Zirconium-catalyzed kinetic resolution of cyclic allylic ethers. An enantioselective route to unsaturated medium ring systems

Full text of DOI:10.1021/ja960263m

J. Am. Chem. Soc. 1996, 118, 3779-3780
3779
Table 1. Zr-Catalyzed Kinetic Resolution of Five- and
Six-Membered Cyclic Allylic Ethersa
Zirconium-Catalyzed Kinetic Resolution of Cyclic
Allylic Ethers. An Enantioselective Route to
Unsaturated Medium Ring Systems
Michael S. Visser, Joseph P. A. Harrity, and
Amir H. Hoveyda*
Department of Chemistry, Merkert Chemistry Center
Boston College, Chestnut Hill, Massachusetts 02167
ReceiVed January 25, 1996
Allylic alcohols and ethers represent an important and
versatile class of substrates that are often used in the preparation
of a wide range of medicinally noteworthy natural products.
Therefore, methods that allow the preparation of these com-
pounds with high enantioselectivity are valuable to chemical
synthesis. Within this context, the well-known Ti-catalyzed
asymmetric epoxidation procedure of Sharpless1 is often used
to synthesize optically pure acyclic allylic alcohols through the
catalytic kinetic resolution of easily accessible racemic mix-
tures.2 However, when the catalytic epoxidation is applied to
cyclic allylic substrates, reaction rates are retarded and notably
lower levels of enantioselectivity are observed. Ru-catalyzed
asymmetric hydrogenation has been employed by Noyori to
effect resolution of five- and six-membered allylic carbinols;3
in this instance, as with the Ti-catalyzed procedure, the presence
of an unprotected hydroxyl function is required. Herein, we
report the results of our initial studies on the kinetic resolution4
of a variety of cyclic allylic ethers effected by asymmetric Zr-
catalyzed carbomagnesation.5 Importantly, in addition to six-
membered ethers, seven- and eight-membered ring systems can
be readily resolved by this catalytic protocol.
a Reaction conditions: 10 mol % of (R)-(EBTHI)Zr-binol, 5.0 equiv
of EtMgCl, THF, 70 °C. Reaction in entry 8 was carried out with 20
mol % of (R)-(EBTHI)Zr-binol. b Conversions determined by GLC
analysis in comparison with an internal standard, by analysis of the 1H
NMR of the reaction mixture, and through isolation (silica gel
chromatography). c The identity of the recovered starting materials was
determined through comparison with authentic enantiomers (see sup-
porting information for details). Enantiomeric excess was determined
by chiral GLC (BETADEX 120 chiral column by Supelco, entries 1-3,
6, and 7; CHIRALDEX-GTA chiral column by Alltech, entries 8 and
9). Recovered starting materials in entries 1, 3, 6, 7, and 8 were first
converted to their derived epoxides and then analyzed. Enantiomeric
excess was determined by chiral HPLC: CHIRALCEL OB-H for entries
4 and 5 and with CHIRALPAK AD for entry 10. Absolute stereo-
chemistry was directly determined for 3, 4, 5, and 8; the remaining
assignments were made by inference.
As illustrated in Table 1, when the benzyl ether of 2-cyclo-
penten-1-ol (1) is treated with 10 mol % of (R)-(EBTHI)Zr-
binol6 in the presence of 5 equiv of EtMgCl (THF, 70 °C), the
recovered starting material is obtained in 52% ee (65%
conversion;7 chiral GLC). When allylic MEM ether 2, derived
(1) (a) Martin, V. S.; Woodard, S. S.; Katsuki, T.; Yamada, Y.; Ikeda,
M.; Sharpless, K. B. J. Am. Chem. Soc. 1981, 103, 6237-6240. (b) Gao,
Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune, H.; Sharpless,
K. B. J. Am. Chem. Soc. 1987, 109, 5765-5780.
(2) For recent reviews of kinetic resolution, see: (a) Kagan, H. B.; Fiaud,
J. C. Top. Stereochem. 1988, 18, 249-330. (b) Finn, M. G.; Sharpless, K.
B. In Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New
York, 1985; 247-308.
from cyclohexenol, is subjected to these conditions, at 63%
conversion the starting material is recovered in 50% ee (chiral
GLC). In contrast, the derived n-butyl ether 3 is resolved more
efficiently (79% ee at 63% conversion). The synthetically more
useful benzyl ether 4 undergoes asymmetric ethylmagnesation
and is recovered with still better enantioselection than was
observed with the above substrates ((S)-4 is recovered in 81%
ee; chiral HPLC);8 the latter substrate reacts faster than 2 or 3
as well (compare entries 2 and 3 with 4). Importantly, as shown
in entry 5, when racemic phenyl ether 5 is treated to the Zr-
catalyzed resolution conditions for 2 h, (S)-5 is obtained in 97%
ee at 60% conversion.9
The resolution data depicted in entries 8-10 in Table 1
indicate the following. (i) Disubstituted allylic ethers can be
resolved through the Zr-catalyzed protocol; phenyl ether 8 is
obtained in 98% ee after 55% conversion. To obtain excellent
levels of enantiomeric purity in the resolution with 8 as the
substrate, 20 mol % of zirconocene (Vs 10 mol %) must be used;
15 mol % of (EBTHI)Zr-binol affords the recovered starting
material in 84% ee at 60% conversion (kfast/kslow ) 8);10 with
10 mol % of precatalyst the reaction does not effectively proceed
(3) Kitamura, M.; Kasahara, I.; Manabe, K.; Noyori, R.; Takaya, H. J.
Org. Chem. 1988, 53, 708-710. For Pd-catalyzed enantioselective synthesis
of cyclic allylic esters, see: Trost, B. M.; Organ, M. G. J. Am. Chem. Soc.
1994, 116, 10320-10321.
(4) For previous reports on kinetic resolution processes through Zr-
catalyzed carbomagnesation, see: (a) Morken, J. P.; Didiuk, M. T.; Visser,
M. S.; Hoveyda, A. H. J. Am. Chem. Soc. 1994, 116, 3123-3124. (b) Visser,
M. S.; Hoveyda, A. H. Tetrahedron 1995, 51, 4383-4394. (c) Visser, M.
S.; Heron, N. M.; Didiuk, M. T.; Sagal, J. F.; Hoveyda, A. H. J. Am. Chem.
Soc. in press.
(5) (a) Morken, J. P.; Didiuk, M. T.; Hoveyda, A. H. J. Am. Chem. Soc.
1993, 115, 6997-6998. (b) Didiuk, M. T.; Johannes, C. W.; Morken, J. P.;
Hoveyda, A. H. J. Am. Chem. Soc. 1995, 117, 7097-7104. For application
of asymmetric Zr-catalyzed carbomagnesation to natural product synthesis,
see: Houri, A. F.; Xu, Z-M.; Cogan, D. A.; Hoveyda, A. H. J. Am. Chem.
Soc. 1995, 117, 2943-2944.
(6) (a) Wild, F. R. W. P.; Waiucionek, M.; Huttner, G.; Brintzinger, H.
H. J. Organomet. Chem. 1985, 288, 63-67. (b) Diamond, G. M.; Rodewald,
S.; Jordan, R. F. Organometallics 1995, 14, 5-7 and references cited therein.
(7) All reactions described herein afford the expected ethylmagnesation
products. In certain cases, due to substrate volatility, the alkylation product
could not be isolated but was detected in the 1H NMR spectrum of the
unpurified mixture. In instances where volatility is not a hindrance, the
product was fully characterized (e.g., with 9 and 10, i and ii were obtained,
respectively). Details (including product stereocontrol) will be provided in
the full account of this study.
(8) All the catalytic resolutions described herein must be carried out with
freshly prepared catalyst batches of high purity. Sluggish reactions and/or
inferior enantioselectivities will be otherwise observed.
(9) The phenyl ether derived from cyclopentenol reacts with the
alkylmagnesium halide in the absence of the catalyst, preempting the
possibility of catalytic kinetic resolution (both at 22 and 70 °C). Nonetheless,
the starting material is recovered in ∼40% ee after 60% conversion.
0002-7863/96/1518-3779$12.00/0 © 1996 American Chemical Society

3780 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Communications to the Editor
Table 2. Zr-Catalyzed Kinetic Resolution of Seven- and
Chart 1
Eight-Membered Cyclic Allylic Ethersa
absence of more extensive studies, it is difficult to identify the
basis (steric or electronic) of the observed variations in
selectivity. However, data shown in the table do imply that
subtle electronic factors can be important; entries 2 and 3 suggest
that electronic effects can influence the resolution efficiency
(n-butyl and MEM groups have near identical steric require-
ments, unless the OMEM unit is coordinated to a Mg ion). As
another example, reactions of 5 Vs 6 or 5 Vs 7, or that of 13 Vs
14, illustrate that alteration of the electronic character of the
allylic heteroatom substituent leads to detectable changes in
enantioselection. Studies designed to clarify the origin of the
above differences in enantioselectivity are underway.
Modes of addition shown in Chart 1 are consistent with extant
mechanistic work5b,12 and accurately predict the identity of the
slower reacting enantiomer. It is noteworthy, however, that
variations in the observed leVels of selectivity as a function of
the steric and electronic nature of substituents and the ring size
cannot be predicted on the basis of these models alone; more
subtle factors are clearly at work. In spite of such mechanistic
dilemma, which will be the subject of future investigations, the
metal-catalyzed resolution protocol provides an attractive option
in asymmetric synthesis. This is because, although the maxi-
mum possible yield is ∼40%, it requires easily accessible
racemic starting materials and conversion levels can be ma-
nipulated so that truly pure samples of substrate enantiomers
are obtained.13 Further studies in the area of Zr-catalyzed
asymmetric carbomagnesation and its applications to the total
synthesis of natural products continue in these laboratories and
will be reported shortly.
a,b See Table 1. c The identity of the recovered starting materials was
determined through comparison with authentic enantiomers (see sup-
porting information for details). Enantiomeric excess was determined
by chiral GLC (ALPHADEX 120 chiral column by Supelco). Absolute
stereochemistry was directly determined for 11, 13, and 14; the
remaining assignment was made by inference.
past 40% conversion. (ii) Bicyclic 1,6-disubstituted ether 9 is
resolved more efficiently than the derived monocyclic substrate
10; this outcome is notable, since we find the corresponding
methoxyisopropyl ether to be an inferior substrate for the Zr-
catalyzed resolution (kfast/kslow ) 1.5).
As entries 1 and 2 of Table 2 illustrate, with seven-membered
allylic ethers, significantly higher levels of enantioselection are
achieved (in comparison to six-membered ring derivatives); 11
and 12 are resolved to afford recovered starting materials in
>99% ee (chiral GLC). Data presented in entries 1 and 4 in
Table 1 and entry 1 in Table 2 thus indicate that, with increasing
the size of the carbocyclic ring structure, larger rate differences
in the carbomagnesations of substrate enantiomers are detected
(kfast/kslow ) 2.8, 7.8, and >25 for 1, 4, and 11). A similar
trend emerges when we compare catalytic ethylmagnesations
of phenyl ethers 5 and 12. The exact reason for the dependence
of the resolution efficiency on ring size must await future
mechanistic studies. Although the observed selectivity patterns
can be reliably extended to larger ring systems, the levels of
resolution efficiency may not be. In the kinetic resolution of
the phenyl ether derived from 2-cycloocten-1-ol, kfast/kslow is
∼8 (Vs >25 for 12). Regardless, because of the ease of
preparation of the racemic cycloheptenol and cyclooctenol
derivatives and due to the paucity of available methods for the
enantioselective synthesis of seven- and eight-membered ring
systems, this procedure should prove useful in applications to
asymmetric synthesis of natural products.
Our attempts to effect the Zr-catalyzed resolution of silyl ether
derivatives of cyclohexenol were not successful; the derived
TES and TBS ethers reacted sluggishly and nonselectively (kfast/
kslow does not exceed 1.8). Parent allylic alcohols cannot be
resolved through this procedure; however, allylic ethers depicted
in Tables 1 and 2, which are resolved effectively by this catalytic
procedure, can be deprotected to yield the parent nonracemic
allylic carbinol.11
Results illustrated in Tables 1 and 2 indicate that simple
structural changes can lead to notable differences in resolution
efficiency; the reaction rate differential between phenyl ether
enantiomers of 5 and benzyl ethers 4 is a case in point. In the
Acknowledgment. This research was generously supported by the
NIH (GM-47480). Additional support was provided by the NSF (CHE-
9258287, NYI Program), Johnson and Johnson, and Pfizer. A.H.H. is
an American Cancer Society Junior Faculty Research Awardee (JFRA-
434), an Alfred P. Sloan Research Fellow, and a Camille Dreyfus
Teacher-Scholar.
Supporting Information Available: Experimental procedures,
stereochemical proofs, and spectral and analytical data for all resolved
compounds (17 pages). This material is contained in many libraries
on microfiche, immediately follows this article in the microfilm version
of the journal, and can be ordered from the ACS, and can be
downloaded from the Internet; see any current masthead page for
ordering information and Internet access instructions.
JA960263M
(12) Hoveyda, A. H.; Morken, J. P. J. Org. Chem. 1993, 58, 4237-
4244.
(13) Representative procedure: 3-Phenoxycyclohept-1-ene 12 (282 mg,
1.50 mmol) was dissolved in 7.5 mL of anhydrous THF, containing
hexadecane (0.50 mL) as an internal standard. After addition of EtMgCl
(5.2 mL, 7.50 mmol) and (R)-(EBTHI)Zr-binol (96 mg, 0.15 mmol), the
reaction mixture was placed in an oil bath heated to 70 °C. Stirring continued
at this temperature for 5 h, after which GLC analysis indicated the reaction
had proceeded to 56% conversion. The reaction was quenched with 10.0
mL of H2O; the aqueous layer was subsequently washed 3 times with 5.0
mL of Et2O. Evaporation of the organic solvents in vacuo and purification
by silica gel chromatography afforded 122 mg (0.65 mmol, 98% yield based
on percent conversion) of 12 which was shown to be >99% ee by chiral
GLC analysis (minor enantiomer not detected).
(10) Relative rates were calculated by an equation reported previously.
See ref 2.
(11) For deprotection of p-anisyl ethers, see: Fukuyama, T.; Laird, A.
A.; Hotchkiss, L. M. Tetrahedron Lett. 1985, 26, 6291-6292. A benzyl
ether of an allylic carbinol can be efficiently removed: Hwu, J. R.; Chua,
V.; Schroeder, J. E.; Barrans, R. E.; Khoudary, K. P.; Wang, N.; Wetzel,
J. M. J. Org. Chem. 1986, 51, 4733-4734.