New experimental conditions for tandem hydroalumination/Cu-catalyzed asymmetric conjugate additions to β-substituted cyclic enones

Article abstract of DOI:10.1021/ol200898c

Readily available alkenylalanes, arising from hydroalumination of unprotected terminal alkynes, have been directly employed for the copper-catalyzed asymmetric conjugate addition (ACA) to β-substituted cyclic enones. The desired products, containing a qua

Full text of DOI:10.1021/ol200898c

ORGANIC
LETTERS
2011
Vol. 13, No. 12
3040–3043
New Experimental Conditions for Tandem
hydroalumination/Cu-Catalyzed
Asymmetric Conjugate Additions to
β-Substituted Cyclic Enones
€
Daniel Muller, Matthieu Tissot, and Alexandre Alexakis*
Department of Organic Chemistry, University of Geneva 30, quai Ernest Ansermet
CH-1211 Geneva 4, Switzerland
alexandre.alexakis@unige.ch
Received April 6, 2011
ABSTRACT
Readily available alkenylalanes, arising from hydroalumination of unprotected terminal alkynes, have been directly employed for the copper-
catalyzed asymmetric conjugate addition (ACA) to β-substituted cyclic enones. The desired products, containing a quaternary stereogenic center,
are generally obtained in good yields and enantioselectivities.
Since the discovery of the hydroalumination reaction of
€
hydroalumination reaction.5,6 The afforded Si-substi-
tuted vinylaluminums were used for NHC-Cu-catalyzed
ACA to β-substituted cyclic enones yielding highly en-
antioenriched compounds. However, when nonprotected
alkenylalanes were used, the enantioselectivity dropped
significantly (30% ee).5
In order to avoid the presence of acetylides, we focused
on the generation of vinylalanes from the corresponding
vinyl halides.7,8 This strategy proved to be particularly
successful for conjugated alkenylalanes such as styrenyla-
lanes but did have the drawback that for simple alkenyl
substituents, such as hexenylalane, a longer and indirect
synthesis of the mixed alane had to be effected, as shown in
Scheme 1.
1
terminal alkynes by Wilke and Muller in 1960, there have
been only few reports of the use of such alkenylalanes in Cu-
catalyzed conjugate addition reactions.2 The use of these
nucleophiles for catalytic asymmetric conjugate addition
reactions is made difficult by the presence of up to 29% of
Al-acetylides in vinylalanes.3 These acetylides might act as
competing ligands in the presence of a chiral copper com-
plex and thus lead to undesired side reactions and
decrease of enantioselectivities.4 A solution to this pro-
blem was very recently reported by Hoveyda et al., who
made use of Si-protected alkynes in order to circumvent
the problem of Al-acetylide formation during the
Therefore, we revisited the tandem hydroalumination-
ACA reaction, keeping in mind that hydroaluminations
€
(1) Wilke, G.; Muller, H. Justus Liebig Ann. Chem. 1960, 629, 222–
240.
(2) For two examples of racemic conjugate addition reactions, see: (a)
Ireland, R. E.; Wipf, P. J. Org. Chem. 1990, 55, 1425–1426. (b) Wipf, P.;
Smitrovich, J. H.; Moon, C. J. Org. Chem. 1992, 57, 3178–3186. For two
examples of tandem hydroaluminationÀACA reactions affording pro-
ducts in 50À73% ee, see:(c) Vuagnoux-d’Augustin, M.; Alexakis, A.
Chem.;Eur. J. 2007, 13, 9647–9662. (d) Palais, L.; Alexakis, A.
Chem.;Eur. J. 2009, 15, 10473–10485.
(6) The use of Si-protected alkynes to circumvent Al-acetylide for-
mation is a well-known strategy; see ref 3.
(7) For one example using (E)-1-iodohex-1-ene as a nucleophile
precursor, see: (a) Hawner, C.; Li, K.; Cirriez, V.; Alexakis, A. Angew.
Chem., Int. Ed. 2008, 47, 8211–8214. For a more general disclosure
(3) Zweifel, G.; Miller, J. A. Org. React. 1984, 32, 375.
€
affording product in up to 96% ee, see:(b) Muller, D.; Hawner, C.;
(4) For observations of the harmful effect of acetylides in ACA
reactions, see: Corey, E. J.; Kwak, Y. Org. Lett. 2004, 6, 3385–3388.
(5) May, T. L.; Dabrowski, J. A.; Hoveyda, A. H. J. Am. Chem. Soc.
2011, 133, 736–739.
Tissot, M.; Palais, L.; Alexakis, A. Synlett 2010, 1694–1698.
(8) Generated Me2Al-vinylalanes do have the advantage of being
stronger Lewis acids than (i-Bu)2Al-vinylalanes and therefore show
higher reactivity.
r
10.1021/ol200898c
Published on Web 05/18/2011
2011 American Chemical Society

Having validated the concerns for the hydroalumination
reactions for different substrates such as primary, second-
ary, and tertiary alkyl-substituted alkynes as well as for
conjugated alkynes, we started optimizing the reaction
conditions. From our experience in the field of ACA
reactions with alkenylalanes, we knew that phosphorami-
dite ligands, and especially electron-rich phosphinamine
ligands, afforded such reactions with good to excellent
enantioselectivities.2c,d,7a,7b Therefore, we envisaged
screening ligands which can be easily prepared on a multi-
gram scale and a second-generation SimplePhos ligand
L3a bearing an alkyl instead of an aryl group, thus
rendering it even more electron rich (Figure 1).14,15
Scheme 1. Direct Use of the (iBu)2Al-vinyl Nucleophile
had to be done in a way as to keep the Al-acetylide
formation as low as possible. For alkynes bearing secondary
or tertiary alkyl substituents, low amounts of the Al-acet-
ylide formation are usually observed, and the generated
alane can be used without precautions.9 We were aware that
for primary alkyl-substituted alkynes depending on the chain
length of the alkyl group significant amounts of the acetylide
are formed.10 Previous studies showed that lower amounts of
alkenylaluminums led to higher enantioselectivities, which is
probably due to the fact that the alane did contain substances
which had a deleterious effect on the active complex.11
Therefore, our initial strategy was to keep the amount of
alane as low as possible but also to employ strong donor
ligands which tightly bind to copper and therefore are less
prone to substitution by a nonchiral acetylide ligand.
Figure 1. Phosphinamine and phosphoramidite ligand library.
First, we optimized the conditions for the most challen-
ging nucleophile, namely hexenylalane, which is known to
contain 6% of deleterious Al acetylide (Table 1).10
Scheme 2. Selective Generation of R- and β-Alkenylalanes12
For all the studied reactions and especially in the
absence of ligand (entry 6), significant amounts of the
1,2-additionÀdehydration products were observed.16
As expected, conversion and the preference for 1,4-
addition was significantly improved with the strong
donor ligand L3a, whereas weaker donor ligands such
as L1 and L4 gave rise to large amounts of 1,2-addition
(entries 1À4). Surprisingly, ligand L2, though affording
high enantioselectivity (entry 2), led only to poor 1,4:1,2
(11) For instance, such a substance might be LiCl, which was shown
to decrease enantioselectivities; see ref 7b.
(12) Gao, F.; Hoveyda, A. H. J. Am. Chem. Soc. 2010, 132, 10961–
10963.
(13) In the absence of a Ni catalyst usually high levels of Al-acetylide
formation (>20%) is observed; see ref 12.
(14) See the Supporting Information for details.
(15) For the synthesis of SimplePhos Ligands, see: (a) Palais, L.;
Mikhel, I. S.; Bournaud, C.; Micouin, L.; Falciola, C. A.; Vuagnoux-d
Augustin, M.; Rosset, S.; Bernardinelli, G.; Alexakis, A. Angew Chem.
Int. Ed. 2007, 46, 7462–7465. (b) Reference 2d. (c) Bournaud, C.;
Falciola, C.; Lecourt, T.; Rosset, S.; Alexakis, A.; Micouin, L. Org.
Lett. 2006, 8, 3581–3584.
Hoveyda and co-workers recently reported the nickel-
catalyzed hydroalumination of conjugated alkynes with
high levels of R- or β-selectivity and even more important
with very low levels (<2%) of Al-acetylide formation
(Scheme 2).12,13
(9) For formation of <2% of Al-acetylide, see ref 3.
(10) 6% of Al-acetylide formation for the hydroalumination of
hexyne; see ref 3.
(16) This has been observed previously; see ref 2c.
Org. Lett., Vol. 13, No. 12, 2011
3041

naphthenate as the most inexpensive organic copper salt
afforded the same selectivities as CuTC (copper thio-
phene 2-carboxylate). Given that this Cu salt can be used
as a stock solution, we favored this Cu source over
CuTC.14 (5) Free alkyne (entry 15) did not have a major
effect on the regioselectivity of the reaction, whereas
upon addition of 20 mol % of the corresponding acet-
ylide the 1,2-addition becomes the predominant reaction
pathway (entry 16). Surprisingly, the presence of Al-
acetylide did not have a great influence on the
enantioselectivity.
With the optimized reaction conditions in hand (entry
14), we focused on enlarging the nucleophile scope. Two
methods were developed (methods A and B) because it was
found that L2 worked equally efficient as L3a when nucleo-
philes other than hexenylalane were used.17 Moreover, we
found that for substrates other than 3-methylcyclohex-2-
enone the established reaction conditions were inefficient
(low conversions). We knew that such transformations were
possible with Me2Al-vinyl species and managed to solve this
problem by activation of the substrate with 1 equiv of
Me3Al.18,19
Table 1. Optimization of Reaction Conditionsa
entry
L*
solvent
convb (%)
2a:4c
ee (2a)d (%)
1
2
L1
Et2O
100
98
65:35
51:49
85:15
53:47
87:13
25:75
92:8
51 (S)
83 (R)
76 (R)
77 (R)
77 (R)
L2
Et2O
3
L3a
L4
Et2O
100
98
4
5e
Et2O
L3a
Et2O
100
75
6
Et2O
7
L3a
L3a
L3a
L3a
L2
THF
80
66 (R)
75 (R)
74 (R)
84 (R)
83 (R)
73 (R)
83 (R)
82 (R)
83 (R)
79 (R)
81 (R)
8
EtOAc
DCM
98
94:6
9
88
42:58
84:16
24:77
52:48
93: 7
91:9
10
11
12
13e
14f
15g
16h
17i
toluene
toluene
hexane
toluene
toluene
toluene
toluene
toluene
97
96
L3a
L3a
L3a
L3a
L3a
L3a
100
72
97
100
96
82:18
32:68
54:46
Table 2. Cu-Catalyzed ACA of Unprotected Alkenylalanes to
β-Substituted Cyclic Enonesa
100
a Reaction performed under Ar atmosphere on a 0.3 mmol scale.
b Determined by GC/MS. c Determination of the corresponding elim-
ination products by GC/MS. d Determined by chiral GC analysis.
e Reaction performed with 1.2 equiv of alkenylalane. f Reaction per-
formed with 13 mol % of Cu(II)naphthenate instead of CuTC. g Addi-
tion of 20 mol % of hex-1-yne. h Addition of 20 mol % of hex-1-
ynyldiisobutylaluminum. i Reaction carried out with 11 mol % of L3a
and 7 mol % of Cu(II)napththenate on a 2.0 mmol scale.
substrate
reagent
(R1)
yieldd eee
(%) (%)
ratios, implying that not only electronic factors influ-
ence the regioselectivity.
entry (R; method)b
n
1,4:1,2c
1f
2
Me; A
Me; A
Me; A
Me; B
Me; A
Me; B
Et; A
0
1
1
1
1
1
1
1
2
n-Bu
n-Bu
(CH)2 (c-C5H9) 85:15 49 (2b) 75
82:18 44 (1a) 34
Because of the high modularity of alkyl-substituted
phospinamines, we synthesized a small library of such
ligands and screened them for the reaction depicted in
Table 1.14 R-Branched alkyl groups (L3b, L3c, L3d, L3e)
were impossible to synthesize, which we believe is due to
the high steric demand. β-Branched alkyl substituents
(L3f, L3g) only affordedmessy reactionsand low enantios-
electivities. Only γ-branched ligands (L3h, L3i) afforded
reasonable levels of enantioselectivity and clean reactions;
however, they did not surpass the selectivities achieved by
their n-Bu counterpart. Forthe linearalkylsubstituentsthe
following trend was observed: L3o < L3j < L3k < L3n <
3m < 3l. Hence, the initially used ligand L3a was the most
selective ligand of its kind for this reaction.
Final optimizations and control experiments were
carried out (entries 7À17, Table 1).14 Several points
regarding these reactions are noteworthy: (1) Donor
solvents such as THF and EtOAc favor the 1,4-addition,
whereas noncoordinating solvents such as hexane and
DCM lead to high levels of 1,2-addition. (2) EtOAc can
be used for such reactions, implying the high functional
group tolerance of such alkenylalanes. (3) Lowering the
amount of alane does slightly favor the 1,4-adduct
(entries 5 and 13). (4) Only 13 mol % of Cu(II)
91:9 72 (2a) 83
3
4
Bn
95:5
94:6
95:5
92:8
61 (2c) 82
91 (2d) 79
79 (2e) 85
5
Cy
6
t-Bu
n-Bu
n-Bu
n-Bu
7f
8f
9f
65 (2f)
55
i-Bu; B
Me; A
63:37 41 (2g) 89
93:7 57 (3a) 79
a Reactions performed under argon on a 0.6 mmol scale. b Method A:
L3a, toluene. Method B: L2, Et2O. c Determined by GC/MS. d Yield of
isolated 1,4-adduct. e Determined by chiral GC. f Activation with 1.0
equiv of Me3Al, Et2O, À20 °C.
As the data in Table 2 indicate, a relatively large scope of
alkyl-substituted alkenes undergo ACA efficiently with
good levels of enantioselectivity. Furthermore, we were
able to achieve high levels of enantioselectivity for
(17) Concerning the selectivities of L2 and L3a, most of the reactions
showed differences in enantioselectivities of <15%. However, L3a gives
higher enantioselectivities for 5- and 7-membered substrates, whereas L2
affords particularly high enantioselectivities for sterically demanding
substrates or nucleophiles.
(18) Unpublished results.
(19) From the work of Wipf, we already knew that activation with
TMS-Cl and BF3 did not accelerate such reactions; see ref 2b.
3042
Org. Lett., Vol. 13, No. 12, 2011

sterically hindered substrates and cycloheptenone (entries
8 and 9).20 Not surprisingly, cyclopentenone adduct 1awas
obtained with low levels of enantioselectivities (entry 1).21
These findings imply that the developed methodology is
somewhat complementary to Hoveyda’s recent work on
Si-substituted vinyl alanes.5
Scheme 3. Further Extension of the Nucleophile Scopea
In order to further extend the nucleophile scope, we
thought about introduction of Z-alkenes (2h), halogen-
containing alkenes (2i), conjugated alkenes (2j), and
R-substituted alkenes (2k). Z-Alkenes are easily acces-
sible via hydroalumination of propargylic ethers,22
whereas for the R-substituted alkene and the conju-
gated alkene we envisaged employing Ni-catalyzed
hydroalumination as recently explored by Hoveyda
and co-workers (Scheme 2).12
We were delighted to see that all of these challenging
nucleophiles did undergo ACA to 3-methylcyclohex-2-
enone with good levels of enantioselectivity (Scheme 3).
Several points for Scheme 3 are noteworthy: (1) For all of
the depicted adducts there isnoprecedence forthe addition
of such nucleophiles in tandem hydroaluminationÀACA
reactions.23 (2) Despite the presence of THF and nickel
salts, good enantioselectivities were obtained for both the
R- and the β-styrenyl adducts. (3) Surprisingly, the R- and
β-styrenyl compounds gave no rise to 1,2-addition. (4) In
the case of the β-styrenyl adduct, we observed no addition
of the R-styrenylalane, which is in accord with previous
observations.24
a Reaction conditions as indicated for Table 2.
b Activation with 1.0 equiv of Me3Al, À20 °C.
protocol is provided for the synthesis of L2 and L3a on a
multigram scale which we hope will stimulate the use of
such ligands in asymmetric transformations.
Acknowledgment. 2cWe thank the Swiss National Re-
search Foundation (Grant No. 200020-126663) and COST
action D40 (SER Contract No. C07.0097) for financial sup-
port as well as BASF for the generous gift of chiral amines.
In conclusion, we have shown that nonprotected in situ
generated alanes can be employed directly for the Cu-
catalyzed ACA. Challenges in such reactions, namely the
presence of Al-acetylides, have been met making use of
several strategies. Moreover, a library of second-genera-
tionSimplePhosligandshas been synthesized, and a simple
Supporting Information Available. Experimental pro-
cedures, NMR spectra, and chiral separations for all
compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.
(22) Alexakis, A.; Duffault, J. M. Tetrahedron Lett. 1988, 29,
6243–6246.
(23) For the β-styrenyl adduct there has been only precedence for
Si-protected alkenyl alanes; see ref 5.
(24) As previously observed, β-styrenylalane adds much faster than
the R-styrenylalane, see ref 7b.
(20) Cycloheptenones prooved to be inefficient in catalysis with
Si-protected alkenyl alanes; see ref 5.
(21) L2 and L4 afforded inferior enantioselectivities for the addition
of trialkylaluminum to β-substituted cyclopentenones compared to the
cyclohexenone counterparts; see refs 2c and 2d.
Org. Lett., Vol. 13, No. 12, 2011
3043