Divergent reactivity via cobalt catalysis: An epoxide olefination

Article abstract of DOI:10.1021/acs.orglett.5b03514

Cobalt salts exert an unexpected and profound influence on the reactivity of epoxides with dimethylsulfoxonium methylide. In the presence of a cobalt catalyst, conditions for epoxide to an oxetane ring expansion instead deliver homoallylic alcohol products, corresponding to a two-carbon epoxide homologation/ring-opening tandem process. The observed reactivity change appears to be specifically due to cobalt salts and is broadly applicable to a variety of epoxides, retaining the initial stereochemistry. This transformation also provides operationally simple access to enantiopure homoallylic alcohols from chiral epoxides without use of organometallic reagents. Tandem epoxidation-homologation of aldehydes in a single step is also demonstrated.

Full text of DOI:10.1021/acs.orglett.5b03514

Letter
pubs.acs.org/OrgLett
Divergent Reactivity via Cobalt Catalysis: An Epoxide Olefination
Megan L. Jamieson, Paul A. Hume, Daniel P. Furkert,* and Margaret A. Brimble*
School of Chemical Sciences and Maurice Wilkins Centre for Molecular Biodiscovery, University of Auckland, 23 Symonds Street,
Auckland 1142, New Zealand
S
* Supporting Information
ABSTRACT: Cobalt salts exert an unexpected and profound influence on the reactivity of epoxides with dimethylsulfoxonium
methylide. In the presence of a cobalt catalyst, conditions for epoxide to an oxetane ring expansion instead deliver homoallylic
alcohol products, corresponding to a two-carbon epoxide homologation/ring-opening tandem process. The observed reactivity
change appears to be specifically due to cobalt salts and is broadly applicable to a variety of epoxides, retaining the initial
stereochemistry. This transformation also provides operationally simple access to enantiopure homoallylic alcohols from chiral
epoxides without use of organometallic reagents. Tandem epoxidation−homologation of aldehydes in a single step is also
demonstrated.
Scheme 2. Ring Expansion of Epoxides 4 to Oxetanes 5 Using
Dimethylsulfoxonium Methylide, Initially Reported by
Okuma,3 and Extended to Oxolanes 6 by Shreiner and Fokin4
ydrolytic kinetic resolution (HKR) employing the
HJacobsen cobalt(III) salen complex A (Scheme 1),
prepared in situ from cobalt(II) precatalyst B, enables robust,
scalable, and economical access to enantiopure epoxides. The
reaction tolerates a wide variety of terminal epoxides and
routinely delivers products of >99% ee on scales suitable for
research and process applications.1 These attributes have
encouraged its use in a wide range of transformations for the
asymmetric synthesis of complex molecular architectures.2
In the context of a synthetic program, we planned to apply the
epoxide to oxetane ring-expansion methodology initially
described by Okuma (Scheme 2),3 and reinvestigated in detail
by Schreiner and Fokin in 2010,4 to chiral epoxides readily
available via Jacobsen HKR.
expansion of oxetanes 5 to oxolanes 6 has also been shown to
proceed in moderate to excellent yields but required higher
temperatures and prolonged reaction times. Density functional
calculations and experimental results have been used to estimate
the energy barriers for formation of the key reaction
intermediates and relative reactivity trends for 2-monosubsti-
tuted and 2,2-disubstituted epoxides and oxetanes.4 Due to the
similar ring-strain energy of oxetanes and epoxides (107 and 114
kJ/mol, respectively),5 a number of epoxide ring-opening
transformations have also been successfully applied to oxetanes
using heteroatomic nucleophiles6 or carbon nucleophiles7 or by
desymmetrization of prochiral 3-substituted oxetanes.8 Oxetanes
are used relatively infrequently in synthesis to date, despite the
many methods available for synthesis of homochiral epoxides or
other precursors,9 although there has recently been increased
interest in 3-substituted derivatives in a drug discovery context.10
The initial conversion of epoxides 4 to oxetanes 5 has been
established to proceed cleanly over 2−5 days, at 50−80 °C, using
2−6 equiv of dimethylsulfoxonium methylide.3 Further ring
Scheme 1. Hydrolytic Kinetic Resolution of Epoxides 1 Using
the Jacobsen Cobalt(III) Salen Complex A
Received: December 10, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b03514
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
a
Scheme 3. Divergent Reactivity of (Top) Racemic ( )-7
Giving Oxetane 8 and (Bottom) Enantiopure (R)-7, Resolved
by Jacobsen HKR, Giving Homoallylic Alcohol 9
Table 2. Lewis Acid Catalyst Screen for Epoxide Olefination
b
entry
catalyst
7 (yield, %)
8 (yield, %)
9 (yield, %)
c
1
2
3
4
5
6
7
BF3·OEt2
Sc(OTf)3
CoBr2
100
c
100
d
d
72
21
d
CoCl2
47
c
FeCl3
100
c
CoBr2/bipy
FeCl3/bipy
100
d
98
a
Reagents and conditions: trimethylsulfoxonium iodide (3 equiv), t-
b
c
BuOK (3 equiv), t-BuOH, 80 °C, 3 h. 5 mol %. Sole component of
product mixture by H NMR; Isolated yield. bipy = 2,2′-bipyridyl.
d
1
conditions using the Co(III) (salen)·OAc catalyst (S,S)-A.
Surprisingly, submission of purified (R)-7 (99% ee) to the
optimized ring-expansion conditions only afforded trace
amounts of the expected oxetane (R)-8, instead delivering
homoallylic alcohol 9 in high yield (Table 1, entry 2). Several
repetitions confirmed that only batches of the resolved
homochiral epoxide (R)-7 underwent the side reaction,
suggesting that a possible cause was a trace amount of catalyst
A remaining after purification by flash chromatography.
To investigate this possibility, a sample of (R)-7 was repurified
by Kugelrohr distillation to remove all traces of catalyst and
submitted to the standard reaction conditions. Only oxetane 8
resulted in 76% yield (Table 1, entry 3). Reaction of racemic 7 in
the presence of 5 mol % of either the Co(III) (salen)·OAc
catalyst A (Table 1, entry 4) or Co(II) (salen) precatalyst B
(Table 1, entry 5) as an additive was found in both cases to afford
homoallylic alcohol 9 as the sole product in high yields.
Interestingly, the reaction catalyzed by the Co(II) (salen)
precatalyst B (Table 1, entry 5) appeared to give a cleaner
product mixture, in slightly higher yield. Taken together, these
data indicate that traces of the cobalt complexes were indeed
responsible for the observed divergent reactivity of epoxide 7
under the ring-expansion conditions. Finally, submission of
racemic oxetane 8 to the same conditions, with the addition of
Co(II) (salen) precatalyst B (Table 1, entry 6), only resulted in
recovered starting material. This last result indicated that oxetane
8 is not an intermediate in the reaction leading to homoallylic
alcohol 9.
Table 1. Confirmation of the Role of Catalytic Cobalt(II) and
-(III) Salts in the Epoxide Olefination
a
entry
sm
catalyst
oxetane 8 (%)
alcohol 9 (%)
1
2
3
4
5
6
( )-7
76
9
0
84
0
b
(R)-7
(R)-7
c
76
0
( )-7
( )-7
( )-8
A
B
89
0
100
d
B
93
0
a
b
c
d
5 mol %. Resolved by HKR. Purified by Kugelrohr distillation. I.e.,
no reaction; starting material ( )-8 recovered.
Scheme 4. Proposed Reaction Pathway for Cobalt-Catalyzed
Epoxide Olefination
On the basis of the studies performed by Shreiner and Fokin, it
appears likely that the mechanism initially proceeds through ring-
opening of epoxide 7 with dimethylsulfoxonium methylide
(Scheme 4) to give an intermediate betaine 10. In the absence of
catalyst, the betaine then preferentially undergoes intramolecular
cyclization of the alkoxide, with the loss of DMSO, to afford the
oxetane 8 that does not undergo further reaction. Under the basic
conditions employed, betaine cyclization theoretically competes
with β-deprotonation and elimination of DMSO to afford allylic
alcohol 12. During this study, however, this compound was only
observed once as a trace byproduct (<2%). Addition of a cobalt
catalyst diverts the course of the reaction, promoting the
exclusive formation of homoallylic alcohol 9. This is likely to
proceed via addition of a second equivalent of dimethylsulfoxo-
nium methylide to betaine 10 to give homologated betaine 11,
although the exact mechanism for this is unclear. β-
Deprotonation and elimination of DMSO from 11 then yields
homoallylic alcohol 9 that does not undergo further reaction.
Interestingly, the possible cyclization product 13 was also not
observed in any reaction during this study.
In order to apply the epoxide to oxetane ring expansion in our
synthetic investigations, benzyl-protected epoxide ( )-7
(Scheme 3) was treated with trimethylsulfoxonium iodide (5
equiv) and potassium tert-butoxide (5 equiv) in tert-butyl
alcohol, according to the literature conditions, to give oxetane
( )-8 in 63% yield after 3.5 days. In our hands, this was improved
to 3 equiv of each reagent at 60 °C for 4 h to give oxetane ( )-8
in 76% yield.
Expecting that (R)-8 would be similarly available from glycidol
(R)-7, epoxide ( )-7 was resolved under standard HKR
B
DOI: 10.1021/acs.orglett.5b03514
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 3. Substrate Scope for the Cobalt-Catalyzed Epoxide Homologation−Olefination*
*
a
Reagents and conditions: trimethylsulfoxonium iodide (3 equiv), B (5 mol %), t-BuOK (3 equiv), t-BuOH, 80 °C, 3 h. Determined by chiral
b
c
HPLC or 19F NMR of a Mosher ester derivative; see the Supporting Information. Product volatile. Relative stereochemistry shown for entry 11.
In an effort to clarify the role of catalytic amounts of cobalt in
diverting the reaction course, a screen of potential Lewis acid
catalysts was carried out using the standard conditions (Table 2).
Interestingly, the presence of boron trifluoride prevented ring
expansion, returning only epoxide starting material (Table 2,
entry 1). Addition of scandium triflate or iron trichloride had no
effect on the reaction, which proceeded as normal to the oxetane
(Table 2, entries 2, 5, and 7). Remarkably, the reaction was
C
DOI: 10.1021/acs.orglett.5b03514
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
selectively diverted to produce homoallylic alcohol 9 instead of
oxetane 8 exclusively in the presence of cobalt salts as the sole
product in nearly all cases (Table 2, entries 3, 4, and 6). The
addition of the bidentate 2,2′-bipyridyl (bipy) ligand appeared to
have no influence on the reaction outcome (Table 2, entries 6
and 7). These data suggest that the observed change in reactivity
is not due to general Lewis acid catalysis but rather is specific to
cobalt salts, out of those in this study. It remains to be clarified
whether this is due to a particular ability of cobalt to promote
addition of a second equivalent of the sulfoxonium ylid to
intermediate betaine 10 to give 11 (Scheme 4) or, more
intriguingly, to the possible formation of a nucleophilic
vinylcobalt species in situ.
In order to explore the generality of the epoxide olefination, a
variety of substrates were investigated (Table 3). Epoxides
possessing a variety of substitution patterns were shown to
efficiently undergo olefination to the corresponding homoallylic
alcohols, in good to excellent yields. Monosubstituted epoxides
(Table 3, entries 1−3, 5, 8, and 9) proved to be excellent
substrates, in many cases affording nearly quantitative yields.
Monosubstituted epoxides possessing bulky substituents gave
reduced conversion to the desired products (Table 3, entries 6
and 10). 1,1-Disubstituted epoxides were also well tolerated
(Table 3, entries 4 and 7). The reaction of 1,2-disubstituted
epoxides also proceeded (Table 3, entry 11), although the
reaction was much slower. Due to the use of dimethylsulfoxo-
nium methylide for Corey−Chaykovsky epoxidation of
aldehydes, we were interested in investigating the possibility of
tandem epoxidation−olefination under cobalt catalysis. Accord-
ingly, an aromatic aldehyde was exposed to the standard reaction
conditions (Table 3, entry 12), successfully affording the
homoallylic alcohol product expected from the epoxidation−
olefination process, in moderate yield. The enantiopurity of the
initial epoxides was transferred unchanged to the homoallylic
alcohol products, as determined for selected examples by Mosher
ester analysis and 19F NMR or chiral HPLC (Table 3, entries 2, 6,
and 8).
In summary, a cobalt-catalyzed homologation of epoxides to
homoallylic alcohols using dimethylsulfoxonium methylide is
reported. The unique catalytic role of cobalt salts was confirmed
through a series of control experiments. The reaction
demonstrated wide substrate scope, with epoxides possessing a
range of substituents being converted to homoallylic alcohols in
good to excellent yields, with retention of starting material
stereochemistry. Tandem epoxidation−olefination of aldehydes
was also demonstrated. Initial investigations showed that the
mechanism does not proceed through an oxetane intermediate,
but further work is required to elucidate the exact role of the
cobalt catalyst in the homologation. Finally, this novel trans-
formation enables convenient synthesis of enantiopure homo-
allylic alcohols from readily available chiral epoxides without the
use of organometallic reagents.
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: d.furkert@auckland.ac.nz.
*E-mail: m.brimble@auckland.ac.nz.
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) (a) Tokunaga, M.; Larrow, J. F.; Jacobsen, E. N.; Kakiuchi, F.
Science 1997, 277, 936−938. (b) Schaus, S. E.; Brandes, B. D.; Larrow, J.
F.; Tokunaga, M.; Hansen, K. B.; Gould, A. E.; Furrow, M. E.; Jacobsen,
E. N. J. Am. Chem. Soc. 2002, 124, 1307−1315.
(2) Review: Pellissier, H. Adv. Synth. Catal. 2011, 353, 1613−1666.
(3) Okuma, K.; Tanaka, Y.; Kaji, S.; Ohta, H. J. Org. Chem. 1983, 48,
5133−5134.
(4) Butova, E. D.; Barabash, A. V.; Petrova, A. A.; Kleiner, C. M.;
Schreiner, P. R.; Fokin, A. A. J. Org. Chem. 2010, 75, 6229−6235.
(5) Pell, A. S.; Pilcher, G. Trans. Faraday Soc. 1965, 61, 71.
(6) (a) Papini, A.; Ricci, A.; Taddei, M. J. Chem. Soc., Perkin Trans. 1
1984, 2261−2265. (b) Chini, M.; Crotti, P.; Favero, L.; Macchia, F.
Tetrahedron Lett. 1994, 35, 761−764. (c) Fernan
́
dez-Per
́
ez, H.; Etayo,
P.; Nunez-Rico, J. L.; Balakrishna, B.; Vidal-Ferran, A. RSC Adv. 2014, 4,
́
̃
58440.
(7) (a) Huynh, C.; Derguini-Boumechal, F.; Linstrumelle, G.
Tetrahedron Lett. 1979, 20, 1503−1506. (b) Carr, S. A.; Weber, W. P.
J. Org. Chem. 1985, 50, 2782−2785. (c) Jung, S. H.; Jang, S. Y. Bull.
Korean Chem. Soc. 2010, 31, 3431−3433. (d) Hernan
́
dez-Cervantes, C.;
́
Alvarez-Corral, M.; Munoz-Dorado, M.; Rodríguez-García, I. Curr. Org.
̃
Chem. 2014, 18, 525−546.
(8) (a) Loy, R. N.; Jacobsen, E. N. J. Am. Chem. Soc. 2009, 131, 2786−
2787. (b) Wang, Z.; Chen, Z.; Sun, J. Angew. Chem., Int. Ed. 2013, 52,
6685−6688.
(9) For a recent review of synthetic methods for oxetane preparation,
see: Davis, O. A.; Bull, J. A. Synlett 2015, 26, 1283−1288.
(10) (a) Stepan, A. F.; Karki, K.; McDonald, W. S.; Dorff, P. H.; Dutra,
J. K.; DiRico, K. J.; Won, A.; Subramanyam, C.; Efremov, I. V.;
O’Donnell, C. J.; Nolan, C. E.; Becker, S. L.; Pustilnik, L. R.; Sneed, B.;
Sun, H.; Lu, Y.; Robshaw, A. E.; Riddell, D.; O’Sullivan, T. J.; Sibley, E.;
Capetta, S.; Atchison, K.; Hallgren, A. J.; Miller, E.; Wood, A.; Obach, R.
S. J. Med. Chem. 2011, 54, 7772−7783. Reviews: (b) Burkhard, J. A.;
Wuitschik, G.; Rogers-Evans, M.; Muller, K.; Carreira, E. M. Angew.
Chem., Int. Ed. 2010, 49, 9052−9067. (c) Wuitschik, G.; Carreira, E. M.;
Wagner, B.; Fischer, H.; Parrilla, I.; Schuler, F.; Rogers-Evans, M.;
̈
Muller, K. J. Med. Chem. 2010, 53, 3227−3246.
̈
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.orglett.5b03514.
1H and 13C NMR spectra for novel compounds, 19F NMR,
and chiral HPLC traces (PDF)
D
DOI: 10.1021/acs.orglett.5b03514
Org. Lett. XXXX, XXX, XXX−XXX