Overlapping peri-implantation phenotypes of ZNHIT1 and ZNHIT2 despite distinct functions during early mouse development

Abstract

Mammalian preimplantation development culminates in the formation of a blastocyst that undergoes extensive gene expression regulation to successfully implant into the maternal endometrium. Zinc-finger HIT domain-containing (ZNHIT) 1 and 2 are members of a highly conserved family, yet they have been identified as subunits of distinct complexes. Here, we report that knockout of either Znhit1 or Znhit2 results in embryonic lethality during peri-implantation stages. Znhit1 and Znhit2 mutant embryos have overlapping phenotypes, including reduced proportion of SOX2-positive inner cell mass cells, a lack of Fgf4 expression, and aberrant expression of NANOG and SOX17. Furthermore, we find that the similar phenotypes are caused by distinct mechanisms. Specifically, embryos lacking ZNHIT1 likely fail to incorporate sufficient H2A.Z at the promoter region of Fgf4 and other genes involved in cell projection organization resulting in impaired invasion of trophoblast cells during implantation. In contrast, Znhit2 mutant embryos display a complete lack of nuclear EFTUD2, a key component of U5 spliceosome, indicating a global splicing deficiency. Our findings unveil the indispensable yet distinct roles of ZNHIT1 and ZNHIT2 in early mammalian embryonic development.

We report that ZNHIT1 and ZNHIT2 are indispensable for mouse embryonic development with knockout embryos dying during peri-implantation stages. Embryos lacking ZNHIT1 or ZNHIT2 result in overlapping phenotypes but through distinct mechanisms.

Graphical Abstract

Graphical Abstract.

Introduction

Early embryonic development is a finely tuned process, characterized by a sequence of precisely timed morphogenetic events. In mice, within the first 3.5 days post-fertilization, zygotes undergo several rounds of cleavage cell division, morula compaction, and initial lineage differentiation resulting in a blastocyst-stage embryo, which consists of two morphologically distinct cell lineages: the inner cell mass (ICM) and trophectoderm (TE). Further differentiation within the ICM leads to the formation of the epiblast (EPI), the precursor cells of the future fetus, and the primitive endoderm (PrE), which contributes to the yolk sac [1]. Fibroblast growth factor 4 (FGF4), the only FGF ligand expressed in ICM, has been found to play a central role in regulating EPI/PrE proportion within the ICM [2–8]. A high dose of FGF4 can convert all ICM cells to PrE identity, while conversely, blocking FGF signaling prevents PrE development [2, 9, 10]. At embryonic day (E) 4.0, blastocysts start to hatch from the zona pellucida, preparing for implantation, an essential step for all placental mammals [11]. This complex developmental process is closely intertwined with precise gene regulation, including chromatin remodeling and RNA splicing [12, 13].

ZNHIT proteins constitute an evolutionarily conserved family with six members in mice and human [14, 15]. Knockout of ZNHIT 3/4/5/6 protein in mice results in embryonic lethality, underscoring the essential roles of ZNHIT family members [16–18]. Interestingly, all six ZNHIT proteins exhibit interactions with the AAA+ ATPase RUVBL2 through their respective zf-HIT domains [15]. Despite this shared characteristic, these proteins participate in different complexes, carrying out diverse functions [15]. ZNHIT1, for instance, was identified as a subunit of the SNF2-related helicase SRCAP complex, analogous to the Saccharomyces cerevisiae SWR1 chromatin remodeling complex, responsible for replacing H2A with H2A.Z (a conserved H2A variant) [19]. The role of ZNHIT1 as a critical transcriptional activator has been shown in intestinal stem cell maintenance, embryonic cardiomyocytes metabolism, and meiotic initiation in spermatogenesis [20–22]. On the other hand, ZNHIT2 has been documented as a co-factor mediating U5 small nuclear ribonucleoprotein (snRNP) assembly, playing a vital role in pre-mRNA splicing [15, 23]. Although these studies identify molecular functions of ZNHIT1 and ZNHIT2 in vitro, their role in early embryonic development remains unexplored.

Here, we elucidate the indispensable roles of ZNHIT1 and ZNHIT2 during peri-implantation development, highlighting the overlapping phenotypes of knockout embryos despite distinct functional mechanisms of the two proteins.

Results

Knockout of Znhit1 or Znhit2 results in post-blastocyst lethality

We first investigated the developmental consequences of the absence of ZNHIT1 or ZNHIT2. Embryos lacking either ZNHIT1 or ZNHIT2 could not be recovered at E7.5. However, we find an increased number of empty decidua when E7.5 litters were dissected from heterozygous intercrosses from each knockout allele, suggesting decidualization may be triggered by mutant embryos (Figure 1A). At E3.5, morphologically normal blastocysts were recovered at Mendelian ratios for both lines (Figure 1A and B). In vitro culture in outgrowth media revealed that Znhit1 mutant embryos successfully hatched from the zona pellucida and formed enlarged blastocysts of comparable size and morphology to control embryos, while Znhit2 mutants were consistently confined within the zona and mostly collapsed/dying after 24 h (Figure 1B).

Figure 1.

Morphological and molecular characterization of Znhit1 and Znhit2 mutant embryos. (A) Table showing numbers and percentages of genotypes recovered at E3.5 and E7.5, and the empty decidua rate at E7.5. NA means not applicable. (B) Representative brightfield images showing gross morphology of control, Znhit1 mutant (n = 10) and Znhit2 mutant (n = 8) embryos flushed at E3.5 (upper panel) and after 24-hr culture in outgrowth media (lower panel). Note that Znhit1 mutants were able to hatch out, while Znhit2 mutants were not. (C) Maximum intensity projections (MaxIP) of representative confocal images showing DAPI (blue), SOX2 (green), and CDX2 (magenta) in KSOM-cultured control, Znhit1 mutant, and 2 Znhit2 mutant blastocysts (note the two different phenotypes observed). The arrowheads point to a scattered DAPI signal. Scale bar, 20 μm. (D) Bar chart comparing the total cell number and (E) percent SOX2⁺ cells for each genotype. The number of circles indicates the number of embryos analyzed in each group. One-way ANOVA was conducted for statistical analysis. ns indicates not statistically significant, ^* indicates P ≤ 0.05, ^** indicates P ≤ 0.01, and ^*** indicates P ≤ 0.001. (F) The upper panel shows the RT-PCR of Mrpl17 (internal control) and Znhit1 in wildtype (Wt), Znhit1 heterozygous (Het), and Znhit1 mutant blastocysts, to validate the knockout of Znhit1. The lower panel shows the RT-PCR of Oct4, Znhit2, and Sox2 in four Znhit2 mutant, two Znhit2 Het, and two Wt embryos. BL Mut: expanded Znhit2 mutant blastocysts. CL Mut: collapsed Znhit2 mutant embryos. Note that Sox2 expression is only completely lost in collapsed mutants. (G) MaxIP of representative confocal images showing localization of ZO-1 (magenta) in cultured whole-mount control (n = 8) and Znhit2 mutant blastocysts (n = 3). Scale bar, 20 μm. The inset highlights the abnormal accumulation of ZO-1 within the mutant embryo.

Immunofluorescence (IF) was then performed to assess the cell number and lineage differentiation in E3.5 + 24 h blastocysts (Figure 1C). While Znhit1 mutants exhibited total cell numbers comparable with control embryos (about 80 cells), Znhit2 mutants displayed a nearly 50% decrease in total cell count, accompanied by apparent fragmented DNA/nuclei, indicated by scattered DAPI signal (arrowheads in Figure 1C and D). Interestingly, both Znhit1 and Znhit2 mutants exhibited a significant reduction in the SOX2⁺ cells. While control embryos displayed 12% SOX2⁺ cells, Znhit1 mutants exhibited only 5% and Znhit2 mutants demonstrated a complete absence of SOX2 expression (Figure 1C and E). This observation suggests defective lineage specification in the ICM of both Znhit1 and Znhit2 mutants.

The absence of Znhit1 and Znhit2 transcripts were validated separately using real-time reverse transcriptase-polymerase chain reaction (RT-PCR) (Figure 1F). Furthermore, RT-PCR indicated a loss of Sox2 mRNA coinciding with the collapse of Znhit2 mutant blastocysts (Figure 1F, lower panel). The collapsing phenotype observed in Znhit2 mutant blastocysts suggested impaired cell junction integrity. Therefore, we specifically examined tight junctions in the expanded Znhit2 mutant blastocysts by ZO-1 staining. We observed irregular accumulation and breaks in ZO-1 membrane localization between cells, indicating TE cell–cell junction defects, a possible hallmark of impending blastocyst collapse (Figure 1G, inset). This evidence emphasizes the requirement of ZNHIT2 in upholding the integrity of the polarized transporting epithelial nature of TE cells.

Znhit1 and Znhit2 mutants exhibit aberrant SOX17 and NANOG expression

We next asked if the segregation within the ICM was affected upon Znhit1 or Znhit2 knockout by performing IF with antibodies against NANOG and SOX17. Consistent with previous results, total cell numbers were significantly reduced in Znhit2 mutants, but not in Znhit1 mutants (Figure 2A and C). Znhit1 mutants showed normal ICM/TE proportions comparable with control embryos, while Znhit2 mutants exhibited a significant reduction in ICM cells (Figure 2D). Interestingly, in both mutants, we observed a substantial reduction in SOX17⁺ cells, with Znhit1 mutants showing a 50% decrease and Znhit2 mutants exhibiting a complete absence of these cells (Figure 2A and E). Additionally, NANOG was observed inappropriately expressed in CDX2+ TE cells in both mutant embryos (Figure 2A and F). These observations suggest defects in both EPI and PrE specification as well as programming defects in TE cells occurring in both Znhit1 and Znhit2 mutants.

Figure 2.

Exogenous FGF4 rescues lineage defects in Znhit1 and Znhit2 mutant embryos. (A, B) MaxIP of representative confocal images showing localization of cells expressing SOX17 (green) and NANOG (magenta) in whole-mount control, Znhit1 mutant, Znhit2 mutant blastocysts, cultured in (A) KSOM + heparin (KSOMh) or (B) KSOM + FGF4 + heparin (KSOMh + FGF4). Scale bar, 20 μm. (C–E) Bar charts comparing the total cell number (C), ICM proportions (D), and SOX17⁺ cell proportion (E) for each genotype with or without exogenous FGF4. Number of circles or triangles indicates the number of embryos analyzed in each group. One-way ANOVA was conducted for statistical analysis. ns indicates not statistically significant, ^* indicates P ≤ 0.05, ^** indicates P ≤ 0.01, and ^*** indicates P ≤ 0.001. ICM cells were determined by the location of DAPI. Note that exogenous FGF4 increased the proportion of SOX17⁺ cells in both Znhit1 and Znhit2 mutant embryos without significantly impacting the total cell count. (F) MaxIP of representative confocal images showing localization of cells expressing NANOG (magenta) and CDX2 (yellow) in whole-mount control, Znhit1 mutant, Znhit2 mutant blastocysts. Note that NANOG⁺CDX2⁺ double positive cells appeared in both mutant embryos. (G) Representative MaxIP confocal images showing protruding/hatching cells (arrowheads) in Znhit2 mutant embryos (n = 2 out of 10) when cultured with FGF4-supplement. Notably hatching is never observed in Znhit2 mutants without FGF4 (N > 10). White dashed circles indicate the location of the zona pellucida. (H) Comparison of the relative intensity of CDX2 (CDX2/DAPI) in wildtype, Znhit1 heterozygous, and Znhit1 mutant blastocysts at E3.5 (embryos in Figure 1C). n indicates the number of embryos included in each group, and each point represents data from one nucleus. One-way ANOVA was conducted for statistical analysis. Znhit1 mutants exhibited a significantly reduced CDX2/DAPI ratio compared with wildtype or heterozygous embryos. (I) Representative brightfield images showing gross morphology (upper) and RT-PCR results (lower) of Znhit1/Znhit2 double knockdown blastocysts. Pooled siRNAs against both transcripts were injected into zygotes, and embryos were cultured 96 h in KSOM. RT-PCR was done on two different embryos for each genotype.

There is a well-established role of FGF4 in PrE specification [2–5, 7, 8, 10]. Therefore, to determine if ZNHIT activity is upstream or downstream of FGF signaling, we attempted to rescue the reduced/absent SOX17⁺ cells by culturing mutants (and controls) with exogenous FGF4. We collected litters from heterozygous intercrosses at E2.5 and cultured with FGF4 for 48 h. Although no significant change in total cell number was observed (Figure 2C), the percentage of SOX17⁺ cells in Znhit1 mutants was restored to normal (Figure 2B and E). Although some Znhit2 mutants showed a few SOX17⁺ cells (Figure 2B and E), FGF4 did not fully rescue the SOX17 phenotype in Znhit2 mutants. Interestingly, FGF4 treatment also largely rescued the inappropriate expression of NANOG in mutant TE cells (Figure 2B), keeping NANOG expression restricted to the EPI in mutants cultured with FGF4. Of note, Znhit2 mutants became more expanded in the presence of FGF4 compared with mutants without FGF4 (Figure 2A and B), with 20% initiating hatching by the end of the culture period, something never observed in Znhit2 mutants without exogenous FGF4 (Figure 2G).

Considering the antagonistic relationship between CDX2 and NANOG [24], we analyzed CDX2 intensity in the TE cells of Znhit1 mutants and littermates. Intriguingly, Znhit1 mutants exhibited significantly reduced CDX2 levels in TE cells compared with heterozygous or wildtype littermates (Figure 2H). This reduction may explain the aberrant NANOG expression in Znhit1 mutant blastocysts. Unfortunately, Znhit2 mutants could not be analyzed due to the collapsing phenotype (described above, Figure 1B).

Given the morphological and molecular similarities between Znhit1 and Znhit2 mutants, we aimed to determine if there is functional redundancy between the two genes. We microinjected pooled Znhit1 and Znhit2 small interfering RNAs (siRNAs) into wildtype zygotes, followed by culture to the blastocyst stage. Double knock-down mutants developed to the blastocyst stage with normal developmental timing dynamics and no obvious phenotype (Figure 2I). Since there is not a more severe phenotype in the double knockdown embryos, these results indicate that either ZNHIT1 and ZNHIT2 have distinct non-overlapping roles prior to blastocyst formation or that ZNHIT1/2 function is simply not required until the late/expanded blastocyst stage.

Znhit1 mutants exhibit compromised implantation both in vitro and in vivo

To examine the ability of Znhit1 mutants to implant, we first used outgrowth assays as an in vitro alternative. Wildtype and heterozygotes formed typical outgrowths with a tightly packed ICM-derived OCT4⁺ colony in the center, closely surrounded by a group of CDX2^high cells resembling extra-embryonic ectoderm (ExE) cells. The periphery is comprised of differentiated trophoblast giant cells (TGCs), identifiable by large nuclei, forming a monolayer attached to the bottom surface (Figure 3A). In contrast, Znhit1 mutant outgrowths exhibited a much smaller and irregular central colony, suggesting a loss of pluripotency, a more dispersed CDX2^high population, and fewer TGCs, suggesting impaired TE differentiation (Figure 3A). To evaluate TGC specification, we stained Placental Lactogen I (PL-1), a protein synthesized specifically by TGCs. While wildtype TGCs showed abundant PL-1 in the cytoplasm, Znhit1 mutants displayed significantly less cytoplasmic PL-1 (Figure 3B). These observations confirm defective TE differentiation and suggest impaired TGC functions during peri-implantation stages.

Figure 3.

Implantation defects in Znhit1 mutant embryos. (A) Representative brightfield and MaxIP confocal images showing gross morphology, OCT4⁺ (green), and CDX2⁺ (magenta) cells in control (n = 10) or Znhit1 mutant outgrowths (n = 4). The red dashed line indicates presumed ICM-derived colony, and the black dashed line indicates the area of trophoblast cells. Scale bar, 100 μm. (B) Representative MaxIP confocal images showing PL-1 (yellow) and DAPI (blue) in control (n = 11) and Znhit1 mutant outgrowths (n = 4). Scale bar, 100 μm. The insets highlight the different sizes of trophoblast nuclei and different intensity of cytoplasmic PL-1 in corresponding outgrowth. (C) Representative MaxIP confocal images showing OCT4⁺ (green) and CDX2⁺ (magenta) cells in whole-mount control (n = 4) and Znhit1 mutant embryos (n = 4) collected at E4.0. Arrowheads point to cells that are OCT4⁺ CDX2⁺ double positive in mutants. Scale bar, 20 μm. (D) Representative sectioned IF images showing localization of OCT4⁺ (orange), CDX2⁺ (green), GATA6⁺ (yellow), Laminin⁺ (magenta) in control and probable Znhit1 mutant embryos (n = 4 out of 20) at E5.25. The same sections were subject to three rounds of IF (three merged images indicate the antibodies used for each round). Arrowheads point to the detached GATA6+ Laminin+ cells in the presumed Znhit1 mutant embryo. Scale bar, 50 μm. (E) Illustration of the localization of trophoblast giant cells, parietal endoderm, and decidua cells. Created with BioRender.com.

To investigate whether Znhit1 mutants could implant in vivo, we performed IF on E4.0 embryos collected from heterozygous intercrosses. E4.0 Znhit1 mutant embryos had normal CDX2 intensity compared with control embryos, but some CDX2⁺ TE cells still retained OCT4 expression, suggesting incomplete TE differentiation (arrowheads, Figure 3C). Additionally, sections of in vivo E5.25 embryos were examined. At this stage, embryos typically form egg-cylinders with OCT4⁺ EPI and CDX2⁺ ExE surrounded by GATA6⁺ VE (as shown in Figure 3D upper panel). The Reichert membrane, formed by GATA6⁺ LAMININ⁺ parietal endoderm (PE) cells, is anchored to the decidua by TGCs (Figure 3D upper panel, cartoon illustration in Figure 3E). Although a lack of working ZNHIT1 antibody prevented mutant identification, 4 out of 20 embryos exhibited a striking phenotypic difference from littermates. These four presumed Znhit1 mutants were smaller in size and consistently exhibited EPI, ExE, and VE cells; however, the GATA6+, LAMININ+ PE cells were not adhered tightly to the maternal decidua, but instead appeared disconnected in uterine crypts (arrowheads in Figure 3D lower panel). This finding is consistent with our observations of defective TGCs in vitro. These results collectively explain the failure of Znhit1 mutants to develop beyond peri-implantation stages resulting in their absence at E7.5 (likely due to embryo death and resorption).

Lack of ZNHIT1 disrupts Fgf4 expression and cell projection organization likely due to insufficient H2A.Z deposition

We next conducted transcriptome analyses on freshly collected stage-matched blastocysts to examine genome-wide changes in Znhit1 mutants. A total of 814 differentially expressed genes (DEGs) were identified, with 78% being downregulated (Figure 4A). Notably, Fgf4 was among those transcripts that were reduced in Znhit1 mutants, showing nearly 2048-fold reduction relative to littermates. This loss of Fgf4 may explain the reduction of SOX17⁺ cells and delayed PrE formation. Gene set enrichment analysis (GSEA) revealed that “regulation of cell projection organization,” a critical feature for TGC function and epithelial invasion, was significantly disturbed (Figure 4B and C), which is consistent with our earlier observations of compromised TGC function and unsuccessful implantation of Znhit1 mutant embryos.

Figure 4.

Genome-wide expression and H2A.Z changes in Znhit1 mutant embryos. (A) Volcano plot showing DEGs in Znhit1 mutant blastocysts (n = 2), compared with control blastocysts (n = 4). The dashed line indicates adjusted P-value = 0.01, Log2 Fold Change = +/−0.3. (B) GSEA showing “regulation of cell projection organization” is disturbed in Znhit1 mutant blastocysts. (C) Heatmap showing the expression of “regulation of cell projection organization” genes that were significantly reduced in Znhit1 mutant blastocysts. (D) Cartoon model of ZNHIT1-H2A.Z machinery. ZNHIT1 is a member of SRCAP complex and facilitates H2A.Z replacement of H2A, promoting gene expression. Created with BioRender.com. (E) Venn diagram showing the overlap between H2A.Z-enriched promoters [25] and downregulated genes in Znhit1 mutants. (F) Metascape [26] gene ontology enrichment analysis of overlapping genes in (E) (n = 205). The asterisk marks one of the most disturbed pathways, “regulation of plasma membrane bounded cell projection organization.” (G) CUT&TAG results showing H2A.Z enrichment peaks at Fgf4, Sema3b, and Skil loci in Znhit1^+/+ and Znhit1  ^−/− embryonic cardiomyocytes [21]. (H) Quantitative RT-PCR showing no change in the expression of H2afz and H2afv transcripts in Znhit1 wildtype, heterozygous, and mutant blastocysts. Expressions were normalized to Gapdh. Number of circles indicates the number of embryos analyzed in each group. One-way ANOVA was conducted for statistical analysis. (I) Representative MaxIP confocal images showing expression of H2A.Z (green) in Znhit1 wildtype, heterozygous, and mutant blastocysts. Scale bar, 20 μm. (J) Bar chart comparing relative intensity of H2A.Z by genotype. n indicates the number of embryos included in each group, and each point represents data from one nucleus. One-way ANOVA was conducted for statistical analysis. ^**** indicates P ≤ 0.0001.

ZNHIT1 has been shown to be an integral member of SRCAP complex, playing a key role in the genome-wide exchange of H2A for H2A.Z (illustrated in Figure 4D) [20–22]. To investigate the relevance of this mechanism in blastocysts, we used data from a recent study that generated ultra-low-input naïve chromatin immunoprecipitation and sequencing (ULI-NChIP-seq) [25], which demonstrated H2A.Z deposition at gene promoters within +/−1 kb of transcription start site (TSS) in early mouse embryos. We compared the downregulated genes in Znhit1 mutants with sites of H2A.Z deposition and found 205 genes or 32% overlap in genomic loci (Figure 4E). These genes, potentially direct targets of the ZNHIT1-mediated H2A.Z deposition, were significantly enriched in genes involved in “regulation of plasma membrane-bounded cell projection organization” pathway (asterisk, Figure 4F), further suggesting defective TGC invasion in Znhit1 mutant embryos. Notably, Fgf4 also emerged as a direct target as the Fgf4 promoter showed abundant H2A.Z deposition throughout early developmental stages, further suggesting Fgf4 regulation through the chromatin remodeling SRCAP complex.

To assess changes in H2A.Z deposition in the absence of ZNHIT1, we reanalyzed the published Cut&Tag dataset from Xu et al. [21], where H2A.Z incorporation was examined in wildtype and Znhit1^−/− embryonic cardiomyocytes. Notably, H2A.Z loss at the promoters of Fgf4 was evident in this data set, as well as 67% of the cell projection organization–related genes that we identified in Znhit1 mutant blastocysts (such as Sema3b and Skil genes shown in Figure 4G). These results further indicate the indispensable role of ZNHIT1 in promoting FGF4 function during preimplantation and cell projections required for successful implantation.

To confirm that the diminished H2A.Z deposition was not due to reduced H2A.Z expression, we conducted RT-qPCR to examine the two hypervariants, H2afz and H2afv, encoded by distinct loci in the mouse genome, yet both encoding H2A.Z [27]. Notably, neither H2afz nor H2afv exhibited reduced expression in Znhit1 mutants (Figure 4H), suggesting that the loss of H2A.Z deposition was not due to reduced H2A.Z transcription but likely from insufficient incorporation at these loci. Importantly, we also documented a dose–response of global H2A.Z protein relative for Znhit1 genotype in blastocysts. We observed that H2A.Z is most abundant in wildtype embryos, intermediate in heterozygous embryos, and lowest in mutant embryos (Figure 4I and J). It has been shown that accumulation of unincorporated histones can have deleterious consequences so that unbound histones always undergo prompt proteolysis [28,29]. Therefore, it is likely that the observed ZNHIT1 dose-dependent reduction of H2A.Z is due to cellular metabolism of unincorporated histones.

ZNHIT2 ensures splicing by facilitating nuclear translocation of U5 protein EFTUD2

Global transcriptional analysis of Znhit2 mutants and controls resulted in 1012 DEGs, with 73% exhibiting downregulation. Similar to the results from Znhit1 mutants, Znhit2 DEGs included critical developmental genes such as Fgf4 and Fgfr1 (Figure 5A). Enrichment analysis uncovered alterations in fundamental cellular functions, including DNA repair and MAP kinase activity (Figure 5B). Moreover, there was a notable upregulation of p53 signaling (Figure 5B), suggesting DNA damage and induction of apoptosis. Interestingly, Znhit2 mutants also displayed a significant increase in genes associated with ribonucleoprotein complex biogenesis and RNA splicing (Figure 5B).

Figure 5.

ZNHIT2 regulates U5-snRNP nuclear entry. (A) Volcano plot showing DEGs in Znhit2 mutant blastocysts (n = 2), compared with control embryos (n = 6). Dashed line indicates adjusted P-value = 0.01, Log2 Fold Change = +/−0.3. (B) Metascape [26] gene ontology enrichment analysis of downregulated (blue) and upregulated (orange/red) genes in Znhit2 mutant blastocysts. (C) Representative MaxIP confocal images showing EFTUD2 (orange) in control (n = 7) and Znhit2 mutant blastocysts (n = 4). Note the complete absence of nuclear EFTUD2 in mutants. Scale bar, 20 μm. (D) Bar chart comparing intronic RNA-sequencing reads (from poly-A selected RNA-seq) in control and Znhit2 mutant blastocysts. Unpaired t-test was conducted for statistical analysis. ^* indicates P ≤ 0.05. (E) Comparison of normalized transcript per million of U5 components Eftud2, Snrnp200, Prp8, and R2TP members Ruvbl1, Ruvbl2, Rpap3, in control and Znhit2 mutant blastocysts. Unpaired t-test was conducted for statistical analysis. Note that none of those transcripts were downregulated in mutants. (F) Representative single-plane confocal images showing expression of EFTUD2 (orange) and SOX17 (white) in whole-mount Znhit2 mutant blastocysts (n = 3) after FGF4-supplemented culture. Scale bar, 20 μm. (G) Cartoon model of ZNHIT2 regulation of nuclear translocation of EFTUD2 (and U5-snRNP). Created with BioRender.com.

RNA splicing, facilitated by the spliceosome, is a crucial mechanism for regulating gene expression within cells. U5 snRNP, an integral component of the spliceosome, undergoes a series of sequential events to become functional. This process initiates with the formation of a cytoplasmic assembly complex that includes the U5-specific proteins EFTUD2 and SNRNP200, accompanied by the cochaperone R2TP/Prefoldin-like complex (R2TP/PFDL) and co-factors such as ZNHIT2 [15, 23]. Subsequently, this complex translocates into the nucleus for final maturation [23]. To assess the impact of Znhit2 knockout on U5 snRNP, we performed IF to examine the localization of U5-specific protein EFTUD2. Strikingly, in Znhit2 mutants, nuclear localization of EFTUD2 was completely abolished (Figure 5C), indicating impaired U5 snRNP translocation and defective spliceosome. To assess splicing, we quantified the percentage of alignments in intronic regions from our polyA-selected RNA sequencing data and found a significant increase in Znhit2 mutants compared with control embryos, confirming splicing defects at loci throughout the genome (Figure 5D). Importantly, no reduction in the mRNA levels of Eftud2 or other U5 proteins such as Snrnp200 and Prp8, R2TP/PFDL members Ruvbl1/2, Rpap3 were detected (Figure 5E). Therefore, the lack of nuclear EFTUD2 and splicing defects are not due to transcriptional regulation of spliceosome components but likely attributed to the failed nuclear translocation in the absence of ZNHIT2.

Since we have shown that exogenous FGF4 was able to rescue some of the developmental defects in Znhit2 mutant embryos, namely, primitive endoderm specification, we also asked if FGF4 signaling could restore U5 splicing function (nuclear EFTUD2). Our results clearly showed that nuclear EFTUD2 was not restored even when SOX17 expression was rescued (Figure 5F), indicating that splicing defects in Znhit2 mutants are not downstream of FGF4 signaling and are likely a direct consequence of the lack of ZNHIT2 function in facilitating the nuclear translocation of EFTUD2 and ensuring the global splicing efficiency in blastocysts (Figure 5G).

Discussion

Here, we report that both ZNHIT1 and ZNHIT2 are essential for early mammalian development. Despite sharing similar overt peri-implantation phenotypes, ZNHIT1 and ZNHIT2 regulate gene expression through distinct mechanisms. Our data indicate that ZNHIT1 facilitates H2A.Z deposition, promoting implantation-related genes, while ZNHIT2 ensures global splicing efficiency by facilitating the nuclear translocation of EFTUD2.

ZNHIT1 regulates H2A.Z deposition

Embryos lacking ZNHIT1 were able to form blastocysts, expand, and hatch out of the zona (Figure 1B), but exhibited reduced SOX2⁺ and SOX17⁺ cells (Figures 1C and 2A). Znhit1 mutant implantation defects, observed in vitro and in vivo, include impaired TE differentiation and defective TGCs (Figure 3). Transcriptional analysis revealed the downregulation of Fgf4 in Znhit1 mutant blastocysts (Figure 4A), potentially explaining the decreased PrE, which can be rescued through exogenous FGF4. In addition, the disruption in genes associated with cell projection organization was significant (Figure 4B and C). One recent study [25] showed that 76% (16/21) of the downregulated “cell projection organization” pathway genes, as well as the FGF4 locus, showed abundant H2A.Z deposition in the promoter regions of normal blastocysts, suggesting direct regulation of H2A.Z deposition by ZNHIT1 at these loci. Furthermore, data from ZNHIT1-deficient embryonic cardiomyocytes confirm the loss of H2A.Z at majority of these loci upon Znhit1 knockout [21]. These findings support the crucial role of ZNHIT1 in promoting Fgf4 expression in ICM and regulating the invasive function of TGCs during implantation through facilitating H2A.Z incorporation.

H2A.Z, the most extensively studied histone variant of H2A, is essential for early embryonic development. H2A.Z mutant embryos became significantly underrepresented at E5.5 and cannot be recovered at E7.5 [30], a similar lethality window that we observed for Znhit1 mutants. Notably, Znhit1 expression closely mirrors H2A.Z expression, both being present in zygotes, becoming undetectable at two-cell stage during zygotic genome activation and reappearing at eight-cell stage through the blastocyst stage ([31,32], https://websites.umass.edu/jmager/early-gene-expression/). Similar to Znhit1 mutants, H2A.Z mutants formed blastocysts, expanded, and hatched without issues, but disruptions in both ICM and TE-derived populations were evident in outgrowth assays [30]. These results collectively lead us to conclude that ZNHIT1 directly facilitates H2A.Z deposition during early development in vivo.

Mouse embryos lacking H2A.Z failed to generate both embryonic and extraembryonic stem cells, underscoring its indispensable function in both lineages [30,31]. Specifically, the reduction of H2A.Z in mouse embryonic stem cells inhibited pluripotency genes [33], a trend consistent with our observation in Znhit1 mutants. These mutants displayed fewer SOX2⁺ cells and a diminished ICM colony in outgrowth assays. Moreover, the critical role of H2A.Z in TGC differentiation is highlighted in a trophoblast stem cell (TSC) study that revealed a dramatic and immediate increase in H2A.Z/pan-H2A as TSC differentiation started, with sustained high expression of H2afz throughout the differentiation process [34]. Our observations of Znhit1 mutants failing to form cell projections align with these findings, emphasizing that timely incorporation of H2A.Z into genomic loci is critical for successful implantation.

Interestingly, H2A.Z transcripts in outgrowths exhibited a roughly 12-fold reduction in the ICM colony compared with trophoblasts [35]; however, similar to a different study [36], our IF results showed similar expression of H2A.Z in ICM and TE in E3.5 blastocysts. This discrepancy likely arises from differences in vivo versus in vitro culture conditions.

ZNHIT2 is required for nuclear translocation of U5 snRNP

Knockout of Znhit2 did not impede blastocyst formation but resulted in tight junction disruption, hatching failure (Figure 1B and G), and the depletion of the critical pluripotency gene SOX2 and PrE marker SOX17 (Figures 1C and 2A). Exogenous FGF4 could successfully induce one or two SOX17⁺ cells by the end of the culture (Figure 2B), with 20% mutants beginning to hatch (Figure 2G), demonstrating that ZNHIT2-deficient embryos have a limited response to FGF signaling. Nearly 80% of DEGs were downregulated in Znhit2 mutants including Fgf4 and Fgfr1. This implies a more extensive disruption to FGF signaling (compared with Znhit1 mutants that show no reduction in Fgfr1), which may explain why FGF4 was unable to rescue Znhit2 mutants as effectively as Znhit1 mutants.

ZNHIT2 has been shown as a bridging factor mediating the interaction between U5 snRNP and its cochaperone complex [15]. Our work, for the first time, demonstrated that the knockout of Znhit2 led to a complete absence of nuclear EFTUD2 (Figure 5C), with no reduction of Eftud2 or other U5 or R2TP/PFDL mRNA levels (Figure 5E). We also showed a significant increase in intronic transcripts present in Znhit2 mutants further confirming global splicing defects (Figure 5D). There have been multiple studies investigating the temporal order and the coordination between splicing and polyadenylation [37–39]; thus, it is possible that transcripts with defective splicing were not captured in our current dataset because their polyadenylation had also been compromised. Furthermore, our Znhit2 mutant transcriptomics show an increase in genes involved in “ribonucleoprotein complex biogenesis” and “RNA splicing” (Figure 5B), suggesting that Znhit2 null cells are compensating for the loss of ZNHIT2 function. This result underscores the critical role of ZNHIT2 in facilitating the translocation of U5 snRNP during early embryonic development.

ZNHIT1 and ZNHIT2 function converge on FGF4 signaling

In both Znhit1 and Znhit2 mutants, we observed reduced SOX2⁺ cells, no change in Sox2 mRNA, and a loss of Fgf4 mRNA. This suggests that the loss of Fgf4 may not be due to insufficient activation of Sox2 [40]. Another intriguing phenotype observed in both mutant phenotypes is the presence of NANOG⁺ TE cells in mutant blastocysts (Figure 2A). One possible explanation in Znhit1 mutants is the reduction in CDX2 intensity in those cells that are CDX2⁺, causing a delay in repression of the Nanog locus by CDX2 in TE (Figure 2H). The rescue of NANOG+ TE cells with exogenous FGF4 in both mutants supports the notion that FGF4 acts as paracrine signal for trophectoderm proliferation or differentiation [41,42]. Two separate studies have shown the collaborative and essential role of FGF signaling in lineage establishment within the ICM [4,43]. Our observations on NANOG mis-expression and rescue by FGF4 offer an interesting future direction to explore the role of FGF signaling in TE progression and function.

Since both ZNHIT1 and ZNHIT2 have been shown to interact through RUVBL1/2, we investigated whether they share any functions in early embryos. However, Znhit1/2 double knockdown embryos still developed to the blastocyst stage (Figure 2I), indicating that they likely do not have redundant roles. One limitation of this work is the challenge to exclude the influence and presence of maternal protein. Several commercial antibodies against ZNHIT1 and ZNHIT2 did not show specific localization, so we could not determine when the proteins were absent in mutant embryos. Nevertheless, our findings highlight the indispensable and unique contributions of ZNHIT1 and ZNHIT2 to early mammalian development. ZNHIT1 promotes implantation likely through the incorporation of H2A.Z into the genome, and ZNHIT2 ensures global splicing by facilitating the nuclear translocation of EFTUD2.

Materials and methods

Animals

Use of animals and all experiments were approved by the University of Massachusetts Institutional Use and Care of Animals Committee (IACUC). Knockout alleles used were produced by the Jackson Laboratory: C57BL/6NJ-Znhit1em1(IMPC)J/Mmjax and C57BL/6NJ-Znhit2em1(IMPC)J/Mmjax. Briefly, the Znhit1 mutant allele was generated by a deletion of 2990 bp in exons 2–6 and 2315 bp of flanking intronic sequence. The Znhit2 mutant allele was generated by a deletion of 1563 bp in exon 1 and 282 bp of flanking intronic sequence. Specific allele details are available at www.informatics.jax.org.

Embryo retrieval, culture, and genotyping

Mutant embryos were generated by heterozygote intercrosses from each line separately (Znhit1 and Znhit2). Males and females were housed together, and the presence of a vaginal plug in the morning indicates E0.5 of development. E2.5 embryos were collected by flushing oviducts with M2 media (EmbryoMax M2 Medium, MilliporeSigma, MR-015-D) for further culture. E3.5 embryos were collected by flushing the uterus with M2 media for culture, outgrowth media (describe below) for outgrowths, or with phosphate-buffered saline (PBS) with 0.03% (w/v) polyvinylpyrrolidone for other purposes. E4.0 embryos were collected by flushing the uterus with PBS. Embryos were cultured in potassium simplex optimized medium (KSOM) (MilliporeSigma, MR-106-D) with oil (EmbryoMax Light Mineral Oil, EMD Millipore, ES-005-C) overlay at 37°C and 5% O₂, 5% CO₂ atmosphere in a humidified incubator. For exogenous FGF4 treatment, we used 500 ng/mL FGF4 (R&D Systems, 235-F4-025) with 1 μg/mL heparin (Millipore Sigma, H3149) as final concentration in KSOM.

Genotypes of individual embryos were determined by PCR using the following primers: Znhit1 wildtype (5′- GATGATGCCAATGAGCGCAG-3′ and 5′- GCGTGTGCAGGAATGAACAG-3′) and mutant (5′- AGGAGACTGGGGCTCTAGG-3′ and 5′- CAGCAGCTAAGAAAGGCAGA-3′). Znhit2 wildtype (5′- TGTGACAGCTGAGGAGATGG −3′ and 5’-CCAGTTCAGCCAGGGAGTTC-3′) and mutant (5′- TCGGAAATGTGCGGAAAGGA −3′ and 5′- TACTGCAAGCGTGTGCCTTA −3′. All results shown were repeated on a minimum of three mutant and control embryos from each knockout line, unless indicated otherwise. For all results shown, a minimum of three embryos of each genotype were assessed, unless labeled otherwise.

RNA extraction, complementary DNA synthesis, RT-PCR analysis, and quantitative RT-PCR

RNA was isolated from embryos using the Zymo Quick-DNA/RNA Microprep Plus Kit (Zymo Research, D7005) and was converted into complementary DNA (cDNA) using the BioRad iScript cDNA Synthesis Kit (BioRad 1708890), according to the manufacturer’s protocol. RT-PCR was performed for 35 cycles of 30 s at 60°C, 72°C, and 95°C with the following primers: Mrpl17 (414 bp), 5’-CCATCTGCTGCGGAACTTG-3′, 5’-TTGCGTCCTGGTTATGGTG-3′; Oct4 (467 bp), 5’-AGCTGCTGAAGCAGAAGAGG-3′, 5’-TGGGAAAGGTGTCCCTGTAG-3′; Sox2 (466 bp), 5′- AGAACCCCAAGATGCACAAC-3′, 5′- ATGTAGGTCTGCGAGCTGGT-3′;Znhit1(228 bp), 5’-CAGCTGGAGGCATTGGAGAA-3′, 5’-GGCACAGGCTGTCAAGTAGT-3′; Znhit2 (171 bp), 5’-GGCCTTTCTGGACTTTGGGA-3′, 5’-ATCTCTGTTCGCGGCATGAT-3′.

Quantitative RT-PCR (qPCR) was performed on a Stratagene MX3005p using Thermo Fisher TaqMan Gene Expression Assays and Quanta Supermix (product #95078) probe-based reactions. All qPCR reactions included Gapdh Cy5-labeled multiplex control. One embryo equivalent of cDNA was used for each qPCR reaction with a minimum of three replicates for all results shown. The Gapdh expression assay was purchased from Integrated DNA Technologies (IDT), Mm.PT.39a.1. TaqMan Gene Expression Assays from Thermo Fisher used are as follows: H2afz VIC-labeled, Mm05916395_g1, H2afv FAM-labeled, Mm01181326_m1.

Outgrowth assays

Blastocysts were harvested at E3.5 and cultured individually in iBiTreat dishes (u-slide 8 well, 80826-90) with Dulbecco Modified Eagle Medium (Lonza, Allendale, NJ, USA) containing 10% (v/v) fetal bovine serum (Atlanta Biologicals, Flowery Branch, GA, USA), 1% (v/v) GlutaMAX (Thermo Fisher Scientific) and 1% (v/v) Penn-strep. Embryos were kept in culture for 72 h at 37°C and 5% O₂, 5% CO₂ atmosphere in a humidified incubator, followed by imaging and genotyping.

Immunofluorescence

For E5.25 dissections, freshly dissected decidua were fixed in 4% paraformaldehyde overnight, dehydrated in an increasing ethanol series (25%, 50%, 70%, 80%, 90%, 100% v/v, each 2× for 30 min), cleared with xylene (2× for 30 min), embedded with paraffin, and sectioned at 7 μm.

Sections were deparaffinized in xylene (2× at 5 min) and rehydrated in decreasing ethanol series (100% 2×, 90%, 80%, 70% v/v, 3 min each), MilliQ water (3 min), antigen retrieval with Tris-EDTA buffer (pH 9) until boiling then 5 min at power 1 (Samsung microwave, MW538W), and cooled to room temperature. The slides were then rinsed twice in PBS/0.01% (v/v) Tween 20 (PBT) for 2 min and blocked with 0.5% (w/v) milk in PBT for 2 h at room temperature in a humid chamber. Primary antibodies were applied in 0.05% milk/PBT in volumes of 200 μL with parafilm on top and incubated at 4°C overnight in a humid chamber. The slides were then rinsed three times with PBT for 15 min, followed by incubation with secondary antibodies at room temperature for 1 h in a humid chamber. The slides were rinsed with PBS for 15 min, three times. Nuclei were counterstained with DAPI (4′,6-diamidino-2-phenylindole) in PBS (1 μg/mL) for 3 min. Slides were rinsed with 1× PBS and sealed with ProLong Gold (Thermo Fisher Scientific, P36934). Fluorescent slides were imaged with Panoramic MIDI II slide scanner (3DHISTECH) or Nikon Ti2 Eclipse inverted microscope.

Freshly dissected E3.5 embryos and outgrowths were fixed in 4% (w/v) paraformaldehyde for 10 or 15 min at room temperature, permeabilized in 1% (v/v) Triton-X, and blocked with 0.1% (w/v) bovine serum albumin (BSA) supplemented with 0.01% (v/v) Tween20 and 20 μg/mL Fab fragment. Primary antibodies were diluted in 0.1% BSA supplemented with 0.01% Tween 20 (block) and incubated overnight at 4°C. The embryos were then rinsed with block for 30 min, followed by incubation with secondary antibodies at room temperature for an hour. Next, embryos were stained with DAPI in PBS (1 μg/mL) for 3 min and rinsed thoroughly. Images were taken with a Nikon Eclipse Ti Series inverted microscope with a C2 confocal attachment.

The primary antibodies used and their dilution include SOX2 (R and D Systems Cat# AF2018, RRID:AB_355110, 1:200), CDX2 (BioGenex Cat# MU392A, RRID:AB_2923402, 1:50 for whole mount and 1:200 for section); ZO-1 (Thermo Fisher Scientific Cat# 33-9100, RRID:AB_2533147, 1:100); SOX17 (R and D Systems Cat# AF1924, 1:200): NANOG (Reprocell Cat# RCAB001P, 1:500): Oct4 (R and D Systems Cat# AF1759, RRID:AB_354975, 1:200 for whole mount and section); PL-1 (Santa Cruz Biotechnology Cat# sc-376436, RRID:AB_11151405, 1:100); GATA6 (R and D Systems Cat# AF1700, RRID:AB_2108901, 1:100); Laminin (Sigma-Aldrich Cat# L9393, RRID:AB_477163, 1:250); H2A.Z (Thermo Fisher Scientific Cat# PA5-21923, RRID:AB_11152068, 1:200); and EFTUD2 (Abcam Cat# ab188327, 1:200).

The secondary antibodies used include Alexa Fluor 488 donkey anti-mouse (Thermo Fisher Scientific, A21202, 1:500); Alexa Fluor 546 donkey anti-rabbit (Thermo Fisher Scientific, A10040, 1:500); Alexa Fluor 647 donkey anti-goat (Thermo Fisher Scientific, A21447, 1:500); and DAPI (Molecular Probes, 1 μg/mL).

Quantification of cell number and protein intensity

The analysis was conducted using the Nikon Elements Analysis 3.1 software. Semiautomated analyses were carried out using the General Analysis 3 (GA3) software. This software provides a range of image processing tools within each channel to facilitate identification of individual objects, which involves combining intensity range, separation, smoothing, and object binary functions to isolate and identify single nuclei within an embryo. Once identified, the program calculated general parameters such as object number and quantification of signal intensity in each specific channel. To facilitate statistical analysis of protein intensities, the intensity of interest protein/DAPI ratios (CDX2/DAPI or H2A.Z/DAPI) were first determined for each cell using NIS Element. Within each litter, one wildtype embryo was selected as the normalizer, and the average of all its ratios was calculated. Subsequently, the ratios of the other littermates were normalized against this average. This approach allowed for the pooling of embryos from multiple litters while mitigating the issue of varying laser power used for different litters to avoid saturation and optimize the intensity range.

Statistical analyses were performed using Prism (Version 10). To assess statistical differences among three or more groups, one-way analysis of variance (ANOVA) was conducted. The assumption of equal standard deviations was tested using Bartlett test. In cases where this assumption was violated, Welch ANOVA was used instead. An unpaired t-test was performed for comparisons between two groups. Statistical significance was set as P = 0.05.

RNA-seq library preparation and sequencing

RNA was extracted from E3.5 embryos individually using the Zymo kit as previously described. Total RNA was analyzed using the Agilent 2100 Bioanalyzer using Eukaryote Total RNA Pico assay (Agilent Technologies Inc). RNaseq libraries were prepared using the NEBNext Single Cell/Low Input RNA Library Prep Kit for Illumina (New England Biolabs) following manufacturer instructions. The quantity of library was assayed using Qubit DNA HS assay (Life Technologies Corp), and quality was analyzed on Bioanalyzer (Agilent Technologies Inc). Libraries were sequenced on Illumina NextSeq 500 platform using NextSeq 500/550 Mid Output v2.5 kit (150 cycles) with 76 bp paired-end sequencing chemistry. RNAseq library preparation and Illumina Next-Generation Sequencing were performed at Genomics Resource Laboratory (RRID:SCR_017907), Institute for Applied Life Sciences, University of Massachusetts, Amherst, MA.

siRNA

TriFECTa Kit DsiRNA Duplex for Znhit1 and Znhit2 was purchased from IDT. Znhit1 duplex includes three DsiRNAs targeting transcript NM_027318, namely, mm.Ri.Znhit1.13.1 (Reference #:447613348), mm.Ri.Znhit1.13.2 (Reference #:447613351), and mm.Ri.Znhit1.13.3 (Reference #:447613354). The Znhit2 duplex includes three DsiRNAs targeting transcript NM_013859, namely, mm.Ri.Znhit2.13.1 (Reference #:447613334), mm.Ri.Znhit2.13.2 (Reference #:447613337), and mm.Ri.Znhit2.13.3 (Reference #:447613340). Control DsiRNA was provided with the kit. All siRNAs were resuspended in RNase-free water to 100 μM solutions.

Microinjection

To prepare zygotes for siRNA microinjection or in vitro culture, B6D2F1 female mice 8–10 weeks old were induced to superovulate with 7.5 IU pregnant mare serum gonadotropin (PMSG, Sigma-Aldrich), followed 48 h later by 7.5 IU human chorionic gonadotropin (hCG, Sigma-Aldrich). Females were mated with B6D2F1 males and euthanized at 20 h post-hCG injection. Oviductal ampullae were dissected to release zygotes, and cumulus cells were removed by pipetting in M2 medium containing hyaluronidase (EMD Millipore), followed by thorough washes in M2 medium (EMD Millipore).

Microinjection was performed in M2 medium using a Nikon inverted microscope equipped with a piezo-driven (Prime Tech, Japan) micromanipulator (TransferMan NK2, Eppendorf, Hamburg, Germany). A volume of 5–10 pl (100 μM) of control or gene specific siRNA was microinjected into the cytoplasm of zygotes using a blunt-ended pipette of 6–7 μm in diameter. After microinjection, zygotes were washed with M2 medium and cultured in KSOM medium for 96 h at 37°C in a humidified atmosphere of 5% CO₂/5% O₂ balanced in N₂. The efficiency of microinjection was assessed by carrying out the RT-PCR of a single embryo at the end of culture.

RNA-seq analysis

FastQC was used to validate read quality before alignment. HISAT2 was used to index the datasets against National Center for Biotechnology Information (NCBI) mouse genome GRCm39 using default values for paired-end alignments. DESeq2 package was used to identify DEGs compared with the control dataset (adjusted P-value <0.05). Volcano plots were created with DEGs using the EnhancedVolcano (v1.14.0) R package. Heatmaps were generated with normalized read counts using the pheatmap (v1.0.12), RColorBrewer (v1.1-3), and dendextend (v1.17.1) R packages. A ranked gene list (RNK) file that was made for GSEA was made from using the DESeq2 package as previously described. In brief, the RNK file contains a single, rank-ordered gene list in a tab-delimited text format. The GSEA gene set was selected from the default top-scoring dataset. We acknowledge our use of the GSEA, GSEA software, and Molecular Signature Database (MSigDB) (http://www.broad.mit.edu/gsea/) [44]. Gene ontology enrichment analysis was generated with Metascape (https://metascape.org) using default parameters. The gene list of promoters that have H2A.Z deposition in early mouse embryos was provided by Dr. Yi-Liang Miao. The Binary Alignment Map files of E13.5 embryonic cardiomyocytes were provided by Dr. Zhongzhou Yang, and H2A.Z deposition was visualized using WashU Epigenome Browser, on mm10 mouse genome. The percentages of intronic reads were calculated using STAR on BasePair, and the P-value was calculated by t-test.

Funding

This work was funded by USDA National Institute of Food and Agriculture/Hatch award 1024792 to WC andNational Institutes of Health award HD083311 to JM.

Conflict of Interest: The authors have declared that no conflict of interest exists.

Data availability

The RNA sequencing data generated in this study are available in Gene Expression Omnibus (GEO) repository under the accession number GSE276496.