The rational design of modified Cinchona alkaloid catalysts. Application to a new asymmetric synthesis of chiral chromanes

Article abstract of DOI:10.1016/j.tetlet.2004.04.090

A new asymmetric synthesis of 2-substituted chiral chromanes has been achieved. The key step is the intramolecular conjugate addition of a phenolic nucleophile on a α,β-unsaturated ester catalyzed by Cinchona alkaloids. The high ee's obtained with cinchonine and its derivatives have been rationalized by ab initio quantum chemistry calculations of transition state structures.

Full text of DOI:10.1016/j.tetlet.2004.04.090

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 4697–4701
The rational design of modified Cinchona alkaloid catalysts.
Application to a new asymmetric synthesis of chiral chromanes
a
b
c
Alain Merschaert,a, Pieter Delbeke, Desire Daloze and Georges Dive
*
ꢀ ꢀ
aGlobal Chemical Product Research & Development, Lilly Research Laboratories, Lilly Development Centre S.A.,
Rue Granbonprꢀe 11, B-1348 Mont-Saint-Guibert, Belgium
bService de Chimie Organique, Facultꢀe des Sciences, Universitꢀe Libre de Bruxelles, Av. F.D. Roosevelt 50, B-1050 Bruxelles, Belgium
cCentre d’Ingꢀeniꢀerie des Protꢀeines, Universitꢀe de Liꢁege, B^atiment B6, Allꢀee de la Chimie 3, B-4000 Liꢁege, Belgium
Received 25 March 2004; revised 14 April 2004; accepted 16 April 2004
Abstract—A new asymmetric synthesis of 2-substituted chiral chromanes has been achieved. The key stepis the intramolecular
conjugate addition of a phenolic nucleophile on a a,b-unsaturated ester catalyzed by Cinchona alkaloids. The high ee’s obtained
with cinchonine and its derivatives have been rationalized by ab initio quantum chemistry calculations of transition state struc-
tures.
Ó 2004 Elsevier Ltd. All rights reserved.
Functionalized chromanes possess potentially useful
biological properties.1 In the scope of a research pro-
gram, we developed earlier the synthesis of 2,6-disub-
stituted-chromanes from 6-substituted chromanones2
(Fig. 1).
utilize a chiral base in the Wittig-oxa-Michael step(Fig.
2). The results of those efforts are reported herein. Ini-
tially very low enantioselectivities (ca. 28% ee with cin-
chonine) were obtained in a tandem process by heating a
mixture of lactol 1, phosphorane, and cinchonine to 80–
100 °C. In this temperature range, we observed that the
phosphorane alone was able to catalyze the cyclization
thus competing with the chiral catalyst. Therefore, we
modified the procedure and performed the reaction
sequentially: (i) the Wittig stepwas carried out at
ambient temperature (ii) the remaining phosphorane
and the triphenylphosphine oxide by-product were
removed by silica gel treatment, and (iii) the chiral
Although high overall yields of racemic material were
obtained, this approach suffered from the need to per-
form an optical resolution in the last step of the syn-
thesis.
In order to improve the productivity of the process
without having to developa new route, we sought to
O
O
O
OH
O
O
H2/Pd-C
DIBALH
CH2Cl2
-78°C
90-95%
THF
20°C
>95%
R
R
R
Ph3P=CH-CO2Et
(cat)
EtONa
85-90%
Toluene
110°C
O
O
O
CO2H
CO2H
NaOH
CO2Et
~100%
25-40%
R
R
R
Figure 1.
Keywords: Alkaloids; oxa-Michael; Asymmetric; Catalysis.
* Corresponding author. Tel.: +32-10-47-70-21; fax: +32-10-47-63-15; e-mail: merschaert_alain@lilly.com
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.04.090

4698
A. Merschaert et al. / Tetrahedron Letters 45 (2004) 4697–4701
O
Table 1. Modifications of cinchonine at carbon-9
Ph3P=CH-CO2R'
Chiral base
80-100°C
O
OH
CO2R'
11
R
R
Max. 38% e.e.
4
10
1a: R = Me
3
Chiral base
Up to 80% e.e.
OH
1b: R = H
Up to 54% e.e.
OH
2
4
CO2R'
1c: R = Boc-NH
7
5
Ph3P=CH-CO2R'
20°C
R
1
N
CO2R'
1
8
6
+
9
H
or (CF3CH2O)2P(O)-CH2CO2Me
R2
R
5'
8'
4'
R
3
KN(TMS)2
-40°C
2
3'
2'
6'
7'
2a/3a: R = R' = Me
2b/3b: R = Me, R' = Et
2c/3c: R = H, R' = Et
2d/3d: R = Boc-NH, R' = Me
2e/3e: R = Boc-NH, R' = Et
2f/3f : R = Boc-NH, R' = tBu
2g/3g: R = Boc-NH, R' = Bn
OH
O
N
1'
CO2Me
CO2Me
Toluene
Me
Me
3a
4
Figure 2.
Entry
Catalysta
(10 mol %)
R1
R2
Conv
(%)b
R/S ratio
(rel%)b
catalyst was added to the crude mixture of olefins
(ꢀ90% E isomer) and the solution heated to 80–100 °C.
Unfortunately, we obtained only slightly better results
(ca. 38% ee with cinchonine). Anyway, simple use of
cinchonine in the Wittig-oxa-Michael stepallowed us to
increase the overall yield by 10–15%. After separation of
the (E)- and (Z)-a,b-unsaturated esters 2e–3e, we found
that these compounds cyclized into the opposite
enantiomers: with cinchonine (5–10 mol %) as catalyst,
the E isomer 2e gave the S-chromane 4 (44% ee) and the
Z olefin 3e afforded the R enantiomer (66% ee). It is
worth mentioning that among Cinchona alkaloids cin-
chonine gave the best results for both E and Z isomers
(cinchonine > quinidine > cinchonidine ꢁ quinine). On
the other hand, we attempted to increase the enantio-
selectivity by modifying the alkyl chain (methyl, tert-
butyl, and benzyl) of the ester function but with little
success: upto 67% ee for the Z benzyl ester 3g and 54%
ee for the E tert-butyl ester 2f.
1
2
3
4
5
6
5
OH
H
H
90
10
5
82:18
68:32
nd
67
78
86
96
109
OH
Me
H
OH
OMe
OAc
H
<1
<1
25
nd
nd
H
H
61:49
a Reactions were carried out in toluene (0.5 M) at 110 °C under N2 for
48 h.
b Determined by HPLC––Column Chiralcel OD, elution with 1 mL/
min n-heptane/EtOH/diethylamine 99/1/0.2, UV detection at 235 nm.
We decided thus to focus on cinchonine (5) and eval-
uated first modifications around chiral carbon-96
(Table 1).
Table 1 shows all the modifications envisaged in cata-
lysts 6–10. It is noteworthy that increase of steric
hindrance and conformational constrain (entry 3), O-
functionalization (entries 4 and 5) and de-hydroxylation
(entry 6) have a large negative effect. The presence of a
conformationally mobile free hydroxy groupis thus
needed in the catalyst to achieve both good kinetics and
enantioselectivity.
We performed a broader screening of chiral bases and
solvents3 but none of the systems studied could improve
the results obtained with readily available cinchonine. In
many experiments performed on the Z substrates, we
observed the formation of the E isomer. Control
experiments showed that this latter product is not
formed by direct isomerization of the Z substrate4 or by
retro-Michael on the final chromane.5 We assumed
therefore that protonation of the ester enolate formed
after cyclization was too slow in aprotic solvents thus
allowing retro-Michael reaction leading to the for-
mation of the E isomer.
We envisaged then the modification of the vinyl side
chain of cinchonine10 5 and selected the aldehyde 12 as
the key intermediate of our approach (Fig. 3).
We found that epimerization occurs at carbon-3 during
the periodate oxidation step11 and developed an efficient
crystallization in diisopropylether to obtain highly pure
HO
O
11
i)Ac2O/Pyridine
3
10
H
ii)
OsO4
HO
O
Ac
H
K3[Fe(CN)6]
K2CO3
Ac
H
NaIO4
SiO2
O
O
N
H
N
H
N
9
8
4'
3'
CH2Cl2/H2O
t-ButOH/H2O
5
11
12
N
N
N
H2/Pd-C
EtOH
R
R
1
R
R
X
2
2
1
E + Z
+
Ph3P -CHR1R2 X-
KN(TMS)2
THF
i)
H
OH
N
H
HO
H
H2/Pd-C
EtOH
HO
N
N
H
ii)
K2CO3
27-30
N
CH3OH
13-23
24-26
N
N
Figure 3. Preparation of C3-modified catalysts.

A. Merschaert et al. / Tetrahedron Letters 45 (2004) 4697–4701
4699
Table 2. Preparation of C3-modified catalysts
Entry Compound R1
Table 3. Cyclization of Z ester model 3a with C3-modified catalysts
R2
Yield (%)a
OH
O
CO2Me
CO2Me
1
13
14
15
16
17
18
19
20
21
22
23
24
25
26
H
Me
25
45
80
70
68
12
30
<5
<5
ꢀ0
<5
95c
85
90
Me
Me
2
Me
H
3a
4
3
H/Etb
i-Pr/Hb
H
Et/Hb
H/i-Prb
Ph
Entry
Catalyst
(10 mol %)a
Conversion (%)b
R/S ratio
(rel%)b
4
5
6
Ph
Me
H
Me
1
5
45
37
30
44
34
23
20
44
75
45
35
35
42
<2
<2
82:18
78:22
83:17
87:13
89:11
88:12
90:10
87:13
86:14
88:12
90:10
65:35
87:13
nd
7
2
13
14
15
16
17
18
19
24
25
26
27
28
29
30
8
Me/Et
Et/Me
3
9
c-Hex
c-Pr
4
10
11
12
13
14
5
Ph
H
Ph
H
6
7
Me
H
Me
Ph
8
9
10
11
12
13
14
15
X
15
16
17
18
27
28
29
30
H
60
95
60
90
CH2OH
CN
CHO
nd
a Isolated unoptimized yields.
a Reactions were carried out in toluene (0.5 M) at 80 °C under N2 for
24 h.
b Determined by HPLC––Column Chiralcel OD, elution with 1 mL/
min n-heptane/EtOH/diethylamine 99/1/0.2, UV detection at 235 nm.
b E/Z mixture (ꢀ1:2 by 1H NMR) could not be separated.
c Commercially available at 97% purity (Aldrich).
12. Catalysts 13–23 (Table 2) were prepared by reaction
of 12 with ca. 3 equiv of the Wittig reagents prepared
from the corresponding phosphonium halides and
LHMDS.12
On the other hand, hydrogenation of the vinyl group
(entries 9–11) increased conversion and enantioselectiv-
ity, especially for the least hindered catalyst (24 > 5).
We obtained about 80% yield for the mono-substituted
compounds as E/Z mixtures (entries 1–6) but only the
11-methyl and 11-phenyl derivatives could be separated
by chromatography on silica gel. With the exception of
the dimethyl compound 19, only traces of the 11,11-
disubstituted derivatives were obtained (entries 8–11). In
these unsuccessful experiments, most recovered alde-
hyde 12 was largely epimerized at carbon-3. Cinchonine
(5), 17, and 19 were hydrogenated with 10% Pd–C to
afford the saturated catalysts 24–26 in high yield (entries
12–14). From the aldehyde 12, we prepared also the
catalysts 28,13 29,13 and 30 in order to evaluate elec-
tronic effects on the quinuclidine nitrogen14 and the
catalyst 2713 to assess lack of substitution at carbon-3 of
the quinuclidine ring.
We assume that the intrinsically higher steric hindrance
of the saturated compounds favors discrimination
between the transition states. Higher conversions are
explained by an increase of the nitrogen basicity due to a
higher þi effect at position-3.17 The poor results ob-
tained with catalyst 27 reinforce our assumption about
the need for steric hindrance at position-3. We have also
shown that introduction of a polar hydroxy function is
tolerated (entry 13) contrary to strong electron-with-
drawing groups, which give significantly poorer results
(entries 14 and 15). Catalysts 29–30 are most probably
not basic enough to deprotonate the phenolic sub-
strate.14
As illustrated in Table 4, the effect of a structural
modification of the catalyst on the E ester 2a cyclization
is relatively similar to that observed for the Z derivative.
These catalysts were tested in the cyclization of model Z
and E esters 2a and 3a.15 Reactions were carried out at a
lower temperature (80 °C) to better discriminate the
effects of our modifications on the conversion and to
minimize possible retro-Michael side reaction16 occur-
ring at higher temperature. As shown in Table 3, mod-
ifications at carbon-3 generally affect conversion and
enantioselectivity. Globally, the introduction of alkyl or
aryl groups on the vinyl side chain of the catalyst (en-
tries 2–8) had a positive effect on enantioselectivity and a
negative impact on conversion. The best enantioselec-
tivities were obtained with the bulkiest groups (entries
4–7) albeit at low conversion. For both methyl and
phenyl groups, the Z geometry gave better enantiose-
lectivity (14 > 13 and 18 > 17). Dialkylation at the ter-
minal position of the olefin was also beneficial (entry 8).
Steric hindrance (entries 2, 4, and 5) and hydrogenation
(entries 6–8) of the vinyl side chain have generally a
positive effect whereas electron-withdrawing groups
(entries 11 and 12) lead to low conversions. As observed
in our earlier studies, cyclization of the E isomer is al-
ways faster but less enantioselective. Of the catalysts
tested, 26 gave the highest enantioselectivities for both Z
(80% ee) and E (52% ee) substrates.
In order to rationalize the results obtained with cin-
chonine and to helpdesign better catalysts, we per-
formed
a
molecular modeling study. Numerous
conformational studies have been conducted on Cin-
chona alkaloids.18 The relative populations of the more

4700
A. Merschaert et al. / Tetrahedron Letters 45 (2004) 4697–4701
Table 4. Cyclization of E ester model 2a with C3-modified catalysts
CO2Me
O
OH
CO2Me
Me
Me
4
2a
Entry
Catalyst
(10 mol %)a
Conversion
(%)b
R/S ratio
(rel%)b
Figure 4. Possible transition state I for Z ester isomer 3a.
1
5
95
58
70
35
60
100
75
82
50
80
3
30:70
31:69
40:60
26:74
28:72
25:75
27:73
24:76
41:59
32:68
nd
2
13
17
18
19
24
25
26
27
28
29
30
3
4
5
6
7
8
9
10
11
12
10
30:70
Figure 5. Possible transition state II for Z ester isomer 3a.
a Reactions were carried out in toluene (0.5 M) at 80 °C under N2 for
24 h.
b Determined by HPLC––Column Chiralcel OD, elution with 1 mL/
min n-heptane/EtOH/diethylamine 99/1/0.2, UV detection at 235 nm.
part in a concerted mechanism (Fig. 5). This pathway
avoiding the formation of a charged ester enolate after
cyclization is energetically favored in apolar solvents.
stable conformations are very sensitive to the environ-
mental variations such as pH, solvent polarity or H-
bonding.19 It is however hazardous to limit the study of
a catalyst–substrate interaction to the most stable con-
formers of the free alkaloid as the relative differences
between their energies are relatively low compared to the
energies of transition states.20
TS II might thus explain the much lower extend of retro-
Michael side reaction observed for cinchonine and its
derivatives as compared to most other aminoalcohols.
As for the less enantioselective E isomer cyclizations we
identified only a structure very similar to I, we tenta-
tively propose II as the preferred transition state for the
cyclization of Z compounds.
Therefore, we decided to use ab initio quantum chem-
istry calculations, which build stepby stepmodels of the
substrate–catalyst complex starting with the key hy-
droxy and amino groups of the alkaloid.21 A potential
energy mapof cinchonine has been calculated for the
torsion angles /1 (C40–C9–C8–N1) and /2 (C30–C40–
C9–C8) by optimization of all the other 3n ꢂ 8 degrees
of freedom. As the differences in free energy between 6
of the 8 minima found are <2.5 kcal/mol, we can con-
clude that several of them are energetically accessible for
the intramolecular oxa-Michael reaction. We found two
possible transition states for the cyclization of the Z
substrate. Both involve the formation of a tight ion-pair
between the basic nitrogen of cinchonine and the phe-
nolic groupof the substrate as well as a large defor-
mation of the quinuclidine ring during cyclization.
Torsion angle /1 of cinchonine remains close to 300°
whereas /2 varies from )20° to +40°. Transition state
(TS) I is stabilized by hydrogen bonding between the
hydroxy groupof the catalyst and the carbonyl function
of the a,b-unsaturated ester. Preferential formation of
the R enantiomer is explained by a lower steric repulsion
between the vinyl and quinoline groups (Fig. 4). Indeed,
the distance between carbons 11 and 60 in the pro-R and
In conclusion, we have achieved a new asymmetric
synthesis of 2-substituted chiral chromanes. The key
stepis the intramolecular conjugate addition of a phe-
nolic nucleophile on a a,b-unsaturated ester. To the best
of our knowledge, it is the first example of asymmetric
catalysis of an exo-trig22 cyclization for an oxygen
nucleophile. An extensive study of the effect of the cat-
alyst’s structure on the enantioselectivity of a model
reaction was achieved. We have shown that the highest
ee’s are obtained with the readily available Cinchona
alkaloids and some of their derivatives. Finally, the
experimental results have been rationalized by ab initio
quantum chemistry calculations of transition state
structures.
Acknowledgements
Dr. Georges Dive is Research Associate of the F.N.R.S.
and is thankful for PAI no P5/33.
ꢂ
pro-S TS I are, respectively, ꢀ9.3 and ꢀ4.8 A. This TS is
References and notes
thus in agreement with the best ee obtained with the
more hindered C-11 substituted catalysts.
1. (a) Browers, W. S.; Ohta, T.; Cleere, J. S.; Marsella, P. A.
Science 1976, 193, 542–547; (b) Schweizer, E. E.; Meeder-
Nycz, O. In Chromenes, Chromanes and Chromones; Ellis,
G. P., Ed.; Wiley-Interscience: New York, 1977; pp 11–
In the other energetically favored TS (II) leading to the
R enantiomer, the hydroxy groupof cinchonine takes

A. Merschaert et al. / Tetrahedron Letters 45 (2004) 4697–4701
4701
139; (c) Ellis, G. P.; Lockhart, I. M. Chromanes and
Tocopherols; Wiley-Interscience: New York, 1981; (d)
Broggini, G.; Folcio, F.; Sardone, N.; Sonzogni, M.;
Zecchi, G. Tetrahedron: Asymmetry 1996, 7, 797–806, and
references cited therein.
obtained by refluxing 12 with NH2OHÆHCl in formic acid
for 1 h followed by deprotection as above. 27 was obtained
by decarbonylation of 12 with (Ph3P)3RhCl in toluene at
70 °C for 20 h followed by deprotection as above.
14. Introduction of electron-withdrawing groups ot carbon-3
has a very significant effect on quinuclidinic nitrogen
basicity. See: Drzewiczak, C.; Suszko-Purzycka, A.; Sko-
lik, J. Pol. J. Chem. 1993, 67, 45–52, Calculated (ACD/
Log D Suite software, Adv. Chem. Dev. Inc.) pKa values
for 27, 28, 29, and 30 are, respectively, 10.09, 9.29, 6.79,
and 9.02 ( 0.4).
15. The Z methyl ester 3a was selected as it is readily obtained
by reaction of 1 with bis(2,2,2-trifluoroethyl)(methoxycar-
bonyl-methyl)phosphonate according to Still, W. C.;
Gennari, C. Tetrahedron Lett. 1983, 24, 4405–4408, It
offered also the best peaks resolution in our chiral HPLC
method.
2. Antoine, L.; Boquel, P.; Borghese, A.; Gorissen, H.;
Khau, V. V.; Martinelli, M.; Merschaert, A.; Ruhter, G.;
Rypens, C. PCT Int. Appl. WO 094334, 2001; Chem.
Abstr. 2001, 136, 37512.
3. Merschaert, A. Ph.D. Thesis, University of Brussels, 2001.
Bases: ephedrine and derivatives (4–34% ee), (S)-prolinol
and derivatives (6–36% ee), cis-1-ethyl-4-hydroxy-L-pro-
linol (20% ee), (R)-3-hydroxyquinuclidine (ꢀ0% ee). Sol-
vents: toluene–chlorobenzene ꢁ THF–dioxane ꢁ EtOH.
4. The E isomer was not detected after refluxing the O-acetyl
Z a,b-unsaturated ester with cinchonine in toluene for
48 h.
5. Neither ring-opening products nor loss of enantiomeric
purity was observed after refluxing the chromane with
cinchonine in toluene for 48 h.
6. Hiemstra, H.; Wijnberg, H. J. Am. Chem. Soc. 1981, 103,
417–430.
16. These conditions were selected based on our initial studies.
We often observed significant decrease of ee above 100 °C.
17. Calculated (ACD) pKa values: 9.99 for 24 and 9.18 for 5.
18. (a) Dijkstra, G. D. H.; Kellogg, R. M.; Wynberg, H.;
Svendsen, J. S.; Marko, I.; Sharpless, K. B. J. Am. Chem
Soc. 1989, 111, 8069–8076; (b) Dijkstra, G. D. H.;
Kellogg, R. M.; Wynberg, H. J. Org. Chem. 1990, 55,
6121–6131; (c) Dijkstra, G. D. H.; Kellogg, R. M.;
Wynberg, H. Rec. Trav. Chim. Pays-Bas 1989, 108, 195–
204; (d) Oleksyn, B. J.; Suszko-Purzycka, A.; Dive, G.;
Lamotte-Brasseur, J. J. Pharm. Sci. 1992, 81, 122–127.
7. Kowalik, J. T.; Lipinska, T.; Oleksyn, B. J.; Sliwinski, J.
Enantiomer 1999, 4, 389–410.
€
8. Lammerhofer, M.; Maier, N. PCT Int. Appl. WO 46557,
1997.
9. Obtained by Pd-catalyzed hydrogenolysis of O-tosyl-
cinchonine prepared as described in Ref. 7.
€
10. (a) Braje, W.; Frackenpohl, J.; Langer, P.; Hoffmann, H.
M. R. Tetrahedron 1998, 54, 3495–3512; (b) Langer, P.;
Hoffmann, H. M. R. Tetrahedron 1997, 53, 9145–9158; (c)
Schrake, O.; Braje, W.; Hoffmann, H. M. R.; Wartchow,
R. Tetrahedron: Asymmetry 1998, 9, 3717–3722; (d)
Schrake, O.; Rahn, V. S.; Frackenpohl, J.; Braje, W. M.;
Hoffmann, H. M. R. Org. Lett. 1999, 1, 1607–1610; (e)
19. (a) Burgi, T.; Baiker, A. J. Am. Chem. Soc. 1998, 120,
12920–12926; (b) Berg, U.; Aune, M.; Matsson, O.
Tetrahedron Lett. 1995, 36, 2137–2140.
20. Uccello-Barretta, G.; Balzano, F.; Quintavalli, C.; Salva-
dori, P. J. Org. Chem. 2000, 65, 3596–3602.
21. The calculations were performed with GAUSSIAN94. For
all the systems, a full geometry optimization was per-
formed at the Hartree–Fock level within the minimal basis
set MINI-10. For the starting models involving water or
simple amino-alcohols as catalysts, a full optimization
within the double 6-31G basis set was also performed. The
transition states (TS) were first optimized and the related
minima were obtained by following the intrinsic reaction
coordinate (IRC) from this TS. For details on similar
approaches, see: Dive, G.; Dehareng, D. Int. J. Quant.
Chem. 1999, 73, 161–174.
€
Arrington, M. P.; Bennani, Y. L.; Gobel, T.; Walsh, P.;
Zhao, S. H.; Sharpless, K. B. Tetrahedron Lett. 1993, 34,
7375–7378.
11. Despite using the optimized conditions described in Ref.
10a, we obtained upto 20 mol % of the unwanted
3-epimer.
12. A large excess of the Wittig reagent is required to achieve
complete conversion of the aldehyde.
13. Catalyst 28 was prepared by reduction of 12 with NaBH4
in methanol followed by deprotection with K2CO3. 29 was
22. Baldwin, J. J. Chem. Soc., Chem. Commun. 1976, 734–737.