New ansamycin derivatives generated by simultaneous mutasynthesis

Article abstract of DOI:10.1021/ol5035639

The conversion from triene- to diene-typed ansamycins is clarified step by step in Streptomyces seoulensis IFB-A01. Such an intertype convertibility is adopted to establish for the first time the simultaneous mutasynthesis of both types of C17-benzene ans

Full text of DOI:10.1021/ol5035639

Letter
pubs.acs.org/OrgLett
New Ansamycin Derivatives Generated by Simultaneous
Mutasynthesis
Ya Nan Song, Rui Hua Jiao, Wen Jing Zhang, Guo Yan Zhao, Huan Dou, Rong Jiang, Ai Hua Zhang,
Ya Yi Hou, Shu Feng Bi, Hui Ming Ge,* and Ren Xiang Tan*
Institute of Functional Biomolecules, State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing, 210093,
China
S
* Supporting Information
ABSTRACT: The conversion from triene- to diene-typed
ansamycins is clarified step by step in Streptomyces seoulensis
IFB-A01. Such an intertype convertibility is adopted to
establish for the first time the simultaneous mutasynthesis of
both types of C17-benzene ansamycins (C17BAs). Three of
the newly generated unnatural compounds showed potent
cytotoxicity.
wing to the unpredictable chemodiversity, natural
been made to generate new C17BAs through “mutasynthetic
Oproducts (viz., secondary metabolites from the biological operations” of the C-11 substituted motif, which is another
determinant for bioactivities of this class of compounds.9 A
viewpoint) have occupied a unique region of chemical space,
previous failure was reported for the restoration of ansatrienin
production by a bacterial strain deprived of the gene (chcA)
encoding 1-cyclohexenylcarbonyl coenzyme A (CHC-CoA)
reductase.10a As illustrated by the trienomycin A (1) structure
(Figure 1b), a cyclohexanecarboxylic acid (CHC)-derived side
chain is anchored via an L-alanine residue to the C-11 position
of a macrolactam. A previous study on the biosynthesis of
ansatrienin A in Streptomyces collinus has revealed the CHC-
CoA is the direct precursor of the cyclohexanate moiety of the
ansatrienin A, and five genes including ans J−M and chcA were
identified to be involved in the formation of the CHC-CoA.10
Thus, the CHC-CoA related biosynthetic genes could be used
as an important genetic target for mutasynthesis of trienomycin
A (1). Additionally, the coproduction of trienomycin A (1) and
benzoxazomycin (2) by Streptomyces sp. MJ672-m311 encour-
aged us to hypothesize that the conversion from triene- to
diene-typed ansamycins might occur in some microbes.
Obviously, this intertype convertibility would permit the
simultaneous mutasynthesis of more (at least two) C17BA
derivatives upon supplementation of a single substrate
(mutasynthon). To test this hypothesis, here we demonstrated
for the first time that conversion from triene to diene was due
to the tandem oxidation and subsequent [4 + 2] intramolecular
Diels−Alder cyclization (Figure 1b). Through supplementing
different CHC-like unnatural precursors, the mutant strain of S.
seoulensis, which lacked the ability to biosynthesize CHC-CoA,
was able to simultaneously produce both triene- and diene-type
anasamycins with the modifiable C-11 dipeptidyl motif (Figure
1). Some of the newly generated unnatural compounds showed
potent cytotoxicity against HepG2 and MCF-7 tumor cell lines.
which is not yet substantially overlapped by commercialized
chemicals/pharmaceuticals and other small molecule libraries.1
Source analysis of approved small molecule drugs indicates that
polyketides are approximately five times more likely to possess
drug activity than other natural product categories, and
macrocyclic polyketides seem to have even higher activity.2
Ansamycins are a unique class of bacterial macrocyclic
polyketides with a broad range of bioactivities.2d,3 As a subclass
of this polyketide family, the C17-benzene ansamycin (C17BA)
such as trienomycin A (1) is architecturally distinct in its
macrolactam formed through an amide linkage to an aryl ring
(Figure 1a).4 The fascinating chemistry and promising
bioactivity of C17BAs have driven collectively the advancement
of the synthetic methodology under the necessity of supplying
and modifying such a family of complex compounds.2d,5 Most
structurally complicated ansamycins continue to be prepared
through fermentation with or without subsequent semi-
synthesis.2d Moreover, chemical derivatization approaches are
limited by the complex functionality of the ansamycin scaffold.
As a result of their synthetic complexity, our understanding of
the structure−activity relationships of the ansamycins remains
incomplete.6 Another bottleneck remains in the drug discovery
owing to the poor structural diversity of compound libraries.7 A
particular case in this regard is the C17BA polyketides that
consist surprisingly of only one diene- and 43 triene-typed
ansamycins (Figure 1a).
To diversify ansamycins, the mutasynthetic approach has
been established by, but so far limited to, the supplementation
of 3-amino-5-hydroxybenzoic acid (AHBA)-derived “unnatural
substrates” (technically called mutasynthon) into the bio-
synthetic machinery of mutated bacterial producers of C15-
benzene ansamycins (C15BAs, strikingly different from
C17BAs) such as geldanamycin.8 However, few attempts have
Received: December 9, 2014
© XXXX American Chemical Society
A
DOI: 10.1021/ol5035639
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Figure 1. Frameworks and triene-to-diene conversions of C17-benzene ansamycins (C17BAs). Transformation of 1 to 2 via 3 and 4 found in the
wild-type strain of S. seoulensis IFB-A01 was repeated in the culture of the CHC biosynthetic genes knockout strain by feeding cyclohexanecarboxylic
acid (CHC). Conversions of 1a−1d to 2a−2d via 3a−3d and 4a−4d were ascertained in cultures of the mutant strain (IFB-A01-C) by
supplementing with cyclopentanecarboxylic, cyclobutanecarboxylic, cyclopropanecarboxylic, and 2-cyclopentylacetic acids, respectively.
Our endeavor was initiated by screening our microbial library
for the strain capable of coproducing triene- and diene-typed
ansamycins. From our microbial library, S. seoulensis strain IFB-
A01 isolated from the gut of Penaeus orientalis (the Chinese
white shrimp) produced both 1 and 2 when fermented using an
optimized medium (Figure 1, see “Supplementary Methods” in
the Supporting Information). To test the convertibility from 1
to 2, S. seoulensis IFB-A01 was regrown with any fluctuation in
the medium monitored by LC-MS analysis. Compound 1
became detectable in the culture on the fourth day followed by
a two-day plateau reflecting a continued accumulation of the
metabolite. After that, the gradual increase of 2 was
accompanied by the decrease of 1 in the culture, confirming
the conversion of 1 into 2 (Figure S1). To understand the
transformation step by step, a scaled-up culture of S. seoulensis
was subjected to an LC-MS guided fractionation of
intermediates formed in the biosynthetic assembly line for 1.
As expected, mycotrienins II (3) and I (4)12 were isolated from
the culture, and both could be the intermediates generated
upon the production of 2 from 1 (Figure 1b). To verify the
assumption, compound 1 was treated with the cell-derived
protein to produce 3 (Figure 2). In view of the redox
equilibrium between hydroquinone and benzoquinone, the
conversion of 3 into 4 was hypothesized to occur in the
presence of oxygen or air. This anticipation was validated by the
conversion of 3 into 4 in a yield of 94% in air-exposed
phosphate buffer (pH = 7) at 28 °C for 10 h (Figure 2).
Interestingly, the subsequent [4 + 2] Diels−Alder cyclization of
Figure 2. LC-MS evaluation for the production of 1−4 in cultures and
reaction solution. (a) Convertibility tests of 1 by treating for 2 or 4 h
with intercellular proteins of the WT strain, and of 3 into 2 via 4 in
PBS buffer (pH = 7) within 0, 5, and 10 h. (b) Test for
transformability of 4 into 2 in different pH solution buffers within
10 h. (c) Medium-dependence of transformability of 4 into 2 (tested
separately in PBS buffer (pH = 7), tetrahydrofuran, and methanol).
4 was shown to proceed at room temperature in phosphate
buffers between pH 5−9 (Figure 2).
Few attempts have been made to generate new C17BAs
despite the incorporatability of 3-amino-5-hydroxybenzoic acid
(AHBA)-derived mutasynthons into C15BAs such as geldana-
mycin.8a−c In the biosynthesis of trienomycin A (1) and
benzoxazomycin (2), CHC-CoA is proposed as the source of
B
DOI: 10.1021/ol5035639
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
the CHC-derived side chain in the C-11 position. Thus, the
mutant strain lacking the ability to biosynthesize CHC-CoA
could coproduce triene (such as 1) and diene (such as 2)-like
compounds by adding a different CHC-like acid to its medium.
To isolate the biosynthetic gene cluster for 1 and 2, a fosmid
library of S. seoulensis IFB-A01 was constructed and screened by
a PCR method using two pairs of primer F/R-AHBA and F/R-
chcA (see “Cloning and analysis of the partial ansatrienin
biosynthetic gene cluster” in the Supporting Information). A
partial biosynthetic gene cluster of trienomycin was cloned
from one fosmid. The gene product showed high identity/
similarity to that of ansatrienin A from S. collinus (Table
S2).10b,c
Table 1. In Vitro Cytotoxicity IC50 (μM) of 1, and 1a−1d
against HepG2 and MCF-7
compounds
HepG2
>20
MCF-7
>20
1
1a
12.18 0.50
8.39 0.14
>20
6.74 1.21
7.33 0.27
>20
1b
1c
1d
6.81 0.21
1.49 0.08
>20
a
doxorubicin
2.18 0.19
a
Positive control; HepG2: human hepatic carcinoma cell; MCF-7:
human breast cancer cell. Results are expressed as mean SEM (n
= 3).
To disrupt the biosynthetic pathway of CHC-CoA, four
genes (ansA1−A4) involved in the CHC-CoA biosynthesis
were inactivated in S. seoulensis IFB-A01 by double homologous
recombination (Figure S2). The desired mutant (IFB-A01-C)
was selected on the basis of an apramycin-sensitive and
kanamycin-resistant phenotype and was further verified by a
polymerase chain reaction followed by diagnostic digestion of
restriction enzymes (Figure S3). The mutant strain is unable to
produce 1−4, but the production of 1−4 can be restorable by
chemical complementation of CHC (Figure S2, see “Supple-
mentary Methods” in the Supporting Information).
To ascertain its mutasynthon scope, the mutant strain, IFB-
A01-C, was cultured separately with a diversity of small
molecule organic acids. As anticipated, compounds 1a−1d and
2a−2d were afforded by supplementing the mutant cultures
with cyclopentanecarboxylic, cyclobutanecarboxylic, cyclo-
propanecarboxylic, and 2-cyclopentylacetic acids, respectively
(see “Structure Elucidation” in the Supporting Information).
However, no mutasynthetic product could be detected if
supplementing bulkier (cycloheptanecarboxylic and 2-cyclo-
hexylacetic acids) or heterocyclic mutasynthons (piperidine-4-
carboxylic, tetrahydro-2H-pyran-4-carboxylic, tetrahydrofuran-
3-carboxylic, and pyrrolidine-3-carboxylic acids). These data
showed that the mutasynthons acceptable for the mutant strain
seem limited to small molecule (≤7 carbons) and heteroatom-
free fatty acids. The finding agrees with the substitution profile
of identified natural ansamycins whose C-11 acyl chains are
restricted to cyclohexanecarbonyl, cyclohex-1-enecarbonyl, 4-
methylpentanoyl, 3-methylbutanoyl, 2-methylbutanoyl, and 2-
methylbut-2-enoyl groups.9 Surprisingly, the mutant strain IFB-
A01-C produces compounds 1a and 2a with a cyclopent-1-ene-
1-carbonyl motif, while culturing with supplementation of
cyclopentanecarboxylic acid. This suggested that the mutant
strain might have produced the α,β-desaturase capable of
transforming cyclopentanecarboxylic acid into cyclopent-1-ene-
1-carboxylic acid.
The chemodiversity of (unnatural) natural products
determines the applicability of the compound library as lead
molecule sources for new drugs and agrochemicals.13 Actino-
mycetes have produced a magnitude of bioactive molecules, but
the characterizable new metabolites from these microbes
generally become fewer and fewer after long-time intense
investigations. To deal with the frustration, a combined
mutasynthesis/semisynthesis approach was introduced to
generate a more diversified compound library using the gene
blocked actinomyces mutant.8a,d,e,14 This work has addressed
the intertype conversion of ansamycins in S. seoulensis IFB-A01,
a new microbial producer that can construct both types of
ansamycin macrolactams. To our surprise, the triene-to-diene
transformation of ansamycins includes two automatic reactions
following the 19-oxygenation (Figure 1b). To overcome the
long existing challenge in mutasynthetic operation of the C-11
motifs, our deletion of the CHC biosynthetic gene cluster gave
the desired mutant strain that was able to produce asamycins
upon exposure to exogenous CHC and other small molecule
acids. This substantiates the simultaneous mutasynthesis of
both types of ansamycins. In concert with the bioassay
procedure, the bioactive molecules were identified from the
ansamycin-based compounds. In aggregation, the simultaneous
mutasynthesis approach established here for the first time is
effective in generating multiply diversified compound libraries
required for the drug discovery effort.
ASSOCIATED CONTENT
■
S
* Supporting Information
Complete description of methods, additional tables, and figures,
including full NMR data for all compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
The production of low-abundance compounds 2a, 3a−3d,
and 4a−4d was ascertained by the high resolution LC-MS
analysis in the cultures of the mutant strain IFB-A01-C (Table
S9) although the scarcity of these compounds in cultures did
not facilitate any fractionation of samples in the study. Despite
the limited number of molecules of the ansamycin analogue
library, the new triene-ansamycin members 1a, 1b, and 1d were
found to be more cytotoxic against HepG2 and MCF-7 tumor
cell lines than trienomycin A (1) (Table 1), and the newly
characterized diene-typed ansamycins 2b and 2c become more
inhibitory against the lipopolysaccharide-induced intracellular
IL-6 production of murine macrophage cell line RAW264.7
(Table S10).
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: hmge@nju.edu.cn (H.M.G.).
*E-mail: rxtan@nju.edu.cn (R.X.T.).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was cofinanced by the NSFC (81172948, 81121062,
21132004, 21302091, and 21072092) and MOST grants
(2013AA092901, 2013AA092903, and 2012ZX09502001-00).
C
DOI: 10.1021/ol5035639
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
REFERENCES
■
(1) Cherblanc, F. L.; Davidson, R. W.; Di Fruscia, P.;
Srimongkolpithak, N.; Fuchter, M. J. Nat. Prod. Rep. 2013, 30, 605−
624.
(2) (a) Newman, D. J.; Cragg, G. M. J. Nat. Prod. 2012, 75, 311−335.
(b) Lachance, H.; Wetzel, S.; Kumar, K.; Waldmann, H. J. Med. Chem.
2012, 55, 5989−6001. (c) Marsault, E.; Peterson, M. L. J. Med. Chem.
2011, 54, 1961−2004. (d) Del Valle, D. J.; Krische, M. J. J. Am. Chem.
Soc. 2013, 135, 10986−10989.
(3) Lourenco̧ , A. M.; Ferreira, L. M.; Branco, P. S. Curr. Pharm. Des.
2012, 18, 3979−4046.
(4) (a) Umezawa, I.; Funayama, S.; Okada, K.; Iwasaki, K.; Satoh, J.;
Masuda, K.; Komiyama, K. J. Antibiot. 1985, 38, 699−705. (b) Kim, C.
G.; Kirschning, A.; Bergon, P.; Zhou, P.; Su, E.; Sauerbrei, B.; Ning, S.;
Ahn, Y.; Breuer, M.; Leistner, E.; Floss, H. G. J. Am. Chem. Soc. 1996,
118, 7486−7491.
(5) (a) Smith, A. B.; Barbosa, J.; Wong, W.; Wood, J. L. J. Am. Chem.
Soc. 1996, 118, 8316−8328. (b) Rossle, M.; Del Valle, D. J.; Krische,
̈
M. J. Org. Lett. 2011, 13, 1482−1485.
(6) Kawamura, T.; Tashiro, E.; Yamamoto, K.; Shindo, K.; Imoto, M.
J. Antibiot. 2008, 61, 303−311.
(7) (a) Wilde, F.; Link, A. Expert Opin. Drug Discovery 2013, 8, 597−
606. (b) Bruns, R. F.; Watson, I. A. J. Med. Chem. 2012, 55, 9763−
9772.
(8) (a) Eichner, S.; Knobloch, T.; Floss, H. G.; Fohrer, J.; Harmrolfs,
K.; Hermane, J.; Schulz, A.; Sasse, F.; Spiteller, P.; Taft, F.; Kirschning,
A. Angew. Chem., Int. Ed. 2012, 51, 752−757. (b) Eichner, S.; Eichner,
T.; Floss, H. G.; Fohrer, J.; Hofer, E.; Sasse, F.; Zeilinger, C.;
Kirschning, A. J. Am. Chem. Soc. 2012, 134, 1673−1679. (c) Franke, J.;
Eichner, S.; Zeilinger, C.; Kirschning, A. Nat. Prod. Rep. 2013, 30,
1299−1323. (d) Harmrolfs, K.; Mancuso, L.; Drung, B.; Sasse, F.;
Kirschning, A. Beilstein J. Org. Chem. 2014, 10, 535−543. (e) Knobloch,
T.; Harmrolfs, K.; Taft, F.; Thomaszewski, B.; Sasse, F.; Kirschning, A.
ChemBioChem 2011, 12, 540−547.
(9) Wrona, I. E.; Agouridas, V.; Panek, J. S. C. R. Chim. 2008, 11,
1483−1522.
(10) (a) Wang, P.; Denoya, C. D.; Morgenstern, M. R.; Skinner, D.
D.; Wallace, K. K.; Digate, R.; Patton, S.; Banavali, N.; Schuler, G.;
Speedie, M. K.; Reynolds, K. A. J. Bacteriol. 1996, 178, 6873−6881.
(b) Cropp, T. A.; Wilson, D. J.; Reynolds, K. A. Nat. Biotechnol. 2000,
18, 980−983. (c) Chen, S.; von Bamberg, D.; Hale, V.; Breuer, M.;
Hardt, B.; Muller, R.; Floss, H. G.; Reynolds, K. A.; Leistner, E. Eur. J.
̈
Biochem. 1999, 261, 98−107.
(11) Hosokawa, N.; Naganawa, H.; Hamada, M.; Iinuma, H.;
Takeuchi, T.; Tsuchiya, K. S.; Hori, M. J. Antibiot. 2000, 53, 886−894.
(12) (a) Sugita, M.; Natori, Y.; Sasaki, T.; Furihata, K.; Shimazu, A.;
Seto, H.; Otake, N. J. Antibiot. 1982, 35, 1460−1466. (b) Sugita, M.;
Sasaki, T.; Furihata, K.; Seto, H.; Otake, N. J. Antibiot. 1982, 35,
1467−1473.
(13) (a) Cragg, G. M.; Newman, D. J. Biochim. Biophys. Acta 2013,
1830, 3670−3695. (b) Cragg, G. M.; Grothaus, P. G.; Newman, D. J.
Chem. Rev. 2009, 109, 3012−3043.
(14) (a) Shier, W. T.; Rinehart, K. L., Jr.; Gottlieb, D. Proc. Natl.
Acad. Sci. U.S.A. 1969, 63, 198−204. (b) Mancuso, L.; Knobloch, T.;
Buchholz, J.; Hartwig, J.; Moller, L.; Seidel, K.; Collisi, W.; Sasse, F.;
̈
Kirschning, A. Chem.Eur. J. 2014, 20, 17541−17551. (c) Kirschning,
A.; Hahn, F. Angew. Chem., Int. Ed. 2012, 51, 4012−4022.
D
DOI: 10.1021/ol5035639
Org. Lett. XXXX, XXX, XXX−XXX