Austropuccinia psidii uses tetrapolar mating and produces meiotic spores in older infections on Eucalyptus grandis.

Austropuccinia psidii is the causal agent of myrtle rust, a fungal disease that infects over 480 species in the Myrtaceae. A. psidii is a biotrophic pathogen that reproduces sexually and asexually. Sexual reproduction has been previously shown on Syzygium jambos and little is known about its reproductive biology on other hosts or whether populations that were formerly structured by host range can outcross on universally susceptible hosts. We investigated if mating genes in three genomes of A. psidii were under selection as a proxy for whether different strains can reproduce sexually on a shared host. We examined three homologs of the STE3.2 gene, sequences of which were near-identical in the three genomes, and the homeodomain locus, which contained two alleles of two homeodomain genes in each genome. A. psidii likely uses tetrapolar mating. Pheromone/receptor loci were distal to homeodomain loci, and based on haplotypes of a phased assembly, mate compatibility is regulated by multiallelic HD genes and biallelic STE3.2 genes; the third homolog of STE3.2 (STE3.2-1) was present in both haplotypes, and our study supports hypotheses this gene does not regulate mate recognition. Populations of A. psidii formerly structured by host range could potentially outcross on universal hosts based on their related mating genes, however this hypothesis should remain theoretical given the implications for biosecurity. Additionally, we searched for core meiotic genes in genomes of A. psidii, four species of Puccinia, and Sphaerophragmium acaciae through comparative genomics based on 136 meiosis-related orthologous genes modeled from Mycosarcoma maydis. Meiotic genes are conserved in rust fungi at family rank. We analyzed the expression of two meiotic and four mitotic genes of A. psidii on E. grandis over a 28-day time course to validate that identified meiotic genes were upregulated in teliospores. Three mitotic genes were significantly downregulated in samples collected 28 days after inoculation (DAI) compared to 14 DAI. Expression of meiotic genes was significantly up-regulated in samples collected 28 DAI compared to 14 DAI, indicating a temporal switch from production of uredinia (mitotic stage) to telia in the life cycle, which we hypothesize may be in response to leaf ageing.

Both R-loop removal and ribonucleotide excision repair activities of RNase H2 contribute substantially to chromosome stability.

Cells carrying deletions of genes encoding H-class ribonucleases display elevated rates of chromosome instability. The role of these enzymes is to remove RNA-DNA associations including persistent mRNA-DNA hybrids (R-loops) formed during transcription, and ribonucleotides incorporated into DNA during replication. RNases H1 and H2 can degrade the RNA component of R-loops, but only RNase H2 can initiate accurate ribonucleotide excision repair (RER). In order to examine the specific contributions of these activities to chromosome stability, we measured rates of loss-of-heterozygosity (LOH) in diploid Saccharomyces cerevisiae yeast strains carrying the rnh201-RED separation-of-function allele, encoding a version of RNase H2 that is RER-defective, but partly retains its other activity. The LOH rate in rnh201-RED was intermediate between RNH201 and rnh201Δ. In strains carrying a mutant version of DNA polymerase ε (pol2-M644G) that incorporates more ribonucleotides than normal, the LOH rate in rnh201-RED was as high as the rate measured in rnh201Δ. Topoisomerase 1 cleavage at sites of ribonucleotide incorporation has been recently shown to produce DNA double strand breaks. Accordingly, in both the POL2 and pol2-M644G backgrounds, the LOH elevation in rnh201-RED was suppressed by top1Δ. In contrast, in strains that incorporate fewer ribonucleotides (pol2-M644L) the LOH rate in rnh201-RED was low and independent of topoisomerase 1. These results suggest that both R-loop removal and RER contribute substantially to chromosome stability, and that their relative contributions may be variable across different regions of the genome. In this scenario, a prominent contribution of R-loop removal may be expected at highly transcribed regions, whereas RER may play a greater role at hotspots of ribonucleotide incorporation.