800 1000 1200

m/z Acquired Current Opinion in Biotechnology

Schematic illustration of standard proteome analysis by 2DE-MS. (a) Proteins are separated by 2DE, stained spots are excised, subjected to in-gel digestion with trypsin, and the resulting peptides are separated by on-line HPLC. (b) An eluting peptide is ionized by electrospray ionization, enters the mass spectrometer, and is

fragmented to collect sequence information (MS/MS spectrum). (c) The MS/MS spectrum from the selected, ionized peptide is compared to predicted tandem mass spectra computer generated from a sequence database to identify the protein.

Alternative separation techniques to 2DE for proteome analysis

chromatography (LC)-MS/MS. Co-eluting peptides can be selected, isolated, and fragmented sequentially and automatically within the MS with no overlap in recorded fragmentation patterns. Sophisticated database searching programs permit further automation of the identification process [15–18]. The sequence information for thousands of peptides can be recorded in a single LC-MS/MS analysis. Comparison of the acquired fragmentation pattern for each peptide with a computer-generated theoretical fragmentation pattern results in peptide, and by association protein, sequence information. The number of proteins that can be identified is only constrained by the length of the analysis and the complexity of the mixtures. Shabanowitz et al. [19•] have been using an MS strategy for highly complex peptide mixtures that employs a single MS scan followed by five MS/MS (sequencing) scans on the five most-intense peptide ions in that scan. To further

To alleviate some of the limitations of 2DE, alternative separation techniques have been integrated with MS as new proteome analysis platforms. Andrews and co-workers [13] used precise mass measurement by matrix-assisted laser desorption ionization (MALDI) MS. Proteins are first separated by isoelectric focusing then analyzed directly to generate ‘virtual’ 2D gels. Oda et al. [14••] utilized preparative high-performance liquid chromatography (HPLC) to fractionate whole yeast lysate prior to separation by SDS-PAGE, which greatly increased the overall protein starting amounts that could be analyzed. A promising alternative method is the direct analysis of highly complex peptide mixtures generated by the digestion of unseparated protein mixtures by liquid

398

Protein technologies and commercial enzymes

Figure 2 (a)

3000

No. of genes

2500

Yeast proteome

2000 1500 1000 500 0

(b)

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

70

2D gel identified

No. of genes

60 50 40 30 20 10 0

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Codon bias Current Opinion in Biotechnology

Current proteome analysis technology based on 2DE does not facilitate the analysis of low-abundance proteins from total cell lysates. Genes encoding highly expressed proteins generally have large codon bias values (>0.2). (a) Codon bias distribution (CBD) of the yeast genome (6139 genes). The interval with the largest frequency of genes is 0.0–0.1 with more than 2500 genes. Light grey bars show probable low-abundance proteins (codon bias value <0.1). (b) CBD of the genes from identified proteins (343) including all major yeast proteome studies to date [4,5••,8•,9•,10]. Only three proteins were found from genes with codon bias values <0.1.

increase the number of peptides sequenced in a single analysis, they complemented this with an elegant implementation of the established ‘peak parking’ method [20,21], which reduces the flow from capillary columns from ~200 nl/min to 20 nl/min to extend elution times for specific peaks. Up to 104 sequencing attempts in a single analysis can then be recorded. To further enhance the peak capacity of the direct LC-MS/MS approach, some researchers have been working on multi-dimensional separation techniques. Link et al. [22•] have reported a proteome analysis technique called DALPC (direct analysis of protein complexes). DALPC uses the orthoganal physical properties of charge and hydrophobicity to resolve complex peptide mixtures before analysis by mass spectrometry. A denatured and reduced protein mixture is first digested to generate a mixture of peptide fragments. The acidified complex peptide mixture is applied to a strong cation exchange (SCX) chromatography column, and discrete fractions of the absorbed peptides are sequentially displaced onto a reversed-phase (RP) chromatography column using a salt step gradient. Peptides are then analyzed by LC-MS/MS. The process successfully cataloged the components of the 80S yeast

ribosome. Comparing their 2D chromatographic technique to 1D (RP only) indicated that more than four times as many peptides were identified (95 different proteins — probably the entire complex) with the 2D technique. In addition to analyzing the approximately equimolar components of large protein complexes, a variation of the 2D chromatography-MS/MS approach has also been a very effective part of a strategy to identify low-abundance protein in yeast lysates [5••]. In this study, we separated 50 mg of whole-cell extracts by preparative SDS-PAGE. A strip containing proteins of sizes 68–85 kDa was excised from the coomassie-stained gel and digested with trypsin. The resulting peptides were separated and analyzed by gradient strong cation exchange chromatography and then capillary RP-LC-MS/MS. Of a total of 193 proteins that were positively identified in the sample, 63 had codon bias values <0.1. In spite of these advances, proteome analysis by direct mixture analysis faces significant limitations. These include extreme complexity of the peptide mixture, matrix effects, and limited dynamic range. All of these can be overcome by careful application and incremental advances to the current technique. The most substantial problem with the direct LC-MS/MS approach to proteome analysis is the complete loss of quantification [23]. The 2DE-MS/MS approach maintains relative quantification of proteins at the level of the 2D gel. Without protein quantification in the gel, only the presence or absence of a particular protein in a mixture can be noted, but not its relative abundance. This past year, a number of laboratories have developed novel techniques, all of which make use of stable isotope dilution theory, to measure the quantity of protein analyzed.

Quantitative proteomics using stable isotope dilution The newest methods for quantitative proteome analysis all make use of the venerable technique of stable isotope labeling [24] in a new context. The method involves the addition to the sample of a chemically identical form of the analyte(s) containing stable heavy isotopes (2H, 13C, 15N, etc.) as internal standards. Because ionization efficiency is highly variable for peptides, the only suitable internal standard for a candidate peptide is that same peptide labeled with stable isotopes. Therefore, protein profiling is accomplished if two protein mixtures from two different conditions or cell states are compared where one serves as the reference sample, containing the same proteins as the other sample but at different abundances and labeled with heavy stable isotopes. In theory, all peptides from the combined samples then exist as analyte pairs of identical sequence but different masses. The peptide pairs have the same physico-chemical properties and behave similarly under any conceivable isolation or separation step. Thus, the ratios between the intensities of the lower and upper mass components of these pairs of peaks provides an accurate measure of the relative abundance of

Measuring gene expression by quantitative proteome analysis Gygi, Rist and Aebersold

Figure 3 legend The ICAT strategy for quantifying differential protein expression. (a) Structure of the ICAT reagent. The reagent consists of three elements: firstly, an affinity tag (biotin), which is used to isolate ICATlabeled peptides; secondly, a linker that can incorporate stable isotopes; and thirdly, a reactive group with specificity toward thiol groups (cysteines). The reagent exists in two forms, namely heavy (containing eight deuteriums) and light (containing no deuteriums). (b) Schematic of the ICAT strategy. The method shows the analysis of a single protein but is equally applicable to total cell lysates. Protein from two different cell states is harvested, denatured, reduced, and labeled at cysteines with the light or heavy ICAT reagents, respectively. The samples are then combined and proteolyzed. ICAT-labeled peptides are isolated via the biotin tag by affinity chromatography and then analyzed by online HPLC coupled to a tandem mass spectrometer. The ratio of the ion intensities for an ICAT-labeled pair quantifies the relative abundance of its parent protein in the original cell state. In addition, a tandem mass spectrum reveals the sequence of the peptide and unambiguously identifies the protein. This strategy results in the quantification and identification of all protein components in a mixture, and it is applicable in theory to protein mixtures as complex as the entire proteome.

Figure 3 ICAT Reagents Heavy reagent: d8-ICAT (X = deuterium) Light reagent: d0-ICAT (X = hydrogen)

(a)

Biotin

(b)

Cell state 1

Cell state 2

ht lig

ICAT label hea vy cysteines

Combine, trypsinize

Affinity isolation

Mass spectrometry Quantitate relative protein levels by measuring peak ratios Relative abundance

100

0

Mass difference from stable isotopes

430

440

Ratio: 0.33

450 460 m/z

470

480

Identify peptide by sequence information (MS/MS scan) Relative abundance

100

NH2-EACDPLR-COOH

0 200

Pasa-Tolic et al. [25•] used stable isotope depleted media to impart a specific isotope signature into proteins. They compared the cadmium stress response in Escherichia coli grown in normal and rare-iosotope depleted (13C-, 15Nand 2H depleted) media. Intact protein mass measurements were carried out by Fourier transform ion cyclotron resonance (FTICR) MS. Although no protein was

Thiol-specific reactive group

Linker (heavy or light)

the peptides (and hence the protein) in the original protein mixtures. Three groups have independently reported measuring protein profiles based on stable isotopes [14••,25•,26••], and two others are preparing manuscripts (Langen H et al., personal communication; James P et al., personal communication). The techniques differ in the method of incorporation of heavy isotopes and in the analytical procedures used. Oda et al. [14••] grew one yeast culture on medium containing the natural nitrogen isotope distribution (14N, 99.6%; 15N, 0.4%), while another culture was grown on the same medium enriched in 15N (>96%). After an appropriate growing period, the cell pools were combined, and proteins of interest were extracted and separated by RP-HPLC and then by SDS-PAGE. In-gel digestion of excised spots of interest resulted in peptide fragments, which were used for protein identification by mass mapping. Each incorporated 15N atom shifted the mass of any given peptide upwards by one mass unit, leading to a pair of peaks from each peptide. The authors measured protein expression of 42 high-abundance proteins derived from two pools of S. cerevisiae that differed only in their ability to express the G1 cyclin CLN2. The percentage error of the experimental technique was found to be excellent (± 10%). The authors went on to also measure differential phosphorylation states in the yeast protein Ste20 by the same technique by extending the technique to phosphopeptides where only unmodified peptides had been analyzed.

399

400

600

800

m/z Current Opinion in Biotechnology

positively identified, the expression ratios for 200 different proteins were compared.

400

Protein technologies and commercial enzymes

Clearly, growth of protein in 15N-enriched or 15N-depleted media permits protein profiling with an important new advantage of precise quantification; however, this method has several disadvantages. First, the method does not allow for the analysis of protein directly from tissue. Second, the stable-isotope-enriched media might themselves affect microbial growth and protein production. Third, stable-isotope-enriched media are costly, and for culturing cells from higher organisms they may be impossible to obtain. Fourth, the increase in nominal mass due to stable-isotope incorporation is not known until the sequence is determined, which can greatly confound database-searching programs and prevent protein identification prior to quantification. We have recently published a novel method for quantitative protein profiling based on isotope-coded affinity tags (ICAT) [26••]. In this method, the stable isotopes are incorporated post isolation by selective alkylation of cysteines with either a heavy (d8) or light (d0) reagent (Figure 3). The two protein mixtures are then mixed. At this point, any optional fractionation technique can be performed to enrich for low-abundance proteins or to reduce the complexity of the mixture, while the relative quantities are strictly maintained. Prior to analysis, the protein mixture is digested with trypsin and passed over a monomeric avidin-agarose column. Because the ICAT label contains the stable isotope information as well as a biotin tag, ICAT-labeled (cysteinecontaining) peptides are selectively isolated for analysis by microcapillary LC-electrospray ionization-MS/MS. The ratio of ion intensities from co-eluting ICAT-labeled pairs permits the quantification, while a subsequent MS/MS scan provides the protein identification. Protein expression profiles were compared from yeast growing on either galactose or ethanol in a single analysis. The advantages of the ICAT strategy are several fold. First, the method is compatible with any amount of protein harvested from bodily fluids, cells or tissues under any growth conditions. Second, the alkylation reaction is highly specific and occurs in the presence of salts, detergents, and stabilizers (e.g. SDS, urea, guanidine-HCl). Third, the complexity of the peptide mixture is reduced by isolating only cysteinecontaining peptides. Fourth, the ICAT strategy permits almost any type of biochemical, immunological, or physical fractionization, which makes it compatible with the analysis of low-abundance proteins. There are two disadvantages to the method. First, the size of the ICAT label (~500 Da) is a large modification that remains on each peptide throughout the MS analysis. This can complicate the database searching algorithms, especially for small peptides (<7 amino acids). Second, the method fails for proteins that contain no cysteines. Only a small percentage of proteins are cysteine-free (8% in yeast), however, and ICAT reagents with specificities to groups other than thiols could be synthesized.

Conclusions 2DE-MS is recognized as the standard approach to proteome analysis. Recent studies have indicated some of the

fundamental limitations of this approach, particularly as they pertain to the detection of low-abundance proteins. It is clear that the 2DE-MS approach is useful for the identification of marker proteins from total cell lysates, but that it is critically limited for proteome analysis because lowabundance proteins are absent. Alternative separation techniques coupled to MS are showing promising results. Protein mixtures as complex as the entire proteome can be analyzed because milligram or gram amounts of protein can be separated. These techniques are extraordinarily powerful when multidimensional orthogonal separation techniques are used prior to MS analysis of peptides; however, any information about protein abundance is lost. Accurate quantification of complex protein mixtures has been accomplished using stable isotope dilution theory. The protein in one mixture contains stable heavy isotopes and serves as the internal standard. Another protein mixture can then be compared to this reference by examining the ratio of peptide ion intensities by MS. Precise quantification of gene expression at the proteome level by the methods discussed is a crucial step forward in the analysis of biological systems. In the near future, global gene expression data measured at both the mRNA and protein levels will provide an intensely accurate and meaningful view of interacting systems and pathways in cells, tissues, and organisms.

Acknowledgements This work was supported by grants from the NIH (HG00041, RR11823, T32HG00035, CA84698, A141109), NSF (BIR 9214821) and Merck Genome Research Institute.

References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as:

• of special interest •• of outstanding interest 1.

Blackstock WP, Weir MP: Proteomics: quantitative and physical mapping of cellular proteins. Trends Biotechnol 1999, 17:121-127.

2.

O’Farrell PH: High resolution two-dimensional electrophoresis of proteins. J Biol Chem 1975, 250:4007-4021.

3.

Patterson SD, Aebersold R: Mass spectrometric approaches for the identification of gel-separated proteins. Electrophoresis 1995, 16:1791-1814.

4.

Shevchenko A, Jensen ON, Podtelejnikov AV, Sagliocco F, Wilm M, Vorm O, Mortensen P, Shevchenko A, Boucherie H, Mann M: Linking genome and proteome by mass spectrometry: large-scale identification of yeast proteins from two dimensional gels. Proc Natl Acad Sci USA 1996, 93:14440-14445.

5. ••

Gygi SG, Corthals GL, Zhang Y, Rochon Y, Aebersold R: Evaluation of two-dimensional gel electrophoresis based proteome analysis technology. Proc Natl Acad Sci USA 2000, in press. This paper thoroughly assesses 2D-gel-based proteome analysis. The paper examines the classes of proteins identified from 2D gels and concludes that low-abundance proteins can not be detected from proteome analysis of unfractionated cell lysates. 6.

Kurland CG: Codon bias and gene expression. FEBS Lett 1991, 285:165-169.

7.

Costanzo MC, Hogan JD, Cusick ME, Davis BP, Fancher AM, Hodges PE, Kondu P, Lengieza C, Lew-Smith JE, Lingner C et al.: The Yeast Proteome Database (YPD) and Caenorhabditis elegans Proteome Database (WormPD): comprehensive resources for the organization and comparison of model organism protein information. Nucleic Acids Res 2000, 28:73-76.

Measuring gene expression by quantitative proteome analysis Gygi, Rist and Aebersold

8. •

Gygi SP, Rochon Y, Franza BR, Aebersold R: Correlation between protein and mRNA abundance in yeast. Mol Cell Biol 1999, 19:1720-1730. This paper examines gene expression on two levels: protein and mRNA. The classes and types of proteins identified are typical of 2DE experiments using total cell lysates. 9. Futcher B, Latter GI, Monardo P, McLaughlin CS, Garrels JI: • A sampling of the yeast proteome. Mol Cell Biol 1999, 19:7357-7368. This paper examines gene expression on two levels: protein and mRNA. The classes and types of proteins identified are typical of 2DE experiments using total cell lysates. 10. Perrot M, Sagliocco F, Mini T, Monribot C, Schneider U, Shevchenko A, Mann M, Jeno P, Boucherie H: Two-dimensional gel protein database of Saccharomyces cerevisiae (update 1999). Electrophoresis 1999, 20:2280-2298. 11. Corthals GL, Molloy MP, Herbert BR, Williams KL, Gooley AA: Prefractionation of protein samples prior to two-dimensional electrophoresis. Electrophoresis 1997, 18:317-323. 12. Berggren K, Steinberg TH, Lauber WM, Carroll JA, Lopez MF, Chernokalskaya E, Zieske L, Diwu Z, Haugland RP, Patton WF: A luminescent ruthenium complex for ultrasensitive detection of proteins immobilized on membrane supports. Anal Biochem 1999, 276:129-143. 13. Loo JA, Brown J, Critchley G, Mitchell C, Andrews PC, Ogorzalek Loo RR: High sensitivity mass spectrometric methods for obtaining intact molecular weights from gel-separated proteins. Electrophoresis 1999, 20:743-748. 14. Oda Y, Huang K, Cross FR, Cowburn D, Chait BT: Accurate •• quantitation of protein expression and site-specific phosphorylation. Proc Natl Acad Sci USA 1999, 96:6591-6596. A significant paper demonstrating quantitative protein profiling of two different cell states in yeast using stable isotopes and MS. 15. Eng J, McCormack AL, Yates JR: An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J Am Soc Mass Spectrom 1994, 5:976-989. 16. Mann M, Wilm M: Error-tolerant identification of peptides in sequence databases by peptide sequence tags. Anal Chem 1994, 66:4390-4399. 17.

Qin J, Fenyo D, Zhao Y, Hall WW, Chao DM, Wilson CJ, Young RA, Chait BT: A strategy for rapid, high-confidence protein identification. Anal Chem 1997, 69:3995-4001.

401

18. Clauser KR, Hall SC, Smith DM, Webb JW, Andrews LE, Tran HM, Epstein LB, Burlingame AL: Rapid mass spectrometric peptide sequencing and mass matching for characterization of human melanoma proteins isolated by two-dimensional PAGE. Proc Natl Acad Sci USA 1995, 92:5072-5076. 19. Shabanowitz J, Settlage RE, Marto JA, Christian RE, White FM, • Russo PS, Martin SE, Hunt DF: Sequencing the primordial soup. In Mass Spectrometry in Biology and Medicine. Edited by Burlingame AL. Totowa: Humana Press; 2000:163-177. An excellent chapter covering the technical difficulties associated with largescale peptide mixture analysis. 20. Davis MT, Lee TD: Rapid protein identification using a microscale electrospray LC/MS system on an ion trap mass spectrometer. J Am Soc Mass Spectrom 1998, 9:194-201. 21. Figeys D, Corthals GL, Gallis B, Goodlett DR, Ducret A, Corson MA, Aebersold R: Data-dependent modulation of solid-phase extraction capillary electrophoresis for the analysis of complex peptide and phosphopeptide mixtures by tandem mass spectrometry: application to endothelial nitric oxide synthase. Anal Chem 1999, 71:2279-2287. 22. Link J, Eng J, Schieltz DM, Carmack E, Mize GJ, Morris DR, • Garvik BM, Yates JR: Direct analysis of large protein complexes using mass spectrometry. Nat Biotechnol 1999, 17:676-682. This paper presents a method for large-scale protein analysis that bypasses 2D gels by using 2D chromatography coupled online to a mass spectrometer. 23. Mann M: Quantitative proteomics? Nat Biotechnol 1999, 17:954-955. 24. De Leenheer AP, Thienpont LM: Application of isotope dilutionmass spectrometry in clinical chemistry, pharmacokinetics, and toxicology. Mass Spectrom Rev 1992, 11:249-307. 25. Pasa-Tolic L, Jensen PK, Anderson GA, Lipton MS, Peden KK, • Martinovic S, Tolic N, Bruce JE, Smith RD: High throughput proteome-wide precision measurements of protein expression using mass spectrometry. J Am Chem Soc 1999, 121:7949-7950. A short paper demonstrating quantitative protein profiling where the expression ratios for 200 E. coli proteins were measured. 26. Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH, Aebersold R: •• Quantitative analysis of protein mixtures using isotope coded affinity tags. Nat Biotechnol 1999, 17:994-999. A novel method using selective derivatization with stable isotopes to permit quantitative protein profiling. Quantitative protein expression profiles were performed on a large scale for yeast growing on either galactose or ethanol as a carbon source.