Cladosporium phlei (George) de Vries belongs to a group of hypomycetous fungus that causes purple eyespot disease in timothy (Phleum pratense). This disease is one of the most common foliar diseases of timothy, which is easily distinguishable by eye-shaped spots of light greyish-fawn centers with purple margins on plant leaves. C. phlei produces a characteristic deep red pigment, phleichrome, in mycelia and culture medium (Yoshihara et al., 1975). Phleichrome, a derivative of 4,9-dihydroxyperylene-3,10-quinone, is a member of a group of fungal perylenequinones. Perylenequinones that share the same 4,9-dihydroxyperylene-3,10-quinone chromophore exhibit photodynamic activities and act as photosensitizers because they contain a core chromophore of phenolic quinone that can absorb light energy and produce reactive oxygen species (ROS) such as the hydroxyl radical (OH^?), superoxide anion (O₂^??), hydrogen peroxide (H₂O₂), and singlet oxygen (¹O₂) (Daub and Ehrenshaft, 2000; Daub et al., 2005; Liao and Chung, 2008). The fact that many phytopathogenic fungi produce perylenequinone pigments suggests that perylenequinone pigments may be virulence factors because they produce ROS, which can kill host cells, thereby facilitating fungal invasion. Perylenequinone pigments are therefore considered to be light-activated, non-host-selective phytotoxins (Daub et al., 2005; Liao and Chung, 2008). Recently, perylenequinones have gained attention because of their therapeutic potential as photosensitizers for photodynamic therapy (PDT) (Hudson and Towers, 1991). Among perylenequinones, the photodynamic activity of phleichrome has been studied in detail (Olivo and Chin, 2006), and it has been used as a pharmacophore to produce various derivatives (Du and Maunder, 2010). In our previous studies, we established a culture system (Lee et al., 2007), genetic manipulation protocol (Kim et al., 2009), and inducer supplementation protocol for mass production of phleichrome (So et al., 2015) from C. phlei. However, the gene responsible for the synthesis of phleichrome has yet to be determined.

Thus, our goal in this study was to identify the gene responsible for biosynthesis of phleichrome. We tested whether Cppks1, which encodes a polyketide synthase (PKS), is the PKS gene responsible for phleichrome biosynthesis by analyzing transcription of Cppks1 in response to a synthetic inducer. Furthermore, we expressed Cppks1 in the chestnut blight fungus Cryphonectria parasitica, which does not produce phleichrome but produces bisanthraquinones such as skyrin, oxyskyrin, and rugulosin via the polyketide pathway (Shibata, 1973).

Fungal strains, culture media, and growth conditions

Cladosporium phlei (ATCC 36193) was stored in the form of frozen agar plugs containing actively growing young hyphae in 5% DMSO solution at ?70°C and maintained on potato dextrose agar (PDA) as described previously (Lee et al., 2007). Culture conditions and methods for phleichrome induction using 150 μM cyclo-(l-Pro-l-Phe) have been described previously (So et al., 2015). Details of the chemical synthesis of the cyclo-(l-Phe-l-Pro) inducer are provided in our previous study (So et al., 2015). Mycelia on cellophane membranes overlaying agar plates were collected and lyophilized as described previously, and stored until use (So et al., 2015). All chemicals were obtained from Sigma-Aldrich Co. (USA), unless otherwise specified.

Analysis of Cppks1 transcript expression

RNA from mycelia grown on cellophane layered on top of PDA media supplemented with 150 μM cyclo-(l-Pro-l-Phe) inducer was extracted as described previously (Park et al., 2004). Northern blot analysis was used to compare the level of Cppks1 (GenBank accession no. JX129223) transcript before and after induction. Semi-quantitative analysis of the accumulation of Cppks1 transcript was performed by RT-PCR using total RNA extracted from cultures at 6 and 18 days after inoculation with and without induction, as described previously (Park et al., 2012). Cppks1 transcript levels were compared with those of another PKS gene encoding a highly-reducing (HR)-PKS, namely Cppks3 (GenBank accession no. JX129225), and normalized to levels of the gene encoding beta-tubulin (β-tub) as an internal control (Choi and Nuss, 1990). Analyses were conducted at least twice, in triplicate for each transcript, from at least two independent RNA preparations of the same sample with primers specific for β-tub, Cppks3, and Cppks1 genes. Primer pairs for each gene are described in Table 1. To determine relative phleichrome content from the corresponding mycelia, phleichrome was extracted from mycelia using ethyl acetate (EtOAc) and compared using thin-layer chromatography (TLC) on a silica gel with a resolving solution (CH₂Cl₂/MeOH = 19:1, v/v), as described previously (Yi et al., 2011).

Construction of a Cppks1 expression cassette for heterologous expression

We selected a genomic fosmid clone containing full-length Cppks1 from our previous study. The deduced Cppks1 sequence contained no introns, and the protein product of CpPKS1 consisted of 2,174 codons, with an estimated molecular mass of 235.1 kDa and a pI of 6.02. To construct an expression vector for C. parasitica, the ORF of Cppks1 was placed between the strong 188-bp promoter element of the C. parasitica cryparin gene (Crp1) (Kwon et al., 2009) and the terminator of A. nidulans trpC gene (Mullaney et al., 1985) from the fungal transforming vector pSilent-Dual1 (Nguyen et al., 2008) using an overlap extension PCR method. The resulting fusion construct was digested with NotI and ligated into NotI-digested recombinant pBluescriptII SK (+) vector carrying the 1.7-kb XhoI fragment of pSilent-Dual1 encoding the geneticin resistance cassette (Nguyen et al., 2008). The resulting expression vector p188CpPKS1 was sequenced and then used to transform C. parasitica EP155/2 strain (ATCC 38755).

Protoplast preparation and transformation were performed as described previously (Churchill et al., 1990; Kim et al., 1995). Transformants were selected from agar plates supplemented with 150 μg/ml geneticin (Invitrogen, USA), passaged three to four times on selective media, and single-spore isolated, as described previously (Baek et al., 2014). PCR and Southern blot analysis were conducted with genomic DNA from the transformants to confirm integration of the p188CpPKS1 vector into the fungal genome (Song et al., 2013).

Analysis of the transformants

Production of phleichrome from recombinant C. parasitica was analyzed using thin-layer chromatography (TLC) followed by liquid chromatography (LC/MS/MS). Briefly, EtOAc was used to extract phleichrome from mycelia on cellophane membranes overlaying agar plates. The crude extract was then resolved using TLC on silica gel with a resolving solution (CH₂Cl₂/MeOH = 19:1, v/v) and purified phleichrome as a control (Yi et al., 2011). The band with the same mobility as that of the purified phleichrome was scraped out from the silica gel and dissolved in methanol, and the presence of phleichrome was further analyzed by LC/MS/MS.

LC/MS/MS analysis

A Surveyor HPLC on line with an Agilent 6410B (Agilent Technology, Wilmington, DE, USA) LC/MS/MS mass spectrometer equipped with an ESI source was employed. Separations were performed on an analytical reverse phase column (Agilent XDB_C18, 5 μm, 2.0 mm × 15 cm). The injection volume was 5 μl. The mobile phases were as follows: A = water, 0.1% formic acid, and B = acetonitrile, 0.1% formic acid. Lipopeptides were eluted using a linear gradient of 10% B to 100% buffer for 20 min at a flow rate of 0.23 ml/min. Ions were generated in positive ionization mode using an electrospray ionization interface. Fragmenter potential was set to 130 V, the interface heater was set to 300°C, and the scan range was m/z 100 to 1000. The metabolic profile of the TLC-purified pigment from the transformed C. parasitica was compared with the profile for purified phleichrome from C. phlei obtained in previous studies (Lee et al., 2007; Yi et al., 2011).

Bioinformatic analyses of a gene involved in phleichrome biosynthesis

Structural comparison with other fungal perylenequinones revealed that phleichrome is a close analogue of cercosporin. The only gross structural differences between the two substances are the two methoxy groups in phleichrome instead of the methylenedioxy group of cercosporin (Arnone et al., 1985). Cercosporin is known to be synthesized via the polyketide metabolic pathway using acetate and malonate subunits (Choquer et al., 2005; Kurobane et al., 1981; Kusari et al., 2009; Liao and Chung, 2008; Okubo et al., 1975). Structural similarity suggests that phleichrome is also synthesized via the polyketide pathway, which in fungi is orchestrated by polyketide synthases (PKS), which are multimeric enzymes that function analogously to fatty acid synthases joining carboxylic acid units in a stepwise fashion (Crawford and Townsend, 2010). All PKSs are currently divided into three general classes according to the organization of their active sites; fungal PKSs are multidomain systems (iterative type I) (Crawford and Townsend, 2010; Hutchinson and Fujii, 1995). In our previous study, we cloned four C. phlei PKS genes, which included three representative subclasses of PKS based on the level of reductive processing during chain assembly (So et al., 2012). In general, fungal aromatic polyketides, including phleichrome, are known to be synthesized by non-reducing (NR)-PKSs (Crawford and Townsend, 2010). Phylogenetic analysis of the four cloned C. phlei PKS genes indicated that the Cppks1 protein product appeared to be a NR-PKS (So et al., 2012). It was most closely related to other fungal PKS genes involved in the synthesis of red and orange perylenequinone pigments, including close analogues of phleichrome such as cercosporin (Choquer et al., 2005) and elsinochromes (Liao and Chung, 2008). In addition, encoded proteins of Cppks1 and two other genes for cercosporin and elsinochrome biosynthesis consisted of five catalytic domains, including keto synthase (KS), acyltransferase (AT), thioesterase/claisen cyclase (TE/CYC), and two consecutive acyl carrier proteins (ACP), that were distributed at comparable distances in the same order and at a similar length (Choquer et al., 2005; Liao and Chung, 2008; So et al., 2012). Altogether, our results suggested that Cppks1 is responsible for the biosynthesis of phleichrome via the fungal polyketide pathway.

Regulation of the expression of Cppks1

In our previous studies, we found that phleichrome production was significantly increased by addition of 150 μM cyclo-(l-Pro-l-Phe) synthetic inducer into the culture media (Fig. 1). Thus, we analyzed levels of the Cppks1 transcript in response to the inducer using semi-quantitative RT-PCR. Total RNA was extracted from cultures at 6 and 18 days postinoculation. These timepoints were chosen because discernable pigmentation was observed on the PDA plates by 6 days postinoculation and the maximum amount of phleichrome was produced by 18 days postinoculation. As shown in Fig. 2, supplementation of the synthetic inducer into culture significantly increased the accumulation of Cppks1 transcript. The level of Cppks1 transcript increased along with incubation time, similar to the temporal pattern of phleichrome production under induction conditions. However, the expression of Cppks3, which is an HR-PKS, did not change in response to the synthetic inducer supplementation. The coordinated response of Cppks1 transcript to the inducer suggests that Cppks1 gene is associated with phleichrome biosynthesis.

Analysis of C. parasitica transformants

Heterologous expression is one of the most efficient methods for functional analysis of a given cloned gene. C. parasitica is a genetically tractable organism (Churchill et al., 1990) capable of producing highly complexed pigments such as skyrin, oxyskyrin, and rugulosin via the polyketide pathway (Shibata, 1973). Therefore, we expressed Cppks1 in C. parasitica to see if the resulting recombinant fungus would produce phleichrome. Thirty putative transformants were selected from the top agar plates supplemented with 150 μg/ml geneticin, passaged three times on selective media supplemented with 50 μg/ml geneticin, and single-spore isolated prior to further analyses. Mitotic stability of each transformant was confirmed by successive transfers alternating on selective and non-selective media. In order to detect integration of the transforming vector into the chromosome of C. parasitica, PCR analysis using Cppks1-specific primers was conducted. All 30 transformants were positive for the 646-bp PCR amplicon, indicating stable integrative transformants.

Because phleichrome is responsible for the characteristic deep red pigmentation in the mycelia and culture medium, we looked for changes in the colour of colonies in single-spored transformants. As shown in Fig. 3, two colonies became pinkish over time, in contrast to the original orange colour, as culture aged. However, no discernible changes in any of the other characteristics, including growth rate and sporulation, were observed in these transformants. Therefore, we assessed the presence of phleichrome in these selected transformants. TLC analysis using an EtOAc extract of the mycelia revealed the presence of yellowish pigment at an Rf value of 0.24, the same Rf as for purified phleichrome (Fig. 4A). However, these pigments were also present in wild-type C. parasitica. These results indicated that TLC was not sensitive enough to differentiate phleichrome from other residual pigments, such as skyrin and oxyskyrin. The amount of heterologous phleichrome produced may also have been too low for detection by TLC.

Thus, pigment spots migrating at the same Rf value as purified phleichrome on TLC plates were scraped off and extracted with methanol, and the presence of phleichrome was determined by LC/MS/MS analysis with purified phleichrome from C. phlei as a control. As shown in Fig. 4B, methanol extract of TLC spots from recombinant C. parasitica had a new peak with the same retention time as the control phleichrome, and the molecular weight of the corresponding peak was the same as that of phleichrome, confirming the presence of phleichrome in extracts. These results clearly indicated that phleichrome is synthesized via the polyketide pathway and that Cppks1 encodes a protein responsible for the biosynthesis of phleichrome. It would be of great interest to carry out functional analysis of the Cppks1 gene product using gene replacement or silencing. In addition to the PKS gene, tailoring proteins that catalyze subsequent reactions including oxidation, hydration, methylation, and hydroxylation are also required for successful cercosporin biosynthesis (Chen and Nuss, 2007). C. parasitica does not produce phleichrome but does produce secondary metabolites such as skyrin, oxyskyrin, and rugulosin via the polyketide pathway (Shibata, 1973).

Genome survey of C. parasitica suggested the presence of at least 31 PKS genes and/or clusters (http://genome.jgipsf.org/Crypa1/Crypa1.home.html). Thus, genes involved in the processing of the core component of C. parasitica may be involved in further processing of phleichrome. In addition, heterologous production of a fungal polyketide in the yeast Saccharomyces cerevisiae requires co-expression of the corresponding PKS gene and a heterologous 4′-phosphopantetheinyl transferase (PPTase) (Kealey et al., 1998; Wattanachaisaereekul et al., 2007). PPTases catalyze the post-translational modification of proteins by the covalent attachment of a 4′-phosphopantetheine moiety of coenzyme A to a conserved serine residue of an inactive form of ACP, resulting in conversion to the active form during the biosynthesis of metabolites (Walsh et al., 1997). Interestingly, no PPTase domain or genes were found in the gene cluster for cercosporin biosynthesis. Therefore, the successful heterologous production of phleichrome suggests that there was an appropriate PPTase that acted in trans in C. parasitica, that converted heterologously expressed inactive apo-CpPKS1 to active holo-CpPKS1 by modifying the ACP domain of CpPKS1 with a 4′-phosphopantetheine moiety. Almost all organisms that utilize more than one 4′-phosphopantetheine-dependent pathway also have more than one PPTase (Gehring et al., 1998; Quadri et al., 1998). A genome survey of C. parasitica suggested the presence of three PPTase genes (http://genome.jgi-psf.org/Crypa1/Crypa1.home.html). However, filamentous fungi such as Neurospora crassa, A. fumigatus, A. nidulans, and Fusarium fujikuroi harbor only a single PPTase (Kim et al., 2015).

Genes involved in the secondary metabolite pathways in filamentous fungi are often organized in clusters (Keller et al., 2005). Therefore, we are currently characterizing the loci surrounding Cppks1 to determine if genes involved in the biosynthesis of phleichrome are clustered. Additionally, Cppks1 may not be the only gene for phleichrome synthesis. The survey of genomic library suggested that more NR-PKS genes remain to be identified in addition to the four PKS genes cloned in a previous study (So et al., 2012). Further studies are required to characterize all cloned PKS genes. Functional analysis of Cppks1 gene will allow determination of the redundancy of the phleichrome gene.