Enantioselective preparation of β,β-disubstituted α-methylenepropionates by MAO promotion of the zinc schlenk equilibrium

Article abstract of DOI:10.1002/anie.200463028

(Chemical Equation Presented) The zinc Schlenk equilibrium, little used since it was first described in 1966, has been promoted through the addition of methylaluminoxane (MAO) to maximize the yield of ZnR₂ from deleterious RZnCl by-products. This process allows an S_N2′- addition approach to the preparation of chiral β,β-disubstituted α-methylenepropionates with high enantioselectivity (see scheme).

Full text of DOI:10.1002/anie.200463028

Angewandte
Chemie
Asymmetric Synthesis
ant to further optimization. With high-throughput screening,
the chiral secondary amines 3 were identified as alternative
ligands for the transformation of 1a to 2a in the presence of
CuII (Table 2). A number of the other 80 species tried
produced significant ligand-accelerated catalysis (LAC)[6]
Enantioselective Preparation of b,b-Disubstituted
a-Methylenepropionates by MAO Promotion of
the Zinc Schlenk Equilibrium**
Table 1: Synthesis of chiral b,b-disubstituted a-methylenepropionates 2.
Paul J. Goldsmith, Simon J. Teat, and
Simon Woodward*
Dedicated to Dr. John M. Brown, FRS
on the occasion of his 65th birthday
In the copper-catalyzed reactions of orga-
nometallic nucleophiles (RM) with allylic
1
Ar
3
R1
R
2
Ar
1
2
=
a
b
c
d
e
f
Ph
a
b
c
d
e
f
g
h
i
Ph
Ph
Ph
H
a
b
c
d
e
f
Ph
halides R CH CR CH2X, two reaction
pathways are possible: either direct SN2
displacement at the a carbon or SN2’
4-(MeO)C6H4
4-MeC6H4
1-Naphthyl
2-BrC6H4
4-FC6H4
4-(CF3)C6H4
4-(NO2)C6H4
Me
CHO
H
H
H
H
H
H
4-(MeO)C6H4
4-MeC6H4
1-Naphthyl
2-BrC6H4
4-FC6H4
4-(CF3)C6H4
4-(NO2)C6H4
c-C6H11
1-Naphthyl
4-FC6H4
4-MeC6H4
4-(MeO)C6H4
3-(MeO)C6H4
displacement
g to the leaving group.
Recently, considerable progress has been
made towards the development of asym-
metric processes that result from copper-
catalyzed SN2’ addition of organozinc or
Grignard reagents.[1] These studies have
generally used allylic halides that lack
functional groups and have focused largely
on (E)-cinnamyl halides or equivalent
phosphate esters (R1 = Ph, R2 = H). Reac-
tions with ester-functionalized electron-
g
h
g
h
j
2-(MeO)C6H4
H
deficient allylic halides (R2 = CO2R, R1 = various, 1) would
offer rapid access to highly useful chiral b,b-disubstituted a-
methylenepropionates 2 (see Table 1). Highly selective addi-
tions of ZnR2 to the regioisomers of 1 have been demon-
strated,[2] but the former substrates themselves have proved
highly resistant to effective asymmetric catalysis.[3]
Stoichiometric groundwork by Xu and Kꢀndig showed the
viability of addition of organozinc reagents to halides of type
1.[4] Using Baylis–Hillman chemistry (to obtain suitable allylic
halides 1), we developed a first-generation catalytic system
for the addition of ZnR2 to allylic chloride 1a by employing a
chiral thioether ligand.[5] However, only low levels of
enantioselectivity were observed, and the system was resist-
Table 2: Addition of ZnEt2 to allylic chloride 1a using CuI and (S,S)-3.[a]
Entry
Ligand
R1
R
Yield 2a [%]
ee 2a [%]
1
2
3
4
5
6
7
8
3a
Ph
Ph
Ph
Ph
H
H
Me
CHO
H
H
H
H
H
50
50
10
18
41
18
28
60
70
69
56
53
65
51
<5
<5
<5
15
47
72
78
74
53
6
3a·HCl
3b·HCl
3c·HCl
3d·HCl
3e·HCl
3 f·HCl
3g·HCl
3h
c-C6H11
1-Naphthyl
4-FC6H4
4-MeC6H4
4-(MeO)C6H4
4-(MeO)C6H4
3-(MeO)C6H4
2-(MeO)C6H4
9
10
11
12
3h·HCl
3i·HCl
3j
H
H
H
[a] Addition of ZnEt2 (0.7 mL of a solution in toluene (1.1m), 0.77 mmol)
to 1a (0.5 mmol) in THF (1 mL) at ꢀ208C in the presence of (S,S)-3
(10 mol%) and CuI (5 mol%). Yield and ee values determined after 1 h
by chiral-GC analysis (see Supporting Information).
[*] P. J. Goldsmith, Dr. S. Woodward
School of Chemistry
The University of Nottingham
Nottingham NG72RD (UK)
Fax: (+44)115-951-3564
E-mail: simon.woodward@nottingham.ac.uk
effects (see Supporting Information), but only compound 3
gave ee values over 35%. Subsequent investigation revealed
that the presence of C2 symmetry is vital for efficient
stereoselection; for example, use of (S)-PhCH(Me)NHBn
gave 2a in high yield, but with less than 20% ee. The presence
of the secondary amino group is also crucial as either its
alkylation or formylation (Table 2, entries 3 and 4, respec-
tively) dramatically reduced the activity and selectivity of the
catalyst. Finally, it was noted that the presence of an aryl
function in the ligand 3 was also vital as 3d·HCl showed no
enantioselectivity and poor conversion of 1a (entry 5). The
Dr. S. J. Teat
Diamond Light Source Ltd.
Rutherford Appleton Laboratory
Chilton, Didcot
Oxfordshire OX110QX (UK)
[**] This work was supported in part by the European Union COST-D24
Program (W6-D24-0003) and by the EPSRC (Grant GR/S56054).
P.J.G. is also grateful for the support of the Grundy Educational
Trust. MAO=methylaluminoxane.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 2235 –2237
DOI: 10.1002/anie.200463028
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2235

Communications
role of the ligand aryl group was further investigated by
(1,2-dimethoxyethane) as solvent (with no additive) leads to a
significant increase in enantioselectivity (88% ee, 47% yield
of 2a at 1 h, ꢀ208C). A combination of the use of MAO and
DME and lowering the temperature led to the attainment of
synthetically useful levels of enantioselectivity in the prepa-
ration of 2a. Use of excess MAO led to ee values greater than
95%, but with low yields of 2a. These results could be
generalized across the range of substrates 1 (Table 3), with
increasing the range of substituents in 3 (Table 2). These were
mostly screened as their HCl salts as these are more
convenient to handle than the free amines. A strong
detrimental steric effect was noted for ligands with a 2-
substituent (entries 6 and 12), but a favorable electronic effect
(either + I or + M) existed at the 4-position (entries 8–10).
Through these studies the 4-methoxy ligand 3h was deter-
mined to be optimal (Table 2, entry 9).
These studies also highlighted that the free amines 3
outperformed their corresponding salts 3·HCl in terms of
product stereoselectivity (compare entries 1 and 9 versus 2
and 10, respectively; Table 2). Alongside this, kinetic studies
had shown that the enantioselectivity of the reaction in THF
was time-dependent: When substrate 1a was treated with
ZnEt2 using ligand 3h·HCl, the ee value fell from 80% at the
onset of the reaction to 74% after one hour (69% yield, the
mass balance being 1a). Coupled together these facts
indicated that EtZnCl, whose concentration increases as the
reaction evolves, impairs the selective transition state. This
could be confirmed by deliberate addition of EtZnCl
Table 3: CuTC/(S,S)-3h-catalyzed addition of ZnR2 (R=Et, Bu) to allylic
chlorides 1 under conditions that promote Schlenk equilibration.[a]
Entry Product Ar
Yield [%] [a]2D3[b] ee [%]
(stereoisomer)[b]
1
2
3
4
5
6
7
8
9
2a
2aa
2b
2c
2d
2e
2 f
Ph
Ph
92
81
ꢀ92.8 87 (R)
ꢀ76.2 78 (R)[c]
ꢀ96.4 90 (R)
ꢀ94.4 89 (R)
ꢀ63.3 86 (R)
ꢀ82.7 80 (S)[d]
ꢀ90.0 87 (R)
ꢀ80.3 83 (R)
ꢀ97.0 76 (R)
4-(MeO)C6H4 80
4-MeC6H4
1-Naphthyl
2-BrC6H4
4-FC6H4
4-(CF3)C6H4
95
53
95
95
93
2g
2h
[7]
4-(NO2)C6H4 95
(prepared from a 1:1 mixture of ZnEt2 and ZnCl2) which
led to an immediate and catastrophic collapse in the catalyst
selectivity (30% ee, 12% yield at 1 h,). We therefore reasoned
that if we could promote the forward reaction (k1) of the zinc
Schlenk equilibrium [Eq. (1)], the enantioselectivity of the
reaction should improve with the lower concentration of
EtZnCl. Conditions under which ZnCl2 is scavenged from the
reaction mixture should fulfil this requirement. After screen-
ing several additives, including crown ethers, we found that
the addition of methylaluminoxane (MAO; [-Al(Me)O-]n)
which is known to strongly scavenge halides[8] led to the
desired effect (83% ee, 48% yield of 2a at 1 h, ꢀ208C). It has
been speculated that MAO can act as a “latent Lewis acid”
and sequester chlorides as shown in structure 4. ZnMe2
[a] Addition of ZnEt2 (1.0 mL of a solution in toluene (1.1m), 1.1 mmol)
to 1 (0.5 mmol) in DME (1 mL) at ꢀ408C in the presence of (S,S)-3h
(10 mol%), CuTC (5 mol%), and MAO (0.5 mL, 15 wt% solution in
toluene). Yields of isolated products quoted. [b] (c 1.0, CHCl3) for [a]2D3 in
all cases. The absolute configuration of (ꢀ)-2a was confirmed crystallo-
graphically as R, and the products 2aa–h were assumed to show the
same sense of stereoinduction. [c] The reaction was performed using
Zn(nBu)2 (1.0 mL of a solution in heptane (1m), 1.0 mmol). [d] Yield
from NMR spectroscopic studies based on isolation of a mixture of 1e
and 2e. The absolute configuration has changed owing to a change in
assignment priority of the aryl substituent.
optimal results attained when CuTC (copper(i) thiophene-2-
carboxylate)[11] was used as the source of copper(i). The origin
of the enantioselectivity attained in this reaction unusually
depends critically on electronic effects. Electron-rich sub-
strates give products with higher enantioselectivities than
electron-poor substrates (compare entry 3 versus 9, Table 3),
which implicates a p-stacking interaction in the transition
state.
The sense of the enantioselectivity in the addition of
ZnEt2 to the halide 1a could be determined by X-ray
crystallographic analysis of the carboxylic acid 2ab, which
resulted from the hydrolysis of 2a, when crystallized as its
(R)-a-methylbenzylamine salt (Scheme 1). Single recrystalli-
zations of the acids derived from 2 and further derivatization
with a-methylbenzylamine improved the enantioselectivity to
> 99% ee for the catalytic addition. This enrichment was
considerably easier and more efficient than direct resolution
of racemic 2ab. The product 2ab of the SN2’ reaction is of
showed similar benefit, presumably through formation of
unreactive MeZnCl, though catalyst turnover was reduced
(84% ee, 35% yield of 2a at 1 h, ꢀ208C).[9]
k1
2 EtZnCl
ZnEt þ ZnCl
ð1Þ
G
H
2
2
k
ꢀ1
To the best of our knowledge,
attempted manipulation of the zinc
Schlenk equilibrium as a synthetic route
to alkyl ZnR2 species has not been
reported before.[10] Alongside these addi-
tive trials, stronger coordinating solvents
were screened, and we found that DME
Scheme 1.
2236
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 2235 –2237

Angewandte
Chemie
conceptual purposes only. For lead overviews, see: a) E. Zurek,
direct use in the preparation of combined ACE–NEP
(angiotensin-converting enzyme and neutral endopeptidase)
inhibitors.[12]
T. Ziegler, Prog. Polym. Sci. 2004, 29, 107 – 148; b) P. L. Bryant,
C. R. Harwell, A. A. Mrse, E. F. Emery, Z. Gan, T. Caldwell,
A. P. Reyes, P. Kuhns, D. W. Hoyt, L. S. Simeral, R. W. Hall,
L. G. Butler, J. Am. Chem. Soc. 2001, 123, 12009 – 12017; c) H.
Hanawa, N. Abe, K. Marouka, Tetrahedron Lett. 1999, 40, 5365 –
5368.
In conclusion, we have described the first method for the
copper-catalyzed enantioselective alkylation of Baylis–Hill-
man-derived electron-deficient allylic chlorides with organo-
zinc reagents thorough the use of a simple chiral secondary
amine and promotion by MAO of the zinc Schlenk equili-
brium. Further intensive studies are underway to identify the
nature of the p-stacking interactions responsible for the
ordered transition state.
[9] The preparation of PhZnEt (from ZnPh2 and ZnEt2) has been
reported by the groups of Fu and others, see: a) M. Fontes, X.
Verdaguer, L. Solꢃ, M. A. Pericꢃs, A. Riera, J. Org. Chem. 2004,
69, 2532 – 2543; b) J. Rudolph, N. Hermanns, C. Bolm, J. Org.
Chem. 2004, 69, 3997 – 4000; c) P. I. Dosa, G. C. Fu, J. Am. Chem.
Soc. 1998, 120, 445 – 446; d) P. I. Dosa, J. C. Ruble, G. C. Fu, J.
Org. Chem. 1997, 62, 444 – 445. We are unaware of any work on
the equivalent halide equilibrium.
[10] a) NMR spectroscopic evidence indicates a value of k1/kꢀ1
<
Experimental Section
0.002 under unpromoted conditions: D. F. Evans, G. V. Fazaker-
ley, J. Chem. Soc. A 1971, 182 – 183; b) Very recently, Knochel
has demonstrated exchange of ArZnI species: F. F. Kneisel, M.
Dochnahl, P. Knochel, Angew. Chem. 2004, 116, 1032 – 1036;
Angew. Chem. Int. Ed. 2004, 43, 1017 – 1021.
General procedure for the CuTC-catalyzed alkylation of allylic
chlorides 1: A dried Schlenk tube was charged with chloride 1
(0.50 mmol), CuTC (4.8 mg, 0.025 mmol), and (S,S)-3h (14.3 mg,
0.05 mmol). Dry DME (1 mL) was introduced, the stirred mixture
was cooled to ꢀ408C, and MAO (0.5 mL of a 15 wt% solution in
toluene; Aldrich) was added. The yellow reaction mixture was stirred
for 5 min at ꢀ408C, and then ZnR2 (for R = Et: 1 mL of a 1.1m
solution in toluene, 1.1 mmol; for R = nBu: 1 mL of a 1m solution
in heptane, 1.0 mmol) was added. Stirring was continued for 45 h at
ꢀ408C, then the reaction was quenched by cautious addition of 2m
HCl (2 mL). The aqueous layer was extracted with Et2O (2 ꢁ 5 mL),
then the organic extracts were dried (MgSO4), and the solvent was
evaporated. The products were isolated by flash chromatography
using mixtures of Et2O/hexanes as eluent.
[11] G. D. Allred, L. S. Liebeskind, J. Am. Chem. Soc. 1996, 118,
2748 – 2749.
[12] N. Inguimbert, P. Coric, H. Poras, H. Meudal, F. Teffot, M.-C.
Fournie-Zaluski, B. P. Roques, J. Med. Chem. 2002, 45, 1477 –
1486.
Received: December 22, 2004
Published online: March 10, 2005
Keywords: allylation · asymmetric catalysis · copper · zinc
.
[1] a) K. Tissot-Croset, D. Polet, A. Alexakis, Angew. Chem. 2004,
116, 2480 – 2482; Angew. Chem. Int. Ed. 2004, 43, 2426 – 2428;
b) A. O. Larsen, W. Leu, C. N. Oberhuber, J. E. Campbell, A. H.
Hoveyda, J. Am. Chem. Soc. 2004, 126, 11130 – 11131; c) U.
Piarulli, P. Daubos, C. Claveric, M. Roux, C. Genari, Angew.
Chem. 2003, 115, 244 – 246; Angew. Chem. Int. Ed. 2003, 42, 234 –
236; d) W.-J. Shi, L.-X. Wang, Y. Fu, S.-F. Zhou, Q.-L. Zhou,
Tetrahedron: Asymmetry 2003, 14, 3867 – 3872; e) H. Malde,
A. W. van Zijl, L. A. Arnold, B. L. Feringa, Org. Lett. 2001, 3,
1169 – 1171; f) F. Dubner, P. Knochel, Tetrahedron Lett. 2000, 41,
9233 – 9237.
[2] a) K. E. Murphy, A. H. Hoveyda, J. Am. Chem. Soc. 2003, 125,
4690 – 4691; b) M. A. Kacrpzynski, A. H. Hoveyda, J. Am.
Chem. Soc. 2004, 126, 10676 – 10681.
[3] The addition of Zn(CH2tBu)2 to 1a (62% yield, 12% ee, 6.8:1
SN2’/SN2) has been reported in a patent: P. Knochel, F. Dꢀbner
(Avecia Ltd., UK), PCT Int. Appl. WO2000012449, 2000 [Chem.
Abstr. 2000, 132, 222063].
[4] L.-H. Xu, E. P. Kꢀndig, Helv. Chim. Acta 1994, 77, 1480 – 1484.
[5] C. Bꢂrner, P. J. Goldsmith, S. Woodward, J. Gimeno, S. Gladiali,
D. Ramazzotti, Chem. Commun. 2000, 2433 – 2434.
[6] D. J. Berrisford, C. Bolm, K. B. Sharpless, Angew. Chem. 1995,
107, 1159 – 1171; Angew. Chem. Int. Ed. Engl. 1995, 34, 1059 –
1070.
[7] a) J. Boersma, J. G. Noltes, Tetrahedron Lett. 1966, 1521 – 1525;
The limited data that exists suggests that under MAO-free
conditions k1/kꢀ1 < 0.01, see: b) D. F. Evans, I. Wharf, J. Chem.
Soc. A 1968, 4, 783 – 787.
[8] The structure of MAO is not known (linear, ring, and cage
structures have all been proposed). Structure 4 is given for
Angew. Chem. Int. Ed. 2005, 44, 2235 –2237
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2237