Exploiting the reaction flexibility of a type III polyketide synthase through in vitro pathway manipulation

Article abstract of DOI:10.1021/ja0441559

A synthetic metabolic pathway has been constructed in vitro consisting of the type III polyketide synthase from Streptomyces coelicolor and peroxidases from soybean and Caldariomyces fumago (chloroperoxidase). This has resulted in the synthesis of the pentaketide flaviolin and its dimeric derivative, and a wide range of pyrones and their coupled derivatives with flaviolin, as well as their halogenated derivatives. The addition of acyl-CoA oxidase to the pathway prior to the polyketide synthase resulted in unsaturated pyrone side chains, further broadening the product spectrum that can be achieved. The approach developed in this work, therefore, provides a new model to exploit biocatalysis in the synthesis of complex natural product derivatives. Copyright

Full text of DOI:10.1021/ja0441559

Published on Web 12/13/2004
Exploiting the Reaction Flexibility of a Type III Polyketide Synthase through
in Vitro Pathway Manipulation
Jae-Cheol Jeong, Aravind Srinivasan, Sabine Gru¨schow,† Horacio Bach,‡ David H. Sherman,*,† and
Jonathan S. Dordick*
Department of Chemical and Biological Engineering, Rensselaer Polytechnic Institute, Troy, New York 12180,
Life Sciences Institute, Department of Medicinal Chemistry, UniVersity of Michigan, Ann Arbor, Michigan 48109
Received September 26, 2004; E-mail: dordick@rpi.edu
Polyketides represent a diverse family of natural products found
in plants, fungi, and soil bacteria, which have found extensive use
as pharmaceuticals (e.g., antibacterials, antivirals, antineoplastics,
among others),1 food ingredients and nutraceuticals,2 pigments,3
and veterinary products.4 Although naturally occurring polyketides
have diverse structures, the majority are produced by three broad
classes of polyketide synthases (PKSs) sharing a common mech-
anism involving sequential decarboxylative condensation reactions.5
While the vast majority of studies have been focused on the
multienzyme type I and iterative type II PKSs, the homodimeric
type III PKSs, which give rise to a range of aromatic compounds
including the flavonoids and chalcones,6 represent an untapped
source of structural diversity. The type III PKS RppA from
Streptomyces griseus, the first bacterial type III PKS to be
characterized, has been shown to possess significant substrate
tolerance.7 RppA catalyzes the sequential decarboxylative conden-
sation, intramolecular cyclization, and aromatization of five units
of malonyl-CoA to give 1,3,6,8-tetrahydroxynaphthalene (THN),
which spontaneously oxidizes to flaviolin (1) upon exposure to air.
THN has been hypothesized to be a key intermediate in the
biosynthesis of numerous natural products (e.g., prenylated naph-
thaquinones, echinochrome, and microbial melanin) that possess
antibacterial and tumor-cytotoxic activity against antibiotic-resistant
pathogens.8 The kinetics of RppA have been previously reported,9
and the crystal structure of the protein has been solved.10 The
structure of the active site region appears to indicate the presence
of a unique cavity that may enable acceptance of a wide range of
starter units with capability for multiple polyketide extensions.
The breadth of natural products arising from RppA suggests that
a still wider array of natural product-like compounds can be
generated by incorporating RppA into a synthetic metabolic pathway
that consists of non-PKS enzymes that also possess broad substrate
specificity. To that end, in the current work we have combined
peroxidase catalysis with the RppA from S. coelicolor11 to expand
the structural diversity of the products from type III PKS catalysis.
The approach developed in this work, therefore, provides a new
model to exploit biocatalysis in the synthesis of complex natural
product derivatives.
Figure 1. RppA catalysis using various starter CoA substrates. Reactions
were performed in 1-mL volumes containing 1 mM malonyl-CoA or 1 mM
each of malonyl-CoA and an alternative acyl-CoA starter unit and 0.1 mg
of RppA. The reaction mixtures were incubated for 2 h at 37 °C after which
50 µL of 4 N HCl was added to stop the reactions. Products were extracted
into 1 mL of ethyl acetate and analyzed by HPLC (see Supporting
Information). (a) For alternate CoA starter units, the reaction was primed
with the acyl-CoA prior to addition of the malonyl-CoA. (b) For the
dehydro-CoA starter units, R ) C5H11, C7H15, and C11H23 were oxidized
by ACO prior to incubating with RppA (bottom scheme) to give 4 and 5.
Compounds 2a, b, f, and h, and 3a, b, f, and h are novel to RppA.
example, lauroyl-CoA and benzoyl-CoA served as starter units to
provide 2f and 3f, and 2h and 3h, respectively; neither starter unit
appears to react in any more than a trace yield with the S. griseus
enzyme.7 Moreover, conversions to RppA-generated products
ranged from 40% for isovaleryl-CoA to 27% for lauroyl-CoA and
octanoyl-CoA to 14% for butyryl-CoA, based on the starter unit
concentration. These relatively high yields lend further support that
the enzyme is inherently flexible toward an array of substrates. In
most cases, a single starter unit yielded two products (2 and 3),
indicating different degrees of condensation. Hence, depending on
the starter unit, RppA can catalyze differential chain extension.
RppA catalysis demonstrated further flexibility by using unsatur-
ated CoA esters generated by acyl-CoA oxidase (ACO). ACO
catalyzes the formation of trans-2,3-dehydroacyl-CoA esters (Figure
1).12 This approach provides potential for creating additional
structural diversity, particularly through the chemistry of the alkene
functionality. To that end, in the presence of ACO hexanoyl-,
octanoyl-, and lauroyl-CoA esters were converted to their respective
trans-2,3-dehydro-CoA esters that subsequently served as starter
units for RppA. Conversions to 4f and 5f, for example, ranged from
5 to 7%, which indicate that RppA is able to accept alkenoyl-CoA
starter units, albeit with low efficiency.
RppA catalyzes the condensation of five units of malonyl-CoA
to 1 in ∼30% yield (Figure 1). We then examined various
commercially available acyl-CoA starter substrates in reactions with
malonyl-CoA as the extender unit. A total of 23 polyketide products
were obtained (Figure 1), and with the exception of the full-length
pentaketide flaviolin, all of the products were pyrones based on
tri-, tetra-, and hexaketides. The reactivity of the S. coelicolor RppA
appears to be broader than that of the S. griseus enzyme.7 For
† Rensselaer Polytechnic Institute.
‡ Current address: Biotechnology Research Laboratories, Taro Pharmaceuticals
USA Inc., New York 10532.
9
64
J. AM. CHEM. SOC. 2005, 127, 64-65
10.1021/ja0441559 CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Figure 3. Halogenation of flaviolin, and butyryl- and isovaleryl-based
pyrones catalyzed by CPO. The yield of 11 from 1 was ca. 10%.
In summary, we have used the type III PKS RppA in Vitro to
generate flaviolin and 22 structurally different pyrones from
malonyl-CoA and other acyl-CoA starter units, including aromatic
and unsaturated acyl moieties. The latter was a result of the coupled
reaction of acyl-CoA oxidase and RppA. The structural diversity
of RppA catalysis was further expanded by coupling RppA with
peroxidase catalysis to yield 14 dimeric, chlorinated, or brominated
compounds. The coupling of oxidative enzymes with RppA
represents an example of the biocatalytic flexibility that extends
natural product structural diversity above and beyond native
pathway endpoints. Further structural diversity may be achieved
by the addition of other enzymes with broad specificity (e.g.,
hydroxylases and transaminases) in both iterative and combinatorial
fashion. This is the subject of our continuing work.
Figure 2. Self- and cross-coupling products catalyzed by SBP from RppA-
generated flaviolin and selected pyrones. The bienzymic reaction was
performed as described in the text for biflaviolin synthesis, except for
reactions with pyrones where a slow feed of H2O2 into the reaction mixture
(33 µL/h) was performed to minimize biflaviolin formation.
The structural diversity of natural polyketides are in large part
due to the various post-PKS tailoring enzymes.6 To expand beyond
the natural polyketides generated solely through RppA catalysis,
we sought other enzymes that could accept the wide range of
structures generated by the PKS and, in the process, generate a
novel in vitro synthetic metabolic pathway. Because flaviolin is a
naphthaquinone derivative, we reasoned that it may be a suitable
substrate for peroxidases, in general, and the highly active soybean
peroxidase (SBP), specifically.13 To demonstrate the reactivity of
flaviolin by SBP, we performed a sequential bienzymic reaction
using RppA and SBP.14 All of the flaviolin generated in the first
reaction was converted to the flaviolin dimer, biflaviolin (6, Figure
2, conversion of ca. 60% of theoretical maximum). Although
biflaviolin has been detected in ViVo,15 this is the first report of
enzymatic synthesis of biflaviolin in Vitro, thereby providing
evidence that peroxidase-mediated coupling may be involved in
the in ViVo production of biflaviolin.
Encouraged by this result, we examined the self- and cross-
coupling of the pyrone products of RppA catalysis, using the pyrone
products from butyryl-CoA and isovaleryl-CoA as substrates for
SBP. In the presence of the butyryl-CoA products (2c and 3c) cross-
coupling of flaviolin (which was present as a coproduct in the RppA
reaction) with the two respective pyrones yielded 7 and 8 (Figure
2). The isovaleryl-CoA product (3g) underwent both cross-coupling
in the presence of the flaviolin coproduct to give 9 and also self-
coupling to give the pyrone dimer 10. The ratio of 9 to 10 was ca.
10:1; hence, cross-coupling was heavily favored. In all cases other
than biflaviolin the conversions were 20% of the theoretical
maximum. These results demonstrate the ability of PKS products
to undergo further transformation to yield higher molecular weight
species that have widely different structures.
Another peroxidase with broad specificity is the chloroperoxidase
(CPO) from Caldariomyces fumago, which catalyzes the chlorina-
tion and bromination of a wide range of aromatic and aliphatic
compounds, including flavones, in the presence of H2O2.16 We
therefore proceeded to perform a sequential bienzymic reaction
similar to that described for SBP, except that in the second stage,
20 µg/mL of CPO was added along with 1 mM of H2O2 and 40
mM of KCl or KBr. To avoid deactivation of CPO by the relatively
high H2O2 concentration, we slowly fed the H2O2 into the reaction
solution. Flaviolin underwent only bromination to give the presumed
3-bromoflaviolin (11). Interestingly, the pyrone products from
butyryl-CoA and isovaleryl-CoA underwent both chlorination and
bromination (Figure 3) to yield several unique products (12 and
13).
Acknowledgment. We acknowledge the financial support of
DARPA and partial support by NSF (BES-0118820 to J.S.D., BES-
0118926 to D.H.S., and CHE-0091892 for LC-MS).
Supporting Information Available: Cloning and purification of
RppA, detailed description of reactions and kinetics, polyketide analysis
by HPLC and LC-MS, MS/MS, and NMR data. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Shen, B. Curr. Opin. Chem. Biol. 2003, 7, 285-295. Kantola, J.; Kunnari,
T.; Mantsala, P.; Ylihonko K. Comb. Chem. High Throughput Screening
2003, 6, 501-512.
(2) Marin, F. R.; Frutos, M. J.; Perez-Alvarez, J. A.; Martinez-Sanchez, F.;
Del Rio, J. A. Stud. Nat. Prod. Chem. 2002, 26, 741-778.
(3) Linnemannstons, P.; Schulte, J. del Mar Prado, M.; Proctor, R. H.; Avalos,
J.; Tudzynski, B. Fung. Gen. Biol. 2002, 37, 134-148.
(4) Kealey, J T.; Liu, L.; Santi, D V.; Betlach, M. C.; Barr, P. J. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 505-509.
(5) Khosla, C.; Harbury, P. B. Nature 2001, 409, 247-252.
(6) Austin, M. B.; Noel, J. P. Nat. Prod. Rep. 2003, 20, 79-110.
(7) Funa, N.; Ohnishi, Y.; Fujii, I.; Shibuya, M.; Ebizuka, Y.; Horinouchi, S.
Nature 1999, 400, 897-899. Funa, N.; Ohnishi, Y.; Ebizuka, Y.;
Horinouchi, S. J. Biol. Chem. 2002, 277, 4628-4635.
(8) Nicolaou, K. C.; Gray, D. L. F. J. Am. Chem. Soc. 2004, 126, 607-612.
Pena-Cabrera, E.; Liebeskind, L. S. J. Org. Chem. 2002, 67, 1689-1691.
Eckermann, S.; Schro¨der, G.; Schmidt, J.; Strack, D.; Edrada, R. A.;
Helariutta, Y.; Elomaa, P.; Kotilainen, M.; Kilpelainen, I.; Proksch, P.;
Teeri, T. H.; Schro¨der, J. Nature 1998, 396, 387-390.
(9) Izumikawa, M.; Shipley, P. R.; Hopke, J. N.; O’Hare, T.; Xiang, L.; Noel,
J. P.; Moore, B. S. J. Ind. Microbiol. Biotechnol. 2003, 30, 510-515.
(10) Austin, M. B.; Izumikawa, M.; Bowman, M. E.; Udwary, D. W.; Ferrer,
J. L.; Moore, B.; Noel, J. P. J. Biol. Chem. 2004 (web version).
(11) RppA was produced by overexpression of the S. coelicolor gene cloned
into Escherichia coli using the NdeI and HindIII restriction sites of the
pET-28b vector. Kanamycin resistance was used for selection of clones,
and protein expression was induced using 0.5 mM IPTG. RppA containing
an N-terminal His6-tag, was purified to near homogeneity using Ni-NTA
column chromatography (see Supporting Information).
(12) Jiang Z.; Thorpe, C. Biochemistry 1983, 22, 3752-3758.
(13) Srinivasan, A.; Wu, X.; Lee, M.-Y.; Dordick, J. S. Biotechnol. Bioeng.
2003, 81, 563-569.
(14) RppA catalysis was initiated first in 1 mL containing 1 mM malonyl-
CoA and 0.1 mg enzyme to generate 0.06 mM flaviolin in 2 h. The second
reaction catalyzed by SBP (5 µg/mL) was then initiated by addition of
0.25 mM H2O2.
(15) Stipanovic, R. D.; Wheeler, H. H. Pesticide Biochem. Physiol. 1980, 13,
198-201.
(16) Yaipakdee, P.; Robertson, L. W. Phytochemistry 2001, 57, 341-347.
JA0441559
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 65