protein quantification by selective isolation and ... · protein quantification by selective...
TRANSCRIPT
Protein Quantification by Selective Isolationand Fragmentation of Isotopic Pairs UsingFT-ICR MS
Hannah Johnson,a Stephen C. C. Wong,a Deborah M. Simpson,b
Robert J. Beynon,b and Simon J. Gaskellaa Michael Barber Centre for Mass Spectrometry, School of Chemistry and Manchester InterdisciplinaryBiocentre, University of Manchester, Manchester, United Kingdomb Proteomics and Functional Genomics Research Group, Faculty of Veterinary Science, University ofLiverpool, Liverpool, United Kingdom
Isolation of tryptic peptide ions, along with their differentially labeled analogs derived from anartificial QconCAT protein, is performed using multiple correlated harmonic excitation fieldsin an FT-ICR cell. Simultaneous fragmentation of the isolated unlabeled and labeled peptidepairs using IRMPD yields specific y-series fragment ions useful for quantification. The massincrement attributed to stable isotope labeling at the C-terminus is maintained in theC-terminal fragment ions, providing multiple measurements of labeled/unlabeled intensityratios during highly selective detection. The utility of this approach has been demonstrated inthe absolute quantification of components of an unfractionated chicken muscle proteinmixture. (J Am Soc Mass Spectrom 2008, 19, 973–977) © 2008 American Society for MassSpectrometry
The absolute quantification of proteins at theglobal level is an important challenge that mustbe addressed, particularly in systems biology [1].
When modeling cellular pathways and networks it isimportant to quantify the constituent proteins. Relativequantification methods are designed to compare pro-tein amounts in matched samples [2]. However, thesemethods are limited to identifying changes in theamount of a protein with respect to a second cellularstate and samples vary even within control sets of thesame organism [3].
When comparing samples analyzed at differenttimes, using different instrumental platforms andwithin different laboratories, absolute quantificationusing a universal stable isotope-labeled internal stan-dard of known concentration allows direct comparison.Currently, the most widely used methods rely on thewell-established principles of stable isotope dilutiontechniques, using tryptic peptides as surrogate analytesfor the proteins of interest [4]. Chemical synthesis andstable isotope incorporation of individual internal stan-dards allow direct signal comparison, enabling theinference of protein quantities [4]. The production of acollection of internal standards in a single step has beenenabled by the design of a DNA construct that istranscribed and translated into a protein concatamer, atechnique referred to as QconCAT [5–7]. The produc-
Address reprint requests to Prof. Simon J. Gaskell, University of Manches-ter, School of Chemistry and Manchester Interdisciplinary Biocentre, John
Garside Building, 131 Princess Street Manchester M1 7DN, UK. E-mail:[email protected]© 2008 American Society for Mass Spectrometry. Published by Elsevie1044-0305/08/$32.00doi:10.1016/j.jasms.2008.03.012
tion of a QconCAT protein thereby enables absolutequantification of multiple proteins using their peptidesurrogates encoded within the QconCAT [7]. In a typi-cal QconCAT workflow those proteins for which abso-lute quantification is required are defined and one ormore specific reference peptides (Q-peptides) are cho-sen for each protein [5]. The DNA coding sequences forthese peptides are concatenated, inserted into a vec-tor, and expressed in Escherichia coli to form anartificial QconCAT protein. Growth of the transfectedE. coli in stable isotope-containing media yields la-beled QconCAT protein [5, 6]. To achieve absolutequantification, the QconCAT protein is introduced, in aknown amount, into the biological sample of interestand proteins are digested with trypsin to yield Q-peptides alongside native peptides. The light-to-heavyratio (L/H) of native peptide to the internal referencepeptide is calculated following mass spectrometricanalysis, allowing inference of the absolute concentra-tion of each native peptide and thus the protein fromwhich it was derived.
The use of tandem mass spectrometry (MS/MS) inrelative quantification has previously been demon-strated to increase both the specificity and sensitivity ofanalyses [8, 9]. In this study, we use multiple correlatedharmonic excitation fields (multi-CHEFs) as a methodof ion isolation within a Fourier transform ion cyclotronresonance (FT-ICR) MS cell. Multi-CHEFs are tailoredwaveforms that are stitched together within the acqui-sition software to allow selective isolation of multiple
precursor ions according to their m/z [10, 11]. To achievePublished online April 8, 2008r Inc. Received January 21, 2008
Revised March 25, 2008Accepted March 25, 2008
974 JOHNSON ET AL. J Am Soc Mass Spectrom 2008, 19, 973–977
quantification, isotopic variants of a peptide of interest(both naturally occurring and stable isotope-containing)are co-isolated from the unfractionated ion population.Nonselective activation of these isolated ions, usinginfrared multiphoton dissociation (IRMPD) [12], gener-ates quantitative and qualitative information pertainingto the differentially labeled Q-peptide pairs. Stableisotope incorporation using 15N or [13C6]-K/R intro-duces a mass increment of 12–24 and 6 Da, respectively,for each Q-peptide. Where [13C6]-K/R has been used,specific quantitative measurements can be derived fromthe y-series fragment ions, which differ by a constant 6Da. Labeling with 15N produces a varying mass differ-ence between labeled and unlabeled counterparts inboth b- and y-series ions, making data processing morechallenging.
This approach significantly reduces the confoundinginfluence of any peptides whose precursor ion signaloverlaps with the analytes of interest. Because of thecomplexity of biological samples, it is important tovalidate the quantification calculated using the precur-sor ion population alone if using limited or no peptide/protein separation pre-MS. In our work, two indepen-dently selected isotopic variants (arising from labeledQconCAT and native chicken protein) constitute a pre-cursor ion population that is subsequently fragmentedby IRMPD [12]. We have demonstrated the utility ofthis approach for absolute quantification of proteinswithin unfractionated chicken muscle protein.
Experimental
Sample Preparation
Chicken skeletal muscle protein and stable isotope [15N]-and [13C6]-K/R-labeled chicken muscle QconCAT proteinwere prepared and quantified as described previously[6, 7]. Labeled QconCAT was added in known amountsto unfractionated chicken muscle protein sample forquantitative analysis. Proteins were reduced, alkylated(with dl-dithiothreitol and iodoacetamide, respec-tively), and incubated with trypsin at 37 °C for 18 h.Peptides were desalted postdigestion using a C18 pep-tide trap (Presearch Ltd, Hampshire, UK) and stored at�20 °C until MS analysis.
FT-ICR Mass Spectrometry
Experiments were performed on a 9.4-T FT-ICR massspectrometer (Apex III™; Bruker Daltonics, Billerica,MA, USA) with an Apollo electrospray source andSynrad 48 series IR laser (� � 1060 �m; Synrad, Mukil-teo, WA, USA). Selective isolation of ions was achievedusing multi-CHEF [11], where each “notch” was cen-tered on the principal isotopic variant of the selectedprecursor ion. Pulsed argon gas was used to facilitatethe confinement of ions by collisional cooling, therebymaximizing overlap of the ion cloud with the IR laser
beam. Multi-CHEF isolation of isotopic envelopes cor-responding to unlabeled and labeled peptide variantswas achieved with 80% efficiency. Losses increased to50% when pulsed gas was not used. The co-isolatedpairs of native and stable isotope-bearing Q-peptideions were excited in a nonselective manner to promotefragmentation. The calculation of light/heavy isotoperatios was performed manually using the peak height ofthe principal isotopic peaks of the light and heavyanalogues.
Results and Discussion
Unlabeled and [13C6]-K/R QconCAT protein weremixed in a 1:1 ratio, digested with trypsin, and analyzedby FT-ICR MS. Each pair of tryptic peptide [M � 2H]2�
ions displayed a 6 Da (3 Th) difference, consistent withthe presence of one lysine or arginine residue. Variousisolation notch widths were evaluated with respect tothe reproducibility of the subsequent determination ofthe signal ratio of labeled to unlabeled analogues.Achievement of the narrowest precursor ion isolationwindow (and therefore highest effective resolution ofthe precursor ion) was achieved using the highestexcitation energy. Initial experiments involved the useof a double-notch waveform to effect specific isolationof the principal isotopic variants of the unlabeled andlabeled peptide ions. This, however, resulted in signif-icant loss of signal intensity, which was proportionatelygreater for the lower mass (higher frequency) ion popula-tion, thereby confounding quantitative analysis. Thusmulti-CHEF was used to isolate the isotopic envelope ofboth the natural and stable isotope-incorporating peptideanalogues.
The combined precursor ion population was thensubjected to IRMPD, producing a series of singlycharged y-ion pairs, with a separation of 6 Th (attrib-uted to the retention of the C-terminal label), displayingL/H ratios useful for quantification. For each pair ofpeptide analogs, a preisolation L/H ratio was calcu-lated along with the L/H ratio upon isolation and themean L/H ratio of the y-series fragment ion doublets.The L/H ratios for 10 peptide pairs pre- and postisola-tion were 0.97 � 0.10 and 0.96 � 0.10 (n � 36) (where nis the number of analyses), indicating that the isolationprocess as applied had no effect on quantification.Postfragmentation, the mean L/H ratio calculated forthe fragment ions in this dataset was 1.08 � 0.11. Foreach product ion spectrum the mean of the coefficient ofvariation for the y-series fragment ions used was 14 �7%, where four to eight fragment ion pairs were usedfor quantification within each experiment. For a singlepeptide ion pair of m/z 929.96/932.97 (T21), the L/Hratios for a 1:1 mixture of light and heavy QconCATwere 0.98 � 0.10 and 0.95 � 0.04 (n � 14), pre- andpostisolation, respectively. For this peptide analyzedonce, the resulting fragment ions produced the L/Hratio 1.05 � 0.07 (f � 7) (where f is the number of
y-series fragment ion doublets). For the same peptide975J Am Soc Mass Spectrom 2008, 19, 973–977 PROTEIN QUANTIFICATION USING SELECTIVE TANDEM MS
pair analyzed 14 times, the mean of the mean fragmention ratios was 1.07 � 0.14 (n � 14).
To assess this approach for determination of isotoperatios across a broader range, [13C6]-K/R-incorporatingQconCAT was mixed in a ratio of 1:10 with unlabeledQconCAT, digested with trypsin, and analyzed byFT-ICR MS. The L/H ratios obtained preisolation, uponisolation, and postfragmentation, for three peptides,were 0.105 � 0.008, 0.104 � 0.003, and 0.130 � 0.006(n � 7). A mean CV of 22 � 2.4% was calculated foreach individual product ion spectrum, where three tofour y-series fragment ion pairs were used for quanti-fication. The observation of increased L/H ratios andhigher CV values reflected reduced signal to noise ofthe unlabeled counterpart. These data suggest satisfac-
600 700 8000.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Rel
ativ
e In
tens
ity
m
Isolation(a)
(c)
*
600 700 8000.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Rel
ativ
e In
tens
it y
m
NativeQ-pep
Figure 1. FT-ICR analyses of a tryptic digest owith [13C6]-K/R QconCAT protein. (a) Mass specwith trypsin. (b) Expanded region correspondinglabeled Q-peptide counterpart. (c) Mass spectrumselective isolation of ions within the range, m/z 9
represents electronic noise. (d) Expanded region, m/z 93tory accuracy for the determination of ratios 1:10, al-though the precision is reduced.
This approach was further validated using unlabeledand [15N]-labeled QconCAT mixed at ratios of 7:1before digestion with trypsin. Because of the nature ofthis labeling method the resulting [15N]-labeled pep-tides have mass differences in comparison with theirunlabeled counterparts, which are dependent on aminoacid composition. For five pairs of labeled and unla-beled peptides the pre- and postisolation and postfrag-mentation ratios were 6.47 � 2.39, 5.98 � 2.15, and 6.25 �0.03 (n � 5). For T19, the labeled peptide isotope patternoverlaps with that of unlabeled T20 peptide; this pre-vents accurate determination of the L/H ratios fromconventional mass spectra (not MS/MS). The apparent
928 930 932 934 9360.0
0.2
0.4
0.6
0.8
Rel
ativ
e In
tens
ity
m/z
1000 1100 1200
928 930 932 934 936 9380.0
0.2
0.4
0.6
0.8
1.0
Rel
ativ
e In
t ens
ity
m/z
nativeT21
Q-peptide T21H
nativeT21
Q-peptide T21H
(b)
928 930 932 934 936
Rel
ativ
e I n
t en s
ity
0.2
0.4
0.6
0.8
928 930 932 934 936
0.2
0.4
0.6
0.8
1.0
1000 1100 1200
and T21H
Rel
ativ
e In
tens
ity
m/z
m/z(d)
ractionated chicken muscle protein supplementedof unfractioned chicken muscle protein digestedubly protonated FGVEQNVPMVFASFIR and itsested, unfractioned, chicken muscle protein after
33. The peak labeled * at approximately m/z 710
900
/z
900
/z
T21 tide
f unftrumto doof dig30–9
0–933, after isolation.
976 JOHNSON ET AL. J Am Soc Mass Spectrom 2008, 19, 973–977
L/H ratio obtained pre- and postisolation for T19 was1.22 and 1.19, respectively. Following fragmentation thecomposite product ion spectrum included a y-ion seriesspecific to T19. Two members of this series (y2 and y3),showing no overlap with other fragments, were se-lected for quantification yielding L/H ratio of 6.21and 7.57 (average 6.89). Thus, in this instance thedetermination of isotope ratios was confoundedwhen using conventional MS analysis alone (withoutchromatographic separation). Deconvolution of over-lapping signals was achieved by precursor ion isola-tion and fragmentation, enabling accurate quantifica-tion. Furthermore, the precision of isotope ratiodetermination was significantly improved when us-ing fragment ion data.
The same MS/MS approach was utilized for thequantification of an unfractionated chicken cell lysate,with [13C6]-K/R-labeled QconCAT as the internal stan-dard. Chicken muscle QconCAT (93 nmol/g of chickenmuscle) was added to chicken muscle soluble proteinfrom a 30-day-old broiler chicken. The mixture wasdigested with trypsin and analyzed by FT-ICR MS. Theconventional mass spectrum (Figure 1a) indicates thecomplexity of the unfractionated cell lysate, with ap-proximately 500 peaks being readily observed. Figure1b shows an expansion of that portion of the spectrumexpected to include both labeled and unlabeled ana-logues of the protonated T21 peptide (sequenceFGVEQNVPMVFASFIR, designed for the quantifica-tion of chicken pyruvate kinase). The complexity of theapparent isotope patterns suggests the overlappingdetection of multiple species, implying the need fordetection of enhanced selectivity. The determination ofsignal intensity ratio (L/H) for the T21 peptide usingconventional MS (Figure 1b) gave an apparent value of0.17, but this is expected to be inaccurate in view of theoverlapping signals observed. The entire isotopic enve-lopes of native chicken peptide T21 and its [13C6]R-analogue were then isolated (Figure 1c). The relativesignal intensities corresponding to the isolated species(Figure 1d) reproduced those observed in the originalspectrum (Figure 1b). Upon fragmentation of the iso-lated species, a series of six y-fragment ion pairs wereobserved, each separated by 6 Th (Figure 2). A labeled/unlabeled signal ratio was determined for all pairsexcept that corresponding to y1; this was excludedbecause of a possible contribution from other trypticpeptides not separated in the mass spectrometric anal-ysis. A mean ratio of 0.24 was calculated for theremaining five y-series ions (Figure 2) with a CV of 10%;this ratio is significantly different from that determinedfrom the conventional mass spectrum, confirming thebenefit of enhanced selectivity of analysis. It is notewor-thy that the apparent L/H ratio for y1 was 0.69, sug-gesting its origin from multiple peptide precursors. Thisparticular experiment was repeated and L/H ratiospreisolation, upon isolation, and the L/H ratio of thefragment ions were calculated at 0.16 � 0.01, 0.16 � 0.01,
and 0.26 � 0.04 respectively (n � 6). These data indicate anabsolute quantity of pyruvate kinase of 24.2 � 4.1 nmol/gof chicken muscle.
Conclusions
FT-ICR MS using the multi-CHEF technique has beenintroduced to simultaneously isolate and fragment lightand heavy Q-peptide pairs. Furthermore, this approachhas been used to quantify unfractionated native chickenmuscle peptides alongside [13C6]-K/R-labeled internalstandards to achieve an increase in specificity. Isolationof ions of interest using multi-CHEF decreases ioninterference, allowing the production of specific y-seriesfragment ions by IRMPD. Quantification using thesey-series ions allows increased accuracy of quantificationwhen there are overlapping species present at the samem/z. The isolation of multiple ions within an ICR cellcan also be achieved using stored waveform inverseFourier transform (SWIFT) [13]. The implementation ofSWIFT as an alternative approach to ion isolation forabsolute quantification is in progress.
In summary, we have reported an approach to theabsolute quantification of proteins using MS/MS,which exploits the selective ion isolation capabilities ofFT-ICR. This approach is of particular value when thecomplexity of the mixture to be analyzed is such thataccurate and precise quantification is compromisedwhen using conventional MS detection. We anticipateselective use of this method, for validation of dataproduced using less rigorous approaches.
AcknowledgmentsThis work was supported by the UK Engineering and Physical
400 6 00 80 0 10 00 120 0
0 .1
0 .2
Rela
tive
Inte
nsity
0.2
0.4
0.6
0.8
1.0
400 600 800 1000 1200
y4-(h)
y5-(h)y3-(h)
y6-(h)
*
m/z
[M+2H]+2
FGVEQNVDMVFASFIR
y8-(h)
++
++
+
++
EQNV-28
y3 y4 y5
y6
y8
Figure 2. Product ion spectrum showing ions derived fromdoubly protonated tryptic peptide, T21 (FGVEQNVPMVFASFIR)and its labeled Q-peptide partner following IRMPD. y-ionsderived from T21 are noted; “yn(h)” indicates a 13C6-containingfragment. The peaks labeled “�” are not derived fromFGVEQNVPMVFASFIR and are attributed to other precursorions of similar m/z. The peak labeled * at approximately m/z 710represents electronic noise.
Sciences Research Council (reference EP/D013615/1) and the
977J Am Soc Mass Spectrom 2008, 19, 973–977 PROTEIN QUANTIFICATION USING SELECTIVE TANDEM MS
Biotechnology and Biological Sciences Research Council (referenceBB/D013615/1). The authors thank the Roslin Institute, Edin-burgh, UK for rearing birds and collecting muscle samples,Jennifer Rivers (University of Liverpool) for the preparation of thechicken muscle protein lysate, and Drs. Claire Eyers and SarahHart for useful discussions.
References1. Aebersold, R.; Mann, M. Mass Spectrometry-Based Proteomics. Nature
2003, 422, 198–207.2. Ross, P. L.; Huang, Y. N.; Marchese, J. N.; Williamson, B.; Parker, K.;
Hattan, S.; Khainovski, N.; Pillai, S.; Dey, S.; Daniels, S.; Purkayastha, S.;Juhasz, P.; Martin, S.; Bartlet-Jones, M.; He, F.; Jacobson, A.; Pappin, D. J.Multiplexed Protein Quantitation in Saccharomyces cerevisiae UsingAmine-Reactive Isobaric Tagging Reagents. Mol. Cell. Proteomics 2004, 3,1154–1169.
3. Gan, C. S.; Chong, P. K.; Pham, T. K.; Wright, P. C. Technical,Experimental, and Biological Variations in Isobaric Tags for Relativeand Absolute Quantitation (iTRAQ). J. Proteome Res. 2007, 6, 821–827.
4. Gerber, S. A.; Rush, J.; Stemman, O.; Kirschner, M. W.; Gygi, S. P.Absolute Quantification of Proteins and Phosphoproteins from CellLysates by Tandem MS. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 6940–6945.
5. Beynon, R. J.; Doherty, M. K.; Pratt, J. M.; Gaskell, S. J. Multiplexed
Absolute Quantification in Proteomics Using Artificial QCAT Proteinsof Concatenated Signature Peptides. Nat. Methods 2005, 2, 587–589.6. Pratt, J. M.; Simpson, D. M.; Doherty, M. K.; Rivers, J.; Gaskell, S. J.;Beynon, R. J. Multiplexed Absolute Quantification for Proteomics UsingConcatenated Signature Peptides Encoded by QconCAT Genes. Nat.Protocols 2006, 1, 1029–1043.
7. Rivers, J.; Simpson, D. M.; Robertson, D. H. L.; Gaskell, S. J.; Beynon,R. J. Absolute Multiplexed Quantitative Analysis of Protein Expressionduring Muscle Development Using QconCAT. Mol. Cell. Proteomics2007, 6, 1416–1427.
8. Warwood, S.; Mohammed, S.; Cristea, I. M.; Evans, C.; Whetton, A. D.;Gaskell, S. J. Guanidination Chemistry for Qualitative and QuantitativeProteomics. Rapid Commun. Mass Spectrom. 2006, 20, 3245–3256.
9. Panchaud, A.; Guillaume, E.; Affolter, M.; Robert, F.; Moreillon, P.;Kussmann, M. Combining Protein Identification and Quantification:C-Terminal Isotope-Coded Tagging Using Sulfanilic Acid. Rapid Com-mun. Mass Spectrom. 2006, 20, 1585–1594.
10. de Koning, L. J.; Nibbering, N. M. M.; van Orden, S. L.; Laukien, F. H.Mass Selection of Ions in a Fourier Transform Ion Cyclotron ResonanceTrap Using Correlated Harmonic Excitation Fields (CHEF). Int. J. MassSpectrom. Ion Process. 1997, 165–166, 209–219.
11. Kruppa, G.; Schnier, P. D.; Tabei, K.; van Orden, S.; Siegel, M. M.Multiple Ion Isolation Applications in FT-ICR MS: Exact-Mass MSInternal Calibration and Purification/Interrogation of Protein-DrugComplexes. Anal. Chem. 2002, 74, 3877–3886.
12. Little, D. P.; Speir, J. P.; Senko, M. W.; O’Connor, P. B.; McLafferty, F. W.Infrared Multiphoton Dissociation of Large Multiply Charged Ions forBiomolecule Sequencing. Anal. Chem. 1994, 66, 2809–2815.
13. Chen, L.; Wang, T. C. L.; Ricca, T. L.; Marshall, A. G. Phase-Modulated
Stored Waveform Inverse Fourier Transform Excitation for Trapped IonMass Spectrometry. Anal. Chem. 1987, 59, 449–454.