Copper catalyzed enantioselective allylic substitution by MeMgX

Article abstract of DOI:10.1016/j.tetlet.2004.07.114

Methyl Grignard undergoes highly regio (>90/10) and enantioselective (ee 91-96%) copper catalyzed allylic substitution on cinnamyl-type chlorides. CuBr (3%) and 3.3% of a chiral phosphoramidite ligand are sufficient for a complete reaction. The synthesis of a precursor of (+)-Naproxen is described. The reaction can be extended to alkyl substituted allylic chlorides (ee 72%).

Full text of DOI:10.1016/j.tetlet.2004.07.114

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 7375–7378
Copper catalyzed enantioselective allylic substitution by MeMgX
Karine Tissot-Croset and Alexandre Alexakis*
Department of Organic Chemistry, University of Geneva, 30, quai Ernest Ansermet, CH-1211 Geneva 4, Switzerland
Received 3 July 2004; revised 18 July 2004; accepted 24 July 2004
Available online 24 August 2004
Abstract—Methyl Grignard undergoes highly regio (>90/10) and enantioselective (ee 91–96%) copper catalyzed allylic substitution
on cinnamyl-type chlorides. 3% of CuBr and 3.3% of a chiral phosphoramidite ligand are sufficient for a complete reaction. The
synthesis of a precursor of (+)-Naproxen is described. The reaction can be extended to alkyl substituted allylic chlorides (ee 72%).
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Grignard or dialkylzinc reagents and 1% Cu-thiophene
carboxylate (and 1.1% L*).3e However, in the case of
MeMgBr, the regioselectivity was in favor of the unde-
sired achiral a-product, although the enantioselectivity
of the c-product was excellent (Scheme 2).
Among the metal-catalyzed allylic substitution reac-
tions,1 copper plays a prominent role due, not only to
the high regio and stereoselectivity it provides, but also
because it allows harder nucleophiles, such as Grignard
or organozinc reagents, to be used.2 Thus, simple alkyl
groups can only be introduced with copper catalysis.
The enantioselective version has found strong interest
in the last 4–5years, both with Grignard3 or diorgano-
zinc reagents.3e,4 High regio and enantioselectivities
could be attained, with a variety of substrates, leaving
groups (LG) or RMgX, and R2Zn groups.
R-MgBr
1% CuTC
1.1% L*
R
+
Ph
Cl
Ph
R
Ph
CH2Cl2
-78°C
R = Me γ/α 36/64 ee 94%
Et 99/1 96%
OMe
The only difficult groupto introduce is the methyl
group, though it is the most valuable, from a synthetic
point of view. The difficulty lies either on the poor enan-
tioselectivity,5 or the low reactivity6 or the bad regio-
selectivity.7 (Scheme 1)
O
O
P
N
OMe
Scheme 2. Enantioselective copper catalyzed allylic substitution.
Me-M
Me
*
+
R
LG
R
Me
R
CuX catalyst
We report herein our successful efforts to improve this
regioselectivity and render this transformation syntheti-
cally useful.
γ-attack
α-attack
Scheme 1. Copper catalyzed allylic substitution.
We have recently disclosed a new phosphoramidite lig-
and, which affords the highest reported enantioselectivi-
ties (91–96%) for cinnamyl-type substrates, using either
2. Results
The regioselectivity of the copper catalyzed allylic sub-
stitution has been extensively studied.2,8 It highly
depends on the copper source, its amount, the solvent,
and various additives, such as BF3ꢀOEt2.9 It is com-
monly believed that copper reagents favor the SN20
adduct, whereas cuprate reagents are not selective, the
Keywords: Asymmetry; Copper; Catalysis; Grignard reagent; Allylic
substitution; Naproxen.
*
Corresponding author. Tel.: +41 22 3796522; fax: +41 22 3793215;
e-mail: alexandre.alexakis@chiorg.unige.ch
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.07.114

7376
K. Tissot-Croset, A. Alexakis / Tetrahedron Letters 45 (2004) 7375–7378
Table 3. Influence of the amount of CuBr catalyst
Table 1. Influence of the copper salta
Entry
Copper salt
Conv.
c/a
ee (%)
Entry
% of Catalyst
Conv.
c/a
ee (%)
1
2
3
4
5
6
CuTCb
CuCl
CuBr
100
93
49/51
60/40
54/46
49/51
52/48
45/55
95
95
95
94
94
46
1
2
3
4
5
1
2
3
4
5
100
100
100
100
100
83/17
87/13
89/11
89/11
89/11
96
96
96
96
95
100
92
CuI
CuBr, Me2S
CuCN
100
10
a Slow addition (2h) at À78ꢁC, in CH2Cl2.
b CuTC stands for copper thiophene carboxylate.
of the undesirable cuprate intermediate. The optimum
value was found to be 3% of CuBr (and 3.3% of chiral
ligand).
Table 2. Influence of the rate of additiona
With the optimized conditions in hand, we screened sev-
eral substrates to generalize the scope of the reaction
(Scheme 3 and Table 4).
Entry
Copper salt
Add. time
Conv.
c/a
ee (%)
1
2
3
4
5
6
CuTCb
CuBr
40min
40min
2h
100
100
100
100
100
100
36/64
38/62
49/51
54/46
81/19
83/17
94
94
95
95
95
96
CuTCb
CuBr
The enantioselectivity of the reaction is always excellent
(eeÕs > 90%), whatever the electron demand of the sub-
2h
4h
CuTCb
CuBr
4h
a Addition at À78ꢁC, in CH2Cl2.
Me-MgBr
3% CuBr
b CuTC stands for copper thiophene carboxylate.
Me
3.3% L*
+
R
Cl
R
Me
R
CH2Cl2
4h, -78°C
formed product is at the least substituted carbon (in our
case, the adduct). Therefore, in a catalytic reaction the
rate of addition of the Grignard reagent is very impor-
tant. A fast addition allows the formation of a cuprate
intermediate. In contrast, a slow addition would prevent
the formation of a cuprate species by letting the organo-
copper intermediate react with the allylic substrate.8c
These parameters were investigated with cinnamyl chlo-
ride as typical substrate and the ligand shown in Scheme
2. Table 1 summarizes our results with various copper
sources, under the standard conditions, and Table 2
shows the results of different addition rates of MeMgBr.
γ-attack
α-attack
Scheme 3. Enantioselective copper catalyzed allylic substitution (Table
4).
Table 4. Enantioselective substitution on various allylic chlorides
Entry
1
R
Conv.
100
c/a
ee (%)
96
89/11
2
90
90/10
95
Cl
Cl
Although it is known that CuCN favors the SN20 pro-
cess,2c,8,10 it was surprising to see that the regioselectivity
was not improved (Table 1, entry 6). Instead, both the
conversion and the enantioselectivity were much lower
than with the other copper salts. It was also interesting
to observe that all the other copper salts gave roughly
similar results, with always good enantioselectivity.
For this reason we selected CuBr as another copper
source (much cheaper), along with CuTC, which was
our favorite until now.3d,e
3
4
5
94
100
100
90/10
90/10
91/9
94
95
91
Cl
Me
MeO
MeO
6
7
100
100
83/17
84/16
90
66
The influence of the addition rate of the Grignard rea-
gent was more instructive (Table 2). Indeed, as expected,
the slower the addition the better was the regioselectiv-
ity. An addition over 2h already brings some improve-
ment (compare entries 3/4 and 1/2), and a syringe
pump addition over 4h gave an acceptable c/a ratio of
83/17, with an excellent enantioselectivity of 96%.
OMe
8
9
100
100
67/33
84/16
92
93
F3C
A last parameter to look at was the amount of catalyst.
Table 3 shows the results obtained under the experimen-
tal conditions stated in entry 6 (Table 2).
O
O
10
11
100
90
90/10
96/4
91
72
The amount of catalyst plays an important role, as it
accelerates the reaction, thus preventing the formation

K. Tissot-Croset, A. Alexakis / Tetrahedron Letters 45 (2004) 7375–7378
7377
use, A. S. E, Ed.; Wiley–VCH: Weinheim, 2001, pp 188–
223; (c) Karlstro¨m, A. S. E.; Ba¨ckvall, J.-E. In Modern
Organocopper Chemistry; Krause, N., Ed.; Wiley–VCH:
Weinheim, 2001, pp 259–288.
stituent on the aromatic group. Both electron withdraw-
ing (entries 2 and 3) or electron-donating group(entries
4 and 5) are tolerated. Only the ortho-methoxy group
(entry 7) seems to interfere, perhaps via a coordination
to the metal, and the ee drops to 66%. The regioselectiv-
ity, although not as high as with normal alkyl Grig-
nard,3e is still at acceptable levels with c-selectivities
ranging from 83% to 91%. The reaction has also been
extended to a nonaromatic group, with a cyclohexyl
substituent (entry 11). The regioselectivity is remarkably
high, although the enantioselectivity is lower than with
the other aromatic substrates.
3. (a) van Klaveren, M.; Persson, E. S. M.; del Villar, A.;
Grove, D. M.; Ba¨ckvall, J.-E.; van Koten, G. Tetrahedron
Lett. 1995, 36, 3059–3062; (b) Karlstrom, A. S. E.; Huerta,
F. F.; Meuzelaar, G. J.; Ba¨ckvall, J.-E. Synlett 2001, 923–
926; (c) Alexakis, A.; Malan, C.; Lea, L.; Benhaim, C.;
Fournioux, X. Synlett 2001, 927–930; (d) Alexakis, A.;
Croset, K. Org. Lett. 2002, 4, 4147–4149; (e) Tissot-
Croset, K.; Polet, D.; Alexakis, A. Angew. Chem., Int. Ed.
2004, 43, 2426–2428.
4. (a) Dubner, F.; Knochel, P. Angew. Chem., Int. Ed. 1999,
¨
¨
38, 379–381; (b) Dubner, F.; Knochel, P.; PCT Int. Appl.;
An interesting simple application was to apply this
method to a formal synthesis of (+)-Naproxen, a well-
known nonsteroidal anti-inflammatory drug (Scheme
4).11 The direct precursor to Naproxen could be
obtained under the established standard conditions.
However, we were not satisfied with the regioselectivity
for such a valuable compound. A simple increase of the
catalyst, from 3% to 5%, allowed the desired improve-
ment, with a 90% c-selectivity and 93% ee.
Avecia Limited, UK, 9 March 2000; Chem. Abstr. 2000,
132, 222063; (c) Bo¨rner, C.; Gimeno, J.; Gladiali, S.;
Goldsmith, P. J.; Ramazzotti, D.; Woodward, S. Chem.
Commun. 2000, 2433–2434; (d) Luchaco-Cullis, C. A.;
Mizutani, H.; Murphy, K. E.; Hoveyda, A. H. Angew.
Chem., Int. Ed. 2001, 40, 1456–1460; (e) Malda, H.; van
Zijl, A. W.; Arnold, L. A.; Feringa, B. L. Org. Lett. 2001,
3, 1169–1171; (f) Piarulli, U.; Daubos, P.; Claverie, C.;
Roux, M.; Gennari, C. Angew. Chem., Int. Ed. 2003, 42,
234–236; (g) Murphy, K. E.; Hoveyda, A. H. J. Am.
Chem. Soc. 2003, 125, 4690–4691; (h) Shi, W.-J.; Wang,
L.-X.; Fu, Y.; Zhu, S.-F.; Zhou, Q.-L. Tetrahedron:
Asymmetry 2003, 14, 3867–3872; (i) Piarulli, U.; Claverie,
C.; Daubos, P.; Gennari, C.; Minnaard, A.; Feringa, B. L.
Org. Lett. 2003, 5, 4493–4496; (j) Van Zijl, A. W.; Arnold,
L. A.; Minnaard, A. J.; Feringa, B. L. Adv. Synth. Catal.
2004, 346, 413–420.
Me-MgBr
x% CuBr, x1.1% L*
Cl
CH2Cl2
MeO
-78˚C
5. A 10% ee is reported for dimethylzinc and cinnamyl
chloride. See Ref. 4b.
Me
6. Me2Zn (5–6equiv) are needed for reasonable conversion,
otherwise poor yields are reported. See Refs. 4f,g,i.
7. c/a Ratio as low as 31/69, and 12/88, were reported in Ref.
3d.
MeO
8. (a) Tseng, C. C.; Paisley, S. D.; Goering, H. L. J. Org.
Chem. 1986, 51, 2884–2891; (b) Tseng, C. C.; Yen, S.-J.;
Goering, H. L. J. Org. Chem. 1986, 51, 2892–2895, and
% of catalyst = 3% γ/α 84/16 ee 92%
5%
90/10
93%
´
references cited therein; (c) Ba¨ckvall, J. E.; Sellen, M.;
Scheme 4. Synthesis of the precursor of Naproxen.
Grant, B. J. Am. Chem. Soc. 1990, 112, 3321–6615, and
references cited therein; (d) Nakamura, E.; Mori, S.
Angew. Chem., Int. Ed. 2000, 39, 3750–3771, and refer-
ences cited therein.
3. Conclusion
9. Yamamoto, Y. Angew. Chem., Int. Ed. 1986, 25, 947–1038,
and references cited therein.
10. Levisalles, J.; Rudler-Chauvin, M.; Rudler, H. J. Organo-
met. Chem. 1977, 136, 103–110.
11. For a review on Naproxen and congeners see: Rieu, J.-P.;
Boucherle, A.; Cousse, H.; Mouzin, G. Tetrahedron 1986,
15, 4095–4131.
In summary, we have disclosed an efficient catalytic sys-
tem for the asymmetric transfer of the methyl group, the
most synthetically valuable.12 The amount of the metal
catalyst and the chiral ligand are still at low enough val-
ues for copper catalysis.
12. Typical procedure (for 1mmol): A dried Schlenk tube was
charged with copper salt (3mol%) and the chiral ligand
(3.3mol%). Dichloromethane (3mL) was added and the
mixture was stirred at room temperature for 30min. The
allylic chloride (1mmol) was introduced dropwise and the
reaction mixture was stirred at room temperature for a
further 5min before being cooled to À78ꢁC in an ethanol-
dry ice cold bath. MeMgBr (3M in diethyl ether,
1.2equiv), diluted in dichloromethane (0.6mL) was added
over 4h via a syringe pump. Once the addition was
complete the reaction mixture was left at À78ꢁC for a
further 12h, at which point gas chromatography of an
aliquot showed that all the starting material had been
converted. The reaction mixture was quenched by addition
of aqueous hydrochloric acid (1N, 2mL). Diethyl ether
(10mL) was added and the aqueous phase was separated
Acknowledgements
We thank the Swiss National Science Foundation (grant
No. 20-068095.02) and COST action D24/0003/01
(OFES contract No. C02.0027) for financial support.
References and notes
1. (a) Magid, R. M. Tetrahedron 1980, 36, 1901–1930; (b)
Trost, B. M.; van Vranken, D. L. Chem. Rev. 1996, 96,
395–422.
2. (a) Posner, G. H. Org. Rect. 1974, 22, 253–400; (b) Breit,
B.; Demel, P. In Modern Organocopper Chemistry; Kra-

7378
K. Tissot-Croset, A. Alexakis / Tetrahedron Letters 45 (2004) 7375–7378
and extracted further with diethyl ether (3 · 3mL). The
oxy)-naphthyl)-1-butene: 1H NMR (400MHz, CDCl3):
7.74–7.17 (m, 6H, ArH), 6.13 (m, 1H, CH2@CH), 5.12 (m,
2H, CH2@CH), 3.96 (s, 3H, OCH3), 3.65 (quint,
J = 6.82Hz, 1H, CHPh), 1.49 (d, J = 7.08Hz, 3H, CH3).
13C NMR (100MHz, CDCl3): 157.3, 143.4, 140.7, 133.2,
129.2, 129.1, 126.9, 126.8, 125.1, 118.7, 113.3, 105.6, 55.3,
combined organic fractions were washed with brine
(5mL), dried over anhydrous sodium sulfate, filtered,
and concentrated in vacuo. The oily residue was purified
by flash column chromatography (eluent = pentane) to
yield the product as a mixture of SN20 and SN2 regioiso-
mers. Gas chromatography or supercritical fluid chroma-
20
43.1, 20.7. ½aꢁD +15.7 (c 1.4, CHCl3) for 93% ee. Ee was
tography on
a
chiral stationary phase showed the
measured by chiral SFC with a Chiralcel OB-H column
(5% MeOH, flow rate 2mL/min) tR: 5.91 (S), 6.48 (R).
enantiomeric excess of SN20 product. (+)-3(S)-(2-(6-Meth-