Two-fold amino acid-based chiral aminophosphine-oxazolines and use in asymmetric allylic alkylation

Article abstract of DOI:10.1016/S0040-4039(03)01561-2

Chiral aminophosphine-oxazoline ligands derived from two different amino acids (tetrahydroisoquinoline carboxylic acid and proline) have been synthesized and examined as chiral auxiliaries in asymmetric allylic alkylation of three substrates. Stereoisomers 1a and 2a are providing the highest enantioselectivities (up to 94% ee) in the allylation of diphenylpropenyl acetate.

Full text of DOI:10.1016/S0040-4039(03)01561-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 6469–6473
Two-fold amino acid-based chiral aminophosphine–oxazolines and
use in asymmetric allylic alkylation
Catherine Blanc, Je´roˆme Hannedouche and Francine Agbossou-Niedercorn*
Laboratoire de Catalyse de Lille, UMR CNRS 8010, ENSCL, BP 108, 59652 Villeneuve d’Ascq Cedex, France
Received 11 April 2003; revised 23 June 2003; accepted 23 June 2003
Abstract—Chiral aminophosphine–oxazoline ligands derived from two different amino acids (tetrahydroisoquinoline carboxylic
acid and proline) have been synthesized and examined as chiral auxiliaries in asymmetric allylic alkylation of three substrates.
Stereoisomers 1a and 2a are providing the highest enantioselectivities (up to 94% ee) in the allylation of diphenylpropenyl acetate.
© 2003 Elsevier Ltd. All rights reserved.
Chiral auxiliaries have become key components in a
great number of catalyzed enantioselective transforma-
tions.1 As such, phosphorous-based chiral ligands are
extremely versatile and have shown stupendous proper-
ties in many catalytic reactions.1 Lately, the chemistry
of P,N-bidentate hybrid ligands has evolved consider-
ably since the original examples reported at the same
time by Helmchen,2 Pfaltz3 and Williams.4 This new
class of ligands is important in the development of
asymmetric catalysis as these auxiliaries have induced
high enantiodifferenciations in many reactions includ-
ing allylation, Heck coupling and other important
transformations.5 Structurally, these auxiliaries are
often related to phosphine–oxazolines5,6 where the chi-
rality is located in the oxazoline moiety, the latter being
constructed from chiral natural amino alcohols. Contri-
butions describing P-oxazolines containing P-het-
eroatom residues have appeared more scarcely even
though the corresponding auxiliaries show equally sig-
nificant properties.7–9
Gilbertson reported the first efficient ligands of that
family where a proline provided the aminophosphine
extremity.12
Herein we report on our first results on the synthesis of
aminophosphine–oxazolines and on their use in asym-
metric allylic alkylation.
The aminophosphine–oxazoline ligands 1a and 1b,
based on tetrahydroisoquinoline carboxylic acid, and
2a and 2b, based on proline (Scheme 1),12 have been
synthesized from the corresponding amino acids. Thus,
commercial (S)-1,2,3,4-tetrahydroisoquinoline car-
boxylic acid has been first converted into its methyl
ester 3-Q using standard procedures. The following step
consisted of preparing the protected aminophosphine
residues (Scheme 2). As a matter of fact, among the
several pathways which can be considered for the syn-
thesis of the target auxiliaries, it appeared that the most
efficient method consisted of preparing the aminophos-
phine moiety, as a protected unit, in the beginning of
the synthesis. We have investigated two protection
modes of the phosphorous, i.e. the sulfide and the
oxide. The possible phosphinoborane methodology13
As part of our ongoing exploration of various synthetic
approaches to new ligands, we have been using the
chiral pool, i.e. amino acids, for the synthesis of
aminophosphine–phosphinite ligands and focused on
their properties in asymmetric catalysis.10 The cyclic
amino acid-based ligands proved to be the most appro-
priate for highly enantioselective hydrogenations of
ketones.11 Thus, we sought to use such chiral cyclic
amino acids in association with an optically pure amino
alcohol to access other types of aminophosphine–oxa-
zoline architectures. During our preliminary study,
* Corresponding
author.
Fax:
+33-(0)3-2043-6585;
e-mail:
francine.agbossou@ensc-lille.fr
Scheme 1.
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)01561-2

6470
C. Blanc et al. / Tetrahedron Letters 44 (2003) 6469–6473
Scheme 2. Reagents and conditions: (a) ClPPh2, NEt3, THF, 40°C, 12 h (90%); (a%) ClP(O)Ph2, NEt3, THF (96%); (b) S8, toluene,
rt, 1 h (85%); (c) LiOH, MeOH, reflux, 24 h (90%); (d) (S)- or (R)-valinol, BOP, NEt3, MeCN (99%); (e) p-TsCl, DMAP, CH2Cl2
(90%); (f) Ni Raney, MeCN (95%); (g) HSiCl3, NEt3 (50%).
has not been retained for this study because of the basic
conditions required by the following steps of the syn-
thesis which would probably lead to a decomplexation
of the BH3 unit from the ‘P·BH3’ moiety of the pro-
tected intermediate. Thus, the sulfide is obtained in two
steps. The intermediate amino–phosphines 5-Q and 6-P
are prepared via phosphinylation of the esters 3-Q and
4-P and then reacted with sulfur providing the P(S)
intermediate 7-Q(S) and 8-P(S). On the other side, the
two esters 5-Q and 6-P can be reacted easily with
ClP(O)Ph2 providing the corresponding P(O) deriva-
tives 7-Q(O) and 8-P(O). Even if the four isolated P(V)
residues have been used successfully for the whole
synthesis, the sulfide appeared more convenient because
of an easier isolation procedure from the final step of
the synthesis (vide infra). The acids 9-Q(S), 9-Q(O),
10-P(S), and 10-P(O) were obtained through hydrolysis
of the corresponding esters. Then, a standard peptide
coupling14 between 9-Q(S), 9-Q(O), 10-P(S), and 10-
P(O) and both enantiomers of valinol provided, quanti-
tatively after workup, the phosphine–sulfide and
phosphine–oxide amidoalcohols 11a-Q(S) and 11a-
Q(O), 12a-P(S), and 12a-P(O) (from (S)-valinol) and
11b-Q(S) and 11b-Q(O), 12b-P(S), and 12b-P(O) (from
(R)-valinol). The oxazoline-based compounds 13a-Q(S),
13a-Q(O), 13b-Q(S), 13b-Q(O), 14a-P(S), 14a-P(O),
14b-P(S), and 14b-P(O) were reached conveniently via
cyclization of the amidoalcohols in the presence of
p-tosylchloride under basic conditions and isolated in
high yields.15 Finally, the amino–phosphine residues
were recovered from 13a-Q(S), 13b-Q(S), 14a-P(S), and
14b-P(S) through reduction with Ni Raney in acetoni-
trile providing 1a, 1b, 2a, and 2b in ca. 95% yield.16 The
phosphine–oxide oxazolines were transformed into the
corresponding phosphine–oxazolines 1a, 1b, 2a, and 2b
in the presence of trichlorosilane/NEt3 in toluene.17,18
The products were isolated in ca. 50% yield. The overall
yield of the synthesis of the chiral auxiliaries through
the sulfide route starting from the amino ester is 58%
which is somewhat higher than the ca. 39% yield
obtained while following the oxyde route. Nevertheless,
the latter can be applied to the synthesis of new ligands
bearing a stereogenic phosphorous atom which can be
recovered by a stereoselective reduction.17 All new com-
pounds have been fully characterized and exhibit classi-
cal spectroscopic features.19
Asymmmetric allylic alkylation is a standard model
reaction for the assessment of new ligands in CꢀC
forming reactions. In order to evaluate these ligands,
three substrates (15–17) have been selected for this
initial study (Scheme 3). The precatalysts were obtained
by reaction of the chloroallylpalladium dimer with the
chosen ligands in the solvent used for catalysis (THF
and MeCN).20
Scheme 3.

C. Blanc et al. / Tetrahedron Letters 44 (2003) 6469–6473
6471
process.21 As a matter of fact, allylic alkylations carried
out in the presence of either [Pd(allyl)(1a)]PF6 or a
combination of Pd2dba3 and the 1a ligand are provid-
ing enantioselectivities within a 2% ee of the value
given here (entry 3). As reported, the allylic palladium
catalytic intermediate exists as two diastereoisomers
designated exo and endo. The addition of a nucleophile
on such species can, in theory, follow four pathways
arising from a reaction on the two allylic C termini of
the two diastereomers. In practice, with P,N type lig-
ands, it has been proposed that the nucleophile adds
preferentially trans to the phosphorous. Indeed, the
stereoelectronic ‘trans-influence’ dictates that the allylic
carbon trans to the P-atom is more electrophylic than
the C-allylic terminus trans to the N-atom.22 The
observed major R diphenyl product arises from the
nucleophilic addition trans to P,N on the exo allylic
intermediate. In theory, if a similar addition pathway is
applied to the corresponding dimethyl allylic
diastereomer, the antipode product is obtained.23 In
practice, this is effectively observed in the presence of
1a for the allylation of 15 and 16 (entries 3 and 14). The
chirality of the oxazoline moiety orients the enantiodif-
ferenciation of the addition to 15. Interestingly, the
match/mismatch behavior which could be expected
from the diastereomeric pair 1a/1b is not observed as
both auxiliaries are providing high ee’s into each
product antipode. The flexibility of the six membered
cycle of ligands 1a and 1b is capable of providing
equally favored conformations able to induce high
enantiodifferenciations in either enantiomer of the
product. Conversely, the pair 2a/2b is exhibiting a
For the transformation of 1,3-diphenyl allyl acetate 15,
the solvent had no significant effect on the selectivity of
the reaction (Table 1, entries 1–5) as ca. 90% ee or
higher were obtained in either acetonitrile or tetra-
hydrofuran. For the substitution of substrates 15 and
16, ligands 1a and 2a exhibit identical properties
(entries 1–8 and 11). For the allylation of 15, a lowering
of the temperature led to a smaller decrease of the
enantioselectivity in the presence of 1a (Dee=3%) com-
pared to 2a (Dee=7%) (entries 3/4 and 11/12). Con-
versely, an increase of 8% ee was obtained while
lowering the temperature during the allylation of 16 in
the presence of both chiral auxiliaries 1a and 2a (entries
14/15 and 17/18). We could notice that a substrate to
catalyst ratio increase to 1000/1 yielded a total transfor-
mation within non optimized 65 h with no loss of the
enantioselectivity in acetonitrile (entries 2/6). However,
a small erosion of the ee occurred in THF (entries 3/8).
The steric course of the reaction was different for the
diastereomeric pair 1a/1b compared to the pair 2a/2b.
As a matter of fact, the two diastereomers 1a and 1b
are providing both antipodes of the allylation product
of 15 with a high level of enantioselectivity. Actually,
78% ee of the (S) enantiomer are achieved in the
presence of 1b and 90% ee of the (R) product with 1a.
The contribution of 2b resulted in the production of the
(S) enantiomer in 22% ee while 2a gave the (R) enan-
tiomer in 94% ee. We think that the chiral auxiliary
remains coordinated in a chelate fashion and the allyl
residue in a h3 mode and that the chlorine atom
initially present on the palladium complex does not stay
on the metal during the enantioselective step of the
Table 1. Palladium-catalysed asymmetric allylic allylationa
Entry
Substrate
Ligand
Solvent
Co-catalyst
S/Pd
Time (h)
Temp. (°C)
Conv. (%)b
Ee (%) (config.)c
1
2
3
4
5
6
7
8
15
1a
1a
1a
1a
1a
1a
1a
1a
1b
1b
2a
2a
2b
MeCN
MeCN
THF
KOAc stoich.
KOAc 10%
KOAc 10%
KOAc 10%
Bu4N+ 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
100
100
100
100
100
1000
1000
1000
250
100
50
3
4
20
20
20
0
20
20
20
0
20
20
20
0
91
100
100
100
73
100
100
100
97
89 (R)
90 (R)
93 (R)
90 (R)
90 (R)
90 (R)
88 (R)
87 (R)
77 (S)
78 (S)
94 (R)
87 (R)
22 (S)
18
18
3.75
18
18
65
5.5
18
18
18
18
THF
MeCN
MeCN
THF
THF
THF
MeCN
THF
THF
9
10
11
12
13
98
100
100
100
100
100
THF
0
14
15
16
17
18
19
16
17
1a
1a
1b
2a
2a
2b
THF
THF
THF
THF
THF
THF
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
KOAc 10%
100
100
100
100
100
100
18
65
18
18
18
18
20
0
20
20
0
77
100
90
100
100
100
39b (S)
47 (S)
25b (S)
37b (S)
45 (S)
5b (S)
0
20
21
1a
2a
THF
THF
KOAc 10%
KOAc 10%
100
100
18
18
20
20
100
100
<5c (n.d.)
<5c (n.d.)
a The catalytic reactions were carried out in the presence of a ligand/Pd ratio=1.2 under the conditions mentioned in Table 1.
b The conversions were determined by 1H NMR on the crude reaction mixture through integration of the methyl residues.
c Enantiomeric ratios were determined by HPLC on Chiralpak AD, hexane/isopropyl alcohol=9/1 for 15, by NMR spectroscopy using the
[Eu(hfc)] shift reagent for 16 and 17.

6472
C. Blanc et al. / Tetrahedron Letters 44 (2003) 6469–6473
match/mismatch behavior (entries 12/13). As well, the
same antipode is produced with both diastereomers 1a
and 1b during the allylation of 16. Finally, the transfor-
mation of the cyclic substrate 17 is occurring with a
very low enantioselectivity.
8. For phosphite–oxazoline, see: (a) Heldmann, D.; See-
bach, D. Helv. Chim. Acta 1999, 82, 1096; (b) Escher, I.;
Pfaltz, A. Tetrahedron 2000, 56, 2879; (c) Alexakis, A.;
Rosset, S.; Allamand, J.; March, S.; Guillen, F.; Ben-
haim, C. Synlett 2001, 1375.
9. For phosphoramidite–oxazolines, see: Hilgraf, R.; Pfaltz,
In summary, chiral aminophosphine–oxazoline auxil-
iaries are easily accessible from natural amino acids and
alcohols. The two phosphorous atom protection modes
utilized constitute valuable tools for the synthesis of
such auxiliaries even those possessing a stereogenic
phosphorous atom. The auxiliaries described are
efficient in the alkylation of diphenyl allyl acetate. Even
if the selectivity is lower with other substrates, it com-
pares well to other systems described. Other ligands of
that family and their use in enantioselective transforma-
tions will be reported soon.
A. Synlett 1999, 1814.
10. (a) Agbossou, F.; Carpentier, J.-F.; Hapiot, F.; Suisse, I.;
Mortreux, A. Coord. Chem. Rev. 1998, 178–180, 1615; (b)
Agbossou-Niedercorn, F.; Suisse, I. Coord. Chem. Rev.,
in press.
11. Pasquier, C.; Naili, S.; Mortreux, A.; Agbossou, F.;
Pe´linski, L.; Brocard, J.; Eilers, J.; Reiners, I.; Peper, V.;
Martens, J. Organometallics 2000, 19, 5723.
12. Xu, G.; Gilbertson, S. R. Tetrahedron Lett. 2002, 43,
2811.
13. Moulin, D.; Bago, S.; Bauduin, C.; Darcel, C.; Juge´, S.
Tetrahedron: Asymmetry 2000, 11, 3939 and references
cited therein.
14. (a) Castro, B.; Dormoy, J. R.; Evin, G.; Selve, C. Tetra-
hedron Lett. 1975, 16, 1219; (b) Fehrentz, J. A.; Castro,
B. Synthesis 1983, 676.
Acknowledgements
15. Sammakia, T.; Latham, H. A. J. Org. Chem. 1996, 61, 5.
16. Gilbertson, S. R.; Wang, X. J. Org. Chem. 1996, 61, 434.
17. Smart, B. E. Org. Synth. 1988, 67, 20.
We thank Professor A. Mortreux and Dr. Lydie Pe´lin-
ski for fruitful discussions. We also thank Nadia Dje-
lassi for some experiments. We are grateful to CNRS
for financial support. C.B. acknowledges MEN for a
doctoral fellowship.
18. The origin of the yield loss experienced during the reduc-
tion of the PꢁO moiety is linked to the purification
procedure. Actually, the necessary filtration over silica gel
had to be carried out under basic conditions (silica gel
washed with NEt3) in order to avoid a cleavage of the
PꢀN bond, the latter being sensitive to the acidic medium
produced by the hydrolysis of the excess of HSiCl3 on the
silica gel.
References
1. (a) Asymmetric Catalysis in Organic Synthesis; Wiley:
New York, 1994; (b) Catalytic Asymmetric Synthesis;
Ojima, I., Ed.; VCH: New York, 1993; (c) Brunner, H.;
Zettlmeier, W. Handbook of Enantioselective Catalysis;
VCH: Weinheim, 1993; (d) Comprehensive Asymmetric
Catalysis; Jacobsen, E. N.; Pfaltz, A.; Yamamoto, H.,
Eds.; Springer: Berlin, 1999; Vols. 1–3.
2. Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34,
1769.
3. Von Matt, P.; Pfaltz, A. Angew. Chem., Int. Ed. Engl.
1993, 32, 566.
19. 1a: 1H NMR (300 MHz, toluene-d8): l 0.53 (d, J=6.8
Hz, 3H, CH3); 0.59 (d, J=6.8 Hz, 3H, CH3); 1.25 (br.
hept., J=6.8 Hz, 1H, CH(CH3)2); 3.07 (dd, J=15.9 and
5.9 Hz, 1H, ArCHtransCHcisCHN); 3.20 (d, J=16.0 Hz,
1H, ArCHtransCHcisCHN); 3.45 (m, 2H, CH2-O); 3.66
(dd., J=6.7 and 7.9 Hz, 1H, NCHCH2O); 4.05 (d, J=
16.4 Hz, 1H, ArCHH%N); 4.43 (d, J=16.1, 1H,
ArCHH%N); 4.57 (br. t, J=6.4 Hz, 1H, CHNP); 6.88 (m,
3H, Harom.); 6.97–7.13 (m, 7H, Harom.); 7.46 (m, 2H,
H
arom.); 7.88 (m, 2H, Harom.). 31P{1H} NMR (121.5 MHz,
4. Dawson, G. J.; Frost, C. G.; Williams, J. M. J.; Cote, S.
CDCl3): l 64.6 ppm (s).
J. Tetrahedron Lett. 1993, 34, 3149.
2a: 1H NMR (300 MHz, C6D6): l 0.74 (d, J=6.7 Hz, 3H,
CH3); 0.81 (d, J=6.7 Hz, 3H, CH3); 1.36 (m, 1H,
CHHCH2CH); 1.53 (br. sept., J=6.6 Hz, 1H,
CH(CH3)2); 1.80 (m, 2H, CHH-CHHCH); 1.96 (m, 1H,
CHH-CH); 2.74 (m, 2H, CHHN); 3.12 (m, 2H, CHH-
N); 3.66 (m, 2H, CH2O); 3.82 (dd, J=5.7 and 7.6 Hz,
1H, NCH-CH2O); 4.66 (m, 1H, CH); 7.27–7.39 (m, 6H,
5. (a) Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33,
336 and references cited therein; (b) Pfaltz, A. J. Hetero-
cyclic Chem. 1999, 36, 1437; (c) Braunstein, P.; Naud, F.
Angew. Chem., Int. Ed. 2001, 40, 681; (d) Espinet, P.;
Soulantica, K. Coord. Chem. Rev. 1999, 193–195, 499; (e)
Gomez, M.; Muller, G.; Rocamora, M. Coord. Chem.
Rev. 1999, 193–195, 769.
H
arom.); 7.93–8.02 (m, 4H, Harom.). 31P{1H} NMR (121.5
6. (a) Ogawara, M.; Yoshida, K.; Kamei, H.; Uozumi, Y.;
Hayashi, T. Tetrahedron: Asymmetry 1998, 9, 1779; (b)
Hou, D. R.; Reibenspies, J. H.; Burgess, K. J. Org.
Chem. 2001, 66, 206.
7. For phosphinite–oxazoline, see: (a) Gomez-Simon, M.;
Jansat, S.; Muller, G.; Panyella, D.; Font-Bardia, M.;
Solans, X. J. Chem. Soc., Dalton Trans. 1997, 3755; (b)
Blankenstein, J.; Pfaltz, A. Angew. Chem., Int. Ed. 2001,
40, 4445. For aminophosphine–oxazoline, see: (c) Cozzi,
P. G.; Zimmermann, N.; Hilgraf, R.; Schaffner, S.;
Pfaltz, A. Adv. Synth. Catal. 2001, 343, 450.
MHz, C6D6): l 48.1 ppm (s).
20. General procedure for the allylic alkylation: Under nitro-
gen, in a 100 mL Schlenk tube equiped with a stir bar,
the aminophosphine oxazoline and [Pd(h3-C3H5)Cl]2
(0.45 equiv.) are reacted in a freshly distilled and
degassed solvent. Then, 1,3-diphenyl allyl acetate (100
equiv.) is added to the solution and the mixture is stirred
at room temperature for 1 h. In a separate Schlenk tube,
a mixture of dimethyl malonate and BSA (both as 3
equiv. with respect to the allylic substrate) and KOAc
(catalytic amount) in 3 mL of the solvent is prepared.

C. Blanc et al. / Tetrahedron Letters 44 (2003) 6469–6473
6473
This mixture is added dropwise to the catalytic solution.
At the end, the reaction mixture was diluted with diethyl
ether and quenched by saturated sodium hydrogenocar-
bonate. The product was extracted with diethyl ether, the
organic phase is washed with brine and dried over magne-
sium sulfate. After filtration and evaporation under
reduced pressure, a white solid was isolated in 95% yield.
The enantiomeric excess was determined by HPLC using
a Chiralpak AD column (iPrOH/hexane: 10/90; flow rate:
1 mL/min; detection UV at 254 nm; tR(R)=8.4 min,
tR(S)=11.6 min).
22. (a) Helmchen, G. J. Organomet. Chem. 1999, 576, 203;
(b) Leutenegger, U.; Umbricht, G.; Fahrni, C.; von
Matt, P.; Pfaltz, A. Tetrahedron 1992, 48, 2143; (c)
Gant, T. G.; Meyers, A. I. Tetrahedron 1994, 50, 2297;
(d) Koch, G.; Lloyd-Jones, G. C.; Loiseleur, O.; Pfaltz,
A.; Pre´toˆt, R.; Schaffner, S.; Schnider, P.; von Matt, P.
Recl. Trav. Chim. Pays-Bas 1995, 114, 206; (e) Allen, J.
V.; Coote, S. J.; Dawson, G. J.; Frost, C. G.; Martin,
C. J.; Williams, J. M. J. J. Chem. Soc., Perkin Trans. 1
1994, 2065.
23. Even if the product antipodes are produced depending
on the substrate, there is no change of the descriptor of
the absolute configuration.
21. Braunstein, P.; Naud, F.; Dedieu, A.; Rohmer, M.-M.;
DeCian, A.; Rettig, S. J. Organometallics 2001, 20, 2966.