A TALE/HOX code unlocks WNT signalling response towards paraxial mesoderm

Mariani, Luca; Guo, Xiaogang; Menezes, Niels Alvaro; Drozd, Anna Maria; Çakal, Selgin Deniz; Wang, Qinhu; Ferretti, Elisabetta

doi:10.1038/s41467-021-25370-4

Download PDF

Article
Open access
Published: 26 August 2021

A TALE/HOX code unlocks WNT signalling response towards paraxial mesoderm

Nature Communications volume 12, Article number: 5136 (2021) Cite this article

4545 Accesses
7 Citations
5 Altmetric
Metrics details

Subjects

Abstract

One fundamental yet unresolved question in biology remains how cells interpret the same signalling cues in a context-dependent manner resulting in lineage specification. A key step for decoding signalling cues is the establishment of a permissive chromatin environment at lineage-specific genes triggering transcriptional responses to inductive signals. For instance, bipotent neuromesodermal progenitors (NMPs) are equipped with a WNT-decoding module, which relies on TCFs/LEF activity to sustain both NMP expansion and paraxial mesoderm differentiation. However, how WNT signalling activates lineage specific genes in a temporal manner remains unclear. Here, we demonstrate that paraxial mesoderm induction relies on the TALE/HOX combinatorial activity that simultaneously represses NMP genes and activates the differentiation program. We identify the BRACHYURY-TALE/HOX code that destabilizes the nucleosomes at WNT-responsive regions and establishes the permissive chromatin landscape for de novo recruitment of the WNT-effector LEF1, unlocking the WNT-mediated transcriptional program that drives NMPs towards the paraxial mesodermal fate.

Eomes and Brachyury control pluripotency exit and germ-layer segregation by changing the chromatin state

Article 02 December 2019

Sox2 levels regulate the chromatin occupancy of WNT mediators in epiblast progenitors responsible for vertebrate body formation

Article Open access 12 May 2022

Wnt/Beta-catenin/Esrrb signalling controls the tissue-scale reorganization and maintenance of the pluripotent lineage during murine embryonic diapause

Article Open access 30 October 2020

Introduction

In mammals, skeletal muscles and vertebrae originate from the paraxial mesoderm, which includes unsegmented pre-somitic mesoderm (PSM) and somites^1,2,3. PSM formation, which represents the earliest event of musculoskeletal development, relies on the availability of expanding neuromesodermal progenitors (NMPs) that also contribute to the spinal cord^3,4,5,6. NMPs reside in the caudal lateral epiblast (CLE) and tailbud, from where they sustain axial elongation^4,7,8. Intricate molecular circuits regulated by WNT signalling and involving the transcription factors (TFs) HOX, CDX, SOX2, BRACHYURY (T-BRA), TBX6 and MSGN1, balance NMP expansion and differentiation^{9,10,11,12,13,14,15,16,17}. NMPs can adopt neural fate by downregulating T-Bra and maintaining Sox2 expression. Conversely, NMPs acquiring PSM fate retain T-Bra expression, downregulate Sox2 and upregulate Tbx6 and Msgn1, transitioning into nascent mesodermal progenitor cells (MPCs) that fuel PSM development^14,18. Ultimately, somite formation depends on the speed by which NMPs leave the CLE to acquire paraxial mesoderm fate¹.

HOX TFs have been historically implicated in axial elongation, with their sequential activation playing a fundamental role in timing PSM development^16,19,20. Posterior Hox genes (Hox10-13) reduce mesoderm ingression by repressing WNT and are responsible for body axis termination^11,21. Conversely, the anterior Hox genes (Hox1-8) control paraxial mesoderm formation by modulating cell ingression into the primitive streak (PS)¹⁹. Nevertheless, the mechanism by which Hox genes promote PSM specification remains largely unknown. TFs belonging to the TALE (three-amino acid loop extension) family like the PBX and MEIS factors have distinct ability to interact with HOX proteins and provide HOX with DNA-binding specificity²². Given that PBX proteins are obligate anterior HOX cofactors, they represent prominent candidates for controlling the distinct response to individual HOX factors²³. PBX1, PBX2 and PBX3 play fundamental roles during organogenesis^22,24,25. Compound deletion of PBX1 and PBX2 (hereafter PBX) uncovers a crucial role for these factors in axial/appendicular patterning^24,25, a surprisingly late phenotype as they are widely expressed at earlier embryonic stages²⁶. Importantly, their role in mesodermal lineage specification remains elusive. Given the broad expression of PBX, it is still unclear how cooperative interactions with HOX provide lineage specificity.

By employing in vitro differentiation systems and in vivo mouse models, we show that the patterning abnormalities in Pbx1/2 double mutant embryos (hereafter Pbx) can be traced back to the PSM specification stage. In Pbx mutants, NMPs exhibit reduced ability to generate paraxial mesoderm, being unable to migrate, remaining trapped in the tailbud in a NMP/MPC transition state. Using single-cell sequencing, chromatin accessibility analyses and iterative rounds of DNA:protein interaction assays, we demonstrate that TALE/HOX proteins generate the DNA-binding context for the recruitment of the WNT-effector LEF1, and consequently promote the expression of PSM genes, including the master regulator Mesogenin1 (Msgn1)^10,15. Ultimately, the activation of the paraxial mesoderm gene regulatory networks (GRNs) rely on the activity of a WNT-HOX integrated code, which drives a multistep process involving changes in the chromatin landscape at the WNT-responsive elements supporting the shift from NMP expansion to paraxial mesoderm differentiation. Thus, TALE/HOX act as a molecular switch that triggers alternative WNT signalling cellular responses, unlocking PSM genes and promoting somite formation.

Results

PBX proteins control the formation of the nascent paraxial mesoderm

Addressing how Hox genes influence WNT signalling is critical for paraxial mesoderm formation and somitogenesis^11,19. The HOX cofactors PBX are broadly expressed in the tailbud of embryonic day (E) 8.5 mouse embryos, including PS and NMPs (Supplementary Fig. 1a), questioning their function in the early stages of somitogenesis. To dissect the role of PBX at early stages, we inactivated Pbx1 in Pbx2 mutant mice using the Sox2^Cre allele to obtain Pbx1^−/−;Pbx2^−/− double-knockout (Pbx-DKO) and Pbx1^−/−;Pbx2^+/− compound mutant (Pbx-com) embryos. Pbx-DKO display shortened trunk, enlarged tailbud and rudimentary somites (Fig. 1a and Supplementary Fig. 1b). Furthermore, FOXC2 and UNCX4.1 immunofluorescence (IF) revealed abnormal posterior somite morphogenesis in E8.5 Pbx-DKO and Pbx-com (Fig. 1b and Supplementary Fig. 1c). To pinpoint the role of PBX proteins in PSM development (Fig. 1c), we performed single-cell RNA sequencing (scRNA-seq) on cells isolated from the tailbuds of control and Pbx mutants at E8.5 and E9.0 (Fig. 1d and Supplementary Data 1). After stringent quality filtering, batch-effect correction and data integration of all sequenced embryos, unsupervised clustering analyses revealed 13 different populations, including NMPs, MPCs and PSM (Fig. 1e, Supplementary Fig. 1d, e and Supplementary Data 2). Each embryo contributed to every cluster, irrespective of the genotype. However, controls and mutants exhibited markedly different proportions of cell types populating each cluster. Specifically, Pbx-DKO and Pbx-com mutant cells were overrepresented in the posterior PSM cluster, while being largely reduced in anterior PSM, suggesting a block on PSM maturation. Quantification of cluster composition confirmed that cells displaying MPC features (T-Bra^pos;Tbx6^pos) significantly accumulate in the posterior PSM cluster of Pbx-DKO and Pbx-com at the expense of anterior PSM (Fig. 1f). Specifically focussing on NMPs, MPCs and PSM cell populations revealed that the Pbx-DKO and Pbx-com NMPs and MPCs/pPSM clusters included cells that exhibited a mixed NMP/MPC profile, expressing high levels of T-Bra, Sox2, Wnt3a, Tbx6 and 5′ Hox genes (Fig. 1g and Supplementary Fig. 1f–h). To understand the transition of NMPs towards PSM, we applied pseudotemporal ordering analyses and confirmed defective paraxial mesoderm differentiation in Pbx-DKO, resulting in inflation of the progenitor reservoir at the expense of mature PSM (Fig. 1h and Supplementary Fig. 1i).

**Fig. 1: PBX proteins control the transition of NMPs towards PSM, underlying the acquisition of paraxial mesoderm identity.**

The impairment of PSM formation and the increased number of MPCs with NMP features in Pbx-DKO and Pbx-com prompted us to address whether the spatial organisation of the progenitor zone was affected²⁷. By IF, we observed an expansion of T-BRA^pos;SOX2^pos NMPs in the Pbx-DKO and Pbx-com tailbuds (Fig. 2a and Supplementary Fig. 2a). By contrast, the mesodermal marker TBX6 was strongly downregulated in the PSM and upregulated in the CLE of Pbx-DKO and Pbx-com mutants (Fig. 2a), confirming the accumulation of aberrant progenitors observed by scRNA-seq. The distribution of progenitor populations assessed in the caudal and medial parts of the CLE, where cells preferentially adopt a PSM fate, revealed that in Pbx-DKO and Pbx-com the number of NMPs (T-BRA^pos;SOX2^pos) and MPCs (T-BRA^pos;TBX6^pos) increased, while the amount of mature PSM (TBX6^pos) cells was dramatically reduced (Fig. 2b, c and Supplementary Fig. 2b). Notably, the proportion of T-BRA^pos;TBX6^pos cells relative to TBX6^pos cells was significantly enhanced in Pbx-DKO and Pbx-com, suggesting an impaired maturation of MPCs (Fig. 2c). Furthermore, NMPs robustly accumulate in the mutant posterior tailbuds, as opposed to the fewer NMPs observed in controls (Fig. 2d). Thus, the analysis of the progenitor allocation in Pbx mutant embryos corroborated the scRNA-seq findings (Fig. 1e–h), and provided evidence that Pbx-DKO and Pbx-com NMPs are hindered in their progress towards PSM, remaining trapped in the CLE in a NMP/MPC state.

**Fig. 2: *Pbx* mutant NMPs fail to acquire paraxial mesodermal fate, do not progress towards PSM, and accumulate in the tailbuds.**

Given that WNT3A supports both NMP maintenance and PSM differentiation^14,18 (Supplementary Fig. 2c), we tested whether the PSM defects in the Pbx mutants were the consequence of perturbed WNT expression, and found that Wnt3a expression domain was increased in the Pbx-com tailbud (Supplementary Fig. 2d), in agreement with the observed progenitor zone expansion (Fig. 2a, d). Therefore, although WNT signalling is a major mediator of PSM specification, the phenotypes of Pbx mutants are not associated with the depletion of the Wnt3a ligand, suggesting that PBX could have a role in modulating how NMPs respond to this pathway. Remarkably, Pbx mutants have an expanded NMP population, implicating PBX in the regulation of the WNT-dependent NMP-to-PSM transition.

PBX proteins promote the transition from NMPs towards PSM

To address PBX function during PSM differentiation, we took advantage of our recently developed in vitro differentiation protocol (Mariani, L. et al., in preparation²⁸) that efficiently drives mouse epiblast stem cells (EpiSCs) towards PSM (Fig. 3a and Supplementary Fig. 3a). This system allowed us to capture, with high temporal resolution, the transition from NMPs to nascent PSM that would be otherwise difficult to pinpoint in vivo due to the limited number of progenitors. ScRNA-seq analyses at different time-points identified NMPs by the co-expression of T-Bra, Sox2, Wnt3a, Nkx1-2 and Cdx2 at 24 h (Fig. 3b, c and Supplementary Data 3). These NMPs expressed Cdx1, Cdh1, Fst and the anterior Hox genes (Hoxa1-b1), a signature characteristic of E8.5 NMPs¹⁸. At 36 and 48 h, NMPs expressing posterior Hox genes (Hox6-9) compatible with a E9.5 NMP or pre-neural signature coexist with cells transitioning towards PSM¹⁸, marked by Tbx6, Foxc2 and Meox1 (Fig. 3b, c). To confirm that our in vitro model replicated genuine in vivo PSM differentiation, we assessed the capacity of these cells to generate somite-derived tissues, by differentiating PSM towards sclerotome and dermomyotome, which give rise to axial skeleton and muscles respectively (Supplementary Fig. 3b). Populations of sclerotome-like cells expressing Foxc2, Pax1 and Pax9, and dermomyotome-like cells marked by Pax3 and Pax7 were readily detectable following 5 days of differentiation, confirming the potency of the progenitors (Supplementary Fig. 3c, d).

**Fig. 3: *Pbx* mutant cells fail to acquire paraxial mesoderm identity in vitro.**

To address PBX role in PSM formation, we established Pbx-DKO cell lines, either by inactivating Pbx1 and Pbx2 using CRISPR/Cas9 (Fig. 3d) or by CRE-mediated recombination of Pbx1^flox/flox;Pbx2^−/− derived EpiSCs (Supplementary Fig. 3e). Real-time quantitative PCR (RT-qPCR) and IF revealed impaired ability of Pbx-DKO cells to generate dermomyotome and sclerotome, mimicking the Pbx-DKO and Pbx-com embryonic defects (Supplementary Fig. 3f–h)^24,25.

We used scRNA-seq analyses along PSM differentiation to test if Pbx-DKO phenotypes are recapitulated in vitro (Fig. 3e), focussing on the transition from NMPs to PSM. Temporal trajectory reconstruction along differentiation showed that wild-type (WT) NMPs efficiently differentiate towards PSM, whereas Pbx-DKO cells fail to reach the PSM state and accumulate as progenitors co-expressing T-Bra and Sox2 (Fig. 3f, g). IF staining and quantification confirmed that PBX loss strongly impairs PSM differentiation, resulting in a higher proportion of NMPs and MPCs at the expense of PSM (Fig. 3h, i), recapitulating the phenotypes observed in vivo and highlighting the critical role of PBX proteins in specifying PSM.

PBX proteins are dynamically recruited to PSM genes

To identify PBX targets during PSM differentiation, we performed chromatin immunoprecipitation followed by deep sequencing (ChIP-seq) using PBX1 and PBX2 antibodies at different time-points (Fig. 4a). The number of PBX1 peaks increased during differentiation (Supplementary Fig. 4a), with a maximum number of binding events occurring at 24 h: the stage when mesodermal fate is established. The functional readout of PBX-binding events was assessed by RNA-sequencing (RNA-seq) of WT and Pbx-DKO cells at 12 and 24 h (Fig. 4a–c and Supplementary Fig. 4b, c). To pinpoint the genes directly modulated by PBX, we integrated the ChIP-seq and RNA-seq datasets and identified 251 differentially regulated targets (Supplementary Data 4). GO-term analyses of the downregulated genes confirmed PBX requirement in PSM differentiation, somite development and segmentation (Fig. 4b and Supplementary Fig. 4b). By contrast, loss of PBX resulted in increased expression of neuronal and NMP genes like Sox2 and Cdx2, suggesting that PBX may be important for balancing the alternative fates of NMPs (Fig. 4c and Supplementary Fig. 4c). As previous reports have suggested that PBX proteins possess pioneer factor activity^29,30,31, we assessed the PBX-dependent changes in chromatin accessibility at the PBX-bound sites. We performed Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq) at different time-points (Fig. 4d) and compared chromatin accessibility of the PBX target genes misregulated in Pbx-DKO cells (Fig. 4e, f). Regulatory regions associated with upregulated genes in Pbx-DKO were open in EpiSCs and remained accessible in differentiated WT and Pbx-DKO cells (Fig. 4f). By contrast, loci associated with downregulated genes in Pbx-DKO were closed in EpiSCs and became increasingly accessible during PSM differentiation peaking at 24 h in WT, but not in Pbx-DKO cells (Fig. 4e). Thus, by enabling chromatin opening at PSM loci, PBX proteins could prime regulatory regions for the recruitment of lineage-specific TFs. These results substantiate the previously suggested role of PBX as pioneer factors^29,30,31, but raise the question of how these broadly expressed proteins act at specific loci. As canonical transcriptional regulation is known to be driven by cooperativity, we searched for factors that could cooperate with PBX to promote chromatin accessibility at key PSM regulatory regions³². De novo motif analyses of the PBX-bound regions revealed dynamic changes in the TF-binding site enrichment during differentiation, suggesting cooperative binding with distinct partners at different time-points (Fig. 4g and Supplementary Data 5). In EpiSCs and at 6 h, when HOX are absent, we identified the PBX3 and PREP1 binding motifs³³, whereas at later time-points, once Hox genes are expressed, the HOX recognition sites featured prominently. Notably during the 12-24 h time window that includes the NMP-to-PSM transition, we identified LEF1 among the top consensus motifs nearby the PBX-bound regions. These data suggest that LEF1 and PBX might bind cooperatively to PSM genes, and provided a possible reason why Pbx-DKO cells cannot complete PSM maturation despite the presence of WNT ligands (Supplementary Fig. 2d).

**Fig. 4: PBX proteins occupy regulatory regions associated with PSM genes and remodel chromatin accessibility.**

PBX proteins are interpreters of LEF1-mediated WNT signalling

Like PBX, the WNT effectors (TCFs/LEF1) exhibit low binding specificity³⁴. Therefore, the observation that PBX and LEF1 could be part of a common regulatory module supported the hypothesis of a PBX-LEF1 DNA-binding cooperativity at PSM genes. LEF1 and TCF7 are the most prominent WNT effectors involved in paraxial mesoderm development, but while the latter is broadly expressed in the tailbud, LEF1 is specifically localised in PSM and posterior somites³⁵ (Fig. 5a). Notably, LEF1 expression is expanded in E8.5 Pbx-DKO tailbuds, suggesting that the Pbx mutant phenotypes are not due to a reduction in LEF1 levels (Fig. 5a). To assess whether PBX and LEF1 TFs cooperatively control common transcriptional targets, we exploited our in vitro differentiation system, and performed ChIP-seq for LEF1. We found that LEF1 is de novo and progressively recruited to chromatin between 12 and 24 h, concomitantly with the formation of MPCs (Fig. 5b). Remarkably, LEF1 recruitment is abolished in mutant cells (Fig. 5b), despite sustained LEF1 expression (Fig. 5a and Supplementary Fig. 5a). By integrating our ChIP-seq datasets for PBX and LEF1 (Fig. 5c and Supplementary Data 6) and RNA-seq of WT and Pbx-DKO cells at 24 h (Supplementary Fig 5b and Supplementary Data 6), we observed that over 60% of the LEF1 targets (including Sox2, T-Bra, Cdx2, Msgn1 and Aldh1a2) were co-bound by PBX (Fig. 5c) and misregulated in Pbx-DKO (Supplementary Fig. 5b). As expected, GO-term analyses underscored targets associated with WNT activity and somitic features (Fig. 5d). We observed that while the occupancy of PBX and LEF1 occurs at variable distances on different regulatory regions, the majority (61%) of the LEF1 ChIP-seq peaks falls in close proximity to the PBX peaks (Supplementary Fig. 5c). ATAC-seq analyses of PBX-LEF1 common targets revealed that chromatin accessibility at LEF1-bound regions decreased for downregulated genes (Supplementary Fig. 5d) and remained unchanged for upregulated genes (Supplementary Fig. 5e), as illustrated for Msgn1, Aldh1a2, Sox2 and Cdx2 (Fig. 5e). These data demonstrate that PBX proteins authorise WNT-dependent transcription by making PSM genes accessible to LEF1. Instead, in the absence of PBX, the inability to activate the PSM fate programme triggers the feedforward mechanisms sustaining NMP expansion.

**Fig. 5: PBX complexes remodel the chromatin environment for LEF1 recruitment.**

Distinct PBX complexes are sequentially recruited to the Msgn1 regulatory region

To establish how the PBX-LEF1 module operates, we focussed on the master regulator of the paraxial mesoderm fate, Msgn1, which controls both cell migration and differentiation¹⁰. Undifferentiated progenitors accumulate in Msgn1 mutant tailbuds comparably to Pbx-DKO embryos^10,14. Notably, Msgn1 expression was lost in Pbx-com embryos (Fig. 6a) and strongly downregulated in Pbx-DKO cells upon PSM differentiation (Fig. 6b). Moreover, the MSGN1 target Snai1¹⁰ was downregulated in E8.5 Pbx-DKO and Pbx-com PSM (Supplementary Fig. 6a), suggesting that epithelial-to-mesenchymal transition (EMT) and cell migration are affected in absence of PBX. To assess the cell migratory capacity of WT and mutant MPCs, we visualised PSM differentiation in time-lapse videos (Supplementary Movies 1 and 2 and Supplementary Fig. 6b). We observed that WT cells leave the NMP zone upon differentiation, while Pbx-DKO cells fail to migrate and accumulate in the expanding progenitor area. Lentiviral re-expression of Msgn1 in Pbx-DKO cells during differentiation resulted in downregulation of NMP genes and upregulation of PSM genes (Supplementary Fig. 6c). These data confirmed that PBX-LEF1 direct PSM development and EMT by controlling Msgn1 expression.

**Fig. 6: Sequential recruitment of PBX/HOX-1 complexes to *Msgn1* promoter changes the chromatin landscape and activates transcription.**

Msgn1 is the earliest expressed paraxial mesoderm gene and is tightly regulated in vivo and in vitro, being transcribed in the 12–36 h window corresponding to the NMP-to-PSM transition (Fig. 6b). ChIP-seq data showed that PBX and LEF1 are dynamically recruited to the Msgn1 promoter, mirroring Msgn1 temporal expression, and LEF1 occupancy is lost in Pbx-DKO cells (Fig. 6c).

To identify the factors that temporally recruit LEF1 on Msgn1 promoter, we focussed on PBX and their combinatorial logic³⁶. Based on the temporal expression of HOX and TALE family members, we identified several candidates such as the heterodimerizing partner Meis2 and group-1 Hox genes (Hox-1) (Supplementary Fig. 7a) that are expressed during the period of PBX occupancy at the Msgn1 promoter. In addition, a second PBX-interacting factor, PREP1, is expressed earlier than Msgn1 (Fig. 6d). Considering the MPC-specific enrichment of Pbx1, Meis2 and Hox-1 (Fig. 6e) these proteins represent the best candidates for regulating Msgn1. We hypothesised that two putative complexes would be sequentially recruited, with PREP1/PBX binding prior to the PBX/MEIS2/HOX-1 complex. Consistent with this idea, ChIP-qPCR showed that PREP1 is recruited to Msgn1 first, at 12 h, while MEIS2 and PBX are recruited later, at 12–24 h (Fig. 6f). Examining the Msgn1 regulatory region, we identified a PBX/HOX-1 responsive site³⁷ within the region of PBX occupancy (Fig. 6g and Supplementary Fig. 7b). Electrophoretic Mobility Shift Assay (EMSA) with either in vitro-translated proteins or 24 h nuclear extracts highlighted the PBX/MEIS2/HOX-1 ternary complex formation on the oligonucleotide containing the PBX/HOX-1-binding site (Msgn1-P2) but not on its mutated variant (Msgn1-P2-M) (Fig. 6h). In agreement with the sequential recruitment to the Msgn1 regulatory region, we found that while the assembly of PBX/MEIS2 on Msgn1-P2 depended on the presence of HOX-1 (Supplementary Fig. 7c), the PREP1/PBX heterodimers could bind in its absence (Supplementary Fig. 7d). These binding properties suggest that the expression of HOX-1 drives the assembly of a new complex at the Msgn1 promoter that would be responsible for LEF1 occupancy and WNT-mediated activation. Consistent with this idea, affinity binding assays confirmed that the PBX/MEIS2/HOX-1 complex has higher affinity for Msgn1-P2 than PBX/PREP1 complexes, suggesting that it could compete for binding once HOX-1 is expressed (Supplementary Fig. 7e–g). Moreover, we observed that changes in the accessibility of Msgn1 regulatory regions were correlated with Hox-1 expression, and concurrent with PBX-LEF1 recruitment. Accordingly, in Pbx-DKO the accessibility of the loci containing the PBX/HOX and LEF1-binding sites was reduced at 24 h (Fig. 6i), highlighting a potential role of PBX/HOX complexes to nucleate their own binding and their ability to destabilise the nucleosomes. These data uncover a hierarchical TF-binding network on the Msgn1 locus, but do not conclusively demonstrate that PBX/HOX are directly involved in chromatin remodelling, prompting us to assess the recruitment of additional pioneer factors.

TALE complexes direct chromatin remodelling at the Msgn1 regulatory region, synergising with T-BRA

T-BRA is a pioneer factor essential for priming mesodermal genes³⁸. ChIP-qPCR analyses showed that T-BRA is already recruited at 12 h to the LEF1-binding region of Msgn1, and its occupancy decreases over time (Fig. 7a). Interestingly, T-BRA occupancy is not abrogated in Pbx-DKO (Fig. 7a), suggesting that the binding of T-BRA on Msgn1 is not sufficient to drive paraxial mesoderm differentiation and that PBX complexes remodel the chromatin independently of T-BRA.

**Fig. 7: Combined pioneering activity of T-BRA and TALE proteins remodels the chromatin, making it accessible at *Msgn1* regulatory regions, thereby directing WNT transcriptional response.**

To establish the direct consequence of the binding of PBX complexes on chromatin accessibility and their effects on LEF1 recruitment, we employed CRISPR/Cas9 technology to generate lines carrying the pMsgn1 point mutations (pMsgn1-mut) studied by EMSA (Fig. 6g, h and Supplementary Fig. 8a, b), preventing the binding of PBX complexes to the Msgn1 promoter but allowing their recruitment to other targets. RT-qPCR analyses showed that the expression of Msgn1 and its downstream targets were affected in pMsgn1-mut cells (Fig. 7b), and PSM differentiation was impaired (Fig. 7c), thus recapitulating at least in part the Pbx-DKO phenotype. However, the mesodermal phenotype of pMsgn1-mut lines was exacerbated, as shown by the absence of T-Bra, the paraxial mesodermal gene Tbx6, and the EMT markers Cdh2 and Snai1. ChIP-qPCR revealed that in pMsgn1-mut neither PBX nor LEF1 were recruited to Msgn1 (Fig. 7d). Interestingly, the recruitment of T-BRA was also affected (Fig. 7d) in pMsgn1-mut, and ATAC-seq analyses showed that both LEF1 and PBX/HOX-binding regions were inaccessible in pMsgn1-mut cells (Fig. 7e), unequivocally confirming the role of PBX complexes as pioneer factors and modulators of chromatin accessibility. Furthermore, the reduced recruitment of T-BRA reveals that target gene selection specificity is achieved through the collaborative effort of multiple pioneer TFs.

In conclusion, paraxial mesoderm induction relies on a WNT-HOX integrated code that is coordinated by the TALE complexes, the WNT-effector LEF1 and the synergistic activity of T-BRA. While T-BRA broadly assesses mesodermal genes, TALE/HOX-1 interpret WNT signalling response by specifically recruiting LEF1 at the PSM genes, resulting in paraxial mesoderm differentiation via a Msgn1-mediated transcriptional programme.

Discussion

In this work, we dissected how TALE complexes control the GRNs that alter the NMP response to WNT signalling and activate the paraxial mesoderm programme. We established that TALE/HOX and WNT signalling work through a common genetic code that simultaneously represses NMP genes and activates the PSM differentiation programme. We found that switching of the PBX-binding partners result in TALE/HOX high-affinity complex formation that destabilises the nucleosomes at WNT-responsive elements enabling de novo recruitment of the WNT-effector LEF1, thereby challenging the consolidated model of WNT-mediated response centred on the pre-bound activity of TCFs/LEF.

This concerted TALE-LEF1 activity represses NMP genes (like Sox2, T-Bra and Cdx2), ensuring silencing of the neural and NMP transcriptional programmes in PSM-fated cells. Concurrently, TALE-LEF1 synergistically drive the expression of PSM genes including the master regulator Msgn1, which in turn controls EMT and the migration of MPCs from the progenitor zone to the anterior PSM. Loss of the TALE activity results in the accumulation of progenitors trapped in a NMP/MPC state that ultimately causes an enlarged tailbud and impaired somitogenesis (Fig. 8a).

**Fig. 8: TALE complexes promote WNT-mediated transcriptional response in paraxial mesoderm.**

In vertebrates, Hox genes are collinearly expressed during axial elongation, establishing the HOX code with a spatio-temporal gradient along the anterior-posterior axis^19,20,39. This sequential activation provides NMPs with positional information and prompts MPCs to progressively adopt a more posterior identity¹¹, ensuring that cell fate and position are intrinsically linked. Our data suggest that group-1 HOX are also indispensable for the initial specification of MPCs and their migration to form PSM; thereby balancing the residence time in the progenitor zone and the flux of PSM patterning the trunk region. In Pbx mutants, 5′ Hox genes are overexpressed (Supplementary Fig. 1h) in NMP/MPCs, suggesting the existence of progenitors with a posterior identity and thereby a potential posteriorisation of the PSM, in agreement with the observation of a rostral shift of the axial Hox gene expression in the Pbx-com mutants²⁵.

In addition to providing HOX with binding specificity³⁷, PBX have also been indicated as pioneer factors^29,31. However, a key unanswered question is how such broadly expressed TFs can exquisitely open lineage-specific loci in a space and time-dependent manner. Exploiting the cooperative interactions among TALE and HOX proteins, and focussing on the tightly regulated spatio-temporal expression of Msgn1, we uncovered the sequential recruitment of TALE factors generating, at least in vitro, a high-affinity (PBX/MEIS2/HOX-1) ternary complex that could promote chromatin accessibility of the Msgn1 promoter (Fig. 8b). While the binding of the PBX/MEIS2 complex occurs concomitantly with increased chromatin accessibility and LEF1 recruitment, the low-affinity PBX/PREP1 dimers are recruited earlier, suggesting that they could be important to scan the DNA or increase the local concentration of PBX proteins, generating the TALE hubs that assist the loading of the TALE/HOX complex⁴⁰. The observation that PBX are pre-bound on regulatory regions prior to gene activation^29,30,31 has been used as proof for their pioneering activity. However, there are no data conclusively showing that they unlock closed chromatin. The pioneer activity of TFs is usually determined by measuring reduced chromatin accessibility in wild-type and knockout cells³⁸, that however can also be caused via indirect mechanisms. Here, using precise CRISPR-mediated mutagenesis, we unequivocally demonstrate that point mutations affecting the binding of PBX/HOX to the endogenous Msgn1 promoter reduce chromatin accessibility and LEF1 recruitment, indicating that PBX directly modulate chromatin accessibility and context-dependent WNT responses.

Analyses of PBX and LEF1 occupancy revealed that the distance between them is flexible and varies at different regulatory regions. This arrangement could support cooperative binding, allowing the specific recruitment of TFs that further stabilise the nucleosome-depleted regions triggered by the PBX complexes. Thus, changes in chromatin accessibility initiated by the PBX complexes could mediate recruitment of additional factors, including ATP-dependent chromatin remodelling factors. PBX/HOX could interact with histone modifying complexes like CBP, TrxG or HDAC and PcG proteins⁴¹ or may affect chromatin conformation via recruitment of histone variants⁴². TF-binding sites located in closed chromatin could provide an additional layer of specificity by restricting the access to highly specific PBX/HOX complexes^43,44. Interestingly, the NFY-binding site appeared among the most enriched motifs identified within the PBX-bound regions at all time points of differentiation⁴⁵. In zebrafish, NFY has been demonstrated to directly interact with the TALE proteins to promote access to enhancers, suggesting that it could play a similar role in paraxial mesoderm differentiation. For all these reasons, it is plausible that chromatin landscape changes are driven by combined effects on DNA-histone affinity and recruitment of nucleosome remodelling factors.

Intriguingly, the TALE-WNT integrated code is conserved in the regulatory regions of PSM genes within eutherians where axial elongation relies on expanding NMPs, but is absent in lower vertebrates where PSM maturation is supported by depleting progenitors^6,46. Of note, the TALE-WNT integrated motif is present at the Sox2 enhancer N1⁴⁷ that is exclusively associated with expansion of NMPs in amniotes. Thus, during vertebrate evolution, WNT, TALE and HOX proteins could have been co-opted by eutherians to balance expansion and differentiation of paraxial mesoderm progenitors for axial elongation.

Prevailing models of WNT activity suggest that TCFs/LEF work through a transcriptional switch, binding and repressing genes in unstimulated conditions and promoting transcription upon WNT induction³⁴. Here we observed that during paraxial mesoderm induction, LEF1 is de novo recruited when WNT signalling is induced and when the PSM genes become accessible. Thus, in this context, the competence for WNT signalling response is established during the transition of NMPs to PSM. This alternative mechanism supports the concept that distinct modes exist to prime chromatin accessibility and lineage choice. Crucially, we found that signalling response is dictated by TFs unlocking the WNT-responsive elements rather than by WNT ligand dosage, questioning the current concentration-dependent morphogen model⁴⁸.

Given that WNT/LEF1 signalling is important for homeostasis and regeneration of the skin⁴⁹, it will be compelling to address if the TALE/HOX motif also unlocks the regenerative transcriptional programmes in the skin, thus having potential therapeutic applications. Furthermore, abnormal activity within this pathway is linked to hyperproliferation and EMT-driven metastatic processes⁵⁰. In this context, it will be important to establish if aberrations in TALE-WNT integrated codes could also be associated with carcinogenesis and cancer invasion⁵¹.

Finally, the observation that TALE complexes collaborate with the pioneer factor T-BRA³⁸ provides a temporal code for WNT-mediated response. While T-BRA has a pan-mesodermal role, the TALE complexes provide specificity towards the paraxial mesodermal fate. Hence, conventional cooperativity can be extended to pioneer factor clusters that synergistically generate a permissive chromatin landscape and collectively operate as key determinants of lineage specification. Given that the GRNs governing paraxial mesoderm differentiation¹⁷ are conserved between human and mouse, these findings can be used to improve the current protocol of hESCs differentiation.

Altogether, this work uncovers a cooperative TALE/HOX-WNT integrated code unlocking the spatio-temporal GRNs that control axial extension. During axial elongation, similar modules could also be employed by 5’ HOX proteins to further fine-tune WNT signalling response to provide MPCs with a more posterior fate. Thus, integrating Hox collinearity and WNT signalling on a common genomic element may represent a powerful strategy that confers mesodermal progenitors with a specific spatio-temporal identity.

Methods

Cell Lines

EpiSC lines were maintained in six-well cell-culture plates (Corning, 3506) pre-coated with 15 µg/ml human Fibronectin (Merck Millipore, FC01015) in N2B27 medium (prepared as described in Ying et al.⁵², supplemented with 20 ng/ml Activin A, 10 ng/ml bFGF and 1 µM XAV939 (EpiSC medium). All lines used in this study were routinely karyotyped and tested for mycoplasma.

Mice

The mutant alleles used in this study were previously described: Pbx1 conditional allele in Koss et al.⁵³ and Pbx2^−/− mice in Selleri et al.⁵⁴ The Cre deleter strain, Sox2^Cre knock-in mouse⁵⁵, was utilised for Pbx1 inactivation in the inner cell mass. All animals used in this study were maintained in laboratory animal housing facilities at macroenvironmental temperature and humidity ranges of 20–24 °C and 45–65%, respectively, with a 12/12 h light/dark cycle. Embryos were dissected at E8.5/E9.0 and genotyped by PCR as described in Koss et al.⁵³. All animal work was carried out in accordance with European legislation. All work was authorised by and carried out under the Project License 2017-15-0201-01255 issued by the Danish Regulatory Authority.

Generation of Pbx-DKO lines using the CRISPR/Cas9 system

EpiSC lines derived from E6.5 mouse embryos (129S6/SvEvTac background) were a kind gift from Dr. M.M. Rothová (The Novo Nordisk Foundation Center for Stem Cell Biology, University of Copenhagen, Copenhagen, Denmark). The lines were derived in EpiSC medium as previously described⁵⁶. Pbx-DKO lines were established via sequential CRISPR/Cas9-mediated gene targeting. First, Pbx2-KO lines were generated by inserting a floxed CMV-hygromycin-Poly (A) cassette into Exon 1 of Pbx2 to disrupt Pbx2 gene expression, employing the CRISPR Nickase system⁵⁷. Guide RNAs (gRNAs) were designed using the CRISPR/Cas9 design tool (crispr.mit.edu) and cloned into pX335 (a gift from Prof. F. Zhang; Addgene, 42335). Cells were co-transfected with a pair of gRNA-pX335 plasmids (gRNA 1: GTCCAGGCTCTGGTCCGTGA; gRNA 2: CCAAGTGAGTGCCCCCACCC; 100 ng each vector) and 1 µg linearised targeting vector using Lipofectamine 2000 (Thermo Fisher Scientific, 11668030), seeded at low density in EpiSC medium in 100 mm tissue-culture dishes (Corning, 430293) pre-coated with 15 µg/ml human Fibronectin, and selected for hygromycin resistance for 8–10 days. Single colonies were screened by diagnostic PCR, Southern blot and western blot analyses. To reduce side effects caused by CRISPR/Cas9 off-targets, we subsequently employed the High-fidelity CRISPR/Cas9 system (Cas9-HF1)⁵⁸ to target Exon1 of Pbx1 in Pbx2-KO EpiSC lines. The gRNA used (GGCCATCCCGACCCCAGCGT) was cloned into a modified version of pX458 (a gift from Prof. F. Zhang; Addgene, 48138) containing the sequence encoding for Cas9-HF1 (a kind gift from Dr. F.V. Roske (Novo Nordisk, Denmark)). Cells were transfected with 300 ng gRNA-pX458 and after 48 h GFP^pos cells were re-seeded at low density in EpiSC medium in Fibronectin-coated 100 mm tissue-culture dishes. After 6-8 days, single colonies were screened by diagnostic PCR and selected Pbx-DKO lines were expanded and validated by Western blot.

Generation of Pbx-DKO lines by CRE-mediated recombination

EpiSC lines were derived in EpiSC medium as previously described⁵⁶ by Dr. J. Martin Gonzalez (Core Facility for Transgenic Mice, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark) from mixed background Pbx1^fl/fl;Pbx2^−/− mouse embryos. Additionally, Yfp (yellow fluorescent protein) cDNA was inserted at the Rosa26 locus preceded by a loxP-flanked stop sequence to generate CRE recombination reporter lines. After isolation, cell lines were cultured as described above for a few passages, transfected with 200 ng pCAG-Cre (a gift from Prof. C. Cepko; Addgene, 13775) using Lipofectamine 2000 and re-seeded at low density in EpiSC medium in 100 mm tissue-culture dishes pre-coated with 15 µg/ml human Fibronectin. After 6–8 days, single YFP^pos colonies were picked and screened by PCR. The selected lines were expanded and their genotype was further validated by western blot.

Generation of pMsgn1-mut lines using the CRISPR/Cas9 system

To generate the pMsgn1-mut lines, CRISPR/Cas9-mediated homology recombination was performed on 129 EpiSCs using a single-stranded DNA (ssDNA) donor template carrying four base substitutions that abrogate the recruitment of PBX to the Msgn1 promoter, as assessed by EMSA (Fig. 6g–h). The gRNA designed for targeting the Msgn1 regulatory region (shown in Supplementary Fig. 8a) was cloned into pX459 (a gift from Prof. F. Zhang; Addgene, 62988). Cells were co-transfected with 200 ng gRNA-pX459 and 500 pM ssDNA donor template using Lipofectamine 2000, re-seeded at low density in EpiSC medium in 100 mm tissue-culture dishes pre-coated with 15 µg/ml human Fibronectin and selected for neomycin resistance. Single colonies were screened by diagnostic PCR and restriction enzyme analysis. The selected lines were expanded and the gene editing was further validated by sequencing.

EpiSC differentiation to PSM and derivatives

All differentiation was conducted in serum-free, feeder-free and monolayer conditions in chemically defined N2B27-based media. Culture conditions followed a modified protocol from Loh et al.⁵⁹. EpiSCs were seeded at ~6 × 10³ cells/cm² in six-well cell-culture plates pre-coated with 15 µg/ml human Fibronectin, and cultured in EpiSC medium for 24 h. EpiSCs were first differentiated towards APS using 30 ng/ml Activin A, 4 µM CHIR-99021, 20 ng/ml bFGF and 100 nM PIK-90 (APS medium) for 12 h, and then to PSM using 1 µM A-83-01, 3 µM CHIR-99021, 250 nM LDN193189 and 20 ng/ml bFGF (PSM medium) for 36 h. For PSM derivatives, cells were kept in PSM medium for additional 24 h, and then differentiated towards ventral somites/sclerotome using 5 nM SAG 21k and 1 µM C59 (sclerotome medium) or dorsal somites/dermomyotome using 3 µM CHIR-99021 and 150 nM GDC-0449 (dermomyotome medium) until day 5. A complete list of cytokines and signalling pathway inhibitors used for cell maintenance and differentiation is provided in Supplementary Table 1.

Lentivirus-mediated exogenous expression of Msgn1

The murine Msgn1-coding sequence was amplified from cDNA, subcloned into the VIRSP lentiviral vector between the restriction sites NheI and XhoI and verified by sequencing. In order to produce the lentivirus carrying the VIRSP-Msgn1 vector, HEK293T cells were seeded at ~9 × 10⁴ cells/cm² in 100 mm tissue-culture dishes in Dulbecco’s Modified Eagle’s Medium (DMEM, Thermo Fisher Scientific, 10566016) + 10% (v/v) Fetal Bovine Serum (FBS, Gibco, 10270-106). The following day, the medium was changed to N2B27, 1 h prior to transfection. Cells were transfected with 15 µg VIRSP-Msgn1, 10 µg psPAX2 (a gift from Prof. D. Trono; Addgene, 12260), 5 µg pCMV-VSV-G (a gift from Prof. R.A. Weinberg; Addgene, 8454) and 90 µl PEI (1 mg/ml, Sigma Aldrich, 765090). The supernatant containing the lentivirus was collected after 72 h and concentrated using ultracentrifugation filter units (Merck Millipore, UFC910008) as per manufacturer’s instructions. The concentrated supernatant was stored at −80 °C. To perform the rescue experiment, WT and Pbx-DKO 129 EpiSCs were seeded in six-well cell-culture plates pre-coated with 15 µg/ml human Fibronectin. In all, 3.5 µl of lentivirus were added 4 h after seeding, and then differentiation to PSM was carried out as described above. Total RNA was extracted from samples harvested after 48 h of differentiation and the rescue efficiency was assessed by RT-qPCR as compared to WT EpiSCs with and without added virus, as well as to Pbx-DKO cells without added virus.

Immunofluorescence staining of adherent cells

EpiSCs were seeded at ~4 × 10³ cells/cm² in EpiSC medium in eight-well microscopy slides (Ibidi, 80826) pre-coated with 15 µg/ml human Fibronectin, and differentiated to PSM as described above. At desired time-points, cells were briefly washed twice in PBS (Sigma, D8537) and fixed in 4% (w/v) paraformaldehyde diluted in PBS from 16% methanol-free formaldehyde stock (Pierce Biotechnologies, 28906) for 15 min at RT. Fixed cells were then permeabilised in PBS containing 0.1% (v/v) Triton X-100 (PBST) for 10 min at RT and blocked in 3% (v/v) Donkey serum in PBST containing 1% (w/v) BSA (blocking solution) for 15 min at RT. Primary and secondary antibody incubations were done in blocking solution overnight at 4 °C and for 2 h at RT, respectively. Antibody incubations were followed by three quick washes in PBST followed by two washes in PBST for 15 min at RT. After the final incubation, cells were stored in PBST containing 1 µg/ml DAPI at 4 °C, packed in Parafilm to prevent evaporation. Cells were always kept in the dark after incubation with secondary antibodies. Confocal imaging was performed using a Leica TCS SP8 microscope and images were processed using the Imaris^TM software (v9.3.0) with default settings. The antibodies used are listed in Supplementary Table 2.

Imaging of embryo morphology

Embryos were dissected in cold PBS, fixed in 4% (w/v) paraformaldehyde for 30 min at RT and washed twice in cold PBS. Images were acquired using a Zeiss AxioZoom.V16 microscope at the Core Facility for Integrated Microscopy (University of Copenhagen, Copenhagen, Denmark).

Immunofluorescence staining of embryos

Embryos were dissected in cold PBS and fixed in 4% (w/v) paraformaldehyde for 30 min at RT. Fixed embryos were rinsed in cold PBS containing 3 mg/ml PVP (PBS/PVP), permeabilised in PBS/PVP + 1% (v/v) Triton X-100 for 10 min at RT and blocked in CAS-Block (Thermo Fisher Scientific, 008120) overnight at 4 °C. Incubation with the primary antibodies listed in Supplementary Table 3 was carried out in CAS-Block for 48 h at 4 °C. Embryos were quickly washed three times with CAS-Block at RT and then overnight at 4 °C. Incubation with secondary antibodies was done in CAS-Block for 24 h at 4 °C. Embryos were quickly washed three times with CAS-Block at RT and then overnight at 4 °C. Following incubation with CAS-Block containing 1 µg/ml DAPI for 1 h at RT, embryos were washed three times with PBS + 0.5% (v/v) Triton X-100 for 15 min at RT and dehydrated in ascending series to 100% methanol. To reduce light scattering, optical clearing was performed by submerging the stained embryos in a 1:2 (v/v) solution of benzyl alcohol:benzyl benzoate (BABB). Images were acquired using a Leica TCS SP8 confocal microscope. Image processing and 3D reconstruction were performed using the Imaris^TM software (v9.3.0). In Figs. 2c, d and 3i and Supplementary Fig. 2b, only cells expressing high levels of either T-BRA, SOX2 or TBX6 were included in the scoring.

In situ hybridisation

Embryos were dissected in cold PBS and fixed in 4% (w/v) paraformaldehyde for 2 h at RT, washed twice in cold PBS with 0.1% Tween-20 and dehydrated in ascending series to 100% methanol. At least three embryos of the same genotype were used. Whole-mount in situ hybridisation was performed following the protocol used in Prof. E. De Robertis’ laboratory (http://www.hhmi.ucla.edu/derobertis/protocol_page/Pdfs/Mouse%20protocols/Whole%20mount%20in%20situ%20hybridization%20on%20mouse%20embryos.pdf). The antisense riboprobes to Msgn1⁶⁰ and Wnt3a⁶¹ were a kind gift from Prof. A.K. Hadjantonakis (Memorial Sloan Kettering Cancer Center, New York, United States).

Time-lapse videos of PSM differentiation

EpiSCs were seeded at ~6 × 10³ cells/cm² in six-well cell-culture plates pre-coated with 15 µg/ml human Fibronectin, and differentiated to PSM as described above. Differentiation was carried out in the IncuCyte S3 Live-Cell Analysis System (Essen Bioscience) at the High-Throughput Cell-based Screens Facility (Biotech Research & Innovation Centre, University of Copenhagen, Copenhagen, Denmark), and images were acquired every h for 3 days. Time-lapse movies with all acquired pictures were further recreated using iMovie (version 10.1.14).

Western blot

In all, 80% confluent cells were washed once with PBS and then lysed directly on the plate by addition of 1x Laemmli sample buffer (2% (w/v) SDS, 10% (v/v) glycerol and 120 mM Tris-HCl pH 6.8) to the culture dish (20 µl/cm²), followed by extensive scraping using a pipette tip. Samples were vortexed, centrifuged briefly and then stored at −20 °C until use. Before loading, samples were boiled for 5 min at 80 °C, sonicated on ice for 10 s at 20% power using a Branson Digital Sonifier Model S-450D (Marshall Scientific), spun for 10 min at 14,000 × g to clear the lysates and then stored at −20 °C until use. Protein concentration was estimated using NanoDrop (Thermo Fisher Scientific) A280 measurements and equal amounts of proteins were diluted in 60 µl of 1x Laemmli sample buffer per sample. After addition of 6 µl of 1 M DTT containing 0.01% (w/v) bromophenol blue, 50 µl of sample were loaded per lane on NuPAGE Novex 4–12% Bis-Tris Protein Gels (Thermo Fisher Scientific, NP0321BOX). Electrophoresis was performed in 1x NuPAGE MES SDS running buffer (Thermo Fisher Scientific, NP0002) at 150 V for 1 h. Resolved proteins were transferred to Amersham Protran Nitrocellulose Blotting membrane 0.45 µm (GE Healthcare, 10600003) at 400 mA for 90 min on ice in 1x NuPAGE transfer buffer (Thermo Fisher Scientific, NP0006). After a quick wash in PBS containing 0.1% (v/v) Tween-20 (PBST), membranes were blocked for 1 h at RT in PBST containing 10% Skim milk powder. All subsequent antibody incubations were performed in PBST containing 5% (w/v) BSA and followed by three washes in PBST for 10 min. The following primary antibodies were used: polyclonal Rabbit anti-PBX1 (CST, 4342, 1:1,000), polyclonal Rabbit anti-PBX2 (Santa Cruz Biotechnology, sc-890, 1:3,000) and monoclonal Mouse anti-H3 (abcam, ab10799, 1:30,000). The secondary antibodies used include: Donkey anti-Rabbit HRP (Thermo Fisher Scientific, A16023, 1:10,000) and Donkey anti-Mouse HRP (Thermo Fisher Scientific, A16011, 1:10,000). For chemiluminescence detection, the membrane was incubated with Amersham ECL Prime Western Blotting Detection reagent (GE Healthcare, RPN2236) and imaged on a ChemiDoc MP (Bio-Rad). For reprobing with different primary antibodies, the membrane was soaked in Restore PLUS Western Blot Stripping Buffer (Thermo Fisher Scientific, 46430) two times for 10 min at RT to inactivate horseradish peroxidase (HRP). To ensure comparable densitometry results, gel electrophoresis, protein transfer and antibody staining were performed in parallel for all samples used to generate each western blot figure panel.

Electrophoretic Mobility Shift Assay

EMSAs were performed as described⁶², using in vitro-translated proteins or nuclear extracts purified from cells. Sequences of the oligonucleotides spanning the PBX/HOX-1-binding site on the Msgn1 promoter used in EMSA are described in Supplementary Fig. 7a. EMSAs with in vitro-translated proteins were performed using 5 μg of nuclear extract or 2 μl of reticulocyte lysate containing the desired combination of proteins mixed with 18 μl of PPH binding buffer (10 mM Tris-Cl pH 7.5, 75 mM NaCl, 1 mM EDTA, 6% (v/v) glycerol, 3 mM spermidine, 1 mM DTT, 0.5 mM PMSF, 1 μg poly (dIdC), 30,000 c.p.m ³²P-labelled oligonucleotide) to a total volume of 20 μl. After 30 min of incubation on ice the reactions were separated by 5% acrylamide gel in 0.5x TBE. For the competition assays 50, 100 or 200-molar excess of unlabelled competitor oligonucleotide were added to the binding reaction 10 min before the labelled probe. Acrylamide gels were pre-run at 300 V for 90 min. Samples were separated on gel at 150 V for 2 h. Gels were dried on gel dryer, exposed overnight and imaged on a Typhoon 9400 imager (GE Healthcare). Based on the known molecular weight of the in vitro-translated proteins, the expected sizes of dimeric and ternary complexes are as follows: Fig. 6h: PBX1A/MEIS2 = ~114 kDa, PBX1B/MEIS2 = ~104 kDa, PBX2/MEIS2 = ~114 kDa, PBX1A/MEIS2/HOXA1 = ~149 kDa, PBX1B/MEIS2/HOXA1 = ~139 kDa, PBX2/MEIS2/HOXA1 = ~149 kDa; Supplementary Fig. 7c: PBX1A/MEIS2 = ~114 kDa, PBX1B/MEIS2 = ~104 kDa, PBX2/MEIS2 = ~114 kDa, PBX3/MEIS2 = ~114 kDa, PBX1A/MEIS2/HOXA1 = ~149 kDa, PBX1B/MEIS2/HOXA1 = ~139 kDa, PBX2/MEIS2/HOXA1 = ~149 kDa, PBX3/MEIS2/HOXA1 = ~149 kDa, PBX1A/MEIS2/HOXB1 = ~146 kDa, PBX1B/MEIS2/HOXB1 = ~136 kDa, PBX2/MEIS2/HOXB1 = ~146 kDa, PBX3/MEIS2/HOXB1 = ~146 kDa; Supplementary Fig. 7d: PBX1A/PREP1 = ~114 kDa, PBX1B/PREP1 = ~104 kDa, PBX2/PREP1 = ~114 kDa, PBX3/PREP1 = ~114 kDa, PBX1A/MEIS2 = ~114 kDa, PBX1B/MEIS2 = ~104 kDa, PBX2/MEIS2 = ~114 kDa, PBX3/MEIS2 = ~114 kDa; Supplementary Fig. 7e: PBX1A/PREP1 = ~114 kDa, PBX1A/PREP1/HOXA1 = ~149 kDa, PBX1A/MEIS2 = ~114 kDa, PBX1A/MEIS2/HOXA1 = ~149 kDa; Supplementary Fig. 7f: PBX3/PREP1 = ~114 kDa, PBX1A/MEIS2/HOXB1 = ~146 kDa; and Supplementary Fig. 7g: PBX1A/PREP1 = ~114 kDa, PBX1A/MEIS2/HOXB1 = ~146 kDa.

Oligonucleotides preparation

Oligonucleotides were resuspended in TE pH 8.0. Double-stranded oligonucleotides were prepared by mixing equal molar ratios of single-strand oligonucleotides. The mixture was heated at 90 °C for 10 min and slowly cooled at RT. Perfect oligo annealing was checked by electrophoresis on 15% acrylamide gel. For oligos longer than 40 bp, double-strand oligonucleotides purification was performed using preparative acrylamide 15% in 0.5x TBE. The oligonucleotides were eluted from the gel in 500 μl TEN (10 mM Tris, 1 μM EDTA, 300 mM NaCl, pH 8.0) with overnight shaking. The oligonucleotides were recovered by ethanol precipitation. The pellet was washed, dried and resuspended in 50 μl TE pH 8.0.

Oligonucleotides 5′ end labelling

In total, 0.5 pmoles of double-stranded oligonucleotides were incubated with 1 μl of 10x PNK buffer 5 Units Polynucleotide Kinase (Roche, 10174645001) and [γ-³²P]ATP 3000 Ci/mMole (PerkinElmer, BLU502A250UC). The reaction was incubated at 37 °C for 40 min, diluted to 50 μl with TE pH 8.0 and purified on a packed Sepharose G-25 column (Roche, 11273949001).

Nuclear extract preparation

Nuclear extracts (NE) for EMSA were prepared from in vitro-differentiated paraxial mesoderm cells. Cells were resuspended in 200 μl lysis buffer (10 mM HEPES pH 7.9, 30 mM KCl, 1.5 mM MgCl₂, 1 mM DTT, supplemented with 1x Protease inhibitor cocktail (Roche, 11836170001)), kept on ice for 10 min and supplemented with addition of Triton X-100 to a final concentration of 0.1%. The nuclei were collected by centrifugation at 4 °C. Nuclear extracts were prepared by resuspending the nuclear pellet in 50 μl nuclear buffer (20 mM HEPES pH 7.9, 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 1 mM DTT, supplemented with 1x Protease inhibitor cocktail). NE preparations were incubated at 4 °C with rotation for 30 min and cleared by centrifugation. Finally, the extracts were quantified by NanoDrop, aliquoted and stored at −80 °C.

Cell-free protein synthesis

The cDNAs were transcribed and translated using the couple T7-TNT reticulocytes lysate system (Promega, L1170) according to the manufacturer’s instructions. All cDNAs of the proteins used in this study (PBX1A, PBX1B, PBX2, PBX3A, PREP1, PREP2, MEIS2, HOXA1 and HOXB1) were cloned in vectors under the control of T7 promoter. The PBX3A (MC202386) and PREP2 (MC204820) constructs were purchased from OriGene Technologies. All the other plasmids were described in Ferretti et al.⁶².

scRNA-seq procedure

scRNA-seq was performed using the Massively Parallel Single-Cell RNA-sequencing technology (MARS-seq⁶³). For in vitro experiments, EpiSCs were seeded at ~6 × 10³ cells/cm² in six-well cell-culture plates pre-coated with 15 µg/ml human Fibronectin, and differentiated to PSM as described above. At desired time-points, cells were dissociated using Accutase (Thermo Fisher Scientific, 00-4555-56) for 3 min at 37 °C and counted. 500,000 cells per condition were washed with ice-cold FACS buffer (10% (v/v) FBS in PBS) and resuspended into 1 ml ice-cold FACS buffer containing 1 µg/ml DAPI. For in vivo experiments, tailbuds were dissected from E8.5/E9.0 embryos and dissociated into 300 μl warmed Trypsin 0.05% (w/v) EDTA for 10 min at 37 °C. Digestion was stopped by addition of 600 μl ice-cold FACS buffer followed by centrifugation at 350 × g for 3 min. Cells were washed with 500 μl ice-cold FACS buffer, centrifuged again at 350 × g for 3 min and resuspended into 250 μl ice-cold FACS buffer containing 1 µg/ml DAPI. Singlecells from in vitro-differentiated cells or embryonic tailbuds were sorted into Eppendorf Polypropylene U-shaped 384-well Twin Tec PCR Microplates (Thermo Fisher Scientific, 10573035), containing 2 μl of lysis solution (0.2% (v/v) Triton X-100) supplemented with 0.4 U/μl RNasin Ribonuclease Inhibitor (Promega, N2515) and 400 nM indexed RT primer from group 1 (1–96 barcodes) or group 2 (97-192 barcodes)⁶³. Additionally, 71 WT EpiSCs were sorted into each plate, as spike-in control for batch-effect correction. Capture plates were prepared on the Bravo automated liquid handling robot station (Agilent) using 384-filtered tips (Agilent, 19133-142). Index sorting was performed using a FACS Aria III cell sorter (BD Biosciences) at the DanStem Flow Cytometry Platform (University of Copenhagen, Copenhagen, Denmark), sorting live cells as singlets. Immediately after sorting, plates were spun down, snap-frozen on dry ice and stored at −80 °C until further processing. Semi-automated library preparation was performed, using 10–12 total cycles of PCR amplification and AMPure XP beads (Beckman Coulter Life Sciences, A63881) for purification⁶³. DNA concentration was measured with a Qubit Fluorometer (Thermo Fisher Scientific, Q32854) and fragment size was determined with a Fragment analyzer (Agilent). Libraries were paired-end sequenced on a Next-Seq 500 Sequencer (Illumina) at the DanStem Genomics Platform (University of Copenhagen, Copenhagen, Denmark). Between 1,146 and 1,528 cells were sequenced per lane.

RNA-seq procedure

EpiSCs were seeded at ~6 × 10³ cells/cm² in six-well cell-culture plates pre-coated with 15 µg/ml human Fibronectin, and differentiated to PSM as described above. At desired time-points, cells were lysed and total RNA was extracted using the RNeasy Mini Kit according to the manufacturer’s instructions (Qiagen, 74106). RNA concentration was quantified by NanoDrop and quality was verified with a Fragment analyzer. Library preparation was carried out using 0.5 μg of RNA with NEBNext Ultra II RNA library Prep Kit (New England BioLabs, E7770S) as per manufacturer’s instructions. Libraries were amplified for 4 total PCR cycles and purified with AMPure XP beads. DNA concentration was measured with a Qubit Fluorometer and fragment size was determined with a Fragment analyzer. All samples were sequenced in biological triplicates on a Next-Seq 500 Sequencer (Illumina) at the DanStem Genomics Platform (University of Copenhagen, Copenhagen, Denmark).

ATAC-seq procedure

ATAC-seq was performed on 50,000 cells per replicate as described in Buenrostro et al.⁶⁴ (with modifications based on Corces et al.⁶⁵), on EpiSCs and PSM-differentiated cell populations at desired time-points. Libraries were generated using the Ad1_noMX and Ad2.1–2.16 barcoded primers⁶⁴ and amplified for 10 total PCR cycles. Libraries were purified with AMPure XP beads to remove contaminating primer dimers and fragments >1,000 bp. Library quality was assessed using the Fragment analyzer and quantitated by Qubit assay. The libraries were sequenced with 50 bp paired-end reads on a Next-Seq 500 Sequencer (Illumina) at the DanStem Genomics Platform (University of Copenhagen, Copenhagen, Denmark).

ChIP-seq procedure

ChIP was performed with a protocol modified from that of Kolasinska-Zwierz et al.⁶⁶. EpiSCs were seeded at ~2 × 10³ cells/cm² in EpiSC medium in 150 mm tissue-culture dishes (Corning, 430599) pre-coated with 15 µg/ml human Fibronectin, and differentiated to PSM as described above. At desired time-points, cells were washed with PBS and fixed directly on the plate by the addition of 20 ml of PBS containing 1% (v/v) Formaldehyde Fixative Ultrapure EM Grade (Rockland, KHF001) for 8–10 min at RT. Fixation was stopped by 5 min incubation with 125 mM glycine at RT. After two washes with cold PBS, fixed cells were harvested in lysis buffer (50 mM Tris-HCl pH 8.1, 100 mM NaCl, 5 mM EDTA, 0.2% (w/v) NaN₃, 0.5% (w/v) SDS, supplemented with 1x Protease inhibitor cocktail and 1 mM PMSF) by extensive scraping using a pipette tip. Following centrifugation at 1,000 × g for 5 min at 4 °C, the pellet was resuspended in 1 ml of FA buffer (50 mM HEPES/KOH pH 7.5, 1 mM EDTA, 1% (v/v) Triton X-100, 0.1% sodium deoxycholate, 150 mM NaCl, supplemented with 1x Protease inhibitor cocktail) per 0.5 ml of pellet. Chromatin was disrupted by sonication using a chilled Bioruptor Plus sonicator UCD-300 (Diagenode) for 8–12 pulses (cycles of 30 s on/45 s off, high setting at 4 °C) to obtain fragments of 200–500 bp in size. The extract was spun for 30 min at 20,000 × g at 4 °C, and the soluble fraction was aliquoted and stored at −80 °C until use. For each IP, 20 μg of chromatin were incubated overnight in 1 ml of FA buffer containing 1% N-lauroylsarkosine sodium salt with the antibodies listed in Supplementary Table 4. 10% of the IP volume of the chromatin was retained as an input reference. After overnight rotation at 4 °C, 30 μl of washed and blocked magnetic Protein A (Thermo Fisher Scientific, 10006D) or Protein G Dynabeads (Thermo Fisher Scientific, 10003D) were added and the incubation continued for 2 additional h. Beads were extensively washed at RT⁶⁶. DNA was eluted twice with 57 μl of elution buffer (1% (w/v) SDS in TE pH 8.0 with 250 mM NaCl) at 65 °C, 15 min each time. Eluted DNA was incubated with 20 μg of RNase A for 30 min at 37 °C and then with 20 μg of Proteinase K for 1 h at 55 °C. Input DNA was also diluted in 114 μl of elution buffer and treated as ChIP samples. Crosslinks were reversed at 65 °C overnight and DNA was purified using the QIAGEN MinElute PCR Purification Kit (QIAGEN, 28006) according to the manufacturer’s instructions (Qiagen). DNA concentration was measured with a Qubit Fluorometer and fragment size was determined with a Fragment analyzer. For ChIP-seq, libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit (New England BioLabs, E7645S) as per manufacturer’s instructions. Libraries were amplified for 5–7 total PCR cycles, purified with AMPure XP beads and sequenced on a Next-Seq 500 Sequencer (Illumina) at the DanStem Genomics Platform (University of Copenhagen, Copenhagen, Denmark). For ChIP-qPCR, immunoprecipitated DNA and input were quantified by real-time qPCR as described below. The following primers and probes (Universal Probe Library, UPL, Roche) were used: for Msgn1 PBX/HOX-1-binding site, Fw, GTGCAGAAATATCCCCTGCT; Rv, TGTCTTCTGGTGCTGTTTGC; Probe, 95; for Msgn1 LEF1-binding site, Fw, GCATCTTCCAGCTAAAATGTCTTT; Rv, GCAGCCTTAATTGGTTTAGTTACAA; Probe, 75. The measures were normalised to the input. All reactions were performed in duplicates.

Real-time quantitative PCR

Total RNA was extracted from 0.5–1 × 10⁶ cells using the RNeasy Mini Kit according to the manufacturer’s instructions. First-strand synthesis was performed on 1 μg of total RNA using random hexamers (Thermo Fisher Scientific, N8080127) and SuperScript III reverse transcriptase (Thermo Fisher Scientific, 18080-044) according to the manufacturer’s instructions. A mix of concentrated cDNA from all samples was used to generate standard curves. Amplification was detected using the Universal Probe Library system on a LightCycler480 Real-Time PCR System (Roche). The measures were normalised to Tbp and Sdha reference genes. All reactions were performed in duplicates. See Supplementary Table 5 for a list of primers and probes used.

scRNA-seq data analysis

R1 and R2 fastq files were generated using bcl2fastq (v2.19.1), and the pooling and well information were extracted from the sequence using umis (v1.0.3) [https://github.com/vals/umis] into a unique fastq file. The reads were then filtered based on the pooling barcodes with 1 mismatch allowed. The poly-Ts at the end of the reads were trimmed using Cutadapt (v1.18)⁶⁷. The reads were mapped to the mouse genome (GRCm38/mm10 together with ERCC92) using HISAT2 (v2.1.0), the alignments were processed with Samtools (v1.7)⁶⁸, and the reads were counted with featureCounts (Subread (v1.5.3))⁶⁹ using Ensembl v93, and the umis using UMI_tools (v1.0.0)⁷⁰. Expression data were analysed using Seurat (v3.1)^71,72. Data filtering, normalisation and scaling were performed using the standard pre-processing workflow⁷¹. Integration of different datasets was performed as described in Stuart et al.⁷². Spiked-in EpiSCs were used as a reference to correct the batch-effect between integrated datasets. Marker genes of each cell cluster were outputted for GO-term analysis to define the cell type. The Monocle package (v2.16.0)⁷³ was used to perform pseudotime analyses. Single-cell trajectories were constructed in a semi-supervised manner, by defining NMPs as cells co-expressing moderate levels of T-Bra/Sox2, MPCs as co-expressing high levels of T-Bra/Tbx6 and PSM as cells co-expressing high levels of Tbx6/Meox1.

RNA-seq data analysis

Fastq files were aligned to the GRCm38/mm10 genome using STAR (v2.5.3a)⁷⁴. Transcript expression levels were estimated with the–quantMode GeneCounts option and GRCm38p5.vM15 annotations. FastQC (v0.11.7) (http://www.bioinformatics.babraham.ac.uk/projects/fastqc) was used for QC metrics, and MultiQC (v1.7)⁷⁵ for reporting. Data analysis was then performed with R/Biocondutcor (v4.0.5) (https://www.R-project.org)⁷⁶. Differentially expressed genes were identified using DESeq2 (v1.24.0)⁷⁷, which applies negative binomial generalised linear model fitting and Wald statistics on the count data. The results were filtered for Benjamini–Hochberg adjusted P ≤ 0.05 and fold-change ≥1.5, as indicated in the figure legends.

GO-term analysis

Biological pathway enrichments were determined using Panther (v15.0)⁷⁸. Statistical significance was assessed using Binomial test with Bonferroni correction for multiple testing (P ≤ 0.05). Only the top12 terms ordered by Fold enrichment were considered and ranked by P-value.

ATAC-seq data analysis

Reads were trimmed with Cutadapt (v1.18)⁷⁹ (Cutadapt -a CTGTCTCTTATA -A CTGTCTCTTATA -q 20 -m 5), mapped to the GRCm38/mm10 mouse reference genome with Bowtie2 (v2.3.4.1)⁸⁰. Mitochondrial reads and reads overlapping blacklisted regions (i.e. regions that are considered artefacts of ATAC-seq and other chromatin assays in mm10, https://github.com/MayurDivate/GUAVA/blob/master/lib/blacklists/JDB_blacklist.mm10.bed and http://mitra.stanford.edu/kundaje/akundaje/release/blacklists/mm10-mouse) were removed using bedtools (v2.27.1)⁸¹, intersect and only uniquely mapped reads were kept using Samtools (v1.9)⁸². Bam files were sorted, deduplicated with Picard (v2.9.1) (http://broadinstitute.github.io/picard/) and indexed. Bigwig files were generated with bamCoverage (v3.2.0) from deepTools⁸³ and the RPGC option. MACS (v2.1.1.20160309) was used to call peaks⁸⁴.

ChIP-seq data analysis

Reads were mapped to the GRCm38/mm10 reference genome with Bowtie2 (v2.3.4.1)⁸⁰ and only uniquely mapped reads were kept using Samtools (v1.9)⁸². Bam files were sorted, deduplicated and indexed. Bigwig files were generated with bamCoverage (v3.2.0) from deepTools⁸³ and the RPGC option. The effective genome size was taken from the deepTools documentation. MACS (v2.1.1.20160309) was used to call peaks⁸⁴. Heatmaps and peak profile were generated from the bigwig files with normalised counts to RPKM using deepTools (v3.2.0)⁸². The relative distance between peaks was calculated using the closest function in bedtools (v2.27.1)⁸¹.

De novo motif enrichment analysis

Homer (v4.1)⁸⁵ de novo motif enrichment analysis was performed on bed files generated by MACS (v2.1.1.20160309)⁸⁴ peak caller on the GRCm38/mm10 reference genome with default settings.

Venn diagrams

Proportional Venn diagrams to show unique and intersecting elements of gene lists were generated using the web tool available at https://www.biovenn.nl.

Statistics and reproducibility

The sample size was estimated based on previous preliminary experiments. No statistical method was used to pre-determine sample size. Experiments and quantifications were not done in a blinded fashion. Unless otherwise noted, each immunofluorescence staining was repeated at least three times independently with similar results. Each western blot and EMSA were performed at least twice independently with similar results. Data are presented in box plots as median ± interquartile range as indicated in the figure legends, when n ≥ 10. Data are presented in histograms when n < 10, and individual data points are plotted. Information about P-values and statistical tests used in this article can be found in the figure legends. Statistics were derived using GraphPad Prism (v8.4.3) (GraphPad Software, San Diego, CA, USA). The raw data underlying all reported averages in graphs and charts, and the uncropped scans of western blots and EMSAs are provided in the Source data file.

Reporting summary

Further information on experimental design is available in the Nature Research Reporting Summary linked to this paper.

Data availability

All data generated during this study are included in this article and its Supplementary Information files. scRNA-seq, bulk RNA-seq, ATAC-seq and ChIP-seq data that support the findings of this study have been deposited in the ArrayExpress database at EMBL-EBI (www.ebi.ac.uk/arrayexpress) under the following accession numbers: scRNA-seq of embryonic tailbuds, E-MTAB-9785; scRNA-seq of in vitro cells, E-MTAB-9774; bulk RNA-seq, E-MTAB-9773; ATAC-seq, E-MTAB-9776; and ChIP-seq, E-MTAB-9775. Source data are provided with this paper.

References

Aires, R., Dias, A. & Mallo, M. Deconstructing the molecular mechanisms shaping the vertebrate body plan. Curr. Opin. Cell Biol. 55, 81–86 (2018).
Article CAS PubMed Google Scholar
Stern, C. D. et al. Head-tail patterning of the vertebrate embryo: one, two or many unresolved problems? Int. J. Dev. Biol. 50, 3–15 (2006).
Article CAS PubMed Google Scholar
Wilson, V., Olivera-Martinez, I. & Storey, K. G. Stem cells, signals and vertebrate body axis extension. Development 136, 1591–1604 (2009).
Article CAS PubMed Google Scholar
Cambray, N. & Wilson, V. Two distinct sources for a population of maturing axial progenitors. Development 134, 2829–2840 (2007).
Article CAS PubMed Google Scholar
Henrique, D., Abranches, E., Verrier, L. & Storey, K. G. Neuromesodermal progenitors and the making of the spinal cord. Development 142, 2864–2875 (2015).
Article CAS PubMed Google Scholar
Steventon, B. & Martinez Arias, A. Evo-engineering and the cellular and molecular origins of the vertebrate spinal cord. Dev. Biol. 432, 3–13 (2017).
Article CAS PubMed Google Scholar
Cambray, N. & Wilson, V. Axial progenitors with extensive potency are localised to the mouse chordoneural hinge. Development 129, 4855–4866 (2002).
Article CAS PubMed Google Scholar
Tzouanacou, E., Wegener, A., Wymeersch, F. J., Wilson, V. & Nicolas, J. F. Redefining the progression of lineage segregations during mammalian embryogenesis by clonal analysis. Dev. Cell 17, 365–376 (2009).
Article CAS PubMed Google Scholar
Amin, S. et al. Cdx and T brachyury co-activate growth signaling in the embryonic axial progenitor niche. Cell Rep. 17, 3165–3177 (2016).
Article CAS PubMed Google Scholar
Chalamalasetty, R. B. et al. Mesogenin 1 is a master regulator of paraxial presomitic mesoderm differentiation. Development 141, 4285–4297 (2014).
Article CAS PubMed PubMed Central Google Scholar
Denans, N., Iimura, T. & Pourquie, O. Hox genes control vertebrate body elongation by collinear Wnt repression. Elife https://doi.org/10.7554/eLife.04379 (2015).
Garriock, R. J. et al. Lineage tracing of neuromesodermal progenitors reveals novel Wnt-dependent roles in trunk progenitor cell maintenance and differentiation. Development 142, 1628–1638 (2015).
Article CAS PubMed PubMed Central Google Scholar
Koch, F. et al. Antagonistic activities of Sox2 and brachyury control the fate choice of neuro-mesodermal progenitors. Dev. Cell 42, 514–526 e517 (2017).
Article CAS PubMed Google Scholar
Nowotschin, S., Ferrer-Vaquer, A., Concepcion, D., Papaioannou, V. E. & Hadjantonakis, A. K. Interaction of Wnt3a, Msgn1 and Tbx6 in neural versus paraxial mesoderm lineage commitment and paraxial mesoderm differentiation in the mouse embryo. Dev. Biol. 367, 1–14 (2012).
Article CAS PubMed PubMed Central Google Scholar
Wittler, L. et al. Expression of Msgn1 in the presomitic mesoderm is controlled by synergism of WNT signalling and Tbx6. EMBO Rep. 8, 784–789 (2007).
Article CAS PubMed PubMed Central Google Scholar
Young, T. & Deschamps, J. Hox, Cdx, and anteroposterior patterning in the mouse embryo. Curr. Top. Dev. Biol. 88, 235–255 (2009).
Article CAS PubMed Google Scholar
Diaz-Cuadros, M. et al. In vitro characterization of the human segmentation clock. Nature 580, 113–118 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Gouti, M. et al. A gene regulatory network balances neural and mesoderm specification during vertebrate trunk development. Dev. Cell 41, 243–261.e247 (2017).
Article CAS PubMed PubMed Central Google Scholar
Iimura, T. & Pourquie, O. Collinear activation of Hoxb genes during gastrulation is linked to mesoderm cell ingression. Nature 442, 568–571 (2006).
Article ADS CAS PubMed Google Scholar
Deschamps, J. & Duboule, D. Embryonic timing, axial stem cells, chromatin dynamics, and the Hox clock. Genes Dev. 31, 1406–1416 (2017).
Article CAS PubMed PubMed Central Google Scholar
Young, T. et al. Cdx and Hox genes differentially regulate posterior axial growth in mammalian embryos. Dev. Cell 17, 516–526 (2009).
Article CAS PubMed Google Scholar
Selleri, L., Zappavigna, V. & Ferretti, E. ‘Building a perfect body’: control of vertebrate organogenesis by PBX-dependent regulatory networks. Genes Dev. 33, 258–275 (2019).
Article CAS PubMed PubMed Central Google Scholar
Rezsohazy, R., Saurin, A. J., Maurel-Zaffran, C. & Graba, Y. Cellular and molecular insights into Hox protein action. Development 142, 1212–1227 (2015).
Article CAS PubMed Google Scholar
Capellini, T. D. et al. Pbx1/Pbx2 requirement for distal limb patterning is mediated by the hierarchical control of Hox gene spatial distribution and Shh expression. Development 133, 2263–2273 (2006).
Article CAS PubMed Google Scholar
Capellini, T. D. et al. Pbx1/Pbx2 govern axial skeletal development by controlling Polycomb and Hox in mesoderm and Pax1/Pax9 in sclerotome. Dev. Biol. 321, 500–514 (2008).
Article CAS PubMed PubMed Central Google Scholar
Sonnet, W., Rezsohazy, R. & Donnay, I. Characterization of TALE genes expression during the first lineage segregation in mammalian embryos. Dev. Dyn. 241, 1827–1839 (2012).
Article CAS PubMed Google Scholar
Mariani, L. et al. Identification of the enhancer signature priming lineage choice at gastrulation. In preparation.
Wymeersch, F. J. et al. Transcriptionally dynamic progenitor populations organised around a stable niche drive axial patterning. Development https://doi.org/10.1242/dev.168161 (2019).
Berkes, C. A. et al. Pbx marks genes for activation by MyoD indicating a role for a homeodomain protein in establishing myogenic potential. Mol. Cell 14, 465–477 (2004).
Article CAS PubMed Google Scholar
Grebbin, B. M. & Schulte, D. PBX1 as pioneer factor: a case still open. Front. Cell Dev. Biol. 5, 9 (2017).
Article PubMed PubMed Central Google Scholar
Magnani, L., Ballantyne, E. B., Zhang, X. & Lupien, M. PBX1 genomic pioneer function drives ERalpha signaling underlying progression in breast cancer. PLoS Genet. 7, e1002368 (2011).
Article CAS PubMed PubMed Central Google Scholar
Penalosa-Ruiz, G., Bright, A. R., Mulder, K. W. & Veenstra, G. J. C. The interplay of chromatin and transcription factors during cell fate transitions in development and reprogramming. Biochim. Biophys. Acta Gene Regul. Mech. 1862, 194407 (2019).
Article CAS PubMed Google Scholar
Penkov, D. et al. Analysis of the DNA-binding profile and function of TALE homeoproteins reveals their specialization and specific interactions with Hox genes/proteins. Cell Rep. 3, 1321–1333 (2013).
Article CAS PubMed Google Scholar
Mosimann, C., Hausmann, G. & Basler, K. Beta-catenin hits chromatin: regulation of Wnt target gene activation. Nat. Rev. Mol. Cell Biol. 10, 276–286 (2009).
Article CAS PubMed Google Scholar
Galceran, J., Farinas, I., Depew, M. J., Clevers, H. & Grosschedl, R. Wnt3a-/-like phenotype and limb deficiency in Lef1(-/-)Tcf1(-/-) mice. Genes Dev. 13, 709–717 (1999).
Article CAS PubMed PubMed Central Google Scholar
Merabet, S. et al. Insights into Hox protein function from a large scale combinatorial analysis of protein domains. PLoS Genet. 7, e1002302 (2011).
Article CAS PubMed PubMed Central Google Scholar
Mann, R. S., Lelli, K. M. & Joshi, R. Hox specificity unique roles for cofactors and collaborators. Curr. Top. Dev. Biol. 88, 63–101 (2009).
Article CAS PubMed PubMed Central Google Scholar
Tosic, J. et al. Eomes and Brachyury control pluripotency exit and germ-layer segregation by changing the chromatin state. Nat. Cell Biol. 21, 1518–1531 (2019).
Article CAS PubMed Google Scholar
Mallo, M. & Alonso, C. R. The regulation of Hox gene expression during animal development. Development 140, 3951–3963 (2013).
Article CAS PubMed Google Scholar
Zeitlinger, J. Seven myths of how transcription factors read the cis-regulatory code. Curr. Opin. Syst. Biol. 23, 22–31 (2020).
Article PubMed PubMed Central Google Scholar
Choe, S. K., Lu, P., Nakamura, M., Lee, J. & Sagerstrom, C. G. Meis cofactors control HDAC and CBP accessibility at Hox-regulated promoters during zebrafish embryogenesis. Dev. Cell 17, 561–567 (2009).
Article CAS PubMed PubMed Central Google Scholar
Dell’Orso, S. et al. The histone variant macroH2A1.2 is necessary for the activation of muscle enhancers and recruitment of the transcription factor Pbx1. Cell Rep. 14, 1156–1168 (2016).
Article PubMed PubMed Central CAS Google Scholar
Mirny, L. A. Nucleosome-mediated cooperativity between transcription factors. Proc. Natl Acad. Sci. USA 107, 22534–22539 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Porcelli, D., Fischer, B., Russell, S. & White, R. Chromatin accessibility plays a key role in selective targeting of Hox proteins. Genome Biol. 20, 115 (2019).
Article PubMed PubMed Central CAS Google Scholar
Ladam, F. et al. TALE factors use two distinct functional modes to control an essential zebrafish gene expression program. Elife https://doi.org/10.7554/eLife.36144 (2018).
Steventon, B. et al. Species-specific contribution of volumetric growth and tissue convergence to posterior body elongation in vertebrates. Development 143, 1732–1741 (2016).
CAS PubMed Google Scholar
Takemoto, T. et al. Tbx6-dependent Sox2 regulation determines neural or mesodermal fate in axial stem cells. Nature 470, 394–398 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Aulehla, A. & Pourquie, O. Signaling gradients during paraxial mesoderm development. Cold Spring Harb. Perspect. Biol. 2, a000869 (2010).
Article PubMed PubMed Central CAS Google Scholar
Phan, Q. M. et al. Lef1 expression in fibroblasts maintains developmental potential in adult skin to regenerate wounds. Elife https://doi.org/10.7554/eLife.60066 (2020).
Niemann, C., Owens, D. M., Schettina, P. & Watt, F. M. Dual role of inactivating Lef1 mutations in epidermis: tumor promotion and specification of tumor type. Cancer Res. 67, 2916–2921 (2007).
Article CAS PubMed Google Scholar
Morgan, R. et al. Antagonism of HOX/PBX dimer formation blocks the in vivo proliferation of melanoma. Cancer Res. 67, 5806–5813 (2007).
Article CAS PubMed Google Scholar
Ying, Q. L. & Smith, A. G. Defined conditions for neural commitment and differentiation. Methods Enzymol. 365, 327–341 (2003).
Article CAS PubMed Google Scholar
Koss, M. et al. Congenital asplenia in mice and humans with mutations in a Pbx/Nkx2-5/p15 module. Dev. Cell 22, 913–926 (2012).
Article CAS PubMed PubMed Central Google Scholar
Selleri, L. et al. The TALE homeodomain protein Pbx2 is not essential for development and long-term survival. Mol. Cell Biol. 24, 5324–5331 (2004).
Article CAS PubMed PubMed Central Google Scholar
Hayashi, S., Lewis, P., Pevny, L. & McMahon, A. P. Efficient gene modulation in mouse epiblast using a Sox2Cre transgenic mouse strain. Mech. Dev. 119, S97–S101 (2002).
Article PubMed Google Scholar
Guo, G. et al. Klf4 reverts developmentally programmed restriction of ground state pluripotency. Development 136, 1063–1069 (2009).
Article CAS PubMed PubMed Central Google Scholar
Ran, F. A. et al. Double nicking by RNA-guided CRISPR Cas9 for enhanced genome editing specificity. Cell 154, 1380–1389 (2013).
Article CAS PubMed PubMed Central Google Scholar
Kleinstiver, B. P. et al. High-fidelity CRISPR-Cas9 nucleases with no detectable genome-wide off-target effects. Nature 529, 490–495 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Loh, K. M. et al. Mapping the pairwise choices leading from pluripotency to human bone, heart, and other mesoderm cell types. Cell 166, 451–467 (2016).
Article CAS PubMed PubMed Central Google Scholar
Yoon, J. K. & Wold, B. The bHLH regulator pMesogenin1 is required for maturation and segmentation of paraxial mesoderm. Genes Dev. 14, 3204–3214 (2000).
Article CAS PubMed PubMed Central Google Scholar
Parr, B. A., Shea, M. J., Vassileva, G. & McMahon, A. P. Mouse Wnt genes exhibit discrete domains of expression in the early embryonic CNS and limb buds. Development 119, 247–261 (1993).
Article CAS PubMed Google Scholar
Ferretti, E. et al. Segmental expression of Hoxb2 in r4 requires two separate sites that integrate cooperative interactions between Prep1, Pbx and Hox proteins. Development 127, 155–166 (2000).
Article CAS PubMed Google Scholar
Jaitin, D. A. et al. Massively parallel single-cell RNA-seq for marker-free decomposition of tissues into cell types. Science 343, 776–779 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y. & Greenleaf, W. J. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods 10, 1213–1218 (2013).
Article CAS PubMed PubMed Central Google Scholar
Corces, M. R. et al. An improved ATAC-seq protocol reduces background and enables interrogation of frozen tissues. Nat. Methods 14, 959–962 (2017).
Article CAS PubMed PubMed Central Google Scholar
Kolasinska-Zwierz, P. et al. Differential chromatin marking of introns and expressed exons by H3K36me3. Nat. Genet. 41, 376–381 (2009).
Article CAS PubMed PubMed Central Google Scholar
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 2011, 10–12 (2011).
Article Google Scholar
Kim, D., Langmead, B. & Salzberg, S. L. HISAT: a fast spliced aligner with low memory requirements. Nat. Methods 12, 357–360 (2015).
Article CAS PubMed PubMed Central Google Scholar
Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930 (2014).
Article CAS PubMed Google Scholar
Smith, T., Heger, A. & Sudbery, I. UMI-tools: modeling sequencing errors in Unique Molecular Identifiers to improve quantification accuracy. Genome Res. 27, 491–499 (2017).
Article CAS PubMed PubMed Central Google Scholar
Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018).
Article CAS PubMed PubMed Central Google Scholar
Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902.e1821 (2019).
Article CAS PubMed PubMed Central Google Scholar
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Article CAS PubMed PubMed Central Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article CAS PubMed Google Scholar
Ewels, P., Magnusson, M., Lundin, S. & Kaller, M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics 32, 3047–3048 (2016).
Article CAS PubMed PubMed Central Google Scholar
Gentleman, R. C. et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 5, R80 (2004).
Article PubMed PubMed Central Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Article PubMed PubMed Central CAS Google Scholar
Thomas, P. D. et al. PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 13, 2129–2141 (2003).
Article CAS PubMed PubMed Central Google Scholar
Martin, D. P., Lemey, P. & Posada, D. Analysing recombination in nucleotide sequences. Mol. Ecol. Resour. 11, 943–955 (2011).
Article PubMed Google Scholar
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Article CAS PubMed PubMed Central Google Scholar
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
Article CAS PubMed PubMed Central Google Scholar
Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Article PubMed PubMed Central CAS Google Scholar
Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 44, W160–W165 (2016).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).
Article PubMed PubMed Central CAS Google Scholar
Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589 (2010).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank the staff of the DanStem Core Facilities: H.M. Neil, M. Michaut, G. De la Cruz, P. van Dieken, J.M. Bulkescher, A. Shrestha and A. Meligkova. We are grateful to M.M. Rothová for sharing the wild-type 129 EpiSC line, L. Selleri for providing the Pbx mutant alleles and the Core Facility for Transgenic Mice at the University of Copenhagen, especially J. Martin Gonzalez for derivation of mouse and EpiSC lines. We thank J.M. Brickman, W. Hamilton and J. Zylicz for critical reading of the manuscript and extensive discussion. This study is supported by the Novo Nordisk Foundation grant number NNF17CC0027852, and the Danish Council for Independent Research project grant 4183-00516A. Work supported by Marie Curie Reintegration Fellowship (H2020-MSCA-IF-2014_RI) to E.F.; and Lundbeckfonden PhD Fellowship (grant number R180-2014-2980) to X.G.

Author information

These authors contributed equally: Luca Mariani, Xiaogang Guo.

Authors and Affiliations

Novo Nordisk Foundation Center for Stem Cell Biology (DanStem), University of Copenhagen, Copenhagen, Denmark
Luca Mariani, Xiaogang Guo, Niels Alvaro Menezes, Anna Maria Drozd, Selgin Deniz Çakal & Elisabetta Ferretti
State Key Laboratory of Crop Stress Biology for Arid Areas and College of Plant Protection, Northwest A&F University, Yangling, Shaanxi, China
Qinhu Wang

Authors

Luca Mariani
View author publications
You can also search for this author in PubMed Google Scholar
Xiaogang Guo
View author publications
You can also search for this author in PubMed Google Scholar
Niels Alvaro Menezes
View author publications
You can also search for this author in PubMed Google Scholar
Anna Maria Drozd
View author publications
You can also search for this author in PubMed Google Scholar
Selgin Deniz Çakal
View author publications
You can also search for this author in PubMed Google Scholar
Qinhu Wang
View author publications
You can also search for this author in PubMed Google Scholar
Elisabetta Ferretti
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

L.M., X.G., N.A.M., A.M.D. and E.F. conceived, performed and analysed the experiments. X.G., N.A.M., L.M. and Q.W. performed the bioinformatics data analyses. X.G. and L.M. designed the CRISPR/Cas9 strategies. A.M.D. performed the lentiviral experiments. S.D.Ç. generated the Pbx1^fl/fl;Pbx2^−/−;Cre^pos EpiSC cell lines. L.M., X.G., N.A.M. and E.F. prepared the figures, wrote and edited the manuscript.

Corresponding author

Correspondence to Elisabetta Ferretti.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks Hugo J Parker, Valerie Wilson and other, anonymous, reviewers for their contributions to the peer review of this work. Peer review reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Supplementary Data 5

Supplementary Data 6

Supplementary Movie 1

Supplementary Movie 2

Description of additional supplementary files

Reporting summary

Source data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Mariani, L., Guo, X., Menezes, N.A. et al. A TALE/HOX code unlocks WNT signalling response towards paraxial mesoderm. Nat Commun 12, 5136 (2021). https://doi.org/10.1038/s41467-021-25370-4

Download citation

Received: 25 February 2021
Accepted: 04 August 2021
Published: 26 August 2021
DOI: https://doi.org/10.1038/s41467-021-25370-4

This article is cited by

A spatio-temporally constrained gene regulatory network directed by PBX1/2 acquires limb patterning specificity via HAND2
- Marta Losa
- Iros Barozzi
- Licia Selleri
Nature Communications (2023)
Shaping faces: genetic and epigenetic control of craniofacial morphogenesis
- Licia Selleri
- Filippo M. Rijli
Nature Reviews Genetics (2023)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Subjects

Abstract

Similar content being viewed by others

Introduction

Results

PBX proteins control the formation of the nascent paraxial mesoderm

PBX proteins promote the transition from NMPs towards PSM

PBX proteins are dynamically recruited to PSM genes

PBX proteins are interpreters of LEF1-mediated WNT signalling

Distinct PBX complexes are sequentially recruited to the Msgn1 regulatory region

TALE complexes direct chromatin remodelling at the Msgn1 regulatory region, synergising with T-BRA

Discussion

Methods

Cell Lines

Mice

Generation of Pbx-DKO lines using the CRISPR/Cas9 system

Generation of Pbx-DKO lines by CRE-mediated recombination

Generation of pMsgn1-mut lines using the CRISPR/Cas9 system

EpiSC differentiation to PSM and derivatives

Lentivirus-mediated exogenous expression of Msgn1

Immunofluorescence staining of adherent cells

Imaging of embryo morphology

Immunofluorescence staining of embryos

In situ hybridisation

Time-lapse videos of PSM differentiation

Western blot

Electrophoretic Mobility Shift Assay

Oligonucleotides preparation

Oligonucleotides 5′ end labelling

Nuclear extract preparation

Cell-free protein synthesis

scRNA-seq procedure

RNA-seq procedure

ATAC-seq procedure

ChIP-seq procedure

Real-time quantitative PCR

scRNA-seq data analysis

RNA-seq data analysis

GO-term analysis

ATAC-seq data analysis

ChIP-seq data analysis

De novo motif enrichment analysis

Venn diagrams

Statistics and reproducibility

Reporting summary

Data availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Additional information

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

This article is cited by

Comments

Search

Quick links