2.2. Transcriptomic features related to histone loss in aged oocytes

We investigated whether the aging‐associated reduction in chromosomal H3–H4 is accompanied by chromatin and transcriptome changes that are caused by histone loss. The previous study showed that in oocytes conditionally deleted of Hira, a global loss of chromosomal H3–H4 is associated with a limited capacity to fine‐tune gene expression levels throughout the genome (Nashun et al., ²⁰¹⁵). To examine the global nucleosomal status, we performed micrococcal nuclease (MNase)‐seq analysis (Hughes & Rando, ²⁰¹⁴) with the MI chromosomes of conditional Hira knockout oocytes, which were obtained by crossing a newly established floxed Hira allele (Figure S2) with the oocyte‐specific Gdf9‐Cre recombinase (Lan et al., ²⁰⁰⁴). Gdf9‐Cre is activated around premeiotic S‐phase, a stage before dictyate arrest, and deletes a target gene with 90% efficiency (Burkhardt et al., ²⁰¹⁶). MNase‐seq analysis showed that the nucleosomal pattern at promotors in conditional Hira knockout oocytes was less distinct than in control oocytes (Figure S3A,B), confirming their severe defects in chromatin integrity (Nashun et al., ²⁰¹⁵). In contrast, the nucleosomal pattern in aged oocytes was not less distinct than in young oocytes (Figure S3A,B), consistent with the idea that aged oocytes carry largely maintained nucleosomal histones. However, transcriptome analysis of our published datasets of aged oocytes (Mishina et al., ²⁰²¹) revealed a significantly increased number of transcribed genes (Figure 2a), suggesting a defect in gene silencing, similar to conditional Hira knockout oocytes (Nashun et al., ²⁰¹⁵). Furthermore, we detected a significantly decreased variance in expression levels among transcribed genes in aged oocytes (Figure 2b), suggesting a limited dynamic range in controlling expression levels, as observed in conditional Hira knockout oocytes (Nashun et al., ²⁰¹⁵). Consistently, we found that genes that were down‐regulated with aging tend to be the ones that were highly expressed in young oocytes (Figure 2c). Thus, several transcriptomic features of aged oocytes are consistent with those caused by histone loss, to which aging can be a contributing factor.

Open in a separate window

FIGURE 2

Transcriptomic features of aged mice. (a) Number of annotated genes (CPM>1) from young and aged oocytes using RamDA‐seq. *: P=.047 using Welch's t‐test. (b) Log10‐scaled variance of gene expression among all annotated genes of each oocyte. *: P=.015 using Wilcoxon's rank‐sum test. (c) Violin plot showing the relationships between the distribution of median expression levels (FPKM) in young oocytes and the DEGs between young and aged oocytes. Significant differences between categories were tested using Welch's t‐test with Holm's correction for multiple comparisons. ***: Aged down versus aged up: P=4.8e−6; aged down versus all genes: P=9.1e−12; n.s.: Aged up versus all genes: P=.26.

2.3. Chromosome segregation abnormalities caused by histone loss are distinct from those associated with aging

Next, we aimed to determine whether the reduction of chromosomal H3–H4 contributes to aging‐associated chromosome segregation errors in oocytes. We speculated that conditional Hira knockout oocytes could be a good, although not perfect, model to address this question. To characterize the chromosomal H3 composition of conditional Hira knockout oocytes, we immunostained them with a specific H3.3 antibody, which confirmed that conditional Hira knockout oocytes carried a significantly decreased chromosomal H3.3 (Figure 3a). In contrast, immunostaining with a H3.1/H3.2‐specific antibody showed a significantly increased level of chromosomal H3.1/H3.2 in conditional Hira knockout oocytes (Figure 3b). This increased H3.1/H3.2 is likely a fraction that persisted abnormally after DNA replication due to the absence of HIRA, as HIRA normally replaces H3.1/H3.2 with H3.3 on chromosomes after DNA replication (Ahmad & Henikoff, ²⁰⁰²; Tagami et al., ²⁰⁰⁴). Thus, conditional Hira knockout oocytes are a model for oocytes with reduced H3.3 and increased H3.1/H3.2.

Open in a separate window

FIGURE 3

Conditional deletion of Hira results in chromosomal abnormalities distinct from those of aged oocytes. (a,b) Conditional Hira deletion results in decreased H3.3 and increased H3.1/H3.2. Oocytes at MI were immunostained with anti‐H3.3 (a) or anti‐H3.1/H3.2 (b) and Hoechst33342. Color code with 16 colors. BDF1 mice at 2months old (young) and 14–22months old (aged) were used. Intensity of immunofluorescence relative to that of Hoechst33342 is shown. Mann–Whitney test. Parentheses show the number of oocytes from three independent experiments. Scale bar, 5μm. (c,d) Conditional Hira knockout oocytes exhibit anaphase bridges. Oocytes collected from mice at 4weeks old were monitored for MI with major‐satellite‐TALE‐mClover (centromere) and H2B‐mCherry (chromosome). Centromere signals were processed for peak enhancement and background subtraction. Time after germinal envelope breakdown (GVBD, entry to M‐phase) is shown (h:mm:ss). The arrowhead indicates the anaphase bridge. Scale bar, 5μm. In (d), oocytes that underwent anaphase were categorized based on anaphase figures. Parentheses show the number of anaphases from six independent experiments.

We acquired 4D datasets of chromosome dynamics with the chromosome marker H2B‐mCherry and the centromere marker major‐satellite‐TALE‐mClover (Miyanari et al., ²⁰¹³) in conditional Hira knockout oocytes throughout meiosis I using live confocal microscopy (Figure 3c). We performed this at a high spatiotemporal resolution capable of detecting major aging‐associated defects, such as premature chromosome separation (Kitajima et al., ²⁰¹¹; Sakakibara et al., ²⁰¹⁵). We found that conditional Hira knockout oocytes frequently exhibited anaphase bridges (Figure 3c,d), consistent with the previous report (Nashun et al., ²⁰¹⁵), and different from aged oocytes, where anaphase bridges are not increased (Sakakibara et al., ²⁰¹⁵). Despite the highly frequent anaphase defects in conditional Hira knockout oocytes, their prometaphase or metaphase did not exhibit premature chromosome separation (0 of 58 oocytes), a major event that precedes anaphase segregation errors in aged oocytes (Sakakibara et al., ²⁰¹⁵). These results show that conditional Hira knockout oocytes result in chromosome segregation abnormalities that are distinct from those caused by aging‐associated defects.

2.4. Hira insufficiency does not accelerate aging‐associated increase in chromosome segregation errors

It is possible that histone loss due to conditional Hira knockout causes inefficient chromosome condensation. Consistently, anaphase bridges found in conditional Hira knockout oocytes resembled those observed in oocytes deficient in chromosome condensation due to condensin dysfunction (Houlard et al., ²⁰¹⁵; Lee et al., ²⁰¹¹). Insufficient condensation can lead to chromosome entanglement, which may have prevented premature chromosome separation in conditional Hira knockout oocytes. We therefore asked whether a mild loss of chromosomal H3.3–H4, as observed in aged oocytes, could facilitate premature chromosome separation. If so, conditional heterozygous Hira knockout (Hira ^+/flox Gdf9‐Cre) could accelerate the aging‐associated increase in premature chromosome separation in oocytes. Conditional heterozygous Hira knockout accelerated aging‐associated reduction of fully grown oocytes in ovaries (Figure 4a), while their chromosomal H3.3 levels were not significantly different from those of control oocytes (Figure 4b). These observations are consistent with the idea that conditional heterozygous Hira knockout oocytes carry a slightly and undetectably reduced amount of chromosomal histones. We then performed live imaging of conditional heterozygous Hira knockout oocytes at young (2–4months old) and middle (6–8months old) ages (Figure 4c), although we could not analyze oocytes from aged mice (12–15months old) due to their severe scarcity (Figure 4a). This analysis showed that Hira insufficiency caused no detectable increase in premature chromosome separation at the middle age (1/86 oocytes in Hira ^+/f vs. 0/71 oocytes in Hira ^+/f Gdf9‐Cre). These data do not provide evidence for a link between reduced chromosomal H3.3 and aging‐associated chromosome segregation errors in oocytes.

Open in a separate window

FIGURE 4

Conditional heterozygous deletion of Hira does not accelerate aging‐associated increase in premature chromosome separation. (a) Conditional heterozygous Hira knockout accelerates a decrease in the number of oocytes. Fully grown oocytes were collected from the ovaries of mice at 2–4months old (young), 6–8months old (middle), and 12–15months old (aged). Dots indicated data from individual mice. 5 (young), 6 (middle) and 2 (aged) independent experiments were performed. (b) Conditional heterozygous Hira knockout results in undetectable decrease in H3.3. Oocytes at MI were immunostained with anti‐H3.3 and Hoechst33342. Color code with 16 colors. Mice at 2months old were used. Intensity of immunofluorescence relative to that of Hoechst33342 is shown. Mann–Whitney test. Parentheses show the number of oocytes from two independent experiments. Scale bar, 5μm. (c) Conditional heterozygous Hira knockout does not exhibit increased premature chromosome separation. Oocytes collected from mice at 2–4months old (young) and at 6–8months old (middle) were monitored for MI with 2mNeonGreen‐CENP‐C (centromere) and H2B‐mCherry (chromosome). Premature chromosome separation was observed in 1/86 oocytes from Hira ^+/f mice and 0/71 oocytes from Hira ^+/f Gdf9‐Cre mice at the middle age; and in 1/76 oocytes from Hira ^+/f mice and 1/66 Hira ^+/f Gdf9‐Cre mice at the young age. Six independent experiments were performed. Centromere signals were processed for peak enhancement and background subtraction. Time after GVBD is shown (h:mm:ss). Scale bar, 5μm.

4.2. Genetic engineering

The Hira conditional knockout (floxed) mice (Accession No.: CDB0105E: https://large.riken.jp/distribution/mutant-list.html) and the Chaf1a floxed mice (Accession No.: CDB0106E) were generated by CRISPR/Cas9‐mediated genome editing in C57BL/6 zygotes using single‐strand oligodeoxynucleotides (ssODN) as previously described (Hashimoto et al., ²⁰¹⁶). gRNA sites were designed using CRISPRdirect (Naito et al., ²⁰¹⁴), and crRNA/tracrRNA and ssODN were chemically synthesized (Fasmac Co., Ltd). The floxed alleles are shown in Figures S2 and S4. Routine genotyping PCR was performed using following primers; 5gtFW (5′‐CAT CCG GCT TCA CCA TCC‐3′) and 5gtREV (5′‐TCT CCT GTG CAC CAC TGG‐3′) (WT: 289bp, 5′‐loxP: 329bp) and 3gtFW(5′‐ GGC CAT ATG ACA ACT AGC‐3′) and 3gtREV(5′‐TGC CTC TAC CTC CCA TGC‐3′) (WT: 267bp, 3′‐loxP: 307bp) for Hira floxed mice, and 5gtFW (5′‐CTC CTC TCT TAG TGC TGC‐3′) and 5gtREV (5′‐TGT GGG TCT CGT GGA GCC‐3′) (WT: 283bp, 5′‐loxP: 323bp), and 3gtFW (5′‐TTG TGA GCC ACC ATG TGG‐3′) and 3gtREV (5′‐GAC ACT GCC TCA GTA GGC‐3′) (WT: 188bp, 3′‐loxP: 228bp) for Chaf1a floxed mice.

4.3. Mouse oocytes

Fully grown oocytes were isolated from the ovaries 48h after injection of pregnant mare's serum gonadotropin (Serotropin, ASKA Animal Health) into female mice, and cultured in M2 medium containing 3‐isobutyl‐1‐metyl‐xanthine (IBMX, Sigma). The mixture of mRNA was introduced by microinjection, and then oocytes were cultured for at least 2h at 37°C. The oocytes were released from IBMX by washing to induce meiotic resumption.

4.4. Transcriptome analysis

We re‐analyzed our previously published dataset of single oocyte transcriptomes from young (2‐month‐old) and aged (12‐month‐old) BDF1 mice (GSE159281) (Mishina et al., ²⁰²¹). The number of annotated genes with counts per million (CPM) >1 was counted for each oocyte, and their difference with respect to the mouse age class was tested using Welch's t‐test. The variance of gene expression of all annotated genes (CPM>0) in each oocyte was compared between the age classes using Wilcoxon's rank sum test because its distribution violated the assumption of Gaussian distribution (P=1.32e−10 using Shapiro–Wilk test). Finally, the relationships between expression levels in oocytes of young mice and differential expression were examined. The list of differentially expressed genes (DEGs) reported in our previous study (Mishina et al., ²⁰²¹) was used. Briefly, these DEGs (false discovery rate, FDR <0.05 calculated using the R package “qvalue”) were identified using the edgeR package (Robinson et al., ²⁰¹⁰) in R, which normalizes library sizes with the trimmed mean of M values (TMM) method, using the dataset of genes expressed in at least five samples.

4.5. Low‐input MNase‐seq

Ten MI oocytes per replicate were sampled into a tube containing the lysis buffer, frozen with liquid nitrogen, and subjected to MNase‐seq as previously described (Sakamoto et al., ²⁰²³). The MNase treatment condition was 21°C for 7.5min. The CAM pipeline (Hu et al., ²⁰¹⁷) was used to assess the data quality of MNase‐seq. Paired‐end sequencing was performed on a Nextseq 2000. The sequencing read information was summarized in Table S1. The data processing was performed in accordance with (Sakamoto et al., ²⁰²³). The sequence data sets were deposited to the Sequence Read Archive (SRA) under accession number PRJNA1122508.

4.6. Immunostaining of oocytes

Oocytes were pre‐extracted in PBT (PBS with 0.1% Triton X‐100) for 1min. The oocytes were fixed with 3.2% formaldehyde (methanol‐free) in PBT for 10min at room temperature. The oocytes were washed with PBT. After blocking with 3% bovine serum albumin (BSA)‐PBT for 30min, the oocytes were incubated with primary antibodies at 4°C overnight. Oocytes were washed with 3% BSA‐PBT and then incubated with secondary antibodies and 20μg/mL Hoechst33342 for at least 2h. Oocytes were washed and imaged using Zeiss LSM780. For immunostaining H4, oocytes were fixed with 1.6% formaldehyde in 100mM PIPES with 1mM MgCl₂ and 0.1% Triton‐X100 for 30min. The oocytes were washed with PBT and fixed again with cold methanol for 15min.

The following primary antibodies were used: a mouse anti‐H3 antibody (1:500, Abcam, ab1791), a mouse anti‐H3.1/3.2 antibody (1:100, Active Motif), a mouse anti‐H3.3 (1:20000, Active Motif), and rabbit anti‐H4 antibody (1:200, no. 13919, Cell Signaling). Alexa Fluor 555 goat anti‐mouse IgG (H+L) (A21424) (1:500, Thermo Fisher) was used as a secondary antibody.

4.7. Quantification of fluorescence signal intensity

Hoechst33342 signals in z‐stack images were subjected to segmentation of whole chromosome regions using the automatic 3D surface rendering of Imaris software (Oxford Instruments). The mean fluorescence intensity of a protein of interest, as well as that of Hoechst33342 as a reference, was obtained within the segmented regions. The mean background intensity for each channel was obtained in cytoplasmic regions using Fiji (https://fiji.sc/). In each channel, the background values were subtracted from the values of the segmented regions. The ratio of the value obtained from the channel for a protein of interest to that of Hoechst33342 was calculated.

4.8. Live imaging

Messenger RNAs were transcribed using mMESSAGE mMACHINE T7 Kit (Ambion). The following RNAs were introduced into oocytes through microinjection: 0.04–0.6pg H2B‐mCherry, 0.2–0.4pg major‐satellite‐TALE‐mClover, and 0.2–0.7pg 2mNeonGreen‐CENP‐C. Live imaging was performed using a Zeiss LSM710 or LSM780 confocal microscope equipped with a GaAsP detector and a ×40 C‐Apochromat 1.2NA water immersion objective lens, controlled by AutoFocusScreen (Rabut & Ellenberg, ²⁰⁰⁴) and MyPiC (Politi et al., ²⁰¹⁸). For chromosome and kinetochore imaging, 17 confocal z‐sections (every 1.5μm) of 256×256 pixel xy images were acquired every 3min. Images were visualized in 3D with Imaris (Bitplane) to detect premature chromosome separation and categorize anaphase segregation.

We thank J. Ellenberg for a microscope automation macro, the imaging and animal facilities of RIKEN BDR for technical support, and all lab members for discussions and comments. O. T. is supported by the RIKEN Special Postdoctoral Researchers Program. This work was supported by RIKEN Intramural grants, JSPS KAKENHI Grant Number 23H04948, 21H02407, and 18H05549 to T. S. K.