Characterization of a fungal thioesterase having claisen cyclase and deacetylase activities in melanin biosynthesis

Article abstract of DOI:10.1016/j.chembiol.2012.10.002

Melanins are a broad class of darkly pigmented macromolecules formed by oxidative polymerization of phenolic monomers. In fungi, melanins are known virulence factors that contribute to pathogenicity. Their biosynthesis generally involves polymerization of 1,8-dihydroxynaphthalene via a 1,3,6,8- tetrahydroxynaphthalene (THN) precursor assembled by multidomain, nonreducing polyketide synthases. Convergent routes to THN have evolved in fungi. Parallel heptaketide and hexaketide pathways exist that utilize conventional C-terminal thioesterase/Claisen cyclase domains and separate side-chain deacylases. Here, in vitro characterization of Pks1 from Colletotrichum lagenarium establishes a true THN synthase with a bifunctional thioesterase (TE) catalyzing both cyclization and deacetylation of an enzyme-bound hexaketide substrate. Chimeric TE domains were generated by swapping lid regions of active sites between classes of melanin TEs to gain insight into this unprecedented catalysis of carbon-carbon bond making and breaking by an α/β-hydrolase fold enzyme.

Full text of DOI:10.1016/j.chembiol.2012.10.002

Chemistry & Biology
Article
Characterization of a Fungal Thioesterase
Having Claisen Cyclase and Deacetylase Activities
in Melanin Biosynthesis
Anna L. Vagstad,1 Eric A. Hill,1 Jason W. Labonte,1 and Craig A. Townsend1,
1Department of Chemistry, The Johns Hopkins University, Baltimore, MD 21218, USA
*Correspondence: ctownsend@jhu.edu
*
http://dx.doi.org/10.1016/j.chembiol.2012.10.002
SUMMARY
1995). Understanding the fundamental principles of melanogen-
esis may enable development of new antifungals to combat
Melanins are a broad class of darkly pigmented infections.
The major pathway for melanin biosynthesis in fungi is through
macromolecules formed by oxidative polymerization
of phenolic monomers. In fungi, melanins are known
virulence factors that contribute to pathogenicity.
Their biosynthesis generally involves polymerization
of 1,8-dihydroxynaphthalene via a 1,3,6,8-tetrahy-
droxynaphthalene (THN) precursor assembled by
multidomain, nonreducing polyketide synthases.
Convergent routes to THN have evolved in fungi.
Parallel heptaketide and hexaketide pathways exist
that utilize conventional C-terminal thioesterase/
polymerization of polyketide-derived 1,8-dihydroxynaphthalene
(DHN, 1) monomers into dark brown or black cell-wall pigments.
Melanins are highly insoluble, and thus are resistant to structural
characterization, but have electrochemical and mechanical
properties that help exert their protective or virulent functions
(Jacobson, 2000). The pivotal intermediate of the biosynthetic
pathways is 1,3,6,8-tetrahydroxynaphthalene (THN, 2), which
undergoes two rounds of enzymatic reduction and dehydration
to 1. The THN skeleton is synthesized by multidomain, nonre-
ducing iterative polyketide synthases (NR-PKSs). In some cases,
Claisen cyclase domains and separate side-chain melanin genes are clustered at one genetic locus, but they can
also be dispersed throughout the genome (Woo et al., 2010).
Although nonessential, melanin pigments are important for
survival and longevity. In many opportunistic plant and human
deacylases. Here, in vitro characterization of Pks1
from Colletotrichum lagenarium establishes a true
THN synthase with a bifunctional thioesterase (TE)
catalyzing both cyclization and deacetylation of an
enzyme-bound hexaketide substrate. Chimeric TE
domains were generated by swapping lid regions of
active sites between classes of melanin TEs to gain
insight into this unprecedented catalysis of carbon–
pathogenic fungi, color-block mutants are significantly less or
not infective (Bell and Wheeler, 1986; Langfelder et al., 2003).
Colletotrichum lagenarium (syn. Colletotrichum orbiculare) is
a dematiaceous fungus that infects members of the cucumber
family and causes anthracnose disease (Tsuji et al., 2003).
Melanin structurally stabilizes the cell wall, facilitating turgor
carbon bond making and breaking by an a/b-hydro-
lase fold enzyme.
pressure to build up in specialized infectious bodies called
appressoria and is essential for host penetration (Pihet et al.,
2009; Takano et al., 1997). Treatment of C. lagenarium with the
melanin inhibitor tricyclazole (which targets THN reductase)
induces an albino phenotype, causes flaviolin (3, the autoxida-
tion product of THN) to accumulate, and prevents infiltration of
INTRODUCTION
Fungal infections afflict millions of people and cause crop losses the plant dermis (Kubo et al., 1985).
in the billions of dollars each year (Gladieux et al., 2011). Many
The importance of THN (2) is highlighted by the existence of
human pathogenic fungi are dematiaceous yeasts that produce convergent biosynthetic routes to this common metabolite in
dark cell-wall pigments called melanins and are among the most different fungi, as outlined in Figure 1. In Aspergillus fumigatus,
resistant to current antifungal treatments (Nosanchuk and Casa- heptaketide YWA1 (4) is the hemiketal of the direct product of
devall, 2006). Melanin is linked to fungal virulence in a variety of the Alb1 NR-PKS (Watanabe et al., 2000). The acetoacetyl
hosts. Human pathogens with melanized cell walls are more side chain is removed by a separate serine hydrolase, Ayg1p,
resistant to damage from the host immune response, promoting, to yield THN (Tsai et al., 2001). In an alternative pathway,
for example, resistance to phagocytosis (Thywißen et al., 2011) WdPks1 from zoopathogenic Wangiella (Exophiala) dermatitidis
and reactive oxygen species (Cunha et al., 2010) mustered in synthesizes a hexaketide, 2-acetyl-1,3,6,8-tetrahydroxynaph-
self-defense by macrophages. Inhibition of melanin production thalene (ATHN, 5), and the acetyl side chain is likewise removed
has been proposed as an effective route to controlling fungal by WdYg1p, an Ayg1p homolog (Wheeler et al., 2008). A handful
infections in susceptible immune-compromised populations of examples of naphtho-g-pyrone synthases have been bio-
(Liu and Nizet, 2009). In phytopathogenic fungi, melanin biosyn- chemically characterized in other fungi with wA from Aspergillus
thetic enzymes are validated targets for fungicides, which are nidulans being a paradigm example. wA synthesizes YWA1,
widely used to prevent crop damage (Wheeler and Klich, which is thought to be polymerized directly into melanin-related
Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved 1525

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
sity is achieved through ‘‘programmed’’ selection of starter unit,
polyketide chain length, regiospecificity of backbone cycliza-
tion(s), and mode of product release. Enormous progress has
been made in the last decade to understand the catalytic func-
tion of the six canonical domains that comprise these enzymes
and how they orchestrate polyketide assembly (Crawford and
Townsend, 2010). The b-ketoacyl synthase (KS), malonyl-CoA:
acyl-carrier protein transacylase (MAT), and acyl-carrier protein
(ACP) were functionally identified by their similarity to domains
from fatty acid metabolism. Like fatty acid biosynthesis, each
round of iterative extension introduces a single C2-unit through
decarboxylative thio-Claisen condensation of malonyl with the
extending chain. Intermediates are tethered as terminal thioest-
ers of a 4’-phosphopantetheinyl moiety attached to the ACP. The
N-terminal domain, dubbed the starter unit: acyl-carrier protein
transacylase (SAT), initiates synthesis with a prescribed starter
unit, typically acetyl from the acyl-CoA (Crawford et al., 2006,
2008b). The product template (PT) domain catalyzes regiospe-
cific cyclization and aromatization of the nascent ACP-bound
b-keto chain (Crawford et al., 2009; Li et al., 2010). In melanin-
type NR-PKSs connectivity between the C2 and C7 backbone
carbons produces a monocyclic intermediate, as shown in Fig-
ure 1. The C terminus constitutes a thioesterase (TE) domain
responsible for product release. The Claisen/Dieckmann cyclase
class of TE domains (TE/CLC), which catalyzes C–C bond forma-
tion in contrast to the classical hydrolytic release of fatty acids,
was originally identified in the wA naphthopyrone synthase
from A. nidulans (Fujii et al., 2001). Further insight into this class
of TEs was advanced through determination of the crystal
structure of a dissected TE/CLC monodomain from the norsolor-
inic acid anthrone synthase, PksA, from Aspergillus parasiticus
(Korman et al., 2010). Inactivating mutation or deletion of TE/
CLC domains generally results in spontaneous O-C cyclization
to release full-length pyrone shunt products (Crawford and
Townsend, 2010).
Historically, metabolites originating from NR-PKSs fit
this model of polyketide assembly; however, Pks1 from
C. lagenarium appeared to be an outlier. THN (2) was identified
as the major product from cultures of Pks1 heterologously ex-
pressed in Aspergillus oryzae (Fujii et al., 1999). Application of
14C-labeled acetyl or malonyl substrates showed no apparent
incorporation of an acetyl starter unit and that pentaketide 2
was produced solely from malonyl units (Fujii et al., 2000).
When the TE domain was inactivated, hexaketide pyrone 7
was produced instead, paralleling the synthetic logic since
shown by other NR-PKS systems (Crawford et al., 2008a; Zhang
et al., 2008). Furthermore, expression of a chimera that spliced
Pks1 to the C terminus of the wA gene at the PT-ACP junction
resulted in isolation of the nonnative hexaketide Claisen cyclase
product, ATHN (5) (Watanabe and Ebizuka, 2002). Based on the
Figure 1. Convergent Biosynthetic Routes to the THN (2) Interme-
diate in Fungal DHN (1) Melanin Biosynthesis by NR-PKSs with
Tandem ACP Domains
Pyrones (6 and 7) are released from heptaketide and hexaketide pathways
as shunt metabolites. Pathways blocked at the THN reductase, for example,
by treatment with tricyclazole, accumulate flaviolin (3), the autoxidation
product of 2.
conidial green pigment, circumventing the THN intermediate prevailing understanding of PKS catalysis at the time of this
(Watanabe et al., 1999). These fungal pathways are distinct publication, the ability to select for the pentaketide chain length
from bacterial type III THN synthases, such as RppA from was ascribed to a special function of the Pks1 TE domain.
Streptomyces griseus, where five malonyl units are condensed Subsequent experiments with NR-PKS family members,
to directly form the pentaketide as a precursor in hexa- however, support the view that chain length is determined by
hydroxyperylenequinone melanin biosynthesis (Funa et al., the KS domain, making this ‘‘TE chain length control’’ theory
2002, 2005).
untenable for Pks1 (Crawford and Townsend, 2010).
In corresponding studies, Ayg1p was identified as the
Fungal NR-PKSs are a family of multidomain enzymes that
generate substituted aromatic natural products. Structural diver- side-chain hydrolase that removes acetoacetyl from YWA1 in
1526 Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
Figure 2. Overview of Syntheses for Melanin
Standards
(A) Synthesis of C12-pyrone 7.
(B) Synthesis of THN (2), flaviolin (3), and ATHN (5).
See also Figure S1 for HPLC separation of stan-
dards.
dures, as detailed in Figure 2 and the
Supplemental Experimental Procedures
(available online). The synthesis of ATHN
(5, Figure 2) initially proved troublesome
but was efficiently achieved by way of
double O-acylation/Fries rearrangements
to generate the asymmetric pentasubsti-
tuted phenyl ester 8.
Methoxymethyl protection of the ester
9 followed by coupling with the Weinreb
amide 10 gave the ketone 11. Enolate
cyclization, removal of the silyl group,
and oxidation led to the desired hexake-
tide pyrone 7 after removal of the methox-
ymethyl groups. Dibenzyl protection of
the ester 12 followed by Friedel-Crafts
A. fumigatus (Tsai et al., 2001). A BlastP search of the sequenced acylation afforded the acetophenone 13. Treatment with
A. oryzae genome (Machida et al., 2005) for proteins similar to methoxide in methanol yielded dibenzyl THN (14), which was
Ayg1p revealed a potential functional complement in this heter- deprotected by hydrogenolysis immediately before use for
ologous host, the transcription of which was supported under high-pressure liquid chromatography (HPLC) analysis of
standard laboratory conditions by detection of an expressed THN (2) or readily oxidized in air to flaviolin (3). The double
sequence tag (AoEST6890) from a conidial cDNA library (Akao O-acylation/Fries rearrangements noted above, selectively
et al., 2007). Furthermore, the purified Pks1 SAT monodomain afforded the asymmetric pentasubstituted phenyl ester 8.
was found to selectively transfer acetyl in disparity with Di-O-benzyl protection and Dieckmann condensation generated
labeled-precursor incorporation experiments (Crawford et al., 3-di-O-benzyl ATHN, which following hydrogenolysis gave
2008b). Taking into account these results and the precedence ATHN (5).
of fungal NR-PKSs catalyzing chain extension prior to cyclization
in the PT and TE domains, we hypothesized that C. lagenarium Protein Deconstruction and Correction of the Pks1
(and heterologous host A. oryzae) contained an enzyme with Coding Sequence
Ayg1p-like activity and that hexaketide ATHN (5) was the direct To study NR-PKSs in vitro, we have adopted a ‘‘deconstruction’’
product of Pks1 (Crawford et al., 2008b). Such a hexaketide approach based on in silico prediction of domain boundaries
precursor was originally proposed by Bardshiri and Simpson to dissect expressible, functional units from target enzymes
(1983) when feeding experiments with labeled acetate failed to (Udwary et al., 2002). This approach offers a versatile method
reveal a ‘‘starter-unit effect’’ for melanin intermediates isolated to assess domains individually or in selected sets, as exemplified
from Phialaphora lagerbergii. However, the supposition of an by the four-part reconstitution of A. parasiticus PksA activity to
Ayg1p-like activity in A. oryzae was undermined when heterolo- synthesize norsolorinic acid anthrone, the first committed inter-
gous expression of the melanin PKS from W. dermatitidis mediate in aflatoxin biosynthesis (Crawford et al., 2008a).
(WdPks1) in this host accumulated ATHN (Wheeler et al., Expression plasmids to study melanin biosynthesis by Pks1
2008), not THN as had been observed for expression of Pks1 from C. lagenarium were patterned on analogous interdomain
under similar conditions. These observations led us to suspect cleavage sites used for deconstruction of PksA. Soluble, active
that something else was occurring. We chose, therefore, to study proteins were readily achieved for all constructs, except those
Pks1 in vitro to establish the biochemical rationale for THN containing the TE domain—a domain typically overexpressed
production in C. lagenarium.
without difficulty.
The current annotation for Pks1 (GenBank accession
BAA18956) designates two short exons within the N-terminal
SAT domain and a third long exon spanning all domains. The
Pks1 protein sequence was aligned to other characterized NR-
RESULTS
Synthesis of Standards
To proceed with this study, synthetic standards of melanin PKSs for comparative analysis. Close inspection revealed that
biosynthetic intermediates and shunt metabolites were prepared all catalytic residues and secondary structural elements ap-
to confidently identify the products of in vitro Pks1 reactions. peared to be intact. Despite its correct appearance, reassess-
Most of the synthetic methods utilized were standard proce- ment of Pks1 gDNA using FGNESH gene prediction software
Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved 1527

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
Figure 3. Two-Part Multidomain Combina-
tion TE Assay: Pks1DTE Reacted with Either
No, Pks1, or wA TE
Reactions were monitored by UV-vis spectro-
photometry scanning every 10 min from 280–
650 nm. Products were separated by HPLC using
a linear gradient. THN (2) initially accumulates in
the Pks1 TE reaction, indicated by increasing
absorbance at 336 nm, and then spontaneously
oxidizes to flaviolin (3), the major product detected
by HPLC. ATHN (5) likewise spontaneously
oxidizes to 3-AF (15). Diode array spectra were
extracted for the major product peak from each
HPLC chromatogram. See also Figure S2 for
WdPks1 parent TE swap control reactions. See
also Tables S2 and S3.
Domain Swapping Reveals
a Thioesterase with Claisen
Cyclase and Deacetylase Activities
The second advantage of the deconstruc-
tion approach is that individually ex-
pressed domains can be ‘‘mixed and
matched,’’ offering a flexible platform for
domain swapping among different NR-
PKSs. In an initial set of reactions, Pks1
SAT–KS–MAT–PT–ACP–ACP (Pks1DTE)
was combined with either the Pks1 or
wA TE monodomain. Reaction progress
was monitored by UV-vis spectrophotom-
etry, and products were separated by
reverse-phase HPLC and identified by
comparison to synthetic standards (Fig-
ure 3). The Pks1 TE reaction rapidly accu-
mulated an absorbance peak at 336 nm,
(http://linux1.softberry.com/berry.phtml) exposed an alternative indicative of THN production (not ATHN as predicted if a second
protein sequence in which the final 17 residues of exon 3 were Ayg1p-like enzyme were invoked). THN (2) spontaneously
replaced by a short 14-amino-acid exon 4. Constructs based converts to the reddish-purple oxidation product flaviolin (3),
on this modified cDNA resulted in soluble expression of the the major species observed by HPLC. Trace amounts of C12
-
functional TE domain. Furthermore, transcript analysis of the pyrone 7 and ATHN (5) were also detected. In contrast, the wA
region straddling the in-silico-predicted pks1 exon 3-4 junction TE reaction developed a lmax at 411 nm, producing mainly 5 to
confirmed that this revised splicing pattern is observed in mirror the product profile observed previously for the full-length
C. lagenarium mRNA, in amendment of the original gene struc- Pks1–wA chimera expressed in A. oryzae (Watanabe and Ebi-
ture report (Takano et al., 1995). In the first study of Pks1 activity, zuka, 2002). ATHN spontaneously oxidizes to 3-acetyl-flaviolin
the cDNA clone used for heterologous expression in A. oryzae (3-AF, 15) over time, albeit at a slower rate than THN oxidation
was based on the original three-exon gene structure but addi- (Wheeler et al., 2008). The control reaction with no TE gradually
tionally contained an extra genomic sequence at its 30 terminus accumulated 7, formed by spontaneous O–C cyclization of the
that encompassed the proposed fourth exon (Fujii et al., 1999). monocyclic ACP-bound product of the PT domain.
The heterologous host A. oryzae was able to complement native
The complementary experiment was conducted using the
mRNA splicing to afford the active enzyme. Additional transcript deconstructed SAT–KS–MAT and PT–ACP–ACP proteins from
evidence in support of a revised gene structure was identified in the ATHN synthase WdPks1. When combined with either no
the melanin PKS (GenBank accession CCF45141) from the TE, its cognate WdPks1 TE, or Pks1 TE, results closely paralleled
closely related species Colletotrichum higginsianum (Ushimaru those for the Pks1 parent enzyme reactions and produced C12
-
et al., 2010), which has a TE domain that shares 94.4% protein pyrone 7, ATHN (5), and THN (2), respectively (Figure S2, avail-
sequence identity with Pks1. Expressed sequence tags reported able online). Both WdPks1 and Pks1, therefore, produce chemi-
for ChPks1 (CH063_03518, Broad Institute) in the Colletotrichum cally equivalent monocyclic intermediates, and their TE domains
genome database confirm that the C. higginsianum gene tran- alone are responsible for the alternative synthetic outcomes.
script contains the same four-exon structure validated here for
C. lagenarium Pks1.
Taken together, these results verify that Pks1 is a hexaketide
synthase, but with an unprecedented bifunctional TE possessing
1528 Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
Table 1. Pks1 Prefers Acetyl Starter Units
Observed
[M+H]+
Relative
Intensity
Observed
[M+Na]+
Relative
Intensity
Product
Peak
M
3-AF (15)
249.0390
0.61
271.0207
272.0247
0.64
M+1
250.0426
1
1
ESI-MS detection of 13C-labeled acetyl starter unit incorporation into
ATHN (5) for the Pks1DTE + wA-TE reaction, measured as peak intensi-
ties for the autoxidation product 3-AF (15).
both Claisen cyclase and deacetylase activities to release the
truncated pentaketide product THN (2) directly. Pks1 is a true
THN synthase and does not require a second hydrolase for poly-
ketide shortening as has been described for melanin biosyn-
thesis in A. fumigatus (Tsai et al., 2001) and W. dermatitidis
(Wheeler et al., 2008).
Incorporation of 13C-acetyl Starter Units
Loss of acetyl by TE-catalyzed cleavage accounts for the
absence of a ‘‘starter-unit effect’’ in the biosynthesis of THN
from classical precursor labeling studies (Bardshiri and Simp-
son, 1983). To confirm that the starter unit of Pks1 was indeed
acetyl, as indicated by the substrate preference of the excised
SAT domain (Crawford et al., 2008b), incorporation of [1-13C]-
Figure 4. Additional Substrates Tested for Side-Chain Removal by
the Pks1 TE Monodomain
No conversion was detected for any compound.
acetate was monitored by mass spectrometry for the Pks1 TE 2008). The present study, corroborated in earlier work by
and wA TE in vitro reactions. As expected, no mass shift was Ebizuka and Fujii (Fujii et al., 2000), confirms that Pks1 is
observed for flaviolin because the starter unit is removed; a THN synthase. Other PKSs included in the sequence analysis
however, the M+1 species indicative of labeled acetyl incorpora- have been linked to pigment biosynthesis through gene disrup-
tion was observed as the major peak (M:M+1, 0.6:1) for ATHN (5), tion, heterologous expression, or their presence in biosynthetic
detected as 3-AF (15) (Table 1).
gene clusters with confirmed melanin pathway enzymes. The
The unenriched products presumably contain starter units direct products of these PKSs are largely undefined; however,
arising from decarboxylation of unlabeled malonyl units intro- phylogenetic analysis demonstrates that TEs of the same known
duced by the MAT domain and can be a physiologically relevant function group within a clade (Figure 5A). Sequences labeled by
source of chain initiators. The ability of Pks1 to access acetyl an asterisk were modified to reflect revised mRNA splicing
starter units from either SAT or MAT domains could improve based on in silico FGNESH predictions, which yielded proteins
the overall catalytic efficiency of this system.
more similar to melanin-producing NR-PKSs than their current
annotation. The proposed protein sequence corrections are
provided in Figure S3 and result in minor changes to their C
Deacetylase Activity Is Coupled to Claisen Cyclization
ATHN (5) treated with the Pks1 TE monodomain or full-length termini.
protein was not converted to THN (2), signifying that deacetylase A homology model of Pks1 TE was prepared using the recently
˚
activity was coupled to Claisen cyclization at the active site. The determined 1.7 A crystal structure of the PksA TE domain
hexaketide pyrone 7 was also inert to ring opening and Claisen (Protein Data Bank [PDB] ID code 3ILS) as the template, which
cyclization by the Pks1 TE. In separate experiments, deacetyla- shares 37% sequence identity and 53% sequence similarity
tion or deformylation of various monocyclic and bicyclic acetyl- with the Pks1 TE (Figure 5B) (Korman et al., 2010). In comparison
and aldehyde-containing substrates, pictured in Figure 4, like- to Pks1, the PksA TE accepts a bicyclic C20 substrate with a hexyl
wise failed to occur.
starter-unit side chain and catalyzes an analogous Claisen cycli-
zation to close the third ring of norsolorinic acid anthrone (Craw-
ford et al., 2008a). The TEs adopt an a/b-hydrolase fold (Nardini
and Dijkstra, 1999) with a flexible lid insert between b6 and b7 of
Primary Sequence and Structural Comparisons of
Melanin-Related Thioesterase Domains
We elected to examine in greater detail which features of the the canonical structure comprised of a helix-loop-helix motif that
Pks1 TE were important for its extra deacetylase function. The caps an internal binding chamber. Following transesterification
protein sequence of Pks1 was compared to other NR-PKSs of the PT-cyclized intermediate from ACP to TE, lid closure is
involved in melanin and conidial pigment biosynthesis. Several thought to facilitate exclusion of water and side-chain localiza-
YWA1-producing PKSs have been functionally characterized, tion to favor C-C bond formation catalyzed by a Ser-His-Asp
including A. fumigatus Alb1 (Watanabe et al., 2000), A. nidulans catalytic triad (Korman et al., 2010). Several lid residues line
wA (Watanabe et al., 1999), Aspergillus niger FwnA (Chiang the roof of the PksA TE active site and are indicated by blue trian-
et al., 2011; Jørgensen et al., 2011), and Gibberella zeae Pks12 gles in the alignment of this region depicted in Figure 5C. Owing
(Frandsen et al., 2011). WdPks1 from W. dermatitidis is the to high sequence and predicted secondary structure similarity
sole validated example of an ATHN synthase (Wheeler et al., among NR-PKSs, the lid is almost certainly a conserved feature
Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved 1529

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
Figure 5. Comparative Analyses of TE Domains
(A) Phylogenetic tree of NR-PKSs with confirmed involvement in melanin or conidial pigment biosynthesis. The direct product of the PKS, when known, is
indicated by either a blue square (THN), red circle (ATHN), or green diamond (YWA1). Bootstrap values are shown next to the branch points. See also Table S1.
(B) Homology model of Pks1 TE based on the PksA TE structure PDB ID code ILS3. The lid region used in chimera swaps is colored blue, and the catalytic triad is
shown as balls and sticks colored by element.
(C) Sequence alignment of the lid region with PksA numbering. Hydrophobic residues lining the PksA active site are indicated by blue triangles. Arrows indicate lid
loop (circle) and complete lid (diamond) regions swapped in the chimeras. The black boxes surrounding Pks1 and WdPks1 residues indicate the regions mutated
in helix swaps. See also Figure S3 for the full TE sequence alignment.
(D) Product distribution of three-part TE assay reactions organized by parent TE. Labels indicate percent peak area for each product. Error bars show standard
deviation from duplicate reactions. LL, Lid loop; L, Lid; Cl, C. lagenarium Pks1; wA, A. nidulans wA; Wd, W. dermatitidis WdPks1. See also Tables S2 and S3.
of this subfamily of TEs. The full-length domain alignment of protein. Because the smaller proteins are expressed in greater
PksA with melanin-type TEs is provided in Figure S3.
yields, these fragments were used to assay the mutant TEs.
Acetyl and malonyl N-acetyl cysteamine (SNAC) thioesters
were used as effective alternative substrates to acyl-CoA mono-
Lid Swap Chimera Alter TE Activity
Of the residues predicted to line the TE active site, the lid was mers. Reactions were monitored by UV-vis spectrophotometric
identified as a potential region of sequence divergence between scanning kinetics for detection of chromophore development
YWA1/ATHN- and THN-producing PKSs. Furthermore, other and by HPLC for profiling product distribution. Reduction in de-
a/b-hydrolase family members have proven amenable to lid acetylase activity was demonstrated by diminished production
variations and in some cases have demonstrated altered of THN (2) in favor of ATHN (5). Decline in Claisen cyclase activity
substrate preferences or reactivities (Chen et al., 2009). There- was indicated by accumulation of the C12-pyrone 7. Other
fore, a series of Pks1 and wA lid chimeras was developed by metabolites were not observed above trace amounts. Product
swapping the alternative lid sequences from the wA (YWA1- distributions of the lid chimera are summarized in Figure 5D
type), WdPks1 (ATHN-type), and Pks1 (THN-type) proteins. and represent the average of duplicate reactions. THN (2), which
Three-part Pks1 SAT–KS–MAT + PT–ACP–ACP + TE reactions elutes as a broad peak around 14 minutes, and its oxidation
accumulated THN in comparable yields to the intact full-length product flaviolin (3) were summed to give the total deacetylated
1530 Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
product. Likewise, peak areas for ATHN (5) and oxidation
product 3-AF (15) were combined.
Swapping the lid loop (LL) regions among different classes of
TEs had no significant effect on either the kinetics of chromo-
phore formation or product distribution. The loop sequence of
Pks1 is four residues longer than either wA or WdPks1 but was
ruled out here as a factor in deacetylation catalysis. Inclusion
of the LL-flanking aL1 and aL2 helices in the lid chimeras caused
a major shift in the distribution of products. Within the Pks1
parent, the WdPks1 lid efficiently complemented Claisen cyclase
activity but showed a 78% reduction in THN relative to wild-type
with a concomitant increase in ATHN. The Pks1 lid-wA variant
displayed reduced Claisen efficiency (pyrone was the favored
product at 82% total peak area) and complete side-chain reten-
tion (with no THN detected). The inverse experiment, where the
Pks1 lid was integrated into the wA parent, was insufficient to
turn on deacetylase activity and produced virtually no Claisen
products, releasing the pyrone almost exclusively at 98% of
the total.
The Pks1 lid-Wd chimera was inspected more closely
because of its reduced ability to catalyze deacetylation without
any apparent loss in Claisen cyclization. The aL1 and aL2 helices
were individually mutated to the WdPks1 sequence (boxed resi-
dues in Figure 4C), resulting in changes to six amino acids per
helix. Product profiles reflect the importance of aL1 for efficient
Figure 6. Proposed Mechanisms for Product Partitioning in WdPks1
and Pks1 TEs, which Synthesize ATHN (5) and THN (2), Respectively
The acetyl starter unit is shown in bold.
removal of the acetyl side chain, which had a distribution virtually DISCUSSION
identical to the complete lid swap, whereas aL2 produced mainly
THN (88%).
No additional catalytic machinery is necessarily required to
Next, the hexaketide pathway homolog TE from WdPks1 was cleave the acetyl pendant by Pks1 TE. Whereas the discrete
examined to see if side-chain removal could be activated in serine hydrolases from A. fumigatus (Ayg1p) and W. dermatitidis
a closer functional relative. Surprisingly, some THN was innately (WdYg1p) melanin pathways can act on YWA1 and ATHN,
produced by the wild-type WdPks1 TE, accounting for 13% of respectively, Pks1 TE showed no conversion of any free
the product output. Substitution with the Pks1 lid enhanced substrate tested for side-chain removal. Lack of activity toward
this basal deacetylase activity. Even though significant product synthetic ATHN implies that TE acts on the immediate product
was siphoned into pyrone derailment, the THN peak area of the Claisen cyclase event, perhaps prior to aromatization to
increased 2-fold over native Wd levels, and the ratio of Claisen the naphthalene core, where overcoming aromatic stabilization
products, THN:ATHN, shifted from 1:6 to 2:1 for this variant.
would hinder side-chain hydrolysis, as depicted in Figure 6. In
Complementation of Claisen cyclization between wA (YWA1- lieu of reprotonation to restore the resting state of the catalytic
type) and WdPks1 (ATHN-type) TEs was also considered. The triad, the serine residue could attack into the side-chain carbonyl
wA lid catalyzed C-C bond formation in the WdPks1 TE with to release THN and acetic acid as products, analogous to b-di-
a marginal increase in the shunt O-C population. The Wd lid ketide hydrolases (E.C.3.7.1.X), of which Ayg1p and WdYg1p
partially complemented cyclase function in the wA TE, releasing are members (Grogan, 2005). The absence of water in the active
a mixture of ATHN (34%) and pyrone (65%).
site of the closed-lid structure of PksA TE (Korman et al., 2010)
In some cases generation of the pyrone by TE lid chimera suggests that side-chain attack is direct rather than water medi-
reactions was much greater than either no TE or Pks1 TE ated by a general base mechanism in Pks1. Although specula-
S2009A loss-of-function mutant controls, where transesterifica- tive, we hypothesize that product partitioning between ATHN
tion of the monocyclic intermediate from ACP to TE cannot and THN is kinetically controlled, based on the fact that both
occur. Spontaneous release of this intermediate from the ACP Pks1 TE and WdPks1 TE synthesize both products but in highly
was extremely slow, and the enzyme likely wields a protective skewed, opposing ratios. According to this model, nucleophilic
role in slowing the favorable O-C cyclization mode. In contrast, attack at C11 would out-compete reprotonation of S2009 to
the wA lid-Pks1 chimera, for example—which was deficient in shift the equilibrium of product formation toward the retro-
Claisen cyclase activity but could presumably still accept the Claisen products in Pks1. Enzyme-assisted stabilization of the
monocyclic substrate—accumulated much higher amounts of prearomatic ‘‘diketide’’ intermediate may further promote this
the pyrone. The foundation for this observed phenomenon could transformation.
be passive, that is, through reduced protection of the monocy-
In an alternative mechanism, deacetylation (retro-Claisen) of
clic intermediate while covalently bound to Claisen-deficient the TE-bound monocyclic intermediate may occur prior to
TEs, or active, through TE-conferred catalytic release of the Claisen cyclization (Dieckmann) with the instantly formed enolate
pyrone. Our experiments did not differentiate between these closing directly to THN. We consider this scenario, however,
alternatives.
less likely because a suitable base could not be identified in
Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved 1531

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
the homology model of the Pks1 TE in the region where the clic hexaketide precursor, explaining the absence of an
side chain is presumed to reside. Structural analyses of observable ‘‘starter-unit effect’’ in this system. Domain and
closely related melanin-TEs might illuminate how small changes secondary structural element swapping experiments could
in their active-site geometries could contribute to product provide a valid route to engineering NR-PKS pathways to
partitioning.
expand the biosynthetic potential of these enzymes and in
Domain-swapping exercises illustrate the interchangeability of understanding the molecular foundation of melanogenesis,
NR-PKS domains among closely related synthases. The wA TE, where it may aid efforts to discover new, effective antifungal
which accepts a native C14-ACP-bound substrate, proficiently therapies. The unusual ability of Pks1 TE to catalyze tandem
catalyzed Claisen cyclization of the Pks1-generated C12 interme- carbon-carbon bond making and breaking reactions in
diate. This promiscuity in side-chain binding indicates that the a single enzymatic step further expands the diverse array
precise identity of the terminal acetyl group of the YWA1 TE- of reactions known for a/b-hydrolase fold enzymes gener-
bound precursor is not the sole determinant for orienting the ally (Carr and Ollis, 2009) and the repertoire of activities
side chain for Claisen/Dieckmann C–C bond formation in wA. exemplified by fungal TE domains in particular.
Furthermore, secondary structural motifs within domains have
proven adaptable to exchange and are useful in deciphering
EXPERIMENTAL PROCEDURES
which protein features are responsible for specific activities
Synthesis of Melanin Standards
and ultimately natural product programming. Through genera-
tion of TE chimera, we have shown that the lid region, specifically
the aL1 helix, is a factor in maintaining the active-site topography
of Pks1 TE necessary to catalyze efficient deacetylation, a previ-
ously unknown function for NR-PKS TE domains. Another recent
example where this approach was successfully applied involved
structural swaps between ketoreductase domains from resor-
cylic acid macrolactone synthases, members of the highly
reducing class of fungal iterative PKSs, to understand the ratio-
nale for differential stereochemistries of their product hydroxyls
(Zhou et al., 2012).
All synthetic details are provided in the Supplemental Experimental
Procedures.
C. Orbiculare cDNA Analysis
C. orbiculare NRRL28842 was cultured in potato-sucrose liquid medium sup-
plemented with 0.2% yeast extract, and mycelia were transferred to 1.2 M
sucrose for 20 hr as previously described to rapidly induce transcription of
melanin biosynthetic pathway genes (Okamoto et al., 2001; Takano et al.,
1995). Mycelia were harvested by vacuum filtration over Miracloth (EMD
Millipore, Billerica, MA, USA), frozen in liquid N2, and stored at ꢀ80ꢁC until
use. Total RNA was harvested from mycelia ground in liquid N2 using an
RNeasy Plant Mini Kit (QIAGEN, Valencia, CA, USA) as recommended by the
manufacturer with the on-column DNase digestion step. First-strand cDNA
was prepared from 1 mg total RNA and random hexamer primer (IDT, Coralville,
IA, USA) using M-MLV reverse transcriptase (Promega, Madison, WI, USA) in
a 20 ml volume. A reaction lacking reverse transcriptase was used as a negative
control. The cDNA synthesis reaction was used directly as a template for
amplification of the region spanning the proposed exon 4 in pks1 using primer
set Pks1e4check-5.1 50-GAAGGCGAGACAGTCCAAAG-30 and Pks1e4check-
Stepwise production of THN via an isolable acyl side-chain
intermediate (YWA1 or ATHN), leads to greater potential for
synthetic diversity. In A. niger, for example, FwnA produces
naphtho-g-pyrones, including YWA1, that can be directly incor-
porated into natural products (like the dimeric aurosperones) or,
through selective use of the Ayg1p homolog OlvA, can be
directed into the DHN melanin pathway (Chiang et al., 2011).
Alternatively, C. lagenarium (and likely many other melanin NR-
PKSs that are closely related) has a bifunctional TE domain
that catalyzes both Claisen cyclization and deacetylation.
Melanin is important for fungal survival and longevity, as well
as virulence in many pathogenic species. Depending on the
selective pressures exerted on different fungal strains, more effi-
cient THN biosynthesis may have evolved in response to envi-
ronmental factors necessitating greater proficiency in cell-wall
pigmentation at the expense of product diversity.
3.1 50-TAAGAGTGTGCATGATCTCGC-30. PCRs contained
2 ml cDNA,
200 mM deoxynucleoside triphosphates, 0.4 mM each primer, 3% dimethyl
sulfoxide, and 2.5 units OneTaq DNA Polymerase (New England BioLabs,
Ipswich, MA, USA) in manufacturer-provided buffer in 50 ml volume. A touch-
down PCR method was employed. The thermocycler conditions were 94ꢁC
0:15 (melt), 60ꢁC to 52ꢁC 0:30 (1ꢁC/cycle incremental temperature drop,
anneal), and 68ꢁC 1:20 (extend) for eight cycles; the annealing temperature
was maintained at 52ꢁC for an additional 20 cycles. Total RNA and PCR-
amplified cDNA were analyzed by agarose gel electrophoresis. The cDNA
amplicon was subcloned into pCRII-TOPO using a TOPO TA Cloning Kit
(Invitrogen, Grand Island, NY, USA), and purified plasmids from four colonies
were individually sequenced to confirm the revised splicing prediction. The
cDNA inserts were aligned in Sequencher v4.0 (Gene Codes Corporation,
Ann Arbor, MI, USA), and the contig is provided in the Supplemental Experi-
mental Procedures.
SIGNIFICANCE
Fungal melanins provide survival advantages in harsh envi-
ronments. They are also virulence factors for many patho-
genic fungi, where pigment disruption often renders these
strains noninfective. Similar heptaketide and hexaketide
pathways have been described that utilize separate deace-
tylases to remove the acetoacetyl and acetyl side chains of
their respective bicyclic substrates released by standard
C-terminal thioesterase/Claisen cyclase domains of NR-
PKSs. In vitro biochemical characterization confirms that
Pks1 efficiently synthesizes THN unaided by a second
enzyme, defining an efficient third fungal route to this impor-
tant melanin precursor. Pks1 TE is both a Claisen cyclase
and deacetylase that processes an enzyme-bound monocy-
Cloning and Expression
DNA manipulations were carried out using standard methods (Sambrook and
Russell, 2001). wA and Pks1 fragments were amplified from A. nidulans
RLMH67 and C. orbiculare (syn. C. lagenarium) NRRL 28842 gDNA, respec-
tively. The W. dermatitidis TE coding sequence DNA was purchased from Gen-
Script (Piscataway, NJ, USA). Point mutants and lid chimeras were generated
by overlap extension PCR using wild-type TE constructs as template DNA.
Subsequently, additional WdPks1 fragments were amplified from
W. dermatitidis UT8656 gDNA. Further cloning details are summarized in the
Supplemental Experimental Procedures (Tables S2 and S3). Recombinant
proteins were heterologously expressed from BL21(DE3) in Luria-Bertani broth
medium and purified by Ni2+-affinity chromatography as previously described
using standard methods (Crawford et al., 2008a).
1532 Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
In Vitro Reactions
sequence pair. Evolutionary distances were computed using the Poisson
correction method.
Reactions were carried out at room temperature with 10 mM protein (each) in
reaction buffer (100 mM K/PO4 pH 7.0 and 1 mM tris[2-carboxyethyl]phos-
phine [TCEP]) and were initiated by the addition of a substrate mix containing
acetyl and malonyl as the CoA or SNAC thioesters. Reactions were monitored
by UV-vis spectrophotometry using scanning kinetics from 280–650 nm on
a Cary 50 instrument (Varian Inc., Walnut Creek, CA, USA). Products were ex-
tracted into ethyl acetate or filtered through a 10k MWCO Microcon centrifugal
filter device (Millipore, Billerica, MA, USA) for HPLC analysis. Products (100 ml
injections) were separated by reverse-phase HPLC (Agilent 1200, Santa Clara,
CA, USA) using a bisolvent linear gradient over a Prodigy ODS3 column (4.6 3
250 mm, 5 mm, Phenomenex, Torrence, CA, USA). Solvent A: H2O + 0.1%
formic acid; Solvent B: acetonitrile + 0.1% formic acid; 5% to 80% B over
30 minutes with a flow rate of 1 ml/minute.
SUPPLEMENTAL INFORMATION
Supplemental Information includes three figures and three tables and can
be found with this article online at http://dx.doi.org/10.1016/j.chembiol.2012.
10.002.
ACKNOWLEDGMENTS
We thank Prof. J.M. Crawford (Yale University, New Haven, CT, USA) for
preliminary DNA work and helpful discussions. Drs. S.A. Rehner (USDA-
ARS, Beltsville, MD, USA) and K. O’Donnell (USDA, Peoria, IL, USA), Prof.
N. P. Keller (University of Wisconsin, Madison, WI, USA), Dr. Z. Wang (Naval
Research Laboratory, Washington, DC, USA), and Prof. P.J. Szaniszlo (Univer-
sity of Texas, Austin, TX, USA) kindly provided spores and/or gDNA from
C. orbiculare, A. nidulans, and W. dermatitidis, respectively. We are grateful
to Katherine Belecki for collecting mass spectrometry data and to K.L. Fiedler
and Prof. R.J. Cotter of the Middle Atlantic Mass Spectrometry Laboratory at
the Johns Hopkins School of Medicine and the Mid-Atlantic Regional Office of
Shimadzu Scientific Instruments, Inc. (Columbia, MD, USA) for use of their Shi-
madzu LC-IT-TOF. This work was supported by the National Institutes of
Health Grant ES 001670.
In initial TE swaps, Pks1DTE was combined with either no TE, Pks1 TE, or
wA TE with 0.5 mM acetyl-CoA and 2 mM malonyl-CoA substrates. After
4 hr, reactions were filtered through Microcon centrifugal filter devices and
analyzed by HPLC. Full-length Pks1 and three-part (SAT–KS–MAT + PT–
ACP–ACP + TE) combinations were likewise tested.
Wild-type, point mutant, and lid-chimera TEs were assayed in three-part
multidomain combinations with Pks1 SAT–KS–MAT and PT–ACP–ACP using
0.25 mM acetyl-SNAC and 1 mM malonyl-SNAC substrates in reaction buffer
additionally containing 10% glycerol. Reactions were monitored by UV-vis
spectrophotometry for 2 hr and filtered through Microcon centrifugal filter
devices. Products were separated by HPLC, and the resulting chromatograms
(A254 nm) integrated to calculate peak areas. Control reactions with WdPks1
SAT–KS–MAT and PT–ACP–ACP were similarly conducted.
Received: April 24, 2012
Revised: September 25, 2012
Accepted: October 2, 2012
Published: December 20, 2012
Deacetylase/Deformylase Assay
ATHN (5), C12-pyrone 7, and a variety of other acetyl and aldehyde side-chain
compounds (Figure 5, 0.5 mM) were separately incubated with the Pks1 TE or
S2009A mutant control (10 mM) overnight. Reactions were filtered through
Microcon centrifugal filter devices and analyzed by HPLC. ATHN (5) was
also tested as a substrate for full-length Pks1.
REFERENCES
Akao, T., Sano, M., Yamada, O., Akeno, T., Fujii, K., Goto, K., Ohashi-Kunihiro,
S., Takase, K., Yasukawa-Watanabe, M., Yamaguchi, K., et al. (2007). Analysis
of expressed sequence tags from the fungus Aspergillus oryzae cultured under
different conditions. DNA Res. 14, 47–57.
Bardshiri, E., and Simpson, T.J. (1983). 13C and 2H labelling studies on the
biosynthesis of scytalone in Phialaphora lagerbergll. Tetrahedron 39, 3539–
3542.
Starter Unit Incorporation Assay
Three-part in vitro reactions containing Pks1 SAT–KS–MAT, Pks1 PT–ACP–
ACP, and either Pks1 TE or wA TE were initiated with 1 mM [1-13C]-acetyl-
SNAC and 1 mM malonyl-SNAC. Following overnight incubation, 500 ml reac-
tions were quenched with 10 ml HCl and extracted twice with 1 ml ethyl acetate.
Pooled extracts were evaporated to dryness and dissolved in 250 ml 50%
Bell, A.A., and Wheeler, M.H. (1986). Biosynthesis and functions of fungal
aqueous acetonitrile. Products were analyzed by HR-ESI-LCMS on
a
melanins. Annu. Rev. Phytopathol. 24, 411–451.
Shimadzu LC-IT-TOF over a Luna C18(2) analytical column (2.0 3 150 mm,
3 mm, Phenomenex) using a linear gradient of 20% to 80% B over 30 minutes
with a flow rate of 0.2 ml/minute. Positive ion peak intensities were compared
for the M and M+1 peaks using LCMS Solutions v3.50 software (Shimadzu,
Columbia, MD, USA).
Carr, P.D., and Ollis, D.L. (2009). a/b hydrolase fold: an update. Protein Pept.
Lett. 16, 1137–1148.
Chen, B., Cai, Z., Wu, W., Huang, Y., Pleiss, J., and Lin, Z. (2009). Morphing
activity between structurally similar enzymes: from heme-free bromoperoxi-
dase to lipase. Biochemistry 48, 11496–11504.
Chiang, Y.-M., Meyer, K.M., Praseuth, M., Baker, S.E., Bruno, K.S., and Wang,
C.C. (2011). Characterization of a polyketide synthase in Aspergillus niger
whose product is a precursor for both dihydroxynaphthalene (DHN) melanin
and naphtho-g-pyrone. Fungal Genet. Biol. 48, 430–437.
Homology Model of Pks1 TE
A homology model for the Pks1 TE domain was generated with the loopmodel
algorithm of MODELER within the Accelrys Discovery Studio Suite v2.1
(Accelrys Inc., San Diego, CA, USA). The PksA TE domain (PDB ID code
3ILS) was set as the template structure. The model was built based on ClustalX
alignments of 37 NR-PKS TE sequences, and overhangs were cut from the
sequences.
Crawford, J.M., and Townsend, C.A. (2010). New insights into the formation of
fungal aromatic polyketides. Nat. Rev. Microbiol. 8, 879–889.
Crawford, J.M., Dancy, B.C.R., Hill, E.A., Udwary, D.W., and Townsend, C.A.
(2006). Identification of a starter unit acyl-carrier protein transacylase domain
in an iterative type I polyketide synthase. Proc. Natl. Acad. Sci. USA 103,
16728–16733.
Sequence Analyses
A list of validated NR-PKSs involved in melanin and conidial pigment bio-
synthesis was compiled (Table S1). Analyses were conducted in MEGA5
(Tamura et al., 2011) and involved 29 proteins. Sequence alignment was
completed with ClustalX2 using default settings. The gDNA from potentially
mismatched protein sequences was submitted to FGNESH gene prediction
software (http://linux1.softberry.com/berry.phtml) to consider alterations to
splicing patterns. In cases in which better sequence similarity to other NR-
PKSs was achieved by revised splicing, these sequences were included in
lieu of the current annotation. The evolutionary history was inferred using
the neighbor-joining method and a bootstrap consensus tree was composed
from 1,000 replicates. All ambiguous positions were removed for each
Crawford, J.M., Korman, T.P., Labonte, J.W., Vagstad, A.L., Hill, E.A., Kamari-
Bidkorpeh, O., Tsai, S.-C., and Townsend, C.A. (2009). Structural basis for
biosynthetic programming of fungal aromatic polyketide cyclization. Nature
461, 1139–1143.
Crawford, J.M., Thomas, P.M., Scheerer, J.R., Vagstad, A.L., Kelleher, N.L.,
and Townsend, C.A. (2008a). Deconstruction of iterative multidomain polyke-
tide synthase function. Science 320, 243–246.
Crawford, J.M., Vagstad, A.L., Whitworth, K.P., Ehrlich, K.C., and Townsend,
C.A. (2008b). Synthetic strategy of nonreducing iterative polyketide synthases
Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved 1533

Chemistry & Biology
Thioesterase as Claisen Cyclase and Deacetylase
and the origin of the classical ‘‘starter-unit effect’’. ChemBioChem 9, 1019–
melanin biosynthesis by Colletotrichum lagenarium. Microbiology 147, 2623–
1023.
2628.
Cunha, M., Franzen, A., Seabra, S., Herbst, M., Vugman, N., Borba, L.,
de Souza, W., and Rozental, S. (2010). Melanin in Fonsecaea pedrosoi:
a trap for oxidative radicals. BMC Microbiol. 10, 80.
Pihet, M., Vandeputte, P., Tronchin, G., Renier, G., Saulnier, P., Georgeault, S.,
Mallet, R., Chabasse, D., Symoens, F., and Bouchara, J.-P. (2009). Melanin is
an essential component for the integrity of the cell wall of Aspergillus fumigatus
conidia. BMC Microbiol. 9, 177.
Frandsen, R.J.N., Schu¨ tt, C., Lund, B.W., Staerk, D., Nielsen, J., Olsson, S.,
and Giese, H. (2011). Two novel classes of enzymes are required for the
biosynthesis of aurofusarin in Fusarium graminearum. J. Biol. Chem. 286,
10419–10428.
Sambrook, J., and Russell, D.W. (2001). Molecular Cloning: A Laboratory
Manual (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press).
Takano, Y., Kubo, Y., Shimizu, K., Mise, K., Okuno, T., and Furusawa, I. (1995).
Structural analysis of PKS1, a polyketide synthase gene involved in melanin
biosynthesis in Colletotrichum lagenarium. Mol. Gen. Genet. 249, 162–167.
Fujii, I., Mori, Y., Watanabe, A., Kubo, Y., Tsuji, G., and Ebizuka, Y. (1999).
Heterologous expression and product identification of Colletotrichum lagena-
rium polyketide synthase encoded by the PKS1 gene involved in melanin
biosynthesis. Biosci. Biotechnol. Biochem. 63, 1445–1452.
Takano, Y., Kubo, Y., Kawamura, C., Tsuge, T., and Furusawa, I. (1997). The
Alternaria alternata melanin biosynthesis gene restores appressorial melaniza-
tion and penetration of cellulose membranes in the melanin-deficient albino
mutant of Colletotrichum lagenarium. Fungal Genet. Biol. 21, 131–140.
Fujii, I., Mori, Y., Watanabe, A., Kubo, Y., Tsuji, G., and Ebizuka, Y. (2000).
Enzymatic synthesis of 1,3,6,8-tetrahydroxynaphthalene solely from malonyl
coenzyme
A by a fungal iterative type I polyketide synthase PKS1.
Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., and Kumar, S.
(2011). MEGA5: molecular evolutionary genetics analysis using maximum like-
lihood, evolutionary distance, and maximum parsimony methods. Mol. Biol.
Evol. 28, 2731–2739.
Biochemistry 39, 8853–8858.
Fujii, I., Watanabe, A., Sankawa, U., and Ebizuka, Y. (2001). Identification of
Claisen cyclase domain in fungal polyketide synthase WA, a naphthopyrone
synthase of Aspergillus nidulans. Chem. Biol. 8, 189–197.
Thywißen, A., Heinekamp, T., Dahse, H.-M., Schmaler-Ripcke, J., Nietsche,
S., Zipfel, P.F., and Brakhage, A.A. (2011). Conidial dihydroxynaphthalene
melanin of the human pathogenic fungus Aspergillus fumigatus interferes
with the host endocytosis pathway. Front. Microbiol. 2, 96.
Funa, N., Ohnishi, Y., Ebizuka, Y., and Horinouchi, S. (2002). Properties and
substrate specificity of RppA, a chalcone synthase-related polyketide syn-
thase in Streptomyces griseus. J. Biol. Chem. 277, 4628–4635.
Funa, N., Funabashi, M., Ohnishi, Y., and Horinouchi, S. (2005). Biosynthesis of
hexahydroxyperylenequinone melanin via oxidative aryl coupling by cyto-
chrome P-450 in Streptomyces griseus. J. Bacteriol. 187, 8149–8155.
Tsai, H.-F., Fujii, I., Watanabe, A., Wheeler, M.H., Chang, Y.C., Yasuoka, Y.,
Ebizuka, Y., and Kwon-Chung, K.J. (2001). Pentaketide melanin biosynthesis
in Aspergillus fumigatus requires chain-length shortening of a heptaketide
precursor. J. Biol. Chem. 276, 29292–29298.
Gladieux, P., Byrnes Iii, E.J., Aguileta, G., Fisher, C.M., Heitman, J., and
Giraud, T. (2011). Epidemiology and evolution of fungal pathogens in plants
and animals. In Genetics and Evolution of Infectious Disease, M. Tibayrenc,
ed. (London: Elsevier), pp. 59–132.
Tsuji, G., Tsuge, S., Shiraishi, T., and Kubo, Y. (2003). Expression pattern of
melanin biosynthesis enzymes during infectious morphogenesis of
Colletotrichum lagenarium. J. Gen. Plant Pathol. 69, 169–175.
Grogan, G. (2005). Emergent mechanistic diversity of enzyme-catalysed beta-
Udwary, D.W., Merski, M., and Townsend, C.A. (2002). A method for prediction
of the locations of linker regions within large multifunctional proteins, and
application to a type I polyketide synthase. J. Mol. Biol. 323, 585–598.
diketone cleavage. Biochem. J. 388, 721–730.
Jacobson, E.S. (2000). Pathogenic roles for fungal melanins. Clin. Microbiol.
Rev. 13, 708–717.
Ushimaru, T., Terada, H., Tsuboi, K., Kogou, Y., Sakaguchi, A., Tsuji, G., and
Kubo, Y. (2010). Development of an efficient gene targeting system in
Colletotrichum higginsianum using a non-homologous end-joining mutant
and Agrobacterium tumefaciens-mediated gene transfer. Mol. Genet.
Genomics 284, 357–371.
Jørgensen, T.R., Park, J., Arentshorst, M., van Welzen, A.M., Lamers, G.,
Vankuyk, P.A., Damveld, R.A., van den Hondel, C.A.M., Nielsen, K.F.,
Frisvad, J.C., and Ram, A.F. (2011). The molecular and genetic basis of
conidial pigmentation in Aspergillus niger. Fungal Genet. Biol. 48, 544–553.
Korman, T.P., Crawford, J.M., Labonte, J.W., Newman, A.G., Wong, J.,
Townsend, C.A., and Tsai, S.-C. (2010). Structure and function of an iterative
polyketide synthase thioesterase domain catalyzing Claisen cyclization in afla-
toxin biosynthesis. Proc. Natl. Acad. Sci. USA 107, 6246–6251.
Watanabe, A., and Ebizuka, Y. (2002). A novel hexaketide naphthalene synthe-
sized by a chimeric polyketide synthase composed of fungal pentaketide and
heptaketide synthases. Tetrahedron Lett. 43, 843–846.
Watanabe, A., Fujii, I., Sankawa, U., Mayorga, M.E., Timberlake, W.E., and
Ebizuka, Y. (1999). Re-identification of Aspergillus nidulans wA gene to code
for a polyketide synthase of naphthopyrone. Tetrahedron Lett. 40, 91–94.
Kubo, Y., Suzuki, K., Furusawa, I., and Yamamoto, M. (1985). Melanin biosyn-
thesis as a prerequisite for penetration by appressoria of Colletotrichum
lagenarium: site of inhibition by melanin-inhibiting fungicides and their action
on appressoria. Pestic. Biochem. Physiol. 23, 47–55.
Watanabe, A., Fujii, I., Tsai, H.-F., Chang, Y.C., Kwon-Chung, K.J., and
Ebizuka, Y. (2000). Aspergillus fumigatus alb1 encodes naphthopyrone syn-
thase when expressed in Aspergillus oryzae. FEMS Microbiol. Lett. 192, 39–44.
Langfelder, K., Streibel, M., Jahn, B., Haase, G., and Brakhage, A.A. (2003).
Biosynthesis of fungal melanins and their importance for human pathogenic
fungi. Fungal Genet. Biol. 38, 143–158.
Wheeler, M.H., and Klich, M.A. (1995). The effects of tricyclazole, pyroquilon,
phthalide, and related fungicides on the production of conidial wall pigments
by Penicillium and Aspergillus species. Pestic. Biochem. Physiol. 52, 125–136.
Li, Y., Image, I.I., Xu, W., Image, I., Tang, Y., and Image, I. (2010).
Classification, prediction, and verification of the regioselectivity of fungal poly-
ketide synthase product template domains. J. Biol. Chem. 285, 22764–22773.
Wheeler, M.H., Abramczyk, D., Puckhaber, L.S., Naruse, M., Ebizuka, Y.,
Fujii, I., and Szaniszlo, P.J. (2008). New biosynthetic step in the melanin
pathway of Wangiella (Exophiala) dermatitidis: evidence for 2-acetyl-1,3,6,8-
Tetrahydroxynaphthalene as a novel precursor. Eukaryot. Cell 7, 1699–1711.
Liu, G.Y., and Nizet, V. (2009). Color me bad: microbial pigments as virulence
factors. Trends Microbiol. 17, 406–413.
Machida, M., Asai, K., Sano, M., Tanaka, T., Kumagai, T., Terai, G., Kusumoto,
K.-I., Arima, T., Akita, O., Kashiwagi, Y., et al. (2005). Genome sequencing and
analysis of Aspergillus oryzae. Nature 438, 1157–1161.
Woo, P.C.Y., Tam, E.W.T., Chong, K.T.K., Cai, J.J., Tung, E.T.K., Ngan, A.H.Y.,
Lau, S.K.P., and Yuen, K.-Y. (2010). High diversity of polyketide synthase
genes and the melanin biosynthesis gene cluster in Penicillium marneffei.
FEBS J. 277, 3750–3758.
Nardini, M., and Dijkstra, B.W. (1999). a/b hydrolase fold enzymes: the family
keeps growing. Curr. Opin. Struct. Biol. 9, 732–737.
Zhang, W., Li, Y., and Tang, Y. (2008). Engineered biosynthesis of bacterial
aromatic polyketides in Escherichia coli. Proc. Natl. Acad. Sci. USA 105,
20683–20688.
Nosanchuk, J.D., and Casadevall, A. (2006). Impact of melanin on microbial
virulence and clinical resistance to antimicrobial compounds. Antimicrob.
Agents Chemother. 50, 3519–3528.
Zhou, H., Gao, Z., Qiao, K., Wang, J., Vederas, J.C., and Tang, Y. (2012).
A fungal ketoreductase domain that displays substrate-dependent stereo-
specificity. Nat. Chem. Biol. 8, 331–333.
Okamoto, S., Sakurada, M., Kubo, Y., Tsuji, G., Fujii, I., Ebizuka, Y., Ono, M.,
Nagasawa, H., and Sakuda, S. (2001). Inhibitory effect of aflastatin A on
1534 Chemistry & Biology 19, 1525–1534, December 21, 2012 ª2012 Elsevier Ltd All rights reserved