Adult Hemicentrotus pulcherrimus were collected around Shimoda Marine Research Center, University of Tsukuba, and around the Marine and Coastal Research Center, Ochanomizu University. This species was collected under the special harvest permission of prefectures and Japan Fishery cooperatives. Gametes were collected by the intrablastocoelic injection of 0.5 M KCl, and the embryos/larvae of H. pulcherrimus was cultured at 15°C, in glass beakers or plastic dishes that contained filtered natural seawater (FSW) with 50 μg/mL kanamycin.

Whole-mount in situ hybridization was performed as described previously (Yaguchi and Katow, 2003) with some modifications. cDNA mix from several embryonic stages was used to make RNA probes based on the H. pulcherrimus genome and transcriptome (Kinjo et al., 2018). The samples were incubated with RNA probes for foxQ2 (HPU_15608), rx (HPU_04689), hbn (HPU_04688), lefty (HPU_15030), neurogenin (HPU_07864), choline acetyltransferase (HPU_01496), msi1 (HPU_03784) and tryptophan 5-hydroxylase (tph; HPU_21307) (Ayala et al., 2007) at a final concentration of 0.4–1.2 ng/μL at 50°C for 5 days. The probes were detected with the Tyramide Signal Amplification Plus System (TSA; Akoya Biosciences, Marlborough, MA, United States) or Alkaline phosphatase-based chromogenic system described previously (Erkenbrack et al., 2019).

Whole-mount immunohistochemistry was also performed as described previously. The samples were blocked with 1% skim milk in PBST for 1 h at RT and incubated with primary antibodies (dilutions: mouse anti-Synaptotagmin B (SynB) (Nakajima et al., 2004), 1:100; mouse anti-ChAT (9), 1:100; rabbit anti-serotonin (#S5545; Sigma-Aldrich), 1:1000) overnight at 4°C.

Microinjection was performed according to a previously described method (Yaguchi, 2019) with injection buffer (24% glycerol, 20 mM HEPES pH 8.0 and 120 mM KCl). The morpholino (Gene Tools, Philomath, OR, United States) sequences and the in-needle concentrations in injection buffer were as follows:

Rx-MO1 (1.9–3.8 mM): 5′- GGG​TGA​TGC​GCT​CCA​TCC​ATT​GTT​A -3′,

Rx-MO2 (1.0–1.9 mM): 5′- TTT​GTG​ACT​GAT​CGT​CTT​TCC​AAA​C -3′,

Msi1-MO1 (0.5–1.0 mM): 5′- AAC​CCT​CAA​CTA​AAA​AGG​CCC​AAT​A-3′,

Msi1-MO2 (1.9 mM): 5′- GAA​TTG​GCA​AAC​GGT​CCT​TCT​TAA​C-3′,

and Hbn-MO1 (0.7 mM): 5′- AAA​ATG​AAC​GGA​ACA​AGT​CCA​GTG​T -3’.

(previously characterized) (Yaguchi and Yaguchi, 2019).

Two non-overlapping translation-blocking morpholinos for Rx and Musashi were used to confirm the specificity of their function (Supplementary Figures S1, S2). For negative control experiments, we injected random MO (2.0 mM: Gene Tools, Supplementary Figure S1) or injection buffer only.

Live samples were observed under a light/fluorescence microscope (IX70, Olympus, Tokyo, Japan). The fixed and stained specimens were observed using a light/fluorescence microscope (IX70, Olympus) and a confocal laser scanning microscope (FV10i, Olympus). All transmission images were taken with an IX70 microscope. The Figure panels and drawings for the Figures were made using Adobe Photoshop and Microsoft PowerPoint.

Previous research has demonstrated the presence of non-serotonergic neurons in the brains of sea urchin larvae (Nakajima et al., 2004). In the species L. variegatus, these neurons were found to be cholinergic (Slota and McClay, 2018). In H. pulcherrimus, one of the commonly studied sea urchins in the Western Pacific, serotonergic neurons (serotonin+/synaptotagminB [SynB]+) appear in the brains of prism larvae at 36 h (hr) (Figures 1A, A′). Non-serotonergic neurons (serotonin-/SynB+) begin to emerge in the early pluteus larvae (Figure 1B). The number of these non-serotonergic neurons increases as development progresses (Figures 1C, C').

To verify whether these non-serotonergic neurons in H. pulcherrimus are also cholinergic, similar to L. variegatus, we detected the messenger RNA (mRNA) of choline acetyltransferase (ChAT) and tryptophan 5-hydroxylase (TPH), a rate-limiting enzyme of serotonin synthesis, in the larvae and examined the details using a confocal microscope. Besides the previously reported expression of chat in ciliary band neurons in H. pulcherrimus larvae (Yaguchi et al., 2022), chat signals were also detected in the brain region at 72 h (Figures 1D, E), as observed in L. variegatus (Slota and McClay, 2018). The presence of ChAT protein in these neurons was confirmed using specific antibodies (Figures 1F–F″).

Additionally, we observed the presence of one to three cholinergic neurons at the edge of the upper lip of the larval mouth, distinct from the brain neurons (yellow arrowheads in Figure 1F, F′). However, due to the complexity of axons and/or dendrites in the brain and mouth regions, tracing the precise spatial patterns of cholinergic neurons and their connections with other neurons in and around the brain region was challenging (Figures 1C, C'). Therefore, we conducted a more precise investigation of the spatiotemporal patterns of cholinergic neurons by observing neurogenic markers during the early pluteus stages. Since Synaptotagmin B (SynB) represents all neurons and their axons/dendrites, and most SynB-neurons at this developmental stage are cholinergic, except for serotonergic neurons in the brain (Figures 1D, E), we initially used Serotonin and SynB antibodies to elucidate the neural patterns at 60 h, 72 h, and 96 h larvae, with a particular focus on the brain and upper lip region.

In 60-h early pluteus larvae, cholinergic neurons were found to be located near serotonergic neurons (Figures 1G–G″). A lateral view, providing a longitudinal optical section, revealed that cholinergic and serotonergic neurons aligned in a single row along the left-right body axis in the brain region (Figures 1H–H″). By 72 h pluteus larvae, the majority of cholinergic neurons appeared to have shifted more towards the ventral side rather than the dorsal side (Figures 1I–I''), which was supported by the lateral view (Figures 1J–J″). This ventral shift of cholinergic neurons became even more pronounced by 96 h pluteus larvae, resulting in the detection of cholinergic neurons only on the ventral side and serotonergic neurons exclusively on the dorsal side (Figures 1K–K″). Detailed observations of optical sections from ventral to dorsal side using a confocal microscope (Figures 1L–L″) revealed that the tips of cholinergic neurons leaned towards the ventral direction, with a complex of neural processes located at the basal side of the epithelial layer (Figures 1M–N‴).

In addition to the cholinergic neurons in the brain (Figure 1N', white arrowheads), cholinergic neurons in the upper lip (Figure 1N'', yellow arrowheads) extended their processes towards the basal complex (Figures 1N''', O). Figure 1P summarizes the patterning of brain serotonergic and cholinergic neurons, as well as the upper lip cholinergic neurons, in 4-day larvae. Based on our observations, two sets of neurons are arranged in the brain of sea urchin larvae: serotonergic neurons on the dorsal side and cholinergic neurons on the ventral side in H. pulcherrimus. Initially, both types of neurons are situated in a single row on the dorsal side of the larvae, and cholinergic neurons appear to shift towards the ventral side during development.

Retinal homeobox (Rx) is a specific transcription factor expressed in the prospective brain region during early development and in some neural progenitor cells in other sea urchin species (Burke et al., 2006; Yaguchi et al., 2012). However, its function remains unknown, although a previous study has suggested a role for Rx in the differentiation of serotonergic neurons based solely on its expression pattern (Wei et al., 2009). Given the importance of Rx in brain formation in other organisms, we hypothesize that it plays a crucial role in sea urchin brain specification. Therefore, we investigated the expression and function of Rx in H. pulcherrimus.

At 18 h in mesenchyme blastulae, mRNA of rx was found to be co-expressed with foxQ2 (Figure 2A). As FoxQ2 is known to be a regional specifier for brain regions in sea urchin embryos/larvae, the co-expression of foxQ2 and rx strongly suggests that FoxQ2 regulates the expression of rx. Indeed, when we attenuated FoxQ2 function using morpholino antisense oligonucleotide (MO), the expression of rx was completely abolished in the brain region (Supplementary Figures S3A, B). While the foxQ2 expression region gradually becomes restricted to the anterior end of embryos as development progresses (Tu et al., 2006; Yaguchi et al., 2008), the rx region remains not only at the anterior end but also in the more dorsal region, similar to the expression pattern of homeobrain (hbn) at 24 h in early gastrula (Figures 2B–D) (Yaguchi and Katow, 2003). We detected rx signal in cells that also expressed tryptophan 5-hydroxylase (tph) at 36 h in larvae (Figure 2E), but the rx signal disappeared in tph-positive cells by 48 h (see Supplementary Figure S5I). Since 36 h is close to the onset of serotonin synthesis (Yaguchi et al., 2000; Ayala et al., 2007), it is plausible that Rx is involved in the differentiation of serotonergic neurons in H. pulcherrimus.

To investigate the precise functions of Rx in the development of serotonergic neurons, we conducted knockdown experiments by injecting Rx-MO. As a result, we did not observe any disturbances in the development of serotonergic neurons in Rx morphants (Figures 2K–M). On the other hand, the number of brain cholinergic neurons was significantly reduced in these morphants (Figures 2K'–M, white arrowhead) compared to controls (Figure 2G') (Supplementary Table S1). Larvae injected with a random morpholino as a control did not show any effects on the development of cholinergic neurons (Supplementary Figures S1C, D). Moreover, Rx-MO2 morphants exhibited identical phenotypes to Rx-MO1 morphants, providing further support for the requirement of Rx in the development of cholinergic neurons in the brain. Strikingly, this phenotype is completely opposite to that of Hbn morphants, in which serotonergic neurons are absent while cholinergic neurons remain (Figures 2N–Q) (Yaguchi and Katow, 2003), despite the almost identical expression patterns of both genes. Rx morphants displayed shorter arms than controls, resembling the phenotype of Hbn morphants (Figures 2F, J, N). Our observations confirmed that the remaining non-serotonergic neurons in the brain region of Hbn morphants were cholinergic (Figures 2O–Q). Notably, cholinergic neurons at the edge of the upper lip and in the non-brain ciliary band region were barely affected in both Rx and Hbn morphants (yellow and orange arrowheads in Figures 2I, M, Q).

Although rx and hbn are co-expressed during early stages (Figure 2C), it is likely that these transcription factors target different genes, at least in the development of sea urchin larval brains. In a previous study using L. variegatus, it was reported that the neurogenin ortholog (Lv-ngn) is expressed throughout the ciliary band, including the brain, and Lv-ngn is necessary for the specification of cholinergic neurons in the ciliary band (Slota and McClay, 2018). We also detected the expression of Hp-ngn throughout the ciliary band and brain from 48 h to 60 h in H. pulcherrimus (Supplementary Figures S4A–D). In early pluteus larvae of H. pulcherrimus, the ngn-positive cells in the brain express chat but not tph, similar to L. variegatus (Supplementary Figure S4E–E‴, F–F‴).

Musashi is an RNA binding protein that is expressed in neural progenitor cells and plays a role in the development of the brain and central nervous system in various organisms (e.g.,25,26). Previous studies have suggested that Musashi is involved in the differentiation of cholinergic neurons, as it is co-expressed with chat in these organisms (Sakakibara et al., 2001; Higuchi et al., 2008; Perry et al., 2012). Sea urchins also possess a Musashi1-like gene (referred to as Msi1) in their genomes (Sodergren et al., 2006; Kinjo et al., 2018; Arshinoff et al., 2022); however, the spatial expression pattern of this gene in larvae has not been reported thus far. Therefore, to elucidate the expression pattern of msi1 in sea urchin larvae, we performed whole-mount in situ hybridization to detect msi1 mRNA expression. In 36-h prism larvae, msi1 was found to be expressed exclusively in the archenteron, which is not the focus of this study (Figure 3A). In 48-h pluteus larvae, msi1 expression was initiated in the brain and continued throughout the pluteus stage (Figures 3B, C). msi1 was co-expressed with foxQ2 (Figure 3D), a determinant of the brain (Yaguchi et al., 2008), but not with rx and tph at this stage (Figures 3E, F). Furthermore, msi1 was expressed in chat-positive cells in the brain region (Figure 3G).

Next, we focused on the relationship between msi1 and rx. As rx initially expresses in the same region as foxQ2 (Figures 2B, B′), it is possible that Rx is involved in the induction of msi1. To investigate this, we injected Rx morpholino and examined the expression pattern of msi1. In Rx morphants, there was no detectable msi1 signal in the brain (Figures 3H, I), and the percentage of msi1-positive larvae was 95.9% (70/73) in the control group and 3.3% (2/61) in the Rx morphant group. In conclusion, Rx regulates msi1 transcription in the brain region of sea urchin larvae, although it remains unclear whether this control is direct or indirect.

Msi1 plays a crucial role in the expression of Chat in the larval brain and the formation of the neural plexus.

To investigate the function of Msi1, we examined the neural pattern of the brain in Msi1 morphants. Compared to the control group, Msi1 morphants exhibited shorter preoral arms (Figures 4A, E). However, similar to the control group, non-serotonergic and serotonergic neurons were present on the ventral and dorsal sides of the brain, respectively (Figures 4B–B‴, F–F‴). In contrast, the axons and dendrites from non-serotonergic neurons were rarely detected in the brain of Msi1 morphants at 96 h (dot-lined in Figures 4C, G), and the chat signal was absent in the brain (Figures 4D, H, Supplementary Table S1). Injecting Msi1-MO2 into embryos yielded the same phenotype as Msi1-MO1 morphants (Supplementary Figures S2A–D), providing further support for the specificity of the morpholino. These findings indicate that while non-serotonergic neurons are present in the brain of Msi1 morphants, they fail to express chat and extend axons and dendrites. Notably, there was no effect on the expression of chat in the upper lip neurons, suggesting that Msi1 is involved in the differentiation of cholinergic neurons specifically in the brain. Interestingly, when we inhibited the function of Rx, the presence of ngn-positive cells was significantly reduced. However, in Msi1 morpholino-treated larvae, the ngn-positive cells appeared to be unaffected and were observed in a normal pattern (Figures 4I–K, white arrowheads).

To investigate the interplay between different neuronal types in and around the brain, we conducted a detailed analysis of tph, chat, rx, zinc finger homeobox 1 (zfhx1), and Go-opsin expression patterns in larvae aged 48–96 h (Figure 5, Supplementary Figure S5). At 60 h, rx signals were observed in bilateral ventral clusters adjacent to the brain, consistent with previous findings in the sea urchin species Strongylocentrotus purpuratus (Valencia et al., 2021) (Supplementary Figures S5J–L). Interestingly, these rx-positive cells in the bilateral clusters did not co-express chat, a marker for cholinergic neurons (Figure 5A), but they did co-express zfhx1, which is known to play a role in specifying neurons in sea urchins (Yaguchi et al., 2012), as well as SynB (Figures 5B, C). Additionally, we confirmed the expression of Go-opsin in these rx-positive cells in the bilateral ventral clusters of H. pulcherrimus (Figures 5D–F, yellow arrows), consistent with the findings in Strongylocentrotus purpuratus (Valencia et al., 2021). These bilateral clusters are prominently identified in immunohistochemistry using SynB antibody (Figure 5G, yellow asterisks). To visualize the morphology of neurons, including neural processes, immunohistochemistry is much better than in situ hybridization since morphology of sea urchin larvae is shrunk during in situ hybridization treatment. Axons from these bilateral clusters are connected to brain neurons (Figure 5G, arrowheads). The bilateral clusters were disappeared in the Rx morphants, supporting the previously reported data (Valencia et al., 2021), although they were present in Msi morphants (Figures 5H, I). In addition to Valencia (2021), in which Rx morphants lost Go-Opsin in the bilateral clusters, the results shown here indicate that Rx is not only required for the specification of the photoreceptor clusters, but also for the characterization of them.