An anion-induced regio- and chemoselective acylation and its application to the synthesis of an anticancer agent

Article abstract of DOI:10.1021/ol016809d

Figure presented An efficient Grignard- and organolithium-induced regio- and chemoselective anionic acylation is reported. A number of tricyclic ketones are prepared in good to excellent yields via this method. This method is complementary to the Frieldel-Crafts acylation for electron-deficient substrates. A novel anisole-based Grignard reagent was developed to effect the cyclization of sterically hindered substrates. This novel reagent has been successfully applied to the synthesis of Sch 66336, a candidate for oncologic treatment.

Full text of DOI:10.1021/ol016809d

ORGANIC
LETTERS
2001
Vol. 3, No. 23
3795-3798
An Anion-Induced Regio- and
Chemoselective Acylation and Its
Application to the Synthesis of an
Anticancer Agent
Marc Poirier, Frank Chen, Charles Bernard, Yee-Shing Wong, and George G. Wu*
Chemical Process Research and DeVelopment, Schering-Plough Research Institute,
1011 Morris AVenue, Union, New Jersey 07083
george.wu@spcorp.com
Received September 24, 2001
ABSTRACT
An efficient Grignard- and organolithium-induced regio- and chemoselective anionic acylation is reported. A number of tricyclic ketones are
prepared in good to excellent yields via this method. This method is complementary to the Frieldel−Crafts acylation for electron-deficient
substrates. A novel anisole-based Grignard reagent was developed to effect the cyclization of sterically hindered substrates. This novel
reagent has been successfully applied to the synthesis of Sch 66336, a candidate for oncologic treatment.
Sch 66336 is a potent farnesyl protein transferase inhibitor1-2
and is currently in clinical trials for the treatment of several
types of cancer. One of its synthetic challenges is the
introduction of both the 3- and 10-bromo substituents in a
convergent fashion. The initial synthesis3 of Sch 66336
requires a total of seven steps for the sequential introduction
of both the 3- and 10-bromo groups starting from the
8-chloroazaketone. These steps are nitration, reduction, and
bromination for the introduction of the 3-bromo and nitration,
reduction, bromination, and deamination for the 10-bromo
group. For the convergent introduction of the 3-bromo group,
we have developed and scaled up successfully an efficient
palladium-catalyzed regioselective carbonylation.4 Recent
reports5 on Grignard-induced cyclizations prompted us to
disclose our results on the regioselective anionic acylation
for the preparation of the 10-bromo-substituted azaketone.6
The synthetic strategy was to dissect the double benzylic
chiral center and divide the molecule into a 3,10-dibromo-
8-chloroazaketone, 1, and a lower part as shown in Figure
1. The azaketone 1 could be prepared from either a direct
acylation of 2a or a directed regioselective cyclization of
2b. As expected, cyclization of 2a under Friedel-Crafts
conditions generated an inseparable 1:1 mixture of two
Figure 1. Retrosynthetic analysis.
(1) Barbacid, M, Annu. ReV. Biochem. 1987, 56, 779.
(2) Leonard, D. M. J Med. Chem. 1997, 40, 2971.
10.1021/ol016809d CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/20/2001

regioisomers, the 8-chloro-10-bromo and the 8-bromo-10-
chloro azaketones. Thus, we turned our attention to an
anionic “Friedel-Crafts” acylation by preactivating the 11-
position with an iodo group.
with s-BuLi followed by alkylation with 2-bromobenzyl-
bromide gave the alkylated acid. The acid group was then
converted to the corresponding amide 4 via an acid chloride.
The other three examples in Table 1 were prepared starting
with 3-methylpyridinic amides. Both n-BuLi and LDA
worked well for the lateral lithiation of 3-methylpyridinic
amides.9 However, LDA is necessary for the lithiation of
5-bromo-3-methylpyridinic amide 8 because of the presence
of the bromo group. In addition, a binary solvent system of
methyl tert-butyl ether and THF was found to work best for
entry 4 in order to minimize the de-iodonation side reaction.
Since only tertiary amides are suitable for the anion-induced
acylation, the secondary amides derived from 5 and 8 were
further converted in good yields to their corresponding
tertiary amides 6, 7, and 2b, using NaH and MeI. Direct
lithiation/alkylation of tertiary amides afforded the same
products but in much lower yields.
Intramolecular anionic Friedel-Crafts equivalent reactions
were first reported by Snieckus’s group for the formation of
tricyclic ketones.7 The reported acylations are generally
initiated by an amide-induced remote lithiation followed by
an intramolecular addition of the newly formed anion to the
amide group. This type of reaction cannot be applied directly
to our synthesis as there are two nondiscriminative lithiation
sites on the aromatic moiety in compound 2a. Thus, it is
necessary to regioselectively activate the 11-position. We
postulated that a selective iodo-lithium or iodo-magnesium
exchange of the 11-iodo substituent in 2b followed by an
intramolecular cyclization would afford the desired tricyclic
ketone 1. This assumption was supported by three very recent
publications in which halogen-lithium and halogen-
magnesium exchange-induced cyclizations are reported.5 All
these reported examples produced good yields for five- and
six-membered heterocycles but poor yields for seven-
membered rings. Furthermore, those substrates do not contain
any other halogen groups that could compete with the desired
iodo-metal exchange. Therefore, there are two key questions
for our proposed reaction. (1) Can we achieve a selective
activation of the iodo group at the hindered 11-positon? (2)
Can we accomplish an effective subsequent cyclization to
form the seven-membered ketones?
The desired 2-iodo-3-bromo-5-chlorobenzyl bromide, 9,
was prepared in excellent overall yield by following the
procedure in Scheme 1. Thus, bromination of 2-amino-5-
Scheme 1. Preparation of Benzyl Bromide 9
First, we studied this type of acylation with simpler
substrates. These substrates were prepared via a lateral
lithiation followed by alkylation with an appropriate elec-
trophile. Some of the alkylated products prepared are
summarized in Table 1. Lateral lithiation of o-toluic acid8
chlorobenzoic acid followed by diazotization and iodide
displacement afforded 2-iodo-3-bromo acid 10 in 87% yield.
Reduction of the acid group with (MeO)3B and BH3-Me2S10
gave the alcohol 11 in 98% yield. 11 was converted to the
Table 1. Lateral Lithiation-Alkylation
(3) (a) Njoroge, F. G.; Taveras, A. G.; Kelly, J.; Remiszewski, S.;
Mallams, A. K.; Wolin, R.; Afonso, A.; Cooper, A. B.; Rane, D. F.; Liu,
Y.; Wong, J.; Vibulbhan, B.; Pinto, P.; Deskus, J.; Alvarez, C. S.; del
Rosario, J.; Connolly, M.; Wang, J.; Desai, J.; Rossman, R. R.; Bishop, W.
R.; Patton, R.; Wang L.; Kirschmeier, P.; Bryant, M. S.; Nomeir, A. A.;
Lin, C. C.; Liu, M.; McPhail, A.; Doll, R. J.; Girijavallabhan, V.; Ganguly,
A. K. J. Med. Chem. 1998. 41, 4890. (b) Njoroge, F. G.; Vibulbhan, B.;
Wong, J. K.; White, S. K.; Wong, S.; Carruthers, N. I.; Kaminski, J. J.;
Doll, R.; Girijavallabhan, V.; Ganguly, A. K. Org. Lett. 1999, 1, 1371.
(4) Wu, G.; Wong, Y.; Poirier, M. Org. Lett. 1999, 1, 745.
(5) (a) Iida, T.; Wada, T.; Tomimoto, K.; Mase, T. Tetrahedron Lett.
2001, 42, 4841. (b) Kondo, Y.; Asai, M.; Miura, T.; Uchiyama, M.;
Sakamoto, T. Org. Lett. 2001, 3, 13. (c) Inoue, A.; Kitagawa, K.; Shinokubo,
H.; Oshima, K. J. Org. Chem. 2001, 66, 4333.
(6) Part of the contents in this Letter are covered by our U.S. patent:
5,998,620, 1999.
(7) (a) Beaulieu, F.; Snieckus, V. J. Org. Chem. 1994, 59, 6508. (b)
Familoni O. B.; Ionica, I.; Bower, J. F.; Snieckus, V. Synlett 1997, 1081.
(c) MacNeil, S. L.; Gray, M.; Briggs, L. E.; L, J. J.; Snieckus, V. Synlett
1998, 419.
(8) Creger, P. L. J. Am. Chem. Soc. 1970, 92, 1396.
(9) (a) Clark, R. D.; Jahangir. J. Org. Chem. 1987, 52, 5378. (b)
Schumacher, D. P.; Murphy, B. L.; Clark, J. E.; Tahbaz, P.; Mann, T. A.
J. Org. Chem. 1989, 54, 2242.
(10) Lane, C. F.; Myatt, H. L.; Daniels, J.; Hopps, H. B. J. Org. Chem.
1974, 39, 3052.
3796
Org. Lett., Vol. 3, No. 23, 2001

corresponding benzyl bromide 9 in 96% yield by using Br2
and PPh3. The 2-bromo-3-bromomethylpyridine was prepared
via a radical reaction11 starting from 2-bromo-3-methyl-
pyridine.
Second, we studied the halogen-lithium and halogen-
magnesium exchanges and the subsequent cyclization. Simple
Grignard reagents such as i-PrMgBr failed to achieve
halogen-magnesium exchanges for 4, 6, and 7. This result
confirms the previous report that i-PrMgBr works well only
for electron-deficient aryl halides or alkenyl halides bearing
an oxygen-containing functionality acting as a metal directing
group.12 Special reagents such as (n-Bu)3MgLi are required
for simple aryl bromide-magnesium exchanges.5b
For substrates 4, 6, and 7, n-BuLi was found to be a very
effective reagent for halogen-lithium exchanges. The comple-
tion of halogen-metal exchange was monitored by proton
NMR after the reaction mixture was quenched with deute-
rium oxide. It is worth noting that the potential lateral
lithiation of the benzylic positions did not take place under
the reaction conditions. Furthermore, all the lithiated inter-
mediates cyclized smoothly to produce the corresponding
tricyclic ketones 12,13 13, and 14 in 77%, 71%, and 91%
isolated yields, respectively. The formation of novel dipy-
ridine-containing tricyclic ketone 13 (Table 2, entry 2) is
Alkyl- and aryllithium reagents are not suitable for
substrate 2b as there is no useful chemoselectivity between
iodo and bromo groups toward halogen-lithium exchanges.
We then focused our attention on iodo-magnesium ex-
change. Simple Grignard reagents such as i-PrMgBr afforded
a mere 10% of the desired azaketone 1. PhMgBr gave ca.
10% of the product together with an intractable mixture, most
likely due to the competitive addition to the amide group.
Mesityl Grignard, a hindered reagent, was found to perform
the iodo-magnesium exchange for both aryl iodides 7 and
2b. However, the subsequent cyclization proceeded only in
40% solution yields for both substrates. The solution yield
was determined by HPLC using an external purified standard.
The major side product was the proton-quenched starting
material.
To improve the yield, we examined the effect of the
leaving group on the cyclization. A variety of amides were
prepared and subjected to the iodo-magnesium exchange
using mesityl Grignard, and their results are summarized in
Table 3. Although, the exchange proceeded well for all
Table 3. Selection of Amide
Table 2. Anion-Induced Regioselective Acylation
a The ratio was determined by 1H NMR. b The solution yield was
determined by HPLC.
amides, the ratio of cyclized ketone vs the proton-quenched
byproduct (1:2a) varied from 0:100 to 63:37. Aryl amides
gave much better ratios than the ratios of the alkyl ones.
For example, PhNMe and 4-FPhNMe amides gave a 63:37
ratio of 1:2a with 40% solution yields. Arylamines, the
byproduct from the cyclization, stabilize anions better than
alkyl ones and therefore are better leaving groups. Con-
versely, electron-donating groups such as 4-MeO in entry 6
destabilize the anion and therefore decrease both the ratio
and the yield. The reason for the relatively good ratio
obtained with piperazine derivatives (entry 3) was not very
clear. PhNMe amide was selected for screening of different
Grignard reagents because of its good results and easy
preparation.
complementary to the Friedel-Crafts acylation for electron-
deficient substrates. Tricyclic ketone 14 is a key intermediate
for the synthesis of loratadine,9b,14 a marketed allergy drug.
(11) Rebek, J.; Costello, T. Heterocycles 1984, 22, 2191.
(12) For reviews, see: (a) Rottlander, M.; Boymond, L.; Berillon, L.;
Lepretre, A.; Varchi, G.; Avolio, S.; Laaziri, H.; Queguiner, G.; Ricci, A.;
Cahiez, G.; Knochel, P. Chem Eur. J. 2000, 6, 767. (b) Boudier, A.; Bromm,
L. O.; Lotz, M.; Knochel, P. Angew. Chem. 2000, 112, 4584; Angew. Chem.,
Int. Ed. 2000, 39, 4414.
(13) Olah, G. A.; Mathew, T.; Farnia, M.; Prakash, G. K. S. Synlett 1999,
1067.
(14) Medichem S. A. U.S. Patent: 6,084,100, 2001.
Org. Lett., Vol. 3, No. 23, 2001
3797

Finally, we prepared a number of Grignard reagents in
order to improve the cyclization yield. The fact that the
bulkier aryl Grignard gave better results led us to speculate
that a stepwise electron transfer mechanism may be operat-
ing. First, the iodo-Gringard exchange takes place and
transfers only one electron to the hindered 11-position. If
the one-electron species has a relatively long half-life, it will
add to the carbonyl group of the amide, leading to a cyclic
ketone. If the rate of the second electron transfer is faster
than that of the cyclization, an inactive anionic intermediate
will form, resulting in a proton-quenched product. Therefore,
the longer the Grignard holds the magnesium, the longer the
half-life of the one-electron species and the better the
cyclization yield. Evidence was observed previously for one-
electron involvement in the reaction of alkyllithium and aryl
iodide.15 Bulkier Grignard reagents such as mesitylMgBr may
slow the second electron transfer. Introduction of magnesium-
coordinating substituents ortho to the MgBr group should
be more effective in slowing down the second electron-
transfer process.
Table 4. Selection of Grignard Reagent
To test our theory, we prepared 2,4,6-(MeO)3PhMgBr
since it contains two coordination sites ortho to MgBr. To
our delight, this reagent worked exceptionally well for
substrate 2b and produced almost exclusively the cyclized
ketone 1 with a 60% solution yield as shown by entry 3 in
Table 4. The requirement of a long reaction time and excess
reagent indicated a strong bidental coordination of the MgBr
and a slow iodo-magnesium exchange. Next, we prepared
a Grignard reagent (entry 4) with only one coordination site
by replacing one of the MeO groups with a methyl. This
monodental Grignard induced the cyclization smoothly in
only 20 min with a 98:2 ratio of 1:2a and a 65% isolated
yield. We then removed the methyl group and prepared four
additional monodental anisole-based Grignard reagents. All
of the four reagents worked very well for the cyclization as
shown in entries 5, 6, 7, and 8 in Table 4. These results also
suggested that the coordination played a more important role
than the steric hindrance.
In summary, we have developed an efficient anion-induced
regio- and chemoselective acylation. We have also introduced
the anisole-based reagents for the cyclization of sterically
hindered substrates and applied them successfully to the
synthesis of an advanced intermediate for an anticancer agent.
This method is complementary to the Friedel-Crafts acyl-
ation for electron-deficient aromatic rings.
2-MeO-5-MePhMgBr (entries 8 and 9) was selected for
optimization on the basis of its good result and easy
preparation. Further studies revealed that addition of 30%
dioxane improves that solution yield from 66% to 78%. A
simple workup procedure16 was also developed to give 1 in
64-67% isolated yield. With this development, we were able
to establish a convergent synthesis of Sch 66336.
Acknowledgment. We thank the staff of the Analytical
Department for assays, Mr. Zhixing Ding for his help in
preparing some of the starting materials, and Drs. Doris
Schumacher and George Love for their proof-reading of the
manuscript.
Supporting Information Available: Experimental pro-
1
cedures and H and 13C NMR spectra for all compounds.
(15) For a review of the mechanism, see: Bailey, W. F. Patricia, J. J. J.
Organomet. Chem. 1988, 352, 1.
(16) After completion, the reaction mixture was poured into water and
the product was extracted with t-BuOMe. The organic layer was washed
with water and concentrated. Addition of t-BuOMe precipitated 1 as off-
white solid.
This material is available free of charge via Internet at
http://pubs.acs.org.
OL016809D
3798
Org. Lett., Vol. 3, No. 23, 2001