ISSN 1990-7508, Biochemistry (Moscow) Supplement Series B: Biomedical Chemistry, 2008, Vol. 2, No. 1, pp. 28–46. © Pleiades Publishing, Ltd., 2008. Original Russian Text © A.T. Kopylov, V.G. Zgoda, 2008, published in Biomeditsinskaya Khimiya.
PROTEOMICS
The Methods of Quantitative Proteomics A. T. Kopylov* and V. G. Zgoda Institute of Biomedical Chemistry, Russian Academy of Medical Sciences, Pogodinskaya ul. 10, Moscow, 119121 Russia; fax: +7 (495) 245 08 57; e-mail:
[email protected] Received May 31, 2007
Abstract—In modern science proteomic analysis is inseparable from other fields of systemic biology. Possessing huge resources quantitative proteomics operates colossal information on molecular mechanisms of life. Advances in proteomics help researchers to solve complex problems of cell signaling, posttranslational modification, structure and funciotnal homology of proteins, molecular diagnostics etc. More than 40 various methods have been developed in proteomics for quantitative analysis of proteins. Although each method is unique and has certain advantages and disadvantages all these use various isotope labels (tags). In this review we will consider the most popular and effective methods employing both chemical modifications of proteins and also metabolic and enzymatic methods of isotope labeling. Key words: quantitative proteomics, mass spectrometry, isotope label, chemical modification. DOI: 10.1134/S1990750808010034
teomics. So this review contains sections dealing with historical background of quantitative proteomics, the main approaches of protein chemistry, description and characterization of targets employed in modification chemistry, technical basis and bioinformation resources, representing the modern approach in the analysis of proteomic data. There are enormous data available in literature and therefore we have focused our attention on the latest results and on consideration of advantages and disadvantages of the most interesting and widely used methods.
INTRODUCTION Proteomics is a branch of science, which specializes in systemic study of structure and function of proteins, protein-protein interactions, determination of levels of gene expression. The development of a technical basis for multimeric separation of proteins and peptides, appearance of rapid and sensitive mass spectrometers and method of bioinformatics resulted in accumulation of large quantity and quality of proteomic data. The technology of chemical modification of protein molecules made the main contribution to the development of modern quantitative proteomics. Since 1960-1970s of the last century it became a basis of the technology of chemical protein modification as the main tool for absolute and relative quantification of protein molecules (including those undergoing posttranslational changes). Appearance of so-called isotope labels widely accepted and used by scientific community gave a tool for solution of various research problems in proteomics.
1. HISTORICAL BACKGROUND Quantitative proteomics specializes on systemic analysis of content of proteins including their modified forms. The method of two-dimensional electrophoresis (2D-PAGE) made substantial contribution to the development of proteomics. Since the middle of 1970s 2D-PAGE played the central role in the methodology of quantitative proteomics for more than 25 years due to its high-resolution capacity. Figure 1 shows the scheme of determination of content of protein products; this was usually achieved by comparative analysis of stained gels with separated protein molecules from cell lysates followed by their identification by mass spectrometry (MS). Separation of molecules involves two parameters typical for each individual protein: charge of a protein molecule (the first dimension) and its electrophoretic mobility (in acrylamide gel of certain density) correlated with molecular mass (the second dimension). Relative quantitative evaluation occurs during comparison of a proteome profile of 2D-maps and densitometry of
In this review we have considered some aspects of the development of methodology of quantitative pro*To whom correspondence should be addressed. Abreviations used: 2D-PAGE—two dimensional poly-acrylamide gel electrophoresis; CID—collision ion dissociation; ECD—electron capture dissociation; ESI—electro spray ionization; ETD— electron transfer dissociation; FD—fluorescence detection; FT-ICR-MS—Fourier transformation ion cyclotrone resonance mass spectrometry; LC—liquid chromatography; MALDI—matrix assisted laser desorbtion/ionization); MS—mass spectrometry; MS/MS—tandem mass spectrometry; PAGE—poly-acrylamide gel electrophoresis; TOF—time of flight.
28
THE METHODS OF QUANTITATIVE PROTEOMICS
29
Two-dimensional electrophoresis and excision of protein spots fro the gel
In-gel enzymatic or chemical cleavage of proteins
Biosample preparation from cell lysates
Recording of mass-spectra by means of a mass-spectrometer
Intens × 105 +MS, 37.7–37.7 min #5485#5488
536.2
1.5 1.0
610.2 685.1
0.5
460.7 743.0
857.0
933.1 1010.5 1089.3
1141.3 1188.6
1258.4
0 500
600
700
800
900
1000
1100
1200
1300 m/z
Fig. 1. Protein identification in biosamples. Proteins of cell lysate are separated by two-dimensional electrophoresis. The protein spots excised from the stained gels are subjected to in-gel proteolysis. The resultant peptides are then analyzed by mass spectrometry.
spots of corresponding proteins. However, it was difficult to achieve high reproducibility of experiments due to different sensitivity of staining agents, conditions for acrylamide gel polymerization and protein separation and also multistep preparations of gels for the first and the second dimensions. Attempts undertaken to overcome these problems resulted in the development of twodimensional difference gel electrophoresis (2D-DIGE) [1–4] (Fig. 2). Control and experimental samples were labeled by means of cyan fluorophores (Cy3 and Cy5) tightly bound to ε-amino group of lysine side chain and then mixed in equimolar ratios (by protein) for subsequent separation within one gel. Quantitative evaluation of differences was carried out by comparing two images
obtained during scanning at different emission wavelengths, typical for each fluorophore. Subsequent use of this method resulted in introduction of the third fluorophore, Cy2 [5, 6], and the development of new versions of Cy3 and Cy5 [7–9] binding cysteine, but this innovation complicated sample preparation and increased price of reagent. The methods based on liquid chromatography played an important role in separation of proteome (the total sum of proteins typical for particular organism) into individual proteins. These include liquid phase isoelectrofocusing [10–12], chromatofocusing [13–17], ion exchange [18] and mixed sorbents excluding molecules by their size and charge; all these methods were used as the first dimension before reverse phase liquid
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
30
KOPYLOV, ZGODA Protein extract B labeled with CY5
Protein extract A labeled with CY3 Sample pooling
Separation by two-dimensional electrophoresis
Scanning of CY3 labeled proteins
Scanning of CY5 labeled proteins
Fig. 2. The scheme illustrating the procedure of two-dimensional difference gel electrophoresis. Control and experimental samples are labeled by different fluorescent labels, pooled together and separated by gel electrophoresis. The gels are then scanned for Cy3 or Cy5 fluorescence using different emission wavelengths, typical for each fluorophore. Similarities and differences are analyzed after image overlapping.
chromatography. Each of these approaches gave reasonably good resolution, however, non of them would be able to separate the whole proteome (more than 104 proteins). It became clear that the future depends on improvement of various technologies for multimer separation. Ion exchange chromatography was quite popular for protein separation before PAGE [19], however, there was one disadvantage: some proportion of sample proteins would not bind to totally positively or negatively charged sorbents due to charge heterogeneity on the surface of protein molecule. In experiments with separation of simple mixtures El Rassi and Horvath [20] demonstrated that effective binding to sorbent and separation may be achieved by means of combined sor-
bents. However, this idea did not receive further development. Although none of the method of chromatographic separation of proteins with high resolution can provide quantitative evaluation of certain protein in a mixture the achievements of this methodology have not been forgotten and they are now employed in the strategy of proteomic analysis alternative to 2D-electrophoresis. Mass-spectrometry combines separation of complex mixtures of proteins and peptides with amino acid sequencing by tandem mass spectrometry (MS/MS). Primary structures of about 270000 protein molecules are known to date; they were obtained by means of various approaches: classical Edman degradation reaction, nuclear-magnetic resonance, small angle X-ray scattering and mass-spectrometry. Nevertheless, mass-spectrometry cannot give final information about quantitative content of proteins in the proteome due to limitations of technical capacities, ionization capacity (niiniaiinoe eiiecaoee), representing individual characteristics of peptide and many other physico-chemical factors. However, now researchers have created several approaches combining chemical modifications of protein molecules and their isotope labeling, which mainly represents additional tool for evaluation of relative and sometimes even absolute protein content in organisms under various conditions. Since 1999 tandem mass-spectrometry is employed in combination with isotope labeling and this provides quantitative evaluation [21–23]. Investigation of two cultures of Sacchoromices cerevisiae was one of the earliest studies on quantitative proteomics employing isotope labeling. Authors used mutant and wild type of yeasts grown in the mediun enriched with 14N or 15N [22–25]. The methods of quantitative evaluation by means of an isotope label are based on determination of ratios of MS and MS/MS peak squares for chemically identical peptides characterized by certain shift in molecular mass. Such procedure includes the following steps: different isotope labeling of protein mixtures, enzymatic cleavage of pooled differentially labeled mixtures, peptide separation by multimer liquid chromatography; MS/MS analysis of separated peptides. There are many labels differing by mass, nature of isotope constituents, binding capacity and specificity to particular amino acids. Such high diversity requires employment of various technologies and approaches for their incorporation into polypeptides and proteins. Introduction of various isotopes occurs via different ways including metabolic, enzymatic, chemical labeling, solid phase methods of isolation etc. Labeling may involve whole protein molecules, as it happens during metabolic labeling; however, such approach is not widely used due to necessity of metabolic availability and it is applicable only to cell cultures in vitro including prokaryotes, yeasts C. elegance and D. melanogaster. It is also possible to use isotopes for selective labeling of rare amino acid residues in polypeptide
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS H N O
X X
X
O
O
X
O
X X
X X
31 S
H N O
HN
NH O
Fig. 3. Structural formula of the ICAT-label: reactive iodoacetate cysteine binding group linked with biotin group via polyether linker region containing 8 deuterium atoms in the heavy chain or 8 hydrogen atoms (markes as X) in the light chain.
chain: cysteine, tyrosine, methionine, tryptophan. The isotope labeling frequently employs 13C, D, and 15N atoms. There are technologies of isotope labeling of peptides by their N-terminal residues or free amino groups of lysine, and also amide groups of glutamine and asparagine. In some cases labeling involves C-terminal residues or free carboxyl groups of peptides. The major reason by which C-terminal labeling is not as popular as the N-terminal one consists in difficulties of discrimination of C-terminal carboxyl group and carboxyl groups of glutamate or aspartate. Consequently, any modifications of carboxyl group would result in multiple incorporations of an isotope label into all sites of an acidic peptide. Isotope coded affinity tag (ICAT) is the most popular method of isotope labeling. The ICAT-labels (chemically identical but differing by isotope composition) are covalently bound to cysteine residues of polypeptides [25]. 2. CHEMICAL METHODS OF ISOTOPE LABELING IN QUANTITATIVE PROTEOMICS 2.1. ICAT (Site-specific Labeling of Cysteine Residues) ICAT is the well-described method used for determination of relative protein content [23, 26, 27]. A prototype of this method was described by Aebersold and Mann in 1999 (Fig. 3) [28, 29]. Originally the ICATcomplex represented a reactive iodoacetate cysteine binding group linked with biotin group via polyether linker region. This significantly simplified isolation of labeled peptides using an avidin column. ICATs containing 8 deuterium atoms [D8] and 8 hydrogen atoms [H8] are defined as heavy and light chain, respectively. Figure 4 shows a scheme illustrating a general plan of experiment using ICAT-label. The method is based on covalent labeling of polypeptide cysteine residue by chemically identical ICAT reagents of different isotope composition. Subsequent proteolysis, multimer liquid chromatography and MS-analysis provided information on sequence of isolated peptides and their relative content [23, 25]. However, some proteins lack cysteine residues and thus ICAT is inapplicable for their identification and quantitative evaluation. In a new version of the ICAT technology there are internal standards labeled by the light chain they may
be used for absolute quantitative determination of proteins in samples labeled with the heavy chain. Although this method is widely used it has some shortcomings. For example, use of deuterium for specific labeling results in chromatographic separation of light and heavy variants of chemically identical peptides. In addition the label is rather large (450 Da) and this also complicates MS/MS analysis of spectra. The procedure of affinity chromatography is often accompanied by irreversible or nonspecific peptide binding. Finally, experimentations with cysteine-containing forms of peptides caused loss of information on posttranslational modifications [30]. In 2002 Applied Biosystems (USA) developed a new cleavable analogue, known as cICAT (Cleavable Isotope Coding Affinity Tag). The cleavable cICAT reagents consist of sulfhydryl reactive moiety, cleavable biotin “tail” and a linker sequences; the presence of nine 12C or 13C atoms underlines belonging to heavy or light chain. Control and experimental samples were labeled by the heavy or the light variant of the cICAT reagent and then were pooled together, enzymatically cleaved, and analyzed by LC-MS. Relative peak height of peptides labeled by 12C and 13C variants of cICAT gave information about relative content of corresponding proteins and various samples. The 13C-labeled peptides had a 9-Da mass shift compared with the 12C-variants. Originally cICAT has been used in 2003 [31–33] and since that time the new ICAT version has been accepted by scientific community. Yu et al. [34] reported about effectiveness of cICAT label, specificity, possible problems originating at the stage of affinity chromatography, chromatographic co-elution of peptides, labeled with heavy or light chain and accuracy of relative quantitative evaluation. Some aspects of reproducibility of and accuracy of experiments with cICAT were also discussed in publications by Parker et al. [35] and Molloy et al. [36]. Subsequently, the most interesting studies employing cICAT included identification of protein thiol groups sensitive to oxidation [37, 38] and use of this strategy for identification of proteins in various subcellular organelles [39–41]. One of the latest reports on a new application of cICAT has been reported by et Jenkins al. [27]. Authors used this technology for the absolute quantification cytochrome P450 isoforms isolated from mouse liver. The cICAT method demonstrated reasonable sensitivity and authors were
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
32
KOPYLOV, ZGODA
Proteins of sample 1 Proteins of sample 2
Reduction of sulfhydryl groups of cysteine residues
Isotope coded labeling with light and heavy ICATTM probes
Pooling of samples and trypsinolysis
Separation of peptides by multimeric chromatography
LC-MS/MS
Relative quantitative evaluation
Protein identification
Fig. 4. The scheme of experiment employing the isotope coded affinity tag technology (ICAT).
able to identify Cyp1a1, which was not detected in control samples. For absolute quantification of 16 isoforms of cytochrome P450 they employed slightly modified cICAT protocol. Using this method it was possible to identify and quantify 50–100 protein simultaneously. Using Fmoc chemistry they synthesized peptides containing cysteine residues and representing fragments of primary structure of the cytochromes studied. Peptides of known concentration were labeled with cICAT light chain and used as internal standards. Data obtained on absolute quantification of cytochrome P450 were presented at the level of subfamilies. Effectiveness of protein quantification depended on amino acid composition of peptides and their elution from the column. Resultant data on absolute and relative quantification were basically consistent to each other and also to expression profile of certain isoforms of cytochrome P450 in control liver samples and liver sample obtained
from mice treated with classic inductor of cytochrome P450, phenobarbital. Only in the case of cytochromes Cyp3a/11/16/41 there were some divergences of data on absolute and relative quantification. Authors indicated that they did not employ the whole potential of this method (at least 50 proteins) it was very sensitive, specific and might be used for parallel analysis of several proteins. Ranish et al. [42] carried out experiments on quantitative proteomics of multicomponent protein complexes. They studied subunit content of the preinitiation complex of RNA polymerase II (PIC II, about 68 subunits). Proteins of control complex purified on a cation exchange column were labeled with the isotopically light and heavy ICAT reagent and then subjected to trypsinolysis. The ICAT-labeled peptides were analyzed by microLC-ESI MS/MS. In the isolated fraction about 58% of identified proteins (189 of 326) were not related to PIC II. Authors indicated that using multimer chromatography and ICAT technology it was possible to identify and analyze proteins involved into transcription and leave out of consideration contaminant proteins. Although these authors quantitatively analyzed 206 proteins, nevertheless some subunits of RNA polymerase II (7 of 68, about 10%) could not be analyzed due to lack of cysteine-containing ions of tryptic peptides with m/z 1+, 2+, 3+. Experts suggest to overcome this problem by employment of N-terminal labeling. Ranish et al. reported about some difference in elution time of peptides labeled with light and heavy cICAT chains during chromatographic procedure. Using cICAT Min Shen’ group [43] also faced numerous problems in their studies. Firstly, MS/MS analysis revealed the presence of peptides lacking cysteine. Interestingly, the signal of such nonspecifically bound peptides was rather potent and comparable to the signal of the cICAT-labeled peptides containing cysteine residues. Secondly, tryptic avidin peptides were clearly registered in these spectra in spite of use of specific trypsin inhibitor and was of avidin column with buffer (pH 5). Thirdly, ICAT adds 450 Da to molecular mass. If a peptide contains more than one cysteine residue its mass will be correspondingly increased and this influences quality of MS/MS spectra. Fourthly, the addition of ICAT reagent tends to increase the charge state of peptide ions and this also does not improve quality of data. However, in spite of all the above-mentioned shortcomings of this method the cICAT technology is still widely employed in proteomic studies due to its highresolution capacity in quantitative analysis, possibility to cover the major part of proteome proteins (provided that cysteine residues are present in their primary structure) and rather simple procedure of isotope labeling.
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
2.2. ICROS (Methionine and Cysteine Specific Labeling) In attempt to overcome inconveniences and shortcomings of the ICAT method Min Shen’n group [43] developed a new and inexpensive method known as ICROS (Isotope Coded Reduction off of Chromatographic Support). Although it is not very complex and original but this is highly effective method. The first step of the ICROS method consists in dissolution of total protein, denaturation and reduction of disulfide bonds followed by subsequent proteolytic proteins under denaturing conditions in the presence of urea. Subsequent steps include removal of salts and disulfide bond reducing agent by means of solid phase extraction (SPE) on a commercially available column and enrichment of cysteine containing peptides using a pyridyl disulfide column and elution of bound peptides. Purified peptides are then subjected to alkylation by means of the deuterium [D]-containing reagents. In that study N-ethyl-iodoacetamide and N-[D5]-ethyl-iodoacetamide were used as the alkylating isotope reagents; employment of these reagents gave a 5 Da molecular mass difference. According to standard procedure after alkylation the peptides were labeled by light and heavy reagents and then were mixed for subsequent LC-MS analysis. Unbound peptides (freely passing the pyridyl disulfide column) were acylated by [D0]- or [D4]-nicotinic acid N-hydroxysuccinimide ester. This approach preserved amino groups and it did not cause negative effect on sensitivity of MS detection. The O-ethers were hydrolyzed using hydroxylamine hydrochloride. Methionine-containing peptides were then isolated using a commercially available column (BioMolecular Technologies, Sunnyvale, USA). Thus, in such experiments both cysteine- and methionine-containing peptides were analyzed and employment of such approach was more informative and ruled out some artifacts typical for the ICAT technology. This experimental study has demonstrated the following characteristics features of the ICROS method: (1) lack of nonspecifically bound peptides; (2) disappearance of potential problems of sample contamination with tryptic fragments of co-eluted avidin-bound peptides, which always accompany an analyzed sample in the case of the ICAT method; (3) alkylation with N-ethyl-iodoacetamide and N-[D5]-ethyl-iodoacetamide gives additional increase of molecular mass by 85 and 90 Da per each cysteine residue of peptide and therefore charge of the analyzed proteins remain unchanged. The effectiveness of label incorporation by the ICROS method was 55–75% in dependence of amino acid sequences of peptides. The ICROS method is very similar to the methods previously used by Wang and Regnier [44]. However the ICROS method is characterized by high sensitivity in simplicity. These authors originally employed isotope labeling by means of succinylation of amino group
33
(with [D0] and [D4] succinic anhydride) and basic groups acquired acidic properties due to attachment of carboxyl group. The problems of poor solubility of pyridyl disulfide reagent was overcome by binding of succinylated peptides onto a reverse phase column playing a role of the reaction chamber in which diluted solution of pyridyl disulfide constantly washed immobilized peptides. These peptides were then eluted from the column and mixed with thiopropyl-Sepharose, on which the pyridyl disulfide labeled peptides formed disulfide bonds with the sorbent. Column was then washed with reducing agents and peptides were simultaneously alkylated with iodacetate. In the case of ICROS peptides are already derivatized with pyridyl disulfide and the step of formation of mixed disulfide is not necessary, because ICROS procedure is accompanied by simultaneous formation of mixed peptides. Subsequent steps on elution of cysteine-containing peptides, isotope labeling and alkylation with N-[D0]- or N-[D5]-ethyl-iodacetamide occur simultaneously and peptide amino groups remain unmodified. Authors indicate that employment of such chemical modifications increases sensitivity of this method compared with the method by Wang and Regnier. Thus, the ICROS method has evident advantages compared with ICAT and therefore it is more convenient in the experimental work. However, ICROS also has some shortcomings. One methodical difference of ICROS from ICAT consists in labeling of peptides after their proteolytic cleavage. This technology has essential disadvantage: isotope labeled peptides are not pooled together up to the final step. This results in separate preparation of control and experimental sample during all preceded step: peptides should be separately cleaved, covalently captured with pyridyl disulfide and subjected to isotope labeling before pooling. This separate work may result in artifacts in subsequent study because of accumulation of errors at each step of experiment. However, subsequent automation would decrease the routine of this work. 2.3. Guanidation (Lysine Specific Labeling) Guanidination is rather simple form of chemical modification. During this modification C-terminal lysine residues of tryptic peptides are converted into homoarginine in the chemical reaction with methylisourea (Fig. 5). This rather simple reaction increases basic properties of peptide ions and increases their detectability. It should be noted that in contrast to many other methods of modification chemistry guanidation may be used for introduction of various isotope labels for peptides and this gives significant advantages and flexibility in the analytical approach. Warwood et al. [45] used guanidation chemistry for relative quantitative proteomic analysis of proteins iso-
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
34
KOPYLOV, ZGODA HOOC CH NH2
HOOC CH NH2
CH2 CH2 CH2 CH2 NH2
HN C
+ H3C
NH2
CH2 CH2
O
CH2
· HX
CH2
· HX NH2
N C NH2
Fig. 5. Guanidation reaction followed by formation of homoarginine from lysine.
lated from FDCP-mix cells of the mouse cell culture. Proteins were subjected to ingel proteolysis and the mixture of resultant peptides were labeled with O-methylisourea, containing the nitrogen isotopes [15N] or [14N]. The reaction yielded arginine analogue, homoarginine (lysine derivative), derivatized with O-[14N]- or O-[15N]methylisourea. Derivatives were pooled together, desalinated and analyzed by MALDI-MS. A 2-Da mass difference between isotope labeled derivatives represents certain problem in the case of overlapping of chemically labeled isotopes and natural isotope distribution. Authors analyzed 46 protein spots on the 2D-gel and identified 42 proteins; after derivatization increased number of matched peptides was found in 36 of 42 proteins and increased percent of coverage of amino acid sequence was found in 30 proteins. Authors attribute such results to the increase of basic properties of proteins after their derivatization, which exhibits positive influence on sensitivity of measurement and extends capacities of protein detection due to increased number of ionized peptides. 2.4. Isotope Labeling of Lysine and N-terminal Amino Groups Many methods of labeling via primary α-amino groups are known in proteomics. Hoang and coauthors [46] described one of them. These authors employed two-step labeling. On the first step they used sulfosuccinimidyl-2-(biotinamido) ethyl-1,3-dithiopropionate, reacting with primary amino groups (of lysine or unmodified N-terminal groups of peptides). During the second step authors used [12CD]- or [13CD3]-methyliodide. Covalent binding with biotin-containing label facilitated isolation peptides of interest by means of a column with immobilized streptavidin. Disulfide bond has two functions in the linker site. Firstly, it provides convenient conditions for peptide elution from the column by means of washing with reducing agents. Secondly, after elution reduced thiol groups are used for methylation with the isotope label. Such mechanism may insert light of heavy labels for subsequent quantitative evaluation.
In a pilot experiment authors used bradykinin as a research object. Derivatization by means of sulfosuccinimidyl-2-(biotinamido) ethyl-1,3-dithiopropionate was rather effective because MALDI-TOF (Matrix Assisted Laser Desorption/Ionization—Time Of Flight) did not demonstrate the presence of peaks corresponding to unmodified bradykinin. Employment of DTT (as a reducing reagent) for column elution caused the decrease of peptide mass by 301.4 Da. This corresponds to biotin group with a linker site fragment. Alkylation added 14 Da to molecular mass (although there was a small proportion of non-alkylated peptides), which gave a 4 Da difference between peptides labeled with heavy and light isotope labels. The method proposed by Hoang et al. is similar to that of ICAT, however it has some characteristic features. Firstly, sulfo-NHS-SS-biotin contains a cleavable linker site, which is removed before MS analysis. Secondly, commercial cICAT with a cleavable linker also exist in two variants: photo-cleavable and acid cleavable. However, cICAT cleavage occurs in two steps and after elution from the avidin column whereas in this method cleavage and elution from the column occur simultaneously and this facilitates and fastens this procedure. Thirdly, cost of sulfo-NHS-SS-biotin and [12CD]- or [13CD3]-methyliodide is incomparably lower than the cost of cICAT. Fourthly, this method targets amino acid residues of lysine and free unmodified terminal amino groups, whereas ICAT modifies cysteine, which occurs rarer in proteins. Lysine labeling may cover larger percent of investigated peptides and consequently may increase probability of protein identification. 2.5. Solid Phase Isotope Tagging (Sit; Leucine) Zhou and Ranish proposed original approach in isotope labeling [47]. They proposed so called site specific isotope labeling of cysteine residues in complex peptide mixtures known as SIT (Solid-phase Isotope Tagging). Figure 6 shows that a o-nitrobenzyl photolabile linker is applied onto aminopropyl-coated glad solid phase. Using a technology applied in the solid phase peptide synthesis leucine containing 7 atoms of hydrogen or deuterium ([D0]-Leu or deuterium [D7]-Leu) is
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
attached to the photolabile linker. Sulfhydryl-specific iodoacetate group is attached to N-terminal leucine. After proteolytic cleavage cysteine-containing peptides of the analyzed samples reacted with iodoacetate group, bound to the solid phase label and then this complex was subjected to mild photo-treatment in near ultraviolet light at 360 nm for minimization of side photo-catalyzed reactions. This resulted in release of isotope containing leucine tail from the solid phase (this tail remains bound to cysteine residues of the analyzed peptide). Experiment with laminin B containing cystein residues and phosphoangiotensin I was carried out in parallel with ICAT. Photoincubation for 1 h cause a shift in intensity of laminin MS signal by 170 kDa corresponding to leucine containing modified group. This was also confirmed by MS/MS analysis. Signal intensity of modified and unmodified laminin B compared with corresponding data for phosphoangiotensin I suggests high specificity of the isotope reagent. The fact of retention of modified laminin B elution compared with the unmodified form (due to hydrophobic leucine nature) also represents indirect evidence indicating success of the labeling reaction. It was also noted that longer period of the photoreaction did not influence release of laminin B. Comparative experiments on proteome of Saccharomices cerevisiae yeasts have shown that the solid phase method demonstrated better results (than cICAT) by such parameter as the number of identified proteins and their quantitative evaluation. For example, using SIT it was possible to determine several proteins involved into galactose cleavage (GAL1, GAL2, GAL7, GALX), whereas employment of the cICAT-label resulted in determination of only one protein, GAL1 [47]. SIT not only has some advantages, but it is more simple, sensitive, effective and reproducible method. Firstly, isolation of cysteine containing peptides and their chemical modification are achieved essentially in a single step. Secondly, covalent capture of cysteine containing peptides to the solid phase (with the isotope tag) permits the use of stringent wash condition to remove non-covalently associated molecules and thus facilitates subsequent purification and elution. Thirdly, the presence of proteolytic enzymes and detergents does not influence reaction rate and therefore there is no need for additional steps to remove such molecules. Thus the solid phase method requires less manual input. It is suitable for studies of low abundance proteins, which may by inducible by some biological processes. Fourthly, this the method may use any other amino acids instead of the couple [D0]-Leu/[D7]-Leu; this could facilitate synthesis of many beds with a range of isotope tags. Fifthly, the mass tag weighs just 170 Da after its binding to cysteine (for [D0]-Leu) and in contrast to ICAT MS/MS spectrum lacks signals of side products of undesirable fragmentation of the label itself. Finally, before photocleavage the covalently
35 X3C
O2N
O
O N H
CH2 O
H N
H3C
CX3 CX
N H
CH2I
OMe O
Fig. 6. The structure of molecule containing the SIT label. A o-nitrobenzyl photolabile linker is applied onto aminopropyl-coated glad solid phase. Leucine contains 7 atoms of hydrogen or deuterium. The sulfhydryl-specific iodoacetate group (X= H or D) is attached to amino group of leucine.
immobilized peptides provide ideal substrates for subsequent chemical and enzymatic reactions if necessary [47]. 3. ENZYMATIC C-TERMINAL (16O/18O) LABELING OF PROTEINS Use of 18O as the label, is known since the work by Spirson and Rittenberg [49]. Studying inhibition of chymotrypsin-catalyzed amide bond hydrolysis by reaction products they found that incubation of carbobenzoxyphenylalanine with chymotrypsin in H218O resulted in incorporation of 18O into carboxy group of carbobenzoxyphenylalanine. Employing 18O/16O yielded chymotryptic peptides labeled by C-end irrespectively to their amino acid sequence. Desidero and Kai [50] originally used this property in quantitative proteomics. They enzymatically labeled C-terminal residues of leucine- and methionine-enkephalin with 18C and used them as an internal standard for quantitative determination of enkephalins isolated from mouse thalamus. Mirgorodskaya et al. [51] also used proteolytic cleavage involving 18O for quantitative proteomics. After their report Stewars et al. [52] published detailed data on conditions of proteolytic cleavage involving 18O. The 18O/16O isotope labeling characterized by such important advantages as rather simple chemistry and strictly determined mass shift became a potent tool in quantitative proteomics. However, due to the small mass difference between isotopes (just 2 Da) employment of this approach requires availability of mass spectrometer of high resolution. The early period of the development of this method employed proteolytic cleavage of proteins in the presence of H218O [53]. However, later proteolysis was carried out in “usual” water and after lyophilization peptide mixtures were dissolved in H218O. In the presence
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
36
KOPYLOV, ZGODA
of trypsin two 18O atoms are incorporated into peptide C-end. Proteolytic 18O labeling may be subdivided into two steps [54]: the first step, trypsinolysis, is carried out in H216O and the second step involves peptide labeling by means H218O, which is carried after vacuum evaporation (of the reaction mixture obtained after the first step). Conditions for each step may be optimized independently. Although this method is quite popular in quantitative proteomics the reaction of 16O/18O isotope labeling has not been standardized yet and many protocols have been described for the enzymatic reaction [54–60]. They are mainly differed by pH, content of salts and organic solvents, use of trypsin-immobilized beads or trypsin solutions. It is known that proteolytic reaction with trypsin increases yield of reaction products in the presence of calcium salts [55]. Addition of acetonitrile or DMSO to the reaction mixture also has positive effect. Lopez-Ferre’s group found [55] that in weakly acidic conditions there is insignificant exchange 18O/16O, which could by further decreased by sample storage at –20°C in H218O after termination of reaction with formic acid. Additional complexity and inconvenience appearing during 18O labeling consist in variability in simultaneous incorporation of one or two atoms of heavy oxygen isotope in to one peptide. The same authors [55] reported that although conditions optimized for this particular method provided effective incorporation of oxygen isotope atoms they could not register simultaneous incorporation of two 18O atoms into peptides. However, others observed this phenomenon. For example, Ong and Mann [61], Julka and Regnier [62] reported about detection of variable incorporation of two oxygen atoms into C-terminal residues of peptides; this caused inconvenience and increased error degree during calculations of 18O/16O ratios. This phenomenon occurs due to different ways of incorporation of oxygen atoms: the first atom is replaced during enzymatic hydrolysis reaction, whereas the second one may be replaced in the result of cleavage and reduction of peptide bond. Under usual conditions for hydrolysis reaction such reaction is rather slow and this is the reason explaining variations in the number of peptides with one or two substitutions of light isotopes of oxygen for the heavy ones. Complexity of quantitative evaluation is quite clear and during calculations 3 peaks should be taken into consideration: (1) monoisotopic peak of 16Opeptides, (2) monoisotopic peak of peptides with one 18O; (3) monoisotopic peak with two 18O atoms. The picture is complicated by the fact that the m/z values of 18O and (16O+2 Da) ions will be the same as in the case 1 of 18O2 m/z which will be overlapped with (18O1+2 Da) and (16O+4 Da). Solution of this problem lies in the development of technology of single or double 18O atom labeling. Of course the first variant requires solution of the problem of inhibition of exchange reaction between 16O and 18O
atoms. Rao et al. [63] found, that employment of the Lys-C protease at pH > 9.5 was accompanied by incorporation of just one oxygen atom. However, the mechanism of this reaction still remains unclear at the moment. Maximal effectiveness employing the second mechanism (incorporation of two atoms of oxygen isotope) requires optimization of conditions for catalysis of the exchange reaction. According to Zang et al. [56] yield of peptides containing two atoms of the heavy isotope 18O was much higher provided that the reaction of trypsinolysis was carried out at pH 6.75 for 20 h. Staes et al. [60] reported about total substitution by two atoms of 18O during the overnight incubation at pH 4.5. The experimental data have shown that employment of acidic conditions almost totally blocks amidase activity of trypsin this enzyme still exhibits some carboxy-oxygen exchange activity. Recent data published by Hajkova et al. [64] indicate that trypsin and Lys-C exhibit pH optimums for the amidase and oxygen exchange reactions within pH 8–9 and 5–6, respectively. The other way for the increased effectiveness of 18O labeling consists in trypsin immobilization [58, 65, 66]. Advantages of this method consist in the following: the immobilized enzyme may regulate with high accuracy the molar ratio substrate/enzyme and therefore regulate yield of peptides labeled with two isotopes of oxygen atoms. The other advantage of trypsin immobilization may be attributed to washout of hydrolysis products and this minimizes probability of the exchange reaction between 18O and 16O. The difference between the single and double labeling with 18O is 4 Da (between 16O and 18O peptides). Achievement of double substitution will allow quantitative evaluation and analysis by monoisotopic 16O and 18O peptides and this will avoid errors in calculations. 4. SILAC ISOTOPE LABELING OF PROTEINS IN CELL CULTURE 4.1. SILAC with Isotope Labeled Amino Acids in the Culture SILAC (Stable Isotope Labeling with Amino acids in Cell culture) is a method for labeling by amino acids with stable isotopes in cell cultures (Fig. 7). The SILAC strategy is based on the phenomenon that mammalian cells cannot synthesize some irreplaceable amino acids. Thus supplementation of cultivation media with such amino acids containing isotopes 15N and 13C will be incorporated into each newly synthesized protein. Foster et al. [67] studied membrane raft proteins of HeLa cells grown in the DMEM medium supplemented with 10% calf serum. In one case the growth medium contained [D0]-Leu, whereas in the other case it contained [D3]-Leu (with three deuterium atoms).
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
Shao-En Ong et al. [68] used [D3]-Leu incorporation into cells of mouse myoblast culture for quantitative measurement of expression level during myoblast cell differentiation. Total incorporation of this label was achieved after five passages. Leucine is often selected because its abundance among other amino acids in protein molecules and also the isotope form of leucine distinguishes it from isoleucine. Schlosser et al. [69] used this method for search and identification of phosphorylated nuclear membrane proteins. The culture of HeLa cells was grown in the presence of the isotopes [13C6]-Arg and[13C6]-Lys and phosphorylation inductor, sodium pervanadate, in the medium. Selection of these amino acids is evident due structure of peptides obtained after trypsinolysis. Such proteolysis yielded peptides containing C-terminal amino acids labeled by 13C-isotope (and 12C-isotope in control). In some populations of control cells phosphorylation was also induced for comparison of phosphoprotein levels in subcultures. Phosphorylated forms of proteins were extracted from mixed lysates of cell cultures by means of immunoprecipitation with phosphotyrosine specific antibodies 4G10 and RC20. Proteins eluted from the immune complex were separated by 1D gel electrophoresis. The proteins extracted from the gel were analyzed by LC-MS/MS (reverse phase chromatography in combination with tandem mass spectrometry). The difference for phosphorylated and nonphosphorylated single charged peptides was 80 Da. For isotope labeled sample the mass difference of phosphorylated single charged peptides with the same number of phospho groups is 6 Da, whereas for a couple of peptides with identical amino acid sequence the mass difference for phosphorylated versus nonphosphorylated states is 80 Da. Peptides from cells labeled with 13C-amino acids and peptides from cells cultivated with labeled with 12C-amino acids represented pairs for calculation of relative quantities of phosphoproteins. In addition cell cultures grown for 12 h with isotope labeled amino acids were then treated with phosphotyrosine phosphatase inhibitors, sodium pervanadate and sodium molibdate. For quantitative evaluation only peptides characterized by significant difference of the isotope ratio were used (e.g., in the peptide pair of the following sequence KIFEYETQR this ratio was 3.5). Only in the case of emerin the ratio between heavy and light chains varied from 2.5 to 6.5; these wide variation is not typical for SILAC. The main result of this study was elucidation of the fact that emerin is a phosphoprotein, playing the key role in assembly of protein complexes in the nuclear membrane. Comparative experiments with mouse cell culture have shown that in mouse and man phosphorylation sites differ, but in most cases this difference represents just a shift in one amino acid residue towards C-terminus of the polypeptide chain. It should be noted that the study by Schlosser et al. is not the only example of employment of SILAC in
37
Leu-D0 culture
Leu-D3 culture
Purification and isolation of proteins
Pooling of fractions and proteolytic digestion
Quantitative evaluation Fig. 7. The scheme of experiment employing the SILAC technology.
quantitative phosphoproteomics. Olsen et al. [70] studied isotope labeling in HeLa cell culture for quantitative analysis and determination of phosphorylation sites of proteins from cells treated with epidermal growth factor. Researchers identified 6600 phosphorylation sites in 2244 proteins. One of the main results of this study was direct demonstration of the fact that many proteins (including ubiquitin ligases, guanine nucleotide exchange factors, and at least 46 different transcriptional regulators) have more than one site for phosphorylation. 4.2. SILAC with Isotope Atoms 14N/15N in the Medium Krijgsveld et al. [71] used 14N/15N isotope labeling of proteins of multicellular organisms, a nematode Caenorhabditis elegans and a common fruit fly Drosophila melanogaster. Authors achieved isotope labeling of basically all proteins of these organisms in vivo. It is clear that the main problem of metabolic labeling consists in label availability for metabolic pathways. In this case almost 100% effectiveness was achieved during a two-step experiment. Initially cultures of S. cerevisiae and E. coli cells were cultivated in the medium supplemented with 15N. Extract of proteins from lyzed cells was separated by means of 2D-electrophoresis and tested for isotope distribution by MALDI-TOF demonstrating more than 98% content of heavy isotopes. During the second step several generations of Caenorhabditis elegans and Drosophila melanogaster were fed with nutrient media supplemented with cultures of yeast and E. coli cells grown in the media with added isotopes.
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
38
KOPYLOV, ZGODA
In the case of Caenorhabditis elegans authors compared proteomes of a wild type strain grown in the medium containing isitope labeled bacterial cells and the glp-4 strain, in which germline cells stop proliferating at an early stage. Equal amount of individuals from both nematode strain were lyzed and the protein extract was separated by means of 2D-electrophoresis. For subsequent studies random protein spots were selected from the gel. Although equal amounts of each strain were mixed before protein extraction at the ratio close to 1 : 1 (but not equal to that) comparison of results required normalization. Experiment was repeated in the reversed order: heavy isotopes were used for labeling of the glp-4 nematode strain. Researchers compared quantitative distribution of 6 proteins related to reproductive system of nematodes. In both experiments authors used the same normalization factor and the same volume and quantitative ratios Comparison of results of both experiments has shown the mean ratio of 0.98. This indicates high reproducibility of experiment for all analyzed proteins. In both experiments (direct and reciprocal) researchers detected quantitative differences in content of the analyzed proteins; these differences varied from 1.4 to 7 times. For additional demonstration of advantages of this method authors also employed liquid chromatography in combination with tandem mass spectrometry (LC-MS/MS). In this experiment they demonstrated that the content of prohibitin (antiproliferative factor localized mainly in mitochondria) was 1.4 times higher in wild type strain s compared with the mutant one; this well correlated with data of MALDI-TOF and immunoblotting. 4.3. SILAC with Isotope Atoms 12C/13C in the Medium Everley et al. [72] used a pair of [12C]-Lys/[13C]-Lys labeled lysines in the cultures of PC3M and PC3M-LN for identification of proteins, which would play some role the development of prostate cancer. This is a first study in which isotope labeling has been used for subsequent comparison of two various cell cultures. It is clear that in various cell lines levels of expression of the same proteins will significantly differ. Using isotope labeling in cell culture authors identified about 1000 proteins and 444 of them were evaluated quantitatively. About 30% of the analyzed proteins have been evaluated by one peptide containing lysine residue and also by one additional peptide containing arginine residue. Several important conclusions may be drawn on the basis of the above-considered results: first, metabolic isotope labeling is really applicable to quantitative analysis; second, this method minimizes variations and errors of routine work; third (and the most important), there is no need in subsequent chemical derivatization after protein extraction; forth, there is not need in purification and isolation of peptides on a column; fifth,
SILAC labels about 50% of all tryptic peptides, whereas ICAT labels just 20%. It should be also noted that lysine is more frequent amino acid than cysteine in proteins and this opens possibility for employment of any isotope amino acid labels. In conclusion we may add the other evidence supporting SILAC: this method is highly reproducible. 5. THE METHOD OF QUANTITATIVE EVALUATION AT THE LEVEL OF MS/MS. iTRAQ (ISOBARIC TAGGING) iTRAQ (isobaric Tagging Reagents Amino-reactive Quantification) is an elegant method, allowing to carry out accurate quantitative simultaneous analysis of several (up to four) samples at the level of MS/MS. It has been developed by Ross et al. [26] in 2004. Authors synthesized a series of reagents identical by their structure but differing by mass during MS/MS analysis. Isobaric tags are linked to N-terminal site of a polypeptide chain or to terminal ε-amino group of lysine. The resultant labeled peptides are characterized by the same m/z values on MS spectra and identical elution time from a column, however, during MS/MS analysis they “signalize” via labels of different masses. Figure 8 shows the main characteristics of this tag: the molecule represents a signaling group of N-methyl piperazine with molecular mass ranged from 114 to 117 Da due to variation of the number of 13C and 15N isotopes in it. Carbonyl balance group ranged from 28 to 31 Da (due to different number of 18O atoms) links this complex with N-hydroxysuccinimide, which is bound to protein molecule by means of ether bond. The balance group maintains constant molecular mass of labels (in MS spectrum), containing different number of isotope atoms. Reporter and balance groups represent so called isobaric tail. Four samples are labeled by one of variants of the isobaric tag, mixed at various rations and peaks areas of m/z of 114.1, 115.1, 116.1, and 117.1 corresponding to ratio of labeled proteins are then analyzed. It should be noted that the region of m/z of 100 is convenient because it is less contaminated with side ions. The mass shift observed during MS/MS analysis is compensated by the balance in MS so that a nominal mass of each peptide remains unchanged in various samples. This significantly increases sensitivity of analysis. In the case of the use of an isobaric tag MS spectrum of peptide ions of an individual protein represents a sum of all MS spectra of peptide ions of this protein in all samples. Since they all have the same molecular mass no cleavage of precursor ion signals and complication of peptide ion patterns for analysis are not observed. The only shortcoming of this method consists in the following: for evaluation of differences in the level of protein expression longer time consumption required for treatment and analysis of MS/MS spectra than that of MS
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
spectra. Employment of this method also requires QTOF type mass spectrometers. Use of such reporter ions described in the considered study provides accurate analysis of data, peptide sequencing and it does not complicate a spectrum with additional peaks. The isobaric reagents described here contain N-methyl piperazine; its presence acquires useful properties. Cyclic amines as peptide derivatives simplify interpretation of MS/MS spectra [73, 74]. These tags exhibit similar behavior during ESI and MALDI ionization with the same tendency in formation of b- and y-ions, used during peptide sequencing. In conclusion, we may say that the method of incorporation of isobaric tags does not require additional steps. Using four variants of the tag it is possible to analyze quantitatively several samples simultaneously. Analysis of expression levels of proteins involves tandem mass spectrometry. During labeling all MS spectra of proteins labeled with isobaric tags of various masses are summarized and do not complicate the mass spectrum pattern, because nominal mass of peptides remains the same. The isobaric tags do not influence elution time of labeled peptides on a chromatographic column. 6. QUANTITATIVE MASS-SPECTROMETRY EVALUATION OF POSTTRANSLATIONAL MODIFIED PROTEINS Posttranslational modification is directed changes of structural and functional properties of a protein molecule by attaching various prosthetic groups, proteolysis or modification of one of several amino acid residues in it. Although this problem is very important study of such protein is rather difficult due to lack of well-developed methods and techniques in this field. Many posttranslational modifications have been found accidentally, during studies of individual protein molecules by means of standard methods of molecular biology [74, 75]. Introduction of such methods as Edman’s chemical degradation and mass spectrometry into research practice rendered an invaluable help in studies of protein modification. Regulation of numerous intracellular processes employs reversible protein phosphorylation. Study of such phosphorylated complexes and mapping of phosphorylation sites represents an important problem in cell biology and biomedicine. Protein phosphorylation degree is often evaluated by means of 32P. Analysis of phosphoproteins and phosphopeptides presenting in cells at very low concentrations requires the development of technology of isotope enrichment (to avoid the effect of ion suppression, determined by the presence of large quantities of nonphosphorylated forms of proteins and peptides). However, posttranslational modifications of proteins are not limited by phosphorylation. Oand N-glycosylation of proteins also plays an important role. Study and identification of such glycoproteins rep-
39
Isobaric tail 145 Da Reporting Reporting group group 114.1–117.1 Da
O O
N N
O
N
O
Balance group 28–31 Da m/z114(+1) m/z115(+2) m/z116(+3) m/z117(+4)
13C 13C 2 13C 15N 2 13C 15N 3
13C 18O 18O 13C
(+3) (+2) (+1) (0)
Fig. 8. The structure of the isobaric tag.
resent even more complicated problem due to heterogeneity and diversity of glycan structure. As in the case of phosphoproteins high concentrations of unmodified molecules significantly complicate analysis of low concentrations of modified proteins due to signal suppression. Additional difficulty comes from the that fact that the same (glyco)protein expressed in various cells may have different glycan structure due to tissue specificity of glycosylation. Attachment of glycosyl-phosphatidylinositol tails to proteins of external plasma membrane layer is not as frequent modification as phosphorylation or glycosylation. Nevertheless such modified form of proteins attracts much interest and may be isolated by means of selective cleavable of the glycosyl-phosphatidylinositol tail and extraction of the released protein. Enzymatic conjugation of ubiquitin with cell proteins also attracts much interest as one of the forms of intracellular modification of proteins. Peng et al. [29] studied such form of protein modification by substituting ubiquitin for a histidine tail followed by subsequent selective purification on a His-tag column and identification of purified proteins. Analysis of protein modifications usually includes comparison of experimental data with known amino acid sequences. Thus, this first step in this direction consists in accurate protein identification by means of antibodies or mass spectrometry. 2D-electrophoresis is also a good tool for study and recognition of modified proteins. For example, phosphorylation changes a charge of protein molecule and, consequently, its electrophoretic mobility. Each of these spots (on 2D-gels) may be then investigated. One of such experiments [76] revealed more than 10 modified forms of enolase. Intensity of spots on stained 2D-gels may provide information about ratio of modified and unmodified forms of proteins.
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
40
KOPYLOV, ZGODA
After identification of protein mapping of sites of its modification involves general scheme. Protein is subjected to proteolysis (because peptides represent a more convenient research object than intact protein molecules). The mixture of peptides is then analyzed by means MALDI-TOF or ion trap spectrometry. Mass spectrometry has been recently introduced into phosphoproteomics and statistics shows that 58% of phosphoprotein studies have been carried out in 2001–2003 [77]. For mapping of protein phosphorylation sites tandem mass spectrometry has originally been used in 1998 [78]. Besides CID (Collision Ion Dissociation) the other method known as END (Electron Capture Dissociation) [79] is now widely used for more detailed information of phosphorylation sites. This method was introduced nine years ago by McLafferty. This is a unique method in which multiply protonated peptide ions and proteins capture low thermal electrons (i.e. electrons with low energy of 1–12 eV). The capture of thermal electrons by a protonated peptide is exothermic process causing fragmentation of a peptide backbone. In contrast to CID the ECD method is more convenient because fragmentation involves the whole peptide backbone irrespectively to amino acid sequence but it does preserve nativity of peptide posttranslational modifying groups. Investigating peptide phosphorylation sites Syka et al. [79] employed slightly modified version of ECD, EOD (Electron Transfer Dissociation). This technology is based on ion-ion reaction between multiply protonated peptides and single-charge anions in a linear quadruple trap. In this section of our review we consider some forms of posttranslational modifications and difficulties of their studies. Quantitative evaluation of modified forms of proteins often employs the above-described methods. The major difference consists in their adaptation and focus on nonprotein groups, which appear after posttranslational modifications of proteins. 6.1. Phosphoproteomics 6.1.1. IMAC (Immobilized Metal Affinity Chromatography). Study of protein phosphorylation process and mapping of such sites is an important part of cell biology studies on intracellular signaling. Huang et al. [80] quantitatively analyzed phosphoproteins my means of the method of IMAC (Immobilized Metal Affinity Chromatography), representing combination of dimethyl labeling with metal affinity chromatography. Alpha and beta-caseins served as model substrates. Isotope dimethyl labeling has been chosen by several reasons: first, it occurs rapidly; second, it involves peptide labeling by phosphorylated amino acid residues, third, such isotope labeling does not influence retention time of such peptides on a column, this means lack of isotope effect.
Protein extract was obtained from pregnant rat uteri. Proteins of tissue homogenate were reduced by cysteine residues followed by alkylation in denaturation buffer. Modified proteins were purified on an IMACcolumn with immobilized Fe(III) ions and then were subjected to trypsinolysis. Peptides were then mixed with 4% [D0]- or [D2]-formaldehyde and then with cyanoborohydride. The labeled peptides were mixed at various ratios: 1 : 1, 1 : 2, 1 : 5, and 1 : 10. The isotopelabeled peptides were purified on an IMAC-column equilibrated with 0.1% acetic acid. Peptides were eluted with acetonitrile and ammonium hydroxide in acetic acid (pH 9.5) and analyzed by means of MALDITOF or ESI-LC-MS/MS. Other researchers [81] used gallium instead of iron on the IMAC column, because gallium ions exhibited higher selectivity to phosphopeptides than IMAC based on iron or aluminum ions. Percent of eluted nonphosphorylated peptides represented just 2% of phosphorylated peptide components eluted from this column. After elution the sample may be directly analyzed by means of MALDI or electrostatic spray. The IMAC method was especially convenient for isolation of phosphoproteins. MALDI analysis did not revealed phosphorylated forms among proteins, which would not pass the IMAC stage of purification. The presence of phosphorylated peptides was confirmed by analysis of spectra of the same peptides pretreated with alkaline phosphatase [80]. In summary, the method of dimethyl labeling in combination with IMAC well fits to phosphoprotein studies. It is possible that some modifications will be included into this process to decrease possible nonspecific bindings: for example, combination of isotope dimethyl labeling with esterification or use of Glu-C instead of trypsin (to reduce number of carboxyl groups in a single peptide) [80–82]. One of shortcomings of this method was binding of not only of negatively charged phosphate groups but also all acidic peptides to the immobilized metal ions. This problem was solved by Ficarro et al. [82], these authors proposed to convert carboxyl groups into methyl ethers. 6.1.2. Isotope free method. Multiprotease approach. For identification of phosphorylation sites in proteins of the inner nuclear membrane Schlosser et al. [69] used two approaches: the multiprotease approach, in which proteins of subcellular fractions of mouse N2a were pooled, separated electrophoretically and subjected to hydrolytic cleavage by various proteases, and also SILAC, in which heavy and light isotope forms of lysine and arginine were added to the HeLa cell cultivation medium. In both cases plasma membrane permeable, highly specific inhibitor of tyrosine phosphatase BiPy was also used. The phosphatase inhibitor and protein kinase inhibitor, staurosporin were added before cell homogenization together with sodium vanadate and molibdate. Thus, inhibitors of tyrosine phosphatase and protein kinase would pre-
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
vent possible changes of phosphorylated proteins after cell homogenization. Homogenate of control and experimental samples were separated in parallel using two gels, one of these gels was then transferred onto a nitrocellulose membrane followed by subsequent immunostaining by phosphotyrosine specific antibodies PY99. However, staining of both samples revealed insignificant changes of phosphotyrosine proteins in both samples. The gel region corresponding to highest intensity of immunoblot staining was excised and proteins were subjected to ingel proteolysis by various proteases: trypsin, elastase, proteinase K, and thermolysin. All four mixtures of peptides were enriched by using a nanocolumn with titan spheres and analyzed by means of nanoLC-MS/MS. In both sample (in the sample with highest phosphotyrosine activity and in control) four proteins were identified. The first one contained an isoform of LAP 2-gamme and emerin, whereas the second sample contained nucleophosmin-1 and cation-dependent receptor of mannose-6-phosphate, and emerin. Although there were proteins corresponding to higher intensity of the immunoblot staining no phosphotyrosine peptides were identified in the first sample. Phosphotyrosine residues were identified only in emerin. Authors proposed the following interpretation of this phenomenon: it is possible that initial mixture contained small undetectable number of phosphotyrosine peptides (compared with other proteins). It was also possible that the phosphotyrosine antibodies exhibited nonspecific reactions. Finally, sometimes use of column with titan spheres possessing good affinity to phosphopeptides is accompanied by co-elution of nonphosphorylated peptides with low pI. 6.2. Glycoproteomics Although there is clear understanding of glycosylation process itself and its importance insignificant progress in glycoproteomics has been achieved. Situation is complicated by huge varieties of glycan forms, their high diversity, and also tissue specificity of glycosylation process. Structural studies of mammalian glycoconjugates mainly employ histochemical methods and monoclonal antibodies [83, 84]. Using such approach it is possible to get information about the presence or absence of glycoconjugates in the cell, however, employment of these methods cannot give provide any information about chemical nature of such conjugates (e.g., whether this conjugate is a glycolipid or N- or O-glycoprotein); such methods also cannot give any information about structure of glycan itself. Nature of glycan and glycosylation site is often studied by means of the method known as IGOT (Isotope coded Glycosylation-site-specific Tagging). This method is based on hydrolysis of glycopeptides by peptidyl-N-glycosidase F in the presence of H218O; this results in incorporation of 18O isotope into each N-glycosylated site of peptides. Situation with O-glycopep-
41
tides is even more complex because of lack of universal glycosidase for glycans with such type of chemical bond. Now lectins are widely used in glycoproteomics. Most of lectins are glycoproteins. In dependence of biological source phytolectins, mycolectins and zoolectins are recognized. These complex molecules may contain up to 20 proteins and often Ca2+ and Mn2+ ions. Lectins exhibit unique property of recognition and binding to carbohydrate components: glucose, mannose, sucrose etc. This property of selective binding to glycoproteins and polysaccharides represents a basis for use of lectin in affinity chromatography. Lectin affinity chromatography is usually carried out in two steps: the first step includes isolation of native glycoproteins (which are then subjected to trypsonolysis) and isolation of peptides containing a glycan group during the second affinity chromatography. This type of chromatography is used in glycoproteomics for elucidation of enantiomer, diastereomer, anomer, and type of bond (1-3/1-4) researchers work. 6.2.1. Derivatization and isotope labeling with [D3]-arginine. Using mouse derma and epidermis as model objects Uematsu et al. [85] proposed own approach in glycoproteomics for qualitative and quantitative characterization of N-glycosylation of tissue proteins. Authors developed a new technique of isotope glycoderivatization. They synthesized new isotope tags, which are covalently bound to proteins and are readily ionized. These include N-((Boc-aminooxy)acetyl)tryptophanylarginine methyl ester, N-((aminooxy)acetyl)tryptophanylarginine methyl ester (aoWR(H)), and [D3]-arginine methyl ester hydrochloride (aoWR(D3)). The epidermis tissue samples were defatted, lyophilized, reduced (by cystaine residues) and S-carmaboyl methylated under denaturing conditions. Deglycosylation involved peptidyl-N-glycosidase F. The lyophilized peptide mixture was dissolved in an aliquot of aoWR(H) and aoWR(D3) and incubated at 90°C for one hour. In comparative experiments the same samples were derivatized by 2-aminopurine as it was earlier proposed in similar studies [86–88]. Each sample derivatized by aoWR(H) and aoWR(D3) was analyzed by means of MALDI-TOF. Importance of incorporation of aoWR(H) and aoWR(D3) employment of N-terminus of peptide for insertion of aminooxy functional group, whereas C-terminus serves for formation of deuterated analogues of methyl esters. MALDI-TOF/TOF analysis of aoWR(H) and aoWR(D3)-derivatives of oligosaccharides studied gives a signal of high intensity at the level of m/z 431 and 434, corresponding to the cleavage of the N-O bond in glycopeptides. The mass difference of aoWR(H) and aoWR(D3)-derivatives of oligosaccharides is 3Da; this provides effective relative quantitative measurement of N-glycan distribution. Glycoproteins with mannose oligosaccharides attracted special attention of researchers. These glycoproteins were purified from the total mix-
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
42
KOPYLOV, ZGODA
ture of proteins by concanavalin A, exhibiting affinity to mannose. The concanavalin A-bound tryptic peptides were then isolated by reverse phase chromatography. For validation of data samples pretreated with peptidyl-N-glycosidase F were also analyzed. For example, there was characteristic distribution of signals with difference of 162 Da (hexose) observed on spectra of the same peptide bound to structurally different glycans containing mannose. After enzymatic hydrolysis, there was only one signal corresponding to mass of amino acid chain of this peptide. Authors also found that glycoproteins with different forms of glycan attached to the same amino acid backbone were characterized by different retention time of column and elution during reverse phase chromatography. Peptides modified by sterically large N-glycans were eluted earlier than those with smaller N-glycans. Researchers found 15 glycoproteins in total; four of them were transmembrane proteins, other had lysosomal or extracellular matrix origin. In conclusion we should underline the main features of this method. First, these isotope labels allow carry out rapid and highly sensitive quantitative measurements of proteome glycoproteins. Second, due to high ionization these chemical derivatives may be selectively detected with minimal number of steps required for addition purification. Third, effectiveness of the method of derivatization allows to compare protein profiles from different tissue samples (derma and epidermis) and from the same sample but from different subcellular organelles. 6.2.2. Solid phase isotope labeling (SIT). Principles of solid phase isotope labeling are applicable for quantitative studies in glycoproteomics. Hui Zhang et al. [48] described a method of quantitative measurements of N-glycoproteins. It is based on covalent crosslinking of proteins with the solid phase by means of hydrazine, isotope labeling of immobilized glycoproteins, their subsequent cleavage into peptides followed by enzymatic detachment from the solid phase catalyzed by peptidyl-N-glycosidase F. The method consists of six steps: glycoproteins are initially treated with periodate, causing conversion of cis-diols of a carbohydrate group into aldehydes; this is important precondition for subsequent reaction with hydrazines, immobilized onto a solid phase. This results in formation of covalent hydrazine bond and proteins becomes spatially fixed, whereas unbound nonglycosylated proteins are removed during wash at the second step. During the third step glycoproteins immobilized on the solid phase are treated with peptidyl-N-glycosidase F and then amino groups of lysine side chain are labeled with [D0]- or [D4]-succinic anhydride. After the isotope labeling N-glycopeptides are eluted from the solid phase by means of peptidyl-Nglycosidase F. The resultant samples are analyzed by LC-MS/MS or MALDI.
This method may be used for identification of glycoproteins, their quantitative analysis and also for determination of glycosylation sites in proteins. Due to its selectivity this method may be used for analysis of cell membranes, tissue fluids and secreted proteins, which are significantly enriched with glycans. Employment of this method may also simplify an analyzed protein mixture, because proteins usually contain not many glycosylation sites. It should be also noted that after minor changes this method is applicable for analysis of O-glycoproteins. 6.2.3. Isotope free method. Study of O-sulfated forms of serine and threonine. Sulfation is the other type of posttranslational modifications, attracting attention of researchers in proteomics. This type of protein modification has been found in all eukaryote species from protozoa to man. Sulfation is a widespread process of enzymatic modification of proteins, which involves both protein and carbohydrate components, and also xenobiotics [89]. This process employs 3'phosphoadenosine-5'-phosphosulfate (PAPS), the only donor of sulfate group [90], and reaction results in formation of several types of bonds, ester (O-sulfation), amide (N-sulfation), and thioester (S-sulfation) bonds. O-Sulfation, transfer of sulfate group onto hydroxyl group or phenolic acceptor prevails in the cell. The major proportion of cell sulfation involves polysaccharides, steroids, catecholamines, and thyroid hormones. Many sulfo-derivatives of proteins cannot regulate expression of own genes. Half-life time of sulfated derivatives is shorted than other derivatives. Medzihradszky et al. [91] investigated such type of modifications. Detection and characterization of such type of posttranslational modification were investigated by means of liquid chromatography followed by subsequent tandem mass spectrometry analysis. Authors described the structure of O-sulfated serine and threonine residues of proteins from evolutionary different classes of animals. These included neuronal intermediate filament and myosin light chain of the mollusk Lymnaea stagnalis, cathepsin C-like protein from the malaria parasite Plasmodium falciparum and human receptor tyrosine kinase Ror2. Using LC-ESI-CID-IS analysis (Liquid Chromatography Electro Spray Ionization Collision Induced Dissociation Mass Spectrometry) it was found the some components of peptide mixture obtained after trypsinolysis had the same molecular mass, but their retention time of the column varied from 1 to 5 min; also there were some differences in ion charge. Careful LCESI-CID-IS analysis revealed the presence of seven such peptides in one of intermediate filament samples. For detailed investigation authors compared profile of chromatographic elution of modified peptides and their unmodified analogues with data of LC-MS analysis. Based on these results they concluded that several modified analogues were isolated; they possessed identical amino acid sequences differing by 80 Da. These modi-
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
fied analogues were stable under conditions of ESI and mass loss was 80 Da; this is also consistent with the presence of sulfate group in peptides. The major result of this paper consists in the fact of identification of proteins sulfation (one of natural form of posttranslational modifications) by serine and threonine amino acid residues of proteins. The characteristics of sulfated aliphatic peptides differ from their phosphorylated analogues. Results suggest that sulfation of serine or threonine is more stable than sulfation of tyrosine. 6.2.4. Frontal affinity chromatography (FAC). This type of affinity chromatography was developed by Kasai and Ishii [92] about 30 years ago. FAC may be used for accurate determination of Kd of biomolecules. Indeed it was used in studies of antigens and antibodies, proteases and their substrate analogues, lectins and glycoproteins. The “old” technology of FAC was time consuming and required sufficient amount of ligand for immobilization onto a column and enough material for elution. However, some time ago Hindsgaul [93] overcame these difficulties by preparing a thin column (0.75 mm × 150 mm) with immobilized cholera toxin. The main innovation was coupling of this column to a mass spectrometer, using this innovation it was possible to analyze MS spectra of oligosaccharide derivatives in the real time mode. Such method was defined as FAC/MS. Hirabayashi et al. [94] modified this system and made it more accurate and specialized on lectin-oligosaccharide interaction. Authors used fluorescent detection (FD) and their method, known as FD/FAC, may detect more than 100 oligosaccharides during short time interval. The development of FD/FAC was necessary to improve effectiveness of existing systems for oligosaccharide analysis. Now FD/FAC employs in parallel two or more capsule type columns (2 × 10 mm); this allows to analyze more than 100 oligosaccharides during relatively short time interval. The method of pyridylamination earlier developed by Hase et al. [95] has certain shortcomings compared with FD/FAC. These include chemical instability of a fluorescent label, lower sensitivity, but this method is characterized by good separation in HPLC systems. 7. ABSOLUTE QUANTITATIVE ANALYSIS OF PROTEINS AND PHOSPHOPROTEINS BY THE METHOD OF TANDEM MASS SPECTROMETRY The method of absolute quantification (known as AQUA) its authors propose absolute qualitative analysis of proteins in proteome [96]. All the methods considered in this review may give relative quantitative evaluation of proteins due to limited technical capacities of mass spectrometry for quantitative evaluations. Gerber et al. [96] proposed rather simple (technically)
43
method of evaluation of absolute quantities of proteins and their modified forms, particularly, phosphoproteins. Authors employed this method for absolute quantitative evaluation of two regulatory proteins Sir2 and Sir4 (involved into histone deacetylation) from yeast cell culture, human separases Ser-1126 and 1501, and horse myoglobin. For realization of this method they synthesized peptides of the exact amino acid sequences identified by MS/MS spectra of peptides obtained from the analyzed proteins. These synthetic peptides used as internal standards contained 13C and 15N as the stable isotope labels. It is also possible to synthesize peptides with covalent modifications like phosphorylation, methylation, or acylation, which would correspond to posttranslational modifications found in cells. Such synthetic are then used as internal standards for accurate absolute evaluation and determination of content of proteins and their modified variants (subjected to proteolysis before analysis). During solid phase synthesis of peptides employed leucine containing six atoms of 13C and one 15N. Synthesis yielded chemically identical peptides; during MS analysis they were easily discriminated from the native ones by a 7 Da-mass shift. Synthetic and native peptides were also compared by such parameters as elution time from a column, effectiveness of ionization and degree of fragmentation. Before the beginning of analysis cell lysate containing from 300 aM to 30 pM myoglobin were separated by 1D SDS electrophoresis, protein bands of interest were excised and subjected to ingel proteolysis in the presence of 500 pM synthetic AQUA peptides followed by subsequent reverse phase chromatography. For cell lysate obtained from 1.8 × 107 cells trypsinolysis was carried out in the presence of 150 fM AQUA peptides. In the case of experiments with human separase Ser1126 trypsinolysis was carried out in the presence of two different synthetic peptides (100 and 150 pM) and in experiments with human separase Ser-1501 200 fM concentration of synthetic peptide was used. Elution behavior and fragmentation of synthetic peptides in MS/MS analysis were absolutely identical to those of native peptides from protein molecules studied. Absolute quantity of proteins studied may be then accurately determined using information on content of internal standards added to the reaction mixture. CONCLUSIONS Quantitative analysis employing tandem mass spectrometry is especially important in system biology studies. Many studies used this method for elucidation of protein-protein protein-peptide interactions. Mass spectrometry methods helped researches to solve such problems as time course of distribution of various forms of posttranslational modifications, their subcellu-
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
44
KOPYLOV, ZGODA
lar localization as well as processes of complex formation. Such fields of quantitative proteomics as glycoproteomics, phosphoproteomics take into consideration variations of protein molecules at the level of posttranslational modifications. Organelle proteomics has become one of the most important applications of quantitative proteomics. Quantitative information on certain protein molecules in various organelles in combination with their dynamic distribution in the cell and the whole organism, induced by cell response to various external factors, natural physiological circulation of proteins between organelles give principally new viewpoint on understanding of mechanisms of cell differentiation, signaling, transport, protein secretion and many other processes related to cell cycle. Mass spectrometry begins to dominate in studies of protein interactions (it even “displaces” a dihybrid system). In contrast to the dihybrid system using the proteomic methods it is possible to investigate various complexes of interacting molecules. However, many interactions are characterized by relatively low affinity mass spectrometry may detect false signals, simulating the existence of relationship between molecules. Such false-positive signal may be removed by means of cross-linkers stabilizing weak bonds with proteins. Quantitative proteomics based on mass spectrometry is a powerful tool for the development of biology, medicine, diagnostics, structural biology etc. However, one should realize that as any other method, mass spectrometry also has limited capacities. Quantitative evaluation of proteins characterized by low concentration in the cell is complicated by the presence of highly abundant proteins and it is not easy to overcome this problem. It is still difficult to carry out quantitative analysis of the whole proteome. Quantitative analysis of absolute content of some components still represents serious problem, but certain approaches have been proposed to overcome these problems. Thus, in spite of all these limitations quantitative proteomics is successfully used in basic and applied fields of science. ACKNOWLEDGMENTS This work was supported by a grant from Russian Foundation for Basic Research (no. 07-04-00803a). REFERENCES 1. Tonge, R., Shaw, J., Middleton, B, Rowlinson, R., Rayner, S., Young, J., Pognan, F., Hawkins, E., Currie, I., and Davison, M., Proteomics, 2001, vol. 1, pp. 377–396. 2. Patton, W.F., J. Chromatography B, 2002, vol. 771, pp. 3–31. 3. Gade, D., Thiermann, J., Markowsky, D., and Rabus, R.J., Mol. Microbiol. Biotechnol., 2003, vol. 5, pp. 240– 251.
4. Ottens, A.K., Kobeissy, F.H., Wolper, R.A., Haskins, W.E., Hayes, R.L., Denslow, N.D., and Wang, K.K.W., Anal. Chem., 2005, vol. 5, pp. 4836–4845. 5. Unlu, M., Morgan, E., and Minden, J.S., Electrophoresis, 1997, vol. 18, pp. 2071–2077. 6. Freeman, W.M., and Hemby, S.E., Neurochem. Res., 2004, vol. 29, pp. 1065–1089. 7. Alban, A., David, S.O., Bjorkesten, L., Andersson, C., Sloge, E., Lewis, S., and Currie, I., Proteomics, 2003, vol. 3, pp. 36–44. 8. Rabilloud, T., Proteomics, 2002, vol. 2, pp. 3–10. 9. Hoffmann, P., Ji, H., Mortiz, R.L., Connoly, L.M., Frecklington, D.F., Layton, M.J., Eddes, J.S., and Simpson, R.J., Proteomics, 2001, vol. 1, pp. 807–818. 10. Lubman, D.M., Kachman, M.T., Wang, H., Gong, S., Yan, F., Hamler, R.L., O’Neil, K.A., Zhu, K., Buchanan, N.S., and Barder, T.J., J. Chromatogr. B, 2002, vol. 183–196. 11. Hamler, R.L., Zhu, K., Buchanani, N.S., Kreunin, P., Kachman, M.T., Miller, F.R., and Lubman, D.M., Proteomics, 2004, vol. 4, pp. 562–577. 12. Wang, H.X., Kachman, M.T., Schwatrz, D.R., Cho, K.R., and Lubman, D.M., Proteomics, 2004, vol. 4, pp. 2476– 2495. 13. Opitek, G.J., Ramirez, S.M., Jorgenson, J.W., and Moseley, M.A., Anal. Biochem., 1998, vol. 258, pp. 349–361. 14. Chong, B.E., Yan, F., Lubman, D.M., and Miller, F.R., Rapid Commun. Mass Spectrom., 2001, vol. 15, pp. 291– 296. 15. Zhang, Z.L., Smith, D.L., and Smith, J.B., Proteomics, 2001, vol. 1, pp. 1001–1009. 16. Lubman, D.M., Kachman, M.T., Wang, H., Gong, S., Yan, F., Hamler, R.L., O’Neil, K.A., Zhu, K., Buchanan, N.S., and Barder, T.J., J. Chromatogr. B, 2002, vol. 183–196. 17. Zhu, K., Kachman, M.T., Miller, F.R., Lubman, D.M., and Zhang, R., J. Chromatogr. A, 2004, vol. 1053, pp. 133–142. 18. Feng, B.B., Patel, A.H., Keller, P.M., and Slemmon, J.R., Rapid Commun. Mass Spectrom., 2001, vol. 15, pp. 821– 826. 19. Krapfenbauer, K., Fountoulakis, M., and Lubec, G., Electrophoresis, 2003, vol. 24, pp. 1847–1870. 20. El Rassi, Z., and Horvath, C., J. Chromatogr. B, 1986, vol. 359, pp. 255–264. 21. Gerber, S.A., Rush, J., Stemman, O., Kirschner, M.W., and Gygi, S.P., Proc. Natl. Acad. Sci. USA, 2003, vol. 100, no. 12, pp. 6940–6945. 22. Oda, Y., Huang, K., Cross, F.R., Cowburn, D., and Chait, B.T., Proc. Natl. Acad. Sci. USA, 1999, vol. 96, pp. 6591–6596. 23. Gygi, S.P., Rist, B., Gerber, S.A., Turecek, F., Gelb, M.H., and Aebersold, R., Nat. Biotechnol., 1999, vol. 17, pp. 994–999. 24. Wang, S. and Regnier, F.E., J. Chromatogr. A, 2001, vol. 924, pp. 345–357. 25. Flory, M.R., Griffin, T.J., Martin, D., and Aebersold, R., Trends in Biotechnology, 2002, vol. 20, no. 12, pp. S23– S29. 26. Ross, P.L., Huang, Y.N., Marcheset, J.N., Williamson, B., Parker, K., Hattan, S., Khainivski, N., Pillai, S., and Sub-
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
THE METHODS OF QUANTITATIVE PROTEOMICS
27.
28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.
43. 44. 45. 46. 47. 48. 49.
hakar, D., Mol. Cell. Proteomics, 2004, vol. 3, pp. 1154– 1169. Jenkins, R.E., Kitteringham, N.R., Hunter, C.L., Webb, S., Hunt, T.J., Elsby, R., Watson, R.B., Williams, D., Pennington, S.R., and Park, K.B., Proteomics, 2006, vol. 6, pp. 1934–1947. Aebersold, R. and Mann, M., Nature, 2003, vol. 422, pp. 198–207. Peng, J. and Gygi, S.P., J. Mass Spectrom., 2001, vol. 36, pp. 1083–1091. Leither, A. and Linder, W., Proteomics, 2006, vol. 6, pp. 5418–5434. Hansen, K.C., Schmitt-Ulms, G., Chakley, F, and Hirsch, J., J. Mol. Cell. Proteomics, 2003, vol. 2, pp. 299–314. Li, J., Steen, H., and Gygi, S.P., J. Mol. Cell. Proteomics, 2003, vol. 2, pp. 1198–1204; Oda, Y., Owa, T., Sato, T., and Boucher, B., Anal. Chem., 2003, vol. 75, pp. 2159–2165. Yu, L.-R., Conrads, T.P., Uo, T., and Isaaq, H.J., J. Proteome Res., 2004, vol. 3, pp. 469–477. Parker, K.C., Pattern, D., Williamson, B., and Marches, J., Mol. Cell. Proteomics, 2004, vol. 3, pp. 625–659. Molloy, M.P., Donohoe, S., Brzezinski, E.E., and Kilby, G.W., Proteomics, 2005, vol. 5, pp. 1204–1208. Sethuraman, M., McComb, M.E., Huanh, H., and Huang, S., J. Proteome Res., 2004, vol. 3, pp. 1228– 1233. Sethuraman, M., McComb, M.E., Heibeck, T., Costello, C.E., and Cohen, R.A., Mol. Cell. Proteomics, 2004, vol. 3, pp. 273–278. Dunkley, T.P.J., Watson, R., Griffin, J.L., Dupree, P., and Lilley, K.S., Mol. Cell. Proteomics, 2004, vol. 3, pp. 1128–1134. Li, K.W., Hornshaw, M.P., van Minnen, J., and Smalla, K.-H., Proteome Res., 2005, vol. 4, pp. 725– 733. Jiang, X.-S., Dai, J, Sheng, Q.-H., and Zhang, L., Mol. Cell. Proteomics, 2005, vol. 4, pp. 12–34. Ranish, J.A., Yi, E.C., Leslie, D.M., Purvine, S.O., Goodlet, D.R., Eng, J., and Aebersold, R., Nature Genetics, 18 February published on line; doi:10.1038/ng1101, 2003. Shen, M., Guo, L., Wallace, A., Fitzner, J., Eisenman, J., Jacobson, E., and Jhonson, R.S., Mol. Cell. Proteomics, 2003, vol. 2, pp. 315–324. Wang, S. and Regnier, F.E., J. Chromatogr. A, 2001, vol. 924, pp. 345–357. Warwood, S., Mohammed, S., Cristea, I.M., Evans, C., Whetton, A.D., and Gaskell, S.J., Rapid Comm. Mass Spectrom., 2006, vol. 20, pp. 3245–3256. Van Hoang, M., Conrads, T.M., Veestra, T.D., Bonder, J., Terunuma, A., Vogel, J.C., and Fisher, R.J., J. Biomolec. Technol., 2003, vol. 14, pp. 216–223. Zhou, H., Ranish, J.A., Watts, J.D., and Aebersold, R., Nat. Biotechnol., 2002, vol. 19, pp. 512–515. Zhang, H., Li, X.-J., Martin, D.B., and Aebersold, R., Nat. Biotechnol., 2003, vol. 21, pp. 660–666. Spirson, D.B. and Rittenberg, D., Nature, 1951, vol. 167, pp. 484–484.
45
50. Desidero, D.M. and Kai, M., Biomed. Mass Spectrum., 1983, vol. 10, pp. 471–479. 51. Mirgorodskaya, O.A., Kozmin, Y.P., Titov, M.I., Korner, R., Sonksen, C.P., and Roepstoroff, P., Rapid Comm. Mass Spectrom., 2000, vol. 14, pp. 1226–1232. 52. Stewart, I.I., Thomson, T., and Figeys, D., Rapid Comm. Mass Spectrom., 2001, vol. 15, pp. 2456–2465. 53. Yao, X., Freas, A., Ramirez, J, Dmitriev, P.A., and Fenselau, C., Anal. Chem., 2001, vol. 73, pp. 2836–2842. 54. Yao, X., Afonso, C., and Fenselau, C., J. Proteome Res., 2003, vol. 2, pp. 147–152. 55. Lopez-Ferrer, D., Ramos-Fernandes, A., Martinez-Bartolome, S., Garcia-Ruiz, P., and Vaskez, J., Proteomics, 2006, vol. 6, pp. S4–S11. 56. Zang, L., Palmer, T.D., Hancock, W.S., Sgroi, D.C., and Karger, B.L., J. Proteome Res., 2004, vol. 3, pp. 604– 612. 57. Heller, M., Mattou, H., Menzel, C., and Yao, X., J. Am. Soc. Mass Spectrom., 2003, vol. 14, pp. 704–718. 58. Qian, W.J., Monroe, M.E., Liu, T., and Jacobs, J.M., Mol. Cell. Proteomics, 2005, vol. 4, pp. 700–709. 59. Bantscheff, M., Dumpelfeld, B., and Kuster, B., Rapid Comm. Mass Spectrum., 2004, vol. 18, pp. 869–876. 60. Staes, A., Demol, H., Van Damm, J., and Martens, L., J. Proteome Res., 2004, vol. 3, pp. 786–791. 61. Ong, S.-E. and Mann, M., Nature Chem. Biology, 2005, vol. 1, 5, pp. 252–262. 62. Julka, S., and Regnier, F.E., J. Proteome Res., 2004, vol. 3, pp. 350–363. 63. Rao, K.S.C., Carruth, R.T., and Miyagi, M., J. Proteome Res., 2005, vol. 4, pp. 507–514. 64. Hajkova, D., Rao, K.S.C., and Miyagi, M., J. Proteome Res., 2006, vol. 5, pp. 1667–1673. 65. Brown, K.J. and Fenselau, C., J. Proteome Res., 2004, vol. 3, pp. 455–462. 66. Chen, X., Cushman, S.W., Panel, L.K., and Hess, S., J. Proteome Res., 2005, vol. 4, pp. 570–577. 67. Foster, L.J., de Hoog, C.L., and Mann, M., Proc. Natl. Acad. Sci. USA, 2003, vol. 100, pp. 5813–5818. 68. Ong, S.-E., Blagoev, B., Kratchmarova, I., Kristensen, D.B., Steen, H., Pandey, A., and Mann, M., Mol. Cell. Proteomics, 2002, vol. 1, pp. 376–386. 69. Schlosser, A., Amanchy, R., and Otto, H., FEBS, 2006, vol. 273, pp. 3204–3215. 70. Olsen, J.V., Blagoev, B., Gnad, F, Macek, B., Kumar, C., and Mortensen, P., and Mann M., Cell, 2006, vol. 127, pp. 635–648. 71. Krijgsveld, J., Ketting, R.F., Mahmoudi, T., Johansen, J., Artal-Sanz, M., Verrijzer, C.P., Plasterk, R.H.A., and Heck, A.J., Nat. Biotechnol., 2003, vol. 21, pp. 927–931. 72. Everley, P.A., Krijgsveld, J., Zetter, B.R., and Gygi, S.P., Mol. Cell. Proteomics, 2004, vol. 3, pp. 729–735. 73. Karimi-Busheri, F., Daly, G., Robins, P., Canas, B., Pappin, D.J., Sgouros, J., Miller, G.G., Fakhrai, H., Davis, E.M., Le Beau, M.M., and Weinfeld, M., J. Biol. Chem., 1999, vol. 274, pp. 24187–24194. 74. Hoss, M., Robins, P., Naven, T.J.P., Pappin, D.J.C., Sgorous, J., and Lindahl, D.A., EMBO J., 1999, vol. 18, pp. 3868–3875.
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008
46
KOPYLOV, ZGODA
75. Mann, M., and Jensen, O.N., Nat. Biotechnol., 2003, vol. 21, pp. 255–261. 76. Larsen, M.R., Larsen, P.M., Fay, S.J., and Roepstoff, P., Electrophoresis, 2001, vol. 22, pp. 566–575. 77. Loyet, K.M., Stults, J.T., and Arnott, D., Mol. Cell. Proteomics, 2005, vol. 4, pp. 235–245. 78. Zubarev, R.A., Kelleher, N.L., and McLafferty, F.W., J. Am. Chem. Soc., 1998, vol. 120, pp. 3265–3266. 79. Syka, J.E., Coon, J.J., Schroeder, M.J., Shabanowitz, J., and Hunt, D.F., Proc. Natl. Acad. Sci. USA, 2004, vol. 101, pp. 9528–9533. 80. Huang, S.-Y., Tsai, M.-L., Wu, C.-J., Hsu, J.-L., Ho, S.-H., and Chen, S.-H., Proteomics, 2006, vol. 6, pp. 1722– 1734. 81. Posewitz, M.C. and Tempst, P., Anal. Chem, 1999, vol. 71, pp. 2883–2892. 82. Ficarro, S.B., McCleland, M.L., Stukenberg, P.T., Burke, D.J., Ross, M.M., Shabanowitz, J., Hunt, D.F., and White, F.M., Nat. Biotechnol., 2002, vol. 20, pp. 301–305. 83. Huang, C.M., Foster, K.W., DeSilva, D., Zhang, J., Zhi, Z., Yusuf, N., Van Kampen, K.R., Elmets, C.A., and Tang, D.C., J. Investig. Dermatol., 2003, vol. 121, pp. 51–64. 84. Blonder, J., Terunuma, A., Conrads, T.P., Chan, K.C., Yee, C., Lucas, D.A., Schaefer, C.F., Yu, L.R., Issaq, H.J., Veenstra, T.D., and Vogel, J.C., J. Investig. Dermatol., 2004, vol. 123, pp. 691–699. 85. Uematsu, R., Furukawa, J.-I., Nakagawa, H., Shinohara, Y., Deguchi, K., Monde, K., and Nishimura, S.-I., Mol. Cell. Proteomics, 2005, vol. 4, pp. 1977–1989.
86. Nakagawa, H., Kawamura, Y., Kato, K., Shimada, I., Arata, Y., and Takahashi, N., Anal. Biochem., 1995, vol. 226, pp. 130–138. 87. Takahashi, N., Nakagawa, H., Fujikawa, K., Kawamura, Y., and Tomiya, N., Anal. Biochem., 1995, vol. 226, pp. 139–146. 88. Yamamoto, S., Hase, S., Fukuda, S., Sano, O., and Ikenaka, T., J. Biochem., 1989, vol. 105, pp. 547–555. 89. Strott, C.A., Endocr. Rev., 2002, vol. 23, pp. 703–732. 90. Robbins, P. and Lippman, F., J. Am. Chem. Soc., 1956, vol. 78, pp. 2652–2653. 91. Medzihradszky, K.F., Darula, Z., Perlson, E., Fainzilber, R.J., Chakley, R.J., Ball, H., Greenbaum, D., Bogyo, M., Tyson, D.R., Bradshow, R.A., and Burlingame, A.L., Mol. Cell. Proteomics, 2004, vol. 3, pp. 429–443. 92. Kasai, K., and Ishii, S., J. Biochem. (Tokyo), 1975, vol. 77, pp. 261–264. 93. Zhang, B., Palcic, M.M., Schriemer, D.C., AlvarezManilla, G., Pierce, M., and Hindsgaul, O., Anal. Biochem., 2001, vol. 299, pp. 173–182. 94. Hirabayashi, J., Arata, Y., and Kasai, K., J. Chromatogr. A, 2000, vol. 890, pp. 261–271. 95. Hase, S., Ibuki, T., and Ikenuka, T., J. Biochem. (Tokyo), 1984, vol. 95, pp. 197–203. 96. Gerber, S.A., Rush, J., Stemmen, O., Kirschner, M.W., and Gygi, S.P., Proc. Natl. Acad. Sci. USA, 2003, vol. 100, pp. 6940–6945.
BIOCHEMISTRY (MOSCOW) SUPPLEMENT SERIES B: BIOMEDICAL CHEMISTRY
Vol. 2
No. 1
2008