DNA-guided transcription factor cooperativity shapes face and limb mesenchyme

Abstract

Transcription factors (TFs) can define distinct cellular identities despite nearly identical DNA-binding specificities. One mechanism for achieving regulatory specificity is DNA-guided TF cooperativity. Although in vitro studies suggest that it may be common, examples of such cooperativity remain scarce in cellular contexts. Here, we demonstrate how "Coordinator," a long DNA motif composed of common motifs bound by many basic helix-loop-helix (bHLH) and homeodomain (HD) TFs, uniquely defines the regulatory regions of embryonic face and limb mesenchyme. Coordinator guides cooperative and selective binding between the bHLH family mesenchymal regulator TWIST1 and a collective of HD factors associated with regional identities in the face and limb. TWIST1 is required for HD binding and open chromatin at Coordinator sites, whereas HD factors stabilize TWIST1 occupancy at Coordinator and titrate it away from HD-independent sites. This cooperativity results in the shared regulation of genes involved in cell-type and positional identities and ultimately shapes facial morphology and evolution.

Keywords: ALX factors; Coordinator; TWIST1; bHLH; cooperativity; face; homeodomain; limb; mesenchyme; neural crest; transcription factor.

Figure 1.. The ‘Coordinator’ motif is active specifically in embryonic face and limb mesenchyme.

A. Schematic of the ‘Coordinator’ motif and its discovery. B. Rankings of Coordinator and its constituent Ebox/CAGATGG motif in enrichment in the top 10,000 distal accessible regions in ENCODE. e1, e2, and n indicate examples detailed in (C). Points jittered to avoid overplotting. C. Top motif clusters for examples of Coordinator-enriched and Coordinator-negative samples, with relevant motifs highlighted. D. Schematic of cell types and tissues in (E). E. Coordinator motif enrichment across additional ATAC-seq datasets.

Figure 2.. TWIST1 binds Coordinator across tissues with diverse homeodomain TF expression.

A. Motif clusters and example motifs aligned to the E-box within Coordinator; bar plots show number of aligned and total motifs per cluster (STAR Methods). Motif origin: C, ChIP; P, PBM; S, SELEX. B. TWIST1 ChIP-seq in human cell types and dissected mouse embryo tissues. TWIST1 peaks ranked from strongest to weakest in bins of 1000. C. As in (A), but for the homeodomain (HD) portion. D. HD TF expression across cell/tissue types with Coordinator enrichment. Colored circles correspond to the schematic and data in (C). E. TF RNA expression in human cranial neural crest cells (hCNCC) and H9 embryonic stem cells (H9ESC).

Figure 3.. Multiple homeodomains co-bind Coordinator motif with TWIST1.

A. Schematic of endogenous TF tagging and knockout. B. Confirmation of TF tagging and depletion upon dTAG^V-1 addition by Western blot. IB, immunoblot. C. Confirmation of ALX4 knockout in three independent clones by Western blot. D. Heatmap of TF binding (ChIP and CUT&RUN) and chromatin accessibility (ATAC) at promoter-distal binding sites for TWIST1 and/or AP-2α. Units: reads per genome coverage, except for ATAC, which is in signal per million reads. E. The top enriched known motif for each TF, with p-values.

Figure 4.. TWIST1 opens chromatin for homeodomain TFs and enhancer acetylation.

A. Schematic of acute depletion experiments. B. Heatmap of Coordinator motif enrichment, TF binding, chromatin accessibility (ATAC), and H3K27ac at distal enhancers grouped by their change in accessibility upon TWIST1 depletion. Units: reads per genome coverage, except for the Coordinator motif (−log₁₀ p-value) and ATAC (signal per million reads). One representative replicate of two independent differentiations. C. Example enhancers with loss, no change, or gain of accessibility upon TWIST1 depletion. Coordinates (hg38): Loss, chr17:70,668,899–70,678,127; No change, chr11:44,958,683–44,968,011; Gain, chr2:172,058,768–172,068,096. D. Top enriched motif clusters in enhancers with loss or gain of accessibility upon TWIST1 depletion compared to those with no change, with p-values.

Figure 5.. Homeodomain TFs stabilize TWIST1 binding at Coordinator sites.

A. Correlation in Log₂ fold change (FC) in accessibility upon loss of ALX1 (long-term dTAG^V-1 treatment) versus ALX4 (knockout). Red line, y = x. B. Change in accessibility upon loss of both ALX1 and ALX4 vs log sum of individual effects. C. Most chromatin accessibility effects of ALX loss (ALX1 and/or ALX4) are concordant with (but are a subset of) those of TWIST1 loss. NS, not significant. D. Top motif enrichments among peaks responsive to TWIST1 and ALX loss. E and F. TWIST1 binding by ChIP-seq quantitatively shifts from Coordinator to double E-box motif sites upon loss of ALX4 (without ALX1 depletion) in hCNCCs (E) or overexpression of TWIST1 alone rather than with ALX4 in HEK293 cells (F). G through I. Volcano plots of differential gene expression upon loss of TWIST1 (G), ALX1 (H), or ALX1 and ALX4 (I). ALX4 is excluded in (I). Selected genes highlighted in darker colors.

Figure 6.. The Coordinator motif guides TWIST1-homeodomain contact and cooperativity.

A. Immunoprecipitation-mass spectrometry (IP-MS) for TWIST1 using the V5 tag, in undepleted (-dTAG, y-axis) versus depleted (+dTAG, x-axis) hCNCC protein extracts. Plotted data are the sum of two biological replicates. B. 3D structure of TWIST1 (aa101–170), TCF4 (aa565–624), and ALX4 (aa210–277) DNA binding domains bound to the Coordinator DNA sequence. DNA bases recognized by the TFs highlighted: cyan for TWIST1, green for TCF4, and magenta for ALX4. C. Zoom-in of contact between ALX4 and TWIST1. D. Sequence alignment of selected homeodomain TF loop sequences with sequence differences from ALX4 in bold, and structural alignment of ALX4 with MSX1 (PDB: 1IG7) and DLX3 (PDB: 4XRS). E. TWIST1 preference for Coordinator motif upon homeodomain overexpression (see Figure S6A for protein levels). TWIST1 peaks ranked from strongest to weakest in bins of 1000 peaks. F. Extent of Coordinator motif binding preference of V5-tagged TWIST1 and various loop mutants expressed in HEK293 cells (see Figure S6A for protein levels) with or without ALX4. Inset: structural alignment of TWIST1 in our structure with the AlphaFold-predicted (AF) or experimentally solved (PDB: 2QL2, 2YPA) of tested bHLH loops. G. EMSA probe sequences, gels, with Hill curve fits and estimated K_d and Hill coefficients (n).

Figure 7.. The roles of Coordinator-binding TFs in facial shape variation.

A-E. Facial shape effects associated with genetic variants at loci encoding Coordinator-binding TFs (A, TWIST1; B, ALX1; C, MSX1; D, ALX4; E, PRRX1). LocusZoom plots (left) show SNPs plotted by p-value of facial shape association and colored by linkage disequilibrium (r²) to the lead SNP in each locus. Note that p-values are with respect to the trait of each lead SNP. Coordinates in hg19. Facial shape effects of each lead SNP near Coordinator-binding TF genes, as normal displacement (displacement in the direction normal to the facial surface) for the facial region (Figure S7A) with highest significance for each lead SNP. F. Facial shape heritability enrichment at TWIST1-dependent regulatory regions. Vertical line indicates enrichment in all hCNCC distal ATAC peaks; flanking dashed lines indicate error bars (s.e.m.).