Catalytic enantioselective allylation at the activated benzylic position of prochiral aryl cyano esters: Access to quaternary stereogenic centers

Article abstract of DOI:10.1016/j.tetlet.2005.01.082

Palladium catalyzed asymmetric allylic alkylation of prochiral aryl cyano esters has been carried out in the presence of various chiral ligands. The base and additives have been varied and allowed to produce the allyl derivative presenting a highly functi

Full text of DOI:10.1016/j.tetlet.2005.01.082

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 1617–1621
Catalytic enantioselective allylation at the activated
benzylic position of prochiral aryl cyano esters: access
to quaternary stereogenic centers
*
´
Audrey Nowicki, Andre Mortreux and Francine Agbossou-Niedercorn
Laboratoire de Catalyse de Lille, UMR CNRS 8010, ENSCL, BP 108, 59652 Villeneuve d’Ascq Cedex, France
Received 7 December 2004; revised 6 January 2005; accepted 18 January 2005
Available online 1 February 2005
Abstract—Palladium catalyzed asymmetric allylic alkylation of prochiral aryl cyano esters has been carried out in the presence
of various chiral ligands. The base and additives have been varied and allowed to produce the allyl derivative presenting a highly
functionalized quaternary stereogenic center. The chiral pocket ligands of Trost appears the most appropriate to produce the
desired chiral derivative.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Catalytic enantioselective allylation via a chiral p-allyl
palladium(II) complex has been quite intensively investi-
gated and remarkable success has been reported.1 Most
of the examples describe the successful introduction of
the chirality into the allylic substrate.1
interact with the prochiral enolate while approaching
the p-allyl carbon on the p-allyl palladium inter-
mediate.6,12
The chiral pocket ligands of Trost have been used suc-
cessfully in the allylation of b-ketoesters13 and b-di-
ketones.14 Recently, advances have been made in the
asymmetric allylation of nonstabilized ketone enolates
assisted by palladium catalysts bearing chiral pocket
type ligands.15,16 In these cases, a-substituted cyclohexan-
ones were the substrates.
In contrast, the enantioselective nucleophilic attack of a
stabilized prochiral carbon nucleophile to the p-allyl–
Pd(II) complex, which is occurring at the side opposite
to the chiral auxiliary, has been fairly less explored
and is not easy to control.2–7 Nonetheless, such a pro-
cess is a valuable and challenging tool for the construc-
tion of quaternary stereocenters8 and has been applied
with success in total synthesis.9
We sought to prepare chirons with highly functionalized
quaternary stereogenic centers via the allylation of cya-
no ester type substrates. To the best of our knowledge,
only one report describes the palladium assisted allyl-
ation of closely related substrates, that is, a-isocyano-
carboxylates, which has been carried out in the
presence of a chiral ferrocenylphosphine ligand bearing
an additional hydroxy function.17 In the best case, the
combination of a base (DBU) and a Lewis acid (ZnCl2)
allowed the formation of the allyl product with 39% ee.
Here, we report on our first results of the allylation of
3,4-dichlorophenyl cyano ester derivatives.
Efficient catalytic systems, which are able to induce high
enantiodifferentiation in such type of allylations, have
been intented for specific substrates. They consist, for
example, in using a bimetallic catalytic system for the
allylation of a chiral rhodium(I) enolate of a-cyanopro-
pionates.4 Another approach is the allylation of a-acet-
amido-b-ketoesters10 or iminoesters in order to produce
new amino acid derivatives.11 Earlier, the allylation of b-
diketones has been carried out with chiral ferrocenyl-
phosphine ligands bearing a functional group able to
The cyano ester 1 has been synthesized quantitatively
through reaction of 3,4-dichloro benzonitrile with di-
methylcarbonate in toluene in the presence of sodium
methoxide (Scheme 1).18 This substrate was then subject
to allylation via asymmetric allylic substitution furnish-
ing the allyl product 2 (Scheme 2).19,20 Four chiral
Keywords: Asymmetric allylation; Palladium; Quaternary centers;
Asymmetric catalysis.
*
Corresponding author. Tel.: +33 3 2043 49 27; fax: +33 3 2043 65
85; e-mail: francine.agbossou@ensc-lille.fr
0040-4039/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.01.082

1618
A. Nowicki et al. / Tetrahedron Letters 46 (2005) 1617–1621
have a notable impact on the enantioselectivity. For
NC
NC
CO Me
2
example, while changing the cation from Li (n-BuLi)
(entry 7) to Na (NaH) (entry 8) and to K (KH) (entry
9), the ee was increasing from 25% to 43%
(Dee = +18%). This significant ee increase can be related
to the size of the cation associated to the enolate which
will react with the allyl–palladium catalytic intermedi-
ate. Thus, the potassium enolate is giving the best result.
With the base LDA (entry 10), an intermediate value of
38% ee is measured. This can be attributed to the amine
produced during the deprotonation of the pronucleo-
phile, which is able to remain associated to the lithium
of the enolate. In consequence, an increase of the steric
hindrance of the nucleophile, which is approaching the
allylic moiety located within the chiral pocket furnished
by the chiral ligand, will lead to an increase of the selec-
tivity when compared to a run where n-Buli is used alone
(Dee = +13%) (entry 7). The steric interactions between
the nucleophile and the chiral ligand have thus to be im-
proved in order to enhance the enantioselectivity. This
subtle steric influence can also be apparent in the pres-
ence of sparteine. Indeed, while associating (ꢀ)-sparte-
ine to n-BuLi, a significant improvement of the
selectivity is observed (Dee = +29%) (entry 11). Very
interestingly, the use of (ꢀ)-sparteine alone is providing
the allylated product with 40% ee (entry 12). The same
reaction carried out at room temperature allowed to
reach an asymmetric induction of 55% (entry 13). This
result is most likely associated to the presence of the
protonated sparteine in the environment of the nucleo-
phile while approaching the allylic moiety than to a
selective deprotonation of the pronucleophile. Interest-
ingly, while combining the (S,S) standard Trost ligand
L6 with (ꢀ)-sparteine, still, the allyl product is isolated
with 40% ee but with the opposite configuration (entry
14). This indicates clearly that the structure of the ion
pair constituting the nucleophile has an important im-
pact on the selectivity of the process and has to be taken
into consideration more than the structure of the
enolate.15
OC(OMe)
NaOMe
toluene
reflux
2
Cl
Cl
1
Cl
Cl
Scheme 1.
NC
COOR
NC
COOR
"Pd-ligand"
X
+
Cl
THF
Base
Cl
additive
Cl
Cl
X=OAc, Cl, OCOCF
R= Me, iPr
3
Scheme 2.
auxiliaries, that is, a phosphinooxazoline (L1),21 BINAP
(L2)22 and the two chiral pocket ligands of Trost (L3 and
L4)23 were available commercially (Scheme 3). The func-
tionalized ferrocenylphosphine (L5) (Scheme 3) has been
prepared according to the reported procedure.6
For our first attempts, we considered the reaction of the
cyano ester 1 with allylacetate using palladium complex
[PdCl(g3-C3H5)]2 associated to chiral ligands L1, L2, and
L5 in the presence of BSA/KOAc (cat),24 LDA or
25
ZnEt2 at room temperature. The highest ee measured
was 20% and selected examples are reported in Table 1
(entries 1–3). The beneficial effect mentioned above for
the allylation of isocyano esters type substrates while
using the DBU/ZnCl2 combination instead of DBU
alone has not been observed here for substrate 1 (entries
4 and 5).17 Rather, the same level of enantioselectivity
was observed. We turned next our attention to the use
of the standard chiral pocket ligand of Trost L3. In
the presence of a BSA/KOAc (cat) mixture, an enantio-
meric excess of 45% was achieved in THF at room
temperature (entry 6).
Zinc enolates can be generated either in the presence of
25
ZnEt2 or by a Li/Zn exchange process taking place
Then, reaction optimization was attempted essentially
by varying the solvent and the base. The solvent had a
moderate effect on the conversion and the enantio-
selectivity of the process. Several solvents have been
examined (CH2Cl2, THF, toluene, DME) with THF
giving the best results.
while using n-BuLi in association with ZnCl2. The
outcome of the reaction involving such enolates clearly
demonstrates the beneficial effect brought by the use of
the latter as 50–52% ee are reached (entries 15 and
16). This is corresponding to an increase of Dee =
+28% compared to the asymmetric induction obtained
with the lithium enolate alone (entry 7). Generally, the
reactions are fast and a total conversion is reached with-
in nonoptimized 30min at ꢀ78 ꢁC and with a substrate/
catalyst ratio of 150.
The base and the presence of additives has a stronger ef-
fect on the selectivity. As a matter of fact, the overall
structure of the enolate and the size of the counterion
H
CH
N
3
OH
O
O
O
O
O
O
O
PPh
PPh
2
2
NH
NH
HN
HN
NH
HN
Fe
PPh
N
PPh CH
2
2
3
PPh
2
PPh Ph P
PPh Ph P
PPh
Ph P
2
2
2
2
2
2
L5
L2
L3
L4
L1
L6
Scheme 3.

A. Nowicki et al. / Tetrahedron Letters 46 (2005) 1617–1621
Table 1. Palladium catalyzed asymmetric allylic alkylation of 1 into 2a
1619
Entry
Allylic derivative
Allyl–OAc
Ligand
Baseb
Additivec
T (ꢁC)d
Time (h)e
Conv (%)f
Ee (%)g
1
L1
L2
L5
L5
L5
L3
L3
L3
L3
L3
L3
L3
L3
L6
L3
L3
L3
L3
L3
L3
L4
L4
L3
L3
L3
BSA
KOAc (0.1)
—
rt
0.5
4
100
100
100
100
100
90
5
18
20
15
17
45
2
ZnEt2
LDA
DBU
rt
3
—
—
rt
1.5
1.5
1.5
1
4
rt
5
BDU
BSA
ZnCl2
KOAc (0.1)
rt
6
rt
7
n-BuLi
NaH
KH
—
—
ꢀ78
ꢀ78 to rt
ꢀ78 to rt
ꢀ78 to rt
ꢀ78
ꢀ78
rt
1.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
1.5
0.5
1.5
1
9025
100
57
8
39
43
38
54
40
55
40
50
52
26
36
9
—
—
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
LDA
98
n-BuLi
(ꢀ)-Sparteine
(ꢀ)-Sparteine
(ꢀ)-Sparteine
ZnEt2
n-BuLi
DBU
(ꢀ)-Sparteine
100
100
100
100
100
100
100
100
8025
8050
9032
100
100
100
100
—
—
—
—
ꢀ78
ꢀ78
ꢀ78 to rt
ꢀ78
ꢀ78
ꢀ78
rt
ZnCl2
—
DBU
—
ZnCl2
—
—
—
—
1
—
—
rt
1
—
BSA
ꢀ78
rt
0.5
0.5
1.5
0.5
15
34
41
48
Allyl–Cl
Allyl–OCOCF3
Allyl–OCOCF3
KOAc
—
—
—
(ꢀ)-Sparteine
ꢀ78
ꢀ78
a The catalytic reactions were carried out in THF and with ligand/Pd/substrate ratio = 1.2/1/150in the conditions mentioned in the table.
b Base: 1.2 equiv/substrate.
c The amount of additive is relative to the substrate.
d For reactions performed between ꢀ78 ꢁC and room temperature (rt), the cold bath is removed right after addition of the catalyst.
e The time was not optimized.
f Conversions were determined by 1H NMR on the crude reaction mixture through integration of the methyl residue.
g Determined by 13C NMR spectroscopy using the [Eu(hfc)] shift reagent. The absolute configuration has not been determined. The sign of the
rotation is (ꢀ) except for entry 14 where it is (+).
A beneficial effect is observed for the use of DBU/ZnCl2
instead of DBU alone (Dee = +10% ) (entries 17 and 18)
at low temperature with a total conversion.17
(entries 19–21). This can be attributed to either a modi-
fication or a decomposition of the catalyst. Nevertheless,
such a behaviour is not observed with the isopropyl
based cyano ester27. Indeed, at room temperature the
allylation in the presence of allylacetate and the auxil-
iary L3 without a base and in 30min, the allyl product
is obtained with a total conversion and in 43% ee. Thus,
the steric hindrance of the ester did not bring any
improvement of the enantioselectivity. Rather, the iso-
propyl ester substrate is allylated with a lower selectivity
than 1 (Dee = ꢀ7%). Thus, the asymmetric inductions
are higher when using substrate 1 and ligand L3 (entries
19 and 20).
The reaction is also feasible without a base (entries 19–
22). Indeed, the substrate is allylated in neutral condi-
tions at low as well as at room temperature. It can be
suggested that the leaving group (OAc) displaced from
the allylic acetate reactant during the ionization step is
capable of acting as a deprotonating agent toward the
pronucleophile. Giambastiani and Poli have examined
very closely such base-free palladium catalyzed alkyl-
ations and showed clearly that neutral conditions are
not limited to allylic carbonates, phenates, and epoxides
but can also apply to allylic acetate based allylations as
long as the pronucleophile presents the suitable pKa.26
Such neutral conditions are attractive from an atom
economy standpoint as no stoichiometric amount of
base is necessary. It has been noticed that the base-free
reaction carried out at room temperature is providing
a better selectivity (50% ee) that at low temperature
(for ligand L3, Dee = +25%) (entries 20and 19). While
using the sterically more hindered Trost ligand L4, no
improvement of the enantioselectivity was observed.
Rather, at low temperature, a decrease to 15% ee is ob-
served (entry 22) when compared to the standard ligand
L3 (entry 20) (Dee = ꢀ10%). Some reactions carried out
on 1 at room temperature are not going to completion
Two other sources of allylic moiety, that is, allylchloride
and allyl trifluoroacetate gave close results (entries 23–
25).
In summary, we have shown that Pd-catalyzed allylic
substitution with arylcyanoesters pronucleophiles can
provide allylated derivatives with interesting levels of
enantioselectivity. The actual structure of the enolate in-
volved in the addition process is not easy to define but
has a critical impact onto the selectivity. Several impor-
tant points have to be taken into consideration in order
to optimize the catalytic system. Beside the structure of
the substrate itself, the base, eventual additives able to
modify the overall steric hindrance of the enolate, and

1620
A. Nowicki et al. / Tetrahedron Letters 46 (2005) 1617–1621
16. You, S.-L.; Hou, X.-L.; Zhu, X.-Z. Org. Lett. 2001, 3, 149.
17. Ito, Y.; Sawamura, M.; Matsuko, M.; Matsumoto, Y.;
Hayashi, T. Tetrahedron Lett. 1987, 28, 4849.
the chiral auxiliary have to be examined very closely. In
our case, the noncyclic nature of the substrate is most
probably responsible for the level of enantioselectivity
observed. Indeed, similar results are often obtained with
pronucleophile leading to enolated which are not part of
a cyclic structure.10,16,28 Also, among the panel of chiral
auxiliaries available, the chiral pocket ligand of Trost
type appears the most appropriate for the targeted
transformation. A further improvement of the selectivity
of the allylation of 1 will certainly require the prepara-
tion of even more specific chiral auxiliaries exhibiting
the chiral pocket concept. Chiral alkylated cyano esters
presenting a quaternary optically pure stereogenic center
are valuable synthons for asymmetric synthesis.29 Stud-
ies in which the strategy described here is applied to the
synthesis of a bio-active molecule will be reported soon.
18. Compound 1: 1H NMR (CDCl3): d 3.82 (s, 3H, CH3), 4.69
(s, 1H, CHCN), 7.31 (dd, J = 2.3 and 8.3 Hz, 1H, HAr),
7.50(d, J = 8.3 Hz, HAr), 7.56 (d, J = 2.3 Hz, 1H, HAr).
13C NMR (CDCl3): d 43 (CHCN), 54 (OCH3), 115 (CN),
127, 129, 130, 131, 133, and 134 (CAr), 164,6 (CO). Anal.
Calcd for C10H7Cl2NO2: C, 49.20; H, 2.89; N, 5.74.
Found: C, 49.18; H, 2.81; N, 5.67.
19. Representative experimental procedure for the allylation
reaction: In a Schlenk tube equipped with a stir bar, the
substrate (1 mmol) is dissolved in freshly distilled THF
(4 mL) and the base is added (1.2 mmol). In a second
Schlenk tube equipped with a stir bar, the palladium
complex [Pd(allyl)Cl]2 (3.7 mg, 0.01 mmol) and the chiral
auxiliary (0.021 mmol) are introduced with THF (2 mL)
followed by the addition of allyl derivative (1.5 mmol).
The mixture is stirred for 15 min and then transferred via
cannula onto the substrate. The catalytic medium is stirred
at the indicated temperature and the evolution of the
reaction is followed through GC or TLC analysis of
aliquots taken from the reaction mixture. At the end of the
reaction, a saturated solution of NH4Cl (5 mL) is added
and the medium stirred for 5 min before addition of water
(10mL). The allylated product is extracted with ethyl
acetate (3 · 15 mL). The organic phase is washed with a
solution of saturated NaHCO3 (25 mL), dried over
MgSO4, and concentrated under reduced pressure. If
necessary, the isolated product is purified through silica
gel chromatography.
Acknowledgements
We thank Bertrand Castro (Sanofi-Chimie) for very
stimulating discussions and Sanofi-Chimie for financial
support (grant to A.N.).
References and notes
1. (a) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96,
395; (b) Trost, B. M. Acc. Chem. Res. 1996, 29, 355; (c)
Frost, C. G.; Howarth, J. M. J. Tetrahedron: Asymmetry
1992, 3, 1089; (d) Consiglio, G.; Waymouth, R. Chem.
Rev. 1989, 3, 257; (e) Williams, J. M. J. Synlett 1996, 705;
(f) Graening, T.; Schmalz, H.-G. Angew. Chem., Int. Ed.
2003, 42, 2580.
20. Compound 2: 1H NMR (CDCl3): d 2.80(dd, J = 13.8 and
6.9 Hz, 1H, CHH0CH@CH2), 2,82 (dd, J = 13.8 and
6.9 Hz, 1H, CHH0CH@CH2), 3.80(s, 3H, CH 3), 5.22–
5.28 (m, 2H, CH@CH2), 5.62–5.76 (m, 1H, CH@CH2),
7.39 (dd, J = 2.2 and 8.3 Hz, 1H, HAr), 7.48 (d, J = 8.3 Hz,
1H, HAr), 7.64 (d, 1H, J = 2.2 Hz, HAr). 13C NMR
(CDCl3): d 42.3 (CH2), 53.3 (CCN), 54.2 (OCH3), 117.1
(CN), 122.0( CH2@CH), 125.7 (CH2@CH), 128.4, 129.8,
131.0, 133.6 and 133.9 (CAr), 167.0(CO). Anal. Calcd for
C13H11Cl2NO2: C, 54.95; H, 3.9; N, 4.93. Found: C, 55.07;
H, 4.08; N, 4.72.
21. (a) Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34,
1769; (b) Von Matt, P.; Pfaltz, A. Angew. Chem., Int. Ed.
Engl. 1993, 32, 566; (c) Dawson, G. J.; Frost, C. G.;
Williams, J. M. J.; Cote, S. J. Tetrahedron Lett. 1993, 34,
3149.
2. Fiaud, J. C.; Hibon de Gournay, A.; Larcheveque, M.;
Kagan, H. J. Organomet. Chem. 1978, 154, 175.
ˆ
´
3. Genet, J. P.; Juge, S.; Achi, S.; Mallart, S.; Ruiz Montes,
J.; Levif, G. Tetrahedron 1988, 44, 5263.
4. Sawamura, M.; Sudoh, M.; Ito, Y. J. Am. Chem. Soc.
1996, 118, 3309.
5. Ito, Y.; Sawamura, M.; Matsuoka, Y.; Hayashi, T.
Tetrahedron Lett. 1987, 28, 4849.
6. Hayashi, T.; Kanehira, K.; Hagihara, T.; Kumada, M.
J. Org. Chem. 1988, 53, 113.
22. Miyashita, A.; Yasuda, A.; Takaya, H.; Toriumi, K.; Ito,
T.; Souchi, T.; Noyori, R. J. Am. Chem. Soc. 1980, 102,
7939.
23. Trost, B. M.; Van Vranken, D. L.; Bingel, C. J. Am.
Chem. Soc. 1992, 114, 9327.
7. Brunel, J. M.; Tenaglia, A.; Buono, G. Tetrahedron:
Asymmetry 2000, 11, 3585.
8. (a) Corey, E. J.; Guzman-Perez, A. Angew. Chem., Int. Ed.
1998, 37, 388; (b) Christoffers, J.; Mann, A. Angew. Chem.,
Int. Ed. 2001, 40, 4591.
24. Trost, B. M.; Murphy, D. J. Organometallics 1985, 4,
1143.
9. Trost, B. M.; Cramley, M. L. Chem. Rev. 2003, 103,
2921.
25. Fuji, K.; Kinoshita, N.; Tanaka, K. Chem. Commun. 1999,
1895.
10. (a) Trost, B. M.; Ariza, X. Angew. Chem., Int. Ed. 1997,
36, 2635; (b) Kuwano, R.; Ito, Y. J. Am. Chem. Soc. 1999,
121, 3236.
11. You, S.-L.; Hou, X.-L.; Dai, L.-X.; Cao, B.-X.; Sun, J.
Chem. Commun. 2000, 1933.
12. Sawamura, M.; Nagata, H.; Sakamoto, H.; Ito, Y. J. Am.
Chem. Soc. 1992, 114, 2586.
13. Trost, B. M.; Radinov; Grenzer, E. J. Am. Chem. Soc.
1997, 117, 7879.
14. Trost, B. M.; Ariza, X. J. Am. Chem. Soc. 1999, 121,
10727.
15. (a) Trost, B. M.; Schroeder, G. M. J. Am. Chem. Soc.
1999, 121, 6759; (b) Trost, B. M.; Schroeder, G. M.;
Kristensen, J. Angew. Chem., Int. Ed. 2002, 41,
3492.
26. Giambastiani, G.; Poli, G. J. Org. Chem. 1998, 63, 9608.
27. Cyano-(3,4)-dichlorophenyl acetic acid isopropyl ester:
this isopropyl ester was prepared quantitatively through a
transesterification of 1 performed in isopropylalcohol
under reflux. 1H NMR (CDCl3): d 1.25 (d, J = 6.1 Hz,
6H, (CH3)2), 4.65 (s, 1H, CHCN), 4.97–5.10(m, 1H,
CH(CH3)2), 7.29 (dd, J = 8.3 and 2.2 Hz, 1H, HAr), 7.49
(d, J = 8.3 Hz, 1H, HAr), 7.54 (d, J = 2.2 Hz, 1H, HAr).
13C NMR (CDCl3): d 21.4 (CH3), 43.0( CH(CH3)2), 72.1
(CN), 127.2, 129.9, 130.0, 131.2, 133.5, 133.8 (CAr), 163.6
(CO). Corresponding allyl derivative (2-cyano-2-(3,4-
dichlorophenyl-pent-4-enoic acid isopropyl ester): 1H
NMR (CDCl3): d 1.18 (d, J = 6.1 Hz, 3H, CH3), 1.28 (d,

A. Nowicki et al. / Tetrahedron Letters 46 (2005) 1617–1621
1621
J = 6.1 Hz, CH3), 2.79 (dd, J = 13.9 and 7.1 Hz, 1H,
CHH0CH@CH2), 3.06 (dd, J = 13.8 and 7.2 Hz, 1H,
CHH0CH@CH2), 5.00–5.09 (m, 1H, CH(CH3)), 5.22–
5.28 (m, 2H, CH@CH2), 5.64–5.72 (m, 1H, CH@CH2),
7.39 (dd, J = 8.3 and 2.0Hz, 1H, H Ar), 7.48 (d, J = 8.6 Hz,
1H, HAr), 7.64 (d, J = 2.0Hz, 1H, H Ar). 13C NMR
(CDCl3): d 21.2 (CH3), 21.4 (CH3), 42.2 (CH2), 53.5
(CCN), 72.0(OCH), 117,2 (CN), 121.8 ( CH2@CH), 125.6
(CH2@CH), 128.4, 130.0, 131.0, 133.4, 133.5, 134.2 (CAr),
165.9 (CO).
28. Sawamura, M.; Nakayama, Y.; Tang, W.-M.; Ito, Y.
J. Org. Chem. 1996, 61, 9096.
´
29. Cativiela, C.; Diaz-de-Villegas, M.; Galvez, J. A. J. Org.
Chem. 1994, 59, 2497.