RNA Silencing in Aspergillus Nidulans Is Independent of RNA-Dependent RNA Polymerases

RNA Silencing in Aspergillus nidulans Is Independent of RNA-Dependent RNA Polymerases

T. M. Hammond and N. P. Keller1 Department of Plant Pathology, University of Wisconsin, Madison, Wisconsin 53706 Manuscript received September 4, 2004 Accepted for publication November 5, 2004

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 5Ј → 3Ј 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.

NA silencing refers to a group of very similar post- ases (RDRPs) are essential components of RNA silenc- R transcriptional gene-silencing mechanisms that ing (e.g., protists, nematodes; Smardon et al. 2000; Sijen have been discovered in a diverse range of eukaryotes et al. 2001; Martens et al. 2002; Simmer et al. 2002), (Pickford et al. 2002; Denli and Hannon 2003; Tang while in others RDRPs appear to be dispensable for this et al. 2003). The core processes of RNA silencing are process (e.g., flies, mammals; Schwarz et al. 2002; Stein highly conserved and involve double-stranded RNA et al. 2003). In plants and fungi, the roles of RDRPs in (dsRNA) processing by an RNAse III domain-containing RNA silencing are not as well defined. For example, the enzyme (Dicer) into 21- to 26-nt small interfering RNAs model plant A. thaliana encodes six putative RDRPs and (siRNAs; Bernstein et al. 2001), which are then incorpo- thus far only two have been partially investigated. Of rated into a ribonucleoprotein complex (RNA-induced these two RDRPs, SGS2/SDE1 is required for RNA si- silencing complex, RISC). RISC recognizes and de- lencing activated by sense transgenes (Beclin et al. grades target mRNAs by complementary base pairing 2002), but not for RNA silencing activated by inverted to the incorporated siRNA (Hammond et al. 2000; repeat transgenes (IRTs) or RNA viruses (Dalmay et al. Elbashir et al. 2001). An essential protein member of 2000; Beclin et al. 2002; Muangsan et al. 2004), and RISC is an argonaute family protein with a PAZ and AtRdRP1 is involved in viral defense (Yu et al. 2003; PIWI domain (PPD; Carmell et al. 2002). Examples Yang et al. 2004). include Rde-1 in Caenorhabditis elegans (Tabara et al. Studies of fungal RDRPs suggest that these enzymes 1999), dAgo2 in Drosophila melanogaster (Hammond et are involved in RNA silencing and a number of other al. 2001), Ago1 in Arabidopsis thaliana (Fagard et al. gene-silencing-related processes in fungi. For example, 2000), Ago1 in Schizosaccharomyces pombe (Volpe et al. the S. pombe RDRP, Rdp1, is required for RNA silencing 2002), and QDE-2 in Neurospora crassa (Catalanotto induced by IRTs (IRT-RNA silencing) and for RNAi- et al. 2002). Recent evidence suggests that the PAZ do- dependent heterochromatin formation at centromeric main of argonaute proteins facilitates transfer of siRNAs regions, mating-type loci, and euchromatic regions to the RISC complex (Lingel et al. 2003; Yan et al. 2003) (Volpe et al. 2002, 2003; Schramke and Allshire 2003; and that the PIWI domain contains the nuclease activity Jia et al. 2004; Verdel et al. 2004). While it is currently responsible for siRNA-guided mRNA cleavage (Song et unknown why the process of IRT-RNA silencing requires al. 2004). an RDRP in S. pombe, current models suggest that RNAi- In some organisms, RNA-dependent RNA polymer- dependent heterochromatin formation requires Rdp1 to create, directly or indirectly, small RNAs used to direct a complex of proteins, referred to as RNA-induced 1Corresponding author: Department of Plant Pathology, University of Wisconsin, 1630 Linden Dr., Madison, WI 53706. initiation of transcriptional gene silencing (RITS) pro- E-mail: [email protected] teins, to specific chromatin regions (Volpe et al. 2002,

Genetics 169: 607–617 (February 2005) 608 T. M. Hammond and N. P. Keller

2003; Schramke and Allshire 2003; Verdel et al. upon IRT-RNA silencing while deletion of a putative 2004). PPD protein, named RsdA, disrupted this process. Possi- In the filamentous fungus N. crassa, there are two ble reasons to account for the apparent difference in a gene silencing processes that require two of three N. RDRP requirement for IRT-RNA silencing in S. pombe crassa RDRPs (Galagan et al. 2003). The first is N. crassa and A. nidulans are discussed. quelling, a type of RNA silencing that is thought to be related to high transgene number (Pickford et al. 2002; Forrest et al. 2004). This process requires the RDRP MATERIALS AND METHODS QDE-1 (Cogoni and Macino 1999a). In vitro studies of Strains, growth conditions, and transformation conditions: QDE-1 activity indicate that it produces both full-length All strains used in this study are listed in Table 1. A. nidulans complementary RNA (cRNA) and 9- to 21-nt cRNAs RJH0128 was transformed with aflR-specific IRTs and control along the length of single-stranded RNA templates transgenes using the method described by Yu and Adams (Makeyev and Bamford 2002), suggesting the possibil- (1999). Standard crossing techniques (Pontecorvo 1953) ity that QDE-1 creates dsRNA for processing by Dicer were then used to introduce different auxotrophic markers from A. nidulans RDIT1.1 into the aflR(IRT) and aflR single- or directly forms siRNAs for incorporation into RISC sense transgene (SST) transformants. Gene replacements during quelling (Makeyev and Bamford 2002). Such were confirmed by Southern blotting with probes specific for activities may be unnecessary when RNA silencing is the internal deleted region and at least one flanking region activated by IRTs, which may explain the recent finding of the deleted gene. A. nidulans RTMH13.C5 (for rsdA) and that QDE-1 is dispensable for IRT-RNA silencing (Cata- A. nidulans RTMH13.F5 (for rrpB and rrpC) were used as the hosts for gene replacements. A. nidulans cultures were grown lanotto et al. 2004). The second N. crassa gene-silenc- in 25 ml of appropriate supplemented liquid or solid minimal ing process requiring an RDRP is meiotic silencing by media or oatmeal agar as previously described (Butchko et unpaired DNA (MSUD; Shiu et al. 2001; Shiu and Met- al. 1999). All strains were cultured under dark, stationary Њ zenberg 2002). This process requires the RDRP SAD-1 conditions at 37 in standard Petri dishes. Liquid cultures were ϫ 6 ف (Shiu et al. 2001; Lee et al. 2003). A third N. crassa inoculated with 1 10 spores/ml and solid cultures were point inoculated with freshly harvested conidia. RDRP, RRP-3, has not yet been attributed with a func- Vector construction: Oligonucleotides used in vector con- tion. Phylogenetic analysis suggests RRP-3 is not part of struction are listed in Table 2. the quelling or MSUD pathways (Galagan et al. 2003; pTMH13.7 [also referred to as aflR(IRT1300)]: This transfor- bp-1300ف mation vector consists of an inverted repeat of two ف Borkovich et al. 2004) and biochemical studies suggests that it is not involved in DNA methylation or heterochro- aflR fragments (Figure 1) separated by an 280-bp spacer, driven by the A. nidulans gpdA promoter and terminated by matin formation (Freitag et al. 2004b). the A. nidulans trpC terminator (Punt et al. 1991). It also The zygomycete Mucor circinelloides may encode an contains a truncated A. nidulans trpC selectable marker for RDRP with an important role in transitive RNA silencing targeted integration of the IRT next to the A. nidulans trpC locus (Mullaney et al. 1985). Oligonucleotides aflh5-mod ف Nicolas et al. 2003). This process, more thoroughly) investigated in plants (Vaistij et al. 2002; Van Houdt and aflr3-BamHI were used to amplify an 1300-bp aflR fragment (5Ј HindIII-aflR), containing the full-length A. nidulans et al. 2003) and nematodes (Sijen et al. 2001), forms aflR coding sequence. This PCR product was cloned into the dsRNA/siRNAs from sequences upstream (3Ј → 5Ј) EcoRV site of pBluescript II SKϪ (pBS, Stratagene, La Jolla, and/or downstream (5Ј → 3Ј) of primary target se- CA) to create pTMH2.3. Oligonucleotides afln5-NcoI and aflr3-BamHI were then used to amplify a similar aflR fragment quences on targeted mRNA, leading to the creation of Ј Ј secondary siRNAs and the spreading of RNA silencing with a different restriction site in the 5 primer (5 NcoI-aflR). This PCR product was cloned into the SmaI site of pBS to create bp-1300ف Denli and Hannon 2003). In M. circinelloides these pTMH3.3. Using multiple cloning steps, these two) secondary siRNAs have been detected, but a specific aflR fragments were then placed in an inverted orientation bp spacer (gf1), creating plasmid-280ف RDRP has yet to be identified (Nicolas et al. 2003). on opposite sides of a Recently, a clear dissimilarity in fungal RDRP function pTMH8.7. The gf1 spacer was amplified from pPRgf-T4 (Zolo- became apparent when examination of a N. crassa strain tukhin et al. 1996) with oligonucleotides 5gfp and 3gfp- BamHI. The aflR IRT was then released from pTMH8.7 and devoid of all its RDRPs showed that, unlike S. pombe cloned between the NcoI and HindIII sites of the high-expres- Rdp1 mutants, it was not affected in DNA methylation sion vector pAN52-3 (gi:474929), creating pTMH10.1. Finally, bp 5Ј EcoRI fragment of A. nidulans trpC was cloned-2400ف or heterochromatin silencing (Freitag et al. 2004b). a Here, in addition to reporting that IRTs efficiently si- into the EcoRI site of pTMH10.1 to give the aflR(IRT1300) lence homologous mRNAs in the model filamentous transgene (Figure 1). pTMH20.7 [also referred to as aflR(IRT900)]: This transforma- fungus Aspergillus nidulans, we report that dissimilarity tion vector differs from pTMH13.7 only in that it consists of bp aflR-1300ف bp aflR fragments instead of two-900ف in fungal RDRP function is also observed in the process two of IRT-RNA silencing. Comparative analysis of all pre- fragments. Oligonucleotides 5gfp-SacII and 3gfp were used to bp fragment of GFP from pPRgf-T4. This-280ف dicted RDRPs in the three sequenced Aspergilli revealed amplify an that A. nidulans encodes two RDRPs and, in contrast to fragment (gf2) was then cloned into the SmaI site of pBS, giving a plasmid containing a SacII-gf2-SacII fragment (pTMH14.8). the related species A. fumigatus and A. oryzae, has lost Plasmid pTMH10.1 was then digested with SacII, thereby re- bp from the center of the aflR IRT, including 1100ف an ortholog of N. crassa QDE-1. Deletion of the re- leasing bp from the 3Ј end of each aflR fragment. This fragment 400ف maining two A. nidulans RDRPs had no detectable effect RNA Silencing in Aspergillus 609

TABLE 1 A. nidulans strains used in this study

Fungal straina Genotype Source RJH0128 biA1; trpC801; ⌬stcE::argB; wA2 J. K. Hicks and N. P. Keller (unpublished data) TTMH13.1 biA1; aflR(IRT1300)::trpC; ⌬stcE::argB; wA2 This study TTMH13.2 biA1; ⌬stcE::argB; wA2 This study TTMH20.8 biA1; aflR(IRT900)::trpC; ⌬stcE::argB; wA2 This study TTMH20.9 biA1; ⌬stcE::argB; wA2 This study TTMH16.9 biA1; aflR(SST1300)::trpC; ⌬stcE::argB; wA2 This study RDIT1.1 pyrG89; argB2; metG1 Tsitsigiannis et al. (2004) RTMH13.B1 aflR(IRT1300)::trpC; ⌬stcE::argB This study RTMH13.B3 ⌬stcE::argB This study RTMH13.C5 aflR(IRT1300)::trpC; ⌬stcE::argB; pyrG89; wA This study RTMH13.F5 aflR(IRT1300)::trpC; ⌬stcE::argB; pyrG89; metG1 This study TTMH65.1 aflR(IRT1300)::trpC; ⌬stcE::argB; ⌬rsdA::pyrG; pyrG89; wA This study TTMH74.12 aflR(IRT1300)::trpC; ⌬stcE::argB; ⌬rrpB::pyrG; pyrG89; metG1 This study TTMH75.17 aflR(IRT1300)::trpC; ⌬stcE::argB; pyrG89; ⌬rrpC::metG; metG1 This study RTMH7475.1A aflR(IRT1300)::trpC; ⌬stcE::argB; ⌬rrpB::pyrG; ⌬rrpC::metG This study RTMH7475.3A ⌬stcE::argB; ⌬rrpB::pyrG; ⌬rrpC::metG This study AAH16 ⌬musN::pyrG; pabaA1; yA2 Hofmann and Harris (2001) RTMH13.D9 aflR(IRT1300)::trpC; ⌬stcE::argB; ⌬musN::pyrG This study RTMH13.D16 ⌬stcE::argB; ⌬musN::pyrG This study a Strains starting with T are original transformants. Strains starting with R are recombinants. All strains contain the veA1 allele (Mooney et al. 1990). was replaced with the SacII-gf2-SacII fragment from pTMH14.8 orientation between the gpdA promoter and trpC terminator to give pTMH17.2. As described above, a trpC selectable of pAN52-3; thus it is referred to as a SST to distinguish it from marker was cloned into the EcoRI site of pTMH17.2 to give the inverted repeat nature of the IRTs. Plasmid pTMH10.1 was bp-1600ف the aflR(IRT900) transgene (Figure 1). digested with BamHI and HindIII to release an pTMH16.3 [also referred to as aflR(SST1300)]: This transfor- BamHI-gf1-Rfla-HindIII fragment. The BamHI and HindIII mation vector contains a single full-length aflR ORF, in a sense ends of the digested plasmid were filled in with Klenow DNA

TABLE 2 Oligonucleotides used in this study

Oligonucleotide Sequence Purpose aflh5-Mod 5Ј-CATATGCAAGCTTCATGGAGC-3Ј aflR(IRT) aflr3-BamHI 5Ј-AAGGATCCGAGGCGTGGCGG-3Ј aflR(IRT/SST) afln5-NcoI5Ј-TTTCCATGGAGCCCCAGCGA-3Ј aflR(IRT/SST) 5gfp 5Ј-ATGACCTAGGGCTATGTGCAGGAG-3Ј aflR(IRT) 3gfp-BamHI 5Ј-GCACGGATCCATCCTCAATGTTG-3Ј aflR(IRT) 5gfp-SacII 5Ј-CATGACCGCGGACTATGTGCAG-3Ј aflR(IRT) 3gfp 5Ј-GCACGGCGCCATCCTCAATG-3Ј aflR(IRT) AnQDE2-5fl5 5Ј-TTGTGGTACCGAGTCTGCCGACCGTGAAATC-3Ј rsdA replacement vector AnQDE2-5fl3 5Ј-CTGATCCAAGCTTCCACCACCACCG-3Ј rsdA replacement vector AnQDE2-3fl5 5Ј-CTGAGCGAATTCACTCATTTCATTAGCTC-3Ј rsdA replacement vector AnQDE2-3fl3 5Ј-CCTTTGGATCCCCTAACTCGGCTTCCCTGG-3Ј rsdA replacement vector AnRDRP1-5f5 5Ј-CGTCAGGTACCTCCAAGTGTTTATACTGCG-3Ј rrpB replacement vector AnRDRP1-5f3 5Ј-TCAAAAAGCTTCCTTGCCTGCACCTC-3Ј rrpB replacement vector AnRDRP1-3f5 5Ј-AAACGAATTCCGGCACAGCACGAGATC-3Ј rrpB replacement vector AnRDRP1-3f3 5Ј-TTTGGATCCAGTTCCTTGCGAAAACCCATATCCC-3Ј rrpB replacement vector AnRDRP2-5f5 5Ј-CAAAGCAACTCGAGCCACCGCAAACC-3Ј rrpC replacement vector AnRDRP2-5f3 5Ј-TAAGTAAAGCTTCTCCCCGTTCTCATCGCAG-3Ј rrpC replacement vector AnRDRP2-3f5 5Ј-CCCCGGGATCCCTAGCCAAGATC-3Ј rrpC replacement vector AnRDRP2-3f3 5Ј-GTCTTCTAGATACACCCACCGCCTAGTCTG-3Ј rrpC replacement vector aflR3END5 5Ј-GTCGCCATGGCTCGGGATAGGTC-3Ј Sense (S) oligo (Figure 3A) aflR3END3B 5Ј-GGCGACGGCTTACCTGAGTCACCAG-3Ј Antisense (AS) oligo (Figure 3B) 610 T. M. Hammond and N. P. Keller polymerase (New England Biolabs, Beverly, MA) and ligated together to give pTMH12.1. As described above, a trpC selectable marker was cloned into the EcoRI site of pTMH12.1 to complete the aflR(SST1300) transgene (Figure 1). Gene replacement vectors: DNA flanking regions were cloned from A. nidulans genomic DNA for creation of rsdA, rrpB, and rrpC gene replacement vectors, using the oligonucleotides listed in Table 2. Oligonucleotide restriction sites were used to place the flanking regions into the matching sites in pBS. A. parasiticus pyrG (Skory et al. 1990; for rsdA and rrpB)orA. nidulans metG (Sienko and Paszewski 1999; for rrpC) selectable markers were placed between the flanking DNA and the resulting plasmids were used in A. nidulans transforma- tions. Northern hybridizations: Total RNA analysis: Trizol reagent Figure 1.—IRT and SST constructs of A. nidulans aflR. (Invitrogen, Carlsbad, CA) was used to isolate total RNA from (Top) A diagram of the 1302-bp A. nidulans aflR coding se- lyophilized A. nidulans cultures per manufacturer’s instruc- quence from the ATG start site to the TGA stop site is shown. tions. RNA was then blotted to Hybond-XL (Amersham, San A SacII site at position 889 is indicated. The region 3Ј to the Francisco). Ambion’s (Austin, TX) Maxiscript kit was used to SacII site was used to make riboprobes for aflR mRNA and make a 3Ј aflR sense-specific ␣-32P-labeled riboprobe (Figure siRNA analysis (diagonal hatching). The aflR(IRT1300) trans- 1, diagonal hatching) for hybridization to total RNA. gene consists of two complete aflR coding regions in an in- bp spacer fragment-280ف Low-molecular-weight (MW) RNA analysis: Low-MW RNA was verted orientation, separated by an isolated as described (Catalanotto et al. 2002). This was (horizontal hatching). B, BamHI site. The aflR(IRT900) trans- separated in a denaturing gel, blotted to a Hybond-XL nylon gene consists of two identical truncated fragments of aflR membrane, and hybridized to a riboprobe as described (Nico- coding region in an inverted orientation, separated by an bp spacer fragment. This is essentially the same IRT as-280ف las et al. 2003), except that Denhardt’s reagent was excluded from the hybridization buffer. The 3Ј sense and antisense aflR- aflR(IRT1300), except that aflR sequences between the SacII specific riboprobes (Figure 1) were prepared as described sites have been removed. The aflR(SST1300) transgene con- above, except that before hybridization the probe was hy- tains the complete open reading frame of aflR with a mutated nt fragments as described (Hamilton and stop codon. All three aflR(IRT) and aflR(SST) transgenes are-50ف drolyzed into Baulcombe 1999). flanked by the A. nidulans gpdA promoter and trpC terminator Norsolorinic acid analysis: From liquid cultures: Twenty-five- (not shown). milliliter A. nidulans cultures were mixed with an equal volume of acetone. The mixture was slightly agitated for 1 hr and then 7.5 ml was transferred to a new container, which was RESULTS then shaken vigorously with an equal volume of CHCl3. The mixture was allowed to separate and a 5-ml aliquot of the IRTs correlate with inhibition of gene expression in CHCl3 layer was transferred to a new tube, evaporated, and ␮ A. nidulans: To test if RNA silencing exists in A. nidulans, redissolved in 200 lofCHCl3. Five-microliter aliquots from two different IRTs were designed with A. nidulans aflR each sample were then loaded on a TLC plate (no. 4410221; Whatman, Brentford, UK). Compounds were separated using sequences (Figure 1, top). AflR is a transcription factor a toluene:ethyl-acetate:acetic acid (80:10:10) solvent system. required for the biosynthesis of a bright orange com- From solid media cultures: A single 1.4-cm-diameter core was pound, norsolorinic acid (NOR), in A. nidulans ⌬stcE removed from the center of a 6-day-old colony, ground in 3 strains (Butchko et al. 1999). Thus loss of aflR expres- ml of 50% acetone, and then mixed with 1.5 ml of CHCl3.A sion leads to loss of NOR production. Transformation 1-ml aliquot of the CHCl3 layer was then transferred to a new tube and evaporated. The residual compounds were redis- of A. nidulans RJH0128 with aflR(IRT1300) or aflR(IRT- solved in 20 ul of CHCl3 and analyzed by TLC as described 900) (Figure 1) resulted in a few transformants that above. did not produce NOR (Figure 2A). Southern analysis Computer-based analyses and database searches: Table 4 indicated that the NORϪ phenotype correlated 100% lists accession numbers for N. crassa, S. pombe, Magnaporthe with the successful integration of an aflR(IRT) into the grisea, and Aspergillus genes used in this study. Putative Asper- genome and that in each case the endogenous aflR gillus RDRPs were identified by a search (blastp and tblastn) Ϫ of the A. nidulans, A. fumigatus, and A. oryzae genome databases locus was not affected (Figure 2, A and B). The NOR with RDRP sequences from N. crassa and the conserved do- phenotype was stable in time course experiments (Fig- main for eukaryotic RDRPs (pfam no. 05183.5). Putative A. ure 2C) and through sexual crosses (data not shown). nidulans PPD proteins and RecQ DNA helicases were identi- In contrast, all 23 transformants resulting from transfor- fied by searching the same databases (blastp) with N. crassa mation of A. nidulans RJH0128 with aflR(SST1300) (Fig- QDE-2 and QDE-3 sequences. Specific database websites are ϩ listed in the acknowledgments section. Similarity comparisons ure 1) were NOR (data not shown). (Clustal W), phylogenetic trees (DrawTree), and ORF analysis To determine if the NORϪ phenotype was a result (SixFrame) were performed on the SDSC Biology Workbench of decreased aflR transcript levels, total RNA extracts (http://workbench.sdsc.edu). A search for a conserved RDRP were analyzed from two strains resulting from the trans- motif (DbDGD) in the putative degenerated A. nidulans rrpA formation of RJH0128 with the aflR(IRT900) transgene. locus was performed by pasting amino acid translations of this ϩ locus in all six reading frames into Microsoft Word and visually A. nidulans TTMH20.9 is a NOR transformant that scanning the document with the aid of the program for the integrated the trpC-selectable marker without the aflR conserved motif. (IRT900) inverted repeat, while A. nidulans TTMH20.8 RNA Silencing in Aspergillus 611

Figure 3.—A. nidulans IRT-RNA silencing is characterized nt siRNAs. The following strains were-25ف by a single class of 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) ␮ ف Figure 2.—A. nidulans aflR IRTs suppress NOR production ϩ was hybridized to 30 g of low-molecular-weight (MW) and aflR expression. (A) A random selection of NOR and RNAs. Approximately 20 pmol of sense (S) and antisense (AS) NORϪ transformants were point inoculated onto oatmeal ,(aflR oligonucleotides, 23 and 25 nt, respectively (Table 2 ف medium and cultured for 5 days to assay NOR production; was used as a control for the probe. It was also mixed with ␮g of low-MW RNAs from TTMH16.9 as a migration 30ف aflR(IRT1300) transformants are numbered 1–6 and aflR (IRT900) transformants are numbered 7 and 8. (B) Southern control to more accurately determine the size of A. nidulans analysis of HindIII-digested genomic DNA from the eight siRNAs. Lanes are labeled with the letter of the oligonucleo- transformants depicted in A and the transformation host strain tide used in the particular lane or with the appropriate lane are shown (lanes 1–H). Lane M shows nonspecific hybridiza- number if an oligonucleotide was not added to the lane. Ethid- tion of the probe to the DNA ladder (top band, 10 kb; bottom ium bromide staining of the highest-concentration RNA spe- band, 6 kb). (C and D) A trpC control transformant (TTMH20.9) cies is shown to demonstrate relative amounts of RNA between and an aflR(IRT900) transformant (TTMH20.8) were cul- lanes. tured 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 do not form significant levels of dsRNA or secondary si- derived from the endogenous aflR locus only. N, NOR stan- RNAs from the 3Ј region of targeted aflR mRNA during dard. IRT-RNA silencing. RNA silencing in A. nidulans requires RsdA, a putative is a NORϪ transformant that integrated both the trpC- argonaute family protein: Two putative PPD proteins Ϫ139 Ϫ31 selectable marker and the aflR(IRT900) inverted repeat. (e and e ) were identified in the A. nidulans genome Analysis of aflR mRNA levels with a riboprobe specific database by searching (blastp) with the predicted N. -bp of aflR (Figure 1, diagonal hatching), crassa QDE-2 amino acid sequence. Deleting the highest 400ف for the 3Ј Ϫ139 a segment of aflR not contained in the aflR(IRT900) matching gene (e ) resulted in a loss of aflR silencing transgene, indicated that aflR transcript levels were sig- (Figure 4 and supplementary Figure S1 at http://www. nificantly decreased in TTMH20.8 relative to TTMH- genetics.org/supplemental/) in the aflR(IRT1300) ge- 20.9 (Figure 2D). netic background (Figure 5). This gene was therefore RNA silencing in A. nidulans is characterized by a named rsdA, for RNA-silencing-deficient A, and it is likely single class of 25 nt siRNAs: To confirm that RNA silenc- a RISC complex protein required for RNA silencing. Ϫ ing is the mechanism responsible for the IRT-induced The second match (e 31) may be an ortholog of N. crassa decrease in aflR mRNA, low-molecular-weight RNAs SMS-2, a QDE-2 paralog required for meiotic silencing were analyzed for the presence of aflR-specific siRNAs. (Lee et al. 2003). This putative protein (referred to as Using sense- and anti-sense-specific riboprobes for the smsA in Table 4) was also identified during a search of 3Ј end of aflR (Figure 1), siRNAs of 25 nt in length the A. nidulans genome with the predicted N. crassa Ϫ were detected in a NORϪ strain containing the aflR SMS-2 sequence (e 15). SmsA was not investigated further (IRT1300) transgene (TTMH13.1) but not in a NORϩ during this study. strain containing the aflR(SST1300) transgene (TTMH- RNA silencing in A. nidulans does not require A. nidu- 16.9), or in a NORϪ strain containing the aflR(IRT900) lans MusN: N. crassa quelling requires QDE-3, a RecQ transgene (TTMH20.8; Figure 3). The absence of si- DNA helicase (Cogoni and Macino 1999b). Therefore RNAs in TTMH20.8 indicates that A. nidulans RDRPs we searched for an A. nidulans QDE-3 ortholog in the 612 T. M. Hammond and N. P. Keller

Figure 4.—A. nidulans rsdA, rrpB, and rrpC. The predicted amino acid sequences of RsdA, RrpB, and RrpC were used to Figure 6.—Three classes of RDRPs exist in N. crassa, M. search NCBI’s conserved domain database. (A–C) The identi- grisea, and Aspergillus species. The neighbor-joining method ﬁed conserved domains are indicated along with the predicted was used with the predicted amino acid sequences of all known starting and stopping points for each domain, the predicted RDRPs and RDRPs identiﬁed in this study from N. crassa, S. length of the protein, and the codons deleted in these studies. pombe, M. grisea, and the three sequenced Aspergilli (Table (A) Predicted PAZ and PIWI domains of A. nidulans RsdA. 4) to create this noniterated, unrooted tree. A. nidulans is (B and C) Predicted RDRP domains of RrpB and RrpC, respec- missing a QDE-1-like RDRP and A. fumigatus is missing a RRP- tively. 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.

A. nidulans genome database using the predicted QDE-3 Loss of a QDE-1 ortholog in A. nidulans: Comparative sequence. The most likely ortholog identified during analysis of all putative RDRPs identified through litera- this search is a previously characterized A. nidulans RecQ ture searches (N. crassa and S. pombe) and database DNA helicase involved in the DNA damage response, Ϫ searches (A. nidulans, A. fumigatus, A. oryzae, and M. MusN (e 174; Hofmann and Harris 2001). The second- grisea) suggests that these enzymes fall into three distinct best match (An5092.2) had a relatively low expect value Ϫ classes in filamentous fungi represented by N. crassa (e 23) and was not investigated further during this study. QDE-1, SAD-1, and RRP-3 (Figure 6). This is in contrast However, the role of MusN in IRT-RNA silencing was to the single RDRP in S. pombe, which is most similar to investigated. A significant portion of musN was deleted the SAD-1 RDRP class (Galagan et al. 2003 and Figure from the A. nidulans genome (Hofmann and Harris 6). Surprisingly, while A. oryzae has conserved members 2001) and the resulting ⌬musN strain (AAH16) was of all three classes, there are no orthologs of QDE-1 in crossed to an aflR(IRT1300) strain (TTMH13.1) to give the A. nidulans genome database or orthologs of RRP-3 progeny with ⌬musN in the aflR(IRT1300) genetic back- in the A. fumigatus genome database. Analysis of the ground. NOR analysis indicated that loss of musN had Aspergilli genomes suggests that the A. nidulans QDE-1 no effect upon IRT-RNA silencing (Figure 5). This find- ortholog was lost during evolution (see below). ing is in agreement with that of Catalanotto et al. Comparison of the chromosomal region adjacent to (2004), who recently reported that QDE-3 is not re- the putative A. fumigatus QDE-1 ortholog (RrpA; Gala- quired for N. crassa IRT-RNA silencing. gan et al. 2003) indicates that a high level of synteny exists between this region and a specific chromosomal region in A. nidulans. A section of this synteny is depicted in Figure 7. These two chromosomal fragments are nearly identical except for the presence of a unique ORF in each, a rearrangement of two ORFs, and the absence of a predicted rrpA ortholog in A. nidulans (Fig- Figure 5.—Genetic analysis of IRT-RNA silencing in A. nidulans. A. nidulans mutants with (top) or without (bottom) the ure 7A and Table 3). aflR(IRT1300) transgene were point inoculated onto 25 ml Guessing that an A. nidulans QDE-1 ortholog might -kb genomic region corre-4.0ف of solid minimal medium and incubated for 6 days. NOR be undetected in the production was analyzed by TLC. Bright orange spots are sponding to A. fumigatus rrpA (Figure 7A, f), a search NOR. NA, not analyzed. Top row, strains are: WT, (blastx) of NCBI’s protein database was performed with RTMH13.B1; ⌬rsdA, RTMH65.1; ⌬musN, RTMH13.D9; and ⌬rrpB ⌬rrpC, RTMH7475.1A. Bottom row, strains are: WT, this sequence. No significant matches were identified RTMH13.B3; ⌬musN, RTMH13.D16; and ⌬rrpB ⌬rrpC, during this process. Contrastingly, a similar search with -kb region spanning the A. fumi-5.9ف RTMH7475.3A. the corresponding RNA Silencing in Aspergillus 613

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.Ge- neric 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 des- ignations above the A. fumigatus rrpA locus and the analogous region in A. nidulans indicate the approximate numbers of bases kb region of A. nidulans genomic DNA depicted in A was-4.0ف between the stop and start sites of the flanking genes. (B) The 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. gatus rrpA locus (Figure 7A, rrpA) identified N. crassa (Figure 5). This finding indicates that A. nidulans RrpB QDE-1 as the most significant match (eϪ96). This suggests and RrpC are not required for IRT-RNA silencing in A. .kb region of A. nidulans-4.0ف that a RDRP is not present in the kb A. nidulans-4.0ف nidulans genomic DNA. Next, the sequence was aligned (Clustal W) with the genomic sequences of 23 predicted or known genes from A. nidu- DISCUSSION lans, A. fumigatus, A. oryzae, and N. crassa and M. grisea. Here we demonstrate that IRTs efficiently induce The results indicated that this region is more similar to RNA silencing in A. nidulans in a process that is indepen- the sequences of the QDE-1-like RDRPs of all included dent of the two A. nidulans RDRPs, RrpB and RrpC. A fungi than to any other sequence included in the analy- third RDRP of the QDE-1 class of fungal RDRPs has sis, including the genomic DNA of the two predicted apparently degenerated over time through changes in A. nidulans RDRPs (Figure 7B and Table 4). This finding the DNA code. In addition, we have shown that RNA kb region-4.0ف is supportive of the hypothesis that this silencing is stable in A. nidulans, is characterized by a of DNA constitutes the remnants of an Aspergillus significant decrease in mRNA expression, results in the QDE-1 ortholog that has degenerated during evolution. accumulation of a single class of 25-nt siRNA molecules, Further support for the hypothesis that this region actu- and requires a PPD protein, RsdA. ally contains a relic RDRP and not a functional RDRP kb-4.0ف comes from a detailed six-frame analysis of the region. This analysis indicated that a conserved RDRP TABLE 3 motif (DbDGD, b is a bulky residue) found in all known RDRPs (Iyer et al. 2003), including the two putative A. Gene types near the A. fumigatus rrpA locus and the nidulans RDRPs, is not present in any of the six frames corresponding locus in A. nidulans (data not shown). Letter code Gene type Expect valuea IRT-RNA silencing in A. nidulans is independent of RDRPs: We next tested whether or not the two A. nidu- a Transporter e 0 Ϫ lans RDRPs, RrpB and RrpC, are essential for IRT-RNA c Transporter e 57 Ϫ216 silencing in A. nidulans. Replacement vectors were de- b Permease e Ϫ187 signed to eliminate the conserved RDRP domains in d Reductase e e Permease eϪ64 both genes (Figure 4 and Figure S1). Transformation rrpA QDE-1 like RDRP NA of A. nidulans RTMH13.F5 with the RDRP gene replace- f No predicted gene NA ment vectors resulted in the RDRP single knockouts, g Clock controlled eϪ142 TTMH74.12 and TTMH75.17 (Figure S1). Double- i Sugar utilization eϪ117 RDRP knock-out strains were then obtained by crossing h and j (unique ORFs) NA TTMH74.12 and TTMH75.17 (data not shown). Our a Expect values were determined by searching (blastp) the results indicated that neither the single nor double mu- A. fumigatus peptide database with the predicted A. nidulans tants were compromised in their ability to silence aflR ORFs depicted in Figure 7. NA, not applicable. 614 T. M. Hammond and N. P. Keller

TABLE 4 with Drosophila and mammals (Schwarz et al. 2002; Identification numbers for protein and nucleotide Roignant et al. 2003; Stein et al. 2003) suggest RDRPs sequences used in this study have no role in gene-silencing processes and, furthermore, are not evolutionary conserved in these species Branch no. Gene Identification no.a (Stein et al. 2003). Our current findings do not support a required role for an A. nidulans RDRP in IRT-RNA 1 M.g. rrp1 MG07682.4 silencing but do not negate a role for these enzymes in 2 N.c. qde-1 28920432,b 4803727c 3 A.n. degen. RrpA See belowd other gene-silencing processes. For example, N. crassa 4 A.o. rrpA 20113.m00115 SAD-1 is required for MSUD and successful progression 5 A.f. rrpA 71.m15624 through the meiotic cycle (Shiu et al. 2001) and our 6 A.n. actin 168004 preliminary results suggest that A. nidulans RrpB may 7 A.o. rrpB 20166.m00248 be required for some aspect of outcrossing (T. M. Ham- 8 A.n. rrpB An4790.2 mond and N. P. Keller, unpublished results). Certainly, 9 A.f. rrpB 59.m09193 the surprising finding that the three currently se- 10 M.g. rrp2 MG02748.4 11 N.c. sad-1 28926016,b 13699900c quenced Aspergilli contain three different combina- 12 A.o. rrpC 20139.m00205 tions of RDRP classes has identified a unique opportu- 13 A.n. rrpC An2717.2 nity to investigate how these different RDRPs might 14 N.c. rrp-3 32423631,b 32418006c affect fundamental gene-silencing processes. 15 M.g. rrp3 MG06205.4 The biological significance of possessing a QDE-1 type 16 A.n. dewA 533424 RDRP is currently unknown. It is possible that the loss 17 A.n. stcA 1235618 of a QDE-1 ortholog in A. nidulans may account for the 18 A.n. pkaR 3170247 19 A.n. cyaA 21326181 lack of reported quelling phenotypes in the A. nidulans 20 A.n. nimP 4102989 literature. Furthermore, our use of an aflR(SST) in this 21 A.n. ppoA 40715887 and previous studies (Shimizu and Keller 2001; Shim- 22 A.n. yA 2425 izu et al. 2003) suggests this form of RNA silencing 23 A.n. aflR 1235618 does not exist in A. nidulans. However, given the small 24 A.n. actin 168004 number of aflR(SST) transformants examined in this 25 N.c. qde-2 7248732 study, the tendency of the transformants to acquire only 26 N.c. qde-3 6934277 27 A.n. rsdA 40744619 a single copy of the transgene (data not shown), and 28 S.p. rdp1 2330856 the small percentage of transformants that stably display 29 A.n. musN 14039839 cosuppression phenotypes in N. crassa (Cogoni et al. 30 N.c. sms-2 23452214 1996; Cogoni and Macino 1997), our analysis of “quell- 31 A.n. smsA 40741781 ing” like RNA silencing in A. nidulans was not thorough 32 N.c. recQ-2 32403964 enough to firmly conclude that this process does not a Aspergillus and Magnaporthe RDRP nucleotide and amino occur in this fungus. acid data were obtained from the Broad Institute, TIGR, and In vitro analysis of N. crassa QDE-1 indicates that this the National Institute of Advanced Industrial Science and enzyme can make full-length dsRNA from a single- Technology (see acknowledgements for web site information) stranded template but that it preferentially makes small and the sequences are accessible using the given identification numbers. All other genes are listed with their GenBank acces- 9- to 21-nt single-stranded RNAs from RNA templates sion numbers. (Makeyev and Bamford 2002). Such a finding suggests b Nucleotide sequence. that QDE-1 may amplify siRNA levels during RNA silenc- c Amino acid sequence. ing. Such amplification could lead to transitive RNA ف d The 4.0-kb A. nidulans degenerate rrpA sequence can be silencing, which seems to occur in the fungus M. circinel- obtained from the A. nidulans genome database (www.broad. mit.edu) contig 1.161 (position 256,699–260,649). loides (Nicolas et al. 2003) but has not been reported in N. crassa. The fact that the low-molecular-weight RNA fraction in an A. nidulans aflR(IRT900)-carrying strain Before this report, RDRPs had not been investigated did not contain aflR siRNAs from the 3Ј-untargeted re- in the Aspergilli or in any other fungi with the exception gion of aflR suggests that 5Ј → 3Ј transitive RNA silenc- of S. pombe and N. crassa. Recent studies indicate that ing does not occur in this organism. Given the in vitro RDRPs are involved in various gene-silencing processes studies of QDE-1, it is possible that the absence of these in these two fungi (Cogoni and Macino 1999a; Shiu siRNAs in A. nidulans could be due to the loss of a et al. 2001; Volpe et al. 2002; Schramke and Allshire QDE-1 ortholog. Alternatively, the lack of secondary 2003; Verdel et al. 2004), protists (Martens et al. 2002), siRNAs could be the result of an inefficient amplifica- nematodes (Smardon et al. 2000; Sijen et al. 2001; Sim- tion of dsRNA by RrpB or RrpC from endogenous aflR mer et al. 2002), and plants (Dalmay et al. 2000; Beclin transcripts. Future studies with A. fumigatus or A. oryzae et al. 2002; Vaistij et al. 2002; Van Houdt et al. 2003; QDE-1 orthologs and more sensitive secondary siRNA Muangsan et al. 2004). On the other hand, studies detection methods should help test these hypotheses. RNA Silencing in Aspergillus 615

The ability of A. nidulans RDRP double mutants and functional RDRPs does not inhibit IRT-RNA silencing the inability of S. pombe Rdp1 mutants to perform IRT- in A. nidulans as it does in S. pombe, we propose that A. RNA silencing highlights a fundamental difference in nidulans does not commingle its RNA-silencing machin- how RDRP mutations affect the RNA-silencing pathway ery with heterochromatin and post-transcriptional si- in these model fungi. Theoretically IRTs should bypass lencing components. Therefore, we further speculate RDRPs in RNA silencing because siRNAs should be de- that RDRPs will be required for IRT-RNA silencing only rived from hpRNA in an RDRP-independent process. in those fungi that use the RNA-silencing machinery for Examples of RNA silencing working independently of both processes. RDRPs when dsRNA is introduced into the cell in an We thank Steve Harris for the A. nidulans ⌬musN strain and Mark RDRP-independent manner are present in the literature Caddick for preliminary analysis of the degenerate rrpA (qde-1) locus (Dalmay et al. 2000; Beclin et al. 2002; Muangsan et in A. nidulans. Genomic data for A. fumigatus was provided by the al. 2004) and the direct formation of hpRNA by an IRT Institute for Genomic Research (www.tigr.org/tdb/e2k1/afu1) and is a logical reason to explain why N. crassa QDE-1 is the Wellcome Trust Sanger Institute (www.sanger.ac.uk/Projects/ A_fumigatus); genomic data for A. nidulans was provided by the Broad not required for IRT-RNA silencing even though it is Institute (www.broad.mit.edu/annotation/fungi/ aspergillus/); and required for quelling. Therefore, why S. pombe, but not genomic data for A. oryzae was provided by the National Institute of A. nidulans, requires an RDRP for IRT-RNA silencing is Advanced Industrial Science and Technology (http://oryzae.cbrc.jp/ a perplexing question. and www.bio.nite.go.jp/dogan/Top). Coordination of analyses of In S. pombe, the RNA-silencing machinery is responsi- these data was enabled by an international collaboration involving more than 50 institutions from 10 countries and coordinated from ble for RNAi-dependent heterochromatin regulation Manchester, United Kingdom (www.cadre.man.ac.uk and www.asper (Volpe et al. 2002; Schramke and Allshire 2003; Ver- gillus.man.ac.uk). This work was supported by Hatch funds and the del et al. 2004). Deleting any of the core RNA-silencing graduate school of the University of Wisconsin-Madison. proteins or a histone methyltransferase (Clr4) disrupts this process (Volpe et al. 2002; Schramke and Allshire 2003). Surprisingly, IRT-RNA silencing is also elimi- LITERATURE CITED nated by a Clr4 mutation (Schramke and Allshire Aufsatz, W., M. F. Mette, J. Van Der Winden, A. J. Matzke and M. 2003). These results suggest that mutations in the RNAi- Matzke, 2002 RNA-directed DNA methylation in Arabidopsis. dependent heterochromatin regulation pathway dis- Proc. Natl. Acad. Sci. USA 99 (Suppl 4): 16499–16506. rupt IRT-RNA silencing and vice versa. Therefore it Beclin, C., S. Boutet, P. Waterhouse and H. Vaucheret, 2002 A branched pathway for transgene-induced RNA silencing in is possible that Rdp1 involvement in RNAi-dependent plants. Curr. Biol. 12: 684–688. heterochromatin regulation contributes to the collapse Bernstein, E., A. A. Caudy, S. M. Hammond and G. J. Hannon, of the IRT-RNA-silencing pathway. Such a collapse may 2001 Role for a bidentate ribonuclease in the initiation step of RNA interference. Nature 409: 363–366. not be encountered in A. nidulans because this fungus Borkovich, K. A., L. A. Alex, O. Yarden, M. Freitag, G. E. Turner may lack RNAi-dependent heterochromatin regulation. et al., 2004 Lessons from the genome sequence of Neurospora Although evidence for RNAi-dependent heterochro- crassa: tracing the path from genomic blueprint to multicellular organism. Microbiol. Mol. Biol. Rev. 68: 1–108. matin regulation has been reported in eukaryotes other Butchko, R. A., T. H. Adams and N. P. Keller, 1999 Aspergillus than S. pombe, such as Drosophila (Pal-Bhadra et al. nidulans mutants defective in stc gene cluster regulation. Genetics 2004) and Arabidopsis (Aufsatz et al. 2002), recent 153: 715–720. Carmell, M. A., Z. Xuan, M. Q. Zhang and G. J. Hannon, 2002 results have shown that N. crassa does not use either its The Argonaute family: tentacles that reach into RNAi, develop- RDRPs or its RNA-silencing machinery for transcrip- mental control, stem cell maintenance, and tumorigenesis. Genes tional silencing by heterochromatin formation (Chicas Dev. 16: 2733–2742. Catalanotto, C., G. Azzalin, G. Macino and C. Cogoni, 2002 et al. 2004; Freitag et al. 2004b). It is unknown if hetero- Involvement of small RNAs and role of the qde genes in the gene chromatin regulation requires RDRPs or any compo- silencing pathway in Neurospora. Genes Dev. 16: 790–795. nent of the RNA-silencing machinery in A. nidulans. Catalanotto, C., M. Pallotta, P. Refalo, M. S. Sachs, L. Vayssie et al., 2004 Redundancy of the two dicer genes in transgene- Currently, the Aspergilli appear to be “hybrid” with re- induced posttranscriptional gene silencing in Neurospora crassa. gard to gene-silencing equipment found in other fungi. Mol. Cell. Biol. 24: 2536–2545. For example, in contrast to the ﬁlamentous fungi stud- Chicas, A., C. Cogoni and G. Macino, 2004 RNAi-dependent and RNAi-independent mechanisms contribute to the silencing of ied so far, e.g., N. crassa (Kouzminova and Selker 2001), RIPed sequences in Neurospora crassa. Nucleic Acids Res. 32: 4237– Ascobolus immerses (Goyon and Faugeron 1989), and Co- 4243. prinus cinereus (Freedman and Pukkila 1993), but like Cogoni, C., and G. Macino, 1997 Isolation of quelling-defective (qde) mutants impaired in posttranscriptional transgene-induced Saccharomyces cerevisiae and S. pombe, Aspergillus species lack gene silencing in Neurospora crassa. Proc. Natl. Acad. Sci. USA signiﬁcant DNA methylation and concomitant methyla- 94: 10233–10238. tion-dependent gene inactivation (Gowher et al. 2001). Cogoni, C., and G. Macino, 1999a Gene silencing in Neurospora crassa requires a protein homologous to RNA-dependent RNA Also, analysis of their genomes shows that the Aspergilli polymerase. Nature 399: 166–169. lack a homolog of N. crassa DNA methylase DIM-2. Yet Cogoni, C., and G. Macino, 1999b Posttranscriptional gene silenc- the Aspergilli contain heterochromatin-maintenance ing in Neurospora by a RecQ DNA helicase. Science 286: 2342– 2344. homologs (e.g., HP1/Swi6) similar to those found in Cogoni, C., J. T. Irelan, M. Schumacher, T. J. Schmidhauser, N. crassa (Freitag et al. 2004a). Given that removal of E. U. Selker et al., 1996 Transgene silencing of the al-1 gene 616 T. M. Hammond and N. P. Keller

in vegetative cells of Neurospora is mediated by a cytoplasmic Muangsan, N., C. Beclin, H. Vaucheret and D. Robertson, 2004 effector and does not depend on DNA-DNA interactions or DNA Geminivirus VIGS of endogenous genes requires SGS2/SDE1 methylation. EMBO J. 15: 3153–3163. and SGS3 and defines a new branch in the genetic pathway for Dalmay, T., A. Hamilton, S. Rudd, S. Angell and D. C. Baulcombe, silencing in plants. Plant J. 38: 1004–1014. 2000 An RNA-dependent RNA polymerase gene in Arabidopsis Mullaney, E. J., J. E. Hamer, K. A. Roberti, M. M. Yelton and W. E. is required for posttranscriptional gene silencing mediated by a Timberlake, 1985 Primary structure of the trpC gene from transgene but not by a virus. Cell 101: 543–553. Aspergillus nidulans. Mol. Gen. Genet. 199: 37–45. Denli, A. M., and G. J. Hannon, 2003 RNAi: an ever-growing puzzle. Nicolas, F. E., S. Torres-Martinez and R. M. Ruiz-Vazquez, 2003 Trends Biochem. Sci. 28: 196–201. Two classes of small antisense RNAs in fungal RNA silencing Elbashir, S. M., W. Lendeckel and T. Tuschl, 2001 RNA interfer- triggered by nonintegrative transgenes. EMBO J. 22: 3983–3991. ence is mediated by 21- and 22-nucleotide RNAs. Genes Dev. 15: Pal-Bhadra, M., B. A. Leibovitch, S. G. Gandhi, M. Rao, U. Bhadra 188–200. et al., 2004 Heterochromatic silencing and HP1 localization in Fagard, M., S. Boutet, J. B. Morel, C. Bellini and H. Vaucheret, Drosophila are dependent on the RNAi machinery. Science 303: 2000 AGO1, QDE-2, and RDE-1 are related proteins required 669–672. for post-transcriptional gene silencing in plants, quelling in fungi, Pickford, A. S., C. Catalanotto, C. Cogoni and G. Macino, 2002 and RNA interference in animals. Proc. Natl. Acad. Sci. USA 97: Quelling in Neurospora crassa. Adv. Genet. 46: 277–303. 11650–11654. Pontecorvo, G., 1953 The genetics of Aspergillus nidulans, pp. 142– Forrest, E. C., C. Cogoni and G. Macino, 2004 The RNA-depen- 235 in Advances in Genetics, edited by M. Demerec. Academic dent RNA polymerase, QDE-1, is a rate-limiting factor in post- Press, New York. transcriptional gene silencing in Neurospora crassa. Nucleic Acids Punt, P. J., N. D. Zegers, M. Busscher, P. H. Pouwels and C. A. Res. 32: 2123–2128. Van Den Hondel, 1991 Intracellular and extracellular produc- Freedman, T., and P. J. Pukkila, 1993 De novo methylation of tion of proteins in Aspergillus under the control of expression repeated sequences in Coprinus cinereus. Genetics 135: 357–366. signals of the highly expressed Aspergillus nidulans gpdA gene. J. Freitag, M., P. C. Hickey, T. K. Khlafallah, N. D. Read and E. U. Biotechnol. 17: 19–33. Selker, 2004a HP1 is essential for DNA methylation in Neuro- Roignant, J. Y., C. Carre, B. Mugat, D. Szymczak, J. A. Lepesant spora. Mol. Cell 13: 427–434. et al., 2003 Absence of transitive and systemic pathways allows Freitag, M., D. W. Lee, G. O. Kothe, R. J. Pratt, R. Aramayo et al., cell-specific and isoform-specific RNAi in Drosophila. RNA 9: 2004b DNA methylation is independent of RNA interference in 299–308. Neurospora. Science 304: 1939. Schramke, V., and R. Allshire, 2003 Hairpin RNAs and retro- Galagan, J. E., S. E. Calvo, K. A. Borkovich, E. U. Selker, N. D. transposon LTRs effect RNAi and chromatin-based gene silenc- Read et al., 2003 The genome sequence of the filamentous ing. Science 301: 1069–1074. fungus Neurospora crassa. Nature 422: 859–868. Schwarz, D. S., G. Hutvagner, B. Haley and P. D. Zamore, 2002 Gowher, H., K. C. Ehrlich and A. Jeltsch, 2001 DNA from Aspergil- Evidence that siRNAs function as guides, not primers, in the lus flavus contains 5-methylcytosine. FEMS Microbiol. Lett. 205: Drosophila and human RNAi pathways. Mol. Cell 10: 537–548. 151–155. Shimizu, K., and N. P. Keller, 2001 Genetic involvement of a cAMP- Goyon, C., and G. Faugeron, 1989 Targeted transformation of dependent protein kinase in a G protein signaling pathway regu- Ascobolus immersus and de novo methylation of the resulting dupli- lating morphological and chemical transitions in Aspergillus nidu- cated DNA sequences. Mol. Cell. Biol. 9: 2818–2827. lans. Genetics 157: 591–600. Hamilton, A. J., and D. C. Baulcombe, 1999 A species of small Shimizu, K., J. K. Hicks, T. P. Huang and N. P. Keller, 2003 Pka, antisense RNA in posttranscriptional gene silencing in plants. Ras and RGS protein interactions regulate activity of AflR, a 286: Science 950–952. Zn(II)2Cys6 transcription factor in Aspergillus nidulans. Genetics Hammond, S. M., E. Bernstein, D. Beach and G. J. Hannon, 2000 165: 1095–1104. An RNA-directed nuclease mediates post-transcriptional gene si- Shiu, P. K., and R. L. Metzenberg, 2002 Meiotic silencing by un- lencing in Drosophila cells. Nature 404: 293–296. paired DNA: properties, regulation and suppression. Genetics Hammond, S. M., S. Boettcher, A. A. Caudy, R. Kobayashi and 161: 1483–1495. G. J. Hannon, 2001 Argonaute2, a link between genetic and Shiu, P. K., N. B. Raju, D. Zickler and R. L. Metzenberg, 2001 biochemical analyses of RNAi. Science 293: 1146–1150. Meiotic silencing by unpaired DNA. Cell 107: 905–916. Hofmann, A. F., and S. D. Harris, 2001 The Aspergillus nidulans Sienko, M., and A. Paszewski, 1999 The metG gene of Aspergillus musN gene encodes a RecQ helicase that interacts with the PI- nidulans encoding cystathionine beta-lyase: cloning and analysis. 3K-related kinase UVSB. Genetics 159: 1595–1604. Curr. Genet. 35: 638–646. Iyer, L. M., E. V. Koonin and L. Aravind, 2003 Evolutionary con- nection between the catalytic subunits of DNA-dependent RNA Sijen, T., J. Fleenor, F. Simmer, K. L. Thijssen, S. Parrish et al., polymerases and eukaryotic RNA-dependent RNA polymerases 2001 On the role of RNA amplification in dsRNA-triggered and the origin of RNA polymerases. BMC Struct. Biol. 3: 1. gene silencing. Cell 107: 465–476. Jia, S., K. Noma and S. I. Grewal, 2004 RNAi-independent hetero- Simmer, F., M. Tijsterman, S. Parrish, S. P. Koushika, M. L. Nonet chromatin nucleation by the stress-activated ATF/CREB family et al., 2002 Loss of the putative RNA-directed RNA polymerase proteins. Science 304: 1971–1976. RRF-3 makes C. elegans hypersensitive to RNAi. Curr. Biol. 12: Kouzminova, E., and E. U. Selker, 2001 dim-2 encodes a DNA 1317–1319. methyltransferase responsible for all known cytosine methylation Skory, C. D., J. S. Horng, J. J. Pestka and J. E. Linz, 1990 Transfor- in Neurospora. EMBO J. 20: 4309–4323. mation of Aspergillus parasiticus with a homologous gene (pyrG) Lee, D. W., R. J. Pratt, M. Mclaughlin and R. Aramayo, 2003 An involved in pyrimidine biosynthesis. Appl. Environ. Microbiol. argonaute-like protein is required for meiotic silencing. Genetics 56: 3315–3320. 164: 821–828. Smardon, A., J. M. Spoerke, S. C. Stacey, M. E. Klein, N. Mackin Lingel, A., B. Simon, E. Izaurralde and M. Sattler, 2003 Struc- et al., 2000 EGO-1 is related to RNA-directed RNA polymerase ture and nucleic-acid binding of the Drosophila Argonaute 2 PAZ and functions in germ-line development and RNA interference domain. Nature 426: 465–469. in C. elegans. Curr. Biol. 10: 169–178. Makeyev, E. V., and D. H. Bamford, 2002 Cellular RNA-dependent Stein, P., P. Svoboda, M. Anger and R. M. Schultz, 2003 RNAi: RNA polymerase involved in posttranscriptional gene silencing mammalian oocytes do it without RNA-dependent RNA polymer- has two distinct activity modes. Mol. Cell 10: 1417–1427. ase. RNA 9: 187–192. Martens, H., J. Novotny, J. Oberstrass, T. L. Steck, P. Postleth- Song, J. J., S. K. Smith, G. J. Hannon and L. Joshua-Tor, 2004 Crys- wait et al., 2002 RNAi in Dictyostelium: the role of RNA- tal structure of Argonaute and its implications for RISC slicer directed RNA polymerases and double-stranded RNase. Mol. Biol. activity. Science 305: 1434–1437. Cell 13: 445–453. Tabara, H., M. Sarkissian, W. G. Kelly, J. Fleenor, A. Grishok et Mooney, J. L., D. E. Hassett and L. N. Yager, 1990 Genetic analysis al., 1999 The rde-1 gene, RNA interference, and transposon of suppressors of the veA1 mutation in Aspergillus nidulans. Genet- silencing in C. elegans. Cell 99: 123–132. ics 126: 869–874. Tang, G., B. J. Reinhart, D. P. Bartel and P. D. Zamore, 2003 A RNA Silencing in Aspergillus 617

biochemical framework for RNA silencing in plants. Genes Dev. Yan, K. S., S. Yan, A. Farooq, A. Han, L. Zeng et al., 2003 Structure 17: 49–63. and conserved RNA binding of the PAZ domain. Nature 426: Tsitsigiannis, D. I., R. Zarnowski and N. P. Keller, 2004 The 468–474. lipid body protein, PpoA, coordinates sexual and asexual sporula- Yang, S. J., S. A. Carter, A. B. Cole, N. H. Cheng and R. S. Nelson, tion in Aspergillus nidulans. J. Biol. Chem. 279: 11344–11353. 2004 A natural variant of a host RNA-dependent RNA polymer- Vaistij, F. E., L. Jones and D. C. Baulcombe, 2002 Spreading of ase is associated with increased susceptibility to viruses by Nicoti- RNA targeting and DNA methylation in RNA silencing requires ana benthamiana. Proc. Natl. Acad. Sci. USA 101: 6297–6302. transcription of the target gene and a putative RNA-dependent Yu, D., B. Fan, S. A. Macfarlane and Z. Chen, 2003 Analysis of RNA polymerase. Plant Cell 14: 857–867. the involvement of an inducible Arabidopsis RNA-dependent Van Houdt, H., A. Bleys and A. Depicker, 2003 RNA target se- RNA polymerase in antiviral defense. Mol. Plant-Microbe Inter- quences promote spreading of RNA silencing. Plant Physiol. 131: act. 16: 206–216. 245–253. Yu, J. H., and T. H. Adams, 1999 Filamentous fungi, pp. 417–434 Verdel, A., S. Jia, S. Gerber, T. Sugiyama, S. Gygi et al., 2004 RNAi- in Manual of Industrial Microbiology and Biotechnology, edited by A. mediated targeting of heterochromatin by the RITS complex. L. Demain and J. E. Davies. American Society for Microbiology Science 303: 672–676. Press, Washington, DC. Volpe, T., V. Schramke, G. L. Hamilton, S. A. White, G. Teng et Zolotukhin, S., M. Potter, W. W. Hauswirth, J. Guy and N. al., 2003 RNA interference is required for normal centromere Muzyczka, 1996 A “humanized” green ﬂuorescent protein function in ﬁssion yeast. Chromosome Res. 11: 137–146. cDNA adapted for high-level expression in mammalian cells. J. Volpe, T. A., C. Kidner, I. M. Hall, G. Teng, S. I. Grewal et al., Virol. 70: 4646–4654. 2002 Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by RNAi. Science 297: 1833–1837. Communicating editor: M. S. Sachs