An asymmetric route to the conanine BCDE ring system. A formal total synthesis of (+)-conessine

Article abstract of DOI:10.1021/ja961903o

The first enantioselective synthesis of the known (+)-conessine precursor (+)-benzohydrindan 23 from the chiral nonracemic bicyclic lactam 1 is described. The key transformation was the highly diastereoselective (3 + 2) cycloaddition of azomethine ylide 1

Full text of DOI:10.1021/ja961903o

9876
J. Am. Chem. Soc. 1996, 118, 9876-9883
An Asymmetric Route to the Conanine BCDE Ring System. A
Formal Total Synthesis of (+)-Conessine
Michael E. Kopach, Andrew H. Fray, and A. I. Meyers*
Contribution from the Department of Chemistry, Colorado State UniVersity,
Fort Collins, Colorado 80523
ReceiVed June 6, 1996X
Abstract: The first enantioselective synthesis of the known (+)-conessine precursor (+)-benzohydrindan 23 from
the chiral nonracemic bicyclic lactam 1 is described. The key transformation was the highly diastereoselective (3 +
2) cycloaddition of azomethine ylide 11a to lactam 7 in order to construct the pyrrolidine E ring system at any early
stage in the synthesis. The requisite pyrrolidine methyl group at C-21 was stereoselectively installed late in the
synthesis by lithiation of N-Boc-pyrrolidine intermediate 20 with sec-BuLi/TMEDA followed by quenching at -90
°C with iodomethane, furnishing the tetracyclic pyrrolidine 21. Reduction of t-Boc 21 with lithium aluminum hydride
affords (+)-N-methyl tetracycle 23 in a concise 13-step synthesis from 1.
Introduction
Stereoselective pyrrolidine syntheses have attracted the at-
tention of synthetic organic chemists for many years since
pyrrolidine rings are found in many alkaloid natural products
including (+)-conessine and regholarrhenine A and C which
are members of the Hollarhena class of Kurchi alkaloids (Figure
1).1 Isolated from the bark of Holarhena antidysenterica, these
steroidal alkaloids have been shown to possess significant
biological activity.2 The best known of these steroidal alkaloids
is (+)-conessine which has been used in the treatment of
dysentery.
Synthetic efforts to reach conessine were quite extensive in
the 1960s when Stork,3 Nagata,4 and Johnson5 all reported
successful total syntheses of the racemic steroidal alkaloid.
Stork,3 Mukharji,6a and Mukherjee6b also completed syntheses
of the racemic advanced intermediate B from accessible steroidal
precursors A. The transformation of B to conessine was also
made more attractive by Velluz7 who demonstrated that the AB
ring could be readily obtained in six steps, a significant
improvement over earlier routes.8
Figure 1.
a more efficient route but also one in which the appropriate
enantiomer of B could be acquired.
From previous work dealing with the synthesis of chiral
cyclopentenones9 and 3,4-disubstituted pyrrolidines10 derived
from chiral bicyclic lactams, it was felt that a potentially efficient
route to the tetracyclic ring system B of (+)-conessine could
be accessed. Thus, a synthetic plan was devised (Scheme 1)
which would utilize a (3 + 2) azomethine ylide addition to the
suitably substituted chiral, nonracemic lactam 7. The resulting
cycloadduct 9 would then be induced to undergo an intramo-
lecular aryllithium addition to the lactam carbonyl, affording
the carbinolamine 13. The latter, upon acidic hydrolysis, should
provide the spiro dicarbonyl derivative 14 which would be
cyclized, in an aldol manner, to the tetracyclic cyclopentenone
15. Precedence for this cyclization paradigm was presented in
a recent report9 to generate chiral, nonracemic benzohydrin-
denones C.
The intended asymmetric route to conessine, Via the tetra-
cyclic pyrrolidine B, appeared to be an attractive one, particu-
larly in view of the lengthy sequence employed earlier.6,7
Furthermore, the frequent low yielding steps, and the incomplete
characterization of B,7 provided the incentive to seek not only
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) Jeger, O.; Prelog, V. In The Alkaloids: Chemistry and Physiology;
Manske, R. H. F., Ed.; Academic Press: New York, 1960; Vol. VII, p
319.
(2) (a) Siddiqui, B. S.; Usmani, S. B.; Begum, S.; Siddiqui, S.; Aftab,
K.; Gilani, A. Heterocycles 1995, 41, 267 and references cited within. (b)
Bhutani, K. K.; Ali, M.; Sharma, S. R.; Vaid, R. M.; Gupta, D. K.
Phytochemistry 1988, 27, 925. (c) Bhutani, K. K.; Vaid, R. M.; Ali, M.;
Kapoor, R.; Soodan, S. R.; Kumar, D. Phytochemistry 1990, 27, 925.
(3) Stork, G.; Darling, S. D.; Harrison, I. T.; Wharton, P. S. J. Am Chem
Soc. 1962, 84, 2018.
(4) Nagata, W.; Terasawa, T.; Aoki, T. Tetrahedron Lett. 1963, 869.
(5) Marshall, J. A.; Johnson, W. S. J. Am. Chem. Soc. 1962, 84, 1484.
(6) (a) Mukharji, P. C.; Sircar, J. C.; Devasthale, V. V. Indian J. Chem.
1971, 9, 515. (b) Mukherjee, M.; Mukherjee, L. M. Indian J. Chem. 1980,
19B, 185. (c) Mukherjee, M.; Mukherjee, L. M. Indian J. Chem. 1981,
20B, 1019.
Results and Discussion
The synthetic sequence which ultimately led to the enan-
tiopure conanine BCDE system 23 began by initially construct-
ing the requisite chiral bicyclic lactam 7 (Scheme 2). The
tethered arylethyl side chain to be affixed to lactam 1 was
prepared routinely from commercially available 3-methoxyphe-
nylacetic acid. Reduction to the alcohol 2, followed by
(7) Velluz, L.; Mathieu, G.; Monine, G. Tetrahedron Suppl. 1966, 8,
495.
(8) Stork, G.; Loewenthal, H. J. E.; Mukharji, P. C. J. Am. Chem. Soc.
1956, 78, 501.
(9) Snyder, L.; Meyers, A. I. J. Org. Chem. 1993, 58, 7507.
(10) (a) Fray, A. H.; Meyers, A. I. Tetrahedron Lett. 1992, 33, 3575.
(b) Fray, A. H.; Meyers, A. I.; J. Org. Chem. 1996, 61, 3362.
S0002-7863(96)01903-8 CCC: $12.00 © 1996 American Chemical Society

Asymmetric Route to the Conanine BCDE Ring System
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9877
Scheme 1. Retrosynthetic Route to the Conanine BCDE
Ring System
Figure 2.
Scheme 3
Scheme 2
With the unsaturated chiral lactam 7 in hand, the azomethine
ylide cycloaddition was next investigated. Treatment of 7 with
the known15-17 azomethine precursor 8 under a variety of acid-
catalyzed and pressurized conditions ultimately gave the cy-
cloadduct 9 in 86-90% yield and as a 15:1 mixture of anti-
syn products (Scheme 3).
The need for high pressure to induce the cycloaddition of 8
to 7 is not without precedent. High pressures18 have proven
effective in 1,3-dipolar cycloaddition reactions of azides,19
nitrones,20 nitronates,21 and diazomethane.22 Therefore, it
seemed reasonable that the (3 + 2) cycloaddition might be
achieved under conditions of higher pressure and temperature.
Accordingly, the crucial (3 + 2) cycloaddition was first achieved
in reasonable yield (50-60%) under conditions of medium
pressure by combining unsaturated lactam 7, azomethine ylide
precursor 8, and catalytic trifluoroacetic acid (10-30 mol %)
in benzene in a sealed Ace pressure tube at 240 °C.23 In this
manner, (3 + 2) cycloaddition occurs, affording pyrrolidine 9
as a 15:1 ratio of diastereomers, and separation of these
diastereomers was readily accomplished by flash chromatog-
raphy.24 The major cycloadduct is assigned so the pyrrolidine
ring exhibits â face fusion based on a 4% NOE enhancement
(anti product, Figure 2) between angular methyl 8b (1.52 ppm)
and the 8a methine hydrogen resonance (2.49 ppm). In addition,
the 13C NMR chemical shift of the angular methyl group at 8b
(28.2 ppm) further suggests that the major cycloadduct has the
transformation to the iodide 3 with triphenylphosphine and
iodine,11 gave the necessary precursor for aromatic bromination.
The latter took place using bromine in CH2Cl2 at -10 °C and
produced the aryl bromide 4 in 95% yield.
The alkylation of the bicyclic lactam 112 was performed using
LDA at -78 °C to form the enolate which was treated with the
aryl iodide 4, affording the monoalkylated lactam 5 in 92% yield
as a 2:1 mixture of diastereomers. In order to achieve this yield,
it was necessary to utilize 2.0 equiv of LDA and 3.0 equiv of
the aryl iodide. Interestingly, no double alkylation occurred
with the use of excessive base and electrophile. Presumably
this was the result of steric crowding about the tertiary proton
as well as a slower rate of electrophilic alkylation.
(15) (a) Padwa, A.; Chen, Y. Y.; Dent, W.; Nimmesgerm, H. J. J. Org.
Chem. 1985, 50, 4006. (b) Padwa, A.; Dent, W. J. Org. Chem. 1987, 52,
235.
(16) Achiwa, K. Imai, N.; Motoyama, T.; Sekiya, M. Chem. Lett. 1984,
2041.
Introduction of the unsaturated linkage into 5 was performed
using either of the following two methods. Initially the use of
9,13,14
KN(SiMe3)2
to generate the enolate of 5 at -78 °C
followed by diphenyl diselenide to give 6 was investigated.
Oxidation of the latter with hydrogen peroxide gave the
unsaturated lactam 7 in 72% yield. It was noted during the
course of optimizing the selenylation-oxidation sequence that
KH (2.0 equiv) also efficiently generated the enolate and after
selenylation-oxidation produced the unsaturated lactam in 75%
yield. The use of KH in place of KN(SiMe3)2 was somewhat
fortunate since the cost of the latter base is exceedingly high.
(17) Vedejs, E.; Martinez, G. R. J. Am. Chem. Soc. 1979, 101, 6453.
(18) For reviews of high-pressure organic reactions, see: (a) Asano, T.;
Lenoible, W. J. Chem. ReV. 1978, 78, 407. (b) Matsumoto, K.; Uchida, T.
Synthesis 1985, 999. (c) Matsumoto, K., Acheson, R. M., Eds. Organic
Synthesis at High Pressures; Wiley: New York, 1991.
(19) Anderson, G. T.; Henry, J. R.; Weinreb, S. M. J. Org. Chem. 1991,
56, 6946.
(20) Deshong, P.; Li, W.; Kennington, J. W.; Ammon, H. L. J. Org.
Chem. 1991, 56, 1364.
(21) Kamernitzky, A. V.; Levina, I. S.; Mortikova, E. I.; Shitkin, V. M.;
El’Yanov, S. B. Tetrahedron 1977, 33, 2135.
(11) Collins, D. J.; Fallon, G. D.; Skene, C. E. Aust. J. Chem. 1992, 45,
71.
(12) Bicyclic lactam 1 was derived from (S)-valinol and levulinic acid
and is commercially available from Ultrafine Chemicals, Manchester, U.K.
(13) Romo, D.; Romine, J. L.; Midura, W.; Meyers, A. I. Tetrahedron
1990, 46, 4951.
(22) Desuray, H.; Weiler, J. Tetrahedron Lett. 1974, 2209.
(23) Vapor pressures of benzene at selected temperatures: 10 atm (178.8
°C), 20 atm (221.5 °C), and 40 atm (272.3 °C) CRC Handbook of Chemistry
and Physics, 56th ed.; Weast, R. C., Ed.; CRC Press: Cleveland, OH, 1976;
p 6.
(24) The ratio of diastereomers was determined by integration of pertinent
resonances corresponding to anti and syn isomers.
(14) Williams, R. M.; Im, M. N. J. Am. Chem. Soc. 1991, 113, 9276.

9878 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Kopach et al.
Scheme 4
manipulation, the carbinolamine was directly hydrolyzed as
previously described.9 It was found that direct hydrolysis of
the THF solution containing carbinolamine using excess tet-
rabutylammonium dihydrogen phosphate buffer led to poor
yields of spirodiketone 14. However, when the crude reaction
mixture containing carbinolamine 13 was diluted with an equal
volume of CH2Cl2 and then treated with excess phosphate buffer
solution28 (15-30 equiv), good yields of spirodiketone 14 were
obtained after refluxing the biphasic mixture for 24-48 h.29
Spirodiketone 14 did not spontaneously cyclize to form ben-
zohydrindenone 15, but upon treatment with sodium ethoxide
in ethanol, cyclization readily occurs to afford 15, thereby
completing construction of the conanine BCDE ring system.
Additionally, it was found to be expedient to convert pyrrolidine
9 to benzohydrindenone 15 in two pots without intervening
purification. By this approach, crude spirodiketone 14 was
submitted directly to base-catalyzed aldol conditions to afford
benzohydrindenone 15 in 82% yield for the three-step sequence
from pyrrolidine 9.
To reach the intended tetracycle 23, it was necessary to reduce
the enone 15 in a 1,2 or 1,4 manner with appropriate stereo-
chemical control. The tetracyclic enone 15 was found to be
resistant to a variety of reducing conditions including highly
reactive magnesium30 or zinc metals.31ab In addition, attempts
to reduce 15 with NaBH4 gave only starting material.
A number of homoannular enones have been efficiently
deoxygenated to the corresponding olefins with AlCl3 and
LiAlH4. Accordingly, when 15 was charged into a mixture of
AlCl3/LiAlH4, deoxygenation readily occurred without double
bond migration to afford 16 as a crystalline solid in 92% yield
(Scheme 5).
pyrrolidine forming anti to the angular methyl group.25 Also,
the 13C NMR chemical shift of the angular methyl group 8b
(21.2 ppm) in the minor isomer is consistent with its being syn
to the pyrrolidine ring.25
Since the (3 + 2) cycloaddition was found to proceed with
“truly catalytic” amounts of TFA,26 it was thought that the
protonated pyrrolidine in the product might serve as the active
catalyst for this reaction. In order to test this hypothesis, a
solution of unsaturated lactam 7, dipole precursor 8, and
protonated pyrrolidine 9 (1%) as a TFA salt were combined in
a sealed tube. Under these conditions a rate acceleration to 6
h was observed as compared to 24 h in the absence of protonated
product.27 Optimized conditions utilized a benzene or benzene-
chloroform solvent containing 0.01 mol % trifluoroacetic acid
and 2.0 equiv of 8. The mixture was then heated to 180-190
°C in a glass (Ace) pressure vessel for 6 h.
Due to the successful construction of the E ring system in 9
by (3 + 2) cycloaddition, the possibility to stereoselectively
install the C21 methyl group in conessine by use of a methyl-
substituted azomethine ylide, 12, was considered (Scheme 4).
Accordingly, the procedure of Padwa15 was used to construct
dipole precursor 10 in a manner analogous to the synthesis of
8. However, it has been observed by Padwa15 that reactions of
azomethine ylide 12 with unsymmetrical dipolarophiles afford
products with regiochemistry resulting from attachment of the
less substituted carbon terminus of the azomethine ylide to the
least substituted terminus of the dipolarophile. Thus, if this
precedent held, the (3 + 2) cycloaddition of 10 to unsaturated
lactam 7 would proceed with the wrong regiochemistry.
Nevertheless, it was hoped that the phenethyl side chain present
in 7 might direct the regiochemistry to the desired regioisomer.
Unfortunately, under optimal conditions for the synthesis of 9
as well as using a variety of Lewis acids, decomposition of the
parent azomethine ylide was the only event observed. At this
stage it was decided to postpone installation of the C21 methyl
group until later in the synthesis.
The next requirement toward the synthesis of 23 was R
addition of hydrogen to saturate the olefin at C14. It was
anticipated that hydrogenation of 16 would occur from the R
face on the basis of well-documented reductions in systems
analogous to 13â-substituted-∆14,15-steroids.7a,8,33 The hydro-
genation of 16 was attempted in the presence of PdCl2 in
ethanol, which after a few hours, produced a 3:1 mixture of 20
and epi-20 (R ) H). 1H NMR examination was unable to
distinguish whether the major product was a result of R or â
addition. An X-ray structure of one of the isolated epimers
would be necessary to clarify this point; however, this mixture
of amines proved to be inseparable by standard chromatography
techniques. The mixture was then carried on to the next step
with the hope of later evaluating the stereochemical outcome
of the C14-hydrogenation. Treatment of the mixture of amines
in 20 (R ) H) with Boc-ON and KH resulted in rapid formation
(28) Owing to the high cost of 1 M tetrabutylammonium dihydrogen
phosphate buffer solution ($90 for 500 mL of 1 M solution, Aldrich), a
buffer recovery protocol was implemented. Upon completion of hydrolysis
of carbinolamine 13, the biphasic mixture was allowed to separate and the
aqueous phase was saved and reused 3-4 times in subsequent syntheses to
afford 15 in better than 80% yield each time.
The C ring of (+)-conessine was next constructed by addition
of excess t-BuLi (2 equiv) to a THF solution of pyrrolidine
cycloadduct 9 over KH. The expected metal-halogen exchange
rapidly occurred, providing a transient alkyllithium species that
spontaneously cyclized to carbinolamine 13. Without further
(29) In some instances 1H NMR of the crude spirodiketone 14 revealed
considerable quantities of unhydrolyzed carbinolamine 13. In these cases
the crude material was redissolved in CH2Cl2 and treated with fresh
phosphate buffer solution and hydrolysis continued at reflux temperatures
until complete consumption of carbinolamine was observed.
(30) Olah, G.; Prakash, G. K. S.; Arvanaghi, M.; Bruce, M. R. Angew.
Chem., Int. Ed. Engl. 1981, 20, 92.
(25) This gamma gauche (γg) substituent effect is greatest in confor-
mationally rigid systems whenever two protons bearing carbons are at
dihedral angles of 60° or less. See: Wehrli, F. W.; Wirthlin, T. Interpretation
of Carbon-13 NMR Spectra; Heydon & Son, Ltd: Chichester, U.K., pp
27-28, 37. For bicyclic lactams, with few exceptions, the chemical shift
for the angular methyl group in the anti isomers 9 (27-29 ppm) is
considerably downfield from that of the syn isomers (19-23 ppm).
(26) Complete decomposition of unsaturated lactam 7 occurs upon use
of BF3‚OEt2 and TiCl4 at higher temperatures than 0 °C.
(27) In a separate experiment 1 mol % pyrrolidine 9 was added to a
solution of dipole precursor 8 and unsaturated lactam 7 followed by trace
TFA (∼0.01%). After heating, the (3 + 2) cycloaddition proceeded to
completion within 6 h at 190 °C.
(31) (a) LeGoff, E. J. Org. Chem. 1964, 29, 2048. (b) Sondengam, B.
L.; Fomum, Z. T.; Charles, G.; Akam, T. M. J. Chem. Soc., Perkin Trans.
1 1983, 1219.
(32) (a) Brown, B. R.; White, A. M. S. J. Chem. Soc. 1957, 3755. (b)
Broome, J.; Brown, B. R.; Roberts, A.; White, A. M. S. J. Chem. Soc.
1960, 1406. (c) Bokadia, M. M.; Brown, B. R.; Cobern, D.; Roberts, A. J.
Chem. Soc. 1962, 1658.
(33) (a) Platner, P. A.; Ruzicka, L.; Heuser, H.; Angliker, E. HelV. Chim.
Acta 1947, 30, 395. (b) Banerjee, D. K.; Chatterjee, S.; Pillai, C. N.; Bhatt,
M. V. J. Am. Chem. Soc. 1956, 78, 3769.

Asymmetric Route to the Conanine BCDE Ring System
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9879
Scheme 5
of tert-butyl carbamate 20 (R ) Boc), also as a 3:1 ratio of
epimers, but separable. The configuration of the minor isomer
of N-Boc-pyrrolidine 20 (R ) Boc) was determined by X-ray
crystallography to have S-configuration (R-H) at C14; thus, the
required epimer 20 was unfortunately the minor product from
the reduction. Apparently, hydrogenolysis of the N-benzyl
group in 16 occurred before CdC hydrogenation, which allowed
competitive hydrogenation from the â face. Consequently, it
was necessary to devise an alternative route to 20 that would
provide predominantly R addition of hydrogen at C14.
It was decided to replace the N-benzyl group of 16 with a
bulky group that would remain intact under the usual hydro-
genolytic conditions. In this regard, replacement of the N-benzyl
group with a carbamate could possibly serve two purposes: (a)
direct the hydrogenation at C14 from the R face; (b) serve as
an activating group for stereoselective carbanionic installation
of the C21 methyl group to reach the final product 23. It was,
therefore, anticipated that N-Boc-pyrrolidine 19 could be ideally
suited for these purposes.
In order to form 19, it was necessary to convert the
N-benzylolefin 16 into a phenyl carbamate which might eject
the phenoxy group with tert-butoxide in direct analogy to the
method described by Comins.34 Accordingly, the N-benzyl
derivative 16 was heated with phenyl chloroformate (X ) H),
furnishing the phenyl carbamate 17. However, when KH was
added to a solution of 17 in refluxing tert-butyl alcohol, cleavage
of the phenyl carbamate group occurred without tert-butoxide
incorporation, producing only the free NH derivative. After
azeotropic removal of t-BuOH the residue was treated with Boc-
ON in ethanol, furnishing the desired N-Boc-olefin 19 in low
yield (30%). Exposure of N-Boc-olefin 19 to an atmosphere
of H2 over activated palladium on carbon afforded, exclusively,
the desired benzhydrindan 2035 in near quantitative yield as a
result of hydrogen delivery to the less hindered R face. The S
(R-H) configuration at C14 was compared to the X-ray data
taken earlier, and both showed identical stereochemistry.
After confirmation of the correct stereochemistry in the
hydrogenation of N-Boc-pyrrolidine 20, improvements in the
sequence from benzohydrindene 16 were sought since the three-
step yield from 16 to 20 was only 30%. It was believed that
use of a more reactive phenyl carbamate might allow for direct
displacement of phenoxide with tert-butoxide as originally
observed by Comins.34 To this end, 16 was treated with
p-nitrophenyl chloroformate and immediately formed phenyl
carbamate 18 (X ) NO2). Addition of the crude carbamate to
a mixture of stirring t-BuOK (5 equiv) in THF followed by the
immediate addition of 1.0 equiv of H2O provided clean
displacement of p-nitrophenol and incorporation of tert-butoxide,
furnishing N-Boc-pyrrolidine 19 in 66% yield. In summary, at
this stage, the benzohydrindan 20 was prepared in 65% yield
by a three-pot process from the N-benzylpyrrolidine 16 with
purification necessary only for the final product. This procedure
also precluded the use of the Boc-ON reagent. It is noteworthy
that 20 (R ) Boc) was prepared in 20% overall yield from
commercially available 1.
The last major hurdle in this synthesis was the stereoselective
introduction of the C21 methyl group which is present in (+)-
conessine and related systems. It is known that simple achiral
N-Boc-pyrrolidines may be metalated R to nitrogen with sec-
BuLi/TMEDA and then alkylated with electrophiles to form
substituted pyrrolidines.36 In addition, Beak has demonstrated
that highly enantioselective alkylations37 can be achieved for
simple achiral N-Boc-pyrrolidines utilizing (-)-sparteine in
place of TMEDA.
Due to the chiral nonracemic nature of the pyrrolidine 20 (R
) Boc), it was felt that a diastereoselective alkylation at C20
might be achieved without any additional chiral components in
the reaction mixture. It was further anticipated that competitive
alkylation at C18 to afford the undesired regioisomer would be
inhibited by hindrance imposed by the adjacent neopentyl center.
Therefore, it was decided to attempt alkylation of N-Boc-
pyrrolidine 20 using Beak’s optimized conditions36 for alkylation
of simple N-Boc-pyrrolidines. Accordingly, a solution of
N-Boc-pyrrolidine 20 in ether was treated with an equimolar
amount of TMEDA followed by sec-BuLi at -78 °C. The
solution was then treated with excess dimethyl sulfate and
1
warmed to room temperature. The H-NMR spectrum of the
(34) Comins, D. L.; Weglarz, M. A.; O’Conner, S. Tetrahedron Lett.
1988, 29, 1751.
(36) (a) Beak, P.; Lee, W. K. J. Org. Chem. 1990, 55, 2578. (b) Beak,
P.; Lee, W. K. J. Org. Chem. 1990, 55, 2578.
1
(35) Although H NMR for 20 showed complete disappearance of the
vinyl proton resonance (5.70 ppm) present in the precursor 19, the N-Boc-
pyrrolidine was found to exist as a mixture of two rotamers at 25 °C.
Collection of spectroscopic data at 55 °C resulted in partial resolution, but
it was necessary to record the spectrum of 20 in toluene-d8 at 100 °C to
effect complete resolution of signals.
(37) (a) Beak, P.; Kerrick, S. T.; Wu, S.; Chu, J. J. Am. Chem. Soc.
1994, 116, 3231. (b) Kerrick, S.; Beak, P. J. Am. Chem. Soc. 1991, 113,
9708. (c) Hoppe, D.; Zschage, O. Angew. Chem., Int. Ed. Engl. 1989, 28,
69. (d) Hoppe, D.; Hintze, F.; Tebben, P. Angew Chem., Int. Ed. Engl.
1990, 29, 1422.

9880 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Kopach et al.
Scheme 6
a syn relationship of the two methine protons at C19 and C20.
For further confirmation of the stereochemistry, a 6% NOE
enhancement was observed between the methine protons at C19
and C20.40 Thus, taken together these data support the
assigment of C20 in 21 as having the appropriate S-configuration
as shown in Figure 2. It is also important to recall that the
desmethyl derivative 20 (R ) Boc) was assigned its structure
by X-ray crystal data. Presently, it is not certain why the C20
methyl entered the â face of pyrrolidine 20, and the reasons
must await further study. However, Beak has proposed37 on
simple, monocyclic N-Boc-pyrrolidines that metalation and
alkylation proceeds with retention. Others have proposed
similar behavior41 as well as alkylation with inversion,42 but
no hard evidence yet exists. Inspection of the X-ray structure
of 20 (R ) Boc) shows that either hydrogen at C20 is roughly
equal with regard to accessibility to the base. Yet, the directing
influence in this instance may be a consequence of the eclipsing
C19 bridgehead methine which may render the R proton at C20
kinetically less accessible. It is also possible that deprotonation
of the R H occurs and the â stereochemistry of the methyl
product is a consequence of an inversion mechanism.43
crude product revealed what appeared to be an impure mixture
of 21 in a 1:1 ratio of inseparable methylated diastereomers,
along with starting material. Attempts to improve the ratio of
methylated products by keeping the alkylation temperature
below -70 °C gave only starting material. Thus, no alkylation
occurred at low temperature. In fact, no alkylation of lithiated
20 took place below -25 °C, and when it did occur, the
selectivity of methyl insertion was ∼1:1. It was also determined,
Via deuteration with methanol-d1 that metalation of 20 had
completely taken place (>95% D) after 8 h at -78 °C; thus,
no concern was registered about the lithio anion being present
during the alkylation.
With the initial attempts at deprotonation of the N-Boc system
20 proving to be disappointing, the formamidine group was
considered to be a viable alternative. It has been shown on
numerous occasions that R-aminocarbanions are readily pro-
duced when treated with organolithium reagents in the presence
of chiral or achiral formamidines. Thus, the N-Boc and
formamidine groups both activate the R-proton of an amino
group but on several occasions lead to stereo- and regiochemi-
cally different results.38d-f Furthermore, the configurational
stability of the R-lithioformamidines was found to be somewhat
greater than the Boc counterparts39 which may also play a role
in the stereoselective approach to 23. The N-Boc-pyrrolidine
20 was transformed into its free amine 22 using TFA-CH2Cl2
and then treated with N,N-dimethyl-N′-tert-butylformamidine
to form the formamidine 24 in 46% unoptimized yield (Scheme
6). When the latter was metalated in THF using sec-BuLi-
TMEDA at -78 °C, the deep red anion appeared and was
alkylated with dimethyl sulfate. Workup revealed that the crude
product was once again a 1:1 mixture of methyl diastereomers.
This was confirmed by removing the formamidine and compar-
ing the spectral data with the NH pyrrolidine 22 generated from
the N-Boc analog 21. The use of methyl iodide, in place of
dimethyl sulfate, showed no change in selectivity in the
methylation process although the reaction was significantly more
rapid at lower temperatures. In view of the enhanced reaction
rate using methyl iodide, it was felt that, by reassessing the
metalation of the N-Boc 20 (R ) Boc) and maintaining a lower
alkylation temperature, a possible improvement in the stere-
oalkylation step may arise. In fact, this proved to be the case
as the N-Boc 20 was metalated with sec-BuLi-TMEDA in ether
at -90 °C and at this temperature treated with methyl iodide.
The monomethyl product 21 was obtained in 50% yield as a
15:1 ratio of diastereomers. Also, a 25% yield of starting
pyrrolidine 20 (R ) Boc) was recovered.
Proceeding to the final target, treatment of methylpyrrolidine
21 with TFA/CH2Cl2 gave the corresponding NH compound in
good yield (88%). This material was submitted to Eschweiler-
Clarke methylation conditions to afford N-methylpyrrolidine 23
in 50% yield. Recrystallization produced a white crystalline
solid (mp 87-8 °C),44 with [R]D +51°. The spectral data
exhibited the presence of a new methyl group (2.10 ppm (1H
NMR), 41.0 ppm (13C NMR)) consistent with N-methylation.
Due to the low yield of the Eschweiler-Clarke methylation an
alternate route was sought. Fortunately, it was found that
45
reduction of 21 with LiAlH4 cleanly furnished the desired
system 23 in 85% yield.
In summary, a 13-step asymmetric synthesis of benzohydrin-
dan 23 is described from the chiral nonracemic bicyclic lactam
1. Significantly, the interesting pyrrolidine E ring was formed
Via an azomethine ylide cycloaddition to unsaturated lactam 7
at an early stage of the synthesis. Stereoselective introduction
of the methyl group in ring E was accomplished late in the
synthesis, and this intermediate was found to correspond to
benzohydrindan 23, reported earlier.6
Experimental Section
3-Methoxyphenethyl iodide, 3. To a 0 °C solution of triphen-
ylphosphine (37.8 g, 144 mmol) and imidazole (9.82 g , 144 mmol) in
CH2Cl2 (250 mL) was added iodine (36.70 g, 144.6 mmol) portionwise
to form a turbid orange solution. The phenethyl alcohol 246 (18.3 g,
120 mmol) over anhydrous sodium sulfate was then cannulated into
the above suspension over 5 min, and the reaction was allowed to
gradually warm to room temperature over 3 h while being stirred. TLC
analysis revealed complete consumption of starting alcohol, and the
(40) Irradiation of the C20 methine shows a 3% NOE with the C21
methyl group. In addition, irradiation of the C21 methyl group shows a 3%
NOE with the C20 methine and does not show any detectable NOE with
the C19 bridgehead methine.
1
The H NMR of the major diastereomer of 21 revealed a
downfield methine resonance (3.93 ppm) with a coupling
constant of 6.3 Hz to the newly installed C21 methyl group
(1.30 ppm) and a 6.0 Hz coupling constant with the C19
bridgehead methine (2.20 ppm). Therefore, these coupling data
confirm the regiochemistry of addition at C20 and also suggest
(41) (a) Hoppe, D.; Paetow, M.; Hintze, F. Angew. Chem., Int. Ed. Engl.
1993, 32, 394. (b) Pearson, W. H.; Lindbeck, A. C.; Kampf, J. W. J. Am.
Chem. Soc. 1993, 115, 2622. (c) Chong, J. M.; Park, S. B. J. Org. Chem.
1992, 57, 2220. (d) Gawley, R. E.; Zhang, Q. J. Am. Chem. Soc. 1993,
115, 7515.
(42) (a) Gawley, R. E.; Zhang, Q. J. Org. Chem. 1995, 60, 5767. (b)
Meyers, A. I.; Dickman, D. A. J. Am. Chem. Soc. 1987, 109, 1263.
(43) We have recently studied the metalation-alkylation of 20 in greater
detail and found, using D-isotope effects, that both steps proceed with
retention (Kopach, M. E., Meyers, A. I. J. Org. Chem., in press).
(44) The melting point for (+)-benzohydrindan 23 was slightly higher
than the literature value for (()-23 (lit.6a,c mp 82 °C).
(38) (a) Meyers, A. I. Aldrichim. Acta. 1985, 18, 59. (b) Meyers, A. I.;
Edwards, P. B.; Rieker, W. F.; Bailey, T. R. J. Am. Chem. Soc. 1984, 106,
3270. (c) Edwards, P. B.; Meyers, A. I. Tetrahedron Lett. 1984, 25, 939.
(d) Meyers, A. I.; Milot, G. J. Am. Chem. Soc. 1993, 115, 6652. (e) Meyers,
A. I.; Milot, G. J. Org. Chem. 1993, 58, 6538. (f) Meyers, A. I.; Nguyen,
T. H. Tetrahedron Lett. 1995, 36, 5873.
(45) Maryanoff, B.; Almond, H. J. Org. Chem. 1986, 51, 3295.
(46) Kanth, J. V. B.; Periasamy, M. J. Org. Chem. 1991, 56, 5964.
(39) Meyers, A. I.; Elworthy, T. Tetrahedron 1994, 6089.

Asymmetric Route to the Conanine BCDE Ring System
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9881
light orange suspension was then washed with saturated sodium bisulfite
solution (2 × 250 mL). The resulting light yellow solution was then
dried over anhydrous magnesium sulfate and concentrated to afford a
light yellow residue which was applied to a silica gel column
presaturated with 1:10 CH2Cl2/hexanes. After elution with the same
solvent ratio, the resulting product fractions were then concentrated to
afford 30.5 g (96.7%) of 3 as a pale yellow oil:11 Rf ) 0.59 (1:1:1
Et2O/CH2Cl2/hexanes); 1H NMR δ 7.22 (t, J ) 7.9 Hz, 1 H), 6.77 (m,
3 H), 3.79 (s, 3 H), 3.33 (t, J ) 7.7 Hz, 2 H), 3.04 (t, J ) 7.7 Hz, 1
H); 13C NMR δ 159.7 (C), 142.1 (C), 129.6 (CH), 120.6 (CH), 114.1
(CH), 112.1 (CH), 55.2 (CH3), 40.4 (CH2), 5.30 (CH2).
1.74 (m, 1 H), 1.47 (s, 3 H), 1.08 (d, J ) 6.7 Hz, 3 H), 0.90 (d, J )
6.6 Hz, 3 H); 13C NMR δ 178.0 (CO), 158.9 (C), 143.5 (CH), 141.1
(C), 139.7 (C), 133.3 (CH), 115.9 (CH), 114.8 (C), 113.8 (CH), 98.7
(C), 74.0 (CH2), 62.2 (CH), 55.4 (CH3), 33.9 (CH2), 33.1 (CH), 25.5
(CH2), 22.1 (CH3), 20.6 (CH3), 19.2 (CH3): IR (film) ν 1711, 1650
cm-1
. Anal. Calcd for C19H24NO3Br: C, 57.88; H, 6.13; N, 3.55.
Found: C, 57.70; H, 6.19; N, 3.46.
N-Benzyl-N-(methoxymethyl)-N-[(trimethylsilyl)methyl]amine, 8.
Padwa’s procedure15b was improved by adding N-benzyl-N-[(trimeth-
ylsilyl)methyl]amine (19.9 g, 103.5 mmol) Via syringe pump to a stirred
solution of 37% aqueous formaldehyde (11.1 mL, 140 mmol) at 0 °C
over 3.5 h. The reaction mixture was then stirred for an additional 0.5
h and was charged with CH3OH (10 mL) and K2CO3 (5 g). After
stirring for 1 h at 0 °C, the reaction mixture was diluted with excess
CH3OH (100 mL) and K2CO3 (15 g) and the temperature reduced to
-20 °C. After stirring for 17 h, excess K2CO3 was filtered and the
mixture was concentrated to afford a pale yellow liquid. Distillation
provided 20 g (88%) of pure 8 as a clear colorless liquid: bp 77-80
°C (0.1 mmHg) (lit.15b bp 78-80 °C (5 mmHg)); 1H NMR δ 7.31 (m,
5 H), 3.99 (s, 2 H), 3.75 (s, 2 H), 3.23 (s, 3 H), 2.18 (s, 2 H), 0.04 (s,
9 H); 13C NMR δ 139.7 (C), 128.7 (CH), 128.1 (CH), 126.8 (CH),
88.4(CH2), 59.5 (CH2), 55.4 (CH3), 42.9 (CH2), -1.5 (CH3)3C.
Pyrrolidine 9. A solution of unsaturated lactam 7 (2.66 g, 6.74
mmol) and dipole precursor 8 (4.1 g, 17.3 mmol) in a 10:1 benzene/
CHCl3 (12 mL) mixture was placed in an Ace glass pressure tube.
Trifluoroacetic acid (3 mg, .03 mmol) was added and an argon blanket
placed over the reaction mixture. The pressure tube was then sealed
and placed in a preheated sand bath (230 °C).48 After 36 h the reaction
mixture was cooled to rt, and TLC data of the dark yellow solution
revealed complete consumption of starting material 7. The mixture
was then partitioned between CH2Cl2 (100 mL) and saturated, aqueous
NaHCO3 (100 mL). The organic extracts were dried (MgSO4) and
concentrated to afford a light brown syrup. In a 1H NMR spectrum of
the crude material, integration of the resolved signals corresponding
to the anti isomer (δ 1.52) and the syn isomer (δ 1.38) indicated an
anti:syn ratio of 15:1 (88% de). The mixture was then purified by
flash chromatography (10% EtOAc/hexanes) to afford 3.0 g (84%) of
diastereomerically pure 9 as a yellow oil: Rf ) 0.45 (1:1:1 Et2O/CH2-
2-Bromo-5-methoxyphenethyl iodide, 4. A solution of Br2 (6.3
mL, 122.5 mmol) in CH2Cl2 (125 mL) was added dropwise to a solution
of phenethyl iodide 345 (30.5 g, 122 mmol) in CH2Cl2 (150 mL) at
-10 °C over 2 h. The reaction mixture was then treated with a solution
of saturated sodium bisulfite (200 mL), and the biphasic mixture was
stirred for 15 min until the dark red color discharged. The mixture
was then diluted with saturated NaHCO3 (150 mL) and extracted with
CH2Cl2 (3 × 250 mL). The combined extracts were dried over
anhydrous MgSO4 and concentrated to afford 4 as a pale yellow syrup
(37.5 g (95%)) which was passed through basic alumina prior to use:
1
Rf ) 0.69 (1:1:1 Et2O/CH2Cl2/hexanes); H NMR δ 7.40 (t, J ) 8.8
Hz, 1 H), 6.78 (d, J ) 3.0 Hz, 1 H), 6.68 (dd, J ) 3.0, 8.8 Hz, 1 H),
3.78 (s, 3H), 3.33 (m, 2H) 3.24 (m, 2H); 13C NMR δ 159.0 (C), 140.7
(C), 133.5 (CH), 116.4 (CH), 114.4 (C), 114.2 (CH), 55.5 (CH3), 40.8
(CH2), 3.10 (CH2).
Monoalkylated Lactam 5. Monoalkylation of lactam 1 was
performed according to the previous method.9 A 2.50 M solution of
n-BuLi in hexanes (10.9 mL, 27.2 mmol) was added dropwise to
diisopropylamine (2.89 g, 3.75 mL, 28.6 mmol) at -10 °C to form a
thick syrup which was diluted with THF (100 mL) and cooled to -78
°C. A solution of bicyclic lactam 1 (2.49 g, 13.5 mmol) in THF (10.0
mL) was added dropwise to the LDA solution followed by a THF (5.0
mL) rinse. After 2 h at -78 °C, stirring was stopped, neat aryl iodide
4 (13.8 g, 40.5 mmol) was rapidly charged into the reaction mixture,
the resulting yellow mixture was warmed immediately to 0 °C, and
stirring was resumed. After an additional 6 h of stirring at 0 °C, the
reaction mixture was quenched with saturated NH4Cl (10 mL) and
diluted with Et2O (200 mL). The organic phase was separated and
washed with saturated NaHCO3 (2 × 150 mL), dried (MgSO4), and
concentrated. This was applied to a silica gel column presaturated with
1:1 CH2Cl2/hexanes and was eluted with the same until the less polar
component (2-bromo-5-methoxystyrene) emerged. The column was
then eluted with 1:1 EtOAc/hexanes to afford 4.95 g (92%) of 5 as a
2:1 ratio of diastereomers. This mixture of diastereomers was used
directly in the next synthetic step. Data for 5: GC/MS m/z (abundance)
1
Cl2/hexanes); [R]D ) +25.1 (c 2.38, benzene); H NMR δ 7.37 (d, J
) 8.7 Hz, 1 H), 7.17-7.35 (m, 5 H), 6.75 (d, J ) 3.0 Hz, 1 H), 6.61
(dd, J ) 3.0, 8.7 Hz, 1 H), 4.16 (dd, J ) 7.4, 8.1 Hz, 1 H), 3.86 (dd,
J ) 5.3, 8.2 Hz, 1 H), 3.74 (s, 3 H), 3.50 (ABq, J ) 13.5 Hz ∆V )
59.4, 2 H), 3.35 (m, 2 H), 2.87 (td, J ) 5.1, 12.5 Hz, 1 H), 2.64 (td,
J ) 5.9, 11.6 Hz, 1 H), 2.49 (dd, J ) 1.9, 8.1 Hz, 1 H), 2.22 (apparent
t, J ) 8.8 Hz, 1 H), 2.12 (apparent d, J ) 8.8 Hz, 1 H),1.94 (m, 2 H),
1.70 (m, 2 H), 1.52 (s, 3 H), 1.05 (d, J ) 6.5 Hz, 3 H), 0.91 (d, J )
6.5 Hz, 3 H); 13C NMR δ 182.2 (CO), 159.1(C), 141.8(C), 138.8 (C),
133.3 (CH), 128.2 (CH), 128.1 (CH), 126.8 (CH), 115.7 (CH), 114.6
(C), 113.7 (CH), 97.4 (C), 70.9 (CH2), 63.1 (CH2), 60.6 (C), 59.1 (CH2),
55.5 (CH), 54.9(CH2), 52.6 (CH3), 35.3 (CH2), 33.1 (CH), 32.9 (CH2),
28.2 (CH3), 20.5 (CH3), 19.2 (CH3); IR (film) ν 2960, 2871, 1706,
1596, cm-1. Data for the minor diastereomer: Rf ) 0.55 (1:1:1 Et2O/
CH2Cl2/hexanes); 1H NMR (selected signals) δ 6.75 (d, J ) 3.0 Hz, 1
H), 6.59(dd, J ) 3.0, 8.7 Hz, 1H), 5.04 (dd, J ) 6.9, 7.8 Hz, 1H), 4.17
(t, J ) 8.2 Hz, 1H), 3.74 (s, 3 H), 3.32 (d, J ) 9.2 Hz , 1 H), 3.06 (d,
J ) 10.2 Hz , 1 H), 2.81 (dd, J ) 7.9, 11.4 Hz, 2 H), 2.70 (d, J ) 6.3
Hz , 1 H), 2.36 (td, J ) 5.1, 13.2 Hz, 1 H), 2.23 (dd, J ) 6.6, 10.2 Hz,
1 H), 1.38 (s, 3 H), 1.13 (d, J ) 6.6 Hz, 3 H), 0.89 (d, J ) 6.6 Hz, 3
H); 13C NMR δ 182.1 (CO), 159.0 (C), 142.2 (C), 138.7 (C), 133.2
(CH), 128.4 (CH), 128.3 (CH), 127.0 (CH), 115.5 (CH), 114.7 (C),
113.7 (CH), 99.6 (C), 69.0(CH2), 64.1 (CH2), 63.7 (CH), 58.7 (C), 58.2
(CH2), 55.44 (CH3), 55.36(CH2), 50.4 (CH), 36.1 (CH2), 34.4 (CH),
32.4 (CH2), 21.0 (CH3), 19.5 (CH3), 19.0 (CH3). Anal. Calcd for
C28H35NO3Br: C, 63.75; H, 6.69; N, 5.31. Found: C, 63.48; H, 6.73;
N, 5.20.
397.3 (M+ + 1), 396.3 (M+); HRMS calcd for C19H26NO3Br (M+
1) 396.1174, found 396.1160.
+
Unsaturated Bicyclic Lactam 7. To a solution of bicyclic lactam
5 (1.90 g, 4.79 mmol) and diphenyl diselenide (1.69 g, 5.41 mmol) at
25 °C in THF (50 mL) was added KH (394 mg, 9.82 mmol) portionwise
over 1 h, causing mild effervescence. After 1.5 h TLC showed
complete consumption of starting material and formation of higher Rf
selenylated intermediates 6. The mixture was quenched with saturated
NH4Cl (2 mL) and diluted with CH2Cl2 (40 mL). The flask was fitted
with a reflux condenser, cold 30% H2O2 (30 mL, 294 mmol) was
quickly added, and the biphasic mixture was stirred at rt for 12 h
(CAUTION: upon addition of H2O2 an exotherm occurs which
maintains reflux temperature for ∼3 h).47 The mixture was diluted
with CH2Cl2 (100 mL) and treated with saturated sodium bisulfite (100
mL), and the organic phase was separated and the aqueous phase
extracted with CH2Cl2 (2 × 150 mL). The organic extracts were
combined, dried (MgSO4), and concentrated to afford a viscous yellow
oil which was purified by column chromatography. Elution with CH2-
Cl2 afforded 1.55 g (82%) of 7 as a yellow oil: Rf ) 0.70 (1:1:1 Et2O/
Spirodiketone 14. A solution of pyrrolidine 9 (156 mg, 0.296
mmol) in THF (5 mL) was stirred over KH (22 mg) while being cooled
to -78 °C to ensure water free conditions. After 40 min, a 1.9 M
1
CH2Cl2/hexanes); H NMR δ 7.38 (d, J ) 8.7 Hz, 1 H), 6.72 (d, J )
3.0 Hz, 1 H), 6.60 (dd, J ) 3.0, 8.8 Hz, 1 H), 6.55 (s, 1 H), 4.30 (dd,
J ) 7.4, 8.9 Hz, 1 H), 4.03 (dd, J ) 6.3, 8.9 Hz, 1 H), 3.74 (s, 3 H),
3.48 (ddd, J ) 6.5, 7.4, 10.1 Hz, 1 H), 2.91 (m, 2 H), 2.53 (m, 2 H),
(48) This procedure involved the use of smaller amounts of dipole
precursor 8 whereas utilizing the conditions described in the discussion (6
h at 190 °C) gave the same yield of product. However, the latter conditions
required the use of larger amounts of dipole precursor 8.
(47) It is recommended that this procedure be conducted behind a blast
shield in a hood.

9882 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Kopach et al.
solution of t-BuLi in pentane (0.33 mL, 0.62 mmol) was added dropwise
and the resulting solution was stirred for 15 min. TLC analysis of the
reaction mixture revealed disappearance of 9 (Rf ) 0.49, 1:1 EtOAc/
hexanes) and formation of UV active carbinol intermediate 13 (Rf )
0.39, 1:1 EtOAc/hexanes). The mixture was quenched at -78 °C with
1 M aqueous (n-Bu)4NH2PO4 (4.5 mL, 4.5 mmol). The resulting frozen
mixture was allowed to warm to rt, diluted with CH2Cl2 (4.5 mL), and
heated to reflux (40 °C) for 13 h. The reaction mixture was poured
into excess saturated aqueous NaHCO3 (20 mL) and extracted with
CH2Cl2 (3 × 20 mL). The organic extracts were dried (Na2SO4) and
concentrated, and the oily residue was purified by radial chromatography
(1:1 EtOAc/hexanes) to afford 78.6 mg (73%) of 14 as an off-white
powder. An analytical sample was obtained by recrystallization from
EtOH: Rf ) 0.31 (EtOAc); mp 125-127 °C; [R]D ) -140 (c 1.03,
benzene); 1H NMR δ 7.94 (d, J ) 8.8 Hz, 1 H), 7.20-7.35 (m, 5 H),
6.80 (dd, J ) 2.5, 8.7 Hz, 1 H), 6.62 (d, J ) 2.4 Hz, 1 H), 3.82 (s, 3
H), 3.67 (ABq, J ) 12.9 Hz ∆V ) 31.2, 2 H), 3.22 (m, 1 H), 3.06-
3.14 (m, 2 H), 2.87-2.97 (m, 2 H), 2.77-2.85 (m, 1 H), 2.58-2.69
(m, 2 H), 2.16 (dt, J ) 4.1, 13.3 Hz, 1 H), 2.09 (s, 3 H); 13C NMR δ
207.4 (CO), 198.8 (CO), 163.7(C), 145.7(C), 138.7 (C), 130.6 (CH),
128.6 (CH), 128.3 (CH), 127.1 (CH), 125.7 (C), 113.4 (CH), 112.3
(CH), 61.8 (CH2), 61.5 (CH), 60.0 (CH2), 56.7 (CH2), 56.2 (C), 55.4
(CH3), 35.7 (CH2), 29.8 (CH3), 27.4 (CH2); IR (film) ν 2919, 2819,
1707, 1666 cm-1. Anal. Calcd for C23H25NO3: C, 76.01; H, 6.93.
Found: C, 75.89; H, 6.98.
(CH), 128.1 (CH), 126.8 (CH), 126.6 (CH), 125.4 (C), 120.1 (CH),
113.1 (CH), 112.5 (CH), 64.5 (CH2), 60.8 (CH2), 60.0 (CH), 57.4 (C),
55.2 (CH3), 47.5 (CH), 40.45 (CH2), 35.1 (CH2), 28.1 (CH2); IR (film)
ν 3029, 2920, 1607, 1568 cm-1. Anal. Calcd for C23H25NO: C, 82.84;
H, 7.56. Found: C, 82.94; H, 7.61.
(1S,7aS)-4′-Methoxybenzo[4,5]hydrindeno[1,7a-c]-1-[[(4-nitro-
phenyl)oxy]carbonyl]pyrrolidine, 18. A solution of 16 (440 mg, 1.32
mmol) in CH2Cl2 was treated with p-nitrophenyl chloroformate (270
mg, 1.34 mmol) and the mixture stirred for 15 min. Evaporation of
the solvent afforded a tan solid whose 1H NMR showed a 1:1 ratio of
benzyl chloride to 18. The crude reaction product was then subjected
to high vacuum (0.1 mmHg) for several hours to remove benzyl
chloride, and the remaining material was purified by column chroma-
tography eluting initially with CH2Cl2 and then 1:1:1 ether/CH2Cl2/
hexanes to afford 347 mg (64.5%) of 18 as a tan semisolid. 1H NMR
data recorded at 25 °C showed the presence of two distinct rotamers,
so characterization data are reported at 55 °C in CDCl3: Rf ) 0.58
(1:1:1 ether/CH2Cl2/hexanes); [R]D ) +42 (c 1.00, benzene); 1H NMR
(55 °C) δ 7.45 (d, J ) 8.6 Hz, 1 H), 7.20-7.35 (m, 4 H), 6.72 (dd, J
) 2.9, 8.6 Hz, 1 H), 6.68 (d, J ) 2.9 Hz, 1 H), 5.79 (s, 1 H), 3.78 (s,
3 H), 3.76 (m, 2 H), 3.50 (m, 2 H), 2.75-3.01 (m, 3 H), 2.71 (br s, 1
H), 2.32 (m, 1 H), 2.05 (dd, J ) 3.0, 12.0 Hz, 1 H), 1.70 (td, J ) 6.0,
12.5 Hz, 1 H); 13C NMR δ 159.1(C), 151.6 (C), 144.6 (C), 142.8 (C),
136.7 (C), 126.5 (CH), 126.0 (CH), 125.0 (2xCH), 124.1 (CH), 113.0
(2xCH), 55.1 (CH3), 55.0 (CH2), 51.5 (CH2), 51.3 (C), 49.7 (CH), 48.7
(CH), 40.3 (CH2), 32.9(CH2), 28.2 (CH2); IR (film) ν 2910, 1727, 1594,
Direct Conversion of 9 to 15. Alternative Procedure. A solution
of lactam 9 (2.0 g, 3.79 mmol) in 10:1 THF/benzene (50 mL) was
stirred over KH (500 mg) while being cooled to -78 °C in the same
manner as above. The organic extracts were dried (MgSO4) and
concentrated, and the oily residue containing 14 was dissolved in
benzene (10 mL) and diluted with EtOH (100%, 135 mL). The reaction
mixture was then charged with a 55% dispersion of NaH in silicone
oil portionwise (5 g). The resulting solution was allowed to stir for
20 h at rt. The mixture was then neutralized by the addition of 1 N
HCl (25 mL), diluted with benzene (125 mL), and concentrated under
reduced pressure. The residue was partitioned between EtOAc (150
mL) and saturated NaHCO3 (125 mL), and the aqueous phase was
extracted thoroughly with EtOAc (3 × 150 mL). The combined extracts
were dried and concentrated to give a crude residue which was purified
by flash chromatography (20% EtOAc/hexanes) to afford 1.15 g (87%)
of enone 15 as a yellow semisolid. In a separate experiment purified
spirodiketone 14 was transformed into enone 15 in 92% yield: Rf )
1522 cm-1
.
(1S,3aS,7aS)-4′-Methoxybenzo[4,5]hydrindano[1,7a-c]-1-(tert-bu-
tyloxycarbonyl)pyrrolidine, 20 (R ) Boc). A solution of N-
benzylpyrrolidine 16 (700 mg, 2.10 mmol) in CH2Cl2 (15 mL) was
treated with p-nitrophenyl chloroformate (270 mg, 1.34 mmol) and the
mixture stirred for 0.5 h. Evaporation of solvent afforded a tan solid,
crude 18, which was dissolved in THF (20 mL) and then added to a
stirred rt solution of 1 M t-BuOK/THF (13.0 mL, 13.0 mmol). H2O
(72 mg, 4.0 mmol) was then added, and the dark red suspension which
immediately formed was stirred for 0.5 h. The mixture was quenched
with 1 N NaOH (10 mL) and diluted with EtOAc (150 mL). The
aqueous layer was removed and the organic layer washed with 1 N
NaOH (3 × 75 mL) to remove p-nitrophenol. The organic layer was
dried (MgSO4) and concentrated, producing a red oil which was
dissolved in EtOH (20 mL) and added to a predried flask under argon
containing 10% Pd/C (300 mg). The argon atmosphere was then
replaced with 1 atm of hydrogen, the mixture was stirred for 17 h and
filtered through Celite, and the filtercake was washed with EtOAc (20
mL). The combined rinsings were subjected to flash chromatography,
giving 480 mg (66%) of 20 as a clear colorless oil which exists as two
distinct N-Boc rotamers at rt. Recrystallization of 20 with heptane/
Et2O provided a crystalline sample. S-stereochemistry was confirmed
by the X-ray crystal structure:49 Rf ) 0.58 (1:1:1 ether/CH2Cl2/hexanes);
1
0.55 (EtOAc); mp 125-127 °C; [R]D ) -109 (c 1.37, benzene); H
NMR δ 7.59 (d, J ) 8.7 Hz, 1 H), 7.20-7.35 (m, 5 H), 6.79 (dd, J )
2.6, 8.7 Hz, 1 H), 6.67 (d, J ) 2.4 Hz, 1 H), 6.29 (s, 1 H), 3.80 (s, 3
H), 3.54 (ABq, J ) 13.5 Hz ∆V ) 16.2, 2 H), 3.27 (d, J ) 8.7 Hz, 1
H), 2.91 (m, 3 H), 2.41 (m, 2 H), 2.15 (m, 2 H), 2.04 (dq, J ) 5.9,
12.6 Hz, 1 H); 13C NMR δ 209.3 (CO), 175.3 (C), 161.7 (C), 140.3
(C), 138.2(C), 129.5 (CH), 128.4 (CH), 128.2 (CH), 126.8 (CH), 122.4
(C), 122.2 (CH), 113.6 (CH), 113.3 (CH), 61.7 (CH2), 59.0 (CH2), 56.7
(CH), 55.6 (CH2), 55.3 (CH3), 52.3 (C), 33.1 (CH2), 27.1 (CH2); IR
(film) ν 1690 cm-1; HRMS calcd for C23H23NO2 (M+ + 1) 346.1807,
found 346.1809.
1
mp 114-16 °C; [R]D ) -66 (c 1.00, benzene); H NMR (100 °C,
toluene-d8) δ 6.95 (d, J ) 9.3 Hz, 1 H), 6.65 (m, 2 H), 3.42 (s, 3 H),
3.39 (br d, J ) 3.1 Hz, 2 H), 3.00 (d, J ) 12.0 Hz, 1 H), 2.91 (d, J )
12.3 Hz, 1 H), 2.70 (m, 2 H), 2.51 (dd, J ) 12.0, 11.0 Hz, 1 H), 1.65-
1.90 (m, 5 H), 1.61 (m, 1 H), 1.45 (m, 1 H), 1.35 (s, 9 H); 13C NMR
(55 °C)50 δ 158.3 (C), 155.2 (C), 137.9 (C), 132.4 (C), 126.6 (CH),
114.0 (CH), 111.5 (CH), 79.1 (C), 55.5 (CH3), 54.3 (CH2), 52.3 (CH2),
50.0 (C), 48.9 (CH), 47.0 (CH), 32.7 (CH2), 31.9 (CH2), 28.8 (CH3)3C,
27.6 (CH2), 25.5 (CH2); IR (film) ν 2928, 1693, 1610, 1399 cm-1. The
X-ray sample was not submitted for combustion analysis.
(1S,7aS)-4′-Methoxybenzo[4,5]hydrindeno[1,7a-c]-1-benzylpyr-
rolidine, 16. According to the procedure of Brown and White32 a 1
M solution of LiAlH4 in THF (10 mL, 10 mmol) was added to a 0 °C
solution of AlCl3 (3.0 g, 22.5 mmol) in Et2O (20 mL). After stirring
for 15 min, a solution of enone 15 (1.1 g, 3.18 mmol) was added and
the mixture was allowed to warm to rt with stirring over 6 h. Excess
EtOAc (50 mL) was slowly added dropwise to quench excess AlH3,
and 1 N NaOH (50 mL) was then slowly added. The aqueous layer
was extracted with EtOAc (3 × 125 mL), and the organics were
combined, dried (MgSO4), and concentrated to give 1.00 g (95.2%) of
16 as an off-white semisolid judged to be pure by 1H NMR. An
analytical sample was prepared by recrystallization from EtOH: Rf )
0.60 (EtOAc); mp 107-9 °C; [R]D ) +47.5 (c 1.06, benzene); 1H NMR
δ 7.46 (d, J ) 8.6 Hz, 1 H), 7.20-7.32 (m, 5 H), 6.71 (dd, J ) 2.6,
8.6 Hz, 1 H), 6.61 (d, J ) 2.6 Hz, 1 H), 5.77 (s, 1 H), 3.78 (s, 3 H),
3.58 (ABq, J ) 13.3 Hz ∆V ) 59.3, 2 H), 2.70-2.95 (m, 5 H, includes
2.82 (d, J ) 9.0 Hz, 1 H)), 2.40-2.60 (m, 3 H), 2.17 (d, J ) 9.0 Hz,
1 H), 1.94 (ddd, J ) 2.0, 5.5, 12.4 Hz, 1 H), 1.82 (td, J ) 5.7, 12.5
Hz, 1 H); 13C NMR δ 158.6 (C), 143.6 (C), 139.2 (C), 137.1 (C), 128.6
(1S,3aS,7aS)-4′-Methoxybenzo[4,5]hydrindano[1,7a-c]-1-(tert-bu-
tyloxycarbonyl)-2(S)-methylpyrrolidine, 21. To a stirred solution of
N-Boc-pyrrolidine 20 (100 mg, 0.291 mmol) in ether (1 mL) over KH
(15 mg) was added TMEDA (45 uL, 0.298 mmol). The reaction
mixture, under dry deoxygenated argon, was stirred at rt for 15 min
and then cooled to -90 °C. The mixture was charged with a 1.20 M
solution of sec-BuLi in cyclohexane (0.75 mL, 0.90 mmol), producing
(49) The X-ray structures have been submitted to the Cambridge
Crystallographic Center. The ORTEP structure for 20 is also included in
the supporting information.
(50) The 13C NMR spectrum, recorded at 100 °C (toluene), did not show
any significant differences, in addition to solvent overlapping with the
sample.

Asymmetric Route to the Conanine BCDE Ring System
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9883
a dark red solution, and stirred for 16 h. Iodomethane (100 uL, 1.61
mmol) was added, and after 11 h at -90 °C the mixture was quenched
with CH3OH (1 mL) dropwise, warmed to rt, diluted with ether (25
mL), and washed with 1 N NaOH (2 × 20 mL). The organic layer
was dried (MgSO4) and concentrated to afford a yellow oil whose crude
1H NMR showed a 2:1 ratio of 21:20. This mixture was purified by
preparative TLC, eluting with 1:1:1 ether/CH2Cl2/hexanes to afford 46
mg (50%) of the more polar 21 as a clear oil: Rf ) 0.60 (1:1:1 ether/
CH2Cl2/hexanes); [R]D ) +12° (c 1.10, benzene); 1H NMR (300 MHz)
δ 6.90 (d, J ) 8.4 Hz, 1 H), 6.60 (m, 2 H), 3.93 (dq, J ) 6.3, 6.0 Hz,
1 H), 3.75 (s, 3 H), 3.05 (d, J ) 12.0 Hz, 1 H), 3.00 (m, 1 H), 2.82 (m,
1 H), 2.75 (d, 1 H, J ) 8.8 Hz), 2.20 (m, 1 H), 1.60-2.15 (m, 7 H),
1.30 (d, J ) 6.3 Hz, 3 H); 1H NMR (500 MHz) δ 6.90 (d, J ) 5.7 Hz,
1 H), 6.60 (m, 2H), 3.93 (dq, J ) 3.9, 3.6 Hz, 1 H), 3.75 (s, 3 H), 3.05
(d, J ) 7.5 Hz, 1 H), 2.90 (m, 1 H), 2.82 (m, 1 H), 2.70 (d, 1 H, J )
7.2 Hz), 2.20 (m, 1 H), 1.60-2.15 (m, 7 H), 1.30 (d, J ) 3.9 Hz, 3 H);
13C NMR δ 157.9 (C), 155.7 (C), 137.7 (C), 132.2 (C), 126.3 (CH),
113.7 (CH), 111.1 (CH), 79.0 (C), 55.3 (CH3), 55.0 (CH), 52.3 (CH2),
50.0 (C), 47.4 (2xCH), 32.3 (CH2), 30.9 (CH2), 28.8 ((CH3)3C), 27.4
(CH2), 24.8 (CH2), 15.0 (CH3); HRMS calcd for C22H31NO3 (M+)
357.2304, found 357.2311.
combined extracts were dried (Na2SO4) and concentrated to afford a
yellow oil which was subjected to preparative TLC, giving 26 mg (85%)
of 23 as a light yellow oil. Recrystallization with hexanes/EtOAc
provided a colorless crystalline sample: Rf ) 0.40 (5% Et3N/95%
1
EtOAc); [R]D ) +51° (c 1.05, benzene) mp 87-8 °C; H NMR δ
6.98 (d, J ) 9.6 Hz, 1 H), 6.64 (m, 2 H), 3.74 (s, 3 H), 2.83-2.96 (m,
2 H), 2.73 (dd, J ) 5.4, 12.3 Hz, 1 H), 2.65 (d, 1 H, J ) 10.5 Hz),
2.45 (dq, J ) 5.9, 6.3 Hz, 1 H), 2.10 (s, 3 H), 1.97-2.09 (m, 2 H),
1.75-1.95 (m, 4 H), 1.46-1.72 (m, 2 H), 1.07 (d, J ) 6.3 Hz, 3 H);
13C NMR δ 157.5 (C), 137.8 (C), 133.0 (C), 126.9 (CH), 113.6 (CH),
110.7 (CH), 64.0 (CH2), 63.9 (CH), 55.1 (CH3), 52.9 (CH), 50.9 (C),
50.1 (CH), 41.0 (CH3), 35.6 (CH2), 27.6 (CH2), 27.4 (CH2), 25.4 (CH2),
14.9 (CH3); HRMS calcd for C18H25NO (M + 1) 272.2014, found
272.2003.
Acknowledgment. The authors are grateful to the National
Institutes of Health for financial support of this work. The
authors also thank Steven Dock for preparation of the starting
materials. Furthermore, we thank Dr. Chris Rithner, Don Dick,
and Susie Miller for technical assistance with NMR, MS, and
X-ray determinations.
(1S,3aS,7aS)-4′-Methoxybenzo[4,5]hydrindano[1,7a-c]-1-methyl-
2(S)-methylpyrrolidine, 23. N-Boc-pyrrolidine 21 (40 mg, 0.112
mmol) was dissolved in dry ether (1.0 mL), and the mixture was charged
with excess LiAlH4 (1 M, 0.75 mL, 0.75 mmol) in THF and stirred at
rt for 8 h. The mixture was carefully quenched with H2O (1 mL),
followed by 3 N NaOH (1 mL) and then H2O (4 mL). The aqueous
phase was thoroughly extracted with EtOAc (3 × 15 mL), and the
Supporting Information Available: 1H and 13C NMR
spectra for 1, 5, 7, 9, 14-16, 20, 21, and 23 and an ORTEP
drawing for 20 (25 pages). See any current masthead page for
ordering and Internet access instructions.
JA961903O