|Home | About | Journals | Submit | Contact Us | Français|
Trypanosoma brucei expresses Variant Surface Glycoprotein (VSG) genes in a strictly monoallelic fashion in its mammalian hosts, but it is unclear how this important virulence mechanism is enforced. Telomere position effect (TPE), an epigenetic phenomenon, has been proposed to play a critical role in VSG regulation, yet no telomeric protein has been identified whose disruption led to VSG derepression. We now identify tbRAP1 as an intrinsic component of the T. brucei telomere complex and a major regulator for silencing VSG expression sites (ESs). Knockdown of tbRAP1 led to derepression of all VSGs in silent ESs, but not VSGs located elsewhere, and resulted in stronger derepression of genes located within 10 kb from telomeres than genes located further upstream. This graduated silencing pattern suggests that telomere integrity plays a key role in tbRAP1-dependent silencing and VSG regulation.
Trypanosoma brucei is a unicellular protozoan parasite that causes Human African Trypanosomiasis. In the bloodstream of its mammalian host, T. brucei periodically switches the major component of its surface coat, the Variant Surface Glycoprotein (VSG; Barry and McCulloch, 2001), thereby evading immune elimination. Although there are more than 1000 VSG genes and pseudogenes in the T. brucei genome (Berriman et al., 2005; Marcello and Barry, 2007), VSGs can only be expressed from one of ~20 (Navarro and Cross, 1996) nearly identical VSG expression sites (ES; Barry and McCulloch, 2001; Hertz-Fowler et al., 2008), which are polycistronically transcribed by RNA Polymerase I (Pol I; Gunzl et al., 2003) and are located immediately upstream of telomeres (de Lange and Borst, 1982). In any ES, the VSG is 0.2–1.6 kb upstream of the telomere DNA repeats, while the promoter is 40–60 kb upstream. Promoter-less VSG genes are also found on minichromosomes within 5 kb of telomeres, or on megabase chromosomes in gene clusters located at subtelomeric regions (Horn and Barry, 2005).
Mechanisms that ensure the monoallelic VSG expression remain elusive, even though several hypotheses have been proposed (Pays et al., 2004). One study showed that ‘silent’ ES promoters are actually moderately active, but that transcription is rapidly attenuated, suggesting that transcription elongation is regulated (Vanhamme et al., 2000). Another study showed that only the active ES colocalizes with Pol I at an extranucleolar site, dubbed the Expression Site Body (ESB; Navarro and Gull, 2001; Landeira and Navarro, 2007), suggesting that recruitment or limitation of a single ES to the ESB is crucial. Recent studies showed that depletion of a T. brucei homologue of the chromatin remodeling factor ISWI (TbISWI) elevated ‘silent’ ES promoter activities without affecting VSG transcription (Hughes et al., 2007), and deletion of tbDot1B, a histone H3 methyltransferase, led to a 10-fold increase in ‘silent’ VSG transcripts (Figueiredo et al., 2008), suggesting that chromatin modifications are important for ES silencing. Finally, as VSG is exclusively expressed from subtelomeric loci (de Lange and Borst, 1982), it has been proposed that the telomere structure plays an important role in VSG expression regulation (Horn and Cross, 1995; Horn and Barry, 2005; Glover and Horn, 2006; Dreesen et al., 2007).
Telomere-specific proteins are indispensable for telomere functions, which include protection of chromosome ends and maintenance of telomere lengths and the specialized telomere chromatin structure. Much is known about the mammalian telomere complex, which contains six core proteins (de Lange, 2005). Among these, TRF2 binds duplex telomere DNA and maintains chromosome end integrity. So far, tbTRF, a functional homologue of TRF2, is the only known T. brucei telomere-specific protein (Li et al., 2005). Another mammalian telomeric protein is RAP1, which does not contact DNA but is recruited to telomeres through its interaction with TRF2 (Li et al., 2000). In contrast, S. cerevisiae RAP1 is the predominant duplex telomere-DNA-binding factor (Shore, 1994). The difference between the DNA binding activities of hRAP1 and scRAP1 can be explained by the fact that scRAP1 has myb and myb-like domains for DNA recognition while hRAP1 has a single myb domain with a negative surface charge on its third helix (Konig and Rhodes, 1997; Hanaoka et al., 2001).
ScRAP1 is essential for the telomeric heterochromatic structure, which can repress the transcription of subtelomeric genes, an epigenetic phenomenon termed telomere position effect (TPE; Gottschling et al., 1990). For genes targeted to subtelomeric loci, TPE is typically stronger at telomere-proximal than telomere-distal regions, spreading continuously inwards from telomeres (Renauld et al., 1993). However, TPE at native chromosome ends can be discontinuous, with a peak of repression at subtelomeric regions not immediately adjacent to telomeres (Pryde and Louis, 1999). TPE has also been observed in T. brucei and mammalian cells when reporter genes are analyzed (Glover et al., 2007; Glover and Horn, 2006; Gao et al., 2007; Pedram et al., 2006).
Many other proteins are involved in TPE, including the NAD+-dependent histone deacetylases scSir2 and mammalian SIRT6 (Gasser and Cockell, 2001; Michishita et al., 2008), and yeast Ku70/80, which binds DNA ends (Mishra and Shore, 1999; Evans et al., 1998; Boulton and Jackson, 1998). The T. brucei Sir2 homologue, SIR2rp1, is also important for TPE (Alsford et al., 2007). Interestingly, the P. falciparum Sir2 homologue is essential for the telomeric heterochromatic structure and necessary for the monoallelic expression of subtelomeric var genes, which are critical virulence genes involved in antigenic variation (Freitas-Junior et al., 2005; Duraisingh et al., 2005). In T. brucei, however, there is no direct evidence supporting the idea that telomeres are important for monoallelic VSG expression. Depletion of tbTRF, telomerase, SIR2rp1, or tbKu80 had no effect on VSG expression (Li et al., 2005; Alsford et al., 2007; Glover et al., 2007; Janzen et al., 2004; Conway et al., 2002).
In this study, we report the identification of a novel T. brucei telomeric protein, tbRAP1, which interacts with tbTRF, associates with telomere DNA, and is essential for silencing ES-associated VSGs.
We previously identified tbTRF as a telomeric protein that is essential for the telomere terminal structure but not VSG expression (Li et al., 2005). To identify additional telomere components, we carried out a yeast 2-hybrid screen using tbTRF as bait. More than ten tbTRF-interacting candidates were identified, including the conserved hypothetical protein Tb11.03.0760. This protein has an N-terminal BRCA1 C-Terminus (BRCT) domain (Figure 1A and 1B), which is present in many proteins involved in DNA damage repair or cell cycle checkpoints (Bork et al., 1997). Since RAP1 is the only known telomeric protein with an N-terminal BRCT domain (Li et al., 2000; Kanoh and Ishikawa, 2001; Park et al., 2002; Callebaut and Mornon, 1997), we compared this protein with known RAP1s more carefully. ClustalW analysis showed that this protein also has two central regions very similar to the myb and myb-like domains in scRAP1 (Figure 1), so we named this protein tbRAP1.
The sequence identity between tbRAP1 BRCT and other RAP1 BRCTs is 8–18% (ClustalW), with the putative β1 β-sheet being the most conserved region (Figure 1B; Zhang et al., 1998). The sequence homology is stronger within the myb and the myb-like domains, with identities of 18–22% and 10–23%, respectively (Figure 1C and 1D). Interestingly, both tbRAP1 and spRAP1 lack the C-terminal RCT protein-protein interaction domain (Figure 1A; Li et al., 2000), suggesting that tbRAP1 uses a different motif to interact with other proteins. Indeed, the N-terminal two thirds of tbRAP1 (aa 2–653) is sufficient and necessary for the interaction with tbTRF (Figure 2A). Compared to tbTRF self-interaction, the affinity of tbRAP1 to tbTRF is lower (Figure 2A). This is similar to the interaction between hRAP1 and hTRF2, which is weaker than hTRF2 self-interaction in a yeast 2-hybrid assay (Broccoli et al., 1997; Li et al., 2000).
To confirm that tbRAP1 interacts with tbTRF in vivo, we carried out co-IP using a rabbit antibody against tbTRF or a monoclonal antibody against the HA epitope tag, in cells containing a FLAG-HA-HA (F2H)-tagged tbRAP1 at its endogenous locus. In several independent experiments, we observed 3–14% of F2H-tbRAP1 co-IP with the endogenous tbTRF and nearly 10% of tbTRF co-IP with the F2H-tbRAP1 (Figure 2B), confirming that tbRAP1 interacts with tbTRF in vivo, though weakly.
To confirm that tbRAP1 is a telomeric protein, we examined its subnuclear localization. Three T. brucei cell lines in which an endogenous tbRAP1 was tagged with an N-terminal Ty1, GFP, or F2H epitope were established. Indirect immunofluorescence (IF) was carried out with anti-tag monoclonal antibodies and an anti-tbTRF rabbit antibody. All the Ty1-, GFP-, and F2H-tagged tbRAP1 proteins partially colocalized with the endogenous tbTRF (Figure 2C; data not shown), which was confirmed by using an anti-tbRAP1 rabbit antibody (data not shown). These data suggest that tbRAP1 is localized at telomeres and possibly in other subnuclear compartments. This is similar to scRAP1, which targets 5% of the promoters in the yeast genome, in addition to telomere DNA (Pina et al., 2003).
To further confirm that tbRAP1 associates with the telomere, we carried out Chromatin ImmunoPrecipitation (ChIP) using a rabbit antibody against tbRAP1. We observed that telomere DNA was enriched in tbRAP1 and tbTRF ChIP (Figure 2D). In contrast, control DNAs, including the tubulin gene array and a random single-copy gene Tb11.0330, were not enriched in tbRAP1 or tbTRF ChIP under the same conditions. Nor was telomere DNA enriched in any ChIP without formaldehyde cross-linking. TbTRF antibody precipitated more telomere DNA than tbRAP1 in normal ChIP (Figure 2D), which is consistent with the observation that all tbTRF but not all tbRAP1 seems to reside at telomeres.
TbRAP1 double-allele knockout cell lines could not be established even though cells having one tbRAP1 allele deleted showed no growth defect (data not shown), suggesting that tbRAP1 is essential for cell growth. We therefore established inducible tbRAP1 RNAi cell lines: Ri-2 in VSG2-expressing SM cells (Wirtz et al., 1999) and Ri-9 in VSG9-expressing PVS3-2/OD1-1 cells (derived from SM cells). Both cell lines have a Ty1-tagged endogenous tbRAP1 allele and constitutively express the T7 RNA polymerase and the tet repressor. Ri-2 and Ri-9 are therefore isogenic but express different VSGs. After inducing tbRAP1 RNAi, western analysis showed that Ty1-tbRAP1 was greatly diminished by 12 hr and was undetectable by 36 hr (Figure 3A), indicating that knockdown of tbRAP1 was effective. Growth of these cells was arrested by ~48 hr after induction (Figure 3B), confirming that tbRAP1 is essential for normal cell growth.
We next examined whether depletion of tbRAP1 led to any telomere defect. To avoid non-specific effects, tbRAP1 RNAi was induced for no more than three days, and three Population Doublings (PDs) passed before growth was arrested. Telomere lengths were compared before and after the RNAi induction by Southern blotting, but no obvious telomere length changes were observed (data not shown).
Since scRAP1 is well known for its essential role in TPE (Grunstein, 1997), we examined the possibility that tbRAP1 may play a role in VSG regulation.
The T. brucei Lister 427 strain contains 14 different VSGs in 15 ESs (Hertz-Fowler et al., 2008). We examined the expression of all 14 VSGs using quantitative real-time RT-PCR. To avoid effects from differences in mRNA stability and PCR efficiency, we compared mRNA levels for individual VSGs before and after induction of tbRAP1 RNAi.
In two independent Ri-2 lines (Table 1A), all VSG mRNA levels (except VSG2) were increased 2–25 or 8–56 fold after 18 or 36 hr of tbRAP1 RNAi induction, respectively. At both time points, Ty1-tbRAP1 levels dropped to less than 6% of wild-type levels (Figure 3A). In contrast, VSG2 expression exhibited a subtle decrease, and rRNA levels were stable (Table 1). In addition to Pol I-transcribed VSGs and rRNA, we examined RNA Polymerase II (Pol II)-transcribed tbTRF, tbTERT, and histone H4 (tbHH4), and found no more than 2-fold changes in their mRNA levels (Table 1B). Since scRAP1 is also known to activate the expression of ribosomal protein and glycolytic protein genes (Pina et al., 2003), we also examined tbRPS15 (Tb927.7.2370), a putative ribosomal protein gene, and tbPGI (Tb927.1.3830), which encodes glycosomal glucose-6-phosphate isomerase. There was no more than a 3-fold increase in the corresponding mRNA levels (Table 1B), indicating that tbRAP1 is not necessary for their transcription. Finally, in parental cells lacking the tbRAP1 RNAi construct, the mRNA level for all tested genes exhibited less than 3-fold variation (Table 1).
To determine whether VSG-derepression was specific to VSG2-expressors, we examined mRNA level of various VSGs in the VSG9-expressing Ri-9 cell line. Northern analysis showed the same derepression for all tested VSGs, including VSGs 2, 3, 11, 13, 15, 17, 18, and 21 (Supplemental Figure 1). Therefore, the VSG-derepression induced by tbRAP1 knockdown is not specific to cells expressing a particular VSG. No VSG derepression was detected in tbTRF RNAi and control cells with empty RNAi vector (Supplemental Figure 1A).
To determine whether a derepressed VSG in tbRAP1 depleted cells was transcribed at a similar level as when it is in an active ES, we used quantitative real-time RT-PCR to compare the mRNA level of derepressed VSG2 in Ri-9 cells to that of VSG2 in Ri-2 cells. In Ri-9 cells, upon tbRAP1 knockdown, even when VSG2 was maximally derepressed (~100 fold), its mRNA was still 70–100 fold less abundant than that in Ri-2 cells. Similarly, when a luciferase gene was located immediately downstream of an ES promoter, it had an activity of 3000–4000 units when the ES was active, but only 1–2 units when the ES was silent (after an in situ ES switch), and a maximum of 40 units when the ES was subsequently derepressed upon tbRAP1 knockdown. Therefore, the derepressed ESs were transcribed at an intermediate level between silent and fully active. This also means that in a ‘silent’ ES, promoter-proximal and telomere-proximal genes are transcribed at ~0.3% and ~0.01% of their fully active levels, respectively.
To test whether VSG derepression requires a promoter, we analyzed the expression of two promoter-less single-copy VSGs in both Ri-2 and Ri-9 cells. VSG5 is at a non-telomeric locus (Horn and Cross, 1997), while VSGG4 is at a subtelomeric locus on a minichromosome (Alsford et al., 2001; data not shown). Although primers for these two VSGs generated specific PCR products, their expression levels were not higher than background either before or after tbRAP1 knockdown (data not shown), and no VSG5 or VSGG4 mRNA could be detected in Ri-9 cells (Supplemental Figure 1A). Therefore, tbRAP1 is only necessary for silencing ES-linked VSGs but not VSGs elsewhere in the genome.
To determine whether proteins were synthesized from derepressed VSG genes, we carried out western analysis using specific antibodies against VSGs 2, 9, and 13. In Ri-2 cells, low levels of VSG9 and VSG13 were detectable by 12 hr post induction while VSG2 was constantly expressed (Figure 4A). Similarly, in Ri-9 cells, VSG2 and VSG13 were detectable at 6 hr and 18 hr, respectively, while the expression of VSG9 remained constant (Figure 4B). In contrast, none of the silent VSG proteins were detectable in respective parent, TRF RNAi, or vector control cells (data not shown).
To determine whether multiple VSGs are expressed simultaneously in individual cells, we carried out IF analysis. In Ri-9 cells, the derepressed VSG2 and VSG13 were detected simultaneously on the cell surface (Figure 4C). Using other combinations of VSG-specific antibodies, we confirmed that VSG9 was constantly synthesized, and that the derepressed VSG2 and VSG13 were present in the same VSG9-expressing cells (data not shown). Although it was impractical for us to test whether more than two VSGs were actually being synthesized simultaneously in individual cells, it is likely that they are.
Wild-type cells have a single active ES, which associates with the only one Pol I-enriched ESB (Navarro and Gull, 2001). Since multiple ESs are derepressed upon tbRAP1 depletion, we examined whether the subnuclear localization of Pol I is affected in Ri-2 cells by IF analysis using an anti-Pol I monoclonal antibody (Navarro and Gull, 2001). We observed that the fraction of cells with more than one extranucleolar Pol I foci increased from 20 to 50% after tbRAP1 RNAi was induced for 36 hr (Figure 4D). Interestingly, some of these foci seem to be much brighter than others, suggesting that different foci contain unequal amounts of Pol I.
To confirm that VSG-derepression was specifically due to tbRAP1 depletion, we introduced an ectopic inducible F2H-tagged tbRAP1 into Ri-2 cells. Adding doxycyclin to these cells led to a decrease in Ty1-tbRAP1 and an increase in F2H-tbRAP1 protein level but no detectable VSG9 or VSG13 (Supplemental Figure 2) or any change in cell growth, indicating that the VSG-derepression phenotype resulted specifically from tbRAP1 deficiency.
Although tbTRF is not essential for VSG silencing (Li et al., 2005), it is possible that tbRAP1 depletion-induced VSG-derepression requires tbTRF function. To test this possibility, we established VSG2-expressing cell lines in which tbTRF and tbRAP1 could be knocked down simultaneously. Upon RNAi induction, VSG9 and VSG13 were still derepressed (Supplemental Figure 3), indicating that tbTRF is not required for VSG-derepression.
Depletion of VSG in T. brucei results in a cell cycle arrest at the precytokinesis stage (Sheader et al., 2005). We questioned whether expression of multiple VSGs leads to a similar cell cycle arrest. Cell cycle analysis by FACS showed that the fraction of cells in S phase is significantly reduced while the fraction in G2/M increased mildly after tbRAP1 depletion (Supplemental Table 2). Therefore, tbRAP1 deficiency led to an altered cell cycle profile that is different from what was observed in VSG-depleted cells.
TbRAP1 knockdown led to derepression of ES-linked silent VSGs, and the fold of increase in VSG mRNA should reflect the strength of tbRAP1-dependent silencing. In all ESs, VSG genes are located 0.2–1.6 kb from telomeres (Hertz-Fowler et al., 2008). As tbRAP1 is a telomeric protein, it is possible that tbRAP1-dependent silencing is stronger closer to telomeres. To determine how telomere-proximal and telomere-distal genes were affected after tbRAP1 depletion, we used quantitative real-time RT-PCR to compare the derepression of unique genes in ES1 that are located at 60 kb (PUR, puromycin-resistance), 7 kb (ΨES1, a VSG pseudogene), and 1 kb (VSG2) upstream of the telomere (Figure 5A; Hertz-Fowler et al., 2008). Since independent inductions can give substantial variations, we used the same RNA samples for all three genes in each RT-PCR reaction and present individual experiment results. We found that PUR is always derepressed by a lesser amount (because, being close to the promoter, it is already somewhat transcribed in ‘silent’ ES) than ΨES1 and VSG2, although ΨES1 was derepressed at a higher level than VSG2 (Figure 5A). Our data indicate that tbRAP1 can silence the whole ES and more strongly closer to telomeres, albeit the silencing level may fluctuate at regions immediately upstream of the telomere. The graduated silencing pattern was confirmed by a similar analysis in ES11, which is silent in both Ri-2 and Ri-9 cells and has a unique ΨES11 (another VSG pseudogene) and VSG16 located at 20 kb and 0.5 kb upstream of the telomere, respectively (Figure 5B; Hertz-Fowler et al., 2008). In both cell lines, ΨES11 was derepressed at lower levels than VSG16 (Figure 5B), indicating that tbRAP1-dependent silencing is stronger at telomere-proximal regions.
Our data from yeast 2-hybrid, co-IP, IF, and ChIP showed that tbRAP1 interacts with tbTRF and associates with telomeres, though not exclusively, indicating that tbRAP1, like scRAP1, is a telomeric protein but also locates elsewhere in the genome (Pina et al., 2003). No drastic telomere length changes were observed within three PDs when tbRAP1 RNAi was induced for three days, suggesting that tbRAP1 does not suppress rapid telomere length changes that have been observed in yeast, human, and T. brucei (Li and Lustig, 1996; Wang et al., 2004; van der Ploeg et al., 1984). However, we cannot exclude the possibility that tbRAP1 participates in telomere length control, because telomere length changes due to telomerase activity or the lack of it are very slow in T. brucei (3–6 bp/PD; van der Ploeg et al., 1984; Dreesen et al., 2005). A small change in three PDs (expected to be 9–18 bp), if there was any, would not be discernable in Southern analyses, when the average telomere sizes are ~10 kb.
The BRCT and myb domains are ubiquitous among RAP1 homologues (Teixeira and Gilson, 2005), but the function of RAP1 BRCT domain is unknown. The myb and the myb-like domains, at least in scRAP1, are required for DNA binding (Konig et al., 1996), as normally a minimum of two myb domains is necessary for DNA recognition (Ogata et al., 1994). As tbRAP1 also contains a myb and a myb-like domain, it might bind telomere DNA directly. In addition, at least 13% of thymidines in T. brucei telomere TTAGGG repeats are replaced by the glucosylated base J in the bloodstream stage (Gommers-Ampt et al., 1991; van Leeuwen et al., 1996), so it will be interesting to determine whether tbRAP1 recognizes J-containing telomeres.
To our knowledge, tbRAP1 is the first telomeric protein identified in T. brucei whose depletion led to disruption of monoallelic VSG expression. All ES-linked VSGs were derepressed upon tbRAP1 knockdown, and this effect was not specific to cells expressing a particular VSG, suggesting that tbRAP1 is a critical ES transcription suppressor. It is possible that the VSG-silencing function of tbRAP1 may be independent of the telomere structure, due to the fact that tbRAP1 is not exclusively at the telomere and the lack of obvious telomere length changes within a short period of time after tbRAP1 depletion. However, depletion of tbRAP1 did not affect the transcription of any tested Pol I- or Pol II-transcribed control genes or non-ES VSGs, indicating that tbRAP1 is not a general transcription regulator. In addition, tbRAP1-mediated silencing is stronger in regions within 10 kb of the telomere than those further upstream, suggesting that telomere structure is essential for this silencing.
The graduated strength of tbRAP1-dependent silencing is different in ES11 and ES1, indicating that silencing in different ESs are not identical. Similar differences have been observed in yeast, where TPE for subtelomeric reporter genes spreads continuously inwards from the telomere (Renauld et al., 1993), while TPE at native telomeres can have a peak of silencing not immediately upstream of the telomere (Pryde and Louis, 1999). In addition, different VSGs are derepessed at various levels upon tbRAP1 knockdown, which appears independent of VSG copy numbers or the distance between VSG and the ES promoter but reflects different levels of silencing in different ESs. Genome structure and chromosome context presumably can influence the degree of silencing. Further investigations will be necessary to determine which ES elements are influential.
In wild-type cells, ‘silent’ ESs are transcribed at a low level for a short distance (Vanhamme et al., 2000), suggesting that the transcription elongation encounters an increasing antagonizing effect when it moves towards the telomere. TbRAP1 knockdown did not affect the active VSG but led to graduated derepression for genes along a silent ES, suggesting that tbRAP1-dependent silencing antagonizes transcription elongation from ES promoters (Figure 6A). Depletion of tbRAP1 lifted the silencing and the entire ESs including VSGs were transcribed (Figure 6B). However, derepressed ESs are transcribed at a 70–100 fold lower level than when they are fully active. Hence we propose that additional mechanisms are necessary to boost an ES to full activation (Figure 6C). This hypothesis is consistent with the observation that depletion of the chromatin modifier tbISWI led to a partial promoter-proximal ES derepression (Hughes et al., 2007). Comparing silencing strength of ES-linked to that of non-ES reporter genes also suggested that a TPE-independent but ES-specific silencing is involved in VSG regulation, further supporting our hypothesis (Glover and Horn, 2006).
It has been proposed that concentrating a limiting amount of cellular Pol I transcription machinery at the ESB is an important mechanism to ensure monoallelic VSG expression (Navarro and Gull, 2001) and that, therefore, two forced fully active ESs have to switch back and forth rapidly and locate next to each other in the nucleus (Chaves et al., 1999). We observed an increase in the number of extranucleolar Pol I foci after tbRAP1 depletion, suggesting that a small amount of Pol I is available for multiple ES transcription at basal level.
Telomeres have been proposed to be involved in VSG regulation ever since the discovery that VSGs are expressed from subtelomeric loci (de Lange and Borst, 1982). Recent studies confirmed that TPE exists in T. brucei but also suggested that telomeres might not be essential for VSG silencing (Horn and Cross, 1995; Horn and Cross, 1997; Glover et al., 2007; Glover and Horn, 2006). We have identified tbRAP1 as a telomeric protein and showed that it is essential for complete ES VSG silencing. Our data strongly support the hypothesis that telomeres are important for the regulation of VSG expression.
In budding yeast, TPE depends on scRAP1, which binds to telomere DNA and recruits Sir proteins to establish a heterochromatic structure at telomeres (Grunstein, 1997). Recent studies indicated that silencing of subtelomeric VSG genes also depends on specific modifications of the chromatin structure (Figueiredo et al., 2008; Hughes et al., 2007). However, tbRAP1, but not SIR2rp1 or tbKu80, is essential for VSG silencing, indicating that tbRAP1-mediated silencing is not exactly the same as TPE in yeast. In addition, TPE in S. cerevisiae was studied using genes transcribed by Pol II (Gottschling et al., 1990), whereas in T. brucei, VSGs are transcribed by Pol I (Gunzl et al., 2003). It is possible that the chromatin structure affects transcription by different polymerases in different ways. Furthermore, in budding yeast, TPE affects initiation of gene transcription, where a promoter is immediately upstream of the reporter gene, while tbRAP1-mediated silencing seems to affect transcription elongation from ES promoters, which are 40–60 kb upstream of VSGs. These observations suggest that not all RAP1 homologues have identical functions and that further studies of tbRAP1 will be required to identify its precise roles in regulating VSG expression.
The full-length tbRAP1 was PCR amplified from T. brucei TREU927 genomic DNA and inserted into pLEW82-BSD (Li et al., 2005) together with an N-terminal F2H tag to generate an inducible tbRAP1 expression construct. The same fragment was inserted into p2T7-TABlue (Alsford et al., 2005) to make an inducible tbRAP1 RNAi construct. The subcloned full-length tbRAP1 was sequenced and the result submitted to GenBank (accession FJ597175). The hygromycin-resistance gene or PUR flanked by tbRAP1 5′ and 3′ UTR were inserted into pBluescript SKII+ to make two tbRAP1 deletion constructs. The phleomycin resistance gene, an α/β tubulin intergenic sequence, and the Ty1 tag (Bastin et al., 1996), together with flanking tbRAP1 5′ UTR and the N-terminal 500 bp of tbRAP1 were inserted into pBluescript SKII+ to generate the tbRAP1 endogenous Ty1 tagging construct. F2H and GFP tagging constructs were similarly generated except PUR was used. Various fragments of tbRAP1 ORF were PCR amplified from T. brucei TREU 927 genomic DNA and inserted into pACT2 (Clontech) or pBTM116 to make yeast 2-hybrid constructs. The luciferase gene plus flanking 3′ and 5′ UTR were excised from pNS11 (Siegel et al., 2005) and inserted into pLF12 (Figueiredo et al., 2008) to generate pPUR-LUC.
The pBTM116-tbTRF-FL plasmid (Li et al., 2005) was transformed into yeast L40 strain (Vojtek et al., 1993), in which the expression of lacZ and HIS3 are driven by minimal GAL1 and HIS3 promoters fused to multimerized LexA binding sites. L40 cells harboring LexABD-tbTRF alone gave a mild transcription of HIS3, which was suppressed by 2.5 mM 3-aminotriazole. A GAD cDNA library generated with T. brucei 427 cDNA (Hoek et al., 2002; J. Munoz-Jordan & G. A. M. Cross, unpublished data) was used for the 2-hybrid screen. 1.6 million primary transformants were screened for HIS3 and lacZ expression. GAD-fusion plasmids recovered from positive transformants were analyzed by restriction digestion. Independent clones were further tested to confirm that they do not give a positive result by themselves and that they interact with LexABD-tbTRF. Clones that passed all screens were sequenced and the insert identities were determined by searching the T. brucei genome database (Berriman et al., 2005). The interactions between different fragments of tbTRF and tbRAP1 were analyzed by liquid assays using ONPG as substrate.
The pPUR-LUC plasmid was transfected into VSG2 (also known as VSG221)-expressing SM cells (Wirtz et al., 1999) to generate Puromycin-resistant clones PUR-LUC1 (PL1). VSG9-expressing PVS3-2 cells were obtained by injecting PL1 cells into naïve Charles River Strain CD-1 mice followed by isolation of in situ VSG switchers (confirmed by Southern blotting). To prevent further in situ switches in culture, the Blasticidin-resistance gene was targeted immediately downstream of VSG9 ES promoter to give rise to PVS3-2/OD1-1 cells, and these cells were maintained with 5 μg/ml of Blasticidin S.
Ri-2 cells were obtained by transfecting SM/Ty1-tbRAP1 cells with p2T7-TABlue-tbRAP1. Ri-9 cells were obtained by transfecting PVS3-2/OD1-1 cells with p2T7-TABlue-tbRAP1 and tbRAP1 Ty1-tagging constructs. All independent tbRAP1 RNAi clones behaved similarly upon induction with 0.1 μg/ml doxycyclin.
Rabbit anti-VSG13, rabbit and chicken anti-VSG2 and the preparation of cross-reacting-determinant- (CRD) depleted sera were described in (Figueiredo et al., 2008). Rabbit anti-VSG9 was kindly provided by Piet Borst, and CRD-depleted antibody was similarly prepared. Chicken antibody 606 was raised against a His6-tbTRF expressed in bacteria. IgY proteins were first purified from egg yolks using EggStract kit (Promega) then affinity purified with His6-tbTRF coupled CNBr-activated beads (GE). Rabbit antibody 597 was raised against a GST-tbRAP1414–855 expressed from bacteria and purified with His6- tbRAP1414–855 coupled CNBr-activated beads.
200 million cells harboring an endogenous F2H-tagged tbRAP1 were lyzed with two rounds of nitrogen cavitation in lysis buffer (50 mM Tris•Cl pH 7.4; 60 mM KCl, 1.5 mM MgCl2, 0.4 M NaCl, 1 mM EDTA, 0.1% Triton X-100, 0.01% SDS, 1 mM DTT). Protein extract was dialyzed against dialysis buffer (20 mM HEPES•KOH, pH 7.9, 0.2 mM EDTA, 0.2 mM EGTA, 100 mM KCl) and pre-cleared with incubation with protein G beads (Sigma) pre-washed with 1 × PBS/1% BSA. The pre-cleared lysate was used for IP using either 2 μg of 12CA5 against HA (Rockefeller-Memorial Sloan Kettering Cancer Center Monoclonal AB facility) or 2 μl of 1260 against tbTRF (Li et al., 2005). IP product was washed sequentially with buffer A (0.1% SDS; 1% Triton X-100; 2 mM EDTA pH 8.0; 20 mM Tris•HCl pH 8.0; 150 mM NaCl), B (same as buffer A but with 500 mM NaCl), C (0.25 M LiCl; 1% NP-40; 1% Na-Deoxycholate; 1 mM EDTA pH 8.0; 10 mM Tris•HCl pH 8.0) and TE. All buffers are supplemented with 1 mM PMSF, protease inhibitor cocktail for mammalian cells (Sigma), 4 μg/ml pepstatin A, and 0.5 mg/ml TLCK immediately before use.
Total RNA was extracted from cells using RNAstat (TEL-TEST, Inc) and purified with Qiagen RNeasy kit and treated with DNase (Qiagen). Reverse transcription was carried out using M-MLV (Promega) according to the manufacturer’s protocol.
Quantitative real-time RT-PCR was carried out using BioRad iTaq SYBR Green Supermix with ROX according to the manufacturer’s protocol. The amount of DNA was quantified by DNA Engine Opticon 2 (BioRad). Sequences for primers used in RT-PCR are listed in Supplemental table 1. Only primers giving specific PCR products were used. The normalized increase in mRNA level was calculated according to the following formula: (mRNAV, n/mRNAV, 0)/(mRNAT, n/mRNAT, o), where mRNAV, n and mRNAV, 0 represent the mRNA level for a particular VSG at n hr or 0 hr after tbRAP1 RNAi induction, and mRNAT, n and mRNAT, o represent the mRNA levels for β-tubulin at corresponding time points.
IF was carried out as published (Lowell and Cross, 2004). ChIP was carried out as previously described (Li et al., 2005), with or without the formaldehyde treatment. The precipitated and input DNA samples were loaded onto a nylon membrane, hybridized with TTAGGG-repeat, tubulin, or Tb11.0330 probes, and quantified with a phosphorimager. The precipitated amount was calculated as a percentage of input material. Under both conditions, DNA was sonicated to an average size of 300–400 bp. FACS was carried out as previously (Li et al., 2005), except the results were analyzed with FlowJo (TreeStar). The cell cycle profile was analyzed using the Watson Pragmatic template. Unpaired T tests were performed in Prism (GraphPad).
We are grateful to Piet Borst for the VSG9 antibody and Keith Gull for the RNA Pol I and tubulin antibodies. We greatly appreciate Titia de Lange and George A. M. Cross for insightful discussions and we thank Margaret J. Irwin, Audrey Lynn, and members of the Li and Cross labs for comments on the manuscript, critical discussions, and technical support. DeltaVision Microscopy was carried out at the RU Bio-Imaging Resource Center with guidance from Alison North and in BGES Dept. at CSU. FACS analysis was done at the Flow Cytometry Resource Center at RU and the Flow Cytometry Core at CCF. This work was supported by National Institute of Health grants to G. A. M. Cross (AI050614 and AI21729) and to B. Li (AI066095).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.