A novel highly stereoselective synthesis of 2,3-disubstituted 3H-quinazoline-4-one derivatives

Article abstract of DOI:10.1021/ol070276c

Figure presented An efficient three-step synthesis of chiral 3H-quinazoline-4-one derivatives from commercial materials is disclosed. The Mumm reaction of imidoyl chloride with α-amino acids followed by reductive cyclization affords enantiomerically pure (ee >93%) quinazoline-4-ones in good overall yield. A comparison with existing approaches indicates that this method is superior for hindered substrates.

Full text of DOI:10.1021/ol070276c

ORGANIC
LETTERS
2007
Vol. 9, No. 7
1415-1418
A Novel Highly Stereoselective
Synthesis of 2,3-Disubstituted
3H-Quinazoline-4-one Derivatives
Paul Zhichkin,*,† Edward Kesicki,‡ Jennifer Treiberg,‡ Lisa Bourdon,†
Matthew Ronsheim,† Hua Chee Ooi,‡ Stephen White,‡ Angela Judkins,‡ and
David Fairfax†
Albany Molecular Research, Inc., 21 Corporate Circle, P.O. Box 15098, Albany,
New York 12212, and ICOS Corp., 22021 20th AVenue SE, Bothell, Washington 98021
paVelz@albmolecular.com
Received February 2, 2007
ABSTRACT
An efficient three-step synthesis of chiral 3H-quinazoline-4-one derivatives from commercial materials is disclosed. The Mumm reaction of
imidoyl chloride with -amino acids followed by reductive cyclization affords enantiomerically pure (ee >93%) quinazoline-4-ones in good
overall yield. A comparison with existing approaches indicates that this method is superior for hindered substrates.
r
Due to their biological activity, 2,3-disubstituted 3H-
quinazoline-4-ones represent one of the most interesting
groups of heterocycles. In particular, quinazoline-4-one
alkaloids such as asperlicin C, possessing cholecystokinin
antagonist properties, and benzomalvins, which are neuro-
kinin receptor antagonists, as well as other similar molecules,
have attracted significant attention.1 These alkaloids are often
biosynthetically derived from anthranilic acid and chiral
amino acids, and as a result contain a chiral center in the
R-position of the 2-substituent. Another chiral compound
based on the quinazoline-4-one scaffold, a kinesin spindle
protein inhibitor ispenisib, is currently in Phase II clinical
trials for cancer.2 Several chiral quinazoline-4-one derivatives
selectively inhibiting p110δ kinase, whose potential applica-
tions range from autoinflammatory disease to leukemia, have
been reported by ICOS scientists.3
Many examples of the total synthesis of chiral quinazoline-
4-one alkaloids have been reported. The closure of the
quinazoline ring while preserving the neighboring chiral
center is a critical step in all of these syntheses. There are
two main methodologies used to effect the cyclization. One
approach, discovered by Mazurkiewicz,4a employs isomer-
ization of 4-imino-4H-3,1-benzoxazines into quinazoline-4-
ones under basic or acidic conditions.4 This transformation
has been used in the stereoselective synthesis of fumiquinazo-
lines A,4g,k B,4g,k C,4j,k E,4j,k F,4f G,4c,d,f H,4j,k I,4g,k and other
(3) (a) Finer, J. T.; Bergnes, G.; Feng, B.; Smith, S. W.; Chabala, J. C.;
Morgans, D. J. WO Patent Application 01/98278, 2001. (b) Sadhu, C.; Dick,
K.; Treiberg, J.; Sowell, S. J.; Kesicki, E. A.; Oliver, A. U.S. Patent
Application 2002/0161014, 2002. (c) Fowler, K. W.; Huang, D.; Kesicki,
E. A.; Ooi, H. C.; Oliver, A. R.; Fuqiang, R.; Treiberg, J. WO Patent
Application 05/113556, 2005.
(4) (a) Mazurkiewicz, R Monatsh. Chem. 1989, 120, 973. (b) Snider, B.
B.; Zeng, H. Heterocycles 2003, 61, 173. (c) Wang, H.; Ganesan, A. J.
Org. Chem. 1998, 63, 2432. (d) He, F.; Snider, B. B. J. Org. Chem. 1999,
64, 1397. (e) Hart, D. J.; Magomedov, N. Tetrahedron Lett. 1999, 40, 5429.
(f) Wang, H.; Ganesan, A. J. Org. Chem. 2000, 65, 1022. (g) Snider, B.
B.; Zeng, H. Org. Lett. 2000, 2, 4103. (h) Witt, A.; Bergman, J. J. Org.
Chem. 2001, 66, 2784. (i) Hart, D. J.; Magomedov, N. A. J. Am. Chem.
Soc. 2001, 123, 5892. (j) Snider, B. B.; Zeng, H. Org. Lett. 2002, 4, 1087.
(k) Snider, B. B.; Zeng, H. J. Org. Chem. 2003, 68, 545. (l) Wang, H.;
Sim, M. M. J. Nat. Prod. 2001, 64, 1497. (m) Wang, H.; Ganesan, A. J.
Comb. Chem. 2000, 2, 186.
† Albany Molecular Researh.
‡ ICOS.
(1) (a) Witt, A.; Bergman, J. Curr. Org. Chem. 2003, 7, 659 and
references cited therein. For the more recent reviews, see: (b) Michael, J.
P. Nat. Prod. Rep. 2004, 21, 350. (c) Michael, J. P. Nat. Prod. Rep. 2005,
22, 627.
(2) (a) Bergnes, G.; Katjusa, B.; Belmont, L. Curr. Top. Med. Chem.
2005, 5, 127. (b) Wood, K.; Bergnes, G. Annu. Rep. Med. Chem. 2004, 39,
173.
10.1021/ol070276c CCC: $37.00
© 2007 American Chemical Society
Published on Web 03/10/2007

fumiquinazoline analogues,4m as well as fiscalin B,4c,d,f
alantrypinone,4e,i glyantrypine,4f circumdatins C4h and F,4h
and verrucines A and B.4l Another widely used cyclization
method is an aza-Wittig cyclization of imines (Eguchi
protocol).5 This protocol has been successfully applied to
the stereoselective synthesis of amauromine and 5-N-
acetylardeemin,6 fumiquinazoline G and its dehydro-deriva-
tive,7 glyantrypine, fumiquinazoline F, and other 2,4-dihydro-
1H-pyrazino[2,1-b]quinazoline-3,6-diones,8 as well as asper-
licin9a and benzomalvin A.9b
Scheme 2. Attempted Preparation of 3a by the Eguchi
Method
As a part of our research on p110δ kinase inhibitors we
found it necessary to prepare a series of 2,3-disubstituted
3H-quinazoline-4-ones possessing a chiral 2-substituent. In
this letter we would like to disclose a novel highly stereo-
selective approach to such molecules. This approach has
shown superior results when directly compared with the other
known cyclization methods on the example of compound
3a.
The benzoxazine approach following the Mazurciewitcz-
Ganesan procedure4b (Scheme 3) afforded 3a with high ee;
however, the overall yield was disappointing (14% from 1).
The difficulty of the formation of the quinazoline ring of
3a is due to the steric hindrance around the 3-nitrogen created
by the o-tolyl substituent, as well as to the fact that another
methyl group, positioned ortho to the carbonyl, prevents the
amide from taking the planar configuration necessary for the
cyclization. As can be expected from the literature on thermal
cyclization,10 the cyclization of 1 and 2 in refluxing toluene
afforded partially epimerized product (Scheme 1):
Scheme 3
Scheme 1
As we were looking for an alternative, more efficient
cyclization method, our attention was attracted by the
reaction of imidoyl chlorides with carboxylic acids (Mumm
rearrangement)11 and the following cyclization of the
resulting imides to 3H-quinazoline-4-ones (Scheme 4).
Our attempts to prepare 3a via the Egichi protocol were
not successful (Scheme 2), as no reaction of amide 4 with
acid chloride 5 was observed in the presence of either
NaHMDS6a or NEt3, DMAP.9a A likely explanation for this
failure is the decreased reactivity of both the amide nitrogen
and the acid chloride carbonyl resulting from their ortho-
substitution.
Scheme 4. Original Mumm synthesis and Its Modification by
Levy and Stephen
(5) (a) Takeuchi, H.; Eguchi, S. Tetrahedron Lett. 1989, 30, 3313. (b)
Takeuchi, H.; Hagivara, S.; Eguchi, S. Tetrahedron 1989, 45, 6375.
(6) (a) Marsden, S. P.; Depew, K. M.; Danishefsky, S. J. J. Am. Chem.
Soc. 1994, 116, 11143. (b) Depew, K. M.; Marsden, S. P.; Zatorska, D.;
Zatorski, A.; Bornmann, W. G.; Danishefsky, S. J. J. Am. Chem. Soc. 1999,
121, 11953.
(7) He, F.; Snider, B. Synlett 1997, 483.
(8) (a) Herna´ndez, F.; Lumetzberger, A.; Avendan˜o, C.; So¨llhuber, M.
Synlett 2001, 1387. (b) Cledera, P.; Avendan˜o, C.; Mene´ndez, J. C. J. Org.
Chem. 2000, 65, 1743. (c) Buenadicha, F. L.; Avendan˜o, C.; So¨llhuber, M.
Tetrahedron: Asymmetry 2001, 12, 3019.
(9) (a) He, F.; Foxman, B. M.; Snider, B. B. J. Am. Chem. Soc. 1998,
120, 6417. (b) Sugimori, T.; Okawa, T.; Eguchi, S.; Kakehu, A.; Yashima,
E.; Okamoto, Y. Tetrahedron 1998, 54, 7997.
(10) Significant enantiomeric erosion was observed during the microwave
synthesis of fumiquinazoline F: Li, J.-F.; Ye, P.; Zhang, B.; Bi, G.; Sargent,
K.; Yu, L.; Yohannes, D.; Baldino, C. M. J. Org. Chem. 2005, 70, 6339.
The prototype of this cyclization was also discovered by
Mumm.12 Asahina and Ohta,13a followed by Stephen and
Levy,13b incorporated a reduction of the nitro-group as the
final step (Scheme 4), further improving the synthesis.
1416
Org. Lett., Vol. 9, No. 7, 2007

Table 1. Synthesis of Chiral Quinazoline-4-one Derivatives 3
yield of
9 (%)
yield of
11 (%)
yield of
3 (%)
ee of
3 (%)
series
X
R1
R2CO2H
R2
a
b
c
d
e
f
6-Me
6-Me
6-Me
H
H
6-F
H
2-methylphenyl
Ph
3,5-difluorophenyl
Bn
Ph
Ph
89
100
58
84
c
99
97
45
Boc-L-alanine
Boc-L-alanine
Boc-L-alanine
Boc-L-alanine
Boc-L-serine
Boc-D-R-Abu-OH
Boc-L-alanine
Boc-L-alanine
S-CH(Me)NHBoc
S-CH(Me)NHBoc
S-CH(Me)NHBoc
S-CH(Me)NHBoc
R-CH(CH2OH)NHBoc
R-CH(Et)NHBoc
S-CH(Me)NHBoc
S-CH(Me)NHBoc
88
80
63
71
40
55
d,e
d,e
58
60
40
71
57
67
49f
20f
>99a
98a
94a
>99a
94b
>98g
93a
g
h
2,6-difluorophenyl
2,6-difluorophenyl
6-Me
>99a
a Ee was determined by removing the Boc-group with TFA/CH2Cl2. The resulting amine was reacted with racemic and with (R)-1-(1-naphthyl)ethyl
isocyanate, and HPLC traces of these two samples were compared. b Ee was determined by reacting 3e with racemic and with (S)-R-methoxyphenylacetyl
chloride and comparing HPLC traces of these two samples. c Commercially available. d Made via KHMDS procedure. e Intermediate not isolated. f Yield
over two steps. g Ee was determined by chiral HPLC.
Similarly to the azide in the Eguchi protocol, the nitro-
group in the Mumm synthesis serves as the amine equivalent.
The few known examples in the literature suggest that these
two methodologies have similar scope. It remains unclear
to us why the Mumm synthesis has been forgotten14 while
the Eguchi protocol, which employs the potentially dangerous
aryl azide, has gained prominence.
On the basis of this research, we envisioned the synthesis
beginning with the preparation of imidoyl chloride followed
by the reaction with a chiral amino acid (Scheme 5).
of the resulting chiral N-acyl-2-nitrobenzamide can then be
effected under mild conditions avoiding epimerization of the
chiral center.
Reaction of 9a and thionyl chloride afforded imidoyl
chloride 10a in quantitative yield. Because 10a is unstable
to air, it was treated directly with Boc-L-alanine, affording
a good yield of 11a (88%). Out of several reducing agents
tried (H2/Pd, H2/Pd(OH)2, Na2S2O4, Al/NH4Cl, Raney Ni,
NiCl2/NaBH4, SnCl2, Fe/HCl, Fe/NH4Cl, Pd/polymethylhy-
drosiloxane) for the cyclization, zinc powder in acetic acid
gave the best yield of 3a (58%). To our great satisfaction, a
single enantiomer of 3a formed almost exclusively (ee
>99%, see Table 1).
Several other 3H-quinazoline-4-ones with different 5- and
3-substituents have been prepared successfully starting with
corresponding anilides (or benzylamide) of 2-nitrobenzoic
acids and Boc-protected amino acids (Table 1). The yields
on both steps are moderate to good, and the ee of the products
is uniformly high (>93%). Somewhat lower yields in the
case of Boc-L-serine are the result of side reactions of the
hydroxy-group.
Scheme 5. Stereoselective Synthesis of 2,3-Disubstituted
3H-Quinazoline-4-one Derivatives via the Mumm Reaction
If desired, chromatographic purification of the intermediate
imide 11 can be avoided, and a crude reaction mixture can
be used in the subsequent reductive cyclization after a
minimal workup. Conversion of 9a into 3a under these
conditions proceeded in a 43% yield as compared to 51%
for the two-step procedure. For the reason of convenience,
the syntheses of 3g,h were conducted without purification
of the intermediate imine.
Importantly, during the Mumm rearrangement with amino
acids the chirality is preserved.15 The reduction-cyclization
(11) (a) Mumm, O. Chem. Ber. 1910, 43, 886. For the mechanism see:
(b) Brady, K.; Hegarty, A. J. Chem. Soc., Perkin Trans. 2 1980, 121.
(12) Mumm, O.; Hesse, H. Chem. Ber. 1910, 43, 2505. For more
examples see: Spath, E.; Platzer, N. Chem. Ber. 1935, 68, 2221.
(13) (a) Asahina, Y.; Ohta, T. Yakugaku Zasshi 1928, 48, 313; Chem.
Abstr. 1928, 22, 29031. (b) Levy, P. B.; Stephen, H. J. Chem. Soc. 1956,
985.
(14) To the best of our knowledge, the Mumm rearrangement or the
following reductive cyclization has never been used in the synthesis of chiral
3H-quinazoline-4-one derivatives. There are only three other articles that
involve preparation of any 3H-quinazoline-4-ones via reduction of N-acyl-
2-nitrobenzamides, and neither article mentions the original authors. Two
of them utilized impractical CO with heating and pressure as the reducing
agent. (a) Motohiro, A.; Kondo, T.; Watanabe, Y. J. Org. Chem. 1993, 58,
310. (b) Nishiyama, Y.; Hirose, M.; Kitagaito, W. Tetrahedron Lett. 2002,
43, 1855. (c) The third paper, dealing with the synthesis of deoxyvasicinone,
appeared during preparation of this paper: Ziaee, V.; Jalalizadeh, H.;
Iranshahi, M.; Shafiee, A. Iran. J. Chem. Chem. Eng. 2004, 23, 33.
(15) Gruszecki, W.; Gruszecka, M.; Bradaczek, H. Liebigs Ann. Chem.
1988, 331.
Further increase of the hindrance around the aromatic ring
R1 from one ortho-substituent (2-methyl, 3a) to two ortho-
substituents (2,6-difluoro, 3g,h) dramatically diminished the
reactivity of the corresponding imidoyl chlorides. Imidoyl
chlorides 10g and 10h failed to react with Boc-L-alanine
under standard conditions in dichloromethane. Heating the
reaction to 85 °C in toluene or addition of catalytic DMAP
had no effect. To remedy this, we deprotonated 9g,h with
KHMDS and then treated the anion with the succinimide
ester of Boc-L-alanine to give 11g,h (Scheme 6). This
approach was successful (albeit low-yielding) even in the
case of 11h when a total of four ortho-substituents were
Org. Lett., Vol. 9, No. 7, 2007
1417

is present only in the amide anion creates an inherent
advantage of our method over the Eguchi protocol in its
applicability to the hindered substrates.
Scheme 6. Alternative Imide Formation via KHMDS
In conclusion, we have found a novel approach to the
synthesis of chiral 2,3-disubstituted 3H-quinazoline-4-one
derivatives. This method is efficient (3 steps from com-
mercial starting materials), affords materials of high enan-
tiomeric purity (ee >93%), and is, particularly for hindered
substrates, superior to other known methods.
present. For comparison, a similar step in the Eguchi protocol
(reaction of chloride 5 with the anion of 4) failed (see Scheme
2). In the reactions of the amide anion and a neutral
electrophile, hindrance around the anion generally affects
the reaction rate much less than hindrance around a neutral
electrophile. This may explain why the Eguchi protocol is
so sensitive to the o-methyl substitution in the acid chloride
(5). The fact that the hindrance in the reaction in Scheme 6
Supporting Information Available: Experimental pro-
cedures and characterization for compounds 1, 3a-h, 7, 8,
9a-h, 10a,d,e, and 11a-f, as well as procedures for ee
determination for compounds 3a-h. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
OL070276C
1418
Org. Lett., Vol. 9, No. 7, 2007