Telomerase biogenesis requires a novel Mex67 function and a cytoplasmic association with the Sm₇ complex

Acceptance summary:

Using an induced Cre-dependent recombination system to replace expression of the endogenous TLC1 gene by an MS2-tagged TLC1 RNA, the authors show that the mRNA export receptor Mex67 is essential for TLC1 RNA biogenesis and stability of newly synthesized transcripts in the nucleus before they are exported by Xpo1, and that MS2-tagged the TLC1 3' mutant unable to bind the Sm₇ complex crucial for processing are unstable and mainly detected in the cytoplasm, indicating that the Sm₇-dependent processing steps occur in the cytoplasm. The manuscript provides new and relevant information to understand telomerase biogenesis related to a new role for the Mex67 RNA export factor.

Decision letter after peer review:

Thank you for submitting your article "Telomerase biogenesis requires a novel Mex67 function and a cytoplasmic association with the Sm₇ complex" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Andrés Aguilera as the Reviewing Editor and James Manley as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Francoise Stutz (Reviewer #1); Vincent Geli (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

Yeast TLC1 RNA biogenesis and processing have mostly been analyzed so far by examining the effect of a variety of mutations on endogenous steady state TLC1 RNA pools. This new study uses an induced Cre-dependent recombination system to replace expression of the endogenous TLC1 gene by an MS2-tagged TLC1 RNA, so that they can distinguish the old untagged TLC1 from the newly synthesized TLC1. This elegant approach allows a detailed description of the steps leading to the formation of a mature TLC1 RNA in the nucleus, and finally to its degradation in the cytoplasm. The study confirms that the export receptor Xpo1 is the only one involved in the nuclear export of TLC1 transcripts, which are then reimported via Kap122 but probably also other import receptors. Importantly, the authors show that the mRNA export receptor Mex67 is essential for TLC1 RNA biogenesis and stability of newly synthesized transcripts in the nucleus before they are exported by Xpo1. Finally, the authors show that MS2-tagged TLC1 3' mutant unable to bind the Sm₇ complex crucial for processing are unstable and mainly detected in the cytoplasm, indicating that the Sm₇-dependent processing steps occur in the cytoplasm and are essential for reimport into the nucleus.

Although several molecular details underlying TLC1 biogenesis have still to be worked out, this study provides a novel more complete view of the TLC1 RNA life cycle with well-designed and rigorously controlled experiments. The manuscript is very clearly written and referenced and is appropriate for publication, provided they respond to the questions raised below.

Essential revisions:

1) Several lines of evidence indeed suggest that TLC1-MS2 is not fully functional: TLC1-MS2 RNA is less abundant (Figure 1E), form less foci (Figure 2B) and is trapped in the cytoplasm for a long period after recombination (Figure 2E). Accordingly the telomeres are short in TLC1-MS2 (Figure 2—figure supplement 1) and in contrast to cells expressing TLC1-Sm2T (Seto et al., 1999), TLC1-Sm2T-MS2 cells senesce and eventually form type II survivors. Could the authors argue about this point?

2) The claim that the Mex67 effect is independent of transcription regulation seems to rely basically on the use of thiolutin as a transcription blocker in an "epistasis" type of experiment. Since it is assumed that TLC1 will get degraded in the nucleus in the mex67-5 mutant, would it be possible to inactivate either the nuclear exosome (Rrp6) or Rat1 and somehow more convincingly demonstrate that it is truly a degradation machinery that is responsible for the loss of TLC1. Alternatively the authors could try to demonstrate (through RNAPII ChIP) that transcription is not altered in the mex67-5 mutant.

3) Could the authors demonstrate that the increased "new" nuclear TLC1 in xpo1-1 mutants is decreased in a mex67-5 mutant.

Essential revisions:

1) Several lines of evidence indeed suggest that TLC1-MS2 is not fully functional: TLC1-MS2 RNA is less abundant (Figure 1E), form less foci (Figure 2B) and is trapped in the cytoplasm for a long period after recombination (Figure 2E). Accordingly the telomeres are short in TLC1-MS2 (Figure 2—figure supplement 1).

This first part of point 1) questions whether tagging the TLC1 RNA leads to a loss of function. As we mentioned and now elaborate in more detail (subsection “Tracking subcellular distribution of telomerase RNA fractions”, third paragraph), the tagged RNA does accumulate to a slightly lower level than the WT. We don’t know the reason for this but wish to emphasize that the slightly lower level of the RNA does not mean loss of functionality of the telomerase RNP. We carried out extensive control experiments on this issue (all published in Gallardo et al., 2011, and Bajon et al., 2015). In all assays, we were unable to detect any loss in functionality of the RNP per se, including telomerase enzymatic activity (Gallardo et al., 2011). Therefore, we concluded in these papers before and here too that the RNP per se is active as WT, but the RNA is somewhat less abundant, explaining the slightly shorter telomeres. For this manuscript here particularly important, the nucleo-cytoplasmic partitioning of the tagged RNA is exactly the same as for WT untagged RNA (see Figure 2—figure supplement 1E).

And in contrast to cells expressing TLC1-Sm2T (Seto et al., 1999), TLC1-Sm2T-MS2 cells senesce and eventually form type II survivors. Could the authors argue about this point?

Now, this second part of point 1) is a different issue. The reason why the cells in the referenced Seto/Cech study of ‘99 look like they maintain telomeres with the tlc1-Sm2T RNA molecule is twofold:

– First, Seto/Cech use a tlc1Δ::LEU2 strain in which they express the various TLC1 alleles, including the tlc1-Sm2T allele, from a plasmid borne gene copy. It is well known that plasmid-based gene expression results in an overexpression of the gene in question. Thus, they slightly overexpress the tlc1-Sm2T allele. All our genes are expressed from the genomic locus of TLC1.

– Secondly, the strain used in the Cech study is also rad52Δ, which thus cannot bypass telomerase problems with HR mediated telomere maintenance. We observe that tlc1-Sm2T expressing strains that are RAD52wt will senesce and form survivors (see Figure 5—figure supplement 1D). Hence, the apparent difference in the results is explained by higher expression and a strong selection of viable cells in the Seto et al., 1999, experiments. We actually think, but have not tested, that expressing the tlc1-Sm2T allele from the genomic locus in a RAD52wt strain would also yield survivors.

2) The claim that the Mex67 effect is independent of transcription regulation seems to rely basically on the use of thiolutin as a transcription blocker in an "epistasis" type of experiment. Since it is assumed that TLC1 will get degraded in the nucleus in the mex67-5 mutant, would it be possible to inactivate either the nuclear exosome (Rrp6) or Rat1 and somehow more convincingly demonstrate that it is truly a degradation machinery that is responsible for the loss of TLC1. Alternatively the authors could try to demonstrate (through RNAPII ChIP) that transcription is not altered in the mex67-5 mutant.

This indeed is an important point with great suggestions for experiments, thanks a bundle for that. We actually carried out both experiments. First, the experiment with a mex67-5 rrp6Δ double mutant strain at restrictive temperature bears out precisely what the reviewer predicted, namely complete suppression of the new RNA loss phenotype in the nucleus (new Figure 4C, E and Figure 4—figure supplement 1F-H). I am still amazed of how well this worked. Secondly, we did the RNA Pol II ChIP and compared Pol II loading on TLC1 at permissive vs. restrictive temperature for the mex67-5 cells. The results show pretty much equal Pol II loading and hence again fully support the notion that transcription is not affected in this condition (new Figure 4—figure supplement 1E). Both experiments are described in the subsection “The nuclear stability of new telomerase RNA transcripts requires Mex67”. Adding these two results to our story allows us to quite confidently conclude that the disappearance of the newly transcribed TLC1 RNA molecules in the mex67-5 cells at high temperature is due to degradation and not a loss of transcription.

3) Could the authors demonstrate that the increased "new" nuclear TLC1 in xpo1-1 mutants is decreased in a mex67-5 mutant.

In essence, the question is whether the mex67-5ts is epistatic to the xpo1-1ts. After much effort and way too much time, we actually were able to generate the double mutant. In brief, the result is exactly as the reviewer suspected: the mex67-5ts xpo1-1ts strain shows the same phenotype as the mex67-5ts alone and mex67-5ts thus is epistatic to xpo1-1ts (see new Figure 4—figure supplement 2A-C; subsection “The nuclear stability of new telomerase RNA transcripts requires Mex67”).