Histone modification dynamics and genome engineering in Magnaporthe oryzae

Abstract

Magnaporthe oryzae (synonym of Pyricularia oryzae) is a filamentous fungal plant pathogen, best known for causing rice (Oryza sativa) and wheat (Triticum aestivum) blast diseases that threaten world food security. The estimated annual loss of rice production due to rice blast is equivalent to the amount needed to feed 60 million people. M. oryzae is also a well-established model to study molecular pathogen-host interactions. In order to suppress plant host immunity, numerous secreted proteins and chemicals, termed effectors, are secreted into host cells to promote disease. Many fungal pathogen effectors identified to-date, including those of M. oryzae, are highly expressed in planta but remain silent during in vitro growth conditions, and are located in highly variable regions of the genome. Little is known about the molecular mechanisms of transcriptional regulation and biased genome distribution and evolution of fungal effectors. Previous reports indicate that trimethylation at H3-Lys27 (i.e., H3K27me3), a repressive histone modification, plays an important role in transcriptional regulation of secondary metabolite gene clusters. Like effectors, secondary metabolites are crucial for growth in changing environments and for specific niche adaptation, but they are dispensable for normal growth. Due to the transcriptional and genomic location similarities between effectors and secondary metabolites, we hypothesized that H3K27me3 plays an important role in silencing effector gene expression in vitro. Moreover, experimental evidence in animal systems suggested that histone modification dynamics involving other modifications, such as acetylation at the H3K27, could be part of a pathway for chromatin switching between inactive and activate states, similar to that observed during infection. Following chromatin immunoprecipitation and sequencing (ChIP-seq) and RNA-sequencing (RNA-seq), we found the enrichment of silent effectors in H3K27me3 domains. Genetic deletion of three core components of Polycomb Repressive Complex 2 (PRC2), resulted in a complete loss of H3K27me3, and derepressed ~10% of transcripts during in vitro growth, including ~44% of high-confidence effectors. Concomitant with the loss of H3K27me3, we found many regions gained the activation mark, H3K27ac, which appeared to play a role in effector activation, accounting for ~30% of increased transcription observed in the PRC2 mutant. ChIP-qPCR from infected materials indicated that some loci underwent H3K27 modification dynamics in planta, while other tested loci were not altered. Our result revealed an unexplored connection between histone modification dynamics and fungal transcriptional regulation, where H3K27me3 is important for effector silencing in vitro, while in planta activated genes may instead be H3K27ac and contributes to effector gene activation. Considering that roughly half of the presumed effectors showed no transcriptional changes in the PRC2 mutants, we additionally considered other genome functions that H3K27me3 may play. One possibility is that H3K27me3 contributes to genome instability, as has been reported in the plant pathogenic fungus Zymoseptoria tritici. Information connecting genome instability and histone modifications remains scarce. To further explore a potential mechanism, we addressed the contribution of DNA repair pathways to mutation outcomes and the potential for intragenomic variation. We first developed a genome editing tool based on CRISPR-Cas12a ribonucleoprotein (RNP) in M. oryzae. Surprisingly, we frequently observed Cas12a RNP edited mutants had abnormal genotyping outcomes. Using hundreds of edited mutants, significant variation in the mutational profile was observed, ranging from small INDELs to several kilobase deletions or insertions. Sequencing outcomes using a range of techniques suggested that multiple DNA double-strand break (DSB) repair pathways were active and responsible for the repair patterns. To exclude that the results were unique to the single tested locus, we expanded the editing to four additional loci. Interestingly, our results suggest that the frequency of the different DSB repair outcomes varies across the tested loci. Taken together, these results suggest that DSBs are repaired in a locus-dependent manner by different DSB repair pathways. Based on these observations, we speculate that choice of DSB repair pathway may contribute to genome instability and biased genome variation. Further efforts to dissect the connection between genomic and epigenomic features that impact DSB repair choices and influence fungal genome evolution will greatly benefit our understanding about fungal genome variation in natural populations.