Step-wise organization of genomic nuclear speckle-associated domains during mammalian embryonic development

Graphical abstract

Dear Editor,

Nuclear speckles, hubs for gene expression and storage sites for RNA processing factors in typical mammalian nuclei (Chen and Belmont, 2019), concentrate numerous functional components (Gordon et al., 2021). Recent evidence suggests that genomic nuclear speckle-associated domains (SPADs) are transcriptionally active, whereas those nuclear lamina-associated domain (LADs) are inactive (Payne et al., 2021; Quinodoz et al., 2018). Disruptions in chromatin organization lead to aberrant gene expression and altered cell fate decisions (Cuartero et al., 2023; Wang et al., 2022). Notably, LAD formation during mouse embryogenesis has been shown to occur de novo through interactions with the nuclear lamina (Borsos et al., 2019). However, the dynamics of chromatin organization at nuclear speckles and their role in regulating gene activation during embryogenesis remain unexplored.

To investigate genome-nuclear speckle interactions in mouse embryos, we developed a low-input CUT&Tag method using an SC35 antibody targeting the nuclear speckle core protein Srrm2. This approach generated high-quality SPADs profiles across developmental stages (Figs. S1 and S2; Table S1). In PN3 zygotes, SPADs were barely detectable. Weak and broad SPADs signals emerged in PN5 zygotes, and more defined smaller SPADs became increasingly prevalent from late 2-cell stage onward (Fig. 1A). SPADs strength significantly increased from late 2-cell to 8-cell stages (Fig. 1B). SPADs signal intensity and strength in embryonic stem cells (ESCs) were comparable to those at blastocyst stage but markedly higher in the prefrontal cortex (PFC) (Fig. 1A and 1B). SPADs genomic coverage gradually expanded during development, consistently exhibiting a mean size exceeding 1 Mb (Fig. S3A). Before 8-cell stage, SPADs underwent significant reorganization, marked by transitions between SPADs and adjacent nonSPADs regions. These transitions could be categorized into seven distinct types (Figs. 1A, S3B, S3C; Table S1). Overall, SPADs undergo profound reorganization following fertilization.

In mice, major zygotic genome activation (ZGA) occurs at late 2-cell stage. Notably, SPADs established at PN5 and early 2-cell stages formed larger domains (Fig. 1A). We classified SPADs formed before ZGA as primary SPADs (pSPADs) and those formed after ZGA as secondary SPADs (sSPADs). The proportion of pSPADs remained above 60% at all stages with minimal decreases throughout development (Fig. 1C). Compared to sSPADs, pSPADs exhibited broader domains, higher CpG density, lower AT content, greater enrichment for genomic elements, higher stable index and higher conservation score (Figs. 1D–F and S3D–F). These findings indicate that SPADs formed before and after ZGA differ significantly in their characteristics.

We further explored parental differences in SPADs. The paternal SPADs were distinct from maternal SPADs at PN5 and 2-cell stages (Figs. 1G, 1H, S4A and S4B). After 4-cell stage, paternal and maternal SPADs became highly correlated (Figs. 1G, 1H, S4A and S4B). Consistently, paternal SPADs showed stronger signals at early stages (Fig. 1I). Nonetheless, distinctions persisted in certain imprinted regions, such as the non-canonical imprinted gene Jade1 (Inoue et al., 2017) (Fig. S4C). At PN5 and 2-cell stages, most SPADs were allelic-specific. From the 4-cell stage onward, common SPADs accounted for over 60% of the total (Fig. S5A). Common SPADs were longer, had higher CpG density, and exhibited greater stability than allelic-specific SPADs (Figs. 1J, S5B and S5C). Notably, the majority of common SPADs were classified as pSPADs (Fig. S5D). SPADs organization also exhibited distinct patterns on the X chromosome. SPADs signals on the maternal X chromosome gradually increased, while those on the paternal X peaked at 4-cell stage before declining (Figs. 1K and S5E). This trend aligns with the paternal X inactivation observed during mouse embryogenesis (Lee and Bartolomei, 2013). Overall, parental SPADs differ in early stages and converge in late stage during embryogenesis.

We then investigated the relationship between SPADs establishment and gene expression. The genes in SPADs exhibit higher expression levels compared to those in nonSPADs (Fig. 1L). Notably, minor and major ZGA genes, housekeeping genes, and numerous transposable elements were significantly enriched in SPADs, particularly in pSPADs (Fig. S6A and S6B). These findings suggest that pSPADs create a favorable environment for early embryonic gene expression.

Genes in newly formed SPADs of each stage exhibited sequential expression patterns and were associated with stage-specific signaling pathways (Figs. 1M and S6C). Most stage-specific expression genes were localized within stage-specific SPADs (Fig. 1N). These data suggest that SPADs are pre-conFig.d and primed for gene activation. Critical genes involved in triggering ZGA and lineage segregation, such as Klf17, Nfya, Yy1, and Obox genes including Obox1, Obox2, Obox3, and Obox5 (Kravchenko and Tachibana, 2025), showed significant enrichment of SPADs binding sites, particularly within pSPADs (Fig. S6D and S6E). These results highlight that SPADs establishment is intricately linked to stage-specific gene activation, fulfilling developmental requirements.

We next explored the correlation between SPADs and epigenetic features in mouse embryos. In general, SPADs were negatively correlated with repressive epigenetic features such as DNA methylation, LADs signals, H2AK119ub1 (H2Aub), and H3K27me3, but they were positively correlated with active epigenetic features such as DNase I hypersensitive sites (DHSs), higher replication timing scores, RNA polymerase II (pol II), H3K36me3, H3K4me3, and H3K27ac (Figs. S7, S8, and S9). These correlations were more pronounced in pSPADs than sSPADs at blastocyst stage. Although, the SPADs are largely inversely related to repressive mark H3K9me3 (Fig. 1O), we observed some H3K9me3 signals within SPADs, and these signals were sequentially removed during development (Fig. 1P), accompanied by the associated genes activation (Fig. 1Q). These associated genes played critical roles in stage-specific biological processes (Fig. 1R), such as Efna4 and Adam15 (Fig. 1S). We further analyzed 293 genes (Table S2) located in SPADs newly formed at 8-cell stage (green box in Fig. 1M). These genes showed a decrease in repressive mark enrichment and an increase in active mark enrichment from 8-cell to morula stage (Fig. 1T), consistent with their higher expression at morula. These findings suggest that SPADs interact with other epigenetic features to finely regulate gene expression.

By integrating our Hi-C data (Ke et al., 2017) with SPADs profiles, we further investigated the relationship between SPADs and 3D chromatin structure. As expected, over 95.5% of SPADs overlapped with A compartments at each stage (Fig. 2A). When TADs are obscure, the pSPADs are already established at the PN5 stage. We found approximately 35% of pSPADs boundaries overlapped with blastocyst TAD boundaries (Fig. 2B). As TAD boundaries became clearer during development (Fig. S10A), insulation scores around pSPADs boundaries were lower than other regions and became more pronounced (Fig. 2C). The smallest distance between pSPADs boundaries and TAD boundaries were observed at the late 2-cell stage (Fig. 2D). Boundaries shared by SPADs and TADs exhibited significantly lower insulation scores compared to other boundaries (Figs. 2E and S10B). Collectively, these data suggest that pre-defined pSPADs might contribute to TAD structure establishment during embryogenesis. We also found that the proportion of frequently interacting regions (FIREs) overlapping with SPADs increased during development, with significant enrichment of FIREs within pSPADs after 4-cell stage (Figs. 2F, S10C and S10D), which indicates that SPADs play a role in FIREs organization. Together, the SPADs are closely associated with 3D chromatin structure during development.

Comparing the A/B compartments of sperm with zygotic SPADs, we observed that 97.53% of zygotic SPADs overlapped with sperm A compartments (Fig. S11A). SPADs also exhibited higher compartment scores, with constant SPADs consistently persisting as A compartments (Fig. S11B and S11C). Regions in sperm A compartments that formed SPADs in the zygote had higher compartment scores, gene density, and CpG density than regions without SPADs formation (Fig. 2G–I). Further subdivision of sperm A compartments into four sub-compartments (see Methods) revealed that the A1.1 sub-compartment had the largest overlap with PN5 SPADs (Fig. 2J and 2K). These findings suggest that sperm A compartments are closely associated with the formation of PN5 SPADs shortly after fertilization. In embryos with a certain extent knockdown of the chromatin structure organization factor Nipbl (Gao et al., 2018) (Fig. S11D and S11E), sSPADs signals were significantly reduced, while pSPADs were not affected (Fig. 2L). This highlights the role of chromatin structural proteins in stage-specific SPADs formation.

To investigate the impact of transcription in SPADs establishment, we treated embryos with α-amanitin efficiently (Lallena, 1997) (Fig. S12A–C). SPADs signals and strength appeared weaker in α-amanitin treated embryos (Fig. 2M and 2N). Notably, most SPADs remaining in α-amanitin-treated embryos were classified as pSPADs (Fig. 2O). Additionally, following α-amanitin inhibition, virtually no maternal SPADs are formed, whereas a few paternal SPADs still emerge (Fig. S12D and S12E). Together, the data indicate that the establishment of sSPADs and maternal SPADs predominantly relies on ZGA.

We next explored the association between SPADs formation and the putative speckle targeting motif (STM) factors, which is essential for chromatin-nuclear speckle interactions (Yu et al., 2023). Focusing on STM genes with known DNA-binding motifs (Table S2), we found that most of them displayed stage-specific expression patterns (Fig. S12F). STM proteins coded from maternally deposited transcripts consistently exhibited binding motif enrichment in pSPADs (Fig. S12G), such as the speckled protein Sp110 (Fraschilla and Jeffrey, 2020) (Fig. S12H and S12I). These findings suggest that maternally deposited STM transcripts may participate in pSPADs formation before ZGA. The STM genes expressed after ZGA (Table S2) showed binding motif enrichment in stage newly formed SPADs, including Gata6 (Schrode et al., 2014) (Fig. 2P). In morula embryos with efficient Gata6 knockdown (Fig. S12J and S12K), signals of newly formed SPADs were significantly reduced (Fig. 2Q–S), while pre-existing SPADs were minimally affected. Collectively, these results indicate that a set of STM genes may participate in stage-specific SPADs formation.

In this study, we generated high-resolution SPADs maps of mouse embryos for the first time. The SPADs are re-established post-fertilization with distinct allelic features until 4-cell stage. The comprehensive analysis suggests dynamic SPADs formation play critical roles in the regulation of gene expression during embryonic development. SPADs were shown to pre-configure gene expression during embryonic development. Interestingly, the heterochromatin marker H3K9me3 within SPADs plays a precise role in tuning gene expression. We also confirmed the closely interplay between chromatin high-order structure and SPADs. Finally, we demonstrated that RNA transcription significantly influences SPADs formation, particularly stage-specific STM genes like the primitive endoderm transcription factor Gata6 involve in stage-specific SPAD formation. This study provides novel insights into the complex epigenetic network regulating embryonic gene expression and highlights the transition of chromatin from a disordered state to an organized structure post-fertilization.

References

[1] Borsos M, Perricone SM, Schauer T et al Genome-lamina interactions are established de novo in the early mouse embryo. Nature 2019;569:729–733.

[2] Chen Y, Belmont AS. Genome organization around nuclear speckles. Curr Opin Genet Dev 2019;55:91–99.

[3] Cuartero S, Stik G, Stadhouders R. Three-dimensional genome organization in immune cell fate and function. Nat Rev Immunol 2023;23:206–221.

[4] Fraschilla I, Jeffrey KL. The Speckled Protein (SP) family: immunity’s chromatin readers. Trends Immunol 2020;41:572–585.

[5] Gao D, Zhu B, Cao X et al Roles of NIPBL in maintenance of genome stability. Seminars in Cell & Developmental Biology 2018;90:181–186.

[6] Gordon JM, Phizicky DV, Neugebauer KM. Nuclear mechanisms of gene expression control: pre-mRNA splicing as a life or death decision. Curr Opin Genet Dev 2021;67:67–76.

[7] Inoue A, Jiang L, Lu F et al Maternal H3K27me3 controls DNA methylation-independent imprinting. Nature 2017;547:419–424.

[8] Ke Y, Xu Y, Chen X et al 3D chromatin structures of mature gametes and structural reprogramming during mammalian embryogenesis. Cell 2017;170:367–381.e20.

[9] Kravchenko P, Tachibana K. Rise and SINE: roles of transcription factors and retrotransposons in zygotic genome activation. Nat Rev Mol Cell Biol 2025;26:68–79.

[10] Lallena MJ, Correas I. Transcription-dependent redistribution of nuclear protein 4.1 to SC35-enriched nuclear domains. J Cell Sci 1997;110:239–247.

[11] Lee JT, Bartolomei MS. X-inactivation, imprinting, and long noncoding RNAs in health and disease. Cell 2013;152:1308–1323.

[12] Payne AC, Chiang ZD, Reginato PL et al In situ genome sequencing resolves DNA sequence and structure in intact biological samples. Science 2021;371:eaay3446.

[13] Quinodoz SA, Ollikainen N, Tabak B et al Higher-order inter-chromosomal hubs shape 3D genome organization in the nucleus. Cell 2018;174:744–757.e724.

[14] Schrode N, Saiz N, Di Talia S et al GATA6 levels modulate primitive endoderm cell fate choice and timing in the mouse blastocyst. Dev Cell 2014;29:454–467.

[15] Wang Y, Elsherbiny A, Kessler L et al Lamin A/C-dependent chromatin architecture safeguards naive pluripotency to prevent aberrant cardiovascular cell fate and function. Nat Commun 2022;13:6663.

[16] Yu R, Roseman S, Siegenfeld AP et al CTCF/cohesin organize the ground state of chromatin-nuclear speckle association. bioRxiv. 2023.

RIGHTS & PERMISSIONS

The Author(s) 2025. Published by Oxford University Press on behalf of Higher Education Press.