N-alkylation of 2,3-dihydroimidazo[2,1-b]quinazolin-1(10)H-5-one. on the cryptoanionic mechanism of N-substitution

Russian Chemical Bulletin, International Edition, Vol. 55, No. 5, pp. 907—919, May, 2006

907

NꢀAlkylation of 2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ5ꢀone.

On the cryptoanionic mechanism of Nꢀsubstitution

A. S. Morkovnik,^ꢀL. N. Divaeva, and T. A. Kuz´menko

Research Institute of Physical and Organic Chemistry, RostovꢀonꢀDon State University,

194/2 prosp. Stachki, 344090 RostovꢀonꢀDon, Russian Federation.

Fax: +7 (863) 243 4028. Eꢀmail: asmork2@ipoc.rsu.ru

Quantum chemical methods involving studies of transition states of the reaction showed

that the main products of Nꢀalkylation of prototropic 2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ

1(10)Hꢀ5ꢀone (1) in the gas phase and under neutral conditions in solution occurring via the

S_N2 mechanism should be N(10)ꢀalkylꢀsubstituted derivatives formed from the 1Hꢀtautomer.

Minor N(1)ꢀsubstituted derivatives in solution can be produced from both tautomers. For the

alkylation of the free Nꢀanion of compound 1, position 1 is attacked first. Validity of concluꢀ

sions concerning the overall regioselectivity of the reaction was confirmed experimentally. In

the absence of solvent, the alkylation proceeds abnormally with a sharp increase in the content

of the 1ꢀsubstituted isomers up to inversion of the regioselectivity of the reaction, which is

explained by the participation in the process of the Hꢀbonded dimer of the substrate (1a)₂,

which undergoes alkylation via the cryptoanionic mechanism.

Key words: 2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ5ꢀone, quantum chemical calꢀ

culations, density functional method, transition states, prototropic tautomerism, Nꢀalkylation,

Hꢀbonded dimers, cryptoanionic reaction mechanism, mechanism of alkyltransferase action.

N(10)ꢀSubstituted 2,3ꢀdihydroimidazo[2,1ꢀb]quinꢀ

azolinꢀ5ꢀones possess diverse biological activity. They

have been synthesized up to presently mainly by ring cloꢀ

sure of the corresponding isatoic anhydrides.^1—6Direct

N(10)ꢀsubstitution in 2,3ꢀdihydroimidazo[2,1ꢀb]quinꢀ

azolinꢀ1(10)Hꢀ5ꢀone (1) under neutral conditions, which

could result in the formation of these compounds, have

not been studied previously. Only the N(1)ꢀalkylation and

N(1)ꢀacylation of compound 1 in the presence of bases

are known^1,7—9and involve, probably, the Nꢀanion of the

substrate as intermediate.

In the present work, the reactivity of different forms of

compound 1 and their role in Nꢀalkylation with allowꢀ

ance for the solvent effect were analyzed by quantum

chemical methods. The calculated data were compared

with the experimental results of Nꢀalkylation, and a preꢀ

parative method for the synthesis of N(10)ꢀsubstituted

quinazolinones 2 was developed.

It is most probable that the main contribution to the

overall yield will be made by the alkylation routes involvꢀ

ing attacks to the nitrogen atoms with nonꢀconjugated

lone electron pairs (LEP).

Evidently, the positional and tautomericꢀsubstrate seꢀ

lectivity of the Nꢀalkylation of imidazoquinazolinone 1 is

determined by the three main factors: nucleophilicity of

the N(1) and N(10) nitrogen atoms in each of its two

tautomers, mobility of the prototropic equilibrium, and

its position. Possible routes of the Nꢀalkylation of tauꢀ

tomers 1a,b and Nꢀanion of 1с in the composition of the

ion pair with the alkaline metal cation via the S_N2 mechaꢀ

nism, transition states (TS) TS1—TS6 corresponding to

these routes, and the corresponding Nꢀalkyl derivaꢀ

tives 2, 3 and salts of their protonated form 4—7 are

presented in Scheme 1.

Quantum chemical studies were performed mainly for

the reaction of electroneutral and Nꢀanionic forms 1a—с

with methyl halides (Hal = Cl, Br, I) and dimethyl sulfate

(DMS). In addition, we studied the benzylation of tauꢀ

tomer 1a with benzyl bromide. The calculations were

mainly performed by the Hartree—Fock (RHF) and denꢀ

sity functional (DFT) methods with the B3LYP funcꢀ

tional in the 6ꢀ31G** basis set.

Results and Discussion

Although annelated quinazolinone 1 should yield only

two types of the final products of Nꢀalkylation, namely,

N(10)ꢀ (2) and N(1)ꢀalkylꢀsubstituted (3) derivatives, the

reaction can include six variants of the electrophilic attack,

because compound 1 has three potentially ambidentate

forms (1Hꢀ (1a), 10Hꢀtautomer (1b), and Nꢀanion (1c)).

The results of calculation of the starting molecules

1a,b, Nꢀanion 1c, and its lithium, sodium, and potassium

salts (Mꢀ1c; M = Li, Na, K), alkylating agents, TS of

Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 5, pp. 876—887, May, 2006.

908

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

Morkovnik et al.

Scheme 1

X = Hal, OSO₃Me; M = Li, Na, K

2—7: R = H (a), Ph (b), pꢀClC₆H₄(c), PhOCH₂(d), pꢀBrC₆H₄OCH₂(e)

reactions, and saltꢀlike products 4—7 (R = H, X = Cl) in

the gas phase are presented in Tables 1 and 2. Imidazoꢀ

quinazolone structures 1a—с have the planar or nearly

planar structures of the tricyclic framework. In this case,

1Hꢀtautomeric form 1a of compound 1, as compared to

10Hꢀtautomer 1b, is characterized by a substantially lower

total energy and thus should be predominant. Imidazoꢀ

quinazolinone 1 forms no 1ꢀ and 10ꢀmetalꢀsubstituted

forms similar to prototropic tautomers 1a,b, because the

metal atoms in its Nꢀmetalꢀsubstituted derivatives are not

localized at any nitrogen atom but are arranged approxiꢀ

mately in the middle between them. In other words, the

Nꢀmetallation of each tautomer 1a,b affords the same

metal derivative with the symmetry C_s. In this case, the

Cryptoanionic mechanism of Nꢀsubstitution

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

909

Table 1. Geometric characteristics and dipole moments of the reactants, products, and transition states and the frequenꢀ

cies of imaginary modes of the TS (ν ) (DFT, B3LYP/6ꢀ31G**)

i

Structure

R in

RCH₂X

X

µ

/D

d/Å

Angle N—C—X^a

/deg

ν

i

calc

/cm^–1

C—N^b

C—X

1a

1b

1c

—

H

Ph

H

Ph

H

H^e

H

Ph ^f

H

—

Cl

Br

3.7

2.5

2.3

6.4

8.8

11.3

2.1

1.9

1.6

4.9

2.3

8.9

10.4

7.9

10.4

9.9

11.4

11.3

5.4

13.4

13.5

14.2

14.1

10.9

11.0

8.5

—

1.80

1.96

2.19

1.45

2.00

—

473.5

470.7

448.3

379.9^d

543.6

550.1

547.3

559.5

456.9^d

507.2

485.0

436.0

424.2

405.6

409.7

440.0

430.4

407.6

409.9

570.4

Liꢀ1c

Naꢀ1c

Kꢀ1c

MeCl

MeBr

MeI^c

Me₂SO₄

BnBr

4

I

—

OSO₃Me

Br

—

Cl

Br

I

Br

1.47

1.51

1.46

1.87

1.94

1.86

2.04

1.76

1.92

1.87

1.92

1.88

2.08

2.22

2.20

2.11

2.22

2.23

2.22

2.12

5

7

TS1

—

2.46

2.59

2.80

2.72

1.94

2.47

2.61

2.87

2.81

2.07

2.43

2.32

2.21

2.19

2.20

2.27

2.19

2.17

2.18

2.35

172.3

168.0

175.6

161.6

170.5

177.5

177.4

159.3

173.8

176.0

179.2

154.6

161.8

167.3

178.4

154.5

161.5

167.0

171.4

OSO₃Me

TS2

Cl

Br

I

Br

OSO₃Me

Cl

TS4

TS5 ^g

Cl

TS5 (M = Li)

TS5 (M = Na)

TS5 (M = K)

TS6 ^g

TS6 (M = Li)

TS6 (M = Na)

TS6 (M = K)

TS7^h

3.0

5.8

8.8

8.0

1.9

3.2

5.3

9.4

^aThe angle formed with participation of the attacked N atom.

^bThe length of the forming bond in the TS and reaction products.

^cThe 6ꢀ311G* basis set was used for the iodine atom in all iodineꢀcontaining structures.

^dThe calculation of ν was performed by the RHF/6ꢀ31G* method.

i

^eData of the calculation by the MP2/6ꢀ31G** methods (for ν , RHF/6ꢀ31G**).

i

^fCalculations were performed by the RHF/6ꢀ31G** methods.

^gTransition state of the free Nꢀanion.

^hCalculations were performed by the RHF/6ꢀ31G method.

metal cation is coordinated by the N(1) and N(10) nitroꢀ

gen atoms forming the fiveꢀmembered chelate cycle. The

metal cation in the chelate cycle is somewhat closer to the

N(1) atom. This fact and a comparison of the calculated

orders of two M—N bonds in all the three salts Mꢀ1c

under study indicate that the M—N(1) bond in them is

somewhat stronger. The structures of salts Liꢀ1c and Kꢀ1с

are shown in Fig. 1. In the series of metals Li, Na, and K,

the sum of orders of two M—N bonds decreases signifiꢀ

cantly due to an increase in the ionic character of Nꢀmetalꢀ

substituted derivatives Mꢀ1c, and the effective posiꢀ

tive charge on the metal atom increases, by contrast,

from +0.45 for Li to +0.72 for K.

When analyzing the data obtained for the methylation

products of compound 1 with methyl chloride, viz., hydroꢀ

chlorides 4—7 (see Table 2), it is noteworthy that the

product of the endothermic reaction of 10Hꢀtautomer 1b

at position 10, namely, hydrochloride 6 (R = H), is therꢀ

910

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

Morkovnik et al.

Table 2. Total energies in the gas phase (E^tot) of the reactants, products, and TS and the energy barriers (∆E^≠) of the Nꢀalkylation of

quinazolinone 1 in the gas phase obtained by the RHF/6ꢀ31G** and DFT (B3LYP/6ꢀ31G**) methods

Structure

R in

RCH₂X

X

–E ^tot(au)

∆E^≠(∆E ^tot)^a/kcal mol^–1

RHF DFT

RHF

DFT

1a

—

622.02709

625.47758

623.992256^b

625.46969

624.91910

632.45766

787.18618

1224.76327

499.97802

499.34180^b

2611.45244

2609.77590^b

6959.28270

778.48437

2842.28569

1125.46557

1125.41244

1125.46583

—

1125.47565

1125.40497

3236.88167

7584.72345

3467.714857

1403.93811

1123.26140^b51.0 (32.1; –3.2^e)

3233.70283^b

—

1b

1c

Liꢀ1c

Naꢀ1c

Kꢀ1c

MeCl

—

H

—

Cl

622.01717

621.45668

628.95222

783.33771

1220.61574

499.06154

—

MeBr

H

Br

2609.50967

MeI ^c

Me₂SO4

BnBr

4

H

Ph

—

H

Ph

H

Ph

H

I

6956.41925

776.00858

OSO₃Me

Br

—

Cl

Br

2838.98590

1121.09183

1121.03743

1121.09374

1120.98489

1121.10266

1121.02711

3231.47861

7578.40212

3460.95270

1398.00018

1121.00738

3231.45846

7578.38371

3460.93716

1397.97624

1120.98450

1121.02464

1120.50632

1127.97208

1282.35969

1719.63997

1120.50969

1127.96940

1282.35858

1719.64026

1742.45077

5

3a^...HCl^d

6

7

TS1

—

38.6 (–2.0)

36.5

27.7

37.8

22.3

31.8 (–6.3)

30.3

23.1

30.4

15.0

I

Br

OSO₃Me

TS2

Cl

Br

I

45.6 (27.0; –6.4 ^e)^b

49.1

39.3

47.6

37.3

59.1

41.0 ^b

—

Br

—

OSO₃Me

Cl

1403.91322

—

30.6

—

28.3 (–17.5)

2.8

TS3

TS4

TS5 ^f

Cl

1125.40256

1124.89259

1132.40919

1287.13799

1724.71595

1124.89575

1132.40683

1287.13767

1724.71720

—

33.9 (–15.0)

7.5

TS5 (M = Li)

TS5 (M = Na)

TS5 (M = K)

TS6 ^f

TS6 (M = Li)

TS6 (M = Na)

TS6 (M = K)

TS7^g

26.2

24.8

23.4

5.4

27.8

25.5

23.2

28.3

16.6

16.4

15.9

0.9

18.1

16.6

15.1

—

Cl

a

∆E^totis the difference of E^totof the products and reactants; when calculating E^totof the products, we used the structures of

hydrochlorides obtained by the IRC method and being local but not necessarily global minima, because their optimization in the

framework of the most favorable localization of the Cl^–anion was not performed.

^bCalculation by the MP2/6ꢀ31G** method.

^cThe 6ꢀ311G* basis set was used for the iodine atom in all structures.

^dThe structure of 5 in the gas phase is unstable and easily transformed into a considerably more stable molecular complex 3a^...HCl.

^eChange in the total energy relatively to complex 3a^...HCl.

^fTransition state for the free Nꢀanion.

^gCalculation by the RHF/6ꢀ31G method.

modynamically very unfavorable compared to both the

starting compounds and isomeric salts 4, 5, and 7. The

main reason for this phenomenon is conjugation in

the partially aromatic pyrimidinone ring. Although the

RHF/6ꢀ31G** method gives the energy minimum for this

product, the more precise DFT calculation shows that

Cryptoanionic mechanism of Nꢀsubstitution

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

911

0.47

0.40

(1.95)

routes of methylation with methyl chloride and are rather

stable to the nucleophilic attack by the Cl^–ion.

Li

0.46

(1.91)

H

For all reactions of Nꢀalkylation, we succeeded to

localize the corresponding TS (TS1—TS6) by the RHF

method. However, the calculation at the higher level

(DFT) did not reveal the most thermodynamically unfaꢀ

vorable TS2 (R = H, X = Cl, Br, I and R = Ph, X = Br)

and TS3 (R = H, X = Cl) for the methylation and

benzylation of tautomers 1a,b, although TS2 (R = H,

X = OSO₃Me) for the reaction with DMS was found. This

result does not mean that some of the TS2 structures are

impossible as transition states, because for R = H,

X = Cl, Br we succeeded to obtain TS2 by the

MP2/6ꢀ31G** method (see Tables 1 and 2). An unsucꢀ

cessful attempt to localize TS3 (R = H, X = Cl) by the

DFT method seems understandable if we take into acꢀ

count the aboveꢀmentioned instability of the assumed

product of TS3 transformation, namely, salt 6, which was

revealed by just this method.

N

C

H

–0.75

–0.68

C

0.95

C

H

C

1.08

1.15

H

(1.34)

(1.33)

0.27

0.00

C

–0.59

C

N

C

H

0.57

O

H

–0.53

A

+0.72

K

0.21

(2.63)

0.24

(2.61)

H

Selected geometric characteristics of the obtained TS,

the frequencies of their imaginary vibrational modes (ν_i),

total energies (E^tot), and the energy activation barriers

∆E^≠calculated from E^totare given in Tables 1 and 2. The

structures of TS1, TS2, TS5, and TS6 for the methylation

of tautomer 1a and Nꢀlithioimidazoquinazolinone Liꢀ1с

with DMS and MeCl, respectively, are shown in Figs 2

and 3 as illustrations for the TS structure.

N

–0.75

C

–0.67

+0.76

C

H

C

1.20

1.32

(1.35)

(1.32)

C

+0.28

+0.01

H

C

–0.56

C

N

C

H

C

+0.57

C

H

As can be seen from the data in Figs 2 and 3 and

Table 1, all the TS under study have the trigonal biꢀ

pyramidal configuration of the reaction center. The TS

containing no metal usually exhibit an insignificant deꢀ

viation of three atoms of the reaction site N, C, Hal from

the ideal linear arrangement (see Table 1). The excepꢀ

tions are the TS of Nꢀbenzylation TS1 and TS2 (R = Ph,

X = Br) for which this deviation is pronounced and

achieves ~20°.* This effect has an electrostatic nature and

is related to the mutual repulsion of the negatively charged

bromine atom and neighboring πꢀelectron cloud of the

phenyl group, which represents a continuous region of

the negative electrostatic potential.¹⁰** Taking into acꢀ

count published data for imidazole and pyridine¹¹and

our results, we can conclude that this effect is probably

inherent in Nꢀbenzylation as a whole and is virtually inꢀ

dependent of the nature of the halogen atom.

H

–0.54

O

B

Fig. 1. Structures of Nꢀlithium (Liꢀ1c; A) and Nꢀpotassium

(Kꢀ1c; B) salts of imidazoquinazolinone 1 (B3LYP/6ꢀ31G**).

For some bonds, their orders and lengths in Å (in parentheses)

are presented; charge values are also given for several atoms.

salt 6 is unstable, at least in the configuration correspondꢀ

ing to methylation. This is caused by its affinity to the

barrierless decomposition to the starting compounds 1b

and MeCl, when the cation of the salt acts toward the

Cl^–ion as the methylating agent. Thus, the methylation

of tautomer 1b with methyl chloride through TS3 in the

gas phase, most likely, is impossible; however, it occurs

with reactants that give less nucleophilic counterions durꢀ

ing the reaction.

Unlike salt 6, isomeric hydrochloride 5 (R = H), which

is also formed by the endothermic attack at the conjuꢀ

gated LEP of the nitrogen atom, corresponds to a miniꢀ

mum on the potential energy surface (PES), although this

compound in the gas phase is prone to exothermic

deprotonation with the chloride ion to form molecular

complex 3a^...HCl. The remaining two methylꢀcontaining

hydrochlorides 4 and 7 correspond to the exothermic

The second case of electrostatic deformation of the

reaction center is observed in metalꢀcontaining TS: TS5

and TS6 (R = H, X = Cl, M = Li, Na, K) corresponding

to the methylation of contact ion pairs (CIP) Mꢀ1c with

* Hereinafter if is not specially mentioned, the data obtained by

the DFT method (B3LYP/6ꢀ31G**) are discussed.

** Validity of this explanation could be verified by studying the

TS containing benzyl halides with strong electronꢀwithdrawing

substituents in the benzene ring for which the electrostatic poꢀ

tential of the πꢀelectron cloud is known¹⁰to decrease sharply.

912

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

Morkovnik et al.

H

–0.51

O

Cl

–0.26

C

0.55

(2.21)

H

+0.59

C

H

0.50

(2.32)

H

C

N

H

C

+0.34

+0.01

+0.31

H

–0.60

–0.27

C

Li

0.27

(2.22)

H

C

+0.88

–0.63

1.30

(1.32)

H

0.50

(1.92)

H

N

C

(1.35)

N

–0.70

N

–0.703

(1.03)

+0.34

0.12

C

–0.71

C

(1.33)

H

C

0.34

(2.04)

C

H

(1.34)

H

+0.82

+0.30

C

(1.87)

–0.11

H

–0.57

H

–0.60

O

1.48

+0.01

C

0.36

C

N

(1.95)

S

+1.36

C

H

–0.55

O

–0.61

C

+0.58

–0.53

(1.53)

C

O

H

O

–0.51

H

–0.10

H

A

C

H

A

Cl

–0.24

H

0.56

(2.19)

–0.50

+0.59

O

0.49

(2.32)

H

C

H

C

H

C

H

C

N

H

C

H

Li

–0.28

–0.55

+0.72

+0.34

+0.03

+0.22

H

0.42

(1.95)

H

0.27

(2.22)

H

C

–0.60

C

–0.57

O

–0.59

N

O

–0.63

C

–0.79

H

(1.08)

H

(1.35)

+0.81

N

C

H

C

H

N

C

N

C

S

(1.32)

H

–0.14

+1.29

+0.35

+0.01

0.39

(1.92)

H

–0.59

H

C

–0.57

–0.53

0.31

(2.07)

C

1.29

(1.53)

O

–0.62

C

N

H

C

H

C

H

C

+0.58

H

C

H

B

H

–0.53

O

Fig. 2. Transition states TS1 (A) and TS2 (B) for the methylaꢀ

tion of tautomer 1a with dimethyl sulfate (B3LYP/6ꢀ31G**).

Designations are the same as in Fig. 1.

B

Fig. 3. Transition states TS5 and TS6 for the Nꢀmethylation of

Nꢀlithiumꢀ2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ5ꢀone (Liꢀ1c;

A and B) with methyl chloride (B3LYP/6ꢀ31G**). Designations

are the same as in Fig. 1.

methyl chloride. Their metal atom, regardless of its naꢀ

ture and direction of the electrophilic attack of the reacꢀ

tant, is always localized at the nitrogen atom adjacent to

the electrophilic carbon atom of alkyl halide. In this case,

the С—Cl bond of methyl chloride and the metal atom lie

virtually in the plane of the heterocyclic framework of the

Nꢀanion. The obvious exception is only TS6 (M = K) in

which both MeCl and the K⁺ion lie substantially higher

than the heterocycle plane due to steric repulsion of the

bulky K⁺ion from the nearest H(9) hydrogen atom. The

deformation of the reaction site in the metalꢀcontaining

TS is also accompanied by a decrease in the N—C—Hal

angles and caused by the shift of the negatively charged

chlorine atom attracting to the metal cation. This deforꢀ

mation, unlike that observed in the TS of Nꢀbenzylation,

manifests the attractive interaction (see Fig. 2). The deꢀ

gree of angular distortion of the N—C—Hal fragment is

close to that detected in the TS of benzylation and deꢀ

pends on the metal nature, decreasing moderately with an

increase in the size of its atom (see Table 1).

In the TS of alkylation under study, the central fragꢀ

ment CH₂R is also deformed, which appears as the shift

of the carbon atom from the trigonal plane toward the

nitrogen atom. The degree of distortion in all TS containꢀ

ing no metal increases, as a whole, with an increase in the

Cryptoanionic mechanism of Nꢀsubstitution

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

913

energy barriers of the reactions ∆E^≠. This effect is maxiꢀ

mum in TS2 (R = H, X = Cl, Br), whose C—H—H—R

dihedral angle is about 20°. At the same time, in the lowꢀ

energy TS1, TS2 (R = H, X = OSO₃Me) and TS5, TS6

(R = H) for the Nꢀanion, which is characterized by the

longest C—N bond (>2 Å), virtually no deformation is

observed (see Table 1). Evidently, this type of distortions

can be explained in the framework of the Hammond posꢀ

tulate determining the relationship of energetic of the

reaction and geometry of its TS.¹²

For all the reactions under study, the DFT method

gives substantially lower values of the internal energy barꢀ

riers ∆E^≠as compared to those found by the RHF method,

which is a consequence of the much higher energy of

electron correlation in the TS than in the starting reꢀ

actants.

of 31.8 kcal mol^–1for the most favorable (as it turned out)

direct Nꢀmethylation at position 10 of although less reacꢀ

tive but predominant tautomeric form 1a. The third prinꢀ

cipally possible reaction route involving the attack of

the tetrahedral N(1) nitrogen atom in 1Hꢀtautomeric

form 1a is not competitive in the gas phase (∆E^≠

=

45.6 kcal mol^–1). Thus, the formation of the N(10)ꢀ (2)

and N(1)ꢀalkylꢀsubstituted (3) derivatives as the main and

secondary products, respectively, is most probable in the

gasꢀphase Nꢀalkylation of the molecular form of the subꢀ

strate.

By contrast to electroneutral forms of the substrate

1a,b, Nꢀanion 1c is characterized by close nucleophilicity

of the N(1) and N(10) atoms and, being an isolated parꢀ

ticle, is extremely reactive toward alkylating agents

(∆E^≠for the reaction with MeCl is ~1 kcal mol^–1). This is

explained, from the one hand, by a higher (as a whole)

nucleophilicity of the Nꢀanions compared to the correꢀ

sponding molecular forms¹⁷and, on the other hand, by

the absence of a counterion and a solvate shell (cf. Ref. 18).

In "bare" Nꢀanion 1c, judging from the obtained data

(see Table 2), the N(1) nitrogen atom is somewhat more

nucleophilic than the N(10) atom (the correspondꢀ

ing ∆E^≠values for methylation with MeCl are 0.9 and

2.8 kcal mol^–1).

Going to considerations of the energetic of alkylation

in the absence of solvation, note that in both tautomers the

nitrogen atoms with nonꢀconjugated LEPs are much more

nucleophilic than those, whose LEPs are conjugated with

the πꢀsystem (the difference in ∆E^≠is ~10—15 kcal mol^–1

;

see Table 2). In this case, the Nꢀalkylation of tautomers

1a,b with halides MeCl, MeBr, and BnBr even to the

most nucleophilic nitrogen atoms is a very slow reaction,

approximately as the gasꢀphase methylation of ammonia

with methyl chloride for which ∆E_calc^≠≈ 32—38 kcal mol^–1

(see Refs 13—16); MeI and DMS are much more reactive

in this case. As a whole, the reactivity of RCH₂X for the

attack to both nucleophilic positions of tautomer 1a

changes in the order MeCl < MeBr ≈ PhCH₂Br < MeI <

< Me₂SO₄in which the activation energy of substitution

for the attack of the N(10) atom decreases from 31.8 to

15.0 kcal mol^–1. In this case, the reactivity of benzyl broꢀ

mide is unexpectedly low, although this halide is usually

more reactive than primary alkyl bromides. This is reꢀ

lated, most likely, to the fact that steric factors created by

a rather bulky (compared to MeBr) BnBr molecule unfaꢀ

vorably affect the energy of TS1 and TS2 and to the

already mentioned angular distortion of the N—С—Br

reaction site in these TS, which impedes overlapping of

the pꢀorbital of the attacked carbon atom with AOs of the

nitrogen and bromine atoms. This effect is absent in TS1

and TS2 in the reactions with MeBr. Note that the highꢀ

est reactivity of DMS toward tautomer 1a is partially due

to the stabilization of TS1 by the NH^…O hydrogen bond,

whose formation involves the N(1) nitrogen atom and

one of the oxygen atoms of DMS (see Fig. 2, A).

In the presence of the contact counterion М⁺, the

nucleophilicity of Nꢀanion 1с decreases sharply, because

the metal cation, as mentioned above, is localized, in this

case, near the sites of electrophilic attack and impedes the

interaction of the CIP with the alkylating agent. This

effect appears quantitatively as an increase in the energy

activation barriers ∆E^≠for the attack of MeCl to each of

the two nucleophilic nitrogen atoms of the CIP to values

about 15 kcal mol^–1. The reactivity of the CIP, which

depends moderately on the metal nature, increases in the

order Li, Na, K in which ∆E^≠ranges from 13.6 to

16.6 kcal mol^–1

.

To the contrary of the free Nꢀanion, the N(10) atom

in its lithium and sodium salts is more nucleophilic, beꢀ

cause the attack of methyl chloride at this atom is conjuꢀ

gated with the cleavage of a weaker M—N(10) bond and,

hence, requires a lower (by 1.5 and 0.2 kcal mol^–1, reꢀ

spectively) activation energy ∆E^≠than the attack at the

N(1) atom. For the potassium salt (Kꢀ1с), this factor is

not decisive because of low orders of the K—N bonds and

their more ionic character, due to which this salt, as well

as the free anion, should react preferentially at position 1

(∆∆E^≠_(TS5,TS6)for Kꢀ1с is –0.8 kcal mol^–1; see Table 2).

Thus, for the free Nꢀanion and its CIP with the potasꢀ

sium cation, the N(1) atom should be a more reactive

position, while an opposite ratio of nucleophilicities of

the N(1) and N(10) atoms can be expected for the lithium

and sodium salts.

Of the two tautomers of compound 1, 10Hꢀtautomer

1b is more nucleophilic, and its ∆E^≠for the attack

by methyl chloride to the more reactive position 1 is

~28.3 kcal mol^–1. Since the formation of this tautomer

from the main 1Hꢀform 1a requires a consumption of

additional ~5.3 kcal mol^–1(see Table 2), the total energy

expenses in this reaction route increase to 33.6 kcal mol^–1

which already exceeds noticeably an energy barrier

,

On going from the gas phase to solutions, the Nꢀalkyꢀ

lation of ammonia, according to available calculated data,

914

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

Morkovnik et al.

is extremely accelerated and becomes strongly exotherꢀ

mic due to the efficient solvation of the TS and cationic

forms of the alkylꢀsubstituted derivatives.^13—16A similar

influence of solvation should be predicted for the Nꢀalkyꢀ

lation of tautomers 1a,b, because the transition states of

the S_N2 type TS1, TS2, and TS4 have considerable dipole

moments, and the starting reactants are lowꢀpolarity (see

Table 1). The alkylation of free Nꢀanion 1c, by contrast,

depends on the solvent to a considerably less extent, beꢀ

cause here not only TS5 and TS6 but also Nꢀanion 1c are

strongly solvated. These conclusions are confirmed by the

calculation of the free activation energies ∆G^≠of Nꢀalkyꢀ

lation of the considered forms of the substrate in nitroꢀ

methane. The polarized continuum model (PCM) and

DFT/B3LYP method (Table 3) were used in the calculaꢀ

tion. Although it is believed¹⁶that the PCM method someꢀ

what underestimates the energy barriers of Nꢀalkylation,

the results obtained by this method can be used for qualiꢀ

tative and comparative estimations. It follows from the

∆G^≠values presented in Table 3 (taking into account that

∆G^≠> ∆E^≠in this case)¹³that all three possible processes

of alkylation of tautomers 1a,b (but not Nꢀanion 1c) are

strongly accelerated indeed. Nevertheless, the most probꢀ

able direction of Nꢀmethylation of the molecular form

of 1 under these conditions remains to be the attack of the

main tautomer 1a to position 10. The N(1)ꢀmethylation

of minor tautomer 1b via the scheme 1a → 1b → 3a,

judging from the overall energy expenses, is less favorable

(by ~1.9 kcal mol^–1). At the same time, unlike the gas

phase, the attack of the main tautomer 1a at the N(1)

nitrogen atom can also be significant, because it demands

only by 5.0 kcal mol^–1more energy than N(1)ꢀsubstituꢀ

tion in the minor tautomer 1b and by 6.9 kcal mol^–1more

energy than in the main tautomer. This is explained by

the maximum polarity of TS2 (R = H, µ_calc≈ 14 D; see

Table 1), whose energy decreases upon solvation more

strongly than in the case of other TS. As shown by the

data in Table 3, a considerable increase in exothermicity

of Nꢀalkylation processes should be expected on going to

solutions. In addition, note that a small difference in

nucleophilicity of the N(1) and N(10) nitrogen atoms in

free Nꢀanion 1с with an advantage for the N(1) atom is

retained in solutions (see Table 3).

Unlike the reactions of tautomers 1a,b, the Nꢀalkylaꢀ

tion of CIP Mꢀ1c should become slower on going to

solutions, because the starting ion pairs are much more

polar than their TS5 and TS6 (see Table 1). However, this

effect will not play, most likely, a substantial role, because

the overall alkylation rate of Nꢀanions in polar media

depends mainly on the presence of solvateꢀseparated ion

pairs (SIP) and free Nꢀanions, whose concentration unꢀ

der standard conditions of such reactions should be much

higher than that of the CIP and, in addition, the reactivity

of the SIP should be higher than that of the CIP. Taking

into account the initially low difference in nucleophilicity

of the N(1) and N(10) nitrogen atoms in Nꢀanion 1с and

its ion pairs and the spatial separation of the anion and

cation upon the transformation of CIP Mꢀ1c into the

corresponding SIP, we can assume that the regioselecꢀ

tivities of the electrophilic attack of the SIP and free

solvated Nꢀanion are close. The contribution to alkylaꢀ

tion of different types of ion pairs and the free Nꢀanion

should depend on the reaction conditions. According to

literature data for the Оꢀ and Sꢀanions, when the solution

simultaneously contains dissociated anions and SIP, situꢀ

ations of alkylation of virtually only free Nꢀanions are

possible, whereas either SIP do not react,¹⁹or mainly

only SIP react due to concentration predomination.²⁰

Thus, it can be assumed that the Nꢀalkylation of the

molecular form of compound 1 in both the solution and

gas phase has approximately the same regioselectivity,

and similar reactions of Nꢀanion 1с in polar media proꢀ

ceed predominantly at the N(1) nitrogen atom regardless

of the counterion nature.

Table 3. Total free energies (G^tot) of the reactants, TS, and

products and the free energies for the reaction in solution ∆G^≠

and ∆G° (PCM/B3LYP/6ꢀ31G**, MeNO₂, 25 °C)

Strucꢀ R in

X

–G ^{tot a}

(au)

∆G^≠(∆G°)

ture

RCH₂X

/kcal mol^–1

1a

1b

1c

MeCl

MeBr

MeI

Me₂SO₄

BnBr

4

5

7

TS1

—

Cl

Br

I

625.448968

625.440048

624.962166

499.966389

2611.440130

6959.26773

778.467981

2842.257023

1125.442390

1125.405446

1125.450712

1125.390765

3236.864904

7584.701289

3467.684522

1403.896651

1125.379783^b

3236.854991^b

1125.387723

1124.916588

1124.917936

—

H

Ph

H

—

15.4 (–17.0)

15.2

9.7

13.5

12.7

Br

OSO₃Me

Cl

TS2

22.3 (6.2)

21.4

11.7 (–27.8)

7.5

Br

Cl

TS4

TS5

TS6

Cl

6.7

^aSomewhat higher (by ~1.0—1.5 kcal mol^–1) ∆G^≠values were

obtained when nitromethane is replaced by methanol; if is not

specially mentioned, the geometry of the structure calculated at

the B3LYP/6ꢀ31G** level was used.

The results of experimental studies of the alkylation of

quinazolinone 1 confirm these conclusions. The action of

both active alkylating agents (DMS, pꢀchlorobenzyl broꢀ

b

Calculated using the TS geometry obtained by the

MP2/6ꢀ31G** method.

Cryptoanionic mechanism of Nꢀsubstitution

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

915

mide) and less reactive βꢀaryloxyethyl bromides in MeNO₂

or DMF on this compound affords mainly N(10)ꢀ (2) and

(to a less extent) N(1)ꢀsubstituted 2,3ꢀdihydroimidꢀ

azo[2,1ꢀb]quinazolinꢀ5ꢀones (3). Here the highest regioꢀ

selectivity is observed for alkylation with DMS and βꢀarylꢀ

oxyethyl bromides for which the content of 1ꢀsubstituted

compounds in a mixture of isomers does not exceed 10%.

The pꢀchlorobenzylation of compound 1 with pꢀchloroꢀ

benzyl bromide is less selective. In MeNO₂the reaction

affords a mixture of 10ꢀ and 1ꢀisomers 2c and 3c in a ratio

of ~2 : 1, while under more drastic conditions (boiling

DMF) this ratio is already ~1.5 : 1, which is related, most

likely, to a decrease in the regioselectivity of the reaction

because of the influence of the temperature factor.

Despite a low selectivity of alkylation, it can be used

successfully for the preparation of 10ꢀalkylquinozolinꢀ

ones 2, which are considerably less soluble than their

1ꢀisomers 3 and can be rather easily purified by recrystalꢀ

lization.

Probably, molecular forms of the reactants act as alkyꢀ

lating agents toward dimer (1a)₂, although for pꢀchloroꢀ

benzylation we cannot exclude the participation of the

pꢀchlorobenzyl cation in the reaction. Although the role

of similar cations is usually insignificant in benzylation

(for example, imidazole and pyridine react with BnBr via

the S_N2 mechanism^11,23), for substrates with a low reacꢀ

tivity this channel can be rather important or even the

main (cf. Ref. 24). Also note that dimer (1a)₂, probably,

takes some part in homogeneous alkylation, especially in

concentrated solutions; however, under these conditions,

the dimeric route of the reaction is evidently lowly comꢀ

petitive.

A possibility of the Nꢀalkylation of the dimer was conꢀ

firmed by the study of a (1a)₂—MeCl bimolecular system

by the RHF/6ꢀ31G method, which made it possible to

find TS7 corresponding to the N(1)ꢀmethylation of assoꢀ

ciate (1a)₂on the PES of this system (Fig. 4). As comꢀ

pared to free dimer (1a)₂, the dimeric fragment of this TS

is considerably distorted, which appears as the loss of the

planar symmetric structure and a considerable elongation

of the N—H distance in the attacked NH group. The

latter is expressed to such an extent that we can speak

about the real migration of a proton in the TS to the

opposite N(1) nitrogen atom of the second monomer.

Evidently, the migration is caused by a comparatively

high acidity of the N—H bond in compound 1a and its

further enhancement during TS formation. As a result of

the transformation of the starting reactant—substrate comꢀ

plex in TS7, the dimer undergoes transformation into the

tight ion pair 1c^...1H⁺consisting of Nꢀanion 1с, at which

the electrophilic attack of the reactant is directed, and the

protonated form of the substrate 1H⁺(Scheme 2). The

total charges on the cation, Nꢀanion, and MeCl are +0.87,

–0.70, and –0.18, respectively. Since the Nꢀanionic

The βꢀaryloxyethylation or pꢀchlorobenzylation of

compound 1 in DMF in the presence of NaH, when

Nꢀanion 1c is subjected to alkylation, affords almost only

1ꢀsubstituted derivatives 3a,с—e.

The Nꢀ(pꢀchlorobenzylation) of the molecular form

of compound 1 occurs rather nontrivial, with inversion of

regioselectivity, in the absence of solvent at 140 °C and

affords almost pure 1ꢀ(pꢀchlorobenzyl)imidazoquinꢀ

azolinone 3c in high yield (an admixture of 10ꢀisomer 2c

is ~5%). We believe that this fact reflects the replacement

of the reaction mechanism on going from the alkylation

of substrate 1 to the alkylation of its Hꢀbonded dimer (1a)₂.

This is caused, most likely, by a low solubility of comꢀ

pound 1 in pꢀchlorobenzyl bromide and the involvement

of only dimer (1a)₂in the heterogeneous reaction, beꢀ

cause the crystalline lattice of compound 1 is formed of

this dimer.* In the dimer, the LEP of the N(10) atom is

blocked by a hydrogen bond and, therefore, the reaction

is directed to the N(1) nitrogen atom, which is free for the

electrophilic attack. It is known²²that ambident ions in

media of solvents prone to hydrogen bonding are usually

alkylated predominantly to less electronegative nucleoꢀ

philic centers for an analogous reason. Under similar conꢀ

ditions, when βꢀphenyloxyethyl bromide is used as the

reactant, a mixture of phenoxyethylꢀsubstituted derivaꢀ

tives 2d and 3d is formed, which is enriched in 1ꢀisoꢀ

mer 3d, whose content, however, is only ~40% in this

case. The lower yield of the N(1)ꢀsubstitution product in

this reaction is caused, most likely, by a higher solubility

of quinazolinone 1 in βꢀphenoxyethyl bromide, due to

which the process also proceeds to a substantial extent in

the homogeneous regime.

H

C

O

H

C

N

C

H

C

N

C

2.12

H

2.35

C

Cl

C

H

1.87

C

H

C

N

H

C

1.80

C

H

C

1.03

N

H

1.03

H

C

N

C

H

N

H

C

H

O

H

* It is known²¹that structurally close 1,2,3,4ꢀtetrahydroꢀ

pyrimido[2,1ꢀb]quinazolinꢀ1(11)Hꢀ6ꢀone exists in the crystalꢀ

line state just as a dimer similar to (1a)₂.

Fig. 4. Transition state TS7 for the Nꢀalkylation of dimer (1a)₂

with methyl chloride (RHF/6ꢀ31G); bond lengths are given in Å.

916

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

Morkovnik et al.

form 1с is much more reactive than the molecular form,

the electrophilic attack at the sp³ꢀN(1) nitrogen atom is

strongly facilitated.

molecular form of compound 1 can form a basis for the

mechanism of catalytic effect of alkyltransferases, viz.,

enzymes providing the transfer of alkyl groups between

biochemical substrates, which usually proceeds via the

scheme of S_N2ꢀsubstitution.^25,26To the present time, the

detailed mechanism of functioning of transferases remains

to be poorly studied, and a possibility of the cryptoanionic

reaction route is not revealed yet. However, the quantum

chemical study of the biochemical Nꢀmethylation of glyꢀ

cine and guanidylacetic acid (GAA)^25—28by the simulaꢀ

tion of the reaction site of the reacting system showed that

in the TS of the enzymatic reactions the alkylating

NH group of the substrate forms the N—H^…O hydrogen

bond with the oxygen atom of the enzyme. It was also

demonstrated^27,28that this bond forms no N^–…H—X⁺

ion pair in the TS, as in the case of dimer (1a)₂, because

of the insufficient basicity of the oxygen atom and acidity

of the NH group; nevertheless, it provides a noticeable

decrease in the energy barrier of the transfer of the methyl

group. According to data of the quantum chemical simuꢀ

lation, the elongation of the N—H bond of GAA upon the

transformation of the starting ternary substrate—reacꢀ

tant—enzyme complex into the TS, unlike the

(1a)₂—MeCl system, is relatively small, being

only ~0.02 Å.²⁸However, although this TS retains the

N—H bond, the Nꢀmethylation of GAA with adenosylꢀ

methionine, which is the last step of the biosynthesis of

creatine, is also accompanied by the deprotonation of the

attacked NH fragment but only after the system has passed

through the TS.²⁸Therefore, the formation of the TS and

deprotonation proceed although via the concerted mechaꢀ

nism but asynchronically.

Scheme 2

Thus, the (1a)₂—MeCl system can be considered as

a simple model that clearly illustrates the principal possiꢀ

bility of functioning of alkyltransferases via the cryptoꢀ

anionic mechanism as applied to substrates with a suffiꢀ

ciently high acidity of the alkylating NH fragment. The

most obvious biochemical target of cryptoanionic Nꢀalkyꢀ

lation could be the amideꢀtype NH groups containing

a mobile proton in the composition of peptides and

proteins.

X = Cl, Br

The data presented suggest the autocatalytic effect,

which is exerted by the second molecule of 1a in the

dimer, and the reaction itself by its mechanism can be

classified as cryptoanionic Nꢀalkylation. The catalytic inꢀ

fluence of the second molecule of the monomer can be

illustrated by the calculated values of the energy barriers

∆E^≠(RHF/6ꢀ31G) for the gasꢀphase N(1)ꢀmethylation of

tautomer 1а with methyl chloride via the scheme with the

*

The present study indicates sufficient adequacy and

prospects of the modern quantum chemical methods for

studies of Nꢀalkylation reactions in the series of nitrogenꢀ

containing heterocycles. The further use of these methods

will unambiguously serve for considerably deeper insight

into fine but substantial details of the processes occurring

in molecular systems during these practically significant

transformations and, in particular, it should provide more

clear understanding of mechanisms of the catalytic action

of different alkyltransferases.

intermediate formation of the dimer (28.3 kcal mol^–1

)

and by the direct methylation (39.6 kcal mol^–1). Thereꢀ

fore, it is not surprising that in TS7, as in other lowꢀ

energy transition states of Nꢀalkylation, the С—N bond is

considerably elongated (~2.12 Å).

It should be mentioned that activation processes of

weak nucleophiles similar to that described above for the

Cryptoanionic mechanism of Nꢀsubstitution

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

917

Table 4. Physicochemical properties of Nꢀalkylquinazolinones 2c—e and 3a,c—e

Comꢀ M.p./°С

¹H NMR (CDCl₃), δ (J/Hz)

pound (solvent)

H(2)

(1 H)

H(3)

(1 H)

H(6)

(1 H)

H(7)

(1 H)

H(8)

(1 H)

H(9)

(1 H)

NCH₂R

(2 H)

Other protons

2c

2d

226—228¹3.93—4.00 4.18—4.23 8.10 (dd, 7.09 (td,

7.43 (t,

J = 7.5)

6.83 (d,

J = 8.3)

5.26

(s)

7.19—7.33 (m, 4 H,

H arom.)

(MeCN)

(m)

J = 7.8,

J = 1.7)

J = 7.5,

J = 0.9)

191—193

(MeOH)

3.94 (t,

J = 9.1)

4.13 (t, 8.09 (dd, 7.11 (t,

7.57 (t,

J = 7.9)

7.35 (d,

J = 8.2)

4.34 (t,

J = 5.5)

4.46 (t, 2 H, ОСH₂,

J = 5.5); 6.84 (d, 2 H,

oꢀH arom., J = 7.9);

6.94 (t, 2 H, pꢀH arom.,

J = 7.5); 7.24 (t, 2 H,

mꢀH arom., J = 7.9)

4.45 (t, 2 H, ОСH₂,

J = 5.7); 6.72 (d, 2 H,

oꢀH arom., J = 9.0);

J = 9.0)

J = 7.9,

J = 1.6)

J = 7.3)

2e

218.5—220.5 3.89—4.00 4.10—4.19 8.10 (dd, 7.12 (td,

7.58 (t,

J = 7.9)

7.31 (d,

J = 7.9)

4.30 (t,

J = 5.6)

(EtOH—

—DMF)

(m)

J = 7.9,

J = 1.8)

J = 7.5,

J = 1.0)

7.34 (d, 2 H, mꢀH arom.,

J = 9.0)

3a

3c

3d

172.5—174.0 3.68 (t,

4.17 (t, 8.12 (dd, 7.16 (td, 7.57 (td,

7.38 (t,

J = 8.2)

3.08*

(s)

—

(iꢀC₈H₁₈

—

J = 8.3)

J = 7.9,

J = 1.2)

J = 7.6,

J = 1.2)

J = 7.7,

J = 1.7)

7.59 (t,

J = 7.7)

—PrⁱOH)

141—142

(MeCN)

3.53 (t,

J = 8.3)

4.15 (t, 8.13 (dd, 7.19 (t,

J = 8.3)

7.41 (d,

J = 8.0)

4.65

(s)

7.28—7.38

(m, 4 H, H arom.)

J = 8.0,

J = 1.6)

J = 7.7)

125—126

3.91 (t,

4.16 (t,

8.12 (d, 7.17 (td, 7.57 (td, 7.37 (d,

3.91 (t,

J = 5.0)

4.28 (t, 2 H, ОСH₂,

J = 5.0); 6.91 (d, 2 H,

oꢀH arom., J = 7.7);

6.97 (t, H, pꢀH arom.,

J = 7.5); 7.29 (t, 2 H,

mꢀH arom., J = 7.5)

4.24 (t, 2 H, ОСH₂,

J = 5.7); 6.80 (d, 2 H,

oꢀH arom., J = 9.1);

7.39 (d, 2 H, mꢀH arom.,

J = 8.9)

(iꢀC₈H₁₈

—

J = 8.6)

J = 8.0)

J = 8.0,

J = 1.1)

J = 7.6,

J = 1.8)

J = 8.1)

—PrⁱOH)

3e

179—180.5

3.89 (t,

4.19 (t, 8.12 (dd, 7.17 (td, 7.58 (td, 7.36 (d,

3.90 (t,

J = 5.2)

(iꢀC₈H₁₈

—

J = 8.2)

J = 8.8)

J = 7.9,

J = 1.2)

J = 7.5,

J = 1.3)

J = 7.6,

J = 1.6)

J = 7.6)

—PrⁱOH)

* Signals from protons of the NMe group.

chemical program package (US).²⁹In the calculation of iodineꢀ

containing structures for the iodine atom, the intrinsic 6ꢀ311G*

basis set was used in all cases.³⁰Stationary points on the PES were

identified as TS on the basis of the presence of the imaginary

vibrational mode in the calculation by the DFT/B3LYP/6ꢀ31G**

method (except for TS7), and the corresponding reactions were

identified using the study of the reaction route in both directions

from the TS. In this case, the RHF/6ꢀ31G method was used

because of voluminous calculations. Since it is difficult to optiꢀ

mize the TS in solutions, nonꢀrelaxed structures of TS and reacꢀ

tants obtained for the isolated molecules were applied in the

calculations by the PCM model. The values of effective charges

on atoms and bond orders calculated according to Mulliken

were used in discussion. All values of total energies and their

changes were determined with corrections to the zeroꢀpoint viꢀ

brational energy (ZPE). In these calculations we used the scalꢀ

ing coefficient values presented in Ref. 31.

Experimental

1

H NMR spectra were recorded on a Varian Unityꢀ300 inꢀ

strument (300 MHz) in the regime of internal stabilization of

the polarꢀresonance line of ²H of the deuterated solvent.

The ratio of isomers during alkylation was determined by the

¹H NMR method. Authentic samples of 10ꢀmethylꢀ (2a) and

10ꢀ(pꢀchlorobenzyl)ꢀsubstituted (2c) derivatives were syntheꢀ

sized by ring closure of the corresponding Nꢀsubstituted isatoic

anhydrides with 4,5ꢀdihydroꢀ2ꢀmethylthioimidazole.¹1ꢀMethylꢀ

substituted isomer 3a was synthesized by analogy to the preparaꢀ

tion of other 1ꢀsubstituted derivatives of compound 1 (see Ref. 1).

The physicochemical properties of all newly synthesized comꢀ

pounds are presented in Table 4. The energies of molecules were

calculated, the TS structure was optimized, and calculations by

the method of internal reaction coordinate were performed usꢀ

ing the PC GAMESS (6.4) version* of the GAMESS quantum

1ꢀMethylꢀ2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ5ꢀone (3а).

Sodium hydride (0.25 g, 11 mmol) (from 0.42 g of a 60% suspenꢀ

sion of NaH in Nujol) was added by portions with stirring for

* Alex A. Granovsky, http://www.classic.chem.msu.su/gran/

gamess/index.html.

918

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

Morkovnik et al.

10 min to a suspension of quinazolinone 1 (1.87 g, 10 mmol) in

anhydrous DMF (30 mL), and then a solution of DMS (0.95 mL,

10 mmol) in DMF (2 mL) was added. The mixture was stored

for 30 min at 25 °С and cooled. Water (100 mL) was added, and

methylꢀsubstituted derivative 3a was filtered off. The yield

was 1.45 g (78%). Found (%): С, 65.93; H, 5.64; N, 20.75.

C₁₁H₁₁N₃O. Calculated (%): C, 65.66; H, 5.51; N, 20.88.

Methylation of 2,3ꢀdihydroimidazo[2,1ꢀb]quinazolinꢀ1(10)Hꢀ

5ꢀone (1). A solution of quinazolinone 1 (1.87 g, 10 mmol) and

DMS (0.95 mL, 10 mmol) in MeNO₂(20 mL) were refluxed for

12 h. After the solvent was distilled off, the residue of salts of

compounds 2а and 3а was treated with a 10% solution of

NH₄ОH, and the precipitate was filtered off. The yield of a

mixture of Nꢀmethylꢀsubstituted derivatives 2а and 3а was

1.85 g (92%), and the ratio of isomers was 2а : 3а = 13 : 1. After

double recrystallization from ethyl acetate, the yield of 10ꢀmeꢀ

thylꢀsubstituted isomer 2а was 1.5 g (75%), m.p. 214—215 °С.¹

A sample of 2a mixed with the authentic sample melts without

depression.

Found (%): С, 69.89; H, 5.87; N, 14.02. C₁₈H₁₇N₃O₂. Calcuꢀ

lated (%): C, 70.34; H, 5.58; N, 13.67.

B. The reaction without solvent was carried out as a similar

process with pꢀchlorobenzyl bromide (entry C) but at 150 °С.

The yield of a mixture of 2d and 3d was 68%, and the ratio was

2d : 3d ≈ 1.5 : 1.

C. The reaction in the presence of NaH was carried out as

described above for a similar process involving pꢀchlorobenzyl

bromide (entry D). The yield of 1ꢀ(βꢀphenoxyethyl)ꢀsubstituted

derivative 3d after recrystallization from an iꢀC₈H₁₈—PrⁱOH

(10 : 1) mixture was 73%. Found (%): С, 70.00; H, 5.74;

N, 13.73. C₁₈H₁₇N₃O₂. Calculated (%): C, 70.34; H, 5.58;

N,13.67.

10ꢀ[β ꢀ(4ꢀBromophenoxy)ethyl]ꢀ2,3ꢀdihydroimidꢀ

azo[2,1ꢀb]quinazolinꢀ5ꢀone (2e) was synthesized from quinꢀ

azolinone 1 and βꢀ(4ꢀbromophenoxy)ethyl bromide by refluxꢀ

ing in MeNO₂similarly to methylꢀsubstituted derivative 2а

in 74% yield. Found (%): C, 55.56; H, 4.02; Br, 20.89.

C₁₈H₁₆BrN₃O₂. Calculated (%): C, 55.97; H, 4.18; Br, 20.69.

pꢀChlorobenzylation of quinazolinone 1. A. A solution of comꢀ

pound 1 (1.87 g, 10 mmol) and pꢀchlorobenzyl bromide (2.05 g,

10 mmol) in MeNO₂(25 mL) was refluxed for 14 h. The solvent

was evaporated, the residue was treated with a 10% solution of

NH₄ОH, and a mixture of pꢀchlorobenzylꢀsubstituted derivaꢀ

tives 2c and 3c was separated from an admixture of the starting

quinazolinone 1 by chromatography on a column with Al₂O₃

(eluent CHCl₃), collecting the fraction with R_f0.4. The yield of

a mixture of the isomers was 2.1 g (67%); ratio 2с : 3с ≈ 2 : 1.

B. A solution of compound 1 (1.87 g, 10 mmol) and

pꢀchlorobenzyl bromide (2.05 g, 10 mmol) in DMF (100 mL)

was refluxed for 3 h, and a mixture of 10ꢀ and 1ꢀpꢀchlorobenzylꢀ

substituted derivatives 2c and 3c in a ratio of ~1.5 : 1 (0.8 g,

25%) was isolated by a standard procedure.

1ꢀ[β ꢀ(4ꢀBromophenoxy)ethyl]ꢀ2, 3ꢀdihydroimidꢀ

azo[2,1ꢀb]quinazolinꢀ5ꢀone (3e) was synthesized from quinꢀ

azolinone 1 and βꢀ(4ꢀbromophenoxy)ethyl bromide in DMF in

the presence of NaH similarly to derivative 3с in 82% yield.

Found (%): С, 56.07; H, 3.91; Br, 20.77. C₁₈H₁₆BrN₃O₂. Calꢀ

culated (%): C, 55.97; H, 4.18; Br, 20.69.

This work was financially supported by the Rusꢀ

sian Foundation for Basic Research and the Administraꢀ

tion of the RostovꢀonꢀDon Region (Project No. 04ꢀ03ꢀ

96804).

References

C. A mixture of quinazolinone 1 (1.87 g, 10 mmol) and

pꢀchlorobenzyl bromide (2.05 g, 10 mmol) were melted together

for 1 h at 140 °С. After cooling, the resulting melt was triturated

with Et₂O (20 mL), a precipitate of hydrobromides of comꢀ

pounds 2c and 3c was filtered off, and 1ꢀsubstituted derivative 3c

(2.6 g, 84%) was isolated as described for entry A; an admixꢀ

ture of 10ꢀisomer 2c was 5%. After recrystallization from a

PrⁱOH—C₆H₁₄(1 : 10) mixture, the yield of 1ꢀ(pꢀchloroꢀ

benzyl)quinazolone 3с was 2.0 g (65%). Found (%): С, 65.43;

H, 4.77; N, 14.04. С₁₇H₁₄ClN₃O. Calculated (%): С, 65.49;

H, 5.30; N, 13.48.

1. G. H. Hardtmann, G. Koletar, and O. R. Pfister, J. Med.

Chem., 1975, 18, 447.

2. G. H. Hardtmann, US Pat. 3833652; http://

www.chemweb.com/databases/patents/.

3. G. H. Hardtmann, US Pat. 3912731; http://

www.chemweb.com/databases/patents/.

4. G. H. Hardtmann, US Pat. 3919210; http://

www.chemweb.com/databases/patents/.

5. G. H. Hardtmann, US Pat. 3969506; http://

www.chemweb.com/databases/patents/.

D. Sodium hydride (0.25 g, 11 mmol) (from 0.42 g of a 60%

suspension of NaH in Nujol) was added by portions with stirring

for 10 min to a suspension of quinazolinone 1 (1.87 g, 10 mmol)

in anhydrous DMF (30 mL), and then pꢀchlorobenzyl bromide

(2.05 g, 10 mmol) was added. The mixture was stored for 30 min

at 25 °С and for 15 min at 40 °С, cooled, and treated with water

(100 mL). A precipitate of 1ꢀ(pꢀchlorobenzyl)quinazolinone 3c

was filtered off. The yield was 2.6 g (84%). A sample of 3c mixed

with the sample obtained in entry C shows no depression of the

melting point.

6. N. P. Peet and Sh. Sunder, US Pat. 4871732; http://

www.chemweb.com/databases/patents/.

7. G. H. Hardtmann, US Pat. 3598823; http://

www.chemweb.com/databases/patents/

8. T. Jen, B. Dienel, H. Bowman, J. Petta, A. Helt, and B. Loev,

J. Med. Chem., 1972, 15, 727.

9. E. Takeuchi and T. Sato, Jpn Pat. 61115083; http://

v3.espacenet.com.

10. M. O. Sinnokrot and C. D. Sherrill, J. Am. Chem. Soc., 2004,

126, 7690.

βꢀPhenoxyethylation of quinazolinone 1. A. A solution of

quinazolinone 1 (1.87 g, 10 mmol) and βꢀphenoxyethyl bromide

(2.05 g, 10 mmol) in MeNO₂(25 mL) was refluxed for 14 h.

A mixture of isomers 2d and 3d was isolated as in the case of

pꢀchlorobenzyl bromide in entry A. The yield was 2.06 g

(67%). The ratio was 2d : 3d ≈ 12 : 1. After recrystallization

from methanol, pure 10ꢀsubstituted isomer 2d was obtained.

11. H.ꢀS. Lee, D.ꢀY. Cheong, S.ꢀD. Yoh, W.ꢀS. Kim, Y.ꢀW.

Kwak, Y.ꢀT. Park, and J.ꢀK. Lee, Bull. Korean Chem. Soc.,

2001, 22, 633.

12. G. S. Hammond, J. Am. Chem. Soc., 1955, 77, 334.

13. T. N. Truong, T.ꢀT. T. Truong, and E. V. Stefanovich,

J. Chem. Phys., 1997, 107, 1881.

14. J. Gao and X. Xia, J. Am. Chem. Soc., 1993, 115, 9667.

Cryptoanionic mechanism of Nꢀsubstitution

Russ.Chem.Bull., Int.Ed., Vol. 55, No. 5, May, 2006

919

15. S. P. Webb and M. S. Gordon, J. Phys. Chem. A, 1999,

103, 1265.

16. M. Sola, A. Lledos, M. Duran, J. Bertran, and J. M. Abboud,

J. Am. Chem. Soc., 1991, 113, 2873.

17. J. March, Advaced Organic Chemistry, Reactions, Mechaꢀ

nisms and Structure, WileyꢀInterscience Publication, New

York, 1985.

18. D. K. Bohme and A. B. Raksit, J. Am. Chem. Soc., 1984,

106, 3447.

19. E. Humeres, R. J. Nunes, V. G. Machado, M. D. G.

Gasques, and C. Machado, J. Org. Chem., 2001, 66, 1163.

20. K. C. Westaway, Y. Gao, and Y. Fang, J. Org. Chem., 2003,

68, 3084.

21. M. L. Glowka, A. Olczak, and L. Korzycka, J. Chem.

Crystallogr., 1994, 24, 725.

25. Y. Takata, Y. Huang, J. Komoto, T. Yamada, K. Konishi,

H. Ogawa, T. Gomi, M. Fujioka, and F. Takusagawa, Bioꢀ

chemistry, 2003, 42, 8394.

26. J. Komoto, T. Yamada, Y. Takata, K. Konishi, H. Ogawa,

T. Gomi, M. Fujioka, and F. Takusagawa, Biochemistry,

2004, 43, 14385.

27. P. Velichkova and F. Himo, J. Phys. Chem. B, 2005,

109, 8216.

28. P. Velichkova and F. Himo, J. Phys. Chem. B, 2005, 109;

ASAP Web Release, Date: 07ꢀDecꢀ2005.

29. 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.

30. M. N. Glukhovstev, A. Pross, M. P. McGrath, and

L. Radom, J. Chem. Phys., 1995, 103, 1878.

22. W. J. Le Noble, Synthesis, 1970, 1.

23. K. Hori, J.ꢀL. M. Abboud, C. Lim, M. Fujio, and Y. Tsuno,

J. Org. Chem., 1998, 63, 4228.

31. K. K. Irikura, R. D. Johnson III, and R. N. Kacker, J. Phys.

Chem. A, 2005, 109, 8430.

24. S. D. Yoh, M.ꢀK. Lee, K.ꢀJ. Son, D.ꢀY. Cheong,

I. Han, and K.ꢀT. Shim, Bull. Korean Chem. Soc., 1999,

20, 466.

Received October 21, 2005;

in revised form February 20, 2006

Article Doi

N-alkylation of 2,3-dihydroimidazo[2,1-b]quinazolin-1(10)H-5-one. on the cryptoanionic mechanism of N-substitution

DOI: 10.1007/s11172-006-0351-7

Source and publish data:

Authors:

Article abstract of DOI:10.1007/s11172-006-0351-7

Full text of DOI:10.1007/s11172-006-0351-7

Products guided by the article

R&D Labs maybe for 40534-69-8

Relevant to this article

Hot Product