Highly diastereoselective three-component vinylogous mannich reaction between isoquinolines, acyl/sulfonyl chlorides, and silyloxyfurans

Article abstract of DOI:10.1021/ol901452d

Reaction of an isoquinoline, a silyloxyfuran, and an acyl or sulfonyl chloride provides easy access to a wide variety of isoquinolinobutyrolactones with excellent yields and diastereoselectivites (R*,R* isomer), even in the case of formation of quaternary

Full text of DOI:10.1021/ol901452d

ORGANIC
LETTERS
2009
Vol. 11, No. 18
4044-4047
Highly Diastereoselective
Three-Component Vinylogous Mannich
Reaction between Isoquinolines,
Acyl/Sulfonyl Chlorides, and
Silyloxyfurans
Philippe Hermange, Marie Elise Tran Huu Dau,
Pascal Retailleau, and Robert H. Dodd*
Institut de Chimie des Substances Naturelles - CNRS UPR 2301, baˆt. 27,
1 aV. de la Terrasse, 91198 Gif-sur-YVette, France
robert.dodd@icsn.cnrs-gif.fr
Received June 26, 2009
ABSTRACT
Reaction of an isoquinoline, a silyloxyfuran, and an acyl or sulfonyl chloride provides easy access to a wide variety of isoquinolinobutyrolactones
with excellent yields and diastereoselectivites (R*,R* isomer), even in the case of formation of quaternary centers (i.e., R3 or R4 ) Me).
Moreover, the use of a chiral auxiliary allowed formation of a single stereoisomer in 96% yield. This represents the first examples of asymmetric
vinylogous Mannich reactions on isoquinolinium salts.
Tetrahydroisoquinolines (THIQ) and the closely related 1,2-
dihydroisoquinolines represent a central unit of a wide variety
of both alkaloids and pharmaceutical products.1 These
moieties are very often observed with a substituent at the
C-1 stereogenic position, and various procedures have been
described in the past to access optically pure C-1-substituted
derivatives.2,3 In this context, the Mannich reaction represents
a very powerful tool for the formation of C-C bonds4 and
although this reaction is more difficult in the case of cyclic
vs noncyclic imines,3h,5 it has been successfully applied to
both 3,4-dihydroisoquinolines and isoquinolines for this
purpose in both racemic6 and asymmetric3g,h,7 versions. In
the last case, the use of chiral auxiliaries attached to the
nitrogen atom has been exploited,7a,8 while more recently,
Jacobsen and co-workers9 have employed chiral thioureas
as organocatalysts for the enantioselective addition of silyl
enol ethers to N-acylisoquinolines with excellent results.
Compared to the simple Mannich reaction, the vinylogous
version would allow introduction of greater chemical com-
plexity at C-1 of the heterocycle with the possibility of
generating two contiguous stereogenic centers in the case
of a nonterminal double bond on the starting silyl enol
ether.10 In this regard, butyrolactones and lactams are
particularly useful reagents for vinylogous Mannich reactions
since their silyl enol ethers are easily prepared and are
available with a variety of substituents on the ring.11 While
very few examples of the addition of silyloxyfurans to cyclic
(1) (a) Bentley, K. W. Nat. Prod. Rep. 2001, 18, 148–170. (b) Scott,
J. D.; Williams, R. M. Chem. ReV. 2002, 102, 1669–1730. (c) Compre-
hensiVe Medicinal Chemistry, 1st ed.; Hansch, C., Sammes, P. G., Taylor,
J. B., Eds.; Pergamon: Oxford, 1990; Vol. 3.
(2) Chrzanowska, M.; Rozwadowska, M. D. Chem. ReV. 2004, 104,
3341–3370
.
10.1021/ol901452d CCC: $40.75
Published on Web 08/13/2009
 2009 American Chemical Society

imines have been reported,12 we described the first vinylo-
gous Mannich reaction of 2-trialkylsilyloxyfurans with 3,4-
dihydroisoquinolinium salts.13 Thus, addition of 1-(tert-
butyldimethylsilyloxy)furan 2a to 3,4-dihydroisoquinolinium
salts 1 derived from acyl or sulfonyl chlorides afforded the
1-isoquinolinyl-5-butyrolactone 3 in high yields but with very
poor diastereoselectivities, typically in the 3:2 range (Scheme
1). In this paper, we describe the extension of this procedure
to completely aromatic isoquinoline derivatives and show
that, using a variety of cyclic silyl enol ethers and acylating/
sulfonylating agents, this vinylogous Mannich reaction is
high yielding and in most cases almost entirely diastereo-
selective in the absence of a chiral environment while in the
presence of a chiral environment, essentially only one of the
four possible stereoisomers is formed.
Scheme 1
of isoquinoline 4a with 2 equiv of acetyl chloride in
acetonitrile at 0 °C for 15 min followed by addition of 1.1
equiv of silyl enol ether 2a afforded product 5a in 80% yield
and with dr ) 92:8 (Table 1, entry 1). An even better result
We first studied the reaction of 2a with isoquinolinium
salts derived from various acid chlorides. Thus, treatment
Table 1. Variation of the Electrophile
(3) For examples and leading references, see: (a) Czarnocki, Z.;
MacLean, D. B.; Szarek, W; J, J. Chem. Soc., Chem. Commun. 1985, 1318–
1319. (b) Lee, A. W. M.; Chan, W. H.; Lee, Y. Tetrahedron Lett. 1991,
32, 6861–6864. (c) Kitamura, M.; Hsiao, Y.; Ohta, M.; Tsukamoto, M.;
Ohta, T.; Takaya, H.; Noyori, R. J. Org. Chem. 1994, 59, 297–310. (d)
Uematsu, N.; Fujii, A.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am. Chem.
Soc. 1996, 118, 4916–4917. (e) Morimoto, T.; Suzuki, N.; Achiwa, K.
Tetrahedron: Asymmetry 1998, 9, 183–187. (f) Lu, S.-M.; Wang, Y.-Q.;
Han, X.-W.; Zhou, Y.-G. Angew. Chem., Int. Ed. 2006, 45, 2260–2263.
(g) Sasamoto, N.; Dubs, C.; Hamashima, Y.; Sodeoka, M. J. Am. Chem.
Soc. 2006, 128, 14010–14011. (h) Dubs, C.; Hamashima, Y.; Sasamoto,
N.; Seidel, T.; Suzuki, S.; Hashizume, D.; Sodeoka, M. J. Org. Chem. 2008,
73, 5859–5871. (i) Funabashi, K.; Ratni, H.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2001, 123, 10784–10785. (j) Itoh, T.; Miyazaki, M.; Fukuoka,
H.; Nagata, K.; Ohsawa, A. Org. Lett. 2006, 8, 1295–1297. (k) Shi, C.;
Ojima, I. Tetrahedron 2007, 63, 8563–8570. (l) Frisch, K.; Landa, A.; Saaby,
S.; Jorgensen, K. A. Angew. Chem., Int. Ed. 2005, 44, 6058–6063. (m)
Garcia, E.; Arrasate, S.; Lete, E.; Sotomayor, N. J. Org. Chem. 2005, 70,
10368–10374. (n) Pedrosa, R.; Andre´s, C.; Iglesias, J. M. J. Org. Chem.
2001, 66, 243–250. (o) Taylor, A. M.; Schreiber, S. L. Org. Lett. 2006, 8,
143–146. (p) Li, Z.; MacLeod, P. D.; Li, C.-J. Tetrahedron: Asymmetry
2006, 17, 590–597. (q) Yamaguchi, R.; Tanaka, M.; Matsuda, T.; Fujita,
K. J. Chem. Soc., Chem. Commun. 1999, 2213–2214. (r) Wang, S.; Seto,
C. T. Org. Lett. 2006, 8, 3979–3982.
a Measured by HPLC. b 1.1 equiv of R-X, 2 equiv of 2a, CH2Cl2, -78
°Cf rt, 18 h. c See the Supporting Information for further examples.
(4) For a review, see: Arend, M.; Westermann, B.; Risch, N. Angew.
Chem., Int. Ed. 1998, 37, 1044–1070.
(5) (a) Friestad, G. K.; Mathies, A. K. Tetrahedron 2007, 63, 2541–
2569. (b) Ferraris, D. Tetrahedron 2007, 63, 9581–9597. (c) Shibasaki, M.;
Matsunaga, S. J. Organomet. Chem. 2006, 691, 2089–2100.
was obtained using phenylacetyl chloride as acylating agent,
compound 5b being formed in 82% yield and dr ) 98:2
(entry 2). With methyl chloroformate, the formation of only
one diastereomer of 5c, obtained in 87% yield, could be
observed by HPLC (entry 3), while benzyl chloroformate
provided compound 5d in 76% yield and dr ) 92:8 (entry
4). Reaction of 2a with the isoquinolinium salt derived from
methylsulfonyl chloride provided an excellent yield (89%)
and high dr (>99:1) of the corresponding N-methylsulfonyl
reaction product 5e (entry 5). Use instead of p-toluenesulfo-
nyl chloride gave a lower product yield (5f, 58%) though a
satisfactory diastereoselectivity (dr ) 93:7) was maintained
(entry 6). No reaction of 2a with the alkylisoquinolinium
salt derived from methyl iodide (entry 7) was observed,
indicating that an electron-withdrawing substituent on the
quaternary nitrogen atom is necessary for the reaction to
proceed.
(6) (a) Akiba, K.; Nakatani, M.; Wada, M.; Yamamoto, Y. J. Org. Chem.
1985, 50, 63–68. (b) Schmidt, A.; Gu¨tlein, J.-P.; Preuss, A.; Albrecht, U.;
Reinke, H.; Langer, P. Synlett 2005, 2489–2487.
(7) (a) Itoh, T.; Nagata, K.; Miyazaki, M.; Kameoka, K.; Ohsawa, A.
Tetrahedron 2001, 57, 8827–8839. (b) Murahashi, S.-I.; Imada, Y.;
Kawakami, T.; Harada, K.; Yonemushi, Y.; Tomita, N. J. Am. Chem. Soc.
2002, 124, 2888–2889.
(8) Itoh, T.; Miyazaki, M.; Nagata, K.; Hasegawa, H.; Ohsawa, A.;
Nakamura, K. T. Heterocycles 1998, 47, 125–128.
(9) Taylor, M. S.; Tokunaga, N.; Jacobsen, E. N. Angew. Chem., Int.
Ed. 2005, 44, 6700–6704.
(10) For a review of the vinylogous Mannich reaction, see: (a) Bur,
S. K.; Martin, S. F. Tetrahedron 2001, 57, 3221–3242. (b) Martin, S. F.
Acc. Chem. Res. 2002, 35, 895–904.
(11) (a) Casiraghi, G.; Rassu, G. Synthesis 1995, 607–626. (b) Rassu,
G.; Zanardi, F.; Battistini, L.; Gaetani, G.; Casiraghi, G. J. Med. Chem.
1997, 40, 168–180. (c) Rassu, G.; Zanardi, F.; Battistini, L.; Casiraghi, G.
Chem. Soc. ReV. 2000, 29, 109–118. (d) Casiraghi, G.; Zanardi, F.;
Appendino, G.; Rassu, G. Chem. ReV. 2000, 100, 1929–1972.
(12) (a) Boto, A.; Hernandez, R.; Suarez, E. Tetrahedron Lett. 2000,
41, 2899–2902. (b) Martin, S. F.; Barr, K. J.; Smith, D. W.; Bur, S. K.
J. Am. Chem. Soc. 1999, 121, 6990–6997. (c) Rassu, G.; Carta, P.; Pinna,
L.; Battistini, L.; Zanardi, F.; Acquotti, D.; Casiraghi, G. Eur. J. Org. Chem.
1999, 1395–1400. (d) Pichon, M.; Hocquemiller, R.; Figade`re, B. Tetra-
hedron Lett. 1999, 40, 8567–8570. (e) Hanessian, S.; McNaughton-Smith,
G. Bioorg. Med. Chem. Lett. 1996, 6, 1567–1572.
The relative configuration of the major diastereoisomer
of compound 5a was determined to be (R*,R*) by X-ray
crystallography (see the Supporting Information).14 We also
investigated the possibility of developing an asymmetric
(13) Razet, R.; Thomet, U.; Furtmu¨ller, R.; Chiaroni, A.; Sigel, E.;
Sieghart, W.; Dodd, R. H. J. Med. Chem. 2000, 43, 4363–4366.
Org. Lett., Vol. 11, No. 18, 2009
4045

version of this reaction using an (S)-valine-based chiral
auxiliary developed by Ohsawa7a for the asymmetric addition
of silyl enol ethers to the C-1 position of isoquinolines. Thus,
when the acid chloride derived from N-phthaloylvaline was
used as the electrophile in the vinylogous Mannich reaction
of 4a with 2a, one of the four possible isomers of 5h was
formed in 96% yield (entry 8).15
We then studied the behavior of substituted isoquinoline
derivatives in the vinylogous Mannich reaction with 2a. In
the case of the N-Cbz salt derived from 1-methylisoquinoline
4b, reaction with 2a under the same reaction conditions as
previously afforded only a low yield (10%) of the expected
1′-methyl-1′-(5-butyrolactone) product 5i though a high
diastereoselectivity was maintained (dr >95:5, Table 2, entry
highest yield of 5i (89%) with no deleterious effect on the
diastereoselectivity (entry 6). However, use of tosyl chloride
instead of Cbz chloride under these same one-pot conditions
did not give the expected addition product 5j (entry 7). These
optimal one-pot reaction conditions were then applied to
isoquinoline derivatives substituted on positions other than
C-1 and using either CbzCl or TsCl as the electrophile (Table
3). Thus, while reaction of 3-cyanoisoquinoline 4c, CbzCl,
Table 3. Variation of the Isoquinoline Substrate
Table 2. Optimization of Reaction Conditions
entry
isoquinoline
R-Cl
product yield (%)
dra
1
2
3
4
5
6
7
8
3-cyano 4c
CbzCl
TsCl
CbzCl
TsCl
CbzCl
TsCl
5k
5l
58
0
>99:1
4-bromo 4d
5-nitro 4e
5m
5n
5o
5p
5q
5r
94
95
90
88
80
81
92:8
90:10
>96:4
87:13
89:11
91:9
entry solvent
R-Cl
2a (equiv) yield (%)
dra
1
2
MeCN CbzCl, 2 equiv
MeCN CbzCl, 1.1 equiv
1.1
2
10
76
41
36
71
89
0
>95:5
>95:5
>95:5
>95:5
>95:5
>95:5
6,7-dimethoxy 4f CbzCl
TsCl
3
4
Et2O
THF
CbzCl, 1.1 equiv
CbzCl, 1.1 equiv
2
2
2
2
a Measured by HPLC.
5
CH2Cl2 CbzCl, 1.1 equiv
MeCN CbzCl, 1.1 equiv
MeCN TsCl, 1.1 equiv
6b
7b
2
and silyloxyfuran 2a provided a good yield (58%) and high
dr (>99:1) of the expected product 5k (entry 1), no reaction
was observed in the presence of TsCl (entry 2), a result
similar to that observed with 1-methylisoquinoline 4b (Table
2, entry 7), suggesting that steric factors inhibit the reaction
in both cases. In support of this, use of the less sterically
hindered 4-bromoisoquinoline 4d as the substrate provided
high yields in the presence of CbzCl (5m, 94%) or of TsCl
(5n, 95%) with dr ) 92:8 and dr ) 90:10, respectively
(entries 3 and 4). Similarly, 5-nitroisoquinoline 4e and 6,7-
dimethoxyisoquinoline 4f reacted with 2a in the presence
of either electrophile to give 80-90% yields of the corre-
sponding products 5o-r with consistently high diastereo-
selectivities (entries 5-8). Taken together, these results
indicate that while steric factors can negatively affect the
outcome of these vinylogous Mannich reactions, the elec-
tronic character of the substituents on the isoquinoline
nucleus has little influence, high yields and dr’s being
obtained whether these are electron-withdrawing or -donat-
ing.
The scope of the reaction with respect this time to the
nature of the vinylogous silyl enol ether was then studied
using isoquinoline 4a and TsCl as components and the same
conditions. As shown in Table 4, no difference in yield (89%,
90%) and dr (93:7) was seen in the formation of product 5f
whether the TBS or TMS ethers 2a or 2b, respectively, were
used (entries 1 and 2), while a slight decrease in yield of 5f
(81%) was observed using the TIPS ether 2c but with no
change in dr (93:7, entry 3). While a methyl group at C3 of
a Measured by 1H NMR. b The three components were mixed together
at the same time.
1).16 The yield of product 5i could be increased to 76%
without affecting the dr by diminishing the quantity of
acylating agent to 1.1 equiv and increasing the quantity of
silyloxyfuran 2a to 2 equiv (entry 2). Replacement, in this
last reaction, of the solvent (acetonitrile) by diethyl ether or
THF led to considerably lower yields (41% and 36%,
respectively) (entries 3 and 4), while dichloromethane
provided a satisfactory yield (71%) of 5i (entry 5).
Finally, it was found that mixing the three reagents
together in one-pot fashion instead of sequentially gave the
(14) The X-ray crystal structure determination of the major isomers of
compounds 5a, 12, 14, 15 (see the Supporting Information) and of 10
(unpublished results) showed that these correspond to the R*,R* diastere-
oisomers. This allowed correlation to 1H NMR data in which the difference
in H1′ and H5 ppm values (∆(δ)) for the major diastereomers was consistenly
larger than for the minor diastereomers, while inversely, the H1′, H5 coupling
constants for the major diastereomers were smaller (usually in the 4.8-5.8
Hz range) than for the minor diastereomer (typically in the 6.5-8.0 Hz
range). The same trends were observed in the vinylogous Mannich products
for which X-ray crystal structures are not available, thereby allowing
attribution of their relative stereochemistry.
(15) While the absolute configuration of the major isomer of compound
5h has not yet been determined with certainty, the observation by Ohsawa
(ref 7a) that the (S)-valine-derived chiral auxiliary gives the Mannich product
having the R configuration at C1 of isoquinoline combined with our
observation that the major product formed in the vinylogous Mannich
reaction has the R*,R* configuration strongly suggests that 5h therefore
has the R,R configuration.
(16) The absence of a proton at C1′ of 5i precludes use of the technique
of ref 15 for attribution of the relative configuration.
4046
Org. Lett., Vol. 11, No. 18, 2009

The introduction of a butyrolactone moiety at C1 of the
isoquinoline ring system allows for considerable synthetic
opportunities. An example is that illustrated in Scheme 2
Table 4. Variation of the Silyl Enol Ether
Scheme 2. Applications
wherein adduct 5d was efficiently converted into the ben-
zo[a]quinazoline 14.19 Thus, palladium/carbon-catalyzed
hydrogenation of 5d in ethyl acetate/acetic acid resulted in
simultaneous reduction of the 3,4-double bonds of the
isoquinoline and lactone moieties and hydrogenolytic cleav-
age of the Cbz group in 64% yield. Treatment of the product
with KOH in refluxing methanol then led to the cyclized
product 14 in 62% yield. Alternatively, reaction of isoquino-
line, 2-bromobutanoyl chloride, and 2b in acetonitrile fol-
lowed by treatment with zinc/iodine provided compound 15,
the core structure of emetine, an antitumor alkaloid.3j,20 The
structures of both 14 and 15 were confirmed by X-ray
crystallographic studies (see the Supporting Information).
In conclusion, we have discovered that the three-component
vinylogous Mannich reaction of silyl enol ethers derived from
butyrolactones and of isoquinolines in the presence of an
acylating or sulfonylating agent is high yielding and highly
diastereoselective even in the case of the formation of a chiral
quaternary center. Use of a chiral acylating agent such as the
acid chloride derived from N-phthaloylvaline then allows almost
complete stereoselectivity with formation of essentially a single
stereoisomer. The transformation of these substrates into a
variety of isoquinoline-based alkaloid natural products is an
objective which is presently being pursued.
a Measured by 1H NMR. b Confirmed by X-ray crystallography (see
the Supporting Information).
the furan ring (6) maintained a high product yield (88%)
and dr (93:7) (entry 4), the presence of the same group at
C4 (7) or C5 (8) resulted in both lower yields (54% and 77%)
and dr’s (60:40 and 76:24) of the corresponding products
11 and 12, respectively (entries 5 and 6). Finally, when the
reaction was conducted with the acyclic vinylogous enol
ether 9, an inseparable mixture of at least four diastereoi-
somers of the expected vinylogous Mannich products was
obtained (13, entry 7).
In order to explain the origin of the observed high diaste-
reoselectivity, the transition-state structures (TS) of the reaction
of silyloxyfuran with N-tosylisoquinolinium and N-tosyldihy-
droisoquinolinium were located and optimized using a semiem-
pirical orbital molecular method at the RHF/AM1 level.17,18
HOMO and LUMO coefficients analysis for favored TS
revealed favorable orbital overlaps between the silyloxyfuran
and the C3′-C4′ double bond of N-tosylisoquinolinium which
is not present in the N-tosyldihydroisoquinolinium TS (see
the Supporting Information). These stabilizing interactions
seem to be responsible for the observed high dr in the case
of the fully aromatic isoquinolinium salts (Figure 1).
Supporting Information Available: Experimental details,
characterization data, and spectra (1H and 13C NMR) for new
compounds, crystallographic data for compounds 5a, 12, 14,
and 15, and HPLC chromatograms for compounds 5. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL901452D
(17) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.
J. Am. Chem. Soc. 1985, 107, 3902–3909.
(18) An attempt to rationalize the diastereoselectivity observed in the
vinylogous Mannich reaction between 2-methoxyfuran and an N-carbam-
oylpyrrolinium salt has been reported. See: Bur, S. K.; Martin, S. F. Org.
Lett. 2000, 2, 3445–3447.
(19) Lee, Y. S.; Kang, S. S.; Choi, J. H.; Park, H. Tetrahedron 1997,
53, 3045–3056.
Figure 1. Calculated preferential transition state for silyloxyfuran
attack on N-tosylisoquinolinium.
(20) For leading references, see: (a) Stearman, C. J.; Wilson, M.; Padwa,
A. J. Org. Chem. 2009, 74, 3491–3499. (b) Tietze, L. F.; Rackelmann, N.;
Mu¨ller, I. Chem.sEur. J. 2004, 10, 2722–2731.
Org. Lett., Vol. 11, No. 18, 2009
4047