Dynamic changes in the interchromosomal interaction of early histone gene loci during development of sea urchin

Animal development starts from a fertilized egg, and the individual animal body consists of tissues and organs that are properly differentiated during development. Accurate spatiotemporal gene expression plays a pivotal role in normal development, and it is controlled not only by transcriptional activation and repression by transcription factors but also by DNA methylation and histone modifications (Margueron and Reinberg, 2010). Furthermore, during the early development of various animals, nuclear size, nucleosome positioning, the types of histone variants, DNA methylation and covalent histone modifications undergo dramatic changes (Clarke et al., 1998; Fulka et al., 2008; Edens et al., 2013). However, the regulatory mechanisms underlying gene regulatory networks to overcome such dynamic changes have not been clarified.

Nuclear organization has been studied in various organisms. Each chromosome occupies a discrete territory within the interphase nucleus (Parada and Misteli, 2002; Cremer and Cremer, 2010). In yeast, Drosophila and many other eukaryotes, chromosomes adopt a Rabl configuration, in which the centromeres of all chromosomes cluster near the nuclear periphery and the telomeres near the opposite side of the nuclear periphery (Hochstrasser et al., 1986; Funabiki et al., 1993; Jin et al., 1998). By contrast, the Rabl configuration is largely absent in mammalian cells (Chuang et al., 2004). Nuclear organization may be closely associated with gene expression. While gene-rich chromosomes tend to locate in the center of the nucleus, gene-poor chromosomes tend to locate in the peripheral region (Croft et al., 1999). Furthermore, actively transcribed protein-coding genes localize preferentially to the peripheries of chromosome territories, and non-coding regions localize to the interior regions of chromosome territories or distribute randomly (Kurz et al., 1996). Interestingly, artificial tethering experiments have showed that the environment at the nuclear periphery can repress gene expression, but that active transcription also can occur at the nuclear periphery (Finlan et al., 2008; Kumaran and Spector, 2008; Reddy et al., 2008). It has been reported that inactive gene loci at the nuclear periphery move to a more-interior site upon activation (Kosak et al., 2002; Zink et al., 2004; Ragoczy et al., 2006; Peric-Hupkes et al., 2010). By contrast, many active genes physically interact with nuclear pore proteins, thus positioning these genes at the nuclear periphery (Brickner and Walter, 2004; Casolari et al., 2004; Rohner et al., 2013); interactions with soluble nuclear pore proteins can also occur in the nucleoplasm in Drosophila and mammals (Capelson et al., 2010; Kalverda et al., 2010; Liang et al., 2013; Pascual-Garcia et al., 2017). Active RNA polymerase II accumulates in the transcriptionally active regions and forms a nuclear complex called the transcription factory (Carter et al., 2008). Some genes extend outward from their chromosome territories and contact transcription factories (Fraser and Bickmore, 2007; Schneider and Grosschedl, 2007). Furthermore, co-regulated genes on the same or different chromosomes share common transcription factories (Osborne et al., 2004, 2007; Schoenfelder et al., 2010; Hakim et al., 2013). Intriguingly, differential nuclear organizations mediated by cell type-specific interchromosomal and intrachromosomal interactions may be related to the formation of cell type-specific gene regulatory networks (Denholtz et al., 2013).

Histones are major components of chromatin in eukaryotes and play key roles in genome functions, such as DNA replication and transcriptional regulation. The majority of histone proteins are synthesized during the S phase of the cell cycle and are expressed from the replication-dependent histone genes (Osley, 1991). In humans, the replication-dependent histone genes are clustered on two different chromosomes. The large cluster, HIST1, on human chromosome 6 contains 55 histone genes, and the two smaller clusters, HIST2 and HIST3, on human chromosome 1 contain 6 and 3 genes, respectively (Marzluff et al., 2002). Drosophila has a single cluster of ∼110 tandemly arrayed repeat units containing replication-dependent histone genes (H1, H2A, H2B, H3 and H4) (Lifton et al., 1978). Although the five histone genes of the newt Notophthaimus viridescens are also clustered, its common repetitive units are widely separated from each other (Stephenson et al., 1981). It has been shown that the replication-dependent histone gene in Drosophila, zebrafish and mammals forms a nuclear body, which is called a histone locus body (HLB), and the HLB creates microenvironments for efficient transcription and histone pre-mRNA processing, a process that is well-characterized in Drosophila and mammals (Liu et al., 2006; Sawyer and Dundr, 2016; Heyn et al., 2017).

Sea urchin is an invertebrate deuterostome that belongs to a sister group of the chordates. Although the molecular mechanisms for transcriptional regulation underlying the embryogenesis have been well-studied in sea urchin (Davidson et al., 2002; Oliveri and Davidson, 2004; Oliveri et al., 2008), little is known about nuclear organization and its correlation with gene expression in sea urchin. Because of their evolutionary position, it is intriguing to investigate the nuclear organization in sea urchin in order to evaluate evolutionary conservation of the nuclear organization in deuterostomes.

Sea urchin has at least four types of histones, early, late, sperm-specific and cleavage-stage (CS), that are sequentially incorporated into chromatin during sea urchin development (Marzluff et al., 2006). Sperm chromatin displays the longest nucleosome repeat length presently determined (∼250 bp) and contains sperm-specific histone variants of H1, H2A and H2B (Easton and Chalkley, 1972; Spadafora et al., 1976; Keichline and Wassarman, 1977, 1979). Unfertilized egg chromatin is composed of CS histone variants (Herlands et al., 1982; Mandl et al., 1997). After fertilization, sperm-specific histones are replaced by maternally stored CS histones, and maternally inherited CS histone mRNAs are translated during the two initial cleavages (Poccia et al., 1984; Green and Poccia, 1985). The early histones become predominant during the morula and blastula stages, and their expression levels reach their maximums during the morula stage (Maxson and Wilt, 1981; Maxson et al., 1983a). The expression of the early histones becomes undetectable by the gastrula stage, during which time the late histones are expressed (Maxson et al., 1983a,b).

Like the replication-dependent histone genes of other metazoans, the early histone genes of sea urchin are clustered as a tandem array of a set of histone genes (H1, H2A, H2B, H3 and H4) repeated several hundred times (Kedes, 1979). Each histone gene is an independent transcription unit, and the regulatory mechanism for transcription has been well-investigated in the sea urchin Paracentrotus lividus. The 30-bp cis-regulatory element (M30) and the transcription factor MBF-1 are required for the activation of the H2A gene (Alessandro et al., 2002; Cavalieri et al., 2009). The sns5 insulator element and its binding factor, COMPASS-like, regulate the H1 gene expression level (Cavalieri et al., 2009, 2013). These promoter and insulator regions showed H3K9 acetylation at the morula stage, when the transcriptional activity is at its maximum level, while H3K9 dimethylation was observed at the gastrula stage when transcription has already been silenced (Di Caro et al., 2007). Furthermore, during early sea urchin development, the nuclear organization, including nuclear size and nucleosome positioning, changes dramatically (Savić et al., 1981; Fronk et al., 1990). However, since little is known about involvement of intranuclear gene positioning and chromosome dynamics in the gene expression in sea urchin, determining the molecular mechanism behind the accurate transcriptional regulation during such drastic nuclear changes is challenging research.

In this study, we performed a three-dimensional fluorescence in situ hybridization (3D-FISH) analysis of the early histone gene loci in the sea urchin Hemicentrotus pulcherrimus. We found that interchromosomal interactions occurred between the early histone gene loci on separate chromosomes at the morula stage of the sea urchin embryo. Furthermore, these interactions were associated with the transcriptionally active states of the early histone genes. Thus, such dynamic interchromosomal interactions may contribute to the efficient synthesis of early histone mRNA during the morula stage of sea urchin development.

To visualize genomic loci in the nuclei of sea urchin embryos, we first performed 3D-FISH on sea urchin embryos. To visualize telomeres, a digoxigenin (DIG)-labeled (CCCTAA)₃ oligonucleotide was prepared and used as a probe for in situ hybridization. A large number of fluorescent spots were observed in the interphase nuclei of the fixed morula embryos in the 3D-FISH analysis (Fig. 1A, right). Furthermore, when the hybridization was carried out on chromosomal spreads prepared from metaphase-arrested morula embryos, a pair of fluorescent spots was detected at each chromosomal terminus (Fig. 1A, left). These indicated that 3D-FISH could be used to visualize telomeric regions in the sea urchin embryos.

Early histone gene loci have tandem repetitive structures, and each one of the histone genes, H1, H2A, H2B, H3 and H4, are included in each unit. Thus, we prepared multiple DIG-labeled probes that targeted histone genes and performed the 3D-FISH analyses. When the hybridization was carried out on metaphase chromosomal spreads, four pairs of fluorescent spots were detected in each cell (Fig. 1B, left). Moreover, four fluorescent spots were seen in the 3D-FISH analysis of gastrula stage embryos (Fig. 1B, right). Thus, the H. pulcherrimus haploid genome contained two non-allelic loci of histone gene clusters

To investigate the intranuclear positioning of the early histone gene clusters during the early development of sea urchin, 3D-FISH analyses were carried out using fixed embryos from the morula through to the gastrula stages. The number of spots changed during the early development as the nuclear volume reduced (Fig. 2A–C). At the morula stage, the proportions of cells exhibiting four, three and two spots were 31%, 38% and 24%, respectively. In the cells exhibiting three or fewer spots, larger fluorescent spots could be observed, suggesting that two or more early histone gene loci were positioned in proximity of each other, leading to interchromosomal interactions (Fig. 2C). However, we could not distinguish heterologous or homologous interchromosomal interactions in this study. Furthermore, because the embryos underwent a rapid cell cycle progression during cleavage, cells showing more than five spots were also detected, which was probably due to chromosomal replication during S phase (Fig. 2C).

As development proceeded, the peak number of fluorescent spots increased from three to four, and the proportion of cells with four spots reached ∼66% at the gastrula stage (Fig. 2C). Thus, although the early histone gene loci exhibited a marked tendency to colocalize with each other, resulting in interchromosomal interactions, at the morula stage, the frequency of the interchromosomal interactions decreased as the early development of sea urchin progressed.

The intranuclear positioning of genes correlates with their transcriptional state, and transcriptional activation repositions genes from the nuclear periphery to the interior (Kosak et al., 2002; Zink et al., 2004; Ragoczy et al., 2006; Peric-Hupkes et al., 2010) or from the interior to the periphery (Brickner and Walter, 2004; Casolari et al., 2004; Rohner et al., 2013). Furthermore, the differentiation of cells is accompanied by the intranuclear repositioning of specific genes (Ragoczy et al., 2006). Thus, we measured the shortest distances from the fluorescent spots on the early histone gene loci to the nuclear boundary. Although the fluorescent spots were preferentially positioned a long distance away from the nuclear boundary at the morula stage (median=1.61 µm), such a preference disappeared as development proceeded (median=0.67 µm at the gastrula stage) (Fig. 2D). However, since nuclear diameters differed among the early developmental stages, we compared these measurements with the distance frequencies predicted by a random distribution (Fig. 2D, open circles; Table 1). At the morula stage, the distribution of the early histone gene loci was not correlated with that predicted by random positioning (R=0.090) (Table 1). However, both distributions became similar as early development progressed, and they well correlated at the gastrula stage (R=0.998) (Table 1). We also measured the shortest distances from telomere signals to nuclear boundaries and found that the distribution of telomeres was highly correlated with that predicted by random positioning at any stage (Fig. S1; Table 1). These results suggest that nuclear positioning of the early histone gene loci in the nuclear interior at the morula stage of the sea urchin development is regulated.

In sea urchin development, minor zygotic gene activation occurs at the one-cell stage, and major zygotic gene activation occurs at the blastula stage (Tadros and Lipshitz, 2009). To investigate the nuclear dynamics of active transcription during the early development of sea urchin, immunostaining of active RNA polymerase II, in which the C-terminal repeat domains are highly phosphorylated at Ser5, was carried out. At the morula stage, several large foci of the active RNA polymerase II were observed (Fig. 3A). However, these large foci gradually disappeared and a large number of small foci appeared from the unhatched blastula to the gastrula stages (Fig. 3A).

To examine whether the positioning of the early histone gene loci is associated with the dynamics of active RNA polymerase II, we visualized the early histone gene loci and active RNA polymerase II simultaneously. The large foci of active RNA polymerase II colocalized with the early histone gene loci at the morula stage (Fig. 3B). Although all of the early histone gene loci were positive for active RNA polymerase II at the morula stage, there seemed to be no correlation between the interchromosomal interaction and the signal intensity. Many of the early histone gene loci were still positive for active RNA polymerase II at the unhatched blastula stage, but such colocalization was not observed at the gastrula stage (Fig. 3B). Because of the occurrence of major zygotic gene activation at the blastula stage onward, it was difficult to evaluate the precise proportion of the active RNA polymerase II-positive loci. Therefore, to evaluate the colocalization, signal intensity profiles of the early histone gene loci and active RNA polymerase II were measured, and their correlation coefficients were calculated at each developmental stage (Fig. S2). These signal intensity profiles were strongly correlated at the morula stage (median of correlation coefficient=0.95), whereas their correlations markedly decreased at the hatched blastula stage and onward (medians of correlation coefficient=0.78, 0.14, 0.01 and −0.32 at the unhatched blastula, hatched blastula, mesenchyme blastula and gastrula stages, respectively) (Fig. 3C). These results suggest that the positioning dynamics of the early histone gene loci are correlated with their gene expression states during the early development of sea urchin.

In another sea urchin species, P. lividus, the M30 modulator is present between H2A and H3 coding sequences and is involved in the activation of these genes (Cavalieri et al., 2009). Moreover, the sns5 insulator is present between the M30 modulator and the downstream H1 gene promoter and blocks the activation of H1 gene transcription by the M30 modulator (Cavalieri et al., 2009). In H. pulcherrimus, the nucleotide sequences of the M30 modulator and BoxA sequence of the sns5 insulator that are required for the enhancer-blocking activity are conserved (Fig. 4A,B). Furthermore, microinjections of M30 oligonucleotides decrease endogenous H2A and H3 mRNA levels at the morula stage, and microinjection of BoxA oligonucleotides disturb the endogenous sns5 insulator function (Cavalieri et al., 2009).

To examine the involvement of these cis-regulatory elements in the behavior of the early histone gene loci, the double-stranded oligonucleotide of either the M30 modulator or BoxA was microinjected into the fertilized eggs of sea urchin as a competitor, and the effects on the intranuclear positioning of the early histone gene loci and the colocalization with active RNA polymerase II were analyzed (Fig. 4C,D). Compared with the microinjection of an M30 mutant sequence (M30 mut), the microinjection of the wild-type M30 sequence resulted not only in a reduction of the signal intensity of active RNA polymerase II [P<0.01 for the difference between average intensities for the M30 mutant and wild-type (WT) M30, t-test; Fig. 4C,E] but also in an increase in the number of fluorescent spots representing the early histone gene loci [P=0.054 for the fitting of the histogram between the M30 mutant and WT M30 as determined by a multinomial exact test of goodness-of-fit (METG), which very nearly reaches the level of significance; Fig. 4F]. However, compared with the microinjection of the BoxA mutant sequence (BoxA mut), the microinjection of the wild-type BoxA sequence increased the signal intensity of the active RNA polymerase II (P<0.01, t-test; Fig. 4D,G), whereas there was no apparent change in the numbers of early histone gene loci (P=0.806, METG; Fig. 4H). Thus, the inhibition of the M30 modulator-binding protein could lead to a reduction in the expression of the early histone genes and in the frequency of the interchromosomal interactions between the early histone gene loci, suggesting that transcriptional activation mediated by the M30 modulator may be involved in the behavior of the early histone gene loci.

Next, we examined the effects of a specific inhibitor of RNA polymerase II, α-amanitin, on the interchromosomal interactions among the early histone gene loci. Treatment of sea urchin embryos with α-amanitin resulted in a reduction in the amount of active RNA polymerase II that colocalized with the early histone gene loci at the morula stage (P<0.01, t-test; Fig. 5A,B). To confirm the reduction of the early histone gene transcription, we performed real-time quantitative reverse transcription PCR (qRT-PCR) and found that the amount of histone H2A transcripts was decreased by approximately half by the α-amanitin treatment (P<0.01, t-test; Fig. S3). Furthermore, α-amanitin-treated embryos showed reduced frequencies of interchromosomal interactions (P<0.01, METG; Fig. 5C). Thus, transcriptional activity may be related to the interchromosomal interactions of the early histone gene loci, and the frequency of the interchromosomal interactions is affected by even a 50% reduction in the number of histone H2A transcripts.

DNA-FISH analyses of the 5S rRNA gene (5S rDNA) in sea urchin P. lividus have been previously reported (Gornung et al., 2005; Caradonna et al., 2007). Here, we reported for the first time on the nuclear dynamics of intranuclear positioning of gene loci, as assessed through 3D-FISH, in a species of sea urchin. In this study, we determined that there are two non-allelic early histone gene loci per H. pulcherrimus genome. The early histone gene loci were regulated to locate to the interior of the nucleus and exhibited interchromosomal interactions during the morula stage, when early histone gene expression levels are at their maximum.

Not much is known about the spatial organization of the sea urchin genome within the nucleus. The spatial distribution of telomeres has been well-studied in other organisms. In yeast and Drosophila, the centromeres tend to cluster near the nuclear periphery at one side of the nucleus while the telomeres tend to cluster at the opposite side, and such organization is known as the Rabl configuration (Hochstrasser et al., 1986; Funabiki et al., 1993; Jin et al., 1998). By contrast, the Rabl configuration is largely absent in mammalian cells, and the telomeres are widely distributed throughout the nucleus although the distribution is dependent on the cell cycle (Chuang et al., 2004). In plants, unlike animals, species that have large (>4800 Mb) genomes, such as wheat, barley, rye and oats, display the Rabl configuration, while species with genomes <1000 Mb, such as sorghum and rice, do not have this configuration, suggesting that the presence of the Rabl configuration depends on genome size (Dong and Jiang, 1998). In this study, we showed that the intranuclear distribution of sea urchin telomeres was highly correlated with random positioning at any developmental stages (Fig. S1). This indicates that, in the sea urchin, which is a basal deuterostome, the telomeres are distributed uniformly throughout the nucleus, at least during the early development.

In this study, we evaluated sea urchin interchromosomal interactions by counting the number of 3D-FISH spots observed within an individual nucleus. In general, when the distances between two spots were less than 1 µm, the two loci were considered to interact with each other. However, because the 3D-FISH spots of the early histone gene loci were too large to be recognized separately at interaction sites, we could not measure the distances between the two interacting loci.

Microinjections of the M30 oligonucleotide decrease the endogenous H2A and H3 mRNA levels at the morula stage (Cavalieri et al., 2009). Here, the microinjection of the M30 oligonucleotide decreased not only the colocalization of active RNA polymerase II with the early histone gene loci but also the frequencies of the interchromosomal interactions between gene loci. Furthermore, the inhibition of RNA polymerase II by α-amanitin treatment also diminished the interchromosomal interactions between the early histone gene loci. Thus, the interchromosomal interactions between the early histone gene loci may be correlated with the active state of transcription at the loci.

Moreover, because the active RNA polymerase II colocalized with the early histone gene loci even when the four loci were positioned separately (Fig. 5B), the interchromosomal interactions are unlikely to be essential for expression. However, morula-stage embryos undergo cleavage followed by rapid cell cycle progression, which requires large quantities of histones to package the newly synthesized DNA. Therefore, sharing the same RNA polymerase II transcription factory between separate gene loci may increase transcriptional efficiency during the short interphase.

Gene positioning and interchromosomal gene clustering have been studied extensively in yeast. Individual genes often move from the nucleoplasm to the nuclear periphery and physically interact with the nuclear pore complex (NPC) upon transcriptional activation (Brickner and Walter, 2004; Casolari et al., 2004). The cis-acting ‘DNA zip codes’ required for such gene positioning and the interaction with the NPC have been identified as being in the promoter region (Ahmed et al., 2010). Furthermore, genes with identical DNA zip codes cluster together, and transcription factors that recognize the DNA zip code and mediate both targeting to the NPC and interchromosomal clustering was identified (Brickner et al., 2012). It has recently been reported that the molecular mechanism controlling targeting to the NPC is distinct from that controlling interchromosomal clustering and that targeting to the nuclear periphery is a prerequisite for gene clustering (Brickner et al., 2012, 2016). In addition, whereas targeting of genes to the NPC is independent of transcription, interchromosomal clustering requires transcription (Brickner et al., 2016). Here, we showed that the disruption of the M30 modulator-binding protein could lead to a reduction in the expression of the early histone genes and in the frequency of the interchromosomal interactions between the early histone gene loci (Fig. 4C,E,F). It is intriguing to examine whether the M30 modulator protein directly controls the positioning of gene loci and interchromosoml interaction.

The early histone gene loci of Strongylocentrotus purpuratus are present as a single repeat in the haploid genome (Marzluff et al., 2006). However, the presence of two non-allelic histone gene clusters was reported for Lytechinus pictus (Cohn and Kedes, 1979). In the DNA-FISH analysis of metaphase chromosomal spreads, four pairs of fluorescent spots were detected in each H. pulcherrimus cell. This clearly indicates that the H. pulcherrimus haploid genome contains two non-allelic loci of histone gene clusters. Furthermore, when we detected four separate spots representing the histone gene loci in the interphase nucleus of morula embryo, all four loci were positive also for active RNA polymerase II. This indicates that both non-allelic histone gene loci are early histone gene clusters that are highly expressed during the morula stage.

Replication-dependent histone genes are expressed during the S phase of the cell cycle during which the genome is replicating. Interchromosomal interactions between the early histone gene loci of sea urchin were frequently observed in the morula, but such interactions were less frequent in the subsequent developmental stages (Fig. 2C). Although almost all interphase cells are in S phase during the cleavage stage, the proportion of cells in S phase is lower in the subsequent developmental stages. Furthermore, despite the repression of the early histone gene transcription, a proportion of cells do show interchromosomal interaction at the gastrula stage (Fig. 2C). This suggests that the interchromosomal interaction of the sea urchin early histone gene loci may be related not only to the transcriptional activity but also to the S phase of the cell cycle. Alternatively, the interchromosomal interactions observed at the gastrula stage could function as a ‘transcription memory’ as described in yeast (Brickner et al., 2007).

The replication-dependent histone mRNAs have unusual properties. In contrast to most protein-coding genes, the replication-dependent histone genes, including the early histone genes of sea urchin (Maxson et al., 1983a), lack introns and their mRNAs are not polyadenylated, but contain well-conserved stem-loop structures in their 3′ untranslated regions (Dominski and Marzluff, 1999). Furthermore, H. pulcherrimus early histone genes contain stem-loop motifs as well as downstream purine-rich sequences, termed histone downstream elements (Fig. S4). The formation of the 3′ ends of replication-dependent histone mRNAs requires two essential factors, Flice-associated huge protein (FLASH) and the U7 small nuclear ribonucleoprotein (snRNP) (Sawyer and Dundr, 2016). These two factors are concentrated in a nuclear body called the HLB during the G1 phase of the cell cycle and increase the transcription of histone mRNA (Sawyer and Dundr, 2016). Because the colocalization of FLASH, U7 snRNP and other components of the HLB with the early histone gene loci was not examined in this study, we cannot conclude whether the early histone gene/active RNA polymerase II signals observed at the morula stage were related to the HLB. The common features of the early histone genes of sea urchin and other replication-dependent histone genes suggest that the early histone genes may form a nuclear body similar to the HLB to effectively produce a large quantity of histones during the rapid cell cycle at the cleavage stage.

In mammalian cancer cell lines with altered cell ploidy, the number of FLASH-positive HLBs correlates with the number of chromosomes 1 and 6, in which the replication-dependent histone genes are clustered, indicating that the HLB forms at each histone gene locus separately (Bongiorno-Borbone et al., 2008). On the other hand, in Drosophila, all interphase nuclei during the blastoderm stages display one or two HLBs (Liu et al., 2006), indicating homologous interchromosomal interactions between replication-dependent histone gene clusters. Therefore, the interchromosomal interactions between the early histone gene loci observed in the early development of sea urchin may play a role in the efficient biogenesis of histone mRNA during development.

Interchromosomal interactions have been reported for many genomic loci. For example, transient homologous pairing of the mouse Oct4 allele is associated with the transition from pluripotency to lineage specification (Hogan et al., 2015). Random inactivation of one female X chromosome in mammalian females requires a transient interchromosomal pairing at the X inactivation center at the onset of X-inactivation (Xu et al., 2006). Moreover, before differentiation of T helper cells, the T helper cell 2 locus control region on chromosome 11 associates with interferon-γ regulatory regions on chromosome 10 (Spilianakis et al., 2005). Thus, interchromosomal interactions of genomic loci seem to have relevant functions in coordinating gene expression during development and differentiation. The interchromosomal interaction observed in this study may play a role in coordinating the expressions of the sea urchin early histone genes located in allelic and non-allelic clusters.

It has recently been reported that long noncoding RNAs are involved in interchromosomal interactions and nuclear body formation (Nakagawa and Hirano, 2014; Chujo et al., 2016). Furthermore, the importance of histone mRNA synthesis for the formation of HLB has been shown in zebrafish (Heyn et al., 2017). At present, it is uncertain whether noncoding RNA or histone gene transcripts are required for the interchromosomal interactions between the early histone gene loci of sea urchin. However, since a reduction of the histone gene transcript level by approximately half resulted in the reduction of the frequency of the interchromosomal interactions (Fig. 5C; Fig. S3), RNA may have a pivotal role in the dynamics of the early histone gene loci. It will be very interesting to investigate the involvement of RNA in the interchromosomal interactions between the early histone gene loci of sea urchin, and this is the next important subject of our future research.

Adult sea urchins (H. pulcherrimus) were collected from the Seto Inland Sea or Tateyama Bay, Japan. Eggs and sperm were obtained by coelomic injections of 0.55 M KCl. Fertilized eggs were subsequently cultured in filtered seawater at 16°C. To inhibit transcriptions by RNA polymerase II, fertilized eggs were cultured for 6 h in filtered seawater containing 50 µM α-amanitin until the development reached the morula stage.

To prepare competitor DNA for the in vivo competition experiments, two complementary strands of oligonucleotides for M30, M30 mut, BoxA or BoxA mut (Table S1) were incubated in 120 mM KCl at 75°C for 5 min and then cooled slowly. The double-stranded oligonucleotide was mixed with glycerol and injected at a final concentration of 5 ng/µl.

At 4 h post fertilization, the 16- and 32-cell stage embryos were treated with 1.0 mg/ml colchicine for 1 h. The resulting embryos were treated in 8% sodium citrate solution and washed three times with Carnoy's fixative (60% ethanol, 30% chloroform and 10% acetic acid). The fixed embryos were placed onto coverslips coated with 0.1% protamine sulfate and dried at room temperature.

To prepare the early histone gene probe solution, DIG-labeled DNA probes were synthesized by using the DIG DNA labeling mix (Roche, Switzerland) by PCR with the primers listed in Table S1. After purification, DIG-labeled DNA probes were mixed in hybridization buffer A [10% dextran sulfate, 70% formamide and 160 ng/µl salmon sperm DNA in 2× saline sodium citrate (SSC)]. For the telomere probe solution, the (CCCTAA)₃ oligonucleotide modified with DIG at the 5′ end was chemically synthesized (Fasmac, Japan) and mixed in hybridization buffer B (10% dextran sulfate and 70% formamide in 2× SSC).

For metaphase chromosomal spreads, the coverslips were washed three times with 2× SSC and then treated with 0.1 mg/ml RNaseA in 2× SSC for 1 h at 37°C. After washing three times with 2× SSC at room temperature, the coverslips were heated to 80°C for 1 min in 50% formamide in 2× SSC to denature the chromosomal DNA. Immediately after the denaturation, coverslips were placed on ice for 5 min and incubated in probe solution at 37°C overnight. Coverslips were washed three times with 2× SSC at room temperature, mounted on slides with SlowFade Gold Antifade Mountant with DAPI (Thermo Fisher Scientific, USA), and sealed with nail polish in preparation for microscopic imaging.

For 3D-FISH of interphase nuclei, sea urchin embryos were fixed with 4% paraformaldehyde, 32.5% filtered seawater, 32.5 mM MOPS pH 7.0 and 162.5 mM NaCl at 4°C overnight. Fixed embryos were washed three times with 1× phosphate-buffered saline (PBS) at room temperature and then treated with 0.5% Triton X-100 in 1× PBS for 20 min at room temperature. Embryos were washed three times with 2× SSC at room temperature and treated with 0.2 mg/ml RNaseA in 2× SSC for 1 h at 37°C. After washing three times with 2× SSC at 42°C, embryos were incubated at 80°C for 5 min in probe solution to denature the genomic DNA. Embryos were immediately placed on ice for 5 min and then incubated at 37°C for 3 days. Embryos were washed three times with 2× SSC at room temperature and were placed onto 0.1% protamine sulfate-coated coverslips. Coverslips were mounted on slides with SlowFade Gold Antifade Mountant with DAPI (Thermo Fisher Scientific, USA) and sealed with nail polish in preparation for microscopic imaging.

For combination of 3D-FISH and immunofluorescence staining, hybridization for 3D-FISH was first performed as described above, and then embryos were washed three times with 1× PBS at room temperature. After treatment with 1% bovine serum albumin (BSA) in 1× PBS for 1 h at room temperature, embryos were incubated with the primary antibody at 4°C overnight. After washing three times with 1× PBS at room temperature, embryos were incubated with the secondary antibody at 37°C for 90 min. After washing with 1× PBS, embryos were mounted as described above for 3D-FISH.

For immunofluorescence, fixed embryos were washed three times with 1× PBS at room temperature and then treated with 0.5% Triton X-100 in 1× PBS for 20 min at room temperature. After washing three times with 1× PBS at room temperature, embryos were treated with blocking solution (1% BSA in 1× PBS) for 1 h at room temperature and then incubated with primary antibody at 4°C overnight. After washing three times with 1× PBS at room temperature, embryos were incubated with secondary antibody at 37°C for 90 min. After washing with 1× PBS, embryos were mounted as described above for 3D-FISH.

Primary antibodies used in this study were mouse anti-digoxigenin antibody (21H8) (Abcam; ab420, 1:500 dilution) and rabbit anti-RNA polymerase II C-terminal repeat domain YSPTSPS (phosphorylated Ser5) antibody (Abcam; ab5131, 1:500 dilution). The secondary antibodies used in this study were Alexa Fluor^® 488-conjugated chicken anti-rabbit IgG (H+L) (Molecular Probes; A-21441, 1:1000 dilution) and Alexa Fluor^® 594-conjugated goat anti-mouse IgG (H+L) (Molecular Probes; A-11032, 1:1000 dilution).

Images were acquired with an LSM700 confocal microscope (Carl Zeiss, Germany) with laser illuminations at 405, 488 and 555 nm, and were analyzed by using ZEN software (Carl Zeiss, Germany). Sequential z-axis images were collected in 0.4-µm steps. Acquired images were processed with a 3D median filter with ZEN software. The detection of fluorescent spots within a nucleus was performed by using the ‘Detect Cell and Vesicle’ function of Imaris (Bitplane) in which the vesicle diameter was set at 0.5 µm (semi-automatic detection). The nuclear volume was estimated from DAPI signals using the ‘Cell’ function of Imaris (Bitplane), with the Cell Smooth Filter Width parameter set at 0.1 µm. The numbers of early histone gene signals within a nucleus were determined by counting the vesicles within a cell.

To evaluate the colocalization between the early histone gene loci and active RNA polymerase II, a single optical section image that exhibited the most intense signal from the early histone gene loci was selected from z-stack images processed with the 3D median filter. The signal intensity profiles of the early histone gene loci and active RNA polymerase II were measured using the ‘profile’ function of the ZEN software. A relative fluorescence intensity of the early histone gene loci over 0.1 was regarded as an actual signal. The correlation between the two signal intensity profiles was calculated by determining the Pearson's correlation coefficient. The intensity of active RNA polymerase II on early histone gene loci in the maximum projection image was measured in Fiji (https://fiji.sc/). R software (http://www.r-project.org) was used for t-tests and multinomial exact test of goodness-of-fit tests (XNomial package) to evaluate the difference in intensity and localization of histone gene loci between WT and mutant embryos.

To examine the amounts of histone gene transcript in control and α-amanitin-treated embryos, total RNA was extracted from embryos at the morula stage by using ISOGEN (Nippongene, Japan) as described in the instruction manual. After purification with a RNeasy Mini kit (QIAGEN, Germany), the total RNA was reverse transcribed by using a ReverTra Ace^® qPCR RT Master Mix with gDNA Remover (TOYOBO, Japan). The qRT-PCR was performed with KOD SYBR^® qPCR Mix (TOYOBO, Japan) and the StepOnePlus™ Real-Time PCR System (Applied Biosystems) using the primers listed in Table S1. To normalize the level of the early histone H2A gene transcript, mitochondrial cytochrome oxidase subunit I (HpMitCOI) mRNA encoded by the mitochondrial genome of H. pulcherrimus was used, because mitochondrial RNA polymerase is insensitive to α-amanitin (Reid and Parsons, 1971; Tsai et al., 1971).

We thank Dr Masato Kiyomoto (Tateyama Marine Laboratory, Ochanomizu University) for supplying live sea urchins. Fluorescence images were acquired in the Gene Science Division, Natural Science Center for Basic Research and Development, Hiroshima University and analyzed in the Research Center for the Mathematics on Chromatin Live Dynamics, Hiroshima University. The H. pulcherrimus early histone gene used in this study was originally cloned by Dr Hiraku Shimada and Dr Koji Akasaka (Hiroshima University).