Predicting the Highest Intensity Ion in Multiple Charging EnvelopesObserved for Denatured Proteins during Electrospray IonizationMass Spectrometry by Inspection of the Amino Acid SequenceKevin A. Douglass and Andre R. Venter*

Department of Chemistry, Western Michigan University, Kalamazoo, Michigan 49008-5413, United States

*S Supporting Information

ABSTRACT: A simple, manual method for predicting thehighest intensity charge states (HICS) of denatured proteinions generated by electrospray ionization based on inspectionof the proteins’ amino acid sequence is proposed. The HICS isaccurately predicted by identifying groupings of nearby basicamino acids in the positive mode or acidic amino acid residuesin the negative mode. The method assumes that the likelihoodof having more than one charge per group becomes less likelydue to Coulombic repulsion of like charges. It is shown empirically that a spacing of at least three noncharged residues is requiredbetween charged amino acids for the charge state with the highest intensity. Verification of this method is presented, and itslimitations are identified. It is fast, inexpensive, and provides similar, although less detailed, information as state-of-the-artmethods that rely on computational calculations. With a few exceptions, the highest intensity charge states were predicted to anaverage of one charge state of the experimental data. For those proteins whose HICS were not accurately estimated, theexperimental values fell short of the predictions. Upon reduction of the disulfide bonds of these proteins, the experimental HICSbecame closer to the predicted values, suggesting that charging lower than the prediction can be attributed to conformationalinflexibility of those proteins.

Electrospray ionization mass spectrometry (ESI-MS) hasbecome an attractive tool for the study of proteins and

other biological macromolecules in part because it is a softionization technique that can preserve solution-phase analytecharacteristics and interactions into the gas phase2 as well asproduce extensive multiple charging so that proteins can beanalyzed even on instruments with limited mass ranges. Forinstance, protein conformation prior to ionization has aprofound effect on the degree of multiple charging observedduring ESI-MS analysis,3 which enables ESI-MS to be used forthe investigation of protein conformation in solution. Proteinmolecules in their native or folded states typically have chargestate distributions (CSD) present at lower m/z values relativeto their denatured, or unfolded, states. The cause of this shiftfrom higher to lower m/z values upon unfolding is attributed toseveral factors, including an increase in the steric accessibility ofpossible charge sites,4−6 a reduction in the electrostaticrepulsion upon increased distance between charge sites,7,8

and the disruption of charge-neutralizing interactions betweenoppositely charged residues.9 Extensive work relating to theionization mechanisms for both folded and unfolded states hasbeen done, and understanding the mechanisms of proteincharging during ESI-MS has enabled the highest observedcharge state (HOCS), with the maximum number of charges, ofproteins to be predicted with confidence.6,9−14

For folded proteins, ionization by ESI is usually described interms of the charge residue model (CRM) originally proposedby Dole and coworkers15 According to this model, the

maximum number of charges on a protein ion is determinedby the number of charges able to be supported on a solventdroplet having a similar surface area as the molecule,10 which inturn is determined by the Rayleigh equation16 and thereforedependent on the solvent surface tension. Additional supportfor this model can be inferred from the good log−logcorrelation between the solvent-accessible surface area andthe average charge state determined for a large set of proteins.13

However, some studies have shown that the multiple chargingof proteins cannot always be explained by surface tensionalone,17−19 indicating that other mechanisms must play a part.One such mechanism is described by the conformation-dependent neutralization theory (CNT),9 where the extent ofmultiple charging relies on the neutralization of oppositelycharged residues, which in folded proteins can be found withinthe protein interior as part of ion−ion pairs. Recently, a theorycompatible with both the CRM and CNT which relies only onthe apparent gas-phase basicity (GBapp) of the protein relativeto that of the solvent was proposed.11

For unfolded proteins the HOCS usually appears limited bythe number of basic residues (lysine, histidine, and arginine)present on the protein,7,20 though charging beyond this valuecan be observed under certain experimental conditions.3,6,21 Itis also generally agreed that the extent of multiple charging is

Received: April 25, 2013Accepted: July 31, 2013

Article

pubs.acs.org/ac

© XXXX American Chemical Society A dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXX

determined by the GBapp of the protein ions relative to the GBof the solvent.11,12,14,22 Using the positive ionization mode asan example, each amino acid residue of a protein moleculepossesses an intrinsic GB, which has been found to be higher by50 to 70 kJ/mol than that of the free amino acid due tostabilizing intramolecular interactions in the polymer chain.23,24

With each additional charge gained by the protein, the GBs ofthe remaining residues are decreased by the resultingCoulombic repulsion. The GBapp of a particular charge stateis assumed to be that of the residue with the highest GB afteraccounting for the additional Coulombic energy added by thepresence of all the existing charges. The HOCS is determinedby the last charge state that has a GBapp higher than that of thesolvent. For several proteins, both folded and unfolded, thechange in GBapp with increasing charge has been measured14,24

and/or calculated11,12,24 to decrease linearly with an increasingnumber of charges. Good agreement between the HOCS ofprotein ions and the GB of the solvent has beendemonstrated.11,12,24 Alternatively, it has been proposed thatthe multiple charging of denatured proteins might occur duringejection from charged droplets in a manner similar to the ionevaporation model (IEM) proposed by Iribarne andThomson,25 where the extent of multiple charging (HOCS)would be determined by the size, or more specifically thecharge density, of the originating droplet.26,27 Upon unfolding,protein molecules migrate immediately to the surface of thedroplet to minimize unfavorable interactions between the polarsolvent and the newly exposed nonpolar residues. One of theterminal ends then emerges from the droplet, followed by thesequential ejection of the rest of the chain. During ejection,charges equilibrate between the droplet and the protein on atime scale much faster than the ejection process.27 Moleculardynamics studies validating this charge ejection model (CEM)were able to accurately reproduce experimental data forcytochrome c and myoglobin without considering the gas-phase basicities of potential charge sites.28

Less work on the origin of the highest intensity charge state(HICS) has been done than for the HOCS. The HICS is ofparticular interest for quantitative protein analysis. Inferencescan be made based on the explanations proposed for the widthof the CSDs for denatured proteins. Typically, CSDs for foldedproteins are much narrower (encompass fewer charge states)than those for unfolded proteins.4 One possible explanationsuggests that the heterogeneity of conformations available tounfolded proteins leads to an increase in the width of thecharge state distribution.3,4,28 While folded proteins possess afew unique conformations, unfolded proteins, or proteins withvarious degrees of unfolding, possess numerous conformationsable to encompass a larger range of protonation.9,29 The mostprominent conformation or subset of conformations resultingin the same extent of multiple charging would determine theHICS. Another hypothesis suggests that since ESI produces arange of droplet sizes which can support a larger or smallernumber of protons depending on how large or small thedroplets are, protein ions which originate from these dropletswould therefore have a larger or smaller degree of protonation,respectively.28 In this case, the HICS would be indicative of themean droplet size.In the present study, we demonstrate that groupings of

nearby basic or acidic residues in the amino acid sequence of aprotein can be used to accurately predict the experimentalHICS for denatured proteins analyzed by ESI-MS without the

use of additional equipment or extensive computationalcalculations.

■ EXPERIMENTAL SECTIONBovine aprotinin, bovine ubiquitin, bovine cytochrome c,bovine hemoglobin, equine myoglobin, bovine serum albumin,and dithiothreitol (DTT) were purchased from Sigma-Aldrich(St. Louis, MO, U.S.A.). β-Lactoglobulin and α-chymotrypsi-nogen were purchased from MP Biomedical (Solon, OH,U.S.A.). Green fluorescence protein (GFP), a mutant form ofthe N-terminal domains 5 and 6 of the Wilson disease protein(WDP-5,6), and a xanthine alkaloid methyltransferase wereexpressed and purified in house. Sequence information forthese proteins is provided in the Supporting Information(Figure S-1). Formic acid and methanol were purchased fromFluka Analytical (St. Louis, MO, U.S.A.). Water was providedin house from a Purelab Ultra water purification system (ElgaLLC, Woodridge, IL, U.S.A.). All reagents were used asreceived without any additional purification. Unless indicatedotherwise, protein sequence data was obtained from theNational Center for Biotechnology Information30 or theRCSB Protein Data Bank (PDB).31

A linear ion-trap mass spectrometer (LTQ, ThermoScientific, Waltham, MA, U.S.A.) equipped with an Ionmaxpneumatically-assisted electrospray source was used for all ESI-MS analysis. Protein samples were prepared to 10 μM in either50/50/0.1 MeOH/H2O/formic acid for positive mode analysisor 80/19/1 MeOH/H2O/NH4OH for negative mode analysis.The instrumental parameters included a spray voltage of 5 kV, acapillary voltage of 20 V, a capillary temperature of 250 °C, asheath gas flow rate of 10 arbitrary units, and a sample deliveryrate of 5 μL/min. The tube lens value is an important mass-dependent ion transmission parameter for the LTQ. Because ofthe large range of protein masses investigated in the presentstudy, this value was set to the optimum for each protein with atypical value of around 150 V. The spectra were generated byaveraging approximately 50 scans for each protein.For proteins treated with DTT prior to analysis, DTT was

added to an aqueous solution of the protein so that theconcentrations of the protein and DTT were 20 μM and 10mM, respectively. The solutions were analyzed after 12−24 h atroom temperature. MeOH and formic acid were added so thatthe final solvent composition was 50/50/0.1 MeOH/H2O/formic acid with a protein concentration of 10 μM.

■ RESULTS AND DISCUSSIONThe extent of multiple charging (HOCS) for a denaturedprotein by ESI has been estimated by considering theelectrostatic modification of the protein’s GBapp upon multiplecharging.11,12,24 It was shown that when basic residues arepresent as a grouping of two or more nearby residues, the GBsof the residues decrease with the presence of charges within thegrouping because of the additional Coulombic repulsionbetween like charges. The additional Coulombic energy termmodifying the GBs is inversely proportional to the distance (r)between the charges (q) on sites i and j according to Coulomb’slaw:

∑πε ε

= −⎛⎝⎜⎜

⎞⎠⎟⎟

qr

GB GB1

4 ijapp int

0

2

where ε0 is the permittivity of free space and ε is the relativepermittivity of the surrounding medium. The largest

Analytical Chemistry Article

dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXXB

Coulombic contributions appear when two charges aresupported on adjacent or nearby residues because of theinverse relationship between the energy and distance betweencharges, which causes the Coulomb potential to varysignificantly at short distances. For instance, the spacingbetween the α-carbons of consecutive amino acid residues isapproximately 4 Å.27 Coulomb potentials for charge separationsof 8 and 12 Å are 86.8 and 57.9 kJ/mol, respectively, assuminga relative permittivity of 2.24

On the basis of these published results, it seems probablethat a grouping of nearby residues on a protein chain wouldend up with only a single charge because of the closeness of thebasic sites, and so we recently proposed that the HICS of anunfolded protein can be predicted without any calculations ormodeling simply by inspection of the amino acid sequencefollowed by an assignment of charges based on charge siteproximity.32 According to this method, the HICS is predicted inthe positive ion mode by assigning the maximum number ofcharges to the basic residues and the N-terminus, with theprovision that at least three residues separate each charge.Similarly for the negative ion mode, the maximum number ofcharges is assigned to the acidic residues aspartic acid, glutamicacid, and the C-terminus, again with at least three residuesseparating each charge. The requirement for a minimum ofthree separating residues was determined empirically as shownin Figure 1.

Charge separations of zero to five residues were evaluated fortheir predictive accuracies. For each charge separation, theaverage difference between the predicted and experimentalHICS (ΔHICS) was calculated for the eight proteins in Table 1between 6 and 27 kDa:

Δ =∑ −= x x

nHICS

( )in

i p i e1 , ,

where xi,p is the predicted HICS for the ith protein, xi,e is theexperimentally observed HICS, and n is the number of proteinsincluded in the calculation. The best results were obtained for aminimum separation of three residues with an average ΔHICSof −0.25 ± 1.83 charge states. At separations less than threeresidues, the HICS becomes increasingly overestimated, forexample, a separation of two results in an average ΔHICS of

+0.88 ± 1.46 charge states, while at separations greater thanthree residues the HICS becomes increasingly underestimated,for example, a ΔHICS of −2.00 ± 2.20 charge states for aseparation of four. Therefore, a minimum charge separation ofthree residues was used within the current study. The relativelysmall difference in the accuracy between the minimum chargeseparations of two or three residues could indicate that theexact separation depends somewhat on the identity of theseparating residues, as this would influence the real distancebetween nearby potential charge sites and their individualbasicities, as discussed below. However, this refinement wouldmake this simple HICS prediction method too complicated toachieve by direct inspection using simple algorithms, and wouldbegin to require extensive computational approaches.Following the example given in Figure 2a for myoglobin, and

starting at the N-terminus side, basic residues G1, K16, H24,R31, H36, and K42 are spaced more than three positions apartand are likely to be protonated in the HICS. Residue K45 is notassigned a charge because there are less than three residuesseparating it from K42. When multiple basic or acidic aminoacids are spaced closer together than three positions on thesequence, it is sometimes possible that more than one chargeassignment exists that results in the same maximum number ofcharges overall. For example, only one of the basic residuesK47−K50 can be charged in order to maintain a minimumthree-residue separation from the other charges, but whichresidue is assigned the charge does not affect the overall count.This method does not contain the necessary level of detail todetermine which specific amino acid in a group will becomecharged. Some flexibility exists in group assignment as long asthe maximum number of charge sites are identified. Forexample, K45 is not shown grouped together with K42 in theexample because doing so would change the appearance ofcharge group 7, but without changing the maximum number ofcharges. This flexibility in group assignment is furtherillustrated in Figure 2b, which shows different ways to placecharges on nearby basic residues for the residues K77−H82 ofthe myoglobin sequence shown in Figure 2a (charge sites 10and 11). Only the bottom three assignments are consideredvalid for predicting the HICS by this method, because the aimis to count the maximum number of possible charge sites usingthe algorithm outlined above. For simplicity’s sake, only onepermutation for charge assignment to each group is shown foreach sequence in Figures 2 and 3. Following this method forthe entire sequence, the HICS for myoglobin is predicted to bethe +20 charge state. This prediction is in good agreement withthe experimental result shown in Figure 3c, which shows anESI-MS spectrum for bovine myoglobin acquired underdenaturing conditions.The simplest way to implement this predictive method in

practice is to start from the N-terminus side of the sequenceand assign a charge to every possible charge site (basic residuesand the N-terminus for positive mode, acidic residues and C-terminus for negative mode), ensuring that at least threeresidues separate each new charge from the previous. Using thismethod the maximum number of charges was always found.The maximum charge does not depend on whether one startsassigning from the N- or C-terminus, although a differentcharge assignment could result. The maximum assignablenumber of charges was shown to be the best predictor for theHICS for the proteins investigated in this study.The remaining panels in Figure 3 show the distributions of

basic amino acid residues in the sequences of four other

Figure 1. Average difference between the predicted HICS andexperimental HICS for eight proteins between 6 and 27 kDa as afunction of the number of residues separating like charges.

Analytical Chemistry Article

dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXXC

proteins and the ESI-MS spectrum of each, also acquired underdenaturing conditions. Basic residues are highlighted in red, andgroupings of nearby basic residues are underlined assuming atleast three uncharged amino acids between each charged aminoacid. The amino acids marked with a caret (∧) indicate chargedamino acids using the simplified charge assignment strategydescribed in the previous paragraph. Bear in mind that thismethod does not provide the means to predict which residue ina group is most likely to be protonated during an actual ESIexperiment, only how many charges will be present in thehighest intensity charge state. For negative mode examples,please see the Supporting Information. Sequence informationfor each protein and a comparison of predicted versusexperimental HICS is presented in Table 1.The HICS for 11 proteins in the positive mode spanning a

mass range of 6−79 kDa were estimated by inspection of theamino acid sequence and compared to their experimentallyobserved HICS. The correlation plot in Figure 4 shows goodagreement between the predicted and experimental HICS forthe eight proteins with molecular masses between 6 and 27kDa. The R2 value for the correlation in this region is 0.94. Forcomparison, a molecular weight fit yields R2 = 0.63 (SupportingInformation Figure S-4). Similarly, the HICS was estimated towithin an average of 1.6 charge states of the experimental valuesfor several of these proteins in the negative ion mode (Figure 4,inset).

The experimental HICS falls short of the predicted HICS forthe proteins in the set with the highest masses: BSA (66 kDa)and apotransferrin (79 kDa). A similar observation was madeby Loo et al. while investigating the correlation between theHOCS and the number of basic residues of a protein.6 In theirstudy, the HOCS for proteins with masses 36 kDa or less wereaccurately predicted, while the HOCS for proteins 66 kDa orlarger were observed below their predicted values. After thelarger proteins were treated with 1,4-dithiothreitol (DTT), areagent used to reduce disulfide linkages, the experimentalHOCS were increased significantly, although they remainedbelow the predicted values. Their experiment demonstrates theimportance of a protein’s complete unfolding in bothminimizing the Coulombic repulsion between like chargessupported on the same protein molecule and providing solventaccess to previously inaccessible areas of the protein. Similarly,we believe that the extent of multiple charging for the largeproteins investigated during our experiment was less thanpredicted because of incomplete denaturing, due either todisulfide linkages creating conformational inflexibility orresidual tertiary structure. Parts e and f of Figure 3 show themass spectrum of BSA before and after treatment with DTT.After treatment, the HICS of BSA shifted from +45 to +58, a39% increase in charging. Although the HICS is still less thanthe predicted HICS of +77, this represents a significant increasein multiple charging. BSA has a unique network of disulfidelinkages: of the 17 disulfide linkages in BSA, 16 form 8 pairs of

Table 1. Information and Experimental Results for the Proteins Investigated in the Current Studya

positive ion mode HICS negative ion mode HICS

protein sequence IDb mass (kDa) predicted exptl predicted exptl

1 aprotinin ref 36 6.6 7 6 4 32 ubiquitin 444791 8.6 11 11 8 73 cytochrome c NP_001039526.1 12.3 14 15 7 104 hemoglobin (α chain) NP_001070890.2 15.1 21 19 11 125 WDP-5,6 16.0 14 136 myoglobin NP_001157488.1 17.0 20 20 16 147 α-chymotrypsinogen 2CGA_A 25.7 15 168 GFP 26.9 24 289 xanthine alkaloid methyltransferase 42.6 38 4710 bovine serum albumin CAA76847.1 66.4 68 4511 human apotransferrin 2HAU_A 79 70 36

aThe data for human apotransferrin was acquired from the literature (ref 1) where the spectra had been obtained from denaturing solutions.bSequences for proteins with “” can be found in the Supporting Information (Figure S-1).

Figure 2. (a) Amino acid sequence of equine myoglobin with the basic residues lysine (K), arginine (R), and histidine (H), as well as the N-terminus, highlighted in red. The numbers and brackets indicate the charge assignments to the charge groupings, which are defined in the text. (b)Possible charge assignments for the residues K67−H72 (charge assignments 10 and 11).

Analytical Chemistry Article

dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXXD

adjacent linkages (Supporting Information Figure S-5). It ispossible that treatment with DTT did not result in completereduction of all the disulfide bonds for BSA. For humanapotransferrin, Thevis et al. observed an increase of the HICSfrom +36 to +89 upon treatment with DTT,1 bringing theHICS to within 11% of the predicted value. This observationmay make it possible to use this method to estimate the degreeof conformational flexibility of proteins based on the differencebetween their predicted and observed HICS and couldpotentially serve as a rapid quality control measure.Our results might also have implications for the origin of the

charge state distribution. It is expected that most conformations

would not resemble the linear chain implied by our model.Charge states lower than the HICS might arise from proteinconformations where fewer potential protonation sites (basicamino acids) are solvent-accessible and the distances betweensome of the exposed potential protonation sites are reduced,due to preserved protein secondary or tertiary structure, whichwould decrease the GBapp of the protein ion. Ion mobilitystudies have shown that protein ions with lower charge statesexhibit more compact formations in the gas phase than thosewith higher charge states.33,34 Charge states higher than that ofthe HICS occur when the number of charges exceeds thenumber of the charging groups. These charge states represent

Figure 3. Spectra for several proteins acquired under denaturing conditions (50/50/0.1 MeOH/H2O/formic acid). Below each spectrum is arepresentation of the respective amino acid sequence, where basic residues are highlighted in red and charge groupings are underlined(representations for all proteins are provided in the Supporting Information, Figures S-2 and S-3). The values in parentheses indicate the ratio of thenumber of groupings to the total number of basic amino acids. For each protein shown, with the exception of BSA, the HICS corresponds to withinone or two charges to the number of groupings of charge sites.

Analytical Chemistry Article

dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXXE

ions where one or more groupings of nearby basic residues onthe amino acid chain contain more than one charged residue.Recall that the HOCS is usually several charge states higherthan the HICS and is determined by the highest availablecharge state with a GBapp higher than that of the solvent.Charge states higher than the HICS are still accessible based ondifferences between the GBapp of the protein and the GB of thesolvent up to the HOCS. Their intensity is reduced relative tothe HICS due to the diminishing difference between the GBappof the charge states and that of the solvent as well as theadditional Coulombic repulsion between nearby like chargesand thus decreased interaction potential8 between charges andpotential charge sites. Gross and Williams determined that theeffect of existing charges on the interaction potential of possiblecharge sites occurs for distances ≤10 Å.8 There is evidence thatfor higher charge states, residues other than lysine, histidine, orarginine can also become charged,7,24,35 which could alsoexplain charge states higher than the HICS, or even the HOCS.

■ CONCLUSIONThe simple count of groupings of nearby basic or acidic aminoacid residues in protein sequences accurately predicted thehighest intensity charge state (HICS) for the majority ofdenatured proteins used in the present study. For thoseproteins whose HICS were not accurately predicted by thismethod, we have shown that this is likely due to someconformational inflexibility of these proteins.This method does not account for all energy contributions

arising from intramolecular interactions or from residualsecondary and tertiary structure, which could decrease thedistance between charges as well as the GBapp of the protein.The accuracy of the model may be improved by considering theactual amino acid residue sequence in each group because ofthe differences in the lengths of the basic residue side groups aswell as their respective gas-phase basicities, topics for furtherinvestigation. However, these considerations would increase the

complexity of the model and begin to resemble some of thesophisticated computational techniques already devel-oped,11,12,27 detracting from the aim of the present work toprovide a fast and easy estimation of the HICS.We foresee several uses for the method presented here. The

prediction of the HICS allows one to determine how much ofthe charge envelope would fall within the mass range of theanalyzer and might prove useful in designing quantitativeprotein analyses. This method also provides informationregarding a protein’s conformation, or degree of conformationalflexibility, based on the agreement between the expected andexperimental HICS. While the method presented in the currentstudy can only be applied to denatured proteins with knownsequences, an increasingly large set of proteins have sequencedata that are publicly available through protein sequencedatabases such as the NCBI or PDB.30,31

■ ASSOCIATED CONTENT*S Supporting InformationAdditional information as noted in text. This material isavailable free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: andre.venter@wmich.edu.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe authors thank Dr. David Huffman and Dr. Todd Barkman,WMU, for providing several protein samples. This work wassupported by the WMU Faculty Research and CreativeActivities Award (FRACAA W2012-049).

■ REFERENCES(1) Thevis, M.; Loo, R. R. O.; Loo, J. A. J. Am. Soc. Mass Spectrom.2003, 14, 635−647.(2) Kebarle, P.; Verkerk, U. H. Mass Spectrom. Rev. 2009, 28, 898−917.(3) Kaltashov, I. A.; Abzalimov, R. R. J. Am. Soc. Mass Spectrom. 2008,19, 1239−1246.(4) Chowdhury, S. K.; Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1990,112, 9012−9013.(5) Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1991, 113, 8534−8535.(6) Loo, J. A.; Edmonds, C. G.; Udseth, H. R.; Smith, R. D. Anal.Chem. 1990, 62, 693−698.(7) Smith, R. D.; Loo, J. A.; Loo, R. R. O.; Busman, M.; Udseth, H. R.Mass Spectrom. Rev. 1991, 10, 359−452.(8) Gross, D. S.; Williams, E. R. J. Am. Chem. Soc. 1995, 117, 883−890.(9) Grandori, R. J. Mass Spectrom. 2003, 38, 11−15.(10) Fernandez de la Mora, J. Anal. Chim. Acta 2000, 406, 93−104.(11) Marchese, R.; Grandori, R.; Carloni, P.; Raugei, S. J. Am. Soc.Mass Spectrom. 2012, 23, 1903−1910.(12) Schnier, P.; Gross, D.; Williams, E. J. Am. Soc. Mass Spectrom.1995, 6, 1086−1097.(13) Kaltashov, I. A.; Mohimen, A. Anal. Chem. 2005, 77, 5370−5379.(14) Peschke, M.; Blades, A.; Kebarle, P. J. Am. Chem. Soc. 2002, 124,11519−11530.(15) Mack, L. L.; Kralik, P.; Rheude, A.; Dole, M. J. Chem. Phys.1970, 52, 4977−4986.(16) Rayleigh, L. Philos. Mag. 1882, 14, 184−186.(17) Samalikova, M.; Grandori, R. J. Mass Spectrom. 2005, 40, 503−510.

Figure 4. Correlation plot of the predicted and experimental HICS fordenatured proteins for both the positive and negative (inset) modes.The dashed line represents a perfect agreement between the predictedand experimental data. Green triangles indicate observed charging forBSA (66 kDa) and human apotransferrin (79 kDa) after treatmentwith DTT.

Analytical Chemistry Article

dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXXF

(18) Samalikova, M.; Grandori, R. J. Am. Chem. Soc. 2003, 125,13352−13353.(19) Lomeli, S. H.; Yin, S.; Ogorzalek Loo, R. R.; Loo, J. A. J. Am. Soc.Mass Spectrom. 2009, 20, 593−596.(20) Covey, T. R.; Bonner, R. F.; Shushan, B. I.; Henion, J.; Boyd, R.K. Rapid Commun. Mass Spectrom. 1988, 2, 249−256.(21) Sterling, H. J.; Cassou, C. A.; Susa, A. C.; Williams, E. R. Anal.Chem. 2012, 85, 138−146.(22) Konermann, L. J. Phys. Chem. B 2007, 111, 6534−6543.(23) Wu, Z.; Fenselau, C. J. Am. Soc. Mass Spectrom. 1992, 3, 863−866.(24) Schnier, P. D.; Gross, D. S.; Williams, E. R. J. Am. Chem. Soc.1995, 117, 6747−6757.(25) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287−2294.(26) Ahadi, E.; Konermann, L. J. Phys. Chem. B 2011, 116, 104−112.(27) Konermann, L.; Rodriguez, A. D.; Liu, J. Anal. Chem. 2012, 84,6798−6804.(28) Konermann, L.; Ahadi, E.; Rodriguez, A. D.; Vahidi, S. Anal.Chem. 2013, 85, 2−9.(29) Testa, L.; Brocca, S.; Grandori, R. Anal. Chem. 2011, 83, 6459−6463.(30) National Center for Biotechnology Information. www.ncbi.nlm.nih.gov (accessed March 2013).(31) RCSB Protein Data Bank. www.rcsb.org (accessed March 2013).(32) Douglass, K. A.; Venter, A. R. J. Am. Soc. Mass Spectrom. 2012,23, 489−497.(33) Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F. J. Am. Chem. Soc.1995, 117, 10141−10142.(34) Shelimov, K. B.; Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F.J. Am. Chem. Soc. 1997, 119, 2240−2248.(35) Kaltashov, I. A.; Fenselau, C. C. J. Am. Chem. Soc. 1995, 117,9906−9910.(36) Kassell, B.; Radicevic, M.; Ansfield, M. J.; Laskowski, M., Sr.Biochem. Biophys. Res. Commun. 1965, 18, 255−258.

Analytical Chemistry Article

dx.doi.org/10.1021/ac401245r | Anal. Chem. XXXX, XXX, XXX−XXXG