|Home | About | Journals | Submit | Contact Us | Français|
A unique characteristic of the protistan parasite Trypanosoma brucei is a multifunctional RNA polymerase I which, in addition to synthesizing rRNA as in other eukaryotes, transcribes gene units encoding the major cell surface antigens variant surface glycoprotein and procyclin. Thus far, purification of this enzyme has revealed nine orthologues of known subunits but no active enzyme. Here, we have epitope tagged the specific subunit RPB6z and tandem affinity purified RNA polymerase I from crude extract. The purified enzyme was active in both a nonspecific and a promoter-dependent transcription assay and exhibited enriched protein bands with apparent sizes of 31, 29, and 27 kDa. p31 and its trypanosomatid orthologues were identified, but their amino acid sequences have no similarity to proteins of other eukaryotes, nor do they contain a conserved sequence motif. Nevertheless, p31 cosedimented with purified RNA polymerase I, and RNA interferance-mediated silencing of p31 was lethal, affecting the abundance of rRNA. Moreover, extract of p31-silenced cells exhibited a specific defect in transcription of class I templates, which was remedied by the addition of purified RNA polymerase I, and an anti-p31 serum completely blocked RNA polymerase I-mediated transcription. We therefore dubbed this novel functional component of T. brucei RNA polymerase I TbRPA31.
The protist Trypanosoma brucei is transmitted by the tsetse fly and lives freely in the bloodstream of its mammalian host, causing the fatal disease African sleeping sickness in humans. This parasite is also the only known eukaryote with a multifunctional RNA polymerase (pol) I which synthesizes both ribosomal and mRNA. T. brucei RNA pol I transcribes the large rRNA gene unit (RRNA) like its counterpart in other eukaryotes and, in addition, the gene units which encode the parasite's major cell surface antigens variant surface glycoprotein (VSG) and procyclin (6, 10, 15, 30). In other eukaryotes, RNA pol I-mediated expression of reporter genes was found to be inefficient (8), and mRNA is exclusively synthesized by RNA pol II. This clear functional division between the two polymerases is most likely due to cotranscriptional mRNA capping, which is mediated by a specific interaction of the capping enzyme with the RNA pol II carboxy-terminal domain (5, 20, 44). In contrast, T. brucei can utilize RNA pol I for effective production of functional mRNA (31, 45), because pre-mRNAs of this parasite are capped posttranscriptionally by trans splicing of the same capped 39-nucleotide-long RNA fragment onto the 5′ end of each mRNA. The trans-spliced sequence is known as spliced leader (SL) or mini-exon and is derived from the 5′ terminus of the SL RNA (reviewed in reference 19).
The multifunctional nature of T. brucei RNA pol I is associated with an unprecedented versatility of this enzyme. During the life cycle of the parasite, RNA pol I is recruited to four structurally different promoters (reviewed in reference 9). Moreover, in bloodstream trypanosomes, RNA pol I is sequestered into two subnuclear compartments, namely, the nucleolus for the synthesis of rRNA and the expression site body for the expression of a single VSG gene located in a telomeric VSG expression site (VSG ES) (26). While this monoallelic VSG expression leads to a uniform cell surface coat, switching to the expression of a different VSG gene results in antigenic variation of the coat, enabling T. brucei to evade the host immune response (recently reviewed in references 3, 12, and 28). Since bloodstream trypanosomes are covered by ~107 VSG molecules, all derived from a single VSG, the parasite's utilization of RNA pol I for VSG expression is most likely due to the high efficiency of this enzyme. Similarly, in tsetse midgut procyclic-stage trypanosomes, RNA pol I transcribes procyclin genes at two chromosome-internal loci.
The versatility of RNA pol I transcription in T. brucei and its role in antigenic variation have raised the possibility that RNA pol I has essential subunits or subunit domains which are unique to the parasite. Several studies have therefore been dedicated to characterizing this enzyme. Eukaryotic RNA polymerases have a minimum of 12 subunits, and TbRPA1, the largest subunit of T. brucei RNA pol I, was initially identified and characterized by homology (13, 38). Subsequently, a partial purification of the enzyme revealed a very large RPA2 subunit which carries a unique ~50-kDa-large N-terminal extension domain, the functional relevance of which is not yet understood (33). With the completion of the T. brucei genome (1), data mining uncovered putative orthologues of eight further subunits, namely, RPB5, RPB6, RPB8, RPB10, RPB12, RPC40, RPC19, and RPA12 (14, 27, 40). Interestingly, trypanosomatid genomes contain two paralogues of the subunits RPB5, RPB6, and RPB10, which are distantly related to each other. Originally they were named with prefixes 1 and 2 (40), but due to the presence of sequence insertions, one paralogous set was renamed RPB5z, RPB6z, and RPB10z, while the other paralogues kept their conventional names (14; for a list of subunits, see reference 27). A first tandem affinity purification (TAP) of T. brucei RNA pol I confirmed the association of RPB5z, RPC40, and RPA12 with the enzyme (40). A 30-kDa-long unknown protein, which specifically copurified with TAP-tagged RPA12, was also identified in this study. The development of a new epitope combination, termed PTP (derived from the tag composition protein A-TEV protease cleavage site-protein C), allowed a more-efficient purification of RNA pol I (35), revealing a stable complex which consisted of the five previously identified subunits and, in addition, the subunits RPB8, RPC19, and RPB10z (27). Furthermore, by epitope tagging and reciprocal coimmunoprecipitation, it was established that RPB6z, but not RPB6, interacted with RNA pol I, albeit in an unstable manner (27). However, the RNA pol I preparations obtained thus far were inactive, suggesting that essential subunits were lost during purification (10, 27, 40).
Here, we report the purification of RNA pol I from T. brucei, which was active in both a nonspecific and a promoter-dependent transcription assay. In comparison to the results of our previous purification, three additional protein bands became apparent. The protein with an apparent size of 31 kDa was identified as the protein which previously copurified with TbRPA12 (40). As we demonstrate, this protein is a subunit of RNA pol I and indispensable for in vitro transcription from three different class I promoters. In accordance with these findings, silencing the expression of this protein affected rRNA abundance and was lethal to trypanosomes.
pPURO-PTP-RPB6z and pRPA31-HA-BLA are constructs for in-allele fusion of the PTP and hemaglutinin (HA) sequences to RPB6z and RPA31, respectively. pPURO-PTP-RPB6z is a derivative of pN-PURO-PTP (35), in which the RPB6z coding region from position 4 to position 242 was inserted using the NotI and ApaI restriction sites. pRPA31-HA-BLA is a derivative of pC-PTP-NEO (35) with the following modifications: the PTP sequence was replaced by the HA sequence, and the RPA31 coding region from position 284 to position 678 was inserted into the KpnI and NotI restriction sites. In addition, the neomycin phosphotransferase coding region was replaced by the corresponding region of the blasticidin acetyltransferase via restriction sites NdeI and BstBI. For conditional RNA interference (RNAi), the RPA31 coding region from position 121 to position 664 was inserted into pZJM (41). RPB6z allele deletion was achieved with the PCR product HYGR-R, in which the hygromycin phosphotransferase coding region was fused to 100 bp of the RPB6z gene flanks on either side. For the in vitro transcription analysis in extracts from 29-13 (RNAi) cells (43), which express T7 RNA polymerase, the gene inserts of the established class I templates GPEET-trm, Rib-trm, and VSG-trm (17) were excised with restriction enzymes KpnI and EcoRI and cloned into the corresponding sites of the T7 promoterless plasmid pUC19. For the class II template SLins19, the EcoRI and HindIII sites were used. These new template constructs were designated with the add-on V.2.
Procyclic T. brucei cell culture, targeted integration of linear DNAs into cells by electroporation, and the generation of stable cell lines by selection and limiting dilution were described in detail previously (10). The TbA8 cell line was obtained by two consecutive rounds of transfection and selection in which one RPB6z allele was replaced by the HYGR-R PCR product and the second allele was modified by targeted insertion of BmgBI-linearized pPURO-PTP-RPB6z. The HA tag was C-terminally fused to RPA31 in TbA8 cells by targeted insertion of SnaBI-linearized pRPA31-HA-BLA to create cell line TbE3. Correct integration of DNAs was confirmed by PCR with one primer positioned outside the cloned region. Transfected cell lines were maintained in medium containing 20 μg/ml of hygromycin, 4 μg/ml of puromycin, and/or 10 μg/ml of blasticidin. 29-13 cells for conditional RNAi (43) containing the pZJM-RPA31 construct were kept in culture with 50 μg/ml of hygromycin, 15 μg/ml of G418, and 2.5 μg/ml of phleomycin. In RNAi experiments, double-stranded RNA (dsRNA) synthesis was induced with 10 μg/ml doxycycline. Cells were counted and diluted to 3 × 106 cells/ml daily.
RNA used in Northern blot analysis was prepared from induced and noninduced cells by a QIAGEN RNeasy mini kit to enrich for mRNA following the manufacturer's protocol for cytoplasmic RNA. RPA31 mRNA was detected by a digoxigenin-labeled PCR product spanning the cDNA region from position −23 to position 118 relative to the translation initiation codon. The α-tubulin probe encompassed the complete coding region and was obtained through random primed labeling of a PCR product using [α-32P]dCTP and the Klenow fragment. For the assessment of rRNA abundance in RNAi-induced and noninduced cells, 2 × 108 cells were lysed in 1 ml of a monophasic, 90°C phenol mix containing 45 mM NaCl, 4.5 mM Tris-HCl, pH 7.5, 2.25 mM EDTA, 0.5% sodium dodecyl sulfate (SDS), and 50% acid phenol. After the initial extraction, the aqueous phase was extracted once with 500 μl acid phenol and once with an equal volume of chloroform. The RNA was ethanol precipitated, resuspended in 50 μl of water, denatured by dimethyl sulfoxide and deionized glyoxal, and finally resolved on a 1.4% agarose gel. rRNA was visualized by ethidium bromide staining, α-tubulin mRNA by Northern detection as described above, and SL RNA and U2 snRNA by primer-extension assays as described previously (18). RNA signals were quantified by densitometry using the ImageJ software package (http://rsb.info.nih.gov/ij/).
PTP-RPB6z was tandem affinity purified from crude TbA8 extract exactly as described previously (35). For identification, purified proteins were separated on a 10 to 20% SDS-polyacrylamide gradient gel and stained with Pierce Gelcode Coomassie stain (Pierce, Rockford, IL). Individual protein bands were excised, digested with trypsin, and analyzed by liquid chromatography-tandem mass spectrometry.
Sedimentation of P-RPB6z and copurified proteins in 4 ml 10 to 40% linear sucrose gradients by ultracentrifugation was carried out as described in our previous study (27).
Coimmunoprecipitation of PTP-RPB6z and RPA31-HA in extracts of TbE3 cells was carried out with immunoglobulin G (IgG) beads, which bind the protein A domain of the PTP tag, as detailed before (27). In the reciprocal experiment, the anti-HA antibody failed to precipitate RPA31-HA, possibly because of steric hindrance (data not shown and see the description of the recombinant protein below).
For the expression and purification of TbRPA31 from Escherichia coli, the complete coding sequence of this protein was C-terminally fused to ProtC, a thrombin cleavage site and the six-His tag in the expression vector pET100/ D-TOPO (Invitrogen, Carlsbad, CA). The recombinant protein was expressed in BL21Star (DE3) cells by induction with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) at 37°C for 2 h. Recombinant TbRPA31 used for antibody production was purified by TALON metal affinity chromatography (BD Biosciences) only, due to the extremely low efficiency of anti-ProtC immunoaffinity chromatography for this particular recombinant protein. The eluate from TALON chromatography was nevertheless of very high yield and virtually free of contaminants. The recombinant protein was used for immunizing rats as detailed elsewhere (32).
Nonspecific RNA polymerization assays were performed in 40-μl reaction mixtures of 60 mM sucrose, 20 mM potassium l-glutamate, 20 mM KCl, 3 mM MgCl2, 20 mM HEPES-KOH, pH 7.7, 20 mM creatine phosphate, 0.48 mg/ml of creatine kinase, 2.5% polyethylene glycol, 0.2 mM EDTA, 0.5 mM EGTA, 4 mM dithiothreitol, 10 mg/ml leupeptin, 4 mg/ml aprotinin, 0.5 mM A/C/GTP, 5 μM UTP, 0.25 μCi of [α-32P]UTP (3,000 Ci/mmol), 10 units of RNase inhibitor (Invitrogen), and 4 μg of activated calf thymus DNA (Sigma). Where appropriate, α-amanitin was added to a final concentration of 0.25 mg/ml. Reaction mixtures were incubated at 27°C for 45 min, and RNA was extracted with buffered phenol-chloroform (1:1, vol/vol) and precipitated with ethanol. For DNase and RNase treatments, the RNA from one reaction was split in two halves and incubated for 10 min either with 5 U of RNase-free DNase at 37°C or 0.25 μg of DNase-free RNase at room temperature. RNA samples were resolved on a 6% polyacrylamide-50% urea gel and visualized by autoradiography.
The in vitro transcription system has been described in detail elsewhere (16, 17). Briefly, standard reaction mixtures of 40 μl containing 8 μl of extract, 20 mM potassium l-glutamate, 20 mM KCl, 3 mM MgCl2, 20 mM HEPES-KOH, pH 7.7, 0.5 mM of each nucleoside triphosphate (NTP), 20 mM creatine phosphate, 0.48 mg/ml of creatine kinase, 2.5% polyethylene glycol, 0.2 mM EDTA, 0.5 mM EGTA, 4 mM dithiothreitol, 10 mg/ml leupeptin, 10 mg/ml aprotinin, 12.5 μg/ml vector DNA, 20 μg/ml GPEET-trm, Rib-trm, or VSG-trm template, and 7.5 μg/ml SLins19 template were incubated for 1 h at 27°C and stopped by adding 300 μl of 4 M guanidine thiocyanate, 25 mM sodium citrate, pH 7.0, and 0.5% N-lauroylsarcosine. In reactions with antisera, the volume of the extract was reduced to 4 μl, and the extract was preincubated with 1 μl of antiserum for 30 min on ice before reactions were started by adding templates and nucleotides. To reconstitute RNA pol I activity in RPA1-depleted extract and extract prepared from RPA31-silenced cells, 4 μl settled volume of anti-ProtC beads with bound P-RPB6z from a standard PTP purification equilibrated in transcription buffer (150 mM sucrose, 20 mM potassium l-glutamate, 3 mM MgCl2, 20 mM HEPES-KOH, pH 7.7, 1 mM CaCl2, 0.5 mM dithiothreitol, 0.1% Tween 20, 10 μg/ml leupeptin) were used per reaction. For transcription assays with extracts from RNAi cells, the V.2 versions of GPEET-trm, Rib-trm, VSG-trm, and SLins19 were used. In vitro transcription reactions were repeated at least twice and quantified using ImageJ.
Total RNA preparations of transcription reaction mixtures were analyzed by primer extension of 32P-end-labeled oligonucleotides Tag_PE (17) and SLtag (11), which hybridize to unrelated oligonucleotide tags of the class I and SLins19 RNAs, respectively. Primer extension products were resolved on 6% polyacrylamide-50% urea gels and visualized by autoradiography.
In all eukaryotic RNA polymerases, RPB6 is a subunit essential for assembly, stability, and activity (24, 29). In our preceding work, we showed that T. brucei RPB6z coprecipitated with RNA pol I in an immunoprecipitation assay but was lost when RNA pol I was tandem affinity purified by C-terminal tagging of the largest subunit, RPA1 (27). To establish that RPB6z has an important function similar to its orthologues in other eukaryotes, we silenced the expression of the single RPB6z gene by conditional RNAi in procyclic cells and found that the induction of RPB6z dsRNA synthesis caused a cell growth defect after 2 days and declining cell numbers after 6 days (data not shown). This finding was in accordance with results independently obtained and reported by another research group (7). We therefore concluded that RPB6z is encoded by an essential gene in T. brucei.
Since yeast RPB6 binding to RNA pol I is dependent on the specific subunits RPA14 and RPA43 (29, 37), whose orthologues have not been identified in T. brucei, we speculated that PTP tagging and purification of RPB6z could result in the identification of additional essential RNA pol I subunits. We therefore created the procyclic cell line TbA8, in which one RPB6z allele was knocked out and the second allele fused to the PTP sequence by targeted integration of construct pPURO-PTP-RPB6z (Fig. (Fig.1A).1A). We chose to N-terminally tag RPB6z, because the C terminus of RPB6 proteins is highly conserved (27). The calculated molecular mass of PTP-RPB6z is 32.5 kDa, which corresponded well with the apparent size of the detected protein in whole TbA8 cell lysate (Fig. (Fig.1B).1B). Since TbA8 cell growth was normal, we assumed that the N-terminal tag did not interfere with the function of RPB6z. We therefore continued our analysis and tandem affinity purified PTP-RPB6z by our established PTP protocol (35). The PTP tag consists of a protein A tandem domain and the protein C epitope (ProtC), which are separated by a cleavage site for tobacco etch virus (TEV) protease (35). In consecutive steps, PTP-RPB6z was purified by IgG affinity chromatography, TEV protease elution, and anti-ProtC immunoaffinity chromatography. Immunoblot monitoring of the purification showed that both chromatography steps were highly efficient and that nearly 20% of the RPB6z present in crude extract was isolated (Fig. (Fig.1C,1C, compare lanes 1 and 6). It should be noted that during purification, TEV protease cleavage reduced PTP-RPB6z to P-RPB6z with an apparent size of 19 kDa. Purified protein was eluted from the anti-ProtC column either in the presence of the ProtC peptide (Fig. (Fig.1C,1C, lane 5) or with an EGTA-containing buffer (Fig. (Fig.1C,1C, lane 6). The latter elution was highly efficient, because the interaction of the monoclonal HPC4 antibody with ProtC depends on the presence of calcium cations (39). To detect copurified proteins, the EGTA eluate was resolved by SDS-polyacrylamide gel electrophoresis and stained with Coomassie stain (Fig. (Fig.1D).1D). Surprisingly, RNA pol I was efficiently copurified with PTP-RPB6z, suggesting that tagging of RPB6z stabilized the interaction between this subunit and RNA pol I. Alternatively, it is possible that the RPA1-PTP fusion used in our previous study had a destabilizing effect on this interaction, leading to a loss of RPB6z during tandem affinity purification. The pattern of stained bands was largely congruent to our previous purifications of RPA1-PTP (27, 35) but exhibited subtle differences. The 19-kDa band stained more strongly due to the comigration of subunit RPA12 and P-RPB6z, a fact which we verified by mass spectrometric analysis and two-dimensional gel electrophoresis (data not shown). Furthermore, while in previous purifications very faint dots of 27 kDa and 31 kDa were detected in a two-dimensional gel analysis (27), PTP purification of RPB6z resulted in clearly detectable protein bands of 27, 29, and 31 kDa, indicating that the stable interaction of these proteins with RNA pol I required the presence of RPB6z (Fig. (Fig.1D1D).
Due to the presence of RPB6z and the other proteins, we tested the possibility that this purification yielded active RNA pol I. First, we used either P-RPB6z peptide eluate or the corresponding EGTA eluate dialyzed against transcription buffer in a promoter-dependent transcription assay (see below) but were unable to detect enzyme activity with either eluate (data not shown). However, when purified RNA pol I was kept on the anti-ProtC beads after the washing steps and equilibration in transcription buffer, the enzyme was active. We showed this first by a nonspecific RNA polymerization assay which takes advantage of the fact that, in general, eukaryotic RNA pols initiate transcription nonspecifically at single-stranded DNA nicks. When purified RNA pol I bound to beads was incubated with nicked calf thymus DNA, a mix of unlabeled nucleotides, and [α-32P]UTP, the reactions yielded radiolabeled RNA of high molecular weight (Fig. (Fig.2A).2A). As expected, RNA polymerization was not affected by α-amanitin, an inhibitor of RNA pols II and III.
Furthermore, we employed a promoter-dependent transcription assay to test whether the purified RNA pol I can direct accurate transcription initiation. The in vitro system is based on a crude procyclic T. brucei extract which supports accurate transcription from the structurally different GPEET procyclin, RRNA, and VSG ES class I promoters with different efficiencies (17). For the RNA pol I analysis, we prepared transcription extract from the previously characterized cell line TbF10 (27), which exclusively expresses RPA1 as a C-terminal PTP fusion, and depleted RPA1 from this extract by IgG chromatography (Fig. (Fig.2B).2B). We then carried out in vitro transcription reactions with mock-treated and RNA pol I-depleted extracts with templates GPEET-trm, Rib-trm, and VSG-trm, which harbor the GPEET/PAG3 procyclin (GPEET), the RRNA, and a VSG ES promoter, respectively. In each reaction, the SL RNA gene (SLRNA) promoter template SLins19 was included as a control, because it is transcribed by RNA pol II. All templates carried unrelated oligonucleotide sequences downstream of the transcription initiation site to enable specific detection of the corresponding RNAs by primer extension assays (Fig. (Fig.2C).2C). While the use of mock-treated extract resulted in both class I and SLins19 transcripts, RPA1 depletion specifically abolished the transcription of GPEET-trm, Rib-trm, and VSG-trm (Fig. (Fig.2C,2C, compare lanes 1, 4, and 7 with lanes 2, 5, and 8, respectively), confirming a previous study in which RNA pol I was depleted by anti-ProtC immunoaffinity chromatography (10). When we added back a comparable amount of PTP-RPB6z-purified RNA pol I, class I transcription was fully restored for all three templates (Fig. (Fig.2C,2C, lanes 3, 6, and 9). Restoration of the RNA pol I activity was not an artifact, because adding back beads alone or beads with immobilized RPA1-PTP-purified RNA pol I did not result in detectable class I transcripts (data not shown). Hence, we concluded that RPB6z-PTP-purified RNA pol I is functional in accurate transcription initiation. This conclusion was further supported by the fact that the extract was depleted of RPA1 and reconstituted by an independent RPB6z purification.
The enrichment of p31, p29, and p27 in active RNA pol I preparations suggested that these proteins have important roles in class I transcription. We were able to identify p31 by liquid chromatography- tandem mass spectrometry (GeneDB accession number Tb10.70.3880) (see Fig. Fig.4)4) but not the other two proteins thus far. To confirm correct identification, we HA-tagged p31 C-terminally in TbA8 cells, prepared extract from these cells, and conducted coimmunoprecipitation assays with IgG beads. Immunoblot detection of PTP-RPB6z revealed that the protein was efficiently precipitated under physiological and high-salt conditions, virtually depleting RPB6z from the supernatant (Fig. (Fig.3A).3A). Interestingly, detection of p31-HA showed that this protein coprecipitated with the same efficiency as PTP-RPB6z, suggesting that p31 in extract is quantitatively and stably associated with RPB6z or an RPB6z complex. The interaction between p31 and RPB6z was specific, because the nuclear protein TFIIB (32) did not detectably coprecipitate (Fig. (Fig.3A)3A) and because, as we have shown previously, HA and PTP tags do not form nonspecific interactions with each other (27). Once we confirmed the correct identification of p31, we expressed the complete protein recombinantly in E. coli (data not shown) and, as a valuable tool, raised a polyclonal antibody against this protein in rats. The antiserum is highly specific in recognizing a single protein band of the correct size in whole-cell lysates of procyclic and bloodstream trypanosomes (Fig. (Fig.3B3B).
Walgraffe et al. have previously identified p31 as a substoichiometric component (identified as p30) of a partial purification of the RNA pol I subunit RPA12 (40), and our own analysis independently identified this protein to copurify with RPB6z. While these purification results suggested that p31 is an RNA pol I subunit, there was still a possibility that the protein copurified with RPB6z and RPA12 independently. To demonstrate that p31 is indeed a subunit of RNA pol I, we sedimented purified P-RPB6z eluate in a sucrose gradient by ultracentrifugation. Immunoblotting of proteins from 20 gradient fractions, taken from top to bottom, clearly showed that the majority of p31 cosedimented in fractions 8 and 9 along with subunits RPA1, RPA2, and P-RPB6z (Fig. (Fig.3C).3C). We therefore concluded that p31 is a subunit of T. brucei RNA pol I and consequently dubbed the protein TbRPA31.
RPA31 is conserved among trypanosomatids (Fig. (Fig.4)4) but has no similarity to any other eukaryotic proteins. Moreover, a bioinformatic analysis did not reveal any recognizable sequence motif in this protein (data not shown), confirming the results of Walgraffe et al. (40).
Since RPA31 appeared to be a parasite-specific RNA pol I subunit, we analyzed its importance for trypanosome growth and class I transcription first by silencing RPA31 expression using an inducible RNA interference system that is based on stable transfection of trypanosomes expressing both the tetracycline repressor and T7 RNA polymerase (43). We cloned 544 bp of the RPA31 coding region into the inducible construct pZJM (41), which harbors T7 but no class I promoters, which are typically used for gene expression in trypanosomes, and transfected the resulting construct into procyclic RNAi cells. We analyzed three independently derived cell lines by inducing RPA31 dsRNA synthesis through the addition of doxycycline to the medium (Fig. (Fig.5A).5A). Induced cells exhibited a growth defect within 24 h and ceased growth soon thereafter, indicating that RPA31 is encoded by an essential gene. A Northern blot analysis confirmed efficient silencing of RPA31 expression in these cells (Fig. (Fig.5B).5B). Accordingly, the level of RPA31 in transcription extract prepared from cells in which RPA31 expression had been silenced for 48 h was greatly reduced compared to extract prepared from noninduced cells (see below) (Fig. (Fig.6A).6A). Our attempts to establish an efficient RNAi cell line in bloodstream trypanosomes have not been successful; the cell lines we obtained exhibited only moderate growth defects and did not show significant reduction of RPA31 protein levels upon induction (data not shown). Hence, these experiments were unsuitable for the investigation of the RPA31 silencing effect on class I transcripts in this life cycle stage and paralleled a recent report of similar ineffective silencing of other essential RNA pol I subunits in bloodstream trypanosomes (7). For further analysis, we prepared total RNA from procyclic RNAi cells at different time points postinduction. When equal amounts of total RNA were resolved on an agarose gel and stained with ethidium bromide, the relative amounts of the RNA pol I-synthesized large 18S, 28Sα, and 28Sβ rRNAs, with lengths of 2.25, 1.84, and 1.57 kilonucleotides, respectively (42), clearly decreased (Fig. (Fig.5C).5C). In contrast, Northern blotting of this RNA revealed an increase of α-tubulin mRNA (Fig. (Fig.5C),5C), which was most likely a consequence of the concomitant reduction of the abundant rRNA in these total RNA samples of equal amount. When normalized with the α-tubulin mRNA signals, the large rRNAs decreased to 67%, 39%, and 27% in cells in which RPA31 expression was silenced for 24, 48, and 72 h, respectively. This decrease was specific to rRNA, because SL RNA and U2 snRNA, which in trypanosomes are synthesized by RNA pol II and III, respectively, were not affected and actually exhibited an increase similar to that of α-tubulin mRNA (Fig. (Fig.5C).5C). Together, the specific decline of class I transcripts in RPA31-silenced cells and our finding that RPA31 is a subunit of RNA pol I suggested that RPA31 has an important role in class I transcription.
To test this hypothesis, we went back to the in vitro transcription system and prepared extract from RNAi cells in which RPA31 silencing was induced for 48 h and from noninduced cells. RPA31 silencing reduced the abundance of the protein in extract; when normalized to the abundance of the transcription factor TFIIB, the RPA31 level has dropped to 28% (Fig. (Fig.6A).6A). In vitro reactions with these extracts revealed that RPA31 silencing reduced the GPEET-trm, Rib-trm, and VSG-trm transcription signals, normalized with the corresponding SLins19 transcription signals, to 30% (±6.0%), 7% (±1.2%), and 13% (±1.9%), respectively (Fig. (Fig.6B,6B, compare lanes 1, 5, and 9 with lanes 2, 6, and 10, respectively). When PTP-RPB6z-purified RNA pol I was added to the extract, the transcription signals were partially restored to 72%, 40%, and 79%, respectively (lanes 4, 8, and 12). These effects were specific and independent of the PTP tag, because PTP-purified TRF4, which was previously shown to be active in reconstituting transcription from the SLRNA promoter (36), did not increase the class I transcription signals (lanes 3, 7, and 11). In contrast to purified RNA pol I, the addition of recombinant RPA31 increased class I transcription marginally (data not shown), suggesting that the protein was not efficiently assembled into the RNA pol I enzyme in extract. Together, these findings strongly indicated that a reduction of the RPA31 level directly and specifically affected the efficiency of class I transcription.
To verify the transcriptional role of RPA31 by an independent criterion, we incubated polyclonal immune sera with extract before carrying out in vitro transcription reactions. We hypothesized that binding of the antibodies to their target proteins inhibits their functional roles in the in vitro system. As shown in Fig. Fig.6C,6C, anti-RPA31 serum specifically abolished transcription from all three class I promoters, whereas the corresponding preimmune serum had no such effect (compare lanes 4, 8, and 12 with lanes 3, 7, and 11). For unknown reasons, this particular preimmune serum, in contrast to other pre- and nonimmune sera we have tested previously, slightly affected the control SLRNA transcription, causing in part aberrant transcription initiation from sites upstream of the correct initiation site (Fig. (Fig.6C,6C, ab. SL). To analyze the specificity of this assay, we included an analogously derived immune serum directed against the general class II transcription factor TbTFIIB (32) in this experiment. In reactions with this immune serum, RNA pol I-mediated transcription was largely unaffected, whereas accurate SLRNA transcription was abolished (lanes 2, 6, and 10). Interestingly, anti-TFIIB immune serum had a positive effect on RRNA promoter transcription, a finding which may correlate with the fact that the proximal RRNA promoter domain IV harbors class II promoter elements (34). Together, these results demonstrated the specificity of the inhibitory effects of the antisera and, aside from confirming the essential role of TFIIB in SLRNA transcription (32), they proved that TbRPA31 is essential for RNA pol I-mediated transcription from GPEET, RRNA, and VSG ES promoters in vitro.
In this study, we identified RPA31 as an essential functional component of T. brucei RNA pol I. Tandem affinity purification of the enzyme, combined with a sedimentation analysis, established that RPA31 is a subunit of RNA pol I. In our functional analysis, we demonstrated that silencing of RPA31 expression was lethal to procyclic trypanosomes and specifically reduced the abundance of the large rRNAs. Moreover, our in vitro transcription analysis established by two independent criteria that RPA31 is essential for transcription from the structurally different GPEET, RRNA, and VSG ES class I promoters: (i) extract from RPA31-silenced cells was ineffective in class I transcription, and (ii) addition of anti-RPA31 immune serum to extract from wild-type cells completely blocked the activity of RNA pol I. Although we were unable to establish effective RPA31 silencing in bloodstream parasites, the fact that RPA31 is expressed in this life cycle stage indicates that this RNA pol I subunit is of equal importance to the infective form of mammals.
While there are orthologues of T. brucei RPA31 in trypanosomatid parasites (Fig. (Fig.4)4) (40), we were unable to identify sequence similarity to proteins of other eukaryotes or a conserved sequence motif. RNA pol I has been studied in great detail in the yeast Saccharomyces cerevisiae, where it is composed of 14 subunits. Two of those subunits, RPA49 and RPA34, have no counterparts in RNA pols II and III, they are not essential for yeast viability, and they are dispensable for basal transcription activity (4). Moreover, these proteins do not interact with yeast RPB6, also known as ABC23. In contrast, TbRPA31 is essential, quantitatively associated with RPB6z (Fig. (Fig.3A),3A), and therefore clearly detectable in PTP-RPB6z (this study) but not in RPA1-PTP purifications which lack RPB6z (27). Hence, it is very unlikely that TbRPA31 is a counterpart of yeast RPA49 or RPA34.
In the T. brucei genome, orthologues of 10 of the remaining 12 yeast subunits were identified and, with the exception of RPB12, the presence of these proteins in RNA pol I complexes was demonstrated (14, 27, 33, 40). Hence, it is possible that TbRPA31 is a counterpart of the remaining yeast subunits RPA43 or RPA14. Interestingly, these two subunits form a heterodimer and bind to RPB6 (29), which is consistent with the observed interaction of TbRPA31 and RPB6z. Although sequences of yeast RPA14 homologues, which include the paralogues RPB4 and RPC17 and the archaeal RNA pol subunit RpoF, are not conserved, these proteins share an HRDC-like domain at their C terminus, which includes a PP motif (21). The trypanosomatid RPA31 orthologues, however, differ strongly at their C termini, and a sequence alignment did not identify a conserved proline residue in their C-terminal halves (data not shown). Moreover, and in contrast to yeast RPA14, TbRPA31 is an essential protein and required for promoter-dependent transcription. While TbRPA31 does share the latter property with yeast RPA43, alignment of yeast RPA43 and trypanosomatid RPA31 sequences was ambiguous and resulted in similarity values of less than 20% (data not shown). In addition, while yeast RPA43 and its RNA pol II-specific paralogue RPB7 exhibit some degree of sequence conservation (21), there is no similarity between TbRPA31 and TbRPB7 sequences (data not shown). In accordance with these results, a structural similarity between TbRPA31 and the resolved structure of yeast RPB7 using the PHYRE Protein Folding Recognition Server could not be identified (data not shown). While these considerations do not rule out that TbRPA31 is a functional homologue of either RPA43 or RPA14, they indicate that TbRPA31 is a novel RNA pol I subunit. If this is the case, TbRPA31 would be the first identified parasite-specific protein essential for the expression of T. brucei's major cell surface proteins procyclin and VSG.
The second major achievement in our study was the purification of RNA pol I, which was active in a nonspecific RNA polymerization assay and efficiently reconstituted accurate transcription initiation at three different class I promoters in an RPA1-depleted extract. Of note, this is the first direct demonstration of RNA pol I transcription initiation at the procyclin gene and VSG ES promoters, which verifies the indirect evidence previously obtained by RNA pol inhibitor studies (6, 15, 30), by depletion of RNA pol I from an in vitro transcription reaction (10), and by RNAi-mediated silencing of TbRPA1 (10).
Currently, we do not understand why RNA pol I activity was lost upon dissociation from the anti-ProtC beads. Although we have to assume that an essential subunit dissociated from the enzyme complex upon protein elution, the sedimentation analysis showed that RPB6z and RPA31 cosedimented with RPA1 and RPA2 (Fig. (Fig.3C).3C). Interestingly, RNA pol I in extract sediments substantially faster than purified RNA pol I (fraction 12 versus fractions 8/9) (see reference 27). Since the only additional proteins we have detected in the PTP-RPB6z purification have apparent sizes of 27 and 29 kDa, it is unlikely that these proteins alone account for the observed difference. On the other hand, when purified RNA pol I was added back to depleted extract, the transcriptional activity was fully restored (Fig. (Fig.2C).2C). Consequently, additional proteins interacting with RNA pol I in extract are either not important for transcriptional activity or are important, but they dissociated from RNA pol I during IgG chromatography, remained in the extract, and complemented purified RNA pol I when the enzyme was added back. Aside from the missing RNA pol I subunit orthologues discussed above, these proteins may include the transcription factor yeast Rrn3p/mammalian TIF-IA (2, 22, 23, 25) or other parasite-specific proteins. Fractionation of the in vitro transcription activity and/or tagging and purification of different RNA pol I subunits may lead to the identification of these proteins.
The purification of an active form of RNA pol I from T. brucei, the identification of a novel subunit essential for multifunctional class I transcription in this parasite, and the development of technology to test for RNA pol I activity are achievements which will facilitate the deciphering and functional analysis of key protein interactions in this important enzyme.
We thank Joost Zomerdijk for advice on purifying active RNA pol I and Mary Ann Gawinowicz (Columbia University, Protein Core Facility) for excellent mass spectrometry.
This work was supported by National Institutes of Health grant AI059377 to A.G.
Published ahead of print on 2 July 2007.