Analysis of non-covalent protein complexes upto 290 kDa using electrospray ionization andion trap mass spectrometry

Yang Wang*†, Michael Schubert, Arnd Ingendoh and Jochen FranzenBruker-Daltonik GmbH, Fahrenheitstrasse 4, 28359 Bremen, Germany

SPONSOR REFEREE: Professor R. E. March, Trent University, Peterborough, Ontario, Canada

Non-covalently-bound subunit complexes of proteins have been measured by an ion trap mass spectrometerequipped with an orthogonal electrospray ionization source. For the analysis of the generated molecularions with high mass/charge ratios, the mass/charge range of the ion trap was extended by increasing its radiofrequency (rf) voltage to 15 kV (V0ÿp) and by resonant ion ejection. Ions of the non-covalent dimer of bovineserum albumin (BSA), as well as of subunit complexes of alcohol dehydrogenase (ADH) from bakers’ yeastand from horse liver, have been detected at mass/charge values between 3000–9000 Th. The maximumobserved molecular weight was that of a non-covalently-bound subunit-octamer of bakers’ yeast ADH (twonon-covalently-bound subunit-tetramers) at ca. 290 kDa. Copyright# 2000 John Wiley & Sons, Ltd.

Received 5 November 1999; Accepted 7 November 1999

Electrospray ionization mass spectrometry (ESI-MS) hasbecome an important analytical technique for the character-ization of biomolecules. With ESI, not only has the massrange for large biomolecules been extended,1–3but, as a softionization technique, ESI has also provided a means forstudying the tertiary structure of proteins as well as non-covalent interactions between biological macromolecularspecies. These include host-guest interactions like enzyme-substrate, receptor-ligand and protein-nucleic binding.

ESI-MS was first applied for investigation on non-covalently-bound complexes by the groups of Henion4,5andof Chait.6 Thus far, non-covalently-bound complexes suchas receptor-ligand,4 peptide-peptide,7 protein-protein inter-action8 and oligonucleotide associations9 have been studiedby ESI-MS. This technique has shown tremendous utilityfor the determination of the quaternary structure of proteins.A comprehensive list of applications on non-covalently-bound complexes by ESI-MS may be found in the reviewsby Loo.10–12

As those samples are stable only under near-neutral pHconditions, they must be sprayed from buffered solutions ofpH.6–7 Consequently, the charge states of the ions producedare much lower than under the usual acidic conditions.Therefore, a mass analyzer with an extended mass/chargerange (>3000 Th) is required.

Ion trap mass spectrometry has been developed to be apowerful analytical technique with high sensitivity, fastscan speed and MSn capability. The theory of operation andperformance enhancements have been described exten-sively in reviews by March and Todd.13,14For the analysisof ions with large mass/charge ratios, the development of

resonant ion ejection, the use of smaller trap electrodes andoperation at lower rf frequencies were essential.15–17 Inearly experiments, a Cs� Secondary Ion Mass Spectrometry(SIMS) source was used to generate large CsI cluster ionswhich were detected in an ion trap. Schluneggeret al.recently described the use of low frequencies to trap andanalyze high mass singly charged ions generated by matrix-assisted laser desorption/ionization in an ion trap.17

Coupling an ESI source to an ion trap mass spectrometerwas first reported by Van Berkel, Glish and McLuckey.18

The general application aspects of a commercial ESI iontrap were described later by Bier and Schwartz.19

The first studies on non-covalent complexes with an iontrap were ion/ion reactions of protein mixtures, reported byStephenson and McLuckey.20–23 Yost et al. have studiedinteractions between cytochrome c and its peroxidase.24 Inthese investigations, the resonant ion ejection method wasapplied in order to achieve extension of the mass/chargerange of the ion trap.

Here, an ion trap mass spectrometer with an orthogonalESI source was used to analyze large biomolecules thatconsisted of non-covalently-bound subunits. The mass/charge range extension was achieved by increasing themaximum rf voltage of the ion trap to 15 kV (V0ÿp) and byresonant ion ejection. Commercial ion traps are driven withmaximum rf voltages of 8.5 kV (V0ÿp)

19 or 12 kV (V0ÿp).21

The operating frequency and the size of the ion trap usedhere remained unchanged. The value of the trappingparameter qz of the trap was kept as large as possible, incontrast to previous studies.16,21,22

The analyzed biomolecules were bovine serum albumin(BSA) and alcohol dehydrogenase (ADH) from bakers’yeast and from horse liver.

EXPERIMENTAL

Mass spectrometer

A modified Bruker-Daltonics/Hewlett-Packard ESQUIRE-

*Correspondence to: Y. Wang, Center for Interdisciplinary MagneticResonance, National High Magnetic Field Laboratory, 1800 East PaulDirac Drive, Florida State University, Tallahassee, FL 32310, USA.E-mail: yangwang@magnet.fsu.edu†Current address: Center for Interdisciplinary Magnetic Resonance,National High Magnetic Field Laboratory, 1800 East Paul Dirac Drive,Florida State University, Tallahassee, FL 32310, USA.

CCC 0951–4198/2000/010012–06 $17.50 Copyright# 2000 John Wiley & Sons, Ltd.

RAPID COMMUNICATIONS IN MASS SPECTROMETRYRapid Commun. Mass Spectrom.14, 12–17 (2000)

LC ion trap massspectrometer was used (Fig. 1). Thenebulizer needleof the ESI source is directedorthogonallyto themassspectrometeraxis.Thesampleflow ratethroughthis grounded needle was 6 mL/min with infusion and thepressureof thenebulizing nitrogen gas30 psi.Thecounter-currentdrying gaswasheld at a flow rateof 5 L/min, at arelatively low temperature of 120°C to avoidthermal stressof theweaknon-covalentbonds.Thepotentials on theend-plate and the non-heated glass capillary entrance wereÿ3600 andÿ4000V, respectively. The potentials on theglass capillary exit and the first and secondskimmerswere 140, 25 and 6 V, respectively. The two octapole ionguides wereoperatedin rf-only modewith anamplitude of100V0ÿp at a frequency of 2.5MHz. Theoffset DC poten-tials of the first andsecond octapoleswere3.5 and2.5V,

respectively. Betweenthesecond octapoleandtheentrancecapof the ion trap,therearetwo lensesheldatÿ5 (lens1)andÿ40V (lens2). To reducethekinetic energyof theionswhen theyenterthetrap,thetrapis filled with heliumgasata pressureof about5� 10ÿ3 mbar. The ion accumulationtime was10–30ms.

A non-linearion trapwith modified angle geometry25–27

wasusedfor the resonant ion ejection.The ring electrodehadro = 10mm, andthefrequency of theappliedrf voltagewas781kHz. With themodifiedangle geometry of thetrapelectrodes, additionalweaknon-linearfieldsweregeneratedandsuperimposed on the main quadrupole field in the iontrap. Theseconditions leadto non-linearresonances for theionsat a number of secular frequencies.26 Theexistenceofthese resonances has been verified experimentally in

Figure 1. Schematicdiagramof the ESI ion trapmassspectrometer.

Figure 2. ESI massspectrumandmolecular-weightdeconvolutionof BSA, SeriesA: non-covalentBSA dimer,seriesB: monomer.The result of the deconvolution of the multiply chargedion peaksshowsin the inset themolecularweightsobtainedfor the BSA monomeranddimer.

Copyright# 2000JohnWiley & Sons,Ltd. Rapid Commun.MassSpectrom.14, 12–17(2000)

NON-COVALENT PROTEIN COMPLEXES BY ESI ION TRAP MS 13

detail.28 Non-linear resonances areparticularly effectiveincoupling additional energy into the ions when the secularfrequencyof theion oscillation is 1/n(n> 1, integers)of therf frequency. To initiate theion ejection,anauxiliarydipolarfield is added by applying an AC voltage of a fixedfrequencyto thetwo end-capsof thetrap.Theamplitudesofboth the quadrupolarandthe dipolar fields arerampedup.Whenthefundamentalsecularion frequency approachesthefrequency of the dipolar field, the ion oscillation isamplified extremelyrapidly and the ions are ejected. Byphase-locking the frequencies of the quadrupolar and thedipolar fields, reproducible mass assignments are en-sured.29,30In ourexperiments,thefrequencyof theauxiliarydipolar AC voltage was 71kHz (781kHz/11), whichcorrespondsto a non-linear resonanceat bz = 2/11. Themaximum detectable mass/charge value in the trap thendepends mainly on the maximum amplitude of thequadrupolar rf voltage. With a maximum rf voltage of15kV (V0ÿp), themass/charge rangewasextendedto 9000Th. Themass/chargescanratewas12000Th/s.The ion areejectedfrom the trapwith increasing molecular weightanddetected in a high energy detector(HED) consistingof ahigh potential conversion electrode (ÿ10 kV) and asecondaryelectronmultiplier (1.6kV).

Sample preparation

The samples of bovine albumin (BSA) and alcoholdehydrogenase (ADH) from both bakers’yeastand horseliver were bought from Sigma and used without furtherpurification. The samplepreparation is critical since thethree-dimensionalstructureof theproteinmustbepreserved

in the solution. As thesenon-covalently-boundmoleculesare stableonly undernear-neutral pH conditions,they aredissolvedin deionizedwaterbufferedwith 10mM ammo-niumacetate (pH 6.5).Thesampleswerepreparedandusedfreshly for one day only. Sample concentration was10pmol/mL for all experiments.

Data processing

The high mass/chargerangeof the ion trap massspectro-meter was calibrated by CsI ion clusters or by singlychargedmellitin ([M � H]� = 2847.5Da), insulin ([M �H]� = 5734.6Da) and ubiquitin ([M � H]� = 8565.9Da).All data were processed by the Bruker-DaltonicsDataAnalysis softwareprogram. The massspectraobtainedrepresent the averagesof 20 scans. The raw data weresmoothed, and the molecularweights were determinedbydeconvolution of the multiply charged ion peaks. Thedeconvolution programusedwasfrom Hewlett-Packard. Inaddition to the instrument calibration described above,themassspectrawererecalibratedafter dataacquisition usingthe data obtained from the BSA monomer as externalstandard. The reasonfor this re-calibration is discussedbelow.

RESULTS AND DISCUSSION

Non-covalently-bound bovine serum albumin (BSA)dimer

Figure 2 shows the ESI mass spectrum of BSA withmultiply chargedion signalsof both the non-covalently-

Figure 3. ESI massspectrumandmolecular-weightdecovolution of ADH from horseliver, SeriesA: non-covalentdimer,seriesB: ADH subunit.Theresultof thedeconvolutionof themultiply chargedion peaksis shownin theinset.Themolecularweightsobtainedare80001Da(re-calibratedmass:79859Da) for thenon-covalenthorseliver ADHdimer and40006Da (re-calibratedmass:39916Da) for its subunit.

Rapid Commun.MassSpectrom.14, 12–17(2000) Copyright# 2000JohnWiley & Sons,Ltd.

14 NON-COVALENT PROTEIN COMPLEXES BY ESI ION TRAP MS

bound dimer and the monomer, labeled A and B,respectively. The multiple-chargestatesare also labeled,e.g.(A � 25) is the chargestate 25 of the BSA dimer. Theresult of the deconvolution of the multiply chargedions isillustratedin theinsetof Fig. 2. Thedeconvolution programcanautomatically or manually find themultiply chargedionpeaksabovea user-defined ion-signalintensity in themassspectrum. Then,a deconvolution algorithm wasappliedtoobtain the molecular weight. The measured molecularweights are 133029Da for the non-covalent BSA dimer,and 66474 Da for the monomer, using the instrumentcalibration from peptidestandards.

Alcohol dehydrogenase(ADH) from horse liver

ADH from horse liver consists of two identical, non-covalently-bound subunitswith molecular weights of 39847Da each. The ESI mass spectrum in Fig. 3 showsmultiply charged ion signals of the intact ADH subunit-dimer as well as thoseof the dissociation into the singlesubunit (labeledseriesA andB, respectively). It is not clearwhetherthisdissociation occursin themassspectrometerorin the samplesolution. The result of the deconvolution isshown in the inset of Fig. 3, revealing an observedmolecular weight of 80001Da for the ADH subunit-dimerand of 40006Da for the subunit, using the instrumentcalibration from peptide standards. The re-calibratedmol-ecular weights, obtained using BSA as external massstandard, were 79859Da for the ADH subunit-dimer and39916Da for the subunit. The observation that the ionsignal intensityof seriesA is higherthanthat of the singlesubunit signal (SeriesB, Fig. 3) is a strongindication thatthe observed dimer is a real native-boundcomplex. In

contrast,the BSA dimer, which is only a cluster formedinsolution, hasa signal intensity lower thanthat of the BSAmonomer.

Alcohol dehydrogenase(ADH) from bakers’ yeast

ADH from bakers’ yeastconsists of four identical, non-covalently-bound subunits with a molecular weight of36814Daeach.In Fig. 4, multiply chargedion peaksof thedissociatedsubunit andthesubunit-dimer, aswell asof theintactADH subunit-tetramer areshown,labeledseriesA, Band C, respectively. The deconvolution of the multiplycharged ion peaks (series C) results in an observedmolecularweight of 147513Da for the tetramer of ADHfrom bakers’ yeast, asshown in the insetof Fig. 4 for theinstrument calibrated with peptide standards.The re-calibrated molecular weight is 147337Da. Moreover,multiply charged subunit-octamer ion peaksof the ADH,i.e. of a dimer of the native subunit-tetramer, are alsoobserved (Fig. 5, chargestatesfrom �39 to �42). Thedeconvolution results in an observedmolecular weight of296720Da for this subunit-octamer with the instrumentcalibratedusing peptides(shownin theinsetof Fig. 5). There-calibratedmolecular weight is 295603Da.

Table 1 summarizesthe observedmolecular weights forall threeproteins.Md is themolecularweightdeterminedonthebasis of a calibrationof the instrument with thepeptidemixture, and Mc representsthe values after external re-calibration of the obtained spectrawith the knownmassofBSA. The theoretical massesareobtained from Ref.31 andfrom calculations using the protein sequencesavailablefrom theNCBI database.ADH from bakers’yeastandhorseliver havepreviously beenmeasured by a magnetic sector

Figure 4. ESImassspectrumof bakers’yeastADH (rawdatawithoutanypost-processingsuchassmoothing)showingmultiply chargedion signalsof the subunit(seriesA), a non-covalentsubunit-dimer(seriesB) and the intact non-covalentsubunit-tetramer(seriesC). Theresultof themolecular-weightdeconvolutionof theseriesC is shownin theinset.Themolecularweightobtained for thenon-covalent tetrameris 147513Da (re-calibratedmass:147337Da).

Copyright# 2000JohnWiley & Sons,Ltd. Rapid Commun.MassSpectrom.14, 12–17(2000)

NON-COVALENT PROTEIN COMPLEXES BY ESI ION TRAP MS 15

instrumentwith an ESI source.31 ADH from baker’s yeastwas also analyzed by an ESI triple quadrupole massspectrometer.32 All results are listed in Table 1 forcomparison.The molecular weights obtained here and inprevious work are very similar. The re-calibrated massesagree betterwith the theoretical valuesthanthoseobtainedusing theinstrumentcalibrationfrom peptidestandards.Thesubunit-dimerof ADH from bakers’ yeastwasnot reportedby Loo,31 but was observedby Caprioli,33–35 and wasobserved herealso.Again, it is not clearherewhetherthisdissociation occurs in the mass spectrometer or in thesample solution. The dimer of the non-covalent bakers’yeastsubunit-tetramer at 290 kDa hasnot beenobservedbefore, to our knowledge.

DISCUSSION

The useof anorthogonal ESI sourcesprayer doesnot causeany problems for the generation and detection of non-covalently-boundcomplex ions. For purposesof compari-son,all proteinshavealsobeenmeasuredusinganAnalytica

of BranfordESI sourcein which thenebulizer needle is on-axis to themassspectrometer.Theresults obtained wereinall casescomparable. Thenon-covalentbondsalsosurvivedtherelativelyhigh He pressureandtheelevatedscanning rfamplitudesapplied here.

The sample preparation and the applied ionizationparameters wereof primary importancefor theobservationof the non-covalent interactions by ESI ion trap massspectrometry. Thetemperatureof thecounter-currentdryinggaswaskeptat a low value to avoidthermal degradation ofthesamples. Thecapillary-exit andskimmer voltageswereadjustedin suchawayasto minimizethedissociationof theweaknon-covalent bonds.On the other hand,thesesourceparameters hadto beoptimized in orderto obtaina narrowmassspectral peakwidth. Themeasuredwidth (FWHM) ofthe multiply charged ion peakswas about 50Th, so theresolution obtainedis higher thanpreviouslyshownusingamagneticsectorinstrument.25 It is alsoremarkablethat themassaccuracyis sostronglyaffectedby theseexperimentalconditions. Loo et al. have calculated that a tetramericproteincomplexof over1300residuesmayhaveattachedas

Figure 5. High mass/chargesectionof theESImassspectrumof bakers’yeastshowingmultiply chargedion signalsof a non-covalently-boundsubunit-octamer.The result of the molecular-weightdeconvolutionof the multiplychargedion peaksis shown in the inset. The molecularweight obtainedfor the non-covalently-boundsubunit-octameris 296720Da (re-calibratedmass:295603Da).

Table 1. Summary and comparison of the theoretical and observedmolecular weights of the subunits and non-covalently-boundsubunitcomplexes

SampleTheoreticalaveraged

MW (Da)Md (Da), ObservedMW

(peptidecalibration)Mc (Da), MW after

re-calibrationwith BSA,Valuesobtainedby Loo

(Ref. 31)Valueobtainedby Van

Dorsselaer(Ref. 32)

BSA, monomer 66430 66474 66443BSA, dimer 132860 133029 132868ADH from horseliver, subunit 39847 40006 39916ADH from horseliver, subunit-dimer 79694 80001 79859 80500ADH from bakers’yeast,subunit-tetramer 147262 147513 147337 147900 147523ADH from bakers’yeast,subunit-octamer 294524 296720 295603

Rapid Commun.MassSpectrom.14, 12–17(2000) Copyright# 2000JohnWiley & Sons,Ltd.

16 NON-COVALENT PROTEIN COMPLEXES BY ESI ION TRAP MS

many as 690 water molecules.31,36 Becauseof the lowdrying temperature and the low nozzle-skimmer voltages,thedesolvationof themolecularionsmightnotbecompletein the gasphase.Onepieceof evidencefor this interpreta-tion is that the measuredmass decreasestowards thetheoreticalonewith increasing valuesof these instrumentalparameters. However, as the dissociation of the non-covalentbondsis simultaneouslyenhanced,theparametershave to be adjusted to a best compromise. Also, themacromoleculesare much more highly chargedthan thesingly chargedpeptides or CsI clusterswhich are usedtocalibrate the instrument. The resulting different spacecharge may lead to a massshift and decreasethe massaccuracy. BSA,whichhasmolecular weightandsizesimilarto those of the non-covalent complexes, was measuredunder the samepreparation and ionization conditions andgenerated similar spacecharges in the trap because ofsimilar ion chargestates.As aconsequence, theexternal re-calibrationof themassspectrawith similar compoundsandunder similar conditions improved the accuracy of thedeterminationof themolecular weight.However,evenwiththis re-calibration, the accurate determinationof the intactmassof the largemolecules in the ion trap is still diffi cultandneedsfurther improvement.

CONCLUSIONS

The experiments haveconfirmedthe capability of the ESIion trapmassspectrometer to analyzelargenon-covalently-boundcomplexes.The orthogonal ESI sourcecangenerateand transfer these molecular ions intact into the massanalyzer. Importantparametersfor theintacttransferarethedrying gas temperature and the nozzle-skimmer voltages.DespitethehighHepressureandthehighrf voltageusedforthe analysis of largecomplex ions, the non-covalentbondssurvive in the ion storageand the mass scansteps. Theobtainablemass/chargerangewasextendedto 9000Th byincreasing themaximumrf voltageupto 15kV (V0ÿp). Ionsof non-covalentsubunitcomplexesupto amolecular weightover 290kDa havebeendetected. An external calibrationwith similar compoundscould increase the accuracyof themassdetermination of those largeprotein complexes.

Acknowledgements

We thankAnnetteVoigt for her assistancewith the experimentsandPaul Goodley (Hewlett-PackardCo., Palo Alto, CA, USA) for hishelpful discussion.In addition, we thank Alain Van DorsselaerandEmmanuelleLeize (Louis PasteurUniversity,Strasbourg,France)forinitiating andsupportingthe investigationson complexproteins.

REFERENCES

1. FennJB,MannM, MengCK, WongSF,WhitehouseCM. Science1989;246: 64.

2. Smith RD, Loo JA, EdmondsCG, BaringeCJ,UdsethHR. Anal.Chem.1990;62: 882.

3. FengR, Konishi Y. Anal. Chem.1992;64: 2090.4. GanemB, Li YT, HenionJD.J. Am.Chem.Soc.1991;113: 6294.5. GanemB, Li YT, HenionJD.J. Am.Chem.Soc.1991;113: 7818.6. Katta V, Chait BT. J. Am.Chem.Soc.1991;113: 8534.7. Loo JA, HolsworthDD, Root-BernsteinRS.Biol. MassSpectrom.

1994;23: 6.8. GoodlettDR, OgorzalekLoo RR, Loo JA, Wahl JH, UdsethHR,

Smith RD. J. Am.Soc.MassSpectrom.1994;5: 614.9. light-WahlKJ,SpringerDL, WingerBE,EdmondsCG,CampDG,

III Thrall BD, Smith RD. J. Am.Chem.Soc.1993;115: 803.10. Loo JA. MassSpectrom.Rev.1997;16: 1.11. Loo JA, Ogorzalek Loo RR. In Electrospray lonization Mass

Spectrometry, Cole RB. (ed).JohnWiley: New York 1997;385.12. Loo JA, Sannes-LoweryKA. In MassSpectrometryof Biological

Materials, LarsenBS, McEwen CN. (eds).Marcel Dekker, Inc:New York, 1998;345.

13. March RE, Hughes RJ, Todd JFJ. Quadrupole Storage MassSpectrometry, Winefordner(ed).JohnWiley: New York, 1989.

14. PracticalAspectsof Ion TrapMassSpectrometry, MarchRE,ToddJFJ(eds).CRCPress:BocaRaton,FL, 1995;1–3.

15. KaiserRE Jr., Louris JN, Amy JW, CooksRG. Rapid Commun.MassSpectrom.1989;3: 225.

16. KaiserRE Jr., CooksRG, StaffordJr. GC, SykaJEP.HembergerPH. Int. J. MassSpectrom.Ion Processes1991;106: 79.

17. SchluneggerUP,StoeckliM, Caprioli RM. RapidCommun.MassSpectrom.1999;13: 1792.

18. Van Berkel GJ,Glish GL, Mcluckey SA. Anal. Chem.1990;63:1284.

19. Bier ME, SchwartzJC. In ElectrospraylonizationMassSpectro-metry, Cole RB (ed).JohnWiley: New York, 1997;235.

20. StephensonJL Jr.,Mcluckey SA. Anal. Chem.1998;70: 3533.21. Mcluckey SA, StephensonJL Jr., AsanoKG. Anal. Chem.1998;

70: 1198.22. StephensonJL Jr., Mcluckey SA. J. Am. Soc. Mass Spectrom.

1998;9: 585.23. StephensonJL Jr., Mcluckey SA. J. Am. Soc. Mass Spectrom.

1998;9: 957.24. RodriguezA, Williams TL, Yost RA, In Proc. 45th ASMSConf.

MassSpectrometryand Allied Topics, Palm Springs,CA, 1997;1411.

25. Wang Y, FranzenJ. Int. J. MassSpectrom.Ion Processes1992;112: 167.

26. Wang Y, FranzenJ, WanczekK-P. Int. J. Mass Spectrom.IonProcesses1993;124: 125.

27. Wang Y, FranzenJ. Int. J. MassSpectrom.Ion Processes1994;132: 155.

28. Alheit R, KleineidamS,VedelF, WerthG. Int. J. MassSpectrom.Ion Processes1996;154: 155.

29. WangY, SchubertM, FranzenJ. In Proc. 44th ASMSConf.MassSpectrometryAllied Topics, Portland,OR, 1996;131.

30. LondryFA, MarchRE. Int. J. MassSpectrom.Ion Processes1995;144: 87.

31. Loo JA. J. MassSpectrom.1995;30: 180.32. Van DorsselaerA, PotierN, Barth P, Trish D, BiellmannJ-F. In

Proc.44th ASMSConf.MassSpectrometryAllied Topics, Portland,OR, 1996;1407.

33. FarmerTB, Caprioli RM. Biol. MassSpectrom.1991;20: 796.34. SuterMJ -F, MooreWT, FarmerTB, Cottrell JS,Caprioli RM. In

Techniquesin ProteinChemistryIII , AngelettiRH (ed).AcademicPress:SanDiego 1992;447.

35. FarmerTB, Caprioli RM. In Techniquesin ProteinChemistryIV,Angeletti RH (ed).AcademicPress:SanDiego,1993;73.

36. Kuhn LA, Siani MA, PiqueME, Tisher CL, Getzoff ED, TainerJA. J. Mol. Biol. 1992;228: 13.

Copyright# 2000JohnWiley & Sons,Ltd. Rapid Commun.MassSpectrom.14, 12–17(2000)

NON-COVALENT PROTEIN COMPLEXES BY ESI ION TRAP MS 17