The yeast telomere length counting machinery is sensitive to sequences at the telomere-nontelomere junction

Abstract

Saccharomyces cerevisiae telomeres consist of a continuous 325 +/- 75-bp tract of the heterogeneous repeat TG1-3 which contains irregularly spaced, high-affinity sites for the protein Rap1p. Yeast cells monitor or count the number of telomeric Rap1p molecules in a negative feedback mechanism which modulates telomere length. To investigate the mechanism by which Rap1p molecules are counted, the continuous telomeric TG1-3 sequences were divided into internal TG1-3 sequences and a terminal tract separated by nontelomeric spacers of different lengths. While all of the internal sequences were counted as part of the terminal tract across a 38-bp spacer, a 138-bp disruption completely prevented the internal TG1-3 sequences from being considered part of the telomere and defined the terminal tract as a discrete entity separate from the subtelomeric sequences. We also used regularly spaced arrays of six Rap1p sites internal to the terminal TG1-3 repeats to show that each Rap1p molecule was counted as about 19 bp of TG1-3 in vivo and that cells could count Rap1p molecules with different spacings between tandem sites. As previous in vitro experiments had shown that telomeric Rap1p sites occur about once every 18 bp, all Rap1p molecules at the junction of telomeric and nontelomeric chromatin (the telomere-nontelomere junction) must participate in telomere length measurement. The conserved arrangement of these six Rap1p molecules at the telomere-nontelomere junction in independent transformants also caused the elongated TG1-3 tracts to be maintained at nearly identical lengths, showing that sequences at the telomere-nontelomere junction had an effect on length regulation. These results can be explained by a model in which telomeres beyond a threshold length form a folded structure that links the chromosome terminus to the telomere-nontelomere junction and prevents telomere elongation.

FIG. 1

Introduction of synthetic telomeres into yeast. (A) Replacement of the left telomere of chromosome VII by homologous integration at the ADH4 locus, using a synthetic telomere adjacent to URA3. Different inserts, shown in panel B, were cloned between URA3 and the TG_1-3 sequences to attempt to alter telomere length regulation. (B) The synthetic telomeres used in this work. The nontelomeric spacers between the TG_1-3 tracts are described in Materials and Methods. The orientation of the TG_1-3 repeats is indicated by arrowheads. YIpADH652-42 has the internal 256-bp TG_1-3 tract in the reverse orientation relative to the terminal 29-bp tract. The specific YIpADHTEF and YIpADHλ telomeres are presented in Fig. 5. (C) Sequences of the TG_1-3 repeats and spacer in the YIpADH256-50 construction. The sites used to construct the plasmids in panel B are indicated. The TG_1-3 repeats are from positions 54 to 309 in the 5′ KpnI site, and the single A in the TG_1-3 repeats is at position 98.

FIG. 2

Some of the internal TG_1-3 sequences are considered part of the telomere across a 50-bp nontelomeric spacer. (A) Predicted sizes of the telomere restriction fragments on Southern blots. Telomere length, i.e., the length of the elongated terminal TG_1-3 tract, was determined by subtracting the StuI-BamHI fragment length from the StuI fragment length measured on the same blot. Analysis on the same blot was important for giving reproducible and accurate length determination for different control telomeres. With lanes 14 and 15 of B as an example, the length of the elongated terminal TG_1-3 sequences (the ? in the diagram) is 150 bp. (B) Representative genomic DNAs from cells bearing the YIpADH256-26, -38, and -50 telomeres and the YIpADH652-42 telomere were cleaved with StuI or StuI plus BamHI and analyzed by Southern blotting using the URA3 fragment in panel A as probe. All telomeres were formed in yeast strain YM708. The 0 spacer is YIpADH35. Each pair of lanes (indicated by a bar under the number) represents an individual transformant. Lane 1, YM708 with no synthetic telomere; lanes 10 and 11, a YIpADH256-50 transformant with short telomeres; lanes 14 and 15, a YIpADH256-50 transformant with long telomeres; lanes 12 and 13, a YIpADH256-50 transformant with short and long telomeres (mixed telomeres). Note that the YIpADH652-42 construction places the BamHI site closer to URA3 (Fig. 1B). Telomere fragment sizes were determined as described in Materials and Methods.

FIG. 3

A 138-bp nontelomeric spacer establishes a new telomere-nontelomere junction. (A) Individual YM708 transformants bearing the YIpADH256-138 telomere were analyzed as for Fig. 2. Lanes 6 and 7, a YIpADH256-138 transformant with long telomeres; lanes 10 to 13, transformants with short telomeres; lanes 8 and 9 and 14 to 19, transformants with short and long telomeres. The transformant in lanes 8 and 9 had lost the BamHI site in the long telomeres and was not used for the analysis in panel C. The chevron-headed arrow indicates the long telomeres, the filled arrowhead indicates the StuI-BamHI long telomere fragment, and the hollow arrowhead indicates the short telomeres. Lane 1 contains DNA from the untransformed strain. (B) Individual transformants bearing the YIpADH256-266 telomere were analyzed as for Fig. 2. Lanes 10, 11, 14, 15, and 18 show transformants with short telomeres, while lanes 6 to 9, 12, 13, 16, and 17 show transformants with short and long telomeres. Arrows are used as in panel A. Lane 1 contains DNA from the untransformed strain. (C) The amount of the internal 256-bp tract counted as part of the elongated TG_1-3 tract plotted against the length of the nontelomeric spacer. The gray box represents the length heterogeneity due to normal variation and was calculated by using seven independent YIpADH35 telomeres. The amount of internal TG_1-3 counted was determined by subtracting the average length of the terminal TG_1-3 tract distal to the BamHI site for different nontelomeric spacer telomeres (the ? in Fig. 2A) from the average modal length of the TG_1-3 tract for the control YIpADH35 telomere (Fig. 2A). The standard error of the telomere length measurements for each spacer telomere was ±10% except for YIpADH652-42 (±53%).

FIG. 4

The transformants with mixed telomeres are colonies of cells with short or long telomeres. (A) Hypothesis for the formation of single transformant colonies bearing mixed telomeres and the predicted outcomes. After integration of the construction, the cell replicates the telomere and divides prior to either elongating the terminal 29 bp TG_1-3 tract or deleting the terminal 29-bp tract and elongating the internal 256-bp tract. After these events, the telomeres are stably maintained in separate cells within the same colony. Subclones derived from single cells from these individual transformants bearing two types of telomeres should have either short or long telomeres but never both. (B) Genomic DNAs from single colonies derived from a single YIpADH256-50 transformant bearing either mixed telomeres (lanes 2 to 11) or only long telomeres (lanes 12 to 16). DNAs were cleaved with PstI, which cleaves in URA3 (see panel A) and analyzed as for Fig. 2. Lane 1 is DNA from a cell transformed with YIpADH35. (C) Genomic DNAs from a single colony derived from a single YIpADH256-138 transformant bearing long and short telomeres (Fig. 3A, lane 14) cleaved with StuI and analyzed as for Fig. 2 (lanes 2 to 15). Lane 1 is DNA from a cell transformed with YIpADH35; lanes 16 to 18 are single-cell subclones of the short telomere transformant in Fig. 3A, lane 12. In panels B and C, the chevron-headed arrows indicate long telomeres that retain the nontelomeric spacer and hollow arrows indicate short telomeres that have lost the spacer. All transformants were in strain YM708.

FIG. 5

Cells count one Rap1p molecule as ∼19 bp of TG_1-3. (A) Telomere constructions contained either phased arrays of Rap1p sites internal to the 29 bp TG_1-3 tract or the same-size fragment of λ DNA. The orientation of the Rap1p site array is indicated by the arrowheads as in Fig. 1. The locations of the BamHI site and the URA3 fragment used to probe the Southern blots are shown. (B) Genomic Southern analysis of telomeres bearing six Rap1p sites spaced one site every 18 bp in strain KR36-6L. DNAs were digested with StuI or StuI plus BamHI and analyzed by Southern blotting using the URA3 probe shown in panel A. Two YIpADHλ108 telomeres were included as negative controls for telomeres bearing inserts of equal size with no Rap1p sites. The leftmost lane of B and C is DNA from the untransformed KR36-6L strain; 35 denotes YIpADH35 (Fig. 1B). (C) Genomic Southern analysis of telomeres bearing six Rap1p sites spaced one site every 35 bp in strain KR36-6L analyzed as for panel B (reprinted from reference with permission). A YIpADHλ210 telomere was included as a negative control for an insert of equal size with no Rap1p sites. (D) Genomic Southern analysis of telomeres in strain KR36-6L bearing six Rap1p sites spaced one site every 35 bp. All DNAs were digested with StuI. All lanes are from the same gel. (E) Genomic Southern analysis of telomeres in strain KR36-6L bearing six Rap1p sites spaced one site every 13 bp. A YIpADHλ88 telomere was included as a negative control for an insert of nearly equal size with no Rap1p sites.

FIG. 6

Phased arrays of Rap1p sites eliminate the length variation between transformants. (A) Structures of steady-state telomeres in independent transformants derived from the YIpADH35 (35) and YIpADHTEF (TEF) constructions. When telomeres are formed with YIpADH35, 90% of the telomere consists of newly synthesized TG_1-3 repeats. Since the TG_1-3 sequences added after the first 11 bp are random (18), each of the independent YIpADH35 transformants will have a different telomere sequence near the telomere-nontelomere junction. In contrast, all of the YIpADHTEF telomeres retain the same nontelomeric Rap1p sites at the junction. The terminal 120 to 150 bp of TG_1-3 sequences for both the YIpADH35 and YIpADHTEF telomeres are randomized during cell growth (48), and so the only differences between the YIpADH35 and YIpADHTEF telomeres are the sequences near their telomere-nontelomere junctions. (B) Genomic DNAs digested with StuI from seven independent YIpADH35 transformants, five independent YIpADHTEF35-6TG transformants, and four independent YIpADHTEF35-6CA transformants (all in strain KR36-6L) were analyzed by Southern blotting as for Fig. 5. All lanes are from the same gel. The length of each telomere was measured at the point indicated by the small white bar in each lane, which is the point of most intense hybridization (the modal telomere length). The lengths of the YIpADH35 telomeres varied over a ∼60-bp range, while the lengths of most of the YIpADHTEF35-6 telomeres were within a ∼10-bp range (see text). (C) The same DNAs as in panel B were digested with StuI and BamHI to show that the length heterogeneity between the independently formed telomeres was due to different lengths of the terminal TG_1-3 sequences.

FIG. 7

A working model for telomere length regulation. Telomere length is monitored by counting Rap1p molecules to keep telomeres within a set range of lengths. We propose that Rap1p molecules are counted by the formation of a transient, highly folded three-dimensional structure stabilized by many weak interactions between Rap1p molecules and negative regulators of telomere length (drawn here as a hairpin for simplicity). When the newly replicated TG_1-3 repeats are 325 ± 75 bp, a highly folded structure that links the telomere-nontelomere junction to the chromosome terminus forms. This structure blocks telomerase access to the chromosome terminus, and thus telomere elongation, either by sequestering the end or by recruiting an inhibitory complex (e.g., a complex formed between the Rap1p C terminus, Cdc13p, and Stn1p [9, 34]). When telomeres are short, the structure is not formed, no inhibition occurs, and telomeres are elongated. Lengthening of short telomeres would be enhanced by the Rap1p molecules which are no longer sequestered by the folded structure (38). Subsequent to these events, telomeric heterochromatin is formed and maintained until the next S phase. A simpler model that linked the telomere-nontelomere junction to the chromosome terminus without counting the intervening sequences had been previously proposed for K. lactis telomeres (31).