Activity of a Trypanosome Metacyclic Variant Surface Glycoprotein Gene Promoter Is Dependent upon Life Cycle Stage and Chromosomal Context

Abstract

African trypanosomes evade the mammalian host immune response by antigenic variation, the continual switching of their variant surface glycoprotein (VSG) coat. VSG is first expressed at the metacyclic stage in the tsetse fly as a preadaptation to life in the mammalian bloodstream. In the metacyclic stage, a specific subset (<28; 1 to 2%) of VSG genes, located at the telomeres of the largest trypanosome chromosomes, are activated by a system very different from that used for bloodstream VSG genes. Previously we showed that a metacyclic VSG (M-VSG) gene promoter was subject to life cycle stage-specific control of transcription initiation, a situation unique in Kinetoplastida, where all other genes are regulated, at least partly, posttranscriptionally (S. V. Graham and J. D. Barry, Mol. Cell. Biol. 15:5945–5956, 1985). However, while nuclear run-on analysis had shown that the ILTat 1.22 M-VSG gene promoter was transcriptionally silent in bloodstream trypanosomes, it was highly active when tested in bloodstream-form transient transfection. Reasoning that chromosomal context may contribute to repression of M-VSG gene expression, here we have integrated the 1.22 promoter, linked to a chloramphenicol acetyltransferase (CAT) reporter gene, back into its endogenous telomere or into a chromosomal internal position, the nontranscribed spacer region of ribosomal DNA, in both bloodstream and procyclic trypanosomes. Northern blot analysis and CAT activity assays show that in the bloodstream, the promoter is transcriptionally inactive at the telomere but highly active at the chromosome-internal position. In contrast, it is inactive in both locations in procyclic trypanosomes. Both promoter sequence and chromosomal location are implicated in life cycle stage-specific transcriptional regulation of M-VSG gene expression.

African trypanosomes are protozoan parasites causing serious and potentially fatal diseases of humans and domestic livestock. They have a digenetic life cycle, with one phase in the tsetse fly vector and another in the tissue fluids and bloodstream of the mammalian host (70). Antigenic variation, the continual switching of the variant surface glycoprotein (VSG) which constitutes the surface coat, allows evasion of mammalian immunity (16). The VSG coat is encoded by around 1,000 genes, most of which are thought to be in long tandem arrays in chromosomes but some of which are at telomeres (68). Only one VSG gene is expressed at a time, and for most genes this is achieved by duplicative transposition, where a copy of a VSG gene is synthesized and inserted into a transcriptionally active, telomeric, bloodstream expression site (3). There are estimated to be around 20 expression sites for VSG genes expressed in the bloodstream (39), and switching between VSGs is accomplished by a number of mechanisms which involve either changing the expression site that is active or replacing the VSG gene in the active expression site (3). Bloodstream expression sites have a common architecture; the VSG gene is at the 3′ end, adjacent to the simple repeats of the telomere, and is coexpressed with various 5′ flanking expression site-associated genes (ESAGs) from a promoter located 40 to 60 kb upstream (51). These complex expression sites are subject to a range of control mechanisms. Within the bloodstream life cycle stage, expression of individual genes within the active expression site is regulated posttranscriptionally (40, 48–50, 73) but switching between expression sites is by regulation of transcription initiation (54). However, in the procyclic stage in the insect, where VSG is not expressed, expression sites are down-regulated, partly at the level of transcription initiation (55) and partly by transcription attenuation close to the promoter itself (14, 37, 48, 49, 55, 69, 73). Thus, in addition to control mechanisms for individual genes within life cycle stages, there are further mechanisms that control between stages.

When trypanosomes are taken up by the tsetse fly, they undergo a rapid differentiation to the procyclic stage. Procyclic trypanosomes are not coated with VSG, expressing instead a new surface coat composed of the protein procyclin, also known as PARP (procyclic acidic repetitive protein) (45, 53). Expression of VSG is reinitiated only during differentiation to the nondividing metacyclic stage in the salivary glands of the tsetse fly (62). This is the stage infective for mammals, and its coating with VSG is thought to be essential for parasite survival and proliferation following transfer into the host. Only a small, specific subset of VSGs are expressed (<28; 1 to 2% of the total repertoire) in the metacyclic population (17, 22, 65). The metacyclic VSG (M-VSG) genes are activated randomly in the metacyclic stage, yielding a polyclonal population, each individual of which expresses only one VSG (62). Further, clonal analysis shows that M-VSG genes are activated in situ at the metacyclic stage, without undergoing duplicative transposition, the mechanism associated with activation of most bloodstream VSG genes (41). This system for randomly and polyclonally expressing M-VSG genes is dominant over the separate bloodstream system and may facilitate establishment of infection in partially immune hosts in the field (4).

M-VSG genes continue to be expressed for up to 7 days following transfer of parasites to the mammal, despite morphological differentiation to bloodstream forms. Further, these early bloodstream trypanosomes, termed metacyclic-derived trypanosomes, continue to express M-VSG genes by the metacyclic-specific in situ mode of activation (28). We have used metacyclic-derived trypanosome populations as a model system for analysis of M-VSG gene expression and have uncovered a number of features in which M-VSG gene expression differs from expression of other trypanosome genes. Whether examined in bloodstream populations derived directly from individual metacyclic cells (26) or derived during VSG switching during chronic bloodstream infections (1), they are expressed as monocistronic transcription units from promoters located within only 3 kb upstream. Further, the telomeres harboring M-VSG genes have little resemblance to bloodstream expression sites; there are no, or very few, of the 70-bp repeats that flank most VSG genes, and there are only limited ESAG-related sequences (27, 41, 43, 60). Indeed, for the two M-VSG gene transcription units that we have studied, there is a transcriptional gap upstream of up to 15 kb (26), a situation unusual in trypanosomes, whose genome is otherwise densely packed with coding sequence. The most prominent difference is that M-VSG genes are under transcriptional regulation during the parasite life cycle (26) whereas in Kinetoplastida in general, all other genes studied thus far have been shown to be regulated, at least partly, at the posttranscriptional level (13, 25). Thus, M-VSG genes represent a new class of genes in Kinetoplastida in that they are transcribed as monocistronic transcription units from otherwise silent telomeric regions, their promoters are located very close to the ends of telomeres, and their expression is most probably under transcriptional regulation during the parasite life cycle.

Previously we identified a transcriptional start site for the 1.22 M-VSG gene and showed, in transient transfection in metacyclic-derived trypanosomes, promoter activity associated with a region encompassing this site. Nuclear run-on analysis revealed the putative promoter to be inactive in both procyclic and bloodstream trypanosomes, but transient transfection analysis revealed a difference between these two stages, yielding only minimal activity in the procyclic stage but very high activity in the bloodstream (26). One explanation for this surprising result might be that dissociation of the promoter from locus-specific, down-regulatory control elements would lead to an apparent activation of the promoter in bloodstream trypanosomes. Now, using stable transformation to integrate into the trypanosome genome, we show that in bloodstream trypanosomes, the 1.22 promoter can drive reporter gene expression in a chromosome-internal locus but not at its endogenous telomeric locus. We carried out similar integrations in procyclic trypanosomes and found that the promoter was inactive at both sites. Our results indicate that regulation of M-VSG gene expression during the parasite life cycle is very stringent and that two different life cycle stage-specific control mechanisms repress M-VSG gene expression outside the metacyclic stage: possibly telomere positioning in bloodstream forms and cis-acting promoter sequences in procyclic trypanosomes.

MATERIALS AND METHODS

Trypanosomes.

A virulent, cloned line of Trypanosoma brucei EATRO 795 which retains fly transmissibility was used in these studies as described previously (28, 64). Maintenance in mammalian hosts was carried out by standard procedures (29). Procyclic culture-form trypanosomes were established from EATRO 795 trypanosomes by standard methods and were maintained in SDM-79 medium (7). Bloodstream forms of stock EATRO 795 were established in axenic culture by the method of Carruthers and Cross (9) and were maintained in HMI-9 medium (9). Although EATRO 795 is not a conventional laboratory-adapted line of T. brucei since it retains fly transmissibility, trypanosomes were monomorphic in culture and retained high virulence for mice even after 5 months in culture.

Differentiation.

Three methods were used to attempt to differentiate cultured bloodstream-form trypanosomes. First, cells were placed at 27°C in the presence of the citric acid cycle intermediates 3 mM citrate and 3 mM cis-aconitate, pH 7.4 (8). Second, cells were incubated in differentiation trypanosome medium (47) supplemented with citrate–cis-aconitate and placed at 27°C. Third, we used an adaptation of the method of Overath et al. (47) designed to differentiate monomorphic bloodstream trypanosomes (6a). Cells were cultured in HMI-9 (9) at 37°C supplemented with citrate–cis-aconitate for 17 h and then placed at 27°C in Cunningham’s SM medium (19) also containing citrate–cis-aconitate. The preincubation, temperature, and medium changes are believed to be necessary for full differentiation of monomorphic trypanosomes (6a).

Recombinant clones.

The clones derived from the basic copy locus of the M-VSG gene ILTat 1.22, λMT1.22B, pMG7.1-1, pMT1.22-BPs, and pMT1.22-HPl (Fig. 1A), have been described previously (15, 26, 43). Construction of plasmids pHD52CAT, p−HD52CAT, p122sHD52CAT, and p122lHD52CAT for use in transient transfection studies has also been described previously (26).

FIG. 1.

Activity of the 1.22 M-VSG gene promoter in transient transfection in bloodstream forms. (A) Map of the 1.22 basic copy telomere (28) and the clone pMG7.1-1 derived from its 3′ end (15). Horizontal black bars show the extent of the short and long promoter-containing fragments cloned in pMT122-BPs and pMT122-HPl. Abbreviations: B, BamHI; H, HindIII; K, KpnI; S, SalI; P, PstI; E, EcoRI. Hatched box, VSG gene region; stippled box, 70-bp repeat region; black box, ingi retroposon sequence; oval, end of the telomere. (B) Short (1.22s) and long (1.22l) versions of the 1.22 M-VSG gene promoter region were tested for the ability to drive expression of a CAT reporter gene, flanked by actin RNA processing signals (35), in transient transfection in bloodstream trypanosomes. Promoter regions tested were as follows: B-ES, the 221 bloodstream expression site promoter (73); 1.22s, the insert in pMT1.22-BPs; 1.22l, a KpnI/PstI fragment derived from the insert in pMT1.22-HPl; and NONE, the CAT gene flanked by actin RNA processing signals but with no promoter upstream. Values are means and deviations from the means for eight experiments.

Plasmid constructs for stable transformation were pt122BC and pr122BC, containing a 1.8-kb fragment encompassing the 1.22 promoter. In pt122BC, the selectable marker cassette was constructed by replacing the chloramphenicol acetyltransferase (CAT) gene, in p5′parpCAT3′parp (44), with a 376-bp fragment containing the ble gene coding region (21) to yield p5′parpble3′parp. The reporter gene cassette was the insert in plasmid p122lCAT3′parp (26), and it was recloned into pBluescript SK− as a KpnI/BamHI fragment to yield restriction sites in the pBluescript polylinker 3′ of the reporter gene cassette (p122lCAT3′parpSK−). The new plasmid was digested with SstI, blunt ended, and then digested with NotI, and the selectable marker cassette was inserted downstream of the reporter gene cassette in the same orientation on a NotI/KpnI-blunted fragment isolated from p5′parpble3′parp. The two cassettes were thus separated by 50 bp of pBluescript polylinker-derived sequence. There is a single SalI site within the 1.22 M-VSG gene promoter region which was used for linearizing the plasmid to promote stable integration into the 1.22 expression telomere (Fig. 2A). Plasmid pr122BC was very similar to pt122BC except that a 712-bp fragment derived from the nontranscribed spacer region of the rRNA locus (71) was inserted upstream of the 1.22 promoter fragment. A 1,148-bp EcoRI fragment containing the ribosomal locus-derived fragment was isolated from pHD430 (71) and cloned into pBluescript SK−. Digestion of this subclone with KpnI yielded the required 712-bp fragment, which was then cloned into KpnI-digested pt122BC to yield pr122BC. For targeting this construct to the nontranscribed spacer region of the rRNA locus, it was to be cleaved with NotI; however, there was already a NotI site in the plasmid located between the two cassettes in the pBluescript polylinker portion. To remove this extra site, pr122BC was partially cleaved with NotI such that only one of the NotI sites was digested and was then blunt ended and religated. The correct plasmid clone was selected by restriction enzyme mapping, and all plasmids were checked by sequencing.

FIG. 2.

Insertion of plasmid pt122BC into the 1.22 endogenous expression telomere. (A) Structure of plasmid pt122BC which has been linearized at the SalI site within the 1.22 promoter region. Abbreviations: Pv, PvuII; Ps, PstI; pBS, pBluescript sequences; CAT, CAT reporter gene; ble, selectable marker gene encoding phleomycin resistance. Stippled boxes, 1.22 promoter region; black box, procyclin/PARP promoter region; open boxes, marker genes; black flag; 1.22 promoter; white flag, procyclin/PARP promoter; dotted line, pBluescript sequence. (B) Partial map of the 1.22 basic copy telomere showing the targeting site, a SalI (S) restriction enzyme site. Also shown are the PvuII and SacI fragments containing the promoter. Abbreviations: K, KpnI; S, SalI; Ps, PstI; Pv, PvuII; Sc, SacI. Stippled box, the 1.22 promoter-containing region; black box, 70-bp repeat region; hatched box, VSG coding region; black flag, 1.22 promoter; oval, end of the telomere. (C) Result of targeting pt122BC into the 1.22 telomere. The new PvuII and SacI fragments, generated by the insertion, which contain the promoter are shown. The horizontal black bar between panels B and C represents the KpnI/PstI fragment used as a probe in hybridizations in panels D and E. (D) Southern blot analysis of PvuII-digested genomic DNA, fractionated on a 0.6% agarose gel, from wild-type trypanosomes (track 1) and from BSFtelo122BC trypanosomes (track 2). (E) Southern blot analysis of SacI-digested genomic DNA fractionated on a FIGE gel of DNA isolated from wild-type trypanosomes (track 1) and from BSFtelo122BC trypanosomes (track 2). Both Southern blots were hybridized with the KpnI/PstI promoter probe shown above panel C in 5× SSC at 65°C and washed to 0.1× SSC at 65°C.

Other plasmid constructs used were pActine, containing a T. brucei actin gene (5), pTb α,β-T1, a clone containing an α/β-tubulin repeat unit (63), and pR4, containing a ribosomal DNA (rDNA) repeat unit (39).

DNA sequence analysis.

Sequencing was performed on denatured double-stranded plasmid DNA by the dideoxy-chain termination method (Sequenase kit; Amersham International). Sequences for both strands of recombinant plasmids were obtained by using the recommended primers for pBluescript or specific primers synthesized on an Applied Biosystems PCR-mate oligonucleotide synthesizer. Computer analysis was carried out by using the Genetics Computer Group sequence analysis software package.

Nuclear run-on analysis.

Preparation and storage of nuclei and run-on reactions were carried out exactly as described previously (40). Run-on reactions were for 5 min at 37°C. Radiolabeled RNA was isolated by using TRIzol reagent exactly as in the protocol for isolation from small samples (Life Technologies). Essentially, reactions were stopped by addition of 800 μl of TRIzol reagent and incubated for 5 min at room temperature to lyse nuclei. Then 160 μl of CHCl₃ was added, and phases were separated by centrifugation at 12,000 × g for 15 min. RNA was precipitated from the aqueous phase with isopropanol. Radiolabeled transcripts were separated from unincorporated nucleotides on NucTrap probe purification columns (Stratagene). Hybridizations were at 55°C in 3× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) for 48 h, and washes were to 0.1× SSC–0.1% sodium dodecyl sulfate (SDS) at 65°C.

FIGE.

Field inversion gel electrophoresis (FIGE) was performed with a switchback pulse controller (Hoefer) and a 1% GTG agarose (Biometra) gel run in 0.5× Tris borate electrophoresis buffer. The gel was run at 150 V (5.2 V/cm) for 22 h at 4°C with a pulse time of 0.6 to 2.0 s and then was run in reverse for 10 min with a forward-to-reverse ratio of 3:1. Molecular size markers (Bio-Rad) were 8 to 48 kb. The gel was depurinated, denatured, and neutralized as for genomic DNA agarose gels (56) and then Southern blotted onto Zetaprobe nylon membrane (Bio-Rad). After blotting, the membrane was baked at 80°C for 2 h.

Purification of nucleic acids, Northern and Southern blotting, and hybridization.

DNA was prepared by using a Nucleon II DNA preparation kit (Scotlab). Genomic DNA was fractionated on 0.6% agarose gels, depurinated in 0.25 M HCl, denatured in 0.5 N NaOH, neutralized, Southern blotted onto a nylon membrane (Hybond-N; Amersham International plc), and then immobilized by UV irradiation (56). RNA was prepared by lithium chloride-urea lysis of trypanosomes followed by phenol extraction (2). RNA was fractionated by electrophoresis on 1% agarose-formaldehyde gels following denaturation of 5 μg of total RNA by incubation for 10 min in the presence of 50% formamide–2.2 M formaldehyde (56). RNA was Northern blotted directly onto a nylon membrane (Hybond-N; Amersham) and immobilized on the filter by UV irradiation. Radiolabeled probes were prepared by random hexanucleotide priming of restriction fragments separated by electrophoresis in low-melting-point gels (23) or by in vitro transcription of the CAT gene cloned into pBluescript KS−, using T3 polymerase (Stratagene protocol handbook). Hybridization with random-primed probes, washing of blots, and removal of hybridized probes were carried out as detailed in the Hybond protocol. Northern blot hybridizations were in 3× SSC–50% formamide at 42°C, and blots were washed to 0.5× SSC at 65°C. Hybridizations with in vitro-transcribed probes were carried out at 55°C in 50% formamide–5× SET (1× SET is 150 mM NaCl, 10 mM Tris-HCl [pH 7.5], and 1 mM EDTA)–5× Denhardt’s solution–50 μg of tRNA per ml–0.5% SDS and washed at 65°C in 0.1× SET–0.1% SDS. Where filters were hybridized with a number of probes sequentially, probes were removed by immersing the filter in a boiling solution of 0.5% SDS, after which filters were autoradiographed to check that no residual hybridization remained.

Transient transfection of bloodstream trypanosomes.

Blood containing trypanosomes at a concentration of 5 × 10⁸ cells/ml was harvested from rats by cardiac puncture and maintained at 37°C. Buffy coat parasites were collected, and 5 μg of supercoiled, CsCl-purified plasmid DNA was electroporated per 3 × 10⁷ trypanosomes per transfection cuvette (0.5 ml) exactly as described previously (73) with a single pulse of 1,500 V, 25-μF capacitance, using a Bio-Rad Gene Pulser. Following electroporation, parasites were transferred to 5 ml of HMI-9 per cuvette and cultured overnight at 37°C. CAT reactions were for 2 h at 37°C, and assays were by xylene extraction (73). Transfections were performed in replicate (two to six times), and results presented are an average of those from a number of experiments; although values for CAT activity varied between experiments, within any one experiment relative values obtained for each construct tested were very reproducible.

Stable transformation.

Electroporation of procyclic trypanosomes was carried out exactly as described above. Five micrograms of linearized, gel-purified plasmid DNA was electroporated into procyclic culture cells derived from stock EATRO 795. Cells were allowed to recover in SDM-79 medium for 18 h in the absence of drug selection, after which phleomycin was added to a final concentration of 10 μg/ml. Cultures were maintained for 1 week to allow stable transformants to grow through (after 1 week, no viable cells remained in control cultures which had been transfected with pBluescript) and then plated onto SDM-79 semisolid agar plates containing phleomycin at 10 μg/ml (9, 10). Colonies were visible after 7 to 10 days, and these were expanded in liquid culture for isolation of nucleic acids. Stable transformation of bloodstream trypanosomes was carried out exactly as described previously (9, 10) except that only 1 μg of gel-purified, linearized, alkaline phosphatase-treated plasmid DNA was transfected per 3 × 10⁷ cells to minimize multiple integrations into the genome. Following recovery of cells overnight in HMI-9 medium at 37°C, stably transformed cells were selected by culturing in phleomycin (2 μg/ml), following which cloned cell lines were selected either by plating onto semisolid agar plates (9, 10) or by doubling dilutions in 96-well plates (minimum of 0.25 trypanosome/well) with phleomycin at a concentration of 1 μg/ml.

PCR amplification.

The primer for the 1.22 promoter region was 5′TGCGGAACTGCCGCTCATTGCACGTT3′, and the primer for the ribosomal promoter region was 5′TAAAGAGCCAGAATGCACCCGCGCTG3′. PCR amplification was performed for 20 cycles of 30 s at 94°C, 1 min at 60°C, and 1 min at 70°C in a final volume of 50 μl containing 50 mM KCl, 10 mM Tris (pH 8.3), 2.5 mM MgCl, 100 μg of bovine serum albumin per ml, and 100 pmol of each primer. PCR products were resolved by gel electrophoresis in 1.5% agarose.

RESULTS

Trypanosome lines.

We have used a trypanosome line (ILTat 1.2) which is partially laboratory adapted; it is highly virulent in mice, and although it does not always switch variable antigen type at the same high rate as trypanosomes recently isolated from the field, it can be transmitted through tsetse flies (64). In this study, we grew trypanosomes in rats for transient transfection studies, but for stable transformation experiments we established a culture-adapted bloodstream trypanosome line by the method of Carruthers and Cross (9). The culture-adapted trypanosomes are monomorphic and retain high virulence in mice, even after 5 months in continuous culture. Growth rates are similar to those reported for culture-adapted bloodstream-form trypanosomes of the laboratory-adapted strain 427 (10): doubling time was 8 to 10 h, with a maximum density of approximately 3 × 10⁶ cells/ml.

The M-VSG gene promoter is active in bloodstream-form trypanosomes.

Previously we found, using nuclear run-on analysis, that the 1.22 M-VSG gene promoter was transcriptionally active only at the right life cycle stage (in metacyclic-derived trypanosomes which continue to express M-VSG by the same metacyclic stage-specific in situ activation mechanism used in the fly [28]); it was inactive in both procyclic and bloodstream parasites (26). Further, we found that in transient transfection experiments with metacyclic-derived trypanosomes, a 420-bp fragment encompassing the transcription initiation site for the gene, and including 167 bp of sequence upstream (the BamHI/PstI insert in pMT1.22-BPs [Fig. 1A]), was almost as active as the 221 bloodstream expression site promoter (26). However, the 1.22 putative promoter exhibited only very low activity in similar experiments using procyclic trypanosomes (26). Hence, we now determined whether the 1.22 promoter was able to drive reporter gene expression in bloodstream trypanosomes. Figure 1B shows that in contrast with previous nuclear run-on experiments (26), the 1.22 promoter fragment in p122sHD52CAT (1.22s) was able to direct very high levels of CAT expression, 190% ± 40% of that obtained with the 221 bloodstream expression site promoter in pHD52CAT (B-ES). In case this fragment did not contain all the sequences necessary for regulated promoter activity in the bloodstream stage, we also tested a larger fragment for the ability to drive reporter gene expression. This was a 1.6-kb fragment (similar to the insert in p1.22-HPl but lacking 200 bp at the 5′ end, stretching from the HindIII site to the KpnI site 200 bp downstream [Fig. 1A]) which contained the entire 420-bp promoter fragment at its 3′ end. Again, the new construct gave high levels of reporter gene expression (Fig. 1B, 1.22l), showing that sequences proximal to the promoter were not responsible for down-regulation in bloodstream trypanosomes, at least as assayed by transient transfection.

Chromosomal context affects 1.22 promoter activity.

To explain the apparent anomaly that the promoter was inactive in nuclear run-on assays but highly active in transient transfection experiments in bloodstream forms, we hypothesized that the transient transfection activity resulted from an escape from control in a chromosomal context. To test this, we designed a construct, pt122BC, for stable integration into the 1.22 M-VSG gene endogenous telomere (Fig. 2A). It contained a selectable marker cassette in which the ble (phleomycin resistance) gene (21) was under the control of the constitutively active PARP B locus promoter (49, 59) and a reporter gene cassette in which the CAT gene was under the control of the same 1.6-kb KpnI/PstI fragment from the 1.22 M-VSG gene promoter region that directed high levels of CAT gene expression in the transient transfection experiments. The stable transformation construct gave levels of CAT activity similar to those of p122lHD52CAT when it was assayed in transient transfection (data not shown). It should be noted that the PARP B promoter is 5- to 10-fold less active in bloodstream trypanosomes than in procyclic cells (6). Cleaving the dual-cassette construct with SalI, a unique site within the 1.22 promoter region (Fig. 2B), allowed targeting to the 1.22 M-VSG gene expression telomere just upstream of the endogenous promoter. We also designed a very similar plasmid, pr122BC (Fig. 3B), that contained 712 bp of sequence derived from the nontranscribed spacer region of the rDNA locus (the rDNA intergenic region in pHD430 [71]) inserted in a 3′-5′ direction upstream of the 1.22 promoter fragment such that, when it was inserted into the genome, the 1.22 promoter was in reverse orientation to the direction of transcription of the rDNA locus. There is a unique NotI site within the nontranscribed spacer-homologous region which, when digested, would allow targeting to the nontranscribed spacer region (Fig. 3B).

FIG. 3.

Insertion of plasmid pr122BC into the ribosomal nontranscribed spacer region. (A) Structure of plasmid pr122BC linearized at the unique NotI site within the ribosomal locus targeting fragment. Abbreviations: Ps, PstI; Pv, PvuII; pBS, pBluescript sequences; CAT, CAT reporter gene; ble, selectable marker gene encoding phleomycin resistance. Dark stippled boxes, ribosomal locus targeting sequence; light stippled box, 1.22 promoter region; black box, procyclin/PARP promoter region; open boxes, marker genes; black flag, 1.22 promoter; white flag, procyclin/PARP promoter; dotted line, pBluescript sequence. (B) Partial map of the ribosomal locus showing the targeting site, containing a unique NotI (N) restriction enzyme site. Dark stippled box, targeting region; cross-hatched box, 18S coding region; cross-hatched flag, ribosomal locus promoter; Ps, PstI. (C) Result of integrating pr122BC into the ribosomal nontranscribed spacer region. Note that pr122BC is designed to insert in reverse orientation with respect to the ribosomal transcription unit. The horizontal black bar beneath the map in panel C represents the 1.6-kb KpnI/PstI 1.22 promoter fragment used as a probe in hybridizations in panel D. The size of the fragment expected when this probe is hybridized to PstI-digested genomic DNA stably transformed with pr122BC is shown above the map in panel C. Horizontal arrows indicate the approximate location of the primers (Mprom and Rprom) used in the PCR shown in panel E to amplify the region between the inserted plasmid and the ribosomal promoter. The size of the expected amplified fragment is shown. (D) Southern blot analysis of genomic DNA cut with PstI and fractionated on a 0.6% agarose gel isolated from track 1 (wild-type trypanosomes and track 2 (BSFribo122BC trypanosomes). The blot was hybridized with the 1.6-kb KpnI/PstI 1.22 promoter region probe shown below panel C in 5× SSC at 65°C and washed to 0.1× SSC at 65°C. (E) Linkage of the 1.22 promoter region and the ribosomal locus promoter tested by PCR. The primers Mprom (5′) and Rprom (3′) (C) were used in an amplification reaction with PstI-cut genomic DNA isolated from wild-type trypanosomes (track 1) and BSFribo122BC trypanosomes (track 2). Track M, 1-kb marker ladder. PCR products were separated by agarose gel (1.5%) electrophoresis, and the gel was stained with ethidium bromide (0.5 μg/ml).

Culture-adapted ILTat 1.2 bloodstream-form cells transformed with either pt122BC or pr122BC were selected by culturing in the presence of phleomycin at 2 μg/ml, a concentration 10-fold higher than required to kill wild-type cells. Three separate stable transformation experiments were carried out with each construct, and stably transformed lines were selected by serial dilution cloning in the presence of phleomycin at 1 μg/ml. We analyzed several uncloned cell populations for each stable transformation event and undertook cloning for those displaying, by Southern blot analysis, the map appropriate for a single integration at the correct site (Fig. 2D, 2E, 3D, and 3E). We analyzed five cloned cell lines for each integration event and observed no differences between clones within single transformation experiments. We chose four cell lines (one from each of four independent stable transformation experiments) for further analysis and determination of 50% inhibitory concentrations (IC₅₀s) (Table 1): BSFtelo122BC (clones t8 and t9) and BSFribo122BC (clones r8 and r10). Similar transformations of procyclic cells yielded four more lines, two of PFtelo122BC (clones c3 and c5) and two of PFribo122BC (clones b2 and b5).

TABLE 1.

IC₅₀s for phleomycin resistance in the four stably transformed cell lines

Plasmid	Line	Stage tested	IC50 (μg/ml)
pt122BC	BSFtelo122BC	Bloodstream	3
	PFtelo122BC	Procyclic	175
pr122BC	BSFribo122BC	Bloodstream	40
	PFribo122BC	Procyclic	600

To show that only one copy of the 1.22 M-VSG gene promoter was inserted in the telomeric locus, SacI-digested genomic DNA was separated on an agarose gel by FIGE, blotted, and probed with the 1.6-kb promoter-containing fragment. In wild-type trypanosomes, the predicted 16-kb band (Fig. 2B) was detected (Fig. 2E, track 2), whereas BSFtelo122BC cloned line t8 (Fig. 2E, track 1) displayed the predicted 23-kb band (Fig. 2C). For insertion into the ribosomal locus, the same promoter probe reveals the predicted appearance of a new 7.0-kb PstI fragment (Fig. 3C; Fig. 3D, track 2), in addition to the 1.8-kb fragment containing the endogenous 1.22 promoter (Fig. 3D, track 1). These two bands in the transformed line BSFribo122BC have similar hybridization intensities, suggesting that only a single copy of pr122BC has integrated (Fig. 3D, track 2). To show linkage of the inserted 1.22 promoter to ribosomal locus sequences, we performed PCR with one primer specific to the 1.22 M-VSG gene promoter and another specific to a sequence just 5′ of the rRNA promoter (arrows in Fig. 3C). Figure 3E shows, in BSFribo122BC clone r10 (track 2) but not in wild-type trypanosomes (track 1), the expected single product of 1,500 bp (Fig. 3C). No additional, higher-molecular-weight products were observed, suggesting again that a single copy of pr122BC had integrated into the genome.

To compare directly the levels of expression of the reporter gene under the control of the 1.22 promoter in both chromosomal locations, we attempted to differentiate the cultured, stably transformed bloodstream cell lines to the procyclic stage by using reduction in temperature to 27°C and the citric acid cycle intermediates citrate and cis-aconitate in SDM-79 medium (7). This was unsuccessful, and so we carried out a new series of differentiation experiments using differentiation trypanosome medium (47). This time, although trypanosomes switched on expression of procyclin/PARP, an event early in differentiation, their morphology and growth rate were abnormal and they could not grow in SDM-79 medium (data not shown). A third series of attempts at differentiating these cell lines was made, using a modification (6a) of the method of Overath et al. (47) where cells are cultured in Cunningham’s medium (19) supplemented with citrate and cis-aconitate at 37°C for 17 h and then placed at 27°C to allow differentiation. This procedure was also unsuccessful, and no dividing population was obtained whether the initial cell population was cultured bloodstream trypanosomes or these same cells which had been expanded in mice. The untransformed culture-adapted trypanosomes were also resistant to differentiation. Since the stably transformed bloodstream trypanosomes could not undergo differentiation to yield a dividing population of cells which were morphologically procyclic, we carried out similar stable integrations, using exactly the same constructs, in procyclic trypanosomes of the same stock, EATRO 795. Southern blot and PCR experiments showed that the integrations with pt122BC and pr122BC had proceeded as expected, with single copies of the plasmids having integrated in the cultured, stably transformed procyclic trypanosome lines PFtelo122BC, cloned line c3, and PFribo122BC, cloned line b5 (data not shown). Because these new lines were not directly derived from the bloodstream stable transformants, it is not possible to compare them directly. However, the integrations obtained were very similar for both the telomeric and chromosome-internal loci in the bloodstream and procyclic trypanosomes, although we cannot rule out the possibility that the construct was inserted in a different position whose restriction pattern is very similar to that in which the bloodstream integration occurred.

To establish a baseline for measuring activity of the integrated 1.22 promoter, we first measured the activities of the inserted procyclin/PARP promoter between different loci in procyclic and bloodstream trypanosomes. The results in Table 1 show that the procyclic cell lines had high resistance to phleomycin but that the bloodstream-form cell lines had very low phleomycin resistance, as expected from the procyclin/PARP promoter being 5- to 10-fold more active (49), procyclin/PARP RNA processing signals being used more efficiently (35) in the procyclic stage, and mRNAs containing procyclin/PARP 3′ untranslated regions being much less stable in bloodstream trypanosomes than in procyclic trypanosomes (33). We found that activity of the procyclin/PARP promoter was much higher in a chromosome-internal location than in a telomeric environment in the bloodstream (the IC₅₀ for BSFtelo122BC is 13-fold higher than that for BSFribo122BC). The promoter was also more active in a chromosome-internal position than in the telomeric site in procyclic trypanosomes, although the difference in activity between the two locations was not so marked (3-fold).

To measure 1.22 promoter activity at its endogenous telomere and in the chromosome-internal position in the stably transformed bloodstream-form trypanosomes, we used Northern blotting and CAT assays. Figure 4A shows that a 1.4-kb transcript hybridizes to a ³²P-labeled CAT antisense RNA probe in track 3, which contains total RNA from BSFribo122BC trypanosomes, but no RNA can be detected in wild-type (track 1) or BSFtelo122BC (track 2) trypanosomes. When the same blot was stripped of probe and rehybridized with the ble probe, no ble RNA was detected in wild-type cells (track 1), a low level was detected BSFtelo122BC cells (track 2), but ble transcripts were abundant in BSFribo122BC cells (track 3). This was consistent with the observation that the line with the construct inserted in the nontranscribed spacer region of rRNA was greater than 10-fold more resistant to phleomycin than the line with a similar insertion at the 1.22 expression telomere. As a control for RNA loading, the actin probe revealed the same signal in all tracks (Fig. 4C). Since there was a possibility that either, or both, of the promoters inserted upstream in the telomeric locus might cause transcription elongation downstream of the insertion, the blot in Fig. 4A was also hybridized with a probe for the 1.22 M-VSG gene, but no VSG mRNA was detected (data not shown): transcription must terminate 5′ of the VSG gene. Preliminary evidence using nuclear run-on experiments indicates that termination occurs within the pBluescript sequence downstream of the selectable marker cassette in the integrated plasmid (data not shown).

FIG. 4.

Northern blot analysis of reporter gene expression driven by the 1.22 promoter in stably transformed trypanosomes. (A to C) Assay with bloodstream stable transformants. Total RNA was isolated from cultured bloodstream-form cells of the wild type (track 1), BSFtelo122BC (track 2), and BSFribo122BC (track 3). (D to F) Assay with procyclic stable transformants. Total RNA was isolated from cultured bloodstream-form cells of the wild type (track 4), PFtelo122BC (track 5), PFribo122BC (track 6), and from cells transiently transfected with p5′parpCAT3′parp and p5′parpble3′parp, constructs where either the CAT or ble gene was expressed from a procyclin/PARP promoter (track 7). RNA was fractionated on a denaturing formaldehyde gel, blotted onto a nylon membrane, and hybridized sequentially with the ³²P-labeled probes shown. Hybridization was in 3× SSC–50% formamide at 42°C. The ble and actin probes were labeled by random priming, and the CAT probe was a ³²P-labeled in vitro-transcribed CAT antisense probe. Blots were washed to 0.5× SSC at 65°C. Following each hybridization, the probe was removed by boiling in 0.5× SSC and the filter was autoradiographed to check that no residual hybridization remained.

Figure 4D shows the results for procyclic stable transformants. The CAT gene is not expressed in wild-type cells (track 4) or in either of the procyclic culture cell lines with the construct integrated at the telomeric (PFtelo122BC; track 5) or chromosome-internal (PFribo122BC; track 6) position. CAT transcripts were detected in RNA from cells transiently transfected with p5′parpCAT3′parp and p5′parpble3′parp simultaneously (track 7). In contrast, Fig. 4E shows that the ble gene is highly expressed both in the telomeric location (track 5) and in the chromosome-internal site (track 6). As before, the actin probe showed that similar amounts of RNA were present in all tracks (Fig. 4F).

To confirm the results of Northern blot analysis, we determined the levels of CAT activity in the eight stably transformed, cloned cell lines. No activity was detected above background levels in any procyclic cell line, and Table 2 shows that none of the bloodstream lines expressed active CAT enzyme significantly above background, except BSFribo122BC (clones r8 and r10), in which CAT activity was around 25- and 40-fold, respectively, above background levels. Other cell lines stably transformed with pr122BC which we studied showed CAT activities that were between 20- and 50-fold above background levels. The difference between these lines may be due to pr122BC inserting at different ribosomal loci which are transcriptionally active at different levels (6).

TABLE 2.

CAT enzyme activity in cloned bloodstream cell lines stably transformed with pr122BC and with pt122BC^a

Plasmid	Clone	CAT activity (% of r10 activity)	CAT activity (cpm/107 cells)c
pr122BC	r10	100	27,651
	r8	66 ± 4	18,250
pt122BC	t8	4.1 ± 0.5	1,134
	t9	4.4 ± 0.5	1,217
None	Wild type	2.6 ± 0.6	719

^aAll cultures were grown to mid-log phase (5 × 10⁵ cell/ml) before being subjected to transient transfection. 

^bCAT activity obtained for wild-type cells with no insertion in the genome is given as background activity. To calculate standard deviations, four experiments were performed for each clone, and each percentage represents the mean of results from triplicate assays. 

^cMean of results of triplicate assays from one representative experiment. 

Taken together, these results suggest that the 1.22 promoter is, at most, only minimally active at its endogenous locus in both procyclic and bloodstream trypanosomes. However, it appears to be highly active, in bloodstream forms alone, if it is removed from this locus either to a chromosomal internal position or onto an episomal vector. To ascertain that CAT expression from the nontranscribed spacer region of rRNA was directed by the 1.22 promoter, we performed nuclear run-on analysis. Figure 5D shows that ³²P-labeled nascent transcripts isolated from BSFtelo122BC trypanosomes hybridized to only the 765-bp fragment 2 of XbaI/HindIII-digested pMT122-HPl (Fig. 5B), which is known to contain the transcription initiation site for the 1.22 M-VSG gene (Fig. 5A). Hybridization was also detected to CAT sequences in p5′parpCAT3′parp (Fig. 5D, track 4) and to pBluescript sequences (Fig. 5D, tracks 3 and 4), indicating that there must be transcriptional readthrough from either the 1.22 or procyclin/PARP promoter. When a similar Southern blot was hybridized with ³²P-labeled transcripts from wild-type trypanosomes, no hybridization to pMT122-HPl or to p5′parpCAT3′parp was detected (Fig. 5E, tracks 3 and 4). Both nascent transcript probes hybridized to rDNA (Fig. 5D and E, tracks 1) and to tubulin sequences (Fig. 5D and E, tracks 2). To avoid excessive signal, there is 1/10 as much DNA loaded in tracks 1 (pR4) as in tracks 2 to 4. Tracks 3 and 4 of the Southern blot in Fig. 5D and E have been subjected to a longer exposure than tracks 1 and 2, but exposure of the Southern blot in Fig. 5D, tracks 3 and 4, is the same as in Fig. 5E, tracks 3 and 4. Since the fragment upstream of the 1.22 promoter region did not hybridize to the nascent transcript probe from the stably transformed trypanosomes, transcription of the CAT gene most likely initiated within the 1.22 promoter region itself.

FIG. 5.

Nuclear run-on analysis of transcription initiation of the CAT gene inserted in the ribosomal locus in stably transformed bloodstream cells. (A) Partial map of the nontranscribed spacer region of rDNA with plasmid pr122BC inserted. (B) Restriction map of plasmid pMT1.22-HPl from which was derived the 1.22 promoter region driving CAT reporter gene expression in pr122BC. The three DNA fragments which should result from digesting plasmid pMT1.22-HPl with HindIII and XbaI are labeled 1, 2, and pBS. Abbreviations: H, HindIII; K, KpnI; X, XbaI; P, PstI; pBS, pBluescript sequence. Dark grey box, 1.22 promoter region; light grey box, 18S rRNA gene; open box, CAT gene; open flag, ribosomal promoter; black flag, 1.22 promoter. (C) Ethidium bromide-stained gel of a PstI digest of pR4, an rDNA repeat unit (39) (track 1), pTbαβ-T1, an αβ-tubulin repeat unit (63), digested with HindIII (track 2), pMT1.22-HPl digested with HindIII/XbaI (track 3), and p5′parpCAT3′parp digested with HindIII/PstI (track 4). (D) Result of hybridizing a Southern blot of the gel in panel C with a ³²P-labeled nascent transcript probe from nuclei isolated from BSFribo122BC trypanosomes, cloned line r10. (E) A blot very similar to that used in panel D, hybridized with a ³²P-labeled nascent transcript probe from nuclei isolated from wild-type (wt) trypanosomes. To avoid excessive signal, there is 1/10 the amount of DNA loaded in tracks 1 (pR4) as in tracks 2 to 4. Tracks 3 and 4 of the Southern blots in panels D and E have been subjected to a longer exposure than tracks 1 and 2, but the exposure of the Southern blot in panel D, tracks 3 and 4, is the same as in panel E, tracks 3 and 4. Hybridizations were in 3× SSC at 55°C, and blots were washed to 0.1× SSC 65°C.

DISCUSSION

Life cycle stage-specific gene expression.

Most genes in Kinetoplastida are organized in polycistronic transcription units and are regulated, at least partly, posttranscriptionally (13, 25). Some modulation of activity also occurs during the life cycle of T. brucei for promoters of the complex polycistronic bloodstream expression sites (55) and the polycistronically transcribed procyclin/PARP loci (6, 49, 69), but metacyclic VSG genes remain the only example of monocistronic transcription units which are transcriptionally controlled during the parasite life cycle (26). The M-VSG gene promoters that we have studied are active only at the metacyclic stage; they are inactive in bloodstream and procyclic trypanosomes in their proper genomic context (26). The metacyclic population makes use of a special subset of VSG genes from which one can be activated by random promoter activation in each trypanosome, thereby presenting an antigenically mixed population, a situation that may be important in facilitating establishment of infection in partially immune hosts in the field (4). We have proposed previously that such random promoter activation is most easily achieved through transcriptional upregulation of M-VSG genes at the metacyclic stage. Investigation of the mechanisms of metacyclic stage-specific transcriptional control is very difficult, owing to the low numbers of metacyclic cells in the salivary glands of tsetse flies. Such direct experiments require fly transmission of our stably transformed cell lines, which we have not yet been able to achieve. However, by studying how the 1.22 M-VSG gene promoter is silenced at other life cycle stages, we have elucidated some mechanisms influencing life cycle stage-specific control. We have studied separate integration events in bloodstream and procyclic trypanosomes of the same stock, since it proved impossible to differentiate the stably transformed in vitro-cultured bloodstream cell lines. The telomeric insertion construct pt122BC can integrate into only one region of the genome, the telomeric haploid expression site for the 1.22 M-VSG gene. However, since there are many copies of the ribosomal repeat unit in T. brucei, it is possible that the construct pr122BC integrated in different rDNA arrays in the bloodstream and procyclic stably transformed cell lines. Thus, it is not possible to make direct comparisons between the results obtained with the bloodstream and procyclic stable transformants, and we deal with the data from each life cycle stage separately. Our results indicate, once again, that the 1.22 M-VSG gene promoter is under transcriptional regulation during the trypanosome life cycle. It is always silent at the procyclic stage, and although it has the potential to act at a high level in bloodstream trypanosomes, it is inactive at its endogenous telomeric location at this stage.

There are two criticisms possible of the constructs we used. First, they contained two juxtaposed expression cassettes. This was because we did not wish to use the antibiotic resistance gene as the reporter gene in case, under selection conditions, we forced activation of an otherwise silent promoter. However, a possible complication of using two linked cassettes is that procyclin/PARP promoter-dependent cis activation of a VSG expression site promoter has been observed when both promoters were inserted close to each other, in the rpo2A locus of trypanosomes (67). We do not believe that the presence downstream of the procyclin/PARP promoter altered the activity of the upstream 1.22 promoter in our experiments, since the effect was originally demonstrated in procyclic trypanosomes and we observed no 1.22 promoter activity at this stage. This may be because the procyclin/PARP promoter is downstream of, and in the same orientation as, the second promoter in our constructs, while for the cis activation of a bloodstream expression site promoter it was upstream and in the opposite orientation, or it may be due to the distance between the two promoters in our constructs being twice that of those inserted in the rpo2A locus (67). Second, the constructs that we used contained procyclin/PARP RNA processing signals which, although used very efficiently in the procyclic form, are inefficient in bloodstream trypanosomes. Use of the same signals was necessary, as we wished to study exactly the same constructs (with procyclin/PARP RNA processing signals) in both life cycle stages, to allow some comparison of promoter activities and rigorous testing of the apparent inactivity of the M-VSG gene promoter at the procyclic stage. Not all of the signals necessary to direct accurate polyadenylation of the reporter and selectable marker genes were present in the constructs, since only the 3′ untranslated region of the procyclin/PARP gene was present and sequences in the procyclin/PARP intergenic region are also required to direct accurate polyadenylation (34, 58). However, we found that the transcripts encoded by the stably integrated plasmids are polyadenylated, but approximately 100 nucleotides downstream of the normal polyadenylation site for the CAT-procyclin chimeric RNA (the end of the CAT message contains sequences from the 5′ end of the procyclin/PARP promoter fragment used in the dual cassette construct) and 115 nucleotides upstream of the normal polyadenylation site for the ble-procyclin chimeric transcripts. We also found that the sites of polyadenylation were very similar in both bloodstream and procyclic trypanosomes (data not shown).

Down-regulation in the bloodstream at a telomeric locus.

The 1.22 promoter can drive high levels of CAT gene expression in transient-transfection experiments in bloodstream trypanosomes, in contrast to nuclear run-on studies, in which promoter activity was undetectable in bloodstream trypanosomes (26). The hypothesis that we have tested is that positioning of the 1.22 promoter in its locus close to the end of the chromosome down-regulates its potential activity in bloodstream forms. We have found that the 1.22 promoter has as much as 40-fold more activity when it is located in the nontranscribed spacer region of rDNA as it displays at its endogenous telomere. The level of derepression of this promoter, when it is removed from the telomere, seems to be greater than that of the procyclin/PARP promoter, which shows around 10-fold up-regulation (Fig. 4B; Table 1). The nontranscribed spacer region of rDNA was chosen as a site for integration because it was known to be transcriptionally silent (55) and transcribed by RNA polymerase I, the polymerase which most probably transcribes VSG genes (11, 12, 36). Although we cannot discount the possibility that this location exerts some positive regulatory effect on the 1.22 promoter itself, the promoter also displays high activity on an episomal vector; thus, one possible explanation is that the observed promoter activation is a result of removal from telomere locus-specific repression.

A putative telomere position effect in trypanosomes was demonstrated in experiments in which a bloodstream expression site promoter, the procyclin/PARP promoter, and a ribosomal promoter, each inserted in a telomere-proximal position in an inactive bloodstream expression site, were repressed, and this effect was stable, reversible, and developmentally regulated (31). However, whether the telomeric location alone is responsible requires further study. In a different study which showed that there was no VSG expression site promoter sequence specificity requirement for expression site switching (54), it was proposed that bloodstream expression site promoters might be subject to control by an epigenetic mechanism, akin to telomere silencing (54). This proposed effect would have to operate far beyond the sorts of distance reported for yeast, in which telomeric silencing extended maximally, following overexpression of the SIR3 silencing protein, only up to 20 kb from the tract of hexanucleotide repeats at the end of the duplex (52, 61). It is becoming apparent that switching between trypanosome VSG gene expression sites in the bloodstream may be regulated, at least in part, by a transcriptional mechanism possibly linked to chromatin-associated effects (20, 32). One important difference between bloodstream expression site promoters and M-VSG gene promoters is that the former are located 40 to 60 kb upstream of the telomere end (18, 30, 38), while the latter are very close to the end of the chromosome (26). In fact, metacyclic VSG gene 3′ ends are within a few hundred base pairs of the telomere tract and their promoters are probably no more than 5 kb away (24a), a distance that is more likely to accommodate a telomere position effect. The chromosome ends from which M-VSG genes are transcribed have a structure very different from that of bloodstream expression sites, being flanked upstream by transcriptionally silent regions 13 to 15 kb in length (26). Further studies have revealed that there are independently controlled transcription units immediately upstream of the M-VSG gene transcription units (28a). The noncoding region between the two different transcription units may represent the extent of the effect of telomere repression at the ends of these large trypanosome chromosomes.

Down-regulation in procyclic trypanosomes.

Whether assayed on an episomal vector by transient transfection or in the genome by nuclear run-on and in stable transformation experiments, the 1.22 metacyclic promoter is found always to be inactive at the procyclic stage. This is in marked contrast to bloodstream expression site promoters, which direct high levels of reporter gene expression in procyclic transient-transfection experiments (37, 72, 73) and display a low level of activity in procyclic nuclear run-on experiments (49) and in experiments using integration of a reporter gene into chromosomes (55). Inactivity of the M-VSG gene promoter in procyclic trypanosomes appears not to be related to its chromosomal positioning, as it is inactive even when placed in the nontranscribed spacer region of rRNA. This finding suggests that cis-acting sequences within the promoter itself mediate down-regulation at this life cycle stage.

Another class of putative metacyclic VSG gene promoters has been identified by virtue of their high activities in procyclic-form transient-transfection experiments (1, 46). These putative promoters were isolated by cloning regions upstream of an M-VSG gene whose in situ expression in bloodstream trypanosomes is rare and was detected only following very extensive selection in mice (42). It is not known whether these putative promoters can act in the metacyclic stage or whether they display life cycle stage regulation in the genome in vivo. The in situ activation and transcriptional regulation mechanisms which we have uncovered for the 1.22 promoter (which is stage regulated and randomly activated) fit well with what was predicted previously from observations at the phenotypic level (62, 66). This promoter appears to represent a novel class of VSG gene promoters which are truly developmentally regulated, displaying inactivity in procyclic and bloodstream trypanosomes and specific up-regulation at the metacyclic stage.

Up-regulation at the metacyclic stage.

We have uncovered two different life cycle stage-specific control mechanisms which contribute to silencing of the metacyclic promoter in its endogenous telomere outside the metacyclic stage: locus-associated in bloodstream forms and involving cis-acting sequences within the promoter itself at the procyclic stage. One consequence is that activation of M-VSG gene expression at the metacyclic stage is stringently controlled. It requires not only lifting of procyclic stage-specific repression or appearance of a metacyclic stage-specific transcriptional activator but also the activation of the mechanism for stochastic activation of one M-VSG gene from the M-VSG repertoire. A candidate for involvement in this mechanism is telomere position effect, one characteristic of which, in yeast, is the reversible transcriptional repression of telomeric promoters (57). Both metacyclic and bloodstream VSG gene promoters are subject to a silencing mechanism in the appropriate life cycle stage that ensures that only one promoter is active at a time. In the bloodstream, it has proven difficult to find a single factor associated with this exclusivity (20, 32), one possible conclusion being that physical interactions between chromosomes are involved. This would be compatible with current evidence showing that in yeast, interaction between telomeres, even on nonhomologous chromosomes, is extensive (24). The metacyclic promoters may have a simpler means of achieving mutually exclusive activation. As the decision to activate M-VSG genes coincides with the random selection of an individual promoter, cross talk between telomeres may not be necessary. Instead, the binding of an activating protein complex by one telomere may exempt it from repression.

It is not surprising that M-VSG gene promoters can be active in bloodstream forms, as we have found, since trypanosomes continue to express metacyclic VSG genes for up to 7 days following transfer to the mammal, despite the parasites having already differentiated to bloodstream forms (28). Metacyclic VSG gene promoters must be able to recruit bloodstream stage-specific transcription factors to allow transcription initiation. Finally, our results suggest that the repression of M-VSG gene expression in the bloodstream is less rigorous than in procyclic cells. Perhaps another layer of control that we have not been able to detect in these experiments is exerted in the bloodstream.