Proteome Analysis of Intact Proteins in Complex Mixtures

Abstract

Analysis of intact protein mixtures by electrospray ionization mass spectrometry requires the resolution of a complex, overlapping set of multiply charged envelopes. To ascertain the ability of a moderate resolution mass spectrometer to resolve such mixtures, we have analyzed the soluble proteins of adult chick skeletal muscle. This is a highly specialized tissue showing a marked bias in expression of glycolytic enzymes in the soluble fraction. SDS-PAGE-resolved proteins were first identified by a combination of matrix-assisted laser desorption ionization time-of-flight (TOF) and electrospray ionization tandem mass spectrometry. Then the mixture of intact proteins was introduced into the electrospray source of a Q-TOF mass spectrometer either by direct infusion or via a C4 desalting trap. In both instances, the complex pattern of peaks could be resolved into true masses, and these masses could in many instances be reconciled with the masses predicted from the known protein sequences when qualified by expected co- and post-translational modifications. These included loss of the N-terminal initiator methionine residue and N-terminal acetylation. The ability to resolve such a complex mixture of proteins with a routine instrument is of considerable value in analyses of protein expression and in the confirmation of post-translational changes in mature proteins. Analysis of intact protein mixtures by electrospray ionization mass spectrometry requires the resolution of a complex, overlapping set of multiply charged envelopes. To ascertain the ability of a moderate resolution mass spectrometer to resolve such mixtures, we have analyzed the soluble proteins of adult chick skeletal muscle. This is a highly specialized tissue showing a marked bias in expression of glycolytic enzymes in the soluble fraction. SDS-PAGE-resolved proteins were first identified by a combination of matrix-assisted laser desorption ionization time-of-flight (TOF) and electrospray ionization tandem mass spectrometry. Then the mixture of intact proteins was introduced into the electrospray source of a Q-TOF mass spectrometer either by direct infusion or via a C4 desalting trap. In both instances, the complex pattern of peaks could be resolved into true masses, and these masses could in many instances be reconciled with the masses predicted from the known protein sequences when qualified by expected co- and post-translational modifications. These included loss of the N-terminal initiator methionine residue and N-terminal acetylation. The ability to resolve such a complex mixture of proteins with a routine instrument is of considerable value in analyses of protein expression and in the confirmation of post-translational changes in mature proteins. A key step in any proteome analysis is the identification of one or more proteins that have been resolved by separation technologies, including one-dimensional and two-dimensional polyacrylamide gel electrophoresis or liquid chromatography. In most instances, the identification begins with complete enzymatic fragmentation of the target proteins with a proteinase of defined specificity such as trypsin. In this way, additional information can be obtained as the properties of the resultant “limit peptides,” notably their mass, can be used as a fingerprint to search databases of all known proteins (1.Cottrell J.S. Protein identification by peptide mass fingerprinting.Pept. Res. 1994; 7: 115-124Google Scholar, 2.James P. Quadroni M. Carafoli E. Gonnet G. Protein identification in DNA databases by peptide mass fingerprinting.Protein Sci. 1994; 3: 1347-1350Google Scholar). This requires a very good match between the experimental data and the theoretical digest map and usually requires that the target protein sequence already be deposited in the database. Searching with peptide maps derived from organisms for which extensive genome data is not available usually requires the gain of further identification data, specifically composition data (3.Pratt J.M. Robertson D.H. Gaskell S.J. Riba-Garcia I. Hubbard S.J. Sidhu K. Oliver S.G. Butler P. Hayes A. Petty J. Beynon R.J. Stable isotope labelling in vivo as an aid to protein identification in peptide mass fingerprinting.Proteomics. 2002; 2: 157-163Google Scholar) or short peptide sequence tags derived by sequencing de novo (4.Mann M. Wilm M. Error-tolerant identification of peptides in sequence databases by peptide sequence tags.Anal. Chem. 1994; 66: 4390-4399Google Scholar, 5.Mann M. A shortcut to interesting human genes: peptide sequence tags, expressed-sequence tags and computers.Trends Biochem. Sci. 1996; 21: 494-495Google Scholar). In peptide mass fingerprinting, the masses of the limit peptides define the positions of the specific amino residues that define the site of cleavage. Proteolysis by trypsin, for example, cleaves the protein into limit peptides typically of ∼10–20 amino acids long. The masses of such peptides are readily matched by the high mass accuracy obtained by the mass analyzers in common use for peptide mass fingerprinting, notably a time-of-flight (TOF) 1The abbreviations used are: TOF, time-of-flight; ACN, acetonitrile; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; EST, expressed sequence tag; MALDI, matrix-assisted laser desorption ionization; MS, mass spectrometry; MS/MS, tandem MS; MaxENT 1, maximum entropy software for deconvolution of multiply charged electrospray envelopes; ESI, electrospray ionization; FTICR, Fourier transform ion cyclotron resonance; BisTris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol. mass analyzer, which can achieve a mass accuracy of 10 ppm and a resolution in excess of 10,000 full-width half-maximum. However, the generation of the set of limit peptides introduces additional steps that can result in sample losses and that diminishes throughput. To circumvent these stages, it has been proposed that the mass of each protein in a proteome can, at an appropriate level of mass accuracy and resolution, be considered to be virtually unique in the data set, which raises the possibility of direct analysis of an intact protein (6.Smith R.D. Pasa-Tolic L. Lipton M.S. Jensen P.K. Anderson G.A. Shen Y. Conrads T.P. Udseth H.R. Harkewicz R. Belov M.E. Masselon C. Veenstra T.D. Rapid quantitative measurements of proteomes by Fourier transform ion cyclotron resonance mass spectrometry.Electrophoresis. 2001; 22: 1652-1668Google Scholar). Indeed the superior mass accuracy and resolution of Fourier transform ion cyclotron resonance (FTICR) mass spectrometers might offer a route to highly accurate mass determination of intact proteins that are electrosprayed as multiply charged species (7.Veenstra T.D. Martinovic S. Anderson G.A. Pasa-Tolic L. Smith R.D. Proteome analysis using selective incorporation of isotopically labeled amino acids.J. Am. Soc. Mass Spectrom. 2000; 11: 78-82Google Scholar, 8.Martinovic S. Veenstra T.D. Anderson G.A. Pasa-Tolic L. Smith R.D. Selective incorporation of isotopically labeled amino acids for identification of intact proteins on a proteome-wide level.J. Mass Spectrom. 2002; 37: 99-107Google Scholar). With such an instrument, proteins between 10,000 and 100,000 Da can be mass measured to an accuracy of around 1 ppm, which can provide data of sufficiently high quality that identification based only on intact mass might in principle be possible. However, protein identification by mass determination of the intact protein is compromised by post-translational changes, including proteolysis, glycosylation, acetylation, phosphorylation, or even disulfide bond formation (9.Lee S.W. Berger S.J. Martinovic S. Pasa-Tolic L. Anderson G.A. Shen Y. Zhao R. Smith R.D. Direct mass spectrometric analysis of intact proteins of the yeast large ribosomal subunit using capillary LC/FTICR.Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5942-5947Google Scholar). All of these modifications elicit a change in the mass of the protein, creating discordance between the measured mass and the mass predicted from the (usually) cDNA-inferred protein sequence. Moreover, FTICR instruments are currently expensive and not yet generally available for proteome analysis. Accordingly we embarked on a study of a complex but natural mixture of proteins using an affordable benchtop mass spectrometer. Specifically we used a preparation of the soluble proteins of chicken skeletal muscle. The adult pectoralis muscle of the chicken (Gallus gallus) is a highly specialized tissue comprising almost exclusively Type IIB fibers. These fibers rely on glycolysis to provide energy for muscle contraction, and the sarcoplasmic protein fraction is rich in glycolytic enzymes. The marked bias in protein expression of this class of proteins means that an unfractionated high speed supernatant fraction of a homogenate of this tissue provides a complex mixture that is nonetheless sufficiently enriched in relatively few proteins to provide an appropriate test material for this analysis. Specifically we wished to assess the extent to which direct mass measurement of the proteins in this sample could be of value in their identification using a readily available instrument (Q-TOF-micro, Micromass, Manchester, UK). Commercial frozen broiler chicken skeletal muscle (5 g) was mechanically homogenized in 25 ml of a buffer consisting of 20 mm BisTris, 50 mm NaCl, pH 6.0 and containing Complete protease inhibitors (Roche Diagnostics). The soluble protein fraction was isolated by centrifugation at 13,000 × g for 45 min at 4 °C. The protein concentration of the resulting supernatant fraction was determined using the Coomassie Plus® Protein assay (Perbio Science UK, Tattenhall, UK), and the supernatant fraction was stored at −20 °C in 200-μl aliquots. The soluble proteins (4–40 μg) were initially separated by one-dimensional 12.5% (v/v) SDS-PAGE and visualized with Coomassie Brilliant Blue stain. In some instances, the gel was scanned, and the band volumes were analyzed by Phoretix 1-D software (Non-linear Dynamics, Newcastle, UK). The correlation coefficient for the relationship between band volume and protein amount was 0.98. For the two-dimensional electrophoresis, the soluble supernatant fraction containing 100 μg of protein was added to 200 μl of rehydration buffer (8 m urea, 2% (v/v) CHAPS, 50 mm dithiothreitol, 0.2% (v/v) pH 3–10 Biolyte ampholytes). This sample was loaded onto 11-cm linear IPG strips, pH 3–10 (Bio-Rad). Passive rehydration was completed overnight, and the strips were then electrophoresed at 250 V for 15 min followed by 250–8000 V for 35,000 V-h using the Protean IEF system (Bio-Rad). The strips were equilibrated in reducing equilibration buffer (6 m urea, 0.375 m Tris, pH 8.8, 2% (w/v) SDS, 20% (v/v) glycerol, 2% (w/v) dithiothreitol) for 15 min and then in alkylating buffer 2 (6 m urea, 0.375 m Tris, pH 8.8, 2% (w/v) SDS, 20% (v/v) glycerol, 2.5% (w/v) iodoacetamide) for a further 15 min before the second dimension separation was performed in 12% gels using the Hoefer EttanTM DALT six vertical gel system at 600 V for 3 h. The proteins were visualized using Coomassie Brilliant Blue stain. A plug of the protein of interest was excised from the gel with a fine glass pipette and transferred to a microcentrifuge tube. To each tube 25 μl of 50 mm ammonium bicarbonate, 50% (v/v) acetonitrile (ACN) was added and incubated at 37 °C for 20 min. This process was repeated until all the stain had been removed. The plugs were then washed in 50 mm ammonium bicarbonate, which was subsequently discarded. 10 was added to each plug and incubated at 37 °C. the supernatant was was added to each and in the for min. The gel was using 25 μl of ACN, and at 37 °C was for 15 min. The supernatant was from the which was to the gel was in 50 mm ammonium containing μl of 100 in 50 mm 50 mm ammonium was added to each and was to at 37 the was by the of 2 μl of of proteins was performed using the All analyses were performed on a instrument and peptide mass were using the search J.S. protein identification by sequence databases using mass spectrometry Scholar, J. E. C. Hubbard S.J. A of chicken 2002; Scholar). were in a with a of in peptide masses in the mass of Da were and used in the The search an of ppm, a of and of The search was to and was mass or All analyses were performed using a tandem mass spectrometer desalting was using a C4 and from which were by an were to the in (v/v) ACN, 0.2% (v/v) and were in (v/v) ACN, 0.2% (v/v) via a capillary into the at 200 were between and with a of to were into a which was subsequently using MaxENT 1 maximum entropy software were between and 100,000 Da at 1 The used to the was Da in all were identified by of the mass with the predicted mass obtained from the for proteins that had been identified by peptide mass All analyses were performed using the were introduced by from at a of from an digest were first by to had mass was between and from which multiply charged were identified as for from their natural isotope The resultant fragmentation were using MaxENT 3 and the peptide sequence was determined using the software in the sequences were the protein using for The was to G. was set to and the was set to were proteins were to both the intact protein and the data already The chicken was used to peptide sequences that not match any of the The was using the and either used the set of sequences or a of sequences specifically derived from chicken skeletal muscle. These were then to any peptide obtained by mass spectrometry. The pattern of proteins in pectoralis muscle of skeletal muscle is relatively and the bias in of protein expression that is a of this one-dimensional proteins can be readily Coomassie Blue The additional resolution of a two-dimensional separation to the mixture to any is some for for the more on the post-translational changes such as the one-dimensional separation is a of the true of the mixture at in the of the most proteins. we embarked on an analysis of this mixture in the of intact we to the proteins by mass spectrometry of peptides of the on the one-dimensional This as the to identification of proteins I. and or 2001; Scholar, protein via tandem mass Mass Spectrom. 2002; 37: was an to The peptide mass were analyzed using the search J.S. protein identification by sequence databases using mass spectrometry Scholar) and were proteins from or from a containing chicken proteins. The proteins that were analyzed in this are marked on the one-dimensional gel with the a to and of the proteins could be for by to known chicken proteins with sequence from for to for identified using peptide maps from and were as and on the of when a of sequences from In both these instances, the 20 from the peptide mass fingerprint search were from species including and a protein, could not be identified with high by peptide mass These protein and further analysis for an of these proteins were de novo by the mass from a analysis of the sample was used to peptides that had been in the digest of the band on the one-dimensional For band the match by peptide mass was to skeletal muscle from for which six of the were matched at a of 250 The predicted mass of this protein, at was with the in the one-dimensional gel to is sequence available for the muscle of the chicken but the match to the chicken was The sequence of the muscle was used to search the chicken skeletal muscle and sequences were that to the of the peptides in the matched to the chick sequence. peptides derived from this protein to were by tandem the by the data but the N-terminal of the these data the of this protein as the muscle of was For band a was obtained 10 and peptides the were matched to the muscle of from six The mass on a gel of this identification as muscle from species has a mass of confirmation of the of this protein from tandem mass spectrometry of peptides not A of peptides were amino residues of protein, and the of this The tandem data a sequencing in an a in the set was by of identification of that of the sequence. protein, band was not readily by peptide mass peptides were by tandem mass amino acids of sequence information All of the peptides extensive sequence to from The of this protein from chicken has not been or but the sequences of both the and of chicken are These were used to search the chick to sequences from skeletal muscle. overlapping and were to an overlapping protein sequence amino acids in A further sequences of chick A were from the protein and The of the tandem data and the of specific peptides in the to sequences this protein as was some of sequencing or further peptides with chick and at the N-terminal of the protein, but is information on this of the chicken A sequence to these identified all of the of this we then on the of the intact proteins in the mass spectrometer. In to the a on the intact mass of the protein as a key usually by such as protein via tandem mass Mass Spectrom. 2002; 37: Scholar, J.M. as a for protein 2002; Scholar, L. J. and mass spectrometric for analysis of Proteome Res. 2002; Scholar). To test the of this to a relatively complex mixture of we electrosprayed the chicken muscle soluble protein fraction into the source of mass spectrometer. was to the intact protein masses of as many of the mixture as was In the first the protein mixture was in (v/v) 0.2% (v/v) and electrosprayed into the source the was very complex with overlapping between 600 and resolution of this into a true mass the analyzed masses of the was not and we used the MaxENT 1 maximum entropy deconvolution software to the true at 10 was used to the of the analysis. This was then to 1 for more mass The to process the data was typically 1 on a 4 at In all instances, the software to a and high quality true mass To the possibility of the protein mixture was and on a × C4 The protein mixture was in (v/v) 0.2% (v/v) and to the The was then washed with the and the mixture was from the with (v/v) 0.2% (v/v) a complex was obtained a high quality true mass The pattern of peaks obtained with the sample or with the sample and on a C4 was very was in either for that can be a of this of analysis S. deconvolution in electrospray Mass Spectrom. Scholar). 20 peaks from this analysis. The masses of these proteins were with the proteins that had been identified by analysis and for which complete protein sequences were available and with chicken skeletal muscle proteins. The most in the true mass a mass of the most protein on a one-dimensional gel of this protein mixture is at on the The sequence of this protein is and the mass of this protein predicted from the cDNA-inferred sequence is the between the and the theoretical masses is This is to the mass of a methionine The the N-terminal initiator methionine which is as a or post-translational (9.Lee S.W. Berger S.J. Martinovic S. Pasa-Tolic L. Anderson G.A. Shen Y. Zhao R. Smith R.D. Direct mass spectrometric analysis of intact proteins of the yeast large ribosomal subunit using capillary LC/FTICR.Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5942-5947Google Scholar, N-terminal of Chem. 2000; Scholar). Indeed many of the for this protein the N-terminal methionine residue in the sequence for example, in which the mass of protein that of the predicted The between the mass and the predicted mass further that sequence in the are either specific to of chicken or are sequencing was to a further 10 proteins to mass peaks in the electrospray mass In some instances, confirmation of the was by a analysis of proteins further by not 2 to by of the N-terminal residue with N-terminal acetylation. of the methionine is followed by N-terminal N-terminal of Chem. 2000; Scholar). Da to the the loss of methionine followed by of an result in a loss of mass of 2 has a measured mass of which is to the mass of methionine identified using intact protein masses determined by post-translational of N-terminal methionine and of N-terminal of N-terminal methionine or of N-terminal of N-terminal of N-terminal methionine of Da of N-terminal methionine and of Da of N-terminal methionine and in a 3 is with the mass of the N-terminal was to In this the between the predicted and masses was a peptide in the map was Da These can be reconciled as loss of N-terminal methionine and a of a residue to a residue However, are residues in this and between the is not in the of tandem data for this The mass of the protein in is with the mass of the N-terminal and be on the of mass we have the of proteins of these masses with on SDS-PAGE that data for the digest the in most instances the loss of the residue was from the mass measurement of the intact protein, in one the N-terminal peptide of was in the of the the N-terminal peptide be expected at of Da with the methionine this peptide was from the but a peptide was at to on the proteins the loss of the N-terminal but the mass are be to the that elicit these mass 10 has an mass of The mass of the protein and followed by of the intact protein and of the digest not provide that this protein is However, are sequence for chicken skeletal muscle in the sequence and G.A. The complete amino sequence of chicken J. Scholar, M. K. K. but not has a and sequence Biochem. Scholar). The sequence of the was derived by and protein the sequence of the second was derived by are of sequence between the However, data provide that in is most to the The extent of of the sequence not of the mass the MaxENT 1 mass is highly we a of were to or not the could be with of the proteins determined Accordingly we a one-dimensional gel of the protein mixture at and each protein by the band volumes were with the obtained by mass spectrometry the correlation was not good to the use of mass spectrometric data for of the use of an by the use of a of the protein of mass have identified such or by the use of to the intact proteins or peptides as for J. and S. J. in These to be by a that an mass separation or by in vivo J.M. Petty J. Riba-Garcia I. Robertson D.H. Gaskell S.J. Oliver S.G. Beynon R.J. of protein a dimension in 2002; which might be with muscle a pattern of protein only of the muscle protein the protein and but the soluble sarcoplasmic proteins a The soluble proteins are glycolytic enzymes as might be expected in a muscle that in this species almost exclusively fibers. This of proteins a one-dimensional or a two-dimensional gel and the proteins that might be expected to be in large However, in any analysis of muscle and protein it is these proteins with the proteins that the of the of in the This study has that a complex mixture of proteins can be analyzed by electrospray ionization mass spectrometer using a benchtop instrument of relatively mixtures of proteins with multiply charged can be readily using the maximum entropy and the masses of the proteins that are identified are measured a few of the predicted on mass measurements of intact proteins have either on relatively mixtures D.H. M.S. Gaskell S.J. Beynon R.J. of proteins in Mass Spectrom. 11: Scholar, R.J. C. Robertson D.H. Gaskell S.J. in in a Chem. 2002; Scholar, J.M. and identification of proteins from complex Am. Chem. Soc. 2002; Scholar) or have L. J. and mass spectrometric for analysis of Proteome Res. 2002; Scholar) or (9.Lee S.W. Berger S.J. Martinovic S. Pasa-Tolic L. Anderson G.A. Shen Y. Zhao R. Smith R.D. Direct mass spectrometric analysis of intact proteins of the yeast large ribosomal subunit using capillary LC/FTICR.Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5942-5947Google Scholar, A protein map of using a two-dimensional liquid 2002; Scholar) separation of protein mixtures to mass we that instruments of can be used to resolve relatively complex mixtures, which might the for at one dimension of the masses of peptides are a relatively a has the added of the protein masses a mass an is on resolution of peaks in the multiply charged but as we have an instrument with a resolution of full-width in this The protein masses predicted from sequence databases be to for post-translational changes, notably the loss of the initiator methionine and of the N-terminal amino by acetylation. A of this study is that we have defined the extent of these post-translational changes for of the proteins in the information that had not been have predicted the mature masses of proteins and and the of the sequences with