BF3-induced cyclobutane-opening of verbenone and its deconjugate homolog. Efficient preparation of o-mentha-1,8-dien-3-one and o-menth-1-en- 3-one in optically active forms

Article abstract of DOI:10.1016/S0957-4166(98)00510-2

Starting with (+)-verbenone, readily obtainable from (+)-nopinone, enantioselective preparation of (S)-(+)-4-isopropenyl-, (S)-(-)-4-isopropyl- and (R)-(+)-4-(1-acetoxy-1-methylethyl)-3-methyl-2-cyclohexen-1-ones was accomplished with little loss of stereochemical integrity via BF₃-induced cyclobutane-opening of (+)4-(methylene)nopinone. As we have developed an efficient chemical transformation of (+)-nopinone into (-)-verbenone, the present syntheses of the above cyclohexenones are formal syntheses of their enantiomers from (+)-nopinone.

Full text of DOI:10.1016/S0957-4166(98)00510-2

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 10 (1999) 339–346
BF3-induced cyclobutane-opening of verbenone and its
deconjugate homolog. Efficient preparation of o-mentha-1,8-dien-
3-one and o-menth-1-en-3-one in optically active forms
Takashi Kusakari, Tsuyoshi Ichiyanagi, Hiroshi Kosugi and Michiharu Kato ∗
Institute for Chemical Reaction Science, Tohoku University, Katahira, Aoba-ku, Sendai 980-8577, Japan
Received 19 November 1998; accepted 16 December 1998
Abstract
Starting with (+)-verbenone, readily obtainable from (+)-nopinone, enantioselective preparation of (S)-(+)-
4-isopropenyl-, (S)-(−)-4-isopropyl- and (R)-(+)-4-(1-acetoxy-1-methylethyl)-3-methyl-2-cyclohexen-1-ones
was accomplished with little loss of stereochemical integrity via BF3-induced cyclobutane-opening of (+)-
4-(methylene)nopinone. As we have developed an efficient chemical transformation of (+)-nopinone into
(−)-verbenone, the present syntheses of the above cyclohexenones are formal syntheses of their enantiomers from
(+)-nopinone. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
2-Cyclohexen-1-ones are potentially useful precursors for the synthesis of carbocycles. In fact, they
have frequently been utilized as acceptors in the conjugate addition reaction with carbon nucleophiles
and as dienophiles for the Diels–Alder reaction.1,2
During the course of our enantioselective synthesis of natural products from (+)-nopinone 1, readily
available in a large quantity by ozonolysis of (−)-β-pinene, a substantial quantity of o-mentha-1,8-diene-
3-one (4-isopropenyl-3-methyl-2-cyclohexen-1-one) 2, and o-menth-1-en-3-one (4-isopropyl-3-methyl-
2-cyclohexen-1-one) 3, in optically active forms were required as chiral starting materials (Fig. 1).
A search of the literature showed a few racemic preparations of these useful compounds, in which
transformation of the Hagemann’s ester into 2 in five steps and ca. 40% overall yield3 and EtAlCl2-
catalyzed cyclization of 8-methylnona-2,7-dien-4-one to give 3 in ca. 80% yield4 may be synthetically
reliable. Regarding optically active 2, Lander et al. have isolated (R)-(−)-2b in ca. 7% yield from a
complicated mixture produced on heating (−)-verbenone 4b with p-toluenesulfonic acid in acetic acid
∗
Corresponding author.
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00510-2

340
T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
Figure 1.
and acetic anhydride, along with a few ring-opened products 7–10 (Eq. 1).5 On the other hand, (S)-(+)-
2a was prepared by Fujisawa et al.,6 in the synthesis of (+)-β-elemenone, from (S)-2-cyclohexen-1-ol in
eight steps and ca. 15% overall yield. To the best of our knowledge, little is known about the preparation
of 3 in optically active forms. These findings prompted us to study an efficient synthesis of the title
compounds in optically active forms. We wish to report that, starting from (+)-verbenone 4a, efficient
preparations of (S)-2a and (S)-3a may be carried out in few steps and with synthetically satisfactory
overall yields.
(1)
2. Results and discussion
We have established that cyclobutane ring opening of 1 and its substituted nopinones 5 with a combined
reagent, BF3·OEt2/Zn(OAc)2/Ac2O, proceeded smoothly by a regioselective cleavage of the C(1)–C(6)
bond at 0°C–room temperature to give enol acetates of cyclohexanones, 117 and 12,8 respectively, in
good to high yields with little loss of optical integrity (Eq. 2). First, as part of a study on cyclobutane
opening of nopinone and its related compounds, we examined the reaction of verbenone 4a with our
combined reagent, because our reaction conditions are milder than those in Lander’s case.5 More
effective production of the cyclobutane-opened product 13, which would serve as a precursor of 2a,
could therefore be expected (Scheme 1). Optically active verbenone 4a is a natural product which is
available commercially. However, this substrate with high optical purity is expensive. Efficient synthesis
of 4a has been achieved by the allylic oxidation of (S)-(−)-α-pinene,9 and by chemical transformation of
1 via (+)-apoverbenone 6 which we have recently reported.10
(2)

T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
341
Scheme 1.
BF3-induced cyclobutane ring opening of 4a was carried out under our reaction conditions; stirring 4a
with BF3·OEt2 (0.5–1.0 equiv.) in the presence of Zn(OAc)2 (1.0 equiv.) in Ac2O7,8 (Scheme 1). The
reaction was very sluggish at 0°C, and the substrate 4a was recovered unchanged in a short reaction
time. On treating at room temperature for 3 h, enol acetates, 13 and 14, and aromatized compound 9
were obtained, albeit in 6, 4 and 2% yields, respectively, along with unreacted 4a. Enol acetate 13 is the
expected product arising from the C(1)–C(6) bond cleavage, whereas formation of enol acetate 14 would
be characteristic of verbenone 4a, and could be accounted for by cleavage of the C(5)–C(6) bond attached
to an enone function. On prolonged reaction (14 h) at room temperature, gradual disappearance of the
products 13 and 14 (TLC analyses) led to a mixture of complicated products, from which enone 2a in 21%
yields and aromatized compounds 9 and 10 in 7 and 3% yields, respectively, were obtained. Structural
elucidation of these products was carried out by comparison of their 1H NMR spectra with those of the
authentic samples. Treatment of 13 and 14 with BF3·OEt2 under the same reaction conditions produced
9 and 10, respectively. These findings indicate that enol acetates 13 and 14 are the initial products in
the ring-opening reaction, and compounds 9 and 10 could be derived from 13 and 14, respectively, by
elimination of an acetoxy group on the side chain followed by aromatization with the aid of acetic acid.
In addition, the enone 2a may act as a precursor for the formation of 9.
After all, BF3-induced cyclobutane-opening reaction of 4a provided directly 2a albeit in 21% yield.11
However, in spite of the careful optimization of reaction conditions, this reaction always produced a
complicated mixture of products. It was assumed that this disappointing result could be attributed not
only to the competing cleavage of the C(5)–C(6) bond of verbenone 4a, but also to elimination of an
acetoxy group at the homoallylic position in the products with formation of aromatized compounds 9
and 10. We thus turned our attention to cyclobutane-opening of enone 15, followed by reconstruction of
a conjugate enone function in the product (Scheme 2), because compound 15 is a deconjugate analog of
4a, and would react in a similar manner as 1 and 5, as seen in Eq. 2.
Enone 15 was prepared from 4a in high yield according to the procedure published by Ohloff et al.12 As
expected, the cyclobutane-opening of 15 under our reaction conditions proceeded smoothly even at 0°C
to give diacetate 16 and monoacetate 17 in 64 and 6% yields, respectively,13 which were readily separable
by silica gel column chromatography. Judging from the fact that the deconjugate enone 15 itself is easily
isomerized to its thermodynamically stable precursor 4a by a base or an acid, it is worth mentioning that,
under our reaction conditions, the ring-opening reaction smoothly proceeds in a regioselective fashion
prior to isomerization of a double bond. In fact, on prolonged reaction, formation of 4a was detected in a
complicated mixture of products.

342
T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
Scheme 2.
Regioselective elimination of an acetoxy group in the side chain of 16 was examined next. After va-
rious attempts at the pyrolysis of 16, it was found that passing a solution of 16 in hexane through a quartz
tube packed with quartz beads at 300°C with the aid of an N2 stream gave 17 in 75% yield as the sole
product.14 The desired enone (+)-2a was prepared from 17 in 82% yield by methanolysis with K2CO3
in methanol and concomitant isomerization of a double bond. Regioselective hydrogenation of 2a was
successfully carried out under the homogeneous conditions employing tris(triphenylphosphine)rhodium
chloride [(PPh3)3RhCl] to give 3a in 83% yield.14 The compounds 2a, 3a, 16, and 17 thus obtained are
stable, and can be stored in a refrigerator.
The enantiomeric purity of 2a and 3a were established by HPLC analysis employing a Chiral Pak AD
column with hexane–isopropyl alcohol as eluant, indicating the enantiomeric purities (ees) of both 2a
and 3a to be 98.9%. When the ee (99%) of the starting material 4a is taken into account, it was confirmed
that there is almost no loss of enantiomeric purity in the chemical transformations of 4a into 2a and 3a.
Finally, methanolysis of 16 in the presence of K2CO3 was carried out. When the reaction was
conducted at 0°C for 30 min, enone 18 was obtained in a nearly quantitative yield, whereas prolonged
exposure resulted in the unfavorable elimination of acetic acid to produce the dienone 7 as the major
product along with 18.
3. Conclusion
In summary, BF3-induced cyclobutane-opening of verbenone 4a provided a complicated mixture of
products, from which two initial products, 13 and 14, in a short reaction time, and enone 2a, on prolonged
exposure, were isolated along with aromatized compounds 9 and 10. On the other hand, the reaction of 15,
a deconjugate homolog of 4a, with BF3·OEt2 proceeded smoothly via an expected ring-opening to give
16 and 17 in good combined yield. On pyrolysis, compound 16 led to 17 in good yield. Enantioselective
preparation of enones (S)-2a and (S)-3a were accomplished by treatment of 17 with a base for the former,
and by regioselective hydrogenation of 2a for the latter. Ultimately, (S)-2a and (S)-3a were prepared
in four and five steps, and ca. 45 and ca. 40% overall yields from (+)-verbenone 4a, respectively. As
we have developed a general and efficient chemical transformation of 1 into (−)-verbenone 4b,15 the

T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
343
present synthesis of (S)-2 and (S)-3 is formal synthesis of their enantiomers, (R)-2 and (R)-3. The present
methodology is useful for the laboratory-scale preparation of 2 and 3 in optically active forms on account
of the availability of starting materials, simplicity of operations, and good overall yields.
4. Experimental section
4.1. General remarks
(+)-Verbenone 4a was obtained by oxidation of (S)-(−)-α-pinene (99% ee) according to the published
procedure by Lajunen et al.9 The deconjugate enone 15 was prepared from (+)-4a according to the
published procedure by Ohloff et al.11 1H NMR spectra were recorded at 400 MHz. J Values are
given in hertz. [α] Values are given in units of 10−1 deg cm2 g−1. All reactions were carried out
under a dry N2 or Ar atmosphere. Extracts obtained on aqueous work-up of the reaction mixtures were
washed successively with water and brine, unless otherwise stated, and Na2SO4 was used for drying of
extracts. Column chromatography was performed on 70–230 mesh silica gel (Merck). Medium-pressure
chromatography (MPLC) utilized a 22œ×300 mm silica gel (10 µm) column. Solvents for elution are
shown in parentheses. Ether refers to diethyl ether.
4.2. (R)-(−)-3-Acetoxy-6-(1-acetoxy-1-methylethyl)-1-methyl-1,3-cyclohexadiene 13, 1-acetoxy-6-(1-
acetoxy-1-methylethyl)-3-methyl-1,3-cyclohexadiene 14 and 1-acetoxy-4-isopropyl-3-methylbenzene 9
To a stirred mixture of 4a (1.0 g, 6.16 mmol) and zinc acetate (1.24 g, 6.16 mmol) in acetic anhydride
(12 ml) at 0°C was added BF3·OEt2 (0.86 ml, 6.16 mmol). After being stirred briefly, the cooling bath
was removed, and stirring was continued at room temperature for 3 h. The reaction mixture was diluted
with water (15 ml), and stirring was continued for an additional 2 h. The product was extracted with
ether:hexane (1:1), and the combined extracts were washed successively with aqueous NaHCO3, water
and brine, and dried. Evaporation of the solvent left a residue which was filtered through a short silica
gel column (hexane:EtOAc, 6:1). Concentration of the filtrate followed by purification of the residue by
MPLC (hexane:EtOAc, 10:1) gave 13 (88 mg, 6%), 14 (60 mg, 4%), 9 (24 mg, 2%) and unreacted 4a
(360 mg, 36%).
20
13: oil, [α]D −194.5 (c 0.34, CHCl3). IR (film) 1763, 1732, 1670, 1618, 1247, 1218, 1183, 1124,
1023, 912, 869, 817 cm−1; 1H NMR (CDCl3) δ 1.52 and 1.54 (3H, s each), 1.92 (3H, s), 1.99 and 2.13
(3H, s each), 2.42 (1H, ddd, J=18.3, 6.6, 1.5), 2.62 (1H, ddd, J=18.3, 10.2, 2.0), 2.89 (1H, d, J=10.2), 5.16
(1H, d, J=6.6, 2.0), 5.66 (1H, s). HRMS m/z calcd for C12H16O2 [M−OAc]+ 192.1149, found 192.1151.
20
14: oil, [α]D +89.8 (c 0.24, CHCl3). IR (film) 1763, 1732, 1682, 1617, 1584, 1250, 1212, 1182,
1132, 1078, 1044, 1019, 952, 982, 884 cm−1; 1H NMR (CDCl3) δ 1.41 and 1.47 (3H, s each), 1.70 (3H,
s), 1.98 and 2.13 (3H, s each), 2.35 (1H, ddd, J=18.6, 6.1, 1.3), 2.71 (1H, ddd, J=18.6, 10.8, 2.9), 3.16
(1H, dd, J=10.8, 1.3), 5.27 (1H, dd, J=6.1, 2.9), 5.68 (1H, s). HRMS m/z calcd for C12H1602 [M−OAc]+
192.1149, found 192.1151.
9: oil, IR (film) 1762, 1211, 1016, 902 cm−1; 1H NMR (CDCl3) δ 1.21 (6H, δ, J=6.8), 2.27 (3H, s),
2.32 (3H, s), 3.09 (1H, h, J=6.8), 6.84 (1H, d, J=2.6), 6.88 (1H, dd, J=8.3, 2.6), 7.23 (1H, d, J=8.3). The
IR and 1H NMR spectral data were identical with those of the authentic sample.4

344
T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
4.3. (S)-(+)-4-Isopropenyl-3-methyl-2-cyclohexen-1-one 2a and 2-acetoxy-1-isopropyl-4-methyl-
benzene 10
To a stirred mixture of 4a (162 mg, 1.0 mmol) and zinc acetate (183 mg, 1.0 mmol) in acetic anhydride
(2 ml) at 0°C was added BF3·OEt2 (0.13 ml, 1.0 mmol). After being stirred briefly, the cooling bath
was removed, and stirring continued at room temperature for 14 h. Aqueous work-up according to the
procedure described above followed by purification of the residue by MPLC gave 2a (31 mg, 21%), 9
(14 mg, 7%), 10 (6 mg, 3%) and unreacted 4a (15 mg, 9%).
2a: oil, [α]D +167.6 (c 0.34, CHCl3), {lit.5 [α]D +171.9 (c 1.01, CHCl3)}. IR (film) 1670, 1645,
20
22
1625, 1248, 1191, 898, 877 cm−1; H NMR (CDCl3) δ 1.81 (3H, s), 1.90 (3H, s), 1.96–2.13 (2H, m),
1
2.26 (1H, ddd, J=16.6, 5.8, 4.8), 2.41 (1H, ddd, J=16.6, 11.2, 4.8), 2.89 (1H, t, J=4.6), 4.74 (1H, s),
4.96 (1H, s), 5.97 (1H, s). The IR and 1H NMR spectral data were identical with those of the authentic
sample.3,5
1
10: oil, IR (film) 1763, 1368, 1209, 1120, 899 cm−1; H NMR (CDCl3) δ 1.18 (6H, δ, J=6.8), 2.31
(6H, s), 2.97 (1H, h, J=6.8), 6.80 (1H, d, J=1.0), 7.03 (1H, dd, J=8.1, 1.0), 7.19 (1H, d, J=8.1). The 1H
NMR spectrum was identical with that of the authentic sample (thymol acetate).
4.4. (R)-(−)-1-Acetoxy-4-(1-acetoxy-1-methylethyl)-5-(methylene)cyclohexene 16 and (R)-(−)-1-
acetoxy-4-isopropenyl-5-(methylene)cyclohexene 17
To a stirred mixture of 15 (150 mg, 1.00 mmol) and zinc acetate (183 mg, 1.00 mmol) in acetic
anhydride (3 ml) at 0°C was added BF3·OEt2 (0.06 ml, 0.50 mmol), and stirring was continued for 3
h. To the reaction mixture, water (7 ml) was added and stirring was continued for an additional 2 h.
The product was extracted with ether, and combined extracts were washed successively with aqueous
NaHCO3, water, and brine, and dried. Removal of the solvent left a residue which was chromatographed
on silica gel (hexane:ether, 9:1) to give 16 (162 mg, 64%) and 17 (11 mg, 6%) along with unreacted 15
(6 mg, 4%).
24
1
16: oil, [α]D −11.8 (c 0.15, CHCl3). IR (film) 1760, 1732, 1646, 1252, 1220, 1118, 1020, 944; H
NMR (CDCl3) δ 1.50 and 1.53 (3H, s each), 1.97 and 2.12 (3H, s each), 2.34 (2H, s with fine splittings),
2.79 (1H, d, J=18.8), 2.82–2.85 (1H, m), 3.04 (1H, d with fine splittings, J=18.8), 4.94 (1H, s), 4.99 (1H,
s), 5.40 (1H, s). Anal. calcd for C14H20O4: C, 66.65; H, 7.99. Found: C, 66.35; H, 7.87.
24
1
17: oil, [α]D −82.5 (c 1.04, CHCl3). IR (film) 1759, 1695, 1646, 1216, 1119, 1080, 1011, 895; H
NMR (CDCl3) δ 1.75 (3H, s), 2.12 (3H, s each), 2.32–2.37 (2H, m), 2.86–2.96 (3H, m), 4.84 (2H, s),
4.88 (1H, s), 4.95 (1H, s), 5.44 (1H, s). Anal. calcd for C12H16O2: C, 74.97; H, 8.39. Found: C, 74.65;
H, 8.24.
4.5. Pyrolysis of 16
A solution of 16 (1.20 g, 4.76 mmol) in hexane (10 ml) was passed through a quartz tube (21×1.5 cm)
packed with quartz beads at 300°C with the aid of an N2 stream, and the vapor was condensed in a dry
ice–acetone bath. The condensate was passed through a silica gel column (hexane:ether, 9:1) to give 17
1
(688 mg, 75%) and unreacted 16 (95 mg, 8%). The IR and H NMR spectral data of 17 were identical
with those of the authentic sample prepared above.

T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
345
4.6. Methanolysis of 17 with formation of 2a
A mixture of 17 (688 mg, 3.58 mmol) and K2CO3 (990 mg, 7.16 mmol) in methanol (5 ml) was stirred
at 0°C for 30 min, and then at room temperature for 1.5 h. Water was added to the reaction mixture,
and the product was extracted with ether. Concentration of the extracts followed by chromatography of
15
the residue on silica gel (hexane:ether, 4:1) gave 2a (439 mg, 82%), [α]D +187.1 (c 1.06, CHCl3),
1
whose IR and H NMR spectral data were identical with those of an authentic sample prepared from
4a. HPLC analysis employing a Chiral Pak AD column with hexane–isopropyl alcohol (49:1) as eluant
demonstrated the ee to be 98.9%.
4.7. (S)-(−)-4-Isopropyl-3-methyl-2-cyclohexen-1-one 3a
The compound 2a (150 mg, 1.0 mmol) and (Ph3P)3RhCl (46 mg, 0.05 mmol) were dissolved in
degassed benzene (2 ml), and the resulting solution was hydrogenated at atmospheric pressure for 6
h. The reaction mixture was added to a silica gel column, and the benzene was removed by use of hexane
19
as eluant, then the product was eluted with hexane:ether (4:1) to give 3a (127 mg, 83%) as an oil, [α]D
−20.8 (c 0.53, CHCl3). IR (film) 1678, 1619, 1253, 1191, 947, 872 cm−1; H NMR (CDCl3) δ 0.84
1
and 1.04 (3H, d, J=6.8 each), 1.86 (1H, m), 1.96 (1H, m), 1.99 (3H, s), 2.17–2.26 (2H, m), 2.28 (1H,
ddd, J=17.0, 10.0, 5.4), 2.46 (1H, ddd, J=17.0, 7.1, 4.7), 5.92 (1H, s). Anal. calcd for C10H16O: C,
78.90; H, 10.59. Found: C, 78.61; H, 10.56. HPLC analysis employing a Chiral Pak AD column with
hexane:isopropyl alcohol (99:1) as eluant demonstrated the ee to be 98.9%.
4.8. (R)-(+)-4-(1-Acetoxy-1-methylethyl)-3-methyl-2-cyclohexen-1-one 18
A mixture of 16 (154 mg, 0.61 mmol) and K2CO3 (337 mg, 2.44 mmol) in methanol (1 ml) was stirred
at 0°C for 30 min. The reaction mixture was quenched by addition of water and extracted with ether.
Removal of the solvent followed by chromatography of the residue on silica gel (hexane:ether, 4:1) gave
18 (89 mg, 96%) as an oil, [α]D24 +78.1 (c 0.92, CHCl3). IR (film) 1732, 1671, 1623, 1251, 1116, 1019,
951, 870; 1H NMR (CDCl3) δ 1.47 and 1.64 (3H, s each), 1.99 (3H, s), 2.07 (3H, 3), 2.02–2.15 (2H, m),
2.31 (1H, dt, J=17.8, 5.8), 2.36 (1H, ddd, J=17.8, 10.5, 5.8), 3.04 (1H, t, J=5.8), 6.01 (1H, s). HRMS m/z
calcd for C10H15O [M−OAc]+: 151.1122; found 151.1122.
Acknowledgements
We thank Soda Aromatic Co. Japan for providing us with samples of alkyl substituted phenols
including thymol.
References
1. Lee, V. J.; Kozlowski, J. A. Comprehensive Organic Synthesis; Trost, B. M., Ed.; Pergamon Press: Oxford, 1991; Vol. 4,
pp. 69–198.
2. Oppolzer, W. Comprehensive Organic Synthesis; Trost, B. M., Ed.; Pergamon Press: Oxford, 1991; Vol. 5, pp. 315–399.
3. Soria, O.; Maldonado, L. A. Synth. Commun. 1982, 12, 1093.
4. Snider, B. B.; Rodini, D. J.; van Straten, J. J. Am. Chem. Soc. 1980, 102, 5872.
5. Lander, N.; Mechoulam, R. J. Chem. Soc., Perkin Trans. 1 1976, 484.
6. Sato, T.; Gotoh, Y.; Watanabe, M.; Fujisawa, T. Chem. Lett. 1983, 1533.

346
T. Kusakari et al. / Tetrahedron: Asymmetry 10 (1999) 339–346
7. Kato, M.; Kamat, V. P.; Tooyama, Y.; Yoshikoshi, A. J. Org. Chem. 1989, 54, 1536.
8. Kato, M.; Watanabe, M.; Vogler, B.; Awen, B. Z.; Masuda, Y.; Tooyama, Y.; Yoshikoshi, A. J. Org. Chem. 1991, 56, 7071.
9. Lajunen, M.; Koskinen, A. M. P. Tetrahedron Lett. 1994, 35, 4461.
10. Kosugi, H.; Ku, J.; Kato, M. J. Org. Chem. 1998, 63, 6939.
11. Comparison of the optical rotation, [α]D +167.6 (CHCl3), of 2a directly derived from 4a with that, [α]D +187.1 (CHCl3),
of 2a obtained via 15 revealed that the former 2a may be somewhat racemized in the course of the ring-opening reaction
(see Experimental).
12. Ohloff, G.; Giersch, W. Helv. Chim. Acta 1977, 60, 1496.
13. On a large scale preparation, a tendency to decrease the combined yield of the ring-opened products 16 and 17 was
observed. We have, at present, no reasonable explanation for this observation.
14. For analogous reaction, see Kato, M.; Watanabe, M.; Tooyama, Y.; Vogler, B.; Yoshikoshi, A. Synthesis 1992, 1055.
15. Watanabe, M.; Awen, B. Z.; Kato, M. J. Org. Chem. 1993, 58, 3923.