Amide bond cleavage in deprotonated tripeptides: a newly discovered pathway to ″b2 ions

RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 2003; 17: 869–875

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rcm.988

Amide bond cleavage in deprotonated tripeptides:

a newly discovered pathway to 00b2 ions

Alex G. Harrison1*, K. W. Michael Siu2 and Houssain El Aribi21Department of Chemistry, University of Toronto, Toronto, Canada2Department of Chemistry & Centre for Research in Mass Spectrometry, York University, Toronto, Canada

Received 28 November 2002; Revised 10 February 2003; Accepted 11 February 2003

The fragmentation reactions of the [M–H]� ions of the tripeptides H-Gly-Leu-Sar-OH, H-Leu-Gly-

Pro-OH and H-Gly-Leu-Gly-OH have been investigated in detail using energy-resolved mass spec-

trometry, isotopic labelling and MS3 experiments. It is shown that the major route to the 00b2 ions

involves loss of a neutral amine from the a3 ([M–H–CO2]�) ion rather than being formed directly

by fragmentation of the [M–H]� ion. When there is no C-terminal amidic hydrogen (Sar, Pro), loss

of a neutral amine is the dominant primary fragmentation reaction of the a3 ion. However, when

there is a C-terminal amidic hydrogen (Gly), elimination of the N-terminal amino acid residue is

the major fragmentation reaction of the a3 ion and formation of the 00b2 ion is greatly reduced in

importance. It is proposed that the 00b2 ions are deprotonated oxazolones. Copyright # 2003

John Wiley & Sons, Ltd.

A common fragmentation reaction of collisionally activated

protonated peptides involves cleavage of an amide bond.1–4

Extensive studies have shown that when the charge resides

on the C-terminal fragment a protonated amino acid (y100)

or protonated peptide (yn00) is formed5,6 but that, when the

charge remains on the N-terminal fragment, the b ions

formed, rather than having the expected acylium ion struc-

ture,1–4 have, in many cases, cyclized to form a protonated

oxazolone.7–11 A recent study by O’Hair and co-workers12

has suggested that in a number of cases alternative cyclic

structures may be more stable than the oxazolone structure

and thus may be preferentially formed. Rodriquez et al.13

have also calculated the relative energies of a number of

structural alternatives to the oxazolone structure.

Nominal amide bond cleavage also occurs for deproto-

nated peptides, as illustrated in Scheme 1, where the yn ions

represent deprotonated amino acids or peptides and the 00bn

ions bear two fewer hydrogen atoms than the corresponding

bn ions formed from protonated peptides. This backbone

cleavage reaction was first observed by Heerma and co-

workers14,15 and elaborated on by Bowie and co-workers,16,17

who proposed the mechanism outlined in Scheme 2

designating the fragmentation leading to the deprotonated

amino acid as a-cleavage and that leading to the charged N-

terminal fragment as b-cleavage. An alternative mechanism

has been put forward by Harrison,18 as outlined in Scheme 3,

where it is proposed that the 00b2 ion is a deprotonated

oxazolone. Recent MS2 and MS3 studies in this laboratory

have shown that an alternative pathway to the b-cleavage

product or 00b2 ion involves elimination of a neutral amine

from the a3 ([M–H–CO2]�) ion. The results of these studies

are presented below.

EXPERIMENTAL

Collision-induced dissociation (CID) studies were carried

out using an electrospray/quadrupole mass spectrometer

(VG Platform, Micromass, Manchester, UK) with CID in the

interface region between the atmospheric pressure source

and the quadrupole mass analyzer. It is well known19,20

that CID can be achieved in this region, so-called cone-vol-

tage CID, and it has been clearly established21–23 that the

average energy imparted to the decomposing ions increases

as the field in the interface region increases. Indeed, recent

work24–27 has shown that, by varying this field in steps,

energy-resolved mass spectra28–30 comparable to those

obtained in variable low-energy CID in quadrupole cells

can be obtained. The results of these cone-voltage CID experi-

ments are presented in the following either as CID mass spec-

tra at a set cone voltage or as breakdown graphs expressing

the percent of total ion abundance as a function of the cone

voltage, a measure of the field in the interface region. MS3

experiments were carried out using an electrospray/triple

quadrupole mass spectrometer (Sciex API 3000, Concord,

Canada). CID in the interface region produced the fragment

ion of interest which was mass-selected by the first quadru-

pole mass analyzer and underwent collisional activation in

the quadrupole collision cell with the fragmentation products

being analyzed by the final quadrupole mass analyzer.

With the single quadrupole instrument, ionization was by

electrospray with the peptide sample, at micromolar con-

centration in 1:1 CH3CN/1% aqueous NH3, being introduced

into the source at a flow rate of 30mL min�1. The electrospray

needle was held at �2.5 to 3.0 kV. Nitrogen, produced by a

Copyright # 2003 John Wiley & Sons, Ltd.

*Correspondence to: A. G. Harrison, Department of Chemistry,University of Toronto, 80 St George Street, Toronto, OntarioM5S 3H6, Canada.E-mail: [email protected]/grant sponsor: Natural Sciences and EngineeringResearch Council, Canada; MDS SCIEX.

Whatman model 75–72 N2 generator (Whatman Inc., Haver-

hill, MA, USA), was used as both nebulizing and drying gas.

By using 1:1 CD3CN/1% ND3 in D2O as the electrospray

solvent, the labile hydrogens were exchanged for deuterium

and the [M–D]� ion formed in the ionization process. Under

these conditions no evidence was seen for back-exchange in

the interface region, although significant back-exchange was

observed when dry air was used as nebulizing and drying

gas. With the triple quadrupole instrument electrospray

ionization was used with a sample flow rate of 3mL min�1 and

dry air was used as nebulizing and drying gas. Nitrogen was

used as the collision gas under multiple collision conditions.

All peptide samples were obtained from BACHEM

Biosciences (King of Prussia, PA, USA). CD3CN (99.8 atom

% D) and D2O (99.9 atom % D) were obtained from Cam-

bridge Isotope Laboratories (Andover, MA, USA) while

ND3OD (26% in D2O,>99 atom % D) was obtained from DCN

Isotopes (Pointe Claire, Quebec, Canada).

RESULTS AND DISCUSSION

Figure 1 shows the breakdown graph obtained for the [M–

H]� ion of the tripeptide H-Gly-Leu-Sar-OH. Three products

are observed at low collision energies, the y1 ion (m/z 88), the

a3 ([M–H–CO2]�) ion (m/z 214) and the 00b2 ion (m/z 169). At

higher collision energies a number of further fragmentation

products, including the c1 ion (m/z 73), are observed. Table 1

presents the CID mass spectra obtained for the [M–H]�, a3

Scheme 1.

Scheme 2.

Copyright # 2003 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2003; 17: 869–875

870 A. G. Harrison, K. W. Michael Siu and H. E. Aribi

and 00b2 ions on the triple quadrupole instrument. Clearly, the

ion signals at m/z 141, 139 and 83 in Fig. 1 originate by further

fragmentation of the 00b2 (m/z 169) ion. While we cannot pre-

clude the possibility that the 00b2 ion originates, in part,

directly by fragmentation of the [M–H]� ion, the MS3 data

of Table 1 show that a major route to the 00b2 ion is by fragmen-

tation of the a3 ion (m/z 214). This route to the 00b2 ion has not

been previously identified. The fragmentation reaction

a3!00b2 involves the loss of a neutral (or neutrals) of 45 Da;

the most likely candidate is dimethylamine. Further informa-

tion comes from CID of the [M–D]� ion of the tripeptide in

which the labile hydrogens have been exchanged for deuter-

ium. Figure 2 compares the cone-voltage CID mass spectrum

of the [M–H]� ion of the unlabelled peptide with that of the

[M–D]� ion of Gly-Leu-Sar-d4. Clearly, the a3 ion cleanly

incorporates three labile hydrogens as expected while the00b2 ion incorporates almost entirely two labile hydrogens,

indicating that the dimethylamine lost in the a3!00b2 frag-

mentation incorporates one labile hydrogen. This leads us

to propose the reaction sequence outlined in Scheme 4. In

Scheme 3.

Table 1. CID of selected ions from H-Gly-Leu-Sar-OH

(20 eV, multiple collisions)

m/z (ion)

Relative intensity (% of base peak)

[M–H]� (258) a3 (214) 00b2 (169)

214(a3) 56.1169(00b2) 100 100152 1.6 4.8141 9.4 55.6 52.1139 4.3 33.7 100108 2.1 11.588(y1) 52.483 1.1 13.4 71.773(c1) 1.1 2.1

Figure 1. Breakdown graph for deprotonated H-Gly-Leu-

Sar-OH.

Amide bond cleavage in deprotonated tripeptides 871


support of the reaction sequence shown in Scheme 4, we note

that CID of the a3 ion at 40 eV collision energy led to signifi-

cant formation of (CH3)2N� (m/z 44) in addition to the pro-

ducts listed in Table 1. In effect, at high collision energy, the

ion/neutral complex of Scheme 4 separates before proton

transfer can occur.

As Fig. 2 shows, the y1 ion in this case incorporates largely

one labile hydrogen. When the C-terminal amide nitrogen

does not bear a hydrogen (i.e., when the C-terminal residue is

Sar or Pro), the pathway in Scheme 2 would lead to

incorporation of no labile hydrogens in y1 while the pathway

in Scheme 3 would lead to incorporation of one labile

Figure 2. Comparison of CID mass spectrum of deprotonated Gly-Leu-Sar with that of

[M–D]� ion of Gly-Leu-Sar-d4.

Scheme 4.



hydrogen in the y1 ion. The latter prediction is in agreement

with the experimental results. The MS3 data of Table 1

indicate that the c1 ion (m/z 73) also originates by fragmenta-

tion of the a3 ion. A plausible mechanism is shown in

Scheme 5. The breakdown graph of Fig. 1 indicates that this is

a high-energy process.

Similar results in all respects were obtained in the study of

deprotonated H-Leu-Gly-Pro-OH. The breakdown graph

(Fig. 3) obtained by cone-voltage CID shows again the y1, a3

and 00b2 ions as low-energy fragmentation products.

Exchange of the labile hydrogens for deuterium showed that

the y1 ion incorporated mainly one labile hydrogen, the 00b2

ion mainly two labile hydrogens, and the a3 ion three labile

hydrogens. The MS2 and MS3 results obtained on the triple

quadrupole instrument are summarized in Table 2 from

which it can be seen that the major fragmentation reaction of

the a3 ion is loss of 71 Da (pyrrolidine) to give the 00b2 ion atm/z

169. A reaction sequence entirely analogous to that of Scheme

4 can be written. Interestingly, the CID mass spectrum of the00b2 ion derived from H-Leu-Gly-Pro-OH (Table 2) differs to

some extent from the CID mass spectrum (Table 1) of the 00b2

ion derived from H-Gly-Leu-Sar-OH. This is consistent with

a different substitution pattern for the deprotonated oxazo-

lones postulated in Scheme 4. The possible structures and

related energetics of putative 00b2 ions are under study both

experimentally and computationally and the results will be

reported in detail in a further communication as will a more

detailed study of the fragmentation reactions of 00b2 ions.

An alternative pathway for the a3!00b2 fragmentation

involves formation of a deprotonated diketopiperazine, as

outlined in Scheme 6. This pathway is very similar to the

pathway proposed by Bowie and co-workers31 for the

elimination of CH3OH from deprotonated dipeptide methyl

esters. Our calculations (to be reported) show that in the

neutral diketopiperazine the amidic hydrogens are the most

acidic. Thus, if the pathway in Scheme 6 was operative, one

would expect to see significant abstraction of an amidic

Table 2. CID of selected ions from H-Leu-Gly-Pro-OH


m/z (ion)


[M–H]� (284) a3 (240) 00b2 (169)

267 1.7240(a3) 38.3169(00b2) 33.1 100152 2.1 13.3141 3.2 35.6139 3.7 100129(c1) 3.4 2.1125 7.4114(y1) 10083 6.782 3.2

Figure 3. Breakdown graph for deprotonated H-Leu-Gly-

Pro-OH.

Scheme 5.

Scheme 6.



(labile) hydrogen in the step leading to the final products.

This would lead to incorporation of only one labile hydrogen

in the 00b2 ion, which is contrary to the experimental

observations. Consequently, we conclude that the pathway

illustrated in Scheme 6 does not play a significant role in the

present systems.

In our earlier study18 of deprotonated peptides containing

H or alkyl a-groups, it was found that a major fragmentation

channel of the a3 ions involved elimination of the N-terminal

amino acid residue. This is illustrated by the breakdown

graph obtained by cone-voltage CID for the [M–H]� ion of H-

Gly-Leu-Gly-OH (Fig. 4) and by the MS2 and MS3 results

obtained on the triple quadrupole instrument (Table 3). The

m/z 143 ion corresponds to elimination of the glycine residue

from the a3 ion; the MS3 data of Table 3 show that this is the

major fragmentation channel for the a3 ion. However, even in

this case, the a3 ion does show elimination of CH3NH2 to form

the 00b2 ion (m/z 169). Exchange of the labile hydrogens for

deuterium showed that the a3 ion retained four labile

hydrogens while the 00b2 ion retained two labile hydrogens,

consistent with Scheme 4. The y1 ion showed largely

incorporation of two labile hydrogens, consistent with the

predictions of Scheme 3 when there is a hydrogen on the C-

terminal amide nitrogen. There was some incorporation of

only one labile hydrogen in the y1 ion indicating that the

pathway of Scheme 2 may be involved to some extent. These

observations are in agreement with earlier results.18 Elimina-

tion of the N-terminal residue does not occur for the a3 ions

derived from H-Gly-Leu-Sar-OH or H-Leu-Gly-Pro-OH. In

the same vein, the a3 ions derived from H-Gly-Gly-Sar-OH

and H-Gly-Gly-Pro-OH did not show elimination of the

glycine residue from the a3 ions but did show a much more

intense 00b2 ion than observed for the tripeptides studied

earlier.18 Clearly, the presence of a hydrogen on the C-

terminal amide nitrogen plays a role in the elimination of the

N-terminal residue from the a3 ion.

CONCLUSIONS

Using isotopic labelling and MS2 and MS3 experiments the

present work has shown that the major route to 00b2 ions

involves elimination of a neutral amine molecule from the

a3 ([M–H–CO2]�) ion. While we cannot rule out the possibi-

lity that 00b2 ions also arise directly by amide bond cleavage in

the [M–H]� ion, as has been proposed previously,16–18 this

appears to be, at best, a minor route to 00b2 ions. When there

is no C-terminal amidic hydrogen (as in Sar and Pro) the

dominant fragmentation mode of the a3 ion involves loss of

a neutral amine to form the 00b2 ion. However, when there is

a C-terminal amidic hydrogen (as in Gly) the major fragmen-

tation mode of the a3 ion involves loss of the N-terminal ami-

no acid residue18 and formation of the 00b2 is much reduced in

importance. The manner in which the presence or absence of

the amidic hydrogen determines the fragmentation pathway

is not clearly understood and further studies are underway.

AcknowledgementsWe thank the Natural Sciences and Engineering Research

Council (Canada) and MDS SCIEX for continual financial

support. The donation of the Platform mass spectrometer to

the University of Toronto by Micromass Canada is gratefully

acknowledged.

REFERENCES

1. Roepstorff P, Fohlman J. Biomed. Mass Spectrom. 1984; 11:601.

2. Biemann K. Biomed. Environ. Mass Spectrom. 1988; 16: 99.3. Biemann K. Methods Enzymol. 1990; 193: 455.4. Papayannopoulos IA. Mass Spectrom. Rev. 1995; 14: 49.5. Mueller DR, Eckersley M, Richter W. Org. Mass Spectrom.

1988; 23: 217.6. Cordero MM, Houser JJ, Wesdemiotis C. Anal. Chem. 1993;

65: 1594.7. Yalcin T, Khouw C, Csizmadia IG, Peterson MR, Harrison

AG. J. Am. Soc. Mass Spectrom. 1995; 6: 1165.8. Yalcin T, Csizmadia IG, Peterson MR, Harrison AG. J. Am.

Soc. Mass Spectrom. 1996; 7: 293.9. Nold MJ, Wesdemiotis C, Yalcin T, Harrison AG. Int. J. Mass

Spectrom. Ion Processes 1997; 164: 137.

Table 3. CID of selected ions from H-Gly-Leu-Gly-OH


m/z (ion)


[M–H]� (244) a3 (200) 00b2 (169)

200(a3) 100182 4.7 3.3169(00b2) 64.7 15.1152 3.5 8.9 5.3143 48.8 100141 2.9 7.2 64.7139 1.0 3.3 100125 3.5 13.1108 11.883 57.874(y1) 40.673(c1) 13.9

Figure 4. Breakdown graph for deprotonated H-Gly-Leu-

Gly-OH.



10. Paizs B, Lendvay G, Vekey K, Suhai S. Rapid Commun. MassSpectrom. 1997; 13: 525.

11. Harrison AG, Csizmadia IG, Tang T-H. J. Am. Soc. MassSpectrom. 2000; 11: 427.

12. Farrugia JM, O’Hair RAJ, Reid GE. Int. J. Mass Spectrom.2001; 210/211: 71.

13. Rodriquez CF, Shoeib T, Chu IK, Siu KWM, Hopkinson AC.J. Phys. Chem. A 2000; 104: 5335.

14. Kulik W, Heerma W. Biomed. Environ. Mass Spectrom. 1989;18: 910.

15. van Setten D, Kulik W, Heerma W. Biomed. Environ. MassSpectrom. 1990; 19: 475.

16. Eckersley M, Bowie JH, Hayes RN. Org. Mass Spectrom.1989; 24: 597.

17. Waugh RJ, Bowie JH. Rapid Commun. Mass Spectrom. 1994;8: 169.

18. Harrison AG. J. Am. Soc. Mass Spectrom. 2001; 12: 1.19. Loo JA, Udseth HR, Smith RD. Rapid Commun. Mass Spec-

trom. 1988; 2: 207.20. Bruins AP. In Electrospray Mass Spectrometry: Fundamentals,

Instrumentation and Applications, Cole RB (ed.). John Wiley:New York, 1997; chap. 3.

21. Voyksner TD, Park T. Rapid Commun. Mass Spectrom. 1991;5: 263.

22. Collette C, DePauw E. Rapid Commun. Mass Spectrom. 1998;12: 165.

23. Collette C, Drahos L, DePauw E, Vekey K. Rapid Commun.Mass Spectrom. 1998; 12: 1673.

24. Harrison AG. Rapid Commun. Mass Spectrom. 1999; 13:1663.

25. van Dongen WD, van Wijk JIT, Green BN, Heerma W,Haverkamp J. Rapid Commun. Mass Spectrom. 1999; 13: 1712.

26. Harrison AG. J. Mass Spectrom. 1999; 34: 1253.27. Makowiecki J, Tolonen A, Uusitalo J, Jalonen J. Rapid Com-

mun. Mass Spectrom. 2001; 15: 1506.28. McLuckey SA, Glish GL, Cooks RG. Int. J. Mass Spectrom.

Ion Phys. 1981; 39: 219.29. Fetterolf DD, Yost RA. Int. J. Mass Spectrom. Ion Phys. 1982;

44: 37.30. McLuckey SA, Cooks RG. In Tandem Mass Spectrometry,

McLafferty FW (ed.). John Wiley: New York, 1983; 203.31. Bradford AM, Waugh RJ, Bowie JH, Vollmer DL, Gross ML.

Int. J. Mass Spectrom. Ion Processes 1994; 136: 143.



Documents

Amide bond cleavage in deprotonated tripeptides: a newly discovered pathway to ″b2 ions