|Home | About | Journals | Submit | Contact Us | Français|
Rap1 (repressor activator protein 1) is a conserved multifunctional protein initially identified as a transcriptional regulator of ribosomal protein genes in Saccharomyces cerevisiae but subsequently shown to play diverse functions at multiple chromosomal loci, including telomeres. The function of Rap1 appears to be evolutionarily plastic, especially in the budding yeast lineages. We report here our biochemical and molecular genetic characterizations of Candida albicans Rap1, which exhibits an unusual, miniaturized domain organization in comparison to the S. cerevisiae homologue. We show that in contrast to S. cerevisiae, C. albicans RAP1 is not essential for cell viability but is critical for maintaining normal telomere length and structure. The rap1 null mutant exhibits drastic telomere-length dysregulation and accumulates high levels of telomere circles, which can be largely attributed to aberrant recombination activities at telomeres. Analysis of combination mutants indicates that Rap1 and other telomere proteins mediate overlapping but nonredundant roles in telomere protection. Consistent with the telomere phenotypes of the mutant, C. albicans Rap1 is localized to telomeres in vivo and recognizes the unusual telomere repeat unit with high affinity and sequence specificity in vitro. The DNA-binding Myb domain of C. albicans Rap1 is sufficient to suppress most of the telomere aberrations observed in the null mutant. Notably, we were unable to detect specific binding of C. albicans Rap1 to gene promoters in vivo or in vitro, suggesting that its functions are more circumscribed in this organism. Our findings provide insights on the evolution and mechanistic plasticity of a widely conserved and functionally critical telomere component.
Multifunctional Rap1 (repressor activator protein 1) was first discovered in the budding yeast Saccharomyces cerevisiae as a positive transcriptional regulator of multiple growth-related genes such as the ribosomal protein genes (23). Other studies identified Rap1 as the major double-strand telomere repeat binding protein in S. cerevisiae and to be essential for maintaining telomere length and structural integrity (5, 10). Upon further analysis, Rap1 was recognized as a key component of the mating-type silencer and shown to be required for transcriptional silencing (7, 43, 44). These disparate observations raised fascinating questions concerning the mechanisms whereby a single protein participates in diverse functions at distinct chromosomal locations. The identification of Rap1 homologues in humans and the fission yeast resulted in yet more surprises (25, 29). Both homologues were shown to be telomere-associated proteins required for proper telomere length regulation. However, instead of binding DNA directly, human and Schizosaccharomyces pombe Rap1 proteins were recruited to telomeres through interaction with other telomere proteins such as TRF2 and Taz1. Moreover, there is no evidence that the human and S. pombe proteins are involved in transcriptional regulation. Thus, although the role of Rap1 in one particular cellular pathway appears to be conserved, its detailed molecular interactions are not.
Perhaps not surprisingly, the S. cerevisiae Rap1 has a complex domain architecture that befits its functional versatility. Near its N terminus is a BRCT domain, a presumed protein interaction domain whose target has not been identified. Located centrally is the DNA-binding domain, which uses a pair of Myb motifs to interact with DNA (18, 53). Interestingly, both fission yeast and human Rap1 contain just a single Myb motif, possibly accounting for their inability to bind DNA directly (25, 29). At the C-terminal end of Rap1 is a purely alpha-helical structure (RCT) that has been shown to mediate interactions with at least four other proteins required for proper telomere structure and function: Sir3, Sir4, Rif1, and Rif2 (12). Finally, a region between the DNA-binding domain and RCT has been ascribed a transcriptional activation function, although deletion of this region has little effect on the expression of some Rap1 targets (43).
Telomeres are specialized nucleoprotein structures that maintain the integrity of eukaryotic chromosomal termini by protecting them from fusion and recombination, and promoting their replication (for reviews, see references 13, 24, and 40). In most organisms, telomeric DNA consists of short repetitive sequences that are rich in G residues on the 3′ end-containing strand. These repeats are maintained by a ribonucleoprotein (RNP) known as telomerase, which acts as an unusual reverse transcriptase (for reviews, see references 3, 9, and 39). Both telomere binding proteins and telomerase are critical for the maintenance of telomere integrity through multiple cell divisions, which in turn is pivotal in supporting genome stability and promoting cellular life span. Interestingly, while the telomeres in many yeast (e.g., Zygomycota, Basidiomycota, and Eurotiomycetes) and metazoan species conform to the canonical TTAGGG repeat unit, the telomeres of the Saccharomycotina subphylum of budding yeast (including Saccharomyces, Kluyveromyces, Dabromyces, and Candida spp.) have been found to exhibit extraordinary sequence diversity (49). The length of the repeat unit ranges from 8 to 25 bp, and the repeat sequence can be rather degenerate. However, a plausible Rap1 homologue can be discerned in most of the genome databases (see below). How this protein can accommodate the variable telomere repeats in budding yeast is just one of the many unanswered questions provoked by the observation of telomere sequence divergence.
Our laboratory has utilized C. albicans as a model organism for understanding telomere structure and regulation (22, 46, 57). Unlike S. cerevisiae, this pathogenic fungus possesses a regular 23-bp telomere repeat unit. Internal to the terminal repeats are several shared subtelomeric elements, including Care-2 and the TLO genes, whose functional significance has not been elucidated (51). Remarkably, in contrast to most other Rap1 homologues, the C. albicans gene is considerably smaller and lacks the C-terminal domain (RCT) implicated in telomere regulation (6, 50). Also in contrast to S. cerevisiae, two previous reports have suggested that the C. albicans RAP1 is not essential for cell viability but may be involved in the regulation of pseudohyphal growth (6, 50). These studies did not address the potential role of C. albicans Rap1 in telomere regulation. We show here that, while not essential, C. albicans RAP1 is critical for maintaining normal telomere length and structure. Consistent with a direct regulatory role, the C. albicans Rap1 binds with high affinity and sequence specificity to the cognate telomere repeat in vitro and localizes to telomeres in vivo. Analysis of combination mutants suggests that Rap1 plays overlapping but nonredundant roles in protecting telomeres against aberrant recombination. Furthermore, we were unable to detect specific binding of C. albicans Rap1 to ribosomal protein genes in vitro or in vivo, suggesting that the protein is mostly dedicated to telomere regulation in this organism. We propose a model for Rap1 evolution to account for its diverse structure and function in different budding yeast.
The C. albicans strain BWP17 (ura3Δ::λimm434/ura3Δ::λimm434 his1::hisG/his1::hisG arg4::hisG/arg4::hisG) was used as the parental strain (55). Deletion rap1−/− strains were generated by subjecting BWP17 to two rounds of transformation and 5-fluoorotic acid selection using a rap1::hisG-URA3-hisG cassette (containing ~830 bp of RAP1 upstream and ~620 bp of downstream sequence) or by using a rap1::URA3-ARG1-URA3 (UAU1) cassette (containing the same RAP1 upstream and downstream sequences) (11, 16). The rap1 stn1, rap1 ten1, rap1 ku70, and rap1 tert combination deletion strains were derived from the respective single mutants by the use of the rap1::hisG-URA3-hisG (URA-Blaster) cassette or the rap1::URA3-ARG1-URA3 (UAU1) cassette. The ku70 single mutant was kindly provided by Lidia Chico and German Larriba (Universidad de Extramadura, Spain). For the reconstituted strains containing full-length or truncated RAP1 gene (with or without an N-terminal GSCP tag), the rap1−/− strain was transformed with pGEM-URA3-RAP1 or pGEM-URA3-GSCP-RAP1 linearized by cleavage with HindIII. The Tbf1-TAP and Cbf1-TAP-containing strains are gifts of the Whiteway lab (27). Strains are all passaged at 30°C either in solid or liquid YPD+uri medium (2% peptone, 1% yeast extract, 2% dextrose, and 80 μg of uridine/ml). Each streak corresponds to ~2 days and ~25 generations of growth.
A PCR fragment encompassing the entire RAP1 open reading frame was cloned in between the NotI and SacI sites of pGEM-URA3 to yield pGEM-URA3-RAP1. Site-specific mutagenesis was then used to create consecutive EcoRV and XhoI sites at the N terminus of RAP1. A triple affinity tag consisting of SBP, CBP, and protein A, followed by a Gly6 linker, was obtained by PCR (from pGEM-URA3-TEN1-GSCP using appropriate primers) and inserted into the EcoRV and XhoI site of pGEM-URA3-RAP1 to yield pGEM-URA3-GSCP-RAP1. Finally, truncated RAP1 genes were generated by PCR using primers containing XhoI and SacI site and used to substitute the corresponding fragment in pGEM-URA3-GSCP-RAP1.
Rap1 homologues from Candida and Saccharomyces spp. were identified from the NCBI (http://blast.ncbi.nlm.nih.gov/blast.cgi), Broad Institute (http://www.broad.mit.edu/annotation/genome/candida_group/blast.html), and Sanger Center (http://www.sanger.ac.uk/cgi-bin/blast/submitblast/c_dubliniensis) databases by BLAST or psi-BLAST searches. The multiple sequence alignment was generated by using the T-COFFEE server (http://www.igs.cnrs-mrs.fr/Tcoffee/tcoffee_cgi/index.cgi) and displayed by using Boxshade (http://www.ch.embnet.org/software/BOX_form.html).
Chromosomal DNAs were isolated from 3 to 5 ml of saturated C. albicans culture by the smash-and-grab method (19), digested with AluI and NlaIII, and fractionated in 0.6 to 0.8% agarose gels. For pulsed-field gel electrophoresis, the DNA fragments were resolved in a 1% agarose gel in 0.5× Tris-borate-EDTA at 250 V for 8 h using an initial switch time of 0.2 s and a final switch time of 0.9 s (CHEF DR II apparatus; Bio-Rad, Inc.). After transfer to nylon membranes, the telomere restriction fragments were detected as previously described by using an oligonucleotide probe that contains two copies of the Candida telomere repeat (45). The hybridization was performed at 50°C to 60°C. Signals from this and other studies involving radioactivity were all analyzed by using a PhosphorImager scanner and ImageQuant software (GE Healthcare, Inc.).
The two-dimensional gel analysis was performed according to the protocol of Brewer and Fangman, as modified by Cohen and Lavi (8). Briefly, the first dimension (0.5% agarose) was run at 0.5 V/cm for 16 h in the absence of ethidium bromide (EtBr). The gel was stained with 0.3 μg of EtBr/ml to visualize the size standards and bulk chromosomal DNA. Gel strips containing DNA in the 0.5- to 15-kb size range are excised and impregnated in a 1.2% agarose gel containing 0.3 μg of EtBr/ml. Electrophoresis was then performed in the orthogonal direction at 5 V/cm for 5 h. The DNAs in the gels were transferred to nylon membrane and probed with labeled oligonucleotides as described before (46).
C. albicans whole-cell extracts were prepared as previously described from 1 liter of YPD+uri culture (45). Binding reactions for unfractionated extracts contained 2 nM probe, a 1 μM concentration of nonspecific competitor oligonucleotide (5′-ACTTCTTGGTGTTGGGATGTCTA-3′/5′-TAGACATCCCAACACCAAGAAGT-3′), 2 μg of salmon sperm DNA, 1 μg of bovine serum albumin, and ~35 μg of extract in 18 μl of 10% glycerol-25 mM HEPES (pH 7.5)-5 mM MgCl2-0.1 mM EDTA-50 mM KCl (7). For recombinant Rap1, full-length RAP1-FLAG was cloned into the pSMT3 vector to enable the expression of a His6-SUMO-Rap1-Flag fusion protein. The CTG triplets encoding amino acids 76, 201, and 251 of the Rap1 protein were mutated to TCG, allowing wild-type protein to be expressed (42). After induction, extracts were prepared, and the fusion protein was purified by Ni-NTA chromatography. The fusion protein was cleaved by ULP1 protease and the Rap1-Flag fragment purified away from the His6-SUMO tag by a second round of Ni-NTA affinity chromatography. Binding reactions for purified Rap1-Flag contained 30 nM telomeric probe in 50 mM Tris-HCl (pH 8.0), 1 mM MgCl2, 1 mM spermidine, 1 mM dithiothreitol, and 10% glycerol. After incubation at 25°C for 20 min, the reaction mixtures were electrophoresed through a 6 to 8% nondenaturing polyacrylamide gel to resolve the free probe from the DNA-protein complex. To assess the binding of Rap1 to promoter regions, an upstream fragment of RPL11 was amplified by PCR using two primers (5′-TGTAGATGGATAAAGTAAGTTGATT-3′ and 5′-CCTTTCAAAACTTTTCAAACGAAAG-3′) and used as the probe. For comparison, a 108-bp fragment of RPL11 promoter region, in which the TBF1 and CBF1 binding sites were replaced by one telomere repeat, was used in binding reactions. This “hybrid” TELO DNA was prepared by annealing two complementary oligonucleotides (5′-aactacagtaactattatgcaacaattgaaagtatcctgtgtACTTCTTGGTGTACGGATGTCTAatgttgatttatttattcctgcacacattgagttttttttcta-3′ and 5′-tagaaaaaaaactcaatgtgtgcaggaataaataaatcaacatTAGACATCCGTACACCAAGAAGTacacaggatactttcaattgttgcataatagttactgtagtt-3′; telomere repeats are in uppercase).
Chromatin immunoprecipitation (IP) was performed using a combination of previously described protocols with some additional modifications (30, 56). Cells were fixed with 1% formaldehyde for 30 min at 30°C, and cross-linking was quenched with 125 mM glycine for 5 min at 30°C. Formaldehyde-fixed or untreated cells were resuspended in lysis buffer (50 mM HEPES [pH 7.5], 1 mM EDTA, 150 mM NaCl, and protease inhibitors) and broken by glass beads. The lysates were sonicated 10 times for 5 s each (constant duty cycle, 35 to 40% output) to shear DNAs to a mean length of ~600 bp. Extracts were adjusted to 1.6 mg of protein/ml in 600 μl of lysis buffer and then diluted with 600 μl of IP dilution buffer (0.01% sodium dodecyl sulfate [SDS], 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl [pH 8.0], 450 mM NaCl, and protease inhibitors). A total of 5% of each cell extract was set aside and used as the input sample. The remainder was subjected to IP using 20 μl of IgG-Sepharose beads at 4°C for 2 h. IP samples were washed for 5 min with rotation in the following buffers: once with buffer A (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl [pH 8.0], and 400 mM NaCl), four times with buffer B (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl [pH 8.0], and 600 mM NaCl), once with buffer C (0.25 M LiCl, 1% NP-40, 1% sodium deoxycholate, 1 mM EDTA, and 10 mM Tris-HCl [pH 8.0]), and once with Tris-EDTA. All of the wash buffers contained protease inhibitors. IP samples were eluted in 500 μl of 1% SDS-0.1 M NaHCO3, and cross-links were reversed at 65°C for 5 h. Samples were treated with RNase A and proteinase K, extracted with phenol-chloroform, precipitated with ethanol, and resuspended in 100 μl of water. The DNA samples were then applied to Hybond-N using a dot blot apparatus, and the membrane was probed with 32P-labeled CaC2 (5′-CATCCGTACACCAAGAAGTTAGACATCCGTACACCAAGAAGTTAGA-3′) corresponding to two copies of the C. albicans telomeric repeat. RPL11 or RPS25 promoter fragments were also used as probes. The fragments were amplified by PCR using appropriate primers (5′-TGTAGATGGATAAAGTAAGTTGATT-3′ and 5′-CCTTTCAAAACTTTTCAAACGAAAG-3′ for RPL11 and 5′-GCTCATCTGCAACTCTTTTTGCCAT-3′ and 5′-GTATCATTATTGAAATTAACAAAAC-3′ for RPS25). Hybridization was performed at 68°C to minimize background.
Two previous reports have noted the unusually small size of C. albicans Rap1 and its apparent loss of the well-conserved C-terminal domain (RCT) (6, 50). To determine how widespread this loss may be in evolution, we queried the NCBI and Broad Institute databases for all potential Saccharomyces and Candida homologues. Multiple sequence alignment suggests that besides C. albicans, the RCT of RAP1 is completely absent from Candida dubliniensis, Candida tropicalis, and Lodderomyces elongisporus (Fig. (Fig.1).1). Moreover, this region of C. parapsilosis Rap1 appears to have undergone a substantial expansion. Because RCT is also present in humans and S. pombe Rap1, this domain is probably part of the ancient budding yeast Rap1. We therefore propose that certain Candida species have experienced loss or remodeling of RCT, following the branching off these species from Candida lusitaniae and Candida guilliermondii (15). In contrast to the RCT divergence, each Rap1 in the Saccharomycotina subphylum possesses an apparent BRCT domain, as well as a duplicated Myb domain, suggesting that the interactions mediated by these domains are better conserved (Fig. (Fig.1B1B and data not shown).
To determine the function of C. albicans RAP1, we attempted to generate rap1 null strains by using either a UAU1- or a URA-Blaster-based disruption cassette (11, 16). The former cassette can yield a null strain through one transformation step, followed by two rounds of selection, while the latter cassette requires two sequential transformation and selection steps. Consistent with previous reports, we found that C. albicans rap1 null strains can be readily generated by either disruption cassette. Thus, in contrast to S. cerevisiae, the C. albicans RAP1 gene is not essential for cell viability. The null mutant strains, however, exhibited slower growth in both solid and liquid media, forming smaller colonies and having longer doubling times than the isogenic parental or the heterozygous strains (Fig. (Fig.22 and data not shown). Furthermore, the percentage of filamentous-form cells in rich medium is significantly higher for the null mutant (~5.5%) than for the parental strain (~0.5%). Although the reasons for this aberrant growth morphology, which was noted previously (6, 50), are not understood, similar aberrations have been described for other C. albicans DNA repair mutants (2, 28), suggesting a shared underlying mechanism.
Southern analysis revealed multiple interesting features of telomeric DNA in the rap1 null mutant. First, we observed extremely long and heterogeneous telomeres in multiple isolates of the null mutant (Fig. (Fig.3A,3A, B, and C). Long and heterogeneous telomeres were detected at the earliest time point after the derivation of the mutants (~100 generations) and were stably maintained for at least 150 generations thereafter (data not shown). In contrast to the parental BWP17, whose telomeres range in size from ca. 1 to 5 kb, the rap1 mutants possess extremely long (>20 kb) and short (<1 kb) telomeres, a finding consistent with the loss of the homeostatic mechanism that normally regulates telomere length. Second, significant fractions of telomere signals in the mutant samples were trapped in the wells, suggesting the existence of unusual DNA structures (Fig. (Fig.3B,3B, cf. lanes 1 and 2 with lanes 3 and 4). The reason for this is not understood, but similar structures appear to be present in the S. pombe rap1 mutant (36). Finally, two-dimensional gel electrophoretic analysis revealed high levels of extrachromosomal telomeric circles (t-circles) in the null mutant (Fig. (Fig.3D).3D). The elevated levels of t-circles in the mutant suggest that the low-molecular-weight telomeric DNA in the one-dimensional gel may be due to small circles rather than short telomeres. To test this notion, we compared the Southern hybridization patterns of untreated and restriction enzyme-digested DNA (Fig. 3B and C). The signal for low-molecular-weight telomeric DNA in the rap1 null mutant was stronger than the wild-type strain even for untreated samples, supporting the existence of small t-circles (Fig. (Fig.3B,3B, cf. lane 1 and 3; Fig. Fig.3C,3C, cf. the light and dark green trace). Then again, the low-molecular-weight signal in the mutant was further enhanced upon restriction enzyme digestion, consistent with the presence of short telomeres (Fig. (Fig.3B,3B, cf. lane 3 and 4; Fig. Fig.3C,3C, cf. the light green and purple trace). We conclude that the rap1 mutant contains an abundance of both small t-circles and short telomeres. The slow-growth, telomere heterogeneity, and t-circle accumulation phenotypes of the null strain were clearly all due to the disruption of RAP1; each abnormality was suppressed by the reintegration of a wild-type RAP1 allele (Fig. (Fig.22 and Fig. 3A and D). Interestingly, both the “reconstituted” strain (which contains one copy of RAP1) and the heterozygous mutant possessed slightly longer telomeres than the parental strain (Fig. (Fig.3A,3A, cf. lanes 1, 2, and 4). The C. albicans RAP1 is thus likely to be haploinsufficient with respect to telomere length regulation.
To characterize the contribution of individual domains of Rap1 to telomere regulation, we reintegrated full-length and three N-terminal truncation mutants of RAP1 back into the null strain and examined the lengths and structure of telomeres in the resulting strains. The three truncation mutants (Δ91N, Δ181N, and Δ234N) were designed to remove successively the very N terminus, the BRCT domain, and the linker preceding the Myb domain (Fig. (Fig.4A).4A). To facilitate biochemical analysis, the reintegrated alleles were each fused at its N terminus to a GSCP tag consisting of multiple-affinity epitopes (see Materials and Methods). Tagging of the full-length RAP1 gene appeared to have minimal effects on its function, as judged by the similar ability of the untagged and tagged gene to suppress the growth and telomere defects of the null strain (Fig. (Fig.22 and and3A).3A). As shown in Fig. Fig.4B,4B, the Δ91N mutant Rap1 was expressed at reduced levels in multiple reintegrant strains, whereas the Δ181N and Δ234N were expressed at slightly higher levels, indicating that the N-terminal domains are not essential for protein stability. Interestingly, telomere Southern and two-dimensional gel analysis indicate that all deletion mutants, including the Δ234N allele that retained just the Myb domain, were able to suppress partially the multiple telomere aberrations of the null strain. The telomeres in all of the truncation mutants were on average just 1.5 kb longer than those in the strain reconstituted with full-length RAP1 (Fig. (Fig.4C).4C). In addition, the levels of t-circles were reduced from ~5 to 6% of total telomeric DNA in the null strain to ~1 to 2% in the reconstituted strains (compare Fig. Fig.3D3D and and4D).4D). The deletion mutants also appeared to suppress the accumulation of complex telomere-containing structures that failed to enter the gel in Southern analysis (Fig. (Fig.4C,4C, arrowhead). Consistent with the suppression of telomere abnormalities, all of the truncation alleles restored the growth of the rap1 mutant to nearly that of wild-type cells (Fig. (Fig.2).2). Altogether, these findings suggest that the bulk of telomeric functions of Rap1 are mediated by the putative DNA-binding Myb domain.
To determine whether the Myb domain is in fact responsible for the telomere binding activity of Rap1, we examined the ability of the truncation mutants to interact with telomeric DNA both in vitro and in vivo. First, cell extracts were prepared from the reconstituted strains and subjected to gel electrophoretic mobility shift analysis using a double-stranded telomere oligonucleotide probe (Fig. (Fig.5A).5A). A predominant telomere DNA-specific binding activity was detected in all extracts. Consistent with being due to Rap1, the mobility of the protein-DNA complex varied as a function of the size of Rap1 protein. Notably, the levels of DNA-binding activity were similar in cell extracts derived from all of the reconstituted strains, arguing that tagging does not affect DNA binding and that the Myb domain alone is sufficient for full DNA-binding activity. Second, the ability of the Rap1 mutants to interact with telomeric DNA in vivo was examined by chromatin IP using IgG-Sepharose, which interacts with the protein A epitope of the GSCP tag (Fig. 5B and C). Again, the full-length and truncation mutants all exhibited tag-dependent and cross-link-dependent interaction with telomeric DNA in this assay, thus confirming their ability to localize to telomeres in vivo. Interestingly, the ChIP signals for the Δ91N mutant were ~50% of those from other strains, which is consistent with the slightly reduced protein level of this mutant (Fig. (Fig.4B4B and and5C).5C). We conclude that the Myb domain of C. albicans Rap1 is sufficient for telomere interaction both in vitro and in vivo.
The phenotypes of the rap1-null mutant resemble closely those of the ku70, stn1, and ten1 mutants; all four mutants exhibit elongated telomeres and elevated levels of t-circles (47a; L. Chico et al., unpublished data). Both Stn1 and Ten1 are subunits of a heterotrimeric protein complex (CST) that binds and protects telomere 3′ overhangs, whereas the Ku70-Ku80 complex is thought to encircle the junction between double and single stranded telomeric DNA (14, 17, 32). The phenotypic similarity between the stn1, ten1, ku70, and rap1 mutants thus suggest a functional connection between proteins that interact with double-stranded and single-stranded telomeres. To further investigate the mechanisms of Rap1 in relation to other telomere protection factors, we generated rap1 ku70, rap1 stn1, and rap1 ten1 combination mutants and compared their telomere dysfunction to those of the single mutants. As shown in Fig. Fig.6A,6A, all of the combination mutants exhibited long and heterogeneous telomeres resembling the single mutants. However, close analysis revealed some interesting quantitative differences (Fig. 6B and D). In the rap1, stn1, and ten1 single mutants, the peak of telomere signals were broad and centered ~ 8kb, whereas in the ku70 mutant, telomeres were somewhat less elongated. In contrast, the peaks of telomere signals in the double mutants were all larger and near the limit of resolution for standard agarose gels (Fig. 6B and D). Pulsed-field gels confirmed that the average telomeres of the combination mutants were substantially longer than the single mutants (Fig. (Fig.6C6C and data not shown). Moreover, all of the combination mutants accumulated higher levels of telomere-containing DNA that were trapped in the well than did the rap1 single mutant (Fig. (Fig.7A,7A, arrowhead). In both the single and double mutants, the kinetics of telomere elongation was quite rapid such that the telomeres have reached their maximal lengths ~100 generations after derivation of the mutant (Fig. (Fig.7B).7B). Similar to the telomere length aberrations, two-dimensional gel analysis indicates that the levels of t-circles in the double mutants were consistently higher than those in the corresponding single mutants (Fig. (Fig.8).8). The only exception was the rap1 stn1 mutant, which did not exhibit an increase in t-circles relative to the stn1 mutant. The possible significance of this is currently unclear. We conclude that Rap1 and other telomere protective factors mediate overlapping and nonredundant roles in regulating telomere lengths and structures.
We have shown earlier that the dramatic lengthening of telomeres in the ten1 mutant is due to both unregulated telomerase and recombination activities (47a). To determine whether Rap1 also regulates both activities, we attempted to generate a rap1 tert and a rap1 rad52 combination mutant. Interestingly, we were unable to delete RAD52 in a rap1 mutant using a UAU1-based cassette despite screening more than 30 Arg+ Ura+ colonies (11). This observation suggests that disruption of RAD52 and RAP1 may be synthetically lethal. On the other hand, several rap1 tert combination mutants can be readily constructed and these mutants exhibited telomeres that are slightly more elongated than the rap1 single mutant (Fig. 6A and E). This observation implies that aberrant recombination is primarily responsible for the telomere lengthening of the rap1 mutant. Curiously, whereas deletion of TERT (named for telomerase reverse transcriptase) in a wild-type strain does not induce t-circle formation, loss of TERT in a rap1 strain results in a further elevation in the levels of t-circles (Fig. (Fig.8).8). Thus, telomerase may play a minor role in suppressing telomere recombination such that the function is only revealed in the absence of other protective factors. An analogous function has been attributed to the S. pombe homologue Trt1 (47).
As noted before, the sequences of telomere repeat units in the Saccharomyces and Candida clades are exceptionally diverse, and yet each genome appears to contain a convincing Rap1 homologue, thus raising interesting questions with regard to the mechanisms of DNA recognition by this protein. Besides S. cerevisiae Rap1, which has been exhaustively studied, only the binding specificity of the Saccharomyces castellii, Saccharomyces dairensis, and Kluyveromyces lactis homologues have been examined to some degree using partially purified preparation (26, 52-54). We thus undertook a detailed characterization of the DNA-binding properties of C. albicans Rap1 to gain insights into the mechanisms of telomere recognition. Full-length C. albicans Rap1 with a C-terminal FLAG tag was expressed as a SUMO-fusion protein in E. coli. After removal of the SUMO tag and further purification, the Rap1-FLAG protein was subjected to a series of EMSA analysis to determine its binding affinity and sequence specificity (Fig. (Fig.9).9). As expected, the C. albicans Rap1 protein binds to the C. albicans telomere repeat with high affinity (Kd app = ~15 nM) (Fig. (Fig.9C).9C). The protein has clearly evolved preference for the cognate telomere repeat unit, because the purely GT repeat of the S. cerevisiae telomere and the more divergent C. pseudotropicalis repeat do not compete efficiently for binding (Fig. 9C and D). On the other hand, the more closely related C. tropicalis and C. maltosa telomere repeat do compete efficiently, indicating that the recognition property of Rap1 is not entirely species specific (Fig. 9C and D).
To define further the sequence specificity of the C. albicans Rap1 protein, we used a series of mutant oligonucleotides, each bearing two or three nucleotide substitutions, as competitors in the binding assays (Fig. (Fig.9B).9B). As shown in Fig. 9E and F, two GT-rich pentanucleotide elements within the telomere repeat unit (GGTGT and GGATG) appear to be most important for sequence-dependent recognition. Substitutions that overlap with these sequence elements all had moderate or severe effects on the abilities of the mutant oligonucleotides to serve as competitors. Notably, the result of this analysis is in excellent agreement with that using heterologous telomere repeats (Fig. 9C and D). For example, the differences between the C. albicans, C. tropicalis, and C. maltosa telomere repeat lie entirely outside of the two pentanucleotide elements, thus accounting for the abilities of the C. tropicalis and C. maltosa repeat to compete efficiently in the binding assays. To assess other features of the telomere sequence necessary for high-affinity interaction with Rap1, we investigated three other permutations of the telomere repeat unit (Fig. 10A). One permutation (“b”) reversed the order of the two elements and increased their spacing by 9 bp. This oligonucleotide did not interact with Rap1 with appreciable affinity (Fig. 10B). Two other permutations (“c” and “d”) retained the spatial relationship between the elements but place one of the elements at the edge of the oligonucleotide. Both of these permutations exhibit reduced binding to Rap1, with oligonucleotide “d” showing a more severe defect (Fig. 10B and C). We conclude that the spatial relationship between the pentanucleotide elements is important for Rap1 binding. In addition, even though nucleotides surrounding of the pentanucleotide elements appear not to be involved in sequence-dependent recognition, they may contribute to the interaction affinity.
Besides our observation of robust telomere binding, the C. albicans Rap1 has previously been reported to bind an RPG box located in the promoter of the S. cerevisiae ENO1 gene, a finding consistent with a potential role in transcriptional regulation (50). However, this sequence element, which exhibits little similarity to the C. albicans telomere repeat unit, failed to interact with our purified C. albicans Rap1 (data not shown). Interestingly, a recent study by Hogue et al. proposes that the transcriptional activation function of Saccharomyces Rap1 at ribosome-related genes is relatively modern and is probably not conserved in Candida species (20). These investigators argue instead that the Candida ribosomal genes are activated by two other transcription factors, Tbf1 and Cbf1. To reinvestigate this issue, we analyzed the interaction between C. albicans Rap1 and the RPL11 or RPS25 promoter. Both S. cerevisiae RPL11 and RPS25 are known to be regulated by Rap1, but their homologues in C. albicans are proposed to be under the control of Tbf1 and Cbf1 (20). We were unable to detect a DNA-protein complex between Rap1 and the RPL11 promoter (Fig. 11A, right panel), even though a hybrid DNA probe containing a telomere repeat unit inserted into a subfragment of the promoter yielded a clear complex (Fig. 11A, left panel). We then attempted to detect localization of Rap1 upstream of the RPL11 and RPS25 genes by ChIP. Despite robust signals for telomeric DNA, the levels of RPL11 and RPS25 promoters in the cross-linked Rap1 IP samples were extremely low, suggesting that Rap1 does not associate with these promoters in vivo (Fig. 11B). In contrast, we were able to demonstrate clear localization of C. albicans Tbf1 and Cbf1 to the RPL11 promoter (Fig. 11C). Thus, our findings are indeed consistent with a more dedicated role for Rap1 at telomeres in C. albicans.
We have shown that despite its unusual, “miniaturized” domain organization, the C. albicans Rap1 has important functions in suppressing aberrant telomere elongation and t-circle formation. Both phenotypes appear to be due to abnormal recombination at telomeres because the rap1 tert combination mutant exhibits the same features. Our findings thus demonstrate for the first time a novel function for Rap1 in suppressing telomere recombination. Notably, K. lactis mutants with aberrant telomere repeats that cannot be bound by Rap1 have also been reported to experience high levels of telomere recombination (4). It is thus tempting to suggest that Rap1 mediates a conserved function in regulating abnormal telomere recombination in C. albicans and K. lactis. Furthermore, analysis of other telomere protein mutants suggests that this regulatory activity of Rap1 is shared by other telomere proteins such as Ku70, Stn1 and Ten1 (47a; Chico et al., unpublished). Although there appears to be some functional overlap between these telomere protection factors, they are clearly not redundant because combination mutants almost invariably have exacerbated phenotypes relative to single mutants (Fig. (Fig.6,6, ,7,7, and and8).8). It is also worth noting that these phenotypes are reminiscent of ALT cancer cells, which evidently maintain telomeres through recombination (37, 38). Altogether, our data suggest that disrupting either the double-strand or the single-strand telomere nucleoprotein complex can lead to an ALT-like state and that the C. albicans mutants may be useful model systems for understanding the molecular basis of ALT.
Another interesting defect of C. albicans rap1 mutant is the accumulation of aberrant telomere-containing DNA structures that cannot enter standard agarose gels. This phenotype appears to be manifested by the S. pombe rap1 mutant as well (36). Interestingly, we observed a slight increase in the level of single-stranded telomere DNA by in-gel hybridization, but the single-stranded DNA signals are mostly resistant to ExoI digestion (data not shown). It is tempting to speculate that this phenotype may be another manifestation of the same aberrant telomere DNA structures.
At S. cerevisiae telomeres, Rap1 has been demonstrated to serve at least two critical functions: suppressing telomere-telomere fusion (t-t fusion) and regulating telomerase activity through a “protein-counting” system (31, 33, 34). Our initial attempts at detecting t-t fusions in C. albicans wild-type and rap1 null strains using the PCR strategies developed previously for S. cerevisiae were unsuccessful (data not shown) (35, 41). In addition, as mentioned before, we were unable to obtain strains missing both RAP1 and RAD52. Consequently, we are at present unable to determine whether the two well-established functions of S. cerevisiae Rap1 are also performed by the C. albicans homologue. Further studies will be necessary to address these issues. However, it is noteworthy that both functions of S. cerevisiae Rap1 are mediated in part by its RCT domain, which is missing in the C. albicans homologue. Thus, it would not surprising if some functional disparity has arisen between the two Rap1 proteins.
A particularly intriguing observation that emerged from our analysis of combination mutants concerns the possible function of telomerase in regulating telomere recombination. In particular, while the tert single mutant exhibited telomere shortening and no detectable increase in t-circles, the rap1 tert combination mutant evidently manifested a more severe telomere lengthening and t-circle accumulation phenotype than the rap1 single mutant (Fig. (Fig.6,6, ,7,7, and and8).8). Thus, it is tempting to speculate that telomerase may play a minor and partly redundant role in suppressing telomere recombination. This possibility is consistent with the conclusion of a recent study in S. pombe (47). Moreover, we have previously demonstrated an enzyme activity-independent function of telomerase in telomere protection, which may be related mechanistically to its function in regulating recombination (21).
Previous analysis of DNA-binding by the Rap1 proteins from several Saccharomyces species has led to the proposal that the two Myb domains recognize two G-rich sequence elements within the diverse telomere repeat units of budding yeast (52-54). Variable distances between the two core elements of different binding sites are accommodated due to a flexible linker between the Myb domains. Moreover, crystal structures of Rap1 bound to different target sites demonstrate that the protein can also use subtle side chain movements at the protein-DNA interface to recognize diverse sequence (48). Thus, the challenge posed by the diverse telomere repeat units in budding yeast could be met by Rap1 through a variety of means without major changes in the DNA-binding residues. A corollary of this hypothesis is that all of the Candida and Saccharomyces Rap1 may have very similar and flexible DNA-binding properties. This notion is supported by multiple sequence alignment of the Rap1 Myb domain; amino acid residues implicated in contacting DNA bases are mostly conserved between Saccharomyces and Candida Rap1 (Fig. (Fig.1B).1B). Also consistent with this notion is our finding that the critical sequence determinants for C. albicans Rap1 binding (i.e., the two pentanucleotide elements [GGTGT and GGATG]) are shared by most telomeres of the Candida clade. On the other hand, some species specificity has clearly emerged during evolution such that the S. cerevisiae and C. pseudotropicalis telomere repeats interact with the C. albicans protein with reduced affinity. In this regard, it would be especially interesting to determine whether the C. pseudotropicalis Rap1 has acquired amino acid changes in order to recognize preferentially the corresponding telomere repeat, which contains a variant pentanucleotide element (GGATT rather than GGATG; see Fig. Fig.9B).9B). Clearly, more studies will be necessary to arrive at a detailed understanding of the molecular basis of DNA-binding by Candida Rap1.
The dispensability of C. albicans RAP1 for cell viability is in accord with previous observations and consistent with current views on the essential role of RAP1 in S. cerevisiae. Notably, an S. cerevisiae strain with “humanized” telomeres that fail to bind Rap1 remains viable, indicating that the telomere function of this gene is not essential for S. cerevisiae cell viability (1). Instead, the transcriptional activation function of Rap1 may account for its indispensability. Thus, the nonessential nature of C. albicans RAP1 can be rationalized by a more limited role in transcriptional regulation. Interestingly, a recent study suggests that much of the transcriptional activation function of Rap1 evolved recently, after the Saccharomyces and Kluyveromyces species branched off from other yeast (20). In the Kluyveromyces and Saccharomyces genera, Rap1 acquired a new function in upregulating the transcription of many growth related genes, including the ribosomal protein and RNA genes. In contrast, the equivalent function in C. albicans and other budding yeast is evidently mediated by two other transcription factors named Tbf1 and Cbf1. Indeed, we were unable to detect an interaction between C. albicans Rap1 and two ribosomal protein gene promoters (Fig. (Fig.11).11). Moreover, genome-wide localization analysis of C. albicans Rap1 revealed a much more limited presence of this protein relative to its Saccharomyces counterpart, which is consistent with a more dedicated role of C. albicans Rap1 in telomere regulation (H. Lavoie et al., personal communication). Thus, the different range of in vivo function for Rap1 can account for the differential effect of RAP1 deletion on cell viability.
Both the telomere and the transcription-related functions of Rap1 have evidently experienced substantial evolutionary changes in budding yeast. It is tempting to speculate that these changes are both consequences of the unusual telomere repeat divergence in these organisms. A possible series of past events that can rationalize Rap1 evolution is presented in Fig. Fig.12.12. First, we hypothesize that in the ancestral budding yeast, Rap1 was tethered to telomeres through an interaction between its RCT domain and TRF2 (as in humans) and mediated a crucial telomere protection function. In accordance with earlier proposals (29, 49), we imagine that multiple telomerase RNA template mutations were somehow fixed in descendants of the yeast. The corresponding telomere sequence alterations in turn caused a major remodeling of the telomere nucleoprotein complex. This remodeling entails both the loss of many canonical telomere binding proteins including TRF2 and the acquisition of a direct-telomere DNA-binding activity by Rap1. Possibly, an accidental duplication of the Myb domain of Rap1 conferred the protein with enhanced potential for DNA binding. Because the RCT of Rap1 was no longer needed for its telomere localization, there was less selection pressure for its retention. The RCT was also in principle free to evolve new interactions and binding partners that might confer selective advantage to the organism. Thus, both the loss of RCT and the acquisition of new activities by RCT (i.e., the Rif1, Rif2, Sir3, and Sir4-interaction activity and the transcription activity of Rap1) in different yeast lineages could be considered natural consequences of the relaxing of constraints on RCT function.
We thank Hugo Lavoie, Malcolm Whiteway, Lidia Chico, and German Larriba for strains and for sharing results prior to publication and Bill Holloman for the use of a CHEF gel apparatus.
This study was supported by NIH (GM-069507) and the STARR Cancer Consortium.
Published ahead of print on 14 December 2009.