Determination of regioselectivity in ring opening of tert-butylaziridinones by a combination of 15N labeling and electrospray ionization-ion trap mass spectrometry

Article abstract of DOI:10.1002/jms.242

The ring opening of 1,3-di-tert-butylaziridinone by tert-butylamine and aniline was investigated by using electrospray ionization and collision-induced dissociation in an ion trap mass spectrometer in conjunction with ¹⁵N labeling of the two amine nucleophiles. Using the MSⁿ capabilities of the ion trap instrument, we were able to monitor the retention of the ¹⁵N label through successive fragmentation steps. Both amines exhibited a remarkable degree of selectivity in that they both cleaved exclusively the 1,3-bond (the alkyl-nitrogen bond). This result is in contrast to that obtained previously with methylamine, which cleaved just the opposite bond, namely, the 1,2-bond (the acyl-nitrogen bond). These contrasting results could not have been predicted by previously published guidelines. Copyright

Full text of DOI:10.1002/jms.242

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2002; 37: 31–40
Determination of regioselectivity in ring opening of
tert-butylaziridinones by a combination of 15N labeling
and electrospray ionization-ion trap mass spectrometry
E. R. Talaty,∗ M. J. Van Stipdonk, M. J. Hague, F. A. Provenzano and C. A. Boese
Department of Chemistry, Wichita State University, Wichita, Kansas 67260-0051, USA
Received 12 January 2001; Revised 21 June 2001; Accepted 21 September 2001; Published online 28 November 2001
The ring opening of 1,3-di-tert-butylaziridinone by tert-butylamine and aniline was investigated by using
electrospray ionization and collision-induced dissociation in an ion trap mass spectrometer in conjunction
with 15N labeling of the two amine nucleophiles. Using the MSn capabilities of the ion trap instrument,
we were able to monitor the retention of the 15N label through successive fragmentation steps. Both
amines exhibited a remarkable degree of selectivity in that they both cleaved exclusively the 1,3-bond
(the alkyl–nitrogen bond). This result is in contrast to that obtained previously with methylamine, which
cleaved just the opposite bond, namely, the 1,2-bond (the acyl–nitrogen bond). These contrasting results
could not have been predicted by previously published guidelines. Copyright  2001 John Wiley & Sons,
Ltd.
KEYWORDS: electrospray ionization mass spectrometry; collision-induced dissociation; 15N label; cleavage of aziridinones;
ion trap mass spectrometry
by virtue of coupling to an adjacent amide type of hydro-
gen (resulting from the partial structure—CO—NH—CH3).
However, even this NMR method cannot provide the answer
in the case of amines that lack a hydrogen atom on the ˛-
carbon, such as tert-butylamine or an arylamine. Before one
can understand or predict the mode of ring opening of
aziridinones with certainty, one needs to study a variety of
nucleophiles and develop a reliable method for choosing
between structures 2 and 3.
INTRODUCTION
The synthetic uses of aziridinones (structure 1 in Fig. 1(a)) are
governed by a high degree of selectivity often encountered
in their reactions, especially with nucleophiles. It has
been reported1,2 that ionic aprotic nucleophiles (Zꢀ) cleave
exclusively the acyl–nitrogen bond (1,2-bond), whereas non-
ionic protic nucleophiles (HZ) yield products derived solely
or mainly from cleavage of the alkyl–nitrogen bond (1,3-
bond) (Fig. 1(a)). Our previous work has provided examples
that clearly contradict these simple rules. Thus, magnesium
methoxide and magnesium iodide, both of which are ionic,
cause scission of diametrically opposite bonds, affording
products arising exclusively from cleavage of the 1,2-
bond and 1,3-bond, respectively.3 A similar contradiction of
previous rules is encountered in the reaction of some protic
nucleophiles such as methylamine which cleaves solely the
1,2-bond, just the opposite of that predicted.4
A distinction between structures 2 and 3 is unambiguous
in cases where Z is oxygen, sulfur or halogen based on sol-
ubility in dilute acid. However, when Z is a basic atom such
as nitrogen, such a clear distinction is not always possible.
For example, when the nucleophile is methylamine, 1H NMR
can differentiate between the two structures by observation
of a doublet signal for its methyl group only in structure 2
In the present paper, we demonstrate that a valuable
method for differentiating between structures 2 and 3
when the nucleophile cleaving the aziridinone is tert-
butylamine or aniline is the use of multiple collision-
induced dissociation (CID) sequences (or MSn) in an ion
trap mass spectrometer following protonated ion formation
by electrospray ionization (ESI). The approach we describe
involves the use of MSn (n D 1–4) in conjunction with
15N labeling of the nucleophiles. The case of ring opening
of 1,3-di-tert-butylaziridinone (1, R1 D R2 D tert ꢀ Bu) by
tert-butylamine is especially interesting because without 15
N
labeling, the two possible products 4 and 5 are identical
(Fig. 1(b)).
EXPERIMENTAL
ESI mass spectrometry
ŁCorrespondence to: E. R. Talaty, Department of Chemistry,
Wichita State University, Wichita, Kansas 67260-0051, USA.
E-mail: talaty@twsuvm.uc.twsu.edu
Contract/grant sponsor: NSF; Contract/grant number:
EPS-9874932; DUE 9751717.
ESI mass spectra were collected using a Finnigan LCQ-Deca
ion-trap mass spectrometer (Thermoquest, San Jose, CA,
USA). Solutions of the products of the tert-butylaziridinone
ring opening reactions, dissolved in a 50 : 50 mixture of
DOI: 10.1002/jms.242
Copyright  2001 John Wiley & Sons, Ltd.

32
E. R. Talaty et al.
Materials
O
(a)
R1CH C Z
1,3-Di-tert-butylaziridinone was prepared by a previously
described procedure.5 tert-Butylamine bearing the 15N label
was obtained by conversion of 15N-labeled potassium
cyanide (KCN) into N-tert-butylformamide via the Ritter
reaction,6 followed by hydrolysis in 80% overall yield. 15N-
labeled KCN and aniline were purchased from Cambridge
Isotope Laboratories (Andover, MA, USA) and used as
received. The adducts of the amines (with and without the
15N label) with the aziridinone were prepared by allowing
equimolar amounts of the two compounds to undergo
reaction in boiling toluene, followed by evaporation of the
solvent and crystallization from toluene–hexane or hexane
alone.
O
NHR2
H
C
2
C
N
R2
O
R1
1
R1CH C NHR2
Z
3
(b)
O
t-Bu-CH-C-15NH-t-Bu
1,2 cleavage
NH-t-Bu
4
O
C2
H
1N t-Bu + t-Bu15NH2
C3
t-Bu
Two derivatives of glycinamide, viz. N-tert-butyl-2-(tert-
butylamino)ethanamide (6) and N-tert-butyl-2-(anilino)etha-
namide (7), that were needed for ascertaining the proposed
fragmentation pathways of the ions, were prepared by the
following procedure:
O
1,3 cleavage
t-Bu-CH-C-NH-t-Bu
15NH-t-Bu
5
The reactions were conducted in boiling benzene and
the products were purified by crystallization from ben-
zene–hexane or hexane alone.
Figure 1. Competitive modes of cleavage of aziridinones.
To produce these compounds and others in which the
exchangeable amine and amide protons were replaced by
deuterons, 1 mM solutions of the compounds were prepared
in a 1 : 1 (v/v) mixture of D2O and CD3OD and allowed to
incubate for >3 h.
methanol and deionized water, were infused into the ESI-
MS instrument using the incorporated syringe pump and a
flow rate of 5 µl minꢀ1. The atmospheric pressure ionization
stack settings for the LCQ (lens voltages, quadrupole and
octapole voltage offsets, etc.) were optimized for maximum
protonated ion abundance by using the auto-tune routine
within the LCQ Tune program. Following the instrument
tune, the spray needle voltage was maintained at C5 kV, the
N2 sheath gas flow at 35 psi and the capillary (desolvation)
RESULTS AND DISCUSSION
Reaction of 1,3-di-tert-butylaziridinone with
tert-butylamine
°
temperature of 200 C. The capillary temperature was chosen
empirically to provide the maximum protonated ion yields
and minimal fragmentation due to thermal dissociation.
The ion trap analyzer was operated at a pressure of
¾1.5 ð 10ꢀ5 Torr (1Torr D 133.3 Pa) as measured using
the remote ion gauge. Helium gas, admitted directly into
the ion trap, was used as the bath/buffer gas to improve
trapping efficiency and as the collision gas for dissociation
experiments. The activation amplitude (proportional to
collision energy) for CID experiments was 30% of the
maximum ‘tickle’ energy available. For all MSn stages, the
activation Q setting was held at 0.25, the activation time was
30 ms and an isolation width of 0.8–1 u was employed. All
mass spectra shown were the result and sum of 20 analytical
scans. Relative fragment ion abundances, where reported,
were determined using a sum of 20 scans at each collision
step, and are shown relative to the most abundant fragment
ion in each particular CID spectrum.
The mass spectrometric study was conducted by comparison
of the product derived from the amine bearing the 15N
label and that without the label. An example of the ESI
mass spectra, MS/MS and MS3 spectra of the protonated
molecular ion derived from the product of the 1,3-di-tert-
butylaziridinone with 15N-labeled tert-butylamine is shown
in Fig. 2. The CID pathways for the labeled and unlabeled
versions of the aziridinone adduct, generated using CID up
to and including MS,4 are shown in Scheme 1. For the 15N-
labeled compound, the fragment ion abundances, relative
to the most abundant peak in the product ion spectrum,
are shown in parentheses. The product ion abundances
resulting from the CID experiments involving the unlabeled
compound were similar to those for the labeled compound.
As noted earlier, each collision step involved the isolation of
the precursor ion using a 0.8–1 u window, activation at 30%
activation amplitude and a 30 ms activation time.
t-BuNH2*
6 (m.p. 157–158 °C)
NEt3
1. SOCl2
BrCH2CONH-t-Bu
(m.p. 97–98 °C)
BrCH2CO2H
2. t-BuNH2,
NEt3
PhNH2*
7 (m.p. 63–64 °C)
NEt3
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

Regioselectivity in ring opening of aziridinones
33
w/15N*
244
Precursor
ion
[B']
*188
(100%)
*87
(6%)
*143 [D']
(13%)
MS/MS
[E]
[A']
*87
(100%)
[D]
*132
*87
(100%)
[B]
74
MS3
MS4
(60%)
(17%)
[C]
*87
(100%)
[A]
Precursor
ion
w/o 15
N
243
86
187
86
MS/MS
142
MS3
74
131
86
MS4
86
Scheme 1. CID pathways for the product obtained by reaction
of 1,3-di-tert-butylaziridinone with tert-butylamine with and
without 15N label.
Figure 2. Typical ESI mass spectra derived from the product
of the reaction of 1,3-di-tert-butylaziridinone with 15N-labeled
tert-butylamine: (a) ESI mass spectrum; (b)tandem mass
spectrum derived from protonated molecule ion; (c) MS3
spectrum derived from the m/z 188 fragment ion.
and (c) transfer of a positive charge to the amide nitrogen
followed by loss of an 15N-substituted aziridinone to furnish
ion C (m/z 74). Since this ion is devoid of 15N, whereas
ions A and B retained the 15N label, ion C was also
observed in the fragmentation of the unlabeled precursor
ion (m/z 243). Competing with the initial loss of isobutylene
from the precursor ion (m/z 244) are two other pathways
involving the loss of substituted formamides via formal four-
centered transition states but more correctly by stepwise
processes (see below): (a) loss of di-tert-butylformamide
furnishes ion E (m/z 87) [Fig. 3(b)]; and (b) loss of tert-
butylformamide followed by loss of isobutylene affords the
same ion (designated D) or ion F depending on which tert-
butyl group is involved [Fig. 3(c)]. Ion D is also an iminium
ion isomeric with ion F. Ion D is probably more stable than F
because it involves a resonance structure that is a secondary
rather than a primary carbocation.
The evidence in support of the pathways proposed in
Fig. 3 and described above is presented herewith along
with comments on possible mechanisms. Since there are
three tert-butyl groups available in the molecular ion for
prevalent fragmentation to an ion via loss of isobutylene,
and this ion (m/z 188) provides the trigger for the majority of
subsequent fragmentations, it is of importance to determine
which tert-butyl group is eliminated during its formation. If
the C-tert-butyl group were to be eliminated, probably by
charge-remote fragmentation, it would furnish an ion having
the structure shown in Fig. 4, viz. attachment of HC to 6. The
structure of 6 is established firmly based on its synthesis from
The fragmentation pathways leading to the product
ions are shown in Scheme 1 and their likely structures
are depicted in Fig. 3. Although these pathways and those
involving aniline just represent plausible pathways, they
are supported as much as possible by the preparation of
compounds having the same structure as probable fragment
ions with and without the 15N label, and also by incorporation
of deuterons. The formation of all of the product ions A–E
can be explained by the assumption that tert-butylamine
cleaves exclusively the 1,3-bond of the aziridinone and
that electrospray of the methanol–water solution of the
ring opening reaction product (an ˛-aminoamide) results in
dominant protonation of the amine group (furnishing A, B,
D, E) rather than the amide group (furnishing C). Based
on the CID patterns, and as shown in Fig. 3(a), we propose
that the precursor ion (the protonated molecular ion at m/z
244) undergoes loss of isobutylene to generate a product
ion B0 at m/z 188, which then partitions itself along three
pathways: (a) loss of a second isobutylene to give ion A0
(loss of both isobutylenes is probably by a charge-remote
fragmentation similar to those proposed by Gross7) followed
by loss of formamide (via a transition state that is only
formally a four-membered ring) leading to ion A (m/z 87),
an iminium ion, which is a well-known resonance-stabilized
intermediate encountered in organic reactions; (b) loss of
N-tert-butylformamide instead of isobutylene as a major
pathway leading to the same ion (m/z 87) designated B;
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

34
E. R. Talaty et al.
m/z 244
(a)
[C]
m/z 74
O
C
t-Bu
+
O
C
C
−(CH3)2C=CH2
CH2-C-NH-t-Bu
O
NH3-t-Bu
C
NH t-Bu
H
+
*
*NH2
t-Bu
O
+NH2
m/z 188
t-Bu
*NH2
t-Bu
C
*
t-Bu
C
NH
(CH3)2C=CH2
H
O
t-Bu
O
C
[B]
+
C
C
NH t-Bu
t-Bu
NH2 t-Bu
H
CH-t-Bu
O
C
[B']
*N
H
H
H
*N
H
H
+
*NH+
m/z 188
H C NH t-Bu
H
H
(CH3)2C=CH2
m/z 87
O
C
[A']
NH2
[A]
t-Bu
CH
CH-t-Bu
*NH+
H
*
H
+ N
H
O
H
H-C-NH2
m/z 132
m/z 87
(b)
(c)
m/z 244
m/z 244
O
C
O
C
t-Bu CH
*
NH t-Bu
t-Bu CH
NH t-Bu
*
+
H
+N
H
H2N
t-Bu
t-Bu
O
O
t-Bu
C
NH t-Bu
H C NH t-Bu
m/z 87
CH
t-Bu
CH
t-Bu
H
*+NH2
[E]
m/z 143
[D']
+NH
C
CH3
CH3
CH2
CH3
CH3
CH2
C
m/z 87
CH
CH2
*+NH
t-Bu
or
t-Bu
*+NH2
[D]
[F]
Figure 3. Fragmentation pathways of the protonated molecular ion derived from the product of the reaction of
1,3-di-tert-butylaziridinone with 15N-labeled tert-butylamine.
bromoacetic acid as described in the Experimental section.
However, as shown in Scheme 2, this ion exhibits a very
different pattern of fragmentation as described in Fig. 4,
possibly because it lacks the steric congestion around the
carbonyl group in comparison with the structure proposed
(B0) in Fig. 3a. In particular, the dissociation of protonated 6
does not furnish ions B and C. Although it does give rise to
an ion having m/z 132 (A0, Fig. 4) by loss of isobutylene in
common with ion B0 of Fig. 3, the A0 ion (A0 in Fig. 4) also
exhibits a different behavior in that it furnishes exclusively
a fragment ion A having m/z 76 shown in Fig. 4 and not the
iminium ion A of Fig. 3. Hence the two ions having a common
value of m/z 132 also must have different structures. It is
therefore clear that the C-tert-butyl group of the molecular
ion in Fig. 3 is retained during its fragmentation to ion B0
and also in its subsequent conversion into ion A0. Out of
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

Regioselectivity in ring opening of aziridinones
35
An important consequence of the conversion of B0 into
B is that it establishes the location of 15N in B0, and
therefore in the molecular ion. Ion B0 bearing either 15N
or 14N loses only 101 Th as shown in Fig. 3 and not 102 Th.
When ion B0 is obtained from the deuterated molecular
ion, its conversion into ion B involves the loss of 103 Th
(corresponding to D—CO—ND—t—Bu), thus supporting
the structure of the neutral fragment. This elimination
reaction then would involve a formal four-centered transition
state. Such eliminations have been encountered, in many
fragmentations, in the present investigation and by others.8 If
they occur in a concerted manner, they would be unfavorable
from orbital-symmetry considerations. Hence they very
likely involve extensive stepwise rearrangements, especially
because of the long retention times in the ion trap. One such
mechanism involves ion–molecule complexes as suggested
by Harrison9 and utilized by others.10
m/z 188
O
C
(6 + H+)
CH2
NH t-Bu
*NH2+
t-Bu
(CH3)2C=CH2
O
O
C
CH2
C
NH t-Bu
CH2
NH2
*NH3
+
*NH+2
m/z 132
[A' ]
t-Bu
(CH3)2C=CH2
Conversion of ion B0 into C involves loss of the 15N
label as stated above. This label could be part of tert-
O
C
[A]
CH2
NH2
Bu(15NH2)C
C O as the neutral fragment or of an isomeric
aziridinone. Deuteration studies showed that one of the
hydrogen atoms in ion C was derived not from carbon
(formal four-centered transition state involving loss of tert-
butylaminoketene) but from the amino nitrogen (loss of an
aziridinone as depicted in Fig. 3(a). A similar competition
between loss of an aziridinone or ketene was considered in
the fragmentation of [Ala–Ala]HC by Cordero et al.,11 and
resolved in favor of the aziridinone as the neutral fragment.
The remaining fragmentation pathways in Fig. 3, namely
the formation of ions A0 (probably by a charge-remote
mechanism), A, D and E, are in agreement with deuteration
studies that lend support to the structures proposed for the
nitrogen-containing neutral fragments. For example, when
ions B0 and A0 are generated from the molecular ion devoid
of N-15, in which the protons joined to nitrogen atoms are
replaced by deuterons (m/z 246), their structures should
be [t-Bu—CH(ND2H)—CO—ND—t—Bu]C (m/z 190) and
[t-Bu—CH(ND2H)CO—NDH]C (m/z 134), respectively, as
required by the pathway proposed in Fig. 3. In that
case, ion B0 should lose either H—CO—ND—t-Bu or
D—CO—ND—t-Bu as the neutral fragment, giving rise to
ion B having m/z D both 87 and 88. Similarly, ion A0 should
lose either HCONDH or DCONDH as the neutral fragment
yielding ion A having m/z D both 87 and 88. These results are
exactly what were observed and further confirm which t-Bu
groups were lost from the molecular ion and in what order.
The loss of t-Bu—CO—ND—t-Bu as the neutral fragment
from the same molecular ion (devoid of N-15) should lead to
ion E having an m/z value of 88, as was observed. Finally,
the pathway proposed for the formation of ions D0 and D
from this molecular ion requires them to have m/z values of
143 and 87, respectively, and this prediction was verified.
One may wonder why the parent ion at m/z 244
apparently undergoes loss of isobutylene from the HŁ2N—t-
Bu group rather than the CO—N—t-Bu group (affording II
instead of I). The most probable conformation of this ion is
the one where tert-butyl groups on adjacent atoms (C and
N*C) are anti to each other as shown in Fig. 5. Conversion into
II rather than I results in a greater gain of torsional freedom
*NH3
+
m/z 76
Figure 4. Fragmentation pathways of the protonated
molecular ion derived from compound 6.
w/15N*
Precursor
ion
188
MS/MS
MS3
[A']
[A]
*76
*132
(20%)
(100%)
*76
(100%)
w/o 15
N
Precursor
ion
187
MS/MS
131
75
75
MS3
Scheme 2. CID pathways for compound 6 with and without
15N label.
the remaining two tert-butyl groups that could be lost from
the molecular ion (m/z 244), the one attached to the amino
nitrogen is the most plausible as shown in the structure
B0, while retaining the one joined to the amide nitrogen (to
be lost in the next step leading to ion B). The chief reason
is that B0 undergoes a subsequent loss of 101 Th as a major
pathway to give ion B and N-tert-butylformamide is the most
straightforward structure of the neutral fragment. Otherwise,
ion B0 would have undergone a loss of 45 Th, implying the
loss of HCONH2 as the neutral fragment.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

36
E. R. Talaty et al.
244
t-Bu
t-Bu
CO-NH-t-Bu
CH3
H
CO-NH2
H
H
-
C
CH2 from CO-NH-t-Bu
H
CH3
H
H
t-Bu
t-Bu
I
CH3
CH3
-
C
CH2 from H2*N-t-Bu
H
188
CO-NH-t-Bu
H
H
II
H
t-Bu
Figure 5. Selectivity in fragmentation of the protonated molecular ion derived from the product of the reaction of
1,3-di-tert-butylaziridinone with tert-butylamine.
is depicted in Scheme 3. For the 15N-labeled compound,
the fragment ion abundances, relative to the most abun-
dant peak in the product ion spectrum, are shown in
parentheses. The product ion abundances resulting from
the CID experiments involving the unlabeled compound
were similar to those for the labeled compound. As in
the previous case, the formation of all of the fragment
ions can be explained by the assumption that aniline
cleaves exclusively the 1,3-bond of the aziridinone and that
protonation of the amino amide reaction product during
electrospray ionization occurs mostly at the amino nitro-
gen (furnishing ions A–E and G) and to a minor extent
at the amide group (furnishing ion F) as shown in Figs 6
and 7.
Although the ions A, B, C, and E produced from
the phenyl-substituted compound are also iminium ions,
as in the previous reaction, they are further stabilized
by conjugation with the aromatic ring. However, their
formation apparently involves a more extensive series
of rearrangements. In Fig. 6(a), the precursor ion (the
protonated molecular ion at m/z 264), which has two tert-
butyl groups, behaves in a fashion analogous to the product
ion (m/z 188) in the previous section, which also has two tert-
butyl groups; that is, it undergoes a neutral loss consistent
with the initial ejection of isobutylene (giving rise to ion
E0) followed by loss of formamide to furnish ion E (m/z
163) or by loss of N-tert-butylformamide to furnish ion E
again, by this alternative route similar to that encountered
in the formation of ion B in the previous section. However,
unlike B, this ion (E) undergoes further fragmentation by
multiple pathways: (a) loss of 56 Th (isobutylene) affords ion
A (m/z 107); (b) loss of 15 Th (methyl radical) produces
ion C (m/z 148), which may be a resonance-stabilized,
(entropy) and is the most likely reason for the selectivity
observed in the formation of ion B0, ultimately leading to
fragments A, B and C.
w/15N*
Precursor
ion
264
*208
(11%)
*163
(100%)
MS/MS
*95
*107 *121 *148 69
(100%)(13%)(15%)(42%)(39%)
MS3
[D]
[A]
[B] [C] [G]
86
(100%)
*163
(60%)
[F]
[E]
w/o 15
N
Precursor
ion
263
207
162
MS/MS
94
106
120
147
69
MS3
162
86
Scheme 3. CID pathways for the product obtained by reaction
of 1,3-di-tert-butylaziridinone with aniline with and without 15N
label.
Reaction of 1,3-di-tert-butylaziridinone with
aniline
A comparison of the CID of the protonated molecule ion
with and without 15N using up to and including MS3
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

Regioselectivity in ring opening of aziridinones
37
conjugated cation-radical; (c) and, as depicted in Fig. 6b,
it perhaps undergoes successive 1,2-methide and hydride
shifts to furnish an isomeric ion (E00). This isomeric ion then
partitions itself (Fig. 6c) by loss of propylene to yield ion
B (m/z 121); by isomerization (involving proton transfer)
followed by loss of dimethylketene to furnish ion D (m/z
95), an ion analogous to ion C in the previous section;
and by loss of aniline to produce a carbocation (G) having
an m/z value of 69. One driving force for these extensive
rearrangements may be the production of a variety of
ions, which gain stabilization by conjugation with the
aromatic ring.
Another process starting from the precursor ion at m/z
264 that does not lead to such conjugation is depicted
(a)
m/z 264
m/z 208
O
O
O
C
CH2 C NH t-Bu
t-Bu
H+*N
CH
C
NH t-Bu
t-Bu
H
CH
*N
NH2
*NH2
H
+
H
m/z 208
+
CH2
−(CH3)2C
−(CH3)2C CH2
Ph
Ph
Ph
[E']
O
O
-HC-NH-t-Bu
NH2
-HC
m/z 163
(CH3)3C-CH
+
CH
m/z 163
[E]
t-Bu
*N+
Ph
H
*NH
+
CH3
-
Ph
CH2
- (CH3)2C
[E]
m/z 148
(CH3)2C
Ph
CH
CH2
m/z 107
*N+
*N+
H
H
Ph
[C]
[A]
(b)
m/z 264
H
O
C
(CH3)3
CH
*N+
t-Bu
C
NH t-Bu
-101
m/z 163
H lost from *NH
O
H*N
H
H
[E]
Ph
+
Ph
-101
t-Bu
-HC NH t-Bu
H lost from CH
t-Bu CH
*N+
C
H*NH
m/z 163
+
Ph
H
Ph
CH3 shift
CH3 +
CH3
CH3
CH3
H
C
C
H
+
C
CH3
C
CH3
Ph
H shift
*NH Ph
*NH
CH3
m/z 163
[E"]
CH3
H
C
H
+
CH3
CH2
H
C
C
CH2
H
CH3
C
H*N:
*NH Ph
Ph
Figure 6. Fragmentation pathways of the protonated molecular ion derived from the product of the reaction of
1,3-di-tert-butylaziridinone with 15N-labeled aniline.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

38 E. R. Talaty et al.
[c]
m/z 69
+
CH3
CH CH3
C
[G]
CH2
CH3
CH3
CH3
+
C
C
-H*NH-Ph
CH3
C
m/z 163
CH3
H
H
CH3
CH3
+
C
CH3
H
CH3
CH3
+
CH2
H
C
H
C
C
+
C
CH2
*
NH Ph
H*N:
[E"]
*NH Ph
Ph
-:C(CH3)2 or
-CH3CH=CH2
H+transfer
CH3
m/z 95
+
C
CH3
C
CH2
*NH2
Ph
m/z 121
CH CH3
H
Ph
*NH2
H
+
or
−(CH3)2C
C
CH2
Ph
[B]
*NH
+
H
H2*N
[D]
+
H
Figure 6. (continued).
in Fig. 7(a). This process also proceeds through ion E0
(m/z 208), but instead of yielding ion E, it takes a more
prominent course involving rearrangement through an
aziridinium ion and subsequent loss of N-phenylformamide
to produce ion F (m/z 86), which is devoid of the 15N
aniline moiety. Consequently, the same ion is produced
m/z 264
(a)
O
C
t-Bu
t-Bu
CH
NH
O
C
t-Bu
+ *NH2
CH
NH
-(CH3)2C=CH2
Ph
+ *NH2
H
from the protonated molecule ion that does not contain 15
N
[E'] m/z 208
Ph
(m/z 263).
An alternative and simpler explanation of the formation
of ion F is to assume that the ring opening of the aziridinone
by aniline gives a mixture of products arising from cleavage
of the 1,3-bond and to some extent also of the 1,2-bond as
depicted in Fig. 7(b). However, this assumption is vitiated
by NMR spectroscopy, which gave no evidence of more than
one reaction product, namely, only one set of twin tert-butyl
signals was detected in the purified sample used in the
present study. 1H NMR (CDCl3): υ 1.09 (s, 9H), 1.29 (s, 9H),
1.58 (s, 1H, NH), 3.22 (s, 1H, COCH), 6.3 (broad, CO-NH,
1H), 6.6–7.2 (m, 5H). The presence of an isomeric product
would have generated an additional set of signals. Of course,
the spectrum does not indicate the location of the two t-
Bu groups or phenyl, but indicates merely their presence.
The mass spectral data, on the other hand, provide crucial
evidence for its structure. Hence, it appears that ring opening
of the aziridinone by aniline also occurs at the 1,3-bond.
As in the case of the product derived from tert-
butylamine, we sought further evidence for the proposed
reaction pathways postulated in Figs 6 and 7 involving
aniline. The molecular ion (m/z 264), in principle, could
H
:NH
C
t-Bu
OH
t-Bu
CH
NH
CH
C
+
OH
: *NH
H
+ *NH
Ph
Ph
m/z 86
H
O
H
H
+
-:C
H
H
+
N
N
NH Ph
*
OH
C
CH
C
or
O
H
t-Bu
t-Bu
[F]
*NH-Ph
HC
-
*NH Ph
Figure 7. Other pathways for fragmentation of the protonated
molecular ion derived from the product of the reaction of
1,3-di-tert-butylaziridinone with 15N-labeled aniline.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

Regioselectivity in ring opening of aziridinones
39
(b)
w/15N*
m/z 264
(CH3)3C
Precursor
ion
208
H
O
C
*NH
C
Ph
+NH2
t-Bu
MS/MS
*152
*107
(100%)
(25%)
- (CH3)2C=CH2
w/o 15N
Precursor
ion
207
O
C
CH2
*NH
Ph
m/z 208
H N
+
H
t-Bu
MS/MS
151
106
O
Scheme 4. CID pathways for compound 7 with and without
- HC-*NH-Ph
15N label.
m/z 208
O
CH2
(7 + H+)
[F]
CH2
C
NH t-Bu
NH+
t-Bu
m/z 86
Figure 7. (continued).
*
+NH2
Ph
eliminate isobutylene from either of the two tert-butyl
groups. If this elimination were to occur from the C-tert-butyl
group, the fragment ion (m/z 208) would have the structure
corresponding to protonated compound 7. However, as
depicted in Scheme 4 and Fig. 8, the ion derived from purified
compound 7 does not exhibit the same behavior as ion E0
(m/z 208). In particular, this ion does not furnish ions E or
F. One possible reason for this difference is that ion E0 has
two bulky groups (t-Bu and Ph) on adjacent atoms, a feature
lacking in the ion from compound 7.
O
(CH3)2C=CH2
H-C-NH-t-Bu
CH2
O
CH2-C-NH2
*N+
*
+NH2
H
Ph
m/z 107
Ph
The pathways leading to ions B, D, F and G in Figs 6 and
7 are in agreement with studies of the deuterated precursors.
In particular, these studies enabled us to follow the course
pursued by the hydrogen that is highlighted to set it apart
from the rest. For example, the tri-deuterated molecular ion
underwent loss of both 102 and 103 Th corresponding to loss
of H—CO—ND—t-Bu as well as D—CO—ND—t-Bu (1,1-
as well as 1,2-elimination), resulting in the production of
two isotopic compositions of ion E [Fig. 6(b)]. By comparison
of the successive fragmentation steps of these two isotopic
ions differing by 1 mass unit, it was determined that the
highlighted hydrogen atom was retained in fragment ions
B and D. Similarly, the fragmentation of the tri-deuterated
version of ion E0 [Fig. 7(a)] is consistent with the pathway
proposed for the formation of ion F, wherein the fate of the
highlighted hydrogen atom could be monitored.
m/z 152
Figure 8. Fragmentation pathways of the protonated
molecular ion derived from compound 7.
of the aziridinone,4 and that aniline behaves in the same
manner as tert-butylamine, it is clear that the previously
cited simple guidelines or rules1,2 for predicting the ring
opening of aziridinones are not generally reliable. It has also
been suggested that the main factor involved in determining
such selectivity is nucleophilicity; the stronger nucleophiles
cleaving the 1,2-bond.12 This conclusion was based on studies
of an aziridinone as an intermediate (1, R1 D Ph or CO2Et,
R2 D Me) and in no case was an aziridinone actually
isolated. If nucleophilicity alone were the determining factor
in our case, then an excellent nucleophile such as iodide
should cleave the 1,2-bond. However, as mentioned earlier,
magnesium iodide cleaved exclusively the 1,3-bond.3 It
therefore appears that a variety of factors, both steric and
electronic, could come into play in a given case, and a
subtle balance between them could determine or change the
Comparison of selectivity in ring opening of
1,3-di-tert-butylaziridinone with other
nucleophiles
Since it has been now established that tert-butylamine and
methylamine cause completely opposite types of cleavage
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40

40
E. R. Talaty et al.
outcome. Further work is needed to evaluate the relative
importance of all these factors.
methylamine or by aniline and methylamine, all of which are
primary, is remarkable and could not have been predicted
by previously postulated guidelines.
Comparison of mass spectrometric method with
NMR spectroscopy
Acknowledgements
As mentioned earlier, 1H NMR spectroscopy cannot dis-
tinguish between products 2 and 3 in Fig. 1(a) when the
aziridinone is cleaved by tert-butylamine or aniline without
incorporation of a special isotope (N-15). However, it is
instructive to compare the relative merits of the mass
spectrometric technique described in the present paper in
combination with 15N labeling with corresponding NMR
techniques. For example, one would anticipate a clear dis-
tinction between compounds 4 and 5 in Fig. 1(b) by 13C NMR
spectroscopy based only on coupling between 15N and the
adjacent carbon. On the other hand, such a distinction is
vitiated by the simultaneous existence of one- and two-bond
13C–15N coupling, both of which can be almost equal,13 espe-
cially if there is an intervening ꢀ-system as in the case of 4.
In contrast, the technique described in the present paper on
mass spectrometry is devoid of such an ambiguity.
The purchase of the LCQ-Deca was made possible, in part, by an
NSF EPSCoR grant. M. J. V. S also acknowledges support from the
WSU College of Liberal Arts and the Sciences and NSF through
Grant EPS-9874932. The 300 MHz NMR spectrometer used for the
characterization of various compounds was acquired with the aid of
NSF Grant DUE 9751717.
REFERENCES
1. Lengyel I, Sheehan JC. Angew. Chem., Int. Ed. Engl. 1968; 7:
25.
2. L’abbe G. Angew. Chem., Int. Ed. Engl. 1980; 19: 276.
3. Talaty ER, Dupuy AE, Jr, Utermoehlen CM. J. Chem. Soc., Chem.
Commun. 1971; 16.
4. Yusoff MM, Talaty ER. Tetrahedron Lett. 1996; 48: 8695.
5. Sheehan JC, Beeson JH. J. Am. Chem. Soc. 1967; 89: 362.
6. Ritter JJ, Kalish J. J. Am. Chem. Soc. 1948; 70: 4048.
7. Gross ML. Int. J. Mass Spectrom. Ion Processes 1992; 118/119: 137.
Gross ML, Int. J. Mass Spectrom. 2000; 200: 611.
8. (a) McLafferty FW, Turecek F. Interpretation of Mass Spectra, (4th
edn). University Science Books: Sausalito, CA, 1993; 205–206;
(b) Plaziak AS, Golankiewicz B, Zeidler J, Golankiewicz K. Org.
Mass Spectrom. 1989; 24: 871.
9. Harrison AG. Org. Mass Spectrom. 1987; 22: 637.
10. Bowen RD. Org. Mass Spectrom. 1993; 28: 1577.
11. Cordero M, Houser JJ, Wesdemiotis C. Anal. Chem. 1993; 65:
1594.
CONCLUSIONS
The present results indicate that the combination of ESI ion
trap mass spectrometry, and in particular MSn experiments,
with 15N labeling is a useful approach for determining the
regioselectivity in ring opening reactions of aziridinones by
amines. These experiments were immeasurably aided by the
ability to monitor the retention of the 15N label by fragment
ions through successive CID steps (up to and including MS4).
The contrasting selectivity exhibited by tert-butylamine and
12. Hoffman RV, Nayyar NK, Chen W. J. Org. Chem. 1995; 60:
4121.
13. Coxon B, Johnson LF. Carbohydr. Res. 1971; 20: 105.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 31–40