Toblerols: Cyclopropanol-Containing Polyketide Modulators of Antibiosis in Methylobacteria

Article abstract of DOI:10.1002/anie.201709056

Trans-AT polyketide synthases (PKSs) are a family of biosynthetically versatile modular type I PKSs that generate bioactive polyketides of impressive structural diversity. In this study, we detected, in the genome of several bacteria a cryptic, architectu

Full text of DOI:10.1002/anie.201709056

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Toblerols, Cyclopropanol-Containing Modulators of
Methylobacterial Antibiosis Generated by an Unusual Polyketide
Synthase
Authors: Reiko Ueoka, Miriam Bortfeld-Miller, Brandon Isamu
Morinaka, Julia Anne Vorholt, and Jörn Piel
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201709056
Angew. Chem. 10.1002/ange.201709056
Link to VoR: http://dx.doi.org/10.1002/anie.201709056
http://dx.doi.org/10.1002/ange.201709056

10.1002/anie.201709056
Angewandte Chemie International Edition
COMMUNICATION
Toblerols, Cyclopropanol-Containing Modulators of
Methylobacterial Antibiosis Generated by an Unusual Polyketide
Synthase
Reiko Ueoka[a], Miriam Bortfeld-Miller[a], Brandon I. Morinaka[a,b], Julia A. Vorholt[a], and Jörn Piel*[a]
Abstract: Trans-AT polyketide synthases (PKSs) are a family of
biosynthetically versatile modular type I PKSs that generate bioactive
polyketides of impressive structural diversity. In this study, we
detected in the genome of several bacteria a cryptic, architecturally
unusual trans-AT PKS gene cluster that eluded automated PKS
prediction. Genomic mining of one of these strains, the model
methylotroph Methylobacterium extorquens AM1, revealed unique
epoxide- and cyclopropanol-containing polyketides named toblerols.
Relative and absolute stereochemistry were determined by NMR
experiments, chemical derivatization, and the comparison of CD data
between derivatized natural product and a synthesized model
compound. Biosynthetic data suggest that the cyclopropanol moiety
is generated by two one-carbon shortening events. Surprisingly, a
knock-out strain impaired in polyketide production showed strong
inhibitory activity against other methylobacteria in contrast to the wild-
type producer. The activity was inhibited by complementation with
toblerols, suggesting that these compounds modulate an as-yet
unknown methylobacterial antibiotic.
suggesting that a reliable prediction is possible for most trans-AT
PKS gene clusters.[4]
In this work, we aimed to study the function of noncanonical
trans-AT PKSs that elude prediction. The rationale was that
structural novelty is likely to be highest when comparison-based
approaches fail. When analyzing the 259 trans-AT PKS clusters
(as defined by the phylogenetic assignment of KS domains, status
2014) deposited at GenBank, one stood out due to its highly
aberrant architecture. The PKS system is encoded by an
unusually small cluster of ca. 20 kb (named tob cluster, Fig. 1A),
exhibited a highly fragmented, largely non-modular architecture,
contained diverse uncommon enzyme components, and was not
detected as a PKS in the genome by the automated prediction
tool AntiSMASH.[5]
Multimodular polyketide synthases (PKSs) are giant bacterial
enzymes that generate a large diversity of complex polyketide
natural products. Many of these play important roles in mediating
bacterial interactions and as basis of drugs.[1] Most multimodular
PKSs fall into two main subfamilies termed cis- and trans-
acyltransferase (AT) PKSs, of which the latter stands out due to
an impressive diversity of biosynthetic components.[2] Although
many of these features remain poorly understood, it has
nevertheless been possible to establish a set of rules that permit
the detailed prediction of polyketide structures from sequenced
trans-AT PKS gene clusters.[3] This method, previously developed
in our lab, is based on ketosynthase (KS) domains that are
located in each PKS module to successively elongate the growing
polyketide chain by one building block. Due to a correlation of KS
domain sequences with structural moieties of their polyketide
intermediate substrates, the analysis of all KSs in a trans-AT PKS
provides information on the collective polyketide core structure.
This colinearity rule has enabled the targeted discovery of new
polyketides with various bioactivities, validating the method and
Figure 1. (A) Architecture of the tob gene cluster (accession numbers
WP_083598533-WP_003598550) from M. extorquens AM1. Vertical triangles
above genes denote ORFs deleted in individual experiments. Only mutations
for filled triangles affected production. Lollipops symbolize point mutations.
Abbreviations: ACP, acyl carrier protein; AL, acyl-CoA or acyl-ACP ligase; DH,
dehydratase; ER, enoylreductase; OXY, FAD-dependent oxygenase; HY,
hydrolase; P450, cytochrome P450; PX, pyridoxal phosphate-dependent
enzyme; SDR, short-chain dehydratase reductase; TE, thioesterase. (B)
Comparative LCMS analysis (total ion chromatography) of extracts from the
wild-type producer (upper) and the ΔtobC deletion mutant (lower). Green
arrows indicate peaks only present in the wild-type extract.
[a]
Dr. Reiko Ueoka, Dr. Miriam Bortfeld-Miller, Dr. Brandon I.
Morinaka, Prof. Dr. Julia Vorholt, Prof. Dr. Jörn Piel.
Institute of Microbiology
Eigenössische Technische Hochschule (ETH) Zurich
Vladimir-Prelog-Weg 1-5/10
8093 Zurich
Switzerland
E-mail: jpiel@ethz.ch
Department of Pharmacy, National University of Singapore
18 Science Drive 4
[b]
Singapore 117543
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201709056
Angewandte Chemie International Edition
COMMUNICATION
2D NMR data supported the presence of a lactone instead of the
cyclopropane ring (Figure 2, Figure S8 and S24-28). Toblerol E
(5) co-eluted with toblerol F (6) by HPLC separation, and toblerol
G (7) co-eluted with toblerol H (8). The HR-ESIMS data revealed
that 5 (Figure S29) and 7 (Figure S30) both had an identical
molecular formula containing one chlorine, the position of which
was determined by 1D and 2D NMR spectra (Figure 2, Figure S8,
and S31-40). Toblerol F (6) had an additional double bond, but
lost an epoxide as compared to 4, which was suggested by its
Due to this noncanonical architecture, it was impossible to
predict the structure of the product by the previously developed
trans-AT PKSs colinearity rule. We identified closely related PKS
loci in genomes of eleven bacterial strains belonging to four
genera (Table S1-2), suggesting that, in spite of extensive
fragmentation, the cluster encodes a functional and biologically
relevant pathway. Interestingly, at least six of these bacteria are
methylotrophs, i.e., organisms that utilize one-carbon compounds
such as methanol and that are almost unexplored regarding their
natural product potential.[6]
1
molecular formula (Figure S41) and H NMR spectrum (Figure
S31). The structure was confirmed by 2D NMR data (Figure 2,
Figure S8, and S32-35). Toblerol H (8) had a molecular formula
of C12H22O4, which was predicted by HR-ESIMS data (Figure S42).
8 lacks two epoxides as compared to 1, but formed an ether ring,
which was elucidated by 1D and 2D NMR data (Figure 2, Figure
S8, and S36-40). The structures of all these congeners (4-8)
suggested interrelationships by opening of epoxides. Responsible
gene candidates could not be identified in the tob gene cluster,
but the diversity of products suggests that chloride attack on
epoxide moieties in 5 and 7 likely occurs spontaneously.
For functional characterization of the PKS cluster, we selected
the model methylotroph Methylobacterium extorquens AM1.[7] To
identify potential polyketides, a knock-out mutant was prepared
by markerless deletion of the PKS gene tobC, suspected to be
essential for biosynthesis. Indeed, HPLC-MS- and 1H NMR-based
comparative analysis of culture extracts revealed the absence of
several compounds in the mutant culture compared to the wild
type (Figure 1B and Figure S1), thus permitting targeted isolation.
Compound 1 had a predicted molecular formula of C12H20O4 as
determined by HR-ESIMS (Figure S2) and was isolated by MS-
guided fractionation. Its 1H NMR spectrum and HSQC data
showed the characteristic signals for two epoxides and a
cyclopropane ring (Figures S3-4). The planar structure was
elucidated by 2D NMR experiments as a highly unusual
diepoxide-cyclopropanol derivative, named toblerol A (Figure 2
and Figures S4-8). Detailed analysis of the LCMS data revealed
another main component, toblerol B (2), which contained one
To determine the absolute stereochemistry, the configuration
of C-4 was first determined by applying the modified Mosher's
method to toblerol
B
(2).[8] The  value revealed 4R
stereochemistry (Figure S43-51). The relative stereochemistry of
the 5,6- and 7,8-epoxides in 2 was determined to be trans- and
cis-epoxide respectively by the characteristic coupling constants
[3J(H5,H6) = 2.2 Hz and 3J(H7,H8) = 4.4 Hz]. The relative
stereochemistry of the ether ring in 8 was identified by NOESY
(Figure S40), which let us suspect the absolute stereochemistry
of epoxides in 2 as 6R, 7R, and 8R based on the hypothesis that
the ether ring in 8 was formed by the opening of 7,8-epoxide in 2
via SN2 reaction. To test this hypothesis, 1 was treated with Red-
Al, expected to result in a product with opened 5,6-epoxide.[9] The
product was identical to 8, which was determined by the
comparison of MS and NMR data for 8 and the reaction mixture
(Figure S42 and S52-55). This suggested the absolute
stereochemistry of epoxides in 2 as 5S, 6R, 7R and 8R. J-based
analysis of 5 determined the relative stereochemistry of 4R', 5R',
and 6S', which provided further evidence (Figure S35, S51, and
S56)).[10] The relative stereochemistry of the cyclopropane ring in
2 was determined to be trans by the small coupling constant
[3J(H10,H12) = 2.8 Hz]. NOESY correlation in the lactone ring of 4-
7 supported the relative stereochemistry of 8S' and 10S' in 4-7
(Figure S28, S35, and S40). This lactone ring is likely formed by
cleavage of the cyclopropane ring and the following attack to the
epoxide. The absolute stereochemistry of C-10 in 2 was therefore
suggested to be S.
1
carbon and one oxygen more than 1 (Figure S9). The H NMR
spectrum of 2 showed an additional singlet signal at 8.2 ppm, and
the 2D NMR data revealed a formyl group attached to the
cyclopropanol moiety (Figure 2, Figures S8, S10-14). Toblerol C
(3) was also isolated as a main hydrophobic component.
Compared to 1, 3 lacked two epoxides and a hydroxyl group on
the carbon chain, but contained one additional double bond, and
a cyclopropanoic acid instead of a cyclopropanol moiety (Figure
2, Figure S8, and S15-21).
Finally, to obtain further evidence for the cyclopropane
stereochemistry in 1, toblerol A (1) was converted to the
corresponding mono-naphthoate ester (9) (Figure S57-58) that
showed a significant Cotton effect (max 238 nm,  -3.9)
associated with chiral perturbation of the naphthoate
chromophore (max 240 nm). A suitable model compound (10) was
prepared from benzyl alkynyl ether (11) which was converted in
five steps to the optically active cyclopropanol (12, see SI and
Figure S59-S68 for details of synthesis). Coupling with 2-
naphthoic acid and deprotection furnished the model compound
Figure 2. Structures of toblerol A-H (1-8).
In addition to 1-3, we identified several minor products.
Toblerol D (4) had a predicted molecular formula of C12H20O5,
suggesting one additional oxygen as compared to 1 (Figure S22).
The 1H NMR spectrum revealed that one epoxide unit was absent
but showed an additional doublet methyl group (Figure S23). The
This article is protected by copyright. All rights reserved.

10.1002/anie.201709056
Angewandte Chemie International Edition
COMMUNICATION
10 (Figure 3 and Figure S69-74). The CD spectra for synthetic 10
matched with that of toblerol mono-naphthoate ester (9), namely
the Cotton effect (max 238 nm,  -4.2) was identical in sign and
magnitude therefore establishing the absolute stereochemistry of
the cyclopropane ring as 10S and 12R (Figure 3).
Only two studies have reported natural products with
cyclopropanol moieties, isolated from a fungus and a fern.[11] The
position of this moiety at the apparent polyketide terminus raised
intriguing questions about its biosynthetic origin. For initial
insights, we performed 13C-labeling studies by feeding wild-type
cultures with sodium [2-13C]- and [1,2-13C2]acetate, respectively.
Growth in acetate-based medium resulted in the dominant
production of toblerol D (4) instead of 1-3 obtained in methanol
medium. The 13C NMR spectrum of 4 isolated from a [2-
13C]acetate culture had 13C-enriched signals for C-1, C-3, C-5, C-
7, C-9, and C-11 (Figure S75). The coupling constants of 13C
signals in 4 isolated from a [1,2-13C2]acetate culture revealed
each acetate units in 4 as shown in Figure 4A (see also Figure
S76). These data suggested that toblerol biosynthesis starts at
the Western terminus and that the lactones (4-7) are formed from
a cyclopropanol precursor containing six acetate units. This result
was unexpected, since it implicated that the terminal carbon (C-
13) in toblerol C (3) is derived from oxidative cleavage of a further
C2 unit. For toblerol A (1) lacking C-13, the data support either a
second or alternative cleavage reaction that generates the
cyclopropanol moiety (Figure 4B). A similar cleavage event, likely
catalyzed by the FAD-dependent oxygenase PedG, has
previously been proposed for pederin biosynthesis.[12] In
agreement, the tob cluster encodes a close PedG homolog, TobD.
For further functional analysis, we generated 10 additional tob
mutants (Figure 1A). Mutations upstream of tobA did not change
the metabolic profile (Figure S77). Deletions of tobB, tobC, tobD,
and tobK resulted in complete abolishment of detectable
polyketide biosynthesis, as did point mutations of predicted
active-site residues in the HY and OXY domains of TobE. For a
mutant lacking the cytochrome P450 gene tobL, LCMS analysis
of the extracts showed accumulation of 3 and a new compound,
Figure 3. (A) Conversion of toblerol A (1) to mono-naphthoate ester (9) and
preparation of model compound (10). (B) CD spectra of 9 (green) and 10 (blue)
in CH3CN at 25°C.
A
B
C
TobG AT ER
DH
13
TobE
PX
TobD
OXY
TobF
?
3
TobC
KS
TobI
TobJ
TE
?
TobH
KR
OXY
14
AL
KS
HY
KS
TobD
TobL
?
1
2
6
PX
-PMP
OXY
PX
pyridox-
8
Cl-
5, 7
4
amine (PMP)
Figure 4. (A) Results of 13C-labeling studies on toblerol D (4). Triangles indicated carbons from acetate C1 and circles indicated carbons from acetate C2. Thick
lines show each acetate units determined by the coupling constants. (B) Compounds accumulated in the tobL strain. (C) Toblerol enzymes and hypothetical
biosynthetic pathway of toblerols.
This article is protected by copyright. All rights reserved.

10.1002/anie.201709056
Angewandte Chemie International Edition
COMMUNICATION
toblerol I (13), while all the other toblerols (1, 2 and 4-8) produced
in the wild-type were absent (Figure 4C and S77-82). Detailed 1H
NMR analysis of these extracts indicated the presence of
additional minor intermediates. NMR-guided fractionation
permitted isolation of the trace component toblerol J (14) (Figure
4B and Figure S83-89). Since none of the three congeners
contained epoxide groups, tobL is likely responsible for
epoxidation. Although these experiments did not reveal the origin
of the cyclopropane ring, we point out the similarities to the cis-
AT PKS-derived ambruticin, a myxobacterial polyketide with an
internal cyclopropane ring.[13] The PKS module proposed to play
a role in cyclopropanation carries a HY and a PX domain, which
are also present in TobE. Both PX homologs carry a Thr instead
of the active-site lysine residue conserved in other pyridoxal
phosphate-dependent enzymes. A Thr-to-Ala point mutation did
not result in a changed metabolic profile, suggesting that this
residue does not participate in catalysis. Based on these
similarities, it is likely that TobE is responsible for toblerol
cyclopropanation.
A hypothetical pathway that would be
Figure 5. (A) Inhibitory interactions of WT (right) and tobC (left) against
Methylobacterium sp. Leaf101. (B) Complementation assay with extracts (left
picture: 2 L, right picture: 5 L) from WT strain (left) and MeOH (right) to tobC
strain. (C) Complementation assay with toblerol C (left, 35 g) and MeOH (right)
to WT (above) and tobC (below) strains.
consistent with these relationships, the PX domain present on
TobE, the TobE OXY domain absent in the ambruticin PKS, and
the mutagenesis and feeding studies, is shown in Fig. 4C.
Based on a previous study showing that cyclopropanols
irreversibly inhibit methanol dehydrogenases,[14] we tested the
hypothesis that toblerols inhibit competing methylobacteria
(Methylobacterium sp. Leaf97, and Leaf101[15]). In contrary to
expectations, however, the non-producing tobC mutant showed
strong inhibitory interactions in methylobacterial overlay assays,
while the wild type producer was only weakly active (Figure 5A).
The inhibitory effect of the tobC strain was complemented in a
concentration-dependent manner by applying the extracts from
the wild-type, but not from the tobC strain (Figure 5B and S90).
Interestingly pure toblerol C (3) also abolished the inhibition
(Figure 5C). Based on these results, we propose that toblerols
negatively regulate antibiosis in M. extorquens AM1, the
mechanism of which remains to be elucidated. Since we observed
that growth of the toblerol producer itself was also slightly inhibited
by this antibiotic (Fig. 5A), regulation of antibiotic production by
toblerol might provide a self-protective mechanism.
Experimental Section
See Supporting Information.
Acknowledgements
We thank D.B. Müller for providing the sensitive methylobacterial
strains, R. Meoded for the valuable discussion, and S. Fuchs for
creative compound naming. We are grateful for funding from the
EU (ERC Advanced Projects SynPlex to J.P. and PhyMo to
J.A.V.) and the SNF (NRP 72 "Antimicrobial resistance",
407240_167051 to J.P. and J.V.). R.U. and B.I.M. were fellowship
recipients of the Alexander von Humboldt foundation.
In conclusion, an in silico-based mining strategy focusing on
non-canonical enzyme homologs revealed a chemically and
functionally unusual minimalistic trans-AT PKS in the model
methylotroph M. extorquens AM1. Characterization of ten novel
polyketides toblerols A-J, combined with gene disruption and
labeling studies, suggest that the unique cyclopropanol polyketide
terminus is in part generated by an oxidative chain shortening
event. Further studies on this pathway, its role in other bacteria,
and the antibiotic activity are currently ongoing.
Keywords: polyketide biosynthesis • genome mining •
cyclopropanes • antibiotics • methylotrophs
[1]
[2]
[3]
C. Hertweck, Angew. Chem. Int. Ed. 2009, 48, 4688-4716.
E. J. N. Helfrich, J. Piel, Nat. Prod. Rep. 2016, 33, 231-316.
T. A. Nguyen, K. Ishida, H. Jenke-Kodama, E. Dittmann, C. Gurgui, T.
Hochmuth, S. Taudien, M. Platzer, C. Hertweck, J. Piel, Nat. Biotech.
2008, 26, 225-233.
[4]
a) R. Teta, M. Gurgui, E. J. N. Helfrich, S. Künne, A. Schneider, G. van
Echten-Deckert, A. Mangoni, J. Piel, ChemBioChem 2010, 11, 2506-
2512. b) A. Nakabachi, R. Ueoka, K. Oshima, R. Teta, A. Mangoni, M.
Gurgui, N. J. Oldham, G. van Echten-Deckert, K. Okamura, K.
Yamamoto, H. Inoue, M. Ohkuma, Y. Hongoh, S. Miyagishima, M. Hattori,
J. Piel, T. Fukatsu, Curr. Biol. 2013, 23, 1478-1484. c) A. Kampa, A. N.
Gagunashvili, T. A. M. Gulder, B. I. Morinaka, C. Daolio, M. Godejohann,
V. P. W. Miao, J. Piel, Ó. S. Andrèsson, Proc. Natl. Acad. Sci. 2013, 110,
E3129-E3137. d) R. Ueoka, A. R. Uria, S. Reiter, T. Mori, P. Karbaum,
E. E. Peters, E. J. N. Helfrich, B. I. Morinaka, M. Gugger, H. Takeyama,
S. Matsunaga, J. Piel, Nat. Chem. Biol. 2015, 11, 705-712.
This article is protected by copyright. All rights reserved.

10.1002/anie.201709056
Angewandte Chemie International Edition
COMMUNICATION
[5]
T. Weber, K. Blin, S. Duddela, D. Krug, H. U. Kim, R. Bruccoleri, S. Y.
Lee, M. A. Fischbach, R. Müller, W. Wohlleben, R. Breitling, E. Takano,
M. H. Medema, Nucl. Acids. Res. 2015, 43, W237-W243.
a) C. G. N. Penalver, D. Morin, F. Cantet, O. Saurel, A. Milon, J. A.
Vorholt, FEBS Letters 2006, 580, 561-567. b) A. W. Puri, A. L. Schaefer,
Y. Fu, E. P. Greenberg, M. E. Lidstrom, J. Bacteriol. 2017, 199, e00773-
16.
[11] a) K. A. Alvi, B. G. Nair, J. Rabenstein, G. Davis, D. D. Baker, J. Antibiot.
2000, 53, 110-113. b) K. P. Kainz, J. Virtbauer, H. Kählig, V. Arion, O.
Donath, G. Reznicek, W. Huber, B. Marian, L. Krenn, Helv. Chim. Acta,
2012, 95, 1531-1537.
[6]
[12] J. Piel, Proc. Natl. Acad. Sci. 2002, 99, 14002-14007.
[13] B. Julien, Z. -Q. Tian, R. Reid, C. D. Reeves, Chem. Biol. 2006, 13, 1277-
1286.
[7]
[8]
[9]
A. M. Ochsner, F. Sonntag, M. Buchhaupt, J. Schrader, J. A. Vorholt,
Appl. Microbiol. Biotechnol. 2015, 99, 517-534.
[14] J. Frank Jr, S. H. van Krimpen, P. E. J. Verwiel, J. A. Jongejan, A. C.
mulder, J. A. Duine, Eur. J. Biochem. 1989, 184, 187-195.
[15] Y. Bai, D. B. Müller, G. Srinivas, R. Garrido-Oter, E. Potthoff, M. Rott, N.
Dombrowski, P. C. Münch, S. Spaepen, M. Remus-Emsermann, B.
Hüttel, A. C. McHardy, J. A. Vorholt, P. Schulze-Lefert, Nature 2015, 528,
364-369.
I. Ohtani, T. Kusumi, Y. Kashman, H. Kakisawa, J. Am. Chem. Soc. 1991,
113, 4092-4096.
T. Murata, M. Sano, H. Takamura, I. Kadota, D. Uemura, J. Org. Chem.
2009, 74, 4797-4803.
[10] a) N. Matsumori, D. Kaneno, M. Murata, H. Nakamura, K. Tachibana, J.
Org. Chem. 1999, 64, 866-876. b) C. Nilewski, R. W. Geisser, M. -O.
Ebert, E. M. Carreira, J. Am. Chem. Soc. 2009, 131, 15866-15876.
This article is protected by copyright. All rights reserved.

10.1002/anie.201709056
Angewandte Chemie International Edition
COMMUNICATION
Entry for the Table of Contents
COMMUNICATION
Bacterial anti-antibiotics: The
structurally and functionally unusual
toblerols were identified as products of
a cryptic polyketide synthase gene
cluster from the methylotrophic model
bacterium Methylobacterium
extorquens AM1. Biosynthetic studies
suggest that the cyclopropanol moiety
is generated by two one-carbon
shortening events.
Reiko Ueoka, Miriam Bortfeld-Miller,
Brandon I. Morinaka, Julia Vorholt, and
Jörn Piel*
tobA
G
B
C
D
E
FAD
ACP KS DH ACP KS
ACP HY PX FAD
SDR
Page No. – Page No.
L
tobF
H
I
J
K
KS (DH) AT ER
AL TE
P450
Toblerols, Cyclopropanol-Containing
Modulators of Methylobacterial
Antibiosis Generated by an Unusual
Polyketide Synthase
This article is protected by copyright. All rights reserved.