A method for trapping intermediates of polyketide biosynthesis with a nonhydrolyzable malonyl-coenzyme A analogue

Article abstract of DOI:10.1002/anie.200501670

(Chemical Equation Presented) Trapped in the middle: To investigate the mechanistic details of polyketide biosynthesis, intermediates are trapped by a nonhydrolyzable malonyl-coenzyme A analogue that reacts with a growing polyketide chain formed by a stilbene synthase. As the reverse transesterification reaction is not possible, the polyketide intermediates accumulate attached to the analogue (see scheme; Enz = enzyme, CoA = coenzyme A).

Full text of DOI:10.1002/anie.200501670

Angewandte
Chemie
Enzyme Reaction Mechanisms
DOI: 10.1002/anie.200501670
A Method for Trapping Intermediates of Poly-
ketide Biosynthesis with a Nonhydrolyzable
Malonyl-Coenzyme A Analogue**
Dieter Spiteller, Claire L. Waterman, and
Jonathan B. Spencer*
There are currently about 10000 natural products that have
been classified as polyketides. Their structures range from
simple molecules, such as 6-methylsalicylic acid,[1] to complex
natural products, such as brevetoxin.[2] Polyketides are an
invaluable source of bioactive compounds with many being
clinically important, for example, the antibiotic erythromycin
and the immunosupressant rapamycin.[3] Despite their diver-
sity, all polyketides share a common biosynthetic pathway; a
starter unit, which is normally acetyl coenzyme A (CoA),
undergoes sequential condensation with an extender unit
(usually malonyl or (S)-methylmalonyl-CoA) in a similar
manner to fatty acid biosynthesis. However, unlike fatty acid
biosynthesis, the carbonyl group is not always completely
reduced after each condensation, which partly accounts for
the huge variety of polyketide structures (Scheme 1).[3]
Scheme 1. Schematic diagram of the mechanism of two rounds of
polyketide biosynthesis by a polyketide synthase (AT: acyl transferase,
KS: ketosynthase, ACP: acyl carrier protein, DH: dehydratase, KR:
ketoreductase). In type III polyketide synthases, such as the stilbene
synthase from Pinus sylvestris, malonyl-CoA is used directly for the
elongation instead of the corresponding ACPs.
Experiments with labeled precursors, in combination with
mechanistic considerations, have been extensively used to
investigate the details of polyketide biosynthesis.[1,3] How-
ever, it has proven very difficult to isolate intermediates
directly as they are covalently attached to the polyketide
synthases (PKSs).[3] One approach that has been possible with
the cloning and sequencing of genes for many PKSs[4] is to
design mutants that halt the biosynthesis at a desired point.
For instance, by genetically engineering the fusion of the
thioesterase from the third subunit of the deoxyerythronolide
synthase onto the first subunit, deoxyerythronolide biosyn-
thesis was stopped after two rounds of (S)-methylmalonyl-
CoA condensations and the intermediate removed in suffi-
cient quantity to be fully characterized.[5] Whereas this
strategy has potential with all type I modular PKSs, it is
very limited when applied to type I iterative, type II, and
type III PKSs, as the same catalytic activities are usually used
repeatedly. For example, if a mutation is made in the
ketosynthase, polyketide synthesis could not occur at all. To
address this point, we sought to develop a chemical approach
that used a non-hydrolyzable malonyl-CoA analogue which
can potentially trap polyketide intermediates from these
classes of PKSs.
Analogues of coenzyme A[6–8] have been used previously
to investigate a range of enzyme reaction mechanisms, such as
the condensation of acetyl-CoA with oxaloacetate catalyzed
by citrate synthase[7] and the enzymic carboxylation of acetyl-
CoA by acetyl-CoA carboxylase.[9] However, malonyl-CoA
analogues have never been used before to investigate the
biosynthesis of polyketides. In our strategy, the malonyl-CoA
analogue would react with the growing polyketide chain but
would not be able to undergo transesterification back onto
the PKS resulting in the accumulation of polyketide inter-
mediates attached to the analogue (Scheme 2a).
[*] C. L. Waterman, Dr. J. B. Spencer
UniversityChemical Laboratory
Universityof Cambridge
Lensfield Road, Cambridge, CB21EW (UK)
Fax: (+44)1223-336-362
E-mail: jbs20@cam.ac.uk
Dr. D. Spiteller
Department of Biochemistry
Universityof Cambridge
80 Tennis Court Road, Cambridge CB21GA (UK)
The analogue of malonyl-CoA 3 used in this study
contains an additional methylene unit between the thiol of
the coenzyme A and the malonate moiety. This methylene
unit could possibly be detrimental to the correct positioning
of the analogue at the active site of a PKS, as the malonyl
moiety is displaced by one carbon atom. However, we felt that
this potential disadvantage was out-weighed by the straight-
forward preparation of 3 in two steps. Commercially available
ethyl 4-chloro-3-oxobutanoate (1) was treated with coenzy-
me A in aqueous 0.11m Li2CO3.[7,8] After purification with
[**] We are indebted to Prof. Dr. Joachim Schröder (Institut für Biologie
II, Biochemie der Pflanzen, Freiburg (Germany)) for providing the
cDNA of stilbene synthase and to Prof. Dr. Fraser (University of
Calgary, Canada) for the gift of a plasmid coding for the succinyl-
CoA:3-ketoacid transferase. D.S. thanks the Deutschen Akademie
der Naturforscher Leopoldina (Germany), for a postdoctoral
fellowship (BMBF-LPD 9901/8-90). Financial support bythe BBSRC
is gratefullyacknowledged.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 7079 –7082
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7079

Communications
Interestingly, we also observed the ions at m/z 958 and
1000 at very low levels in the control samples without the
addition of the enzyme. Although not totally unexpected, to
our knowledge, the non-enzymic formation of polyketides
from malonyl-CoA has not been previously reported.
Several closely eluting peaks, all exhibiting m/z 1000, were
observed in some LC-MS traces of 10 instead of a single one,
thus suggesting the presence of isomers of 10 probably formed
by enolization and/or cyclization.
The MS/MS spectrum of both intermediates that exhibit
([M+H]+) peaks at m/z 958 and 1000 showed characteristic
fragments for the presence of coenzyme A derivatives (see
Figure 2and the Supporting Information). In the case of the
triketide intermediate 10, the fragments at m/z 591 and 428
are characteristic for the cleavage of the phosphodiester bond
of the coenzyme A unit of the molecule (Figure 2). Further-
more, a fragment at m/z 493 was also observed that corre-
sponds to loss of the 3’-phospho adenosine diphosphate
moiety from the alkyl chain. The observed fragmentation
pattern for the intermediates of m/z 958 and 1000 is in
agreement with a study of the MS/MS fragmentation of other
coenzyme A derivatives by Burns et al.[14]
Scheme 2. a) Trapping intermediates of polyketide biosynthesis with a
malonyl-CoA analogue 3; b) synthesis of 3. Enz=enzyme.
RP-18 HPLC, the ethyl ester 2
was hydrolyzed to 3 with pig liver
esterase (PLE; Scheme 2b).[10] As
anticipated,
3
was found to
undergo decarboxylation more
readily than malonyl-CoA (4)
during purification, therefore it
was used directly after precipita-
tion of the PLE.
The potential of 3 to trap
intermediates of polyketide bio-
synthesis was investigated with
the type III PKS stilbene synthase
(STS) from Pinus sylvestris[11,12] as
a model system. This class of PKS
is the simplest to investigate this
approach with as it uses 4 directly,
in contrast to type I and II PKSs
that carry out condensations with
malonyl acid carrier proteins
(ACPs).[3] STS has a broad sub-
strate specificity with regard to
the starter unit, thus producing a
number of stilbene and pyrone
Scheme 3. Formation of the pyrone-type and stilbene-type products 7 and 8, respectively, by stilbene
synthase from Pinus sylvestris using either 5 or 6 as the starter unit and 4 for chain elongation.
products.[13] If cinnamoyl-CoA (6) is used as the starter unit,
the enzyme catalyzes its sequential condensation with three
molecules of 4 to form the stilbene pinosylvin (8),[12] and with
4-hydroxyphenylacetyl-CoA (5), only two condensations with
4 occur to give the pyrone 7 (see Scheme 3).
For our initial experiments, 5 was used as a starter unit and
the ratio of 3 to 4 was varied. Formation of the products was
followed by LC-MS, and across a range of conditions (a ratio
of 3:4 of 3/4 was found to give the highest yield) we observed
ions at m/z 958 and 1000 that correspond to one and two
condensation steps, in addition to the ion at m/z 219 for the
normal pyrone product 7 (Figure 1).
To provide further evidence for the isolation of these
intermediates, assays were carried out with [13C3]malonyl-
CoA.[15] The incorporation of the labeled substrate should
shift the mass of the [M+H]+ ion to 1002amu in the case of
10, as one molecule of [13C3]malonyl-CoA is utilized in its
formation. Figure 2clearly shows that the MS/MS spectra
complement and confirm the MS/MS spectra of the unlabeled
assay. All signals at m/z 1002, 984, 593, 575, 495, and 477,
which contain the polyketide chain in their fragments, are
shifted by 2amu, whereas the signals at m/z 428 and 410,
which correspond to the 3’-phospho adenosine diphosphate
moiety, remained unchanged.
7080 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7079 –7082

Angewandte
Chemie
fore, the addition of 3 to the assay could trap three
different intermediates during the stilbene forma-
tion. However, no intermediates were detected
above the background levels under a range of
conditions. It is possible that the positioning of the
analogue in the active site of the STS is better
matched with 5 as a starter unit because it has one
less carbon atom than 6 and 3 has one more carbon
atom than 4. This observation indicates that the
positioning of the growing polyketide chain bound
to the active-site cysteine unit is strongly influenced
by the structure of the starter unit. A smaller starter
unit may allow the necessary flexibility in the active
site for the analogue to be able to be correctly
positioned and react.
In summary, we have shown for the first time
that intermediates of polyketide biosynthesis can be
trapped for detailed analytical studies with a non-
hydrolyzable malonyl-CoA analogue. This finding
potentially opens the way to the determination of
the timing of crucial events, such as cyclization of the
polyketone chain. It is highly likely that the ana-
logue can be transferred to ACPs with Sfp, a
phosphopantetheinyl transferase from Bacillus sub-
Figure 1. UV chromatogram and ion traces m/z 219, 958, and 1000 observed
after the reaction of the stilbene synthase with 5 and 4 in presence of 3. Peaks:
Coenzyme A (12); decarboxylated malonyl-CoA analogue 11; ethylmalonyl-CoA
analogue 2; STS intermediate formed from two condensations condensations
10; 4-hydroxyphenylacetyl-CoA 5; STS intermediate formed from one condensa-
tion 9; 4-hydroxyphenylacetylpyrone (7).
These results show that the malonyl-CoA analogue 3 can
be successfully used to trap the growing polyketide chain. It is
also interesting that the tetraketide is not observed. This
result indicates that the triketide formed from the starter unit
and two malonyl-CoA molecules either rapidly cyclizes to the
observed pyrone 7 or is not readily transferred back onto the
STS after the final condensation.
tilis, as this enzyme has been shown to have a wide substrate
specificity.[16]
This approach should allow type I and II PKSs to be
investigated. Experiments to apply this concept to other
polyketide synthases and the synthesis of alternative ana-
logues that may be better steric matches for malonyl-CoA are
currently under investigation.
To see if longer intermediates could be trapped, we used 6
as starter unit because the favored product in this case is the
stilbene 8 formed from the tetraketide intermediate. There-
Received: May 14, 2005
Published online: October 6, 2005
Figure 2. ESI-MS/MS of 10 formed bythe stilbene synthase using 5, 3, and either a) 4 or b) [13C3]malonyl-CoA.
Angew. Chem. Int. Ed. 2005, 44, 7079 –7082 ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7081

Communications
Keywords: biosynthesis · coenzymes · mass spectrometry ·
.
polyketides · synthases
[1] A. J. Birch, P. A. Massy-Westropp, C. J. Moye, Aust. J. Chem.
1955, 8, 539 – 544.
[2] Y. Y. Lin, M. Risk, S. M. Ray, D. Vanengen, J. Clardy, J. Golik,
J. C. James, K. Nakanishi, J. Am. Chem. Soc. 1981, 103, 6773 –
6775.
[3] J. Staunton, K. J. Weissman, Nat. Prod. Rep. 2001, 18, 380 – 416.
[4] F. H. Malpartida, D. A. Hopwood, Nature 1984, 309, 462– 464; J.
Cortes, S. F. Haydock, G. A. Roberts, D. J. Bevitt, P. F. Leadlay,
Nature 1990, 348, 176 – 178; S. Donadio, M. J. Staver, J. B.
McAlpine, S. J. Swanson, L. Katz, Science 1991, 252, 675 – 679.
[5] J. Cortes, K. E. H. Wiesmann, G. A. Roberts, M. J. B. Brown, J.
Staunton, P. F. Leadlay, Science 1995, 268, 1487 – 1489; C. M.
Kao, G. L. Luo, L. Katz, D. E. Cane, C. Khosla, J. Am. Chem.
Soc. 1995, 117, 9105 – 9106.
[6] Y. M. Zhao, M. Michenfelder, J. Retey, Can. J. Chem. 1994, 72,
164 – 169; K. W. Vogel, D. G. Drueckhammer, J. Am. Chem. Soc.
1998, 120, 3275 – 3283; P. K. Mishra, D. G. Drueckhammer,
Chem. Rev. 2000, 100, 3283 – 3309.
[7] T. Ciardelli, C. J. Stewart, A. Seeliger, T. Wieland, Liebigs Ann.
Chem. 1981, 828 – 841.
[8] L. A. Paige, G. Q. Zheng, S. A. Defrees, J. M. Cassady, R. L.
Geahlen, J. Med. Chem. 1989, 32, 1665 – 1667.
[9] J. Nikawa, S. Numa, T. Shiba, C. J. Stewart, T. Wieland, FEBS
Lett. 1978, 91, 144 – 148.
[10] W. Boland, C. Frꢀssl, M. Lorenz, Synthesis 1991, 1049 – 1072.
[11] J. Fliegmann, G. Schrꢀder, S. Schanz, L. Britsch, J. Schrꢀder,
Plant Mol. Biol. 1992, 18, 489 – 503.
[12] S. Schanz, G. Schrꢀder, J. Schrꢀder, FEBS Lett. 1992, 313, 71 – 74.
[13] H. Morita, H. Noguchi, J. Schrꢀder, I. Abe, Eur. J. Biochem.
2001, 268, 3759 – 3766.
[14] K. L. Burns, L. T. Gelbaum, M. C. Sullards, D. E. Bostwick, S. W.
May, J. Biol. Chem. 2005, 280, 16550 – 16558.
[15] P. M. Jordan, J. B. Spencer, D. L. Corina, J. Chem. Soc. Chem.
Commun. 1986, 911 – 913.
[16] L. E. N. Quadri, P. H. Weinreb, M. Lei, M. M. Nakano, P. Zuber,
C. T. Walsh, Biochemistry 1998, 37, 1585 – 1595.
7082 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7079 –7082