Double-stranded (ds) RNA induces potent gene silencing, termed RNA interference (RNAi). At an early step in RNAi, an RNaseIII-related enzyme, Dicer (DCR-1), processes long-trigger dsRNA into small interfering RNAs (siRNAs). DCR-1 is also required for processing endogenous regulatory RNAs called miRNAs, but how DCR-1 recognizes its endogenous and foreign substrates is not yet understood. The C. elegans RNAi pathway gene, rde-4, encodes a dsRNA binding protein that interacts during RNAi with RNA identical to the trigger dsRNA. RDE-4 protein also interacts in vivo with DCR-1, RDE-1, and a conserved DExH-box helicase. These findings suggest a model in which RDE-4 and RDE-1 function together to detect and retain foreign dsRNA and to present this dsRNA to DCR-1 for processing (Tabara, 2002).
RNAi and related posttranscriptional gene silencing (PTGS) pathways have been implicated in silencing transposons and/or viruses, suggesting that these pathways may function as a form of sequence-directed immunity. The two strongest mutants discovered in screens carried out to discover C. elegans mutants deficient in RNAi, (rde-1, encoding an Argonaute family member and rde-4) were both completely deficient in RNAi but failed to exhibit transposon activation or any other discernible phenotypes. Subsequent genetic studies suggested that rde-1 and rde-4 may function upstream in the initiation of RNAi (Tabara, 2002 and references therein).
To shed more light on the role of the RDE-4 protein in RNAi, the corresponding gene was positionally cloned. The rde-4 gene is predicted to encode a 385 amino acid protein that contains two copies of a conserved motif found in dsRNA binding proteins. The genetic lesions in rde-4 appear likely to be null mutations, and consistent with this idea, the ne299 allele, when placed in trans to a chromosomal deficiency, exhibits an RNAi-deficient phenotype and no other additional phenotypes (Tabara, 2002).
The RDE-4 protein contains two copies of a dsRNA binding motif (dsRBM) that is found in numerous other proteins that interact with dsRNA in a sequence-nonspecific manner. Therefore, it was first asked whether recombinant RDE-4 could bind to dsRNA by using an in vitro gel shift assay. Labeled dsRNA corresponding to a portion of a GFP cDNA was incubated with GST-fused RDE-4, or with the GST protein alone, and was then run on a native polyacrylamide gel. The GST-fused RDE-4 efficiently bound to the dsRNA, causing a mobility shift, while the GST protein did not. In competition assays, dsRNA but not single-stranded RNA efficiently competed for RDE-4 binding. Similarly, it was asked whether the labeled dsRNA could bind to the recombinant RDE-4 immobilized on a membrane. This North-Western blot analysis also showed that the dsRNA binds efficiently to the GST-fused RDE-4 but not to the GST protein alone (Tabara, 2002).
It was then asked if RDE-4 interacts with dsRNA during RNAi in vivo. To address this question, polyclonal antibodies were raised against a recombinant RDE-4 protein. Animals were then exposed to dsRNA targeting the gene pos-1. Extracts from these animals and from control animals not exposed to dsRNA were prepared, and RDE-4-specific antibodies were used to precipitate the RDE-4 protein complex. The precipitates were extracted to purify any associated RNA and analyzed by agarose and acrylamide gel electrophoresis followed by Northern blotting using pos-1 sense and antisense radiolabeled RNA probes. Both the sense and antisense probes detected approximately equal amounts of pos-1 RNA that migrated below 350 bases on an agarose gel and migrated in a size range of 50 to 200 bases on an acrylamide gel. The coprecipitated pos-1 RNA was resistant to treatment with a mixture of the single-strand-specific RNaseA and RNaseT1 enzymes, suggesting that the pos-1 RNA is double stranded. Little or no pos-1 RNA coimmunoprecipitated with RDE-4, which was purified from populations either not exposed to dsRNA or exposed for only 15 min. As expected, rde-4(ne299) mutants failed to exhibit pos-1 dsRNA in the precipitate. The RDE-4 immune complex recovered from the rde-1(ne300) and rde-3(ne298) mutant strains exhibited greatly reduced quantities of pos-1 dsRNA (Tabara, 2002).
dsRNA is thought to exist and function at multiple steps in RNAi: (1) it functions in the initiation step when long dsRNA is recognized as foreign; (2) it functions in the execution step when target mRNA is bound by siRNA and cleaved, and (3) it functions in a potential amplification step in which target mRNA sequences are copied by an RNA-dependent RNA polymerase activity. In the above experiments, the probes and techniques used were designed to detect long dsRNA sequences contained within the trigger dsRNA, and thus could not address the question of whether siRNAs or target mRNA sequences might also coimmunoprecipitate with RDE-4. To address these questions, sequences in the 3′ UTR of pos-1 that were not included in the dsRNA trigger were screened for. These experiments failed to detect significant coprecipitating RNA, indicating that sequences within the target mRNA are not stable components of the RDE-4 immune complex (Tabara, 2002).
Several studies have suggested that RNAi may involve a mechanism for the amplification of the dsRNA signal and provide evidence that sequences in the target mRNA that lie 5′ of the trigger sequence can be amplified. In order to ask if the RDE-4 immune complex contains antisense RNA sequences derived from regions of the mRNA located 5′ of the trigger dsRNA, transgenic animals were utilized that expressed dsRNA corresponding to a portion of unc-22 gene. RNA probes were then used to assay for sequences in the RDE-4 immune complex that were derived either from the trigger region or from the region located just 5′ of the trigger. As expected, the RDE-4 immune complex contained unc-22 antisense RNA corresponding to the targeted region of unc-22. In contrast, no significant hybridization was observed when a sequence derived from the region located just 5′ of the trigger dsRNA was used as a probe. As expected, precipitates prepared from nontransgenic wild-type animals did not exhibit significant hybridization to the probes designed to detect unc-22 antisense RNA (Tabara, 2002).
Finally, in order to ask if siRNA sequences precipitate with RDE-4, the RNA species in the RDE-4 immune complex and in the RDE-4-depleted supernatant were fractionated by acrylamide gel electrophoresis. RNA species smaller than 150 bases were analyzed by Northern blotting using a pos-1 sense probe. It was found that the RNA species in the supernatant included abundant molecules of less than 150 bases, including a prominent signal corresponding in size to siRNAs, that migrated at approximately 24 to 25 bases. RNA species of 24 to 25 bases were absent in the RDE-4 immune complex, suggesting that RDE-4 preferentially interacts with longer dsRNAs. Taken together, these studies indicate that RDE-4 binds to long dsRNA sequences that are restricted to the region present within the trigger dsRNA. These findings support a model in which RDE-4 interacts with the trigger dsRNA and functions in the initiation step of RNAi (Tabara, 2002).
In order to determine what fraction of the total dsRNA was bound to RDE-4 and to examine the requirements for the stability of the dsRNA bound to RDE-4, RNA extracts from wild-type and rde mutant strains were analyzed. Wild-type animals showed an accumulation of antisense and sense RNA corresponding to the pos-1 trigger dsRNA sequence that appeared to be very similar to the dsRNA sequences found in the RDE-4 immune complex. This RNA species was similar in size but slightly smaller, on average, than the major pos-1 RNA species detected in the bacterial expression strain used to induce RNAi, suggesting that some digestion of the bacterially expressed dsRNA may have occurred during transit through the intestine of the animals. Single-stranded RNA-specific nucleases digested the endogenous pos-1 mRNA but failed to digest the accumulated RNA species, suggesting that the accumulated RNA species is indeed double-stranded (Tabara, 2002).
Interestingly, the analysis of total RNA extracts showed that animals lacking rde-1, rde-3, and rde-4 activities exhibit markedly reduced levels of dsRNA. In contrast, animals lacking rde-2 and mut-7 accumulate dsRNA to levels that are intermediate (rde-2) or similar to wild-type (mut-7). The finding that the accumulation of this dsRNA species requires some but not all rde(+) activities supports the idea that this RNA species is an intermediate in the RNAi process (Tabara, 2002).
To ask what fraction of the accumulating dsRNA in the total RNA extract from wild-type animals is associated with RDE-4, the immunoprecipitations were repeated with either RDE-4-specific antibodies or with control nonspecific IgG obtained from preimmune serum and the amount of the dsRNA present in the immune complex and supernatant of each lysate was compared. Approximately 60% of the pos-1 dsRNA species present in the total RNA coprecipitates with the RDE-4 immune complex, indicating that a majority of the accumulating dsRNA is associated with RDE-4 (Tabara, 2002).
Previous studies have suggested that RDE-1 and RDE-4 function at an upstream step in the initiation of RNAi. This observation, and the finding that RDE-1 and RDE-4 are both required for the accumulation of dsRNA, prompted a test to see if RDE-4 forms a complex in vivo with RDE-1. For these studies a transgenic strain was generated expressing RDE-1 tagged with the HA epitope and RDE-4 tagged with the FLAG epitope. RDE-4 was immunoprecipitated with RDE-4-specific polyclonal antibodies, and RDE-1 via the epitope tag, and the precipitates were analyzed by immunoblotting. In reciprocal assays, RDE-1 and RDE-4 were found to coprecipitate. Based on a comparison of the RDE-1 present in total extracts and in the RDE-4 immune complex, it is estimated that approximately 5% of the RDE-1 molecules stably associate with the RDE-4 complex (Tabara, 2002).
The interaction between RDE-1 and RDE-4 occurred in animals that were not exposed to exogenous dsRNA, and the interaction was not abolished by treatment with either double-stranded RNA-directed nuclease, RNaseV1, or a mixture of RNaseA and T1 that are single-stranded RNA-specific nucleases. These findings suggest that RDE-1 and RDE-4 form a complex in the absence of dsRNA, but the possibility cannot be ruled out that dsRNA promotes the interaction or is required for the interaction but is protected from digestion in the immune complex (Tabara, 2002).
In order to identify additional proteins that interact with RDE-4 in vivo, large quantities of the RDE-4 immune complex were isolated and then its components were resolved using SDS polyacrylamide gel electrophoresis (SDS-PAGE). rde-4(ne299) mutant animals were used that were rescued via an RDE-4::GFP fusion protein. Protein extracts prepared from these transgenic animals and from control wild-type animals were then passed over a column conjugated with anti-GFP monoclonal antibody. After electrophoresis, the gel was silver stained to detect components of the RDE-4::GFP immune complex. These experiments identified prominent bands migrating at 210 kilodaltons (kDa) and 110 kDa as well as several additional polypeptides that were present in the RDE-4::GFP immune complex but absent in wild-type control animals. A minor band at 125 kDa was identified, corresponding in size to RDE-1, but it contained insufficient protein for further analysis. The 210 kDa and 110 kDa bands were subjected to tryptic digestion, and the resulting fragments were then analyzed by mass spectrometry. The peptide mass fingerprints of p210 and p110 were compared with in silico-digested sequences from protein databases. A total of 21 peptides from the tryptic digestion of p210 matched DCR-1, while 11 peptides from p110 matched a DExH-box RNA helicase-related protein F15B10.2 (Tabara, 2002).
The finding that RDE-4 forms a complex in vivo with DCR-1 suggests that RDE-4 may function to present the foreign trigger dsRNA to DCR-1 for processing. DCR-1 also acts to process the stem-loop precursors of the developmental regulators lin-4 and let-7. However, rde-4 mutants do not exhibit developmental defects, and thus are not required for the activities of lin-4 or let-7. Furthermore, no lin-4 or let-7 precursor RNAs associated with the RDE-4 immune complex were detected. Thus, although RDE-4 interacts strongly with DCR-1 in vivo, RDE-4 is not essential for the developmental functions of DCR-1 and may rather serve as an adaptor protein that recruits DCR-1 to the RNAi pathway (Tabara, 2002).
The RDE-4 protein interacts with a 110 kDa protein, which as described in the previous section corresponds to the predicted gene F15B10.2, one of two closely related C. elegans DExH-box RNA helicase genes. F15B10.2, along with its homolog C01B10.1, map directly adjacent to one another on chromosome IV and appear likely to be expressed together from a single promoter. Consistent with this idea, it was found that some F15B10.2 cDNAs were spliced to the trans-spliced leader sequence (sl2), which is often spliced onto the 5′ end of transcripts encoded by downstream genes in an operon. The C01B10.1 gene and a cDNA clone (yk226c6) both appear to encode a frame-shifted open reading frame predicted to terminate the protein prior to the helicase domain, raising the question of whether or not this represents a functional gene. An alternative conceptual splicing of C01B10.1, or utilization of an alternative initial methionine, could encode a protein with overall 74% identity to F15B10.2 (Tabara, 2002).
These C. elegans genes have close homologs in mammals, including a gene named RHIV-1 induced in response to viral infection in pigs. Notably, these helicase proteins are more similar to the DCR-1 helicase domain than to any other helicase-related proteins in C. elegans or other organisms. Consequently, these genes have been named drh-1 and drh-2 (for F15B10.2 and C01B10.1, respectively) which stands for dicer-related helicase (Tabara, 2002).
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.