Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/unpaywall/a-novel-rap1p-interacting-factor-rif2p-cooperates-with-rif1p-to-3CNlJ4u73Q
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- Unpaywall
- ISSN
- 0890-9369
- DOI
- 10.1101/gad.11.6.748
- Publisher site
- See Article on Publisher Site
Abstract
Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press A novel Raplp-interacting factor, Rif2p, cooperates with Riflp to regulate telomere length in Saccharomyces cerevisiae Davi d Wotton* and Davi d Shore^'^ Department of Microbiology, Columbia University College of Physicians and Surgeons, New York, New York 10032 USA; ^Department of Molecular Biology, University of Geneva, CH-1211 Geneva 4, Switzerland The Saccharomyces cerevisiae Rapl protein binds with high affinity to sites within the poly(Ci_3A) tracts at telomeres, where it plays a role in both telomere length regulation and the initiation of telomeric silencing. Raplp initiates silencing at telomeres by interacting through its carboxy-terminal domain with Sir3p and Sir4p, both of which are required for repression. This same domain of Raplp also negatively regulates telomere elongation, through an unknown mechanism. We have identified a ne w Rapl-interacting factor (Rif2p) that plays a role in telomere length regulation. RifZp has considerable functional similarities with a Raplp-interacting factor (Riflp) identified previously. Mutations in RIFl or RJF2 (unlike mutations in the silencing genes SIRS and SIR4) result in moderate telomere elongation and improved telomeric silencing. However, deletion of both RIFl and RIF2 in the same cell results in a dramatic increase in telomere length, similar to that seen with a carboxy-terminal truncation of Raplp. In addition, overexpression of either jRIFl or RIF2 decreases telomere length, and co-overexpression of these proteins can reverse the telomere elongation effect of overexpression of the Raplp carboxyl terminus. Finally, we show that Riflp and Rif2p can interact wit h each other in vivo. These results suggest that telomere length regulation is mediated by a protein complex consisting of Riflp and Rif2p, each of which has distinct regulatory functions. One role of Raplp in telomere length regulation is to recruit these proteins to the telomeres. [Key Words: Raplp; Riflp; Rif2p; telomere length; transcriptional silencing; Saccharomyces cerevisiae] Received November 4, 1996; revised version accepted January 31, 1997. Telomeres , the ends of eukaryotic chromosomes, are (Lustig and Petes 1986; for review, see Greider 1996; Za specialized protein-DNA complexes typically based kia n 1995a). upo n a simpl e DNA-repeat structure (TTAGGG in many Th e telomeric repeats in S. cerevisiae form high-affin multicellula r organisms). Telomeres in the yeast Sac ity binding sites for th e essential repressor/activator pro charomyces cerevisiae, which have an irregular repeat tei n 1 (Raplp) (Shore and Nasmyt h 1987; Buchma n et al. structur e commonly abbreviated as Ci_3A, are essential 1988; Longtine et al. 1989) as often as once per 18 bp for chromosom e stability (Sandell and Zakia n 1993). Th e (Gilson et al. 1993). Rapl p plays two roles at telomeres, maintenanc e of telomeres in yeast (Singer and bot h of which are mediated by its carboxy-terminal do Gottschlin g 1994), and probably most other organisms, mai n (for review, see Shore 1995; Zakian 1995a): It is requires a unique reverse transcriptase, called telomer- involved in transcriptional silencing, or telomere posi ase, tha t can add telomeri c repeats onto the chromosome tio n effect (TPE) (Kyrion et al. 1993; Morett i et al. 1994), ends (for review, see Blackburn 1994; Zakia n 1995b). Th e an d it controls telomere length (Conrad et al. 1990; lengt h of telomeric repeats is variable between strains Lustig et al. 1990; Sussel and Shore 1991; Kyrion et al. and individual clones of a given strain and can sponta 1992; Liu et al. 1994). Rapl p is also involved in silencing neousl y increase or decrease (Shampay and Blackburn at the HM mating-type loci (Kurtz and Shore 1991; Sus 1988), suggesting a balanc e between telomerase addition sel an d Shore 1991; Kyrion et al. 1993). Th e role of Raplp and degradation that is under complex genetic control in transcriptional silencing appears to be th e recruitment of specific repressor proteins (Sir3p and Sir4p) t o th e telo mere s and HM silencers via direct protein-protein inter ^Present address: Memorial-Sloan Kettering Cancer Center, Cell Biology action s (Moretti et al. 1994; Cockell et al. 1995; Hech t et and Genetics, New York, New York 10021 USA. al. 1996). In addition to its specialized roles at telomeres ^Corresponding author. E-MAIL [email protected]; FAX 41 22 702 68 68. and HM silent mating-type loci, Rapl p also binds to the GENES & DEVELOPMENT 11:748-760 © 1997 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/97 $5.00 Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Telomere length control in yeast upstream regions of many genes, including a large num synergistic effect on telomere length, resulting in telo ber of ribosomal protein and glycolytic enzyme genes, mere elongation similar to that seen in rapl^ cells. We where the protein appears to function as a transcrip argue from these and other data that Riflp and Rif2p tional activator (for review, see Shore 1994). Activation form a functional complex capable of regulating telo of transcription by Raplp may be mediated by a region mere length when recruited to telomeres by Raplp. immediately amino-terminal to the silencing and telo mere domain (Hardy et al. 1992a) and involves, at least in some cases, specific interactions with other DNA-bind- Results ing activator proteins (Tornow et al. 1993; Drazinic et al. Isolation of RIF2 1996). Point mutations (e.g., rapl'') and frameshift mutations To identify proteins that interact with Raplp and play a (e.g., rapl") within the Raplp carboxyl terminus result in role in mediating Raplp-dependent functions, we have varying degrees of telomere tract elongation, suggesting used the two-hybrid system to screen yeast genomic that this domain regulates telomere elongation (Sussel DNA libraries with a LexA/Rapl fusion encoding amino and Shore 1991; Kyrion et al. 1992; Liu et al. 1994). How acids 635-827 of Raplp. Previously we reported the iso ever, this function of Raplp does not appear to be medi lation of clones containing SIRS and SIR4, two genes ated directly by the Sir proteins. In fact, mutation of directly involved in transcriptional silencing (Moretti et SIRS or SIR4 actually results in slight telomere shorten al. 1994). Here we describe the isolation and character ing (Palladino et al. 1993), suggesting that the normal ization of a new gene from this two-hybrid screen, which function of these proteins in some way supports telo we call RIF2 {Raplp-inteiacting factor 2). As shown in mere elongation. Instead, a number of studies indicate Table 1, the GAD/Rif2p hybrid encoded by the library that other proteins interact with the Raplp carboxyl ter plasmid interacts specifically with LexA/Rapl(635-827), minus to negatively regulate telomere growth. For ex but not with LexA alone or an unrelated LexA hybrid ample, overexpression of the Raplp carboxyl terminus in (LexA/lamin). Furthermore, when the reading frame be the absence of its centrally located DNA-binding domain tween the GAD sequence and the RIF2 insert was dis (Conrad et al. 1990; Hardy 1991) or introduction of extra rupted (creating GAD/rif2fs) no activation of transcrip telomere repeats into cells (Runge and Zakian 1989) re tion was observed. Thus, the interaction with LexA/ sults in telomere elongation, consistent with the idea Rapl(635-827) is dependent on the GAD/Rif2 fusion that titratable Rap 1-interacting factors control telomere protein. length. One candidate for such a factor is Riflp, which LexA/Rapl(635-827) contains not only the Raplp si was identified in a two-hybrid screen as a protein that lencing domain, but also most of a transcriptional acti interacts with the Raplp carboxyl terminus (Hardy et al. vation domain, which spans amino acids 630-695 (Hardy 1992b). Interestingly, the telomere elongation defect of et al. 1992a). In addition, this hybrid protein has been rapr mutants may result from an inability to bind Riflp: shown to be a transcriptional activator in cells mutated Mutant rapP proteins interact weakly or not at all with for either SIR2, SIRS, SIR4, or RIFl (Moretti et al. 1994). Riflp in two-hybrid assays, and disruption of RIFl re Therefore, we wanted to determine whether expression sults in telomere elongation similar to that observed in of the GAD/Rif2 fusion was able to cause increased ac severe lapr mutants (Sussel and Shore 1991; Hardy et al. tivation from other LexA fusion proteins that are them 1992b). However, several lines of evidence point to the selves weak or cryptic activators. Both LexA/Gcr 1(4- existence of other Raplp-interacting factors involved in 419) and LexA/Clb2( 15-491) activate transcription from telomere length regulation. First, rapV mutants (which LexA operators to a low level. However, no increase in completely lack the carboxyl terminus) have a much transcriptional activity of these fusions was observed in more severe telomere elongation phenotype than either the presence of GAD/Rif2 (Table 2), further demonstrat rapr or lifl mutants (Kyrion et al. 1992). Second, point ing that the interaction of GAD/Rif2 with Raplp is spe mutations in the Raplp carboxyl terminus with weak cific. Sequence analysis of the GAD/Rif2 plasmid re telomere elongation phenotypes display additive effects vealed that the insert encoded a large open reading frame when present together or in combination with a RIFl mutation (Liu et al. 1994; L. Sussel, S. Buck, and D. Shore, unpubl.). Finally, overexpression of the Rapl car boxyl terminus causes further telomere elongation in Table 1. GAD/Rif2 interacts specifically with LexA/Rapl RIFl mutant cells, suggesting the existence of other GAD GAD/Rif2 GAD/rif2fs« Raplp-interacting factors that regulate telomere length (Wiley and Zakian 1995). LexA/Rapl (635-827) 7.1 249 6.9 Here we present the identification and characteriza 3.6 LexA 3.5 3.7 LexA/lamin 3.7 3.7 3.4 tion of a novel protein, which we call Rif2p (Rapl-inter acting factor 22). This protein interacts with the carboxyl P-Galactosidase activity in CTY10-5D cells determined as de terminus of Raplp in the two-hybrid system, and is also scribed previously (Moretti et al. 1994). able to interact with Riflp. Rif2p has striking functional *'GAD/rif2fs contains the same Rif2 insert as GAD/Rif2 but has similarities to Riflp, despite a lack of sequence similar a frameshift mutation introduced between the GAL4 activation ity. However, deletion of both RIF genes has a strong domain sequences and the RIF2 sequences. GENES & DEVELOPMENT 749 Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Wotton and Shore Table 2. GAD/Rif2 does not increase the function of other quire these other Rap 1-interactin g factors. To test this weak activators idea more directly, we examined the GAD/Rif 2 interac tio n wit h LexA/Rapl(647-827) in strains containing mu GAD GAD/Rif2 Fold increase tation s in either SIR3, SIR4, or RIFl. We chose this par LexA/RapP 7.5 271 ticula r LexA/Rapl hybrid because it still interacts with 36.2 LexA/Gcrr 39.3 29.2 0.7 Sir3p and Riflp hybrids bu t does no t become a transcrip LexA/Clb2^ 44.6 33.3 0.8 tiona l activator when SIR genes or RIFl are mutated (Moretti et al. 1994). As shown in Table 3, mutatio n of For determination of P-galactosidase activity, see Table 1 foot SIRS or SIR4 results in a -12 - and 16-fold increase, re note. spectively, in the apparent interaction between Raplp ''The LexA fusions used encoded amino acids 635-827 of Raplp, 4-419 of Gcrlp, and 15-491 of Clb2p. an d Rif2p. However, mutatio n of RIFl results in a muc h mor e dramatic (> 100-fold) increase in the Raplp-Rif2p interaction . These data clearly indicate that the Raplp - Rif2p interaction is not dependent upon native Sir3p, (see below and Fig. 2A) fused in frame, at amino acid 14, Sir4p, or Riflp, and provide further support for the idea wit h the GAL4 activation domain sequences. tha t thi s is a direct interaction. In addition, they raise the possibility that all three proteins (Sir3p, Sir4p, and Riflp) compet e with Rif2p for binding to the Raplp carboxyl Localization of a minimal Rif2p-interacting domain terminus . within Raplp SirSp, Sir4p, and Rif I p interact wit h a large portio n of the carboxyl terminus of Raplp encompassing part of the DBD Sil. putativ e activation domain and all of the sequences car- boxy-termina l to it (Hardy et al. 1992b; Moretti et al. 1994). To determine whether a similar region was re p-gal activity quired for interaction with Rif2p, we tested a series of GAD GAD/Rif 2 LexA/Rap l fusions with increasing amino-terminal de 8.1 33 1 letions . Proteins of the expected size for all LexA/Rapl fusions were visible by Western blotting with a LexA- 3.6 20. 9 specific antiserum (data not shown). As shown in Figure 1, GAD/Rif2 was able to interact with amino-terminal 5.5 44. 4 truncation s up to and including LexA/Rapl(679-827) 65 3 (the numbers in parentheses indicate the Raplp amino 4.6 29. 5 acids present in the hybrid), but failed to interact with 65 5 hybrids beginning at amino acid 691 or 702 in Raplp. 4.8 39. 6 Thi s pattern is very similar to that of GAD/Sir3(307- LexA fc. 66 7 978) and GAD/Rifl(1614-1916), but unlike that of 4.6 14 8 GAD/Sir4( 1204-1358), which interacts only with the LexA LB'i.i;ifo.-..:,k4?.- 67 9 larger LexA/Rapl(635-827) hybrid (Moretti et al. 1994). 2.4 3.4 LexA T o determine the carboxy-terminal boundary of the 69 1 Rif2p-interacting region of Raplp we used a series of LexA 1.4 1.4 LexA/Rap l fusions with a commo n amino-terminal fu 70 2 sion point, at amino acid 667. In contrast to the results 66 7 obtaine d with Sir3p and Riflp hybrids (Moretti et al. 1.9 31. 3 1994), in which mutation at amino acid 825 or deletion 82 7 t o amino acid 799 of Raplp severely weakened the in 2.7 77. 1 teraction s wit h these proteins , the same LexA/Rap l mu 82 5 tation s actually increased the signal obtained with 96. 0 3.1 GAD/Rif 2 (Fig. 1). Th e tw o shorter Rapl p fusions tested, encoding amin o acids 667-756 and 667-716, were unable 16.5 14.6 to interact with GAD/Rif 2. Thus, the region of Raplp 75 6 required for interaction with Rif2p lies between amino I LexA 13.8 12.8 acids 679 and 799. In contrast, the Raplp interaction wit h Riflp or Sir3p requires amin o acids 679-827, and an Figure 1. GAD/Rif2 interacts with amino acids 679-799 of even larger region is required for interaction with Sir4p Raplp. Two series of LexA/Rapl fusions with increasing (Moretti et al. 1994). amino- and carboxy-terminal truncations were assayed for in Th e results described above indicate that Rif2p can teraction with GAD/Rif2 in CTY10-5D cells. p-Galactosidase interac t with a smaller region of the Rapl p carboxyl ter activities were determined from liquid cultures as described minu s than that required by either Sir3p, Sir4p, or Riflp, previously (Moretti et al. 1994). Each fusion was also tested suggesting that the Rif2p-Raplp interaction does not re with GAD alone as a control. GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Telomere length control in yeast Table 3. Deletion of genes encoding knoMm Raplp- Rif2p are th e presence of a lysine-rich region (21.3% ly inteiacdng factors increases the interaction of LexA/Rapl sine) spanning amino acids 15-75 and a second basic re with GAD/Rif2 gion at the carboxyl terminus (17.1% lysine and 14.3% arginine over 35 amino acids). GAD/Rif2 GAD T o determine which region of Rif2p was responsible Wild type 20.9 3.6 for the interaction with Raplp and Riflp, we created a sir3::HIS3 259 5.7 series of LexA/Rif2 fusions and tested their interactions sir4::HIS3 332 5.9 wit h GAD/Rap l and GAD/Rifl . As show n in Figure 2B, rifl::HIS3 2244 4.9 truncatio n of the carboxyl terminus of Rif2p to amino For determination of p-galactosidase activity, see Table 1 foot acid 388 did not affect the interaction with Raplp or note. In all cases, the DNA-binding domain hybrid is LexA/ Riflp , but further truncation (to amino acid 332) abol Rapl (647-827). The wild-type strain is CTY10-5D, and the mu ished the interaction with both proteins. Both GAD/ tants were derived from this strain by gene disruption (Moretti Rap l and GAD/Rifl were able to interact with Rif2p et al. 1994). fusions starting at amin o acid 2 or 18. However , deletion of the first 49 amino acids of Rif2p abolished the inter actio n with Raplp (Fig. 2B). In contrast, the LexA/ Rif2(50-395) hybrid interacted strongly wit h GAD/Rifl, Rif2p interacts with Riflp i n fact giving more than twice as man y p-galactosidase A n alternative explanation of th e results described above unit s as the LexA/Rif2(2-395) hybrid, which interacts (Table 3) is that Rif2p normally interacts with the Sir wit h both Raplp and Riflp. A further deletion of 40 protein s or w^ith Riflp (in addition to Raplp), and that amino-termina l residues, giving LexA/Rif2(90-395), re th e bindin g of GAD/Rif2 to thes e protein s somehow pre sulted in a hybrid unable to interact with either the vent s or weakens a functional interaction with LexA/ Rapl p or th e Riflp hybrid. Thus, Rapl p and Riflp inter Rap l in th e two-hybrid system. In this case, mutatio n of act with largely overlapping regions of Rif2p, although th e SIR or RIFl genes might leave more GAD/Rif2 hy ther e does appear to be a difference between the regions brid free to interact productively with LexA/Rapl. To of Rif2p required for interaction wit h these tw o proteins. explore this possibility we created a LexA/Rif2 fusion Th e fact that GAD/Rifl interacts with LexA/Rif2(50- protei n containing amino acids 2-395 of Rif2p. This fu 395), whereas GAD/Rap l is unable to, suggests that the sion protei n did no t activate transcription on its ow n and Riflp-Rif2p interaction is independent of Raplp. n o significant increase in p-galactosidase activity was We further analyzed these interactions using cells in observed upon coexpression of either GAD/Sir3 or whic h either RIFl or RIF2 had been deleted. A RIF2 dis GAD/Sir4 , suggesting tha t these proteins do no t interact ruptio n was created by replacing the sequence encoding wit h Rif2p (Table 4). However, a GAD/Rifl fusion was amin o acids 18-389 of Rif2p with the HISS gene. Cells able to interact with LexA/Rif2(2-3 95), giving a signal lacking the RIF2 gene were viable and appeared to grow >20-fold over the background wit h GAD alone (Table 4). normally . Deletion of either RIFl or RIF2 did not result Thus , it appears that Raplp , Riflp, and Rif2p are all able i n loss of any of the interactions described above be t o interact with each other. In light of these results, the twee n Raplp, Riflp, and Rif2p, suggesting that none of simples t interpretation of the effect of SIR mutations on thes e proteins is bridging the interactions between any th e Rif2p-Raplp interaction (Table 3) would be that othe r pair (data not shown). Sir3p and Sir4p compete for binding to the Raplp car- boxyl terminu s wit h Rif2. In th e case of Riflp, because of th e observed interaction with Rif2p, we favor the idea RIF2 mutation affects silencing at both HMR and tha t the RIFl mutation improves the Raplp-Rif2 inter telomeres actio n by freeing GAD/Rif 2 to interact more produc Th e known Raplp-interacting proteins, Sir3p and Sir4p tively wit h the LexA/Rapl hybrid. However, we cannot on the one hand, and Riflp on the other, clearly play rule out the possibility that Riflp also competes with different roles in transcriptional silencing. Null muta Rif2p for binding to Raplp. tion s of either SIRS or SIR4 completely abolish silencing at bot h telomeres and HM loci (Ivy et al. 1986; Rine and Rif2p and its interaction with Riflp and Raplp Analysis of the amino-acid sequence of Rif2p (Fig. 2A), Table 4. Rif2p interacts with Riflp reveals that it encodes a protei n of 395 amin o acids with a predicted molecular mass of 46 kD . (The RIF2 gene is GAD GAD/Sir3 Gad/Sir4 GAD/Rifl o n chromosome XII-R, hypothetical protein YLR453c, LexA 5.1 4.8 5.3 6.0 GenBank accession no. U22382.) Wester n analysis of epi- LexA/Rif2 4.5 9.1 7.1 91.7 tope-tagged Rif2p, expressed from its own promoter, re vealed a protein of the expected size (data not shown). For determination of |3-galactosidase activity, see Table 1 foot Thi s protein has no significant homology to other note . Th e GAD fusions are as described in Morett i et al. (1994). know n sequences and no convincing structural motifs. Th e LexA/Rif2 fusion encodes amino acids 2-395 of Rif2p and Th e only notable features of th e amino acid sequence of wa s generated by PCR (see Materials and Methods). GENES SL DEVELOPMEN T 751 Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Wotto n and Shore MEHVDSIgaPIEI^SKKVVDSIKIVECftlglXLBOKN^^ 80 EU3IirajSISYVPSLSKFIiSKNU^a«KI^IVFHKVEH^ 160 BCIE3CffiSYSQQLDFEASEKPS!SmSDU<M4VMRKIMID^ 240 Ka^OAFKWLIYSMWSISSLLSISLKKKimKYTVFE^ 320 E3iLAIFFEFIJWFF«PFTYUFK^YIEIBA3SRIF^^ 395 GAD/Rap1 GAD/Rifl GAD Figure 2. Rif2p residues required for the interactions with Raplp and Riflp. [A] 3.0 180 95.3 The predicted amino-acid sequence of 3.4 176 87.9 Rif2p. The position of the original GAD fusion isolated is indicated by an arrow. A 6.1 4.1 4.0 lysine-rich region near the amino termi 3.2 4.3 5.0 nus and a basic carboxy-terminal stretch are underlined. [B] A series of LexA/Rif2 3.5 3.8 4.2 fusions were created to identify the region 4.8 9.5 of the protein required for interaction with Raplp and Riflp. ^-Galactosidase activi 2.5 3.8 3.1 ties were determined as in Fig. 1. Herskowit z 1987; Aparicio et al. 1991). RIPl mutations, however, actually improve silencing at telomeres while weakenin g repression at HMR loci containing a mutate d HMR-E silencer (Hardy et al. 1992b; Kyrion et al. 1993). It is thought, therefore, that SirSp and Sir4p are abso lutel y required for silencing, consistent with the obser vatio n that these proteins appear to be structural com ponent s of silent chromatin (Hecht et al. 1995, 1996). Rifl p appears to play a regulatory role in silencing by affecting the balance between telomeric and HM locus silencing, a conclusio n that is supported by studies of the unusua l properties of rapT mutations, which create a defect in the Raplp-Riflp interaction (Buck and Shore 1995). We therefore examined the effect of RIF2 mutation on silencing at HMR and at a genetically marked telomere. T o examine silencing at HMR, we used a sensitive re porter gene (TRPl) and an HMR-E silencer containing a mutatio n of the origin recognition complex (ORG) bind ing site {HMRAA::TRP1] (Sussel and Shore 1991). rapF and rifl mutations result in derepression at HMR only in th e context of this weakened silencer (Sussel and Shore ^^ •r. enn 'j| ^ • */»;**' •*" 1991; Hardy et al. 1992b). In an otherwise wild-type strai n background, the hmrAA::TRPl reporter is strongly repressed, as indicated by the absence of growth on me diu m lacking tryptophan (SC-Trp, Fig. 3A). Little or no effect of a RIF2 mutation alone is observed, whereas de Figure 3. Deletion of R1F2 affects silencing. [A] Strains were letio n of RIFl results in a clear increase in th e expression created with an hmrAA::TRPl reporter and gene disruptions of of the TRPl reporter gene at HMRAA, as reported previ either RIFl, RIF2, or both together. YDW76 [RIFl RIF2], YDW123 (iifl::URA3 RIF2), YDW77 {RIFl nf2::HIS3), and ously (Hardy et al. 1992b). Interestingly, despite the ab YDW124 {nfl::URA3 nf2::HIS3) were assayed for TRPl expres sence of a clear effect of the hf2 mutation, introduction sion (see Materials and Methods). The relevant genotypes are of th e rif2 disruption into a rifl HMRAAr.TRPl strain led shown on the left. (B) Telomeric silencing in strains {YDW84, to a significant further derepression of the reporter gene, YDW85, YDW86, and YDW87) with a telomeric ADE2 reporter as indicated by th e ability of essentially all of th e cells to was assayed by colony color. Repression of the telomeric ADE2 grow in the absence of tryptophan (Fig. 3A). Thus, al gene results in red sectors or colonies and ADE2 expression thoug h a nf2 mutation alone has little or no effect on results in white colonies. A single white colony of each strain HMR, in combination with a rifl mutation it has a syn was grown overnight in YPD and 200-300 cells were plated onto ergistic effect on silencing at this locus. YPD plates. Cells were photographed after 2-3 days at 30°C and 2 days at 4°C. T o examine the effects of RIF mutants on telomere 752 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Telomere length control in yeast positio n effect (TPE), w e assayed the expression of a telo- WT rif2 rifl ['/fl meri c ADE2 reporter gene by observing colony color. Ex pression of ADE2 results in white colonies, whereas in cells that do no t express the ADE2 gene red colonies are observed on plates containing minimal adenine. Colo nies containing red and whit e sectors represent a mixed populatio n of cells in which the ADE2 gene is repressed «• • •»•»• « as well as expressed. For each strain, a single white colony was picked and grown overnight in rich medium, after which 200-300 cells were spread onto YPD plates and the colony color observed. In wild-type cells very few sectored or red colonies were observed (<1%), sug gesting tha t the ADE2 gene was expressed in most cells, and that this derepressed state was relatively stable (Fig. 3B). Deletio n of RIF2 resulted in a smal l bu t reproducible increase in th e numbe r of sectored colonies to ~11 % (Fig. 2323 - 3B). Mutatio n of RIFl has been shown previously to in crease TPE (Kyrion et al. 1993), and as shown in Figure 3B, this effect is larger than that seen in rif2 mutants 1929 (-34% sectored colonies). In cells lacking bot h Rif I p and Rif2p, a large proportion of the colonies was either com pletely red or contained red sectors (>50%; Fig. 3B). Thus , although mutatio n of RIF2 increased telomeric si lencing only slightly, it had a synergistic effect when combine d with a hfl mutation. In summary , th e hfl phenotype is stronger tha n tha t of rif2 in bot h HMR silencing and TPE assays. Furthermore, despite the relatively weak effect of a hf2 mutation in eithe r assay, it has a strong synergistic effect together Figure 4. Increased telomere length in RIF-rautant cells. DNA wit h a hfl mutation on the balance between telomeric was isolated from wild-type, lifl, rif2, and rifl rif2 cells and HMR silencing, significantly strengthening the (YDW76, YDW123, YDW 77, and YDW124) and was digested former and weakening the latter. with Xliol and hybridized to reveal telomere repeat sequences, as described in Materials and Methods. In wild-type cells, this digest releases a broad band of -1.2 kb containing the telomeric RIFl and RIF2 mutaUons have a synergistic effect on repeat sequences from the majority of telomeres (see Zakian telomere elongation and chromosome loss 1995a). The relevant genotypes are indicated above the lanes. On e possible explanation for the strong effect of the rif2 mutatio n on silencing only whe n present together wit h a rifl mutation is that rif2 significantly exacerbates the telomer e elongation effect of rifl mutations (Hardy et al. Rifl p in two-hybrid assays) is similar to tha t seen in rifl 1992b). Extreme telomere elongation (caused by rapV cells suggests that this allele of RAPl retains some abil mutations ) can increase TPE in wild-type cells (Kyrion et ity to regulate telomere length. Consistent wit h this, w e al. 1993) whil e exerting an opposite effect on silencing at found that Rif2p is able to interac t wit h rapT mutants in HMR (Buck and Shore 1995). We therefore examined the th e two-hybrid assay (data not shown). However, this length of th e telomeric repeats in rif2 and rifl rif2 double interactio n was reduced to -50 % of that seen wit h wild- mutan t cells. As shown in Figure 4, deletion of RIF2 typ e Rapl p fusions, indicating that the rapT mutations alone results in an increase of -100 bp in the average do not specifically affect the Raplp-Riflp interaction, bu t instead may be having a more general effect on the lengt h of Y'-containing telomeres (which yield th e broad carboxy-terminal region of Raplp . This reduced rapPp - lower band in th e Xhol digest shown). For comparison, a Rif2p interaction ma y explain wh y rapT rifl double mu RIFl deletion resulted in a somewha t greater increase in tant s have a mor e severe telomer e elongation phenotype average length and heterogeneity of telomeres, as shown tha n either single mutant (L. Sussel, S. Buck, and D. previously (Hardy et al. 1992b). In DN A from cells lack Shore, unpubl.). An additional possibility is that some ing both RIFl and RIF2, the length of the telomeric re Rifl protein is recruited to telomeres in rapr cells via peat sequences is increased by at least 600 bp, and up to th e Rif Ip-Rif2 p interaction. 2.5 kb (Fig. 4), which is considerably greater than the increase for either single RIF mutant. Thus, Riflp and As mentioned above, previous work has shown that a Rif2p probably perform different functions in telomere large carboxy-terminal truncation of Rapl p (in particular lengt h regulation. th e rapV allele rapl-17, in which amino acids 663-827 Th e fact that the telomere elongation observed in se are missing) results in a dramatic increase in telomere vere rapT mutants (which are unable to interact with lengt h (Kyrion et al. 1992). Th e protein produced by the GENES & DEVELOPMENT 753 Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Wotton and Shore T o further address the functional relationships be RAP! rap1-17 RAP! twee n Rif Ip , Rif2p, and th e Rapl p carboxyl terminus , we measure d chromosome loss rates in RIF single mutants an d the lifl iif2 double mutant. Previous studies have show n that lapl'' mutants display elevated chromosome loss rates, perhaps as a result of impaired telomere func tio n (Kyrion et al. 1992). As show n in Table 5, w e found tha t chromosome III loss rates in rifl and nf2 mutants wer e elevated by -7.5- and 3.5-fold, respectively, com pared wit h wild-type. However, just as for silencing and telomer e length measurements, the rifl hf2 double mu tan t displayed a synergistic effect, giving a chromosome III loss rate >30-fold greater tha n wild type. Thi s relative increase is comparable to that determined for the most severe rapV mutant [iapl-17] using a different chromo som e loss assay (Kyrion et al. 1992). Additionally, cells mutated for both RIFl and RIF2 hav e a significantly longer doubling tim e tha n wild-type cells (Table 5). In contrast, mutation of either RIF gene alone does not significantly alter the growth rate, pro viding further evidence for synergistic action of Riflp an d Rif2p. Overexpression of Riflp and Rif2p reduces telomere length and reverses the effect of Raplp 1371 ^ carboxy-terminal overexpression 1264- § m It has been demonstrated previously that overexpression Figure 5. Telomere length in iapl-17 cells and in cells lacking of the carboxyl terminus of Raplp, in the absence of its both RIFl and RIF2. DNA was isolated and treated as in Fig. 4. DNA-bindin g domain, increases telomere length (Con The genotypes of the cells are shown at the top. (+) Wild type,- (-) rad et al. 1990; Hardy 1991). Th e likely explanation of mutant for RIFl or RIF2 as shown. The RAPl alleles were either thi s effect is tha t the overexpressed carboxyl terminu s of wild-type or rapl-17, which expresses a Rapl protein lacking Rapl p interacts with proteins involved in telomere the carboxy-terminal 164 amino acids. The strains used in this lengt h regulation, titrating them away from the telo analysis (YDW126-YDW131) were generated by sporulation mere s and thus causing telomere elongation. We rea and dissection of the diploid YDW125, which is heterozygous for the iapl-17 allele and disruptions of both RIFl and RIF2. soned that if Riflp and Rif2p were responsible for telo mer e length regulation, and in particular acted as nega tive regulators of telomere elongation, overexpression of thes e proteins might shorten telomeres in wild-type rapl-17 allele would be unable to interact with either cells and restore normal telomere length to cells in Rifl p or Rif2p, as judged by our two-hybrid analyses whic h the carboxyl terminus of Rapl p is overexpressed. (Hardy et al. 1992b; Fig. 1). It was therefore of interest to As show n in Figure 6 (lanes 2,3), th e telomeric repeats in determin e whether the changes in telomere length in a rifl rif2 double mutan t are comparable to those seen in a rapl-17 strain. To compare these strains more easily we Table 5. Phenotypes of cells lacking Riflp and Rif2p crossed a rapl-17 strain with a strain mutated for both RIFl and RIF2 to create a diploid that was heterozygous Telomer e length Doublin g Chromosom e for each mutation . This diploid was sporulated to obtain increase'' time'^ loss rate*^ (min) (per generation) strain s with combinations of these three mutations. As Genotyp e (bp) show n in Figure 5, a similar increase in telomere length Wild type N.A . 90 4.3 X 10-^ an d heterogeneity is observed when Raplp is truncated 200-60 0 92 3.2 X 10-^ rifl or whe n both RIFl and RIF2 are mutated. Additionally, Tif2 100 90 1.5 X 10-^ n o further increase in telomere length was observed in rifl rif2 600-250 0 108 1.5 X 10-5 cells carrying the rapl-17 allele in combination with "All strains are isogenic to W303 except for the RIF mutations mutation s in RIFl and RIF2 (Fig. 5). Thes e results indi indicated. cate tha t the Rapl p carboxyl terminu s and th e Rifl/Rif 2 ^Derived from Figs. 4 and 5. protei n pair act in the same pathway to regulate telo '^Growth rates were measured for exponentially growing cells in mer e length. In other words, the function of the Raplp rich medium (YPD) at 30°C. carboxyl terminu s can be explained, at least in principle, ''Based on determination of rate of loss of chromosome 111 (see by its ability to recruit Riflp and Rif2p to the telomere. Materials and Methods for details). GENES &, DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Telomere length control in yeast multi-copy nificantly reduced telomere lengths to nearly wild-type low-copy levels, whereas RIP2 had a small effect. We also examined the effect on telomere length of the addition of single extra copies of th e RIP and SIR genes in th e presence of carboxy-terminal overexpression of Rap1 C-ter:.... + + + + + + + + + + + + Raplp . Th e addition of RIFl on a centromere-based plas- mi d was able to partially suppress the effect of th e LexA/ Rapl(630-827) plasmid. Strikingly, whe n RIFl and RIF2 wer e present together on centromeric plasmids in these cells, a reduction in telomere length and heterogeneity '* * ?l 1^ !»•*• almos t to that of wild-type cells was observed (Fig. 4, lanes 10-13). N o effect of either SIRS or SIR4 alone was l««i ii JHPWw'^" 1 ^ observed, although some suppression of telomere elon gation was observed wit h both together (Fig. 6, lanes 14- m^ 16). Taken together, these results suggest tha t the levels of Riflp and Rif2p can have a direct effect on the regu 4822- 4324- —1& -' *^;iai: latio n of telomere length, and provide further indication tha t these two proteins play a critical role in this process. —• , — :: _'_ .,i. 3675- —•>• , Discussion Several lines of evidence indicate that Raplp, which bind s to multiple high-affinity sites within the poly - a 2323 — (Ci_3A) tracts at telomeres, plays a critical yet complex role in telomere length regulation. Temperature-sensi 1 929 — tive RAPl mutants, grown under semi-permissive con ditions, have shorter telomeres (Conrad et al. 1990; Lustig et al. 1990), suggesting tha t one function of Raplp ma y be simply to protect the telomere from degradation 1371 — by exonucleases. However, this is unlikel y to be th e sole 1264 — •I driyi""ntii^" function of Raplp in telomere length regulation. Lustig an d colleagues have shown that carboxy-terminal trun catio n mutations of RAPl, which retain the DNA-bind- ing domain and are viable, exhibit an extreme telomere- 1 2 3 4 5 6 7 8 9 10 12 14 16 elongation phenotyp e (Kyrion et al. 1992). It is presumed, therefore, tha t this nonessentia l domain of Raplp , which Figure 6. Effects of overexpression of Riflp and Rif2p on telo is also required for telomere position effect and HM lo mere length. Wild-type W303-1B cells [MATa] were transformed with plasmids containing the genes encoding the proteins indi cus silencing (Kyrion et al. 1993; Moretti et al. 1994), cated at the top. Riflp, Rif2p, Sir3p, and Sir4p were expressed as negatively regulates telomere elongation. This migh t oc full-length proteins from their own promoters, present on either cur, for example, by direct inhibition of telomerase or by multicopy (2ia-based) or low-copy (CEN/ARS-based) plasmids. controlling the access of telomerase to the chromosome (ctrl) The presence of a control plasmid lacking an insert. (Rapl end. Whatever the mechanism of telomere length regu C-ter) The presence (+) of a plasmid encoding amino acids 630- latio n by Raplp , it is likely to wor k through interactions 827 of Raplp fused to LexA. This fusion, which lacks the Raplp wit h other proteins because overexpression of th e Raplp DNA-binding domain, is expressed from the strong ADHl pro carboxyl terminus, in the absence of its normal DNA- moter on a multicopy {2]x] plasmid. DNA was prepared and bindin g domain, results in telomere elongation (Conrad telomere repeat lengths analyzed as in Fig. 4. et al. 1990; Hardy 1991). Thes e data can be most easily explained by the existence of titratable factors that in terac t wit h th e Rapl p carboxyl terminu s to mediat e telo mer e length regulation. The first candidate for such a wild-type cells carrying either full-length RIF2 or RIFl factor to be identified was Riflp (Raplp-interacting fac on a multicopy plasmid were reduced in length. No ef tor 1), whic h was isolated by a two-hybrid screen using fect of a multicopy S/i?4-containing plasmid was ob th e carboxy-terminal 175 amin o acids of Rapl p (Hardy et served (Fig. 6, lane 4). In contrast, cells expressing LexA/ al. 1992b). However, the effect on telomere length of Rapl(630-827) from the strong ADHl promoter (on a mutatin g RIFl is relatively small by comparison with multicop y plasmid) had slightly elongated telomeres th e Raplp truncation mutants. (Fig. 6, lane 5), consistent with previous studies (Conrad et al. 1990; Hardy 1991). Strikingly, overexpression of Riflp and Rif2p are required to mediate the telomere eithe r RIFl or RIF2, but no t SIR4, resulted in a reduction length regulation function of Raplp in telomere length in the presence of LexA/Rap 1(630- 827) (Fig. 6, lanes 6-9). Th e RIFl high-copy plasmid sig- Her e we have described a new Raplp-interacting factor GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Wotto n and Shore (Rif2p) with properties very similar to those of Riflp. 1992b; Kyrion et al. 1993), the precise function of this Both of these proteins interact with the Raplp carboxyl protein was unclear. The identification and characteriza terminus in two-hybrid assays, and mutation of either tion of Rif2p reported here greatly clarifies this picture protein results in moderate telomere elongation and an by providing evidence for a specialized mechanism of increase in TPE. Significantly, cells lacking both of these telomere length regulation by Raplp, separate from its proteins have extremely elongated telomeres that are in silencing function. The fact that Rif2p is required to distinguishable from those in cells containing a rapV gether with Riflp for telomere length control, and that mutation, which removes the Raplp carboxyl terminus. these two proteins interact with each other in two-hy The simplest interpretation of these observations is that brid assays, supports the idea that their primary role is to the Raplp carboxyl terminus recruits both Riflp and negatively regulate telomere elongation. Several lines of Rif2p to telomeres where they are required to regulate evidence indicate that the changes in silencing brought telomere elongation. One prediction of this model, about by RIF mutations are likely to be secondary con which we are at present trying to test, is that Rif proteins sequences of their effects on telomere length and struc are localized at telomeres in vivo. Our results do not ture. To begin with, telomere elongation in wild-type indicate whether or not Riflp and Rif2p are sufficient to cells is sufficient to cause increased TPE (Kyrion et al. mediate the telomere length regulation function of 1993) and decreased silencing at HMR (Buck and Shore Raplp, but they do indicate that both are necessary. If 1995). In principle, therefore, the effects of RIF muta other Rifs are involved in this regulatory function, their tions on silencing could be an indirect consequence of action must be dependent upon either RIFl, RIF2, or their effects on telomere structure. Second, Rif proteins, both. Finally, the fact that neither RIFl nor RIF2 muta unlike the Sir proteins, are not required for silencing, but tion exacerbates the telomere elongation phenotype of rather affect the balance between telomeric and HMR the rapl-17 allele suggests that the function of these two locus silencing mediated by competition for Sir proteins proteins in telomere length regulation requires the (Buck and Shore 1995). As such, the Rif proteins can be Raplp carboxyl terminus. viewed as regulators of silencing, but are not required components of silent chromatin, like Sir3p and Sir4p Because Riflp and Rif2p can interact with each other (Hecht et al. 1995, 1996). Finally, it is important to keep in the two-hybrid system, these two proteins may act as in mind that the effects of SIRS and SIR4 mutations on a complex that is recruited to the telomere. If this is the telomeres are exactly opposite those of RIFl and RIF2 case, it might seem odd that the two proteins have dis mutations. In the SIR mutants TPE is abolished (Apari- tinct functions, as indicated by the synergistic effect of a cio et al. 1991) and telomere tracts become slightly lifl hf2 double mutation on telomere length. One pos shorter (Palladino et al. 1993), whereas RIF mutations sible explanation for these observations is that, once re improve TPE (Fig. 3B; Kyrion et al. 1993) and increase cruited to the telomere, Riflp and Rif2p interact with telomere length (Fig. 4; Hardy et al. 1992b). different proteins, and thus genetically and biochemi cally define two regulatory pathways for telomere length In a separate report (Marcand et al. 1997) we present regulation. Alternatively, Riflp and Rif2p might have a evidence that telomere length is regulated by a negative common target. In this case each protein would by itself feedback mechanism that can sense the precise number provide partial regulation by interacting independently of Rapl carboxyl termini at the chromosome end. The with this target. More information concerning the data presented here suggest that the Riflp/Rif2p com mechanism of action of both proteins will be required plex, bound to Raplp at telomeres, mediates this length- before these models can be distinguished. Another inter sensing and regulation function of Raplp, which is itself esting possibility raised by the interaction of Riflp with antagonized by Raplp-Sir interactions. Putting these Rif2p is that Riflp at one telomere may interact with and other data together, we propose a model in which Rif2p at another. Thus, the Rif proteins may play a role Raplp-Sir and Raplp-Rif complexes are naturally parti in the clustering of telomeres into groups that has been tioned to opposite ends of the telomeric Ci_3A repeats observed by immunolocalization of telomere-associated (see Fig. 7). We imagine that Raplp-Sir interactions are proteins and in situ hybridization with telomere-specific favored at the proximal end of the telosome (Wright et al. probes (Klein et al. 1992; Palladino et al. 1993; Gotta et 1992) (the telosome/nucleosome junction) by coopera al. 1996). Finally, we should point out that it is unclear at tive interactions between Sir proteins themselves (Mor- present what, if any, role Rifl and Rif2 might play in a etti et al. 1994) and between Sir proteins and histones H3 novel mechanism of telomere length control, called and H4 (Hecht et al. 1995). Because TPE is likely to re "telomeric rapid deletion" (Li and Lustig 1996), which sult from a continuous spreading of Sir complexes along can reduce extremely elongated telomeres to wild-type the nucleosome fiber, it follows that the Raplp mol lengths in what appears to be a single-division event. ecules at the junction between the telosome and nucleo- somal DNA would be most critical for establishing si lencing. On the other hand, the Riflp/Rif 2p complex is Rif and Sir proteins mediate different functions of likely to act at the telomere end, and we speculate that Raplp at telomeres Rif protein assembly at the distal end of the telosome Because the initial characterization of RIFl revealed a might be promoted by interactions with telomere end- complex set of effects on telomeric and HM locus silenc binding proteins. We note that a tendency of Rif com ing as well as telomere length control (Hardy et al. plexes to assemble at the distal end of the telomere 756 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Telomere length control in yeast Figure 7. A speculative model for the arrange ment of Rif and Sir protein complexes at telo meres. Partitioning of the telosome (Wright et al. 1992) into centromere-proximal Sir and telo- mere-proximal Rif domains is diagramed, to gether with the effects of increased [A] and de Increased Rif expression Decreased Rif ex pr ess ion creased [B] expression of Rif proteins. We pro pose that the extent of the Raplp-Rif protein complex may play a direct role in regulating telomere length. In the case of Rif protein over- expression we imagine that Rif complex spreads from the telomere end and results in a gradual fCDCIi) shortening of the Cj^aA tract until a new stable equilibrium value is reached in which Rif com plex addition and telomere shortening are in bal ance, Under conditions of decreased Rif expres sion we propose a loss of Rif complex at the telosome and a corresponding increase in Sir complex (which results in improved TPE). This reduction of Rif complex results in telomere elongation, which, depending upon the severity of the defect, may lead to a longer, stable telo mere length [rifl or rif2 single mutants) or es ISir3p/Sir4p (^^Riflp/Rif2p sentially unregulated telomere elongation {rifl CZ) Rap Ip rif2 double mutant). migh t help to explain recent observations in Kluyveio- How do Rif proteins regulate telomere lengths myces lactis that suggest that RapIp-binding sites near est to th e telomere end play a critical role in lengt h regu Ou r data can b e mos t simply explained by proposing that latio n (Krauskopf and Blackburn 1996). Nevertheless , we Rif proteins are recruited to telomeres by Raplp where woul d emphasize tha t this model is speculative, and that the y either block telomere elongation directly or pro th e available data do not rule out a model in which Sir mot e degradation of th e ends. T o explain ho w this regu and Rif complexes are interdigitated along th e telosome. latory mechanis m can be highly sensitive to th e number It seems likely that a direct biochemical and structural of Raplp molecules (and ultimately Rif complexes) characterizatio n of the telosome will be required to dis boun d at an end, w e woul d propose tha t either a stochas tinguis h between these two models. tic process controls events at the end through interac tion s wit h a Raplp/Ri f complex, as suggested previously Whateve r the precise arrangement of Sir and Rif com for the yeast K. lactis (McEachern and Blackburn 1995), plexes along the telosome, the critical parameter con trolling telomere length regulation would appear to be or the Raplp/Rif complex creates a length-sensitive th e amount of Rif complex assembled there. The short switc h to control the access of factors to the telomere ening of telomeres caused by the overexpression of Rif end or their activity once bound there (Marcand et al. protein s (Fig. 6) woul d thus result from an extension of 1997). On e obvious candidate for Rif action is th e telom th e Rif complex along the telosome that results in telo erase enzyme itself, whic h has been detected recently in mer e shortening, possibly by inhibition of telomerase vitr o (Cohn and Blackburn 1995; Lin and Zakian 1995), (Fig. 7A). Conversely, mutation of one of the RIF genes and whose RNA template is now known (Singer and ma y result in a decrease in the amount of the telomere Gottschlin g 1994). Another candidate for Rif action is boun d by Rif protein, allowing telomerase activation and th e product of the PIFl gene, a helicase that inhibits telomer e elongation (Fig. 7B). Th e removal of Sir3p or telomer e elongation (Schulz and Zakian 1994). One Sir4p by mutation would allow free access of the Rifl/ could imagine that Rif proteins target this enzyme to Rif2 complex to the telosome, which would cause in telomeres . However, genetic epistasis tests suggest that creased telomerase inhibition and result in telomere Pifl p acts in a different pathway from the Raplp car- shortening, which is precisely what has has been ob boxyl terminus (Schulz and Zakian 1994). Alternatively, served (Palladino et al. 1993). Our model for telomere th e Rif proteins may interact with one or more recently lengt h regulation (Marcand et al. 1997) predicts that the identified telomere end-binding proteins to regulate exten t of telomere shortening in SIRS or SIR4 mutants telomerase . The Estl (Virta-Pearlman et al. 1996) and (-50 bp) would reflect the amount of Ci_3A tract occu Cdcl3(Est4) (Lin and Zakian 1996; Nugent et al. 1996) pied by Raplp-Sir complexes (thus excluding Rifl/Rif2) protein s both bind to single-strand TGi_3 telomeric se in wild-type cells. Assuming that Raplp binding sites quence s and appear by genetic criteria to be essential occur approximately every 18 b p along th e telomere, this regulators of telomerase activity (Lundblad and Szostak woul d translate to about three Raplp-Sir complexes. 1989; Garvik et al. 1995; Nugent et al. 1996). A newly GENES &. DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Wotto n and Shore identifie d Cdcl3p-interactin g protei n wit h simila r func fusions, GAD/Rifl, GAD/Sir3, and GAD/Sir4, are as described tions , Stnlp , is anothe r potentia l Rif target (Grandin et previously (Hardy et al. 1992a; Moretti et al. 1994). Wild-type RIPl, RIF2, SIRS, and SIR4 were expressed from their own pro al . 1997). Given our present limited understanding of moters , and were present on CEN/ARS- or 2-]Li-based pRS vec telomer e replication , it seem s likel y tha t ther e are other tors (Sikorski and Hieter 1989). Th e full-length RW2 gene, in possibl e mechanism s of Rifl action. Identifying factors cluding 495 bp of sequence 5' to th e initiatio n codon and 120 bp othe r tha n Rapl p tha t interac t wit h th e Rif proteins , an d 3 ' to th e translational stop, was isolated by PCR and cloned into studyin g geneti c and biochemica l interactions between pBS. Th e RIF2 deletion/insertion mutation was created in pBS th e RIF genes and know n telomere replication factors, by replacing the sequences encoding amino acids 18-389 with shoul d provide importan t clues to understanding how th e HISS gene. RIFl, SIRS, and SIR4 disruption mutations have th e Rif protein s work. been described previously (Hardy et al. 1992b; Moretti et al. 1994). Sequences of oligonucleotides used in PCR-based cloning steps and details of all plasmid constructs are available upon Material s and methods request. Plasmids Yeast strains All LexA fusions were expressed from pBTM116 (2 fim origin, TRPl, pADHl-lexA; a gift of P. Bartel and S. Fields, State Uni Th e yeast strains used in this study are all derivatives of either versity of Ne w York, Stony Brook). Th e LexA/Rapl fusions are W303 (Thomas and Rothstein 1989) or the two-hybrid assay as described previously (Moretti et al. 1994), except for LexA/ strain CTY10-5D, and are hsted in Table 6. Rap 1(679-756), which was created by PCR. The GAD-fusion library screened with LexA/Rapl was created in pGAD 3 (Chien Yeast manipulations et al. 1991) and was a generous gift of P. Bartel and S. Fields. GAD/rif2fs was created by digesting GAD/Rif2 with Mlul, end- Growt h and manipulatio n of yeast strains was carried out using filling wit h Klenow, and religating. LexA/Rif 2 fusions were gen standard procedures (Rose et al. 1990). Screening of the GAD- erated by PCR, except for those encoding amino acids 18-389 fusion library wit h LexA/Rapl was carried out in CTY10-5D, as and 18-298, which were created by subcloning fragments from described previously (Moretti et al. 1994). Spot assays were per GAD/Rif 2 into pBS (Stratagene), then into pBTM116. GAD/ formed by spotting 5ial aliquots of 10-fold serial dilutions of a Rap l encoded amino acids 478-827 and was isolated by screen saturate d overnight culture onto plates containing th e appropri ing a two-hybrid library wit h LexA/Rif2(2-395). The Gbd/Rapl at e selective media. Plates were photographed after 2-3 days. Tabl e 6. Yeast strains used in this study W303-1A HMLa MATa HMRa ade2-l canl-100 his3-ll,15 leu2-3,112 trpl-1 uraS-l W303-1B HMLoL MATa HMRa ade2-l canl-100 hisS-11,15 leu2-3,112 trpl-1 uraS-l CTY10-5D MATa ade2-l trpl-901 leu2-3,112 his3-200 gal4 gal80 URA3::LexA op-lacZ PMlO l CTY10-5D siT3::HIS3 PM102 CTY10-5D sir4::HIS3 PM103 CTY10-5D riflr.HISS AJL394-la W303-1A rapl-17 telVII::ADE2/URA3 YLS607 W303-1B hmriAA::ADE2 riflr.URAS YDW62 CTY10-5D rif2::HIS3 YDW76 W303-1B hmrAA::TRPl telVII::URA3 YDW77 YDW76 rif2::HIS3 YDW123 YBW76 iifl::URA3 YDW124 YDW76 riflr.URAS rif2::HIS3 YDW80 W303 HMLa MATa hmrAA::ADE2 HMLa MATa hmrAA::ADE2 YDW81 YDW80 rifl::URAS/rifl::URA3 YDW82 YDW80 rif2::HISS/rif2::HIS3 YDW83 YDW80 rifl::URAS/rifl::URA3 rif2:: HIS3/rif2::HIS3 YDW84 W303-1A hmrAA::TRPl telVII::ADE2/URA3 YDW85 YDWM rifl::URA3 YDW86 YDW84 rif2::HISS YDW87 YDW84 riflr.URAS rif2rHISS YDW125 W303 HMLa MATa hmrAArTRPl rapl-17 riflr.URAS rif2r.HlS3 HMLa MATa HMRa RAPl RIFl RIF2 YDW126 W303-1A hmrAArTRPl YDW127 W303-IA rapl-17 YDW128 YDW127 rifl URA3 YDW129 YDW127 rif2:. HISS YDW130 YDW127 rifl: URAS rif2::HIS3 YD W13 1 YDW116 rifl: URA3 rif2::HIS3 758 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Telomere length control in yeast P-galactosidase assays were carried out as described previously twee n HMR and telomeres in yeast. Genes and Dev. 9: 370 - (Moretti et al. 1994). 384. Chi , M.-H. and D. Shore. 1995. SUMl-1, a dominan t suppressor of SIR mutations in Sacchawmyces cerevisiae, increases Chiomosome loss assays transcriptiona l silencing at telomeres and HM mating-type loci and decreases chromosome stability. Mol. Cell. Biol. Determinatio n of th e rate of loss of chromosome III wa s carried 16:4281-4294 . ou t as described previously (Chi and Shore 1995). Briefly, single Chien , C.-T., P.L. Bartel, R. Sternglanz, and S. Fields. 1991. Th e colonies were resuspended in 1 m l of YPD and sonicated. Dilu two-hybrid system: A metho d to identify and clone genes for tion s were plated onto YPD medium to assess the number of protein s that interact with a protein of interest. Proc. Natl. viable cells. Half of each colony was then incubated with a Acad. Sci. 88: 9578-9582. 100-fold excess of a MATa tester strain for 8 hr at 30°C. Cells wer e the n spread onto SD plates to select for colonies tha t arose Cockell, M., F. Palladino, T. Laroche, G. Kyrion, C. Liu, A.J. from a mating event. The chromosome rate loss was derived Lustig, and S.M. Gasser. 1995. Th e carboxy termini of Sir4 according to the following formula: (0.4343 x F)/log N-log NQ, and Rapl affect Sir3 localization: Evidence for a multicom- wher e P = th e frequency of mating events, N = th e number of ponen t complex required for yeast telomeric silencing. /. cells in th e colony and NQ = th e numbe r of cells from which the Cell Biol. 129: 909-924. colony arose. For this analysis, we used diploid strains that Cohn , M. and E.H. Blackburn. 1995. Telomerase in yeast. Sci lacked any mating-type information at HMR. ence 269:396-400. Conrad, M.N., J.H. Wright, A.J. Wolf, and V.A. Zakian. 1990. RAP l protein interacts with yeast telomeres in vivo: Over DNA sequencing productio n alters telomere structure and decreases chromo som e stability. Cell 63: 739-750. All sequencing was carried out by use of the dideoxy chain Drazinic, CM. , J.B. Smerage, M.C. Lopez, andH.V . Baker. 1996. terminatio n method, using Sequenase (Amersham). Activatio n mechanism of the multifunctional transcription factor Repressor-Activator Protein 1 (Raplp) . Mol. Cell. Biol. Telomere blots 16:3187-3196 . Garvik, B., M. Carlson, and L. Hartwell. 1995. Single-stranded DN A was isolated from overnight yeast cultures and 1 pg was DN A arising at telomeres in cdcl3 mutants may constitute digested wit h Xhol for 4 hr. DN A was electrophoresed through a specific signal for the RAD9 checkpoint. Mol. Cell. Biol. 0.8% agarose and transferred to nylon membranes (Hybond). 15:6128-6138 . Membrane s were hybridized at 50°C in 6x SSC, 5x Denhardt's Gilson, E., M. Roberge, R. Giraldo, D. Rhodes, and S.M. Gasser. solution, with a poly[d(G-T)] probe labeled by random priming. 1993. Distortion of the DNA double helix by RAPl at si Membrane s were washed twice in 2x SSC at 55°C for 45 min lencers and multiple telomeric binding sites. /. Mol. Biol. and autoradiographed with Kodak X-AR5 film. 231:293-310 . Gotta , M., T. Laroche, A. Formenton, L. Maillet, H. Scherthan, Acknowledgments and S.M. Gasser. 1996. Th e clustering of telomeres and co- localization with Rapl, Sir3, and Sir4 proteins in wild-type We are grateful to A. Lustig for providing strains and to P. Bartel Saccharomyces cerevisiae. J. Cell Biol. 134: 1349-1363. and S. Fields for generously providing the pGAD3 library. We Grandin, N., S.I. Reed, and M. Charbonneau. 1997. Stnl , a ne w woul d also like to thank H. Laman, L. Pemberton, D. Vannier, Saccharomyces cerevisiae protein, is implicate d in telomere and S. Marcand for critical discussion and helpful suggestions; size regulation in association with Cdcl3. Genes & Dev. S. Marcand for sharing unpublished results and ideas; and R. 11:512-527 . Rothstei n for the use of tetrad dissection microscopes. This Greider, C.W. 1996. Telomere length regulation. Annu. Rev, wor k was supported by grants from the National Institutes of Biochem. 65: 337-365 . Healt h (GM40094) and th e America n Cancer Society (VM-62) to Hardy, C.F.J. 1991. "Studie s on RAPl and th e RAPl interacting D.S. D.W. was a recipient of a long-term fellowship from the factor, RIFl. " Ph.D . thesis, Columbi a University, Ne w York, Huma n Frontiers in Science Program. NY. Th e publication costs of this article were defrayed in part by Hardy, C.F.J., D. Balderes, and D. Shore. 1992a. Dissection of a paymen t of page charges. This article mus t therefore be hereby carboxy-terminal region of the yeast regulatory protein marke d "advertisement" in accordance with 18 USC section RAP l with effects on both transcriptional activation and 1734 solely to indicate this fact. silencing. Mol. Cell. Biol. 12: 1209-1217. Hardy, C.F.J., L. Sussel, and D. Shore. 1992b. A RAPl-interact ing protein involved in silencing and telomere length regu References lation. Genes & Dev. 6: 801-814. Aparicio, O.M., B.L. Billington, and D.E. Gottschling. 1991. Hecht , A., T. Laroche, S. Strahl-Bolsinger, S.M. Gasser, and M. Modifiers of position effect are shared between telomeric Grunstein . 1995. Histone H3 and H4 N-termini interact and silent mating-type loci in S. cerevisiae. Cell 66: 1279- wit h SIR3 and SIR4 proteins: A molecular model for the 1287. formation of heterochromatin in yeast. Cell 80: 583-592. Blackburn, E.H. 1994. Telomeres : N o end in sight. Cell 77: 621 - Hecht , A., S. Strahl-Bolsinger, and M. Grunstein. 1996. Spread 623. ing of transcriptional repressor SIR3 from telomeric hetero Buchman , A.R., N.F. Lue, and R.D. Romberg. 1988. Connec chromatin . Nature 383: 92-96 . tion s between transcriptional activators, silencers, and telo Ivy, J.M., A.J.S. Klar, and J.B. Hicks . 1986. Cloning and charac mere s as revealed by functional analysis of a yeast DNA- terizatio n of four SIR genes from Saccharomyces cerevisiae. bindin g protein. Mol Cell. Biol. 8: 5086-5099. Mol. Cell. Biol. 6: 688-702. Buck, S.W. and D. Shore. 1995. Action of a RAPl C-terminal Klein, F., T. Laroche, M.E. Cardenas, J.F.-X. Hofmann, D. Sch- silencing domain reveals an underlying competition be weizer, and S.M. Gasser. 1992. Localization of RAPl and GENES & DEVELOPMENT 759 Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press Wotto n and Shore topoisomeras e II in nuclei and meiotic chromosomes of meri c DNA sequences into Saccharomyces cerevisiae re yeast. /. Cell Biol. 117: 935-948 . sult s in telomere elongation. Mol. Cell. Biol. 9: 1488-1497. Krauskopf, A. and E.H. Blackburn. 1996. Control of telomere Sandell, L.L. and V.A. Zakian. 1993. Loss of a yeast telomere: growt h by interactions of RAPl with the most distal telo- Arrest, recovery and chromosome loss. Cell 75: 719-7^9. meri c repeats. Nature 383: 354-357 . Schulz, V.P. and V.A. Zakian. 1994. The Saccharomyces PIFl Kurtz, S. and D. Shore. 1991. Th e RAPl protein activates and DN A helicase inhibit s telomere elongation and de nov o telo silences transcription of mating-type genes in yeast. Genes mer e formation. Cell 76: 145-155. &. Dev. 5: 616-628 . Shampay, J. an d E.H. Blackburn. 1988. Generation of telomere- Kyrion, G., K.A. Boakye, and A.J. Lustig. 1992. C-termina l trun lengt h heterogeneity in Saccharomyces cerevisiae. Proc. cation of RAPl results in the deregulation of telomere size, Natl. Acad. Sci. 85: 534-538 . stability, and function in Saccharomyces cerevisiae. Mol. Shore, D. 1994. RAPl: A protean regulator in yeast. Trends Cell. Biol. 12:5159-5173. Genet. 10: 408^12 . Kyrion, G., K. Liu, C. Liu, and A.J. Lustig. 1993. RAPl and . 1995. Telomere position effect and transcriptional si telomer e structure regulate telomere position effects in Sac lencing in th e yeast Saccharomyces cerevisiae. In Telomeres charomyces cerevisiae. Genes &. Dev. 7: 1146-1159. (ed. E.H. Blackburn and C.W. Greider), pp. 139-191. Cold Li, B. and A.J. Lustig. 1996. A novel mechanism for telomere Spring Harbor Laboratory Press, Cold Spring Harbor, NY. size control in Saccharomyces cerevisiae. Genes & Dev. Shore, D. and K. Nasmyth . 1987. Purification and cloning of a 10:1310-1326 . DN A binding protein from yeast that binds to both silencer Lin, J.J. and V.A. Zakian. 1995. An in vitro assay for Saccharo and activator elements. Cell 51: 721-732. myce s telomerase requires ESTl. Cell 81: 1127-1135. Sikorski, R. and P. Hieter . 1989. A syste m of shuttl e vectors and . 1996. The Saccharomyces CDC13 protein is a single- yeast host strains designed for efficient manipulation of stran d TGI-3 telomeric DNA-binding protein in vitro that DN A in Saccharomyces cerevisiae. Genetics \T1: 19-27. affects telomere behavior in vivo. Proc. Natl. Acad. Sci. Singer, M.S. andD.E . Gottschling. 1994. TLCl: Template RNA 93:13760-13765 . componen t of Saccharomyces cerevisiae telomerase. Sci Liu, C., X. Mao, and A.J. Lustig. 1994. Mutational analysis de ence 266:404-409. fines a C-terminal tail domain of RAPl essential for telo Sussel, L. and D. Shore. 1991. Separation of transcriptional ac meri c silencing in Saccharomyces cerevisiae. Genetics tivatio n and silencing functions of th e i?AP J-encode d repres 138:1025-1040 . sor/activato r protein 1: Isolation of viable mutant s affecting Longtine, M.S., N.M. Wilson, M.E, Petracek, and J. Herman. bot h silencing and telomere length. Proc. Natl. Acad. Sci. 1989. A yeast telomere binding activity binds to tw o related 88 : 7749-7753. telomer e sequence motifs and is indistinguishable from Thomas , B.J. and R. Rothstein. 1989. Elevated recombination RAPl . Curr. Genet. 16:225-240. rates in transcriptionally active DNA . Cell 56: 619-630. Lundblad, V. and J.W. Szostak. 1989. A mutan t with a defect in Tomow , J., X. Zeng, W. Gao, and G.M. Santangelo. 1993. GCRl , telomer e elongation leads to senescence in yeast. Cell a transcriptional activator in Saccharomyces cerevisiae, 57:633-643 . complexes with RAPl and can function without its DNA Lustig, A.J. and T.D. Petes. 1986. Identificatio n of yeast mutants binding domain. EMBO J. 12: 2431-2437. wit h altered telomere structure. Proc. Natl. Acad. Sci. Virta-Pearlman, V., D.K. Morris, and V. Lundblad. 1996. Estl 83:1398-1402 . has the properties of a single-stranded telomere end-binding Lustig, A.J., S. Kurtz, and D. Shore. 1990. Involvement of the protein . Genes 8k Dev. 10: 3094-3104 . silencer and UAS bindin g protei n RAPl in regulation of telo Wiley, E.A. and V.A. Zakian. 1995. Extra telomeres, but not mer e length. Science 250: 549-553 . interna l tracts of telomeric DNA, reduce transcriptional re Marcand, S., E. Gilson, and D. Shore. 1997. A protein-counting pression at Saccharomyces telomeres. Genetics 139: 67-79 . mechanis m for telomere length regulation in yeast. Science Wright, J.H., D.E. Gottschling, and V.A. Zakian. 1992. Saccha 275: 986-990. romyces telomeres assume a non-nucleosomal chromatin McEachern, M.J. and E.H. Blackburn. 1995. Runaway telomere structure . Genes & Dev. 6: 197-210. elongation caused by telomerase RNA gene mutations . Na Zakian , V.A. 1995a. Saccharomyces telomeres: Function, struc ture 376: 403-409. ture, and replication. In Telomeres (ed. E.H. Blackburn and Moretti , P., K. Freeman, L. Coodley, and D. Shore. 1994. Evi C.W. Greider), pp. 107-137. Cold Spring Harbor Laboratory dence that a complex of SIR proteins interacts with the si Press, Cold Spring Harobr, NY. lencer and telomere-binding protein RAPl. Genes &. Dev. . 1995b. Telomeres ; Beginning to understan d the end. Sci 8:2257-2269 . ence 270: 1601-1607. Nugent , C.I., T.R. Hughes, N.F. Lue, and V. Lundblad. 1996. Cdcl3p : A single-strand telomeric DNA-binding protein wit h a dual role in yeast telomere maintenance. Science 274 : 249-252. Palladino, F., T. Laroche, E. Gilson, A. Axelrod, L. Pillus, and S.M. Gasser. 1993. S1R3 and S1R4 proteins are required for th e positioning and integrity of yeast telomeres. Cell 75:543-555 . Rine, J. and I. Herskowitz. 1987. Four genes responsible for a positio n effect on expression from HML and HMR in Sac charomyces cerevisiae. Genetics 116: 9-22. Rose, M.D., F. Winston, and P. Hieter. 1990. Methods in yeast genetics: A laboratory course manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Runge, K.W. and V.A. Zakian. 1989. Introductio n of extra telo 760 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 3, 2021 - Published by Cold Spring Harbor Laboratory Press A novel Rap1p-interacting factor, Rif2p, cooperates with Rif1p to regulate telomere length in Saccharomyces cerevisiae. D Wotton and D Shore Genes Dev. 1997, 11: Access the most recent version at doi:10.1101/gad.11.6.748 This article cites 54 articles, 34 of which can be accessed free at: References http://genesdev.cshlp.org/content/11/6/748.full.html#ref-list-1 License Receive free email alerts when new articles cite this article - sign up in the box at the top Email Alerting right corner of the article or click here. Service Copyright © Cold Spring Harbor Laboratory Press
Journal
Genes & Development – Unpaywall
Published: Mar 15, 1997