Characterization of Hypovirus-Derived Small RNAs Generated in the Chestnut Blight Fungus by an Inducible DCL-2-Dependent Pathway

Abstract

The disruption of one of two dicer genes, dcl-2, of the chestnut blight fungus Cryphonectria parasitica was recently shown to increase susceptibility to mycovirus infection (G. C. Segers, X. Zhang, F. Deng, Q. Sun, and D. L. Nuss, Proc. Natl. Acad. Sci. USA 104:12902-12906, 2007). We now report the accumulation of virus-derived small RNAs (vsRNAs) in hypovirus CHV1-EP713-infected wild-type and dicer gene dcl-1 mutant C. parasitica strains but not in hypovirus-infected dcl-2 mutant and dcl-1 dcl-2 double-mutant strains. The CHV1-EP713 vsRNAs were produced from both the positive and negative viral RNA strands at a ratio of 3:2 in a nonrandom distribution along the viral genome. We also show that C. parasitica responds to hypovirus and mycoreovirus infections with a significant increase (12- to 20-fold) in dcl-2 expression while the expression of dcl-1 is increased only modestly (2-fold). The expression of dcl-2 is further increased (∼35-fold) following infection with a hypovirus CHV1-EP713 mutant that lacks the p29 suppressor of RNA silencing. The combined results demonstrate the biogenesis of mycovirus-derived small RNAs in a fungal host through the action of a specific dicer gene, dcl-2. They also reveal that dcl-2 expression is significantly induced in response to mycovirus infection by a mechanism that appears to be repressed by the hypovirus-encoded p29 suppressor of RNA silencing.

RNA silencing, the RNA-mediated, sequence-specific suppression of gene expression, in animals, plants, and fungi has been described previously (21, 45, 51, 64). The basic elements of the RNA-silencing pathways are conserved across kingdoms and involve the processing of structured or double-stranded RNA (dsRNA) into short duplex RNAs of 21 to 24 nucleotides (nt) by RNase III-like endonucleases termed dicers. With the aid of the Argonaute family of proteins, these short RNA duplexes are then incorporated into RNA-induced silencing complexes (RISC) that guide the sequence-specific cleavage or translational repression of homologous RNA molecules or transcriptional silencing through the modification of homologous DNA (reviewed in references 63 and 68).

RNA silencing serves a crucial role as an antiviral defense mechanism in plants (reviewed in references 16, 17, 37, and 65). Multiple mechanisms appear to be available for the generation of virus-specific small RNAs (vsRNAs) that guide the subsequent degradation of cognate viral RNA in plants. vsRNAs are derived primarily from highly structured regions of viral single-stranded RNA in the cases of Cymbidium ringspot virus (43), Turnip crinkle virus (27), Potato virus X (43), and Tobacco mosaic virus (43). In contrast, vsRNAs are derived equally from both strands in the cases of Turnip mosaic virus (27) and Cucumber yellows virus (67), indicating the processing of complementary viral dsRNAs. vsRNAs are also generated from branched stem-loop regions of 35S leader RNA produced by the plant DNA virus Cauliflower mosaic virus (42) and from structured regions of viroid RNA (29). The clarification of the role of specific plant dicers in antiviral defense has been complicated by the existence of multiple dicers, e.g., four dicers in Arabidopsis species (2, 3). While all four plant dicers have been implicated in the production of vsRNAs (4, 42), there is increasing evidence to indicate that dicer 4 is primarily responsible for the biogenesis of vsRNAs in the plant antiviral defense pathway (4, 5, 15, 22).

The importance of RNA silencing for antiviral defense in plants is reflected in the fact that most plant viruses encode suppressors of RNA silencing (reviewed in reference 37). The majority of these suppressors bind to dsRNA and are thought to interfere with vsRNA-RISC assembly (35). The targeting of silencing pathway components such as dicer (15) and an Argonaute protein (70) show that multiple steps in the RNA-silencing process are subject to virus-mediated suppression.

Parallels to the interactions between viruses and RNA silencing in plants have also been observed in invertebrates. Several members of the nodavirus family that infect both insects and vertebrates, e.g., Flock house virus and Nodamura virus, encode suppressors of RNA silencing (20, 30, 38, 59), and the disruption of the RNA-silencing pathways in adult Drosophila flies or mosquitoes has resulted in increased susceptibility to infection with a number of different viruses (23, 33, 66).

The role of RNA silencing in antiviral defense in vertebrates is less obvious. It has been suggested previously that the development of the interferon response to virus infection in vertebrates has effectively replaced the need for RNA silencing for protection against virus infection (14). However, there are clear indications of interactions between vertebrate viruses and the RNA-silencing pathway (reviewed in references 14, 37, 39, and 53). Several mammalian DNA viruses encode transcripts containing hairpins that are processed into small RNAs, microRNAs (miRNAs), that influence host gene expression (6, 49, 50, 52, 60). Cellular miRNAs interact with several mammalian viruses in ways that influence viral replication (31, 36). Mammalian viruses encode suppressors of RNA silencing (reviewed in reference 39). A direct role for dicer in protecting against influenza A virus infection was recently reported (41).

Virus infections are widespread in the kingdom Fungi. Evidence that RNA silencing functions as an antiviral defense mechanism in fungi has recently been reported for the chestnut blight fungus Cryphonectria parasitica (55). This fungus has been shown to support the replication of members of five virus families: Hypoviridae, Reoviridae, Narnaviridae, Partitiviridae, and Chrysoviridae (reviewed in reference 26). The positive-strand RNA virus members of the family Hypoviridae have been studied extensively (46) because they attenuate the virulence of their fungal host, causing hypovirulence, and because reverse-genetics systems for several hypoviruses have been developed. The C. parasitica-infecting dsRNA reoviruses also cause hypovirulence but confer a distinct set of associated phenotypic changes in the fungal host (26). The papain-like protease p29, encoded by hypovirus CHV1-EP713, was recently shown to augment reovirus MyRV1-9B21 accumulation (61) and to suppress RNA silencing in the natural fungal host and in a heterologous plant system (54). The disruption of one of two C. parasitica dicer genes, dcl-2, resulted in increased susceptibility to infections with the hypovirus CHV1-EP713 and the reovirus MyRV1-9B21 (55). Here, we present evidence that hypovirus-specific small RNAs are derived from both positive and negative RNA strands by a DCL-2-dependent mechanism. Moreover, we show that mycovirus infection triggers a large increase in dcl-2 expression, which is further enhanced following infection with a hypovirus mutant lacking the p29 suppressor of RNA silencing.

MATERIALS AND METHODS

Fungal strains and growth conditions.

C. parasitica strains were maintained on potato dextrose agar (PDA; Difco, Detroit, MI) at 22 to 24°C with a 12-h light-dark cycle at 1,300 to 1,600 lx. C. parasitica cultures used for RNA preparations were grown for 7 days under similar conditions on PDA overlaid with a cellophane membrane. The origins and sources of wild-type C. parasitica strain EP155 (ATCC 38755) and the isogenic hypovirus CHV1-EP713-containing strain EP713 (ATCC 52571) have been reviewed by Chen and Nuss (12). Reovirus MyRV1-Cp9B21-infected strain EP155 was provided by Bradley Hillman (Rutgers, The State University of New Jersey, New Brunswick). Null mutants of C. parasitica strain EP155 containing disruptions of dcl-1 (Δdcl-1), dcl-2 (Δdcl-2), and both dcl-1 and dcl-2 (Δdcl-1 Δdcl-1) and the dcl-2 functionally complemented strain Δdcl-2/Com have been described by Segers et al. (55).

RNA isolation and small-RNA cloning.

Total RNAs from fungal cultures were prepared as previously described (62). Low-molecular-weight RNAs were further enriched from the total RNAs as described by Catalanotto et al. (8). Small RNAs in the range of 16 to 28 nt were cloned and sequenced by following the protocol described by Chappell et al. (10) with slight modifications. The 5′ and 3′ adaptors described by Elbashir et al. (19) were used in place of those described by Chappell et al. (10). The 5′ PCR primer contained an NheI site, while the 3′ PCR primer contained an XhoI site. The T4 DNA ligase concatamerization step was eliminated, and the NheI/XhoI-digested PCR DNA products were cloned into a pGEM-T Easy vector modified to contain an NheI/XhoI cloning site. Sequence information for cloned hypovirus-derived small RNAs in the 18- to 24-nt range is listed in Table S1 in the supplemental material.

Northern blot analysis.

Northern blot analyses of small RNAs were performed as described by Hamilton and Baulcombe (24) with some modifications. Samples consisting of 20 μg of total RNA or 30 μg of low-molecular-weight RNA were subjected to electrophoresis in 0.5× TBE buffer (1× TBE is 89 mM Tris [pH 8.3], 89 mM boric acid, 2.5 mM EDTA) through a 10% denaturing polyacrylamide gel (7 M urea) along with a [γ-³²P]ATP-labeled 10-bp DNA ladder according to the instructions of the manufacturer (Invitrogen, CA) and then electrotransferred in 0.5× TBE onto GeneScreen Plus filters (PerkinElmer, MA). The membranes were fixed by UV light cross-linking. The sense and antisense hybridization probes were generated from two PCR fragments spanning nt 10845 to 11799 of the 12.7-kb CHV-EP713 viral genome RNA (57); each of these fragments contained the T7 promoter sequence at the 5′ or 3′ end. Single-strand RNA probes were transcribed with [α-³²P]UTP and the T7 MAXIscript in vitro transcription kit (Ambion, CA). The labeled RNA probes were then digested with DNase I to remove the DNA template. To hydrolyze the probes to an average size of 50 nt, 300 μl of hydrolyzation buffer (80 mM sodium bicarbonate, 120 mM sodium carbonate) was added to the probes, the reaction mixture was incubated at 60°C for 3 h, and a volume of 3 M sodium acetate, pH 5.0, equal to the volume of the reaction mixture was added to stop the reaction. Membranes were prehybridized at 55°C in Church buffer (0.5 M sodium phosphate [pH 7.2], 7% sodium dodecyl sulfate [SDS], and 1 mM EDTA [pH 7.0]) for at least 1 h and then hybridized in the same buffer at 55°C with the labeled probes overnight. Hybridized membranes were washed twice with 2× SSC (0.15 M NaCl, 0.015 M sodium citrate, pH 7.0)-0.1% SDS for 15 min each time and then twice with 0.1% SSC-0.1% SDS for 15 min each time. The washed membranes were exposed to storage phosphor screens (Kodak, NY). The screens were scanned and analyzed by a Storm 840 scanner and ImageQuant 5.1 software (Molecular Dynamics, CA).

Semiquantitative RT-PCR.

Semiquantitative reverse transcription PCR (RT-PCR) was performed on an Applied Biosystems 7300 fast real-time PCR system using the protocol described by Suzuki and Nuss (62). Calculations of transcript accumulation values were performed with the comparative threshold cycle method using 18S rRNA values to normalize for variations in template concentrations.

RESULTS

Virus-derived small RNAs accumulate in hypovirus-infected wild-type and dcl-1 mutant strains but not in dcl-2 mutant strains.

While the generation of small interfering RNAs (siRNAs), a hallmark of RNA silencing in plants and animals (68), has been reported to be associated with RNA silencing in several filamentous fungi, e.g., Neurospora crassa (8), Magnaporthe oryzae (32), and Aspergillus nidulans (25), the accumulation of virus-derived vsRNA in mycovirus-infected fungal hosts has not been reported previously. Northern analysis was performed on total RNA isolated from virus-free wild-type strain EP155 and the isogenic strain EP713, which harbors hypovirus CHV1-EP713, to assay for the accumulation of hypovirus-specific small RNAs. Positive- and negative-strand-specific probes corresponding to a 955-nt region of the 12.7-kb CHV1-EP713 genome RNA (CHV1-EP713 map positions 10845 to 11799 [56]) both hybridized to a distinct band migrating around 21 nt in the RNA samples prepared from hypovirus-infected strain EP713 but not in the samples prepared from the uninfected wild-type strain EP155 (Fig. 1A and B). These results are consistent with the accumulation of vsRNAs derived from both viral RNA strands in hypovirus-infected C. parasitica.

FIG. 1.

(A and B) Northern blot analysis showing the accumulation of virus-derived small RNAs in hypovirus CHV1-EP713-infected C. parasitica. Lanes 1 and 2 received 20 μg of total RNA extracted from uninfected and CHV1-EP713-infected cells of C. parasitica strain EP155, respectively. Lane 3 received loading buffer only, and lane 4 received 10 ng of a 124-bp oligonucleotide homologous to the hybridization probe. The migration positions of the 20- and 100-nt DNA size markers (lanes M) are shown on the left. The RNA blots were hybridized with a hydrolyzed riboprobe corresponding to an ∼1-kb region (map positions 10,845 to 11,799) of the 12.7-kb CHV1-EP713 genome RNA. Membrane A was hybridized with a positive-strand probe, while membrane B was hybridized with a negative-strand probe. (C) Northern blot analysis of vsRNA accumulation in CHV1-EP713-infected dicer gene mutants. For this analysis, samples were enriched with low-molecular-weight RNA to increase the signal-to-noise ratios for the vsRNAs. Thirty micrograms of RNA was loaded for each sample. The membrane was hybridized with a plus-strand probe. Lane 1 contained the 10-bp DNA ladder (Invitrogen). Lane 2 received RNA isolated from uninfected C. parasitica strain EP155. RNA from CHV1-EP713-infected strain EP155 was loaded into lane 3, while RNA from CHV1-EP713-infected dicer mutant strains was loaded into lanes 4 through 7. Lane 4, infected Δdcl-1 mutant; lane 5, infected Δdcl-2 mutant; lane 6, infected Δdcl-1 Δdcl-2 double mutant; lane 7, infected complemented Δdcl-2 mutant Δdcl-2/Com. The vsRNA bands were broader than those observed for membranes in panels A and B because the tracking dye was omitted.

We recently reported that the disruption of the C. parasitica dicer gene dcl-2 results in greatly increased susceptibility to mycovirus infection but that the disruption of the second dicer gene, dcl-1, has no effect on the response of C. parasitica to virus infection (55). A Northern blot analysis of low-molecular-weight RNA isolated from hypovirus-infected wild-type and dicer gene mutant strains (Fig. 1C) showed that the hypovirus vsRNAs accumulated in strain EP713, the dcl-1 gene mutant strain Δdcl-1, and the functionally complemented dcl-2 gene mutant strain Δdcl-2/Com. In contrast, hypovirus vsRNAs were not detected in the infected dcl-2 mutant Δdcl-2 or the Δdcl-1 Δdcl-2 double dicer mutant. These results extend our previous findings that dcl-2 is required for functional RNA silencing and that dcl-1 is dispensable (55) to include the requirement for dcl-2 but not dcl-1 for vsRNA biogenesis.

Hypovirus vsRNAs are derived from both viral RNA strands with a nonrandom distribution along the length of the viral genome.

Small RNAs isolated from hypovirus CHV1-EP713-infected mycelia were cloned and sequenced to further investigate the sizes and origins of the hypovirus-derived vsRNAs detected by Northern analysis. The sequencing of 233 cloned small RNAs revealed 171 clones (73%) that were homologous to the CHV1-EP713 hypovirus RNA sequence. The majority of the vsRNAs (70%) were found to be within the 20- to 22-nt range, with a clear peak at 21 nt (Fig. 2). In contrast, the 62 cloned RNAs that were not homologous to the hypovirus sequence exhibited fairly even distribution over the observed 16- to 28-nt range. These small RNAs were derived primarily from rRNA and are likely to be products of RNA degradation pathways.

FIG. 2.

Size distribution of sequenced small-RNA cDNA clones in the 16- to 28-nt size range. The sizes of the small RNAs are indicated below the columns, and the numbers on the left and those above each column indicate the number of small RNAs of a particular size. Small RNAs with identity to the CHV1-EP713 sequence are indicated by dark gray bars, while RNAs with no homology to the viral sequence are indicated by light gray bars.

Virus-derived small RNAs of 18 to 24 nt were mapped to the 12.7-kb CHV1-EP713 viral RNA according to their polarities and patterns of alignment with the viral RNA sequence. The ratio of sequences derived from the plus strand to those derived from the minus strand was 3:2. As indicated in Fig. 3, the vsRNA cDNA clones were distributed along the entire plus and minus viral RNA strands, with uneven distribution in some areas. For example, only positive-polarity vsRNA cDNA clones were found in association with the terminal region containing the 5′-terminal noncoding region of the positive strand that extends to map position 495. Very few vsRNA clones, all of positive polarity, corresponded to the region extending from map positions 7,500 to 11,000, which encodes the viral polymerase and helicase domains (57). Several vsRNAs were cloned multiple times. These included five clones within map positions 12,587 to 12,607, corresponding to the terminus containing the 3′ noncoding region of the coding strand. Four independent clones were obtained for map positions 451 to 471, corresponding to the 5′ noncoding region, and for positions 3896 to 3922, corresponding to the region just 3′ of the p48 protease coding domain in the second viral open reading frame, open reading frame B. A total of five vsRNAs were represented by three independent cDNA clones each, and 23 vsRNAs were cloned twice.

FIG. 3.

Origins and polarities of vsRNAs in the 18- to 24-nt range. The positions along the 12.7-kb CHV1-EP713 genome RNA from which vsRNAs originated are indicated. Virus-derived vsRNAs originating from the positive strand are indicated above the line representing the viral RNA genome, and vsRNAs originating from the negative strand are indicated below the line.

Unexpectedly, only 65 of the 171 vsRNA sequences in the 18- to 24-nt range were a perfect match to the viral RNA sequence. A substantial number (87) contained 1-nt (61) or 2-nt (26) mismatches at the terminus of the vsRNA sequence. The majority of these vsRNA clones contained an A (42) or AA (15) at the 3′ end, but 1- to 2-nt mismatches involving the other nucleotides were found at both the 3′ and 5′ ends for a minority of the cloned vsRNA sequences (Table 1). Only nine vsRNAs were found to have an internal single-nucleotide mismatch.

TABLE 1.

vsRNAs containing mismatches relative to the CHV1-EP713 sequence

Mismatch type (total no. of vsRNAs)	5′-terminal nucleotide(s) in mismatch sequence	No. of vsRNAs with indicated nucleotide(s)	3′-terminal nucleotide(s) in mismatch sequence	No. of vsRNAs with indicated nucleotide(s)
One-nucleotide mismatch (61)	A	2	A	42
	U	5	U	1
	C	0	C	4
	G	0	G	7
Two-nucleotide mismatch (26)	UU	3	AA	15
	CG	1	GA	2
	UA	1	GU	1
	AG	1	GC	1
	CU	1
Three-nucleotide mismatch (4)	None		AAU	1
			CAA	1
Four-nucleotide mismatch (2)	None		ACGC	1
			UAAA	1
Five-nucleotide mismatch (1)	UUACA	1	None
	UAU	1	A	2
Mismatch at both ends (3)	UUA	1	CA	1
	AG	1
Single internal mismatches with no terminal mismatch (6)	None		None
Single internal mismatch plus a terminal mismatch (3)	None		AA	3

The accumulation of dcl-2 gene transcripts is dramatically increased upon viral infections.

Given the apparent requirement for dcl-2 and the dispensability of dcl-1 for the modulation of mycovirus infection (55) and vsRNA biogenesis (Fig. 1C), it was of interest to examine the influence of mycovirus infection on dicer gene expression. Consequently, dcl-1 and dcl-2 transcript levels were monitored by real-time semiquantitative RT-PCR analysis following infection with hypovirus CHV1-EP713 and mycoreovirus MyRV1-Cp9B21. As indicated in Fig. 4, dcl-1 transcript levels increased 1.4- to 2.2-fold following infection with either mycovirus. In contrast, dcl-2 transcript levels increased in excess of 10- and 15-fold following CHV1-EP713 and MyRV1-Cp9B21 infection, respectively. Interestingly, parallel infection with a CHV1-EP713 mutant virus that lacked the p29 suppressor of RNA silencing invariably resulted in a 20- to 30-fold additional increase in dcl-2 transcript levels relative to that observed in CHV1-EP713-infected strains (34-fold in Fig. 4).

FIG. 4.

Semiquantitative RT-PCR analysis of dicer gene transcript levels relative to levels of 18S rRNA. The values on the y axis were normalized to the dicer transcript levels in strain EP155 (set at 1), with the standard deviations, based on three independent measurements of two independent RNA preparations, indicated by the error bars. Open bars, dcl-1 transcript accumulation; closed bars, dcl-2 transcript accumulation. The identities of the infecting mycoviruses are indicated below the columns: hypovirus CHV1-EP713, reovirus MyRV1-Cp9B21, and Δp29, a CHV1-EP713 mutant isolate that lacks the p29 suppressor of RNA silencing.

DISCUSSION

RNA silencing has only recently been shown to serve as an antiviral defense mechanism in fungi (55). The results of this study demonstrate the production of virus-derived small RNAs in a mycovirus-infected fungus and identify a specific fungal dicer as being responsible for vsRNA biogenesis. They further reveal that mycovirus infection activates a fungal RNA-silencing pathway by causing a large increase in the expression of the dicer gene responsible for vsRNA biogenesis through a mechanism that is sensitive to a mycovirus-encoded suppressor of RNA silencing.

The sequence profiles and strand polarities observed for hypovirus vsRNAs are not predicted by the prevailing view of vsRNA biogenesis in plants. Results of studies with several plant positive-strand RNA viruses (27, 43), viroids (29), and the double-stranded DNA Cauliflower mosaic virus (42) suggest that vsRNAs originate primarily from highly structured regions of viral single-stranded RNA rather than from double-stranded viral RNA replication intermediates. This asymmetry was not observed for hypovirus vsRNAs, with over 40% of these vsRNAs derived from the negative strand (Fig. 4). Although hypoviruses are clearly positive-strand viruses, synthetic copies of the coding strand are infectious (11); these fungal viruses are distinguished from plant positive-strand viruses by the absence of a coat protein (57). As a result, hypoviruses do not form discrete virus particles. Moreover, the ratio of positive- to negative-strand RNA is lower for hypoviruses (10 to 1) (discussed in reference 62), resulting in a significant portion of the viral RNA's being observed in the form of large dsRNAs. This trait may contribute to the more symmetric production of vsRNAs. Interestingly, Ho et al. (27) have reported a similar pattern of 60% positive and 40% negative polarity for vsRNAs derived from Turnip mosaic virus, a potyvirus phylogenetically related to the hypoviruses (34). Since the characterization of vsRNAs from other mycoviruses has not yet been reported, it is currently unclear whether the polarity ratio observed for hypovirus RNAs is a function of the mechanisms that operate in fungi for generating vsRNAs or is related to some aspects of the replication cycle used by hypoviruses and closely related plant viruses.

The clustered distribution of the positive-polarity vsRNAs, particularly evident for the highly structured (57) 5′ noncoding leader region (Fig. 3), is consistent with the processing of structured regions of the coding strand. Moreover, with one exception, the vsRNAs that were cloned three times or more had positive polarity. However, the vsRNAs with negative polarity were also unevenly distributed, which is not the result expected if the vsRNAs were generated from long viral dsRNA replicative intermediates. The possibility that vsRNAs with negative polarity are derived from structured regions of the negative RNA strand requires further investigation.

The paucity of vsRNAs derived from the region spanning map positions 7,500 to 11,000 is intriguing. This region, which encodes the viral polymerase and helicase domains, may be inaccessible to the RNA-silencing machinery as a result of local secondary-structure constraints, as suggested for regions of certain cellular (47) and viroid (29) RNAs, or as a result of the binding of protective virus-encoded RNA binding proteins. The propensity of CHV1-EP713 to generate internally deleted defective RNAs may provide an alternative or additional contributing explanation. The prominent 8- to 10-kb defective dsRNAs commonly found in CHV1-EP173-infected strains were shown to retain at least 3.5 kb of each terminus (56). High-resolution-mapping studies of the deletion boundaries, now in progress, indicate that the region extending from map positions 7,500 to 11,000 is preferentially deleted in the defective RNA population (X. Zhang and D. L. Nuss, unpublished observations), thereby providing lower levels of the substrate for vsRNA biogenesis from this region.

The observation that a relatively high proportion (over 50%) of hypovirus vsRNAs did not show a perfect match with the progenitor viral RNA sequence due to the presence of a single- or double-nucleotide terminal mismatch was also unexpected. The facts that the mismatches were found predominantly at the 3′ terminus and overwhelmingly consisted of one or two adenosine residues argue against the introduction of errors during the reverse transcription and PCR amplification stages of the cloning process, which should introduce mismatches in a random manner. The cloning and sequencing of cDNA generated from purified CHV1-EP713 dsRNA revealed no evidence of significant levels of sequence heterogeneity in the virus RNA population (data not shown). One possibility to consider in future studies is that the vsRNAs showing a perfect match with the viral genome sequence are generated directly by dicer cleavage while the vsRNAs with the terminal mismatches are generated by transitive or secondary silencing (1, 42), during which they are subjected to postdicer modifications. The addition of non-template-encoded uridine or adenosine residues to the 3′ terminus of RISC-mediated cleavage sites in plants (48, 58) and Chlamydomonas reinhardtii (28) has been reported previously, and extra nucleotides were reported to be added to the 3′ end of miRNAs and siRNAs in Arabidopsis plants carrying a mutation in the hen1 methyltransferase gene that is responsible for the 3′-end methylation of these small RNAs (40). In this regard, the hypovirus p29-related potyvirus suppressor of RNA silencing, helper component-protease, has been reported to mimic the hen1 mutation for virus-specific but not for endogenous siRNAs (18).

Additional cloning and sequencing of other mycovirus vsRNAs will be required to determine whether fungi use a mechanism that differs from that used predominantly for single-stranded RNA plant viruses or whether multiple mechanisms exist for generating mycovirus vsRNAs. These studies will be facilitated by the recent sequencing and characterization of C. parasitica viruses with encapsulated dsRNA genomes (Reoviridae and Chrysoviridae viruses) and unencapsidated single-stranded RNA genomes (Narnaviridae viruses) (reviewed in reference 26).

The loss of vsRNA production in dcl-2 deletion strains (Fig. 1) is consistent with the findings in our previous report that the disruption of dcl-2 results in increased mycovirus-mediated symptoms and viral RNA levels (55). Although a role for other fungal dicers in antiviral defense has not yet been established, the C. parasitica dcl-1 and dcl-2 homologues in the closely related filamentous fungus N. crassa have been shown to exhibit a level of redundancy in RNA silencing (8). We did not see any effect of dcl-1 deletion on the response to virus infection (55), and Northern analysis indicated little if any vsRNA production in dcl-2 deletion mutant strains (Fig. 2). The involvement of a single dicer gene in vsRNA biogenesis provides opportunities for studying the antiviral role of RNA silencing in the absence of contributions from multiple dicers, as found in plant systems.

miRNAs that are produced in plants and animals from genome-encoded RNA hairpins and are involved in developmental and cellular regulation (reviewed in references 7 and 69) have not been detected in fungi (9, 44). Thus, the role of RNA silencing in fungi has been assumed to be primarily defense against invasion by foreign nucleic acids. The low level of dcl-2 transcripts observed in uninfected C. parasitica cells (data not shown) and the large increase in dcl-2 gene expression in response to infection with two unrelated mycoviruses (Fig. 4) are consistent with this view. In this regard, Choudhary et al. (13) recently reported that the production of hairpin dsRNA in N. crassa results in the induced expression of the Argonaute gene qde-2 and the dicer gene dcl-2, the homologue of the C. parasitica dcl-2 gene. A similar response in dicer gene expression following virus infection in plants or animals has not yet been reported, possibly due to complications posed by multiple dicers in plants and the prominence of the interferon response in animals.

The superinduction of dcl-2 expression following infection with the CHV1-EP713 mutant virus that lacks the p29 suppressor of RNA silencing was unexpected. Virus-encoded RNA-silencing suppressors interfere with RNA silencing by sequestering siRNAs, one of the more common mechanisms; by preventing siRNA formation by binding structured RNA substrates; or by targeting components of the RNA-silencing machinery (reviewed in reference 37). The hypovirus p29 suppressor has been shown to interfere with hairpin silencing in fungi as well as virus-induced and agroinfiltration-induced RNA silencing in plants (54). Thus, the novel suppression of the transcriptional induction of the specific dicer responsible for vsRNA biogenesis reported here may complement more-conventional mechanisms used by this suppressor to counter fungal antiviral defense.