Elucidation of the Herbicidin Tailoring Pathway Offers Insights into Its Structural Diversity

Article abstract of DOI:10.1021/acs.orglett.9b00066

The biosynthetic gene clusters for herbicidins (hbc) and aureonuclemycin (anm) were identified in Streptomyces sp. KIB-027 and Streptomyces aureus, respectively. The roles of genes possibly involved in post-core-assembly steps in herbicidin biosynthesis i

Full text of DOI:10.1021/acs.orglett.9b00066

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Elucidation of the Herbicidin Tailoring Pathway Offers Insights into
Its Structural Diversity
Hai-Xue Pan,† Zhang Chen,‡ Tianfang Zeng,∥ Wen-Bing Jin,† Yujie Geng,§ Geng-Min Lin,‡ Juan Zhao,†
Wei-Tao Li,† Zijun Xiong,⊥ Sheng-Xiong Huang,⊥ Xin Zhai,∥ Hung-wen Liu,*,‡,§
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
‡Department of Chemistry and §Division of Chemical Biology and Medicinal Chemistry, College of Pharmacy, University of Texas
at Austin, Austin, Texas 78712, United States
∥Key Laboratory of Structure-Based Drug Design & Discovery, Ministry of Education, Shenyang Pharmaceutical University,
Shenyang 110016, China
⊥State Key Laboratory of Phytochemistry and Plant Resources in West China, Center for Excellence in Molecular Plant Sciences,
Kunming Institute of Botany, CAS, Kunming 650201, China
S
* Supporting Information
ABSTRACT: The biosynthetic gene clusters for herbicidins
(hbc) and aureonuclemycin (anm) were identified in
Streptomyces sp. KIB-027 and Streptomyces aureus, respectively.
The roles of genes possibly involved in post-core-assembly
steps in herbicidin biosynthesis in these clusters and a related
her cluster were studied. Through systematic gene deletions,
structural elucidation of the accumulated intermediates in the
mutants, and in vitro verification of the encoded enzymes, the
peripheral modification pathway for herbicidin biosynthesis is
now fully established.
erbicidins (HBCs, 1−8) are undecose nucleoside
Hantibiotics isolated from different strains of Streptomyces
and exhibit a wide range of biological activities, including
herbicidal, antibacterial, antifungal, and antiparasitic activ-
ities.1−9 They share a common core but differ in their
methylation patterns at C2′-OH and C11′-OH and the
acylation mode at C8′-OH (Figure 1). In contrast,
aureonuclemycin (ANM, 9) isolated from Streptomyces aureus
var. suzhoueusis is unique because it bears only the bare
nucleoside core of 1−8 (Figure 1).10
Because of their complex chemical structures and diverse
biological activities, HBCs have attracted considerable
attention as targets for chemical synthesis.11−13 However, the
biosynthetic pathways for HBCs were unknown until a recent
report that described the identification of the biosynthetic gene
cluster (BGC) her in Streptomyces sp. L-9-10 and the initial
Figure 1. Structures of HBCs. Compounds 1, 2, 7, and 10 were
characterization of the biosynthetic pathway for HBC A (1).14
isolated from Streptomyces sp. KIB-027. 3, 5, 6, 8, and 11 were isolated
Further investigation of the assembly of the tricyclic undecose
core may reveal new mechanisms for C−C bond formation,15
and elucidation of the peripheral modifications is crucial for
diversification of HBCs through pathway engineering. Herein,
two BGCs for HBCs (hbc) and ANM (anm) were identified
from Streptomyces sp. KIB-027 and S. aureus, respectively. The
entire peripheral modification pathway of HBCs was identified
by gene-deletion experiments, structural elucidation of
from the mutants. 9 was isolated from Streptomyces lividans
anmBCDE.
Received: January 7, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b00066
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
accumulated intermediates, in vitro characterizations of the
expressed enzymes, and biochemical verification of their
catalytic functions.
Meanwhile, genome mining for an analogous hbc/her cluster
in S. aureus, which produces only 9, was also conducted to
identify the essential genes for core assembly. As a result, a
putative ANM gene cluster (anm) containing four genes
(anmB/C/D/E) homologous to both hbcB/C/D/E and her4/
5/7/6 was found (Figure 2A and Table S2). Interestingly,
BLAST analysis using hbcB/C/D/E as a query showed the
existence of more than 20 hbc/her/anm-like clusters in bacteria
(Figure S23). To verify the biosynthetic ability of the anm
cluster, the anmB−E locus was cloned into pSET152
downstream to an ermEp* promoter for heterologous
expression in S. lividans. LC−MS analysis showed that the
anm-expressing strain S. lividans anmBCDE indeed produced
ANM (9) as indicated by LC−MS (Figure S24) and confirmed
by NMR analysis of the isolated product (Figures S25 and
S26).12
Streptomyces sp. KIB-027 was isolated from soil samples
collected at Gaoligong Mountain, Yunnan province, China.
Through LC−MS analysis of its fermentation broth, four
products with UV spectra and molecular weights identical to
those of the known HBCs were detected (Figure S1).
Subsequent large scale fermentation of KIB-027 led to the
isolation of three known HBC congeners, HBC A (1), HBC B
(2), and HBC H (7), and one new compound, HBC M (10)
(Figure 1). The structures of 1, 2, and 7 were confirmed by
comparing the NMR data with those reported previously
(Figures S2−S13). The structure of 10 was determined by
analysis of its HRMS, one-dimensional, and two-dimensional
NMR data (Table S1 and Figures S14−S22).
Comparison of the results presented above helped to
distinguish the enzymes catalyzing skeleton formation (such
as those in anm) from those involved in post-core-assembly
modifications. It was thus hypothesized that the remaining
genes hbcA/F/G/H/I in the hbc cluster and their counterparts
her3/8/10/9/11 in the her cluster are responsible for the
peripheral modifications of HBCs. These genes in KIB-027
were each knocked out by in-frame deletion to study their
functions. While KAS III, HbcA was originally thought to
catalyze the esterification reaction at C8′-OH,16−18 the hbcA-
deletion mutant (ΔhbcA) still produced HBCs with yields
comparable to those generated by the wild type strain (Figure
2B, i and ii). Hence, hbcA is unlikely to be involved in the
biosynthesis of HBCs. Another functionally unattributed
enzyme, the serine hydrolase HbcH (Her9), may instead be
responsible for C8′-OH acylation.
Next, the genome of KIB-027 was sequenced and a BGC
(hbc) similar to the her cluster found in Streptomyce sp. L-9-
1014 was located. The hbc cluster contains genes encoding two
methyltransferases (MTs, HbcF and HbcG) and a cytochrome
P450 monooxygenase (HbcI) that may be responsible for the
late-stage decoration of 1. The cluster also contains 12 other
genes encoding a ketoacyl-ACP synthase III (KAS III, HbcA),
an S-adenosylhomocysteine (SAH) hydrolase (HbcB), two
NAD-binding oxidoreductases (HbcC and HbcD), a B12-
dependent radical S-adenosyl-L-methionine (SAM) enzyme
(HbcE), a serine hydrolase (HbcH), a short-chain dehydro-
genase (incomplete, HbcJ), and five regulatory proteins
(HbcR1−R5) (Figure 2A and Table S2).
As predicted, the production of 1, 2, 7, and 10 was
completely abolished when hbcH was deleted (ΔhbcH).
Similar outcomes were also observed for other gene knockout
mutants (ΔhbcF, ΔhbcG, and ΔhbcI). Interestingly, some new
metabolites, which are likely HBC congeners, were produced
by these mutants (Figure 2B, iii−vi). All of these compounds
were isolated and structurally characterized by HRMS and
NMR analysis. Specifically, the ΔhbcF mutant produced HBC
G (6) with a tiglyl group at C8′ (see Figure 2B, iii, Table S3,
and Figures S27−S34); the ΔhbcG mutant mainly produced
HBC K (8) having a tiglyl group at C8′ and a methoxyl group
at C11′ (Figure 2B, iv, Table S4, and Figures S35−S42); the
ΔhbcH mutant produced HBC C (3) bearing a methoxyl
group on C11′ but lacking an acyl group at C8’ (Figure 2B, v,
Table S5, and Figures S43−S50); and the ΔhbcI mutant
produced HBC F (5), containing a tiglyl group at C8′ and two
methoxyl groups at C2′ and C11′ (Table S6 and Figures S51−
S58), and a small amount of 8 (Figure 2B, vi). Collectively,
these results indicated that HbcF and HbcG are responsible for
the methylation of C11′-OH (6 → 8) and C2′-OH (8 → 5),
respectively, HbcH is associated with the uploading of the
tiglyl group at C8′ (9 → 6), and HbcI catalyzes the
hydroxylation of the tiglyl group of 5 to give 1.
In addition to HBC K (8), the ΔhbcG mutant also produced
another minor product, 11 (Figure 2B, iv), whose molecular
weight is greater than that of 8 by 16 Da. Compound 11 was
later determined to be a new HBC K (8) analogue carrying a
hydroxyl group at C5″ of the tiglyl moiety and was designated
as HBC N (Figure 1, Table S7, and Figures S59−S66). Further
in vivo experiments showed that the double-deletion mutant of
hbcG and hbcI (ΔhbcGI) produced only 8 (Figure S67). Thus,
Figure 2. (A) Organizations of the hbc, her, and anm biosynthetic
gene clusters. (B) HPLC analyses of fermentations with Streptomyces
sp. KIB-027 the wild type strain and the mutants.
B
DOI: 10.1021/acs.orglett.9b00066
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
in addition to its primary function of catalyzing the
hydroxylation of 5 (5 → 1), HbcI can also process 8 (8 →
11), albeit less effectively. On the contrary, compound 6 is not
a substrate of HbcI because no hydroxylated product of 6 was
produced by the ΔhbcF mutant. Hence, methylation of the
C11′ hydroxyl group of 6 (6 → 8) appears to be a prerequisite
for HbcI-catalyzed hydroxylation. Further methylation of the
C2′ hydroxyl group of 8 (8 → 5) renders it a better substrate
for HbcI.
On the basis of the in vivo studies mentioned above, HbcH/
Her9-mediated esterification (9 → 6) and HbcF/Her8-
catalyzed methylation (6 → 8) are expected to occur early
in the post-core-assembly process. However, it was puzzling
that neither the ΔhbcF mutant nor the ΔhbcH mutant
produced the presumed precursor 9. To address such
obscurity, these reactions were then investigated in vitro.
HbcF/G/H and the HbcH homologue Her9 were each
overexpressed in Escherichia coli and purified as N-His6-tagged
proteins (Figures S68 and S82). The Her8/10 proteins were
obtained as previously described,14 but an attempt to obtain
soluble HbcI was unsuccessful. The methylation activity of
HbcF and Her8 was tested using 6, and conversion to 8 was
observed in both cases (Figure 3 , i and ii, Figure S69 for a
different HPLC method, and Figure S86). HbcF could also
methylate 9 to give 3 (Figure 3, iii and iv). The capability of
HbcF to convert 9 to 3 could explain why 9 was not detected
in the fermentation of the ΔhbcH mutant. Interestingly, the
HbcF homologue Her8 has a more strict substrate specificity
and shows no methylation activity on 9 (Figure S83).
With the role of HbcF/Her8 having been established, efforts
were directed to verify HbcH/Her9 as the enzyme responsible
for the acylation of 9. The acyl donor tiglyl-CoA was
chemically prepared from tiglic acid and HSCoA (Figure
S71). Enzymatic assays indicated that HbcH could catalyze the
loading of the tiglyl group onto 9 to form 6 (Figure 3, v and
vi). The same was also observed for HbcH homologue Her9
(Figure S84). However, when 3 was used as the acceptor, only
a trace amount of tiglylated 8 was observed (Figure S70, v and
vi). Hence, the reaction of HbcH/Her9 most likely occurs
prior to any methylation of 9. In addition, the previously
reported 3 may only be a shunt product, but not an
intermediate, in HBC biosynthesis.
Next, the HbcG/Her10-catalyzed methylation taking place
after the HbcH/Her9 and HbcF/Her8 reactions was studied in
vitro. Though the HbcG protein tends to precipitate in the
reaction system, its ability to catalyze the methylation at C2′-
OH of 8 to form 5 could still be demonstrated (Figure 3, vii
and viii). HbcG has stringent substrate selectivity because it
could not methylate other substrates with free C2′-OH (9, 6,
and 3) (Figure S72). These results also showed that C8′-O-
acylation and C11′-O-methylation are prerequisites for HbcG
reaction. The same conclusion was also reached for Her10
based on the results of in vivo her10 deletion (Figure S81) and
the in vitro incubation of Her10 with compounds 9, 6, and 8
(Figures S83, S86, and S87). Lastly, the activity of Her11, a
homologue of HbcI, was tested by feeding 5 to a herbicidin
nonproducing strain Streptomyces albus J1074 overexpressing
Her11. The result revealed 5 was successfully converted to 1
by Her11 in vivo (Figure S88).
Taken together, a complete tailoring pathway of 1 can be
proposed as depicted in Figure 4A. Unlike HbcG, HbcF can
catalyze not only the methylation of 6 and 9 but also the
conversion of other compounds containing C11′-COOH (7
and 10) to their methyl ester forms (1 and 2, respectively)
(Figure S73, i−iv). When 3 and 8 containing C11′-COOMe
were used as substrates, no methylation of other hydroxyl
groups by HbcF was noted (Figure S73). Instead, 3 and 8 were
found to undergo spontaneous hydrolysis of the C11′-methyl
ester producing a small amount of 9 and 6, respectively, which
could be remethylated in the presence of HbcF and SAM
(Figure S73, v−viii). The substrate-tolerant HbcF may play a
role in ensuring methyl esterification of the carboxylic acid of
HBCs.
The substrate spectrum of acyltransferase HbcH/Her9 was
also investigated with compounds containing a bare C8′-OH
as substrates and tiglyl-CoA as the acyl donor. It was found
that 2 was inactive but 10 was tiglylated to yield a new
compound 19 (Figures S70, vii−xi, and S85). In combination
with the results of the acylation of 9 and 3 by HbcH (Figure 3,
iii and iv, and Figure S84), the preferred acyl acceptors for
HbcH/Her9 reaction are likely the C11′-carboxyl HBCs. In
addition, we tested the catalytic activity of HbcH using 9 as the
acceptor toward various acyl-CoAs donors, including acetyl-,
propionyl-, butyryl-, isobutyryl-, isovaleryl-, benzoyl-, and
malonyl-CoA and the chemically synthesized angelyl-CoA, a
cis isomer of tiglyl-CoA. It was found that HbcH could catalyze
Figure 3. HPLC analyses of the enzymatic reactions involved in the
HBC tailoring pathway. HbcF-catalyzed reaction with (i) 6 or (iii) 9
as the substrate. (v) HbcH-mediated reaction with 9 as the substrate.
(vii) HbcG-catalyzed reaction with 8 as the substrate. (ii, iv, vi, and
viii) Control reactions. (ix) Authentic standards. The results obtained
with the Her enzymes are shown in Figures S83, S84, S86, and S87.
C
DOI: 10.1021/acs.orglett.9b00066
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Figure 4. Summary of the biosynthesis of HBCs. (A) Proposed herbicidin A biosynthetic pathway. (B) Proposed biosynthetic network for
decorated HBCs derived from 9.
the transfer of these acyl groups, except malonyl-CoA, to form
a series of derivatives of 9 (Figures S70, i and ii, S74, and S75),
a feature useful for generating new HBC analogues.
more HBC analogues could be generated using the newly
characterized tailoring enzymes.
The fact that HbcH is homologous to serine hydrolase also
prompted us to investigate whether HbcH has hydrolytic
activity for HBCs 1 and 5−8, which all contain a tiglyl or a 5″-
hydroxytiglyl ester at C8′. In these experiments, spontaneous
hydrolysis of the methyl ester at C11′ and the acyl ester group
at C8′ was observed for 1, 5, and 8 after overnight reaction,
and the presence of HbcH further promoted the hydrolysis of
the C8′ tiglyl group of 8 but not others (Figures S76−S78). In
addition, HbcH could facilitate hydrolysis of an otherwise
stable acyl group at C8′ of 6 and 7. These findings
demonstrated that the C8′ acyl ester in C11′-OMe HBCs
can undergo hydrolysis spontaneously in vitro, while the same
activity requires HbcH participation for C11′-OH HBCs.
When HSCoA was present in the reaction mixture, the
hydrolytic activity of HbcH was significantly inhibited (Figures
S78−S80). The fact that HbcH has both acyl loading and
hydrolysis activity suggests that some HBC homologues such
as HBC B (2) may be produced by spontaneous or enzymatic
hydrolysis of HBC A (1) as previously suggested.14,19 The data
presented above allow the full constitution of the biosynthetic
network for the generation of all HBC congeners, in which
HbcH, HbcF, and spontaneous hydrolysis contribute to the
diversity of HBCs (Figure 4B).
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.9b00066.
Materials and methods and supplementary tables and
figures (PDF)
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: h.w.liu@mail.utexas.edu.
*E-mail: gltang@sioc.ac.cn.
ORCID
Sheng-Xiong Huang: 0000-0002-3616-8556
Hung-wen Liu: 0000-0001-8953-4794
Gong-Li Tang: 0000-0003-3149-4683
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
In summary, we have characterized the complete tailoring
pathway of herbicidin A (1), including the serine hydrolase-
catalyzed tiglyl loading, two steps of SAM-dependent
methylation, and the P450-catalyzed hydroxylation reaction.
To the best of our knowledge, HbcH and Her9 are the first
enzymes found in bacteria to catalyze the transfer of a tiglyl
group from tiglyl-CoA, a primary catabolite of isoleucine. In
addition, this study placed most of the reported ANM-derived
HBCs in the HBC modification network and suggested that
■
The authors thank Prof. Li-Ming Tao of East China University
of Science and Technology for providing the ANM-producing
strain. This work was supported by grants from the NNSFC
(81473124) and the CAS (Youth Innovation Promotion
Association 2016235, XDB20000000, QYZDJ-SSW-SLH037,
and K. C. Wong Education Foundation) to H.-X.P. or G.-L.T.
and from the National Institutes of Health (GM040541) and
Welch Foundation (F-1511) to H.-w.L.
D
DOI: 10.1021/acs.orglett.9b00066
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
REFERENCES
■
(1) Arai, M.; Haneishi, T.; Kitahara, N.; Enokita, R.; Kawakubo, K.;
Kondo, Y. J. Antibiot. 1976, 29, 863.
(2) Haneishi, T.; Terahara, A.; Kayamori, H.; Yabe, J.; Arai, M. J.
Antibiot. 1976, 29, 870.
(3) Takiguchi, Y.; Yoshikawa, H.; Terahara, A.; Torikata, A.; Terao,
M. J. Antibiot. 1979, 32, 862.
(4) Takiguchi, Y.; Yoshikawa, H.; Terahara, A.; Torikata, A.; Terao,
M. J. Antibiot. 1979, 32, 857.
(5) Terahara, A.; Haneishi, T.; Arai, M.; Hata, T.; Kuwano, H.;
Tamura, C. J. Antibiot. 1982, 35, 1711.
(6) Choi, C. W.; Choi, J. S.; Ko, Y. K.; Kim, C. J.; Kim, Y. H.; Oh, J.
S.; Ryu, S. Y.; Yon, G. H. B. Bull. Korean Chem. Soc. 2014, 35, 1215.
(7) Chai, X.; Youn, U. J.; Sun, D.; Dai, J.; Williams, P.; Kondratyuk,
T. P.; Borris, R. P.; Davies, J.; Villanueva, I. G.; Pezzuto, J. M.; Chang,
L. C. J. Nat. Prod. 2014, 77, 227.
(8) Zhang, J. C.; Yang, Y. B.; Chen, G. Y.; Li, X. Z.; Hu, M.; Wang,
B. Y.; Ruan, B. H.; Zhou, H.; Zhao, L. X.; Ding, Z. T. J. Antibiot. 2017,
70, 313.
(9) Chen, J. J.; Rateb, M. E.; Love, M. S.; Xu, Z.; Yang, D.; Zhu, X.;
Huang, Y.; Zhao, L. X.; Jiang, Y.; Duan, Y.; McNamara, C. W.; Shen,
B. J. Nat. Prod. 2018, 81, 791.
(10) Dai, X.; Li, G.; Wu, Z.; Lu, D.; Wang, H.; Li, Z.; Zhou, L.;
Chen, X.; Chen, W. Chem. Abstr. 1989, 111, 230661f.
(11) Hager, D.; Paulitz, C.; Tiebes, J.; Mayer, P.; Trauner, D. J. Org.
Chem. 2013, 78, 10784.
(12) Hager, D.; Mayer, P.; Paulitz, C.; Tiebes, J.; Trauner, D. Angew.
Chem., Int. Ed. 2012, 51, 6525.
(13) Ichikawa, S.; Shuto, S.; Matsuda, A. J. Am. Chem. Soc. 1999,
121, 10270.
(14) Lin, G. M.; Romo, A. J.; Liem, P. H.; Chen, Z.; Liu, H. W. J.
Am. Chem. Soc. 2017, 139, 16450.
(15) Wyszynski, F. J.; Lee, S. S.; Yabe, T.; Wang, H.; Gomez-
Escribano, J. P.; Bibb, M. J.; Lee, S. J.; Davies, G. J.; Davis, B. G. Nat.
Chem. 2012, 4, 539.
(16) Bretschneider, T.; Zocher, G.; Unger, M.; Scherlach, K.; Stehle,
T.; Hertweck, C. Nat. Chem. Biol. 2012, 8, 154.
(17) Li, J. N.; Xie, Z. J.; Wang, M.; Ai, G. M.; Chen, Y. H. PLoS One
2015, 10, e0120542.
(18) Pickens, L. B.; Kim, W.; Wang, P.; Zhou, H.; Watanabe, K.;
Gomi, S.; Tang, Y. J. Am. Chem. Soc. 2009, 131, 17677.
(19) Yoshikawa, H.; Takiguchi, Y.; Terao, M. J. Antibiot. 1983, 36,
30.
E
DOI: 10.1021/acs.orglett.9b00066
Org. Lett. XXXX, XXX, XXX−XXX