Total synthesis of gypsetin, deoxybrevianamide E, brevianamide E, and tryprostatin B: Novel constructions of 2,3-disubstituted indoles

Article abstract of DOI:10.1021/ja9925249

A concise and efficient total synthesis of the acyl-CoA:cholesterol acyltransferase inhibitor gypsetin (1) is described. The route features a straightforward method for the introduction of a reverse prenyl group into the C2-position of an N-phthaloyl-protected tryptophan (11). The total synthesis of gypsetin was completed by the dimethyldioxirane-promoted double- oxidative cyclization of a prefashioned diketopiperazine (19). Total syntheses of deoxybrevianamide E (24) and brevianamide E (25) following similar procedures are also described. The reaction of nucleophiles with in situ-generated 3-chloroindolenines provides a route to 2,3-disubstituted indoles from 3-substituted precursors. Indications of the scope and limitations of such reactions are provided. A total synthesis of tryprostatin B (41), a diketopiperazine derived from an L-tryptophan derivative (bearing a prenyl group at the α position of the indole) and L-proline, was accomplished. The key step involved the introduction of the prenyl function onto a protected tryptophan congener (11). A route for the prenylation of ketones with virtually no competitive reverse prenylation is also provided.

Full text of DOI:10.1021/ja9925249

11964
J. Am. Chem. Soc. 1999, 121, 11964-11975
Total Synthesis of Gypsetin, Deoxybrevianamide E, Brevianamide E,
and Tryprostatin B: Novel Constructions of 2,3-Disubstituted Indoles
Jeffrey M. Schkeryantz, Jonathan C. G. Woo, Phieng Siliphaivanh,
Kristopher M. Depew, and Samuel J. Danishefsky*
Contribution from the Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer
Research, 1275 York AVenue, New York, New York 10021, and Department of Chemistry,
Columbia UniVersity, HaVemeyer Hall, New York, New York 10027
ReceiVed July 16, 1999
Abstract: A concise and efficient total synthesis of the acyl-CoA:cholesterol acyltransferase inhibitor gypsetin
(1) is described. The route features a straightforward method for the introduction of a reverse prenyl group
into the C2-position of an N-phthaloyl-protected tryptophan (11). The total synthesis of gypsetin was completed
by the dimethyldioxirane-promoted double-oxidative cyclization of a prefashioned diketopiperazine (19). Total
syntheses of deoxybrevianamide E (24) and brevianamide E (25) following similar procedures are also described.
The reaction of nucleophiles with in situ-generated 3-chloroindolenines provides a route to 2,3-disubstituted
indoles from 3-substituted precursors. Indications of the scope and limitations of such reactions are provided.
A total synthesis of tryprostatin B (41), a diketopiperazine derived from an L-tryptophan derivative (bearing
a prenyl group at the R position of the indole) and L-proline, was accomplished. The key step involved the
introduction of the prenyl function onto a protected tryptophan congener (11). A route for the prenylation of
ketones with virtually no competitive reverse prenylation is also provided.
The development of cholesterol-lowering drugs remains an
ultimately leads to foam cell formation and presumptively to
myocardial infarction. Thus, an ACAT inhibitor might be useful
in lowering cholesterol levels and attenuating the risk of
coronary heart disease.
5
active area of research and development. It is widely accepted
that hypercholestemia is one of the major risk factors leading
1
to coronary heart disease. Containment of excessive levels of
plasma cholesterol has generally been achieved via controlled
dietary absorption of cholesterol-containing foodstuffs as well
as with therapeutic agents that interfere with the de novo
biosynthesis of cholesterol itself. Indeed, a cornerstone in the
treatment of hypercholestemia was the discovery that lovastatin
and other compactin analogues are powerful inhibitors of the
enzyme HMG-CoA reductase, the rate-limiting enzyme in the
Given these considerations, the reports concerning gypsetin
(1, Figure 1) evoked particular interest. The compound was
isolated from the fungal source N. gypsea Var. incurVata IFO
9228. It was found to be a promising competitive inhibitor of
ACAT with respect to oleoyl-CoA, with a Ki value of 5.5 µM.
Furthermore, inhibition of cholesterol ester formation in cultured
6
macrophages with an IC50 of 0.65 µM was observed. In light
2
cholesterol biosynthetic pathway. Agents based on the inhibi-
of the novel structure of gypsetin, its structural relationship to
the ardeemins, brevianamide, and tryprostatin systems (vide
tion of HMG-CoA have effectively lowered LDL and total
cholesterol levels in hypercholestemic patients. In this connec-
tion, squalene synthase inhibitors such as the zaragozic acids
and CP-225,917/263,114 have been identified as possible
therapeutic agents; however, their apparent toxicity and difficult
7
infra) as well as its potential as a lead in drug discovery, this
alkaloid emerged as a target structure for total synthesis.
Upon inspection, it is seen that the structure of gypsetin (1)
embodies reverse prenyl groups at C5a and C13a of the hexahy-
dropyrroloindole nucleus. It seemed not improbable that a simple
and direct method for the incorporation of the reverse prenyl
groups at C5a and C13a could be helpful in completing total
syntheses of gypsetin as well as the brevianamides (vide infra).
This capability could also be helpful in a drug lead discovery
3
availability has hampered their further development.
In recent years, notable attention has been directed to the
possibility of inhibition of the enzyme acyl-CoA:cholesterol
acyltransferase (ACAT). ACAT is the rate-limiting enzyme in
the absorption of cholesterol. Accordingly, inhibition of ACAT-
catalyzed cholesterol esterification could result in blocking the
accumulation of intracellular cholesterol esters in macro-
(5) Kimura, T.; Takase, Y.; Hayashi, K.; Tanaka, H.; Ohtsuka, I.; Takao,
S.; Kogushi, M.; Yamada, T.; Fujimori, T.; Saitou, I.; Akasaka, K. J. Med.
Chem. 1993, 36, 1630 and references therein.
4,5
phages. It is postulated that build-up of these cholesterol esters
(6) (a) Shinohara, C.; Hasumi, K.; Takei, Y.; Endo, A. J. Antibiot. 1994,
(
1) Kannel, W. B.; Castelli, W. P.; Gordon, T.; McNamara, P. M. Ann.
47, 163. (b) Nuber, B.; Hansske, F.; Shinohara, C.; Miura, S.; Hasumi, K.;
Endo, A. J. Antibiot. 1994, 47, 168. (c) Fukuyama, T. F.; Chen, X.; Peng,
G. J. Am. Chem. Soc. 1994, 116, 3127 and references therein.
Intern. Med. 1971, 74, 1. Brown, M. S.; Goldstein, J. L. Angew. Chem.,
Int. Ed. Engl. 1986, 25, 583.
(
2) (a) Grundy, S. M. Cholesterol and Atherosclerosis, Diagnosis and
(7) (a) For a preliminary communication on gypsetin, see: Schkeryantz,
J. M.; Woo, J. C. G.; Danishefsky, S. J. J. Am. Chem. Soc. 1995, 117,
7025. (b) For a preliminary account of the synthesis of tryprostatin B, see:
Depew, K. M.; Danishefsky, S. J.; Rosen, N.; Sepp-Lorenzino, L. J. Am.
Chem. Soc. 1996, 118, 12463. (c) For the total syntheses of amauromine
and 5-N-acetylardeemin, see: Depew, K. M.; Marsden, S. P.; Zatorska,
D.; Zatorski, A.; Bornmann, W. G.; Danishefsky, S. J. J. Am. Chem. Soc.
1999, 121, 11953-11963 and references therein.
Treatment; Tower Medical,: New York, 1990; p 4. (b) Grundy, S. M. New
Engl. J. Med. 1988, 319, 24. (c) Spector, A. A.; Mathur, S. N.; Kaduce, T.
L. Prog. Lipid Res. 1979, 18, 31. (d) Suckling, K. E.; Stange, F. J. Lipid
Res. 1985, 26, 647.
(
3) Abe, I.; Tomesch, J. C.; Wattanasin, S.; Prestwich, G. D. Nat. Prod.
Rep. 1994, 11, 279 and references therein.
4) Sliskovic, D. R.; White, A. D. Trends Pharm. Sci. 1991, 12, 194.
(
1
0.1021/ja9925249 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/09/1999

NoVel Constructions of 2,3-Disubstituted Indoles
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 11965
5
Figure 1.
context. Moreover, a comprehensive strategy to accomplish the
total synthesis goals must also offer provision for incorporating
the required oxygen functionality at the bridgehead positions
(see C8a and C16a) of the pyrroloindole system.
Closer examination of the gypsetin structure reveals an
interesting duality governing the stereochemical connectivity
of the central diketopiperazine ring with the respective junction
functionalities of the two pyrroloindole moieties. Not surpris-
ingly, in each case the pyrroloindole linkage is cis fused.
However, the backbone relationships between these functions
and the two hydrogen atoms at the two fusion points of the
core pyrrolodiketopiperazine ring are different. For instance,
the stereochemical connectivity between the C16a hydroxyl and
its nearest junction hydrogen (C15a) is syn. By contrast, the
connectivity between the C8a hydroxyl and its nearest junction
(C7a) hydrogen is anti. Put differently but equivalently, the
Figure 2.
central diketpiperazine ring can be construed as arising from
the union of two steroechemically distinct amino acids, 2 and
3
. These structures can, in principle, be derived from the
essential amino acid L-tryptohan. However, 2 and 3 differ in
the “absolute” configurations of their pyrroloindole sectors. It
is readily seen that in compound 2, the absolute configuration
of the carbon bearing the bridgehead hydroxyl group (destined
to become C8a) is R, whereas in 3, the corresponding center
(destined to become C16a) is S.
Analysis of the problem in terms of the hypothetical subunits
2
and 3 suggests, in principle, an obvious (though perhaps
difficult to achieve) formal route to gypsetin. The program
would call for synthesizing 2 and 3 in differentially protected
forms (cf. 2′ and 3′), which would eventually allow for the
hetero coupling of the two constructs en route to 1. The success
of such a program would require, ideally, stereoselective access
to the two diastereometrically distinct oxidatively and alkyla-
tively cyclized L-tryptophan derivatives, 2′ and 3′. The likely
course of progression of a hypothetical synthesis along these
lines of conjecture would commence with building a suitable
reverse prenylated L-tryptophan derivative (cf. 4). It would then
be necessary to devise methodology by which oxidative
cyclization of 4 could be used to reach either 2 or 3. This goal
can be envisioned in terms of kinetically controlled oxidative
cyclizations (see hypothesized pathways 4 f 2′ and 4 f 3′).
The prospects of developing two stereospecific kinetic oxidative
cyclization solutions, each culminating in opposite diastereo-
facial outcomes, were daunting.
Figure 3.
protonation solution might involve thermodynamic equilibration
of one of the kinetic oxidative cyclization products to produce
its acyl side chain diastereomer (for the sake of argument, ent
2′ f 3′). It is recognized that, to reach 3′ en route to gypsetin
by either a second-stage “corrective” kinetic or thermodynamic
solution, the oxidative cyclization would have to have been
accomplished in the D-tryptophan series. Epimerization (of ent
2′) at the tryptophan acyl center would transpose the system to
the L-amino acid series but with the required stereochemistry
at the pyrroloindole junction (i.e., 3′).
While it is certainly possible that these generalized concepts
could be reduced to practice with suitable study, we also
considered a rather more concise strategy although, admittedly,
one which was unlikely to produce a stereoselective solution.
It is recognized that, in principle, a preformed, suitable
diketopiperazine could undergo two-fold oxidative cyclization
to furnish gypsetin. We started with the simplifying assumption
that the two cyclization events are independent; i.e., the
diastereofacial course of the first oxidative cyclization would
not influence, in a significant way, the outcome of the second
process. Under these circumstances, if the overall oxidative
cyclization event is stereospecific in either the syn or anti sense,
such a scheme could not deliVer gypsetin, since two opposite
An alternative possibility would entail inversion of config-
uration of the tryptophan acyl center, after the formation of the
kinetic oxidatively cyclized product 2′ or 3′. For the sake of
discussion, we postulate a scenario where the only stereoselec-
tive capability which is accomplished with excellent kinetic
control allows one to progress stereoselectively from 4 f 2′
(Figure 3). In principle, however, deprotonation at the acyl center
of 2′ followed by kinetic quenching could produce a diastere-
omer different from the direct product of oxidative cyclization
(cf. ent 2′ f 3′). An alternative to such a second-stage kinetic

11966 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Schkeryantz et al.
Figure 5. Preparation of 2,3-disubstituted indoles by addition of
nucleophiles to 3-chloroindolenines.
Scheme 1a
a
3
Reagents and conditions: (a) tert-butylhypochlorite, Et N, THF,
-78 °C, 30 min. (b) prenyl 9-BBN (3 equiv), - 78 °C, 6 h, 95%.
Figure 4.
such as austamide, echinulin, and neoechinulin, with the only
outcomes are required (see discussion above). If, however, each
of the disconnected oxidative cyclization reactions is stereo-
random, then the prospects for reaching gypsetin are excellent.
Toward this end, there would be required one syn and one anti
closure. In the limiting case of stereorandomness in both the
first and second cyclization steps, the expected ratio would be
gypsetin (syn:anti ) anti:syn, 2):6 (syn:syn, 1):7 (anti:anti, 1)
practical method arising from various Claisen rearrangements.8
Of the existing methods available for the synthesis of 2,3-
disubstituted indoles, none seemed particularly attractive with
9
respect to the particular objective required here.
Recalling the oxidative conversion of 2,3-disubstituted indoles
to chloroindolenines and the use of such chloroindolenines for
10a
functionalization of the carbon benzylic to C2, we asked
(Figure 4). The expected preference for reaching gypsetin over
whether an intermediate such as 8 might have transient viability
either 6 or 7 arises, of course, from the symmetry of the system
wherein syn:anti (s,a) is identical with anti:syn (a,s). Therefore,
two of the four equally probable permuted outcomes converge
on gypsetin.
10b
even if C2 were unsubstituted. We hoped that 8 might suffer
nucleophilic attack at C2 leading to 9 and thence, after tau-
tomerization, to 10 (Figure 5). Development of this protocol
would allow for a rapid and general preparation of 2,3-
disubstituted indoles.
In this paper, we describe a remarkably concise total synthesis
7
of gypsetin based on the diketopiperazine double-cyclization
Accordingly, the synthesis of gypsetin started with N-
concept. In so doing, it was first necessary to learn how to
introduce a reverse prenyl group in to the 2-position of a
11
phthaloyltryptophan methyl ester (11). Treatment of this
compound with tert-butylhypochlorite and triethylamine led to
3
-substituted indole. This problem was quite different from the
the formation of an unstable 3-chloroindolenine, presumably
challenge which we faced in the total synthesis of 5-N-acetyl-
ardeemin, wherein the goal required introduction of a reverse
prenyl group into the 3-position of a pyrroloindole.7c We will
also relate the application of this methodology to the total
synthesis of deoxybrevianamide E (24) and brevianamide (25).
We then describe extension of the new method for C2-alkylation
of indoles to encompass nucleophiles other than reverse prenyl
groups.
12
1
2. Happily, addition of freshly prepared prenyl-9-BBN to this
intermediate led to 13 in 95% yield. Thus, the reverse prenyl
group had been cleanly incorporated at C2 of the indole ring
(Scheme 1). It was particularly gratifying to find that the reverse
prenylation reaction at the R-position of the indole in suitably
protected tryptophan derivative had been achieved without
noticeable racemization.
Interestingly, both prenyl tri(n-butyl)stannane and prenyl
trimethylsilane failed to react with the chloroindolenine under
the influence of several Lewis acid promoters, such as BF3‚
Et2O, ZnCl2, and SnCl4. Since there are relatively few direct
methods for the preparation of 2,3-disubstituted indoles, studies
During the course of these studies, another natural product,
tryprostatin B (41), was isolated. This compound was reported
to disrupt progression of ts-ft210 cells at the G2/M phase barrier
(vide infra). The goal of a total synthesis of tryprostatin B
implicitly raised another issue, i.e., that of introduction of a
prenyl, rather than a reverse prenyl group, at C2 of an indole
and, by extension, to other systems. The attainment of a
nucleophilic prenylating protocol and its application to the total
synthesis of tryprostatin B is described in the concluding phase
of this report.
(8) (a) Tomita, K.; Terada, A.; Tachikawa, R. Heterocycles 1976, 4, 733.
b) Godtfredsen, W. O.; Vangedal S. Acta Chem. Scand. 1956, 10, 1414.
c) Owellen, R. J. J. Org. Chem. 1974, 39, 69. (d) Pleininger, H.; Kraemer,
(
(
H.-P.; Sirowej, H. Chem. Ber. 1974, 107, 3915. (e) Takamatsu, N.; Inoue,
S.; Kishi, Y. Tetrahedron Lett. 1971, 4661. (f) Takamatsu, N.; Inoue, S.;
Kishi, Y. Tetrahedron Lett. 1971, 4665. (g) Hutchison, A. J.; Kishi, Y. J.
Am. Chem. Soc. 1979, 101, 6786.
(
9) Saulnier, M. G.; Gribble, G. W, J. Org. Chem. 1982, 47, 2810.
Results and Discussion
(10) (a) Kuehne, M. E.; Hafter, R. J. Org. Chem. 1978, 43, 3702 and
references therein. (b) For the only published example of a carbon-based
nucleophilic addition to a C2-unsubstituted chloroindolenine, see: Parsons,
R. L.; Berk, J. D.; Kuehne, M. E. J. Org. Chem. 1993, 58, 7482.
Installation of the Reverse Prenyl Group. At the outset of
the program, we focused on an efficient method for introducing
the reverse prenyl group at the R-position of the indole moiety
of a suitably protected tryptophan. This type of problem had
received attention in connection with the synthesis of alkaloids
(
11) Prepared from L-tryptophan methyl ester by modification of the
procedure listed in the following: Bodansky, M.; Bodansky, A. The Practice
of Peptide Synthesis; Springer-Verlag: Berlin, 1984; p10.
(12) Kramer, G. W.; Brown, H. C. J. Organomet. Chem. 1977, 132, 9.

NoVel Constructions of 2,3-Disubstituted Indoles
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 11967
Scheme 2. Oxidative Ring Closure of Tryptophan
Scheme 3a
Compounds
a
Reagents and conditions: (a) NH
O, Et N, THF, 1 h, quantitative. (c) LiOH/
O, room temperature, 3 h, 100%. (d) 18 (0.91 equiv),
BOP-Cl (3 equiv), CH Cl , -78 f 0 °C, 1 h. (e) TFA, CH Cl , room
temperature, 1 h. (f) NH , MeOH, reflux, 12 h, 73% from 18.
2 2
NH , EtOH, room temperature,
3
days, 65%. (b) (Boc)
THF/MeOH/H
2
3
2
2
2
2
2
of the reactions of other nucleophiles with transiently generated
C2-unsubstituted C3-chloroindolenines were of interest. The
results of that investigation are presented later in this disclosure
3
these structures would have to be differentiated, and a difficult
separation of diastereomers would have to be overcome. Given
the absence of stereoselectivity in the reactions with DMDO,
and given the gloomy prognosis from literature precedents which
documented other related oxidative cyclizations, our hopes to
gain stereospecific access to 2′ and 3′ would not be readily
achievable and were, for the moment, set aside. The realities
of our situation strongly favored the prebuilt diketopiperazine
approach (see Figure 4). The advantage of the double cyclization
is that the protecting group issues would be much simplified.
The predicted stereorandomness of the reaction would have been
turned to some advantage from the standpoint of conciseness
and could benefit additionally from the stochastic analysis
offered above. We therefore embarked on this course.
Total Synthesis of Gypsetin. Fashioning of the diketopip-
erazine portion of gypsetin required the removal of the phthal-
imide protecting group from 17. This result was accomplished
via hydrazinolysis in ethanol at ambient temperature for 3 days
to provide 17 in 65% yield after chromatography. Attempts to
increase the yield and decrease the reaction times by heating
13 with hydrazine in refluxing methanol resulted in concurrent
hydrazinolysis of the methyl ester.
(vide infra).
Oxidative Ring Closure To Form Systems of the Type 2
and 3. With a direct solution for the introduction of a reverse
prenyl moiety at C2 of a tryptophan ring system in hand, our
attention turned to the gross feasibility and stereoselectivity in
the formation of pyrrolo[2,3-b]indoline systems such as 2′ and
3
′. As discussed above, the synthesis of gypsetin would require
access to both of these hypothetical amino acids for hetero
coupling. The ideal method for obtaining these intermediates
would obviously involve kinetically controlled stereoselective
oxidative cyclizations of suitably protected tryptophans (cf. 4),
leading specifically to either 2′ or 3′. We envisioned formation
of systems such as 2′ and 3′ by oxidative cyclization of Nb of
a suitable amino side chain mounted at C2 of the indole with
concomitant hydroxylation at C3 through the agency of a formal
+
hydroxenium ion ( OH). Oxidative cycloaromatizations of
tryptophans are well precedented,1 but only a few examples
3
of oxidative cyclizations, where the C3 hydroxyl survives, had
been practiced. One such instance involves the use of singlet
oxygen1
4,15
(vide infra). We chose, instead, the mild oxidant
16
dimethyldioxirane (DMDO) as a possible source of the desired
hydroxenium equivalent due to the extensive background of our
laboratory in the use of this agent in the epoxidation of sensitive
A five-step procedure was developed to provide the required
diketopiperazine from the amino ester 13 (Scheme 3). First, the
BOC derivative (14) was prepared from 13 via 17. Saponifica-
tion of the methyl ester of 14 gave the acid, 18. Coupling of
this compound to amino ester 17 was accomplished under
mediation of BOP-Cl. Removal of the t-BOC group with
trifluoroacetic acid was followed by ammonia-catalyzed cy-
clization to provide 19 in 73% yield over the three steps. The
1
4c
glycals.
The three substrates 14, 15, and 16 had been synthesized by
utilization of the reverse prenylation protocol discussed above.
In each case, oxidative cyclization17 via DMDO did occur
apparently in high yield, as indicated by rough purification and
NMR analyses of the reaction products (Scheme 2). However,
NMR analyses of the reaction products indicated the formation
of a ca. 1:1 mixture of diastereomers. This result is per se not
surprising, since the results of Kametani and Ottenheijm
(15) Kametani and co-workers achieved a 2:1 selectivity in their oxidative
cyclization by using singlet oxygen for the synthesis of brevianamide E.
The major product was the desired natural product. (a) Kametani, T.;
Kanaya, N.; Ihara, M. J. Am. Chem. Soc. 1980, 102, 3974. (b) Kametani,
T.; Kanaya, N.; Ihara, M. J. Chem. Soc., Perkin Trans. 1 1981, 959. (c)
For a report of a 3:2 selectivity ratio in the same reaction, see: Sanz-Cervera,
J. F.; Glinka, T.; Williams, R. M. Tetrahedron, 1993, 49, 8471. (d)
Ottenheijm et al. obtained a 1:1 ratio of diastereomers in their approach
toward sporidesmin: Plate, R.; Akkerman, M. A. J.; Ottenheijm, H. C. J.
J. Chem. Soc., Perkin Trans. 1 1987, 2481.
suggested that, at best, only very modest stereoselectivity is
achieved in this kind of cyclization15 (see, for example, the
synthesis of brevianamide E below).
To accomplish a hetero coupling, in formation of the
diketopiperazine moiety, the protecting groups embodied in
(
13) Ohno, M.; Spande, T. F.; Witkop, B. J. Am. Chem. Soc. 1968, 90,
521.
14) (a) Nakagawa, M.; Watanabe, H.; Kodato, S.; Okajima, Y.; Hino,
T.; Flippen, J. L.; Witkop, B. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 4730.
b) Nakagawa, M.; Yokoyama, Y.; Kato, S.; Hino, T. Tetrahedron 1985,
1, 2125. (c) Danishefsky, S. J.; Bilodeau, M. T. Angew. Chem., Int. Ed.
Engl. 1996, 35, 1380.
(16) (a) Murray, R. W. Chem. ReV. 1989, 89, 1187. (b) Adam, W.; Curci,
R.; Edward, J. O. Acc. Chem. Res. 1989, 22, 205. (c) Adam, W.;
Hadjiarapoglou, L. Top. Curr. Chem. 1993, 164, 45.
(17) For the reaction of Na-acylated indoles with dimethyldioxirane,
see: (a) Zhang, X.; Foote, C. S. J. Am. Chem. Soc. 1993, 115, 8867. (b)
Adam, W.; Ahrweiler, M.; Peters, K.; Schmiedeskamp, B. J. Org. Chem.
1994, 59, 2733 and references therein.
6
(
(
4

11968 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Schkeryantz et al.
Scheme 4. One-Pot Synthesis of Diketopiperazine 19 from
Scheme 5a
1
7
synthesis became especially concise following the finding that
diketopiperazine (19) can be prepared, albeit only in 30%
yield, by pyrolysis of 17 at 140 °C. This capability saved
four steps relative to the synthetic sequence described above
(Scheme 4).
At this stage in the synthesis, we hoped to effect a culminating
one-step oxidative conversion of a diketopiperazine to gypsetin
1). Indeed, we would attempt the reaction on 19 itself, wherein
(
the indoline amino group was unprotected. As noted above, an
approach involving stereochemically random oxidative cycliza-
tion, would benefit from the truism that the hypothetical ratio
of gypsetin:“syn:syn” product:“anti:anti” product would be 2:1:
a
2 2
Reagents and conditions: (a) dimethyldioxyrane (4 equiv), CH Cl /
acetone, -78 f 0 °C.
Scheme 6a
1
. In the event, reaction of 19 with 4 equiv of DMDO afforded
a 40% isolated yield of fully synthetic gypsetin (1). This was
accompanied by the formation of two isomers, referred to as
2
0 and 21, in 18 and 20% yields, respectively. At the present
time, we favor the “syn:syn” structure for 20 and the “anti:
anti” arrangement in 21, though these assignments are not
definitive. When conducted in large scale (>0.5 g), this
cyclization produces essentially the same ratio of products but
in only ca. 65% combined yield.
The spectral properties ( H NMR, 13C NMR, MS, and IR)
of synthetic gypsetin (1) were the same as those of the natural
material. The chromatographic properties were also identical,
and the melting point (159 °C) was in accord with that reported
for the naturally derived material. The optical rotation of
synthetic gypsetin was [R]24D -113.4° (c 0.20, CHCl3) was in
1
2
4
good agreement with that reported for natural gypsetin, [R]
116.9° (c 0.14, CHCl3).
We briefly investigated the use of other potential “oxenium
D
-
ion” equivalents to achieve two-fold cyclization of 19. In
practice, successful oxidative cyclization on this substrate turned
out to be restricted to the set of conditions we tried first. Thus,
attempts to utilize mCPBA, Pb(OAc)4, N-phenylselenophtha-
lomide, t-BuOCl, or Br2 led to substantially decomposed starting
material with no indication of formation of either 20, 21, or 1
a
Reagents and conditions: (a) Boc-proline (1.1 equiv), BOP-Cl (3
Cl , - 78 f 0 °C, 1 h. (b) TFA, CH Cl , room temperature,
h then NH , MeOH, reflux, 12 h, 52% from 17. (c) Dimethyldioxirane
4 equiv), CH Cl /acetone, -78 f 0 °C, 76% for both 25 and 26.
equiv), CH
2
2
2
2
2
(
3
2
2
(
Scheme 5). Whether these failures reflect a particular nuance
cyclization, afforded deoxybrevianamide E (24) in 52% yield
from 17. To our knowledge, this is the most concise and efficient
total synthesis of deoxybrevianamide E of those which have
been reported.
Since compound 24 had previously been converted to 25, a
formal total synthesis of brevianamide E had also, in effect,
been accomplished. Nonetheless, another opportunity to study
the DMDO-induced oxidative cyclization of indolyl-substituted
diketopiperazines had presented itself. Hence, the matter was
pursued for purposes of learning more about this reaction.
Addition of 4 equiv of DMDO to deoxybrevianamide E (24)
of substrate 19, or whether they are general, was not determined.
Synthesis of the Brevianamides. At this point, it seemed
probable that the reactions devised for the synthesis of gypsetin
could be applicable toward the synthesis of other reverse prenyl-
containing alkaloids, such as deoxybrevianamide (24) and
brevianamide E (25). Deoxybrevianamide E is a metabolite of
19
Aspergillus ustus first isolated by Steyn, while brevianamide
E had been isolated from the culture medium of Penicillium
20
breVicompactum by Birch and Wright. Their syntheses had
been the subject of several investigations.1
5a,b,18
Our synthesis
commenced with coupling of the previously described amino
ester 17 to N-BOC-protected L-proline to afford 22. Deprotection
of this compound with TFA (see compound 23), followed by
provided a mixture of compounds in a ratio of ∼5:1 (Scheme
1
6
). Based on H NMR and optical rotation data, the major
product is bis-epi-brevianamide E (26), while the minor product
is brevianamide E (25). This result is in striking contrast to
Kametani’s reported result, wherein it was claimed that subjec-
tion of deoxybrevianamide E to singlet oxygen yielded brevi-
anamide E (25) and bis-epi-brevianamide E (26) in a ratio of
(
(
18) Ritchie, R.; Saxton, J. E. J. Chem. Soc., Chem. Comm. 1975, 611.
19) (a) Steyn, P. S. Tetrahedron Lett. 1971, 3331. (b) Steyn, P. S.
Tetrahedron 1973, 29, 107.
20) (a) Isolation: Birch, A. J.; Wright, J. J. J. Chem. Soc., Chem. Comm.
969, 644. (b) Structure determination: Birch, A. J.; Wright, J. J.
Tetrahedron 1970, 26, 2329.
(
1
1
5
∼2:1.

NoVel Constructions of 2,3-Disubstituted Indoles
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 11969
a
Table 1. Addition of Nucleophiles to 3-Chloroindolenines 30
Figure 6. Proposed pathway of oxidation of deoxybrevianamide E.
By monitoring the consequences of the reaction of DKP (24)
with dimethyldioxirane by thin-layer chromatography, we were
able to note the rapid disappearance of starting material with
formation of a minor and a major component as judged by TLC
“
spot” sizes. Upon continued stirring, the minor product
remained seemingly unchanged. Its spectral characteristics,
following workup, were consistent with those of brevianamide
E (25). By contrast, the major TLC-identified product progressed
to bis-epi-brevianamide E (26) only after prolonged reaction
time (ca. 12 h).
We emphasize that the structure of the precursor of compound
26 has not been established. Attempts to purify this compound
were not successful. A possible explanation for these qualitative
findings is as follows. We postulate that the initial product of
dimethyldioxirane oxidation of 24 is the corresponding epoxide,
which suffers rapid valance isomerization to the stereoisomeric
hydroxyiminium species 27 and 28 (Figure 6). Apparently,
stereoisomer 27 cyclizes very rapidly to give rise to the observed
25. By contrast, the cyclization reaction of presumed 28 is rather
slow, allowing this labile intermediate to accumulate and be
detected by TLC analysis en route to 26.
a
If this line of conjecture has merit, then two conclusions seem
to be required. First, in contrast to the singlet oxygen case of
Kametani and in contrast to the DMDO-mediated oxidations
of diketopiperazine 19 or model compounds 14-16, oxidation
of 24 is significantly stereoselective in favor of 28. Furthermore,
it would be necessary to presume that cyclization of 28 is
considerably slower than is the case with hypothetical 27.
Conceivably, this differential rate could be the consequence of
hindrance to cyclization en route to 26. Such a cyclization
requires forcing the prolyl acyl portion of the diketopiperazine
into the endo face of the “cup-shaped” pyrroloindole structure.
These matters are the subject of continuing investigations.
Synthesis of 2,3-Disubstituted Indoles. Given the methodol-
ogy which was implemented for the installation of reverse prenyl
groups at the C2-indolic position of substituted tryptophans, we
wondered about extensions of the general concept. The thought
was that a chloroindolenine of the type 8 should be susceptible
to nucleophilic attack by a range of other nucleophiles. As in
the case of reverse prenylation, attack would be likely to occur
at C2, with expulsion of chloride ion. Eventual tautomerization
would lead to a 2,3-disubstituted indole. As a probe structure
for this purpose, we utilized the readily available N,N-diben-
zyltryptophan methyl ester (29). The tertiary nature of the
nitrogen precluded cyclization of the required chloroindolenine
intermediate. Compound 29 is more conveniently prepared than
is 11 and avoids the problem of competitive electrophilicity from
Typical procedure: tert-butylhypochlorite (1.2 equiv) was added
to a cold (-78 °C) solution of starting tryptophan (1 equiv) and Et
3
N
(
1.2 equiv) in 1 mL of THF. After 30 min, the nucleophile was added
2.0 equiv), followed by BF O (2.0 equiv). The solution was then
(
3
‚Et
2
b
allowed to warm to ambient temperature over 12 h. A promoter was
not used with this nucleophile.
the imide linkage. In the event, compound 29 was treated in
several protocols with tert-butylhypochlorite to generate the
presumed chloroindolenine derivative, 30. Reaction of 30 with
prenyl 9-BBN as described before afforded 31. Surprisingly, a
range of other nucleophiles did not react well with the presumed
30 in the absence of an additional Lewis acid catalyst. However,
it was found that, in the presence of BF3‚OEt2, nucleophilic
attack became more general. We first investigated this reaction
with allyl trimethylsilane. Curiously, we were unable to obtain
the 2-allylated isomer. However, through the use of tri(n-butyl)
allylstannane, BF3‚OEt2, a 79% yield of compound 32 was
obtained.
Other nucleophiles surveyed in this preliminary inquiry (see
Table 1) resulted in the formation of products 33-39 as shown.
As seen, trimethylsilyl cyanide reacted in good yield to afford
the C2-cyanotryptophan derivative. Trimethlsilyl azide, on the
other hand, afforded a complex mixture of unidentified products.
A significant extension of the method arose from the finding
that the silyl keteneacetals and enamines reacted in serviceable
yields to afford the ester and ketones, respectively (see

11970 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Schkeryantz et al.
Scheme 7
Scheme 8
42a
43
8
40
41
42b
44
12
complementary to the Gribble method, wherein suitably N-
protected, 3-substituted indoles can be used to fashion 2-lithio
45
46
9,22
derivatives. These indolyllithium agents can be alkylated with
appropriate electrophiles, thereby producing 2,3-disubstituted
indoles. The possibility of successfully extending the published
2-lithioindole method in the context of tryptophan congeners
seemed to be remote. Rather, we preferred to explore the
possibility of introducing the C2 group as a nucleophile.
Accordingly, our first approach was directed toward synthe-
sizing a suitable and regioispecific reverse prenylborane nu-
cleophile, generalized as 42. The thought was that such an agent
might react with a chloroindolenine of the type 8. Following
the expected allylic transposition of the nucleophilic moiety, a
prenyl group would appear at C2 of the indole (43). The earliest
attempts along these lines involved reactions of prenyltri(n-
butyl)stannane with 9-BBN bromide or triflate. We hoped to
generate the 9-BBN version of the reverse prenyl borane
derivative (see 42b). The latter might undergo an “allylic
rollover”, thereby delivering its prenyl product (see 43).
Preliminary attempts to reduce this method to practice were not
satisfying. One interpretation of those early experiments was
that 42 (BR2 ) borabicyclononane) was, in fact, generated.
However, its allylic transposition leading to 44 is quite rapid,
even in the absence of reaction with an electrophile. Species
44 serves to deliver a reverse prenyl function to the chloroin-
dolenine. In practice, reverse prenylation leading to 46 was
seriously competitive with formation of the desired prenylation
product, 45, in ca. a 1:1 ratio. Moreover, the combined yields
of 45 and 46 were low.
compounds 34-36). Disappointingly, 2-trimethylsilyloxyfuran
afforded only a low yield of the butenolide, although it was
interesting that the olefin isomerized into conjugation with the
indole ring rather than remain in conjugation with the ester
moiety. As seen, the lithium salt of phenylacetylene afforded
3
8 in moderate yield. The full scope of this kind of reaction
with other organolithium derivatives remains to be determined.
Thus, in some instances, there is apparent reductive cleavage
of 30 to produce 29. Indeed, indole itself reacted with in situ-
generated 30 to produce the 8a-(3′-indolyl)tryptophan methyl
ester 39 in 68% yield. We emphasize that these reactions have
not been optimized in this survey phase. However, even at this
stage, a promising strategy to synthesize tryptophan derivatives
that are otherwise difficult to access has emerged.
Synthesis of Tryprostatin B. While these studies were well
in progress, we noted a report by Osada and co-workers
describing the isolation of two natural products, tryprostatin A
2
1
(
40) and B (41) from Aspergillus fumigatus (Scheme 7). The
presence of a diketopiperazine sector and the presence of a
prenyl function at the 2-position of the indole are features of
the tryprostatins that did not escape our attention. From a
synthetic point of view, at the level of the indole substructure,
the tryprostatins pose a challenge related to yet different from
that faced in the gypsetin and brevianamide targets. In the
tryprostatin project, we would be concerned with the introduc-
tion of a prenyl group at the C2-position of a suitable tryptophan
derivative. By contrast, a reverse prenyl function was required
in the gypsetin and brevianamide programs. Similarly, the key
challenge in the 5-N-acetylardeemin target7 was that of intro-
ducing a reverse prenyl function at the 3-position of a pyrro-
loindole core. In the tryprostatin case, there is no comparable
issue of pyrroloindole formation. Our interest in the tryprostatin
problem was enhanced by the claims that 40 and 41 are cell
cycle progression inhibitors in ts-ft210 cells at the G2/M phase
In seeking a workable solution to the goal of prenylating
indoles via chloroindolenines, we were much influenced by
23
earlier contributions from the laboratories of Keck, Yama-
c
24
25
26
27
moto, Denmark, Thomas, and Wardell. These critical
background disclosures suggested the possibility of using
reactions of allylic organometallic agents with boron trichloride
to generate transient allyl boron reagents, possibly structures
of the type 42c (R ) Cl, n ) 2) or -ate versions thereof (42d,
R ) Cl, n ) 3). With this subgoal in mind, we rapidly treated
tri(n-butyl)prenylstannane and the chloroindolenine 12 at -78
°C with BCl3. Following workup, an 83% yield of the desired
compound 45 was obtained (Scheme 8). Only traces of undesired
46 could be detected. We speculated that, perhaps, the prenyl-
2
1
barrier.
At the chemical level, we favored a solution which would
adapt the logic already utilized in the gypsetin venture but would
now involve attack of a suitable nucleophilic prenylating agent
onto a chloroindolenine (e.g., 8, Scheme 7). We focused on
tryprostatin B, because we could start with readily available
L-tryptophan. Viewed differently, we sought a methodology
(
23) (a) Keck, G. E.; Abbot, D. E.; Boden, E. P.; Enholm, E. J.
Tetrahedron Lett. 1984, 25, 3927. (b) Keck, G. E.; Andrus, M. B.; Castellino,
S. J. Am. Chem. Soc. 1989, 111, 8136.
(24) Yamamoto, Y.; Maeda, N.; Maruyama, K. J. Chem. Soc., Chem.
(
21) (a) Cui, C.-B.; Kakeya, H.; Okada, G.; Onose, R.; Ubukata, M.;
Commun. 1983, 742.
(25) Denmark, S. E.; Wilson, T.; Willson, T. M. J. Am. Chem. Soc. 1988,
110, 984.
Takahashi, I.; Isono, K.; Osada, H. J. Antibiot. 1995, 48, 1382. (b) Cui,
C.-B.; Kakeya, H.; Osada, G.; Onose, R.; Osada, H. J. Antibiot. 1996, 49,
5
27. (c) Cui, C.-B.; Kakeya, H.; Osada, H. J. Antibiot. 1996, 49, 534.
22) For another preparation of 2-substituted indoles, see: Fukuyama,
T.; Chen, X.; Peng, G. J. Am. Chem. Soc. 1994, 116, 3127.
(26) McNeill, A. H.; Thomas, E. J. Synthesis 1994, 322.
(27) Harston, P.; Wardell, J. L.; Marton, D.; Tagliavini, G.; Smith, P. J.
Inorg. Chim. Acta 1989, 162, 245.
(

NoVel Constructions of 2,3-Disubstituted Indoles
Table 2. Prenylation of Indoles and Ketones
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 11971
Scheme 9a
a
Reagents and conditions: (a) hydrazine hydrate (3.5 equiv), 3:1
Cl (0.1 M), 24 h, 82%; (b) cyanuric fluoride, pyridine,
, -15 °C; (c) CH Cl , NaHCO , H O 94%; (d) TMSI (1.2 equiv),
MeCN, 0 °C; (e) NH /MeOH, 20 h, 67% from 61.
MeOH/CH
CH Cl
2
2
2
2
2
2
3
2
3
Though the matter has not been sorted out in detail, these
conditions apparently triggered significant and unproductive
chemistry at the double bond of the prenyl function. Success,
however, was achieved through the action of TMSI on 61. This
treatment gave rise to the N-deprotected prolyl tryptophan
derivative 62 (Scheme 9). Finally, cyclization was achieved by
reaction of 62 with methanolic ammonia. There was thus
produced tryprostatin B (41). The high-field NMR spectrum of
fully synthetic tryprostatin B was identical to that provided by
stannane reacts with BCl3 to generate, transiently, a reverse
prenyl intermediate of the type 42c or d. Transfer of a prenyl
function through allylic transposition would provide 45. It is
conceivable, though certainly not proven, that prenyl delivery
had occurred via an intramolecular transfer reaction from 47,
in turn generated from 42c and 12 by ligand exchange.
The reaction was extended to two other indole cases. Once
again, the starting 3-substituted indoles 48 and 49 were
converted (by chloroindolenine formation and prenylation at C2)
to 50 and 51, respectively (Table 2). It was also of interest to
probe the applicability of the prenylation method to ketonic
electrophiles. In so doing, we would be addressing the problem
of avoiding the formation of mixtures of prenyl and reverse
3
0
Professor H. Osada. The optical rotation of fully synthetic
27.5
tryprostatin was substantially the same ([R] D -74.6° (c 0.64,
2
7
CHCl3) vs lit. [R] D -71.1° (c 0.63, CHCl3)) as that reported.
The overall yield of tryprostatin B from N-phthalolyl tryptophan
methyl ester was a gratifying 46%.
2
8
prenyl products provided by other allylic metal systems. In
Elsewhere, we have described the results of investigations
7
b
this very preliminary phase, we examined three substrates, 52-
into the biological profile of tryprostatin B. While the
cytotoxicity reported in the original disclosure could be nicely
duplicated in a collaborator’s laboratory, cell proliferation
inhibition by the agent in our assays occurred only at higher
concentrations (50 µg/mL) than those reported. Proper analysis
of the matter was further complicated by the finding of some
shelf instability for tryprostatin B. These findings raised the
question as to whether part of the discrepancy in the assays
may, in fact, be the consequence of transformation products of
41, rather than the compound itself. Clarification of these issues
would require careful side-by-side evaluation of natural and fully
synthetic 41, in addition to transformation products of 41 of as
yet undetermined structure.
5
4. As before, the ketone and tri(n-butyl)prenylstannane at -78
°C were treated with BCl3. This step was followed by quenching
and standard workup. In this way, cyclohexanone (52) itself
gave rise to 55, while its 4,4-dimethyl derivative (53) afforded
5
6. The impact of a preexisting chiral center on this reaction
was examined in only one case. Thus, 2-phenylcyclohexanone
54), when subjected to this protocol, reacted smoothly to
(
provide an 80% yield of 57. No other products were obtained.
At the interpretive level, it can be argued that our protocol
simply generates a de facto equivalent of 42c or 42d which
transfers a prenyl group to the keto group in a bimolecular
reaction. An alternative, more interesting, but certainly unproven
interpretation envisions ligand exchange leading to 58. This step
is followed by intramolecular transfer prenylation.
Summary
With a sound and seemingly broadly applicable method de-
veloped for the nucleophilic prenylation of indoles accom-
plished, we turned our attention to completion of the total
synthesis of tryprostatin B. This program started with hydrazi-
nolysis of 45, thereby giving rise to 59 (Scheme 9). Its coupling
partner was generated by conversion of N-Boc-L-proline to its
N-Boc acid fluoride derivative 60, using cyanuric fluoride,
following the tenets laid down by Carpino et al.29 Coupling of
Several new synthetic methods have been developed and
applied to interesting problems in alkaloid total synthesis. A
key unifying factor in the investigation was the finding that
2
-unsubstituted chloroindolenines can be used, rather broadly,
in reactions with various nucleophiles to generate 2,3-disubsti-
(28) We note recent examples of prenylation via a prenylbarium
species: (a) Yanagisawa, A.; Ogasawara, K.; Yasue, K.; Yamamoto, H. J.
Chem. Soc., Chem. Commun. 1996, 367. (b) Yanagisawa, A.; Habaue, S.;
Yasue, K.; Yamamoto, H. J. Am. Chem. Soc. 1994, 116, 6130.
5
9 and 60 was conducted under essentially Shotten-Baumann
(
29) Carpino, L. A.; Mansour, E.-S. M. E.; Sadat-Aalaee, D. J. Org.
Chem. 1991, 56, 2611.
30) We thank Professor Osada for furnishing us with the spectra of
authentic tryprostatin B.
conditions, giving rise to seco derivative 61.
Attempted acidic cleavage of the Boc function through the
agency of trifluoroacetic acid was surprisingly complicated.
(

11972 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Schkeryantz et al.
J ) 11.2, 16.3 Hz, 1 H), 3.11 (dd, J ) 14.2, 8.6 Hz, 1 H); 13C NMR
tuted indole derivatives. This general concept allowed for
introduction of a reverse prenyl function at C2 in substituted
tryptophans. This chemistry was used in a total synthesis of
gypsetin. The final cyclization step in that total synthesis
involved two-fold DMDO-induced oxidative ring closure of a
diketopiperazine derivative (see 19 f 1, Scheme 5). The
stereochemistry of this induced cyclization reaction seemed to
be governed largely by stochastic considerations. Extension of
the logic of the gypsetin synthesis was attempted in the total
syntheses of deoxybrevianamide E (24) and brevianamide E
(CDCl
3
, 100 MHz) δ 172.9, 139.5, 136.1, 128.8, 128.2, 127.4, 126.9,
1
3
22.7, 121.8, 119.2, 118.7, 112.1, 61.4, 54.7, 51.0, 26.2; IR (CHCl
3
)
-
1
+
3
420, 1729, 1455 cm ; MS (DCI, NH ) m/V 399 (M + H), 268, 198;
27 2 2
HRMS calculated for C26H N O
399.2072, found 399.2065.
Methyl (2S)-3-[2-(1,1-Dimethyl-2-propenyl)-1H-indol-3-yl]-2-(1,3-
dioxo-1,3-dihydro-2H-isoindol-2-yl)propanoate (13). To a cold (-78
3
°C) solution of 11 (1.99 g, 5.70 mmol) and Et N (0.95 mL, 6.84 mmol)
in 20 mL of THF was added tert-butyl hypochlorite (0.83 mL, 6.84
mmol). After the solution was stirred for 0.5 h at -78 °C, a 1.0 M
solution of prenyl-9-BBN12 (11.4 mL, 11.4 mmol) in THF was added
dropwise. The solution was allowed to warm slowly over 6 h to ambient
(25). In the latter case, DMDO-induced cyclization led, in a
moderately stereoselective fashion, to the bis-epi derivative of
the natural agent as the major product (see 26). The chloroin-
dolenine method was generalized to produce a range of 2-sub-
stituted tryptophan congeners, not easily obtained by other
means. Finally, extension of the method involving introduction
of a prenyl function led to a total synthesis of tryprostatin B
2 3
temperature, after which 5 mL of a saturated solution of K CO (aq)
was added. The layers were separated, and the aqueous layer was
extracted with ethyl acetate (3 × 10 mL). The organics were combined
4
and dried (MgSO ), filtered, and concentrated in vacuo. Chromatog-
raphy (10% EtOAc/Hex) of the residue afforded 2.26 g (95%) of pure
25
1
3 as a pale yellow foam, R
f
) 0.70 (50% EtOAc/Hex): [R]
); H NMR (CDCl , 400 MHz) δ 7.93 (s, 1 H), 7.68
m, 2 H), 7.60 (m, 2 H), 7.26 (d, J ) 7.9 Hz, 1 H), 7.12 (d, J ) 8.1
D
-180.8°
1
(c ) 3.9, CHCl
3
3
(
41). Undoubtedly, the concise access provided by this new
(
methodology to novel indole derivatives, including substituted
tryptophans, will find application in other synthetic under-
takings.
Hz, 1 H), 6.9 (t, J ) 7.1 Hz, 1 H), 6.71 (t, J ) 7.1 Hz, 1 H), 6.19 (dd,
J ) 10.5, 17.5 Hz, 1 H), 5.18 (m, 3 H), 3.87 (dd, J ) 3.8, 15.2 Hz, 1
1
3
H), 3.77 (s, 3 H), 3.67 (dd, J ) 11.3, 15.2 Hz, 1 H), 1.57 (s, 6 H);
NMR (CDCl , 100 MHz) δ 169.9, 167.7, 145.8, 140.1, 133.9, 133.8,
31.8, 129.7, 123.2, 121.1, 119.1, 117.7, 112.1, 110.2, 106.2, 53.5,
C
3
1
5
Experimental Section
-
1
2.7, 39.2, 27.6, 27.5, 24.4; IR (CHCl
3
) 3416, 1742, 1713, 1391 cm
;
+
General. All melting points were determined using a Hoover
capillary melting point apparatus and are uncorrected. Reagents and
starting materials were obtained from commercial suppliers and used
without further purification unless otherwise indicated. Unless specified
otherwise, solvents were freshly distilled prior to use: tetrahydrofuran
MS (DCI, NH
for C25
3
) m/V 417 (M + H), 386, 324, 309; HRMS calculated
H
24
N
2
O
4
416.1736, found 416.1733.
Methyl (2S)-2-Amino-3-[2-(1,1-dimethyl-2-propenyl)-1H-indol-
3-yl]propanoate (17). Hydrazine monohydrate (2.44 g, 48.3 mmol,
3.0 equiv) was added to a solution of 13 (5.92 g, 14.2 mmol) in 120
mL of ethanol. The solution was stirred at ambient temperature for 3
days, after which the ethanol and hydrazine were removed in vacuo.
(THF) and diethyl ether were distilled under nitrogen from sodium metal
utilizing benzophenone as an indicator; benzene, toluene, pyridine,
dichloromethane, diisopropylamine, tetramethylethylenediamine (TME-
DA), and chloroform were distilled from powdered calcium hydride;
dimethyl formamide was distilled at reduced pressure from barium
oxide. Flash column chromatography was carried out on silica gel
obtained from EM Science (230-400 mesh). HPLC grade solvents were
used for all chromatography. Analytical thin-layer chromatography
The residue was taken up in 5 mL of H
were separated. The aqueous layer was extracted with EtOAc (3 × 5
mL), and the organics were combined, washed with H O (10 mL) and
brine (10 mL), dried (Na SO ), and concentrated in vacuo. Chroma-
tography (20% EtOAc/Hex, hexamethyldisilazane treated silica) of the
residue afforded 4.06 g (69%) of pure 17 as a pale yellow foam, R
2
O/EtOAc (1:1), and the layers
2
2
4
f
)
(TLC) was conducted on precoated plates: silica gel 60 F-254, 0.25
0.20 (50% EtOAc/Hex). This reaction was also run on a 5.66 mmol
mm thickness, manufactured by E. Merck & Co., Germany. Air- and/
or moisture-sensitive reactions were carried out under an atmosphere
of dry nitrogen in oven-dried glassware equipped with a tightly fitted
rubber septum. All syringes and needles were oven-dried and transferred
to a desiccator for cooling immediately before use. A Bruker AM-400
spectrometer was used to obtain proton (400 MHz) and carbon (100
MHz) nuclear magnetic resonance spectra in deuteriochloroform unless
otherwise indicated. The chemical shifts are reported as δ values in
scale for only 12 h. An identical workup and purification procedure
afforded 17 in 56% yield: [R]25
12.4° (c 1.9, CHCl
); H NMR
1
D
3
(CDCl , 400 MHz) δ 8.04 (s, 1 H), 7.57 (d, J ) 7.7 Hz, 1 H), 7.31 (m,
3
1 H), 7.11 (m, 2 H), 6.17 (dd, J ) 5.1, 10.2, 1 H), 5.20 (m, 3 H), 3.90
(m, 1 H), 3.69 (s, 3 H), 3.36 (dd, J ) 5.0, 14.5 Hz, 1 H), 3.09 (dd, J
) 9.4, 14.4 Hz, 1 H), 1.62 (br s, 2 H), 1.58 (s, 6 H); 13C NMR (CDCl
,
3
100 MHz) δ 175.8, 145.9, 140.4, 134.1, 129.7, 121.5, 119.3, 118.5,
112.1, 110.4, 106.8, 55.7, 51.9, 39.1, 31.1, 27.8, 27.7; IR (CHCl ) 3405,
) m/V 287 (M + H) , 272, 198; HRMS
287.1760, found 287.1771.
3
1
13
-1
+
ppm relative to deuteriochloroform (7.26 for H, 77.0 for C).
1730, 1202 cm ; MS (DCI, NH
calculated for C17
3
1
3
J-Modulated spin-echo Fourier transform (JMOD) C NMR experi-
ments are reported as (+) for CH and CH or (-) for CH and C and
23 2 2
H N O
3
2
Methyl (2S)-2-[(tert-Butoxycarbonyl)amino]-3-[2-(1,1-dimethyl-
2-propenyl)-1H-indol-3-yl]propanoate (14). Di-tert-butyl dicarbonate
(2.14 g, 9.80 mmol) was added to a solution of 17 (1.4 g, 4.90 mmol)
in 25 mL of THF. The solution was stirred at ambient temperature for
are used as an alternative to off-resonance decoupling. Mass spectra
were determined on a Finnigan spectrometer. Reactions at 0 °C were
carried out in an ice/water bath. Reactions at -78 °C were carried out
in a dry ice/acetone bath. The term “concentrated in vacuo” refers to
the removal of volatile materials on a rotary evaporator at reduced
pressure (30-80 Torr).
2 h, after which H O (25 mL) was added. The aqueous layer was
2
extracted with EtOAc (3 × 25 mL). The organics were washed with
H O (10 mL) and brine (10 mL), dried (Na SO ), and concentrated in
2
2
4
Methyl (2S)-2-(Dibenzylamino)-3-[1H-indol-3-yl]propanoate (29).
Benzyl bromide (2.4 g, 13.7 mmol) was added dropwise to a cool (0
C) solution of tryptophan methyl ester hydrochloride (1.0 g, 6.2 mmol)
vacuo. Chromatography (25% EtOAc/Hex) of the residue afforded 1.43
25
g (76%) of pure 14 as white solid, R
f
) 0.30 (25% EtOAc/Hex): [R]
D
1
°
11.3° (c ) 0.8, CHCl
3
); H NMR (CDCl , 400 MHz) δ 8.03 (s, 1 H),
3
and H u¨ nig’s base (2.8 g, 3.83 mL, 21.8 mmol) in 10 mL of CH Cl
2
2
.
7.49 (d, J ) 7.4 Hz, 1 H), 7.27 (m, 1 H), 7.09 (m, 2 H), 6.14 (dd, J )
10.5, 17.4 Hz, 1 H), 5.20 (m, 2 H), 5.09 (d, J ) 7.4 Hz, 1 H), 4.57
(app. q, J ) 7.5 Hz, 1 H), 3.57 (s, 3 H), 3.30 (m, 1 H), 3.24 (m, 1 H),
The ice-bath was removed, and the solution was stirred for 48 h, after
which water (20 mL) was added. The layers were separated, and the
aqueous layer was extracted with EtOAc (3 × 25 mL). The organics
1.56 (s, 3 H), 1.55 (s, 3 H), 1.53 (s, 9 H); 13C NMR (CDCl
, 100 MHz)
3
were combined and dried (MgSO
4
), filtered, and concentrated in vacuo.
δ (rotamers) 173.5, 155.0, 146.7, 146.0, 140.5, 134.2, 129.7, 121.5,
Chromatography (10 f 25% EtOAc/Hex) of the residue afforded 1.6
119.4, 118.2, 112.2, 110.4, 105.6, 85.2, 79.6, 54.5, 52.1, 39.1, 28.5,
2
5
-1
g (65%) of pure 29, R
f
) 0.65 (25% EtOAc/Hex): [R]
D
-104.4° (c
28.2, 27.7, 27.4; IR (CHCl
3
) 3379, 1699, 1366, 1167 cm ; MS (DCI,
1
+
)
3
3.0, CHCl ); H NMR (CDCl , 400 MHz) δ 7.91 (s, 1 H), 7.23-
3
3
NH ) m/V 387 (M + H) , 369, 331; HRMS calculated for C22
H N O
30 2 4
7.35 (m, 11 H), 7.13-7.17 (m, 2 H), 6.97 (dd, J ) 7.1, 7.9 Hz, 1 H),
6.90 (d, J ) 2.1 Hz, 1 H), 4.04 (d J ) 13.9 Hz, 2 H), 3.83 (dd, J )
5.6, 9.0 Hz, 1 H), 3.71 (s, 3 H), 3.57 (d, J ) 13.9 Hz, 2 H), 3.41 (dd,
386.2209, found 386.2206.
(2S)-2-[(tert-Butoxycarbonyl)amino]-3-[2-(1,1-dimethyl-2-prope-
nyl)-1H-indol-3-yl]propanoic Acid (18). Solid LiOH monohydrate

NoVel Constructions of 2,3-Disubstituted Indoles
0.77 g, 18.5 mmol) was added to a solution of 14 (1.43 g, 3.7 mmol)
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 11973
(
13a,15a,16,16a-octahydroindolo[3′′′,2′′′:4′′,5′′]pyrrolo[1′′,2′′:4′,5′]-
pyrazino[2′,1′:5,1]pyrrolo[2,3-b]indole-7,15(5H,7aH)-dione (Gypsetin,
1). After chromatography, the residue was triturated with hexane to
in 38 mL of a THF/MeOH/H O (8/1/1) solution. The solution was
2
stirred at ambient temperature for 3 h, after which 1 N HCl was added
to neutralize the solution. The aqueous layer was extracted with ether
afford 163 mg (32%) of pure gypsetin as a white solid, R
EtOAc/Hex, plate eluted twice): [R]25
-113.4° (c ) 0.2, CHCl
NMR (CDCl , 400 MHz) δ 7.28 (d, J ) 7.3 Hz, 1 H), 7.21 (d, J ) 7.5
f
) 0.27 (15%
1
(
3 × 50 mL). The organics were washed with H
brine (25 mL), dried (Na SO ), and concentrated in vacuo to afford
.38 g (100%) of pure 18 as a white foam, R ) 0.26 (75% EtOAc/
); H NMR (CDCl , 400 MHz) δ
.95 (m, 1 H), 7.60 (m, 1 H), 7.27 (d, J ) 7.5 Hz, 1 H), 7.10 (m, 2 H),
.15 (dd, J ) 10.5, 17.4 Hz, 1 H), 5.21 (m, 2 H), 5.18 (m, 1 H), 4.61
2
O (2 × 25 mL) and
D
3
); H
2
4
3
1
f
Hz, 1 H), 7.13 (app. q, J ) 7.4 Hz, 2 H), 6.79 (m, 2 H), 6.73 (d, J )
7.9 Hz, 1 H), 6.64 (d, J ) 7.9 Hz, 1 H), 6.4 (br s, 1 H), 6.35 (dd, J )
10.8, 17.6 Hz, 1 H), 5.95 (dd, J ) 10.8, 17.6 Hz, 1 H), 5.12 (dd, J )
10.8, 17.7 Hz, 2 H), 4.92 (m, 2 H), 4.04 (d, J ) 10.9 Hz, 1 H), 3.59
(dd, J ) 7.4, 11.0 Hz, 1 H), 3.35 (d, J ) 13.7 Hz, 1 H), 2.71 (m, 2 H),
2.51 (dd, J ) 7.4, 13.1 Hz, 1 H), 2.25 (br s, 1 H), 2.06 (br s, 1 H),
2
5
1
Hex): [R]
7
6
D
4.5° (c ) 1.15, CHCl
3
3
(
m, 1 H), 3.44 (m, 1 H), 3.23 (dd, J ) 9.6, 14.7 Hz, 1 H), 1.57 (s, 6
H), 1.49, 0.94 (singlets, rotomers, 9 H); IR (CHCl ) 3384, 1716, 1699,
) m/V 373 (M + H), 333, 317, 257,
372.2049, found 372.2050.
3S,6S)-3,6-Bis{[2-(1,1-dimethyl-2-propenyl)-1H-indol-3-yl]methyl}-
,5-piperazinedione (Pre-Gypsetin Diketopiperazine) (19). Bis-(2-
3
-
1
+
13
1
1
367, 1165 cm ; MS (DCI, NH
98; HRMS calculated for C21
3
1.33 (s, 3 H), 1.31 (s, 3 H), 0.67 (s, 3 H), 0.60 (s, 3 H); C NMR
H
28
N O
2 4
3
(CDCl , 100 MHz) δ 170.5 (-), 168.1 (-), 148.6 (-), 148.4 (-), 144.2
(
(+), 144.1 (+), 130.8 (+), 130.7 (+), 130.2 (-), 130.1 (-), 124.6
(+), 123.6 (+), 120.5 (+), 120.0 (+), 113.4 (-), 112.9 (-), 111.3
(+), 111.2 (+), 93.72 (-), 90.9 (-), 88.8 (-), 87.8 (-), 61.2 (+),
59.2 (+), 44.7 (-), 44.0 (-), 35.8 (-), 35.1 (-), 27.2 (+), 25.9 (+),
2
oxo-3-oxazolidinyl)phosphonic chloride (BOP-Cl) (2.92 g, 11.1 mmol)
was added to a cool (0 °C) solution of Et N (1.49 g, 2.06 mL, 14.8
3
mmol), acid 18 (1.34 g, 3.70 mmol), and amine 17 (1.16 g, 4.07 mmol)
in 18.5 mL of THF. The solution was stirred at ambient temperature
23.1 (+), 22.0 (+); IR (CHCl
3
) 3365, 1672, 1609, 1468, 1366, 1109,
-
1
+
1090, 894 cm ; MS (DCI, NH
HRMS calculated for C32
3
) m/V 541 (M + H) , 471, 388, 279;
4
for 12 h, after which H
extracted with EtOAc (3 × 25 mL). The organics were combined and
washed with H O (50 mL) and brine (50 mL), dried (Na SO ), and
concentrated in vacuo to afford a light yellow solid. The solid was
suspended in 6 mL of CH Cl . The suspension was cooled to 0 °C,
2
O (50 mL) was added. The aqueous layer was
H
37
N
4
O
541.2815, found 541.2818.
The third compound to elute was the trans,trans isomer, (5aS,7aS,-
8aR,13aS,15aS,16aR)-5a,13a-bis(1,1-dimethyl-2-propenyl)-8a,16a-
dihydroxy-5a,8,8a,13,13a,15a,16,16a-octahydroindolo[3′′′,2′′′:4′′,5′′]-
pyrrolo-[1′′,2′′:4′,5′]pyrazino[2′,1′:5,1]pyrrolo[2,3-b]indole-
7,15(5H,7aH)-dione (21). After chromatography, the residue was
triturated with hexane to afford 83 mg (16%) of pure trans,trans isomer
2
2
4
2
2
and 34 mL of TFA was added. The ice bath was removed and the
clear orange solution allowed to warm to ambient temperature. After
the solution was stirred for 1 h, the volatiles were removed in vacuo,
21 as a white solid, R
[R]25
-400° (c ) 0.2, CHCl ); H NMR (CDCl , 400 MHz) δ 7.23
3 3
f
) 0.20 (15% EtOAc/Hex, plate eluted twice):
1
a 7 M solution of NH
3
in MeOH (50 mL) was added, and the solution
D
brought to reflux. The reaction was complete after 3 h. It was best to
(d, J ) 7.4 Hz, 1 H), 7.12 (t, J ) 7.5 Hz, 1 H), 6.76 (t, J ) 7.5 Hz,
1 H), 6.62 (d, J ) 7.8 Hz, 1 H), 6.36 (dd, J ) 10.8, 7.6 Hz, 1 H), 6.16
(s, 1 H), 5.16 (d, J ) 17.7 Hz, 1 H), 5.09 (d, J ) 10.9 Hz, 1 H), 3.43
(dd, J ) 7.5, 11.2 Hz, 1 H), 2.76 (dd, J ) 11.5, 12.9 Hz, 1 H), 2.62
(dd, J ) 7.6, 13.1 Hz, 1 H), 2.25 (s, 1 H), 1.36 (s, 3 H), 1.13 (s, 3 H);
1
monitor the final cyclization reaction by H NMR. The volatiles were
removed in vacuo, and the residue was dissolved in 25% EtOAC/Hex
and a couple of drops of THF and loaded onto a silica gel column.
Chromatography (25 f 33 f 50% EtOAc/Hex) afforded 0.85 g (73%)
1
3
of pure 19 as a white amorphous solid, R
f
) 0.49 (75% EtOAc/Hex):
); H NMR (CDCl , 400 MHz) δ 8.17
s, 1 H), 7.47 (d, J ) 7.7 Hz, 1 H), 7.27 (d, J ) 7.9 Hz, 1 H), 7.09 (m,
3
C NMR (CDCl , 100 MHz) δ 171.4, 166.0, 148.7, 144.2, 130.9, 129.8,
2
5
1
[R]
D
-39.6° (c ) 1.0, CHCl
3
3
123.7, 120.1, 113.3, 90.0, 89.0, 61.3, 45.1, 35.5, 27.1, 23.0; IR (CHCl )
3
-1
(
3375, 1679, 1608, 1469, 1361, 1318, 1109, 1082, 911 cm ; MS (DCI,
+
2
4
H), 6.10 (dd, J ) 10.4, 17.5 Hz, 1 H), 5.78 (s, 1 H), 5.16 (m, 2 H),
.34 (d, J ) 10.2 Hz, 1 H), 3.70 (dd, J ) 3.2, 14.5 Hz, 1 H), 3.22 (dd,
NH
C
3
) m/V 541 (M + H) , 471, 401, 340, 270; HRMS calculated for
32
H
37
N
4
O
4
541.2815, found 541.2803.
13
3
J ) 11.8, 14.5 Hz, 1 H), 1.54 (s, 3 H), 1.52 (s, 3 H); C NMR (CDCl ,
1
Methyl N-Phthaloyl-2-(3-methyl-2-butenyl)-L-tryptophan (45). To
00 MHz) δ 167.2, 145.6, 141.6, 134.2, 128.7, 122.0, 120.0, 118.0,
a 0 °C solution of 11 (101 mg, 0.290 mmol) and Et N (0.040 mL,
3
1
12.4, 110.8, 104.4, 54.8, 39.0, 29.9, 28.0, 27.8; MS (FAB) m/V 509
0.290 mmol) in CH Cl (2.9 mL) was added tert-butylhypochlorite
2
2
+
(M + H) , 386, 373, 315, 301, 287; HRMS calculated for C32
H
37
N
4
O
2
(0.696 mL, 0.5 M in CCl ) via syringe pump over a 20 min period.
4
5
09.2917, found 509.2913.
The solution was then chilled to -78 °C. Tri(n-butyl)prenylstannane
(0.423 mL, 1.16 mmol) was added, followed by rapid addition of BCl3
(0.580 mL, 1.0 M in CH Cl ). After 3 min, the solution was poured
Gypsetin (1). An acetone solution of dimethyldioxirane (DMDO)
(
53 mL, ∼3.9 mmol, ∼4 equiv) was added to a cold (-78 °C) solution
2
2
of diketopiperazine 19 (500 mg, 0.98 mmol) in 20 mL of THF. The
resultant yellow solution was warmed to -30 °C over 1 h, and the
volatiles were removed in vacuo. The residue was dissolved in 50 mL
of THF and stirred for 24 h at ambient temperature. The volatiles were
again removed in vacuo. Chromatography (TLC grade silica, 15%
EtOAc/Hex) of the residue afforded three compounds.
3 2 2
into saturated NaHCO to quench and extracted with CH Cl . The
organic layer was washed successively with saturated NaHCO and
3
saturated KF. Both layers were filtered through Celite with CH Cl .
2
2
The organic layer was separated and washed again with saturated KF,
dried (Na SO ), filtered, concentrated, and chromatographed (silica gel,
2
4
95/5 to 60/40 Hex/EtOAc) to give a yellow-green amorphous solid
weighing 100.1 mg (83%): R ) 0.64 (60/40 Hex/EtOAc); [R] -253°
); H NMR (400 MHz, CDCl ) δ 7.73 (m, 2 H), 7.71
The first compound to elute was the cis,cis isomer, (5aR,7aS,8aS,-
f
D
1
1
3aR,15aS,16aS)-5a,13a-bis(1,1-dimethyl-2-propenyl)-8a,16a-dihy-
droxy-5a,8,8a,13,13a,15a,16,16a-octahydroindolo[3′′′,2′′′:4′′,5′′]pyrrolo-
1′′,2′′:4′,5′]pyrazino[2′,1′:5,1]pyrrolo[2,3-b]indole-7,15(5H,7aH)-
dione (20). After chromatography, the residue was triturated with
hexane to afford 75 mg (15%) of pure 20 as a white solid, R ) 0.36
246.5° (c ) 0.2, CHCl );
, 400 MHz) δ 7.17 (d, J ) 7.4, 1 H), 6.96 (t, J ) 6.7
Hz, 1 H), 6.71 (t, J ) 7.3 Hz, 1 H), 6.31 (dd, J ) 10.8, 17.6 Hz, 1 H),
.05 (d, J ) 7.8 Hz, 1 H), 5.90 (br s, 1 H), 5.09 (dd, J ) 10.9, 17.7
Hz, 1 H), 4.22 (d, J ) 10.4 Hz, 1 H), 3.42 (d, J ) 13.7 Hz, 1 H), 2.55
dd, J ) 10.5, 13.6 Hz, 1 H), 1.85 (s, 1 H), 1.28 (s, 3 H), 1.22 (s, 3 H);
(c ) 4.0, CHCl
3
3
(br d, J ) 1.2 Hz, 1 H), 7.64 (m, 2 H), 7.48 (d, J ) 7.6 Hz, 1 H), 7.17
(dd, J ) 8.0, 0.8 Hz, 1 H), 7.01 (ddd, J ) 8.0, 7.2, 1.2 Hz, 1 H), 6.95
(ddd, J ) 7.6, 7.2, 1.2 Hz, 1 H), 5.20 (dd, J ) 10.2, 5.4 Hz, 1 H), 5.12
(tquin, J ) 7.2, 1.4 Hz, 1 H), 3.79 (s, 3 H), 3.69 (dd, J ) 14.8, 5.4 Hz,
[
f
25
(
15% EtOAc/Hex, plate eluted twice): [R]
D
3
1 H), 3.63 (dd, J ) 14.8, 10.4 Hz, 1 H), 3.40 (ddd, ν ) 52 Hz, J )
1
13
H NMR (CDCl
3
16.4, 7.2 Hz, 2 H), 1.67 (s, 6 H); C NMR (75 MHz, CDCl
3
) δ 169.6,
167.4, 135.7, 135.0, 134.7, 133.8, 131.7, 128.4, 123.2, 121.0, 120.1,
119.2, 117.7, 110.3, 105.8, 52.7, 52.4, 25.6, 24.9, 24.0, 17.7; FTIR
(film) 3398, 3058, 3028, 2954, 2916, 2852, 1774, 1743, 1714, 1615,
6
(
1462, 1438, 1390, 1255, 1206, 1116, 1100, 1028, 920, 868, 746, 719,
13
+
3
C NMR (CDCl , 100 MHz) δ 161.4, 142.5, 139.5, 126.2, 124.4, 119.3,
666, 608, 581, 530; HRMS (FAB) calculated for C25
24 2 4
H N O M
1
3
14.6, 108.2, 106.2, 89.0, 83.5, 54.5, 40.3, 29.7, 21.2, 18.1; IR (CHCl
3
)
)
416.1736, found 416.1730.
-
1
372, 1663, 1611, 1469, 1375, 1320, 1091, 913 cm ; MS (DCI, NH
3
Ethyl 2-(3-Methyl-2-butenyl)-3-indoleacetate (50). To a -78 °C
+
m/V 541 (M + H), 471, 401, 331, 313, 261; HRMS calculated for
solution of ethyl 3-indoleacetate (48, 149 mg, 0.733 mmol) and Et N
3
C
32
H N O
37 4 4
541.2815, found 541.2803.
The second compound to elute was (5aR,7aS,8aR,13aS,15aS,16aS)-
a,13a-bis(1,1-dimethyl-2-propenyl)-8a,16a-dihydroxy-5a,8,8a,13,-
(0.102 mL, 0.733 mmol) in CH Cl (7.3 mL) under argon was added
2
2
via syringe pump over a 20 min period a tert-butyl hypochlorite solution
(1.8 mL, 0.5 M in CCl ). Tri(n-butyl)prenylstannane (1.07 mL, 2.93
5
4

11974 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Schkeryantz et al.
mmol) was then added, followed by rapid addition of BCl
1.45 mL, 1.0 M in CH Cl ), which produced a deep orange solution
color. After 3 min, the reaction was poured into saturated NaHCO
and extracted with CH Cl . The organic layer was washed with saturated
NaHCO and then with saturated KF, filtered over Celite and then
washed again with saturated KF, dried (Na SO ), filtered, concentrated,
and chromatographed on silica gel (90/10 to 70/30 Hex/EtOAc) to
recover a clear oil (130 mg, 65%), which turns a bright yellow color
upon exposure to air. The product was stored neat at -20 °C under
3
solution
and chromatographed (95/5 to 80/20 Hex/EtOAc) to recover a clear
(
2
2
oil (276.2 mg, 88%), R
MHz, CDCl ) δ 5.25 (tquin, J ) 8.1, 1.5 Hz, 1 H), 2.15 (d, J ) 7.8
Hz, 1 H), 1.75 (s, 3 H), 1.65 (s, 3 H), 1.53-1.44 (m, 4 H), 1.20-1.18
(m, 4 H), 0.93 (s, 3 H), 0.88 (s, 3 H); 13C NMR (125 MHz, CDCl
) δ
135.0, 119.0, 71.4, 40.7, 34.8, 33.2, 30.8, 29.5, 26.0, 25.3, 17.9; FTIR
(CHCl ) 3694, 3604, 3010, 2932, 2855, 1668, 1602, 1453, 1378, 1365,
f
) 0.40 (90/10 Hex/EtOAc): 1H NMR (500
3
3
2
2
3
3
2
4
3
1260, 1235, 1174, 1118, 1103, 1065, 1048, 990, 967, 879, 849; HRMS
+
(FAB) calculated for C13
H
24O M 196.1827, found 196.1820.
argon to prevent further air oxidation, R
H NMR (400 MHz, C D
6 6
7
f
) 0.67 (70/30 Hex/EtOAc):
) δ 7.79 (dd, J ) 7.0, 1.6 Hz, 1 H), 7.24-
.18 (m, 2 H, H5), 7.01-6.99 (m, 2 H, NH), 5.18 (dquin, J ) 7.1 1.3
Hz, 1 H), 3.88 (q, J ) 7.2 Hz, 2 H), 3.64 (s, 3 H), 3.29 (d, 2 H), 1.61
br d, J ) 1.2 Hz, 3 H), 1.52 (s, 3 H), 0.87 (t, J ) 7.2 Hz, 3 H); 13
NMR (125 MHz, C ); δ 171.5, 135.6, 135.5, 133.6, 129.4, 121.5,
21.1, 119.9, 118.9, 110.8, 104.7, 60.4, 30.7, 25.6, 25.5, 17.7, 14.2;
FTIR (C ) 3434, 3059, 2980, 2930, 1916, 1879, 1733, 1587, 1568,
461, 1443, 1368, 1342, 1300, 1263, 1239, 1158, 1098, 1036, 1014,
(()-trans-1-(3-Methyl-2-butenyl)-2-phenylcyclohexan-1-ol (57).
To a -78 °C solution of 2-phenylcyclohexanone (100 mg, 0.574 mmol)
and tri(n-butyl)prenylstannane (0.419 mL, 2.30 mmol) in CH Cl (5.7
1
2
2
mL) was added rapidly BCl solution (0.631 mL, 1.0 M in CH Cl ).
3
2
2
(
C
After 5 min, the reaction was poured into saturated NaHCO3 and
extracted with CH Cl (two times). The combined organic layers were
6
D
6
2
2
1
2 4
washed with saturated KF (two times), dried (Na SO ), filtered,
concentrated, and chromatographed (95/5 to 80/20 hexane/EtOAc) to
D
6 6
1
9
2
recover a clear oil (112.6 mg, 80%), R ) 0.46 (90/10 Hex/EtOAc):
f
+
1
10, 741; HRMS (FAB) calculated for C17
71.1564.
H21NO
2
M 271.1572, found
H NMR (500 MHz, CDCl ) δ 7.29-7.23 (m, 4 H), 7.20-7.15 (m, 1
3
H), 5.07 (tquin, J ) 7.7, 1.5 Hz), 2.52 (dd, J ) 12.9, 3.5 Hz, 1 H),
2.03 (dq, J ) 13.1, 3.6 Hz, 1 H), 1.92-1.81 (AB-dd, J ) 18.8, 7.6
Hz, 2 H), 1.79-1.74 (m, 1 H), 1.73-1.67 (m, 2 H, H5), 1.65 (br d, J
) 1.1 Hz, 3 H), 1.60-1.55 (m, 2 H, H3e), 1.40 (s, 3 H, H4), 1.40 (m,
2
-(3-Methyl-2-butenyl)-O-tert-butyldiphenylsilyltryptophol (51).
To a -48 °C solution of tryptophol TBDPS ether (49, 204 mg, 0.512
mmol) and Et N (0.071 mL, 0.512 mmol) in CH Cl (5.0 mL) under
argon was added over a 20 min period via syringe pump a tert-butyl
hypochlorite solution (1.23 mL, 0.5 M in CCl ). The solution was
chilled to -78 °C, and tri(n-butyl)prenylstannane (0.784 mL, 2.05
mmol) was then added, followed by rapid addition of BCl solution
1.00 mL, 1.0 M in CH Cl ). After 3 min, the reaction was poured into
saturated NaHCO and extracted with CH Cl (two times). The organic
layers were combined and washed with saturated NaHCO and then
with saturated KF. The emulsion thus formed was filtered over Celite,
rinsed with CH Cl , washed again with saturated KF, dried (Na SO ),
filtered, concentrated, and chromatographed on silica gel (100/0 to 80/
0 Hex/EtOAc) to recover a clear viscous oil (195 mg, 81%) which
turns bright yellow on exposure to air or upon standing for extended
periods of time (1-2 h) in halogenated solvents, R ) 0.62 (80/20 Hex/
) δ 7.80 (m, 5 H), 7.54 (m, 2 H),
.47 (m, 4 H), 7.39 (d, J ) 7.7 Hz, 1 H), 7.34 (d, J ) 8.0 Hz, 1 H),
3
2
2
1 H), 1.31 (qt, J ) 13.2, 3.7 Hz, 1 H); NOE (H2′) H2, H4a, 4’t-CH
3
,
H6a; NOE (H2) H2′, H3e, H6a; 13C NMR (125 MHz, CDCl
) δ 143.2,
4
3
134.9, 129.1, 128.0, 126.3, 119.1, 73.3, 51.9, 40.3, 37.0, 29.2, 26.5,
26.1, 21.6, 17.9; FTIR (film) 3563, 3488, 3084, 3060, 3025, 2931, 2856,
2727, 2669, 1946, 1879, 1808, 1729, 1602, 1492, 1446, 1376, 1271,
3
(
2
2
3
2
2
1234, 1184, 1131, 1088, 1070, 1032, 979, 958, 894, 860, 830, 765,
+
3
702, 516; HRMS (FAB) calculated for C17
244.1827.
H
24O M 244.1829, found
2
2
2
4
Methyl 2-(3-Methyl-2-butenyl)-L-tryptophan (59). To a solution
of 45 (437 mg, 1.05 mmol) in 3:1 MeOH/CH Cl (7.9:2.6 mL) was
2
2
2
added hydrazine hydrate 0.178 mL, 3.67 mmol). The flask was capped
and the solution stirred at ambient temperature for 24 h, during which
time a white precipitate formed. The solution was poured into water
(50 mL) and extracted with CH Cl (4 × 10 mL). The combined organic
f
1
EtOAc): H NMR (400 MHz, CDCl
3
2
2
7
7
5
2 4
extracts were dried (Na SO ), filtered, concentrated, and chromato-
.21 (dt, J ) 6.9, 1.1 Hz, 1 H), 7.12 (dt, J ) 6.9, 6.9, 1.0 Hz, 1 H),
.36 (b t, J ) 7.3 Ha, 1 H), 4.00 (t, J ) 7.8, 7.6 Hz, 1 H), 3.48 (d, J
graphed (silica gel treated with (TMS) NH, 90/10 to 20/80 Hex/EtOAc)
2
to yield 59 as an opaque syrup weighing 248 mg (83%), R ) 0.37
f
1
)
7.3 Hz, 1 H), 3.15 (t, J ) 7.8, 7.6 Hz), 1.89 (s, 3 H), 1.85 (s, 3 H),
(95/5 CHCl /MeOH): [R]D +9.5° (c ) 1.94, CHCl ); H NMR (400
3
3
1
3
1
1
1
3
1
1
.22 (s, 9 H); C NMR (125 MHz, CDCl
3
) δ 135.4, 135.0, 134.9,
MHz, CDCl ) δ 7.94 (br s, 1 H), 7.54 (d, J ) 7.2 Hz, 1 H), 7.27 (d, J
3
34.2, 134.0, 129.5, 128.9, 127.6, 120.8, 120.5, 119.0, 118.2, 110.2,
07.4, 64.4, 27.7, 26.9, 25.7, 25.0, 19.1, 17.8; FTIR (film) 3411, 3070,
051, 2930, 2857, 2737, 1960, 1888, 1827, 1667, 1621, 1589, 1568,
487, 1462, 1428, 1384, 1361, 1338, 1307, 1285, 1244, 1210, 1189,
) 7.8 Hz, 1 H), 7.10 (m, 2 H), 5.30 (tquin, J ) 7.6, 1.4 Hz, 1 H), 3.82
(dd, J ) 8.2, 5.0 Hz, 1 H), 3.71 (s, 3 H), 3.49 (d, J ) 7.2 Hz, 2 H,),
3.25 (dd, J ) 14.4, 5.0 Hz, 1 H), 2.98 (dd, J ) 14.4, 8.2 Hz, 1 H),
1.79 (s, 3 H), 1.76 (s, 3 H); 13C NMR (75 MHz, CDCl ) δ 175.7, 135.9,
3
111, 1065, 1030, 1010, 940, 823, 739, 702, 614, 505; HRMS (FAB)
135.2, 134.7, 128.7, 121.1, 120.2, 119.3, 119.1, 110.4, 106.2, 55.3,
51.9, 30.0, 25.7, 26.1, 17.9; FTIR (film) 3387, 3180, 3055, 2950, 2925,
2854, 1736, 1578, 1461, 1438, 1376, 1343, 1308, 1285, 1202, 1175,
1098, 1034, 1010, 916, 845, 742; HRMS (FAB) calculated for
+
calculated for C31
H
37NOSi M 467.2644, found 467.2642.
1
-(3-Methyl-2-butenyl)cyclohexan-1-ol (55). To a -78 °C solution
of cyclohexanone (215 mg, 2.19 mmol) and tri(n-butyl)prenylstannane
1.6 mL, 8.8 mmol) in CH Cl (14.6 mL) was added rapidly BCl
solution (2.4 mL, 1.0 M in CH Cl ). After 5 min, the reaction was
poured into saturated NaHCO and extracted with CH Cl (two times).
The combined organic layers were washed with saturated KF (two
times), dried (Na SO ), filtered, concentrated, and chromatographed
95/5 to 80/20 Hex/EtOAc) to recover a clear oil (287 mg, 78%), R
(
2
2
3
C
17
H
22
N
2
O
2
286.1681, found 286.1688.
2
2
Methyl N-tert-Butoxycarbonyl-L-proline-L-2′-(3-methyl-2-bute-
3
2
2
nyl)tryptophan (61). To a -30 °C solution of N-Boc-L-proline (279
mg, 1.30 mmol) and pyridine (0.104 mL, 1.30 mmol) in CH Cl (13
2
2
2
4
mL) was added cyanuric fluoride (0.350 mL, 3.89 mmol). The solution
was stirred for 1 h between -30 and -5 °C and then poured onto ice.
The emulsion was extracted quickly with CH Cl (3 × 15 mL). The
(
0
f
)
1
3
.39 (90/10 Hex/EtOAc): H NMR (500 MHz, CDCl ) δ 5.22 (tquin,
2
2
J ) 8.1, 1.5 Hz, 1 H), 2.13 (d, J ) 8.1 Hz, 1 H), 1.73 (d, J ) 1.1 Hz,
H), 1.62 (s, 3 H), 1.60-1.35 (m, 10 H), 1.28-1.18 (m, 1 H); 13
NMR (125 MHz, CDCl ) δ 135.2, 118.9, 71.8, 40.8, 37.3, 26.05, 25.8,
2 4
combined organic layers were dried (Na SO ), filtered, and concen-
3
C
2 2
trated to afford crude 60. The concentrate was redissolved in CH Cl
3
(2 × 5 mL) and transferred to a vigorously stirring biphasic solution
2
1
9
1
2.2, 17.9; FTIR (film) 3385, 2933, 2858, 2727, 2666, 1672, 1448,
of 59 (248 mg, 0.864 mmol) in CH Cl (10 mL) containing NaHCO
2
2
3
376, 1356, 1318, 1297, 1264, 1169, 1148, 1115, 1067, 1036, 972,
(145 mg, 1.73 mmol) in water (10 mL). The reaction was stirred for
10 h before the layers were separated. The aqueous layer was washed
with CH Cl (2 × 5 mL). The combined organic layers were dried
+
14, 888, 845, 779, 726; HRMS (FAB) calculated for C11
68.1514, found 168.1519.
H20O M
2
2
1
-(3-Methyl-2-butenyl)-4,4-dimethylcyclohexan-1-ol (56). To a
78 °C solution of 4,4-dimethylcyclohexanone (201 mg, 1.58 mmol)
Cl (16
). After
and extracted
(two times). The combined organic layers were washed
with saturated KF (two times), dried (Na SO ), filtered, concentrated
2 4
(Na SO ), filtered, concentrated, and chromatographed (silica gel, 85/
-
15 to 30/70 Hex/EtOAc) to give 391 mg (94%) of 61 as a white
amorphous solid, R -38.3° (c )
) 0.19 (60/40 Hex/EtOAc): [R]30
); H NMR (500 MHz, CDCl , 50 °C) δ 7.80 (br s, 1 H),
and tri(n-butyl)prenylstannane (1.16 mL, 6.34 mmol) in CH
2
2
f
D
1
mL) was added rapidly BCl
min, the reaction was poured into saturated NaHCO
with CH Cl
3
solution (1.7 mL, 1.0 M in CH
Cl
2 2
2.0, CHCl
3
3
5
3
7.47 (d, J ) 7.4 Hz, 1 H), 7.24 (m, 1 H), 7.09 (m, 2 H), 5.31 (tt, 1 H),
4.82 (dd, 1 H), 4.18 (br s, 1 H), 3.64 (s, 3 H), 3.46 (d, 1 H), 3.30 (m,
1 H), 3.21 (ddd, ν ) 32 Hz, J ) 14.7, 6.8 Hz, 2 H), 3.12 (br s, 1 H),
2
2
2
4

NoVel Constructions of 2,3-Disubstituted Indoles
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 11975
1
1
.94 (br s, 1H), 1.80 (s, 3 H), 1.77 (s, 3 H), 1.66 (bm, 1 H), 1.48 (m,
1114, 1099, 1012, 919, 753, 666; HRMS (FAB) calculated for
1
3
+
H), 1.41 (s, 9 H), 1.27 (m, 1 H); C NMR (75 MHz, CDCl
3
, 50 °C)
C
21
H
26
N
3
O
2
(M + H) 352.2025, found 352.2021.
δ 172.3, 172.2, 154.9, 135.9, 135.2, 134.8, 129.2, 121.2, 120.1, 119.6,
1
2
1
17.7, 110.4, 105.3, 80.3, 60.7, 53.0, 52.0, 46.9, 28.2, 27.2, 25.6, 25.1,
3.7, 17.8; FTIR (film) 3393, 3314, 2975, 2928, 2880, 1743, 1676,
Acknowledgment. This work was supported by the NIH
Grants No. CA 28824 and HL25848). J.C.G.W. gratefully
(
512, 1458, 1439, 1391, 1366, 1247, 1215, 1163, 1124, 744; HRMS
acknowledges the NIH for a postdoctoral fellowship (HL09187).
The authors thank Professor A. Endo of Tokyo Noko University
for an authentic sample of gypsetin (1). K.M.D. gratefully
acknowledges a Lilly predoctoral fellowship. George Sukenick
(SKI NMR Core Facility, CA-08748) is gratefully acknowledged
for mass spectral analyses. We gratefully acknowledge Vinka
Parmakovich and Barbara Sporer of the Columbia Mass Spectral
Facility and separately to Dr. H. Osada of the Institute of
Physical and Chemical Research (RIKEN) for copies of NMR
spectral data of tryprostatin B.
+
(FAB) calculated for C27
H N
37 3
O
5
M
483.2733, found 483.2730.
Tryprostatin B (41). To a 0 °C solution of 61 (50.0 mg, 0.103
mmol) in freshly distilled MeCN (2.0 mL) under argon was added TMSI
0.018 mL, 0.126 mmol). After 45 min, the reaction was quenched
(
with a few drops of methanol. The solvent was removed, and the
orange-brown concentrate (crude 62) was redissolved in ammonia-
saturated methanol. The light yellow solution was stirred for 20 h,
concentrated, and chromatographed on (TMS)
2
NH-treated silica gel (90/
1
0 to 0/100 Hex/EtOAc) to obtain 24.1 mg (67%) of tryprostatin B
2
7.5
(
41) as a white amorphous solid: [R]
D
-74.6° (c ) 0.64, CHCl
) δ 7.94 (s, 1 H), 7.48 (d, J ) 7.7 Hz, 1
H), 7.31 (d, J ) 8.0 Hz, 1 H), 7.16 (ddd, J ) 8.1, 7.1, 1.0 Hz, 1 H),
3
);
1
H NMR (500 MHz, CDCl
3
Supporting Information Available: Experimental details
relative to the syntheses of deoxybrevianamide E and brevi-
anamide E (24-26), as well as experimental details relative to
the syntheses and characterizations of 2,3-disubstituted indoles,
listed in Table 1 (31-39) (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
7
4
1
.10 (ddd, J ) 7.9, 7.1, 0.9 Hz, 1 H), 5.61 (br s, 1 H), 5.32 (m, 1 H),
.37 (dd, J ) 11.0, 2.6 Hz, 1 H), 4.06 (t, J ) 8.0 Hz, 1 H), 3.69 (m,
H), 3.67 (m, 1 H), 3.59 (ddd, J ) 9.3, 8.7, 3.3 Hz, 1 H), 3.49 (m, 2
H), 2.96 (dd, J ) 15.1, 11.4 Hz, 1 H), 2.34 (m, 1 H), 2.03 (m, 2 H),
13
1
.91 (m, 1 H), 1.79 (s, 3 H), 1.75 (s, 3 H); C NMR (125 MHz, CDCl
δ 169.3, 165.8, 136.4, 135.4, 135.4, 127.9, 121.8, 119.9, 119.7, 117.7,
10.8, 104.6, 59.2, 54.5, 45.4, 45.4, 28.3, 25.7, 25.6, 25.1, 22.6, 18.0;
FTIR (film) 3295, 2959, 2925, 2854, 1666, 1461, 1429, 1305, 1216,
3
)
1
JA9925249