evidence for enzymatic activity in the absence of solvent in gas-phase complexes of lysozyme and...

Evidence for enzymatic activity in the absence of solvent ingas-phase complexes of lysozyme and oligosaccharides

Fei He, Javier Ramirez, Carlito B. Lebrilla*

Department of Chemistry, University of California, Davis, CA 95616, USA

Received 19 November 1999, accepted 16 July 1999

Abstract

Hen egg-white lysozyme complexed to substrate and nonsubstrate oligosaccharides are examined by electrospray ionizationmass spectrometry. Heated capillary dissociation and collision induced dissociation (CID) are used to characterize thecomplexes. The relative order of stability obtained in gas phase agrees well with their solution-phase association constants. Aproton transfer reaction occurs during the CID of substrate complexes that is not observed with non-substrate complexes.Oligosaccharide fragments are also observed with lysozyme–chitohexaose complex during CID. The possibility of lysozymeactivity in the gas phase, or a gas-phase enzymatic activity, is explored. (Int J Mass Spectrom 193 (1999) 103–114) © 1999Elsevier Science B.V.

Keywords:CID; Lysozyme; Gas phase; Enzyme

1. Introduction

Mass spectrometry has recently been used to in-vestigate the higher-order structures of proteins.McLafferty and co-workers [1] first reported thefolding and unfolding of protein in vacuo. Since thenseveral papers on folding and unfolding of proteins inthe gas phase, including cytochromec and lysozyme[2–8] have been reported. A key question remains,however, as to whether the gas-phase structures arerelated to those in solution. Several reports addressthis question by different means. Smith and co-workers studied peptide and protein dimers by twodifferent electrospray ionization (ESI) interfaces, onewith the countercurrent flow of bath gas and the other

with heated metal capillary, both coupled with colli-sion induced dissociation (CID) [9,10]. The authorsconcluded that at least some forms of noncovalentassociations found in solution are preserved duringthe ESI process. Williams and co-workers utilizedproton transfer reactivity to investigate isolatedcharge states of hen egg-white lysozyme [2]. It wasfound that the 91 and 101 charge states havereactivity consistent with their crystal structures. Re-cently, Loo et al. examined the structure of RES-701-1, a hexadecapeptide isolated fromStreptomyces,and its synthetic analog [11]. The naturally occurringpeptide has a unique highly ordered structure, whichcan not be duplicated synthetically. They reportedsignificant differences in the two peptides in theircollisional induced dissociation (CID) and solution-phase H/D exchange. The authors surmised that thenatural peptide can withstand the solution-to-gas*Corresponding author. E-mail: [email protected]

1387-3806/99/$20.00 © 1999 Elsevier Science B.V. All rights reservedPII S1387-3806(99)00167-0

International Journal of Mass Spectrometry 193 (1999) 103–114

phase transition and preserve its solution-phase struc-ture.

In this article, we use ESI Fourier transform massspectrometry (FTMS) to investigate enzyme substratecomplexes. If enzyme–substrate complexes are foundto have enzymatic activity in vacuum, it would be themost direct evidence to indicate whether solution-phase structures are preserved in the gas phase. Thelysozyme used in this research, hen egg-white ly-sozyme, is a single protein chain of 129 residues. It isa compact, highly stable protein with four disulfidebonds. Therefore, it serves as a good candidate forstudying the stability of enzyme–substrate complexesand enzymatic activity in the gas phase by electro-spray mass spectrometry.

Hen egg-white lysozyme was chosen as the modelenzyme to examine gas-phase enzymatic activities.Lysozyme is widely found in living organism and wasone of the first enzymes whose crystal structure wasdetermined [12,13]. In solution, it hydrolyzesb(1–4)-glycosidic bonds in bacterial cell wall carbohydrates.Oligosaccharides composed ofN-acetylglucosamineoligomers (commonly known as chitin) [14] are alsonatural substrates of lysozyme. It is well accepted thatlysozyme contains a grove able to bind up to sixcontiguous saccharide residues. An essential require-ment for the specific binding of oligosaccharides tolysozyme is theN-acetyl group on the sugar residue,which interacts strongly with the enzyme by formingtwo hydrogen bonds. The aspartic acid 52 (Asp 52)and glutamic acid 35 (Glu 35) participate directly inthe catalysis [14–17]; the COOH group of Asp 52donates a H1 to the glycosidic oxygen to cleave theglycosidic bond through a positively chargedcarbo-nium ion intermediate. This process is facilitated bythe resulting negative charge on the Asp 52, whichelectrostatically stabilizes the positive charge. Thenascentcarbonium ionintermediate then reacts withH2O from the solution and produces the hydrolysisproducts. Henion and co-workers studied lysozymecomplexes with N-acetylglucosamine oligosaccha-rides [18]. By electrospraying enzyme substrate solu-tions, they were able to determine the relative abun-dance of complex peaks in the mass spectra. A goodagreement was found between the complex intensity

and their solution-phase binding constants. Further-more, they investigated the complexation when tetra-N-acetylchitotetraosed-lactone, one of the lysozymeinhibitors, was added to the solution. They concludedthat the close correlation observed between relativeion abundance and substrate or inhibitor associationconstants are unlikely to be a random aggregation ofoligosaccharides on the exterior surface of the en-zyme. However, the probe of the gas-phase structureof the complexes has not been reported. Nonetheless,this study was important for understanding structuresand behavior of noncovalent complexes produced byelectrospray ionization.

2. Experimental

Hen egg-white lysozyme (L) and all the glucosepolymers: maltose (M2), maltotriose (M3), maltotet-raose (M4), maltopentaose (M5), and maltohexaose(M6) were obtained from Sigma Chemical Co. (St.Louis, MO) and used without further purification. AllN-acetylglucosamine oligosaccharides: chitobiose(O2), chitotriose (O3), chitotetraose (O4), chitopen-taose (O5), and chitohexaose (O6) were obtained fromSeikagaku Corporation (Tokyo, Japan) and used with-out further purification. The electrospray solutionswere prepared by dissolving the protein and oligosac-charides (1:5) in 50:50 water:methanol solvent at alysozyme concentration of 1.03 1025 M, unlessotherwise specified.

The experiments were carried out on a home-builtelectrospray FTMS described in an earlier publication[19]. The ions produced from the needle travelthrough a heated capillary and a set of skimmers andlenses. Upon exiting the source, the ions are guidedthrough the fringing field of the 3.0 tesla supercon-ducting magnet by using rf-only quadrupole rods tothe ion cyclotron resonance (ICR) cell where the ionsare detected. The analyte solution is transported fromthe syringe pump to the electrospray needle at a flowrate of 5 mL/h in all experiments unless otherwisestated.

104 F. He et al./International Journal of Mass Spectrometry 193 (1999) 103–114

2.1. Heated capillary dissociation

Heated capillary dissociation (HCD) experimentswere carried out on a differentially heated capillary.The apparatus is shown in Scheme 1. The capillary ismounted inside a home-built electrospray sourcewhose operation is described elsewhere [20,21]. Onlythe relevant features of the newly designed capillarywill be described here. The capillary is made ofstainless steel with 1.9 mm o.d., 0.5 mm i.d., and 25cm length. The capillary is divided into two segmentsof equal length, with the one close to the electrosprayneedle (front end) designated as segment1, and theone close to the ICR cell (back end) designated assegment2. Both segment1 and segment2 are resis-tively heated by two dc power supplies. Two copperblocks connect the capillary and the leads on eachsegment. Each segment has a J-type thermocoupleconnected to the middle of the segment to monitor thetemperature during the experiment. Because of thedynamic nature of the system, a temperature gradientexists in the capillary. The temperature is measured inthe middle of each segment to monitor the hottestpoint. The front end of the capillary is attached to acounter electrode which is charged to 71 V. An O-ringseals the atmosphere from the stage 1 of the differen-tial pumping system. The back end of the capillarypoints to the hole of skimmer1, which opens toskimmer2 and pumping stage 2.

The temperature response of one segment when theother is heated has been characterized [21]. To ensurethat the dissociation is a gas-phase rather than asolution phase process, the HCD experiments are

performed by setting the temperature of the firstsegment to desolvate ions. The temperature on thesecond segment is then increased to perform thedissociation. We have earlier shown that the dissoci-ation process is a gas-phase process [21]. During theHCD experiments, all other conditions including nee-dle, capillary, lens, and skimmer voltages are heldconstant, unless specified. The temperature on seg-ment2 is monitored and the temperature at which thecomplex is no longer observed (,1%) is designatedas the dissociation temperature (Td). We have earliershown that Td is a relative measure of complexbinding strength [20,21].

For the HCD and CID experiments, an intensesignal of the intact complex is required. For theoligomers chitobiose, chitotriose, and chitotetraosepremixing the lysozyme and the oligosaccharide inthe same solution was sufficient. However, chitopen-taose and chitohexaose reacted too quickly to producethe intact complex. To obtain stronger signals ofintact complexes of chitopentaose and chitohexaose, a“Tee” type connector was used to mix the twosolutions immediately before the electrospray, asshown in Scheme 2. The Tee has a 0.75 mm bore(fitted for 1/16 in. tubing) with a length (horizontal) of23 mm.

The length from the center of the Tee connector tothe needle tip is 20 mm, and the inner diameter of thestainless steel needle is 100mm. With the typical flowrate of 5 mL/h, the mixing period of the two solutionsis about 0.5 s. Under these conditions, only a smallportion of the oligosaccharides are hydrolyzed before

Scheme 1. Differentially heated capillary.

Scheme 2. “Tee” configuration.

105F. He et al./International Journal of Mass Spectrometry 193 (1999) 103–114

introduction into the gas phase. With this setup,abundant signals corresponding to the desired com-plexes of chitopentaose and chitohexaose with ly-sozyme are observed.

2.2. Collision-induced dissociation

There are basically two types of CID experimentsin FTMS: on-resonance and sustained off-resonanceirradiation (SORI) [22]. Both were performed on thelysozyme complexes. For both methods, the complexions are isolated with a series of pulses and burstsproduced by an arbitrary waveform generator as partof the IonSpec data system. Ions are excited toincrease the translational energy and allowed to col-lide with gas pulsed to a maximum pressure of 531027 Torr. During collisions, translational energiesare converted into internal energies resulting in thedissociation of the complex ions. With on-resonanceCID, the parent ion was excited by a 0.5 ms rf pulsewith frequency corresponding to the cyclotron fre-quency. Ions were detected 6 s after the excitation toallow sufficient time for collisions. With SORI-CID,the ion was excited by a rf a 800 Hz above thecyclotron frequency for a duration of 1 s. Twonitrogen pulses in 500 ms intervals were applied tomaintain a pressure of about 53 1027 Torr duringthe CID event. SORI-CID results in the sequentialactivation of ions by multiple, low energy (,10 eV)collisions [22]. As a consequence, only small incre-ments of internal energy are deposited into the ionthroughout the duration of the event. This mode ofactivation is analogous to infrared multiphoton disso-ciation using low power lasers; it can be used as aprobe of the lowest decomposition pathway of ions.

3. Results and discussion

3.1. Complex formation

b-1,4 linked oligomers ofN-acetylglucosamine arenatural substrates of lysozyme. Table 1 lists theassociation constants [23] and the relative hydrolysisrate constants [24] of chitin oligosaccharides when

reacted with lysozyme. For chitotetraose (O4), chito-triose (O3), and chitobiose (O2), the hydrolysis ratesare very slow. Indeed, after mixing O4, O3, and O2

with lysozyme in 50:50 H2O/methanol solution, nohydrolysis products are observed even after 70 min.Fig. 1(a) shows the spectrum of lysozyme mixed withchitotetraose with a capillary temperature of 215 °C.The only complex peak observed belongs to the18charge state. Complexes of other charge states (19,17) are not observed in this spectrum but are ob-served under conditions of low capillary temperature(125 °C). However, they remain with very low abun-dances and dissociate readily when the temperature ofthe capillary is increased. Chitotriose (O3) and chito-biose (O2) yield similar results.

When lysozyme is mixed with chitopentaose (O5)and chitohexaose (O6), hydrolysis products are imme-diately observed. Fig. 1(b) shows the spectrum oflysozyme and chitohexaose (O6) dissolved in 50:50H2O/methanol at a capillary temperature of 220 °C.The hydrolysis products corresponding to O3 and O4

are observed complexed to lysozyme. The O4 com-plex is observed primarily in the18 charge state,while the O3 complex is observed in the17 chargestate. The intact complexes of chitohexaose [L:O6 17H]17 and [L:O6 1 8H]18 have very low intensities.Another product, O2, is observed complexed to ly-sozyme at low capillary temperatures (below 90 °C,spectrum not shown).

Fig. 2(a) shows the spectrum of lysozyme andchitohexaose using the Tee arrangement. The intact

Table 1Association constantsKassocand relative rate constants ofhydrolysis of hen egg white lysozyme andN-acetylglucosamineoligosaccharides complexes in solution; the deviation in thedissociation temperature is60.5 °C

Oligosaccharide Kassoc(M21)aRelative rateof hydrolysis Td (°C)

(GlcNAc)2 3.23 103 0 262(GlcNAc)3 1.13 105 1 335(GlcNAc)4 1.83 105 8 373(GlcNAc)5 ;(GlcNAc)4 4 000 405(GlcNAc)6 ;(GlcNAc)4 30 000 409(GlcNAc)8 ;(GlcNAc)4 30 000

a See [23].


complex peak corresponding to the19 charge state,[L:O6 1 9H]19, is the most abundant peak. The18charge state of the intact complex is about 40% of thebase peak. Also observed are the complexes of hy-drolysis products [L:O4 1 9H]19 and [L:O3 18H]18. The [L:O3 1 8H]18 is about 50% relative tothe base peak, while the intensity of [L:O4 1 9H]19

is significantly lower. It is clear that using the Teegreatly reduces the hydrolysis before ionization andenhances the signal intensity of the intact complexpeaks. In this article, all the HCD and CID resultswith chitopentaose and chitohexaose were obtained inthis manner.

For comparison, nonsubstrate oligosaccharides(glucose and its oligomers, Mn, n 5 2, 3, . . . , 6)were examined. It is known that lysozyme does nothydrolyze these glucose oligomers in solution. Whenthey are mixed with lysozyme and electrosprayed into

the gas phase, their complexes with lysozyme are alsoobserved. However, no hydrolysis products are ob-served even with the maltopentaose and maltohexaose(M5 and M6). Fig. 2(b) shows the spectrum oflysozyme with maltotetraose (M4). No hydrolysisproducts are observed. In addition to single adductcomplexes, e.g. [L:M4 1 9H]19 and [L:M4 18H]18, the double adduct species, [L:(M4)2 1 9H]19

and [L:(M4)2 1 8H]18 are also observed. Maltotrioseand maltose also show similar adducts however withsignificantly lower intensities.

The typical condition for lysozyme enzymaticactivity involves water at a pH 5.0 [12]. Complexformation followed by ESI were also performed underthese conditions. In general, all spectra obtained fromaqueous solution (pH 5.0) are similar to those ob-tained from 50:50 H2O/methanol solution. Abundantproducts are observed with O5 and O6, however, no

Fig. 1. ESI-FTMS spectra of lysozyme (13 1025 M) in 50:50H2O/methanol with (a) chitotetraose (O4) at capillary temperatureof 215 °C and (b) chitohexaose (O6) at capillary temperature of 220°C.

Fig. 2. ESI-FTMS spectra of lysozyme using the Tee (Scheme 2)with (a) chitohexaose (O6) and (b) maltotetraose (M4). See text foradditional information.


hydrolysis products are observed with the rest ofoligosaccharides, as before. Using the Tee connectoralso greatly increases the signal intensities of intactcomplex peaks of O5 and O6.

For comparison, denatured lysozyme was alsoexamined. Lysozyme has four disulfide bonds thatkeep the protein in a very compact conformation.Reducing the four disulfide bonds unfolds the protein[25]. The denatured lysozyme is prepared by boiling asolution of lysozyme in 0.02 M dithiothreitol for 30min immediately before it was electrosprayed. Aspectrum of denatured lysozyme with chitotetraose isshown in Fig. 3. The charge state distribution ofdenatured lysozyme in the mass spectrum is shifted tohigh charge states. The dominant peak is the114charge state rather than the18 charge state (Fig. 2)indicating that the denatured protein has an unfoldedconformation. The molecular weight obtained fromthe ESI mass spectrum is eight mass unit higher,resulting from the reduction of four disulfide bonds.Interestingly, there are no complexes of denaturedlysozyme and oligosaccharides observed. Evidently,denatured lysozyme does not form complexes witheither type of oligosaccharides.

4.2. Heated capillary dissociation

Fig. 4 shows the result of a typical HCD experi-ment. A solution of lysozyme was mixed with both O3

and O4 (1:2 lysozyme:oligosaccharide) and the spec-

trum obtained at different temperatures. At a temper-ature of 202 °C, lysozyme complexes of both oligo-saccharides are observed as the18 charge state [Fig.4(a)]. Complexes of other charge states such as17and 19 were observed at lower temperatures, albeitwith very low abundances. They dissociated readilywhen the temperature is increased. When temperatureis increased further to 220 °C, [L:O3 1 8H]81 isalmost fully dissociated [Fig. 4(b)]. The lysozymecomplex with chitotetraose [L:O4 1 8H]81 fully dis-sociates at a higher temperature (257 °C) Fig. 4(c). Nofragmentation of the individual moieties (protein andoligosaccharides) was observed. It is believed thatonly the non-covalent complex dissociates underthese conditions. For clarity, “fragment” will refer tofragmentation of individual moieties and “dissociate”will refer to the dissociation of the complex into twointact moieties.

Note that the [L:O4 1 8H]81 peak is already moreintense than [L:O3 1 8H]81 at low temperature [Fig.

Fig. 3. ESI-FTMS spectrum of denatured hen egg-white lysozymewith chitotetraose (O4) in H2O/methanol solution.

Fig. 4. ESI-FTMS spectra of lysozyme with chitotriose and chito-tetraose at different temperatures of segment2. (a) 202, (b) 220, (c)257 °C.


4(a)]. It should be pointed out that the lower intensityof the complex [L:O3 1 8H]81 does not necessarilygive it a lower Td. We have observed less intensesignals that produced higherTd than more intenseones in the HCD of peptides complexed to cyclodex-trins [20,21]. The relative intensities are likely areflection of the relative solution-phase affinities. It isknown that chitotetraose forms a stronger complexthan chitotriose with lysozyme in solution (Table 1).It has been suggested that the peak intensities in themass spectra are related to the relative binding affinityof noncovalent complexes in solution. Attempts toobtain solution-phase association constants by mea-suring peak intensities have been made [18]. Thecorrelation between solution-phase association con-

stants and the peak intensities further supports thisnotion.

By plotting the complex peak intensities versustemperatures, the HCD plot is obtained. Fig. 5 showsthe HCD behavior of the complex [L:O4 1 8H]81.This species fully dissociates at about 278 °C. Asimilar plot constructed for [L:O3 1 8H]81 producesa lower dissociation temperature of 220 °C. The plotsclearly show that [L:O3,4 1 8H]81 and [L:O4 18H]81 have distinct dissociation temperatures,Td.The summary of the results is tabulated in Table 2.The higherTd of the chitotetraose complex relative tothe chitotriose complex is consistent with their rela-tive stability in the solution.

Two intriguing trends are found in Table 2. First,

Fig. 5. HCD plot of lysozyme complexed with chitotriose [L:O3 1 8H]18 and chitotetraose [L:O4 1 8H]18. The dissociation temperature(Td) corresponds to 220 °C for O3 and 278 °C for O4.

Table 2Dissociation temperatures (Td) of substrate and nonsubstrate sugars complexed with hen white lysozyme, determined by heated capillarydissociation (HCD); the values have an uncertainty of65 °C

Sugar M2 M3 M4 M5 M6

Td (°C) ,200 274 277 327 331

Sugar O2 O3 O4 O5 O6

Td (°C) 262 335 373 405 409


the dissociation temperatures of complexes with sub-strate oligosaccharides (On) are consistently higherthan the corresponding nonsubstrate oligosaccharides(Mn). For instance, the maltohexaose complex disso-ciates at a temperature of 331 °C, while chitohexaosedissociates at 409 °C. The other oligosaccharides inTable 2 show similar behavior.

Second, the relative order of stability of substrateoligosaccharide agrees with theKassoc(Table 1). Therelative gas-phase stability follows the order O6 .

O5 . O4 . O3 . O2, which is the same order asKassoc. The fact that the results obtained in gas phaseagree well with those in solution phase might be anindication that lysozyme complexes have similarstructures and relative stability in both the gas phaseand the solution phase. As Dobson and co-worker[26] reported using1H nuclear magnetic resonance(NMR) studies, the specific binding are the conse-quence of interactions including hydrogen bonds totheN-acetyl group, rather than to the pyranose rings.They further concluded that the complex formationemploys an induced fit mechanism with the maindriving force being hydrogen bonding. This mayexplain the consistency of the relative stability be-tween gas and solution phase. It also suggests thatalthough solvation is one of the major driving forcesof complex formation of many noncovalent systems,it may not be as important in the lysozyme complexes.

4.3. Collision-induced dissociation

Both on-resonance and sustained off-resonanceirradiation (SORI) CID were performed on the ly-sozyme complexes. Fig. 6(a) shows an on-resonanceCID spectrum of the lysozyme-maltotriose (M3) com-plex. The parent peak of [L:M4 1 8H]81 was isolatedby a series of sweeps and bursts. An on-resonanceexcitation pulse with 12.5 V (base to peak) wasapplied to the parent peak at the same frequency as itsnatural cyclotron frequency. The only product ob-served was the [L1 8H]81 species, as shown in:

[L:M 3 1 8H]81OO3CID[L 1 8H]81 1 M3 (1)

Neither fragmentation of the lysozyme nor of theoligosaccharide is observed. Only dissociation occurswith the oligosaccharide leaving as a neutral species,which cannot be detected by mass spectrometry. Thesame results are obtained with lysozyme complexes ofother non-substrate oligosaccharides M2, M4, M5, andM6.

It is believed that the complexes of lysozyme andnonsubstrate oligosaccharides are not specific. It wassuggested by Henion that these nonspecific adductsare simple aggregation of the oligosaccharides on theprotein surface [18]. With no specific interactionbetween the enzyme and oligosaccharide, the nonsub-strate oligosaccharide dissociates from the proteinduring CID.

When CID is performed on the complexes[L:O3 1 8H]81 as shown in Fig. 6(b), the majorproduct observed was the deprotonated lysozyme

Fig. 6. On-resonance CID spectrum of (a) lysozyme–maltotriosecomplex with 12.5 V (base-peak) amplitude and (b) lysozyme–chitotetraose complex with 12.5 V (base-peak) amplitude.


[L 1 7H]71. The results correspond to a net chargetransfer from the lysozyme to the oligosaccharideduring the dissociation:

[L:O3 1 8H]81OO3CID[L 1 7H]71 1 [O3 1 H]1

(2)

However, the singly charged chitotriose was notobserved in the CID spectrum under the currentexperimental conditions. There are several reasons forthe absence of the protonated oligosaccharide. Theexperimental conditions optimized for detecting high-mass species may not work as well for low-massspecies. Similarly, the multiply charged protein mayexert a strong columbic repulsion towards the singlycharged oligosaccharide after it dissociates. Thelighter fragment then escapes from the ICR cell.Another possible reason is that the singly chargedoligosaccharide undergoes proton transfer with traceamount of water or methanol (from electrospraysource) in the ICR cell. The charged water andmethanol molecules are too small to be detected undercurrent conditions. It is not clear at this point which ofthese possibilities exists. The other substrate oligosac-charides O2, O4, and O5 yield the same dissociationbehavior.

SORI-CID was also performed on the same com-plexes. The same dissociation behavior was obtainedwith SORI-CID as with on-resonance CID. The factthat both on-resonance CID and SORI-CID yield thesame results suggests that “complex dissociation” isthe lowest pathway.

The reason that substrate and nonsubstrate oligo-saccharide complexes give different products duringCID can be rationalized by the structural differencesbetween specific and nonspecific complexes [26]. Inthe hydrolysis catalyzed by lysozyme, proton dona-tion is the first step of the process. It happens insidethe cavity where the Glu 35 group from the proteindonates a proton to the glycosidic bond of the sub-strate sugar, cleaving the bond as shown in Scheme 3[24]. This process occurs only when the substrate iscomplexed to the active site. In the gas phase, thisendothermic proton transfer reaction is facilitated by

the CID event. Similar endothermic proton transferreactions on other noncovalent complexes have alsobeen observed in this laboratory [27]. While protontransfer does occur with the smaller oligomers (O2,O3, O4, and O5), there is no glycosidic bond cleavageobserved. With the exception of O5, the oligomers aretoo short to be cleaved because cleavage generallyoccurs between the fourth and fifth residues. Theseobservations are consistent with solution-phase resultsthat showN-acetylglucosamine oligosaccharides withless than 5 units have significantly slower hydrolysisrates (Table 1). For both O5 and O6, the Glu 35 ispositioned close to a glycosidic bond so that bondcleavage occurs immediately after the proton transferin solution. In the gas phase, O6 shows cleavage of theoligosaccharide bond (vida supra); however, O5 doesnot. In solution, the O6 oligomer is cleaved at a ratethat is an order of magnitude larger than O5. Thedifferences in the gas phase fragmentation behavior ofthe two compounds may simply be a reflection of thelarge differences in the solution kinetics.

4.4. CID fragmentation of lysozyme–chitohexaose(O6) complex

CID of the lysozyme–chitohexaose complex yieldsfragments different from O2, O3, O4, and O5. WhenO6 is mixed with lysozyme by using the Tee config-uration, two abundant charge states of the intactcomplex are obtained: [L:O6 1 9H]19 and [L:O6 18H]18. Fig. 7 shows the CID results of the [L:O6 19H]19 complex. Note that the isolation of the com-

Scheme 3. Hydrolysis of substrate sugar by lysozyme.


plex peak in Fig. 7(a) is incomplete; the [O6 1 Na]1

and [L 1 10H]110 peaks are also present. There werea total of more than ten peaks including differentcharge states of lysozyme and its complexes presentin the original spectrum. Several species have cyclo-tron frequencies that are close to the desired peak. Itwas difficult to completely isolate the desired com-plex peak without attenuating the signal. The isolationspectrum shown in Fig. 7(a) was a compromisebetween complete isolation and optimal signal inten-sity. Nevertheless, the unejected ions do not interferewith the CID results. The two peaks [L1 10H]110

and [O6 1 Na]1 could not produce the fragmentsobserved.

Fig. 7(b) and (c) show the CID spectra at excitationamplitudes of 3.8 V (base-peak) and 4.0 V, respec-tively. Protonated chitohexaose [O6 1 H]1 and two

fragments B5 and B4 are observed. The proton on[O6 1 H]1 comes undoubtfully from the enzyme.The B5 and B4 are fragments of the oligosaccharide,corresponding to a five residue and a four residuesequence, respectively. The oligosaccharide fragmentproducts in the gas phase differ from the hydrolysisproducts in solution because there is no water mole-cule available. For yet undetermined reasons, thecomplement [L1 8H]18 product is not observed.

There are two possible pathways for the formationof the B5 and B4 fragments. They may fragmentdirectly from the complex after the proton transfer.Alternatively, they may be secondary fragments fromthe [O6 1 H]1 that has dissociated from the complex.The [O6 1 H]1 species from the complex may haveexcess energy allowing it to undergo subsequentfragmentation. It is not clear at this point whichmechanism is responsible for the B5 and B4 frag-ments.

CID on the18 charge state of the chitohexaosecomplex [L:O6 1 8H]18 was also performed. Thespectra at an amplitude 3.5 V (base-to-peak) is shownin Fig. 8. Neither dissociation nor fragmentation isobserved. Only ion loss is observed when the excita-tion amplitude is increased. The results suggest thatthere are some structural differences between the19and 18 charge states of the lysozyme–chitohexaosecomplexes. Williams and co-workers [2] have sug-gested that the conformer of the19 ion of lysozymeis consistent with the crystal structure. However, the

Fig. 7. CID of the lysozyme–chitohexaose complex [L:O6 19H]19. (a) Isolation. (b) SORI-CID with 3.8 V amplitude (base-peak). (c) SORI-CID with 4.0 V amplitude (base-peak).

Fig. 8. CID spectrum of lysozyme–chitohexaose complex [L:O6 18H]18 with 3.5 V amplitude (base-peak). No fragments are ob-served. Only the disappearance of the complex is observed withhigh amplitudes.


18 conformer has a similar structure to a denaturedone. In their studies, the19 ions were produceddirectly from the electrospray whereas the18 ionswere produced by charge stripping19 ions. Theyfurther concluded that charge stripping denatures theintermediate charge states. In this article, both the19and18 complexes are formed directly from electro-spray. It is possible that the18 charge state isinherently denatured.

5. Conclusion

The presence of the complex in the mass spectrumwith electrospray ionization is not a sufficient crite-rion for the presence of a specific protein–substrateinteraction in solution. The substrate chitose oli-gomers and the nonsubstrate maltose oligomers bothshow strong intensities for the lysozyme in theESI-MS spectra. However, HCD and CID are shownto be useful probes for characterizing gas-phase non-covalent complexes. The dissociation temperature ofthe substrate oligosaccharides are generally higherthan the similarly length nonsubstrate oligosaccha-rides. In addition, the relative order of stability ob-tained by HCD agrees well with solution phaseassociation constants for the substrate oligosaccha-rides suggesting that HCD may be generally useful forpredicting the relative strengths of association insolution.

Collision-induced dissociation of the lysozymesubstrate complex is also consistent with solutionphase results. Nonsubstrate oligosaccharides, pre-sumed to be nonspecifically bound, dissociate duringCID. Substrate oligosaccharides, presumed to be spe-cifically bound, dissociate by first transferring a pro-ton from the more basic protein to the less basicoligosaccharide. This endothermic proton transferreaction has been recently reported in other systems.For example, cyclodextrin–amino acid complexesproduce protonated cyclodextrin fragments upon CID.The reaction in the enzyme lysozyme is driven bycollision energy but is consistent with the putativemechanism for enzymatic degradation in solution.The reaction of chitohexaose, a known substrate for

lysozyme, in the gas phase is consistent with theknown mechanism in the solution phase. Fragmenta-tion of the oligosaccharide is observed presumablythrough the proton transfer and the proton catalyzedfragmentation of the glycosidic bond. The resultsfurther suggest that once the complex is formed, thereaction occurs even in the absence of solvent mole-cules.

Acknowledgement

The authors wish to thank the National ScienceFoundation for their continuing support.

References

[1] T.D. Wood, R.A. Chorush, F.M. Wampler III, D.P. Little,P.B. O’Connor, F.W. McLafferty, Proc. Natl. Acad. Sci. USA92 (1995) 2451.

[2] D.S. Gross, P.D. Schnier, S.E. Rodriguez-Cruz, C.K. Fager-quist, E.R. Williams, Proc. Natl. Acad. Sci. USA 93 (1996)3143.

[3] A. Miranker, C.V. Robinson, S.E. Radford, R.T. Aplin,Science 262 (1993) 896.

[4] H.R. Morris, P. Pucci, M. Panico, G. Marino, Biochemical J.268 (1990) 803.

[5] S.J. Valentine, D.E. Clemmer, J. Am. Chem. Soc. 119 (1997)3558.

[6] H. Vandael, P. Haezebrouck, L. Morozova, C. Aricomuendel,C.M. Dobson, Biochemistry 32 (1993) 11886.

[7] P.G. Wolynes, Proc. Natl. Acad. Sci. USA 92 (1995) 2426.[8] U.A. Mirza, B.T. Chait, Int. J. Mass Spectrom. Ion Processes

162 (1997) 173.[9] B.E. Winger, K.J. Light-Wahl, A.L. Rockwood, R.D. Smith,

J. Am. Chem. Soc. 114 (1992) 5897.[10] R.D. Smith, K.J. Light-Wahl, Biol. Mass Spectrom. 22 (1993)

493.[11] J.A. Loo, J.X. He, W.L. Cody, J. Am. Chem. Soc. 120 (1998)

4542.[12] P. Jolles, J. Jolles, J. Mol. Cell. Biochem. 63 (1984) 165.[13] J. Fernandez-Sousa, J. Gavilanes, A. Municio, A. Perez-

Aranda, R. Rodriguez, Eur. J. Biochem. 72 (1977) 25.[14] T. Imoto, L.N. Johnson, A.C.T. North, D.C. Phillips, J.A.

Rupley, The Enzymes, Academic, New York, 1972.[15] L.N. Johnson, J. Cheetham, P.J. McLaughlin, K.R. Acharya,

D. Barford, D.C. Phillips, Curr. Topic Microbiol. Immunol.139 (1988) 81.

[16] L.O. Ford, L.N. Johnson, P.A. Machin, D.C. Phillips, R.Tjian, J. Mol. Biol. 88 (1974) 349.


[17] N.C.J. Strynadka, M.N.G. James, J. Mol. Biol. 220 (1991) 401.[18] B. Ganem, Y.T. Li, J.D. Henion, J. Am. Chem. Soc. 20 (1991)

7818.[19] E.E. Gard, M.K. Green, H. Warren, E.J.O. Camara, F. He,

S.G. Penn, C.B. Lebrilla, Int. J. Mass Spectrom. Ion Processes158 (1996) 115.

[20] S.G. Penn, F. He, M.K. Green, C.B. Lebrilla, J. Am. Soc.Mass Spectrom. 8 (1997) 244.

[21] F. He, J. Ramirez, B.A. Garcia, C.B. Lebrilla, Int. J. MassSpectrom. Ion Processes 182/183 (1999) 261.

[22] J.W. Gauthier, T.R. Trautman, D.B. Jacobson, Anal. Chim.Acta 246 (1991) 211.

[23] M. Schindler, Y. Assaf, N. Sharon, D.M. Chipman, Biochem-istry 16 (1977) 423.

[24] L. Stryer, Biochemistry, Freeman, New York, 1995.[25] W. Fiedler, C. Borchers, M. Macht, S.O. Deininger, M.

Przybylski, Bioconjugate Chem. 9 (1998) 236.[26] K.J. Lumb, C.M. Dobson, J. Mol. Bio. 227 (1992) 9.[27] F. He, J. Ramirez, C.B. Lebrilla, J. Am. Chem. Soc. 121

(1999) 4726.


evidence for enzymatic activity in the absence of solvent in gas-phase complexes of lysozyme and...

Documents