An Enantioselective Access to 1-Alkyl-1,2,3,4-tetrahydroisoquinolines. Application to a New Synthesis of (-)-Argemonine

Article abstract of DOI:10.1021/jo0357869

Potassium ferricyanide oxidation of salt 1 gave isoquinolinone 7 whose treatment with Grignard reagents resulted in a high-yield formation of substituted isoquinolinium salts 5. The selectivity of the reduction of these salts to give derivatives 6 has been studied. Particularly good selectivities (82-84%) were observed when R is a benzylic group. On the basis of these results, a practical and enantioselective synthesis of the natural alkaloid (-)-argemonine is presented.

Full text of DOI:10.1021/jo0357869

An En a n tioselective Access to
1-Alk yl-1,2,3,4-tetr a h yd r oisoqu in olin es. Ap p lica tion to a New
Syn th esis of (-)-Ar gem on in e†
J ean-J acques Youte,‡ Denis Barbier,‡ Ali Al-Mourabit,‡ Dino Gnecco,§ and
Christian Marazano*,‡
Institut de Chimie des Substances Naturelles, CNRS, 1 Avenue de la Terrasse, 91198 Gif-sur-Yvette,
France, and Centro de Quimica del Instituto de Ciencias, BUAP Cd., Universitaria Puebla,
Pue. C. P. 72570, Me´xico
marazano@icsn.cnrs-gif.fr
Received December 5, 2003
Potassium ferricyanide oxidation of salt 1 gave isoquinolinone 7 whose treatment with Grignard
reagents resulted in a high-yield formation of substituted isoquinolinium salts 5. The selectivity of
the reduction of these salts to give derivatives 6 has been studied. Particularly good selectivities
(82-84%) were observed when R is a benzylic group. On the basis of these results, a practical and
enantioselective synthesis of the natural alkaloid (-)-argemonine is presented.
We recently reported1 that isoquinolinium salts 1, now
readily available using the Zincke procedure,2 can be
alkylated by Grignard reagents to give, after reduction,
chiral 1-alkyl-1,2,3,4-tetrahydroisoquinolines 2. This pro-
cess offers the advantage of starting from readily avail-
able 1-phenylethylamine as a chiral auxiliary, but is
limited in some cases by the relatively poor diastereose-
lectivity of the Grignard attack, ranging from 28% to
80%. Use of a phenylethanol chiral auxiliary (salt 3)
resulted in increased selectivities (38-90%) in favor of
derivatives 4. In both cases the lowest selectivities, 32%
or less, were observed using benzylic Grignard reagents.
To improve the efficiency of this approach, we anticipated
that reduction of isoquinolinium salts of general structure
5 would give a more selective access to tetrahydroiso-
quinolines 6. Indeed, this reaction would take advantage
of 1,3-allylic strain effects in salts 5. These effects are
likely to block the rotation of the chiral auxiliary, fixing
the R group and the C-H bond in a syn relationship for
steric reasons, and thus differentiating the two diastereo-
topic faces of the isoquinolinium ring. Due to these effects
(attack of the hydride from the less hindered side)
formation of isoquinolines 6 as major isomers was
expected.3
to 84%.4 The selectivities now obtained with benzylic
Grignard reagents have allowed a new and practical
enantioselective synthesis of the natural pavine alkaloid
(-)-argemonine.5
Our approach was based on the facile oxidation of
isoquinolinium salts to isocarbostyril derivatives using
potassium ferricyanide oxidation6,7 (Scheme 1).
Thus, treatment of salt 1 with potassium ferricyanide
in alkaline medium afforded isoquinolinone 7 in 77%
yield. We observed that heating of salt 1 in the presence
In this paper we report a convenient access to salts 5
and their stereoselective reduction to tetrahydroisoquino-
lines 6 with diastereoisomeric excesses ranging from 74%
(4) For a related approach see: Kaufmann, C. R.; Polniaszek, R. P.
J . Am. Chem. Soc. 1989, 111, 4859-4863.
† This article is dedicated to the memory of Franc¸ois Frappier,
Directeur de recherche au CNRS, Muse´um National d'Histoire Na-
turelle, Paris, who passed away on February 23, 2003.
‡ Institut de Chimie des Substances Naturelles.
§ Centro de Quimica.
(5) For a previous enantioselective synthesis of (-)-argemonine
see: Munchhof, M. J .; Meyers, A. I. J . Org. Chem. 1996, 61, 4607-
4610. For racemic syntheses see: Orito, K.; Satoh, Y.; Nishizawa, H.;
Harada, R.; Tokuda, M. Org. Lett. 2000, 2, 2535-2537. Ruchirawat,
S.; Namsa-Aid, A. Tetrahedron Lett. 2001, 42, 1359-1361 and refer-
ences therein.
(1) Barbier, D.; Marazano, C.; Riche, C.; Das, B. C.; Potier, P. J .
Org. Chem. 1998, 63, 1767-1772.
(2) Barbier, D.; Marazano, C.; Das, B. C.; Potier, P. J . Org. Chem.
1996, 61, 9596-9598.
(6) Brown, D. B.; Dyke, S. F. Tetrahedron 1966, 22, 2429-2435.
(7) For previous use of this oxidation process in the pyridinium series
see: Gnecco, D.; Marazano, C.; Enr´ıquez, R.; Tera´n, J .; Sa´nchez, S.
M.; Galindo, A. Tetrahedron: Asymmetry 1998, 9, 2027-2029.
(3) For a review of 1,3-allylic strain effects see: Hoffmann, R. W.
Chem. Rev. 1989, 89, 1841-1860.
10.1021/jo0357869 CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/10/2004
J . Org. Chem. 2004, 69, 2737-2740
2737

Youte et al.
SCHEME 2. Syn th esis of (-)-Ar gem on in ea
SCHEME 1. P r ep a r a tion a n d Ster eoselective
Red u ction of Isoqu in olin iu m Sa lts 5a
a
Reagents and conditions: (a) K3Fe(CN)6, MeOH, KOH, 1 h, 0
°C, then H2O-toluene, 45 °C. (b) (i) CeCl3, toluene; (ii) RMgX,
THF; (iii) HBr, H2O. (c) NaBH4, MeOH.
TABLE 1. Syn th esis a n d Ster eoselective Red u ction of
Sa lts 5
a
% yield of diastereoisomeric mixtures of 6-8 after filtration
over alumina. Calculated from 1H NMR and GC analysis of the
b
crude reaction mixture.
of potassium hydroxide resulted in rapid racemization.
Fortunately, prior treatment with potassium ferricyanide
allowed recovery of product 7 with excellent enantiomeric
purity (>98% ee) as shown by HPLC on a chiral column
(Chiracel OD).
The results of Grignard reactions of derivative 7 in the
presence of CeCl3 followed by treatment with HBr to give
salts 6 are presented in Table 1.8
a
Reagents and conditions: (a) K3Fe(CN)6, MeOH, KOH, 1 h, 0
°C, then H2O-toluene, 45 °C (62% yield). (b) (i) CeCl3, toluene;
(ii) (MeO)2PhCH2MgCl, THF; (iii) HBr, H2O (92% yield). (c) LiAlH4,
THF, -78 °C (77% yield for (-)-13a ). (d) HCO2H, H3PO4 (77%
yield). (e) H2, Pd/C, H+ (77% yield). (f) HCHO, NaBH4 (95% yield).
Major isoquinoline derivatives 6a -c,e and minor iso-
quinolines 8a -c,e were characterized by comparison
with the enantiomeric authentic samples previously
obtained from the addition of the same Grignard reagents
on salt 1.1
This new enantioselective and efficient approach to
1-benzyl isoquinoline derivatives was used for a synthesis
of the natural pavine alkaloid (-)-argemonine according
to Scheme 2.
Potassium ferricyanide oxidation of salt (-)-92 afforded
isoquinolinone (-)-10, which was recovered in 62% yield.
Treatment of the latter with 3,4-dimethoxybenzylmag-
nesium chloride under the conditions used for prepara-
tion of salts 5a -e then gave salt 11 in excellent yield.
The reduction of salt 11 to give the dihydro derivatives
12a and 12b was accomplished at -78 °C, using LiAlH4
in THF. The ratio of isomers 12a /12b was estimated to
From these results it can be concluded that, in the case
of methyl or phenyl Grignard reagents, the observed
selectivities are in a similar range compared to those
obtained from salts 1 or 3, using the same reagents (74%
selectivity for 6a and 6b compared to 28% (R ) Me) and
74% (R ) Ph) for the corresponding derivatives 2 and
76% (R ) Me) and 90% (R ) Ph) for the corresponding
derivatives 4). By contrast these selectivities are signifi-
cantly higher in the case of benzylic Grignard reagents.
For example, preparation of isomer 6c proceeded with
an excess of 84% while the selectivities observed for the
corresponding derivatives 2 (R ) (3-OMe)-Ph-CH2-) and
4 (R ) (3-OMe)-Ph-CH2-) were as low as 24% and 38%,
respectively.
1
be 93/7 by H NMR spectroscopy but the instability of
these compounds precluded further studies. For this
reason these derivatives were cyclized directly in acidic
medium9 to give stable pavine derivatives (-)-13a and
(+)-13b. The ratio of these isomers was confirmed to be
93/7 by 1H NMR spectroscopy and comparison with
authentic samples (vide infra). The structure of the major
(8) Experimental details for the preparation of isomer 6d are
presented as a typical procedure.
(9) Battersby, A. R.; Binks, R. J . Chem. Soc. 1955, 2888-2896.
2738 J . Org. Chem., Vol. 69, No. 8, 2004

New Synthesis of (-)-Argemonine
SCHEME 3 a
Pure isoquinolinone 10 was obtained as a pale yellow oil (0.31
g, 1 mmol, 62% yield): [R]D -313 (c 2.3, CHCl3); IR (CHCl3,
cm-1) 1649, 1596; 1H NMR (CDCl3, 300 MHz) δ 1.76 (d, J ) 7
Hz, 3H), 3.98 (s, 3H), 4.02 (s, 3H), 6.39 (d, J ) 7.5 Hz, 1H),
6.58 (q, J ) 7 Hz, 1H), 6.84 (s, 1H), 6.88 (d, J ) 7.5 Hz, 1H),
7.12 (m, 5H), 7.87 (s, 1H); 13C NMR (CDCl3, 100 MHz) δ 18.8,
52.2, 56.1, 56.2, 105.9, 106.0 (2C), 108.1, 120.0, 126.7, 127.4
(2C), 127.7, 128.7, 132.0, 140.8, 149.2, 153.4 (2C), 161.4; MS
(CI, isobutane) m/z 310 (MH+), 308, 206; HRMS (CI, isobutane)
calcd for C19H20NO3 (MH+) 310.1459, found 310.1444.
Isoqu in olin iu m Sa lt 11. A solution of isoquinolinone 10
(400 mg, 1.3 mmol) in dry toluene (5 mL) was treated with
cerium chloride (1.6 g, 4.2 mmol) and 3,4-dimethoxybenzyl-
magnesium chloride (0.11 M in THF, 47 mL) followed by HBr
(16% in H2O, 5 mL) in the conditions used for the preparation
of salt 5d . Salt 11 was obtained as a yellow powder (624 mg,
1.2 mmol, 92% yield): 1H NMR (CDCl3, 300 MHz) δ 2.02 (d, J
) 6.9 Hz, 3H), 3.84 (s, 6H), 4.09 (s, 3H), 4.18 (s, 3H), 5.37 (d,
J ) 17 Hz,1H), 5.44 (d, J ) 17 Hz, 1H), 6.37 (dd, J ) 1.7 Hz,
8 Hz, 1H), 6.48 (q, J ) 6.9 Hz, 1H), 6.74 (d, J ) 1.7 Hz, 1H),
7.07 (m, 2H), 7.14 (d, J ) 1.7 Hz, 1H), 7.32 (m, 3H), 7.69 (s,
1H), 7.85 (s, 1H), 8.17 (d, J ) 7 Hz, 1H), 8.54 (d, J ) 7 Hz,
1H); 13C NMR (CDCl3, 100 MHz) δ 21.3, 34.4, 55.8, 56.2, 56.8,
57.4, 64.2, 105.7, 106.9, 111.4, 111.9, 119.5, 124.1, 124.6, 127.0,
129.3 (5C), 131.1, 136.0 (2C), 136.7, 148.5, 153.5, 154.1, 157.7;
MS (electrospray) m/z 444 (M+).
Dih yd r oisoqu in olin e 12a . To a solution of salt 11 (200
mg, 0.38 mmol), suspended in THF (5 mL), was added
dropwise, at -78 °C, an excess of LiAlH4 (1 M solution in THF,
1.5 mL). After being stirred for 2 h at -78 °C, the resulting
mixture was quenched carefully with cold acetone (2 mL)
followed by NaOH (2 N H2O solution, 5 mL). The reaction
products were extracted with Et2O. Usual workup yielded a
a
Reagents and conditions: (a) (MeO)2PhCH2MgX, THF. (b)
HCO2H, H3PO4. (c) H2, Pd/C, H+. (d) HCHO, NaBH4.
mixture of derivatives 12a and 12b in
a 93/7 ratio as
determined by integration of methyl signals in the 1H NMR
spectrum of the crude mixture. These unstable dihydroiso-
quinolines were used without further purification: MS (CI,
isobutane) m/z 446 (MH+), 340, 294, 151. Major isomer 12a :
1H NMR (CDCl3, 300 MHz) δ 1.48 (d, J ) 6.9 Hz, 3H), 2.57
(dd, J ) 5.1, 12.6 Hz, 1H), 2.76 (dd, J ) 9.1, 12.6 Hz, 1H),
3.44 (s, 3H), 3.61 (s, 3H), 3.76 (s, 6H), 4.36 (m, 2H), 5.38 (d, J
) 7.2 Hz, 1H), 5.73 (s, 1H), 6.03 (dd, J ) 1.1, 7.1 Hz, 1H), 6.15
(d, J ) 1.7 Hz, 1H), 6.34 (dd, J ) 1.7, 8.1 Hz, 1H), 6.45 (s,
1H), 6.63 (d, J ) 8.1 Hz, 1H), 7.15 (m, 5H); 13C NMR (CDCl3,
100 MHz) δ 21.1, 38.9, 55.8, 55.8 (4 C), 61.4, 61.7, 98.5, 105.8,
110.8, 113.5, 120.6, 122.0, 125.3, 127.2, 128.5 (6C), 131.7,
131.1, 139.1, 143.3, 147.3, 147.4 (8C); MS (CI, isobutane) m/z
446 (MH+), 340, 294, 151.
Der iva tive (-)-13a . Crude dihydroisoquinoline 12a (pre-
pared from 200 mg of salt 11 and accompanied by 7% of isomer
13c), was dissolved in formic acid (85% in H2O, 5.3 mL) and
orthophosphoric acid (99%, 2 mL). This solution was heated
at 100 °C for 2 h. The resulting mixture was diluted with H2O
(10 mL) and washed twice with diethyl ether. The aqueous
phase was alkalinized with 2 N NaOH and extracted with
CHCl3. Removal of solvent under reduced pressure afforded a
mixture of (-)-13a accompanied by (+)-13b in a 92/8 ratio as
determined by GC analysis. Chromatography over silica gel
with a mixture of AcOEt/heptane (20/80) allowed recovery of
pure (-)-13a (130 mg, 0.29 mmol, 77% yield): [R]D -145 (c
1.4, CHCl3); 1H NMR (CDCl3, 300 MHz) δ 1.46 (d, J ) 6.5 Hz,
3H), 2.54 (d, J ) 16.1 Hz, 2H), 3.30 (dd, J ) 5.2, 16.1 Hz, 2H),
3.68 (q, J ) 6.5 Hz, 1H), 3.78 (s, 6H), 3.81 (s, 6H), 4.20 (d, J
) 5.2 Hz, 2H), 6.46 (s, 2H), 6.55 (s, 2H), 7.27 (m, 5H); 13C NMR
(CDCl3, 100 MHz) δ 22.5, 33.9 (2C), 52.4 (2C), 55.8 (2C), 56.1
(2C), 59.1, 110.4 (2C), 111.7 (2C), 127.0, 127.4, 128.5 (5C),
125.1 (2C), 130.5 (2C), 146.5, 147.6 (2C), 147.9 (2C); HRMS
(IE) calcd for C28H31NO4 445.2253, found 445.2238.
isomer 13a was unambiguously established after hydro-
genolysis of the chiral auxiliary to give secondary amine
(-)-14, which was methylated to finally afford natural
(-)-argemonine.
To unambiguously characterize isomers 12b and 13b
we also treated salt (+)-9 (Scheme 3) with 3,4-dimethoxy-
benzylmagnesium chloride under the conditions used in
our previous paper.1 The reaction afforded the enanti-
omer of 12b, ent-12b, as the major isomer but, as
expected, with poor selectivity (30% de).
Isomers (+)-13a and (-)-13b are difficult to separate
with classical chromatographic methods, but could be
separated with reverse-phase HPLC (eluent MeOH/H2O/
NEt3:60/40/0.3). Isomer (-)-13b gave as expected second-
ary amine (-)-14 and once again natural (-)-argemonine.
These last results illustrate the superiority of the
approach depicted in Scheme 2. While this approach
required two more steps compared to that depicted in
Scheme 3, this is largely compensated by the better
selectivities obtained.
The reported results completed our reported approach,
in particular good stereoselectivities are now available
for the enantioselective synthesis of 1-benzyl tetrahydro-
isoquinolines. This allowed a nonracemic synthesis of
natural pavine alkaloids.
Exp er im en ta l Section
Syn th esis of (-)-Ar gem on in e: 2-(1S)-(-)-2-(1-P h en yl-
eth yl)-6,7-d im eth oxyisoqu in olin on e 10. Isoquinolinium
salt 9 (0.55 g, 1.67 mmol), in MeOH (20 mL), was treated with
potassium ferricyanide (6 g, 18.4 mmol) and KOH (1.4 g, 26
mmol) followed by toluene (20 mL) and H2O (20 mL), following
the procedure used for the preparation of isoquinolinone 7.
A sample of isomer (-)-13b was prepared, according to the
procedure presented in Scheme 3, for authentication with the
above minor isomer (+)-13b. Isomer (-)-13b: [R]D -111 (c 1.2,
CHCl3); 1H NMR (CDCl3, 300 MHz) δ 1.37 (d, J ) 6.3 Hz, 3H),
J . Org. Chem, Vol. 69, No. 8, 2004 2739

Youte et al.
2.54 (d, J ) 16.2 Hz, 2H), 3.37 (dd, J ) 5.7, 16.3 Hz, 2H), 3.70
(q, J ) 6.3 Hz, 1H), 3.78 (s, 6H), 3.81 (s, 6H), 4.21 (d, J ) 5.6
Hz, 2H), 6.47 (s, 2H), 6.54 (s, 2H), 7.35 (m, 5H); 13C NMR
(CDCl3, 100 MHz) δ 23.0, 34.2 (2C), 51.8 (2C), 55.8 (2C), 56.0
(2C), 58.5, 110.1 (2C), 111.6 (2C), 127.0, 127.5, 127.6 (5C),
124.8 (2C), 130.5 (2C), 145.7, 147.48 (2C), 147.9 (2C); HRMS
(IE) calcd for C28H31NO4 445.2253, found 445.2252.
Der iva tive (-)-14. Isomer (-)-13a (116 mg, 0.26 mmol)
was dissolved in AcOEt (5 mL), EtOH (5 mL), and an aqueous
solution of HCl (2.4 N). The resulting solution was hydroge-
nated for 15 h, using 10% Pd/C (10%) as a calatyst. After
filtration over Celite and removal of solvent under reduced
pressure, H2O was added. The resulting solution was washed
with diethyl ether. The aqueous phase was then alkalinized
with 2 N NaOH and extracted with CHCl3. Removal of solvent
under reduced pressure afforded pure secondary amine (-)-
14 (60 mg, 0.18 mmol, 67% yield): [R]D -161 (c 0.44, CHCl3);
1H NMR (CDCl3, 300 MHz) δ 2.41 (s, 1H), 2.67 (d, J ) 16.1
Hz, 2H), 3.27 (dd, J ) 5.6, 16.1 Hz, 2H), 3.75 (s, 6H), 3.85 (s,
6H), 4.39 (d, J ) 5.6 Hz, 2H), 6.46 (s, 2H), 6.62 (s, 2H); 13C
NMR (CDCl3, 100 MHz) δ 37.4 (2C), 50.1 (2C), 55.8 (2C), 56.0
(2C), 109.6 (2C), 111.9 (2C), 124.7 (2C), 130.9 (2C), 147.5 (2C),
148.0 (2C); HRMS (CI, methane) calcd for C20H24NO4 (MH+)
342.1705, found 342.1735.
2 mL) and aqueous formaldehyde (37%, 1 mL). The resulting
mixture was heated at 90-95 °C for 2.5 h. The crude reaction
mixture was alkalinized with NaOH (2 N) and the product
extracted with CHCl3. Usual workup furnished pure (-)-
argemonine (colorless foam, 47 mg, 0.12 mmol, 75% yield): [R]D
1
-205 (c 0.5, CHCl3) [lit.10 [R]D -209 (c 0.5, CHCl3)]; H NMR
(CDCl3, 300 MHz) δ 2.53 (s, 3H), 2.58 (d, J ) 16.3 Hz, 2H),
3.40 (dd, J ) 5.7, 16.3 Hz, 2H), 3.76 (s, 6H), 3.85 (s, 6H), 4.00
(d, J ) 5.7 Hz, 2H), 6.45 (s, 2H), 6.61 (s, 2H); 13C NMR (CDCl3,
100 MHz) δ 33.7 (2C), 41.0, 55.8 (2C), 56.1 (2C), 56.5 (2C),
110.2 (2C), 111.6 (2C), 124.0 (2C), 130.11 (2C), 147.6 (2C),
148.0 (2C); HRMS (CI, isobutane) calcd for C20H24NO4 (MH+)
356.1865, found 356.1839.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedure for the preparation of derivatives 7, 5a -e, and 6a -e
1
and copies of H NMR and 13C NMR spectra of derivatives 7,
5a -e, 6d , 10-14, and (-)-argemonine. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O0357869
(-)-Ar gem on in e. Primary amine (-)-14 (60 mg, 0.18
mmol) was dissolved in a solution of formic acid (85% in H2O,
(10) Sternitz, F. R.; Seiber, S. N. J . Org. Chem. 1966, 31, 2925-
2933.
2740 J . Org. Chem., Vol. 69, No. 8, 2004