SOX transcription factors direct TCF-independent WNT/beta-catenin transcription

WNT/ß-catenin signaling regulates gene expression across numerous biological contexts including development, stem cell homeostasis and tissue regeneration, and dysregulation of this pathway has been implicated in many diseases including cancer. One fundamental question is how distinct WNT target genes are activated in a context-specific manner, given the dogma that most, if not all, WNT/ß-catenin responsive transcription is mediated by TCF/LEF transcription factors (TFs) that have similar DNA-binding specificities. Here we show that the SOX family of TFs direct lineage-specific WNT/ß-catenin responsive transcription during the differentiation of human pluripotent stem cells (hPSCs) into definitive endoderm (DE) and neuromesodermal progenitors (NMPs). Using time-resolved multi-omics analyses, we show that ß-catenin association with chromatin is highly dynamic, colocalizing with distinct TCFs and/or SOX TFs at distinct stages of differentiation, indicating both cooperative and competitive modes of genomic interactions. We demonstrate that SOX17 and SOX2 are required to recruit ß-catenin to hundreds of lineage-specific WNT-responsive enhancers, many of which are not occupied by TCFs. At a subset of these TCF-independent enhancers, SOX TFs are required to both establish a permissive chromatin landscape and recruit a WNT-enhanceosome complex that includes ß-catenin, BCL9, PYGO and transcriptional coactivators to direct SOX/ß-catenin-dependent transcription. Given that SOX TFs are expressed in almost every cell type, these results have broad mechanistic implications for the specificity of WNT responses across many developmental and disease contexts.

1 WNT/ß-catenin signaling regulates gene expression across numerous biological contexts 2 including development, stem cell homeostasis and tissue regeneration, and dysregulation of this 3 pathway has been implicated in many diseases including cancer. One fundamental question is 4 how distinct WNT target genes are activated in a context-specific manner, given the dogma that

207
We next assessed whether SOX17 was required to recruit ß-catenin to chromatin, 208 particularly in genomic regions not co-occupied by TCFs. To test this, we generated a CRISPR-209 mediated SOX17-null mutant iPSC line [ Fig S4A] and asked whether recruitment of ß-catenin to 210 DE-specific genomic loci was compromised. Immunostaining and western blots confirmed a loss 211 of SOX17 and showed that total and nuclear ß-catenin protein levels were unaltered in SOX17 212 knockout (KO) cells [Fig S4 B

258
Together these data demonstrate that SOX17 is required to recruit ß-catenin to a subset 259 endoderm-specific WNT-responsive enhancers, many of which have no evidence of TCF binding.

260
Loss of SOX17 leads to widespread relocalization of ß-catenin genomic binding, in many cases 261 recruitment to mesodermal enhancers by TCFs, suggesting that SOX17 and TCFs might compete

266
There is evidence that SOX TFs can act as pioneering factors by directly engaging

306
Lineage-specific recruitment of ß-catenin is a general feature of SOX TFs.

307
Next, we investigated whether other SOX TFs also had the ability to regulate ß-catenin 308 chromatin binding and lineage-specific WNT-responsive transcription. To test this, we used

338
To identify direct SOX2 and WNT target genes, we then performed RNA-seq on WT and 339 SOX2KD cells as well as on cultures where CHIR was replaced with C59 to inhibit WNT signaling

352
Next, we evaluated TCF occupancy of the SOX2/ß-catenin co-bound loci by ChIP-seq for 353 TCF7L1 and LEF1, the two TCFs most highly expressed in NMPs. We found that 50% 358 Surprisingly, we did not observe appreciable differences in ATAC-seq signal at loci that had

12
To understand how SOX17 and ß-catenin activate transcription, we focused on a subset 369 of SOX17/ß-catenin regulated TCF-independent enhancers. We tested a -60kb CXCR4 and a -370 33kb BMP7 enhancer; both these genes were coregulated and co-occupied by SOX17 and ß-371 catenin, but had little evidence of TCF occupancy. DNA sequence analysis confirmed that both 372 these putative enhancers had several SOX17 but no TCF binding sites [see Methods,

387
In WT DE, the SP5 enhancer displayed robust activity, and this was not altered in 388 SOX17KO cells. Motif analysis showed evidence of multiple TCF as well as SOX17 binding sites.

389
Mutating the SOX17 sites did not affect enhancer activity, while as expected, mutating TCF sites

423
Collectively, our data suggests that SOX17 is required to recruit ß-catenin to a subset of     The notion that TFs other than TCF might function to integrate the multiple components of this 534 enhanceosome has been proposed previously 11 . For example, TBX3 associates with the WNT 535 enhanceosome through interactions with BCL9 22 and the RUNX family of TFs can interact with 536 the ChiLS complex 9 . However, the extent to which these TFs can regulate ß-catenin genomic 537 recruitment is unknown.

538
Our experiments show that on TCF-independent WNT responsive enhancers, SOX17 is 539 required for the recruitment of BCL9, PYGO2, as well as transcriptional coactivators p300, BRG1, 540 MED12 and SMC2, which physically interact with the ß-catenin C-terminal transactivation domain.

541
However, the hierarchy and the sequential order in which SOX17 and the WNT enhanceosome 542 complex are assembled on DNA remains to be determined. An attractive hypothesis supported 543 by our data is that SOX17 acts in a sequential manner to first increase the chromatin accessibility 544 at specific enhancers, perhaps acting as a pioneer TF. This then sets the scene for SOX17 to 545 recruit enhanceosome components like Pygopus which facilitates the loading ß-catenin, similar 546 to a model that's been proposed to prime ß-catenin/TCF enhancers to respond to WNT 547 activation 9 . Ultimately, elucidation of the SOX17/ ß-catenin/DNA complex structure, coupled with 548 proteomics and biochemical assays will be critical to dissect the mechanisms assembling a 549 SOX17/ß-catenin transcription complex.

796
PCR was performed as follows: 1 cycle of 72°C for 5min, 98°C for 30s, 5 cycles of 98°C for 10s, 797 63°C for 30s, 72°C for 1min. 5ul of the amplified DNA was then used to perform qPCR to 798 determine the optimal number of additional cycles to prevent amplification saturation of DNA 799 libraries. In all cases, either 4 or 5 additional cycles of PCR was performed at: 98°C for 10s, 63°C

852
In the absence of replicates, peak calling was performed using macs2 98 using -call-summits and 853 a qvalue cutoff of 0.05. HOMER annotatePeaks.pl was then used to annotate these peaks.

854
Differential binding analysis was performed using the DiffBind package 99,100 using default 855 parameters. Differentially bound ß-catenin and ATAC peaks were identified using a fold

918
Analysis of public data: The following public datasets were reanalyzed as described above:         Quantification of ChIP-seq tag density for ß-catenin (A), SOX17 (B), TCF7L2 (C), TCF7 (D), TCF7L1 (E) and LEF1 (F) at the following peak categories: All ß-catenin peaks, peaks bound only be ß-catenin and TCF. Peaks co-bound by ß-catenin, SOX17 and at least one TCF, peaks bound by ß-catenin and SOX17 but not TCFs.
Dotted lines represent the approximate read density corresponding to the peak calling threshold. G. De-novo DNA-binding motif analyses of the above-described peak categories.     Background is all genes in the genome. ** p = 9.8 x 10 -208 for Wnt regulated genes, p = 6.26 x 10 -187 for SOX17 regulated genes, p < 2.