INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Successful infection of plants by viral pathogens requires a series of compatible interactions between the host plant and the viral invader. Virus-host interactions such as replication of the virus genome at the single-cell level, cell-to-cell movement through plasmodesmata, and long-distance translocation in vascular tissue are required for the virus to successfully invade the host plant and cause disease (6, 7). A resistant or immune response can be caused by a passive mechanism when essential host components are missing (13). This kind of incompatible reaction could be due to the inability of a virus to associate with host-specific proteins for replication (17, 23) or movement (30, 37). Plants have also developed the ability to recognize and resist many invading pathogens by inducing a set of defense responses. These resistant responses, such as hypersensitive cell death (12), systemic acquired resistance (36, 44), or posttranscriptional gene silencing (PTGS) (33), can prevent the virus infection.

A hypersensitive response (HR) is characterized by the development of a necrotic lesion at the site of infection which results in confinement of the invader to this area. It is induced by a direct or indirect interaction between the product of the host plant R gene and the matching avirulence product of the corresponding pathogen. Different virus proteins, including RNA-dependent RNA polymerase (15, 16, 24, 28, 32), movement protein (29, 48, 49), and the coat protein (CP) (2, 3, 14, 46), have been identified as avirulence determinants. In addition to proteins, polysaccharides and other low-molecular-weight compounds have also been shown to act as specific elicitors in different nonviral host-pathogen systems (22).

Avirulence determinants can be responsible for host specificity of viruses, and they are often identified by genomic exchanges between related viruses having different host ranges and symptoms. The availability of infectious cDNA clones of different well-characterized tombusviruses having altered host ranges offers an excellent experimental system to recognize host specificity determinants.

The genome of a tombusvirus is a linear, single-stranded monopartite RNA molecule of mRNA polarity. The 4.7-kb-long genomic RNA acts as mRNA for the translation of a 33-kDa protein (p33, from open reading frame1 [ORF1]) and a readthrough product of a 92-kDa protein (p92, from ORF2). It was previously demonstrated that both of them are required for replication (11, 25, 31, 43). The 41-kDa CP is synthesized from the 2.1-kb subgenomic RNA 1 (sg1) (35), and two small nested ORFs (ORF4 and ORF5) coding for a 22-kDa (p22) and a 19-kDa (p19) protein are translated from the 0.9-kb subgenomic RNA 2 (sg2) (35). The p22 protein is required for cell-to-cell movement (11, 34, 40), and it also has a role in symptom development on certain hosts (41). p19 plays a crucial role in necrotic symptoms (4, 11, 34, 41) and participates in virus spread in a host-specific manner (42). Furthermore, p19 is able to suppress the virus-induced PTGS in Nicotiana benthamiana (47).

The aim of the present study was to determine virus-encoded host specificity determinants of tombusviruses. We used chimeras of previously described biologically active cDNA clones of cymbidium ringspot virus (CymRSV) (11) and tomato bushy stunt virus pepper strain (TBSV-P) (19). Although the genomic RNAs of the two viruses have very similar primary structures and genome organizations, there are significant differences in their host ranges. TBSV-P readily infects Capsicum annum and Datura stramononium plants systemically, while CymRSV is not able to infect these plants. Here we report that a stretch of 860 nucleotides (nt) of the RNA sequence from the CymRSV CP coding region is responsible for the HR-like resistance of D. stramonium plants against CymRSV. Furthermore, evidence is provided that the elicitor of the HR-like response resides in this 860-nt-long RNA sequence itself and not in the encoded protein. To our knowledge, this is the first report which shows that an RNA itself can act as an avirulence determinant.

View larger version (65K): [in this window] [in a new window]   FIG. 1.   Symptoms and virus replication in C. annum and D. stramonium plants inoculated with CymRSV, TBSV-P, and CymRSV-TBSV-P chimeras. (A) Symptoms on D. stramonium plants 3 weeks after inoculation with CymRSV (a) and TBSV-P (b) and on C. annum plants 2 weeks after inoculation with CymRSV (c) and TBSV-P (d). (B) Press blot analysis of D. stramonium leaves inoculated with either mock inoculum (1) or CymRSV (2 and 3). Inoculated leaves were harvested either at 1 h postinfection (2) or at 3 dpi (1 and 3). Filters were hybridized with radiolabeled, CymRSV-specific probes. The photographs were prepared by overlaying the filters with the appropriate autoradiograms showing the sites of infected cells. Panel 3a shows the autoradiogram of filter 3 without the filter in order to make faint signals visible. The green line indicates the edge of the inoculated leaf. (C) Symptoms induced on the inoculated leaves of D. stramonium by mock inoculum (a), CymRSV (b), TBSV-P (c), L6 (d), L841 (e), L12 (f), G11AbglII (g), G11Aval (h), and TBSV-AvaBg (i). Insets in panels b, h, and i are enlarged portions of the leaves shown.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmid constructs. DNA manipulations and cloning were carried out using standard molecular biology techniques (38) unless otherwise described.

DNA sequencing. Plasmid DNAs were purified as previously described (18). DNA was sequenced by the dideoxy chain termination method (39) using T7 DNA polymerase (Sequenase; U.S. Biochemicals).

In vitro transcription and inoculation into plants. Plasmids G11, TBSV-P, and their derivatives were linearized by SmaI, and the RNA transcriptions were performed as previously described (11). For inoculation, the appropriate synthetic transcripts were mixed, and then the RNA was diluted with an equal volume of inoculation buffer (20) and applied to three leaves (15 µl/leaf) of N. benthamiana plants with a sterile glass spatula. D. stramonium plants were inoculated with the plant sap of infected leaves of N. benthamiana.

RNA extraction and Northern blot analysis. Total RNA from leaf tissue was isolated essentially according to the method of White and Kaper (50), with some modifications (11). Northern blot analysis was performed as follows. RNA samples were denatured with formamide and formaldehyde, electrophoresed in 1.5% agarose gels, and transferred to nylon membranes (38). The following DNA fragments were labeled with [-³²P]dCTP by random priming following the manufacturer's instructions (Prime-a-Gene kit; Promega): the PflMI-SmaI fragment of TBSV-P, corresponding to nt 3850 to 4776 of the TBSV-P genome, and the PflMI-SmaI fragment of G11, corresponding to nt 3806 to 4733 of the CymRSV genome. The two fragments were labeled separately and were mixed prior to use according to their specific activity.

In situ hybridization. The press blot of inoculated leaves was carried out according to the method of Szilassy et al. (45). In brief, inoculated leaves were removed, and the lower side was abraded with Celite using a brush. The leaves were rinsed in water and were placed onto a Hybond N membrane (Amersham) with the abraded side down. The membrane was floated on W5 solution (125 mM CaCl₂ · 2H₂O, 154 mM NaCl, 5.5 mM glucose, 10.7 mM KCl) containing 0.7% cellulase (Onozuka R-10) and 0.25% macerozyme (R-10) and incubated in the dark overnight at room temperature. The membrane, together with the leaves, was removed and roller pressed between Whatman 3MM papers. Leaf remains were removed by rinsing the membrane with 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), and the membrane was dried and UV cross-linked according to the manufacturer's instructions. The presence of viral RNA was detected with randomly primed, ³²P-labeled probes synthesized from the nt 3806-to-4733 PflMI-SmaI insert of CymRSV.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Symptoms and replication of CymRSV and TBSV-P on C. annum and D. stramonium plants. TBSV-P and CymRSV were used to infect C. annum and D. stramonium plants. TBSV-P-infected C. annum plants reacted with characteristic chlorotic lesions followed by the appearance of systemic symptoms, which consisted of a yellow and dark green mosaic pattern (Fig. 1A, panel d). The TBSV-P-infected D. stramonium plants displayed expanded necrotic lesions on the inoculated leaves (Fig. 1C, panel c) followed by systemic necrotic spots and twisting of the top leaves (Fig. 1A, panel b). However, CymRSV was not able to systemically infect either C. annum or D. stramonium plants (Fig. 1A, panels a and c), and the inoculated leaves of D. stramonium showed a typical HR-like necrotic response (very small, needle prick-sized necrotic lesions) at 3 days postinoculation (dpi) (Fig. 1C, panel b). This result suggested that the observed resistance against CymRSV infection is due to a very fast HR and that as a consequence of the HR-like necrotic local lesions the virus was not able to spread from the initially infected cells. To test this suggestion, press blot analysis was done on CymRSV-inoculated D. stramonium leaves showing HR-like necrotic lesions. Virus-specific RNAs could not be detected on leaves blotted at 1 hour postinoculation. In contrast, when leaves were blotted at 3 dpi, those containing HR-like necrotic local lesions reacted with the CymRSV-specific probe (Fig. 1B), indicating virus accumulation in the initially infected cells. Attempts to detect CymRSV RNAs in the total RNA extracts of these leaves at 3 and 7 dpi using Northern analysis were unsuccessful. The lack of detection was likely due to the fact that only a few cells could be infected with CymRSV (Fig. 1B). However, the same RNA extracts were able to induce infection on N. benthamiana plants. These results suggested that CymRSV was able to replicate in the primary infected cells but that the plant quickly confined the virus by an HR-like response.

The CP gene of tombusviruses determines the host specificity. To localize the viral genes responsible for the host specificity, hybrid viruses were constructed by exchanging the corresponding genomic regions between TBSV-P and CymRSV (Fig. 2A). The viability of hybrid viruses was tested by inoculating them on N. benthamiana. All chimeras were able to infect N. benthamiana plants systemically and caused similar symptoms and accumulated to the same level as the parental viruses did (Fig. 2B). These results confirmed that all chimeras retained basic competence for infection. In order to identify those viral factors which are responsible for host specificity, selective hosts (C. annum and D. stramonium) were inoculated with plant sap prepared from N. benthamiana plants infected with corresponding hybrids. Virus RNA accumulation and symptom development were analyzed on the inoculated leaves of C. annum and D. stramonium. The results obtained showed that replacement of the two 3' proximal nested genes (constructs L1 and L2) and the ORF1 (constructs L3 and L4) did not modify the symptoms of wild-type (wt) CymRSV and TBSV-P infections (Fig. 2A). Similarly, the exchange of sequence between StuI and NheI sites in ORF2 (constructs L9 and L10) did not alter the symptoms of wt virus infections. However, modifications which affected the CP-encoding sequence (ORF3) (constructs L5 and L6) resulted in a significant alteration in the symptoms of infected D. stramonium (Fig. 2A). Northern blot analysis revealed that only those constructs (L1, L4, L6, L8, and L9) that carry the CP gene and the last 143 nt of the replicase gene of TBSV-P were able to infect D. stramonium plants (Fig. 2A and C). Some variation in the accumulation of viral RNA was observed, but these differences were likely due to sampling effects and did not result in alterations in symptoms. Constructs (L2, L3, L5, L7, and L10) containing the corresponding region (the CP gene and the last 143 nt of the replicase gene) of CymRSV induced HR in D. stramonium (Fig. 2A and C). The symptoms of infectious constructs on D. stramonium (e.g., L6 in Fig. 1C, panel d) and C. annum (not shown) were similar to those of TBSV-P (Fig. 1C, panel c), causing expanding chlorotic and necrotic lesions. This observation may suggest that similar viral factors are involved in the host specificity determination on C. annum and D. stramonium, although C. annum did not exhibit HR-like necrotic lesions upon CymRSV infection. In this study C. annum was not used for further investigation. The obtained results suggested that the CP and/or the last 143 nt of the replicase gene of TBSV-P are required to infect D. stramonium and C. annum plants, and alternatively, that the CymRSV CP and/or the last 143 nt of the replicase gene represent an avirulence factor to these hosts.

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Viral symptoms and RNA accumulation in plants inoculated with CymRSV, TBSV-P, and CymRSV-TBSV-P chimeras. (A) Schematic representation and induced symptoms of the CymRSV-TBSV-P chimeras. The organizations of CymRSV and TBSV-P genomic RNAs are shown above with the ORFs (boxes) and the approximate molecular masses of the encoded proteins. The common restriction endonuclease sites used for constructing chimeras are indicated. The designations of the chimeras are on the left. Symptoms of parental and hybrid viruses are indicated at the right. Three plants were inoculated with each inoculum, and the experiment was repeated three times. Symbols: +, efficient virus infection with the typical symptoms detailed in the text; , no symptoms appeared for up to 4 weeks; HR, small necrotic local lesions which limited the virus spread; n.d., not determined. N. b., N. benthamiana; C. a., C. annum; D. s., D. stramonium. The accumulation of hybrid viral RNAs in N. benthamiana (B) and D. stramonium (C) plants is shown. Total RNAs were extracted at 6 to 8 dpi from systemically infected leaves of N. benthamiana and inoculated leaves of D. stramonium. Northern blots were hybridized with 32P-labeled DNA mixed probes of the 3'-terminal 1,000 nt of CymRSV and TBSV-P RNAs. G, genomic RNA; sg1, subgenomic RNA 1; sg2, subgenomic RNA 2.

CymRSV CP or the encoding RNA sequence acts as an avirulence factor eliciting an HR-like resistance response in D. stramonium. The inability of CymRSV to infect D. stramonium plants can be explained in two different ways: (i) the TBSV-P CP gene is required for successful infection or (ii) the CymRSV CP gene acts as an avirulence factor eliciting an HR-like resistance response. To distinguish between these two possibilities, we prepared TBSV-PCP, L1CP, L4CP, L6CP, L8CP, and L9CP deletion mutants (Fig. 3A) and tested them on both N. benthamiana and D. stramonium plants. As was expected from previous observations (10, 11), all of the CP deletion mutants were able to infect N. benthamiana, and they accumulated to the same level as (data not shown) and caused symptoms similar to those previously reported for a CymRSV CP deletion mutant (G11AbglII) (10, 11). Surprisingly, all of the TBSV-P CP deletion mutants were also able to infect D. stramonium plants (Fig. 3B) and induced expanding light chlorotic lesions on the inoculated leaves (Fig. 1C, panel e) which were milder than those induced by TBSV-PCP (data not shown). These results clearly indicate that the CP of TBSV-P was not required for the infection of D. stramonium and support the alternative hypothesis that CymRSV CP inhibits the virus infection on D. stramonium by eliciting an incompatible HR response.

View larger version (32K): [in this window] [in a new window]   FIG. 3.   Symptoms and RNA accumulation in plants infected with CP deletion mutants of TBSV-P, CymRSV, and CymRSV-TBSV-P chimeras. (A) Diagrams and induced symptoms of the CP deletion mutant chimeras. The restriction endonuclease sites used for constructing the CP deletion mutant chimeras are indicated. Symptoms of the parental and CP deletion mutant chimeras are indicated at the right. Symptoms were evaluated at 6 to 8 dpi. Three plants were inoculated with each inoculum, and the experiment was repeated three times. Dotted lines indicate the deleted sequences, and solid lines represent presumed untranslated sequences. Numbers shown above the CP of CymRSV indicate the predicted length of the translated polypeptides. Other symbols are defined in the legend to Fig. 2A. (B) Accumulation of viral RNAs in the inoculated leaves of D. stramonium plants inoculated with the CP deletion mutants of TBSV-P and the indicated chimeras, respectively. Total RNAs were extracted at 6 to 8 dpi and subsequently subjected to Northern analysis. (C) Accumulation of CymRSV CP deletion mutant viral RNAs in D. stramonium plants. Total RNAs were extracted at 6 to 8 dpi from inoculated leaves of D. stramonium, and Northern analysis was performed using the same DNA probes as defined in the legend to Fig. 2B. G and sg1 indicate the positions of the genomic RNA and the subgenomic RNA 1 of the CP deletion mutant chimeras, respectively. Other symbols are described in the legend to Fig. 2B.

A stretch of 860 nt of RNA in CymRSV CP coding sequence acts as an avirulence factor. Further studies were conducted to determine whether the CymRSV CP or the encoding RNA is the elicitor of the HR-like resistance response of D. stramonium. CymRSV and L5 CP frameshift mutants were prepared (Fig. 4A). The CP frameshift mutant chimeras (G11FsCP and L5FsCP) contain a 4-nt insertion at the AvaI (nt 2666) restriction enzyme site. Theoretically, these mutants are able to produce a 61-aa peptide in which the first 15 aa are identical to those of the CymRSV CP. In addition, G11AvaI and L5AvaI mutants were produced by deleting 5 nt at the same AvaI position, and they may produce a 58-aa peptide (their first 13 aa are identical to those of the CP). In the inoculation experiments, all of the CymRSV CP frameshift mutants (G11FsCP, L5FsCP, G11AvaI, and L5AvaI) were able to infect N. benthamiana plants, and they accumulated to the same level (Fig. 4C) as the CP mutants of wt CymRSV did and caused similar symptoms (10). The regions containing the mutations were reverse transcriptase-PCR amplified and sequenced. The results of sequence analysis confirmed that the progeny of all frameshift mutants were identical with synthetic transcripts used to inoculate N. benthamiana, and the coding region of the CymRSV CP was not restored. D. stramonium plants were inoculated with saps prepared from G11FsCP-, L5FsCP-, G11AvaI-, and L5AvaI-infected N. benthamiana. Surprisingly, none of these CP frameshift mutant viruses were able to infect D. stramonium plants; instead, they elicited a typical HR-like resistance response (Fig. 1C, panel h), as the wt CymRSV did. Northern blot analysis confirmed that the CymRSV CP frameshift mutants were not able to accumulate in D. stramonium (Fig. 4D). This finding strongly suggests that the RNA in the CymRSV CP coding region downstream of the AvaI site at position 2666 acts as an avirulence factor on resistant D. stramonium plants. We analyzed further whether this RNA region can elicit the HR-like reaction in a different genetic background. The TBSV-P CP coding sequence between the NotI (nt 2725) and SalI (nt 3670) restriction enzyme sites was replaced by three different parts of the CymRSV CP coding sequence (Fig. 4B). These were AvaI-BglII (860-nt long), AvaI-ApaI (686-nt long), and ApaI-BglII (174-nt long) fragments of CymRSV CP. The resulting chimeras are able to produce only short peptides (24 aa) instead of functional CP. It was shown above (Fig. 3) that the deletion of AvaI-BglII or ApaI-BglII fragments of CymRSV CP abolished the elicitor activity of CymRSV CP on D. stramonium plants. Therefore, it was expected that the insertion of the AvaI-BglII CymRSV CP fragment or part of it into the corresponding position of TBSV-P would result in an HR-like resistance response in D. stramonium. All chimeras (TBSV-AvaBg, -AvApa, and -ApaBg) were able to infect N. benthamiana plants and accumulated to levels similar to those of the TBSV-P CP mutant (data not shown). In addition, the regions of interest of the progeny of chimeras (TBSV-AvaBg, -AvApa, and -ApaBg) were reverse transcriptase-PCR amplified and sequenced. The results of sequence analysis confirmed that the progeny of all chimeras retained the original sequences. Northern blot analysis revealed that constructs carrying the AvaI-ApaI (TBSV-AvApa) or ApaI-BglII (TBSV-ApaBg) fragment of CymRSV CP accumulated efficiently in the inoculated leaves of D. stramonium (Fig. 4D). These chimeras also induced light chlorotic spots on the inoculated leaves (data not shown). In contrast, the TBSV-AvaBg chimera carrying the CymRSV CP AvaI-BglII fragment was not detectable by Northern analysis (Fig. 4D). It induced an HR-like local necrotic response, and thus the virus infection was confined. However, there was a little delay in the virus localization, so the necrotic lesions became a little larger than with CymRSV infection (Fig. 1C, panel i versus panel b). Sequence comparison of the two viruses using the Bestfit program of the Wisconsin package, version 9.1, Genetics Computer Group, Madison, Wis., revealed that within this 860-nt-long RNA region there is only 55% similarity between CymRSV and TBSV-P. This is the least conserved region in the genomic RNAs compared to the 81% similarity of the rest of the sequences. Despite the relatively low similarity of these 860-nt-long RNA regions, we did not find significant differences between the predicted secondary structure of the two RNA stretches. In conclusion, our results demonstrate that the 860-nt-long RNA sequence located between the AvaI and BglII sites in the CymRSV CP is an avirulence factor. Moreover, we provide evidence that this RNA segment is sufficient and also required to elicit an HR-like resistance response in D. stramonium.

View larger version (31K): [in this window] [in a new window]   FIG. 4.   Viral symptoms and RNA accumulation in plants inoculated with the CP frameshift mutants of CymRSV, L5, and TBSV-P chimeras carrying different parts of the coding sequence of the CymRSV CP. (A) Schematic diagrams and induced symptoms of the frameshift mutants. Restriction enzyme sites used for constructing chimeras are indicated. The inserted nucleotides at the AvaI site are shown by boldface letters. The positions of the five deleted nucleotides (CCGAG) are indicated by the symbol . The developed symptoms of the inoculated plants are shown on the right. Three plants were inoculated with each inoculum, and the experiment was repeated three times. The nucleotide (above) and amino acid (below) sequences of the region of interest of the CymRSV CP in the G11 and L5 genomes are shown. Numbers at the starts and ends of the sequences indicate the nucleotide and amino acid positions in the CymRSV genome and the CP sequence, respectively. Dots indicate the nonmodified part of the sequences. Asterisks indicate the termini of wt and mutant CymRSV CP. (B) Diagrams of the TBSV-P genome carrying different parts of the CymRSV CP. The developed symptoms of the inoculated plants are shown at the right. The large open boxes show the short 24-aa-long coding region of the N-terminal part of the TBSV-P CP. Thick dark lines indicate the CymRSV CP, and thin open boxes show the out-of-frame sequences of TBSV-P. Dotted lines indicate the deleted sequences. Three plants were inoculated with each inoculum, and the experiment was repeated three times. (C) The accumulation of viral RNAs in the inoculated leaves of N. benthamiana plants infected with chimeras indicated above the lanes. Total RNAs were extracted at 6 to 8 dpi and subsequently subjected to Northern analysis. The DNA probes used and other symbols are defined in the legend to Fig. 2. (D) Northern analysis of total RNAs extracted at 6 to 8 dpi from the inoculated leaves of D. stramonium plants infected with the chimeras indicated above the lanes. The DNA probes used and other symbols are defined in the legend to Fig. 2.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we described the use of CymRSV-TBSV-P chimeras to determine viral factors involved in the determination of host specificity of tombusviruses. We have shown here that an 860-nt-long RNA sequence of CymRSV is the elicitor of a very rapid HR-like response of D. stramonium. This finding provides the first example that an untranslatable RNA can trigger an HR-like resistance response in virus-infected plants. This finding represents a new feature of molecular strategies employed by plants for defense against pathogens.

Mapping of avirulence factors of tombusviruses. Mapping of host-range determinants of plant viruses by gene exchange and reverse genetics has the major advantage of the exchanged genomic sequences being put into a very similar background to the one they originated from. Foreign virus vectors (PVX or TMV) have also been used for the same purposes (41); however, the expression of a viral gene in a different genetic background may switch a virulent gene to an avirulent one (26). In addition, the compartmentalization of the studied viruses (5) and the use of viral expression vectors could modify the plant response. The facts that CymRSV and TBSV-P have a similar primary structure and the same genome organization and that they replicate in the same compartments (27, 35) offer an excellent opportunity to study viral avirulence factors. In addition, we found significant differences in the host ranges of CymRSV and TBSV-P. TBSV-P readily infects C. annum and D. stramonium plants systemically, while CymRSV is not able to infect these plants systemically. The results of press blot analysis clearly indicated that the failure of CymRSV to infect D. stramonium plants is not caused by its inability to replicate in this host. Most likely, the alterations caused by the replicating virus in the initially infected cells induced a very fast HR-like response, which arrested further spread of the virus. Therefore, the differences in the host ranges of the two tombusviruses were the consequence of different host responses of the plants infected with TBSV-P or CymRSV. It was shown by several studies that different virus proteins with different functions are able to elicit HR as an effective resistance response (8, 9). The tombusvirus proteins p22 and p19 have been shown to elicit HR-like necrotic responses in selected resistant plant species (41, 42). In contrast, the replacement of the nested coding regions of p19 and p22 has no effect on the differential response of C. annum and D. stramonium to TBSV-P and CymRSV (e.g., see symptoms of L5 and L7 or L6 and L8), confirming that different hosts can recognize different avirulence factors of the same pathogen (9). Similarly, the virus-encoded replicase subunits (p33 and p92) did not influence the host specificity (e.g., see symptoms of L11 and L12 on C. annum and D. stramonium).

Can RNAs of molecular pathogens be recognized as avirulence factors? Cellular organisms have evolved different strategies to recognize and combat molecular parasites. In eukaryotes, recognition of double-stranded (ds) replication intermediates of RNA viruses leads to the release of different defense reactions, including PTGS (1) and interferon-mediated responses (21). PTGS degrades in a sequence-specific manner all mRNAs homologous to the dsRNA. In vertebrates, dsRNA molecules trigger interferon-induced cellular antiviral responses which might lead to apoptosis. Importantly, dsRNAs are recognized independently from their sequences.