Passalora arachidicola isolate ATCC18667 was obtained from the American Type Culture Collection (Manassas, Virginia) and maintained at Massey University as strain NZ46. Growth was on potato dextrose agar (PDA) at 22 °C unless otherwise stated. For broth cultures, one 5-mm diameter plug of mycelium was macerated with a pestle in H₂O and used to inoculate between two and four 125 mL flasks containing 25 mL potato dextrose (PD) broth for genomic DNA preparation or Dothistroma medium (DM; [18]) for monitoring growth and gene expression. Cultures were grown at 22 °C and 160 rpm for 14 days (gDNA preparation) or for up to 30 days with three biological replicates per time point (growth experiment), then filtrates and mycelia harvested by filtration through Miracloth (Calbiochem Corporation, La Jolla, CA). Mycelia for RNA extraction were snap-frozen in liquid nitrogen and those for DNA extraction or biomass quantification were freeze-dried. For quantification, dothistromin was extracted from 5 mL filtrate using three successive treatments with 5 mL chloroform, air-dried and dissolved in 50 µL of DMSO. TLC was carried out as described in [22] with serial dothistromin standards. The densitometric values for each dothistromin band on the TLC plate were determined using a BioRad Gel Doc Documentation (BioRad) and Gel-Pro Analyser (Media Cybernetics, Inc., Bethesda, MD).

Genomic DNA was isolated from NZ46 using a CTAB method [19]. Fragments of putative P. arachidicola cypA, dotA, vbsA and pksA genes were PCR amplified using degenerate primers designed from alignments of homologous A. parasiticus, A. nidulans and D. septosporum genes. PCRs were performed in 25 µl reactions containing 0.5 U Taq DNA polymerase (Invitrogen, Carlsbad, CA), 1× Invitrogen PCR buffer, 1.5 mM MgCl₂, 50 mM dNTPs, 0.4 mM of each primer (see Supplementary Table S1) and 10 ng NZ46 genomic DNA. Amplification was done with a gradient Mastercycler (Eppendorf, Hamburg) with an initial step of 94 °C for 2 min, followed by three cycles of 94 °C for 30 s, 66 °C for 60 s, and 72 °C for 80 s. In subsequent cycles the c 3 °C every three cycles down to 42 °C, followed by 72 °C for 5 min. PCR products were purified using a QIAquick PCR Purification Kit (Qiagen, Hilden, Germany), ligated into pGEM-T easy vector (Promega Corp., Madison, WI), transformed into E. coli Top10 (Invitrogen), plasmids extracted using a QIAprep spin miniprep kit (Qiagen) and sequenced using an ABI Prism Big-Dye Terminator cycle sequencing ready reaction kit and ABI3730 Genetic Analyzer (Applied Biosystems, Foster City, CA).

Clones containing genes clustered alongside the dothistromin gene fragments identified by degenerate PCR were obtained by construction and screening of size-fractionated genomic libraries. Southern hybridization was firstly carried out to determine the best libraries and fractions to use for each available gene fragment and to assess copy number. P. arachidicola NZ46 genomic DNA (2 µg) was digested with EcoRI, SalI, BamHI or ScaI and fractionated on a 0.7% agarose gel. Southern hybridization using DIG-labeled probes (Roche) amplified from dothistromin genes dotA, cypA, pksA and vbsA was done as described previously [10]. For partial genomic library construction, 6–10 kb and 10–14 kb EcoRI-digested DNA fractions were recovered using a QIAquick gel extraction kit (Qiagen), ligated into plasmid vector pIC19H and cloned in E. coli Top10. Libraries were screened by plate colony hybridization as described [10] using the same DIG-labeled probes as above. Plasmid DNAs were isolated and sequenced on both strands as above, using a primer walking strategy.

Based on these sequenced plasmids and the degenerate PCR products, inverse PCR was carried out as described by Ochman et al. [20] to extend the sequenced regions. Genomic DNA was digested with restriction enzymes EcoRI, HindIII, KpnI and NcoI, respectively, then self-ligated with T4 DNA ligase (Roche) at 4 °C overnight. PCR reactions were carried out as described in Section 2.2 using primers SZDP-82 and SZDP-83 with EcoRI ligation for extension from Pa-dotA, primers MGPA22 and MGPA35 with KpnI ligation for Pa-pksA, primers MGPA66 and MGPA69 with NcoI ligation for Pa-moxA and primers MGPA46 and MGPA79 with HindIII ligation for both Pa-hexA and Pa-vbsA ends (for primers, see Supplementary Table S1). PCR products were cloned using a pGEM-T easy vector and then sequenced.

Total RNA was isolated from frozen mycelia using TRIzol reagent (Invitrogen) and quantified using a NanoDrop spectrometer (NanoDrop Technologies, Wilmington, Delaware). RNA quality was assessed on a sodium dodecyl sulfate (SDS; 0.3%) tris-acetate-EDTA (TAE) buffer agarose gel before treatment with TURBO DNase (Ambion, Austin, TX). For first strand cDNA synthesis, 500 ng DNase-treated total RNA was reverse transcribed using random hexamer primers and superscript^TM III RT (Invitrogen) in a 20 µL reaction, following the manufacturer’s instructions. Semi-quantitative RT-PCR was performed to measure the expression of Pa-dotA, Pa-pksA, Pa-vbsA and control β-tubulin (Pa-tub; accession number HM587712) genes using primers listed in Supplementary Table S1. To exclude false positive PCR products from contaminating genomic DNA, at least one primer of each pair for each gene was designed to exon regions flanking an intron. PCR reactions were set up as in Section 2.2 above except with 1 µL of diluted cDNA reaction from each time point as template. The dilution of the individual cDNA samples was determined empirically to show exponential phase amplification for each gene at 30 cycles, and assessed in triplicate. Cycling conditions were 2 min at 94 °C; 30 cycles of 30 s at 94 °C, 30 s at 60 °C, and 60 s at 72 °C; followed by 5 min at 72 °C. PCR samples were electrophoresed on 2.5% agarose TBE gels, stained with ethidium bromide (10 µg/mL) and densitometric values for each PCR product determined as described in Section 2.1. RT-PCR values are presented as a ratio of the dothistromin gene signal relative to that of β-tubulin.

The Pa-dotA complementation plasmid pR280 was constructed with a D. septosporum dotA promoter region (transcriptional fusion) using a three-step procedure. Primer sequences are shown in Supplementary Table S1. Firstly a region including the open reading frame (904 bp) and 3’ UTR (372 bp) of the Pa-dotA gene was PCR amplified using primers SZDP-137 and SZDP-130 and 962 bp of the promoter region of the Ds-dotA gene was amplified with primers SZDP-135 and SZDP-136. Primers SZDP-136 and SZDP-137 overlapped both amplicons and were complementary to each other. In the second step, the two PCR products were combined in a further round of PCR using primers SZDP-135 and SZDP-130. Finally, the combined 2.2 kb product was cloned using the pGEM-T easy vector (Promega) to make plasmid pR280, and the sequence confirmed by DNA sequencing. For complementation of D. septosporum vbsA KO FJT12, plasmid pR283 containing the 5.8 kb EcoRI fragment (556–6321 bp, GenBank accession No. GU566731), constructed initially for dothistromin gene cloning, was used directly.

Protoplasts of D. septosporum dothistromin-deficient mutants FJT1 (∆dotA) and FJT12 (∆vbsA) were prepared and transformed using methods described previously [21,22]. FJT1 was transformed with pR280 (Pa-dotA construct) and FJT12 with pR283 (Pa-vbsA construct); in both cases circular plasmids were co-transformed with plasmid pBC-phleo (pR224, [23]) to enable selection on DM medium containing 7 µg/mL phleomycin (Apollo Scientific Ltd., Stockport, UK). After single spore purification, putative complemented transformants were analyzed for the presence of Pa-dotA or Pa-vbsA genes by PCR as shown in Supplementary Figure S1. To assay dothistromin production, D. septosporum mutants complemented with the Pa-dotA or Pa-vbsA gene were grown in Dothistroma broth [8] as described above for P. arachidicola. After 11 days mycelia were harvested by filtration and the growth media analyzed for dothistromin content with TLC as described previously [22].

Sequence data were assembled into contigs using VectorNTI (Invitrogen) and analyzed using VectorNTI or MacVector (Accelrys Inc., San Diego, CA). Sequence comparisons were performed at NCBI (http://www.ncbi.nlm.nih.gov/) using BLAST [24]. GenBank accession numbers for the P. arachidicola NZ46 dothistromin gene cluster sequences are GU566729, GU566730 and GU566731.

PCR with degenerate primers targeted to cypA, dotA, pksA and vbsA genes yielded products of the expected size (Supplementary Table S1) and with predicted amino acid sequence identities of at least 90% compared to the respective D. septosporum homologs. Southern blot analysis confirmed that each sequence was unique and present as a single copy in the P. arachidicola genome (Supplementary Figure S1). These PCR products were therefore probably derived from P. arachidicola dothistromin biosynthetic genes that have been designated dotA (ketoreductase), cypA (averufin monooxygenase) pksA (polyketide synthase) and vbsA (versicolorin B synthase). To clearly distinguish between genes from P. arachidicola and D. septosporum, the gene names are prefixed by Pa- or Ds-, respectively, in this article.

Full sequences for Pa-cypA, Pa-dotA, Pa-vbsA genes and a partial sequence for Pa-pksA were obtained from library clones. For each of these, nucleotide and amino acid identities were much higher when compared to D. septosporum homologs than to A. parasiticus (AF cluster) or A. nidulans (ST cluster) homologs, as would be expected for vertical inheritance (Table 1; genes in bold type).

Functional analysis was performed to confirm that P. arachidicola genes are involved in dothistromin biosynthesis. No transformation system has been developed for P. arachidicola and attempts to make gene replacement mutants in this species were not successful (results not shown). Instead, two of the P. arachidicola genes were analyzed by functional complementation of previously characterized dothistromin-deficient D. septosporum gene-replacement mutants [10,21]. Transformation of Pa-dotA into a Ds-dotA mutant and of Pa-vbsA into a Ds-vbsA mutant was confirmed by PCR (Supplementary Figure S2). PCR products of Pa-vbsA and Pa-dotA genes integrated into the D. septosporum transformants were sequenced to ensure that they matched P. arachidicola and not D. septosporum dothistromin genes. Transformation of either Pa-dotA or Pa-vbsA in the respective D. septosporum mutants restored the ability to produce dothistromin (Figure 1), confirming that the Pa-dotA and Pa-vbsA gene products are functional in dothistromin biosynthesis. Complementation was not done for Pa-cypA as no corresponding D. septosporum mutant was available, or for Pa-pksA because only partial gene sequence was obtained during this study.

The production of dothistromin by P. arachidicola and expression of dothistromin genes were assessed over a time-course in culture (dothistroma broth; DB). Figure 2 shows that both the rate of dothistromin biosynthesis (µg/mg DW biomass) and expression of Pa-dotA, Pa-vbsA and Pa-pksA genes were highest during early exponential growth phase. These results are very similar to those reported for D. septosporum [8], where an unusual early growth-stage expression of dothistromin biosynthesis was shown. The experiment was repeated with potato dextrose broth instead of DB and similar results (not shown) obtained to those found in D. septosporum [8]. Although further studies under different growth conditions need to be assessed, P. arachidicola and D. septosporum do appear to show a different pattern of regulation for dothistromin biosynthesis compared to that for aflatoxin biosynthesis by Aspergillus spp., which predominantly involves late-exponential/stationary phase production [9].

The discovery that P. arachidicola, like D. septosporum, produces dothistromin during early exponential growth stage in culture suggests that the pattern of regulation may have some functional importance for the role of dothistromin. The timing of dothistromin biosynthesis in planta is not known for D. septosporum, but the appearance of red bands in pine needles is secondary to the appearance of necrotic lesions. This, and other reasoning, led us to propose that dothistromin is mainly produced during a period of rapid growth by D. septosporum that occurs following nutrient release from dead pine tissue, and that the role of dothistromin may be to inhibit the growth of other needle-dwelling competitor fungi [6]. Although many aspects of this hypothesis remain to be investigated, the toxicity of dothistromin to some needle-dwelling fungi has been demonstrated [6]. Whether dothistromin has any biological role in the P. arachidicola-peanut interaction is even less well known. It is not known if dothistromin is produced in peanut leaves infected with P. arachidicola, although P. arachidicola does have a latent phase in planta prior to its necrotrophic phase [17].

P. arachidicola genomic library clones that hybridized in Southern blots to Pa-dotA (14.7 kb clone), Pa-vbsA (5.8 kb clone), Pa-cypA (6.3 kb clone) and Pa-pksA (5.2 kb clone) (Supplementary Figure S1) were sequenced. The Pa-cypA and Pa-pksA clones were found to be contiguous as they both overlapped with a Pa-pksA gene fragment produced by degenerate PCR (Section 3.1). Sequences were extended by inverse PCR as described above to yield the three Pa-DOT contigs shown in Figure 3.

Examination of the predicted genes in the P. arachidicola clones revealed a remarkable similarity with dothistromin mini-clusters in D. septosporum. As shown in Figure 3, the order and orientation of putative dothistromin genes (colored) in the P. arachidicola clones and D. septosporum mini-clusters are highly conserved, with only two notable differences. The first main distinguishing feature is that the moxA gene is in a different orientation with respect to the rest of the grouped genes, being divergently transcribed from Pa-epoA in P. arachidicola but in the same orientation as Ds-epoA in D. septosporum. The reason for this distinction is unclear. The intergenic regions between the epoA and moxA genes are of similar length (529 bp in D. septosporum and 485 bp in P. arachidicola), show 53% nucleotide identity and have no obvious repeated sequences or unusual GC composition (Supplementary Figure S3). However, the adjacent Pa-epoA gene sequence contains evidence of genomic rearrangement, as discussed below, and this may be related to the inversion of Pa-moxA.

A second notable difference in gene synteny between the two species occurs next to the dotA-dotD gene mini-cluster region (Figure 3, right). Whilst Ds-dotD is flanked by a 10.6 kb region devoid of genes [10], Pa-dotD is only 140 bp away from a putative MFS transporter gene (Pa-mfsA), which is closely followed by another open reading frame (hyp, of unknown function). There is a high level of nucleotide identity (54%) between the 140 bp Pa-dotD: Pa-mfs intergenic region and the corresponding region next to Ds-dotD and no clear difference in GC composition in this region (see Supplementary Figure S3 and [10]). The Pa-mfsA gene is unrelated to the Ds-dotC MFS transporter gene, with only 13% amino acid identity.

As well as an overall conservation of dothistromin gene synteny, D. septosporum and P. arachidicola dothistromin genes are similar in sequence, with nucleotide identities ranging from 66–86% and predicted amino acid identities of 73–96% (Table 1). The most notable exception to this is between the epoA genes, with only 30% nt and 22.5% aa identity. The Ds-epoA gene has no homolog in the AF or ST clusters but is predicted to encode a putative 420 aa epoxide hydrolase. Whilst its function is unknown, hypotheses of a role in biosynthesis or detoxification have been proposed [22]. In contrast, Pa-epoA is a pseudogene with three stop codons in the predicted coding region and replacement of approximately 650 bp of expected epoA coding sequence with a 200 bp repeat-rich region. In addition to this, although both Ds-epoA and Pa-epoA each have one predicted intron, the positions and sizes of these introns differ. Repeat-rich regions have been associated with other fragmented secondary metabolite gene clusters such as an indole-diterpene cluster in the grass endophyte Neotyphodium lolii [12].

Like D. septosporum, the putative P. arachidicola dothistromin genes have sequences matching binding sites for the AF/ST pathway-specific regulatory protein AflR. In both species, all but one of the dothistromin genes have putative AflR-binding sites just upstream of the coding region, or in an intergenic region shared with a divergently transcribed gene (Supplementary Figure S3). The only genes devoid of upstream AflR sites are Ds-avfA and Pa-avfA. Since Ds-avfA is co-regulated along with other dothistromin genes [10] this suggests that additional regulatory mechanisms are in place. A putative aflR ortholog has recently been found in the newly completed D. septosporum genome sequence but whether it is involved in the regulation of dothistromin biosynthesis is not yet known.

The high level of sequence similarity between P. arachidicola and D. septosporum genes is consistent with the very close phylogenetic relatedness of the Dothistroma and Passalora genera recently suggested by Crous et al. [25]. In contrast, A. nidulans and A. parasiticus, although belonging to different sections of the genus Aspergillus, have a very different gene order (Figure 3) and amino acid identities for homologs of the dothistromin genes range from 55–90% [26]. In general, P. arachidicola genes showed a higher level of identity to A. parasiticus than to A. nidulans genes (Table 1), but this trend was not observed with all genes examined.

The conservation of gene order and sequence between D. septosporum and P. arachidicola is remarkable. It seems that a very similar ‘fragmented cluster’ of dothistromin genes is present in these two species. Whether the fragmented clusters are present in equivalent regions in the two species is not known, although in D. septosporum the three mini-clusters are all located on one 1.3 Mb chromosome [22]. The fragmentation may be even more severe than previously supposed [10]: a recent re-evaluation of D. septosporum dotB, dotD and epoA genes sheds doubt on whether they should be classified as ‘dothistromin genes’ [27]. The discovery that Pa-epoA is a pseudogene adds support to notion that Ds-epoA may not be required for dothistromin biosynthesis. The availability of the genome sequence of D. septosporum (sequenced by the Joint Genome Institute) will enable us to examine the full extent of dothistromin gene organization, assess the regulation of these genes and develop new hypotheses about the evolution of secondary metabolite gene clusters.