pubmed.ncbi.nlm.nih.gov

Comprehensive study of nuclear receptor DNA binding provides a revised framework for understanding receptor specificity - PubMed

  • ️Tue Jan 01 2019

Comprehensive study of nuclear receptor DNA binding provides a revised framework for understanding receptor specificity

Ashley Penvose et al. Nat Commun. 2019.

Abstract

The type II nuclear receptors (NRs) function as heterodimeric transcription factors with the retinoid X receptor (RXR) to regulate diverse biological processes in response to endogenous ligands and therapeutic drugs. DNA-binding specificity has been proposed as a primary mechanism for NR gene regulatory specificity. Here we use protein-binding microarrays (PBMs) to comprehensively analyze the DNA binding of 12 NR:RXRα dimers. We find more promiscuous NR-DNA binding than has been reported, challenging the view that NR binding specificity is defined by half-site spacing. We show that NRs bind DNA using two distinct modes, explaining widespread NR binding to half-sites in vivo. Finally, we show that the current models of NR specificity better reflect binding-site activity rather than binding-site affinity. Our rich dataset and revised NR binding models provide a framework for understanding NR regulatory specificity and will facilitate more accurate analyses of genomic datasets.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1

Characterizing NR-DNA binding with PBMs. a Schematic of spacer preferences for NRs to direct repeats (DRs) and half-sites (HS). b Canonical spacer preferences of NRs indicate preferred spacer lengths from the literature (Supplementary Data 2 and 3). Published PWM models are shown in colored dots that indicate the methodology used to derive the model (Supplementary Data 3). c Schematic of PBM probes, SNV probe organization and SNV-based motif generation for a single seed sequence. d Scatter plot of z-scores for RARβ:RXRα experiments detected with antibodies against each heterodimer partner. Dots represent average over ~5 replicates for all 10,728 unique SNV probes (black dots) and 500 background probes (gray dots) e PBM replicate averaged z-score distributions for PPARγ:RXRα to all SNV probes. Z-scores for consensus DR1 and reported functional binding sites are highlighted (Supplementary Data 2). f DR1 DNA-binding logo for PPARγ:RXRα generated from all DR1 full-site models from the PBM experiments. g Comparison of PPARγ:RXRα PBM z-scores and competition EMSA-determined relative Kd measurements for binding sites spanning a wide affinity range. Relative Kd values are normalized to the highest affinity sequence (P1) and represent mean over two independent experiments (error bars = STDEV). Identifiable DR half-sites in each binding sequence are shown in bold. Mutations introduced to ablate the 5’ half-site of P3 (P3 5’Abl) or the 3’ half-site of P3 (P3 3’Abl) are shown in red. Source data are provided as a Source Data file

Fig. 2
Fig. 2

NR-binding specificity and DR preferences. PBM-derived DNA-binding logos for 12 NRs at all examined DR spacer lengths. Half-site logos identified for each NR on either the 5’ half-site (5’HS) or 3’ half-site (3’HS) are shown. Logos based upon a single significant (z-score > 3.0) seed sequence are indicated (∗). Source data are provided as a Source Data file

Fig. 3
Fig. 3

NR half-site binding mode. a Schematic of NR full-site or half-site binding modes. b, c For three seed sequences bound with different modes, the impact of SNVs on LXRα heterodimer binding and the corresponding DNA-binding logos are shown. Binding perturbation for each SNV is shown as a Δz-score from the median z-score of all four base variants at each position. Colors correspond to base identity indicated in logos below. d DNA-binding logos for all 12 NRs generated for the single DR1 seed sequence shown. e Amino acid sequence of zinc finger 1 for the wild-type RXRα, RXRα DNA-binding domain mutant, wild-type PPARγ, and the PPARγ DNA-binding domain mutant. Altered amino acids are highlighted in gray. f DNA-binding logos for individual seed sequences (shown) for which the binding mode was either altered (left) or maintained (right) for the PPARγ:RXRα-DNA binding domain mutant. Source data are provided as a Source Data file

Fig. 4
Fig. 4

NR-binding affinity and mode for sequences at each DR spacer length. At each spacer length, the replicate averaged z-score of the highest scoring SNV for each seed sequence is shown; seed sequences with z-score < 3 are not represented. Colors indicate binding mode for each seed sequence. For each NR, box plots show the z-score distributions for all sites that are bound in half-site modes across all direct repeat spacer lengths (the aggregate of all gray dots). Center line: median; box limits: upper and lower quartiles; whiskers: last datum within 1.5x interquartile range. Source data are provided as a Source Data file

Fig. 5
Fig. 5

NR specificity differences. a Scatter plots of LXRα and PPARγ binding to DR1 and DR4 sites. Each spot is the average of ∼5 replicates for each unique DNA sequence (∼1600 at each spacer length) on the PBM. DR1 and DR4 spacer-variant sequences are shown in the box below panel. b Binding logos generated for LXRα and PPARγ for the spacer-variant seed sequences from a are shown. c Scatter plots as in 5a of VDR and PXR binding to DR1 and DR4 sites. d Scatter plots as in 5a of VDR and PXR binding to DR1 and DR3 sites. e Scatter plots as in 5a of LXRα and PXR binding to DR1 and DR4 sites. f DR4 z-score logos, directly representing Δz-scores of SNV binding, are shown for LXRα and PXR. Δz-scores are calculated (separately for each NR) as the difference from the median of all SNV variants. g Differential binding of NRs to spacer-sequence variants. (Top panel) Binding is shown for five NRs to the DR4 seed sequence 5’-AGGTCATAGGAGGTCA-3’ and all 12 SNVs of the spacer region (spacer region in bold). Δz-scores are calculated as in 5f. (Bottom panel) Binding is shown for five NRs to the DR3 sequence 5’-AGGTCAGAGAGGTCA-3’ and all nine corresponding SNVs of the spacer region (spacer region in bold). Examples of highly variant spacer sequences are indicated. Source data are provided as a Source Data file

Fig. 6
Fig. 6

Genomic enrichment of NR-binding motifs. a Receiver-operating characteristic (ROC) curves for PPARγ motif enrichment in ChIP-seq data. ROC curves and area under the curve (AUC) values for different PBM-derived NR-binding models are shown, along with the results for best-performing published PPARγ DR1 motif (HOCOMOCO-f1, HMf1). Motif enrichment for all models had p-values < 10−46, using a Wilcoxon rank sum test with continuity correction and Bonferroni corrected for multiple hypotheses. b ROC curves for LXRα motif enrichment in ChIP-seq data. ROC curves and AUC values for different LXRα binding models are shown. Results for best-performing published LXRα DR4 motif (JASPAR MA0494.1) are shown. c ROC curves and AUC values are shown for PPARγ DR1 motif enrichment in reproducibly-bound PPARγ ChIP-seq peaks (solid lines, Methods), and for those peaks occurring within 10 kb upstream of differentially expressed genes (i.e., active peaks). d ROC curves and AUC values are shown for LXRα DR4 motif enrichment in reproducibly-bound LXRα ChIP-seq peaks (solid lines, Methods), and for those peaks occurring within 10-kb upstream of differentially expressed genes (i.e., active peaks). Source data are provided as a Source Data file

Fig. 7
Fig. 7

Activity versus affinity for distinct classes of NR-binding sites. a LXRα-dependent activity and binding affinity of a sequence bound in a half-site mode. Luciferase reporter gene activation, and corresponding z-scores, are shown for the DR1.7 sequence, which is bound in a half-site mode on PBM, and sequences with each half-site ablated (DR1.17 and DR1.18), sequences shown in b. Fold-change reporter expression indicates luciferase activity in HEK293T cells over-expressing LXRα and RXRα normalized to cells not over-expressing these proteins. Fold-change expression is shown for cells treated with DMSO (vehicle), agonist (T0901317), or antagonist (GSK2033), and values represent mean over nine replicate measurements (error bars = SEM). Reporter gene p-values: * < 0.01, *** < 0.0001 (calculated using Student’s two-tailed t-test). b Logo for LXRα heterodimer binding to DR1.7, and sequences for DR1.7, DR1.17, and DR1.18 discussed in a. c LXRα- and PPARγ-dependent activity and PBM-derived binding scores to select DR1 and DR4 sites. Fold-change expression for LXRα is as described in a. Fold-change for PPARγ is shown for cells treated with DMSO (vehicle), agonist (rosiglitazone), or antagonist (T0070907), and values represent the mean over nine replicate measurements (error bars = SEM). d Overview of relation between NR in vitro binding, in vivo binding, and function. Source data are provided as a Source Data file

Similar articles

Cited by

References

    1. de Aguiar Vallim TQ, Tarling EJ, Edwards PA. Pleiotropic roles of bile acids in metabolism. Cell Metab. 2013;17:657–669. doi: 10.1016/j.cmet.2013.03.013. - DOI - PMC - PubMed
    1. Evans RM, Mangelsdorf DJ. Nuclear receptors, RXR, and the big bang. Cell. 2014;157:255–266. doi: 10.1016/j.cell.2014.03.012. - DOI - PMC - PubMed
    1. Kliewer SA, Lehmann JM, Willson TM. Orphan nuclear receptors: shifting endocrinology into reverse. Science. 1999;284:757–760. doi: 10.1126/science.284.5415.757. - DOI - PubMed
    1. Calkin AC, Tontonoz P. Transcriptional integration of metabolism by the nuclear sterol-activated receptors LXR and FXR. Nat. Rev. Mol. Cell Biol. 2012;13:213–224. doi: 10.1038/nrm3312. - DOI - PMC - PubMed
    1. Potthoff MJ, Kliewer SA, Mangelsdorf DJ. Endocrine fibroblast growth factors 15/19 and 21: from feast to famine. Genes Dev. 2012;26:312–324. doi: 10.1101/gad.184788.111. - DOI - PMC - PubMed

Publication types

MeSH terms

Substances