Biochemistry-Guided Prediction of the Absolute Configuration of Fungal Reduced Polyketides

Article abstract of DOI:10.1002/anie.202110658

Highly reducing polyketide synthases (HR-PKSs) produce structurally diverse polyketides (PKs). The PK diversity is constructed by a variety of factors, including the β-keto processing, chain length, methylation pattern, and relative and absolute configurations of the substituents. We examined the stereochemical course of the PK processing for the synthesis of polyhydroxy PKs such as phialotides, phomenoic acid, and ACR-toxin. Heterologous expression of a HR-PKS gene, a trans-acting enoylreductase gene, and a truncated non-ribosomal peptide synthetase gene resulted in the formation of a linear PK with multiple stereogenic centers. The absolute configurations of the stereogenic centers were determined by chemical degradation followed by comparison of the degradation products with synthetic standards. A stereochemical rule was proposed to explain the absolute configurations of other reduced PKs and highlights an error in the absolute configurations of a reported structure. The present work demonstrates that focused functional analysis of functionally related HR-PKSs leads to a better understanding of the stereochemical course.

Full text of DOI:10.1002/anie.202110658

Angewandte
Research Articles
Chemie
How to cite:
International Edition: doi.org/10.1002/anie.202110658
German Edition: doi.org/10.1002/ange.202110658
Polyketides
Hot Paper
Biochemistry-Guided Prediction of the Absolute Configuration of
Fungal Reduced Polyketides
Junya Takino, Akari Kotani, Taro Ozaki, Wenquan Peng, Jie Yu, Yian Guo, Susumu Mochizuki,
Kazuya Akimitsu, Masaru Hashimoto, Tao Ye,* Atsushi Minami,* and Hideaki Oikawa*
Abstract: Highly reducing polyketide synthases (HR-PKSs)
produce structurally diverse polyketides (PKs). The PK
diversity is constructed by a variety of factors, including the
b-keto processing, chain length, methylation pattern, and
relative and absolute configurations of the substituents. We
examined the stereochemical course of the PK processing for
the synthesis of polyhydroxy PKs such as phialotides, phome-
noic acid, and ACR-toxin. Heterologous expression of a HR-
PKS gene, a trans-acting enoylreductase gene, and a truncated
non-ribosomal peptide synthetase gene resulted in the forma-
tion of a linear PK with multiple stereogenic centers. The
absolute configurations of the stereogenic centers were deter-
mined by chemical degradation followed by comparison of the
degradation products with synthetic standards. A stereochem-
ical rule was proposed to explain the absolute configurations of
other reduced PKs and highlights an error in the absolute
configurations of a reported structure. The present work
demonstrates that focused functional analysis of functionally
related HR-PKSs leads to a better understanding of the
stereochemical course.
structure of biologically active polyketides (PKs), including
cholesterol-lowering lovastatin, virulence factor T-toxin, and
the squalene synthase inhibitors zaragozic acids. HR-PKSs
have a single set of catalytic domains whose organization
resembles with that of fatty acid synthases (FASs).[1–4]
Although both HR-PKSs and FASs use the catalytic domains
repeatedly in their chain elongation processes, they differ in
the programming of their b-keto processing domains. In the
case of HR-PKSs, the b-keto processing domains operate to
varying extents during iteration, thereby affording structur-
ally diverse PKs.[1,2] The structural diversity can also be
attributed to a variety of other factors, including the chain
length and methylation pattern. In bacterial PKSs, the
absolute configurations of the stereogenic centers are pre-
dicted by motif analysis and the programming rule. However,
a similar corresponding rule has not been established for
fungal HR-PKSs.
During PK processing, HR-PKSs generate acyl carrier
protein (ACP)-tethered intermediates with multiple stereo-
genic centers. Understanding the stereochemical course
leading to fungal PKs is important in the functional analysis
of HR-PKSs. Direct prediction of the stereochemical course
from the structure of natural PKs is difficult because: 1) The
hydroxy group installed by a ketoreductase (KR) domain is
frequently lost during PK processing. As a result, in many
cases fungal PKSs give simple polyenes or saturated fatty
acids with few substituents. 2) Post-PKS oxidation installs
additional oxygen functionalities, making it difficult to
determine the origin of the oxygen atom without experimen-
tal support. 3) In some cases, a single KR domain produces
the products having hydroxy groups with different configu-
rations,[5,6] preventing understanding of the stereochemical
rule for HR-PKS catalysis. To overcome these problems,
administration of isotopically labeled precursors and in vitro
enzymatic reactions using a recombinant protein have
provided insight into the stereospecificity of the individual
HR-PKS.[7,8] However, to our knowledge, unified under-
standing of the stereochemical course of the HR-PKS syn-
thesis of structurally and biosynthetically related fungal PKs
is limited because of the lack of focused analysis of function-
ally related HR-PKSs, which hampers the proposal of
a general mechanism for HR-PKSs.
Introduction
Fungal highly reducing polyketide synthases (HR-PKSs)
are multifunctional enzymes for synthesis of the backbone
[*] J. Takino, A. Kotani, Dr. T. Ozaki, Dr. A. Minami, Prof. Dr. H. Oikawa
Department of Chemistry, Faculty of Science, Hokkaido University
Sapporo 060-0810 (Japan)
E-mail: aminami@sci.hokudai.ac.jp
hoik@sci.hokudai.ac.jp
W. Peng, Dr. J. Yu, Dr. Y. Guo, Prof. Dr. T. Ye
State Key Laboratory of Chemical Oncogenomics, Peking University
Shenzhen Graduate School
Xili, Nanshan District, Shenzhen 518055 (China)
E-mail: yet@pkusz.edu.cn
Dr. S. Mochizuki, Prof. Dr. K. Akimitsu
International Institute of Rare Sugar Research and Education &
Faculty of Agriculture, Kagawa University
Kagawa, 761-0795 (Japan)
Prof. Dr. M. Hashimoto
Faculty of Agriculture and Life Science, Hirosaki University
Hirosaki 036-8561 (Japan)
The most reliable method to elucidate the stereospecific-
ity of HR-PKSs is in vitro analysis.[8] However, this method
requires protein expression optimization and purification.
Recently, heterologous expression of HR-PKSs in fungal
versatile hosts has become a practical method for functional
analysis.[9] This is an effective alternative to in vitro analysis
Dr. J. Yu, Dr. Y. Guo
School of Biotechnology and Health Sciences, Wuyi University
Jiangmen 529020 (China)
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202110658.
&&&&
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Figure 1. Chemical structures of fungal polyhydroxy polyketides.
because of the high versatility of PKS expression. Based on
the background described above, we hypothesized that a small
group of functionally related HR-PKSs share the same
stereochemical course, and therefore the focused analysis of
such HR-PKSs may provide a general mechanism of PKS
processing.
homologous NRPS (PhomC, 39% identity) was found near
the proposed BGC of 2. Although the expression profile of
the NRPS gene was different from that of PKS2,[13] the NRPS
gene is conserved in other homologous BGCs found in
Stagonospora nodorum and Metharhizium acridum.[13] Taken
together, we hypothesized that PhiaA-E are candidates
participating in the biosynthesis of 1.
To determine whether the identified genes are sufficient
for production of the PK chain of 1, we constructed trans-
formants, AO-phiaABCD, AO-phiaABC, AO-phiaAB, and
AO-phiaAC. The AO-phiaABCD produced new product 5a
(C29H50O6), which was named as prophialotide (Figure 2). The
1H and 13C NMR data of 5a resembled those of the PK part of
1. Extensive 2D NMR analysis revealed a planar structure.
In this article, we describe the functional analysis of HR-
PKSs that produce polyhydroxy PKs with multiple stereo-
genic centers, such as phialotide A (1),[10] phomenoic acid
(PMA, 2),[11] and ACR-toxin (3a)[12] (Figure 1). Determina-
tion of the relative and absolute configurations of an
unmodified PK by degradation studies enabled us to propose
a general stereochemical rule for the PK processing. This rule
was successfully applied to related PKS families that produce
different polyhydroxy PKs; however, violations of the rule
were found for several reported configurations in thermo-
lides. This prompted us to examine computational chemical
shift predictions and synthetic studies of the model com-
pound, resulting in the proposal of a revised structure. We
also discuss the extensive application of the rule to other HR-
PKSs.
Results and Discussion
Recently, a putative biosynthetic gene cluster (BGC),
phom, that is involved in the biosynthesis of 2 was identified
by gene silencing experiments and quantitative reverse
transcriptase PCR analysis.[13] The HR-PKS gene, phomA
(previous name: PKS2), accompanies four genes: a trans-ER
(t-ER) gene (phomB), a cytochrome P450 gene (phomE),
a functionally unknown gene (phomD), and a transcription
factor. Considering the structural similarity between 1 and 2
that possess multiple numbers of substituents, including
a hydroxy group, an E-olefin, and a branched methyl group,
which are regularly installed on the single PK chain to form an
alternately repeating a dehydrated and a non-dehydrated
substructure (Figure 1), we hypothesized that a HR-PKS gene
as well as modification genes are highly conserved in the BGC
of 1. A BLAST search allowed us to identify a putative BGC
(phia) in the genome of Pseudophialophora sp. BF-0158,
which produces 1 (Figure S1 and Table S1). The BGC
includes the four homologous genes described above: HR-
PKS (PhiaA); 58% (identity), t-ER (PhiaB); 63%, unknown
protein (PhiaD); 42%, and cytochrome P450 (PhiaE); 33%.
An NRPS gene, phiaC, was also identified in phia BGC. The
Figure 2. UPLC-MS profiles of the metabolites from transformants
harboring biosynthetic genes of A) phialotide, B) phomenoic acid, and
C) ACR-toxin. i) AO-WT, ii) AO-phiaABC, iii) AO-phiaAB, iv) AO-phiaAC,
v) AO-phiaABC(HAxxDG mut.), vi) AO-phiaABC(AHxxDG mut.), vii)
AO-phiaABC(AAxxDG mut.), viii) AO-WT, ix) AO-phomABC, x) AO-WT,
and xi) AO-ACRTS2.
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
MS/MS analysis showed fragmentation ion peaks that re-
flected the repeated substructure (Figure S2). 5a has all the
oxygen and methyl substituents found in 1, suggesting they
are installed by the collaborative action of PhiaA and PhiaB.
To determine the relative and absolute configurations, we
prepared two alcohols, 7-1 and 8-1 (single isomer), by
ozonolysis of 5a followed by a reductive workup with NaBH4
(Figures 3 and S3). In the case of 7-1, we synthesized four
diastereomers as synthetic standards (STDs) from known
compounds (Scheme S1). Chiral GC/MS analysis showed that
the relative configuration of 7-1 was determined to be
(2S,3R,4S) by comparison with the STDs 7-1a–7-1d. This
the producer (Figure 2 and Scheme 1).[11] The structure was
confirmed by 1D and 2D NMR analysis. Similar to the process
used for the analysis of 5a, we conducted an ozonolysis of 6 to
obtain four fragments, 8-1, 9-1, 10-1 (single isomer), and 11-1,
to determine their absolute configurations (Figures 3 and S3).
The absolute configuration of 8-1 was determined to be
(2S,3S). Similar to the procedures used for 8-1, we applied
a stepwise derivatization of 10-1 and determined the absolute
configuration as 2S. For 9-1, we synthesized four chiral STDs,
9-1a–9-1d, and the racemic mixture 9-1e. Chiral GC-MS
analysis showed that 9-1 had the same retention time as 9-1a,
which had a (2S,4S,6S) configuration. UPLC-MS analysis of
the (S)-MaNP derivative further supported the absolute
configuration of 9-1. In the case of 11-1, we synthesized two
diastereomers, 11-1a and 11-1b. They were derivatized with
(S)-MaNP acid to afford 11-2a and 11-2b, which were
separable by UPLC-MS with a non-chiral column. The
chromatographic behavior of 11-2 was the same as 11-2a,
but different from the corresponding (R)-MaNP derivative
11-3a, demonstrating that 11-1 had an (3R,5S) configuration.
Taken together, the absolute configuration of 6 was deter-
mined to be (3R,5S,9S,13S,16S,17S,22S,24S,26S).
1
was further supported by H NMR analysis (Figure S3). The
relative configuration of 8-1 was determined in a similar
manner using synthetic STDs 8-1a (anti) and 8-1b (syn)
(Scheme S1). The retention time of the benzylidene acetal 8-2
was the same as that of 8-2a. The absolute configurations of
the degradation products were then determined after the
derivatization to both (R)-MaNP and (S)-MaNP esters.[14]
UPLC-MS analysis showed that 7-2 and 8-3 were identical to
7-2a and 8-3a, respectively. We therefore concluded
that
the
absolute
configuration
of
5a
was
(4S,5S,8S,9S,12S,13S,16S,17R,18S).
To obtain further insight into the stereochemical course of
this type of PK, we then conducted a heterologous expression
of the ACRTS2 involved in the biosynthesis of 3a because
ACRTS2 exhibits homology with PhiaA (55% identity) and
PhomA (68%). Because t-ER is not essential for synthesis of
the PK backbone of 3a, we introduced an ACRTS2 gene into
A. oryzae to construct AO-ACRTS2. Metabolite analysis
showed the production of 3b (C18H30O4) as a major product
(Figure 2 and Scheme 1), which is known as a decarboxylation
product of ACR-toxin.[12] The NMR data and optical rotation
value ([a]D22 =+ 58 [lit. [a]D21 =+ 44]) were in good agree-
ment with those previously reported. The proposed decar-
boxylation mechanism to afford 3b is shown in Scheme S2.
Currently, we speculate that an intramolecular non-enzymatic
cyclization affords 3a as in the case of shimalactone
biosynthesis.[16]
Based on the above results, we were able to propose the
PK processing leading to 5, 6, and 3a (Schemes 1 and S3). The
biosynthetic Scheme to synthesize 5a and 5b is summarized in
Scheme 1. The characteristic repeating substructure is con-
structed by successive dehydration (DH) skipping and enoyl
reduction (ER) skipping during PK processing. During the
PK processing of 5 and 6, t-ER reduces the double bond of the
a,b-unsaturated thioester in the initial stage to yield a satu-
rated thioester. Considering that 6 has a 1,3-syn dimethyl
structure, this methyl-terminal enoyl reduction is likely
a characteristic feature of this type of PK. After processing,
the ACP-tethered PK subsequently undergoes hydrolysis by
a truncated NRPS to afford a linear PK. To our knowledge,
a truncated NRPS that catalyzes hydrolysis is rare in fungal
PK biosynthesis.
UPLC-MS analysis of the metabolites from other trans-
formants showed that AO-phiaABC produced 5a while AO-
phiaAB did not, suggesting that the hypothetical protein
PhiaD is not essential for the production of 5a, but a truncated
NRPS, PhiaC, is indispensable. AO-phiaAC, which lacks
a trans-ER gene (phiaB) produced a new metabolite 5b
(C29H48O6) (Figure 2). 1D and 2D NMR analyses revealed
that this is a non-reduced product of 5a, possessing an E-
olefin at the methyl-terminal position (Scheme 1). MS/MS
analysis supported this proposed structure.
We then turned our attention to the hydrolysis of the
ACP-tethered intermediate because PhiaA has no chain-
releasing domain such as a thioesterase (TE) domain. Given
the fact that 5a was produced only in the transformants
harboring phiaC, we hypothesized that PhiaC, a truncated
NRPS lacking a thiolation (T) domain, mediates the hydrol-
ysis reaction instead of a canonical condensation reaction.
The proposed hydrolysis was examined by mutational anal-
ysis, with particular focus on the active site HHxxxDG motif
of the C domain[15] because the C domain catalyzes reactions
À
such as intramolecular lactonization (C O bond formation)
and amide bond formation using an amino acid.[15] Three
mutants (HAxxxDG, AHxxxDG, and AAxxxDG) were
constructed and co-expressed with PhiaA and PhiaB in A.
oryzae. The transformant possessing an HAxxxDG mutant
produced 5a, while the other transformants did not, strongly
suggesting that the C domain mediates the hydrolysis reaction
(Figure 2).
Identification of three essential enzymes for 5a produc-
tion enabled us to examine the heterologous production of
the PK chain of 2. We prepared two plasmids, pDP801:pho-
mA and pDP201:phomBC, and introduced them into A.
oryzae to construct AO-phomABC. UPLC-MS analysis of the
metabolites from AO-phomABC showed the production of
new metabolite 6 (C34H58O7), of which molecular formula is
identical to that of deoxy derivative of 2 lactone isolated from
We then turned our attention to the stereochemical course
used to synthesize 5, 6, and 3a (Scheme 2). To simplify
discussions about the stereogenic centers, we are applying
o
a newly defined R/oS model that focuses on the direction of
the PK chain elongation and does not consider the absolute
configurations of other substituents located nearby. Following
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Figure 3. A) Chemical degradation of 5a and 6. Degradation product 8-1 was obtained as diastereomeric mixture at the C4 position labeled by an
asterisk, which was installed by the NaBH4 reduction of the ozonolysis product. B) Derivatization of the ozonolysis products. C) Determination of
the absolute configuration of 7-1. D) Determination of the absolute configuration of 8-1. To simplify the discussions, we focused on the major
isomers, which are labeled with #. E) Chemical structures of the synthetic standards.
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Scheme 1. A) Proposed biosynthetic pathways of 5a and 5b. Domain organizations of PhiaA-C are enclosed by a dotted line. B) Chemical
structures of 6 and 3b synthesized by PhomABC and ACRTS2, respectively. Their proposed biosyntheses are summarized in the Supplementary
Information. Carbons with a substituent installed by MT, KR, and ER domain are labeled with blue, red, and green circle, respectively. These labels
are used in the following figures and schemes.
o
reduction by t-ER gives IntE possessing an SER-methyl
group. This stereochemical course is basically the same as that
of fatty acid biosynthesis.[7,17]
Polyhydroxy PKs are an emerging family of fungal PKs,
although they are a relatively small group (Figures 4 and S4).
Polyhydroxy PKs can be divided into three groups according
to the PK chain-releasing mechanism. The first PMA group
consists of linear PKs with a terminal carboxylic acid such as
1 and 2 (Figure S4), which is hydrolyzed by a truncated NRPS.
The second TML group is a macrolactam family PKs
represented by thermolide (TML),[18] which are composed
of a polyhydroxy PK and an amino acid (Figure S4). The PK
construction is catalyzed by HR-PKS with an active ER
domain. The characteristic macrolactam moiety is construct-
ed by a complete NRPS composed of condensation (C),
adenylation (A), thiolation (T), and terminal condensation-
like (CT) domains (Figure S5). This group includes metacri-
Scheme 2. Stereochemical course leading to polyketides classified into
PMA, TML, and PSL groups. In our stereochemical model, we defined
damides,[19] georatusin,[20] and valactamides.[21] The third PSL
that a carboxy terminal chain (RC) has higher priority than a methyl
terminal chain (RM). The configuration is described by a capital letter
and the domain used to install the substituent is described by
a subscript.
group is a macrolactone family of PKs represented by
phaeospelide A (PSL)[22] (Figure S4). The characteristic mac-
rocyclization is catalyzed by a stand-alone TE. Unlike HR-
PKSs for PMA and TML PKs, HR-PKSs in the macrolactone
family lack the MT domain. This domain organization is
consistent with the fact that representative macrolactone PKs,
such as decarestrictine C1 (HR-PKS; DcsA),[23] AKMLA
(AkmlA),[24] and CIMLA (CimlA),[24] have no branched
methyl group.
To obtain insight into the biosynthetic relationship of
these three polyhydroxy PK groups, we conducted a phyloge-
netic analysis as reported in our recent studies on PKS-NRPS
hybrids[25] (Figure S6). When focusing on functionally char-
acterized HR-PKSs, we find three distinct clades correspond-
the formation of IntA, the methyltransferase (MT) domain
o
installs an RMT-methyl group at the a position to give IntB,
which is reduced by the KR domain to yield IntC with an
oRKR-hydroxy group. The hydroxythioester then undergoes
syn-elimination to afford IntD with an E-olefin by the action
of a dehydratase (DH) domain. This stereochemical course
o
leading to an E-olefin via an R-hydroxythioester indicates
a close stereochemical and biosynthetic relationships between
the three domains in the PK chain processing. The final enoyl
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Figure 4. Chemical structures of PKs synthesized by HR-PKSs in the A) PSL clade and B) TML clade. The putative ACP-tethered intermediate is
described in the brackets. A mismatched stereochemistry is labeled by an asterisk. Three sereogenic centers found in thermolide that were re-
examined by ab initio NMR chemical shift calculations, are highlighted by the grey color.
ing to the PMA (linear PKs), TML (macrolactam PKs), and
PSL (macrolactone PKs) groups. In the case of the former
PMA and TML clades, either NRPS or the truncated NRPS
catalyze the cleavage of the PK chain from HR-PKS. In
contrast, in the PSL clade, a stand-alone TE catalyzes this
cleavage. Taken together, we found that phylogeny-based
classification of HR-PKSs is useful for predicting the type of
chain-releasing mechanism, which directly reflects the struc-
ture of the PKs. Indeed, the PMA clade includes HR-PKSs
from a producer of cubensic acid (strain; Xylaria cubensis)
and TMC-171 (Clonostachys rosea) (Figure 1), of which the
structures of the PK moiety resemble those of 1–3.
Ac-pTML (14), is (2S,8S,9R,11R,12S). The configurations at
positions 9, 11, and 12 violate the proposed stereochemical
rule. To re-examine the absolute configuration on the macro-
cyclic ring, we performed DFT-based NMR chemical shift
calculations[29] for model compounds simplified the side chain
with an isopropyl group using Hehreꢀs protocol. DP4
probability analysis showed that, among 16 possible dia-
steromers, the probability for model compound B, which has
the reported configuration (9R,11S,12S), was only 0.3% when
Goodmanꢀs parameters (s = 2.306 ppm, n = 11.38) were used.
The analysis also suggested an exclusively high probability for
the predicted (9S,11R,12R) model compound A (87.2%)
(Figure S7).
To obtain conclusive evidence that model compound A
represents the relative configuration of the corresponding
fragment of the natural substance, an asymmetric synthesis of
A was carried out as shown in Scheme 3. The synthesis thus
commenced with the stereoselective hydrogenation of the
known a,b-unsaturated imide 15[30] over Pd/Al2O3 to afford 16
as a mixture of two diastereomers (4:1 by 1H NMR).[31] After
chromatographic purification, the major diastereomer 16 was
obtained in 72% yield. The hydroxyl group of 16 was next
protected as a Bn ether with benzyl trichloroacetimidate (17),
followed by reductive removal of the chiral auxiliary with
NaBH4 to provide the primary alcohol 18 in 91% yield over
two steps. Treatment of alcohol 18 with TCCA and TEM-
PO[32] afforded the corresponding aldehyde, which was
subjected to a vinylogous aldol reaction under Krꢁger and
Carreiras copper-catalyzed conditions[33] with silyloxy diene
19 to give rise to b-hydroxy dioxinone 20 in 65% yield and
a diastereomeric ratio (> 99:1) favoring the desired syn
The PMA group of PKs, such as 5, 6, and 3a, commonly
o
o
have RMT
,
oRKR, and SER configurations installed by HR-
PKS/t-ER (Scheme 2). This stereochemical rule can explain
the absolute configurations of the structurally related TMC-
171 (4),[26] of which its producer, C. rosea, has a single BGC
classified into the PMA clade (Figure 1). Presuming that the
PMA clade BGCs found in Xylaria sp. are involved in the
biosynthesis of structurally related PKs such as cubensic acid
(producer: Xylaria cubensis),[27] malaysic acid (Xylaria sp.),[28]
berteric acid (Xylaria sp.),[28] and cameronic acid (Xylaria
sp.)[28] (Figure S1), we can predict the absolute configurations
based on the proposed stereochemical course (Figure S4).
Based on the established backbone structures of macro-
cyclic PKs in the TML and PSL groups (Figures 4 and S4), we
also propose that the stereochemical course can basically be
applied to PKs such as metacridamide A (TML group).
Exceptional PK is thermolides.[18] The reported absolute
configuration of the macrolactam moiety of thermolides,
prothermolide (pTML, 12), 16-O-Ac-pTML (13), and 18-O-
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Scheme 3. Synthesis of model compound A. TCCA=trichloroisocyanuric acid, TBAT=tetrabutylammonium difluorotriphenylsilicate, TCBC=2,4,6-
trichlorobenzoyl chloride, TASF=tris(dimethylamino)sulfonium difluorotrimethylsilicate.
adduct. Protection of 20 as its TBS ether and hydrogenolytic
removal of the benzyl ether to afford the corresponding
alcohol 21, which was then condensed with Cbz-Ala under
Yamaguchi conditions, furnished 22 in 80% yield over three
steps. Hydrogenolysis of the Cbz protecting group cleanly
afforded the amine, which was heated in toluene at reflux
under dilute conditions resulted in macrocyclization to give
the expected b-ketoamide[34] followed by deprotection of the
TBS ether with TASF, provided A in 68% yield over three
steps. The 13C-NMR spectrum of A was nearly identical to
those of 12–14 (macrocyclic ring). The deviations of the
13C NMR chemical shifts of the synthetic model compound A
with 12–14 are within Æ 0.7 ppm (Figure S8). Based on the
results, we propose a structural revision of the absolute
configurations on the macrolactam moiety of the thermolides.
This is a good example that the stereochemical rule is useful
for predicting the absolute configurations of known PKs that
have been classified into TML and PSL clades.
clades (Figure S6). Given the fact that HR-PKSs in PMA/
TML/PLS clades synthesize a PK chain via the R-hydrox-
o
ythioester, HR-PKSs classified into those new clades may
catalyze PK processing in the same stereospecific manner. In
contrast to the MT, KR, and DH domains, the ER domain can
act independently, as seen in the case of mFAS and bacterial
o
multimodular PKS.[7,17] Although the SER configuration is
o
predominant (Figure S11), there are PKs possessing all RER
configurations, such as fumonisin B1, betaenone, and scy-
phostatin (Figure S12). This optional selectivity has been
retained during evolution of the HR-PKSs to synthesize
structurally diverse PKs.
There are a few exceptions to the proposed stereochem-
ical rule that have mismatched stereochemistry (Figures 5, S4,
and S13). For example, phaeospelide A is a PSL-type PK that
o
possesses an SKR-configuration at the methyl-terminal posi-
tion (Figure S4). Given the fact that all of the remaining
o
hydroxy groups have RKR configuration, the stereochemical
Given the close relationship between the MT, KR, and
DH domains, the stereospecificity of the ketoreduction is
a key issue for understanding the stereochemical course
catalyzed by HR-PKSs. Based on the well-established ste-
reochemical model for bacterial PKSs that A type and B type
violation occurs only in the first round of PK processing. A
similar observation is found in resorcylic acid lactones and
structurally related derivatives such as hypothemycin,[5,6,40]
dehydrocurvularin,[41] and monocerin[42] (Figure S13). Stereo-
chemical violation at a methyl-terminal position is also
observed in other domains. The DH domain of Fus1 involved
in fusarin biosynthesis gives a Z-olefin at the methyl-terminal
position,[43] while the same DH domain generates an E-
configurated polyene backbone (Figure 5). Considering that,
in general, a Z-olefin is synthesized by syn-elimination of an
oSKR-hydroxythioester during PK processing,[44] the methyl-
terminal Z-olefin found in fusarin is likely constructed via an
oSKR-hydroxythioester. Similar Z-olefin formation has also
been reported in the biosynthesis of aminoacylated prod-
ucts.[45] In addition, t-ERs in the biosynthesis of atpenin[46] and
KRs install SKR- and RKR-configurations, respectively,[35] we
examined motif analysis of the KR domains of fungal HR-
PKSs. We found that the KR domains of most HR-PKSs,
including HR-PKSs in PMA, TML, and PSL clades, as well as
mammalian FAS, have characteristic motifs found in bacterial
B-type KRs (Figure S10). Representative HR-PKSs with a B-
type_KR domain include SQTKS (natural product, squales-
tatin),[9b] CazF (chaetoviridin),[36] LovB and LovF (lovasta-
o
o
tin),[37] GLPKS4 (pneumocandin),[38] and EasB (emericella-
o
mide)[39] (Figure 5), which synthesize a PK chain with RKR
-
-
OH. These results suggest that most HR-PKSs afford an oRKR
aspyridone[47] generate an RER-methyl group in the first
o
OH during PK processing to synthesize an E-olefin (Fig-
round of PK processing (Figure S13), although they generate
oSER-configurations in later steps. Detailed functional analysis
of each domain would provide insight into the site-specific
configurational violations found in those PKs. However, when
we exclude these minor examples, the prediction rule is
otherwise applicable to fungal reduced PK metabolites.
ure S11). Indeed, Cox and co-workers demonstrated that the
DH domain of SQTKS catalyzes a syn-elimination of an oRKR
-
hydroxythioester to give an E-olefin.[8] It should also be
notable that functionally characterized HR-PKSs described
above make distinct clades different from PMA/TML/PSL
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Figure 5. A) Chemical structures of PKs synthesized by HR-PKSs in other clades. B) Examples of PKs with mismatched stereochemistry. The
corresponding ACP-tethered intermediate is described in the brackets.
supported under Contract No. DE-AC02-05CH11231 (H.O.).
T.Y. acknowledges financial support from the Shenzhen
Conclusion
Peacock Plan (KQTD2015071714043444).
In this study, we elucidated a unified stereochemical
course for polyhydroxy PKs, phialotides, phomenoic acid, and
ACR-toxin by determining the relative and absolute config-
urations of the unmodified PKs synthesized by HR-PKS. The
proposed stereochemical rule explains the absolute config-
urations of other polyhydroxy PKs in PMA, TML, and PSL
groups, and highlights errors in the reported absolute config-
urations on the macrolactam moiety of thermolides. Compu-
tation-assisted chemical shift prediction and a synthetic study
of model compound A, which lacks the side chain of
prothermolide, allowed us to propose a structural revision.
The proposed stereochemical rule is helpful for the structural
revision and prediction of fungal polyhydroxy PKs. Our
present work demonstrates that focused functional analysis of
functionally related HR-PKSs based on the phylogenetic
analysis is effective for understanding the stereochemical
course of PK processing.
Conflict of Interest
The authors declare no conflict of interest.
Keywords: bioinformatics analysis · fungal polyketides ·
heterologous expression · polyketide synthase ·
stereochemical course
[1] R. J. Cox, Org. Biomol. Chem. 2007, 5, 2010 – 2026.
[2] Y.-H. Chooi, Y. Tang, J. Org. Chem. 2012, 77, 9933 – 9953.
[3] D. A. Herbst, C. A. Townsend, T. Maier, Nat. Prod. Rep. 2018,
35, 1046 – 1069.
[4] J. Wang, J. Liang, L. Chen, W. Zhang, L. Kong, C. Peng, C. Su, Y.
Tang, Z. Deng, Z. Wang, Nat. Commun. 2021, 8, 867.
[5] H. Zhou, Z. Gao, K. Qiao, J. Wang, J. C. Vederas, Y. Tang, Nat.
Chem. Biol. 2012, 8, 331 – 333.
[6] W. Ding, Y. Li, Q. Zhang, ACS Chem. Biol. 2015, 10, 1590 – 1598.
[7] J. D. BuꢀLock, T. J. Simpson in Biosynthesis, Vol. 6 (Eds.: J. D.
BuꢀLock), Royal Society of Chemistry, London, 1980, pp. 1 – 39.
[8] a) E. Liddle, A. Scott, L.-C. Han, D. Ivison, T. J. Simpson, C. L.
Willis, R. J. Cox, Chem. Commun. 2017, 53, 1727 – 1730; b) D. M.
Roberts, C. Bartel, A. Scott, D. Ivison, T. J. Simpson, R. J. Cox,
Chem. Sci. 2017, 8, 1116 – 1126.
[9] a) J. Kennedy, K. Auclair, S. G. Kendrew, C. Park, J. C. Vederas,
C. R. Hutchinson, Science 1999, 284, 1368 – 1372; b) R. J. Cox, F.
Glod, D. Hurley, C. M. Lazarus, T. P. Nicholson, B. A. M. Rudd,
T. J. Simpson, B. Wilkinson, Y. Zhang, Chem. Commun. 2004,
Acknowledgements
We are grateful to Prof. Hiroshi Tomoda for giving us
a phialotides producer, Pseudophialophora sp. BF-0158. This
work was financially supported by Japan Society for the
Promotion of Science Grants [16H06446 (A.M.), 17K19214
(A.M.), 19H02835 (A.M.), and 19H02891 (H.O.)]. The work
also conducted by the U.S. Department of Energy Joint
Genome Institute, a DOE Office of Science User Facility,
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
2260 – 2261; c) S. M. Ma, J. W.-H. Choi, H. Zhou, K. K. M. Lee,
V. A. Moorthie, X. Xie, J. T. Kealey, N. A. Da Silva, J. C.
Vederas, Y. Tang, Science 2009, 326, 589 – 592; d) T. Ugai, A.
Minami, R. Fujii, M. Tanaka, K. Gomi, H. Oikawa, Chem.
Commun. 2015, 51, 1878 – 1881.
[28] M. O. Adeboya, R. L. Edwards, T. Laessoee, D. J. Maitland, J.
Chem. Res. Synop. 1995, 9, 356 – 357.
[29] D. A. Forsyth, A. B. Sebag, J. Am. Chem. Soc. 1997, 119, 9483 –
9494.
[30] S. Shirokawa, M. Kamiyama, T. Nakamura, M. Okada, A.
Nakazaki, S. Hosokawa, S. Kobayashi, J. Am. Chem. Soc. 2004,
126, 13604 – 13605.
[10] A. Yagi, R. Uchida, K. Kobayashi, H. Tomoda, J. Antibiot. 2020,
73, 211 – 223.
[11] a) M. Devys, J. P. Fꢂrꢂzou, R. S. Topgi, M. Barbier, J. Chem. Soc.
Perkin Trans. 1 1984, 2133 – 2137; b) M. Devys, R. S. Topgi, J.-P.
Fꢂrꢂzou, L. Quaino, J.-F. Bousquet, A. Kollmann, M. Barbier,
Phytochemistry 1986, 25, 531 – 532.
[31] T. Nakamura, M. Harachi, T. Kano, Y. Mukaeda, S. Hosokawa,
Org. Lett. 2013, 15, 3170 – 3173.
[32] L. De Luca, G. Giacomelli, S. Masala, A. Porcheddu, J. Org.
Chem. 2003, 68, 4999 – 5001.
[12] a) J. M. Gardner, Y. Kono, J. H. Tatum, Y. Suzuki, S. Takeuchi,
Phytochemistry 1985, 24, 2861 – 2867; b) J. M. Gardner, Y. Kono,
J. H. Tatum, Y. Suzuki, S. Takeuchi, Agric. Biol. Chem. 1985, 49,
1235 – 1238.
[33] a) J. Krꢁger, E. M. Carreira, J. Am. Chem. Soc. 1998, 120, 837 –
838; b) B. L. Pagenkopf, J. Krꢁger, A. Stojanovic, E. M. Car-
reira, Angew. Chem. Int. Ed. 1998, 37, 3124 – 3126; Angew.
Chem. 1998, 110, 3312 – 3314.
[13] C. E. Elliott, D. L. Callahan, D. Schwenk, M. Nett, D. hoffmeis-
ter, B. J. Howlett, Fungal Genet. Biol. 2013, 53, 50 – 58.
[14] In this study, we used MaNP acid instead of a-methoxy-a-
(trifluoromethyl)phenylacetic acid (MTPA) as a chiral derivatiz-
ing agent because separation of the resultant diasteromers is
always superior to that of the corresponding MTPA esters. a) Y.
Kasai, J. Naito, S. Kuwahara, M. Watanabe, A. Ichikawa, N.
Harada, J. Synth. Org. Chem. Jpn. 2004, 62, 1114 – 1127; b) N.
Harada, Molecules 2016, 21, 1328.
[34] R. K. Boeckman, J. R. Pruitt, J. Am. Chem. Soc. 1989, 111, 8286 –
8288.
[35] a) A. Baerga-Ortiz, B. Popovic, A. P. Siskos, H. M. OꢀHare, D.
Spiteller, M. G. Williams, N. Campillo, J. B. Spencer, P. F.
Leadlay, Chem. Biol. 2006, 13, 277 – 285; b) A. T. Keatinge-Clay,
Nat. Prod. Rep. 2016, 33, 141 – 149.
[36] J. M. Winter, M. Sato, S. Sugimoto, G. Chiou, N. K. Garg, Y.
Tang, K. Watanabe, J. Am. Chem. Soc. 2012, 134, 17900 – 17903.
[37] a) J. Kennedy, K. Auclair, S. G. Kendrew, C. Park, J. C. Vederas,
C. R. Hutchinson, Science 1999, 131, 8388 – 8389; b) X. Xie, M. J.
Meehan, W. Xu, P. C. Dorrestein, Y. Tang, J. Am. Chem. Soc.
2009, 131, 8388 – 8389.
[38] Y. Li, N. Lan, L. Xu, Q. Yu, Appl. Microbiol. Biotechnol. 2018,
102, 9881 – 9891.
[39] Y.-M. Chiang, E. Szewczyk, T. Nayak, A. D. Davidson, J. F.
Snachez, H.-C. Lo, H. Wen-Yueh, H. Simityan, E. Kuo, A.
Praseuth, K. Watanabe, B. R. Oakley, C. C. C. Wang, Chem.
Biol. 2008, 15, 527 – 532.
[40] C. D. Reeves, Z. Hu, R. Reid, J. T. Kealey, Appl. Environ.
Microbiol. 2008, 74, 5121 – 5129.
[41] R. V. K. Cochrane, Z. Gao, G. R. Lambkin, W. Xu, J. M. Winter,
S. L. Marcus, Y. Tang, J. C. Vederas, ChemBioChem 2015, 16,
2479 – 2483.
[42] L. C. Axford, T. J. Simpson, C. L. Willis, Angew. Chem. Int. Ed.
2004, 43, 727 – 730; Angew. Chem. 2004, 116, 745 – 748.
[43] E.-M. Niehaus, K. Kleigrewe, P. Wiemann, L. Studt, C. M. K.
Sieber, L. R. Connolly, M. Freitag, U. Gꢁldener, B. Tudzynski,
H.-U. Humpf, Chem. Biol. 2013, 20, 1055 – 1066.
[44] a) X. Guo, T. Liu, C. R. Valenzano, Z. Deng, D. E. Cane, J. Am.
Chem. Soc. 2010, 132, 14694 – 14696; b) D. Gay, Y.-O. You, A.
Keatinge-Clay, D. E. Cane, Biochemistry 2013, 52, 8916 – 8928.
[45] M.-C. Tang, C. R. Fischer, J. V. Chari, D. Tan, S. Suresh, A. Chu,
M. Miranda, J. Smith, Z. Zhang, N. K. Garg, R. P. St. Onge, Y.
Tang, J. Am. Chem. Soc. 2019, 141, 8198 – 8206.
[15] R. D. Sꢁssmuth, A. Mainz, Angew. Chem. Int. Ed. 2017, 56,
3770 – 3821; Angew. Chem. 2017, 129, 3824 – 3878.
[16] I. Fujii, M. Hashimoto, K. Konishi, A. Unezawa, H. Sakuraba, K.
Suzuki, H. Tsushima, M. Iwasaki, S. Yoshida, A. Kudo, R. Fujita,
A. Hichiwa, K. Saito, T. Asano, J. Ishikawa, D. Wakana, Y. Goda,
A. Watanabe, M. Watanabe, Y. Masumoto, J. Kanazawa, H. Sato,
M. Uchiyama, Angew. Chem. Int. Ed. 2020, 59, 8464 – 8470;
Angew. Chem. 2020, 132, 8542 – 8548.
[17] D. H. Kwan, F. Schulz, Molecules 2011, 16, 6092 – 6115.
[18] a) J.-P. Guo, C.-Y. Zhu, C.-P. Zhang, Y.-S. Chu, Y.-L. Wang, J.-X.
Zhang, D.-K. Wu, K.-Q. Zhang, X.-M. Niu, J. Am. Chem. Soc.
2012, 134, 20306 – 20309; b) J.-M. Zhang, H.-H. Wang, X. Liu, C.-
H. Hu, Y. Zou, J. Am. Chem. Soc. 2020, 142, 1957 – 1965.
[19] S. B. Krasnoff, U. Englich, P. G. Miller, M. L. Shuler, R. P. Glahn,
B. G. G. Donzelli, D. M. Gibson, J. Nat. Prod. 2012, 75, 175 – 180.
[20] Y.-M. Shi, C. Richter, V. L. Challinor, P. Grꢁn, A. G. del Rio, M.
Kaiser, A. Schꢁffler, M. Piepenbring, H. Schwalbe, H. B. Bode,
Org. Lett. 2018, 20, 1563 – 1567.
[21] K. D. Clevenger, J. W. Bok, R. Ye, G. P. Miley, M. H. Verdan, T.
Velk, C. Chen, K. Yang, M. T. Robey, P. Gao, M. Lamprecht,
P. M. Thomas, M. N. Islam, J. M. Palmer, C. C. Wu, N. P. Keller,
N. L. Kelleher, Nat. Chem. Biol. 2017, 13, 895 – 901.
[22] Y. Morishita, H. Zhang, T. Taniguchi, K. Mori, T. Asai, Org. Lett.
2019, 21, 4788 – 4792.
[23] D.-W. Gao, C. S. Jamieson, G. Wang, Y. Yan, J. Zhou, K. N.
Houk, Y. Tang, J. Am. Chem. Soc. 2021, 143, 80 – 84.
[24] Y. Morishita, Y. Aoki, M. Ito, D. Hagiwara, K. Torimaru, D.
Morita, T. Kuroda, H. Fukano, Y. Hoshino, M. Suzuki, T.
Taniguchi, K. Mori, T. Asai, Org. Lett. 2020, 22, 5876 – 5879.
[25] A. Minami, T. Ugai, T. Ozaki, H. Oikawa, Sci. Rep. 2020, 10,
13556.
[46] M. Ohtawa, S. Ogihara, K. Sugiyama, K. Shiomi, Y. Harigaya, T.
nagamitsu, S. Omura, J. Antibiot. 2009, 62, 289 – 294.
[47] Z. Wasil, K. A. K. Pahirulzaman, C. Butts, T. J. Simpson, C. M.
Lazarus, R. J. Cox, Chem. Sci. 2013, 4, 3845 – 3856.
[26] J. Kohno, Y. Asai, M. nishio, M. Sakurai, K. Kawano, H.
hiramatsu, N. Kameda, N. Kishi, T. Okuda, S. Komatsubara, J.
Antibiot. 1999, 52, 1114 – 1123.
Manuscript received: August 8, 2021
[27] R. L. Edwards, D. J. Maitland, A. J. S. Whalley, J. Chem. Soc.
Perkin Trans. 1 1991, 1411 – 1417.
Accepted manuscript online: August 26, 2021
Version of record online: && &&
,
&&&&
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Research Articles
Polyketides
J. Takino, A. Kotani, T. Ozaki, W. Peng,
J. Yu, Y. Guo, S. Mochizuki, K. Akimitsu,
M. Hashimoto, T. Ye,* A. Minami,*
H. Oikawa*
&&&& — &&&&
Biochemistry-Guided Prediction of the
Absolute Configuration of Fungal
Reduced Polyketides
We performed a focused analysis of
functionally related polyketide synthases
(PKSs) which produce linear polyketides
(PKs) such as prophialotide and other
polyhydroxy PKs. The relative and abso-
lute configurations were determined by
degradation studies. Based on the
results, we propose a stereochemical rule
during the chain-elongation process of
the reduced PKs.
Angew. Chem. Int. Ed. 2021, 60, 2 – 11
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!