A study on the conversion of indanones into carbostyrils

Article abstract of DOI:10.1016/S0968-0896(03)00116-0

We have surveyed the utility of Beckmann rearrangement for the conversion of indanones into carbostyrils. Initial attempts at the conversion of 6-methoxy indanone oxime under classical conditions resulted in the formation of the two unusual products: 2-sulfonyloxyindanone and the dimeric product. This unusual rearrangement was also observed by the treatment of some metal triflates species. Further investigation has led to the development of reliable conditions starting from oxime mesylate (not oxime tosylate), in which some strong Lewis acid catalyst (ZrCl₄) was employed in either a conventional or non-conventional solvent system. The advantage of the new protocol is highlighted by the simple work up and direct isolation of the product in 65% isolated yield.

Full text of DOI:10.1016/S0968-0896(03)00116-0

Bioorganic & Medicinal Chemistry 11 (2003) 2205–2209
A Study on the Conversion of Indanones into Carbostyrils
Yasuhiro Torisawa,* Takao Nishi and Jun-ichi Minamikawa
Process Research Laboratory, Second Tokushima Factory, Otsuka Pharmaceutical Co., Ltd.,
Kawauchi, Tokushima, 771-0182, Japan
Received 5 December 2002; accepted 11 February 2003
Abstract—We have surveyed the utility of Beckmann rearrangement for the conversion of indanones into carbostyrils. Initial
attempts at the conversion of 6-methoxy indanone oxime under classical conditions resulted in the formation of the two unusual
products: 2-sulfonyloxyindanone and the dimeric product. This unusual rearrangement was also observed by the treatment of some
metal triflates species. Further investigation has led to the development of reliable conditions starting from oxime mesylate (not
oxime tosylate), in which some strong Lewis acid catalyst (ZrCl ) was employed in either a conventional or non-conventional sol-
4
vent system. The advantage of the new protocol is highlighted by the simple work up and direct isolation of the product in 65%
isolated yield.
#
2003 Elsevier Science Ltd. All rights reserved.
Some years ago, we started a project aimed at elabora-
tion of a new synthetic pathway to the substituted car-
bostyril (3,4-dihydroquinolin-2-one, 2) via Beckmann
rearrangement (BR) of the substituted oxime (1)
possible by the careful selection of the acid catalysts
2
(ZrCl ) in proper solventsys te ms.
4
(
Scheme 1). This new pathway was not only interesting
Reaction with Eaton Reagent and PPA
for the process development for our important drug
substances but also useful for general medicinal chem-
istry research.
A conversion of the dimethoxyindanone oximes into the
isocarbostyril derivatives (undesired isomer) was repor-
ted by the acid catalyst: P O –MeSO H (Eaton
2
5
3
1
a,3
While the BR has found broad application from the
manufacturing process to a variety of laboratory synth-
esis, not many reports have described the practical
synthesis of substituted carbostyrils (CS). One of the
problems might stem from the selectivity of this rear-
rangement.
reagent).
In contrast to this report, reaction of 3 with
Eaton reagent gave a mixture of the two polar spots,
both of which were unlike the desired CS structure.
The major productwas assigned as a-sulfonyloxy deri-
vative (4, 33%) as shown in Scheme 1. Along with 4, a
polar product( 5: 15%) was obtained. Reaction with the
tosylate (6) also furnished the same products in a
slightly higher yield. Product 4 (now 50% from 6) lost
OTs group of the starting material (6).
In the reported example of BR of unsubstituted inda-
none oximes, the products were obtained only in low
yield (ꢀ20% by PPA). On the other hand, in the sub-
stituted indanone oximes, yields varied depending on the
1
substituents.
Careful inspection of another product (5) revealed its
dimeric structure (30% from 6) on the basis of spectro-
scopic analysis. Two sets of C¼O and OMe are present
in one molecule. We have as yetbeen unable ot find
another example of such dimerizartion during BR by
such classical protic reagents.
Our investigation has shed light on the unusual reactiv-
ity of indanone oximes and further on the reactions of
these oxime sulfonates with some metal triflates. Suc-
cessful transformation to carbostyril was finally made
The reaction with Eaton reagent must have an inter-
molecular aspect (reaction with solvent MsOH), since
both products incorporated MsO group absent in the
*Corresponding author. Tel.:+88-665-2126; fax:+88-637-1144;
e-mail: torisawa@fact.otsuka.co.jp
0
968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00116-0

2206
Y. Torisawa et al. / Bioorg. Med. Chem. 11 (2003) 2205–2209
Scheme 2. Survey with Rh catalyst.
To our delight, however, the reaction of the sulfonate
(
(
5
8) with RhCl(PPh3)3 –TfOH in ethylene dichloride
EDC) atreflux ( >80 ) gave a single polar product( 9,
ꢁ
5% isolated yield) containing OTs group migrated in
the 2-position of indanone. The results are obviously
better than the case of Bi(OTf) and Hf(OTf) .
3
4
The 1,3-migration of the sulfonate was very sluggish in
CH Cl , in which typical BR was most frequently car-
2
2
Scheme 1. Initial plan and result.
1
ried out. Reaction without TfOH was somewhat slow.
The preferable amount of TfOH was 2–4 equiv to the
metal catalyst Rh and Pd, so that the active species
mightform in ht e following equilibrium in Scheme 3.
Reaction with TfOH gave a messy mixture.
starting materials. It seems likely that these reactions
proceed through the enamine tautomer, and the impor-
tant aspect is that the OMs group could be introduced
by intermolecular addition with concomitant elimina-
tion of some leaving group on oxime nitrogen.
We further attempted similar transformation with some
methoxy-substituted oxime tosylate (10), and found that
this 1,3-rearrangement to 11 was a general reaction
pathway and independent on the position of a methoxy
group. As a metal catalyst, RhCl(PPh ) and Pd(benzo-
As shown in Scheme 1, the treatment of the oxime (3) wit h
PPA ataround 100 C for 1 h produced the dimer (5) along
ꢁ
with a small amount of indanone and polar materials.
3
3
Same reaction of the tosylate (6) in PPA was very sluggish
(mostly starting material and unidentified fragments).
nitrile) Cl were the most effective among examined
2 2
(with less than 20 mol% catalyst to consume all the
starting material) but RhCl , RuCl and PdCl were
3
3
2
also applicable at higher loading (with more than 20
mol% catalyst). Usually 2–4 equiv of TfOH to metal
catalyst was required. Reaction completed within 0.5 h
with yields 55 (with 8) to 75% (with 10) along with
indanone produced through hydrolysis. There was very
little influence of the small water present, but a large
amount( >1 equiv) of water simply accelerated hydro-
lysis and no BR productwas produced. Itshould finally
be emphasized that the reaction of 10 was much faster
than 8. Generally, reactions with oxime tosylates were
faster than those with oxime mesylates.
Reaction with Transition Metal Catalysts: Mechanistic
Implication
The unusual reactivity under BR conditions described
above was intriguing from the mechanistic point of
view. Mechanistic speculation strongly suggested the
isomerization of oxime double bond to enamine tauto-
mer, and transition metal catalyst was most likely to
catalyze such isomerization. Our attention thus focused
on the interaction of oxime double bond with other
H-metal species (H-M-X), in which the metal is a tran-
sition metal.
Another Rh catalyst system has been successfully
employed for the catalytic Beckmann rearrangement of
4
As summarized in Scheme 2, indanone oxime (7) was
intact with such catalyst as Pd(OAc) , RhCl(PPh ) ,
acyclic oximes.
2
3 3
NiCl (PPh ) in AcOH or MeSO H as with typical
2
3 2
3
Lewis acids. Only observable phenomena were the E, Z
isomerization of oxime OH and hydrolysis to indanone.
We were then interested in the reaction of transition
metal triflate (MOTf) formed in TfOH. In our previous
experiments, some metal triflates such as Bi(OTf) and
3
Hf(OTf) were actually effective for the conversion of
4
2
indanone oximes butin very low yield. There was no
indication of the use of such transition metal triflate for
oxime–enamine interconversions.
Scheme 3. Reaction with transition metal catalysts.

Y. Torisawa et al. / Bioorg. Med. Chem. 11 (2003) 2205–2209
2207
ꢁ
Although we could neither propose a plausible reaction
mechanism nor isolate any pivotal intermediates in our
studies, we are speculating the interaction between
oxime C¼N bond and metal (or acid) catalyst to pro-
mote isomerization of the C¼N bond (from exo to
endo) affording enamine intermediate for the anomalous
reaction pathway described above.
was done at0 C to rt. with exothermic way in EDC.
Simple extractive isolation by AcOEt gave a crude pro-
duct, which was easily triturated by AcOEt–hexane (or
ethylal, see Experimental) to afford a powdered material
(14a) as a firstcrop. Fur ht er amountof 14a was recov-
ered from the mother liquor.
Reaction with isomeric 15 was also successful to furnish
the corresponding carbostyril. Thus we have established
a preferable pathway to carbostyrils by utilizing oxime
mesylate derivatives of indanones, that is superior to the
previous protocol with oxime tosylate, which required a
Reaction with Lewis-Acid Catalysts
After these abortive investigations, we have been inter-
ested in the BR promoted by strong metal halide cata-
lysts. Although preliminary reactions of 3 with some
7
tedious chromatographic separation.
5
Lewis acids in CH Cl or a solvent-free system led to a
2
To attain more easy and effective work-up, we devel-
oped a new setof Lewis acid and solventcombina ti on,
in which BF OEt was utilized as a solvent in place of
2
mixture through decomposition into unidentified polar
products, some literatures precedent indicated the use of
3
2
6
AlCl in this stubborn rearrangement.
3
EDC. As detailed in Experimental, the new protocol
without conventional halogenated solvent (EDC) made
it possible to isolate the crude product by simple dilu-
tion with water. As a first crop, we isolated 14a in up to
65% yield, which is superior than the reaction in EDC.
This new protocol does not require any extractive oper-
ation and amenable to scalable synthesis.
A clue to our pursuit was finally found in the report by
Chi, who developed optimized reaction conditions for
the indanone oxime tosylate (8) with AlCl in CH Cl
2
3
2
ꢁ
7
under careful temperature control (ꢂ40 t o rt ).This
protocol was reproducible in our hands (as summarized
in Scheme 4). Fortunately, we found that other Lewis
acids (TiCl , ZrCl and HfCl ) were also able to catalyze
Thus, we have improved the original Chi protocol by
employing the mesylate (13b) as a substrate with ZrCl4
as a key Lewis acid catalyst. Advantage of the Zr cata-
lysts can be emphasized again in that the reaction was
carried out at rt without any cryogenic conditions and
no chromatographic purification was required.
4
4
4
the reaction. We then attempted to modify Chi’s origi-
nal protocol so as to obtain the product without any
chromatographic purification.
We focused on the 4-methoxy series (13) after the suc-
cessful conversion with 12, particularly in relation to the
8
study for OPC-14523. Fortunately, it was found that
As shown in Scheme 5, the novel combination with
ZrCl4 in BF OEt in place of conventional solvent
(EDC) was very useful for further process optimization.
the reaction of the oxime mesylate (13b) was more facile
than the corresponding tosylates (8) towards various
Lewis acid catalysts in CH Cl . For example, the reac-
3
2
2
2
tion of 13b with excess FeCl or SnCl gave nearly 50%
3
4
yield of the carbostyril (14), while the corresponding
tosylate (8) gave almost recovery of the starting oxime
under identical conditions. We then examined various
solvent systems other than CH Cl , and found that eth-
2
2
ylene dichloride (EDC) was a superior one to carry out
ꢁ
the reaction at around 5–10 C.
In order to prevent exothermic reaction at the addition of
AlCl , we utilized ZrCl for the key catalyst. Reaction
3
4
Scheme 4. Reaction with Lewis acids.
Scheme 5. Survey with oxime mesylates.

2208
Y. Torisawa et al. / Bioorg. Med. Chem. 11 (2003) 2205–2209
Conclusion
We have demonstrated that AlCl and ZrCl were the
From the filtrate, a further amount of 14a (15%) was
obtained by a column chromatography (CC) separation
from the polar isomer isoCS (ca. 5%). A part of iso-CS
was actually elusive in aqueous layer in these work up
operations. The average ratio of carbostyril:isomer was
usually ꢀ5:1 in the extractive residue.
3
4
key reagents for the Beckmann rearrangement of inda-
none oximes. The final conclusion we have shown here
7
is virtually same as those reported by Chi and others.
However we had a chance to uncover some novel reac-
tivities of indanone derivatives under classical BR con-
ditions. In our survey, oximes (3, 6, 10) were all sluggish
towards conventional BR conditions. They were pre-
sumably prone to isomerize, by the action of an acid,
and the unexpected dimer (5) was formed through the
intermediacy of 4. For the intramolecular migration of
the oxime sulfonate (10) t o11, we developed new effec-
tive reagents generated from common Rh or Pd catalyst
under the influence of TfOH.
If the stirring was inefficient, a large amount of inda-
none was produced through simple hydrolysis. Addition
of toluene as a co-solvent was helpful to obtain less
colored mixture. Reaction in CH Cl was also success-
2
2
ꢁ
ful at5–10 C, which gave a dark colored crop (14a) in
less than 40% yield.
BR by ZrCl4 in BF OEt . The pure oxime mesylate
3
2
powder (13a, recrystallized, 17.0 g, 66.6 mmol) was
stirred in the mixture of BF OEt (distilled, 60 mL) and
3
2
ꢁ
a firstshotof ZrCl 4 powder (20 g initially) was delivered
All of these results simply indicated that the selection of
acid catalyst is crucial for the conversion of indanone
into carbostyrils. The new protocol of carefully opti-
mized conditions using BF OEt as solventc aa tlyst
H O (0.5 mL) and keptat10 C by water-ice bath, while
2
ꢁ
for 15 min, before the second shot of ZrCl (20 g) in one
in one portion. The mixture was kept stirring at 10 C
3
2
4
was superior to the classical system. It offers an
optional synthetic pathway for carbostyril core of
OPC-14523 and Aripiprazole, provided that the cat-
alyst-solvent combination is further optimized and
starting indanone is obtainable in a reasonable cost as
a starting material.
portion. After stirring 1.5 hr without cooling at rt, the
resulting mixture was slowly poured into ice-water (1.5
L) and stirred at rt for further 2 h. The mixture was then
kept standing overnight to settle the product and fur-
ther diluted with cold water, before filtration.. Thus the
8
9
obtained solid product was washed with H O to afford
2
the first crop of the carbostyril (14a, 65% after drying at
7
rt) as a yellow powder.
Experimental
From the filtrate, a further amount of 14a (10%) was
obtained by a CC separation from the polar isomer
isoCS (ca. 10%).
General comments. Chemicals were purchased from the
commercial firms indicated and used without further
purification except for the case of brief drying with MS-
BR by AlCl . The oxime mesylate (13a, 4.5 g) was dis-
3
solved well in dry EDC (90 mL) and kepts it rring vig-
orously under ice-bath cooling (5 C). The firstshotof
ꢁ
AlCl (Wako, 9.0 g) was delivered in one portion and
4
AlCl ] prior to use. TLC analysis was carried out using
A (solvents) or crushing into fine powder [ZrCl and
4
3
3
1
13
ꢁ
further amount of AlCl (0.5 g) was delivered to ensure
Merck silica gel 60 F254 plate (Art 5715). H- and
C
the resulting mixture was stirred at 5–25 C for 0.5 h. A
NMR spectra were measured at 300 and 75 MHz,
respectively, with tetramethylsilane (TMS) as the inter-
nal standard. The products of BR reaction were char-
3
the reaction. The resulting mixture was quenched after
TLC analysis and worked up as above. The crude pro-
duct was triturated from AcOEt–n-hexane in the cold to
give nearly pure carbostyril (14a, 1.5 g, 49%) as a first
7
acterized by the comparison with the reported deta.
BR by ZrCl in EDC. The oxime mesylate (13a) was
4
prepared from the oxime in a conventional way through
the treatment of MsCl and Et N in CH Cl under cool-
ing. The pure oxime mesylate powder (13a, recrys-
crop.. Reaction of the tosylate (16) with AlCl in EDC
3
also gave similar resul ts butwe could notob ta in a good
solid product from the extractive residue as in the case
of the mesylate (13a).
3
2
2
tallized, 17.0 g, 66.6 mmole) was dissolved in dry EDC
(
ꢁ
this was added a first shot of ZrCl powder (Merck, 20 g
dry 250 mL) and keptat <10 C by water-ice bath. To
4
initially, 85 mmole) in one portion. A yellow mixture
ꢁ
Acknowledgements
was kepts it rring at10
C for 10 min before the second
shotof ZrCl 4 (20 g, 85 mmol) in one portion. After
stirring 1 h with water bath temperature, the resulting
yellow-brown mixture was slowly poured into ice-water
We thank Mr. I. Miura in our laboratory for his helpful
assistance in NMR assignment. We also thank Pro-
fessor D. Y. Chi (Inha University) for the helpful dis-
cussion in manuscriptprepara ti on and encouragemen t.
(
extracted from AcOEt after neutralization by Na CO
1 l) and stirred at rt for 1 h. The mixture was then
2
3
powder (10 g). Crude extracts were filtered through a
short pad of Celite to remove polar materials. Resulting
filtrate was then concentrated to give crude products,
which were triturated by ethylal (TCI, 400 mL) to afford
the first crop of the carbostyril (14a, 5.7 g, 48%) as a
yellow powder.
References and Notes
1
1
. For recentreviews, see: (a) Gawley, R. E. Organic Reaction
988, 35, 1 and references cited therein. (b) Maruoka, K.;
Yamamoto, H. In Comp. Org. Synth; Trost, B. M., Fleming,
7
I., Eds.; Pergamon Press: Oxford, 1991; p 763.

Y. Torisawa et al. / Bioorg. Med. Chem. 11 (2003) 2205–2209
2209
2. Preliminary results have already been published as a short
communication: Torisawa, Y.; Nishi, T.; Minamikawa, J.
Bioorg. Med. Chem. Lett. 2002, 387.
6. Tanga, M. J.; Reist, E. J. Heterocyclic Chem. 1986, 23, 747
and references cited therein.
7. Lee, B. S.; Chu, S.; Lee, I. Y.; Lee, B.-S.; Song, C. E.; Chi,
D. Y. Bull. Kor. Chem. Soc. 2000, 21, 860.
8. Oshiro, Y.; Sakurai, Y.; Sato, S.; Kurahashi, N.; Tanaka,
T.; Kikuchi, T.; Tottori, K.; Uwahodo, Y.; Miwa, T.; Nishi,
T. J. Med. Chem. 2000, 43, 177.
3
3
4
5
. Eaton, P. E.; Carlson, G. R.; Lee, J. T. J. Org. Chem. 1973,
8, 4071.
. Arisawa, M.; Yamaguchi, M. Org. Lett. 2001, 3, 311.
. (a) Khodaei, M. M.; Meybodi, F. A.; Rezai, N.; Salehi, P.
Synth. Commun. 2001, 31, 2047. (b) Thakur, A. J.; Boruah, A.;
Prajapati, D.; Sandhu, J. S. Synth. Commun. 2000, 30, 2105.
9. Oshiro, Y.; Sato, S.; Kurahashi, N.; Tanaka, T.; Kikuchi, T.;
Tottori, K.; Uwahodo, Y.; Nishi, T. J. Med .Chem. 1998, 41, 658.