Preparation of 1,4-dihydroquinolines bearing a chiral sulfoxide group: New highly enantioselective recyclable NADH mimics

Article abstract of DOI:10.1055/s-2005-862350

The stereoselective preparation of 1-,4-dihydroquinolines possessing a chiral sulfoxide group at C-3 is reported. These novel biomimetic NADH models (R)-1a,b have been shown to be highly enantioselective in the reduction of methyl benzoylformate, producing (R)-methyl mandelate in up to 95% ee. The corresponding quinolinium salts 4a,b have been recovered in good yields. The regenerated models 1a,b could be reused without any significant erosion of the enantioselectivity. Georg Thieme Verlag Stuttgart.

Full text of DOI:10.1055/s-2005-862350

LETTER
441
Preparation of 1,4-Dihydroquinolines Bearing a Chiral Sulfoxide Group:
New Highly Enantioselective Recyclable NADH Mimics
E
nantioselecti
t
ve Rec
é
yclable
N
A
p
D
HMimics hane Gaillard, Cyril Papamicaël, Francis Marsais, Georges Dupas, Vincent Levacher*
Laboratoire de Chimie Organique Fine et Hétérocyclique, UMR 6014, IRCOF, CNRS, Université et INSA de Rouen, B.P. 08,
76131 Mont Saint Aignan Cédex, France
Fax +33(2)35522962; E-mail: vincent.levacher@insa-rouen.fr
Received 22 November 2004
We first focused on the preparation of 1a by halogen–
metal exchange reaction of 3-bromoquinoline (2a) fol-
lowed by treatment of the resulting 3-metallated quinoline
with (1R,2S,5R)-(–)-(S)-menthyl p-toluenesulfinate.
Abstract: The stereoselective preparation of 1,4-dihydroquinolines
possessing a chiral sulfoxide group at C-3 is reported. These novel
biomimetic NADH models (R)-1a,b have been shown to be highly
enantioselective in the reduction of methyl benzoylformate, produc-
ing (R)-methyl mandelate in up to 95% ee. The corresponding quin- Preliminary attempts to achieve bromide–magnesium
olinium salts 4a,b have been recovered in good yields. The
regenerated models 1a,b could be reused without any significant
erosion of the enantioselectivity.
exchange with isopropylmagnesium bromide at different
temperatures (–78 °C to 0 °C) failed. In any case, 3-bro-
moquinoline and (S)-menthyl p-toluenesulfinate were
recovered together with isopropyl phenyl sulfoxide. The
desired (R)-3-p-tolyl sulfoxide quinoline (3a) was finally
obtained in modest yield (30%) and excellent stereo-
selectivity (ee >95%)3,4 by lithium–bromide exchange
using t-BuLi in Et2O at –78 °C (Scheme 1). Attempts to
improve the yield by means of ate complexes5 afforded 3a
in somewhat higher yield (55%). However, the stereo-
selectivity is significantly affected under these conditions,
limiting seriously the synthetic interest of these ate com-
plexes in the stereoselective preparation of sulfoxides
from chiral sulfinates esters (Scheme 1).
Key words: chiral sulfoxide, NADH models, quinoline, asymmet-
ric reduction
Biomimetic NADH models possessing a sulfoxide group
as a chiral inducer have proved to be highly efficient, al-
lowing the reduction of various prochiral substrates with
a high level of induction.1 The main drawback of these co-
enzyme mimics is their poor stability towards nucleo-
philes and water, making them difficult to handle and
hampering the possibility of being used in catalytic pro-
cess. To circumvent this problem and further extend the
potential of these mimics, we report herein the synthesis
of annelated models 1a,b in the quinoline series.2 The
presence of an additional fused phenyl ring is thereby ex-
pected to improve the stability of the reagent by masking
the reactivity of the enamine double bond (Figure 1).
lithium–bromide
exchange
O
Br
S
(a)
p-Tol
30%
ee >95%
N
N
H O
H
¨
(R)-3a
2a
S
p-Tol
p-Tol
R1COR2
(b)
50%
ee = 50%
Mg
Li
N
R
O
side-reactions
¨
HO
R1
H
N
3
S
R2
ate complexe
high ee
N
R
Scheme 1 Reagents and conditions: (a) t-BuLi, Et2O, –78 °C, then
(1R,2S,5R)-(–)-(S)-menthyl p-toluenesulfinate; (b) Bu3MgLi, toluene,
–10 °C, 2.5 h, then (1R,2S,5R)-(–)-(S)-menthyl p-toluenesulfinate.
H O
¨
S
H
R
R
annelated
models
p-Tol
N
1a R = H
1b R = OMe
Quaternization of (R)-3a with methyl triflate in CH2Cl2
afforded quinolinium salt (R)-4a6 in quantitative yield. At
this stage, first attempts to reduce regioselectively quino-
linium salt (R)-4a under classical conditions (Na2S2O4/
Na2CO3) failed, resulting in the formation of tarry materi-
als presumably due to desulfination side reactions.7 While
the reduction conducted with NaBH4 gave rise to an
inseparable mixture of 1,2- and 1,4-regioisomers, we were
pleased to observe the exclusive formation of 1,4-di-
Me
Figure 1 Simple biomimetic NADH models bearing a sulfoxide
group as chiral inducer and their annelated analogues.
SYNLETT 2005, No. 3, pp 0441–0444
Advanced online publication: 17.01.2005
DOI: 10.1055/s-2005-862350; Art ID: D34304ST
1
6.
0
2.
2
0
0
5
© Georg Thieme Verlag Stuttgart · New York

442
S. Gaillard et al.
LETTER
H
H O
gen exchange reaction took place smoothly to furnish the
corresponding Grignard intermediate which was sub-
sequently quenched with MeOD. An optimization of the
reaction conditions showed that 3 equivalents of isopro-
pylmagnesium bromide are required to drive the reaction
to completion, providing 3-deuterated-6,7-dimethoxy-
quinoline in 100% yield. Treatment of the Grignard
intermediate with (1R,2S,5R)-(–)-(S)-menthyl p-toluene-
sulfinate resulted in the formation of (R)-7b13 in 40%
yield as a single enantiomer (ee >98%).14 Quinoline 7b
was further converted into 1,4-dihydroquinoline 1b16 by
quaternization to give 8b15 which was regioselectively
reduced with BNAH (Scheme 4).
O
S
S
p-Tol
p-Tol
N
N
O
(R)-3a
Me
S
(R)-1a
p-Tol
(a)
100%
(b)
95%
N
TfO
(R)-4a
Me
Scheme 2 Reagents and conditions: (a) MeOTf, CH2Cl2, 20 °C, 2 h;
(b) BNAH, CH2Cl2, 20 °C, 2 h.
hydroquinoline 1a8 in high yield and under mild condi-
tions by using N-benzyl-1,4-dihydronicotinamide
(BNAH, Scheme 2).
O
MeO
MeO
Br
MeO
MeO
S
(a)
We then turned our attention to the preparation of model
1b. We anticipated that the presence of two strong elec-
tron-donating methoxy groups might increase the reduc-
ing properties of the 1,4-dihydroquinoline ring and
thereby should improve the performance of these annelat-
ed models. To this end, various methods were investigated
for the synthesis of the new 3-bromo-6,7-dimethoxyquin-
oline 2b (Scheme 3). We first considered the possibility to
have access to this key intermediate 2b via a Curtius re-
arrangement of the known carboxylic acid 5b.9 This was
accomplished by treatment of the former with diphe-
nylphosphorazide (DPPA) in refluxing t-BuOH affording
6b in 41% yield. Diazotization of 6b with HBr–NaNO2
followed by addition of molecular bromine provided the
required key intermediate 2b,10 however, in modest yield
(20%). In search of a more straightforward route, we next
investigated an alternative approach by construction of
the quinoline ring from 3,4-dimethoxyaniline and an ap-
propriate 1,3-dielectrophile such as a-bromoacrolein11 or
2-bromomalondialdehyde.12 In both cases, bromoquino-
line 2b was isolated in 30% yield. Although the yield re-
mains rather modest, this approach presents the advantage
of furnishing the key intermediate 2b in a single step
process from readily available reagents (Scheme 3).
p-Tol
40%
N
N
2b
(R)-7b
(b)
100%
H
H O
O
MeO
S
MeO
MeO
S
p-Tol
p-Tol
(c)
95%
MeO
N
N
(R)-1b
Me
TfO
(R)-8b
Me
Scheme 4 Reagents and conditions: (a) i-PrMgCl (3 equiv), THF,
20 °C, 3 h, then (1R,2S,5R)-(–)-(S)-menthyl p-toluenesulfinate, 24 h;
(b) MeOTf, CH2Cl2, 20 °C, 2 h; (c) BNAH, CH2Cl2, 20 °C, 2 h.
Performances of models 1a,b were assessed in the course
of the reduction of methyl benzoylformate (Scheme 5).
For comparison purpose with literature data,1 models
were tested under standard conditions, i.e. in the presence
of magnesium perchlorate in acetonitrile.9 Under these
conditions, both models 1a,b produced (R)-methyl man-
delate in 50% and 70% yield, respectively. As anticipated
above, the presence of electron donating groups in 1b im-
proves to some extent the reducing properties of the mod-
el. More importantly, one may notice that the level of
enantioselectivity of these new 1,4-dihydroquinolines is
as high as those observed with non-annelated NADH
models, affording (R)-methyl mandelate in up to 95% ee.
One can conclude that an additional phenyl ring does not
affect the high degree of stereocontrol already observed
with dihydropyridine NADH mimics developed earlier.1
Additionally, this annelation process offers the main
advantage of stabilizing the reagent and hence to recycle
the models. Thus, quinolinium salts (R)-4a and (R)-8b
recovery was accomplished in 90% and 85% yields, re-
spectively, after flash chromatography. Both recovered
quinolinium salts were subsequently reduced to give
models 1a,b17 in 95% yields. The recycled reagents were
involved in the reduction of methyl benzoylformate with-
out significant lost of their performances in terms of reac-
tivity and enantioselectivity. One attractive perspective of
this work is the potential for immobilization of these
reagents on insoluble support, which should make
recycling easier (Scheme 5).
In contrast to 3-bromoquinoline, when 3-bromo-6,7-
dimethoxyquinoline (2b) was treated with isopropylmag-
nesium bromide in THF at room temperature, metal–halo-
MeO
MeO
COOH
MeO
MeO
NHCOOt-Bu
(a)
41%
N
N
5b
6b
Br
CHO
(b)
20%
(c)
30%
MeO
MeO
MeO
MeO
Br
(d)
O
30%
NH2
N
Br
2b
OH
Scheme 3 Reagents and conditions: (a) (PhO)2PON3, Et3N, t-
BuOH, reflux, 24 h; (b) HBr (48%), Br2, NaNO2, –5 °C, 1 h; (c) Br2,
glacial acetic acid, reflux 1 h; (d) EtOH, HCl, reflux, 12h.
Synlett 2005, No. 3, 441–444 © Thieme Stuttgart · New York

LETTER
Enantioselective Recyclable NADH Mimics
443
(6) Analytical data for (R)-4a: 1H NMR (300 MHz, CDCl3): d =
2.39 (3 H, s), 4.73 (3 H, s), 7.43 (2 H, d, J = 8 Hz), 7.81 (2
H, d, J = 8 Hz), 8.12 (1 H, t, J = 9 Hz), 8.36 (1 H, t, J = 9
Hz), 8.52 (2 H, m), 9.42 (1 H, s), 9.62 (1 H, s). 13C NMR (75
MHz, MeOD): d = 21.8, 47.5, 120.6, 124.3, 127.2, 131.1,
132.4, 132.8, 132.9, 139.2, 141.3, 141.7, 142.8, 144.9,
145.8, 147.8. 19F NMR (282 MHz, CDCl3): d = –80.5.
HRMS (CI): m/z calcd for C17H16NOS: 282.0953. Found:
282.0957.
H
H O
Ph
R
S
CONH2
O
p-Tol
MeOOC
R
N
N
Cl
Me
Me
BNA+
(R)-1a R = H
(R)-1b R = OMe
(b)
CONH2
(a)
O
H
H
R
S
(7) Similar side reactions have already been observed during the
reduction of 3-sulfoxide pyridinium salt under these
conditions. See: Imanishi, T.; Hamano, Y.; Yoshikawa, H.
T.; Iwata, C. J. Chem. Soc., Chem. Commun. 1988, 473.
(8) Analytical data for (R)-1a: 1H NMR (300 MHz, CDCl3): d =
2.76 (3 H, s), 3.45 (1 H, d, J = 19 Hz), 3.59 (3 H, s), 4.13 (1
H, d, J = 19 Hz), 7.05 (1 H, d, J = 8 Hz), 7.27 (2 H, d, J = 8
p-Tol
Ph
H
R
N
OH
N
MeOOC
TfO
Me
Me BNAH
(R)-3a R = H
(R)-3b R = OMe
(R)-1a: yield = 50%; ee = 95% (R)
(R)-1b: yield = 70%; ee = 95% (R)
Hz), 7.47 (1 H, m), 7.65 (3 H, m), 7.87 (2 H, d, J = 8Hz). 13
C
Scheme 5 Reagents and conditions: (a) Mg(ClO4)2, MeCN, r.t., 24
h; (b) MeOTf, CH2Cl2, 20 °C, 2 h; (c) BNAH, CH2Cl2, 20 °C, 2 h.
NMR (75 MHz, CDCl3): d = 21.7, 23.2, 38.9, 109.7, 112.9,
121.7, 123.3, 125.3, 127.7, 130.0, 130.2, 139.2, 139.3,
140.7, 140.8. HRMS (CI): m/z calcd for C17H17NOS:
283.1031. Found: 283.1035.
Acknowledgment
(9) Charpentier, P.; Lobrégat, V.; Levacher, V.; Dupas, G.;
Quéguiner, G.; Bourguignon, J. Tetrahedron Lett. 1998, 39,
4013.
We thank the CNRS, the région Haute-Normandie and the
CRIHAN for financial and technical support.
(10) Analytical data for 2b: 1H NMR (300 MHz, CDCl3): d = 3.98
(3 H, s), 4.00 (3 H, s), 6.92 (1 H, s), 7.34 (1 H, s), 8.10 (1 H,
s), 8.68 (1 H, s). 13C NMR (75 MHz, CDCl3): d = 56.5, 56.6,
104.5, 108.3, 115.6, 125.3, 135.8, 143.9, 149.2, 150.9,
153.0. Anal. Calcd for C11H10BrNO2: C, 49.28; H, 3.76; N,
5.22. Found: C, 49.35; H, 3.82; N, 5.21.
(11) Smith, A. B.; Levenberg, P. A.; Jerris, P. J.; Scarborough, R.
M.; Wovkulich, P. M. J. Am. Chem. Soc. 1981, 103, 1501.
(12) Trofimenko, S. J. Org. Chem. 1963, 28, 3243.
(13) Analytical data for (R)-7b: 1H NMR (300 MHz, CDCl3): d =
2.29 (3 H, s), 3.93 (6 H, s), 7.02 (1 H, s), 7.20 (2 H, d, J = 8
Hz), 7.33 (1 H, s), 7.51 (2 H, d, J = 8 Hz), 8.29 (1 H, s), 8.62
(1 H, s). 13C NMR (75 MHz, CDCl3): d = 21.8, 56.6, 56.7,
105.8, 108.3, 123.6, 125.4, 130.6, 131.4, 137.5, 142.1,
142.5, 144.2, 146.7, 151.1, 154.3. HRMS: m/z calcd for
C18H17NO3S: 327.0929. Found: 327.0933..
References
(1) For leading references of NADH models bearing a sulfoxide
group, see: (a) Li, J.; Liu, Y.-C.; Deng, J.-G. Tetrahedron:
Asymmetry 1999, 10, 4343. (b) Kazuyuki, M.; Nishimoto,
N.; Hidenobu, M.; Asuka, M.; Satoshi, O.; Yasuko, I.;
Toshimasa, I.; Takeshi, I. Chem. Commun. 1996, 2535.
(c) Obika, S.; Nishiyama, T.; Tatematsu, S.; Miyashita, K.;
Iwata, C.; Imanishi, T. Tetrahedron 1997, 53, 593.
(d) Obika, S.; Nishiyama, T.; Tatematsu, S.; Miyashita, K.;
Imanishi, T. Chem. Lett. 1996, 853. (e) Obika, S.;
Nishiyama, T.; Tatematsu, S.; Nishimoto, M.; Miyashita, K.;
Imanishi, T. Heterocycles 1998, 261. (f) Imanishi, T.;
Obika, T.; Nishiyama, T.; Nishimoto, M.; Hamano, Y.;
Miyashita, K.; Iwata, C. Chem. Pharm. Bull. 1996, 44, 267.
(2) For early reports on annelated NADH models, see:
(a) Levacher, V.; Dupas, G.; Quéguiner, G.; Bourguignon, J.
Trends Heterocycl. Chem. 1995, 4, 293. (b) Dupas, G.;
Levacher, V.; Quéguiner, G.; Bourguignon, J. Heterocycles
1994, 39, 405. (c) Vitry, C.; Vasse, J.-L.; Levacher, V.;
Dupas, G.; Quéguiner, G.; Bourguignon, J. Tetrahedron
2001, 57, 3087.
(14) Enantiomeric excesses were determined by HPLC analysis
using a Chiralpak AD column (250 × 4.6 mm; 10 mm).
Chromatographic conditions: eluent: heptane–2-propanol =
85:15; flow rate: 1 mL min–1; pressure: 300 psi; temperature:
19 °C; UV detection: l = 230 nm; tR = 36 min [(S)-
enantiomer] and 40 min [(R)-enantiomer].
(15) Analytical data for (R)-8b: 1H NMR (300 MHz, MeOD): d =
2.39 (3 H, s), 4.06 (3 H, s), 4.20 (3 H, s), 4.62 (3 H, s), 7.42
(2 H, d, J = 8Hz), 7.61 (1 H, s), 7.76 (3 H, m), 9.09 (1 H, s),
9.28 (1 H, s). 13C NMR (75 MHz, MeOD): d = 26.7, 51.7,
62.4, 63.3, 130.7, 113.6, 130.8, 132.3, 136.5, 143.7, 143.9,
144.8, 145.5, 146.6, 149.4, 158.6, 165.2. 19F NMR (282
MHz, CDCl3): d = – 80.25. HRMS: m/z calcd for
(3) Analytical data for (R)-3a: 1H NMR (300 MHz, CDCl3): d =
2.30 (3 H, s), 7.21 (2 H, d, J = 8 Hz), 7.51 (3 H, m), 7.73 (1
H, m), 7.85 (1 H, d, J = 8 Hz), 8.05 (1 H, d, J = 8 Hz), 8.53
(1 H, s), 8.76 (1 H, s). 13C NMR (75 MHz, CDCl3): d = 21.8,
125.6, 127.7, 128.3, 128.8, 129.9, 130.7, 131.7, 133.2,
139.5, 141.7, 142.9, 146.2, 149.1. Anal. Calcd for
C16H13NOS: C, 71.88; H, 4.90; N, 5.24. Found: C, 71.78; H,
4.82; N, 5.10.
(4) Enantiomeric excesses were determined by HPLC analysis
using a Chiracel OJ column (250 × 4.6 mm; 10 mm).
Chromatographic conditions: eluent: heptane–2-propanol =
90:10; flow rate: 1 mL min–1; pressure: 300 psi; temperature:
19 °C; UV detection: l = 230 nm; tR: 22 min [(S)-
enantiomer] and 26 min [(R)-enantiomer].
C19H20NO3S: 342.1164. Found: 342.1165.
(16) Analytical data for (R)-1b: 1H NMR (300 MHz, CDCl3): d =
2.32 (3 H, s), 2.94 (1 H, d, J = 18 Hz), 3.17 (3 H, s), 3.63 (1
H, d, J = 18 Hz), 3.67 (3 H, s), 3.78 (3 H, s), 6.24 (1 H, s),
6.33 (1 H, s), 6.84 (1 H, s), 7.21 (2 H, d, J = 8Hz), 7.43 (2 H,
d, J = 8Hz). 13C NMR (75 MHz, CDCl3): d = 21.8, 21.9,
39.2, 56.5, 56.6, 98.6, 108.5, 113.1, 113.4, 125.4, 130.0,
132.7, 139.2, 140.6, 140.9, 154.2, 148.3. HRMS (CI):
m/z calcd for C19H21NO3S: 343.1242. Found: 343.1249.
(17) Typical Procedure for the Reduction of Methyl
Benzoylformate with Mimics 1a,b:
(5) (a) Dumouchel, S.; Mongin, F.; Trécourt, F.; Quéguiner, G.
Tetrahedron Lett. 2003, 44, 2033. (b) Dumouchel, S.;
Mongin, F.; Trécourt, F.; Quéguiner, G. Tetrahedron 2003,
44, 8629.
In a flask, flushed with argon, were introduced model 1b
(0.283 g, 1 mmol), MeCN (3 mL), methyl benzoylformate
(142 mL, 1 mmol) and Mg(ClO4)2 (224 mg, 1 mmol). The
Synlett 2005, No. 3, 441–444 © Thieme Stuttgart · New York

444
S. Gaillard et al.
LETTER
resulting solution was stirred at r.t. for 24 h in the dark. After
addition of H2O (10 mL), the organic solvent was evaporated
under reduced pressure and the resulting aqueous phase was
extracted with CH2Cl2 (3 × 10 mL). After drying (MgSO4)
and evaporation of the solvent, the residue was purified by
chromatography on silica gel (eluent: Et2O–cyclohexane =
2:1). Yield: 50%. Enantiomeric excesses were determined
by HPLC analysis using a Chiracel OD column (250 × 4.6
mm; 10 mm). Chromatographic conditions: injection: 20 ml
(0.5 mg of methyl mandelate in 10 mL of hexane); eluent:
hexane–2-propanol = 90:10; flow rate: 1 mL min–1; pressure:
300 psi; temperature: 22 °C; UV detection: l = 235 nm; tR =
9.2 min [(S)-enantiomer] and 14.8 min [(R)-enantiomer].
Synlett 2005, No. 3, 441–444 © Thieme Stuttgart · New York