low-energy collision-induced dissociation of b1-type sugar ions formed from peracetylated methyl...

JOURNAL OF MASS SPECTROMETRYJ. Mass Spectrom. 34, 346È353 (1999)

Low-energy Collision-induced Dissociation of B1-

Sugar Ions Formed from PeracetylatedTypeMethyl Pentosides and Methyl 6-Deoxyhexosides

P. Vouros,1¤, D. R. Mu� ller,*2 and W. J. Richter2”1 Department of Chemistry, Northeastern University, Boston, Massachusetts 02115, USA2 Novartis Pharma AG, Functional Genomics Area, Protein Sciences, CH-4002 Basle, Switzerland

Low-energy collision-induced dissociation (CID) was explored for the purpose of di†erentiating between isomericpertrideuteroacetylated aldopentosides, 6-deoxyaldohexosides and their respective furanosidic and pyranosidic ana-logues. The 1-O-methylated derivatives were analyzed by gas chromatography/chemical ionization tandem(CH

4)

mass spectrometry. ions (formally C(1) carbenium ions), which are expected to reÑect the core structure of theB1parent sugar and which were produced by the elimination of methanol from the protonated glycoside derivatives,

were used as precursor ions to obtain the corresponding CID spectra. In general, strong similarities which impedean unequivocal di†erentiation of the individual constituents were found in the spectral patterns of the di†erentisomeric pairs. However, examination of the pseudo-breakdown behavior of the ions over the range 5–35 eVB

1revealed improved distinction windows which provide for a fuller di†erentiating capability between Ðve-and six-membered ring isomers including even the most problematic ribose derivatives. The reliability of the analyses isenhanced by the high reproducibility (s Æ 10% ) of these ratios. The considerable similarity between the tandemmass spectra of certain furanosidic and pyranosidic ions is highly suggestive of a partial loss of structuralB

1integrity, mainly due to ring contraction, which occurs to di†erent extents in the di†erent isomers. CopyrightJohn Wiley & Sons, Ltd.( 1999

KEYWORDS: peracetylated methyl pentosides ; peracetylated methyl 6-deoxyhexosides ; low-energy collision-induceddissociation ; ring size isomers ; isomer distinction

INTRODUCTION

Di†erentiation of ring size isomers of monosaccharidicglycosides can often be accomplished by electron ion-ization (EI) mass spectrometry after preparation of suit-able chemical derivatives. In the familiar O-peracetatesof hexopyranosides and hexofuranosides, highly diag-nostic a-cleavages are triggered under these conditionsby the ring-O-atoms that induce losses of either C(6) orC(5)] C(6) in the form of orCH2OAc

radicals, respectively, and therebyCH(OAc)CH2OAcindicate six-or Ðve-membered ring carbohydrate struc-tures (Scheme 1).1,2 An analogous behavior is observedfor peracetylated ribose3 but, in general, these impor-tant ions are not always sufficiently abundant to allowsafe assignments of sugar ring size when only smallamounts of sample of moderate or low purity are avail-

¤ On leave of absence from the Department of Chemistry, North-eastern University, Boston MA 02115, USA.

* Correspondence to : R. Novartis Pharma AG, Func-D. Mu� ller,tional Genomics Area, Protein Sciences, K-127.1.34, CH-4002 Basle,Switzerland.

E-mail : dieter.muller=pharma.novartis.com” Part of this paper was completed after the passing of WJR. The

coauthors hope that any subsequent revisions also reÑect the opinionof WJR.

able. By contrast, sugar ions of the type, i.e. formallyB1C(1) carbenium ions resulting from cleavage of the gly-cosidic substituents (Scheme 2), can often be producedfrom mono-and smaller oligosaccharide peracetates invery high yield, especially when relatively soft ionizationtechniques are used.4h6 While ions as such do notB1indicate sugar ring size directly, their subsequent frag-mentation, which was for instance the basis for the dif-ferentiation of peracetylated methyl hexosides,7h10should o†er a good possibility of allowing distinctionby speciÐc spectral di†erences. With respect to theanalysis of composite samples or samples of limitedpurity, it was therefore of special interest to investigatewhether collision-induced dissociation (CID) of ionsB1can provide an e†ective alternative to EI. The obviousprerequisite of such an approach is, of course, that thelong-lived ions mass selected for CID in tandemB1mass spectrometry (MS/MS) retain their original ringstructure or, at least, resemble that structure to a suffi-cient extent.

In this study, capillary gas chromatography (GC)/MS/MS was employed for the chromatographicresolution of a/b-anomeric mixtures of the methyl-pyranoside educts and the corresponding furanosidesoften formed along with the anomers during 1-O-methylation. This also ensured that these isomers wereanalyzed under exactly the same conditions. MethaneCI (GC/CI-MS/MS) was used for generating the pre-B1

CCC 1076È5174/99/040346È08 $17.50 Received 29 September 1998Copyright ( 1999 John Wiley & Sons, Ltd. Accepted 29 October 1998

LOW-ENERGY CID OF B1-TYPE SUGAR IONS 347

Scheme 1.

cursor ions in the MS/MS experiments. To facilitate dis-crimination between similar reaction paths in isomers,CID at low collision energies (5È30 eV) was chosen,

Scheme 2.

Scheme 3. Structures of pentose and deoxyhexose isomersinvestigated (R

1¼CD

3CO; R

2¼CH

3)

since this energy regime allows for subtle variation ofion internal energies. Thus, otherwise small spectral dif-ferences may possibly be enhanced in order to optimizedistinction criteria. Di†erentiation of ring-size isomers(and stereoisomers) based on MS/MS of ions has, inB1fact, been reported earlier for the peracetylated fouraldopentoses (D-ribose, D-arabinose, D-xylose and D-lyxose) using metastable ion decompositions. Inter-estingly, use of high-energy CID (collision energies inthe 10 keV range) probed in the same work was foundunsuitable for di†erentiating between these isomers.

The aldopentose and 6-deoxyaldohexose derivativesinvestigated here under low-energy CID conditions areshown in Scheme 3 (furanosidic and/or pyranosidicstructures IÈIII and IVÈVI), and included the Ðrst threeof the above pentoses and one 6-deoxyaldohexose as p/fpairs. The two sets of monosaccharides having threenon-anomeric hydroxyl groups, are discussed in thatsequence. For simplicity, these compounds will some-times be referred to in terms of the underivatized parentsugar molecules from which they are derived (e.g.“riboseÏ), but their 1-O-methyl 2,3,4-or 2,3,5-tri-O-acylderivatives are always implied. Per(trideutero)acetylinstead of unlabeled peracetyl derivatives were usedwith advantage because otherwise overlapping fragmen-tation channels became resolved. This is in factobserved for parallel “isobaricÏ losses of (CD3CO)2Oand Whenever possible, 1-O-CD3COOH/CD2CO.methyl pentosides were employed in order to ensureregiospeciÐc formation of C(1) carbenium ions by massselecting [MH[ MeOH]` precursors for CID. Whenanalysing [M ] H [ AcOH]` ions of corresponding1,2,3,4-or 1,2,3,5-tetra-O-acetates (R1\ R2 \CD3COin Scheme 3) of pyrano-and furanopentosides, respec-tively, such “regiospeciÐcÏ mass selection would not be apriori warranted.

EXPERIMENTAL

Methyl pentosides and methyl-6-deoxyhexosides wereeither obtained from Sigma (Buchs, Switzerland) or pre-pared as mixtures of furanosides and pyranosides bymethylation of the respective parent compounds over-night at room temperature using 1 M MeOHÈHCl.After neutralization with and evaporation toAg2CO3dryness under a stream of argon, samples were reactedwith perdeuteroacetic anhydrideÈpyridine, neutralizedwith sodium hydrogencarbonate and extracted with

Neutralization and extraction were repeatedCHCl3.twice and residual pyridine was removed by evapo-ration with toluene. Furanosides and pyranosides wereassigned by GC/MS comparison with derivatives of oneauthentic isomer using a Finnigan TSQ 70 triple-quadrupole mass spectrometer. Isomers were separatedon a glass capillary column coated with OV-1701 usinghelium as carrier gas. After 1 min isothermal chroma-tography at 50 ¡C, temperature was increased to 220 ¡Cat a rate of 5 ¡C min~1. The GC effluent was ionized byCI (70 eV electron energy) at a pressure opti-(CH4)mized to keep the ratio m/z 15 : 17 at \0.05. Selectedparent ions were dissociated in the GC/MS/MS experi-ments by collisions with argon at a pressure of 1 mTorr(indicated) (1 Torr\ 133.3 Pa) . MS/MS correction was

Copyright ( 1999 John Wiley & Sons, Ltd. J. Mass Spectrom. 34, 346È353 (1999)

348 P. VOUROS, D. R. MU� LLER AND W. J. RICHTER

Figure 1. 15 eV CID spectra of ions (m /z 268) of pyranose and furanose isomers of the aldopentose derivatives I–III.B1

set to zero. Product ions were scanned from m/z 20 to300 in 1 s.

RESULTS

CID behavior of ions derived fromB1trideuteroacetylated 1-O-methyl pentofuranosides and

-pyranosides (I–III)

The 15 eV CID spectra of the C(1) carbenium ions gen-erated from the p-and f-isomers of IÈIII ions at m/z(B1

268, produced by CI as shown in Scheme 2 are(CH4)shown in Fig. 1. The 15 eV collision energy was chosenhere as an intermediate value representative of thegeneral behavior of these derivatives within the 5È30 eVenergy range investigated. It is also the energy at whichthe relative abundance of the major fragment ion [B1(m/z 142) tends to maximize for all[ 2CD3COOH]`substrates. The spectrum of only one anomer isincluded as the ions of a-and b-forms showed, for theB1most part, almost identical spectra.

Conspicuously at Ðrst sight, the spectra of all threefuranosides I( f ), II( f ) III( f ) and look almost the same inspite of di†erent stereochemistry. Very similar to the



Figure 2. Pseudo-breakdown curves of ions (m /z 268) of pyranose and furanose isomers of the aldopentose derivatives I–III in theB1

range 5–35 eV ( Ã,m /z 97; m /z 98; m /z 142; m /z 161; m /z 205; m /z 268).>, …, +, …, =,

spectra of the furanosides is also that of the ribopyrano-side I(p), while those of the two other six-membered ringisomers, the arabino-and xylopyranosides II(p) andIII(p), show increasingly pronounced di†erences. In par-ticular, it is observed that ions at m/z 205, 161 and 97increase steadily within the pyranosides relative to afairly constant m/z 98 and, of course, m/z 142 (basepeak). On this basis, abundance ratios for pairs of ionsrepresenting the “constantÏ part of the spectrum relativeto the “variableÏ part, e.g. m/z 142 : 205 or m/z 98 : 97,would appear to provide the most promising criteria fordi†erentiating some of the isomers from certain othersin a more detailed analysis. For instance, the m/z 98 : 97and m/z 142 : 205 ratios change distinctly from about 3to 1.5 to 0.75 and from about 41 to 12 to 4, respectively,on going from I(p) via II(p) and III(p) i.e. they reÑectmarkedly the variations in stereochemistry and ring sizeexcept for I(p) vs. I( f ) (or the other furanosides).

In line with the latter considerations, the full di†eren-tiating capability of the approach can be better evalu-

ated when the additional dimension of data availablefrom the collision energy variation between 5 and 35 eVis included with the objective of locating possiblyimproved distinction windows. Figure 2 summarizes thebreakdown behavior of the ions of the p/f pairs ofB1the three pentoses and includes all major fragment ionsexcept m/z 46 which gains signiÐcance only(CD3CO`),at energies in the upper range. Conclusions similar tothose above can be drawn from these extended data,indicating that the 15 eV Ðndings were highly represen-tative of the general behavior of the ions within theB1probed energy range. In particular, it was found that (i)ion (m/z 205) is more prominent at[B1[ CD3COOH]low energies in the spectra of all isomers and decaysgradually above 5 eV. As already observed at 15 eV, m/z205 is favored in the spectra of the six-membered ringover the Ðve-membered ring isomers throughout theentire energy range, although again to a much lesserextent for I(p) than for II(p) and especially III(p). (ii) Theprincipal fragment ion (m/z 142,[B1[ 2CD3COOH]



Figure 3. Peak intensity ratio of m /z 142 : 205 ions (ÍB1

in the CID spectra of theÉ2CD3COOHË½ :ÍB

1ÉCD

3COOHË½) B

1ions of the aldopentose derivatives I–III in the range 5–25 eV.

base peak in all spectra) reaches maximum abundancevalues at about 15 eV for all isomers. Maximum TICcontributions (%&) of this ion decrease from the furano-sides (D50%) to the pyranosides (46% down to 38%).This trend is opposite and, thus, complementary to thatof m/z 205. (iii) Within the six isomers the m/z 98 : 97abundance ratios for the furanosides I( f ) to III( f )remain the highest throughout the energy range (risingfrom 1.4 to 10 with increasing energy) and are closelyparalleled by that of I(p). Conversely, the xylopyrano-

Figure 4. Energy dependence of the ratio of summed ion currents(I(m /z 142) ½ I(m /z 98))/(I(m /z 205) ½ I(m /z 161) ½ I(m /z 97)) inthe CID spectra of the ions of the aldopentose derivatives I–IIIB

1in the range 10–30 eV.

side III(p) maintains its much lower m/z 98 : 97 ratio (0.6up to about 1.3) throughout, while the arabinopyranoseII(p) keeps an intermediate ratio.

(iv) The m/z 142 : 205 ratio was found to provide themost suitable windows for di†erentiating the six isomersunder consideration when the energy range 30È35 eV isexcluded because of very low abundance of m/z 205.This is illustrated in Fig. 3, which shows the m/z142 : 205 ratio plotted as a function of collision energyfor each of the three p/f pairs. Even for ring-size di†er-entiation within the difficult I(p)/I( f ) pair (upper panel)these ratios di†er by a factor of D2. Except for very lowand very high collision energies, they provide distinctand conveniently measurable di†erences over most ofthe investigated energy range (cf. discussion ofreproducibility).

In general, the complete breakdown behavior empha-sizes again the relatively close spectral similarity alreadynoted at 15 eV for the furanosides I( f ), II( f ), III( f ) andthe pyranoside I(p) compared with each other, and thepronounced spectral di†erences that prevail withrespect to each of the other two pyranosides, II(p) andIII(p), over the whole energy range. This similarity“clusterÏ of I( f ), II( f ), III( f ) and the pyranoside I(p) vsthe “outliersÏ II(p) and III(p) is further emphasized whenratios of the sums of the obviously correlating ions[&(m/z 142 ] 98) and &(m/z 205 ] 161 ] 97)] are con-sidered. These ratios show remarkably little change withincreasing collision energy and reÑect virtually constantcontributions of the sums to fragment ion currentdespite the widely varying contributions of the individ-ual ions. The ratios are higest and closely “stackedÏ forthe furanosides and the ribopyranoside, but markedlydistant from the similarly constant, but much lower,values for the remaining isomers (cf. Fig. 4 and Dis-cussion section).

CID behavior of ions derived fromB1trideuteroacetylated 1-O-methyl 6-deoxyhexosides

(IV–VI)

Of the chosen deoxyhexoses (D-6-deoxyglucose, L-rhamnose, L-fucose ; for structures see Scheme 3) onlydeoxyglucose was available to us as a p/f pair (IV(p) andIV( f )). Comparison of the 15 eV CID spectra of the B1ions of these substrates (m/z 282, Fig. 5) with those ofthe above pentoses (m/z 268, Fig.1) shows generally aparallel behavior between the two sets of studied com-pounds. Note, for example, the low relative abundanceof the m/z 219 ion in the spectrum([B1[ CD3COOH])of the deoxyglucofuranoside IV( f ) compared with thoseof the three pyranosides IV(p), V(p) and VI(p).

In fact, a more detailed consideration of thepseudo-breakdown curves of the 6-deoxyhexoses (datanot shown) revealed that, in analogy with the above m/z142 : 205 ratio ([B1[ 2CD3COOH] : [B1now the ratio m/z 156 : 219 provides a[ CD3COOH]),potential window for accentuating spectral di†erencesbetween these isomers. A plot of this latter ratio overthe range from 10È25 eV (Fig. 6) illustrates this pointand shows that now the p-isomers resemble each otherfairly closely. (Energies in excess of 25 eV are not



Figure 5. 15 eV CID spectra of ions (m /z 282) derived from the 6-deoxyaldohexose derivatives IV–VI .B1

included because of the extremely low abundance of them/z 219 ion beyond this point.)

Reproducibility of data

Given the small di†erences in signal intensities in someof the spectral patterns, the reproducibility of the low-energy CID spectra becomes a matter of prime impor-tance. Since ratios such as m/z 142 : 205 or m/z 156 : 219are found useful for isomer di†erentiation, we under-took a rigorous examination of the relevant data. Thus,given that, from a practical point of view, it is not pos-sible to conduct analyses over a large range of CIDenergies, we limited the assessment of reproducibilityprimarily to that at 15 eV, including only one of thevery similar furanosides.

Figure 6. Peak intensity ratio of m /z 156 : 219 ions (ÍB1

in the CID spectra of theÉ2CD3COOHË½ : ÍB

1ÉCD

3COOHË½)

ions of the 6-deoxyaldohexose derivatives IV–VI in the rangeB1

10–25 eV.



Table 1. Peak intensity ratio of m/z 142 : 205 ions producedby CID (15 eV) of m/z 268 in the spectra ofpentose isomers

Compound m /z 142 : 205 s %s

Ribofuranose I(f ) 104.1a 6.7 6.4 (n ¼3)

124.7a 8.8 7.1 (n ¼4)

Ribopyranose I(p) 46.3a 1.57 3.4 (n ¼6)

Arabinopyranose II(p) 8.8 0.23 2.6 (n ¼4)

Xylopyranose III(p) 3.3a 0.11 3.4 (n ¼5)

3.2a 0.14 4.3 (n ¼4)

a Two different anomers.

The results obtained from a series of repeat measure-ments, all carried out under identical conditions of ionenergy and collision gas pressure, are summarized inTables 1 and 2. High reproducibility of low-energy CIDwas observed (except when the collision gas pressurewas changed and readjusted) and was reÑected inremarkably low relative standard deviations. Relativestandard deviations well below 10% appeared to be thenorm. Thus, even the distinction of the I( f )/I(p) pair(m/z 142 : 205 ratios di†ering by far more than 50% overmost of the energy range) can be reliably accomplishedas long as one reference isomer is available. For thedata given in Tables 1 and 2, repeat measurements at 15eV collision energy were performed independently fromthe single measurements underlying Figs 1È3 andcarried out over the whole energy range. When therepeat measurements are considered, it is clear that low-energy CID on ions, if performed under well con-B1trolled conditions, can be highly e†ective indi†erentiating pentose and deoxyhexose isomers at agood sensitivity level. This appears not to be the casefor the (a priori less sensitive) high-energy CID alterna-tive reported earlier.3

DISCUSSION

For the aldopentoses IÈIII the main features evidentfrom Figs 1 to 3 are the close similarity between thethree furanoses and ribopyranose when compared witheach other, and the increasing dissimilarities bothwithin and between the f/p pairs of arabinose andxylose. This situation prevails throughout the entireenergy range and is highly suggestive of “compositeÏ B1ions containing, for instance, two ion structures in pro-portions that vary from “very similarÏ (p/f pair of I)

Table 2. Peak intensity ratio of m/z 156 : 219 ions producedby CID (15 eV) of m/z 282 in the spectra of 6-deoxyhexose isomers

Compound m /z 156 : 219 s %s

Deoxyglucofuranose IV(f ) 36.7a 0.95 2.6 (n ¼3)

32.6a 1.2 3.7 (n ¼3)

Deoxyglucopyranose IV(p) 7.2a 0.24 3.3 (n ¼4)

7.5a 0.055 0.7 (n ¼4)

a-L-Rhamnose V(p) 7.8 0.36 4.6 (n ¼3)

a-L-Fucose a-VI(p) 5.9 0.40 6.8 (n ¼3)

b-L-Fucose b-VI(p) 6.0 0.42 7.0 (n ¼3)

a Two different anomers.

towards increasingly “di†erentÏ (p/f pairs of II and III). Asimple working hypothesis may, e.g., assume that di†er-ing two-component compositions for ions arise fromB1partial ring interconversion occurring at di†erent rates(six- ] Ðve-membered ring faster, Ðve- ] six-memberedring slower) prior to or during formation byB1rearrangements of the original ring structures to theirring-size isomers to extents di†ering substantially forI(p), II(p) and III(p), but much less so for I( f ), II( f ) andIII( f ). The Ðnding that all three furanosides yield verysimilar (i.e. almost identical) spectra may thereforesuggest that their ions have largely retained the orig-B1inal Ðve-membered ring structures of their molecularprecursors, provided that the di†erent diastereoisomersof these ions fragment in a very similar fashion. Conse-quently, for the pyranosides high extents of ring con-traction to the corresponding Ðve-membered ring B1structures and stereochemically non-distinct fragmenta-tion would have to be assumed. Based on these prem-ises, namely that low-energy CID mainly “seesÏ ring sizerather than stereochemistry, it would follow, in particu-lar, that ions in I(p) are almost completely convertedB1into a corresponding Ðve-membered ring structure,whereas in I( f ) the original and probably more stableÐve-membered ring structure remains largely intact.Less efficient although still high six-] Ðve-memberedring conversion, i.e. substantial conservation of the orig-inal ring structures, would follow for the arabino-andxylopyranosides II(p) and III(p), respectively.

If such a mechanistic model basically applies as anapproximation, the fragment ions at m/z 142 and 98

](B1] ] [B1[ 2CD3COOH]` [B1 [ 2CD3COOHin the spectra of I( f ), II( f ), III( f ) and[ CD2CO]`)

also I(p) would largely represent intact Ðve-memberedring structures. In contrast, the ions at m/z 205,B1161 and 97 (B1] [B1[ CD3COOH]`] [B1 [

andCD3COOH[ CD2CO]` [B1[ (CD3CO)2O]`),which gain importance in the pyranosides on goingfrom I(p) via II(p) to III(p), would represent the varyingproportions of six-membered ring structures either con-served in the pyranosides or formed less efficiently byÐve- ] six-membered ring expansion in the furanosides.The fact that m/z 142 remains the base peak in allsix pentoside spectra would consistently indicate thatÐve-membered ring ion structures are the mainB1constituents of fragment ion current irrespectiveof the original eductÏs ring size and conÐguration.Scheme 4 rationalizes the hypothesis for such indepen-dent decomposition routes of concomitantly formedsix-and Ðve-membered ring precursor ion structures.The near constancy of the contributions of Ðve-and six-membered ring educts to fragment ion current over thewhole energy range (Fig. 4) would rule out major ringinterconversion at a later stage as a consequence ofCID. This same rationale also appears applicable to thelow-energy CID behavior of the ions produced fromB1the deoxyhexoses IVÈVI (Figs 5 and 6). This is in agree-ment with expectations as no great e†ects are to beexpected from substitution of for H in the “remoteÏCH36-position.

It is interesting that for the reported metastable iondecomposition3 of ions derived from aldopentosesB1(m/z 259 for unlabeled acetates) the corresponding

ratios (m/z[B1[ 2CH3COOH] : [B1[ CH3COOH]



Scheme 4.

139 : 199) are also in agreement with these assignmentsto two independent fragmentation channels. In all threereported Ðve-/six-membered ring pairs (ribose, arabi-nose and xylose) the pyranosides show single

loss far more efficiently than double loss,CH3COOHwhereas for the furanosides the opposite is true.

CONCLUSION

Prior to CID, the original cyclic structures of the pre-cursors IÈVIII have lost, to varying extents, ring-sizeintegrity. Especially the pyranosidic six-membered ringstructures appear to have undergone ring contraction toÐve-membered ring isomers with considerable ease (e.g.VI ] V and particularly II ] I).

Despite the apparent loss of structural integrity of theions, interconversions proceed largely isomer-B1speciÐcally and with high reproducibility.

On this basis, structural assignments for neutral pre-cursors can be made with the necessary reliability in anumber of cases. Even closely related isomers can bedistinguished by selected abundance ratios of speciÐcfragment ions. Thus, judicious consideration of seem-ingly minor di†erences in the primary CID spectra maybe utilized to accentuate pronounced di†erences in thebehavior of these isomeric species.

Acknowledgements

We gratefully acknowledge the skilful technical assistance of GiselaRollinson.

REFERENCES

1. K. Biemann, D. C. DeJongh and H. K. Schnoes, J. Am. Chem.Soc. 85, 1763 (1963).

2. T. Radford and D. DeJongh, in Biochemical Applications ofMass Spectrometry , edited by G. R. Waller, p. 314. Wiley,New York (1972).

3. L. Blok-Tip, W. Heerma and J. Haverkamp, Org. MassSpectrom. 28, 139 (1993).

4. A. M. Hogg and T. L. Nagabushan, Tetrahedron Lett. 47, 4827(1972).

5. H. Egge and J. Peter-Katalinic, Mass Spectrom. Rev. 6, 331(1987).

6. A. Dell, Adv. Carbohydr. Chem. Biochem. 45, 19 (1987).

7. W. J. Richter, W. Blum, U. P. Schlunegger and M. Senn, inTandem Mass Spectrometry , edited by F. W. McLafferty, p.417. Wiley, New York (1983).

8. R. Guevremont and J. L. C. Wright, Rapid. Commun. MassSpectrom. 1, 12 (1987).

9. D. R. Mu� ller, B, Domon and W. J. Richter, Spectrosc. Int. J. 7,11 (1989).

10. W. J. Richter, D. R. Mu� ller and B. Domon, Methods Enzymol. ,607 (1990).

11. L. Blok-Tip, W. Heerma and J. Haverkamp, Rapid Commun.Mass Spectrom. , 7, 688 (1993).


low-energy collision-induced dissociation of b1-type sugar ions formed from peracetylated methyl...

Documents