Distinguishing between recruitment and spread of silent chromatin structures in Saccharomyces cerevisiae

Essential revisions:

All reviewers have found this manuscript interesting and the data of high quality. However, several concerns are raised and discussed among reviewers. All reviewers agree that the following two control experiments are essential for the revision.

1) The overexpression of Sir3 should be confirmed using western blot analysis.

We have added a Western blot for the Sir3-M.EcoGII overexpression experiments as a new Figure 3B (and added the source data with full uncropped blots as Figure 3–Source Data 1) along with the following text in the results:

“We confirmed the overexpression of SIR3-M.ECOGII in these strains by protein immunoblotting (Figure 3B)”

We have also updated the Materials and Methods to add the new antibody we used. (Thanks to Namrita Dhillon and Rohinton Kamakaka for the Sir3 antibody we used for the new Western and ChIP-seq experiments!)

2) Figure 3 – In the overexpression experiments, the authors are making conclusions about heterochromatin formation based on the level of methylation. It seems important to look at the different classes of telomeres by more than one method in order to fortify and expand this conclusion. First, looking at the level of recruitment of the Sir3-M.EcoGII by ChIP would tell us if the difference in methylation is caused at the level of recruitment. This is a Class A experiment as either result is interesting – either there's differential recruitment of Sir3 between chromosome 6 and the others or there's differential ability of Sir3 to spread/methylate. Second, if Sir3-M.EcoGII is recruited at a higher level, it would be worthwhile to look at the possible silencing of genes near the telomeres to see if their expression is reduced. This is especially important given the results presented in Figure 4 where a small change in association appears to increase silencing.

We have followed the reviewers’ suggestion and backed up the methylation data upon overexpression of Sir3-EcoGII with ChIP data on the binding of Sir3-EcoGII. We feel confident concluding that the ChIP-seq results at HML, HMR, and telomeres parallel the Nanopore data quite well, suggesting that the increase in methylation signal by Nanopore sequencing at telomere 6R is due to Sir3 binding/contacting chromatin more frequently.

We have included ChIP-seq results for the overexpression experiments in a new Figure 3—figure supplement 1 and added the following text to the Results section:

“The changes in methylation upon overexpression of Sir3-M.EcoGII matched changes in occupancy of the fusion protein by ChIP-seq, suggesting that methylation reflects binding of the protein well (Figure 3—figure supplement 1).”

“As at HML and HMR, the changes in methylation seen with Nanopore sequencing matched with changes in occupancy measured by ChIP-seq (Figure 3—figure supplement 1).”

We have also updated the Materials and Methods to add the new antibody we used and submitted the new ChIP-seq data to the GEO to add to accession 189038.

The reviewer invited us to look at the impact of greater Sir3 spreading on gene expression. It turns out that nearly the entire body of knowledge on the impact of Sir3 overexpression on repression of telomeric genes comes from telomere 6R, which is one of those our work here established the physical basis of. In earlier work from the lab mapping Sir protein distribution and silencing on all yeast telomeres (Ellahi et al., 2015), we established that there are few subtelomeric genes susceptible to silencing, and most of these are in gene families, making it difficult to attribute the modest effects observed to any one copy of that gene.

Please also address or respond to all the other concerns and suggestions made by reviewers in your revised manuscript and in your response to reviews.

Reviewer #1 (Recommendations for the authors):

1. As the authors mentioned, the Sir3 BAH domain deletion results in silencing defects and lack of Sir3 spreading. Importantly, the study reports that sir3-bah∆ can still bind to silencers/silencing nucleation sites. So it appears that transcription silencing is associated with Sir3 spreading. Interestingly, in the Sir3 temperature sensitive experiment, the authors showed that transcriptional repression is already established prior to Sir3 spreading. How would the authors interpret the relationship between Sir3 spreading and transcription silencing? What are the roles of the BAH domain in transcriptional repression beyond Sir3 spreading?

Indeed, sir3-bah∆ resulted in silencing defects at both HML and HMR, but there was still a small amount of repression at HML and an approximately 10-fold repression at HMR compared to sir3∆ (Figure 2C). In the Discussion section, we discussed the results of the bah∆ and time course experiments and concluded that some repression is possible without Sir3 spreading, but spreading is required for full transcriptional silencing.

2. The authors described several identified functional domains in Sir3. In addition to the BAH domain, there are AAA+ and wH domains on Sir3. However, only BAH domain deletion was investigated. Could the functions of the other domains be investigated using this Sir3- M.EcoGII approach? The AAA+ domain interacts with Sir4 and the wH domain mediates dimerization, hence both are relevant in the investigation of Sir3 mediated silencing and spreading.

Great point! We had the same thought during this project and attempted to make and test various mutants in the BAH, wH, and AAA+ domains of Sir3: D17G (in BAH, no/little interaction with nucleosomes), K657A/K658A (in AAA+, no interaction with Sir4), and wH∆ (no dimerization). Unfortunately, these approaches failed for various reasons. In particular, the D17G and wH∆ mutants we made resulted in lower or completely absent Sir3 protein levels (by Western, Author response image 1). Previous work using these mutants showed wild-type levels of Sir3 (wH∆: Oppikofer et al., 2013, D17G: Buchberger et al., 2008, K657A/K658A: Ehrentraut et al., 2011). We are unsure why our results are not consistent with these findings. The above mutant alleles were expressed from single-copy plasmids, whereas ours were expressed from the endogenous SIR3 locus, so perhaps therein lies the difference.

Author response image 1

Download asset Open asset

Reviewer #3:

1. lines 114-117 and Figure 1B – It's not clear why silencing was assayed by fluorescence here, when it was assayed by RT-qPCR everywhere else. If there is an advantage, please provide it. Otherwise, it would be better to include RT-qPCR here as well to provide a basis for comparison with the later experiments. Also, some readers will not be familiar with using fluorescent reporters to assay silencing so if this stays, the authors should cite Dodson and Rine and give an estimate of the sensitivity of this assay compared to the older methods of measuring RNA levels or mating.

We have replaced the fluorescent assay results in Figure 1B with RT-qPCR results for each strain. The take-home message is the same, but the reviewer makes the point that the RT-qPCR data are probably more accessible to a wider audience.

2. line 157, Figure 1D, E – The authors discuss a periodicity of reads as shown in Figure 1E. However, there is also a periodicity evident in Figure 1D. Do the peaks from the two methods match up? Please address this in this section.

Thank you for pointing this out. You are correct that there appears to be a periodicity in Figure 1D, but the peaks in 1D are not the same ‘peaks’ as in 1E. Otherwise, there would be many more peaks in 1D. The data visualization in 1D (and all of the other aggregate nanopore data figures) is achieved using by applying a loess function, and thus the apparent periodicity in Figure 1D (and others) is likely a result of the same phenomenon seen in the single read data. If we zoom into the aggregate data and increase the granularity of the loess smoothing, more peaks and valleys appear in the aggregate data that match up with linker regions (Figure 1—figure supplement 3A).

3. Figure 3 – In the overexpression experiments, the authors are making conclusions about heterochromatin formation based on the level of methylation. It seems important to look at the different classes of telomeres by more than one method in order to fortify and expand this conclusion. First, looking at the level of recruitment of the Sir3-M.EcoGII by ChIP would tell us if the difference in methylation is caused at the level of recruitment. This is a Class A experiment as either result is interesting – either there's differential recruitment of Sir3 between chromosome 6 and the others or there's differential ability of Sir3 to spread/methylate. Second, if Sir3-M.EcoGII is recruited at a higher level, it would be worthwhile to look at the possible silencing of genes near the telomeres to see if their expression is reduced. This is especially important given the results presented in Figure 4 where a small change in association appears to increase silencing.

This point was addressed above in the “Essential Revisions” section. We reiterate the resulting changes here:

“The changes in methylation upon overexpression of Sir3-M.EcoGII matched changes in occupancy of the fusion protein by ChIP-seq, suggesting that methylation reflects binding of the protein well (Figure 3—figure supplement 1).”

“As at HML and HMR, the changes in methylation seen with Nanopore sequencing matched with changes in occupancy measured by ChIP-seq (Figure 3—figure supplement 1).”

4. lines 282-283, Figure 4B – The authors wrote that, after the 37 to 25 shift, the Sir3-8 protein levels grew "to match" the constitutive 25C levels. However, their results show the level at 150 minutes to be 0.63 of the constitutive level. I suggest revising and simply giving the real value in the text, something like, "Over the course of 150 minutes, the protein levels of sir3-8 slowly increased to approximately 63% that of the level in constitutive 25C growth conditions (Figure 4B)."

We modified the text to include this more specific statement.