ANALYTICAL BIOCHEMISTRY 244, 1–15 (1997)ARTICLE NO. AB969877

REVIEW

The Quadrupole Ion Trap MassSpectrometer —A Small Solutionto a Big Challenge

Karen R. Jonscher1 and John R. Yates, IIIKaren R. Jonscher

Department of Molecular Biotechnology Box 357730,John R. Yates, IIIUniversity of Washington, Seattle, Washington 98195-7730

Innovations in ion trap mass spectrometry have ex- those of standard amino acid derivatives. The tech-nique requires a free amino terminus and a homoge-tended the applicability of this technique to the analy-

sis of biological molecules. This tutorial review dis- neous sample. Typical sample quantities are at the 1-to 10-pmol level (1). Posttranslational modifications tocusses basic ion trap theory and provides practical

examples of how the theory is used to perform different amino acid residues cause anomalous retention times(2) and may be difficult to identify. Cycle times areÇ30types of experiments such as molecular weight mea-

surements, high resolution, and multiple stages of min/amino acid (1); thus; a peptide containing 25 aminoacids would take 12.5 h to sequence.mass spectrometry (MSn). Peptides generated from en-

zymatic digestion of a-casein, recombinant tissue plas- By contrast, mass spectrometry is a high-sensitivity,high-throughput technique used to acquire both molec-minogen activator, and a cellular extract of proteins

from Haemophilus influenzae illustrate the utility of ular weight and sequence information for proteins andpeptides. Typical sample quantities are at the low- toion trap mass spectrometers for the analysis of bio-

chemical problems. mid-femtomole level (3). Time-of-flight instrumentscan be employed to obtain protein molecular weightsAnalysis of biochemical systems comprised of inter-

acting proteins and peptides typically involves eluci- with two orders of magnitude improvement in massaccuracy over gel electrophoresis. Triple quadrupoledating the molecular weights of the biological mole-

cules and obtaining their amino acid sequences. mass spectrometers are utilized to analyze enzymaticdigests of proteins and tandem mass spectrometry canCovalent modifications to the primary sequence may

affect protein function and must also be determined by be performed to elucidate amino acid sequence infor-mation for peptides through the use of collision-aidedsome means. Several complementary techniques are

typically used to obtain this sort of information. Gel dissociation. The presence of posttranslational modifi-cations is directly determined. Data may be generatedelectrophoresis provides an excellent visualization of

the complexity of the biological process under study; in less than 1 min. Data interpretation has historicallylimited the throughput of the mass spectrometry ap-however, anomalies in migration afford poor mass ac-

curacy and the presence of posttranslational modifica- proach; however, a number of algorithms have beendeveloped to automatically interpret tandem masstions may be difficult to ascertain. Amino acid sequence

information has historically been obtained using auto- spectra (4).Because of these advantages, mass spectrometry ismated Edman degradation. Chemical reagents are em-

ployed to remove one amino acid at a time from the developing into an essential technique for biochemicaland biological research. The range of problems thatamino terminus of an intact protein or peptide. The

resulting amino acid derivative is purified and identi- mass spectrometry is currently being applied to in-cludes the analysis of posttranslational modificationsfication is obtained in a straightforward manner byof proteins (5); noncovalent protein–protein, protein–comparing the HPLC retention time of the sample toDNA, and protein–RNA interactions (6–9); study ofpeptides implicated in the functioning of the immune1 Present address: National Jewish Center for Immunology and

Respiratory Medicine, 1400 Jackson St., Denver, CO 80206. system (10–13); and the study of proteins involved in

10003-2697/97 $25.00Copyright q 1997 by Academic Press, Inc.All rights of reproduction in any form reserved.

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES2

FIG. 1. Rendering of the ion trap electrode assembly showing the ring electrode and the two endcap electrodes.

signal transduction pathways (14–16). A sensitive and of tandem mass spectrometry (MS12) have been per-formed (18); and mass resolution that can allow theversatile analytical system, capable of detecting both

large and small molecules and determining aspects of separation of ions of m/z 106 and m/z 106 / 1 has beenimplemented (19). Quadrupole ion trap mass spectrom-molecular structure, is required to address the complex

mixtures of molecules found in these types of biological eters are also exquisitely sensitive. Molecular weightinformation has been recorded with as few as 1.5 mil-problems. Of fundamental importance to the biochem-

ist and biologist is the existence of robust, easy-to-use, lion peptide molecules (20). Although not all of thesefeatures can be applied simultaneously, a judiciousand inexpensive instrumentation for application to

their studies. choice of parameters can afford sensitive molecularweight measurements and structural analyses of bio-Developments over the past 10 years have made the

quadrupole ion trap mass spectrometer an excellent polymers. The goal of this tutorial is to review the the-ory of ion trap operation and give examples of the prac-tool for biomolecular analysis. A quadrupole ion trap

is a mass analyzer roughly the size of a tennis ball tical application of ion trap mass spectrometry tobiomolecular analysis.whose size is inversely proportional to its versatility.

Three hyperbolic electrodes, consisting of a ring andtwo endcaps, form the core of this instrument (Fig. 1).

HISTORY OF THE DEVELOPMENT OF ION TRAPSUsing theory to drive instrument development, thenominal mass range of the instrument has been ex- In the early 1950s, Wolfgang Paul and co-workers

invented two instruments that could be used to deter-tended from m/z 650 to m/z 70,000 (17); up to 12 stages

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 3

TABLE 1

Time Line of Ion Trap Technology Development

1953 Invention of quadrupole mass filter and quadrupole ion trap by Paul.1959 Storage of single microparticles.1959 Used as a mass spectrometer. Detection by power absorbance.1962 Single ions stored at low temperatures to set frequency standards.1968 Used as a mass spectrometer with external detection.1972 Characterization of the ion trap: Chemical ionization, study of ion/molecule kinetics. Used as a storage device with a

quadrupole mass filter employed for mass analysis.1976 Ions collisionally focused.1978 Used as a selective ion reactor.1979 Ions resonantly ejected.1980 Used as a GC detector.1982 Multiphoton dissociation of ions.1983 Development of mass-selective instability mode of operation.1984 Commercialization of ion trap detector (ITD).1985 Commercialization of ion trap mass spectrometer (ITMS).1987 High-performance mass spectrometry: Multiple stages of mass spectrometry, chemical ionization, photodissociation,

external ion injection, mass range extension.1990 Electrospray ionization of biopolymers.1991 High resolution1991 Discovery of nonlinear effects.1992 Matrix-assisted laser desorption ionization of biopolymers.1993 Biological problem-solving using ion trap mass spectrometry.

mine mass-to-charge (m/z) ratios of ions (21, 22). The bility mode of operation that had been previously em-ployed. This new method for operating the ion trapfirst was the quadrupole mass filter that rapidly was

applied to a wide range of analytical problems (23). The simplified the use of the instrument. Stafford’s groupnext discovered that a helium damping gas ofÇ1 mtorrsecond was the quadrupole ion trap, consisting of a ring

electrode and two endcap electrodes with hyperbolic within the trapping volume greatly improved the massresolution of the instrument (32). Both of these discov-surfaces. As is shown in Table 1 (24), the quadrupole

ion trap was primarily used by the physics community, eries led to the successful development of a commercialion trap mass spectrometer. In later work, the additionnotably Hans Dehmelt at the University of Washing-

ton, to investigate the properties of isolated ions (25– of helium was observed to significantly improve trap-ping efficiencies, especially for externally injected ions28). The ion trap was operated at that time in a ‘‘mass-

selective stability’’ mode of operation. In this mode, (33). Subsequent innovations have been rapid. Cooksand co-workers at Purdue University have pioneeredanalogous to the operation of a quadrupole mass filter,

rf and dc voltages applied to the ring electrode were high-performance techniques such as external injectionof ions (33), mass range extension (17, 34), MSn (18),ramped to allow stability, hence storage, of a single

(increasing) value of m/z in the ion trap (24). Ions were and high resolution (19) that improved the perfor-mance of the ion trap and created interest in its appli-detected by resonance absorption from an external

power source (29) or were ejected using a dc pulse ap- cation to biological molecules.plied to an endcap and detected using an electron mul-tiplier (30). Due to limited mass range and resolution, HOW THEORY IS PUT INTO PRACTICEthese methods of mass measurement were not practical

The Theoryfor many analytical purposes.The chemistry community’s interest in the trap was Quadrupole ion traps are dynamic mass analyzers

that use an oscillating electric potential applied to theconfined to several research groups until 1983 whenGeorge Stafford and co-workers at Finnigan MAT made ring electrode, called the ‘‘fundamental rf,’’ to focus ions

toward the center of the trap. This is accomplished bytwo major advances. First, they developed the mass-selective instability mode of operation (31). The funda- creating a parabolic potential, shaped like a saddle

(35), inside the trapping volume. The strength of themental difference between this mode of operation andprevious methods is that all ions created over a given restoring force linearly increases as the ion trajectory

deviates from the central axis, focusing the ion back totime period were trapped and then sequentially ejectedfrom the ion trap into a conventional electron multi- the center of the trapping volume. This is demonstrated

in Fig. 2, a simulation of ion trajectories created usingplier detector. Thus, all ions were stored while massanalysis was performed, unlike the mass-selective sta- SIMION 3D version 6.0 (36). A population of trapped

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES4

FIG. 2. Simulation of ion trajectories in the ion trap using Simion 3D. The ion trajectories quickly collapse toward the center of the trap.

ions is therefore observed to occupy only the space near higher-order fields within the trapping volume. Theeffects of nonlinear resonances produced in thethe center of the trap due to the focusing effect of the

oscillating electric fields. Assuming a cylindrically sym- stretched trap have been actively studied for the pastfew years and have led to many new insights regardingmetric system, the potential an ion experiences at any

point in the ion trap is given by the fundamental performance characteristics of the iontrap mass spectrometer (37–43).

The force on an ion, given by the electric field, isF(r, z) Å (U 0 V cos vt)

2 Fr2 0 2z2

r2o

G obtained by

Fr

(r, z) Å Er

(r, z) Å 0eÇr

F(r, z) Å mar

(r, z) [2]/ (U 0 V cos vt)

2[1]

and, using Newton’s law, is proportional to the acceler-ation an ion of charge e experiences due to that force.where U is the amplitude of a dc potential applied toEquation [2] may be placed in the form of the Mathieuthe endcap electrodes with reference to the ring elec-equation (44) in the radial and axial directions whentrode, V is the amplitude of the fundamental rf appliedthe substitutionsto the ring electrode, v is the angular frequency of the

rf potential, and ro is the closest distance between thecenter of the trap and the ring electrode (24). The clos- az Å

08eUmr2

ov2 , qz Å

4eVmr2

ov2 ,

est distance between the center of the trap and theendcap electrode is given by zo . To obtain an ideal quad-rupolar field, ro is equal to the square root of 2zo . The ar Å 0az/2, qr Å 0qz/2, and j Å vt

2[3]

actual geometry of the commercial ITMS2 is ‘‘stretched’’and ro is equal to 0.781zo , leading to the presence of

are made. Ion trajectories are determined by solutionsto the Mathieu equation and are oscillating functions

2 Abbreviations used: ITMS, ion trap mass spectrometry; MALDI, with regions of stability described by the parametersmatrix-assisted laser desorption ionization; SWIFT, stored waveform az and qz . Thus, the stability of ion motion dependsinverse Fourier transform; CID, collision-induced dissociation; DAC,

upon the mass and charge of the ion (m), the size ofdigital-to-analog converter; TOF, time-of-flight; ITCL, ion trap in-strument control language. the ion trap (ro), the oscillating frequency of the funda-

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 5

trometer is operated on the line az Å 0. This corre-sponds to the case of maximizing the range of m/z val-ues that may be stably trapped. Ion trajectories becomeunstable in the axial direction (between the endcapelectrodes) but remain stable in the radial directionwhen qz Å 0.908. Ions are ejected through holes in theendcap electrode and are typically detected using anelectron multiplier.

Trapped ions of a given m/z oscillate at a frequencyknown as the secular frequency that is proportional tothe angular frequency of the applied signal, v. Theconstant of proportionality is given by bz,r . For valuesof qz õ 0.4, bz may be approximated by (45)

b2z Å az / q2

z/2 [4]

which reduces to bzÅ qz/√2 for the mass-selective insta-

bility mode of operation. Resonance conditions are in-duced by matching the frequency of a supplementarypotential applied to the endcap electrodes to the secularfrequency of the ion. The ion will absorb energy fromthe applied field and the trajectory will linearly in-crease toward the endcap electrodes until the ion be-comes unstable and is ejected (24).

FIG. 3. Diagram showing the regions of stability in the quadrupoleion trap parameterized in terms of the operating voltages and fre-quencies.

mental rf (v), and the amplitudes of the applied dc (U )and rf (V ) voltages. One region of stability in whichradial and axial stability overlap is shown in Fig. 3.An ion of a given mass-to-charge ratio will be stablytrapped anywhere within that region. The position ofthe ion within the stability region can be moved bychanging the amplitude of the applied dc and rf volt-ages to change the values of az and qz , termed the‘‘working points’’ of the ion. Values of the workingpoints are chosen to ensure stability or instability of anion trajectory of interest. For the case of the commercialFinnigan ion traps (ITD, ITMS), ro Å 1 cm, v/2p Å 1.1MHz, and V ranges from 0 to 7500 V0–p.

As an example, consider three working points for anion of m/z 1500, shown in Fig. 4. Values of the ampli-tudes for the applied dc and rf potentials are shown inparentheses. The corresponding az and qz values are

FIG. 4. Selected working points for an ion of m/z 1500. The applieddelineated in the figure legend. It is clear that a judi-dc and rf potentials are shown in parentheses (U, V ). The correspond-cious choice for the amplitude of the applied potentialsing (az , qz) values are as follows: (0100 V, 1000 V) r (0.0108, 0.0539),is required to ensure stability for all ions within a mass similarly (01000 V, 3000 V) r (0.108, 0.162) and (0100 V, 6000 V) r

range of interest. The mass-selective instability mode (0.0108, 0.323). A judicious choice of conditions is required to ensuretrajectory stability for a wide range of m/z values.of operation utilizes no dc voltage; thus, the mass spec-

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES6

able ionization techniques, i.e., electrospray ioniza-tion and matrix-assisted laser desorption ionization,to the ion trap. These externally created ions need tobe injected into the ion trap and efficiently trapped.Ions are focused by an einzel lens system and allowedinto the ion trap during the ionization period. A gat-ing lens pulses from positive to negative voltages torepel or attract ions toward the entrance endcap ap-erture. The time during which ions are allowed intothe trap is set to maximize the signal while minimiz-ing ‘‘space-charge’’ effects, resulting from too manyions in the trap, that lead to an overall reductionin performance. The ion trap is typically filled withhelium to a pressure of Ç1 mtorr. Collisions withhelium reduce the kinetic energy of the ions andserve to quickly contract trajectories toward the cen-ter of the ion trap, enabling trapping of injected ions.This cooling effect is demonstrated in Fig. 6 wherethe ion population forms a ‘‘packet’’ near the centerof the trap.

Ion trapping. Ions of different m/z values may havestable orbits at the same time, as shown in Fig. 7. Fromthe expression for qz in Eq. [3], we see that

mz

}Vqz

. [5]

FIG. 5. Molecular weight (top) and MS/MS (bottom) scan functions Larger values of m/z will have smaller values of qzfor the quadrupole ion trap mass spectrometer. and smaller values of m/z will have larger qz values.

Since ion trajectories become unstable when qz Å0.908, a well-defined low-mass cutoff is created for aThe Practicegiven value of the amplitude of the applied rf voltage,

To measure the m/z value of a molecule in an ion V. No ions below that mass will be trapped, but ionstrap the molecule must be ionized, focused into the above that mass will be trapped with trapping effi-ion trap, trapped, ejected, and detected. Structural ciency decreasing for larger m/z values (35). Low-information is obtained by collision-induced dissocia- mass cutoffs for various amplitudes of the appliedtion with a helium damping gas and a mass spectrum fundamental rf potential are listed in Fig. 7. Theis generated by sequentially ejecting fragment ions trapping efficiency for an ion of interest depends, infrom low m/z to high m/z. The mass-selective instabil- part, upon the value of the low-mass cutoff, or theity mode is utilized for ion ejection. The mass-selec- so-called exclusion limit (39). This can be a problemtive instability line is the locus of qz values where az when using ionization methods that generate manyis set to zero and maximizes the mass range that may low-mass matrix ions since the ion trap can accommo-be stably trapped. Operation of the ion trap consists date on the order of 105 ions before space-charge seri-of the construction of a scan function used to manipu- ously impairs the performance of the instrument. Forlate the working points of ions of interest. The scan example, the model peptide human angiotensin I (Mrfunction sets the amplitude of the fundamental and 1296) may be most efficiently trapped at a low-masssupplementary potentials and sets the time taken cutoff of 85 u. Matrix-assisted laser desorption ion-for each step. Typical scan functions for molecular ization (MALDI) generates matrix ions above thisweight analyses and MS/MS experiments are shown cutoff in a ratio of Ç1:106. In this situation, the highin Fig. 5. sensitivity of the ion trap can be most effectively uti-

lized if the ion of interest is selectively injected intoIon injection. Ion traps were initially utilized toanalyze volatile samples by electron impact or chemi- the ion trap. Current efforts revolve around selective

injection utilizing shaped excitation waveforms (46)cal ionization. In this case, ions were created insidethe trapping volume. An interest in the analysis of or filtered noise fields (47) to cause all ions but the

ion of interest to have unstable trajectories. Otherbiological molecules led to the need to interface suit-

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 7

FIG. 6. Simulations show that collisions with the helium damping gas lead to the creation of an ion packet near the center of the trap.

approaches include ramping the amplitude of the Ion ejection. Shown in Fig. 7 is an example of therelative positions of three ions of differing m/z ratiosfundamental rf during injection to increase trapping

efficiency, even at low pressures of the helium damp- on the mass-selective instability line, az Å 0. Threedifferent values for the amplitude of the fundamentaling gas (48), as well as the addition of a quadrupole

mass filter to afford selective injection of ions of inter- rf signal are given. As the voltage is increased, the qz

value for the ion also increases. Figure 7c shows thatest into the ion trap (49).

FIG. 7. Relative positions of ions with three different mass-to-charge ratios along the mass-selective instability line, az Å 0. The effect ofincreasing the amplitude of the fundamental rf voltage is shown in (a) through (c).

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES8

FIG. 8. The same conditions as in Fig. 7 except a resonance point at qz Å 0.227 has been imposed to increase the effective mass rangeby a factor of 4. A region of instability is created that affords the ejection of ions at lower voltages than would normally be required;therefore, ions of large m/z can be ejected from the ion trap and detected.

at 6000 V, the ion of m/z 500 has been ejected from the voltage is ramped from low to high amplitudes, all ofthe ions ‘‘fall through the hole’’ and are ejected fromion trap. At the maximum amplitude of 7500 V, at m/

z 1500, the qz value has only reached 0.404; thus, that the trap and detected.ion cannot be ejected from the ion trap and detected. Ion isolation. In a typical multiple-stage mass spec-As noted above, a resonance condition may be induced trometry experiment, the ion of interest is isolated be-by matching the frequency of an applied oscillating sig- fore undergoing resonance excitation or charge statenal to the secular frequency of an ion in the trap (17). determination using high resolution. Isolation in theThis will cause the ion to gain energy and the ampli- Finnigan ITMS may be accomplished in two ways, de-tude of the trajectory to linearly approach the endcap picted in Fig. 9. One method, illustrated in Fig. 9a,electrodes until the ion is ejected from the trap. Ejec- includes the combined use of dc and rf potentials totion can therefore be made to occur at voltages lower bring the qz and az values of the ion to an apex of thethan those required for ejection at qz of 0.908, extending stability diagram; all other ions will be unstable (50,the nominal mass range of the ion trap. Conceptually, 51). The other method is shown in Fig. 9b and consiststhis may be viewed as creating a ‘‘hole’’ in the stability of scanning the amplitude of the fundamental rf voltagediagram. The position of the hole is dependent upon in a reverse-then-forward manner while applying a res-the frequency of the supplementary potential while the onance signal (32, 34). This allows ejection of ions withsize of the hole depends upon the amplitude of the sig- m/z greater than the ion of interest followed by ejectionnal. This effect is illustrated in Fig. 8 where an ellipse of ions having m/z smaller than the ion of interest.represents a resonance point that extends the mass Both isolation methods are used; however, the effects

of space-charge and field nonlinearities on the shaperange by a factor of 4. At 1000 V none of the ions haveof the stability diagram may degrade performanceqz values approaching that of the resonance point; thus,when the dc/rf isolation method is employed. A recentnone will be detected. At 3000 V, m/z 500 has beenrefinement includes the use of the stored waveformejected and m/z 1000 is in the process of being ejected.inverse Fourier transform (SWIFT) technique (46, 52)The qz value for m/z 1500 is smaller than 0.227; thus,and filtered noise fields (47) to isolate ions usingthat ion will not be ejected. At 6000 V, the qz valuesnotched waveforms.for all of the ions are greater than 0.227, the qz value

of the resonance point. This example shows that when Ion dissociation. As discussed above, when an ionapproaches a region of instability in the axial direction,resonance ejection is used and the amplitude of the

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 9

and the frequency and amplitude of the tickle voltagemust be carefully tuned to optimize fragmentation. Theauxiliary frequency generator outputs a single-fre-quency sinusoidal signal that is not sufficient to excitethe envelope of ion signals resulting from isotopic abun-dances for ions with large m/z values. Stored waveforminverse Fourier transform techniques (52) and the ap-plication of random noise (55) have been successfullyused to excite a broad range of ion secular frequencies.In addition, shifting the qz value of the ion and increas-ing the amplitude of the tickle pulse have substantiallyincreased the amount of fragmentation observed forlarge peptides (56).

High resolution. The mass resolution of the ion trapmass spectrometer is a function of the number of rfcycles that the ion spends interacting with the trappingfield (57). Resolution is increased by reducing the am-plitude of the resonance ejection signal and reducingthe ejection scan speed, nominally 5555 u/s for the Fin-nigan ITMS. The scan speed is attenuated utilizing anetwork of resistors placed in series with the digital-to-analog converter (DAC) that controls the amplitudeof the rf voltage applied to the ring electrode (17). Thefixed scanning rate of the DAC is applied to smaller‘‘windows’’ of rf voltages with a concomitant gain inFIG. 9. Methods of isolating a single m/z in an ion trap. (a) Athe number of data points taken per unit mass. A dccombination of dc and rf potentials are applied to bring the az andpotential is used as an offset to position the rf voltage,qz values of the ion of interest to the apex of the stability diagram.

Neighboring ions have working points that fall outside of the region or mass window. This is schematically illustrated inof stability. (b) Reverse-then-forward scanning of the amplitude of Fig. 10. Figure 10a shows the unattenuated mass win-the fundamental rf voltage in conjunction with the application of an dow resulting from scanning the amplitude of the rfauxiliary signal to create a resonance point affords ion isolation. (i)

voltage from 346 to 7500 V while applying a supple-Reverse scanning resonantly ejects ions from high to low m/z. (ii)mentary frequency at 120 kHz to extend the massForward scanning resonantly ejects ions from low to high m/z. (iii)

Resultant isolation of one value of m/z. range by a factor of 3. This increases the mass scanspeed to 16665 u/s. Attenuation of the scan speed by afactor of 10 reduces the size of the mass window by

the deviation of its trajectory from the center of the the same factor; thus, Figs. 10b–10d represent a 186-trap will increase. When instability is induced by a u mass window created by the attenuation. The differ-resonance signal, the amplitude of the resonance signal ent dc offset voltages serve to position the mass windowcan be adjusted to cause collisionally induced dissocia- in different regions within the mass window. In Fig.tion (CID) of the ion with the helium damping gas 10b, the mass window is positioned at 267 u, in 10c itrather than ejection from the ion trap (32). An esti- is 800 u, and in 10d it is 1600 u; therefore, differentmated 10,000 low-energy collisions (23) transfer regions of the mass window are accessed. Attenuationenough energy into peptide ions to cause random frag- by a factor of 100–300 is typically required to resolvementation along the peptide backbone in a manner the isotopes for singly to triply charged peptide ionsanalogous to that obtained using a triple quadrupole to achieve resolutions of 10,000–30,000 at m/z valuesmass spectrometer. CID efficiency typically ranges ranging between 500 and 2000.from 40 to 80%, although it approaches 100% for somefavorable cases (53). The amplitude of the rf signal that An Examplesets the qz value of the isolated ion during resonanceexcitation, termed the ‘‘tickle mass,’’ must be judi- The pulsed nature of the quadrupole ion trap makes

it particularly well suited to pulsed ionization tech-ciously set as it will serve to eject all ions with m/zvalues below the tickle mass. This limits the amount niques such as MALDI. A MALDI ion trap that has

been described previously (58) was used for mappingof low m/z fragmentation information obtained. A com-plete set of complementary b- and y-type ions (54) is the tryptic peptides from tPA. A 1/2-ml aliquot of the

digest, corresponding to Ç1 pmol, was loaded onto atypically not obtained unless multiple stages of massspectrometry are performed. The qz value of the ion probe tip and cocrystallized with 1 ml of a saturated

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES10

cally used to obtain both MS and MS/MS spectra (61,63), similar to results obtained by triple quadrupoleand TOF mass spectrometry. Lower levels are possible,but are not routine at present. Sensitivity is improvedby varying the ion collection time and selectively in-jecting the ion of interest. The ion trap, with MALDI,has shown equivalent performance to TOF mass spec-trometers at low-mass range with the added advantageof exact precursor ion selection and MSn. It is unlikelythe ion trap will be as suitable for ultrahigh mass anal-ysis as TOF mass spectrometers due to hardware limi-tations of the auxiliary frequency generator used toextend the mass range. High-molecular-weight spectraobtained for singly charged proteins (30- to 50-kDarange) have shown results comparable to those ob-tained using a linear TOF mass spectrometer withoutthe implementation of delayed extraction techniques.

There are several limitations of the performance ofquadrupole ion trap mass spectrometers. The alternate

FIG. 10. Extending the resolution on the quadrupole ion trap massspectrometer. (a) A normal resonance ejection scan from 90 to 1950u. (b) The scan speed is attenuated by a factor of 10 resulting in a10-fold decrease in the width of the scanning mass window. The dcoffset is utilized to position the scanning window throughout themass range. An offset of 100 V positions the window at 267 u, (c)300 V positions the window at 800 u, and (d) 600 V positions thewindow at 1600 u.

solution of a-cyano-4-hydroxycinnamic acid in 1:1 0.1%trifluoroacetic acid:acetonitrile. The probe tip was irra-diated using a nitrogen laser (337 nm). The resultingmass spectrum of the digest is shown in Fig. 11. Severalof the peaks corresponding to tryptic peptides are la-beled. Approximately 75% of the expected peptides fall-ing within the measurement mass range were detected.Extensive fragmentation of peptide ions generated byMALDI has been observed (59, 60) and many of thesignals in the mass spectrum in Fig. 11 result fromfragmentation upon injection into the mass spectrome-ter. The analytical potential of MALDI–ITMS contin-ues to be explored and shows great promise for applica-tion to biological molecules (61, 62).

Comparison with Other Methods

As an ion storage device, an ion trap has the capabil-ity for high mass resolution, mass range, sensitivity,and MSn that translates into versatile performance as amass spectrometer. In comparison to triple quadrupole FIG. 11. MALDI–ITMS mass spectrum of a tryptic digest from

recombinant tPA. Ions were gated into the trap during a 5-ms period.and TOF mass spectrometers the ion trap is unique inThe rf exclusion limit was set to 95 u. A 5-ms pulse at 7500 V0–pits ability to perform MSn. All three techniques arewas employed to eject low-mass matrix ions. Ions were resonantlyabout equal in terms of mass accuracy and sensitivity. ejected using a supplementary signal at 59,124 Hz, 8 Vp–p, corre-

When utilizing electrospray ionization or MALDI, mid- sponding to a mass range extension of a factor of 6. Several of thetryptic peptides are labeled.femtomole to low-picomole levels of sample are typi-

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 11

scan modes of triple quadrupole mass spectrometers tickle voltages, and voltage ramping for acquisition ofthe mass spectrum (see Fig. 5). Once a scan functionsuch as precursor ion and neutral loss scans are cur-

rently not possible. Furthermore, the number of ions is established for a given experiment it can be usedagain but some parameters may need to be changedinjected into the ion trap must be carefully controlled

since space-charging can degrade the performance of based on the m/z value of the ion of interest. For exam-ple, MS, MS/MS, MSn, and high-resolution experimentsthe instrument. This problem is solved through a rapid

prescan that assesses the ion current injected into the all require the construction of unique scan functions.Significant interaction and expertise with the softwaretrap for Ç50 ms then sets the ionization time to max-

imize the signal while minimizing space-charge. Fi- were required with the older ion traps. The LCQ wasdeveloped with ion trap instrument control languagenally, when MS/MS is performed, all ions with qz values

below that of the resonance point will be ejected from (ITCL), a computer language that controls all of theelements of the scan function. For example, the hypo-the ion trap; therefore, a complete sequence of comple-

mentary b- and y-type ions typically cannot be ob- thetical ITCL command ‘‘hires 1200’’ would set up thescan function to isolate the ion at m/z 1200, then slowtained. Cotter et al. have recently shown that using

low qz values in conjunction with a heavier target gas the scan rate to achieve high mass resolution. All pa-rameters required are automatically set with the oneaffords full tandem mass spectra; consequently, the

ejection of low m/z fragment ions during CID is not a ITCL command, compared with the necessity to manu-ally set a number of parameters using the ITMS soft-fundamental limitation of the ion trap (64). Perhaps

one major advantage of the ion trap not easily over- ware.ITCL also enables the user to perform data-depen-looked is the size of the instrument. As lab space be-

comes tighter, the size of the ion trap and ease of main- dent experiments. A mass scan can return to the com-puter program all the information it acquires duringtenance become a considerable advantage.the scan. For example, a command such as ‘‘hiresmass(1)’’ would perform a high-resolution mass scan

THE NEW GENERATION OF ION TRAPSon the most intense ion returned from the previousmass scan. Very complicated data-dependent routinesIn the past, the ITMS has not been an instrument

well suited to the robust and routine analyses required such as ‘‘on-the-fly’’ tandem mass spectrometry can beperformed by stringing together commands in the formby biochemists and biologists. High-performance inno-

vations to the ITMS developed over the past several of a computer program. A graphical user interface isemployed to simplify the use of ITCL and to edit theyears have been used to build a new generation of ion

trap mass spectrometer, the Finnigan MAT LCQ. This type of experiment desired during the course of an anal-ysis. An m/z measurement, followed by a high-resolu-instrument has been carefully designed to interface

with atmospheric pressure ionization techniques that tion scan to separate the isotopes of the desired ion forcharge state determination, followed by tandem massare optimal for the analysis of biomolecules. The op-

erating characteristics of the instrument have been spectrometry, is achieved by selecting the experimentthrough the user interface. The software can automati-changed by using a fundamental rf of 760 kHz instead

of 1.1 MHz, an electrode spacing of 0.707 cm instead cally select precursor ions based on some predefinedcriteria such as abundance, presence, or absence of anof 1.0 cm, and a qz value of 0.83 instead of 0.908 for

resonance ejection of ions (65). Ion injection into the ion in a predefined list. No user intervention in theprocess is required except for the initial setup of theion trap has been optimized using a lensing system

that consists of two rf-only octopoles, resulting in a analysis. This level of control is unprecedented in massspectrometry. In fact, the reliance on embedded soft-narrow spatial and energy distribution of the injected

ions (49, 66). Selective injection, trapping, and excita- ware control is so great that instrument upgrades willessentially require downloading software from a CD-tion of ions are performed using tailored waveforms,

analogous to the SWIFT technique (67). Unit mass res- ROM to change operational parameters, obviating theneed for expensive additions of hardware.olution, or the ability to separate an m/z value of 1500

from 1501, is maintained over the 2000-dalton mass A number of different automated, data-dependentexperiments are possible including full-range MS atrange with a mass accuracy of 0.015% (68). These fig-

ures of merit are comparable to the performance of unit resolution, MSn with n Å 1 to 10, single-ion moni-toring (SIM), and single-reaction monitoring (SRM),current triple quadrupoles. It is expected that the mass

range of the LCQ will increase to 5000 daltons in the charge-state determination (utilizing the ‘‘ZoomScan’’)of up to /4 ions, and unit resolution isolation up to m/next year.

The most striking feature of the new ion trap is the z 1200. An example of the application of these experi-ments is illustrated in Figs. 12 and 13.software control of instrument operation. Ion traps are

operated through the use of a scan function that sets The benefit of performing multiple stages of massspectrometry is demonstrated in Fig. 12 where MS4the ion injection time, trapping voltages, cooling time,

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES12

FIG. 12. MS4 on the doubly charged ion of m/z 880 from a tryptic digest of the model protein a-casein. A 0.5 pmol/ml sample solution wasinfused into a home-built microspray ionization source at a flow rate of 200 nl/min. The ionization time was automatically set usingautomated gain control for the first three stages of mass spectrometry. The AGC was disabled for the fourth stage and the ionization timewas set to 400 ms to compensate for the loss in sensitivity due to the performance of multiple stages of mass spectrometry. Ten scans weresummed for the MS, MS2, and MS3 experiments. Fifteen scans were summed for the MS4 experiment. (a) Mass assignments for sequenceions corresponding to MS2 of m/z 880, MS3 of m/z 436, and MS4 of m/z 266. (b) The precursor at m/z 880 displayed in the top panel waschosen for fragmentation. The b/1

4 fragment ion at m/z 436, shown in the second panel, was chosen for a further stage of fragmentationand the resulting mass spectrum is exhibited in the third panel. The b/1

2 fragment ion at m/z 266 was chosen to obtain the very low-massend of the fragmentation spectrum. Results are displayed in the bottom panel.

was performed on an ion from a trypsin-generated pep- in a complex biological mixture. Cellular proteins fromH. influenzae were fractionated using ion-exchangetide from the model protein a-casein. The top panel

shows the unit resolution full mass-range spectrum of chromatography (MonoQ). One of the fractions was di-gested using trypsin then concentrated and buffer-ex-the entire digest. The ion of m/z 880 was chosen for

further investigation. MS/MS on that ion provided the changed using Centricon filters. An aliquot was loadedonto a 500-mm POROS R2 packed column and sepa-mass spectrum shown in the second panel. Ions below

245 u were ejected upon the application of the reso- rated by reverse-phase high-performance liquid chro-matography using a 50-min gradient from 0–40% Bnance excitation pulse. A third stage of mass spectrom-

etry, depicted in the third panel, provides additional followed by 40–60% B in 10 min at a flow rate of 50ml/min. Solvent A was H2O/AcOH in a ratio of 100:0.5low-mass sequence ions while the sequence is com-

pleted using a fourth stage of mass spectrometry. The and solvent B was ACN/H2O/AcOH in a ratio of80:20:0.5. Shown in the top panel of Fig. 13 is the unitamino acid sequence was deduced to be HQGLPQEVL-

NENLLR. resolution full-range mass spectrum corresponding toscan number 1083 in the ion chromatogram (data notA final example demonstrates the automated appli-

cation of the instrument to the analysis of components shown). A number of coeluting species are observed.

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 13

TVTEIVELIAAMEEK and is derived from a ribosomalprotein homologous to the L7/L12 protein found inEscherichia coli.

CONCLUSION

The quadrupole ion trap is an extremely versatileinstrument capable of performing multiple stages ofmass spectrometry with one mass analyzer. High-reso-lution techniques afford easy charge-state determina-tion, facilitating the interpretation of data generatedby electrospray ionization. The sensitivity and perfor-mance characteristics of the instrument, especially theautomated experiments developed for the newly com-mercialized ion traps, make quadrupole ion trap massspectrometry an attractive technique to apply to theanalysis of biological and biochemical problems.

ACKNOWLEDGMENTS

The authors thank John Stults of Genentech for graciously provid-ing the tPA sample. Jon DeGnore and Richard Yost from the Univer-sity of Florida provided the wonderful ion trap renderings. Thesecan be accessed on their web page at http://analytic15.chem.ufl.edu/anim1.html. Edwin Carmack provided the ribosomal protein data.

APPENDIX

Ion Trap Jargon

ac voltage: also called supplementary or auxiliaryFIG. 13. On-line data-dependent analysis of tryptic peptides from potential, is a voltage placed on the endcap electrodes.an H. influenzae cellular protein extract. The full-range mass spec- Bath gas, damping gas, target gas: helium gas in thetrum from scan number 1083 is shown in the top panel. The ion at

trapping volume at a pressure of Ç1 mtorr.m/z 839 was automatically chosen for further analysis. Charge-stateFundamental rf: a (typically) 1.1-MHz potential ap-determination, shown in the middle panel, was accomplished by

slowing the mass scan speed and enhancing the resolution. TheÇ0.5- plied to the ring electrode.u distance between the isotopic peaks indicates the ion is doubly High resolution: an experiment in which peaks corre-charged. MS/MS was performed and the resulting fragmentation sponding to carbon isotopes may be resolved.mass spectrum is illustrated in the bottom panel. Computer-aided

Resonance: an ac voltage is applied to the endcapidentification of the peptide indicated it was a fragment from ribo-somal protein homologous to the L7/L12 protein found in E. coli. and the qz value of an ion of interest is changed until

the secular frequency of the ion matches the frequencyof the applied ac voltage. A high-amplitude ac voltagewill cause resonance ejection, while a low-amplitude acThe ion at m/z 839 was chosen automatically for

charge-state determination followed by MS/MS. The voltage will cause resonance excitation.Secular frequency: the frequency, dependent uponresult of the ZoomScan is demonstrated in the middle

panel. The spacing between the isotope peak centroids the qz value, with which an ion oscillates in the trap.Space-charge: too many ions in the trap distort theis Ç0. 5 u, indicating the ion has a charge state of /2.

The width of the peak at half the maximum intensity electric fields, leading to significantly impaired perfor-mance.(FWHM) for the signal at m/z 839.1 is 0.179 u, provid-

ing a resolution of 4688 at that mass. Resolution in- Tickle voltage: an ac voltage applied to the endcapelectrodes during an excitation period. The amplitudecreases with increasing m/z and resolutions ap-

proaching 20,000 have been observed. The automated of the voltage is generally small so as to enable frag-mentation of the ions by collisions with the heliumfragmentation mass spectrum is depicted in the bottom

panel. Some selected fragment ions are labeled. The damping gas rather than ejection.Working points: the values of the Mathieu parame-SEQUEST database searching algorithm (69) was uti-

lized to identify the amino acid sequence of the peptide ters az and qz . For an ion of a given m/z, the positionof the working points depends upon the amplitudes ofand determine its origin. The result of the computer

analysis indicates that the peptide has the sequence rf and dc potentials applied to the ring electrode.

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

JONSCHER AND YATES14

M. O., and Sinclair, R. M., Eds.), May 27–31, 1991, pp. 16–22,REFERENCESWorld Scientific, Singapore, Santa Fe, NM.

1. Stolowitz, M. L. (1993) Curr. Opin. Biotechnol. 4, 9–13. 29. von Rettinghaus, V. G. (1967) Z. Angew Phys. 22, 4, 321–326.2. Wettenhall, R. E. H., Aebersold, R. H., and Hood, L. E. (1991) 30. Dawson, P. H., and Whetten, N. R. (1968) J. Vac. Sci. Technol.

Methods Enzymol. 201, 186–199. 5, 11–18.3. Hunt, D. F., Henderson, R. A., Shabanowitz, J., Sakaguchi, K., 31. Stafford, G. C., Jr., Kelley, P. E., Syka, J. E. P., Reynolds, W. E.,

Michel, H., Sevilir, N., Cox, A. L., Appella, E., and Engelhard, and Todd, J. F. J. (1984) Int. J. Mass Spectrom. Ion Proc. 60,V. H. (1992) Science 255, 1261–1263. 85–98.

4. Yates, J. R., III, McCormack, A. L., Link, A. J., Schieltz, D., Eng, 32. Louris, J. N., Cooks, R. G., Syka, J. E. P., Kelley, P. E., Stafford,J., and Hays, L. (1996) Analyst 121, 7, R65–R76. G. C., Jr., and Todd, J. F. J. (1987) Anal. Chem. 59, 1677–1685.

5. Burlingame, A. L., Boyd, R. K., and Gaskell, S. J. (1996) Anal. 33. Louris, J. N., Amy, J. W., Ridley, T. Y., and Cooks, R. G. (1989)Chem. 68, 12, 599R–651R. Int. J. Mass Spectrom. Ion Proc. 88, 97–111.

6. Smith, D. L., and Zhang, Z. Q. (1994) Mass Specrom. Rev. 13, 34. Kaiser, R. E., Jr., Louris, J. N., Amy, J. W., and Cooks, R. G.5–6, 411–429. (1989) Rapid Commun. Mass Spectrom. 3, 7, 225–229.

7. Smith, R. D., and Light-Wahl, K. J. (1993) Biol. Mass Specrom. 35. McLuckey, S. A., Van Berkel, G. J., Goeringer, D. E., and Glish,22, 493–501. G. L. (1994) Anal. Chem. 66, 13, 689A–696A.

8. Smith, R. D., Cheng, X. S., B. L., Chen, R., and Hofstadler, S. A. 36. Dahl, D. A. (1995) Lockheed Martin Idaho Technologies, Idaho(1996) ACS Symp. Ser. 619, 294–314. Falls, ID.

9. Przybylski, M., and Glocker, M. O. (1996) Angew Chem. Int. Ed. 37. Paradisi, C., Todd, J. F. J., Traldi, P., and Vettori, U. (1992) Org.Engl. 35, 806–826. Mass Spectrom. 27, 251–254.

10. Henderson, R. A., Michel, H., Sakaguchi, K., Shabanowitz, J., 38. Paradisi, C., Todd, J. F. J., Traldi, P., and Vettori, U. (1992)Appella, E., Hunt, D. F., and Engelhard, V. H. (1992) Science Rapid Commun. Mass Spectrom. 6, 641–646.255, 1264–1266.

39. Williams, J. D., Reiser, H.-P., Kaiser, R. E., Jr., and Cooks, R. G.11. Wang, W., Meadows, L. R., Haan, J. M. M. D., Sherman, N. E., (1991) Int. J. Mass Spectrom. Ion Proc. 108, 199–219.

Chen, Y., Blokland, E., Shabanowitz, J., Agulnik, A. I., Hendrick-40. Williams, J. D., Cox, K. A., Cooks, R. G., McLuckey, S. A., Hart,son, R. C., Bishop, C. E., Hunt, D. F., Goulmy, E., and Engelhard,

K. J., and Goeringer, D. E. (1994) Anal. Chem. 66, 725–729.V. H. (1995) Science 269, 1588–1590.41. Guidugli, F., and Traldi, P. (1991) Rapid Commun. Mass Spec-12. Cox, A. L., Skipper, J., Chen, Y., Henerson, R. A., Darrow, T. L.,

trom. 5, 343–348.Shabanowitz, J., Engelhard, V. H., Hunt, D. F., and Slingluff,42. Eades, D. M., Johnson, J. V., and Yost, R. A. (1993) J. Am. Soc.C. L., Jr. (1994) Science 264, 716–719.

Mass Spectrom. 4, 917–929.13. Slingluff, C. L., Jr., Hunt, D. F., and Engelhard, V. H. (1994)43. Cox, K. A., Williams, J. D., Cooks, R. G., and Kaiser, R. E., Jr.Curr. Opin. Immunol. 6, 733–740.

(1992) Biol. Mass Spectrom. 21, 226–241.14. Watts, J. D., Affolter, M., Krebs, D. L., Wang, R. L., Samelson,44. McLachlan, N. W. (1947) Theory and Applications of MathieuL. E., and Aebersold, R. (1994) J. Biol. Chem. 269, 47, 29520–

Functions, Clarendon, Oxford.29529.45. Wuerker, R. F., Shelton, H., and Langmuir, R. V. (1959) J. Appl.15. Watts, J. D., Affolter, M., Krebs, D. L., Wange, R. L., Samelson,

Phys. 30, 3, 342–349.L. E., and Aebersold, R. (1996) ACS Symp. Ser. 619, 381–407.46. Soni, M. H., and Cooks, R. G. (1994) Anal. Chem. 66, 2488–2496.16. Affolter, R., Watts, J. D., Krebs, D. L., and Aebersold, R. (1994)

Anal. Biochem. 223, 74–81. 47. McLuckey, S. A., Goeringer, D. E., and Glish, G. L. (1991) J. Am.Soc. Mass Spectrom. 2, 11–21.17. Kaiser, R. E., Jr., Cooks, R. G., Stafford, G. C., Jr., Syka, J. E. P.,

and Hemberger, P. H. (1991) Int. J. Mass Spectrom. Ion Proc. 48. Doroshenko, V. M., and Cotter, R. J. (1993) Rapid Commun.Mass Spectrom. 7, 9, 822–827.106, 79–115.

18. Louris, J. N., Brodbelt-Lustig, J. S., Cooks, R. G., Glish, G. L., 49. Jonscher, K. R., and Yates, J. R., III (1996) Anal. Chem. 68, 4,659–667.Van Berkel, G. J., and McLuckey, S. A. (1990) Int. J. Mass Spec-

trom. Ion Proc. 96, 117–137. 50. Dawson, P. H., Hedman, J., and Whetten, N. R. (1969) Rev. Sci.Instrum. 40, 1444–1450.19. Cooks, R. G., Hoke, S. H., II, Morand, K. L., and Lammert, S. A.

(1992) Int. J. Mass Spectrom. Ion Proc. 118/119, 1–36. 51. Mather, R. E., Waldren, R. M., and Todd, J. F. J. (1978) Dyn.Mass Spectrom. 5, 71–85.20. Kaiser, R. E., Jr., Cooks, R. G., Syka, J. E. P., and Stafford, G. C.,

Jr. (1990) Rapid Commun. Mass Spectrom. 4, 1, 30–33. 52. Julian, R. K. (1993) Anal. Chem. 65, 1827–1833.21. Paul, W., and Steinwedel, H. (1956) German Patent 944,900. 53. Cooks, R. G., and Kaiser, R. E., Jr. (1990) Acc. Chem. Res. 23, 7,

213–219.22. Paul, W. (1990) Angew Chem. Int. Ed. Engl. 29, 739.23. Cooks, R. G., McLuckey, S. A., and Kaiser, R. E. (1991) Chem. 54. Roepstorff, P., and Fohlman, J. (1984) Biomed. Mass Spectrom.

11, 601.Eng. News 69, 12, 26–41.24. March, R. E., and Hughes, R. J. (1989) in Chemical Analysis: A 55. McLuckey, S. A., Goehringer, D. E., and Glish, G. L. (1992) Anal.

Chem. 64, 1455–1460.Series of Monographs on Analytical Chemistry and Its Applica-tions (Winefordner, J. D., and Kolthoff, I. M., Eds.), Vol. 102, 56. Qin, J., and Chait, B. T. (1995) Proceedings of the 43rd ASMSWiley, New York. Conference on Mass Spectrometry. Allied Topics, May 21–26,

Atlanta, GA, p. 1100.25. Dehmelt, H. (1995) Phys. Scripta Vol. T T59, 87–92.26. Dehmelt, H. G. (1967) Adv. At. Mol. Phys. 3, 53–72. 57. Fischer, E. (1959) Z. Phys. 156, 1.

58. Jonscher, K. R., Currie, G., McCormack, A. L., and Yates, J. R.,27. Dehmelt, H. G. (1969) Adv. At. Mol. Phys. 5, 109.III (1993) Rapid Commun. Mass Spectrom. 7, 20–26.28. Dehmelt, H. (1991) in Santa Fe Workshop. Foundations of Quan-

tum Mechanics (Black, T. D., Nieto, M. M., Pilloff, H. S., Scully, 59. Jonscher, K. R., and Yates, J. R., III (1994) Proceedings of the

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba

QUADRUPOLE ION TRAP MASS SPECTROMETRY 15

42nd ASMS Conference on Mass Spectrometry. Allied Topics, the 43rd ASMS Conference on Mass Spectrometry. Allied Topics,May 21–26, Atlanta, GA, p. 1114.May 29–June 3, Chicago, IL, pp. 216–217.

66. Bier, M. E., Schwartz, J. C., Zhou, J., Taylor, D., Syka, J., James,60. Qin, J., and Chait, B. T. (1995) J. Am. Chem. Soc. 117, 19, 5411–M., Fies, B., and Stafford, G. (1995) Proceedings of the 43rd5412.ASMS Conference on Mass Spectrometry. Allied Topics, May61. Qin, J., and Chait, B. T. (1995) Proceedings of the 43rd ASMS21–26, Atlanta, GA, p. 1117.Conference on Mass Spectrometry. Allied Topics, May 21–26,

67. Taylor, D., Schwartz, J., Zhou, J., James, M., Bier, M., Korsak,Atlanta, GA, p. 989.A., and Stafford, G. (1995) Proceedings of the 43rd ASMS Confer-62. Jonscher, K. R., and Yates, J. R., III (1996) J. Biol. Chem. Inence on Mass Spectrometry. Allied Topics, May 21–26, Atlanta,press.GA, p. 1103.

63. Yates, N. A., Shabanowitz, J., and Hunt, D. F. (1994) Abstr. Pa- 68. Land, A. P., Wheeler, K., Mylchreest, I. C., Sanders, M., andpers Am. Chem. Soc. 208, 132-ANYL. Jardine, I. (1995) Proceedings of the 43rd ASMS Conference on

64. Cotter, R. (1996) in Photonics West Conference, January 27– Mass Spectrometry. Allied Topics, May 21–26, Atlanta, GA, p.February 3. 653.

69. Eng, J. K., McCormack, A. L., and Yates, J. R., III (1994) J. Am.65. Schwartz, J. C., Bier, M. E., Taylor, D. M., Zhou, J., Syka,J. E. P., James, M. S., and Stafford, G. C. (1995) Proceedings of Soc. Mass Spectrom. 5, 976–989.

AID AB 9877 / 6m21$$$121 12-06-96 01:41:00 aba