1 Mass Spectrometry of Amino Acids and Proteins

50 Pages·2011·1.41 MB·English

j
1
1
Mass Spectrometry of Amino Acids and Proteins
SiminD.MalekniaandRichardJohnson
1.1
Introduction
1.1.1
MassTerminology
Likemostmatter(withtheexceptionof,say,neutronstars),proteinsandpeptidesare
mostly made of nothing – an ephemeral cloud of electrons with very little mass
surroundingtinyandverydenseatomicnucleithatcontainnearlyallofthemass(i.e.,
peptidesandproteinsaremadeofatoms).Atomshavemassandtheunitofmassthat
ismostconvenienttouseiscalledtheatomicmassunit(abbreviatedamuoru)orin
biologicalcirclesaDalton(Da).Overtheyears,physicistsandchemistshaveargued
aboutwhatstandardtousetodefineanatomicmassunit,buttheissueseemstohave
beensettledin1959whentheGeneralAssemblyoftheInternationalUnionofPure
andAppliedChemistrydefinedanatomicmassunitasbeingexactly1/12ofthemass
of the most abundant carbon isotope (12C) in its unbound lowest energy state.
Therefore,oneatomof12Chasamassof12.0000u.Usingthisasthestandard,one
proton has a measured mass of 1.00728u and one neutron is slightly heavier at
1.00866u.One12Catomcontainssixprotonsandsixneutrons,thesumofwhichis
clearlymorethanthemassof12.0000u.Acarbonatomislessthanthesumofits
parts,andthereasonisthattheprotonsandneutronsinacarbonnucleusareina
lower energy state than free protons and neutrons. Energy and mass are inter-
changeableviaEinstein’sfamousequation(E¼mc2),andsothis“massdefect”isa
resultofthenuclearforcesthatholdneutronsandprotonstogetherwithinanatom.
This mass defect also serves as a reminder of why people like A. Q. Khan are so
dangerous[1].
Eachelementisdefinedbythenumberofprotonspernucleus(e.g.,carbonatoms
alwayshavesixprotons),buteachelementcanhavevariablenumbersofneutrons.
Elementswithdifferingnumbersofneutronsarecalledisotopesandeachisotope
possessesadifferentmass.Insomecases,theadditionalneutronsresultinstable
isotopes,whichareparticularlyusefulinmassspectrometry(MS)inamethodcalled
AminoAcids,PeptidesandProteinsinOrganicChemistry.Vol.5:AnalysisandFunctionofAminoAcidsandPeptides.
FirstEdition.EditedbyAndrewB.Hughes.
(cid:1)2012Wiley-VCHVerlagGmbH&Co.KGaA.Published2012byWiley-VCHVerlagGmbH&Co.KGaA.
j
2 1 MassSpectrometryofAminoAcidsandProteins
isotope dilution. Examples in the proteomic field that employ isotope dilution
methodologyincludetheuseofthestableisotopes2H,13C,15N,and18O,asapplied
in methods such as ICAT (isotope-coded affinity tags) [2], SILAC (stable isotope
labeling with amino acids in cell culture) [3], or enzymatic incorporation of 18O
water [4]. Whereas some isotopes are stable, others are not and will undergo
radioactive decay. For example, hydrogen with one neutron is stable (deuterium),
but if there are two additional neutrons (a tritium atom) the atoms will decay to
helium (two protons and one neutron) plus a negatively charged b-particle and a
neutrino.Generally,iftherearesufficientamountsofaradioactiveisotopetoproduce
anabundantmassspectralsignal,thesampleislikelytobeexceedinglyradioactive,
theinstrumentationwouldhavebecomecontaminated,andtheoperatorwouldlikely
cometoregrethavingperformedtheanalysis.Therefore,massspectrometristswill
typically concern themselves with stable isotopes. Each element has a different
propensitytotakeondifferentnumbersofneutrons.Forexample,fluorinehasnine
protonsandalways10neutrons;however,brominewith35protonsisevenlysplit
between possessing either 44 or 46 neutrons. There are most likely interesting
reasonsforthis,buttheyarenotparticularlyrelevanttoadescriptionoftheuseofMS
intheanalysisofproteins.
What is relevant is the notion of “monoisotopic” versus “average” versus
“nominal” mass. The monoisotopic mass of a molecule is calculated using the
massesofthemostabundantisotopeofeachelementpresentinthemolecule.For
peptides,thismeansusingthespecificmassesfortheisotopesofeachelementthat
possessthehighestnaturalabundance(e.g.,1H,12C,14N,16O,31P,and32Sasshown
inTable1.1).The“average”or“chemical”massiscalculatedusinganaverageofthe
isotopesforeachelement,weightedfornaturalabundance.Forelementsfoundin
mostbiologicalmolecules,themostabundantisotopecontainsthefewestneutrons
Table1.1 Massandabundancevaluesforsomebiochemicallyrelevantelements.
Element Averagemass Isotope Monoisotopicmass Abundance(%)
Hydrogen 1.008 1H 1.00783 99.985
2H 2.01410 0.015
Carbon 12.011 12C 12 98.90
13C 13.00335 1.10
Nitrogen 14.007 14N 14.00307 99.63
15N 15.00011 0.37
Oxygen 15.999 16O 15.99491 99.76
17O 16.99913 0.04
18O 17.99916 0.200
Phosphorus 30.974 31P 30.97376 100
Sodium 22.990 23Na 22.98977 100
Sulfur 32.064 32S 31.97207 95.02
33S 32.97146 0.75
34S 33.96787 4.21
36S 35.96708 0.02
j
1.1 Introduction 3
and the less abundant isotopes are of greater mass. Therefore, the monoisotopic
masses calculated for peptides are less than what are calculated using average
elementalmasses.Theterm“nominalmass”referstotheintegervalueofthemost
abundantisotopeforeachelement.Forexample,thenominalmassesofH,C,N,and
O are 1, 12, 14, and 16, respectively. A rough conversion between nominal and
monoisotopicpeptidemassesisshownas[5]:
M ¼1:000495(cid:2)M ð1:1Þ
c n
D ¼0:03þ0:02(cid:2)M =1000 ð1:2Þ
m n
whereM istheestimatedmonoisotopicpeptidemasscalculatedfromanominal
c
mass, M . D is the estimated standard deviation at a given nominal mass. For
n m
example,peptideswithanominalmassof1999wouldbeexpected,onaverage,to
haveamonoisotopicmassofaround1999.99withastandarddeviationof0.07u.
Therefore, 99.7% of all peptides (3 standard deviations) at a nominal mass 1999
wouldbefoundatmonoisotopicmassesbetween1999.78and2000.20(Figure1.1).
Figure1.1 Predictingmonoisotopicfrom weightsof1998.99,1999.99,and2000.99with
nominalmolecularweights.Usingthe standarddeviationsof0.07.Thedifference
equationsfromWoolandSmilansky[5], betweenmonoisotopicandnominalmassesis
peptideswithnominalmolecularweightsof calledthemassdefectandthisvaluescaleswith
1998,1999,and2000wouldonaveragebe mass.
expectedtohavemonoisotopicmolecular
j
4 1 MassSpectrometryofAminoAcidsandProteins
As can be seen, at around mass 2000, the mass defect in a peptide molecule is
just about one whole mass unit. Most of this mass defect is due to the large
number of hydrogen atoms present in a peptide of this size. The mass defect
associatedwithnitrogenandoxygentendstocancelout,andcarbonbydefinition
hasnomassdefect.Theotherimportantobservationthatcanbemadefromthis
example is that 99.7% of peptides with a nominal mass of 1999 will be found
between1999.78and2000.20.Therefore,amoleculethatisaccuratelymeasured
tobe2000.45cannotbeastandardpeptideandmusteithernotbeapeptideatall
or is a peptide that has been modified with elements not typically found in
peptides.
1.1.2
ComponentsofaMassSpectrometer
Atminimum,amassspectrometerhasanionizationsource,amassanalyzer,anion
detector,andsomemeansofreportingthedata.Forthepurposeshere,thereisno
needtogointoanydetailatallregardingtheiondetectionandalthoughthereare
manyhistoricallyinterestingmethodsofrecordingandreportingdata(photographic
plates, UV-sensitive paper, etc.), nowadays one simply uses a computer. The ioni-
zationsourceandthemassanalyzerarethetwocomponentsthatneedtobewell
understood.
Historically,ionizationwaslimitedtovolatilemoleculesthatwereamenableto
gasphaseionizationmethodssuchaselectronimpact.Overtime,othertechniques
weredevelopedthatallowedforionizationoflargerpolarmolecules–techniques
such as fast atom bombardment (FAB) or field desorption ionization. However,
thesehadrelativelypoorsensitivityrequiring0.1–1nmolofpeptideand,withthe
exceptionofplasmadesorptionionization–atechniquethatusedtoxicradioactive
californium,weregenerallynotcapableofionizinglargermoleculeslikeproteins.
Remarkably,twodifferentionizationmethodsweredevelopedinthelate1980sthat
didallowforsensitiveionizationoflargermolecules–electrosprayionization(ESI)
andmatrix-assistedlaserdesorptionionization(MALDI).Posterspresentedatthe
1988AmericanSocietyforMassSpectrometryconferencebyJohnFenn’sgroup
showedmassspectraofseveralproteins[6,7],whichrevealedthegeneralnatureof
ESIofpeptidesandproteins.Namely,aseriesofheterogeneousmultiplyproton-
ated ions are observed, where the maximum number of charges is roughly
dependenton thenumber of basicsites in the protein or peptide.Conveniently,
thisputstheionsatmass-to-charge(m/z)ratiostypicallybelow4000, whichisa
rangesuitableforjustaboutallmassanalyzers(seebelow).Inaseriesofpapers
between 1985 and 1988, Hillenkamp and Karas described the essentials of
MALDI[8–10].Also,TanakapresentedaposterataJointJapan–ChinaSymposium
on Mass Spectrometry in 1987 showing a pentamer of lysozyme using laser
desorption from a glycerol matrix containing metal shavings [11]. These early
resultsshowedthegeneralnatureofMALDI –singlychargedionspredominate
andthereforethemassanalyzermustbecapableofmeasuringionswithveryhigh
m/zratios.
j
1.1 Introduction 5
Itisdesirableforuserstohavesomebasicunderstandingofthedifferenttypes
ofmassanalyzersthatareavailable.Atone timemultisectoranalyzers[12]were
well-liked(backwhenFABionizationwaspopular),butquicklybecamedinosaurs
for protein work after the discovery of ESI. It was too difficult to deal with the
electrical arcs that tended to arise when trying to couple kiloelectronvolt source
voltages with a wet acidic atmospheric spray. ESI was initially most readily
coupled to quadrupole mass filters, which operated at much lower voltages.
Quadrupole mass filters [13], as the name implies, are made from four parallel
rods where at appropriate frequency and voltages, ions at specific masses can
oscillatewithoutrunningintoarodorescapingfrombetweentherods.Givena
littlepush(afewelectronvoltspotential)theoscillatingionswillpassthroughthe
lengthoftheparallelrodsandbedetectedattheotherend.Bothquadrupolemass
filters and multisector instruments suffer from slow scan rates and poor sensi-
tivity due to their low duty cycle. Instrument vendors have therefore been busy
developingmoresensitiveanalyzers.Theiontraps[14,15]arelargelygovernedby
the same equations for ion motion as quadrupole mass filters, but possess a
greater duty cycle (and sensitivity). For those unafraid of powerful super cooled
magnets, and who possess sufficiently deep pockets to pay for the initial outlay
and subsequent liquid helium consumption, Fourier transform ion cyclotron
resonance (FT-ICR) provides a high-mass-accuracy and high-resolution mass
analyzer [16]. In this case, the ions circle within a very high vacuum cell under
theinfluenceofastrongmagneticfield.Theoscillatingionsinduceacurrentina
pair of detecting electrodes, where the frequency of oscillation is related to the
m/z ratio. Detection of an oscillating current is also performed in Orbitrap
instruments [17, 18], except in this case the ions circle around a spindle-shaped
electrode rather than magnetic field lines. The time-of-flight hybrid (TOF) mass
analyzer [19, 20] is, at least in principle, the simplest analyzer of all – it is an
emptytube.Ionsareaccelerateddowntheemptytubeand,asthenameimplies,
the TOF is measured and is related to the m/z ratio (big ions move slowly and
little ones move fast).
TandemMSisaconceptthatisindependentofthespecifictypeofmassanalyzer,
butshouldbeunderstoodwhendiscussingmassanalyzers.Asthenameimplies,
tandemMSemploystwostagesofmassanalysis,wherethetwoanalyzerscanbe
scannedinvariouswaysdependingontheexperiment.Inthemostcommontypeof
experiment,thefirstanalyzerisstaticallypassinganionofaspecificmassintoa
fragmentation region, where the selected ions are fragmented somehow (see
below) and the resulting fragment ions are mass analyzed by the second mass
analyzer.Theseso-calleddaughter,orproduct,ionscansareusuallywhataremeant
whenreferringtoan“MS/MSspectra.”However,thereareothertypesoftandem
MS experiments that are occasionally performed. One is where the first mass
analyzerisstaticallypassingaprecursorion(asintheaforementionedproduction
scan)andthesecondanalyzerisalsostaticallymonitoringone,orafew,specific
fragment ions. This so-called selected reaction monitoring (SRM) experiment is
particularly useful in the quantitation of known molecules. There are other less
frequentlyusedtandemMSscans(e.g.,neutrallossscans)anditshouldbenoted
j
6 1 MassSpectrometryofAminoAcidsandProteins
that only certain combinations of specific analyzers are capable of performing
certainkindsofscans.
There are various combinations of mass analyzers used in different mass
spectrometers. One of the more popular has been the quadrupole/TOF hybrid
(quadrupole/time-of-flight hybridQ-TOF) [21], which uses the quadrupole as a
mass filter for precursor selection and the TOF is used to analyze the resulting
fragmentions.Iontrap/time-of-flighthybridsarealsosoldandprovideadditional
stagesoftandemMScomparedtothequadrupole/lineariontraphybrid(Q-trap).
TheQ-TOFhybrid[22]isauniqueinstrumentinthatitcanbethoughtofasatriple-
quadrupoleinstrumentwherethethirdquadrupolecanalternativelybeusedasa
linear ion trap. There is consequently a great deal of flexibility in the types of
experimentsthatcanbedoneonsuchamassspectrometer.ThetandemTOF(TOF-
TOF)[20]isaninstrumentthatallowsacquisitionoftandemmassspectraorsingle-
stage mass spectra of MALDI-generated ions. A timed electrode is used for
precursorselection,whichsweepsawayallionsexceptthosepassingatacertain
time(i.e.,m/z)whentheelectrodeisturnedoffmomentarily.Theselectedpacketof
ions is then slowed down, possibly subjected to collision-induced dissociation
(CID), and reaccelerated for the final TOFmass analysis of the fragments. The
Orbitrapanalyzerispurchasedasalineariontrap/Orbitraphybridandthesame
vendorsellstheirioncyclotronresonanceICRinstrumentasalineariontrap/ICR
hybrid.Itisbeyondthescopeofthischaptertogointoanyfurtherdetailsregarding
theoperationofthemassanalyzers.Furthermore,itseemslikelythatthefieldwill
continue to change in the coming years, where instrument vendors will make
furtherchanges.
1.1.3
ResolutionandMassAccuracy
Regardlessofthemassspectrometer,theuserneedstounderstandtheircapabilities
andlimitations.Sensitivityhasbeenadrivingforceforthedevelopmentofmanyof
thenewermassspectrometers.Itisalsoadifficultparametertoevaluate,andonehas
to be careful not to simply evaluate the ability and tenacity of each vendor’s
applicationchemistwhensendingtestsamplesout.Dynamicrangeisaparameter
thatisusefulinthecontextofquantitativemeasurementsandformostinstrumentsit
isaround104.Someinstrumentscanperformuniquescantypes(e.g.,theQ-trap),or
aremoresensitiveatperformingSRMquantitativeexperiments(triple-quadrupole
andQ-trapinstruments).ThescanspeedorrateofMS/MSspectraacquisitionisan
instrument parameter that is relevant when attempting a deeper analysis of a
complexmixtureinagivenamountoftime.Thislatterissueisparticularlyimportant
whenanalyzingcomplexproteomicsamples.
Twoanalyzer-dependentparametersareparticularlyimportant–massaccuracy
and resolution. Resolution is defined as a unit-less ratio of mass divided by the
peak width andis typically measuredhalfway up the peak. Figure 1.2shows the
peak shapes calculated for the peptide glucagon at various resolution values.
Atthismass,aresolutionof10000issufficienttoprovidebaselineseparationof
j
1.1 Introduction 7
Figure1.2 Effectofmassspectrometric atvariousresolutionvalues:30000(innermost
resolutiononpeakshape.Shownarethe narrowpeaks),10000(outermostbroadpeak),
calculatedpeakshapesforthe(MþH)þionof 3000(outermostbroadpeak),and1000(outer
porcineglucagon(monoisotopicmassof mostbroadpeak).
3481.62Daandaveragemassof3483.8Da)
eachisotopepeakandthehigherresolutionof30000resultsinthenarrowingof
eachisotopepeak.Astheresolutiondropsbelow10000thevalleybetweeneach
isotopebecomeshigheruntilat3000theisotopeclusterbecomesasinglebroad
unresolvedpeak.Astheresolutiondropsfurther(blue),thesinglebroadpeakgets
even fatter. Resolution is important to the extent that one needs to know if it is
sufficient to separate the isotope peaks of a particular sample. If not, then a
centroidofabroadunresolvedpeak(e.g.,1000or3000forglucagon)isgoingtobe
closesttothepeptidemasscalculatedusingaverageelementalmass.Alternatively,
iftheresolutionissufficienttoresolvetheisotopepeaks,anditispossibleforthe
data system to accurately and consistently identify the monoisotopic 12C peak,
then this observed peptide mass will be closest to that calculated using mono-
isotopic elemental masses.
Whydohighresolutionandhighmassaccuracygohandinhand?Onedoesnot
hearoflow-resolution,high-mass-accuracyinstruments,forinstance.Thereareat
least two reasons. First, it is not possible to determine a very accurate average
elementalmass,whichisweightedforisotopeabundance.Chemicalandphysical
fractionationprocessesoccurringinnatureresultinvariableamountsofeachisotope
indifferentsamples.Forexample,thedifferentphotosyntheticprocesses(e.g.,C3
andC4)willfractionate13Cslightlydifferently.Hence,cornwilltendtohaveaslightly
higherpercentageof13Cthanatree.Therefore,incontrasttomonoisotopicmasses,
averageelementalmassescomewithfairlysubstantialerrorbars.Thesecondreason
j
8 1 MassSpectrometryofAminoAcidsandProteins
Figure1.3 Roleofhighmassaccuracyin (500ppm).Ifthemassmeasurementsare
reducingfalse-positivesfromdatabase accurateto0.01Da(5ppm),whichareroutinely
searches.Thishistogram(from[23])showsthe availableforOrbitrapandcertainQ-TOF
numberoftrypticpeptidesatdifferentmass instruments,thenumberofpossibletryptic
accuraciesfora1996GenPeptdatabase.Fora peptidesinthedatabasedropstoonetoa
nominalmolecularweightof1000,thereare couplehundred,dependingonthespecific
around5000trypticpeptidesifthe masswindow.
measurementsareaccurateto0.5Da
why higher resolution usually results in higher mass accuracy is that as a mono-
isotopicallyresolvedpeakbecomesnarrower,anyslightvariationinthepeakposition
isalsoreduced.Duetofactorssuchasoverlappingpeaksandionstatistics,itisnot
possible to consistently and accurately measure a much wider unresolved isotope
cluster at low resolution. Hence, the type of mass analyzer will determine the
resolutionandmassaccuracy.
Therearethreetypesofresolution(andmassaccuracy)fortandemMSthatare
associated with the precursor ion, precursor selection, and fragment ions. The
precursor and fragment ion resolution and accuracy may be identical (e.g., for
Q-TOForiontraps)ordifferent(e.g.,foriontrap-FT-MShybridsorTOF-TOF).The
importanceofbeingabletomoreaccuratelydeterminepeptidemasseswasclearly
demonstrated by Clauser et al. [23] as shown in Figure 1.3, which depicts a
histogram of the number of tryptic peptides at different mass accuracies. For a
1996GenPeptdatabase,therearearound5000trypticpeptidesatanominalmassof
1000withatoleranceof(cid:3)0.5Da(500ppm).However,ifthetoleranceistightened
up to(cid:3)0.05Da, then the number oftrypticpeptides drops by an amount that is
dependentonthemass.Therearefewerpeptidesateitherthelow-orhigh-mass
endofthehistogram,suchthatthereareonlytwotrypticpeptidesinthedatabaseat
ameasurementof1000.3(cid:3)0.05Da.Likewise,thereareonly30–40peptideswitha
massof1000.7(cid:3)0.05Da.Mostofthetrypticpeptidesatanominalmassof1000are
intherangeof1000.45–1000.65,soatoleranceof(cid:3)0.05Dainthemiddleofthis
histogram will reduce the number of possible tryptic peptides from 5000 to
2000–3000.Whenusingadatabasesearchprogramthatidentifiespeptidesfrom
j
1.1 Introduction 9
their MS/MS spectra, a tighter precursor mass tolerance will result in fewer
candidate sequences, which has the desirable effect of reducing the chances of
anincorrectidentification.
Databasesearchprograms(e.g.,Mascot[24]orSEQUEST[25])assumethatthereis
onlyasingleprecursorandthatallofthefragmentionsarederivedfromthatone
precursorion.Formorecomplicatedsamplesitisquitepossiblethatmorethanone
precursorisselectedatatimeandthelikelihoodofthishappeningisdependenton
the precursor selection resolution. Typical ion traps select the precursor using a
windowthatisthreeorfourm/zunitswide,Q-TOFsaresimilar,andTOF-TOFshave
aprecursorresolutionofaround400(e.g.,atm/z1000,anypeakat997.5willhaveits
transmissionreducedbyhalf).Theshapeofthisprecursorselectionwindowisalso
important – a sharp cutoff to zero transmission is good and a slow taper is not.
Sometimesanextraneouslow-intensityprecursorisnotaproblem,aslongasmostof
thefragmentionintensityisassociatedwiththemajorprecursorandtheprecursor
mass that is associated with the resulting MS/MS spectrum is from the correct
precursorion.Searchprogramswillstillidentifythemajorpeptide,sincetherewill
onlybeafewlow-intensityfragmentionsleftover.However,onecanreadilyimagine
severalscenarioswheremassselectionofmultipleprecursorswouldbeaproblem.
For example, suppose a minor precursor fragments really well, but the major
precursordoesnot.Inthiscase,theMS/MSspectrumcontainsfragmentionsfrom
theminorprecursor,buttheprecursormassthatisusedinthedatabasesearchis
derivedfromthemajorone.Or,alow-intensityprecursortriggersadata-dependent
MS/MS acquisition, but another very intense ion that is a few m/z units away
contributes much of the fragment ion intensity. In such instances, where the
fragment ions are derived from more than one precursor, search programs may
getthewronganswerbecausethewrongprecursormasswasusedortherearetoo
manyleftoverfragmentionsandthescoringalgorithmpenalizesoneofthecorrect
sequences.Tighterselectionwindowswithabruptcutoffs(highprecursorselection
resolution) reduce the likelihood of this occurring. Improved database search
algorithmswouldalsohelp.
Oneofthemajorchallengesinproteomicsishigh-throughputanalysis.Thehigh
resolvingpowerofFT-ICRinstrumentsofferslessthan1ppmmassmeasurement
accuracyandthepeptideidentificationprotocolofaccuratemasstags(AMTs)now
affordsproteinidentificationwithouttheneedfortandemMS/MS.Combiningthe
AMT information with high-performance liquid chromatography (HPLC) elution
timesandMS/MSisreferredtoaspeptidepotentialmassandtimetags(PMTs)[26].
Thisapproachexpeditestheanalysisofsamplesfromthesameproteomethrough
shotgun proteomics – a method of identifying proteins in complex mixtures by
combining HPLC and MS/MS [27]. Once a peptide has been correctly identified
through AMTand MS/MS with an assigned PMT, the information is stored in a
database.Thisstrategygreatlyincreasesanalysisthroughputbyeliminatingtheneed
for time-consuming MS/MS analyses. Accurate mass measurements are now
routinely practiced in applications involving organisms with limited proteomes,
including proteotyping the influenza virus [28], and the rapid differentiation of
seasonalandpandemicstains[29].
j
10 1 MassSpectrometryofAminoAcidsandProteins
1.1.4
AccurateAnalysisofESIMultiplyChargedIons
Itisimportanttobrieflydescribethedeconvolutionalgorithmsusedtotranslatem/z
ratiosofmultiplychargedionsgeneratedduringESItozero-chargemolecularmass
values.Theaccurateassignmentofmultiplychargedionsissignificantinproteomics
applications,bothintheanalysisoftheintactproteinsandfortheidentificationof
fragment ions by MS/MS. Forlow-resolution massspectra, algorithms were orig-
inallydevelopedbyassumingthenatureofcharge-carryingspeciesorconsidering
onlyalimitedsetofchargecarryingspecies(i.e.,proton,sodium)[30,31].Fortwo
ions (m /z and m /z ) that differ by one charge unit and both contain the same
a a b b
charge-carryingspecies,thechargeoniona(z )isgivenbyEq.(1.3),wherem isthe
a p
massofaproton,andthemolecularweightisderivedfromEq.(1.4):
z ¼ðm =z (cid:4)m Þ=ðm =z (cid:4)m =z Þ ð1:3Þ
a b b p b b a a
molecular weight¼z ðm =z Þ(cid:4)z m ð1:4Þ
a a a a p
Theadvantageofhigh-resolutionelectrospraymassspectraisthattheioncharge
canbederiveddirectlyfromthereciprocalofthemass-to-chargeseparationbetween
adjacentisotopicpeaks(1/Dm/z)foranymultiplychargedion–referredtoasthe
isotope spacing method [32]. Although the isotope spacing method is direct,
complexities arising from spectral noise and overlapping peaks may result in
inaccurateionchargedetermination;furthermore,distinguishing1/zand1/(z þ 1)
forhighchargestateions(z>10),wouldrequiremassaccuraciesofafewpartsper
million, which is not achieved routinely. To overcome some of these limitations,
algorithmsofZscore[33]andTHRASH[34]combinedpatternrecognitiontechni-
questotheisotopespacingmethod.Forexample,theTHRASHalgorithmmatches
the experimental abundances with theoretical isotopic distributions based on the
modelaminoacid“averagine”(C H N O S )[35];however,this
4.938 7.7583 1.3577 1.4773 0.0417
requirementrestrictsitsapplicationtoaspecificgroupofcompoundsandelemental
compositions (i.e., proteins). The AID-MS [36] and PTFT [37] algorithms further
advanced the latter algorithms by incorporating peak-finding routines to locate
possibleisotopicclustersandtoovercometheproblemsassociatedwithoverlapping
peaks.
Auniquealgorithm,CRAM(chargeratioanalysismethod)[38–40],deconvolutes
electrospraymassspectrasolelyfromthem/zvaluesofmultiplychargedions.The
algorithmfirstdeterminestheionchargebycorrelatingtheratioofm/zvaluesforany
two(i.e.,consecutiveornonconsecutive)multiplychargedionstotheuniqueratiosof
twointegers.Themass,andsubsequentlytheidentityofthechargecarryingspecies,
isthendeterminedfromm/zvaluesandchargestatesofanytwoions.Fortheanalysis
ofhigh-resolutionelectrospraymassspectra,CRAMcorrelatesisotopicpeaksthat
sharethesameisotopiccompositions.Thisprocessisalsoperformedthroughthe
CRAMprocessaftercorrectingthemultiplychargedionstotheirlowestcommonion
charge.CRAMdoesnotrequirepriorknowledgeoftheelementalcompositionofa