Reply to: Pitfalls in using phenanthroline to study the causal relationship between promoter nucleosome acetylation and transcription

We appreciate the comments of Zencir et al. regarding our manuscript on the impact of transcription on genome-wide histone acetylation patterns 1 . We agree that our paper shifts our understanding of how acetylation is targeted to transcribed genes and thus the more scrutiny the results receive, the better. With this being said, we respectfully disagree that the loss of histone acetylation in cells treated with the transcription inhibitor, 1,10-phenanthroline (1,10-pt), is due to the disruption of histone acetyltransferase (HAT) binding to promoters. Although we do not dispute that the binding of some transcription activators, and thus HATs, may be impacted by this drug, the NuA4/piccolo HAT complex subunit, Epl1, is not lost at most promoters following 1,10-pt treatment. Using DESeq2’s binomial test 2 with a lenient cutoff ( p -value < 0.1, 1.5 fold-change), we observe a statistically significant loss of Epl1 binding at only 3.6% (198/5542) of promoters (Fig. 1a ). This re-analysis of our data is completely consistent with the heatmap in our original manuscript (Figure S5A) 1 , which shows that upon transcription inhibition: (1) Epl1 binding is stable at most promoters, albeit with a small fraction showing decreased binding, and (2) Epl1 is ubiquitously lost from 5′ gene body regions, consistent with our interpretation of transcription-dependent targeting of Epl1 to gene bodies. While the binding of Epl1 to both promoters and gene bodies is difficult to differentiate in the gene-dense yeast genome, the contrasting impact of Epl1 mutation and 1,10-pt treatment on Epl1 binding to these features confirms the specificity of our ChIP-seq experiments and strengthens our finding that Epl1 binding is stable to 1,10-pt at the vast majority of promoters. While Zencir et al. raise good points about the pitfals of metaplots, the heatmaps we included in our original paper address many of the points they now claim to uncover. Despite the limited impact of 1,10-pt on Epl1 occupancy, we observed loss of H4K8ac and H4K12ac at the majority of adjacent +1 nucleosomes (Fig. 1b, c ). We do observe Epl1 loss at a larger portion of promoters with Epl1-peaks (126/562 promoters, Fig. 1d ), but Fig. 1e, f show that acetylation is lost at nucleosomes adjacent to all promoters in this class, even those exhibiting increased Epl1 occupancy, suggesting that loss of acetylation cannot be simply explained by loss of HAT recruitment. These data provide indirect support that the binding of transcription factors to most promoters is undisturbed upon 1,10-pt treatment, consistent with the results of Poramba-Liyanage et al., which characterized the impact of 1,10-pt on the occupancy of general transcription factors at a transcribed locus 3 . Interestingly, Fig. 1a, b, d in Zencir et al. also reinforce this point. At the genes shown, Rpb3 binding to promoters, presumably via recruitment by transcription activators, is not impacted by 1,10-pt. Regardless, the key point is that Epl1 occupancy at most genes is largely unaffected by 1,10-pt, which is inconsistent with the Zencir et al. statement that this drug has profound genome-wide effects on the localization and activity of transcription activators, as well as HAT and HDAC complexes. Again, we do not dispute that some promoters might be impacted, such as those of genes encoding ribosomal proteins and proteins involved in ribosome biogenesis, however, although these genes, represent a large portion of RNAPII initiation events, they are a small fraction of total genes (~350 in total), and thus account for a very small percentage of acetylated histones in the cell. Zencir et al. also speculate that increased targeting of the HDAC, Rpd3, could explain the decrease in acetylation following 1,10-pt treatment. However, we tested this possibility in our original paper (Supplementary Fig. 2e) 1 , finding that the decrease in bulk histone acetylation was not due to increased HDAC activity. Fig. 1: Transcription inhibition by 1,10-phenanthroline (1,10-pt) does not impact Epl1 occupancy at most promoters. Scatter plots comparing the effect of 1,10-pt on fragment midpoint counts from Epl1 ( a , d ), H4K8ac ( b , e ), or H4K12ac ( c , f ) ChIP-seq, over promoters ( a and d ) or associated +1 nucleosomes ( b , c , e , and f ) of all (5542) genes ( a , b , c ) or of 562 genes with NDR Epl1 peaks identified in MNase ChIP-seq in control cells ( d , e , f ). Promoters with >1.5 fold-change in Epl1 using an adjusted p -value < 0.1 (DESeq2’s binomial test 2 ) are shown in red (increased) or blue (decreased). In b , c , e , and f , the line is: y = x + 0. Full size image

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.