Synthesis of cyclohexenols and cycloheptenols via the regioselective reductive ring opening of oxabicyclic compounds

Article abstract of DOI:10.1021/ja964361j

The reductive ring opening of oxabicyclic compounds has been achieved. While RMgBr/MgBr₂ works in a few limited substrates, diisobutylaluminum hydride reacts with oxabicyclic[3.2.1]- and -[2.2.1]alkenes to provide cycloheptenols and cyclohexeno

Full text of DOI:10.1021/ja964361j

6478
J. Am. Chem. Soc. 1997, 119, 6478-6487
Synthesis of Cyclohexenols and Cycloheptenols via the
Regioselective Reductive Ring Opening of Oxabicyclic
Compounds
Mark Lautens,* Shihong Ma, and Pauline Chiu
Contribution from the Department of Chemistry, UniVersity of Toronto, Toronto,
Ontario, Canada M5S 3H6
ReceiVed December 19, 1996X
Abstract: The reductive ring opening of oxabicyclic compounds has been achieved. While RMgBr/MgBr2 works
in a few limited substrates, diisobutylaluminum hydride reacts with oxabicyclic[3.2.1]- and -[2.2.1]alkenes to provide
cycloheptenols and cyclohexenols in good yield and in some cases in good regioselectivity. With some substrates
further reduction of the alkene was observed which led us to examine transition metals such as nickel which are
known to accelerate the hydroalumination reaction. The reaction with Ni(COD)2 (COD ) cyclooctadiene) gave the
best reactivity-selectivity profile, and significantly higher yields were obtained with minimal overreduction of the
alkene. Phosphine ligands were shown to dramatically improve the regioselectivity of ring opening of bridgehead-
substituted oxabicyclic compounds. The best ligand was 1,4-bis(diphenylphosphino)butane which gave up to 380:1
selectivity. A series of deuterium quenching experiments showed that the selectivity of the hydroalumination varies
according to the reaction protocol and ligand-metal ratio.
Transition metal catalyzed hydrometalations play an important
novel routes to natural products with diverse stereochemical
motifs.6 While carbon nucleophiles were found to efficiently
react in an SN2′ fashion, there existed no efficient source of
hydride to efficiently react with these systems.7-9a We therefore
investigated the possibility of using a hydrometalation-elimina-
tion sequence to achieve this goal. We envisaged that the
initially produced organometallic species would undergo â-e-
limination to form the ring-opened compound (eq 1).9,10 We
now describe our successful development of an efficient
reducing system including a determination of the reactivity and
regioselectivity of the reaction and present our preliminary
studies relating to the mechanism.
role in organic synthesis due to the utility of the resulting
organometallic compounds.1 The combination of the metal
catalyst and its accompanying ligands offers the opportunity to
tune the reactivity and prepare enantiomerically enriched
products. This strategy is amply illustrated by rhodium- and
ruthenium-catalyzed hydroborations, palladium- and rhodium-
catalyzed hydrosilations, and palladium-, rhodium-, and mo-
lybdenum-catalyzed hydrostannations.2-4
Nickel-catalyzed hydroaluminations of unsaturated hydro-
carbons have been known for many years, particularly the
hydroalumination of alkynes to yield vinylic alanes. However,
the utility of the hydroalumination of alkenes has been limited
due to the low reactivity of the alkylalane and the propensity
of the carbon aluminum bond to undergo racemization.5
Oxabicyclo[3.2.1]octenes and oxabicyclo[2.2.1]heptenes have
been used as templates to control the regio- and stereoselectivity
in the synthesis of polyfunctionalized 7- and 6-membered rings.
A variety of ring opening strategies have been reported providing
Results and Discussions
X Abstract published in AdVance ACS Abstracts, June 1, 1997.
(1) The Chemistry of the Metal Bonds; Hartley, F. R., Ed.; John Wiley
and Sons: New York, 1987; Vol. 4.
(2) Metal-catalyzed hydroboration: Mannig, D.; Noth, H. Angew. Chem.,
Int. Ed. Engl. 1985, 24, 878. For a review, see: Burgess, K.; Ohlmeyer,
M. J. Chem. ReV. 1991, 91, 1179. For discussions on the mechanism, see:
Evans, D. A.; Fu, G. C.; Anderson, B. A. J. Am. Chem. Soc. 1992, 114,
6679 and references therein.
(3) For recent reviews of metal-catalyzed hydrosilation, see: (a) Hiyama,
T.; Kusumoto, T. In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Oxford, 1991; Vol. 8, p 763. (b) Mikolajczyk, M.;
Drabowicz, J.; Kielbasinski, P. StereoselectiVe Synthesis; Houben-Weyl,
4th ed.; E21e, p 5733.
(4) Metal-catalyzed hydrostannation: (a) Ichinose, Y.; Oda, H.; Oshima,
K.; Utimoto, K. Bull. Chem. Soc. Jpn. 1987, 60, 3468. (b) Kikukawa, K.;
Umekawa, H.; Wada, F.; Matsuda, T. Chem. Lett. 1988, 881. (c) Zhang,
H. X.; Guibe, F.; Balavoine, G. J. Org. Chem. 1990, 55, 1857. (d) Miyake,
H.; Yamamuram, K. Chem. Lett. 1991, 1099. (e) Koerber, K.; Gore, J.;
Vatele, J.-M. Tetrahedron Lett. 1991, 32, 1187. (f) Casson, S.; Kocienski,
P. Synthesis 1993, 1133. (g) Lautens, M.; Kumanovic, S.; Meyer, C. Angew.
Chem., Int. Ed. Engl. 1996, 35, 1329 and references therein.
(5) For a recent review of alkene and alkyne hydroalumination, see:
Eisch, J. J. In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Oxford, 1991; Vol. 8, p 733.
Identifying Diisobutylaluminum Hydride (DIBAL-H) as
a Good Reducing Reagent. The first system we identified for
the reductive ring opening was the combination of an organo-
magnesium or organolithium reagent bearing â-hydrogens and
excess MgBr2.9b In general, these reactions are sluggish,
proceed in moderate to good yield, and are mainly limited to
(6) For reviews describing the synthesis, reactions, and utility of
oxabicyclic compounds, see: (a) Chiu, P.; Lautens, M. Top. Curr. Chem.
1997, 190, 3. (b) Keay, B. A.; Woo, S. Synthesis 1996, 669. (c) Shipman,
M. Contemp. Org. Synth. 1995, 2 (1), 1. (d) Lautens, M. Synlett 1993, 177.
(e) Ager, D. J.; East, M. B. Tetrahedron 1993, 49, 5683. (f) Lautens, M.
Pure Appl. Chem. 1992, 64, 1873 and references therein. (g) Vogel, P.;
Fattori, D.; Gasparini, F.; Le Drian, C. Synlett 1990, 173. (h) Harmata, M.
In AdVances in Cycloadditions; Lautens, M., Ed.; JAI Press: Greenwich,
CT; Vol. 4, p 41. (i) West, F. G. Ibid; p 1.
(7) Lautens, M.; Ma, S.; Yee, A. Tetrahedron Lett. 1995, 36, 4185.
(8) Lautens, M.; Belter, R. K. Tetrahedron Lett. 1992, 33, 2617.
(9) (a) Lautens, M.; Chiu, P.; Colucci, J. T. Angew. Chem., Int. Ed. Engl.
1993, 32, 281. (b) Lautens, M.; Chiu, P. Tetrahedron Lett. 1991, 32, 4827.
(10) Lautens, M.; Chiu, P.; Ma, S.; Rovis, T. J. Am. Chem. Soc. 1995,
117, 532.
S0002-7863(96)04361-2 CCC: $14.00 © 1997 American Chemical Society

ReductiVe Ring Opening of Oxabicyclic Compounds
Table 1. Reductive Ring-Opening Reactions Using DIBAL-H
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6479
Table 2. Regioselectivity of DIBAL-H Reductive Ring Opening
a DIBAL-H (5 equiv) in refluxing hexanes. b DIBAL-H (6 equiv)
in refluxing hexanes. c MeLi (1.2 equiv) then 5 equiv of DIBAL-H in
1
refluxing hexanes. d Ratio was determined by H NMR. e Ratio was
determined by GC. f Yield of both regioisomers. g Isolated yield of
product obtained along with 53% recovered starting material.
a DIBAL-H (6 equiv) in refluxing hexanes. b DIBAL-H (6 equiv)
in refluxing ether. c Yields of isolated and analytically pure material.
d Could not be separated from 16% of overreduced product, yield
estimated from the 1H NMR spectrum of the mixture. e 10% of
overreduced product.
head substituents gave exclusively the cycloheptenol 11 in 89%
yield without any overreduction product (entry 3). For com-
pound 12, the additional reactivity of the bicyclic[2.2.1] structure
allowed the reductive ring opening to occur under milder
conditions (entry 4).
oxabicyclic[2.2.1] compounds. Furthermore, low regioselec-
tivity is observed for unsymmetrical substrates (eq 2).
Regioselectivity of DIBAL-H-Promoted Ring Opening.
We also examined the regioselectivity of reductive ring opening
of unsymmetrical [3.2.1] substrates. As shown in Table 2, the
trisubstituted cycloheptenol a will be obtained if the hydride
adds distal to the bridgehead substituent, whereas the disubsti-
tuted cycloheptenol b will be produced if the hydride attacks
the position of the double bond proximal to the substituent.
When the hydroxyl group in the substrate was protected, the
reaction favored the formation of regioisomer b which resulted
from the delivery of the bulky alkylaluminum to the less-
hindered reaction site distal to the bridgehead substituent in
accord with the regioselectivity anticipated for hydroalumination
(entries 1 and 2). In contrast, treatment of 18 with DIBAL-H
alone resulted in virtually no selectivity (entry 3). However,
when 18 was deprotonated with MeLi prior to hydroalumination,
the major product was the regioisomer 19a (entry 4). This
reversal of regioselectivity demonstrated that the remote endo
alkoxide group exerts a directing effect on the mode of the
hydroalumination reaction. It is worth noting that the double
bonds in oxabicyclic compounds with an endo alkoxide function
at C3 are more reactive toward organolithium reagents providing
further support of an endo alkoxide activation.8
The limitations of the RM-MgBr2 system as a general
method to achieve the reductive ring opening led us to seek a
more efficient and versatile reducing agent. Inspired by
Katzenellenbogen’s report of the reduction of vinyl epoxides
using DIBAL-H (eq 3), we examined these conditions with the
oxabicyclic substrates.10,11
We showed that treatment of 3 with 7.0 equiv of DIBAL-H
in hexanes at reflux afforded the ring-opened product 4 in 50%
yield along with the overreduction product 5 in 27% yield (eq
4).6b,9a,10,12 Decreasing the amount of DIBAL-H to 1.5 equiv
The high regioselectivity of the ring opening of unsym-
metrical substrates required an endo hydroxyl group at C3.
Moreover, the overreduction following the ring opening ob-
served in some of the symmetrical substrates and the potential
of an enantioselective ring opening prompted us to investigate
transition metal catalyzed hydroalumination processes.
Ni(COD)2-Catalyzed Reductive Ring Opening (COD )
cyclooctadiene). An early study of hydroalumination of alkenes
by triisobutylaluminum reported that the most active catalysts
are Ti(IV), Zr(IV), and Ni(0), and to a lesser extent, Co(II),
Fe(III), and Rh(III).13 Hydroalumination of olefins by lithium
aluminum hydride has traditionally been catalyzed by TiCl4 and
ZrCl4.14 It was also reported that dialkylaluminum hydride
addition to olefins are catalyzed by titanium and nickel.15
avoided the formation of 5 and gave 4 in 65% yield along with
25% recovered starting material. A number of oxabicyclo[2.2.1]
and -[3.2.1] compounds were shown to undergo ring opening
(Table 1). The amount of overreduced product varied as a
function of the substrate. For example, 6 which lacked
substituents gave a 64% yield of ring opening product and 16%
of overreduced product (entry 1). Compound 10 with bridge-
(11) Lenox, R. S.; Katzenellenbogen, J. A. J. Am. Chem. Soc. 1973, 95,
957.
(12) Keay also showed that excess DIBAL-H in CH2Cl2 leads to the
regioselective ring opening of an oxabicyclic substrate to give a ring-opened
product in 28% yield, see: Woo, S.; Keay, B. A. Tetrahedron Lett. 1992,
33, 2661.
Several metals including nickel, titanium, and zirconium were
examined for their ability to catalyze the hydroalumination. A
D2O quenching experiment was performed after treating sub-
(13) Asinger, F.; Fell, B.; Janssen, R. Chem. Ber. 1964, 97, 2515.

6480 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Lautens et al.
Table 3. Survey of Metals for Catalysis of Hydroalumination
Scheme 1
entry
catalyst (mol %)
reaction timea
yield (%)
D (%)b
1
2
3
4
5
6
7
Cp2ZrCl2 (5)
12 h
1 h
45 min
15 min
7 min
30 min
7 min
0
57
70
77
88
67
75
Cp2TiCl2 (10)
Cp2TiCl2 (60)
Ni(acac)2 (10)
Ni(acac)2 (60)
Ni(COD)2 (10)
Ni(COD)2 (60)
65
45
74
20
83
83
opened product 4 in one of two ways. Simply heating a solution
of 20 at 70 °C for 12-16 h led to a mixture of 4, 5, and 23.
Much better results were obtained when 20 was heated in the
presence of a Lewis acid. Diisobutylaluminum chloride (DIBAL-
Cl, 5 equiv) was the most efficient and gave 4 in 76% yield,
with minimal formation of 5 (<5%) arising from hydroalumi-
nation of 4.
a Reaction time of the metal-catalyzed hydroalumination before
workup. b Deuterium incorporation in the product as analyzed by H
1
NMR.
strate 3 with DIBAL-H and a fixed amount of catalyst in toluene
at room temperature (eq 5). The deuterium incorporation in
An example of the dramatic improvement in the efficiency
of the nickel catalyzed hydroalumination is illustrated in the
reaction of an unsubstituted oxabicyclic compound 24 (Scheme
2).7 Using 5 equiv of DIBAL-H in the absence of catalyst, a
1:1 mixture of the cycloheptene diol 25 and cycloheptane diol
26 was obtained. Addition of nickel avoided the formation of
the alkane 26 but the ring opening was inefficient and an
equimolar amount of 27 was isolated. However, the combina-
tion of a nickel catalyst and DIBAL-H then DIBAL-Cl gave
25 in 85% yield with only a trace of 27 which was easily
separated.
Effect of Phosphine Ligands on the Ring Opening of
Unsymmetrical Oxabicyclic Substrates. We had demonstrated
that nickel(0)-catalyzed hydroalumination of oxabicycles clearly
offered advantages over the noncatalyzed reaction, but several
issues needed to be probed, including the regioselectivity of
the ring opening in the case of unsymmetrical oxabicyclic
compounds. We found that with substrates lacking a bridgehead
substituent the hydroalumination of the alkene occurred readily
but the subsequent ring-opening step had to be carried out in
the presence of a Lewis acid and heat as mentioned above.
Conversely, ring opening occurred spontaneously for the
unsymmetrical substrates bearing a substituent on the bridgehead
position even though the initial rate of the hydroalumination
was reduced compared to that of the unsubstituted substrate
(Table 4). We observed that unsymmetrical oxabicycles 14 and
18 afforded two regioisomers in good yields under noncatalyzed
hydroalumination conditions with a preference for the tertiary
alcohols (entries 1 and 3). A greater proportion of product “a”
was isolated in the presence of 10 mol % of Ni(COD)2, but the
yield was unacceptably low and a product analogous to a
“hydrogenation” of the starting material was obtained (entries
2 and 4). (note in entry 2 the reaction still favors “b”).
The effect of adding ligands to the nickel-catalyzed reductive
ring opening was examined next. Eisch noted that the addition
of phosphines to the nickel-catalyzed hydroalumination of a
substituted indene significantly enhances the regioselectivity of
the reaction (Scheme 3).15c Presumably, the reaction proceeds
Via initial hydronickelation in which the hydride is delivered
to the more hindered site, and the more sterically encumbered
metal species resides in the more accessible position. However,
electronic effects also favor this regioisomer since formation
of a benzylnickel species should be preferred over an alkylnickel
compound.
the product 21 was determined by 1H NMR and was used as a
measure of the efficiency of the catalyst and the integrity of
the C-Al bond (Table 3). Exposure of the organoalane 20 to
oxygen gave the exo alcohol 22 in 42% yield, and this was
assumed to correlate to the stereochemistry of the hydroalumi-
nation step.16
Hydroalumination of 3 using a zirconium complex failed
which was consistent with Negishi’s observation that catalytic
Cp2ZrCl2 does not accelerate hydroalumination.17 Cp2TiCl2 and
Ni(acac)2 both gave moderate levels of 21 using 10 mol % of
catalyst (entries 2 and 4), and increasing the amount of Cp2-
TiCl2 or Ni(acac)2 led to lower deuterium incorporation (entries
3 and 5). These experiments indicated that the hydroaluminated
intermediate was protonated under the reaction conditions before
the D2O quench, and the extent of this internal protonation
depended on the amount of catalyst used. By whatever pathway
quenching occurs, it decreases the amount of hydroaluminated
substrate available to undergo ring opening to the desired
product. In contrast, high levels of deuterium incorporation
were consistently obtained with either high or low levels of Ni-
(COD)2 (entries 6 and 7). Ni(COD)2 was identified as the
catalyst of choice for effecting the hydroalumination step in
the overall reductive ring opening.
Hydroalumination of oxabicyclic alkenes occurs at temper-
atures as low as -78 °C in the presence of Ni(COD)2 (Scheme
1). Reaction of 3 with 1.1 equiv of DIBAL-H in the presence
of 5-15 mol % of Ni(COD)2 is complete within a few minutes
at room temperature, whereas reaction of 3 with 5 equiv of
DIBAL-H in the absence of nickel required extended heating
at 50-70 °C. The organoalane 20 was converted to the ring-
Of the monodentate ligands examined (Ph3P, Cy3P (Cy )
cyclohexyl), (o-tolyl)3P and (i-PrO)3P), Ph3P was found to give
the highest regioselectivity for isomer a. However, the presence
of the ligand was detrimental to the reaction rate, and a higher
ratio of nickel catalyst (31 mol %) to substrate was required
(Table 4, entry 5). We wondered if the excess DIBAL-H which
was present in the reaction mixture reacted with the catalyst
slowly to terminate the catalytic cycle.15d Thus, if we could
(14) (a) Sato, F.; Sato, S.; Kodama, H.; Sato, M. J. Organomet. Chem.
1977, 142, 71. (b) Isagawa, K.; Tatsumi, K.; Otsuji, Y. Chem. Lett. 1976,
1145.
(15) (a) Asinger, F.; Fell, B.; Theissen, F. Chem. Ber. 1967, 100, 937.
(b) Fischer, K.; Jonas, K.; Misbach, P.; Stabba, R.; Wilke, G. Angew. Chem.,
Int. Ed. Engl. 1973, 12, 943. (c) Eisch, J. J.; Fichter, K. C. J. Am. Chem.
Soc. 1974, 96, 6815. (d) Eisch, J. J.; Sexsmith, S. R.; Fichter, K. C. J.
Organomet. Chem. 1990, 382, 273.
(16) The catalyst used in this experiment was Ni(acac)2.
(17) Negishi, E.; Yoshida, T. Tetrahedron Lett. 1980, 21, 1501.

ReductiVe Ring Opening of Oxabicyclic Compounds
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6481
Scheme 2
Table 4. Reductive Ring Opening of 14 and 18 Under Catalyzed
and Noncatalyzed Conditions
Table 5. The Effect of the Ph3P/Ni Ratio on the Regioselectivity
of Reductive Ring Opening of 18
yield
ratio a/bb (%)c
15a/15b 1:6.4 72
15a/15b 1:2.0 24
19a/19b 1:1.3 75
19a/19b 4.8:1 38
equiv of Ph3P
(based on Ni(COD)2)
entry substrate
reaction conditionsa
entry
ratio 19a/19ba
yield (%)
1
2
3
4
5
6
14
14
18
18
18
18
d
1
2
3
4
5
6
0
1
2
3
4
5
83:17
88:12
95:5
96.6:3.4
98:2
38b
66c
71b
74b
75b
50c
10 mol % Ni(COD)2
d
10 mol % Ni(COD)2
31 mol % Ni(COD)2, 62 mol % Ph3P 19a/19b 7.0:1 58
10 mol % Ni(COD)2,e 20 mol % Ph3P 19a/19b 19:1 71
98:2
a DIBAL-H (1.2 equiv) used for 14, 2.5 equiv of DIBAL-H used
for 18, toluene, otherwise specified. b Determined by GC or 400 MHz
1H NMR. c Isolated yield. d DIBAL-H (5.5 equiv) hexanes. e DIBAL-H
was added over 3 h.
a Ratio was determined by GC (carbowax column). b Isolated yield.
c GC yield.
triphenylphosphine/nickel ratio increased, and the optimum yield
and regioselectivity was reached when 4 equiv of triphenylphos-
phine per nickel were used (entries 2-5). Additional triph-
enylphosphine did not affect the regioselectivity, although the
yield was sharply decreased (entry 6). Surprisingly, com-
mercially available Ni(Ph3P)4 under the same reaction conditions
was less regioselective (19a/19b 45:55) and gave a lower yield
(37%) along with a major byproduct arising from hydrogenation
of the starting material.
Scheme 3
Given that different phosphines have different steric and
electronic properties, the nature of the phosphine was expected
to play a crucial role in the selectivity of the reaction.19 We
briefly looked at bidentate phosphines. When 2.0 equiv of dppe
(dppe ) 1,2-bis(diphenylphosphino)ethane) were added to the
reaction, no reaction occurred. In contrast, 2.0 equiv of dppb
(dppb ) 1,4-bis(diphenylphosphino)butane) gave even higher
regioselectivity (19a/19b 99.6:0.4) and yield (82%) than Ph3P.20
To date, dppb is the most effective ligand in regioselective ring
opening of oxabicyclic compounds we have found.
Scope and Limitations of the Nickel Catalyzed Regiose-
lective Ring Opening. A series of oxabicylic substrates were
examined, and the results are summarized in Table 6. The
methyl and TBDMS (TBDMS ) tert-butyldimethylsilyl) ethers,
28 and 14, respectively, underwent ring opening, demonstrating
that the presence of the free hydroxyl group is not a requirement
for high selectivity (entries 1 and 2). For some substrates we
found that the time of addition of DIBAL-H to the reaction
with 28 could be shortened to 10 min without significant
deterioration in the regioselectivity. However, a longer addition
time (2 h) was necessary to ensure high regioselectivity in the
opening of substrate 14.
decrease the amount of DIBAL-H relative to the catalyst in the
reaction, then the catalyst should be active until the starting
material was fully consumed. Indeed, when the DIBAL-H was
added to the mixture slowly Via a syringe pump, 10 mol % of
Ni(COD)2 and 20 mol % of phosphine was sufficient and the
product was isolated in 71% yield with even higher regiose-
lectivity (Table 4, entry 6). A most remarkable observation
was that the intermediate from hydroalumination was not
detected under the slow addition conditions. We have previ-
ously noted a similar relationship between the rate of addition
of DIBAL-H and the enantiomeric excess in a study on
enantioselective ring opening reactions.10
Tolman and co-workers have shown that Ni(Ph3P)4 dissociates
in solution and the equilibrium existing between Ni(Ph3P)4 and
Ni(Ph3P)3/Ph3P lies entirely on the side of the tricoordinate
nickel species.18 We varied the ligand/nickel ratio in order to
explore the effect on the regioselectivity of the reaction. The
alcohol 18 was chosen as the model compound and a set of
typical reaction conditions was established (Table 5). Different
amounts of triphenylphosphine were transferred to 10 mol %
of Ni(COD)2 in toluene (0.1 M) and the mixture was stirred
for 1 h at room temperature. The oxabicyclic alcohol 18 was
premixed with 1.2 equiv of DIBAL-H and the resulting
alkoxyalane was transferred to the Ni(COD)2/Ph3P solution.
Another 1.3 equiv of DIBAL-H was then added over 1-3 h
Via a syringe pump, and the reaction was worked up after 16 h
at room temperature. The regioselectivity increased when the
The reaction is also sensitive to the nature of the bridgehead
substituent. The methyl- and phenyl-substituted compounds
(19) Tolman, C. A. J. Am. Chem. Soc. 1970, 92, 2953.
(20) We found that the bidentate chiral ligand BINAP, which also has
four carbons between the two phosphine atoms, gave the best ee in the
asymmetric ring opening of oxabicyclic[2.2.1] substrates. See: ref 10.
(21) Lautens, M.; Klute, W. Angew. Chem., Int. Ed. Engl. 1996, 35, 442.
(18) Tolman, C. A. J. Am. Chem. Soc. 1970, 92, 2956.

6482 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Lautens et al.
Table 6. Nickel-Catalyzed Reductive Ring-Opening Reactions
In order to shed light on the reaction pathway, we conducted
several deuterium quenching experiments with substrate 28
(Table 7). When DIBAL-H was added to 28 in the presence
of Ni(COD)2 in the absence of a phosphine ligand (entries 1
and 2), 45 and 46 were obtained in a combined 76% yield, and
the ring-opened products were obtained in 18% combined yield
after 5 min (entry 1). Under these conditions, isomer 29a was
the major ring-opened product. Deuterium incorporation in the
oxabicyclic alkanes was 94%, and the major regioisomer was
46 (46/45 ) 6.8:1). If the reaction was stirred for 3 days prior
to the deuterium quench, the ring-opened products were obtained
in 39% yield in a ratio of 2.7:1 favoring 29b (entry 2). When
the reaction was conducted in the presence of dppb (with fast
addition of DIBAL-H), a low yield of 45 + 46 was obtained
(i.e., less than the mole percent of the catalyst used) (entries 3
and 4). In comparison with the ligand-free experiment, the
presence of the ligand significantly decreased the rate of
hydrometalation (<5% starting material remaining Vs 73%,
entries 1 and 3) but the ratio of ring-opened product to
hydrometalated product increased (18:76 Vs 20:5). The regi-
oselectivity of the ring-opened products was much higher in
the presence of the ligand (68:1 Vs 6.5:1) but still lower than
that obtained when the reaction was run under the “optimized”
conditions (entry 4). More surprisingly, the regioselectivity of
the deuterated products was modest and opposite to that
expected based on the final product (1:2.9 and 1:6.8 for 45/46
Vs 6.5:1 and 68:1 for 29a/29b)! In all cases the amount of 45
isolated was relatively small, implying that the organoalane that
leads to 45 undergoes fast opening once it is formed.
a Typical conditions: 10 mol % Ni(COD)2, 20 mol % dppb, 1.2-
3.0 equiv of DIBAL-H, toluene. b Isolated yield of the major regioi-
1
somer. c Ratio determined by GC. d Ratio determined by H NMR.
undergo highly regioselective ring opening in excellent yield
(entries 1, 2, and 4), whereas the trimethylsilyl-substituted 34
was much less reactive under the standard Ni(COD)2/dppb
reaction conditions (entry 5). Only when 100 mol % of Ni-
(COD)2 and 5.0 equiv of DIBAL-H were employed did 34 react
to give 35a and 35b in a 1.3:1 ratio and in 82% yield. It is
interesting to note that in the absence of ligand the two
regioisomers were still formed in an approximately 1:1 ratio
but the conversion was low. When the bridgehead group was
increased in size to tert-butyl (entry 6), no ring opening was
observed regardless of the amount of Ni(COD)2, of the Ni/P
ratio, or when a smaller hydrometalation agent such as Et2AlH
was used. Clearly there is a maximum size of the substituent
which can be accommodated at the bridgehead.
Among the issues these experiments were designed to address
were the following: Does the major reaction pathway occur
by a net hydrometalation-elimination sequence, and if so, is
the reaction a hydronickelation-reductive elimination or an
aluminonickelation-reductive elimination? Is the reaction
pathway the same for all substrates? Is the mechanism the same
when a ligand is present? Why are the major products from
the regioselective ring opening those containing a more sub-
stituted olefin which implies that the hydride added to the carbon
which is adjacent to a more substituted carbon?
The mechanistic pathways we considered are illustrated
Scheme 4: (1) a Lewis acid induced weakening or cleavage of
the bridging C-O bond to give 47 (X ) Ni-H or H) followed
by inter- or intramolecular hydride delivery, (2) hydronickelation
to form 48 followed by â-elimination of the bridging oxygen
or reductive elimination of nickel(0) from a C-Ni-Al species,
(3) aluminonickelation to form 49 followed by reductive
elimination of nickel(0) from a C-Ni-H species. Complicating
this discussion is the uncertainty surrounding the intermediacy
of aluminonickel hydrides and the possibility that nickel does
not insert into the Al-H bond.22 We are currently unable to
resolve this latter issue, and the discussion which follows will
assume that a nickel hydride species of some type is formed
during the reaction.
Under the reaction conditions, the tricyclic compound 37
(entry 7) gave the bicyclic diol 38a as the major product (19:1)
in high yield. Oxabicyclic[2.2.1] substrates with exo or endo
substituents (39 and 41) undergo the ring opening and the
trisubstituted cyclohexenols are obtained in good yields with
excellent regioselectivities (entries 8, 9). Reactions with
oxabicyclic[3.2.1] systems normally required several hours to
go to completion, but the reaction of oxabicyclo[2.2.1] com-
pounds was complete as soon as the final portion of DIBAL-H
was added. When the oxygenated substrate 43 and the
corresponding silyl ether 44 were subjected to the reaction
conditions, complex mixtures were obtained (eq 6). The isolated
product was less polar than the starting material and lacked a
methoxy group which was somehow eliminated during the
reaction conditions.
(22) Po¨rschke, K.-R.; Kleimann, W.; Tsay, Y.-H.; Kru¨ger, C.; Wilke,
G. Chem. Ber. 1990, 123, 1267.

ReductiVe Ring Opening of Oxabicyclic Compounds
Table 7. Trapping Experiments for Substrate 28
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6483
tion is now much faster or that the reaction takes a different
course (i.e., Via 47 or from 48 directly to the product). The
ratio of regioisomeric products is significantly altered by the
combination of the nickel catalyst and phosphine ligand, clearly
demonstrating that the metal, ligand, and substrate are all present
at the regioselectivity-determining step. The most perplexing
result is the preference for 46 at the hydrometalation stage but
the preference for 29a at the ring-opening stage. One possibility
is that the organoalane which when quenched gives 45,
eliminates very rapidly compared to the regioisomer alane. It
is also possible that the metalation is reversible.
The selective formation of regioisomers of type “a” in Table
6 in the presence of the ligand can be explained by an
aluminonickelation, which would position the less bulky di-
alkylaluminum proximal to the bridgehead and the bulky
hydridophosphinonickel distal to the bridgehead, or an electron
deficient aluminum species which complexes to the bridging
oxygen and weakens the C-O bond to form the more-stabilized
allylic cation. The change in regioselectivity as a function of
phosphine implies the former scenario is occuring, since it is
difficult to explain why the regioselectivity would improve in
the presence of increasing amounts of phosphine if the triggering
event is cleavage of the bridging C-O bond prior to delivery
of the hydride moiety. The question of reversibility has not
been proven but a reversible hydronickelation and preferential
elimination of the minor regioisomer would also explain the
results we have observed.
a Conditions: 10 mol % Ni(COD)2 with or without dppb, 1.1 equiv
of DIBAL-H added over 10 min, toluene. After the time indicated,
the reaction was quenched with 10% D2SO4/D2O. b GC yield. c Ratio
2
determined by H NMR after isolation of 45 and 46. d Percentage of
1
deuterium incorporation in 45 and 46 was determined by H NMR.
e The reaction was stirred for 16 h after the addition was complete.
Scheme 4
In conclusion, we have demonstrated that DIBAL-H in the
presence of nickel catalyst is effective in the reductive ring
opening of oxabicyclic compounds and the use of phosphine
ligands significantly enhances the regioselectivity in the reaction
of unsymmetrical substrates. This methodology, combined with
the oxidative cleavage reactions of the cycloheptene or cyclo-
hexene, provides a route to acyclic compounds with synthetically
useful arrays.
Experimental Section
(1S*,5S*,6S*,7R*)-6-(Benzyloxy)-5,7-dimethylcyclohept-3-en-
1-ol (4) and (1S*,2R*,3S*,4S*)-3-(Benzyloxy)-2,4-dimethylcyclo-
heptan-1-ol (5). (A) Excess DIBAL-H. DIBAL-H (neat, 0.23 mL,
1.3 mmol) was added dropwise at 0 °C to a solution of 3 (47.8 mg,
0.19 mmol) in hexanes (1.4 mL). The clear solution was heated at
reflux for 6.5 h, and the reaction was quenched. Workup yielded a
yellow oil which was purified by flash chromatography on silica gel
(10% EtOAc/hexanes) to afford 4 (24.1 mg, 50%) and 5 (12.9 mg,
27%). For 4: white crystals; mp 53-54 °C; IR (CCl4) 3634, 3027,
2965, 2935, 2877, 1455, 1344, 1155, 1108, 1099, 1066, 1027, 944 cm-1
;
1H NMR (200 MHz, CDCl3) δ 7.31 (5H, m), 5.71 (1H, dddd, J )
10.4, 6.1, 5.4, 2.3 Hz), 5.43 (1H, dd, J ) 10.5, 4.4, 1.2 Hz), 4.66 (2H,
s), 3.57 (1H, td, J ) 9.2, 4.8 Hz, 1H), 3.48 (1H, d, J ) 2.5 Hz), 2.60
(1H, m), 2.39 (1H, ddd, J ) 14.3, 13.0, 3.7 Hz), 2.31 (1H, dddd, J )
14.3, 6.6, 1.9, 1.4 Hz), 1.81 (1H, dqd, J ) 9.5, 6.8, 2.6 Hz), 1.48 (1H,
exchanges with D2O, s), 1.20 (3H, d, J ) 6.9 Hz), 1.18 (3H, d, J ) 7.4
Hz); 13C NMR (50 MHz, CDCl3) δ 139.3, 136.1, 128.1, 127.2, 127.1,
125.4, 86.2, 75.4, 70.2, 49.6, 38.9, 37.4, 20.8, 17.5.; HRMS calcd for
C16H22O2 246.1620, found 246.1627. Anal. Calcd for C16H22O2: C,
The isolation of 45/46 and 29a/29b and the increase in the
yield of 29a/29b over time (Table 7, entries 1 and 2) suggest
that a two-step process involving formation of 50 is one pathway
that can occur and is certainly the major pathway when
DIBAL-H is added quickly. However, the rate at which the
â-elimination of the aluminum alkoxide occurs is clearly
influenced by the substrate, the position of the aluminum relative
to the bridgehead, and the reaction conditions. Ring opening
of the hydroaluminated product can be slow, and Lewis acid is
then required for substrates which lack bridgehead substituents.
When a substituent is placed at the bridgehead position and a
phosphine ligand is present, it is often difficult to observe the
hydrometalated species which suggests that either the â-elimina-
1
78.01; H, 9.00. Found: C, 77.98; H, 9.28. For 5: colorless oil; H
NMR (200 MHz, CDCl3) δ 7.31 (5H, m), 4.66 (2H, s), 3.63 (1H, m),
3.43 (1H, s), 2.00 (1H, m), 1.60 (7H, m), 1.29 (1H, s), 1.22 (3H, d, J
) 6.9 Hz), 1.03 (3H, d, J ) 6.8 Hz); 13C NMR (50 MHz, CDCl3) δ
139.8, 128.2, 127.1, 126.8, 88.6, 75.9, 75.7, 46.6, 40.0, 36.7, 32.7, 22.2,
21.3, 19.8.
(B) One Equivalent of DIBAL-H. DIBAL-H (0.52 mL, 2.97
mmol) was added dropwise at 0 °C to a solution of 3 (724 mg, 2.97
mmol) in hexanes (29.7 mL). The clear solution was heated at reflux
for 3.5 h, and the reaction was quenched. Workup yielded a yellow

6484 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Lautens et al.
oil which was purified by flash chromatography on silica gel (10%
EtOAc/hexanes) to afford cycloheptenol 4 (470.2 mg, 65%) and 180.1
mg of starting material.
3.76 (1H, ddddd, J ) 10.7, 10.5, 4.4, 3.3, 2.6 Hz), 2.45 (1H, ddm, J )
13.7, 10.5 Hz), 2.21 (4H, m), 1.81 (3H, dd, J ) 1.6, 1.5 Hz), 1.66
(1H, dd, J ) 13.1, 10.7 Hz), 1.24 (3H, s); 13C NMR (50 MHz, CDCl3)
δ 139.2, 121.0, 70.7, 65.3, 54.4, 43.3, 39.9, 31.7, 26.3; HRMS calcd
for C8H13O2 (M - CH3)+ 141.0916, found 141.0925.
(C) Nickel-Catalyzed DIBAL-H Addition. To Ni(COD)2 (2.6 mg,
0.0095 mmol) was added 3 (23.0 mg, 0.094 mmol) in toluene (0.9
mL), and then DIBAL-H (0.14 mL, 0.78 mmol) was added dropwise.
The progress of the reaction was followed by TLC. The starting
material was consumed after 30 min. To the reaction mixture was
added 2.0 mL of dry hexanes, followed by DIBAL-Cl (0.1 mL, 0.51
mmol). The reaction mixture was heated at reflux for 2 h. The reaction
was quenched by the addition of saturated NH4Cl aqueous solution,
and then enough 10% H2SO4 was added to make the aqueous layer
transparent. The organic layer was separated, and the aqueous layer
was extracted three times with ether and two times with ethyl acetate.
The combined organics were dried over MgSO4. Removal of the
solvent in Vacuo yielded a product mixture that was purified by
chromatography with 10% EtOAc in hexanes to afford 4 (18.3 mg,
76%).
General Procedure for the Reductive Opening of Oxabicyclic
Substrates Using DIBAL-H. The substrate was dissolved in hexane
(0.1 M solution), and the solution was cooled to 0 °C prior to addition
of 6.0 equiv of neat DIBAL-H. The ice bath was then removed, and
the reaction mixture was heated to reflux. When the reaction was
complete, the reaction mixture was diluted with ether and cooled to 0
°C and the reaction was quenched with a saturated solution of NH4Cl.
Fifteen minutes of stirring at room temperature produced a white gel
which was dissolved by the dropwise addition of 10% H2SO4. The
organic layer was separated, and the aqueous layer was extracted three
to five times with ethyl acetate. The combined organics were dried
over MgSO4, and the solvent was removed in Vacuo. The crude product
was purified by flash chromatography on silica gel.
(1S*,3S*)-Cyclohept-5-ene-1,3-diol (7). Alcohol 6 (31.1 mg, 0.25
mmol) was suspended in dry hexanes (2.2 mL) (6 was not completely
soluble in hexanes). DIBAL-H (neat, 0.26 mL, 1.5 mmol) was added
dropwise at 0 °C. The cloudy mixture was heated to reflux. After
warming, the reaction mixture became a colorless solution. The reaction
was heated at reflux overnight. The crude oil obtained upon workup
was purified by chromatography with 80% EtOAc/hexanes to give an
inseparable 4:1 mixture of 7 and the corresponding overreduced product
totaling 25.3 mg. The yield of 7 was calculated to be 64%. Spectral
data for the mixture: white solid; IR (nujol) 3295, 3019, 2930, 2908,
1376, 1363, 1336, 1296, 1249, 1052, 1031 cm-1. For 7: 1H NMR
(200 MHz, CDCl3) δ 5.77 (2H, m), 4.00 (2H, m), 2.39 (4H, m), 2.06
(2H, t, J ) 5.5 Hz), 1.62 (2H, s); 13C NMR (50 MHz, CDCl3) δ 128.6,
65.6, 47.8, 36.1; HRMS calcd for C7H10O (M - H2O)+ 110.0732, found
110.0732.
(1R*,5S*,6R*)-5,6-Bis(benzyloxymethyl)cyclohex-3-enol (13). To
12 (43.3 mg, 0.129 mmol) in freshly distilled ether (1.2 mL) was added
DIBAL-H (neat, 0.07 mL, 0.39 mmol) at 0 °C. The reaction mixture
was heated at reflux overnight. The reaction was not complete;
therefore, another 0.07 mL (0.39 mmol) of DIBAL-H was added at 0
°C. It was heated at reflux for 4 h. The crude product obtained upon
workup was purified by chromatography with 10-20% EtOAc/hexanes
to give 13 (34.8 mg, 80%). For 13: colorless oil; IR (neat) 3419,
3026, 2908, 2875, 2865, 1496, 1454, 1363, 1214, 1095, 1075, 1028,
1
736, 697 cm-1; H NMR (200 MHz, CDCl3) δ 7.30 (10H, m), 5.68
(1H, ddd, J ) 9.9, 6.2, 3.4 Hz), 5.52 (1H, dm, J ) 9.9 Hz), 4.49 (1H,
d, J ) 12.1 Hz), 4.44 (1H, d, J ) 11.8), 4.42 (1H, d, J ) 12.1 Hz),
4.36 (1H, d, J )11.8 Hz), 4.27 (1H, d, J ) 10.0 Hz, exchanges with
D2O), 3.91 (1H, dtd, J ) 10.0, 5.2, 2.7 Hz), 3.63 (1H, dd, J ) 9.3, 7.8
Hz), 3.53 (1H, dd, J ) 9.3, 7.5 Hz), 3.42 (1H, dd, J )9.5, 4.6 Hz),
3.35 (1H, dd, J ) 9.4, 4.9 Hz), 2.65 (1H, m), 2.46 (1H, dddd, J )7.8,
7.5, 7.1, 2.7 Hz), 2.31 (1H, dm, J ) 17.9 Hz), 2.14 (1H, dm, J ) 18.0
Hz); 13C NMR (50 MHz, CDCl3) δ 138.0, 137.5, 128.4, 128.4, 128.0,
127.8, 127.8, 127.7, 127.6, 127.6, 127.0, 125.8, 73.5, 73.3, 69.7, 69.5,
66.8, 39.9, 37.4, 33.6; HRMS calcd for C22H26O3 338.1882, found
338.1867.
(1S*,5S*,6S*,7R*)-6-(tert-Butyldimethylsiloxy)-4,5,7-trimethylcy-
clohept-3-en-1-ol (15a) and (1R*,5R*,6R*,7R*)-6-(tert-butyldimeth-
ylsiloxy)-1,5,7-trimethylcyclohept-3-en-1-ol (15b). To 14 (23.3 mg,
0.0826 mmol) in hexanes (0.80 mL) was added DIBAL-H (0.08 mL,
0.45 mmol) dropwise at room temperature. The reaction mixture was
heated at reflux for 2.5 h, after which the reaction was quenched. The
crude oil obtained upon workup was purified by flash chromatography
using 5% EtOAc/hexanes to give 15a (2.3 mg, 10%) and 15b (14.7
mg, 62%). For 15a: crystalline solid; mp 44-45 °C; IR (CCl4) 3629,
2956, 2931, 2857, 2886, 1471, 1462, 1442, 1381, 1257, 1154, 1091,
1073, 1061, 1023, 968, 939, 880, 861, 837 cm-1; 1H NMR (400 MHz,
CDCl3) δ 5.42 (1H, tm, J ) 6.0 Hz), 3.88 (1H, dd, J ) 4.2, 2.2 Hz),
3.56 (1H, m), 2.42 (1H, qm, J ) 7.1 Hz), 2.32 (2H, t, J ) 6.1 Hz),
1.83 (1H, dqd, J ) 8.1, 7.1, 4.2 Hz), 1.68 (3H, d, J )1.1 Hz), 1.58
(1H, br s), 1.12 (3H, d, J ) 7.3 Hz), 0.91 (3H, d, J ) 7.2 Hz), 0.89
(9H, s), 0.06 (6H, s); 13C NMR (100 MHz, CDCl3) δ 141.4, 119.6,
75.3, 72.3, 49.2, 44.9, 35.5, 26.0, 23.9, 18.3, 16.3, 16.0, -3.8, -4.4;
HRMS calcd for C12H23O2Si (M - Bu)+ 227.1467, found 227.1469.
For 15b: crystalline solid; mp 81-82 °C; IR (CCl4) 3617, 3478, 3028,
2931, 2857, 1471, 1462, 1388, 1371, 1253, 1161, 1110, 1051, 1030,
1
1009, 944, 869, 842 cm-1; H NMR (200 MHz, CDCl3) δ 5.71 (1H,
(1S*,2R*,3S*,4S*)-2,4-Dimethylcyclohept-5-ene-1,3-diol (9). To
8 (46.8 mg, 0.304 mmol) in dry hexanes (2.7 mL) was added DIBAL-H
(neat, 0.32 mL, 1.8 mmol) at 0 °C. The reaction mixture was heated
at reflux for 12 h. The crude oil obtained upon workup was separated
with 20% EtOAc/hexanes to afford 9 (39.5 mg, 83%) and 5.0 mg of
the corresponding overreduced product (10%). For 9: white crystalline
solid; mp 64-65 °C; IR (CCl4) 3634, 3607, 3581, 2967, 2934, 2899,
dddd, J ) 10.3, 7.9, 5.9, 2.0 Hz), 5.44 (1H, ddd, J ) 10.3, 5.2, 1.8
Hz), 3.70 (1H, m), 2.58 (1H, m), 2.43 (1H, ddm, J ) 13.8, 5.7 Hz),
2.24 (1H, dd, J ) 13.8, 7.9 Hz), 1.80 (1H, qd, J ) 7.2, 2,6 Hz), 1.65
(1H, br s), 1.18 (3H, s), 1.13 (3H, d, J ) 7.4 Hz), 1.07 (3H, d, J ) 7.2
Hz), 0.89 (9H, s), 0.06 (6H, s); 13C NMR (100 MHz, CDCl3) δ 138.8,
125.9, 78.8, 72.7, 52.7, 42.9, 39.4, 26.2, 24.4, 20.0, 18.6, 15.4, -3.3,
-4.4; HRMS calcd for C16H31O2Si (M - H)+ 283.2093, found
283.2113.
(1S*,5S*,6S*,7R*)-6-(Benzyloxy)-4,5,7-trimethylcyclohept-3-en-
1-ol (17a) and (1R*,5R*,6R*,7R*)-6-(Benzyloxy)-1,5,7-trimethylcy-
clohept-3-en-1-ol (17b). To 16 (49.1 mg, 0.19 mmol) in hexanes (1.7
mL) was added DIBAL-H (0.17 mL, 0.95 mmol) dropwise at 0 °C.
The reaction mixture was allowed to warm up and then heated to reflux.
After 6.5 h, the reaction was quenched. The crude oil obtained upon
workup was assessed by GC as a 5.6:1 mixture of 17b to 17a. The oil
was purified by flash chromatography using 10% ether/hexanes to give
20.9 mg of 17b and 17a as a 6:1 mixture (GC analysis) along with
26.1 mg of starting material. This represents a 90% yield of product
based on recovered starting material. For 17a: 13C NMR (50 MHz,
CDCl3) δ 141.4, 139.2, 128.2, 127.4, 127.3, 126.9, 120.1, 83.9, 74.1,
72.0, 48.0, 42.2, 36.0, 23.2, 17.2, 16.2. For 17b: 1H NMR (200 MHz,
CDCl3) δ 7.31 (5H, m), 5.74 (1H, dddd, J ) 10.3, 8.0, 5.9, 2.2 Hz),
5.49 (1H, ddd, J ) 10.2, 4.9, 1.8 Hz), 4.66 (2H, s), 3.47 (1H, s), 2.66
(1H, m), 2.38 (1H, ddm, J ) 13.8, 5.9 Hz), 2.26 (1H, dd, J ) 13.9,
8.1 Hz), 1.90 (1H, qd, J ) 7.1, 2.6 Hz), 1.55 (1H, exchanges with
2880, 1459, 1444, 1375, 1194, 1156, 1070, 1027, 968, 960, 939 cm-1
;
1H NMR (400 MHz, CDCl3) δ 5.88 (1H, dddd, J ) 10.3, 8.3, 5.4, 2.6
Hz), 5.32 (1H, dddd, J ) 10.3, 4.1, 2.4, 1.0 Hz), 3.46 (1H, d, J ) 7.0
Hz), 3.36 (1H, ddd, J ) 9.3, 9.3, 3.2 Hz), 2.62 (1H, qm, J ) 7.2 Hz),
2.38 (1H, ddd, J ) 14.1, 8.4, 3.2 Hz), 2.32 (1H, ddddd, J ) 14.1, 9.3,
5.4, 2.3, 1.1 Hz), 1.67 (1H, br, exchanges with D2O), 1.66 (1H, dqd, J
) 9.3, 6.9, 2.4 Hz), 1.30 (1H, br d, J ) 9.8 Hz, exchanges with D2O),
1.18 (3H, d, J ) 7.3 Hz), 1.16 (3H, d, J ) 6.8 Hz); 13C NMR (50
MHz, CDCl3) δ 135.4, 128.7, 78.3, 68.8, 49.8, 38.4, 37.6, 20.4, 17.6;
HRMS calcd for C9H14O (M - H2O)+ 138.1045, found 138.1041. Anal.
Calcd for C9H16O2: C, 69.19; H, 10.32. Found: C, 68.81; H, 10.18.
(1S*,3S*)-1,5-Dimethylcyclohept-5-ene-1,3-diol (11). To 10 (46.0
mg, 0.299 mmol) in dry hexanes (2.6 mL) was added DIBAL-H (neat,
0.32 mL, 1.8 mmol) at 0 °C. The colorless and transparent solution
was allowed to reflux overnight. The crude product obtained upon
workup was separated with 20-30% EtOAc/hexanes to afford 11 (41.7
mg, 89%). For 11: colorless oil; IR (neat) 3400-3200, 2964, 2925,
2880, 2845, 1452, 1434, 1376, 1328, 1261, 1239, 1113, 1084, 1031,
1
1003, 910, 867 cm-1; H NMR (200 MHz, CDCl3) δ 5.41 (1H, m),

ReductiVe Ring Opening of Oxabicyclic Compounds
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6485
D2O, br s), 1.20 (3H, s), 1.19 (3H, d, J ) 7.4 Hz), 1.15 (3H, d, J ) 7.1
Hz); 13C NMR (50 MHz, CDCl3) δ 139.5, 136.7, 128.2, 127.1, 126.8,
126.3, 86.5, 76.1, 72.6, 52.2, 43.1, 38.6, 23.9, 20.5, 15.0; HR GC/MS
of the mixture, major component (17b) calcd for C17H24O2 260.1776,
found 260.1765, minor component (17a) calcd for C17H24O2 260.1776,
found 260.1774.
Hz), 1.03 (3H, d, J ) 7.3 Hz), 0.91 (3H, d, J ) 7.2 Hz); 13C NMR (50
MHz, CDCl3) δ 138.7, 128.3, 127.3, 126.9, 87.0, 80.3, 79.1, 75.9, 72.6,
39.0, 38.5, 38.2, 13.2, 13.1.
General Procedure for Survey of Transition Metal Catalysts. A
solution of the substrate 3 in toluene was added to the catalyst followed
by dropwise addition of DIBAL-H. When the starting material was
consumed, the reaction was quenched with D2SO4/D2O. The organic
layer was separated from the aqueous layer, which was then extracted
three times with ether. The combined organics were dried over MgSO4.
Removal of the solvent in Vacuo yielded a product mixture that was
(1S*,2R*,3S*,4S*)-2,4,5-Trimethylcyclohept-5-ene-1,3-diol (19a)
and (1R*,2R*,3R*,4R*)-1,2,4-Trimethylcyclohept-5-ene-1,3-diol (19b).
To 18 (42.7 mg, 0.254 mmol) in hexanes (2.0 mL) was added DIBAL-H
(0.25 mL, 1.4 mmol) at 0 °C. There was a rapid evolution of gas. The
reaction mixture was allowed to reflux overnight. The reaction was
worked up, and the crude oil thus obtained was separated with 10%
EtOAc/hexanes to give 19a (14.2 mg, 33%) and 19b (18.1 mg, 42%).
For 19a: white crystalline solid; mp 68-69 °C; IR (CCl4) 3645, 3632,
2968, 2944, 2904, 2883, 2879, 1462, 1444, 1380, 1181, 1157, 1088,
1
analyzed for deuterium incorporation by H NMR. Integration of the
peak at 1.65 ppm was used to assess the deuterium incorporation.
Ni(COD)2 (10 mol %). Following the general procedure, to Ni-
(COD)2 (4.7 mg, 0.017 mmol) were added 3 (37.4 mg, 0.153 mmol)
in 1.0 mL of toluene and then DIBAL-H (neat, 0.04 mL, 0.224 mmol)
at -20 °C. After 30 min, the reaction was complete and was quenched
with 10% D2SO4/D2O. After workup, extraction, and drying over K2-
CO3, removal of solvent yielded a residue that was separated with 10%
Et2O/hexanes to give 21 and 23 (25.0 mg, 67%). Integration of the 1H
NMR spectrum showed an 83% deuterium incorporation.
1
1025, 964 cm-1; H NMR (200 MHz, CDCl3) δ 5.71 (1H, dddq, J )
8.2, 5.8, 2.0, 1.4 Hz), 3.44 (1H, br d, J ) 5.0 Hz), 3.28 (1H, br td, J
) 9.7, 3.7 Hz), 2.79 (1H, qd, J ) 6.6, 1.8 Hz), 2.36 (1H, dm, J ) 13.8
Hz), 2.25 (1H, ddd, J ) 13.8, 8.5, 3.7 Hz), 1.71 (3H, d, J ) 1.0 Hz),
1.65 (2H, 1H exchanges with D2O, dqd, J ) 9.6, 6.8, 2.7 Hz), 1.17
(4H, 1H exchanges with D2O, d, J ) 6.9 Hz), 1.13 (3H, d, J ) 6.6
Hz); 13C NMR (100 MHz, CDCl3) δ 140.5, 122.7, 79.4, 69.6, 50.2,
40.5, 36.9, 22.4, 17.5, 17.3; HRMS calcd for C10H18O2 170.1307, found
170.1307. Anal. Calcd for C10H18O2: C, 70.55; H, 10.66. Found:
C, 70.33; H, 10.53. For 19b: white crystalline solid; mp 68-69 °C;
IR (CCl4) 3646, 2967, 2927, 2880, 2868, 1452, 1375, 1320, 1162, 1142,
cis-Cyclohept-5-ene-1,3-diol (25). DIBAL-H (1.0 M in toluene,
18.23 mL, 18.23 mmol) was added to a solution of 24 (1.0 g, 7.93
mmol) in toluene (12 mL) at 0 °C which led to the rapid evolution of
H2 gas. The resulting pale yellow solution was warmed to room
temperature and transferred Via a cannula to a flask containing Ni-
(COD)2 (230 mg, 0.84 mmol). The reaction mixture was stirred for
16 h at room temperature. DIBAL-Cl (7.7 mL, 39.44 mmol) was added
to the reaction mixture, and the temperature was raised to 60 °C and
stirred for an additional 6.5 h. The reaction mixture was cooled to 0
°C, and the reaction was quenched with saturated aqueous NH4Cl. The
aqueous layer was acidified with 10% HCl to pH 3 and saturated with
NaCl, and the mixture was extracted with EtOAc 10 times. The
combined organic solutions were dried over Na2SO4 and concentrated,
and the residue was purified by flash chromatography on silica gel
(Et2O to EtOAc) to give 25 (865 mg, 85%) as a white solid: mp 106-
1
1103, 963, 941 cm-1; H NMR (200 MHz, CDCl3) δ 5.87 (1H, dddd,
J ) 10.4, 8.2, 5.6, 2.5 Hz), 5.40 (1H, ddd, J ) 10.3, 4.7, 2.3 Hz), 3.50
(1H, dd, J ) 8.4, 2.6 Hz), 2.67 (1H, m), 2.43 (1H, ddd, J ) 13.5, 5.6,
2.2 Hz), 2.23 (1H, dd, J ) 13.6, 8.6 Hz), 1.77 (1H, qd, J ) 7.1, 2.5
Hz), 1.33 (1H, exchanges with D2O, br s), 1.21 (3H, s), 1.15 (3H, d, J
) 6.6 Hz), 1.13 (4H, 1H exchanges with D2O, d, J ) 7.0 Hz); 13C
NMR (50 MHz, CDCl3) δ 135.4, 129.2, 78.3, 71.8, 51.9, 43.5, 38.1,
23.7, 20.1, 14.6; HRMS calcd for C10H18O2 170.1307, found 170.1300.
To 18 (43.4 mg, 0.26 mmol) in dry hexanes (1.3 mL) was added
MeLi (low halide, 1.4 M in Et2O, 0.29 mL, 0.406 mmol) at 0 °C. There
was effervescence, and the formation of a white precipitate was
observed. The reaction mixture was warmed with a hot water bath for
about 10 min to ensure complete deprotonation. The reaction mixture
was then cooled to 0 °C, and DIBAL-H (0.23 mL, 1.3 mmol) was
added. No effervescence was observed, but the reaction mixture
became slightly cloudy. After the mixture was heated to reflux
overnight, another 0.12 mL (0.67 mmol) of DIBAL-H was added, and
the reaction mixture was heated to reflux for another night. The crude
oil obtained upon workup showed a 9.5:1 selectivity for 19a to 19b as
assessed by GC. The oil was purified using 10-20% EtOAc/hexanes
to give 19a (32.6 mg, 73%) and 19b (2.1 mg, 5%).
107 °C (Et2O); IR (nujol) 3218, 2907, 1455, 1363, 1005, 955 cm-1
;
1H NMR (200 MHz, CD3CN) d 5.70 (2H, m), 3.46 (2H, m), 2.51 (2H,
br s), 2.21 (5H, m), 1.61 (1H, dd, J ) 20.9, 10.4 Hz); 13C NMR (50
MHz, CD3CN) d 129.1, 67.6, 51.6, 37.9; HRMS calcd for C7H13O2
(M + H)+ 129.0916, found 129.0923. Anal. Calcd for C7H12O2: C,
65.60; H, 9.44. Found: C, 65.83; H, 9.43.
General Procedure for the Reductive Ring Opening Using Ni-
(COD)2/dppb. The reaction scale was typically 0.3-0.5 mmol. Ni-
(COD)2 (0.1 equiv) in toluene (1.0 mL) was added to dppb (0.2 equiv),
and the resulting light brown solution was stirred at room temperature
for 1 h. Substrates bearing a free hydroxyl moiety were premixed with
DIBAL-H (1.0 M in hexanes, 1.1 equiv) in toluene (1.0 mL) to form
the aluminum alkoxide. Protected alcohols were dissolved in toluene
(1.0 mL) and added directly to the flask containing Ni(COD)2/dppb.
DIBAL-H (1.1-1.9 equiv) was added Via a syringe pump over the
indicated time. The reaction mixture was stirred at room temperature
overnight and cooled to 0 °C. The reaction was quenched by the
addition of saturated aqueous NH4Cl, and sufficient 10% HCl was added
to make the aqueous layer transparent. The organic layer was separated,
and the aqueous layer was extracted three times with Et2O. The
combined organics were dried over MgSO4. Removal of the solvent
in Vacuo yielded a product mixture that was purified by chromatog-
raphy. If the reaction product was a diol, the reaction was quenched
by 1.0 mL of 1.1 M Rochelle salt solution. The resulting suspension
was stirred at room temperature for 4 h and filtered, and the white gel
was washed with hot EtOAc several times. The combined filtrate was
dried over Na2SO4. Removal of the solvent in Vacuo yielded a product
mixture that was purified by chromatography.
(1R*,2S*,3S*,4R*,5S*,6S*)-3-(Benzyloxy)-2,4-dimethyl-8-
oxabicyclo[3.2.1]octane-6-d1 (21) and (1R*,2S*,3S*,4S*,5R*,6S*)-
3-(Benzyloxy)-2,4-dimethyl-8-oxabicyclo[3.2.1]octan-6-ol (22). To
3 (51.5 mg, 0.211 mmol) and Ni(acac)2 (4.3 mg, 0.024 mmol) in toluene
(1.0 mL) was added DIBAL-H (0.26 mL, 1.0 M in toluene, 0.26 mmol)
at room temperature. After 15 min, the starting material was consumed.
O2 was bubbled through the dark brown solution for 25 min. Then,
the reaction was worked up by the addition of D2SO4/D2O. The
aqueous layer was extracted three times with ether, and the combined
organics were dried over MgSO4. The volatiles were removed, and
the residue was purified by chromatography with 5% EtOAc/hexanes
to afford 21 with 23 (14.6 mg, 29%), 22 (23.3 mg, 42%). For 21 (as
a mixture with 23): oil; IR (CCl4) 3031, 2953, 2931, 2875, 1496, 1453,
1372, 1359, 1341, 1315, 1193, 1168, 1095, 1068, 1036, 973, 947, 919,
1
734, 695 cm-1; H NMR (200 MHz, CDCl3) δ 7.33 (5H, m), 4.50
(2H, s), 4.01 (2H, m), 3.44 (1H, t, J ) 4.0 Hz), 2.27 (1H, d, J ) 9.8
Hz), 2.21 (1H, s), 2.09 (2H, m), 1.68 (1H, m), 0.96 (6H, d, J ) 7.2
2
Hz); H NMR (60 MHz, CHCl3) d 1.65; 13C NMR (50 MHz, CDCl3)
(1S*,2R*,3S*,4S*)-2,4,5-Trimethylcyclohept-5-ene-1,3-diol (19a).
The reaction was carried out as in the general procedure using dppb
(23.6 mg, 0.055 mmol) and Ni(COD)2 (7.6 mg, 0.028 mmol). Substrate
18 (46.4 mg, 0.28 mmol) had been premixed with DIBAL-H (1.0 M
in hexanes, 0.33 mL, 0.33 mmol). Additional DIBAL-H (0.36 mL,
0.36 mmol) was added Via a syringe pump over 4 h. The crude product
was purified by chromatography on silica gel (6% 2-propanol/CH2-
Cl2) to give a mixture of 19a and 19b (38.3 mg, 82%). The ratio of
δ 139.2, 128.2, 127.1, 126.8, 80.8, 78.5, 77.2, 75.8, 39.8, 24.5, 24.2
(t), 13.4. For 22: oil; IR (CCl4) 3415, 2956, 2935, 2876, 1452, 1372,
1343, 1089, 1066, 1053, 1028, 980, 961, 942, 905, 734, 698 cm-1; 1H
NMR (200 MHz, CDCl3) δ 7.31 (5H, m), 4.68 (1H, dd, J ) 7.1, 2.1
Hz), 4.48 (2H, s), 4.15 (1H, dd, J ) 7.8, 3.4 Hz), 3.85 (1H, d, J ) 3.7
Hz), 3.41 (1H, t, J ) 4.0 Hz), 2.89 (1H, dd, J ) 13.2, 7.0 Hz), 2.04
(2H, m), 1.89 (1H, br d, J ) 6.2 Hz), 1.52 (1H, ddm, J ) 13.3, 7.9

6486 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Lautens et al.
products was determined by GC (carbowax column, 140 °C, retention
time for 19a 37.25 min, for 19b 20.65 min) to be a 300:1 mixture of
19a/19b.
mmol) was added 34 (53.0 mg, 0.23 mmol) and DIBAL (1.0 M in
hexanes, 0.23 mL, 0.23 mmol) in 1.0 mL of toluene. A second portion
of DIBAL-H (0.94 mL, 0.94 mmol) was added Via a syringe pump
over 45 min. The reaction mixture was stirred at room temperature
for 2 h, and the reaction was quenched by Rochelle salt solution (1.1
M, 3.0 mL). The resulting suspension was stirred at room temperature
for 4 h, the organic layer was separated, and the aqueous layer was
extracted three times with ethyl acetate. The combined organics were
dried over Na2SO4. Removal of the solvent in Vacuo yielded a product
mixture that was purified by chromatography on silica gel (50% EtOAc/
hexanes) to give a 1.3:1 mixture of 35a and 35b (43.8 mg, 82%) as
determined by integration of the peaks corresponding to vinylic protons.
For 35a: mp 116-117 °C (50% Et2O in pentane); IR (CCl4) 3636,
3578, 2965, 2900, 2882, 1552, 1539, 1460, 1250, 1157, 1034, 1008,
(1S*,5S*,6S*,7R*)-6-Methoxy-4,5,7-trimethylcyclohept-3-en-
1-ol (29a). The reaction was carried out as in the general procedure
using dppb (45.9 mg, 0.108 mmol), Ni(COD)2 (14.8 mg, 0.054 mmol),
and substrate 28 (97.9 mg, 0.54 mmol). DIBAL-H (1.0 M in hexanes,
0.60 mL, 0.60 mmol) was added dropwise over 15 min. The crude
product was purified by chromatography on silica gel (50% Et2O/
hexanes) to give 29a and 29b (79.6 mg, 80%). The ratio of products
was determined by GC (chiral γ-TA column, 135 °C, retention time
for 29a 21.41 min, for 29b 14.54 and 15.14 min) to be a 380:1 mixture
of 29a/29b. An authentic sample of minor regioisomer was prepared
by another method.21 For 29a: colorless liquid; IR (neat) 3380, 2927,
1
1445, 1373, 1188, 1097, 1025, 972, 956 cm-1; H NMR (400 MHz,
1
968, 951, 909 cm-1; H NMR (400 MHz, CDCl3) δ 6.30 (1H, m),
CDCl3) δ 5.50 (1H, m), 3.47 (3H, s), 3.43 (1H, m), 3.27 (1H, dd, J )
4.0, 1.5 Hz), 2.64 (1H, dd, J ) 15.0, 7.4 Hz), 2.31 (1H, m), 2.23 (1H,
ddd, J ) 14.6, 8.4, 2.9 Hz), 1.88 (1H, m), 1.72 (3H, s), 1.19 (1H, m),
1.18 (3H, d, J ) 7.3 Hz), 1.14 (3H, d, J ) 7.0 Hz); 13C NMR (100
MHz, CDCl3) δ 141.5, 120.0, 87.3, 71.4, 61.3, 48.4, 41.4, 36.2, 23.0,
17.5, 16.2; HRMS calcd for C11H20O2 184.1463, found 184.1458.
(1S*,5S*,6S*,7R*)-6-(tert-Butyldimethylsiloxy)-4,5,7-trimethylcy-
clohept-3-en-1-ol (15a). The reaction was carried out as in the general
procedure using dppb (10.5 mg, 0.025 mmol), Ni(COD)2 (3.4 mg, 0.012
mmol), and substrate 14 (34.9 mg, 0.12 mmol). DIBAL-H (1.0 M in
hexanes, 0.15 mL, 0.15 mmol) was added Via a syringe pump over 2
h. The crude product was purified by chromatography on silica gel
(6% EtOAc in hexanes) to give 15a (26.7 mg, 76%). The reaction
product was analyzed by 400 MHz 1H NMR, and the minor regioisomer
15b was not detected.
3.54 (1H, dd, J ) 8.8, 2.2 Hz), 3.34 (1H, m), 2.91 (1H, m), 2.47 (1H,
d, J ) 7.0 Hz), 2.45 (1H, m), 1.75 (1H, qm, J ) 7.0 Hz), 1.48 (1H, br
s), 1.32 (3H, d, J ) 7.3 Hz), 1.15 (3H, d, J ) 7.0 Hz), 1.08 (1H, d, J
) 9.5 Hz), 0.12 (9H, s); 13C NMR (100 MHz, CDCl3) δ 147.2, 139.3,
78.2, 69.4, 49.7, 42.1, 38.4, 19.6, 16.7, 0.7; HRMS calcd for C12H23O2-
Si (M - H)+ 227.1467, found 227.1461. Anal. Calcd for C12H24O2-
Si: C, 63.10; H, 10.59. Found: C, 63.32; H, 10.69. For 35b: mp
73-74 °C (20% Et2O/pentane); IR (CCl4) 3592, 2965, 2935, 2899,
1550, 1456, 1374, 1249, 1214, 1165, 1139, 1072, 959 cm-1; 1H NMR
(400 MHz, CDCl3) δ 5.67 (1H, m), 5.49 (1H, dm, J ) 11.0 Hz), 3.58
(1H, dm, J ) 9.0 Hz), 2.89 (1H, br m), 2.56 (1H, dddd, J ) 15.4, 5.5,
1.8, 1.1 Hz), 2.34 (1H, dd, J ) 15.4, 6.9 Hz), 1.97 (1H, qd, J ) 7.3,
3.6 Hz), 1.58 (1H, s), 1.35 (3H, d, J ) 9.2 Hz), 1.17 (3H, d, J ) 7.7
Hz), 1.12 (1H, d, J ) 7.3 Hz), 0.10 (9H, s); 13C NMR (100 MHz,
CDCl3) δ 136.0, 127.0, 78.7, 70.7, 52.4, 38.7, 36.5, 20.0, 16.8, -1.6;
HRMS calcd for C11H21O2Si (M - CH3)+ 213.1311, found 213.1319.
Anal. Calcd for C12H24O2Si: C, 63.10; H, 10.59. Found: C, 63.20;
H, 10.54.
(1S*,3S*)-5-Methylcyclohept-5-ene-1,3-diol (31a). The reaction
was carried out as in the general procedure using dppb (35.4 mg, 0.083
mmol) and Ni(COD)2 (11.4 mg, 0.041 mmol). Substrate 30 (58.0 mg,
0.41 mmol) was premixed with DIBAL-H (1.0 M in hexanes, 0.45
mL, 0.45 mmol). Additional DIBAL-H (0.58 mL, 0.58 mmol) was
added Via a syringe pump over 1 h. The crude product was purified
by chromatography on silica gel (Et2O) to give a mixture of 31a and
31b (42.2 mg, 72%). The ratio of product was determined by GC
(carbowax column, 140 °C, retention time for 31a 36.62 min, for 31b
20.23 min) to be a 167:1 mixture of 31a/31b. An authentic sample of
the minor regioisomer was prepared by a less-selective reaction without
dppb. For 31a: white crystalline solid; mp 77-78 °C (Et2O); IR (KBr)
3299, 2930, 2896, 1447, 1344, 1297, 1190, 1077, 1045, 1018, 903,
(1S*,2S*,6R*,9S*)-5-Methylbicyclo[4.2.1]non-4-ene-2,9-diol (38a).
The reaction was carried out as in the general procedure using dppb
(15.2 mg, 0.036 mmol) and Ni(COD)2 (4.9 mg, 0.018 mmol). Substrate
37 (29.6 mg, 0.18 mmol) was premixed with DIBAL-H (1.0 M in
hexanes, 0.20 mL, 0.20 mmol). A second portion of DIBAL-H (0.25
mL, 0.25 mmol) was added Via a syringe pump over 3 h. The crude
product was purified by chromatography on silica gel (50% EtOAc/
hexanes) to give 38a and 38b (24.4 mg, 82%). The ratio of products
was determined to be a 19:1 mixture of 38a to 38b by integration of
the peaks corresponding to vinylic protons. An authentic sample of
the minor regioisomer was prepared by a less-selective reaction without
dppb. For 38a: white crystalline needle; mp 134.5-135.5 °C (EtOAc);
IR (KBr) 3384, 3302, 2920, 2848, 1446, 1370, 1328, 1300, 1274, 1095,
1075, 1017, 898, 883 cm-1; 1H NMR (400 MHz, CDCl3) δ 5.50 (1H,
dm, J ) 8.8 Hz), 4.17 (1H, dt, J ) 10.6, 7.2 Hz), 4.02 (1H, dm, J )
11.0 Hz), 2.56 (1H, m), 2.44 (1H, t, J ) 6.6 Hz), 2.35 (1H, m), 2.20
(1H, dddd, J ) 15.0, 9.1, 4.8, 1.2 Hz), 2.02 (1H, m), 1.90 (1H, ddd, J
) 13.4, 10.2, 3.3 Hz), 1.80 (1H, d, J ) 11.0 Hz), 1.78-1.52 (2H, m),
1.73 (3H, dd, J ) 2.5, 1.4 Hz), 1.31 (1H, d, J ) 4.4 Hz); 13C NMR
(50 MHz, CDCl3) δ 142.3, 119.7, 76.1, 65.1, 50.2, 45.8, 32.6, 28.1,
27.6, 21.0; HRMS Calcd for C10H16O2 168.1150, found 168.1147. Anal.
Calcd for C10H16O2: C, 71.39; H, 9.59. Found: C, 71.40; H, 9.82.
1
866, 831 cm-1; H NMR (400 MHz, CDCl3) δ 5.48 (1H, tm, J ) 6.6
Hz), 3.95 (2H, m), 2.34 (2H, m), 2.02 (2H, t, J ) 5.7 Hz), 1.96 (2H,
br s), 1.77 (3H, s); 13C NMR (50 MHz, CDCl3) δ 137.7, 120.9, 66.0,
65.5, 48.2, 41.7, 35.6, 26.8; HRMS calcd for C8H14O2 142.0994, found
142.0994. Anal. Calcd for C8H14O2: C, 67.57; H, 9.92. Found: C,
67.74; H, 10.16.
(1S*,2R*,3S*,4S*)-2,4-Dimethyl-5-phenylcyclohept-5-ene-1,3-
diol (33a). The reaction was carried out as in the general procedure
using dppb (21.7 mg, 0.051 mmol) and Ni(COD)2 (7.0 mg, 0.025
mmol). Substrate 32 (58.5 mg, 0.26 mmol) was premixed with
DIBAL-H (1.0 M in hexanes, 0.26 mL, 0.26 mmol). DIBAL-H (0.51
mL, 0.51 mmol) was added Via a syringe pump over 6 h. The crude
product was purified by chromatography on silica gel (40% hexanes/
Et2O) to give 33a (55.0 mg, 93%). The ratio of products was >49:1
as determined by integration of the peaks corresponding to the vinylic
protons. The minor compound was assumed to be the regioisomer but
an authentic sample was not prepared. For 33a: white solid; mp 98-
99 °C (Et2O); IR (neat) 3408, 2930, 1448, 1383, 1117, 1019, 969, 845,
(1R*,5R*,6R*)-5,6-Bis(methoxymethyl)-4-methylcyclohex-3-en-
1-ol (40a). The reaction was carried out as in the general procedure
using dppb (22.0 mg, 0.052 mmol), Ni(COD)2 (7.1 mg, 0.026 mmol),
and substrate 39 (51.1 mg, 0.26 mmol). DIBAL-H (1.0 M in hexanes,
0.31 mL, 0.31 mmol) was added dropwise over 15 min. The crude
product was purified by flash chromatography on silica gel (75% Et2O/
hexanes) to give a mixture of 40a and 40b (40.0 mg, 78%). The ratio
of products was determined to be 39:1 mixture of 40a to 40b by GC
(chiral Supelco column, 155 °C, retention time for 40a 21.83 min, for
40b 15.64 and 15.85 min). An authentic sample of the minor
regioisomer was prepared by another method.21 For 40a: colorless
oil; IR (neat) 3429, 2890, 1450, 1190, 1107 cm-1; 1H NMR (400 MHz,
CDCl3) δ 5.69 (1H, m), 5.52 (1H, m), 4.25 (1H, d, J ) 9.9 Hz), 3.89
(1H, m), 1.57 (1H, dd, J ) 9.2, 8.1 Hz), 3.51 (1H, dd, J ) 9.2, 7.5
Hz), 3.35 (3H, s), 3.34 (2H, m), 3.33 (3H, s), 2.59 (1H, m), 2.32 (2H,
m), 2.15(1H, m); 13C NMR (50 MHz, CDCl3) δ 126.9, 125.8, 72.1,
1
765, 704 cm-1; H NMR (400 MHz, CDCl3) δ 7.37 (5H, m), 5.84
(1H, tm, J ) 7.2 Hz), 3.71 (1H, dd, J ) 8.4, 2.6 Hz), 3.49 (1H, m),
3.13 (1H, dd, J ) 15.0, 7.4 Hz), 2.54 (1H, dd, J ) 8.1, 2.2 Hz), 2.53
(1H, d, J ) 7.0 Hz), 1.87 (1H, m), 1.72 (1H, br s), 1.62 (1H, br s),
1.24 (3H, d, J ) 6.9 Hz), 1.01 (3H, d, J ) 7.3 Hz); 13C NMR (100
MHz, CDCl3) δ 146.6, 142.4, 128.3, 127.6, 126.54, 126.49, 77.7, 69.8,
49.4, 41.6, 37.0, 18.7, 17.0; HRMS calcd for C15H20O2 232.1463, found
232.1468.
(1S*,2R*,3R*,4R*)-2,4-Dimethyl-5-(trimethylsilyl)cyclohept-5-
ene-1,3-diol (35a) and (1S*,2R*,3R*,4R*)-2,4-Dimethyl-1-(trimeth-
ylsilyl)cyclohept-5-ene-1,3-diol (35b). To Ni(COD)2 (64.5 mg, 0.23

ReductiVe Ring Opening of Oxabicyclic Compounds
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6487
66.5, 58.9, 58.8, 39.8, 37.3, 33.6. HRMS calcd for C11H21O3 (M +
H)+ 201.1491, found 201.1482.
resulting suspension was stirred at room temperature for 16 h, the
organic layer was separated, and the aqueous layer was extracted with
Et2O three times. The combined organic layers were dried over MgSO4,
and the volatiles were removed in Vacuo. The crude mixture was
analyzed by GC (γ-TA column): 26% ring opening products and major
to minor regioisomer is 68:1, 4.8% hydrogenation product. The
hydrogenation product was isolated after the crude mixture was
ozonolyzed. The ratio of 45 to 46 was 1:2.9 which was determined
by 2H NMR (1.23 and 1.73 ppm respectively); 86% deuterium
incorporation was determined by integration of the corresponding peaks
in the 1H NMR. The spectra of hydrogenation product: colorless oil;
IR (neat) 2927, 2825, 1464, 1375, 1342, 1301, 1257, 1216, 1175, 1102,
(1R*,5S*)-5-(Hydroxymethyl)-4-methylcyclohex-3-en-1-ol (42a).
The reaction was carried out as in the general procedure using dppb
(36.6 mg, 0.086 mmol) and Ni(COD)2 (11.8 mg, 0.043 mmol).
Substrate 41 (60.0 mg, 0.43 mmol) was premixed with DIBAL-H (1.0
M in hexanes, 0.47 mL, 0.47 mmol). Additional DIBAL-H (0.60 mL,
0.60 mmol) was added dropwise Via a syringe pump over 30 min. The
crude product was purified by flash chromatography on silica gel (Et2O)
to give a mixture of 42a and 42b (53.5 mg, 88%). The ratio of products
was determined to be >49:1 mixture of 42a to 42b using 400 MHz 1H
NMR by integration of the peaks corresponding to the vinylic protons.
The minor compound was tentatively assigned to be 42b, although an
authentic sample was not prepared. For 42a: white solid; mp 58-59
°C (Et2O); IR (KBr) 3265, 2901, 1495, 1445, 1371, 1343, 1058, 1037,
1
1077, 1032, 991, 926, 890, 829 cm-1; H NMR (400 MHz, CDCl3) δ
3.93 (1H, dd, J ) 7.7, 3.3 Hz), 3.32 (3H, s), 3.06 (1H, t, J ) 3.7 Hz),
2.20 (1H, ddd, J ) 11.7, 9.9, 4.4 Hz), 2.07 (1H, m), 1.99 (1H, qt, J )
7.3, 3.3 Hz), 1.73 (2H, m), 1.23 (1H, ddm, J ) 11.7, 4.4 Hz), 1.20
(3H, s), 0.92 (3H, d, J ) 7.3 Hz), 0.89 (3H, d, J ) 7.3 Hz); 13C NMR
(100 MHz, CDCl3) δ 83.7, 82.0, 79.4, 62.3, 44.9, 40.2, 30.5, 25.8,
25.0, 13.13, 13.09.
1
1017, 982, 951, 862, 810 cm-1; H NMR (400 MHz, CDCl3) δ 5.44
(1H, m), 4.10 (1H, m), 3.69 (2H, m), 2.35 (1H, m), 1.96 (1H, m), 1.74
(1H, m), 1.71 (3H, s), 1.55 (1H, br s), 1.35 (1H, br s); 13C NMR (50
MHz, CDCl3) δ 133.0, 122.0, 64.4, 64.0, 41.3, 34.4, 33.9, 21.6.
(1R*,2R*,3R*,4R*,5S*)-3-Methoxy-1,2,4-trimethyl-8-oxabicyclo-
[3.2.1]octane. To dppb (47.1 mg, 0.111 mmol) was added a solution
of Ni(COD)2 (15.2 mg, 0.055 mmol) in 2.0 mL toluene. The resulting
light brown solution was stirred at room temperature for 1 h. Substrate
28 (100.6 mg, 0.55 mmol) was added to the flask containing Ni(COD)2/
dppb. DIBAL-H (1.0 M in hexanes, 0.61 mL, 0.61 mmol) was added
dropwise over 10 min. The reaction mixture was stirred at room
temperature for an additional 5 minutes and cooled to 0 °C, and the
reaction was quenched by addition of 2.0 mL of 20% D2SO4/D2O. The
Acknowledgment. The E. W. R. Steacie Memorial Fund,
Natural Science and Engineering Research Council (NSERC)
of Canada, Bio-Mega/Boehringer Ingelheim Inc., Eli Lilly
Grantee Program, and the Petroleum Research Fund, adminis-
tered by the ACS, are thanked for financial support.
JA964361J