Parallel kinetic resolution of racemic amines using
3-N,N-diacylaminoquinazolin-4(3H)-ones
Abdullah G. Al-Sehemi, Robert S. Atkinson and Christopher K. Meades
Department of Chemistry, Leicester University, Leicester, UK LE1 7RH
Received (in Cambridge, UK) 11th October 2001, Accepted 9th November 2001
First published as an Advance Article on the web 4th December 2001
The two pseudoenantiotopic N-acyl groups of enantiopure
3-[N-(3-m-acetoxyphenylpropanoyl)-N-(3-phenylpropanoy-
l)amine]-2-isopropylquinazolin-4(3H)-one 7 each react with
a different enantiomer of 2-methylpiperidine giving rise to
efficient parallel kinetic resolution (ee > 95%).
acyl groups6 i.e. the two N-acyl groups in DAQ1 5 were playing
the rôle of the two pseudoenantiomeric reagents in a PKR (see
above).3
To maximise the opportunity for PKR to be obtained in these
amine acylations, we elected to use enantiopure DAQ1 7 which
contains the N–N axis as the only chiral element. The
isoenergetic transition states (TSs) in which each amine
enantiomer reacts with its respective pseudoenantiotopic N–
acyl group can be represented as in Scheme 2(a) and (b).
Although the sites occupied by S, M and L (small, medium,
large) groups of the amine are arbitrarily assigned, Scheme 2
incorporates the expected exo/endo"endo/exo conformational
equilibrium for the imide moiety of DAQ1 74 and our
conclusion is that attack of the amine takes place on the face of
the exo-oriented imide carbonyl group syn to the quinazolinone
carbonyl group (exo+cis to Q).7 Clearly it is important for the
success of the PKR that the ratio of these two imide
conformations present is 1+1 and remains so throughout the
reaction.
In a normal kinetic resolution (KR) the enantiopure reagent
reacts preferentially with one enantiomer of the racemic
substrate. Under conditions of stoichiometry (2 eq. amine: 1 eq.
reagent) this enantiopreference of the reagent must accom-
modate the diminishing concentration of the more reactive to
the less reactive enantiomer as the reaction proceeds. Although
high ees are possible under these conditions using enzymes1 or
chemzymes,2 more often in practice the substrate is used in
excess and only the derivatised enantiomer is isolated.
Vedejs has shown that different enantiomers of a racemic
substrate can each be reacted at the same rate with pseudoe-
nantiomeric reagents, designed to allow separation of the two
derivatised products (parallel kinetic resolution, PKR).3 The ees
of both these derivatives (88% and 95%) are raised because the
concentrations of both unreacted enantiomers decline at the
same rate throughout the reaction.
We have shown previously that 3-N,N-diacylaminoquinazo-
linones (DAQs) e.g. 1 bearing a chiral centre and different N-
acyl groups are separable into diastereoisomers because the N–
N bond is a chiral axis.4 DAQ1 1 (Scheme 1) was obtained by
reaction of 3-N-benzoylamino-quinazolinone (MAQ1) 4 with
2-methylpropanoyl chloride–pyridine and the oily and crystal-
line diastereoisomers separated by chromatography.5
The synthesis of DAQ1 7 as a single enantiomer was
accomplished as in Scheme 3.
DAQ2 6 is chemospecifically attacked by amines at the
cinnamoyl carbonyl group.‡ After two kinetic resolutions using
(R)-2-methylpiperidine (2 eq. DAQ: 1 eq. amine), the recovered
DAQ2 6 had [a]D 102 (c = 1, CH2Cl2) and showed a single
peak on a Chiralcel column under conditions where racemic
material showed two peaks. Catalytic hydrogenation gave a
sample of DAQ 7 of unknown configuration [a]D 1 (c = 1.0)
but believed to be enantiopure.
Both these enantiopure diastereoisomers of DAQ 1 react
chemospecifically with 2-methylpiperidine to give the N-
benzoylamide 3; both also react enantioselectively with this
amine even under conditions of stoichiometry (1 eq. DAQ: 2 eq.
amine). Since the oily and crystalline diastereoisomers differ in
configuration at their N–N axis and produce (R)-amide 3 (95%
ee) and (S)-amide 3 (82% ee), respectively, it is the configura-
tion of the N–N axis which controls the sense of enantiose-
lectivity.
Reaction of DAQ2 7 (1 eq.) with racemic 2-methylpiperidine
(1 eq.) was carried out at 5 °C for 48 h. Separation of the crude
amide 8 and MAQ2 9 which both contain the phenyl ring from
their m-acetoxyphenyl ring-containing analogues 10 and 11
(Scheme 4) was accomplished by chromatography.
Efficient separation of the amides 8 and 10 from each of their
respective MAQ2s was then accomplished by trituration with
The corresponding DAQ1 5 was obtained completely diaster-
eoselectively by reaction of MAQ1 4 with acetyl chloride-
pyridine.† 2-Methylpiperidine reacts with DAQ1 5 at both the
benzoyl and acetyl groups (ratio 2.5+1); from the high ee of the
(S)-3 amide and from the low specific rotation of the unreacted
2-methylpiperidine recovered, it appeared that the two amine
enantiomers were each reacting preferentially with different N-
Scheme 2
Scheme 3 Reagents and conditions: i, PhCHNCHCOCl, pyr. (81%); ii, m-
acetoxyphenylpropanoylchloride, CH2Cl2, pyr. DMF (40%); iii, 6 (2 eq.),
(R)-2-methylpiperidine (1 eq.), 215 °C, 24 h (KR); repeat; iv, Pd/C, H2,
EtOAc (90%).
Scheme 1
2684
Chem. Commun., 2001, 2684–2685
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b109297a