|Home | About | Journals | Submit | Contact Us | Français|
Orthologs of the yeast telomere protein Stn1 are present in plants, but other components of the Cdc13/Stn1/Ten1 (CST) complex have only been found in fungi. Here we report the identification of Conserved Telomere maintenance Component 1 (CTC1) in plants and vertebrates. CTC1 encodes an ~140 kDa telomere-associated protein predicted to contain multiple OB-fold domains. Arabidopsis mutants null for CTC1 display a severe telomere deprotection phenotype accompanied by a rapid onset of developmental defects and sterility. Telomeric and subtelomeric tracts are dramatically eroded, and chromosome ends exhibit increased G-overhangs, recombination, and end-to-end fusions. AtCTC1 both physically and genetically interacts with AtSTN1. Depletion of human CTC1 by RNAi triggers a DNA damage response, chromatin bridges, increased G-overhangs and sporadic telomere loss. These data indicate that CTC1 participates in telomere maintenance in diverse species and that a CST-like complex is required for telomere integrity in multicellular organisms.
The terminus of a linear chromosome must be distinguished from a double-strand break to avoid deleterious nucleolytic attack and recruitment into DNA repair reactions. Telomeres prevent such actions by forming a protective cap on the chromosome end. This cap consists of an elaborate, higher-order DNA architecture and a suite of telomere-specific proteins. The formation of a t-loop of telomeric DNA is thought to play an important role in sequestering the terminal single-strand G-overhang from harmful activities (de Lange, 2004; Wei and Price, 2003), while double-strand (ds) and single-strand (ss) telomeric DNA binding proteins coat the chromosome terminus to further distinguish it from a double-strand break (Palm and de Lange, 2008)
In Saccharomyces cerevisiae, telomeres are bound by a trimeric protein complex, termed CST, composed of Cdc13, Stn1 and Ten1 (Gao et al., 2007; Lundblad, 2006). The three proteins interact to form an RPA-like complex with specificity for ss telomeric DNA. Cdc13 and Stn1 harbor at least one oligonucleotide-oligosaccharide binding (OB) fold, which in the case of Cdc13 is exploited to bind to the G-overhang (Guo et al., 2007; Mitton-Fry et al., 2002). Stn1 and Ten1 associate with the overhang primarily via interactions with Cdc13. The CST complex plays a key role in telomere length regulation (Bianchi and Shore, 2008). Cdc13 recruits the telomerase RNP via a direct interaction with the Est1 component of telomerase (Bianchi et al., 2004; Chandra et al., 2001), while Stn1 is thought to inhibit telomerase action by competing with Est1 for Cdc13 binding (Li et al., 2009; Puglisi et al., 2008). In addition, Cdc13 and Stn1 contribute to coupling of G- and C-strand synthesis through interactions with DNA polymerase α (Grossi et al., 2004; Qi and Zakian, 2000).
The CST complex is also essential for chromosome end-protection. Mutations in any one of the CST components result in degradation of the C-strand, accumulation of ss G-rich telomeric DNA and late S/G2 cell cycle arrest (Garvik et al., 1995; Grandin et al., 2001; Grandin et al., 1997). Telomere protection appears to be facilitated primarily by Stn1 and Ten1, and overexpression of Stn1 and Ten1 can rescue the lethality of Cdc13 depletion (Grandin et al., 2001; Petreaca et al., 2007; Puglisi et al., 2008). Finally, Cdc13 and Stn1 also inhibit telomere recombination (Iyer et al., 2005; Petreaca et al., 2006; Zubko and Lydall, 2006).
Mammalian telomeres are bound by shelterin, a six-member complex that, unlike CST, binds both ss and ds telomeric DNA (Palm and de Lange, 2008). The shelterin proteins TRF1 and TRF2 coat ds telomeric DNA, while POT1 binds the ss G-overhang. The TRF1/TRF2-interacting protein TIN2 and the POT1-interacting protein TPP1 associate with each other, providing a bridge between the duplex and ss regions of telomeric DNA. RAP1 associates with telomeres via interaction with TRF2. The majority of shelterin components are implicated in telomere capping, although TRF2 and POT1 appear to play pivotal roles in this process. TRF2 associates with telomeric DNA via a myb-like DNA binding domain. Loss of telomere-bound TRF2 results in immediate degradation of the G-overhang and end-to-end chromosome fusions (Celli and de Lange, 2005), while certain dominant negative alleles cause rapid telomere shortening with extrusion of extra-chromosomal telomeric circles via homologous recombination (Wang et al., 2004).
Like components of the CST complex, POT1 and its partner TPP1 harbor OB-folds. POT1 binds directly to the overhang through two adjacent OB-folds, thus sequestering the DNA 3′ terminus and reducing access to telomerase (Lei et al., 2004; Lei et al., 2005). TPP1 does not bind DNA directly, but dimerization with POT1 increases the DNA-binding affinity of POT1 by ~10 fold (Wang et al., 2007). Knockdown of human POT1 by RNAi causes a fairly mild phenotype characterized by impaired proliferation, an increase in chromosome fusions, decreased G-overhang signals and an increase in telomere length (Hockemeyer et al., 2005; Veldman et al., 2004; Yang et al., 2005; Ye et al., 2004). Disruption of the POT1 gene leads to more dire consequences (Churikov et al., 2006; Hockemeyer et al., 2006; Wu et al., 2006) including activation of a strong ATR-mediated DNA damage checkpoint, G-overhang elongation, rapid telomere growth, elevated telomere recombination and ultimately cell death (Churikov and Price, 2008; Denchi and de Lange, 2007; Guo et al., 2007).
Telomere protein composition may be more conserved than previously surmised (Linger and Price, 2009). At least one shelterin component, Rap1, is present in S. cerevisiae, although unlike vertebrate RAP1, ScRap1p directly binds ds telomeric DNA through two myb-like DNA binding domains and contributes to telomere length regulation and telomere silencing (Lundblad, 2006). Likewise, fission yeast contain several shelterin orthologs including Taz1, an ortholog of mammalian TRF1/TRF2 proteins (Cooper et al., 1997), and Pot1 (Baumann and Cech, 2001). Furthermore, recent purification of SpPot1-associated proteins identified Tpz1, a presumed ortholog of vertebrate TPP1 (Miyoshi et al., 2008). Like TPP1, Tpz1 contains an OB-fold, and physical association of SpPot1 and Tpz1 is required for chromosome end protection (Miyoshi et al., 2008; Xin et al., 2007). The Pot1-Tpz1 complex recruits two additional proteins, Ccq1 and Poz1. Poz1 serves as a bridge linking the Pot1-Tpz1 complex to the ds telomere proteins Rap1 and Taz1 in a manner similar to the shelterin component TIN2 (Miyoshi et al., 2008). Altogether, these findings argue that the core components of the shelterin complex are evolutionary conserved.
Emerging data indicate that CST components are also widespread. Although Cdc13 orthologs have yet to be uncovered outside of S. cerevisiae, a Stn1/Ten1 capping complex was recently described for S. pombe (Martin et al., 2007). Both proteins localize to telomeres and are essential for chromosome end protection from exonucleases and telomere fusions. Notably, no direct physical association between Stn1/Ten1 and Pot1 has been observed (Martin et al., 2007) and mass spectrometry of SpPot1-associated factors failed to identify Stn1 or Ten1 (Miyoshi et al., 2008). These findings suggest that CST and shelterin components may constitute distinct telomere complexes.
Plants also appear to harbor both shelterin and CST components. Several Myb-containing TRF-like proteins from Arabidopsis bind telomeric dsDNA in vitro (Zellinger and Riha, 2007) and in rice genetic data implicate one of these, RTBP1, in chromosome end protection (Hong et al., 2007). Arabidopsis encodes three OB-fold bearing POT1-like proteins (Shakirov et al., 2005; Surovtseva et al., 2007)(A. Nelson, Y. Surovtseva and D. Shippen, unpublished data). Interestingly, while over-expression of a dominant negative allele of AtPOT1b or depletion of AtPOT1c lead to a telomere uncapping phenotype similar to a pot1 deficiency in yeast and mammals (Shakirov et al., 2005)(A. Nelson, Y. Surovtseva and D. Shippen, unpublished data), AtPOT1a is dispensable for chromosome end protection and instead is required for telomerase function (Surovtseva et al., 2007). Currently, orthologs for TIN2, RAP1 and TPP1 cannot be discerned in any plant genome.
Recently, a distant homolog of the CST component STN1 was uncovered in Arabidopsis (Song et al., 2008). AtSTN1 bears a single OB-fold and localizes to telomeres in vivo. Deletion of AtSTN1 results in the rapid onset of growth defects and sterility, coupled with extensive exonucleolytic degradation of chromosome ends, increased telomere recombination, and massive end-to-end chromosome fusion (Song et al., 2008).
Here we report the identification of a new telomere protein, termed CTC1 (Conserved Telomere maintenance Component 1), that physically and genetically interacts with AtSTN1. We show that AtCTC1 localizes to telomeres in vivo and, as for AtSTN1, loss of AtCTC1 triggers rapid telomere deprotection resulting in gross developmental and morphological defects, abrupt telomere loss, telomere recombination and genome instability. Although not as severe as an Arabidopsis ctc1 null mutant, the consequences of CTC1 knockdown in human cells include a DNA damage response, formation of chromatin bridges, increased G-overhang signals and loss of telomeric DNA from some chromosome ends. Altogether, these data argue that CTC1 is a component of a CST-like complex in multicellular organisms that is needed for telomere integrity. Notably, we found that mammalian CTC1 and STN1 correspond to the two subunits of alpha accessory factor (AAF), a protein complex previously shown to stimulate mammalian DNA pol α-primase (Casteel et al., 2009; Goulian and Heard, 1990). Thus, the CST-like complex from plants and mammals may resemble the S. cerevisiae CST by providing a link between telomeric G- and C-strand synthesis.
In an effort to identify mutations in AtPOT1c, we examined lines within a TILLING collection of EMS-mutagenized Arabidopsis plants. A mutant was uncovered that showed a profound telomere uncapping phenotype (described below). However, this phenotype did not segregate with nucleotide changes in AtPOT1c and therefore map-based cloning was employed to identify the lesion responsible for the phenotype. A single-nucleotide transition (G to A) was found in At4g09680, which co-segregated with telomere uncapping. At4g09680 lies on chromosome 4, while AtPOT1c resides on chromosome 2. At4g09680 was designated CTC1 (Conserved Telomere maintenance Component 1) and the point mutant was termed ctc1-1. CTC1 is a single copy gene and sequence analysis of CTC1 cDNA from wild type plants revealed a large ORF with 16 exons that encodes a previously uncharacterized 142 kDa protein (Figure 1A). RT-PCR demonstrated that CTC1 is widely expressed in both vegetative and reproductive organs (Figure S1A). Further analysis of the CTC1 protein sequence is discussed below.
To determine whether CTC1 associates with telomeres in vivo, an N-terminal CFP-tagged version of CTC1 protein was expressed in transgenic Arabidopsis and immunolocalization experiments were performed on different tissues. Nuclear CFP signal was detected in plants expressing CFP-CTC1, but not in untransformed controls (Figure 1B, Figure S1B and data not shown). Telomere distribution was analyzed by fluorescence in-situ hybridization (FISH) using a telomere probe. In Arabidopsis, telomeres lie at the nucleolar periphery (Armstrong et al., 2001) (Song et al., 2008) and, as expected, telomeric FISH signals were positioned in this location. Similarly, CFP-CTC1 was distributed in a punctate pattern surrounding the nucleolus. A merge of these images showed that much of the CFP-CTC1 co-localized with Arabidopsis telomeres (Figure 1B and Figure S1B). CTC1 association with telomeres was quantitated in flowers and seedlings, which contain cycling cells. On average, 51% (n=38, SD=+/-26%) of the telomere signals overlapped with CFP-CTC1. To determine if CTC1 co-localization with telomeres was retained in non-cycling cells, we examined the apical half of rosette leaves that were at least two weeks old and arrested in G1 (Donnelly et al, 1999). In these cells, 44.1% (n=28, standard deviation=+/-24.5%) of the telomeres displayed an overlapping signal with CFP-CTC1. These data argue that CTC1 associates with telomeres throughout the cell cycle.
We next examined the impact of CTC1 inactivation on plant morphology. Sequence analysis of CTC1 cDNA from ctc1-1 mutants revealed that the G(1935)A point mutation resulted in a nonsense codon within the ninth exon (Figure 1A). Two additional CTC1 alleles, ctc1-2 and ctc1-3, bearing T-DNA insertions in the sixth exon or tenth intron, respectively, were identified within the SALK database (Figure 1A). RT-PCR analysis showed that no CTC1 full length mRNA was produced in either ctc1-2 or ctc1-3, indicating that these lines are null alleles of AtCTC1 (Figure S1C).
All three ctc1 mutants displayed a rapid onset of severe morphological defects in the first generation (Figure 1C), confirming that CTC1 lesions are responsible for telomere uncapping. The large majority of ctc1 plants had grossly distorted floral phyllotaxy with an irregular branching pattern and fasciated (thick and broad) main and lateral stems and siliques (Figure 1C). Although most mutants produced an influorescence bolt, this structure was highly variable in size, ranging from very short to wild type (Figure 1C, compare middle and bottom right panels). Flowers and siliques were often fused, and seed yield was typically reduced to ~10% of wild type. The germination efficiency of the few seeds that could be recovered was extremely low, making propagation to the next generation almost impossible.
Terminal restriction fragment (TRF) analysis was performed to examine bulk telomere length in ctc1 plants derived from a single self-pollinated heterozygous parent. In contrast to the telomeres of their wild type and heterozygous siblings, which spanned 2-5 kb in length (Figure 2A, lanes 1 to 4), telomeres in homozygous ctc1-1 mutants were severely deregulated (Figure 2A, lanes 5 and 6). The longest ctc1-1 telomeres were in the wild type range, but a new population of shorter telomeres emerged, the shortest of which trailed to 0.5 kb. Homozygous ctc1-2 and ctc1-3 mutants showed a similar aberrant telomere length phenotype (Figure S2A).
We investigated how individual telomeres were affected by CTC1 loss using subtelomeric TRF analysis with probes directed at specific chromosome termini. As expected (Shakirov and Shippen, 2004), sharp bands were produced from wild type telomeres (Figures 2B and S2B). In contrast, telomeres in ctc1 mutants gave rise to a broad heterogeneous hybridization signal spanning 1.5 kb (Figures 2B and S2B). Primer extension telomere repeat amplification (PETRA) also generated broad smears in ctc1 mutants (Figure 2C), confirming that the length of individual telomere tracts was grossly deregulated. Telomere shortening and increased heterogeneity at individual telomere tracts in ctc1 mutants is not due to a reduction in telomerase activity. Quantitative telomere repeat amplification (Q-TRAP) revealed no significant difference in the in vitro telomerase activity levels in ctc1 mutants relative to wild type (Figure S3).
Next we studied the G-overhang status in ctc1 mutants using non-denaturing in-gel hybridization. Strikingly, the G-overhang signal was ~three times greater in ctc1 mutants relative to wild type (3.5±0.7) (Figure 2D). A similar increase in G-overhang signal is observed in Arabidopsis stn1 mutants (Song et al. 2008). Exonuclease treatment reduced the G-overhang signal in ctc1 mutants by approximately 95%, indicating that the majority of ss telomeric DNA is associated with the chromosome terminus (Figure 2D, left panel).
To investigate whether telomeres in ctc1 mutants are subjected to increased recombination, we used t-circle amplification (TCA) (Zellinger et al., 2007) to look for evidence of extra-chromosomal telomeric circles (ECTC), a by-product of t-loop resolution. In this procedure, telomere sequences are amplified by phi29, a polymerase with strand displacement activity that generates high molecular weight ssDNA products from a circular template. As a positive control, TCA was performed on DNA from ku70 mutants previously shown to accumulate ECTCs (Zellinger et al., 2007). A high molecular weight DNA band was detected in both ku70 and ctc1 DNA samples, but not in wild type (Figure 3A). To verify the presence of ECTCs in ctc1 mutants, we employed the bubble trapping technique (Mesner et al., 2006), which relies on the ability of linear DNA fragments to enter the gel, while circular DNA cannot. A telomeric signal was detected in the well with DNA from ctc1 and ku70 mutants, but not with wild type (Figure 3B). These data confirm that ECTCs accumulate in the ctc1 background and argue that loss of CTC1 results in elevated rates of homologous recombination at telomeres. Altogether, these results indicate that the architecture of the chromosome terminus is perturbed in the absence of CTC1.
In Arabidopsis, telomeres shorter than 1 kb are prone to end-to-end chromosome fusions (Heacock et al., 2007). Since a substantial fraction of ctc1 telomeres dropped below this critical threshold, we looked for evidence of mitotic abnormalities. Anaphase bridges were scored in four individual ctc1-1 mutants and in their wild type siblings. As expected, there was no evidence of genome instability in wild type plants, but in all four ctc1-1 mutants a high fraction of mitotic cells (up to 39%) exhibited anaphase bridges (Figure 3C and Table S1). Many anaphases contained multiple bridged chromosomes as well as instances of unequal chromosome segregation (Figure 3C). FISH using a mixture of probes from nine subtelomeric regions produced signals in 20/23 anaphase bridges, indicating that the bridges represent end-to-end fusions (Table 1). FISH probes from eight chromosome ends were individually applied to chromosome preparations from a single ctc1-1 flower cluster. Signals from each probe were observed in anaphase bridges suggesting that all chromosome arms participated in chromosome fusions (Table 1).
Telomere fusion PCR confirmed end-to-end chromosome fusion. Abundant telomere fusion products were generated from ctc1-1 homozygous plants, but not from heterozygous or wild type siblings (Figure 3D and data not shown). Sequence analysis of 27 cloned fusion junctions failed to detect joining events involving direct fusion of telomere repeats. Instead, telomere-subtelomere fusions (14%) and subtelomere-subtelomere fusions (86%) were recovered (Figure 3D), which were characterized by extensive loss of subtelomere sequences (792 bp average loss). In contrast, in G9 tert mutants, telomere-subtelomere fusions are the most prevalent (78%), and the average loss of subtelomeric DNA sequences is only 290 bp (Heacock et al., 2004). Thus, chromosome ends are subjected to dramatic DNA loss prior to fusion in ctc1 mutants.
Since the rapid telomere uncapping phenotype associated with loss of AtCTC1 is remarkably similar to AtSTN1 deficiency (Song et al., 2008), we asked whether the two proteins act in the same genetic pathway for chromosome end protection. Plants heterozygous for ctc1-1 were crossed to stn1-1 heterozygotes and F1 progeny were self-pollinated to generate homozygous ctc1-1 stn1-1 mutants, and their ctc1-1 and stn1-1 single mutant siblings. The ctc1 stn1 double mutants were viable, and the severity of morphological defects was similar to the single mutants (Figure S4A). TRF analysis and PETRA revealed the same heterogeneous, shortened telomere profile in double mutants as in the single mutants (Figures 4A and S4B). Similarly, G-overhang signal intensity and the level of ECTC were comparable, implying that ctc1-1 stn1-1 double mutants did not undergo additional telomeric DNA depletion or increased telomere recombination (Figures 4B and S4C). Finally, the frequency of anaphase bridges was similar in double mutants and their ctc1 and stn1 siblings (Table S2). Altogether these findings indicate that AtCTC1 and AtSTN1 act in the same pathway for chromosome end protection.
We looked for evidence of a physical association between AtCTC1 and AtSTN1 proteins. Full length AtSTN1 and truncation fragments of AtCTC1 were expressed in rabbit reticulocyte lysate as T7-tagged proteins or radiolabeled with 35S methionine. Immunoprecipitation experiments showed no interaction between AtSTN1 and fragments A-CTC1 or D-CTC1. However, AtSTN1 bound the B-CTC1 and C-CTC1 fragments in reciprocal immunoprecipitation assays (Figure 4C). The STN1/C-CTC1 interaction was confirmed in a yeast two hybrid assay (data not shown). These data indicate that AtSTN1 and AtCTC1 directly interact in vitro and hence may also associate with each other in vivo.
TBLASTN and EST database searches revealed CTC1 homologs in a wide range of plant species, and searches using PSI-BLAST and HHpred uncovered putative CTC1 orthologs in many vertebrates (see supplemental material for details). Although the putative plant and animal orthologs exhibited considerable sequence divergence, a global profile-profile alignment indicated that the secondary structures had similarity throughout the length of the protein. Further analysis indicated that the C-terminal domain of human and Arabidopsis CTC1 shows homology to OB-fold regions from RPA orthologs, while the N-terminal domain may contain an OB-fold that is distantly related to OB2 from human POT1 (Figures 5A and S5).
Interestingly, the mammalian ortholog of CTC1 is identical to one subunit of Alpha Accessory Factor (AAF-132), while the second subunit of AAF (AAF-44, also known as OBFC1) corresponds to the mammalian ortholog of Stn1 (Casteel et al., 2009; Martin et al., 2007). AAF is a heterodimeric protein that was originally identified as a factor that stimulates Pol α-primase. It was subsequently shown to enhance Pol α-primase association with ssDNA, allowing the enzyme to prime and extend DNA in a reiterative fashion without falling off the DNA template (Goulian and Heard, 1990). Genes encoding the two subunits of AAF were identified recently and AAF-44 was predicted to contain OB-folds resembling those from RPA32 (Casteel et al., 2009).
To investigate whether the human CTC1 protein is important for telomere integrity, we examined the effect of knocking down CTC1 expression in human cells. HeLa and MCF7 cells were subject to two rounds of transfection with individual siRNAs and the level of CTC1 transcript was analyzed by quantitative real-time RT-PCR. Out of eight siRNAs tested, six routinely gave a 60-80% knockdown (Figures 5B and and6C,6C, data not shown). The effect of CTC1 knockdown was monitored after the cells had recovered from the dual transfection.
FACS analysis of DNA content revealed that CTC1 knockdown affected cell cycle progression in MCF7 cells, with cultures showing an accumulation of cells in G1 and a decrease in the S/G2 fraction (Figure S6A). Microscopy of DAPI stained cells revealed that CTC1 knockdown perturbed chromosome segregation. For HeLa cells, we observed an ~2-fold increase in the frequency with which interphase cells remained connected by chromatin bridges (Figures 5C, 5D and S6B). Although the incidence of chromatin bridges was lower in MCF7 cells, there was an increase in the number of cells with micronuclei (Figure S6C). These micronuclei probably reflect anaphase or interphase bridges that were later resolved (Hoffelder et al., 2004). We were unable to determine whether CTC1 knockdown caused an increase in anaphase bridges as the frequency of mitotic cells was too low. However, the cut-like phenotype with interphase bridges is similar to what was observed after POT1 knockdown in HeLa cells (Veldman et al., 2004), suggesting that like Arabidopsis CTC1, human CTC1 is needed to prevent chromosome fusions.
To determine whether the defects in chromosome segregation led to a DNA damage response, we looked for the appearance of γH2AX foci. Treatment with CTC1 siRNA caused an increase in foci in both HeLa and MCF7 cells. These foci were fewer in number and larger than the foci observed after UV irradiation. Moreover, they persisted for the duration of the knockdown whereas UV-induced foci were resolved after a few hours (data not shown). We looked for co-localization of γH2AX and TRF2 staining but this was not readily apparent (data not shown) suggesting that either the DNA damage was not telomeric or that disruption of CTC1 results in complete loss of the telomeric tract from a subset of telomeres. Overall, our results indicate that loss of human CTC1 causes a DNA damage response and genome instability.
To determine whether CTC1 knockdown has a direct effect on telomere structure, we used non-denaturing in-gel hybridization to examine the status of the G-overhang. CTC1 depletion caused a modest but consistent increase in ss G-strand DNA in both HeLa and MCF7 cells (Figure 6 and data not shown). In MCF7 cells, the G-strand signal increased by 33 to 41% relative to the non-silencing control siRNA (Figure 6). This increase was statistically significant. Treatment with Exonuclease I removed essentially all the G-strand signal from the control DNAs, but a small amount remained in the samples from CTC1 depleted cells (Fig. 6A). Thus, removal of CTC1 causes an increase in G-overhang length and may also result in internal regions of ssG-strand DNA.
Given the failure of the γH2AX foci to co-localize with TRF2 after CTC1 knockdown, we analyzed metaphase spreads to determine whether depletion of CTC1 lead to sporadic telomere loss. Metaphase spreads were prepared from siRNA-treated HeLa and 293T cells and hybridized with Cy3-labeled (TTAGGG)3 PNA probe. Subsequent analysis of individual chromosomes revealed an increase in signal free ends (Figure 5 E and F). This increase was statistically significant in 4 out of 6 trials, with the greatest frequency of signal free ends correlating with the deepest CTC1 knockdown (Figure S6D). We therefore conclude that like Arabidopsis CTC1, human CTC1 is required to maintain telomere integrity.
Although overall telomere architecture and the general mechanism of telomere replication are well conserved, telomere protein sequence and composition have evolved rapidly (Bianchi and Shore, 2008; Linger and Price, 2009). The resulting divergence has complicated telomere protein identification and it is currently unknown whether the full complement of dedicated telomere proteins has been established for any organism. It is also unclear whether additional telomere-specific factors are required to address the unique problems associated with replicating the DNA terminus. In this study we used a forward genetic approach to identify CTC1, a new telomere protein that is required for genome integrity in higher eukaryotes. The CTC1 gene is predicted to encode a large protein (142 kDa in Arabidopsis and 134.5 kDa in humans) that has orthologs dispersed widely throughout the plant and animal kingdoms. Both Arabidopsis and human CTC1 interact with STN1, an ortholog of S. cerevisiae Stn1 that was recently found at Arabidopsis and human telomeres (this study and (Casteel et al., 2009; Dejardin and Kingston, 2009; Song et al., 2008)). Moreover, we discovered that the mammalian CTC1/STN1 complex corresponds to the recently identified DNA polymerase Alpha Accessory Factor (AAF), previously shown to stimulate Pol α-primase (Casteel et al., 2009). Thus, CTC1 appears to be a specialized replication factor that is required for telomere protection and/or telomere replication.
In Arabidopsis, the phenotype of a ctc1 null mutant reflects rapid and catastrophic deprotection of all chromosome ends. Telomere tracts are grossly deregulated in both length and terminal architecture and are subjected to increased recombination and extensive loss of both telomeric and subtelomeric sequences prior to end-to-end fusion. The dramatic effect of CTC1 depletion contrasts with the gradual loss of telomeric DNA in tert mutants and the correspondingly later onset of developmental defects (Fitzgerald et al., 1999; Riha et al., 2001). It is striking that plants null for CTC1 are viable because in other model organisms loss of telomere capping proteins activates an ATM or ATR-mediated DNA damage checkpoint and is a lethal event (e.g. loss of CDC13, STN1 or TEN1 in budding yeast, STN1, TEN1 or POT1 in fission yeast, and TRF2 or POT1 in vertebrates (Churikov and Price, 2008; Grandin et al., 1997; Palm and de Lange, 2008). The extraordinary tolerance of plants to telomere uncapping may reflect a difference in pathways used to monitor genome integrity (Gutierrez, 2005) or the partial duplication of the Arabidopsis genome which permits some degree of aneuploidy. In addition, developmental plasticity may mitigate the consequences of genome instability by allowing healthy cells to assume the functions of their more severely compromised neighbors.
Depletion of the human CTC1 mRNA revealed a more modest, but significant role for this protein in promoting telomere integrity. Several cell lines exhibited hallmarks of genome instability such as chromatin bridges, micronuclei and γH2AX staining. Moreover, telomere architecture was perturbed with cells showing an increase in G-overhang signal and sporadic telomere loss. The milder phenotypes associated with HsCTC1 depletion relative to Arabidopsis may reflect the partial knockdown. Plants that are heterozygous for CTC1 show no deleterious phenotypes, thus only low levels of protein may be needed to maintain telomere integrity. This is the case for vertebrate POT1 as the knockdown causes a less severe phenotype than the full gene knockout (Churikov et al., 2006). It is also possible that the function of HsCTC1 only partially overlaps that of AtCTC1. In Arabidopsis, POT1 variants seem to be telomerase subunits rather than stable components of the telomere (C. Cifuentes-Rojas et al. in preparation)(Surovtseva et al., 2007). Thus, plant CTC1 may have evolved to function both in chromosome end protection and telomere replication. In contrast, mammalian CTC1 may function only in telomere replication.
How CTC1 promotes telomere integrity in higher eukaryotes is unknown, but important clues come from recent studies of AAF (HsCTC1/STN1) (Casteel et al). AAF-44 (HsSTN1) contains an OB-fold that is required for AAF to bind ssDNA and stimulate Pol α-primase activity. Thus, as in the budding yeast Cdc13/Stn1/Ten1 (CST) complex, the mammalian CTC1/STN1 complex binds ssDNA and provides a link to the lagging strand replication machinery. This connection also appears to be conserved in plants as AtCTC1 physically interacts with both AtSTN1 (this study) and the DNA pol α catalytic subunit (X. Song and D. Shippen, unpublished data). These findings raise the intriguing possibility that plant and mammalian CTC1 and STN1 are part of a CST complex that, like budding yeast CST, functions in telomere capping and/or coordination of G- and C-strand synthesis during telomere replication. If CTC1 functions in a CST-like complex we would expect higher eukaryotes to possess a Ten1-like protein. Indeed, a putative TEN1 ortholog has been identified in humans (F. Ishikawa, personal communication) and Arabidopsis (X. Song, K. Leehy and D. Shippen, unpublished data). Like its counterpart in budding yeast, the Arabidopsis TEN1 protein exhibits strong affinity for AtSTN1 in vitro.
The observation that both S. cerevisiae CST and mammalian CTC1/STN1 (AAF) modulate DNA pol α-primase is particularly striking. In yeast, both Cdc13 and Stn1 interact with Pol α subunits and are proposed to couple telomeric G- and C-strand synthesis (Grossi et al., 2004; Puglisi et al., 2008; Qi and Zakian, 2000). This coupling prevents accumulation of long G-strand overhangs following G-strand extension by telomerase or C-strand resection by nuclease. Previous studies of mammalian CTC1/STN1 (AAF) only explored Pol α-primase stimulation in vitro and did not investigate in vivo telomeric function or interactions with telomeric DNA (Casteel et al., 2009; Goulian and Heard, 1990). Thus, this work did not address whether CTC1/STN1 promotes general DNA replication or telomere replication. Our results reveal a clear role for CTC1/STN1 in telomere maintenance. However, we cannot rule out additional non-telomeric functions. Indeed, the non-telomeric γH2AX staining after CTC1 knockdown is consistent with a role in DNA replication or repair. One possibility is that CST acts as a specialized replication/repair factor that is needed to reinitiate DNA synthesis by DNA Pol α if a replication block causes uncoupling of polymerase and helicase activity at the replication fork (Heller and Marians, 2006; Yao and O'Donnell, 2009). Such a function might explain the residual exonuclease resistant G-strand signal after CTC1 depletion.
Many of the telomere defects observed after CTC1 depletion can be explained by defects in lagging strand replication either at the chromosome terminus or within the telomeric tract. For example, failure to fill-in the C-strand following telomerase action or C-strand resection would lead to long G-overhangs. Damage to the G-strand might, in turn, result in telomere loss and/or telomere fusions. Likewise, failure to re-initiate lagging strand synthesis after replication fork stalling could lead to loss of large stretches of telomeric DNA and signal free ends.
Given the role played by the S. cerevisiae CST complex, one attractive model for CTC1/STN1 function is that it serves to recruit Pol α-primase to the telomeric G-strand after telomerase action and/or C-strand processing. Pol α appears to be recruited to replication forks by Mcm10, which may in turn interact with the Cdc45/Mcm2-7/GINS replicative helicase (Warren et al., 2008). However, since the G-strand overhang cannot support a conventional replication fork, telomeres appear to require a specialized mechanism to recruit Pol α-primase for C-strand fill-in. Further studies are needed to test this model for CTC1/STN1 function. Additional work will also be required to determine the extent to which the telomeric function of CTC1/STN1 stems from its role in telomere replication versus a more passive function in G-overhang protection. Perhaps the balance between these activities will differ between organisms. For example, the Arabidopsis and S. cerevisiae complexes may function in both capacities, while the mammalian complex is specialized for telomere replication.
The ctc1-1 line was identified in the TILLING collection (Till et al., 2003). ctc1-2 and ctc1-3 lines were found in the SALK database (stock lines SALK_114032 and SALK_083165, respectively). Genotyping is described in supplemental methods. The stn1-1 line was previously described (Song et al., 2008). A genetic cross was performed between plants heterozygous for stn1-1 and for ctc1-1. For localization studies, a genomic copy of CTC1 was cloned into the pB7WGC2 Gateway vector (Karimi et al., 2005). The resulting N-terminal CFP fusion was transformed into wild type Arabidopsis (Surovtseva et al., 2007). Cloning, telomere assays and cytology, including FISH, are described in supplemental methods.
Map-based cloning was performed essentially as described (Lukowitz et al., 2000). Briefly, a mutant line (Columbia ecotype) was out-crossed to wild type Arabidopsis Landsberg erecta ecotype. F1 plants were self-propagated to F2. Pools of wild type and mutant plants were generated (~50 plants in each pool) for bulked segregant analysis. CIW5 and CIW6 markers were identified as markers linked to the mutation. 150 individual mutant plants were used to find recombinants in the genomic interval between CIW5 and CIW6. The region containing the mutation was mapped by creating and analyzing new markers. Primer sequences of mapping markers are available upon request.
HeLa, MCF7 or 293T cells were subject to two rounds of transfection 24 hrs apart using Lipofectamine2000, Oligofectamine or CaPO4. The final concentration of siRNA duplex (see supplemental methods for sequences) was 50 nM (Ambion) or 100 nM (EZBiolab) for each transfection. The efficiency of CTC1 knockdown was assessed using quantitative real-time RT-PCR with SYBR Green. Regions of CTC1 and GAPDH mRNAs were amplified for each RNA sample. The GAPDH mRNA level was used as an endogenous control to normalize the level of CTC1 mRNA for each sample. The normalized values were plotted relative to the mock-transfected control that was set to 100%. All reactions were performed in duplicate.
We thank Fuyuki Ishikawa for communicating results prior to publication. We also thank Jung Ro Lee for two hybrid analysis and Mary Chaiken, Geoffrey Kapler, Rachid Drissi, Erik Hendrickson, Paul Andreasen and Wayne Versaw for helpful comments and gifts of reagents. This work was supported by NIH grants GM065383 to D.E.S., GM041803 and GM73169 to C.M.P., and a Ruth L. Kirschstein National Research Service Award (GM800052) to J.C.L.
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.