An Unusual Dehydratase Acting on Glycerate and a Ketoreducatse Stereoselectively Reducing α-Ketone in Polyketide Starter Unit Biosynthesis

Article abstract of DOI:10.1002/anie.201406602

Polyketide synthases (PKSs) usually employ a ketoreductase (KR) to catalyze the reduction of a β-keto group, followed by a dehydratase (DH) that drives the dehydration to form a double bond between the α- and β-carbon atoms. Herein, a DH?-KR? involved in FR901464 biosynthesis was characterized: DH? acts on glyceryl-S-acyl carrier protein (ACP) to yield ACP-linked pyruvate; subsequently KR? reduces α-ketone that yields L-lactyl-S-ACP as starter unit for polyketide biosynthesis. Genetic and biochemical evidence was found to support a similar pathway that is involved in the biosynthesis of lankacidins. These results not only identified new PKS domains acting on different substrates, but also provided additional options for engineering the PKS starter pathway or biocatalysis. Biochemical characterization of a dehydratase and a ketoreductase-like domain (DH?-KR?) in the loading module of FR901464 polyketide synthase revealed that DH? catalyzes the dehydration of an acyl carrier protein (ACP)-tethered glycerate to an ACP-linked pyruvate. The KR? domain then carries out α-ketone reduction to yield L-lactyl-S-ACP, which serves as a starter unit for polyketide biosynthesis. Genetic and biochemical evidence was found to support a similar pathway that is involved in the biosynthesis of lankacidins. These results not only identified new PKS domains acting on different substrates, but also provided additional options for engineering the PKS starter pathway or biocatalysis.

Full text of DOI:10.1002/anie.201406602

Angewandte
Chemie
DOI: 10.1002/anie.201406602
Polyketide Biosynthesis
An Unusual Dehydratase Acting on Glycerate and a Ketoreducatse
Stereoselectively Reducing a-Ketone in Polyketide Starter Unit
Biosynthesis**
Hai-Yan He, Hua Yuan, Man-Cheng Tang, and Gong-Li Tang*
Dedicated to Professors Chengye Yuan and Lixin Dai on the occasion of their 90th birthdays
Abstract: Polyketide synthases (PKSs) usually employ a ketor-
eductase (KR) to catalyze the reduction of a b-keto group,
followed by a dehydratase (DH) that drives the dehydration to
form a double bond between the a- and b-carbon atoms.
Herein, a DH*-KR* involved in FR901464 biosynthesis was
characterized: DH* acts on glyceryl-S-acyl carrier protein
(ACP) to yield ACP-linked pyruvate; subsequently KR*
reduces a-ketone that yields l-lactyl-S-ACP as starter unit for
polyketide biosynthesis. Genetic and biochemical evidence was
found to support a similar pathway that is involved in the
biosynthesis of lankacidins. These results not only identified
new PKS domains acting on different substrates, but also
provided additional options for engineering the PKS starter
pathway or biocatalysis.
potency,[4] have been well-studied. Most KRs utilize
NADPH to stereoselectively reduce the b-keto group of
a b-ketoacyl-S-ACP intermediate generated by the KSs; then,
DHs catalyze the dehydration of the resulting intermediate to
form a double bond between the a- and b-carbons.
FR901464 (Scheme 1A) is a natural antitumor agent and
representing a new class of potent anticancer small molecules
that target the spliceosome inhibiting both splicing and
nuclear retention of pre-mRNA.[5] Previous studies have
M
odular polyketide synthases (PKSs), also known as type I
PKSs, have been well-established to catalyze the biosynthesis
of a large group of polyketide natural agents with remarkably
structural diversity using a thiotemplated assembly line. Each
module contains a b-ketoacyl synthase (KS), an acyltransfer-
ase (AT), and an acyl carrier protein (ACP), which catalyze
one cycle of chain elongation that extend the growing
polyketide chain by a C2 unit.[1] Other chemically diverse
polyketide skeletons arise from processing domains, such as
ketoreductase (KR), dehydratase (DH), and enoyl reductase
(ER) domains, which may be present in a module in various
combinations to control the oxidation state and stereochem-
istry of the growing polyketide chain.[1] The DH and KR
domains, including their structure-based catalytic mecha-
nism,[2] regio- and stereoselectivity,[3] and biocatalytic
[*] Dr. H.-Y. He, Dr. H. Yuan, Dr. M.-C. Tang, Prof. Dr. G.-L. Tang
State Key Laboratory of Bioorganic and Natural Products Chemistry
Shanghai Institute of Organic Chemistry
Chinese Academy of Sciences
345 Lingling Road, Shanghai, 200032 (China)
E-mail: gltang@sioc.ac.cn
Prof. Dr. G.-L. Tang
Shanghai Collaborative Innovation Center for
Biomanufacturing Technology
130 Meilong Road, Shanghai 200237 (China)
Scheme 1. Loading modules of PKSs in FR901464 and lankacidin
biosynthesis. A) The molecular structure and the loading module (M-
L). DH, dehydratase; KR, ketoreductase; GAT (or FkbH), glyceryl
transferase/phosphatase; ACP, acyl carrier protein. The three-carbon
starter units are boxed. B) Proposed reactions catalyzed by the M-L,
the reactions by DH*-KR* are highlighted in a box. C) Proposed starter
unit biosynthesis of lankacidin (1) and lankacyclinol (2).
[**] We thank Prof. Zixin Deng’s Lab at Shanghai JiaoTong University for
support in obtaining MS data. This work was financially supported
by grants from the 973 Program (2010CB833200) and NNSFC
(31200054 and 31330003).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201406602.
Angew. Chem. Int. Ed. 2014, 53, 11315 –11319
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11315

.
Angewandte
Communications
revealed that it is biosynthesized by “trans-AT” PKSs
hybridized with nonribosomal peptide synthetase (NRPS)
and an isoprenoid-like b-branching pathway.[6] In the loading
module, a DH- and a KR-like didomain (DH*-KR*, that did
not have an exact match throughout the sequence) were
proposed to catalyze the biosynthesis of unusual PKS starter
unit (Scheme 1A).[6] Although a similar domain organization
and suggested biosynthetic pathway were also described in
the biosynthesis of bryostatin (Supporting Information, Fig-
ure S1), an anticancer polyketide from an uncultivated
bacterial symbiont, based on the bioinformatic analysis,[7]
the enzymatic logic has remained unknown. Herein, we
report the biochemical elucidation of the physiological role of
this type of novel domains: DH* first acting on glycerate;
KR* then stereoselectively reducing a-ketone, which differs
from known PKSs.
It has been established that a glyceryl transferase (GAT)
and an ACP in the loading module incorporate d-1,3-
bisphosphoglycerate (1,3-BPG) to afford ACP-tethered glyc-
erate 3.[6] The following enzymatic pathway has been pro-
posed: first, DH* catalyzes the dehydration of 3 to an ACP-
linked enoylpyruvate 4, which could spontaneously rearrange
to form ACP-bound pyruvate 5; then, KR* carries out a-
ketone reduction to yield l-lactyl-S-ACP 6, which serves as
the starter unit for PKS in FR901464 biosynthesis (Sche-
me 1B).[7] To validate this postulation, the DH* domain was
firstly expressed and purified from E. coli BL21 (DE3)
(Supporting Information, Figure S2). Considering that the
retention time of 3 and holo-ACP is almost the same in high-
performance liquid chromatography (HPLC) analysis,[6] we
chose another ACP-tethered glycerate 3b (Figure 1A, with
a different retention time to the respective holo-ACP), from
the quartromicin (QMN) biosynthetic pathway,[8] as a sub-
strate mimic to facilitate detection. After 3b was generated
in situ catalyzed by QmnD1 (a GAT coupled with QmnD2 in
QMN biosynthesis, Figure 1B-I), it was incubated with Fr9C-
DH* and the reaction products were then subjected to HPLC
analysis. The result shows that the peak height of 3b
decreased and that of holo-ACP increased, but no new peak
signals were detected (Figure 1B-II). In a control assay
containing the boiled enzyme, the substrate 3b did not show
any changes over several hours (Figure 1B-III). When the
reaction products were detected using high-resolution MS
(HRMS), a trace of a new protein possessing identical
molecule weight with ACP-bound pyruvate 5b was observed
(Figure 1C-III). The a-keto ester 5b is difficult to detect
owing to its instability towards hydrolysis to give the a-keto
acid 5c and holo-ACP (Figure 2A), as described by Calder-
one.[9] To solve this problem, we added derivative reagent 4-
dinitrophenylhydrazine (DNPH) to the reactions to detect a-
keto acid (Figure 2A).[10] As expected, pyruvate derivative 5d
was observed in the Fr9C-DH* assay only (Figure 2B). These
results indicate that Fr9C-DH* catalyzes b-dehydration of
ACP-tethered glycerate 3b to yield pyruvoyl-S-ACP 5b
(Figure 1A).
Figure 1. Biochemical characterization of Fr9C-DH*-KR*. A) Reactions
in these assays. B) HPLC analysis: I, generation of 3b in situ; II, 3b
with Fr9C-DH*; III, 3b with boiled Fr9C-DH*; IV, 3b with Fr9C-DH*-
KR* and NADPH; V, 3b with Fr9C-DH*-KR* and NADH for 30 min;
VI, 3b with Fr9C-DH*-KR* and NADH for 2 h. C) HRMS analysis: I,
holo-ACP; II, generation of 3b; III, 3b with Fr9C-DH*; IV, 3b with
Fr9C-DH*-KR* and NADPH.
Figure 2. Chemical derivatives and assays of KR*. A) The derivative
reaction of pyruvate. B) HPLC analysis: I, standard control; II, full
assay; III, assay with boiled Fr9C-DH*. C) The derivative reaction of
lactyl unit. D) HPLC analysis: I, standard control; II, assay with Fr9C-
DH*-KR* and NADPH; III, assay with Orf19 and NAD(P)H. E) Reac-
tion catalyzed by KR. F) HPLC analysis: I and II, standard; III, full
assay with Fr9C-DH*-KR* and NADPH.
Next, we investigated the function of Fr9C-KR*. DH*-
KR* didomain protein was expressed and purified for
biochemical assays, since expression of the independent
KR* domain was not successful owing to its insolubility in
11316 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 11315 –11319

Angewandte
Chemie
E. coli BL21 (DE3) (Supporting Information, Figure S2).
Similarly, after 3b was generated by in situ, DH*-KR* and
cofactor NADPH were added to the reaction. HPLC analysis
revealed that the amount of 3b decreased and a new peak
emerged (Figure 1B-Vand VI). HRMS analysis reported that
the molecular weight of this new peak is consistent with lactyl-
S-ACP 6b (Figure 1C-IV); however, when NADPH was
replaced by NADH, production of 6b was not observed, even
though the reaction lasted 2 h. Most of 3b had been
converted, which is similar to the result of only adding
Fr9C-DH* (Figure 1B-IV). Furthermore, if d-phenylalanine
methyl ester as derivative reagent was added to the Fr9C-
DH*-KR* assays, it will react with 6b to give 6c (Fig-
ure 2C).[11] We chemically synthesized diastereoisomer 6c
and 6d for potential product standards (Supporting Informa-
tion, Figure S3). As expected, generation of 6c, the lactyl
chirality of which is consistent with that of final product
FR901464, was detected in the derivative assay, but 6d with
opposite chirality cannot be observed (Figure 2D-II). We also
synthesized N-acetyl cysteamine (NAC) thioester 5e to mimic
the ACP-bound substrate 5b of KR* domain and stereoiso-
mer 6e and 6 f for potential product standard (Figure 2E).
After incubation of 5e with NADPH and the DH*-KR*,
HPLC and MS analyses showed that the anticipated reductive
product 6e could be detected, but 6 f with opposite chirality
was not observed (Figure 2F); though, owing to instability of
5e in its aqueous phase, conversion efficiency of 5e to 6e was
very low (about 5–10%). These results demonstrated that the
Fr9C-KR* domain could use NADPH, but not NADH, as
a hydrogen donor to reduce the pyruvoyl unit to l-lactyl unit
tethered on ACP stereospecifically (Figure 1A). Further-
more, consumption of the unstable intermediate pyruvoyl-S-
ACP 5b into 6b by KR* could accelerate the dehydration by
DH*, which hints at the functional correlation between two
domains.
cluster;[7] however, there is no any genetic or biochemical
evidence to support this hypothesis to date. Thus, in vivo
experiments were carried out and we found that production of
1 and 2 were terminated completely in knockout mutants of
orf19 and orf22. When orf19 and orf22 were complemented to
the mutant respectively, production of 1 and 2 was regained,
though the yield was lower than that of the wild type
(Supporting Information, Figure S4 and S5). These results
indicate that orf19 and orf22 are essential for the biosynthesis
of 1 and 2. Next, we used in vitro analyses to verify our
speculation. Three proteins Orf19, Orf21, and Orf22 were
expressed and purified in E. coli BL21 (DE3) (Supporting
Information, Figure S2). Similily, apo-Orf21 was converted
completely into active holo-Orf21 by Sfp (Figure 3A-I and
3B-I). Then 1,3-BPG and bifunctional glyceryl transferase/
Figure 3. Biochemical characterization of Orf19, Orf21, and Orf22.
A) HPLC analysis: I, generation of holo-Orf21; II, generation of 3a
catalyzed by Orf 22; III, assay of 3a with Orf19; IV, full assay of 3a
with Orf19 and NAD(P)H. B) HRMS analysis: I, holo-ACP; II, gener-
ation of 3a; III, assay of 3a with Orf19 and NAD(P)H.
DH*-KR*-GAT-ACP organization is rarely used as a PKS
starter module for the biosynthesis of polyketide natural
products. Other than bryostatin, additional such cases can be
found in thailanstatin,[12] the symmetric polyketide dimer
SIA7248[13] and the tartrolons[14] biosynthetic pathways (Sup-
porting Information, Figure S1). Additionally, a homologous
sequence was discovered in Genbank, which is located within
a linear plasmid in Streptomyces rochei 7434AN4.[15] This
homologous sequence contains three independent genes,
orf19, orf21 and orf22, which encode a DH*-KR* didomain
protein (Orf19), an ACP (Orf21) and a GAT-like protein
(Orf22), respectively (Scheme 1A). These genes are adjacent
to the gene cluster of lankacidin which is biosynthesized by
a hybrid PKS/NRPS system.[16] The structural difference
between lankacidin A (1) and its analogue lankacyclinol A
(2) is that the three-carbon unit is linked to the amino group
with pyruvoyl and l-lactyl group, respectively (Scheme 1A).
Although two groups have reported the biosynthetic studies
of 1 and 2,[16] the origin and pathway of the three-carbon unit
is still unknown. Considering the FR901464 starter pathway,
we believed that the genes orf19, orf21, and orf22 may be
related to the biosynthesis of the three-carbon units of 1 and 2
(Scheme 1C). This proposal had also been discussed by
Sherman and Haygood in reporting on the bryostatin gene
phosphatase Orf22 were added to the reaction. Glyceroyl-S-
ACP 3a was detected (Scheme 1C, Figure 3A-II and 3B-II),
illustrating that Orf22 is fully functional. The DH*-KR*
didomain protein Orf19 was next added into the reaction
using 3a as substrate. As expected, the peak of 3a disap-
peared and the amount of holo-Orf21 increased (Figure 3A-
III), which is analogous to the phenomena observed in the
dehydration reaction. When NADPH or NADH were added
as hydrogen donors, a new peak emerged with molecular
weight consistent with that of 6a (Figure 3A-IV and B-III);
this assay was also dealt with derivative reagent, product 6c
was detected (Figure 2D-III). By combining in vivo and
in vitro results, we established the essential roles of orf19,
orf21, and orf22 in the biosynthesis of lankacidins and
identified the missing link of the starter pathway of 1 and 2
(Scheme 1C).
In modular PKS systems, DH and KR domains are
essential components for the structural diversity of various
products.[1] Recent structural studies have revealed that Asp
residue donates a proton to the b-hydroxy group and His
residue abstracts an a-proton in the DH domain, which play
vital roles in dehydration, resulting in an a,b-unsaturated
intermediate.[2] DH domains have two hotdog folds and the
His residue is located on the small cap of the N-terminal; the
Angew. Chem. Int. Ed. 2014, 53, 11315 –11319
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11317

.
Angewandte
Communications
Asp is on a helix of the C-terminal hotdog.[17] In the present
work, it was confirmed that the two DH*-like proetins (Fr9C-
DH* and Orf19-DH*) function as DHs acting on the glyceryl-
S-ACP substrate; however, further bioinformatic analysis of
these DH*-like enzymes suggested that these proteins belong
to a family of MaoC-like dehydratases (pfam 01575, Support-
ing Information, Figure S6), which usually act as (R)-specific
enoyl-CoA hydratase and have a conserved catalytic DxxxxH
motif.[18] The Asp residue activates a water to attack the C3
carbon of the substrate, and the His residue donates a proton
to the C2 carbon. Interestingly, multiple sequence alignment
analysis revealed that the DH* sequences also have a con-
served DxxxxH motif (Supporting Information, Figure S6).
Considering the reversible reaction properties of these
enzymes,[19] we inferred that this motif is also essential for
the dehydration activity in Fr9C-DH* and Orf19-DH*. As
expected, biochemical assays showed that both the His and
Asp mutants lost their dehydration activity (Supporting
Information, Figure S8). Thus, the DH*-like domains may
be a new family of DHs that could be integrated into the PKS
modules and have the catalytic DxxxxH motif as in (R)-
specific enoyl-CoA hydratases but catalyze the b-dehydration
of glyceryl-S-ACP (Figure 4A).
expected, mutation of the key Tyr430 into Ala abolished the
activities of Fr9C-KR* (Supporting Information, Figure S8).
Further neighbour-joining tree analysis revealed that these
KR*s could also be classified into a new family of KRs,
although they have sequence homology to SDR (Figure 4B).
It is well-established that b-KRs can be classified into A- and
B-types according to their sequences, which catalyze the
opposite configuration of products.[23] Owing to the structure
of the lactyl unit, we inferred that a-KR* that play a role in
the biosynthesis of FR901464, thailanstatins, lankacinins,
SIA7248, and tartrolons are all responsible for l-OH
formation, while only a-KR* that plays a role bryostatin
biosynthesis produces d-OH (Supporting Information, Fig-
ure S1). However, multiple alignment analysis of these a-
KR*s did not reveal any clues to predict the product
configuration, which do not resemble the well studied PKS
b-KRs. Several mutants of Fr9C-KR* were constructed to
alter the configuration of reduction products, but all tries
were unsuccessful (data not shown).
Glyceroyl-S-ACP is an unusual extender unit used in PKS
which could be further transformed into hydroxymalonyl-S-
ACP or methoxymalonyl-S-ACP.[24] Herein we discovered
a new conversion by a DH*-KR* bifunctional protein into
lactyl-S-ACP, which serves as a starter unit for PKS in
FR901464 and for hybrid NRPS/PKS in lankacidin biosyn-
thesis. Further bioinformatic analysis allowed us to identify at
least seven other biosynthetic systems containing the DH*-
KR* like enzymes based on the genome sequence, although
the respective biosynthetic pathways or products have never
been explored (Supporting Information, Table S4). Thus, our
characterizations of a different starter pathway in polyketide
biosynthesis not only provide a new strategy for combinato-
rial biosynthesis, but also could impact in the discovery of
novel natural products by genome mining.
Received: June 26, 2014
Published online: August 27, 2014
Figure 4. A neighbor-joining tree cladogram of DHs (A) and KRs (B).
~
*
*
A) DH*-like ( ); MaoC-family ( ); FabA-family ( ); DHs involved in
Keywords: dehydratase · FR901464 · ketoreductase ·
PKS starter · a-ketone reduction
.
*
*
PKSs and fatty acid synthases (&). B) KR*-like ( ); SDR-family ( );
KRs involved in PKSs and fatty acid synthases (&); a-KRs involved in
NRPSs ( ); a/b-KR in PksJ ( ). The details for every gene/protein are
~
!
provided in the Supporting Information.
[1] a) E. S. Sattely, M. A. Fischbach, C. T. Walsh, Nat. Prod. Rep.
2008, 25, 757 – 793; b) C. Hertweck, Angew. Chem. Int. Ed. 2009,
48, 4688 – 4716; Angew. Chem. 2009, 121, 4782 – 4811.
[2] a) A. T. Keatinge-Clay, Nat. Prod. Rep. 2012, 29, 1050 – 1073;
b) D. Gay, Y.-O. You, A. Keatinge-Clay, D. E. Cane, Biochem-
istry 2013, 52, 8916 – 8928.
[3] a) D. E. Cane, J. Biol. Chem. 2010, 285, 27517 – 27523; b) D. H.
Kwan, F. Schulz, Mol. 2011, 16, 6092 – 6115.
[4] a) S. Bali, H. M. OꢀHare, K. J. Weissman, ChemBioChem 2006,
7, 478 – 484; b) S. Bali, K. J. Weissman, ChemBioChem 2006, 7,
1935 – 1942.
[5] a) H. Nakajima, B. Sato, T. Fujita, S. Takase, H. Terano, M.
Okuhara, J. Antibiot. 1996, 49, 1196 – 1203; b) D. Kaida, H.
Motoyoshi, E. Tashiro, T. Nojima, M. Hagiwara, K. Ishigami, H.
Watanabe, T. Kitahara, T. Yoshida, H. Nakajima, T. Tani, S.
Horinouchi, M. Yoshida, Nat. Chem. Biol. 2007, 3, 576 – 583.
[6] F. Zhang, H.-Y. He, M.-C. Tang, Y.-M. Tang, Q. Zhou, G.-L.
Tang, J. Am. Chem. Soc. 2011, 133, 2452 – 2462.
Unlike b-KRs, a-KRs are rare in PKS systems. In the
biosynthesis of bacillaene, a KR domain in PksJ is thought to
be a bifunctional enzyme that could catalyze both a- and b-
ketoreduction.[9] Other reports have revealed that such a-
KRs are either integrated into NRPS loading modules with
domain organization of A-KR-PCP (cereulide and valinomy-
cin)[20] or integrated into an extension module with C-A-KR-
PCP (antimycin).[21] Sequence analysis showed that these a-
KRs and PKS b-KRs belong to the same family of the KR
domain (PF08659), while the present KR* belong to the
family of short-chain dehydrogenase/reductases (SDR,
PF00106). Previous studies have established that the SDRs
possess a catalytic tetrad of N-S-Y-K residues,[22] and multiple
alignment revealed these KR*s have a similar conserved S-Y-
K catalytic triad (Supporting Information, Figure S7). As
11318 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 11315 –11319

Angewandte
Chemie
[7] M. Hildebrand, L. E. Waggoner, H. Liu, S. Sudek, S. Allen, C.
Anderson, D. H. Sherman, M. Haygood, Chem. Biol. 2004, 11,
1543 – 1552.
[8] H.-Y. He, H.-X. Pan, L.-F. Wu, B.-B. Zhang, H.-B. Chai, W. Liu,
G.-L. Tang, Chem. Biol. 2012, 19, 1313 – 1323.
[9] C. T. Calderone, S. B. Bumpus, N. L. Kelleher, C. T. Walsh, N. A.
Magarvey, Proc. Natl. Acad. Sci. USA 2008, 105, 12809 – 12814.
[10] D. H. Scharf, P. Chankhamjon, K. Scherlach, T. Heinekamp, M.
Roth, A. A. Brakhage, C. Hertweck, Angew. Chem. Int. Ed.
2012, 51, 10064 – 10068; Angew. Chem. 2012, 124, 10211 – 10215.
[11] F. Kopp, U. Linne, M. Oberthꢁr, M. A. Marahiel, J. Am. Chem.
Soc. 2008, 130, 2656 – 2666.
[12] X. Liu, S. Biswas, M. G. Berg, C. M. Antapli, F. Xie, Q. Wang,
M.-C. Tang, G.-L. Tang, L. Zhang, G. Dreyfuss, Y.-Q. Cheng, J.
Nat. Prod. 2013, 76, 685 – 693.
[16] a) K. Arakawa, F. Sugino, K. Kodama, T. Ishii, H. Kinashi,
Chem. Biol. 2005, 12, 249 – 256; b) J. S. Dickschat, O. Vergnolle,
H. Hong, S. Garner, S. R. Bidgood, H. C. Dooley, Z. Deng, P. F.
Leadlay, Y. Sun, ChemBioChem 2011, 12, 2408 – 2412.
[17] J. W. Labonte, C. A. Townsend, Chem. Rev. 2013, 113, 2182 –
2204.
[18] T. Hisano, T. Tsuge, T. Fukui, T. Iwata, K. Miki, Y. Doi, J. Biol.
Chem. 2003, 278, 617 – 624.
[19] Y. M. Qin, A. M. Haapalainen, S. H. Kilpelꢂinen, M. S. Marttila,
M. K. Koski, T. Glumoff, D. K. Novikov, J. K. Hiltunen, J. Biol.
Chem. 2000, 275, 4965 – 4972.
[20] N. A. Magarvey, M. Ehling-Schulz, C. T. Walsh, J. Am. Chem.
Soc. 2006, 128, 10698 – 10699.
[21] a) Y. Yan, L. Zhang, T. Ito, X. Qu, Y. Asakawa, T. Awakawa, I.
Abe, W. Liu, Org. Lett. 2012, 14, 4142 – 4145; b) M. Sandy, Z.
Rui, J. Gallagher, W. Zhang, ACS Chem. Biol. 2012, 7, 1956 –
1961.
[22] U. Oppermann, C. Filling, M. Hult, N. Shafqat, X. Wu, M. Lindh,
J. Shafqat, E. Nordling, Y. Kallberg, B. Persson, H. Jçrnvall,
Chem.-Biol. Interact. 2003, 143 – 144, 247 – 253.
[23] a) P. Caffrey, ChemBioChem 2003, 4, 654 – 657; b) A. T. Kea-
tinge-Clay, Chem. Biol. 2007, 14, 898 – 908.
[13] Y. Zou, H. Yin, D. Kong, Z. Deng, S. Lin, ChemBioChem 2013,
14, 679 – 683.
[14] S. I. Elshahawi, A. E. Trindade-Silva, A. Hanora, A. W. Han,
M. S. Flores, V. Vizzoni, C. G. Schrago, C. A. Soares, G. P.
Concepcion, D. L. Distel, E. W. Schmidt, M. G. Haygood, Proc.
Natl. Acad. Sci. USA 2013, 110, E295 – E304.
[15] S. Mochizuki, K. Hiratsu, M. Suwa, T. Ishii, F. Sugino, K.
Yamada, H. Kinashi, Mol. Microbiol. 2003, 48, 1501 – 1510.
[24] Y. A. Chan, A. M. Podevels, B. M. Kevany, M. G. Thomas, Nat.
Prod. Rep. 2009, 26, 90 – 114.
Angew. Chem. Int. Ed. 2014, 53, 11315 –11319
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11319