|Home | About | Journals | Submit | Contact Us | Français|
In the protist parasite Trypanosoma brucei, the small nuclear spliced leader (SL) RNA and the large rRNAs are key molecules for mRNA maturation and protein synthesis, respectively. The SL RNA gene (SLRNA) promoter recruits RNA polymerase II and consists of a bipartite upstream sequence element (USE) and an element close to the transcription initiation site. Here, we analyzed the distal part of the ribosomal (RRNA) promoter and identified two sequence blocks which, in reverse orientation, closely resemble the SLRNA USE by both sequence and spacing. A detailed mutational analysis revealed that the ribosomal (r)USE is essential for efficient RRNA transcription in vivo and that it functions in an orientation-dependent manner. Moreover, we showed that USE and rUSE are functionally interchangeable and that rUSE stably interacted with an essential factor of SLRNA transcription. Finally, we demonstrated that the T.brucei homolog of the recently characterized transcription factor p57 of the related organism Leptomonas seymouri specifically bound to USE and rUSE. Since p57 and its T.brucei counterpart are homologous to SNAP50, a component of the human small nuclear RNA gene activation protein complex (SNAPc), both SLRNA and RRNA transcription in T.brucei may depend on a SNAPc-like transcription factor.
Eukaryotic RNA polymerase (pol) I exclusively transcribes the large ribosomal gene unit (RRNA), denoted as class I transcription, and is specifically recruited solely to the RRNA promoter. The parasite Trypanosoma brucei, causative pathogen of African trypanosomiasis, belongs to the protist family Trypanosomatidae and is the only known species which utilizes RNA pol I for both rRNA synthesis and transcription of some of its protein-coding genes, namely those encoding its major surface antigens procyclin and variant surface glycoprotein (VSG). Four different types of class I promoters have been characterized in T.brucei. Besides the RRNA promoter, these are the metacyclic and bloodstream form VSG gene expression site promoters and the promoters of the procyclin transcription units [reviewed in Günzl (1)]. There is no obvious sequence homology among these promoter types, and structurally they fall into two classes. The two expression site promoters are very short, extending only to position –67 relative to the transcription initiation site (2–4), whereas the procyclin gene and RRNA promoters have a four-domain structure extending approximately to position –250 (5–7). The details known about the latter two promoters closely resemble the structure of the RRNA promoter of Saccharomyces cerevisiae. The yeast promoter has been characterized in detail and consists of a proximal core promoter, denoted as domain I, and a bipartite upstream element (USE) comprising domains II and III (8,9). The distal domain IV corresponds to the Reb1p-binding element (Reb1) centered at position –215. In T.brucei, procyclin gene promoter domains I–IV and RRNA promoter domains I and II have been exactly mapped by block substitution analyses (5–7,10) and the presence of RRNA promoter domains III and IV has been indicated by progressive 5′ deletions (7,11). The elements in these two promoters are similar in size to their yeast counterparts and are located at corresponding positions, suggesting that they may be functionally analogous.
Like rRNA, spliced leader (SL) RNA is an essential structural RNA which trypanosomes need continuously in large amounts for protein-coding gene expression. Trypanosoma brucei and related organisms transcribe their protein-coding genes polycistronically, and individual mRNAs are processed from large precursors by trans splicing and polyadenylation. In trans splicing, the 39 nt long SL is cleaved from the 5′ terminus of the SL RNA and fused to the 5′ end of each mRNA. This SL addition trans splicing is an obligatory mRNA processing step in trypanosomes and requires the consumption of one SL RNA molecule for the maturation of one mRNA molecule. Hence, the pathogen crucially depends on strong constitutive SL RNA gene (SLRNA) expression throughout its life cycle. Each trypanosome contains approximately 200 SLRNAs which are organized in tandem repeats of 1.35 kb and which accommodate the high rate of SL RNA synthesis. As has been demonstrated in the related organism Leptomonas seymouri, SLRNA transcription is mediated by RNA pol II (12). The structure of the SLRNA promoter has been meticulously characterized in the three trypanosomatid species: T.brucei (13), L.seymouri (14,15) and Leishmania tarentolae (16). In all three cases, two USEs, here denoted as USE1 and USE2, were essential for SLRNA transcription. The two sequence blocks form a bipartite USE because minimal changes of the distance between the two blocks severely affected transcription efficiency (16,17). In L.seymouri, a transcription factor which binds specifically to the USE has been characterized (18). The factor was termed promoter-binding protein 1 (PBP-1) and shown to be part of a functional SLRNA transcription initiation complex (19). PBP-1 consists of three subunits with apparent Mrs of 57, 46 and 36 kDa. Purification of PBP-1 led to the identification and cloning of two subunits (19). Whereas p46 has no homology to any known transcription factor or to sequences of other trypanosomatid genome databases, p57 is homologous to the SNAP50 subunit of the human small nuclear RNA (snRNA)-activating protein complex (SNAPc). Human SNAPc is an essential factor for RNA pol II- or III-mediated transcription of genes encoding spliceosomal uridylic acid-rich (U) snRNAs [reviewed in Hernandez (20)]. No other function has been reported yet. Appropriately, the trypanosome SL RNA resembles a spliceosomal U snRNA because it has the same size, it is predominantly located in the nucleus (21) and it assembles in a corresponding ribonucleoprotein particle by binding a set of common proteins (22).
In this study, we discovered that T.brucei RRNA promoter domain IV harbors two sequence elements which closely resembled the bipartite SLRNA USE. Astonishingly, this ribosomal (r)USE was essential for efficient RRNA transcription in transiently transfected cells and could be functionally replaced by the SLRNA USE. Furthermore, it specifically bound the T.brucei homolog of SNAP50 (TbSNAP50), suggesting that a SNAPc-like complex is involved in T.brucei class I transcription.
Transcription template constructs SLins19, Rib-trm and GPEET-trm have been described in detail previously (23) as well as SLins19 linker scanner mutations LS –71/–62 and LS –53/–42 (13). Construct RibCAT was made for transient transfection analysis and is a derivative of pJP44, a T.brucei transfection vector, in which the procyclin gene promoter GPEET and flanking regions drive the expression of the chloramphenicol acetyltransferase gene [CAT (5)]. RibCAT was constructed by replacing the GPEET promoter in pJP44 by the RRNA promoter from construct Rib-trm using KpnI and SmaI restriction sites. For the block substitution constructs RibCAT1–6 and constructs RibCAT +4, +11, REV and USE, mutated RRNA promoter sequences were generated by single-step or overlapping PCR and cloned into KpnI–SmaI sites of RibCAT replacing the wild-type promoter. Plasmid TbSNAP50-TAP was used to epitope-tag TbSNAP50 and contains two sequence units. The first unit consists of 462 bp of the 3′-terminal TbSNAP50-coding region followed by the tandem affinity purification (TAP) tag sequence (24) and 482 bp of the 3′-flanking region of TbRPA1, the gene encoding the largest subunit of RNA pol I. An engineered NotI site connects the TbSNAP50-coding region and the TAP tag. As a selectable marker, the second unit harbors the neomycin phosphotransferase gene flanked 5′ and 3′ by the intergenic regions of heat shock protein 70 (HSP70) genes 2 and 3, and of β- and α-tubulin genes, respectively.
Cultivation of procyclic forms of T.brucei brucei strain 427 and preparation of transcription-competent extracts were carried out as described (23). For the generation of cell line TbC8, 10 µg of pTbSNAP50-TAP were linearized with BstBI inside the TbSNAP50-coding region and transfected by electroporation into 108 trypanosomes (25,26). Transfected cells were cloned by limiting dilution and selected at 40 µg/ml of the antibiotic G418. Correct integration of the plasmid at the TbSNAP50 locus was confirmed by Southern blotting of BstNI- and HincII-digested genomic DNA of TbC8 and non-transfected cells (data not shown). Cell line TbD11 was generated in an analogous way expressing the TAP tag at the homolog of the human U1 snRNA-specific 70K protein.
Transient transfection, RNA preparation and RNA detection were essentially conducted as described previously (27). In brief, cells were co-transfected with 50 µg of plasmid RibCAT and 20 µg of construct TU281 containing a tagged U2 snRNA gene (28). At 16 h after transfection, total RNA was isolated and CAT and TU281 RNAs were detected by primer extension of 32P-end-labeled oligonucleotides CAT5 (5′-GCCATTGGGATATATCAACGG-3′) and U2-Btag (5′-GATCCTTGCGGGATCCCG-3′), respectively. Primer extension products were separated on 6% polyacrylamide–50% urea gels, visualized by autoradiography and quantified by densitometry. A minimum of three independent experiments was conducted for each construct.
In vitro transcription assays have been described in detail elsewhere (10,23). A standard in vitro transcription reaction was carried out in a volume of 40 µl for 60 min at 27°C and contained 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 NTP, 20 mM creatine phosphate, 0.48 mg ml–1 of creatine kinase, 2.5% polyethylene glycol, 0.2 mM EDTA, 0.5 mM EGTA, 4 mM dithiothreitol (DTT), 10 mg ml–1 of leupeptin, 10 mg ml–1 of aprotinin and 40 µg ml–1 of template DNA. For competition experiments, linear competitor DNAs were produced by PCR using the High Fidelity Expand System (Roche). For wild-type RRNA and SLRNA competitors, RibCAT and SLins19 served as templates, respectively, while rUSE1-mut was generated with RibCAT4, rUSE2-mut with RibCAT2, USE1mut with SLins19 LS –71/–62, and USE2-mut with SLins19 LS –53/–42. The competitor fragments were separated by agarose gel electrophoresis, and purified from gel slices with the QIAquick gel extraction kit (Qiagen) according to the manufacturer’s protocol. The transcription competition experiments were carried out by pre-incubating competitor DNA, cell extract and reaction components on ice for 15 min before transcription was started by transferring reactions to 27°C and adding template DNA and nucleoside triphosphates. Transcription was terminated after 60 min by addition of guanidinium thiocyanate solution. Subsequently, total RNA was prepared and Rib-trm and SLins19 RNAs were specifically detected by primer extension of 5′-end-labeled oligonucleotides Tag_PE (23) and SLtag (13), respectively. In control reactions, endogenous U2 snRNA was detected with either oligonucleotide U2f or oligonucleotide U2k (29).
Biotinylated promoter DNA fragments were generated by PCR using a 5′-biotinylated sense oligonucleotide. For the generation of fragments GPEET –246/–162, RRNA –257/–162 and rUSE1-mut, the 5′-biotinylated T7 sense oligonucleotide was used which added 41 bp of vector sequence 5′ to the promoter region, whereas for fragments SLRNA –126/–18 and USE1-mut, the biotinylated oligonucleotide SL14 was used which is sense to SLRNA promoter positions 126 to 107 and did not add extra base pairs to the fragment. For each reaction, 500 ng of biotinylated DNA fragments were coupled to 10 µl (100 µg) of RNase-free, paramagnetic M-280 Streptavidin Dynabeads (Dynal) according to the manufacturer’s protocol. Consistently, we observed a DNA binding efficiency of >90% (data not shown). After binding, the beads were equilibrated and blocked for 30 min at room temperature in TK20 buffer (150 mM sucrose, 20 mM HEPES-KOH, pH 7.7, 20 mM potassium l-glutamate, 20 mM KCl, 3 mM MgCl2, 2.5% (w/v) polyethylene glycol, 0.2 mM EDTA, 0.5 mM EGTA, 4 mM DTT, 10 µg ml–1 leupeptin, 10 µg ml–1 aprotinin) containing 5 mg ml–1 bovine serum albumin and 5 mg ml–1 polyvinylpyrrolidone. The beads were washed twice with 0.5 ml of TK20 buffer. For TbSNAP50 binding, the beads were incubated in a 40 µl in vitro transcription reaction first on ice for 15 min and then at 27°C for 15 min. Subsequently, the beads were washed three times with 0.5 ml of TK20 buffer and once with 0.5 ml of TN40 buffer (150 mM sucrose, 20 mM Tris pH 8.0, 40 mM NaCl, 3 mM MgCl2, 0.5 mM DTT, 10 µg ml–1 leupeptin, 10 µg ml–1 aprotinin). For protein elution, beads were resuspended in standard SDS gel loading buffer and incubated for 5 min at 70°C. Half of the eluate was separated on an SDS–8% polyacrylamide gel and electroblotted onto a PVDF membrane. TbSNAP50-TAP was detected by the PAP reagent (Sigma) in combination with the BM Chemiluminescence Blotting substrate (Roche).
In a previous study, we unexpectedly observed that SLRNA transcription in vitro was efficiently competed by a linear RRNA promoter fragment extending from position –257 to –3 (–257/–3), suggesting that the RRNA promoter sequence was able to stably bind a trans-activator of SLRNA transcription (10). This competitive effect was specific for the RRNA promoter fragment and not seen with corresponding fragments of other T.brucei class I promoters (10). When we compared the sequences of RRNA and SLRNA promoters, we found two sequence blocks in the distal part of the RRNA promoter which closely resembled USE1 and USE2 of the SLRNA promoter by both sequence and spacing (Fig. (Fig.1).1). However, in the RRNA promoter, these sequence elements are in opposite orientation to the direction of transcription. In the ribosomal sequence, 13 bp out of 16 bp are identical to USE1 and 6 bp out of 10 bp are identical to USE2. The first 4 bp of USE1 do not match the ribosomal sequence, indicating that they may not be relevant to transcription. The spacing between the two ribosomal sequence blocks is 8 bp and identical to that of USE1 and USE2 in the SLRNA promoter. Hence, we named these two elements according to their SLRNA counterparts as rUSE1 and rUSE2 (Fig. (Fig.1).1). The sequence consensus is confined to these two blocks and not present in the spacer or flanking regions. We could not detect any similarity to the SL RNA sequence in the ribosomal spacer, indicating that it does not contain an SLRNA-like gene.
To determine whether rUSE1 and rUSE2 are responsible for the competitive effect on SLRNA transcription, we first dissected the RRNA –257/–3 competitor fragment and analyzed its effect on both RRNA and SLRNA transcription. The in vitro transcription system we employed is based on a crude cytoplasmic extract with high non-specific labeling activity. To avoid excessive background labeling, we inserted an unrelated tag sequence into our template constructs downstream of the transcription initiation site and detected transcripts from these templates specifically by primer extension of a 5′-end-labeled oligonucleotide complementary to the tag (23). For RRNA promoter transcription, we used the construct Rib-trm and for SLRNA promoter transcription the construct SLins19 (23). Rib-trm and SLins19 RNAs with correct 5′ ends gave rise to 127 and 71 nt primer extension products, respectively (Fig. (Fig.2).2). As previously observed, competition of Rib-trm transcription with a 10-fold molar excess of the linear DNA promoter fragment RRNA –257/–3 strongly reduced the Rib-trm transcription signal (Fig. (Fig.2,2, compare lanes 1 and 2). The competitor fragment was then split in two. Fragment RRNA –162/–3 spanning promoter domains I–III competed nearly as efficiently as the full-length competitor, whereas fragment RRNA –257/–162, containing the bipartite rUSE, had no detectable effect on RRNA transcription (Fig. (Fig.2,2, lanes 3 and 4). The latter result was expected because we had shown that RRNA promoter domain IV is dispensable for transcription in vitro [(23) and data not shown]. When SLRNA transcription was competed with these DNA fragments, the opposite result was obtained. Whereas fragment RRNA –257/–162 retained the competitive effect of the full-length competitor, RRNA –162/–3 had no influence on SLRNA transcription, demonstrating that an essential SLRNA trans-activator interacted with RRNA promoter domain IV (Fig. (Fig.2,2, lanes 6–8).
In a second set of experiments, we mutated rUSE1 and rUSE2 in the RRNA –257/–162 competitor fragment and competed SLins19 transcription in vitro (Fig. (Fig.3).3). Mutation of rUSE1 completely abolished the competitive effect of RRNA –257/–167, whereas mutation of rUSE2 retained some of the DNA fragment’s competitive ability, suggesting that rUSE1 is the main sequence element binding to the transcription factor (Fig. (Fig.3B,3B, compare lanes 1–4). We prepared corresponding competitor fragments from the SLRNA promoter to analyze whether USE1 and USE2 were capable of competing SLRNA transcription in a similar fashion. The wild-type competitor fragment SLRNA –126/–18 competed SLins19 transcription as efficiently as the wild-type ribosomal fragment (Fig. (Fig.3B,3B, compare lanes 1, 2 and 5), whereas SLRNA –126/–18 competitors with mutations in either USE1 or USE2 did not compete at all (Fig. (Fig.3B,3B, lanes 6 and 7). We have reproducibly seen that mutation of USE2 had a stronger effect on SLRNA transcription competition than mutation in rUSE2, indicating that factor binding is not completely equivalent in the two different promoters. Taken together, these data demonstrated that USE and rUSE stably interacted with and sequestered a trans-activating factor of SLRNA transcription.
Next, we asked whether rUSE1 and rUSE2 function in RRNA transcription. Mutation of these sequence elements, however, did not detectably affect in vitro transcription of the construct Rib-trm (data not shown). This result was in accordance with our previous finding that deletion of the distal RRNA promoter portion did not significantly affect in vitro transcription efficiency (23). In contrast, deletion of RRNA promoter domain IV to position –181 reduced transient reporter gene expression to 14% in a different study (30). Accordingly, we investigated whether rUSE is responsible for this effect. Since in trypanosomes expression of protein-coding genes requires a splice site and a polyadenylation signal we made the construct RibCAT in which the RRNA promoter drives CAT expression and the procyclin gene-flanking regions provide the RNA processing signals. To analyze the complete domain IV of the RRNA promoter, we mutated the region from position –257 to –182 in six adjacent blocks (Fig. (Fig.4A).4A). The constructs were transiently transfected into procyclic trypanosomes and CAT expression was analyzed at the level of mRNA by primer extension assays using a 5′-end-labeled, CAT-specific oligonucleotide. To control transfection, RNA preparation and primer extension efficiencies in our experiments, we co-transfected a tagged version of the U2 snRNA gene and analyzed its expression correspondingly. Mutation of rUSE2 in plasmid RibCAT2 reduced the CAT mRNA level in comparison with experiments with the wild-type construct to 23% and the two block mutations of rUSE1 in constructs RibCAT4 and RibCAT5 diminished the level to 10 and 9%, respectively (Fig. (Fig.4B4B and C). These results clearly demonstrated that both rUSE1 and rUSE2 were essential for efficient transcription of the ribosomal gene unit inside the cells. In contrast, mutation of the rUSE spacer and flanking regions did not strongly affect CAT expression. Hence, the promoter-relevant sequences in RRNA promoter domain IV are confined to rUSE1 and rUSE2.
Having identified the bipartite rUSE as an important RRNA promoter element, we altered promoter domain IV in construct RibCAT in several ways to learn more about its operating mode (Fig. (Fig.5A).5A). First, we increased the distance between rUSE and the transcription initiation site by 4 and 11 bp. In transient transfection experiments, these manipulations did not interfere with RRNA promoter function and even increased RRNA promoter-driven CAT expression to some extent, suggesting that there is some flexibility in the position of rUSE (Fig. (Fig.5B,5B, lanes WT, +4 and +11). Furthermore, we made construct RibCAT-REV in which rUSE was replaced by its reverse complement. This manipulation reduced CAT expression to 26% of the wild-type level (Fig. (Fig.5B5B and C, lane REV). Since this reduction is in the range of what was observed with mutating rUSE2 (Fig. (Fig.4),4), we concluded that rUSE functions in an orientation-dependent manner. Finally, the sequence homology between rUSE and USE, and the capability of rUSE to sequester an essential SLRNA transcription factor in vitro suggested that the two bipartite promoter elements are functionally equivalent. To test this hypothesis in the RRNA promoter, we substituted rUSE by USE in the construct RibCAT-USE (Fig. (Fig.5A).5A). Astonishingly, the SLRNA promoter element was able to functionally replace rUSE in the RRNA promoter and even increased CAT expression significantly above the wild-type level (Fig. (Fig.5B5B and C, lane USE). We therefore concluded that USE has the functional property of rUSE. To the best of our knowledge, this is the first report of a small RNA gene promoter element functioning in eukaryotic class I transcription.
We also conducted the reciprocal experiment and investigated whether rUSE can functionally replace USE in the SLRNA promoter. In Figure Figure6,6, in vitro transcription results of the unaltered SLins19 construct and three derivatives are shown. The latter comprised linker scanner mutations of USE1 and USE2, and the construct SLins19-rUSE in which USE was replaced by its ribosomal counterpart. The two linker scanner mutations had been analyzed before in a nuclear extract and dramatically reduced SLRNA transcription efficiency, with LS –53/–42, the construct harboring the USE2 mutation, having an even stronger effect than the USE1 mutation LS –71/–62 (13). For comparison, these constructs were re-tested in our cytoplasmic transcription extract and revealed results similar to those in the previous study. Compared with the wild-type signal, mutation of USE1 caused a reduction of the transcription signal by 82%, whereas mutation of USE2 nearly abolished transcription (Fig. (Fig.6,6, compare lane 1 with lanes 2 and 3). In construct SLins19-rUSE, the ribosomal sequence was able to promote significantly more SLRNA transcription than either USE mutation (P < 0.005, t-test), but 66.3% less than the wild-type promoter, revealing that rUSE can only partially replace USE in the SLRNA promoter, again indicating that the interaction of the trans-activating factor with the RRNA and the SLRNA promoter is not equivalent.
The data obtained thus far suggested that the same factor interacted with both SLRNA promoter and RRNA promoter. Characterization of the L.seymouri SNAP50 homolog as a component of an SLRNA transcription factor binding to USE1 (19) enabled us to directly test this hypothesis. As a prerequisite to study the T.brucei SNAP50 homolog (TbSNAP50), we isolated, cloned and sequenced its complete cDNA (GenBank accession no. AJ581666). The sequence consists of the SL, a 94 bp 5′-untranslated region (UTR), 1347 bp of coding region, 484 bp of 3′ UTR and the poly(A) tail. The encoded protein comprises 448 amino acids with a predicted Mr of 51 kDa and a theoretical pI of 5.58. A pair-wise alignment using the program ClustalW (31) revealed that TbSNAP50 is 38% identical and 57% similar to its homolog in L.seymouri and 19% identical and 40% similar to its human counterpart [(19) and data not shown]. Furthermore, based on eight different digests of genomic DNA, Southern analysis showed that TbSNAP50 is encoded by a single copy gene (data not shown). As a tool to investigate binding of TbSNAP50 to SLRNA and RRNA promoters, we epitope-tagged TbSNAP50 C-terminally with the TAP tag (24). Tagging was achieved by targeted insertion of construct pTbSNAP50-TAP into one TbSNAP50 allele (Fig. (Fig.7A).7A). The calculated mass of TAP-tagged TbSNAP50 is 72 kDa, and immunoblot analysis with a peroxidase-labeled IgG domain recognizing the protein A epitopes within the TAP tag specifically detected a polypeptide of this size in cell line TbC8 (Fig. (Fig.7B,7B, compare lanes 1 and 2). We then prepared a transcription extract from TbC8 cells and employed a pull-down assay using immobilized promoter DNA fragments to analyze binding of TbSNAP50 to USE and rUSE. In a negative control, we showed that procyclin gene promoter domain IV which did not compete SLRNA transcription in a previous study (10) was unable to bind TbSNAP50 in the pull-down assay (Fig. (Fig.7B,7B, lane 3). In contrast, tagged TbSNAP50 specifically bound to the upstream region of the SLRNA promoter and domain IV of the RRNA promoter (Fig. (Fig.7B,7B, lanes 4 and 6). In L.seymouri, the transcription factor harboring the SNAP50 homolog p57 binds to USE1. In accordance with this observation, mutation of USE1 and rUSE1 abolished binding of TbSNAP50 to the SLRNA and RRNA promoter fragments, respectively (Fig. (Fig.7B,7B, lanes 5 and 7). In a control experiment, cell line TbD11 was generated in which the TAP tag was fused to an snRNP-specific protein with no known transcriptional function. Pull-down assays with TbD11 extract revealed that the tagged protein did not bind to RRNA and SLRNA promoter DNA, excluding the possibility that the TAP tag is responsible for the observed TbSNAP50-binding phenotype (data not shown). Furthermore, we have reproducibly seen that the ribosomal DNA bound TbSNAP50 more efficiently than the SLRNA fragment, which is in contrast to the transcription competition assays where both promoter DNAs competed SLRNA transcription equally well. An explanation for this discrepancy may be that in the SLRNA fragment, the distance between the immobilizing biotin group and USE is shorter and that in this fragment USE1 is located towards the biotin group, possibly causing a steric problem for efficient binding of a multi-subunit transcription complex. Nevertheless, our results clearly showed that TbSNAP50 specifically binds to the USE of the SLRNA promoter and, in addition, to the rUSE of the RRNA promoter. Taking into account that RRNA promoter domain IV is capable of efficiently competing SLRNA transcription, this finding strongly indicates that in T.brucei synthesis of both rRNA and SL RNA depends on the same SNAPc-like transcription factor.
We have found that the distal part of the T.brucei RRNA promoter contains the bipartite sequence element rUSE which closely resembles the USE of the SLRNA promoter and which binds an essential SLRNA transcription factor in vitro. Although rUSE and USE apparently bind the same transcription factor, these elements serve different functions in their respective promoters. In the SLRNA promoter, USE is required for efficient SLRNA transcription initiation in vivo and in vitro, indicating that this bipartite element is directly involved in the formation of a transcription initiation complex (13). This hypothesis is supported by the finding that changing the USE location affected the site of transcription initiation (16,17). Conversely, in the RRNA promoter, rUSE is dispensable for in vitro transcription, suggesting that this element and its binding factor facilitate transcription initiation indirectly. At a similar distance from the transcription initiation site, the RRNA promoters of yeast and vertebrates harbor an element which has the same property as rUSE. In yeast, the corresponding element Reb1 is only necessary for efficient transcription within its chromosomal context (32). Investigation of Reb1p binding revealed that the protein protects only 15–20 nt from enzymatic or chemical degradation but clears ~200 bp from nucleosomes (33), indicating that Reb1/Reb1p is involved in chromatin remodeling. In vertebrate RRNA promoters, an element termed the proximal terminator T0 is located between positions –150 and –200. The function of T0 has been meticulously investigated in the mouse system where it has been shown to bind the transcription termination factor I [TTF-1 (34)] By analyzing transcription of naked DNA and pre-assembled chromatin templates in vitro, it was elegantly demonstrated that binding of TTF-1 to T0 induced chromatin remodeling and relieved transcriptional repression (35). Analogously, chromatin remodeling may be the predominant function of RRNA promoter domain IV in T.brucei.
Strikingly, both the yeast Reb1 and the mouse T0 sequences are RNA pol I transcription termination signals also present at the 3′ end of the ribosomal transcription unit. It has been shown that RNA pol I-mediated read-through transcription initiated upstream of an RRNA promoter at so-called spacer promoters can displace assembled transcription factors from the actual promoter (36). In vertebrate systems, T0 protects the promoter by terminating transcription of incoming RNA pol I (37). However, in trypanosomes, the region upstream of the RRNA promoter is transcriptionally silent and it remains to be determined whether rUSE functions in RNA pol I transcription termination. The presence of the terminator sequences at both promoter and 3′ end regions has led to the hypothesis that they functionally link transcription initiation and termination possibly through DNA looping. This DNA configuration should then facilitate efficient recycling of RNA pol I from the termination to the initiation site. This model is supported by micrographs of chromatin spreads in which active RRNA transcription units have been visualized as loops separated by intergenic spacers (38). Moreover, TTF-1 has the property to induce DNA looping because it can oligomerize and simultaneously interact with two separate DNA fragments containing its binding site (39). To function in a similar manner, the T.brucei rUSE would have to be located at the 3′ end of the RRNA transcription unit which has been mapped by nuclear run-on assays and an S1 nuclease protection analysis just downstream of the 3′-terminal coding sequences (40). Thus far, we were unable to identify a motif similar to the rUSE/USE consensus sequence at the putative termination region or within the whole RRNA repeat. The information about the nucleotides required for binding the TbSNAP50 complex is limited and the binding motif may be too degenerate for detection by sequence comparison. However, it will be possible to employ SLRNA transcription competition assays using putative RRNA terminator sequences as competitors to determine whether the TbSNAP50 factor binds downstream of the RRNA transcription unit.
rUSE is in opposite orientation to the transcription direction of USE in the SLRNA promoter and its function is orientation dependent. This arrangement of rUSE is reminiscent of the conserved head-to-head organization of tRNA and small RNA genes in trypanosomatids [(41), reviewed in Günzl (1) and Nakaar et al. (42)]. In these gene associations, the tRNA gene-internal A and B box promoter elements are important for both tRNA and small RNA gene transcription. A detailed study of the associated U6 snRNA and threonine tRNA genes in T.brucei revealed that the A box which binds TFIIIB and recruits RNA pol III to the tRNA gene is essential for U6 snRNA gene transcription in vivo and in vitro, and that the position of this element relative to the U6 transcription initiation site is critical (43). Conversely, the B box, which in other systems was shown to be involved in chromatin remodeling, was only essential in vivo and its position relative to the transcription initiation site was flexible. In comparison, rUSE has the properties of the B box and not of the A box, because its function did not depend on its exact location and became apparent only in vivo. At the threonine tRNA/U6 snRNA gene locus, transcription is initiated in both directions, suggesting that this may also be the case in the RRNA promoter. However, it was shown previously that the RRNA spacer region is transcriptionally silent (40) and, accordingly, we were unable to detect a specific transcript from this region by employing northern blotting and RT–PCR (data not shown).
Unlike the two types of VSG expression site promoters which are very short and do not contain a promoter domain IV, the procyclin gene promoter has a four-domain structure (1). Procyclin gene promoter domain IV appears to be functionally equivalent to its RRNA counterpart because it is required in vivo but not in vitro and it resides at a similar position (6,10). Moreover, procyclin–RRNA hybrid promoters were fully functional in transient reporter gene assays, indicating that procyclin gene promoter domain IV can functionally substitute its RRNA counterpart (7). However, the procyclin gene promoter does not interact detectably with TbSNAP50 in our assays and most probably binds a different trans-acting factor. It is thus possible that the parasite utilizes domain IV of the RRNA and procyclin gene promoters to differently regulate transcription of these gene units. For example, while rUSE may promote constitutive RRNA expression in different life cycle stages, its procyclin counterpart may be responsible for the known up- and downregulation of procyclin gene transcription in procyclic and bloodstream form trypanosomes, respectively (44).
Finally, the question arises of how conserved is rUSE among trypanosomatid organisms. The RRNA promoter of Leishmania donovani has been investigated in detail and, surprisingly, it structurally resembles the bloodstream form VSG expression site promoter and not the RRNA promoter of T.brucei (45). It has the same small size and two-domain structure upstream of the transcription initiation site, and there is no indication of an rUSE element. In other trypanosomatid species with known promoter sequences, there is no convincing sequence conservation between SLRNA and RRNA promoters, and it remains to be determined experimentally whether rUSE is a common feature among trypanosomatids. Alternatively, rUSE may be a T.brucei invention to facilitate differential regulation of multi-functional class I transcription.
In conclusion, our data strongly suggest that in T.brucei, a SNAPc-like transcription factor is essential for both RRNA and SLRNA transcription. Intriguingly, such a factor would enable the parasite to regulate global gene expression simultaneously at the level of protein synthesis and RNA maturation. We have initiated in vivo experiments to investigate these possibilities and begun to purify and functionally characterize the TbSNAP50 complex.
This work was supported by a grant from the German Research Foundation (DFG) to A.G. (Gu 371) and an institutional fund of the University of Connecticut Health Center.
DDBJ/EMBL/GenBank accession no. AJ581666