Biomimetic synthesis of the antimalarial flindersial alkaloids

Article abstract of DOI:10.1021/ja301387k

A biomimetic strategy for the synthesis of the antimalarial flindersial alkaloids is described. Flinderoles A, B, and C, desmethylflinderole C, isoborreverine, and dimethylisoborreverine were all synthesized in three steps from tryptamine. The key step is an acid-promoted dimerization of the natural product borrerine. This approach is thought to mirror the biosynthesis of these compounds.

Full text of DOI:10.1021/ja301387k

Communication
pubs.acs.org/JACS
Biomimetic Synthesis of the Antimalarial Flindersial Alkaloids
Ravikrishna Vallakati and Jeremy A. May*
Department of Chemistry, University of Houston, 136 Fleming Building, Houston, Texas 77204-5003, United States
S
* Supporting Information
all of the bioactive members of the family. Additionally, these
studies shed light on the probable biosynthetic pathway for
these alkaloids.
ABSTRACT: A biomimetic strategy for the synthesis of
the antimalarial flindersial alkaloids is described. Flinder-
oles A, B, and C, desmethylflinderole C, isoborreverine,
and dimethylisoborreverine were all synthesized in three
steps from tryptamine. The key step is an acid-promoted
dimerization of the natural product borrerine. This
approach is thought to mirror the biosynthesis of these
compounds.
The report of the structure of the flinderoles and their
bioactivity quickly drew synthetic interest, culminating in two
strategies consisting of 146 and 197 synthetic steps. Concurrent
with the initiation of those efforts, we became interested in
synthesizing the flinderoles and their constitutional isomers, the
borreverines.8 Studies of the likely biosyntheses of borreverine
and isoborreverine were conducted in the 1970s.9 These
alkaloids have an obvious biosynthetic origin in naturally
occurring borrerine (1). The proposed mechanism of their
formation begins with acidic opening of borrerine to an
equilibrium mixture of iminium 9 and indolodiene 10, which is
followed by Diels−Alder cycloaddition (Scheme 1).10 Sub-
istorically, successful malaria treatments have been
H
derived from natural products.1 Emerging malarial drug
resistance compels additional investigations in this area, and
natural product research remains a promising ground for
finding new treatments.2 In a recent disclosure of the isolation
and structural determination of flinderole B (3) and C (5)
(Figure 1), Avery and co-workers reported significant
Scheme 1. Biosynthesis of Borreverines
sequent N1 or C3 attack by the indole of 10 at C12′ in 9 would
produce isoborreverine or borreverine, respectively. Exper-
imentally, treatment of borrerine with trifluoroacetic acid
(TFA) at 65 °C was reported by Koch to produce a 50:50
mixture of isoborreverine and borreverine in 80% yield.8
Prolonged exposure to acid was said to produce a greater
percentage of the former product, indicating that it is favored
thermodynamically. Together with the observation that these
compounds naturally occur as racemates, these studies strongly
supported a biosynthesis of the borreverines from borrerine.
We hypothesized that the flinderole skeleton could also arise
from a dimerization of borrerine through reaction of
indolodiene 10 with imine 9 or 11. However, this would
proceed through a formal [3+2] cycloaddition pathway in lieu
Figure 1. The flindersial alkaloids.
antimalarial activity for the alkaloids 2, 3, and 5−7 of the
plant genus Flindersia.3 A follow-up paper reported more
detailed biological activities and suggested that these alkaloids
interrupt parasitic hemoglobin metabolism through a different
mechanism than that of chloroquine.4 Consequently, these
alkaloids show parasitic toxicity even in chloroquine-resistant
strains of Plasmodium falciparum. To gain access to these
compounds and generate structural analogues in significant
amounts for further testing, we have developed a biomimetic
synthesis5 of alkaloids 2−7 in three synthetic steps from
tryptamine. This route is significantly shorter than reported
strategies for synthesizing the flinderoles and allows access to
Received: February 10, 2012
Published: April 10, 2012
© 2012 American Chemical Society
6936
dx.doi.org/10.1021/ja301387k | J. Am. Chem. Soc. 2012, 134, 6936−6939

Journal of the American Chemical Society
Communication
of the initial [4+2] cycloaddition assumed for the borreverines
(Scheme 2). Further support for this possibility was found in
Table 1. Treatment of 1 with TFA
Scheme 2. Proposed Flinderole Biosynthesis
a
entry TFA equiv
solvent
T (°C)
time
2 h
2:4:6:8
b
1
2
3
4
5
6
7
8
9
1.0
2.0
2.0
3.0
3.0
6.0
6.0
2.0
2.0
benzene
benzene
benzene
benzene
benzene
benzene
benzene
CHCl3
65
0:0:0:0
65
40 min
80 min
2 h
41:30:28:1
38:33:29:0
34:32:34:0
28:28:43:1
16:13:34:0
17:17:21:0
55
40
65
30 min
24 h
b
b
0−22
40
30 min
30 min
48 h
b
45
28:28:7:1
none
65
0:0:100:0
a
1
Determined by integration of H NMR peaks. The combined yields
b
were all above 95%. The balance of the material was 1.
Dethe’s recent report of a biosynthesis-inspired [3+2] cyclo-
addition between two strategically protected alkene-appended
indoles.6 We began to wonder whether the flinderoles had
actually been produced in the biosynthetic study of the
borreverines9 but were overlooked as the flinderoles’ structures
were not known at the time.
To test whether flinderole formation occurs alongside
borreverine production, borrerine was first synthesized
(Scheme 3). This was readily accomplished using Sakai’s
integration indicated that it was produced in only trace
amounts. In retrospect, this is unsurprising, as the stereo-
chemistry of borreverine would have to arise from a less-
favorable exo-Diels−Alder reaction in the initial step (Scheme
1). This observation may have relevance to the alkaloids’
biosynthesis, as Avery and co-workers did not observe any
borreverine in the Flindersia species that produced the
flinderoles.3 Changing the solvent to chloroform gave the
flinderoles as a majority of the crude mixture (entry 8). The use
of tetrahydrofuran (THF), dioxane, or N,N-dimethylformamide
(DMF) as the solvent resulted in no reaction. Interestingly, the
omission of solvent resulted in exclusive formation of
isoborreverine.
Scheme 3. Synthesis of Borrerine (1)
A variety of Brønsted and Lewis acids were also introduced
to the reaction (Table 2). Many of these acids either failed to
promote the reaction or caused decomposition of the material
to mixtures of unidentifiable products (entries 1−5). When
dimerization cleanly occurred, isoborreverine was almost
exclusively produced (entries 6−9). The use of BF3·OEt2 for
an extended time led to the formation of only isoborreverine
Table 2. Effects of Acids on Dimerization
approach.11 In fact, by optimizing this sequence, we obtained 1
in 77% overall yield from tryptamine in two steps.12
The synthetic borrerine was treated with 2 equiv of TFA in
benzene at 65 °C for 40 min (Table 1, entry 2) and was
consumed, as observed by thin-layer chromatography. Two new
closely spaced product spots appeared, and 1H NMR analysis of
the crude mixture showed evidence of isoborreverine as one of
the major products. The remaining 1H NMR peaks
corresponded to the spectra of flinderole A and desmethyl-
flinderole C. In fact, the flinderole diastereomers appeared to be
produced in roughly the same amount as isoborreverine! Contrary
to the previous report, only trace amounts of borreverine (8)
were observed. While isoborreverine was readily isolated via
chromatography, the separation of the flinderoles required the
use of reversed-phase HPLC. This purification allowed us to
confirm the alkaloid identities and the ratios of their formation
a
b
entry
acid
TfOH
solvent
T (°C)
time
2:4:6:8
c
1
benzene
benzene
benzene
benzene
benzene
benzene
benzene
benzene
CH2Cl2
methanol
none
0−22
60
40 min
40 min
60 min
40 min
120 min
40 min
40 min
60 min
3 days
2 h
0:0:0:0
2
TfOH
decomp.
c
3
pTsOH
0−22
60
0:0:0:0
c
4
TMSOTf
AlCl3
0:0:0:0
c
5
0−22
60
0:0:0:0
c
6
AlCl3
0:0:40:0
0:0:38:0
0:0:50:0
c
c
7
Sc(OTf)3
Tf2O
60
8
0
9
BF3·OEt2
1 N HCl
22
0:0:100:0
d
10
11
55
43:26:30:1
e
c
CH3COOH
55
18 h
38:33:0:0
1
as observed in the H NMR spectra of the crude mixtures.
a
b
In an effort to reproduce the original report of the formation
Unless otherwise noted, 2 equiv of acid was used. Determined by
integration of H NMR peaks. The combined yields were all above
95%. The balance of the material was 1. 6 equiv of acid. 150 equiv
of borreverine,13 a range of TFA equivalents, reaction times,
1
1
c
d
e
and reaction temperatures were screened (entries 1−7). H
NMR peaks associated with 8 were occasionally observed, but
of acid.
6937
dx.doi.org/10.1021/ja301387k | J. Am. Chem. Soc. 2012, 134, 6936−6939

Journal of the American Chemical Society
Communication
with complete conversion (entry 9). Methanolic HCl, however,
provided the flinderoles as the major product (entry 10).
Impressively, neat acetic acid produced only the flinderoles
(entry 11)! This selectivity is particularly notable given that
neat TFA favors isoborreverine formation.
60 mg yield of dimethylisoborreverine was obtained as the sole
product from the reaction of 100 mg of borrerine with 1 equiv
of MeOTf and then 2 equiv of TFA for 35 min at room
temperature.12 When this latter reaction was run for less time,
20 mg of dimethylisoborreverine, 13 mg of flinderole C, and 13
mg of flinderole B were obtained.
In conclusion, a biomimetic synthesis of all of the
antimalarial flindersial alkaloids has been established. Variations
in the reaction conditions allow for the acid-promoted
dimerization of borrerine to produce isoborreverine or the
flinderoles selectively. We have demonstrated that the outcome
of this reaction is different from that previously reported. A
novel strategy utilizing an initial methylation followed by acid
treatment allows for the synthesis of the flindersial alkaloids
appended with tertiary amines. Remarkably, the sequence
allows selective access to these alkaloids in only three synthetic
steps. Furthermore, the use of tryptamine analogues, alternate
α,β-unsaturated aldehydes, and a variety of acylating agents will
enable the synthesis of analogues of isoborreverine and the
flinderoles.
Thus, conditions for selective formation of either the
flinderoles or isoborreverine were found. Apparently, only a
few acids promote just enough reactivity to form the
flinderoles. Other acids are either too weak to initiate the
reaction or are so reactive that the flinderoles are bypassed for
the more thermodynamically stable isoborreverine.7 8 In fact,
b,
prolonged reaction times generally produce greater percentages
of isoborreverine. Furthermore, isolated samples of the
flinderoles are converted to isoborreverine in strong acid,
illustrating the preference for that structure under thermody-
namic conditions.
As flinderole B (3), flinderole C (5), and dimethylisoborre-
verine (7) exhibit the strongest antimalarial activity,4 these were
next targeted for synthesis. While reductive amination to add an
additional methyl group to the secondary amine of the isolated
products 2, 4, and 6 is certainly feasible, a more direct approach
would be to add the methyl concomitantly with borrerine
dimerization. Thus, we postulated that borrerine could be
exposed to a methylating agent to generate an ammonium salt.
Upon treatment with acid and an increase in the temperature,
ring opening would form intermediates similar to 9−11 but
having a pendant tertiary amine. Dimerization could then
directly provide 7 (Scheme 4) as well as 3 and 5.
ASSOCIATED CONTENT
* Supporting Information
■
S
Additional optimization data, experimental procedures, and
new characterization data for all compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
jmay@uh.edu
■
Scheme 4. Direct Methylative Dimerization
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the Welch Foundation (Grant E-1744) and the
University of Houston for financial support in conducting this
research. We are also grateful to the Gilbertson lab for the use
of their preparative HPLC instrument with a reversed-phase
column and Dr. Anton Agarkov for HPLC assistance.
REFERENCES
■
(1) (a) Midiwo, J. O.; Clough, J. M. In Aspects of African Biodiversity,
Proceedings of the Pan Africa Chemistry Network Biodiversity
Conference, Nairobi, Kenya, Sept 10−12, 2008; Midiwo, J. O.,
Clough, J. M., Eds.; Royal Society of Chemistry: Cambridge, U.K.,
2009; pp 11−19. (b) Kumar, N.; Sharma, M.; Rawat, D. S. Curr. Med.
Chem. 2011, 18, 3889. (c) Fattorusso, E.; Taglialatela-Scafati, O.
Phytochem. Rev 2010, 9, 515. (d) Wells, T. N. Malar. J. 2011, 10
(Suppl. 1), S3.
Indeed, when borrerine was first treated with methyl triflate
in chloroform at 0 °C and then treated with 2 equiv of TFA and
warmed to 22 °C, the dimethylamine alkaloids were formed
directly in a combined 70% yield. Dimethylisoborreverine
constituted 30% of the mixture, and flinderoles B and C were
present as 21 and 19% of the mixture, respectively.12,14 A longer
exposure to TFA afforded dimethylisoborreverine as the only
product. These results may have implications for the biosyn-
thesis of flinderoles B and C, which were isolated from the
same plant species as dimethylisoborreverine.3b In that species,
bisindole alkaloid synthesis may be initiated by methylation.
To demonstrate that this biomimetic synthesis can practically
provide access to these compounds, the most selective
conditions were repeated on a larger scale to generate isolable
quantities of each alkaloid. For example, when 50 mg of
borrerine was treated with BF3·OEt2 in benzene for 3 days, 20
mg of pure isoborreverine was isolated. Similarly, 28 mg of
flinderole A and desmethylflinderole C was obtained from the
reaction of 50 mg borrerine with TFA in benzene at 65 °C. A
(2) (a) de Azevedo Calderon, L.; Silva-Jardim, I.; Zuliani, J. P. J. Braz.
Chem. Soc. 2009, 20, 1011. (b) Guantai, E.; Chibale, K. Malar. J. 2011,
10, S2. (c) Ginsburg, H.; Deharo, E. Malar. J. 2011, 10, S1.
(d) Nogueira, C. R.; Lopes, L. M. X. Molecules 2011, 16, 2146.
(3) (a) Fernandez, L. S.; Jobling, M. F.; Andrews, K. T.; Avery, V. M.
Phytother. Res. 2008, 22, 1409. (b) Fernandez, L. S.; Buchanan, M. S.;
Carroll, A. R.; Feng, Y. J.; Quinn, R. J.; Avery, V. M. Org. Lett. 2009,
11, 329.
(4) Fernandez, L. S.; Sykes, M. L.; Andrews, K. T.; Avery, V. M. Int. J.
Antimicrob. Agents 2010, 36, 275.
(5) (a) Nguyen, H.; Ma, G.; Fremgen, T.; Gladysheva, T.; Romo, D.
J. Org. Chem. 2011, 76, 2. (b) Al-Mourabit, A.; Zancanella, M.; Tilvi,
S.; Romo, D. Nat. Prod. Rep. 2011, 28, 1229. (c) Li, C.; Johnson, R. P.;
Porco, J. A., Jr. J. Am. Chem. Soc. 2003, 125, 5095. (d) Roche, S. P.;
Cencic, R.; Pelletier, J.; Porco, J. A., Jr. Angew. Chem., Int. Ed. 2010, 49,
6938
dx.doi.org/10.1021/ja301387k | J. Am. Chem. Soc. 2012, 134, 6936−6939

Journal of the American Chemical Society
Communication
6533. (e) Razzak, M.; De Brabander, J. K. Nat. Chem. Biol. 2011, 7,
865.
(6) Dethe, D. H.; Erande, R. D.; Ranjan, A. J. Am. Chem. Soc. 2011,
133, 2864.
(7) Zeldin, R. M.; Toste, F. D. Chem. Sci 2011, 2, 1706.
(8) (a) Pousset, J.; Kerharo, J.; Maynart, G.; Monseur, X.; Cave,
́
A.;
Goutarel, R. Phytochemistry 1973, 12, 2308. (b) Pousset, J.-L.; Cave,
́
A.; Chiaroni, A.; Riche, C. J. Chem. Soc., Chem. Commun. 1977, 261.
(c) Tillequin, F.; Koch, M.; Rabaron, A. J. Nat. Prod. 1985, 48, 120.
́
(9) Tillequin, F.; Koch, M.; Pousset, J.-L.; Cave, A. J. Chem. Soc.,
Chem. Commun. 1978, 826.
(10) Lee, V. Tetrahedron 1996, 52, 9455.
(11) (a) Demaindreville, M.; Levy, J.; Tillequin, F.; Koch, M. J. Nat.
Prod. 1983, 46, 310. (b) Yamanaka, E.; Shibata, N.; Sakai, S.
Heterocycles 1984, 22, 371.
(12) See the Supporting Information for full experimental details,
compound characterization, and representative NMR spectra.
(13) Equivalents of acid and reaction concentration were not
specified in the original procedure.
(14) The balance of the material was made up of a fourth dimeric
component that appears to be the result of initial indole methylation;
complete structural elucidation of this compound is ongoing.
6939
dx.doi.org/10.1021/ja301387k | J. Am. Chem. Soc. 2012, 134, 6936−6939