Enantioselective Total Synthesis of (-)-Martinellic Acid

Article abstract of DOI:10.1002/anie.201507849

An enantioselective total synthesis of martinellic acid is described. The pyrroloquinoline alkaloid core is efficiently prepared from a quinoline, employing a method which relies on a newly developed Cu-catalyzed enantioselective alkynylation using the ch

Full text of DOI:10.1002/anie.201507849

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201507849
German Edition: DOI: 10.1002/ange.201507849
Natural Products
Enantioselective Total Synthesis of (À)-Martinellic Acid
Mukesh Pappoppula and Aaron Aponick*
Abstract: An enantioselective total synthesis of martinellic acid
is described. The pyrroloquinoline alkaloid core is efficiently
prepared from a quinoline, employing a method which relies
on a newly developed Cu-catalyzed enantioselective alkynyla-
tion using the chiral imidazole-based biaryl P,N ligand
StackPhos to establish the absolute stereochemistry. The
remaining carbon atoms are then installed by means of
a diastereoselective Pd-catalyzed decarboxylative allylation
and the synthesis is completed after straightforward functional-
group manipulation. This new synthetic method enables the
most concise enantioselective synthesis of this important class
of molecules to date.
first total synthesis was reported by Ma et al. in 2001.[6]
Additional syntheses of the natural products in both race-
mic[7] and nonracemic[8] form continue to appear.[9] Batey and
Powell reported an elegant nine-step synthesis of (Æ)-
martinelline, but described their structures as “deceptively
simple”.[7b] However, if synthetic efficiency is assessed in
terms of the step count, enantioselective synthesis has proven
considerably more difficult. Although there are only three
stereogenic centers, the most recent synthesis was reported by
Davies et al. in 2013 and required 20 steps.[8d] This step count
is common among the enantioselective syntheses reported to
date and the relatively high number is indicative of the
inherent synthetic challenges embedded in what speciously
appear to be relatively simple structures. Considering pre-
vious approaches, we felt that a direct route to these alkaloids
from the corresponding quinoline would be advantageous;
however, our strategy would necessitate the development of
new synthetic method (see below). Herein we report a concise
total synthesis of martinellic acid 1 that relies on a new
enantioselective Cu-catalyzed alkynylation using the imida-
zole-based biaryl ligand StackPhos[10] to establish the absolute
stereochemistry in the early stages of the synthesis.
T
he pyrroloquinoline alkaloids (À)-martinellic acid 1 and
(+)-martinelline 2 were isolated by a group at Merck[1] who
were studying the root bark extracts of Martinella iquitoensis
used by South American indigenous peoples for treating eye
infirmities such as conjunctivitis (Figure 1).[2–4] Witherup,
Our initial retrosynthetic analysis required the efficient
generation of ketone 4, which we believed would be rapidly
and easily converted into martinellic acid by standard
synthetic methods, thus streamlining the synthesis. As such,
introduction of the prenyl-substituted guanidines would occur
at a late stage leading back to the differentially protected
martinella core structure 3 (Scheme 1). It seemed prudent to
design a synthesis that could differentiate the three nitrogen
atoms and the carboxylate for ease in substituting these
positions to probe biological activity. The pyrrolidine ring of
Figure 1. Martinella alkaloids.
Varga, and co-workers determined the Martinella alkaloid
structures 1 and 2, which are composed of a pyrroloquinoline
core structure that was previously unknown in natural
products chemistry.[1] These compounds were found to
possess a variety of different biological activities. They were
the first reported nonpeptide bradykinin receptor antagonists
(mm), exhibit anti-a-adrenic activity (nm), and are moderately
effective antibiotics against both Gram-positive and Gram-
negative bacteria.[1]
The martinella alkaloids have attracted significant con-
tinued attention from the synthetic community, likely because
of the combination of interesting biological activity and
unique structural features. These efforts initially centered on
methods to prepare the novel pyrroloquinoline core[5] and the
[*] M. Pappoppula, Prof. A. Aponick
Department of Chemistry
Center for Heterocyclic Compounds, University of Florida
Gainesville, FL 32611 (USA)
E-mail: aponick@chem.ufl.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201507849.
Scheme 1. Retrosynthetic analysis. Pg=protecting group.
Angew. Chem. Int. Ed. 2015, 54, 15827 –15830
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
15827

.
Angewandte
Communications
the core would be derived from a double reductive amination
of a dicarbonyl compound that should be available from the
a-allyl ketone 4. The 1,2-trans stereochemistry required here
should be readily accessible by a variety of ketone a-allylation
methods, and a diastereoselective Pd-catalyzed decarboxyla-
tive allylation[11] leads back to the allylic carbonate 5. Further
simplification of 5 to the corresponding quinoline 7 and
alkyne 6 was postulated to be a good disconnection, however,
at the outset no such enantioselective alkynylation reaction
was known[12] and it was unclear if an allylic carbonate would
be compatible.
As part of a program directed at using the imidazole-
based chiral biaryl ligand StackPhos, we recently developed
a highly enantioselective Cu-catalyzed dearomative quinoline
alkynylation reaction similar to what would be needed
here.[13] The method focused on using simple quinolines, and
it was unknown what effect the 4-OR substituents, specifically
a 4-allyl carbonate, would have on both the reactivity and
selectivity. However, we felt that if the product could be
formed, the allylic carbonate should survive the mild reaction
conditions. Furthermore, it seemed possible to form the
carbonate in situ by reaction of excess chloroformate with
a quinolone, which would be an exceptionally convenient
starting material. To probe this hypothesis, the study com-
menced with preliminary experiments on a simplified system
using 2 equivalents of ethyl chloroformate (10), quinolone 8,
and alkyne 9 [Scheme 2; Eq. (1)]. Under catalytic conditions
with (R)-StackPhos (5.5 mol%), CuBr (5 mol%), and
DIPEA (N,N-diisopropylethylamine), the desired product
12 was isolated in 40% yield and 88% ee along with
carbamate 11. While these results were quite encouraging,
substantial efforts to minimize the formation of 11 and favor
12 were unsuccessful. Further studies demonstrated that, once
formed, carbamate 11 does not undergo further reaction
[Scheme 2, Eq. (2)], but that, once formed, carbonate 13 is
a viable substrate for the formation of 12 [Scheme 2, Eq. (3)].
Interestingly, this suggested that the outcome was governed
by the partitioning between 11 and 13 in the first acylation
step and also demonstrated that a C4-carbonate-substituted
quinoline could be a suitable substrate. Further optimization
with respect to both yield and selectivity would be needed,
but more importantly it was unclear if the allylic carbonate
would be stable under the reaction conditions, or if there
might be O to N crossover by acyl transfer.
To this end, the allylic carbonate 14 was prepared in three
steps from benzocaine, an inexpensive and convenient start-
ing material that is available in large quantities. The reaction
between 14 and the N-protected propargyl amine 9, which
had been demonstrated to work in the initial experiments, was
optimized with the StackPhos/CuBr conditions (Table 1).
Fortunately, under these conditions, minimal amounts of acyl
transfer products were detected and the allylic carbonate
persisted to form the desired product 15 in all cases. As
a starting point, the reaction (employing 2 equivalents of 9
and 2 equivalents of 10) was conducted at 08C with 0.1m of 14
(Table 1, entry 1). These conditions afforded 15 in 86% ee
and the reaction progressed to 95% conversion after 15 h.
Lowering the temperature predictably decreased the con-
version but the selectivity was improved to 89% ee (entry 2).
Scheme 2. Initial enantioselective alkynylation experiments. Conditions
[Eq. (1)]: CuBr (5 mol%), (R)-StackPhos (5.5 mol%), DIPEA
(2.8 equiv). Conditions [Eqs. (2,3)]: 9 (2 equiv), 10 (1 equiv), CuBr
(5 mol%), (R)-StackPhos (5.5 mol%), DIPEA (1.4 equiv). Bn=benzyl.
Table 1: Optimization of the Cu-catalyzed alkynylation.
entry[a]
14 [m] Temperature Time
conversion [%][b]
ee [%][c]
1
2
3
4
5
0.1
0.1
0.2
0.4
0.8
08C
15 h
80 h
42 h
48 h
24 h
95
50
86
89
89
91
87
À258C
À258C
À258C
À258C
60
70[d]
30
[a] Conditions: 9 (2 equiv), 10 (2 equiv), CuBr (5 mol%), (R)-StackPhos
(5.5 mol%), DIPEA (2.8 equiv). [b] Determined by 1H NMR analysis of
the crude reaction mixture. [c] Determined by HPLC with a chiral
stationary phase. [d] Yield of isolated product.
Increasing the concentration of 14 first to 0.2m (entry 3) and
then 0.4m (entry 4) restored reactivity and the product was
isolated in an acceptable 70% yield in 91% ee under these
conditions. Interestingly, further increasing the concentration
(entry 5) did not appear to significantly benefit the reaction.
Carbamate 15 appeared to be a good starting material as it
was readily available, but more importantly provided four
different carbonyl groups (ester, carbonate, carbamate, Cbz;
15828 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 15827 –15830

Angewandte
Chemie
Cbz = benzyloxycarbonyl) that could all be selectively manip-
ulated in distinct ways.
To complete the synthesis of martinellic acid 1, introduc-
tion of the guanidine side-chains, deprotection of the nitro-
gen, and saponification of the ester remained. With the ester
and carbamate in place in 18, the primary and secondary
amines should have differential reactivity that could be
exploited for sequential introduction of the side-chains to
form compounds that may prove useful to probe biological
activity. Gratifyingly, it was found that the primary amine
could be selectively functionalized in 66% yield with the bis-
Boc reagent 19 (Scheme 4; Boc = tert-butoxycarbonyl).[7b]
With the development of a reliable synthetic method for
15, the reaction was scaled up to provide sufficient material to
move forward. It was found that on larger scales the catalyst
loading could be reduced to 2.0 mol% without negatively
impacting the yield or ee value, although a prolonged reaction
time was required (Scheme 3). Under these conditions, 15 was
Scheme 3. Synthesis of the pyrroloquinoline core structure. a) EtOC-
(O)Cl (2.0 equiv), CuBr (2.0 mol%), StackPhos (2.2 mol%), DIPEA
(2.8 equiv), CH2Cl2, À308C, 90 h, 73%, 91% ee; b) [Pd(PPh3)4]
(2.5 mol%), THF, RT, 3 h, 80%, d.r. ꢀ25:1; c) O3, CH2Cl2, À788C,
10 min; DMS À788C to RT, 18 h, 66%; d) BnNH2·HCl, NaCNBH3,
MeOH, RT, 18 h, 72%; e) Pd(OH)2, H2, EtOAc, MeOH, 50 psi, RT,
18 h, 82%. DMS=dimethylsulfide.
Scheme 4. Completion of the synthesis. a) Et3N, CH3CN, RT, 14 h,
66%; b) HgCl2, Et3N, DMF, RT, 3 h, 74%; c) 1n NaOH, MeOH, 658C,
16 h, 87%; d) TFA, anisole, CH2Cl2, RT, 18 h, 70%. DMF=dimethyl-
formamide; TFA=trifluoroacetic acid.
prepared in 73% yield and 91% ee. This set the stage for the
decarboxylative allylation step. Although Stoltz et al. and
others have reported many syntheses that rely on enantiose-
lective decarboxylative allylation to set the absolute stereo-
chemistry, diastereoselective reactions are much less fre-
quently encountered in total synthesis.[14] Treatment of 15
with [Pd(PPh3)4] (2.5 mol%) smoothly provided the a-allyl
ketone 16 in 80% yield with a ꢀ 25:1 ratio of diastereomers,
favoring the desired trans isomer. The olefin of the allyl group
was then selectively cleaved under ozonolysis conditions to
provide a 1,4-diacarbonyl compound that was subsequently
transformed into the pyrrolidine 17 in 72% yield by a double
reductive amination using benzyl amine and sodium cyano-
borohydride.[15] Under hydrogenation conditions, the alkyne
was reduced to an alkane and the three nitrogen protecting
groups cleaved in 82% yield to reveal the martinella alkaloid
core structure 18.
Using the more reactive mono-Boc reagent 21,[7b] the
remaining secondary amine in 20 could then be guanidiny-
lated to form 22 in 74% yield. Deprotection of the carbamate
and saponification of the ester proceeded in 87% yield,
completing a formal synthesis of martinelline.[7b] Finally,
deprotection of the Boc groups[6] yielded martinellic acid 1 as
the trifluoroacetic acid (TFA) salt. The spectroscopic charac-
terization of this compound completely matched previously
reported spectral data.[6–8] The observed optical rotation
verified that we had prepared the correct enantiomer of the
natural product, (À)-martinellic acid 1.[16]
In conclusion, we have reported a catalytic enantioselec-
tive total synthesis of (À)-martinellic acid that proceeds in
twelve steps from benzocaine and employs a new strategy to
Angew. Chem. Int. Ed. 2015, 54, 15827 –15830
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 15829

.
Angewandte
Communications
Ma, Tetrahedron Lett. 2002, 43, 9405 – 9409. For racemic formal
syntheses, see: d) Y. He, R. Moningka, C. J. Lovely, Tetrahedron
Lett. 2005, 46, 1251 – 1254; e) Y. He, H. Mahmud, R. Moningka,
C. J. Lovely, H. V. R. Dias, Tetrahedron 2006, 62, 8755 – 8769;
f) O. Miyata, A. Shirai, S. Yoshino, T. Nakabayashi, Y. Takeda, T.
Kiguchi, D. Fukumoto, M. Ueda, T. Naito, Tetrahedron 2007, 63,
10092 – 10117; g) M. Ueda, S. Kawai, M. Hayashi, T. Naito, O.
Miyata, J. Org. Chem. 2010, 75, 914 – 921; h) Z. Rong, Q. Li, W.
Lin, Y. Jia, Tetrahedron Lett. 2013, 54, 4432 – 4434.
assemble the martinella alkaloid core. Key steps include an
enantioselective Cu-catalyzed alkyne addition to an allyl
carbonate substituted quinolone, followed by diastereoselec-
tive decarboxylative allylation, and a double reductive
amination to set both the absolute and relative stereochem-
istry. This route is the most concise enantioselective synthesis
of (À)-martinellic acid to date and highlights the power of the
new catalytic enantioselective alkyne addition method using
StackPhos as the ligand.
[8] For enantioselective total syntheses see Ref. [6] and: a) S. Ikeda,
M. Shibuya, Y. Iwabuchi, Chem. Commun. 2007, 504 – 506; b) A.
Shirai, O. Miyata, N. Tohnai, M. Miyata, D. J. Procter, D.
Sucunza, T. Naito, J. Org. Chem. 2008, 73, 4464 – 4475; c) T.
Naito, Pure Appl. Chem. 2008, 80, 717 – 726; d) S. G. Davies,
A. M. Fletcher, J. A. Lee, T. J. A. Lorkin, P. M. Roberts, J. E.
Thomson, Org. Lett. 2013, 15, 2050 – 2053. For enantioselective
formal syntheses, see: e) D. Ma, C. Xia, J. Jiang, J. Zhang, W.
Tang, J. Org. Chem. 2003, 68, 442 – 451; f) V. Badarinarayana,
C. J. Lovely, Tetrahedron Lett. 2007, 48, 2607 – 2610; g) Y.
Yoshitomi, H. Arai, K. Makino, Y. Hamada, Tetrahedron 2008,
64, 11568 – 11579; h) S. G. Davies, A. M. Fletcher, J. A. Lee,
T. J. A. Lorkin, P. M. Roberts, J. E. Thomson, Tetrahedron 2013,
69, 9779 – 9803.
Acknowledgements
We thank the National Science Foundation (CHE-1362498)
and the University of Florida for their generous support of
our programs. M.P. thanks Ji Liu (UF) for helpful discussions.
Keywords: alkynes · copper · martinellic acid · natural products ·
total synthesis
How to cite: Angew. Chem. Int. Ed. 2015, 54, 15827–15830
Angew. Chem. 2015, 127, 16053–16056
[9] For reviews of synthetic studies towards Martinella alkaloids,
see: a) M. Nyerges, Heterocycles 2004, 63, 1685 – 1712; b) C. J.
Lovely, V. Badarinarayana, Curr. Org. Chem. 2008, 12, 1431 –
1453.
[10] F. S. P. Cardoso, K. A. Abboud, A. Aponick, J. Am. Chem. Soc.
2013, 135, 14548 – 14551.
[11] a) D. C. Behenna, B. M. Stoltz, J. Am. Chem. Soc. 2004, 126,
15044 – 15045; b) B. M. Trost, J. Y. Xu, J. Am. Chem. Soc. 2005,
127, 2846 – 2847. For comprehensive reviews on decarboxylative
allylation, see: c) J. T. Mohr, B. M. Stoltz, Chem. Asian J. 2007, 2,
1476 – 1491; d) J. D. Weaver, A. Recio III, A. J. Grenning, J. A.
Tunge, Chem. Rev. 2011, 111, 1846 – 1913.
[12] For leading references, see: a) D. A. Black, R. E. Beveridge,
B. A. Arndtsen, J. Org. Chem. 2008, 73, 1906 – 1910; b) A. M.
Taylor, S. L. Schreiber, Org. Lett. 2006, 8, 143 – 146; c) W. Lin, T.
Cao, W. Fan, Y. Han, J. Kuang, H. Luo, B. Miao, X. Tang, Q. Yu,
W. Yuan, J. Zhang, C. Zhu, S. Ma, Angew. Chem. Int. Ed. 2014,
53, 277 – 281; Angew. Chem. 2014, 126, 281 – 285, and references
therein.
[13] M. Pappoppula, F. S. P. Cardoso, B. O. Garrett, A. Aponick,
Angew. Chem. Int. Ed. 2015, 54, DOI: 10.1002/anie.201507848;
Angew. Chem. 2015, 127, DOI: 10.1002/ange.201507848.
[14] a) T. Tanaka, N. Okamura, K. Bannai, A. Hazato, S. Sugiura, K.
Tomimori, K. Manabe, S. Kurozumi, Tetrahedron 1986, 42,
6747 – 6758; b) K. C. Nicolaou, G. Vassilikogiannakis, W. Mager-
lein, R. Kranich, Angew. Chem. Int. Ed. 2001, 40, 2482 – 2486;
Angew. Chem. 2001, 113, 2543 – 2547; c) D. A. Carcache, Y. S.
Cho, Z. Hua, Y. Tian, Y. M. Li, S. J. Danishefsky, J. Am. Chem.
Soc. 2006, 128, 1016 – 1022; d) R. M. McFadden, B. M. Stoltz, J.
Am. Chem. Soc. 2006, 128, 7738 – 7739; e) J. A. Enquist, Jr.,
B. M. Stoltz, Nature 2008, 453, 1228 – 1231; f) C. J. Gartshore,
D. W. Lupton, Angew. Chem. Int. Ed. 2013, 52, 4113 – 4116;
Angew. Chem. 2013, 125, 4207 – 4210; g) F. Horeischi, C.
Guttroff, B. Plietker, Chem. Commun. 2015, 51, 2259 – 2261.
[15] B. Lei, C. Ding, X. Yang, X. Wan, X. Hou, J. Am. Chem. Soc.
2009, 131, 18250 – 18251.
[1] K. M. Witherup, R. W. Ransom, A. C. Graham, A. M. Bernard,
M. J. Salvatore, W. C. Lumma, P. S. Anderson, S. M. Pitzen-
berger, S. L. Varga, J. Am. Chem. Soc. 1995, 117, 6682 – 6685.
[2] A. Gentry, K. Cook, J. Ethnopharmacol. 1984, 11, 337 – 343.
[3] A. Boralkar, P. Dindore, R. Fule, B. Bangde, M. Albel, A. Saoji,
Indian J. Ophthalmol. 1989, 37, 94 – 95.
[4] Interestingly, Alexander Anderson described the ophthalmic
properties of Martinella root extract in a manuscript from 1791
that unfortunately was not published. For complete details, see
Ref. [2].
[5] a) C. J. Lovely, H. Mahmud, Tetrahedron Lett. 1999, 40, 2079 –
2082; b) B. B. Snider, Y. Ahn, B. M. Foxman, Tetrahedron Lett.
1999, 40, 3339 – 3342; c) M. Nyerges, I. Fejes, L. Toke, Tetrahe-
dron Lett. 2000, 41, 7951 – 7954; d) M. Hadden, M. Nieuwen-
huyzen, D. Osborne, P. J. Stevenson, N. Thompson, Tetrahedron
Lett. 2001, 42, 6417 – 6419; e) H. Mahmud, C. J. Lovely, H. V. R.
Dias, Tetrahedron 2001, 57, 4095 – 4105; f) Y. He, H. Mahmud,
B. R. Wayland, H. V. R. Dias, C. J. Lovely, Tetrahedron Lett.
2002, 43, 1171 – 1174; g) A. Viranyi, M. Nyerges, G. Blask, L.
Toke, Synthesis 2003, 2655 – 2660; h) M. Nyerges, A. Viranyi, L.
Toke, Heterocycl. Commun. 2003, 9, 239 – 242; i) J. S. Yadav,
B. V. Subba Reddy, V. Sunitha, K. Srinivasa Reddy, K. V. S.
Ramakrishna, Tetrahedron Lett. 2004, 45, 7947 – 7950; j) M.
Hadden, M. Nieuwenhuyzen, D. Osborne, P. J. Stevenson, N.
Thompson, A. D. Walker, Tetrahedron 2006, 62, 3977 – 3984;
k) P. Y. Ng, C. E. Masse, J. T. Shaw, Org. Lett. 2006, 8, 3999 –
4002; l) O. Miyata, A. Shirai, S. Yoshino, Y. Takeda, M. Sugiura,
T. Naito, Synlett 2006, 893 – 896; m) Z. Zhang, Q. Zhang, Z. Yan,
Q. Liu, J. Org. Chem. 2007, 72, 9808 – 9810; n) M. Neuschl, D.
Bogdal, M. Potacek, Molecules 2007, 12, 49 – 59; o) S. Comesse,
M. Sanselme, A. Daìch, J. Org. Chem. 2008, 73, 5566 – 5569;
p) M. Boomhoff, A. K. Yadav, J. Appun, C. Schneider, Org. Lett.
2014, 16, 6236 – 6239; q) Y. Yang, H. Zhang, S. Zhu, P. Zhu, X.
Hui, Org. Lett. 2014, 16, 5048 – 5051.
[16] See the Supporting Information for full details.
[6] D. Ma, C. F. Xia, J. Q. Jiang, J. Zhang, Org. Lett. 2001, 3, 2189 –
2191.
[7] For racemic total syntheses, see: a) B. B. Snider, Y. Ahn, S. M.
OHare, Org. Lett. 2001, 3, 4217 – 4220; b) D. A. Powell, R. A.
Batey, Org. Lett. 2002, 4, 2913 – 2916; c) C. F. Xia, L. S. Heng, D.
Received: August 21, 2015
Revised: September 26, 2015
Published online: November 17, 2015
15830 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 15827 –15830