Mechanism of the reaction of neutral and anionic N-nucleophiles with α-halocarbonyl compounds

Article abstract of DOI:10.1007/s11172-007-0182-1

N-Acylalkylation of neutral and anionic N-nucleophiles with α-halocarbonyl compounds was investigated by quantum chemical methods in terms of the density functional theory and by experimental methods for 2,3-dihydroimidazo[2,1-b]quinazolin-1(10)H-5-one, its N-anion, and simpler model structures. High reactivity of these reagents is determined primarily by stabilization of transition states (TS) by bridge bonds involving halogen or nitrogen atoms rather than by conjugation, as has been commonly accepted. Bridged TS are formed by both the substitution mechanism S _N 2 and the addition-elimination mechanism. α-Haloalkyl-substituted zwitterions, which are potential intermediates of stepwise N-acylalkylation of neutral N-nucleophiles, do not exist in the isolated state, but they are rather efficiently stabilized upon solvation. These zwitterions, as well as analogous O-anions generated from anionic N-nucleophiles, can serve as intermediates of N-acylalkylation, as was demonstrated by localization of the corresponding TS.

Full text of DOI:10.1007/s11172-007-0182-1

Russian Chemical Bulletin, International Edition, Vol. 56, No. 6, pp. 1194—1209, June, 2007
1194
Mechanism of the reaction of neutral and anionic Nꢀnucleophiles
with αꢀhalocarbonyl compounds
A. S. Morkovnik,ꢀ L. N. Divaeva, and V. A. Anisimova
Institute of Physical and Organic Chemistry, Southern Federal University,
194/2 prosp. Stachki, 344090 RostovꢀonꢀDon, Russian Federation.
Fax: +7 (863) 243 4028. Eꢀmail: asmork2@ipoc.rsu.ru
NꢀAcylalkylation of neutral and anionic Nꢀnucleophiles with αꢀhalocarbonyl compounds
was investigated by quantum chemical methods in terms of the density functional theory and
by experimental methods for 2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ5ꢀone, its Nꢀanion,
and simpler model structures. High reactivity of these reagents is determined primarily by
stabilization of transition states (TS) by bridge bonds involving halogen or nitrogen atoms
rather than by conjugation, as has been commonly accepted. Bridged TS are formed by both
the substitution mechanism SN2 and the addition—elimination mechanism. αꢀHaloalkylꢀsubꢀ
stituted zwitterions, which are potential intermediates of stepwise Nꢀacylalkylation of neutral
Nꢀnucleophiles, do not exist in the isolated state, but they are rather efficiently stabilized upon
solvation. These zwitterions, as well as analogous Oꢀanions generated from anionic Nꢀnucleoꢀ
philes, can serve as intermediates of Nꢀacylalkylation, as was demonstrated by localization of
the corresponding TS.
Key words: acylalkylation, transition states, bridge bonds, αꢀhalocarbonyl comꢀ
pounds, 2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ5ꢀone, Nꢀanion of 2,3ꢀdihydroimidꢀ
azo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ5ꢀone, tetrahedral intermediates, addition—elimination
mechanism.
When studying Nꢀacylalkylation of 2,3ꢀdihydroimidꢀ
by the carbonyl group,16 high electron affinity of the reꢀ
agent (see Ref. 17),* and, finally, the formation of addiꢀ
tional nonclassical bonds between the nucleophile or (and)
azo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ5ꢀone (1) with bielectroꢀ
philic αꢀhaloketones (αꢀacylalkyl halides (AH)) 2, we
noticed that there is uncertainty in the interpretation of
the mechanism of these synthetically important transꢀ
formations.1a,b For example, the character and the numꢀ
ber of reaction steps, the role of the second electrophilic
center of AH (the CO group) in Nꢀacylalkylation, and the
factors responsible for higher reactivity of AH compared
to that of usual alkyl halides remain unclear.
Taking into account the second reaction order,2—4 the
Nꢀacylalkylation, like the acylalkylation as a whole, is
most commonly considered as the nucleophilic substituꢀ
tion by the SN2 mechanism (see, for example, Refs 5—8).
In this case, high activity of AH is generally explained by
stabilization of the transition state (TS) of the reaction by
the CO group, stabilization being provided not directly
by the electronꢀwithdrawing character of this group
(cf. Ref. 9) but by its conjugation with the pꢀelectron pair
of the attacked carbon atom of AH.10—14 It was also sugꢀ
gested that this effect could lead to the transformation of
the reagent in TS into the enolate form.15 In addition,
stabilization of TS was sometimes accounted for by such
factors as the promotion of the indirect interaction beꢀ
tween the frontier orbitals of the substrate and the reagent
* In the present study, the terms "reagent" and "substrate" are
used in the sense opposite to that used in SN2 reactions based on
the classification of the reaction under consideration as acylꢀ
alkylation.
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 6, pp. 1150—1164, June, 2007.
1066ꢀ5285/07/5606ꢀ1194 © 2007 Springer Science+Business Media, Inc.

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1195
Scheme 1
R = Ar, Alk, Alk2N; X = Hal
the halogen atom and the CO group,6,7,15,18,19 which imꢀ
part the bridged character to TS. The lastꢀmentioned facꢀ
tor can be illustrated by bridged structures TS1a—TS3a
and TS1b—TS3b for neutral and anionic Nꢀnucleophiles,
respectively.
nism has recently been confirmed by kinetic experiꢀ
ments with pyridine, benzylamine, and their derivaꢀ
tives.22—24 In this case, TS4 corresponding to the inꢀ
tramolecular substitution of the halogen atom was sugꢀ
gested as the major transition state for the transformation
of zwitterion 3 into Nꢀacylalkylation products. Like
SN2ꢀtype transition state TS1a, TS4 has the Nꢀcentered
bridge bond (Scheme 1). However, the existence of strucꢀ
tures TS4 was not confirmed by their quantum chemical
localization.
The SNAE mechanism of acylalkylation is indirectly
confirmed by the fact that the CO group of AH is comꢀ
petitive as the center of the nucleophilic attack, which is
convincingly demonstrated for such nucleophiles as
hydroxylamine,1a thiosemicarbazide,1b and a series of
Oꢀ and Cꢀanions (see, for example, Refs 1 and 25—31) by
the preparative isolation of the transformation products of
tetrahedral intermediates, which are not associated with
acylalkylation, in good yields. Since these data refer also
to anionic substrates, it is reasonable to suggest that the
Nꢀacylalkylation of anionic nucleophiles can also proꢀ
ceed through the initial nucleophilic addition at the
CO group of AH with the only difference that Oꢀanions 5
Besides, the possible stepwise mechanism was proꢀ
posed for the acylalkylation involving the addition of
the nucleophile at the CO group of AH to form zwitterꢀ
ionic tetrahedral intermediate 3* as the first step folꢀ
lowed by the transformation of the latter into the final
product either in one step via the intramolecular substiꢀ
tution of the halogen atom concerted with the carbon—nuꢀ
cleophile bond cleavage1,5,21—24 or via cyclization to
amino oxirane 4 followed by its nucleophilic cleavage.22,23
Hereinafter, we will refer to this nonconcerted reacꢀ
tion pathway as SNAE, i.e., the nucleophilic substituꢀ
tion via the addition and elimination. Direct experiꢀ
mental evidence in support of the possible formation of
zwitterions 3 is lacking. However, the stepwise mechaꢀ
* Analogous tetrahedral intermediates were assumed20 for the
reactions of Nꢀnucleophiles with carbonyl compounds as a
whole.

1196
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
rather than zwitterions serve as tetrahedral intermediates
(see Scheme 1).
tropic TS6 (see Scheme 1) due to instability of the
unsolvated zwitterions.32a,33 Consequently, it can be conꢀ
cluded that, under the conditions of Nꢀacylalkylation,
the substrate would undergo Nꢀhydroxyalkylation to a
greater or lesser extent to give the corresponding hydroxyꢀ
alkyl derivative 6. Apparently, the latter is yet another
potential intermediate of Nꢀacylalkylation, because it
should give Nꢀacylalkyl derivatives via the cyclization to
highly reactive aziridinium cations40 followed by their
cleavage (see Scheme 1).
Evidently, regardless of the mechanism of Nꢀacylꢀ
alkylation, the same structural factors from those menꢀ
tioned above can make a contribution to stabilization of
primary TS and the increase in the reactivity of AH. As
for the nitrogen bridge bonds considered as one of such
factors, the formation of the latter is, apparently, unlikely
in the case of neutral TS generated from molecular
Nꢀnucleophiles, because the nitrogen atom in such
Nꢀnucleophiles bears only one lone electron pair with the
narrowly directed sp3 or sp2 orbital and, consequently, it
is rather difficult to form two rather strong N—CHal and
N—CO bonds. In SN2ꢀtype TS, the formation of an addiꢀ
tional N—CO bond is, to a certain extent, hindered also
by the total positive charge of the Nꢀnucleophile facilitatꢀ
ing its repulsion from the positively charged carbon atom
of the CO group. In addition, doubly bridged transition
states TS3a are somewhat destabilized by the nonlinear
configuration of the N—C—Hal reaction unit, which is
unfavorable for the overlapping between the p orbital of
the attacked carbon atom and the orbitals of the nitrogen
and halogen atoms.
On the contrary, the nitrogen bridge (such as in TS1b)
can, in principle, be formed in anionic TS of Nꢀacylꢀ
alkylation, which are generated from Nꢀanions, because
the nucleophilic nitrogen atom in such species generally
contains two lone electron pairs.
The transition states of Nꢀacylalkylation can be stabiꢀ
lized also by halogen bridge bonds, in particular, in the
case of neutral TS, because the hindrance analogous to
those described above for nitrogen atoms is absent for
halogen atoms. In addition, these TS contain the elecꢀ
tronꢀdeficient center (the CO group), which is typical of
classical halogenꢀbridged structure, such as halonium catꢀ
ions,41—43 the 1ꢀchloroallyl cation,44,45 and dimers of aluꢀ
minum halides Al2X6 (X = Hal).
Therefore, the SNAE mechanism is characterized by
the initial capture of the Nꢀnucleophile by the CO group
as the second electrophilic center of the reagent to form
zwitterionic or Oꢀanionic tetrahedral intermediates. Since
this transformation is not accompanied by the bond cleavꢀ
age, it would be expected that, in the case of the favorable
thermodynamics, the lowest activation barrier should corꢀ
respond to this transformation, and the SNAE mechanism
would be most probable for highly nucleophilic substrates.
Although zwitterions 3 have not been characterized,
it can be predicted that they are absolutely unstable
in the isolated state. Semiempirical and ab initio
(RHF/4ꢀ31G(3ꢀ21G)) calculations showed32a,b,33 that
the simplest zwitterion of carbonyl compounds,
H3N+—CH2—O–, does not correspond to a minimum on
the potential energy surface in the absence of solvation
and it does not exist as an isolated species because the
repulsion between the reagent and the substrate domiꢀ
nates, resulting in the barrierless cleavage of the zwitterꢀ
ion to form NH3 and CH2O. At the same time, the calꢀ
culations demonstrated that oligohydrates of the
H3N+—CH2—O– zwitterion tend to be stabilized, and
this tendency increases with increasing number of solvaꢀ
tion water molecules.32a,33 The possibility of the existence
of zwitterions under the conditions of solvation is conꢀ
firmed also by experimental data. For example, it was
demonstrated that 4ꢀdimethylaminobutyraldehyde and its
derivatives exist in solution in equilibrium with rather
stable cyclic zwitterionic pyrrolidineꢀtype species stabiꢀ
lized due to the absence of prototropically active hydroꢀ
gen and the intramolecular character of their formation.
Zwitterions, which are rather stable under the conditions
of solvation, can be generated also in bimolecular reacꢀ
tions of carbonyl compounds with tertiary Nꢀnucleoꢀ
philes.34—37 At the same time, zwitterions of primary or
secondary Nꢀnucleophiles are extremely unstable because
they undergo very rapid isomerization to give Nꢀhydroxyꢀ
alkyl derivatives (αꢀaminocarbinols).20 For example, the
formation of the zwitterion of piperazine in the reaction
of the latter with pyridylꢀ4ꢀaldehyde proceeds with the
rate constant of ~107 L mol–1 s–1, whereas the prototropic
isomerization of this zwitterion into the corresponding
αꢀaminocarbinol in the spontaneous reaction (transition
state TS5) has the rate constant of ~106 s–1 38
. The isomerꢀ
Evidently, the validity of these considerations and
the actual factors responsible for stabilization of TS
of Nꢀacylalkylation could be estimated by quantum
chemical calculations. However, these investigations
have not been performed so far. Only TS of degenerate
chloroꢀ and fluoroacylalkylation of X– ions with haloꢀ
aldehydes XCH2CHO (X = Cl (2a) or F (2b)) were studꢀ
ied (MINDO and RHF/4ꢀ31G calculations).10,12,13,46 It
should be noted that bridge bonds were not found in the
latter transition states, but it was demonstrated that
ization of the zwitterion proceeds even faster under acid
or base catalysis.38 In a rather strong protonꢀdonating
medium, such processes can, evidently, proceed as the
direct protonation of the zwitterion at the oxygen atom39
characterized by rather high basicity.34,39 Therefore, the
aboveꢀconsidered data show that the Nꢀhydroxyalkylation
in solution proceeds through zwitterions intermediates.
This is the difference from the analogous gasꢀphase reacꢀ
tion, which is concerted and proceeds through 1,3ꢀprotoꢀ

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1197
these TS are characterized by the virtually right X—C—C
angles and the perpendicular orientation of the C—X
bonds with respect to the plane of the aldehyde group
accompanied by a slight shortening of the C—C bond and
an elongation of the C=O bond compared to those in free
compounds 2a,b. These data were interpreted as evidence
for the presence of the stabilizing conjugation between
the p electrons of the attacked carbon atom and the adjaꢀ
cent CO group in TS, and the increase in the rate of
nucleophilic substitution was accounted for by the latter
fact. However, this conclusion seems unwarranted beꢀ
cause the orthogonal arrangement of the C—X bonds and
the CHO group, as well as the direction of motion of
X– ions in the case of formation of TS, can be a conseꢀ
quence of the wellꢀknown steric requirements determinꢀ
ing the optimal trajectory of nucleophiles in the vicinity
of the carbonyl group.47,48
In the present study, we investigated Nꢀacylmethylaꢀ
tion of quinazolinone 1 and its Nꢀanion, as well as of
ammonia and amide ions, as representatives of neutral
and anionic Nꢀnucleophiles predominantly by quantum
chemical methods. The results of our study show that the
reaction can proceed both in the concerted (SN2) and
nonconcerted (SNAE) modes, and the probability of the
latter mechanism increases with increasing nucleophilicꢀ
ity of the substrate. These results confirm also (at least, as
applied to the Nꢀacylalkylation) that it is unreasonable to
attribute high reactivity of AH to stabilization of TS via
the conjugation.8,10—14
Results and Discussion
NꢀAcylmethylation of quinazolinone 1 and its Nꢀanion
with αꢀhaloketones. To study the acylmethylation by exꢀ
perimental methods, we chose haloketones 2c—f as typiꢀ
cal and most widely used acylalkylating agents (Scheme 2).
Due to very low solubility and the presence of the
deactivating electronꢀwithdrawing CO group, quinazoꢀ
linone 1 almost does not react with phenacyl bromide 2c
in acetonitrile or nitromethane at room temperature but
reacts with 2c under reflux to give a mixture of isomeric
N(10)ꢀ and N(1)ꢀphenacyl derivatives 7c and 8c, with the
former predominating (see Scheme 2). Individual comꢀ
pound 7c can be readily prepared by recrystallization of
this mixture; pure isomer 8c, by a procedure described
below.
Both isomers were identified taking into account a
much stronger deshielding of the protons of the exocyclic
Scheme 2
R = H, X = Cl (a); R = H, X = F (b); R = Ph, X = Br (c); R = 4ꢀClC6H4, X = Br (d); R = 2,6ꢀBut2ꢀ4ꢀOHC6H2, X = Br (e); R = But, X = Br (f)

1198
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
1
Table 1. Melting points and H NMR spectra of Nꢀacylmethylquinazolinones 7c—f and 8c
Comꢀ M.p./°C
1H NMR (CDCl3), δ (J/Hz)
pound (solvent)
H(2)
H(3)
H(6)
H(7)
H(8)
H(9) NCH2R
2 H
Other protons
2 H
1 H
7с
218—222 3.86—3.94 4.14—4.21 8.12 (dd, 7.10 (td,
7.45 (t,
6.64 (d,
J = 8.3)
5.55
(s)
7.50—7.58 (m, 2 H, mꢀH
arom.); 7.63—7.66 (m, 1 H,
pꢀH arom.); 8.04—8.08
(m, 2 H, оꢀH arom.)
(EtOH)
(m)
(m)
Jo = 7.8, Jo = 7.6, J = 7.8)
Jm = 1.6) Jm = 0.9)
7d•
•HBr
312—315 3.84—3.92 4.10—4.20 8.10 (d,
7.09 (t, 7.40—7.47 6.60 (d,
J = 7.3) (m) J = 8.4)
5.48
(s)
7.48 (d, 2 H, mꢀH arom.,
J = 8.5); 7.97 (d, 2 H,
оꢀH arom., J = 8.5)
(with
(m)
(m)
J = 7.8)
decomp.)
(EtOH)
306—308
7e•
3.90
(m)
4.18
(m)
8.11 (dd, 7.09 (t,
Jo = 7.8, J = 7.5) Jo = 7.8, J = 8.4)
Jm = 1.6) m = 1.7)
7.78 (td, 6.98 (d,
7.45 (td, 6.69 (d,
5.51
(s)
1.47 (s, 18 H, But); 4.78
(s, 1 H, ОH); 7.92
(s, 2 H, оꢀH arom.)
1.43 (s, 9 H, But);
•HBr (MeNO2)
J
7f•
•HBr
273—275 4.16 (t,
(with
decomp.)
(MeNO2)
4.45 (t, 8.31 (dd, 7.50 (t,
J = 9.2) Jo = 7.9, J = 7.5) Jo = 7.9, J = 8.5) (br.s) 12.05 (br.s,
6.43
J = 9.2)
Jm = 1.5)
J
m = 1.5)
1 H, N+H)
8с
193—197 3.85 (t,
(AcOEt) J = 8.3)
4.28 (t,
J = 8.4)
8.14 (d,
J = 7.9)
7.17 (t,
J = 7.6)
7.64 (t,
J = 7.4)
7.31 (d,
J = 8.2)
4.98
(s)
7.48—7.58 (m, 3 H,
mꢀH arom., pꢀH arom.);
8.03 (d, 2 H, оꢀH arom.,
J = 7.8)
N(10)—CH2 group in Nꢀsubstituted dihydroimidazoquinꢀ
azolinones49a compared to the analogous group at posiꢀ
tion 1. The difference in the chemical shifts of the aboveꢀ
mentioned groups in compounds 7c and 8c (Table 1,
~0.6 ppm) is approximately equal to that observed for
isomeric Nꢀmethyl, Nꢀbenzyl, and Nꢀβꢀphenoxyethyl deꢀ
rivatives of imidazoquinazolinone 1.49a It should be noted
that the N(10)ꢀisomers of some of these structures were
synthesized not only by the Nꢀalkylation but also by the
independent synthesis from the corresponding Nꢀsubꢀ
stituted isatoic anhydrides.49a,b
The regioselectivity of Nꢀphenacylation of comꢀ
pound 1 is similar to that observed in the Nꢀalkylation of
this compound with monoelectrophilic alkylating reagents
under analogous conditions.49a The percentage of derivaꢀ
tive 7c in a mixture with substitution product 8c is ~70%
(1H NMR spectroscopic data). This fact, taking into acꢀ
count the results reported earlier,49a is indicative of the
predominant involvement of the major 1H tautomer of
substrate 1a in the substitution, whereas the 10H tauꢀ
tomer of 1b plays a minor role.
The Nꢀacylalkylation of compound 1 with phenacyl
bromides 2d,e and with αꢀbromopinacolone (2f) proceeds
analogously. The aboveꢀdescribed transformations are of
particular preparative interest because 10ꢀacylmethylimidꢀ
azoquinazolinones of type 7, unlike other N(10)ꢀRꢀ2,3ꢀ
dihydroimidazo[2,1ꢀb]quinazolinꢀ5ꢀones, cannot be synꢀ
thesized by cyclocondensation of the corresponding
Nꢀsubstituted isatoic anhydrides with 2ꢀmethylthioꢀ4,5ꢀ
dihydroimidazole.50
The Nꢀphenacylation of quinazolinone 1 with broꢀ
mide 2c in the presence of bases (KOH or NaH), like the
baseꢀcatalyzed Nꢀalkylation of compound 1 with monoꢀ
electrophilic alkylating reagents,49a,b is characterized by
the opposite regioselectivity, which is associated with the
fact that, under these conditions, the reaction with the
involvement of the Nꢀanion of substrate 1c proceeds by
the ionic mechanism. The reaction affords only 1ꢀsubstiꢀ
tuted isomer 8c (38% yield) corresponding to the attack
on the most nucleophilic center of the Nꢀanion.49a The
low yield of compound 8c is, apparently, attributed to the
competitive decomposition of the reagent under the acꢀ
tion of alkali giving rise to colored byꢀproducts, which
were not observed in the reaction under neutral conꢀ
ditions.
Quantum chemical investigations
of Nꢀacylalkylation
NꢀAcylalkylation by the SN2 mechanism. In addition to
the major 1H tautomer of imidazoquinazolinone 1a, we
chose ammonia, which is the simplest representative of
neutral Nꢀnucleophiles, as the model substrate, and the
simplest Nꢀformylmethylating reagents, viz., haloacetꢀ
aldehydes 2a,b, as AH. The Nꢀformylmethylation reacꢀ
tions under study are presented in Scheme 2. In particuꢀ
lar, the corresponding transition states TS7—TS9 and the
saltꢀlike and basic forms of the resulting Nꢀformylmethyl
derivatives 7—11 are shown.

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1199
The results of calculations of the geometric and energy
characteristics, some bond orders, dipole moments, and
imaginary vibrational frequencies of the reagents, TS, and
Nꢀformylmethylation products (DFT, B3LYP/6ꢀ31G**)
are summarized in Tables 2 and 3. The data on the
bond orders show that neutral SN2ꢀtype transition states
TS7—TS9 (Figs 1 and 2), as suggested above, contain no
Nꢀcentered bridge bonds, because the nucleophilic nitroꢀ
gen atom is bound to the methylene group (the bond
order is 0.37—0.47), and the N—CO bond is absent. Moreꢀ
over, as evidenced from the overlap populations of the
corresponding AOs, the interaction between the nitrogen
atom and the carbon atom of the CO group is antibonding.
The antibonding character of this interaction is confirmed
also by an increase in the N—CH2—CO angles in the
reaction unit by 7—17° relative to the right angle characꢀ
teristic of undistorted trigonalꢀbipyramidal SN2ꢀtype TS.
Unlike the nitrogen atom, the chlorine and fluorine
atoms in TS7—TS9 are bound not only to the methylene
group but also to the carbonyl group and, consequently,
form halogen bridges, resulting in a substantial decrease
in the X—CH2—CO angles from 90° (to 60—79°) (see
Table 2). The order of the additional X—CO bond varies
from 0.19 to 0.45, which is indicative of its energy imporꢀ
tance. The strongest X—CO bond is present in fluorineꢀ
containing TS, where it can have an even higher
order than the CH2—F bond (see Table 2). It should be
noted that, according to the Mulliken analysis, the orꢀ
ders of these bonds in TS of ammonia TS9 (X = Cl or F)
are 0.39, 0.35 (CH2—X) and 0.19, 0.39 (CO—X),
Table 2. Geometric characteristics (bond lengths (d) and bond angles (ω)) and bond orders for the reagents, intermediates, and TS of
acylmethylation of 1Hꢀquinazolinone 1a, ammonia, and their Nꢀanions with haloaldehydes 2a,b (B3LYP/6ꢀ31G** calculations)
Reagents
and TS
X
d/Å
CH2—X N—CH2—X
ω/deg
Bond order
C—Na
N—CH2—CO X—CH2—CO N—СH2 (N—CO) C—X OС—X
2a
2b
Cl
Cl
Cl
—
—
1.80
1.38
1.83
1.78
1.82
1.81
1.81
—
1.86
1.85
1.87
2.37
1.85
2.50
1.94
2.43
—
—
—
—
—
—
—
—
0.73
—
(1.05)
(0.85)
(0.88)
(0.77)
(0.81)
(0.07)
(0.15)
0.37
0.97
0.92
0.93
—
0.96
0.99
0.99
—
0.89
0.90
0.88
0.43
0.38
0.33
0.32
0.39
—
—
—
—
—
—
—
—
—
12•4H2O
13•20H2Ob Cl
15
16
17
19
20
21
1.58
1.59
1.44
1.50
1.48
1.52
1.57
2.72
2.37
1.96
1.96
1.81
1.88
1.88
91.4
93.3
146
146.5
146.3
—
152.7
163.8
167.2
178.4
169.6
169.5
167.0
172.7
36.8
36.8
34.9
35.9
35.3
31.4
40.5
67.0
61.8
100.0
103.8
97.9
107.1
97.0
113.5
110.0
111.1
111.2
111.4
—
112.9
107.3
108.5
78.5
66.2
75.0
60.4
78.8
Cl
Cl
Cl
—
Cl
Cl
Cl
Cl
F
—
—
22
TS7
TS7
TS8
TS8
TS9
0.20
0.38
0.23
0.45
0.19
0.36
0.47
0.40
0.42
Cl
F
Cl
(0.35)c (0.14)c
TS9
F
1.89
1.58
1.71
1.70
1.85
1.75
1.71
1.51
1.49
2.14
2.20
2.12
2.13
1.90
1.83
1.81
1.82
1.82
1.82
2.64
1.83
2.62
—
160.9
95.2
150.8
97.7
147.8
98.0
141.9
100.0
141.4e
—
98.2
36.0
42.4
36.3
40.2
37.1
56.6
34.0
33.3
118.0
63.5
91.6
85.2
65.2
113.4
109.9
113.5
107.6
112.3
89.9
113.5
92.9
—
0.39
(0.74)
(0.56)
(0.60)
(0.52)
(0.58)
0.54 (0.75)
(0.88)
0.35
0.92d
0.98d
0.94d
0.98d
0.94d
0.26
0.92
0.27
—
0.39
—
—
—
—
—
—
—
—
—
—
0.07
—
TS10
TS11
TS12a
TS12b
TS12c
TS13
TS14
TS15
TS16
TS17
TS18
TS19
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
—
(0.87)
0.26 (0.58)
0.21 (0.43)
0.31
Cl
Cl
Cl
2.15
2.28
2.26
172.8
175.4
176.0
109.8
90.1
93.6
0.60
0.51
0.52
0.29 (0.06)
a For TS, the distance between the attacked carbon atom and the attacking nitrogen atom of the nucleophile.
b The PM3 calculations.
c The bond orders calculated at the CCSD(T)/6ꢀ311G**//B3LYP/6ꢀ31G** level of theory are given in parentheses.
d The N—H, O—H, and C—O bond orders in the reaction unit are 0.54, 0.37, 1.21 (TS10); 0.53, 0.33, 1.34 (TS11); 0.52, 0.35, 1.23
(TS12a); 0.43, 0.41, 1.26 (TS12b); 0.47, 0.38, 1.24 (TS12c).
e For the O—CH2—Cl angle.

1200
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
Table 3. Total energies (Etot), the dipole moments (µcalc) of the reagents, intermediates, and
TS of Nꢀacylmethylation, the energy barriers for the gasꢀphase reaction (∆E≠), and the
imaginary vibrational frequencies of TS (ν )
i
Reagents,
products, and TS
X
–Etota/a.u.
∆E≠
µ
calc/D
ν /cm–1
i
/kcal mol–1
NH3
1a49a
1с49a
2a
2b
11
12•4H2O
13•20H2Ob
15
16
17
19
20
21
22
TS7
TS7
TS8
TS8
—
—
—
Cl
F
—
Cl
Cl
Cl
Cl
Cl
—
Cl
Cl
Cl
Cl
F
56.488220
625.477586
624.919106
613.227262
252.887503
208.9901314
975.241327
349.078082
669.728150
1238.697608
1238.720423
209.295187
669.129127
1238.170869
1238.172220
1238.678400
878.316741
1238.654994
878.289428
669.678839
309.313324
669.671276
1238.691266
1238.645545
1238.641749
1238.647202
669.667937
975.211337
975.192612
265.779155
669.116992
1238.155523
1238.161015
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
1.8
3.7
2.3
1.1
1.1
3.0
3.0
3.5
1.1
1.9
1.8
6.0
3.1
1.1
2.5
7.0
4.0
9.5
8.0
9.9
7.7
3.4
3.5
1.7
3.1
1.6
10.1
3.0
5.6
6.9
1.7
6.3
7.1
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
440.6
547.0
380.7
489.1
405.6
530.2
1336.0
839.0
1570.7
1859.7
1773.3
430.8
1500.2
439.0
422.2
400.4
413.9
417.9
16.6
30.3
31.3
47.5
23.0
39.2
27.7
08.5
37.2
39.6
36.2
37.8
18.8c
30.6
02.7
07.6
9.6 (–15.4)d
7.0 (–16.2)d
Cl
F
Cl
F
TS9
TS9
TS10
TS11
TS12a
TS12b
TS12c
TS13
TS14
TS15
TS16
TS17
TS18
TS19
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
—
Cl
Cl
Cl
a With the ZPE correction.
b Calculations by the PM3 method; the energy is not corrected for ZPE.
c Relative to the structure of 12•4H2O presented in this table rather than relative to the
higherꢀenergy structure formed in the reaction.
d The activation energy for the prereaction reagent—substrate complex is given; the energy
of this complex relative to the reagents is given in parentheses.
respectively. The formation of halogen bridges in
TS is confirmed also by MP2/6ꢀ31G** and
CCSD(T)/6ꢀ311G**//B3LYP/6ꢀ31G** calculations for
TS9 (X = Cl), which gave the similar bond orders (0.40
and 0.35 for CH2—Cl; 0.15 and 0.14 for CO—Cl, respecꢀ
tively).
The formation of bridges causes a substantial decrease
in the dipole moments of TS (µcalc) and the charges on
the halogen atoms due to donation of the electron density
to the CO group, as can be seen from a comparison of
atoms in TS7 and TS8 (X = Cl) calculated by the
B3LYP/6ꢀ31G** method are –0.45 and –0.52, respecꢀ
tively, whereas these charges in the aboveꢀmentioned anaꢀ
logs are substantially higher (–0.73 and –0.67, respecꢀ
tively).
Undoubtedly, the additional CO—X bond is the key
factor of stabilization of neutral TS of SN2ꢀNꢀacylꢀ
alkylation responsible for an increase in the reactivity
of AH in the SN2 reactions with nitrogen nucleophiles.
In addition, the CO...HN hydrogen bonds beꢀ
tween the nucleophile and electrophile can also make
a contribution to stabilization of TS in the case of
µ
calc for TS7 and TS8 (see Table 3) and their formylꢀfree
analogs studied earlier.49a The charges on the chlorine

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1201
0.32
1.94
F
H
Cl
47.3
60.4
72.3
0.45
1.77
a
b
C
H
0.20
78.5
0.43
2.37
35.2
2.53
118.5
0.40
1.88
H
66.3
0.89
1.50
1.75
1.23
H
C
86.3
1.62
1.24
O
O
H
H
N
C
122.6
110.1
109.6
0.16
1.79
H
H
94.9
N
0.94
H
C
H
106.8
1.49
0.37
1.96
C
151.0
H
C
0.76
1.03
C
H
116.8
H
N
C
N
C
C
H
H
123.4
110.2
C
H
C
N
N
C
C
H
H
C
C
C
C
H
C
C
C
C
O
H
O
H
C
H
H
H
Fig. 1. Transition states TS7 (X = Cl) (a) and TS8 (X = F) (b) for the SN2 formylmethylation of tautomer 1a with aldehydes 2a,b
(B3LYP/6ꢀ31G** calculations). Here and in Figs 2—7, the bond orders (upper values), bond lengths/Å (lower values), and bond
angles/deg (indicated by arcs) are given.
F
a
b
Cl
0.19
2.60
46.6
0.35
1.90
0.39
1.86
H
34.6
1.81
1.22
H
C
129.0
0.39
2.43
H
123.4
H
65.2
66.6
68.1
78.8
O
86.9
C
H
0.93
1.51
1.72
1.23
0.90
1.49
0.07
2.20
C
97.0
98.2
H
C
O
0.39
1.89
0.42
1.88
H
112.9
N
0.83
1.03
0.84
1.02
N
H
H
H
H
H
Fig. 2. Transition states TS9 for the SN2 formylmethylation of ammonia with chloride 2a (a) and fluoride 2b (b) (B3LYP/6ꢀ31G**
calculations).
Nꢀnucleophiles containing the primary or secondary
amino group. These hydrogen bonds are present, in parꢀ
ticular, in transition states TS7—TS9 (X = Cl and F) (see
Figs 1 and 2).* The total stabilizing effect of the carbonyl
group in these TS due to the aboveꢀmentioned two factors
is significant and can be estimated from the activation
energy difference (∆∆E≠) between the Nꢀalkylation of the
substrates with methyl halides49a and haloaldehydes 2
(see Table 3). For substrates 1a and NH3 with X = Cl in
the absence of solvation, ∆∆E≠ is ~10—15 kcal mol–1. In
the presence of solvation, this difference and, conseꢀ
quently, the difference in the reactivity of AH and alkyl
halides should be substantially smaller due to a much
higher polarity and solvation stabilization of TS of
Nꢀmethylation compared to TS for formylꢀcontaining
analogs (TS of the reaction of 1a with MeCl have
µ
calc ~13.4 and ~10.4 D).49a It should be emphasized that
double bridge bonds analogous to those present in hypoꢀ
thetical structures TS3 were found in none of the TS
under consideration.
* The hypothesis about the presence of this type of hydrogen
bonds in TS of Nꢀacylalkylation of amines has been made
earlier.51

1202
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
Further studies of structures TS7—TS9 showed that
there is virtually no conjugation with the carbonyl group,
which was generally considered as a factor responsible for
stabilization of TS of acylalkylation in modern investigaꢀ
tions (see above). Actually, the C—CO bond order
in TS7—TS9 is even smaller than unity (0.89—0.96).
For comparison, the C—C bond order in the enolate
anion CH2=CH—O– calculated by the B3LYP/6ꢀ31G**
method is 1.59. Like in the aboveꢀmentioned TS
[Hal...CH2(CHO)...Hal]–,10 the C—C and C=O bond
lengths in TS7—TS9 (1.48—1.51 and 1.22—1.24 Å, reꢀ
spectively) are intermediate between the calculated bond
lengths in free aldehydes 2a,b (1.52 and 1.21 Å, respecꢀ
tively) and the enolate anion CH2=CH—O– (1.38 and
1.27 Å, respectively). However, this is not due to conjuꢀ
gation but due to the presence of halogen bridges in TS.
In particular, a slight elongation of the C=O bond comꢀ
pared to that in free aldehydes 2a,b is associated with a
decrease in the order of this bond as a result of additional
bonding between the CO group and the halogen atom.
The conjugation with the carbonyl fragment is hinꢀ
dered because the corresponding pꢀAO of the carbon atom
of the methylene group is already involved in the bonding
with the entering and leaving groups and, to a lesser exꢀ
tent, because of a deviation of the C—Hal bond from the
optimal (orthogonal with respect to the plane of the formyl
group) direction (by ~20—30°). Apparently, these factors
are also responsible for the virtually complete absence of
conjugation in SN2ꢀtype TS involving allyl halides, viz.,
C analogs of AH 2.52
sible under gasꢀphase conditions (see Schemes 1 and 3).
This is confirmed by the results of investigation of the
NH3—2a and 1a—2a systems. For these systems, there
are stationary points TS10—TS12 and 15—17 on the poꢀ
tential energy surface corresponding to TS of Nꢀhydroxyꢀ
alkylation and haloalkylaminocarbinols generated from
these TS, respectively (see Scheme 3 and Tables 2 and 3).
Transition states TS10 and TS12 (Fig. 3) correspond to
the Nꢀhydroxyalkylation of ammonia and position 1 of
quinazolinone 1a. These transition states have a structure
with the fourꢀmembered 1,3ꢀprototropic reaction unit
typical of such transformations33,47 and a rather high enꢀ
ergy (see Table 3). To the contrary, transition state TS11
for hydroxyalkylation of quinazolinone 1a at the N(10)
atom (see Fig. 3) contains the weakly strained sixꢀmemꢀ
bered 1,5ꢀprototropic reactive ring formed with the inꢀ
volvement of the adjacent NH group and, consequently,
has a rather low energy relative to the reagents. The naꢀ
ture of this TS is indicative of a rather unusual reaction
mechanism characterized by the attack on the tertiary
N(10) atom and the prototropic shift of the H atom from
the N(1) atom rather than from the attacked N atom,
which is observed in the case of the usual Nꢀhydroxyꢀ
alkylation by the 1,3ꢀprototropic mechanism.33,47 The
activation energies ∆E≠ for transition states TS10—TS12
are 27.7, 8.5, and 36.2 kcal mol–1, respectively. Unlike
TS11, isomeric TS12 is diastereomeric because it conꢀ
tains the tetrahedral N(1) atom and was localized for the
stereo forms R(N),S(C) (TS12a) and R,R (TS12b,c)
(see Tables 1 and 2). The R,R structures TS12b,c are
formally conformers of each other and differ in the angle
of rotation of the formyl group characterized by the
O=C—C—Cl dihedral angle (–82 and 178°, respecꢀ
tively). All three forms are similar in energy, and structure
TS12c corresponding to the reaction pathways with the
minimal energy consumption has the lowest energy (see
Table 3).
According to the results of calculations for chloroꢀ
aldehyde 2a, the latter is characterized by a much higher
reactivity in the Nꢀformylmethylation by the SN2
mechanism compared to fluoroaldehyde 2b (∆∆E≠
≈
–14—–16 kcal mol–1, see Table 3).
Therefore, the SN2 mechanism accounts for the rather
efficient Nꢀacylalkylation of nucleophiles and directly reꢀ
lates the high reactivity of AH to particular structural
features of TS of the reaction.
The fact that the SN2 Nꢀacylalkylation and the Nꢀhydrꢀ
oxyalkylation of quinazolinone 1a with chloroaldehyde
2a in the absence of solvation are mutually competitive
can be estimated by a pairwise comparison of the total
energies or the energies ∆E≠ of the corresponding isoꢀ
meric TS. For four TS involved in these reactions, these
energies decrease in the series TS12c > TS8 > TS7 >
> TS11, in which the energies of the states TS7, TS8, and
TS12c are 8.5, 22.8, and 27.7 kcal mol–1, respectively,
relative to the most stable structure TS11. Therefore, in
the case of the attack of tautomer 1a on the N(10) atom,
the SN2ꢀsubstitution reaction is noncompetitive with
NꢀAcylalkylation by the SNAE mechanism. Calculaꢀ
tions by the RHF/6ꢀ31G** and DFT (B3LYP/6ꢀ31G**)
methods showed that the potential intermediates of
Nꢀformylmethylation of substrates 1a and NH3 by the
SNAE mechanism, viz., αꢀhaloalkylꢀsubstituted zwitterꢀ
ions 12—14 (Scheme 3), like the parent zwitterion
H3N+—CH2O–,32a,33 are absolutely unstable in the abꢀ
sence of solvation and do not exist as isolated species.
Most likely, this is also true for zwitterions of nitrogen
bases with carbonyl compounds as a whole. Conseꢀ
quently, it can be concluded that the mechanism inꢀ
volving the intermediate formation of zwitterions is imꢀ
possible for the gasꢀphase Nꢀacylalkylation of neutral
Nꢀnucleophiles.
the Nꢀhydroxyalkylation (∆∆E≠ = ∆E≠
– ∆E≠
=
S
2
Add
8.5 kcal mol–1). This is not surprising takiNng into account
the aboveꢀmentioned specific features of TS11. To the
contrary, based on the negative energy differences ∆∆E≠
(–4.9 and –4.7 kcal mol–1), the SN2ꢀacylalkylation reacꢀ
tion dominates in the case of the attack of quinazolinone
However, the Nꢀacylalkylation through the direct forꢀ
mation of haloalkylaminocarbinols is, in principle, posꢀ

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1203
Scheme 3
1a on the less nucleophilic N(1) atom, as well as in the
case of the attack of ammonia (and, apparently, of other
Nꢀnucleophiles). In the latter case, the Nꢀacylalkylation
through the Nꢀhydroxyalkylation becomes even more
noncompetitive because the second step of this reaction,
viz., the intramolecular nucleophilic substitution in
haloalkylaminocarbinols, is even more unfavorable beꢀ
cause of the presence of the strained threeꢀmembered
ring in TS. Actually, the activation energy ∆E≠ of the
simplest TS of this type (TS13) for haloalkylaminocarbinol
15 is ~38 kcal mol–1, which is almost 15 kcal mol–1 higher
than ∆E≠ for the direct Nꢀacylalkylation of ammonia with
chloride 2a by the SN2 mechanism (see Table 3). In the
absence of solvation, TS13 is transformed, not into the
expected chloride of aziridinium cation 18, but immediꢀ
ately into the Nꢀformylmethylation product, to be more
precise, into its hydrochloride associate (11•HCl), which
was demonstrated by the intrinsic reaction coordinate
(IRC) method. Cyclic structure 18 is generated in the
course of the reaction but it does not correspond to the
energy minimum. This structure immediately undergoes
the barrierless Oꢀdeprotonation with the Cl– counterion
accompanied by the threeꢀmembered ring opening, so
that all transformations are involved in a single elemenꢀ
tary reaction event (see Scheme 3). To the contrary, salt
18 in solutions, where the nucleophilicity of the anions is

1204
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
sharply decreased, most likely corresponds to the miniꢀ
mum on the potential energy surface and serves as an
intermediate of the reaction.
O
H
H
H
a
C
C
C
H
N
C
C
The fundamental difference of the Nꢀacylalkylation
reaction in solution is that zwitterions stabilized by solvaꢀ
tion can be formed under these conditions, and these
zwitterions can serve as both intermediates of the reaction
and precursors of different byꢀproducts. Calculations for
zwitterionic structures 12—14 by the RHF/6ꢀ31G**
method taking into account solvation (the PCM model;
MeNO2 as the solvent) confirmed that these structures in
solution correspond to energy minima. However, it apꢀ
peared that zwitterions 12—14 cannot be optimized in
terms of the PCM model in calculations by the more
precise DFT (B3LYP/6ꢀ31G**) method. This is because
the energy minima corresponding to these zwitterions are
shallow and, consequently, they are not revealed, because
the solvation energy of zwitterions as well as of other
compounds, which are prone to specific interactions with
the solvent or hydrogen bonding, are underestimated in
terms of the PCM model (cf. Refs 53—55). This concluꢀ
sion is confirmed by the fact that even the tetrasolvates of
zwitterion 12 with H2O, MeNO2, and CHCl3 having the
minimum solvation shell are characterized by the energy
minima in calculations by both of the aboveꢀmentioned
methods without the use of the PCM model. All three
tetrasolvates are stabilized by hydrogen bonds between
the solvents and the NH3+ or CO group of the zwitterion
(cf. Ref. 33) (Fig. 4). According to the results of calculaꢀ
tions by the semiempirical PM3 method, both icosaꢀ
hydrate of zwitterion 12 containing 20 H2O molecules
and icosahydrate of the analogous pyridine zwitterion
(C5H5N+—CH(O–)—CH2Cl) are stabilized, even despite
the absence of a hydrogenꢀbond donor (the NH group) in
pyridine. The length of the resulting C—N bond in these
C
H
H
1.26
1.33
1.17
1.33
C
C
N
0.53
1.17
C
H
N
C
H
0.56
1.71
C
0.33
O 1.34
H
H
1.34
1.30
H
C
H
Cl
H
C
H
b
O
H
H
H
C
C
N
H
C
0.58
1.75
N
Cl
C
C
C
C
0.47
1.24
O
H
C
H
H
0.38
1.32
C
N
H
C
C
H
H
Fig. 3. 1,5ꢀPrototropic transitions states TS11 (a) and TS12c (b)
for the addition of 1Hꢀimidazoquinazolinone 1a at the carbonyl
group of chloroaldehyde 2a (B3LYP/6ꢀ31G** calculations).
Cl
a
b
H
H
O
C
H
H
H
H
N
C
0.88
1.51
H
O
0.97
1.40
H
Cl
N
H
0.73
1.58
H
H
O
O
0.44
1.20
0.93
1.53
0.47
1.23
H
H
H
H
C
C
H
H
1.19
1.33
O
H
0.43
1.25
0.40
1.30
H
O
H
H
O
H
Fig. 4. Tetrahydrate of zwitterion 12 (a) and prototropic transition state TS14 (b) for its catalytic isomerization giving tetrahydrate of
hydroxyalkyl derivative 15 (only the catalyzing water molecule is shown) (B3LYP/6ꢀ31G** calculations).

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1205
two zwitterions is 1.59 and 1.63 Å, respectively, which is
evidence of its substantial weakening.
form, which is presented in Tables 2 and 3 and shown
in Fig. 4.
Let us emphasize that we failed to find TS for the
formation of the zwitterion hydrate 12•4H2O from the
corresponding threeꢀcomponent system NH3—2a—4H2O
by the restricted Hartree—Fock method. This is, apparꢀ
ently, because only of technical difficulties (the complex
character of the potential energy surface in the vicinity
of TS, the low energy barrier of the reaction, and the
necessity of varying a large number of geometric paramꢀ
eters reflecting the mutual arrangement of the reagent
and solvent molecules) or because the solvation shell is
insufficiently representative. We also did not find TS for
the direct transformation of 12•4H2O into tetrahydrate of
Nꢀacylalkyl derivative 11, which is analogous to TS4.
This is circumstantial evidence that the existence of such
nitrogenꢀbridged neutral structures is highly improbable,
as has been mentioned above. At the same time, we loꢀ
cated TS14—TS16 (Figs 4 and 5) corresponding to the
formation of Nꢀformylmethyl derivative 11 from 12•4H2O
in the threeꢀstep reaction involving the prototropic
isomerization of tetrahydrate 12•4H2O into tetrahydrate
of haloalkylaminocarbinol 15 catalyzed by one water molꢀ
ecule, the cyclization of 15 to tetrahydrate of amino
oxirane 19, and the nucleophilic cleavage of 19 (see
Scheme 3). Investigations of the character of the reaction
pathway accompanied by the minimal energy consumpꢀ
tion in the vicinity of TS14 demonstrated that the cataꢀ
lytic isomerization of zwitterions proceeds as the conꢀ
certed twoꢀproton transfer, during which one of the proꢀ
tons of the ammonium group in the zwitterion migrates to
the H2O molecule acting as the catalyst, whereas one of
protons of the latter molecule migrates to the O atom of
the zwitterion. This transformation is characterized by
∆E≠ of 7.4 and 18.8 kcal mol–1 relative to the starting
configuration of the tetrahydrate 12•4H2O (TS14 is diꢀ
rectly generated from this tetrahydrate) and its more stable
The activation energies ∆E≠ for the processes
12•4H2O → 19•4H2O and 19 → 11 are 30.6 and
2.7 kcal mol–1, respectively. Therefore, the aboveꢀconꢀ
sidered results show that the cyclization of the zwitterion
to amino oxirane is most likely the rateꢀdetermining step
of the amino oxirane pathway of the Nꢀacylalkylation
involving zwitterions. However, TS of this step (TS15),
even despite the partially solvated character, has a much
higher energy relative to the reagents than unsolvated
highly polar transition state TS9 (23.0 kcal mol–1) of
direct SN2 formylmethylation of ammonia. Hence, the
Nꢀacylalkylation accompanied by the intermediate forꢀ
mation of haloalkylcarbinol is presumably the noncomꢀ
petitive reaction channel in solution as well.
NꢀAcylmethylation of Nꢀanions. For this type of
Nꢀacylalkylation, the principal possibility of the SNAE
mechanism and its ability to compete with the substituꢀ
tion by the SN2 mechanism were demonstrated. Investiꢀ
gations of TS of the reactions of the amide ion and the
Nꢀanion of imidazoquinazolinone 1c with AH 2a demꢀ
onstrated that the Nꢀacylmethylation of the unsolvated
Nꢀanions proceeds by either the SN2 or SNAE mechanism
depending on the nature of the anionic substrate, primaꢀ
rily, on its nucleophilicity. There is no sharp boundary
between two mechanisms because their TS are structurꢀ
ally similar and the Nꢀacylalkylation in both processes
starts with the more or less pronounced electrophilic atꢀ
tack of the CO group on the reagent. The direct SN2
substitution, which excludes this attack, at least in the
step of formation of the prereaction complex, apparently
cannot proceed.
–
Based on the data for the NH2 ion, it can be conꢀ
cluded that highly nucleophilic Nꢀanions are acylalkylated
by the twoꢀstep SNAE mechanism through the anionic
tetrahedral intermediate. For the NH2–—2a system,
O
H
a
b
H
H
H
H
N
O
H
0.56
1.74
H
O
H
H
H
H
C
H
O
0.94
1.39
1.66
N
C
H
H
H
1.00
1.44
2.14
110.3
1.30
1.47
161.2
C
0.27
2.62
H
0.87
1.49
86.0
43.2
50.8
1.89
H
H
C
N
O
0.25
2.08
Cl
O
H
H
H
H
H
Fig. 5. Transition state TS15 for the cyclization of zwitterion hydrate 12•4H2O giving amino oxirane hydrochloride tetrahydrate 19 (a)
and transition state TS16 for the nucleophilic ring opening in unhydrated cation 19 (b) (B3LYP/6ꢀ31G** calculations).

1206
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
Scheme 4
Barrierless
R = H, Ph
Oꢀanion 20 is formed as such an intermediate (Scheme 4,
Fig. 6). Due to high nucleophilicity of the amide anion,
the C—N bond in anion 20 is rather strong, although it is
substantially loosen (the bond length is 1.57 Å and the
bond order is 0.81). The formation of tetrahedral interꢀ
mediate 20 is strongly exothermic and barrierless, whereas
its transformation into Nꢀformylmethylation product 11
proceeds through lowꢀbarrier (∆E≠ = 7.6 kcal mol–1) TS17
(see Fig. 6). The latter contains the nitrogenꢀcentered
bridge and is structurally very similar to the aboveꢀdisꢀ
cussed hypothetical neutral TS4,22—24 the N—CO bond
in TS17 being substantially stronger than the N—CH2
bond (the bond orders are 0.43 and 0.21, respectively).
Therefore, anionic Nꢀnucleophiles, unlike neutral
nucleophiles, can react with AH in the absence of solvaꢀ
tion by the twoꢀstep SNAE mechanism. Taking into acꢀ
count the barrierless character of the initial step of the
H
H
H
a
b
N
H
N
0.43
2.03
0.21
2.20
H
0.81
1.57
H
1.50
1.29
1.68
1.25
O
H
C
C
H
1.48
0.94
0.84
1.57
C
O
C
0.89
1.85
H
0.60
2.15
H
Cl
Cl
Fig. 6. Tetrahedral intermediate 20 for the Nꢀformylmethylation
of the amide ion with chloroaldehyde 2a by the SNAE mechaꢀ
nism (a) and TS17 for the second reaction step relating intermeꢀ
diate 20 to Nꢀformylmethyl derivative 11 (b) (B3LYP/6ꢀ31G**
calculations).

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1207
and 22 (see Scheme 4). Both these complexes, like tetraꢀ
hedral intermediate 20, correspond to the attack of the
CO group of compound 2a on the lone electron pair of
the nitrogen atom and are stabilized not only by the ionꢀ
dipole interaction but also by the weak N—CO bond (the
order is 0.07 and 0.15, respectively). Evidently, these comꢀ
plexes can be considered, on the one hand, as structures
analogous to the prereaction complexes of the usual SN2
alkylation56,57 and, on the other hand, as strongly loosen
analogs of tetrahedral intermediate 20.
Cl
0.52
2.26
H
H
C
0.96
1.48
85.2
58.4
0.29
2.13
H
C
O
H
N
C
36.5
H
0.06
2.49
C
C
C
101.9
H
N
Apparently, the reactions with weakly nucleophilic
Nꢀanions cannot proceed by the SNAE mechanism, at
least in the absence of solvation, because these Nꢀanions
are, most likely, unable to form tetrahedral intermediates.
For anion 1c, this is confirmed by investigations of the
structures corresponding to its addition at the CO group
of chloroaldehyde 2a. It appeared that these structures
show characteristic features of the dominant repulsion
between the substrate and the reagent and the tendency to
undergo barrierless decomposition.
H
C
C
N
C
C
H
C
H
C
H
H
O
Fig. 7. Transition state TS19 for the Nꢀformylmethylation of the
Nꢀanion of quinazolinone 1c with chloroaldehyde 2a by the SN2
mechanism (B3LYP/6ꢀ31G** calculations).
Hence, neutral Nꢀnucleophiles cannot, in principle,
undergo Nꢀacylalkylation by the addition—elimination
(SNAE) mechanism involving zwitterionic intermediates
in the absence of solvation because of absolute instability
of these intermediates. In solution, zwitterions can be
formed and involved in the Nꢀacylalkylation, at least via
the amino oxirane pathway. However, this reaction pathꢀ
way is noncompetitive with the SN2 mechanism. Dependꢀ
ing on the nucleophilicity, Nꢀanionic substrates react with
AH in the absence of solvation by either the addiꢀ
tion—elimination or SN2 mechanism. The high activity of
AH in reactions with Nꢀnucleophiles is not associated
with the presence of conjugation in the acylalkyl halide
fragment of TS but is determined by stabilization of TS
due to the formation of halogen or nitrogen bridges as
well as in the presence of necessary structural prerequiꢀ
sites, viz., hydrogen bonding between the substrate and
the reagent.
reactions involving highly nucleophilic Nꢀanions analoꢀ
gous to NH2–, it is reasonable to suggest that there is no
special TS corresponding to the substitution by the SN2
mechanism for these processes.
The Nꢀacylalkylation of the weakly nucleophilic
Nꢀanion of quinazolinone 1c is radically different. This
anion reacts with AH 2a by the concerted SN2 mechanism
through isomeric TS18 and TS19 (Fig. 7; see Scheme 4).
The latter transition state corresponding to the substituꢀ
tion at the N(1) atom is 3.5 kcal mol–1 more favorable,
which is consistent with the aboveꢀmentioned fact that
anion 1c undergoes the N(1)ꢀ rather than N(10)ꢀphenꢀ
acylation with phenacyl bromide (2c).
Transition state TS19 contains the nitrogen bridge
with the weak N—CO bond (the bond order is 0.06) and
with the substantially stronger N—CH2 bond (the order
is 0.29). Nevertheless, even this bond with the carbonyl
group causes a noticeable deformation of the N—C—CO
bond angle in the reaction unit in the direction opposite
to that observed in the halogenꢀbridged TS. As a result,
this angle in TS19 is smaller than 90° (~85°), whereas
the Cl—C—CO angle is, on the contrary, larger
than 90° (~94°).
Transition state TS18 contains the halogen bridge with
the weak Cl—CO bond (the order is 0.07) instead of the
nitrogen bridge, which, at first glance, is somewhat unexꢀ
pected. However, this structure of TS is, apparently, atꢀ
tributed to a particularly strong conjugation between the
second lone electron pair of the N(10) atom and the
π system in Nꢀanion 1c, which does not allow this lone
pair to form an additional bond with the CO group.
It should be noted that both transition states of ion 1c
are formed from prereaction ionꢀmolecular complexes 21
Experimental
The 1H NMR spectra were recorded on a Varian XLꢀ300
instrument in CDCl3. The energy calculations of the molecules,
the structure optimization of TS, and calculations by the IRC
method were carried out with the use of the quantum chemical
GAMESS (US) program package58 (PC GAMESS version).ꢀ
The stationary points on the potential energy surface were idenꢀ
tified as TS from the presence of the imaginary vibrational mode
in DFT B3LYP/6ꢀ31G** calculations and in calculations with
the use of the IRC method in both directions from TS. Because
of a large body of calculations, most of the latter calculations
were performed by the RHF/6ꢀ31G method. However, in the
most important cases (for TS17 and TS18) and also for all
ꢀ
A. A. Granovsky, http://www.classic.chem.msu.su/gran/
gamess/index.html.

1208
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
Morkovnik et al.
reactions with the involvement of NH3 and NH2–, the calculaꢀ
tions were carried out at the B3LYP/6ꢀ31G** level of theory.
The transition states, reagents, and intermediates were described
using the Mulliken bond orders and charges. The force matrices
for all structures were calculated by the B3LYP/6ꢀ31G** method.
All energy characteristics are given with the zeroꢀpoint enꢀ
ergy (ZPE) correction obtained using the scaling coefficient
of 0.961.59
Compound 1 was synthesized by analogy with its 7ꢀchloro
derivative60 starting from isatoic anhydride and 2ꢀmethylthioꢀ
4,5ꢀdihydroimidazole. The yield was 75%, m.p. 264—266 °C
(PriOH—DMF). Found (%): C, 63.92; H, 4.99; N, 22.66.
purified by the treatment with boiling MeNO2 (10 mL), and
undissolved quinazolinone 1 (0.13 g) was filtered off. After coolꢀ
ing of the filtrate, hydrobromide of compound 7e was obtained
in a yield of 1.34 g (52.0%). Found (%): C, 60.6; H, 6.5; N, 8.3.
C26H32BrN3O3. Calculated (%): C, 60.7; H, 6.3; N, 8.2.
10ꢀPivaloylmethylquinazolinone (7f). A mixture of quinꢀ
azolinone 1 (0.94 g, 5 mmol) and bromopinacolone (1 mL,
7.5 mmol) in MeCN (30 mL) was refluxed for 28—30 h. After
completion of the reaction, the solvent was distilled off, and the
residue was successively treated with Me2CO (10 mL) and boilꢀ
ing MeNO2 (20 mL). Poorly soluble starting quinazolinone 1
(0.4 g) was separated. After cooling, hydrobromide of comꢀ
pound 7f was filtered off from the filtrate in a yield of 0.99 g
(54.0%). Found (%): C, 54.5; H, 6.5; N, 12.0. C16H22BrN3O.
Calculated (%): C, 54.6; H, 6.3; N, 11.9.
C
10H9N3O. Calculated (%): C, 64.16; H, 4.85; N, 22.45.
1H NMR (CDCl3), δ: 3.84 and 4.30 (both t, 2 H each, CH2, J =
8.3 Hz); 6.88 (br.s, 1 H, NH); 7.19 (t, 1 H, H(7), J = 7.0 Hz);
7.25 (d, 1 H, H(9), J = 7.8 Hz); 7.58 (t, 1 H, H(8), J = 7.7 Hz);
8.13 (d, 1 H, H(6), J = 7.9 Hz).
This study was financially supported by the Russian
Foundation for Basic Research (Project No. 04ꢀ03ꢀ96804)
and the Administration of the Rostov Region.
The physicochemical constants of the newly synthesized
compounds are given in Table 1.
Phenacylation of the molecular form of quinazolinone 1.
A. A solution of quinazolinone 1 (1.87 g, 10 mmol) and phenacyl
bromide (2c) (2.00 g, 10 mmol) in MeCN (20 mL) was refluxed
for 29 h. The solvent was distilled off, the residue was treated
with a concentrated ammonia solution (15 mL), and the reacꢀ
tion product was extracted with chloroform and chromatoꢀ
graphed on an alumina column. The resulting mixture of isoꢀ
meric substitution products was recrystallized from ethanol, and
10ꢀphenacylꢀsubstituted product 7c was obtained in a yield of
1.80 g (58.8%). Found (%): C, 70.55; H, 5.17; N, 13.85.
C18H15N3O2. Calculated (%): C, 70.81; H, 4.95; N, 13.76.
B. The reaction was carried out analogously but in nitroꢀ
methane for 8 h. A mixture of isomers 7c and 8c was obtained in
75.0% yield.
Phenacylation of the Nꢀanion of quinazolinone 1. Quinꢀ
azolinone 1 (1.87 g, 10 mmol) and phenacyl bromide (2c) (5.97 g,
30 mmol) were successively added with stirring to a solution of
KOH (0.67 g, 12 mmol) in DMSO (20 mL) at ~20 °C. After
10 min, water (50 mL) was added to the reaction mixture, and
the reaction product was extracted with chloroform (2×20 mL).
Chloroform was distilled off under reduced pressure. The resiꢀ
due was thoroughly triturated with MeCN (10 mL), and
1ꢀphenacyl derivative 8c was filtered off in a yield of 1.16 g
(38.0%). Found (%): C, 70.74; H, 5.17; N, 14.03. C18H15N3O2.
Calculated (%): C, 70.81; H, 4.95; N, 13.76.
References
1. (a) A. W. Erian, S. M. Sherif, and H. M. Gaber, Molecules,
2003, 8, 793; (b) I. K. Moiseev, M. N. Zemtsova, and N. V.
Makarova, Khim. Geterotsikl. Soedin., 1994, 867 [Chem.
Heterocycl. Compd., 1994, 30, 745 (Engl. Transl.)].
2. J. B. Conant and W. R. Kirner, J. Am. Chem. Soc., 1924,
46, 232.
3. J. B. Conant and R. E. Hussey, J. Am. Chem. Soc., 1925,
47, 476.
4. J. B. Conant, W. R. Kirner, and R. E. Hussey, J. Am. Chem.
Soc., 1925, 47, 488.
5. R. G. Pearson, S. H. Langer, F. V. Forrest, and W. J.
McGuire, J. Am. Chem. Soc., 1952, 74, 5130.
6. A. J. Sisti and S. Lowell, Can. J. Chem., 1964, 42, 1896.
7. J. W. Thorpe and J. Warkentin, Can. J. Chem., 1973, 51, 927.
8. D. M. Kalendra and B. R. Sickles, J. Org. Chem., 2003,
68, 1594.
9. F. G. Bordwell and W. T. Brannen, J. Am. Chem. Soc., 1964,
86, 4645.
10. F. Carrion and M. J. S. Dewar, J. Am. Chem. Soc., 1984,
106, 3531.
11. D. J. McLennan and A. Pross, J. Chem. Soc., Perkin Trans. 2,
1984, 981.
pꢀChlorophenacylquinazolinone (7d). A mixture of quinꢀ
azolinone 1 (0.94 g, 5 mmol) and pꢀchlorophenacyl bromide
(2d) (1.71 g, 5 mmol) in MeCN (25 mL) was refluxed for 42 h,
during which compound 1 was gradually transferred into soluꢀ
tion followed by precipitation of the reaction products. After
completion of the reaction, the precipitate was filtered off and
recrystallized from ethanol. Hydrobromide of compound 7d was
obtained in a yield of 1.20 g (53.3%). Found (%): C, 51.39;
H, 3.59; N, 9.99. C18H15BrClN3O2. Calculated (%): C, 50.07;
H, 3.61; N, 10.21.
2,6ꢀDiꢀtertꢀbutylꢀ4ꢀhydroxyphenacylquinazolinone (7e).
A mixture of compound 1 (0.94 g, 5 mmol) and bromo ketone 2e
(1.64 g, 5 mmol) in MeCN (20 mL) was refluxed for 22 h. The
reaction mixture was cooled, and the precipitate that formed
was filtered off and washed with a small amount of MeCN.
Hydrobromide 7e containing an impurity of the starting comꢀ
pound 1 was obtained in a yield of 1.97 g. The product was
12. R. D. Bach, B. A. Coddens, and G. J. Wolber, J. Org. Chem.,
1986, 51, 1030.
13. D. Kost and K. Aviram, J. Am. Chem. Soc., 1986, 108, 2006.
14. F. A. Carey and R. J. Sundberg, Advanced Organic Chemisꢀ
try, Structure and Mechanisms, 4th ed., Kluwer Press, New
York—Boston, 2000, A, 302.
15. T. I. Yousaf and E. S. Lewis, J. Am. Chem. Soc., 1987,
109, 6137.
16. V. Gineityte, J. Mol. Struct. (Theochem.), 2003, 663, 47.
17. S. Shaik, J. Am. Chem. Soc., 1983, 105, 4359.
18. P. D. Bartlett and E. N. Trachtenberg, J. Am. Chem. Soc.,
1958, 80, 5808.
19. J. March, Advanced Organic Chemistry, Reactions, Mechaꢀ
nisms and Structure, Wiley Intersci. Publ., New York, 1985.
20. T. H. Lowry and K. S. Richardson, Mechanism and Theory
in Organic Chemistry, Harper and Row, New York, 1987,
704; 707; 709.

Mechanism of Nꢀacylalkylation of Nꢀnucleophiles
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 6, June, 2007
1209
21. J. W. Baker, Trans. Faraday Soc., 1941, 37, 632.
22. H. J. Koh, K. L. Han, H. W. Lee, and I. Lee, J. Org. Chem.,
2000, 65, 4706.
23. I. Lee, H. W. Lee, and Y.ꢀK. Yu, Bull. Korean Chem. Soc.,
2003, 24, 993.
43. G. A. Olah, G. K. S. Prakash, R. E. Williams, I. D. Field,
and K. Wade, Hypercarbon Chemistry, WileyꢀInterscience,
New York—Chichester—Brisbane—Toronto—Singapore,
1987.
44. J. M. Bollinger, J. M. Brinich, and G. A. Olah, J. Am. Chem.
Soc., 1970, 92, 4025.
45. D. Kilemet, Z. Mihalic, I. Novak, and H. Vancik, J. Org.
Chem., 1999, 64, 4931.
46. S. Wolfe, D. J. Mitchell, and H. B. Schlegel, Can. J. Chem.,
1982, 60, 1291.
24. K. S. Lee, K. K. Adhikary, H. W. Lee, B.ꢀS. Lee, and I. Lee,
Org. Biomol. Chem., 2003, 1, 1989.
25. Organic Syntheses. An Annual Publication of Satisfactory
Methods for the Preparation of Organic Chemicals, Ed. L. S.
Hegedus, Wiley, Hoboken (New Jersey), 2003, 79, 228.
26. M. Masaki, K. Fukui, and M. Ohta, J. Org. Chem., 1967,
32, 3564.
47. R. M. Minyaev and E. A. Lepin, Mendeleev Commun.,
1997, 189.
27. T. I. Temnikova and E. N. Kropacheva, Zh. Obshch. Khim.,
1949, 19, 1917 [J. Gen. Chem. USSR, 1949, 19 (Engl.
Transl.)].
48. The Chemistry of Functional Groups. Suppl. A3, The Chemistry
of DoubleꢀBonded Functional Groups, Ed. S. Patai, J. Wiley
and Sons, 1997, 1113.
28. C. L. Stevens, W. Malik, and R. Pratt, J. Am. Chem. Soc.,
1950, 72, 4758.
29. C. L. Stevens and E. Farkas, J. Am. Chem. Soc., 1957,
79, 3448.
49. (a) A. S. Morkovnik, L. N. Divaeva, and T. A. Kuz´menko,
Izv. Akad. Nauk, Ser. Khim., 2006, 876 [Russ. Chem. Bull.,
Int. Ed., 2006, 55, 907]; (b) G. H. Hardtmann, G. Koletar,
and O. R. Pfister, J. Med. Chem., 1975, 18, 447.
30. D. T. Mowry, Chem. Rev., 1948, 42, 189.
31. M. Yasuda, T. Ohata, I. Shibata, A. Baba, and H. Matsuda,
J. Chem. Soc., Perkin Trans. 1, 1993, 859.
32. (a) K. Ya. Burstein and A. N. Isaev, J. Mol. Struct.
(Theochem.), 1985, 133(26), 263; (b) A. N. Isaev, Izv. Akad.
Nauk, Ser. Khim., 1994, 227 [Russ. Chem. Bull., 1994, 43,
206 (Engl. Transl.)].
33. I. H. Williams, J. Am. Chem. Soc., 1987, 109, 6299.
34. R. McGrindle and A. J. McAlees, J. Chem. Soc., Chem.
Commun., 1983, 61.
50. G.
E.
Hardtmann,
US
Pat.
4020062;
http://v3.espacenet.com.
51. W. Forster and R. M. Laird, J. Chem. Soc., Perkin Trans. 2,
1982, 135.
52. A. Streitwieser, E. G. Jayasree, S. S.ꢀH. Leung, and G. S.ꢀC.
Choy, J. Org. Chem., 2005, 70, 8486.
53. V. I. Minkin, A. D. Garnovskii, J. Elguero, A. R. Katritzky,
and O. V. Denisko, Adv. Heterocycl. Chem., 2000, 76, 203.
54. F.ꢀT. Hung, W.ꢀP. Hu, T.ꢀH. Li, C.ꢀC. Cheng, and P.ꢀT.
Chou, J. Phys. Chem. A, 2003, 107, 3244.
35. F. L. Weisenborn and J. S. P. Schwarz, US Pat. 3360560;
http://v3.espacenet.com.
55. D. Jacquemin, J. Preat, V. Wathelet, M. Fontaine, and E. A.
Perpete, J. Am. Chem. Soc., 2006, 128, 2072.
36. L. V. Saloutina, M. I. Kodess, and A. Ya. Zapevalov, Izv.
Akad. Nauk, Ser. Khim., 1994, 2177 [Russ. Chem. Bull., 1994,
43, 2057 (Engl. Transl.)].
37. M. L. M. Schilling, H. D. Roth, and W. C. Herndon, J. Am.
Chem. Soc., 1980, 102, 4271.
56. J. K. Laerdahl and E. Uggerud, Int. J. Mass Spectrom., 2002,
214, 277.
57. A. Streitwieser, G. S.ꢀC. Choy, and F. AbuꢀHasanayn, J. Am.
Chem. Soc., 1997, 119, 5013.
58. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S.
Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A.
Nguyen, S. J. Su, T. L. Windus, M. Dupuis, and J. A.
Montgomery, J. Comput. Chem., 1993, 14, 1347.
59. K. K. Irikura, R. D. Johnson, III, and R. N. Kacker, J. Phys.
Chem. A, 2005, 109, 8430.
38. H. Diebler and R. N. F. Thorneley, J. Am. Chem. Soc., 1973,
95, 896.
39. W. P. Jenks, Acc. Chem. Res., 1976, 9, 425.
40. P. E. Fanta, in The Chemistry of Heterocyclic Compounds,
Part 1, Heterocyclic Compounds with Threeꢀ and Fourꢀ
membered Rings, Ed. A. Weissberger, Interscience, New
York—London—Sydney, 1964, 548.
60. T.
Jen
and
B.
Loev,
US
Pat.
3745216;
http://v3.espacenet.com/.
41. G. A. Olah, P. W. Westerman, G. Melby, and Y. K. Mo,
J. Am. Chem. Soc., 1974, 96, 3565.
42. G. A. Olah, Halonium Ions, WileyꢀInterscience, New
York, 1975.
Received November 9, 2006;
in revised form March 9, 2007