Two histone deacetylases, FfHda1 and FfHda2, are important for Fusarium fujikuroi secondary metabolism and virulence

Abstract

Histone modifications are crucial for the regulation of secondary metabolism in various filamentous fungi. Here we studied the involvement of histone deacetylases (HDACs) in secondary metabolism in the phytopathogenic fungus Fusarium fujikuroi, a known producer of several secondary metabolites, including phytohormones, pigments, and mycotoxins. Deletion of three Zn(2+)-dependent HDAC-encoding genes, ffhda1, ffhda2, and ffhda4, indicated that FfHda1 and FfHda2 regulate secondary metabolism, whereas FfHda4 is involved in developmental processes but is dispensable for secondary-metabolite production in F. fujikuroi. Single deletions of ffhda1 and ffhda2 resulted not only in an increase or decrease but also in derepression of metabolite biosynthesis under normally repressing conditions. Moreover, double deletion of both the ffhda1 and ffhda2 genes showed additive but also distinct phenotypes with regard to secondary-metabolite biosynthesis, and both genes are required for gibberellic acid (GA)-induced bakanae disease on the preferred host plant rice, as Δffhda1 Δffhda2 mutants resemble the uninfected control plant. Microarray analysis with a Δffhda1 mutant that has lost the major HDAC revealed differential expression of secondary-metabolite gene clusters, which was subsequently verified by a combination of chemical and biological approaches. These results indicate that HDACs are involved not only in gene silencing but also in the activation of some genes. Chromatin immunoprecipitation with the Δffhda1 mutant revealed significant alterations in the acetylation state of secondary-metabolite gene clusters compared to the wild type, thereby providing insights into the regulatory mechanism at the chromatin level. Altogether, manipulation of HDAC-encoding genes constitutes a powerful tool to control secondary metabolism in filamentous fungi.

Fig 1

Inhibition of Zn(II)-dependent histone deacetylases (HDACs) results in decreased gibberellin (GA) biosynthesis in F. fujikuroi. (A) The wild type was grown in liquid synthetic medium under GA-inducing conditions (ICI medium with 6 mM glutamine) with 1 μM (red line) and without (black line) trichostatin A (TSA). After 7 days, the culture filtrate was extracted and analyzed as described in Materials and Methods. The quantified bioactive GA₃ and its precursors, GA₄ and GA₇, are labeled in the chromatogram. mAU, milli-absorbance units. (B) Quantification of gibberellin accumulation in the WT without TSA and in TSA-treated WT samples. The amount of gibberellins was related to the biomass in order to exclude falsification in the quantification of gibberellin biosynthesis due to altered growth behavior. The experiment was performed in triplicate; mean values and standard deviations (referring to overall gibberellin accumulation) are given. Asterisks above the bars denote significant differences in the measurement of the TSA-treated samples compared to the untreated WT. ∗∗, P < 0.01.

Fig 2

Contribution of single Zn(II)-dependent histone deacetylases (HDACs) in Fusarium fujikuroi to the overall HDAC activity. The wild type (WT) and the three single HDAC deletion mutants, Δffhda1, Δffhda2, and Δffhda4, were grown for 24 h in synthetic ICI medium under gibberellin-inducing conditions (6 mM glutamine). A crude nuclear extract was isolated as described in Materials and Methods and directly used for quantification of HDAC activity. A HeLa cell nuclear extract served as a positive control (+), and for the negative control (−), the HeLa cell nuclear extract was directly incubated with the HDACi trichostatin A (TSA) according to the manufacturer's instructions. HDAC activity is given in pmol min⁻¹ mg⁻¹. Experiments were done in quadruplicate. Mean values and standard deviations are given. Asterisks above the bars denote significant differences in the measurements of the indicated strains compared to the WT. ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Fig 3

Deletion and constitutive expression of ffhda1 have diverse effects on secondary-metabolite biosynthesis in Fusarium fujikuroi. (A) The wild type (WT) and the Δffhda1 mutant were grown for 7 days under biosynthesis-inducing conditions for the following five secondary metabolites: bikaverin (BIK) (6 mM glutamine), fusarubins (FSR) (6 mM sodium nitrate), gibberellic acids (GA) (6 mM glutamine), fusaric acid (FU) (120 mM sodium nitrate), and fusarins (FUS) (60 mM glutamine). After 7 days, a culture filtrate was taken, and samples for HPLC analyses were prepared as described in Materials and Methods. Experiments were performed in triplicate for each metabolite. Accumulation of metabolites was related to the biomass to exclude falsifications in the quantification due to an altered growth behavior of the mutant compared to the WT. Production of the WT was set to 100%. Mean values and standard deviations are given. Asterisks above the bars denote significant differences in the measurements of the indicated strains compared to the WT. ∗∗, P < 0.01; ∗∗∗, P < 0.001. (B) Detection of bikaverin under normally repressing conditions (60 mM glutamine) in the Δffhda1 mutant. (C) Northern blot analysis of the WT and the Δffhda1 mutant after 3 days of growth in ICI medium under conditions of nitrogen starvation (6 mM glutamine) and nitrogen sufficiency (60 mM glutamine). The biosynthetic gene bik2 (O-methyltransferase) was used for probing. (D) The WT and an overexpression mutant (gpd::ffhda1) were grown under biosynthesis-inducing conditions as described above for panel A. Experiments were performed in triplicate for each metabolite. Mean values and standard deviations are given. Production of the WT was set to 100%. Asterisks above the bars denote significant differences in the measurements of the indicated strains compared to the WT. ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Fig 4

Chromatin immunoprecipitation reveals distinct alterations in histone acetylation at the gibberellin (GA) and bikaverin (BIK) gene clusters in Fusarium fujikuroi. The wild type (WT) and the Δffhda1 mutant were grown in synthetic ICI medium with either 6 mM (nitrogen starvation) or 60 mM (nitrogen sufficiency) glutamine. After 3 days, the mycelium was used for ChIP as described in Materials and Methods. ChIP assays were performed by using an H3K9 acetylation-specific antibody. Subsequent next-generation sequencing was carried out on an Illumina GAII genome analyzer by single-end 36- to 50-nt sequencing, and quantitative real-time PCR was carried out as described in Materials and Methods. (A) Chromosomal location of the gibberellin and the bikaverin gene clusters. (B and C) ChIP-seq analysis of the GA gene cluster (B) and the bikaverin gene cluster (C) in the ffhda1 deletion mutant compared to the WT under both nitrogen conditions (6 mM glutamine and 60 mM glutamine). Experiments were done twice, each in duplicate. Red lines indicate significantly enriched reads. (D and E) Quantification of precipitated DNA at the gibberellin (D) and the bikaverin (E) gene clusters. The gibberellin biosynthetic gene cps/ks and the pathway-specific transcription factor bik5 were chosen for verification of the ChIP-seq data. In each case, the amount of precipitated DNA in the WT under low-nitrogen conditions (6 mM glutamine) was arbitrarily set to 1. Mean values and standard deviations are given.

Fig 5

FfHda1 and FfHda2 have common and distinct effects on secondary-metabolite biosynthesis in Fusarium fujikuroi. (A) The wild type (WT) and the Δffhda2 mutant were grown for 7 days under biosynthesis-inducing conditions for the following five secondary metabolites: bikaverin (BIK) (6 mM glutamine), fusarubins (FSR) (6 mM sodium nitrate), gibberellic acids (GA) (6 mM glutamine), fusaric acid (FU) (120 mM sodium nitrate), and fusarins (FUS) (60 mM glutamine). Experiments were performed in triplicate for each secondary metabolite. Accumulation of the respective metabolites was related to the biomass to exclude falsifications in the quantification due to an altered growth behavior of the mutant compared to the WT. The WT value was set to 100%. Mean values and standard deviations are given. Asterisks above the bars denote significant differences in the measurements of the indicated strains compared to the WT. ∗∗, P < 0.01; ∗∗∗, P < 0.001. (B and C) The WT and the Δffhda2 mutant were grown for 4 days (in the case of Northern blot analysis) and 7 days (for HPLC analysis) under fusarubin biosynthesis-repressing conditions (acidic pH) (ICI medium with 6 mM glutamine). Accumulation of fusarubins and bikaverin in the liquid culture is highlighted by red boxes. The biosynthetic genes fsr1 (PKS) and fsr2 (O-methyltransferase) were used for probing in the Northern blot. (D) The WT and the Δffhda1 Δffhda2 mutant were grown for 7 days under biosynthesis-inducing conditions and used for HPLC analysis as described above for panel A. Experiments were done in triplicate for each secondary metabolite. Standard deviations are given. Production of the WT was set to 100% for each secondary metabolite investigated. Asterisks above the bars denote significant differences in the measurements of the indicated strains compared to the WT. ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Fig 6

FfHda1 and FfHda2 are both required for gibberellin-induced bakanae disease on rice. Rice seedlings and the indicated fungal strains, the wild type (WT) and the Δffhda1, Δffhda2, Δffhda4, and Δffhda1 Δffhda2 mutants, were cocultivated for 7 days at 28°C with 80% humidity, as described in Materials and Methods. Noninfected rice seedlings and rice seedlings treated with 100 ppm of the bioactive gibberellic acid GA₃ served as negative and positive controls, respectively. For determination of bakanae symptoms on grown rice plants, the length between internodes was measured (highlighted by black arrows). (A) Pictures of rice plants after 7 days of cocultivation with the indicated fungal strains. (B) Measurement of internode elongation. The experiment was performed in triplicate. Mean values and standard deviations are given. Asterisks above the bars denote significant differences in the measurements of the indicated strains compared to the WT. ∗∗∗, P < 0.001.

Fig 7

Hypothetical model for decreased gibberellin (A) and bikaverin (B) biosynthesis in the Δffhda1 and gpd::ffhda1 mutants of Fusarium fujikuroi. (A) For sufficient gene activation in the wild type (WT), an as-yet-unknown transcription factor (TF) is deacetylated upon reception of an environmental stimulus by FfHda1 (1). Only the deacetylated transcription factor, e.g., the GATA transcription factor AreA, which is required for sufficient gibberellin gene expression, can then bind to the gibberellin (GA) gene cluster, resulting in the recruitment of the histone acetyltransferase (HAT) complex (2). The HAT is now able to acetylate histone proteins in the vicinity of the gibberellin gene cluster (3), resulting in accessibility of the genomic region for chromatin-remodeling complexes that further open the chromatin landscape, enabling the binding of further pathway-specific transcription factors and the transcription machinery, thus leading to gene activation. The concomitant presence of FfHda1 at the gibberellin gene cluster keeps the balance between histone acetylation and deacetylation (4). Upon deletion of ffhda1, the global transcription factor cannot be deacetylated anymore, resulting in hypoacetylation and the inability to bind to the cluster due to steric hindrance, thus resulting in gene silencing (1). When ffhda1 is overexpressed, the gibberellin-specific transcription factor is deacetylated (1), localizes to the gibberellin gene cluster, and recruits the HAT (2). While the HAT acetylates the histone in the vicinity of the gibberellin gene cluster, leading finally to gene activation (3), overrepresentation of FfHda1 deacetylates the histones (4), resulting putatively in hypoacetylation and thereby in downregulation of gene expression. (B) A similar mechanism of gene transcription happens at the bikaverin gene cluster. A bikaverin-specific transcription factor binds to the bikaverin gene cluster (1) and recruits a HAT (2), which then acetylates histones (3) in the vicinity of the bikaverin gene cluster, thereby allowing accessibility of the genomic DNA for chromatin-remodeling complexes that further open the genomic region. This enables binding of further transcription factors, e.g., the pathway-specific regulator BIK5, and subsequently the transcriptional machinery, finally resulting in gene activation. Hyperacetylation of the histones due to HAT activity is counteracted by HDAC activity (4). The global transcription factor is not deacetylated by FfHda1 prior to binding; however, whether either another HDAC is involved in this process or the transcription factor can bind in the acetylated state or is not subject to acetylation in the first place is not known (indicated by “?”). Therefore, acetylation of histone in the vicinity of bikaverin genes still takes place upon deletion of ffhda1 (1 to 3); however, histones are not deacetylated anymore to create a balanced state between HAT and HDAC activities, thereby leading to hyperacetylation (4). The lack of ffhda1 possibly also results in hyperacetylation of histones in the vicinity of the hypothetical bikaverin-specific repressor, resulting in its activation (5). The repressor then binds to the bikaverin gene cluster, thereby leading to gene silencing (6). Similar to the gibberellin gene cluster, overexpression of ffhda1 possibly results in hypoacetylation and thus downregulation of bikaverin gene expression due to overrepresentation of FfHda1.