Investigation of the early stages in soraphen A biosynthesis

Article abstract of DOI:10.1039/b303542p

The unusual benzoate starter unit in soraphen A derives from phenylalanine via cinnamate in a β-oxidative (plant-like) pathway; 3-phenyl-3-hydroxypropanoate incorporates directly into soraphen by loading onto module 2 of the PKS and indirectly from the β-

Full text of DOI:10.1039/b303542p

Investigation of the early stages in soraphen A biosynthesis
Alison M. Hill†,* Betty L. Thompson, Jonathan P. Harris and Roger Segret
Department of Chemistry, King’s College London, Strand, London, UK WC2R 2LS.
E-mail: a.m.hill@ex.ac.uk; Fax: +44 (0)1392 263434; Tel: +44 (0)1392 263467
Received (in Cambridge, UK) 28th March 2003, Accepted 15th April 2003
First published as an Advance Article on the web 16th May 2003
The unusual benzoate starter unit in soraphen A derives
from phenylalanine via cinnamate in a b-oxidative (plant-
like) pathway; 3-phenyl-3-hydroxypropanoate incorporates
directly into soraphen by loading onto module 2 of the PKS
and indirectly from the b-oxidative pathway to generate
benzoyl CoA.
sum.§ For all of these compounds we observed incorporation of
13C label into position 17 of soraphen (Table 1). We observed
that the incorporation of benzoate, while significant, was much
lower than for cinnamate. The most likely explanation for this is
that oxidative cleavage of cinnamate gives benzoyl CoA
directly and that the low levels of benzoate incorporation into
soraphen A arises from low levels of ligase activity and/or from
toxicity (benzoate is a known antibacterial agent). A benzoyl
CoA ligase has been reported in enterocin biosynthesis, in
addition to the b-oxidative pathway from phenylalanine, which
provides two possible routes to the starter unit.2b No specific
benzoyl CoA ligase has been reported in S. cellulosum to date1
and consequently transformation of benzoate to benzoyl CoA
would require sequestering a ligase from primary metabolism.
To overcome this problem, benzoyl N-acetyl cysteamine (NAC)
thioester was prepared¶ (these thioesters have been shown to
facilitate incorporation of precursors onto polyketide syn-
thases).7 Unfortunately, we found that NAC thioesters were
incompatible with the adsorber resin used in the production
media and growing the organism in the absence of resin resulted
Soraphen A 1 is a novel 18-membered macrolide characterised
by an unsubstituted phenyl group and by a cyclic hemiketal
moiety. It is assembled by a type I polyketide synthase (PKS),
the genes for which have been cloned and sequenced.1 The first
module in the sorPKS contains both the loading domain and the
first extension module. This ‘starter module’ has the function of
loading the benzoyl CoA starter unit and a malonyl CoA
extender unit, condensing the two units together and reducing
the b-keto moiety to produce the diketide intermediate,
3-hydroxy-3-phenylpropanoate 2, bound to the PKS as a
thioester. Benzoate is a rare starter unit in polyketide bio-
synthesis and has been found to date in only one other bacterial
polyketide biosynthetic pathway, the type II enterocin PKS
from the marine bacterium Streptomyces maritimus.2 The
enterocin benzoate starter unit is derived from phenylalanine via
cinnamate. Five activities are required for this transformation
which generates benzoyl CoA using a b-oxidative pathway; the
putative soraphen diketide 2, but not benzoic acid, is an
intermediate in this pathway (Fig. 1). In anaerobic bacteria,
phenylalanine is converted to benzoate via phenylpyruvate;
cinnamate is not an intermediate.3 We had previously observed
labelling of the phenyl ring from [1,3-13C2]-glycerol consistent
with the shikimate pathway4 and Höfle and Reichenbach
reported that phenylalanine, but not cinnamate or benzoate,
labelled the phenyl ring of 1.5 However, a hybrid soraphen/
erythromycin bimodular PKS was able to produce a phenyl
substituted lactone product only when benzoyl CoA was
provided within the cell as a specific precursor.6 To investigate
the origin of the starter unit and diketide intermediate in
soraphen A biosynthesis, 13C- and 2H-labelled precursors were
administered to Sorangium cellulosum. The results of these
labelling experiments are reported herein.
in a reduction of the soraphen titre from ca. 30 to 3 mg litre21
.
Consequently, we observed only 5 fold enrichment of [1-13C]-
benzoyl NAC thioester into 1. In conclusion, we have shown
that the benzoate starter unit in soraphen A biosynthesis is
derived from the deamination of phenylalanine to give cinna-
mate followed by b-oxidative cleavage to furnish benzoyl CoA
in a plant like pathway (Fig. 1). This is the same biosynthetic
route observed for the type II polyketide metabolite enter-
ocin.2
The benzoyl CoA starter unit is then extended and processed
by the starter module of the soraphen PKS to give the diketide,
3-hydroxy-3-phenylpropanoate 2 bound to the PKS as a
thioester; note that 2 is also an intermediate in the conversion of
cinnamate to benzoyl CoA (Fig. 1). We decided to label the
2
carbinol site with both 13C and H so that we could delineate
between intact incorporation of 2 (resulting from the substrate
being accepted by the PKS) and incorporation via degradation
of 2 to benzoyl CoA; while it is possible for the 13C label to be
incorporated into soraphen by either pathway, the deuterium
label will be incorporated into 1 only if the diketide precursor
loads onto the PKS directly. Despite the problems encountered
Initial experiments focused on the b-oxidative pathway to
produce benzoate and 3-[13C]-phenylalanine, [2,3-13C2]-cinna-
mate‡ and 1-13C-benzoic acid were administered to S. cellulo-
Table 1 Incorporation of precursors into C-17 of soraphen A 1
13C Enrich-
ment at
C-17 of 1a
1H : 2Hb
(ppm)
c
Compound
DdC
[3-13C]-Phenylalanine
23
65
10
5
2.5
3.3
3
[2,3-13C2]-Cinnamic Acid
[1-13C]-Benzoic Acid
[1-13C]-Bz NAC thioester
rac-[3-13C,2H]-Diketide 2a
rac-[3-13C,2H]-Diketide 2b
rac-[3-13C,2H]-Diketide 2c
2.3 : 0.2
20.34
a Expressed as number of fold increase in the height of the carbon-13
resonance relative to unlabelled 1, with correction by normalising each
signal to C-3A,5A in both labelled and unlabelled samples. b Ratio of carbon
peak intensities for a-isotope shifted signal. c a-Isotope shift due to 2H.
Fig. 1 Early stages in soraphen 1 biosynthesis.
† Current address: School of Chemistry, University of Exeter, Stocker
Road, Exeter, UK EX4 4QD.
1358
CHEM. COMMUN., 2003, 1358–1359
This journal is © The Royal Society of Chemistry 2003

with the use of NAC thioesters described above, it was thought
that their ability to facilitate incorporation onto the PKS coupled
with lower substrate degradation would outweigh the lower
soraphen titre. 3-(R,S)-[3-13C,2H]-3-hydroxy-3-phenylpropa-
noate NAC thioester 2c was prepared (Scheme 1b) and
administered to the producing organism. As expected, the
soraphen titre was low and 0.5 mg of purified 1 was isolated
from a 500 mL culture. Analysis of this sample with 13C NMR
resulted in a spectrum with only two signals: We were able to
detect 13C enrichment at C-17 of soraphen A (ca. 3 fold) in
addition to the chloroform peak. No upfield triplet correspond-
ing to intact incorporation of the 13C–2H bond was detected. In
the 13C proton and deuterium decoupled NMR spectrum, no
additional peak was detected upfield from the C-17 signal. The
deuterium NMR spectrum was also inconclusive from this
sample. Hence, we decided to repeat the experiment but this
time presenting the diketide to the producing organism as the
corresponding free acid 2b and ethyl ester 2a. This would
enable us to use the adsorber resin during the growth of the
micro-organism increasing the titre back to 20–30 mg litre21
.
However, gain in soraphen production would be counteracted
by degradation of the substrate. We chose a literature proce-
dure2a to synthesise the target molecule, 3-(R,S)-[3-13C,2H]-
3-hydroxy-3-phenylpropanoate ethyl ester 2a, from 1-[13C,
2H]-benzaldehyde∑ and ethyl bromoacetate; saponification of
the ethyl ester gave the corresponding acid 2b (Scheme 1a). As
expected, these feeding experiments resulted in the isolation of
between 10 and 15 mg of soraphen from a 500 mL culture.
Fig. 2 Intact incorporation of 3-(R,S)-[3-13C,2H]-diketide ethyl ester 2a into
soraphen A 1.
Intact incorporation of the 13C–2H bond was observed in the 13
C
proton decoupled NMR spectrum of the soraphen sample
isolated from the 3-(R,S)-[3-13C,2H]-diketide ethyl ester 2a
feeding experiment (Fig. 2): a triplet (J = 22 Hz) a-shifted 0.3
ppm upfield was observed which collapsed to a singlet when the
deuterium atom was decoupled. We also observed significant
degradation of the substrate as a 2.3 fold enrichment of the C-17
signal was observed. For the soraphen sample isolated from the
corresponding acid 2b feed, only 13C enrichment at C-17 was
observed (Table 1). We could see no triplet upfield from this
signal, and upon simultaneous deuterium decoupling, no new
signal upfield from C-17 was detected.
Sorangium cellulosum strains M15 and SJ3 were provided by
Syngenta, Research Triangle Park, North Carolina, USA. We
thank Dr. T. Schupp (Syngenta, Switzerland) for providing
Probion S, Mrs. J. Hawkes and Mr J. Cobb (KCL) and Dr. H.
Toms (QMW) for NMR support. 13C NMR (1H and 2H
decoupled) were obtained using the Bruker AMX 600 Uni-
versity of London Intercollegiate Research Service for NMR at
Queen Mary Westfield College. This work was financially
supported by funding from King’s College London, EPSRC,
and The Royal Society.
The intact incorporation of the 13C–2H bond of the racemic
diketide ethyl ester 2a into soraphen A demonstrates that it is
possible to by-pass the starter module and load an advanced
precursor directly onto the second module of the soraphen PKS.
Incorporation of the 13C label only from all three diketide
precursors demonstrates that the major pathway of incorpora-
tion is via the b-oxidative pathway to generate the benzoyl CoA
starter unit as 3-hydroxy-3-phenylpropanoyl CoA is an inter-
mediate between cinnamate and benzoyl CoA; presumably, the
anabolic and catabolic pathways are differentiated by the
stereochemistry of the carbinol site.
In summary, feeding experiments with labelled intermediates
have shown that the starter unit in soraphen A biosynthesis is
benzoyl CoA which is biosynthesised by the plant-like pathway
from phenylalanine. Intact incorporation of the racemic ad-
vanced intermediate, 3-hydroxy-3-phenylpropanoate, has dem-
onstrated that it is possible to by-pass the ‘starter module’ of the
soraphen PKS.
Notes and references
‡ [2,3-13C2]-Cinnamic acid was prepared from a Horner–Emmons reaction
between 13C-benzaldehyde and [1-13C]-triethylphosphonoacetate followed
by saponification using sodium hydroxide.
§ Cultures of a kanamycin and streptomycin resistant strain of S. cellulosum
(mutant over-producing strain SJ3) were used for the diketide 2a and 2b
experiments and the kanamycin resistant strain of S. cellulosum (wild-type
strain M15) was used for the starter unit experiments following the
established protocol.4 [3-13C]-Phenylalanine (250 mg), [2,3-13C2]-cinnamic
acid (100 mg), and 1-[13C]-benzoic acid (1 g) were added as sterile aqueous
solutions with a few drops of NaOH to aid dissolution. The diketides 2a
(100 mg) and 2b (60 mg) were dissolved in water with a minimum amount
of ethanol to aid dissolution. The [1-13C]-benzoyl NAC thioester (500 mg)
and diketide 2c (208 mg) were added as 50% aqueous ethanolic solutions.
¶ [1-13C]-benzoyl NAC thioester was prepared from carboxy-13C-benzoic
acid and N-acetylcysteamine with DCC and DMAP.
∑ [1-13C,2H]-benzaldehyde (99% D) was prepared by reduction of [1-13C]-
benzaldehyde with LiAlD4 followed by oxidation with MnO2.
1 J. Ligon, S. Hill, J. Beck, R. Zirkle, I. Molnár, J. Zawodny, S. Moey and
T. Schupp, Gene, 2002, 285, 257–267.
2 (a) C. Hertweck and B. S. Moore, Tetrahedron, 2000, 56, 9115–9120; (b)
L. Xiang and B. S. Moore, J. Bacteriol., 2003, 185, 399–404.
3 S. Schneider, M. E. Mohamed and G. Fuchs, Arch. Microbiol., 1997, 168,
310–320.
4 A. M. Hill, J. P. Harris and A. P. Siskos, Chem. Commun., 1998,
2361–2362.
5 G. Höfle and H. Reichenbach in Sekundärmetabolismus bei Mikroorga-
nismen, ed. W. Kuhn and H.-P. Fiedler, Attempto Verlag, Tübingen,
Germany, 1995, pp. 61–78.
6 C. Wilkinson, E. Frost, J. Staunton and P. F. Leadlay, Chem. Biol., 2001,
8, 1197–1208.
7 (a) C. R. Hutchinson, S. Yue, J. S. Duncan and Y. Yamamoto, J. Am.
Chem. Soc., 1987, 109, 1253–1255; (b) D. E. Cane and C. C. Yang, J. Am.
Chem. Soc., 1987, 109, 1255–1257.
Scheme 1 Synthesis of 3-(R,S)-[3-13C,2H]-diketides 2a–c. (i) LiAlD4; (ii)
MnO4; (iii) BrCH2CO2Et, CrCl2, LiI; (iv) 1M NaOH; (v) LDA; (vi)
Ph13CDO; (vii) t-BuPh2SiCl, imidazole; (viii) TFA; (ix) N-acetyl cystea-
mine (NAC), DCC, DMAP; (x) HF–py/THF/py.
CHEM. COMMUN., 2003, 1358–1359
1359