A new asymmetric synthesis of trans-hydroisoquinolones

Article abstract of DOI:10.1021/ol015696v

Matrix presented A convenient enantioselective synthesis of trans-hydroisoquinolones is described. This synthesis capitalizes on the ready availability of enantioenriched 2-substituted cyclohexenols by exploiting the asymmetry of an allylic carbon-oxygen

Full text of DOI:10.1021/ol015696v

ORGANIC
LETTERS
2001
Vol. 3, No. 8
1229-1232
A New Asymmetric Synthesis of
trans-Hydroisoquinolones
Asayuki Kamatani† and Larry E. Overman*
Department of Chemistry, 516 Rowland Hall, UniVersity of California,
IrVine, California 92697-2025
leoVerma@uci.edu
Received February 11, 2001
ABSTRACT
A convenient enantioselective synthesis of trans-hydroisoquinolones is described. This synthesis capitalizes on the ready availability of
enantioenriched 2-substituted cyclohexenols by exploiting the asymmetry of an allylic carbon−oxygen σ bond to control carbon−carbon bond
formation in pinacol-terminated cyclizations of N-acyliminium cations.
Decahydroisoquinoline rings are found in structurally diverse
isoquinoline alkaloids as well as several important clinical
agents.1,2 Morphine (1) and reserpine (2) are well-known
members of these groups. Because of the wide occurrence
and pharmacological importance of trans-hydroisoquinolines,
the development of new asymmetric routes to this ring
system remains an important objective in organic synthesis.
companying communication in this issue details reactions
wherein pinacol rearrangement of a ring carbon terminates
a cationic cyclization process.5 Much less developed are
cyclization-pinacol reactions concluded by hydride migra-
tion.6 A new sequence of this latter type, which we envisaged
(4) Recent illustrative examples include (a) MacMillan, D. W. C.;
Overman, L. E. J. Am. Chem. Soc. 1995, 117, 10391-10392. (b) Minor,
K. P.; Overman, L. E. Tetrahedron 1997, 53, 8927-8940. (c) Hanaki, N.;
Link, J. T.; MacMillan, D. W. C.; Overman, L. E.; Trankle, W. G.; Wurster,
J. A. Org. Lett. 2000, 2, 223-226. (d) Overman, L. E.; Pennington, L. D.
Org. Lett. 2000, 2, 2683-2686. (e) Molina-Ponce, A.; Overman, L. E. J.
Am. Chem. Soc. 2000, 122, 8672-8676.
(5) Cohen, F.; MacMillan, D. W. C.; Overman, L. E.; Romero, A. Org.
Lett. 2001, 3, 1225-1228; accompanying paper in this issue.
(6) For other Prins-pinacol reactions that involve hydride migrations,
see: (a) Cloninger, M. J.; Overman, L. E. J. Am. Chem. Soc. 1999, 121,
1092-1093. (b) Overman, L. E.; Pennington, L. D. Can. J. Chem. 2000,
78, 732-738.
In recent years, we have invented a suite of carbon-carbon
bond-forming ring constructions that couple pinacol rear-
rangements with cationic cyclization reactions.3,4 The ac-
(7) Several catalytic asymmetric reduction procedures would be possible;
at the time this work was initiated, oxazaborolidine-catalyzed borane
reduction was particularly attractive.8
(8) (a) Itsuno, S.; Ito, K.; Hirao, A.; Nakahama, S. J. Chem. Soc., Chem.
Commun. 1983, 469-470. (b) Corey, E. J.; Bakshi, R. K.; Shibata, S. J. J.
Am. Chem. Soc. 1987, 109, 5551-5553. (c) For a review, see: Itsuno, S.
Org. React. 1998, 52, 395-576.
(9) Kamatani, A.; Overman, L. E. J. Org. Chem. 1999, 64, 8743-8744.
(10) For a review of the Suzuki reaction, see: Miyaura, N.; Suzuki, A.
Chem. ReV. 1995, 95, 2457-2483.
(11) For recent reviews of N-acyliminium ion chemistry, see: (a)
Hiemstra, H.; Speckamp, W. N. In ComprehensiVe Organic Synthesis; Trost,
B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 2, Chapter 4.5.
(b) de Koning, H.; Speckamp, W. N. In Methods of Organic Chemistry
(Houben-Weyl); Helmchen, G., Hoffmann, R. W., Mulzer, J., Schaumann,
E., Eds.; Thieme: Stuttgart, 1995; Vol. E21b, Chapter D.1.4.5.
† Banyu Pharmaceutical Company, Process Research Division, 3-9-1
Kami-Mutsuna, Okazaki, Aichi 444-0858, Japan.
(1) Bentley, K. W. Nat. Prod. Rep. 1999, 16, 367-388 and references
therein.
(2) Lednicer, D.; Mitscher, L. A. The Organic Chemistry of Drug
Synthesis; Wiley: New York, 1997; Vol. 7 and earlier volumes in this series.
(3) For brief reviews, see: (a) Overman, L. E. Acc. Chem. Res. 1992,
25, 352-359. (b) Overman, L. E. Aldrichim. Acta 1995, 28, 107-120. (c)
Overman, L. E. In SelectiVities in Lewis Acid-Promoted Reactions; NATO
ASSI Series 289; Schinzer, D., Ed.; Kluwer Academic: Dordrecht, The
Netherlands, 1989; pp 1-20.
10.1021/ol015696v CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/23/2001

could provide a concise enantioselective entry to trans-
hydroisoquinolones having axial substituents at C1, is
illustrated in Figure 1. The heart of this plan is to set the
Scheme 1. Enantioselective Synthesis of Homoallylic
Carbamate Cyclization Precursors
Figure 1. Prins-pinacol synthesis of enantioenriched trans-
octahydroisoquinolones.
absolute configuration by catalytic asymmetric reduction of
2-iodocyclohexenone (3).7,8 The asymmetry of the allylic
carbon-oxygen σ-bond would then be employed to regulate
carbon-carbon bond formation in a Prins-pinacol reaction.
Specifically, silyl protection of the enantioenriched allylic
alcohol, followed by Suzuki â-aminoethylation,9,10 would
provide unsaturated carbamates 4. Under appropriate condi-
tions, condensation of 4 with an aldehyde should generate
N-acyliminium intermediates that we expected would cyclize
as depicted in 5 to deliver 6.11 Hydride migration and loss
of the silyl protecting group from 6 would lead to trans-5-
oxooctahydroisoquinoline 7 having an axial substituent at
C1. Stereoselection in the pivotal cyclization step was
anticipated to arise from two factors: (a) preferential
cyclization of the (E)-N-acyliminium ion stereoisomer so as
to avoid developing A1,3 interactions in the 1-substituted-2-
acylhydroisoquinoline product12 and (b) preferential approach
of the iminium ion electrophile from the cyclohexene face
opposite the bulky siloxy group.13
or tert-butyldiphenylsilyl (TBDPS) groups to yield 12-14.
Cross-coupling of these products with the 9-borabicy-
clononane adduct of benzyl vinylcarbamate (15) provided
homoallylic carbamates (R)-16, 17, and 18 in high yields.9
HPLC analysis of the alcohols derived from 16-18 con-
firmed that there was no loss of enantiomeric purity during
the Suzuki coupling step.14b Enantioselective reduction of 3
using the R enantiomer of 9 delivered (S)-16, 91% ee, in
comparable yield.
We initially examined formation of unsubstituted hy-
droisoquinolones from acid-promoted reaction of homoallylic
carbamate (R)-16 with paraformaldehyde (Scheme 2). When
carried out in toluene at room temperature in the presence
of 1.6 equiv of trifluoroacetic acid and Na2SO4, the reaction
of paraformaldehyde and (R)-16 provided a single 5-oxooc-
Scheme 2. Enantioselective Synthesis of
trans-5-Oxooctahydroisoquinolines
To test this sequence, homoallylic carbamates were
assembled from 2-iodocyclohexenones 3 and 8 as sum-
marized in Scheme 1 for the R cyclization precursors.
Enantioselective reduction of 3 or 8 with the oxazaborolidine
catalyst (S)-9 introduced by Corey and co-workers provided
1014a and 1114b in nearly quantitative yield and 93-94%
enantiomeric purity. Because preliminary survey experiments
had shown that the pivotal cyclization step took place in
higher yield when a robust silyl protecting group was used,
these products were protected with triisopropylsilyl (TIPS)
(12) Hoffmann, R. W. Chem. ReV. 1989, 89, 1841-1860.
(13) Cyclization from the face of the siloxy substituent, anti to the allylic
C-H σ-bond, should be favored for electronic reasons.5 However,
destabilizing nonbonded interactions between the developing axial C1
substituent and the bulky siloxy would be severe in such a cyclization
transition state.
(14) Enantiomeric purity was determined (a) by capillary GLC analysis
using a J&W CyclodexB column or (b) by HPLC analysis using a Daicel
OD-H column. In all cases, the analysis was calibrated with a sample of
the racemate.
1230
Org. Lett., Vol. 3, No. 8, 2001

tahydroisoquinoline product 20 in 66% yield. That 20 was
the expected trans stereoisomer was signaled by the di-
agonostic triplet of doublets (J ) 11.1, 3.2 Hz) observed
for H-4a at δ 2.26 ppm in the 1H NMR spectrum.15 However,
the enantiomeric purity of 20 was a disappointing 66% ee.16
That the 4aS,8aS enantiomer of 20 had been formed was
established unambiguously by Mosher analysis.17,18 On the
assumption that the loss of enantiopurity arose from compet-
ing acid-promoted conversion of the starting cyclohexenyl
ether to an achiral allyl cation, we turned to the nonacidic
conditions first reported by Chamberlin and Chung for
generating N-acyliminium cations.19 Reaction of (R)-16 with
excess paraformaldehyde and 2 equiv of Cs2CO3 in dry THF
gave R-hydroxycarbamate 19, which when activated at room
temperature with methanesulfonyl chloride and Et3N in CH2-
Cl2 delivered (4aS,8aS)-20 in 76% overall yield and 86%
ee. Under identical conditions, the enantiomeric ketone (4aR,-
8aR)-20 was formed from (S)-16 with a similar high degree
of stereospecificity, whereas cyclization of dimethyl deriva-
tive 18 took place with complete stereospecificity to give
21 in 67% overall yield.17
room temperature provided carbamates 24 (80% yield) and
25 (85% yield) as ∼1:1 mixtures of ethoxy epimers. After
examining several common Lewis acids, it was found that
24 and 25 cyclized in highest yield when exposed to 1 equiv
of BF3‚OEt2 in CH2Cl2 at 0 °C in the presence of the protic
acid scavenger 2,6-di-tert-butyl-4-methylpyridine (DTBMP).
Under these conditions, both silyl ether intermediates gave
rise to a single 1-substituted 5-oxooctahydroisoquinoline
product 26. Both yield and enantioselection in this conversion
were improved by use of the more acid stable tert-
butyldiphenylsilyl protecting group, with 25 providing 26
in 91% yield and 88% ee.14 The relative configuration of 26
1
followed unambiguously from H NMR studies,21 whereas
the 1R,4aS,8aS absolute configuration of 26 was determined
on a derivative by the advanced Mosher method.17
Results of our initial survey of this new enantioselective
synthesis of 1-substituted trans-hydroisoquinolones are sum-
marized in Table 1. Aliphatic, aromatic, or R,â-unsaturated
Table 1. Enantioselective Synthesis of 1-Substituted
trans-5-Oxooctahydroisoquinolines
To expand this sequence to the formation of 1-substituted
5-oxooctahydroisoquinolines, we turned to cyclizations with
other aldehydes. Homoallylic primary amines 22 and 23 were
first prepared by selective removal of the benzyloxycarbonyl
protecting group from (R)-16 and 17 under transfer hydro-
genolysis conditions (Scheme 3). Condensation of these
Scheme 3. Enantioselective Synthesis of 1-Substituted
trans-5-Oxooctahydroisoquinoline 26
carbamate 27 5-oxohydroisoquinolinea
yield,
% compd
yield,
%
ee,
%
entry amine
R
SiR3
1
2
3
4
5
6
7
8
9
22
n-Pr
TIPS
72 28
87
75
69
91
63
72
53
34
51
64
0
87
ent-23 n-Pr
TBDPS 99 ent-28
TIPS 80 26
TBDPS 85 26
TBDPS 99 29
TBDPS 59 30
87
39
88
85b
88
90
88
90
22
23
23
23
23
23
23
i-Pr
i-Pr
Ph
p-OMeC6H4
(E)-MeCHdCH TBDPS 61 31
t-Bu
CH2SPh
CO2Et
TBDPS 44 32
TBDPS 56c 33
TBDPS 80 34
10 23
11 23
0-20
p-OMe-m-OBn- TBDPS 60 35
C6H3CH2
a Absolute configuration was determined with the N-methyl equatorial
alcohol derivative by the method of Mosher and Kakisawa (ref 18).
b Cyclization was conducted at -78 °C; ee was 76% when the cyclization
was carried out at 0 °C. c Contained unidentified impurities
aldehydes can be used to provide, in high enantiopurity (85-
90% ee), trans-5-oxooctahydroisoquinolines having axial
amines with isobutryaldehyde in CH2Cl2 at room temperature
using a 1:1 mixture of MgSO4 and K2CO3 as a water
scavenger delivered the corresponding imines. After removal
of CH2Cl2 and addition of dry ethanol, reaction of these
intermediates with 1.2 equiv of diethyl pyrocarbonate20 at
(15) To collapse carbamate rotamers, this analysis was carried out in
DMSO-d6 at elevated temperature.
(16) Enantiomeric purity was determined by chromatographic analysis14a
of the N-methyl derivative obtained by reduction of the Prins-pinacol
product with LiAlH4, followed by Swern oxidation of the resulting amino
alcohol product.
Org. Lett., Vol. 3, No. 8, 2001
1231

alkyl, aryl, or alkenyl substituents at C1. When R was
primary alkyl, secondary alkyl, or aryl (entries 1-6), overall
yields from amine 23 (or ent-23) ranged from 43% to 74%.
Yields for both steps were somewhat lower in reactions with
trans-crotonaldehyde, pivaldehyde, or 2-(phenylthio)acetal-
dehyde (entries 7-9). In favorable cases, the TIPS- and
TBDPS-protected allylic alcohols performed comparably
(entries 1 and 2); however, the TBDPS derivative is generally
preferred. In one case (entry 5) it was demonstrated that
chirality transfer was slightly higher when the cyclization
step was carried out at -78 °C. It is likely that cyclization
at lower temperature will be preferred in many cases.
Several limitations of the sequence summarized in Table
1 were also revealed. For example, ethyl glyoxylate (entry
10) provided a single trans-5-oxooctahydroisoquinoline 34
in useful yield. However, this product was nearly racemic
(0-20% ee). To pursue the origin of racemization in this
case, a 1:1 mixture of R-ethoxy carbamate 25 and the
analogous intermediate derived from ethyl glyoxylate was
cyclized under standard conditions for 30 s at 0 °C to provide
26 in 64% yield; 34 was not detected. That racemization in
the glyoxylate case involves to a significant extent formation
of an achiral cyclohexenyl carbocation is consistent with the
deuterium scrambling observed in the reactions reported in
eqs 1 and 2. The slow conversion of ethyl glyoxylate-derived
deficient N-acyliminium cation in this case, apparently allows
“background” racemization of the starting material to occur
at a competitive rate. Although R-alkoxycarbamates formed
from 23 and acetone or 2-(3-benzyloxy-4-methoxyphenyl)-
acetaldehyde (Table 1, entry 11) could be generated in useful
yield, neither intermediate was converted to the correspond-
ing trans-5-oxooctahydroisoquinoline when exposed to BF3‚
OEt2. Unfortunately, this latter failure precludes direct use
of this chemistry for enantioselective synthesis of opium
alkaloids.
In summary, this communication discloses a conceptually
new strategy for asymmetric construction of trans-hydroiso-
quinolines. The synthesis exploits the wide availability of
enantioenriched 2-substituted cyclohexenols by catalytic
asymmetric reduction of cyclohexenone precursors. The
central step in this sequence translates the asymmetry of the
allylic C-O σ bond of the cyclohexenol precursors by a
Prins-pinacol sequence to three adjacent stereocenters of
the hydroisoquinoline products. Using this chemistry, a
variety of trans-5-oxooctahydroisoquinolines having axial
alkyl, aryl, or alkenyl substituents at C1 can be prepared in
high enantiopurity.
Acknowledgment. This research was supported by a
Javits Neuroscience Investigator Award from NIH NINDS
(NS-12389). Merck, Pfizer, Roche Biosciences, and Smith-
Kline Beecham provided additional support. NMR and mass
spectra were determined at UCI using instruments acquired
with the assistance of NSF and NIH shared instrumentation
grants.
Supporting Information Available: Representative ex-
perimental procedures and characterization data; details of
the determination of absolute configuration of 28. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL015696V
(18) (a) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem.
Soc. 1991, 113, 4092-4096. (b) Dale, J. A.; Dull, D. L.; Mosher, H. S. J.
Org. Chem. 1969, 34, 2543-2549.
(19) Chamberlin, A. R.; Chung, J. Y. L. J. Am. Chem. Soc. 1983, 105,
3653-3656.
R-ethoxy carbamate 27a to hydroisoquinolone 34, which
must reflect the relative instability of the highly electron-
(20) Hiemstra, H.; Fortgens, H. P.; Speckamp, W. N. Tetrahedron Lett.
1985, 26, 3155-3158.
(21) The 1H NMR spectrum at 100 °C (DMSO-d6) exhibited a diagnostic
triplet of doublets (J ) 11.8, 3.8 Hz) for H-4a at δ 2.47 ppm, and the
NOESY spectrum showed a strong correlation between H-1 and H-8a.
Similar experiments established the structure of the other 1-substituted
5-oxooctahydroisoquinolines reported in Table 1.
(17) Absolute configuration was determined by the advanced Mosher
method using the N-methyl equatorial alcohol formed by sequential reduction
of the Prins-pinacol product with NaBH4 (EtOH, -30 °C) and LiAlH4;
details are provided in Supporting Information.
1232
Org. Lett., Vol. 3, No. 8, 2001