Trypanosoma brucei Orc1 Is Essential for Nuclear DNA Replication and Affects Both VSG Silencing and Switching

Summary

Binding of the Origin Recognition Complex (ORC) to replication origins is essential for initiation of DNA replication, but ORC has non-essential functions outside of DNA replication, including in heterochromatic gene silencing and telomere maintenance. Trypanosoma brucei, a protozoan parasite that causes human African trypanosomiasis, uses antigenic variation as a major virulence mechanism to evade the host’s immune attack by expressing its major surface antigen, the Variant Surface Glycoprotein (VSG), in a monoallelic manner. An Orc1/Cdc6 homolog has been identified in T. brucei, but its role in DNA replication has not been directly confirmed and its potential involvement in VSG repression or switching has not been thoroughly investigated. In this study, we show that TbOrc1 is essential for nuclear DNA replication in mammalian-infectious bloodstream and tsetse procyclic forms (BF and PF). Depletion of TbOrc1 resulted in derepression of telomere-linked silent VSGs in both BF and PF, and increased VSG switching particularly through the in-situ transcriptional switching mechanism. TbOrc1 associates with telomere repeats but appears to do so independently of two known T. brucei telomere proteins, TbRAP1 and TbTRF. We conclude that TbOrc1 has conserved functions in DNA replication and is also required to control telomere-linked VSG expression and VSG switching.

Introduction

Replicating nuclear DNA once per cell cycle is critical for genome stability. In eukaryotic cells, DNA replication is initiated with the assembly of a pre-replication complex (pre-RC) at specific chromosomal locations termed replication origins, establishing replication competence for the next S phase (Mechali, 2010). Origin Recognition Complex (ORC), a heterohexamer of Orc1-Orc6, binds to replication origins and recruits licensing factors to establish the pre-RC (Bell, 2002). Once the pre-RC – including ORC, Cdc6, Cdt1, and MCM proteins – is assembled, the replication origin is licensed to initiate DNA replication (DePamphilis et al., 2006).

In addition to DNA replication, subunits of the ORC complex have been implicated in various other cellular functions, including cell-cycle checkpoints (Gibson et al., 2006; Ide et al., 2007), heterochromatin organization (Prasanth et al., 2010; Chakraborty et al., 2011), and centrosome and kinetochore function (Prasanth et al., 2004; Prasanth et al., 2002). In budding yeast, recruitment of Sir1 by Orc1 N-terminal Bromo-Adjacent Homology (BAH) domain appears to be critical for mating-type silencing (Zhang et al., 2002; Triolo and Sternglanz, 1996; Gardner et al., 1999; Rusche et al., 2002; Hsu et al., 2005; Hou et al., 2005). Yeast Orc1 also binds to replication origins located at subtelomeric regions (Pryde and Louis, 1997), and N-terminal acetylation of Orc1 is required for telomeric silencing (Geissenhoner et al., 2004). Mammalian ORC subunits interact with the telomere-binding protein TRF2, which is important for the assembly of pre-RC at the telomere (Deng et al., 2009; Tatsumi et al., 2008). Interestingly, the roles of Orc1 in gene silencing do not always overlap with its function in DNA replication initiation (Bell et al., 1995).

The hexameric ORC complex is largely conserved among eukaryotes, including yeast (Bell and Stillman, 1992), plants (Diaz-Trivino et al., 2005), insects (Gossen et al., 1995), and mammals (Dhar and Dutta, 2000). In Trypanosoma brucei, a protozoan parasite that causes sleeping sickness in humans, we have just begun to understand DNA replication initiation. An Orc1/Cdc6 homolog, an Orc4 homolog, an Orc1-like protein, Orc1b, and a couple of novel ORC complex subunits have recently been identified (Godoy et al., 2009; Dang and Li, 2011; Tiengwe et al., 2012b). TbOrc1/Cdc6, which will be referred to as TbOrc1 henceforth, is essential for normal cell proliferation. A recent study determined TbOrc1 binding sites along all megabase chromosomes, and 20% of those sites appeared to be replicated in early S phase and are considered to be DNA replication origins (Tiengwe et al., 2012a). However, there is no direct evidence that TbOrc1 is involved in de novo nuclear DNA replication.

While inside a mammalian host, bloodstream form (BF) T. brucei cells regularly switch their major surface antigen, the Variant Surface Glycoprotein (VSG), to evade the host’s immune response (Barry and McCulloch, 2001), a phenomenon known as antigenic variation. Although T. brucei has more than 1,000 VSG genes in its genome, only one VSG is expressed at any time. T. brucei is transmitted among different mammalian hosts via its insect vector, the tsetse (Glossina spp.). When inside the mid-gut of tsetse, T. brucei differentiates into the procyclic form (PF) in which all VSGs are silent while cells express a small family of procyclins abundantly on their surface.

In BF trypanosomes, VSGs are expressed exclusively from ‘Bloodstream-form VSG Expression Sites’ (BESs), which are large polycistronic units transcribed by RNA Polymerase I (Gunzl et al., 2003) and are located at subtelomeres with the VSG gene positioned immediately adjacent to the telomere repeats and the promoter 50–60 kb upstream (de Lange and Borst, 1982; Hertz-Fowler et al., 2008; Berriman et al., 2005). Tight regulation of VSG silencing and switching ensures the effectiveness of antigenic variation and maximizes its efficiency. Recent studies have shown that BES and VSG silencing requires several factors that are involved in heterochromatin organization, including chromatin remodeling (Figueiredo et al., 2008; Hughes et al., 2007; Stanne and Rudenko, 2010; Denninger et al., 2010; Figueiredo and Cross, 2010; Rudenko, 2010), nucleosome packaging (Stanne and Rudenko, 2010; Figueiredo and Cross, 2010), telomere structure (Yang et al., 2009) and DNA replication (Tiengwe et al., 2012a). Interestingly, inhibition of DNA replication led to derepression of genes adjacent to silent BES promoters without affecting downstream VSG expression (Sheader et al., 2004). Similarly, a few chromatin-remodeling factors including ISWI, Spt16, DAC3 and a transcription factor NLP are also necessary for proper BES promoter silencing but not VSG silencing (Hughes et al., 2007; Denninger et al., 2010; Wang et al., 2010; Narayanan et al., 2010).

VSG switching occurs by two main mechanisms, principally involving recombination pathways that largely depend on factors required for protection of genome stability (McCulloch and Barry, 1999; Hartley and McCulloch, 2008; Kim and Cross, 2010; Kim and Cross, 2011). One mechanism that does not involve DNA rearrangement is a so-called in-situ switch, where the originally active BES is turned off and a silent BES is transcriptionally activated. VSG silencing and in-situ VSG switching may share common players because BES promoters are regulated in both events. Interestingly, depletion of a cohesin subunit, TbScc1, increased in-situ VSG switching (Landeira et al., 2009). It was hypothesized that the premature dissociation of sister chromatids in TbScc1-depleted cells led to a less stable inheritance of the expression status for the active BES. A recent study identified NUP-1 as a lamin-like nucleoskeletal protein and depletion of NUP-1 led to a slightly higher level of subtelomeric VSG mRNA, probably due to elevated VSG switching frequency, according to IF analysis (DuBois et al., 2012).

The ORC complex is required for genome stability and gene silencing in other organisms. In T. brucei, a recent study showed that TbOrc1 is located at subtelomere regions and TbOrc1 depletion led to derepression of metacyclic VSGs in procyclic cells (Tiengwe et al., 2012a). In this study, we examined the roles of TbOrc1 in nuclear DNA replication, subtelomeric bloodstream form VSG silencing and VSG switching, to explore functional links between antigenic variation and DNA replication in T. brucei. We found that TbOrc1 is required for nuclear DNA replication and complete silencing of BES-linked VSGs, and it associates with telomere repeats. Depletion of TbOrc1 also affected VSG switching events that involve BES promoter regions. Our data indicate that TbOrc1 is required for both VSG silencing and VSG switching.

Results

TbOrc1 is required for nuclear DNA replication

Several components of ORC complex, including an Orc1/Cdc6 homolog, have been identified in T. brucei and are expected to play essential roles in DNA replication. However, no direct evidence has been reported to demonstrate that TbOrc1 is indeed necessary for DNA replication, although previous studies have shown that TbOrc1 is required for cell cycle progression in PF and BF parasites (Godoy et al., 2009; Tiengwe et al., 2012b) and that some Orc1 binding sites overlap with DNA sequences that are replicated early in S phase (Tiengwe et al., 2012a). In addition, Orc1 homologs are known to play important roles in cellular processes other than DNA replication. Therefore, we felt compelled to firmly establish a role for TbOrc1 in DNA replication before setting out to identify its other possible functions.

To study TbOrc1’s function, we established tightly regulated inducible TbORC1 RNAi cell lines that produce a stem-loop dsRNA. One endogenous TbORC1 allele was also N-terminally tagged with FLAG-HA-HA (F2H) in these cells to monitor the F2H-TbOrc1 protein level. In the BF stage, induction of TbORC1 RNAi led to depletion of the F2H-TbOrc1 protein in two independent clones and caused a growth defect (Fig. 1, A & B). When the same TbORC1 RNAi construct was introduced into TbORC1 single-allele knockout (SKO) cells, RNAi induction caused a more severe defect, arresting cell growth within one day (Fig. 1B, orange line), which is similar to previous observations (Tiengwe et al., 2012b). The uninduced cells grew normally, indicating that the tagged TbOrc1 is functional. For all subsequent experiments, we used TbORC1 RNAi in wild-type background instead of in SKO, because the latter has extremely severe cell growth and cell morphological defects that rendered the BrdU labeling experiment impractical.

Figure 1.

TbOrc1 is essential for BF cell growth. (A) Depletion of TbOrc1 in RNAi cells. Whole cell lysates from two independent cell lines with FLAG-HA-HA (F2H)-tagged TbOrc1, B4 and B5, were analyzed by western blotting using anti-HA antibody at the indicated time points. Tubulin is shown as a loading control. (B) Growth curves for WT, TbORC1 RNAi cell lines in wild-type or in TbORC1^−/+ (SKO) heterozygous background. TbORC1 RNAi was induced using 0.1 μg/ml doxycyclin (DOX). The average population doublings (PDs) versus days of induction and standard deviations were calculated from three independent experiments. (C) Depletion of TbOrc1 led to a cell cycle profile change. Top, gating schematic of uninduced TbORC1 RNAi cells in flow cytometry analysis. PI, propidium iodide staining intensity. Bottom, quantification of populations of cells at different cell cycle stages before (day 0) and after (days 1, 2, and 3) depletion of TbOrc1. Unpaired t-tests were done to compare Day 1, 2, and 3 values with the Day 0 value. One asterisk, 0.009<P≤0.05; two asterisks, 0.0001<P≤ 0.009; three asterisks, P≤0.0001.

To further investigate the cell cycle progression in TbORC1 RNAi cells, we carried out flow cytometry analysis of propidium iodide-stained cells (Fig. 1C and Fig. S1). Uninduced cells exhibited a normal distribution of cells with 2C (G1, ~56%), S (~8%), and 4C (G2/M, ~30%) DNA contents. However, in TbOrc1-depleted cells (at day 3), we observed a significant decrease in the number of 2C cells (~34%) and a significant increase in the number of 4C (~48%) and S cells (~12%). This abnormal cell cycle profile suggests that TbOrc1’s function is necessary for normal cell cycle progression and is consistent with the hypothesis that TbOrc1 plays an important role in DNA replication.

The knockdown in PF was very efficient as growth arrest was observed faster (by day 3–4, Fig. S2, A & B) than previously studied cell lines (Godoy et al., 2009; Tiengwe et al., 2012b). We also verified that silencing of TbOrc1 in PF cells impairs normal cell cycle progression by examination of the nucleus-kinetoplast (N-K) content in induced and uninduced cells by DAPI staining and detection of basal bodies (Fig. S2, C & D). We observed a dramatic increase in 1N2K cells and zoids (cells lacking a nucleus) when TbOrc1 was depleted, suggesting that these cells efficiently completed kDNA replication but were unable to complete nuclear DNA replication on time. These cell cycle phenotypes of TbOrc1-depleted cells are consistent with previous reports (Godoy et al., 2009; Tiengwe et al., 2012b).

To directly examine roles of TbOrc1 in DNA replication, we labeled bulk DNA with DAPI and newly synthesized DNA with 2′-bromo-5′-deoxyuridine (BrdU) in both PF and BF cells (Fig. 2). PF parasites were labeled with BrdU for the last 3 hours of total TbORC1 RNAi induction time. In uninduced control PF cells (Fig. 2, A & B), the 1N1K cells were either BrdU-negative or BrdU-positive, as these should be in G1 or S phase of the cell cycle, respectively, while nearly all 1N2K nuclei (G2/M phase) were labeled with BrdU as expected. In TbOrc1-depleted cells, however, about 60 – 70% of 1N2K cells were BrdU-negative (days 4 & 6 in Fig. 2B), and the number of BrdU-positive 1N1K cells decreased from 24.5% to 7.5% after 6 days of TbORC1 RNAi induction (Fig. 2B), indicating that nuclear DNA replication was impaired in the absence of TbOrc1. Consistent with the previous reports (Godoy et al., 2009; Tiengwe et al., 2012b), there was a dramatic increase in 1N2K cells during TbOrc1 depletion with a corresponding increase in zoids (Fig. S2 & S3B). Therefore, our interpretation is that DNA replication defects resulting from TbOrc1 depletion leads to prolonged S phase with low level of BrdU incorporation, and this cell cycle disruption leads to an accumulation of zoid cells (Fig. S2D). As expected, kDNA still incorporated BrdU during TbOrc1 silencing (Fig. 2A, Day 4, asterisk). More images of BrdU stained PF cells are shown in Fig. S3A and quantification of cell cycle profiling using the N-K content is shown in Fig. S3B (data corresponding to the BrdU quantification of PF cells shown in Fig. 2A & 2B).

Figure 2.

TbOrc1 is required for PF and BF nuclear DNA replication. DNA replication was analyzed by comparing BrdU incorporation in TbORC1 RNAi uninduced and induced cells. Immunostaining of BrdU-incorporated cells with anti-BrdU antibody and bulk DNA that was counterstained with DAPI (A, C). DAPI fluorescence indicates cells at different stages of the cell cycle. Representative images are shown. Asterisk, kDNA only BrdU-positive; Arrow, nucleus and kDNA BrdU-positive; Arrowhead, nucleus only BrdU-positive. Percentage of BrdU-positive and BrdU-negative cells in control and RNAi induced samples (B, D). Data are presented as the mean percent of 200 cells (for PF experiments) and 150 cells (for BF experiments) counted per time point from two independent experiments. Error bars represent the variation range. Only 1N1K and 1N2K cells are included in the analysis of replicating cells. The scale bar in (A) and (C) is 10 μm.

Similarly, BF parasites were labeled with BrdU for the last 24 hours of total TbORC1 RNAi induction time. During this labeling period, we expect almost all replicating cells to be BrdU-positive while cells that had arrested before the labeling period would remain BrdU-negative. As expected, 90% of the uninduced BF cells incorporated BrdU (Fig. 2C), and these showed the normal distribution of cells with 1N1K (80%), 1N2K (9%), and 2N2K (5.5%) configurations (Fig. S3D). In this uninduced situation, a small percent of 1N1K cells are BrdU-negative (~10%), while ~90% of 1N1K cells apparently have entered S phase and have normal DNA replication (so that they are BrdU-positive) (Fig. 2D). However, after TbORC1 RNAi induction, 28% (Day 1) or 73% (Day 3) of 1N1K cells are BrdU-negative, indicating that a great percentage of the cells had DNA replication defects during the period of TbOrc1 depletion (Fig. 2C & D). As expected and similar to what we observed in PF, kDNA still incorporated BrdU during the labeling period, while the nucleus was not (asterisk in Fig. 2C, day 3 BrdU image asterisk). Additionally, there was an increase in aberrant cells including zoids and cells with multiple nuclei and kinetoplasts both at day 1 and 3, however increase in zoids was not as prominent as in the TbOrc1-depleted PF cells (Fig. S3D). Collectively, we have directly demonstrated for the first time that TbOrc1 is indeed required for nuclear DNA replication in T. brucei cells at both BF and PF stages.

TbOrc1 is required for complete silencing of BES-linked subtelomeric VSGs

Orc1 participates in the transcriptional silencing at mating-type loci and at native XI-L telomere in budding yeast (Hsu et al., 2005; Hou et al., 2005; Pryde and Louis, 1999), Orc1 is linked to heterochromatin organization in mammals (Chakraborty et al., 2011), and TbOrc1 is necessary for silencing metacyclic VSGs in PF trypanosomes (Tiengwe et al., 2012a). We therefore examined whether TbOrc1 plays any roles in BES-linked VSG silencing in both BF and PF. The mRNA levels of several BES-linked silent VSGs were measured by quantitative real-time PCR (Q-RT-PCR) before and after the TbORC1 RNAi induction, first, in BF cells. We observed a moderate derepression of the BES-linked subtelomeric VSGs (Fig. 3A). The derepression levels for different VSGs varied, ranging from 3-fold (VSG9) to as high as 6-fold (VSG11) on day 2 and from 4-fold (VSG16) to 12-fold (VSG6) on day 4 of TbORC1 RNAi induction. We did not observe any elevated expression of the originally active VSG2 gene nor several control genes located in the chromosome-internal loci: rRNA, TbTERT (telomerase protein component gene), TbPGI (Tb927.1.3830, which encodes glycosomal glucose-6-phosphate isomerase), and TbRPS15 (Tb927.7.2370, a putative ribosomal protein gene). In addition, we were able to detect expression of two previously silent VSG proteins, VSG9 and VSG13, after 4 days of TbORC1 RNAi induction in western analysis (Fig. 3B).

Figure 3.

TbOrc1 is required for complete silencing of subtelomeric BES-linked VSGs. (A) Depletion of TbOrc1 in BF cells leads to derepression of BES-linked silent VSG genes. Steady state mRNA levels of several silent VSGs, the active VSG2, and control genes were analyzed by Q-RT-PCR, using tubulin gene as an internal loading control. The fold changes in mRNA level after RNAi induction are plotted. Day 0 values were set to 1. rDNA, TbTERT, TbPGI, and TbRPS15 genes were analyzed as controls. Fold changes were averaged from at least three independent experiments. Standard deviations are shown as error bars. (B) Expression levels of silent VSG proteins and telomere proteins in TbOrc1-depleted BF cells. (C) BES-linked silent VSGs are derepressed in TbOrc1-depleted PF cells. The analysis was performed as described in (A). (D) Two silent VSG proteins can be expressed simultaneously on the surface of a single TbOrc1-depleted BF cell. VSG2-expressing BF cells were examined by immunofluorescence at indicated time points after TbORC1 RNAi induction. Cells co-expressing the silent VSG3 and VSG13 were visible four days post TbORC1 RNAi induction. A single layer of the z-stack images was shown for each time point to better illustrate the expression of derepressed VSGs and their localization.

To test whether derepression of multiple VSGs can occur simultaneously within the same cells, we carried out immunofluorescence (IF) analysis using antibodies specific to two silent VSGs, VSG3 and VSG13. As shown in Fig. 3D (day 0), VSG3 or VSG13 expression was undetectable before induction of TbORC1 RNAi. In contrast, at day 4, we observed cells that were co-stained with both VSG3 and VSG13 antibodies, indicating that multiple silent VSG loci can be simultaneously derepressed in a single BF trypanosome cell in the absence of TbOrc1.

Expression of the active VSG2 was not affected by TbOrc1 depletion (Fig. S4). VSG2 expression was easily observed at day 0 when stained with a rabbit antibody in IF (Fig. S4A). To detect cells expressing the active and silent VSGs together, we co-stained the cells with rabbit antibodies against either VSG3 (silent) or VSG13 (silent), and a chicken anti-VSG2 (active) antibody at a low concentration to avoid a strong VSG2 staining that could mask the staining of silent VSG3 or VSG13 (Fig. S4, B & C). At day 4 of RNAi induction, we observed cells expressing the active and a silent VSG (VSG13 in Fig. S4B and VSG3 in Fig. S4C) simultaneously, and we did not detect any change in VSG2 staining at day 0 or day 4. Occasionally, we observed VSG13-expressing cells that no longer expressed the originally active VSG2 (data not shown), indicating that these were switched from VSG2 to VSG13. In the uninduced control cells, silent VSG proteins or switched variants were not observed. From these data, we conclude that TbOrc1 is required for VSG silencing and may be also required for VSG switching regulation.

PF cells repress all VSGs and coat their surface with procyclins. Therefore, we examined whether TbOrc1 is similarly required for VSG silencing in PF cells by performing Q-RT-PCR. TbOrc1 depletion caused milder derepression of BES-linked VSG genes, compared to BF cells, and the level of VSG derepression also varied in PF, ranging from 2- to 6-fold (Fig. 3C). Expression levels of VSG2 and VSG11 were not affected in PF cells. No changes in mRNA levels were observed for control genes.

TbTRF is a mammalian TRF2 homolog that binds the duplex telomere DNA (Li et al., 2005) and directly interacts with TbRAP1, which is required for complete repression of all BES-linked silent VSGs (Yang et al., 2009). Human Orc1 interacts with TRF2 (Atanasiu et al., 2006; Tatsumi et al., 2008; Deng et al., 2009); therefore, we asked whether derepression of silent VSGs in TbOrc1-deficent cells is due to the instability of telomere proteins. However, as shown in Fig. 3B, protein levels of TbRAP1 and TbTRF were unchanged in TbOrc1-depleted BF cells. Next, we examined whether the depletion of TbOrc1 could inhibit the binding of TbTRF and TbRAP1 to telomeres in BF. We found that TbTRF or TbRAP1 associated with telomere regardless of the TbOrc1 protein level (Fig. S5). Therefore, our data suggest that TbOrc1 contributes to VSG silencing independently of these telomere factors. However, the VSG derepression phenotype needs to be analyzed in cells depleted of both TbOrc1 and TbRAP1 to determine whether the two proteins can act in the same genetic pathway.

TbOrc1 controls specific VSG switching mechanisms

VSG13-expressing cells that no longer expressed the originally active VSG2 (switchers) were occasionally observed when TbOrc1 was depleted in BF cells, but not in RNAi uninduced samples, suggesting that TbOrc1 may be required for VSG switching. To further investigate potential roles of TbOrc1 in VSG switching, the TbORC1 RNAi construct was introduced in a switching reporter cell line where a blasticidin-resistance gene (BSD) had been inserted at the promoter driving the VSG2 BES and a puromycin-resistance gene fused to Herpes simplex virus thymidine kinase (PUR-TK) downstream of the 70-bp repeats in the same BES (Kim and Cross, 2010) (Fig. 4A). VSG switchers are expected to lose the PUR-TK gene or repress its expression and become resistant to ganciclovir (GCV), a nucleoside analogue that kills TK-expressing cells. To monitor the depletion of TbOrc1, one TbORC1 allele was C-terminally tagged with three repeats of HA (3xHA). The 3xHA tagged version (TbOrc1-HA) is functional, as cells expressing one allele of TbORC1-HA did not show any growth defect.

Figure 4.

TbOrc1 controls VSG switching mechanisms. (A) VSG switching mechanisms. Diagram exhibits four major switching events; in-situ (transcriptional inactivation of the active and activation of a silent BES promoter), ES GC + ES loss (recombination near BES promoters, or active BES loss associated with in-situ switching), VSG GC (gene conversion near VSGs), and crossover (XO) (exchanges of chromosome ends containing the active and a silent VSG). (B) Expression levels of TbOrc1-HA, histone H3, and tubulin in the TbORC1 RNAi cells. (C) VSG switching frequency increased about 3-fold in the TbOrc1-depleted cells. Switchers were enriched and selected after two days of induction followed by one day of recovery (see text for details). Three independent cultures each were examined. Standard deviations are shown as error bars. Unpaired T tests were done and means of switching frequencies were significantly different (P=0.0124). (D) TbOrc1 suppresses VSG switching at the BES promoter. Fifty-eight and 135 switchers from three uninduced and three induced cultures, respectively, were analyzed for their switching mechanisms (see text for details). Percentage of individual mechanisms is indicated in the sector. Approximately 1% of clones appeared to have undergone in-situ and homologous recombination because VSG2 is lost but BSD does not appear to be expressed from the active ES even though the BSD gene is still present.

Cells were grown in the presence of blasticidin and puromycin to maintain a homogeneous population of VSG2 expressors. TbORC1 RNAi was induced for two days and recovered for one day without induction and in the absence of selection, to allow switching to occur. Uninduced samples were grown in the absence of selection for two days. As TbOrc1 depletion causes cell growth arrest, RNAi uninduced and induced samples reached to the same density after two and three days of total incubation time, respectively. Three independent cultures were analyzed for VSG switching phenotypes. As shown in Fig. 4B, TbOrc1-3xHA protein was not detected after TbORC1 RNAi induction. To measure VSG switching frequency, unswitched VSG2-expressing cells were depleted by magnetic-activated cell sorting (MACS) (Boothroyd et al., 2009) using anti-VSG2 antibody, and the flow-through fraction enriched with switchers was distributed into 96-well plates in the presence of GCV to select for the switched variants. Switching frequency was determined as the ratio of GCV-resistant cells to the total number of viable cells applied to the MACS column. As shown in Fig. 4C, overall VSG switching frequency increased about 3-fold in TbOrc1-depleted cells, suggesting that TbOrc1 has a role in VSG switching.

To better understand the function of TbOrc1 in VSG switching, we examined 193 cloned switchers to determine their switching mechanisms as described previously (Kim and Cross, 2010). Genes that are located closer to silent BES promoters are not completely repressed (Vanhamme et al., 2000). Therefore, in-situ switchers can grow at a lower concentration of blasticidin but cannot at a higher one (100 μg/ml) (Kim and Cross, 2010), at which we test all switchers. In-situ switching occurs by inactivating the active BES and activating one of the silent BESs. Therefore, in-situ switchers will preserve BSD and VSG2 genes but repress their expression, and they will be sensitive to a high concentration of blasticidin. ES gene conversion (ES GC) occurs through recombination near the BES promoter, which allows duplication and translocation of an entire silent BES to the VSG2 subtelomere. The ES loss can result in switching either through ES GC or through in-situ switching (‘ES GC+ES loss’). These switchers will lose BSD and VSG2 genes and will be sensitive to blasticidin. Recombination near VSGs can result in either duplicative gene conversion of a new VSG to the VSG2 BES (VSG GC) or in VSG crossover (VSG XO) switching. In both cases, the BSD gene will be at the active promoter and the switchers resistant to blasticidin, but these two recombinants can be distinguished by VSG2 absence (VSG GC) or presence (crossover).

We analyzed 58 and 135 cloned switchers isolated from uninduced and TbORC1 RNAi induced cultures as described above (Fig. 4D). Results from PCR genotyping of BSD and VSG2, and blasticidin sensitivity of all switchers are summarized in Tables S1–S6. Examples of BSD and VSG2 PCR analyses of switchers are shown in Fig. S6. Newly activated VSGs were cloned from several of the switched variants and representatives of each switching mechanism were further analyzed by pulse-field gel analysis (Fig. S7). Consistent with previous in-vitro VSG switching studies (Kim and Cross, 2010; Boothroyd et al., 2009), when TbOrc1 was present, VSG switching occurred mainly by recombination (92%) including VSG GC, VSG crossover, and ‘ES GC+ES loss’, but in-situ switching was infrequent (5%). In contrast, in-situ VSG switching was the preferred event (38%) in TbOrc1-depleted cells, exhibiting about 20-fold increase in frequency. This suggests that TbOrc1 is not only involved in repression of silent VSGs (Fig. 3) but also in the transition of their transcriptional status. Interestingly, TbOrc1 depletion caused an ~4-fold higher frequency of ‘ES GC+ES loss’. These data suggest that TbOrc1 suppresses transcriptional switching and recombinatorial VSG switching near BES promoter.

In TbORC1 RNAi uninduced cells, VSG crossover (XO) appeared to be higher than normally observed in WT cells (Kim and Cross, 2010). 6 and 4 XO switcher clones were obtained from cultures #1 and #3, and none from #2. However, all 6 from culture #1 expressed VSG18 and all 4 from culture #3 expressed VSG17 (Table S1 and S3), suggesting that these might not be independent switches, but rather derived from an early switch event in the population.

TbOrc1 associates with the telomere DNA

Mammalian ORC proteins interact with telomere proteins and TERRA – the RNA transcripts of the telomere DNA (Atanasiu et al., 2006; Deng et al., 2007; Deng et al., 2009; Tatsumi et al., 2008). TbOrc1 is enriched at subtelomere loci (Tiengwe et al., 2012a) but it is not known if it binds to telomeres. Depletion of TbOrc1 increased silent VSG expression and VSG switching. We therefore reasoned that TbOrc1 might affect antigenic variation by associating with the telomere repeats. We performed Chromatin IP (ChIP) in BF cells expressing F2H-TbOrc1 (Fig. 5, A & B) and in PF cells expressing a sole allele of TbOrc1-PTP (the normal growth of this PF cell line indicates that TbOrc1-PTP is functional; Fig. 5, C & D). The PTP tag contains two protein A and one protein C peptide separated by a TEV cleavage site (Schimanski et al., 2005). TbOrc1 proteins were immunoprecipitated using anti-HA antibodies (F2H-TbOrc1) or IgG coupled sepharose beads (TbOrc1-PTP). Precipitated DNA was analyzed by Southern blot using telomere probes and quantified as a percent of the input. We found that telomere DNA was significantly enriched in HA antibody or IgG-beads precipitated fractions in cells expressing F2H-TbOrc1 (BF) or TbOrc1-PTP (PF), compared to WT cells expressing no tagged proteins (Fig. 5). TbTRF was used as a positive control, which showed higher level of telomere association. Therefore, TbOrc1 appears to associate with the telomere DNA.

Figure 5.

TbOrc1 is located at telomeres. Quantification of TbOrc1 ChIP results in BF (A) or PF (C) cells. Probes used in different hybridizations were indicated at the top. Average enrichments were calculated from 15 (BF) or 4 (PF) independent experiments, and standard variations are shown as error bars. P values of unpaired t-tests are shown for indicated pairs of data. Representative Southern analyses of the TbOrc1 ChIP in BF (B) or PF (D) cells are shown. DNA precipitated with HA antibody (HA), TbTRF antibody (TbTRF), or no antibody (No AB) from BF cells, and DNA precipitated with IgG beads or protein G beads from PF cells, and the input DNA sample (20%) were hybridized with either a TTAGGG-repeat probe, a 177-bp repeat probe (BF), or a 50 bp repeat probe as indicated.

Next, we examined whether TbOrc1 is tethered to the telomere through interactions with telomere proteins TbTRF (Li et al., 2005) or TbRAP1 (Yang et al., 2009), by performing yeast two-hybrid analyses. LexA-binding domain (LexABD)-TbOrc1 and Gal4-activation domain (GAD)-TbTRF were co-expressed in a yeast strain with the β-galactosidase reporter (Fig. S8C). The β-galactosidase activity was measured using ortho-nitrophenyl-β-galactoside (ONPG) as the substrate. No direct interaction was observed between TbOrc1 and TbTRF (Fig. S8A). Similarly, LexABD-TbRAP1 and GAD-TbOrc1 were co-expressed (Fig. S8C), but no direct interactions were observed (Fig. S8A). It is possible that GAD-TbOrc1 and LexABD-TbOrc1 might not be fully functional. However, we did not detect an interaction between TbOrc1 and TbTRF or TbOrc1 and TbRAP1 by co-immunoprecipitation experiments (Fig. S8B). Therefore, TbOrc1 binds to telomere repeats but likely not through interactions with TbTRF or TbRAP1.

Discussion

TbOrc1 in antigenic variation

Here we showed that TbOrc1 is not only involved in DNA replication but also in antigenic variation, both VSG silencing and VSG switching. Similarly, in other organisms, the ORC complex has a major role in DNA replication and is also required for non-essential cellular processes (Sasaki and Gilbert, 2007), including heterochromatin organization and gene silencing (Chakraborty et al., 2011). As we have used an RNA knockdown strategy to study TbOrc1 function in T. brucei, it is difficult to clearly state whether the DNA replication function of TbOrc1 overlaps with its functions in antigenic variation.

DNA damage and replication inhibition led to derepression of silent BES promoters in a dosage dependent manner, with a minor effect on expression of silent VSGs (Sheader et al., 2004). On the other hand, TbOrc1 depletion led to significant increases of all silent VSGs examined in BF cells, suggesting that TbOrc1 may have separate functions in VSG silencing and DNA synthesis and/or DNA repair.

When VSG switching phenotypes were examined in TbOrc1 depleted BF cells, the majority of switching events appeared to occur via either BES instability (ES GC+ES loss associated events) or in-situ switching that transcriptionally turns off the active BES promoter and turns on one of the silent BES promoters, suggesting that TbOrc1 may have roles in maintaining the integrity of promoter regions of BESs. If loss of BES silencing is linked with an elevated rate of in-situ switching, aphidicolin treated cells are expected to transcriptionally switch at a higher rate. However, treating cells with sublethal dose of aphidicolin (1 ng/ml) or hydroxyurea (an inhibitor of dNTP synthesis) did not significantly change the overall VSG switching rate (Kim and Cross, 2010). Silent BES promoter derepression was more prominent when a higher amount of aphidicolin (30 μM or ~10,000 ng/ml) was used (Sheader et al., 2004). At a higher concentration, the effects of aphidicolin can include chromosome aberrations rather than simply DNA synthesis inhibition-induced DNA damage (Glover et al., 1984), indicating that loss of BES silencing might be due to DNA damage. Therefore, it is difficult to directly compare outcomes of TbORC1 RNAi to those of DNA synthesis inhibitors, and extrapolate mechanisms of actions. Nonetheless, considering our limited understanding on the molecular mechanisms of antigenic variation and DNA replication in T. brucei, it is important to note that our results and previous work revealed crucial clues regarding mechanistic links between genomic instability at VSG expression sites and antigenic variation in T. brucei.

Although previous studies showed that mammalian Orc2 and TRF2 suppress telomere-telomere recombination (Deng et al., 2007), TbOrc1 does not appear to play an important part in recombination-mediated VSG switching, as VSG GC switchers were only 20% in TbOrc1-depleted samples, but 51% in uninduced samples (given that overall switching rate increased 3 fold, this indicates only trivial changes in the GC rate).

The increased VSG switching in TbOrc1-depleted cells was largely due to the increased rate of in-situ switching (from 5% to 38%). Depletion of a cohesin subunit, TbScc1, also increased in-situ VSG switching rate (Landeira et al., 2009). A recent study showed that cohesins are transiently recruited to the replication origin and spread along DNA as the replication fork progresses in budding yeast (Tittel-Elmer et al., 2012). It is possible that T. brucei cohesin has a similar function, and depletion of TbOrc1 may cause uncoupling of DNA replication and cohesion establishment, leading to improper sister chromatid pairing and resulting in the elevated rate of in-situ VSG switching. However, TbScc1 depletion did not affect VSG silencing, whereas TbOrc1 depletion significantly increased silent VSG expression. Whether or not TbOrc1 plays multiple independent roles in T. brucei antigenic variation and how TbOrc1 and TbScc1 functionally interact to control in-situ VSG switching remain to be determined.

Recently, Tiengwe et al. (Tiengwe et al., 2012a) showed that depletion of TbOrc1 led to significant derepression of metacyclic VSG expression (5–13 fold) in PF cells but only a very mild or no derepression of BES-linked VSGs (1.5 fold in PF and 2.5 fold in BF cells). In contrast, we observed significant changes in VSG derepression levels. Although we used the same T. brucei strain, Lister 427, we observed nearly complete depletion of TbOrc1 in both BF and PF cells as demonstrated in western analyses, while TbOrc1 depletion was only 70% (PF) and 40% (BF) in the previous report. In addition, we observed elevated expression of various silent VSGs not only at the mRNA levels (by Q-RT-PCR), but also at the protein levels (by western and IF analyses). Furthermore, we also observed cells that express multiple silent VSGs simultaneously, indicating that VSGs derepression is a true phenotype of TbOrc1 depletion.

Depletion of TbOrc1 led to derepression of BES-linked silent VSGs both in BF and PF, but to a lesser degree in PF, indicating that TbOrc1 is required for VSG silencing in both life-cycle stages but detailed mechanisms may be different. We have also observed that TbOrc1 is localized at telomeres. Together, these data suggest that T. brucei Orc1 has a conserved function in telomeric silencing, one of the fundamental components that control T. brucei antigenic variation.

TbOrc1 at the telomeres

We found that TbOrc1 is enriched at telomeres, and recently it was shown that TbOrc1 appears to associate subtelomeric regions containing VSG expression sites at a subset of T. brucei megabase chromosomes (Tiengwe et al., 2012a). The association of TbOrc1 with VSG expression sites and telomeres may be important for TbOrc1’s function in T. brucei antigenic variation.

Mammalian ORC associates with telomeres via interaction with TRF2, the duplex telomere DNA-binding protein (Atanasiu et al., 2006; Deng et al., 2007; Tatsumi et al., 2008; Deng et al., 2009), and the mammalian Orc2 and TRF2 appear to suppress telomere-telomere recombination (Deng et al., 2007). Depletion of TbOrc1 did not show prominent effects on recombination events near VSGs or telomeres, such as VSG gene conversion (VSG GC) or VSG exchange (crossover), suggesting that TbOrc1 may not regulate VSG switching through influence of subtelomeric DNA recombination events. Rather, the association of TbOrc1 with expression sites and telomeres may contribute to the maintenance of the repressive status of silent VSG loci.

In Plasmodium falciparum, another protozoan parasite that undergoes antigenic variation, the Orc1 homolog is also located at the telomeres (Mancio-Silva et al., 2008) and is essential for silencing subtelomeric virulence gene expression (Deshmukh et al., 2012). It is possible that TbOrc1 has a similar function as its Plasmodium homolog.

TbOrc1 is involved in VSG switching regulation and the function of TbOrc1 appears to be focused on the BES promoter, as a significant portion of the VSG switching events in TbOrc1-depleted cells are in-situ switches or ES GC/ES loss associated events. One possibility is that removal of TbOrc1 allows more loosely packed chromatin at the subtelomeric BES loci and this leads to both derepression of silent VSGs and elevation of in-situ VSG switching frequency. It was previously shown that RAD51, a protein that is required for strand invasion step during homologous recombination in other organisms (Karpenshif and Bernstein, 2012), is required for VSG switching (McCulloch and Barry, 1999). Particularly, in the absence of TbRAD51, ‘ES GC+ES loss’ switching frequency decreased (Kim and Cross, 2011), indicating that recombination is required for this type of switching to occur. On the other hand, this type of switching was preferred in cells lacking TbOrc1, suggesting that in the absence of TbOrc1, recombination factors may have easier access to the BES promoter region, promoting recombination initiation that could result in ES GC or ES loss, therefore VSG switching. Analysis of chromatin compaction by formaldehyde-assisted isolation of regulatory elements (FAIRE) (Giresi and Lieb, 2009) did not detect any significant changes in chromatin structure of derepressed BESs in the absence of TbOrc1 (data not shown). The FAIRE assay, however, is probably not sensitive enough to detect subtle changes in chromatin structure in TbOrc1-defective cells.

Whether or not the TbOrc1 binding sites mapped in the BESs and telomeres are replication origins is not clear. Tiengwe et al. has found that TbOrc1 binding is enriched at multiple places along the chromosome, but only a subset (~20%) of these ~100 loci (within chromosomal core regions) are replicated in early S phase and hence can act as DNA replication origins (Tiengwe et al., 2012a). The analysis to identify replication origins was performed by comparing relative enrichment of DNA contents between FACS-sorted early S-phase and G2 cells, and telomeres appear to be replicated late in S phase, similar to that in other organisms. Therefore, the analysis suggested that subtelomeric TbOrc1-binding sites are not utilized as early-firing replication origins. Considering the high density of TbOrc1 binding sites present in subtelomeres, it is likely that most of them serve cellular functions other than replication origins. Although we were unable to detect any significant changes by FAIRE, employment of more sophisticated epigenetic assay may allow us to identify TbOrc1-mediated chromatin changes at VSG loci.

TbOrc1 in nuclear DNA replication

We demonstrated that TbOrc1 is required for nuclear DNA replication by performing cell cycle analysis of the N-K contents, flow cytometry, and also by detecting newly synthesized DNA incorporated with BrdU, which has not been successfully performed in previous studies of T. brucei Orc1. In TbOrc1-depleted PF cells, the number of zoids increased dramatically, an indicative of a nuclear DNA replication defect, as cytokinesis is uncoupled from nuclear DNA synthesis in PF T. brucei and cells can divide even when nuclear DNA replication is incomplete (Ploubidou et al., 1999). BrdU experiments provide more direct evidence demonstrating the requirement of TbOrc1 in DNA replication, where significantly smaller number of 1N1K (G1-S) and 1N2K (late S-G2) cells incorporated BrdU in the absence of TbOrc1. This clearly indicates that TbOrc1 is required for de novo DNA synthesis.

In contrast to the previous report, where multinucleated cells accumulated in TbOrc1-depleted BF cells (Tiengwe et al., 2012b), we did not observe any change in 6C or 8C population. One possible reason for this difference could be differences in the penetrance and sequences of RNAi constructs used in these studies. Nevertheless, both studies are consistent with the notion that TbOrc1 is important for DNA replication.

Defective DNA replication could delay S phase progression, resulting in delayed nuclear division. The accumulation of 1N2K and 1N1K cells in the TbOrc1-depleted cultures agrees with an S phase delay, as kDNA replication and division is independent of that of nuclear DNA. A similar phenotype was observed in yeast, where orc2-1 mutants arrested as large budded cells with only one DAPI positive staining focus (Foss et al., 1993). Loss of TbOrc1 function would prevent most replication origins from firing, hence arresting G1 or S phase cells in the S phase, which would lead to an elevated percentage of cells arrested at S phase or with a prolonged S phase. Similarly, when asynchronous orc2-1 mutant yeast cells were analyzed by FACS, a significant elevation in the number of S phase cells was observed (Bell et al., 1993). Although we only observed 4% more cells in S phase in TbOrc1 depleted cells, the 50% change (from 8% in WT to 12% in TbOrc1-depleted cells) was statistically significant. However, to rigorously confirm whether there is an S phase delay in TbOrc1-depleted cells, a more detailed cell cycle analysis using a synchronized trypanosome population is required, but this is not satisfactorily achievable in T. brucei at this time.

We are now beginning to understand T. brucei DNA replication with recent identification of replication origins and functional characterization of several replication proteins (Godoy et al., 2009; Dang and Li, 2011; Tiengwe et al., 2012b). Our study underscores TbOrc1’s role in T. brucei antigenic variation in addition to its role in DNA replication.

Experimental Procedures

Plasmids

To generate pFLAG-HA-HA (F2H)-ORC1 (the TbOrc1 N-terminal F2H-tagging construct) the 5′UTR (~500 bp) of TbORC1 (Tb11.02.5110) and full-length TbORC1 were separately PCR amplified from T. brucei 927 genomic DNA and inserted into pBlueScript to flank a cassette containing the puromycin-resistance gene (PUR), the α/β tubulin intergenic sequence, and the F2H tag. To create pORC1-PTP-NEO, TbORC1 C-terminal coding sequence (1,098 bp) was PCR amplified from T. brucei 927 genomic DNA and inserted into pC-PTP-NEO (Schimanski et al., 2005) at the ApaI and NotI sites. TbORC1 knockout construct was generated by inserting a hygromycin resistance gene (HYG) flanked by genomic DNA fragments upstream and downstream of TbORC1 ORF (each of ~ 500 bp) into the pBlueScript vector. To generate pStLORC1 (the TbORC1 RNAi construct), 613 bp of TbORC1 coding sequence (nucleotides 1,046 –1,660) was PCR amplified from T. brucei 927 genomic DNA to generate the two fragments for subsequent cloning steps in the construction of a stem-loop vector as previously described by (Wang et al., 2000). To generate pStLBFORC1, the same 613 bp region of TbORC1 was used for construction of a stem-loop vector for bloodstream form transfection, substituting pT7-stl (Brandenburg et al., 2007). For yeast 2-hybrid analysis, pACT2-TbORC1 (the GAD-TbORC1 fusion construct) was generated by inserting the PCR amplified TbORC1 ORF into the BamHI and EcoRI sites of pACT2. pBTM116-TbORC1 (the LexABD-TbORC1 fusion protein expression construct) was generated by inserting the PCR amplified TbORC1 ORF into the BamHI and EcoRI sites of pBTM116. All cloning primer sequences are available upon request.

T. brucei Cell lines

Lister 427 procyclic and bloodstream antigenic type MITat 1.2 clone 221a T. brucei (VSG2-expressor) were used. VSG2 is also known as VSG221. Cells harboring the T7 polymerase and the tetracycline (Tet) repressor, ‘single marker’ (SM, BF) and 29–13 (PF) (Wirtz et al., 1999), were used for conditional expression of the TbORC1 double-stranded RNA (dsRNA). BF and PF cells were cultured in HMI-9 and SDM-79 medium, respectively, with appropriate drug selection.

Cells carrying one endogenous allele of N-terminal F2H or C-terminal PTP tagged TbORC1 were generated by transfecting KpnI/SacII-digested pF2H-ORC1 or SalI-digested pORC1-PTP-NEO into SM or 29–13 cells, respectively. C-terminal HA-tagging of TbORC1 was achieved by transfecting a PCR product containing the 3xHA and HYG flanked by targeting sequences homologous to the C-terminal 76 bp and the 3′ UTR of TbORC1 as described in (Oberholzer et al., 2006).

TbORC1 RNAi cell lines were established by transfecting SacII-digested pStLORC1 construct into PF 29–13 cells as described in (Chandler et al., 2008) or by transfecting SacII-linearized pStLBFORC1 into BF SM cells. Switching assay cell line (HSTB-611) was established as follows: HSTB-261 (Kim and Cross, 2010) containing two reporters, a blasticidin-resistance gene (BSD) at the active BES promoter and a puromycin-resistance gene conjugated with Herpes simplex virus thymidine kinase (PUR-TK) at the 70-bp repeats were transfected with a TbORC1 RNAi plasmid (pBLSOrc1) linearized by SacII digestion. One allele of TbORC1 was then epitope-tagged in-situ with 3xHA at the C-terminus by PCR (Oberholzer et al., 2006). All transfections were performed using an AMAXA Nucleofector (Lonza) according to manufacturer’s protocol.

Immunofluorescence analysis

Immunofluorescence was carried out as described previously (Lowell and Cross, 2004) with minor modifications. Briefly, BF T. brucei cells were harvested and fixed with 2% formaldehyde and permeabilized with 0.2% NP40 in PBS. After blocking with PBS/Gelatin, cells were incubated with Alexa-488 conjugated anti-VSG3 (also known as VSG224, a kind gift from Nina Papavasiliou) and Alexa-647 conjugated anti-VSG13 (also known as VSG113, a kind gift from Nina Papavasiliou) in Fig. 3. Alternatively, cells were incubated with chicken (For Fig. S4, B&C) or rabbit (For Fig. S4A) antibodies specific for VSG2, rabbit antibodies specific for VSG13 or VSG3 followed by incubation with secondary antibodies in Fig. S4. Basal bodies were labeled using the monoclonal antibody YL1/2 (Abcam). Cell images were captured by using a DeltaVision image restoration microscope (Applied Precision/Olympus), deconvoluted by using measured point spread functions, and edited with Photoshop. A series of images was taken for each time point at different layers (z stack images). A single layer of the images was shown to better illustrate the localization of derepressed VSGs.

BrdU labeling experiments

Uninduced (Day 0) and induced (Day 2, 4 and 6) TbORC1 RNAi PF cells were incubated for 3 hours with 50 μM 5-bromo-2-deoxyuridine (BrdU) and 50μM 2′-deoxycytidine in a pulse labeling experiment. Cells were settled on poly-L-lysine coated slides before fixation in 4% paraformaldehyde. Cells were permeabilized in methanol overnight followed by incubation in 1.5 N HCl for 20 minutes. Subsequent anti-BrdU antibody and secondary antibody incubation was carried out as described above.

Incorporation of BrdU into TbORC1 RNAi BF parasites was assayed after incubation with 25 mM BrdU and 50 μM 2′-deoxycytidine for 24 hours (Modified from (Tu and Wang, 2004; Muller et al., 2002)). Cells were harvested (1,300 × g for 10 min at 4°C) washed in 1× PBS with 0.1 % (w/v) glucose (PBS-G), and resuspended in 1× PBS-G with. Cells were then fixed in 2% paraformaldehyde for 5 min followed by adhesion to poly-L-lysine coated slides. Cells were permeabilized with 0.2 % NP-40 in 1× PBS for 10 min followed by incubation with 2 N HCl for 20 min at RT. Cells were incubated with anti-BrdU mouse monoclonal antibody PRB-1 (1:50) for 60 min followed by secondary Alexa-Fluor 594 goat anti-mouse (1:50) for 60 min. Cells were then washed, stained with 3 μg/mL DAPI for 5 min and mounted in Vectashield. For incorporation of BrdU into PF, 200 cells per time point were analyzed and for BF incorporation, 150 cells per time point were analyzed using a Nikon E600 microscope and Diagnostic Instruments Spot-RT CCD camera.

Immunoprecipitation

1×10⁸ PF WT cells or cells carrying an TbORC1-PTP at its endogenous locus were lysed by vortexing cell suspension with acid washed glass beads three times, each for 40 seconds followed by chilling on ice for 40 sec. Protein extract was pre-cleared with pre-washed protein G beads (Sigma) followed by incubation with IgG Sepharose 6 Fast Flow (GE Healthcare). IP product was washed five times with TST buffer (50 mM Tris-buffer pH 7.6, 150 mM NaCl, 0.05 % Tween 20) and once with 5 mM NH₄Ac, pH 5.0 then eluted with 2M glycine pH 2.2. 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. Samples were analyzed using western blotting.

Chromatin IP (ChIP)

ChIP was carried out as described previously (Siegel et al., 2009) with minor modifications. 1×10⁸ cells were harvested and fixed with 1% formaldehyde for 20 minutes at room temperature. BF cell lysates were sonicated using a Bioruptor®300 for 4 cycles with 30s on/off each at high output and incubated with Dynabeads (Invitrogen) and specific antibodies to immunoprecipitate protein-DNA complexes. PF cell lysates were sonicated using a Bioruptor®300 for 10 cycles with 30s on/off each at medium output and incubated with IgG Sepharose 6 Fast Flow (GE Healthcare). Immunoprecipitated DNA was purified using the Qiaquick spin PCR purification Kit (QIAGEN) and analyzed by Southern hybridization using telomere specific probes.

Quantitative RT-PCR was carried out as described in (Yang et al., 2009).

Yeast 2-hybrid analyses were carried out as described in (Li et al., 2005).

Western blot

Western blotting was carried out using the following antibodies: for HSTB-611 cells, TbOrc1-HA, histone H3, and tubulin were detected using anti-HA (Sigma, rabbit-polyclonal), anti-H3 (Abcam, rabbit-polyclonal), and anti-beta tubulin TAT-1 (a generous gift from K. Gull (Sherwin and Gull, 1989)) antibodies. In TbORC1 RNAi cells, F2H-TbOrc1 was detected using 3F10 (Roche Applied Science) or 12CA5 (MSKCC, AB core facility), monoclonal antibodies against HA. HSP70 was detected using a specific antibody against the C. fasciculata protein (1:10,000) (Effron et al., 1993) followed by secondary chicken anti-rabbit IgG-HRP (1:10,000). Rabbit antibodies specifically against VSG2, VSG13 (as described in (Yang et al., 2009)) and VSG9 (also known as VSG VO2, a kind gift from Piet Borst) were used to detect expression of these VSGs. Expression of LexA Binding Domain- and Gal4 Activation Domain-fusion proteins were detected using monoclonal antibody against LexA (Santa Cruz Biotechnology, Inc.) and monoclonal antibody against GAL4 (Abcam Inc.).

VSG switching assay

VSG-switching phenotypes were determined by transiently knocking down TbORC1 transcripts. HSTB-611 cells were maintained in the presence of blasticidin and puromycin to homogenize the cell population (VSG2-expressors). Cells were then allowed to switch in the absence of drug selection for two days, with or without TbORC1 RNAi induction (1μg/ml tetracyclin) in triplicate. Plating efficiency was assessed to determine cell viability after TbORC1 knockdown and this was taken into account to calculate VSG switching frequency. TbORC1 RNAi induced cells were recovered for one additional day. Cells were collected and incubated with anti-VSG2 antibodies. Unswitched VSG2-expressing cells were depleted by magnetic-activated cell sorting (MACS) method (Boothroyd et al., 2009). The column flow-through enriched with switchers was diluted in medium containing 10 μg/ml GCV and distributed into 96-well plates. Switching frequency was determined as the ratio of GCV-resistant cells to the total number of viable cells applied to the MACS column. Cloned switchers were analyzed for blasticidin sensitivity at 100 μg/ml. PCR experiments were carried out to determine the presence of BSD and VSG2 genes in the switchers. All sequences of primers used in this study are available upon request.

Flow Cytometry was carried out according to (Li et al., 2005).