Range extender mediates long-distance enhancer activity

Abstract

Although most mammalian transcriptional enhancers regulate their cognate promoters over distances of tens of kilobases, some enhancers act over distances in the megabase range¹. The sequence features that enable such long-distance enhancer-promoter interactions remain unclear. Here we used in vivo enhancer-replacement experiments at the mouse Shh locus to show that short- and medium-range limb enhancers cannot initiate gene expression at long-distance range. We identify a cis-acting element, range extender (REX), that confers long-distance regulatory activity and is located next to a long-range limb enhancer of Sall1. The REX element has no endogenous enhancer activity. However, addition of the REX to other short- and mid-range limb enhancers substantially increases their genomic interaction range. In the most extreme example observed, addition of REX increased the range of an enhancer by an order of magnitude from its native 73 kb to 848 kb. The REX element contains highly conserved [C/T]AATTA homeodomain motifs that are critical for its activity. These motifs are enriched in long-range limb enhancers genome-wide, including the ZRS (zone of polarizing activity (ZPA) regulatory sequence), a benchmark long-range limb enhancer of Shh². The ZRS enhancer with mutated [C/T]AATTA motifs maintains limb activity at short range, but loses its long-range activity, resulting in severe limb reduction in knock-in mice. In summary, we identify a sequence signature associated with long-range enhancer-promoter interactions and describe a prototypical REX element that is necessary and sufficient to confer long-distance activation by remote enhancers.

Fig. 1. Transplanted enhancers lack long-range limb activity in KI mice.

a, For each limb enhancer (coloured blocks), pseudobulk ZPA scATAC–seq, E11.5 forelimb H3K27ac chromatin immunoprecipitation–sequencing (ChIP–seq) and placental conservation tracks are shown. mHS72 and mHS1516 are the mouse homologues of the HS72 and HS1516 enhancers. b, The corresponding enhancer activities in transgenic E11.5 mouse embryos. Images were acquired from the VISTA enhancer database. Magnified forelimb buds are shown on the right. The white dotted lines demarcate the location of the ZPA. c, Mouse genomic map showing the location of corresponding enhancer regions and their endogenous target genes (dark blue). The curved lines indicate E–P links supported by capture Hi-C, multiome analysis of E11.5 mouse limb buds or matched enhancer activity and gene expression in limb buds of E11.5 or E10.5 embryos (Methods). The H3K27ac ChIP–seq signal from E11.5 forelimbs is shown underneath each region (grey). d, Schematic of mouse Shh loci with the ZRS enhancer replaced by the HS72, MM1564, HS1516 and MM1492 limb enhancers or a fragment of the lacZ sequence. e, Comparative Shh mRNA in situ hybridization analysis in WT and homozygous KI mouse embryos during forelimb bud development (first column). Corresponding skeletal preparations of E18.5 mouse embryos are shown in the second column. s, scapula; h, humerus; r, radius; u, ulna; a, autopod. The number of embryos that exhibited the representative limb phenotype over the total number of embryos with the genotype is indicated. Further details, including hindlimb analysis, are provided in Extended Data Fig. 2. Scale bars, 1 mm (b (left) and e (right)), 250 μm (b, right) and 500 μm (e, left).

Fig. 2. Transplanted enhancers maintain open chromatin structure and drive functional Shh expression in the limb at short range.

a, Allele-specific ZPA ATAC–seq profiles at the transplanted HS1516 (green box) and HS72 (orange box) enhancers and the corresponding WT ZRS locus (top). Bottom, ATAC–seq profiles at endogenous mHS1516 (left; chromosome 16: 95847911–95849564; mm10) and mHS72 (right; chromosome 8: 89454508–89455383; mm10) enhancers (green and orange striped boxes). b, LacZ-stained transgenic E11.5 mouse embryo carrying the HS72 limb enhancer upstream of the Shh promoter (light blue) and lacZ reporter gene (dark blue). The number of embryos with robust LacZ staining in limb buds over the total number of transgenic embryos screened is indicated. c, Schematic of the Shh^HS72(+2kb) KI allele, in which the HS72 enhancer was inserted around 2 kb upstream of the Shh TSS. Skeletal preparation of a forelimb from Shh^{HS72(+2kb)/WT} mouse is shown below on the left. The star symbols indicate extra digits (polydactyly). Hindlimb images are shown in Extended Data Fig. 4b. Quantitative PCR (qPCR) analysis of gene expression in anterior and posterior E11.5 embryonic hindlimb buds in Shh^{HS72(+2kb)/WT} (n = 8 embryos) and WT (n = 6 embryos) mice. Expression is normalized to the average Shh expression in the WT ZPA. Statistical analysis was performed using two-sided Wilcoxon rank-sum tests with no adjustments for multiple comparisons; **P = 0.000311. Scale bars, 1 mm (b (left) and c) and 250 μm (b, right).

Fig. 3. The REX element is necessary and sufficient for long-range activation of Shh by a heterologous limb enhancer.

a, An evolutionary conserved element of unknown function is located adjacent to the human HS72 enhancer. The HS72 enhancer region is shown together with evolutionary conservation tracks. b, Replacement of the ZRS with an extended version of the HS72 enhancer (chromosome 16: 51623658–51625572; hg38) containing the REX element results in initiation of Shh expression in developing limb buds and full limb outgrowth in ZRS^{HS72+REX/HS72+REX} KI mice. Hindlimb and E11.5 Shh expression analysis is shown in Extended Data Fig. 5. c, The REX element lacks classical enhancer activity in E11.5 transgenic embryos when placed upstream of the Shh promoter and lacZ reporter gene. The number of transgenic embryos with no LacZ activity in the limb over the total number of transgenic embryos is indicated. d, Replacement of the ZRS with a chimeric cis-regulatory element consisting of the short-range MM1492 enhancer and the REX element from the HS72 enhancer region results in the initiation of Shh expression in developing limb buds full limb outgrowth in ZRS^{MM1492+REX/MM1492+REX} KI mice. The stars indicate extra digits (polydactyly). Hindlimb Shh expression analysis is provided in Extended Data Fig. 5. Scale bars, 1 mm (c (left) and b and d (right)), 500 μm (b and d, middle) and 250 μm (c, right).

Fig. 4. The REX element contains conserved LHX motifs that are critical for its function and globally linked to long-range regulation.

a, The position and evolutionary conservation of predicted TF motifs within the REX element. The conserved core of the REX element (chromosome 16: 51624707–51624984; hg38) is aligned with orthologous sequences from 12 mammalian species. Sequences matching TF-binding preferences (below) are highlighted. b,c, Forelimb phenotype in ZRS^{HS72+REX∆LEF1/HS72+REX∆LEF1} (b) and ZRS^{HS72+REX∆LHX/HS72+REX∆LHX} (c) KI mice. The star symbols indicate extra digits (polydactyly). d, The top-most enriched TF motifs in long-range (400 kb to 2 Mb) compared with short-range (10 kb to 200 kb) enhancers for experimentally verified VISTA limb enhancers assigned to target genes by scATAC–seq/scRNA-seq or Capture Hi-C. Only the farthest target gene (within 2 Mb range) was considered for each enhancer. TF motifs are grouped by their similarity. [C/T]AATTA HD motifs are highlighted by red shading. False-discovery rate values are shown on the right. Further details are provided in the Methods and Extended Data Fig. 6, and a complete list of motifs is provided in Supplementary Table 2. e, The distribution of E–P distances for VISTA limb enhancers assigned target genes by either Hi-C or Multiome (n = 233) with 0–1 or ≥2 conserved [C/T]AATTA HD motifs. Statistical analysis was performed using the two-sided Wilcoxon rank-sum test with no adjustment for multiple comparisons; P = 0.0073. The counts for each group are displayed in the outlined boxes at the base of the plot. The box plot shows the interquartile range (IQR; quartile 1 to quartile 3) (box limits), the median (centre line) and the minimum (Q1 − 1.5 × IQR) and maximum (Q3 + 1.5 × IQR) values (whiskers). Scale bars, 1 mm (b and c).

Fig. 5. [C/T]AATTA HD motifs are required for long-range gene activation.

a, The position of previously identified TF-binding sites within the mouse ZRS core (chromosome 5: 29314881–29315667; mm10)^,. Conserved [C/T]AATTA motifs are highlighted in blue and their multispecies alignment is shown below. The blue boxes demarcate regions that were mutagenized. b, A dual-enSERT transgenic construct containing WT mouse ZRS driving mCherry (red) and the ZRS with mutated [C/T]AATTA HD motifs (ZRS∆HD) driving eGFP (green) separated by a strong synthetic insulator. Hindlimb bud images from a representative transgenic embryo are shown below. The white dotted line encircles the region that was quantified in c. c, The normalized mean fluorescence intensity in embryos containing the ZRS–mCherry/ZRS∆HD–eGFP or ZRS∆HD–mCherry/ZRS–eGFP (n = 3 embryos) and control ZRS–mCherry/Empty–eGFP (n = 3 embryos) constructs. Statistical analysis was performed using a two-sided paired t-test test with no adjustment for multiple comparisons; *P = 0.0182. d, Skeletal forelimb preparation from E18.5 ZRS^∆HD/∆HD KI mice with three mutated [C/T]AATTA sites within the endogenous ZRS enhancer. e, Addition of the REX element to ZRS∆HD partially rescues limb outgrowth in ZRS^{∆HD+REX/∆HD+REX} KI. f, The proposed model of long-distance enhancer activation in the developing limb buds. Scale bars, 250 μm (b) and 1 mm (d and e).

Extended Data Fig. 1. Multiome analysis (scATAC-seq + scRNAseq) in a mouse hindlimb bud.

(a) Single-cell multiomics in developing E11.5 mouse hindlimb bud. The UMAP plot depicts cell type clusters based on integrated scRNA-seq and scATAC-seq data, along with a cartoon mapping the different regions of the E11.5 hindlimb bud. (b) Single-cell multi-omics quality control violin plots for a number of mRNA counts, ATAC read counts, transcription start site (TSS) enrichment, and nucleosome signal by each cell type cluster. (c) Dot plot showing expression of cell-type-cluster-specific genes (left) and putative target genes of enhancers tested in this study (right) across cell clusters. Colour represents expression level and the size of the circle depicts the percent of cells expressing each gene. The boxed region (Distal Posterior) encompasses the posterior limb bud region containing the Zone of Polarizing Activity (ZPA), where the ZRS normally activates Shh. Apical ectodermal ridge (AER). (d) Schematic pipeline for genome-wide identification of putative enhancer-promoter interactions in the hindlimb. (e) Example of predicted E-P interaction between mHS919 limb enhancer and Rad21. The Rad21 locus (chr15:5000000-5300000; mm10) is shown with Hi-C data (top), pseudobulk chromatin accessibility tracks and a violin plot for Rad21 expression by cell type (bottom). Arches indicate mHS919 enhancer-centric E–P interactions from enhancer capture Hi-C (red) or multiome (blue).

Extended Data Fig. 2. Generation and characterization of mice with transplanted enhancers.

(a) Schematic overview of enhancer replacement strategy. A 4.5 kb mouse genomic region containing the ZRS enhancer is shown together with the vertebrate phyloP conservation (dark blue). The donor vector contained two homology arms with vector-specific sequences for genotyping (green) and a corresponding replaced region containing the transplanted enhancer and mutagenized sgRNA recognition site (black, 5’-agtaccatgcgtgtgtTtTagCC-3’). PCR primers used for genotyping are shown as arrows. See Methods and Kvon et al. for more details. (b) Shown are the results of PCR genotyping for each knock-in mouse line. One representative sample is shown for each genotype consistent with the results seen for all embryos and mice of the lines represented that were analysed in this study. For gel source data, see Supplementary Fig. 1. (c) Shh mRNA whole mount in situ hybridization analysis in wild type and knock-in mouse embryos (first two columns). The corresponding hind limb skeletal preparations of E18.5 wild type and knock-in mouse embryos are shown (third column). The number of embryos that exhibited representative limb phenotype over the total number of embryos with the genotype is indicated. Genotypes for each row are displayed on the left. (d) E18.5 skeletal staining showing the forelimb (left panel) and hindlimb (right panel) phenotype in heterozygous knock-in embryos.

Extended Data Fig. 3. Pairwise alignment of the human HS72 and HS1516 enhancers.

Pairwise sequence alignments of the HS1516 (a) and HS72 (b) enhancers with their respective mouse homologues are shown.

Extended Data Fig. 4. The HS72 enhancer can activate Shh at short-range.

(a) Transgenic E13.5 mouse embryo with the HS72 limb enhancer placed upstream of the Shh promoter and ORF and integrated at H11 safe-harbour locus (light blue). Close-up images of limbs are shown on the right. Asterisks indicate extra digits (polydactyly). Numbers of embryos with limb polydactyly in both forelimb and hindlimb buds over the total number of transgenic embryos screened are shown. FL, forelimb. HL, hindlimb. (b) Hindlimb skeletal staining of P0 Shh^{HS72(+2KB/wt)} mice. (c) PCR gel electrophoresis confirming the correct integration of the left and right homology arms in the Shh^H272(+2KB) line. One representative sample is shown for each genotype consistent with the results seen for all embryos and mice analysed in this study. (d) Model of the targeting construct used to insert the HS72 enhancer approximately 2 KB upstream of Shh. The donor vector contained two homology arms and a corresponding replaced region containing the transplanted enhancer and removing the sgRNA recognition sites (purple, 1: 5’-gggatcatgaggctggccacAGG-3’ and 2: 5’-aaaggccacatttcttcctgTGG-3’). PCR primers used for genotyping are shown as arrows. For gel source data, see Supplementary Fig. 1. See Methods for more details.

Extended Data Fig. 5. The REX element is required and sufficient for long-range activity at the Shh locus.

(a) Shh gene expression analysis using mRNA whole mount in situ hybridization in homozygous ZRS^{HS72+REX/HS72+REX} knock-in mouse embryos at E10.5 (left two panels). Arrows point to the Shh expression in the ZPA. The corresponding hindlimb skeletal preparations at E18.5 are shown (third column). The number of embryos that exhibited representative limb phenotype over the total number of embryos with the genotype is indicated. (b) Forelimb (first panel) and hindlimb (second panel) skeletal phenotypes in heterozygous ZRS^HS72+REX/wt knock-in embryos at E18.5. (c) Comparative Shh mRNA in situ hybridization analysis in wild type (top row) and homozygous ZRS^{HS72+REX/HS72+REX} knock-in (bottom row) mouse embryos during limb bud development in E11.75 embryos. Black arrows point to areas of Shh expression in the ZPA. Red arrows point to ectopic expression in the anterior portion of the limb bud. * Extra digits. (d) Schematic showing the sequences from the HS72 (chr16:51623658-51623900 and chr16:51624689-51625572; hg38) and MM1492 (chr4:154706480-154712089; mm10) loci that make up the chimeric MM1492 + REX element. The orange bars indicate the expanded region flanking the core HS72 enhancer that contains the REX element. The purple bar marks the MM1492 enhancer sequence. The core HS72 enhancer was swapped out for the MM1492 enhancer to generate the MM1492 + REX construct (bottom). (e) Shh gene expression analysis using mRNA whole mount in situ hybridization in homozygous ZRS^{MM1492+REX/MM1492+REX} knock-in mouse embryos at E10.5. (f) Forelimb and hindlimb skeletal phenotypes of heterozygous ZRS^{MM1492+REX/wt} embryos at P0.

Extended Data Fig. 6. Putative TF binding sites in the REX element and other limb enhancers.

(a) Multiple sequence alignment (from UCSC genome browser) across 64 vertebrate species showing conserved LHX2, LEF1, and LHX9 TF motifs in the REX element. (b) Hind limb skeletal staining from ZRS^{HS72+REX∆LEF1/HS72+REX∆LEF1} and ZRS^{HS72+REX∆LHX/HS72+REX∆LHX} P0 mice. (c-e) Motifs enriched in long-range (400 kb - 2MB) compared to short-range (10 kb - 200 kb) enhancers assigned to their longest interacting target gene. In (c) all enhancer regions predicted by scATAC-seq/scRNA-seq were used while in (d) target genes were assigned to experimentally validated limb enhancers by Hi-C while in (e) target genes were only assigned to experimentally validated limb enhancers if they were identified by both Hi-C and scATAC-seq/scRNA-seq (Multiome). (f-i) Distribution of enhancer-promoter distances for putative enhancers with 0-1 compared to more than 1 [C/T]AATTA HD motifs. The box plots show the interquartile range (IQR; Q1–Q3) (box edges), the median (middle line), and the minimum (Q1 − 1.5 × IQR) and maximum (Q3 + 1.5 × IQR) values (whiskers). P-values are reported on the charts from a two-sided Wilcoxon rank sum test with no adjustments for multiple comparisons. (f) For experimentally validated limb enhancers assigned to their nearest interacting target gene by Hi-C or Multiome (n = 233 independent enhancers-promoter pairs). (g-i) Assignment to the longest target identified is shown in the top panel, while shortest is shown in the bottom. (g) Shows experimentally validated limb enhancers assigned a target gene through Hi-C (n = 138 independent enhancers-promoter pairs). (h) Shows experimentally validated limb enhancers where the target gene was assigned by both Hi-C and scATAC-seq/scRNA-seq (Multiome) (n = 46 independent enhancers-promoter pairs) (i) Shows all predicated limb enhancers assigned target genes by Multiome (n = 13,464 independent enhancers-promoter pairs). (j) CTCF ChIP-seq from E11.5 fore- and hindlimb buds showing CTCF binding at Sall1 locus. HS72 and the REX element are highlighted in orange. (k) Shows the location of [C/T]AATTA HD motifs called by HOMER in representative long-range limb enhancers assigned the same target gene by both Hi-C and scATAC-seq/scRNA-seq together with evolutionary conservation track.

Extended Data Fig. 7. Mutagenesis analysis of [C/T]AATTA sites within the ZRS enhancer.

(a) Schematic of the ZRS-mCherry/Empty-eGFP dual-enSERT vector (b-d) Fluorescent images of E11.5 hindlimbs of transgenic mice with a single integration of the ZRS-mCherry/Empty-eGFP dual-enSERT (b) ZRS-mCherry/ZRS∆HD-eGFP (n = 2) (c) and ZRS∆HD-mCherry/ZRS-eGFP (n = 1) (d) constructs at H11 locus. While dotted lines encircle the region quantified in Fig. 5. (e-f) Schematic of a knock-in strategy used to generate ZRS^∆HD (e) and ZRS^∆HD+REX (f) mice. Primers F5 and R5 are designed to anneal to the mutated [C/T]AATTA HD sites but not the endogenous ZRS sequence. (g) Genotyping results of ZRS^∆HD knock-in mice. (h) hindlimb skeletal preparations of ZRS^∆HD/∆HD. (i) Genotyping results of ZRS^∆HD+REX knock-in mice. (j) Hindlimb skeletal preparations from ZRS^{∆HD+REX/∆HD+REX} P0 mice. For the PCR results shown in (g, i) one representative sample is shown for each genotype consistent with the results seen for all embryos and mice of the lines represented that were analysed in this study. For gel source data, see Supplementary Fig. 1.