Additional Issues and Analysis

Histone H4 depletion

1. Do the results from the histone H4 depletion experiment reflect the effects of histone loss per se, or a homeostatic response to the loss of a specific histone?
   The nucleosome levels are first detected to be significantly lower 1 hour into the depletion experiment (17), coinciding with the onset of gene expression changes. The changes in gene expression levels we observe over the time course closely parallel the changes in nucleosome levels. Thus, we cannot eliminate the possibility that some effects are due to a homeostatic response to the loss of histone H4, but this seems unlikely to account for most of the results.
   
   
2. When nucleosome density is low, do nucleosomes move about? Does this explain why the % of genes repressed is almost as large as the % of genes activated?
   The depletion of histone H4 results in a 50-60% loss of nucleosomes (17). The remaining nucleosomes appear to be distributed randomly along the DNA (16). However we do not believe that the distribution and/or mobility of nucleosomes upon histone H4 depletion explains why the percent of genes repressed is almost as large as the percent of genes activated. This hypothesis would predict that the genes affected by histone depletion would vary between experiments, since the results would depend upon the random distribution of nucleosomes. However, we have compared separate, independent experiments of histone H4 depletion (see manuscript, page 3, paragraph 1) and the results are very reproducible, arguing against this hypothesis.
   
   
3. Why are the kinetics of gene activation and repression different in the histone H4 depletion time course?
   Examination of the kinetics of changing mRNA levels in Figure 1 reveals that the increases in mRNA levels were generally apparent 60 minutes prior to decreases in mRNA levels. This is the expected result if induction of one set of genes and loss of expression of the other set occurred simultaneously. An increase in transcription can be detected almost immediately, particularly for genes normally expressed at low levels, whereas the loss of transcription is detected only after significant decay of existing mRNA has occurred.
   
   
4. Are genes previously implicated in telomeric silencing affected by histone H4 depletion?
   SIR2,3,4 and RAP1 transcript levels are not significantly affected by histone H4 depletion. The yeast Ku genes are induced 3-fold and CAC1 is induced 20-fold by histone H4 depletion. No other known telomeric silencing factors, including RIF1, SAS2,3; RAD6, TEL1,2; and RPD3 (6,12), appear to be affected by histone H4 depletion.
   
   
5. How do genes previously identified as upregulated upon histone H4 depletion behave in the microarray data?
   Five genes have been previously identified as being upregulated upon histone H4 depletion. These genes are GAL1, PHO5, CYC1, HIS3, and CUP1 (15-16,18). The microarray data indicates that GAL1 is induced 6-fold at the 6-hour time point, and PHO5 is unaffected at the 6-hour time point, but is induced 4-fold at the 4-hour time point. The HIS3 gene was unaffected in our experiment, but the growth conditions used here differ from those used previously. The two remaining genes, CYC1 and CUP1, were previously shown to be unaffected by histone H4 depletion when the genomic wild-type gene was examined (15,18).
   
   
6. How many genes are affected by histone H4 depletion if we decrease the stringency of the analysis by lowering the fold change cutoff to 2 fold?
   Using a less stringent 2-fold cutoff in analyzing the 6 hour histone H4 depletion time point gives the following results: 1258 genes derepressed, 1263 genes repressed, and 3373 genes unaffected by histone H4 depletion. Thus, even with less stringent criteria for determining the number of genes affected by histone H4 depletion, a majority of genes (57%) are unaffected by histone H4 depletion.
   
   
7. Are genes that are downregulated by histone H4 depletion anything other than the effects of slow growth or death?
   While the literature clearly documents instances in which nucleosomes potentiate the expression of some genes, it is likely that a substantial number of genes that are downregulated by histone H4 depletion are due to secondary effects such as slowed or arrested growth of the mutant cells. We have tried to determine the degree to which slow growth and cell-cycle phenotypes exhibited by cells undergoing histone depletion contribute to changes in gene expression with two approaches. First, we compared our data to the diauxic shift experiment (25) and found that approximately 46% (260/569) of the genes repressed by histone depletion are also repressed >2-fold as cells slow in growth during the final stage (OD600 = 7.3) of the diauxic shift, and thus can be attributed to indirect effects of the histone depletion.
   
   Second, we performed the following experiment to control for cell-cycle effects. Wild-type cells were arrested in G2/M using the drug Nocadozole. The cells were harvested 4 hours after the addition of drug, and the changes in genome-wide expression were monitored using Affymetrix arrays. Nocadozole arrest has been used previously to mimic the cell-cycle arrest phenotype of histone H4 depleted cells (18). Comparing this data (for data set click here) to our histone H4 depletion data, we find that only an additional 7% of the genes repressed by histone H4 depletion are also repressed in arrested cells. Thus, only half of the genes that experience a reduction of expression during histone depletion could be so affected because of the indirect effects of slow growth and cell cycle arrest.
   
   

Telomere-proximal gene expression

1. Is derepression of telomere-proximal genes upon histone H4 depletion continuous from the telomere?
   The telomeric silencing does not appear to be continuous at most telomeres. Many genes which are in the midst of a cluster of silenced, telomere-proximal genes are expressed at detectable levels and are unaffected by histone H4 depletion.
   
   
2. Are all telomeres silenced?
   Telomere silencing appears to be more pronounced at some telomeres than at others. At least 50% of telomere-proximal genes are repressed in 17 of the 32 telomeres. Although all telomeres have at least 1 gene that is derepressed by histone H4 depletion within 20 kb of the telomere, there is 1 telomere with less than 10% of genes affected by nucleosome depletion.
   
   
3. Why is there differing extents of derepression of genes at different telomeres? Why do the SIR2, SIR3, SIR4, and RAP1 deletions not completely overlap in terms of which telomeres are affected?
   A recently published paper indicates that a subset of yeast telomere ends are silenced more readily than others (26). Furthermore, previous studies have shown that telomeric silencing is metastable (5). Thus, a lineage of cells has a specific subset of telomeres repressed, and this subset can change with time. We believe that this may not only explain why a subset of telomeres is affected, but also why the SIR2, SIR3, SIR4 and RAP1 mutants, while affecting approximately the same fraction of telomeric genes in duplicate experiments, do not perfectly overlap.
   
   
4. Does the DNA replication timing of a chromosomal region influence how it is affected by histone H4 depeletion?
   Telomere-proximal genes are replicated late in S-phase. Hence, it is conceivable that replication timing may play a role in the gene expression affects we see at telomeres. However, this hypothesis is not supported upon examination of other regions of the genome that are also late replicating but are not telomeric, including the region surrounding the late replicating ARS1412 on Chr XIV (23) and the region telomere-distal to the late replicating ARS501 on Chr V (24). These non-telomeric late replicating regions do not show any significant clustering of genes affected by histone H4 depletion. This suggests that the timing of DNA replication does not play a role in the effects we see upon histone H4 depletion.
   
   
5. What percentage of telomere-proximal genes are repressed upon SIR3 loss?
   Approximately 2% (6/267) of the telomere-proximal genes are repressed upon SIR3 deletion. This is very similar to fraction of genes repressed SIR3 deletion throughout the entire genome (1.8%), and is considerably less than the number of telomere-proximal genes derepressed by SIR3 deletion (8%).
   
   
6. Is there a mating type effect (due to the a1/alpha2 repressor) that is responsible for the different extent of the telomere-proximal effects observed with Sir and histone depletion experiments?
   To eliminate mating type effects, we compared the expression profile of haploid cells lacking SIR4 and the silent HMRa locus to wild type cells. In these mutant cells, the deletion of SIR4 does not activate the a1/alpha2 repressor, since the a1 gene has been deleted. We found that the effect of SIR4 deletion on telomeric genes is independent of the presence of the silent mating locus. For data from this experiment, click here.

Data analysis

1. How was the expression data from wild-type and mutant strains normalized?
   As described in the protocols (polyA mRNA preparation), equal amounts of five poly-adenylated B.subtilis genes (polyA-controls) were added to 1 mg of total RNA from the wild-type and mutant strains. The signals from the polyA controls were used to normalize expression data between experiments. This normalization procedure allows us to detect bulk changes in the mRNA population.
   
   
2. Click here for expanded version of table 1.