Identification of crucial bottlenecks in engineered polyketide biosynthesis

Article abstract of DOI:10.1039/c9ob00831d

The concept of combinatorial biosynthesis promises access to compound libraries based on privileged natural scaffolds. Ever since the elucidation of the biosynthetic pathway towards the antibiotic erythromycin A in 1990, the predictable manipulation of type I polyketide synthase megaenzymes was investigated. However, this goal was rarely reached beyond simplified model systems. In this study, we identify the intermediates in the biosynthesis of the polyether monensin and numerous mutated variants using a targeted metabolomics approach. We investigate the biosynthetic flow of intermediates and use the experimental setup to reveal the presence of selectivity filters in polyketide synthases. These obstruct the processing of non-native intermediates in the enzymatic assembly line. Thereby we question the concept of a truly modular organization of polyketide synthases and highlight obstacles in substrate channeling along the cascade. In the search for the molecular origin of a selectivity filter, we investigate the role of different thioesterases in the monensin gene cluster and the connection between ketosynthase sequence motifs and incoming substrate structures. Furthermore, we demonstrate that the selectivity filters do not apply to new-to-nature side-chains in nascent polyketides, showing that the acceptance of these is not generally limited by downstream modules.

Full text of DOI:10.1039/c9ob00831d

View Article Online
View Journal
Organic &
Biomolecular
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: M. Grote, S.
Kushnir, N. Pryk, D. Möller, J. Erver, A. Ismail-Ali and F. Schulz, Org. Biomol. Chem., 2019, DOI:
10.1039/C9OB00831D.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Volume 14 Number
1 7 January 2016 Pages 1–372
Organic &
Biomolecular
Chemistry
www.rsc.org/obc
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1477-0520
COMMUNICATION
Takeharu Haino et al.
Solvent-induced emission of organogels based on tris(phenylisoxazolyl)
benzene
rsc.li/obc

Page 1 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
Identification of Crucial Bottlenecks in Engineered Polyketide
Biosynthesis
Marius Grote, Susanna Kushnir, Niclas Pryk, David Möller, Julian Erver, Ahmed Ismail-Ali, Frank
Schulz*
Affiliation
Organische Chemie 1
AG Naturstoffchemie und -biochemie
Fakultät für Chemie und Biochemie
Ruhr-Universität Bochum
Universitätsstraße 150
44780 Bochum, Germany
ABSTRACT
The concept of combinatorial biosynthesis promises access to compound libraries based on privileged
natural scaffolds. Ever since the elucidation of the biosynthetic pathway towards the antibiotic
erythromycin A in 1990, the predictable manipulation of type I polyketide synthase megaenzymes was
investigated. However, this goal was rarely reached beyond simplified model systems. In this study,
we identify the intermediates in the biosynthesis of the polyether monensin and numerous mutated
variants using a targeted metabolomics approach. We investigate the biosynthetic flow of
intermediates and use the experimental setup to reveal the presence of selectivity filters in polyketide
synthases. These obstruct the processing of non-native intermediates in the enzymatic assembly line.
Thereby we question the concept of a truly modular organization of polyketide synthases and highlight
obstacles in substrate channeling along the cascade. In the search for the molecular origin of a
selectivity filter, we investigate the role of different thioesterases in the monensin gene cluster and
the connection between ketosynthase sequence motifs and incoming substrate structures.
Furthermore, we demonstrate that the selectivity filters do not apply to new-to-nature side-chains in
nascent polyketides, showing that the acceptance of these is not generally limited by downstream
modules.
INTRODUCTION
Biosynthesis of polyketides on type I polyketide synthases (PKS) follows the colinearity rule, in which
the structure of the biosynthetic product is determined by the type and location of respective catalytic
domains.1 The monensin PKS is a member of this enzyme family and comprises a loading module and
twelve extender modules on eight proteins with 2,239 to 4,133 amino acids and a total molecular
weight of about 4 MDa.2, 3 In a PKS, each module catalyzes typically one elongation step on the nascent
polyketide chain as well as an individually different number of reductive steps (Fig. 1B). The minimal
1

Organic & Biomolecular Chemistry
Page 2 of 20
View Article Online
set of catalytic domains in a canonical module towards a reduced polyketide such as monensin is
composed of a ketosynthase (KS), an acyltransferase (AT) and an acyl carrier proteDinOI:(A10C.1P03).9/CT9hOeB00831D
oxidation state is determined by the optional presence of a ketoreductase (KR), a dehydratase (DH)
and an enoylreductase (ER) domain which reduce the primary β-ketothioester product in a stepwise
fashion, analogous to fatty acid biosynthesis.4
This apparent modular organization of PKS has been utilized for the engineered biosynthesis of
polyketide derivatives. Combinatorial biosynthesis potentially enables the generation of highly diverse
and otherwise hardly accessible new-to-nature compounds, yet it often suffers from low fermentation
yields and poor predictability of the experimental outcome.5, 6
The monensin PKS orchestrates 71 catalytic steps on the way from a malonyl-coenzyme A starter unit
towards the fully assembled backbone of the natural product, the shunt product premonensin
described previously.7 However, this biosynthetic pathway is remarkably specific and after extensive
post-PKS processing, delivers monensin A and B as the dominant products. At the same time, the
monensin PKS reveals a significant catalytic promiscuity, most prominently in the AT domain in
module 5 (monAT5), which was shown to accept a range of different extender units if these are
supplied to the fermentation medium.8-10 On the other hand, the targeted inactivation of reducing
domains in the monensin PKS often resulted in a drop of fermentation yields towards the modified
products.11
The moderate to high fidelity on PKS assembly lines is ascribed to either the substrate specificity of
individual catalytic domains or a precise domain-domain interaction pattern which ensures the
channeling of the nascent polyketide chain in the correct order of catalytic events.5, 12-14 The exclusive
use of targeted point mutations in the engineering of the monensin PKS, however, minimizes the
structural disturbance of the large PKS multienzyme complexes,15 therefore, diminished fermentation
yields in such experiments could potentially be attributed to an intrinsic substrate specificity of
individual PKS segments.
We launched a comparative investigation of differently mutated variants of Streptomyces
cinnamonensis to investigate potential obstacles in the channeling of the nascent polyketide chain in
an engineered biosynthesis. Alterations targeted both the redox pattern and the extender unit
flexibility of the monensin PKS, using the shunt product premonensin as a model system.
Consequently, two different experimental approaches seemed applicable: Synthetic chain terminators
to capture intermediates16-18 or the direct detection of spontaneously or enzymatically released
intermediates as hop-off products.19, 20
RESULTS
Firstly, monensin biosynthesis in the wild-type strain S. cinnamonensis ATCC 15413 was analyzed for
spontaneous hop-off products from the assembly line via LC-HRMS. The biosynthetic pathway,
including a complex post-PKS processing towards monensin, is described to proceed via fully ACP-
attached intermediates. Product release is assumed to run via a non-canonical thioesterase2, 3, 21 and
no nascent polyketide intermediates towards monensin were detectable. However, when the mutated
strain S. cinnamonsis A495 (S. cinnamonsis ATCC 15413 ΔmonCIΔmonBI/II7) was fermented in an
analogous manner, the release of the previously described new-to-nature product premonensin by
2

Page 3 of 20
Organic & Biomolecular Chemistry
View Article Online
this variant was not as efficient as for monensin in the case of the wild type. Instead, the intermediate
DOI: 10.1039/C9OB00831D
product for each module from module 4 upwards was detectable in the fermentation extract (Fig. 1).
Figure 1: A. Extracted ion chromatograms ([M+Na]+ ± 0.005 m/z) indicating the hop-off products released from
the assembly line as free acids. The hop-off products from different modules (M) were detected via LC-HRMS in
extracts from fermentations of S. cinnamonensis A495. Reductive domains labeled with an asterisk are
evolutionarily inactivated in the native assembly line. The fermentation products were absorbed in situ onto
XAD16 resin in the fermentation broth, which was subsequently eluted using ethyl acetate. In module 5, a
promiscuous acyl transferase incorporates two different extender units into the nascent polyketide chain,
leading to premonensin A and B (PreA/PreB). Likewise, intermediates from module 5 upwards arise as A (red)-
and B (blue)-derivatives (premonensin B: R = methyl; premonensin A: R = ethyl), exemplarily shown for the M5
intermediate (C). B. Modular assembly line of the monensin gene cluster consisting of a loading module (LM) and
12 elongation modules encoded on 8 separate proteins (MonA1-MonA8). Instead of passing on, spontaneous
hydrolysis of the nascent polyketide results in hop-off products M4 to M11. Hop-off products M6A/B were
previously isolated and fully characterized by NMR spectroscopy.9 Early intermediates labelled with “–” do not
3

Organic & Biomolecular Chemistry
Page 4 of 20
View Article Online
bind reliably to XAD16 resins and, thus, are not included in the analysis. KS: Ketosynthase. DH: Dehydratase. ER:
DOI: 10.1039/C9OB00831D
Enoylreduktase. KR: Ketoreductase. AT: Acyl transferase. ACP: Acyl carrier protein. D. The β-ketone hop-off
product M9 was only detected in its decarboxylated form (M9A/B). Intermediates released from module 11 were
detected in the form of the corresponding lactones, similar to that in the release of the final products PreA and
PreB. Linear intermediates were detected in variable, but low, abundancies for both. E. HRMS data from the main
adduct [M+Na]+ of each intermediate. Accurate masses for further adducts, mSigma values22-24 and MS/MS-
fragmentation patterns are described in the SI (chapter 5).
The inactivation of post-PKS enzymes in monensin biosynthesis not only leads to the anticipated
product alteration but also appears to introduce a strong congestion in the pathway, leading to a
release of intermediates through the spontaneous competing hydrolysis en route to downstream
modules. This supports the model of the dedicated thioesterase (TE) monCII for product release in
monensin biosynthesis21 and suggests that the shunt product premonensin is formed by spontaneous
lactonization and, hence, cleavage from the ACP in the final module.
The data presented here unequivocally confirm the biosynthetic pathway for premonensin and
support the current model for ACP-aided post-PKS processing (vide supra). This is in accord with the
comparably low yields in premonensin fermentation as the release of this shunt product from a trans-
acting ACP proceeds spontaneously and is not TE-mediated. The most abundant intermediates arise
from hydrolysis after modules 4, 8 and 9 (Fig. 1A), henceforth named M4, M8 and M9. All shunt
products described here are absent in the wild-type strain, indicating that the release of the final
product is rate-limiting in the modified biosynthetic pathway towards premonensin.
Intermediates in Redox Derivative Biosynthesis
We next turned our attention to the biosynthesis of the recently reported premonensin
redox-derivative ER20.11 It is biosynthesized via a mutated monensin PKS, carrying two adjacent point
mutations in the second extension module through which its ER domain loses most of its activity. As
the main products of the fermentation are the redox derivatives, the loss-of-activity mutation causes
the reduction of the intermediate C=C bond to be significantly slower than the transfer of the nascent
polyketide derivative to the downstream module 3. The typical fermentation yield of premonensin
ER20 is at about 10 % of the value obtained for the unmodified version, rendering it the most
productive variant generated in a library of premonensin redox-derivatives.11
When the fermentation products of S. cinnamonensis A495-ER20 were analyzed using the same
methodology, an analogous yet significantly different pattern was found in the abundance of
intermediates (Fig. 2).
4

Page 5 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
Figure 2: A. Extracted ion chromatograms ([M+Na]+ ± 0.005 m/z) indicating the hop-off products released from
different modules (M) in a fermentation of S. cinnamonensis A495-ER20. The redox derivatives of premonensin
are labelled as ER20PreA/B (A: R = ethyl, B: R = methyl). While the relative abundance of ER20M4 in comparison to
the final product rises, hop-off products from modules 6 and 7 could not be detected, and only trace amounts
were found for M10 and M11. B. HRMS-data of the hop-off products (ER20M) detected and the final redox
derivatives ER20PreA/B. Further characterization is found in the SI, chapter 5. C. Clipping from the assembly line
indicating the ER-inactivating point mutation in module 2. This results in the incomplete reduction of the
polyketide before it is passed on to the downstream module. D. Comparison of the MS signal intensity of
individual intermediates in S. cinnamonensis A495 and A495-ER20, normalized to the signal intensity of the
respective biosynthetic end products. Error bars represent the mean standard deviation over four independent
clones.
The abundance of the early intermediate ER20M4 was significantly increased over the later products.
Intermediates ER20M6 and ER20M7 were released below the detection limit, in contrast to the
fermentation of unmodified premonensin, while the relative abundance of ER20M5A/B was slightly
higher than its reduced counterpart in the control fermentation of unmodified premonensin.
Intermediates beginning from module 8 could again be detected, albeit only in traces of ER20M10A/B
and ER20M11A/B. This suggests a significant rate-limitation in early modules, leading to a lowered
occupation of ACP in downstream modules.
5

Organic & Biomolecular Chemistry
Page 6 of 20
View Article Online
DOI: 10.1039/C9OB00831D
Ketoreductase Null Mutants
This prompted us to investigate the biosynthesis of those premonensin derivatives in variants of
S. cinnamonensis A495 that we previously identified as non- or low-productive.11 We began with KR
variants from modules 4 to 6, of which KR40 and KR60 yielded the corresponding premonensin
derivatives (premonensin KR40A/B and KR60A/B) in non-isolatable amounts, whereas variant KR50 did not
yield any detectable quantity of premonensin KR50A/B. The intermediates in the respective mutated
module were identified in all three examples (KR40M4, KR50M5A/B and KR60M6 A/B). No hop-off
products from downstream modules were detected identifying the transfer of the derivatized
intermediate to the respective downstream module or its elongation in that module as rate-limiting
steps (Fig. 3).
Figure 3: A. Extracted ion chromatograms ([M+Na]+ ± 0.005 m/z) showing the hop-off products in fermentations
of different KR0 variants of S. cinnamonensis A495. As expected, the β-ketoacids were completely detected as
decarboxylated compounds. The native intermediates were detectable until ahead of the mutated module.
KR40Pre: premonensin KR40A/B. KR60Pre: premonensin KR60 B. HRMS data of the decarboxylated hop-off
A/B
products (KR40M4, KR50M5A/B, KR60M6A/B) and the final shunt products premonensin KR40A/B and premonensin
KR60A/B. Further HRMS-characterization of all compounds can be found in the SI chapter 5. C. Structures of the
redox derivatives KR40Pre and KR60Pre. D. Structures of the decarboxylated intermediates released from the
assembly line in the three KR0 mutants.
6

Page 7 of 20
Organic & Biomolecular Chemistry
0
Intermediate KR60M6A/B accumulated strongly in S. cinnamonensis A495 KR6 , in contrast to
fermentations of both S. cinnamonensis A495 and the ER20 variant, emphasizing the rDaOteI:-1l0im.10i3t9a/tCi9oOnB00831D
induced by the redox-derivatization. Hydrolysis of the respective intermediate largely outcompetes its
transfer to module 7. This appears to limit the productivity of the overall assembly line towards the
new-to-nature fermentation product in each of the KR0 variants investigated.
View Article Online
This led us to the introduction of analogous KR0 mutations in the final modules 11 and 12, which had
not been reported previously. The KR110M11A/B intermediates in variant KR110 show some
accumulation ahead of their transfer to module 12 in compliance with the observations of targeted
earlier modules; the biosynthesis afterwards proceeds to the predictable final products (SI chapters 4
and 5).
No specific selectivity filter can be deduced for the mutation in module 12 as the final elongation
module and, according to the expectations, the engineered biosynthesis delivers the products
anticipated in an intensity which is comparable to the premonensin fermentation in the control strain
(SI chapters 4 and 5). This indicates that the introduced point-mutation in KR0 domains does not
necessarily compromise the assembly line’s structure, but instead, a substrate specific filter appears
to exist.
Dehydratase Null Mutants
We next analyzed the cause of the fermentation in DH0 variants of the premonensin assembly line. DH0
variants yielded the predicted new-to-nature fermentation products in the majority of cases tested in
a previous report,11 in sharp contrast to KR0 variants. The productivity, however, was usually too low
to allow for the seamless isolation of the premonensin derivatives from the fermentation broths.
Fermentations of S. cinnamonensis A495 DH20, DH40,DH50, DH70 and DH80 were analyzed as described
above. In these cases, the limitation induced by the redox-derivatization is not as severe as for
KR0 mutants. In each case, the corresponding premonensin derivative can be detected, however, at
significantly diminished intensities. In the case of targeting the DH domain in module 5, both
DH50M5A/B intermediates accumulate strongly similar to analogous KR0 variants. After the completion
of the PKS assembly line, DH50PreB can be detected in a very low intensity, while DH50PreA falls below
the detection limit (for chromatograms and HRMS analysis see SI chapters 4 and 5). This indicates a
difference in the processing of the two DH50M5 intermediates in downstream modules. Each DH0
variant was also found to be “leaky” and produced varying amounts of premonensin A and B as a by-
products. This demonstrates that the mutation used for DH domains cannot fully inactivate these,
which is in accordance with previous findings.25
Enoylreductase Null Mutants
We then turned our attention to ER0 variants in modules 6 and 8. Both S. cinnamonensis A495 ER60
and ER80 deliver the corresponding premonensin derivatives.11 The respective intermediates are found
throughout the assembly line, with the exception of ER60M10A/B (for chromatograms and HRMS
analysis see SI chapters 4 and 5). The results indicate little limitation from an in-built selectivity filter
on ER0 variant productivity, which is in accordance with our previous observation of a high productivity
in these variants.
7

Organic & Biomolecular Chemistry
Page 8 of 20
View Article Online
DOI: 10.1039/C9OB00831D
The General Effect of Null Mutations on Polyketide Synthase Redox Domains
The combined results from mutations in KR, DH and ER domains indicate that a redox-derivatization
introduces new rate-limitations in the engineered biosynthesis of premonensin derivatives, which
correlates with the low yields in the isolation of the corresponding derivative from fermentation broths
(an overview of all detected hop-off intermediates is found in Figure S2). ER0 redox variants are
structurally closest to the wild-type product and are typically tolerated throughout the whole assembly
line, albeit they converted with a reduced catalytic processivity. This tendency is still visible for DH0
variants, however, significant rate-limitations can be identified in several cases. The exact localization
of the limiting effect cannot be deduced by the detection of intermediates. However, the productivity
loss of the full assembly line is most pronounced in KR0 variants, which show a strong accumulation of
intermediates in the targeted module.
The Effect of Combined Null Mutations on the Assembly Line
We next investigated the combination of redox variations in engineered premonensin biosynthesis.
Mutation ER20 was individually combined with mutations KR40, DH50, KR80 and DH80. This miniature
library combined the productive ER20 mutation with one productive and one unproductive DH0 and
KR0 variant, respectively. The assembly line regarding variants ER20KR80 and ER20DH50 ended in the
targeted modules 8 and 5, respectively (for chromatograms and HRMS see SI chapter 4 and 5). By
contrast, variants ER20KR40 and ER20DH80 gave rise to the corresponding doubly-modified
premonensin derivatives (Fig. 4A, E, F), of which premonensin ER20DH80 was chosen for structural
characterization. Analysis of the fermentation extract revealed an intrinsic instability of the products
by dehydratization, driven by the strong conjugation of the resulting C=C bonds in the respective
products (see Fig. 4).
8

Page 9 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
Figure 4: A. Extracted ion chromatograms ([M+Na]+ ± 0.005 m/z) showing the hop-off products released from
the assembly line in a fermentation of S. cinnamonensis A495 ER20DH80. The redox derivatives of premonensin
carrying two alterations are labelled as ER20DH80PreA/B. Driven by conjugation, spontaneous dehydratization was
found via HRMS analysis. B. HRMS data of the specific products in ER20DH80PreA/B. Intermediates ER20M7A/B are
found in this double-mutant while it was not detected in fermentations of the ER20 variant, presumably due to
an additional rate-limitation introduced by the second mutation. C. Structural comparison of the dehydratized
1
double mutant product ER20DH80PreA/B, ER20PreA/B and native PreA/B. Numbers indicate H-NMR signals as
highlighted in D. Excerpt of the 1H-NMR spectra of premonensin A/B and its ER20A/B and dehydratized ER20DH80
A/B
derivatives in the region between 3.5 and 7.5 ppm. The signals are numbered according to the atom numbers in
C. A full NMR-based characterization is found in SI chapter 6. E. Structure of the second derivative produced by
a double mutant, however, in very low abundancies F. HRMS data on the final products released from the
assembly line in a fermentation of S. cinnamonensis A495 ER20KR40. Additional MS-characterization is shown in
the SI chapter 5.
A 1.8-L fermentation of S. cinnamonensis A495 ER20DH80 in SM-16 medium yielded 1.0 mg and 0.7 mg
of the B and A forms, respectively, after flash chromatography and preparative HPLC. The product was
identified by LC-HRMS and 1D- and 2D-NMR characterization as the dehydratization product of
premonensin ER20DH80 (Fig. 4C). The key structural motifs were evident from the NMR signals of the
newly introduced vinylic protons and confirmed by 2D-NMR couplings (Fig. 4D). The proton signal at
C11 is shifted downfield strongly, reflecting the conjugation that led to the dehydratization of the
primary products ER20DH80PreA/B. Assignment was carried out via the long-range dd-multiplicity of the
signal (see SI chapter 6 for full NMR characterization).
9

Organic & Biomolecular Chemistry
Page 10 of 20
View Article Online
The analysis showed that the combination of mutations in the premonensin assembly is feasible,
provided the single mutations give rise to the respective product. In the case of DHD5O0,I: t10h.1e03s9i/nCg9lOeB00831D
mutation delivered the final product in very low abundance and in combination with ER20, no
fermentation product was detectable. In analogy to the first generation of redox derivatives, hop-off
products were found in all doubly-mutated variants, indicating similar rate-limitations; however, the
individual population of modules can vary. Furthermore, extensive modification of reduced
polyketides can lead to an intrinsic chemical instability of the new-to-nature products, which limits the
scope of engineered biosynthesis.
Potential Limitations in Precursor-directed Extender Unit Modification
In addition to mutagenesis-induced redox derivatizations, structural alterations in premonensin and
monensin were described to be amendable by precursor-directed biosynthesis.8-10, 26, 27 Consequently,
synthetic extender unit analogues were accepted by monAT5 and incorporated into the polyketide
backbone. We investigated the incorporation of synthetic malonic acid-based extender units into
premonensin to reveal possible limitations of precursor-directed derivatization.
S. cinnamonensis A495 was fermented in the presence of allyl(All)-, propargyl(Prg)-, propyl(Pr)-,
butyl(Bu)-, and chloropropyl(Cl)-substituted malonic acid diethyl ester, as described previously. As
expected, all previously reported corresponding premonensin derivatives (PreAll, PrePrg, PrePr and PreBu)
were detected by LC-HRMS, furthermore, very small quantities of the previously undescribed
chloropropylpremonensin (PreCl, Figure S15, Table S50, S51).
The experiments on the premonensin redox derivatives suggested a selectivity filter in modules
directly or further downstream of a mutated PKS domain. This led us to the expectation that passing
the M5-intermediate carrying a non-native side chain downstream or its elongation in module 6 might
be a bottleneck, resulting in an accumulation of released intermediate derivatives in the feeding
experiments. Contradictory to these expectations, no over-proportional intensities of those hop-off
products were detected (for chromatograms see SI chapter 4). Analysis of most feeding experiments,
besides the ones leading to PreCl and PreBu, showed an intermediate pattern comparable to the control
fermentation of S. cinnamonensis A495 with an apparent slight tendency towards the accumulation of
the respective M8-11 intermediates and a concomitantly lower abundancy in modules 5-7. In
fermentations resulting in trace amounts of PreCl, intermediate M10Cl was the only detectable one and
PreBu formation was accompanied by the detectable release of the late-stage intermediates M8Bu,
M10Bu and M11Bu (Fig. 5).
10

Page 11 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
Figure 5: A. Extracted ion chromatogram ([M+Na]+ ± 0.005) indicating the final products and hop-off
intermediates in an extract from a fermentation of S. cinnamonensis A495 supplemented with 10 mM of
diethyl butylmalonate. Comparably small amounts of butylpremonensin (PreBu) were formed via incorporation
of the new-to-nature extender unit in module 5. Hop-off products are shown in the corresponding color of the
final product (premonensin A: blue; premonensin B: red; butylpremonensin: green). B. Ten-times extension of
the chromatogram showing hop-off products released from different modules. C. HRMS data of PreBu and the
detectable hop-off products released from modules 8, 10 and 11.
The biosynthetic flow towards premonensin is split into three competing products, depending on
which artificial building block is incorporated in module 5 by the promiscuous AT domain. In contrast
to mutagenesis-induced redox derivatizations, extender unit variations in the monensin assembly line
are not limited by selectivity filters in the adjacent module 6. Instead, the machinery in module 6
appears to rely on the built-in selectivity of the AT5 domain and the correct protein-protein interaction
in between the two modules. The incorporation of new-to-nature extender units does not lead to an
incomplete reduction of the growing chain in module 5, as no irregular intermediates were released
above the detection limit. However, a lower abundance of non-natural intermediates generally
correlated with an increase of the final derivatized product abundance, pointing at processivity
limitations at various positions in the assembly line. These results suggest that the major limitation in
the precursor-directed variation of extender units is the competition with endogenous analogues,
which requires further engineering of the respective AT domain to achieve an improved shift in
specificity.10, 26, 28-31
Specificity Sources in PKS Assembly Lines: Trans-acting Hydrolases as Potential Proofreading
Enzymes
The results raised the question: Which factor would induce the apparent engineering limitation of the
monensin PKS? Different specificity sources in PKS are being currently discussed in various different
systems. The intrinsic specificity of KS domains towards an incoming nascent polyketide chain is well
described for trans-AT PKS systems.32-34 Additionally, trans-acting hydrolases in the same sub-family of
11

Organic & Biomolecular Chemistry
Page 12 of 20
View Article Online
PKS were identified which are capable of removing aberrant intermediates from the ACP in the
DOI: 10.1039/C9OB00831D
respective module.35-40
As discussed in previous studies, downstream domains in PKS assembly lines appear to possess an
intrinsic substrate selectivity not only in trans-AT but also in cis-AT systems.41-43 However, this
selectivity does not seem to be equally strong in cis-AT PKS, as it tolerates a derivatization of the
nascent polyketide chain in a variety of different cases with moderate yields and cis-AT KS sequences
do not strictly correlate with substrate structures.42, 44, 45 Zhang et al. recently described fingerprint
motifs in KS synthase domains,43 which we find to apply moderately to olefin-accepting KS domains
and showed no correlation with any other type of substrate structure in the monensin and related
polyether PKS KS domains (for phylogenetic tree and multiple sequence alignment see SI chapter 7).
We particularly could not identify a fingerprint motif for β-ketoacyl-accepting KS domains, which would
otherwise have explained the poor processivity in KR0 variants (Figure S50). Furthermore, ACP domains
may contribute to selectivity.35-40
The observation that a rate-limitation does not necessarily only lie in the neighboring downstream
module suggests that in addition to the possible effects of substrate-specific KS domains, one or more
presumably trans-acting proofreading enzymes might exist. In S. cinnamonensis, these would
recognize aberrant intermediates and remove them from the assembly line.
TE I domains are responsible for the acyl chain release of the nascent polyketides in modular PKS.46-48
However, many modular PKS gene clusters contain supplementary TE genes (referred to as type II TEs
or TE II) in close proximity to the PKS genes.20, 49, 50 Deletion of these genes was shown to decrease the
biosynthesis of tylosin51 and rifamycin,52 increase the production of precolibactins,20 while the yields
of pikromycin53 and surfactin54 were unaffected. In vitro studies on type II TEs showed that some are
capable of hydrolyzing aberrant acyl-SNAC49 and acyl-ACP thioesters,55 assigning them an editorial role.
A closer inspection of the monensin biosynthetic gene cluster in S. cinnamonensis showed three
different hydrolases with an incompletely studied function in the biosynthetic pathway (Fig. 6A). It has
been stated previously that these might act as type II TEs (MonAIX and MonAX)21, 49 or as lipase
(Orf31).2 We decided to explore the activity of these putative enzymes (Fig. 6A/B) in vitro on substrates,
which would resemble polyketide biosynthesis intermediates on the triketide stage (Fig. 1C).
12

Page 13 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
Figure 6: A. Localization of putative hydrolases in the monensin PKS gene cluster.2 B. Activity profiling of the
hydrolases using simple model substrates by means of a DTNB assay (SI chapter 8). Error bars result from the
standard deviation of three independent measurement; the different substrates are shown in different shades
of grey. C. Activity of the three hydrolases in the hydrolysis of the SNAC esters of M2 and ER20M2 to interrogate
a possible discrimination of redox variations on the assembly line. D. Synthesis of the wild-type triketide
intermediate in monensin biosynthesis as a model substrate, corresponding to the M2-intermediate in the
biosynthetic assembly line. a) t-BuOK, THF, 0 °C to rt, (57%);56 b) NaIO4, H2O, 0 °C to rt, (75%);57 c) LiAlH4, THF, 0
°C to rt, then 65 °C, (93%); d) amano lipase from Pseudomonas fluorescens, vinyl acetate, rt (41%, 98% ee, Figure
S55, S56);58 e) 1. DMP, CH2Cl2, 0 °C to rt (88%) 2. MeMgBr, Et2O, 0°C (51%); f) oxalyl chloride, DMSO, Et3N, CH2Cl2,
-64 to -30 °C, then NaClO2, Na2HPO4, 2-methyl-2-butene, t-BuOH:H2O (1:1) (28%);59 g) DPPA, SNAC, Et3N, DMF,
0°C to rt (24%).60 tBuOK = potassium tert-butoxide, THF = tetrahydrofuran, DMSO = dimethyl sulfoxide,
Et3N = triethylamine,
DMF = N,N-dimethylformamide,
MeMgBr = methyl
magnesium
bromide,
Et2O = ethoxyethane,
DPPA = diphenyl phosphoryl azide,
SNAC = N-acetyl
cysteamine,
DMP = Dess-Martin periodinane. E. Synthesis of the SNAC-activated ER20-M2 intermediate as a model substrate.
a) LDA, MeI, THF, -65 °C to rt (77%);61 b) Imidazole, TBS-Cl, DMF, rt (94%);62 c) DIBAL-H, toluene:CH2Cl2 (2:1), -63
to -20 °C (98%); d) oxalyl chloride, DMSO, Et3N, CH2Cl2, -60 °C to rt (81%);
e) ethyl 2-(triphenylphosphoranylidene)propionate, toluene, 70 °C (97%);60 f) K2CO3, MeOH:H2O (3:1), reflux
(95%); g) DPPA, SNAC, Et3N, DMF, 0 °C to rt (89%);60 h) PPTS, MeOH, 50 °C (93%); i) DMP, CH2Cl2, 0 °C (78%).
LDA = lithium diisopropylamide, TBS-Cl = tert-butyldimethylsilyl chloride, DIBAL-H = diisobutylaluminium
hydride, PPTS: pyridinium p-toluenesulfonate, MeI: iodomethane. Further details on synthetic procedures are
found in the SI, chapter 9.
13

Organic & Biomolecular Chemistry
Page 14 of 20
View Article Online
DOI: 10.1039/C9OB00831D
Firstly, the hydrolytic activity of MonAIX, MonAX and Orf31, purified from E. coli expression cultures,
was monitored on five SNAC-activated thioesters as mimics for the ACP-bound intermediates (Fig. 5B).
All three enzymes showed activity on several substrates and hydrolyzed test substrates with different
redox patterns and irregular chain length (compounds 1-3). MonAIX also converted the branched
substrates 4 and 5. In previous reports, MonAIX and MonAX were unsuccessfully tested for activity on
longer, fatty acid-like substrates.21
The enzymes show a broad activity profile while also revealing some limited substrate specificity. We
next decided to interrogate the catalytic properties in vitro on more realistic substrates that would
resemble intermediates in module 2 of the monensin PKS. Consequently, we chemically synthesized
the wild-type intermediate (6) and the intermediate 7 from a PKS variant with an inactivated ER2
domain (Fig. 6 D, E), each in the form of the SNAC thioester.
A synthesis of the wild-type thioester was carried out. Stereoselectivity was introduced by a lipase-
mediated desymmetrization58 of the meso-intermediate 12. The analogous ER20-triketide was
obtained through a 9-step synthesis, starting from ethyl (S)-3-hydroxybutyrate 16 (Fig. 6D/E). The
stereospecificity required was controlled via a diastereoselective Frater-Seebach alkylation61 and an E-
selective Wittig reaction.60
Hydrolysis experiments show a difference in the catalytic efficiency of the two compounds between
the three enzymes. MonAIX and MonAX accept both substrates, however, slightly prefer the redox
derivative ER20. Orf31 does not accept the wild-type intermediate as a substrate while hydrolyzing the
ER20 derivative with significant activity. These in vitro results suggest a potential proofreading activity
of these enzymes. However, this hypothesis called for in vivo verification and, hence, we proceeded
with the deletion of all hydrolases in the genome of S. cinnamonensis A495.
In Vivo Role of Hydrolases in Premonensin Biosynthesis
Firstly, the three hydrolases were separately in-frame deleted in the chromosome of S. cinnamonensis
A495. Fermentation of the resulting variants A495 ΔmonAIX and A495 Δorf31 led to insignificant
alterations in premonensin productivity in accordance with previous reports of monAIX in the
monensin producer.21 However, while the deletion of monAX has been previously reported to lead to
a strongly varying drop in monensin productivity, the analogous mutation in the premonensin
producer S. cinnamonensis A495 ΔmonAX fully abolished productivity over all four clones tested. In
contrast to the properties of the previously reported type II thioesterase clpQ in the colibactin gene
cluster,20 none of the deletions resulted in a decrease in the abundance of intermediates.
We decided to individually delete the hydrolases MonAIX and Orf31 in all variants having a single null
mutation of a reductive domain of modules 2 and 6 to identify both module and redox state specificity
of a potential proofreading activity. Both modules possess a complete reductive loop with all three
types of reductive domains. Furthermore, each of these domains was successfully inactivated in a
previous study with strong processivity variations in the resulting PKS.11
MonAIX and orf31 genes were successfully deleted in all variants targeted, except in the case of the
orf31 deletion in the KR60 variant. Fermentation of the resulting 11 S. cinnamonensis variants revealed
no significant alterations in premonensin derivative productivity over at least four independent clones.
14

Page 15 of 20
Organic & Biomolecular Chemistry
View Article Online
The deletion of the two hydrolases were combined in S. cinnamonensis A495 ΔmonAIXΔorf31 and its
corresponding redox variants in modules 2 and 6. This experiment would reveaDl OaI: 10p.1o0t3e9n/Ct9iaOlB00831D
complementation of a single deletion by the remaining hydrolase.
Comparative fermentations have not shown an increase in productivity for any premonensin redox
derivative after the deletion of one or both of the non-essential hydrolases Orf31 or MonAIX.
Furthermore, the relative abundance of intermediates was not affected in any of the clones.
The effect of the deletion of the hydrolases on the incorporation of artificial extender units was
investigated in parallel experiments in variants of S. cinnamonensis A495 carrying single and combined
deletions of monAIX and orf31 using the non-natural malonic acid derivatives listed above. The
abundance of the corresponding premonensin derivatives, each normalized to the sum of
premonensin A and B, increased in the case of the butyl-substituted extender unit (Figure 7).
Alterations were insignificant in the other cases. PreCl formation was found in one out of three clones
in the control, whereas all three deletion clones gave rise to the derivative.
Figure 7: Effect of hydrolase double-deletions on the relative abundance of the side-chain derivatives of
premonensin, normalized to the combined abundance of premonensin A and B in the same fermentation extract.
The control strain S. cinnamonensis A495 and variant S. cinnamonensis A495 ΔmonAIXΔorf31 were tested in
three independent clones each. Deletion of the hydrolases had a significant positive effect on the productivity
level of butyl premonensin (p = 0.05). On the other hand, all deletion clones produced the corresponding
chloropropyl derivative, while only one wild-type clone delivered measurable quantities of this compound. Both
single deletions showed no significant effect.
The relative abundance of derivatized intermediates in each case increased proportionally to the final
derivatized product. However, a proofreading activity of a hydrolase would possibly reveal itself in a
change in the intermediate pattern upon its deletion. As this is not discernible, the molecular reason
for the possibly increased productivity remains elusive and requires further experiments.
CONCLUSION
We here report on the identification of key intermediates in monensin biosynthesis and variously
engineered biosynthetic assembly lines.
The experiments highlight a significant substrate specificity in downstream domains as nascent
polyketide intermediates with new-to-nature β-keto groups typically accumulate and are not further
15

Organic & Biomolecular Chemistry
Page 16 of 20
View Article Online
processed after the mutated module. Other non-native redox states appear to be converted with
DOI: 10.1039/C9OB00831D
lowered processivity instead, leading to a gradual decrease in product formation rather than a
complete shut-down of the assembly line.
When the catalytic promiscuity of an AT domain is exploited for the introduction of non-natural
extender units, the module directly downstream of this event does not select against this
incorporation. Instead, nascent intermediates with aberrant side chains tend to accumulate in distant
modules, implying so far nebulous obstacles in substrate channeling.
The detailed mechanism underlying the rate and, thereby, yield limitation in engineered polyketide
biosynthesis remains to be identified. The contribution of downstream KS domains and potentially
trans-acting hydrolases might turn out to be causative for the decreased yield upon the engineering of
polyketide biosynthetic pathways. KS sequence alignments have not yet revealed discernible
fingerprint sequences that would identify substrate preferences on a broader basis nor have cis-AT
type I PKS gene clusters readily revealed proofreading enzymes.
The results presented in this manuscript suggest that concomitant targeted metabolomics
experiments to identify bottlenecks in the artificial system are advisable in each case of engineered
polyketide biosynthesis to reveal specific limitations. Counter-strategies will have to be developed,
currently on a case-by-case basis, until more enzymological knowledge on PKS is gathered. In contrast
to oversimplified earlier views on polyketide assembly lines, cis-AT type I PKS possess a substrate
specificity that turns out to limit the engineered biosynthesis of reduced polyketides on a routine basis.
MATERIALS AND METHODS
Chromosomal deletions: Constructs based on pKC1139 were employed for the deletion of the
hydrolases.63 codA (E. coli) was introduced into pKC1139 and exconjugants cultivated in the presence
of 5-fluorocytosin to enable rapid selection for the second crossover.64 After selection, suitable
colonies were identified by colony PCR and sequencing. Oligonucleotide sequences are shown in the
SI chapter 8.
Fermentation: Precultures in TSB medium (2 ml, 300 µg/ml phosphomycin, 2 d, 30 °C, 180 rpm) in PP
tubes supplied with glass beads were inoculated from colonies on GYM agar. An amount of 15 ml SM16
medium (20 g/L XAD16, 250 ml Erlenmeyer flasks, glass beads, 30 °C, 180 rpm) was inoculated with
5 % preculture and cultivated for 5 d.
Sample preparation: Cell paste and XAD16 resin were collected by centrifugation and frozen at -20 °C.
The solids from one flask were extracted with 6 ml ethylacetate (12 h, 19 °C, 150 rpm). The organic
layer was subsequently recovered and the solvent removed in vacuo. The residue obtained was
dissolved in 3 ml acetonitrile (HPLC grade) and stored at -80 °C. Samples were centrifuged (12,000 x g,
4 °C, 30 min) prior to LC-HRMS analysis.
LC-HRMS: An amount of 2 µl or 5 μL (redox derivatives) of each sample was analyzed by LC-HRMS/MS2.
LC-HRMS analysis was run on an Ultimate 3000 HPLC System (consisting of a pump, autosampler,
column oven and UV detector) coupled to a compact mass spectrometer (BRUKER DALTONIK GmbH,
Life Sciences, Bremen, Germany) using the standard electrospray ionization source. All solvents were
LC-MS grade (Chromasolv). A Nucleodur C18 Isis column (Macherey&Nagel, 150/2, 1.8 µm), was used
for chromatographic separation. Further details are described in the SI chapter 2.
16

Page 17 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
Author contributions
F. S. and S. K. conceived the idea and designed the experiments. S. K., M. G., N. P. performed
experiments with the help of J. E., D. M., and A. I.-A., M. G. and F. S. conducted the data analysis with
the help from S. K., and F. S. and M. G. wrote the manuscript with input from all authors. F. S.
supervised the research.
Conflicts of interest
There are no conflicts to declare.
Acknowledgments
Christian Merten (RUB) is thanked for VCD-analysis of compound 13. Part of this research was funded
by the DFG Research Training Group GRK 2341 “Microbial Substrate Conversion” (MiCon). M. G. and
A. I.-A. are fellows of the International Max Planck Research School of Chemical and Molecular Biology
and M. G. thanks the Studienstiftung des deutschen Volkes. N. P. is a fellow of the Manchot
Foundation. D. M. and F. S. acknowledge the German Research Foundation for support within the
Cluster of Excellence RESOLV EXC1069. This work is supported by the „Center for Solvation Science
ZEMOS“ funded by the German Federal Ministry of Education and Research BMBF and by the Ministry
of Culture and Research of Nord Rhine-Westphalia.
17

Organic & Biomolecular Chemistry
Page 18 of 20
View Article Online
DOI: 10.1039/C9OB00831D
References
1.
2.
J. Staunton and K. J. Weissman, Natural Product Reports, 2001, 18, 380-416.
M. Oliynyk, C. B. Stark, A. Bhatt, M. A. Jones, Z. A. Hughes-Thomas, C. Wilkinson, Z. Oliynyk,
Y. Demydchuk, J. Staunton and P. F. Leadlay, Mol Microbiol, 2003, 49, 1179-1190.
Z. A. Hughes-Thomas, C. B. W. Stark, I. U. Bohm, J. Staunton and P. F. Leadlay, Angew Chem
Int Edit, 2003, 42, 4475-4478.
3.
4.
5.
6.
7.
C. Hertweck, Angewandte Chemie International Edition, 2009, 48, 4688-4716.
B. Lowry, X. Li, T. Robbins, D. E. Cane and C. Khosla, ACS Central Science, 2016, 2, 14-20.
K. J. Weissman and P. F. Leadlay, Nature Reviews Microbiology, 2005, 3, 925.
A. Bhatt, C. B. W. Stark, B. M. Harvey, A. R. Gallimore, Y. A. Demydchuk, J. B. Spencer, J.
Staunton and P. F. Leadlay, Angewandte Chemie International Edition, 2005, 44, 7075-7078.
D. Möller, S. Kushnir, M. Grote, A. Ismail-Ali, K. R. M. Koopmans, F. Calo, S. Heinrich, B. Diehl
and F. Schulz, Lett. Appl. Microbiol., 2018, 67, 226-234.
8.
9.
A. Ismail-Ali, E. K. Fansa, N. Pryk, S. Yahiaoui, S. Kushnir, M. Pflieger, A. Wittinghofer and F.
Schulz, Org. Biomol. Chem., 2016, 14, 7671-7675.
10.
11.
K. Bravo-Rodriguez, A. F. Ismail-Ali, S. Klopries, S. Kushnir, S. Ismail, E. K. Fansa, A.
Wittinghofer, F. Schulz and E. Sanchez-Garcia, Chembiochem, 2014, 15, 1991-1997.
S. Kushnir, U. Sundermann, S. Yahiaoui, A. Brockmeyer, P. Janning and F. Schulz, Angewandte
Chemie International Edition, 2012, 51, 10664-10669.
12.
13.
14.
15.
16.
C. Bayly and V. Yadav, Molecules, 2017, 22, 235.
E. Kalkreuter and G. J. Williams, Current Opinion in Microbiology, 2018, 45, 140-148.
M. Klaus and M. Grininger, Natural Product Reports, 2018, DOI: 10.1039/C8NP00030A.
K. J. Weissman, Natural Product Reports, 2016, 33, 203-230.
M. Tosin, L. Betancor, E. Stephens, W. M. Ariel Li, J. B. Spencer and P. F. Leadlay,
ChemBioChem, 2010, 11, 539-546.
17.
18.
19.
20.
21.
J. Havemann, M. E. Yurkovich, R. Jenkins, S. Harringer, W. Tao, S. Wen, Y. Sun, P. F. Leadlay
and M. Tosin, Chemical Communications, 2017, 53, 1912-1915.
M. Tosin, Y. Demydchuk, J. S. Parascandolo, C. Blasco Per, F. J. Leeper and P. F. Leadlay,
Chemical Communications, 2011, 47, 3460-3462.
J. Moldenhauer, X.-H. Chen, R. Borriss and J. Piel, Angewandte Chemie International Edition,
2007, 46, 8195-8197.
Z.-R. Li, J. Li, J.-P. Gu, J. Y. H. Lai, B. M. Duggan, W.-P. Zhang, Z.-L. Li, Y.-X. Li, R.-B. Tong, Y. Xu,
D.-H. Lin, B. S. Moore and P.-Y. Qian, Nature Chemical Biology, 2016, 12, 773.
B. M. Harvey, H. Hong, M. A. Jones, Z. A. Hughes-Thomas, R. M. Goss, M. L. Heathcote, V. M.
Bolanos-Garcia, W. Kroutil, J. Staunton, P. F. Leadlay and J. B. Spencer, ChemBioChem, 2006,
7, 1435-1442.
22.
S. Ojanperä, A. Pelander, M. Pelzing, I. Krebs, E. Vuori and I. Ojanperä, Rapid Communications
in Mass Spectrometry, 2006, 20, 1161-1167.
23.
24.
25.
A. L. Rockwood, S. L. Van Orden and R. D. Smith, Analytical Chemistry, 1995, 67, 2699-2704.
A. Tenhosaari, Organic Mass Spectrometry, 1988, 23, 236-239.
J. Wu, T. J. Zaleski, C. Valenzano, C. Khosla and D. E. Cane, Journal of the American Chemical
Society, 2005, 127, 17393-17404.
26.
K. Bravo-Rodriguez, S. Klopries, Kyra R. M. Koopmans, U. Sundermann, S. Yahiaoui, J. Arens,
S. Kushnir, F. Schulz and E. Sanchez-Garcia, Chemistry & Biology, 2015, 22, 1425-1430.
M. Grote and F. Schulz, ChemBioChem, 2019, in press.
I. Koryakina, C. Kasey, J. B. McArthur, A. N. Lowell, J. A. Chemler, S. Li, D. A. Hansen, D. H.
Sherman and G. J. Williams, ACS Chemical Biology, 2017, 12, 114-123.
E. M. Musiol-Kroll, F. Zubeil, T. Schafhauser, T. Härtner, A. Kulik, J. McArthur, I. Koryakina, W.
Wohlleben, S. Grond, G. J. Williams, S. Y. Lee and T. Weber, ACS Synthetic Biology, 2017, 6,
421-427.
27.
28.
29.
30.
C. C. Ladner and G. J. Williams, Journal of Industrial Microbiology & Biotechnology, 2016, 43,
371-387.
18

Page 19 of 20
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C9OB00831D
31.
32.
U. Sundermann, K. Bravo-Rodriguez, S. Klopries, S. Kushnir, H. Gomez, E. Sanchez-Garcia and
F. Schulz, ACS Chemical Biology, 2013, 8, 443-450.
M. Jenner, S. Frank, A. Kampa, C. Kohlhaas, P. Pöplau, G. S. Briggs, J. Piel and N. J. Oldham,
Angewandte Chemie International Edition, 2013, 52, 1143-1147.
J. Piel, Natural Product Reports, 2010, 27, 996-1047.
M. Jenner, J. P. Afonso, H. R. Bailey, S. Frank, A. Kampa, J. Piel and N. J. Oldham, Angewandte
Chemie, 2015, 127, 1837-1841.
33.
34.
35.
36.
K. J. Weissman, Natural Product Reports, 2015, 32, 436-453.
C. Anselmi, M. Grininger, P. Gipson and J. D. Faraldo-Gómez, Journal of the American
Chemical Society, 2010, 132, 12357-12364.
37.
38.
S. Dutta, J. R. Whicher, D. A. Hansen, W. A. Hale, J. A. Chemler, G. R. Congdon, A. R. H.
Narayan, K. Hakansson, D. H. Sherman, J. L. Smith and G. Skiniotis, Nature, 2014, 510, 512-
517.
J. R. Whicher, S. Dutta, D. A. Hansen, W. A. Hale, J. A. Chemler, A. M. Dosey, A. R. H. Narayan,
K. Hakansson, D. H. Sherman, J. L. Smith and G. Skiniotis, Nature, 2014, 510, 560-564.
K. J. Weissman, Nat Chem Biol, 2014, 11, 660-670.
C. Nguyen, R. W. Haushalter, D. J. Lee, P. R. L. Markwick, J. Bruegger, G. Caldara-Festin, K.
Finzel, D. R. Jackson, F. Ishikawa, B. O/'Dowd, J. A. McCammon, S. J. Opella, S.-C. Tsai and M.
D. Burkart, Nature, 2014, 505, 427-431.
39.
40.
41.
42.
43.
T. Nguyen, K. Ishida, H. Jenke-Kodama, E. Dittmann, C. Gurgui, T. Hochmuth, S. Taudien, M.
Platzer, C. Hertweck and J. Piel, Nature Biotechnology, 2008, 26, 225.
A. C. Murphy, H. Hong, S. Vance, R. W. Broadhurst and P. F. Leadlay, Chemical
Communications, 2016, 52, 8373-8376.
L. Zhang, T. Hashimoto, B. Qin, J. Hashimoto, I. Kozone, T. Kawahara, M. Okada, T. Awakawa,
T. Ito, Y. Asakawa, M. Ueki, S. Takahashi, H. Osada, T. Wakimoto, H. Ikeda, K. Shin-ya and I.
Abe, Angewandte Chemie International Edition, 2017, 56, 1740-1745.
H. Jenke-Kodama, T. Börner and E. Dittmann, PLOS Computational Biology, 2006, 2, e132.
A. Ginolhac, C. Jarrin, P. Robe, G. Perrière, T. M. Vogel, P. Simonet and R. Nalin, Journal of
Molecular Evolution, 2005, 60, 716-725.
44.
45.
46.
47.
R. Aggarwal, P. Caffrey, P. F. Leadlay, C. J. Smith and J. Staunton, J Chem Soc Chem Comm,
1995, DOI: DOI 10.1039/c39950001519, 1519-1520.
K. J. Weissman, C. J. Smith, U. Hanefeld, R. Aggarwal, M. Bycroft, J. Staunton and P. F.
Leadlay, Angewandte Chemie International Edition, 1998, 37, 1437-1440.
R. Pieper, G. Luo, D. E. Cane and C. Khosla, Nature, 1995, 378, 263-266.
M. L. Heathcote, J. Staunton and P. F. Leadlay, Chemistry & Biology, 2001, 8, 207-220.
L. Du and L. Lou, Natural Product Reports, 2010, 27, 255-278.
A. R. Butler, N. Bate and E. Cundliffe, Chemistry & Biology, 1999, 6, 287-292.
Y. Doi-Katayama, Y. J. Yoon, C. Y. Choi, T. W. Yu, H. G. Floss and C. R. Hutchinson, Journal of
Antibiotics, 2000, 53, 484-495.
48.
49.
50.
51.
52.
53.
54.
S. Chen, J. B. Roberts, Y. Xue, D. H. Sherman and K. A. Reynolds, Gene, 2001, 263, 255-264.
P. Cosmina, F. Rodriguez, F. Ferra, G. Grandi, M. Perego, G. Venema and D. Sinderen,
Molecular microbiology, 1993, 8, 821-831.
55.
56.
57.
58.
59.
B. S. Kim, T. A. Cropp, B. J. Beck, D. H. Sherman and K. A. Reynolds, Journal of Biological
Chemistry, 2002, 277, 48028-48034.
T. Ishikawa, R. Kadoya, M. Arai, H. Takahashi, Y. Kaisi, T. Mizuta, K. Yoshikai and S. Saito, The
Journal of Organic Chemistry, 2001, 66, 8000-8009.
M. L. Wolfrom and J. M. Bobbitt, Journal of the American Chemical Society, 1956, 78, 2489-
2493.
J. S. Yadav, D. N. Chary, N. N. Yadav, S. Sengupta and B. V. S. Reddy, Helv Chim Acta, 2013, 96,
1968-1977.
X. Guo, T. G. Liu, C. R. Valenzano, Z. X. Deng and D. E. Cane, Journal of the American Chemical
Society, 2010, 132, 14694-14696.
19

Organic & Biomolecular Chemistry
Page 20 of 20
View Article Online
DOI: 10.1039/C9OB00831D
60.
61.
62.
63.
64.
K. Kinoshita, P. G. Williard, C. Khosla and D. E. Cane, Journal of the American Chemical
Society, 2001, 123, 2495-2502.
D. Gallenkamp and A. Fürstner, Journal of the American Chemical Society, 2011, 133, 9232-
9235.
J. C. Barrish, H. L. Lee, T. Mitt, G. Pizzolato, E. G. Baggiolini and M. R. Uskokovic, The Journal
of Organic Chemistry, 1988, 53, 4282-4295.
M. Bierman, R. Logan, K. O'Brien, E. T. Seno, R. Nagaraja Rao and B. E. Schoner, Gene, 1992,
116, 43-49.
M.-P. Dubeau, M. G. Ghinet, P.-É. Jacques, N. Clermont, C. Beaulieu and R. Brzezinski, Applied
and Environmental Microbiology, 2009, 75, 1211-1214.
20