Exploration of the molecular origin of the azinomycin epoxide: Timing of the biosynthesis revealed

Article abstract of DOI:10.1021/ol8018852

(Equation Presented) Streptomyces sahachiroi whole cell feeding experiments, utilizing putative precursors labeled with stable isotopes, established that the epoxide unit of the DNA cross-linked agents, azinomycin A and B, proceeds via a valine-dependent pathway and that hydroxylation and dehydration precedes formation of the terminal epoxide. Sodium 3-methyl-2-oxobutenoate, formed through a transimination reaction, was shown to be the penultimate precursor incorporated into the azinomycin epoxide.

Full text of DOI:10.1021/ol8018852

ORGANIC
LETTERS
2008
Vol. 10, No. 21
4815-4818
Exploration of the Molecular Origin of
the Azinomycin Epoxide: Timing of the
Biosynthesis Revealed
Vasudha Sharma, Gilbert T. Kelly, and Coran M. H. Watanabe*
Department of Chemistry, Texas A&M UniVersity, College Station, Texas 77842
watanabe@mail.chem.tamu.edu
Received August 13, 2008
ABSTRACT
Streptomyces sahachiroi whole cell feeding experiments, utilizing putative precursors labeled with stable isotopes, established that the epoxide
unit of the DNA cross-linked agents, azinomycin A and B, proceeds via a valine-dependent pathway and that hydroxylation and dehydration
precedes formation of the terminal epoxide. Sodium 3-methyl-2-oxobutenoate, formed through a transimination reaction, was shown to be the
penultimate precursor incorporated into the azinomycin epoxide.
Azinomycin B is a potent antitumor agent produced by
certain soil-dwelling Streptomyces species.1 The metabolite
interacts with DNA, forming interstrand cross-links within
the major groove without prior activation. In vitro studies
reveal that the agent interacts with the duplex DNA sequence
5′-d(PuNPy)-3′, forming interstrand covalent linkages be-
tween the electrophilic C10 and C21 carbons of azinomycin
The overall structure of azinomycin B is suggestive of a
mixed biosynthetic origin based upon a polyketide synthase
(PKS)/nonribosomal peptide synthetase (NRPS) skeleton
(Figure 1). Formation of the naphthoate ring system can be
2
and the N7 positions of suitably disposed purine bases.
Fluorescence imaging with the natural product revealed
localization within the nuclei of yeast and experiments with
DNA microarrays resulted in transcriptional effects that were
closely associated with DNA damage and repair responses
providing a direct correlation of in vitro DNA cross-linking
with an in vivo cellular response.3
Figure 1. Structure of azinomycins A and B.
(1) (a) Nagaoka, K.; Matsumoto, M.; Oono, J.; Yokoi, K.; Ishizeki, S.;
Nakashima, T. J. Antibiot. 1986, 39, 1527–32. (b) Yokoi, K.; Nagaoka, K.;
Nakashima, T. Chem. Pharm. Bull. 1986, 34, 4554–61. (c) Hata, F.; Koga,
Y.; Sano, K.; Kanamori, A.; Matsumae, R.; Sugawara, T.; Hoshi, T.; Shimi,
S.; Ito, S.; Tomizawa, S. J. Antibiot. 1954, 7, 107–112.
rationalized by the successive condensation of acetate and
malonate units by a PKS. Further functionalization of the
natural product by the action of an NRPS and various
tailoring enzymes would give the epoxide, enol, and azabi-
cycle. Data available on the biosynthesis has been limited,
largely due to difficulties with the culture method and
(2) (a) Lown, J. W.; Majumdar, K. C. Can. J. Biochem. 1977, 55, 630–
635. (b) Coleman, R. S.; Perez, R. J.; Burk, C. H.; Navarro, A. J. Am.
Chem. Soc. 2002, 124, 13008–13017. (c) Lepla, R. C.; Landreau, A. S.;
Shipman, M.; Jones, G. D. D. Org. Biomol. Chem. 2005, 3, 1174–1175.
(d) Armstrong, R. W.; Salvati, M. E.; Nguyen, M. J. Am. Chem. Soc. 1992,
14, 3144–3145. (e) Zang, H.; Gates, K. S. Biochemistry 2000, 39, 14968–
14975. (f) Fujiwara, T.; Saito, I.; Sugiyama, H. Tetrahedron Lett. 1999,
40, 315–318.
(3) Kelly, G. T.; Liu, C.; Smith, R., III; Coleman, R. S.; Watanabe,
C. M. H. Chem. Biol. 2006, 13, 1–8.
10.1021/ol8018852 CCC: $40.75
Published on Web 10/09/2008
2008 American Chemical Society

securing a consistent source of the natural product. Initially,
some gains were made to establish the polyketide origin of
the naphthoate moiety,4 and a cell-free system was developed
to support synthesis of azinomycin B in vitro.5 Recent
breakthroughs have led to optimization of culture conditions
by nutrient limitation resulting in marked improvement in
azinomycin production. This methodology enables a reliable
culture method for isotopic feeding studies as illustrated by
the definitive assignment of threonine as the most advanced
intermediate accepted by the NRPS machinery in final
processing and construction of the enol moiety.6
In this paper, we explore the biosynthetic route to the
azinomycin epoxide, where exact timing of individual
enzymatic transformations and intermediacy of metabolites
were substantiated by whole cell feeding studies with
isotopically labeled substrates. Structural evaluation and
results from a recent cell-free study would suggest that
L-valine serves as a logical precursor (Figure 2).5 If this is
indeed the case, several scenarios can be envisaged. One
possibility is that valine or respective advanced precursors
are converted to the epoxide while tethered to the NRPS
(Figure 2, path A), where catalysis is achieved by modifying
enzymes either contained within NRPS domains or elsewhere
within the gene cluster. In cerulide biosynthesis, for example,
a reductase module found within NRPS adenylation domains,
reduces tethered R-keto acids (R-ketoisocaproic acid and
R-ketoisovaleric acid) to R-hydroxy acids, while still bound
to their respective PCP domains.7 Alternatively, it is
conceivable that the epoxide moiety is fully constructed by
other biosynthetic pathway enzymes prior to activation by
its respective adenylation module. For instance, the epoxide
might be derived from radical cyclization of a terminal
alcohol (Figure 2, path B).8 In contrast, the epoxide might
originate from reaction with molecular oxygen via oxygen
insertion into an olefin.9 This olefin might in turn be derived
from a terminal alcohol (Figure 2, path C).9 In either case
(Figure 2, path B or C), the amino group of valine also
requires biochemical conversion to give the R-hydroxy group
of the epoxide fragment 9, presumably arising by transimi-
nation and subsequent reduction of the resulting ketone to
give the alcohol. The exact timing of these biosynthetic
transformations and feasibility of the overall route can only
be resolved through experimentation. Therefore, to test our
hypotheses, we exogenously fed valine and a variety of
Figure 2. Possible biosynthetic routes to the epoxide moiety in
the azinomycins.
synthesized advanced precursors in isotopically labeled form
to whole cells (Table 1).
When [1-13C]-valine was administered to cultures of S.
sahachiroi, site-specific enrichment was observed at C-17
of the natural product, corroborating our cell-free experiments
with [1-14C]-valine. Analysis of the 13C NMR spectrum
revealed a 8.71% incorporation at C-17, 164.0 ppm (see
Table 1 and Supporting Information). Encouraged by this
result, we synthesized the advanced precursors shown in
Table 1.
The synthetic routes were designed to allow easy and
efficient preparation, in high yields and with minimal
chromatographic separations, providing a cost-effective ap-
proach to these molecules in isotopically labeled form. In
addition to the biosynthetic studies described here, the
availability of synthetic routes to unnatural amino/R-hydroxy
acid is of pharmaceutical importance as they can serve as
peptidomimetic substrates10 as well as building blocks in the
synthesis of natural product pharmacophores.11
(4) (a) Corre, C.; Lowden, P. A. S. Chem. Commun. 2004, n/a, 990–
991. (b) Corre, C.; Landreau, C. A. S.; Shipman, M.; Lowden, P. A. S.
Chem. Commun. 2004, n/a, 2600–2601.
(5) Liu, C.; Kelly, G. T.; Watanabe, C. M. H. Org. Lett. 2006, 8, 1065–
1068.
(6) Kelly, G. T.; Sharma, V.; Watanabe, C. M. H. Bioorganic Chem.
2008, 36, 4–15.
(7) Magarvey, N. A.; Ehling-Schulz, M.; Walsh, C. T. J. Am. Chem.
Soc. 2006, 128, 10698–10699.
(8) (a) Yan, F.; Munos, J. W.; Liu, P.; Liu, H. W. Biochem. 2006, 45,
11473–11481. (b) Munos, J. W.; Moon, S. J.; Mansoorabadi, S. O.; Change,
W.; Hong, L; Yan, F.; Liu, A.; Liu, H. W. Biochemistry 2008, 47, 8726–
35.
Biosynthetic precursors 11 and 12 (Scheme 1) were
synthesized as follows. Commercially available 2-methyl-
(9) (a) Gru¨schow, S.; Sherman, D. H. The Biosynthesis of Epoxides. In
Aziridines and Epoxides in Organic Synthesis, Yudin, A. K., Ed.; Wiley-
VCH Verlag GmbH & Co. KGaA: Weinhem, 2006; pp 349-398 and
references within. (b) Watanabe, C. M. H.; Townsend, C. A. J. Org. Chem.
1996, 61, 1990–1993.
(10) (a) Hanessian, S.; Auzzas, L. Acc. Chem. Res. 2008, 41, 1241-
1251.
(11) (a) Aina, O. H.; Liu, R.; Sutcliffe, J. L.; Marik, J.; Pan, C.-X.;
Lam, K. S. Mol. Pharm. 2007, 4, 631–651.
4816
Org. Lett., Vol. 10, No. 21, 2008

Scheme 2
Table 1. Percent Incorporationc at C-17 of Azinomycin B and
the Epoxyamide (Samples Were Prepared As Detailed in Kelly
et al.6)
Allylic oxidation with activated MnO213 in ether afforded
the R,ꢀ-unsaturated ester 16 which upon hydrolysis with
esterase at pH 8.0 gave sodium 3-methyl-2-oxobut-3-enoate
12.
Epoxides 8 and 9 (Scheme 26) were synthesized from ethyl
bromoacetate 17 converting it to its corresponding phospho-
nate, which upon coupling under standard Horner-
Wadsworth-Emmons conditions afforded the acrylic ester
18 in good yields. The ester was hydrolyzed and converted
to the benzyl ester, which upon Sharpless asymmetric
dihydroxylation with AD-mix R afforded a chiral diol system
19.14 Mesylation followed by reflux with anhydrous K2CO3
generated the chiral epoxide 20, which was carefully opened
with catalytic amounts of anhydrous pTsOH to generate a
chiral allylic alcohol 21. Treatment with VO(acac)2 and
t-butyl hydroperoxide gave the desired benzyl epoxy ester
22, which upon hydrogenation in the presence of 10% Pd/C
generated the precursor 9. Mild oxidation of the allyic alcohol
with 2-iodoxybenzoic acid (IBX) gave the corresponding R-
ketoester, which upon hydrogenation facilitated debenzyla-
tion giving the desired product 8.
a
L-Valine TFA salt showed an overall incorporation of 6.97%. b This
R-keto carboxylic acid degraded rapidly. c The percent incorporation ) [(A
- B)/B] × 1.10 where A ) intensity of labeled carbon normalized to the
intensity of the 3′OCH3 of azinomycin B and the C18 of the epoxyamide;
B ) intensity of normalized unlabeled carbon; 1.10 ) natural abundance
of 13C. The table entry reads n.d. if incorporation was not detected by 13C
NMR.
Isodehydrovaline 10 (Scheme 1 of the Supporting Infor-
mation) was synthesized via a modified route detailed by
Baldwin et al.,15 and γ-hydroxyvaline 3 was generated by
an extension of earlier work by Kazmaier et al. (Scheme 2
of the Supporting Information).16
The γ-hydroxy R-keto acid 5 (Scheme 3) was generated
by condensing ethyl pyruvate with the ylide of trieth-
ylphosphonoacetate via a Horner-Wadsworth-Emmons
condensation. Hydrolysis followed by dehydration, regiose-
lective reduction of anhydride followed by hydrogenation
afforded the lactone 26. R-Hydroxylation using Vedejs’
reagent, oxidation, and subsequent hydrolysis provided 5 in
good yields.
Scheme 1
Each of the modified valine derivatives were synthesized
in 13C-labeled form (at C-1) and fed separately to whole cell
suspension cultures as detailed previously.6 Table 1 provides
2-propenal was treated with sodium cyanide and acetic acid
to afford cyanohydrin 14,12 which upon refluxing with
methanolic HCl generated the R-hydroxy methyl ester 15.
Additionally, hydrolysis with LiOH generated the R-hydroxy-
3-methylbut-3-enoic acid 11.
(13) Capps, N. K.; Davies, G. M.; Loakes, D.; Mecabe, R. W.; Young,
D. W. J. Chem. Soc., Perkin. Trans. 1 1991, 12, 3077–3086.
(14) Bryant, H. J.; Dardonville, C. Y.; Hodgkinson, T. J.; Hursthouse,
M. B.; Abdul Malik, K. M.; Shipman, M J. Chem. Soc., Perkin Trans. 1
1998, 1249–1255.
(15) Baldwin, J. E.; Haber, S. B.; Hoskins, C.; Kruse, L. I. J. Org. Chem.
1977, 42, 1239–1241.
(12) Faulkner, D. J.; Petersen, M. R. J. Am. Chem. Soc. 1973, 95, 553–
63.
(16) (a) Kazmaier, U.; Krebs, A. Tetrahedron Lett. 1999, 40, 479–482.
(b) Kazmaier, U.; Krebs, A. Angew. Chem., Int. Ed. Engl. 1995, 34, 2012.
Org. Lett., Vol. 10, No. 21, 2008
4817

against many cancers, compound toxicity and selectivity is
an issue. The azinomycins are no exception, and significant
effort has already been made to synthesize chemical ana-
logues of these compounds.19 In this investigation, we
demonstrate timing and intermediacy of several metabolites
along the biosynthetic route to the epoxide moiety (Figure
3). We were intrigued to find that the majority if not all of
Scheme 3
the feeding results for all of the amino acid precursors. In
addition to valine 2, only substrates 3, 5, and 12 resulted in
site-specific incorporation above background at 164.0 ppm.
Interestingly, incorporation was also observed at 168.7 ppm
corresponding to the epoxyamide (Table 1), a metabolite that
frequently accompanies production of the azinomycins.1b We
were gratified to find that L-γ-hydroxyvaline 3 was unam-
biguously incorporated, substantiating its involvement in
either forming the epoxide directly (Figure 2, path B) or
generating an olefin where subsequent oxygen insertion
would give the epoxide (Figure 2, path C). As expected, only
the L-isomer 3 served as a substrate over its corresponding
D-isomer 3*, confirming the stereospecific nature of these
reactions. The site-specific incorporation of R-keto hydroxy
acid 5 was also observed and is suggestive that hydroxylation
of valine (to 3) precedes transimination. Reconstitution of
the enzymes involved in these transformations will, however,
be needed to rigorously establish this notion. The most
advanced putative precursor shown to be processed by the
azinomycin biosynthetic machinery was 3-methyl-2-ox-
obutenoic acid 12 negating direct formation of the epoxide
from the alcohol (Figure 2, path B). In contrast, isodehy-
drovaline 10 failed to incorporate, further substantiating the
order of biosynthetic steps (favoring Figure 2, path C), where
dehydration of the γ-alcohol to the double bond is suggested
to follow transimination.
Interestingly, neither of the epoxides showed incorporation
into the natural product. This is likely attributed to their
instability in aqueous medium, which increased dramatically
over time (Figure 1 of Supporting Information), with ring-
opened products and considerable lactonization occurring
over a 24 h period at room temperature (data not shown).
Notwithstanding the instability of epoxides 8 and 9, the
lifetimes of the other amino acid derivatives in aqueous
media were sufficient, under the conditions of our feeding
regimen (two separate and equal aliquots fed 24 h apart pH
7.1-7.5, 30 °C), to yield reliable incorporation data.6
The azinomycins are a structurally unique class of soil-
derived antitumor antibiotics that have shown promising
activity against carcinomas, reticulosarcoma, and terablas-
tomas in clinical trials.17 While DNA modifying agents (e.g.,
mitomycin C, calicheamin)18 are often a first line of defense
Figure 3. Biosynthetic route to the epoxide moiety in the azino-
mycins where failure to incorporate 11 suggests that 3-methyl-2-
oxobutenoic acid 12 is epoxidized to 8 which is then reduced to 9.
the enzymatic steps required to generate the epoxide fragment
occur prior to loading onto the NRPS machinery (invalidating
Figure 2, path A).
With the recent discovery of the azinomycin biosynthetic
cluster,20 it will now be possible to assign specific genes to
these and other enzymatic activities of the pathway. Such
experiments will pave the way for future genetic engineering
and/or chemoenzymatic manipulation of the biosynthetic
operon in the design of more effective anticancer agents.
Acknowledgment. We thank Dr. Chaomin Liu (former
postdoctoral researcher in the Watanabe laboratory) for
preliminary work on the synthesis and the American Cancer
Society and the Robert A. Welch Foundation (A-1587) for
financial support.
Supporting Information Available: Experimental details
and spectral data. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL8018852
(18) (a) Nygren, P.; Larsson, R. J. Intern. Med. 2003, 253, 46. (b)
Bernold, D. M.; Sinicrope, F. A. Clin. Gastroenterol. Hepatol. 2006, 4,
808. (c) van den Bent, M. J.; Hegi, M. E.; Stupp, R. Eur. J. Cancer 2006,
42, 582. (d) Molina, J. R.; Adjei, A. A.; Jett, J. R. Chest 2006, 130, 1211.
(19) (a) Coleman, R. S.; Tierney, M. T.; Cortright, S. B.; Carper, D. J.
J. Org. Chem. 2007, 72, 7726–7735. (b) Coleman, R. S.; Woodward, R. L.;
Hayes, A. M.; Crane, E. A.; Artese, A.; Ortuso, F.; Alcaro, S. Org. Lett.
2007, 9, 1891–1894. (c) LePla, R. C.; Landreau, C. A. S.; Shipman, M.;
Hartley, J. A.; Jones, G. D. D. Bioorg. Med. Chem. Lett. 2005, 15, 2861–
2864. (d) LePla, R. C.; Landreau, C. A. S.; Shipman, M.; Jones, G. D. D.
Org. Biomol. Chem. 2005, 3, 1174–1175. (e) Casely-Hayford, M. A.; Kerr,
N. O.; Smith, E.; Gibbons, S.; Searcey, M. Med. Chem. 2005, 1, 619–627.
(f) Casely-Hayford, M. A.; Pors, K.; James, C. H.; Patterson, L. H.; Hartley,
J. A.; Searcey, M. Org. Biomol Chem. 2005, 3, 3585–3589. (g) Hodgkinson,
T. J.; Kelland, L. R.; Shipman, M.; Suzenet, F. Bioorg. Med. Chem. Lett.
2000, 10, 239–241.
(20) Zhao, Q.; He, Q.; Ding, W.; Tang, M.; Kang, Q.; Yu, Y.; Deng,
W.; Zhang, Q.; Fang, J.; Tang, G.; Liu, W. Chem. Biol. 2008, 15, 693–
705.
(17) (a) Shimada, N.; Uekusa, M.; Denda, T.; Ishii, Y.; Iizuka, T.; Sato,
Y.; Hatori, T.; Fukui, M.; Sudo, M. J. Antibiot. 1955, 8, 67–76.
4818
Org. Lett., Vol. 10, No. 21, 2008