2-Thioindoles as precursors to spiro-fused indolines: Synthesis of (±)-dehaloperophoramidine

Article abstract of DOI:10.1002/anie.200601278

(Chemical Equation Presented) Closing the ring: The cyclization of a 2-thioindole with an electrophilic substituent serves as the key transformation in a synthetic strategy for the generation of indole alkaloids of the perophoramidine and communesin class. DBU = 1,8-diazabicyclo[5.4.0]undec-7-ene, Ms = methanesulfonyl chloride.

Full text of DOI:10.1002/anie.200601278

Angewandte
Chemie
Natural Product Synthesis
illustrated in Scheme 2.[8] We envisioned that the intramolec-
ular coupling between the 3-position of a 2-thioindole with a
pendant electrophile would serve to generate one of the
perophoramidine spiro-fused rings; the resulting thioimidate
DOI: 10.1002/anie.200601278
2-Thioindoles as Precursors to Spiro-Fused
Indolines: Synthesis of
(ꢀ)-Dehaloperophoramidine**
Amir Sabahi, Alexei Novikov, and Jon D. Rainier*
A large number of complex indole- and indoline-containing
alkaloids have been isolated from the marine environment
over the past decade.[1] Among these compounds is perophor-
amidine, which was isolated from the colonial ascidian
Perophora namei by Ireland and co-workers.[2] The hexacyclic
bisamidine skeleton of perophoramidine (see Scheme 1),
Scheme 2. Retrosynthetic analysis of perophoramidine.
would be nicely set up to undergo a subsequent cyclization
with the pendant amine. The result of this cascade would be
the generation of two rings, five of perophoramidineꢀs six
rings, and 1. Incorporation of the remaining quaternary center
and the A ring would provide perophoramidine. If successful,
we were confident that the strategy outlined in Scheme 2
would not only enable us to efficiently synthesize perophor-
amidine but would also allow us to target other members of
this family (namely, communesin B)[9] along with analogues
for biological evaluation.
Scheme 1. Structures of perophoramidine and communesin B.
Clearly, the success of the strategy outlined in Scheme 2
depended upon the nucleophilicity of 2-thioindoles. As an
indication of their reactivity, we had previously demonstrated
that 2-thioindoles undergo “interrupted” Pictet–Spengler
cyclizations to spiro-fused indolines (namely, 4) when treated
with the appropriate aldehyde in the presence of 4-Š
molecular sieves (MS) at room temperature [Eq. (1)].[10] It
is noteworthy that these conditions are milder than those used
for related reactions of indoles that lack the thioether at the 2-
position.[11]
With this in mind, we set out to investigate whether
reactions related to that outlined in Equation (1) could be
carried out to generate the dehaloperophoramidine skeleton.
The synthesis of the requisite cyclization precursor is illus-
trated in Scheme 3. The coupling of tryptamine or benzyl-
which contains vicinal quaternary centers, was established
with the help of extensive 2D NMR experiments and was later
validated by the total synthesis developed by Fuchs and
Funk.[3] Preliminary biological data indicated that perophor-
amidine is cytotoxic to HMT 116 colon carcinoma cells and is
able to induce apoptosis by poly(adenosine-5’-diphosphate-
ribose)polymerase (PARP) cleavage.[4]
Our interest in the synthetic chemistry of 2-thioindoles,[5–7]
combined with the novel structural features of perophorami-
dine and its preliminary biological data, induced us to initiate
a program to target its synthesis. Described herein are our
preliminary results in this area and the synthesis of (ꢀ )-
dehaloperophoramidine.
We considered several synthetic routes to the perophor-
amidine skeleton, but ultimately settled upon the approach
[*] A. Sabahi, Dr. A. Novikov, Prof. J. D. Rainier
Department of Chemistry
University of Utah
315 South, 1400 East, Salt Lake City, UT 84112 (USA)
Fax: (+1)801-581-8433
E-mail: rainier@chem.utah.edu
[**] This research was supported by NIH GeneralMedicalSciences
(GM 61608) and the University of Utah. The authors thank Dr.
Atta M. Arif, Dr. Charles L. Mayne, and Elliot M. Rachlin for help
with the X-ray analysis, NMR spectroscopy, and mass spectrometry,
respectively. Professor Chris M. Ireland is acknowledged for pro-
viding the spectraldata of dehaol perophoramidine.
Scheme 3. Synthesis of cyclization precursors 10 and 11: a) N-Boc-isatin (7),
THF, RT, 7 h (70%); b) PhSCl, CH2Cl2, 08C!RT, 12 h (97%); c) NaBH4, MeOH,
08C, 0.5 h.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 4317 –4320
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4317

Communications
protected tryptamine with N-tert-butyloxycarbonyl (N-Boc)-
KOtBu as the base. Quenching the resulting ketene acetal
with allyl iodide led to 16 as a single diastereomer in 89%
yield.
protected isatin followed by thioether incorporation and
reduction with NaBH4 provided cyclization precursors 10 and
11, respectively.[12]
Pentacycle 16 proved to be amenable to single-crystal X-
ray analysis, thus firmly establishing that the reaction of 15
had resulted in the desired C4 and C20 diastereomer needed
for the synthesis of perophoramidine (Figure 1).
With a facile route to their synthesis in hand, we were
prepared to examine the ability of 10 and 11 to undergo the
key cyclization reaction. As a further indication of the
nucleophilicity of 2-thioindoles, the desired cyclization occur-
red under the conditions used for the generation of the
mesylate from 11 to give spirocycle 13 as a 1:1 mixture of C4
diastereomers. In contrast to the results with 11, unprotected
amide 10 did not undergo the desired cyclization but instead
underwent competitive decomposition upon extended expo-
sure to the reaction conditions. Presumably, this latter result is
due to the inability of 10 to overcome the propensity of the
amide to exist as its s-trans rotamer.[13]
The fact that 13 existed as a mixture of C4 diastereomers
was inconsequential; treatment of the mixture with 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU) led to equilibration at
C4, and only one of the two diastereomers underwent a
subsequent cyclization to give the C12amidine 15 in 79%
overall yield from 9. Following optimization studies, we found
it best to carry out both transformations in a single flask by
concentrating the reaction mixture containing 13, dissolving
the resulting residue in CH2Cl2, and subjecting the resulting
solution to DBU (Scheme 4).
With pentacycle 15 in hand, we examined the generation
of the C4 quaternary center as a precursor to the A ring
[Eq. (2)]. To achieve this goal, we ultimately settled upon an
N,O-ketene acetal allylation reaction similar to that employed
by Artman and Weinreb in their perophoramidine studies;[14]
although they had used NaH, we found it best to employ
Figure 1. ORTEP drawing of 16 from X-ray crystallographic analysis.
With ready access to 16, all that remained to complete the
synthesis of dehaloperophoramidine was the generation of
the A ring (Scheme 1). We anticipated accomplishing this aim
by converting the olefin in 16 into the corresponding amine,
activating the lactam, and carrying out the requisite cycliza-
tion reaction to the C24 amidine. To accomplish these goals,
we first needed to remove the benzyl group on the lactam.
Although 16 decomposed when subjected to dissolving-metal
conditions, we were able to affect the removal of the benzyl
group when the N11 Boc group was absent. Thus, removal of
the Boc group from 16 and subjecting the resulting compound
to dissolving-metal conditions gave 17 after the selective
formation of the o-nitrobenzensulfonyl (nosyl)-protected
amidine. Interestingly, as evidenced by the downfield shift
of the C15 proton in 17 relative to 16, this latter reaction
resulted in the selective formation of the N13 nosyl derivative.
Oxidative cleavage of the alkene and reductive amination
gave 19. Conversion of the secondary amine into the
Scheme 4. Cyclization to the perophoramidine pentacycle: a) MsCl,
pyridine, 08C, 4 h; b) DBU, CH2Cl2, 08C!RT (R=Bn: 79%, 3 steps;
R=H: 0%). Ms=methanesulfonyl chloride.
4318
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4317 –4320

Angewandte
Chemie
corresponding nosyl derivative gave 21 after activation of the
lactam through its conversion into the corresponding methyl
imidate (Scheme 5).
aforementioned inability to generate the requisite acyclic
amidine precursor.
The significant difference between the successful sub-
strate prepared by Fuchs and Funk (namely, 24) and ours was
the oxidation state of C12; they had employed a C12 aminal
to our amidine. As we believe steric congestion at C24 to be
the problem with imidate 21, we suspected that the success of
Fuchs and Funk with 24 might be related to its ability to
undergo a reversible ring-opening reaction to temporarily
relieve some of the congestion about C24 in the course of its
conversion into 27 (Scheme 6).[16]
Scheme 5. Generation of the perophoramidine A-ring precursor:
a) TFA, 08C!RT, 12 h; b) Li, NH3, THF, ꢁ788C, 10 min; c) Boc2O,
NEt3, THF, 08C!RT, 8 h (69%, 3 steps); d) OsO4, NaIO4, NaOAc,
THF, H2O, 08C, 3 h; e) CH3NH2·HCl, NaOAc, MeOH, RT, 7 h; NaBH4,
08C, 1 h; f) NosCl, NEt3, THF, 08C!RT, 12 h (42%, 3 steps);
g) Me3OBF4, iPr2NEt, CH2Cl2, 08C, 1 h (44%). Nos=nosyl.
Scheme 6. Proposed mechanism for the cyclization to amidine 27
developed by Fuchs and Funk. DMF=dimethylformamide.
In the last critical transformation to dehaloperophorami-
dine, attempts to form the A ring under the conditions
developed by Fuchs and Funk were completely unsuccessful
in our hands.[15] As an indication of its lack of reactivity, we
also were unable to convert the methyl imidate (or the
corresponding imido triflate) in 21 into the corresponding
amidine [Eq. (3)] through its intermolecular coupling with
methyl amine or ammonia.
With this hypothesis driving our efforts, we synthesized
the C12aminal cyclization precursor from the allyl alkylation
adduct 16 (Scheme 7). Oxidative cleavage and reductive
amination gave 28. Incorporation of the Boc group onto the
secondary amine and reduction of the C12amidine gave 30.
Note that reduction of the amidine did not occur during the
reductive amination reaction. Aminal generation at C12and
global removal of the Boc groups gave 31 after reintroduction
of the Boc moiety onto the secondary amine. Reductive
removal of the benzyl group and selective incorporation of
the methyl carbamate at N11 gave 32.[17] Finally, conversion of
the amide into the corresponding methyl imidate with
trimethyloxonium tetrafluoroborate gave the targeted cycli-
zation precursor 33.
Interestingly, when we attempted to generate the azido
substrate that corresponded to 22 as a precursor to a C24
amidine from an aza-Wittig reaction, we isolated O-alkylated
adduct 23 [Eq. (4)]. Our attempts to use a similar reaction to
form an acyclic amidine were unsuccessful as a result of our
With 33 in hand, it remained to remove the Boc group and
induce cyclization. We were pleased to find that exposure of
33 to trifluoroacetic acid (TFA) not only resulted in the
removal of the Boc group but also led to the generation of the
desired C24 amidine in 95% yield [Eq. (5)].
Angew. Chem. Int. Ed. 2006, 45, 4317 –4320
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4319

Communications
Scheme 8. Completion of the preparation of (ꢀ)-dehaloperophorami-
dine: a) KOH, MeOH, H2O, 1008C, 6 h (71%); MnO2, CH2Cl2, RT,
0.5 h (64%).
Received: April 1, 2006
Published online: May 30, 2006
Keywords: communesin · cyclization · natural products ·
.
perophoramidine · thioindoles
[1] For recent reviews, see: a) W. Gul, M. T. Hamann, Life Sci. 2005,
78, 442; b) J. W. Blunt, B. R. Copp, M. H. G. Munro, P. T.
Northcote, M. R. Prinsep, Nat. Prod. Rep. 2006, 23, 26, and
references therein.
[2] S. M. Verbitski, C. L. Mayne, R. A. Davis, G. P. Concepcion,
C. M. Ireland, J. Org. Chem. 2002, 67, 7124.
Scheme 7. Conversion of pentacycle 16 into imidate cyclization precur-
[3] J. R. Fuchs, R. L. Funk, J. Am. Chem. Soc. 2004, 126, 5068.
[4] J.-B. Denault, G. S. Salvesen, Chem. Rev. 2002, 102, 4489.
[5] a) J. D. Rainier, A. R. Kennedy, J. Org. Chem. 2000, 65, 6213;
b) J. D. Rainier, A. R. Kennedy, E. Chase, Tetrahedron Lett.
1999, 40, 6325.
[6] a) A. R. Kennedy, M. H. Taday, J. D. Rainier, Org. Lett. 2001, 3,
2407; b) A. N. Novikov, A. R. Kennedy, J. D. Rainier, J. Org.
Chem. 2003, 68, 993; c) A. M. Nyong, J. D. Rainier, J. Org. Chem.
2005, 70, 746.
sor 33: a) OsO4, NaIO4, NaOAc, THF, H2O, 08C, 3 h (72%);
b) CH3NH2·HCl, NaOAc, MeOH, RT, 7 h; NaBH4, 08C, 1 h; c) Boc2O,
NEt3, THF, 08C!RT, 12 h (82%, 2 steps); d) NaBH4, EtOH, 08C!RT,
18 h (89%); e) TFA, CH2Cl2, 08C!RT, 8 h; f) Boc2O, NEt3, THF (97%,
2 steps); g) Li, NH3, THF, NH4Cl, ꢁ788C; h) ClCO2CH3, pyridine,
CH2Cl2, temp, time (88%, 2 steps); i) Me3OBF4, NaHCO3, CH2Cl2, 08C,
1 h (81%).
[7] For the use of 2-thioindoles to generate indolines with quater-
nary substitution at C3, see: K. S. Feldman, D. B. Vidulova, A. G.
Karatjas, J. Org. Chem. 2005, 70, 6429.
[8] Our attempts to employ thionium ylide rearrangements to the
perophoramidine skeleton were unsuccessful; for examples of
this reaction, see reference [6].
[9] A. Numata, C. Takahashi, Y. Ito, T. Takada, K. Kawai, Y. Usami,
E. Matsumura, M. Imachi, T. Ito, T. Hasegawa, Tetrahedron Lett.
1993, 34, 2355.
[10] A. R. Kennedy, PhD Dissertation, University of Arizona, 2001.
[11] See R. B. Woodward, M. P. Cava, W. D. Ollis, A. Hunger, H. U.
Daeniker, K. Schenker, Tetrahedron 1963, 19, 247.
[12] J. F. M. Da Silva, S. J. Garden, A. C. Pinto, J. Braz. Chem. Soc.
2001, 12, 273.
[13] W. E. Stewart, T. H. Siddall III, Chem. Rev. 1970, 70, 517.
[14] G. D. Artman III, S. M. Weinreb, Org. Lett. 2003, 5, 1523.
[15] We also explored a number of other basic and acidic conditions
without success.
[16] An examination of molecular models of 21 and 24 indicates that
approach of the amine onto the C24 imidate carbon is blocked
by the C18 proton; the formation of 25 allows the aromatic ring
to rotate, thus alleviating this problem.
[17] When the aminal was protected with a Boc group, reductive
removal of the benzyl group resulted in the decomposition of the
perophoramidine ring system.
To complete the synthesis of dehaloperophoramidine, it
remained to remove the methyl carbamate and oxidize the
aminal to the corresponding amidine. This occurred unevent-
fully; exposure of 34 to KOH and MeOH followed by
oxidation with MnO2 under the conditions developed by
Fuchs and Funk gave (ꢀ )-dehaloperophoramidine, whose
spectral data matched the data reported by Ireland and co-
workers (Scheme 8).[2]
In conclusion, we have successfully generated (ꢀ )-deha-
loperophoramidine from a highly efficient spirocyclization
reaction of a readily available 2-thiotryptamine derivative.
We are currently examining this reaction in the synthesis of
other interesting alkaloids, including members of the com-
munesin family. These efforts will be reported in due course.
4320 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4317 –4320