Isolation and characterisation of amphotericin B analogues and truncated polyketide intermediates produced by genetic engineering of Streptomyces nodosus

Article abstract of DOI:10.1039/b922074g

Amphotericin B is a powerful but toxic drug used against fungal infections and leishmaniases. These diseases would be treated more effectively if non-toxic amphotericin derivatives could be produced on a large scale at low cost. Genetic manipulation of the amphotericin B producer, Streptomyces nodosus, has previously led to the detection and partial characterisation of 8-deoxyamphotericin B, 16-descarboxyl-16-methyl-amphotericin B, 15-deoxy-16-descarboxyl-16-methyl-15-oxo-amphotericin B, 7-oxo-amphotericin B and pentaene analogues. Here we report improved production and purification protocols that have allowed detailed chemical analyses of these compounds. The polyketide synthase product 8-deoxy-16-descarboxyl-16-methyl-amphoteronolide B was identified for the first time. In addition, the ketoreductase 10 domain of the polyketide synthase was specifically inactivated by targeted gene replacement. The resulting mutants produced truncated polyketide intermediates as linear polyenyl-pyrones.

Full text of DOI:10.1039/b922074g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Isolation and characterisation of amphotericin B analogues and truncated
polyketide intermediates produced by genetic engineering of Streptomyces
nodosus†
Barry Murphy,‡a Katie Anderson,a Charles Borissow,a Patrick Caffrey,b Gerald Griffith,a Jessica Hearn,a
Odubunmi Ibrahim,a Naseem Khan,b Natalie Lamburn,a Michael Lee,a Katherine Pugha and
Bernard Rawlings*a
Received 22nd October 2009, Accepted 18th May 2010
First published as an Advance Article on the web 23rd June 2010
DOI: 10.1039/b922074g
Amphotericin B is a powerful but toxic drug used against fungal infections and leishmaniases. These
diseases would be treated more effectively if non-toxic amphotericin derivatives could be produced on a
large scale at low cost. Genetic manipulation of the amphotericin B producer, Streptomyces nodosus,
has previously led to the detection and partial characterisation of 8-deoxyamphotericin B,
16-descarboxyl-16-methyl-amphotericin B, 15-deoxy-16-descarboxyl-16-methyl-15-oxo-amphotericin
B, 7-oxo-amphotericin B and pentaene analogues. Here we report improved production and
purification protocols that have allowed detailed chemical analyses of these compounds. The polyketide
synthase product 8-deoxy-16-descarboxyl-16-methyl-amphoteronolide B was identified for the first
time. In addition, the ketoreductase 10 domain of the polyketide synthase was specifically inactivated
by targeted gene replacement. The resulting mutants produced truncated polyketide intermediates as
linear polyenyl-pyrones.
molecules and cause cell death.6 There is also some affinity for
cholesterol, which along with low serum solubility contributes to
severe side effects such as nephrotoxicity.7
Introduction
Amphotericin B 1 (Fig. 1) is produced by the soil bacterium
Streptomyces nodosus and, despite considerable toxicity, it has
been a leading broadspectrum antifungal antibiotic for more
than fifty years.1 Resistance has not become a problem in this
long period of clinical use, unlike all other established antifungal
therapies.2 Amphotericin B has also emerged as the first line of
treatment against some Leishmania parasites,3 and has potential
in the treatment of HIV infection and prion diseases.4
Traditionally, 1 was injected as a solution in DMSO or as a
complex with deoxycholate. Recent liposomal formulations have
greatly reduced toxicity, but are sufficiently expensive to require
patient selection on pharmacoeconomic considerations, and are
major components of the drugs bills of some NHS Trusts.8
Visceral leishmaniasis (kala azar) is caused by protozoan
parasites transmitted by the sand fly, and is fatal if left untreated.
It remains a ‘most neglected’ disease with 50,000 deaths each year,
many in high-level poverty areas in Asia and South America.3
Conventional 1 is now the first line of treatment in Bihar, India
and in neighbouring Nepal. Efficacy is close to 100% but there
are nephrotoxic side effects and administration requires hospi-
talisation for a complicated infusion regime of fifteen injections
over thirty days. Despite considerable subsidy from the WHO
(US$3,000 per treatment reduced to US$200), the cost of the less
toxic liposomal formulation Ambisome severely restricts its use in
these regions, some of the poorest parts of the world. Thus there
is a need for an affordable effective analogue with greatly reduced
toxicity.3
Amphotericin B toxicity is only partly alleviated by costly lipo-
somal formulation. Improvement could be achieved by structural
alteration. Chemical studies have generated derivatives such as
MFAME 2 (Fig. 2) that has been shown to be an effective non-
toxic antibiotic but is much too expensive for clinical use.9
The activity of MFAME shows that much improved analogues
of 1 exist in chemical space, the challenge is to produce them at a
low cost. Due to the complex structure of 1, chemical modification
is unlikely to deliver an affordable non-toxic derivative. Such a
compound might instead be obtained by engineering biosynthetic
Fig. 1 Amphotericin B.
The antibiotic mode of action of 1 is complexation with
ergosterol in fungal membranes and related sterols in parasite
membranes.5 The complexes self-assemble to form transmem-
brane pores that allow uncontrolled loss of ions and small
aDepartment of Chemistry, University of Leicester, University Road, Leices-
ter, UK. E-mail: bjr2@le.ac.uk; Fax: (+44) 0116 2523789; Tel: +44 (0)116
2522093
bSchool of Biomolecular and Biomedical Science and Centre for Synthesis
and Chemical Biology, University College Dublin, Belfield, Dublin 4, Ireland.
E-mail: Patrick.caffrey@ucd.ie; Fax: +353 1 716 1183; Tel: +353 1 7161396
† Electronic Supplementary Information (ESI) available: Data for com-
pounds 1, 3, 5, 8, 9, 10, 11, 16, 20, 21 and 24. See DOI: 10.1039/b922074g/
‡ A major proportion of work and intellectual input into new isolation and
purification methods appearing in this article was from Dr Barry Murphy.
3758 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Fig. 2 MFAME.
genes in S. nodosus.10 The parent wild-type strain can produce high
levels (4 g L-1) of 1 in low cost fermentations, demonstrating the
presence of a very efficient metabolic flow of the required biosyn-
thetic building blocks, and efficient catalytic processes. Thus,
given sufficient understanding of the genetic and biosynthetic
machinery, there is potential to achieve low-cost production of
engineered compounds that can replace 1 in therapy.11
Scheme 1 Biosynthesis of amphotericin B 1.
Isotope labelling studies have shown that the aglycone core 3
is assembled from sixteen acetate and three propionate units, in
eighteen acyl chain extending cycles.12 All the genes responsible
for amphotericin B biosynthesis, export and regulation are found
in a cluster.13 The aglycone 3 is assembled by a modular polyke-
tide synthase (PKS) composed of the AmphABCIJK proteins.
AmphA loads the starter unit, AmphB adds on two propionate-
derived units, each with ketoreduction, to form the triketide
(Fig. 3). Hexamodular AmphC incorporates six acetate-derived
building blocks, each 3-oxo group reduced and dehydrated, to give
a conjugated hexaenoyl thioester. Within AmphC, module 5 has a
partially active ER domain that is responsible for formation of the
tetraene analogue, amphotericin A (28,29-dihydroamphotericin
B). The hexamodular AmphI adds on six more building blocks
(two acetate-, one propionate-, then three more acetate-derived
units). The first is reduced and dehydrated to complete the
conjugated heptaene (or tetraene) moiety. At some stage the
hydroxyl group formed in cycle 11 lactolises with the ketone
introduced incycle 13. This lactolring is acommonfeature inmany
related polyene compounds, and if formed at the tetradecaketide
stage may be important in folding and controlling the very long
growing acyl chain as it passes through the PKS proteins. AmphJ
and AmphK complete the growing acyl chain, and a thioesterase
on AmphK catalyses lactonisation to form 3.13
Scheme 2 Biosynthesis of GDP-D-mycosamine.
C-3 gives GDP-D-mycosamine, the substrate for the glycosyltrans-
ferase AmphDI.
In preliminary work, we have identified several biosyn-
thetic intermediates of 1 produced by genetic engineering of
S. nodosus. For example, inactivation of amphL and amphN
gave 8-deoxyamphotericin B (5) and 16-descarboxyl-16-methyl-
amphotericin B (8) (Fig. 4) respectively.14,15 Removal of two
PKS modules from AmphC gave material with the absorbance
spectrum of a pentaene.16 Inactivation of the ketoreductase (KR)
domain in module 16 gave the corresponding 7-oxo-analogue 9
(Fig. 5) whilst inactivation of KR12 gave traces of 15-oxo-heptaene
products.17 In this paper, we describe enhanced production,
isolation, purification and characterisation of these products,
essential for progress towards an affordable effective analogue.
The PKS product, aglycone 3, was purified for the first time. We
also investigate the inactivation of PKS KR domains, which can
deliver carbonyl-containing analogues with improved solubility
or potential for chemical or biochemical transformation.18,19 The
Fig. 3 Role of AmphABCIJK.
Aglycone (3) formation is followed by oxidation of the
methyl unit at C-16 to –COOH by the AmphN cytochrome
P450 monooxygenase to give 8-deoxy-amphoteronolide B (4).
Attachment of a mycosaminyl unit at C-19 by the AmphDI
glycosyltransferase gives 5 which is finally oxidised at C-8 by the
AmphL cytochrome P450 monooxygenase to 1 (Scheme 1), this
order of late modifications is preferred but not obligatory.14
The mycosaminyl sugar is assembled by the dehydration of
GDP-D-mannose to 4,6-dideoxy-4-oxo-D-mannose 6 by the 4,6-
dehydratase AmphDIII and isomerisation to the 4-hydroxy-3-oxo
regioisomer 7 (Scheme 2). AmphDII catalysed transamination at
Fig. 4 16-Descarboxyl-16-methyl-amphotericin B (8).
Fig. 5 7-Oxo-amphotericin B (9).
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3758–3770 | 3759
©

View Article Online
KR16 and KR12 mutants were constructed as the corresponding
7-oxo and 15-oxo analogues would be expected to have increased
serum solubility, and reduced toxicity. Prior to this study, inac-
tivation of KR10 was expected to result in biosynthesis of C-19
carbonyl containing amphoteronolides. Such aglycones might be
chemically derivatised to give libraries of semi-synthetic analogues.
Nicolaou has prepared protected amphoteronolide heptaenones
and chemically converted them into 1.20 Here we report that KR10
inactivation had unexpected biosynthetic consequences and led
to the synthesis of truncated polyketide intermediates that were
isolated as linear polyenyl-pyrones.
The amphL mutant
The final step in the biosynthesis of 1 is the AmphL cytochrome
P450 monooxygenase catalysed insertion of a hydroxyl group at
C-8 (Scheme 1). Zotchev et al. have isolated and purified the
related NysL cytochrome P450 monooxygenase that provides the
final hydroxylation step in nystatin biosynthesis.24 Inactivation
of NysL gave 10-deoxynystatin. Similar late hydroxylation also
occurs in biosynthesis of the related heptaenes, candidin and
mycoheptin.25
The amphL gene has been disrupted by insertional inacti-
vation, targeted integration of a KC515 phage vector carrying
a thiostrepton-resistance gene.16 However, even when grown in
thiostrepton, the mutant produced not only 8-deoxy-amphotericin
B (5) but also 1, indicating that the phage derived DNA could
readily undergo precise excision and that the resulting strain with
a functional AmphL enzyme was viable, despite the presence of
the antibiotic. All subsequent S. nodosus mutants were generated
by gene replacement (in which the chromosomal DNA is perma-
nently replaced with an engineered version) and cannot undergo
reversion in this way.
In the absence of thiostrepton in the media, most (>90%)
product was 1. However, when high levels (50 mg L-1) of freshly
made thiostrepton solution in DMSO (50 g L-1) were added to
all media, 5 was produced as the dominant heptaene (>80%).
Previous work on the amphL mutant identified 5.16 Improved
strains have now been isolated by strain selection of inoculum
cultures based upon yield of subsequent production cultures, and
production yields (>100 mg L-1) improved by the inclusion of
XAD16 resin in the production medium.
Results and Discussion
Genetically engineered strains were grown on a multilitre scale in
grooved Erlenmeyer flasks. Most analogues were located in the
sediment after centrifugation of production cultures. Improved
yields were obtained by including the non-ionic polymeric ab-
sorbent resin XAD16 in the production medium. Amphotericin
B is extracted from whole broth using butanol, however this has
proved problematic with the lower titres of low purity analogues.
Methanolic extraction of this sediment gives membrane lipid
and dextrin derived oligosaccharides along with heptaenes and
tetraenes. Partial removal of the methanol in vacuo from the crude
extract can result in a yellow precipitate containing heptaene (1–
80% depending on mutant), the supernatant is now largely mycelia
derived water.
Polyenes can display considerable affinity for both lipid and
oligosaccharide co-extractants, hindering purification. For most
analogues stability and solubility can dramatically alter depending
upon level of purity, due to this close affinity with these contam-
inating materials. In general, tetraenes were more water soluble
aiding separation from heptaene.
Heptaenes have a characteristic intense multi-band pattern in
the UV spectrum. The amount of heptaene present in any extract
or sample was estimated by UV absorbance at 405 nm.21 However,
any such assay requires caution as any value depends on the
‘state of oligomerisation’ of the sample. The monomeric form
contributes to the UV peak at 406 nm whilst aggregate contributes
to a broad peak at 340 nm.22
Preliminary work on KR1617 focussed upon synthetic derivati-
satisation of crude precipitates to the N-acetyl methyl ester to
assist with purification but yields were disappointing. Isolates
from the M57 (amphC) mutant were also methylated initially, but
again yields were low. This led to a focus on non-chromatographic
isolation of underivatised analogues, which was more likely in the
long term to result in cost-effective commercial production of any
promising analogue.
Isolation of 8-deoxyamphotericin B (5) from the amphL mutant
Extraction of sediment with methanol was inefficient due to
insolubility of 5. Schreier et al. have recently shown that chaotropes
(water structure breakers) such as thiocyanate can increase the
solubility of 1.22 Addition of saturated sodium thiocyanate in
methanol to sedimented mycelia and resin dramatically improved
the overall extraction procedure. The first extract was discarded.
An equal volume of water was added to the second extraction
◦
and left at 4 C overnight to give a yellow solid (1 g L-1 culture)
containing by UV assay crude heptaene (100 mg L-1). This
assay may be an underestimate due to polyene oligomerisation
in the presence of lipid. The yellow powder can be washed with
methanol–chloroform or EtOAc to partially remove lipid, or
redissolved in thiocyanate saturated methanol, with subsequent
water addition as before, to give a yellow solid containing 5
(100 mg L-1; ca 90% w/w by UV assay and NMR analysis). This
could be further purified by RP-HPLC for NMR spectroscopic
analysis that showed a very high level of similarity to that for 1.
However, whilst 7-C in 1 resonates at 29 ppm, heterocorrelation
NMR spectroscopy on 5 showed a methylene carbon at 21.37 ppm
correlating to protons at 1.73 and 1.01 ppm. This is in full
agreement with the corresponding 7-C carbon in the 8-deoxy
aglycone 3, and appears characteristic of an 8-deoxy analogue.
Commercial and medical samples of 1 are 50–90% pure
preparations that contain lipid and multiple other polyenes.23 We
found that many analogues could be readily isolated in purities
comparable to that of commercial 1, however, complete removal of
residual lipid, and in cases, other polyenes, was problematic. The
wide range of analogues shows that the biosynthetic machinery
in S. nodosus is flexible and efficient. Genetic engineering and
improvement of culture conditions should increase titre levels to
near wild-type (4 g L-1) and lead to commercial production of any
promising analogue at a low cost.
The amphNM+perDIDII mutant
The amphN gene encodes the cytochrome P450 monooxygenase
that oxidises the C-16 methyl group to a carboxyl group. This
3760 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
amphN gene was deleted from the S. nodosus chromosome
along with the neighboring amphM ferredoxin coding gene.15
Previously reported work showed that this mutant produced 16-
descarboxyl-16-methyl-amphotericin B (8), and that despite the
loss of the 16-carboxyl group, this material retained biological
activity with reduced hemolytic activity.15 Burke et al. have
chemically synthesised 8 and confirmed its promising therapeutic
potential.26 Zotchev et al. have inactivated NysN and obtained
the nystatin analogue of 8 which also shows good antifun-
gal activity and low toxicity in a mouse model of systemic
mycosis.27
In the present study, work on a derivative of the amphNM
mutant led to the discovery of aglycone 3 for the first time.
Our interest in engineering polyene glycosylation led us to
investigate biosynthesis of perimycin. This aromatic heptaene
differs from the other polyene macrolides in that it lacks an
exocyclic carboxyl group and contains perosamine instead of
mycosamine. D-Perosamine is the 4-amino-3-hydroxyl regioisomer
of D-mycosamine.28 The perDII perosamine synthase and perDI
perosamine specific glycosyltransferase genes have been isolated
from Streptomyces aminophilus. These genes were inserted into the
amphNM mutant in early attempts to engineer the biosynthesis
of 16-descarboxyl-16-methyl-19-(O)-perosaminyl amphoterono-
lides. This strain, amphNM+perDIDII, is capable of synthesising
both GDP-D-mycosamine and GDP-D-perosamine, however, no
perosaminylated analogues were identified in production cultures.
We have recently found that efficient biosynthetic replacement
of mycosamine with perosamine on 1 requires both inactivation
of amphDII and introduction of a hybrid glycosyl transferase.29
In extensive studies reported here, production cultures of am-
phNM+perDIDII were unpredictable and could yield 8 or the
aglycone 3, or both, in varying proportions. This instability may
result from recombination between directly repeated homolo-
gous amphDI-DII and perDI-DII sequences, which would lead
to loss of functional glycosylation genes. Use of high levels
of thiostrepton increased production of 8, whilst subculturing
amphNM+perDIDII on thiostrepton-free media increased pro-
duction of 3.
Isolation of 8-deoxy-16-descarboxyl-16-methyl-amphoteronolide B
(3) from the amphNM+perDIDII mutant
Repeated subculturing of the amphNM+perDIDII mutant in
thiostrepton-free GYE media led to formation of increased levels
of aglycone 3 in the subsequent production cultures. Methanolic
extraction of sediment followed by partial removal of volatiles in
vacuo gave a yellow precipitate of 3 (>100 mg L-1) containing
less than 10% other polyenes and a relatively small (<10% w/w)
amount of lipid. Repeated washing with methanol improved
purity, but did not remove lipid completely. Despite being initially
extracted into methanol from mycelia, the purified aglycone was
quite insoluble, illustrating how the properties can change with
increasing purity. Despite the low solubility of 3, most carbon and
proton resonances could be assigned. Heterocorrelation NMR
spectroscopy revealed a carbon resonance at 23.7 ppm correlated
with protons at 1.77 and 0.99 ppm, characteristic of an 8-deoxy
analogue.
Reaction of 3 with benzoyl chloride initially gave a heptabenzoy-
lated derivative, which after continued reaction time converted into
a semicrystalline octabenzoylated derivative. 13C NMR spectro-
scopic analysis showed a ketone resonance at 203 ppm indicating
that the lactol had been trapped in the ring opened form as the
benzoyl ester 10 (Fig. 6). The positive electrospray gave a peak at
1582.7 (M + water)+ and FAB mass spectra at 1564.67 (M+) rather
than the expected protonated analogues. The sample examined
by FAB-MS contained some heptabenzoylated derivative, which
FAB showed as the (M - water)+ peak, indicating that the final
position benzoylated is 17-hydroxyl as the ring opened lactol.
Fig. 6 Octabenzoylated aglycone (10).
Borowski et al. have previously trapped the lactol of protected
1 as the O-methyl oxime,30 and MacPherson et al. the oxime.31
Isolation of 16-descarboxyl-16-methyl-amphotericin B (8) from the
amphNM+PerDIDII mutant
Analysis of pentaenes from the M57 mutant lacking polyketide
synthase modules 5 and 6 in AmphC
The amphNM+perDIDII mutant was grown in the presence of
thiostrepton (20 mg mL-1). The sediment was extracted with
methanol, and the volatiles partially removed. The resulting crude
precipitate containing 8 was largely purified from tetraenes by
water washes, and membrane lipid by repeated selective dissolu-
tion with sonication in methanol and subsequent precipitation
with EtOAc, facilitating preparative HPLC purification. Positive
ion ESMS gave peaks at 916.5 (M+Na)+, or after loss of water
at 876.5, corresponding to the expected mass of 8 (893.5 Da).
Heterocorrelation NMR spectroscopy did not show any cross-
peak at ca. 21 ppm to 1.7 and 1.0 ppm characteristic of an 8-
deoxy compound, but instead a carbon resonance at 31.1 ppm
reminiscent of 1. Thus AmphL has efficiently oxidised (at C-8)
the 16-methyl glycosylated heptaene, showing that a 16-carboxyl
group is not necessary for AmphL catalysed oxidation, and that
AmphL does not require AmphM ferredoxin.
Zotchev et al. engineered the nystatin PKS to produce a hexaene
analogue by eliminating one module from NysC.32 He has also
reported an octaene homologue due to stuttering in the nystatin
PKS.33 The tailoring enzymes were active with both hexaene and
octaene analogues.
In a previous report, we probed the flexibility of the am-
photeronolide synthase by the precise deletion of two PKS
modules from AmphC that normally assembles the heptaene
portion of 1.16 Extracts of the M57 mutant gave a UV absorption
spectrum with maxima at 310, 318, 333 and 352 nm, in agreement
with reported pentaenes.25
These pentaenes have now been absorbed from the culture
supernatant by the addition of XAD16, along with large quantities
of contaminants from the broth. Addition of diazomethane en-
abled purification of pentaenes by silica chromatography. Normal
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3758–3770 | 3761
©

View Article Online
phase HPLC revealed several major pentaenes, whose RMM
and NMR spectra were consistent with the methyl ester of
pentaene aglycones with the propionate derived methyl group
oxidised to a carboxylate, but without AmphDI glycosylation or
AmphL oxidation. There was no evidence for polyenes other than
pentaenes. NMR analysis of the major pentaene was consistent
with 11 (as its methyl ester 12) (Fig. 7) along with a small
level of unknown aromatic impurity. The stereochemistry shown
is assumed by analogy with 1. The carbon NMR showed ten
double bond resonances as expected at 130–140 ppm, a hemiacetal
resonance at 97.5 ppm, and lactone or ester resonances at 171
and 173. There were three methine and eight visible methylenes
between 20 and 49 ppm, including a methylene at 21.6 ppm
characteristic of a ‘deoxy’ compound. There is expected to be a
ninth methylene under the DMSO resonance at around 47 ppm by
analogy with the NMR of aglycone 3. There are nine heteroatom
linked methines at 50–80 ppm (7 ¥ hydroxy, CHCOOH and lactone
alcohol), but two have shifted 5–10 ppm upfield relative to those
in the aglycone (3) NMR spectrum.
Isolation of 7-oxoamphotericin B (9) from the KR16 mutant
The underivatised analogue 9 has now been obtained (ca
100 mg L-1; ca 75% of extract levels) by multiple washing of the
original crude yellow precipitate with water, followed by washing
with chloroform. Tetraenes in the crude extract are also removed
by these washings. Whilst the crude material (bound to lipid)
was water insoluble, this purified material is relatively soluble
in water (ca 1 mg mL-1). A major factor in the toxicity of 1 is
its lack of serum-solubility that causes aggregation in serum and
accumulation in the kidneys.
7-Oxo-polyene formation (ca. 250 mg L-1) shows that the
remaining modules have indeed recognised the unreduced car-
bonyl and continued processing the growing acyl chain as usual
(Scheme 3). The unreduced 3-oxothioester in 13 is recognised by
KS17 and chain elongated to the 3,5-dioxothioester 14. Despite
the enhanced enolisation and reduced conformational flexibility
expected in its modified substrate, KR17 reduces the 3-oxo car-
bonyl in 14, allowing for the final addition of a C2 unit, reduction
by KR18 and lactonisation (TE) to give the 7-oxo aglycone 15.
The sp3 flexibility normally present at ‘C-7’ can in this case be
substituted by the more constrained planar carbonyl group. This
aglycone (15) appears to be efficiently transformed by AmphN,
AmphDI and AmphL (despite the neighbouring carbonyl) to 9.
Fig. 7 Pentaene (11) and methyl ester (12).
Despite the removal of two modules, the PKS has successfully
completed the acyl chain assembly, and the thioesterase formed
the ring contracted lactone.
The M57 mutant has also been grown in the presence of XAD16,
and in this case, the pentaenes extracted from the sediment. After
methanol extraction and partial solvent removal, examination
of the resultant precipitate by LCMS indicated some (<10%)
glycosylated material (with and without hydroxylation by AmphL
at C-6); most other peaks were isomeric with pentaene 11 with full
AmphN oxidation but the conspicuous absence of any AmphL
hydroxylation. This indicates that oxidation of the methyl group
by AmphN is facile and occurs first, that AmphDI has reluctantly
recognised the pentaene aglycone, and that AmphL catalysed
oxidation only occurs after this glycosylation.
The KR16 mutant
The heptaenes candidin and mycoheptin have ketone groups at C-7
and C-3, respectively, and have all three post-PKS modifications as
for 1.34 This suggests that heptaene PKSs can tolerate inactivation
of KR domains in modules 16 to 18. and, that the resulting
oxoaglycones can be readily modified by all three post-PKS
enzymes. Zotchev et al. describe the inactivation of KR16 and
KR17 in the nystatin PKS, along with NysN cytochrome P450
monooxygenases to obtain a range of oxo- and descarboxyl-
analogues of nystatin.27
Scheme 3 Biosynthesis of 7-oxoamphotericin B (9).
A sample was purified by HPLC and examined by 13C NMR
spectroscopy to reveal a ketone resonance at 212.75 ppm.
Resonances between 65 and 80 ppm are considerable altered
from those in amphotericin B itself, indicating conformational
alteration upon the introduction of the sp2 centre at 7-C into the
macrolactone ring.
We have previously replaced the KR16 active site tyrosine
with phenylalanine, abolishing ketoreduction in module 16, and
rendering superfluous the DH16 and ER16 enzymes, resulting in
9 isolated as the N-acetyl methyl ester.17
Polyene products from the DNM KR12 mutant
The KR12 domain was inactivated in the amphNM mutant.17
HPLC and LCMS analysis of sediment extracts indicated
3762 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
formation in poor yield (ca 2 mg L-1) of 16-descarboxyl-15-deoxy-
16-methyl-15-oxo-amphotericin B (16) and 16-descarboxyl-8,15-
dideoxy-16-methyl-15-oxo-amphotericin B (20), the expected
products of KR12 inactivation, followed by AmphDI glycosy-
lation, with and without AmphL oxidation.
to that previously used with the KR12 and KR16 domains17 and
as detailed in the Experimental section (Fig. 8).
The hemiketal ring of the polyene normally results from reaction
of the hydroxyl group introduced in cycle 11 with the ketone
remaining after cycle 13. Though it remains unclear at which
step this occurs, hemiacetal ring formation at the end of cycle
13 (processive) may be essential in folding the growing acyl chain
for subsequent chain extension and processing.
There is no KR13 in the amphotericin PKS, so at the end
of cycle 13 the KR12 mutant will generate a 3,5-dioxothioester
18 (Scheme 4). Hemiacetal ring formation may be slowed by
the increased enolisation of this 3,5-dioxo intermediate 18 and
by the additional sp2 centre in the hemiacetal ring of 19. If
chain extension of 18 by KS14 now competes with the impaired
hemiacetal formation of 18, then an acyclic 3,5,7-trioxothioester
should appear prior to ketoreduction in cycle 14. This enolised
and conformationally rigid intermediate might not be recognised
or processed efficiently by downstream activities, blocking the
synthase. The product 16 was obtained in poor yield, without
evidence of any truncated polyene products released from the PKS.
Fig. 8 Evidence for inactivation of KR10 coding sequence. PCR primers
KR10CF and KR10CR2 were used to amplify the KR10 coding region
from S. nodosus (lane 1) and the KR10-1 mutant (lane 2). Treatment with
HindIII revealed that the DNA amplified from S. nodosus was resistant to
digestion (lane 3) whereas that from the mutant contained the expected site
(lane 4). In control digests, Bcl I cut the PCR products from both strains
at an internal site (lanes 5 and 6).
The original aim was to engineer in vivo synthesis of substantial
quantities of 19-deoxy-19-oxo-amphoteronolides for subsequent
conversion into a library of analogues. Mycelium from the first
gene replacement mutant (KR10-1) was extracted with methanol
to give an orange solution with UV maxima at 422, 411, 385
and 309 that gave largely one peak on analysis by HPLC. After
removal of solvent from the KR10-1 extract, the resulting solid
(mass 135 mg L-1) was readily soluble in ethyl acetate, and
could be purified by flash chromatography to give an orange
product (15 mg L-1 unoptimised). Negative ion electrospray mass
spectrometry showed RMM of 386, as well as some singly charged
dimer and trimer. ESMS indicates low levels of additionally
oxygenated material with RMM 402. NMR spectroscopy showed
three methyl doublets, ten vinylic protons, only four other protons,
and only twenty-three carbon resonances, including three at 157,
166 and 167 ppm and two at 99 and 100 ppm consistent with a
pyrone structure. Correlation spectroscopy showed connectivity
from 9¢-H through to 17¢-H. This data is consistent with the
decaketide 21. The stereochemistry shown for 21 is that expected
for an intermediate in the biosynthesis of 1.
There are only a few instances where inactivation of a polyketide
biosynthetic gene has led to formation of truncated pyrone
intermediates. A polyene-pyrone heptaketide was generated by
the lovastatin producer Aspergillus terreus when LovC (an enoyl
reductase) was disrupted.35 Overexpression and isolation of LovB
gave protein that could assemble penta-, hexa- and heptaketide
pyrones. The absence of SAM, and consequent lack of methylation
at the tetraketide stage, has caused derailment and offloading
of subsequent partially assembled acyl chains.36 Intermediates
in rifamycin biosynthesis ranging from the tetraketide to the
decaketide were isolated following deletion of the rifF gene for
Scheme
oxo-amphotericin B (16).
4
Biosynthesis
of
deoxy-16-descarboxyl-16-methyl-15-
Truncated polyketide intermediates from KR10-1 and KR10-2
mutants
The KR10 coding sequence in the S. nodosus genome was replaced
with a version encoding an inactive enzyme using a strategy similar
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3758–3770 | 3763
©

View Article Online
the lactam-forming chain-terminating enzyme. The tetraketide
intermediate was obtained as both a pyrone and a linear carboxylic
acid.37 Prematurely released intermediates have been obtained by
deletion of domains in the trans-AT PKS responsible for bacillaene
biosynthesis.38
The structure of 21 is reminiscent of the chloropyrrole polyene-
pyrone rumbrin produced by the fungus Auxarthon umbrinum and
the furan polyene-pyrone gymnoconjugatin B produced by the soil
ascomycete Gymnoascus reessii.39
The structure of 21 suggested that two modules of AmphC
had been lost by a spontaneous deletion (precise excision) in
addition to the targeted inactivation of KR10. To investigate
this further, primer pairs were designed for the amplification of
the DNA sequences coding for ACP3, ACP4, ACP5, ACP6 and
ACP7 domains (Fig. 9). PCR analysis indicated that the ACP5
and ACP6 coding sequences were absent from the chromosome
of the KR10-1 mutant, which is consistent with the absence of a
dihydroanalogue of 21 that would result from reduction of one the
double bonds in the polyene unit by ER5.
Scheme 5 Biosynthesis of KR10-1 decaketide product (21).
and 34.0 ppm in the carbon NMR spectrum. The yield of 24 was
estimated at over 20 mg L-1 by NMR purity and signal to noise
ratio.
No KR10 mutant was isolated that could produce a full-length
macrolactone despite extensive screening. The KR10-1 strain has
undergone a deletion equivalent to precise excision of modules
5 and 6. In the KR10 mutants, module 11 is still able to chain
extend 22 to 23 or 25 to 26 generating the corresponding 3,5-
dioxothioester system. However, this is no longer processed as
usual by the otherwise functionally active KR11. This may be
due to either lack of binding, incorrect folding in the active site
due to the extra 5-oxo group, or due to extra enolisation that
extends the system of conjugated double bonds. Ketoreduction
may simply now be much slower than the competing non-enzyme
catalysed pyrone formation and release from the PKS. Module
11 of the amphotericin PKS incorporates a propionate extender
and is thought to epimerise the initial 2R-methyl branch in its
product prior to ketoreduction. It has been suggested that cis enols
feature as intermediates in epimerisation reactions in modules
that generate the same (anti) methyl and alcohol stereochemistry
as module 11.40 Such an intermediate (27) might now favour
spontaneous pyrone formation prior to ketoreduction by KR11
(Scheme 6). Alternatively, the decaketide and dodecaketide chains
may be released from ACP11 by AmphE, a discrete editing
thioesterase,18 with spontaneous pyrone formation occurring later.
Whatever the mechanism of termination, the isolation of 21 and
24 suggest that the intermediates are efficiently off-loaded from
the PKS.
Fig. 9 PCR analysis of AmphC ACP coding sequences in S. nodosus and
the KR10-1 mutant. Lanes 1, 3, 5, 7 and 9, DNA amplified from S. nodosus
genomic DNA with primers for sequences encoding ACPs 3, 4, 5, 6 and
7, respectively. Lanes 2, 4, 6, 8 and 10, DNA amplified from S. nodosus
KR10-1 genomic DNA with primers for sequences encoding ACPs 3, 4, 5,
6 and 7, respectively. The ACP5 and ACP6 coding sequences are missing
from the KR10-1 mutant (lanes 6 and 8).
The structure also suggested that the 3-oxoacyl intermediate
22 normally reduced by KR10 can instead be efficiently chain
extended by KS11 to give 3,5-dioxo intermediate 23 (Scheme 5).
However, KR11 has not reduced 23, and neither has any further
chain elongation occurred. This was surprising because we do
not observe the formation of linear pyrones by the KR16 mutant
that also generates a 3,5-dioxothioester intermediate 14 in cycle
17 (Scheme 3). KR17 and chain extension operates efficiently on
this intermediate.
Further KR10 gene replacement mutants were sought in an
effort to obtain a strain synthesising full-length heptaenone
macrolactones. This gave strain KR10-2 that generated a mixture
of heptaene and tetraene dodecaketides released from PKS module
11 as pyrones. Extracts of the mycelia with ethyl acetate gave
compounds that were difficult to quantify or purify. Partial
purification by flash chromatography gave dark red fractions
that when analysed by ESMS gave mass predominantly 437.3,
and fractions with mass predominantly 439.4 corresponding to
the tetraene-pyrone 24. Tetraene-pyrone (24) containing fractions
were analysed by NMR spectroscopy to reveal spectra similar to
that for 21 except for the presence of four allylic protons at 2.20
and 2.23 ppm in the proton NMR and two allylic carbons at 33.5
The polyenyl-pyrones 21 and 24 were tested for antifungal
activity as described previously.15 Neither compound showed
activity even at high concentrations (>2 mg mL-1).
Conclusions
A wide range of analogues of 1 have been produced in vivo, many
with titres over 100 mg L-1. This indicates that the biosynthetic
machinery has sufficient flexibility to be able to produce promising
3764 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
amphotericin PKS might be used for precursor-directed synthesis
of new pyrones.
Experimental
General
UV assays of polyenes were run on a Shimadzu UV-2401PC
spectrophotometer. For heptaenes, an absorbance e of 1.7 ¥
105 M-1 cm-1 at 405 nm was used, based upon the value reported
for monomeric amphotericin, and for tetraenes an absorbance e of
0.78 ¥ 105 M-1 cm-1 at 318 nm was used.21 Pentaenes were assayed
using e of 0.85 ¥ 105 M-1 cm-1 at 333 nm.25
HPLC analyses were performed on a Varian Prostar diode array
with Galaxie workstation and software. Analytical HPLC were
run using using 4.6 mm ¥ 25 cm Supelco silica or C-18 silica
(5 mm) with 1 mL min-1 flow rate and semipreparative HPLC
using 21.2 mm diameter columns at 14.8 mL min-1 flow rate using
silica, C-8 or C-18 columns. In general, every passage through
silica or HPLC dramatically reduced yield.
NMR spectra were collected on a Bruker AV500 MHz Avan-
ceIII, or DRX400 spectrometer. Chemical shifts are quoted in
ppm. Coupling constants quoted in near-first order sytems are
‘observed’ values. Proton resonances are (1 H, m) unless otherwise
indicated. Attached proton test (APT) or ‘DEPT135¢ used in 13
C
NMR multiplicity assignments.
Electrospray mass spectrometry was performed on a triple
quadropole Micromass Quattro LC machine or Kratos Concept
1H double focussing high resolution spectrometer. FAB used 3-
nitrobenzyl alcohol as matrix. Quadropole time of flight mass
spectrometry (QTOFMS) was run on a XEVO instrument with
Waters Accuity UP-LC.
Scheme 6 Biosynthesis of KR10-2 dodecaketide tetraene (24).
analogues at levels and cost comparable to 1. Many compounds
displayed strong affinity for membrane lipid and low solubility,
complicating full purification and analysis at current titre levels.
Most analogues have been readily prepared in purity similar to
commercial samples of 1 using non-chromatographic methods
that could be affordably scaled up. The compounds isolated
support the non-obligate preferred order of final biosynthetic steps
being oxidation at C-16 by AmphN, glycosylation at O-19 by
AmphDI, followed by oxidation at C-8 by AmphL. The AmphL
oxidation occurs more readily on glycosylated compounds, whilst
AmphDI glycosylation can readily occur on 16-descarboxyl-16-
methyl analogues. Deletion of two PKS modules from AmphC
gave fully assembled pentaene-carboxylate compounds, demon-
strating not only the potential flexibility of the downstream
modules including the final lactonising thioesterase, but also the
AmphN oxidation enzyme.
Disruption of a ketoreductase (KR) close to the end of the
PKS acyl chain assembly process led efficiently to the expected
7-oxo-product with full subsequent acyl chain extension and
lactonisation. In contrast, KR12 disruption gave only low yields
of the corresponding 15-oxo-products suggesting that hemiacetal
formation is ‘processive’ and moderated by the introduction of
a nearby carbonyl group. Disruption of KR10 resulted in no
detectable polyene macrolide. Direct acyl chain extension of the
KS10 intermediate by KS11 to a 3,5-dioxothioester has resulted
in efficient cyclisation to linear polyenyl-pyrones in preference to
reduction by the next domain, KR11, and further chain extension
to a macrolide. The truncated linear polyenyl-pyrones synthesised
by the viable KR10 mutants are analogous to bioactive polyenyl-
pyrones synthesised by several fungal species. The truncated
ATR infra red spectroscopy was run on a Perkin Elmer
‘Spectrum One’ instrument.
Cultures are grown in ‘trigrooved’ Erlenmeyer flasks (250 mL
or 2 L; washed in deionised water) in environmentally controlled
shakers (New Brunswick Series 25 with 25 mm gyrotary and Sanyo
Gallenkamp ‘Orbisafe’ 32 mm gyrotary). Media components are
from Sigma and made up in deionised water, and the flask necks
plugged with a sponge before autoclaving.
Thiostrepton was added to cultures of strains containing a
thiostrepton resistance marker. Stock solution of 100 mg mL-1
were made up in DMSO and frozen at -26 ◦C until needed.
Working concentrations of thiostrepton were 5 mg mL-1 in
liquid culture and 50 mg mL-1 in solid culture unless stated.
Centrifugation was performed using a Sorvall RC5B centrifuge
with a GSA (13,000 rpm) or SS34 rotor (20,000 rpm).
All cultures are maintained at -80 ◦C as glycerol deeps. The
contents (1.5 mL) are added to a GYE preculture medium of
glucose (10 g L-1), yeast extract (10 g L-1) and deionised water
(100 mL) adjusted to pH 7 with sodium hydroxide (2 M) and
◦
incubated (28 C; 200 rpm) for 48 h. Aliquots (10 mL) are then
used to inoculate production media (28 ¥ 2 L trigrooved flasks
each containing 250 mL broth) containing fructose (20 g L-1),
Dextrin (corn, Type II; 60 g L-1), Soybean flour (Type I not roasted;
30 g L-1), CaCO3 (10 g L-1) and Amberlite XAD16 resin (50 g L-1)
autoclaved at 121 ◦C for 20 min. The remaining preculture was
retained, and if polyene production levels were found to be high,
the contents were used to make fresh growth and/or glycerol deeps.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3758–3770 | 3765
©

View Article Online
The inoculated production media was incubated at 28 ◦C for
4–5 days to give a distinctive earthy odour, and copious mycelia
that when shaken slid slowly down the sides of the Erlenmeyer
flask. The mycelia and XAD16 resin was collected (GSA; 10 min).
All water baths used in rotary evaporation of extracts or
concentration of extracts (‘in vacuo’) were kept below 40 ◦C.
with GenBank accession number AF357202) fused to the perDI
and perDII genes. Both amphDI and perDI genes have an NcoI
site (5¢ CCATGG 3¢) surrounding the start codon. The S. nodosus
DNA was ligated to the S. aminophilus DNA through this con-
served site. This positioned the perDI start codon at the optimum
distance from the amphDI promoter and ribosome-binding site.
The KC-Per1 phage integrated into the S. nodosus DamphNM
chromosome by recombination between the homologous amphDI
upstream sequences. This gave a thiostrepton-resistant lysogen
S. nodosus DamphNM + perDI-DII which contains the related
perDI-perDII and amphDI-DII regions in direct repeat orienta-
tion. The genotype was verified by Southern hybridisation.
8-Deoxyamphotericin B (5)
All media used in production of polyenes from the amphL mutant
contained thiostrepton (50–100 mg mL-1). The mycelia and resin
from production media (7 L) was added to a solution of sodium
thiocyanate (900 g) in methanol (1.5 L) and shaken (200 rpm) in
trigrooved flasks for 2 h to give a dark coloured solution containing
heptaene (350 mg). The mycelia and beads were collected (GSA, 10
min) and the extraction repeated with fresh methanol/thiocyanate
solution (1.5 L). Assay of the second extract indicated heptaene
(1.4 g) in a much clearer solution. An equal volume of water
was added to this second extract, and left at 4 ◦C overnight. The
resulting precipitate was collected (GSA, 10 min), washed with
water (2 ¥ 250 mL), redissolved in thiocyanate saturated methanol
and precipitated with an equal volume of water as before. The
resulting precipitate was washed with water (2 ¥ 250 mL) and
freeze dried to give a yellow solid (8 g) that by UV assay in
DMSO/methanol contained heptaene (800 mg). The yellow solid
was dissolved in DMSO (5 ¥ 50 mL) and any remaining insoluble
material was removed by centrifugation, and then ether (250 mL)
added to give a precipitate. Addition of DMSO to the precipitate
gave an oil that contained concentrated heptaene (10–15 mg
mL-1). Addition of EtOAc (8 vol) gave a clear supernatant and a
precipitate that was washed (MeOH; 3 ¥ 100 mL) to give 5 (0.9 g;
90% heptaene w/w by UV assay) with reduced contamination by
1. HPLC (C-18; water–methanol containing methanoic acid 0.1%
v/v) was used to obtain a sample free of 1.
16-Descarboxyl-16-methyl-amphotericin B (8)
The mycelia and XAD16 resin sedimented from the am-
phNM+perDIDII mutant [7
L growth; with thiostrepton
(20 mg L-1) in both preculture and production broth] was soaked
in methanol (3 ¥ 7 L; 3 ¥ 6 h) to give an extract containing
heptaene (1.13 g) and tetraene (1.1 g). Partial removal of combined
solvent in vacuo gave a yellow precipitate (600 mg heptaene and
700 mg tetraene in 6 g). The precipitate was dissolved in methanol
(100 mL), water added (300 mL), left at 4 ◦C for 2 h, and the
precipitate collected (400 mg heptaene and 300 mg tetraene in
1.6 g). This was dissolved in meth◦anol (ca 25 mL), EtOAc slowly
added (20 volumes) and left at 4 C for 4 h to give a precipitate
(110 mg heptaene and less than 5 mg tetraene). More heptaene can
be recovered from the supernatant by repeating this procedure. A
sample was purified by preparative HPLC (C-18, water–methanol
plus methanoic acid 0.1%) and examined by NMR spectroscopy.
lmax (MeOH)/nm 405, 383 and 362.
Partial dH(500 MHz; d4-MeOH) 1.037 (3 H, d, J 7.1, 41-H),
1.039 (3 H, d, J 6.1, 39-H), 1.14 (3 H, d, J 6.3, 40-H), 1.22 (3 H,
d, J 6.4, 38-H), 1.25, 1.31 (3 H, br d, J 5.7, 6¢-H), 1.36, 1.39, 1.42,
1.47, 1.51, 1.61, 1.73 and 1.74, 1.93 (36-H), 2.00 (1 H, dd, J 12.2
and 4.7), 2.23 (1 H, ABX, 2b-H), 2.24 (18a-H), 2.30 (1 H, ABX,
2a-H), 2.41 (34-H), 3.10 (1 H, br d, 8-H), 3.20, 3.22 (35-H), 3.30
(5¢-H), 3.39 (9-H), 3.59, 3.63, 3.74, 3.93 and 3.99, 4.20 (1 H, tt,
J 9.6 and 3.0), 4.35, 4.46 (19-H), 4.64 (1 H, br s, 1¢-H), 5.40 (2 ¥ 1
H, 2 ¥ m, 33-H and 37-H), 6.04 (1 H, dd, J 15.2 and 8.9, 20-H),
6.15-6.5 (12 H, m).
dH(500 MHz; d6-DMSO) 0.92 (3 H, d, J 6.9, 39-H), 1.01 (7b-H),
1.04 (3 H, d, J 5.9, 40-H), 1.08 (6b-H), 1.10 (14b-H), 1.12 (3 H,
d, J 6.3, 38-H), 1.17 (8b-H), 1.18 (3 H, d, J 4.9, 6¢-H), 1.24 (2 ¥
1 H, 2 ¥ m, 6a-H and 8a-H), 1.31 (4b-H), 1.34 (2 H, m, 10ab-H),
1.38 (4a-H), 1.51 (12b-H), 1.56 (18b-H), 1.58 (H-12a), 1.73 (7a-H),
1.74 (36-H), 1.91 (14a-H), 1.94 (16-H), 2.10 (18a-H), 2.16 (2 H,
m, 2ab-H), 2.29 (34-H), 2.85 (3¢-H), 3.10 (35-H), 3.18 (4¢-H), 3.21
(5¢-H), 3.46 (5-H), 3.58 (9-H), 3.77 (2¢-H), 4.02 (15-H), 4.06 (3-H),
4.20 (17-H), 4.23 (11-H), 4.40 (1 H, br s, 19-H), 4.46 (1 H, br s,
1¢-H), 5.21 (1 H, ca. dq, 37-H), 5.43 (1 H, dd, J 14.9 and 9.9,
33-H), 5.95 (1 H, dd, J 14.8 and 8.8, 20-H), 6.04–6.39 (11 H, m),
6.45 (1 H, ca dd).
Partial dC(125.8 MHz; d4-MeOH) 12.2 (39-C), 13.5 (41-C), 17.1
(38-C), 17.6 (6¢-C), 19.0 (40-C), 31.1 (t), 33.3 (18-C), 36.2 (t), 38.2
(t), 40.7 (t), 41.7 (36-C), 42.7 (2-C), 43.6 (16-C), 44.1 (34-C), 44.7
(t), 45.6 (t), 48.1 (t), 57.2 (8-C), 68.9 (d), 69.6 (d), 69.7 (d), 70.3
(d), 70.6 (9-C), 70.7 (37-C), 71.1 (d), 72.6 (d), 74.5 (5¢-C), 75.7 (d),
75.9 (d), 79.4 (19-C), 79.8 (35-C), 99.1 (1¢-C).
Partial dC(125.8 MHz; d6-DMSO) 11.9 (39-C), 16.8 (38-C), 17.5
(6¢-C), 18.2 (40-C), 21.4 (7-C), 36.9 (18-C), 38.5 (6-C), 38.8 (8-C),
39.5 (36-C), 41.9 (2-C), 42.2 (34-C), 43.6 (10-C), 44.1 (14-C), 44.6
(4-C), 46.3 (12-C), 55.1 (3¢-C), 57.2 (16-C), 65.2 (17-C), 65.5 (15-
C), 66.1 (3-C), 68.3 (11-C), 68.8 (37-C), 68.9 (2¢-C), 69.1 (5-C),
69.9 (4¢-C), 71.3 (9-C), 72.7 (5¢-C), 74.8 (19-C), 77.2 (35-C), 96.3
(1¢-C), 128.9 (d), 130.9 (21-C), 136.2 (20-C), 136.8 (33-C).
m/z (+ve ES) 908 (MH)+
m/z (+ve ES) 916.5 (M+Na+), 876.5 (M-water+H+).
m/z (FAB) Found: 894.51939 (MH+). C47H76NO15 requires
894.51959.
8-Deoxy-16-descarboxyl-16-methyl-amphoteronolide B (3)
The mycelia and XAD16 resin sedimented from am-
phNM+perDIDII mutant (7 L growth grown with thiostrepton
in the innoculum but not the production medium) was repeatedly
(x 6) soaked in methanol (5 L) overnight with occasional shaking.
After each overnight soaking, the sediment was collected (GSA
10,000) for further soaking. Partial removal of combined solvent
m/z (-ve ES) 906.3 (M-H)-
Construction of S. nodosus DamphNM + perDI-DII
Phage KC-Per1 contains a StuI-BamHI insert consisting of the
amphDI upstream region (nucleotides 65861 to 68195 of sequence
3766 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
in vacuo and storage at 4 ◦C gave a yellow precipitate which
was collected by centrifugation. The precipitate was resuspended
in MeOH (500 mL) vigorously shaken with sonication, and the
remaining yellow powder collected by Buchner filtration. This
process was repeated until washings are clear, to give a yellow
powder (500 mg) found to be analytically pure, but estimated by
UV assay at 95% purity by weight, and containing some (< 5%)
membrane debris by NMR analysis.
and 3.1), 2.02, 2.08, 2.14, 2.16, 2.31, 2.34, 2.65 (1 H, dd J 16.1, 6.5,
2-H), 2.67 (2-H), 2.72 (34-H), 2.74 (1 H, dd J 16.6, 3.5), 2.84 (1 H,
dd, J 16.6, 9.2), 2.89 (1 H, dd, J 18.2, 6.8), 3.05 (1 H, dd, J 18.2,
4.3), 5.00 (2 ¥ 1 H, 2 ¥ ca d), 5.16, 5.24, 5.40, 5.42, 5.45 (33-H),
5.47 (37-H), 5.48 (19-H), 5.55, 5.80 (1 H, dd J 9.3 and 9.1, 20-H),
6.22–6.54 (12 H, m, 21 to 32-H), 7.22–7.35 (8 H, m, Ar-meta),
7.41–7.52 (12 H, m, ArH 4 ¥ para + 8 ¥ meta), 7.55–7.62 (4 H,
m, para), 7.78–7.82 (2 H, ca t, ortho), 7.88–7.92 (6 H, m, ortho),
8.01–8.08 (6 H, m, ortho), 8.13–8.17 (2 H, m, ortho).
Found C, 67.32; H, 8.69%. C41H63O11 requires C, 67.28; H,
8.68%;
lmax (MeOH)/nm 404, 381, 362.
dC(125.8 MHz; CDCl3) 10.4 (q), 11.1 (q), 16.3 (q), 18.2 (q), 21.0
(t), 34.3 (t), 34.5 (t), 34.8 (t), 38.21 (t), 38.24 (t), 38.8 (t), 39.0 (d),
40.6 (d), 41.2 (d), 46.7 (t), 47.1 (t), 68.6 (d), 68.7 (d), 69.0 (d), 69.7
(d), 71.3 (d), 71.46 (d), 71.50 (d), 73.7 (d), 79.3 (d), 128.13 (d),
128.16 (d), 128.2 (d), 128.3 (d), 128.44 (d), 128.49 (d), 128.53 (d),
129.5 (d), 129.6 (d), 129.65 (d), 129.7 (d), 129.85 (d), 129.9 (d),
130.0 (d), 130.1 (d), 130.2 (s), 130.3 (s), 130.9 (d), 132.3 (d), 132.5
(d), 132.6 (d), 132.7 (d), 132.77 (d), 132.80 (d), 132.9 (d), 133.0 (2
C, d), 133.1 (d), 133.3 (2 C, d), 133.4 (2 C, d), 133.6 (d), 133.7 (d),
133.9 (d), 134.0 (d), 134.1 (d), 134.3 (d), 134.7 (d), 135.0 (d), 136.7
(d), 149.0 (d), 165.49 (s), 165.51 (s), 165.58 (s), 165.64 (s), 166.01
(s), 166.02 (s), 166.19 (s), 166.35 (s) 169.0 (s), 203.0 (s).
dH(500 MHz; d6-DMSO) 0.88 (3 H, d, J 6.3, 41-H), 0.92 (3 H, d,
J, 7.0, 39-H), 0.99 (7b-H), 1.04 (3 H, d, J 6.3, 40-H), 1.06 (16-H),
1.08 (6b-H), 1.11 (3 H, d J 6.0, 38-H), 1.13 (14b-H), 1.16 (8b-H),
1.23 (8a-H), 1.29 (H-4b), 1.31 (6a-H), 1.32 (2 H, m, 10ab-H), 1.38
(4a-H), 1.48 (12b-H), 1.52 (18b), 1.54 (18a-H), 1.73 (36-H), 1.77
(7a-H), 1.83 (1 H, ca dd, 14a-H), 1.90 (1 H, ca dd, 18a-H), 2.10 (1
H, ABX, 2a-H), 2.16 (1 H, ABX, 2b-H), 2.30 (34-H), 3.09 (35-H),
3.38 (15-H), 3.43 (1 H, br s, 5-H), 3.57 (9-H), 3.85 (1 H, ca t, J
9.2, 17-H), 4.07 (3-H), 4.19 (19-H) 4.21 (11-H), 4.50 (1 H, d, J 4.7,
5-OH), 4.61 (1 H, d, J 4.7, 5-OH), 4.71 (1 H, d, J 4.1, 19-OH),
4.78 (1 H, ca d, 35-OH), 4.79 (1 H, ca d, 35-OH), 4.99 (1 H, d, J
4.1, 9-OH), 5.21 (1 H, ca dq, 37-H), 5.42 (1 H, dd, J 15.0 and 10.0,
33-H), 5.49 (1 H, br s, 11-OH), 5.65 (1 H, br s, 13-OH), 5.96 (1 H,
dd, J 15.1 and 10.1, 20-H), 6.06 to 6.43 (19 H, m, 21-H to 33-H).
dC(125.8 MHz; d6-DMSO) 14.1 (39-C), 15.2 (41-C), 18.9 (38-
C), 20.4 (40-C), 23.7 (7-C), 40.8 (6-C), 41.0 (8-C), 41.5 (36-C) 42.1
(18-C), 43.9 (2-C), 44.42 (34-C), 44.43 (16-C), 45.7 (10-C), 46.9
(4-C), 47.1 (14-C), 48.7 (12-C), 68.1 (3-C), 69.9 (15-C), 70.3 (11-
C), 70.7 (37-C), 70.8 (17-C), 70.9 (5-C), 71.3 (19-C), 73.2 (9-C),
79.2 (35-C), 98.8 (13-C), 129.3 (31-C), 133.1, 133.2, 133.7, 134.0,
134.2, 134.4, 134.5, 135.0, 135.5, 135.6, 135.7, 138.8 (33-C), 142.2
(20-C), 172.6 (1-C).
m/z (+ve ES) 1582.7 (M+water)+
m/z (FAB) 1564.67 (M)+, 1442 (Heptabenzoyl-water)+.
(1R,3S,5S,9R,11R,15S,16R,17R,18S,21E,23E,25E,27E,29R,
31S,32R,33S)-1,3,5,9,11,17,29,33-Octahydroxy-15,16,18-
trimethyl-13-oxo-14,35-dioxabicyclo[29.3.1]pentatriaconta-
19,21,23,25,27-pentaene-32-carboxylic acid (11) and methyl
(1R,3S,5S,9R,11R,15S,16R,17R,18S,21E,23E,25E,27E,29R,
31S,32R,33S)-1,3,5,9,11,17,29,33-octahydroxy-15,16,18-
trimethyl-13-oxo-14,35-dioxabicyclo[29.3.1]pentatriaconta-
19,21,23,25,27-pentaene-32-carboxylate (12)
m/z (+ve ES) 755.5 (M+Na+).
The M57 mutant was grown (8 ¥ 250 mL) in the absence of XAD16
resin. Sediment was soaked in in MeOH (800 mL L-1 culture) for
18 h, but found to contain only low levels of pentaene (<20 mg L-1).
UV Assay of the combined supernatant (1.6 L) indicated pentaene
(465 mg L-1).
m/z (-ve ES) 731.3 (M-H+), 713.5 (M-H3O+).
(3R,5R,9S,11S,15R,17S,19S,35S)-Octabenzoyloxy-13-oxo-
(16S,34S,36S)-trimethyl-(20E,22E,24E,26E,28E,30E,32E)-
octatriacontaheptaeno-(37S)-lactone 10
XAD16 resin (100 g) was washed with water (3 ¥ 50 mL) and
MeOH (3 ¥ 50 mL) and then added batchwise (3 ¥ 30 g) to the
supernatant, stirred (30 min), and filtered. Methanol (4 ¥ 50 mL)
was added to the beads batchwise with stirring to obtain the
pentaenes (630 mg), the beads were filtered and the volatiles were
then removed in vacuo. The resulting solid was dissolved in DMSO
(10 mL), methanol (800 mL) was then added followed by excess
ethereal diazomethane. Removal of volatiles in vacuo gave a dark
brown oil to which MeOH (50 mL) and then ether (100 mL) was
added. After discarding the upper ethereal layer, EtOAc (200 mL)
was added to the methanolic layer. The resulting solid was removed
and extracted again (200 mL; 20% MeOH–EtOAc) to give total
pentaene (550 mg in 400 mL). The volatiles were removed in vacuo,
the remaining solid redissolved in methanol (10 mL) and dry
loaded for purification by flash chromatography (15–20% MeOH–
EtOAc) to give a pentaene (150 mg) containing fraction which
was analysed by normal phase HPLC (MeOH–EtOAc) as several
different closely running pentaenes, all free of other polyenes. The
first major peak was partially purified by HPLC to give methyl
ester 12 (13 mg).
Crude 8-deoxy-16-descarboxyl-16-methyl-amphoteronolide B 3
(323 mg heptaene in 500 mg; 0.44 mmol) was suspended in DCM
(125 mL) and pyridine (25 mL) was added. Benzoyl chloride
(13 mL, 0.09 mmol) was added dropwise at RT, followed by
DMAP (10 mg, 0.08 mmol). The reaction was followed by TLC
(40% EtOAc–hexane). The reaction mixture went a clear yellow
solution after 30 min, was stirred for 3 days at RT then washed with
water (150 mL), sodium hydrogen carbonate (150 mL; saturated),
copper sulfate (150 mL; saturated) and water (150 mL). After
removal of volatiles in vacuo, the residue was dry loaded onto a
flash column which was flushed (hexane) and then heptaene eluted
(30% EtOAc–hexane) to give title compound (325 mg), which was
analysed by silica HPLC (20% EtOAc–hexane).
lmax (MeOH)/nm 407, 384 and 365.
umax (ATR)/cm-1 2928 (w), 1787 (w), 1713 (s), 1690 (s), 1602
(m), 1584 (m), 1266 (s) and 706 (s).
dH(500.1 MHz; CDCl3) 0.75 (3 H, d, J 7.2, 41-H), 0.90, 1.00 (3
H, J 6.5, 40-H), 1.03 (3 H, d, J 6.9, 39-H), 1.19 (3 H, d, J 6.5,
38-H), 1.30, 1.35, 1.60, 1.65 (2 H, m), 1.87 (1 H, ddd, J 15.0, 5.9
lmax (MeOH)/nm 350, 332, 317 and 303(sh) nm.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3758–3770 | 3767
©

View Article Online
1H NMR (400 MHz; d4-MeOH, partial) 1.02 (3 H, d, J 7.1,
35-H). 1.07 (3 H, d, J 6.6, H-36), 1.22 (3 H, d, J 6.3, 34-H), 2.06
(1 H, dd, J 12.4 and 5.0), 2.43 (2 H, m), 5.15 (ca t, J 5.1), 5.56 (1
H, dd, J 14.3 and 9.3), 5.94 (1 H, dd, J 14.9 and 8.4), 6.13–6.40
2.26 (2a-H), 2.28 (14a-H), 2.40 (34-H), 2.53 (1 H, dd, J 17.5 and
3.0, 6b-H), 2.88 (1 H, dd, J 17.5 and 9.2, 6a-H), 3.11 (1 H, dd, J
9.3 and 2.9, 4¢-H), 3.21 (1 H, dd, J 9.3 and 2.1, 35-H), 3.34 (5¢-H),
3.87 (1 H, d, J 2.2, 8-H), 4.00 (1 H, br d, J 3.2, 3¢-H), 4.22 (1 H,
dt, J 2.2 and 7.0, 9-H), 4.26 (11-H), 4.29 (2 ¥ 1 H, 2 ¥ m, 3-H
and 17-H), 4.37 (1 H, ddd, J 9.2, 3.0 and 3.0, 5-H), 4.39 (11-H),
4.41 (17-H), 4.42 (19-H), 4.59 (1 H, br s, 1¢-H), 5.36 (37-H), 5.38
(33-H), 6.02 (1 H, dd, J 15.2 and 8.8, 20-H), 6.12–6.43 (12 H, m,
21-H to 32-H).
=
(CH ).
13C NMR (100 MHz; d4-MeOH) 12.8 (q), 16.9 (q), 17.5 (q), 21.7
(t), 37.3 (t), 37.7, 40.5 (t), 40.6 (t), 41.8 (t), 42.4 (d), 43.3 (d), 43.5
(t), 46.4 (t), 47 (under solvent), 48.4 (d), 50.8 (d), 56.9 (d), 66.0
(d), 66.1 (d), 67.0 (d), 67.9 (d), 70.0 (d), 71.1 (d), 78.0 (d), 97.80
(lactol), 130.3, 130.6, 131.06 (2 C),, 131.9, 133.1, 133.37, 133.45,
136.2, 138.0, 171.2 (s), 173.3 (s).
Partial dC(125.8 MHz; d6-DMSO, 308 K) 12.6 (q), 17.46 (q),
18.3 (q), 18.9 (q), 40.2 (d), 40.4 (d), 40.6 (t), 42.3 (t), 43.0 (d), 43.9
(t), 44.6 (t), 47.0 (wk, t), 47.5 (t), 56.3 (d), 58.6 (d), 65.6 (d), 65.9
(d), 66.0 (d), 66.1 (d), 66.5 (2 C, 2 ¥ d), 68.4 (d), 69.6 (d), 70.3
(d), 70.5 (d), 73.2 (d), 75.8 (d), 78.1 (d), 79.8 (d), 97.0 (d), 97.7 (s),
129.7 (d), 131.5 (d), 132.3 (d), 132.5 (d), 132.5 (d), 132.6 (d), 132.6
(d), 133.9 (2 C, 2 ¥ d), 134.0 (d), 134.56 (d), 134.65 (d), 136.7 (d),
137.2 (d), 171.2 (s), 212.8 (s).
High Res FAB Found: 747.39300. C38H60O13Na requires
747.39319.
A second and third methylated pentaene fraction was isolated by
HPLC and examined by ESMS ‘Pentaene 2’: ESMS (-ve) 723 (M-
H) and ‘Pentaene 3’ ESMS (–ve) 723. FAB (+ve) 747 (M+Na)+.
NMR analysis of these crude fractions were consistent with related
pentaene analogues.
m/z (+ve ES) 938.5 (MH+).
The M57 mutant was then grown (1 L) in the presence of XAD16
resin. Extraction of sediment with methanol (1 L) gave pentaene
(assay indicated 200 mg). Partial removal of volatiles in vacuo gave
a precipitate which when examined by LCMS showed that early
fractions contained glycosylated pentaenes most later fractions
were isomeric with 11 with little evidence of AmphL catalysed
hydroxylation, or compounds in which the methyl group had not
been oxidised to a carboxyl by AmphNM.
m/z (-ve ES) 936.5.
m/z (QTOF) Found 938.4752 (MH+) C47H72NO18 requires
938.4749.
DNM KR12 mutant
Mycelia and XAD16 resin from the DNM KR12 mutant (1 L;
grown in the absence of thiostrepton), was extracted with methanol
(1 L; o/n; RT). UV assay of the extract indicated heptaene (9 mg)
with negligible levels of tetraene present. Partial removal of solvent
in vacuo (to less than 20 mL) gave an oil that was removed from
remaining solvent. The oil was resuspended in MeOH (50 mL) and
assayed to contain heptaene (2.5 mg). Analysis by LCMS gave
15-deoxy-16-descarboxyl-16-methyl-15-oxo-amphotericin B (16)
[m/z (+ve ES) 892.7 (MH+) and 874.6 (M-water+H+), m/z (–ve
ES) 891.7 (M-H+) and 936.7 (M+HCOO-)] and 16-descarboxyl-
8,15-dideoxy-16-methyl-15-oxo-amphotericin B (20) [m/z (+ve
MS) 876.7 (MH+) and m/z (–ve ES) 920.7 (M+HCOO-)].
7-Oxo-amphotericin B (9)
Sedimented mycelia and XAD16 resin from the KR16 mutant (7 L
broth) was soaked (3 ¥ 18 h) in methanol (3 ¥ 3 L) with occasional
agitation. UV assay of the combined supernatants indicated that
heptaenes (850 mg) and tetraenes (850 mg) were both present.
Partial removal of volatiles in vacuo (to ca 300 mL) followed by
storage (4 ◦C; 18 h) gave a yellow precipitate that was collected by
centrifugation. Partial dissolution in methanol (1 L) with vigorous
agitation and sonication, removal of the insoluble debris, and
then removal of volatiles gave heptaene (700 mg) plus tetraene
(400 mg) as a waxy yellow solid containing heptaene (1–20%
w/w). The solid was vigorously washed with water (10 ¥ 300 mL).
The remaining solid was partially redissolved in methanol (1 L),
insoluble debris was removed by centrifugation and volatiles then
removed in vacuo to give a yellow solid (600 mg heptaene, 50 mg
tetraene, 40–80% pure by weight). The solid was vigorously shaken
and sonicated with chloroform (500 mL), the resulting suspension
filtered through filter paper to give a yellow precipitate and orange
filtrate. The precipitate was washed with chloroform (500 mL)
to give a yellow solid (600 mg heptaene, 20 mg tetraene, 65–
85% purity by UV assay and NMR analysis), which was further
purified by RP-HPLC (water–methanol+0.1% v/v formic acid)
for spectroscopic characterisation.
Construction of KR10 mutants
The 3¢ region of the KR10 coding sequence was amplified by PCR
with primers KR10F2 and KR10R2. The amplified DNA was di-
gested with EcoRI and HindIII and cloned into plasmid pUC118
to give pUC-KR103¢. The 5¢ region of the coding sequence was
amplified with KR10F1 and KR10R1, digested with HindIII
and inserted into the HindIII site of pUC-KR103¢. A plasmid
clone containing the second insert in the correct orientation
was named pUC-KR10. In this construct the sequence AGC-
GCC-TAC encoding Ser3088-Ala3089-Tyr3090 of AmphI was replaced
by GCA-AGC-TTC specifying Ala-Ser-Phe. Substitution of the
active site tyrosine (Tyr3090) with phenylalanine abolishes KR
activity without disrupting PKS protein structure or diminishing
product yield.17 The mutagenesis also introduces a HindIII site
that can be used as a marker. The modified coding sequence
was excised with BglII and PstI and cloned into phage KC-
UCD1 to give KC-KR10. This recombinant phage was used to
replace the chromosomal KR10 coding sequence by homologous
recombination as described previously.17,29 Mutants were first
identified by PCR with oligonucleotides KR10CF and KR10CR1
that only amplify the modified sequence. To confirm that gene
lmax (MeOH)/nm 406, 383, 364.
dH(500.1 MHz; d4-MeOH) 1.03 (3 H, d, J 7.2, 39-H), 1.15 (3 H,
d, J 6.4, 40-H), 1.22 (3 H, d, J 6.3, 38-H), 1.29 (3 H, d, J 5.7, 6¢-H),
1.30 (18b-H), 1.39 (1 H, dt, J 14.1 and 3.0, 4b-H), 1.52 (1 H, dt, J
14.1 and 9.8, 4a-H), 1.62–1.68 (3 H, m, 12b-H and 10ab-H), 1.69
(14b-H), 1.75 (1 H, dd, J 14.7 and 10.9, 12a-H), 1.82 (1 H, ca ddq,
J 1.6, 1.6 and 7.3, 36-H), 2.00 (1 H, dd, J 12.4 and 4.8, 18a-H),
2.04 (1 H, t, J 10.5 16-H), 2.19 (1 H, dd, J 17.1 and 3.2, 2b-H),
3768 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
replacement had occurred, the KR10 coding region was amplified
with primers KR10CF and KR10CR2 and tested for the presence
of a HindIII site (Fig. 8).
The sequences of the oligonucleotide primers (5¢ – 3¢) were:
KR10F1, TTTTAAGCTTCTGCAGACGAGCGTTACGAC-
CACAACAC; KR10R1, GATCAAGCTTGCCTGGTCGC-
CGCTGCCCCAGACGC; KR10F2, GATCAAGCTTCGGC-
GCTGCCAACGCCTACCTCGAC; KR10R2, AAAAGAATT-
CGGATCCAGCAGTTCGAGGATCTCGTCGTG; KR10CF,
CACCGAGAACGCCGACAACAC; KR10CR1, GCGTTGG-
in MeOH (50 mL), and then an equal volume of EtOAc was
slowly added and any solid removed by centrifugation. Volatiles
were removed from the supernatant in vacuo, the resultant solid
was sonicated and shaken vigorously in methanolic ethyl acetate
(10% v/v), any debris was removed, and the volatiles removed in
vacuo from the supernatant to give a crude solid. Purification by
flash chromatography (70% hexane in EtOAc, gradient to 100%
EtOAc) gave tetraene analogue 24 (>10 mg L-1 based on NMR
purity and signal to noise) eluted with 70% EtOAc–hexane and a
trace amount of the corresponding heptaene analogue (eluted at
90% EtOAc–hexane), lmax (heptaene, MeOH)/nm 440) coeluting
with aromatic impurities.
CAGCGCCGAAGCTT;
GCAGTTCGAC.
KR10CR2,
GAGGTGCTTGC-
The following oligonucleotides were used to detect ACP coding
sequences in strain KR10-1 (5¢ – 3¢): ACP3F, TTCGTGCTCTTC
TCGTCCGTC; ACP3R, CCGAAGAAGTCGGCGTCGA-
AG; ACP4F, ACCGATGAGGCCGCTCTCGT; ACP4R,
lmax (MeOH)/nm 399, 386 and 289.
dH(500.1 MHz; d4-MeOH) 0.84 (3 H, d, J 7.1, 13-Me), 1.01
(3 H, d, J 6.7, 15-Me), 1.13 (3 H, d, J 6.3, 19-H), 1.73 (17-H), 1.91
(3 H, br s, 3-Me), 2.20 (2 H, m, 9 or 10-H), 2.23 (2 H, m, 9 or
10-H), 2.41 (1 H, ddq, J 7.2, 3.4 and 7.2, 15-H), 3.39 (1 H, dd, J
8.9 and 3.4, 16-H), 4.00 (1 H, dq, J 6.2 and 6.2, 18-H), 5.61 (1 H,
ddd, J 14, 7.2 and 7.2, 8 or 11-H), 5.67 (14-H), 5.84 (11 or 8-H),
6.04, 6.07 (2 ¥ 1 H), 6.17 (2 ¥ 1 H, 2 ¥ ca dd, J 15 and 12), 6.26
(1 H, dd, J 14.8 and 10.9), 6.35, 6.38, 6.40, 6.56 (1 H, dd, J 14.6
and 10.8), 7.08 (1 H, dd, J 15.0 and 11.1).
GATCACGATCGGGTCGTCGT;
ACP5F
AACGCCTT-
CCTGGACGCACT; ACP5R, GAGCAGTCGCCACAGG-
TCCT; ACP6F, CAACGCGGGTCAGGCCAACTA; ACP6R,
GAAGCTCGTCCAGGATGAAGT; ACP7F, TCTGGAC-
CTGGACGCGTTCAT; ACP7R, GGCTCCGAAGAGTTC-
GTCGTG.
dC(125.8 MHz; d4-MeOH) 8.7 (q), 11.6 (q), 13.2 (q), 19.1 (q),
33.5 (t), 34.0 (t), 40.5 (d), 43.6 (d), 70.6 (d), 79.3 (d), 100.6 (s),
102.1 (d), 122.5 (d), 131.0 (d), 131.8 (2 C, 2 ¥ d), 132.3 (d), 132.5
(d), 132.6 (d), 136.3(d), 137.5 (2 C, 2 ¥ d), 137.7 (d), 139.8 (d),
158.3 (s), 167.7 (s), 168.0 (s).
6-[(1E,3E,5E,7E,9E,11S,12S,13S,14S)-12,14-Dihydroxy-11,13-
dimethyl-1,3,5,7,9-pentadecapenten-1-yl]-4-hydroxy-3-methyl-2-
pyrone (21)
Sedimented mycelia and XAD16 resin (KR10-1 broth; 7 L) was
extracted with methanol (3 ¥ 10 L). Partial removal of volatiles in
vacuo followed by cooling (4 ◦C, 18 h) gave an orange precipitate
which was collected by centrifugation, washed (water) and dried in
vacuo (1.0 g). Purification by flash chromatography (60% EtOAc–
hexane) gave 21 (100 mg).
m/z (–ve ESMS) 439.
Acknowledgements
This work was supported by BBSRC grant BB/D017270/1 and a
summer studentship from the Nuffield Foundation (URB/35829)
to KP. NK received a PhD studentship from the Irish Higher
Education Authority Programme for Research in Third Level In-
stitutions. PC received the financial support of Science Foundation
Ireland under grant number 09/RFP/GEN2132.
m
max/cm-1 422, 400(max), 380(sh) and 311.
dH(500.1 MHz; d4-MeOH) 0.85 (3 H, d, J 6.9 17¢-C) 1.04 (3 H,
d, J 6.8 16¢-Me), 1.14 (3 H, d, J 6.3, 15¢-H), 1.74 (13¢-H), 1.90 (3
H, s, 7-C), 2.48 (1 H, dq, J 3.2 and 6.8, 11¢-H), 3.41 (1 H, dd, J
8.9 and 3.4, 12¢-H), 4.06 (1 H, dq, J 6.3 and 6.3, 14¢-H), 5.89 (1
H, dd, J, 15.2 and 7.8, 10¢-H), 6.07 (1 H, s, 5-H), 6.16, 6.18 (9¢-H),
6.27, 6.34, 6.34, 6.38, 6.43, 6.59 (1 H, dd, J 14.6 and 11.0), 7.08 (1
H, dd, J 15.0 and 11.3).
References
dC(125.8 MHz; d4-MeOH) 7.3 (7-C), 10.2 (17¢-C), 11.7 (16¢-C),
17.7 (15¢-C), 39.5 (11¢-C), 42.3 (13¢-C), 69.1 (14¢-C), 77.7 (12¢-C),
99.2 (3-C), 100.6 (5-C), 121.2 (d), 129.9 (9¢-C), 130.7 (d), 130.9 (d),
131.7 (d), 134.9 (d), 135.2 (d), 136.2 (d), 138.4 (d), 140.2 (10¢-C),
156.9 (s), 166.2 (s), 166.6 (s).
1 B. P. Matthew and M. Nath, ChemMedChem, 2009, 4, 310; O. A.
Cornely, J.-J. Vehreschild and A. J. Ullmann, Curr. Opin. Infect. Dis.,
2006, 19, 565.
2 D. Ellis, J. Antimicrob. Chemother., 2002, 49(Supp S1), 7.
3 F. Chappuis, S. Sundar, A. Hailu, H. Ghalib, S. Rijal, R. W. Peeling,
J. Alvar and M. Boelaert, Nat. Rev. Microbiol., 2007, 5, S7; H. W.
Murray, Am. J. Trop. Med. Hyg., 2004, 71, 787; S. Sundar, M. Rai,
J. Chakravarty, D. Agarwal, N. Agrawal, M. Vaillant, P. Olliaro and
H. W. Murray, Clin. Infect. Dis., 2008, 47, 1000.
m/z (-ve ES) 385.1 (M-H+), 771.3 (M2-H+), 1157.3 (M3-H+).
m/z (+ve ES) 387.1, 773.0 (M2H)+, 1160.1 (M3H)+
m/z (FAB) 387.21637 (MH+). C23H31O5 requires 387.21628.
4 L. Soler, P. Caffrey and H. E. M. McMahon, Biochim. Biophys. Acta,
Gen. Subj., 2008, 1780, 1162; J. Jones, B. R. Kosloff, E. N. Benveniste,
G. M. Shaw and O. Kutsch, Virology, 2005, 331, 106.
5 F. C. Odds, A. J. P. Brown and N. A. R. Gow, Trends Microbiol.,
2003, 11, 272; H. I. Al-Mohammed, M. L. Chance and P. A. Bates,
Antimicrob. Agents Chemother., 2005, 49, 3274.
6 M. Baran, E. Borowski and J. Mazerski, Biophys. Chem., 2009, 141,
162; N. Matsumori, Y. Sawada and M. Murata, J. Am. Chem. Soc.,
2005, 127, 10667; B. De Kruiff and R. A. Demel, Biochim. Biophys.
Acta, Biomembr., 1974, 339, 57.
7 G. Deray, J. Antimicrobial chemotherapy, 2002, 49(Supp S1), 37.
8 R. Chopra, J. Antimicrobial Chemotherapy, 2002, 49(Supp S1), 43; J.
Adler-Moore and R. T. Proffitt, J. Antimicrobial Chemotherapy, 2002,
49(Supp S1), 21; P. J. Cagnoni, J. Antimicrobial Chemotherapy, 2002,
6-[(1E,3E,5E,7E,11E,13E,15S,16S,17S,18S)-16,18-Dihydroxy-
15,17-dimethyl-1,3,5,9,11,13-heptadecahexen-1-yl]-4-hydroxy-3-
methyl-2-pyrone (24)
Sedimented mycelia and XAD16 resin from KR10-2 culture (2 L)
was extracted with methanol (2 ¥ 4 L) for 3 h. The volume of the
supernatant was reduced in vacuo (to 100 mL) to give a thick dark
◦
orange suspension, which was cooled to 4 C to give an orange
precipitate (ca 500 mg). The precipitate was partially dissolved
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3758–3770 | 3769
©

View Article Online
49(Supp S1), 81; B. Dupont, J. Antimicrobial Chemotherapy, 2002,
49(Supp S1), 31.
23 J. D. Cleary, S. W. Chapman, E. Swiatlo and R. Kramer, J. Antimicrob.
Chemother., 2007, 60, 1331; P. M. Kelly, J. Chromatogr., A, 1988, 437,
221.
9 J. Szlinder-Richert, J. Mazerski, B. Cybulska, J. Grzybowska and
E. Borowski, Biochim. Biophys. Acta, Gen. Subj., 2001, 1528, 15;
J. Szlinder-Richert, B. Cybulska, J. Grzybowska, J. Bolard and E.
Borowski, Farmaco, 2004, 59, 289; B. Cybulska, I. Gadomska, J.
Mazerski, J. Grzybowska, E. Borowski, M. Cheron and J. Bolard,
Acta Biochim. Polonica, 2000, 47, 121; J. Szlinder-Richert, B. Cybulska,
J. Grzybowska, E. Borowski and R. Prasad, Acta Biochim. Polonica,
2000, 47, 133.
24 O. Volokhan, H. Sletta, T. E. Ellingsen and S. B. Zotchev, Appl. Environ.
Microbiol., 2006, 72, 2514.
˜
25 S. Omura, and H. Tanaka, in Macrolide Antibiotics, Chemistry, Biology,
˜
and Practice, Ed. S. Omura, Academic Press, Orlano, 1984, Chapter 9,
pp 351–404.
26 D. S. Palacios, T. M. Anderson and M. D. Burke, J. Am. Chem. Soc.,
2007, 129(13), 804.
10 A. M. Hill, and J. Staunton, ‘Modular Polyketide Synthases’, Compre-
hensive Natural Products II Chemistry and Biology, L. Mander, H.-W.
Lui, ed.; Elsevier: Oxford, 2010; volume 1, 386-452; A. M. Hill, Nat.
Prod. Rep., 2006, 23, 256; J. Staunton and K. J. Weissman, Nat. Prod.
Rep., 2001, 18, 380; B. J. Rawlings, Nat. Prod. Rep., 2001, 18, 190; B. J.
Rawlings, Nat. Prod. Rep., 2001, 18, 231.
27 T. Brautaset, H. Sletta, A. Nedal, S. F. Borgos, K. F. Degnes, I. Bakke,
O. Volokhan, O. N. Sekurova, I. D. Treshalin, E. P. Mirchink, A. Dikiy,
T. E. Ellingsen and S. B. Zotchev, Chem. Biol., 2008, 15, 1198.
28 J. Pawlak, P. Sowin´ski and E. Borowski, J. Antibiot., 1995, 48, 1034.
29 E. Hutchinson, B. Murphy, T. Dunne, C. Breen, B. Rawlings and P.
Caffrey, Chem. Biol., 2010, 17, 174.
11 P. Caffrey, J. F. Aparicio, F. Malpartida and S. B. Zotchev, Curr. Top.
Med. Chem., 2008, 8, 639.
30 A. Czerwin´ski, W. A. Ko¨nig, P. Sowin´ski, L. Falkowski, J. Mazerski
and E. Borowski, J. Antibiot., 1990, 43, 1098.
12 C. M. McNamara, S. Box, J. M. Crawforth, B. S. Hickman, T. J.
Norwood and B. J. Rawlings, J. Chem. Soc., Perkin Trans. 1, 1998,
83.
31 D. T. MacPherson, D. F. Corbett, B. C. Costello, M. J. Driver, A. R.
Greenlees, W. S. MacLachlan, C. T. Shanks, and A. W. Taylor, Roy.
Soc. Chem., Special Publ., ‘Recent Adv, in the Chem. of Anti-infective
Agents’, 1993, Chapter 14, 205.
32 T. Brautaset, P. Bruheim, H. Sletta, L. Hagen, T. E. Ellingsen, A. R.
Strøm, S. Valla and S. B. Zotchev, Chem. Biol., 2002, 9, 367.
33 P. Bruheim, S. E. F. Borgos, P. Tsan, H. Sletta, T. E. Ellingsen, J.-M.
Lancelin and S. B. Zotchev, Antimicrob. Agents Chemother., 2004, 48,
4120.
13 P. Caffrey, S. Lynch, E. Flood, S. Finnan and M. Oliynyk, Chem. Biol.,
2001, 8, 713.
14 B. Byrne, M. Carmody, E. Gibson, B. Rawlings and P. Caffrey, Chem.
Biol., 2003, 10, 1215.
15 M. Carmody, B. Murphy, B. Byrne, P. Power, D. Rai, B. Rawlings and
P. Caffrey, J. Biol. Chem., 2005, 280, 34420.
16 M. Carmody, B. Byrne, B. Murphy, C. Breen, S. Lynch, E. Flood, S.
Finnan and P. Caffrey, Gene, 2004, 343, 107.
17 P. Power, T. Dunne, B. Murphy, L. N. Lochlainn, D. Rai, C. Borissow,
B. Rawlings and P. Caffrey, Chem. Biol., 2008, 15, 78.
34 E. Borowski, J. Golik, J. Zielin´ski, L. Falkowski, P. Kolodziecjczyk, J.
Pawlak and Y. Shenin, J. Antibiot., 1978, 31, 117.
35 J. Kennedy, K. Auclair, S. G. Kendrew, C. Park, J. C. Vederas and C. R.
Hutchinson, Science, 1999, 284, 1368.
18 S. Donadio, M. J. Staver, J. B. McAlpine, S. J. Swanson and L. Katz,
Science, 1991, 252, 675.
19 R. Reid, M. Piagentini, E. Rodriguez, G. Ashley, N. Viswanathan, J.
Carney, D. V. Santi, C. R. Hutchinson and R. McDaniel, Biochemistry,
2003, 42, 72.
20 K. C. Nicolaou, T. K. Chakraborty, Y. Ogawa, R. A. Daines, N. S.
Simpkins and G. T. Furst, J. Am. Chem. Soc., 1988, 110, 4660; K. C.
Nicolaou, R. A. Daines, J. Uenishi, W. S. Li, D. P. Papahatjis and T. K.
Chakraborty, J. Am. Chem. Soc., 1988, 110, 4762; K. C. Nicolaou,
R. A. Daines, T. K. Chakraborty and Y. Ogawa, J. Am. Chem. Soc.,
1988, 110, 4685; K. C. Nicolaou, R. A. Daines, Y. Ogawa and T. K.
Chakraborty, J. Am. Chem. Soc., 1988, 110, 4696.
21 M. Kawabata, M. Onda and T. Mita, J. Biochem., 2001, 129, 725.
22 D. R. Worthen, M. Jay and P. M. Bummer, Drug Dev. Ind. Pharm.,
2001, 27, 277; M. T. Grijalba, M. Che´ron, E. Borowski, J. Bolard and
S. Schreier, Biochim. Biophys. Acta, Gen. Subj., 2006, 1760, 973.
36 S. M. Ma, J. W. H. Li, J. W. Choi, H. Zhou, K. K. M. Lee, V. A.
Moorthie, X. Xie, J. T. Kealey, N. A. Da Silva, J. C. Vederas and Y.
Tang, Science, 2009, 326, 589.
37 T.-W. Yu, Y. Shen, Y. Doi-Katayama, L. Tang, C. Park, B. S. Moore,
C. R. Hutchinson and H. G. Floss, Proc. Natl. Acad. Sci. U. S. A.,
1999, 96, 9051; A. Stratmann, C. Toupet, W. Schilling, R. Traber, L.
Oberer and T. Schupp, Microbiology, 1999, 145, 3365.
38 J. Moldenhauer, D. C. G. Go¨tz, C. R. Albert, S. K. Bischof, K.
Schneider, R. D. Su¨ssmuth, M. Engeser, H. Gross, G. Bringmann and
J. Piel, Angew. Chem. Int. Ed., 2010, 49, 1465.
39 B. R. Clark, R. J. Capon, E. Lacey, S. Tennant and J. H. Gill, Org. Lett.,
2006, 8, 701; B. R. Clark and C. D. Murphy, Org. Biomol. Chem., 2009,
7, 111; Y. Yamagishi, M. Matsuoka, A. Odagawa, S. Kato, K. Shindo
and J. Mochizuka, J. Antibiotics, 1993, 46, 884.
40 A. Starcevic, M. Jaspars, J. Cullum, D. Hranueli and P. F. Long,
ChemBioChem, 2007, 8, 28.
3770 | Org. Biomol. Chem., 2010, 8, 3758–3770
This journal is
The Royal Society of Chemistry 2010
©