Ligand tuning in asymmetric hydrovinylation of 1,3-dienes: A stereoselective route to either steroid-C20 (S) or -C20 (R) derivatives

Article abstract of DOI:10.1021/ja711475f

1,3-Dienes derived from steroidal D-ring C₁₇-ketones undergo Ni(II)-catalyzed hydrovinylation to give 1,2- or 1,4-addition of ethylene. Using finely tuned phosphoramidite ligands, it is possible to synthesize either the C₂₀ (R)- or (S)-derivatives without mutual contamination. The proportion of the 1,4-adduct, which is also formed stereoselectively, can be minimized by optimizing the reaction conditions. Because the two alkenes in the resultant dienes have differing steric demands for many potential reactions, and are ideally juxtaposed for further D-ring functionalization, these intermediates could be useful for the preparation of biologically important compounds such as vitamin D analogs and various antitumor steroidal glycosides.

Full text of DOI:10.1021/ja711475f

Published on Web 06/21/2008
Ligand Tuning in Asymmetric Hydrovinylation of 1,3-Dienes:
A Stereoselective Route to Either Steroid-C20 (S) or -C20 (R)
Derivatives
Biswajit Saha, Craig R. Smith, and T. V. RajanBabu*
Department of Chemistry, The Ohio State UniVersity, 100 West 18th AVenue,
Columbus, Ohio 43210
Received January 3, 2008; E-mail: rajanbabu.1@osu.edu
Abstract: 1,3-Dienes derived from steroidal D-ring C17-ketones undergo Ni(II)-catalyzed hydrovinylation
to give 1,2- or 1,4-addition of ethylene. Using finely tuned phosphoramidite ligands, it is possible to synthesize
either the C20 (R)- or (S)-derivatives without mutual contamination. The proportion of the 1,4-adduct, which
is also formed stereoselectively, can be minimized by optimizing the reaction conditions. Because the two
alkenes in the resultant dienes have differing steric demands for many potential reactions, and are ideally
juxtaposed for further D-ring functionalization, these intermediates could be useful for the preparation of
biologically important compounds such as vitamin D analogs and various antitumor steroidal glycosides.
Introduction
A chiral side chain carrying a methyl group is a very common
structural motif in many terpenoids, and this side chain is often
attached at a stereogenic center of a ring. Examples can be found
in simple sesquiterpenes such as juvenile hormone juvabione,1
or in more complex structures such as pseudopterosins,2
elisabethin A,3 ophiobolin C,4 steroid hormone calcitriol (1-R-
2,5-dihydroxyvitamin D3) and its analogs,5 antitumor agents
cephalostatins,6 and various cytotoxic steroidal glycosides7
(Figure 1). Several creative solutions to the problem of
installation of these stereogenic centers have been developed
over the years, even though no broadly applicable method that
uses readily available precursors has emerged.8 The problem is
especially acute for the synthesis of the unnatural 20(S)-epimers.
Consider, for example, precursors (e.g., 1) for calcitriol analogs
with exocyclic C20 (S)-configuration, which have been shown
to have significant biological activity.9 These molecules are
currently prepared by circuitous routes that involve the equili-
bration of the aldehyde 2, obtained from vitamin D2 and
subsequent reactions of the minor isomer isolated from the
mixture.10
Figure 1. Important classes of natural products with exocyclic chirality
carrying a methyl-bearing carbon.
We recently showed11 that the asymmetric hydrovinylation
of readily available 1-vinylcycloalkene offers a potential general
(1) For an enantioselective synthesis and leading references, see: Wa-
tanabe, H.; Shimizu, H.; Mori, K. Synthesis 1994, 1249.
(2) Look, S. A.; Fenical, W.; Jacobs, R. S.; Clardy, J. Proc. Natl. Acad.
Sci. U.S.A. 1986, 83, 6238.
(8) For a leading reference, discussion, and citations of earlier work, see:
(a) Guevel, A.-C.; Hart, D. J. J. Org. Chem. 1996, 61, 465, and
references cited therein. For a pedagogical view of this problem and
a historical account of syntheses of erythro- and threo-juvabiones,
see: (b) Carey, F. A.; Sundberg, R. J. AdVanced Organic Chemistry,
2nd ed.; Kulwer: New York, 2001; Vol. 2, pp 848-859. (c) Trost,
B. M.; Verhoeven, T. R. J. Am. Chem. Soc. 1976, 98, 630. (d) Snider,
B. B. Acc. Chem. Soc. 1980, 13, 426. (e) Snider, B. B.; Deutsch, E. A.
J. Org. Chem. 1983, 48, 1823. (f) Wulkovich, P. M.; Barcelos, A.;
Sereno, J. F.; Baggiolini, E. G.; Hennesey, B. M.; Uskokovic, M. R.
Tetrahedron 1984, 40, 2283. (g) Andersen, N. H.; Hadley, S. W.;
Kelly, J. D.; Bacon, E. R. J. Org. Chem. 1985, 50, 4144. (h) Mikami,
K.; Kawamoto, K.; Nakai, T. Tetrahedron Lett. 1985, 26, 5799. (i)
Mikami, K.; Loh, T.-P.; Nakai, T. J. Chem. Soc., Chem. Commun.
1988, 1430. (j) Houston, T. A.; Tonaka, Y.; Koreeda, M. J. Org. Chem.
1993, 58, 4287. For recent references, dealing with side-chain of the
anticancer natural product OSW-1, see (k) Yu, W.; Jin, Z. J. Am. Chem.
Soc. 2002, 124, 6576.
(3) For a review, see: Zanoni, G.; Franzini, M. Angew. Chem., Int. Ed.
2004, 43, 4837.
(4) (a) Rowley, M.; Tsukamoto, M.; Kishi, Y. J. Am. Chem. Soc. 1989,
111, 2735. (b) Boeckman, R. K., Jr.; Arvanitis, A.; Voss, M. E. J. Am.
Chem. Soc. 1989, 111, 2737.
(5) For a leading reference, see: (a) Hijikuro, I.; Doi, T.; Takahashi, T.
J. Am. Chem. Soc. 2001, 123, 3716. For a review of vitamin D
chemistry, see: (b) Dai, H.; Posner, G. H. Synthesis, 1994, 1383. (c)
Posner, G. H.; Lee, J. K.; White, M. C.; Hutchings, R. H.; Dai, H.;
Kachinski, J. L.; Dolan, P.; Kensler, T. W. J. Org. Chem. 1997, 62,
3299.
(6) Pettit, G. R.; Tan, R.; Xu, J.-P.; Ichihara, Y.; Williams, M. D.; Boyd,
M. R. J. Nat. Prod. 1998, 61, 955, and references cited therein.
(7) Mimaki, Y.; Yokosuka, A.; Sashida, Y. J. Nat. Prod. 2000, 63, 1519.
9
9000 J. AM. CHEM. SOC. 2008, 130, 9000–9005
10.1021/ja711475f CCC: $40.75
2008 American Chemical Society

Stereoselective Route to Steroid-C20 (S) or -C20 (R)
A R T I C L E S
Scheme 1. Exocyclic Stereocenters via Asymmetric
Hydrovinylation
solution to this exocylic stereochemistry problem (Scheme 1).
Both achiral and chiral ligands (Figure 2) carrying suitable
hemilabile groups such as o-benzyloxyphenyldiphenylphosphine
(L3), phospholanes (e.g., L4), and phosphoramidites (e.g., L5)
gave nearly quantitative yields of the hydrovinylation products,
which could be subjected to further functionalization via
diastereoselective reactions.
While attempting to apply the diene hydrovinylation for the
functionalization of a steroid D-ring (eq 1), it was observed,
that several excellent ligands we had initially employed either
did not react (Figure 2: Ph3P, L1, L2) or gave mixtures (L3,
L4) of stereo- and regioisomers. We anticipate this lack of
selectivity to be a recurring problem in the context of this and
other future synthetic objectives in which hydrovinylations of
key chiral intermediates will be involved. It is entirely conceiv-
able that the inherent diastereoselectivity in such intermediates
could be low or even opposite to what would be desired. Thus,
from a synthetic perspective, either enhancing the inherent
selectivity or overriding such an outcome with the use of a
tunable asymmetric catalyst will be a highly desirable goal.
Looking for a general solution to this problem, we decided to
examine the effect of ligands on the selectivity of the hydro-
vinylation reactions of 1,3-dienes 3 and 4, derived from two
prototypical steroids, estrone and 3-epiandrosterone. In this
paper, we report the results of these studies, which demonstrate
that in these steroids it is possible to install, with complete
stereoselectiVity, either C20 (R) or C20 (S) configuration by proper
choice of ligands and reaction conditions. A limited study of
the Ru-catalyzed hydrovinylation of 3 published earlier12 did
not address the key issue of control of stereoselectivity at this
stereogenic center.
conditions14 using either [(allyl)NiBr]2/Ph3P/AgOTf or [(al-
lyl)NiBr]2/(L)/(NaBARF) [L ) L1, L2; BARF ) tetrakis-(3,5-
bis-trifluoromethylphenyl)borate] at temperatures between -55
and 25 °C under 1 atm of ethylene gave no products. This lack
of reactivity is quite surprising in this otherwise broadly
applicable hydrovinylation protocol. Upon further examination
of the reaction using other ligands L3-L12, most notably the
phosphoramidites,15 under a variety of conditions, it was found
that synthetically useful levels of selectivity could be achieved
(eqs 1–3). The results are listed in Table 1.
Hydrovinylation using an achiral ligand L3 (o-benzyloxy-
phenyldiphenylphosphine) gives a mixture of C20 (S) [5] and
C20 (R) [7] epimers in a ratio of 1.0:2.5 (entry 1) along with
1,4-adducts 6 and 8 (see the following paragraphs and Sup-
porting Information for details of structural assignments).16 This
inherent selectivity for the formation of the C20 (R) adduct can
be reversed, albeit modestly, with the use of a chiral ligand,
the phospholane L4, which yields a ratio of 3:1 for 5:7 (entry
2). In addition to the formation of the byproducts, 6 and 8 (the
1,4-adducts), these reactions are also complicated by minor
isomerization of the primary products giving what appears to
be conjugated dienes. The first indication that this unwanted
isomerization can be completely blocked and exclusive selectiv-
ity for the 20(S) compound can be achieved came from ligands
L5 (RaScSc) and L10 (ScSc), which gave a clean mixture of 5
and 6, with no trace of the 20(R)-epimer 7, the 1,4-adduct 8, or
isomerization products (eq 2, entries 3 and 5-8). Typically,
the reaction is done as follows:15h a solution of 0.0025 mmol
of [(allyl)2NiBr]2 in 0.5 mL of solvent (usually CH2Cl2) and a
solution of the ligand in 0.5 mL solvent are mixed in a drybox.
This solution is added to a suspension of NaBARF (1 equiv
Results and Discussion
Our initial studies were conducted with the diene 3, readily
prepared from estrone as described previously.13 Nickel-
catalyzed hydrovinylation of 3 under our initially reported
(9) (a) Honzawa, S.; Suhara, Y.; Nihei, K.; Saito, N.; Kishimoto, S.;
Fujishima, T.; Kurihara, M.; Sugiura, T.; Waku, K.; Takayamaa, H.;
Kittaka, A. Bioorg. Med. Chem. Lett. 2003, 13, 3503, and references
cited therein. One modified D-ring analog with C20 (S)- configuration
(e.g., Ro 26-9228) was reported to increase bone mineral density
while showing reduced calciuria and calcemia compared to a C20 (R)
analog. Related compounds were found to stimulate HL-60 cell
differentiation, and hence could find application for treatment of
leukemia. For leading references, see: (b) Kabat, M. M.; Garofalo,
L. M.; Daniewski, A. J.; Hutchings, S. D.; Liu, W.; Okabe, M.;
Radinov, R.; Zhou, Y. J. Org. Chem. 2001, 66, 6141.
(14) (a) Nomura, N.; Jin, J.; Park, H.; RajanBabu, T. V. J. Am. Chem. Soc.
1998, 120, 459. (b) RajanBabu, T. V.; Nomura, N.; Jin, J.; Nandi,
M.; Park, H.; Sun, X. J. Org. Chem. 2003, 68, 8431.
(15) (a) Feringa, B. L. Acc. Chem. Res. 2000, 33, 346. (b) Arnold, L. A.;
Imbos, R.; Mandoli, A.; de Vries, A. H. M.; Naasz, R.; Feringa, B. L.
Tetrahedron 2000, 56, 2865. (c) Alexakis, A.; Rosset, S.; Allamand,
J.; March, S.; Guillen, F.; Benhaim, C. Synlett 2001, 1375. Hydro-
vinylations using phosphoramidite ligands. (d) Francio´, G.; Faraone,
F.; Leitner, W. J. Am. Chem. Soc. 2002, 124, 736. (e) Park, H.;
Kumareswaran, R.; RajanBabu, T. V. Tetrahedron 2005, 61, 6352.
(f) Smith, C. R.; RajanBabu, T. V. Org. Lett. 2008, 10, 1657. (g) For
a scalable procedure for the synthesis of phosphoramidites, see: Smith,
C. R.; Mans, D.; RajanBabu, T. V. Org. Synth. 2008, in press. (h)
Smith, C. R.; Zhang, A.; Mans, D.; RajanBabu, T. V. Org. Synth.
2008, in press.
(10) (a) Fujishima, T.; Konno, K.; Nakagawa, K.; Kurobe, M.; Okano, T.;
Takayama, H. Bioorg. Med. Chem. 2000, 8, 123. (b) Ferna´ndez, B.;
Pe´rez, J. A. M.; Granja, J. R.; Castedo, l.; Mourino, A. J. Org. Chem.
1992, 57, 3173. See also: (c) Ferna´ndez, C.; Go´mez, G.; Lago, C.;
Moma´n, E.; Fall, Y. Synlett 2005, 2163. (d) See also ref 5.
(11) (a) Zhang, A.; RajanBabu, T. V. J. Am. Chem. Soc. 2006, 128, 54.
For a recent review of hydrovinylation, see: (b) RajanBabu, T. V.
Chem. ReV. 2003, 103, 2845.
(12) He, Z.; Yi, C. S.; Donaldson, W. A. Org. Lett. 2003, 5, 1567.
(13) De Riccardis, F.; Meo, D.; Izzo, I.; Di Filoppo, M.; Casapullo, A.
Eur. J. Org. Chem. 1998, 1965.
(16) See Supporting Information for 1H and 13C NMR spectra.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 9001

A R T I C L E S
Saha et al.
Figure 2. Ligands for hydrovinylation of steroidal 1,3-dienes.
Table 1. Ni-Catalyzed Hydrovinylation of the Steroidal Diene 3a
Table 2. Ni-Catalyzed Hydrovinylation of the Steroidal Diene 4a
no.
ligand/conditions
yield (%)
5(20S):6:7(20R):8b
no.
ligand/conditions
yield (%)
9 (20S):10:11(20R):12b
1.
2.
3.
4.
5.
6.
L3 22 °C, 14 h
L4 22 °C, 14 h
76
76
64
86
78
18:16:45:5c
45:18:15:14c
30:70:0:0
1.
2.
3.
4.
5.
6.
L3 0-22 °C, 14 h
73
74
86
83
87
82
28:15:42:<5c
75:25:0:0
12:80:0:0c
80:20:0:0
68:30:0:0
20:80:0:0
L5 (RaScSc) 22 °C, 14 h
L6 (RaSc-i-Pr) 22 °C, 14 h
L10 (ScSc) 0-10 °C, 2 h
L10 (ScSc) 0-22 °C, 4 h
L10 (ScSc) 0 °C, 10 min,
then warmed to 20 °C, 14 h
L11 (SaScSc) 22 °C, 14 h
L12 (SaRcRc) 22 °C, 14 h
L5 (RaScSc) 22 °C, 14 h
L6 (RaSc-i-Pr) 0-22 °C, 14 h
L10 (ScSc) -10 °C, 14 h
L10 (ScSc)
(a) CH2Cl2, 0-22 °C, 4 h
(b) tol.,d 0-22 °C, 14 h
L10 (ScSc) 0-10 °C, 1 h
L10 (ScSc), 0 °C, 10 min,
then warmed to 20 °C, 14 h
L11 (SaScSc) 22 °C, 14 h
L12 (SaRcRc) 22 °C, 14 h
∼7:80:0:0c
83:15:0:0
85
66
71:38:0:0
88:12:0:0
90:10:0:0
22:72:0:0
7.
8.
81
69
0:<5:66:10c
0:0:65:35
7.
8.
84e
82
a See eqs 1, 4, and 5. b Ratio as %; no isomerization unless indicated.
c Contains other olefins (δ 5.60-5.75).
9.
10.
83
84
trace:trace:75:8c
trace:0:70:30
preparatively useful runs (e.g., entry 5 in Table 1 and entry 4
in Table 2), the reaction mixture is maintained at the temper-
atures and times indicated. Under these conditions the highest
proportion of the 1,2-adduct is obtained. For example, in 14 h
at -10 °C, 83% of the 20(S)-compound was formed from 3
using ligand L10 (entry 5, Table 1). Attempts to shorten the
reaction time by running the reaction at 0-10 °C for 1 h gave
a slightly higher proportion (90%) of the desired product, but
the conversion remained low (∼40%, entry 7, Table 1). One
useful variable was the solvent. Reactions done in toluene (entry
6, Table 1) gave 88% of the desired product in 14 h. Ethyl
acetate and hexaflurobenzene gave no products. Mixing the
reagents at low temperature (0 °C) and immediately warming
the solution to room temperature in 10 min resulted in the
a See eqs 1–3. See text for details of the experiments. For entries
1-7, 9, and 10, the reaction was started at the indicated temperature and
was warmed to the final temperature with the cold bath in place. b Ratio
as %; no isomerization products unless indicated. c Rest isomerization
products (δ 5.60-5.85). d No reaction in EtOAc, C6F6. e Yield based on
recovered starting material (40% conversion).
with respect to Ni). After stirring for 1.5 h, the mixture is filtered
through a small plug of Celite into a Schlenk flask, which is
taken out of the drybox. The catalyst is brought to the
appropriate temperature, and under 1 atm of ethylene the
substrate is added in 1.2 mL of solution. The mixture is stirred
for the times indicated in Tables 1 and 2 and then quenched
with saturated ammonium chloride before work up. In the
9
9002 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Stereoselective Route to Steroid-C20 (S) or -C20 (R)
A R T I C L E S
spectrum (500 MHz, CDCl3) shows the most discernible peaks
at δ 5.843 (ddd, J ) 18, 10, and 8 Hz, 1 H, vinyl,
C16-CHdCH2), 5.309 (dq, J ) 7, 2 Hz, 1 H, vinyl, C20-H),
5.062 (s, 2 H, OCH2Ph), 5.044 (d, J ) 18 Hz, 1 H, vinyl,
C16-CHdCHt Hc), 4.994 (d br, J ) 10 Hz, 1 H, vinyl,
C16-CHdCHt Hc), 3.423 (dd, J ) 8, 8 Hz, 1 H, C16-H),
2.80-3.00 (m, 3 H, benzyl H’s), 1.607 (dd, 7 Hz, 1 Hz, 3 H,
C20-CH3), 0.860 (s, 3 H, C18). The doublet of quartert at δ
5.309 (C20-H), the dd at δ 3.423 for the C16-H, and the doublet
of doublet at δ 1.607 for the C20-CH3 are especially valuable
in identifying this product. The NOESY spectrum indicates
proximity of the bisallylic hydrogen (C16-H) on the D-ring to
the C20-Me suggesting an E-configuration for the alkene. This
is further supported by NOE contact between the vinylic C20-H
and C-ring hydrogens rather than the D-ring hydrogens. The
stereochemistry (R-vinyl appendage at C16) is deduced from the
fact that 6 is the only other product formed with the 20(S)
compound 5 and, thus, must originate from the same allyl-Ni
intermediate arising from the R-face addition of the cationic
Ni-H to the starting diene (Scheme 2).18a–c Further support
for this rationale comes from the corresponding exclusiVe
formation of the C16-ꢀ-vinyl side product (8) concomitant with
the formation of the C20 (R)-epimer (7) [vide infra]. The most
stable conformation for the 1,3-diene is assumed in the
construction of the models shown in Scheme 2. It is premature
to propose models for these reactions, especially considering
the two possible orientations of the square planar complex [L∼X
Ni-(olefin)H]+ are possible (L ) phosphorus; X ) hemilabile
group: a dioxalane O in L4 or the Ph group attached to the
N-benzyl substituent17). Clearly each of the two ligands L10
or L12 that give respectively the 20(S) and the 20(R) product,
must form a complex which matches the chirality of the starting
material, only when the appropriate face of the diene is
coordinated, resulting in the high selectivity observed. Support
for the aryl group acting as a hemilabile ligand comes from
Leitner’s DFT-based computational studies on the Ni-phos-
phoramidite-catalyzed hydrovinylation.17
formation of mostly the 1,4-adduct (entry 8). The ratio of the
1,2- vs 1,4-adducts is also dependent on the structure of the
ligands. For example, the ligand L6 with an N-i-Pr group (or
the corresponding N-benzyl derivative L7) instead of the N-R-
methylbenzyl substituent of L5 gave a good yield of the 1,4-
adduct 6 (entry 4). Surprisingly, a ligand without R-methyl-
benzyl substituents (L9) gave no product. Ligands derived from
(S)-binaphthol (L11 and L12) gave 20(R)-adduct in moderate
yields (eq 3, Table 1, entries 9 and 10), with only traces (2-3%)
of the 20(S)-epimer.16 These experimental observations are based
on multiple trials, and this lack of reactivity of L9 complexes
is highly reproducible in the hydrovinylation of dienes and as
well as in vinyarenes.15g One possible explanation for the lack
of reactivity of L9 (vis-a`-vis L10 or L12) might be the absence
of the Me substitution at the benzylic position, which is essential
to limit the degrees of freedom for the N-substituent facilitating
hemilabile coordination of the Ph-unit as conjectured by
Leitner.17 This would be somewhat akin to the “Thorpe-Ingold”
effect, which operates in certain cyclization reactions.
The 20(R) epimer 7 is similarly prepared using the ligand
L12 (Table 1, entry 10). The corresponding 1,4-addition product
8 was formed in 25% as the sole side product. Adduct 7 has 1H
and 13C spectra closely resembling 5 except for characteristic
differences listed below: the C20-methyl in 7 appears as a doublet
δ 1.199 (d, J ) 7 Hz), 0.029 ppm upfield compared to that for
the 20(S)-compound 5 (δ 1.228, J ) 7 Hz). Similar differences
in chemical shifts of 20(S)- and 20(R)-methyl compounds have
been reported in the literature for structurally related compounds
with C16-C17 unsaturation.19 We also note that the 13C NMR
signal for C20-CH3 for the R-derivative also appear at higher
field (20S: 20.74; 20R: 20.28). The chemical shift of C16–H in
the two compounds also distinguishes the two isomers (5: 5.446;
7: 5.432). The minor product 8 shown in eq 2 is very similar to
6, yet distinctly different, showing the following diagnostic
peaks in the NMR. The C20-vinyl hydrogen in 8 appears at δ
5.259 as a dq compared to the corresponding peak at δ 5.309
(dq, J ) 7, 2 Hz, 1 H, vinyl, C20-H) in 6. The other major
differences are in the chemical shifts of the allylic hydrogen
C16-H (8: 3.320 ddd, 8 Hz, 8 Hz, 8 Hz; 6, 3.423, dd, 8 Hz, 8
Hz) and C18-H3’s (8:0.844; 6: 0.860). The almost identical
chemical shift of the vinyl C20-Me signals (6: 1.607; 8: 1.604)
and the significant differences in chemical shifts (δ 6: 3.423
and 8: 3.320) and coupling patterns of C16-H’s in the two
compounds (6: dd; 8: ddd) also provide indirect support for these
two structures in which the only difference is the configuration
Preparatively, the most useful reactions involve the use of
ligands L10 and L12, which give the 20(R) or the 20(S)
compound, respectively, along with minor amounts of a 1,4-
adduct (eqs 2 and 3). The highly stereoselective formation of
the otherwise scarce C20 (S)-isomer uncontaminated with the
corresponding (R)-epimer is particularly noteworthy. Thus, the
diene 3 reacts with ethylene (1 atm) in the presence of a
catalyst14,15g,h prepared from [(allyl)NiBr]2/L10/NaBARF (2
mol%) at -10 °C giving 78% yield of hydrovinylation products
5 and 6 (Table 1, entry 5). The hydrovinylation product 5 is
identified as the 20(S)-derivative by comparison of 1H and 13
C
NMR spectra with those of a compound described in the
literature.12 The connectivity of atoms in the minor 1,4-adduct
6 is ascertained from the 1H NMR features. The 1H NMR
(17) Ho¨lscher, M.; Francio´, G.; Leitner, W. Organometallics 2004, 23,
5606.
(18) Discussion of a model for asymmetric induction in nickel-catalyzed
asymmetric hydrovinylation, see: (a) Zhang, A.; RajanBabu, T. V. Org.
Lett. 2004, 6, 1515. (b) Saha, B.; RajanBabu, T. V. J. Org. Chem.
2007, 72, 2357. (c) For the role of hemilabile ligands and counter
ions in asymmetric hydrovinylation, see: Nandi, M.; Jin, J.; RajanBabu,
T. V. J. Am. Chem. Soc. 1999, 121, 9899.
(19) Schmuff, N. R.; Trost, B. M. J. Org. Chem. 1983, 48, 1404.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 9003

A R T I C L E S
Saha et al.
Scheme 2. Origin of 1,2- and 1,4-Adducts in the Hydovinylation Reactions
of C16. The upfield shift for C18H3 in the 13C NMR of 8 with a
C16-ꢀ-substituent compared to the C18H3 in 6 (C16-R-substituent)
is also consistent with the γ-effect observed in cis- dialkyl
substituted cycloalkanes.20
Scheme 3. D-Ring Functionalization via Hydrovinylation Adducts
two epimeric compounds (9: 5.376; 11: 5.360). The NMR
features that distinguish the minor products 10 and 12 mirror
what is seen for the byproducts, 6 and 8, from the estrone-
derived system. Thus the C20-Me groups appear at δ 1.554
and 1.557 (almost identical) whereas the C16-H’s appear at δ
3.335 (dd) and 3.228 (ddd) [significantly different]. The
mechanistic rationale outlined in Scheme 2 for the formation
of the two sets of compounds applies here as well.
The steroid D-ring can be elaborated in myriad ways using
the diene functionality in 5 and 9. For example, selective
hydroboration of the monosubstituted olefin gives an alcohol
(13 or 14), which could serve as a precursor for more advanced
intermediates. Catalytic hydrogenation of each of these alcohols
gives a single product, (15 or 16), whose structure is inferred
as arising from R-face addition of hydrogen from well-
precedented steroid examples.19,21 The endocyclic π-bond
(C16-C17) will also be a useful handle for oxygenation of the
D-ring, a key feature in many important steroidal glycosides,7
including potent anticancer agent OSW-1.8k
Functional group compatibility of the Ni(II)-catalyzed hydro-
vinylation allows the functionalization of diene 4, derived from
3-acetylepiandrosterone (eqs 4 and 5). The observed ligand
effects parallel the results obtained with the diene 3. These are
listed in Table 2. The C20 (S) compound 9 and the corresponding
C20 (R) compound 11 are produced without any mutual
contamination using ligands L10 and L12 respectively (entries
4 and 8). As with the estrone derivatives, the C20 (S) compounds
are characterized by the downfield chemical shifts of the C20-
CH3 signals in both the 1H (9: 1.177; 11: 1.096) and 13C NMR
(9: 20.66; 11: 20.22).16 Among other distinguishing features is
the diagnostic chemical shift difference of the C16-Η in the
Conclusions
In summary, here we disclose a new highly stereoselective,
ligand-dependent protocol for the installation of exocyclic
stereocenters in a steroid D-ring via asymmetric hydrovinylation.
Phosphoramidites derived from enantiopure (S)-1,1′-bisphos-
phinoxynaphthyl ligands containing (-)-bis[(R)-1-phenyl-
ethyl]amine (L12) gives exclusively 20(R)-hydrovinylation
adduct, where as enantiomeric ligand (RaScSc, L5) gives the
(20) Lambert, J. B.; Shurvell, H. F.; Lightner, D. A.; Cooks, R. G. Organic
Structural Spectroscopy; Printice-Hall: Upper Saddle River, NJ, 1998;
p 49.
(21) Trost, B. M.; Verhoeven, T. R. J. Am. Chem. Soc. 1978, 100, 3435.
9
9004 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Stereoselective Route to Steroid-C20 (S) or -C20 (R)
A R T I C L E S
20(S) product. In the case of the corresponding ligand with the
biphenyl scaffolding, L10, bis[(S)-1-phenylethyl]amine sub-
stituent induces 20(S)-selectivity. This result is consistent with
the observation that the chiral amine component dictates
the atropisomeric nature of the fluxional biphenyl unit
when the (SS)-amine is resident, leading to (R) axial chirality
at the biphenyl22 and the attendant 20(S)-selectivity in the
hydrovinylation reaction. The two alkenes in the resultant diene
have differing steric demands for several potential reactions and
are ideally juxtaposed for further D-ring functionalization for
elaboration along the chain. Such studies are in progress. A
slight modification of the reaction also allows the stereoselective
preparation of C16-vinyl derivatives.
Acknowledgment. Financial assistance for this research by the
U.S. National Science Foundation (CHE-0610349) and the National
Institutes of Health (General Medical Sciences, R01 GM075107)
is gratefully acknowledged.
Supporting Information Available: Full experimental details
for the preparation of the substrates and protocols for the
1
hydrovinylation reactions; H and 13C NMR spectra of key
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(22) For a related result in Cu-catalyzed conjugate addition of diethylzinc
reagents to enones, see ref 15c.
JA711475F
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 9005