Trypanosoma brucei TIF2 and TRF Suppress VSG Switching Using Overlapping and Independent Mechanisms

Trypanosoma brucei causes debilitating human African trypanosomiasis and evades the host’s immune response by regularly switching its major surface antigen, VSG, which is expressed exclusively from subtelomeric loci. We previously showed that two interacting telomere proteins, TbTRF and TbTIF2, are essential for cell proliferation and suppress VSG switching by inhibiting DNA recombination events involving the whole active VSG expression site. We now find that TbTIF2 stabilizes TbTRF protein levels by inhibiting their degradation by the 26S proteasome, indicating that decreased TbTRF protein levels in TbTIF2-depleted cells contribute to more frequent VSG switching and eventual cell growth arrest. Surprisingly, although TbTIF2 depletion leads to more subtelomeric DNA double strand breaks (DSBs) that are both potent VSG switching inducers and detrimental to cell viability, TbTRF depletion does not increase the amount of DSBs inside subtelomeric VSG expression sites. Furthermore, expressing an ectopic allele of F2H-TbTRF in TbTIF2 RNAi cells allowed cells to maintain normal TbTRF protein levels for a longer frame of time. This resulted in a mildly better cell growth and partially suppressed the phenotype of increased VSG switching frequency but did not suppress the phenotype of more subtelomeric DSBs in TbTIF2-depleted cells. Therefore, TbTIF2 depletion has two parallel effects: decreased TbTRF protein levels and increased subtelomeric DSBs, both resulting in an acute increased VSG switching frequency and eventual cell growth arrest.


Introduction
Trypanosome brucei is a protozoan parasite that causes fatal African trypanosomiasis in humans and nagana in cattle. While proliferating in the extracellular spaces of its mammalian balancing subtelomere stability and plasticity is important for parasite survival, and factors that influence the amount of subtelomere DSBs will influence VSG switching frequency.
So far, T. brucei TIF2 is the only protein that has been shown to influence the amount of subtelomeric DSBs [19]. TbTIF2 is an intrinsic component of the T. brucei telomere protein complex [19]. It interacts tightly with the duplex telomere DNA binding factor, TbTRF, which has been shown to play an essential role in maintaining the terminal telomere structure [29]. Telomere proteins are well-known for their roles in protecting the chromosome ends from illegitimate DNA processes including degradation, fusion, and recombination [30]. However, the functions of telomere proteins in maintaining subtelomere integrity are not clear. We found that depletion of TbTIF2 resulted in a significant increase in the amount of DSBs at subtelomeres and subsequent elevated VSG switching frequency [19], demonstrating for the first time that telomere proteins also play important roles in maintaining subtelomere integrity. However, the underlying mechanism of TbTIF2 function is not clear. Interestingly, we found that TbTRF also suppresses VSG switching, and the telomeric DNA binding activity of TbTRF is essential for this function [18]. Importantly, depletion of either TbTIF2 or TbTRF increases the amount of GC-mediated VSG switching events mostly involving the whole ES, suggesting that they may function in a same genetic pathway in influencing VSG switching [18,19]. In mammals, TIN2, the functional homologue of TbTIF2 [19,31] interacts with both TRF1 and TRF2, which bind the duplex TTAGGG repeats [32][33][34] and are functional homologues of TbTRF [18,29]. In addition, the interaction between TIN2 and TRF1 is essential for preventing TRF1 from being ADP-ribosylated by a TRF1-interacting factor, Tankyrase 1 [35,36], which releases TRF1 from the telomere DNA [37] and allows subsequent ubiquitination of TRF1 by SCF Fbx4 [38] followed by proteasome-mediated degradation [38,39]. Functions of TIN2, TRF1, and TRF2 are partly overlapping because TIN2 interacts with both TRF1 and TRF2 and is the central protein in the six-membered mammalian telomere complex called Shelterin [40].
Because TbTIF2 interacts with TbTRF, and because depletion of either protein leads to an acute increase in VSG switching frequencies with similar switching mechanisms and eventual cell death [18,19], we decided to examine whether these two proteins function in the same pathway, which will help reveal the underlying mechanism of TbTIF2's function in maintaining subtelomere stability. In this work, we found that depletion of TbTIF2 decreased TbTRF protein levels but did not affect its mRNA levels. We further found that TbTRF was degraded by the proteasome upon depletion of TbTIF2. However, depletion of TbTRF did not increase the amount of DSBs inside the subtelomeric VSG ESs. Furthermore, expression of an ectopic TbTRF WT allele delayed TbTRF degradation, mildly improved cell growth, and partially suppressed the phenotype of elevated VSG switching frequency in TbTIF2 RNAi cells. However, this did not suppress the phenotype of increased amount of subtelomeric DSBs in TbTIF2 RNAi cells. Therefore, our observations indicate that increased amounts of subtelomeric DSBs and decreased TbTRF protein levels are two independent and parallel consequences of TbTIF2 depletion, both contributing to increased VSG switching frequencies and eventual cell growth arrest.

Chromatin IP (ChIP)
ChIP was performed exactly the same way as described in [18]. Briefly, Cells were fixed with 1% Formaldehyde for 30 min at room temperature and cell lysate was sonicated in a Bioruptor (Diagenode Corp.) at medium level for 6 cycles (30 sec on/30sec off per cycle) at 4°C. IP was carried out using protein G Dynabeads (Life Technologies) coupled with a TbTRF antibody [29] followed by wash, elution, reverse crosslinking, and DNA isolation. ChIP products were hybridized with a TTAGGG repeat or a 50 bp repeat probe in slot blot Southern analysis.

VSG Switching Assay
Switching assays were performed according to [18,41]. Detailed switching assay protocol is included in the supplemental materials.

Ligation mediated PCR (LMPCR)
LMPCR were performed the same way as described previously [19,27]. Briefly, in each ligation reaction, 2 μg of genomic DNA was either treated or not treated with 2 μl of T4 DNA Polymerase (3000 U/ml, New England BioLabs) in the presence of 200 μM dNTP and then ligated with 10 μl annealed adaptor. Three 1:3 serial dilutions of the ligated products were prepared and used in subsequent PCR using Hotstart Platinum 1 Taq DNA Polymerase (Life Technologies) and a touchdown PCR program.

MG-132 treatment
T. brucei cells were incubated in medium with 25 μM of MG-132 for 6 hours before proteins were extracted for western analysis. For RNAi induced samples, MG-132 was added to the medium 6 hours prior to the time point of harvesting cells.

TbTIF2 protects TbTRF from being degraded by the 26S proteasome
We observed a similar growth arrest phenotype upon RNAi induction in TbTIF2 and TbTRF RNAi cells [19,29]. In addition, a transient TbTIF2 or TbTRF RNAi induction resulted in a significant increase in VSG switching frequency with most VSG switchers arising from subtelomeric gene rearrangements that resulted in the loss of the whole active ES [18,19]. Therefore, we hypothesized that TbTIF2 and TbTRF function in the same genetic pathway in maintaining subtelomere stability.
We first tested whether TbTRF protein levels were affected by TbTIF2. Indeed, induction of TbTIF2 RNAi led to depletion of not only the endogenous FLAG-HA-HA (F2H) tagged TbTIF2 ( Fig 1A) but also the endogenous TbTRF protein (Fig 1A). In contrast, protein levels of TbRAP1, another intrinsic component of the T. brucei telomere complex [42], were not affected ( Fig 1A). Northern blotting analysis showed that only the TbTIF2 mRNA was knocked-down by TbTIF2 RNAi, while TbTRF mRNA levels were not affected (Fig 1B), indicating that TbTIF2 is required for maintaining TbTRF protein levels.
To investigate whether TbTRF is degraded by the proteasome upon TbTIF2-depletion, we treated TbTIF2 RNAi cells with and without MG-132, which is a specific proteasome inhibitor and has previously been successfully used in T. brucei [45]. We found that in cells not treated with MG-132, TbTIF2 was depleted upon induction of RNAi (the relative levels of TbTIF2-F2H are 100%, 81%, and 79% at 0, 24, and 40 hrs after induction, respectively), which led to a quick TbTRF depletion (Fig 1C). In contrast, in cells treated with MG-132, although TbTIF2 was again depleted by RNAi at a very similar rate (the relative levels of TbTIF2-F2H are 100%, 82%, and 72% at 0, 24, and 40 hrs, respectively), TbTRF protein levels were not decreased ( Fig  1C), indicating that TbTRF is degraded by the proteasome when TbTIF2 is absent. As a control, we also examined the protein levels of TbPSA6 and TbAUK1. TbPSA6 is the A6 subunit of the 20S proteasome and is not degraded by the proteasome [46], while TbAUK1, the T. brucei Aurora like kinase, is degraded by the proteasome [47]. As expected, adding MG-132 blocks the activity of the proteasome and the TbAUK1 protein level was higher in these cells than in cells not treated with MG-132 ( Fig 1C). In contrast, the TbPSA6 protein level was not affected by MG-132 ( Fig 1C). In this particular induction, the TbAUK1 protein level appears to be slightly increased upon TbTIF2 depletion in the absence of MG-132 ( Fig 1C). To better examine the effect of TbTIF2 depletion on TbAUK1 protein levels, we repeated the induction three more times and quantified the TbAUK1-3HA protein levels in Adobe Photoshop using tubulin or EF-2 as a loading control. On average (calculated from four independent inductions), TbAUK1-3HA protein levels changed from 100% at 0 hr to 106.8% at 24 hrs and 103.8% at 40 hrs after induction, indicating that TbTIF2 does not affect TbAUK1 protein levels significantly. Therefore, TbTIF2 regulates TbTRF protein levels by inhibiting its degradation by the proteasome.
We also examined both TbTRF and TbTIF2 protein levels in TbTRF RNAi cells. Upon RNAi induction, we only observed depletion of TbTRF, while the TbTIF2 protein level was (B) Northern blotting shows that TbTRF mRNA levels were not affected in TbTIF2-depleted cells. rRNA precursors were shown as a loading control. (C) Western blotting was performed using whole cell extracts prepared from TbTIF2 RNAi cells treated with and without MG-132 (Sigma), an inhibitor of the 26S proteasome. TbAUK1 is normally degraded by the 26S proteasome while TbPSA6 is not. (D) Western analysis showed that TbTIF2 protein levels are not sensitive to TbTRF depletion. In western analyses, TbTIF2-F2H and TbAUK1-3HA were detected using an HA monoclonal antibody (F-7, SantaCruz Biotechnology). TbTRF [29], TbRAP1 [42], TbPSA6 [43], and tubulin [44] were detected by their respective antibodies. EF-2 was detected using a goat polyclonal antibody against human EF-2 (Santa Cruz Biotechnology). All primary antibodies were diluted 1,000 fold. In this and other figures, proteins from 15 million cells were loaded each lane for western blotting, except when detecting tubulin, proteins from only 0.5 million cells were loaded each lane. stable for at least 24 hrs. The mild decrease in TbTIF2 protein levels after 36 hrs of TbTRF RNAi induction is likely due to the severe growth defect in TbTRF-depleted cells [29].
TbTRF does not affect the amount of subtelomere DNA double strand breaks (DSBs) The fact that depletion of TbTIF2 leads to depletion of TbTRF further suggests that TbTIF2 and TbTRF function in the same pathway to influence VSG switching and allow normal cell proliferation. To further investigate this possibility, we tested whether depletion of TbTRF results in more subtelomeric DSBs as we observed in TbTIF2 RNAi cells. Due to extremely limited availability of specific antibodies against γH2A, which has been shown to be deposited at the chromatin with DSBs [48], we used Ligation-mediated PCR (LMPCR) analysis (Fig 2A) to detect and estimate the amount of DSBs in active and silent VSG ESs as we did previously [19]. LMPCR has a higher resolution than γH2A Immunofluoresence analysis (IF) so that we can determine whether DSBs are in subtelomeric VSG ESs, while IF can only reveal whether DSBs are in telomere vicinity in general (can be either telomeric or subtelomeric or both). Surprisingly, although DSBs were still detected, the amount of DSBs within VSG ESs was not increased when TbTRF was depleted (Fig 2). We examined the single-copy active VSG2 ( Fig  2B) and silent VSG21 (Fig 2C) and multi-copy ES promoter (S1A Fig) and 70 bp repeat regions (S1B Fig), and the results are reproducible. Quantification of LMPCR product levels indicated that there is no significant change observed in the DSB levels inside VSG ESs before and after depletion of TbTRF (Fig 2D). In addition, depletion of TbTRF did not affect the DSB levels at a chromosome internal SNAP50 gene locus (S1C Fig). Therefore, depletion of TbTIF2 appears to have two independent effects: decreasing TbTRF protein levels and increasing subtelomeric DSB levels.
Induced expression of an ectopic TbTRF allele slightly improves cell growth in TbTIF2 RNAi cells To further investigate the functional relationship between TbTRF and TbTIF2 in maintaining subtelomere stability and in cell survival, we introduced an ectopic F2H-tagged WT allele of TbTRF into the TbTIF2 RNAi strain used previously for examining TbTIF2's effect on VSG switching, S/TIF2i [19]. We have shown that F2H-TbTRF has essential TbTRF functions and can rescue mutant phenotypes in TbTRF RNAi cells [18,29]. TbTIF2 RNAi and expression of F2H-TbTRF in S/TRFi+F2H-TbTRF cells can be induced simultaneously by adding doxycycline [49]. Upon adding doxycycline, we observed induced F2H-TbTRF expression within the first 48 hours (Fig 3A). However, expression of F2H-TbTRF was not stably maintained 72 hrs after the induction, because TbTIF2 was depleted soon after induction of TbTIF2 RNAi, as shown in northern blotting analysis (Fig 3B). In addition, we also observed that the endogenous TbTRF protein level decreased significantly 48 hrs after inducing TbTIF2 RNAi (Fig 3A). Even so, at 24-48 hrs after induction, the TbTRF protein level in S/TIFi+F2H-TbTRF cells has not significantly decreased as in S/TIF2i cells, suggesting that expression of the ectopic F2H-TbTRF allele resulted in more TbTRF proteins in the cell that presumably require longer time to be completely degraded after removal of TbTIF2.
Consistent with our observation of delayed TbTRF depletion in S/TIF2i+F2H-TbTRF cells, we found that these cells showed mildly better growth than S/TIF2 cells (Fig 3C). In S/TIF2i cells, cell growth slowed down significantly by 24 hrs and was completely arrested by 72 hrs after induction of TbTIF2 RNAi, while in S/TIF2i+F2H-TbTRF cells, cell growth slowed down by 48 hrs and was completely arrested by 96 hrs after induction. Therefore, maintaining TbTRF protein levels for a longer time slightly delayed the cell growth arrest phenotype in TbTIF2 RNAi cells. This further supports the idea that depletion of TbTIF2 caused two parallel effects that both lead to cell growth arrest.
We previously found that when the DNA binding activity of TbTRF is significantly impaired [less than 20% of TbTRF is associated with the telomeric chromatin in Chromatin IP (ChIP) experiment], the VSG switching frequency is increased [18]. To examine whether TbTRF's telomere binding is affected in S/TIF2i and S/TIF2i+F2H-TbTRF cells, we performed ChIP analysis using a TbTRF antibody [29]. In S/TIF2i cells, the association between TbTRF and the telomere chromatin decreased significantly upon TbTIF2 depletion (Fig 3D,  top), as TbTRF is degraded. In S/TIF2i+F2H-TbTRF cells, within 30 hrs after induction of TbTIF2 RNAi, the association of TbTRF with the telomeric DNA was decreased mildly, but 75% of TbTRF still remained at the telomere (Fig 3D, bottom), indicating that expression of the ectopic F2H-TbTRF helped to keep sufficient TbTRF at the telomere within a short frame of time after TbTIF2 was depleted. This mild decrease of TbTRF binding at the telomere is not expected to affect VSG switching frequencies based on our previous observations [18].

Expression of ectopic TbTRF does not suppress the phenotype of more subtelomeric DSBs in TbTIF2 RNAi cells
We found that depletion of TbTRF does not increase the subtelomeric amount of DSBs (Fig 2), suggesting that more subtelomeric DSBs in TbTIF2-depleted cells are independent of decreased TbTRF protein levels. If this is the case, expression of the ectopic F2H-TbTRF in TbTIF2 RNAi cells will not suppress the phenotype of increased DSB levels. To confirm this, we performed the LMPCR analysis (Fig 2A) in S/TIF2i+F2H-TbTRF cells to detect DSBs at subtelomeres. Indeed, 24 hrs after induction of TbTIF2, when TbTRF level had not decreased significantly and more than 75% of TbTRF were still associated with the telomere, we observed a significant increase in the amount of subtelomeric DSBs (Fig 4) the same as in S/TIF2i cells [19]. We tested the active VSG2 (S3A Fig) and ψ ES1 pseudogene (Fig 4A), the silent VSG21 ( Fig 4B) and ψ ES11 pseudogene (data not shown), and 70 bp repeats (S3B Fig) and observed similar results at all tested loci. To compare the fold change in DSB levels in S/TIF2i+F2H-TbTRF cells with that in S/TIF2i cells [19], we also quantified the fold change in DSB levels at multiple gene loci including the active VSG2 and ψ ES1 loci, the silent VSG21 and ψ ES11 loci, 70 bp repeats, and the chromosome internal SNAP50 locus (Fig 4C). In S/TIF2i+F2H-TbTRF cells, depletion of TbTIF2 led to~1.5-3.5 fold increase in subtelomeric DSB levels, which is the same as in S/ TIF2i cells [19]. This observation also confirmed that induction of RNAi in S/TIF2i+F2H-TbTRF cells efficiently depleted TbTIF2 as we did previously in S/TIF2i cells.

Expression of ectopic TbTRF in TbTIF2 RNAi cells partially suppresses the phenotype of increased VSG switching frequency
Our observations suggest that the two effects of TbTIF2 depletion, increased subtelomeric DSB levels and decreased TbTRF protein levels, are independent of each other. Still, both are expected to contribute to subsequent increased VSG switching frequency. To test this, we estimated the VSG switching frequencies in S/TIF2+F2H-TbTRF cells the same way we did previously [18]. Both S/TIF2i and S/TIF2+F2H-TbTRF strains were derived from the HSTB261 strain, in which the active VSG2-containing ES are marked with a Blasticidin resistance (BSD) marker immediately downstream of the ES promoter and a Puromycin resistance (PUR)-Thymidine Kinase (TK) fusion gene between the active VSG2 and the 70 bp repeats [15].
To be able to compare current results with previous observations, we performed the switching assay in S/TIF2i, S/TIF2+F2H-TbTRF, and S/ev control cells that carry an empty RNAi construct, using exactly the same switching protocol as before [18]. Cells were induced for 30 hrs. Then doxycycline was removed by extensive washing [19]. This allows cells to recover from the growth arrest so that VSG switchers can subsequently be obtained. To enrich for VSG switchers, cells were first mixed with MACS beads conjugated with a VSG2 monoclonal antibody. Non-switchers that still express VSG2 are trapped with MACS beads, while switchers were in the flow through fraction, which is further selected by ganciclovir [15]. VSG switchers should no longer express the TK gene targeted immediately upstream of the active VSG2, and cells not expressing TK are resistant to ganciclovir, a nucleoside analog [50].
We found that VSG switching frequencies in S/TIF2i+F2H-TbTRF cells are lower than that in S/TIF2i cells but are higher than that in S/ev cells (Fig 5A), supporting the idea that expression of an ectopic F2H-TbTRF allele partially suppressed the phenotype of increased VSG switching frequency in TbTIF2 depleted cells.
By determining the expression status and genotypes of the markers in the obtained VSG switchers, we can reveal the underlying mechanisms of switching events [15,18,19]. Previously, we found that although VSG switchers that lost the original active ES are already the most frequent switching events (70%), depletion of TbTIF2 further increased the percentage of such VSG switchers (80-90%) [19]. Here we estimated the VSG switching frequency in S/TIF2i cells using the MACS-based VSG switching assay and observed the same phenotype, where VSG switchers that lost the original active ES consist of 90% of all switchers (Fig 5B). In addition, expression of the ectopic F2H-TbTRF reduced the percentage of this type of switchers to 80% (Fig 5B), which is between the levels in S/TIF2i cells and S/ev cells [19]. Therefore, examination of VSG switching in S/TIF2i+F2H-TbTRF cells further indicates that expressing F2H-TbTRF and maintaining cellular TbTRF protein levels at the time of performing switching assays partially suppresses phenotypes in TbTIF2 RNAi cells.

Discussion
Human African trypanosomiasis is fatal without treatment. However, only few drugs are available for treating T. brucei infections, all with severe side effects. We previously showed that both TbTIF2 and TbTRF are essential for normal cell proliferation and have weak sequence homology with their mammalian homologues only within functional domains [19,29]. Therefore, TbTIF2 and TbTRF could serve as anti-parasite drug targets. However, the underlying mechanisms of TbTIF2 and TbTRF functions are unclear, and whether TbTIF2 and TbTRF function in the same genetic pathway is unknown. Our current investigation reveals that the functions of these two telomere proteins are not identical but are partially overlapping.
The mechanisms of TbTIF2 and TbTRF functions are at least partially overlapping. TbTRF interacts with TbTIF2, and both suppress subtelomeric gene conversion events involving the whole active VSG ES [18,19]. More importantly, depletion of TbTIF2 also leads to depletion of TbTRF (Fig 1). Clearly, decreased TbTRF protein levels in TbTIF2-depleted cells is an important contributing factor for the growth defect and increased VSG switching frequency, as TbTRF is essential for normal cell growth and also suppresses VSG switching [18,29]. products using primers specific to the TbRAP1 gene (as a loading control) are shown at the bottom. (C) Quantification of the change in the amounts of LMPCR products (with T4 DNA pol treatment using 54 ng input gDNA) from three independent experiments. Average values are shown. Error bars represent standard deviations. Numbers represent P values of unpaired t-tests between pairs of data groups as indicated.
doi:10.1371/journal.pone.0156746.g004 TbTIF2 has an additional important function in maintaining subtelomere integrity [19], which we now show is independent of TbTRF. Depletion of TbTRF does not affect the amount of DSBs inside VSG ESs (Fig 2); expressing an ectopic WT allele of TbTRF in S/TIF2i cells does not suppress the phenotype of more subtelomere DSBs but mildly improved the cell growth and partially suppressed the phenotype of increased VSG switching frequency. Therefore, functions of TbTIF2 and TbTRF are not identical.
Based on our observations, we conclude that the functional interaction pathways between TbTIF2 and TbTRF are as follows: TbTIF2 depletion has two independent effects: increased amount of subtelomere DSBs and decreased TbTRF protein levels. Both effects are known to contribute to growth defect and increased VSG switching frequencies [18,19,29]. First, DSBs induced in the active VSG ES are detrimental to T. brucei cells and cause >85% of cell death [28]. At the same time, DSBs induced near the 70 bp repeats upstream of the active VSG gene are also a potent inducer for VSG switching [27,28]. Therefore, the increased amount of subtelomere DSBs in TbTIF2 RNAi cells is an important contributor for cell growth arrest and increased VSG switching frequency. Second, TbTRF is known to have an essential role in maintaining the terminal telomere structure, which is likely the reason why it is essential for normal cell proliferation [29]. TbTRF also suppresses subtelomere gene conversion events that lead to VSG switching [18]. Therefore, decreased TbTRF protein levels induced by TbTIF2 depletion are also an important factor contributing to an acute elevated VSG switching frequency and eventual growth arrest.
It is interesting that TbTIF2 is essential for maintaining subtelomere integrity, while its interacting factor TbTRF is not. Since TbTRF binds the duplex telomere DNA directly [29], it is possible that TbTRF is more important for maintaining telomere than subtelomere integrity. Our previous observation that TbTRF is important for maintaining the terminal telomere structure is consistent with this view [29]. Mammalian TRF2, the functional homologue of TbTRF [29], is also essential for maintaining the telomere terminal G-overhang structure and for preventing chromosome end-to-end fusions through the non-homologous end-joining (NHEJ) pathway [51][52][53]. We speculate that loss of TbTRF may also result in chromosome end fusions and subsequent breakage-fusion-bridge cycle, which often result in large terminal chromosome deletions [54]. Because VSG is essential, if the chromosome end containing the active VSG ES is lost, then only VSG switchers can survive. As a consequence, TbTRF depletion would result in increased VSG switching [18]. In this case, the DNA break sites can locate well upstream of the active VSG ES, hence were not detected in the LMPCR analysis where ES-specific probes were used. Although the essential player in NHEJ, DNA ligase IV, has not been identified in T. brucei [55], micro-homology mediated end-joining has been observed in T. brucei [55,56], which can also mediate telomere end fusions. Due to the fact that T. brucei chromosomes do not condense during mitosis and the limited sensitivity of our current molecular tools, telomere end-to-end fusions have not been detected in T. brucei yet. Developing more sensitive molecular tools would help their detection in the future.
Previous studies indicated that DSBs downstream of the active VSG gene mostly lead to switchers that lost the whole active ES [28]. Therefore, it is also possible that TbTRF depletion may cause DSBs only in the telomere repeats that lead to the loss of the active VSG ES and subsequent VSG switching. Although these cells have telomerase activity that can elongate telomere sequences after telomere breaks [57,58], it has been shown that the active telomere are more prone to large telomere truncations than silent telomeres [59], presumably because the active telomere is expressed [60], which make it more fragile [61]. DSBs inside the telomere repeats may not be easily detected by the LMPCR analysis, as the most telomere-proximal locus-specific probe is located at the 5' end of the VSG gene, and DSBs located a few kilobases downstream may not be efficiently detected by LMPCR due to limited PCR efficiency.
Depletion of TbTRF does not increase the amount of DSBs in subtelomeric ESs, while depletion of TbTIF2 does, indicating that TbTIF2 has essential functions independent of TbTRF. Therefore, it would be interesting to investigate whether recruiting TbTIF2 to telomeres is completely dependent on TbTRF, as TbTRF has a duplex telomere DNA binding activity [29] while TbTIF2 does not have any functional domain suggesting DNA binding activity [19]. Although we have not identified any telomere proteins other than TbTRF [29], TbTIF2 [19], and TbRAP1 [42] in T. brucei, it is very possible that the T. brucei telomere complex has more than just three protein members. For example, T. brucei has a telomere G-rich 3' overhang at the very end [63]. Although this overhang is short, we expect it to be bound and protected by proteins with single stranded DNA binding activity. It has been shown that Leishmania RPA1 can bind telomeric single stranded DNA [64]. Since T. brucei and Leishmania are closely related, it is possible that T. brucei RPA1 can also bind telomeric G-overhang. If TbTIF2 interacts with single-stranded telomere DNA binding factors, it can also be recruited to the telomere independently of TbTRF.
Our studies further validate that both TbTIF2 and TbTRF are essential for normal T. brucei proliferation. Since these proteins are only weakly homologous to their mammalian counterparts [19,29], they are good potential targets for anti-parasite agents. In addition, our results indicate that targeting TbTIF2 would be more efficient than targeting TbTRF, as TbTIF2 depletion causes simultaneous depletion of TbTRF in addition to increased subtelomeric DSB numbers. In human cells, TIN2 stabilizes TRF1 proteins by preventing Tankyrase 1 from modifying TRF1 [35,36], which in turn prevents TRF1 from being ubiquitinated by SCF FBX4 [38] and subsequent degradation by the 26S proteasome [38,39]. However, tankyrase 1 and SCF FBX4 homologs have not been identified in T. brucei, and whether TbTRF is ubiquitinated before it is degraded by the proteasome is unknown. Further studies on TbTRF modifications will help to reveal the detailed mechanisms of how TbTIF2 regulates TbTRF protein levels.