Biosynthesis of the Tetramic Acids Sch210971 and Sch210972

Article abstract of DOI:10.1021/acs.orglett.5b00715

A biosynthetic pathway to fungal polyketide-nonribosomal peptide natural products, Sch210971 (1a) and Sch210972 (1b) from Hapsidospora irregularis, was characterized by reconstitution and heterologous expression in Fusarium heterosporum. Using genetic, biochemical, and feeding experiments, we show that the incorporated amino acid 4-hydroxyl-4-methyl glutamate (HMG) is synthesized by an aldolase, probably using pyruvate as the precursor.

Full text of DOI:10.1021/acs.orglett.5b00715

Letter
pubs.acs.org/OrgLett
Biosynthesis of the Tetramic Acids Sch210971 and Sch210972
Thomas B. Kakule,† Shuwei Zhang,‡ Jixun Zhan,‡ and Eric W. Schmidt*,†
†Department of Medicinal Chemistry, L. S. Skaggs Pharmacy Institute, University of Utah, Salt Lake City, Utah 84112, United States
‡Department of Biological Engineering, Utah State University, Logan, Utah 84322, United States
S
* Supporting Information
ABSTRACT: A biosynthetic pathway to fungal polyketide−nonribosomal peptide
natural products, Sch210971 (1a) and Sch210972 (1b) from Hapsidospora irregularis,
was characterized by reconstitution and heterologous expression in Fusarium
heterosporum. Using genetic, biochemical, and feeding experiments, we show that the
incorporated amino acid 4-hydroxyl-4-methyl glutamate (HMG) is synthesized by an
aldolase, probably using pyruvate as the precursor.
onproteinogenic amino acids (those not normally found in
proteins) are abundant components of secondary
PKS synthesizes the acetate-derived portion, including the
decalin, while the NRPS adds the amino acid. An auxiliary enoyl
reductase (ER) protein and Diels−Alderase are essential to
decalin formation as well (Scheme 2).8−10 Therefore, it could be
predicted that 1 would be synthesized by a PKS-NRPS and
auxiliary ER. Two likely possibilities could be envisioned for
HMG incorporation (Scheme 1). The amino acid could be
derived from leucine, for example, via α-ketoglutarate-dependent
oxidation that is well-known in NRPS amino acid biochemistry.1
Such an oxidase would act on an amino acid either before or after
N
metabolites, contributing greatly to natural chemical diversity.1
Among these, 4-hydroxyl-4-methyl glutamate (HMG, 2) has
long been known because it is concentrated in several plants.2 In
the 1960s, a biosynthetic pathway to plant HMG was proposed,3
wherein two units of pyruvate would undergo aldol condensation
to 4-hydroxyl-4-methyl-2-oxoglutarate (HMOG, 3) (Scheme 1).
Scheme 1. Hypotheses for HMOG Biosynthesis
Scheme 2. Biogenesis of 1, Showing Isolated Shunt Product 4
Subsequently, transamination would yield HMG. Aldolases were
isolated in the late 1960s that were thought to catalyze HMOG
synthesis in vivo.3,4 However, in vitro, they only catalyzed the
reverse reaction, degrading HMOG to pyruvate. Because this
reaction also proceeds spontaneously in the presence of
magnesium, isotope exchange assays were used to provide
indirect evidence that the reaction could proceed in the forward
direction.4
Later, HMG was also discovered as an amino acid
incorporated into co-occurring fungal secondary metabolites,
Sch210971 (1a), Sch210972 (1b) and other close structural
relatives.5,6 Compound 1b is a decalin-containing tetramic acid
derivative that inhibits the cytokine receptor. Only its relative
configuration has been reported. Over the past decade, the
biosynthesis of several decalin-tetramic acids has been studied,
revealing that their core structures are synthesized by dimodular
polyketide synthase−nonribosomal peptide synthetase (PKS-
NRPS) proteins.7 In these large hybrid biosynthetic enzymes, the
Received: March 11, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b00715
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
incorporation by the PKS-NRPS and would represent a quite
novel route to HMG. Alternatively, an aldolase−transaminase
route may be possible, as proposed for the plant metabolites but
not further investigated in recent decades.
In a discovery program employing novel fungi, we
rediscovered the diastereomers 1a and 1b, but in a fungus not
previously known to produce the compound (Hapsidospora
irregularis). Because of our interest in tetramic acids and the
novel nonproteinogenic HMG subunit, we chose to investigate
its biosynthesis.
biosynthetic hypotheses without further testing. Hence, all the
genes were cloned into vectors by yeast homologous
recombination, and heterologously expressed in the recently
reported Fusarium heterosporum expression platform.11,12 The
cluster also contained homologues of two genes universally
found in decalin-tetramic acid clusters, including an amidohy-
drolase and a Diels−Alderase.10 Close homologues of these
genes are found in the F. heterosporum genome, and in our
experience, these do not need to be transferred to F. heterosporum
to produce closely related decalin-tetramates.11,13
The gene cluster for Sch210971/2 was identified by
sequencing. The genome of H. irregularis was sequenced and
assembled to provide 28.4 Mbp (1924 contigs, 55.8% GC).
Automated annotation predicted 14 573 proteins. BLAST
searching for PKS-NRPS and auxiliary ER proteins unveiled 3
PKS-NRPS genes, but only one with an accompanying ER,
leading us to propose that this was the correct cluster, tas
(tetramic acid Sch210971/2; Figure 1A and Table 1). The tas
Heterologous expression was used to demonstrate that tas was
correctly identified as the Sch210971/2 biosynthetic locus
(Figure 1B). We used the fungus F. heterosporum because it
synthesizes sufficient recombinant products for simple chemical
characterization.11 The following combinations of genes were
coexpressed: (1) PKS-NRPS tasS and ER tasC; (2) tasS, tasC,
and aldolase tasA; (3) tasS, tasC, and transaminase tasG; (4) all
four genes tasS, tasC, tasA, and tasG. No compounds were
detected in condition 1, when just the decalin-tetramate
synthesizing genes were present, nor were compounds detected
in condition 3, with the addition of transaminase. However, in
conditions 2 and 4, which both included the aldolase, authentic
Sch210971 and Sch210972 were obtained, as validated by
HPLC, including coinjection with an authentic standard, and by
MS experiments. In some conditions, we also obtained a minor
amount of compound 4, which is similar to an intermediate
previously isolated in heterologous PKS-NRPS experiments and
which represents the pre-Diels−Alder PKS structure of 1
(Scheme 2).12 Compound 4 was purified and characterized by
NMR from a 100 mL culture (see Supporting Information). It is
likely that the host aspartate aminotransferase is involved in
transforming HMOG into HMG; homologues of this enzyme are
known to be promiscuous in other systems.14,15
The requirement for aldolase enzyme, and the lack of a
requirement for the clustered oxidase tasH, strongly implicated
the aldolase mechanism rather than leucine oxidation in the
synthesis of 1. To further investigate this possibility, we
synthesized HMOG from pyruvate16 and fed it to F. heterosporum
expressing just tasS and tasC, lacking the aldolase (Figure 2).
Indeed, feeding with HMOG fully restored production of 1,
while an identical control culture lacking HMOG did not
produce any recombinant compounds.
Figure 1. (A) H. irregularis tas gene cluster schematic. (B) Coexpression
of genes from H. irregularis tas cluster in heterologous host F.
heterosporum, followed by extraction and HPLC-DAD analysis. The
blue line indicates presence of Sch210971/2 only when aldolase is
coexpressed.
To further verify that HMOG was a true intermediate of the
pathway and to rule out alternative explanations, we synthesized
HMOG using 1-13C pyruvate, which was selected because the
label on carboxyl was unlikely to wind up in the polyketide
portion of the molecule even if HMOG was degraded to pyruvate
in the culture. We used five different ratios of labeled/unlabeled
pyruvate in chemical synthesis of HMOG, including 100%, 75%,
50%, 25%, and 0% labeled pyruvate. This provided either pure
Table 1. Predicted Function of tas Genes
gene name
length (aa, nt)
predicted function
tasH
tasG
tas3
tasS
tasK
tasR
tasC
tasA
324, 1070
325, 1675
397, 1194
4061, 12450
981, 3152
566, 1701
385, 1158
270, 889
oxidase
transaminase
Diels−Alderase
PKS-NRPS
amidohydrolase
regulator
labeled or unlabeled HMOG, containing either zero or two 13
C
C
units, or it provided statistical mixtures of zero, one, or two 13
ER
labels. It was envisioned that these mixtures would enable us to
rule out confounding factors. Labeled HMOG was fed to F.
heterosporum containing tasS + tasC (Figure 3). Cultures were
extracted and analyzed by HPLC-MS. In the event, 1 with
unlabeled pyruvate provided a major ion at m/z 446, while that
with 100% 13C label provided a major ion at m/z 448.
Fermentations derived from 25%, 50%, or 75% labeled
HMOG contained the predicted mixtures of peaks at m/z 447,
448, and 449. These results revealed that HMOG was
incorporated intact into 1 and that metabolism into individual
pyruvate subunits within the fungus did not interfere with
aldolase
gene cluster was deposited in GenBank, accession number
KP835202. In addition, we identified coclustered genes that were
predicted to be involved in forming HMG, including genes
encoding an oxidase, an aldolase, and an aminotransferase.
Overall, tas was very similar to an uncharacterized cluster in the
genome sequence of Acremonium chrysogenum ATCC 11550.
On the basis of the proposed activities of the cluster genes, it
was impossible to eliminate either the aldolase or oxidase
B
DOI: 10.1021/acs.orglett.5b00715
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
shown to act as an aldolase in vitro is required for HMG
incorporation in vivo.
In summary, we report the biosynthetic pathway to the
cytokine receptor inhibitors Sch210971 and Sch210972,
identified by genome sequencing and heterologous pathway
expression. We demonstrate the novel aldolase route to a
nonproteinogenic amino acid found in plants and filamentous
fungi, expanding the known routes to nonproteinogenic amino
acids in NRPS metabolism. This also adds to the known amino
acid precursors in fungal PKS-NRPS metabolism, which
currently include proteinogenic amino acids serine, threonine,
tryptophan, phenylalanine, and tyrosine.
ASSOCIATED CONTENT
* Supporting Information
■
S
Figure 2. Incorporation of synthetic HMOG. Culture extracts were
analyzed by HPLC-MS. Shown are filtered chromatograms (m/z
445.7−446.7). When aldolase tasA is present (top), compound 1 is
produced. Without tasA (bottom and middle), 1 is only produced with
addition of synthetic HMOG (middle).
Plasmid construction; fungal transformation procedure; chem-
ical analysis of mutants; and HRESIMS and NMR spectral data.
This material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: ews1@utah.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This project was supported by NSF 0957791 and a Utah State
University Research Catalyst grant.
■
REFERENCES
■
(1) Walsh, C. T.; O’Brien, R. V.; Khosla, C. Angew. Chem., Int. Ed. 2013,
52, 7098.
(2) Grobbelaar, N.; Pollard, J. K.; Steward, F. C. Nature 1955, 175, 703.
(3) Shannon, L. M.; Marcus, A. J. Biol. Chem. 1962, 237, 3342.
(4) Marcus, A.; Shannon, L. M. J. Biol. Chem. 1962, 237, 3348.
(5) Neumann, K.; Kehraus, S.; Gutschow, M.; Konig, G. M. Nat. Prod.
Commun. 2009, 4, 347.
(6) Yang, S.-W.; Mierzwa, R.; Terracciano, J.; Patel, M.; Gullo, V.;
Wagner, N.; Baroudy, B.; Puar, M.; Chan, T.-M.; McPhail, A. T.; Chu,
M. J. Nat. Prod. 2006, 69, 1025.
(7) Chooi, Y. H.; Tang, Y. J. Org. Chem. 2012, 77, 9933.
(8) Kakule, T. B.; Sardar, D.; Lin, Z.; Schmidt, E. W. ACS Chem. Biol.
2013, 8, 1549.
(9) Kennedy, J.; Auclair, K.; Kendrew, S. G.; Park, C.; Vederas, J. C.;
Richard Hutchinson, C. Science 1999, 284, 1368.
Figure 3. Labeled HMOG incorporation. HMOG was synthesized using
different ratios of 13C labeled and unlabeled pyruvate (right). These
synthetic HMOG derivatives were fed to cultures containing tasS + tasC,
but lacking aldolase. Analysis by HPLC-MS in continuous mode (left)
shows expected isotope incorporation. Isotope ratio % expected is
calculated from the expected values for different ratios of unlabeled,
singly labeled, and doubly labeled HMOG, normalized to 100% for the
expected largest peak.
(10) Kato, N.; Nogawa, T.; Hirota, H.; Jang, J.-H.; Takahashi, S.; Ahn,
J. S.; Osada, H. Biochem. Biophys. Res. Commun. 2015, DOI: 10.1016/
j.bbrc.2015.03.011.
(11) Kakule, T. B.; Jadulco, R. C.; Koch, M.; Janso, J. E.; Barrows, L. R.;
Schmidt, E. W. ACS Synth. Biol. 2014, DOI: 10.1021/sb500296p.
(12) Kakule, T. B.; Lin, Z.; Schmidt, E. W. J. Am. Chem. Soc. 2014, 136,
17882.
incorporation. Moreover, the stoichiometry of the label was
consistent with incorporation cleanly into the amino acid portion
of 1 (one, and only one, HMOG unit was incorporated into the
compound).
We investigated the catalytic activity of the aldolase, which was
recombinantly expressed in E. coli as the His-tagged construct. In
all conditions attempted, selective HMOG fission could be
achieved, but the forward reaction to synthesize HMOG from
pyruvate was not observed. This was identical to the previous
reverse reaction observed with plant aldolases in the 1960s and
reinforces these earlier findings associating such aldolases with
HMOG synthesis in nature.3,4 It also confirms that the protein
acts as aldolase in vitro. In this instance, genetic data by
heterologous expression in fungi enables us to provide evidence
that HMG is produced via the aldolase route since a protein
(13) Jadulco, R. C.; Koch, M.; Kakule, T. B.; Schmidt, E. W.; Orendt,
A.; He, H.; Janso, J. E.; Carter, G. T.; Larson, E. C.; Pond, C.; Matainaho,
T. K.; Barrows, L. R. J. Nat. Prod. 2014, 77, 2537.
(14) Baldwin, J. E.; Dyer, R. L.; Ng, S. C.; Pratt, A. J.; Russell, M. A.
Tetrahedron Lett. 1987, 28, 3745.
(15) Senuma, M.; Nakamichi, K.; Nabe, K.; Nishimoto, S.; Tosa, T.
Appl. Biochem. Biotechnol. 1989, 22, 141.
(16) Hel
Catal. 2001, 343, 692.
́
aine, V.; Rossi, J.; Gefflaut, T.; Alaux, S.; Bolte, J. Adv. Synth.
C
DOI: 10.1021/acs.orglett.5b00715
Org. Lett. XXXX, XXX, XXX−XXX