Regioselective intramolecular N1-C3 cyclizations on pyrrole-proline to ABC tricycles of dibromophakellin and ugibohlin

Article abstract of DOI:10.1016/j.tetlet.2004.11.066

Pyrrole N1-C and C3-C regioselective linkage to the fused tricycle ABC system present in the marine metabolites ugibohlin and dibromoisophakellin is described. The cyclization is closely dependent on the electrophilic function, bromination degree of the pyrrole moiety and pH conditions. The mechanism of the functionalization of the ABC olefin by oxidative agents was found to occur through an N-acyliminium intermediate as showed by the natural chemical connection between ugibohlin and dibromoisophakellin.

Full text of DOI:10.1016/j.tetlet.2004.11.066

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 249–252
Regioselective intramolecular N1–C3 cyclizations on
pyrrole–proline to ABC tricycles of dibromophakellin and ugibohlin
a
Nathalie Travert, Marie-Therese Martin,
a
´ `
Marie-Lise Bourguet-Kondrackib and Ali Al-Mourabita,*
aInstitut de Chimie des Substances Naturelles, BP1, CNRS, 91198 Gif-sur-Yvette Cedex, France
´
Laboratoire de Chimie, UMR 5154 CNRS, Museum National d’Histoire Naturelle, 63 rue Buffon, 75005 Paris, France
b
Received 4 August 2004; revised 12 November 2004; accepted 12 November 2004
Available online 26 November 2004
Abstract—Pyrrole N1–C and C3–C regioselective linkage to the fused tricycle ABC system present in the marine metabolites ugiboh-
lin and dibromoisophakellin is described. The cyclization is closely dependent on the electrophilic function, bromination degree of
the pyrrole moiety and pH conditions. The mechanism of the functionalization of the ABC olefin by oxidative agents was found to
occur through an N-acyliminium intermediate as showed by the natural chemical connection between ugibohlin and
dibromoisophakellin.
Ó 2004 Published by Elsevier Ltd.
1. Introduction
Agelasida, Axinellida and Halichondrida. It is currently
admitted that these alkaloids are taxonspecific of at least
Agelasida order and play the role of chemical markers
of these phylogenetically related sponges.5 Pyrrole-2-
aminoimidazole metabolites seem to share a common
biogenetic chemical pathway and are probably derived
from proline and ornithine.6
Marine pyrrole-2-aminoimidazole secondary metabo-
lites, exemplified by dibromophakellin 1,1 ugibohlin 2,2
dibromoisophakellin 3a and monobromoisophakellin
3b,3 (Fig. 1) have attracted the attention of synthetic
organic chemists4 as a result of their challenging struc-
tures and potent biological activities.
In the context of the synthesis of pyrrole 2-aminoimi-
dazole natural alkaloids and their bioactive derivatives,
we are interested in pyrrole N1/C3 nucleophilicity per-
mitting a number of strategic connections as in natural
compounds 1–3. These compounds are characterized
by the presence of the common tricyclic ABC or A0B0C0
moieties (Fig. 1). The selective pyrrole N1–C or C3–C
cyclizations seem to be an important step in the forma-
tion of a number of natural metabolites belonging to the
pyrrole-2-aminoimidazole family. Interesting synthetic
approaches studying the stereoselective access to ABC
tricycle of dibromophakellin have been recently
described.7
More than 70 derivatives have been extracted from dif-
ferent species of marine sponges belonging to the orders
Figure 1. Structures of the marine metabolites containing the N1–C or
C3–C tricyclic ABC or A0B0C0 moieties.
In our earlier communication on the strategy for build-
ing the bicyclic core of longamide,8 we have outlined the
importance of the pyrrolic N1–C and C3–C intramole-
cular cyclization. Further investigations of the 2-
pyrrolecarboxamideacetals reactivity appeared in the
literature.9 The reactivity of the common pyrrole–pro-
line equivalents 4–6 (Scheme 1), their regioselective
Keywords: Dibromoisophakellin; Ugibohlin; Bromopyrrole; N-Acyli-
minium; Bromination; Regioselective cyclization.
*
Corresponding author. Tel.: +33 1 69 82 45 85; fax: +33 1 69 07 72
47; e-mail: ali.almourabit@icsn.cnrs-gif.fr
0040-4039/$ - see front matter Ó 2004 Published by Elsevier Ltd.
doi:10.1016/j.tetlet.2004.11.066

250
N. Travert et al. / Tetrahedron Letters 46 (2005) 249–252
Further investigation showed that when 4 or 5 were trea-
ted under standard conditions like polyphosphoric acid,
or methansulfonic acid, C3–C six-membered ring isomer
12 was not detected at all. Internal Friedel–Crafts acyl-
ation via the acyl chloride of 5 in presence of Lewis acids
did not produce the desired fused 5/6 rings 12. The size
of the cycle and the terminal function of the side chain
seem to play together a determinant role in the ring clo-
sure. To our knowledge, only one example of such cycli-
zation using flash vacuum pyrolysis at 950°C was
described in the literature.10 The reaction of sodium
methylate with 11 in refluxing methanol produced the
methyl ester 4 but not the desired tricyclic product 12.
This result is contrasting with the synthesis of the previ-
ously described homologous pyrroloazepine seven mem-
bers systems.11 It is interesting to note the similarity of
our results with those obtained in indole derivatives12
revising Ro¨derÕs results.13 Therefore we decided to
reduce ester 4 into the corresponding aldehyde 6 and
to investigate its intramolecular cyclization.
Scheme 1. Targeted ABC and A0B0C0 regioisomers from the pseudo-
peptides 4–6.
cyclization to the desired regioisomers 7 and 9 or 8 and
10, the influence of the bromine substitutions on the
cyclization became the focus in our studies.
In this letter we report the clear acid/base-mediated
N1/C3 regioselective cyclization to ABC or A0B0C0 sys-
tems. The functionalization towards the synthesis of
ugibohlin and the participation of the determinant N-
acyliminium in the reactivity of the enamides 9 (Scheme
1) are also discussed.
After some exploratory attempts including reduction of
4 with DIBAH reagent, reduction of acid 5 was accom-
plished through Weinreb intermediate 13 using LiAlH4
at 0°C. Importantly, the reaction led to the pH depen-
dent C–C compounds 1414 and15 by 1M HCl treatment
and to C–N carbinolamine 16, which gave 17 in 72%
yield after treatment with methansulfonic acid (Scheme
3). N1–C regioisomer 16 was also obtained by Lindel
and co-workers.7a
2. Results and discussion
We started our investigation with the preparation of the
pyrrole–proline 4.6b In this case, the commonly used
synthesis of amides from primary amines and the 2-(tri-
chloroacetyl)pyrrole in CH2Cl2, DMF or acetonitrile
did not produce the expected amide. Sterically hindered
secondary amide 4 (Scheme 2) was obtained upon EDCI
condensation of the pyrrole-2-carboxylic acid and L-
proline methyl ester with catalytic amount of DMAP
in good yield.
Believing that compound 15 could be derived from 14 or
vice versa, we have made two complementary experi-
ments from the pure 14 and 15. Treatment of the unex-
pected compound 14 in acidic media did not give the
elimination compound 15 nor did addition of N,O-di-
methylhydroxylamine to 15 in acidic conditions give the
addition compound 14. Thus, the formation of 14 when
For a close examination of the N1–C and C3–C regiose-
lectivity, several trials have been carried out in different
conditions. The subsequent N1–C cyclization of the
ester 46b to pyrazine-1,4-dione derivative 11 was per-
formed most cleanly in THF in presence of NaH
(Scheme 2). The unsymmetrical compound 11 is very
sensitive to nucleophilic agents. In presence of metha-
nol, 11 gave the ester 4 immediately back.
Scheme 3. Cyclization after reduction through Weinreb intermediate.
NOEÕs of 14 are indicated.
Scheme 2. Direct pyrrole N1–C cyclization.

N. Travert et al. / Tetrahedron Letters 46 (2005) 249–252
251
the reaction mixture was treated with HCl 1M indicates
the Weinreb-complex participation in the mechanism
of the cyclization. Compound 14 was probably formed
from the corresponding aldehyde after a complete
reduction and hydrolysis of the Weinreb complex.
Reduction/cyclization of the brominated intermediate
18 was also investigated to determine if bromine affects
the regioselectivity. Compound 18 was prepared directly
from 13 in presence of 2equiv of NBS in 97% yield
(Scheme 4). In contrast with the above results, reduction
and hydrolysis with both of water or aqueous 1M HCl
gave exclusively the N1–C cyclized carbinolamine 19.
Dehydration of 19 using methansulfonic acid gave 20
in 37% non-optimized yield from 18.
Scheme 6. Bromination and N-acyliminium mechanism.
Apparently, the acid mediated regioselective cyclization
is totally masked by the bromine electronic effect. It is
clear that in this case, steric effect does not influence
the orientation of the cyclization.
nary a-hydroxyketone 30 through a twofold epoxidation
(Scheme 7). Again, the mechanism of the reaction sug-
gests a participation of the N-acyliminium intermediate
27.
After these results, we decided to study the functional-
ization of the olefin of the already cyclized non-bromi-
nated 15 and to investigate the possible participation
of an N-acyliminium 21 in the transformation of
dibromoisophakellin 3a into ugibohlin 2 (Scheme 5).
The electron-rich enol 28 is much more reactive than the
olefin 15. Thus treatment with mCPBA (1equiv or
excess) gave the same compound 30 (46%) along with
the starting material. We were not able to isolate any
intermediate 26–29. The observed over oxidation is
facilitated by the rapid rearrangement of 26 into 28
through the N-acyliminium 27. To our knowledge, the
pyrrolic C3–C six-membered ring ketone of type 30 is
described here for the first time.
At first test, bromination of 15 with 1equiv of Br2 gave a
mixture of mono- and di-brominated compounds. Use
of 2 or 3equiv or excess of bromine in acetic acid gave
the dibrominated compound 24 in 80% yield (Scheme
6). The tribrominated compound 25 was obtained in
60% yield from 24 with excess of bromine in DMSO.
Compounds 15, 24, 25 and 30 could be further function-
alized by metal-catalyzed reactions yielding C–N or C–C
cross-coupled derivatives.
It is also noteworthy that the reaction of 15 with
mCPBA gave rise to the formation of the stable quater-
We were pleased to find that oxidation with bromine or
mCPBA occurs through the valuable N-acyliminium but
preliminary attempts to introduce the guanidine moiety
revealed the necessity to investigate the reaction system-
atically. The mechanism of the bromination (epoxida-
tion) of the olefin 15 seems to involve stepwise the
bromonium bridge 22 (epoxide 26), then the N-acylimin-
ium 23 (27) and finally the brominated double bound 25
(28). Aiming to progress in the synthesis of the natural
product ugibohlin 2, both of these reactions were carried
out in the presence of guanidine but did not enable the
addition of the guanidine on the olefin. It is clear that
Scheme 4. Selective cyclization after pyrrole bromination.
Scheme 5. Important N-acyliminium mechanism connecting the nat-
ural dibromoisophakellin and ugibohlin.
Scheme 7. Twofold epoxidation by mCPBA.

252
N. Travert et al. / Tetrahedron Letters 46 (2005) 249–252
3. (a) Fedoreyev, S. A.; Utkina, N. A.; Ilyin, S. G.;
Reshetnyak, M. V.; Maxinov, O. B. Tetrahedron Lett.
1986, 27, 3177–3180; (b) Assmann, M.; Van Soest, R. W.
M.; Ko¨ck, M. J. Nat. Prod. 2001, 64, 1345–1347.
the kinetic opening of the bromonium or the epoxide
into the corresponding iminium form will lay a serious
difficulty for the introduction of guanidine. Further
work directed to the synthesis of the ugibohlin 2,
dibromoisophakellin 3a and monobromoisophakellin
3b using the above studies is currently underway and
will be reported soon.
4. (a) Foley, L. H.; Buchi, G. J. Am. Chem. Soc. 1982, 104,
¨
1776–1777; (b) Wiese, K. J.; Yakushijin, K.; Horne, D. A.
Tetrahedron Lett. 2002, 43, 5135–5136; (c) Dilley, A. S.;
´
Romo, D. Org. Lett. 2001, 3, 1535–1538; (d) Belanger, G.;
Hong, F.-T.; Overman, L. E.; Rogers, B. N.; Tellew, J. E.;
Trenkle, W. C. J. Org. Chem. 2002, 67, 7880–7883; (e)
Anderson, G. T.; Chase, C. E.; Koh, Y.; Stien, D.;
Weinreb, S. M. J. Org. Chem. 1998, 63, 7594–7595; (f)
Little, T. L.; Webber, S. E. J. Org. Chem. 1994, 59, 7299–
7305; (g) Starr, T. J.; Koch, G.; Carreira, E. M. J. Am.
Chem. Soc. 2000, 122, 8793–8794.
The potent antibiotic, antifungal and anti-inflammatory
activities of the most new compounds were evaluated.
Only compounds 15 and 20 exhibited a significant anti-
fungal activity towards Candida tropicalis (/ inhibition
of 16 and 27mm at 100lg/disk, respectively). In a preli-
minary assay using a fluorimetric method,15 compound
24 inhibited both cobra and bee venom phospholipase
A2 with an ID50 value of 10 and 2lg/mL, respectively.
5. For chemotaxonomy, see: Braekman, J.-C.; Daloze, Q.;
Stoller, C.; Soest, R. W. M. V. Biochem. Syst. Ecol. 1992,
20, 417–431.
6. (a) Al-Mourabit, A.; Potier, P. Eur. J. Org. Chem. 2001,
237–243; (b) Travert, N.; Al-Mourabit, A. J. Am. Chem.
Soc. 2004, 35, Web released; (c) Kitagawa, I.; Kobayashi,
M.; Kitanaka, K.; Kido, M.; Kyogoku, Y. Chem. Pharm.
Bull. 1983, 31, 2321–2328; (d) Linington, R. G.; Williams,
D. E.; Thir, A.; van Soest, R.; Andersen, R. J. Org. Lett.
2003, 15, 2735–2738.
Acknowledgements
We gratefully acknowledge Pr. Pierre Potier for his
encouragement throughout this investigation. We thank
Arlette Longeon (MNHN) for antibacterial tests.
7. (a) Jacquot, D. E. N.; Hoffmann, H.; Polborn, K.; Lindel,
T. Tetrahedron Lett. 2002, 43, 3699–3702; (b) Poullennec,
K. G.; Kelly, A. T.; Romo, D. Org. Lett. 2002, 4, 2645–
2648.
Supplementary data
8. Marchais, S.; Al-Mourabit, A.; Ahond, A.; Poupat, C.;
Potier, P. Tetrahedron Lett. 1999, 40, 5519–5522.
9. Sosa, A. C. B.; Yakushijin, K.; Horne, D. A. Tetrahedron
Lett. 2000, 41, 4295–4299.
10. Qiao, G. G.; Meutermans, W.; Wong, M. W.; Tra¨ubel,
M.; Wentrup, C. J. Am. Chem. Soc. 1996, 118, 3852–3861.
11. Kasum, B.; Prager, R. H.; Tsopelas, C. Aust. J. Chem.
1990, 43, 355–365.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tet-
let.2004.11.066. Detailed experimental procedures and
characterization for 4–5, 11, 13–14, 17–18, 20, 24–25
and 30.
References and notes
12. Suzuki, H.; Shinpo, K.; Yamazaki, T.; Niwa, S.; Yoko-
yama, Y.; Murakami, Y. Heterocycles 1996, 42, 83–86.
1. (a) Sharma, G. M.; Burkholder, P. R. J. Chem. Soc.,
Chem. Commun. 1971, 151–152; (b) Sharma, G. M.;
Magdoff-Fairchild, B. J. Org. Chem. 1977, 42, 4118–4124.
2. Goetz, G. H.; Harrigan, G. G.; Lokos, J. J. Nat. Prod.
2001, 64, 1581–1582.
13. Pigulla, J.; Roder, E. Liebigs Ann. Chem. 1978, 1390–1398.
14. The relative diastereochemistry of the racemic 14 was
determined by NOESY experiments.
15. Radvanyi, F.; Jordan, L.; Russo-Marie, F.; Bon, C. Anal.
Biochem. 1989, 177, 103–109.
¨