Production of cell lines stably expressing tetR and T7 RNA polymerase

Plasmid pLEW13, a construct designed to target the T. brucei β-tubulin locus, contains both T7 RNA polymerase and tetR genes with neo as a selectable marker [7] (Fig. 1A). We electroporated T. cruzi CL-Brener epimastigotes with circular pLEW13 DNA (a gift from George Cross) and selected recombinant parasites on 200 μg ml^-1 G418. Stably transformed parasites were obtained after six weeks, even though this vector contains no T. cruzi-derived sequences. Southern analysis showed that the transformants contained multiple copies of the input construct organised in a tandem array (data not shown). Circular DNA in transformed T. cruzi usually replicates as an episome of head-to-tail repeats of the input construct [21]. However, since the vector contained T. brucei β-tubulin coding sequences, which are very similar to the corresponding T. cruzi gene (88% overall nucleotide identity, up to 96% in some regions), it was important to establish whether the pLEW13 tandem array was a circular episome or had resulted from multiple integrations into the tubulin locus.

Circular molecules show aberrant migration on pulsed field gels as their movement is independent of their molecular mass, in contrast to linear chromosomes. DNA from CL-Brener epimastigotes transformed with pLEW13 (CL-Brener [pLEW13]) was therefore subjected to contour-clamped homogenous electric field gel electrophoresis analysis (CHEFE) under differing separation conditions (Fig. 1B). Using parameters designed to separate the larger molecules (up to 3 Mb), the T7 RNA polymerase probe hybridised to a band of approximately 2 Mb and to a smear of higher molecular weight material (Fig. 1B lane 3). When the DNA was fractionated under conditions optimal for separation of molecules of between 300 kb and 1 Mb, the hybridising band ran at 680 kb, again accompanied by a smear of apparently higher molecular weight material (Fig. 1B lane 6). The migration of the pLEW13 construct is therefore independent of its molecular weight indicative of a circular episome containing multiple copies of the T7 RNA polymerase and tetR genes. Southern analysis of genomic DNA also indicated no linkage between the T. cruzi α-tubulin genes and the T7 RNA polymerase (data not shown).

To check expression of the transfected genes, RNA was prepared and analysed by northern blotting. This showed that both T7 RNA polymerase and the tetR gene were expressed at high levels (Fig. 1C). In trypanosomes each mRNA is processed by trans-splicing which results in the addition of a 5' spliced leader sequence of 39 nucleotides [22]. Since the RNA processing signals in pLEW13 were derived from T. brucei it was necessary to establish that the transgenic mRNAs were correctly spliced. For each gene, primers were designed to sequences approximately 150–250 bp into the ORF and used in conjunction with a primer to the T. cruzi spliced leader in an RT-PCR reaction (Methods). The resulting products were cloned and sequenced. Each splice addition site could be mapped to an AG dinucleotide located downstream of a polypyrimidine tract (Fig. 1D). In the case of the T. brucei actin intergenic sequence upstream of neo, three separate splice sites were identified, all of which were upstream of the start codon. Only one of these was identified in the tetR mRNA which has the same flanking sequence. In the case of the procyclin splice acceptor site upstream of the T7 RNA polymerase, only the site previously mapped in T. brucei was utilised [23]. This analysis indicated that the T. brucei RNA processing signals were being correctly utilised by T. cruzi and that the T7 RNA polymerase and tetR mRNAs were therefore likely to be functional.

Features of the tetracycline inducible expression vector

The inducible expression vector pTcINDEX (Fig. 2A) was designed to integrate into the non-transcribed ribosomal RNA spacer region upstream of the pol I-mediated transcription start site [24] (Methods). We targeted this region because, to our knowledge, it is the only section of the T. cruzi genome so far identified as being transcriptionally silent. In addition, the level of sequence conservation at this locus suggested that the construct could be targeted to the corresponding region in multiple parasite strains. The targeting fragment is cloned as a Sac I cassette which can be readily replaced to allow integration elsewhere in the genome.

As a drug selectable marker we used the hygromycin B phosphotransferase (hyg) gene under the control of a non-repressible T7 promoter, thus converting antibiotic resistance into a digenic trait. In pLEW13 transformed cells that constitutively express the T7 RNA polymerase, this arrangement serves a second function. In the presence of hygromycin, the requirement for T7 RNA polymerase to drive expression of hyg removes the necessity for the continued use of G418 to maintain the pLEW13 construct and selects for trypanosomes with active T7 RNA polymerase. The inducible expression cassette in the pTcINDEX vector contains a tetracycline-dependent T7 promoter, with the tet operator sequence (tetO, cCTATCAgTGATAGa, where upper case indicates bases important in tetR binding) placed immediately downstream. The multiple cloning site is flanked at its 3'-end by the intergenic sequence from the T. cruzi actin locus to provide a polyadenylation signal, and at the 5'-end by the splice acceptor site from the ribosomal protein P2β locus. Sequences from this region have been shown to enhance the expression of transfected genes [25]. Finally, we incorporated a T7 RNA polymerase transcription terminator into the construct to block run-through transcription of sequences downstream of the integration site.

To test the capability of the vector to mediate tetracycline-regulatable expression we cloned the genes encoding firefly luciferase (Luc) and red fluorescent protein (RFP) into the multiple cloning site (Fig. 2, Methods). Spe I linearised forms of the resulting constructs (pTcINDEX-Luc and pTcINDEX-RFP) were then used to transform CL-Brener [pLEW13] epimastigotes that constitutively express T7 RNA polymerase and tetR. Integration into the rRNA locus (illustrated in Fig. 3A) was confirmed by Southern analysis. This showed linkage of both of the transgenes to the endogenous 18S rRNA gene, (for examples, see Fig. 3). The appearance of novel fragments in the lanes containing DNA from the transformants (9.5 kb with the luciferase and 18S rRNA probes (Fig. 3B, lanes 2 and 4), 6 kb with the 18S rRNA and RFP probes (Fig. 3B, lanes 6 and 8)), which were absent from the CL-Brener [pLEW13] lanes, were diagnostic of targeted integration into the non-transcribed spacer region upstream of the 18S rRNA gene.

Induction of luciferase in pTcINDEX-Luc transformants

We first investigated the induction of luciferase RNA in a polyclonal line of pTcINDEX-Luc transformed cells. Tetracycline was added once to epimastigotes in early mid-logarithmic growth phase (approximately 10⁶ parasites ml^-1) and aliquots removed every 24 hours for RNA purification. No further tetracycline was added during this period, as we wished to see if the gene returned to a repressed state. Northern analysis was performed (Fig. 4A) and the relative level of luciferase RNA measured at each time point using a phosphorimager. In the lane containing RNA from non-treated cells, the signal detected was not significantly above the background measured from an irrelevant piece of the membrane. This indicates a tightly regulated system with a very low level of "leaky" transcription. 24 hours after the addition of tetracycline, the level of luciferase mRNA was found to have increased dramatically (Fig. 4A). The mRNA levels at later time points declined gradually. The change in luciferase RNA levels was mirrored in the level of luciferase activity (Fig. 4B). The enzyme level increased considerably over 24 hours and continued to increase up to 48 hours. Thereafter it declined gradually. When the cells were washed after 24 hours exposure to tetracycline and resuspended in tetracycline-free medium, the luciferase activity reached a peak at 48 hours but had declined almost to background levels three days later (Fig 4B, dashed line).

To examine the relationship between tetracycline concentration and the induction of luciferase activity, a culture of pTcINDEX-Luc transformed CL-Brener [pLEW13] epimastigotes was divided into 10 individual flasks and tetracycline added at a range of concentrations (Fig. 4C). The cells were incubated for 24 hours, then harvested and the luciferase activity measured (Methods). There was negligible increase in luciferase activity over the level in non-treated cells at concentrations up to and including 1 ng ml^-1. The increased activity became significant following treatment with 5 ng ml^-1 and continued to increase with concentration before levelling off at 500 ng ml^-1 (Fig. 4C). At tetracycline concentrations of 1 or 2 μg ml^-1, the extent of induction decreased approximately two-fold, an effect that was reproducible. At higher levels of tetracycline there were detectable increases in parasite doubling time. The optimal increase of luciferase activity that was achieved was approximately 60-fold over background. This is less than the increased level of the corresponding transcript (Fig. 4A) and may indicate the presence of control mechanisms at the level of translation or instability of the luciferase protein in this context.

Cell-by-cell examination of the induction process using RFP

To determine how individual cells responded to tetracycline, we examined clones isolated following transformation with pTcINDEX-RFP (Methods). Epimastigotes of clone CL:RFP C2 were maintained in tetracycline-free medium, then an aliquot was fixed onto a slide. Tetracycline was added to the remainder of the culture. An aliquot of cells was removed and fixed onto a slide every 24 hours for eight days. It was apparent that RFP expression had been highly induced by the third day, since the cell pellet had a red tinge visible to the eye. The cells were stained for DNA and examined by confocal microscopy. No red fluorescence was visible in the uninduced population (Fig. 5A). After 24 hours a few cells displayed faint fluorescence, and after three days some cells were extremely bright. The number of visibly fluorescent cells increased over time. After eight days the majority of cells exhibited some red fluorescence (Fig. 5B), although there was variation in the level. This suggested that induction, as measured using this parameter, does not occur at the same rate or to the same degree in all cells of a given clonal population.

To examine this variation further we quantified the distribution of induced fluorescence in the population by FACS analysis (Fig. 6). In this experiment the tetracycline treatment was carried out at 0.5 μg ml^-1, as this concentration was optimal for expression, at least in the case of luciferase (Fig. 4C). FACS analysis showed that in the first 24 hours post-induction, there was a significant shift in the fluorescence profile, with 35% of cells showing a 10–1000 fold increase in intensity (Fig. 6, blue line). The profile shifted rightwards over time but did not sharpen, indicating that fluorescence intensity varied between individual cells, confirming the observation made by microscopy (Fig. 5). The maximal shift was seen on day 5 (Fig. 6, green line), when 14 % of the cells were found to exhibit a 1000–10000 fold increase in fluorescence. Even at this stage however, 34% of cells remained in the 0–10 arbitrary fluorescence units (AFU) range.

Addition of an rRNA promoter to pLEW13 results in higher background expression levels

In an attempt to produce a more homogeneous induction profile, we constructed a derivative of pLEW13 in which the T7 RNA polymerase and tetR genes were transcribed from the T. cruzi rRNA promoter (Methods). Cells were transformed with this plasmid (pTcrRNA-T7tet) and selected at 100 μg ml^-1 G418. The transformants were resistant to 2 mg ml^-1 G418, with no lag phase, indicative that the rRNA promoter was driving high level expression of the neo gene. These cells were then electroporated with the inducible vector pTcINDEX-RFP. Parasites were cloned immediately after electroporation. FACS analysis of several independent clones confirmed that expression of RFP was tetracycline-regulated, but again the response was heterogeneous within clonal populations (Fig. 7).

Expression and localisation of epitope-tagged superoxide dismutase

To assign a biological role to a protein it is necessary to know its subcellular location. Generation of specific antibodies against every protein of interest is costly, time-consuming and not always successful. We therefore made a derivative of pTcINDEX with a c-myc epitope tag inserted next to the polylinker to facilitate localisation of induced proteins (pTcINDEX-C-myc, Fig. 2B). This could also provide a simple method to follow the induction by western blotting. To test the vector, we chose the T. cruzi superoxide dismutase A gene (TcSOD A), which encodes an isoform with a predicted mitochondrial targeting sequence. The T. brucei orthologue of this gene is targeted to the mitochondrion [27, 28].

The TcSOD A gene was inserted into pTcINDEX-C-myc such that the epitope tag was located at the carboxyl-terminus of the fusion protein (Fig. 2B). CL-Brener [pLEW13] epimastigotes were transformed as previously. Two clones were characterised (A1 and A2). With both, an induced band was visible after western blotting (Fig. 8A). In the induced cells, the corresponding bands were visible on a Coomassie stained gel (Fig. 8B). The upregulated SOD was enzymatically active (Fig. 8C), with induced cells showing a 14- and 18-fold increase over the control lines, respectively. This represents the total cellular SOD activity. Since there are four distinct isoforms in trypanosomatids [27, 28], it is clear that the level of SOD A overexpression considerably exceeds 18-fold.

It was important to confirm that the SOD A was targeted correctly since overexpression might lead to mis-targeting or blocking of the trafficking pathway. Cells were stained with an antibody against the carboxyl-terminal epitope tag and examined by microscopy (Fig. 9). The immunofluorescence showed targeting of the induced protein to the single lattice-like mitochondrion of the trypanosome with a concentration in a rod-like structure next to, or on top of, the kinetoplast (mitochondrial) DNA. The exact nature of this structure is unclear, but the consistent proximity to the kinetoplast suggests a possible role in protection of the replicating kDNA from reactive oxygen species.

Parasite maintenance and genetic manipulation

Epimastigotes of T. cruzi CL-Brener were maintained at 27°C in RPMI-1640 medium as described previously [46], except that we used 5% tetracycline-free fetal calf serum (Autogen Bioclear). Parasites were transformed by electroporation using a Bio-Rad Gene Pulser II, placed into fresh medium and incubated for 48 hours to allow expression of the drug-selectable marker. The appropriate drug was then added (G418 at 100–200 μg ml^-1 or hygromycin at 100 μg ml^-1) and the cells incubated for a further four to six weeks to allow selection of transformants. For direct cloning, parasites were resuspended in 24 ml of fresh medium directly after electroporation. 1 ml was then transferred to each well of a 24-well plate and the cells allowed to grow for 48 hours prior to addition of the selective drug. Typically, between 2 and 5 clones were generated per 24-well plate.

Plasmid construction

The inducible expression vector pTcINDEX (Fig. 2) was based on the T. brucei RNAi vector pZJM [8](a gift from Paul Englund). First, the inverted promoter fragment of pZJM, which contains bi-directional T7 promoters, was isolated by digestion with Kpn I and Bam HI. This fragment was then subcloned into pGEM3zf+ (Promega) to produce vector pGEMT7Tet2. In parallel, a hyg gene flanked by the processing signals from the T. cruzi glycosomal glyceraldehdye-3-phosphate dehydrogenase gene[21, 46] was inserted into the Eco RV site of pBluescript KS(-). A constitutive T7 promoter was then inserted upstream of the hyg gene after generation of the appropriate fragment by PCR. This 2.4 kb cassette was isolated following Kpn I and Sac I digestion and sub-cloned into pGEMT7Tet2, upstream of the fragment derived from pZJM, to create pGEMhygT7-3. Since the pGEM backbone contains an additional unwanted T7 promoter, the whole insert fragment was liberated by Bam HI/Sac I digestion and inserted into pUC19ΔH (pUC19 with the Hind III site deleted by end-filling).

The ribosomal RNA non-transcribed spacer and promoter region were amplified from genomic DNA of T. cruzi in two pieces of 1.8 kb and 0.3 kb to allow the introduction of a unique Spe I site for vector linearization prior to transfection. These pieces were generated using the primer pairs:

5'-TTTACTAGTAGCTCGGTGCACCCTG, 5'-GGGGAGCTCACACAAATGGACGGTTA and, 5'-GGGGAGCTCATTTGTGTCTAGTACATC, 5'-GGGACTAGTCTGAGGCATGCATGGCTA.

The fragments were ligated together then cloned into the Sac I site immediately downstream of the hyg cassette (as shown in Fig. 2) to produce plasmid pTcIRi. To construct the inducible expression vector, we modified pTcIRi by removing the antisense promoter and adding a polylinker and RNA processing signals. Briefly, pTcIRi was partially digested with Sac I and Hind III. The Sac I/Hind III fragment containing the hyg gene and the sense strand inducible promoter was cloned into pGEM3zf+ (Promega) to create pGEMTcI. The T7 transcriptional terminator from pTcIRi was amplified with primers 5'-TTTCTCGAGCGGCCGCCGGCATCGATACGCGTGGATCCTCGCGAATCAGGTGCTAGCCCGCT and 5'-TTTAAGCTTGATCCCCGGATATAGTT. This fragment, which also incorporated a multiple cloning site (underlined, Fig. 2A), was inserted into Xho I/Hind III digested pGEMTcI.

The splice acceptor site from the ribosomal protein P2β locus was amplified from pTREX-n [25] (a gift from Mariano Levin), using primers which added an Xho I site to the 5' end and a Not I site to the 3'end. This 212 bp fragment was cloned into the corresponding sites of the polylinker. The actin intergenic sequence [Acc. No. U20234], which contains a putative polyadenylation signal, was amplified from genomic DNA of T. cruzi using primers: 5'-CCCGGATCGTCGCGAGGCAGGCCCAAGCA and 5'-CCCGATATCGTCAGACATCCTTAGAA. The resulting 424 bp fragment was then digested with Bam HI and Eco RV and cloned into the Bam HI and Nru I sites of the polylinker. The entire insert was again transferred to pUC19 via a partial Sac I/Hind III digest to produce the final expression construct pTcINDEX (Fig. 2).

The luciferase coding sequence was obtained from pGEM-luc (Promega). The plasmid was digested with Sal I and end-filled by the Klenow fragment. The gene was then excised by digestion with Bam HI. pTcINDEX was digested with Bam HI and Nru I and the luciferase gene was ligated into the vector to produce pTcINDEX-Luc. To obtain the gene encoding the red fluorescent protein (RFP), construct pTEX-Red [47] was first digested with Bam HI and Bgl II and re-ligated to delete awkward restriction sites. The modified plasmid was then cut with Spe I and the ends filled by Klenow treatment. The RFP gene was liberated by digesting the linear plasmid with Cla I and cloned into Nae I/Cla I sites of pTcINDEX to create pTcINDEX-RFP.

pLEW13 was modified to include a T. cruzi rRNA promoter to drive expression of the T7 RNA polymerase, neo and tetR genes. Briefly the Sac I fragment carrying the rRNA promoter and upstream spacer region was removed from pTcINDEX and subcloned into pUC19 to make pUC-TcrRNA. The T7 RNA polymerase, neo and TetR genes were removed from pLEW13 on a 5.9 kb fragment using EcoRV and Stu I. This fragment was then cloned into the unique Spe I site in the rRNA promoter fragment of pUC-TcrRNA, such that the transcription initiation point was upstream of the T7 polymerase gene. This derivative of pLEW13 was named pTcrRNA-T7Tet.

We created an epitope tagging vector by cloning the BPP1-myc fusion gene from pTEX-BPP1-9E10 into Bam HI/Nru I digested pTcINDEX [48]. This fusion gene contains a unique Eco RV site between the BPP1 ORF and the c-myc tag such that any gene of interest can replace the BPP1 coding sequence and be cloned in-frame with the tag (Fig. 2B). The epitope tag encodes the sequence EQKLISEEDL*, where * indicates a translational stop. This vector was named pTcINDEX-C-myc where the uppercase C denotes that the tag is fused to the carboxyl terminal of the protein of interest. To make an inducible tagged copy of TcSOD A, the gene (>Tc00.1047053509775.40 [49]) was amplified from genomic DNA of the CL-Brener strain using the following primers:

SOD A F: gggggatccATGTTGAGACGTGCGGTGAA

SOD A R: ggttgatatcTTTTATTGCCTGCGCAT

where underlining indicates restriction sites introduced for ease of cloning. The 699 bp product was digested with Bam HI and Eco RV and ligated into Bam HI/Eco RV digested pTcINDEX-C-myc, such that the SOD A ORF was in-frame with the carboxyl terminal epitope tag under the control of the inducible T7 promoter. The construct was confirmed by DNA sequencing.

Nucleic acid analysis

DNA and RNA were prepared and purified using Qiagen kits as per manufacturer's instructions. RNA was quantified using a 2100 Bioanalyzer with RNA 6000 Nano Labchip (Agilent). Southern and northern blotting were carried out using standard protocols. Reverse transcriptase PCR (RT-PCR) was carried out using the Access RT-PCR kit (Promega) and primers:

Spliced Leader sense 5'-GGGGGATCCACAGTTTCTGTACTATATTG

T7 Polymerase antisense 5'-TCGTAAGACTCATGCTCAA

Neo antisense 5'-CCTCGTCCTGACAGTTCAT

tetR antisense 5'-TGCCTATCTAACATCTCA

The products were cloned into pGEM-T (Promega) and sequenced using a dye terminator cycle-sequencing kit (Applied Biosystems) and an ABI Prism 377 DNA sequencer. For CHEFE analysis parasite blocks were made as described [50]. The chromosomes were resolved on a CHEFmapper system (Bio-Rad) using the auto-algorithm and conditions as detailed in figure legends.

Induction of gene expression

For induction experiments, epimastigotes were cultured in 25 cm³ flasks at 27°C and maintained in logarithmic growth phase (10⁶ – 10⁷ cells ml^-1). Control cells were grown in tetracycline-free medium, whilst the induced cells were cultured in medium supplemented with the stated concentration of tetracycline. We found that the doubling time of wild type parasites was unaffected by low concentrations of tetracycline, but was increased by 13% at 5 μg ml^-1 and by 30% at 10 μg ml^-1. Inductions were carried out over variable time courses as stated in figure legends.

Luciferase assays

Epimastigotes transformed with pTcINDEX-Luc were grown as described above. At each time point an aliquot was removed, pelleted and washed in PBS (137 mM NaCl, 4 mM Na₂HPO₄, 1.7 mM KH₂PO₄, 2.7 mM KCl). Cell pellets were frozen in liquid nitrogen and stored at -80°C. For the luciferase assay, the pellet was resuspended in 500 μl of cell culture lysis reagent (Promega). Lysates were vortexed for 15 seconds and the debris removed by centrifugation. Activity was measured using the luciferase assay system (Promega) and light emission measured on a β-plate counter (Wallac). The linear detection limits of the counter were measured using serial dilutions of Quanti-Lum recombinant luciferase (Promega). Protein concentrations were determined by the BCA assay (Pierce) using equivalent amounts of cells lysed in PBS, as the lysis reagent is incompatible with the protein assay.

Fluorescence microscopy and FACS analysis of RFP expression

RFP expression was examined by confocal microscopy on a Zeiss LSM 510 Axioplan microscope. Transformed parasites were induced as described above. At each time point, an aliquot of cells (10⁷) was removed, pelleted, washed in PBS and then fixed for 30 minutes in 4% paraformaldehyde/PBS. Cells were then washed and resuspended in 5 ml PBS. 20 μl of the suspension was dotted onto a single well of a 12-well slide. DNA was stained by adding 50 nM TOTO-3 (Molecular Probes) in 10 mg ml^-1 RNAse A/0.1% saponin/PBS to each well, incubating at room temperature for 20 minutes, then washing twice in PBS. Slides were mounted in 1:1 PBS/glycerol. For FACS analysis, cells were fixed as above and finally resuspended at 10⁷ parasites ml^-1. 5 × 10³ – 10⁴ cells per time point were counted on a FacScan or FacsCalibur (Becton Dickinson). Data were analysed using Cellquest™ software (BD Sciences).

Protein extraction and analysis

For western blot and SOD activity assays, cells were pelleted, and washed once in PBS. The cells were pelleted again and resuspended in lysis buffer (PBS supplemented with proteinase inhibitors, Roche). The cell suspension was freeze-thawed three times in liquid nitrogen then sonicated. Membrane debris was removed by centrifugation (10,000 g for 10 mins). The supernatant was removed to a sterile tube and stored at -80°C. SDS-PAGE and western blotting were carried out as per standard protocols. The western blots were probed with mouse monoclonal c-Myc (9E10) (cat no. sc-40, Santa Cruz Biotechnology Inc.) diluted 1:2000. For SOD activity assays the Bioxytech™ SOD 525 (Oxis Research) kit was used as per manufacturer's instructions.

Immunofluorescence

To check the localisation of the tagged SOD A, epimastigotes were fixed in 4% paraformaldehyde and dried onto slides. The slides were stained with mouse monoclonal c-Myc (9E10) (diluted 1:200) and then Alexafluor 488 conjugated goat anti-mouse (diluted 1:400 Molecular Probes). DNA was stained with DAPI. Slides were examined on a Zeiss LSM 510 confocal laser scanning microscope.