KAS-seq profiling captures transcription dynamics during oocyte maturation

All animal experiments adhered to the regulations and protocols established by both the local animal ethics committee and the Animal Care and Use Committee of Nanjing Medical University (ethics approval no. 2110009). The research protocol received approval from the Animal Care and Use Committee of Nanjing Medical University. The mice were accommodated in cages maintained under specific pathogen-free conditions and were provided unrestricted access to both water and standard rodent food. Female ICR mice, aged three weeks, were procured from Charles River Laboratories, China Inc.

To acquire Fully Grown Oocytes (FGOs), mice aged 3–4 weeks were subjected to intraperitoneal injections of pregnant mare's serum gonadotropin (PMSG; 7 IU), followed by human chorionic gonadotropin (hCG; 7 IU) administration 46–48 h later. Germinal vesicle (GV) stage oocytes (> 70 μm) were collected 48 h after PMSG treatment. Germinal vesicle breakdown (GVBD) and metaphase II (MII) stage oocytes were collected 3 h and 13 h post hCG injection, respectively. To distinguish between Surrounded Nucleus (SN) and Not Surrounded Nucleus (NSN) oocytes without the need for DNA staining, GV oocytes were cultured in vitro with IBMX (3-isobutyl-1-methylxanthine) for 1 h. Subsequently, GV oocytes were classified into two groups based on the presence or absence of a visible Perivitelline Space (PVS). Oocytes with PVS were designated as SN oocytes, while those lacking PVS were categorized as NSN oocytes. For MII oocytes, the zona pellucida and first polar body was carefully removed through treatment with Pronase E (Solarbio P8360).

Oocytes were labeled with N3-Kethxal, and labeled genomic DNA was subsequently extracted using the Quick-DNA Microprep Plus Kit (Zymo D4074).After labeling, gDNA was fragmented using Tn5 transposase (Vazyme TD501) under incubation conditions of 37 °C and 500 rpm for 30 min. Following fragmentation, biotinylation labeling was performed using DBCO-PEG4-biotin (DMSO solution, Sigma 760,749). The biotinylated DNA was then cleaned up using DNA Clean & Concentrator-5 kit (Zymo, D4013), and 5 µL of labeled DNA was retained as input, while the remaining 50 µL of DNA was used for enrichment. For bead-based enrichment, 1 × Binding and Wash buffer was used to wash 5 µL of Dynabeads™ MyOne™ Streptavidin C1 (Thermo 65,001) three times. Pre-washed beads were then resuspended in 2 × Binding and Wash buffer. The beads were mixed with the 50 µL fragmented DNA obtained from the previous steps and rotated slowly at room temperature for 15 min. After incubation, the mixture was placed on a magnetic rack to remove the supernatant, and the beads were washed five times with 1 × Binding and Wash buffer. DNA-conjugated beads and their corresponding inputs were used for library PCR using i5 and i7 index primers (Vazyme TD205) and VAHTS HiFi Amplification Mix (Vazyme N616). The PCR reactions were initiated with a 5 min incubation at 72 °C, followed by 10 min at 95 °C, and then amplified for 14 cycles (10 s at 98 °C, 30 s at 60 °C, 1 min at 72 °C). Finally, the libraries were cleaned up by using DNA Clean & Concentrator-5 kit (Zymo, D4013).

Paired-end reads of 150-bp were generated on the Illumina novaseq 6000 platform in this study (sequenced by Annoroad). For KAS-seq data, raw reads were trimmed using trim galore (v0.6.10) (parameters: –quality 20 –phred33 –length 30 –stringency 3 –paired) and mapped to mouse reference genome (GRcm38) using Bowtie2 (v2.5.1) (parameters: -X 2000 -I 10). All unmapped reads, non-uniquely mapped reads and PCR duplicates were removed. The signal tracks and heatmap were both generated by deeptools (v3.5.2). The MACS2 (macs2 2.2.7.1) were used for peak calling (parameters: –broad). HOMER (v4.10) was used for motif enrichment analysis. Annotation of peaks was performed using Staccseeker (v1.36.0).

The mRNA levels of Padi6, Gdf9, Bmp15, and Zp2 were assessed through quantitative real-time polymerase chain reaction (qRT-PCR) employing the comparative CT method, as previously outlined in our research. Total cellular RNA was extracted from a pool of 50 oocytes from the specified experimental group, employing the RNAprep Pure Micro Kit (TIANGEN DP420). Subsequently, first-strand cDNA was synthesized following the manufacturer's protocols using the HiScript III RT SuperMix for qPCR (Vazyme R323).Real-time PCR assays were conducted with SYBR Green in a final reaction volume of 10 μL, utilizing an ABI QuantStudio™7 Flex PCR system (Applied Biosystems). GAPDH served as the internal reference control for all samples, and each experiment was carried out in triplicate. The primer sequences are provided bellow:

The EU stock solution was diluted to 100 mM with ddH2O. Subsequently, the EU solution (100 mM) was diluted 1:100 with M2, and SN/NSN oocytes were added, followed by incubation at 37 °C for 45 min. The oocytes were then placed in 4% PFA at 4 °C overnight. Subsequently, the oocytes were permeabilized by incubation in 0.5% Triton X-100 for 15 min. Afterward, they were washed three times with PBST, with each wash lasting 3–5 min. Next, the oocytes were incubated for 30–60 min at room temperature with the Click-iT reaction cocktail (Click-iT™ RNA Alexa Fluor™ 488 Imaging Assay Kit, ThermoFisher C10329). Following this, they were washed once with a raise buffer for 3–5 min and then washed once with a wash buffer for 3–5 min. DNA staining was performed using Hoechst 33,342 at a dilution of 1:1000 in PBST for 15 min. Finally, the oocytes were washed three times with a wash buffer, with each wash lasting 3–5 min.

Oocytes were added into 1 μL of Sample Buffer (prepared by adding 2 μL of RNase Inhibitor (Vazyme R301-01) to 18 μL of Lysis Buffer containing 0.2% (vol/vol) Triton X-100). Care was taken to ensure that the liquid volume in the micropipette was approximately 0.5–1 μL, and then it was supplemented with Nuclease-free H2O to a total volume of 3.5 μL. Subsequently, the cDNA synthesis reaction was prepared as follows: the above sample (3.5 μL) was mixed with 1 μL of Oligo(dT) VN Primer (1 μM) and 1 μL of dNTP Mix (Thermo R0912). The reaction mixture was incubated at 72 °C for 3 min and immediately placed on ice. Next, the reverse transcription reaction was prepared as follows: the above product (5.5 μL) was mixed with 2 μL of First Strand Buffer (Vazyme R721-01), 0.5 μL of DTT (Vazyme R721-01), 0.5 μL of RNase Inhibitor, and 1 μL of Sc Reverse Transcriptase (Vazyme R721-01). The reaction was carried out in a PCR machine with the following program: 42 °C for 90 min; 50 °C for 2 min, 42 °C for 2 min (return to the second step, repeat for 2–11 cycles); 70 °C for 15 min; and finally, hold at 4 °C.Subsequently, the cDNA amplification reaction was prepared as follows: 10 μL of the first-strand cDNA synthesis product was mixed with 2 μL of Nuclease-free H2O, 0.5 μL of ISPCR Primer (5 μM), and 12.5 μL of 2 × Amplification Mix (Vazyme N616-01). The reaction was carried out in a PCR machine with the following program: 98 °C for 1 min; 98 °C for 10 s, 65 °C for 15 s, 72 °C for 6 min (return to the second step, repeat for 15 cycles); 72 °C for 5 min; and finally, hold at 4 °C.The cDNA amplification products were purified using VAHTS DNA Clean Beads (Vazyme N411). Subsequently, quality control of the cDNA amplification products was performed, including measuring cDNA concentration using Qubit 4 and assessing cDNA fragment distribution using Qsep 400.Following this, sequencing libraries were prepared using the TruePrep DNA Library Prep Kit V2 for Illumina (Vazyme TD503). The sequencing libraries were then purified using VAHTS DNA Clean Beads (Vazyme N411). And then, quality control of the sequencing libraries was performed, including measuring library concentration using Qubit 4 and assessing library fragment distribution using Qsep 400.

For Pol2 and ATAC-seq data, raw reads were trimmed using trim_galore and mapped to mouse reference genome (GRcm38) using Bowtie2 (v2.5.1) [20] (parameters: -X 2000 -I 10). All unmapped reads, non-uniquely mapped reads and PCR duplicates were removed. The signal tracks and heatmap were both generated by deeptools (v3.5.2) [21]. The peaks were identified by using MACS2 [22].

Statistical data were presented as the mean ± SD in bar charts. For boxplots, middle lines indicated the median, the boxes indicated the 25th/75th percentiles. KAS-seq and RNA-seq was repeated twice. qPCR and immunofluorescence experiments were repeated at least three times, and P values were determined by two-sided Student’s t-test using GraphPad Prism (version 8.0.1). Significant differences were shown with “***” for indicating P < 0.001. “#” denotes no significance.

Azide-tagged kethoxal (N3-kethoxal), a modified kethoxal [23], retains not only its elevated reactivity and specificity towards guanine residues within single-stranded nucleic acids, but also provides a bio-orthogonal handle that can be easily functionalized with biotin (Fig. 1). Using the optimized KAS-seq assay, we systematically performed the library construction and ssDNA analysis of GV, GVBD, and MII oocytes (Fig. 2A, Supplementary Table 1). The experiment was conducted in two replicates with 300 oocytes per sample, a remarkable reduction in sample quantity compared to previous study [17]. KAS-seq exhibits robustness and reproducibility, demonstrating a notable enrichment efficiency and strong correlation, as well as a substantial overlap of peaks across replicates (Figs. S1, S2 and Fig. 2B). As expected, the ssDNA signal is enriched in transcriptionally active NSN oocytes, with over 57,000 peaks detected. With the transition in chromatin configuration, there is a sharp decline in ssDNA signals in SN-configured oocytes. Simultaneously, the ssDNA signal further attenuates accompanying with meiotic resumption (Fig. 2C). We particularly observed that in MII oocytes, only six peaks were detected (Fig. S3A), which represent extremely small ssDNA regions relative to the entire mouse genome. In NSN oocytes, KAS-seq reads are considerably enriched at gene-coding regions, especially at promoter regions and transcriptional termination areas. However, the six peaks in MII oocytes are primarily located within promoters, introns, and exons, with no distribution in transcription termination regions (Fig. 2D). These results suggest that the transcriptional activity in MII oocytes is exceedingly low. In NSN oocytes, KAS-seq profiles of gene-coding regions unveiled distinct features, including a prominent and sharp peak proximal to the transcription start site (TSS), comparatively faint and extensive signals spanning the entire gene body, and a robust yet broad peak emerging from the transcription termination site (TTS) extending into downstream regions (Fig. 2E). The distribution pattern of KAS-seq signals in NSN oocytes is consistent with that observed in somatic cells [17], serving as a direct indicator of increased transcriptional activity. In SN oocytes, although the KAS-seq signal is weak, the distribution pattern of ssDNA regions within gene-coding regions was similar to NSN oocytes(Fig. 2C, D). Indicating the presence of relatively weak gene transcription in SN oocytes. Simultaneously, we observed that in GVBD oocytes, KAS-seq profile on gene-coding regions is distinct from that in other stages. The ssDNA signal is relatively weak around the TSS region, but it gradually intensifies along the gene body, reaching a strong and broad peak near the TTS (Fig. 2E, Fig. S3B and C). This observation suggests that following meiotic resumption, transcription gradually ceases until reaching the GVBD stage. As transcription progresses, Pol II migrates from the TSS region towards the TTS region. Meanwhile, due to chromatin condensation, there is a reduction in de novo transcription, which leads to the gradual accumulation of ssDNA signals in the TTS region.

Previous studies showed that SN oocytes possess a higher developmental potential, as they are better equipped to support early embryonic development than NSN oocytes [24, 25]. Understanding the intricate transcriptional dynamics between NSN and SN oocyte is fundamental to elucidate the molecular processes underlying their developmental potentials. We previously performed the analysis of chromatin accessibility by ATAC-seq in SN and NSN oocytes (data not published). Pol II Stacc-seq was also employed to investigate the differences in transcriptional activity between SN and NSN oocytes [26]. Here, we conducted a comparative analysis between KAS-seq and these two sequencing techniques to delve deeper into the transcriptional characteristics SN nd NSN oocytes. This allowed us to elucidate whether KAS-seq technology offers advantages over the other two methods. Initially, we collected NSN and SN oocytes using the previously established method (Fig. S4A). The results indicate that this method effectively distinguishes NSN from SN oocytes (Fig. S4B). We subjected these oocytes to KAS-seq analysis. Subsequently, we performed a comparative analysis by juxtaposing the results obtained through KAS-seq with those from ATAC-seq and Pol II Stacc-seq (Fig. 3A, B). In both NSN and SN oocytes, KAS-seq results correlate well with the findings from Pol II Stacc-seq. This observation suggests that KAS-seq, in comparison to ATAC-seq, provides a more accurate representation of the transcriptional pattern in oocytes. Through an analysis of signals within gene-coding regions, we observed that in NSN oocytes, both Pol II signals and ATAC signals are considerably enriched at gene-coding regions, especially at gene promoters and transcription termination regions, exhibiting a distribution pattern consistent with the features observed in KAS-seq signals (Fig. 3C-E). In SN oocytes, the signals in the gene coding regions for KAS-seq, Pol II Stacc-seq, and ATAC-seq have all experienced a drastic reduction compared to NSN oocytes. However, there are differences in the signals identified by these three techniques in SN oocytes, particularly in the case of KAS-seq and ATAC-seq. As shown in the peak plots, the signal intensity is primarily concentrated around the transcription start site TSS, with a relatively weak signal increase near the TTS, exhibiting a pattern similar to that of Pol II Stacc-seq signals. In contrast to KAS-seq and Pol II Stacc-seq, the ATAC-seq signal is predominantly localized around the TSS with a robust signal but not around TTS region in SN oocytes. Additionally, quantitative statistical analysis of ATAC-seq peaks signals showed that 4,968 genes in the accessible chromatin region were covered in SN oocytes. While, the peaks density within the gene coding regions of SN oocytes is relatively weak, particularly in the TTS region, which is consistent with the condensed chromatin state of SN oocytes. This minimal chromatin accessibility makes it challenging to determine the presence of transcriptional activity in SN oocytes. KAS-seq data provide a more objective reflection of the transcriptional characteristics of SN oocytes. We can observe that while the transcriptional activity in SN oocytes is reduced compared to NSN oocytes, with a limited number of genes showing transcription, in line with the results of 5-ethynyl uridine staining (EU staining, a method to visualize newly synthesized RNA) (Fig. S4C, D). Subsequently, we conducted a functional enrichment analysis of the upregulated transcripts identified by RNA-seq in SN oocytes. We found that the upregulated genes in SN oocytes were enriched in processes related to meiotic nuclear division, chromosome organization, and sister chromatid cohesion (Fig. S4E), all of which are crucial for the resumption and maturation of meiosis. In summary, KAS-seq method, compared to ATAC-seq technique, offers a superior reflection of the transcriptional activity in oocytes.

In the above results, we investigated the differences of transcriptional activity between SN and NSN oocytes through KAS-seq profile, we next explored the correlation between KAS-seq and the transcriptome. Firstly, transcriptome analysis was performed upon chromatin configuration transition. Despite the differences in transcriptional activity between SN and NSN oocytes, their transcriptome differences are minimal, and the RNA-seq results show highly positive correlation with each other (Fig. S5A). There are only 333 differential transcripts (FC > 2, adjust p < 0.01) between SN and NSN oocytes out of tens of thousands of transcripts (Fig. 4A), which is inconsistent with the differences in their transcriptional activity. However, there is a substantial difference in the specific peaks identified in SN and NSN oocytes through KAS-seq profile (Fig. 4B, S5B). We then ranked all transcripts of SN and NSN oocytes into three groups according to their expression levels based on RNA-seq data [17] (Fig. 4C), and showed that the strength of KAS-seq signals increase notably in genes with high expression levels in NSN oocytes (Fig. 4D). This suggests that in NSN oocytes, the KAS-seq results are consistent with the RNA-seq data and can directly measures transcripts generation. In contrast to NSN oocytes, where highly abundant transcripts exhibit a strong KAS-seq signal, SN oocytes show a notably weak KAS-seq signal corresponding to highly abundant transcripts. Therefore, during the chromatin configuration transition from NSN to SN, there is an inconsistent dynamic change between the transcript’s abundance and the intensity of KAS-seq signals (Fig. 4E). We subsequently validated the correlation between KAS-seq data and transcripts. The quantitative PCR results indicate that gene expression is not significantly different between SN and NSN oocytes, although there is a substantial difference in KAS-seq signals (Fig. 4F, J). The above results indicate that during chromatin configuration transition, although there are significant changes in the dynamic of single-stranded DNA, accompanied by a decrease in transcriptional activity, the transcripts stored in oocyte cytoplasm remain relatively stable.

To further elucidate the potential function of KAS-seq signals that were gained, lost, or maintained during the transition in chromatin configuration, ssDNA signals in SN and NSN oocytes were categorized into four distinct clusters by using k-means algorithms (Supplementary Table 2): two stage-open domains (cluster 1), NSN-open domains (cluster 2), SN-open domains (cluster 3) and two stage-close domains (cluster 4) (Fig. S5C). Given that cluster 3 comprises only five domains, we conducted a more focused analysis on the domains within cluster 1 and cluster 2 (Fig. 5A, B). Gene Ontology (GO) analysis indicated that domains in cluster 1 was deposited for genes involved in regulation of mRNA metabolic process; whereas domains in cluster 2 were deposited for genes in cell cycle process, chromosome organization and meiotic nuclear division (Fig. 5A). These results indicate that genes with sustained expression in both SN and NSN oocytes may be involved in maintaining transcript stability. Due to the shared domains in both SN and NSN oocytes, we asked whether the common domains harbor motifs for transcription factors that involved in the sustained transcription of these genes. Notably, the ZFP143 factor is strongly enriched in the shared domains, and it is also highly expressed in both SN and NSN oocytes (Fig. 5C). Previous research has shown that the ZFP143 knockout leads to infertility, indicating a critical role for ZFP143 in the reproductive process of mice. However, the specific mechanism still requires further exploration.