Spatiotemporal dynamics and selectivity of mRNA translation during mouse pre-implantation development

Abstract

Translational regulation plays a pivotal role during pre-implantation development. However, the mechanisms by which messenger RNAs (mRNAs) are selectively regulated over time, along with their dynamic utilization and fate during this period, remain largely unknown. Here, we performed fraction-resolved polysome profiling and characterized translational dynamics across oocytes and early embryo development. This approach allowed us to examine the changes in translation during pre-implantation development in high resolution and uncover previously unrecognized modes of translational selectivity. We observed a stage-specific delay in translation, characterized by the postponed recruitment of stored mRNAs-either unbound or associated with light ribosomal fractions-into actively translating polysomes (heavy fraction). Comparative analysis of translatome with proteomics, RNA N⁶-methyladenosine modifications, and mRNA features further revealed both coordinated and distinct regulatory mechanisms during pre-implantation development. Furthermore, we identified a eukaryotic initiation factor 1A domain containing 3, Eif1ad3, which is exclusively translated at the two-cell stage and is essential for embryonic development by regulating ribosome biogenesis and protein synthesis. Collectively, our study provides a valuable resource of spatiotemporal translational regulation in mammalian pre-implantation development and highlights a previously uncharacterized translation initiation factor critical for early embryos.

Graphical Abstract

Graphical Abstract.

Introduction

Translational control of messenger RNA (mRNA) is vital across various cellular processes, enabling rapid protein synthesis essential for both immediate cellular needs and longer-term physiological changes [1, 2]. During mammalian pre-implantation development, the degradation of maternal stored mRNAs and activation of embryonic genome are precisely regulated, in large part by the translational and posttranslational regulation [3]. For example, fully grown oocytes rely entirely on the translation of stored mRNA for maturation and fertilization, given their lack of active transcription [4]. However, a central gap in our understanding of post-transcriptional regulation exists, particularly how mRNAs are selected for spatial and temporal regulation and their dynamics of mRNA utilization and fate during oocyte maturation, fertilization, zygotic genome activation (ZGA), and early differentiation. Thus, understanding the selective mRNA translation during these critical transitions is essential for elucidating the mechanisms underlying successful embryo development.

In the past decade, the dynamics of genome-wide transcription, translation, and protein expression during mammalian pre-implantation embryo development have been characterized; however, limitations remain. First, the mRNA expression from transcriptomic profiling does not represent their functional status [5]. Second, despite the fact that proteomic analysis of oocytes and embryos has been achieved in several mammalian species by mass spectrometry [6–9], the methodology faces challenges such as low sensitivity and limited coverage due to the scarcity of material available. Third, recent studies have developed low-input ribosome profiling approaches (LiRibo-seq and Ribo-ITP) and provided the translational dynamics of mouse pre-implantation development [10, 11]. However, these studies were confined to an analysis of ribosome bound mRNAs as a whole, while ignoring the variation encountered in the different numbers of ribosome-bound mRNAs and what are the specific mRNAs preferentially selected for translation, underscoring the need for improved techniques to better understand the dynamics of mRNA usage and fate during these crucial stages.

To decipher comprehensive translational regulation in mouse pre-implantation development, we have further extended our previously standardized Scarce Sample Polysome (SSP) profiling protocol (see the ‘Materials and methods’ section) [2, 12] on mouse oocytes and pre-implantation embryos. This methodology has successfully provided systematic monitoring of non-translated RNA (free mRNA), mRNA prepared for translation (monosomes-bound mRNA), and actively translating mRNA (polysomes-bound mRNA) in a bovine embryo model [13]. However, this analysis was confined to polysome-bound mRNA to study translation efficiency, similar as those two ribosome profiling studies by LiRibo-seq or Ribo-ITP [10, 11], leaving their spatial dynamics and selectivity of translation unexplored.

To address this gap in knowledge, we employ fraction-resolved polysome profiling and generate a high-resolution (single fraction), spatiotemporal translational dynamics during mouse oocyte maturation, and pre-implantation embryo development. This comprehensive resource enables the precise characterization of endogenous mRNA states whether untranslated, actively translated, or degraded and reveals the stage-specific translational regulation of individual mRNAs throughout pre-implantation development. Among these, we highlight the Eif1ad3 and its paralogs, which are exclusively translated at the two-cell stage. This temporally restricted expression plays a critical role in driving protein synthesis immediately following ZGA, underscoring its importance in orchestrating early embryonic development.

Materials and methods

Animal care and use

All animal experiments were performed in accordance to guidelines and protocols approved by the University of Florida Institutional Animal Care and Use Committee (IACUC) under protocol number IACUC202300000558, and the Laboratory of Biochemistry and Molecular Biology of Germ Cells at the Institute of Animal Physiology and Genetics in Czech Republic under Act No. 246/1992 on the protection of animals against cruelty, issued by the Ministry of Agriculture experimental project #67756/2020MZE-18134.

Mice

Institute of Cancer Research (ICR) females (produced in house, IAPG, Czech Republic) and CF-1 females (Envigo, Indianapolis, IN, USA) used for oocytes and embryo collection for the SSP profiling and knock down experiments. B6SJLF1/J males used for breeding were obtained from Jackson Laboratory.

Antibodies and reagents

A list of all primary antibodies, reagents, as well as oligonucleotides used in this study can be found in Supplementary Table S1.

Oocyte and embryo isolation, cultivation, and inhibitor treatment

Oocytes were collected from ICR female mice (6 weeks old) maintained under a 12 h light/dark cycle. For germinal vesicle (GV) oocytes, females received an intraperitoneal injection of 5 IU pregnant mare serum gonadotropin (PMSG) 46 h before collection. Fully grown GV oocytes were isolated in transfer medium (TM) containing 100 μM 3-isobutyl-1-methylxanthine (IBMX) to inhibit meiotic resumption. After denudation, oocytes were cultured in M16 medium without IBMX at 37°C, 5% CO₂ for 0 h (GV stage) or 12 h [metaphase II (MII) stage]. For embryo collection, PMSG-primed females received 5 IU human chorionic gonadotropin (hCG) and were mated overnight with ICR males. The day a vaginal plug was detected was designated embryonic day (E) 0.5. Pronuclear-stage zygotes were recovered from oviducts at 18 h post-hCG and cultured in K-Simplex Optimised Medium (KSOM) at 37°C, 5% CO₂. Embryos were harvested at specific developmental stages: two-cell (∼24–30 h post-fertilization, E1.5), four-cell (∼44–52 h, E2.5), eight-cell (∼56–64 h, E2.75–E3.0), morula (∼72–84 h, E3.0–E3.25), and blastocyst (∼96–108 h, E3.5–E4.5).

In particular, since major genome activation occurs at the two-cell stage, the timing of embryo collection reflects the transcriptionally active state. We collected mid-two-cell stage embryos, characterized by active transcriptional processes, as confirmed by α-Amanitin treatment (1.0 μg/ml, Sigma–Aldrich) for 2 h (Supplementary Fig. S1A and B). Two-cell embryos were treated with 5 or 10 μM of W-7 hydrochloride (Merck), and untreated cells were considered as the control group. Embryos were kept in the presence of the inhibitor for the rest of the experiment, and the development status was checked for each stage.

Isolation of ribosome-bound mRNA

Approximately 200 oocytes (GV or MII oocyte) or embryos at different developmental stages (zygote, two-, four-, eight-cell, morula, and blastocyst) were combined with lysis buffer containing 10 mM HEPES (pH 7.5), 5 mM KCl, 5 mM MgCl₂, 2 mM dithiothreitol (DTT), 1% Triton X-100, 100 μg/ml cycloheximide, complete ethylenediaminetetraacetic acid (EDTA)-free protease inhibitor (Roche), and 40 U/ml RNase inhibitor (Ribolock, Thermo). Oocytes and embryos were disrupted by zirconium silica beads (Sigma) in the mixer mill apparatus MM301 (shake frequency 30, total time 45 s, Retsch). Lysates were cleaned by centrifugation in 10 000 × g for 5 min at 4°C and the supernatants were loaded into 10%–40% linear sucrose gradients containing 10 mM HEPES (pH7.5), 100 mM KCl, 5 mM MgCl₂, 2 mM DTT, 100 μg/ml cycloheximide, complete EDTA-free protease inhibitor, and 5 U/ml RNase inhibitor. Ultracentrifugation was carried out with a SW55Ti rotor and Optima L-90 Ultracentrifuge (Beckman Coulter). Polysome profiles were recorded by ISCO ultraviolet absorbance reader. Ten equal fractions were then recovered and subjected to RNA isolation by Trizol reagent (Sigma).

Library preparation and RNA sequencing

The RNA sequencing (RNA-seq) libraries were generated from individual fractions by using the Smart-seq2 v4 kit with minor modification from manufacturer’s instructions. Briefly, individual samples were lysed, and mRNA was captured and amplified with the Smart-seq2 v4 kit (Clontech). After AMPure XP beads purification, amplified RNAs were quality checked by using Agilent High Sensitivity D5000 kit (Agilent Technologies). High-quality amplified RNAs were subject to library preparation (Nextera XT DNA Library Preparation Kit; Illumina) and multiplexed by Nextera XT Indexes (Illumina). The concentration of sequencing libraries was determined by using Qubit dsDNA HS Assay Kit (Life Technologies) and KAPA Library Quantification Kits (KAPA Biosystems). The size of sequencing libraries was determined by means of High Sensitivity D5000 Assay in at Tapestation 4200 system (Agilent). Pooled indexed libraries were then sequenced on the Illumina HiSeq X platform with 150-bp paired-end reads.

A pool of 20 oocytes or preimplantation embryos (n = 4) selected from the same batch in each developmental stage used for polysome profiling were used to profile transcriptomes by RNA-seq following Smart-seq2 protocol as above described.

RNA-seq data analysis

Multiplexed sequencing reads that passed filters were trimmed to remove low-quality reads and adaptors by TrimGalore (v.0.6.7) (-q 25 –length 20 –max_n 3 –stringency 3). The quality of reads after filtering was assessed by fastQC, followed by alignment to the mouse genome (GRCm39) by HISAT2 (v.2.1.0) with default parameters. The output SAM files were converted to BAM files and sorted using SAMtools (v.1.17). Read counts of all samples were quantified by using FeatureCount (v.2.0.1) with mouse genome as a reference and were adjusted to provide Counts per million mapped reads. Principal component analysis (PCA) and cluster analysis were performed by using R (a free software environment for statistical computing and graphics). Differentially expressed genes (DEGs) were identified by using DESeq2 (v.4.2.1) in R. The genes were considered differentially expressed if they provided a false discovery rate of <0.05 and fold change >2. Clustering of time series gene expression data was performed by using Mfuzz (v.3.19) in R. ClusterProfiler (v.4.6.1) was used to reveal the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in R.

Following datasets were used for Integrative analysis: the poly(A) tail length dataset (GSE228001) from GV, MII, zygote, and two-cell stages; the N⁶-methyladenosine (m⁶A) dataset (GSE192440) from GV, MII, zygote, two-cell, eight-cell, and blastocyst stages; and the two proteomic datasets (PXD003315, PXD018777) from different stages of pre-implantation development.

Knockdown in mouse embryos

The microinjection platform was equipped with a microinjector (FemtoJet 4i, Eppendorf), an inverted microscope (Leica, DMi8), and micromanipulators (Narishige). Microinjections were performed between 20 and 24 h after hCG administration. For the Eif1ad3 knockdown (KD) experiments, one-cell embryos were collected from CF1 (Envigo) mice and microinjected with either control (RfxCas13d-IVT mRNA, 500 ng/μl) or Eif1ad KD guide RNA (gRNA), 2 μM + RfxCas13d-IVT mRNA, 500 ng/μl). Embryos from both control and KD groups were cultured to the blastocyst stage to observe development. For RNA-Seq, two-cell stage embryos (from both control and KD groups) were collected.

In vitro transcription of Eif1ad3

To generate Eif1ad3-6xHis mRNA, total RNA from two-cell embryos was extracted from a pool of mouse embryos, followed by first strand cDNA synthesis using SuperScript™ IV VILO™ Master Mix (Thermo Fisher Scientific, Waltham, MA). Primers were designed based on the current genome annotation (GRCm39; National Center for Biotechnology Information) of Eif1ad3 (Supplementary Table S1). Polymerase chain reaction (PCR) was conducted using Q5 Hot Start High-Fidelity 2× Master Mix (New England Biolabs, Ipswich, MA) with an initial denaturation step at 98°C for 30 s followed by 30 cycles at 98°C for 10 s, annealing at 58°C for 30 s and extension at 72°C for 30 s and a final extension at 72°C for 2 min. The purified PCR products were served as DNA template for in vitro transcription using HiScribe^® T7 ARCA mRNA Kit with tailing following manufacture’s instruction. The yield and integrity of resulting mRNA were assessed using Qubit 4 (Thermo Fisher Scientific, Waltham, MA) and Tapestation 4150 (Agilent Technologies, Santa Clara, CA). To overexpress Eif1ad3, in vitro transcribed mRNAs (IVT-mRNAs) were microinjected into zygotes at final concentration of 200 ng/μl.

RNA immunoprecipitation sequencing

RIP experiments were conducted using the EZ-Magna RIP Kit (Millipore Corporation, Billerica, MA). Briefly, two-cell stage embryos (over expressed 200 cells and control 200 cell per replicate) were harvested, cross-linked with 0.5% paraformaldehyde for 10min at room temperature, quenched with glycine (final concentration 125 mM), lysed, and then incubated with RIP buffer containing magnetic beads conjugated with anti-His tag antibody (Millipore) or corresponding negative control IgG (Millipore). After the antibody was recovered using protein A/G beads, the purified RNA was used to perform RNA-seq.

Immunoblotting

An exact number of cells (15–30 oocytes) were washed in polyvinyl alcohol (PVA)/phosphate buffered saline (PBS) and frozen to −80°C. Prepared samples were lysed in NuPAGE LDS Sample Buffer (NP0007, Thermo Fisher Scientific) and NuPAGE Sample Reducing Agent (NP0004, Thermo Fisher Scientific) and heated at 100°C for 5 min. Proteins were separated on precast gradient 4%–12% sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) gel (Thermo Fisher Scientific) and blotted to Immobilon P membrane (Millipore) in a semidry blotting system (Biometra GmbH) at 5 mA/cm² for 25 min. Membranes were then blocked in 5% skimmed milk dissolved in 0.05% Tween-Tris buffer saline (TTBS), pH 7.4 for 1 h. Membranes were incubated overnight at 4°C with relevant primary antibodies (Supplementary Table S1) diluted in 1% milk/TTBS. Appropriate peroxidase-conjugated secondary antibodies were used (711-035-152 Anti-Rabbit Donkey, or 715-035-151 Anti-Mouse Donkey, both Jackson Immuno research) at a 1:7500 dilution in 1% milk/TTBS for 1 h at room temperature. ECL (Amersham) was used for visualization of immunodetected proteins on Azure Biosystems.

Immunofluorescence

Fixed embryos (15 min in 4% paraformaldehyde (PFA), Sigma–Aldrich) were permeabilized for 10 min in 0.1% Triton X-100, washed in PBS with PVA (Sigma–Aldrich), and incubated overnight at 4°C with primary antibodies (Supplementary Table S1) diluted 1:250 in PVA/PBS. Immunofluorescence analysis was performed according to a published protocol [14]. Samples were visualized using an inverted confocal microscope in 16-bit depth (TCS SP5; Leica). Images were assembled and quantified by FIJI software (version 1.8.0_172).

Measurement of overall and in situ protein synthesis

To measure total protein synthesis, 50 mCi of ³⁵S-methionine (Hartman Analytic) was added to methionine-free culture medium, for 60 min and then embryos were lysed in sodium dodecyl sulphate-buffer and subjected to SDS–PAGE. The labeled proteins were visualized by autoradiography with BasReader (FujiFilm). GAPDH was used as a loading control. For nascent protein synthesis embryos were cultured in the methionine-free medium (Gibco) supplemented with 1% dialyzed fetal bovine serum (10 000 MW; Sigma) and 50 μM homopropargylglycine (HPG) for 60 min. HPG was detected by using Click-iT Cell Reaction Kit (Life Technologies). Expression of H2B-GFP was used as a microinjection control. Samples were visualized using an inverted confocal microscope in 16-bit depth (TCS SP5; Leica). Images were assembled and quantified by FIJI software (version 1.8.0_172).

Quantification and statistical analysis

Statistical differences between pairs of datasets were analyzed by two-tailed unpaired t-tests. Values of P <.05 were considered statistically significant. Where applicable, all quantitative data are presented as the mean ± standard error of the mean. Repeated number was indicated as ‘n’ in figure legends.

Results

Fraction-resolved polysome profiling defines translational dynamics across mouse oocytes and pre-implantation development

By employing fraction-resolved polysome profiling and sequencing all 10 gradient fractions [12, 13], we analyzed translational profiles of mRNAs bound by different number of translating ribosomes in mouse oocytes at the GV and metaphase II (MII) stages, as well as pre-implantation embryos at the zygote, two-cell, four-cell, eight-cell, morula, and blastocyst stages (Fig. 1A and Supplementary Dataset S1). For each stage, 200 oocytes or embryos were utilized, and the experiment was performed with three biological replicates per stage. Unlike previous studies that considered the ribosome protected fragments as a whole, we collected continuous samples from 10 RNA–ribosome fractions [from unbound fractions (UF) to heavy fractions (HF): F1–F10] following ultracentrifugation on sucrose gradients for low-input RNA-seq analysis (see the ‘Materials and methods’ section). The different ribosome numbers (weight) can be distinguished by fraction, which encompasses all RNA types and their translation status. To systematically explore the correlation of global mRNA transcription and translation, RNA-seq analysis of total mRNA was conducted with 20 oocytes (GV and MII stages) (n = 4) and 20 embryos (n = 4) at each developmental stage collected from the same batches used for polysome profiling. Reproducibility of sequencing datasets was confirmed by PCA, showing that all replicate samples from the same biological groups in both transcriptomic and ribosome profiles consistently aligned with each other (Fig. 1B and Supplementary Figs S2 and S3). From the translatome perspective, we found that the GV, MII oocytes, and zygotes were clustered together, separating from the two-cell embryos, when the major ZGA occurs [15] (Fig. 1B and Supplementary Figs S2and S3). Subsequently, a marked shift was observed in four-cell embryos and the stages from four-cell onwards showed a continuous but steady progression up to the blastocyst stage (Fig. 1B and Supplementary Figs S2 and S3). Additionally, a clear difference between transcriptomic and translatomic profiles was evident across all developmental stages (Fig. 1B), indicating that the transcriptome data do not reflect real-time mRNA translation, as has been found in other studies [13, 16, 17].

Figure 1.

Fraction-resolved polysome profiling of mouse oocytes and pre-implantation embryos. (A) Scheme of high-resolution polysome profiling of mouse pre-implantation development. (B) PCA of translatome (F1–F10) (n = 3) and transcriptomes (n = 4) of mouse oocytes and pre-implantation embryos. (C) Distinct patterns showing the distribution of ribosome protected RNAs within 10 fractions: highly enriched in unbound fractions (UF; green), highly enriched in light fractions (LF; yellow), highly enriched in heavy fractions (HF; red). Heatmap showing the expression level of genes during the cell fate transition between GV and MII (first panel); MII and zygote (second panel); zygote to two-cell (third panel); eight-cell to blastocyst (forth panel). (D) Upset graph showing the genes in different patterns between GV and MII stages. Genes fall in UF enriched (UFe) at GV stage and HF enriched (HFe) at MII stage, as well as genes fall in LF enriched (LFe) at GV stage and HFe at MII stage were highlighted, respectively. (E) Specific GO/KEGG terms enriched from the genes highlighted in panel (D). (F) Upset graph showing the genes in different patterns between MII and zygote stages. Genes fall in UFe at MII stage and HFe at zygote stage, as well as genes fall in LFe at MII stage and HFe at zygote stage were highlighted, respectively. (G) Specific GO/KEGG terms enriched from the genes highlighted in panel (F). (H) Upset graph showing the genes in different patterns between zygote and two-cell stages. Genes fall in UFe at zygote stage and HFe at two-cell stage, as well as genes fall in LFe at zygote stage and HFe at two-cell stage were highlighted, respectively. (I) Specific GO/KEGG terms enriched from the genes highlighted in panel (H). (J) Upset graph showing the genes in different patterns between eight-cell and blastocyst stages. Genes fall in UFe at eight-cell stage and HFe at blastocyst stage, as well as genes fall in LFe at eight-cell stage and HFe at blastocyst stage were highlighted, respectively. (K) Specific GO/KEGG terms enriched from the genes highlighted in panel (J).

Fraction-resolved polysome profiling allowed us to identify five distinct patterns across pre-implantation development, which were further classified into three major categories (Fig. 1C, left panel, Supplementary Fig. S4A, and Supplementary Dataset 2): (i) mRNA highly enriched in unbound fractions (UF: F1 and F2), which represents RNA that is free from ribosome and not translated; (ii) and (iii) highly enriched in LF (F3–F5) and/or UF, which represent RNA that is stored and untranslated; (iv) and (v). broadly enriched at HF (F6–F10) and/or LF, which represent RNA that is actively translated. To simplify the analysis, fractions from each of UF, LF, and HF profiles were then combined individually for downstream analysis. Across different stages, the PCA plot indicated distinct separation between HF-bound mRNA and LF-bound mRNA at each stage, highlighting that despite both being ribosome-bound, a clear gap exists (Supplementary Fig. S4A). Notably, a significant variance between the different fractions was observed at the two-cell stage, which decreased at later stages (Supplementary Fig. S4A), reflecting the selective gene translation associated with ZGA. We also confirmed that oocyte marker genes such as Zar1, Ooep, and Mos, and trophectoderm markers such as Eomes, Krt8, and Cdx2, are specifically translated at the corresponding stages, demonstrating the consistent trends across datasets (Supplementary Fig. S4B). When comparing to a previous translatomic dataset from matched stages that profiles only translated mRNAs by using Ribo-lite-seq [11], we detected a positive correlation between Ribo-lite dataset and our HF and LF profiles among all the stages (Supplementary Fig. S4C). Although both HF and LF represent ribosome-occupied mRNAs, we only detected a positive correction between translational efficiency (TE) and HF across all stages (Supplementary Fig. S4D). Collectively, our results demonstrate the advantage of the fraction-resolved polysome profiling in obtaining high-resolution, spatial translational dynamics.

The identified three distinct patterns provided us with the detailed translational status of each single gene and the translational fate of mRNA is dedicated across stages, which allows us to investigate the fate of mRNA in crucial stage transitions during early development. By analyzing the common mRNAs fall into all three patterns, we revealed their translational status in each developmental transition. First, we found that most but not all the genes remain in the same pattern during GV to MII transition (Fig. 1C, maturation). In this transition, 227 genes switched from UFe in GV stage to HFe, and 47 genes from LFe to HFe (Fig. 1D). These stored mRNAs became actively translated after oocyte maturation, implying their important roles during this process. GO enrichment analysis of these genes showed that they are involved in key events including meiotic cell cycle, translation, and RNA splicing (Fig. 1E). Second, during fertilization, translation of 93 genes (UFe–HFe: 8, LFe–HFe: 85) became active and they were involved in translation, meiotic cell cycle, and cellular component disassembly (Fig. 1F and G). Third, 86 genes (UFe–HFe: 66, LFe–HFe: 20) were identified to be switched from inactive to active translation during ZGA, with essential cellular functions including translation, mRNA export from nucleus, and protein location in nucleus (Fig. 1H and I). Finally, during lineage separation from eight-cell to blastocyst stage, 166 genes (UFe–HFe: 48, LFe–HFe: 118) were actively translated and involved in regulating metabolic process and cell polarity (Fig. 1J and K).

To validate the genes in these defined patterns and their functions for embryo development, we selected several candidate genes (Ncl,Ezh2, andSpic) for western blotting analysis, and demonstrated that their protein abundances are consistent with their shifts in translational status during these stage transitions (Supplementary Fig. S5A–C). To further explore stage-specific polysome activation, we focused on well-known candidates Calm1 and Calm2, which encode calmodulin, a key regulator of intracellular Ca²⁺ levels and essential for embryonic development [18, 19]. Both genes were enriched in HF at both two-cell and morula/blastocyst stages (Supplementary Fig. S5D), suggesting their involvement in ZGA and lineage specification. It has been reported that calmodulin antagonist W-7 has an inhibitory effect on the first cleavage during mouse preimplantation development [20]. To overcome the arrest of first cleavage, we cultured embryos with W-7 after first cleavage and maintained them until the end of culture (Supplementary Fig. S5E). We found that the developmental rate is not affected until morula stage, but the morula rate is significantly decreased in a dose dependent manner, and none of them could proceed to blastocyst at both concentrations (5 and 10 μM) (Supplementary Fig. S5F and G), highlighting the critical role of Calm1 and Calm2 translation in regulating first lineage specification. These results validated our data analysis and demonstrated the significant value of this unprecedented resource for prioritizing genes for further characterization during the critical stage of pre-implantation development.

Translational fate of mRNAs in pre-implantation development

Because RNAs enriched in adjacent polysome fractions often reflect distinct translational states, we next focused on characterizing all mRNAs that bound by ribosomes across the developmental stages. By applying the fuzzy k-means algorithm on all mRNAs bounded to LF and HF separately, we identified 12 subclusters of genes with stage specific patterns in both HF- and LF-occupied RNA profiles (Fig. 2A and B and Supplementary Dataset 3), offering a clearer insight into the activation and suppression of genes as development advances. TE analysis showed a similar trend as the polysome cluster across stages (Supplementary Fig. S6), implying the positive correlation between HF and TE. While most genes exhibited stage-specific expression, many LF-bound genes either degraded from HF-bound or transitioned to HF-bound to support immediate protein synthesis for specific stages (Fig. 2C). This variation illustrates how mRNAs are regulated based on developmental needs. GO analysis of stage-specific genes revealed the dominant biological functions for each RNA fraction occupation (Supplementary Fig. S7A), showing both consistent and inconsistent functions across the different fractions. For example, microtubule-based movement related genes that are associated with the LF at the GV and MII oocytes shifted to two-cell in the HF. Similarly, ribosome biogenesis and mRNA splicing genes that are prominent in the LF profiles at the four- and eight-cells, were more prominent in the HF at the morulae, mainly due to the inconsistency between the LF and HF profiles.

Figure 2.

The dynamics of mRNA translational fate during mouse pre-implantation development. (A, B) Fuzzy c-means clustering identified 12 distinct temporal patterns of HF-bound mRNA (A) and LF-bound mRNA (B). The ‘x’ axis represents developmental stages in time course, while the ‘y’ axis represents log₂-transformed, normalized intensity ratios in each stage. (C) The heatmap showing the number of genes overlapped from the identified stage-specific clusters between heavy- and light-fraction layers based on Fig. 2A and B. The color scheme, from white to green, indicates the number of intersection genes from low to high. The genes intersected between HF peaking at two-cell stage (H_Cluster7) and LF peaking at zygote (L_Cluster3)/two-cell (L_Cluster5)/four-cell (L_Cluster7) were highlighted and further explored. (D) Heatmap showing the expression levels of representative genes (Rep. genes) across the HF, LF, and transcriptome profiles. (E) An illustration graph summarizing three different mRNA fates during early embryo development. Example given here is focusing on two-cell stage.

Given that ZGA triggers extensive gene activation and translation, we specifically focused on the mRNA translational fate in the two-cell embryos. We identified that genes in HF-bound (Clusters 5, 6, and 7) are actively translated at the two-cell stage (Fig. 2A and Supplementary Fig. S7B), suggesting their critical function for the ZGA. We also noticed three distinct patterns: (i) 30 genes are intersected between HF_Cluster 7 and LF_Cluster 3 (Fig. 2C, highlighted in top dashed box), representing mRNAs that are initially not translationally active, becoming highly active during ZGA and subsequently returning to a quiescent state (Fig. 2D, top panel, 10 representative genes shown); (ii) 266 genes are intersected between HF_Cluster 7 and LF_Cluster 5 (Fig. 2C, highlighted in middle dashed box), representing mRNAs that remain low in abundance and are rarely loaded into ribosome until a spike in transcription and translation in the two-cell stage (Fig. 2D, middle panel, 10 representative genes shown); and (iii) 47 genes are intersected between HF_Cluster 7 and LF_Cluster 7 (Fig. 2C, highlighted in bottom dashed box), representing mRNAs that are actively translated in the two-cells, then transition to LF-bound status afterwards, either being degraded or reserved for later stages (Fig. 2D, bottom panel, 10 representative genes shown). These patterns demonstrate the translational fate of mRNA during linear progression of early embryonic development and emphasize the strategic regulation of mRNA storage and translation (Fig. 2E), which is critical for the subsequent cellular processes in pre-implantation development.

We also extended our analysis across all stages to investigate the translational fate of mRNAs, and observed widespread presence of the delayed translational activation (for a given mRNA species, its active translation occurs one or more developmental stages later than its binding to LF). We were able to categorize genes into the following three groups (Supplementary Fig. S8and Supplementary Dataset 3): (i) those overlapping between LF and HF occupied RNA at the same stage (red), (ii) those HF -occupied one stage after being LF-occupied (blue), and (iii) those HF-occupied two stages after LF occupancy (green). We also explored the biological functions of genes in these groups (Supplementary Fig. S8). For example, the overlapping genes at the same stages regulate essential biological processes sequentially, such as meiosis in oocytes, mitotic cell cycle, and transcription in two-cells, ribosomal large subunit biogenesis in four-cell, and stem cell differentiation in blastocysts. On the other hand, the genes with delayed translational activation suggested their involvement in crucial later embryonic processes. These include mRNA stability regulation in zygotes, mitotic cell cycle phase transition in two-cell, RNA and histone modifications in morula, and Rho protein signal transduction in blastocyst.

Collectively, our analysis suggests that during pre-implantation development, discrepancies arise not only between the transcriptome and the translatome, but also within the translatome itself, as mRNAs selectively bind to LF or HF. Oocytes and embryos regulate mRNA translational through a dynamic process. Some mRNAs are stored along with light ribosomes for later activation, while others can be continuously translated by binding to both mono- and poly-ribosomes at the same stage or being stored with light-ribosomes after early activation. Our fraction-resolved polysome profiling enables us to track mRNA translational dynamics and offers deeper insights into translational regulation during early development.

Comparative analysis of translatome, proteomics, RNA m⁶A modifications, and mRNA features, during pre-implantation development

It’s well known that the poly(A) tail length and m⁶A are key regulators for translation [21–23]. To further explore translational regulation during pre-implantation development, we integrated our translatomic dataset with the publicly available datasets of poly(A) tail length [24], RNA m⁶A modification [25], and proteomics [7, 26] at the matched developmental stages (Fig. 3A–D, Supplementary Fig. S9A–D, and Supplementary Dataset 4). Overall, we observed that the genes from UFe, LFe, and HFe groups exhibited low to high of average protein abundances, across MII to eight-cells (Fig. 3E), suggesting the protein abundances directly correlate with their gene translational status during maternal to zygotic transition, and ZGA where massive transcription and translation occur. In both GV and blastocyst, higher protein abundances were seen in both UFe and HFe compared to LFe (Fig. 3E). Similarly, an increased m⁶A density from UFe to HFe genes was observed throughout the pre-implantation development except for GV and blastocyst stage (Fig. 3F). In blastocyst stage, the highest m⁶A density were seen in untranslated genes in UFe group (Supplementary Fig. S9D). This unique trend may be attributed to the combined analysis of the two distinct lineages, trophectoderm and inner cell mass, presenting in blastocyst-stage embryos. Interestingly, an increased trend of poly(A) tail length was evidenced in genes enriched in HFe while decreased in genes enriched in LFe during developmental progression from GV to two-cell stage (Fig. 3G).

Figure 3.

Translational and epigenetic regulation in mouse oocytes and early embryos. Heat maps showing expression level of ribosome occupied genes within UFe (green), LFe (yellow), and HFe (red) (panel one), as well as their protein level (second panel), poly(A) length (third panel), and m6A modification (forth panel) at GV (A), MII (B), zygote (C), and two-cell (D) stages. The clusters of genes with specific features were highlighted and specific GO terms with genes were listed below. (E) Bar plot showing the average protein level in distinct patterns at stages from GV till blastocyst stage. (F) Bar plot showing the m⁶A modification rate in distinct patterns at stages from GV till blastocyst stage. (G) Bar plot showing the average poly(A) length in distinct patterns at stages from GV till two-cell stage. (H) Scatter plot comparing the HF-occupied mRNA data and protein expression level at each developmental stage. (I) Scatter plot comparing the HF-occupied mRNA data (four-/eight-cell) and protein expression level (morula). (J) Heatmap showing the correlation between HF-occupied mRNA level and protein level across developmental stages. The color scheme, from white to red, indicates the correlation score from low to high. (K) Heatmap showing the correlation between HF-occupied mRNA level and poly(A) tail length across developmental stages. The color scheme, from white to blue, indicates the correlation score from low to high. (L) Scheme of the delayed correlation between HF-occupied mRNA profile and proteomic date, RNA m⁶A data, and poly(A) tail length profiles. Boxplots showing the length of 3′UTR (M) and 5′UTR (N) in distinct patterns across mouse oocyte and preimplantation development. Data are presented as the mean ± standard deviation (SD). ∗P<.05, ∗∗P<.01, ∗∗∗P<.001, ∗∗∗∗P<.0001 from one-way analysis of variance followed by Tukey’s multiple comparisons test.

We also observed specific groups of genes following interesting patterns. For example, in the GV stage, one gene cluster in untranslated UFe group had significantly longer poly(A) tail length and lower m⁶A density; however, the protein level is not distinct from other genes within the group (highlighted in Fig. 3A). This inconsistency might reflect that they are maternal stored but not newly translated proteins that playing essential roles at GV stage. From MII to two-cells, a cluster of genes from highly translated HFe group had enriched protein level and significantly longer poly(A) tail (highlighted in Fig. 3B–D). Analysis of the functions of genes revealed that they are involved in mRNA splicing, translation, negative regulation of cell cycle in both MII and zygotes (Fig. 3B and C), regulating transcription and positive regulation of cell cycle in two-cells (Fig. 3B and C). This analysis suggests that prior to ZGA, for specific group of genes, the length of poly(A) mRNA is positively correlated with heavy ribosome occupancy and, consequently, with protein abundance-indicating that poly(A) tail length plays a regulatory role in key cellular functions during specific developmental stages.

Although the HFe genes have the highest average protein level compared to LF and UF as expected (Fig. 3E), surprisingly, we found it is temporal disconnected from protein output across pre-implantation development (Fig. 3H). Specifically, we observed that HF occupancy and protein expression were not positively correlated at the four-cell and eight-cell stages (Fig. 3H). We detected a delayed correlation between polysome levels at the four-/eight-cell stages and protein levels at the morula/blastocyst stages (Fig. 3I and J), suggesting that during the second and third rounds of cell division, protein maturation is delayed until embryo compaction. Additionally, by integrating datasets of poly(A) tail length [24] and actively translation in HF profile, we found a delayed positive correlation between poly(A) tail length at the two-cell stage and polysome levels at the four-/eight-cell stages (Fig. 3K). This implies that embryos prepare for translation at later stages through re-adenylation during ZGA [24]. Collectively, these results demonstrate that embryos optimize timing and energy by preparing mRNA for rapid translation and functional activation through re-adenylation, m⁶A modification, and ribosome binding (Fig. 3L).

The length of 3′UTR and 5′UTR also play critical roles in translation regulation [27, 28]. We observed an increase of 3′UTR length in genes from LFe and HFe groups before (GV to zygote) and after (four-cell to blastocyst) ZGA stage, while an opposite trend in transcriptionally active two-cell stage (Fig. 3M). There is no clear trend of 5′UTR length across the stages, indicating it is not essential for translation regulation during mouse early development (Fig. 3N).

Two-cell specific translation of Eif1ad3 and its variants are essential for embryo development

Our fraction-resolved polysome profiling allows the identification of genes that are translationally activated in the specific developmental stages, underlying their critical roles in early development. We identified a group of genes including Eif1ad2, Eif1ad3, Eif1ad4, Eif1ad6, Eif1ad7, Eif1ad8, Eif1ad16, and Eif1ad19, to be exclusively translated at the major ZGA stage while their roles in embryogenesis remain uncharacterized (Fig. 4A). These genes, in particular Eif1ad2, Eif1ad3, Eif1ad4, Eif1ad6, Eif1ad7, and Eif1ad8, have highly conserved sequences with only few nucleotide mismatches, but differ from well-studied translation initiation factors Eif1 and Eif1ad (Fig. 4B and Supplementary Fig. S10A). We therefore focused on investigating the functions of Eif1ad3 and its variants in mouse pre-implantation development.

Figure 4.

Function of Eif1ad3 for mouse embryo development. (A) Heatmap showing the expression level of Eif1 family genes. The color spectrum, ranging from red through white to blue, indicates high to low levels of gene expression. (B) Multiple sequence alignment (Omega) of EIF1AD amino acid sequence. (C) Experimental scheme of Eif1ad3 knock down experiment. (D) A representative image of mouse embryos from control and KD groups. (E) Four-cell developmental rate of Eif1ad3 KD embryos compared to control group. (F) Volcano plots showing the number of upregulated or downregulated genes in control compared to Eif1ad3 KD at two-cell stage, the top GO terms of upregulated and downregulated genes are shown respectively. (G) Venn diagram showing the number of genes and the top representative GO terms that are overlapped between control enriched and two-cell specific genes from HF profile in Fig. 2A. (H) Venn diagram showing the number of genes and the top representative GO terms that are overlapped between control enriched and four-cell specific genes from HF profiles in Fig. 2A. (I) Rank the DEGs according to the log₂ fold change. Each circle represents a gene, and the highlighted circle represents the Rpl and Eif genes separately. The genes are annotated with detailed information listed in Supplementary Dataset 5. (J) Representative confocal images of two-cell embryos with incorporation of methionine analog (HPG). From three biological replicates, n ≥ 21; HPG, gray and red; H2B-GFP, green; 4′,6-diamidino-2-phenylindole (DAPI), blue; scale bar, 20 μm. NTC, negative technical control where HPG was omitted. (K) Quantification of HPG signal from panel (J). Values are presented as mean ± SD; Student’s t-test: **P <.01. (L) Analysis of global de novo proteosynthesis by incorporation of ³⁵S-methionine in embryos with downregulation of Eif1ad3. Data from four biological replicates. GAPDH was used as a loading control. (M) Quantification of ³⁵S-methionine incorporation from panel (L). Values from control KD were set as 100%. Data presented as mean ± SD; Student’s t-test: *P <.01.

Due to the high sequence similarity among these genes, we employed the CRISPR/RfxCas13d system to achieve KD of Eif1ad3 and its variants (Fig. 4C). We found that almost all embryos in the KD group were arrested at the two-cell stage (Fig. 4D and E), suggesting they are critical for ZGA. RNA-seq analysis of individual embryos (both KD and control) at the two-cell stage observed a total of 2642 genes downregulated and 3828 genes upregulated in KD groups compared to control (Fig. 4F and Supplementary Dataset 5). The downregulated genes mainly related to ribosome biogenesis, RNA processing and cell division, which are essential for embryo activation during ZGA, while upregulated genes were associated with mitochondrial function, cellular respiration and metabolite formation, indicating apoptosis and embryo death (Fig. 4F).

During ZGA, new mRNAs are synthesized and at the same time translation is reprogrammed via ribogenesis [10]. To further assess the importance of the Eif1ad3 regulated genes at two-cell stage, we compared stage-specific translated genes from Fig. 2B (H_Cluster 5, 6, 7) with the control-enriched genes, we identified 316 common genes, and they were involved in developmental growth, cell fate commitment, and cell maturation (Fig. 4G). Interestingly, the top GO terms such as ribosome biogenesis and ribosomal RNA (rRNA) processing that enriched in control groups were found in the overlapped genes specific in four-cell stage instead of two-cell stage (Fig. 4H and Supplementary Fig. S10B). Further analysis of RNA-seq data revealed that the expression of ribosomal genes, as well as translational initiation and elongation factors, was significantly disrupted in Eif1ad3 KD embryos, further confirming its role in translational regulation (Fig 4I and Supplementary Dataset 5).

To characterize the phenotype resulting from disrupted ribogenesis and altered expression of translational factors, we employed in situ detection of protein synthesis using a methionine analog HPG (Fig. 4J and K) and performed a methionine incorporation assay (Fig. 4L and M) to visualize and quantify protein synthesis in two-cell embryos. Both assays revealed a dramatic reduction in protein synthesis in Eif1ad3 KD embryos, indicating an impairment of the translational machinery, highlighting the essential role of Eif1ad3 expression during early embryogenesis.

Eif1ad3 regulates genes involved in ribosome assembly and ZGA

Eukaryotic initiation factors (eIFs) play a crucial role in initiation of translation [29, 30]. To investigate the regulation of mRNA targets mediated by the specific expression of Eif1ad3 in two-cell embryos, we microinjected in vitro transcribed His-tagged Eif1ad3 mRNA (Eif1ad3-His IVT-mRNA, Fig. 5A, B) into zygotes, given the absence of a specific Eif1ad3 antibody. Notably, the expression of Eif1ad3-His did not affect development up to the blastocyst stage (Fig. 5C and D). We then performed RNA immunoprecipitation sequencing (RIP-seq) on two-cell stage embryos using anti-His antibody and compared the results between the Eif1ad3-His group and a control IgG group (Supplementary Dataset 6). PCA results showed a clear separation between groups, except for IgG, likely due to nonspecific interactions (Supplementary Fig. S10C). After eliminating nonspecific interactions, we identified 3708 genes physically associated with Eif1ad3, indicating its broad involvement in initiating mRNA translation during ZGA (Fig. 5E). These genes were primarily involved in processes such as Golgi vesicle transport, RNA processing, RNA modification, and protein maturation, all closely related to translation (Fig. 5F). Importantly, a significant number of mRNAs coding for ribosomal protein were found to be bound to Eif1ad3, suggesting its role in ribosome assembly and protein synthesis (Fig. 5G). To further assess the importance of the Eif1ad3 regulated gene at two-cell stage, we narrowed down genes that are enriched in the two-cell stage from HF profiles (Fig. 2B, Clusters 5, 6, 7) to 727 genes (Fig. 5H). These genes regulate DNA-binding TFs activity and affect BMP and Ras signaling pathways (Fig. 5I). Since TFs are crucial for genome activation across mammals [31], we analyzed the TFs within these 727 genes (Fig. 5J). These TFs were highly involved in ZGA, regulating BMP signaling, cell proliferation, and cell fate commitment (Supplementary Fig. S10D).

Figure 5.

Eif1ad3 regulates translation of functional genes during ZGA. (A) Experimental scheme of Eif1ad3-His overexpression experiment. (B) Representative protein levels of His after injecting Eif1ad3-His mRNA at different concentrations into mouse zygotes. (C) Representative image of mouse blastocysts from ctrl and Eif1ad3-His groups. (D) Quantification of blastocyst rates of embryos with Eif1ad3-His compared to ctrl. (E) Volcano plots showing the number of upregulated or downregulated genes in RIP-His compared to RIP-IgG group at two-cell stage. (F) Main biological functions regulated by RIP-His group enriched genes. (G) Heatmap in HF profile, LF profile, and transcriptome profile showing the expression level of genes regulating ribosome biogenesis. (H) Venn diagram showing the number His-enriched genes from RIP-seq compared to two-cell peaked genes from HF profile explored in in Fig. 2A. (I) Representative GO terms enriched from overlapped genes shown in Fig. 2H, as well as the specific genes in each term. (J) Heatmap showing the expression level of transcription factors within overlapped genes shown in Fig. 2H.

In summary, Eif1ad3 is essential for ZGA during mouse pre-implantation development by regulating the translation of mRNAs encoding factors that maintain developmental competence.

Discussion

The mRNA translation landscape and particularly the translational control operating on specific mRNAs during the rapid period of pre-implantation development is poorly understood. Ribo-seq has been used to characterize the translational landscapes of mouse oocytes and pre-implantation embryos [10, 11, 32]. However, these analyses preclude exploration of the spatial landscape of the dynamic mRNA translation process due to the lack of single-fraction resolution of ribosome-occupied RNAs. Here, by utilizing a low-input, high-resolution, optimized SSP protocol [12] integrated with RNA sequencing (RNA-seq), we were able to overcome this critical limitation. Instead of exploring ribosome-occupied RNA as a pool, we analyzed 10 individual fraction profiles based on the number of bounding ribosomes. It not only significantly increased the depth of the sequencing, but also enabled us to track the dynamics of translational fate for mRNAs. The spatial map of mRNA translational fate represents another level of translational regulation and is critical for understanding the mechanisms and strategies that oocyte and embryo employ to rapidly respond to the demands arising from successive but distinct stages during pre-implantation development. Furthermore, by comparing mRNA profiles associated with heavy and light ribosomal fractions, we identified key translationally activated genes that play essential roles in oocyte maturation, the oocyte-to-embryo transition, embryonic genome activation, and lineage specification.

Our study captured diverse, albeit partially empirical, modes of translational selectivity for individual transcripts. The results suggest that the three major mRNA fates observed in oocytes and early embryos reflect strategic mechanisms employed by the embryo to optimize energy efficiency and minimize the time required for transcript re-synthesis. These mechanisms likely meet the specific translational demands of oocyte and pre-implantation embryo development. Maternal mRNAs are known to be important for pre-implantation development [31]. However, it remains unclear whether translation of specific maternally stored mRNAs is essential for ZGA or later stages. Here, by analyzing the genes from HF and LF profiles, we identified a group of maternal RNAs that bind to the monosome at the zygote stage but exclusively to the polysome at the two-cell stage (Fig. 2F top panel), undoubtedly indicating their specific functions during ZGA. Furthermore, this database enabled us to reveal the fate of each maternal mRNA across the different developmental stages: It is either degraded without ribosome binding, stored by LF binding, or translated upon HF occupancy. A comparative analysis of datasets from oocytes and pre-implantation embryos occupied with UF, LF, and HF showed dynamic changes in the stage-specific activated translatome landscape. Interestingly, only a small fraction of HF-enriched RNAs were found to maintain actively translated across all stages, especially after the zygote stage. Most of the HF-enriched RNAs are activated in a stage-specific manner by groups with significant and nonsignificant levels of LF enrichment, reflecting precise regulation for the development.

The length of poly(A) tails, which varies widely across different mRNAs, has a significant impact on both their stability and translational activity [33, 34], therefore plays a crucial role in coordinating the translation during maternal-zygotic transition. Additionally, polyadenylation, occurring at different stages, helps fine-tune gene expression by selectively stabilizing certain transcripts while allowing others to degrade [21, 35]. Our study highlighted a notable delayed positive correlation between poly(A) tail length at the two-cell stage and polysome association at the four-/eight-cell stages. This suggests that while poly(A) tails are elongated at the two-cell stage, their functional impact on translation becomes more evident in subsequent stages of development. Furthermore, this delay could indicate a temporal control mechanism, where poly(A) tail lengthening prepares transcripts for translation only after key developmental checkpoints are passed, ensuring precise expression in the highly dynamic activation of specific mRNAs. This observation aligns with other studies that highlight extensive poly(A) tail modification as a form of translational control that governs oocyte meiosis and embryonic development, and establishes the foundation for cell lineage decisions [36, 37]. Interestingly, our data revealed no significant differences in 5′UTR length between ribosomal fractions, suggesting that 5′UTR length is not a key determinant of translational regulation in oocytes and early embryos. Instead, these findings underscore the importance of specific cis-regulatory motifs within 5′UTRs as drivers of translational control during developmental stages. On the other hand, we observed a positive correlation between 3′UTR length and translation efficiency, as reflected by the enrichment of transcripts in the heavy ribosomal fractions in oocytes and zygotes. In zebrafish, 3′UTRs shorten globally after fertilization [38], suggesting that longer 3′UTRs may harbor multiple binding sites for RNA-binding proteins [39] or microRNAs [38, 40]. This allows for more intricate regulation of maternal mRNAs and facilitates their storage in a quiescent state, active translation, stabilization, or degradation within the transcriptionally silent stages.

We have identified a key eukaryotic initiation factor, Eif1ad3, which is translated exclusively at the two-cell stage and plays an essential role in ZGA. eIFs are fundamental to the translation initiation process and act as primary regulators of gene expression patterns during development. While it has been found a number of genes in the eIF family play essential roles in the pre-implantation development, including eIF1A [41], eIF3 [42], eIF4E [30, 38], and eIF4E1b [43], we extended this list by providing the essential role of other eIF factors, Eif1ad3, and its variants, in the mammalian pre-implantation development. Interestingly, Eif1ad3 exhibits a translation pattern distinct from Eif1a and Eif1ad and is specifically translated at the two-cell stage. Deficiency of Eif1ad3 leads to embryonic arrest at the two-cell stage. Surprisingly, the large number of genes affected by Eif1ad3 KD regulates ribosome biogenesis and associated rRNA synthesis. These genes are only marginally transcribed but are extensively translated from the four-cell stage, which is consistent with the transition from the maternal to the embryonic stage. Consequently, the two-cell arrest due to a downregulated protein synthesis required for initiation of ZGA, highlighting the need for further investigation.

In summary, our study reveals a previously underappreciated level of translational regulation in mouse oocytes and early embryos and provides comprehensive data sets on the translational fates of mRNA during development. We show that both maternal and newly synthesized mRNAs follow several specific trajectories that allow optimizing energy use and navigating efficiently through developmental transitions, which is potentially correlated with epigenetic regulators including poly(A), m⁶A, and UTR. In particular, stage-specific translation of Eif1ad3 and its variants at the two-cell stage emerges as a critical factor for embryogenesis, highlighting the value of these unique resources for the identification of additional candidates or mechanisms in future studies.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

Funding to pay the Open Access publication charges for this article was provided by Institutional Research Concept (RVO67985904) and NIH Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD102533). This work was supported by the NIH Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD102533 to Z.J), Institutional Research Concept (RVO67985904), The Czech Science Foundation (22-27301S to M.D. and A.S.; PPLZ-L200452502 to M.D.; 23-07532S and 25-18241S to K.K.), Research programme Strategy AV21 FUTURE OF ASSISTED REPRODUCTION (ART), and The Ministry of Education, Youth and Sports (EXCELLENCECZ.02.1.01/0.0/0.0/15_003/0000460 OP RDE).

Data availability

The raw FASTQ files and normalized read accounts per gene are available at Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) under the accession numbers GSE263902 and GSE279822. The previously published datasets including proteomics (PXD003315, PXD018777), polyA (GSE228001) and m⁶A modification (GSE192440) were downloaded from ProteomeXchange Consortium and SRA database, respectively, and re-analysed in this study. The mouse reference genome GRCm39 (Release 113) was downloaded from the Ensembl (https://ftp.ensembl.org/pub/release-113/fasta/mus_musculus/).