Cross-Module Enoylreduction in the Azalomycin F Polyketide Synthase

Article abstract of DOI:10.1002/anie.202011357

The colinearity of canonical modular polyketide synthases, which creates a direct link between multienzyme structure and the chemical structure of the biosynthetic end-product, has become a cornerstone of knowledge-based genome mining. Herein, we report genetic and enzymatic evidence for the remarkable role of an enoylreductase in the polyketide synthase for azalomycin F biosynthesis. This internal enoylreductase domain, previously identified as acting only in the second of two chain extension cycles on an initial iterative module, is shown to also catalyze enoylreduction in trans within the next module. The mechanism for this rare deviation from colinearity appears to involve direct cross-modular interaction of the reductase with the longer acyl chain, rather than back transfer of the substrate into the iterative module, suggesting an additional and surprising plasticity in natural PKS assembly-line catalysis.

Full text of DOI:10.1002/anie.202011357

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: Cross-module Enoylreduction in the Azalomycin F Polyketide
Synthase
Authors: Guifa Zhai, Wenyan Wang, Wei Xu, Guo Sun, Chaoqun Hu,
Xiangming Wu, Zisong Cong, Liang Deng, Yanrong Shi, Peter
F. Leadlay, Heng Song, Kui Hong, Zixin Deng, and Yuhui Sun
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202011357
Link to VoR: https://doi.org/10.1002/anie.202011357

Angewandte Chemie International Edition
10.1002/anie.202011357
RESEARCH ARTICLE
Cross-module Enoylreduction in the Azalomycin F Polyketide
Synthase
[a]
[b]
[a,d]
[a]
[a]
[a]
[b]
Guifa Zhai, Wenyan Wang, Wei Xu,
Guo Sun, Chaoqun Hu, Xiangming Wu, Zisong Cong,
[a]
[a]
[c]
[b]
[a]
[a]
[a]
Liang Deng, Yanrong Shi, Peter F. Leadlay, Heng Song, Kui Hong, Zixin Deng, Yuhui Sun*
[
a]
G. Zhai,[+] Dr. W. Xu, G. Sun, C. Hu, X. Wu, L. Deng, Y. Shi, Prof. Dr. K. Hong, Prof. Dr. Z. Deng, Prof. Dr. Y Sun
Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Wuhan University), Ministry of Education, and School of Pharmaceutical Sciences
Wuhan University
No. 185 East Lake Road, Wuhan 430071, People’s Republic of China
E-mail: yhsun@whu.edu.cn
[
+]
[
b]
c]
W. Wang, Z. Cong, Prof. Dr. H. Song
College of Chemistry and Molecular Sciences
Wuhan University
No. 299 Bayi Road, Wuhan 430072, People’s Republic of China
Prof. Dr. P. F. Leadlay
[
Department of Biochemistry
University of Cambridge
No. 80 Tennis Court Road, Cambridge CB2 1GA, United Kingdom
Current address: Singapore Institute of Food and Biotechnology Innovation, Agency for Science, Technology, and Research (A*STAR), Singapore
These authors contributed equally to this work.
[
[
d]
+]
Abstract: The colinearity of canonical modular polyketide synthases,
which creates a direct link between multienzyme structure and the
chemical structure of the biosynthetic end-product, has become a
cornerstone of knowledge-based genome mining. Here we report
genetic and enzymatic evidence for the remarkable role of an
reduction of the β-keto group, and subsequently transferred to the
downstream module for the next round of chain elongation. Finally,
the full-length intermediate is released by
a C-terminal
thioesterase (TE) domain, usually by hydrolysis or
macrocyclisation.[
1]
enoylreductase in the polyketide synthase for azalomycin
F
A second important group of modular PKSs has evolved that
also adopts an assembly-line mechanism, but here the extension
modules lack an integral AT domain, and all of the extension
modules are serviced by a free-standing AT enzyme acting in
trans.[4,5] A free-standing enzyme may even act selectively at
different stages of chain growth, as in the action of the ER LovC
during lovastatin biosynthesis on a fungal iterative unimodular
PKS.[ The trans-AT PKSs of bacteria also tend to show other
obvious deviations from colinearity, making the deciphering of
their assembly lines far from easy.
biosynthesis. This internal enoylreductase domain, previously
identified as acting only in the second of two chain extension cycles
on an initial iterative module, is shown here also to catalyse
enoylreduction in trans within the next module. The mechanism for
this rare deviation from colinearity appears to involve direct cross-
modular interaction of the reductase with the longer acyl chain, rather
than backtransfer of the substrate into the iterative module,
suggesting an additional and surprising plasticity in natural PKS
assembly-line catalysis.
6]
Strikingly, certain canonical (or cis-AT) PKSs also exhibit rare
deviations from colinearity. For example, an extension module
may 'stutter', that is, catalyse more than one successive identical
chain elongation.[7-12] In other cases, a module is 'skipped' to
Introduction
provide a shortened polyketide.[9c,13,14] In certain PKS systems,
the aberrant behaviour concerns a single domain. For example,
the function of a single "missing" DH domain may actually be
accomplished by a DH domain residing in a later-acting module.
There are also cases of a PKS AT domain working iteratively
on other modules.[9,18]
In previous work, we identified an unusual toggling ER domain
in the first iterative extension module during PKS-catalysed
biosynthesis of the 36-membered macrocyclic azalomycin F
Modular polyketide synthase (PKS) multienzymes are
versatile biosynthetic machines catalyzing the formation of a very
large number of antibiotics and related bioactive compounds via
an assembly line mechanism.[1] The assembly-line paradigm, first
established for the erythromycin-producing PKS,[2,3] established
that each step is catalysed by a distinct enzyme domain and that
these domains are arrayed in the multienzyme essentially in the
order they are used. This colinearity has been found to hold for a
large number of bacterial modular PKS. In each round of chain
[15-17]
(
AZL) (Figure 1A), in which the ER is inactive in the first but active
elongation, an extender unit is selected by
a
specific
[19]
in the second chain-elongation, respectively.
This stuttering
acyltransferase (AT) domain, transferred from acyl-CoA onto the
acyl carrier protein (ACP) domain, and then condensed with the
growing polyketide chain by the β-ketoacyl synthase (KS) domain.
The elongated intermediate remains tethered to ACP and is
successively and selectively processed by the (optionally present)
ketoreductase (KR) domain, dehydratase (DH) domain and
enoylreductase (ER) domain to adjust appropriately the extent of
module is referred to here as module 1/2. We wished to examine
a further non-colinearity in the operation of the AZL PKS: during
the third elongation, catalysed by extension module 3, the
polyketide chain should be fully reduced based on the known AZL
structure. However, module 3 contains no ER domain (Figure 1A).
This function is unlikely to be supplied by an external ER enzyme
since the AZL biosynthetic gene cluster alone produces AZL when
1
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202011357
RESEARCH ARTICLE
AzlA
AzlB
AzlE
A
B
LM
(
module 1/2
module 1/2
module 3
module 4
(the 4th elongation)
module 15
the 1st elongation) (the 2nd elongation)
(the 3rd elongation)
(the 15th elongation)
ER1/2
DH1/2
KR1/2
ER1/2
DH1/2
ER15
KR1/2
DH3 KR3
DH15
KR15
KR4
ACPL KS1/2 AT1/2 ACP1/2
KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
KS4 AT4 ACP4
KS15 AT15 ACP15
12
13
LM
module 1/2
ER1/2
module 3
3
6
37
DH1/2
KR1/2
DH3 KR3
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
AzlA
AzlB
AzlA
AzlB
D
AzlA
AzlB
AzlE
C
module 1/2
module 1/2
module 3
module 4
module 1/2
module 1/2
module 3
module 4
module 15
LM
LM
(
the 1st elongation) (the 2nd elongation)
(the 3rd elongation)
(the 4th elongation)
(the 1st elongation) (the 2nd elongation)
(the 3rd elongation)
(the 4th elongation)
(the 15th elongation)
ER1/2
DH1/2
KR1/2
ER1/2
DH1/2
KR1/2
ER1/2
DH1/2
KR1/2
ER1/2
DH1/2
KR1/2
ER？
KR3
ER15
DH3 KR3
DH3
DH15
KR15
KR4
KR4
ACPL KS1/2 AT1/2 ACP1/2
KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
KS4 AT4 ACP4
ACPL KS1/2 AT1/2 ACP1/2
KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
KS4 AT4 ACP4
KS15 AT15 ACP15
3
7
6
36
37
3
Figure 1. Alternative models for non-colinear enoylreduction in the AZL PKS. A) The non-colinearity between the AZL F3a structure
and domain organization in the initial extension modules of the AZL PKS. B) In this model, the intermediate α,β-unsaturated tetraketide
product of module 3 is further elongated until it reaches ACP15 where ER15 catalyses enoylreduction of both C12-C13 and C36-C37.
C) In this model, the unreduced intermediate of module 3 is transferred back to ACP1/2 for reduction by ER1/2 then transferred forward
(
without condensation) onto KS
4
for further elongation. D) ER1/2 and/or ER15 catalyses enoylreduction in trans of the α,β-unsaturated
during the third round of elongation.
intermediate of module 3 when it is tethered to ACP
3
heterologously expressed.[11] There are only two ER domains in
the AZL PKS, residing in module 1/2 and module 15, respectively.
We therefore considered alternative models for enoylreduction in
module 3 involving these enzymes. In one mechanism, similar to
the dehydration in module 4 of the iso-migrastatin PKS by the DH
domain of module 10,[16] the intermediate product of module 3 is
elongated until it reaches module 15, where ER15 catalyses two
successive enoylreductions to give the expected product (Figure
NADPH binding site was mutated from GGVG to AAVA in either
the ER1/2 or the ER15 domain of the AZL gene cluster in
Streptomyces sp. 211726, (Figures S1-S2). Analysis of the
fermentation products of these mutants (referred to here as
ΔER1/2 and ΔER15, respectively) by liquid chromatography
coupled with electrospray ionization high-resolution mass
spectrometry (LC-ESI-HRMS) showed that inactivation of either
ER1/2 or ER15 completely abolished the production of AZL. For
ΔER15 this led to the appearance of a new peak whose molecular
weight was 2 Da less than that of AZL F3a from the wild-type
strain (Figure 2). In contrast, for the ΔER1/2 mutant the new peak
had a molecular weight 4 Da less than that of AZL F3a (Figure 2).
Large-scale fermentation was conducted to obtain enough of
these F3a-related compounds for NMR analysis. The 13C NMR
1
B). Alternatively, the α,β-unsaturated intermediate product of
elongation on module 3 is transferred back to the ACP of module
/2 for enoylreduction, catalysed by ER1/2 and is then transferred
forward without condensation (i.e. module 3 is skipped) onto KS
1
4
of module 4 for subsequent elongation (Figure 1C). A further
possibility is that either ER1/2 or ER15 catalyses the enoylreduction
required in module 3, acting directly in trans independent of the
assembly-line (Figure 1D).
C
chemical shift changes of C12 (δ from 33.7 to 133.2) and C13
C
(δ from 30.7 to 134.6) revealed the presence of a double bond
between C12 and C13 in the structure of 1 from ΔER15 compared
with F3a of wild-type (Figures S3-S9 and Table S1). This result
revealed that inactivation of ER15 only affected the enoylreduction
on the intermediate of module 15, which unequivocally ruled out
the possibility that ER15 is involved in the enoylreduction on
module 3 (Figure 1B). In contrast, the structure of 2 purified from
Results and Discussion
To discriminate between the alternative models (Figure 1) for
enoylreduction of the tetraketide product of AZL module 3, the
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202011357
RESEARCH ARTICLE
ΔER1/2 contained two extra double bonds compared to wild-type,
in conjugation with the C40-C41 double bond, evidenced by the
A
+
1
3
changes in C NMR chemical shift of C36 (δ
C37 (δ from 34.6 to 130.2), C38 (δ from 28.0 to 127.7), and C39
from 33.7 to 135.1); and the corresponding changes in 13C
C
from 35.1 to 135.1),
(3)
Module 3 (AzlB)
methylmalonyl-CoA
C
C
(δ
C
α,β-reduction
NMR chemical shift of the adjacent carbon atoms, C35 (δ
C
from
from
8
1
C C
0.7 to 84.3), C40 (δ from 130.2 to 133.1), and C41 (δ
32.6 to 134.7) (Figures S3, S10-S15 and Table S2). As well as
(
4)
(5)
being fully consistent with the previously-reported role of the
toggling ER1/2 domain in the iterative module 1/2,[19] these results
provide direct genetic evidence that ER1/2 is the necessary and
sufficient catalyst for double bond reduction at C36-C37.
LM-Module 1/2 (AzlA)
Module 3 (AzlB)
DH3 KR3
B
C
D
E
F
KS3 AT3 ACP3
AzlA
AzlB
AzlE
ER1/2
module 1/2
module 1/2
module 3
module 15
the 15 elongation)
DH1/2
KR1/2
DH3 KR3
LM
(the 1st elongation) (the 2nd elongation)
(the 3rd elongation)
th
(
+
+
+
+
+
+
+
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
ER1/2
KR1/2
ER1/2
DH1/2
KR1/2
ER
ER15
DH1/2
DH
3
KR
3
DH15
KR15
ACPL KS1/2 AT1/2 ACP1/2
KS1/2 AT1/2 ACP1/2
KS3
AT
3
ACP3
KS15 AT15 ACP15
ER1/2
DH1/2
KR1/2
DH3 KR3
AzlA(ΔER1/2
)
Obs.: 1068.6566
Cal.: 1068.6577
F3a
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
AzlE(ΔER15
)
wild-type
0
5
10
15
ER1/2
Time (min)
2
DH1/2
KR1/2
DH3 KR3
Obs.: 1064.6268
Cal.: 1064.6264
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
ΔER1/2
0
5
10
15
Time (min)
Obs.: 1066.6404
Cal.: 1066.6421
1
ER1/2
DH1/2
KR1/2
DH3 KR3
ΔER15
0
5
10
15
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
Time (min)
Figure 2. LC-ESI-HRMS analysis of ΔER1/2 and ΔER15 mutants.
Inactivated domains are labelled in black with a cross, and the
corresponding changes in the main products are highlighted in
yellow. The structures of azalomycin F3a and its derivatives from
ΔER1/2 and ΔER15 were confirmed by LC-ESI-HRMS and NMR
ER1/2
DH1/2
KR1/2
DH3 KR3
G
H
I
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
ER1/2
DH1/2
KR1/2
DH3 KR3
(
Figures S4-S15 and Tables S1-S2).
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
ER1/2
DH1/2
KR1/2
DH3 KR3
To examine the role of ER1/2 in vitro, the PKS multienzyme AzlA,
ACPL KS1/2 AT1/2 ACP1/2
KS3 AT3 ACP3
comprising the loading module (LM) and module 1/2, and module
of multienzyme AzlB with an intact N-terminus were each
expressed as recombinant proteins in Escherichia coli BAP1
15
16
17
18
19
20
Time (min)
3
Figure 3. In vitro reconstitution of cross-module enoylreduction.
A) The synthetic N-acetylcysteamine thioester 3 serves as a
substrate mimic for module 3 and is condensed with elongation
unit methylmalonyl-CoA. B) the tetraketide extension product was
then released from module 3 by KOH hydrolysis and monitored
by LC-ESI-HRMS using m/zꢀ240.1706 for 4 and m/zꢀ242.1863 for
(
Figure S16).[20] In our previous work, module 1/2 chain elongation
[19]
was reconstituted in vitro,
by using Azl4 (ligase) and Azl5
(
acyltransferase) to activate and deliver the starter unit to AzlA.
Here, we synthesised the product of module 1/2 as the N-
acetylcysteamine thioester (3) (Scheme S1and Figures S17-S22)
and used it as a triketide substrate mimic for module 3 (Figure 3A). 5. The same reaction was carried out in the presence of C) wild-
An assay containing 3 and AZL PKS module 3 was performed in
the presence of methylmalonyl-CoA and NADPH and the
type AzlA, D) AzlA(ΔER1/2) and E) AzlA(ΔACP1/2); and also with
AzlA domain or domain sets F) ER1/2, G) ER1/2-KR1/2, H) DH1/2
-
thioester-bound product was subsequently released by alkaline
ER1/2-KR1/2 and I) DH1/2-ER1/2-KR1/2-ACP1/2. As a negative control,
methylmalonyl-CoA was omitted from the reaction mixture.
Inactivated domains are shown in black with a cross.
hydrolysis (Figure 3B).[
19,21]
LC-ESI-HRMS and MS/MS
fragmentation[ analysis revealed a peak corresponding to the
tetraketide 10-guanidino-2-methyl-2,6-dien-decanoic acid (4),
which represents one round of chain elongation without
enoylreduction (Figure 3B and Figures. S23A-S23B). When the
same reaction was carried out in the presence of AzlA, a major
compound (10-guanidino-2-methyl-6-en-decanoic acid, 5) was
detected whose mass corresponds to one round of chain
elongation with ensuing enoylreduction (Figure 3C and Figures
S23C-S23D). Even though AzlA comprises seven domains
22]
performed in the presence of methylmalonyl-CoA, NADPH and
the mutated AzlA, the reaction product was unsaturated
tetraketide, with less than 1% α,β-saturated product detected
(Figure 3D). Even this trace amount of reduced product was
eliminated when a different AzlA mutant was used whose NADPH
binding site was mutated by changing the sequence GGVG to
SPVG (Figure S24).[23] These in vitro results further support the
conclusion that ER1/2 within AzlA is responsible for enoylreduction
of the tetraketide intermediate in AZL biosynthesis.
(
ACPLoading-KS1/2-AT1/2-DH1/2-ER1/2-KR1/2-ACP1/2), the obvious
catalyst for the enoylreduction is the ER domain. To confirm its
role in this conversion, we inactivated the ER1/2 of AzlA by
mutating the NADPH binding site sequence GGVG to AAVA
These in-vitro results do not allow a distinction between a
mechanism (Figure 1C) in which the α,β-unsaturated intermediate
of module 3 is transferred back to the ACP of module 1/2 for
(
Figure S1). When the assay containing 3 and module 3 was
3
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202011357
RESEARCH ARTICLE
enoylreduction catalysed by ER1/2, and one (Figure 1D) in which
the ER acts in trans upon the unsaturated tetraketide while it
remains tethered to the ACP of module 3. We therefore
inactivated ACP1/2 in the AzlA protein, by using site-directed
mutagenesis to replace the active site serine residue by alanine
In summary, this study revealed an unprecedented modular PKS
enoylreductase domain that is used in not only “switchable” but
also cross-module enoylreduction during polyketide chain
elongation, and provided evidence in favour of
a direct
intermodular interaction with the substrate anchored in the
downstream module, instead of a "stuttering" mechanism in which
the acyl chain is backtransferred for reduction. Our findings
provide the starting point for detailed structural and mechanistic
investigation of the factors favouring this mechanism, and add to
our growing appreciation of the plasticity of assembly-line
polyketide biosynthesis, and of the insights examples of non-
colinearity may provide into plausible pathways by which these
systems have evolved.[
(
Figure S25).[24] AzlA(ΔACP1/2) was purified from E. coli BAP1 and
supplied to the module 3 chain elongation reaction (Figure S16).
The result (Figure 3E) showed that the elongated product of
module 3 is reduced to the same extent as in the unmutated
control (Figure 3C), indicating that the intermediate does not
require back-transfer to ACP1/2 and likely remains tethered to
27]
ACP
results together, ER1/2 is recruited in trans by module 3 to catalyse
α,β-double bond reduction of the intermediate tethered onto ACP
3
for enoylreduction. Taking the above genetic and enzymatic
3
.
Acknowledgements
The factors that promote this - to our knowledge - unprecedented
cross-modular catalysis remain to be determined, but may include
a gatekeeping function of the KS domain of the downstream
module 4, which may be highly selective for recruitment of the
reduced tetraketide over the α,β-unsaturated tetraketide; and the
presence of ER1/2 on a separate protein may allow more flexibility
than if modules 1-3 were housed in the same multienzyme.
We were interested to determine whether the action of ER1/2
This work was supported by National Key R&D Program of China
(2018YFA0903203) and the Open Funding Project from State
Key Laboratory of Microbial Metabolism (MMLKF18-11). The
authors are grateful to J. Zheng at Shanghai Jiao Tong University
for helpful discussion.
on the neighbouring module would be seen with the isolated ER1/2
domain; or whether other domains might be involved. Based on
the known boundaries of PKS ER domains,[25,26] ER1/2 was
expressed as a recombinant protein, purified (Figure S16) and
used in the module 3 chain-elongation assay. No reduced product
was detected (Figure 3F). This result might reflect either incorrect
folding of the recombinant protein, or a need for the presence of
additional domains to provide protein:protein interactions
essential for the intermodular activity. As confirmed by the crystal
structure of the KR-ER didomain from the second module of the
spinosyn type I PKS, the ER domain is sandwiched between a
structural and a catalytic subdomain of the KR domain.[27] If the
same intimate contacts are maintained in the AZL ER1/2-KR1/2
didomain, it may promote proper folding of the ER1/2, allowing the
didomain to be active when the ER1/2 domain alone is not. The
effects of inclusion of other flanking domains are less predictable.
When recombinant subsets of module 1/2 containing additional
Keywords: natural products • biosynthesis • polyketide
synthases • enoylreductase domain • iterative domain
[
1]
T. Robbins, Y. C. Liu, D. E. Cane, C. Khosla, Curr. Opin. Struct. Biol.
2016, 41, 10–18.
[2]
J. Cortes, S. F. Haydock, G. A. Roberts, D. J. Bevitt, P. F. Leadlay,
Nature 1990, 348, 176–178.
[3]
[4]
[5]
S. Donadio, M. J. Staver, J. B. McAlpine, S. J. Swanson, L. Katz, Science
1
991, 252, 675–679.
Y. Q. Cheng, G. L. Tang, B. Shen, Proc. Natl. Acad. Sci. USA 2003, 100,
149–3154.
3
T. A. Nguyen, K. Ishida, H. Jenke-Kodama, E. Dittmann, C. Gurgui, T.
Hochmuth, S. Taudien, M. Platzer, C. Hertweck, J. Piel, Nat. Biotechnol.
2008, 26, 225–233.
[6]
B. D. Ames, C. Nguyen, J. Bruegger, P. Smith, W. Xu, S. Ma, E. Wong,
S. Wong, X. Xie, J. W.-H. Li, et al., Proc. Natl. Acad. Sci. USA 2012, 109,
11144–11149.
[
7]
8]
B. Wilkinson, G. Foster, B. A. M. Rudd, N. L. Taylor, A. P. Blackaby, P.
J. Sidebottom, D. J. Cooper, M. J. Dawson, A. D. Buss, S. Gaisser, et al.,
Chem. Biol. 2000, 7, 111–117.
domains (ER1/2-KR1/2, DH1/2-ER1/2-KR1/2 and DH1/2-ER1/2-KR1/2
-
ACP1/2) were assayed (Figure S16), the LC-ESI-HRMS results
showed that they all supported inter-modular enoylreduction
[
N. Gaitatzis, B. Silakowski, B. Kunze, G. Nordsiek, H. Blöcker, G. Höfle,
R. Müller, J. Biol. Chem. 2002, 277, 13082–13090.
(
Figures 3G-3I). Notably, the presence of the KR1/2 domain seems
[9]
a) J. He, C. Hertweck, ChemBioChem 2005, 6, 908–912; b) Y. Sugimoto,
K. Ishida, C. Hertweck, Angew. Chem. Int. Ed. 2014, 53, 1560–1564; c)
H. Peng, K. Ishida, C. Hertweck, Angew. Chem. Int. Ed. 2019, 58,
necessary for the in-trans ER1/2 activity on the substrate analogue
tethered to module 3. But the inactivation of KR1/2 domain has no
detectable effect on ER1/2 domain catalysing the cross-module
enoylreduction, which suggest that KR1/2 domain is important in
preserving contacts with ER1/2 and stabilising the ER1/2 in its
native conformation (Figure S26). Also, the ER1/2 can interact with
its substrate tethered to module 3 even in the absence of specific
ACP-KS contacts (Figure 3H); although the extent of reduction is
greater when the ACP1/2 domain is present (Figure 3I). The
phylogenetic analysis and the 3D modelling revealed no obvious
feature in which ER1/2 differs structurally from typical PKS ER
domains (Figures S27-S29). Further structural and functional
experimentation will be required to identify the exact
protein:protein interactions that contribute to this enhancement.
18252–18256.
[
[
[
10] C. Olano, B. Wilkinson, S. J. Moss, A. F. Braña, C. Méndez, P. F. Leadlay,
J. A. Salas, Chem. Commun. 2003, 22, 2780–2782.
11] H. Hong, Y. Sun, Y. Zhou, E. Stephens, M. Samborskyy, P. F. Leadlay,
Beilstein J. Org. Chem. 2016, 12, 2164–2172.
12] O. Bilyk, M. Samborskyy, P. F. Leadlay, PLoS One 2019, 14, e0215958.
[13] Y. Xue, D. H. Sherman, Nature 2000, 403, 571–575.
[14] C. J. Rowe, I. U. Böhm, I. P. Thomas, B. Wilkinson, B. A. M. Rudd, G.
Foster, A. P. Blackaby, P. J. Sidebottom, Y. Roddis, A. D. Buss, et al.,
Chem. Biol. 2001, 8, 475–485.
[
[
[
[
15] L. Tang, S. Ward, L. Chung, J. R. Carney, Y. Li, R. Reid, L. Katz, J. Am.
Chem. Soc. 2007, 126, 46–47.
16] B. Zhang, Z. Xu, Q. Teng, G. Pan, M. Ma, B. Shen, Angew. Chem. Int.
Ed. 2017, 56, 7247–7251 (2017).
17] W. D. Fiers, G. J. Dodge, D. H. Sherman, J. L. Smith, C. C. Aldrich, J.
Am. Chem. Soc. 2016, 138, 16024–16036.
18] M. Otsuka, K. Ichinose, I. Fujii, Y. Ebizuka, Antimicrob. Agents Chemother.
Conclusion
2004, 48, 3468–3476.
4
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202011357
RESEARCH ARTICLE
[
[
[
[
[
[
[
[
[
19] W. Xu, G. Zhai, Y. Liu, Y. Li, Y. Shi, K. Hong, H. Hong, P. F. Leadlay, Z.
Deng, Y. Sun, Angew. Chem. Int. Ed. 2017, 56, 5503–5506.
20] B. A. Pfeifer, S. J. Admiraal, H. Gramajo, D. E. Cane, C. Khosla, Science
2
001, 291, 1790–1792.
21] S. E. O’Connor, H. Chen, C. T. Walsh, Biochemistry 2002, 41, 5685–
694.
22] H. Hong, T. Fill, P. F. Leadlay, Angew. Chem. Int. Ed. 2013, 52, 13096–
3099.
5
1
23] S. Donadio, J. B. McAlpine, P. J. Sheldon, M. Jackson, L. Katz, Proc.
Natl. Acad. Sci. USA. 1993, 90, 7119–7123.
24] C. Liu, J. Zhu, Y. Li, J. Zhang, C. Lu, H. Wang, Y. Shen, ChemBioChem
2015, 16, 998–1007.
25] L. Zhang, J. Ji, M. Yuan, Y. Feng, L. Wang, Z. Deng, L. Bai, J. Zheng,
ACS Chem. Biol. 2018, 13, 871–875.
26] J. Zheng, D. C. Gay, B. Demeler, M. A. White, A. T. Keatinge-Clay, Nat.
Chem. Biol. 2012, 8, 615–621.
27] S. J. Moss, C. J. Martin, B. Wilkinson, Nat. Prod. Rep. 2004, 21, 575–
593.
5
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202011357
RESEARCH ARTICLE
Polyketide synthase assembly line of Azalomycin F
LM module 1/2
module 3
(
the 1st elongation) (the 2nd elongation)
(the 3rd elongation)
ER
ER
ER
DH
KR
DH
KR
DH
KR
ACPL
KS AT ACP
KS AT ACP
KS AT ACP
Borrowing: A switchable enoylreductase (ER) domain borrowed by neighbouring module in macrolide antibiotic Azlomycin F
biosynthesis shows modular polyketide synthase (PKS)'s versatility that has never been known. This internal ER domain not only
serves as an unusual switch in an iterative module for enoylreduction but also function surprisingly in a cross-module mode during
polyketide chain elongation. This discovery further expands our understanding of the wonderful nature of PKS.
6
This article is protected by copyright. All rights reserved.