Stereoselective isomerisations of 4-(2′,5′-dimethoxyphenyl)-2,5-dimethyl-1,3-dioxolanes and their 2′-chloro-5′-methoxyphenyl analogues. Temperature-dependent diastereoselective formation of isochromanes

Article abstract of DOI:10.1039/a704543c

Stereoselective isomerisations of rel-(2R,4S,5R)-4-(2′,5′-dimethoxyphenyl)-2,5-dimethyl-1,3-dioxolane 1 and the 2:1 epimeric mixture of rel-(2S,4R,5R)- and rel-(2R,4R,5R)-4-(2′,5′-dimethoxyphenyl)-2,5-dimethyl-1,3-dioxolanes 2 and 3 with titanium tetrachl

Full text of DOI:10.1039/a704543c

View Article Online / Journal Homepage / Table of Contents for this issue
Stereoselective isomerisations of 4-(2Ј,5Ј-dimethoxyphenyl)-2,5-
dimethyl-1,3-dioxolanes and their 2Ј-chloro-5Ј-methoxyphenyl
analogues. Temperature-dependent diastereoselective formation of
isochromanes
Robin G. F. Giles,*,a Rodney W. Rickards*,b and Badra S. Senanayakea,b
a Chemistry Department, Murdoch University, Murdoch, WA 6150, Australia
b Research School of Chemistry, Australian National University, Canberra, ACT 0200,
Australia
Stereoselective isomerisations of rel-(2R,4S,5R)-4-(2Ј,5Ј-dimethoxyphenyl)-2,5-dimethyl-1,3-dioxolane 1
and the 2:1 epimeric mixture of rel-(2S,4R,5R)- and rel-(2R,4R,5R)-4-(2Ј,5Ј-dimethoxyphenyl)-2,5-
dimethyl-1,3-dioxolanes 2 and 3 with titanium tetrachloride at Ϫ78 ЊC afford rel-(1R,3R,4S)- and rel-
(1S,3R,4R)-4-hydroxy-5,8-dimethoxy-1,3-dimethylisochromanes 25 and 30, respectively. The yields are
only moderate owing to the competing influence of the 2Ј-methoxy group in the starting dioxolanes, and
are improved significantly when this group is replaced by a 2Ј-chloro substituent. rel-(2R,4S,5R)-4-(2Ј-
Chloro-5Ј-methoxyphenyl)-2,5-dimethyl-1,3-dioxolane 13 under similar conditions is isomerised smoothly
to rel-(1R,3R,4S)-5-chloro-4-hydroxy-8-methoxy-1,3-dimethylisochromane 34 as the sole reaction
product. In contrast the C-2 epimers rel-(2R,4R,5R)- and rel-(2S,4R,5R)-4-(2Ј-chloro-5Ј-methoxyphenyl)-
2,5-dimethyl-1,3-dioxolanes 14 and 15 each favour the formation of rel-(1R,3R,4R)-5-chloro-4-hydroxy-8-
methoxy-1,3-dimethylisochromane 36 at Ϫ78 ЊC, while at Ϫ95 ЊC the 1S product 35 predominates.
Dioxolane 14 isomerises more rapidly than its C-2 epimer 15, and both these reactions are under kinetic
not thermodynamic control.
We have recently established that the isomerisation of 4-(3Ј,5Ј-
dimethoxyphenyl)-2,5-dimethyl-1,3-dioxolanes with titanium
tetrachloride is both high yielding and highly stereoselective.1
MeO
MeO
4-Hydroxy-6,8-dimethoxy-1,3-dimethylisochromanes
are
O
O
R1
R2
O
2
4
5
formed at low temperatures (Ϫ78 to Ϫ30 ЊC), and in the case of
the all-cis-dioxolane the diastereoselectivity at C-1 of the prod-
ucts is temperature-dependent. At higher temperatures (0 ЊC),
both the parent dioxolanes and the derived isochromanes
undergo further stereoselective isomerisation to afford dihydro-
isobenzofurans. Whereas the previous study used phenyl dioxo-
lane substrates in which the aromatic ring was symmetrically
substituted with two methoxy groups meta to the dioxolane
substituent, in this paper we extend the study to include 2Ј,5Ј-
dimethoxy- and 2Ј-chloro-5Ј-methoxy-phenyldioxolanes, as
these provide better models for the potential synthesis of natur-
ally occurring naphthopyrans and their quinones.2,3
MeO
MeO
O
2 R1 = Me, R2 = H
3 R1 = H, R2 = Me
1
MeO
MeO
R
4 R = CHO
5 R =
O
9 R =
Results and discussion
10 R =
O
Synthesis of the aryl dioxolanes
6 R =
The initial synthetic targets were the stereoisomeric 2Ј,5Ј-
dimethoxyphenyldioxolanes 1, 2 and 3. The starting aldehyde 4
was treated with ethylidenetriphenylphosphorane to give an
84% yield of a mixture of the (Z)- and (E)-alkenes 5 and 6 in a
4:1 ratio. When this mixture was reacted with bis(acetonitrile)-
dichloropalladium(),4 this ratio was reversed, but to no better
than 1:3. It has been noted previously that ortho substitution of
styrenes can prevent the otherwise total transformation of such
E/Z mixtures to the stereochemically pure E isomer.4 Since the
initial Wittig product was stereochemically more homogeneous
than after isomerisation with the palladium complex, it was
used in subsequent transformations with a view to separation
of derived stereoisomers at a later stage in the reaction
sequence.
OCOC6H4Cl
11 R =
12 R =
OH
OH
Cl
7
R =
OH
OH
OH
8 R =
OH
tetroxide,5 which yielded a 4:1 mixture of the erythro- and
threo-diols 7 and 8. This mixture was then enriched to 7:1 in
favour of the erythro-isomer 7 by chromatography. The stereo-
chemical assignments were supported by 1H NMR spectro-
scopy, the benzylic proton of the erythro-diol 7 resonating
further downfield (δ 4.89) and having a smaller vicinal coupling
The 4:1 mixture of alkenes 5 and 6 was reacted with N-
methylmorpholine N-oxide and a catalytic amount of osmium
J. Chem. Soc., Perkin Trans. 1, 1997
3361

View Article Online
constant (J 4.5 Hz) than that of the threo-diol 8 (δ 4.53,
J 7.1 Hz).6,7 Treatment of this enriched mixture of diols 7 and
8 with 1,1-dimethoxyethane in the presence of the catalyst
camphorsulfonic acid, followed by chromatography of the
crude product, gave the pure all-cis-dioxolane 1 in 72% yield.
To obtain the alternative 4,5-trans-dioxolanes 2 and 3, the
4:1 mixture of (Z)- and (E)-alkenes 5 and 6 was treated with m-
chloroperbenzoic acid in the presence of solid sodium hydrogen
carbonate. The stereochemically pure cis product 9 was the sole
epoxide isolated, in 66% yield, together with a diastereomeric
mixture (1:1) of hydroxy esters 11 formed by opening of the
epoxide ring at the benzylic position by m-chlorobenzoic acid.
The cis stereochemistry of 9 was confirmed by the relatively
large coupling constant (J 4.9 Hz) between the epoxide ring
protons.1 Ring-opening of the epoxide 9 by aqueous potassium
hydroxide in dimethyl sulfoxide1 proceeded with complete
stereoselectivity through nucleophilic inversion to afford the
threo-diol 8 in 76% yield. Acetalation of this diol 8 as above
afforded the 4,5-trans-dioxolanes 2 and 3 in 77% yield as a 2:1
mixture of epimers at C-2. The major, but not the minor, isomer
was obtained pure by chromatography. Individual assignments
of stereochemistry to the dioxolanes 2 and 3 were not made.
The analogous 2Ј-chloro-5Ј-methoxyphenyldioxolane stereo-
isomers 13, 14 and 15 were prepared from 2-chloro-5-methoxy-
the other hand, gave a mixture of the required 4,5-trans-
dioxolanes in an approximate 2:1 ratio, which could be separ-
ated by careful chromatography into the major and minor
1
isomers 14 and 15, pure by both GC and H NMR spectro-
scopy, in yields of 55 and 21%, respectively. Stereochemical
assignments at C-2 for this pair of epimers 14 and 15 were
based on NOE difference spectroscopy, as were the respective
preferred conformations 23 and 24. Thus, for the major isomer
OMe
MeO
Cl
H6′
H2
Cl
O
O
Me
H5
H5
H4
Me
O
O
H2
H4
Me
Me
23
24
14 with the conformation 23, irradiation of 2-H effected an 8%
enhancement of 5-H, and 9% was observed for the reverse
process. Irradiation of 2-H also caused a 16% enhancement of
6Ј-H. For the minor isomer 15 with the conformation 24,
irradiation of the C-5 methyl effected a 7% enhancement of
2-H and a 9% enhancement of 4-H. Irradiation of 2-H caused a
5% enhancement of 4-H. There was no observable enhance-
ment of 5-H when the C-2 methyl was irradiated, or vice versa.
MeO
MeO
Isomerisation of the 2Ј,5Ј-dimethoxyphenyldioxolanes
Treatment of the all-cis-dimethoxyphenyldioxolane
methylene dichloride with one equivalent of titanium tetra-
chloride at Ϫ78 ЊC for 30 min gave, aside from starting material
(20%), the isochromane 25 (18%) and a diastereomeric mixture
R1
R2
O
O
O
O
1 in
Cl
Cl
13
14 R1 = H, R2 = Me
15 R1 = Me, R2 = H
MeO
MeO
1
O
3
4
R
MeO
OH
Cl
16 R = CHO
17 R =
25
O
20 R =
21 R =
(1:1) of the chlorohydrins 12 (45%). The mass spectrum of
isochromane 25 showed a molecular ion at m/z 238, major
fragment ions at m/z 223 and 220, and a base peak at m/z 205.
These fragment ions result from losses of a methyl radical and
water, and establish the product as an isochromane rather than
an isomeric dihydroisobenzofuran.1 A further major ion at m/z
194 arose from an alternative retro Diels–Alder fragmentation
of the molecular ion with loss of acetaldehyde. The stereo-
chemistry of the isochromane 25 was defined by 1H NMR
spectroscopy, which showed resonances for the 1-H, 4-H and
3-H protons at δ 5.01 (q, J 6.6 Hz), δ 4.55 (d, J 7.9 Hz) and
δ 3.98 (dq, J 7.9 and 6.2 Hz), respectively. The large vicinal
coupling constant between 3-H and 4-H indicated axial and
pseudoaxial orientations, confirming that the C-3 methyl and
C-4 hydroxy groups were equatorial and pseudoequatorial.1
The chemical shift for 3-H (δ 3.98) in particular supported the
assignment of the C-1 methyl configuration as pseudoaxial,1,9,10
as did the absence of homoallylic coupling between 1-H and
the pseudoaxial 4-H. Such coupling (JaЈ,aЈ) would have been
expected had the C-1 methyl been pseudoequatorial and 1-H,
therefore, pseudoaxial.1
OH
18 R =
19 R =
OH
OH
22 R =
O
OH
benzaldehyde 16.8 Reaction with ethylidenetriphenylphos-
phorane gave an 80% yield of a mixture of the (Z)- and (E)-
alkenes 17 and 18 in a 1:1 ratio, which was improved to 1:13
in favour of the (E)-isomer 18 by isomerisation with bis-
(acetonitrile)dichloropalladium().4 This enriched mixture was
converted with m-chloroperbenzoic acid into a 1:13 mixture of
the cis- and trans-epoxides 19 and 20 in 90% yield, and thence,
through epoxide ring-opening and chromatography, into the
pure erythro-diol 21 in 86% yield. Alternatively, the 1:13 mix-
ture of alkenes 17 and 18 was hydroxylated as before with
osmium tetroxide, chromatography then affording the pure
threo-diol 22 in 80% yield. The expected configurations for
these diols were again confirmed by 1H NMR spectroscopy, the
benzylic proton of the erythro-diol 21 having a larger chemical
shift and smaller vicinal coupling constant (δ 5.22, J 3.2 Hz)
than the corresponding proton in the threo-epimer 22 (δ 4.90,
J 5.5 Hz).6,7
The 1H NMR spectrum of the diastereomeric chlorohydrins
12 was unexceptional. The GC–MS analysis of this mixture,
however, was noteworthy in that it showed three pairs of iso-
meric compounds. These were the chlorohydrins 12, with a
molecular ion pair at m/z 232 and 230 (ratio 1:3), the known
cis-epoxide 9 and its trans-diastereomer 10 with molecular ions
at m/z 194, and the structurally isomeric ketones 26 and 27 with
Acetalation of the erythro-diol 21 afforded the all-cis-
dioxolane 13 in 92% yield. Acetalation of the threo-diol 22, on
3362
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
small vicinal coupling (J 1.6 Hz) to 4-H defines the latter as
MeO
MeO
MeO
MeO
pseudoequatorial. The long-range homoallylic coupling (J 1.5
Hz) between 1-H and 4-H in the quinone 31 is consistent with
the configuration of 1-H as pseudoaxial, since that of 4-H is
pseudoequatorial. It is known11,13 that the coupling constant
O
between such pseudoaxial and pseudoequatorial protons (JaЈ,eЈ
)
O
is ~1.5 Hz, while two pseudoaxial protons have JaЈ,aЈ ~3.5 Hz,
and two pseudoequatorial protons give rise to JeЈ,eЈ ~0 Hz.
Thus, for quinone A and its dimethyl ether,3,11,12 where 1-H is
pseudoequatorial and 4-H is pseudoaxial (the reverse of the
arrangement in quinone 31), JaЈ,eЈ is 1.5 Hz, while for quinone
AЈ and its dimethyl ether, JeЈ,eЈ is not observed, a situation which
would have arisen if the C-1 methyl in quinone 31 were also
pseudoaxial.
When the isomerisaton of the dioxolane 1 or of the mixed
dioxolanes 2 and 3 was undertaken for short periods at
higher temperatures, either Ϫ30 or 0 ЊC, the chlorohydrin
diastereomers 12 were again produced, together with the iso-
chromene 32. For the mixture of dioxolanes 2 and 3, the
respective yields of the products 12 and 32 were 38 and 36%.
The dehydration product 32 arises through the influence of the
methoxy substituent peri to the C-4 hydroxy group in each of
the intermediate isochromanes 25 and 30 (shown for 30 in
Scheme 1).
26
27
MeO
MeO
+
+
O
MeO
MeO
28
29
molecular ions again at m/z 194. These ketones showed the
expected fragment ions 28 and 29, respectively, as base peaks
at m/z 165 and 151. It appeared that, at the GC injector
temperature of 250 ЊC, hydrogen chloride was eliminated from
the chlorohydrins 12 to afford the epoxides 9 and 10 which then
rearranged thermally to the ketones 26 and 27. In support of
this hypothesis, the chlorohydrins 12, prepared by treatment of
the epoxide 9 with titanium tetrachloride, gave a similar GC–
MS pattern.
MeO
MeO
O
O
32
Treatment of the 2:1 mixture of 4,5-trans-dioxolanes 2 and 3
with titanium tetrachloride, under the same conditions used for
the all-cis-dioxolane 1, gave rise to the new isochromane 30 as
H
MeO
+
MeO
OH
TiCl4
MeO
MeO
O
O
MeO
MeO
30
O
O
O
Scheme 1
The relatively low yields of the isochromanes 25 and 30
derived from these 2Ј,5Ј-dimethoxyphenyldioxolanes 1, 2 and 3,
in comparison with those from their 3Ј,5Ј-dimethoxy isomers,1
are ascribed to the influence of the 2Ј-methoxy substituent.
Being ortho to the dioxolane ring, this group promotes the
alternative C-4–O-3 bond cleavage upon reaction with the
Lewis acid (shown for intermediate 33 from dioxolane 1 in
Scheme 2), with formation of the chlorohydrins 12. This effect
OH
OH
30
31
32
the sole product of isomerisation together with the diastereo-
meric mixture of chlorohydrins 12 as previously described,
in yields of 45 and 31%, respectively. The isochromane
30 showed the characteristic mass spectroscopic pattern of
molecular ion at m/z 238, fragment ions at m/z 223 and 220, and
base peak at m/z 205,1 together with the retro Diels–Alder
fragment at m/z 194. 1H NMR spectroscopy confirmed the
structure 30 and revealed the stereochemistry. The three
heterocyclic methine protons 1-H, 4-H and 3-H resonated at δ
4.93 (q, J 6.3 Hz), δ 4.58 (d, J 1.6 Hz) and δ 3.56 (dq, J 1.6 and
6.3 Hz), respectively. The assignment of the C-1 methyl con-
figuration as pseudoequatorial followed both from NOE differ-
ence data, which showed reciprocal enhancements between
pseudoaxial and axial 1-H and 3-H protons (each 8%), with no
enhancement of 3-H on irradiation of the C-1 methyl group,
and also from the chemical shift of the 3-H proton at δ
3.56, which is in the range typical for cis-1,3-dimethyliso-
chromanes.1,9,10 The small coupling constant between 3-H and
4-H indicated that, with 3-H axial, 4-H was pseudoequatorial
and therefore the C-4 hydroxy group was pseudoaxial.1 A two-
dimensional NOESY spectrum was completely consistent with
these conclusions.
MeO
_
TiCl4
+
O
1
MeO
O
33
MeO
_
Cl
TiCl3
O
12
MeO
+
O
Scheme 2
Further evidence for the assigned relative stereochemistry for
the isochromane 30 was adduced through its oxidation to the
isochromanequinone 31, for which the 1-H, 4-H and 3-H pro-
tons resonated at δ 4.64 (dq, J 1.5 and 6.0 Hz), δ 4.38 (dd, J 1.5
and 1.6 Hz) and δ 3.57 (dq, J 1.6 and 6.3 Hz), respectively. The
chemical shift of the 3-H proton is consistent with those for 3-H
in eleutherin11 and its derivatives,12 whose C-1 and C-3 methyl
substituents are cis and equatorial; 3-H is thus axial, and its
is absent in the dioxolanes 13, 14 and 15 in which the 2Ј-
methoxy substituent is replaced by chlorine. Furthermore, the
chlorine substituent, while maintaining the asymmetric substi-
tution of the aromatic ring, would still allow for its ready oxid-
ation to afford quinones, or alternatively could be removed by
reductive dechlorination. The latter factors could be useful in
the synthesis of naturally occurring naphthopyrans and their
J. Chem. Soc., Perkin Trans. 1, 1997
3363

View Article Online
Table 1 The effect of reaction conditions on isomerisation of dioxolane 13
Ratio of products (%)
Conc.a/
Dioxolane
Isochromane
Diol
Entry
Conditionsa
mol lϪ1 × 10Ϫ3
13b
34b
21c
1
2
3
4
5
Ϫ95 ЊC,d 15 min
Ϫ95 ЊC,d 60 min
Ϫ78 ЊC,d 30 min
Ϫ78 ЊC,d 30 min
Ϫ78 to 0 ЊC,d 30 min
6
6
6
6
3
78
38
5
5
3
20
52
93
91
93
2
10
2
4
4
1
a Dioxolane 13 in methylene dichloride treated with titanium tetrachloride (2 equiv.). b Estimated from GC analysis. c Estimated from H NMR
spectral analysis. d Temperature at which the reaction was quenched with methanol, or at which an aliquot was removed for quenching in entry 1.
Table 2 The effect of reaction conditions on isomerisation of dioxolane 14
Ratio of
dioxolane
products
14:15b
Ratio of
Ratio of products (%)d
isochromane
products
35:36c
Dioxolanes
Isochromanes
35 ϩ 36
Isochromene
38
Entry
Conditionsa
14 ϩ 15
1
2
3
4
Ϫ95 ЊC,e 2 min
Ϫ95 ЊC,e 12 min
Ϫ95 ЊC,e 60 min
Ϫ95 ЊC, 12 min
Ϫ78 ЊC,e 30 min
Ϫ78 ЊC,e 2 min
Ϫ78ЊC,e 30 min
Ϫ78 ЊC, 30 min
0 ЊC,e 30 min
2:1
1:1
—
5.8:1
4.6:1
28:1
63
21
0
37
79
100
0
0
0
—
—
—
—
3.8:1
1:3.2
1:3.1
1:3.1
3
0
0
0
97
100
100
94
0
0
0
6
5
6
7
8
Ϫ78 ЊC,e 30 min
0 ЊC, 30 min
—
1:2.9
0
66
34
r.t.,e 60 min
9
10
Ϫ78 ЊC,e 2 min
Ϫ78 ЊC,e 30 min
1:1
1:1
3
3
97
97
0
0
—
a Dioxolane 14 in methylene dichloride at a concentration of 6 × 10Ϫ3 mol lϪ1 except entries 9 and 10 which are 3 × 10Ϫ3 mol lϪ1, treated with
titanium tetrachloride (2 equiv.). b Determined by 1H NMR spectral analysis. c Determined by GC analysis. d Estimated from GC analysis.
e Temperature at which the reaction was quenched with methanol, or at which an aliquot was removed for quenching in entries 1 and 2.
quinones.2,3 The isomerisation of the chlorinated dioxolanes
13, 14 and 15 was therefore investigated.
manes 35 and 36, epimeric at C-1, with the former being pre-
dominant. At Ϫ78, Ϫ30 and 0 ЊC, however, the epimer 36
became the major product. The mass spectrum of each isomer
resembled that of the isochromane 34. Individual stereochemi-
cal assignments for these isochromanes 35 and 36 were again
Isomerisation of the 2Ј-chloro-5Ј-methoxyphenyldioxolanes
Isomerisation of the all-cis-dioxolane 13 with two equivalents
of titanium tetrachloride gave rise to the isochromane 34 as a
MeO
MeO
MeO
O
O
O
Cl
OH
R
OH
Cl
OH
36 R = Cl
37 R = H
35
34
based on their 3-H chemical shifts of δ 3.68 and 4.11, respect-
ively, and on supportive NOE difference spectra. In the latter
experiments, reciprocal enhancements between the proximate
axial protons 1-H and 3-H were both 7% for isochromane 35.
For compound 36, irradiation of the pseudoaxial C-1 methyl
gave rise to a 10% enhancement of the axial 3-H. In each case,
the small coupling constant between 3-H and 4-H (1.2 Hz for
35 and 1.7 Hz for 36) confirmed the pseudoequatorial nature
of 4-H. Two-dimensional NOESY data supported these
assignments.
These preliminary results with mixtures of the dioxolanes 14
and 15 were confirmed when treatment of the pure epimer 14
for 60 min at Ϫ95 ЊC with the Lewis acid (Table 2, entry 3)
resulted in an almost exclusive preference (28:1) for the iso-
chromane 35. Earlier sampling of an incomplete reaction at
Ϫ95 ЊC (Table 2, entries 1 and 2) showed much lower product
diastereoselectivity; this is believed to reflect an uncontrolled
increase in the temperature of the sample taken and the sensi-
single diastereomer under all the conditions investigated, at
reaction temperatures ranging from Ϫ95 to 0 ЊC (Table 1). Its
mass spectrum showed a molecular ion pair at m/z 244/242
reflecting the presence of chlorine isotopes, and major fragment
ion pairs at m/z 229/227, 211/209 and 200/198, arising from loss
of a methyl radical, water and acetaldehyde and indicative of
the isochromane structure. The relative stereochemistry was
confirmed by criteria similar to those used previously, viz. the
large coupling constant (J 6.0 Hz) between 3-H and 4-H, the
chemical shift (δ 4.16) of the 3-H proton,1,9,10 an NOE differ-
ence spectrum in which irradiation of the C-1 methyl afforded
an enhancement (11%) of the 3-H proton and a two-
dimensional NOESY spectrum.
In contrast to the all-cis-dioxolane 13, the course of isomeri-
sation of the mixture of C-2 epimers 14 and 15 with two
equivalents of titanium tetrachloride was temperature depend-
ent. At Ϫ95 ЊC, the only products formed were the isochro-
3364
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
Table 3 The effect of reaction conditions on isomerisation of dioxolane 15
Ratio of
dioxolane
products
14:15b
Ratio of
Ratio of products (%)d
isochromane
products
35:36c
Dioxolanes
Isochromanes
35 ϩ 36
Diol
Entry
Conditionsa
14 ϩ 15
22b
1
2
3
4
5
6
7
Ϫ95 ЊC,e 12 min
Ϫ95 ЊC,e 60 min
Ϫ95 ЊC,e 60 min
Ϫ78 ЊC, e 5 min
Ϫ78 ЊC,e 30 min
Ϫ78 ЊC,e 5 min
Ϫ78 ЊC,e 30 min
1:5
1:5
1:5
11.6:1
9.8:1
9.0:1
1:3.3
1:3.4
1:1.7
1:1.6
90
69
64
6
~1
0
7
30
36
89
95
94
91
~3
~1
—
5
~4
~6
~9
—
—
—
0
a Dioxolane 15 in methylene dichloride at a concentration of 6 × 10Ϫ3 mol lϪ1 except entries 6 and 7 which are 3 × 10Ϫ3 mol lϪ1, treated with titanium
tetrachloride (2 equiv.). b Determined by 1H NMR spectral analysis. c Determined by GC analysis. d Estimated from GC analysis. e Temperature at
which the reaction was quenched with methanol.
_
TiCl4
tivity of the reaction diastereoselectivity to temperature. This
Ar
Ar
view was supported by an experiment commenced at Ϫ95 ЊC
and then warmed to Ϫ78 ЊC (Table 2, entry 4). When the entire
reaction was performed at Ϫ78 ЊC, the ratio of isochromanes 35
to 36 changed dramatically to ~1:3, the isochromane 36 becom-
ing the major product (Table 2, entries 5 and 6). At this tem-
O
O
+
_
Me
H
TiCl4
O
O
+
H
Me
Me
Me
1
39
40
perature, GC and H NMR spectral analyses showed that the
reaction was complete after only 2 min (entry 5) and that
the ratio of products was unchanged after 30 min (entry 6).
With higher temperatures and prolonged reaction times, the
isochromene dehydration product 38 became increasingly sig-
in complex 39 is more compelling than that for the correspond-
ing 2,5-diaxial interaction between hydrogen and the smaller
methyl group in complex 40.14
Secondly, in the incomplete reactions at Ϫ95 ЊC (Table 2,
entries 1 and 2 and Table 3, entries 1, 2 and 3), each of the
epimerically pure dioxolanes 14 and 15 was recovered as a
mixture of both these C-2 epimers. This epimerisation must
arise through reversible Lewis acid-catalysed opening of the
dioxolane ring to form a planar oxocarbenium ion. While the
isomerisation of the dioxolanes to isochromanes involves cleav-
age of the C-2–O-3 bond, it is probable that this competing
epimerisation of the dioxolanes involves specific cleavage of the
C-2–O-1 bond to give the formal intermediate 41. We have pre-
MeO
O
Cl
38
nificant (entries 7 and 8). Reactions carried out at higher dilu-
tion (entries 9 and 10) gave the isochromanes in ~1:1 ratio,
suggesting that the diastereoselectivity is concentration
dependent.
MeO
+
O
Related isomerisations were carried out on the C-2 epimeric
dioxolane 15, and similar observations were recorded. Thus, at
Ϫ95 ЊC, isochromane 35 was the major product (Table 3, entries
1, 2 and 3). At Ϫ78 ЊC, there was a dramatic reversal, with the
alternative isochromane 36 favoured (entries 4 and 5). The dia-
stereoselectivity was reduced when the dioxolane 15 was isom-
erised under conditions of higher dilution (entries 6 and 7).
The 5-chloro-8-methoxyisochromanes 35 and 36 isolated
from isomerisation of the dioxolanes 14 and 15 represent
kinetic rather than thermodynamic products, since an inde-
pendent experiment showed that there was no isomerisation of
isochromane 35 into its C-1 epimer 36 when the former was
treated with the Lewis acid at Ϫ78 ЊC before standing at 0 ЊC
for 30 min. This is in contrast to the situation observed for a
6,8-dimethoxyisochromane, where the combined effect of the
two methoxy groups is sufficient to cleave the C-1–O-2 bond
reversibly.1
A comparison of Tables 2 and 3 reveals two noteworthy
effects. First, under the same conditions the dioxolane 14 reacts
much more rapidly than does its epimer 15. Thus, at Ϫ95 ЊC
after both 12 and 60 min, the reactions of compounds 14 and
15 had proceeded to very different extents (Table 2, entries 2
and 3 compared with Table 3, entries 1, 2 and 3, respectively). In
fact, the reaction of 14 in 2 min proceeded to approximately the
same extent as that of isomer 15 in 60 min (Table 2, entry 1
compared with Table 3, entries 2 and 3). This rate difference can
be rationalised in terms of the respective titanium-coordinated
intermediates 39 and 40, for which relief of the 2,4-diaxial
interaction between the hydrogen and the bulky aromatic ring
_
Cl
OTiCl4
41
viously suggested that such ions cannot, for stereoelectronic
reasons, cyclise to dihydroisobenzofurans and therefore reform
the dioxolanes.1 Furthermore, the ratios of epimeric dioxolanes
recovered from incomplete reactions of 14 (Table 2) show a
decreasing proportion of this isomer with time, whereas for the
C-2 epimer 15 (Table 3), the ratio favours that isomer by a
factor of 5:1 after both 12 and 60 min. These results suggest
that there is a preference for the isomer 15 at Ϫ95 ЊC in the
presence of titanium tetrachloride, which can be understood
in terms of the lesser diaxial interactions in the correpond-
ing complex 40 than in the complex 39, as discussed above.
In contrast, the acid-catalysed formation of the pair of free
dioxolanes 14 and 15 from the corresponding threo-diol 22 at
the temperature of boiling methylene dichloride (~40 ЊC)
favours the dioxolane 14.
Preliminary investigations on the isomerisation of naphthyl-
dioxolanes have shown a preference for the formation of angu-
lar naphthopyrans.15 The use of halogen as a blocking group
and its subsequent removal could possibly be used to direct the
synthesis of linear naturally occurring naphthopyrans2 such as
glucoside B.3 The removal of chlorine from the isochromane 36
derived from isomerisation of the chlorinated dioxolanes 14
and 15 would complete a model for this process. In practice it
J. Chem. Soc., Perkin Trans. 1, 1997
3365

View Article Online
was found that photochemically induced reductive dechlorin-
ation of 36 could be achieved smoothly by Beckwith’s method16
to afford the new isochromane 37 in 58% yield.
(Z)- and (E)-1-(2Ј,5Ј-Dimethoxyphenyl)prop-1-enes 5 and 6
Sodium hydride (55–60% dispersion in mineral oil, 200 mg) was
stirred in dimethyl sulfoxide (0.5 ml) and tetrahydrofuran (2 ml)
at 60 ЊC for 2 h in an atmosphere of argon. The mixture was
cooled to room temperature and ethyltriphenylphosphonium
bromide (1.8 g, 5.20 mmol) in tetrahydrofuran (10 ml) was
added and stirred for 15 min. The aldehyde 4 (500 mg, 3.01
mmol) in tetrahydrofuran (20 ml) was added and the mixture
stirred at 60 ЊC for a further 1 h. The reaction mixture was
centrifuged and the supernatant layer separated. This organic
layer was washed with dilute hydrochloric acid (10%) and the
residue obtained upon work-up was chromatographed (7%
ethyl acetate–hexane) to give a 4:1 mixture of the (Z)- and (E)-
alkenes 5 and 6 as a colourless oil (450 mg, 84%) (Found: C,
74.4; H, 7.9. C11H14O2 requires C, 74.1; H, 7.9%); δH(for 5) 6.83
(1H, d, J 2.8, 6Ј-H), 6.81 (1H, dd, J 7.8 and 2.8, 4Ј-H), 6.68 (1H,
d, J 7.8, 3Ј-H), 6.47 (1H, dq, J 11.4 and 1.2, 1-H), 5.88 (1H, dq,
J 11.4 and 7.1, 2-H), 3.80 (6H, s, 2 × OCH3) and 1.85 (3H, dd,
J 7.1 and 1.2, CH3); δC(for 5) 152.9 and 151.3 (C-2Ј, C-5Ј),
127.3, 127.1 and 125.1 (C-1, C-1Ј, C-2), 116.2 (C-6Ј), 112.1 and
111.2 (C-3Ј, C-4Ј), 56.0 and 55.7 (OCH3) and 13.9 (CH3);
δH(for 6) 6.98 (1H, d, J 2.8, 6Ј-H), 6.79 (2H, m, 3Ј-H and
4Ј-H), 6.68 (1H, dq, J 16.0 and 1.2, 1-H), 6.24 (1H, dq, J 16.0
and 6.6, 2-H), 3.80 (6H, s, 2 × OCH3) and 1.91 (3H, dd, J 6.6
and 1.2, CH3); m/z (for 5 and 6) 178 (Mϩ, 100%), 163 (10), 135
(10), 121 (14) and 103 (20).
Conclusions
We have shown previously that 4-(3Ј,5Ј-dimethoxyphenyl)-2,5-
dimethyl-1,3-dioxolanes are stereoselectively isomerised to
4-hydroxy-6,8-dimethoxy-1,3-dimethylisochromanes in high
yield by titanium tetrachloride at low temperature.1 The present
work demonstrates that whereas similar isomerisations of the
corresponding 2Ј,5Ј-dimethoxyphenyldioxolanes 1, 2 and 3
proceed in only mediocre yield, owing to the competing influ-
ence of the 2Ј-methoxy substituent in the starting materials, the
analogous 2Ј-chloro-5Ј-methoxyphenyldioxolanes 13, 14 and 15
are converted smoothly into isochromanes. In contrast to the
6,8-dimethoxyisochromanes derived from the previous 3Ј,5Ј-
dimethoxyphenyldioxolanes, the 5,8-dimethoxy- and 5-chloro-
8-methoxy-isochromanes of the present work lack the add-
itional activation provided by the 6-methoxy substituent and do
not isomerise further to dihydroisobenzofurans at higher
temperatures.
In all these isomerisations, the vicinal stereochemistry at C-4
and C-5 of the dioxolanes is transferred unaltered to C-4 and
C-3, respectively, of the isochromanes. The third stereogenic
centre C-1 of the isochromanes is derived from C-2 of the
parent dioxolanes. In some cases in both the previous and pres-
ent work, the same relative stereochemistry of the isochromane
C-1 methyl substituent remains favoured over the temperature
range examined, whereas in others the preference can be
reversed by altering the reaction temperature. In the example
of the C-1 epimeric 5-chloro-8-methoxyisochromanes 35 and
36 formed from the dioxolanes 14 and 15, the dramatic dia-
stereochemical reversal between Ϫ95 and Ϫ78 ЊC is kinetic not
thermodynamic, and is also concentration dependent. The fac-
tors which determine this relative stereochemistry at C-1 are
complex and not yet understood. It is noteworthy that in those
cases where a single stereoisomer of the isochromane was
formed, the substituents at C-1 and C-4 are cis-related, i.e. one
is pseudoaxial while the other is pseudoequatorial. This is also
true of the major stereoisomer where both epimeric iso-
chromanes were formed in a ratio which did not vary with tem-
perature. Finally, where both epimeric isochromanes were
formed in a ratio which reversed with temperature, the same cis-
relationship holds for the stereoisomer which is favoured at the
lower temperature. This phenomenon is being investigated
further.
rel-(1S,2R)- and rel-(1R,2R)-1-(2Ј,5Ј-Dimethoxyphenyl)-
propane-1,2-diols 7 and 8
The 4:1 mixture of alkenes 5 and 6 (320 mg, 1.54 mmol) in
a 2:1 mixture of acetone–water (9 ml) was treated with N-
methylmorpholine N-oxide (246 mg, 2.10 mmol) and osmium
tetroxide (5 mg) in tert-butyl alcohol (1 ml) at 0 ЊC. After
stirring for 24 h, the acetone was removed under vacuum at
room temperature. The remaining aqueous layer was poured
into dilute hydrochloric acid (2 , 5 ml) and the organic
materials extracted into ethyl acetate (5 × 20 ml). The residue
(310 mg) obtained upon work-up was chromatographed (50%
ethyl acetate–hexane) to yield a 7:1 mixture of the diols 7 and
8 as an orange oil (270 mg, 69%) (Found: C, 62.0; H, 7.9.
C11H16O4 requires C, 62.25; H, 7.6%); δH(for 7) 7.00 (1H, d,
J 2.2, 6Ј-H), 6.80 (2H, narrow m, 3Ј-H and 4Ј-H), 4.89 (1H, d,
J 4.5, 1-H), 4.12 (1H, dq, J 4.5 and 6.4, 2-H), 3.78 and 3.77
(each 3H, s, OCH3) and 1.09 (3H, d, J 6.4, CH3); δC(for 3) 153.1
and 150.6 (C-2Ј, C-5Ј), 129.6 (C-1Ј), 113.9, 112.9 and 111.3
(C-3Ј, C-4Ј, C-6Ј), 73.8 (C-1), 70.0 (C-2), 55.8 and 55.5 (OCH3)
and 17.2 (CH3); δH(for 8) 6.87 (1H, d, J 2.6, 6Ј-H), 6.80 (2H,
narrow m, 3Ј-H and 4Ј-H), 4.53 (1H, d, J 7.1, 1-H), 3.95 (1H,
dq, J 7.1 and 6.3, 2-H), 3.80 and 3.77 (each 3H, s, OCH3) and
1.07 (3H, d, J 6.3, CH3); m/z (for 7 and 8) 212 (Mϩ, 12%), 168
(22), 167 (100), 139 (40), 137 (32) and 124 (23).
Experimental
Elemental analyses were carried out by the ANU Microanalyti-
1
cal Services Unit. Unless otherwise stated H and 13C NMR
spectra were measured at 300 and 75 MHz, respectively, on
Varian Gemini-300 and VXR-300 spectrometers for solutions
in [2H]chloroform with tetramethylsilane as internal reference.
J Values are given in Hz. Electron impact mass spectra were
recorded at 70 eV on a VG Micromass 7070 spectrometer, while
high resolution measurements were carried out on an AEI MS
902 instrument. GC–MS analyses were performed on a Hewlett
Packard 5970 spectrometer and 5890 chromatograph. GC
analyses employed a Varian 3400 chromatograph with flame
ionisation detection, and isomers were assumed to give identi-
cal responses. Melting points were determined on a Reichert
hot-stage microscope and are uncorrected. Flash column
chromatography was performed on silica gel (Merck, mesh
0.004–0.063 mm) as described by Still et al.17 Light petroleum
refers to the fraction bp 60–80 ЊC. The phrase ‘residue obtained
upon work-up’ refers to the residue obtained when an organic
phase was separated, dried (MgSO4) and the solvent evaporated
under reduced pressure.
rel-(2R,4S,5R)-4-(2Ј,5Ј-Dimethoxyphenyl)-2,5-dimethyl-1,3-
dioxolane 1
The 7:1 mixture of diols 7 and 8 (200 mg, 0.94 mmol) in dry
methylene dichloride (20 ml) was treated with 1,1-dimethoxy-
ethane (0.13 ml) and ( )-camphorsulfonic acid (10 mg, 0.042
mmol) and the solution was boiled for 1 h. The reaction was
quenched with saturated aqueous sodium hydrogen carbonate.
The organic layer was separated and the aqueous layer ex-
tracted with more methylene dichloride. The residue (190 mg)
obtained upon work-up was chromatographed (10% ethyl
acetate–hexane) to give the dioxolane 1 as a colourless oil (162
mg, 72%) (Found: C, 65.5; H, 7.9. C13H18O4 requires C, 65.5; H,
7.6%); δH 7.07 (1H, d, J 2.1, 6Ј-H), 6.79 (2H, narrow m, 3Ј-H
and 4Ј-H), 5.35 (1H, d, J 7.2, 4-H), 5.18 (1H, q, J 4.9, 2-H), 4.44
(1H, dq, J 7.2 and 6.3, 5-H), 3.78 and 3.77 (each 3H, s, OCH3),
1.54 (3H, d, J 4.9, 2-CH3) and 0.81 (3H, d, J 6.3, 5-CH3); δC
153.5 and 149.9 (C-2Ј, C-5Ј), 126.2 (C-1Ј), 113.9, 112.5 and
110.5 (C-3Ј, C-4Ј, C-6Ј), 99.6 (C-2), 75.5 and 74.8 (C-4, C-5),
3366
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
55.1 (2 × OCH3), 19.5 (CH3-2) and 15.5 (CH3-5); m/z 238 (Mϩ,
35%), 194 (84), 179 (22), 165 (41), 163 (100), 151 (18), 135 (86)
and 44 (99).
(Z)- and (E)-1-(2Ј-Chloro-5Ј-methoxyphenyl)prop-1-enes 17
and 18
Sodium hydride (55–60% dispersion in mineral oil, 400 mg) was
stirred with dimethyl sulfoxide (1 ml) and tetrahydrofuran (4
ml) at 60 ЊC for 2 h in an atmosphere of argon. The mixture was
cooled to room temperature and ethyltriphenylphosphonium
bromide (3.2 g, 8.80 mmol) in tetrahydrofuran (10 ml) was
added and stirred for 15 min. The aldehyde 16 (1 g, 6.02 mmol)
in tetrahydrofuran (40 ml) was added and the mixture stirred at
60 ЊC for another 1 h. The reaction mixture was centrifuged and
the supernatant layer separated. The residue obtained upon
work-up was chromatographed (7% ethyl acetate–hexane) to
give a 1:1 mixture (confirmed by GC analysis) of the (Z)- and
(E)-olefins 17 and 18 as a colourless oil (850 mg, 80%). This
mixture in ethanol (10 ml) with bis(acetonitrile)dichloro-
palladium() (10 mg) was heated under reflux for 2 h. The
catalyst was removed by filtration and the filtrate evaporated to
afford a residue which was chromatographed (7% ethyl acetate–
hexane) to give the (E)-alkene 18 (93% by GC) contaminated
with the (Z)-olefin 17 (7% by GC) (Found: C, 65.8; H, 6.1; Cl,
19.4. C10H11ClO requires C, 65.8; H, 6.2; Cl, 19.3%); δH(for
18) 7.25 (1H, d, J 8.8, 3Ј-H), 7.03 (1H, d, J 2.9, 6Ј-H), 6.78
(1H, dq, J 16.0 and 1.8, 1-H), 6.72 (1H, dd, J 8.8 and 2.9, 4Ј-
H), 6.24 (1H, dq, J 16.0 and 6.6, 2-H), 3.81 (3H, s, OCH3) and
1.95 (3H, dd, J 6.6 and 1.8, CH3); δC(for 18) 158.2 (C-5Ј),
136.6 and 133.8 (C-1Ј, C-2Ј), 130.1, 128.5 and 127.3 (C-1, C-2,
C-6Ј), 113.8 and 111.4 (C-3Ј, C-4Ј), 55.4 (OCH3) and 18.7
(CH3); m/z 184 [Mϩ(37Cl), 31%], 182 [Mϩ(35Cl), 100], 147 (90)
and 132 (13). Inspection of the 1:1 mixture indicated the fol-
lowing data for the (Z)-olefin 17; δH 7.32 (1H, d, J 8.8, 3Ј-H),
6.91 (1H, d, J 2.9, 6Ј-H), 6.72 (1H, dd, J 8.8 and 2.9, 4Ј-H),
6.55 (1H, dq, J 11.5 and 1.8, 1-H), 5.96 (1H, dq, J 11.5 and
7.1, 2-H), 3.84 (3H, s, OCH3) and 1.85 (3H, dd, J 7.1 and 1.8,
CH3).
cis-1-(2Ј,5Ј-Dimethoxyphenyl)-1,2-epoxypropane 9
m-Chloroperbenzoic acid (500 mg, 2.91 mmol) in chloroform
(30 ml) was added dropwise to the 4:1 mixture of alkenes 5 and
6 (450 mg, 2.52 mmol) in chloroform (10 ml) at 0 ЊC and the
reaction mixture stirred with sodium hydrogen carbonate (100
mg) for 24 h. This mixture was filtered and the filtrate poured
into saturated aqueous sodium hydrogen carbonate (15 ml).
The organic phase was separated and the aqueous phase
extracted with chloroform (3 × 10 ml). The residue (800 mg)
obtained upon work-up was chromatographed (10% ethyl
acetate–hexane) to afford first the epoxide 9 as a white solid
(300 mg, 66%), mp 65–66 ЊC (light petroleum) (Found: C,
68.0; H, 7.6. C11H14O3 requires C, 68.0; H, 7.3%); δH 6.83
(1H, d, J 2.2, 6Ј-H), 6.78 (2H, narrow m, 3Ј-H and 4Ј-H),
4.16 (1H, d, J 4.9, 1-H), 3.80 and 3.76 (each 3H, s, OCH3),
3.38 (1H, dq, J 4.9 and 5.2, 2-H) and 1.06 (3H, d, J 5.2,
CH3); δC 153.0 and 152.0 (C-2Ј, C-5Ј), 125.0 (C-1Ј), 113.4,
113.0 and 110.9 (C-3Ј, C-4Ј, C-6Ј), 55.8 and 55.7 (OCH3),
55.1 and 54.7 (C-1, C-2) and 12.6 (CH3); m/z 194 (Mϩ, 46%),
179 (11), 165 (20), 163 (43), 151 (32) and 135 (100). This was
followed by the mixture of hydroxy esters 11 (13 mg, 7%)
which showed significant resonances at δH 8.07, 7.97, 7.41,
7.38, 6.92, 6.84, 6.78 (14H, Ar-H of both isomers), 6.38 and
6.26 (each 1H, d, J 3.9 and 5.9 respectively, 1-H of each
isomer), 4.23 (2H, m, 2-H of both isomers), 3.84 and 3.72
(each 6H, s, OCH3) and 1.20 (6H, d, J 6.5, CH3 of both
isomers); m/z 350 (Mϩ, 2%), 306 (9), 167 (23), 151 (6), 140
(28), 139 (100) and 111 (24).
rel-(1R,2R)-(2Ј,5Ј-Dimethoxyphenyl)propane-1,2-diol 8
The epoxide 9 (300 mg, 1.55 mmol) in dimethyl sulfoxide (15
ml) and potassium hydroxide (0.4 , 6 ml) was stirred at 80 ЊC.
After 24 h, the resulting solution was cooled to room temper-
ature, poured into water and extracted with ethyl acetate
(4 × 20 ml). The residue obtained upon work-up was chrom-
atographed (50% ethyl acetate–hexane) to give the diol 8 as
a light orange oil (250 mg, 76%) (Found: C, 62.6; H, 7.9.
C11H16O4 requires C, 62.25; H, 7.6%); δH 6.87 (1H, d, J 2.6,
6Ј-H), 6.80 (2H, narrow m, 3Ј-H and 4Ј-H), 4.53 (1H, d, J 7.1,
1-H), 3.95 (1H, dq, J 7.1 and 6.3, 2-H), 3.80 and 3.77 (each 3H,
s, OCH3) and 1.07 (3H, d, J 6.3, CH3); m/z 212 (Mϩ, 11%), 168
(22), 167 (100), 139 (52), 137 (34) and 124 (23).
trans-1-(2Ј-Chloro-5Ј-methoxyphenyl)-1,2-epoxypropane 20
m-Chloroperbenzoic acid (600 mg, 3.49 mmol) in chloroform
(20 ml) at 0 ЊC was added dropwise to the 1:13 mixture of
olefins 17 and 18 (450 mg, 2.71 mmol) in chloroform (10 ml)
and the solution stirred with sodium hydrogen carbonate (100
mg) for 24 h. The reaction mixture was filtered and the filtrate
poured into saturated aqueous sodium hydrogen carbonate (15
ml). The organic phase was separated and the aqueous phase
extracted with cold chloroform (3 × 10 ml). The residue (500
mg) obtained upon work-up was carefully chromatographed
(10% ethyl acetate–hexane) to afford the 1:13 mixture of
epoxides 19 and 20 as an orange oil (440 mg, 90%). A small
portion of this was rechromatographed to yield the pure
epoxide 20 (Found: C, 60.7; H, 5.3; Cl, 17.6. C10H11ClO2
requires C, 60.5; H, 5.6; Cl, 17.85%); δH 7.23 (1H, d, J 8.5,
3Ј-H), 6.76 (1H, d, J 2.3, 6Ј-H), 6.74 (1H, dd, J 8.5 and 2.3, 4Ј-
H), 3.88 (1H, d, J 2.0, 1-H), 3.76 (3H, s, OCH3), 2.87 (1H, dq,
J 2.0 and 5.1, 2-H) and 1.50 (3H, d, J 5.1, CH3); δC 159.3 (C-5Ј),
137.2, 130.3 and 130.2 (C-1Ј, C-2Ј, C-3Ј), 115.5 and 111.1 (C-4Ј,
C-6Ј), 59.3 and 57.6 (C-1, C-2), 56.1 (OCH3) and 18.4 (CH3);
m/z 200 [Mϩ(37Cl), 13%], 198 [Mϩ(35Cl), 35], 183 (11.5), 169
(32), 167 (55), 154 (14) and 119 (100).
rel-(2S,4R,5R)- and rel-(2R,4R,5R)-4-(2Ј,5Ј-Dimethoxyphenyl)-
2,5-dimethyl-1,3-dioxolanes 2 and 3
The diol 8 (230 mg, 1.08 mmol) in dry methylene dichloride (25
ml) was treated with 1,1-dimethoxyethane (0.16 ml) and ( )-
camphorsulfonic acid (10 mg) in a manner similar to that for
the diol mixture 7 and 8. The product (230 mg) was chromato-
graphed (10% ethyl acetate–hexane) to give a ca. 2:1 mixture
of dioxolanes 2 and 3 as a colourless oil (200 mg, 77%); δH 7.12
(2H, d, J 2.5, 6Ј-H of both isomers), 6.80 (4H, narrow m, 3Ј-H
and 4Ј-H of both isomers), 5.38 (2H, m, 2-H of both isomers),
4.96 and 4.87 (each 1H, d, J 6.0, 4-H of each isomer), 3.93
(2H, dq, J 7.0 and 6.4, 5-H of both isomers), 3.79 and 3.75
(each 6H, s, OCH3 of both isomers), 1.50 and 1.42 (each 3H, d,
J 5.0, 2-CH3 of each isomer) and 1.36 and 1.34 (each 3H, d,
J 6.4, 5-CH3 of each isomer). Chromatographic separation of
the mixture afforded a single diastereoisomer (40 mg, 40%)
(Found: C, 65.5; H, 7.6. C13H18O4 requires C, 65.5; H, 7.6%); δH
7.12 (1H, d, J 2.5, 6Ј-H), 6.80 (2H, narrow m, 3Ј-H and 4Ј-H),
5.38 (1H, q, J 5.0, 2-H), 4.96 (1H, d, J 6.0, 4-H), 3.97 (1H, dq,
J 6.0 and 6.4, 5-H), 3.79 and 3.75 (each 3H, s, OCH3), 1.50 (3H,
d, J 5.0, 2-CH3) and 1.36 (3H, d, J 6.4, 5-CH3); m/z 238 (Mϩ,
14%), 194 (36), 179 (10), 165 (18), 163 (39), 151 (8), 135 (46) and
44 (100).
rel-(1S,2R)-1-(2Ј-Chloro-5Ј-methoxyphenyl)propane-1,2-diol 21
The 1:13 mixture of epoxides 19 and 20 (350 mg, 1.77 mmol) in
dimethyl sulfoxide (15 ml) and potassium hydroxide (0.4 , 6
ml) was stirred at 80 ЊC. After 24 h, the resulting solution was
cooled to room temperature, poured into water and extracted
with ethyl acetate (4 × 20 ml). The residue obtained upon work-
up was chromatographed (50% ethyl acetate–hexane) to afford
the diol 21 as a light orange oil (330 mg, 86%) (Found: C, 55.7;
H, 6.2; Cl, 16.2. C10H13ClO3 requires C, 55.4; H, 6.05; Cl,
16.4%); δH 7.27 (1H, d, J 8.7, 3Ј-H), 7.23 (1H, d, J 2.9, 6Ј-H),
6.81 (1H, dd, J 8.7 and 2.9, 4Ј-H), 5.22 (1H, d, J 3.2, 1-H), 4.26
(1H, dq, J 3.2 and 6.5, 2-H), 3.85 (3H, s, OCH3) and 1.10 (3H,
J. Chem. Soc., Perkin Trans. 1, 1997
3367

View Article Online
d, J 6.5, CH3); δC 158.3 (C-5Ј), 139.0 (C-2Ј), 129.7 (C-3Ј) 123.2
(C-1Ј), 114.3 and 113.3 (C-4Ј, C-6Ј), 73.4 (C-1), 69.1 (C-2),
55.4 (OCH3) and 15.9 (CH3); m/z 218 [Mϩ(37Cl), 2%] 216
[Mϩ(35Cl), 5], 174 (21), 173 (15), 172 (69), 171 (29), 143 (37),
137 (34), 109 (38), 77 (51) and 44 (100).
5.02 (1H, d, J 5.4, 4-H), 4.03 (1H, dq, J 5.4 and 6.5, 5-H), 3.80
(3H, s, OCH3), 1.54 (3H, d, J 4.8, 2-CH3) and 1.43 (3H, d, J
6.5, 5-CH3); mass spectral fragmentation pattern identical to
isomer 14.
rel-(1R,3R,4S)-4-Hydroxy-5,8-dimethoxy-1,3-dimethyliso-
chromane 25
rel-(2R,4S,5R)-4-(2Ј-Chloro-5Ј-methoxyphenyl)-2,5-dimethyl-
1,3-dioxolane 13
To the dioxolane 1 (55 mg, 0.22 mmol) in dry methylene
dichloride (35 ml) at Ϫ78 ЊC was added titanium tetrachloride
(25.3 µl, 0.22 mmol). After stirring at Ϫ78 ЊC for 30 min, the
mixture was quenched with methanol (0.1 ml) and saturated
aqueous sodium hydrogen carbonate (0.5 ml) was added. The
resultant mixture was poured into water, the organic layer sep-
arated and the aqueous layer extracted with methylene di-
chloride (3 × 10 ml). The residue obtained upon work-up was
chromatographed (25% ethyl acetate–hexane) to give unreacted
dioxolane 1 (11 mg, 20%) (1H NMR and mass spectra in agree-
ment with those quoted above) followed by the isochromane 25
as a colourless oil (10 mg, 18%) (Found: Mϩ, 238.1205.
C13H18O4 requires M, 238.1205); δH 6.82 and 6.72 (each 1H, d,
J 9.0, 6-H and 7-H), 5.01 (1H, q, J 6.6, 1-H), 4.55 (1H, d, J 7.9,
4-H), 3.98 (1H, dq, J 7.9 and 6.2, 3-H), 3.85 and 3.76 (each
3H, s, OCH3), 1.55 (3H, d, J 6.6, 1-CH3) and 1.49 (3H, d, J 6.2,
3-CH3); m/z 238 (Mϩ, 15%), 223 (53), 220 (15), 205 (100), 195
(16), 194 (46) and 179 (67). Subsequent fractions afforded the
(1:1) mixture of the chlorohydrins 12 as a colourless oil (25 mg,
45%); δH 7.15 and 7.05 (each 1H, d, J 2.1, 6Ј-H of each isomer),
6.82 (4H, m, 3Ј-H and 4Ј-H of both isomers), 5.43 and 5.42
(each 1H, d, J 6.0 and 6.4, 1-H of each isomer), 4.15 (2H, dq
overlapped, 2-H of both isomers), 3.80 and 3.76 (each 6H,
2 × OCH3 of each isomer) and 1.25 and 1.14 (each 3H, d, J 6.0
and 6.4, CH3 of each isomer). GC–MS analysis of the chloro-
hydrin mixture 12 showed six components: two isomeric chloro-
hydrins 12 with retention times 13.4 and 13.5 min and identical
spectra, m/z 232 [Mϩ(37Cl), 10%], 230 [Mϩ(35Cl), 31], 194 (27),
188 (28), 186 (84), 171 (36) and 151 (100); the epoxides 9 and 10
with retention times 11.4 and 11.6 min and identical spectra,
m/z 194 (Mϩ, 98%), 179 (13), 165 (47) and 151 (100); the ketone
26 with retention time 11.5 min and m/z 194 (Mϩ, 56%), 165
(100) and 150 (60); and the ketone 27 with retention time 11.8
min and m/z 194 (Mϩ, 80%), 151 (100) and 121 (53).
The diol 21 (250 mg, 1.16 mmol) in dry methylene dichloride
(25 ml) was treated with 1,1-dimethoxyethane (0.16 ml) and
( )-camphorsulfonic acid (10 mg, 0.04 mmol) as for the prepar-
ation of dioxolane 1. The crude product was purified by chrom-
atography (10% ethyl acetate–hexane) to give the dioxolane 13
as a colourless oil (260 mg, 92%) (Found: Mϩ, 242.0711.
C12H15ClO3 requires M, 242.0710); δH 7.30 (1H, d, J 8.7, 3Ј-H),
7.17 (1H, d, J 3.1, 6Ј-H), 6.85 (1H, dd, J 8.7 and 3.1, 4Ј-H), 5.48
(1H, d, J 7.2, 4-H), 5.26 (1H, q, J 4.7, 2-H), 4.60 (1H, dq, J 7.2
and 6.4, 5-H), 3.89 (3H, s, OCH3), 1.64 (3H, d, J 4.7, 2-CH3)
and 0.95 (3H, d, J 6.4, 5-CH3); δC 158.3 (C-5Ј), 137.2 (C-2Ј),
129.6 (C-3Ј), 123.0 (C-1Ј), 114.3 and 113.6 (C-4Ј, C-6Ј), 100.0
(C-2), 76.5 and 75.2 (C-4, C-5), 55.4 (OCH3), 19.7 (CH3-2) and
16.3 (CH3-5); m/z 244 [Mϩ(37Cl), 3%], 242 [Mϩ(35Cl), 9], 200
(18), 198 (55), 169 (40), 167 (77), 154 (14), 119 (38), 77 (16), 72
(100) and 44 (78).
rel-(1R,2R)-1-(2Ј-Chloro-5Ј-methoxyphenyl)propane-1,2-diol 22
The 1:13 mixture of olefines 17 and 18 (200 mg, 1.08 mmol) in
a 2:1 mixture of acetone–water (6 ml) was treated with N-
methylmorpholine N-oxide (155 mg, 1.32 mmol) and osmium
tetroxide (5 mg) in tert-butyl alcohol (0.5 ml) at 0 ЊC. After
stirring for 24 h, acetone was removed under vacuum at room
temperature. The remaining aqueous layer was poured into
dilute hydrochloric acid (2 , 5 ml) and extracted into ethyl
acetate (5 × 20 ml). The residue obtained upon work-up was
chromatographed (50% ethyl acetate–hexane) to give the diol
22 as a light orange oil (200 mg, 80%) (Found: Mϩ,
216.0553. C10H13ClO3 requires M, 216.0552); δH 7.25 (1H,
d, J 8.9, 3Ј-H), 7.03 (1H, d, J 3.0, 6Ј-H), 6.77 (1H, dd, J 8.9 and
3.0, 4Ј-H), 4.90 (1H, d, J 5.5, 1-H), 3.92 (1H, dq, J 5.5 and 6.6,
2-H), 3.78 (3H, s, OCH3) and 1.19 (3H, d, J 6.6, CH3); δC 158.5
(C-5Ј), 139.8 (C-2Ј), 130.1 (C-3Ј), 123.8 (C-1Ј), 114.7 and 113.2
(C-4Ј, C-6Ј), 74.4 (C-1), 71.5 (C-2), 55.5 (OCH3) and 18.7
(CH3); m/z 218 [Mϩ(37Cl), 1%], 216 [Mϩ(35Cl), 3], 174 (27), 172
(87), 145 (16), 143 (55), 137 (47), 109 (50), 108 (69), 77 (58) and
44 (100).
rel-(1S,3R,4R)-4-Hydroxy-5,8-dimethoxy-1,3-dimethyliso-
chromane 30
To the 2:1 mixture of dioxolanes 2 and 3 (40 mg, 0.22 mmol) in
dry methylene dichloride (15 ml) at Ϫ78 ЊC was added titanium
tetrachloride (18.4 µl, 0.22 mmol). After stirring at Ϫ78 ЊC for
30 min, the reaction mixture was quenched with methanol (0.1
ml) and saturated aqueous sodium hydrogen carbonate (2 ml)
was added. The resultant mixture was poured into water, the
organic layer separated and the aqueous layer extracted with
methylene dichloride (3 × 10 ml). The combined organic
extracts were dried and evaporated to give a colourless oil (38
mg), shown by GC and GC–MS analysis to contain a number
of products. This mixture was chromatographed (25% ethyl
acetate–hexane) to give several compounds: unreacted dioxol-
anes 2 and 3 (8 mg, 20%) (1H NMR and mass spectra in agree-
ment with those quoted above); the chlorohydrin mixture 12 as
a colourless oil (12 mg, 31%) (1H NMR and GC–MS data are
given above); and the isochromane 30 as a colourless oil (18 mg,
45%) (Found: C, 65.6; H, 7.9. C13H18O4 requires C, 65.5; H,
7.6%); δH 6.77 and 6.73 (each 1H, d, J 8.5, 6-H and 7-H), 4.93
(1H, q, J 6.3, 1-H), 4.58 (1H, d, J 1.6, 4-H), 3.84 and 3.78 (each
3H, s, OCH3), 3.56 (1H, dq, J 1.6 and 6.3, 3-H), 1.58 (3H, d,
J 6.3, 1-CH3) and 1.39 (3H, d, J 6.3, 3-CH3); δC 151.5 and 150.2
(C-5, C-8), 129.3 and 126.9 (C-4a, C-8a), 112.3 and 108.9 (C-6,
C-7), 72.2, 71.1 and 63.8 (C-1, C-3, C-4), 55.8 and 55.6 (OCH3),
21.6 (1-CH3) and 16.7 (3-CH3); m/z 238 (Mϩ, 13%), 223 (7), 220
(26), 206 (13), 205 (100), 194 (56), 190 (20), 179 (16) and 175
(30).
rel-(2R,4R,5R)- and rel-(2S,4R,5R)-4-(2Ј-Chloro-5Ј-methoxy-
phenyl)-2,5-dimethyl-1,3-dioxolanes 14 and 15
Diol 22 (200 mg, 0.93 mmol) in dry methylene dichloride (20
ml) was treated with 1,1-dimethoxyethane (0.10 ml) and ( )-
camphorsulfonic acid (10 mg, 0.043 mmol) as for the prepar-
ation of the dioxolane 1. The crude product (200 mg) was
chromatographed (10% ethyl acetate–hexane) to give a 2:1
diastereomeric mixture of dioxolanes 14 and 15 as a colourless
oil (190 mg, 85%). Further chromatography of this mixture
(10% ethyl acetate–hexane) afforded first the single dioxolane
14 as a colourless oil (105 mg, 55%) (Found: C, 59.2; H, 6.2;
Cl, 14.6. C12H15ClO3 requires C, 59.4; H, 6.2; Cl, 14.6%); δH
7.24 (1H, d, J 8.8, 3Ј-H), 7.03 (1H, d, J 3.1, 6Ј-H), 6.77 (1H,
dd, J 8.8 and 3.1, 4Ј-H), 5.49 (1H, q, J 4.7, 2-H), 4.97 (1H, d, J
7.4, 4-H), 3.91 (1H, dq, J 7.4 and 6.1, 5-H), 3.80 (3H, s,
OCH3), 1.49 (3H, d, J 6.1, 5-CH3) and 1.46 (3H, d, J 4.7, 2-
CH3); δC 158.6 (C-5Ј), 138.0 (C-2Ј), 130.3 (C-3Ј), 123.7 (C-1Ј),
114.5 and 112.9 (C-4Ј, C-6Ј), 102.2 (C-2), 81.5 and 80.4 (C-4
and C-5), 55.5 (OCH3), 20.8 (CH3-2) and 17.5 (CH3-5); m/z
244 [Mϩ(37Cl), 2%], 242 [Mϩ(35Cl), 6], 200 (9), 198 (28), 169
(19), 167 (33), 154 (14), 119 (15), 77 (16), 72 (80) and 42 (100).
This was followed by the dioxolane 15 as a colourless oil (40
mg, 21%); δH 7.24 (1H, d, J 8.8, 3Ј-H), 7.14 (1H, d, J 3.1, 6Ј-
H), 6.78 (1H, dd, J 8.8 and 3.1, 4Ј-H), 5.42 (1H, q, J 4.8, 2-H),
3368
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
rel-(1S,3R,4R)-4-Hydroxy-1,3-dimethylisochromane-5,8-
quinone 31
stirred soluton of a 2:1 mixture of dioxolanes 14 and 15 (50
mg, 0.2 mmol) in dry methylene dichloride (60 ml) at Ϫ78 ЊC
in an atmosphere of argon. After 30 min the reaction was
quenched with methanol (0.1 ml). The resultant solution was
neutralised with saturated aqueous sodium hydrogen carbon-
ate, washed with water, dried and evaporated. Analysis of
the resulting crude residue by GC and 1H NMR spectro-
scopy indicated a 1:2 mixture of isochromanes 35 and 36.
Chromatography (20% ethyl acetate–hexane) afforded first the
isochromane 35 as a colourless oil (13 mg, 26%) (Found: C,
59.35; H, 6.4; Cl, 14.95. C12H15ClO3 requires C, 59.4; H, 6.2; Cl,
14.6%); δH 7.28 (1H, d, J 8.8, 6-H), 6.78 (1H, d, J 8.8, 7-H), 4.93
(1H, q, J 6.2, 1-H), 4.56 (1H, d, J 1.2, 4-H), 3.81 (3H, s, OCH3),
3.68 (1H, dq, J 1.2 and 6.3, 3-H), 1.56 (3H, d, J 6.2, 1-CH3) and
1.40 (3H, d, J 6.3, 3-CH3); δC 154.7 (C-8), 135.1 (C-5), 130.0,
127.9 and 125.7 (C-4a, C-6, C-8a), 111.0 (C-7), 72.0, 71.0 and
66.1 (C-1, C-3, C-4), 55.3 (OCH3), 21.6 (1-CH3) and 16.8
(3-CH3); m/z 244 [Mϩ(37Cl), 2.5%], 242 [Mϩ(35Cl), 8], 229 (8),
227 (25), 211 (11), 209 (30), 200 (31), 198 (100), 169 (52). This
was followed by the isochromane 36 as white crystals (25 mg,
50%), mp 107–108 ЊC (methylene dichloride–hexane) (Found:
C, 59.7; H, 6.4; Cl, 14.4. C12H15ClO3 requires C, 59.4; H, 6.2;
Cl, 14.6%); δH 7.28 (1H, d, J 8.8, 6-H), 6.75 (1H, d, J 8.8, 7-H),
5.09 (1H, q, J 6.6, 1-H), 4.50 (1H, d, J 1.7, 4-H), 4.11 (1H, dq,
J 1.7 and 6.5, 3-H), 3.81 (3H, s, OCH3), 1.48 (3H, d, J 6.6,
1-CH3) and 1.40 (3H, d, J 6.5, 3-CH3); δC 153.9 (C-8), 133.9
(C-5), 129.7, 128.1 and 126.0 (C-4a, C-6, C-8a), 110.6 (C-7),
68.5, 66.2 and 65.0 (C-1, C-3, C-4), 55.4 (OCH3), 17.8 (1-CH3)
and 16.9 (3-CH3); m/z 244 [Mϩ(37Cl), 4%], 242 [Mϩ(35Cl), 11],
229 (12), 227 (35), 211 (13), 209 (38), 200 (22), 198 (68) and 42
(100).
To the isochromane 30 (20 mg, 0.08 mmol) and silver() oxide
(50 mg) in dioxolane (2 ml) was added nitric acid (6 , 0.2 ml).
The reaction mixture was stirred for 5 min and a mixture of
chloroform (6 ml) and water (2 ml) was then added. The residue
obtained upon work-up was chromatographed (20% ethyl
acetate–hexane) to give the quinone 31 as a yellow oil (13 mg,
75%); δH 6.80 and 6.74 (each 1H, d, J 10.0, 6-H and 7-H), 4.64
(1H, dq, J 1.5 and 6.0, 1-H), 4.38 (1H, dd, J 1.6 and 1.5, 4-H),
3.57 (1H, dq, J 1.6 and 6.3, 3-H), 1.53 (3H, d, J 6.0, 1-CH3) and
1.37 (3H, d, J 6.3, 3-CH3); δC 187.3 and 186.3 (C-5 and C-8),
145.3 and 139.8 (C-4a, C-8a), 137.3 and 136.1 (C-6, C-7), 72.2,
69.8 and 61.5 (C-1, C-3, C-4), 20.3 (1-CH3) and 15.9 (3-CH3);
m/z 210 (Mϩ ϩ 2H, 10%), 208 (5), 193 (27), 164 (100) and 136
(64).
5,8-Dimethoxy-1,3-dimethyl-1H-isochromene 32
The 2:1 mixture of dioxolanes 2 and 3 (40 mg, 0.22 mmol) was
treated with titanium tetrachloride as above and the reaction
mixture stirred at 0 ЊC for 10 min before quenching with
methanol (0.1 ml). 1H NMR analysis of the crude residue
showed a 3:1 mixture of the chlorohydrins 12 (20 mg, 38%)
and the isochromene 32 (20 mg, 36%) (Found: Mϩ, 220.1099.
C13H16O3 requires M, 220.1099); δH 6.65 and 6.58 (each 1H, d,
J 9.0, 6-H and 7-H), 5.81 (1H, s, 4-H), 5.61 (1H, q, J 6.0, 1-H),
3.79 and 3.76 (each 3H, s, OCH3), 1.91 (3H, s, 3-CH3) and 1.35
(3H, d, J 6.0, 1-CH3); m/z 220 (Mϩ, 9%), 205 (21), 149 (27) and
57 (100).
rel-(1R,3R,4S)-5-Chloro-4-hydroxy-8-methoxy-1,3-dimethyl-
isochromane 34
Titanium tetrachloride (47.0 µl, 0.4 mmol) was added to a
stirred solution of dioxolane 13 (50 mg, 0.20 mmol) in dry
methylene dichloride (33 ml) at Ϫ95 ЊC in an atmosphere of
argon. After 15 min an aliquot (10 ml) of the reaction mixture
was removed by syringe and immediately quenched with
methanol (0.1 ml). The rest of the reaction mixture was kept at
Ϫ95 ЊC for 60 min before quenching with methanol (0.1 ml).
Duplicate portions of dioxolane 13 (15 mg, 0.06 mmol) in
methylene dichloride (10 ml) were stirred with titanium tetra-
chloride (14.1 µl, 0.12 mmol) at Ϫ78 ЊC for 30 min in an atmos-
phere of argon before quenching with methanol (0.1 ml) at
Ϫ78 ЊC. To a fourth portion (15 mg, 0.06 mmol) at lower con-
centration in methylene dichloride (20 ml) at Ϫ78 ЊC was added
titanium tetrachloride (14.1 µl, 0.12 mmol), and the temper-
ature raised to 0 ЊC for 30 min before quenching. The quenched
reactions were neutralised with saturated aqueous sodium
hydrogen carbonate, and the organic phases washed with water,
dried and evaporated. Analysis of the crude residues by GC and
1H NMR spectroscopy indicated mixtures of the dioxolane 13,
diol 21 and isochromane 34 (see Table 1).
The residues obtained from the duplicate experiments at
Ϫ78 ЊC were combined and purified by chromatography on
silica gel (25% ethyl acetate–hexane) to afford isochromane 34
as white crystals (23 mg, 77%), mp 108–110 ЊC (methylene
dichloride–hexane) (Found: C, 59.2; H, 6.0; Cl, 14.9.
C12H15ClO3 requires C, 59.4; H, 6.2; Cl, 14.6%); δH 7.25 (1H, d,
J 8.8, 6-H), 6.75 (1H, d, J 8.8, 7-H), 4.96 (1H, q, J 6.3, 1-H),
4.57 (1H, d, J 6.0, 4-H), 4.16 (1H, dq, J 6.0 and 6.5, 3-H), 3.81
(3H, s, OCH3), 1.56 (3H, d, J 6.3, 1-CH3) and 1.27 (3H, d, J 6.5,
3-CH3); δC 153.9 (C-8), 133.0 (C-5), 130.5, 128.2 and 125.7
(C-4a, C-6, C-8a), 110.5 (C-7), 69.9, 68.3 and 66.2 (C-1, C-3,
C-4), 55.4 (OCH3), 19.6 (3-CH3) and 17.6 (1-CH3); m/z 244
[Mϩ(37Cl), 8%], 242 [Mϩ(35Cl), 24], 229 (35), 227 (100), 211 (13),
209 (34), 200 (19), 198 (58), 183 (56) and 169 (29).
Isomerisation of dioxolane 14
To the dioxolane 14 (100 mg, 0.41 mmol) in stirred dry
methylene dichloride (66 ml) was added titanium tetrachloride
(75.2 µl, 0.80 mmol) at Ϫ95 ЊC in an atmosphere of argon.
Portions (5 ml) of the reaction mixture were removed by syringe
and immediately quenched with methanol (0.1 ml) after 2 min
and 12 min. The rest of the reaction mixture was warmed to
Ϫ78 ЊC, kept for 30 min and then quenched with methanol (0.1
ml).
Further aliquots of the dioxolane 14 (15 mg, 0.06 mmol) in
methylene dichloride (10 or 20 ml) were stirred with titanium
tetrachloride (14.1 µl, 0.12 mmol) and quenched under the con-
ditions of temperature and time detailed in Table 2. The
quenched reaction solutions were neutralised with saturated
aqueous sodium hydrogen carbonate, washed with water, dried
and evaporated. The residues were analysed by GC, GC–MS
1
and H NMR spectroscopy with the results given in Table 2.
The isochromene 38 formed in entries 7 and 8 was identified by
GC–MS (Found: Mϩ, 226/224. C12H13ClO2 requires M, 226/
224); m/z 226 [Mϩ(37Cl), 8%], 224 [Mϩ(35Cl), 25], 211 (34), 209
(100), 196 (7) and 194 (22).
Isomerisation of dioxolane 15
Aliquots of the dioxolane 15 (15 mg, 0.06 mmol) in methylene
dichloride (10 or 20 ml) were stirred with titanium tetrachloride
(14.1 µl, 0.12 mmol) and quenched under the conditions of
temperature and time detailed in Table 3. The quenched reac-
tion solutions were neutralised with saturated aqueous sodium
hydrogen carbonate, washed with water, dried and evaporated.
The residues were analysed by GC, GC–MS and 1H NMR
spectroscopy with the results given in Table 3.
Attempted isomerisation of isochromane 35
Isochromane 35 (15 mg, 0.06 mmol) in methylene dichloride (10
ml) was treated with titanium tetrachloride (14.1 µl, 0.12 mmol)
at Ϫ78 ЊC. After 2 min the reaction mixture was warmed to 0 ЊC
and stirred for 30 min in an atmosphere of argon before
quenching with methanol (0.1 ml) at 0 ЊC. The residue obtained
rel-(1S,3R,4R)- and rel-(1R,3R,4R)-5-Chloro-4-hydroxy-8-
methoxy-1,3-dimethylisochromanes 35 and 36
Titanium tetrachloride (47.0 µl, 0.4 mmol) was added to a
J. Chem. Soc., Perkin Trans. 1, 1997
3369

View Article Online
upon work-up was shown by GC, GC–MS and 1H NMR
spectroscopy to contain unchanged isochromane 35.
III, Recent Advances, Chapman and Hall, London, 1987; (c) R. H.
Thomson, Naturally Occurring Quinones IV, Recent Advances, 4th
edn., Blackie Academic and Professional, Chapman and Hall,
London, 1997.
3 D. W. Cameron, R. I. T. Cromartie, D. G. I. Kingston and Lord
Todd, J. Chem. Soc., 1964, 51.
4 R. G. F. Giles, V. R. Lee Son and M. V. Sargent, Aust. J. Chem.,
1990, 43, 777.
5 (a) V. VanRheenen, R. C. Kelly and D. Y. Cha, Tetrahedron Lett.,
1976, 1973; (b) R. Ray and D. S. Matteson, Tetrahedron Lett., 1980,
21, 449.
6 G. Casiraghi, M. Cornia and G. Rassu, J. Org. Chem., 1988, 53,
4919.
7 M. K. Meilahn, C. N. Statham, J. L. McManaman and M. E.
Munk, J. Org. Chem., 1975, 40, 3551.
8 A. Bhati, J. Chem. Soc., 1963, 730.
9 T. Kometani, Y. Takeuchi and E. Yoshii, J. Chem. Soc., Perkin Trans.
1, 1981, 1197.
10 R. G. F. Giles, I. R. Green, V. I. Hugo, P. R. K. Mitchell and
S. C. Yorke, J. Chem. Soc., Perkin Trans. 1, 1983, 2309.
11 D. W. Cameron, D. G. I. Kingston, N. Sheppard and Lord Todd,
J. Chem. Soc., 1964, 98.
12 R. G. F. Giles, I. R. Green, V. I. Hugo, P. R. K. Mitchell and
S. C. Yorke, J. Chem. Soc., Perkin Trans. 1, 1984, 2383.
13 M. Karplus, J. Chem. Phys., 1960, 33, 1842.
14 P. A. Bartlett, W. S. Johnson and J. D. Elliott, J. Am. Chem. Soc.,
1983, 105, 2088. For a discussion of related effects, see S. E.
Denmark and N. G. Almstead, J. Am. Chem. Soc., 1991, 113, 8089.
15 R. G. F. Giles, I. R. Green, L. S. Knight, V. R. Lee Son and
S. C. Yorke, J. Chem. Soc., Perkin Trans. 1, 1994, 865.
16 A. L. J. Beckwith and S. H. Goh, J. Chem. Soc., Chem. Commun.,
1983, 907.
rel-(1R,3R,4R)-1,3-Dimethyl-4-hydroxy-8-methoxyisochromane
37
Following the procedure of Beckwith and Goh,16 a mixture of
the isochromane 36 (10 mg, 0.04 mmol), di-tert-butyl peroxide
(4 µl, 0.02 mmol) and lithium aluminium hydride (50 mg) in dry
tetrahydrofuran (10 ml) was irradiated with a 250 W high-
pressure Hg lamp for 5 h, cooled, diluted with aqueous hydro-
chloric acid and extracted with diethyl ether. The residue
obtained upon work-up gave a colourless oil (5 mg, 58%) iden-
1
tified by GC–MS and H NMR spectroscopy as the isochro-
mane 37 (Found: Mϩ, 208. C12H16O3 requires M, 208); δH 7.25
(1H, t, J 8.8, 6-H), 7.00 and 6.80 (each 1H, d, J 8.8, 5-H and
7-H), 5.09 (1H, q, J 6.6, 1-H), 4.18 (1H, d, J 1.7, 4-H), 4.11
(1H, dq, J 1.7 and 6.5, 3-H), 3.81 (3H, s, OCH3), 1.50 (3H, d,
J 6.6, 1-CH3) and 1.45 (3H, d, J 6.5, 3-CH3); m/z 208 (Mϩ,
15%), 193 (27), 175 (45) and 164 (100).
Acknowledgements
We gratefully acknowledge receipt of an Australian Post-
graduate Research Award (B. S. S.), expert technical assistance
from Mr A. J. Herlt, and the provision of mass spectra by
Ms J. M. Rothschild.
17 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923.
References
1 R. G. F. Giles, R. W. Rickards and B. Senanayake, J. Chem. Soc.,
Perkin Trans. 1, 1996, 2241.
2 (a) R. H. Thomson, Naturally Occurring Quinones, Academic Press,
London, 1971; (b) R. H. Thomson, Naturally Occurring Quinones
Paper 7/04543C
Received 27th June 1997
Accepted 31st July 1997
3370
J. Chem. Soc., Perkin Trans. 1, 1997