Chromatin architecture reorganization in murine somatic cell nuclear transfer embryos

Abstract

The oocyte cytoplasm can reprogram the somatic cell nucleus into a totipotent state, but with low efficiency. The spatiotemporal chromatin organization of somatic cell nuclear transfer (SCNT) embryos remains elusive. Here, we examine higher order chromatin structures of mouse SCNT embryos using a low-input Hi-C method. We find that donor cell chromatin transforms to the metaphase state rapidly after SCNT along with the dissolution of typical 3D chromatin structure. Intriguingly, the genome undergoes a mitotic metaphase-like to meiosis metaphase II-like transition following activation. Subsequently, weak chromatin compartments and topologically associating domains (TADs) emerge following metaphase exit. TADs are further removed until the 2-cell stage before being progressively reestablished. Obvious defects including stronger TAD boundaries, aberrant super-enhancer and promoter interactions are found in SCNT embryos. These defects are partially caused by inherited H3K9me3, and can be rescued by Kdm4d overexpression. These observations provide insight into chromatin architecture reorganization during SCNT embryo development.

Fig. 1. Landscape of chromatin organization in early SCNT embryo development.

a Illustration of SCNT procedure and sample collections. According to the classical procedure of mouse cloning, donor nuclei from cumulus cells were injected into enucleated oocytes. Reconstructed embryos were chemically activated 1 h post-injection by culture in the Ca²⁺-free CZB medium including strontium chloride. b Hi-C produces a genome-wide contact matrix. The normalized Hi-C interaction frequencies (100-kb bin, chromosome 19) in each sample. Zoomed-in views (40-kb bin) are also shown. Each pixel represents all interactions between a 100-kb (40-kb) locus and another 100-kb (40-kb) locus; intensity corresponds to the ICE normalized value (0–10).

Fig. 2. Mitotic metaphase to meiosis metaphase II-like chromatin transition.

a Immunofluorescence staining of 0.5-hpi, 1-hpi, 1-hpa and 6-hpa 1-cell embryos and 8-cell embryos. DNA: DAPI (blue) and microtubules: α-tubulin (green). Scale bar: 10 μm. Data are representative of three independent experiments. b Both the curves and heatmaps showing the intrachromosomal interaction probabilities (P(s)) relative to genomic distance that are normalized by the P(s) ~ s⁻¹ values. The peaks represent the rapid fall-off at ~4 Mb and ~10 Mb in secondary meiotic and mitotic chromatin, respectively. Source data are provided as a Source Data file. c The chromatin contact probabilities (P(s)) relative to genomic distance for CC, 0.5-hpi, and 1-hpi embryos. The P(s) ~ s⁻¹ curve representing the predicted fractal globule state is shown for reference. The P(s) ribbon is bounded by minimum and maximum (P(s)) calculated from all replicates of Hi-C data sets. Source data are provided as a Source Data file.

Fig. 3. Dissolution and reestablishment of 3D chromatin architecture.

a Correlations of intrachromosomal interaction frequency patterns between any two regions along chromosome 19 (300-kb bin). The first principle component (PC1) values are shown under each heatmap as compartments A (magenta) and B (cyan). b The top heatmaps show the mean normalized interaction frequencies for the TADs and their flanking regions with a TAD length of ±0.5. The TADs were defined in the ICM stage and used for all stages. The bottom heatmaps show the difference in the interaction frequencies between consecutive stages. Source data are provided as a Source Data file.

Fig. 4. Differential dynamic of chromatin structure reorganization between normal and SCNT embryos.

a Heatmaps showing the intrachromosomal interaction probabilities relative to genomic distance that are normalized by the P(s) ~ s⁻¹ values. Dashed boxes indicate the developmental stages and genomic distance with differential interaction probabilities between SCNT and fertilization-derived embryos. Source data are provided as a Source Data file. b Ratios of interaction frequencies between compartments A & B to those between compartments A & A or B & B (n = 323). Boxes show 25th, 50th and 75th percentiles and whiskers show 1.5× the inter-quartile range. The p values were calculated by a two-sided Wilcoxon rank-sum test with BH multiple testing correction Source data and exact p-value are provided as a Source Data file. c RTI across all stages (n = 2057). Boxes show 25th, 50th and 75th percentiles and whiskers show 1.5× the inter-quartile range. The two-sided p values were calculated by the Kruskal–Wallis test with Dunn’s multiple comparison test and adjusted by default with the holm method (N.S., not significant). Source data and exact p-value are provided as a Source Data file.

Fig. 5. Defects of promoter and super-enhancer interactions in SCNT embryos.

a The top heatmap shows the differential interaction frequencies at the Zscan4d locus between fertilization-derived and SCNT early-2-cell embryos. The dashed circle indicates the SE–P interactions in fertilization-derived embryos but not in SCNT embryos. The bottom tracks show gene expression in IVF and SCNT embryos, H3K27ac signals in fertilization-derived 2-cell embryos, and Refseq genes, respectively. Source data and exact p-value are provided as a Source Data file. b The normalized interaction frequencies between the super-enhancer and promoter of Zscan4d. The q-values were calculated in Fit-Hi-C by applying Benjamini-Hochberg correction to the p-values, as computed by the binomial distribution model employed.

Fig. 6. H3K9me3 is a potential barrier of TADs reprogramming.

a Heatmaps of chromatin interaction frequencies showing a CC TAD unreprogrammed example (magenta) and a CC TAD reprogrammed (cyan) example during SCNT embryo development. The bottom track shows the H3K9me3 signal in CCs. b RTIs of H3K9me3-maked and H3K9me3-unmarked TADs identified in CCs. RTI values in CC, 6-hpa, 12-hpa and early-2-cell embryos are shown (n = 2184). Boxes show 25th, 50th and 75th percentiles and whiskers show 1.5× the inter-quartile range. The two-sided p values were calculated by the Kruskal–Wallis test with Dunn’s multiple comparison test and adjusted by default with the holm method. Source data and exact p-value are provided as a Source Data file. c Heatmaps showing normalized chromatin interaction frequencies (100-kb bin, chromosome 16) in SCNT, SCNT with Kdm4d mRNA injection, SCNT with TSA treatment and fertilization-derived early 2-cell embryos. Both ICE and Binless normalization approaches are used. d RTIs of H3K9me3-maked TAD in early 2-cell embryos. RTI values in CC, SCNT, SCNT with Kdm4d mRNA injection, SCNT with TSA treatment and fertilization-derived are shown (n = 2184). Boxes show 25th, 50th and 75th percentiles and whiskers show 1.5× the inter-quartile range. The two-sided p values were calculated by the Kruskal–Wallis test with Dunn’s multiple comparison test and adjusted by default with the holm method (N.S., not significant). Source data and exact p-value are provided as a Source Data file. e Virtual 4C test showing the interaction frequencies between Zscan4d promoter and its adjacent region in chromosome 7. Black dots represent observed values, red line represents fitted value by Binless.

Fig. 7. A schematic model showing the 3D chromatin structure dynamic during SCNT embryo development.

a CCs exhibit interphase-state chromatin characterized by mature compartments and TADs, both of which are dissolved quickly after injection into enucleated oocytes. Consequently, 1-hpi and 1-hpa embryos exhibit a mitotic and secondary meiotic metaphase-like chromatin state lacking compartments and TADs, respectively. Compartments and TADs emerge in early-2-cell embryos, exist in later stages, and become mature in the ICM stage. TADs with enriched H3K9me3 signals are resistant to reprogramming. The differential reprogramming of chromatin architecture results in the loss of SE–P loops of  regulating genes critical to development.