Investigations into the biosynthesis of the antifungal strobilurins

Article abstract of DOI:10.1039/c8ob00608c

The strobilurins are important antifungal metabolites isolated from a number of basidiomycetes and have been valuable leads for the development of commercially important fungicides. Isotopic labelling studies with early and advanced intermediates confirm for the first time that they are produced via a linear tetraketide, primed with the rare benzoate starter unit, itself derived from phenylalanine via cinnamate. Isolation of a novel biphenyl metabolite, pseudostrobilurin B, provides evidence for the involvement of an epoxide in the key rearrangement to form the β-methoxyacrylate moiety essential for biological activity. Formation of two bolineol related metabolites, strobilurins Y and Z, also probably involves epoxide intermediates. Time course studies indicate a likely biosynthetic pathway from strobilurin A, with the simplest non-subsubstituted benzoate ring, to strobilurin G with a complex dioxepin terpenoid-derived substituent. Precursor-directed biosynthetic studies allow production of a number of novel ring-halogenated analogues as well as a new pyridyl strobilurin. These studies also provide evidence for a non-linear biosynthetic relationship between strobilurin A and strobilurin B.

Full text of DOI:10.1039/c8ob00608c

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Investigations into the biosynthesis of
the antifungal strobilurins†
Cite this: DOI: 10.1039/c8ob00608c
a
a
a
a
Zafar Iqbal,‡§ Li-Chen Han,§ Anna M. Soares-Sello,§ Risa Nofiani,¶
b
b
c
a
Gerald Thormann, Axel Zeeck, Russell J. Cox,
Thomas J. Simpson
Christine L. Willis* and
a
*
The strobilurins are important antifungal metabolites isolated from a number of basidiomycetes and
have been valuable leads for the development of commercially important fungicides. Isotopic label-
ling studies with early and advanced intermediates confirm for the first time that they are produced
via a linear tetraketide, primed with the rare benzoate starter unit, itself derived from phenylalanine
via cinnamate. Isolation of a novel biphenyl metabolite, pseudostrobilurin B, provides evidence for
the involvement of an epoxide in the key rearrangement to form the β-methoxyacrylate moiety
essential for biological activity. Formation of two bolineol related metabolites, strobilurins Y and Z,
also probably involves epoxide intermediates. Time course studies indicate a likely biosynthetic
pathway from strobilurin A, with the simplest non-subsubstituted benzoate ring, to strobilurin G with
a complex dioxepin terpenoid-derived substituent. Precursor-directed biosynthetic studies allow
production of a number of novel ring-halogenated analogues as well as a new pyridyl strobilurin.
These studies also provide evidence for a non-linear biosynthetic relationship between strobilurin A
and strobilurin B.
Received 12th March 2018,
Accepted 22nd May 2018
DOI: 10.1039/c8ob00608c
rsc.li/obc
lisation as “mucidermin” for the treatment of skin infec-
tions. In 1977, Anke and co-workers isolated two antifungal
Introduction
3
Strobilurins are a group of bioactive metabolites produced metabolites, strobilurin A 2 and strobilurin B 3 from
1
4
by various fungi. Mucidin 1 was the first to be isolated in Strobilurus tenacellus. The former was clearly identical to
1
2
965 from the basidiomycete Oudemansiella mucida. The mucidin which was thus re-assigned the E,Z,E configuration.
triene system was initially assigned the all E,E,E configur- Many other strobilurin analogues have been isolated from
ation, and its potent antifungal activity led to its commercia- other basidiomycetes. Structural variations include oude-
mansin A 4 from Oudemansiella mucida in which the central
9
/10 double bond of the triene system has formally had
5
methanol added,
Flaviolaschia sp. which contains a methyl enol ether and
14-hydroxystrobilurin A 6 from a Petrula sp. containing a
9-methoxystrobilurin
A
5
from
a
School of Chemistry, University of Bristol, Cantocks Close, Bristol, BS8 1TS, UK.
6
E-mail: tom.simpson@bristol.ac.uk, chris.willis@bristol.ac.uk
b
Institut für Organische und Biomolekulare Chemie, Georg-August Universität,
7
Tammannstraße 2, 37077 Göttingen, Germany
hydroxymethyl. A number of strobilurins have complex
c
Institut für Organische Chemie Chemistry, Schneiderberg 1B, Leibniz Universität,
dioxepin substituents containing two highly modified
3
0167 Hannover, Germany
prenyl moieties on the phenyl ring, e.g. strobilurin G 7 from
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
8
Bolinea lutea. Bolineol 8 in which the methoxyacrylate sub-
c8ob00608c
‡
Current address: Department of Agriculture, Faculty of Nutrition Sciences, The stituent of strobilurin A has been replaced by a methyl
University of Agriculture Peshawar, Pakistan.
3-hydroxy-propionate moiety has also been isolated from
B. lutea.
The key β-methoxyacrylate toxophore targets the Qo site of
complex III of the mitochondrial electron transport chain and
§
¶
These authors contributed equally to the work.
9
Current address: Department of Chemistry, Faculty of Mathematics and
Natural Sciences, University of Tanjungpura, Jl. Prof. Dr H. Hadari Nawawi,
Pontianak, Indonesia 78124.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
1
0
prevents ATP synthesis. The major class of β-methoxyacrylate
agricultural fungicides were developed from the structures of 2
and 3 with the aim of increasing photo-stability and selectivity. Feeding studies using singly and doubly labelled [ C]-acetates,
Results and discussion
1
3
1
3
2
13
13
Thus, compounds such as azoxystrobin 9 (Syngenta) and [methyl- C]-methionine, [ H ]-, [3- C]- and [2,3- C ]-phenyl-
8
2
Kresoxim-methyl 10 (BASF) are among the most widely used alanines, supplemented to cultures of S. tenacellus and B. lutea
fungicides worldwide, used as effective treatments against a resulted in the production of isotopically enriched strobilurin
11,12
broad range of destructive fungal plant pathogens.
The strobi- A 2 (Scheme 1a). The results are consistent with the pathway
lurin fungicides are estimated to have been worth $3.4 billion per summarised in Scheme 1b, in which phenylalanine undergoes
1
8,19
annum in 2015 and make up 25% of the fungicide market and loss of ammonia to give cinnamic acid 13,
6
followed by
13
.7% of the total crop protection market. Other antifungal degradation to benzoate. Although benzoate is a relatively
natural products containing the β-methoxyacrylate toxophore uncommon starter unit in polyketide biosynthesis, benzoyl
include myxothiazole 11 and cyrmenin A 12 from the myxobac- CoA serves as a component in the formation of a number of
14,15
20
21
teria, Myxococcus fulvus and Cystobacter armeniaca respectively.
important plant metabolites including taxol and cocaine.
The biosynthesis of the strobilurins, particularly that of the The hexaketide moiety in squalestatin (zaragozic acid) biosyn-
β-methoxyacrylate toxophore, remains obscure despite the fact thesis in fungi is also formed from a benzoate starter, now
that the strobilurins and analogues are amongst the most com- known to be produced from phenylalanine via cinnamic acid
2
2,23
mercially important fungal metabolites known. Early labelling 13.
The bacterial metabolites enterocin (Streptomyces mari-
2
4
25
studies on mucidin were consistent with a polyketide pro- timus) and soraphen A (Sorangium cellulosum) also have
duced from a benzoate starter, itself derived via phenylalanine, benzoate priming their polyketide synthases.
extended by successive condensations with three malonates,
In strobilurin A 2, the labelling pattern supports a pathway
1
6
and a C-methylation from S-adenosyl methionine (SAM). The in which benzoyl CoA 14 would undergo three successive
1
3
labelling of both C-11 and C-12 from [1- C]-acetate suggested chain elongations with one C-methylation via the putative PKS-
that rearrangement of an originally linear polyketide chain bound diketide, triketide and tetraketide intermediates 15, 16
occurs. B. lutea has been described as the only ascomycete to and 17 to give prestrobilurin 18 which would then undergo
1
7
produce the strobilurin/oudemansin family of metabolites.
rearrangement of the epoxide 19 via 20 to the formyl-carboxylic
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
2
13
Scheme 1 (a) Incorporation of H and C-labelled precursors into strobilurin A 2 in S. tenacellus and B. lutea. (b) Proposed assembly and
rearrangement of strobilurin tetraketide.
13
Scheme 2 Reagents and conditions: (a) [2- C]-malonic acid (1.02 equiv.), Na
2 4
SO (0.15 equiv.), pyridine, piperidine, reflux, 4 h, 91%; (b) DCC (1.13
equiv.), DMAP (0.04 equiv.), HSNAC (1.43 equiv.), DCM, 0 °C, 2 h, 23 °C, 16 h, 58%.
acid 21/22. Finally, two O-methylations using SAM as cofactor precursor-directed biosynthesis studies (discussed in detail
1
9
complete the biosynthesis of strobilurin A 2. A number of below) with F-labelled benzoates and cinnamates gave con-
these points are discussed below. sistently higher incorporation of fluorine label into fluorostro-
Cinnamate 13 is in the unusual position of potentially bilurin analogues from cinnamate (78%) than from benzoate
acting as the source of the benzoate starter, but also as a poly- (30%) for the 4-fluoro SNAC thiolester analogues. This poten-
ketide chain elongation intermediate (15, Scheme 1b). Feeding tial contradiction was re-examined by synthesis (Scheme 2)
1
3
13
13
of [2,3- C
from C-3) being incorporated, indicating that cinnamate and its SNAC thiol ester 24. SNAC thiol esters have been com-
cannot be incorporated intact into the strobilurins. However, monly used in studies of polyketide biosynthesis to mimic the
2 2
]-phenylanine resulted in only a single C label and feeding of [2,3- C ]-cinnamate both as the free acid 23
(
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
thiol ester linkage of co-enzyme A and acyl carrier protein (ACP) methylation of ethyl acetoacetate followed by condensation
and are usually incorporated into biosynthetic pathways more with 3-fluorobenzaldehyde and cyclisation to give the
26,27
readily than the parent carboxylic acids.
In this case, 3-hydroxy-lactone 25. Following reductive removal of the
however, free cinnamic acid 23 was incorporated significantly hydroxyl group, ring opening of the resulting unsaturated
more efficiently (85%) than the SNAC thiolester 24 (40%, lactone using TBAF gave acid 26 which was esterified to the
13
Fig. S1†). In contrast, [carboxyl- C]-benzoate is incorporated corresponding thiol ester 27 in 94% yield. The Z-geometry of
more efficiently as its SNAC thiol ester compared to the free acid. the 2,3-double bond was confirmed by nOe studies.
Most fungal PKS belong to the Type I iterative class consist-
Both precursors 26 and 27 were pulse fed to cultures of
ing of a single multi-domain protein encoded by a single B. lutea after 24, 36 and 48 hours according to the protocol
28
gene. They have been classified according to the increasing used for previous feeding studies. LCMS analysis of the culture
degree of reductive modifications and C-methylations they extract, however, failed to detect any fluorinated strobilurin
carry out. One major class, the highly-reducing Type II (HR II) analogues but the peaks corresponding to 26 and 27 remained
2
9
have a trans-acting stand-alone ER domain while a minority with no obvious degradation of triketide to benzoate occurring
2
9
are Type III systems which do not catalyse β-processing reac- as had been observed to occur for the free acid and SNAC thio-
3
0
tions. The strobilurin PKS is likely to be a Type I PKS and cat- lester of the diketide (cinnamate) 15.
alyse three iterations of chain-extension and β-processing to
Tetraketide 18 (Scheme 1) is proposed to be the first
produce the three enzyme bound intermediates 15–17 shown enzyme-free product of the PKS and its fluoro-analogue was
in Scheme 1b. In order to investigate the intermediacy of these also synthesised as both the free acid 30 and SNAC thiol ester
compounds we decided to feed analogues of 16 and 17 to cul- 31 (Scheme 3B). Acid 26 was reduced to primary alcohol 28
tures of B. lutea. To aid metabolite analysis they were syn- and, following oxidation to an aldehyde was chain extended
thesised with a fluorinated phenyl ring as fluorine has been under Horner–Wadsworth–Emmons conditions to ester 29
3
1
demonstrated to be a useful tracer for biosynthetic studies.
with the 2E double bond. Hydrolysis of ester 29 gave acid 30
Accordingly, 3-fluorobenzaldehyde was converted to the 2Z,4E which was converted to thiol ester 31. The isomeric acid 33
isomer of 2-methyl-5-phenylpenta-2,4-dienoic acid 26 and its with the 2Z-double bond was prepared using a similar strategy
SNAC thiol ester 27 (Scheme 3A). Key synthetic steps included but with a Still–Gennari reaction for the chain extension.
Scheme 3 (a) NaH (60%, 1.08 equiv.), THF, 23 °C, 0.5 h, 50 °C, 20 h, 56%; (b) DIPA (2.5 equiv.), n-BuLi (2.5 equiv.), HMPA (1 equiv.), 3-fluorobenzal-
dehyde (1.1 equiv.), THF, −78 °C, 3 h; (c) 1 M KOH(aq), 23 °C, 14 h, then 6 M HCl(aq), 0 °C, 65% (over 2 steps); (d) DIPEA (1.5 equiv.), Tf
CH Cl , −78 °C, 1 h, 97%; (e) Pd(PPh (0.01 equiv.), Et SiH (2.0 equiv.), DMF, 60 °C, 2 h, 99%; (f) TBAF (1 M, 5 equiv.), THF, 23 °C, 2 h, 97%; (g)
EDCI·HCl (1.6 equiv.), DMAP (1.2 equiv.), HSNAC (1.5 equiv.), DCM, 0 °C, 2 h, 23 °C, 14 h, 94%; (h) Et N (2 equiv.), EtOCOCl (1.3 equiv.), THF, 0 °C,
.5 h; (i) NaBH (2.5 equiv.), MeOH, −78 °C, 4 h, 60% (over 2 steps); ( j) Dess–Martin periodinane (0.3 M, 1.3 equiv.), CH Cl , 23 °C, 1.5 h; (k) NaH
60%, 1.4 equiv.), (EtO) P(O)CH CO Et (1.7 equiv.), THF, 0 °C, 10 min, 23 °C, 16 h, 67% (over 2 steps); (l) (CF CH O) P(O)CH CO Et (2.1 equiv.),
2
O (1.1 equiv.),
2
2
3
)
4
3
3
0
4
2
2
(
2
2
2
3
2
2
2
2
KHMDS (15 wt%, 2 equiv.), 18-crown-6 (2.4 equiv.), THF, −78 °C, 6.5 h, 64% (2Z : 2E::5 : 1, over 2 steps); (m) 1 M NaOH(aq), THF, 23 °C, 91%; (n)
EDCI·HCl (2 equiv.), DMAP (2.4 equiv.), HSNAC (8.9 equiv.), DCM, 23 °C, 16 h, 60%; (o) 1 M LiOH(aq), MeOH, 23 °C, 16 h, 72%; (p) DIPA (2 equiv.),
n-BuLi (2 equiv.), InCl
3 2
(0.6 equiv.), THF, −78 °C, 2 h, 23 °C, 1 h, 34 (65%, over 2 steps), 35 (67%, over 2 steps); (q) NaH (60%, 1.4 equiv.), (EtO) P(O)
CH CO Et (1.7 equiv.), THF, 0 °C, 10 min, 23 °C, 16 h, 81% (over 2 steps).
2
2
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
the signal at δ 1.98 ppm for the three protons of the 14-methyl.
This confirms that tetraketide 17 is indeed the final enzyme-
bound product of the strobilurin PKS.∥
Conversion of tetraketide 18 (thus named prestrobilurin A)
to strobilurin A is proposed to proceed via epoxidation and
rearrangement to give aldehyde 21, enolisation to 22 and
methylation
Scheme 1b). To further investigate this process we synthesised
the epimeric epoxides 34 and 35 with a fluorine label
Scheme 3C). The allylic alcohol 29 was oxidized by Dess–
forming
the
β-methoxyacrylate
moiety
(
(
Martin periodinane, and then the unstable aldehyde was
treated with the methyl ester or SNAC thiolester of
α-bromoacetate under Darzens reaction conditions. However,
on feeding to B. lutea no incorporation into strobilurins, direct
Fig. 1 LCMS analysis shows that the tetraketide 30 fed to culture of or indirect could be detected. In addition, fermentation of
B. lutea yields more of the enriched metabolite than the corresponding B. lutea in the presence of varying concentrations of the known
SNAC thiol ester 31.
32
cytochrome P450 monooxygenase inhibitor ancymidol had
no effect on strobilurin production. Thus no direct evidence
for an epoxide-mediated arrangement is available.
Acid 30 and thiol ester 31 were pulse fed to B. lutea in sep-
However, on investigating minor components of a second
arate experiments. On extraction, and LCMS analysis (Fig. 1), a strain of B. lutea (strain F23523, which in our hands produced
new metabolite was detected at 23.7 min, between the peaks strobilurin B 3 as its major metabolite) we observed a minor
for strobilurin A (23.3 min) and strobilurin B (24.4 min) with a peak eluting after strobilurin B 3 with a molecular ion
+
+
37
molecular ion [M + Na] 299 Da, 18 units heavier than strobi- [M + H] 291 Da with a Cl isotopic peak at 293 Da. We puri-
lurin A 2. Interestingly, no fluorinated analogue of strobilurin fied this metabolite by preparative HPLC, and isolated 0.8 mg
B 3 was detected in either experiment, suggesting that fluorine of pure compound for NMR characterisation. Signals for a
in the phenyl ring inhibits hydroxylation. The yield of the new 3-methoxy-4-chloro-phenyl ring as in strobilurin B 3 were
metabolite was ca. three times higher from the free acid 30 evident, but those for the triene side chain were absent, being
than from the SNAC ester 31 (Fig. 1). The metabolite 3-fluoros- replaced by signals consistent with the presence of a second
1
trobilurin 36 was purified and its structure confirmed by H 1,2,4-trisubstituted aromatic ring. COSY and HMBC corre-
19
and F NMR analysis. The proton resonances for the trienoic lations (Fig. S3†) were consistent with the biphenyl structure
acid side chain are essentially identical to those of strobilurin 39 (Scheme 4). We propose that this novel compound is pro-
A 2 as expected, confirming the 7E,9Z,11E configuration, but duced by an alternative rearrangement of the key epoxide inter-
the aromatic ring shows the coupling pattern predicted for a mediate 37 (cf. 19 in Scheme 1b) where the initial carbocation
3
1
-fluoro substituent, in particular vicinal H–F couplings of intermediate from epoxide ring opening undergoes electrophi-
0.1 and 8.8 Hz respectively to H-4 and H-2, and a 4 bond H–F lic addition to the C-6/C-7 double bond, to give the cyclohexe-
1
9
coupling (5.8 Hz) to H-5. The F NMR spectrum shows a nyl cation 38, which is then aromatised by proton loss and de-
signal at −114 ppm as a double, double, doublet with the hydration to give 39 which we have named pseudostrobilurin
same coupling constants (Fig. S2b†).
B. In support of this hypothesis is the fact that 37 must
As the trienoic acid 18 has not been detected in extracts of possess the predicted 4Z,6E geometry for the cyclisation to
B. lutea, it suggests that on release of 17 from the PKS, this occur.
first enzyme-free intermediate is rapidly converted to strobilur-
Further detailed analysis of LCMS chromatograms of
ins (Scheme 1b) as discussed in more detail below. At this B. lutea extracts revealed the presence of two new minor com-
stage we could not rule out that the observed incorporation of pounds (each ca. 1 mg L− ) with UV characteristics typical of
fluorine label was occurring via degradation of 30 or 31 to strobilurin A and B. These were isolated and their structures
1
3
-fluorobenzoate and reincorporation, as observed for cinna- 41 and 42 determined from their mass spectra and full NMR
mate. We thus synthesised an isotopomer of 30 (Scheme 3B) analysis (Fig. S4†). They were named strobilurin Y 41 and stro-
with a trideuteriomethyl at C-14 in addition to the original C-3 bilurin Z 42 respectively. Their formation can be rationalised
fluorine label. Incubation of the isotopomer with B. lutea and by methanolysis of the epoxide 40 derived by a P450 or similar
LCMS analysis of the extract showed a new peak at 24.3 min mono-oxygenase-mediated oxidation of strobilurins A and B
(
(
strobilurin A 23.9 min) with the correct molecular ion (Scheme 5). In order to rule out formation of these compounds
[M + H] 280 Da) for the trideuterio-labelled analogue of
+
3
-fluorostrobilurin 36 to indicate intact incorporation of 30.
The structure was again confirmed by full NMR analysis. The
∥
The 2Z isomer 33 was also synthesised and fed as a 5 : 1 mixture with the 2E
isomer. A small incorporation into 36 was observed, but this is likely to have
for 36, as was the H NMR spectrum except for the absence of been due to the small amount of E-isomer present in the fed mixture.
1
9
F NMR spectrum was identical to that previously obtained
1
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 4 Conversion of epoxide 37 to prestrobilurin B 39.
Scheme 5 Formation of strobilurins Y 42 and Z 43 via methanolysis of epoxide 40.
from methanol used as solvent in the original LCMS purifi- extracts after partial purification by thin layer chromatography
cation, we repeated the isolation using acetonitrile in place of and the results are summarised in Fig. 2. Mono- and di-fluori-
methanol. Their continued production confirmed that they are nated analogues were incorporated and the structures con-
the result of enzymatic transformations.
firmed by NMR analysis, but individual analogues were not
In addition, a number of other strobilurins previously iso- always isolated. Similar results were obtained with both
lated from B. lutea were observed after optimisation of fermen- B. lutea F23523 and S. tenacellus, but variations were observed
tation conditions (Fig. S5†). Besides the major metabolites, in that relative efficiencies of benzoate and cinnamate incor-
strobilurins A 2 and B 3, these include strobilurins F1 43, F2 porations were sometimes reversed with benzoates being pre-
4
4, G 7, H 45 and bolineol 8. We also observed a number of ferred by S. tenacellus.
very minor compounds which were identified by NMR and However, feeding of 2-, 3- and 4-fluorobenzoate/cinnamate
mass spectrometric characteristics. These are strobilurin C 46 gave the corresponding fluorostrobilurins 48, 36 and 49 in
3
3
and strobilurin I 47 previously isolated from Xerula longipes
and an agaricus species
S. tenacellus in relative yields (compared to levels of strobilurin
3
4
respectively. Strobilurin C 46 A) of 8–30% with 2-fluorostrobilurin 48 consistently the
appears to be the earliest prenylated strobilurin, presumably poorest yield. The 3,4 and 3,5-difluoro-analogues 50 and 51
formed from strobilurin F1 43, whereas strobilurin I 47 is were formed from the corresponding benzoates, and feeding
likely to be an intermediate to strobilurin G 7. The likely 3-fluoro-4-methyl- and 3-fluoro-4-methoxybenzoates gave the
overall biosynthetic interrelationships are summarised in corresponding fluorostrobilurins 52 and 53. Feeding of nic-
Scheme 6.
otinic acid gave the novel 3-aza-strobilurin A analogue 60 and
Precursor directed biosynthesis, in which analogues of supplementation of the medium with KBr produced the
natural substrates are fed to either WT or mutant producing bromo-analogue 4-dechloro-4-bromostrobilurin B 56. Other
cultures has proved an effective method for producing ana- attempts to introduce halogens, e.g. by supplementation with
logues of microbial natural products, particularly for produ- chloride, bromide or iodide, or by adding chloro- or bromo-
3
5
cing halogenated analogues. The presence of halogens often benzoates gave only trace amounts (by LCMS), at best, of alter-
has a beneficial effect for enhancing biological activities. For natively halogenated compounds. Similar results were
example, the presence of a halogen is required for bioactivity observed in feeding a range of aromatic analogues to squales-
3
8
in the case of salinosporamide, neomangicols and rebeccamy- tatin-producing cultures of Phoma sp. In this study a similar
3
6,37
cin,
their non-halogenated analogues being inactive.
range of mono- and difluorinated benzoates were incorporated
As indicated above, fluorinated precursors can be efficiently but no other halogens were accepted.
incorporated into strobilurins. With this in mind we made a
With B. lutea F23523, feeding of 2-fluorocinnamate gave an
more systematic study aimed at producing novel fluorinated, interesting result. Production of strobilurin A 2 was completely
chlorinated and brominated analogues. The outcomes of these inhibited, the fermentation giving strobilurin B 3 along with
experiments were dependent on the producing strain used. about 50% 2-fluorostrobilurin B 48, 10% 2-fluorostrobilurin G
Our initial studies used B. lutea strain F24510. A series of 59, and about 1% 2-fluorostrobilurin H 57 (Fig. S6†). All struc-
fluorinated cinnamates and benzoates were fed both as the tures were confirmed by NMR and mass spectrometric ana-
free acid and SNAC thiol esters. The products were observed by lysis. 3-Fluorocinnamate again produced 3-fluorostrobilurin A
1
19
mass spectrometric and H and F NMR analysis of the 36, 4-fluorostrobilurin B 55 and 4-fluorostrobilurin C 58.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 6 Proposed biosynthetic inter-relationships among strobilurin related metabolites in B. lutea.
Fig. 2 Strobilurin analogues produced by precursor-directed biosynthesis.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 7 Proposed biosynthetic conversion of acid 61 to diol 62.
Finally 4-fluorocinnamate gave strobilurin A 2 along with tetraketide, prestrobilurin A 18. The potential role of cinna-
-fluorostrobilurin A 49 only with no strobilurin B, presumably mate as both the precursor to the starter unit and an assembly
fluorine at C-4 inhibiting oxidative hydroxylation. intermediate in its own right has been clarified. Prestobilurin
4
The inhibition of strobilurin A production with 2-fluorocin- A 18 undergoes oxidative rearrangement to form the core
namate reflected the result of another experiment. The simple β-methoxyacrylate moiety, but no direct evidence for the
triketide analogue 61 was chemically synthesised (Scheme S1†) process has yet been observed. However, indirect evidence for
as a potential substrate analogue for the epoxidase believed to the intermediacy of an epoxide in the key rearrangement to
be responsible for the rearrangement to give the methoxyacry- form the methoxyacrylate toxiphore has been obtained here
late toxiphore. On feeding to B. lutea it was rapidly metab- via the isolation pseudostrobilurin B 39 which is likely formed
olised (Fig. S7†) to give a new compound shown to be diol 62. via the common epoxide intermediate 37. Precursor-directed
The formation of 62 can be rationalised as summarised in biosynthesis experiments demonstrated that a wide range of
Scheme 7 by a cycle of polyketide condensation, methylation mainly fluorinated starter units can be tolerated by the PKS to
and keto-reduction to give 63. Concerted decarboxylation and provide several halogenated strobilurin analogues. Finally and
dehydration to triene 64 followed by epoxidation and hydro- unexpectedly, the production of strobilurin A 2 can be selec-
lysis gives diol 62, whose structure was established by 1D and tively inhibited with no effect on strobilurin B 3 biosynthesis.
2
D NMR analysis (Fig. S8†).
This suggests that they are produced by parallel pathways and
In this experiment, strobilurin B 3 production remains that strobilurin A 2 is not an intermediate to strobilurin B 3.
unaffected but no strobilurin A 2 is detected. Time course Further work on the genetic and molecular basis for strobi-
experiments (Fig. S9†) clearly indicate that the triketide ana- lurin biosynthesis is underway and will be reported elsewhere.
logue 61 is rapidly converted to the diol and that the diol itself
is responsible for inhibiting strobilurin A production. In both
these experiments, feeding acid 61 and 2-fluorocinnamate,
strobilurin B production is unaffected. This strongly suggests
Experimental
that the pathways to strobilurin A 2 and strobilurin B 3 are Full experimental details are supplies in the ESI.†
different and that strobilurin A 2 is not converted directly to
strobilurin B 3, so that the pathways must diverge at an early
1
3
stage. On feeding [2,3- C
2
]-cinnamate to cultures inhibited as
Conflicts of interest
above, no restoration of strobilurin A 2 production is observed,
but strobilurin B 3 is produced and labelled from the precur- There are no conflicts of interest to declare.
sor. This suggest that in these experiments, incorporation
post-benzoyl CoA into strobilurin A is being inhibited. This
suggests that benzoyl CoA may first be hydroxylated to 3-hydro-
xybenzoyl CoA before elaboration into the strobilurin B
Acknowledgements
pathway (Scheme 6).
AMS-S thanks the Commonwealth Scholarship Commission
for a postgraduate scholarship. L-CH thanks BrisSynBio, the
Centre for Synthetic Biology (BB/L01386X/1) for fellowship
support. RN thanks the Directorate General of Resources for
Science and Higher Education (Beasiswa Pendidikan
Conclusions
The strobilurins and analogues are amongst the most commer- Pascasarjana Luar Negeri Direktorat Jenderal Sumber Daya
cially important fungal metabolites known. Despite this their Ilmu Pengetahuan dan Pendidikan Tinggi) (BPP-LN Ditjen SD
biosynthesis has been relatively little studied since the initial Iptek-Dikti), Republic of Indonesia, for an Overseas
1
6
work of Nerud et al. In this study we have demonstrated that Postgraduate Scholarship and the Alumni foundation of the
phenylalanine is converted to cinnamate, and then onwards to University of Bristol for financial support. ZI thanks the
benzoate which forms the starter unit for formation of a linear Higher Education Commission of Pakistan for a post-graduate
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
scholarship. RJC thanks EPSRC (EP/F066104/1) and DFG (INST 19 L. Xian and B. S. Moore, J. Bacteriol., 2003, 185, 399–404.
87/621-1) for LCMS equipment. GT and AZ are grateful to 20 E. Baloglu and D. G. I. Kingston, J. Nat. Prod., 1999, 62,
Prof. Tim Anke for providing us with Strobilurus tenacellus. 1448–1472.
Dr Y. O’Connell is thanked for synthesis of labelled cinnamate. 21 A. J. Bjorklund and E. Leete, Phytochemistry, 1992, 31,
883–3887.
1
3
2
2 B. Bonsch, V. Belt, C. Bartel, N. Duensing, M. Koziol,
C. M. Lazarus, A. M. Bailey, T. J. Simpson and R. J. Cox,
Chem. Commun., 2016, 52, 6777–6780.
References
1
2
T. Anke, Can. J. Bot., 1995, 73, 940–945.
23 N. Liu, Y.-S. Hung, S.-S. Gao, L. Hang, Y. Zou, Y.-H. Chooi
and Y. Tang, Org. Lett., 2017, 19, 3560–3563.
24 J. Piel, C. Hertweck, P. R. Shipley, D. M. Hunt,
M. S. Newman and B. S. Moore, Chem. Biol., 2000, 7, 943–
955.
V. Musílek, J. Černá, V. Šašek, M. Semerdžieva and
M. Vondráček, Folia Microbiol., 1969, 14, 377–388.
H. Sauter, W. Steglich and T. Anke, Angew. Chem., 1999, 38,
3
4
5
6
7
8
9
1
328–1349.
T. Anke, F. Oberwinkler, W. Steglich and G. Schramm, 25 A. M. Hill, B. L. Thompson, J. P. Harris and R. Segret,
J. Antibiot., 1977, 10, 806–810.
Chem. Commun., 2003, 1358–1359.
T. Anke, H. J. Hecht, G. Schramm and W. Steglich, 26 S. Yue, J. S. Duncan, Y. Yamamoto and C. R. Hutchinson,
J. Antibiot., 1979, 32, 1112–1117.
J. Am. Chem. Soc., 1987, 109, 1253–1255.
G. Schramm, W. Steglich, T. Anke and F. Oberwinkler, 27 L. Xian and B. S. Moore, J. Biol. Chem., 2002, 277, 32505–
Chem. Ber., 1978, 111, 2779–2784.
32509.
M. Engler, T. Anke, D. Klostermeyer and W. Steglich, 28 R. J. Cox, Org. Biomol. Chem., 2007, 5, 2010–2026;
J. Antibiot., 1995, 48, 884–885.
A. Fredenhagen, P. Hug and H. H. Peter, J. Antibiot., 1990,
Y.-H. Chooi and Y. Tang, J. Org. Chem., 2012, 77, 9933–
9953.
29 T. J. Simpson, Nat. Prod. Rep., 2014, 31, 1247–1252.
4
3, 661–666.
C. T. Bedford, D. Perry and R. K. Sharma, Nat. Prod. Res., 30 M. Hashimoto, T. Nonaka and I. Fujii, Nat. Prod. Rep.,
008, 22, 1535–1539.
2104, 31, 1306–1317.
0 W. F. Becker, G. Von Jagow, T. Anke and W. Steglich, FEBS 31 D. S. J. McKeown, C. McNicholas, T. J. Simpson and
2
1
Lett., 1981, 132, 329–333.
1 J. M. Clough, Nat. Prod. Rep., 1993, 10, 565–574.
2 D. W. Bartlett, J. M. Clough, J. R. Godwin, A. A. Hall,
N. J. Willett, J. Chem. Soc., Chem. Commun., 1996, 301–302.
32 A. E. Desjardins, R. D. Plattner and M. N. Beremand, Appl.
Environ. Microbiol., 1987, 53, 1860–1865.
1
1
M. Hamer and B. Parr-Dobrzanski, Pest Manage. Sci., 2002, 33 T. Anke, H. Besl, U. Mocek and W. Steglich, J. Antibiot.,
5
8, 649–662.
1983, 143, 661–666.
1
1
3 PhillipsMcDougall Product Directory, 2015.
4 W. Trowitzsch, G. Reifenstahl, V. Wray and K. Gerth,
J. Antibiot., 1980, 23, 1480–1490.
5 F. Sasse, T. Leibold, B. Kunze, G. Höfle and
H. Reichenbach, J. Antibiot., 2003, 56, 827–831.
6 P. Nerud, P. Sedmera, Z. Zouchova, V. Musílek and
M. Vondráček, Collect. Czech. Chem. Commun., 1982, 47,
34 S. Zapf, T. Anke, H. Dasenbrock and W. Steglich,
Bioengineering, 1993, 1, 92.
35 C. Portmann, C. Prestinari, T. Myers, J. Scharte and
K. Gademann, Org. Biomol. Chem., 2009, 7, 644–646.
36 P. E. Rodrigues, L. Belin, M. Sancelme, M. Prudhomme,
M. Ollier, M. Rapp, D. Severe, J. Riou, D. Fabbro and
T. Mayer, J. Med. Chem., 1996, 39, 4471–4477.
37 M. K. Renner, P. R. Jensen and W. Fenical, J. Org. Chem.,
1998, 63, 8346–8354.
38 R. J. P. Cannell, M. J. Dawson, R. S. Hale, R. M. Hall,
D. Noble, S. Lynn and N. L. Taylor, J. Antibiot., 1993, 46,
1381–1389.
1
1
1
020–1025.
1
1
7 A. Fredenhagen, A. Kuhn, H. H. Peter, V. Cuomo and
U. Giuliano, J. Antibiot., 1990, 43, 655–660.
8 C. Hertweck and B. S. Moore, Tetrahedron, 2000, 56, 9115–
9
120.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.