Selective Functionalization of Antimycin A Through an N-Transacylation Reaction

Article abstract of DOI:10.1021/acs.orglett.6b00882

Acylation of 3-(N-formylamino)salicylic acids resulted in transacylation with loss of the formyl moiety. The reaction proceeds through a bis-N-acylated intermediate, which undergoes facile deformylation. This transacylation reaction has been employed for the site-specific functionalization of the mitochondrial poison antimycin A, affording several novel derivatives. The selective cytotoxicity of some of these derivatives toward cultured A549 human lung epithelial adenocarcinoma cells, in comparison with WI-38 normal human lung fibroblasts, illustrates one application of this transacylation reaction.

Full text of DOI:10.1021/acs.orglett.6b00882

Letter
pubs.acs.org/OrgLett
Selective Functionalization of Antimycin A Through an
N‑Transacylation Reaction
Arnaud Chevalier, Yanmin Zhang, Omar M. Khdour, and Sidney M. Hecht*
Biodesign Center for BioEnergetics and School of Molecular Sciences, Arizona State University, Tempe, Arizona 85287, United States
S
* Supporting Information
ABSTRACT: Acylation of 3-(N-formylamino)salicylic acids resulted in transacylation with loss of the formyl moiety. The
reaction proceeds through a bis-N-acylated intermediate, which undergoes facile deformylation. This transacylation reaction has
been employed for the site-specific functionalization of the mitochondrial poison antimycin A, affording several novel derivatives.
The selective cytotoxicity of some of these derivatives toward cultured A549 human lung epithelial adenocarcinoma cells, in
comparison with WI-38 normal human lung fibroblasts, illustrates one application of this transacylation reaction.
ntimycin A denotes a family of structurally related
natural products produced by Streptomyces; they differ
Scheme 1. Acylation of Antimycin A: An Unexpected
Result
A
only by the aliphatic substituents appended to the bislactone
moiety.1 These compounds have been shown to inhibit the
activity of complex III of the mitochondrial electron transport
chain, resulting in inhibition of autophagy2 and induction of
apoptosis3 in cultured mammalian cells. Accumulating
evidence supports the involvement of mitochondrial bioen-
ergetics and signaling in tumorigenesis and suggests that
agents that target these functions may find utility in the
chemotherapy of some tumors.4 Several classes of molecules,
including antimycin A, are of interest in this regard.5
Hockenbery and co-workers described the effect of 2-methoxy
antimycin A3 as an experimental antitumor agent; this
compound was reported to have minimal effects on the
respiratory chain but to compete with a proapoptic Bak BH3
peptide for binding to recombinant Bcl-2. Consequently, cells
expressing Bcl-xL, which are usually resistant to anticancer
drugs, were hypersensitive to antimycin A.6 In addition to
antimycin A, several types of other molecules targeting Bcl-2
or Bcl-xL have been in (pre)clinical trials,7 but no further
efforts to develop therapeutic agents based on the structure of
antimycin A have been described.
In an effort to develop new antimycin A derivatives by
functionalization of the natural product,8 we attempted the O-
acylation of antimycin A, in analogy with the reported6a O-
alkylation. As shown in Scheme 1, attempted O-acylation with
hexanoic acid resulted instead in replacement of the N-formyl
moiety. As a consequence we began to investigate this
unexpected result. The N-transacylation of formylated aniline
derivatives does not seem to have been reported using metal
free methods. Literature examples of methods to effect the
alteration of alkylcarboxamide structures involve metal
catalysts such as palladium in aminocarbonylation reactions,10
ruthenium catalyzed olefin hydrocarbamoylation reactions,11
and more recently cobalt catalyzed activation of aldehydic C−
H bonds.12 Moreover, those reactions were all performed at
high temperatures (>100 °C), reinforcing the unusual nature
of the transformation outlined in Scheme 1, which occurred at
room temperature.
Since the sample of antimycin A that we had employed was
actually a mixture of congeners,8 a simpler model system was
employed initially to study the N-transacylation reaction.
Scheme 2 presents the synthesis of an antimycin A model
compound in three steps. Hydrogenation of 3-nitrosalicylic
acid under pressure afforded the respective aniline derivative.
Formylation of this derivative was performed using the
conditions reported by Pettit et al.13 to afford compound 2a
in 85% overall yield for the two steps. Amidation in the
presence of n-butylamine gave model compound 2b in 48%
yield, along with the deformylated compound 2c (44% yield).
The latter was reformylated by treatment with formamide,
Received: March 26, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b00882
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
conversion. Important mechanistic information was obtained
by using an O-benzylated analogue of 2b as a starting material
for the transacylation reaction. Even using the conditions that
we determined to be optimal for 2b (entry 4), the reaction
appeared to stop at intermediate 3 (Table 1, entries 10 and
11). Clearly, the phenolic moiety is involved in the
deformylation reaction leading from 3 to 1c since in the
absence of this free OH group no product was formed. While
the precise mechanism for deformylation is presently unclear,
it may be noted that incubation of 1 mM 3a in a 1:2 mixture
of acetonitrile and 10 mM phosphate buffer, pH 7.4, resulted
in complete conversion to 1c within 1 h.
Scheme 2. Synthesis of AMA Model Compound 2b
In addition to its role in providing insights into the
transacylation mechanism, it may be noted that 3a represents
a plausible prodrug for the final transacylated product, which
may find utility in biological/therapeutic contexts.
The optimized reaction conditions (Table 1) were then
applied to define the scope of this transacylation reaction. As
shown in Table 2, ten structurally diverse carboxylic acids
exactly as described for the initial preparation of compound
2a. With compound 2b in hand, we studied its reactivity with
activated esters of hexanoic acid under a variety of conditions
(Table 1).
a
Table 1. Study of the Transacylation Reaction
a
Table 2. Scope of the Transacylation Reaction
a
All reactions were performed using DIEA (3 equiv) as base at a 0.2 M
concentration of 2b at room temperature under an argon atmosphere.
b
The yields represent pure compound isolated following purification
c
d
on a silica gel column. 3a, R = H; 3b, R = Bn. Two equivalents of
DIEA were added to the reaction mixture of entry 2 after 24 h, and the
reaction was maintained for an additional 6 h until it was complete.
It was quickly apparent that the use of PyBrOP resulted in
lower yields than was realized using HBTU or BOP. The
latter provided the N-transacylation product 1c in good yield
after 15 h of reaction in CH2Cl2 at room temperature. It is
important to note that the low yield of this product using
PyBrOP was mainly a consequence of recovering significant
amounts of N-formylated analogue 3,14 which did not happen
using the other two coupling reagents. That the N-bisacylated
species 3 is an intermediate on the reaction pathway from 2b
to 1c was established by adding two additional equivalents of
DIEA after 24 h of reaction (Table 1, entry 3); after an
additional 6 h, 1c was obtained in good yield, essentially
equivalent to that obtained using HBTU (entry 4). The use of
a solvent other than CH2Cl2 (e.g., THF, DMF, or CH3CN)
did not give significantly different results, indicating that the
transacylation reaction should be compatible with a variety of
organic solvents. Only a kinetic effect was observed, especially
for THF, which required at least 24 h to achieve complete
a
All reactions were performed using DIEA (3 equiv), 2b (0.2 M), and
HBTU (1.2 equiv) in CH2Cl2 at room temperature under an argon
atmosphere for 15 h. DMF was used as a solvent to optimize the
solubility of the pyridinium salt.
b
afforded the anticipated transacylation products (1c, 4a-i) in
isolated yields ranging from 85% to 99%, confirming the
generality of the transformation.
The same reaction conditions were then employed for the
attempted derivatization of antimycin A, using the commer-
cially available mixture of this natural product.8 The use of
acetic acid and 6-phenylhexanoic acid afforded the expected
B
DOI: 10.1021/acs.orglett.6b00882
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
products 5a and 5b in 75 and 99% yields, respectively (Table
3).15 Numerous reports have established the ability of cationic
Table 3. Application to the Selective N-Acylation of
a
Antimycin A
Figure 1. Cytotoxicity assays of compounds 5a−5f in comparison
with antimycin A for human lung cancer (A549) and normal lung
fibroblast (WI-38) cells. Each of seven test compounds was
employed at 10 μM (left bar) and 50 μM (right bar) concentrations.
a
All reactions were performed using DIEA (3 equiv), 2b (0.2 M), and
HBTU (1.2 equiv) in CH2Cl2 at room temperature under an argon
atmosphere for 15 h. DMF was used as a solvent to optimize the
b
solubility of the pyridinium salt.
1 and respirantin,13 as well as other natural products bearing
the same 3-N-formylaminosalicylic acid moiety, might be
amenable to functionalization using the same strategy.
functional groups to effect the delivery of small molecules to
the mitochondria.16 Since the mitochondria represent a major
locus of action of antimycin A, triphenylphosphonium and
pyridinium-containing carboxylic acids were prepared and
used to provide access to antimycin A derivatives 5c−5f in
good yields.15 Overall, the six antimycin A derivatives were
obtained in yields ranging from 73 to 99%. The somewhat
lower yields obtained for the derivatives containing a
pyridinium moiety (4h, 5c, and 5d) were all associated with
the use of DMF as solvent to facilitate dissolution of the
carboxylic acid; this solvent was associated with a somewhat
lower yield during the initial optimization experiments (Table
1, entry 8). The lower yields may also be related to small
losses of material during the aqueous washing in the workup
process.
A preliminary investigation of the biological properties of
AMA analogues 5a−5f has been carried out, involving their
cytotoxicity toward cultured mammalian cells. The assay
involved a comparison of the effects of these compounds on
A549 human lung epithelial adenocarcinoma cells, in
comparison with WI-38 normal human lung fibroblasts. The
results are reported in Figure 1. While limited differences in
activity were noted for these compounds after 24 h of
treatment, longer incubation times gave increasing divergence
in expressed cytotoxicity among the individual analogues, and
a clear selectivity for cytotoxicity toward cancer cells. These
effects were especially pronounced for compound 5b.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b00882.
Experimental procedures and characterization data
(PDF)
NMR spectra (PDF)
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: sidney.hecht@asu.edu. Phone: (480) 965-6625. Fax:
(480) 965-0038.
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) Strong, F. M.; Dickie, J. P.; Loomans, M. E.; van Tamelen, E.
E.; Dewey, R. S. J. Am. Chem. Soc. 1960, 82, 1513.
(2) Ma, X.; Jin, M.; Cai, Y.; Xia, H.; Long, K.; Liu, J.; Yu, Q.; Yuan,
J. Chem. Biol. 2011, 18, 1474.
(3) Park, W.-H.; Han, Y.-W.; Kim, S.-W.; Kim, S.-H.; Cho, K.-W.;
Kim, S.-Z. Cancer Lett. 2007, 251, 68.
(4) (a) Wallace, D. C. Nat. Rev. Cancer 2012, 12, 685. (b) Sullivan,
L. B.; Chandel, N. S. Cancer Metab. 2014, 2, 17. (c) Weinberg, S. E.;
Chandel, N. S. Nat. Chem. Biol. 2015, 11, 9.
(5) Fantin, V. R.; Leder, P. Oncogene 2006, 25, 4787.
(6) (a) Tzung, S.-P.; Kim, K. M.; Basanez, G.; Giedt, C. D.; Simon,
J.; Zimmerberg, J.; Zhang, K. Y. J.; Hockenbery, D. M. Nat. Cell Biol.
2001, 3, 183. (b) Schwartz, P. S.; Manion, M. K.; Emerson, C. B.;
In summary, the convenient procedure described here for
the functionalization of antimycin A should facilitate the
further study of this mitochondrial poison, particularly at the
levels of defining nonmitochondrial loci of action, and
controlling trafficking to appropriate cellular compartments.
It may also be noted that cyclodepsipeptides such as kitastatin
C
DOI: 10.1021/acs.orglett.6b00882
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Fry, J. S.; Schulz, C. M.; Sweet, I. R.; Hockenbery, D. M. Mol. Cancer
Ther. 2007, 6, 2073. (c) Schwartz, P. S.; Hockenbery, D. M. Cancer
Biol. Ther. 2007, 6, 465.
(7) (a) Lessene, G.; Czabotar, P. E.; Colman, P. M. Nat. Rev. Drug
Discovery 2008, 7, 989. (b) Kang, M. H.; Reynolds, C. P. Clin. Cancer
Res. 2009, 15, 1126.
(8) The commercial sample of antimycin A that we employed for
our studies was actually a mixture consisting predominantly of
antimycins A1, A2, A3 and A4, i.e. differing in the aliphatic
substituents on the bislactone ring (cf Figure S1 of the Supporting
Information).9 For the experiments described here, we employed this
mixture as received and assayed the product mixtures for their
cytotoxic effects. Additionally, a sample of the commercial antimycin
A mixture was separated by HPLC and the individual constituents
were shown to have quite similar biological properties on the
respiratory chain, and as cytotoxic agents.
(9) (a) Liu, W.-C.; Strong, F. M. J. Am. Chem. Soc. 1959, 81, 4387.
(b) Viegelmann, C.; Margassery, L. M.; Kennedy, J.; Zhang, T.;
O’Brien, C.; O’Gara, F.; Morrissey, J. P.; Dobson, A. D. W.; Edrada-
Ebel, R. Mar. Drugs 2014, 12, 3323.
(10) Sawant, D. N.; Wagh, Y. S.; Bhatte, K. D.; Bhanage, B. M. J.
Org. Chem. 2011, 76, 5489.
(11) Li, B.; Park, Y.; Chang, S. J. Am. Chem. Soc. 2014, 136, 1125.
(12) Bai, C.; Yao, X.; Li, Y. ACS Catal. 2015, 5, 884.
(13) Pettit, G. R.; Smith, T. H.; Feng, S.; Knight, J. C.; Tan, R.;
Pettit, R. K.; Hinrichs, P. A. J. Nat. Prod. 2007, 70, 1073.
(14) The assignment of a N,N-bisacylated structure to compound
3a, rather than a N,O-biosacylated species, was based both on the
1
appearance of a phenolic H in the H NMR spectra at ∼ 13.1 ppm
for entries 1, 2, 6 and 8 in Table 1, and on the fluorescence of
compound 3a (as compared with 3b, entries 10 and 11 in Table 1),
which requires the presence of a free phenolic OH group.
(15) In each case, high resolution mass spectrometric analysis of
the isolated mixture revealed the presence of the expected molecular
ion for each of the products derived from antimycins A1, A2, A3 and
A4.
(16) (a) Rin Jean, S.; Tulumello, D. V.; Wisnovsky, S. P.; Lei, E. K.;
Pereira, M. P.; Kelley, S. O. ACS Chem. Biol. 2014, 9, 323. (b) Ma,
J.; Lim, C.; Sacher, J. R.; Van Houten, B.; Qian, W.; Wipf, P. Bioorg.
Med. Chem. Lett. 2015, 25, 4828.
D
DOI: 10.1021/acs.orglett.6b00882
Org. Lett. XXXX, XXX, XXX−XXX