A non-canonical RNA degradation pathway suppresses RNAi-dependent epimutations in the human fungal pathogen Mucor circinelloides

Abstract

Mucorales are a group of basal fungi that includes the casual agents of the human emerging disease mucormycosis. Recent studies revealed that these pathogens activate an RNAi-based pathway to rapidly generate drug-resistant epimutant strains when exposed to stressful compounds such as the antifungal drug FK506. To elucidate the molecular mechanism of this epimutation pathway, we performed a genetic analysis in Mucor circinelloides that revealed an inhibitory role for the non-canonical RdRP-dependent Dicer-independent silencing pathway, which is an RNAi-based mechanism involved in mRNA degradation that was recently identified. Thus, mutations that specifically block the mRNA degradation pathway, such as those in the genes r3b2 and rdrp3, enhance the production of drug resistant epimutants, similar to the phenotype previously described for mutation of the gene rdrp1. Our genetic analysis also revealed two new specific components of the epimutation pathway related to the quelling induced protein (qip) and a Sad-3-like helicase (rnhA), as mutations in these genes prevented formation of drug-resistant epimutants. Remarkably, drug-resistant epimutant production was notably increased in M. circinelloides f. circinelloides isolates from humans or other animal hosts. The host-pathogen interaction could be a stressful environment in which the phenotypic plasticity provided by the epimutant pathway might provide an advantage for these strains. These results evoke a model whereby balanced regulation of two different RNAi pathways is determined by the activation of the RNAi-dependent epimutant pathway under stress conditions, or its repression when the regular maintenance of the mRNA degradation pathway operates under non-stress conditions.

Fig 1. fkbA sRNAs are present in the FK506 resistant epimutants isolates of the r3b2Δ mutant strains.

The numbers of the isolates correspond to those in S1 Table. With the exception noted below, the isolates analyzed here do not include those confirmed to harbor Mendelian mutations. (A) The sRNA enriched samples (35 μg) from all of the r3b2Δ FK506 resistant isolates lacking mutations in the target genes were obtained after incubation for 48 hours on MMC media supplemented with 1 μg/ml of FK506. sRNA blots were hybridized with an antisense-specific probe to detect fkbA sRNA (see Methods). 5S rRNA probe was used as a loading control. Abundant sRNAs were detected in all of the strains with the exception of isolate 7 of the MU429 strain that was confirmed to have a Mendelian mutation. (B) Signals from blots presented in panel (A) were quantified by densitometry and the results were plotted. The bars indicate the standard deviation.

Fig 2. Exogenously activated silencing frequency of the qip, rdrp3, rnhA, and sexM null mutants.

The color of the colonies was observed after subculture of the original transformants on YNB pH = 4.5 plates and incubation for 4 to 5 days in the light. Colonies with white patches were scored as white transformants.

Fig 3. Transgene hairpin silencing is defective in qip null mutant.

(A) Phenotype of the descendants of white colonies obtained from transformation of wild type (R7B, right) and the qipΔ mutant (left) with plasmid pMAT1253, which expresses a carB mRNA hairpin. The color of the colonies was analyzed after 48 hours growth on YNB media in the light. (B) sRNA enriched samples (50 μg) from 11 non-silenced transformants in the qipΔ background carrying pMAT1253 were obtained after 48 hour incubation in liquid YNB media pH = 4.5. A carB probe that specifically detects antisense RNA was used for hybridization in a northern blot assay. A silenced transformant with the same plasmid but in a WT background was used as a positive control. Ribo 3 (antisense, AS) and carB25 (sense, S) primers were used as size and sense markers (see Methods). tRNA were stained with ethidium bromide and served as a loading control.

Fig 4. M. circinelloides RnhA protein (RNA helicase).

(A) Structure of the putative M. circinelloides RnhA helicase. Three conserved domains can be found by blast with the NCBI database (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi): DEXDc domain, UvrD_C_2 domain, and a NF-X1-zinc-finger domain. (B) Conserved putative helicases are found in related fungal species close to the sex locus. Four different related fungal species are depicted: Rhizopus oryzae, Phycomyces blakesleeanus, Mucor circinelloides, and Mucor mucedo.

Fig 5. Detection of fkbA sRNAs in the FK506 resistant epimutant isolates of rdrp3Δ mutant strains.

(A) The numbers of the isolates correspond to those in S1 Table. The isolates analyzed here do not include those confirmed to harbor Mendelian mutations. The sRNA enriched samples (35 μg) from the rdrp3Δ FK506 resistant isolates lacking mutations in the target genes were obtained after 48 hour incubation at room temperature on MMC media pH = 4.5 supplemented with 1 μg/ml of FK506. sRNA blots were hybridized with an antisense-specific probe to detect fkbA sRNA (see Methods). 5S rRNA probe served as a loading control. (B) Signals from blots presented in panel (A) were quantified by densitometry and the results were plotted. The bars indicate the standard deviation.

Fig 6. RdRP3 and RnhA participate in the rdrp-dependent dicer-independent RNA degradation pathway.

(A) Accumulation of mRNAs in wild type and silencing mutants. Northern blots of high molecular weight RNAs corresponding to genes regulated by the rdrp-dependent RNA degradation pathway (genes P1 and P2) were carried out using total RNA (50 μg) extracted from wild type (R7B), dcl1Δ/2Δ (MU411), r3b2Δ (MU412), rdrp1Δ (MU419), rdrp2Δ (MU420), and rdrp3Δ (MU438) mutant strains grown for 24 hours in liquid MMC medium. Samples were separated in 1.2% denaturing agarose gels, transferred to membranes, and hybridized with gene specific probes (S2 Table). Target genes correspond to the following gene products: P1: ID 26072, pyruvate decarboxylase, P2: ID 92956, actin binding protein. The membranes were reprobed with a 28S rRNA probe as loading control. Images are representative of three independent experiments. (C) Similarly, total RNA extracted from qip (MU430) and rnhA (MU437) mutant strains was analyzed in Northern blots assays. (B) and (D) Densitometric analysis of expression data shown in (A) and (C). Signal intensities were quantified and normalized to rRNA levels. All data were again normalized with respect to the expression value of the wild type strain (R7B) for each gene.

Fig 7. Models for the different RNAi pathways operating in M. circinelloides.

The siRNA RNAi pathway (left) is a defense mechanism against invasive nucleic acids such as plasmids, transposons, and viruses [36]. Aberrant transcripts from these invasive agents are used by RdRP1 to generate dsRNA molecules, which are cleaved by Dcl2 to produce siRNAs that are transferred to Ago1. The RNase III-like protein R3B2 participates in the biogenesis of these siRNAs, although its function in this pathway is still unknown. RdRP2 generates secondary dsRNA and amplifies this pathway [18]. The epimutation RNAi pathway (middle) silences target endogenous transcripts under stress conditions to generate epimutant strains that are better adapted. In this pathway, the generation of both primary and secondary dsRNA might be under the control of RdRP2, because rdrp2 mutants are incapable of producing epimutant strains resistant to the antifungal drug FK506 [4]. The function of RnhA might be mediated via its catalytic activity in unwinding ds-esRNAs; however, its precise role remains to be established. The non-canonical RNAi pathway (right) targets highly expressed mRNAs for degradation. RdRP1, RdRP2 [5] and RdRP3 (this study) could interact with these highly expressed mRNAs to synthesize complementary strands that signal these mRNAs for degradation by R3B2, which produces the resulting rdRNAs. RnhA (this study) has also a positive role in this pathway, although its specific function is not yet known.