Genes & Genetic Systems
Online ISSN : 1880-5779
Print ISSN : 1341-7568
ISSN-L : 1341-7568
Special reviews
Telomere-binding factors in the regulation of DNA replication
Hisao Masai Yutaka KanohKenji MoriyamaSatoshi YamazakiNaoko YoshizawaSeiji Matsumoto
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML

2017 Volume 92 Issue 3 Pages 119-125

Details
ABSTRACT

Recent studies have indicated new roles for telomere-binding factors in the regulation of DNA replication, not only at the telomeres but also at the arm regions of the chromosome. Among these factors, Rif1, a conserved protein originally identified in yeasts as a telomere regulator, plays a major role in the spatiotemporal regulation of DNA replication during S phase. Its ability to interact with phosphatases and to create specific higher-order chromatin structures is central to the mechanism by which Rif1 exerts this function. In this review, we discuss recent progress in elucidating the roles of Rif1 and other telomere-binding factors in the regulation of chromosome events occurring at locations other than telomeres.

INTRODUCTION

Telomeres, special chromatin structures at the ends of linear chromosomes, serve two major purposes. First, at each cycle of DNA replication, telomere length is shortened. To maintain telomere length, telomere repeat unit sequences are added by RNA-dependent DNA synthesis, which is particularly important in stem cells. Second, telomere ends need to be protected from attack by the DSB repair/checkpoint machinery. Telomeres protect the chromosome ends by suppressing end-processing and checkpoint reactions (Blackburn et al., 2015). At telomeres, a conserved protein complex, shelterin, is assembled. Shelterin communicates with telomerase and the checkpoint machinery to protect or “cap” the chromosome ends (de Lange, 2005).

Rif1 (Rap1-interacting factor 1) was originally discovered as a factor binding to Rap1, a budding yeast shelterin component (Hardy et al., 1992; Mattarocci et al., 2016), and later Junko Kanoh and Fuyuki Ishikawa also identified rif1 in fission yeast (Kanoh and Ishikawa, 2001). Telomeres are elongated in the rif1 null mutant in both budding and fission yeast cells, indicating that Rif1 negatively regulates telomere elongation. Rif1 is conserved in higher eukaryotes, but its association with telomeres has not been clearly shown (Silverman et al., 2004; Xu and Blackburn, 2004). Instead, human Rif1 plays roles in DNA damage responses (Buonomo et al., 2009; Kumar et al., 2012; Chapman et al., 2013; Di Virgilio et al., 2013; Escribano-Díaz et al., 2013; Feng et al., 2013; Zimmermann et al., 2013). A role for Rif1 in cytokinesis was also recently reported (Hengeveld et al., 2015; Zaaijer et al., 2016).

A more conserved role for Rif1 in the regulation of DNA replication was reported. In both fission yeast and mammalian cells, spatial and temporal regulation of DNA replication was almost completely abrogated in rif1∆ cells (Cornacchia et al., 2012; Hayano et al., 2012; Yamazaki et al., 2012, 2013). Taz1 is another telomere-binding protein in fission yeast (an ortholog of human TRF1/TRF2 with a DNA-binding Myb domain; Cooper et al., 1997; Deng et al., 2015). Telomeres are elongated in taz1∆ cells. In fission yeast, Rif1 is recruited to telomeres through an interaction with Taz1, but not with Rap1 (Kanoh and Ishikawa, 2001). Taz1 was also reported to suppress the firing of a subset of late origins in fission yeast cells (Tazumi et al., 2012).

In this short review, we will summarize recent findings on the link between telomere-binding proteins and the regulation of replication origin firing, and discuss potential mechanisms by which telomere-binding proteins exert their effect on the spatiotemporal regulation of DNA replication.

Rif1 regulates DSB repair

Rif1 was originally discovered as a telomere-binding factor. It binds to Rap1 in budding yeast (Hardy et al., 1992), whereas it binds to Taz1 in fission yeast (Kanoh and Ishikawa, 2001). However, in mammalian cells, it does not appear to bind to telomeres, but functions in the repair of DSBs (Silverman et al., 2004; Xu and Blackburn, 2004; Buonomo et al., 2009). Rif1 is recruited to DSB sites through recognizing phosphorylated 53BP1 (the N-terminal phospho-SQ/TQ domain), and antagonizes its 5’→3’ resection activity, thus inhibiting homologous recombination-dependent repair and stimulating non-homologous end joining (Chapman et al., 2013; Escribano-Díaz et al., 2013; Feng et al., 2013). Rif1 accumulation at DSBs is strongly antagonized by BRCA1 and its interacting partner CtIP. Rif1(–/–) mice are severely compromised for 53BP1-dependent class switch recombination and fusion of dysfunctional telomeres (Di Virgilio et al., 2013; Zimmermann et al., 2013).

In budding yeast, Rif1 is also recruited to DSBs, but, in contrast to mammals, limits the access of Rad9, an ortholog of 53BP1, to DSBs and promotes the resection of DSBs, facilitating homologous recombination-dependent repair (Martina et al., 2014).

Identification of rif1+ as a bypass suppressor of hsk1 (Cdc7), the major regulator of replication origin firing

Cdc7 (Dbf4-dependent kinase; DDK) is a conserved kinase that triggers the initiation of DNA replication at each origin. It is conserved from yeasts to human, and it is now widely accepted that the Cdc7-Dbf4 kinsae complex plays a central role in the initiation of DNA replication in the eukaryotic kingdom (Masai et al., 2010). In fission yeast, hsk1+ and dfp1+/him1+ encode, respectively, the Cdc7 and Dbf4 (activation subunit of Cdc7 catalytic subunit) homologs (Masai et al., 1995; Brown and Kelly, 1998; Takeda et al., 1999). We reported that hsk1, although essential for growth under normal growth conditions, becomes dispensable for growth under certain genetic and growth conditions (Hayano et al., 2011; Matsumoto et al., 2011). Most notably, replication checkpoint mutations, such as cds1∆ and mrc1∆, bypass the requirement of Hsk1 kinase for growth; hsk1cds1∆ and hsk1mrc1∆ are viable. In these mutants, checkpoint-dependent late origin suppression is abrogated and replication potential is generally high. We also noted that hsk1∆ cells are viable at a high temperature under which late/dormant origins are weakly fired (Matsumoto et al., 2011). Thus, we surmised that a search for mutations that enable hsk1∆ to grow would uncover new factors regulating the origin firing program. This screening led to the identification of rif1+, whose mutation efficiently restored the growth of hsk1-89 and hsk1∆ (Hayano et al., 2012).

Rif1 suppresses a substantial fraction of late-firing origins in fission yeast cells

We then examined replication timing in rif1∆ cells (Hayano et al., 2012). We discovered that late/dormant origins are extensively activated in rif1∆ cells (Fig. 1). One hundred eighty-nine out of 637 late/dormant origins in the wild-type cells are activated early in rif1∆ cells. Deregulation of late/dormant origins is observed even during the unperturbed S phase. On the other hand, substantial numbers of early-firing origins are suppressed in rif1∆ cells. This may reflect unknown functions of Rif1 in regulating early-firing origins. Alternatively, it may be due to competition for limited initiation factors that have been recruited to the late-firing origins in rif1∆ cells. These results supported our earlier prediction that abrogation of the replication timing regulation would cause bypass of hsk1 function for initiation of DNA replication. It was later shown that rif1∆ partially suppresses cdc7ts in budding yeast as well (Peace et al., 2014).

Fig. 1.

Deregulation of replication over a long distance following the loss of a single Rif1-binding site in fission yeast cells. Dark blue, Rif1 binding signal; red, pre-replicative complex signal; green, BrdU incorporation in cells released from M phase arrest in the presence of HU in various genetic backgrounds. Signals at 4,160–4,360 kb of fission yeast chromosome II are shown. In Rif1BSII:4255 mut cells, the Rif1 binding signal at II:4255 is completely absent (not shown). The bracket indicates the area where DNA replication is affected in this mutant. Note that the BrdU incorporation at II:4255 is not affected in Rif1BSI:2663 mut cells.

Taz1 also regulates subsets of late/dormant origins in fission yeast cells

Taz1 is known to bind to fission yeast telomeric DNA in a sequence-specific manner (Spink et al., 2000). Hisao Masukata’s group identified a DNA element (RTC, for replication timing control) near a late/dormant origin, AT2088/ars745, that confers the late-firing property on an early-firing origin when placed nearby. They dissected RTC and identified the essential sequence, which contains two tandem repeats of a telomeric repeat. They showed that Taz1 binds to the telomeric repeats near the internal late origins. Genome-wide analyses showed that nearly half of the late origins, including those in subtelomeres, are regulated by Taz1. They further demonstrated that the presence of two tandem copies of the RTC essential sequence (each containing two copies of telomeric repeats) is sufficient for recruitment of Taz1 and for repression of nearby origins (Tazumi et al., 2012). It was also shown that Taz1-mediated origin suppression depends on Rif1. Thus, in fission yeast, late origin firing is regulated by the Taz1-Rif1 pathway and by a pathway involving only Rif1. It is of interest that Taz1 is required for recruitment of Rif1 to telomeres. In contrast, Rif1 recruitment to the chromosome arms does not require Taz1 (Hayano et al., 2012). Thus, an interesting possibility is that Taz1 at the telomere facilitates the physical and functional communication between telomeres and the telomeric repeats on the arms, which may lead to the suppression of late/dormant origins on the arms through Rif1 protein. At the moment, it is not known whether the Taz1 homologs TRF1 and TRF2 play a similar role in replication timing regulation in mammalian cells.

Rif1 regulates genome-wide replication timing in mammalian cells

Analyses of mammalian Rif1 revealed that the ability of Rif1 to regulate replication timing is also conserved in mammalian cells. In Rif1-depleted cells, the genome-wide replication timing domain profile was dramatically altered in both human and mouse cells (Cornacchia et al., 2012; Yamazaki et al., 2012, 2013). Rif1 binds to chromatin at late M to early G1 and this binding is DNaseI-resistant. Rif1 is bound to highly insoluble nuclear structures, presumably the nuclear matrix. It is bound to chromatin throughout S phase, but dissociates at G2/M. Rif1 is enriched at the nuclear periphery and in the region surrounding the nucleoli. This is reminiscent of the distribution of mid-S replication foci. Indeed, the absence of Rif1 results in selective loss of the mid-S replication foci pattern. This suggests the involvement of Rif1 in establishing the spatiotemporal chromatin domain specifying mid-S replication. Interestingly, the sizes of chromatin loops associated with the nuclear periphery significantly increased upon Rif1 depletion in human cells (Yamazaki et al., 2012). On the basis of these results, it was proposed that Rif1 promotes the formation of higher-order chromatin structure that is related to the mid-S replication timing domain by facilitating chromatin loop formation at the nuclear periphery (see below).

Recruitment of phosphatase by Rif1

Comparison of the amino acid sequences of Rif1 from various species revealed the presence of the conserved SILK/RVxF, the PP1 phosphatase interaction motif (Sreesankar et al., 2012). This prompted studies on the potential role of a phosphatase in origin regulation (Davé et al., 2014; Hiraga et al., 2014; Mattarocci et al., 2014). It was found that Rif1 recruits PP1 in a SILK/RVxF-dependent manner, and the recruited PP1 counteracts phosphorylation events conducted by Cdc7 kinase, most notably that of Mcm2 and Mcm4. This prevents the firing of replication origins, causing late replication of nearby origins (Fig. 2). Recently, it was reported that human Rif1 also recruits PP1, and this recruitment inhibits the origin activation process by reducing the phosphorylation of Mcm that is required for initiation (Hiraga et al., 2017; Alver et al., 2017), as found in yeast. Unexpectedly, Rif1-PP1 protects ORC1 protein from untimely phosphorylation and consequent degradation by the proteasome. Depletion of Rif1 led to destabilization of ORC1, thus reducing pre-replicative complex (pre-RC) formation (Hiraga et al., 2017).

Fig. 2.

A model for the action of fission yeast Rif1. Rif1 binds to intergenic G4 structures to generate a replication-suppressive chromatin domain near the nuclear periphery. Rif1 recruits phosphatase (PP1), which antagonizes the phosphorylation events executed by Cdc7 kinase (DDK). Chromatin compartments generated by Rif1-mediated chromatin loops define mid-S/late replication domains. Gray crescents indicate nuclear scaffold/matrix structures that are associated with nuclear membranes (not shown).

Chromatin binding of Rif1: genome-wide analyses of Rif1-binding sites

In fission yeast, Rif1 binds to chromatin at early G1. The binding increases during G1, and decreases during S phase. Initial ChIP-chip analyses indicated 155 Rif1-binding sequences on the entire complement of fission yeast chromosome arms, in addition to strong and abundant binding at the telomeres (Hayano et al., 2012), but more recent ChIP-Seq analyses revealed the presence of 35 efficient Rif1-binding sites (Rif1BSs) on the arms (Kanoh et al., 2015). They are located in the intergenic segments, but do not exactly overlap with pre-RC locations or the promoters. There is only one Rif1BS on chromosome III, which is largely early-replicating. Bioinformatics analyses indicated that Rif1BSs are located closer to the late/dormant origins, especially to those that become early-firing in rif1∆ cells, than to early-firing origins, suggesting that Rif1 bound close to the late/dormant origins somehow prevents them from firing early in S phase.

Genome-wide analyses of Rif1 in mouse embryonic stem (ES) cells showed that Rif1 binds to large domains which coincide with the late-replicating domain and lamin B1-assoicated domain (Foti et al., 2016). This is consistent with the result from fission yeast, supporting the notion that chromatin-bound Rif1 suppresses origin firing.

Rif1 is recruited to chromatin through recognizing the G-quadruplex

Sequence analyses of Rif1BSs on the fission yeast genome led to the identification of Rif1CS, a common consensus sequence (CNANGTGGGGG). In many Rif1BSs, notably in the strong Rif1-binding sites, two (or more) copies of Rif1CS are present. The two copies are arranged in head-to-tail orientation in 75% of these cases. Mutation of the Rif1CS on a Rif1BS resulted in the loss of Rif1 binding and local deregulation of origin firing, indicating the functional significance of this sequence in recognition by Rif1 in the cells (Fig. 1; Kanoh et al., 2015).

These results indicate that Rif1 may directly recognize the Rif1BS in a sequence-specific manner. However, purified Rif1 binds to double-stranded DNA in a non-sequence-specific manner. A typical Rif1BS contains not only the G-rich Rif1CS but also other G-tracts containing three or four stretches of guanines. Guanine-rich sequences are known to form non-B DNA structures including the G-quadruplex (hereafter referred to as G4). Indeed, the G-rich DNA strand derived from Rif1BS was shown to generate a higher-order DNA structure characteristic of the G4. Mutations that abrogated the chromatin binding of Rif1 disabled G4 formation in vitro, suggesting that the ability of the Rif1BS to form a G4 structure is important for Rif1 to bind to its targets. Indeed, upon heat denaturation and renaturation of duplex Rif1BS DNA in the presence of KCl, it adopts a distinct DNA structure, which is selectively bound by Rif1 in vitro. Rif1 from mammalian cells also selectively binds to G4 structures in vitro, indicating that G4 binding is evolutionarily conserved for Rif1 (Moriyama et al., submitted).

Rif1 may regulate replication timing through modulating chromatin architecture

The finding that Rif1 depletion affects chromatin loop sizes in human cells suggests the possibility that Rif1 regulates higher-order chromatin structure by regulating chromatin association (Yamazaki et al., 2012, 2013). In fission yeast, the loss of Rif1 binding at a single Rif1BS, through a point mutation that disrupted G4 formation, affected origin firing over nearly 100 kb (Kanoh et al., 2015). This long-range effect on DNA replication supports the idea that Rif1 exerts its inhibitory effect in a three-dimensional manner. Biochemical characterization of fission yeast and human Rif1 showed that it can simultaneously bind to multiple DNA molecules and form multimers (Moriyama et al. and Masai et al., unpublished data). Multimer formation was also previously reported for budding yeast Rif1. Interaction of Rif1 with multiple molecules of target DNA through multimerization may provide the molecular basis for Rif1-mediated chromatin loop formation.

Chromosome conformation capture sequencing (4C-Seq) analyses of Rif1 knockout ES cells provided direct evidence for Rif1’s role in chromatin organization (Foti et al., 2016). It was shown that Rif1 knockout results in increased interaction among replication timing domains, suggesting that Rif1 restricts the interactions between different replication timing domains. This is consistent with the loosening of the chromatin loop structures upon loss of Rif1. In fact, we observed increased dynamics and decreased condensation of the chromosomes in Rif1-depleted cells (Masai et al., unpublished results; Fig. 3). These observations support the idea that Rif1 organizes chromatin architecture to regulate DNA replication and possibly other chromosome transactions.

Fig. 3.

Chromatin architecture in the presence and absence of Rif1. Rif1 generates condensed and more static chromatin near the nuclear periphery through chromatin loops. In the absence of Rif1, chromatin loops are lost and chromatin is decondensed and becomes more dynamic (Masai et al., unpublished data). The 4C-Seq assays indicated that chromatin interactions between different timing domains increase in the absence of Rif1 (Foti et al., 2016), reflecting a more mobile nature of the chromosomes. Red and blue lines indicate, respectively, mid-S/late- and early-replicating chromosome segments. Double-headed arrows indicate chromatin interactions.

Regulation of transcription by Rif1 protein

As described above, Rif1 plays an important role in DNA repair choice in mammalian cells. Earlier studies indicated the absence of Rif1 at normal telomeres, excluding its direct role in telomere maintenance in mammalian cells (Silverman et al., 2004; Xu and Blackburn, 2004). However, it was reported that Rif1 may regulate telomere length homeostasis in mouse ES cells through transcriptional regulation of Zscan4, which can induce telomere extension by recombination and by upregulation of meiosis-specific homologous recombination genes (Dan et al., 2014). Zscan4 expression is dramatically increased by Rif1 knockdown, leading to telomere abnormalities including terminal hyperrecombination, telomere length heterogeneity, and chromosomal fusions. It was shown that Rif1, through interacting with and stabilizing histone H3K9 methyltransferase, elevates H3K9me3 level and negatively regulates Zscan4 expression. Zscan4 is a part of a gene cluster, and Rif1 negatively regulates transcription of genes present in gene clusters (Yoshizawa et al. and Yamazaki et al., unpublished results).

Roles of telomere-binding proteins in transcriptional silencing have been known for a long time (Hardy et al., 1992). Recent studies in fission yeast shed light on such silencing mechanisms. Taz1 and Rif1 are required to generate heterochromatin-euchromatin boundaries. They also impose the late replication property, suggesting that both transcription and replication are concomitantly regulated by telomere-binding factors (Toteva et al., 2017). Taz1, Rif1 and other shelterin components were also implicated in the generation of heterochromatin near late/dormant origins on the chromosome arms (Zofall et al., 2016). These results point to novel roles of telomere factors in inducing silencing of gene expression concomitant with inhibition of origin firing on the arm segment. It remains to be seen if Rif1 can similarly induce heterochromatin silencing in higher eukaryotes.

Summary and outstanding questions

Recent studies have implicated telomere-binding proteins in the regulation of chromosome transactions at chromosome arm segments. The most striking example is Rif1, which has been shown to be a conserved regulator of origin firing. Rif1 probably regulates firing of late/dormant origins through two distinct functions. The first is its ability to recruit phosphatase, which would antagonize the phosphorylation events essential for initiation. The second is its ability to bind to the G4 and generate higher-order chromatin architecture near the nuclear periphery, which would define a compartment for spatiotemporal regulation of DNA replication. The combination of these two activities enables Rif1 to efficiently and globally inhibit origin firing, and possibly to regulate transcription, repair and other processes.

Taz1, the fission yeast homolog of TRF1/2, is also required for suppression of origin firing, although the mechanism of its action is still unclear. Taz1-dependent origin suppression requires Rif1, which is known to interact with Taz1. Physical and functional interactions between telomere repeat sequences on the telomere and arms mediated by Taz1-Rif1 may contribute to the mechanisms ensuring suppression of some late origins.

Outstanding questions include the following. 1) What is the structural basis of Rif1-G4 interaction, and how does it contribute to the formation of nuclear architecture defining the spatiotemporal regulatory units for DNA replication and potentially other chromosome activities? 2) Is there any coordinated mechanism with which Rif1 would concomitantly regulate replication, transcription, repair, recombination and other chromosome transactions? 3) How is the Rif1-mediated chromatin architecture regulated during the cell cycle? 4) What is the mechanism for coregulation of origin firing and heterochromatin formation on the chromosome arms by telomere-binding factors, and is this regulation conserved in higher eukaryotes? 5) What is the nature of telomere-arm interactions, if any, that would enable telomere-binding factors to regulate chromosome events on the arm? 6) What are the developmental roles of telomere-binding proteins and the chromosome architecture governed by them? Active ongoing research in many laboratories should provide answers to these questions in the near future.

ACKNOWLEDGMENTS

We would like to thank Dr. Junko Kanoh for her invitation to write this article. We also thank Naoko Kakusho for editorial assistance, and all the members of our laboratory and other collaborators for helpful discussion and support.

REFERENCES
 
© 2017 by The Genetics Society of Japan
feedback
Top