Directed Biosynthesis of Iso-aclacinomycins with Improved Anticancer Activity

Article abstract of DOI:10.1021/acs.orglett.9b04069

A four-enzyme catalyzed hydroxy regioisomerization of anthracycline was integrated into the biosynthetic pathway of aclacinomycin A (ALM-A), to generate a series of iso-ALMs via directed combinatorial biosynthesis combined with precursor-directed mutasynthesis. Most of the newly acquired iso-ALMs exhibit obviously (1-5-fold) improved antitumor activity. Therefore, we not only developed iso-ALMs with potential as clinical drugs but also demonstrated the utility of this tailoring tool for modification of anthracycline antibiotics in drug discovery and development.

Full text of DOI:10.1021/acs.orglett.9b04069

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Directed Biosynthesis of Iso-aclacinomycins with Improved
Anticancer Activity
Yu Hu,† Zhuan Zhang,† Yue Yin, and Gong-Li Tang*
State Key Laboratory of Bio-organic and Natural Products Chemistry, Center for Excellence in Molecular Synthesis, Shanghai
Institute of Organic Chemistry, University of Chinese Academy of Sciences (CAS), CAS, Shanghai 200032, China
S
* Supporting Information
ABSTRACT: A four-enzyme catalyzed hydroxy regioisomerization of anthracycline was integrated into the biosynthetic
pathway of aclacinomycin A (ALM-A), to generate a series of iso-ALMs via directed combinatorial biosynthesis combined with
precursor-directed mutasynthesis. Most of the newly acquired iso-ALMs exhibit obviously (1−5-fold) improved antitumor
activity. Therefore, we not only developed iso-ALMs with potential as clinical drugs but also demonstrated the utility of this
tailoring tool for modification of anthracycline antibiotics in drug discovery and development.
its analogs 2−4 (Figure 1A),2−4 has been marketed for the
he antitumor anthracyclines, used as single agents or in
treatment of stomach, lung, ovarian carcinoma; malignant
Tcombination therapy, have been introduced in clinical use
for around 50 years to treat a wide variety of cancers.1 Apart
lymphoma; and acute leukemia. To develope optimized drug
candidates, novel formulation and discovery of new derivatives
from the famous doxorubicin family (e.g., daunorubicin,
about anthracyclines have received continuous attention in the
idarubicin, epirubicin, and so on, Figure 1A), aclacinomycin
past 40 years.5−12
A (ALM-A, 1), isolated from Streptomyces galilaeus along with
It is well-known that biosynthesis of anthracyclines is guided
by the type II polyketide synthase (PKS) system, which usually
generates an anthracycline core with a C-4 hydroxy group in
the D-ring (Figure 1A).13 Nonetheless, kosinostatin (KST, 5),
a rare anthracycline antibiotic with one hydroxy group at C-1
instead of C-4, was isolated from marine Micromonospora sp.
TP-A0468 bearing remarkable structural features and strong
cytotoxicity against various cancer cell lines.14 Recently,
characterization of the KST biosynthetic pathway not only
established a hybrid type II PKS-nonribosomal peptide
synthetase (NRPS) platform15 but also discovered a four-
enzyme-catalyzed hydroxy regioisomerization: the 4-hydroxy-
anthraquinone intermediate 7 is hydroxylated by hydroxylases
KstA15/KstA16 to yield a 1,4-dihydroxy product 8, which
undergoes an asymmetric reduction−dearomatization cata-
lyzed by one short-chain dehydrogenase/reductase (SDR)
KstA11 and dehydration mediated by another SDR KstA10 to
afford the 1-hydroxy anthraquinone 9 (Figure 1B).16
Hydroxy and methoxy groups in the anthracyclines play a
critical role in their antitumor effects and point out a direction
in searching for new drug candidates.17 For example, idarubicin
Figure 1. Relative antitumor anthracyclines mentioned in this
research. (A) Chemical structures. (B) Previously characterized
pathway of hydroxy regioisomerization in KST biosynthesis.
Received: November 13, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b04069
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(6, Figure 1A), a clinical anticancer agent chemically derived
from daunorubicin by removal of the 4-methoxy group at C-4
in ring D, displays advantages over daunorubicin in both
bioactivity and cardiotoxicity.18 Nevertheless, the addition and
removal of the hydroxy group on the complex anthracycline
skeleton with high efficiency and regioselectivity is still
challenging. Accordingly, with a distinguished tailoring
modification involving hydroxy regioisomerization directly
acting on the previously seldom-touched anthracycline core
in hand, we anticipate that this four-enzyme cascade may serve
as a facilitation tool to modify other anthracyclines.
The following observations led to the hypothesis that the
four-enzyme system could be used in engineering the ALMs to
give iso-ALMs (C-4 deoxy-C-1 hydroxy-ALMs). On the one
hand, the clinically used ALM-A and ALM-B have a very
similar anthracycline core skeleton with KST. The only
difference is that ethyl group substitution occurs on C-9 in
ALMs while it is the methyl group in KST. On the other hand,
1-hydroxy ALMs are produced together with ALMs in the S.
galilaeus wild-type (WT) strain,3,4 implying the glycosyltrans-
ferases (GTs) of akn cluster probably have a wide substrate
scope toward the anthracycline skeleton. Finally, the AclR/
AclQ pair performing the transformation from 10 to 1119
displays about 60% identity with hydroxylases KstA15/Kst16
responsible for the first reaction in our regioisomerization,
which indicates that our four-enzyme system is probably
compatible with the ALMs biosynthesis.
deoxy-C-1 hydroxyaklavinone 12 (Figure 2B). Compounds
11 and 12 were then prepared by the large-scale reactions and
elucidated by NMR analysis and comparison to the NMR data
of 10 (Tables S4−S5 and Figures S6−S16). Our in vitro
results demonstrated that hydroxy regioisomerization of
aklavinone 10 can be achieved by our four-enzyme cascade
system to afford 12 effectively, which spurs us to carry out the
further in vivo engineering of ALM-A biosynthetic pathway to
generate iso-ALMs.
Based on this, we envisaged that expression of kstA15/A16
and kstA10/A11 in S. galilaeus WT would generate iso-ALMs.
Unfortunately, no anticipated compounds were detected in the
recombinant strain S. galilaeus::kstA15/A16/A10/A11 (Figure
S17), probably due to the preferred recognition of GTs toward
the natural substrate 10 rather than 12. Such a supposed
competition with the natural building block forced us to
perform a precursor directed mutasynthesis approach as a
roundabout strategy.
To overcome this, we first need to construct a deletion
mutant which completely abolishes the formation of 10 to
effectively incorporate its analogue 12. Previous studies have
revealed that aknB and aknC encode ketosynthase (KS)-α and
-β respectively, which play a key role in generating the core
structure of ALMs (Figure 3A).20 To obtain a desired
To prove our hypothesis, aklavinone 10 was first employed
as substrate to carry out the enzymatic reactions. After
fermentation of S. galilaeus MA144-M1 (ALMs producer),
ALMs were hydrolyzed with hydrochloric acid to afford 10
(Figure 2A), which was characterized by NMR spectra (Table
S3 and Figures S1−S5). When 10 was subjected to the
enzymes assay (Figure 2A), as we expected, KstA15/KstA16
could effectively catalyze the hydroxylation of 10 to produce
1,4-dihydroxy intermediate 11, which underwent one-pot
reaction involving successive reduction and dehydration
acted by KstA11 and KstA10 respectively to yield C-4
Figure 3. Summary of the route to generate iso-ALMs 14−17. (A)
Previously biosynthetic pathway of ALMs 1−4. (B) The strategies
used in this study: directed combinatorial biosynthesis combined with
precursor-directed mutasynthesis.
engineering organism for a precursor feeding experiment,
aknB and aknC were knocked out together via in-frame
deletion to yield mutant S. galilaeus TG1724 (ΔaknB/C, Mut-
B, Figure 3B and Figure S18), which abolished the production
of ALMs completely (Figure 4A-II). Meanwhile, we intended
to manipulate the ALMs gene cluster through combinatorial
biosynthesis strategies to obtain sufficient altered precursor 12,
which could be used as precursor to feed the Mut-B (Figure
3B). Given the fact that GT AknS and the partner protein
AknT were genetically and biochemically characterized to
transfer TDP-rhodosamine to 10 (Figure 3A),21−23 aknS and
aknT were deleted together to construct mutant S. galilaeus
TG1723 (ΔaknS/T, Mut-A, Figure 3B and Figure S19)
Figure 2. Biochemical characterization of KstA15/A16-KstA11-
KstA10 catalyzed hydroxy regioisomerization of aklavinone (10).
(A) Preparation of 10 via fermentation-chemical hydrolysis and
reactions catalyzed by KstA15/A16-KstA11/A10 using 10 as
substrate. (B) HPLC analysis of enzymatic reactions with UV
detected at 254 nm: (I) substrate 10 as control; (II) 10 with KstA15/
A16; (III) standard of 11; (IV) 11 with KstA11/A10. In these assays,
NADH is used as the cofactor, and the reaction time is 30 min.
B
DOI: 10.1021/acs.orglett.9b04069
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Figure 4. Identification of the newly generated iso-ALMs 14−17 as well as comparison of solubility and metabolic stability between them and their
analogues. (A) HPLC analysis of the relative metabolites with UV detected at 432 nm. (I) S. galilaeus MA144-M1 (WT); (II) Mutant-B: S. galilaeus
TG1724 (ΔaknB/C); (III) Mutant-A: S. galilaeus TG1723 (ΔaknS/T); (IV) Mutant-C: S. galilaeus TG1725 (ΔaknS/T::kstA10/A11); (V)
Mutant-B fed with compound 12. (B) Chemical structures of compounds 13−17. (C) Aqueous solubility assessed through miniaturized shake flask
approach. Error is SEM, n ≥ 3. (D) Metabolic stability was tested in human liver microsomes at 37 °C for 2 h, and the remaining percentage was
quantified relative to t0 using an internal standard. Error is SEM, n ≥ 3.
a
Table 1. Summary of Cytotoxic Activities against Cancer Cell Lines
a
A549, human lung carcinoma; Hela, human epitheloid cervix carcinoma cell; HepG2, human liver hepatocellular carcinoma cell; HL-60, human
promyelocytic acute leukemia cell; MCF-7, human breast adenocarcinoma cell; LOVO, human colonic carcinoma. Cell Counting Kit-8 method was
used to assay the cytotoxicity activity. The time of the drug acting on the cancer cell was 48 h.
intending to accumulate a large amount of 10. To our surprise,
although 10 was produced with low yield in this mutant, we
detected a dominant peak corresponding to 11 based on its
retention time and MS spectrum (Figure 4A-III), implying
AclR/AclQ could effectively convert 10 to 11 in Mut-A. When
kstA10 and kstA11 were subsequently introduced into this
mutant to yield S. galilaeus TG1725 (ΔaknS/T::kstA10/A11,
Mut-C, Figure 3B), production of 11 was abolished along with
the appearance of two new peaks (Figure 4A-IV). The major
one was determined to be desired compound 12 through LC-
MS analysis and comparing retention time with standard.
Further NMR characterization of the minor compound
demonstrated a new metabolite 13 (Figure 4B) almost
identical to 12 but with dehydroxylation at C-7 (Table S6
and Figure S20−S26), suggesting 13 was also a shunt product
as in the case of in vitro characterization of KstA10/KstA11 in
KST biosynthesis.16
C
DOI: 10.1021/acs.orglett.9b04069
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
After large-scale fermentation of Mut-C, a sufficient amount
of 12 was obtained. When 12 was fed into Mut-B, a series of
new compounds with the same molecular weight but different
retention time as ALMs were produced as expected (Figure 3B
and 4A-V). Subsequently, detailed MS and extensive NMR
analyses (Tables S7−S10 and Figures S27−S58) confirmed
that four major products of them were structurally assigned to
be iso-ALMs 14−17 (Figure 4B).
With four iso-ALMs 14−17 and corresponding natural
products 1−4 (Figures S59−S61) in hand, we first evaluated
their antitumor activities using various cancer cell lines
(Figures S62−S67). Despite the fact that 16 exhibited slightly
lower activities than 3 against human lung carcinoma (A549),
epitheloid cervix carcinoma (Hela), promyelocytic acute
leukemia (HL60), and colonic carcinoma (LOVO) cell lines,
the IC50 values of 15, 14, and 17 against all six cell lines were
lower than those of their natural analogues 1, 2, and 4 (Table
1), implying that the transfer of the hydroxy group from C-4 to
C-1 can improve the cytotoxic activities. Among these
compounds, 17 displayed the best antitumor activities in
most cases except for both human breast adenocarcinoma
(MCF-7) and LOVO cell lines, against which 15 performed
best (Table 1). In addition, 14 exhibited significantly improved
bioactivities against most tested cancer cell lines, especially the
MCF-7 cell line with almost an order of magnitude increase. In
summary, these results point at the significance of the hydroxy
regioisomerization of ALMs for enhanced activity.
Furthermore, we measured the solubility of eight com-
pounds using the miniaturized shake-flask method.24 Com-
pared to their natural counterparts, iso-ALMs 15 and 17 both
showed more than 5-fold increased aqueous solubility (Figure
4C and Figure S68). By carrying out a human liver microsome
assay,25 we then tested their metabolic stability. Both 16 and
17 clearly were more metabolically stable when compared to 3
and 4 respectively (Figure 4D). Finally, we occasionally
discovered the stability differences of ALMs and iso-ALMs in
DMSO after they were left at room temperature for 10 days. It
is truly inspiring to detect that all iso-ALMs showed
significantly greater stability than corresponding ALMs (Figure
S69). DMSO is often used as a solvent for tested compounds
in cellular and physiological experiments, so the preferable
stability in DMSO gives iso-ALMs greater potential in follow-
up studies. In consequence, iso-ALMs have improved activity,
aqueous solubility, and stability. Especially for 17, dramatic
improvement was observed in all tests with almost the best
anticancer activity, solubility, and stability among the eight
compounds, advancing it for clinical application and worthy of
further investigation.
In conclusion, guided by biochemical verification, we
developed an S. galilaeus mutant by combinatorial biosynthesis
to directly produce C-4 deoxy-C-1 hydroxyaklavinone (12).
Next, 12 was subsequently fed into another mutant which is
deficient in the key enzymes responsible for biosynthesis of 10
to generate iso-ALMs through precursor directed mutasyn-
thesis. Notably, the iso-ALMs show improved antitumor
activity, aqueous solubility, and stability, thereby hinting at
the potency for drug development. In recent years, it required
enormous effort in the combinatorial biosynthesis of type II
PKS to diversify complex aromatic polyketides via engineering
of a starter unit (tetracycline, chartreusin),26,27 nonenzyme
process (jadomycins, isatropolones),28,29 and glycoside (mi-
thramycin).30,31 In this work, we adopted an alternative
modification directly acting on the anthracycline core,
providing a favorable supplement to achieve more structural
diversification. The production of iso-ALMs utilized a
complementary strategy integrating combinatorial biosynthesis
and precursor-directed mutasynthesis. Moreover, 12 can be
considered as an important biosynthetic building block
immediately used in the engineering of the biosynthesis of
other anthracycline antibiotics. The resulting iso-ALMs
displayed partial improvement in antitumor activity. Taken
together, our findings established this hydroxy regioisomeriza-
tion as a rational tool, possessing potential to modify other
anthracyclines.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.orglett.9b04069.
Material and methods, supplementary tables and figures
(PDF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: gltang@sioc.ac.cn.
ORCID
Gong-Li Tang: 0000-0003-3149-4683
Author Contributions
†Y.H. and Z.Z. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are thankful for the supporting grants from the National
Key Research and Development Program of China
(2018YFA0901902), National Natural Science Foundation of
China (21632007 and 21621002), the Strategic Priority
Research Program of Chinese Academy of Sciences
(XDB20000000), and K. C. Wong Education Foundation.
REFERENCES
■
(1) Monneret, C. Eur. J. Med. Chem. 2001, 36, 483.
(2) Oki, T.; Matsuzawa, Y.; Yoshimoto, A.; Numata, K.; Kitamura,
I.; Hori, S. J. Antibiot. 1975, 28, 830.
(3) Oki, T.; Kitamura, I.; Yoshimoto, A.; Matsuszawa, Y.;
Shibamoto, N.; Ogasawa, T.; Inui, T.; Takamatsu, A.; Takeuchi, T.;
Masuda, T.; Hamada, M.; Suda, H.; Ishizuka, M.; Sawa, T.; Umezawa,
H. J. Antibiot. 1979, 32, 791.
(4) Oki, T.; Kitamura, I.; Matsuzawa, Y.; Shibamoto, N.; Ogasawa,
T.; Yoshimoto, A.; Inui, T.; Naganawa, H.; Takeuchi, T.; Umezawa,
H. J. Antibiot. 1979, 32, 801.
(5) Tikhomirov, A. S.; Shtil, A. A.; Shchekotikhin, A. E. Recent Pat.
Anti-Cancer Drug Discovery 2018, 13, 159.
(6) Shiokawa, T.; Hattori, Y.; Kawano, K.; Ohguchi, Y.; Kawakami,
H.; Toma, K.; Maitani, Y. Clin. Cancer Res. 2005, 11, 2018.
(7) Oki, T.; Yoshimoto, A.; Matsuzawa, Y.; Takeuchi, T.; Umezawa,
H. J. Antibiot. 1980, 33, 1331.
(8) Niemi, J.; Ylihonko, K.; Hakala, J.; Parssinen, R.; Kopio, A.;
Mantsala, P. Microbiology 1994, 140, 1351.
(9) Kim, H.-S.; Hong, Y.-S.; Kim, Y.-H.; Yoo, O.-J.; Lee, J.-J. J.
Antibiot. 1996, 49, 355.
(10) Kim, H.-S.; Kim, Y.-H.; Yoo, O.-J.; Lee, J.-J. Biosci., Biotechnol.,
Biochem. 1996, 60, 906.
̈
̈
̈
(11) Wang, Y.; Niemi, J.; Mantsala, P. FEMS Microbiol. Lett. 2002,
208, 117.
D
DOI: 10.1021/acs.orglett.9b04069
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(12) Ryu, J. S.; Lee, H. S.; Hong, Y.-S.; Lee, J. J.; Sohn, U. D.; Kim,
T. Y. Cancer Chemother. Pharmacol. 2006, 57, 811.
́
(13) Olano, C.; Mendez, C.; Salas, J. A. Nat. Prod. Rep. 2009, 26,
628.
(14) Igarashi, Y.; Higuchi, H.; Saito, N.; Oki, T.; Furumai, T. J.
Antibiot. 2002, 55, 128.
(15) Ma, H.-M.; Zhou, Q.; Tang, Y.-M.; Zhang, Z.; Chen, Y.-S.; He,
H.-Y.; Pan, H.-X.; Tang, M.-C.; Gao, J.-F.; Zhao, S.-Y.; Igarashi, Y.;
Tang, G.-L. Chem. Biol. 2013, 20, 796.
(16) Zhang, Z.; Gong, Y.-K.; Zhou, Q.; Hu, Y.; Ma, H.-M.; Chen, Y.-
S.; Igarashi, Y.; Pan, L.; Tang, G.-L. Proc. Natl. Acad. Sci. U. S. A.
2017, 114, 1554.
(17) Thompson, R. H. Naturally Occurring Quinones IV: Recent
Advances; Chapman & Hall: 1997; p 309.
(18) Minotti, G.; Menna, P.; Salvatorelli, E.; Cairo, G.; Gianni, L.
Pharmacol. Rev. 2004, 56, 185.
̈
̈
̈
(19) Beinker, P.; Lohkamp, B.; Peltonen, T.; Niemi, J.; Mantsala, P.;
Schneider, G. J. Mol. Biol. 2006, 359, 728.
̈
(20) Raty, K.; Kantola, J.; Hautala, A.; Hakala, J.; Ylihonko, K.;
̈
̈
̈
Mantsala, P. Gene 2002, 293, 115.
̈
̈
̈
̈
(21) Raty, K.; Kunnari, T.; Hakala, J.; Mantsala, P.; Ylihonko, K.
Mol. Gen. Genet. 2000, 264, 164.
(22) Lu, W.; Leimkuhler, C.; Gatto, G. J., Jr.; Kruger, R. G.;
Oberthur, M.; Kahne, D.; Walsh, C. T. Chem. Biol. 2005, 12, 527.
̈
(23) Leimkuhler, C.; Fridman, M.; Lupoli, T.; Walker, S.; Walsh, C.
T.; Kahne, D. J. Am. Chem. Soc. 2007, 129, 10546.
(24) Bingham, T. W.; Hernandez, L. W.; Olson, D. G.; Svec, R. L.;
Hergenrother, P. J.; Sarlah, D. J. Am. Chem. Soc. 2019, 141, 657.
(25) Venkataramanan, R.; Zang, S.; Gayowski, T.; Singh, N.
Antimicrob. Agents Chemother. 2002, 46, 3091.
̈
̈
(26) Ueberschaar, N.; Xu, Z.; Scherlach, K.; Metsa-Ketela, M.;
̈
Bretschneider, T.; Dahse, H.-M.; Gorls, H.; Hertweck, C. J. Am. Chem.
Soc. 2013, 135, 17408.
̌
̌
̌
̌
(27) Lesnik, U.; Lukezic, T.; Podgorsek, A.; Horvat, J.; Polak, T.;
̌
Sala, M.; Jenko, B.; Harmrolfs, K.; Ocampo-Sosa, A.; Martínez-
̌
Martínez, L.; Herron, P. R.; Fujs, S.; Kosec, G.; Hunter, I. S.; Muller,
̈
R.; Petkovic, H. Angew. Chem., Int. Ed. 2015, 54, 3937.
́
(28) Robertson, A. W.; Martinez-Farina, C. F.; Smithen, D. A.; Yin,
H.; Monro, S.; Thompson, A.; McFarland, S. A.; Syvitski, R. T.;
Jakeman, D. L. J. Am. Chem. Soc. 2015, 137, 3271.
̈
(29) Cai, X.; Shi, Y.-M.; Pohlmann, N.; Revermann, O.; Bahner, I.;
Pidot, S. J.; Wesche, F.; Lackner, H.; Buchel, C.; Kaiser, M.; Richter,
̈
C.; Schwalbe, H.; Stinear, T. P.; Zeeck, A.; Bode, H. B. Angew. Chem.,
Int. Ed. 2017, 56, 4945.
(30) Wang, G.; Pahari, P.; Kharel, M. K.; Chen, J.; Zhu, H.; Van
Lanen, S. G.; Rohr, J. Angew. Chem., Int. Ed. 2012, 51, 10638.
́
́
́
(31) Nun
̃
ez, L. E.; Nybo, S. E.; Gonzalez-Sabín, J.; Perez, M.;
́
Menendez, N.; Brana, A. F.; Shaaban, K. A.; He, M.; Morís, F.; Salas,
̃
́
J. A.; Rohr, J.; Mendez, C. J. Med. Chem. 2012, 55, 5813.
E
DOI: 10.1021/acs.orglett.9b04069
Org. Lett. XXXX, XXX, XXX−XXX