π-facial diastereoselection in the 1,2-addition of allylmetal reagents to 2-methoxycyclohexanone and tetrahydrofuranspiro(2-cyclohexanone)

Article abstract of DOI:10.1021/ja9536835

The stereochemical course of the 1,2-addition of several allylmetal reagents and of the Normant Grignard [ClMgO(CH₂)₃MgCl] to 2-methoxycyclohexanone and tetrahydrofuranspiro-(2-cyclohexanone) has been determined. In four of the six substrates examined, a 4-tert-butyl group is present to serve as a conformational anchor. The neighboring methoxyl substituent is shown to be capable of engaging effectively in chelation, although special circumstances can dictate otherwise. Experiments involving the allylindium reagent as the nucleophile in aqueous solution reveal that the presence of water does not inhibit the operation of chelation control, which often exceeds that attainable with the corresponding magnesium, cerium, and chromium reagents in anhydrous media by significant margins. The extent to which cooperation between the α-oxygen atom and control of π-facial nucleophilic attack reaches a maximum ( > 97:3) is when the system is conformationally rigid and the 2-methoxy and 4-tert-butyl groups are both oriented equatorially. As the steric bulk about the oxygen is increased, the ability of indium to anchor onto the heteroatom is significantly lessened. The results of competition experiments are detailed. The prospects for useful synthetic applications of indium catalysis in water or water/THF mixtures appear to be very promising.

Full text of DOI:10.1021/ja9536835

J. Am. Chem. Soc. 1996, 118, 1917-1930
1917
π-Facial Diastereoselection in the 1,2-Addition of Allylmetal
Reagents to 2-Methoxycyclohexanone and
Tetrahydrofuranspiro-(2-cyclohexanone)
Leo A. Paquette* and Paul C. Lobben
Contribution from the EVans Chemical Laboratories, The Ohio State UniVersity,
Columbus, Ohio 43210
ReceiVed NoVember 1, 1995X
Abstract: The stereochemical course of the 1,2-addition of several allylmetal reagents and of the Normant Grignard
[ClMgO(CH2)3MgCl] to 2-methoxycyclohexanone and tetrahydrofuranspiro-(2-cyclohexanone) has been determined.
In four of the six substrates examined, a 4-tert-butyl group is present to serve as a conformational anchor. The
neighboring methoxyl substituent is shown to be capable of engaging effectively in chelation, although special
circumstances can dictate otherwise. Experiments involving the allylindium reagent as the nucleophile in aqueous
solution reveal that the presence of water does not inhibit the operation of chelation control, which often exceeds
that attainable with the corresponding magnesium, cerium, and chromium reagents in anhydrous media by significant
margins. The extent to which cooperation between the R-oxygen atom and control of π-facial nucleophilic attack
reaches a maximum (>97:3) is when the system is conformationally rigid and the 2-methoxy and 4-tert-butyl groups
are both oriented equatorially. As the steric bulk about the oxygen is increased, the ability of indium to anchor onto
the heteroatom is significantly lessened. The results of competition experiments are detailed. The prospects for
useful synthetic applications of indium catalysis in water or water/THF mixtures appear to be very promising.
The capture by conformationally rigid cyclohexanones of
sterically unhindered nucleophiles preferentially from the axial
direction has been extensively documented.1 The central
importance of this phenomenon to our understanding of π-facial
selectivity has resulted in the proposal of several mechanistic
models. Dauben’s early explanation in terms of product
development control2 and Felkin’s original suggestion of
torsional strain3 have more recently been joined by theoretically-
based treatments. Klein’s analysis in terms of the unsymmetrical
distribution of π orbitals,4 Cieplak’s concept of “transition-state
stabilization by electron donation into the vacant σ*q orbital
associated with the incipient bond,”5 and the frontier orbital
analyses offered by Houk and Paddon-Row,6 Reetz,7 Dannen-
berg,8 Coxon,9 and Boyd10 attest to the challenging nature of
the problem. Ab initio calculations have also appeared.11
The impact of a neighboring polar substituent on the control
of diastereofacial stereoselection has commanded considerable
attention.12a With R-alkoxy cyclohexanones, Grignard reagents
add with chelation-controlled selectivity (see 1) to give 2
predominantly.12b TiCl4-promoted additions proceed via equa-
torial attack under conditions where the intervention of a chelate
complex has been established.13a Although RTi(Oi-Pr)3 reagents
are incapable of chelation, the R group is again delivered
preferably to the face opposite to that occupied by the alkoxy
group (see 3).13b In fact, bonding from the axial surface can
be reliably achieved only by prior formation of a complex
between the ketone and the bulky methylaluminum bis(2,6-di-
tert-butyl-4-methylphenoxide) (MAD) reagent.14 Under these
circumstances, MAD so sterically encumbers the less congested
side of the carbonyl carbon (see 4) that nucleophilic attack
proceeds as shown to provide 5.
X Abstract published in AdVance ACS Abstracts, February 15, 1996.
(1) (a) Li, H.; le Noble, W. J. Recl. TraV. Chim. Pays-Bas 1992, 111,
199. (b) Houk, K. N.; Wu, Y.-D. In Stereochemistry of Organic and
Bioorganic Transformations; Bartmann, W.; Sharpless, K. B., Eds., VCH:
Weinheim, Germany, 1987; pp 247-260. (c) Boone, J. R.; Ashby, E. C.
Top. Stereochem. 1979, 11, 53.
Although an R-chloro substituent can be expected to exhibit
diminished chelating ability relative to alkoxy, stereoselectivity
(2) Dauben, W. G.; Fonken, G. J.; Noyce, D. S. J. Am. Chem. Soc. 1956,
78, 2579.
(3) (a) Che´rest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 2199.
(b) Che´rest, M.; Felkin, H. Tetrahedron Lett. 1968, 2205.
(4) Klein, J. Tetrahedron Lett. 1973, 4307.
(5) (a) Cieplak, A. S. J. Am. Chem. Soc. 1981, 103, 4540. (b) Cieplak,
A. S.; Tait, B. D.; Johnson, C. R. J. Am. Chem. Soc. 1989, 111, 8447.
(6) (a) Paddon-Row, M. N.; Wu, Y.-D.; Houk, K. N. J. Am. Chem. Soc.
1992, 114, 10638. (b) Wu, Y.-D.; Houk, K. N.; Paddon-Row, M. N. Angew.
Chem., Int. Ed. Engl. 1992, 31, 1019. (c) Wong, S. S.; Paddon-Row, M. N.
J. Chem. Soc., Chem. Commun. 1991, 327.
(8) (a) Huang, X. L.; Dannenberg, J. J. J. Am. Chem. Soc. 1993, 115,
6017. (b) Huang, X. L.; Dannenberg, J. J.; Duran, M.; Bertra´n, J. J. Am.
Chem. Soc. 1993, 115, 4024.
(9) (a) Coxon, J. M.; Houk, K. N.; Luibrand, R. T. J. Org. Chem. 1995,
60, 418. (b) Coxon, J. M.; Luibrand, R. T. Tetrahedron Lett. 1993, 34,
7093, 7097.
(10) Shi, Z.; Boyd, R. J. J. Am. Chem. Soc. 1993, 115, 9614.
(11) Mori, S.; Nakamura, M.; Nakamura, E.; Koga, N.; Morokuma, K.
J. Am. Chem. Soc. 1995, 117, 5055.
(7) (a) Frenking, G.; Ko¨hler, K. F.; Reetz, M. T. Angew. Chem., Int. Ed.
Engl. 1991, 30, 1146. (b) Frenking, G.; Ko¨hler, K. F.; Reetz, M. T.
Tetrahedron 1993, 49, 3971, 3983.
(12) (a) Ashby, E. C.; Laemmle, J. T. Chem. ReV. 1975, 75, 521. (b)
Maruoka, K.; Itoh, T.; Sakurai, M.; Nonoshita, K.; Yamamoto, H. J. Am.
Chem. Soc. 1988, 110, 3588.
0002-7863/96/1518-1917$12.00/0 © 1996 American Chemical Society

1918 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
can apparently be achieved under the proper circumstances. A
dramatic example has been uncovered in the tetraphenyl-
stibonium bromide-promoted addition of tin enolates to six-
ring R-chloro ketones. The resulting chlorohydrins are formed
exclusively in that manner in which the Cl and OH groups are
cis-disposed.15
Scheme 1
As a consequence of the need to utilize organometallics or
moisture-sensitive Lewis acids in the condensations discussed
above, all of these many reactions have been routinely conducted
in the strict absence of water. In recent years, the intriguing
observation has been made that indium is capable of promoting
addition reactions to carbonyl compounds in water as the
reaction medium.16,17 Given that essentially no information is
available on the stereochemical response of cyclohexanones in
general to metal-based reagents in aqueous environments, we
set out to undertake an in-depth study in which C-alkylation
and related condensations were accomplished by a variety of
methods including aqueous indium protocols. The systems
examined here include 2-methoxycyclohexanone (6), its con-
formationally rigidified 4-tert-butyl homologues 7 and 8, and
the three tetrahydrofuranspiro-substituted examples 9-11.
6a and 6e is sufficiently low that the involvement of 6e in
chelation control should not be energetically impeded.
Treatment of 6 with allylmagnesium chloride in THF at 0
°C afforded 12 and 13 (Scheme 1) in a ratio of 2.3:1 (Table 1,
entry 1).25 Equatorial attack was similarly favored when
recourse was made to the organocerate26 and allylchromium
reagents27 (entries 2 and 3). The proportion of the less polar
axial alcohol was greatest in the CrCl2-promoted example. This
distribution was closely approximated when recourse was made
to the Normant reagent ClMgO(CH2)3MgCl (entry 4).28 The
diols 14 and 15, produced in a 1.5:1 ratio, were alternatively
available by hydroboration of 12 and 1325 and were identified
in this manner.
Coupling of 6 to allyl bromide in the presence of indium
resulted in a significantly heightened increase in the level of
12 irrespective of the solvent employed (entries 5-8). The best
selectivity (14.1:1) was observed in 1:1 THF-H2O, and the
yields were maximized when the reaction medium was purely
water. Product composition was modestly affected when the
Results
2-Methoxycyclohexanone (6). The conformational equilib-
rium defined by 6 was initially reported by Robinson in 1974
to be heavily dominated by the axial form regardless of
(16) The reactivity of allylindium reagents was first recognized in organic
solvents by Butsugan. For selected references, see: (a) Araki, S.; Ito, H.;
Butsugan, Y. J. Org. Chem. 1988, 53, 1831. (b) Araki, S.; Ito, H.;
Katsumura, N.; Butsugan, Y. J. Organomet. Chem. 1989, 369, 291. (c)
Araki, S.; Katsumura, N.; Ito, H.; Butsugan, Y. Tetrahedron Lett. 1989,
30, 1581. (d) Araki, S.; Butsugan, Y. J. Chem. Soc., Perkin Trans. 1 1991,
2395. (e) Araki, S.; Katsumura, N.; Butsugan, Y. J. Organomet. Chem.
1991, 415, 7. (f) Araki, S.; Shimizu, T.; Johar, P. S.; Jin, S.-J.; Butsugan,
Y. J. Org. Chem. 1991, 56, 2538. (g) Araki, S.; Imai, A.; Shimizu, K.;
Butsugan, Y. Tetrahedron Lett. 1992, 33, 2581. (h) Jin, S.-J.; Araki, S.;
Butsugan, Y. Bull. Chem. Soc. Jpn. 1993, 66, 1528. For a general review
of selective reactions involving allylic metal reagents, see: Yamamoto, Y.;
Asao, N. Chem. ReV. 1993, 93, 2207.
solvent.18 Subsequent refinements by H NMR showed 6a to
1
(17) For reviews which focus attention on allylation in aqueous media,
consult: (a) Li, C. J. Chem. ReV. 1993, 93, 2023. (b) Chan, T. H.; Li, C.
J.; Lee, M. C.; Wei, Z. Y. Can. J. Chem. 1994, 72, 1181. (c) Lubineau, A.;
Auge´, J.; Queneau, Y. Synthesis 1994, 741.
be favored (63%) in CCl4 but appreciably disfavored (20%) in
a more polar solvent.19 This widely disseminated finding,20
believed to be in conformity with decreased hyperconjugative21
and dipole-dipole interactions22 as well as reduced A(1,3)
strain,23 has recently been further reevaluated at 500 MHz.24
The percent of 6a was found in this setting to range from 57
(in C6D12) to 16 (CD3CN). The important conclusion that can
be drawn from these studies is that the energy barrier between
(18) Robinson, M. J. T. Tetrahedron 1974, 30, 1971.
(19) Cantacuzene, D.; Tordeux, M. Can. J. Chem. 1976, 54, 2759.
(20) Lambert, J. B. In The Conformational Analysis of Cyclohexenes,
Cyclohexadienes, and Related Hydroaromatic Compounds; Rabideau, P.,
Ed.; VCH: Weinheim, Germany, 1989; Chapter 2.
(21) Eisenstein, O.; Anh, N. T.; Jean, Y.; Devaquet, A.; Cantacuzene,
J.; Salem, L. Tetrahedron 1974, 30, 1717.
(22) Corey, E. J.; Burke H. J. J. Am. Chem. Soc. 1955, 77, 5418.
(23) Johnson, F. Chem. ReV. 1968, 68, 375.
(24) Fraser, R. R.; Faibish, N. C. Can. J. Chem. 1995, 73, 88.
(25) This reaction has been previously conducted in refluxing ether.
Under these circumstances, the 12/13 composition is 6.6:1 [Negri, J. T.;
Paquette, L. A. J. Am. Chem. Soc. 1992, 114, 8835].
(13) (a) Reetz, M. T.; Kesseler, K.; Schmidtberger, S.; Wenderoth, B.;
Steinbach, R. Angew. Chem., Int. Ed. Engl. 1983, 22, 989. (b) Reetz, M.
T.; Organotitanium Reagents in Organic Synthesis; Springer-Verlag: Berlin,
1986.
(14) (a) Maruoka, K.; Oishi, M.; Yamamoto, H. Synlett 1993, 683. (b)
Maruoka, K.; Oishi, M.; Shiohara, K.; Yamamoto, H. Tetrahedron 1994,
50, 8983.
(26) Imamoto, T.; Takiyama, N.; Nakamura, K. Tetrahedron Lett. 1985,
26, 4763.
(15) Yasuda, M.; Oh-hata, T.; Shibata, I.; Baba, A.; Matsuda, H.; Sonoda,
N. Tetrahedron Lett. 1994, 35, 8627.
(27) Hiyama, T.; Okude, Y.; Kimura, K.; Nozaki, H. Bull. Chem. Soc.
Jpn. 1982, 55, 561.

1,2-Addition of Allylmetal Reagents to Cyclohexanones
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1919
Table 1. Facial Selectivity in Nucleophilic Additions to 6 and 9a
Scheme 2
Scheme 3
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments. b The allyl
bromide and indium powder were refluxed in THF for 1 h and cooled
to 25 °C before the ketone was added.
allylindium reagent was preformed in refluxing THF (entry 8)
prior to addition of the ketone (compare entry 5).
1-Oxaspiro[4.5]decan-6-one (9). A change in the nature of
the polar substituent to a spirocyclic tetrahydrofuran ring
introduces several conformationally counterbalancing influences.
The geminal disubstitution requires that the oxygen atom and a
methylene carbon be concomitantly projected axially and
equatorially. For 16, the O-axial isomer has been shown to
predominate (ca. 68%) in CS2 solution at 35 °C.29 This ∆G3°08
of 0.46 kcal/mol favoring 16a has been attributed to the
reduction in syn-axial compression that materializes when the
oxygen atom is projected axially. Although the conformational
properties of 930 have not been comparably scrutinized, it would
appear logical to assume that the extent to which 9a is populated
9 to adopt the O-equatorial conformation relative to 6, thereby
diminishing the attractiveness of metal chelation.
The experimental results appear to support this hypothesis.
Thus, when the same battery of experiments was applied to 9
(Scheme 2), the level of diastereoselection in favor of the cis
isomers (17 and 19) was noticeably more normalized (entries
9-16). Although the indium-promoted couplings generally
exhibited heightened chelate/non-chelate behavior as before, the
more favorable distribution was now encountered in THF (entry
13) and not in aqueous THF (entry 14). Quite unexpected was
the discovery that the Normant reagent was notably effective
in eliciting a very respectable (8.8:1) non-chelate-controlled
response from 9 (entry 12). Due to the conformational
flexibility of products 17-20, unambiguous assignment of
relative stereochemistry was accomplished by hydroborative
interconversion. Diol 19 has previously been transformed into
the Cs symmetric dispiro heterocycle by dehydrative cycliza-
tion.30
Conformationally Disparate 2-Methoxy-4-tert-butylcyclo-
hexanones 7 and 8. Significant reduction in the conformational
mobility of both 6 and 9 was viewed as an invaluable tool for
gaining insight into the stereochemical course of the reactions
described above. The classical tactic of incorporating a 4-tert-
butyl substituent onto the cyclohexanone ring was adopted. The
preparation of 7 and 8 began by conversion of 21 into the silyl
enol ether 22 (Scheme 3). Subsequent oxidation of this
intermediate with iodosobenzene in cold (-70 °C) methanol
containing 2 equiv of boron trifluoride etherate led directly to
7 and 8 in 60% yield.31,32 The more dominant equatorial isomer
7 (3:1), which was readily separated from 8 by chromatography
is significantly greater than that of 6a. In light of the fact that
the A values for ethyl and methoxyl are 1.8 and 0.6 kcal/mol,
respectively, simple additivity does not appear to be applicable.
This may be because other structural deformations are also
operative. In any event, it should prove more difficult to entice
(28) Cahiez, G.; Alexakis, A.; Normant, J. F. Tetrahedron Lett. 1978,
3013.
(29) (a) Picard, P.; Moulines, J. Tetrahedron 1978, 34, 671. (b) Picard,
P.; Moulines, J. Tetrahedron Lett. 1970, 5133.
(30) Paquette, L. A.; Negri, J. T.; Rogers, R. D. J. Org. Chem. 1992,57,
3947.

1920 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
Table 2. Facial Selectivity in Nucleophilic Additions to 7 and 8a
on silica gel, provided the highest face selectivity observed in
this study (entries 17-24, Table 2). In all experiments, 23 or
25 proved to be so dominant in the reaction mixtures (>97:3)
1
that the products of axial attack were not observed upon H
NMR analysis (Scheme 4). Only in the CeCl3-promoted
coupling was 24 formed to a detectable level that permitted its
characterization (Table 2).
DEPT experiments33 were particularly useful in defining the
stereochemical features of 23 and 24. This technique takes
advantage of heteronuclear three-bond coupling, with the degree
of signal enhancement achieved by pulsing a selected hydrogen
frequency and observing a 13C nucleus being maximized when
the dihedral angle relationships are at 0° or 180°. As for
carbinol 23, models show the dihedral angle between H-2a and
C-7 to be approximately 60° (see A) and the same angle for 24
to be significantly widened to about 180° (see B). Only in the
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments. b The allyl
bromide and indium powder were refluxed in THF for 1 h and cooled
to 25 °C before the ketone was added.
Scheme 4
latter instance was C-7 measurably coupled to H-2a. The
methoxyl carbon was seen to exhibit an enhanced signal in both
products, the result of heteronuclear coupling through the oxygen
atom.
Addition of the Normant reagent to 7 proved to be a highly
stereocontrolled process as well, giving rise to 25 in 96% yield
(entry 20). Since hydroboration-oxidation of 23 also led to
25, its stereochemical assignment was considered secure.
Cyclization of 25 by initial regioselective tosylation of its
primary hydroxyl followed by intramolecular SN2 displacement25
provided the spirotetrahydrofuran 27.
The conformational inflexibility of 7 virtually guarantees that
the methoxyl oxygen is projected equatorially. This advanta-
geous situation so orients the nonbonded electron pairs that a
cooperatively matched arrangement exists. The stereoisomeric
nature of 8 positions the methoxyl substituent axially when in
the chair form (8ch), thereby discounting the possibility of
chelation control. However, whereas 7 is certain to be confor-
mationally rigid, 8 is likely to be only conformationally biased
toward 8ch.34 Thus, its methoxyl group need not be confined
to an axial disposition by virtue of equilibration with the twist
conformer 8tw. In this geometry, the ethereal oxygen is
projected pseudoequatorially and now resides in close proximity
to the carbonyl group. In simple cyclohexanes, the twist form
(31) Haight, A. R. Ph.D. Thesis, University of WisconsinsMadison,
1990. We thank Prof. E. Vedejs for making the relevant part of this thesis
available to us.
(32) (a) Moriarty, R. M.; Prakash, O.; Duncan, M. P.; Vaid, R. K.;
Musallam, H. A. J. Org. Chem. 1987, 52, 150. (b) House, H. O.; Czuba, L.
J.; Gall, M.; Olmstead, H. D. J. Org. Chem. 1969, 34, 2324. (c) Cousins,
R. P. C.; Simpkins, N. S. Tetrahedron Lett. 1989, 30, 7241. (d) Cain, C.
M.; Cousins, R. P. C.; Coumbarides, G.; Simpkins N. S. Tetrahedron 1990,
46, 523.
(33) (a) Bax, A. J. Magn. Reson. 1983, 52, 330; 1983, 53, 517. (b) Rutar,
V. J. Am. Chem. Soc. 1983, 105, 4095. (c) Rutar, V. J. Magn. Reson. 1984,
56, 87. (d) Wimperis, S.; Freeman, R. J. Magn. Reson. 1984, 58, 348. (e)
Lin, L.-J.; Cordell, G. A. J. Chem. Soc., Chem. Commun. 1986, 377.
(34) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Carbon
Compounds; John Wiley and Sons, Inc.: New York, 1994; Chapter 11.
is somewhat more than 5 kcal/mol less stable than the chair.
For the parent cyclohexanone, the twist form is only 2.7 kcal/
mol above the chair,35 a value considerably lower because the
ring happens to be flattened at the carbonyl carbon and the

1,2-Addition of Allylmetal Reagents to Cyclohexanones
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1921
Scheme 5
We had originally hoped to apply to 28 and 29 the same
DEPT analysis that proved so successful earlier. The Karplus
angle between H-2e and C-7 in 28 can reasonably be expected
to be close to 60° (see C), while that in 29 should be
significantly smaller because of unfavorable 1,3-diaxial inter-
actions involving the allyl substituent and likely hydrogen
bonding of the hydroxyl to the methoxyl oxygen (see D).
opposite end.36,37 The difference in ∆G° between 8ch and 8tw
is very likely of even lesser magnitude because the steric
interaction of the axial methoxyl in 8ch is relieved when
proceeding to the twist geometry.
However, in neither case was heteronuclear coupling observed
between H-2e and C-7. This may simply be a reflection of the
fact that equatorial/equatorial and axial/equatorial couplings are
significantly smaller than those of the axial/axial type, with the
result that the associated signals are of much lower intensity.
Alternatively, 1H/1H COSY 90 experiments performed on 29
in C6D6 solution confirmed H-6a to reside at δ 1.69, its
downfield location stemming from deshielding contributions
arising from the nearby methoxyl oxygen. Irradiation of H-6a39
under DEPT conditions caused C-7 to exhibit a strong signal
in the carbon subspectrum. Further confirmation of the axial
orientation of the allyl group was derived from an NOE study
in which 10% integral enhancements were seen for H-3a and
H-5a when H-7/H-7′ were irradiated. In 29, the H-7 protons
appear at δ 2.28 as a narrowly split (∆µ ) 15.8, JAB ) 14.3
Hz) component of an ABX pattern. In contrast, H-7 and H-7′
in 28 are much more widely spaced (δ 2.40 and 2.16) since the
magnetic environment of one of them is significantly pertured
by the proximal methoxy group.
Reaction of 8 with the Normant reagent provided a 1.3:1
mixture of 30 and 31 in 90% yield (entry 28). The structural
features of these 1,4-diols were established as before by the
hydroboration-oxidation of 28. The resulting 30 proved
identical in all respects to the major constituent obtained earlier.
Evaluation of the Haptophilic Properties of the Tetrahy-
drofuran Ring in 10 and 11. When chelation operates,
nucleophilic attack by the organometallic reagent is skewed
toward addition from the equatorial surface of the carbonyl as
a consequence of the haptophilic influence of the neighboring
ether oxygen. An interesting relevant question inquires whether
the coordinating capability of methoxyl oxygen is superior to
or less than that of the tetrahydrofuran ring oxygen. Also, is
the trend comparable for different metals? A quick glance at
the experimental data for 6 and 9 (Table 1) might lead one to
conclude in favor of methoxyl. However, the differing con-
formational dynamics of this pair of R-oxygenated cyclohex-
anones, as pointed out earlier, could bias the experimental
observations. For this reason, the selectivities observed for the
7/10 and 8/11 pairs could prove informative. Although subtle
differences in the conformational predisposition of these systems
are evident, an analysis of the extent of haptophilic control was
considered to be necessary and worthwhile.
With the exception of the allylmagnesium chloride and
allylcerium examples (entries 25 and 26), the ability of 8 to
engage in chelate-controlled metal-mediated allylation reactions
is very respectable (entries 27-32). The stereoselectivity of
the 1,2-addition process when indium is involved proved to be
consistently very good. The change in solvent from THF to
H2O modestly erodes the proportion of 28 formed (Scheme 5).
This phenomenon has previously been observed with 9. The
emergence of the allylchromium addition as a highly stereo-
controlled process merits comment. When R-alkoxy aldehydes
are involved,38 crotylchromium reagents have been shown to
attack with a high level of anti selectivity. Thus, the chelation-
controlled response of 7 and 8 is not entirely unexpected,
although the phenomenon is not observed in all of the examples
studied in this investigation.
(35) Anet, F. A. L.; Chmurny, G. N.; Krane, J. J. Am. Chem. Soc.
1973,95, 4423.
(36) Lambert, J. B.; Carhart, R. E.; Corfield, P. W. R. J. Am. Chem.
Soc. 1969, 91, 3567.
(37) 2-tert-Butylcyclohexanone is reported to reside predominantly in
the twist conformation: ref 34, p 733.
(38) (a) Martin, S. F.; Li, W. J. Org. Chem. 1989, 54, 6129. (b) Mulzer,
J.; de Lasalle, P.; Freissler, A. Liebigs Ann. Chem. 1986, 1152. (c) Hiyama,
T.; Okude, Y.; Kimura, K.; Nozaki, H. Bull. Chem. Soc. Jpn. 1982, 55,
561. (d) Lewis, M. D.; Kishi, Y. Tetrahedron Lett. 1982, 23, 2343. (e)
Nagaoka, H.; Kishi, Y. Tetrahedron 1981, 37, 3873. (f) Hiyama, T.; Kimura,
K.; Nozaki, H. Tetrahedron Lett. 1981, 22, 1037. (g) Buse, C. T.; Heathcock,
C. H. Tetrahedron Lett. 1978, 1685. (h) Okude, Y.; Hirano, S.; Hiyama,
T.; Nozaki, H. J. Am. Chem. Soc. 1977, 99, 3179.
(39) In actuality, the signal for H-6a is overlapped by those for H-3e
and H-6e. However, these protons cannot exert an effect on C-7 resulting
from heteronuclear coupling.

1922 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
Scheme 6
With pure samples of cyclohexanones 10 and 11 in hand, it
was initially assumed that they might be distinguished on the
basis of standard NOE measurements. Analysis was highlighted
by the fact that the major diastereomer displayed four protons
between δ 4.2 and 1.8, while the minor constituent exhibited
five protons between δ 3.8 and 2.1 at 300 MHz. For added
corroboration, 1H/1H COSY 90 studies were carried out on both
compounds. As seen in Figure 1, H-9 and H-9′ in 10 appear at
appreciably lower field than the remaining protons, with H-9
residing well into the deshielding cone of the ketone carbonyl.
This identification allowed H-8 and H-8′, and subsequently H-7
and H-7′, to be uncovered. Three of these protons appear as
multiplets in the heavily overlapped δ 1.63-1.43 region, while
H-8′ is centered at δ 1.38. As expected, H-6e is characterized
by one large geminal and two small vicinal couplings. The
proximity of H-6e and H-6a to H-5e and H-5a provided key
information relative to their location in the spectrum. Clearly
indicated was the fact that the signal for H-5e was overlapped
with those of H-7/H-7′ and H-8. Also, H-5a is sufficiently
shielded that it constitutes the ring proton to highest field. The
methine proton H-4 could not be traced from H-5a or H-5e.
However, because H-5e shows strong W-plan coupling to H-3e
and the latter is demonstrably coupled to H-4 and H-3a, the
remaining protons were easily located. The principal adoption
of a chairlike conformation is supported by the existence of a
diagnostic W-coupling between H-3e and H-5e.
Full assignment to 11 could be made in an entirely analogous
way (Figure 2). Additionally, the significant deshielding
experienced by H-4 and H-6a as a consequence of their 1,3-
diaxial relationship to the tetrahydrofuranyl oxygen is striking.
As seen with 10, the normal sequencing in rigid six-membered
rings is for equatorial protons to appear downfield of their axial
counterparts. This phenomenon is also reflected in the chemical
shifts of H-5e/H-5a and H-3e/H-3a in 11.
The route selected for the preparation of 10 and 11 is outlined
in Scheme 6. The conversion of 21 to 32 was quickly
With the stereochemistry of 10 and 11 established, the
Normant reagent deployed in their synthesis was seen to exhibit
a kinetic preference for equatorial attack. A logical rationaliza-
tion of this finding would be to involve intramolecular delivery
from the equatorial oxygen of the acetal as in E. The preferred
intervention of E is believed to reflect the usual steric advantages
associated with chelation to the less encumbered oxygen atom.
determined to be problematic because of competitive formation
of the dibenzylidene derivative under conventional aldol
condensation conditions.40a In order to curtail the formation
of this product, the aldol reaction mixture was treated with
methanesulfonyl chloride 5 s after introduction of the
benzaldehyde.40b The subsequent addition of triethylamine was
followed either by a reflux period of 1 h or overnight stirring
at room temperature. Product purification was most easily
achieved after ketalization. Once this step had been performed,
33 was obtained in 50% overall yield by direct crystallization
from 95% ethanol. Ozonolysis of 33 resulted in oxidative
cleavage of the double bond to give 34 (91%). Implementation
of Normant technology at this stage gave rise in 84% yield to
a 4.5:1 mixture of diols. The fact that axial attack had
predominated was made clear following monotosylation of the
two diols and intramolecular SN2 cyclization of these function-
alized intermediates to furnish 35 and 36. For convenience,
separation of the diastereomers was deferred until after deket-
alization with concentrated HCl in hot aqueous acetone.
The two diastereomeric allylation products of 10 proved to
be conveniently amenable to chromatographic separation. The
less polar diastereomer (Rf 0.35), when subjected to 1H/1H
COSY 90 and NOE experiments, was confirmed to be 37 (see
F). Comparable analysis of the more polar homoallylic alcohol
(Rf 0.18) was consistent with its formulation as 38 (see G). With
the unequivocal identification of H-6a, semiselective long-range
DEPT studies established its trans relationship to C-10 (3J).
The polarity distinction defined above is very pervasive and,
in our view, can serve as a very reliable indicator of stereo-
chemistry in this series. We have noted that those cyclohexanols
which preferentially adopt an axial hydroxyl are invariably less
polar than their equatorial counterparts (Table 3). This is
believed to be a reflection of the inability of the sterically more
crowded axial OH to bind as tightly to the adsorbent. Interest-
ingly, a polarity reversal has been noted once the same oxygen
becomes incorporated into a tetrahydrofuran ring and a second
(40) (a) Sanghvi, Y. S.; Rao, A. S. Ind. J. Chem. 1987, 263, 671. (b)
Heathcock, C. H.; Buse, C. T.; Kleschick, W. A.; Pirrung, M. C.; Sohn, J.
E.; Lampe, J. J. J. Org. Chem. 1980, 45, 1066.

1,2-Addition of Allylmetal Reagents to Cyclohexanones
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1923
1
Figure 1. 300 MHz H/1H COSY 90 spectrum of 10 in CDCl3 solution.
reflect some overriding of this steric effect by modest chelation
to oxygen. The result is the formation of “reduced” amounts
of 38, although this diastereomer continues to predominate.
The behavior detailed above clearly does not extend to the
allylations promoted by indium (entries 37-40). In all four
condensations studied, the equatorial homoallylic alcohol 37
proved to be the major product. Thus, 10 reacts with the
allylindium reagent with much greater chelation control than
the other organometallic reagents, irrespective of whether the
reaction is conducted in THF or water.
The studies involving diastereomer 11 were equally revealing.
The structural distinction between 41 and 42 was made by means
paralleling those developed earlier for 37 and 38. For example,
the axial alcohol 41 was less polar than 42 (Table 3) and
exhibited the confirmatory NOE effects given in H. Also,
whereas W-coupling between H-6a and H-10 was not evident
in 41, it was clearly apparent in the spectrum of 42 (see I).
Semiselective long-range DEPT experiments again confirmed
the trans-diaxial relationship of H-6a and C-10.
ether functionality resides on an adjacent carbon. We speculate
that this effect may originate from the greater dipole moments
present in those molecules that feature at least one axial C-O
bond. Beyond that, the stereochemical projection of the second
ether oxygen appears to be inconsequential.
As shown in Table 4, the substitution of an equatorial
methoxyl by a spirotetrahydrofuran ring has extensive conse-
quences on π-facial stereoselectivity (Scheme 7). While the
allyl Grignard reagent attacks from both possible directions with
approximately equal facility (entry 33), the cerate and chromium
reagents exhibit an almost 2-fold preference for axial attack
(entries 34 and 35). The Normant reagent gives rise predomi-
nantly to 40, and by a very large margin (14.2:1, entry 36). We
are not aware of reports in which an axial preference of this
magnitude has been observed previously.
The stereoselectivity of the last reaction may be fostered by
projection of the methylene of the heterocyclic ring in the axial
direction resulting in steric impedance to bonding from the
adjacent equatorial surface of the carbonyl. If this is the case,
then the experiments defined in entries 33-35 (Table 4) could
Although 43 and 44 proved difficult to separate chromato-

1924 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
1
Figure 2. 300 MHz H/1H COSY 90 spectrum of 11 in CDCl3 solution.
graphically, they could be independently prepared by hydro-
boration of 41 and of 42. Structural proof, realized in this
manner, also permitted configurational assignment to be made
with confidence to the dispiro ethers 45 and 46 (Scheme 8).
Despite the axial disposition of the ether oxygen in 11, the
Grignard and cerium reagents add with a modest preference from
the diaxial direction (entries 41 and 42). All of the other
organometallics, and particularly the indium reagents, exhibit
a respectable kinetic preference for the bonding to the equatorial
face of the carbonyl group (entries 43-48). Proper understand-
ing of these facts is dependent on a reasonable knowledge of
the extent to which conformations 11ch and 11tw become
involved in the rate-determining transition states. We have
earlier discussed how small the free energy difference between
these conformers might be. If chelation to 11tw operates to a
significant degree, heightened levels of equatorial attack should
be operational as is seen.
Competition Experiments. Eliel, Frye, and co-workers41
have recently called attention to the fact that in the absence of
kinetic studies it becomes impossible to distinguish between
chelation as a product-determining event or an unproductive
reversible process.42 In their words, “if chelation is the cause
of the high stereoselectivity observed in additions to alkoxy
ketones, it must also be true that the chelated transition state
provides a lower energy pathway to products than the less
stereoselective pathway not involving chelation which is avail-
able to all ketones”. In order to probe this important issue in
the present context, we have conducted a series of competition
experiments designed to elucidate whether those examples
exhibiting increased stereoselectivity are indeed mediated by
more reactive indium chelates. In the interpretation of these
findings, we have assumed that chelation/non-chelation is the
only kinetic consideration at issue. Possible contributions
stemming from differing field/inductive effects, orbital overlap
factors, electronegativity effects, and steric contributions are
assumed to play a lesser role, but strictly speaking,43 this need
not be the case. However, since the order of reactivity does
follow the observed stereoselection order and since the ketones
carry only a single R-oxygen atom, it can be argued that the
correlation is reasonable.
Several revealing pieces of information have emerged from
this aspect of the study. Thus, direct competition between 6
and 7 has revealed these two ketones to be almost equally
reactive toward the allylindium reagent (entry 49, Table 5). Since
both of these substrates project their R-methoxyl substituent
equatorially for possible complexation to the incoming organo-
metallic reagent and comparable levels of chelation are antici-
pated, it would be necessary that quite similar rates of addition
be observed if complexation to methoxyl oxygen is kinetically
relevant. The sensitivity of the allylindium reagent to steric
screening is made evident in the direct competition between
ketones 7 and 10 (entry 51). Although both of these ketones
project their ether oxygens equatorially, only 7 reacts with the
(41) (a) Chen, X.; Hortelano, E. R.; Eliel, E. L.; Frye, S. V. J. Am. Chem.
Soc. 1990, 112, 6130. (b) Frye, S. V.; Eliel, E. L.; Cloux, R. J. Am. Chem.
Soc. 1987, 109, 1862.
(42) Laemmle, J.; Ashby, E. C.; Neumann, H. M. J. Am. Chem. Soc.
1971, 93, 5120.
(43) Das, G.; Thornton, E. R. J. Am. Chem. Soc. 1993, 115, 1302.

1,2-Addition of Allylmetal Reagents to Cyclohexanones
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1925
Table 3. Polarities of Selected Cyclohexanols and Spirotetrahydrofuran Derivatives (Rf Values)
a The symbolism O(CH2)3 is meant to indicate that the oxygen atom resides at R2 ()equatorial), while (CH2)3O denotes axial attachment at R3.
organometallic reagent to the point where it is completely
consumed. Evidently, therefore, the incorporation of the
R-oxygen atom into a tetrahydrofuran ring does not allow for
the generation of a kinetically relevant chelate structure as
rapidly as that associated with 7. This does not mean that 10
is unable to engage in chelation. However, it does so much
less effectively than 7.
parallelism between reactivity and stereoselectivity unequivo-
cally operates in water as the reaction medium.
Conclusions
Allylation reactions performed with indium under aqueous
conditions have generally been shown to be more stereoselective
toward R-alkoxy cyclohexanones than related reactions involv-
ing other organometallics under anhydrous conditions. In fact,
this new reagent system ranks as the most highly selective
allylating agent yet reported for aqueous systems, as long as
steric effects do not gain heightened importance. In the
companion paper,44 neighboring unprotected hydroxyl substit-
uents are shown to be capable of still greater directing effects.
Consequently, this chemistry clearly constitutes a synthetically
useful operation well suited for application to more complex
problems in organic syntheses.
The fortuitous situation that the rate effects arising from
chelation are kinetically dominant has allowed us to probe the
orientational consequences of flanking C-O bond geometry on
stereoselective attack at a cyclohexanone carbonyl. It turns out
that good agreement with a chelate transition state model is seen,
from which heuristic value may be derived. Studies are
Entries 50 and 52 point out the differences in relative rate
associated with axial and equatorial projection of the flanking
ether oxygen. Since chelation to 10 is sterically impeded, it is
not surprising to find that 10 reacts only 2.5 times faster than
11. The extent to which the latter ketone may adopt a twist-
boat conformation in order to engage in chelation is not known.
However, the latter option appears rather unlikely when viewed
in the context of the less than maximal complexing capacity of
10. More revealing in this context is the data associated with
entry 50. In this instance, the two possible epimers are
substituted by methoxyl and are not sterically disadvantaged.
The exclusivity with which 7 reacts in the presence of 8 denotes
the substantive kinetic advantage associated with equatorial
projection of the methoxyl substituent. The corresponding
highly stereoselective behavior of 7, which gives only the
product predicted by Cram’s chelate rule, is in complete accord
with the proposal that allylation proceeds via the chelated
transition state. A provocative aspect of this finding is that the
(44) Paquette, L. A.; Mitzel, T. M. J. Am. Chem. Soc. 1996, 118, 1931.

1926 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
Table 4. Facial Selectivity in Nucleophilic Additions to 10 and
Scheme 8
11a
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments. b The allyl
bromide and indium powder were refluxed in THF for 1 h and cooled
to 25 °C before the ketone was added. c Based on recovered starting
material.
Scheme 7
Experimental Section
Melting points were determined in open capillaries and are uncor-
rected. Mass spectra were recorded at The Ohio State University
Chemical Instrumentation Center. Elemental analyses were performed
by the Scandinavian Microanalytical Laboratory, Herlev, Denmark, and
Atlantic Microlab, Inc., Norcross, GA. All organic solvents were
predried by standard methods. Unless otherwise indicated, all separa-
tions were carried out under flash chromatography conditions on silica
gel 60 (230-400 mesh, 60 Å) using the indicated solvents. The organic
extracts were dried over anhydrous magnesium or sodium sulfate. The
purity of all compounds was shown to be g95% by high-field 1H NMR
analysis.
Additions Involving Allylmagnesium Chloride. A. 2-Methoxy-
cyclohexanone. Magnesium turnings were washed with 10% HCl,
rinsed several times with water, slurried in acetone, and dried under
vacuum before use. A 446 mg (18.3 mmol) sample of magnesium
turnings was flushed with dry nitrogen, covered with dry THF (15 mL),
warmed gently, and treated slowly with allyl chloride (0.872 mL, 10.7
mmol) at a rate sufficient to maintain reflux. After 1 h, the Grignard
solution was cooled to 0 °C. The reaction mixture was stirred at this
temperature for 1.5 h, quenched with saturated NH4Cl solution, and
extracted with ether. The combined organic layers were washed with
brine, dried, and evaporated, and the product alcohols were separated
by chromatography on silica gel (elution with 10% ether in petroleum
ether). The spectral properties of 12 and 13 are identical to those
previously reported.25
underway which should ascertain the extent to which other types
of effects can possibly contribute to rate enhancement and
stereoselection.
The present work establishes for the first time that ketones
are indeed reactive in indium-promoted allylations conducted
in aqueous media. Further, relative reactivities are unquestion-
ably governed by chelation effects, which in turn are especially
sensitive to nonbonded steric congestion. The same correlation
is exhibited by aldehydes44 as well as ketones. In effect, an
unencumbered R-oxygen which is not deactivated by acylation
or silylation is capable of controlling π-facial diastereoselection
in water. The potential synthetic utility of this stereocontrol is
made clear by direct comparison with the diastereoselectivities
of other allylations performed by different metals in anhydrous
organic solvents.
B. 1-Oxaspiro[4.5]decan-6-one. Ketone 9 reacted with equal
1
facility to give 17 and 18, the high-field H and 13C spectra of which
are identical to those previously reported.30
C. cis-2-Methoxy-4-tert-butylcyclohexanone. Submission of 7 to
the same reaction conditions afforded alcohols 23 and 24, which were
separated by MPLC on silica gel (elution with 10% ethyl acetate in
hexanes).
23: colorless oil; IR (neat, cm-1) 3472, 1639; 1H NMR (300 MHz,
C6D6) δ 5.98-5.84 (m, 1 H), 5.10-5.02 (m, 2 H), 3.08 (s, 3 H), 2.75
(dd, J ) 11.2, 4.6 Hz, 1 H), 2.45 (ddt, J ) 13.5, 6.5, 1.3 Hz, 1 H),
2.25 (dd, J ) 13.6, 8.3 Hz, 1 H), 1.98 (d, J ) 2.3 Hz, 1 H), 1.87-1.80

1,2-Addition of Allylmetal Reagents to Cyclohexanones
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1927
Table 5. Competitive Indium-Promoted Allylations in Water at 25
°C
34.2, 31.8, 27.5, 25.3, 23.8; MS m/z (M+) calcd 226.1933, obsd
226.1959. Anal. Calcd for C14H26O2: C, 74.29; H, 11.58. Found:
C, 74.22; H, 11.59.
E. (5R*,9R*)-9-tert-Butyl-1-oxaspiro[4.5]decan-6-one (10). Analo-
gous addition of allylmagnesium chloride to 10 gave rise to 37 and
38, separation of which by chromatography on silica gel (elution with
19:1 f 6:1 hexanes/ethyl acetate) provided the pure diastereomers.
37: colorless oil; IR (neat, cm-1) 3478, 1065; 1H NMR (300 MHz,
C6D6) δ 6.22-6.08 (m, 1 H), 5.14-5.08 (m, 2 H), 3.60-3.53 (m, 2
H), 2.37 (ddt, J ) 13.8, 5.4, 1.6 Hz, 1 H), 2.27 (br s, 1 H), 2.02-1.87
(m, 3 H), 1.63 (td, J ) 12.8, 3.8 Hz, 1 H), 1.57-1.29 (m, 6 H), 1.06-
0.74 (m, 2 H), 0.85 (s, 9 H); 13C NMR (75 MHz, C6D6) δ 135.9, 116.7,
88.4, 74.3, 67.6, 46.2, 41.4, 36.0, 33.7, 32.6, 32.3, 27.7, 27.4, 21.9;
MS m/z (M+) calcd 252.2089, obsd 252.2085. Anal. Calcd for
C16H28O2: C, 76.14; H, 11.18. Found: C, 76.34; H, 11.21.
38: colorless oil; IR (neat, cm-1) 3476, 1365, 1068; 1H NMR (300
MHz, C6D6) δ 6.12-5.98 (m, 1 H), 5.13-5.06 (m, 2 H), 3.76-3.69
(m, 1 H), 3.62 (dd, J ) 14.5, 7.8 Hz, 1 H), 2.76 (ddd, J ) 14.6, 6.1,
1.4 Hz, 1 H), 2.50 (dd, J ) 14.6, 8.5 Hz, 1 H), 2.31 (ddd, J ) 12.1,
8.7, 5.4 Hz, 1 H), 1.86-1.80 (m, 1 H), 1.74-1.53 (m, 3 H), 1.51-
1.21 (m, 6 H), 1.05-0.98 (m, 1 H), 0.78 (s, 9 H); 13C NMR (75 MHz,
C6D6) ppm 135.5, 117.6, 88.6, 75.5, 67.7, 45.8, 38.4, 36.4, 33.8, 32.1,
31.9, 27.7, 27.3, 23.3; MS m/z (M+) calcd 252.2089, obsd 252.2079.
Anal. Calcd for C16H28O2: C, 76.14; H, 11.18. Found: C, 75.97; H,
11.23.
F. (5R*,9S*)-9-tert-Butyl-1-oxaspiro[4.5]decan-6-one (11). Com-
parable treatment of 11 produced a mixture of 41 and 42, which were
separated chromatographically on silica gel (elution with 10:1 hexanes/
ethyl acetate).
41: colorless solid, mp 46-47 °C; IR (film, cm-1) 3498, 1467, 1437,
1064; 1H NMR (300 MHz, CDCl3) δ 5.96-5.82 (m, 1 H), 5.17-5.07
(m, 1 H), 3.88-3.73 (m, 2 H), 2.32 (dd, J ) 13.6, 7.1 Hz, 1 H), 2.26-
2.16 (m, 2 H), 1.95-1.72 (series of m, 2 H), 1.65-1.18 (series of m,
10 H), 0.84 (s, 9 H); 13C NMR (75 MHz, CDCl3) δ 134.3, 118.8, 86.3,
73.7, 68.0, 42.5, 40.3, 35.1, 33.9, 33.2, 31.9, 27.4, 26.3, 21.5; MS m/z
(M+) calcd 252.2089, obsd 252.2083. Anal. Calcd for C16H28O2: C,
76.14; H, 11.18. Found: C, 76.24; H, 11.17.
42: colorless oil; IR (neat, cm-1) 3563, 1365, 1055; 1H NMR (300
MHz, CDCl3) δ 6.01-5.87 (m, 1 H), 5.12-5.05 (m, 2 H), 3.92-3.79
(m, 2 H), 2.31 (dd, J ) 14.2, 8.4 Hz, 1 H), 2.24-2.12 (m, 2 H), 1.95-
1.81 (m, 3 H), 1.72 (dt, J ) 13.7, 3.0 Hz, 1 H), 1.64-1.34 (series of
m, 5 H), 1.12 (t, J ) 13.2 Hz, 1 H), 1.04 (dq, J ) 12.9, 3.6 Hz, 1 H),
0.84 (s, 9 H); 13C NMR (75 MHz, CDCl3) δ 134.1, 117.2, 87.8, 74.0,
67.7, 42.1, 37.9, 34.1, 34.0, 31.8, 31.2, 27.5, 25.6, 23.6; MS m/z (M+)
calcd 252.2089, obsd 252.2088. Anal. Calcd for C16H28O2: 76.14;
H, 11.18. Found: C, 75.79; H, 11.13.
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments. b The relative
rate ratio was ascertained by quantitative analysis of the alcohol products
formed. c The relative rate ratio was determined by quantitative
assessment of the levels of unreacted ketones remaining.
(m, 2 H), 1.53 (dq, J ) 12.7, 3.6 Hz, 1 H), 1.40-1.32 (m, 2 H), 1.06
(tq, J ) 13.6, 2.2 Hz, 1 H), 0.85 (s, 9 H), 0.79 (tt, J ) 12.4, 3.1 Hz,
1 H); 13C NMR (75 MHz, C6D6) δ 135.3, 117.3, 83.0, 72.5, 56.2, 46.5,
45.3, 34.4, 32.4, 27.7, 26.9, 22.1; MS m/z (M+ - C3H5) calcd 185.1542,
obsd 185.1546. Anal. Calcd for C14H26O2: C, 74.29; H, 11.58.
Found: C, 74.48; H, 11.60.
24: colorless crystals, mp 47-48 °C; IR (film, cm-1) 3484, 1639,
1
1467, 1393, 1366, 1328, 1245, 1186; H NMR (300 MHz, C6D6) δ
Prototypical CeCl3-Mediated Coupling. A 364 mg (0.98 mmol)
sample of cerium trichloride heptahydrate was dried according to
standard protocol.45 The cooled salt was slurried with anhydrous THF
(3.6 mL) for 3 h under N2 at room temperature (rt). A solution of
allylmagnesium chloride (1.1 mL of 1 M in THF, 1.1 mmol), prepared
in the predescribed manner, was added via syringe into this slurry, and
the reaction mixture was stirred at 0 °C for 20 min. The ketone (0.15
mmol) dissolved in dry THF (1.0 mL) was next introduced slowly at
0 °C. Upon completion of the addition (monitored by TLC, most
commonly 45 min), aqueous NH4Cl solution was added and the usual
workup was applied.
5.99-5.85 (m, 1 H), 5.09-5.03 (m, 2H), 3.07 (s, 3 H), 2.68 (dd, J )
11.2, 4.9 Hz, 1 H), 2.45 (dd, J ) 13.6, 6.4 Hz, 1 H), 2.26 (dd, J )
13.6, 8.4 Hz, 1 H), 1.97 (s, 1 H), 1.90 (dt, J ) 13.5, 3.1 Hz, 1 H),
1.86-1.48 (m, 4 H), 0.93 (br t, J ) 12.9 Hz, 1 H), 0.83 (s, 9 H),
0.87-0.65 (m, 1 H); 13C NMR (75 MHz, C6D6) δ 135.3, 117.3, 82.3,
73.3, 56.1, 45.8, 41.7, 35.9, 32.0, 27.6, 25.9, 25.3; MS m/z (M+
-
C3H5) calcd 185.1541, obsd 185.1548. Anal. Calcd for C14H26O2: C,
74.29; H, 11.58. Found: C, 74.06; H, 11.56.
D. trans-2-Methoxy-4-tert-butylcyclohexanone. Comparable treat-
ment of 8 provided 28 and 29, which were separated by chromatography
on silica gel (elution with 10% ethyl acetate in hexanes).
28: colorless oil; IR (neat, cm-1) 3475, 1365, 1100; 1H NMR (300
MHz, CDCl3) δ 5.74-5.44 (m, 1 H), 5.18-5.09 (m, 2 H), 3.74 (s, 3
H), 3.04 (br t, J ) 1.6 Hz, 1 H), 2.40 (dd, J ) 13.6, 7.3 Hz, 1 H), 2.16
(dd, J ) 13.6, 7.9 Hz, 1 H), 1.93-1.87 (m, 1 H), 1.64-1.25 (m, 6 H),
0.86 (s, 9 H); 13H NMR (75 MHz, CDCl3) ppm 133.8, 119.1, 81.3,
71.6, 56.1, 43.8, 40.3, 33.6, 32.0, 27.4, 24.2, 21.9; MS m/z (M+) calcd
226.1933, obsd 226.1924. Anal. Calcd for C14H26O2: C, 74.29; H,
11.58. Found: C, 73.68; H, 11.55.
Prototypical CrCl2-Promoted Coupling. Freshly distilled thionyl
chloride (45 mL) was added dropwise during 10 min to finely divided
chromium trichloride hexahydrate (17.5 g, 65.7 mmol). The magneti-
cally stirred mixture was slowly warmed to reflux, maintained at this
temperature for 2.5 h, and freed of the excess thionyl chloride under
reduced pressure to leave a purple-rose powder. This solid was stored
overnight in a desiccator charged with KOH pellets and free of
oxygen.46
29: colorless oil; IR (neat, cm-1) 3566, 1391, 1366, 1094; 1H NMR
(300 MHz, CDCl3) δ 6.02-5.84 (m, 1 H), 5.12-5.03 (m, 2 H), 3.35
(s, 3 H), 3.19 (br t, J ) 1.8 Hz, 1 H), 2.82 (br s, 1 H), 2.28 (ABq of
ABX, ∆µ ) 15.8 Hz, JAB ) 14.3 Hz, JAX ) 6.9 Hz, JBX ) 7.4 Hz, 2
H), 2.00 (br dq, J ) 14.1, 3.1 Hz, 1 H), 1.69-1.51 (m, 3 H), 1.31 (br
tt, J ) 12.6, 3.1 Hz, 1 H), 1.18-0.97 (m, 2 H), 0.87 (s, 9 H); 13C
NMR (75 MHz, CDCl3) δ 133.7, 117.4, 82.2, 72.4, 56.4, 39.9 (2 C),
A slurry of anhydrous chromium trichloride (4.32 g, 23.7 mmol) in
cold (0 °C), dry THF (54 mL) was cautiously treated in five portions
with lithium aluminum hydride (517 mg, 13.6 mmol), stirred for 15
(45) (a) Imamoto, T.; Takiyama, N.; Nakumura, K. Tetrahedron Lett.
1985, 26, 4763. (b) Encyclopedia of Reagents for Organic Synthesis;
Paquette, L. A., Ed.; 1995; Volume 2, pp 1031-1034.
(46) Pray, A. R. Inorg. Synth. 1954, 5, 153.

1928 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
min at this temperature, and allowed to warm to 20 °C during another
20 min, at which point the reaction mixture became characteristic dark
brown in color. The mixture was recooled to 0 °C, the ketone (3.90
mmol) was introduced via syringe, and stirring was maintained for 15
min before allyl bromide (1.42 g, 11.7 mmol) was added dropwise.
After 3 h at 0 °C and 30 min at rt, the reaction mixture was quenched
with saturated NaHCO3 solution and diluted with ether. The aqueous
phase was extracted with CH2Cl2 (3×), and the combined organic
extracts were washed with brine (2×), dried, and concentrated.
Prototypical Indium-Promoted Couplings. A. In Dry THF. To
a magnetically stirred suspension of indium powder (168 mg, 1.46
mmol) in dry THF (5 mL) was introduced allyl bromide (0.126 mL,
1.46 mmol) via syringe, and the mixture was agitated for 5 min prior
to introduction of neat ketone (0.97 mmol). Reaction was allowed to
proceed for 3 h at room temperature, 10% hydrochloric acid was then
added, and the usual workup conditions were applied.
B. In 50% Aqueous THF. To a reaction vessel containing allyl
bromide (26 mg, 0.21 mmol), indium powder (37 mg, 0.32 mmol),
and the ketone (0.19 mmol) were added THF (2.0 mL) and water (2.0
mL). The reaction mixture was stirred at rt until ketone consumption
was complete (TLC analysis), quenched with 10% HCl, and extracted
with ether. The combined organic phases were washed once with brine,
dried, and concentrated to leave an oily residue that was purified
chromatographically.
C. In Water. A mixture of allyl bromide (30 mg, 0.25 mmol),
indium powder (29 mg, 0.25 mmol), ketone (0.15 mmol), and water
(2.0 mL) was stirred at rt until ketone consumption was complete (TLC
analysis). A workup identical to that in B was employed.
D. In THF with Prior Formation of the Allylindium Reagent.
A magnetically stirred mixture of allyl bromide (31 mg, 0.26 mmol),
indium powder (28 mg, 0.25 mmol), and THF (0.71 mL) was gently
refluxed for 1 h, cooled to rt, and treated with the ketone (0.181 mmol)
dissolved in 0.71 mL of THF. The reaction mixture was stirred for 1
h, filtered through a small pad of silica gel (elution with ethyl acetate),
and concentrated. Diastereomer separation was accomplished by
chromatography on silica gel.
Prototypical Normant Alkylations. A cold (0 °C), magnetically
stirred solution of 3-chloro-1-propanol (3.78 g, 40 mmol) in dry THF
(40 mL) was treated dropwise with a solution of isopropylmagnesium
chloride in ether (20 mL of 2.0 M, 40 mmol) during 20 min and
subsequently warmed to room temperature. Flame-dried magnesium
turnings (1.46 g, 60 mmol) were quickly added, and the mixture was
refluxed for 1 h and left to stand overnight without stirring in order to
allow the excess magnesium to settle. The concentration of active
Grignard reagent was established by titration of a small portion of the
supernatant with menthol and 1,10-phenanthroline according to estab-
lished procedure.47
To a cold (0 °C), magnetically stirred solution of the ketone (1.93
mmol) in dry THF (7.5 mL) was added dropwise the Normant reagent
(5.88 mL of 0.365 M, 2.15 mmol). The reaction mixture was stirred
at 0 °C for 1 h, quenched with 10% HCl, and diluted with ether. The
organic phase was washed with brine (2×), dried, and concentrated.
The spectral properties for 14 and 1525 and of 19 and 2030 are
identical to those previously reported.
NMR (75 MHz, C6D6) δ 82.4, 72.1, 62.5, 55.3, 40.7, 36.6, 34.4, 31.0,
26.7, 26.3, 25.0, 24.2; MS m/z (M+) calcd 244.2038, obsd 244.2034.
Anal. Calcd for C14H28O3: C, 68.81; H, 11.55. Found: C, 68.98; H,
11.38.
(1R*,2S*,4R*)-1-(3-Hydroxypropyl)-2-methoxy-4-tert-butylcyclo-
hexanol (30): colorless solid, mp 114-115 °C; IR (CCl4, cm-1) 3334,
1
1549, 1252, 1218, 1098, 1005; H NMR (300 MHz, CDCl3) δ 3.73-
3.60 (m, 2 H), 3.24 (s, 3 H), 3.19 (s, 1 H), 2.40 (br s, 2 H), 1.92 (br d,
J ) 13.5 Hz, 1 H), 1.77-1.25 (series of m, 10 H), 0.86 (s, 9 H); 13C
(75 MHz, CDCl3) δ 80.9, 71.8, 63.5, 56.1, 40.3, 36.4, 33.8, 32.0, 27.4,
25.7, 24.0, 21.4; MS m/z (M+) calcd 244.2038, obsd 244.2041. Anal.
Calcd for C14H28O3: C, 68.81; H, 11.55. Found: C, 68.98; H, 11.40.
(1S*,2S*,4R*)-1-(3-Hydroxypropyl)-2-methoxy-4-tert-butylcyclo-
hexanol (31): colorless oil; IR (neat, cm-1) 3388, 1460, 1393, 1365,
1
1097; H NMR (300 MHz, CDCl3) δ 3.67-3.63 (m, 2 H), 3.35 (s, 3
H), 3.21 (br d, J ) 1.7 Hz, 1 H), 2.68 (br s, 2 H), 2.02 (dq, J ) 14.0,
3.1 Hz, 1 H), 1.74-1.51 (series of m, 5 H), 1.31 (tt, J ) 12.6, 3.1 Hz,
1 H), 1.20 (t, J ) 6.9 Hz, 1 H), 1.18-0.87 (series of m, 3 H), 0.85 (s,
9 H); 13C NMR (75 MHz, CDCl3) δ 82.5, 72.6, 63.3, 56.4, 40.1, 33.7,
32.1, 31.8, 27.5, 26.1, 25.5, 23.9; MS m/z (M+) calcd 244.2038, obsd
244.2033. Anal. Calcd for C14H28O3: C, 68.81; H, 11.55. Found:
C, 68.40; H, 11.42.
(5R*,6R*,9R*)-9-tert-Butyl-6-hydroxy-1-oxaspiro[4.5]decane-6-
propanol (39): colorless solid, mp 98.9-99.5 °C; IR (film, cm-1) 3428,
1
1468, 1365, 1065; H NMR (300 MHz, C6D6) δ 3.66-3.46 (m, 4 H),
2.05-1.73 (m, 4 H), 1.68-1.19 (series of m, 10 H), 0.93-0.76 (m, 2
H), 0.86 (s, 9 H); 13C NMR (75 MHz, C6D6) δ 88.7, 74.4, 67.3, 63.6,
46.0, 36.0, 33.1, 32.8, 32.7, 32.3, 27.7, 27.5, 27.1, 21.9; MS m/z (M+)
calcd 270.2195; obsd 270.2197. Anal. Calcd for C16H30O3: C, 71.07;
H, 11.18. Found: C, 70.98; H, 11.10.
(5R*,6S*,9R*)-9-tert-Butyl-6-hydroxy-1-oxaspiro[4.5]decane-6-
propanol (40): colorless solid, mp 106-107 °C; IR (film, cm-1) 3404,
1459, 1365, 1067, 1027, 1002; 1H NMR (300 MHz, C6D6) δ 3.77 (td,
J ) 7.6, 5.0 Hz, 1H), 3.65 (td, J ) 7.8, 6.7 Hz, 1 H), 3.56-3.44 (m,
2 H), 2.38-2.29 (m, 1 H), 2.06 (br s, 2 H), 1.94-1.25 (series of m, 12
H), 1.03-0.93 (m, 2 H), 0.80 (s, 9 H); 13C NMR (75 MHz, C6D6) δ
89.2, 75.5, 67.6, 63.4, 45.8, 36.4, 33.2, 32.1, 31.9, 29.9, 27.7, 27.3,
26.6, 23.5; MS m/z (M+) calcd 270.2195, obsd 270.2192. Anal. Calcd
for C16H30O3: C, 71.07; H, 11.18. Found: C, 71.11; H, 11.12.
(5R*,6S*,9S*)-9-tert-Butyl-6-hydroxy-1-oxaspiro[4.5]decane-6-
propanol (43): colorless solid, mp 109 °C; IR (KBr, cm-1) 3384, 1415,
1364, 1197, 1059, 1003, 972; 1H NMR (300 MHz, C6D6) δ 3.73-3.59
(m, 2 H), 3.50-3.43 (m, 1 H), 3.38-3.31 (m, 1 H), 2.18 (qd, J ) 8.9,
5.9 Hz, 1 H), 2.04 (br s, 1 H), 1.79-1.28 (series of m, 15 H), 0.92 (s,
9 H); 13C NMR (75 MHz, C6D6) δ 86.9, 73.7, 68.0, 63.5, 42.8, 35.3,
33.4, 33.2, 32.6, 32.1, 27.7, 26.7, 26.3, 22.1; MS m/z (M+) calcd
270.2195, obsd 270.2196. Anal. Calcd for C16H30O3: C, 71.07; H,
11.18. Found: C, 71.51; H, 10.97.
(5R*,6R*,9S*)-9-tert-Butyl-6-hydroxy-1-oxaspiro[4.5]decane-6-
propanol (44): colorless solid, mp 99.8-100.2 °C; IR (film, cm-1
)
1
3417, 1366, 1055; H NMR (300 MHz, C6D6) δ 3.70-3.54 (m, 3 H),
3.48 (dd, J ) 15.4, 7.6 Hz, 1 H), 2.7-2.2 (br, 2 H), 2.07 (ddd, J )
12.0, 9.3, 9.3 Hz, 1 H), 1.97 (dt, J ) 13.0, 3.3 Hz, 1 H), 1.90-1.79
(m, 1 H), 1.71-1.35 (series of m, 9 H), 1.16-1.09 (m, 1 H), 1.03-
0.89 (m, 2 H), 0.80 (s, 9 H); 13C NMR (75 MHz, C6D6) δ 88.1, 74.3,
67.6, 63.3, 42.6, 34.4, 33.7, 31.8, 31.3, 30.1, 27.7, 26.8, 25.8, 24.3;
MS m/z (M+) calcd 270.2195, obsd 270.2189. Anal. Calcd for
C16H30O3: C, 71.07; H, 11.18. Found: C, 70.88; H, 11.13.
(1R*,2R*,4R*)-1-(3-Hydroxypropyl)-2-methoxy-4-tert-butylcyclo-
hexanol (25): colorless solid, mp 86 °C; IR (film, cm-1) 3734, 1644,
1
1422, 1265; H NMR (300 MHz, C6D6) δ 3.51-3.42 (m, 2 H), 3.08
(s, 3 H), 2.67 (dd, J ) 11.2, 4.6 Hz, 1 H), 1.93 (dt, J ) 13.8, 3.3 Hz,
1 H), 1.87-1.26 (series of m, 10 H), 0.90 (dd, J ) 13.6, 4.3 Hz, 1 H),
0.86 (s, 9 H), 0.84-0.74 (m, 1 H); 13C NMR (75 MHz, C6D6) δ 84.3,
72.2, 63.5, 56.5, 46.4, 37.2, 34.0, 32.4, 27.7, 27.4, 27.1, 22.1; MS m/z
(M+) calcd 244.2038, obsd 244.2038. Anal. Calcd for C14H28O3: C,
68.81; H, 11.55. Found: C, 68.96; H, 11.50.
(2S*,4S*)-2-Methoxy-4-tert-butylcyclohexanone (7) and (2R*,4S*)-
2-Methoxy-4-tert-butylcyclohexanone (8). A nitrogen-blanketed,
magnetically stirred suspension of iodosobenzene48 (20.5 g, 0.094
mmol) in dry methanol (500 mL) was treated with boron trifluoride
etherate (20.39 mL, 0.169 mol) via syringe to give a clear solution
which was subsequently cooled to -78 °C. A solution of 2232 (20.0
g, 0.088 mol) in dry methanol (200 mL) was cooled to -78 °C and
introduced slowly dropwise via cannula. The reaction mixture was
stirred for 3 h at -78 °C, allowed to stir at rt for 2 h, and freed of
methanol under reduced pressure. The residue was diluted with CH2Cl2
and washed three times with saturated NaHCO3 solution and once with
(1S*,2R*,4R*)-1-(3-Hydroxypropyl)-2-methoxy-4-tert-butylcyclo-
hexanol (26): colorless solid, mp 104 °C; IR (film, cm-1) 3943, 3430,
1630; 1H NMR (300 MHz, C6D6) δ 3.47-3.42 (m, 2 H), 3.07 (s, 3 H),
2.59 (dd, J ) 11.2, 4.8 Hz, 1 H), 1.98 (dt, J ) 13.4, 3.1 Hz, 1 H),
1.81-1.31 (series of m, 10 H), 0.85 (s, 9 H), 0.86-0.64 (m, 2 H); 13
C
(47) (a) Furniss, B. S.; Hannaford, A. J.; Smith, P. W. G.; Tatchell, A.
R. In Vogel’s Textbook of Practical Organic Chemistry, 5th ed.; Long-
man: Essex, U.K., 1989; pp 443-444. (b) Watson, S. C.; Eastham, J. F. J.
Organomet. Chem. 1967, 9, 165.
(48) Saltzman, H.; Sharefkin, J. S. Organic Synthesis; Wiley: New
York,1973; Collect. Vol. V, p 658.

1,2-Addition of Allylmetal Reagents to Cyclohexanones
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1929
brine prior to drying. Chromatography on silica gel, gradient elution
with 15:1 f 9:1 hexanes/ethyl acetate, afforded 2.00 g (12%) of 8 and
8.15 g (50%) of 7. The spectral properties of these ketones are identical
to those previously reported.31
δ 111.6, 76.1, 65.9, 65.3, 63.3, 44.2, 36.4, 32.7, 32.2, 30.3, 27.6, 26.8,
24.5; MS m/z (M+) calcd 272.1987, obsd 272.2000. Anal. Calcd for
C15H28O4: C, 66.14; H, 10.36. Found: C, 66.15; H, 10.33.
48: colorless crystals, mp 107 °C; IR (film, cm-1) 3398, 1365, 1189,
1
6-[(E)-Benzylidene]-8-tert-butyl-1,4-dioxaspiro[4.5]decane (33).49
A cold (-78 °C), nitrogen-blanketed solution of diisopropylamine (43.2
mL, 0.311 mmol) in dry THF (650 mL) was treated with n-butyllithium
(210.5 mL of 1.6 M in hexanes, 0.337 mol), allowed to warm to rt
during 1 h, and returned to -78 °C before a solution of 21 (40.0 g,
0.259 mol) in dry THF (400 mL) was slowly introduced via cannula.
After this mixture had stirred in the cold for 1 h, benzaldehyde (26.3
mL, 0.233 mol) was added, followed 5 s later by methanesulfonyl
chloride (40.0 mL, 0.508 mol). The reaction mixture was warmed to
rt during 2 h, treated with triethylamine (300 mL), and refluxed for 1
h or stirred overnight at rt prior to dilution with water (400 mL) and
extraction with ether (4 × 100 mL). The combined organic layers
were washed with saturated NH4Cl solution and brine, dried, and
evaporated to leave 31.5 g of a yellow oil which was used without
further purification.
1151, 1088; H NMR (300 MHz, C6D6) δ 3.56-3.42 (m, 6 H), 2.13
(td, J ) 13.0, 4.0 Hz, 1 H), 2.04 (dt, J ) 13.2, 3.0 Hz, 1 H), 1.80-
1.53 (m, 7 H), 1.45-1.30 (m, 1 H), 1.33 (t, J ) 12.8 Hz, 1 H), 0.90
(s, 9 H); 13C NMR (75 MHz, C6D6) δ 111.5, 75.5, 65.4, 65.1, 63.6,
41.4, 34.4, 32.2, 32.1, 31.3, 27.8, 26.6, 24.3; MS m/z (M+) calcd
272.1987, obsd 272.2010. Anal. Calcd for C15H28O4: C, 66.14; H,
10.36. Found: C, 66.11; H, 10.36.
The above oil was taken up in ethylene glycol (215 mL, 5.73 mol)
and trimethyl orthoformate (200 mL, 1.83 mol), treated with p-
toluenesulfonic acid (1.24 g, 6.52 mmol), stirred overnight, quenched
with one drop of triethylamine, diluted with 50% saturated NaHCO3
solution, and extracted with ether. The combined organic phases were
dried and evaporated to leave a thick oil which was crystallized from
95% ethanol. There was obtained 37.0 g (50%) of 33 as a fluffy cream-
colored solid: mp 87.5 °C; IR (film, cm-1) 1479, 1366, 1188, 1098;
1H NMR (300 MHz, C6D6) δ 7.50-7.47 (m, 2 H), 7.41-7.37 (m, 3
H), 7.29-7.23 (m, 1 H), 3.96-3.83 (m, 2 H), 3.81-3.73 (m, 2 H),
3.27 (br d, 1 H), 2.35-2.24 (m, 2 H), 2.11-1.97 (m, 1 H), 1.94-1.86
(m, 2 H), 1.44-1.34 (m, 1 H), 0.99 (s, 9 H); 13C NMR (75 MHz, C6D6)
δ 141.6, 138.1, 129.3, 128.4, 126.7, 122.1, 109.0, 65.4, 63.6, 49.2, 38.1,
32.7, 28.7, 27.5, 25.1; MS m/z (M+) calcd 286.1933, obsd 286.1930.
Anal. Calcd C19H26O2: C, 79.68; H, 9.15. Found: C, 79.37; H, 9.21.
8-tert-Butyl-1,4-dioxaspiro[4.5]decan-6-one (34). A solution of 33
(5.00 g, 17.5 mmol) in a 1:1 mixture of methanol and CH2Cl2 (150
mL) containing 1% pyridine was ozonolyzed at -78 °C until a faint
blue color appeared. The reaction mixture was purged with oxygen,
treated with dimethyl sulfide (8 mL), allowed to warm to rt for 2 h,
diluted with 1% NaHCO3 solution, and extracted with CH2Cl2. The
combined organic layers were dried and concentrated, and the residue
was chromatographed on silica gel (elution with 8:1 hexanes/ethyl
acetate) to give 3.37 g (91%) of 34 as a faintly yellow oil: IR (neat,
cm-1) 1731, 1479, 1367, 1103, 1063, 1029; 1H NMR (300 MHz, C6D6)
δ 3.99 (dt, J ) 14.1, 7.3 Hz, 1 H), 3.63 (dt, J ) 13.1, 7.3 Hz, 1 H),
3.47-3.36 (m, 2 H), 2.44-2.31 (m, 2 H), 1.99 (dt, J ) 13.5, 3.2 Hz,
1 H), 1.74-1.63 (m, 1 H), 1.59-1.43 (m, 2 H), 1.15-1.04 (m, 1 H),
0.63 (s, 9 H); 13C NMR (75 MHz, C6D6) δ 205.1, 106.6, 66.1, 64.2,
48.3, 41.4, 35.7, 32.3, 27.1, 24.1; MS m/z (M+) calcd 212.1412, obsd
212.1388. Anal. Calcd for C12H20O3: C, 67.89; H, 9.50. Found: C,
67.52; H, 9.51.
(6R*,12R*)-12-tert-Butyl-1,4,7-trioxadispiro[4.0.4.4]tetradecane
(35). To a magnetically stirred solution of 47 (2.07 g, 7.60 mmol) in
dry CH2Cl2 (35 mL) was added trimethylamine (2.4 mL, 17.1 mmol),
p-toluenesulfonyl chloride (1.39 g, 9.89 mmol), and DMAP (46 mg,
0.38 mmol). The reaction mixture was stirred overnight at 45 °C,
cooled, diluted with ether, and washed with brine prior to drying and
concentration. Purification of the residue by chromatography on silica
gel (elution with 8:1 hexanes/ethyl acetate) afforded 1.54 g (79%) of
35 as a viscous, colorless oil: IR (film, cm-1) 1514, 1468, 1442, 1394,
1366; 1H NMR (300 MHz, CDCl3) δ 4.21-4.12 (m, 1 H), 4.00-3.88
(m, 3 H), 3.87-3.75 (m, 2 H), 2.11-1.79 (m, 3 H), 1.72-1.45 (m, 6
H), 1.38-1.31 (m, 1 H), 1.14-1.02 (m, 1 H), 0.89 (s, 9 H); 13C NMR
(75 MHz, CDCl3) δδ 111.2, 87.2, 68.2, 65.9, 65.7, 45.7, 37.4, 34.4,
32.7, 32.1, 27.7, 26.7, 23.9; MS m/z (M+) calcd 254.1882, obsd
254.1890. Anal. Calcd for C15H20O3: C, 70.83; H, 10.30. Found:
C, 70.75; H, 10.25.
(6R*,12S*)-12-tert-Butyl-1,4,7-trioxadispiro[4.0.4.4]tetradecane
(36). Analogous cyclization of 48 (2.69 g, 9.87 mmol) furnished 1.73
g (69%) of 36 as a faintly yellow oil following chromatographic
purification on silica gel (elution with 8:1 hexanes/ethyl acetate): IR
(neat, cm-1) 2951, 2871, 1478, 1435, 1366, 1186; 1H NMR (300 MHz,
C6D6) δ 3.98 (dd, J ) 14.5, 7.6 Hz, 1 H), 3.81 (ddd, J ) 15.0, 7.4, 4.8
Hz, 1 H), 3.60-3.49 (series of m, 4 H), 2.18-2.04 (m, 2 H), 1.82-
1.56 (series of m, 7 H), 1.44-1.30 (m, 2 H), 0.87 (s, 9 H); 13C NMR
(75 MHz, C6D6) δ 111.6, 85.8, 68.9, 65.1, 64.5, 42.6, 38.4, 32.6, 32.0,
31.1, 27.8, 26.6, 24.4; MS m/z (M+) calcd 254.1882, obsd 254.1883.
Anal. Calcd for C15H26O3: C, 70.83; H, 10.30. Found: C, 70.78; H,
10.35.
(5R*,9R*)-9-tert-Butyl-1-oxaspiro[4.5]decan-6-one (10). A mag-
netically stirred solution of 35 (1.54 g, 6.01 mmol) in acetone (15 mL)
and water (15 mL) was treated with concentrated HCl (0.1 mL), refluxed
for 12 h, quenched with a few drops of triethylamine, and extracted
with ether. The combined organic phases were dried and concentrated
to leave a residue that was chromatographed on silica gel (elution with
8:1 hexanes/ethyl acetate). There was isolated 1.11 g (88%) of 10 as
a colorless solid: mp 65 °C; IR (film, cm-1) 3014, 2694, 2870, 1718,
1216, 1074; 1H NMR (300 MHz, C6D6) δ 4.08 (dd, J ) 14.2, 7.1 Hz,
1 H), 3.81 (ddd, J ) 13.9, 7.3, 2.2 Hz, 1 H), 2.24 (ddd, J ) 14.8, 4.2,
2.6 Hz, 1 H), 1.93 (ddd, J ) 14.7, 13.5, 6.1 Hz, 1 H), 1.78 (ddd, J )
12.9, 3.3, 3.3 Hz, 1 H), 1.68 (dd, J ) 12.4, 12.4 Hz, 1 H), 1.63-1.42
(m, 4 H), 1.37-1.28 (m, 1 H), 1.12 (dddd, J ) 12.2, 12.2, 3.4, 2.8
Hz,1 H), 1.02 (dddd, J ) 12.4, 12.4, 4.3, 1.1 Hz, 1 H), 0.69 (s, 9 H);
13C NMR (75 MHz, CDCl3) δ 211.3, 87.7, 68.4, 45.7, 40.5, 38.8, 35.2,
32.2, 27.5, 27.4, 25.2; MS m/z (M+) calcd 210.1620, obsd 210.1622.
Anal. Calcd for C13H22O2: C, 74.24; H, 10.54. Found: C, 74.20; H,
10.53.
(6R*,8R*)-8-tert-Butyl-6-hydroxy-1,4-dioxaspiro[4.5]decane-6-
propanol (47) and (6R*,8S*)-8-tert-Butyl-6-hydroxy-1,4-dioxaspiro-
[4.5]decane-6-propanol (48). A magnetically stirred solution of 34
(1.21 g, 5.66 mmol) in dry THF (75 mL) was cooled to 0 °C and treated
dropwise via syringe with a solution of the Normant reagent28 in THF
(14.15 mL, 6.79 mmol). After completion of the addition, the reaction
mixture was stirred for 15 min at 0 °C, quenched with saturated NH4Cl
solution, and extracted with ether. The combined organic phases were
washed with brine, dried, and evaporated. Chromatography of the
residue on silica gel (elution with 90% ethyl acetate in hexanes) afforded
1.33 g (84%) of a 4.4:1 mixture of 47 and 48. A second chromatog-
raphy under more controlled conditions afforded pure samples of the
two 1,4-diols.
47: colorless crystals, mp 62-66 °C; IR (film, cm-1) 3406, 1449,
1
1366, 1184; H NMR (300 MHz, C6D6) δ 3.78-3.42 (series of m, 6
(5R*,9S*)-9-tert-Butyl-1-oxaspiro[4.5]decan-6-one (11). Analo-
gous treatment of 36 (461 mg, 1.80 mmol) provided 274 mg (74%) of
H), 2.89 (br s, 1 H), 2.43 (br s, 1 H), 2.09-2.03 (m, 1 H), 1.99 (dt, J
) 12.8, 3.6 Hz, 1 H), 1.92-1.73 (m, 2 H), 1.69-1.57 (m, 2 H), 1.54-
1.44 (m, 1 H), 1.32 (t, J ) 12.3 Hz, 1 H), 1.34-1.22 (m, 1 H), 1.04
(tt, J ) 12.5, 3.0 Hz, 1 H), 0.83 (s, 9 H); 13C NMR (75 MHz, C6D6)
11 as a clear oil following chromatographic purification: IR (neat, cm-1
)
1721, 1480, 1428, 1366, 1234, 1096, 1049; 1H NMR (300 MHz, C6D6)
δ 3.63-3.56 (m, 1 H), 3.48-3.41 (m, 1 H), 2.83 (ddd, J ) 14.1, 13.3,
5.9 Hz, 1 H), 2.12 (ddd, J ) 13.3, 8.5, 5.1 Hz, 1 H), 2.23 (ddd, J )
(49) We thank Mr. Anthony Lombardo for performing this experiment.

1930 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Lobben
1
13.2, 6.7, 3.7 Hz, 1 H), 1.97-1.85 (m, 2 H), 1.74-1.63 (m, 2 H),
1.56-1.43 (m, 1 H), 1.19 (dd, J ) 12.9, 12.9 Hz, 1 H), 1.13-0.97 (m,
2 H), 0.73 (s, 9 H); 13C NMR (75 MHz, C6D6) δ 208.6, 86.4, 68.0,
43.0, 40.5, 38.0, 31.9, 30.9, 28.2, 27.6, 26.1; MS m/z (M+) calcd
210.1620, obsd 210.1608. Anal. Calcd for C13H22O2: C, 74.24; H,
10.54. Found: C, 74.37; H, 10.64.
Hydroboration-Oxidation of 23. A cold (0 °C), nitrogen-
blanketed, magnetically stirred solution of 23 (198 mg, 0.87 mmol) in
dry THF (2.0 mL) was treated dropwise with the borane-dimethyl
sulfide complex (0.48 mL of 2.0 M in THF, 0.963 mmol) via syringe,
and the reaction was allowed to proceed at 0 °C for 2 h and at rt for
2 h. With external ice cooling, 3 M NaOH (1 mL) and 30% hydrogen
peroxide (1 mL) were introduced and the product was extracted into
ethyl acetate. The combined organic layers were washed with brine,
dried, and concentrated to leave a residue which was chromatographi-
cally purified (silica gel, elution with 2:1 ethyl acetate/hexanes). There
was isolated 101 mg (47%) of 25 as a fluffy white solid, mp 86 °C.
Hydroboration-Oxidation of 28. Analogous treatment of 28 (36
mg, 0.16 mmol) furnished 22 mg (57%) of 30 as a white solid, mp
114-115 °C.
46: colorless solid, mp 53-54 °C; H NMR (300 MHz, C6D6) δ
4.03 (ddd, J ) 8.7, 7.8, 6.4 Hz, 1 H), 3.87 (td, J ) 7.4, 3.7 Hz, 1 H),
3.77-3.61 (m, 2 H), 2.31 (tdd, J ) 13.6, 4.1, 1.0 Hz, 1 H), 2.00-1.90
(m, 2 H), 1.84 (tt, J ) 12.5, 3.6 Hz, 1 H), 1.78-1.43 (series of m, 7
H), 1.28-1.17 (m, 2 H), 1.00 (t, J ) 12.9 Hz, 1 H); 13C NMR (75
MHz, C6D6) δ 87.5, 86.7, 68.8, 67.7, 42.1, 38.7, 33.4, 32.7, 32.2, 31.9,
27.7, 26.6, 26.5, 25.1; MS m/z (M+) calcd 252.2089, obsd 252.2107.
Anal. Calcd for C16H28O2: C, 76.14; H, 11.18. Found: C, 76.16; H,
11.17.
(5R*,6R*,12R*)-12-tert-Butyl-1,7-dioxadispiro[4.0.4.4]tetra-
decane (49). Analogous processing of 39 (100 mg, 0.370 mmol) during
36 h afforded 75 mg (80%) of 49 as a colorless oil: IR (neat, cm-1
)
1
1440, 1365, 1060; H NMR (300 MHz, C6D6) δ 4.05 (dd, J ) 14.9,
7.8 Hz, 1 H), 3.89 (td, J ) 7.4, 3.6 Hz, 1 H), 3.83-3.68 (m, 2 H), 2.10
(t, J ) 12.5 Hz, 1 H), 2.03-1.93 (m, 1 H), 1.82-1.56 (m, 8 H), 1.50-
1.14 (series of m, 4 H), 1.07 (tt, J ) 12.6, 3.4 Hz, 1 H), 0.94 (s, 9 H);
13C NMR (75 MHz, C6D6) δ 88.2, 85.7, 68.4, 67.5, 46.4, 37.2, 34.9,
33.2, 32.4, 32.0, 27.9, 26.8, 26.7, 22.6; MS m/z (M+) calcd 252.2089,
obsd 252.2092. Anal. Calcd for C16H28O2: C, 76.14; H, 11.18.
Found: C, 75.82; H, 11.07.
Hydroboration-Oxidation of 37. Reaction of 37 (292 mg, 1.15
mmol) with the borane-dimethyl sulfide complex (0.634 mL of 2.0
M in THF, 1.27 mmol) at 0 °C in the predescribed manner (1 h 45
min) afforded 188 mg (60%) of 39.
Hydroboration-Oxidation of 38. From 54 mg (0.21 mmol) of
38 and 0.234 mmol of the borane-dimethyl sulfide complex (0 f 20
°C overnight), there was isolated 41 mg (72%) of 40.
Hydroboration-Oxidation of 41. From 114 mg (0.453 mmol) of
41 and 0.543 mmol of the borane-dimethyl sulfide complex (0 f 20
°C over 2 h), there was obtained 92 mg (75%) of 43.
(5R*,6S*,12S*)-12-tert-Butyl-1,7-dioxadispiro[4.0.4.4]tetra-
decane (50). Parallel cyclization of 40 (31 mg, 0.12 mmol) delivered
21 mg (73%) of 50 as a colorless oil: IR (neat, cm-1) 1448, 1394,
1
1366, 1080, 1034; H NMR (300 MHz, C6D6) δ 3.81-3.70 (m, 4 H),
2.45-2.34 (m, 2 H), 2.01-1.83 (m, 2 H), 1.74-1.37 (series of m, 8
H), 1.34 (t, J ) 12.6 Hz, 1 H), 1.06-0.83 (m, 2 H), 0.80 (s, 9 H); 13
C
NMR (75 MHz, C6D6) δ 87.9, 87.6, 68.1, 67.9, 45.9, 39.1, 37.2, 32.1,
32.0, 31.0, 27.7, 27.6, 27.4, 25.1; MS m/z (M+) calcd 252.2089, obsd
252.2089. Anal. Calcd for C16H28O2: C, 76.14; H, 11.18. Found:
C, 76.07; H, 11.13.
Hydroboration-Oxidation of 42. Reaction of 42 (303 mg, 1.20
mmol) in the predescribed manner (0 °C for 5 h) led to the isolation of
44 (292 mg, 90%).
(5R*,6R*,8R*)-8-tert-Butyl-6-methoxy-1-oxaspiro[4.5]decane (27).
To a nitrogen-blanketed, magnetically stirred solution of 25 (900 mg,
0.368 mmol) in dry CH2Cl2 (2 mL) were added 4-(dimethylamino)-
pyridine (2 mg), triethylamine (0.154 mL, 1.11 mmol), and p-
toluenesulfonyl chloride (105 mg, 0.552 mmol). After 36 h, the reaction
mixture was diluted with water and 10% HCl (1.5 mL) and extracted
with ether. The combined organic phases were washed with 3% NaOH
solution (2 × 1.5 mL) and brine, dried, and concentrated. Chroma-
tography of the residue on silica gel (elution with 6:1 hexanes/ethyl
Competition Experiments. A mixture of 10 (105 mg, 0.50 mmol),
11 (105 mg, 0.50 mmol), allyl bromide (91 mg, 0.75 mmol), indium
powder (63 mg, 0.55 mmol), and water (5.5 mL) was stirred at 25 °C
in a stoppered flask for 48 h. Following the addition of 10% HCl, the
reaction mixture was extracted with ether, and the combined organic
phases were washed with brine, dried, and evaporated. The resulting
oil was subjected to chromatography on silica gel (gradient elution with
16:1 f 4:1 hexanes/ethyl acetate). There were isolated 31 mg of 10
and 77 mg of 11, representing essentially quantitative recovery (viz.
51%) of the unreacted ketones.
In entry 49, the two starting ketones proved difficult to separate
cleanly. However, the pairs of allylated epimers formed from each
ketone possess distinctively different Rf values and analysis was
therefore accomplished in this manner. For example, when 0.50 mmol
quantities of 6 and 7 were treated with allyl bromide and indium in
water as described above and the crude product mixture was subjected
to silica gel chromatography, there were isolated 62 mg of 23/24 and
40 mg of 12/13. The combined total represents an essential quantitative
recovery of possible product alcohols.
acetate) provided 64 mg (77%) of 27 as a colorless oil: IR (neat, cm-1
)
1456, 1366, 1138, 1100, 1054; 1H NMR (300 MHz, C6D6) δ 3.98 (dd,
J ) 14.6,7.0 Hz, 1 H), 3.76 (ddd, J ) 7.4, 7.4, 5.5 Hz, 1 H), 3.18 (s,
3 H), 2.78 (dd, J ) 11.4, 4.2 Hz, 1 H), 2.11 (ddd, J ) 11.7, 8.6, 6.8
Hz, 1 H), 2.00-1.90 (m, 1 H), 1.89-1.82 (m, 1 H), 1.72-1.51 (m, 4
H), 1.48-1.41 (m, 1 H), 1.32 (ddd, J ) 11.7, 8.7, 5.7 Hz, 1 H), 1.17-
1.07 (m, 1 H), 0.95 (ddd, J ) 12.2, 12.2, 3.4 Hz, 1 H), 0.87 (s, 9 H);
13C NMR (75 MHz, C6D6) δ 85.8, 82.9, 68.9, 56.6, 47.1, 37.8, 34.6,
32.6, 27.6, 26.8, 23.5; MS m/z (M+) calcd 226.1933, obsd 226.1952.
Anal. Calcd for C14H26O2: C, 74.28; H, 11.58. Found: C, 74.12; H,
11.54.
(5R*,6S*,12R*)-12-tert-Butyl-1,7-dioxadispiro[4.0.4.4]tetra-
decane (45) and (5R*,6R*,12S*)-12-tert-Butyl-1,7-dioxadispiro-
[4.0.4.4]tetradecane (46). Comparable cyclization of a 7.6:1 mixture
of 43 and 44 (53 mg, 0.20 mmol) during 24 h followed by gradient
elution chromatography on silica gel (16:1 f 5:1 hexanes/ethyl acetate)
gave 34 mg (68%) of 45 and 6 mg (12%) of 46.
45: colorless oil; IR (neat, cm-1) 1469, 1392, 1365, 1058; 1H NMR
(300 MHz, C6D6) δ 3.73-3.60 (m, 4 H), 1.91-1.77 (m, 3 H), 1.71-
1.36 (series of m, 12 H), 0.92 (s, 9 H); 13C NMR (75 MHz, C6D6) δ
86.4, 85.1, 67.6 (2 C), 43.0, 35.8, 35.0, 33.7, 33.1, 32.1, 27.8, 26.5,
26.4, 23.0; MS m/z (M+) calcd 252.2089, obsd 252.2095. Anal. Calcd
for C16H28O2: C, 76.14; H, 11.18. Found: C, 76.26; H, 11.22.
Acknowledgment. Funds in support of this research were
generously provided by the Emissions Reduction Research
Center at the New Jersey Institute of Technology.
JA9536835