Solvent effects on stereoselectivity in 2,3-dimethyl-4-chromanone cyclization by quinine-catalyzed asymmetric intramolecular oxo-Michael addition

Article abstract of DOI:10.1016/S0957-4166(00)00461-4

Examination of the quinine-catalyzed asymmetric intramolecular oxo-Michael addition of o-tigloylphenol in various solvents led to high stereoselectivity in chromanone cyclization when chlorobenzene was used as a solvent, giving cis-2,3-dimethyl-4-chromano

Full text of DOI:10.1016/S0957-4166(00)00461-4

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 11 (2000) 4633–4637
Solvent effects on stereoselectivity in
2,3-dimethyl-4-chromanone cyclization by quinine-catalyzed
asymmetric intramolecular oxo-Michael addition
Tomohiro Tanaka, Takuya Kumamoto and Tsutomu Ishikawa*
Faculty of Pharmaceutical Sciences, Chiba University, 1-33 Yayoi, Inage, Chiba 263-8522, Japan
Received 10 October 2000; accepted 16 November 2000
Abstract
Examination of the quinine-catalyzed asymmetric intramolecular oxo-Michael addition of o-tigloylphe-
nol in various solvents led to high stereoselectivity in chromanone cyclization when chlorobenzene was
used as a solvent, giving cis-2,3-dimethyl-4-chromanone with 60% de and 98% ee. © 2001 Elsevier Science
Ltd. All rights reserved.
(+)-Calanolide A 1, isolated as a strong anti-HIV-1 active coumarin from Calophyllum
lanigerum var. austrocoriaceum (Guttiferae),1 is presently being examined as a possible candidate
for an AIDS drug at clinical level in USA.2 The (10R,11S,12S) stereochemistry of the D ring in
(+)-calanolide A 1 is suggested to be the most responsible function for anti-HIV-1 activity.1,3 We
have approached the enantioselective construction of the chromanone ring, easily leading to the
corresponding chromanol function by hydride reduction, by intramolecular oxo-Michael addi-
tion (IMA) of an o-tigloylphenol in the presence of chincona alkaloids such as (−)-quinine, and
effective asymmetric induction (up to 87% ee) was achieved in cis-chromanone cyclization in
model studies using a coumarin lacking a 4-propyl group.4 However, not only was there no
diastereoselectivity between cis- and trans-products but also the enantioselectivity in the
trans-chromanone cyclization was poor. Examination of the quinine-catalyzed IMA of 7-
hydroxy-5-methoxy-4-propyl-8-tigloylcoumarin 3 in various solvents toward (+)-calanolide A 1
synthesis led to predominant formation (60% de) of (+)-enantiomer-rich cis-chromanone (98%
ee) when chlorobenzene was used as a solvent. In this communication we present remarkable
solvent effects on both diastereo- and enantioselectivities for a 2,3-dimethyl-4-chromanone ring
construction in the quinine-catalyzed IMA.
* Corresponding author. Tel/fax: +81-43-290-2910; e-mail: benti@p.chiba-u.ac.jp
0957-4166/00/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00461-4

4634
T. Tanaka et al. / Tetrahedron: Asymmetry 11 (2000) 4633–4637
In synthetic trials for the total synthesis of (+)-calanolide A 1 we planned a retrosynthesis
shown in Scheme 1, in which a 2,2-dimethylpyran ring (the ring C) would be prepared after the
IMA of 5-methoxycoumarin derivative 3 because of successful IMA using a structurally related
methoxycoumarin in the model experiments.4
Me
Me
6
Me
O
Me
Me
Me
7
5
MeO
MeO
C
IMA
4
MeO
8
3
A
B
9
O
O
1
O
2
O
O
O
HO
Me
O
O
D
10
Me
12
MeO
O
O
OH
11
Me
O
O
4
Me
(+)-calanolide A (1)
Me
Me
3
2
Scheme 1.
The starting o-tigloylphenol 3 for the IMA was similarly prepared according to the reported
method5 (Scheme 2). Friedel–Crafts reaction of 1,3,5-trimethoxybenzene with butyryl chloride,
selective demethylation, and treatment6 with carbethoxymethylenetriphenylphosphorane
afforded 5,7-dimethoxy-4-propylcoumarin 4. Regioselective introduction7 of a tigloyl function
into 4 followed by demethylation at the 7-methoxy group with magnesium iodide8 yielded the
o-tigloylphenol 3. Furthermore, an isomeric o-angeloylcoumarin9 8 was prepared by the
photoirradiation8 of 3 in acetone using a 400 W mercury lamp.
Scheme 2. Reagents and conditions: (a) butyryl chloride, SnCl4 in CH2Cl2, −15 to −10°C, 2 h (96%); (b) BCl3 in
CH2Cl2, 0–5°C, 2.5 h (85%); (c) Ph3PꢀCHCO2Et in PhNEt2, 215°C, 6.5 h (82%); (d) tigloyl chloride, SnCl4 in CH2Cl2,
0°C, 6 d (60%); (e) MgI2 in PhH, reflux, 1 d (62%); (f) hn (400 W Hg) in acetone, 0°C, 40 min (60%)
We have suggested that ion pair formation in transition states as shown in Fig. 1 could play
an important role for the effective stereoselectivity of cis-chromanone cyclization in the
quinine-catalyzed IMA.4 The use of polarizable solvents would be expected to cause stabilizisa-
tion of the chiral complex. Thus, we examined the IMA in the presence of 10 mol% of
(−)-quinine10 in various solvents (Table 1). According to a model experiment4 o-tigloylphenol 3
was firstly subjected to the IMA in tetrahydrofuran for 7 days to afford a 56:44 mixture of cis-
and trans-chromanones in 87% isolated yield (run 1). Each isomer was separated by preparative
TLC.11 As expected, cis-derivative cis-2 was produced with relatively high enantioselectivity
(77% ee).12 A similar result was obtained when acetone was used as a solvent (run 2). Slight rate
accelerations were observed in the reactions using chloroform or ethanol (runs 3 and 4) and the
use of 2-trifluoroethanol led to large rate acceleration (run 5). However, enantioselectivities were
lowered in these cases.

T. Tanaka et al. / Tetrahedron: Asymmetry 11 (2000) 4633–4637
4635
OMe
O
N
MeO
O
O
O-
H
O
Me
Me
H
N+
Figure 1.
Table 1
Solvent effects on quinine-catalyzed asymmetric IMA of o-tigolylphenol 3 or o-angeloylphenol 8 to chromanone
cyclization
On the other hand, aromatic solvents remarkably affected the IMA (runs 6–14). Interest-
ingly the use of either halobenzenes or anisole greatly improved diastereo- and/or enantioselec-
tivities of the cyclized products, leading the predominant production of cis-chromanone cis-2
in ca. 60% de (runs 8–11). In particular, the use of chlorobenzene resulted in satisfactory
asymmetric induction (98% ee) with smooth chemical conversion (run 10). Similar high
selectivities were observed when trifluoromethylbenzene (run 8) and anisole (run 11) were used

4636
T. Tanaka et al. / Tetrahedron: Asymmetry 11 (2000) 4633–4637
as solvents, indicating that polarizability13 of solvents rather than the electronical character of
substituent(s) on the benzene ring could affect the quinine-catalyzed IMA. Furthermore,
although perfluorobenzene or di(trifluoromethyl)benzenes resulted in improved stereoselectivi-
ties, slow chemical conversions were observed in these cases (runs 12–14). Thus, we concluded
that chlorobenzene would be the most suitable solvent for the quinine-catalyzed IMA.14 The
absolute stereochemistry of an optically active cis-2 with 98% ee obtained in run 10 was
determined to be (8R,9S) configuration by comparison of its [h]D +156.7 (CHCl3) with that of
a model cis-chromanone 9, [h]D +132.5 (CHCl3).4,15
MeO
R
O
O
O
8
Me
O
9
Me
cis-2: R=Pr; 9: R=H
We applied the above IMA in chlorobenzene to o-angeloylphenol 8 as a comparable
experiment. Heating 8 at 50°C for 10 days expectedly afforded a 32:68 mixture of cis- and
trans-2 in 64% yield, in which trans-2 was obtained in 78% ee (run 15).16 Resistance of the
reaction17 in this case may be attributable to the predominant conformation as an s-cis isomer
due to compensation of steric repulsion between the 7-hydroxy group and 3%-methyl one in the
angeloyl residue. These facts indicated that the tigloylphenol 3 was a more suitable substrate
than the angeloylphenol 8 for quinine-catalyzed intramolecular oxo-Michael reaction not only in
the efficiency of the reaction but also in the selectivity of chromanone formation, in spite of
predominant production of undesired cis-2 for (+)-calanolide A 1 synthesis.15
In conclusion, the use of chlorobenzene in the quinine-catalyzed asymmetric IMA led to
satisfactory diastereo- and enantioselectivities in chromanone cyclization. Thus, cis-2,3-
dimethyl-4-chromanone cis-2 could be derived from o-tigloylphenol 3 with 60% de and 98% ee.
Acknowledgements
We thank the Futaba Electron Memorial Foundation for partial financial support of this
work.
References
1. Kashman, Y.; Gustafson, K. R.; Fuller, R. W.; Cardellina II, J. H.; McMahon, J. B.; Currens, M. J.; Buckheit
Jr., R. W.; Hughes, S. H.; Cragg, G. M.; Boyd, M. R. J. Med. Chem. 1992, 35, 2735–2743; ibid. 1993, 36, 1110.
2. Xu, Z.-Q.; Hollingshead, M. G.; Borgel, S.; Elder, C.; Khilevich, A.; Flavin, M. T. Bioorg. Med. Chem. Lett.
1999, 9, 133–138.
3. Zembower, D. E.; Liao, S.; Flavin, M. T.; Xu, Z.-Q.; Stup, T. L.; Buckheit, Jr. R. W.; Khilevich, A.; Mar, A.
A.; Sheinkmann, A. K. J. Med. Chem. 1997, 40, 1005–1017 and references cited therein.
4. Ishikawa, T.; Oku, Y.; Tanaka, T.; Kumamoto, T. Tetrahedron Lett. 1999, 40, 3777–3780.
5. (a) Ishikawa, T.; Oku, Y.; Kotake, K.-I.; Ishii, H. J. Org. Chem. 1996, 61, 6484–6485; (b) Ishikawa, T.; Oku, Y.;
Kotake, K.-I. Tetrahedron 1997, 53, 14915–14928.
6. Ishii, H.; Kenmotsu, K.; Dopke, W.; Harayama, T. Chem. Pharm. Bull. 1992, 40, 1770–1772.

T. Tanaka et al. / Tetrahedron: Asymmetry 11 (2000) 4633–4637
4637
7. The coumarin 7 obtained in this reaction had been isolated from natural source [McKee, T. C.; Fuller, R. W.;
Covington, C. D.; Cardelline, II, J. H.; Gulakowski, R. J.; Krepps, B. L.; McMahon, J. B.; Boyd, M. R. J. Nat.
Prod. 1996, 59, 754–758].
8. Palmer, C. J.; Josephs, J. L. J. Chem. Soc., Perkin Trans. 1 1995, 3135–3152.
9. Up-shield shifted signals due to 3%-Me (l 1.54) and 3%-H (l 5.57–5.63) in 8 was observed compared to those
(3%-Me: l 1.84 and 3%-H: l 6.14–6.17) in 3 in their 1H NMR spectra. In addition, the presence of a strong
hydrogen bond in 8 was suggested by appearance of a signal at l 13.71. The corresponding signal in 3 was
resonated at l 11.86.
10. A commercially available (−)-quinine [purchased from Nacalai Tesque Co. Ltd (Japan)] was used after its
azeotropic dehydration using toluene and then drying.
11. A mixed solvent of CHCl3 and EtOAc (20:1) was used for the separation. The cis- and trans-isomer were
obtained as a more polar and a less polar component, respectively.
12. Each enantiomer of cis-2 was observed at retention times of 22.2 and 28.4 min, respectively, when CHIRALCEL
AS (Daicel Co. Ltd) was used as a column under the following conditions; eluent: n-hexane:EtOH=9:1, flow
rate: 1.0 ml/min, detection: 254 nm.
13. The dipole moment of chlorobenzene is reported to be 1.69 D, whereas that of anisole is 1.38 D. (Handbook of
Chemistry and Physics, 55th ed. West, R. C., Ed.; CRS Press: Ohio, 1974; E-63)
14. Chlorobenzene-assisted enantioselection was recently noted in a review of base-catalyzed asymmetric Diels–Alder
reaction using 3-hydroxy-2-pyrone [Okamura, H.; Iwagawa, T.; Nakatani, M. J. Synth. Org. Chem. (Jpn) 1999,
57, 84–91].
15. Recently we succeeded in the total synthesis of (+)-calanolide A 1 by application of the quinine-catalyzed
intramolecular oxo-Michael reaction of an o-tigloylphenol followed by MgI2-mediated isomerization of cis-chro-
manone to the trans-one (Tanaka, T.; Kumamoto, T; Ishikawa, T. Tetrahedron Lett., in press).
16. Each enantiomer of trans-2 was observed at retention times of 12.2 and 13.4 min, respectively, when CHIRAL-
PAK AD-RH (Daicel Co. Ltd) was used as a column under the following conditions; eluent: MeOH, flow rate:
1.0 ml/min, detection: 254 nm.
17. No reaction was observed at room temperature.
.
.