The Neo-X Does Not Form a Barr Body but Shows a Slightly Condensed Structure in the Okinawa Spiny Rat (Tokudaia muenninki) Free

In female somatic cells of placental mammals, one of the two X chromosomes becomes transcriptionally silenced, compensating for the difference in the transcription of genes located on the X chromosome between the two sexes. This mechanism, called X chromosome inactivation (XCI), is common in placental mammals [Lyon, 1961]. XCI is a well-established process in most mammalian species; therefore, little is known about the initial state in the acquisition of XCI.

The genus Tokudaia (Muridae, Rodentia) consists of three species, each of which has unique sex chromosomes. The Okinawa spiny rat (T. muenninki, 2n = 44) possesses sex chromosomes fused with a pair of autosomes. The short arms derived from autosomes have been called “neo-X” and “neo-Y” regions on the X and Y chromosomes, respectively. These regions are homologs of mouse chromosomes 11 and 16 [Murata et al., 2012]. In addition, the neo-X region is homologous to the long arm of chromosome 3 of Tokudaia osimensis and chromosome 10 of Tokudaia tokunoshimensis [Nakamura et al., 2007]. Our previous study showed genetic differentiation between the pericentromeric regions of neo-X and neo-Y regions in T. muenninki as well as an accelerated rate of evolution in the neo-Y region through the detection of male-specific substitutions by gene sequencing in multiple males and females, and each neo-sex-derived BAC sequencing [Murata et al., 2015]. However, based on an RNA-seq analysis of 55 genes in the pericentric region of the neo-X, there is no difference in expression levels between males and females [Zushi et al., 2017]. On the other hand, X-linked non-coding X-inactive-specific transcript (Xist) RNA on the inactive X chromosome is localized on the entire long arm (ancestral X) and part of the short arm (neo-X) [Zushi et al., 2017]. DNA-FISH using the LINE-1 sequence, which is correlated with Xist RNA localization [Lyon, 1998; Bailey et al., 2000], has shown fluorescence signal accumulation only on the long arm of the X chromosome and the region around the centromere, with no signals in the neo-X [Zushi et al., 2017]. These results suggest that dosage of genes located on the neo-X chromosome was not compensated, even though Xist RNA is partially localized. Furthermore, it is still not clear whether the species shows heterochromatinization of the neo-X, initiated by Xist RNA.

XCI is triggered by Xist RNA. Briefly, Xist RNA is transcribed from and subsequently coats a random X chromosome in female somatic cells. The coated X chromosome exhibits gene silencing [Brown et al., 1991, 1992; Clemson et al., 1996]. The accumulation of histone H3 lysine 27 trimethylation (H3K27me3) is one of the earliest epigenetic changes on the X chromosome, which is inactivated after Xist RNA accumulation [Plath et al., 2002; Heard and Disteche, 2006; Marks et al., 2009]. Lysine 27 of histone H3 is methylated by polycomb repressive complex 2 (PRC2), a multi-subunit methyltransferase enzyme [Hasegawa et al., 2010; Chu et al., 2015; McHugh et al., 2015; Minajigi et al., 2015]. Xist RNA recruits noncanonical Polycomb group RING finger 3/5-PRC1, in turn inducing the mono-ubiquitylation of histone H2A lysine 119 (H2AK119ub). H2AK119ub recruits noncanonical PRC1 and PRC2 [Almeida et al., 2017]. PRC2 induces the trimethylation of H3K27. H3K27me3 is a repressive histone modification and accumulates on the whole inactivated X chromosome.

The inactivated X chromosome is condensed during interphase, taking the form of a Barr body [Barr and Bertram, 1949; Barr and Carr, 1962]. The Barr body appears as dark or dense spots around the nucleus and adjacent to the nucleolus or the nuclear envelope in female cells in many mammals [Klinger et al., 1957; Kenney and Ittwoch, 1965]. However, mice have many chromocenters, which hide the condensed structures of inactive X chromosomes, like Barr bodies [Barr and Carr, 1962]. Therefore, in mice, the condensed heterochromatic structure of the inactive X chromosome can be observed by using DNA probes or other labeling methods [De La Fuente et al., 2011; Deng et al., 2015]. On the other hand, the Barr body has been observed in the nerve cells of rats, which are also rodents [Ohno et al., 1959; Beis et al., 1980]. It forms by chromosome folding via the binding of SMCHD3 and HP1 to the H3K9me3 and H3K27me3 scaffolds, respectively [Almeida et al., 2017]. HP1 has a highly conserved chromo domain and a chromo shadow domain and binds to H3K9me3 [Bannister et al., 2001; Lachner et al., 2001], and HP1 is one component of the Barr bodies [Chadwick and Willard, 2003]. Heterochromatic regions including the inactivated X chromosome are replicated in late S phase [Taylor, 1960; Priest et al., 1967].

In this study, we confirmed the chromatin state of inactive X chromosomes in T. muenninki by R-banding by acridine orange (RBA), H3K27me3 and HP1 immunostaining, and Xist RNA-FISH. We also observed the formation of the inactive X chromosome in interphase by cross-species fluorescence in situ hybridization (Zoo-FISH) and BAC-FISH.

Fibroblast cells derived from the tail of T. muenninki were cultured in 20% AmnioMax-II Complete Medium (Thermo Fisher Scientific, Waltham, MA, USA) within 20% FBS/DMEM at 37°C in the presence of 5% CO₂. BrdU (10 mg/mL, 0.1 N NaOH) was added to the culture medium at a final concentration of 50 μg/mL. After 9–10.5 h of culture, 40 μg/mL colcemid was added in the last 30 min of culture. Then, cells were collected by trypsinization and hypotonically treated with 75 mm KCl solution at room temperature for 15 min. The cells were fixed by Carnoy’s solution (methanol:acetic acid = 3:1). Fixation was repeated at least three times. Preparations were obtained by the air-drying method [Rothfels and Siminovitch, 1957]. Sörensen’s buffer solution was prepared with acridine orange in a Coplin jar. Dried glass slides were placed in the solution and allowed to stand for at least 1 h. After staining, they were rinsed gently with water and momentarily immersed in an emulsion jar containing Sörensen’s buffer solution. After rinsing the slide glass with buffer solution, a cover glass was placed on the slide and sealed with nail polish.

The sterilized cover glasses (18 × 18 mm) were placed on a Petri dish for cell culture. The glass slides were then removed and placed on a 35-mmØ treated dish, containing 4% paraformaldehyde /PBS (stored at 4°C, 2 mL/dish) at room temperature for 10 min. After fixation, the cells were treated with 0.5% paraformaldehyde/PBS (2 mL/dish) for 5 min at room temperature. Then, permeabilized cells were washed twice with PBS (2 mL/dish, 10 min at room temperature). As a primary antibody, α-H3K27me3-mouse IgG (MBL, Medical and Biological Laboratories Co., Ltd, Tokyo, Japan) and H3K9me3-rabbit IgG (Abcam, Cambridge, UK) were used. The antibody was diluted in 0.5% BSA (fraction V) (Sigma-Aldrich, St. Louis, MO, USA) at 1:800 (v/v) and 1:50 (v/v). For double-immunostaining, α-H3K27me3-mouse IgG (MBL) and α-HP1beta-rabbit IgG (Abcam, Cambridge, UK) were mixed in 0.5% BSA/PBS at 1:400 and 1:50, respectively. The cells were treated with the antibody mixture in a humidified box at 4°C overnight. Then, the cells were washed with 0.05% Tween-20/PBS. As secondary antibodies, α-mouse IgG-Cy3 (Stemgent, Cambridge, MA, USA) was used for α-H3K27me3-mouse IgG and α-rabbit IgG-FITC (Invitrogen, Waltham, MA, USA) was used for α-H3K9me3-rabbit IgG and α-HP1beta-rabbit IgG. The cells were treated with the secondary antibody mixture in a humidified box at 37°C for 1 h. Samples were placed in PBS at 37°C and left for 10–15 min. A drop of VECTASHIELD Mounting Medium (H-1000, antibody contact material and encapsulant; Vector Laboratories, Inc., Newark, CA, USA) mixed with 1 μg/mL Hoechst 33342 was applied to a tip, placed on a glass slide, and covered with a cover glass. The slides were sealed with nail polish.

Growing fibroblast chromosomes were treated with colcemid (0.04 μg/mL) for 15 h prior to trypsinization. After trypsinization, collected cells were resuspended in 2 mL of 75 mm KCl hypotonic solution. After maintenance on ice for at least 10 min, cell suspensions were allowed to stand at 0°C for at least 10 min and then subjected to the cytospin method [Gooderham and Jeppesen, 1983]. Various volumes of the cell suspension were added to 2 mL of 75 mm KCl in the chamber of SC-02 centrifuge (Tomy Seiko Co., Ltd., Tokyo, Japan) and centrifuged at 1,450 rpm for 10 min to prepare cell slides. Immediately after centrifugation, the slides were transferred to emulsion jars filled with KCM solution (120 mm KCl, 20 mm NaCl, 10 mm Tris-HCl, pH 7.5, 0.5 mm EDTA, and 0.1% [v/v] Triton X-100) and immersed for 30 min to 12 h at room temperature. Following incubation, cells were treated with α-H3K27me3-mouse IgG (MBL) in 0.5% BSA/PBS at 1:250 dilution for 18 h in a humidified box at 4°C. Cells were washed in KCM. Then, cells were treated with α-mouse IgG-Cy3 (Stemgent) diluted in 0.5% BSA/PBS in a humidified chamber at 37°C, cells were finally fixed in KCM containing 10% (v/v) formaldehyde for 10 min in the dark. Cells were mounted with VECTASHIELD containing 1 μg/mL Hoechst 33342.

Xist RNA-FISH was performed as described in a previous study [Zushi et al., 2017]. A partial fragment of mouse Xist containing the A-repeat region was used as an RNA-FISH probe [Zushi et al., 2017].

Zoo-FISH was carried out as described in a previous study [Murata et al., 2012]. The signal distribution of two areas in nucleus was measured by ImageJ. Statistical analysis to compare the value was performed using unpaired, two-tailed t tests.

For BAC-FISH, the preparation of R-banded chromosomes and FISH was performed as described previously [Kobayashi et al., 2007; Kimura et al., 2014]. BACs of T. muenninki were the same as those used in a previous study [Murata et al., 2015]. The BAC clones and genes were as follows: TMB1-165M17 (Foxk2), 064N22 (Nptx1), 089D20 (Doc2b), 038H17 (Gemin4), 081H15 (Emp2), 016L10 (Lmf1), 354D04 (Tnfrsf17), 061I05 (Litaf), 266E06 (Parn), 113A21 (Snn), and 110N16 (Nde1). The distance between the furthest signals in each region was measured by ImageJ. Statistical analysis to compare the value was performed using unpaired, two-tailed t tests.

To detect the heterochromatic region on the neo-X of the inactive X chromosome of T. muenninki, the RBA method was used. In this method, BrdU is continuously added to cells from the middle to latter half of the S phase, thus labeling heterochromatic regions that replicate in the late S phase [Priest et al., 1967]. The euchromatic region stained with acridine orange shows bright green fluorescence, while heterochromatic regions characterized by late replication are observed in dark red. On one of the two X chromosomes, the ancestral X region (Xq) was observed in dark red, indicating that it is inactive (Xi in Fig. 1). The neo-X region (Xp) showed bright fluorescence on both inactive and active X chromosomes, indicating that the neo-X region does not contain heterochromatin.

H3K27me3 is a histone modification enriched in heterochromatic regions. We performed immunostaining of H3K27me3 in T. muenninki chromosomes. On one of the two X chromosomes, we detected fluorescein signals in the ancestral-X region (Fig. 2a, b; online suppl. Fig. S1; for all online suppl. material, see https://doi.org/10.1159/000531275), indicating that this was the inactive X chromosome (Xi in Fig. 2c). By contrast, the neo-X region did not show significant differences in signal strength between inactive and active X chromosomes, indicating a lack of heterochromatinization in the neo-X region of the inactive X chromosome.

The signals observed by immunostaining of H3K27me3 in interphase chromosomes were divided into two regions (Fig. 3a–c). The distribution of Xist RNA in interphase chromosomes was the same as that in metaphase chromosomes (Fig. 3d–f). These signals were referred to as R1 and R2 in the ancestral X region.

To confirm the condensed configuration of ancestral X of the inactive X chromosome, we performed Zoo-FISH using a mouse X chromosome painting probe. We detected signals in the ancestral X region of the metaphase chromosomes (Fig. 4a–c). These results were consistent with those of a previous study [Murata et al., 2012]. In interphase chromosomes, one of the two X-territories showed more condensed staining (Fig. 4d–i), suggesting that this was the inactive X chromosome (Xi in Fig. 4f, i). We measured the signal distribution of two areas in nucleus, one is larger, suggesting that this is Xa and another one is smaller, indicating that is Xi. A total of 12 nuclei were observed, and the average values are shown in Figure 4j. A significant difference was confirmed in the areas of the signal distributions in both, suggesting that Xa and Xi could be distinguished.

Double-immunostaining for the detection of H3K27me3 and HP1 in interphase chromosomes was performed to confirm Barr body formation. HP1 is one component of the Barr body [Chadwick and Willard, 2003]. H3K27me3 signals showed the same distribution as shown in Figure 3a–c and could be divided into two regions, R1 and R2 (Fig. 5a). The distribution of HP1 signals was similar to that of H3K27me3 (Fig. 5b), and their signals were merged in R1 and R2 (Fig. 5c, d). HP1 is known to be colocalized with H3K9me3 [Bannister et al., 2001; Lachner et al., 2001]. H3K9me3 signals also showed the similar distribution of HP1 that are divided into two regions (Fig. 5e–h). These results indicate that the Barr body structure is composed of two components, R1 and R2, in the ancestral X region, and the neo-X region does not contribute to Barr body formation in the inactive X chromosome.

To confirm the detailed structures of the ancestral X and neo-X, we performed two-color BAC-FISH using BACs located on the T. muenninki X chromosome (Fig. 6a–d). Two BACs were hybridized to interphase nuclei of T. muenninki, one BAC containing the forkhead box K2 (Foxk2) gene located on the distal end of Xp (neo-X) [Murata et al., 2015] and another containing the Xist gene located on Xq12–q21 (ancestral X) [Zushi et al., 2017] (Fig. 6e). For 12 of 22 nuclei observed in this study, the long and short distances between the two signals (red indicating Foxk2 and green indicating Xist) were observed in each nucleus (Fig. 6d).

Next, we performed BAC-FISH using 11 BACs containing genes on the neo-X region (Fig. 7a–f). The localization of each BAC on the neo-X (Xp) is shown in Figure 7g. In metaphase, these signals were distributed in several clusters from the centromeric region to the telomeric region of the neo-X (Fig. 7a–c). In 30 nuclei observed in this study, the signals of these BACs showed two distinct distribution patterns: in a narrow or dispersed region within interphase nuclei (Fig. 7d–f). The signals concentrated in a narrow region were assumed to represent the inactive X (Xi in Fig. 7f), and signals dispersed over a wide region were assumed to represent the active X (Xa in Fig. 7f). We measured distance between the furthest signals in a narrow region and a wide region in each nucleus, and the average values are shown in Figure 7h. A significant difference was confirmed in the signal distance between both.

Hoechst staining did not clearly show Barr bodies in about half of cells. Therefore, to investigate the location of the Barr body in the nucleus, we evaluated female T. muenninki cells by immunostaining for H3K27me3 combined with Hoechst 33342 staining. We observed 134 nuclei. Among Barr bodies, 58.9% were located at the nuclear periphery (Fig. 8a, b), 35.8% were near a nucleolus (Fig. 8c, d), and 5.2% were not close to any particular structure (Fig. 8e, f).

Although Xist RNAs were partially localized to the neo-X region in a previous study of Tokudaia [Zushi et al., 2017], RBA banding and immunostaining of H3K27me3 in this study indicated that heterochromatinization does not occur in the neo-X region. We therefore hypothesized that XCI, which is initiated by Xist RNA, is incomplete in the neo-X region. The male:female gene expression ratio in the neo-X region was evaluated in a previous study [Zushi et al., 2017], revealing mean and median ratios (neo-X male/2 neo-X female) of approximately 0.5, indicating that XCI has not yet evolved in the neo-X region at the gene dosage level. We conclude that there is no gene dosage compensation by heterochromatinization in the neo-X region.

We detected structural condensation of the ancestral X of the inactive X by Zoo-FISH. As determined by double-immunostaining, H3K27me3 and HP1 were colocalized and could be divided into two groups, R1 and R2, in the ancestral X. This result indicates that the Barr body structure is constructed from two components, R1 and R2, of the ancestral X region. A structural analysis of cells of several mammalian species has shown that the entire inactive X chromosome folds into a unique bipartite structure composed of two megadomains showing very little local structure, separated by a boundary element containing the DXZ4/Dxz4 macrosatellite [Rao et al., 2014; Deng et al., 2015; Minajigi et al., 2015: Darrow et al., 2016; Giorgetti et al., 2016; Dossin and Heard, 2022]. The DXZ4/Dxz4 locus is heterochromatic in the active X chromosome, whereas the allele in the inactive X chromosome is euchromatic and shows high binding affinity to CTCF [Chadwick, 2008; Horakova et al., 2012]. We therefore suppose that the two components R1 and R2 observed in this study correspond to the bipartite structure, with Dxz4 located at the boundary. A comparative genome analysis has shown that the mouse homolog of microsatellite DXZ4 is located downstream of Pls3 in diverse mammals [Horakova et al., 2012]. In our previous comparative FISH mapping analysis using mouse cDNA clones of 22 X-linked genes [Murata et al., 2012], all genes were mapped to the ancestral X chromosome region of T. muenninki. Comparative analyses of gene order in mouse and T. muenninki X chromosomes can reveal the location of Pls3. Furthermore, if the genome sequence is obtained in the future, the Dxz4 locus in the ancestral X chromosome region in T. muenninki can be determined, and the structural relations of the locus and R1 and R2 components can be revealed.

We did not observe HP1 signals in the neo-X region, suggesting that this region does not contribute to Barr body formation in the inactive X chromosome. However, a BAC-FISH analysis showed signals on the neo-X region of the inactive X chromosome concentrated in a narrow region. We hypothesized that the neo-X region of the inactive X chromosome does not form a complete Barr body but forms a slightly condensed structure. It is not clear why this slightly condensed structure forms, despite the lack of a Barr body structure. However, incomplete inactivation states occur in the neo-X region (e.g., partial binding of Xist RNA), and this may represent an early chromosomal state in the acquisition of an inactivation mechanism. In our previous study, the sequence differentiation was identified in genes on pericentric region of the neo-X and neo-Y [Murata et al., 2015]. However, if amino acid sequences are conserved between alleles on the neo-X and neo-Y, the gene dosage does not need to be compensated. This indicates that chromosome formation involved in XCI precedes functional differentiation between alleles on neo-sex chromosome regions. To support this idea, further analysis for chromosome formation and expression analysis for genes on entire neo-X and neo-Y regions are needed in the future.

In rodents, heterochromatin regions on autosomes are generally condensed during interphase, and Barr bodies cannot be distinguished in all types of cells observed [Kenney and Ittwoch, 1965]. Stained areas in the nucleus that do not match the H3K27me3 signal are thought to be condensed autosomal heterochromatin regions, once called chromocenters [Jones, 1970]. Autosomes may become less condensed in a time-specific manner or in response to other factors, and only the condensed Barr bodies are observed. Since there are many heterochromatin blocks on the autosomes in T. muenninki [Murata et al., 2010], Barr bodies might not be clearly visible for the same reason.

The mechanism underlying XCI is well established in most mammals, limiting evolutionary analyses of the early stage of XCI acquisition. Mammalian species in which neo-X chromosomes arose relatively recently, such as T. muenninki, can improve our understanding of the evolution of XCI.

The Okinawa spiny rat (T. muenninki) is endangered (the IUCN Red List of Threatened Species, DOI: e.T21972A22409515) and has been protected by the Japanese government as a natural monument and national endangered species of wild fauna and flora since 1972 and 2016, respectively. With permission from the Agency for Cultural Affairs and the Ministry of the Environment in Japan, all animals were released at their capture sites after a small piece was cut from the tip of the tail. All animal experiments described in this study were approved by the Institutional Animal Care and Use Committee of National University Corporation, Hokkaido University (20-0038) and were performed following the Guidelines for the Care and Use of Laboratory Animals issued by Hokkaido University. This study did not involve any human participants or specimens.

The authors have no conflicts of interest to declare.

This work was supported by JSPS KAKENHI Grant No. 22H02667.

Conceptualization: Asato Kuroiwa; data curation: Ikuya Yoshida; investigation: Ryoma Kudo; project administration: Asato Kuroiwa; resources: Yoko Kuroki and Takamichi Jogahara; supervision: Ikuya Yoshida, Shusei Mizushima, and Asato Kuroiwa; writing and original draft preparation: Ryoma Kudo and Asato Kuroiwa; writing, review and editing: Ikuya Yoshida, Luisa Matiz Ceron, Shusei Mizushima, Yoko Kuroki, and Takamichi Jogahara.

All data generated or analyzed during this study are included in this article and its online supplementary material.