Available online at www.sciencedirect.com

www.elsevier.com/locate/yfgbi

Fungal Genetics and Biology 45 (2008) 719–727

DCL-1 colocalizes with other components of the MSUDmachinery and is required for silencing

William G. Alexander a,1, Namboori B. Raju b,1, Hua Xiao a,1, Thomas M. Hammond a,Tony D. Perdue c, Robert L. Metzenberg d, Patricia J. Pukkila c, Patrick K.T. Shiu a,*

a Division of Biological Sciences, University of Missouri, Columbia, MO 65211, USAb Department of Biological Sciences, Stanford University, Stanford, CA 94305, USAc Department of Biology, University of North Carolina, Chapel Hill, NC 27599, USA

d Department of Biology, California State University, Northridge, CA 91330, USA

Received 13 September 2007; accepted 10 October 2007Available online 22 October 2007

Abstract

In Neurospora, a gene present in an abnormal number of copies is usually a red flag for mischief. One way to deal with these potentialintruders is by destroying their transcripts. Widely known as RNA interference (RNAi), this mechanism depends on the ‘‘dicing’’ of adouble-stranded RNA intermediate into small-interfering RNA, which in turn guide the degradation of mRNA from the target gene.Quelling is a vegetative silencing system in Neurospora that utilizes such a mechanism. Quelling depends on the redundant activity oftwo Dicer-like ribonucleases, DCL-1 and DCL-2. Here, we show that Meiotic Silencing by Unpaired DNA (MSUD), a mechanism thatsilences expression from unpaired DNA during meiosis, requires the dcl-1 (but not the dcl-2) gene for its function. This result suggeststhat MSUD operates in a similar manner to Quelling and other RNAi systems. DCL-1 colocalizes with SAD-1 (an RdRP), SAD-2, andSMS-2 (an Argonaute) in the perinuclear region.� 2007 Elsevier Inc. All rights reserved.

Keywords: Dicer; Epigenetics; Meiotic silencing by unpaired DNA (MSUD); Neurospora crassa; Post-transcriptional gene silencing

1. Introduction

Neurospora crassa is a multicellular organism made upof a network of hyphae called a mycelium. Since the sep-tum between two adjacent cells is normally incomplete,nuclei as well as other cellular contents can migrate freelyalong their shared cytoplasm through protoplasmicstreaming. N. crassa is thus especially vulnerable to inva-sion by viruses or other selfish DNA elements. Accord-ingly, several mechanisms have evolved that silence theexpression of repetitive sequences. For example, Quellingis a vegetative silencing system that destroys the transcriptsmade from hyperhaploid transgenes (Catalanotto et al.,

1087-1845/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.fgb.2007.10.006

* Corresponding author. Fax: +1 573 882 0123.E-mail address: shiup@missouri.edu (P.K.T. Shiu).

1 These authors contributed equally to this work.

2006). Repeat-induced point mutation (RIP), which oper-ates before karyogamy, is a mechanism that targets dupli-cated sequences for mutation (Galagan and Selker, 2004).We recently discovered a third silencing mechanism termedMeiotic Silencing by Unpaired DNA (MSUD), which pro-tects the genome during the sexual cycle (Shiu et al., 2001).In this system, a gene not paired during the meiotic homo-log pairing stage produces a signal that silences all copies ofthat gene during meiosis.

Quelling is related to an RNA-based silencing phenom-enon often referred to as RNA interference (RNAi; Melloand Conte, 2004). In Quelling, the presence of multiplecopies of a transgene presumably alerts the host defensemechanism, triggering the production of a single-strandedaberrant RNA (Catalanotto et al., 2006). QDE-1, anRNA-directed RNA polymerase (RdRP), converts theaberrant RNA into a double-stranded species (dsRNA).

720 W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727

The dsRNA molecules are processed by Dicers into small-interfering RNA (siRNA) of 21–25 nt, which subsequentlyguide the qde-2-encoded Argonaute nuclease (and its inter-acting protein, QIP) to degrade homologous mRNA (Cata-lanotto et al., 2002; Maiti et al., 2007). Since an RdRP(SAD-1) and an Argonaute protein (SMS-2) are alsorequired for meiotic silencing (Shiu and Metzenberg,2002; Lee et al., 2003), it is possible that this sexual silenc-ing process operates in a similar fashion. For the Quellingmechanism, it is known that the QDE-1 RdRP can synthe-size short RNA directly from single-stranded RNA, a sub-set of which has size similar to that of siRNA (Makeyevand Bamford, 2002). It is unclear whether the SAD-1RdRP can also produce such small RNA species, andwhether such an activity can bypass the requirement ofDicers, as once suggested for the Quelling mechanism. Inthis paper, we have asked if the two Dicer genes encodedin the Neurospora genome are involved in meiotic silencing.

2. Materials and methods

2.1. Strains and media

Neurospora strains used in this study are listed in Table1. Auxotrophic strains used in this study were obtainedthrough the Fungal Genetics Stock Center (FGSC;McCluskey, 2003). Descriptions of the dcl-1D and dcl-2D

deletion alleles can be found in Catalanotto et al. (2004).The dcl-1fs (frameshift) allele was created by a restrictioncut at the unique AgeI site (positions 1210–1215; Fig. 1)followed by a Klenow fill-in. Descriptions of other lociused in this study can be found in Perkins et al. (2001). Cul-turing and crossing media were prepared as previouslydescribed (Vogel, 1964; Westergaard and Mitchell, 1947).Crosses, genetic selection, and other routine manipulation

Table 1Neurospora strains used in this study

Strain Genotype

23-23 his-3+::dcl-1fs; dcl-2D::hph mat A

24-15 his-3+::hH1-gfp; dcl-2D::hph mat a

25-31 his3+::hH1-gfp; dcl-2D::hph; pyr-1 dcl-1D::hph arg-2 mat a + [G]

28-29 dcl-2D::hph mat A

28-30 dcl-2D::hph mat a

29-01 dcl-1D::hph mat A

29-02 dcl-1D::hph mat a

31-32 dcl-2D::hph; dcl-1D::hph mat A

40-27 Oak Ridge wild-type mat A

48-36 rid his-3+::hH1-gfp; dcl-1D::hph mat A

53-14 ad-3B cyh-1 mat am1 (Griffiths helper [G])

69-13 rid his-3+::dcl-1fs; mus-52D::bar mat a

81-02 fl mat a

96-02 Sad-1D::hph mep mat A

N2281 rid his-3+::hH1-gfp mat a

P5-70 rid his-3+::sad-1-gfp mat A

P6-62 rid his-3+::sad-2-RFP; inv Sad-2RIP mat a

P10-18 rid his-3+::dcl-1-gfp; dcl-1D::hph mus-52D::bar mat A

P10-43 rid his-3+::sad-2-gfp; inv Sad-2RIP mat A

P11-08 rid his-3+::sms-2-rfp; sms-2D::hph mat a

were performed according to Davis and DeSerres (1970).Homokaryon isolation from heterokaryotic transformantswas performed using the protocol of Ebbole and Sachs(1990).

2.2. DNA manipulation and transformation

Cloning and other molecular biology procedures were asdescribed by Sambrook and Russell (2001). The DNeasyPlant Mini Kit (Qiagen, Valencia, CA) was used for Neu-

rospora DNA isolation. Custom oligonucleotides wereobtained from Invitrogen (Carlsbad, CA). Polymerasechain reaction (PCR) was performed in a PTC-100 PeltierThermal Cycler (MJ Research, Waltham, MA), usingeither the FastStart High Fidelity PCR System or theExpand Long Template PCR System (Roche Applied Sci-ence, Indianapolis, IN). The pCRII-TOPO vector (Invitro-gen, Carlsbad, CA) was used for the cloning of PCR-amplified fragments. Plasmid DNA was purified from bac-terial culture with the HiSpeed Plasmid Midi Kit (Qiagen).DNA sequencing services were provided by the Universityof Missouri DNA core (Columbia, MO).

Primers involved in the construction of dcl-1fs, dcl-1-gfp,and sms-2-rfp are listed in Table 2, with the resulting PCRproducts inserted into pBM61 (Margolin et al., 1997),pMF272 (Freitag et al., 2004) and pMF334 (Freitag andSelker, 2005), respectively. Fungal gene placement at thehis-3 locus, using electroporation of washed conidia, wasperformed according to Margolin et al. (1997).

2.3. Reverse-transcriptase (RT)-PCR

Total RNA was extracted as previously described (Log-emann et al., 1987; Shiu and Glass, 1999) and enriched forpoly(A)+ using the Oligotex mRNA kit (Qiagen, Valencia,CA). For RT-PCR, poly(A)+-enriched RNA was used forcDNA synthesis using the first-strand cDNA synthesis kit(Amersham Biosciences, Piscataway, NJ). Primers used inamplification of the cDNA regions spanning the intronsare listed in Table 2. The PCR products of the cDNA(without introns), as compared to that of the DNA, areas follows: 159 bp vs. 221 bp for dcl-1 intron 1, 319 bpvs. 371 bp for dcl-1 intron 2, 207 bp vs. 389 bp for dcl-2

introns 1–3, and 550 bp vs. 600 bp for dcl-2 intron 4. Allintron identities were confirmed by DNA sequencing ofRT-PCR products from both vegetative mRNA (from40-27) and perithecial mRNA (from 40-27 · 81-02).

2.4. Cytological methods

For crosses involving the histone H1-gfp gene (hH1-gfp),strains containing hH1-gfp were used as the male parent,whenever possible, to limit the green fluorescence to ascog-enous hyphae and asci. At least 10 perithecia were dissectedat 1–2 day intervals in a drop of 10% glycerol between 3and 8 days after crossing. The rosettes of asci were gentlysquashed under a cover glass. At least 1000 asci were exam-

Fig. 1. The nucleotide and amino acid sequence of dcl-1 (NCU08270) of N. crassa. The two introns are italicized, whereas sequences matching theconsensus for the TATA box and the transcriptional start site are underlined. The DNA sequence corresponds to nucleotides 225299–219300 of contig7.53 of the Neurospora genome (http://www.broad.mit.edu).

W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727 721

Table 2Primers used in this study

Primer Sequence (50 to 30) Amplification of

Dcl1-5207F 368 GAGAATATCAGGTGGAGCTTTTTGAACG 395 cDNA spanning dcl-1 intron 1Dcl1-5427R 588 CAATCCGTCTGGGTAAACCCTTTTCTC 562 cDNA spanning dcl-1 intron 1Dcl1-7911F 3072 GACGATACGCTGAGGGATGAGGA 3094 cDNA spanning dcl-1 intron 2Dcl1-8281R 3442 TCAAGAAGGAGTCACCAAGAAAC 3420 cDNA spanning dcl-1 intron 2Dcl2-32248F 202 GCCTTGACCGCGAGGGCTTATCAAC 226 cDNA spanning dcl-2 introns 1–3Dcl2-32636R 590 TTGTGATTGCAAAACTCGGTGTTGTTG 564 cDNA spanning dcl-2 introns 1–3Dcl2-34775F 2729 CTCCCCTCAGCCTCCCCTTGTGA 2751 cDNA spanning dcl-2 intron 4Dcl2-35374R 3328 TAACCGTTCATGGTGGTTTGGAA 3306 cDNA spanning dcl-2 intron 4Dcl1-4387F �453 TCTGTTGCGAGTGTTTGTATAAAAGCG �427 dcl-1 for frameshift constructionDcl1-10100R 5261 TTCACGAGGCTTACAGGATTTCCATAC 5235 dcl-1 for frameshift constructionDcl1-4816F �24 GGAAAA-ACTAGT-CAGCTCGTCACCATGGCCGTAG 10 dcl-1 for GFP localizationDcl1-9720R 4881 GTAGAAT-TTAATTAA-AACCGCCGTGCCATGTACC 4848 dcl-1 for GFP localizationSms2-107713F 107713 CTCTTGCCTCAACCAGTACC-ACTAGT-ATGTCTGCTCCTGG 107752 sms-2 for RFP localizationSms2-110724R 110724 GCCAAAGCGACCAAG-TCTAGA-CCCACCACATGGTGTTGTG 110685 sms-2 for RFP localization

The SpeI (ACTAGT), PacI (TTAATTAA), and XbaI (TCTAGA) site are underlined. Nucleotide positions are listed according to Figs. 1 and 2.

722 W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727

ined for each cross for determining whether hH1-gfp wasexpressed or silenced in the developing asci and ascospores.A Nikon Microphot FX fluorescence microscope fittedwith an excitation/dichroic mirror/long pass filter set upwas used, as described in Freitag et al. (2004). See Raju(1980) for a review on normal ascus development in N.

crassa.For localization studies involving GFP and RFP, peri-

thecia were fixed in freshly prepared 4% paraformaldehyde,90 mM PIPES pH 6.9, 10 mM EGTA, 5 mM MgSO4 for20 min at room temperature. They were rinsed briefly inPBS, and the perithecial contents were teased out into adrop of 90% glycerol, 10% 100 mM potassium phosphate(pH 8.7), 10 lg/ml DAPI, and 100 mg/ml 1,4-diazabicy-clo[2,2,2]octane. Contents of 3–5 perithecia were dispersedunder a cover slip by gentle tapping, slides were sealedusing clear nail polish, and either viewed immediately orafter storage at �20 �C. Samples were imaged using a ZeissLSM510META confocal laser scanning microscope. Allimages were collected using a PlanNeofluar 40· (NA1.3)oil immersion objective. Multi-fluorophore images werescanned sequentially. Visualization of the GFP wasachieved by use of a 488 nm Argon laser line for excitationand a BP505-530 emission filter. RFP visualization wasachieved by use of a 543 nm HeliumNeon laser line forexcitation and the META detector set to collect emissionbandwidth 560–625 nm in Channel Mode. DAPI visualiza-tion was achieved by use of 364 nm Argon laser excitationand a BP385-470 emission filter. All images were collectedusing standard Zeiss software.

3. Results

3.1. Detection of dcl-1 and dcl-2 mRNA transcript by RT-

PCR

dcl-1 is located on the right arm of Linkage Group (LG)IV, between arg-14 and his-4, whereas dcl-2 is located onthe right arm of LG II, between aro-3 and preg. Whilethe dcl-1 ORF was correctly predicted by the Neurospora

Genome Database (Fig. 1), dcl-2 was found to contain 4(instead of the predicted 3) introns by our RT-PCR exper-iment (Fig. 2). Translation of the dcl-2 ORF yields a 1539-amino-acid (aa) polypeptide (instead of the predicted 1421-aa polypeptide).

The cDNA products of dcl-1 and dcl-2 mRNA (the iden-tities of which were confirmed by sequencing) could bedetected in mycelia and in two perithecial preparations (4and 6 days after fertilization), suggesting that the two genesare expressed in both vegetative and sexual tissues (see Sec-tion 2.3).

3.2. Mutation in dcl-1 affects sexual development

Two Dicer-like genes, dcl-1 and dcl-2, are involved inQuelling in N. crassa (Catalanotto et al., 2004). The twogenes are redundant for the Quelling function, and defectsin silencing can only be observed in a dcl-1D dcl-2D doublemutant. We did not observe any obvious aberrant pheno-types in the dcl-1D, dcl-2D, or dcl-1D dcl-2D mutants duringvegetative growth. Sexual development in a cross homozy-gous for the dcl-2D mutation also appears to be normal.However, a cross homozygous for the dcl-1D mutation iscompletely barren. Although pigmented perithecia are pro-duced in a dcl-1D · dcl-1D cross, they are slightly smallerthan those from a wild-type (WT) cross. The beaks of theperithecia are small and under-developed. Dissection ofthe perithecia shows that they contain a very small ballof tissue, which is mostly made up of paraphyseal cells.There are no recognizable ascogenous tissues, croziers, orasci within the perithecia. These data suggest that thedevelopment of perithecia is arrested at an early stage ina dcl-1D · dcl-1D cross.

3.3. dcl-1D, dcl-2D, and dcl-1D dcl-2D do not act as dominant

suppressors of meiotic silencing

sad-1 and sad-2 each encode a component of the meioticsilencing machinery (Shiu et al., 2001, 2006). Deletion andheavily mutated strains of sad-1 and sad-2 act as dominant

Fig. 2. The nucleotide and amino acid sequence of dcl-2 (NCU06766) of N. crassa. The four introns are italicized, whereas sequences matching theconsensus for the TATA box and the transcriptional start site are underlined. The DNA sequence corresponds to nucleotides 97103–103102 of contig 7.33of the Neurospora genome (http://www.broad.mit.edu).

W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727 723

724 W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727

suppressors of meiotic silencing. In a cross, these Sad

mutant alleles presumably prevent the wild-type sad-1 genefrom proper pairing, thus silencing the silencer and creatinga negative feedback loop for the meiotic silencing mecha-nism. To determine if such dominant suppression can beobserved in heterozygous crosses of Dicer mutants, weexamined the ability of Dicer deletion mutants to preventthe meiotic silencing of an unpaired copy of hH1-gfp

inserted at the his-3 locus (his-3+::hH1-gfp). When crossedto a his-3+::hH1-gfp strain, dcl-1D, dcl-2D, and dcl-1D dcl-2D

behave like wild-type (Fig. 3A), in that the unpaired hH1-

gfp gene is silenced throughout meiosis and postmeioticmitosis. These observations suggest that unlike Sad-1 andSad-2 (Fig. 3B), dcl-1D, dcl-2D, and dcl-1D dcl-2D mutantsdo not act as dominant suppressors of meiotic silencing.Additionally, meiotic silencing of his-3+::hH1-gfp appearsto function normally in a cross heterozygous for dcl-1D

Fig. 3. Rosettes of asci from various dcl-1 and dcl-2 crosses, showing that the d

gene. (A) Wild-type (40-27) · hH1-gfp (N2281), 4 days after fertilization. The uthe ascus rosette, but is completely silenced during meiosis and postmeiotic mitafter fertilization. Meiotic silencing of hH1-gfp is suppressed by the self-silencascus development. Initially, the nuclei fluoresce in all eight young ascospores,ascospores, because of protein turnover. (C) dcl-1D dcl-2D (31-32) · dcl-2D hH1-g

dominant suppressor of meiotic silencing. A few very young asci show fluorescmitosis do not show expression of hH1-gfp. (D) dcl-1fs dcl-2D (23-23) · dcl-1D dc

of hH1-gfp, apparently because of a low Dicer environment (dcl-1 knockdownafter fertilization, A rosette of asci showing the expression of unpaired hH1-gfp

is just beginning to fade in four of the eight ascospores in a couple of asci. Evide2D (28-29) · dcl-2D hH1-gfp (24-15), 5 days after fertilization. The unpaired hH

and homozygous for dcl-2D (dcl-1D dcl-2D · his-3+::hH1-

gfp dcl-2D; Fig. 3C).

3.4. A cross deficient in dcl-1 and dcl-2 gene products is

defective in meiotic silencing

In the previous experiment, we showed that the Dicerdeletion mutants do not act as dominant suppressors ofmeiotic silencing. These results suggest that either (1) thetwo Dicer genes are not involved in meiotic silencing, or(2) an unpaired copy of a dcl gene is not sufficient to silenceits own expression. To determine the role of Dicers in mei-otic silencing unequivocally, we need to assay the expres-sion of an unpaired gene in a dcl-1D dcl-2D · dcl-1D dcl-2D

background. Because dcl-1D is required for early sexualdevelopment (see Section 3.2), this cross is technicallyimpossible.

cl-1 function is important for the meiotic silencing of the unpaired hH1-gfp

npaired hH1-gfp gene is expressed in the ascogenous hyphae at the base ofosis in the developing asci. (B) Sad-1D (96-02) · hH1-gfp (N2281), 6–7 daysing effect of the unpaired sad-1 gene, and the nuclei fluoresce throughoutbut the fluorescence gradually fades away in four of the eight non-hH1-gfp

fp (24-15), 4 days after fertilization. Unlike Sad-1D, dcl-1D does not act as aence (albeit faintly), but most of the older asci in meiosis and postmeioticl-2D hH1-gfp (25-31), 5 days after fertilization. There is no meiotic silencing

and dcl-2 knockout). (E) dcl-1fs (69-13) · dcl-1D hH1-gfp (48-36), 6 daysthroughout meiosis and in all eight young ascospres. The hH1-gfp protein

ntly, gene knockdown of dcl-1 alone can suppress meiotic silencing. (F) dcl-

1-gfp is silenced even in the absence of dcl-2.

W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727 725

To address this problem, we devised an experiment inwhich dcl-1 could be expressed at an early stage of the sex-ual cycle and inactivated at later stages by MSUD. Accord-ingly, we have constructed a cross homozygous for dcl-2D,heterozygous for dcl-1D, heterozygous for an insertion ofdcl-1fs (a frameshift null allele), and heterozygous for aninsertion of hH1-gfp at the his-3 locus, i.e. his-3+::dcl-1fs

dcl-2D · his-3+::hH1-gfp dcl-1D dcl-2D. There are no dcl-

2+ copies in this cross. In addition, the lone dcl-1+ copyis subject to meiotic silencing triggered by two unpairingevents (dcl-1+ vs. dcl-1D::hph at the endogenous dcl-1 locusand dcl-1fs vs. hH1-gfp at the his-3 locus). Our results showthat the unpaired hH1-gfp gene is indeed expressedthroughout meiosis in such a background, indicating thatmeiotic silencing is deficient in a low Dicer environment(Fig. 3D). This result suggests that Dicers play a role inmeiotic silencing.

Fig. 4. Colocalization of MSUD proteins in the perinuclear region.Micrographs illustrate asci in meiotic prophase. (A–D) dcl-1-gfp

(P10-18) · sms-2-rfp (P11-08). (E–H) dcl-1-gfp (P10-18) · sad-2-rfp (P6-62). (I–L) sad-2-gfp (P10-43) · sms-2-rfp (P11-08). (M–P) sad-1-gfp

(P5-70) · sms-2-rfp (P11-08). Chromatin was visualized with DAPI. Thebar equals 5 lm.

3.5. dcl-1 is involved in Meiotic Silencing

Our previous results establish the fact that Dicers areinvolved in the mechanism of meiotic silencing. To deter-mine whether dcl-1 and dcl-2 provide redundant functionfor meiotic silencing, as is the case for their roles in Quell-ing, we tested the dcl-1D and dcl-2D mutations individuallyfor their ability to suppress the silencing of unpaired hH1-

gfp. In a cross deficient only in dcl-1 gene products, i.e. his-

3+::dcl-1fs · his-3+::hH1-gfp dcl-1D, we observed theexpression of the unpaired hH1-gfp gene throughout meio-sis (Fig. 3E). However, we did not observe any expressionof hH1-gfp in a cross deficient only in dcl-2 gene products,i.e. dcl-2D · his-3+::hH1-gfp dcl-2D (Fig. 3F). These resultssuggest that while dcl-1 is important for meiotic silencing,dcl-2 is not required for the process.

3.6. DCL-1 colocalizes with other MSUD proteins

SAD-1 and SAD-2, two components of the meioticsilencing machinery, colocalize in the perinuclear region(Shiu et al., 2006). Since DCL-1 and SMS-2 are alsoinvolved in meiotic silencing, we tested whether the twoproteins also colocalize in the same region. We made vari-ous fusion constructs using vectors encoding green and redfluorescent proteins (GFP and RFP) and determined thelocalization of their gene products in several pairwisecrosses. Our results demonstrate that DCL-1 and SMS-2colocalize with SAD-1 and SAD-2 in the perinuclear region(Fig. 4).

4. Discussion

Two Dicer-like genes, dcl-1 and dcl-2, are found in Neu-

rospora. The two Dicer-like genes are apparently redun-dant in function for Quelling, since the double knockoutmutant (dcl-1D dcl-2D) is deficient in silencing while singleknockout mutants (dcl-1D and dcl-2D) are not (Catalanottoet al., 2004). In this study, we demonstrate that the twogenes are not redundant for the meiotic silencing function,with only dcl-1 required for the process.

Previously, we have shown that sad-1, which encodes anRNA-directed RNA polymerase (RdRP), is required formeiotic silencing (Shiu and Metzenberg, 2002). Therequirement of an RdRP implies the involvement of a dou-ble-stranded RNA species in the pathway. Our currentmodel suggests that such a dsRNA species is made froma single-stranded aberrant RNA, which in turn is tran-scribed from an unpaired DNA region during the silencingprocess. The requirement of an Argonaute protein for mei-otic silencing is further indication that this mechanism isdependent on the siRNA-guided destruction of mRNAtranscripts (Lee et al., 2003). Our results here demonstratethat the DCL-1 protein is likely the enzyme responsible forproducing siRNA from the dsRNA species.

Sad-1D and Sad-2D act as dominant suppressors ofmeiotic silencing (Shiu et al., 2001, 2006). We reason that

726 W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727

in a SadD · sad+ cross, the wild-type sad gene is unpairedduring meiosis. Subsequently, the sad silencer wouldsilence itself, thus preventing an abundant productionof the sad gene product and suppressing the silencingmechanism. Unlike Sad-1 and Sad-2, the dcl-1D mutantdoes not act as a dominant suppressor of meiotic silenc-ing. One possible explanation is that, in a dcl-1D · dcl-1+

cross, the one unpairing event is insufficient in silencingthe dcl-1+ expression. Because dcl-1 products are neededfor early sexual development, we cannot create a dcl-1-null environment through a cross homozygous for thedcl-1 mutation. We addressed this problem by puttingone functional unpaired dcl-1+ copy and a second,non-functional unpaired dcl-1fs copy through a cross.The dcl-1+ copy provides enough gene product for earlysexual development. However, because an unpaired,ectopically-inserted dcl-1fs copy is also present in thecross, the lone dcl-1+ gene will be silenced by twounpaired copies (dcl-1+ and dcl-1fs) at later stages, thuspreventing meiotic silencing from functioning properly.We predict that this strategy can be used to knock downthe function of any particular gene of interest specificallyduring the later stages of sexual development.

Our previous study demonstrates that SAD-1 and SAD-2 colocalize in the perinuclear region, where siRNA areshown to localize in mammalian cells (Shiu et al., 2006).SAD-1 is localized in the cytosol in the absence of SAD-2, suggesting that SAD-2 may function to recruit SAD-1to the perinuclear region. The importance of their localiza-tion is not clear. It is possible that, when single-strandedaberrant RNA are exported from the nucleus, the gauntletof SAD-1 RdRP molecules can convert them into dsRNAbefore they have a chance to reach the exonucleases or thetranslational machinery in the bulk cytosol. Different Dic-ers are located in different compartments of the cell; someare localized in the nucleus, while others are localized in thecytoplasm (Papp et al., 2003; Billy et al., 2001). Our datademonstrate that DCL-1 colocalizes in the perinuclearregion with all the known factors of meiotic silencing,including SAD-1 RdRP, SAD-2, and SMS-2 Argonaute(Sk-2 and Sk-3 have not been molecularly characterized;Raju et al., 2007). This result suggests that the perinuclearregion is the center of RNAi activity for meiotic silencing.Interestingly, some components of the RNA-silencingmachinery have been shown to localize in the perinuclearregion of Drosophila and mouse germ cells (Lim and Kai,2007; Pane et al., 2007; Kotaja and Sassone-Corsi, 2007).Colocalization of RNAi proteins may allow the couplingof various reactions, thereby increasing the efficiency ofgene silencing.

Like sad-1 and sad-2, a cross homozygous for dcl-1D isbarren. The progression of meiosis is blocked at the samestage for sad-1 and sad-2, namely at early prophase (Shiuet al., 2001, 2006). This similarity may be attributed tothe fact that the SAD-1 and SAD-2 proteins work inti-mately together and that SAD-2 controls the perinuclearlocalization of SAD-1. Although asci are made in a Sad · -

Sad cross, the same cannot be said for a cross homozygousfor dcl-1D. In such a cross, the sexual development isblocked at a much earlier stage, and the perithecia haveno observable sexual tissue. These results suggest that thebarrenness of a dcl-1D · dcl-1D cross could be due to thedefects of a different cellular pathway. One possibility isthat DCL-1 is important for processing endogenous hair-pin dsRNA into microRNA required for sexual develop-ment. In plants and animals, a microRNA species canbind to its complimentary mRNA, leading to their degra-dation or translational blockage (Bartel, 2004). Since micr-oRNA are yet to be discovered in fungi, it would beinteresting to determine if such non-coding RNA exist inNeurospora and whether they require DCL-1 formaturation.

Acknowledgments

We are grateful to the members of the Shiu Laboratoryfor their technical assistance, to Carlo Cogoni and Gius-eppe Macino for the dcl-1D and dcl-2D strains, and to Mi-chael Freitag and Eric Selker for the pBM61, pMF272,and pMF334 plasmids. This work was sponsored by Na-tional Science Foundation Grants MCB0417282 (in sup-port of N.B.R.), MCB0234421 (to R.L.M.), EF0412016(to P.J.P.), and MCB0544237 (to P.K.T.S.). We dedicatethis article to the memory of our esteemed colleague Rob-ert Metzenberg (1930–2007).

References

Bartel, D.P., 2004. MicroRNAs: genomics, biogenesis, mechanism, andfunction. Cell 116, 281–297.

Billy, E., Brondani, V., Zhang, H., Muller, U., Filipowicz, W., 2001.Specific interference with gene expression induced by long, double-stranded RNA in mouse embryonal teratocarcinoma cell lines. Proc.Natl. Acad. Sci. USA 98, 14428–14433.

Catalanotto, C., Azzalin, G., Macino, G., Cogoni, C., 2002. Involvementof small RNAs and role of the qde genes in the gene silencing pathwayin Neurospora. Genes Dev. 16, 790–795.

Catalanotto, C., Pallotta, M., ReFalo, P., Sachs, M.S., Vayssie, L.,Macino, G., Cogoni, C., 2004. Redundancy of the two Dicer genes intransgene-induced posttranscriptional gene silencing in Neurospora

crassa. Mol. Cell. Biol. 24, 2536–2545.Catalanotto, C., Nolan, T., Cogoni, C., 2006. Homology effects in

Neurospora crassa. FEMS Microbiol. Lett. 254, 182–189.Davis, R.H., DeSerres, F.J., 1970. Genetic and microbiological research

techniques for Neurospora crassa. Methods Enzymol. 27A, 79–143.Ebbole, D., Sachs, M., 1990. A rapid and simple method for isolation of

Neurospora crassa homokaryons using microconidia. Fungal Genet.Newsl. 37, 17–18.

Freitag, M., Hickey, P.C., Raju, N.B., Selker, E.U., Read, N.D., 2004.GFP as a tool to analyze the organization, dynamics and function ofnuclei and microtubules in Neurospora crassa. Fungal Genet. Biol. 41,897–910.

Freitag, M., Selker, E.U., 2005. Expression and visualization of redfluorescent protein (RFP) in Neurospora crassa. Fungal Genet. Newsl.52, 14–17.

Galagan, J.E., Selker, E.U., 2004. RIP: the evolutionary cost of genomedefense. Trends Genet. 20, 417–423.

Kotaja, N., Sassone-Corsi, P., 2007. The chromatoid body: a germ-cell-specific RNA-processing centre. Nat. Rev. Mol. Cell Biol. 8, 85–90.

W.G. Alexander et al. / Fungal Genetics and Biology 45 (2008) 719–727 727

Lee, D.W., Pratt, R.J., McLaughlin, M., Aramayo, R., 2003. Anargonaute-like protein is required for meiotic silencing. Genetics 164,821–828.

Lim, A.K., Kai, T., 2007. Unique germ-line organelle, nuage, functions torepress selfish genetic elements in Drosophila melanogaster. Proc. Natl.Acad. Sci. USA 104, 6714–6719.

Logemann, J., Schell, J., Willmitzer, L., 1987. Improved for the isolationof RNA from plant tissues. Anal. Biochem. 163, 16–20.

Maiti, M., Lee, H.C., Liu, Y., 2007. QIP, a putative exonuclease,interacts with the Neurospora Argonaute protein and facilitatesconversion of duplex siRNA into single strands. Genes Dev. 21,590–600.

Makeyev, E.V., Bamford, D.H., 2002. Cellular RNA-dependent RNApolymerase involved in posttranscriptional gene silencing has twodistinct activity modes. Mol. Cell 10, 1417–1427.

Margolin, B.S., Freitag, M., Selker, E.U., 1997. Improved plasmids forgene targeting at the his-3 locus of Neurospora crassa by electropor-ation. Fungal Genet. Newsl. 44, 34–36.

McCluskey, K., 2003. The fungal genetics stock center: from molds tomolecules. Adv. Appl. Microbiol. 52, 245–262.

Mello, C.C., Conte Jr., D., 2004. Revealing the world of RNAinterference. Nature 431, 338–342.

Pane, A., Wehr, K., Schupbach, T., 2007. zucchini and squash encode twoputative nucleases required for rasiRNA production in the Drosophila

germline. Dev. Cell 12, 851–862.Papp, I., Mette, M.F., Aufsatz, W., Daxinger, L., Schauer, S.E., Ray, A.,

van der Winden, J., Matzke, M., Matzke, A.J.M., 2003. Evidence for

nuclear processing of plant micro RNA and short interfering RNAprecursors. Plant Physiol. 132, 1382–1390.

Perkins, D.D., Radford, A., Sachs, M.S., 2001. The Neurosporacompendium: chromosomal loci. Academic Press, San Diego.

Raju, N.B., 1980. Meiosis and ascospore genesis in Neurospora. Eur. J.Cell Biol. 23, 208–223.

Raju, N.B., Metzenberg, R.L., Shiu, P.K., 2007. Neurospora spore killersSk-2 and Sk-3 suppress meiotic silencing by unpaired DNA. Genetics176, 43–52.

Sambrook, J., Russell, D.W., 2001. Molecular Cloning: A LaboratoryManual, third ed. Cold Spring Harbor Laboratory, Cold SpringHarbor, NY.

Shiu, P.K.T., Glass, N.L., 1999. Molecular characterization of tol, amediator of mating-type-associated vegetative incompatibility inNeurospora crassa. Genetics 151, 545–555.

Shiu, P.K.T., Metzenberg, R.L., 2002. Meiotic silencing by unpaired DNA:properties, regulation, and suppression. Genetics 161, 1483–1495.

Shiu, P.K.T., Raju, N.B., Zickler, D., Metzenberg, R.L., 2001. Meioticsilencing by unpaired DNA. Cell 107, 905–916.

Shiu, P.K.T., Zickler, D., Raju, N.B., Ruprich-Robert, G., Metzenberg,R.L., 2006. SAD-2 is required for Meiotic Silencing by Unpaired DNAand perinuclear localization of SAD-1 RNA-directed RNA polymer-ase. Proc. Natl. Acad. Sci. USA 103, 2243–2248.

Vogel, H.J., 1964. Distribution of lysine pathways among fungi: evolu-tionary implications. Am. Nat. 98, 435–446.

Westergaard, M., Mitchell, H.K., 1947. Neurospora V. A syntheticmedium favoring sexual reproduction. Am. J. Bot. 34, 573–577.