Effect of stereochemistry on the electrospray ionization tandem mass spectra of transition metal chloride complexes of monosaccharides

JOURNAL OF MASS SPECTROMETRYJ. Mass Spectrom. 2005; 40: 1044–1054Published online 17 June 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.879

Effect of stereochemistry on the electrospray ionizationtandem mass spectra of transition metal chloridecomplexes of monosaccharides†

K. P. Madhusudanan,∗ Sanjeev Kanojiya and Brijesh Kumar

Sophisticated Analytical Instrument Facility, Central Drug Research Institute, Lucknow-226001, India

Received 10 December 2004; Accepted 10 April 2005

The effect of stereochemistry on the complexation of aldohexoses (glucose, mannose, galactose, alloseand talose) and ketohexoses (fructose, tagitose and sorbose) with transition metal chlorides (CoCl2,NiCl2, MnCl2 and ZnCl2) has been investigated by electrospray ionization tandem mass spectrometry.Electrospray ionization of methanolic solutions of hexoses containing metal chlorides gave abundantions corresponding to [M + MetCl]+ and [2M + MetCl]+ which on collision-induced dissociation gavecharacteristic fragment ions. The fragmentation pathways have been confirmed by examining methylglucoside and several isotopically labeled glucoses. Eliminations of H2O and HCl, C–C cleavages andelimination of metalhydroxychloride are the competing fragmentation pathways observed. All thesepathways seem to be influenced by the stereochemistry of the molecule. The fragmentation of the dimericcomplexes, [2M + MetCl]+, is also controlled by the stereochemistry of the molecule. The abundance ofthe product ions corresponding to elimination of HCl is found to increase with increasing number of axialhydroxyl groups in aldohexoses. [2M + MetCl]+ dissociates by elimination of HCl followed by C2H4O2 inaldohexose complexes and by elimination of HCl followed by C3H6O3 in ketohexose complexes. Copyright 2005 John Wiley & Sons, Ltd.

KEYWORDS: monosaccharides; electrospray ionization; tandem mass spectrometry; stereochemical differentiation

INTRODUCTION

Carbohydrates are involved in a wide variety of cellularinteractions.1 Determination of molecular structure of carbo-hydrates has been an active research area. Nuclear magneticresonance (NMR) and mass spectrometry (MS) play impor-tant roles in carbohydrate analysis. Initially, gas phase ioniza-tion techniques combined with synthetic modifications havebeen used for obtaining structural and stereochemical infor-mation from mass spectrometry.2,3 There have been severalreports on the use of desorption ionization techniques suchas fast atom bombardment (FAB) for the determination ofsequences and stereochemical linkages of oligosaccharides.4,5

More recently, electrospray ionization (ESI)6 and matrix-assisted laser desorption ionization (MALDI)7 have foundwide application in carbohydrate analysis.8 – 10

Adduct ions have been widely used for signal enhance-ment and promotion of fragmentation in soft ionizationtechniques.11 – 13 Metal ions can form adducts with a varietyof compounds and subsequently be transferred to the gasphase.14 – 16 Sugars are no exception, and a large number ofreports indicate that metal cationization of carbohydrates

ŁCorrespondence to: K. P. Madhusudanan, SophisticatedAnalytical Instrument Facility, Central Drug Research Institute,Lucknow-226001, India. E-mail: [email protected],[email protected]†CDRI Communication No. 6690.

during ESI followed by MS/MS produces data rich in struc-tural and stereochemical information.17 – 22 Stable cationizedmolecules can be generated reproducibly by suitable choiceof metallic reagent ions. Upon collision-induced dissocia-tion (CID) such metal ion adducts give more cross ringcleavage products than those involving proton addition.It is the cross-ring cleavages that give information aboutthe linkage position in oligosaccharides. These are thoughtto involve charge-remote processes.23 Fragmentation notinvolving cross-ring cleavages are expected to show stere-ochemical differences. For example, loss of water could bestereospecific as axial OH is known to be lost preferentially.

AgC-, CuC- and Pb2C-cationization of monosaccharideshas been found to give useful structural and stereochemicalinformation.24 – 27 Leary et al.28,29 have pioneered the useof auxiliary ligands along with metal cationization todifferentiate stereoisomers of mono and disaccharides.However, data on the use of transition metal chloridesalone to achieve stereochemical differentiation is scarce inliterature. Tabet et al.30 – 32 recently studied the ESI tandemmass spectra of [M C FeCl]C and [2M C FeCl]C ions andreported stereochemical differentiation of glucose, mannose,galactose and talose. No data on other transition metalssuch as Co, Ni, Mn and Zn are available. Moreover, sincethese workers employed ion trap mass spectrometry, no datawas available in the mass range below m/z 180. The focusof the present study is to examine the MS/MS spectra of

Copyright 2005 John Wiley & Sons, Ltd.

Complexation of monosaccharides with transition metal chlorides 1045

ionic complexes formed by CoCl2, NiCl2, MnCl2 and ZnCl2

with aldohexoses D-glucose (1), D-mannose (2), D-galactose(3), D-allose (4), D-talose (5), methyl-˛-D-glucopyranoside(6) and ketohexoses D-fructose (7), D-tagatose (8) and L-sorbose (9) and to explore the possibility of stereochemicaldifferentiation of these diastereomers (Table 1). In orderto confirm the different fragmentation pathways, severalisotopically labeled glucose have also been employed togenerate MS/MS data. These include D-glucose-d5 (10), [1-13C]-D-glucose (11), [2-13C]-D-glucose (12), [6-13C]-D-glucose(13) [1,2-13C2]-D-glucose (14), [1,6-13C2]-D-glucose (15) and[6,6-d2]-D-glucose (16).

EXPERIMENTAL

The aldohexoses (1–5), methyl-˛-D-glucopyranoside (6),ketohexoses (7–9), the isotopically labeled aldohexoses(11–16) and methanol-d employed in this study wereobtained from Sigma (Sigma Aldrich, Milwaukee, WI, USA).They were used as such. D-glucose-d5 (10) was generated byexchange with methanol-d.

The electrospray mass spectra and tandem mass spectrawere recorded on a Micromass Quattro II triple-quadrupolemass spectrometer (Micromass, UK) equipped with an ESIsource. Heated dry nitrogen at 80 °C was used as thenebulizing and drying gas at flow rates of 10 and 250 l h�1

respectively. The sample solutions were prepared in HPLCgrade methanol containing the respective transition metalchloride at a concentration of approximately 10�4 M. Theconcentration of the sample and the metal salt was in theratio 10 : 1. For D-labelling studies, methanol was substitutedwith MeOD. The samples were infused into the ESI sourcefrom a Harvard Apparatus Model 11 syringe pump at aflow rate of 5 µl min�1. The ESI source potentials were:capillary 3.5 kV, lens at 0.5 kV, cone at 40 V unless otherwisestated and skimmer lens offset at 5 V. The mass spectrometerwas scanned at the rate of 300 m/z units per second. Dataacquisition and processing were carried out using MassLynx3.3 software supplied with the instrument. Data acquisition

Table 1. Ring substituent configuration of aldohexoses andketohexoses

Monosaccharide C2-OH C3-OH C4-OH

D-Glucose (1) Equatorial Equatorial EquatorialD-Mannose (2) Axial Equatorial EquatorialD-Galactose (3) Equatorial Equatorial AxialD-Allose (4) Equatorial Axial EquatorialD-Talose (5) Axial Equatorial AxialD-Fructose (7) – Equatorial EquatorialD-Tagatose (8) – Axial EquatorialL-Sorbose (9) – Axial Axial

was performed in the multichannel analyzer (MCA) modeand the spectra reported were accumulated over 10 scans.Calibration was carried out using NaI/CsI solution. CID ofvarious ions was performed in an r.f-only hexapole lens.Argon was admitted at a pressure so as to achieve 30%transmission of the precursor ion. The collision energy was8–14 eV and was set for each compound at a value that gavea precursor ion intensity of approximately 50% of that of themost abundant product ion. The tandem mass spectra werecollected by scanning the second quadrupole analyzer at ascan rate of 300 m/z units per second in the MCA mode. Thespectra were accumulated over 25 scans.

RESULTS AND DISCUSSION

The ESI mass spectra of the monosaccharides recordedin presence of metal chlorides such as CoCl2, NiCl2,MnCl2 and ZnCl2, showed a number of ionic complexes.Prominent among them were the doubly charged ions[Mn C Met]2C (where n is in the range 2–5 and Metrepresents metal) and the singly charged ions [M C Met �H]C, [M C MetCl]C, [2M C Met � H]C, [2M C MetCl]C and[3M C MetCl]C. In most cases, the [2M C MetCl]C ionswere the most abundant among these. Other multimericspecies such as [3M C MetCl]C or [4M C MetCl]C were notprominent. It appears that MetClC is able to effectivelychelate with two molecules of hexose resulting in a stableionic complex. The normal ESI mass spectra were similarand therefore could not be used for gleaning structuraland stereochemical information. It was therefore decidedto explore the possibilities of obtaining structurally andstereochemically useful information from CID experiments.The tandem mass spectra of [Mn C Met]2C and [3M CMetCl]C ions only showed the loss of monosaccharide unitand hence were unable to provide any useful information,whereas [M C MetCl]C and [2M C MetCl]C ions did givediagnostic fragment ions upon CID. Hence these spectra arediscussed separately.

[M + MetCl]+

The MS/MS spectra of [M C MetCl]C (where Met representsCo, Ni, Mn and Zn) ions were complex giving a numberof fragment ions. As representative examples, the MS/MSproduct ion spectra of [M C CoCl]C ions of (1–9) aregiven in Fig. 1. Glucose (1) is taken as an example forexplaining the fragmentation pathways. As the resolutionof the first quadrupole analyzer was sufficiently lowered toaccommodate both 35Cl and 37Cl isotopes in the selectedwindow, the tandem mass spectra gave both 35Cl and37Cl isotopic peaks wherever these were present. Hencethe presence of chlorine in the fragments could be easilydiscerned in the tandem mass spectra of the metal complexes.This is applicable to CoClC and MnClC complexes, whereas inthe case of NiClC and ZnClC complexes there is interferencefrom Ni and Zn isotopes. MetClC can, in principle, bindto any of the electron-rich centers of the monosaccharides.Accordingly, MetClC will be complexed with the varioushydroxyl groups and the ring oxygen in different waysresulting in a mixed population of structures. Among the

Copyright 2005 John Wiley & Sons, Ltd. J. Mass Spectrom. 2005; 40: 1044–1054

1046 K. P. Madhusudanan, S. Kanojiya and B. Kumar

30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300

m/z

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

214190178

8573

15414512797 109103 162 174

184202

196 208

274220 256238

22085

69 73

178145127

9791 109 118 154 174163

190184 214202196

274256238

190184178

97918573 127109 154145 162 174

220

214202196

274238 256

19018414512785

57 73 9791 109103 115 154 178163 172

274256220202

196 214208238

238

22019018485 12791 202274

256

214190184

1781541278573 97 109 145 168162202

196 210

288256220

234 270

238190184

1451278573 97 178154 163220202

196 214208

256274

238

18485 14512797 163154 178

190 220202

208 214

274256

238

190184145127 220202274

256

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

12 eV

14 eV

14 eV

13 eV

14 eV

12 eV

13 eV

12 eV

12 eV

Figure 1. Product ion tandem mass spectra of [M C CoCl]C of m/z 274 of (a) 1, (b) 2, (c) 3, (d) 4, (e) 5, (g) 7, (h) 8 and (i) 9 and (f) ofm/z 288 of 6 at collision energies given in the figure.

Table 2. Product ion tandem mass spectral data from isotopically labeled hexoses, m/z values followed by abundances inparentheses

[M C CoCl]C Neutral losses

Compd. C2H4O2 HCl C H2O C CH2O C3H6O3 HCl C C2H4O2 C4H8O4

D-glucose-d5 (10) 217 (100) 192 (60) 186 (83) 180 (48) 155 (18)[1-13C]-D-glucose (11) 214 (100) 191 (85) 184 (64) 178 (50),179 (45) 154 (24)[2-13C]-D-glucose (12) 214 (100) 191 (86) 184 (65) 178 (50),179 (48) 154 (22)[6-13C]-D-glucose (13) 215 (100) 190 (65) 185 (60) 178 (12),179 (60) 154 (5),155 (13)[1,2-13C2]-D-glucose (14) 214 (100) 192 (86) 184 (78) 178 (37),180 (43) 154 (22)[1,6-13C2]-D-glucose (15) 215 (100) 191 (63) 185 (52) 179 (48),180 (18) 154 (3),155 (13)[6,6-d2]-D-glucose (16) 216 (100) 190 (65) 186 (57) 178 (12),180 (60) 154 (8),156 (18)

five hydroxyl groups in D-glucose, the anomeric hydroxyl isknown to be the most acidic.33,34 Because of the flexibility,the hydroxy-methyl groups will be involved in complexationalong with the ring oxygen and other hydroxyl groups.35

The reactivity of these complexes will depend upon thetypes of coordination possible and the preservation of theintramolecular hydrogen bonds.35 In order to confirm thedifferent fragmentation pathways, tandem mass spectra werealso collected from labeled glucose (10–16), and the majorions and their abundances are tabulated in Table 2.

From an examination of the spectra the followingcompeting fragmentation pathways involving neutral lossescould be envisaged: (1) elimination of H2O, (2) eliminationof HCl, (3) elimination of CnH2nOn resulting from C–Ccleavages and (4) elimination of metal hydroxychloridefollowed by loss of H2O. No bare reduced metal is observedin the tandem mass spectra of the metal complexes as thehydroxyl groups of the hexoses provide a variety of metal

binding sites involving three or more oxygens to coordinatewith the metal. The methyl glucoside gave a spectrum similarto that of D-glucose except for the fact that elimination ofMeOH is seen in place of loss of H2O.

Loss of H2OAs the molecule contains several hydroxy groups, multiplelosses of water could be expected. Therefore, the fragmentions of m/z 256, 238, 220 and 202 could possibly correspondto successive eliminations of water (Fig. 1a). In the D-exchanged spectrum, the m/z values of these ions areshifted to m/z 259, 239/242, 222 and 202, respectively. Itis clear, therefore, that the ion of m/z 238 in the tandemmass spectrum of [M C CoCl]C is a composite ion consistingof both [M C CoCl � 2H2O]C and [M C CoCl � HCl]C or[M C Co � H]C. The ions of m/z 220 and 202 do not seemto contain chlorine as its isotopic pattern is absent andhence these ions correspond to losses of H2O and 2H2O,



Scheme 1. Fragmentation pathways of [1 C CoCl]C of m/z 274 and [6 C CoCl]C of m/z 288.

respectively from [M C Co � H]C. The ion of m/z 220 couldalso arise by elimination of HCl from [M C CoCl � H2O]C.This was confirmed by precursor ion analysis of theions of m/z 220, which showed both [M C Co � H]C and[M C CoCl � H2O]C as its precursors.

It is reasonable to assume that the anomeric hydroxylis eliminated as water from [M C CoCl]C ion resultingin 1,2-anhydro sugar complex, similar to the reportedelimination of H2O from AgC-complexes of glucose andglucopyranosides.24,25 This was confirmed by the followingobservations. In the D-exchange spectrum, D2O eliminationis seen in place of H2O. There was no significant loss ofH2O in the tandem mass spectrum of methyl glucoside (6).In the spectrum of (6), the corresponding loss is MeOHresulting in an ion having the same m/z value of 256.However, in the other diastereomers H2O loss may occurfrom different sites in each complex depending on therelative axial and equatorial configurations of the hydroxylgroups.36 The CID spectra of [M C CoCl � H2O]C of m/z 256from (1) and [M C CoCl � MeOH]C of m/z 256 from (6) weresimilar (data not shown). The major fragmentations wereelimination of HCl leading to the ion of m/z 220 followed byloss of CH2O involving C6, as confirmed from experimentswith [6-13C]-D-glucose (13) and [6,6-d2]-D-glucose (16). Theresulting ion of m/z 190 upon CID gave the cross-ring

cleavage peaks at m/z 130 and 118 (Scheme 1). In Scheme 1and in the subsequent schemes, the entries in parenthesesare the m/z values of the ions in the D-exchanged spectrum.The indicated cleavages are confirmed by CID of thecorresponding ion from labeled glucose. The cross-ringcleavage ions have m/z values of 119/130, 118/131, 118/132and 118/130, respectively in (10), (11), (14) and (16).

Loss of HClAs stated earlier, the ion having an m/z value of 238could be due to elimination of two molecules of wateror it could be due to loss of HCl. This was further con-firmed by examining the tandem mass spectra of both[M C Co35Cl]C and [M C Co37Cl]C by raising the resolutionof the precursor ion selection to accept only a one mass unitwindow. The tandem mass spectrum of [M C Co37Cl]C ionof m/z 276 shows two product ions of m/z 240 and 238 cor-responding to elimination of 2H2O and H37Cl, respectively.Similar experiments were carried out for all the compounds,and the ratios [M C Met37Cl–H37Cl]C/[M C Met37Cl–H2O]C

and [M C Met37Cl–H37Cl]C/[M C Met37Cl–2H2O]C are tab-ulated in Table 3. Only data from CoClC and MnClC com-plexes are included because of possible interference fromNi and Zn isotopes in the data from their complexes. Fromthe data in Table 3 it is clear that elimination of HCl is



Table 3. Ion abundances and the abundance ratios of [M C Met37Cl–H37Cl]C/[M C Met37Cl–H2O]C and[M C Met37Cl–H37Cl]C/[M C Met37Cl–2H2O]C in the product ion tandem mass spectra of [M C Met37Cl]C ions of (1)–(5) and(7)–(9)

Compounds

Abundance ratio 1 2 3 4 5 7 8 9

[M C Co37Cl–H37Cl]C/[M C Co37Cl–H2O]C 0.24 0.41 1.07 0.09 16.2 1.67 1.14 6.7[M C Co37Cl–H37Cl]C/[M C Co37Cl–2H2O]C 0.93 2.75 4.67 0.37 17.2 8.45 4.8 14.8[M C Mn37Cl–H37Cl]C/[M C Mn37Cl–H2O]C 0.05 0.15 0.48 0.12 20.4 0.36 0.62 0.67[M C Mn37Cl–H37Cl]C/[M C Mn37Cl–2H2O]C 0.17 0.66 1.56 0.5 16.4 3.66 2.17 2.2

Scheme 2. Elimination of HCl from [1 C CoCl]C.

preferred in (5), among the aldohexoses. Similarly, HCl elim-ination is the most facile in (9) among the ketohexoses. Itappears that the presence of several axial hydroxyl groupsleads to increased elimination of HCl. The stereochemicalimplications are discussed later.

Loss of HCl also triggers further fragmentation since theionic complex now corresponds to [M C Met � H]C. Thesefragmentations involve losses of H2O and CH2O and theelimination of C2H4O2. The ion of m/z 190 corresponds to theformer. As already discussed, CH2O elimination involves C6and the two unexchangeable hydrogens on C6 as observedearlier for fragmentation of FeClC complexes.30 – 32 Precursorion analysis shows that the ion of m/z 190 has severalprecursors (m/z 220, 238, 256 and 274) suggesting therebythat both stepwise and concomitant processes are operating.The precursor of m/z 238 corresponds to elimination of HCland not 2H2O as the precursor ion has an m/z value of242 corresponding to elimination of DCl in D-exchangedspectrum. Elimination of C2H4O2 after loss of HCl leads to theion of m/z 178. This loss corresponds to C2H2D2O2 in the D-exchanged spectrum suggesting thereby that two replaceablehydrogens are involved in this loss. Three possibilities can beenvisaged for the elimination of C2H4O2, involving C1–C2,C3–C4 and C5–C6 as in the illustrative example shownin Scheme 2. From the data in Table 2 it is clear that thiselimination of C2H4O2 following loss of HCl can involveeither C1–C2, C3–C4 or C5–C6 as the corresponding peak isobserved at m/z 178, 179 and 180 depending on the isotopicsubstitution. It appears, therefore, that HCl elimination

occurs from multiple sites, and the resulting [M C Co � H]C

ion can eliminate C2H4O2 from C1–C2, C3–C4 or C5–C6depending upon the site of cobalt attachment.

C–C cleavagesAnother competing pathway is the C–C cleavages leadingto the direct elimination of 60-Da (C2H4O2), 90-Da (C3H6O3)and 120-Da (C4H8O4) and resulting in the ions of m/z 214,184 and 154 respectively, in the aldohexoses. Only the lattertwo processes (eliminations of C3H6O3 and C4H8O4) occurin the ketohexoses. But these are not as prominent as thosein the aldohexoses. Similar eliminations have been reportedearlier in the FAB36 and ESI MS37 of monosaccharides. Theseeliminations could correspond to C2–C3, C3–C4 and C4–C5cleavages resulting in 0,2A, 0,3A and 0,4A ions,38 respectively.Complexes of carbohydrates with an auxiliary ligand andtransition metals are known to promote C–C cleavages.39

From the data of isotopically labeled compounds given inTable 2 it is clear that elimination of C2H4O2 from [1 C CoCl]C

involves C1 and C2 exclusively and hence corresponds to C2-C3 cleavage. The C2–C3 cleavage, being very specific, has thepotential for stereochemical differentiation (discussed later).Similarly, elimination of C3H6O3 involves C1, C2 and C3,whereas loss of C4H8O4 involves C1–C2 and either C3–C4 orC5–C6. However, elimination of C5-C6 appears to be muchless probable than that of C3–C4. As expected, the ions ofm/z 214, 184 and 154 fragment by elimination of HCl to givethe ions of m/z 178, 148 and 118, respectively. The C2–C3



[3+CoCl]+

0

2

4

6

8

10

12

14

16

1 3 5 7 9 11 13 15 17 19 21 23 25

Collision energy, eV

Abu

ndan

ce

[2+CoCl]+

0

2

4

6

8

10

12

14

16

18

1 3 5 7 9 11 13 15 17 19 21 23 25


Abu

ndan

ce

[4+CoCl]+

02468

101214161820

1 3 5 7 9 11 13 15 17 19 21 23 25


Abu

ndan

ce

[1+CoCl]+

02468

101214161820

1 3 5 7 9 11 13 15 17 19 21 23 25

Collsion energy, eV

Abu

ndan

ce(a) (b)

(c) (d)

m/z 274

m/z 184m/z 154

m/z 190

m/z 214

m/z 256

m/z 274

m/z 256

m/z 238

m/z 184

m/z 190

m/z 154

m/z 214

m/z 274

m/z 256

m/z 190

m/z 184

m/z 154m/z 145

m/z 274

m/z 256m/z 214

m/z 184m/z 154

m/z 238

[8+CoCl]+

0

5

10

15

20

25

30

1 3 5 7 9 11 13 15 17 19 21 23 25


Abu

ndan

ce

[9+CoCl]+

0

5

10

15

20

25

30

35

40

1 3 5 7 9 11 13 15 17 19 21 23 25


Abu

ndan

ce

[7+CoCl]+

0

5

10

15

20

25

30

35

1 3 5 7 9 11 13 15 17 19 21 23 25


Abu

ndan

ce

(f)

(g) (h)

[5+CoCl]+

0

5

10

15

20

25

30

35

40

45

1 3 5 7 9 11 13 15 17 19 21 23 25


Abu

ndan

ce

(e) m/z 274m/z 238

m/z 256m/z 184

m/z 274m/z 238

m/z 190

m/z 256 m/z 154

m/z 274

m/z 256 m/z 238

m/z 190

m/z 274

m/z 238

m/z 256

m/z 190

m/z 184

m/z 145

Figure 2. Selected ion abundances in the product ion tandem mass spectra of [M C CoCl]C of m/z 274 as a function of collisionenergy: (a) 1, (b) 2, (c) 3, (d) 4, (e) 5, (f) 7, (g) 8 and (h) 9. ♦, �, �, ð, *, ž, C, -represent the ion series m/z 145, 154, 184, 190, 214,238, 256 and 274 in (a–e). ♦, �, �, ð, *, ž, C represent the ion series m/z 145, 154, 184, 190, 238, 256 and 274 in (f–h).



and C3-C4 cleavage products also fragment by loss of CH2Oresulting in the ions at m/z 184 and 154, respectively. Thecorresponding peaks appeared at m/z 185 and m/z 154/155in (13) and at m/z 186 and m/z 154/156 in (16), respectively.It is clear, therefore, that the elimination of CH2O after C2-C3cleavage involves only C3, whereas after C3-C4 cleavage, lossof CH2O can involve either C4 or C6. This is in contrast to theintact molecule losing CH2O exclusively from C6 followingeliminations of HCl and H2O from [M C CoCl]C.

Loss of metalThere is evidence for the direct elimination of metal from[M C MetCl]C ions. Elimination of metal as Met(OH)Clinvolves the abstraction of OH by the metal leading tothe characteristic gluconium ion of m/z 163, which furtherfragments by consecutive eliminations of H2O resulting inthe ions of m/z 145 and 127 similar to the gluconium iongenerated from ammonium adduct of glucose. In the D-exchanged spectra these ions have m/z values of 167, 147 and128 confirming the assignment. Precursor ion analysis showsthe direct formation of these ions from [M C MetCl]C ions.Co and Mn complexes show this ion of m/z 163 only in (2)and (4) among the aldohexoses. The metal loss fragments aremost abundant in ZnClC complexes, whereas Ni complexesgive abundance values in between. The ketohexoses arecharacterized by fairly abundant gluconium ions from Co,Ni, Mn and Zn complexes. This is understandable in viewof the fact that in the ketohexoses, the hydroxyl group is lostfrom a tertiary carbon.

Effect of collision energySelected ion abundances in the tandem mass spectra of[M C CoCl]C of m/z 274 as a function of collision energy(breakdown graph) is given in Fig. 2 for aldohexoses (1)–(5)and ketohexoses (7–9). In order to enhance the readabilityof the plots, the abundance range is adjusted according tothe abundances of the fragment ions. The abundance ofthe precursor ion (m/z 274) is in the range of 70–80% at acollision energy of 5 eV (laboratory frame), but drops downsharply to below 5% as the collision energy is increased to14 eV except for mannose and talose which require highercollision energies of 16 and 17 eV, respectively to reachthe 5% level. In glucose–CoClC complex, the eliminationof H2O (m/z 256) and C2-C3 cleavage (m/z 214) appear tobe the lowest energy fragmentations as these eliminationsare the only fragmentations that are prominent in the lowcollision energy range up to 8 eV, after which C3-C4 cleavage(m/z 184) and elimination of HCl followed by losses of waterand CH2O (m/z 190) become prominent. Finally, in the laststage C4-C5 cleavage (m/z 154) takes over. The same trendis observed in the mannose complex. However, the C2-C3 cleavage is less prominent than H2O loss. In galactosecomplex the lowest energy processes appear to be lossesof H2O and HCl and having equal prominence, whereas inallose complex H2O loss is the lowest energy process. Inthe talose complex, elimination of HCl is the most facileprocess and it is approximately six-fold more abundant thanelimination of H2O. Elimination of (HCl C H2O C CH2O)and C3-C4 cleavage are the most prominent processes in the

galactose complex. Elimination of metal hydroxychlorideand H2O are facile in mannose and allose complexes.

Among the ketohexoses, elimination of HCl is the mostprominent in sorbose, followed by fructose and tagitose,whereas loss of H2O follows the order tagitose > fructose >sorbose. Losses of H2O and HCl appear to be the lowestenergy processes in fructose and tagitose, whereas in sorboseonly elimination of HCl appears to be the lowest energyprocess.

Effect of metalThe availability of MS/MS data for the CoClC, NiClC,MnClC and ZnClC complexes enables one to study theeffect of metal on the different fragmentation pathways.The fragmentation processes were similar with all the abovemetal complexes. However, the relative contributions of thevarious fragmentation processes were different. The effect ofcollision energy on fragmentation in NiClC complex (data notshown) is similar to that observed for the CoClC complex andthe lowest energy processes are loss of H2O (m/z 255) andC2-C3 cleavage (m/z 213). This is followed by C3-C4 cleavage(m/z 183), elimination of (HCl C H2O C CH2O), m/z 189 andby elimination of HCl, which is the least favoured inZnClC complexes and most favored in CoClC complexes.A similar behavior has previously been reported for Co andZn-coordinated oligosaccharides.21 Higher collision energyleads to abundant C4-C5 cleavage product (m/z 153). Thesame trend is observed in MnClC adducts. C–C cleavages arequite abundant in CoClC and NiClC complexes. The propertyof divalent Co and Ni to produce cross-ring cleavages hasbeen highlighted in previous reports.26,27 The abundances ofthe ions resulting from elimination of metal hydroxychlorideare plotted in Fig. 3 for the MetClC complexes of Co, Ni, Mnand Zn. The elimination of metal as metal hydroxychlorideis the most favored in ZnClC complexes. According toIrving–Wallace series, the formation of Zn-O bond is favoredover the formation of other metal-O bonds.40 This mayexplain the facile loss of ZnCl(OH). In general, eliminationof H2O and metal hydroxychloride are more prominent inketoses than in aldoses.

[2M + MetCl]+

As stated earlier, these hexoses give highly abundant dimericcomplexes with metal chlorides. It was thought that CID

Elimination of MetCl(OH)

0

2

4

6

8

10

12

14

Co Ni Mn Zn

Metal

Abu

ndan

ce

1

2

3

4

5

7

8

9

8

8

4 822

Figure 3. Elimination of Met(OH)Cl expressed as % totalionization in the tandem mass spectra of MetClC complexes of(1)–(5) and 7–9.



Table 4. Ion abundances in the tandem mass spectra of [2M C CoCl]C at m/z 454 of aldohexoses andketohexoses (collision energy, 8 eV)

Neutral losses m/z 1 2 3 4 5 7 8 9

HCl 418 5.3 17.2 45.9 35.2 69.9 48.6 42.5 38.4HCl, C2H4O2 358 26.2 14.4 12.5 18.2 4.7 – – –HCl, C3H6O3 328 – – – – – 18.9 13.1 18.8HCl, C4H8O4 298 8.1 2.4 3.3 4.5 2.1 – – –M 274 26.5 19.0 10.3 7.5 12.0 13.3 18.6 29.0

of these complexes should give stereospecific cleavages.A study of the effect of collision energies on the tandemmass spectra of [2M C MetCl]C ions showed that consistentstereochemical differences could be observed in the collisionenergy range 5–11 eV. Hence, the data were compared at acollision energy of 8 eV as given in Table 4 for [2M C CoCl]C

complexes. The breakdown pattern of NiClC complexeswas the same as of CoClC complexes, whereas the MnClC

and ZnClC dimeric complexes dissociated predominantlyto the monomeric complexes. Unlike the complex tandemmass spectra of [M C MetCl]C ions, the spectra of [2M CMetCl]C ions are simple with easily assigned fragments.A similar behavior of monomeric and dimeric flavonoidmetal complexes has been observed by Brodbelt et al.41 Theelimination of C2H4O2 is a fragmentation characteristic ofthe aldohexoses, whereas it is completely absent in theketohexoses. In its place, elimination of C3H6O3 is seen inthe tandem mass spectra of the [2M C MetCl]C complexes ofketohexoses. This is understandable in view of the fact thatthere is a highly specific cleavage of the C2-C3 or C3-C4 bondin aldohexoses and ketohexoses, respectively. Elimination ofC4H8O4 is less significant in aldohexoses and is of very minorimportance in the ketohexoses.

From the data of the isotopically labeled sugars it is clearthat the elimination of C2H4O2 following loss of HCl involvesC1 and C2 as the resulting ion has an m/z value of 360 in (14)showing the loss of label on C1 and C2, whereas eliminationof the second molecule of C2H4O2 may involve C6 to theextent of 10% as the second elimination of C2H4O2 resultsin two peaks at m/z 299 and 300 in the ratio 1 : 10 in (13),m/z 298 and 300 in the ratio 1 : 10 in (14), at m/z 300 and 301in the ratio 1 : 5 in (15) and at m/z 300 and 302 in the ratio1 : 10 in (16). From the data of the mixed dimers of (14) with(11) and (13) (Scheme 3) it is seen that the first loss of C2H4O2

could occur from either of the molecule, whereas the secondloss of C2H4O2 could occur from either the same molecule orthe other.

Stereochemical effectsThe tandem mass spectra of the metal chloride complexesof hexoses show stereospecific fragmentation resulting insignificant abundance variations. This should result fromthe various preferred sites for coordination possible in thesediastereomers. Leavell et al.42 have reported from ion mobil-ity studies and density functional theory calculations onZn-ligand-hexose diastereomers that glucose and galactosecomplexes may have metal coordination to the O(6) hydroxylgroup, whereas mannose may have metal coordination to the

Scheme 3. Elimination of HCl and C2H4O2 from [11 C 14 CCoCl]C and [13 C 14 C CoCl]C of m/z 457.

O(2) hydroxyl group. The breakdown graphs given in Fig. 2reflect the differences in the population of metal coordi-nated structures for the CoClC adducts of (1)–(5) and (7–9).The most significant and consistent difference that couldbe noticed is in the elimination of C2H4O2. Elimination ofC2H4O2 is absent in the ketohexose complexes, whereas itis favored to different extents in the aldohexose complexes.The abundances of C2-C3 cleavage ions in the tandem massspectra of [M C MetCl]C ions of (1)–(5) are plotted in Fig. 4afor the metals Co, Ni, Mn and Zn. As already stated, the elim-ination of C2H4O2 is a very specific one comprising C1 andC2 and involving cleavages of C1-O and C2-C3. The stereo-chemistry of the different hydroxyl groups appears to playan important role in controlling the fragmentation. The abun-dance of the resulting ion seems to decrease with increasingnumber of axial hydroxyl groups. Among aldohexoses it isthe most significant in glucose and least significant in talose,and this dependence of fragmentation on stereochemistrycould help in stereochemical differentiation of aldohexoses.The differences are consistent and significant in the case ofCoClC, NiClC and MnClC complexes where the C2-C3 cleav-age follows the order glucose >mannose >galactose >allose>talose. Similar trends are seen in ZnClC complexes, butthe differences are less significant. The presence of an axialhydroxyl at C4 has previously been reported to favor C3-C4 cleavage during CID of galctose-Zn-diethylenetriaminecomplex.37 A similar trend is also seen in the present study.The ratios of the abundances of the product ions of C3-C4and C2-C3 cleavages in the tandem mass spectra of CoClC

adducts of (1)–(5) measured at a collision energy of 12 eVwere: 0.7, 0.7, 3.1, 5.6 and 4.1 in (1), (2), (3), (4), (5), respec-tively. The presence of a C4 or C3 axial hydroxyl group



Table 5. Ion abundances ratios HCl loss and HCl loss followed by C2H4O2 or C3H6O3 in the product iontandem mass spectra [2M C MetCl]C ions of (1)–(5) and (7)–(9) (collision energy, 8 eV)

Ratio of the losses Metal 1 2 3 4 5 7 8 9

HCl/HCl, C2H4O2 Co 0.2 1.2 3.7 1.9 14.9HCl/HCl, C3H6O3 2.8 3.2 2.0HCl/HCl, C2H4O2 Ni 0.1 0.9 2.3 0.6 7.7HCl/HCl, C3H6O3 1.7 1.9 1.3HCl/HCl, C2H4O2 Mn 0.2 0.6 1.2 0.4 1.9HCl/HCl, C3H6O3 0.6 1.0 0.7

Elimination of C2H4O2

0

5

10

15

20

Co Ni Mn Zn

Metal

Abu

ndan

ce

(a)

1

23

4

Elimination of HCl

0

20

40

60

80

Co Ni Mn

Metal

1

2

3

4

5

(b) 5

Figure 4. (a) Elimination of C2H4O2 expressed as % totalionization in the tandem spectra of [M C MetCl]C ions of (1)–(5)at a collision energy of 12 eV and (b) Elimination of HClexpressed as % total ionization in the tandem spectra of[2M C MetCl]C ions of (1)–(5) at a collision energy of 8 eV.

seems to favor C3-C4 cleavage and accordingly these ratiosare higher in galactose, allose and talose, which contain eitherC4 or C3 axial hydroxyl group.

The competing losses of water and HCl are expected tobe stereospecific. The ratios of the losses of H2O and HClare different in the diastereoisomers and can help in theirdifferentiation using any of the metal complexes studied.The ratio of losses of HCl and H2O from [M C CoCl]C variesfrom 0.09 in allose to 16.2 in talose, whereas the changeis not so conspicuous in the ketohexoses (Table 3). Similardifferences can also be seen in the MnClC complexes. Theratio of losses HCl and 2H2O also show difference amongthe diastereomers.

The CoClC, NiClC and MnClC dimeric adducts givecharacteristic tandem mass spectra showing significantdifferences among the diastereomers. It is clear from Table 4that elimination of HCl from [2M C MetCl]C resulting in[2M C Co � H]C is controlled by the stereochemistry of themolecule, that is, by the number of axial hydroxy groupsas observed for the monomeric complex. The abundances of

[2M + Met – H]C ions in the tandem mass spectra of [2M+ MetCl]C ions of aldohexoses are plotted in Fig. 4b for themetals Co, Ni and Mn. The loss of HCl is least favoredin glucose, whereas it is increasingly favored in mannose,allose, galactose, and talose. The trend is just the reverseof the one observed for C2-C3 cleavage from [M C MetCl]C

except for the change of the order of galactose and allose.It appears that formation of [2M C Co � H]C by eliminationof HCl is favored by the presence of axial hydroxy groups.Among the C2-, C3- and C4-hydroxy groups, the C4-hydroxygroup has the maximum influence on the elimination of HClas galactose gives an abundance that is next to that of talose,which has two axial hydroxy groups at C2 and C4. The sameorder is seen in Ni and Mn complexes except for the factthat allose and mannose were indistinguishable when themetal was Mn. The stereochemical differences are maximalwhen the metal is Co. Loss of HCl is quite abundant inketohexoses, but the differences are not significant in (7)–(9)for all the metal complexes studied. Table 5 gives the ratiosof losses of HCl and HCl followed by C2H4O2 for (1)–(5)and of HCl and HCl followed by C3H6O3 for (7)–(9). SinceHCl elimination is not at all favored in Zn complexes,data from Zn complexes are not included in the table. Itis seen that significant differences could be observed forthe diastereomers of aldohexoses. Among the aldohexosesthe ratio increases in the order glucose <mannose <allose<galactose <talose when the metal is Co.

CONCLUSIONS

From our study on the ESI tandem mass spectra oftransition metal chloride complexes of monosaccharides, itis clear that the stereochemistry of the molecule influencesthe complexation and the fragmentation of the resultingcomplexes. Upon CID of [M C MetCl]C ions, elimination ofH2O is preferred in mannose and allose, whereas eliminationof HCl is favored in talose. C2-C3 cleavage gives rise tothe most abundant fragment in glucose, whereas C3-C4cleavage is preferred in galactose. C2-C3 cleavage seems tobe favored as the number of axial hydroxyl groups decreases.Elimination of HCl from [2M C MetCl]C also follow almostthe same order. The ketohexoses are characterized byfavored eliminations of HCl from both [M C MetCl]C and[2M C MetCl]C ions. The aldohexoses and ketohexoses canalso be easily distinguished by the favored C2-C3 cleavage inaldohexoses and C3-C4 cleavage in ketohexoses. Among themetal complexes studied CoClC complexes seem to offer themaximum stereochemical differences. Elimination of metal



hydroxychloride is favored in ZnClC complexes owing to theease of formation of Zn-O bond.

AcknowledgementGrateful acknowledgement is made to SAIF, CDRI, Lucknow wherethe mass spectrometric studies were carried out.

REFERENCES

1. Stryer L. Biochemistry. W.H. Freeman: New York, 1988; 331.2. Richter WJ, Mueller DR, Domon B. Tandem mass spectrometry

in structural characterization of oligosaccharide residues inglycoconjugates. Methods Enzymol. 1990; 193: 607.

3. Chaplin MF, Kennedy JF (eds). Carbohydrate Analysis: A PracticalApproach, 2nd edn. Oxford: New York. 1994.

4. Hofmeister GE, Zhou Z, Leary JA. Linkage position determina-tion in lithium cationized disaccharides: Tandem mass spec-trometry and semiempirical calculations. J. Am. Chem. Soc. 1991;113: 5964.

5. Perrault H, Costello CE. Stereochemical effects on the massspectrometric behaviour of native and derivatised trisaccharideisomers: comparison with results from molecular modeling. J.Mass Spectrom. 1999; 34: 184.

6. Fenn JB, Mann M, Meng CK, Wong SF. Electrospray ionization-principles and practice. Mass Spectrom. Rev. 1990; 9: 37.

7. Karas M, Bachmann D, Bahr U, Hillenkamp F. Matrix assistedultraviolet laser desorption of non-volatile compounds. Int. J.Mass Spectrom. Ion Processes 1987; 78: 53.

8. Domon B, Mueller DR, Richter WJ. High performance tandemmass spectrometry for sequence branching and interglycosidiclinkage analysis of peracetylated oligosaccharides. Biol. Environ.Mass Spectrom. 1990; 19: 390.

9. Zaia J. Mass spectrometry of oligosaccharides. Mass Spectrom.Rev. 2004; 23: 161.

10. Quemener B, Ralet M-C. Evidence for linkage positiondetermination in known feruloylated mono- and disaccharidesusing electrospray ion trap mass spectrometry. J. Mass Spectrom.2004; 39: 1153.

11. Teesch LM, Adams J. Intrinsic interactions between alkalineearth metal ions and peptides: a gas-phase study. J. Am. Chem.Soc. 1991; 113: 812.

12. Alvarez EJ, Brodbelt JS. Evaluation of metal complexation asan alternative to protonation for electrospray ionization ofpharmaceutical compounds. J. Am. Soc. Mass Spectrom. 1998;9: 463.

13. Teesch LM, Adams J. Fragmentations of gas-phase complexesbetween alkali metal ions and peptides: metal ion binding tocarbonyl oxygens and other neutral functional groups. J. Am.Chem. Soc. 1990; 112: 4110.

14. Aduru S, Chait BT. Californium -252 plasma desorption massspectrometry of oligosaccharides and glycoconjugates: controlof ionization and fragmentation. Anal. Chem. 1991; 63: 1621.

15. Rodgers MT, Campbell SA, Beauchamp JL. Site-specific lithiumion attachment directs low energy dissociation pathways ofdinucleotides in the gas phase. Applications to nucleic acidsequencing by mass spectrometry. Int. J. Mass Spectrom. IonProcesses 1997; 161: 193.

16. Teesch LM, Adams J. Metal ions as special reagents in analyticalmass spectrometry. Org. Mass Spectrom. 1992; 27: 931.

17. Zhou Z, Ogden S, Leary JA. Linkage position determinationin oligosaccharides: MS/MS study of lithium cationizedcarbohydrates. J. Org. Chem. 1990; 55: 5444.

18. Kohler M, Leary JA. LC/MS/MS of carbohydrates with postcolumn addition of metal chlorides using a triaxial electrosprayprobe. Anal. Chem. 1995; 67: 3501.

19. Orlando R, Bush CA, Feneselau C. Structural analysis ofoligosaccharides by tandem mass spectrometry: collisionalactivation of sodium adduct ions. Biol. Environ. Mass Spectrom.1990; 19: 747.

20. Konig S, Leary JA. Evidence for linkage position determinationin cobalt coordinated pentasaccharides using ion trap massspectrometry. J. Am. Soc. Mass Spectrom. 1998; 9: 1125.

21. Sible EM, Brimmer SP, Leary JA. Interaction of 1st rowtransition metals with ˛-1,3-, ˛-1,6-mannotriose and conservedtrimannosyl core oligosaccharides: a comparative electrosprayionization study of doubly and singly charged complexes. J. Am.Soc. Mass Spectrom. 1997; 8: 32.

22. Harvey DJ. Ionization and collision-induced fragmentation ofN-linked and related carbohydrates using divalent cations. J.Am. Soc. Mass Spectrom. 2001; 12: 926.

23. Cancilla MT, Wong AW, Voss LR, Lebrilla CB. Fragmentationreactions in the mass spectrometry analysis of neuraloligosaccharides. Anal. Chem. 1999; 71: 3206.

24. Salpin J-Y, Boutreau L, Haldys V, Tortajada J. Gas-phasereactivity of glycosides and methyl glycosides with CuC, AgCand Pb2C ions by fast-atom bombardment and tandem massspectrometry. Eur. J. Mass Spectrom. 2001; 7: 321.

25. Berjeaud JM, Couderc F, Prome JC. Sterochemically controlleddecomposition of silver cationized methyl glucosides. Org. MassSpectrom. 1993; 28: 455.

26. Salpin J-Y, Tortajada J. Structural characterization of hexoses andpentoses using lead cationization. An electrospray ionizationtandem mass spectrometric study. Eur. J. Mass Spectrom. 2002; 8:379.

27. Boutreau L, Leon E, Salpin J-Y, Amekraz B, Moulin C,Tortajada J. Gas-phase reactivity of silver and coppercoordinated monosaccharide cations studied by electrosprayionization and tandem mass spectrometry. Eur. J. Mass Spectrom.2003; 9: 377.

28. Desaire H, Leary JA. The effects of coordination number andligand size on the gas phase dissociation and stereochemicaldifferentiation of cobalt-coordinated monosaccharides. Int. J.Mass Spectrom. 2001; 209: 171.

29. Gaucher SP, Leary JA. Determining anomericity of the glycosidicbond in Zn(II)-diethylenetriamine-disaccharide complexesusing MSn in a quadrupole ion trap. J. Am. Soc. Mass Spectrom.1999; 10: 269.

30. Carlesso V, Fournier F, Tabet J-C. Stereochemical differentiationof four mono saccharides using transition metal complexes byelectrospray ionization/ion trap mass spectrometry. Eur. J. MassSpectrom. 2000; 6: 421.

31. Carlesso V, Fournier F, Afonso C, Tabet J-C. Halogen counter-ion effect on the dissociation of monosaccharide-iron complexesgenerated by electrospray ionization combined with an ion trapmass spectrometer. Eur. J. Mass Spectrom. 2001; 7: 331.

32. Carlesso V, Afonso C, Fournier F, Tabet J-C. Stereochemicaleffects from doubly charged iron clusters for the structuralelucidation of diastereomeric monosaccharides using ESI/IT-MS. Int. J. Mass Spectrom. 2002; 219: 559.

33. Cai Y, Cole RB. Stabilization of anionic adducts in negative ionelectrospray ionization mass spectrometry. Anal. Chem. 2002; 74:985.

34. Salpin J-Y, Tortajada J. Gas-phase acidity of D-glucose. A densityfunctional theory study. J. Mass Spectrom. 2004; 39: 930.

35. Zeng YJ, Ornstein RL, Leary JA. A density functional theoryinvestigation of metal binding sites in monosaccharides. J. Mol.Structure (Theochem.) 1997; 233: 389.

36. Smith G, Leary JA. Mechanistic studies of diastereomericnickel(II) N-glycoside complexes using tandem massspectrometry. J. Am. Chem. Soc. 1998; 120: 13 046.

37. Gaucher SP, Leary JA. Stereochemical differentiation ofmannose, glucose, galactose and talose using zinc(II)diethylenetriamine and ESI-Ion trap mass spectrometry. Anal.Chem. 1998; 70: 3009.

38. Domon B, Castello C. A systematic nomenclature forcarbohydrate fragmentation in FAB MS/MS of glycoconjugates.Glycoconj. J. 1988; 5: 397.

39. Gaucher SP, Leary JA. Influence of metal ion and coordinationgeometry on the gas-phase dissociation and stereochemical



differentiation of N-glycosides. Int. J. Mass Spectrom. 2000; 197:139.

40. Lippard JS, Berg JM. Principles of Bioinorganic Chemistry.University Sciences Books: Mill Valley, 1994.

41. Davis BD, Brodbelt J. Determination of glycosylation site offlavonoid monoglucosides by metal complexation and tandemmass spectrometry. J. Am. Soc. Mass Spectrom. 2004; 15: 1287.

42. Leavell MD, Gaucher SP, Leary JA, Taraszka JA, Clemmer DE.Conformational studies of Zn-ligand-hexose diastereomersusing ion mobility measurements and density functional theorycalculations. J. Am. Soc. Mass Spectrom. 2002; 13: 284.


Effect of stereochemistry on the electrospray ionization tandem mass spectra of transition metal chloride complexes of monosaccharides

Documents