In Vitro Analysis of α-Amanitin-Resistant Transcription from the rRNA, Procyclic Acidic Repetitive Protein, and Variant Surface Glycoprotein Gene Promoters in Trypanosoma brucei

Abstract

In Trypanosoma brucei, transcription resistant to the mushroom toxin α-amanitin is not restricted to the rRNA genes (rDNA), as in higher eukaryotes, but extends to genes encoding the major cell surface proteins variant surface glycoprotein (VSG) and procyclin or procyclic acidic repetitive protein (PARP). Here, we report the development of a homologous cell extract from procyclic T. brucei cells in which rDNA and PARP A and VSG gene promoters drive efficient, accurate, and α-amanitin-resistant transcription. A comparative analysis revealed that transcription from the three promoters generally required identical reaction conditions for maximal efficiency. Nevertheless, PARP promoter transcription proved to be exceptional by its high efficiency, its lag phase, a high template DNA concentration optimum, and its tolerance to increasing concentrations of Mn²⁺. Mutational analysis for both the PARP and rDNA promoters showed that the proximal and distal core elements were essential for efficient transcription in vitro. Deletion of the upstream control regions (UCRs), however, had a different effect. Whereas PARP UCR deletion reduced transcription efficiency almost 10-fold, deletion of the rDNA UCR had only a minor effect on transcription efficiency.

The cell surface coat of Trypanosoma brucei consists mostly of identical protein molecules. In the procyclic insect form trypanosome, the major surface protein is procyclin or procyclic acidic repetitive protein (PARP), and in the bloodstream form trypanosome it is 1 of about 1,000 variant surface glycoproteins (VSG). The importance of the surface proteins to the parasite is emphasized by the fact that they contribute about 1 to 10% of the total amount of protein in T. brucei cells (7, 10, 12). Moreover, in the bloodstream of its mammalian host, T. brucei evades the immune system by antigenic variation, a process in which trypanosomes change their surface coat by expressing a different VSG gene (recently reviewed in references 3, 28, and 38). VSG and PARP gene expression has been a focus of research in the past (reviewed in reference 24). One outstanding finding was that transcription of these genes is, in contrast to most other protein-coding genes of T. brucei, resistant to α-amanitin, a characteristic of RNA polymerase I (Pol I)-mediated transcription (9, 21, 32). In higher eukaryotes, including the hosts of T. brucei, RNA Pol I transcribes only the polycistronic gene unit encoding the large rRNAs, while RNA Pol II transcribes all protein-coding genes. The reason for this clear distinction appears to be the association of RNA Pol II with guanylyltransferase, which adds the essential guanosine cap to the 5′ end of each mRNA molecule. In T. brucei and related organisms, however, nuclear mRNA capping occurs via RNA trans splicing, a process in which the short and capped 5′-terminal spliced leader (SL) is cleaved from its donor, the SL RNA, and is fused to the 5′ end of each mRNA molecule. Transcription of protein-coding genes is thus not limited to RNA Pol II. Although the final proof of RNA Pol I transcription of PARP and VSG genes is still missing, a vast body of evidence which supports this hypothesis has been collected (reviewed in reference 24).

In T. brucei, the promoters for the PARP, VSG, and rRNA genes (rDNA) have been characterized in detail by transfection studies (4, 19, 29, 35, 39). Although there is no obvious sequence homology among the promoters, they do have a similar structure. Each promoter has a bipartite core region with two separate elements within the 70-bp region just upstream of the transcription initiation site. Whereas in the VSG promoter the two elements suffice for full transcriptional activity, the rDNA and PARP promoters require in addition an upstream control region (UCR) extending approximately to position −250 relative to the transcription initiation site. In T. brucei cell extracts, specific single-stranded or double-stranded DNA-protein complexes were characterized for all three core promoters (5, 20, 30, 39). Thus far, however, an in vitro transcription system which would facilitate a functional analysis of trans-acting factors has not been established. Recently, such systems were developed for the SL RNA and U2 and U6 snRNA genes in T. brucei, Leptomonas seymouri, and Leishmania tarentolae (14, 15, 18, 25, 36).

In this study, we report accurate, efficient, and α-amanitin-resistant transcription of gene constructs containing either an rDNA or the PARP A or a VSG gene expression site promoter in a homologous cell system based on an extract of procyclic T. brucei cells. A comparison of transcriptions from the three promoters revealed that, with the exception of the template DNA concentration, maximal efficiencies were achieved under identical reaction conditions. Nevertheless, PARP promoter transcription was different. Compared to rDNA and VSG promoter transcription, it was more efficient, required a higher DNA template concentration to reach optimal efficiency, exhibited a lag phase in the kinetic analysis, and was tolerant to elevated Mn²⁺ concentrations. Moreover, efficient transcription in vitro from the PARP promoter depended strongly on its UCR, while the rDNA UCR only had a minor influence on transcription efficiency.

MATERIALS AND METHODS

DNA oligonucleotides and plasmid construction.

The following DNA oligonucleotides were used: minirib1, 5′-GGGGCGACGACATAAACGCGC-3′; minirib2, 5′-GCGGATCCGAGTGAATGATGATAGATTTGAAGCTTCCCGTATCGCATTGCGC-3′; parptrm1, 5′-GCGGATCCTTCTGTGCCCATCACTGG-3′; parptrm2, 5′-GCTCGAGTGATATTTGTTTG-3′; VSG1, 5′-GCCGGTACCCCATCCAAGCGGAATAAC-3′; VSG2, 5′-GGGGACGGCACGCTCAGGCCAA-3′; Tag_PE, 5′-GAGTGAATGATGATAGATTTG-3′; and SLtag (15). The gene constructs are derivatives of the previously described vector pHD50, in which the rDNA promoter is connected to a reporter gene (19). For generation of the construct Rib-trm, the reporter gene of pHD50 between restriction sites SmaI and BamHI was replaced by a small rDNA fragment which was amplified from T. brucei genomic DNA with oligonucleotides minirib1 and minirib2. This restored the SmaI site of pHD50, extended the transcribed rDNA region to position +100, and added a HindIII restriction site and a 21-bp tag to the ribosomal sequence. In a second cloning step, a 1,067-bp DNA fragment of the putative PARP gene termination region (positions 3649 to 4715 [1]) was amplified by PCR from genomic DNA with oligonucleotides parptrm1 and parptrm2 and was inserted after the tag sequence. To generate the construct PARP-trm, the rRNA promoter of Rib-trm was replaced with the PARP A gene promoter of construct pJP44 (35) by using restriction enzymes KpnI and SmaI. For the construct VSG-trm, an expression site promoter was amplified from genomic DNA by PCR with oligonucleotides VSG1 and VSG2 and cloned analogously. Except for the two single base pair changes G-48 to A and T-81 to G, the VSG-trm promoter is identical to the 221 expression site promoter (42). In the construct PARP-trm/25.35 positions −71 to −62 of the PARP A promoter were replaced with 5′-TCAGAGGCCT-3′, and in the construct PARP-trm/44.38 positions −41 to −32 were replaced with 5′-AGAGGCCTAG-3′. In constructs Rib-trm/153 and Rib-trm/156, rDNA promoter positions −62 to −52 and −32 to −22 were mutated to 5′-CGGAGGCCTA-3′ and 5′-CTCGAGTCGT-3′, respectively.

Trypanosome culture and extract preparation.

Cultures of the procyclic form of T. brucei brucei strain 427 (11) were grown at 28°C in SDM-79 medium (6) supplemented with 10% (vol/vol) fetal calf serum and 5 mg of hemin/liter. For a standard preparation, a 2-liter culture of procyclic cells was grown in a 5-liter flat-bottom, spherical glass flask under constant stirring to a density of 10⁷ cells per ml. Cells were harvested at 4°C, yielding a packed cell volume of around 3 ml. The cell pellet was washed twice in 10 ml of ice-cold wash solution (20 mM Tris-HCl [pH 7.4], 100 mM NaCl, 3 mM MgCl₂, 1 mM EDTA) and once in 10 ml of ice-cold transcription buffer (150 mM sucrose, 20 mM potassium l-glutamate, 3 mM MgCl₂, 20 mM HEPES-KOH [pH 7.7], 2 mM dithiothreitol, leupeptin [10 μg/ml]). The pellet was finally resuspended in a 1.5 times packed cell volume of transcription buffer and incubated for 10 min on ice. Cells were broken in a 7-ml Dounce homogenizer with a type A pestle by applying rapid strokes continuously for about 5 min, until more than 75% of the cells were broken. This suspension was aliquoted in 900-μl portions, shock-frozen in liquid nitrogen, and stored at −70°C. For a whole-cell extract preparation, a 900-μl aliquot was thawed, mixed with 100 μl of transcription buffer containing 1.5 M KCl, and incubated for 20 min on ice. Subsequently, the extract was spun at 21,000 × g for 10 min at 4°C. The supernatant was transferred to a new reaction tube, diluted with a 0.5 volume of transcription buffer to reduce the KCl concentration to 100 mM, and concentrated three- to fourfold in a Centricon 10 concentrator (Millipore). The final cell extract was aliquoted, shock-frozen, and stored at −70°C. The protein content of extracts varied between 20 and 30 mg/ml.

In vitro transcription and RNA analysis.

A standard transcription reaction was carried out in a volume of 40 μl containing 8 μl of whole-cell extract. The reaction mixture contained 20 mM potassium l-glutamate, 20 mM KCl, 3 mM MgCl₂, 20 mM HEPES-KOH [pH 7.7], 0.5 mM ATP, 0.5 mM GTP, 0.5 mM CTP, 0.5 mM UTP, 20 mM creatine phosphate, creatine kinase [0.48 mg/ml], 2.5% polyethylene glycol, 0.2 mM EDTA, 0.5 mM EGTA, 4.25 mM dithiothreitol, leupeptin [10 μg/ml], and PARP-trm [40 μg/ml] or Rib-trm or VSG-trm [7.5 μg/ml] template in the form of circular plasmid DNA. The reaction mixture was preincubated on ice for 10 min in the absence of nucleoside triphosphates. Subsequently, the reaction was started by nucleotide addition, incubated at 28°C for 60 min, and stopped with 250 μl of solution D (4 M guanidinium thiocyanate, 25 mM sodium citrate [pH 7.0], 0.5% N-lauroylsarcosine). RNA was prepared by the single-step method of Chomczynski and Sacchi (8). Briefly, reactions were extracted with 1 volume of water-saturated phenol, RNA was precipitated from the aqueous phase with ethanol, and residual amounts of DNA were degraded with DNase I. For the detection of in vitro-synthesized RNA, primer extension reactions of the RNA preparations were carried out with ³²P-end-labeled oligonucleotide Tag_PE and extension products were separated on 6% polyacrylamide–50% urea gels. Transcription signals were visualized by autoradiography and quantitated by the E.A.S.Y Win32 imaging system (Herolab). For the time course experiments, the standard reaction was scaled up 10 times, and 40-μl aliquots were removed at specified time intervals. Cotranscription reaction mixtures contained additional pGSLins19 DNA at a concentration of 5 μg/ml.

RESULTS

In vitro transcription system.

For the development of a homologous in vitro system for α-amanitin-resistant transcription in T. brucei, we chose to use a gene construct with the rDNA promoter, because this promoter is the typical RNA Pol I promoter shared by all eukaryotes and it drives constitutive expression of the rRNA genes in all life cycle stages of T. brucei. The construct Rib-trm contained the rDNA promoter from positions −258 to +100 relative to the transcription start site, an unrelated 21-bp tag sequence, and part of the putative PARP transcription termination region (Fig. 1). The tag sequence was introduced to allow specific detection of the Rib-trm transcript by primer extension of the 5′-end-labeled, complementary oligonucleotide Tag_PE and the terminator sequence to potentially facilitate dissociation and reutilization of RNA Pol I complexes on circular template DNA. To determine the suitability of Rib-trm for in vitro assays, we first analyzed its in vivo expression by transient transfection of procyclic cells. In these experiments, Rib-trm RNA was readily detectable in total RNA preparations by a primer extension signal of correct length, demonstrating that transcription had started at the correct site and that synthesis and stability of Rib-trm RNA were sufficient for detection (data not shown).

FIG. 1.

Schematic drawing (to scale) of Rib-trm, PARP-trm, and VSG-trm gene constructs. Sequences from rDNA and the PARP and VSG genes are represented by striped, open, and checkered boxes, respectively. The tag is drawn as a black rectangle, and its sequence is shown above the drawing of Rib-trm. The putative PARP transcription termination region (trm) is not drawn to scale. Drawings of the three constructs are aligned according to transcription start sites, indicated by flags. Numbers denote positions of promoter 5′ ends relative to transcription start sites, and arrowheads point to unique restriction sites KpnI (K), SmaI (S), HindIII (H), BamHI (B), and XhoI (X).

Next, a homologous cell extract was prepared from procyclic cells competent in transcribing the Rib-trm template in vitro. We decided to generate a whole-cell extract (see Materials and Methods), because in our hands, lysis of T. brucei cells with either a Stansted cell disrupter, a glass Dounce homogenizer, or a French press always leads to strong leakage of nucleoplasm into the cytoplasmic fraction, as judged by the distribution of uridylic acid-rich snRNAs (data not shown). In addition, it is known that in higher eukaryotes the cytoplasm itself contains substantial amounts of RNA Pol I and essential RNA Pol I-specific transcription factors (34). By varying extract preparation and transcription reaction conditions, we were finally able to detect a faint primer extension signal of correct size with the Rib-trm template (an example is shown in Fig. 2, lane 3). Subsequently, constructs PARP-trm and VSG-trm were generated by replacing the rDNA promoter in Rib-trm up to position +32 by the PARP A promoter and a VSG gene expression site promoter (see Materials and Methods), respectively (Fig. 1). The PARP promoter comprised the region from position −254 to +26, and the VSG promoter comprised the region from position −100 to +51. In this way, the tag became located at different distances from the transcription start site, and primer extension products of PARP-trm, Rib-trm, and VSG-trm RNA with the 5′-end-labeled oligonucleotide Tag_PE were expected to have lengths of 121, 127, and 146 nucleotides, respectively. Transcription of PARP-trm and VSG-trm resulted in specific in vitro transcription signals of correct lengths (Fig. 2, lanes 2 and 4). The PARP-trm signal was considerably stronger than those of the other two templates (data not shown; see below). By titrating potassium chloride, magnesium chloride, template DNA, polyethylene glycol, dithiothreitol, and EGTA, we defined the conditions of the standard transcription reactions. Except for the template DNA concentration, transcription of the three templates required the same concentrations of these reaction components for optimal efficiency. However, by varying the template DNA concentration, we found a clear difference: Whereas a maximum of PARP-trm transcription was achieved at a DNA concentration of 40 μg/ml, Rib-trm and VSG-trm transcriptions were most efficient at a DNA concentration of 7.5 μg/ml (data not shown). In Fig. 2, transcriptional signals obtained by primer extension analysis of standard reactions are shown. No signal was detectable in a control reaction with vector DNA as the template (lane 1). In contrast, transcription from the PARP, rDNA, and VSG promoters resulted in each case in one major band of correct size (lanes 2 to 4). An additional minor band was reproducibly seen in reactions with PARP-trm, indicating a second in vitro transcription start site in the PARP promoter located approximately 10 bp further upstream. Since primer extensions of PARP-trm, Rib-trm, and VSG-trm RNA were carried out with the same radiolabeled oligonucleotide, the strengths of the transcription signals could be directly compared. Quantitation of the signals revealed that transcription of PARP-trm was about fourfold more efficient than transcription of Rib-trm or VSG-trm (Fig. 2B). The high efficiency of PARP-trm transcription was not due to the presence of the PARP termination region, because deletion of this sequence in PARP-trm, Rib-trm, and VSG-trm decreased transcription of these constructs uniformly by approximately 50% (data not shown).

FIG. 2.

In vitro transcription of PARP-trm is stronger than that of Rib-trm or VSG-trm. Standard in vitro transcription reactions were carried out with templates PARP-trm, Rib-trm, and VSG-trm and with vector DNA as a control (Ctrl). (A) RNA prepared from these reaction mixtures was analyzed by primer extension of the 5′-end-labeled oligonucleotide Tag_PE. Reaction products were separated on a 6% polyacrylamide–50% urea gel and visualized by autoradiography. Numbers on the left indicate molecular weights (in thousands) of marker fragments. M, marker (MspI-digested pBR322). (B) Diagram of transcription signal strengths obtained with templates PARP-trm, Rib-trm, and VSG-trm in five independent experiments. The strongest signal was given a value of 100. Error bars represent standard deviations.

Effect of RNA Pol inhibitors.

In T. brucei, the effect of α-amanitin on the transcription of various genes has been analyzed by nuclear run-on assays (9, 21, 32), in a permeabilized cell system (37), and for the U2 and SL RNA genes by in vitro transcription experiments (14, 15). Independent of the reaction system used, the observed inhibition pattern was the same: RNA Pol II-mediated transcription of protein-coding genes was very sensitive to α-amanitin even at concentrations of 2 μg/ml; transcription of small RNA genes, including the SL RNA gene, exhibited an intermediate sensitivity, with inhibitory effects seen only at α-amanitin concentrations above 5 μg/ml, and transcription of the rDNA, and VSG and PARP genes was resistant to the toxin. To examine the effect of α-amanitin in our in vitro system, we performed cotranscription experiments in which PARP-trm, Rib-trm, or VSG-trm was mixed with pGSLins19, a construct containing a tagged SL RNA gene (15). This construct was accurately transcribed in the whole-cell extract (data not shown). The transcription reactions were carried out in the presence of α-amanitin at 0, 5, or 200 μg/ml. Transcription of PARP-trm, Rib-trm, and VSG-trm was not affected by α-amanitin at 200 μg/ml (Fig. 3, lanes 4, 7, and 10), and even a 1-mg/ml concentration of α-amanitin did not result in a reduction of transcription efficiency (data not shown). In contrast and as expected, transcription of pGSLins19 was inhibited in an intermediate way: a 5-μg/ml concentration of α-amanitin did not affect SL RNA gene transcription (lanes 3, 6, and 9), whereas a 200-μg/ml concentration of the inhibitor reduced SL RNA synthesis by more than 90% (lanes 4, 7, and 10). We conclude that in vitro transcription of Rib-trm, PARP-trm, and VSG-trm was resistant to α-amanitin and that the inhibition pattern obtained is congruent with that of other reaction systems.

FIG. 3.

Effect of α-amanitin on PARP-trm, VSG-trm, and Rib-trm transcription. In vitro transcription reactions were carried out in the presence of α-amanitin (α-ama) at 0, 5, or 200 μg/ml and contained template mixes of pGSLins19 and PARP-trm (PARP + SL), Rib-trm (Rib + SL), or VSG-trm (VSG + SL). A control reaction was carried out with vector DNA only (lane C). RNAs of these reaction mixtures were extracted and subjected to primer extension analysis with the 5′-end-labeled oligonucleotides SLtag and Tag_PE, which are complementary to the tags of SLins19 and trm RNAs, respectively. Arrows indicate expected lengths of specific VSG-trm (VSG), Rib-trm (Rib), PARP-trm (PARP), and SLins19 products. M, marker (MspI-digested pBR322).

In nuclear run-on assays, transcription elongation on rRNA, PARP, and VSG genes but not on other genes was insensitive to the detergent Sarkosyl at a concentration of 0.6% (33). In contrast, addition of only 0.2% Sarkosyl completely abolished in vitro transcription of the three genes and of the SL RNA gene construct (data not shown). This finding is most likely due to differential effects of Sarkosyl on transcriptional initiation and elongation (16, 31).

Mapping of transcription initiation sites.

The initiation sites of PARP-trm, Rib-trm, and VSG-trm transcription were mapped to nucleotide positions by analyzing linear amplification sequencing reactions of the template DNAs in parallel with primer extension reactions of in vitro-synthesized transcripts. For both assays, the same 5′-end-labeled oligonucleotide Tag_PE was used (Fig. 4). The in vivo transcription start site of the PARP promoter was previously mapped within the sequence 5′-GTGAG-3′ to the first (4), the second (27), or the third (35) G residue. The in vitro transcription start site of PARP-trm mapped to the second G (Fig. 4A). To confirm correct transcription initiation, we transiently transfected PARP-trm into procyclic cells and analyzed in vivo- and in vitro-synthesized PARP-trm RNA in parallel. Specific primer extension products of both RNA preparations comigrated and demonstrated that transcription of PARP-trm initiated at the same position in vivo and in vitro (data not shown). The minor transcription signal seen in PARP-trm in vitro reactions (for example, Fig. 2, lane 2) was not reproduced in Fig. 4 but was mapped to the G residue at position −10 with a longer exposure of the autoradiograph. The initiation sites of Rib-trm (Fig. 4B) and VSG-trm (Fig. 4C) corresponded to those mapped previously in vivo on the rDNA and VSG promoters (27, 40), respectively. Hence, we conclude that in vitro transcription from all three promoters initiated accurately.

FIG. 4.

Mapping of in vitro transcription initiation sites. PARP-trm (A), Rib-trm (B), and VSG-trm (C) DNAs were subjected to linear amplification sequencing using the 5′-end-labeled oligonucleotide Tag_PE (lanes G, A, T, and C). Primer extension products obtained with the same oligonucleotide from in vitro-synthesized PARP-trm (PARP), Rib-trm (Rib), and VSG-trm (VSG) RNA and from RNA of control reactions carried out with vector DNA (Ctrl) were analyzed in parallel on 6% polyacrylamide–50% urea gels. On the right of each panel, the double-stranded sequences surrounding the transcription initiation sites are shown. The sequences on the left are from the transcribed strands and correspond to the sequencing ladders. The start residues at positions +1 are aligned with the sequencing ladders as indicated by the arrows. M, marker (MspI-digested pBR322).

Kinetics of RNA synthesis.

To analyze the reaction kinetics of PARP-trm, Rib-trm, and VSG-trm transcription in vitro, the standard reaction was scaled up 10-fold. Aliquots were taken from these reaction mixtures at given time points, and RNA extractions of these aliquots were analyzed by primer extension assays (Fig. 5). The kinetics of the three transcription reactions were similar in that a linear increase in the transcription signal was observed for 40 to 60 min, reaching a maximal signal strength after 60 min followed by a slow decrease in signal strength after this peak. As a further specific characteristic of PARP-trm transcription, we observed in several independent experiments a lag phase of approximately 10 min. In contrast, Rib-trm and VSG-trm produced a detectable transcription signal soon after the reaction was started (Fig. 5A [compare lanes 5, 10, and 20 min]). The reduction in signal seen with incubations longer than 1 h indicated the presence of low RNase activity, which was most likely of endogenous origin (17).

FIG. 5.

Kinetic analysis of PARP-trm, Rib-trm, and VSG-trm transcription in vitro. (A) Tenfold scale up of the standard transcription reaction was carried out with templates PARP-trm (PARP), Rib-trm (Rib), or VSG-trm (VSG). Aliquots were taken from each reaction mixture after 0, 5, 10, 20, 30, 45, 60, 90, 120, and 150 min of incubation, and RNA was analyzed by primer extension of 5′-end-labeled oligonucleotide Tag_PE. Extension products were separated on 6% polyacrylamide–50% urea gels and visualized by autoradiography. M, marker (MspI-digested pBR322). (B) Quantitation of transcription signals. For each reaction, the strongest signal was set to 100.

Effect of Mn²⁺.

A major argument against RNA Pol I-mediated transcription of VSG genes stems from nuclear run-on experiments with bloodstream form trypanosomes, in which transcription elongation on a VSG gene was resistant to Mn²⁺ concentrations up to 4 mM in contrast to that on rDNA (13). To determine whether in vitro transcription is similarly affected, manganese chloride was titrated into standard reaction mixtures. We found that transcription of both Rib-trm and VSG-trm was sensitive to Mn²⁺ concentrations of 3 mM and higher (Fig. 6). The discrepancy between results from Grondal et al. (13) and ourselves may be explained by a differential effect of Mn²⁺ on transcription initiation and transcription elongation. However, in vitro transcription reactions were performed in an extract prepared from procyclic cells and the nuclear run-on experiments were conducted with nuclei from bloodstream form cells. Thus, it is possible that a factorial difference in procyclic and bloodstream form trypanosomes is responsible for a different effect of Mn²⁺ on transcription of the VSG gene in these two trypanosome life cycle stages. Interestingly, PARP-trm transcription in the procyclic extract was not affected by Mn²⁺ even at a concentration of 4 mM (Fig. 6). This finding supports our previous observations that transcription of PARP-trm differs somehow from that of VSG-trm and Rib-trm, potentially by binding of a PARP gene-specific transcription factor. The differential effect on PARP, rDNA, and VSG promoters was Mn²⁺ specific and not observed in Mg²⁺ titrations. Increasing the concentration of Mg²⁺ up to 10 mM resulted in a slight decrease of transcription from all three promoters (data not shown).

FIG. 6.

Effect of Mn²⁺ on PARP-trm, Rib-trm, and VSG-trm transcription in vitro. Standard transcription reactions were conducted with templates PARP-trm (PARP), Rib-trm (Rib), and VSG-trm (VSG) in the presence of 0, 1, 2, 3, and 4 mM manganese chloride. RNAs were analyzed by primer extension of 5′-end-labeled oligonucleotide Tag_PE.

Effect of promoter mutation.

Previously, effects of 10-bp substitutions of the PARP and rDNA core promoter regions on reporter gene expression were characterized by transient transfection of procyclic T. brucei cells (19, 35). In the PARP promoter analysis, the distal (positions −62 to −71) and proximal (positions −32 to −41) core elements were mutated in the constructs pJP25.35 and pJP44.38, respectively. With pJP25.35, reporter gene expression was reduced to 5%, and with pJP44.38 it was nearly abolished (35). In the rDNA promoter analysis, mutation of the distal core element in pHD153 (positions −53 to −62) and the proximal core element in pHD156 (positions −23 to −32) reduced reporter gene expression to the same extent (19). To determine how these promoter mutations affect transcription in vitro, we replaced the wild-type promoters of PARP-trm and Rib-trm with the mutated promoters. The PARP constructs were accordingly named PARP-trm/25.35 and PARP-trm/44.38, and the rDNA promoter constructs were named Rib-trm/153 and Rib-trm/156 (Fig. 7A). These constructs were analyzed by the cotranscription assay with template pGSLins19. Transcription signals obtained by primer extension analysis from three independent sets of experiments were quantitated and standardized with the SLins19 signals, and one set of experiments is shown in Fig. 7B. In comparison to PARP-trm (lane 2), transcription of PARP-trm/25/35 dropped to 12.8% (standard deviation [s] = 1.4%) (lane 3) and that of PARP-trm/44.38 dropped to background levels (lane 4). A very similar picture was obtained with the ribosomal constructs. Compared to Rib-trm (lane 6), transcription of Rib-trm/153 was diminished to 5.4% (s = 3.8%) (lane 7) and no Rib-trm/156 transcription signal was detectable (lane 8). Hence, core promoter mutations have similar effects on transcription in vivo and in vitro.

FIG. 7.

In vitro transcription analysis of rDNA and PARP promoter mutations. (A) Schematic drawing (not to scale) of the promoter region. Each of the gene constructs PARP-trm and Rib-trm carries two wild-type core promoter elements (solid rectangles) and a less defined UCR (long rectangles). The distal core promoter element was mutated in constructs PARP-trm/25.35 and Rib-trm/153, and the proximal element was mutated in constructs PARP-trm/44.38 and Rib-trm/156 (open rectangles). In constructs PARP-trm/−84 and Rib-trm/−84, promoter 5′ ends were deleted to position −84 relative to the transcription initiation site (flag). (B) In vitro transcription reactions were conducted with PARP and ribosomal template DNAs as indicated, and with vector DNA as the control (Ctrl). In addition, all reaction mixtures contained the pGSLins19 template to standardize transcription efficiency. RNA was analyzed by primer extension of 5′-end-labeled oligonucleotides Tag_PE and SLtag. Arrows point to the expected lengths of PARP-trm (PARP), Rib-trm (Rib), and GSLins19 transcription signals. M, marker (MspI-digested pBR322).

In contrast to VSG genes, efficient in vivo expression of PARP genes and rDNA depends on sequences upstream of the core promoter extending to position −250. Deletion of these UCRs reduced transcriptional activity from PARP and rDNA promoters by about 95% (4, 19, 35, 41). To assess the effect of these regions on transcription in vitro, we deleted the 5′-terminal region of the PARP-trm and Rib-trm promoters to position −84 and generated constructs PARP-trm/−84 and Rib-trm/−84 (Fig. 7A). Transcription of these constructs gave different results. The signal strength of PARP-trm/−84 transcription was only 11% (s = 7.3%) of the PARP-trm signal (Fig. 7B, lane 5), demonstrating that the UCR of the PARP promoter is essential for efficient transcription in vitro. In contrast, deleting the UCR of the rDNA promoter only had a minor effect on transcription efficiency, reducing the signal strength to 88% (s = 5.4%) (lane 9).

In sum, the core promoters of the PARP gene and rDNA were essential for in vitro transcription, and the mutational effects correlated with those determined in vivo. Whereas the UCR of the PARP promoter stimulated in vitro transcription about ninefold, the UCR of the rDNA promoter exhibited only a minor positive effect.

DISCUSSION

We have established and characterized an in vitro system for T. brucei in which rDNA, PARP, and VSG promoters drive correct and α-amanitin-resistant transcription. In vitro transcription from the three promoters required, with the exception of the template DNA concentration, identical optimal reaction conditions. This may indicate that these promoters have common trans-acting factors. On the other hand, we showed that transcription from the PARP promoter, in comparison to that from rDNA and VSG promoters, was more efficient, exhibited a lag phase in the kinetic analysis, had a higher template concentration optimum, and was resistant to increased Mn²⁺ concentrations. It has been previously observed in vivo that PARP promoter-mediated transcription of vector sequences was inefficient (2, 22, 23). Therefore, it is possible that vector sequences in our constructs caused termination of PARP promoter-driven transcription but not of transcription initiated at the rDNA and VSG promoters. As a consequence, RNA Pol complexes may have dissociated from the circular PARP-trm template at a high rate, and reutilization of these complexes may have increased the transcriptional efficiency. Although we cannot rule out this possibility at present, the lag phase and the resistance to elevated Mn²⁺ concentrations of PARP-trm transcription suggest a qualitative distinctness of the PARP transcription complex and the existence of one or more PARP gene-specific transcription factors. The existence of such a factor is supported by transfection and nuclear run-on experiments. PARP promoter-driven expression of reporter genes was active in both bloodstream and procyclic forms, but the expression level was 5- to 10-fold higher in the procyclic form (2, 26, 32). Biebinger et al. (2) stably transfected reporter gene constructs which exhibited no posttranscriptional regulation and showed that the high PARP promoter-driven expression level in procyclic cells was independent of the chromosomal integration site. Thus, it is very likely that the increase of reporter gene expression was regulated at the level of transcription initiation. In addition, bloodstream form trypanosome expression of the same constructs was generally down-regulated when integrated at different chromosomal loci. Taken together, these data and our results suggest that procyclic trypanosomes express a PARP gene-specific trans-activating transcription factor.

Transcription reactions with all three promoters were maximal at potassium concentrations of 40 mM, whereas in the T. brucei permeabilized-cell system, rRNA synthesis increased up to a KCl concentration of 150 mM, the highest concentration tested (37). For SL RNA synthesis, however, a potassium concentration of 40 mM was optimal in both permeabilized cells and a homologous cell extract (15, 37). A possible explanation for the discrepancy seen with rDNA, PARP, and VSG promoters is that in our cell extract but not in the permeabilized cell system, trans-acting factors have to bind to the newly added template DNA to form transcription preinitiation complexes and that this binding step may be salt sensitive. Support for this view comes from the work of Janz et al. (20), who characterized specific protein binding to the proximal core element of the PARP promoter and found that formation of the specific DNA-protein complex was sensitive to potassium concentrations above 50 mM.

In conclusion, we have established an in vitro transcription system for the parasite T. brucei in which three different promoters, namely, the rDNA and PARP and VSG gene promoters, drive efficient, accurate, and α-amanitin-resistant transcription. A comparison of the three promoters revealed several differences between PARP promoter transcription on the one hand and rDNA and VSG promoter transcription on the other hand. The in vitro system will facilitate a biochemical characterization of the transcriptional machinery and will reveal whether parasite-specific transcription factors interact with the PARP and VSG promoters.