Control of Chemo-, Regio-, and Enantioselectivity in Copper Hydride Reductions of Morita-Baylis-Hillman Adducts

Article abstract of DOI:10.1021/acs.orglett.6b03464

Nonracemically ligated copper hydride can be used to effect tandem S_N2′/1,2-reductions of racemic Morita-Baylis-Hillman (MBH) acetates to access enantioenriched chiral allylic alcohols with defined olefin geometry. MBH esters, including those w

Full text of DOI:10.1021/acs.orglett.6b03464

Letter
pubs.acs.org/OrgLett
Control of Chemo‑, Regio‑, and Enantioselectivity in Copper Hydride
Reductions of Morita−Baylis−Hillman Adducts
Roscoe T. H. Linstadt, Carl. A. Peterson, Carina I. Jette, Zarko V. Boskovic, and Bruce H. Lipshutz*
Department of Chemistry and Biochemistry, University of California, Santa Barbara, California 93106, United States
S
* Supporting Information
ABSTRACT: Nonracemically ligated copper hydride can be used
to effect tandem SN2′/1,2-reductions of racemic Morita−Baylis−
Hillman (MBH) acetates to access enantioenriched chiral allylic
alcohols with defined olefin geometry. MBH esters, including those
with β-substitution, can be transformed to stereodefined enoates by
taking advantage of a bulky, oligomeric, in situ generated
trialkoxysiloxane leaving group. Finally, an atypical conversion of easily arrived at MBH alcohol derivatives to nonracemic
allylic alcohols is disclosed.
orita−Baylis−Hillman (MBH) adducts are well-known,
versatile intermediates in synthesis owing to their
Table 1. Optimization Studies
M
multiple electrophilic sites.1 Their potential for use in reactions
with copper hydride (CuH), especially nonracemically ligated
CuH, has yet to be realized. Since Stryker’s seminal reports,2a−c
CuH-mediated methods for numerous types of reduction have
reached considerable sophistication in recent years, with catalytic
systems allowing for highly chemo- and enantioselective 1,2- or
1,4-reductions under mild conditions.3a−g While substitution
reactions based on copper have been extensively developed for
carbon-, boron-, and silicon-type nucleophiles,4a−d substitutions
involving hydride have been far less studied.5a−d
Examining simple MBH acetate 1a (Table 1) immediately
poses the interesting question of regioselectivity, where CuH
(prepared from Cu(OAc)2 and a silane) could, in principle,
deliver hydride to the activated allylic site or add hydride in a 1,4-
sense to the enoate as a Michael acceptor. CuH reductions are
widely believed to give rise to an O−Cu enolate after 1,4-hydride
delivery,6a,b,7 and a 1,4-addition pathway in this case could
potentially lead to a silyl ketene acetal or undergo anti-
elimination to give rise to a new enoate, 2. Subjecting 1a to
CuH (THF, rt) led to only enoate 2 and the corresponding over-
reduced ester 3, the former with 8:1 E/Z selectivity. These results
suggest that allylic substitution, in this case, is faster than 1,4-
reduction. The difficulty in separating the E-enoate from both its
Z isomer and over-reduced product 3 would likely preclude use
of this chemistry as a general method to arrive at α-substituted
enoates. Therefore, the process was optimized to minimize
formation of 3 and improve E/Z selectivity.8
a
Reactions were performed under an inert atmosphere at 0.4 M in
b
THF. Determined by GCMS of the crude reaction mixture.
c
Determined by 1H NMR of the pure product. DEMS =
diethoxymethylsilane. PMHS = polymethylhydrosiloxane
Control experiments revealed that the reaction is not induced
by silane alone, nor does a silane in the presence of catalytic
phosphine ligand lead to reduction; copper is clearly required for
conversion (Table 1, entries 1−3). Attempts to mitigate over-
reduction by varying temperature, time, and leaving group were
unsuccessful. Several of the commonly used ligands for CuH
either behaved sluggishly or led to an undesirable amount of
over-reduced product (entries 5−7). Ligand L3a, on the other
hand, gave rise to the desired enoate 2 in 6:1 E/Z selectivity; no
detectable over-reduction (entry 8) was observed even in the
presence of excess silane or prolonged reaction times. The use of
less expensive PMHS, as opposed to DEMS, gave comparable
results.9 Although in the absence of a leaving group it was
Received: November 20, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b03464
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
expected that 1,4-reduction would predominate, to our surprise,
with alcohol 1b as substrate, vigorous gas evolution was noted
and product enoate 2 was obtained quantitatively and with
dramatically higher (>20:1) E/Z selectivity (entry 10). This
result implies that the reaction of 1b initially proceeded through a
dehydrogenative silylation10 to afford a PMHS-bound silyl ether
where 1,4-addition and subsequent elimination gives rise to the
desired product. The higher E/Z selectivity can be rationalized
given the much larger size of the oligomeric11 O-PMHS-bound
leaving group, forcing a highly anti-selective elimination (Scheme
1).
congested nature at the alcohol site, dehydrogenative silylation
remains kinetically favored over 1,4-addition, thus providing a
comparatively mild albeit nontraditional method to access
tetrasubstituted olefins such as 9 and 10. Educts possessing
substitution at the β-site of the Michael acceptor were also tested,
giving rise to enoates 11 and 12. While both were obtained
uneventfully with high levels of stereocontrol, addition of t-
BuOH14 was needed to enhance catalyst regeneration.
Since silyl ethers are not traditionally thought of as competent
leaving groups in substitution chemistry, we questioned whether
the high reaction efficiency might be a result of the electronics of
the proposed trialkoxysilyl ether intermediate and/or the
activated nature of MBH adducts. The reaction of CuH on
TBS ether 13 (Scheme 3) gave a mixture of 1,4-reduction
Scheme 1. Proposed Reaction Mechanism
Scheme 3. Probing Silyl Ethers as Leaving Groups*
These encouraging results prompted us to examine the scope
of these reductions for several MBH alcohols (Scheme 2). In
general, isolated yields were quite good for most examples. In all
cases, E/Z selectivity for 2° alcohols was observed to be >20:1,
with no detectable over-reduction.
Scheme 2. Stereoselective Reductions of MBH Esters*
*
Conditions A: 6 mol % of Xantphos/Cu(OAc)2·H2O, 1.25 equiv of
(EtO)2MeSiH or EtO(Me)2SiH (0.2 M), Et2O, rt, 2 h. Conditions B:
a
1.35 equiv of n-HexMgBr, −78 °C to rt, overnight. Determined by
GCMS of crude reaction mixture. Determined by H NMR of pure
product.
b
1
product 14, allylic substitution product 2, and starting material as
confirmed by GCMS, suggesting that steric bulk alone of the silyl
ether is insufficient for substitution.15 Likewise, copper-catalyzed
dehydrogenative silylation10 on p-bromocinnamyl alcohol using
either DEMS or dimethylethoxysilane was carried out. Once the
intermediate silyl ether was fully formed, the solutions were
treated at cryogenic temperature with Grignard reagent to effect
a copper-catalyzed allylic alkylation. Only DEMS, which gives
rise to a trialkoxysilyl ether intermediate, gave allyllic substitution
product as a mixture of linear and branched regioisomers, lending
support to our hypothesis that a trialkoxysilyl ether is crucial for
these displacement reactions.12a,b
With various substitution patterns on MBH esters having been
examined, we turned our attention to the much more intriguing
question of regioselectivity with MBH ketones, since we had
previously shown that enones bearing an α-substituent can
redirect nonracemically ligated CuH to react in a 1,2-sense to
arrive at chiral allylic alcohols.3c,f Hence, if CuH could be directed
to react in a sense similar to that seen with MBH esters (vide
supra), the resulting α-substituted enone would subsequently be
expected to undergo further asymmetric 1,2-reduction to
ultimately arrive at a nonracemic allylic alcohol. This seemed
plausible, as the BIPHEP series of ligands are competent for use
in these reductions of MBH esters and had already been shown
to be highly enantioselective for the anticipated 1,2-reductions.
Indeed, subjecting acetate 17 to previously developed conditions
with ligand L3c led to allylic alcohol 26 in 92% isolated yield and
*
Yields in parentheses are the percentage of product isolated after
a
b
c
chromatography. Reaction time 18 h. Reaction time 24 h. With 3
equiv of t-BuOH and reaction time 72 h.
Most methodology developed for functionalization of MBH
adducts relies on substrates accessible from traditional
conditions. (e.g., acrylate, aldehyde, DABCO). On the other
hand, Ramachandran’s hydroalumination of ynoates13 signifi-
cantly expands the scope of MBH derivatives that can be
prepared, including adducts possessing both tertiary alcohols and
those with β-substitution. Accordingly, four different MBH
adducts were prepared by this method and tested as the alcohol
or as the corresponding acetate. The reactions with tertiary MBH
alcohols are particularly interesting cases, as despite their
B
DOI: 10.1021/acs.orglett.6b03464
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
93% ee with 100% control of olefin geometry (Table 2).16 Thus,
as is the case with MBH esters, allylic reduction is a much faster
completely reversed the mode of reduction back to the expected
allylic substitution/1,2-reduction pathway, affording 34 albeit in
substantially reduced enantioselectivity (55% ee). That
regioselectivity can be completely inverted by modest alterations
in ligand architecture highlights the subtleties associated with
substrate recognition where the reactive functionality is forced
into close proximity.
Table 2. Enantioselective Reductions of MBH Ketones
With the successful implementation of the tandem SN2′/1,2-
reductions of MBH acetates, the functionality in the MBH
adducts seemed to offer good prospects for multiple manipu-
lations. In particular, a three-step process might include a
preliminary dehydrogenative silylation to activate the alcohol,
prior to double reduction. Although CuH could potentially react
with 36 in either a 1,4 or 1,2-sense, or at the alcohol, only an
initial silylation would arrive at the requisite leaving group for
further substitution (Scheme 4, top). In the event, subjecting 36
Scheme 4. Enantioselective Tandem Reactions
to our standard conditions afforded allylic alcohol 26 as the major
product in essentially the same level of enantioselectivity as seen
using the corresponding acetate as educt. Thus, in terms of an
overall transformation, a racemic and readily available substrate,
and lacking alkene stereochemistry, was transformed over three
consecutive steps (68% average yield per step) by a single catalyst
to an enantioenriched allylic alcohol with defined double-bond
geometry.
Lastly, to further demonstrate the potential of this method-
ology to prepare chiral synthons, and for its use under green
chemistry conditions, we subjected MBH acetate 17 to
Suginome’s silylborane under micellar conditions followed by
addition of PMHS to accomplish both introduction of a
stereodefined allylic silane and enantioselective 1,2-reduction
in water at room temperature to arrive at 37 in good yield
(Scheme 4, bottom).17,18a,b
In summary, by harnessing each of the three major reactivity
modes of CuH (i.e., dehydrogenative silylation, 1,4-addition, and
1,2-addition), racemic MBH adducts can be efficiently trans-
formed into stereodefined enoates or nonracemic allylic alcohols
via controlled tandem reduction sequences. Evidence is provided
that an atypical, in situ generated trialkoxysilyl ether as leaving
group is important for efficient substitution. When derived from
oligomeric PMHS, excellent stereocontrol of the resulting
alkenyl derivatives can be obtained, presumably due to the steric
demand of these -O-PMHS ethers. The versatility of CuH has
been further extended to include a previously unknown and
potentially very useful sequence that converts a racemic MBH
a
Conditions: 3 mol % of CuOAc2, 3 mol % of L3c, 4 equiv of DEMS,
b
Et2O (0.4 M), −25 °C, then quench NH4F/MeOH. Reaction
conducted at rt in aqueous TPGS-750-M using PMHS as H-source.
process than 1,4-reduction, and in this case, 1,2-reduction
notwithstanding the α-substituent in the starting material. No
isomerization of the product allylic alcohols was detected.
The scope of this tandem allylic substitution/asymmetric 1,2-
reduction was then further examined (Table 2). Electron-rich 18
was reduced uneventfully, as was electron-deficient 19. Ethyl
ketones 20 and 21 possessing benzyl-protected phenols
displayed remarkable enantioselectivity, with the latter product
obtained in essentially enantiopure form. Naphthyl and hydro-
cinnamyl substrates reacted smoothly to form products 31 and
32. Diastereomeric substrate 24 was converted to 33 with high
diastereoselectivity, implying that the influence of the distal
methyl group was not significant. While the s-cis conformation of
enones has been previously suggested to favor 1,2-reduction by
CuH in the presence of bulky biaryl bis-phosphines such as L2,3f
employing cyclic substrate 25 with an s-trans conformation led
predominantly to the 1,2-adduct 35. Remarkably, L3c
C
DOI: 10.1021/acs.orglett.6b03464
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
and variable E/Z ratios are often observed. See: Nagaoka, H.; Kishi, Y.
Tetrahedron 1981, 37, 3873.
(9) Aldrich prices as of Apr 13, 2016: PMHS [cat. no. 176206] was
calculated at $0.11/mmol H− vs DEMS [cat. no. 66612] $2.27/mmol
H−, assuming 60 g mol−1 of effective hydride for PMHS.
(10) Ito, H.; Watanabe, A.; Sawamura, M. Org. Lett. 2005, 7, 1869.
(11) Lawrence, N. J.; Drew, M. D.; Bushell, S. M. J. Chem. Soc., Perkin
Trans. 1 1999, 3381.
alcohol into a nonracemic allylic alcohol with defined olefin
geometry. Further subtleties and applications associated with
CuH chemistry will be reported in due course.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.orglett.6b03464.
(12) (a) We ascribe the appearance of the branched isomer to the
involvement of the soluble Cu(I) catalyst, as formation of branched
products are relatively rare in the absence of copper, with Alexakis’ NHC
ligands being a notable exception. See: Jackowski, O.; Alexakis, A. Angew.
Chem., Int. Ed. 2010, 49, 3346. (b) Further experimentation showed that
the yield of this reaction could be improved to 56% with exclusively
linear product being formed by switching the ligand to BDP.
(13) While there are numerous protocols in the literature developed to
accommodate additional functionality/substitution in MBH adducts,
many of these required lengthy syntheses, specialty catalysts, or gave
unreliable results. By contrast, Ramachandran’s hydroalumination
protocol uses readily available reagents and in our hands has proven
quite reliable and easy to perform. See: Ramachandran, P. V.; Rudd, M.
T.; Burghardt, T. E.; Reddy, M. V. R. J. Org. Chem. 2003, 68, 9310.
(14) t-BuOH is confirmed to accelerate 1,4-reductions by a more rapid
quenching of a copper enolate. See: Lipshutz, B. H.; Servesko, J. M.;
Taft, B. R. J. Am. Chem. Soc. 2004, 126, 8352. (b) Its role in this
particular substitution is less clear and other modes of acceleration may
be operative, as the copper enolate presumably leads to rapid loss of the
O-PMHS silyl ether and catalyst regeneration is quite facile for reactions
of simpler MBH alcohols.
Detailed experimental procedures along with analytical
and spectral data for all new compounds (PDF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: lipshutz@chem.ucsb.edu.
ORCID
Bruce H. Lipshutz: 0000-0001-9116-7049
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the NSF (SusChEM-1561158) for support of this
work. R.L. warmly thanks Dr. Nicholas Isley (Scripps) for helpful
discussions.
(15) Reaction of MBH alcohol under silane-free conditions with either
stoichiometric PPh3CuH and catalytic L3a or catalytic BDP/CuCl/KO-
t-Bu under 1 atm of H2 gas gave predominantly recovered starting
material along with traces of the 1,4-reduction product, suggesting that
the intermediate copper alkoxide does not function as a leaving group
and hinders further reduction.
REFERENCES
■
(1) For reviews, see: (a) Basavaiah, D.; Reddy, B. S.; Badsara, S. S.
Chem. Rev. 2010, 110, 5447. (b) Basavaiah, D.; Venkateswara, K.; Reddy,
R. J. Chem. Soc. Rev. 2007, 36, 1581.
(2) (a) Brestensky, D. M.; Huseland, D. E.; McGettigan, C.; Stryker, J.
M. Tetrahedron Lett. 1988, 29, 3749. (b) Mahoney, W. S.; Brestensky, D.
M.; Stryker, J. M. J. Am. Chem. Soc. 1988, 110, 291. (c) Brestensky, D.
M.; Stryker, J. M. Tetrahedron Lett. 1989, 30, 5677.
(3) For selected examples of 1,4- and 1,2-reductions, see: (a) Lipshutz,
B. H.; Noson, K.; Chrisman, W. J. Am. Chem. Soc. 2001, 123, 12917.
(b) Lipshutz, B. H.; Shimizu, H. Angew. Chem., Int. Ed. 2004, 43, 2228.
(c) Moser, R.; Boskovic, Z. V.; Crowe, C. S.; Lipshutz, B. H. J. Am. Chem.
Soc. 2010, 132, 7852. (d) Deutsch, C.; Krause, N.; Lipshutz, B. H. Chem.
Rev. 2008, 108, 2916. (e) Moritani, Y.; Appella, D. H.; Jurkauskas, V.;
Buchwald, S. L. J. Am. Chem. Soc. 2000, 122, 6797. (f) Voigtritter, K. R.;
Isley, N. A.; Moser, R.; Aue, D. H.; Lipshutz, B. H. Tetrahedron 2012, 68,
3410. (g) Albright, A.; Gawley, R. E. J. Am. Chem. Soc. 2011, 133, 19680.
(4) For reviews and selected recent examples, see: (a) Vyas, D. J.;
Oestreich, M. Angew. Chem., Int. Ed. 2010, 49, 8513. (b) Ito, H.; Ito, S.;
Sasaki, Y.; Matsuura, K.; Sawamura, M. J. Am. Chem. Soc. 2007, 129,
14856. (c) Shido, Y.; Yoshida, M.; Tanabe, M.; Ohmiya, H.; Sawamura,
M. J. Am. Chem. Soc. 2012, 134, 18573. (d) Langlois, J.-B.; Alexakis, A.
Top. Organomet. Chem. 2011, 38, 235.
(16) The use of low-temperature conditions for reactions of the
acetates of MBH ketones is apparently satisfactory for high E-selectivity
in the initial substitution step.
(17) (a) Suginome, M.; Matsuda, T.; Ito, Y. Organometallics 2000, 19,
4647. (b) During the course of this study, another group independently
reported a similar silylation of MBH alcohols. See: Xuan, Q.-Q.; Zhong,
N.-J.; Ren, C.-L.; Liu, L.; Wang, D.; Chen, Y.-J.; Li, C.-J. J. Org. Chem.
2013, 78, 11076.
(18) CuH reductions in aqueous surfactant have been described
previously. See: (a) Huang, S.; Voigtritter, K. R.; Unger, J. B.; Lipshutz,
B. H. Synlett 2010, 2010, 2041. (b) We found the use of smaller TMS-
terminated PMHS oligomers (average Mn = 390, CAS: 63148-57-2)
produced much more reliable results by reducing excessive foaming/
clumping of the reaction mixture that was problematic with the higher
oligomers.
(5) (a) Deutsch, C.; Lipshutz, B. H.; Krause, N. Angew. Chem., Int. Ed.
2007, 46, 1650. (b) Deutsch, C.; Lipshutz, B. H.; Krause, N. Org. Lett.
2009, 11, 5010. (c) Brummond, K. M.; Lu, J. J. Am. Chem. Soc. 1999,
121, 5087. (d) Daeuble, J. F.; McGettigan, C.; Stryker, J. M. Tetrahedron
Lett. 1990, 31, 2397.
(6) The initially formed copper enolate can also be quenched with
other electrophiles (e.g., carbonyl compounds) to perform a net
reductive aldol. See: (a) Lipshutz, B. H.; Amorelli, B.; Unger, J. B. J. Am.
Chem. Soc. 2008, 130, 14378. (b) Deschamp, J.; Chuzel, O.;
Hannedouche, J.; Riant, O. Angew. Chem., Int. Ed. 2006, 45, 1292.
(7) If the copper enolate is trapped as the silyl ketene acetal, it can be
regenerated in the presence of (EtO)3SiF and CuF for use in a
subsequent aldol reaction. See: Oisaki, K.; Suto, Y.; Kanai, M.; Shibasaki,
M. J. Am. Chem. Soc. 2003, 125, 5644.
(8) While these educts could be prepared by established olefination
chemistry, it should also be noted that α-substitution in Horner−
Emmons reactions does not as a rule give isomerically pure E-enoates,
D
DOI: 10.1021/acs.orglett.6b03464
Org. Lett. XXXX, XXX, XXX−XXX