RNA silencing in Aspergillus nidulans is independent of RNA-dependent RNA polymerases

Abstract

The versatility of RNA-dependent RNA polymerases (RDRPs) in eukaryotic gene silencing is perhaps best illustrated in the kingdom Fungi. Biochemical and genetic studies of Schizosaccharomyces pombe and Neurospora crassa show that these types of enzymes are involved in a number of fundamental gene-silencing processes, including heterochromatin regulation and RNA silencing in S. pombe and meiotic silencing and RNA silencing in N. crassa. Here we show that Aspergillus nidulans, another model fungus, does not require an RDRP for inverted repeat transgene (IRT)-induced RNA silencing. However, RDRP requirements may vary within the Aspergillus genus as genomic analysis indicates that A. nidulans, but not A. fumigatus or A. oryzae, has lost a QDE-1 ortholog, an RDRP associated with RNA silencing in N. crassa. We also provide evidence suggesting that 5' --> 3' transitive RNA silencing is not a significant aspect of A. nidulans IRT-RNA silencing. These results indicate a lack of conserved kingdom-wide requirements for RDRPs in fungal RNA silencing.

Figure 1.—

IRT and SST constructs of A. nidulans aflR. (Top) A diagram of the 1302-bp A. nidulans aflR coding sequence from the ATG start site to the TGA stop site is shown. A SacII site at position 889 is indicated. The region 3′ to the SacII site was used to make riboprobes for aflR mRNA and siRNA analysis (diagonal hatching). The aflR(IRT1300) transgene consists of two complete aflR coding regions in an inverted orientation, separated by an ∼280-bp spacer fragment (horizontal hatching). B, BamHI site. The aflR(IRT900) transgene consists of two identical truncated fragments of aflR coding region in an inverted orientation, separated by an ∼280-bp spacer fragment. This is essentially the same IRT as aflR(IRT1300), except that aflR sequences between the SacII sites have been removed. The aflR(SST1300) transgene contains the complete open reading frame of aflR with a mutated stop codon. All three aflR(IRT) and aflR(SST) transgenes are flanked by the A. nidulans gpdA promoter and trpC terminator (not shown).

Figure 2.—

A. nidulans aflR IRTs suppress NOR production and aflR expression. (A) A random selection of NOR+ and NOR− transformants were point inoculated onto oatmeal medium and cultured for ∼5 days to assay NOR production; aflR(IRT1300) transformants are numbered 1–6 and aflR(IRT900) transformants are numbered 7 and 8. (B) Southern analysis of HindIII-digested genomic DNA from the eight transformants depicted in A and the transformation host strain are shown (lanes 1–H). Lane M shows nonspecific hybridization of the probe to the DNA ladder (top band, 10 kb; bottom band, 6 kb). (C and D) A trpC control transformant (TTMH20.9) and an aflR(IRT900) transformant (TTMH20.8) were cultured in liquid minimal media and analyzed for (C) NOR production and (D) aflR expression over 96 and 72 hr, respectively. The riboprobe was specific for aflR sequences not present in the aflR(IRT900) transgene to detect aflR transcripts derived from the endogenous aflR locus only. N, NOR standard.

Figure 3.—

A. nidulans IRT-RNA silencing is characterized by a single class of ∼25-nt siRNAs. The following strains were cultured for 72 hr in liquid minimal medium: TTMH16.9, aflR(SST1300); TTMH13.1, aflR(IRT1300); and TTMH20.8, aflR(IRT900); designated SST, IRT, and IR, respectively. A riboprobe specific for (A) aflR sense sequences or (B) antisense aflR sequences (see Figure 1 for aflR region specificity) was hybridized to ∼30 μg of low-molecular-weight (MW) RNAs. Approximately 20 pmol of sense (S) and antisense (AS) aflR oligonucleotides, 23 and 25 nt, respectively (Table 2), was used as a control for the probe. It was also mixed with ∼30 μg of low-MW RNAs from TTMH16.9 as a migration control to more accurately determine the size of A. nidulans siRNAs. Lanes are labeled with the letter of the oligonucleotide used in the particular lane or with the appropriate lane number if an oligonucleotide was not added to the lane. Ethidium bromide staining of the highest-concentration RNA species is shown to demonstrate relative amounts of RNA between lanes.

Figure 4.—

A. nidulans rsdA, rrpB, and rrpC. The predicted amino acid sequences of RsdA, RrpB, and RrpC were used to search NCBI's conserved domain database. (A–C) The identified conserved domains are indicated along with the predicted starting and stopping points for each domain, the predicted length of the protein, and the codons deleted in these studies. (A) Predicted PAZ and PIWI domains of A. nidulans RsdA. (B and C) Predicted RDRP domains of RrpB and RrpC, respectively.

Figure 5.—

Genetic analysis of IRT-RNA silencing in A. nidulans. A. nidulans mutants with (top) or without (bottom) the aflR(IRT1300) transgene were point inoculated onto 25 ml of solid minimal medium and incubated for 6 days. NOR production was analyzed by TLC. Bright orange spots are NOR. NA, not analyzed. Top row, strains are: WT, RTMH13.B1; ΔrsdA, RTMH65.1; ΔmusN, RTMH13.D9; and ΔrrpB ΔrrpC, RTMH7475.1A. Bottom row, strains are: WT, RTMH13.B3; ΔmusN, RTMH13.D16; and ΔrrpB ΔrrpC, RTMH7475.3A.

Figure 6.—

Three classes of RDRPs exist in N. crassa, M. grisea, and Aspergillus species. The neighbor-joining method was used with the predicted amino acid sequences of all known RDRPs and RDRPs identified in this study from N. crassa, S. pombe, M. grisea, and the three sequenced Aspergilli (Table 4) to create this noniterated, unrooted tree. A. nidulans is missing a QDE-1-like RDRP and A. fumigatus is missing a RRP-3-like RDRP. N.c., N. crassa; A.n., A. nidulans; A.o., A. oryzae; A.f., A. fumigatus; M.g., M. grisea; S.p., S. pombe.

Figure 7.—

Genome analysis indicates that A. nidulans rrpA has been lost during evolution. (A) Synteny exists between the A. fumigatus rrpA region and an analogous region in A. nidulans. Generic gene-type predictions were obtained from the draft annotation of the A. nidulans and A. fumigatus genomes and are listed in Table 3. Expect values were obtained by searching (blastp) the A. fumigatus genome database with the predicted amino acid sequences of the corresponding A. nidulans ORF depicted in this diagram. The nucleotide length designations above the A. fumigatus rrpA locus and the analogous region in A. nidulans indicate the approximate numbers of bases between the stop and start sites of the flanking genes. (B) The ∼4.0-kb region of A. nidulans genomic DNA depicted in A was compared to genomic sequences of all known and predicted fungal RDRPs from N. crassa, A. fumigatus, A. oryzae, A. nidulans, and M. grisea and 10 random genes from A. nidulans. Genomic sequences (including introns) between the known or predicted start and stop sites of the genes listed in Table 4 were compared using the neighbor-joining method and are depicted in a noniterated, unrooted tree. Numbers on the branches correspond to the genes listed in Table 4.