Decomposition of protonated threonine, its stereoisomers, and its homologues in the gas phase: Evidence for internal backside displacement

Article abstract of DOI:10.1021/ol049700p

Protonated threonine and its allo diastereomer exhibit different proportions of collisionally activated dissociation (CAD) product ions. N-Methylation attenuates these differences. Water loss from protonated allo-threonine gives protonated trans-3-methyla

Full text of DOI:10.1021/ol049700p

ORGANIC
LETTERS
2004
Vol. 6, No. 10
1561-1564
Decomposition of Protonated Threonine,
Its Stereoisomers, and Its Homologues
in the Gas Phase: Evidence for Internal
Backside Displacement
Scott V. Serafin,† Kangling Zhang,† Luigi Aurelio,‡ Andrew B. Hughes,‡ and
,†
Thomas Hellman Morton*
Department of Chemistry, UniVersity of California-RiVerside,
RiVerside, California 92521-0403, and Department of Chemistry,
La Trobe UniVersity, Bundoora, Victoria 3086, Australia
morton@citrus.ucr.edu
Received February 19, 2004
ABSTRACT
Protonated threonine and its allo diastereomer exhibit different proportions of collisionally activated dissociation (CAD) product ions. N-Methylation
attenuates these differences. Water loss from protonated allo-threonine gives protonated trans-3-methylaziridinecarboxylic acid via an internal
SN2 pathway, rather than protonated vinylglycine.
Protonated amino acids readily expel water in the gas phase.
In most cases, loss of water from the carboxylic group is
accompanied by CO loss,1 producing an iminium ion,2 as
Scheme 1 depicts. In a few instances (notably threonine3),
loss of water from the side chain predominates.
can be differentiated by mass spectrometry. This letter reports
the collisionally activated dissociation (CAD) spectra of the
conjugate acid ions from threonine, its allo diastereomer,
and their N-methyl homologues.
N-Methylthreonine has also been found in peptides,5 but
the N-methyl allo diastereomer has yet to be reported. As
molecules incorporating this latter isomer may occur in
nature, we describe its preparation (starting from the amino
acid, using a method previously reported6) and its CAD. The
threo and allo isomers exhibit different fragmentation pat-
Scheme 1. Dissociation of a Protonated Amino Acid
(1) Kulik, W.; Heerma, W. Biomed. EnViron. Mass Spectrom. 1988, 15,
419-427.
(2) Dookeran, N. N.; Yalcin, Y.; Harrison, A. G. J. Mass Spectrom. 1996,
31, 500-508.
Threonine has two asymmetric carbons. The diastereomer
not involved in protein biosynthesis is found in a variety of
naturally occurring cyclic peptides.4 This fact prompts the
question as to whether the two diastereomeric amino acids
(3) Van Dongen, W. D.; de Koster, C. G.; Heerma, W.; Haverkamp, J.
Rapid Commun. Mass Spectrom. 1995, 9, 845-850.
(4) (a) Bender, C. L.; Alarco´n-Chaidez, F.; Gross, D. C. Microbiol. Mol.
Biol. ReV. 1999, 63, 266-292. (b) Matsunaga, S.; Fusetani, R. Curr. Org.
Chem. 2003, 7, 945-966.
† University of California-Riverside.
(5) Aurelio, L.; Box, J. S.; Brownlee, R. T. C.; Hughes, A. B.; Sleebs,
M. M. J. Org. Chem. 2003, 68, 2652-2667.
‡ La Trobe University.
10.1021/ol049700p CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/09/2004

Table 1. Proportions of Neutrals Lost in MS/MS (12 eV Collision Energy) for Selected Parent Ions (Relative to Loss of H2O + CO)
fragment ions from loss of:
neutral precursor
threonine
m/z of parent
H2O
NH3
2H2O
CO
CO + 2H2O
0.12 ( 0.01
120
102
120
102
134
116
134
116
102
102
102
102
120
102
0.58 ( 0.10
0.85 ( 0.05
3.19 ( 0.14
2.08 ( 0.22
1.24 ( 0.12
0.49 ( 0.15
3.05 ( 0.17
0.54 ( 0.14
0.20 ( 0.03
0.05 ( 0.04
0.32 ( 0.07
0.09 ( 0.03
0.48 ( 0.09
allo-threonine
1.34 ( 0.12
0.56 ( 0.06
0.81 ( 0.06
0.28 ( 0.03
0.21 ( 0.04
0.35 ( 0.06
N-methylthreonine
N-methyl-allo-threonine
vinylglycine
0.33 ( 0.05
1-aminocyclo-propane-1-COOH
aziridine ester 4
azetidine-2-COOH
homoserine
0.37 ( 0.07
2.12 ( 0.10
0.02 ( 0.00
0.49 ( 0.03
1.7 ( 0.5
0.35 ( 0.01
3.3 ( 0.8
0.04 ( 0.00
0.11 ( 0.00
terns, significantly expanding the list of acyclic stereoisomers
that mass spectrometry can distinguish.7 Three different
hypotheses have been explored to account for this difference.
The first hypothesis looks to the two five-membered cyclic
transition states drawn in Scheme 2. If they corresponded to
The experiments described below do not confirm that
prediction: the allo isomer displays the greater proportion
of single water loss. O’Hair and Reid have reported deuter-
ium labeling studies that suggest that the rate-limiting step
comes after proton transfer to the side chain oxygen.8 Scheme
4 illustrates two plausible pathways.
Scheme 2. Transition States for H2O Loss from ThreonineH+
Scheme 4. Alternative Mechanisms for Side-Chain Water
Loss
One pathway, originally proposed by O’Hair and Reid,
proceeds via internal SN2 displacement to give protonated
aziridinecarboxylic acid 1. This mechanism implies that
diastereomeric threonines ought to produce isomeric ions:
threo should yield cis-1, and allo should yield trans-1. An
alternative pathway would involve internal E2 elimination
via five-membered cyclic transition states to yield protonated
vinylglycine 2 or protonated R-aminocrotonic acid 3, which
MP2 calculations9 predict to be 1.9 kcal mol-1 and 6.2 kcal
mol-1 ((Z)-isomer), respectively, more stable than trans-1.
The dissociation patterns of the protonated amino acids
were studied using an orthogonal quadrupole/time-of-flight
mass spectrometer with electrospray sample introduction.
Table 1 summarizes the abundance of other neutral losses
relative to the dissociation in Scheme 1 (which all R-amino
acids have in common). The threonine diastereomers show
quantitative differences in fragment ion abundances. These
patterns vary with the collision energy. The ratios of the two
most intense fragment ions are plotted in Figure 1 as a
the rate-limiting steps for the two most prominent pathways,
then competition between them would predict measurable
differences between diastereomers. Assuming that the acidic
proton resides on the nitrogen to begin with, varying the
relative stereochemistry of the two asymmetric centers should
affect the likelihood of loss of side chain OH. As Scheme 3
depicts, the transition state for the allo puts the two substit-
uents cis to one another in the five-membered ring. For steric
reasons, then, the aforementioned hypothesis predicts that
loss of a single water ought to be more favorable for the
threo than for the allo.
Scheme 3. H2O Loss Transition States for allo-ThreonineH+
(6) (a) Aurelio, L.; Brownlee, R. T. C.; Hughes, A. B.; Sleebs, B. E.
Aust. J. Chem. 2000, 53, 425-433. (b) Aurelio, L.; Brownlee, R. T. C.;
Hughes, A. B. Org. Lett. 2003, 4, 3767-3769.
(7) Applications of Mass Spectrometry to Organic Stereochemistry;
Splitter, J. S., Turecˇek, F., Ed.; VCH: New York, 1994.
1562
Org. Lett., Vol. 6, No. 10, 2004

Scheme 5. Rotamers of Protonated Threonine and Their
Dissociation Products
Scheme 6. Rotamers of Protonated allo-Threonine and
Dissociation Products
Figure 1. Ratios of H2O loss to H2O + CO loss from D- and
L-threonine (open symbols) and their allo-diastereomers (solid
symbols) as a function of collision energy of m/z 120 parent ions.
but instead reflects the rotameric distribution of the proto-
nated amino acids prior to dissociation. The effect of meth-
ylation tests this second hypothesis. Schemes 5 and 6 portray
these equilibria, including ab initio enthalpy differences.8 The
most favored structures have the NH3+ hydrogen-bonded to
both the side chain oxygen and the carbonyl oxygen. Only
side-chain water loss can take place from this geometry. Less
+
favored rotamers have the NH3 hydrogen-bonded to the
side-chain oxygen and to the carboxylic OH group. Loss of
water from either position can occur from this geometry.
N-Methylation increases the enthalpy difference (∆∆H)
between allo and threo isomers. The calculated ∆∆H for
the amino acids is 1.37 kcal mol-1 (the difference between
∆EMP2 + ∆ZPE in Scheme 6 and ∆EMP2 + ∆ZPE in Scheme
5). The corresponding difference for the N-methyl amino
acids is ∆∆H ) 1.66 kcal mol-1. If the rotameric equilibrium
by itself determined the dissociation pattern, one would have
expected N-methylation to increase the proportion of allo
side-chain water loss relative to threo. Comparison of Figures
1 and 2 demonstrates that the exact opposite result is
obtained.
The observed consequences of N-methylation do agree
with expectation, if the rate-limiting step for side-chain water
loss corresponds to the pathways depicted in Scheme 4. The
question then arises whether that reaction operates via the
internal SN2 pathway or via the internal E2. O’Hair and Reid8
have argued in favor of the former, on the basis of their deu-
terium labeling experiments. MS/MS of the ions from skim-
mer-induced side-chain water loss substantiates their inter-
pretation.
Figure 2. Ratios of H2O loss to H2O + CO loss for D- and L-N-
methylthreonine (open symbols) and their allo-diastereomers (solid
symbols) as a function of energy of m/z 134 parent ions.
function of collision energy for all four stereoisomers. Figure
2 displays the same type of plot for all four stereoisomers
of N-methylthreonine. Given the pronounced increase of
single-water loss in the allo relative to threo, the first question
concerns whether that dissociation represents elimination
only from the side chain. This was assessed by labeling the
allo with both carboxylic oxygens replaced by 18O (in the
same manner as doubly 18O-labeled threo has been pre-
pared3). The CAD spectrum of labeled allo shows that loss
of a single water comes exclusively from the side chain,
while loss of H2O plus CO comes exclusively from the
carboxyl group (consistent with Scheme 1).
As will be discussed below, additional experimental data
corroborate the SN2 mechanism shown in Scheme 4. Before
presenting that evidence, though, an alternative hypothesis
should be weighed; namely, that the difference between dia-
stereomers does not result from competing transition states
As Table 1 summarizes, CAD of the ions from water loss
(m/z 102 from the threonines; m/z 116 from the N-methyl
derivatives) shows significant differences between the ions
(8) (a) O’Hair, R. A. J.; Reid, G. E. Rapid Commun. Mass Spectrom.
1998, 12, 999-1002. (b) Reid, G. E.; Simpson, R. J.; O’Hair, R. A. J. Am.
Soc. Mass Spectrom. 2000, 11, 1047-1060.
Org. Lett., Vol. 6, No. 10, 2004
1563

originating from threo and allo isomers. Such an outcome
is consistent with the SN2 pathway. The stereospecificity of
backside displacement predicts that the threo ought to yield
a cis aziridine, while the allo should give a trans aziridine.
of side-chain water from N-methylthreonine can produce the
two protonated aziridines 5a and 5b. The latter ion has all
substituents cis and is therefore likely to be formed much
less often than is 5a. By contrast, both of the protonated
aziridines from N-methyl-allo-threonine, 5c and 5d, have one
cis and two trans substituents on the ring, as does ion 5a.
The difference between protonated threonine and proto-
nated allo-threonine can be ascribed to the steric hindrance
introduced by placing two substituents cis on the developing
ring during displacement of side-chain water. In the N-methyl
homologues, 5a, 5c, and 5d all have nearly the same degree
of steric hindrance. While the torsional strain in the aziridine
rings does not fully develop in the SN2 transition state, it
does tend to diminish the difference between threo and allo
isomers. If ions 5a, 5c, and 5d all form at the same rate,
then (on the basis of a na¨ıve statistical argument) side-chain
water loss in the N-methyl-allo should be twice as abundant
as in the N-methyl-threo. The ratio of the two curves in
Figure 2 ranges from 2.3 (at low collision energies) to 2.9,
not far from this prediction.
The distinctions described above hinge on the stereochem-
istry of backside displacement, thus enlarging the repertoire
of stereospecific dissociations by which diastereomers can
be differentiated using MS/MS. Methods previously reported
from this laboratory11 for discriminating acyclic diastereo-
mers have relied on syn eliminations via four-membered
transition states. Are there other mechanisms by which MS/
MS can tell acyclic stereoisomers apart? Preliminary negative
ion data suggest there might be.
Anions from threonine display the same CAD patterns for
both diastereomers. By contrast, MS/MS of negative ions
from isoleucine do exhibit a statistically significant differ-
ence. The isoleucine and allo-isoleucine M - 1 anions
display different extents of dissociation via CO + H2O loss,
relative to the parent ion (although the M + 1 positive ions
show no differences). No variations are to be found among
the competing fragmentation pathways themselves.
The example of isoleucine illustrates a less robust way to
differentiate between diastereomers than the competition
among pathways described above for the M + 1 positive
ions of the threonines. Current efforts seek methodologies
for dipeptides and other modified amino acids.
To test this stereochemical assignment, the authentic trans
aziridine ion was synthesized from the epoxide of tert-butyl
trans-crotonate using established methods,10 as Scheme 7
Scheme 7. Preparation of tert-Butyl
trans-3-Methylaziridine-2-carboxylate 4,10 Which Expels
Isobutene in the Mass Spectrometer
outlines. Because of the known lability of free aziridinecar-
boxylic acids, tert-butyl ester 4 was prepared. Under elec-
trospray conditions, the aziridine ester exhibits not only the
protonated parent ion (m/z 158) but also an intense, skimmer-
induced m/z 102 peak, which comes from isobutene loss from
the protonated parent via the McLafferty rearrangement.
The CAD fragmentation pattern of the m/z 102 ion from
4 is the same as that of the m/z 102 ion derived from allo-
threonine. By contrast, protonated vinylglycine 2 (an ion that
would have resulted from the E2 pathway), gives a very
different CAD pattern, as Table 1 summarizes. The other
m/z 102 ions listed in Table 1 also give decomposition pat-
terns that differ greatly from those of the protonated
aziridinecarboxylic acids. Table 1 does not include protonated
azetidine-3-carboxylic acid, which primarily loses the ele-
ments of CH2dNH and water to give m/z 55.
As noted above, N-methylation attenuates the difference
between diastereomers. This is consistent with aziridine
formation, as outlined in Scheme 8. N-Alkylation introduces
an additional stereogenic center (the protonated, methylated
nitrogen) into the three-membered ring. SN2 displacement
Scheme 8. Side-Chain Water Loss From Protonated N-Methyl
Amino Acids
Acknowledgment. This work was supported by NSF
Grant CHE0316515.
Supporting Information Available: Synthesis and spec-
tra of N-methylthreonines and of ester 4; tabulated DFT and
ab initio energies. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL049700P
(9) Geometries were optimized at the B3LYP and MP2(FC) levels using
the 6-31G** basis. Computed stabilities include DFT zero-point energies.
(10) Legters, J.; Thijs, L.; Zwanesburg, B. Recl. TraV. Chim. Pays-Bas
1992, 111, 1-15. The authors are grateful to Prof. Francois Mathey for
suggesting tris(2-cyanoethyl)phosphine in place of triphenylphosphine.
(11) (a) Zhang, K.; Bouchonnet, S.; Serafin, S. V.; Morton, T. H. Int. J.
Mass Spectrom. 2003, 227, 175-189. (b) Morizur, J.-P.; Taphanel, M.-H.;
Mayer, P. S.; Morton, T. H. J. Org. Chem. 2000, 65, 381-387. (c) Taphanel,
M.-H.; Morizur, J.-P.; Leblanc, D.; Borchardt, D.; Morton, T. H. Anal.
Chem. 1997, 69, 4191-4196.
1564
Org. Lett., Vol. 6, No. 10, 2004