Molecular analysis of the cercosporin biosynthetic gene cluster in Cercospora nicotianae

Authors

  • Huiqin Chen,

    1. Citrus Research and Education Center, Institute of Food and Agricultural Sciences (IFAS), University of Florida, 700 Experiment Station Road, Lake Alfred, FL 33850, USA.
    Search for more papers by this author
  • Miin-Huey Lee,

    1. Department of Plant Pathology, National Chung-Hsing University, Taichung 402, Taiwan.
    Search for more papers by this author
  • Margret E. Daub,

    1. Department of Plant Biology, North Carolina State University, Raleigh, NC 27695, USA.
    Search for more papers by this author
  • Kuang-Ren Chung

    Corresponding author
    1. Citrus Research and Education Center, Institute of Food and Agricultural Sciences (IFAS), University of Florida, 700 Experiment Station Road, Lake Alfred, FL 33850, USA.
    2. Department of Plant Pathology, IFAS, University of Florida, Gainesville, FL 32611, USA.
    Search for more papers by this author

*E-mail krchung@ufl.edu; Tel. (+1) 863 956 1151; Fax (+1) 863 956 4631.

Summary

We describe a core gene cluster, comprised of eight genes (designated CTB1–8), and associated with cercosporin toxin production in Cercospora nicotianae. Sequence analysis identified 10 putative open reading frames (ORFs) flanking the previously characterized CTB1 and CTB3 genes that encode, respectively, the polyketide synthase and a dual methyltransferase/monooxygenase required for cercosporin production. Expression of eight of the genes was co-ordinately induced under cercosporin-producing conditions and was regulated by the Zn(II)Cys6 transcriptional activator, CTB8. Expression of the genes, affected by nitrogen and carbon sources and pH, was also controlled by another transcription activator, CRG1, previously shown to regulate cercosporin production and resistance. Disruption of the CTB2 gene encoding a methyltransferase or the CTB8 gene yielded mutants that were completely defective in cercosporin production and inhibitory expression of the other CTB cluster genes. Similar ‘feedback’ transcriptional inhibition was observed when the CTB1, or CTB3 but not CTB4 gene was inactivated. Expression of four ORFs located on the two distal ends of the cluster did not correlate with cercosporin biosynthesis and did not show regulation by CTB8, suggesting that the biosynthetic cluster was limited to CTB1–8. A biosynthetic pathway and a regulatory network leading to cercosporin formation are proposed.

Introduction

Many phytopathogenic Cercospora species produce a host non-selective, photoactivated phytotoxin, called cercosporin that is required for high levels of virulence by these fungi on plants (Daub, 1982a; Daub et al., 2005). Cercosporin is a perylenequinone that absorbs light and reacts with oxygen molecules to generate reactive oxygen species, including singlet oxygen (1O2) and superoxide radicals (O2) (Daub and Hangarter, 1983). The reactive oxygen species cause peroxidation of cell membranes and electrolyte leakage in host plants (Daub, 1982b; Daub and Briggs, 1983; Daub and Ehrenshaft, 2000). In addition, cercosporin is able to damage major cellular components including nucleic acids, proteins and lipids, and has also been shown to exert broad toxicity to bacteria, many fungi and mice (Yamazaki et al., 1975; Ito, 1981; Daub, 1982a). Extensive investigations have focused on understanding the chemistry and mode of action of cercosporin as well as on the mechanisms involved in fungal self-defence against singlet oxygen and related photosensitizing compounds (Daub et al., 1992; Sollod et al., 1992; 2000; Jenns and Daub, 1995; Ehrenshaft et al., 1998; 1999; Chung et al., 1999; 2003a). Cercosporin has recently been demonstrated to play a crucial role in pathogenicity of C. nicotianae and lesion formation on tobacco (Choquer et al., 2005; 2007; Dekkers et al., 2007).

In contrast to the large body of information on cercosporin toxicity and resistance, considerably less is known about the cercosporin biosynthetic pathway and its regulation. Accumulation of cercosporin in culture is affected by environmental and developmental factors (Jenns et al., 1989; Daub and Ehrenshaft, 2000). Light has been shown to be the most critical factor, not only for toxicity, but also for cercosporin production. Cercosporin is only produced in the light, and its production is completely suppressed under complete darkness (Ehrenshaft and Upchurch, 1991); brief exposure to light immediately triggers its biosynthesis. A Zn(II)Cys6 transcription factor, CRG1, has been identified with a role in regulating both cercosporin resistance and production (Chung et al., 1999; 2003a). The mechanism of CRG1 regulation and its relation to signalling pathways and environmental signals is not yet understood.

A substrate-feeding study conducted by Okubo et al. (1975) suggested that cercosporin is synthesized via the polyketide pathway using acetate and malonate subunits. The first gene putatively identified in the pathway was CFP (cercosporin facilitator protein) encoding a major facilitator superfamily (MFS) transporter required for exportation of cercosporin out of the mycelium (Callahan et al., 1999). In order to identify genes involved in cercosporin biosynthesis, we identified several classes of cercosporin deficiency mutants in C. nicotianae through a restriction enzyme-mediated integration (REMI) mutagenesis approach (Chung et al., 2003b). A fungal polyketide synthase encoding gene, named CTB1 (Cercosporin Toxin Biosynthesis), was subsequently recovered from one of the REMI mutants and was demonstrated to be required for cercosporin production through genetic and molecular analysis (Choquer et al., 2005). Further study revealed that another gene (CTB3), immediately adjacent to the CTB1, was also required for cercosporin biosynthesis (Dekkers et al., 2007), suggesting that the biosynthetic pathway genes are clustered.

A number of fungal secondary metabolites, including aflatoxins, compactin, ergot alkaloids, fumonisins, gibberellins, gliotoxin, HC toxin, indole-diterpenes, loline alkaloids, penicillin, sterigmatocystin, sirodesmin and trichothecenes, are synthesized by genes found in clusters in the genome (Hohn et al., 1993; Brown et al., 1996; 2004; Brakhaage, 1998; Young et al., 2001; 2006; Abe et al., 2002; Ahn et al., 2002; Proctor et al., 2003; Gardiner et al., 2004; 2005; Yu et al., 2004; Haarmann et al., 2005; Spiering et al., 2005; Tudzynski, 2005). We thus sequenced the regions flanking CTB1 and CTB3 as a means to identify the putative biosynthetic genes. Here we report on the identity of a core cluster of six genes in addition to CTB1 and CTB3 that were co-ordinately regulated and highly induced under cercosporin-producing conditions. In addition, we confirm the function of two of these genes, CTB2 and CTB8, in cercosporin biosynthesis using targeted gene disruption and genetic complementation. This work provides a genetic framework for better understanding the metabolic synthesis and regulation of cercosporin.

Results

Chromosomal walking, sequence analysis and identification of putative open reading frames

We previously characterized the CTB1 gene encoding a fungal polyketide synthase (Choquer et al., 2005) and the CTB3 gene encoding a dual O-methyltransferase and FAD-dependent monooxygenase (Dekkers et al., 2007) from C. nicotianae to be responsible for cercosporin biosynthesis. Sequences flanking the CTB1 and CTB3 genes were obtained by polymerase chain reaction (PCR) using a chromosomal walking strategy with CTB1 and CTB3-specific primers. New primers were designed based on the end sequences of compiled sequences and used for further rounds of PCR. Overlapping sequences were carefully assembled. The CTB1, CTB2, CTB3, CTB4 and CTB8 open reading frames (ORFs) were determined by sequencing cDNA clones, whereas other ORFs were predicted using the computer software.

In addition to CTB1 and CTB3, 10 putative ORFs were identified in a span of approximately 36 kb analysed (Table 1; Fig. 1A). blastx searches in the NCBI database were used to identify function for the predicted polypeptide. The conceptually translated CTB2 protein is similar to many methyltransferases (see below for details). The CTB4 protein is highly similar to many hypothetical proteins and MFS transporters of fungi and bacteria, with greatest similarity to an Aspergillus fumigatus membrane transporter. The CTB5 protein has homology to numerous hypothetical proteins and to oxygen, FAD/FMN-dependent oxidoreductases of fungi and bacteria including A. fumigatus, Neurospora crassa and Thermobifida fusca. CTB6 encodes a putative protein with similarity to a wide range of NADPH-dependent dehydrogenases or oxidoreductases such as d-lactaldehyde dehydrogenase of Cryptococcus neoformans and ketoreductase of A. fumigatus. The CTB7 protein is similar to many FAD/FMN-dependent oxidoreductases of bacteria, such as Xanthomonas axonopodis pv. citri, Streptomyces coelicolor and Pseudomonas syringae. The CTB8 protein has similarity to many fungal transcription factors containing zinc finger DNA-binding domains (see below for details).

Table 1.  The cercosporin biosynthetic gene cluster in Cercospora nicotianae.
Genea (accession number)Length (bp)Intron numberPredicted functionAmino acidsClosest blast match (accession number)Identity (%)E-value
  • a. 

    Genes with putative roles in the cercosporin pathway are designated CTB (Cercosporin Toxin Biosynthesis), whereas other flanking genes with no predicted function in cercosporin production are tentatively designated ORFs 9, 10, 11 and 12.

CTB1 (AY649543)70368Polyketide synthase2196Polyketide synthase PksP (XP_756095)370
CTB2 (DQ991505)14391O-methyltransferase462O-methyltransferase (XP_748078)
(AAS90087) (AAS90065)
(AAS66017) (BAE71329)
263e-31
2e-19
1e-19
CTB3 (DQ355149)27312O-methyltransferase/
FAD-dependent monooxygenase
871O-methyltransferase (XP_748078);
FAD-dependent monooxygenase (NP_901496)
37
30
1e-26
4e-32
CTB4 (DQ991506)18163Major Facilitator Superfamily
(MFS) Transporter
512Membrane transporter(XP_756023)483e-112
CTB5 (DQ991507)13800Oxygen, FAD/FMN-dependent
oxidoreductase
459Oxidoreductase, FAD-binding
(XP_755780) (XP_961030)
(YP_288121)
372e-69
CTB6 (DQ991508)10740NADPH-dependent oxidoreductase357d-lactaldehyde dehydrogenase
(XP_570921) (XP_755361)
291e-26
CTB7 (DQ991509)14011FAD/FMN-dependent oxidoreductase450Oxidoreductase (AAM36534)
(NP_624584) (NP_790149)
331e-45
CTB8 (DQ991510)12451Zinc fingeritranscription factor397AFLR_EMENI Sterigmatocystin biosynthesis
regulatory protein
(XP_681089) (XP_411957)
(AAM02988) (AAM46650)
(AAM02989) (AAW32181)
254e-11
ORF9 (DQ993249)9763Conserved eukaryotic protein of
unknown function
271Hypothetical protein with DUF850 domain
(XP_747337)
609e-63
ORF10 (DQ993250)13853Hypothetical protein395Unnamed protein product (BAE64725) (Q00808)402e-50
ORF11 (DQ993251)9990Truncated transcription factor332AFLR (AAM02976)360.001
ORF12 (DQ993252)9600Truncated transcription factor319Hypothetical protein (EAT83777)331e-09
Figure 1.

The cercosporin toxin biosynthesis (CTB) gene cluster in Cercospora nicotianae.
A. Transcription map of the CTB cluster genes and the adjacent open reading frames (ORFs). Arrows indicate the orientation of transcription based on sequence analysis and/or deduced from comparisons of the related genes in the databases.
B. Expression of the CTB cluster genes assessed by Northern-blot analysis and production of cercosporin by the wild-type C. nicotianae. Total RNA was purified from fungal mycelia grown on potato dextrose agar (PDA) or complete medium (CM) under continuous light (LT) or darkness (DK), electrophoresed in formaldehyde-containing gels, blotted to nylon membranes and hybridized to a specific probe as indicated. Sizes of hybridizing bands are indicated in kilobase pairs (kb). Cercosporin was extracted with ethyl-acetate and analysed by thin-layer chromatography (TLC). Cercosporin (Rf 0.58) was indicated by an arrow.

Two ORFs (9 and 10) downstream of CTB4 did not appear to encode proteins involved in metabolic functions. The predicted ORF9 protein contains a DUF850 domain and has amino acid similarity to many conserved eukaryotic proteins of unknown function. The ORF10 protein has similarity to many hypothetical proteins of fungi and proteins containing a GTP-binding motif that are presumably involved in vegetative compatibility in Podospora anserine. Two small ORFs (11 and 12) showing similarity to GAL4-like Zn(II)Cys6 transcription regulators were found upstream of CTB8. ORF11 appears to encode a truncated zinc binding protein due to missing two cysteine residues in the first zinc cluster (xxxxxxxxxC-----CxxCxxxxxxC). ORF12 displays amino acid similarity to CTB8 and ORF11, but completely lacks the zinc binding domain.

In the present study CTB2 and CTB8 were characterized in detail. Comparisons of the CTB2 genomic and cDNA sequences revealed that the CTB2 gene coding region contains 1389 nucleotides with a 53 bp intron. The predicted translation product (462 amino acids) of CTB2 displayed a strong similarity to numerous fungal hypothetical proteins derived from genome sequencing projects and to O-methyltransferases of Aspergillus  flavus, Aspergillus nomius, Aspergillus parasiticus and Aspergillus oryzae. CTB2 contains a conserved S-adenosyl methionine (SAM)-binding domain similar to a wide range of O-methyltransferases. CTB2 amino acids also display similarity to the N-terminal methyltransferase domain of CTB3.

Comparisons of the CTB8 genomic and cDNA sequences identified a 1185 bp ORF and the presence of a 51 bp intron. The translated 397 amino acids of CTB8 contained a GAL4-like Zn(II)Cys6 binuclear cluster DNA-binding domain (data not shown), and exhibited similarity to the sterigmatocystin biosynthesis regulatory proteins in Emericella nidulans, and the aflatoxin biosynthesis AFLR transcriptional activators in Aspergillus species. These results suggested that CTB8 may function as a transcriptional regulator involved in cercosporin biosynthesis.

Coexpression of the CTB1–8 genes and cercosporin accumulation

Expression of the CTB1-CTB8 genes and four putative ORFs was performed by Northern-blot hybridization of total RNA of the wild-type C. nicotianae grown on PDA or CM in the light or under darkness to determine if they were co-ordinately regulated under conditions highly favourable (PDA and light), moderately favourable (CM and light) or unfavourable (dark) for cercosporin accumulation (Fig. 1B). Accumulation of transcripts for CTB1, CTB2, CTB3, CTB4, CTB5, CTB6, CTB7 and CTB8 was observed when the fungus was grown on PDA in the light, whereas their expression was completely inhibited when grown under complete darkness. Accumulation of the eight CTB1-CTB8 transcripts was significantly downregulated when grown on CM, compared with PDA, in the light and was almost not detected under darkness. Production of cercosporin was well co-ordinated with the patterns of gene expression of the eight CTB1-CTB8 genes (Fig. 1B). Large amounts of cercosporin were detected when the fungus was grown on PDA, with markedly reduced levels in cultures grown on CM in the light. Cercosporin was nearly undetectable when grown under complete darkness in either medium. In contrast, the ORF9, ORF10 and ORF12 transcripts were constitutively expressed and did not appear to respond to light or medium. Expression of ORF11 was slightly regulated by light, but did not show the pronounced medium-regulation of the CTB1-CTB8 genes.

Promoter analysis of the CTB1–8 genes

Because the CTB clustered genes were co-ordinately regulated, the promoters of the CTB1-CTB8 genes and their adjacent ORFs were analysed for potential common regulatory elements. Analysis of the 650 bp sequences upstream of the putative ATG translational start codon revealed that the consensus sequence CAAT was found in all the promoter regions except for ORF10 (Table 2). The consensus TATA sequence was found in the promoters of CTB1, CTB2, CTB4, CTB5, CTB7 and ORF11 but not in CTB3, CTB6, ORF9 and ORF12. All the promoter regions have one or multiple putative GATA motifs found in the AreA (nitrogen regulatory protein) gene of E. nidulans (Wilson and Arst, 1998), whereas none of the CTB clustered gene promoters has the CreA (carbon regulatory protein) binding site (Panozzo et al., 1998). All but CTB1, CTB2, CTB5 and ORF10 have a putative pH regulatory protein PacC binding motif (Espeso and Arst, 2000). A palindrome sequence 5′-TCG(N3−6)CGA-3′ was identified in all the CTB1–8 and ORF10 promoters, but was absent in the promoters of ORF9, ORF11 and ORF12 (Table 2).

Table 2.  Promoter analysis of genes in the cercosporin biosynthetic gene cluster of Cercospora nicotianae.
GeneTranslation initiation site
CA(C/A)(A/C)ATGGCa
TATAbCAATbpH regulatory
GCCA(A/G)G
AreA or
WC1/WC2:GATAb
Palindrome TCG(N3−6)CGA
  • a. 

    The translation initiation site is underlined.

  • b. 

    Numbers of the consensus sequences with 100% matches as indicated.

CTB1CACCATGGA1302TCG (N3)CGA
CTB2AAAAATGGT1501TCG (N6)CGA
CTB3CACACATGAT0311TCG (N6)CGA
CTB4CCACAATGGC1512TCG (N3)CGA
CTB5AATCCATGGG1703TCG (N3)CGA
CTB6AGTTATGGC0312TCG (N5)CGA
CTB7AGACATGGC1111TCG (N6)CGA
CTB8AACCATGGC1213TCG (N6)CGA
ORF9AACCATGGC0211
ORF10AGTGATGCG1001TCG (N5)CGA
ORF11GCTGATGAG1212
ORF12GTCCATGCG0411

Functional characterization of CTB2 and CTB8 involved in cercosporin biosynthesis

To determine the involvement of CTB2 and CTB8 in cercosporin formation, disruption plasmids, pΔctb2 and pΔctb8, were constructed to replace the entire coding regions of CTB2 and CTB8, respectively, via homologous integration. Because cercosporin is a visible red pigment, all transformants recovered were visually screened for cercosporin production on solid potato dextrose agar medium. After transforming with split marker fragments obtained from pΔctb2 (Fig. 2A), four of 52 transformants (7.7%) failed to produce visible cercosporin. Southern-blot analysis of the EcoRI/NcoI-digested genomic DNA from wild type and two putative ctb2 knockouts (Δctb2-D5 and D18) confirmed that gene disruption specifically occurred at the CTB2 locus (Fig. 2B). The two ctb2 mutants did not produce any visible cercosporin, but accumulated a yellow pigment (Fig. 2C). Transformation of a full-length CTB2 gene cassette into a Δctb2 null mutant restored cercosporin production.

Figure 2.

Targeted disruption of the CTB2 gene encoding a putative O-methyltransferase in Cercospora nicotianae.
A. Physical map of CTB2 in wild type (WT) and in a Δctb2 disruptant carrying a hygromycin phosphotransferase B gene (HYG) marker integrated within CTB2 via homologous recombination. Sizes of hybridizing fragments from wild type and the Δctb2 disruptant are also indicated.
B. Southern-blot analysis of genomic DNA from wild type and two Δctb2 disruptants (D5 and D18). Fungal DNA was digested with NcoI and EcoRI, electrophoresed, blotted and hybridized to a CTB2-specific probe. Band patterns show a successful disruption of the CTB2 gene.
C. TLC analysis of cercosporin produced by wild type, two Δctb2 disruptants (D5 and D18), and two complemented strains. Cercosporin and putative intermediates were extracted with ethyl acetate and analysed by TLC.

To disrupt CTB8, split marker fragments amplified from the disruption construct, pΔctb8 (Fig. 3A), were transformed into the wild-type C. nicotianae strain, resulting in six cercosporin-defective mutants out of 55 transformants screened (11%). Targeted gene disruption of the CTB8 gene in C. nicotianae was verified by Southern-blot analysis of the XhoI/PuvII-digested genomic DNA from wild type and five mutants using a CTB8-specific probe (Fig. 3B). TLC analysis of culture extracts indicated that production of cercosporin was completely abolished in the Δctb8 knockouts (Fig. 3C). Transformation of a full-length CTB8 gene and its promoter region into a Δctb8 mutant restored cercosporin production (Fig. 3C).

Figure 3.

Targeted gene disruption of CTB8 encoding a putative Zn(II)Cys6 transcription regulator using a split-marker strategy in Cercospora nicotianae.
A. Restriction maps of the CTB8 gene in the genome of wild type (WT) and in the Δctb8-disrupted mutant. Two truncated CTB8 fragments fused with an overlapped HY/YG (hygromycin resistance gene cassette) were amplified by PCR with oligonucleotide primers (TF1 paired with hyg4; TF2 paired with hyg3) as indicated. Note: drawing is not to scale.
B. Southern-blot analysis of genomic DNA from wild type (WT) and five Δctb8 disrupted mutants (Δctb8-D1, 3, 4, 10 and 47). Fungal DNA was digested with PuvII and XhoI, electrophoresed, blotted onto a nylon membrane and hybridized with a CTB8-specific probe. Hybridizing patterns indicate disruption of the CTB8 gene.
C. TLC analysis of cercosporin produced by wild type, six Δctb8 disruptants (D1, D3, D4, D10, D41 and D47), and five complemented strains (C5, C6, C7, C14 and C32).

The Δctb2 or Δctb8 disruptants grew slightly faster than wild type, but retained the wild-type level of resistance to exogenous cercosporin or other photosensitizers such as eosin Y, haematoporphyrin, methylene blue and toluidine blue (data not shown). Pathogenicity assays using detached tobacco leaves also revealed that the Δctb2 and Δctb8 knockouts incited fewer lesions compared with wild type (data not shown), consistent with the crucial role of cercosporin in fungal virulence and lesion formation.

Gene regulation in cercosporin biosynthesis

As stated above, the CTB1-CTB8 genes were regulated by light and medium corresponding to the conditions conducive to cercosporin accumulation, indicating that CTB1-CTB8 may represent the core cluster for cercosporin biosynthesis. Northern-blot analysis of total RNA from wild type and two Δctb8 null mutants grown on PDA in the light showed that expression of the CTB1-CTB7 genes was nearly abolished in two Δctb8 null mutants (Fig. 4), indicating that the CTB8 transcriptional activator controls cercosporin production by controlling gene transcript levels. In contrast, levels of the ORF9-ORF12 transcripts were not reduced in the Δctb8 mutants as compared with wild type. Regulation of the CTB1, CTB3, CTB7 and CTB8 genes was also examined in a mutant (205C3) disrupted for the CRG1 transcription regulator; this mutant produces approximately 50% of the cercosporin produced by wild type and exhibits partial sensitivity to exogenous cercosporin (Chung et al., 2003a). Northern-blot hybridization showed a drastic reduction of transcripts of the CTB1, CTB3, CTB7 and CTB8 genes in the 205C3 mutant (Fig. 5A). Accumulation of the ORF10 transcript was slightly reduced in the 205C3 mutant compared with wild type. However, there was no significant difference in expression of the CRG1 gene between the Δctb8 null mutants and wild type (Fig. 5B).

Figure 4.

Northern-blot hybridization indicates cosuppression of transcripts of the CTB cluster genes but not the adjacent ORFs in the ctb8 disruptants of Cercospora nicotianae. Total RNA purified from wild type (WT) and two Δctb8 disruptants (D1 and D4) was electrophoresed in formaldehyde-containing gels, blotted onto nylon membranes and hybridized to the probes at 65°C as indicated. The membranes were washed at high stringency, and sizes of hybridizing bands are indicated in kilobase pairs (kb).

Figure 5.

Epistatic analysis between the CTB8 and CRG1 transcriptional regulators in Cercospora nicotianae.
A. Northern-blot analysis of expression of the CTB genes in wild type (WT) and a previously generated 205C3 strain that is defective in a putative Zn(II)Cys6 transcriptional regulator, CRG1, and is required for normal cercosporin production and resistance (Chung et al., 1999; 2003a).
B. Gene expression of the CRG1 gene in wild type (WT) and two Δctb8 disruptants of C. nicotianae. In addition to the CRG1 transcript (2.6 kb), a 4.5 kb transcript is produced due to production of a dicistronic mRNA of CRG1 and an adjacent putative uracil transcript gene (PUT1) (Chung et al., 2003c). Ethidium bromide-stained rRNA indicating the relative loading of the samples is also shown. Total RNA was isolated from fungal strains grown on potato dextrose agar under continuous light for 5 days, electrophoresed in formaldehyde-containing gels, blotted onto nylon membranes and hybridized to the probes at 65°C as indicated.

To further determine coregulation of the CTB genes in the cluster, accumulation of the CTB gene transcripts also was examined in the Δctb2 null mutants as well as in the Δctb1 and Δctb3 null mutants identified in the previous studies (Fig. 6). The results showed that accumulation of the CTB3 and CTB8 gene transcripts in the Δctb1 knockouts was drastically reduced in four knockouts tested (Fig. 6A). In contrast, regulation of expression of CTB2 and CTB4 in the Δctb1 knockouts was apparently variable because two Δctb1-D1 and D11 knockouts had the wild-type levels of expression of CTB2 and CTB4, whereas expression of the two genes was nearly undetectable in the Δctb1-D6 and D7 knockouts. Analysis of two Δctb2 null mutants also revealed that transcripts of the CTB1, CTB2, CTB3, CTB4 and CTB8 genes were completely undetectable when CTB2 was disrupted (Fig. 6B). A small, but truncated transcript (less than 1.2 kb) was detected by the CTB2 probe (Fig. 6B) in the ctb2 knockout (D5), likely resulting from expression driven by the trpC promoter in the HYG cassette. When the CTB3 coding region was disrupted by a hygromycin gene cassette, expression of the CTB1 gene was markedly reduced and expression of the CTB2, CTB4 and CTB8 genes was nearly abolished (Fig. 6C).

Figure 6.

Northern-blot analysis indicates a feedback inhibition of the CTB clustering genes in Cercospora nicotianae.
A. Accumulation of transcripts of the CTB genes in wild type (WT) and four Δctb1 disruptants (D1, D6, D7 and D11).
B. Expression of the CTB genes in wild type and two Δctb2 disruptants (D5 and D18). A truncated transcript (small than 1.2 kb) was detected by the CTB2 probe in the Δctb2-D5 mutant, probably resulting from expression driven by the trpC promoter in HYG.
C. Expression of the CTB genes in wild type and three Δctb3 disruptants (D4, D5 and D16). Total RNA was isolated from fungal strains grown on potato dextrose agar under continuous light for 5 days, electrophoresed in formaldehyde-containing gels, blotted onto nylon membranes and hybridized to the probes as indicated. Ethidium bromide-stained rRNA is shown to indicate the relative loading of the samples.

Differential expression of the CTB1–8 genes

A prior study indicated that expression of the CTB3 gene was affected by changing the carbon and/or nitrogen sources or pH in the medium (Dekkers et al., 2007). Northern-blot analyses were performed to further determine if expression of the CTB1, 2, 4, 5, 6, 7 and 8 genes also was influenced by the carbon and nitrogen sources and pH in culture (Fig. 7A). Except for CTB4, accumulation of the CTB1, 2, 5, 6, 7 and 8 gene transcripts was markedly elevated when the fungus was cultured on a synthetic medium containing mannitol, instead of glucose, as the sole carbon source (lanes 1 and 2). Substitution of nitrate with ammonium as the sole nitrogen source or depletion of nitrogen in the medium caused a drastic reduction of the CTB1, 2, 4, 5, 6, 7 and 8 gene transcripts (lanes 3 and 4). All the CTB genes tested, except for CTB4, were expressed abundantly at pH 4, pH 7, and non-buffered PDA (pH 5.6). The CTB4 gene transcript was not detected when the fungus was grown on CM under light or in acidic conditions. The amounts of cercosporin accumulated in the medium with different carbon and nitrogen sources and pH did not correlate with the levels of the CTB gene transcripts (Fig. 7B). Noticeably, all the CTB genes were highly expressed as the fungus was grown on PDA under light, whereas accumulation of cercosporin was drastically reduced on PDA buffered with a citrate-phosphate solution. Compared with the amounts of cercosporin produced on non-buffered PDA (pH 5.6; > 120 nmole cercosporin per agar plug), accumulation of cercosporin on CM (pH 5.3; < 10 nmole per agar plug) was much lower, and did not show significant differences between treatments. Depletion of the nitrogen source from CM slightly enhanced the production of cercosporin, but failed to support high levels of expression of the CTB genes.

Figure 7.

Differential expression of the CTB genes in Cercospora nicotianae.
A. Northern-blot analysis of expression of the CTB genes in the C. nicotianae wild type grown on complete medium (CM with glucose and calcium nitrate as the carbon and nitrogen sources respectively; pH 5.3) (lane 1), mannitol as the sole carbon source (lane 2), ammonium chloride as the sole nitrogen source (lane 3), no nitrogen source (lane 4), and on PDA buffered with citrate-phosphate solution to pH 4 (lane 5), or pH 7 (lane 6), and non-buffered PDA (pH 5.6) (lane 7) under continuous light. Total RNA was isolated, electrophoresed in formaldehyde-containing gels, blotted onto nylon membranes and hybridized to the probes as indicated. Sizes of hybridizing bands are indicated in kb. Ethidium bromide-stained rRNA is shown to indicate the relative loading of the samples.
B. Accumulation of cercosporin by the wild-type C. nicotianae grown in the conditions as described above. Cercosporin was purified with 5 N KOH and quantified by a spectrophotometer at 480 nm.

Discussion

Previous characterization of the CTB1 gene encoding a polyketide synthase and the CTB3 gene encoding dual O-methyltransferase and FAD-dependent monooxygenase domains revealed the requirement of both genes in cercosporin biosynthesis and fungal pathogenesis (Choquer et al., 2005; Dekkers et al., 2007). The genes involved in secondary metabolite pathways in filamentous fungi are often organized in a cluster (Keller and Hohn, 1997; Keller et al., 2005). Thus, we hypothesized that similar clustering organization of the genes for cercosporin biosynthesis may also occur in Cercospora spp. Clustering of the biosynthetic genes may provide a great advantage in coordinated gene regulation at transcriptional levels during cercosporin biosynthesis. Analysis of genomic sequence flanking the CTB1 and CTB3 genes identified 10 additional ORFs. All the ORFs were expressed as examined by Northern-blot analysis. However, only accumulation of transcripts of the CTB1-CTB8 genes correlated with conditions conducive for cercosporin production. We propose that these genes represent the core cercosporin biosynthesis gene cluster in C. nicotianae. The core gene cluster for cercosporin biosynthesis includes genes encoding a polyketide synthase (CTB1), two O-methyltransferases (CTB2 and CTB3 N terminus), a monooxygenase (CTB3 C terminus), an MFS transporter (CTB4), three oxidoreductases (CTB5, 6 and 7) and a Zn(II)Cys6 transcription regulator (CTB8). When the CTB8 gene was disrupted, expression of all of the core cercosporin cluster genes was eliminated, or significantly reduced, suggesting that CTB8 did play a major role in regulation of the pathway.

Cercosporin is a polyketide compound. Although the biochemical functions of the CTB proteins remain to be determined, we propose a pathway for cercosporin biosynthesis in a route that resembles fatty acid synthesis based on putative functions deduced from sequence similarities and studies of the biosynthesis of fungal polyketides (Rawlings et al., 1989; Watanabe and Ebizuka, 2004). The proposed pathway shall provide a framework for further investigation of cercosporin biosynthesis. As depicted in Fig. 8, the early steps for biosynthesis of cercosporin have been predicted to begin with condensation of acetyl-CoA (starter) and malonyl-CoA (extender) subunits (Okubo et al., 1975) by the function of the polyketide synthase encoded by CTB1. CTB1 contains a set of active site domains including a keto synthase (KS), an acyltransferase (AT), a thioesterase (TE) and two acyl carrier protein (ACP) domains (Choquer et al., 2005) that act cooperatively to form the polyketomethylene backbone of cercosporin. The malonyl-CoA subunit is assumed to attach to the ACP domains of CTB1 by formation of phosphopantotheine (PPT) (Rawlings et al., 1989; Watanabe and Ebizuka, 2004). The AT domain is in turn responsible for transferring the acetate unit from acetyl-CoA to the PPT of the ACP domain. The KS domain of CTB1 functions to condense the malonyl- and acetyl-CoAs by decarboxylation. After each cycle of condensation, the malonyl keto group is reduced. The putative polyketide synthase encoded by CTB1 iteratively catalyses the synthesis by incorporating two carbons in each cycle to form a linear polyketide that is released by the function of TE. Unlike fatty acid synthesis, biosynthesis of cercosporin may not involve β-ketoreduction, dehydration and enoyl reduction because CTB1 lacks β-ketoreductase (KR), dehydratase (DH) and enolyreductase (ER) domains (Choquer et al., 2005).

Figure 8.

Summary of a probable biosynthetic pathway led to the formation of cercosporin, showing hypothesized functions of the CTB gene products in Cercospora nicotianae. The keto synthase (KS) domain and acyl carrier protein (ACP) of the CTB1 likely involved in chain elongation are also indicated. More details are discussed in the text.

The polyketide generated by the function of CTB1 must undergo successive ring closure, oxidation, hydration and methylation reactions to form the polyketomethylene backbone of cercosporin. Cyclization of the aromatic ring of cercosporin is likely accomplished via the Claisen condensation as proposed for other polyketide compounds (Birch, 1967), and would be mainly catalysed by the function of TE domain in the CTB1. The N-terminal conserved sequence (LFGDQ) and a conserved histidine residue in the C terminus of CTB1 may also involve cyclization as shown for the WA protein involved in naphthopyrone production (Fujii et al., 2001). The polypeptides encoded by CTB3 and CTB5 show similarities to FAD/FMN-dependent monooxygenases/oxidoreductases that are likely involved in the polyketide oxidations. The CTB6 and CTB7 products are similar to oxidoreductases/hydrogenases and are likely responsible for hydration. Methylation at C2 and C11 of cercosporin is presumably completed by the translational products of CTB2 and/or CTB3. Cercosporin contains two methyl groups at positions C2 and C11, and methylation has also been proposed to be involved in its biosynthesis (Okubo et al., 1975). Thus, the translated products of the CTB2 and CTB3 genes may likely catalyse the addition of one or two methyl groups into the cercosporin backbone. Substrate feeding analysis by coculturing the ctb2 and ctb3 disruptants failed to form any cercosporin (data not shown). However, cercosporin was observed when a ctb1 disruptant was paired with a ctb2 or ctb3 disruptant by forming a distinct red band where the mycelia of the mutant colonies were in contact (Dekkers et al., 2007 and unpubl. data). Thus, the functions of CTB2 and CTB3 as an O-methyltransferase were not redundant and both were required for cercosporin biosynthesis.

Cercosporin has a bilateral symmetrical structure (Fig. 8). Thus, formation of the mature cercosporin is likely mediated by dimerization of two identical polyketomethylene units. The enzymatic reaction directly contributing to dimerization is currently unknown. It is also possible that the CTB3, 5, 6, and/or 7 translational products may be involved in dimerization. Alternatively, dimerization of polyketomethylene units may proceed non-enzymatically as noted for trichothecene biosynthesis in Fusarium (McCormick et al., 1990).

We propose that exportation of cercosporin outside fungal cells is mediated by a putative 12-transmembrane MFS transporter encoded by CTB4 (Choquer et al., 2007). Previous studies have argued that CFP, a putative 14-transmembrane MFS transporter, is the major transporter responsible for cercosporin secretion in Cercospora kikuchii (Callahan et al., 1999). Interestingly, comparative analysis showed that CFP has no similarity to CTB4, and the CFP gene is not linked with the CTB gene cluster. The respective roles of these two transporters are not known at this time. Finally, the CTB core gene cluster was apparently coregulated through CTB8, which encodes a Zn(II)Cys6 transcriptional activator, as expression of the CTB1-CTB7 genes was completely or nearly abolished in two Δctb8 disruptants.

Unlike the CTB1-CTB8 genes, the ORF9 and ORF10 transcripts were present in all conditions tested, thus representing the end of the CTB cluster on the right. ORF11, immediately adjacent to CTB8, is located in the opposite end of the cluster. Expression of ORF11 appeared to be light regulated, but was unaffected by medium composition. The deduced polypeptide of ORF11 showing low similarity to Zn(II)Cys6 transcriptional regulators has only one zinc-binding motif, suggesting that ORF11 may encode a non-functional protein. Attempts to disrupt ORF11 using similar split marker approach failed to identify any cercosporin deficiency mutants out of 385 transformants screened, suggesting that ORF11 likely plays no role in cercosporin biosynthesis. The conceptually translated ORF12 also has low similarity to GAL4-like, Zn(II)Cys6 transcriptional regulators, but completely lacks the zinc-binding domain. CTB8 did not regulate expression of ORFs 9–12 as accumulation of these gene transcripts in two Δctb8 disruptants had no significant difference from wild type. The results support CTB1-CTB8 as the core biosynthetic cluster for cercosporin biosynthesis.

Many microorganisms including Cercospora spp. that produce biologically toxic secondary metabolites have mechanisms for self-protection to avoid suicide (Daub et al., 2005). In some cases, the genes contributing to self-defence are also situated in the biosynthetic gene cluster. For example, genes for longevity assurance factors and an ABC transporter have been identified in the fumonisin biosynthetic gene cluster in Gibberella moniliformis (Proctor et al., 2003). TRI12, responsible for toxin transport and self-protection in Fusarium sporotrichioides clusters with the trichothecene biosynthetic genes (Alexander et al., 1999). SIRA in Leptosphaeria maculans, located in the sirodesmin biosynthetic cluster, is not required for production, but is involved in self-protection (Gardiner et al., 2005). In Cercospora, genes have been shown to have a dual role in cercosporin resistance and biosynthesis. For example, inactivation of the CFP gene encoding a putative cercosporin transporter in C. kikuchii yielded a mutant deficient in cercosporin production as well as in resistance (Callahan et al., 1999). Further, disruption mutants for CRG1, encoding a putative Zn(II)Cys6 transcriptional activator, resulted in a parallel reduction in both cercosporin production and the ability to tolerate cercosporin toxicity (Chung et al., 1999; 2003a). To date, however, we have no evidence that the CTB biosynthetic cluster contains genes involved in resistance. Disruption of the CTB8 gene resulted in a mutant that failed to express the related genes in the cluster, but retained normal resistance to exogenous cercosporin and to other singlet oxygen-generating photosensitizers (data not shown). As auto-resistance has frequently been associated with membrane transporters, we specifically tested a disruption of CTB4 encoding a putative MFS transporter. Disruption of CTB4 yielded a mutant that was partially defective in cercosporin production but was unchanged in cercosporin sensitivity (Choquer et al., 2007), and ruled out its role in self-protection against cercosporin.

Production of cercosporin is regulated by a wide range of factors such as light, medium composition, cultural age, growth stage and developmental stage (Jenns et al., 1989; Ehrenshaft and Upchurch, 1991; Daub and Ehrenshaft, 2000), and also appears to be regulated by a signalling network interplay between Ca2+/calmodulin and a MAP kinase pathway (Chung, 2003; Shim and Dunkle, 2003). Further, there is a regulatory relationship between cercosporin production and resistance, as evidenced by the dual phenotype of crg1 disrupted mutants (Chung et al., 1999; 2003a). These findings suggest a complex and intertwined regulatory network, where CTB8 is a pathway-specific regulator and CRG1 has broader regulatory ability to activate transcript levels of the genes responsible for both biosynthesis and resistance. Figure 9 shows a proposed model for regulation. The fungus receives signal cues, particularly from light, through membrane receptors that subsequently activate the Ca2+/calmodulin and/or the MAP kinase signalling pathways. CRG1, which is unlinked to the CTB gene cluster, appears to exert its regulation of cercosporin biosynthesis, at least in part, by activating the transcription of CTB8; disruption of the CRG1 gene in the 205C3 mutant caused a marked reduction in the CTB1, CTB3, CTB7 and CTB8 gene transcripts (Fig. 5), indicating that CRG1 may directly and indirectly trigger expression of the CTB8 transcript. CTB8 is specifically responsible for activating expression of the other CTB (17) genes, which eventually leads to cercosporin production. Biosynthesis of secondary metabolites in fungi is often regulated by various environmental factors (Yu and Keller, 2005). In addition to CRG1, many positive- or negative-acting globe regulatory factors such as AreA (nitrogen regulatory protein; Marzluf, 1997), PacC (pH regulatory protein; Espeso and Arst, 2000), WC (light regulatory proteins; Iwasaki and Dunlap, 2000), and/or AP-1 (oxidative stress-responsive transcription activator; Toone and Jones, 1999) might be involved in cercosporin biosynthesis and/or self-defence as well. Recently, a novel transcriptional regulator, LaeA, has been shown to be involved in a global regulation of the gene clusters for a wide range of secondary metabolites including gliotoxin, lovastatin, penicillin and sterigmatocystin in Aspergillus spp. (Bok and Keller, 2004). Whether or not those transcription orthologues are involved in the regulation of cercosporin biosynthesis in Cercospora spp. awaits further investigation by disrupting the corresponding genes.

Figure 9.

Proposed signal transduction and regulatory controls for cercosporin biosynthesis and self-protection via two Zn(II) Cys6-containing transcription regulators, CRG1 and CTB8, in Cercospora nicotianae. Other transcriptional regulators such as AP-1, AreA, LaeA, PacC and WC may also be involved in the regulation of the biosynthetic pathway leading to cercosporin production. Solid arrows indicate known pathways, whereas dotted arrows indicate possible regulation. Question marks indicate other hypothesized yet unknown factors or signals. A cross mark indicates an inhibition of cercosporin, likely via blocking expression of CTB8, and the entire biosynthetic cluster genes. A blocked line indicates the feedback inhibitory activity of the CTB1–8 genes.

The data presented in Fig. 6 suggest the presence of a feedback inhibition mechanism, in which disruption of one of the biosynthetic genes in the pathway blocked cercosporin production and in turn inhibited expression of the other CTB genes in the cluster. However, the inhibition may not be absolutely stringent as evidenced that disruption of the CTB1 gene only partially reduced accumulation of the CTB3 and CTB8 gene transcripts in four Δctb1 disruptants tested. Expression of the CTB2 and CTB4 genes was completely abolished in two Δctb1 disruptants (D6 and D7) but unchanged in other two Δctb1 disruptants (D1 and D11). This could be the result of ‘leaky’ expression of those genes in the D1 and D11 disruptants, further indicating low stringency of the feedback regulation. Disruption of the CTB2 gene, however, completely inhibited expression of the CTB1, CTB2, CTB3, CTB4 and CTB8 genes. In contrast, a previous study revealed that disruption of the CTB4 gene resulted in a mutant that accumulated less cercosporin in culture but retained normal expression of the CTB1, CTB2, CTB3 and CTB8 genes (Choquer et al., 2007). The feedback regulations appear to have various effects on the expression of the CTB cluster genes. It is very likely that expression of the CTB cluster genes, collectively as a cluster or individually, might be regulated by other transcription factors.

Promoter analysis was conducted to identify putative regions involved in regulation. Many of the Zn(II)Cys6-harbouring transcriptional activators recognize and bind a palindrome DNA sequence with inverted repeats of CGG elements separated by a spacer with various nucleotides [5′-CGG(N4−11)CCG]-3′ (Marmorstein and Harrison, 1994; Li and Kolattukudy, 1997). However, AFLR in both E. nidulans and A. parasiticus tends to bind the palindromic sequence 5′-TCG(N5)CGA-3′ (Fernandes et al., 1998; Ehrlich et al., 2002). Promoter analysis showed that all the CTB genes except for CTB2 have several sets of the palindromic sequence 5′-GCC(N4−11)GGC-3′ in the promoter regions. All the CTB genes have the palindromic sequence 5′-TCG(N3−6)CGA-3′ that was not found in ORFs 9, 11 and 12. It will require more experimental evidence to determine if CTB8 can bind to the palindromic sequence 5′-TCG(N3−6)CGA-3′ and/or 5′-GCC(N4−11)GGC-3′. Other regulatory regions were also identified. The promoter regions in all CTB1-CTB8 genes have one to three putative AreA binding sequences (GATA), suggesting that biosynthesis of cercosporin might be regulated by nitrogen. Given that the light regulation of the CTB1–8 gene expression was more obvious than the nitrogen regulation, one would assume that the AreA consensus sequences were also likely the binding sites for WC1/WC2 light regulatory complex that posses GATA-type DNA-binding domains in N. crassa (Ballario et al., 1996; Linden and Macino, 1997). A putative PacC conserved sequence was found in the promoter regions of CTB3, CTB4, CTB6, CTB 7 and CTB8, but was absent in the CTB1, CTB2 and CTB5 promoters. None of the CTB genes has the CreA (carbon regulatory protein) (Dowzer and Kelly, 1991) binding site in the promoter regions, suggesting that expression of the CTB genes may not be directly responsive to the carbon sources.

In the present study, Northern-blot analyses revealed that the CTB1, 2, 5, 6, 7 and 8 gene transcripts were accumulated to high levels when the fungus was grown on CM with mannitol and calcium nitrate as the carbon and nitrogen sources, respectively (Fig. 7). Depletion of nitrogen or substitution of nitrate with ammonium in the medium apparently blocked their expression. Similar results were observed in the expression of CTB3 (Dekkers et al., 2007). Despite that there is no CreA binding site in the promoters of the CTB1–8 genes, substitution of glucose with mannitol as the sole carbon source increased accumulation of the CTB1–8 gene transcripts, yet did not enhance cercosporin production. Expression of the CTB1–8 genes and production of cercosporin could also be affected by pH. However, this study revealed that accumulation of the CTB1–8 gene transcripts and cercosporin were more likely affected by the citrate-phosphate buffer rather than by the pH values (Fig. 7). Although several CTB1–8 gene promoters have putative PacC binding motifs (Table 2), there were no direct correlations between cercosporin production and accumulation of the CTB1–8 gene transcripts in different pH. Accumulation of the CTB1–8 gene transcripts and cercosporin was, however, co-ordinately regulated by light. All CTB1-CTB8 gene promoters also have CAAT consensus sequences that could be recognized by many other transcriptional regulators. These results imply a regulatory complexity in the biosynthetic pathway to form cercosporin.

Our studies are continuing to further define the pathway for cercosporin biosynthesis. Characterization of intermediates accumulated in various mutants will likely provide more definitive data on the pathway. Disruption of CTB8 blocked cercosporin production but did not lead to the accumulation of obvious intermediates, most likely because Δctb8 mutants were downregulated for the entire pathway. Similarly, disruption of CTB1 encoding the polyketide synthase failed to detect any pigments (Choquer et al., 2005), as this gene encodes the first step in the pathway. By contrast, disruption of CTB2 or CTB3 resulted in accumulation of a yellow pigment (Fig. 2 and Dekkers et al., 2007), whereas disruption of CTB4 yielded a brown pigment (Choquer et al., 2007). Preliminary analysis of the yellow-brown pigments that were accumulated by the CTB2, CTB3 or CTB4 disruptants failed to identify any distinct peaks by spectrophotometry or HPLC (data not shown). It is presently unknown if those pigments are the intermediates for cercosporin biosynthesis. Continued characterization of the CTB1–8 genes and pathway intermediates will lead to a fuller understanding of the biosynthetic pathway for this important polyketide toxin and for pathway regulation.

Experimental procedures

Biological materials and cultural conditions

The wild-type (ATCC18366) Cercospora nicotianae (Ellis & Everh.) and genetically modified strains were maintained on a complete medium (CM) (Jenns et al., 1989). The Δctb1 (D1, D6, D7 and D11) and Δctb3 (D4, D5 and D16) disruptants were created in previous studies (Choquer et al., 2005; Dekkers et al., 2007). Fungal mutants defective in cercosporin biosynthesis were screened and identified by the lack of red pigmentation (due to cercosporin production) when grown on potato dextrose agar (PDA, Difco) plates by the method described (Chung et al., 2003b). Assays for sensitivity to photosensitizing compounds (cercosporin, eosin Y, haematoporphyrin, methylene blue or toluidine blue) were performed by growing fungal isolates on CM medium containing 10 or 100 μM of the test compound under continuous light as described by Jenns and Daub (1995). Cercosporin and photosensitizing compounds were purchased from Sigma-Aldrich (St. Louis, MO), and dissolved in acetone or water to make a 10 or 100 mM stock solution as appropriate. The pH of medium was adjusted appropriately with 0.1 M citric acid and 0.2 M dibasic sodium phosphate buffer.

Purification and quantification of cercosporin

Cercosporin and the biosynthetic intermediates were extracted with 5 N KOH or with ethyl acetate from agar plugs cut from mycelial cultures as described previously (Chung, 2003; Choquer et al., 2005). Cercosporin and the biosynthetic intermediates in the KOH extracts were quantified by measuring absorbance at 480 nm using a model Genesys 5 spectrophotometer (Spectronic Instruments, Rochester, NY). The ethyl acetate extracts were separated on thin-layer chromatographic (TLC) plates coated with a 60 F254 fluorescent silica gel using ethyl acetate: hexane: methanol: H2O (6:4:1.5:1, v/v) mixture as the solvent.

Chromosomal walking and sequence analysis

Genomic DNA of C. nicotianae was isolated using a DNeasy Plant Mini kit (Qiagen, Valencia, CA). A DNA library of C. nicotianae was constructed from genomic DNA digested with DraI, EcoRI, PvuI and StuI, and ligated to adaptors using a Universal GenomeWalker kit (BD Biosciences, Palo Alto, CA) according to the manufacturer's instructions. To obtain unknown genomic regions, primers were designed to complement sequenced regions and used for two rounds of PCR amplification with adaptor primers using a Titanium or Advantage 2 DNA polymerase (BD Biosciences). The amplified DNA fragments were purified with a DNA purification kit (Mo Bio Laboratories, Carlsbad, CA) and either directly sequenced or cloned into pGEM-T easy vector (Promega, Madison, WI) for sequence analysis from both directions at Eton Bioscience, (San Diego, CA). PCR primers were synthesized by Integrated DNA Technologies (Coralville, IA) and Allele Biotechnology and Pharmaceuticals, (San Diego, CA). blast similarity searches (Altschul et al., 1997) were performed at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). ORF and exon/intron positions were deduced from comparisons of genomic and cDNA sequences and/or predicted using the fgenesh gene-finding software at http://www.softberry.com. A search for functional domains was conducted using the prosite database in the ExPASy Molecular Biology Server (http://us.expasy.org) and Motif/ProDom and Block programs (http://motif.genome.jp/). Analysis of the CTB promoter regions was conducted using regulatory sequence analysis tools (van Helden, 2003) at http://rsat.ulb.ac.be/rsat/ and the output threshold was set at 1 (no mismatches were allowed). Palindrome searches were performed at http://bioweb.pasteur.fr/seqanal/interfaces/palindrome.html.

Disruption of CTB2 and CTB8 genes

Fungal protoplasts were prepared and transformed using CaCl2 and polyethylene glycol by previously described methods (Chung et al., 2002). To disrupt CTB8, a 2.4 kb DNA fragment encompassing the entire CTB8 ORF and its 5′- and 3′ untranslated regions was amplified using a high fidelity DNA polymerase (Roche Applied Science) with primers TF1 (5′-ctggatcaccgcagaggaagc-3′) and TF2 (5′-gcctgagcgcccactgagta-3′), and cloned into the pGEM-T easy vector to yield pCTB8. A 1.1 kb KpnI–BglII fragment in pCTB8 was replaced with a 1.6 kb BamHI end-filled fragment harbouring the hygromycin phosphotransferase B gene (HYG) cassette from pUCATPH (Lu et al., 1994) to generate a disruption construct named pΔctb8 (Fig. 3). PCR products containing truncated HYG and CTB8 fusion fragments were amplified from pΔctb8 for gene disruption. A 1.5 kb fragment containing 5′-CTB8 fused with 3′HYG was amplified with primers TF2 and hyg3 (5′-ggatgcctccgctcgaagta-3′). A 2.0 kb fragment containing 5′HYG fused with 3′CTB8 was amplified with primers TF1 and hyg4 (5′-cgttgcaagaactgcctgaa-3′). PCR fragments, overlapping within the HYG region, were purified with a PCR clean-up kit (Mo Bio Laboratories), mixed, and directly transformed into the C. nicotianae wild-type strain for gene disruption.

To disrupt CTB2, a 2.7 kb fragment containing the full-length CTB2 gene was amplified from C. nicotianae genomic DNA by PCR with primers mts1 (5′-ggattcccgatcttggcgtaag-3′) and mts2 (5′-catgcgagatggtttggggattgt-3′) and cloned into pGEM-T easy vector to form pCTB2 (Fig. 2). A 1.1 kb fragment in pCTB2 was replaced with the HYG cassette to generate the disruption construct pΔctb2. The CTB2 fragments fused with the HYG marker were amplified with primers mts2 and hyg4 (5′-cgttgcaagaactgcctgaa-3′) and primers mts1 and hyg3 (5′-ggatgcctccgctcgaagta-3′), respectively, and used for gene disruption.

Genetic complementation

Genetic complementation was conducted by cotransformation of a PCR fragment containing the full-length CTB2 or CTB8 ORF and its endogenous promoter with plasmid pBARKS1 carrying the BAR gene responsible for Ignite/basta (bialaphos) resistance under control of the Aspergillus nidulans trpC promoter [Pall and Brunelli, 1993; obtained from the Fungal Genetics Stock Center (FGSC)] or with the pCB1532 plasmid carrying the Magnaporthe grisea acetolactate synthase gene (SUR) cassette for sulfonylurea resistance (Sweigard et al., 1997; obtained from the FGSC) into the Δctb8-D4 or Δctb2-D3 null mutant. Putative transformants were selected on medium containing 250 μg ml−1 hygromycin (Roche Applied Science), 5 μg ml−1 sulfonylurea (chlorimuron ethyl; Chem Service, West Chester, PA) or 50 μg ml−1 bialaphos (Phytotechnology Laboratory, Lenexa, KS) as appropriate and assessed for cercosporin production on PDA plates as described previously (Chung, 2003).

Reverse transcriptase (RT)-PCR and cDNA isolation

Fungal RNA was isolated with TRIZOL reagent (Invitrogen). The poly (A+) mRNA was purified with an Oligotex kit (Qiagen). Double strand cDNA of C. nicotianae was generated with the BD SMART PCR cDNA synthesis kit (BD Biosciences) according to the manufacturer's instructions. Each cDNA fragment encompassing the entire ORF and its partial 5′- and 3′ regions was subsequently amplified from the cDNA pools with a TITANIUM Taq polymerase (BD Biosciences) and the respective gene primers, and the fragments were purified and directly subjected to sequence analysis.

Manipulation of nucleic acids

All plasmids were propagated in Escherichia coli DH5α bacterial cells and purified using a Wizard DNA purification kit (Promega). Standard procedures were used for endonuclease digestion of DNA, electrophoresis, and Southern-blot hybridizations (Sambrook and Russell, 2001). For Northern-blot hybridizations, total RNA was denatured in a formaldehyde-containing agarose and buffer solution as described previously (Chung et al., 2003c), blotted onto a positively charged nylon membrane, and hybridized to a PCR-generated DNA probe as appropriate. The hybridization probes were generated by PCR with gene-specific primers to insert a digoxigenin (DIG)-11-dUTP into the CTB1–8 DNA fragments. The conditions and procedures used for probe labelling, hybridization, post-hybridization washing and immunological detection of the probe using a CSPD chemofluorescent substrate for alkaline phosphatase were carried out according to the manufacturer's recommendations (Roche Applied Science).

Nucleotide sequences

Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under Accession nos. AY649543 (CTB1), DQ991505 (CTB2), DQ355149 (CTB3), DQ991506 (CTB4), DQ991507 (CTB5), DQ991508 (CTB6), DQ991509 (CTB7), DQ991510 (CTB8), DQ993249 (ORF9), DQ993250 (ORF10), DQ993251 (ORF11) and DQ993252 (ORF12).

Ancillary