Cellular zinc status alters chromatin accessibility and binding of p53 to DNA

Abstract

Zn²⁺ is an essential metal required by approximately 850 human transcription factors. How these proteins acquire their essential Zn²⁺ cofactor and whether they are sensitive to changes in the labile Zn²⁺ pool in cells remain open questions. Using ATAC-seq to profile regions of accessible chromatin coupled with transcription factor enrichment analysis, we examined how increases and decreases in the labile zinc pool affect chromatin accessibility and transcription factor enrichment. We found 685 transcription factor motifs were differentially enriched, corresponding to 507 unique transcription factors. The pattern of perturbation and the types of transcription factors were notably different at promoters versus intergenic regions, with zinc-finger transcription factors strongly enriched in intergenic regions in elevated Zn²⁺ To test whether ATAC-seq and transcription factor enrichment analysis predictions correlate with changes in transcription factor binding, we used ChIP-qPCR to profile six p53 binding sites. We found that for five of the six targets, p53 binding correlates with the local accessibility determined by ATAC-seq. These results demonstrate that changes in labile zinc alter chromatin accessibility and transcription factor binding to DNA.

Figure 1.. MCF10A cells are susceptible to perturbations in nuclear Zn²⁺.

(A) Background-corrected FRET ratio traces for MCF10A cells expressing NLS-ZapCV2. Addition of ZnCl₂ for 30 min results in an increase in labile Zn²⁺ from 150 pM to 75 nM. Addition of the Zn²⁺ chelator TPA followed by the addition of Zn²⁺ and pyrithione at 75 min allows for calibration of the sensor and subsequent quantification of labile Zn²⁺. (B) Quantification of Zn²⁺ at rest and at the time point of ATAC-seq. (C) Pseudo-colored ratio images of representative cells at the indicated points. (D) FRET ratio traces of MCF10A cells treated with TPA to deplete labile Zn²⁺, followed by calibration. (E) Quantification of Zn²⁺ at rest. The asterisk indicates that Zn²⁺ cannot be accurately quantified because it is at the lower detection limit of the sensor. Zn²⁺ is estimated to be ∼1 pM. (F) Pseudo-colored ratio images of representative cells at the indicated points. Each trace represents a single cell in the field of view. Lookup table values refer to the FRET ratio (background-corrected FRET channel/background-corrected CFP channel). Scale bar = 20 μm.

Figure 2.. Perturbing cellular Zn²⁺ causes global changes in chromatin accessibility.

(A) Volcano plots depicting genomic regions that exhibit a change in chromatin accessibility for cells treated with 50 μM TPA (left) or 30 μM ZnCl₂ (right). A positive log₂FoldChange indicates an increase in accessibility, whereas a negative log₂FoldChange indicates a decrease in accessibility. (B) DESeq2 differential accessibility analysis shows that most peaks are uniquely accessible depending on Zn²⁺ status. The overlap in the Venn diagrams indicates the peaks which are inversely accessible between the denoted treatments. (C) Genomic tracks showing the top hit for differential accessibility for TPA (left) and ZnCl₂ (right) treated cells. Green boxes denote accessible chromatin as annotated using HMMRATAC. Source data are available for this figure.

Figure 3.. Genomic regions encompassing the Zn²⁺ regulatory genes MT2A and MT1E show increased accessibility with ZnCl₂ treatment.

Top, the ∼154 kb region of chromosome 16 where all MT isoforms reside. Bottom, zoomed in views of the MT2A locus (left) and the MT1E locus (right).

Figure 4.. Perturbing cellular Zn²⁺ causes changes in TF motif enrichment that differ at promoters versus distal regulatory regions.

(A, B) Enrichment plot at promoters (A) and distal regulatory regions (B) showing differential enrichment in ZnCl₂ versus TPA. Dark blue = differentially enriched in both ZnCl₂ and TPA, light blue = differentially enriched in ZnCl₂, green = differentially enriched in TPA. Closed circles are ZF TFs. Open circles are not ZF TFs. (C) Summary of the changes. (D) Analysis of the motifs that are differentially enriched in ZnCl₂ in both intergenic (pink) and promoter (green) region. (E) Analysis of the motifs that are differentially enriched in TPA in both intergenic (pink) and promoter (green) region. Overlapping motifs were ranked by their P_adj values. Source data are available for this figure.

Figure 5.. p53 motifs show differential enrichment upon Zn²⁺ perturbation.

(A) Barcode plots that show enrichment of the p53 motif upon treatment of MCF10A cells with TPA (top). Addition of exogenous ZnCl₂ results in depletion of the same motif (bottom). The barcode plots represent each quartile (Q1–Q4) of the enrichment curves generated for the p53 motif. Red indicates more enrichment of the motif; yellow is less enrichment of the motif. (B) A subset of p53 binding sites from a ChIP-seq dataset (GSM3378513) are differentially accessible depending on cellular Zn²⁺ status. ATAC-seq reads were mapped to 2,164 ChIP-seq peaks and subjected to differential accessibility analysis using DESeq2. With TPA treatment, 62 peaks showed significant (P_adj ≤ 0.1) changes in accessibility (51 decreased, 11 increased). With ZnCl₂ treatment, seven peaks showed significant changes in accessibility (one increased, five decreased). Source data are available for this figure.

Figure 6.. Selected regions of interest for ChIP-qPCR assays with ATAC-seq coverage tracks for MCF10A cells treated with either 50 μM TPA, a media-only Control (Ctrl), or 30 μM ZnCl₂.

Also shown are the PRO-seq coverage tracks for MCF10A cells treated with 10 μM Nutlin-3A for 3 h (GSE227931), the annotated region from the GSM3378513 p53 Nutlin-3A ChIP-seq dataset, and the predicted amplicon from ChIP-qPCR. (A, B, C, D, E, F) Coverage tracks as noted above for the (A) ERGIC1, (B) NFIB, (C) SFN, (D) EGR1, (E) PLD5, and (F) LRIG3-DT regions.

Figure 7.. ChIP-qPCR results for putative p53 targets.

(A) %IP of each p53 target from ChIP-qPCR samples incubated with (+) or without (−) p53 antibody. Each data point is a biological replicate averaged between two technical replicates. (A, B) Same as (A) but for the negative and positive qPCR controls. (C) Signal-to-noise ratio of each ChIP-qPCR p53 target. Dotted line denotes a SNR value of 1.5. Controls for p53 binding are shown in the gray box. (D) Fold enrichment relative to the control for each ChIP-qPCR p53 target (left). Dashed line indicates enrichment ratio of 1 (no enrichment), and lines within each dataset represent the mean. Comparison between the ATAC-seq prediction of increased or decreased accessibility (>1 or <1, respectively) and experimental fold enrichment for each gene target by ChIP-qPCR (right).