Molecular basis for telomere repeat divergence in budding yeast

Abstract

Telomerase is a ribonucleoprotein enzyme that adds repetitive sequences to the ends of linear chromosomes, thereby counteracting nucleotide loss due to incomplete replication. A short region of the telomerase RNA subunit serves as template for nucleotide addition onto the telomere 3' end. Although Saccharomyces cerevisiae contains only one telomerase RNA gene, telomere repeat sequences are degenerate in this organism. Based on a detailed analysis of the telomere sequences specified by wild-type and mutant RNA templates in vivo, we show that the divergence of telomere repeats is due to abortive reverse transcription in the 3' and 5' regions of the template and due to the alignment of telomeres in multiple registers within the RNA template. Through the interpretation of wild-type telomere sequences, we identify nucleotides in the template that are not accessible for base pairing during substrate annealing. Rather, these positions become available as templates for reverse transcription only after alignment with adjacent nucleotides has occurred, indicating that a conformational change takes place upon substrate binding. We also infer that the central part of the template region is reverse transcribed processively. The inaccessibility of certain template positions for alignment and the processive polymerization of the central template portion may serve to reduce the possible repeat diversification and enhance the incorporation of binding sites for Rap1p, the telomere binding protein of budding yeast.

FIG. 1

(A) The telomere consensus sequence does not correspond to perfect repeats of the sequence specified by the TLC1 template RNA. (B) TLC1 template mutants used in this study. Mutant nucleotides are highlighted. In order to facilitate the cloning, the NcoI restriction site 3′ of the template was destroyed in the mutants by ligation with compatible BspHI ends. This results in a single ⁴⁵⁵A→G nucleotide change.

FIG. 2

At template position ⁴⁷⁰C, the reverse transcription can be continued (upper right) or aborted (bottom). With the ⁴⁶⁹A→U mutant telomerase, the different possibilities can be distinguished based on the incorporated telomere sequence (indicated below the TLC1 template sequence; newly incorporated nucleotides are highlighted). For reasons of simplicity, all 5′-TGGGTGTGGGT-3′ sequences were attributed to alignment at position ⁴⁷⁶C (bottom left). However, product dissociation at position ⁴⁷³C followed by realignment with ⁴⁷⁹C would also result in the incorporation of the same telomeric sequence. Therefore, the number given on the bottom left is probably an overestimate of the alignment frequency at ⁴⁷⁶C. The number of events indicated for alignment 3′ of position ⁴⁷⁹C (bottom right) comprises both position ⁴⁸³C and position ⁴⁸¹C. Only the latter is represented in the scheme. Repeats not listed in the figure resulted from product dissociation 3′ of position ⁴⁷⁰C.

FIG. 3

Alignment possibilities for telomeric 3′ ends containing a GG dinucleotide along the WT template region. The alignment positions within TLC1 are indicated on the left. Underlined nucleotides indicate positions of ambiguity for the alignment (see text). Incorporated telomere sequences resulting from the alignment indicated on the left are highlighted in the middle. Relative frequencies of the indicated telomere sequence in the pooled WT telomeres (genome sequence, WT TLC1, and tlc1 ΔNcoI) are indicated on the right. n.d., not determined.

FIG. 4

Telomerase assays using WT and mutant telomerases in vitro. Reaction mixtures contained equal amounts of TLC1 RNA. 5′-GTGTGTGTGGG-3′ was used as a substrate. (A) Lane 1, substrate extended with [α-³²P]ddATP using terminal deoxynucleotidyltransferase; lane 2, telomerase reaction from WT cells pretreated with RNase A; and lanes 3 to 11, reactions performed with mutant telomerases as indicated in panel B. The open arrowhead indicates the position of a labeled 32-mer oligonucleotide that served as control for precipitation efficiency and gel loading. The filled arrowhead indicates extension to position ⁴⁷⁰C. (B) Relative amounts of +1 to +4 products. The different numbers of incorporated radioactive nucleotides were taken into account.

FIG. 5

Model of the S. cerevisiae telomerase reaction cycle. (State 1) Telomerase before substrate binding. Template positions ⁴⁷⁹CAC⁴⁷⁷ and ⁴⁷³CAC⁴⁷¹ are not accessible for alignment. (State 2) Upon substrate binding, a conformational change occurs and all template positions are available for base pairing during reverse transcription. Dissociation of a partially extended telomere occurs with moderate frequency. (State 3) The central part of the TLC1 template region is reverse transcribed processively. The results presented in this paper exclude product dissociation at position ⁴⁷⁶C. However, the region of processive reverse transcription may cover template positions ⁴⁷⁸A to ⁴⁷²A to favor incorporation of the telomeric 5′-TGGGTGT-3′ core sequence. (State 4) Template positions close to the 5′ boundary are reverse transcribed with moderate frequency. Only half of the products are extended beyond template position ⁴⁷¹C.