Biosynthesis of HSAF, a tetramic acid-containing macrolactam from Lysobacter enzymogenes

Article abstract of DOI:10.1021/ja105732c

HSAF was isolated from Lysobacter enzymogenes, a bacterium used in the biological control of fungal diseases of plants. Structurally, it is a tetramic acid-containing macrolactam fused to a tricyclic system. HSAF exhibits a novel mode of action by disrupt

Full text of DOI:10.1021/ja105732c

COMMUNICATION
pubs.acs.org/JACS
Biosynthesis of HSAF, a Tetramic Acid-Containing Macrolactam
from Lysobacter enzymogenes
Lili Lou,†,# Guoliang Qian,‡,# Yunxuan Xie,† Jiliang Hang,† Haotong Chen,† Kathia Zaleta-Rivera,†
Yaoyao Li,†,§ Yuemao Shen,§ Patrick H. Dussault,† Fengquan Liu,*,‡ and Liangcheng Du*,†
†Department of Chemistry, University of Nebraska-Lincoln, Lincoln, Nebraska 68588, United States
‡Department of Plant Pathology, Nanjing Agricultural University, Nanjing 210095, China, and
§State Key Laboratory of Microbial Technology, School of Life Science, Shandong University, Jinan 250100, China
S Supporting Information
b
ABSTRACT: HSAF was isolated from Lysobacter enzymo-
genes, a bacterium used in the biological control of fungal
diseases of plants. Structurally, it is a tetramic acid-contain-
ing macrolactam fused to a tricyclic system. HSAF exhibits a
novel mode of action by disrupting sphingolipids important
to the polarized growth of filamentous fungi. Here we
describe the HSAF biosynthetic gene cluster, which con-
tains only a single-module polyketide synthase/nonriboso-
mal peptide synthetase (PKS/NRPS), although the biosyn-
thesis of HSAF apparently requires two separate polyketide
chains that are linked together by one amino acid (ornithine)
via two amide bonds. Flanking the PKS/NRPS are six genes
that encoding a cascade of four tightly clustered redox enzymes
on one side and a sterol desaturase/fatty acid hydroxylase and
a ferredoxin reductase on the other side. The genetic data
demonstrate that the four redox genes, in addition to the PKS/
NRPS gene and the sterol desaturase/fatty acid hydroxylase
gene, are required for HSAF production. The biochemical data
show that the adenylation domain of the NRPS specifically
activates L-ornithine and that the four-domain NRPS is able to
catalyze the formation of a tetramic acid-containing product
from acyl-S-ACP and ornithinyl-S-NRPS. These results reveal a
previously unrecognized biosynthetic mechanism for hybrid
PK/NRP in prokaryotic organisms.
e previously isolated a novel antifungal compound, HSAF
(1), from Lysobacter enzymogenes C3, which is a bacterium
W
used for the biological control of fungal diseases of plants.1 HSAF
exhibits potent inhibitory activities against a wide range of fungal
species and shows a novel mode of action by disrupting the
biosynthesis of fungal sphingolipids.2 Sphingolipids represent an
attractive new target for the development of novel antifungal
drugs because their structure in fungal cells is distinct from that in
mammalian cells.3 HSAF is a tetramic acid (2,4-pyrrolidinedione)-
containing macrolactam (Figure 1).1 Tetramic acid is the key
structural feature of many bioactive heterocycles that exhibit a wide
range of biological activities, including antibiotic and anticancer
activities.4 The macrolactam structure is also found in many bioactive
natural products. HSAF combines these two structural features. In
addition, it contains a 5,5,6-tricyclic system fused to the macrolactam.
Such structural features are shared by a group of natural products
Figure 1. (A) Structure of HSAF and related tetramic acid-containing
macrolactams. On the basis of their possible biosynthetic mechanisms,
the compounds can be divided into two groups: the hexaketide-
ornithine-hexaketide group and the pentaketide-ornithine-hexaketide
group. Both groups can have a 17-membered macrolactam fused to a
tricyclic system or a 21-membered macrolactam fused to a bicyclic system.
(B) Map of the HSAF gene cluster (EF028635.2). The ORFs in the
blocked region are annotated here; the annotation of the rest of the ORFs is
included in Figure S6.
(2-7) that were isolated from marine invertebrates and/or asso-
ciated microorganisms (Figure 1). No biosynthetic studies on these
Received: June 29, 2010
Published: December 20, 2010
r
2010 American Chemical Society
643
dx.doi.org/10.1021/ja105732c J. Am. Chem. Soc. 2011, 133, 643–645
|

Journal of the American Chemical Society
COMMUNICATION
marine metabolites have been reported. Previously, we reported four
HSAF biosynthetic genes, including a hybrid polyketide synthase/
nonribosomal peptide synthetase (PKS/NRPS). Since then, we have
been searching for additional PKS genes around this PKS/NRPS
locus because the biosynthesis of HSAF apparently requires two
separate polyketide chains. We now provide the complete sequence of
the gene cluster based on the genome sequences. The most striking
finding is that only a single PKS/NRPS is present in the gene cluster.
We have also expressed the NRPS module and demonstrated in vitro
that it specifically activates L-ornithine and catalyzes the formation of
a tetramate product.
In L. enzymogenes C3, we previously identified three genes
downstream from the PKS/NRPS. Gene disruption data showed
that only the gene encoding sterol desaturase/fatty acid hydro-
xylase is required for HSAF production.1 Because biosynthetic
genes for PK/NRP in microorganisms are almost always clus-
tered together, we searched the region upstream from the PKS/
NRPS, where we found four genes (OX1 to OX4, EF028635.2)
that encode a cascade of NADP/FAD-dependent oxidoreduc-
tases. OX1, OX2, and OX3 are similar to each other, while OX4 is
distinct (Figure 1). The OX genes appear to form an operon with
the PKS/NRPS because they are either translationally coupled or
have only a small intergenic region between them. To test their
relevance to HSAF biosynthesis, we disrupted each of the OX
genes. The disruption mutants of OX1, OX2, and OX3 did not
produce HSAF, whereas the disruption mutants of OX4 pro-
duced a small amount of a metabolite with [M þ H]þ at m/z
511.2789, which is two mass units smaller than HSAF and
coincident with the mass of maltophilin, a putative precursor
of HSAF (Figure S1 in the Supporting Information). To
eliminate the potential polar effect caused by insertion of the
conjugal vector, we then generated four in-frame deletion
mutants by deleting a small sequence (6-11 residues) at the
highly conserved active sites of the OXs (Figures S2-S5). The
results showed that the deletion mutants of OX1, OX2, and OX3
did not produce HSAF while the OX4 deletion mutants still
produced the same putative precursor. Together, the results
show that the OX genes are required for the biosynthesis of the
final product.
The Liu group has been using L. enzymogenes OH11 as a
biocontrol agent for plant fungal diseases in China.5 We recently
completed the draft genome sequence of strain OH11. The
results showed that strain OH11 has all of the HSAF genes found
in strain C3. Within the HSAF genes (20280 bp), the two strains
have a 96% identity. At the amino acid sequence level, the indivi-
dual open reading frames (ORFs) have an identity of 98-99%.
With the genome sequence, we were able to search for “extra
PKS” genes juxtaposed to the HSAF locus. Surprisingly, no such
gene was found. The center of the gene cluster is the hybrid PKS/
NRPS. Flanking the PKS/NRPS are genes encoding six redox
enzymes, arginase, and putative transporters and receptors/
regulators (Figure 1B and Figure S6). Beyond the 55 kb region
shown in Figure 1B, we found only genes encoding rRNA,
ribosome proteins, tRNA, and DNA-repairing proteins, which
apparently are not related to HSAF biosynthesis. The HSAF
biosynthetic gene cluster contains only a single-module PKS/
NRPS, a cascade of redox enzymes, and several putative trans-
porter/regulators. Recently, Clardy and co-workers also ob-
served similar results in gene clusters mined from a number of
Streptomyces genomes.6 They found that this type of single-
module hybrid PKS/NRPS is conserved among phylogenetically
diverse bacterial species.
Figure 2. In vitro synthesis of tetramic acid-containing product 8 using
the purified HSAF NRPS. (A) Schematic illustration of the reactions.
The acyl-S-ACP substrate was prepared by incubation of ACP and
stearoyl-CoA with Svp, and Orn-S-NRPS was prepared by incubation of
NRPS and CoA with Svp followed by addition of L-ornithine and ATP .
(B) SDS-PAGE gel showing the expression and purification of the four-
domain NRPS. The expected size of the NRPS was 148.6 kDa. Lane 1,
markers; lane 2, purified NRPS; lane 3, soluble fraction of protein
extracts; lane 4, total protein extracts upon IPTG induction; lane 5, total
protein extracts before IPTG induction. (C) LC-MS detection of the
expected product 8. The top spectrum is for a control reaction without
the NRPS. The middle spectrum is for a complete reaction and shows
the production of the expected product with [M þ H]þ at m/z
647.6080. The bottom spectrum is that of the chemically synthesized
standard 8 (mixture of diastereomers).
To test whether the single NRPS module could accept two
separate acyl chains to produce the characteristic tetramic acid-
containing product, we conducted a series of in vitro experi-
ments. We first expressed the adenylation domain of the PKS/
NRPS in Escherichia coli and used the purified 66.4 kDa protein
(Figure S7) for an ATP-PPi exchange assay. Among the 16
amino acids tested, the A domain specifically activates L-ornithine
and, to a lesser extent, L-lysine (Figure S7). We then tested the
in vitro activity of the entire NRPS module (Figure 2). The four-
domain NRPS (C-A-PCP-TE) was expressed in E. coli
as a His6-tagged protein with an expected size of 148.6 kDa
(Figure 2B). The promiscuous 40-phosphopantetheine transfer-
ase Svp was used to prepare stearoyl-S-ACP and holo-NRPS.7
The holo-NRPS was then loaded with L-ornithine in the presence
of ATP, producing the ornithinyl-S-NRPS substrate (Figure 2A).
When stearoyl-S-ACP and ornithinyl-S-NRPS were incubated
together, we detected an [M þ H]þ ion at m/z 647.6080 in the
reactions (Figure 2C), which is consistent with the predicted
mass of the product 8. This molecular ion was not present in the
control reaction without the NRPS protein. The putative tetra-
mic acid structure was correlated with an authentic sample
prepared as illustrated in Scheme 1 (also see Scheme SI-1 in
the Supporting Information). The boroxazolidinone adduct of
L-ornithine (9) was selectively acylated at the terminal amino
group with octadecanoyl chloride to furnish a monoamide.8
Deprotection of the boroxazolidinone with ethylenediamine
followed by reaction with thionyl chloride in methanol furnished
methyl ester 10. Conversion of the amino acid ester to the
tetramic acid was accomplished through a modification of a
published approach.9 Condensation of the amino ester with
644
dx.doi.org/10.1021/ja105732c |J. Am. Chem. Soc. 2011, 133, 643–645

Journal of the American Chemical Society
COMMUNICATION
Scheme 1. Preparation of Tetramic Acid 8
chains that are linked together via two amide bonds on the same
amino acid and are subsequently folded/cyclized into a complex
scaffold, probably by a cascade of redox enzymes. These results
open a new area of biosynthetic research to study hybrid
PK/NRP products with a very unusual scaffold and interesting
biological activities.
’ ASSOCIATED CONTENT
S
Supporting Information. Details of experimental proce-
b
dures, HSAF gene cluster annotation, gene disruption/deletion,
chemical synthesis, and protein expression and activity assays.
This material is available free of charge via the Internet at http://
pubs.acs.org.
2-benzyloxycarbonyloctacadecanoic acid (11), which was pre-
pared in two steps from dibenzyl malonate, furnished the
R,δ-bisamide 12 as an inseparable ∼1:1 mixture of diastereo-
mers. Reaction with excess sodium hydride resulted in tandem
Claisen condensation and deacylation to furnish tetramic acid 8
as a 1:1 mixture of epimers at C3 of the 3,5-dioxopyrrolidine. The
retention time and mass spectra of the synthetic tetramic acid
sample exactly matched those of the enzymatically synthesized
material (Figure 2C).
’ AUTHOR INFORMATION
Corresponding Author
fqliu20011@sina.com; ldu@unlserve.unl.edu
Author Contributions
#These authors contributed equally.
These results show that the two amide functionalities in HSAF
are formed between two acyl chains and ornithine, one at the
R-amino group and the other at the δ-amino group of ornithine.
The results also suggest that the HSAF NRPS module is able to
accept two polyketide acyl-S-ACPs, making two amide bonds on
the same amino acid via two Claisen condensation reactions and
forming the tetramic acid ring via a Dieckmann-type reaction.
Hybrid PKS/NRPS have been found in gene clusters for several
tetramic acid-containing natural products, such as tenellin, aspyr-
idone A, equisetin, and cyclopiazonic acid.10 All these PKS/NRPS
contain a reductase (R) domain in the end of the enzyme, which is
responsible for the cyclization of tetramic acid and the release of the
hybrid PK/NRP chain via a Dieckmann-type condensation. In
HSAF PKS/NRPS, the TE domain is the most likely candidate for
this activity. In spite of the functional similarity, the R domain's
substrate is a thioester attached to the PCP domain, whereas the TE
domain's substrate is an oxoester attached to TE itself.11 Another
interesting question in HSAF biosynthesis is whether the C domain
catalyzes the formation of both amide bonds or the TE domain is also
involved. We are currently working on the enzymes to answer this
question.
’ ACKNOWLEDGMENT
This work was supported in part by the NIH (AI073510),
Nebraska Research Initiatives, the NSFC (31028019), and the
National High Technology Research and Development Program
of China (2006AA10A211). We thank Ron Cerny, Kurt Wulser,
and Wei Zhang for technical assistance. The research was
performed in facilities renovated with support from the NIH
(RR015468-01).
’ REFERENCES
(1) Yu, F.; Zaleta-Rivera, K.; Zhu, X.; Huffman, J.; Millet, J. C.;
Harris, S. D.; Yuen, G.; Li, X. C.; Du, L. Antimicrob. Agents Chemother.
2007, 51, 64–72.
(2) Li, S.; Du, L.; Yuen, G.; Harris, S. D. Mol. Biol. Cell 2006, 17,
1218–1227.
(3) Thevissen, K.; Francois, I. E.; Aerts, A. M.; Cammue, B. P. Curr.
Drug Targets 2005, 6, 923–928.
(4) Schobert, R.; Schlenk, A. Bioorg. Med. Chem. 2008, 16, 4203–4221.
(5) Qian, G. L.; Hu, B. S.; Jiang, Y. H.; Liu, F. Q. Agric. Sci. China
2009, 8, 68–75.
(6) Blodgett, J. A. V.; Oh, D. C.; Cao, S.; Currier, C. R.; Kolter, R.;
Clardy, J. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 11692–11697.
(7) Sanchez, C.; Du, L.; Edwards, D. J.; Toney, M. D.; Shen, B. Chem.
Biol. 2001, 8, 725–738.
(8) Dent, W. H.; Erickson, W. R.; Fields, S. C.; Parker, M. H.;
Tromiczak, E. G. Org. Lett. 2002, 4, 1249–1251.
(9) Page, P. C. B.; Hamzah, A. S.; Leach, D. C.; Allin, S. M.; Andrews,
D. M.; Rassias, G. A. Org. Lett. 2003, 5, 353–355.
(10) (a) Halo, L. M.; Marshall, J. W.; Yakasai, A. A.; Song, Z.; Butts,
C. P.; Crump, M. P.; Heneghan, M.; Bailey, A. M.; Simpson, T. J.;
Lazarus, C. M.; Cox, R. J. ChemBioChem 2008, 9, 585–594. (b)
Bergmann, S.; Schumann, J.; Scherlach, K.; Lange, C.; Brakhage, A. A.;
Hertweck, C. Nat. Chem. Biol. 2007, 3, 213–217. (c) Sims, J. W.;
Schmidt, E. W. J. Am. Chem. Soc. 2008, 130, 11149–11155. (d) Liu, X.;
Walsh, C. T. Biochemistry 2009, 48, 8746–8757.
(11) Du, L.; Lou, L. Nat. Prod. Rep. 2010, 27, 255–278.
(12) Cox, R. J. Org. Biomol. Chem. 2007, 5, 2010–2026.
(13) (a) Shao, L.; Qu, X. D.; Jia, X. Y.; Zhao, Q. F.; Tian, Z. H.;
Wang, M.; Tang, G. L.; Liu, W. Biochem. Biophys. Res. Commun. 2006,
345, 133–139. (b) Wenzel, S. C.; Gross, F.; Zhang, Y.; Fu, J.; Stewart,
A. F.; Muller, R. Chem. Biol. 2005, 12, 349–356.
Clardy and co-workers6 found that all HSAF-type gene
clusters contain a central PKS/NRPS flanked by genes encoding
a cascade of redox enzymes. In HSAF, we found that all four OX
genes, in addition to the genes for PKS/NRPS and sterol
desaturase/fatty acid hydroxylase, are required for HSAF bio-
synthesis. These genes could be involved in the formation of the
5,5,6-tricyclic system and the conversion of maltophilin to HSAF.
Although the disruption of the ferredoxin reductase gene and the
arginase gene did not eliminate HSAF production in the mutants,
these genes may play a pathway-specific role (as a reducing
partner for the redox enzymes or in the synthesis of ornithine
from arginine via the urea cycle). The MFS transporter and the
TonB-dependent outer membrane receptor could play a role in
resistance and regulation.
The iterative use of a single-module PKS/NRPS is common in
fungi but rare in bacteria.12 Only a few bacterial modular PKSs
have been reported to act iteratively.13 Our present results show
that HSAF PKS/NRPS represents one of the first two examples
in which a single-module PKS assembles two separate polyketide
645
dx.doi.org/10.1021/ja105732c |J. Am. Chem. Soc. 2011, 133, 643–645