Enantiospecific Formation of Trans 1,3-Disubstituted Tetrahydro-β-carbolines by the Pictet-Spengler Reaction and Conversion of Cis Diastereomers into Their Trans Counterparts by Scission of the C-1/N-2 Bond

Article abstract of DOI:10.1021/jo951170a

The factors which effect the stereoselective formation of trans-1-alkyl-2-benzyl-3-(alkoxycarbonyl)-1,2,3,4-tetrahydro-β-carbolines and trans-3-(alkoxycarbonyl)-1-alkyl-2-(diphenylmethyl)-1,2,3,4-tetrahydro-β- carbolines by the Pictet-Spengler cyclization were examined by heating tryptophan derivatives with aldehydes of varied steric bulk under aprotic and acidic conditions, followed by determination of the ratio of cis to trans diastereomers so formed. The presence of a benzyl group at the N_b-nitrogen atom alters the diastereochemical outcome of this condensation to provide 100% trans stereoselectivity when the cyclization is carried out with cyclohexanecarboxaldehyde. Furthermore, when N_b-(diphenylmethyl)tryptophan isopropyl ester was condensed with aldehydes of any size, trans diastereomers are formed with 100% stereoselectively. The trans N_b-substituted diastereomers are thermodynamically more stable than their cis congeners as shown by equilibration experiments in TFA. Conversion of the cis diastereomers into the more stable trans diastereomers is believed to occur under acidic conditions by cleavage of the carbon (C-1)-nitrogen (N-2) bond with complete retention of configuration at the C-3 stereocenter. Evidence from deuterium exchange experiments as well as optical rotations support this model for epimerization. In addition, when cis diastereomer 66a was allowed to stir in CF₃COOD, the trans isomer 66b was isolated in 90% yield, while treatment of cis 66a with CF₃COOH/NaBH₄ provided a mixture of the ring cleaved [scission across C(1)-N(2) bond] product 67 and the trans isomer 66b. Treatment of 66b (control experiment) with NaBH₄/CF₃COOH under the same conditions returned only starting trans 66b in excellent yield. The Pictet-Spengler reaction of substrates with sufficiently large substituants, followed by treatment with acid, permits the 100% enantiospecific formation of trans-1,3-disubstituted-1,2,3,4-tetrahydro-β-carbolines for alkaloid total synthesis.

Full text of DOI:10.1021/jo951170a

44
J . Org. Chem. 1997, 62, 44-61
En a n tiosp ecific F or m a tion of Tr a n s 1,3-Disu bstitu ted
Tetr a h yd r o-â-ca r bolin es by th e P ictet-Sp en gler Rea ction a n d
Con ver sion of Cis Dia ster eom er s in to Th eir Tr a n s Cou n ter p a r ts
by Scission of th e C-1/N-2 Bon d
Eric D. Cox, Linda K. Hamaker, J in Li, Peng Yu, Kevin M. Czerwinski, Li Deng,
Dennis W. Bennett, and J ames M. Cook*
Department of Chemistry, University of WisconsinsMilwaukee, Milwaukee, Wisconsin 53201
William H. Watson and Mariusz Krawiec
Department of Chemistry, Texas Christian University, Fort Worth, Texas 76129
Received J uly 25, 1995 (Revised Manuscript Received October 23, 1996X)
The factors which effect the stereoselective formation of trans-1-alkyl-2-benzyl-3-(alkoxycarbonyl)-
1,2,3,4-tetrahydro-â-carbolines and trans-3-(alkoxycarbonyl)-1-alkyl-2-(diphenylmethyl)-1,2,3,4-
tetrahydro-â-carbolines by the Pictet-Spengler cyclization were examined by heating tryptophan
derivatives with aldehydes of varied steric bulk under aprotic and acidic conditions, followed by
determination of the ratio of cis to trans diastereomers so formed. The presence of a benzyl group
at the Nb-nitrogen atom alters the diastereochemical outcome of this condensation to provide 100%
trans stereoselectivity when the cyclization is carried out with cyclohexanecarboxaldehyde.
Furthermore, when Nb-(diphenylmethyl)tryptophan isopropyl ester was condensed with aldehydes
of any size, trans diastereomers are formed with 100% stereoselectively. The trans Nb-substituted
diastereomers are thermodynamically more stable than their cis congeners as shown by equilibration
experiments in TFA. Conversion of the cis diastereomers into the more stable trans diastereomers
is believed to occur under acidic conditions by cleavage of the carbon (C-1)-nitrogen (N-2) bond
with complete retention of configuration at the C-3 stereocenter. Evidence from deuterium exchange
experiments as well as optical rotations support this model for epimerization. In addition, when
cis diastereomer 66a was allowed to stir in CF3COOD, the trans isomer 66b was isolated in 90%
yield, while treatment of cis 66a with CF3COOH/NaBH4 provided a mixture of the ring cleaved
[scission across C(1)-N(2) bond] product 67 and the trans isomer 66b. Treatment of 66b (control
experiment) with NaBH4/CF3COOH under the same conditions returned only starting trans 66b
in excellent yield. The Pictet-Spengler reaction of substrates with sufficiently large substituents,
followed by treatment with acid, permits the 100% enantiospecific formation of trans-1,3-
disubstituted-1,2,3,4-tetrahydro-â-carbolines for alkaloid total synthesis.
In tr od u ction
this enantiospecific Pictet-Spengler reaction has since
facilitated the total synthesis of a number of indole
alkaloid natural products including (-)-sauveoline,5,6 (-)-
raumacline,5,6 (-)-Nb-methylraumacline,5,6 (-)-alston-
erine,7 and (+)-macroline.8 The central intermediate in
the syntheses of all of these monomeric indole alkaloids
is the tetracyclic ketone 3; the enantiospecific Pictet-
Spengler reaction is the key step in its formation.9,10
Underlying interest in the enantiospecific Pictet-
Spengler reaction dates back to the work of Ungemach
et al.4,11 He was able to show that introduction of a
benzyl function at the Nb-nitrogen moiety permitted
control of the stereochemical outcome of this cyclization.
Condensation of (()-Nb-benzyltryptophan methyl ester
with cyclohexanecarboxaldehyde under nonacidic aprotic
The Pictet-Spengler reaction has long been an impor-
tant reaction for the syntheses of both indole and iso-
quinoline alkaloids.1 Discovered in 1911 by Ame´ Pictet
and Theodor Spengler, the reaction was first utilized to
prepare simple tetrahydroisoquinolines by heating phen-
ethylamine, phenylalanine, and tyrosine with methylal,
individually, in the presence of hydrochloric acid.2 It was
Tatsui, apparently, who first utilized this method with
indole bases, specifically for the synthesis of 1-methyl-
1,2,3,4-tetrahydro-â-carboline (tetrahydroharmane).3 The
condensation by Ungemach et al. of Nb-benzyltryptophan
methyl ester with aldehydes of large steric size was
significant for it led to the 100% stereoselective formation
of trans-1,3-disubstituted diastereomers (Scheme 1) in
the methyl ester series.4 Furthermore, development of
(5) Fu, X.; Cook, J . M. J . Am. Chem. Soc. 1992, 114, 6910.
(6) Fu, X.; Cook, J . M. J . Org. Chem. 1993, 58, 661.
(7) Zhang, L. H.; Cook, J . M. J . Am. Chem. Soc. 1990, 112, 4088.
(8) Bi, Y.; Cook, J . M. Tetrahedron Lett. 1993, 34, 4501.
(9) Zhang, L. H.; Bi, Y.; Yu, F.; Menzia, G.; Cook, J . M. Heterocycles
1992, 34, 517. Previously a typographical error led to the report of the
rotation of the trans isomer in the 3(R)10 series as -36° rather than
the correct -56 ( 2°. The optical rotation of 64b [3(S) series] was
+54.8°; moreover, the tetracyclic ketone prepared from the trans isomer
in both the 3(R) and 3(S)10 series had the same optical rotation as
reported, although the sign was opposite.
(10) Zhang, L. H.; Cook, J . M. Heterocycles 1988, 27, 1357, 2795.
(11) Ungemach, F.; Soerens, D.; Weber, D.; DiPierro, M.; Campos,
O.; Mokry, P.; Cook, J . M.; Silverton, J . V. J . Am. Chem. Soc. 1980,
102, 6976.
X Abstract published in Advance ACS Abstracts, December 15, 1996.
(1) Whaley, W. M.; Govindachari, T. R. In Organic Reactions;
Adams, R., Ed.; J ohn Wiley and Sons: New York, 1951; Vol. VI, pp
151-190. Czerwinski, K. M.; Cook, J . M. In Advances in Heterocyclic
Natural Products Synthesis; Pearson, W., Ed.; J AI Press: Greenwich,
CT, 1996; Vol. 3, pp 217-277. Hamaker, L. H.; Cook, J . M. In
Alkaloids: Chemical and Biological Perspectives; Pelletier, S. W., Ed.;
Elsevier Science: New York, 1995; Vol. 9, pp 23-85.
(2) Pictet, A.; Spengler, T. Ber 1911, 44, 2030.
(3) Tatsui, G. J . Pharm. Soc. J pn. 1928, 48, 92.
(4) Ungemach, F.; DiPierro, M.; Weber, R.; Cook, J . M. J . Org. Chem.
1981, 46, 164.
S0022-3263(95)01170-4 CCC: $14.00 © 1997 American Chemical Society

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 45
Sch em e 1. En a n tiosp ecific Syn th esis of th e Tetr a cyclic Keton e 3 by th e P ictet-Sp en gler Con d en sa tion
Ta ble 1. Cis:Tr a n s Ra tiosa of 1,2-Disu bstitu ted a n d 1,2,3-Tr isu bstitu ted Tetr a h yd r o-â-ca r bolin es
compd no.
R1
R2
R3
CH3
CH3
CH3
CH3
CH3
CH3
CH(CH3)2
CH(CH3)2
CH(CH3)2
CH3
CH3
CH3
CH(CH3)2
CH(CH3)2
CH(CH3)2
nonacidic aprotic cis:trans
acidic cis:trans
4a ,4b
5a ,5b
6a ,6b
7a ,7b
8a ,8bc
9
10a ,10b
11a ,11b
12
13a ,13b
14
15
16
17
CH3
H
H
H
Bn
Bn
Bn
Bn
Bn
Bn
NA
NA
75:25
43:57b
41:59b
12:88
11:89c,e
0:100
13:87
12:88
0:100
0:100
0:100
0:100
0:100
0:100
0:100
CH2CH3
C6H11
40:60
26:74
23:77
0:100
23:77
13:87
0:100
10:90
0:100
0:0d
CH3
CH2CH2CH3
C6H11
CH3
CH2CH2CH3
C6H11
CH3
CH(Ph)2
CH(Ph)2
CH(Ph)2
CH(Ph)2
CH(Ph)2
CH(Ph)2
CH2CH2CH3
C6H11
CH3
0:100
CH2CH2CH3
C6H11
0:0d
18
0:0d
a
b
(3% as determined by integration of the 1H NMR spectrum (compounds labeled “a” are the cis diastereomer and “b” the trans). See
ref 11. c Carried out in the optically active series. These transformations yielded only starting materials under nonacidic aprotic conditions
but were converted to tetrahydro-â-carbolines upon addition of TFA to the reaction medium. e When the equilibration was effected directly
on the reaction mixture in benzene in the presence of aldehyde the ratio was determined as reported here. Pure 8a in CF3COOH was
completely converted into 8b in CH2Cl2.
d
conditions resulted in the 100% stereoselective formation
of (()-trans-2-benzyl-3-carbomethoxy-1-cyclohexyl-1,2,3,4-
tetrahydro-â-carboline as noted above.12,13 The advan-
tage of this method centered on exclusive formation of
the trans isomer after which the benzyl group could be
easily removed by hydrogenolysis. The stereochemical
configuration at C-1 could then be determined by 13C
NMR spectroscopy.11 This afforded one diastereomeric
tetrahydro-â-carboline in high yield as opposed to the
mixture of diastereomers traditionally formed in Pictet-
Spengler condensations.1 The use of a benzyl group as
a substituent at the Nb-nitrogen atom inverted the
diastereochemical outcome of the Pictet-Spengler reac-
tion from preferential formation of the cis isomer1 to the
trans isomer stereoselectively. Although diastereochemi-
cal product ratios had previously been reported11,14,15 for
the Pictet-Spengler reaction of Na-H, Nb-H derivatives,
systematic variation of the steric bulk of the substituents
in the Pictet-Spengler condensation of Nb-benzyltryp-
tophan derivatives was undertaken to determine the
stereochemical scope of the process.16,17 Examination of
the results of these studies, presented here, permit the
100% stereospecific formation of any 3-(alkoxycarbonyl)-
1-alkyltetrahydro-â-carboline regardless of the size of the
desired alkyl substituent at position-1. This is especially
important since both D-(+)-tryptophan and L-(-)-tryp-
tophan are readily available resulting in the enantiospe-
cific synthesis of either antipode via this process. Fur-
thermore, evidence is presented here to suggest that
conversion of the thermodynamically less stable cis
diastereomers into their trans isomers occurs by cleavage
of the carbon (position-1)-nitrogen (position-2) bond
followed by recyclization to the trans isomer.
Resu lts a n d Discu ssion
Previous reports highlighted the variance in the dias-
tereochemical product ratios obtained when employing
the Pictet-Spengler reaction.11,14,15 Three such examples
are shown in Table 1 (4-6). The ratios of diastereomers
when acetaldehyde, propionaldehyde, and cyclohexane-
carboxaldehyde were reacted with tryptophan methyl
ester under acidic conditions reportedly varied from 75:
25-cis:trans (4a :4b), to a more equal ratio of 43:57-cis-
trans (5a :5b), to 41:59-cis-trans (6a :6b). (The terms cis
and trans here refer to the stereocenters at C-1 and C-3
according to the â-carboline numbering convention; two
representative structures are shown at the top of Table
1.) The only significant difference between these reac-
tions is the size of the substituent at position C-1 of the
tetrahydro-â-carboline.
(12) Brown, R. T.; Chapple, C. L. J . Chem. Soc., Chem. Commun.
1973, 886.
(13) De Silva, K. T. D.; King, D.; Smith, G. N. J . Chem. Soc., Chem.
Commun. 1971, 908.
(14) Soerens, D.; Sandrin, J .; Ungemach, F.; Mokry, P.; Wu, G. S.;
Yamanaka, E.; Hutchins, L.; DiPierro, M.; Cook, J . M. J . Org. Chem.
1979, 44, 535.
(15) Brossi, A.; Focella, A.; Teitel, S. J . Med. Chem. 1973, 16, 418.
(16) Deng, L.; Czerwinski, K. M.; Cook, J . M. Tetrahedron Lett. 1991,
32, 175.
To examine the effect the steric bulk of the incoming
aldehyde had upon the diastereomeric ratio realized in
the Pictet-Spengler condensation in the presence of a
(17) Czerwinski, K. M.; Deng, L.; Cook, J . M. Tetrahedron Lett. 1992,
33, 4721-4724.

46 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
benzyl group at the amino nitrogen function, Nb-benzyl-
tryptophan methyl ester was prepared18 and reacted with
acetaldehyde, butyraldehyde, and cyclohexanecarboxal-
dehyde individually under the conditions of Soerens et
al.4 This afforded a series of 1-alkyl-2-benzyl-3-(meth-
oxycarbonyl)tetrahydro-â-carbolines in which the sub-
stituent at position-1 varied in steric bulk from the
relatively small methyl group, to propyl, to the larger
cyclohexyl moiety.
dehyde (23:77-cis 8a :trans 8b) revealed that the ratio of
cis to trans for both was essentially the same. As the
steric bulk of the substituents increased to cyclohexyl,
however, the cis isomer was excluded from formation.
Ungemach proposed this phenomenon was the result of
a steric interaction between the ester function at C-3 and
the alkyl group at C-1 in the syn spiroindolenine inter-
mediate, as well as A-strain in the six-membered rear-
ranged ring.4
The reactants were heated in benzene at reflux under
nonacidic aprotic conditions for 36 h. All the condensa-
tions were stopped after 36 h to provide uniform experi-
mental conditions, thus enabling an unbiased comparison
of the stereochemical effects of the various substituents.
Chemical yields are thus unoptimized. The reactions
utilizing acetaldehyde were carried out in a sealed glass
tube to avoid loss of the aldehyde. Each aldehyde was
purified by vacuum distillation and the IR spectrum
measured to verify that the reactants were indeed free
of acid. In all cases, after 36 h a small aliquot was
removed and the proton NMR spectrum recorded. It was
sometimes difficult to accurately measure the diastereo-
meric ratio due to impurities which masked the signals.
In these cases the reaction mixture was passed through
a short wash column of silica gel to remove the impuri-
ties, and the NMR spectrum was recorded once again.
Comparison of the spectra before and after chromatog-
raphy assured that the diastereochemical ratio had not
been altered. Flash chromatography on silica gel permit-
ted separation of the cis and trans diastereomers, and
close examination of the NMR spectra of the individual
diastereomers clearly illustrated which proton signals
differed significantly between diastereomers. These
signals were then used to determine the diastereomeric
ratio of cis to trans isomers in the spectra of the mixtures.
Although CDCl3 was initially employed as the solvent for
NMR spectroscopy, it was found that acidic impurities
were in high enough concentration to effect epimerization
of the cis diastereomers into their trans congeners. The
cause of this epimerization will be expanded upon later.
These impurities could be removed but it was found
easier to employ C6D6 as the solvent for spectroscopy. Not
only was it free of acidic impurities but its use generally
provided a greater anisotropy between the critical signals
in the proton NMR spectra of the cis and trans diaster-
eomers. Catalytic transfer hydrogenation of the 1,2,3-
trisubstituted species afforded 1,3-disubstituted-tetrahy-
dro-â-carbolines that are generally well known compounds,
consequently, the stereochemistry of the individual dia-
stereomers could be unequivocally established by com-
parison of their 13C NMR spectra.11 Determination of the
stereochemistry of these individual diastereomers had
been carried out previously according to the method of
Ungemach et al., a method which has been validated by
independent sources.11,19
Since aldehydes similar in size to an n-butyl group
provided a mixture of diastereomers the synthetic po-
tential of the process was limited. If Ungemach’s hy-
pothesis were correct, however, any increase in steric
interactions between the ester and alkyl groups should
shift the diastereomeric ratios toward increased trans
stereoselectivity. To this end, replacement of the methyl
ester function at C-3 with the larger isopropyl group was
undertaken. The Nb-benzyltryptophan isopropyl ester
was synthesized in a fashion analogous to the methyl
ester,14 after which the previously mentioned series of
condensations was repeated, and the diastereomeric
ratios were examined by proton NMR spectroscopy.
Comparison of the diastereochemical ratios obtained in
both series (7-9 and 10-12) showed that the condensa-
tion of Nb-benzyltryptophan isopropyl ester with butyral-
dehyde yielded an increased amount of the trans isomer.
In the methyl ester series 8, the ratio of cis 8a to trans
8b was 23:77 while in the isopropyl ester series 11 the
quantity of trans isomer 11b was increased to 13:87-cis:
trans. The stereochemical configuration of the 1-cyclo-
hexyl-3-isopropyl ester 12 was confirmed as trans by
single crystal X-ray crystallography.
The previous work in this area clearly demonstrated
that introduction of a benzyl group at the Nb-nitrogen
function favored formation of the trans diastereomer.
Substituents at the Nb-nitrogen atom have come under
scrutiny previously, and it has been shown by Sandrin20
et al. that electronic effects21 play a significant role in
the stereochemical outcome of the reaction. The effect
the steric bulk of the Nb-alkyl function had upon the
diastereochemical ratio had not been studied in detail.
In order to increase the size of the substituent at the Nb-
nitrogen atom without altering the electronic character
of the group a diphenylmethyl moiety was employed.
Introduction of this group was accomplished by trans-
imination of the appropriate ester hydrobromide or
hydrochloride salt with benzophenone imine according
to the procedure of Polt and O’Donnell.22 Reduction
(Scheme 2) of the resultant imine with sodium cy-
anoborohydride in methanol under acidic conditions
provided the desired Nb-diphenylmethyl-substituted tryp-
tophan derivative. This protocol provided excellent
yields, could easily be scaled up to the 10 g level, and
was not subject to disubstitution or to the formation of
quaternary ammonium salts. With the methyl and
isopropyl esters of Nb-(diphenylmethyl)tryptophan (27,
28) in hand, the series of condensations was repeated
(benzene at reflux), and the diastereomeric ratios were
measured by NMR spectroscopy. In the methyl ester
series, the condensation with acetaldehyde furnished the
The condensation of Nb-benzyltryptophan methyl ester
with acetaldehyde and butyraldehyde (individually) re-
sulted in the formation of mixtures of cis and trans
diastereomers. This condensation using cyclohexanecar-
boxaldehyde furnished only the trans diastereomer as
was expected.4 Examination of the product ratios for the
condensations of Nb-benzyltryptophan methyl ester with
acetaldehyde (26:74-cis 7a :trans 7b) and with butyral-
1
desired products (13a , 13b) as evidenced by H and 13C
(20) Sandrin, J .; Hollinshead, S. P.; Cook, J . M. J . Org. Chem. 1989,
54, 5636.
(18) Zhang, L. H. Ph.D. Thesis, University of WisconsinsMilwaukee,
1990.
(19) To´th, G.; Sza´ntay, C.; Kalaus, G.; Thaler, G.; Snatzke, G. J .
Chem. Soc., Perkin Trans. 2 1989, 1849.
(21) Hermkens, P. H. H.; van Maarseveenk, J . H.; Cobben, P. L.
H.; Ottenheijm, H. C. J .; Kruse, C. G.; Scheeren, H. W. Tetrahedron
1990, 46, 833-846.
(22) O’Donnell, M. J .; Polt, R. L. J . Org. Chem. 1982, 47, 2663.

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 47
Sch em e 2. Syn th esis of 1,2,3-Tr isu bstitu ted
Sch em e 3
Tetr a h yd r o-â-ca r bolin es
NMR spectroscopy after initial removal of the non-indolic
byproducts. However, attempts to isolate the cis dias-
tereomer 13a were unsuccessful. Only the trans dias-
tereomer 13b was found upon examination of the chro-
matography fractions. Moreover, the reaction of Nb-
(diphenylmethyl)tryptophan isopropyl ester with acet-
aldehyde provided only the trans diastereomer 16. In
the methyl ester series the reaction of butyraldehyde also
provided only the trans diastereomer 14. In three of the
cases (15, 17, and 18) no reaction was observed under
nonacidic aprotic conditions after 36 h. Prolonged heat-
ing only increased the amount of byproducts which were
formed. In order to effect cyclization it was necessary to
add TFA to the reaction medium. The addition of TFA
to catalyze the formation of 15, 17, and 18 resulted in
the formation of only the trans diastereomer for each.
Although TFA is known to cleave some diphenylmethyl
groups,23-25 at low concentrations it failed to do so here
as evidenced by control reactions executed in the absence
of aldehyde. For example, compounds 15, 17, 18, 27, and
28 were individually stirred under conditions identical
to those employed in their formation in the absence of
aldehyde and were monitored daily by TLC for a period
of five days with no changes observed in retention factor
or proton NMR spectrum. In all of the reactions which
employed the diphenylmethyl group, the chemical yields
are somewhat lower than the corresponding reactions in
the benzyl cases. This is not unreasonable since the
presence of such bulky moieties should retard the rate
of the Pictet-Spengler condensation.23-25 Under these
conditions it was also observed that the amount of
polymeric material increased due to self condensation of
the aldehydes. This made chromatographic separations
more difficult and further contributed to the decreased
chemical yields.
the cis isomer as the major product. Introduction of a
benzyl function at the Nb-nitrogen atom inverted the
stereochemical outcome to provide the trans diastereomer
under nonacidic aprotic conditions (7--26:74--cis:trans) as
the major product. Furthermore, when the Nb-substitu-
ent was increased to the large diphenylmethyl group
under similar conditions, a 100% stereoselective forma-
tion of the trans diastereomer 16 was observed in the
isopropyl ester case.
This high trans diastereoselectivity under nonacidic
aprotic conditions presumably results from the increased
energy of the transition states involving the cis spiroin-
dolenine intermediates (Scheme 3). Two distinct spiroin-
dolenine intermediates may result from attack on the
iminium ion double bond. In the Nb-(diphenylmethyl)-
tryptophan methyl ester case, attack of the iminium ion
32 from the face opposite the ester function would provide
the anti spiroindolenine 33. Attack from the same face
as that occupied by the ester would result in a spiroin-
dolinine of syn configuration such as 34. Molecular
mechanics calculations combined with conformational
searching (MacroModel 2.5 - MM2 force field) revealed
that the anti configuration is 2.1 kcal/mol lower in energy
than the syn configuration.17 It is the anti isomer which
rearranges to provide the trans diastereomer.
A comparison of the cyclizations which employed
acetaldehyde show a marked trend. In the Nb-H case (4-
-75:25--cis:trans), the Pictet-Spengler reaction yielded
The data from the experiments under nonacidic aprotic
conditions reflect ratios which are the result of kinetic
trapping experiments. As noted earlier, however, if the
mixtures of diastereomers realized in the nonacidic
(23) Pless, J . Helv. Chim. Acta 1976, 59, 499.
(24) Hanson, R. W.; Law, H. D. J . Chem. Soc. 1965, 7285.
(25) Zervas, L.; Photaki, I. J . Am. Chem. Soc. 1962, 84, 3887.

48 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
Sch em e 4. Ep im er iza tion of th e Ster eocen ter a t
C-1 by Clea va ge of th e C-1/N-2 Bon d u n d er Acid ic
Con d ition s
Sch em e 5. Olefin P r oton a tion Mech a n ism for
Ep im er iza tion of th e Ster eocen ter a t C-1
aprotic Pictet-Spengler reaction were exposed to a
proton source, the diastereomeric ratios shifted to further
increase the amount of trans isomer. This generality was
confirmed by either the addition of TFA to a small aliquot
of the reaction mixture or as in cases 15, 17, and 18, by
reaction in the presence of TFA to catalyze the cycliza-
tion. Examination of the ratios of 7a to 7b, 8a to 8b,
10a to 10b, and 13a to 13b clearly indicated that the
ratio of cis diastereomer to trans diastereomer formed
under nonacidic aprotic conditions was higher than that
realized under acidic conditions. The method by which
the conversion of the cis diastereomers into the more
thermodynamically stable trans diastereomers takes
place can now be rationalized.
The work of Zhang10 suggested that scission of the C-1/
N-2 bond was involved in the cis to trans isomerization.
He isolated a mixture of two byproducts formed in the
Pictet-Spengler condensation of (D)(+)-Na-methyl-Nb-
benzyltryptophan methyl ester with methyl 3-formylpro-
pionate. These byproducts, an olefin and a methyl ether,
were heated as a mixture in 1% anhydrous methanolic
hydrogen chloride after which the trans diastereomer was
isolated as the sole product in 70% yield. The most
probable intermediate from which all four of the isolated
species could arise is one that is carbocationic in nature.
Zhang proposed10 (Scheme 4) that the mechanism by
which these byproducts were formed was preceeded by
protonation of the Nb-nitrogen atom with concomitant
cleavage of the carbon (position-1)-nitrogen (position-2)
bond. The carbocation so generated could undergo
elimination to yield an olefin, attack by the solvent to
form the ether, or after rotation of the C-1/C-10 carbon-
carbon bond, recyclize to provide the trans diastereomer
now epimeric at C-1. In addition, this appears to be the
same mechanism of isomerism whereby reserpine is
converted into isoreserpine as well as the epimerization
of the C-1 stereocenter of (-)-1,2,3,4-tetrahydroroe-
harmine enroute to racemic material.10,26,27
Sch em e 6
does not account for the formation of the two byproducts
which Zhang observed, it could not be discounted without
further examination.
In order to test the model for carbon-nitrogen bond
cleavage versus the olefinic protonation model, two
experiments were carried out. In the first, optically
active Nb-benzyltryptophan methyl ester was synthesized
from D-(+)-tryptophan according to the method employed
by Zhang for the analogous Na-methyl compound.18 This
material was heated under the nonacidic aprotic condi-
tions of benzene at reflux under an atmosphere of
nitrogen gas with freshly distilled butyraldehyde to
provide the kinetic ratio of optically active diastereomers
as a 23:77-cis:trans mixture. Separation of the diaster-
eomers on silica gel with CH2Cl2 by gravity chromatog-
raphy provided the pure cis 8a and trans 8b isomers. The
specific rotation [R]25 of the (-)-trans diastereomer 8b
D
was measured as -50.4° (c ) 1, benzene) and the cis 8a
as [R]25 ) -13.3° (c ) 1, benzene). The (-)-cis diaste-
D
reomer 8a was then taken up in CH2Cl2 and stirred with
2 equiv of TFA (Scheme 6). Examination of the mixture
by TLC after 12 h revealed that 8a had been completely
converted into the trans compound 8b. Workup with
mild base (NaHCO3) and chromatography on a short
wash column of silica gel (CH2Cl2) yielded the pure (-)-
Another possible mechanism for this cis-trans epimer-
ization involves the initial protonation of the indole
3-position and subsequent formation of an olefinic species
(Scheme 5). This olefin could then be attacked prefer-
entially from one face to yield the thermodynamically
more stable trans diastereomer. Although this model
trans diastereomer. The specific rotation [R]25 of this
D
sample was -50.4° (c ) 1.6, benzene). The specific
rotations of the two (-)-trans compounds were identical
(as well as the NMR and mass spectra) and presumably
the epimerization that Zhang observed10 across the C(1)-
N(2) had occurred here as well. Epimerization of the
stereocenter at C-3 had not occurred; had epimerization
occurred at C-3 then the enantiomer of 8b should have
(26) Searle, P. A.; Molinski, T. F. J . Org. Chem. 1994, 59, 6600-
6605.
(27) Reddy, M. S.; Cook, J . M. Tetrahedron Lett. 1994, 35, 5413.

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 49
has rapidly increased.30 In 1988 Zhang proposed that
an intermediate such as 44 (X ) H) was involved in the
epimerization of cis Nb-benzyl-1,3-disubstituted tetrahy-
dro-â-carbolines into their trans diastereomers when
heated in methanolic hydrogen chloride.10 Although
three potential mechanisms would account for this
isomerization, Zhang trapped byproducts related to 44
(X ) H) in poor yield and converted this material under
analogous conditions (CH3OH, 1% HCl, ∆) into the trans
diastereomer exclusively.10
The three potential intermediates depicted in Figure
1 are related to three different mechanistic pathways
earlier investigated by J oule with regard to the isomer-
ization of reserpine into isoreserpine.31 In 1989 evidence
from our laboratory was provided to indicate the isomer-
ization of reserpine into isoreserpine took place via C(1)-
N(2) bond session related to 44 rather than an iminium
ion related to 46.32 Experiments with reserpine in CF3-
COOD pioneered by J oule31 unequivocally ruled out the
intermediacy of an olefinic intermediate related to 45
during the isomerization of reserpine to isoreserpine.
Deuterium would have been incorporated at C(1) of the
tetrahydro-â-carboline nucleus during regeneration of the
2,3-indole double bond. In fact, in all of the epimerization
reactions described here (see below), stirring a cis-Nb-
benzyl-1,3-disubstituted tetrahydro-â-carboline in CF3-
COOD provided the trans isomer exclusively with no
incorporation of deuterium at C(1). These results rule
out an intermediate such as 45 during the process of
epimerization.
More recently, the optically active 6-methoxy-Na-meth-
yl-Nb-benzyltryptophan ethyl ester 47 was heated with
a slight excess of R-ketoglutaric acid to provide the trans
diastereomer 48 stereospecifically and in excellent yield
(Scheme 7).29 It is believed the electrons on the 6-meth-
oxy moiety of the indole contributed to the stabilization
of intermediate 49A via oxonium ion contributor 49B as
well as iminium ion form 49C. If the cis diastereomer
did form during the Pictet-Spengler reaction, it was felt
the methoxy moiety would help to stabilize intermediate
49A via 49B and 49C to facilitate C(1)-N(2) cleavage
and epimerization to the thermodynamically more stable
trans diastereomer 48.29
Since the carbocationic intermediate 49A was stabi-
lized by 49B and 49C, attention turned toward the
synthesis of a 6-methoxytryptophan analog which might
retard electronic stabilization of these resonance forms.
The Na-sulfonamido-protected 6-methoxy-Nb-benzyl-(D)-
tryptophan ethyl ester 5033 was chosen for this purpose.
The sulfonamide moiety on the indole Na-nitrogen func-
tion should withdraw electron density from the indole
Na-nitrogen atom and decrease the stability of carboca-
tionic intermediates 49D and 49E. The Na-sulfonami-
dotryptophan ethyl ester 50 prepared earlier33 was
stirred with benzaldehyde at room temperature followed
by reduction of the resulting imine with sodium borohy-
dride at -5 °C (low temperature to ensure against
racemization10) to provide the Nb-benzyl-6-methoxytryp-
tophan analog 51 (Scheme 8).
F igu r e 1. Potential intermediates.
been formed and the rotation would then have been
opposite in sign, namely [R]25 ) +50.4°.
D
To test the olefinic mechanism of cis-trans epimer-
ization, the same experiment was repeated with 2 molar
equiv of pure CF3COOD. If the olefin protonation mech-
anism was in operation, then deuterium would be incor-
porated into the tetrahydro-â-carboline at C-1 upon
reprotonation (D+) of the olefin to reform the tetrahydro-
â-carboline. Repke et al. have shown that incorporation
of deuterium at position-1 is found to occur when an
olefinic mechanism of isomerization is found to be in
operation.28 We found no evidence for deuterium incor-
1
poration at C-1 by integration of the H NMR spectrum
nor by mass spectrometry.
It is clear that the isomerization of the cis diastereo-
mers to the more thermodynamically stable trans dia-
stereomers does not take place by the olefinic interme-
diate 45 (Figure 1). In the experiments which involve
Na-H derivatives 8a and 8b, it was observed that upon
treatment of the cis isomer 8a with acid, cis 8a was
completely converted into the trans diastereomer 8b.
However, treatment of the trans diastereomer 8b under
similar conditions provided only trans 8b. Furthermore,
when optically active trans isomer 8b was heated in TFA
for 12 h and the reaction monitored hourly by TLC, cis
isomer 8a was not observed. After prolonged heating no
cis isomer 8a was observed in the mixture. Moreover,
there was no evidence of the original starting material
(Na-H-Nb-benzyltryptophan methyl ester) which would be
indicative of a retro-Pictet-Spengler process.29 From
these experiments and the epimerizations illustrated in
Table 1, it was concluded that the trans diastereomers
are thermodynamically more stable than their cis coun-
terparts in the Na-H, Nb-alkyl series. When optically
pure Nb-benzyl-D-(+)-tryptophan methyl ester and bu-
tyraldehyde were again heated to reflux in benzene until
TLC indicated the consumption of all the indolic starting
material, it was found that the trans diastereomer 8b
could be formed exclusively upon addition of an excess
of TFA to the reaction mixture at room temperature. The
epimerization of the cis diastereomer 8a to the trans
diastereomer 8b was monitored by TLC (74% isolated
yield).
With the advent of asymmetric control,1,30 the impor-
tance of the Pictet-Spengler condensation, as pointed
out, for the stereospecific synthesis of indole alkaloids
The Pictet-Spengler reaction of sulfonamide 51 and
R-ketoglutaric acid (1.05-2.10 equiv) in toluene/dioxane
was allowed to reflux for 44 h to provide a mixture of cis
(28) Repke, D. B.; Artis, D. R.; Nelson, J . T.; Wong, E. H. F. J . Org.
Chem. 1994, 59, 2164.
(29) Hamaker, L. H.; Cook, J . M. 208th American Chemical Society
National Meeting, Washington D.C., 1994, Abstract No. 269. Hamaker,
L. K. Ph.D. Thesis, University of WisconsinsMilwaukee, 1995.
(30) Cox, E. D.; Cook, J . M. Chem. Rev. 1995, 95, 1797.
(31) Gaskell, A.; J oule, J . Tetrahedron 1967, 23, 4053.
(32) Zhang, L.-H.; Gupta, A.; Cook, J . M. J . Org. Chem. 1989, 54,
4708.
(33) Allen, M. S.; Hamaker, L. K.; Laloggia, A.; Cook, J . M. Synth.
Commun. 1992, 22, 2077.

50 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
Sch em e 7
Sch em e 8
(52a ) and trans (52b) acids in a ratio of 49:51 (Scheme
2), respectively. When the reaction mixture was heated
for a longer period of time (91 h), the ratio of trans acid
52b to cis acid 52a increased slightly; however, the cis
isomer 52a diminished until it was no longer detected
(TLC). It should be noted that the actual yield of the
trans acid 52b did not increase as the reaction solution
was heated at length presumably due to the decomposi-
tion of the cis isomer 52a rather than epimerization of
cis 52a into trans 52b. In contrast to the results in the
Na-methyl-6-methoxytryptophan case 47,29 in the Na-
sulfonamido series the addition of excess R-ketoglutaric
acid did not promote the acid-catalyzed epimerization of
the cis isomer 52a into the trans diastereomer 52b. The
sulfonamide group had effectively retarded the acid-
catalyzed epimerization of the cis isomer 52a into trans
compound 52b in the Pictet-Spengler reaction itself in
stark contrast to the 100% stereoselective formation29 of
the trans isomer in the Na-methyl-Nb-benzyl-6-methoxy-
tryptophan 48 series.
gen chloride in the presence of triethyl orthoformate for
45 min to furnish the mixture of esters 53a (cis) and 53b
(trans) still in an approximately 1:1 ratio. Attempts were
again made to effect the epimerization [at C(1)] of the
cis diester 53a into the Na-sulfonamido trans diester 53b.
When the mixture of cis and trans diesters (53a ,b) was
heated in refluxing 2% ethanolic hydrogen chloride for 3
h to effect epimerization, analysis of the reaction mixture
(TLC) did not indicate the presence of increased amounts
of the trans diester 53b but rather 6-methoxy-Na-sul-
fonamido-Nb-benzyltryptophan ethyl ester 51 was formed
in 76% yield. This material was identical to the ethyl
ester 51 prepared by an independent method.33 The
sulfonamido group on the Na-nitrogen function of the cis
isomer 11a had clearly prevented epimerization presum-
ably by destabilization of a carbocationic intermediate
related to 49D. The cis 53a and trans 53b diesters were
then heated separately in ethanolic hydrogen chloride for
3 h, and workup of each reaction provided the same Nb-
benzyltryptophan derivative 51. Since the sulfonamido
group withdrew electron density from the indole nitrogen
function and destabilized the carbocationic intermediate
The mixture of cis 52a and trans 52b acids was
esterified upon heating in refluxing 2% ethanolic hydro-

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 51
Sch em e 9
Sch em e 10
which would result from C(1)-N(2) cleavage, a retro
Pictet-Spengler (type) process (see 56 f 57 f 58) has
instead occurred under these conditions, as illustrated
in Scheme 9. Donation of the electron density from the
7-methoxy group presumably facilitated protonation at
the R-carbon atom of the indole double bond to provide
56. This protonated indole could then undergo C-C bond
cleavage to furnish iminium ion 58 (path a) or bond
migration (path b, see 57) followed by rearrangement into
the same iminium ion 58. Hydrolysis of the CdN bond
of iminium ion 58 in ethanol would furnish the Nb-benzyl-
6-methoxytryptophan ethyl ester analog 51 (Scheme 9).
Even upon heating in excess acid (at length) at no time
did any of the cis Na-H (see 37) or cis Na-CH3 diastere-
omers described herein provide the related Na-H or Na-
CH3-Nb-benzyltryptophan derivatives which would arise
from the retro-Mannich (Pictet-Spengler) process. This
provides evidence (although indirect) that the iminium
ion pathway 56 to 58 is not operating in any of these
epimerizations with the exception of the Na-sulfonamido-
substituted (deactivated) series of Scheme 9.
For comparison purposes, the Pictet-Spengler reaction
of Na-methyl-Nb-benzyl-6-methoxy-(D)tryptophan ethyl
ester29 with an aldehyde under conditions of kinetic
trapping in nonacidic aprotic media was investigated.
Analogous to the method of Zhang,9,10 ethyl 3-formyl
propionate 59 was prepared for the Pictet-Spengler
reaction and dissolved with the Na-methyl-Nb-benzyl-6-
methoxytryptophan derivative 47 in benzene (Scheme
10). The solution was heated at reflux for 10 h, and this
resulted in a mixture of trans 60b to cis 60a diesters in
a ratio of 61:39 (Scheme 10). More importantly, when the
cis and trans diesters 60a ,b were dissolved in a mixture
of methylene chloride/trifluoroacetic acid and stirred at
room temperature (90 minutes), a 96% yield of the trans
diester 60b was realized (TLC, silica gel, Rf cis ) 0.28,
Rf trans ) 0.32, EtOAc/hexane, 30:70).
realization of 100% trans diastereoselectivity in this
Pictet-Spengler condensation. Analysis of this data
further supported the proposed mechanism of epimer-
ization [C(1)-N(2) cleavage] at C(1) of cis into trans-1,2,3-
trisubstituted tetrahydro-â-carbolines (see Schemes 4 and
7) and prompted additional investigation of the mecha-
nism of this isomerization in tetrahydro-â-carbolines
derived from nonactivated (desmethoxy) Nb-benzyltryp-
tophan derivatives.
As outlined at the top of Table 2, heating Nb-benzyl-
(D)tryptophan methyl ester 61 with R-ketoglutaric acid
gave a mixture of cis 62a and trans 62b diastereomers
in a ratio of 1:7. Conversion of the mixture of acids into
the corresponding esters was accomplished under neutral
conditions to provide the cis 63a and trans 63b diesters
in the Na-H, Nb-benzyl series. When the cis isomer 63a
was stirred at room temperature with 2.4 equiv of CF3-
COOH in CH2Cl2 for 10 days, the stereospecific formation
of the trans diastereomer 63b was realized (90% yield).
The optical rotation of trans 63b was identical to that of
pure 63b obtained earlier; epimerization had occurred
only at C(1). When the mixture of cis 63a and trans 63b
isomers was treated under the same conditions, again a
90% yield of the desired trans isomer 63b was realized.
The rate of the epimerization could be drastically in-
creased by heating the solution (see Experimental Section
for details). As illustrated in Table 2, the percent
When the above results in nonacidic aprotic media
were compared to the results from reaction of 6-meth-
oxytryptophan analog 47 with an excess (1.00-1.05
equiv) of R-ketoglutaric acid, it was felt that protonation
of the Nb-benzyl nitrogen group was essential for the

52 J . Org. Chem., Vol. 62, No. 1, 1997
Ta ble 2. Con ver sion of Cis Isom er 63a in to Tr a n s Isom er 63b
Cox et al.
start. mater., mg
temp/time, d
catalyst
solvent (mL)
product (%)
63b (90)
63b (85)
63b (60), 63a (20)
63b (0), 63a (85)
63b (80)
63a (cis), 20
63a (cis), 20
63a (cis), 20
63a (cis), 20
63b (trans), 20
63a +63b (1:1), 20
rt/10
rt/10
rt/10
rt/10
rt/10
rt/10
TFA
TFA
TFA
TFA
TFA
TFA
CH2Cl2 (1.0)
CHCl3 (1.0)
PhH (1.0)
THF (1.0)
CH2Cl2 (1.0)
CH2Cl2 (1.0)
63b (90)
Sch em e 11
conversion of 62a into 62b in the presence of TFA took
place more readily in the polar aprotic solvent CH2Cl2
than in benzene. Moreover, when this same transforma-
tion was attempted in THF, no epimerization was
observed.
absence of anhydrous ZnCl2 returned only cis isomer 64a
in quantitative yield. A similar experiment with the dry
pure hydrochloride salt of cis isomer 64a gave the salt
of the trans analog 64b again in support of intermediate
44 over 45 or 46 (Scheme 11).
Similar results were observed in CF3COOH/CH2Cl2 in
the Na-methyl, Nb-benzyl series. Pictet-Spengler reac-
tion of Na-methyl-Nb-benzyl-(L)tryptophan methyl ester
with methyl 3-formylpropionate in refluxing benzene
furnished a mixture of cis 64a and trans 64b diastere-
omers in a ratio of 28:72 in 90% yield as expected.9,10
When the cis isomer 64a was stirred in CH2Cl2/CF3COOD
for several days at room temperature, the trans isomer
64b was obtained in optically pure form (92% yield) with
no deuterium incorporation at C(1) or C(3). Again,
intermediate 45 was unequivocally ruled out; moreover,
the mixture [cis (64a ) 28: trans (64b) 72] was converted
into optically active trans 64b on heating for 5 h in CF3-
COOD/CH2Cl2.
In order to form intermediate 46 (Figure 1) in this
process, protonation of the indole double bond (at the
R-carbon atom) must precede iminium ion formation,31,32
consequently the cis isomer 64a was heated in dry
distilled chloroform (EtOH, HCl free) with anhydrous
ZnCl2 (4.7 equiv). This process provided the trans isomer
64b in 92% yield. It was felt the Lewis acid promoted
the formation of carbocationic intermediate 44 and not
iminium ion 46 since the solution was devoid of protons.
The control reaction with 64a in refluxing CHCl3 in the
Evidence above supported the C(1)-N(2) scission mech-
anism (see 44) illustrated in Figure 1 for this isomeriza-
tion prompting a similar set of experiments in a different
series (Scheme 12). When optically active Nb-benzyl-(D)-
tryptophan methyl ester 61 was heated with aldehyde
65 in refluxing benzene, a mixture of cis 66a and trans
66b isomers was formed in 90% yield in a ratio of 3:7,
as expected.9,10 When the mixture was stirred in CF3-
COOD, the trans isomer 66b was formed in high yield;
moreover, no deuterium incorporation was observed.
Epimerization of 66a had taken place at C(1) presumably
via intermediate 44 or 46. The cis isomer 66a was
allowed to stir in CF3COOH in the presence of NaBH434
which furnished a mixture of the reduced ring-cleaved
intermediate 67 [C(1)-N(2) cleavage] and the trans
isomer 66b in approximately equal amounts. The struc-
ture of 2-substituted indole 67 was deduced based on HR
NMR including 2D COSY and TOCSY experiments. No
cis diastereomer 66a was observed in this mixture nor
was any product isolated which would correspond to
either reduction of/or hydrolysis of an iminium ion related
to that in intermediate 46. Moreover, when the thermo-
(34) Gribble, G.; Nutaitis, C. Org. Prep. Proced. Int. 1985, 17, 317.

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 53
Sch em e 12
dynamically more stable trans isomer 66b was allowed
to stir with CF3COOH/NaBH4 under exactly the same
pendent upon the steric bulk of the incoming carbonyl
compound, the substituents on the Nb-nitrogen atom, and
the size of the ester function. As the size of the aldehyde
increased, so too did the ratio of trans diastereomer to
cis isomer. Under the nonacidic aprotic conditions of
benzene at reflux, aldehydes such as acetaldehyde un-
dergo the Pictet-Spengler condensation to provide only
modest diastereoselectivity (7a , 7b--26:74). Cyclohexane-
carboxaldehyde, however, reacts to provide a high yield
of solely the trans diastereomer 9 with 100% stereose-
lectivity. Increased steric interaction from the ester
function imparted a slight increase in the amount of trans
diastereomer which was formed. This was most evident
in the nonacidic aprotic condensations of butyraldehyde
with Nb-benzyltryptophan methyl ester (8a :8b--cis:trans-
-23:77) and Nb-benzyltryptophan isopropyl ester (11a :
11b--cis:trans--13:87). Replacement of the benzyl func-
tion by the larger diphenylmethyl group also increased
the trans stereoselectivity. When the Nb-diphenyl-
methyl substituent was introduced into the nonacidic
aprotic condensation it was evident that an appreciable
amount of steric hindrance was present since no reaction
was observed in the cases with bulky aldehydes. How-
ever, in those cases which did react, substantial increases
in the diastereomeric excess were observed.
34
conditions no product 67 of ring opening was observed.
The trans isomer 66b was recovered from this process
in greater than 90% yield. The failure of trans isomer
66b to provide 67 under these (control) conditions sug-
gests that the origin of 67 (from cis isomer 66a ) arises
from reduction of a carbocation “like” intermediate
related to 44 (or the corresponding iminium ion resonance
form similar to 49C in the desmethoxy series) and not
borane-mediated cleavage of the C(1)-N(2) bond. If the
latter event had occurred, then the isolation of 67 from
trans isomer 66b would have also been expected. The
isolation of thioacetal 67 from treatment of cis 66a with
CF3COOH/NaBH4 as well as conversion of 66a into 66b
in CF3COOD (with no deuterium incorporation) provide
the best evidence, to date, that epimerization of cis-1,3-
disubstituted-1,2,3,4-tetrahydro-â-carbolines into their
trans counterparts is mediated via scission across the
C(1)-N(2) bond (see 44)10 and not by intermediates 45
or 46 (Figure 1). In addition, conversion of cis 64a into
trans 64b mediated by ZnCl2 also supports the carboca-
tion C(1)-N(2) scission mechanism proposed here. The
lack of evidence to support the intermediacy of com-
pounds such as 45 or 46 provides additional support for
the C(1)-N(2) scission mechanism earlier proposed for
the isomerization of reserpine into isoreserpine reported
from our laboratory.32
The ratios observed under nonacidic aprotic conditions
are the result of kinetic trapping experiments. Upon
exposure to acid, the diastereomeric ratios shift to provide
increased amounts of the trans diastereomers. The
marked increases in the amount of the trans diastere-
omer formed in each case upon addition of acid and the
fact that the cis diastereomer 8a epimerizes irreversibly
to the trans congener 8b indicate that the trans diaste-
reomer in the Nb-alkyl cases is the thermodynamically
more stable diastereomer. Furthermore, there is strong
evidence to support the mechanism of C-1-N-2 bond
cleavage which results in the epimerization of cis 8a to
trans 8b. The isolation of 67 from the carbocation-
trapping experiment (NaBH4/CF3COOH) with 66a
strongly supports this hypothesis.
Con clu sion
Pictet-Spengler condensation of aldehydes with either
an Nb-benzyl or Nb-diphenylmethyl-substituted tryp-
tophan derivative resulted in the preferential formation
of 1,3-trans diastereomers. This is contrary to the
traditional condensations devoid of a substituent at the
Nb-nitrogen atom whereby cis diastereomers are often
times formed preferentially.1 The diastereochemical ratio
of the products obtained in the Pictet-Spengler cycliza-
tion of Nb-substituted-tryptophan alkyl esters was de-

54 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
Sch em e 13
extraction process; there is a definite risk of epimeriza-
tion at C(1) of the tetrahydro-â-carboline nucleus.
Exp er im en ta l Section
All starting materials were purchased from Aldrich Chemi-
cal Co. unless otherwise noted. All aldehydes were distilled
at reduced pressure under an atmosphere of nitrogen gas to
remove acidic impurities as evidenced by IR spectroscopy. Thin
layer chromatography was performed on Merck silica gel 60
F-254 TLC plates. Visualization of the indolic products was
achieved by spraying with an acidic cerric ammonium sulfate
solution and development in a stream of warm air for several
minutes. Unless otherwise noted, all spectral and analytical
data were recorded as described in reference 6. Optical
rotations were measured with a J ASCO DIP-370 polarimeter.
Spectral data for compounds 4,36 7,35 and 914 were identical to
those previously reported.
Gen er a l P r oced u r e for Ca ta lytic Tr a n sfer Hyd r oge-
n a tion (CTH).37 The Nb-alkyltetrahydro-â-carboline to be
hydrogenated was placed into a round-bottomed flask, and to
it was added solid, dry ammonium formate (20 mol equiv). This
mixture was dissolved in a minimal amount of absolute
methanol. Freshly activated 10%Pd/C (1 mass equiv relative
to the compound to be hydrogenated) was carefully slurried
in absolute methanol and the ammonium formate solution
added. This mixture was stirred under an atmosphere of dry
nitrogen gas until complete removal of the alkyl group was
indicated by TLC (silica gel; 1:9, ethyl acetate:benzene). The
solution was filtered through Celite and the Celite thoroughly
washed with ether. The organic solution which resulted was
washed with brine to remove the residual ammonium formate.
Solvent removal was achieved under reduced pressure to yield
the Nb-hydrogen bases in yields that ranged between 90% and
97%.
cis-3-(Meth oxyca r bon yl)-1-m eth yl-1,2,3,4-tetr a h yd r o-
9H-p yr id o[3,4-b]in d ole (4a ) a n d tr a n s-3-(Meth oxyca r bo-
n yl)-1-m e t h yl-1,2,3,4-t e t r a h yd r o-9H -p yr id o[3,4-b]in -
d ole (4b). Upon subjection of compounds 7a and 7b to CTH
(catalytic transfer hydrogenation), a 94% and 96% yield of 4a
and 4b, respectively, was realized, the 1H NMR spectra of each
being identical to those published previously.15,36 4a : 1H NMR
(CDCl3) δ 1.49 (d, J ) 6.6 Hz, 3H), 1.90 (br, 1H), 2.86 (m, 1H),
3.13 (m, 1H), 3.80 (s, 3H), 3.85 (dd, J ) 4.5 and 11.5 Hz, 1H),
4.28 (br, 1H), 7.12 (br, 2H), 7.33 (br, 1H), 7.48 (br, 1H), 7.83
(br, 1H). 4b: 1H NMR (CDCl3) δ 1.49 (d, J ) 6.6 Hz, 3H),
2.10 (br, 1H), 2.90 (m, 1H), 3.12 (m, 1H), 3.77 (s, 3H), 3.97
(dd, J ) 4.5 and 11.5 Hz, 1H), 4.41 (br, 1H), 7.10 (br, 2H),
7.35 (br, 1H), 7.50 (br, 1H), 7.86 (br, 1H).
Recently, Reddy et al. reported27,30 the synthesis of two
Roemaria alkaloids roeharmine and (-)-1,2,3,4-tetrahy-
droeharmine (68) isolated by Gozler et al. The 100%
trans stereoselectivity in the Pictet-Spengler reaction of
6-methoxytryptophan 47 with R-ketoglutaric acid pro-
vided much needed insight for the enantiospecific syn-
thesis of the 6,7-dimethoxy indole alkaloid 68 by Reddy.
The specific rotation for the natural product 68 had been
reported to be -4.0° (c ) 0.12, MeOH). Consideration of
the mechanism of epimerization described above in the
6-methoxy series (see Scheme 10) suggested that 68
might readily undergo C(1)-N(2) bond cleavage under
the acidic conditions of isolation by Gozler.27,30 Therefore,
Reddy designed the synthesis of THâC 68 utilizing the
Moody azide pyrolysis process and the Pictet-Spengler
condensation under nonacidic aprotic conditions.14,30 The
spectral properties of synthetic 68 were identical to the
natural product except for the optical rotation which was
-18.0° (c ) 1.04, MeOH). It was believed that the
natural 1,2,3,4-tetrahydroeharmine 68 had undergone
partial racemization during the acid/base-mediated isola-
tion procedure facilitated by the 7-methoxy group. This
hypothesis was tested by exposing optically pure 68 to
trifluoroacetic acid in CH2Cl2 at room temperature as
illustrated in Scheme 13. The 1H NMR and Rf of the
alkaloid which resulted were unchanged; however, the
optical rotation of the product was approximately -0.8°.
Racemization was believed to occur through the C(1)-
N(2) scission pathway related to 44 in which the 6-meth-
oxy group (7 in THâC nomenclature) was felt to contrib-
ute to stabilization of the ring-cleaved [C(1)-N(2)]
carbocation. When optically pure 68 was allowed to stir
with CF3COOD in CH2Cl2 at 25 °C (Scheme 13), the
optical rotation of the reaction product 68 was again
found to be near zero as expected. Although deuterium
incorporation had occurred at C(5) and C(8) no deuterium
(1R,3R)-(-)-2-Ben zyl-3-(m et h oxyca r b on yl)-1-p r op yl-
1,2,3,4-tetr ah ydr o-9H-pyr ido[3,4-b]in dole (8a) an d (1S,3R)-
(-)-2-Ben zyl-3-(m et h oxyca r b on yl)-1-p r op yl-1,2,3,4-t et -
r a h yd r o-9H-p yr id o[3,4-b]in d ole (8b). Optically active Nb-
benzyl-(D)tryptophan methyl ester was prepared according to
the method of Zhang18 and is known.38 A mixture of this ester
(6.0 g, 0.020 mol), dry benzene (100 mL), and butyraldehyde
(2.8 g, 0.039 mol) was brought to reflux and reacted under a
nitrogen atmosphere for 72 h. The solvent was removed
under reduced pressure and the cis and trans diastereomers
separated by gravity chromatography on silica gel (CH2Cl2)
to yield 8b (4.9 g) as a white solid and 8a (0.54 g) as a yellow
2
was found at C(1) as illustrated in Scheme 13 (see H-
oil for a combined yield of 77%. 8a : [R]25 ) -13.3° (c ) 1,
D
68). An alternate mechanism would result from forma-
tion of an olefinic intermediate related to 45 whereupon
protonation at C(1) would provide racemic material. If
this alternate pathway had operated during this epimer-
ization, the above reaction with CF3COOD would have
provided 68 in which a deuterium atom would appear at
C(1). This was not observed. Analysis of the deuterium
labeling experiment supports the cationic-mediated mech-
anism via scission across the C(1)-N(2) bond after which
protonation from both faces led to racemic material.27
These results should therefore be considered a warning
to those who isolate ring-A methoxylated indole alkaloids
when strongly acidic conditions are employed in the
1
benzene); IR (KBr) 3320-3180, 1718 cm-1; H NMR (C6D6) δ
0.76 (t, J ) 7.0 Hz, 3H), 1.38-1.60 (m, 4H), 2.94 (dd, J ) 15.9
and 5.8 Hz, 1H), 3.18 (dd, J ) 16.0 and 5.0 Hz, 1H), 3.41 (s,
3H), 3.74 (t, J ) 4.9 Hz, 1H), 3.92 (broad, 3H), 7.05-7.52
(broad, 9H), 7.65 (s, 1H); 13C NMR (C6D6) δ 14.3, 19.8, 19.9,
36.8, 51.3, 58.0, 58.9, 59.3, 106.3, 111.1, 118.6, 119.6, 121.9,
127.3, 127.6, 128.4, 128.9, 129.2, 134.5, 136.7, 139.5, 140.3,
(35) Ishida, A.; Nakamura, T.; Irie, K.; Ohishi, T. Chem. Pharm.
Bull. 1985, 33, 3237.
(36) Deng, L. M.S. Thesis, University of WisconsinsMilwaukee,
1991.
(37) Anwer, M. K.; Spatola, A. F. Synthesis 1980, 929.
(38) Shimizu, M.; Ishikawa, M.; Komoda, Y.; Nakajima, T.; Yamagu-
chi, K.; Sakai, S.-i. Chem. Pharm. Bull. 1984, 32, 1313.

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 55
174.7; MS (CI, CH4) m/z (relative intensity) 363 (M + 1, 100%).
Anal. Calcd for C23H26N2O2: C, 76.20; H, 7.23; N, 7.73.
Found: C, 76.55; H, 6.99; N, 8.01. 8b: mp 149.2-152.4 °C.
of the individual crops yielded 101.67 g (79.3%) of the pure
ester hydrobromide salt (20): mp 228.1 °C dec; IR (KBr) 3200,
1
1735 cm-1; H NMR (CDCl3, free base) δ 1.18 (d, J ) 6.4 Hz,
[R]25 ) -50.4° (c ) 1, benzene); IR (KBr) 3340-3260, 1700
3H), 1.20 (d, J ) 6.4 Hz, 3H), 1.61 (broad, 2H), 2.99 (dd, J )
14.4 and 7.8 Hz, 1H), 3.26 (dd, J ) 14.2 and 4.9 Hz, 1H), 3.76
(broad, 1H), 5.00 (heptet, J ) 6.3 Hz, 1H), 7.04 (s, 1H), 7.07-
7.21 (m, 2H), 7.34 (d, J ) 7.5 Hz, 1H), 7.36 (d, J ) 7.5 Hz,
1H), 8.18 (s, 1H); 13C NMR (DMSO-d6, HBr salt) δ 20.9, 21.3,
26.2, 52.7, 69.6, 106.4, 111.5, 118.0, 118.5, 121.1, 124.7, 126.7,
136.1, 166.8; MS (CI, CH4), m/z (relative intensity) 247 (M +
1, 33). HRMS calculated for C14H18N2O2: 246.1368. Found:
246.1396. Anal. Calcd for C14H18N2O2‚HBr: C, 51.39; H, 5.85;
N, 8.56. Found: C, 51.31; H, 5.82; N, 8.49.
D
1
cm-1; H NMR (C6D6) δ 0.75 (t, J ) 6.85 Hz, 3H), 1.30-1.47
(m, 4H), 2.96 (dd, J ) 6.8 and 17.5 Hz, 1H), 3.16 (dd, J ) 9.2
and 15.8 Hz, 1H), 3.33 (s, 3H), 3.52 (d, J ) 13.7 Hz, 1H), 3.70
(m, 1H), 3.95 (m, 2H), 6.45 (s, 1H), 7.16 (m, 6H), 7.45 (d, J )
7.2 Hz, 2H), 7.58 (d, J ) 6.1 Hz, 1H); 13C NMR (C6D6) δ 14.2,
19.1, 21.6, 36.6, 51.2, 53.7, 55.7, 57.0, 107.2, 111.0, 116.5, 119.7,
121.8, 127.3, 127.6, 128.0, 128.4, 129.5, 135.1, 140.5, 173.0;
MS (CI, CH4) m/z (relative intensity) 363 (M + 1, 100%). Anal.
Calcd for C23H26N2O2: C, 76.20; H, 7.23; N, 7.73. Found: C,
76.40; H, 7.25; N, 7.73.
Isop r op yl Nb-Ben zyl-(()-tr yp top h a n oa te (26). Benzal-
dehyde (11.38 g, 0.107 mol) was added to a solution of
methanol (140 mL) and tryptophan isopropyl ester 20 (24 g,
0.098 mol). This solution was allowed to stir for 11 h at 25 °C
after which it was cooled to 0 °C, and NaBH4 (0.9 g, 0.048 mol)
was added portionwise over 30 min. After 6 h, ice-water (2.1
mL) was added and the volume reduced under vacuum. The
residue was dissolved in CHCl3 (180 mL) and the solution
washed with brine (2 × 60 mL) and dried over Na2SO4.
Purification by flash chromatography (ethyl acetate:hexanes-
30:70 v/v on silica gel) yielded 20 g (61%) of pure Nb-
benzyltryptophan isopropyl ester 26: mp 55.5-56.8 °C; IR
(1R,3R)-(+)-3-(Met h oxyca r bon yl)-1-p r op yl-1,2,3,4-t et -
r a h yd r o-9H-p yr id o[3,4-b]in d ole. CTH of 8a provided the
optically active Nb-H isomer (R1 ) CH2CH2CH3, R3 ) CH3) in
91% yield, the spectral properties of which [13C NMR (CDCl3)
δ 56.40 (C-1), 52.45 (C-3)], were identical to the (1S,3S)
enantiomer.39,42 Anal. Calcd for C16H20N2O2: C, 70.56; H,
7.40; N, 10.29. Found: C, 70.73; H, 7.90; N, 9.98.
(1S,3R)-(+)-3-(Met h oxyca r b on yl)-1-p r op yl-1,2,3,4-t et -
r a h yd r o-9H-p yr id o[3,4-b]in d ole. CTH of 8b provided the
optically active Nb-H isomer (R1 ) CH2CH2CH3, R3 ) CH3) in
96% yield, the spectral properties of which [13C NMR (CDCl3)
δ 52.44 (C-1), 49.99 (C-3)], were identical to the (1R,3S)-
enantiomer.39,42 Anal. Calcd for C16H20N2O2: C, 70.56; H,
7.40; N, 10.29. Found: C, 70.40; H, 7.25; N, 10.53. When the
cis isomer 8a was stirred in CF3COOD it was converted into
the trans isomer 8b [R]25D ) -50.4° (c ) 1, benzene). The trans
isomer 8b did not contain deuterium (NMR, MS).
tr a n s-1-Cycloh exyl-3-(m et h oxyca r b on yl)-1,2,3,4-t et -
r a h yd r o-9H-p yr id o[3,4-b]in d ole (6b). CTH of 9 provided
a 95% yield of 6b; mp 147-148 °C the spectral data of which
were identical to those reported previously:11
Isop r op yl (()-tr yp top h a n oa te (20). Anhydrous HBr (g)
was passed through dry 2-propanol (1.1 L) at 0 °C to form a
saturated solution. In one portion, (D/L)-tryptophan (80.0 g,
0.392 mol) was added to the viscous solution. This stirred
suspension was warmed to reflux, and the resulting homoge-
neous solution took on a deep red coloration. After 7 h the
flask was cooled and the precipitate isolated by vacuum
filtration. Additional material was obtained from the mother
liquors and recrystallization from 2-propanol. Combination
1
(KBr) 3720, 3070, 2950, 1700 cm-1; H NMR (CDCl3) δ 1.04
(d, J ) 6.3 Hz, 3H), 1.14 (d, J ) 6.3 Hz, 3H), 1.81 (broad, 1H),
3.13 (d, J ) 6.9 Hz, 2H), 3.60 (t, J ) 6.7 Hz, 1H), 3.64 (d, J )
6.7 Hz, 1H), 3.81 (d, J ) 13.1 Hz, 1H), 4.95 (heptet, J ) 6.3
Hz, 1H), 7.02-7.35 (m, 8H), 7.59 (d, J ) 7.9 Hz, 1H), 7.99 (s,
1H); 13C NMR (CDCl3) δ 21.5, 21.8, 29.3, 52.0, 61.4, 66.1, 111.0,
111.2, 118.8, 119.2, 121.9, 122.8, 126.9, 127.5, 128.1, 128.3,
136.1, 139.7, 174.5; MS (CI, CH4) m/z (relative intensity) 337
(M + 1, 100%). HRMS calculated for C21H24N2O2: 336.1838.
Found: 336.1809. Anal. Calcd for C21H24N2O2‚HCl: C, 67.65;
H, 6.71; N, 7.52. Found: C, 67.07; H, 6.89; N, 7.27.
cis-2-Ben zyl-3-(isop r op oxyca r bon yl)-1-m eth yl-1,2,3,4-
t et r a h yd r o-9H-p yr id o[3,4-b]in d ole (10a ) a n d tr a n s-2-
Ben zyl-3-(isop r op oxyca r bon yl)-1-m eth yl-1,2,3,4-tetr a h y-
d r o-9H-p yr id o[3,4-b]in d ole (10b). Acetaldehyde (0.7 g, 14.6
mmol), Nb-benzyltryptophan isopropyl ester 26 (0.464 g, 1.38
mmol), and dry benzene (10 mL) were sealed under vacuum
in a glass tube and heated for 36 h in an oven at 80 °C. The
reaction mixture was cooled and transferred to a round-
bottomed flask where the solvent was removed under reduced
pressure. A short silica gel wash column provided 0.430 g
(86%) of a mixture of the trans and cis isomers. Further
chromatography (CHCl3, silica gel) was used to separate the
isomers. Anal. Calcd (for the mixture of diastereomers)
C23H26N2O2: C, 76.21; H, 7.23; N, 7.73. Found: C, 76.45; H,
7.25; N, 7.73. 10a : 1H NMR (CDCl3) δ 1.20 (d, J ) 6.3 Hz,
3H), 1.23 (d, J ) 6.3 Hz, 3H), 1.33 (d, J ) 6.9 Hz, 3H), 2.97
(dd, J ) 7.1 and 5.5 Hz, 1H), 3.20 (dd, J ) 15.7 and 8.1 Hz,
1H), 3.86 (t, J ) 6.2 Hz, 1H), 3.95 (d, J ) 15.1 Hz, 1H), 4.03
(d, J ) 15.0 Hz, 1H), 4.27 (q, J ) 6.9 Hz, 1H), 4.98 (heptet, J
) 6.3 Hz, 1H), 7.00-7.55 (m, 9H), 7.65 (s, 1H); 13C NMR
(CDCl3) δ 18.4, 21.8, 22.4, 53.2, 54.4, 60.6, 68.3, 106.7, 110.7,
118.2, 119.5, 121.5, 126.8, 128.2, 128.3, 135.4, 135.9, 140.3,
173.1; MS (CI, CH4) m/z (relative intensity) 363 (M + 1, 100)
419 (M + 28, 18.7). HRMS calculated for C23H26N2O2: 362.1994.
Found: 362.1959. CTH of 10a provided a compound (Nb-H)
which had spectral properties identical to 29a .42 10b: 1H
NMR (CDCl3) δ 1.18 (d, J ) 6.2 Hz, 3H), 1.19 (d, J ) 6.2 Hz,
3H), 1.42 (d, J ) 6.6 Hz, 3H), 2.98 (dd, J ) 15.7 and 5.4 Hz,
1H), 3.09 (dd, J ) 15.7 and 7.1 Hz, 1H), 3.77 (d, J ) 14.4 Hz,
1H), 3.88 (d, J ) 14.3 Hz, 1H), 3.92 (dd, J ) 7.1 and 5.5 Hz,
1H), 4.14 (q, J ) 6.6 Hz, 1H), 5.01 (heptet, J ) 6.2 Hz, 1H),
7.00-7.55 (multiplet, 9H), 7.62 (s, 1H); 13C NMR (CDCl3) δ
21.3, 21.5, 22.3, 51.0, 53.8, 57.0, 68.0, 106.4, 110.7, 118.1, 119.4,
121.4, 126.9, 128.3, 128.4, 136.1, 136.2, 140.1, 172.7; MS (CI,
CH4) m/z (relative intensity) 363 (M + 1, 100). HRMS
calculated for C23H26N2O2: 362.1994. Found: 362.1965. CTH
of 10b provided a compound (Nb-H) which had spectral
properties identical to 29b.42 Anal. Calcd (as a mixture of
diastereomers) for C23H26N2O2‚HCl‚3/4H2O: C, 66.99; H, 6.73;
N, 6.80. Found: C, 67.26; H, 6.66; N, 6.45.
(39) Bailey, P. D.; Hollinshead, S. P.; McLay, N. R.; Morgan, K.;
Palmer, S. J .; Prince, S. N.; Reynolds, C. D.; Wood, S. D. J . Chem.
Soc., Perkin Trans. 1 1993, 4, 431.
(40) The author has deposited atomic coordinates for this structure
with the Cambridge Crystallographic Data Centre. The coordinates
can be obtained, on request, from the Director, Cambridge Crystal-
lographic Data Centre, 12 Union Road, Cambridge, CB2 1EZ, UK.
(41) Not only does the Nb-benzyl group promote trans stereoselec-
tivity in the Pictet-Spengler reaction itself (see results in PhH, ∆)4
but also promotes trans stereochemistry during the acid-mediated
equilibration at C(1). In the racemic series Ungemach demonstrated
that stereospecific formation of ii (from i) occurred as reported in the
Nb-benzyl series under acidic or nonacidic conditions. However after
the trans diastereomer ii was converted into the corresponding Nb-H
trans isomer iii via catalytic debenzylation, this pure trans isomer iii
(Nb-H) was transformed into an approximate 40:60 mixture of cis to
trans diastereomers (iv and v) on heating under acidic conditions. In
the absence of the Nb-benzyl group a considerable amount of cis-isomer
iv formed under these acidic conditions. Ungemach, F. MS Thesis,
University of Wisconsin-Milwaukee, 1978.
(42) K. Czerwinski, Ph.D. Thesis, University of Wisconsins
Milwaukee, Milwaukee, WI, 1995.

56 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
cis-2-Ben zyl-3-(isop r op oxyca r bon yl)-1-p r op yl-1,2,3,4-
t et r a h yd r o-9H-p yr id o[3,4-b]in d ole (11a ) a n d tr a n s-2-
Ben zyl-3-(isop r op oxyca r bon yl)-1-p r op yl-1,2,3,4-tetr a h y-
d r o-9H-p yr id o[3,4-b]in d ole (11b). Butanal (0.432 g, 6.0
mmol) was added to a solution of Nb-benzyl tryptophan
isopropyl ester 26 (1.0 g, 3.0 mmol) in dry benzene (50 mL).
This solution was heated to reflux, and after 36 h the reaction
was cooled to 25 °C, the solvent was removed under vacuum
and a small amount of the crude material removed for
spectroscopy. The mixture of cis and trans diastereomers
(0.927 g, 82%) was separated by flash chromatography (CH2-
Cl2, silica gel). Anal. Calcd (for the mixture of diastereomers)
C25H30N2O2‚1/4H2O: C, 76.04; H, 7.73; N, 7.09. Found: C,
75.88; H, 7.76; N, 7.10. 11a : 1H NMR (CDCl3) δ 0.76 (t, J )
7.2 Hz, 3H), 1.20 (d, J ) 6.3 Hz, 3H), 1.25 (d, J ) 6.5 Hz, 3H),
1.40-1.80 (m, 4H), 2.94 (dd, J ) 15.7 and 5.8 Hz, 1H), 3.18
(dd, J ) 15.7 and 5.0 Hz, 1H), 3.78 (t, J ) 5.0 Hz, 1H), 3.92
(broad, 3H), 4.94 (heptet, J ) 6.3 Hz, 1H), 7.05-7.52 (broad,
9H), 7.65 (s, 1H); 13C NMR (CDCl3) δ 14.0, 19.8, 20.7, 21.7,
21.8, 36.1, 58.1, 58.2, 59.7, 66.3, 106.7, 110.6, 118.1, 119.4,
121.5, 124.0, 127.1, 128.1, 128.7, 134.8, 136.0, 139.8, 173.2;
MS (CI, CH4) m/z (relative intensity) 391 (M + 1, 100%) 419
(M + 28, 18.7%). HRMS calculated for C25H30N2O2: 390.2307.
Found: 390.2294. CTH of cis 11a provided a compound (Nb-
H) which had spectral properties identical to 30a .42 11b: mp
3.30 (dd, J ) 6.0 and 7.3 Hz, 1H), 4.14 (t, J ) 6.8 Hz, 1H),
4.83 (m, J ) 6.3 Hz, 1H), 6.73-7.72 (m, 15H), 10.76 (s, 1H);
13C NMR (DMSO-d6) δ 21.4, 29.0, 66.5, 67.5, 109.8, 111.1,
118.0, 120.7, 123.6, 127.3, 127.9, 130.3, 135.5, 135.9, 138.9,
169.1, 170.7; MS (CI, CH4) m/z (relative intensity) 411 (M +
1, 100%). Anal. Calcd for C27H26N2O2: C, 79.03; H, 6.38; N,
6.82. Found: C, 78.83; H, 6.31; N, 6.63.
Isop r op yl Nb -(Dip h en ylm et h yl)-(()-t r yp t op h a n oa t e
(28). A solution of anhydrous methanol (500 mL) and imine
24 (11.80 g, 0.029 mol) was stirred at pH 6 (adjusted by the
addition of glacial acetic acid). Sodium cyanoborohydride (3.62
g, 0.058 mol) was added portionwise while the reaction was
monitored by TLC. Additional glacial acetic acid was added
to maintain the pH at 6. The observation that the starting
material was no longer visible by TLC (silica gel; 1:9 ethyl
acetate:benzene v/v) prompted the addition of excess distilled
water until the solution took on a white color. The pH was
adjusted to 8 with NaHCO3 and the solution extracted with
CHCl3. Removal of the solvent under reduced pressure
furnished a tan oil. The oil was taken up in ether and the
hydrochloride salt formed by the addition of 1.5 mol equiv of
anhydrous HCl (g) dissolved in ether. The stoppered solution
was held overnight at 0 °C, and the resultant white precipitate
was collected by vacuum filtration and thoroughly washed with
cold ether to yield, upon drying, 12.33 g (95%) of 28 as a white
powder: mp 212.6 °C dec; IR (KBr) 3280, 1740 cm-1; 1H NMR
(500 MHz, C6D6) δ 0.82 (d, J ) 6.2 Hz, 3H), 0.92 (d, J ) 6.3
Hz, 3H), 3.20 (m, 2H), 3.76 (t, J ) 6.6 Hz, 1H), 4.96 (m, J )
6.3 Hz, 1H), 5.05 (s, 1H), 6.52 (d, J ) 2.3 Hz, 1H), 6.68 (broad,
1H), 6.98-7.19 (m, 10H), 7.36 (d, J ) 7.0 Hz, 2H), 7.44 (d, J
) 7.1 Hz, 2H), 7.67 (d, J ) 8.0 Hz, 1H); 13C NMR (DMSO-d6)
δ 20.5, 21.2, 59.3, 64.4, 69.3, 108.1, 111.4, 118.0, 118.4, 121.1,
124.2, 128.2, 128.8, 136.2, 168.9; MS (CI, CH4) m/z (relative
intensity) 413 (M + 1, 100%). Anal. Calcd for C27H28N2O2‚
HCl: C, 72.23; H, 6.51; N, 6.24. Found: C, 72.54; H, 6.63; N,
6.17.
1
137.2-138.1 °C; H NMR (CDCl3) δ 0.73 (t, J ) 7.4 Hz, 3H),
1.19-1.80 (m, 10H), 2.96 (dd, J ) 16.2 and 5.3 Hz, 1H), 3.10
(dd, J ) 15.8 and 9.0 Hz, 1H), 3.56 (d, J ) 13.8 Hz, 1H), 3.81-
3.86 (m, 2H), 3.96 (dd, J ) 9.0 and 5.3 Hz, 1H), 5.06 (heptet,
J ) 6.2 Hz, 1H), 7.08-7.40 (broad, 8H), 7.51 (d, J ) 7.0 Hz,
1H), 7.62 (s, 1H); 13C NMR (CDCl3) δ 14.0, 18.8, 21.1, 21.9,
36.8, 53.2, 55.4, 56.7, 68.1, 107.4, 110.7, 118.1, 119.4, 121.4,
126.9, 127.1, 128.1, 129.0, 135.4, 136.0, 139.8, 172.8; MS (CI,
CH4) m/z (relative intensity) 391 (M + 1, 100%) 419 (M + 28,
19.2%). Anal. Calcd for C25H30N2O2‚1/4H2O: C, 76.04; H, 7.73;
N, 7.09. Found: C, 75.88; H, 7.76; N, 7.10. CTH of trans 11b
provided a compound (Nb-H) which had spectral properties
identical to 30b.42
Meth yl Nb-(Dip h en ylm eth yl)-(()-tr yp top h a n oa te (25).
Tryptophan methyl ester hydrobromide salt (11.318 g) was
converted into the imine 21 by the method of Polt and
O’Donnell.22 The product of this preparation was then im-
mediately dissolved in anhydrous methanol and the pH of the
solution adjusted to 6 with glacial acetic acid. Addition of
NaCNBH3 (7.50 g) over a 1 h period and further stirring while
the reaction was monitored by TLC (silica gel; 1:9 ethyl
acetate:benzene v/v) indicated reaction completion in 4 h.
Concomitant additions of acetic acid over the course of the
reduction were necessary to maintain the pH at 6. Distilled
water was added until a cloudy white precipitate persisted.
The pH was adjusted to 8 with concentrated aqueous ammonia
and the mixture extracted with CHCl3. The combined organic
extracts were dried over K2CO3. The filtrate was concentrated
under reduced pressure and dissolved in a minimal amount
of anhydrous ether. Addition of an ether solution of 1.5 mol
equiv of HCl and storage at -20 °C overnight yielded a fluffy
white powder which was filtered off and washed with cold
ether. The white solid was dried to provide 21.4 g (80% from
the tryptophan methyl ester hydrobromide salt) of the sub-
tr a n s-2-Ben zyl-1-cycloh exyl-3-(isop r op oxyca r b on yl)-
1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (12). Cyclohex-
anecarboxaldehyde (0.51 g, 4.5 mmol) was added to a solution
of Nb-benzyltryptophan isopropyl ester 26 (1.0 g, 3.0 mmol) in
dry benzene (100 mL). This solution was heated to reflux
using a Dean-Stark trap for water removal. After 36 h the
reaction was cooled, and a small amount of the crude material
was removed for spectroscopy. Removal of the solvent under
reduced pressure and recrystallization from methanol afforded
1.16 g (89%) of the pure material with mp 185.0-187.3 °C:
IR (KBr) 3450, 1730, 1400 cm-1; 1H NMR (DMSO-d6) δ 0.60-
1.30 (m, 5H), 1.28 (d, J ) 6.2 Hz, 3H), 1.30 (d, J ) 6.2 Hz,
3H), 1.40-1.90 (broad, 5H), 2.20-2.30 (broad, 5H), 2.85 (dd,
J ) 14.2 and 5.5 Hz, 1H), 2.96 (dd, J ) 14.2 and 10.0 Hz, 1H),
3.20 (d, J ) 8.5 Hz, 1H), 3.29 (d, J ) 13.8 Hz, 1H), 3.66 (d, J
) 13.8 Hz, 1H), 4.09 (q, J ) 5.0 Hz, 1H), 4.98 (heptet, J ) 6.0
Hz, 1H), 6.95-7.45 (broad, 9H), 10.59 (s, 1H); 13C NMR (CDCl3)
δ 20.7, 21.9, 26.3, 30.4, 42.0, 53.1, 57.0, 61.4, 68.1, 107.6, 110.6,
118.1, 119.3, 121.5, 126.9, 127.9, 129.1, 134.4, 135.8, 139.8,
172.7; MS (CI, CH4) m/z (relative intensity) 431 (M + 1, 100%).
HRMS calculated for C28H34N2O2: 430.2620. Found: 430.2635.
Anal. Calcd for C28H34N2O2: C, 78.10; H, 7.96; N, 6.51.
Found: C, 78.27; H, 8.23; N, 6.57. The stereochemical
configuration of 12 was confirmed as trans by single crystal
X-ray crystallography.40
stituted amine 25: mp 254 °C dec; IR (neat) 3428, 1728 cm-1
;
1H NMR (CDCl3, free base) δ 2.23 (s, 1H), 3.16 (d, J ) 6.4 Hz,
2H), 3.58 (s, 3H), 4.80 (s, 1H), 6.99-7.35 (m, 15H), 7.52 (d, J
) 7.8 Hz, 1H), 7.96 (s, 1H); 13C NMR (DMSO-d6) δ 29.5, 52.6,
66.2, 70.6, 110.1, 112.8, 118.0, 118.7, 121.3, 123.3, 127.5-128.8
(m), 130.3, 135.9, 136.0, 168.9; MS (CI, CH4) m/z (relative
intensity) 385 (M + 1, 45.7%). Anal. Calcd for C25H24N2O2‚
HCl: C, 71.33; H, 5.99; N, 6.65. Found: C, 71.01; H, 6.34; N,
6.22.
Isopr opyl Nb-(Diph en ylm eth ylen e)-(()-tr yptoph an oate
(24). Into a 1 L flask were placed a suspension of tryptophan
isopropyl ester hydrobromide salt (14.08 g, 0.043 mol) in CH2-
Cl2 (500 mL) and an equimolar amount of benzophenone imine
(7.80 g). This mixture was stirred for 24 h under nitrogen.
Filtration to remove NH4Br and solvent removal under
reduced pressure yielded a tan oil. The oil was taken up in
ether and the solution filtered, washed with distilled water,
and dried over K2CO3. Filtration and solvent removal followed
by recrystallization from ether/hexane yielded 15.9 g (90%) of
24 as white solid: mp 118.9-120.8 °C; IR (KBr) 3455, 3190,
tr a n s-3-(Meth oxyca r bon yl)-1-m eth yl-2-(d ip h en ylm eth -
yl)-1,2,3,4-tetr a h yd r o-9H-p yr id p [3,4-b]in d ole (13b). A so-
lution of 23 (76 mg, 0.2 mmol) and dry benzene (7 mL) was
placed in
a glass tube and cooled to 0 °C after which
acetaldehyde (0.166 mL, 3.0 mmol) was added to it. (The
volume of the solution was not allowed to exceed one-tenth
the total volume of the glass tube.) The tube was sealed at
atmospheric pressure and carefully thawed. After heating at
80 °C for 36 h, the tube was cooled to 25 °C prior to opening,
and the solvent was removed under reduced pressure to
1725 cm-1
;
1H NMR (DMSO-d6) δ 1.04 (d, J ) 6.3 Hz, 3H),
1.19 (d, J ) 6.3 Hz, 3H), 3.02 (dd, J ) 6.5 and 7.3 Hz, 1H),

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 57
provide a yellow oil. The residue was subjected to a wash
column (silica gel, CH2Cl2) followed by separation of the
diastereomers by flash chromatography (silica gel, toluene) to
yield 13b (28 mg, 34%) as a pale yellow oil: IR (neat) 3520,
NMR (C6D6) δ 19.5, 20.8, 21.6, 21.7, 49.4, 54.9, 67.8, 72.0,
106.1, 110.6, 118.7, 119.7, 121.6, 127.5, 128.1, 128.5, 128.9,
129.0, 130.2, 132.0, 135.4, 136.7, 143.7, 144.5, 173.7; IR (neat)
1400, 1710, 2875, 3405; MS (CI, CH4) m/z (relative intensity)
439 (M + 1, 97%). Anal. Calcd for C29H30N2O2‚HCl: C, 73.33;
H, 6.58; N, 5.90. Found: C, 72.89; H, 6.84; N, 5.99. CTH of
16 provided a compound (Nb-H) which had spectral properties
identical to trans 29b [R1 ) CH3, R3 ) CH(CH3)2].42
1
1750 cm-1; H NMR (C6D6, free base) δ 1.22 (d, J ) 7.0 Hz,
3H), 2.82 (q, J ) 7.0 Hz, 1H), 3.24 (s, 3H), 3.27 (d, J ) 14.2
Hz, 1H), 4.07 (q, J ) 6.0 Hz, 2H), 5.23 (s, 1H), 6.17 (s, 1H),
7.05 (m, 8H), 7.56 (m, 6H); 13C NMR (C6D6) δ 19.4, 20.4, 21.4,
49.4, 51.1, 54.4, 72.1, 106.0, 111.0, 118.7, 119.7, 121.7, 127.4,
127.5, 127.8, 128.2, 128.3, 128.5, 129.0, 129.0, 129.3, 135.2,
136.7, 143.5, 144.6, 174.5; MS (CI, CH4) m/z (relative intensity)
411 (M + 1, 56.8%). Anal. Calcd for C27H26N2O2‚HCl: C,
72.55; H, 6.09; N, 6.27. Found: C, 72.59; H, 6.02; N, 6.74.
CTH of 13 provided a compound which had spectral properties
identical to trans-4b.
tr a n s-3-(Meth oxyca r bon yl)-2-(d ip h en ylm eth yl)-1-p r o-
p yl-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (14). A so-
lution of 23 (1.17 g, 3.05 mmol), butyraldehyde (0.242 g, 3.36
mmol), and dry benzene (50 mL) was heated to reflux under a
nitrogen atmosphere for 36 h. Analysis by TLC indicated very
little product formation. A further period of heating (5 d)
provided tetrahydro-â-carboline for study. Removal of the
solvent under reduced pressure and purification by flash
chromatography yielded 0.720 g (59%) of a waxy foam 14: IR
(neat) 3430, 1736 cm-1; 1H NMR (C6D6, free base) δ 0.87 (t, J
) 6.4 Hz, 3H), 1.16 (m, 1H), 1.33 (m, 1H), 1.60 (m, 1H), 1.88
(m, 1H), 2.10 (s, 1H), 2.85 (dd, J ) 5.9 and 8.1 Hz, 1H), 3.28
(d, J ) 13.4 Hz, 1H), 3.36 (s, 3H), 3.74 (dd, J ) 5.4 and 4.3
Hz, 1H), 4.11 (d, J ) 6.8 Hz, 1H), 4.80 (s, 1H), 6.18 (s, 1H),
6.97-7.27 (m, 10H), 7.38 (d, J ) 7.1 Hz, 1H), 7.63 (t, J ) 7.0
Hz, 2H); 13C NMR (C6D6) δ 15.1, 18.0, 21.4, 21.9, 38.8, 51.9,
54.7, 55.1, 73.4, 105.9, 111.4, 119.3, 120.2, 122.3, 126.1, 128.8,
134.6, 137.1, 144.4, 144.7, 174.7; MS (CI, CH4) m/z (relative
intensity) 439 (M + 1, 61%). Anal. Calcd for C29H30N2O2‚
HCl: C, 73.33; H, 6.58; N, 5.90. Found: C, 73.01; H, 6.66; N,
5.96. CTH of 14 provided a compound which had spectral
properties identical to the trans Nb-H isomer related to 5b with
(R1 ) CH2CH2CH3, R3 ) CH3).42
tr a n s-1-Cycloh exyl-3-(m eth oxyca r bon yl)-2-(d ip h en yl-
m eth yl)-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (15). A
solution of 23 (1.01 g, 2.63 mmol), anhydrous benzene (50 mL),
and cyclohexanecarboxaldehyde (0.324 g, 2.88 mmol) was
placed into a round-bottomed flask. To this solution TFA
(0.329 g, 2.89 mmol) was added. The solution was heated to
reflux and after 36 h the reaction mixture was cooled, washed
with 15% aqueous ammonia, brine, and dried over K2CO3.
Filtration and solvent removal under reduced pressure yielded
a yellow oil. Flash chromatography on silica gel (toluene)
resulted in 0.391 g of 15 (31%): IR (neat) 3444, 1735 cm-1; 1H
NMR (C6D6, free base) δ 1.39 (m, broad 7H), 2.04 (d, J ) 5.8
Hz, 4H), 2.51 (s, 1H), 2.29 (m, 2H), 2.32 (s, 3H), 3.66 (t, J )
6.2 Hz, 1H), 5.08 (s, 1H), 6.31 (s, 1H), 6.92 (s, 1H), 7.19 (m,
7H), 7.44 (m, 4H), 7.73 (d, J ) 7.5 Hz, 1H); 13C NMR (C6D6) δ
21.5, 26.7, 27.6, 28.6, 28.7, 30.0, 33.5, 51.2, 60.2, 66.1, 110.6,
111.7, 117.3, 119.5, 120.0, 122.4, 125.7, 127.3, 127.7, 128.4,
128.5, 128.6, 128.7, 128.8, 129.4, 137.5, 139.4, 143.6, 145.0,
175.3; MS (CI, CH4) m/z (relative intensity) 479 (M + 1, 45%).
Anal. Calcd for C32H34N2O2‚HCl: C, 74.62; H, 6.85; N, 5.44.
Found: C, 74.40; H, 7.23; N, 6.01. CTH of 15 provided a
compound which had spectral properties identical to trans 6b
(R1 ) C6H11, R3 ) CH3).42
tr a n s-2-(Dip h en ylm et h yl)-3-(isop r op oxyca r b on yl)-1-
p r op yl-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (17). A
solution of the hydrochloride salt of 24 (1.29 g, 2.9 mmol),
butyraldehyde (1.03 g, 14.3 mmol), TFA (0.65 g, 5.7 mmol),
and dry benzene (75 mL) was heated to reflux. After 36 h the
solution was cooled and the solvent removed under reduced
pressure. The residue was dissolved in CH2Cl2, washed with
14% ammonia (aq) and brine, and dried (K2CO3). Flash
chromatography on silica gel (CH2Cl2) yielded 17 (0.27 g, 20%
1
yield): IR (neat) 3450, 3110, 1772 cm-1; H NMR (C6D6, free
base) δ 0.90 (m, 6H), 1.04 (d, J ) 6.1 Hz, 3H), 1.18 (m, 1H),
1.30 (m, 1H), 1.58 (m, 1H), 1.91 (m, 1H), 2.20 (s, 1H), 2.86 (d,
J ) 6.0 and 8.2 Hz, 1H), 3.30 (d, J ) 14.0 Hz, 1H), 3.81 (dd,
J ) 5.2 and 4.2 Hz, 1H), 4.16 (d, J ) 7.1 Hz, 1H), 4.86 (s, 1H),
4.98 (m, 1H), 6.22 (s, 1H), 6.90-7.24 (m, 10H), 7.38 (d, J ) 7.4
Hz, 1H), 7.59 (t, J ) 7.2 Hz, 2H); 13C NMR (C6D6) δ 15.6, 18.1,
21.3, 21.7, 21.9, 29.8, 39.0, 51.7, 54.8, 56.1, 72.9, 106.1, 112.1,
120.3, 120.7, 122.2, 127.0, 128.7, 134.1, 136.9, 144.9, 145.2,
175.1; MS (CI, CH4) m/z (relative intensity) 467 (M + 1, 100%).
Anal. Calcd for C31H34N2O2‚HCl: C, 74.00; H, 6.81; N, 5.57.
Found: C, 73.60; H, 6.67; N, 5.92. CTH of 17 provided a
compound (Nb-H) which had spectral properties identical to
trans 30b [R1 ) CH2CH2CH3, R3 ) CH(CH3)2].42
tr a n s-1-Cycloh exyl-2-(d ip h en ylm eth yl)-3-(isop r op oxy-
ca r bon yl)-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (18).
A mixture of the hydrochloride salt of 24 (1.50 g), cyclohex-
anecarboxaldehyde (0.56 g), dry benzene (70 mL), and TFA
(0.38 g) was heated to reflux for 36 h. After cooling, the solvent
was removed under vacuum and the tan residue taken up in
CH2Cl2. This solution was washed with NaHCO3 (saturated)
and brine and then dried (MgSO4). Filtration and removal of
the solvent under reduced pressure gave a tan oil. Flash
chromatography on silica gel (CH2Cl2) yielded a yellow oil
(0.423 g, 25%): IR (neat) 3417, 3123, 1770 cm-1 1H NMR
;
(C6D6, free base) δ 0.95 (d, J ) 6.2 Hz, 3H), 1.03 (d, J ) 6.2
Hz, 3H), 1.11-1.82 (m, 7H), 2.07 (dd, J ) 4.9 and 6.0 Hz, 4H),
2.51 (s, 1H), 3.29 (t, J ) 6.7 Hz, 2H), 3.82 (t, J ) 6.4 Hz, 1H),
5.01 (m, J ) 6.2 Hz, 1H), 5.09 (s, 1H), 6.33 (s, 1H), 6.92 (s,
1H), 7.02-7.61 (m, 11H), 7.76 (d, J ) 6.9 Hz, 2H); 13C NMR
(C6D6) δ 21.6, 21.8, 27.6, 28.5, 28.6, 30.0, 33.5, 53.4, 60.4, 66.0,
67.9, 110.5, 111.9, 117.3, 119.6, 119.9, 122.3, 127.3, 127.8,
127.9, 128.6, 128.7, 128.9, 129.4, 130.0, 130.2, 132.1, 137.4,
139.3, 143.5, 145.1, 174.4; MS (CI, CH4) m/z (relative intensity)
507 (M + 1, 100%). Anal. Calcd for C34H38N2O2‚HCl: C, 75.19;
H, 7.24; N, 5.16. Found: C, 74.75; H, 7.00; N, 5.19. Spectral
data obtained after CTH of 18 were identical to those of trans
31 [R1 ) C6H11, R3 ) CH(CH3)2].42
(+)-6-Meth oxy-Na-m eth yl-Nb-ben zyl-(D)tr yptoph an Eth -
yl Ester (47). To a solution of (-)-Na-methyl-6-methoxy-(D)-
tryptophan ethyl ester29,33 (4.1 g, 15 mmol) in anhydrous
methanol (25 mL) was added freshly distilled benzaldehyde
(1.59 g, 15 mmol). The solution which resulted was stirred
for 2.5 h at room temperature. The mixture was then cooled
in a large ice-salt bath to -5 °C (reaction mixture tempera-
ture), and sodium borohydride (0.72 g, 18.9 mmol) was added
in 2 portions over a period of 45 min without allowing the
temperature to rise above 0 °C. The reaction was monitored
by TLC on silica gel with MeOH/EtOAc (10:90) as the develop-
ing solvent for the formation of the imine (Rf ) 0.6) and EtOAc/
hexane (50:50) for the formation of the amine 47 (Rf ) 0.3).
The solution was allowed to stir for an additional 0.5 h followed
by the addition of ice-water (1 mL). The solvent was removed
under reduced pressure. The residue was dissolved in CH2-
Cl2 and extracted with water (2 × 20 mL). The organic layer
was washed with brine (50 mL) and dried (MgSO4). The
solvent was removed in vacuo to provide the crude Nb-
benzyltryptophan 47 which was purified by flash chromatog-
raphy on silica gel with EtOAc/hexane (30:70) to yield the free
tr a n s-2-(Dip h en ylm et h yl)-3-(isop r op oxyca r b on yl)-1-
m eth yl-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (16). A
solution of 24 (0.50 g, 1.3 mmol) and dry benzene (7 mL) was
placed in a tube for sealing. After cooling the solution to 0
°C, acetaldehyde (0.20 mL, 3.6 mmol, also at 0 °C) was added.
The solution was cooled further in liquid nitrogen, sealed under
vacuum, and carefully thawed. The solution was heated for
36 h at 80 °C followed by removal of the solvent under reduced
pressure to give a brown oil. Flash chromatography on silica
gel (CH2Cl2) yielded 16 (0.165 g, 31%) as a white foam: mp
1
235 °C dec; H NMR (C6D6, free base) δ 0.91 (d, J ) 6.3 Hz,
3H), 1.06 (d, J ) 6.3 Hz, 3H), 1.29 (d, J ) 7.0 Hz, 3H), 2.85 (q,
J ) 7.0 Hz, 1H), 3.33 (d, J ) 15.6 Hz, 1H), 4.09 (q, J ) 6.3 Hz,
2H), 4.87 (m, J ) 6.2 Hz, 1H), 5.33 (s, 1H), 6.18 (s, 1H), 7.10
(m, 8H), 7.61 (d, J ) 8.9 Hz, 5H), 7.69 (d, J ) 7.1 Hz, 1H); 13
C

58 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
base 47 as an oil (5.1 g, 93%); [R]25D ) +4.75° (c ) 1.6, CHCl3);
IR (neat) 2936, 2835, 1730, 1035 cm-1; 1H NMR (CDCl3) δ 1.14
(t, J ) 7.1 Hz, 3H), 1.77 (br, s, 1H), 3.08 (dd, J ) 6.9 and 2.7
Hz, 2H), 3.61 (m, 3H), 3.65 (s, 3H), 3.80 (d, J ) 13.2 Hz, 1H),
3.84 (s, 3H), 4.07 (q, J ) 7.1 Hz, 2H), 6.72 (m, 3H), 7.25 (m,
5H), 7.40 (d, J ) 9.0 Hz, 1H); CI MS (CH4) m/e (relative
intensity) 367 (M+ + 1, 100). Anal. Calcd for C22H26N2O3: C,
72.11; H, 7.15; N, 7.64. Found: C, 71.93; H, 7.10; N, 7.53.
The optical purity of Nb-benzyltryptophan 47 was assessed in
the exact same manner as in the Na-sulfonamido case 51
employing the europium chiral shift reagent described in
references 29 and 33. The methyl signal in the 1H NMR
spectrum (δ 1.14) did not split even after 10 additions of the
chiral shift reagent, thereby indicating an optical purity of
greater than 98% ee.
C27H28N2O5S: C, 65.85; H, 5.69; N, 5.69. Found: C, 65.96; H,
5.81; N, 5.62.
The optical purity of 51 was determined by H NMR in the
1
presence of the chiral shift reagent tris[3-[(trifluoromethyl)-
hydroxymethylene]-(+)-camphorato]europium III.29,33 The 1H
NMR of racemic 51 in CDCl3 was run as the standard
spectrum. The 1H NMR spectrum was taken after the addition
of 2 drops of the chiral shift reagent in CDCl3 (50% solution),
and this process was repeated for five additions. The methyl
triplet was split immediately (δ 1.07) when the first two drops
were added to (()-51. The separation of the triplets became
greater as more chiral shift reagent was added until the signal
was integrated. When the above benzyl amine 51 was
analyzed in the same manner, the methyl triplet (δ 1.07) never
split, indicating an optical purity of at least greater than 98%
ee. This was checked vs addition of 3% of spiked material (().
cis- (52a ) a n d tr a n s-(1S,3R)-2-Ben zyl-3-(eth oxyca r bo-
n yl)-1-(2′-ca r boxyeth yl)-7-m eth oxy-9-(ben zen esu lfon yl)-
1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (52b). The (-)-
6-methoxy-Na-benzenesulfonyl-Nb-benzyltryptophan ethyl ester
51 (1.03 g, 2.2 mmol) was dissolved in toluene (40 mL) and
dioxane (40 mL) in a three-neck round-bottom (250 mL) flask
equipped with a Dean-Stark trap (DST) and a reflux con-
denser. The solution was heated to reflux and followed by the
dropwise addition of a solution of R-ketoglutaric acid (0.295
g, 2.32 mmol, 1.05 equiv) in dioxane (10 mL). After 7 h at
reflux, another portion of R-ketoglutaric acid (0.295 g, 2.32
mmol, total of 2.1 equiv) in dioxane (10 mL) was added. The
reaction mixture was held at reflux for 44 h with continuous
removal of water via a Dean-Stark trap. The mixture was
allowed to cool followed by the removal of solvent under
reduced pressure to provide a mixture of cis to trans acids 52a
(49) and 52b (51), respectively. Further purification was
achieved with a short wash column on silica gel with EtOAc/
hexane (50:50) to yield the mixture of diastereomers 52a ,b (72
mg, 60%).29
tr a n s-(1S,3R)-(-)-2-Ben zyl-3-(et h oxyca r b on yl)-1-(2′-
ca r b oxyet h yl)-7-m et h oxy-9-m et h yl-1,2,3,4-t et r a h yd r o-
9H -p yr id o[3,4-b]in d ole (48). The (+)-6-methoxy-(D)-Na-
methyl-Nb-benzyltryptophan ethyl ester 47 (21.5 g, 0.0587 mol)
was dissolved in toluene (800 mL) and dioxane (800 mL) in a
three-neck round-bottom (3 L) flask equipped with a Dean-
Stark trap (DST) and a reflux condenser. The solution was
heated to reflux followed by the dropwise addition of a solution
of R-ketoglutaric acid (8.6 g, 0.0588 mol) in dioxane (90 mL).
The reaction mixture was held at reflux for 11 h with
continuous removal of water. The solution was allowed to cool
followed by the removal of solvent under reduced pressure to
provide exclusively the trans acid (13C NMR). Further puri-
fication of the crude solid was carried out on a short wash
column of silica gel with EtOAc/hexane (50:50) to yield the
pure trans acid 48 as an amorphous solid (21.9 g, 83.0%): [R]25
D
) -24.0° (c ) 1.2, CHCl3); IR (CHCl3) 3254, 2937, 1732, 1706
1
cm-1; H NMR (CDCl3) δ 1.35 (t, J ) 7.1 Hz, 3H), 1.93-2.04
(m, 2H), 2.45 (qt, J ) 30.0, 17.5 and 5.8 Hz, 1H), 3.09 (d, J )
8.2 Hz, 1H), 3.45 (d, J ) 13 Hz, 1H), 3.53 (s, 3H), 3.63 (d, J )
6.3 Hz, 1H), 3.67 (s, 3H), 3.96 (d, J ) 13 Hz, 1H), 4.09 (t, J )
8.5 Hz, 1H), 4.28 (dq, J ) 7.2 and 3.6 Hz, 2H), 6.79 (s, 1H),
6.82 (d, J ) 2.1 Hz, 1H), 7.22-7.44 (m, 6H); 13C NMR (CDCl3)
δ 14.25, 20.48, 27.91, 29.74, 30.96, 52.92, 54.43, 55.95, 56.16,
60.99, 93.58, 106.56, 108.79, 118.78, 121.07, 127.39, 128.36,
129.48, 133.65, 138.65, 138.51, 156.60, 172.27, 177.17; EIMS
(70 eV) m/e (relative intensity) 450 (M+, 6.3), 377 (100), 213
(41.2). Anal. Calcd for C26H30N2O5: C, 69.30; H, 6.72; N, 6.22.
Found: C, 68.90; H, 6.61; N, 6.24.
tr a n s 52b: 1H NMR (CDCl3, 250 MHz) δ 1.28 (t, 3H, J )
7.1 Hz), 2.05 (m, 1H), 2.41-2.50 (m, 3H), 2.96 (d, 2H, J ) 7.6
Hz), 3.37 (d, 1H, J ) 13.7), 3.85 (d, 1H, J ) 13.7 Hz), 3.87 (s,
3H), 4.05 (t, 1H, J ) 8.1 Hz), 4.10-4.32 (m, 2H), 6.89 (dd, 1H,
J ) 8.6 and 2.2 Hz), 7.22-7.68 (m, 13H), 7.70 (d, 1H, J ) 2.1
Hz); 13C NMR (CDCl3) δ 14.17, 20.63, 28.90, 31.72, 53.44,
54.92, 55.84, 57.73, 61.12, 100.16, 112.69, 117.46, 119.00,
123.42, 126.38, 127.53, 128.30, 128.87, 129.14, 133.68, 134.52,
138.16, 138.39, 158.27, 171.97, 178.76.
(-)-6-Met h oxy-Na -(b en zen esu lfon yl)-Nb -b en zylt r yp -
top h a n Eth yl Ester (51). To a solution of (-)-Na-(benzene-
sulfonyl)-6-methoxytryptophan ethyl ester 5033 (11.4 g, 0.0284
mol) in anhydrous methanol (45 mL) was added benzaldehyde
(3.31 g, 0.312 mol). The solution which resulted was stirred
for 3 h at rt. The mixture was then cooled in a large ice-salt
bath to -5 °C (solution temperature) and sodium borohydride
(1.104 g, 0.0292 mol) was added in portions over a period of 1
h (without allowing the temperature to rise above 0 °C). The
reaction was monitored by TLC on silica gel with MeOH/EtOAc
(10:90) as the eluent for the formation of the imine (Rf imine
) 0.65, Rf starting mat. ) 0.30) and EtOAc/hexane (50:50) for
the formation of the amine 51 (Rf imine ) 0.50, Rf amine )
0.40). The solution was allowed to stir for an additional 2.5 h
followed by the addition of ice-water (5 mL). The solvent was
removed under reduced pressure. The residue was dissolved
in CH2Cl2 and extracted with H2O (20 mL). The organic layer
was washed with brine (2 × 50 mL) and dried (K2CO3). The
solvent was removed under reduced pressure to furnish Nb-
benzylamine 51 which was purified by flash chromatography
on silica gel with EtOAc/hexane (40:60) to yield the free base
cis- (53a ) a n d tr a n s-(1S,3R)-2-Ben zyl-3-(eth oxyca r bo-
n yl)-1-[2′-(eth oxyca r bon yl)eth yl]-7-m eth oxy-9-(ben zen e-
su lfon yl)-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (53b).
The mixture of cis and trans acids 53a and 53b (50 mg, 0.08
mmol) was dissolved in anhydrous EtOH (10 mL) containing
2% HCl (w/w) and triethyl orthoformate (0.3 mL). The mixture
was heated to reflux for 45 min under an atmosphere of N2.
The solvent was removed under reduced pressure, and the
residue was brought to pH 9 with aqueous NH3 (14%) and then
extracted with CH2Cl2 (3 × 10 mL). The combined organic
layers were dried (K2CO3), and the solvent was removed under
reduced pressure to afford a yellow oil which was purified via
flash chromatography (silica gel, EtOAc/hexane, 30:70) to
provide the cis and trans diethyl esters 53a ,b as a colorless
oil (45 mg, 86%) in an approximate 1:1 ratio: 1H NMR (CDCl3,
250 MHz) of the trans isomer 53b δ 1.21 (t, 3H, J ) 7.1 Hz),
1.25 (t, 3H, J ) 7.1 Hz), 2.05 (m, 1H), 2.30-2.51 (m, 3H), 2.90
(d, 2H, J ) 6.7 Hz), 3.32 (d, 1H, J ) 14.1 Hz), 3.79 (d, 1H, J
) 14.0 Hz), 3.83 (s, 3H), 4.06 (q, 4H, J ) 7.3 Hz), 4.19 (m,
2H), 6.86 (dd, 1H, J ) 8.5 and 2.2 Hz), 7.25 (d, 1H, J ) 6.9
Hz), 7.32 (br s, 7H), 7.39 (d, 1H, J ) 8.0 Hz), 7.51 (d, J ) 5.8
Hz), 7.55 (d, 1H, J ) 7.1 Hz), 7.72 (d, 1H, J ) 2.2 Hz); 13C
NMR (CDCl3) δ 14.09 (q), 14.14 (q), 20.63 (t), 29.34 (t), 31.34
(t), 53.26 (t), 54.77 (d), 55.73 (q), 57.41 (d), 60.05 (t), 60.83 (t),
100.10 (d), 112.47 (d), 117.35 (s), 118.85 (d), 126.31 (d), 126.68
(d), 128.11 (d), 128.55 (d), 129.03 (d), 133.56 (d), 135.03 (s),
138.81 (s), 158.07 (s), 172.16 (s), 173.48 (s).
51 (12.2 g, 87.3%); [R]26 -16.2° (c ) 1.0, CHCl3); IR (neat)
D
1
3339, 2980, 1729 cm-1; H NMR (CDCl3, 250 MHz) δ 1.08 (t,
3H, J ) 7.12 Hz), 1.83 (br s, 1H), 2.96 (d, 2H, J ) 6.7 Hz),
3.53 (t, 1H, J ) 7.1 Hz), 3.61 (d, 1H, J ) 13.1 Hz), 3.77 (d, 1H,
J ) 13.1 Hz), 3.65 (s, 3H), 4.01 (dq, 2H, J ) 7.2 and 2.5 Hz),
6.81 (dd, 1H, J ) 8.7 and 2.3 Hz), 7.21-7.51 (m, 11H), 7.79
(d, 2H, J ) 7.3 Hz); 13C NMR (CDCl3) δ 14.03, 28.97, 51.96,
55.68, 60.26, 60.67, 98.06, 112.15, 118.73, 120.00, 122.91,
124.69, 126.51, 126.98, 128.02, 128.27, 129.07, 133.56, 136.12,
138.14, 139.49, 158.03, 174.23; EIMS (70 e/V) m/e (relative
intensity) 492 (M+, 2.0), 192 (100), 160 (78). Anal. Calcd for
Attem p ted Cis (53a ) to Tr a n s (53b) Ep im er iza tion of
cis- a n d tr a n s-(1S,3R)-2-Ben zyl-3-(eth oxyca r bon yl)-1-[2′-
(eth oxyca r bon yl)eth yl]-7-m eth oxy-9-(ben zen esu lfon yl)-
1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (53). The mix-

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 59
ture of cis and trans diethyl esters 53a ,b (120 mg, 0.20 mmol)
was dissolved in anhydrous EtOH (30 mL) containing 2% HCl
(w/w). The mixture was allowed to reflux for 3 h under an
atmosphere of nitrogen. The solvent was removed under
reduced pressure, the residue was brought to pH 9 with
aqueous NH3 (14%) and then extracted with CH2Cl2 (3 × 20
mL). The combined organic layers were dried (K2CO3), and
the solvent was removed under reduced pressure to afford a
yellow oil which was purified via flash chromatography (silica
gel, MeOH/hexane, 5:95) to provide the ring-cleaved benzyl-
amine 51 as the major product (92 mg, 76%). The spectral
properties of 51 were identical to that reported above for the
benzylation of 50. Note: the pure trans diester 53b and the
pure cis diester 53a were subjected to the above conditions
(24 h) and also provided the same Nb-benzylamine 51.
oil (19.2 mg, 96%). The spectral properties of 60b were
identical to that reported immediately above for the trans
diester 60b.
t r a n s-(1R ,3R )-(-)-2-Be n zyl-3-(m e t h oxyca r b on yl)-1-
[(m eth oxyca r bon yl)eth yl]-1,2,3,4-tetr a h yd r o-9H-p yr id o-
[3,4-b]in d ole (63b) a n d cis-(1S,3R)-(+)-2-Ben zyl-3-(m eth -
o x y c a r b o n y l)-1-[(m e t h o x y c a r b o n y l)e t h y l]-1,2,3,4-
tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (63a ). Optically active
Nb-benzyl-(D)(+)-tryptophan methyl ester 61 (5.0 g, 16.2 mmol)
was dissolved in C6H6 (50 mL) and dioxane (50 mL) in a 250
mL round-bottom flask which was equipped with a Dean-
Stark trap (DST) and a reflux condenser. To this solution,
R-ketoglutaric acid (2.6 g, 17.8 mmol) was added. The reaction
mixture was heated to reflux for 3 h and then allowed to cool
to rt. The solution was diluted with EtOAc (100 mL) and
washed with H2O (2 × 50 mL) and brine (1 × 50 mL). The aq
layer was extracted with EtOAc (2 × 30 mL). After removal
of solvent from the combined organic layers under reduced
pressure followed by purification with flash chromatography
(silica gel, EtOAc/hexane, 50/50, v/v), the ester acids were
obtained as a mixture of cis 62a and trans 62b diastereomers
(4.99 g, 85%). The diastereomeric acids were dissolved in CH2-
Cl2 (50 mL) and the resulting solution cooled to -20 °C. To
this stirred solution an excess of ethereal diazomethane was
added over a 10 min period at -20 °C. The mixture which
resulted was then stirred at rt for an additional 1.5 h. After
removal of solvent under reduced pressure followed by separa-
tion via flash chromatography (silica gel, EtOAc/hexane, 10/
90, v/v) the trans diester 63b (4.0 g, 83% from ester acid) and
the cis diester 63a (560 mg, 12% from ester acid) were
obtained.
Eth yl 3-F or m ylp r op ion a te (59). A solution of γ-butyro-
lactone 54 (86 g, 1 mol) and concd H2SO4 (2 g) in EtOH (500
mL) was stirred at 20 °C for 33 h, after which calcium
carbonate (15 g) was added, and the mixture which resulted
was stirred for an additional 2 h. The reaction mixture was
filtered and the solution poured through a bed of K2CO3. (It
was found that this reaction was partially reversible if the
solvent was removed when the solution was acidic; therefore,
the pH of the solution should be neutral before the removal of
solvent by first pouring the solution through a bed of K2CO3.)
The solvent was then removed under reduced pressure at 18
°C to provide 4-hydroxybutyrate (65 g, 55%). To a solution of
the EtOH free residue (40 g, 0.30 mol) and dry CH2Cl2 (500
mL) was added pyridinium dichromate (265 g, 0.70 mol), and
this mixture was stirred via a mechanical stirrer for 48 h at
rt. The dark mixture which resulted was diluted with CH2-
Cl2 (300 mL) and then filtered through a short bed of silica
gel. The liquid that remained was purified by fractional
distillation (90 °C, 0.5 mmHg) to provide pure ethyl 3-formyl-
63a : [R]28 ) +0.58° (c ) 2.5, CHCl3), lit.38 [R]17 ) -1.3°
D
D
(c ) 1.0, CHCl3); IR (KBr) 3470, 1730 cm-1; 1H NMR (CDCl3)
δ 1.74-1.90 (m, 1H), 1.90-2.01 (m, 1H), 2.43-2.56 (m, 2H),
2.99 (dd, 1H, J ) 6.3 and 15.8 Hz), 3.24 (dd, 1H, J ) 3.6 and
15.8 Hz), 3.56 (s, 3H), 3.62 (s, 3H), 3.76-3.94 (m, 2H), 3.84
(dd, 1H, J ) 14.0 Hz), 3.94 (dd, 1H, J ) 14.0 Hz), 7.02-7.44
(m, 8H), 7.52 (d, 1H, J ) 7.2 Hz), 8.01 (brs, 1H); 13C NMR
(CDCl3) δ 20.03, 29.38, 30.40, 51.48, 51.75, 56.37, 58.82, 59.42,
110.88, 118.24, 119.43, 121.76, 127.30, 128.38, 128.97, 133.47,
136.50, 139.05, 173.88, 174.56; MS (EI) m/z (relative intensity)
406 (M+, 15%), 347 (14), 319 (100), 259 (13), 169 (25). Anal.
Calcd for C24H26N2O4: C, 70.91; H, 6.45; N, 6.89. Found: C,
70.87; H, 6.32; N, 6.75.
propionate 59 (25.3 g, 63%): IR (neat) 2987, 1734, 1718 cm-1
;
1H NMR δ 1.11 (t, 3H, J ) 7.1 Hz), 2.46 (t, 2H, J ) 6.6 Hz),
2.65 (t, 2H, J ) 6.4 Hz), 3.99 (q, 2H, J ) 7.1 Hz), 9.65 (s, 1H).
Note: It was found the best way to monitor these two reactions
1
was analysis via the H NMR spectrum of samples taken from
the reaction mixtures.
cis- 60a a n d tr a n s-(1S,3R)-2-Ben zyl-3-(eth oxyca r bo-
n yl)-1-[2′-(et h oxyca r b on yl)et h yl]-7-m et h oxy-9-m et h yl-
1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (60b). To a so-
lution of (+)-6-methoxy-Na-methyl-Nb-benzyl-(D)tryptophan
ethyl ester 47 (500 mg, 1.6 mmol) in refluxing benzene (5 mL)
was added ethyl 3-formylpropionate 59 (213 mg, 1.6 mmol)
dropwise over several minutes. The mixture was held at reflux
for 10 h with continuous removal of H2O via a Dean-Stark
trap (DST), after which the solvent was removed under
reduced pressure to provide a brown oil. A portion of the crude
mixture was employed to determine the ratio of cis to trans
63b: mp 152-153 °C (lit.38 150-151 °C); [R]28 ) -35.5° (c
D
) 1.09, CHCl3) lit.38 [R]13D ) -38.0° (c ) 1.0, CHCl3); IR (KBr)
3310, 1731, 1707 cm-1; 1H NMR (CDCl3) δ 1.85-2.15 (m, 2H),
2.20-2.50 (m, 2H), 3.03 (dd, 1H, J ) 5.3 and 15.8 Hz), 3.12
(dd, 1H, J ) 8.8 and 15.8 Hz), 3.48 (s, 3H), 3.75 (s, 3H), 3.59
(d, 1H, J ) 13.6 Hz), 3.84 (d, 1H, J ) 13.6 Hz), 3.87-3.93 (m,
1H), 3.98 (dd, 1H, J ) 5.3 and 8.8 Hz), 7.07-7.35 (m, 8H),
7.43 (d, 1H, J ) 7.2 Hz), 7.98 (br s, 1H); 13C NMR (CDCl3) δ
21.39, 28.94, 29.88, 51.45, 51.85, 53.51, 54.81, 56.79, 110.99,
118.16, 119.51, 121.77, 127.08, 127.13, 128.28, 129.13, 134.26,
136.50, 139.41, 173.42, 174.18; MS (EI) m/z (relative intensity)
406 (M+, 60%), 347 (45), 319 (100), 169 (50). Anal. Calcd for
C24H26N2O4: C, 70.91; H, 6.45; N, 6.89. Found: C, 70.88; H,
6.47; N, 6.91.
1
isomers by H and 13C NMR spectroscopy (cis 60a /trans 60b
) 39/61). The remainder of the material (338 mg, 52%) was
purified via flash chromatography [silica gel, EtOAc/hexane,
20:80; the trans isomer 60b was eluted first (TLC on silica
gel; Rf cis ) 0.26, Rf trans ) 0.30; MeOH/hexane, 5:95,
developed twice)]. The spectral data for trans 60b were
identical to that described earlier from the Fischer esterifica-
tion of the trans acid to provide 60b.29 The yield of this
sequence was not maximized but could be improved greatly
employing excess aldehyde.
Acid -Ca t a lyzed E p im er iza t ion of cis-(1S,3R)-(+)-2-
Ben zyl-3-(m eth oxycar bon yl)-1-[(m eth oxycar bon yl)eth yl]-
1,2,3,4-tetr ah ydr o-9H-pyr ido[3,4-b]in dole (63a) in to tr a n s-
(1R ,3R )-(-)-2-Be n zyl-3-(m e t h oxyca r b on yl)-1-[(m e t h -
oxyca r bon yl)eth yl]-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]-
in d ole (63b). General procedure: To the cis diastereomer 63a
(20 mg, 0.05 mmol) in dry CH2Cl2 (1.0 mL) was added CF3-
COOH (10 mL, 0.12 mmol, 2.4 equiv) via syringe. The solution
was allowed to stir at rt under N2 and monitored by TLC (silica
gel, EtOAc/hexane, v/v, 20/80). After stirring for 10 d the
reaction was stopped (TLC indicated that all cis isomer 63a
had been consumed) and cooled in an ice bath, after which
the mixture was diluted with CH2Cl2 (10 mL) and brought to
pH ) 8 with aq NaHCO3 solution (10%). The aq layer was
separated and extracted with CH2Cl2 (2 × 10 mL). The
combined organic layers were washed with brine (20 mL) and
dried (K2CO3), and the solvent was removed under reduced
pressure to furnish an oil. The oil which resulted was purified
Ep im er iza tion of cis 60a in to tr a n s-(1S,3R)-2-Ben zyl-
3-(eth oxyca r bon yl)-1-[2′-(eth oxyca r bon yl)eth yl]-7-m eth -
oxy-9-m et h yl-1,2,3,4-t et r a h yd r o-9H -p yr id o[3,4-b]in d ole
(60b). The mixture of cis and trans diesters 60a and 60b (20
mg, 0.042 mmol) obtained from the reaction immediately above
was dissolved in CH2Cl2 (10 mL) to which trifluoroacetic acid
(TFA, 100 µL) was added. The reaction mixture was stirred
at rt for 1.5 h after which additional CH2Cl2 (10 mL) was added
and the solution basified with aq NH4OH (14%). The aq layer
was washed with CH2Cl2 (2 × 20 mL), and the combined
organic extracts were washed with brine (20 mL) and dried
(MgSO4). The solvent was removed under reduced pressure
to afford exclusively the pure trans 60b isomer as a colorless

60 J . Org. Chem., Vol. 62, No. 1, 1997
Cox et al.
by flash chromatography (silica gel, EtOAc/hexane, v/v, 20/
80) to furnish pure trans 63b (18 mg) as white crystals in 90%
yield. This ester 63b was spectrometrically identical to that
of an authentic sample of trans 63b including the optical
rotation.
To the cis diastereomer 63a (20 mg, 0.05 mmol) in dry
CHCl3 (1.0 mL) was added CF3COOH (10 mL, 0.12 mmol, 2.4
equiv) via syringe. This solution was stirred at rt under N2
and monitored by TLC (silica gel, EtOAc/hexane, v/v, 20/80).
After stirring for 10 d (TLC indicated that all cis isomer 63a
had been consumed) the reaction was stopped, worked up, and
purified identically to the above procedure to furnish pure
trans 63b (16 mg) in 80% yield. This material was spectro-
metrically identical to that of an authentic sample of trans
63b including the optical rotation.
To the cis diastereomer 63a (20 mg, 0.05 mmol) in dry C6H6
was added CF3COOH (10 mL, 0.12 mmol, 2.4 equiv) via
syringe. This solution was stirred at rt under N2 and
monitored by TLC (silica gel, EtOAc/hexane, v/v, 20/80). After
stirring for 10 d (TLC indicated that about 70% of the cis
isomer 63a had been consumed) the reaction was stopped,
worked up, and purified identically to the above procedure to
furnish pure trans 63b (12 mg) in 60% yield. This material
was spectrometrically identical to that of an authentic sample
of trans 63b including the optical rotation. In addition, 4 mg
of starting cis isomer 63a were recovered.
To the cis diastereomer 63a (20 mg, 0.05 mmol) in dry THF
(1.0 mL) was added CF3COOH (10 mL, 0.12 mmol, 2.4 equiv)
via syringe. The solution was stirred at rt under N2 and
monitored by TLC (silica gel, EtOAc/hexane, v/v, 20/80). After
stirring for 10 d (TLC indicated that no trans isomer 63b had
been formed), the reaction was stopped, worked up, and
purified identically to the above procedure to return 17 mg of
starting cis 63a (85%) which was spectrometrically identical
to that of an authentic sample of cis 63a including the optical
rotation.
To the trans diastereomer 63b (20 mg, 0.05 mmol) in dry
CH2Cl2 (1.0 mL) was added CF3COOH (10 mL, 0.12 mmol, 2.4
equiv) via syringe. The solution was stirred at rt under N2
and monitored by TLC (silica gel, EtOAc/hexane, v/v, 20/80).
After stirring for 10 d (TLC indicated that no new component
had been formed), the reaction was stopped, worked up, and
purified identically to the above procedure to return 17 mg of
starting trans 63b (85%) which was spectrometrically identical
to that of an authentic sample of trans 63b including the
optical rotation.
To a mixture (1:1) of cis diastereomer 63a and trans
diastereomer 63b (20 mg, 0.05 mmol) in dry CH2Cl2 (1.0 mL)
was added CF3COOH (10 mL, 0.12 mmol, 2.4 equiv) via
syringe. This solution was stirred at rt under N2 and
monitored by TLC (silica gel, EtOAc/hexane, v/v, 20/80). After
stirring for 10 d (TLC indicated that all the cis isomer had
been consumed), the reaction was stopped, worked up, and
purified identically to the above procedure to furnish pure
trans 63b (18 mg) in 90% yield. This material was spectro-
metrically identical to that of an authentic sample of trans
63b including the optical rotation.
Acid -Ca t a lyzed E p im er iza t ion of cis-(1S,3S)-(-)-2-
Ben zyl-3-(m eth oxycar bon yl)-1-[(m eth oxycar bon yl)eth yl]-
9-m eth yl-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (64a )
t o tr a n s-(1R,3S)-(+)-2-Ben zyl-3-(m et h oxyca r b on yl)-1-
[(m et h oxyca r b on yl)et h yl]-9-m et h yl-1,2,3,4-t et r a h yd r o-
9H-p yr id o[3,4-b]in d ole (64b). Deuterotrifluoroacetic acid
(1.4 g, 12.3 mmol) was added to a solution of the cis isomer
64a (258 mg, 0.614 mmol) in CH2Cl2 (16 mL), and the reaction
mixture was stirred for 10 d at rt and progress monitored by
TLC (silica gel). The reaction mixture was cooled in an ice
bath, diluted with CH2Cl2 (30 mL), and brought to pH 8 with
aq NaHCO3 solution (10%). The aq layer was separated and
extracted with CH2Cl2 (2 × 15 mL). The combined organic
layers were washed with brine (20 mL) and dried (K2CO3), and
the solvent was removed under reduced pressure to provide
an oil. This material was purified by flash chromatography
(silica gel, v/v, EtOAc/hexane, 20:80) to furnish 64b (237 mg)
in 91.8% yield. This white solid was spectrometrically identi-
cal to that of an authentic sample of trans isomer 64b
including the magnitude of the optical rotation. Trans isomer
64b: [R]28 ) +54.8° (c ) 7.62, CHCl3);9 mp 119-120 °C [lit.9
D
mp 119-120 °C]; IR (KBr) 1735 cm-1
;
1H NMR (250 MHz,
CDCl3) δ 1.85-2.00 (2H, m), 2.38 (dt, 1H, J ) 17.5, 5.6 Hz),
2.61 (ddd, 1H, J ) 17.5, 9.6, 5.6 Hz), 3.05 (dd, 1H, J ) 15.8,
5.5 Hz), 3.12 (dd, 1H, J ) 15.8, 11.0 Hz), 3.39-3.81 (ABq, 2H,
J ) 13.1 Hz), 3.50 (3H, s), 3.65 (3H, s), 3.51 (dd, 1H, J ) 15.7,
5.2 Hz), 3.84 (3H, s), 4.10 (dd, 1H, J ) 11.0, 5.5 Hz), 7.12-
7.40 (8H, m), 7.60 (d, 1H, J ) 8 Hz); 13C NMR (CDCl3) δ 20.25,
27.90, 29.60, 29.71, 51.26, 52.00, 52.79, 53.32, 56.12, 106.29,
108.90, 118.12, 119.11, 121.32, 126.52, 126.96, 128.14, 129.29,
135.67, 137.46, 139.25, 173.34, 173.87; MS (CI, CH4) m/z
(relative intensity) 421 (M + 1, 100).
Trifluoroacetic acid (543 mg, 4.76 mmol) was added to a
solution of the cis isomer 64a and trans isomer 64b (cis, 280
mg, 0.667 mmol; trans, 720 mg, 1.71 mmol) in C6H6 (20 mL),
and the reaction mixture was heated at reflux for 5 h.
Progress was monitored by TLC (silica gel). The reaction
mixture was cooled in an ice bath, diluted with C6H6 (40 mL),
and brought to pH 8 with aq NaHCO3 solution (10%). The aq
layer was separated and extracted with C6H6 (2 × 25 mL).
The combined organic layers were washed with brine (40 mL)
and dried (K2CO3), and the solvent was removed under reduced
pressure to provide an oil. This material was purified by flash
chromatography (silica gel, v/v, EtOAc/hexane, 20:80) to
furnish trans 64b (880 mg) in 88% yield. This white solid was
spectrometrically identical to that of an authentic sample of
trans isomer 64b including the optical rotation, [R]28D ) +54.6°
(c ) 1.42, CHCl3).
Zin c Ch lor id e-Ca ta lyzed Ep im er iza tion of th e cis-
(1S,3S)-(-)-2-Ben zyl-3-(m eth oxyca r bon yl)-1-[(m eth oxy-
ca r bon yl)eth yl]-9-m eth yl-1,2,3,4-tetr a h yd r o-9H-p yr id o-
[3,4-b]in d ole (64a ) in to th e tr a n s-(1R,3S)-(+)-2-Ben zyl-
3-(m e t h oxyca r b on yl)-1-[(m e t h oxyca r b on yl)e t h yl]-9-
m eth yl-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (64b).
The cis isomer 64a (200 mg, 0.48 mmol) was dissolved in dry
distilled CHCl3 (8 mL). Anhydrous ZnCl2 (99.999%, 300 mg,
2.20 mmol) was added to the above solution in one portion.
The mixture was allowed to reflux for 30 h under Ar, after
which time it was cooled to rt and filtered. The organic layer
was washed with aq NaHCO3 (10%) and brine (20 mL) and
dried (K2CO3). The solvent was removed under reduced
pressure to provide an oil. The oil was purified by passage
through a short wash column (silica gel, v/v, EtOAc/hexane,
20:80) to provide the trans isomer 64b (185 mg, 92.5%) which
was spectrometrically identical to that of an authentic sample
of trans isomer 64b including the optical rotation, [R]28
)
D
+54.8° (c ) 1.97, CHCl3). As a control experiment the cis
isomer 64a was heated under exactly the same conditions in
dry distilled CHCl3 (8 mL) in the absence of ZnCl2. Under
these conditions no trans isomer 64b was observed, and only
starting material 64a was recovered. (The acid-free dry,
distilled CHCl3 was prepared by drying reagent grade CHCl3
over KOH, followed by refluxing over sodium hydride and then
distillation under Ar.)
Ep im er iza tion of th e cis-(1S,3S)-(-)-2-Ben zyl-3-(m eth -
oxycar bon yl)-1-[(m eth oxycar bon yl)eth yl]-9-m eth yl-1,2,3,4-
tetr a h yd r o-9H-p yr id o[3,4-b]in d ole (64a ) Hyd r och lor id e
Sa lt to th e tr a n s-(1R,3S)-(+)-2-Ben zyl-3-(m eth oxyca r bo-
n yl)-1-[(m eth oxycar bon yl)eth yl]-9-m eth yl-1,2,3,4-tetr ah y-
d r o-9H-p yr id o[3,4-b]in d ole (64b). The anhydrous hydro-
chloride salt 64a ‚HCl was prepared as follows: the cis isomer
64a (200 mg, 0.48 mmol) was dissolved in dry CH3OH (10 mL)
which contained 2% HCl (w/w). The mixture was allowed to
stir 10 min at rt to form the hydrochloride salt. The solvent
was removed under reduced pressure at rt to provide an oil.
The oil was dissolved in dry CH3OH (1 mL), and cold ether
(20 mL) was added to precipitate out the hydrochloride salt of
64a . The white solid which formed was filtered by vacuum
filtration. The solid (64a ‚HCl) was washed with dry cold ether
(2 × 10 mL) and dried. The HCl salt of 64a [analysis by TLC
(silica gel) indicated that no epimerization of the HCl salt had
occurred] was dissolved in CH2Cl2 (10 mL) and the solution
stirred for 18 d at rt under N2. The reaction mixture was
washed with 10% aq NaHCO3 (20 mL) and brine (20 mL) and
dried (K2CO3). The solvent was removed under reduced

Trans 1,3-Disubstituted Tetrahydro-â-carbolines
J . Org. Chem., Vol. 62, No. 1, 1997 61
pressure to provide an oil which was purified by flash chro-
matography (silica gel, v/v, EtOH/hexane, 20:80) to afford trans
isomer 64b (178 mg, 89%) and this material was spectrometri-
cally identical to that of an authentic sample of trans 64b
gel, 99:1, CHCl3/CH3OH) indicated that all of the cis isomer
66a had been converted into the trans compound 66b. The
reaction mixture was diluted with CH2Cl2 (20 mL), washed
with 10% aq NH4OH and brine, dried (K2CO3), and the solvent
was removed under reduced pressure. The residue was passed
through a small plug of silica gel to provide 345 mg of the trans
diastereomer 66b as a white solid (90%). The spectral proper-
ties and optical rotation of this compound were identical to
those for authentic trans isomer 66b.
including the optical rotation, [R]28 ) +55.2° (c ) 1.62,
D
CHCl3).
cis-(1R,3R)-(-)-2-Ben zyl-1-(1,3-d it h ia n -2-ylm et h yl)-3-
(m eth oxyca r bon yl)-1,2,3,4-tetr a h yd r o-9H-p yr id o[3,4-b]-
in dole (66a) an d tr a n s-(1S,3R)-(-)-2-Ben zyl-1-(1,3-dith ian -
2-ylm eth yl)-3-(m eth oxyca r bon yl)-1,2,3,4-tetr a h yd r o-9H-
p yr id o[3,4-b]in d ole (66b). Optically pure Nb-benzyl-(D)(+)-
tryptophan methyl ester 61 was prepared according to the
method of Zhang et al.9,10 A mixture of this ester 61 (3.0 g,
9.7 mmol), dry C6H6 (15 mL), and 1,3-dithiane-2-acetaldehyde
(65) (1.55 g, 9.56 mmol) was heated to reflux under N2 for 14
h at which time analysis by TLC (silica gel, 4:6 EtOAc:hexanes)
indicated the complete consumption of aldehyde. The solvent
was removed under reduced pressure and the residue passed
through a wash column (silica gel) to provide a mixture of cis
66a and trans 66b diastereomers (4.1 g) in a combined yield
of 93% in a ratio of 30:70, respectively (1H NMR). The cis and
trans diastereomers were separated by flash chromatography
(silica gel, 99:1, CHCl3/hexanes). cis 66a : Rf ) 0.14 (TLC, 99:
Ep im er iza tion of th e Cis Isom er 66a in to th e Tr a n s
Isom er 66b in CF 3COOD/CH2Cl2. To the cis diastereomer
66a (10 mg, 0.02 mmol) in dry CH2Cl2 (0.5 mL) was added
CF3COOD (5 µL, 0.06 mmol, 2.9 equiv) via micropipet. The
solution which resulted was stirred at room temperature under
N2 and monitored by TLC (silica gel, 99:1, CHCl3/CH3OH).
After 48 h the reaction was stopped and worked up identically
to the above procedures. Note: a small amount of cis isomer
66a was still present at the time the reaction was stopped.
Analysis of the resulting trans isomer 66b by low and high
1
resolution mass spectroscopy as well as H NMR indicated no
deuterium incorporation in 66b.
Evid en ce for a Ca r boca tion In ter m ed ia te: Tr ea tm en t
of th e cis Isom er 66a w ith CF 3COOH/Na BH4 To P r ovid e
th e C(1)-N(2) Scission P r od u ct 3-[2-(Ben zyla m in o)-2-
(m et h oxyca r bon yl)p r op yl]-2-[2-(1,3-d it h ia n -2-yl)et h yl]-
in d ole (67). To an oven-dried 5 mL round-bottom flask
containing dry CH2Cl2 (1.0 mL) was added TFA (80 µL, 1.04
mmol) via syringe, and the solution was cooled to 0 °C followed
by the addition of NaBH4 (12 mg, 0.32 mmol).34 The solution
bubbled vigorously and was stirred under N2 at 0 °C for 0.5 h
and then allowed to warm to rt. The cis diastereomer 66a
(12 mg, 0.03 mmol) was added, and the solution which resulted
was stirred at rt for 1.5 h at which time analysis by TLC
indicated the formation of a new product. Another 0.4 equiv
of TFA was then added to the solution, and it was allowed to
stir at rt overnight (∼16 h). Analysis of the solution by TLC
(silica gel, 3:7, EtOAc/hexanes) indicated approximately 50%
of the cis isomer 66a had been converted into a new compound.
The material which remained was identical in Rf to the trans
isomer 66b. The reaction was diluted with CH2Cl2 and washed
with H2O, 10% aq NH4OH, and brine, and dried (K2CO3), and
the solvent was removed under reduced pressure to yield a
clear oil. The residue was subjected to flash chromatography
(silica gel, 1:9, EtOAc/hexanes) to provide 3.4 mg of 67 (25%)
as a clear oil. 67: Rf ) 0.22 (TLC, 3:7, EtOAc/hexanes); IR
(KBr) 3369, 1728 cm-1; 1H NMR (500 MHz, CDCl3) δ 1.79 (m,
1H), 2.05 (m, 3H), 2.75 (m, 4H), 2.92 (m, 2H), 3.09 (m, 2H),
3.53-3.63 (m, 5H), 3.75 (m, 1H), 3.89 (t, 1H, J ) 7.2 Hz), 7.03
(t, 1H, J ) 8.0 Hz), 7.10 (t, 1H, J ) 8.0 Hz), 7.15-7.27 (m,
6H), 7.52 (d, 1H, J ) 8.0 Hz), 8.03 (s, 1H); 13C NMR (CDCl3)
δ 22.9, 26.0, 28.8, 29.7, 30.2, 35.2, 46.4, 51.7, 52.3, 61.6, 106.0,
110.4, 118.5, 119.4, 121.5, 126.9, 128.0, 128.3, 128.7, 135.1,
135.4, 139.8, 152.0, 155.7, 175.5; MS (CI, CH4) m/z (relative
intensity) 454 (M + 1, 100%), 276 (55). A control experiment
(CF3COOH/NaBH4) with trans isomer 66b carried out under
identical conditions to that described immediately above
returned only trans isomer 66b in 90% yield.
1, CHCl3/CH3OH); [R]25 ) -14.4° (c ) 1.4, CH2Cl2); white
D
solid, mp 100 °C; IR (KBr) 3388, 2900, 1736, 1652, 1455 cm-1
;
1H NMR (CDCl3) δ 2.03 (m, 4H), 2.88 (m, 5H), 3.27 (dd, 1H, J
) 1.17 and 16.23 Hz), 3.67 (s, 3H), 3.83 (dd, 1H, J ) 1.8 and
6.4 Hz), 4.02 (q, 2H, J ) 11.6 Hz), 4.31 (m, 1H), 4.57 (m, H),
7.16 (m, 2H), 7.33 (m, 4H), 7.49 (d, 1H, J ) 6.7 Hz), 7.55 (dd,
1H, J ) 1.5 and 7.5 Hz), 7.92 (s, 1H); 13C NMR (CDCl3) δ 18.7,
26.2, 29.6, 30.2, 40.3, 44.6, 52.1, 54.6, 56.6, 60.2, 106.9, 110.8,
118.3, 119.5, 121.9, 127.2, 127.3, 128.4, 128.8, 133.1, 136.2,
139.4, 173.9; MS (CI, CH4) m/z (relative intensity) 452 (M +
1, 100%), 319 (85). Anal. Calcd for C25H28N2O2S2: C, 66.34;
H, 6.23; N, 6.19. Found: C, 66.44; H, 6.19; N, 6.41. trans
66b: Rf ) 0.22 (TLC, 99:1, CHCl3/CH3OH); [R]25 ) -12.1° (c
D
) 1.4, CH2Cl2); white crystals, mp 179-180 °C; IR (KBr) 3381,
2924, 1729, 1462, 1251 cm-1; 1H NMR (CDCl3) δ 1.75 (m, 1H),
2.06 (m, 3H), 2.62 (m, 2H), 2.84 (m, 2H), 3.07 (m, 2H), 3.45 (d,
1H, J ) 13.4 Hz), 3.82 (s, 4H), 4.08 (m, 3H), 7.14-7.42 (m,
8H), 7.56 (d, 1H, J ) 7.0 Hz), 8.04 (s, 1H); 13C NMR (CDCl3)
δ 20.5, 25.9, 29.7, 30.2, 40.8, 44.3, 52.0, 53.3, 56.7, 106.1, 111.0,
118.3, 119.7, 122.0, 127.1, 128.2, 129.3, 133.5, 136.3, 139.5,
173.1; MS (CI, CH4) m/z (relative intensity) 452 (M + 1, 100%).
Anal. Calcd for C25H28N2O2S2: C, 66.34; H, 6.23; N, 6.19.
Found: C, 66.26; H, 6.12; N, 6.18. The structure of this isomer
66b was verified by single crystal X-ray analysis.
Ep im er iza tion of th e cis Isom er 66a in to th e tr a n s
Isom er 66b in CF 3COOH/CH2Cl2. To the cis diastereomer
66a (155 mg, 0.34 mmol) in dry CH2Cl2 (6 mL) was added via
syringe TFA (55 mL, 0.71 mmol, 2.1 equiv). The solution
which resulted was stirred at rt under N2 for 17 h at which
time analysis by TLC (silica gel, 99:1, CHCl3/CH3OH) indicated
the absence of starting cis isomer 66a . The reaction mixture
was diluted with CH2Cl2 (15 mL), washed with 10% aq NH4-
OH and brine, dried (K2CO3), and the solvent was removed
under reduced pressure. The residue was passed through a
small plug of silica gel to provide 145 mg of the trans
diastereomer 66b as a white solid (94%). The spectral proper-
ties and optical rotation of this compound were identical to
those for authentic trans isomer 66b.
Ep im er iza tion of a Mixtu r e of cis 66a a n d tr a n s 66b
Isom er s in to th e tr a n s Isom er 66b in CF 3COOH/CH2Cl2.
To a mixture of cis 66a and trans 66b diastereomers (380 mg,
0.84 mmol) 30:70 in dry CH2Cl2 (12 mL) was added via syringe
TFA (136 mL, 1.76 mmol, 2.1 equiv). The solution which
resulted was stirred at rt under N2 until analysis by TLC (silica
Ack n ow led gm en t. We thank Mr. Frank Laib for
providing mass spectroscopic data and Mr. Keith Krum-
now for elemental analyses. We are also indebted to
the Graduate School at the University of Wisconsins
Milwaukee and the NIMH (MH 46851) for financial
support.
J O951170A