The introduction ten years ago of RNA interference (RNAi) as a tool for molecular exploration in Trypanosoma brucei has led to a surge in our understanding of the pathogenesis and biology of this human parasite. In particular, a genome-wide RNAi screen has recently been combined with next-generation Illumina sequencing to expose catalogues of genes associated with loss of fitness in distinct developmental stages. At present, this technology is restricted to RNAi-positive protozoan parasites, which excludes T. cruzi, Leishmania major, and Plasmodium falciparum. Therefore, elucidating the mechanism of RNAi and identifying the essential components of the pathway is fundamental for improving RNAi efficiency in T. brucei and for transferring the RNAi tool to RNAi-deficient pathogens. Here we used comparative genomics of RNAi-positive and -negative trypanosomatid protozoans to identify the repertoire of factors in T. brucei. In addition to the previously characterized Argonaute 1 (AGO1) protein and the cytoplasmic and nuclear Dicers, TbDCL1 and TbDCL2, respectively, we identified the RNA Interference Factors 4 and 5 (TbRIF4 and TbRIF5). TbRIF4 is a 3′-5′ exonuclease of the DnaQ superfamily and plays a critical role in the conversion of duplex siRNAs to the single-stranded form, thus generating a TbAGO1-siRNA complex required for target-specific cleavage. TbRIF5 is essential for cytoplasmic RNAi and appears to act as a TbDCL1 cofactor. The availability of the core RNAi machinery in T. brucei provides a platform to gain mechanistic insights in this ancient eukaryote and to identify the minimal set of components required to reconstitute RNAi in RNAi-deficient parasites.
RNA interference (RNAi), a naturally-occurring pathway whereby the presence of double-stranded RNA in a cell triggers the degradation of homologous mRNA, has been harnessed in many organisms as an invaluable molecular biology tool to interrogate gene function. Although this technology is widely used in the protozoan parasite Trypanosoma brucei, other parasites of considerable public health significance, such as Trypanosoma cruzi, Leishmania major, and Plasmodium falciparum do not perform RNAi. Since RNAi has recently been introduced into budding yeast, this opens up the possibility that RNAi can be reconstituted in these pathogens. The key to this is getting a handle on the essential RNAi factors in T. brucei. By applying comparative genomics we identified five genes that are present in the RNAi-proficient species, but not in RNAi-deficient species: three previously identified RNAi factors, and two novel ones, which are described here. This insight into the core T. brucei RNAi machinery represents a major step towards transferring this pathway to RNAi-deficient parasites.
Citation: Barnes RL, Shi H, Kolev NG, Tschudi C, Ullu E (2012) Comparative Genomics Reveals Two Novel RNAi Factors in Trypanosoma brucei and Provides Insight into the Core Machinery. PLoS Pathog8(5): e1002678. https://doi.org/10.1371/journal.ppat.1002678
Editor: John M. Mansfield, University of Wisconsin-Madison, United States of America
Received: December 9, 2011; Accepted: March 16, 2012; Published: May 24, 2012
Copyright: © 2012 Barnes et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by Public Health Service grants AI28798 and AI56333 to EU. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
RNA interference (RNAi) was first described in 1998 and within a short period of time tremendously facilitated the analysis of gene function, especially in organisms where classical genetic approaches are not available. This is particularly evident in the human pathogen Trypanosoma brucei, where RNAi has become a primary tool to interrogate its biology aided by the availability of an inducible and heritable system , . The effectiveness of RNAi in T. brucei is documented by over 500 publications in the past ten years and the availability of the genome sequence has opened up the possibility for genome-wide RNAi screens . In particular, the very recently introduced RIT-Seq method (RNA Interference Target Sequencing) took advantage of the power of genome-wide RNAi screens and combined it with the strength and depth of next-generation Illumina sequencing . This strategy provided the scientific community with a catalogue of genes whose knock-down is detrimental to the parasite under a variety of developmental conditions and is likely to find numerous applications in RNAi-positive parasites.
Although RNAi has flourished immensely in T. brucei, as highlighted by the above brief synopsis, it was rather disappointing to recognize both experimentally and at the genome sequence level that other protozoan parasites with a major impact on mankind, including T. cruzi , Leishmania major, L. donovani  and Plasmodium falciparum , were lacking a functional RNAi pathway. On the other hand, the genome sequence of L. (Viannia) braziliensis  predicted the existence of the RNAi pathway in this leishmania subgenus, and this prediction was recently experimentally validated . The realization that T. cruzi and old world leishmanias are RNAi-negative was quite unfortunate, but it has been argued that once the core genes involved in T. brucei RNAi are identified it might be possible to try to reconstruct the pathway in RNAi-negative trypanosomatids . Indeed, the recent success of David Bartel's group in reconstructing the RNAi pathway in S. cerevisiae  provides a proof of principle and informs us that it is realistic to try to achieve the same goal in T. cruzi or old world Leishmania sp.
Comparison of RNAi mechanisms in different model organisms  suggests that the common, minimal machinery consisted of firstly, a Dicer endonuclease of the RNase III family that processes long dsRNAs into duplex small interfering RNAs (siRNAs) with characteristic two-nucleotide 3′ overhangs; secondly, an Argonaute (AGO) “slicer” endonuclease that in a complex with a single-stranded (ss) guide siRNA (as opposed to the passenger strand that is discarded) cleaves target mRNA; and thirdly, a Dicer cofactor with a dsRNA-binding domain that facilitates siRNA biogenesis and loading into AGO. The AGO-siRNA complex forms the catalytic engine of the RNA-induced silencing complex or RISC . An additional factor found only in some RNAi-proficient organisms is an RNA-dependent RNA polymerase  that creates secondary siRNAs to amplify the initial RNAi response.
The players identified so far in T. brucei RNAi are a single AGO slicer, TbAGO1 , , and two Dicer homologs, TbDCL1  and TbDCL2 . This set of genes is also present in L. (V) braziliensis , although their precise function needs to be experimentally addressed. In T. brucei we have provided evidence that the nuclear Dicer TbDCL2 is the first line of defence against dsRNAs originating from retroposons (ingi and SLACS) and satellite-like repeats (CIR147), whereas in the cytoplasm TbDCL1 processes intermediate-sized dsRNA molecules generated by TbDCL2, as well as dsRNA that may escape from the nucleus or enter the cytoplasm from the exterior milieu , . The RISC loading mechanism in T. brucei is not known; no RNAi-specific dsRNA-binding protein has been identified yet . In Drosophila ,  and Neurospora  it has been shown that mutations of AGO slicer catalytic residues prevent cleavage of both the siRNA passenger strand and the target RNA, and siRNAs associated with the corresponding AGO are double-stranded. In contrast, mutations that affect target RNA degradation by TbAGO1 in vivo do not affect the maturation of siRNAs from duplex to single-stranded form and TbAGO1 is associated with single-stranded siRNAs , .
The above observations hinted at an additional activity involved in the formation of the TbAGO1/ss-siRNA complex. To test this hypothesis we compared the genomes of RNAi-proficient trypanosomatid protozoa with those that have lost the RNAi machinery. In addition to the three characterized RNAi proteins in T. brucei, we identified two novel factors, both of which we have shown to be active in RNAi. RNA Interference Factor 5 (TbRIF5) is required for cytoplasmic, but not nuclear RNAi, and appears to act in conjunction with TbDCL1. TbRIF4 contains a C-terminal domain related to 3′-5′ exonucleases of the DnaQ superfamily. In the absence of TbRIF4 protein or function, duplex siRNAs accumulate, and are not associated with TbAGO1. Human AGO2 can replace both TbRIF4 and TbAGO1 functions, suggesting that the slicing function of human AGO2 in RISC maturation is replaced in trypanosomes by the TbRIF4 exonuclease.
TbRIF4 and TbRIF5 are two novel T. brucei RNAi factors identified by comparative genomics
In order to identify the core RNAi genes we compared the genomes of RNAi-deficient (T. cruzi and L. major) and RNAi-proficient (T. brucei, T. congolense and L. braziliensis) trypanosomatids and selected ORFs that are exclusively present in all the RNAi-positive organisms, but absent from the RNAi-negative parasites. This analysis only returned five genes. In addition to TbAGO1, TbDCL1 and TbDCL2, the search identified two putative RNAi factors (Tb927.10.8880 and Tb927.10.10730 in T. brucei), which we named T. brucei RNA Interference Factor 4 and 5 (TbRIF4, 490 amino acids long, and TbRIF5, 1509 amino acids long).
At the primary sequence level the C-terminus of TbRIF4 contains a domain that resembles the DnaQ superfamily of 3′-5′ exonucleases (Figure 1). This structural feature is shared with Neurospora crassa QIP (QDE2 Interacting Protein), which functions in removal of the passenger strand siRNA fragments, after it has been cleaved by the AGO slicer NcQDE2 . This protein family is characterized by three motifs encompassing the active site residues DEDDh that are present in TbRIF4, as noted in a bioinformatic screen by Trudeau and colleagues . The active site residues are conserved in all RIF4 homologs identified in RNAi-proficient trypanosomatids (Figure S1A) and using the alignment interface of SWISS-MODEL , structural predictions (Figure S1B) revealed significant similarity with the exonuclease domain of DnaQ . Besides the exonuclease motifs, TbRIF4 and NcQIP do not share any other sequence similarity. TbRIF5 is specific to the trypanosomatid lineage and bioinformatic analysis did not reveal any known domains (Figure S2).
Schematic representation of TbRIF4 (the drawing is not to scale) and alignment of the three motifs characteristic of 3′-5′ exonucleases of the DEDDh superfamily from E. coli DnaQ , N. crassa QIP , and T. brucei RIF4. Asterisks and bold characters indicate conserved residues.
TbRIF4 is essential for RNAi in both the cytoplasm and the nucleus. First, rif4−/− cells (Figure S3A) had a much reduced ability to respond to transfection with α-tubulin dsRNA (Table 1, Figure 2A), an assay that monitors the cytoplasmic RNAi response and in wild-type cells results in degradation of α-tubulin mRNA. Second, in rif4−/− cells the levels of long and heterogeneous repeat-derived (CIR147) transcripts were substantially increased as compared to wild-type cells (Figure 2B, lane 2, and Figure 2C, lane 3), a previously-reported hallmark of nuclear RNAi deficiency . A TbRIF4-GFP cassette, integrated into the tubulin locus of rif4−/− cells (rif4c cell line, Figure S3A), restored TbRIF4 levels and complemented the mutant phenotypes, leading to wild-type response to dsRNA transfection (Table 1, Figure 2A) and restoration of long retroposon- and repeat-derived transcript levels (Figure 2B, lane 3, and Figure 2C, lane 4). Consistent with the involvement of TbRIF4 in both the nuclear and cytoplasmic arms of the RNAi pathway, the TbRIF4-GFP fusion protein was distributed uniformly between the nucleus and cytoplasm (Figure 2D).
(A) Response of rif4−/− cells and cells expressing TbRIF4 catalytic mutants or HsAGO2 to transfection of α-tubulin dsRNA. Cells were electroporated with different amounts (in micrograms) of α-tubulin dsRNA, as indicated above each lane or with 5 µg poly(I-C) (first lane in each panel set), and total RNA was prepared 2 h after electroporation and analyzed by Northern hybridization with an α-tubulin-specific probe (α-tubulin panels). α-tubulin mRNA hybridization was quantitated by PhosphorImager analysis, normalized to the load control and expressed as % mRNA remaining, setting as 100% the amount of α-tubulin mRNA present in the samples that received poly(I-C). Load; hybridization to fumarate hydratase mRNA. (B) Steady-state levels of repeat-derived (CIR147) transcripts are increased in rif4−/− cells but not rif5−/− cells. Total RNA isolated from various cell lines, as indicated above each lane, was analyzed by Northern hybridization with a CIR147-specific probe. Load; hybridization to α-tubulin mRNA. Filled squares indicate the positions of the large ribosomal RNAs, as determined by methylene blue staining of the membranes. (C) Semi-quantitative RT-PCR analysis of CIR147 transcript levels. cDNA derived from various cell lines (as indicated above each lane; lanes 1–10) was used as a template for 22 cycles of PCR using oligonucleotides that amplify the CIR147 tandem repeat transcripts, producing a ladder of fragments (top two panels), or a portion of histone H4 (bottom two panels). Dilutions of genomic DNA were used as a positive control and an indicator for non-saturated PCR (lanes 11–13); mock cDNA synthesis without reverse transcriptase (second and fourth panels, -RT) served as a negative control. (D) Cells were visualized using differential interference microscopy (DIC), GFP imaging, and Hoechst (Nucleus and Kinetoplast are indicated by N and K, respectively); GFP and Hoechst images were merged (merge).
TbRIF5, on the other hand, only appears to be required for the cytoplasmic branch of the RNAi pathway. rif5−/− cells (Figure S3B) transfected with double-stranded α-tubulin RNA showed a severe defect in the cytoplasmic RNAi response (Table 1). However, levels of CIR147 transcripts were not increased by TbRIF5 ablation (Figure 2B, lane 7), suggesting that nuclear RNAi was unaffected. A TbRIF5-BB2-HA cassette, integrated into the tubulin locus of rif5−/− cells (rif5c cell line, Figure S3B), complemented the cytoplasmic RNAi defect (Table 1). Due to the low level of expression of the epitope-tagged protein, the cellular localization of TbRIF5 could not be ascertained by IFA. However, cell fractionation experiments indicated that most of the protein partitioned in the cytoplasm (data not shown).
TbRIF4 functions downstream of the dicer step of RNAi, while TbRIF5 acts with TbDCL1
To determine the stage in the RNAi pathway at which TbRIF4 and TbRIF5 act, we first asked whether siRNAs were produced in their absence. Northern blot analysis revealed that in rif4−/− cells (Figure 3A, lane 2) ingi- and CIR147-derived siRNAs accumulated to levels greater than wild-type (lane 1) or rif4c (lane 3) trypanosomes, although their size distribution was more heterogeneous than in wild-type cells, possibly due to nuclease nibbling as these siRNAs are not associated with TbAGO1 (see below). An in vitro dicing assay programmed with long dsRNA  indicated that siRNA processing was as active in rif4−/− as in wild-type cell extracts (Figure 3B), suggesting that TbRIF4 acts downstream of the dicing step.
(A) Steady-state siRNA levels in rif4−/− and rif5−/− cells. Low-molecular weight RNAs were separated by denaturing gel electrophoresis and analyzed by Northern hybridization with an ingi-, SLACS- or CIR147-specific probe, as indicated next to each panel. Load; hybridization to 5S rRNA. 26 nt, DNA marker. (B) Dicing is unaffected in ago1−/− and rif4−/− T. brucei extracts, but shows an altered pattern in rif5−/− extracts. Whole cell extracts from cell lines, as indicated above the lanes, or buffer alone (−) were incubated with a 83-nt, internally labelled dsRNA substrate. After incubation the products of digestion were separated on a denaturing polyacrylamide gel. Size marker positions are indicated. (C) Native gel analysis of siRNAs from rif4−/− and rif5−/− cells. Low molecular weight RNAs isolated from various cell lines, as indicated above each lane, were resolved on a native polyacrylamide gel without (−) or with (+) heating the samples to 100°C for 2 minutes prior to electrophoresis and analyzed by Northern blotting with a CIR147 probe. A radiolabelled synthetic RNA was included as a control (lanes 7 and 8). Load; hybridization to 5S rRNA.
siRNAs also accumulated in rif5−/− parasites. However, as in a dcl1−/− cell line , while siRNAs derived from the CIR147 tandem repeats accumulated in rif5−/− cells to a level comparable to wild-type or rif5c trypanosomes (Figure 3A), SLACS retroposon siRNA levels were reduced in the mutant cell line (lane 5), but were restored by complementation (lane 6). A similar phenotype was observed for the ingi-derived siRNAs (data not shown). These data suggested a link to the cytoplasmic Dicer and the following experiments provided corroborating evidence. Firstly, the in vitro dicing assay showed that the siRNAs generated by a rif5−/− cell extract were the same size as those created by a dcl1−/− extract, which, as has been shown previously by our group, are one nucleotide smaller than those generated by a dcl2−/− extract (; Figure 3B). Secondly, we co-immunoprecipitated the two proteins (Figure S4) and found that a proportion of TbDCL1 is in a complex with TbRIF5. We expressed TbDCL1-BB2-FLAG and TbRIF5-BB2-HA in the same cell line. Immunoprecipitation of TbDCL1 (top panel) or TbRIF5 (bottom panel) was carried out using anti-FLAG or anti-HA antibodies, respectively, and the fate of both TbDCL1 and TbRIF5 was monitored by Western blotting for the common BB2 epitope. It should be noted that TbDCL1-BB2-FLAG is at least two to three times more abundant than TbRIF5-BB2-HA in this cell line (see input lane in Figure S4). When TbDCL1 was pulled down, the majority of TbRIF5 was found in the immunoprecipitated fraction. In contrast, when TbRIF5 was immunoprecipitated, approximately 50% of TbDCL1 was detected in the precipitate, in agreement with the lower abundance of TbRIF5 (the identity of the immunoprecipitated proteins was verified by Western blotting with HA monoclonal antibody for TbRIF5, or FLAG monoclonal antibody for TbDCL1; data not shown). The interaction of these two factors was further supported by: i. co-fractionation on a Superdex-200 sizing column; ii. mass-spectrometry of FLAG-immunoprecipitated TbDCL1, which revealed several peptides derived from TbRIF5 as well as TbDCL1 (data not shown). Thus, we concluded that TbDCL1 and TbRIF5 interact either directly or indirectly.
siRNAs in TbRIF4 null trypanosomes accumulate in duplex form
An essential step in the RNAi pathway is the formation of the RNA-induced silencing complex or RISC , consisting of an AGO “slicer”, an endonuclease of the RNase H family, in a complex with an siRNA, which guides target cleavage by AGO. Biochemical studies have shown that human and Drosophila AGO2 are initially loaded with duplex siRNAs and, following passenger strand nicking by AGO, the resulting fragments dissociate . Specific factors required for this transition have been identified, such as the multimeric endonuclease C3PO (Component 3 Promoter of RISC) in Drosophila and humans ,  or the Neurospora 3′-5′ exonuclease QIP , , . Thus, to test whether TbRIF4 and/or TbRIF5 were involved in RISC formation, we next asked whether the rif4−/−- and rif5−/−-derived siRNAs were single- or double-stranded. We purified RNA from wild-type, rif4−/−, rif4c, rif5−/−, and rif5c S100 extracts and analyzed the samples on a native polyacrylamide gel with or without denaturation at 100°C (Figure 3C). A radiolabelled double-stranded synthetic siRNA was used as a control (lanes 7 and 8). Following Northern blot hybridization with a CIR147 probe, most siRNAs from wild-type cells appeared single-stranded (lane 1, see ). A similar result was obtained with siRNAs from rif5−/− and rif5c cells (lanes 11 and 13). In contrast, by this analysis non-denatured rif4−/−-derived siRNAs displayed the mobility expected for duplex molecules (compare lanes 3 and 7), but were converted to ss-siRNAs upon denaturation (lane 4), whereas siRNAs from the rif4c sample again reproduced the single-stranded nature present in wild-type cells (lane 5).
To provide additional evidence that siRNAs isolated from rif4−/− trypanosomes were double-stranded, we incubated the RNA samples at different temperatures (20° to 100°C) prior to loading them on a native gel (Figure S5A). Ingi siRNA strands dissociated between 60°C and 65°C, consistent with a duplex structure. Lastly, we assessed the sensitivity of siRNAs isolated from wild-type and rif4−/− cells to treatment with RNase T1 and RNase A, which preferentially cleave ss RNA in high salt (Figure S5B). Wild-type CIR147 siRNAs were mostly degraded by the ribonucleases (lane 2), with trace amounts resistant to digestion, but siRNAs derived from rif4−/− cells were largely resistant to degradation (lane 5) and were only cleaved efficiently following denaturation (lane 6).
The above results strongly indicated that rif4−/− siRNAs are predominantly in duplex form. We think it is unlikely that the duplex siRNAs are an artefact of the RNA isolation procedure, since very little duplex siRNA can be observed in wild-type samples prepared under the same conditions (Figure 3C, lane 1). We showed previously by Northern blotting  that both sense and antisense siRNAs from the retroposons ingi and SLACS accumulate in T. brucei. Recent deep-sequencing of siRNAs associated with TbAGO1 confirmed these early observations and extended them to include CIR147 (unpublished observations). Although wild-type siRNAs have the potential to anneal to form duplex structures, they do not appear to do so under the experimental conditions used.
siRNAs are not associated with TbAGO1 in TbRIF4 null cells
Next we asked whether siRNAs in rif4−/− cells were associated with TbAGO1. To this end, we introduced TAP-tagged TbAGO1, which fully complements TbAGO1 deficiency , into rif4−/− cells. It should be noted that the level of TAP-tagged TbAGO1 was much lower in rif4−/− cells as compared to wild-type trypanosomes (Figure 4A, top panel, compare lanes 4 and 8). This dramatic diminution of TbAGO1 levels was also observed for endogenous TbAGO1 in rif4−/− cells (Figure 4B, middle panel, compare lanes 1 and 2), but returned to normal upon complementation (lane 3). By semi-quantitative Western blot analysis we estimated that the level of endogenous TbAGO1 was reduced approximately 20-fold in rif4−/− cells (data not shown).
(A) Cytoplasmic extracts from cells expressing TAP-tagged TbAGO1 in a wild-type (lanes 1–4) or rif4−/− (lanes 5–8) background were immunoprecipitated with IgG beads, and the indicated amounts of supernatant (S, lanes 1 and 5) and immunoprecipitated material (P, lanes 2–4 and 6–8) were analyzed by Western blotting with a polyclonal anti-TbAGO1 antibody (upper panel, TbAGO1). CIR147 siRNAs in the soluble and precipitated materials were revealed by Northern hybridization (middle panel, CIR147 siRNA). RNA loading was quantified by hybridization to 5S rRNA (bottom panel, 5S RNA). (B) TbRIF4 and TbAGO1 steady-state levels in various cell lines. Total protein extracts from 5×106 cells from various cell lines, as indicated above each lane, were resolved by SDS-PAGE and Western blotted with a polyclonal anti-TbRIF4 antibody (top panel, TbRIF4) or a polyclonal anti-TbAGO1 antibody (middle panel, TbAGO1). Load; a cross-reacting band to the TbAGO1 antibody (bottom panel).
TAP-tagged TbAGO1 was immunoprecipitated from wild-type and rif4−/− S100 extracts and varying amounts from the different samples were analysed by Western (Figure 4A, top panel) or Northern blotting (middle panel). This experiment showed that when roughly similar amounts of TbAGO1 protein were loaded (compare 90% of the immunoprecipitated material from rif4−/− cells with 9% or even 1% of the material from the wild-type sample), no siRNAs could be detected in the pellet from the rif4−/− sample (lanes 6–8), although these molecules were clearly visible in the wild-type immunoprecipitate (lanes 2–4). We concluded that the majority of CIR147 siRNAs in rif4−/− cells were not associated with TbAGO1.
Biochemical investigation of TbRIF4 exonuclease activity
Our experiments so far suggested that TbRIF4 plays a role in the conversion of double- to single-stranded siRNAs. To test whether the predicted 3′-5′ exonuclease domain in the C-terminus of TbRIF4 was involved in this process, we generated recombinant GST-TbRIF4 fusion proteins (Figure S6A) and tested its activity in vitro. Incubation of 25-nt duplex α-tubulin siRNA-315 , labelled at the 5′ end of the sense strand, with increasing amounts of GST-TbRIF4 produced a distinct ladder of fragments that progressively decreased in size with increasing protein concentration (Figure 5A, lanes 4–11) and with time (Figure S6B, lanes 2–7). Mutation of conserved exonuclease active site residue H472 to alanine drastically reduced the detected exonuclease activity (Figure 5A, lane 12 and Figure S6B, lane 8). With the above substrate we observed accumulation of a 7-nt fragment, but this was not the case when the antisense strand of the duplex substrate was labelled (Figure S6C), indicating that digestion by TbRIF4 is somewhat affected by the substrate sequence. Nevertheless, these results suggested that in vitro both strands are accessible to TbRIF4 digestion and that the enzyme has a distributive rather than processive mode of action.
(A) Titration of GST-RIF4 activity. 1 pmol of a 25 nt-long, 5′-end labelled synthetic dsRNA was incubated with 0.01–30 pmol GST-RIF4 (lanes 4–11; molar ratios from 10∶1 to 1∶30) or 30 pmol GST-TbRIF4 carrying mutation H472A (lane 12). Lane 3, input. AH, alkaline hydrolysate ladder. M, marker oligonucleotides. Samples were separated on a denaturing 20% polyacrylamide gel. (B) Action of GST-TbRIF4 on a RNA substrate blocked at the 3′ end of one strand. 1 pmol of a 25 nt-long, 5′-end labelled synthetic dsRNA with a dideoxycytosine residue at the 3′ end of the labelled strand was incubated with 0.03–30 pmol GST-TbRIF4 (lanes 3–6), or buffer alone (lanes 1 and 2). Samples were separated on a non-denaturing 16% polyacrylamide gel. As a control for the separation of double-stranded and single-stranded species, the substrate incubated with buffer alone was resolved on the same gel without (lane 1, native) or with (lane 2, denat.) heating to 100°C for 2 minutes prior to electrophoresis.
To further characterize TbRIF4 activity, we first used substrates with no, single or double 3′ nucleotide overhang (the label was placed at the 5′ end of the sense strand) and found that they were also processed, suggesting that the 2-nt 3′ extension is not important for substrate recognition by recombinant TbRIF4 in vitro (Figure S6D), consistent with a distributive mode of action. Secondly, we showed that recombinant TbRIF4 acted preferentially on duplex as opposed to single-stranded RNA (Figure S6E) and that catalysis required magnesium ions (Figure S6F) and a free 3′ end (Figure S6G), as processing of the RNA substrate was inhibited by EDTA or by blocking the 3′ end with a 2′,3′-dideoxycytosine, respectively. Thirdly, we asked whether GST-TbRIF4 can convert the RNA duplex into ss-RNA. Recombinant TbRIF4 does not exhibit clear strand selectivity (Figure S6C), thus either strand can be degraded and any released ssRNA will base-pair with available complementary products to form dsRNA. To circumvent this problem, we carried out an in vitro assay where one strand of the double-stranded RNA substrate was labelled at the 5′ end and the 3′ end of the same strand was blocked by dideoxy-C to prevent it from being a substrate for TbRIF4 (Figure 5B). This mimics the effect of a bound protein protecting one end of the duplex, a likely in vivo scenario. The products of this reaction were separated on a native gel, showing that many of the molecules had been converted to the single-stranded form. Thus, TbRIF4 has an intrinsic 3′-5′ exonuclease activity in vitro, which acts preferentially on double-stranded RNA, and its activity is dependent on a predicted active site residue.
Active site residues are important for TbRIF4 function in vivo
Next we assessed whether TbRIF4's conserved DnaQ exonuclease family DEDDh residues (Figure 1) were essential for in vivo function by mutating each amino acid to alanine in the context of the TbRIF4-GFP complementation cassette, and introducing the resulting mutants into rif4−/− cells. With the exception of mutant E296A (Figure 4B, lane 5), the other proteins (D294A, D392A, H472A and D477A, lanes 4, 6, 7 and 8) were expressed close to the level of wild-type TbRIF4-GFP (lane 3). Upon transfection of 5 µg α-tubulin dsRNA, which in wild-type cells results in approximately 84% degradation of target mRNA, in cells lacking TbRIF4 or expressing mutants D294A, D392A or H472A there was very low or no detectable degradation of α-tubulin mRNA (Figure 2A). In contrast, transfection of dsRNA into cells expressing mutant D477A showed that at each dsRNA concentration tested, the cytoplasmic RNAi efficiency was reduced about 50% as compared to the wild-type level. Thus, it appeared that the D477 residue is not as important as the other predicted catalytic residues for TbRIF4 activity in vivo. However, our results indicated that mutation of residues D294, D392 and H472 severely impaired the ability of cells to respond to dsRNA transfection. Similarly, nuclear RNAi as measured by RT-PCR of CIR147 repeats was partially restored in cells expressing mutant D477A (Figure 2C, lane 8), in agreement with the results of Figure 2A, whereas cells expressing mutants D294A, D392A and H472A appeared significantly impaired in nuclear RNAi (lanes 5–7). Finally, expression of active-site TbRIF4 mutants D294A, D392A, H472A and D477A failed to restore TbAGO1 to wild-type levels (Figure 4B, middle panel, lanes 4–8), although the amount of TbAGO1 was increased as compared to rif4−/− cells.
TbRIF4 is part of a complex containing TbAGO1 and siRNAs
The finding that siRNAs were not detectably associated with TbAGO1 in rif4−/− cells could indicate a physical interaction between the two proteins. To test this hypothesis, we performed immunoprecipitations. Pull-downs of either wild-type or H472A TbRIF4-GFP revealed the presence of TbAGO1 in the immunoprecipitates (Figure 6A), with more TbAGO1 in the H472A TbRIF4-GFP sample compared to the wild-type sample (compare lanes 3 and 6, bottom panel). We next asked whether siRNAs could be immunoprecipitated with wild-type or H472A TbRIF4-GFP (Figure 6B). In cells expressing wild-type TbRIF4-GFP, CIR147 siRNAs were below the limit of detection (lane 2). In contrast, a fraction of CIR147 siRNAs were found in the immunoprecipitated material after pull-down of H472A TbRIF4-GFP (lane 4). These results indicated that a proportion of a catalytically inactive TbRIF4 was in a complex containing TbAGO1 and siRNAs. We take these results as supportive evidence that TbRIF4 and TbAGO1 interact in vivo, either directly or indirectly, and that the interaction can be stabilized by mutation of the TbRIF4 exonuclease active site, although the significance of the latter observation remains to be investigated further.
(A) Cytoplasmic extracts from rif4−/− cells expressing wt TbRIF4-GFP (lanes 1–3) or mutant H472A TbRIF4-GFP (lanes 4–6) were subjected to immunoprecipitation with anti-GFP antibody. 0.1% of the supernatant (S, lanes 1 and 4), 0.1% (lanes 2 and 5) and 0.4% (lanes 3 and 6) of the immunoprecipitated material were analyzed by Western blotting with a polyclonal anti-RIF4 antibody (top panel) or a polyclonal anti-TbAGO1 antibody (bottom panel). The efficiency of immunoprecipitation is shown in the antibody titration experiment in Figure S7. (B) The presence of CIR147 siRNAs in the various samples as described in panel (A) was analyzed by Northern blot hybridization with a CIR147 probe. Load; hybridization to 5S rRNA. (C) Immunoprecipitated HsAGO2 is associated with duplex and single-stranded siRNAs. Cytoplasmic extracts from cell lines expressing HA-FLAG-HsAGO2 in an ago1−/−:rif4−/− genetic background, were immunoprecipitated with anti-FLAG antibodies. RNA associated with HsAGO2 was separated on a native polyacrylamide gel and subjected to Northern hybridization with a CIR147 probe (upper panel). Total RNA from rif4−/− trypanosomes was used as a control for the position of duplex siRNAs (T, lane 1). Load; hybridization to 5S rRNA (middle panel). Equal amounts of the supernatant (S, lane 2) and immunoprecipitated material (P, lane 3) were analyzed by Western blotting with anti-HA antibody (HsAGO2 panel).
TbRIF4 substitutes for the slicer activity of human AGO2 in passenger strand removal
Our past and present data indicated that the transition from ds- to ss-siRNAs in T. brucei is uncoupled from TbAGO1 slicer function and is carried out by the TbRIF4 exonuclease. We previously reported that human AGO2 slicer (HsAGO2), but not HsAGO1, which lacks target transcript slicing activity, can partially complement TbAGO1 null cells . Since HsAGO2 slicing activity is involved in the cleavage of both the passenger strand of duplex siRNAs and the target RNA , , we hypothesized that the restoration of the RNAi response by HsAGO2 may occur by a TbRIF4-independent mechanism. Thus, HA-FLAG-tagged HsAGO2 was introduced into rif4−/−:ago1−/− cells. Expression of HsAGO2 resulted in a partial cytoplasmic RNAi response measured by transfection with dsRNA (Figure 2A), as previously described for cells expressing HsAGO2 and TbRIF4, but nuclear RNAi was not restored (Figure 2C, lane 9), as evidenced by the accumulation of CIR147 transcripts. To monitor the physical state of the siRNAs, HsAGO2 was immunoprecipitated and the associated siRNAs were analyzed by native gel electrophoresis (Figure 6C). In agreement with the partial RNAi response, a fraction of the siRNAs was single-stranded (lane 3). These results indicated that in trypanosomes expressing HsAGO2 maturation of siRNAs from duplex- to ss-form, takes place independently of TbRIF4 and underscored a model where TbRIF4 in a complex with TbAGO1 degrades the siRNA passenger strand, thus substituting for HsAGO2 passenger cleavage activity in ss-siRNA biogenesis in T. brucei.
The availability of genome sequences from trypanosomatids that are either RNAi proficient or not, provided us with an opportunity to tease out genes specific to the RNAi mechanism. Quite surprisingly, the comparative genomics analysis revealed a limited set of five genes: the previously characterized TbAGO1, TbDCL1 and TbDCL2, as well two additional candidates, TbRIF4 and TbRIF5. Here, we showed that both TbRIF4 and TbRIF5 are essential for RNAi, and carry out distinct functions. TbRIF4 is implicated in the formation of a stable TbAGO1/guide siRNA complex through conversion of duplex siRNAs to their single-stranded form, whereas TbRIF5's function is less clearly defined due to a lack of recognizable domains, but this factor appears to be working in concert with TbDCL1. The latter conclusion is based on the observations that TbRIF5's ablation resulted in cytoplasmic RNAi deficiency and that the protein associated with TbDCL1, the cytoplasmic Dicer. We therefore propose a role for TbRIF5 as a cofactor in TbDCL1-mediated processing of long dsRNA. It is unlikely that TbRIF5 is involved in nuclear RNAi as CIR147 siRNAs, which are generated by TbDCL2, are not affected by TbRIF5 ablation, and we have recently shown that RNAi of small nucleolar RNAs similarly requires TbDCL2, but is independent of TbDCL1 . The specific reduction of SLACS and ingi siRNAs in TbRIF5 null cells points to the possibility that TbRIF5 is required at some point in the biogenesis of such siRNAs in the cytoplasm, either by stimulating TbDCL1 cleavage activity and/or by functioning as a bridge between TbDCL1, the siRNAs and TbAGO1. RNase III enzymes, such as Dicer and Drosha, in general require the assistance of dsRBP cofactors . For instance, in Drosophila R2D2 forms a complex with DCR-2, enhances the incorporation of siRNAs into RISC  and acts as a sensor for strand asymmetry of siRNAs . In human cells the interaction of DGCR8 with Drosha is essential for excision of pre-miRNAs from primary miRNA containing transcripts . Although TbRIF5 lacks a recognizable dsRNA binding domain, it may be that such a motif exists but is highly divergent.
TbRIF4 contains a predicted 3′-5′ exonuclease domain. In vivo TbRIF4 knockout results in both nuclear and cytoplasmic RNAi deficiency and in the accumulation of duplex siRNAs that are not detectably associated with TbAGO1 by co-immunoprecipitations. As predicted from its domain structure, TbRIF4-GST is endowed with in vitro 3′-5′ exonuclease activity, and processes duplex siRNAs. Complementation of rif4−/− cells with TbRIF4-GFP carrying mutations of catalytic residues D294A, D392A and H472A, which also impair exonuclease activity in vitro (data not shown), did not restore RNAi competency of TbRIF4 null cells, indicating that TbRIF4 catalytic activity is required for RNAi function.
TbRIF4 and TbAGO1 interact, either directly or indirectly, and form a complex that can be recovered in greater amounts from mutant H472A TbRIF4-GFP than from wild-type cells and contains a small proportion of siRNAs. Although at present it is unclear how this observation relates to TbRIF4 function, it indicates that mutant TbRIF4 becomes “stuck” in a complex with TbAGO1. It is possible that the interaction between the two wild-type proteins is transient and that after the siRNA is converted to single-stranded form within this complex (either before or after loading into TbAGO1), TbRIF4 dissociates from TbAGO1. TbRIF4 on its own does not appear to show any strand selectivity; therefore, the strand removed is possibly determined by the orientation in which the duplex is presented by Dicer, as has been reported in Drosophila . The use of chaperone proteins, for example heat shock proteins, in RISC loading has been reported in higher eukaryotes –. It is possible that TbRIF4 acts in this way in addition to its role as an exonuclease.
We also observed that TbAGO1 levels are diminished in cells lacking functional TbRIF4. However, we note that there was no straightforward correlation between the amount of TbAGO1 present in the various mutant cell lines and their degree of RNAi competency, suggesting that catalytically impaired TbRIF4s can to some extent promote TbAGO1 accumulation through a mechanism that is presently unknown. As TbAGO1 mRNA levels are only marginally affected by TbRIF4 deletion (data not shown), either TbAGO1 translation is severely inhibited or the protein is made and rapidly degraded. So far inhibition of the proteasome and lysosome degradation pathways has not restored TbAGO1 levels (unpublished observations), pointing to the possibility that translational control regulates the amount of TbAGO1 that is synthesized.
T. brucei RIF4 shares with Neurospora QIP a C-terminal 3′-5′ exonuclease domain of the DnaQ superfamily. Although both proteins are essential RNAi factors and interact with the corresponding Argonaute, they are functionally distinct in several respects. First, in Neurospora qipko cells duplex siRNAs are stably loaded into the AGO homolog QDE-2 and, after QDE-2 slicer-mediated cleavage of the passenger strand, remain bound to QDE-2 in duplex, nicked form . In contrast, siRNAs can be successfully loaded into TbAGO1 and TbRIF4 can accomplish the transition from double- to single-stranded siRNAs in the absence of TbAGO1 slicer activity , . Secondly, we showed that TbRIF4 activity is required for TbAGO1 protein accumulation. This finding is not mirrored in Neurospora or Drosophila, where mutation of passenger strand fragment removal factors QIP or C3PO has no effect on the corresponding AGO levels , . In light of these data, it is evident that TbRIF4 and NcQIP are functionally distinct RNAi factors.
In trypanosomes expressing HsAGO2, maturation of siRNAs takes place independently of TbRIF4, suggesting a model where TbRIF4 substitutes for AGO slicing activity in the generation of ss-siRNAs. This would represent a departure from strategies of RISC maturation in higher eukaryotes, which favor slicer activity marking the passenger strand for degradation, although bypass pathways can operate in the absence of slicer activity , , .
Taken together with our earlier studies, the results presented here provide a detailed depiction of the core RNAi machinery in the early divergent parasitic protozoan T. brucei. In these organisms, the RNAi pathway is initiated by two distinct Dicer-like enzymes, namely TbDCL1, mostly found in the cytoplasm , and TbDCL2, a mainly nuclear Dicer responsible for processing dsRNAs originating from retroposons and satellite-like repeats . siRNAs generated by both Dicers are channelled into a single member of the Argonaute family of proteins, namely TbAGO1 , . The identification and characterization of TbRIF4 and TbRIF5 most likely completes the collection of core components shared by RNAi-positive trypanosomatids. However, it is likely that, as described in other organisms, other cellular factors interact with the core RNAi machinery. Some of these factors may perform housekeeping functions, for instance the Hsp90/70 family of chaperons that functions in RISC assembly in higher eukaryotes, and be present in all trypanosomatids irrespective of their RNAi proficiency, while others may be organism-specific. Indeed, we have recently discovered that the T. brucei homologue of the plant HEN1 2′-O methyltransferase  is required for siRNA 3′ end modification (unpublished observation). However, siRNA modification is not an absolute requirement for trypanosomatid RNAi, as in L. braziliensis the HEN1 gene is absent, and deletion of TbHEN1 is still compatible with a robust RNAi response (unpublished observations). It could be argued at this point that attempting to reconstitute RNAi in RNAi-deficient pathogens is quite daunting with a minimum of five core components. However, our past and current mechanistic studies have highlighted that not all components are necessary. In particular, we have shown previously that TbDCL2-deficient cells are more responsive to RNAi triggers than wild-type cells  suggesting that this factor does not need to be included in a reconstitution attempt. In addition, in the present study we have shown that human AGO2 can functionally replace both TbAGO1 and TbRIF4, indicating that the TbRIF4 function can be bypassed by using an appropriate AGO protein. It will be interesting to see whether human AGO2 can similarly function in L. braziliensis. The factor remaining a mystery is TbRIF5 and its function will be challenging to ascertain due to a complete lack of a possible functional indicator. Nevertheless, the available proof of principle that RNAi can be reconstituted in S. cerevisiae by using only AGO1 and DCR1 from S. castelli  in combination with our functional studies raises the intriguing possibility that a similar strategy might be applicable in RNAi-negative trypanosomatids.
Materials and Methods
Previously published procedures were followed for culturing trypanosome YTat 1.1 cells , generation of knockout cell lines by PCR-based methods , Northern blots of total RNA , Western blots , S100 preparation , immunoprecipitations , and the Dicer in vitro assay .
5 µg of total RNA from various cell lines was treated with 2 units of DNase RQ (Promega), phenol extracted and ethanol precipitated. DNase-treated RNA was used in a reverse transcription reaction using the manufacturer's protocol for the Superscript II enzyme (Invitrogen), phenol extracted, ethanol precipitated, and resuspended in 100 µl water. Mock reactions without the reverse transcriptase enzyme were carried out in parallel. 22 cycles of PCR were performed using the Phusion enzyme (Finnzyme) and the GC buffer provided, according to manufacturer's instructions: 50°C annealing temperature, 10 sec extension time for the CIR147 reactions, 30 sec extension time for the histone 4 reactions.
Preparation of small RNAs
Small RNAs were extracted from S100 extracts by incubation in polysome buffer plus 20 mM EDTA, 2% SDS and 0.1 µg/µl proteinase K for 1 hour at room temperature. An equal volume of phenol and 1/10 volume 20× SET [pH 7.4] were added and the mixture was centrifuged at 13,200× g for 5 min. The supernatant was incubated with 600 µl isopropanol and 30 µg GlycoBlue (Ambion) at −20°C for 30 min and centrifuged at 13,200× g for 10 min at 4°C. The pellet was washed with 70% ethanol, dried and resuspended in water (for native gels) or urea loading buffer (for denaturing gels). RNA was transferred to a positively-charged nylon membrane (Hybond N+, GE Healthcare) and UV cross-linked at 120 mJ/cm2. The membrane was incubated in boiling water, which was allowed to cool to room temperature, twice prior to probe hybridization. This treatment significantly improved detection of dsRNA.
Preparation of probes
DNA oligos were 5′ end-labelled with T4 polynucleotide kinase (NEB) using [γ-32P]ATP and purified using P6 columns (BioRad). Hybridizations were carried out in ExpressHyb solution (Clontech) overnight at 30°C. Washes in 2× SSC, 0.2% SDS were carried out at room temperature. Blots were analyzed by PhosphorImager.
Purification of recombinant protein
TbRIF4 was cloned into the pGEX4T-1 vector. Inductions of 500 ml DH5α bacteria were carried out by adding 0.2 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) overnight at 16°C. Cells were washed in PBS containing protease inhibitors, lysed by sonication in 5 ml PBS containing protease inhibitors and GST-TbRIF4 was purified using a SpinTrap column (GE). Proteins were eluted in 50 mM Tris-HCl [pH 8.0], 10 mM glutathione, before being applied to a desalting column and stored in assay buffer (20 mM HEPES-KOH [pH 7.9], 150 mM sucrose, 20 mM potassium L-glutamate, 3 mM MgCl2, 2 mM DTT).
Exonuclease assays were carried out for 30 min at 28°C in 10 µl assay buffer including RNase Inhibitor (Roche) and 0.01 µg/µl bovine tRNA. Reactions were terminated by the addition of 4× SET, 20 mM EDTA, 30 µg GlycoBlue and isopropanol, stored at −20°C for 30 min and spun at 13,200× g for 10 min at 4°C. The pellet was washed with 70% ethanol, dried and resuspended in urea loading buffer. Samples were resolved on a 20% denaturing polyacrylamide gel and analyzed by PhosphorImager.
TbRIF4 sequence analysis. (A) Sequence alignment of TbRIF4 protein sequences from T. brucei (Tbru), T. congolense (Tcon), T. vivax (Tviv) and L. (V.) braziliensis (Lbra) created using ClustalX (www.clustal.org) and shaded using boxshade (http://www.ch.embnet.org/software/BOX_form.html). Conserved motifs are outlined, and conserved active site residues are indicated by asterisks. (B) Ribbon diagram of the secondary structure of E. coli DnaQ (left) adapted from  and of the predicted model of the exonuclease domain of TbRIF4 (right) using the alignment interface of SWISS-MODEL .
TbRIF5 sequence analysis. Sequence alignment of TbRIF5 protein sequences from T. brucei (Tbru), T. congolense (Tcon), T. vivax (Tviv) and L. (V.) braziliensis (Lbra) created using ClustalX (www.clustal.org) and shaded using boxshade (http://www.ch.embnet.org/software/BOX_form.html).
TbRIF4 and TbRIF5 knockout and complementation cell lines. (A) rif4−/− cells do not express TbRIF4 mRNA (top panel) or protein (bottom panel). Total RNA isolated from various cell lines, as indicated above each lane, was resolved by formaldehyde-agarose gel electrophoresis and analyzed by Northern blot hybridization with a TbRIF4-specific probe (top panel). Load; α-tubulin hybridization. Total protein extracts were resolved by SDS-PAGE and probed with a polyclonal anti-TbRIF4 antibody (third panel). Load; a cross-reacting band. (B) rif5−/− cells do not express TbRIF5 mRNA (top panel). Total RNA isolated from various cell lines, as indicated above each lane, was resolved by formaldehyde-agarose gel electrophoresis and analyzed by Northern blot hybridization with a RIF5-specific probe (top panel). Load; α-tubulin hybridization. Ten-fold less RNA was loaded in the rif5c lane compared to the other two lanes.
Physical association of TbDCL1 and TbRIF5. Cytoplasmic extracts from rif4−/− cells expressing DCL1-FLAG-BB2 and RIF5-HA-BB2 were subjected to immunoprecipitation with anti-FLAG antibody (upper panel) or anti-HA antibody (lower panel). Equal numbers of cell equivalents of the input (I), supernatant (S), and immunoprecipitated material (P) were analyzed by Western blotting with a monoclonal anti-BB2 antibody.
siRNAs in rif4−/− trypanosomes are double-stranded. (A) Native gel analysis of siRNAs in rif4−/− cells. RNA isolated from a rif4−/− S100 extract was resolved on a native polyacrylamide gel after heating the samples at a range of temperatures as indicated above each lane for 2 minutes prior to electrophoresis and analyzed by Northern blotting with an ingi probe. (B) siRNAs in rif4−/− cells are protected from digestion with RNase. S100-derived RNAs from wild-type (lanes 1–3) and rif4−/− (lanes 4–6) cells were treated with RNase T1 (5 U/ml) and RNase A (20 µg/ml) for 1 hr at 25°C, either without (lanes 2 and 5) or with (lanes 3 and 6) denaturation by boiling prior to enzyme addition. RNAs incubated with buffer alone (lanes 1 and 4) were included as a control. Following digestion with proteinase K, the samples were electrophoresed on a 16% denaturing gel and analyzed by Northern hybridization with a CIR147 probe (top panel). Load; hybridization to 5S rRNA (bottom panel).
Recombinant TbRIF4 3′-5′ exonuclease activity in vitro. (A) Purification of GST-TbRIF4. 30 pmol of recombinant GST-TbRIF4 was resolved by SDS-PAGE and stained with Coomassie Brilliant Blue. (B) Time-course of GST-TbRIF4 activity. 1 pmol of a 25 nt-long, 5′-end labelled synthetic dsRNA was incubated with 30 pmol GST-RIF4 for 1 to 30 min (lanes 2–7). The activity of GST-TbRIF4 carrying mutation H472A incubated for 30 min is shown in lane 8, and a no-protein control is in lane 1. (C) GST-TbRIF4 activity is not restricted to one of the duplex RNA strands. 1 pmol of synthetic siRNA-315 labelled at the 5′ end of either the “sense” or “antisense” strand was incubated with wild-type GST-TbRIF4 (wt; lanes 2 and 5), GST-TbRIF4 carrying mutation H472A (H472A; lanes 3 and 6) or buffer alone (lanes 1 and 4). (D) Action of GST-TbRIF4 on siRNA-like substrates with shorter 3′ overhangs. 1 pmol of a 25 nt-long, 5′-end labelled synthetic dsRNA with 3′ overhangs of 2, 1 or 0 nt was incubated with 0.04 to 25 pmol GST-TbRIF4 (lanes 2–6, 8–12 and 14–18) or buffer alone (lanes 1, 7 and 13). (E) GST-TbRIF4 preferentially cleaves double-stranded RNA. 1 pmol synthetic ssRNA (lanes 1 and 2) or dsRNA (lanes 3 and 4) was incubated with (+; lanes 2 and 4) or without (−; lanes 1 and 3) 30 pmol wild-type GST-TbRIF4. (F) GST-TbRIF4 activity is abolished by addition of EDTA. 1 pmol synthetic dsRNA was incubated with 30 pmol wild-type GST-RIF4 (wt; lanes 2 and 5), GST-TbRIF4 carrying mutation H472A (H472A; lanes 3 and 6) or buffer alone (lanes 1 and 4). 5 mM EDTA was added in the reactions for lanes 4–6. (G) GST-TbRIF4 activity requires a free 3′ end. 1 pmol synthetic dsRNA either without (lanes 1–3) or with (lanes 4–6) a terminal 2′,3′-dideoxycytidine residue was incubated with 30 pmol GST-TbRIF4 (wt; lanes 2 and 5), GST-TbRIF4 carrying mutation H472A (H472A; lanes 3 and 6) or buffer alone (lanes 1 and 4).
Efficient pull-down of TbRIF4-GFP. Cytoplasmic extracts from rif4−/− cells expressing wt TbRIF4-GFP were subjected to immunoprecipitation with anti-GFP antibody. Equal volumes of the input (I), supernatant (S) and immunoprecipitated material (P) were analyzed by Western blotting with a polyclonal anti-TbRIF4 antibody.
Conceived and designed the experiments: RLB HS NGK CT EU. Performed the experiments: RLB HS NGK CT EU. Analyzed the data: RLB HS NGK CT EU. Wrote the paper: RLB CT EU.
- 1. Kolev NG, Tschudi C, Ullu E (2011) RNA interference in protozoan parasites: achievements and challenges. Eukaryot Cell 10: 1156–1163.NG KolevC. TschudiE. Ullu2011RNA interference in protozoan parasites: achievements and challenges.Eukaryot Cell1011561163
- 2. Atayde VD, Tschudi C, Ullu E (2011) The emerging world of small silencing RNAs in protozoan parasites. Trends Parasitol 27: 321–327.VD AtaydeC. TschudiE. Ullu2011The emerging world of small silencing RNAs in protozoan parasites.Trends Parasitol27321327
- 3. Alsford S, Turner DJ, Obado SO, Sanchez-Flores A, Glover L, et al. (2011) High-throughput phenotyping using parallel sequencing of RNA interference targets in the African trypanosome. Genome Res 21: 915–924.S. AlsfordDJ TurnerSO ObadoA. Sanchez-FloresL. Glover2011High-throughput phenotyping using parallel sequencing of RNA interference targets in the African trypanosome.Genome Res21915924
- 4. DaRocha WD, Otsu K, Teixeira SM, Donelson JE (2004) Tests of cytoplasmic RNA interference (RNAi) and construction of a tetracycline-inducible T7 promoter system in Trypanosoma cruzi. Mol Biochem Parasitol 133: 175–186.WD DaRochaK. OtsuSM TeixeiraJE Donelson2004Tests of cytoplasmic RNA interference (RNAi) and construction of a tetracycline-inducible T7 promoter system in Trypanosoma cruzi.Mol Biochem Parasitol133175186
- 5. Robinson KA, Beverley SM (2003) Improvements in transfection efficiency and tests of RNA interference (RNAi) approaches in the protozoan parasite Leishmania. Mol Biochem Parasitol 128: 217–228.KA RobinsonSM Beverley2003Improvements in transfection efficiency and tests of RNA interference (RNAi) approaches in the protozoan parasite Leishmania.Mol Biochem Parasitol128217228
- 6. Baum J, Papenfuss AT, Mair GR, Janse CJ, Vlachou D, et al. (2009) Molecular genetics and comparative genomics reveal RNAi is not functional in malaria parasites. Nucleic Acids Res 37: 3788–3798.J. BaumAT PapenfussGR MairCJ JanseD. Vlachou2009Molecular genetics and comparative genomics reveal RNAi is not functional in malaria parasites.Nucleic Acids Res3737883798
- 7. Peacock CS, Seeger K, Harris D, Murphy L, Ruiz JC, et al. (2007) Comparative genomic analysis of three Leishmania species that cause diverse human disease. Nat Genet 39: 839–847.CS PeacockK. SeegerD. HarrisL. MurphyJC Ruiz2007Comparative genomic analysis of three Leishmania species that cause diverse human disease.Nat Genet39839847
- 8. Lye LF, Owens K, Shi H, Murta SM, Vieira AC, et al. (2010) Retention and loss of RNA interference pathways in trypanosomatid protozoans. PLoS Pathog 6: e1001161.LF LyeK. OwensH. ShiSM MurtaAC Vieira2010Retention and loss of RNA interference pathways in trypanosomatid protozoans.PLoS Pathog6e1001161
- 9. Drinnenberg IA, Weinberg DE, Xie KT, Mower JP, Wolfe KH, et al. (2009) RNAi in budding yeast. Science 326: 544–550.IA DrinnenbergDE WeinbergKT XieJP MowerKH Wolfe2009RNAi in budding yeast.Science326544550
- 10. Liu Q, Paroo Z (2010) Biochemical principles of small RNA pathways. Annu Rev Biochem 79: 295–319.Q. LiuZ. Paroo2010Biochemical principles of small RNA pathways.Annu Rev Biochem79295319
- 11. Hammond SM, Bernstein E, Beach D, Hannon GJ (2000) An RNA-directed nuclease mediates post-transcriptional gene silencing in Drosophila cells. Nature 404: 293–296.SM HammondE. BernsteinD. BeachGJ Hannon2000An RNA-directed nuclease mediates post-transcriptional gene silencing in Drosophila cells.Nature404293296
- 12. Cerutti H, Casas-Mollano JA (2006) On the origin and functions of RNA-mediated silencing: from protists to man. Curr Genet 50: 81–99.H. CeruttiJA Casas-Mollano2006On the origin and functions of RNA-mediated silencing: from protists to man.Curr Genet508199
- 13. Durand-Dubief M, Bastin P (2003) TbAGO1, an argonaute protein required for RNA interference, is involved in mitosis and chromosome segregation in Trypanosoma brucei. BMC Biol 1: 2.M. Durand-DubiefP. Bastin2003TbAGO1, an argonaute protein required for RNA interference, is involved in mitosis and chromosome segregation in Trypanosoma brucei.BMC Biol12
- 14. Shi H, Djikeng A, Tschudi C, Ullu E (2004) Argonaute protein in the early divergent eukaryote Trypanosoma brucei: control of small interfering RNA accumulation and retroposon transcript abundance. Mol Cell Biol 24: 420–427.H. ShiA. DjikengC. TschudiE. Ullu2004Argonaute protein in the early divergent eukaryote Trypanosoma brucei: control of small interfering RNA accumulation and retroposon transcript abundance.Mol Cell Biol24420427
- 15. Shi H, Tschudi C, Ullu E (2006) An unusual Dicer-like1 protein fuels the RNA interference pathway in Trypanosoma brucei. RNA 12: 2063–2072.H. ShiC. TschudiE. Ullu2006An unusual Dicer-like1 protein fuels the RNA interference pathway in Trypanosoma brucei.RNA1220632072
- 16. Patrick KL, Shi H, Kolev NG, Ersfeld K, Tschudi C, et al. (2009) Distinct and overlapping roles for two Dicer-like proteins in the RNA interference pathways of the ancient eukaryote Trypanosoma brucei. Proc Natl Acad Sci U S A 106: 17933–17938.KL PatrickH. ShiNG KolevK. ErsfeldC. Tschudi2009Distinct and overlapping roles for two Dicer-like proteins in the RNA interference pathways of the ancient eukaryote Trypanosoma brucei.Proc Natl Acad Sci U S A1061793317938
- 17. Shi H, Tschudi C, Ullu E (2006) Functional replacement of Trypanosoma brucei Argonaute by the human slicer Argonaute2. RNA 12: 943–947.H. ShiC. TschudiE. Ullu2006Functional replacement of Trypanosoma brucei Argonaute by the human slicer Argonaute2.RNA12943947
- 18. Okamura K, Ishizuka A, Siomi H, Siomi MC (2004) Distinct roles for Argonaute proteins in small RNA-directed RNA cleavage pathways. Genes Dev 18: 1655–1666.K. OkamuraA. IshizukaH. SiomiMC Siomi2004Distinct roles for Argonaute proteins in small RNA-directed RNA cleavage pathways.Genes Dev1816551666
- 19. Kim K, Lee YS, Carthew RW (2007) Conversion of pre-RISC to holo-RISC by Ago2 during assembly of RNAi complexes. RNA 13: 22–29.K. KimYS LeeRW Carthew2007Conversion of pre-RISC to holo-RISC by Ago2 during assembly of RNAi complexes.RNA132229
- 20. Maiti M, Lee HC, Liu Y (2007) QIP, a putative exonuclease, interacts with the Neurospora Argonaute protein and facilitates conversion of duplex siRNA into single strands. Genes Dev 21: 590–600.M. MaitiHC LeeY. Liu2007QIP, a putative exonuclease, interacts with the Neurospora Argonaute protein and facilitates conversion of duplex siRNA into single strands.Genes Dev21590600
- 21. Shi H, Ullu E, Tschudi C (2004) Function of the Trypanosome Argonaute 1 protein in RNA interference requires the N-terminal RGG domain and arginine 735 in the Piwi domain. J Biol Chem 279: 49889–49893.H. ShiE. UlluC. Tschudi2004Function of the Trypanosome Argonaute 1 protein in RNA interference requires the N-terminal RGG domain and arginine 735 in the Piwi domain.J Biol Chem2794988949893
- 22. Shi H, Chamond N, Djikeng A, Tschudi C, Ullu E (2009) RNA interference in Trypanosoma brucei: role of the N-terminal RGG domain and the polyribosome association of argonaute. J Biol Chem 284: 36511–36520.H. ShiN. ChamondA. DjikengC. TschudiE. Ullu2009RNA interference in Trypanosoma brucei: role of the N-terminal RGG domain and the polyribosome association of argonaute.J Biol Chem2843651136520
- 23. Zhang D, Xiong H, Shan J, Xia X, Trudeau VL (2008) Functional insight into Maelstrom in the germline piRNA pathway: a unique domain homologous to the DnaQ-H 3′-5′ exonuclease, its lineage-specific expansion/loss and evolutionarily active site switch. Biol Direct 3: 48.D. ZhangH. XiongJ. ShanX. XiaVL Trudeau2008Functional insight into Maelstrom in the germline piRNA pathway: a unique domain homologous to the DnaQ-H 3′-5′ exonuclease, its lineage-specific expansion/loss and evolutionarily active site switch.Biol Direct348
- 24. Bordoli L, Kiefer F, Arnold K, Benkert P, Battey J, et al. (2009) Protein structure homology modeling using SWISS-MODEL workspace. Nat Protoc 4: 1–13.L. BordoliF. KieferK. ArnoldP. BenkertJ. Battey2009Protein structure homology modeling using SWISS-MODEL workspace.Nat Protoc4113
- 25. Hamdan S, Carr PD, Brown SE, Ollis DL, Dixon NE (2002) Structural basis for proofreading during replication of the Escherichia coli chromosome. Structure 10: 535–546.S. HamdanPD CarrSE BrownDL OllisNE Dixon2002Structural basis for proofreading during replication of the Escherichia coli chromosome.Structure10535546
- 26. Czech B, Hannon GJ (2011) Small RNA sorting: matchmaking for Argonautes. Nat Rev Genet 12: 19–31.B. CzechGJ Hannon2011Small RNA sorting: matchmaking for Argonautes.Nat Rev Genet121931
- 27. Matranga C, Tomari Y, Shin C, Bartel DP, Zamore PD (2005) Passenger-strand cleavage facilitates assembly of siRNA into Ago2-containing RNAi enzyme complexes. Cell 123: 607–620.C. MatrangaY. TomariC. ShinDP BartelPD Zamore2005Passenger-strand cleavage facilitates assembly of siRNA into Ago2-containing RNAi enzyme complexes.Cell123607620
- 28. Liu Y, Ye X, Jiang F, Liang C, Chen D, et al. (2009) C3PO, an endoribonuclease that promotes RNAi by facilitating RISC activation. Science 325: 750–753.Y. LiuX. YeF. JiangC. LiangD. Chen2009C3PO, an endoribonuclease that promotes RNAi by facilitating RISC activation.Science325750753
- 29. Ye X, Huang N, Liu Y, Paroo Z, Huerta C, et al. (2011) Structure of C3PO and mechanism of human RISC activation. Nat Struct Mol Biol 18: 650–657.X. YeN. HuangY. LiuZ. ParooC. Huerta2011Structure of C3PO and mechanism of human RISC activation.Nat Struct Mol Biol18650657
- 30. Lee DW, Millimaki R, Aramayo R (2010) QIP, a component of the vegetative RNA silencing pathway, is essential for meiosis and suppresses meiotic silencing in Neurospora crassa. Genetics 186: 127–133.DW LeeR. MillimakiR. Aramayo2010QIP, a component of the vegetative RNA silencing pathway, is essential for meiosis and suppresses meiotic silencing in Neurospora crassa.Genetics186127133
- 31. Xiao H, Alexander WG, Hammond TM, Boone EC, Perdue TD, et al. (2010) QIP, a protein that converts duplex siRNA into single strands, is required for meiotic silencing by unpaired DNA. Genetics 186: 119–126.H. XiaoWG AlexanderTM HammondEC BooneTD Perdue2010QIP, a protein that converts duplex siRNA into single strands, is required for meiotic silencing by unpaired DNA.Genetics186119126
- 32. Djikeng A, Shi H, Tschudi C, Ullu E (2001) RNA interference in Trypanosoma brucei: cloning of small interfering RNAs provides evidence for retroposon-derived 24-26-nucleotide RNAs. RNA 7: 1522–1530.A. DjikengH. ShiC. TschudiE. Ullu2001RNA interference in Trypanosoma brucei: cloning of small interfering RNAs provides evidence for retroposon-derived 24-26-nucleotide RNAs.RNA715221530
- 33. Meister G, Landthaler M, Patkaniowska A, Dorsett Y, Teng G, et al. (2004) Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs. Mol Cell 15: 185–197.G. MeisterM. LandthalerA. PatkaniowskaY. DorsettG. Teng2004Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs.Mol Cell15185197
- 34. Gupta SK, Hury A, Ziporen Y, Shi H, Ullu E, et al. (2010) Small nucleolar RNA interference in Trypanosoma brucei: mechanism and utilization for elucidating the function of snoRNAs. Nucleic Acids Res 38: 7236–7247.SK GuptaA. HuryY. ZiporenH. ShiE. Ullu2010Small nucleolar RNA interference in Trypanosoma brucei: mechanism and utilization for elucidating the function of snoRNAs.Nucleic Acids Res3872367247
- 35. Liu Q, Rand TA, Kalidas S, Du F, Kim HE, et al. (2003) R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway. Science 301: 1921–1925.Q. LiuTA RandS. KalidasF. DuHE Kim2003R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway.Science30119211925
- 36. Tomari Y, Matranga C, Haley B, Martinez N, Zamore PD (2004) A protein sensor for siRNA asymmetry. Science 306: 1377–1380.Y. TomariC. MatrangaB. HaleyN. MartinezPD Zamore2004A protein sensor for siRNA asymmetry.Science30613771380
- 37. Gregory RI, Yan KP, Amuthan G, Chendrimada T, Doratotaj B, et al. (2004) The Microprocessor complex mediates the genesis of microRNAs. Nature 432: 235–240.RI GregoryKP YanG. AmuthanT. ChendrimadaB. Doratotaj2004The Microprocessor complex mediates the genesis of microRNAs.Nature432235240
- 38. Preall JB, He Z, Gorra JM, Sontheimer EJ (2006) Short interfering RNA strand selection is independent of dsRNA processing polarity during RNAi in Drosophila. Curr Biol 16: 530–535.JB PreallZ. HeJM GorraEJ Sontheimer2006Short interfering RNA strand selection is independent of dsRNA processing polarity during RNAi in Drosophila.Curr Biol16530535
- 39. Johnston M, Geoffroy MC, Sobala A, Hay R, Hutvagner G (2010) HSP90 protein stabilizes unloaded argonaute complexes and microscopic P-bodies in human cells. Mol Biol Cell 21: 1462–1469.M. JohnstonMC GeoffroyA. SobalaR. HayG. Hutvagner2010HSP90 protein stabilizes unloaded argonaute complexes and microscopic P-bodies in human cells.Mol Biol Cell2114621469
- 40. Iwasaki S, Kobayashi M, Yoda M, Sakaguchi Y, Katsuma S, et al. (2010) Hsc70/Hsp90 chaperone machinery mediates ATP-dependent RISC loading of small RNA duplexes. Mol Cell 39: 292–299.S. IwasakiM. KobayashiM. YodaY. SakaguchiS. Katsuma2010Hsc70/Hsp90 chaperone machinery mediates ATP-dependent RISC loading of small RNA duplexes.Mol Cell39292299
- 41. Iki T, Yoshikawa M, Nishikiori M, Jaudal MC, Matsumoto-Yokoyama E, et al. (2010) In vitro assembly of plant RNA-induced silencing complexes facilitated by molecular chaperone HSP90. Mol Cell 39: 282–291.T. IkiM. YoshikawaM. NishikioriMC JaudalE. Matsumoto-Yokoyama2010In vitro assembly of plant RNA-induced silencing complexes facilitated by molecular chaperone HSP90.Mol Cell39282291
- 42. Miyoshi T, Takeuchi A, Siomi H, Siomi MC (2010) A direct role for Hsp90 in pre-RISC formation in Drosophila. Nat Struct Mol Biol 17: 1024–1026.T. MiyoshiA. TakeuchiH. SiomiMC Siomi2010A direct role for Hsp90 in pre-RISC formation in Drosophila.Nat Struct Mol Biol1710241026
- 43. Rand TA, Petersen S, Du F, Wang X (2005) Argonaute2 cleaves the anti-guide strand of siRNA during RISC activation. Cell 123: 621–629.TA RandS. PetersenF. DuX. Wang2005Argonaute2 cleaves the anti-guide strand of siRNA during RISC activation.Cell123621629
- 44. Leuschner PJ, Ameres SL, Kueng S, Martinez J (2006) Cleavage of the siRNA passenger strand during RISC assembly in human cells. EMBO Rep 7: 314–320.PJ LeuschnerSL AmeresS. KuengJ. Martinez2006Cleavage of the siRNA passenger strand during RISC assembly in human cells.EMBO Rep7314320
- 45. Yu B, Yang Z, Li J, Minakhina S, Yang M, et al. (2005) Methylation as a crucial step in plant microRNA biogenesis. Science 307: 932–935.B. YuZ. YangJ. LiS. MinakhinaM. Yang2005Methylation as a crucial step in plant microRNA biogenesis.Science307932935
- 46. Ngo H, Tschudi C, Gull K, Ullu E (1998) Double-stranded RNA induces mRNA degradation in Trypanosoma brucei. Proc Natl Acad Sci U S A 95: 14687–14692.H. NgoC. TschudiK. GullE. Ullu1998Double-stranded RNA induces mRNA degradation in Trypanosoma brucei.Proc Natl Acad Sci U S A951468714692
- 47. Arhin GK, Shen S, Ullu E, Tschudi C (2004) A PCR-based method for gene deletion and protein tagging in Trypanosoma brucei. Methods Mol Biol 270: 277–286.GK ArhinS. ShenE. UlluC. Tschudi2004A PCR-based method for gene deletion and protein tagging in Trypanosoma brucei.Methods Mol Biol270277286