Abstract

   During development, a group of cells called organizers plays critical
   roles by sending signals to adjacent cells and controlling embryonic
   and tissue patterning. Recent studies suggest that these inductive
   cells facilitate the downstream signaling pathways conserved across
   organisms. However, what makes these cells fundamentally inductive is
   little understood. In this study, we demonstrate that the micromeres of
   the sea urchin, one of the known organizers, have distinct metabolic
   properties compared to the rest of the embryo. The specific metabolic
   inhibitors for sugar metabolism (2-DG), fatty acid synthesis
   (cerulenin), and N-linked glycosylation (tunicamycin) compromise
   micromeres’ regulatory capacity, altering the downstream germ layer
   patterning in the resultant embryos. Notably, the endoplasmic reticulum
   (ER) asymmetrically localizes during asymmetric cell division,
   resulting in the enrichment of ER and Wnt protein at the vegetal cortex
   of micromeres. Metabolic inhibition appears to compromise ER activity
   in Wnt particle distribution. We propose that the micromere ER is
   sensitive to specific metabolic regulation, contributing to the
   inductive signaling activity. This study provides a paradigm of how ER
   and metabolic regulation contribute to the inductive capability of the
   cells.

   Subject terms: Metabolism, Embryonic induction, Endoplasmic reticulum,
   Gastrulation
     __________________________________________________________________

   During sea urchin development, the micromeres form via asymmetric cell
   division and act as organizers to induce gastrulation. Here they show
   that micromeres have unique metabolic characteristics and that
   disrupting their metabolic activity blocks gastrulation induction.

Introduction

   Axis formation during early development represents one of the most
   fundamental challenges faced by the embryo. Without correctly
   establishing up, down, front, back, left, and right, an organism would
   have no way of faithfully constructing a functional adult. Many of the
   genetic and signaling mechanisms, including gene and protein
   expressions, are involved in the process. However, we are still far
   from a complete understanding of how the process works. One such
   example is the organizer’s function during embryogenesis. Organizers
   are specialized groups of cells that signal to adjacent cells to direct
   their cell fate, thereby governing the complex morphogenesis of
   tissues. The most well-known example is Spemann-Mangold’s organizer in
   amphibian embryos, which was discovered over a century ago^[42]1. The
   similar organizer function of micromeres in the sea urchin embryo was
   also found around the same time^[43]2–[44]5. Since its initial
   discovery, organizers or organizer-like inductive cells have been
   reported to play critical roles in embryonic patterning, tissue
   morphogenesis, and tissue regeneration in various multicellular
   organisms. Further, more recent studies suggest that these cells drive
   conserved signaling pathways downstream, such as Wnt and hedgehog
   pathways^[45]6–[46]11. However, what makes these cells fundamentally
   unique and inductive is yet to be understood. In this study, we use
   micromeres of the sea urchin embryo as a model system to address the
   remaining question in developmental biology.

   Micromeres are formed through a single asymmetric cell division at the
   16-cell stage in the sea urchin embryo. Micromeres have a number of
   unique features compared to the rest of the blastomeres in the embryo,
   such as a smaller size, slower cell cycle, increased Ca^2+ level,
   distinct membrane properties, autonomously committed cell lineages, and
   a unique set of gene and protein expressions^[47]12–[48]20.
   Furthermore, as soon as they are formed at the 16-cell stage,
   micromeres start functioning as organizers. When micromeres are placed
   at an ectopic location, such as at the animal cap at the 16-cell stage,
   a secondary axis is induced, resulting in the formation of a secondary
   gut at gastrulation. Therefore, micromeres are capable of signaling to
   adjacent cells and inducing the endomesoderm lineage even at an ectopic
   location in the embryo^[49]21,[50]22.

   It is still largely unknown how micromeres gain these unique properties
   at the 16-cell stage through a single asymmetric cell division.
   However, some of the properties described above (e.g., unique membrane
   composition and Ca^2+ elevation in micromeres) are closely linked to
   metabolic regulation^[51]13,[52]20. With the hypothesis that micromeres
   undergo distinct metabolic regulation to gain their inductive
   capability, this study performed comprehensive metabolomics at the
   16-cell stage in the presence or absence of specific metabolic
   inhibitors. The results demonstrate that micromeres are enriched with
   lipids and sugar metabolites. Perturbation of these metabolic pathways
   by sugar metabolism and fatty acid synthesis inhibitors compromised
   micromeres’ inductive signaling, resulting in failed embryonic
   patterning. We also found that the endoplasmic reticulum (ER) is
   asymmetrically enriched toward the micromere side of the spindle during
   asymmetric cell division, and metabolic inhibition appears to
   compromise the regulation of signaling molecules. Therefore, unique
   metabolic regulation and enriched ER in micromeres appear to contribute
   to efficient inductive signaling at the 16-cell stage. Such unique
   metabolic activities may also serve as fundamental mechanisms of
   inductive cell signaling in other organisms.

Results

Micromeres have a unique metabolic profile

   To test if micromeres have distinct metabolic properties, we performed
   metabolomics for micromeres and non-micromeres at the 16-cell stage of
   sea urchin (Strongylocentrotus purpuratus; Sp) embryos, using gas
   chromatography-mass spectrometry (GC-MS). For this experiment, embryos
   were dissociated and fractionated into micromere and non-micromere
   populations just after micromere formation, using a sucrose gradient
   (Fig. [53]1A). Each population was then subjected to metabolome
   analysis via GC-MS. Of note, three biological replicates used in this
   experiment are derived from three pairs (batches) of sea urchins from
   wild populations in the ocean. Therefore, strong genetic variations
   across batches are expected, yet unexpectedly, the resultant analyses
   consistently showed a clear separation between micromeres and
   non-micromere groups across biological replicates (Fig. [54]1B, C;
   Data [55]S1).

Fig. 1. Metabolomic characterization of micromeres.

   [56]Fig. 1
   [57]Open in a new tab

   A Sample preparation schematic showing a representative whole embryo at
   the 16-cell stage, just after micromere formation. These embryos were
   dissociated and separated over a sucrose gradient to generate a
   micromere-enriched fraction and a non-micromere fraction. The resulting
   fractions were then processed for Gas Chromatography and Mass
   Spectrometry analysis (GC-MS). The experiment was performed in
   triplicate, resulting in six samples. B Partial least squares
   discriminant analysis (PLS-DA) on the six samples shows a separation of
   the micromere and non-micromere samples. C Heatmap visualizing the
   differentially identified metabolites between the micromere and
   non-micromere samples. Red indicates higher abundance, and blue
   indicates lower abundance based on the GSEA statistic: −log10(p-value)
   * sign(log2(FC)). Rows (individual metabolites) were ordered by GSEA
   score, with each metabolite annotated by a barcode indicating
   associated pathways. D Significantly enriched pathways in micromeres or
   non-micromeres are shown with their respective adjusted p-values from
   GSEA and directionality.

   The pathway enrichment analysis on metabolites detected in this
   analysis shows that the micromere-enriched population exhibits more
   significant activities associated with fatty acid synthesis
   (Fig. [58]1d). Further, individual metabolite analysis suggests that
   multiple sugar metabolites are enriched in micromeres (Fig. [59]S1a).
   The processes found to be less enriched in micromeres are
   overwhelmingly related to amino acid metabolism. These results appear
   to be driven by higher levels of many amino acids in non-micromeres
   (Figs. [60]1D and [61]S1b). Micromeres exhibit a higher rate of protein
   synthesis relative to non-micromeres^[62]23. As such, they are likely
   to have lower levels of free amino acids compared to other cells since
   early embryos heavily rely on maternally loaded products, including
   metabolites, to support their rapid development (e.g., every 40 minutes
   of cell divisions during embryogenesis of the sea urchin). This may be
   the reason amino acid metabolism is enriched in non-micromeres, yet
   this needs to be further investigated in the future. Notably, the
   metabolites involved in ABC transporters are also found to be less
   enriched in micromeres, which is consistent with previous reports
   (Fig. [63]1D)^[64]24,[65]25. These results suggest that micromeres
   indeed have quite distinct metabolic profiles compared to the rest of
   the cell lineages. Furthermore, the enrichment of fatty acid synthesis
   and sugar metabolism in the micromeres led us to hypothesize that these
   metabolic processes may be essential for micromere function.

Metabolic inhibition during micromere formation compromises embryonic
patterning by Day 2

   To test the above hypothesis, we next treated embryos with inhibitors
   of fatty acid synthesis and sugar metabolism that are enriched in
   micromeres (Fig. [66]2a). These include Cerulenin (Cer) and
   2-deoxy-D-glucose (2-DG), which inhibit fatty acid synthesis and
   glucose/sugar metabolism, respectively. Furthermore, we included sodium
   azide (NaN[3]), an inhibitor of mitochondrial respiration, to compare
   the results, as oxidative phosphorylation (Oxphos) is less enriched in
   micromeres. Embryos were acutely treated with each inhibitor during the
   4th cell division, which gives rise to micromeres (Fig. [67]2b). This
   treatment scheme allowed us to determine whether inhibition of these
   processes impacted micromere formation, function, or both. In almost
   all cases, micromere formation was not found to be impaired by these
   inhibitors (Fig. [68]2c). The only inhibitor that appeared to inhibit
   micromere formation was the mitochondrial respiration inhibitor NaN[3]
   (Fig. [69]2d), which is reported to arrest cell cycle progression in
   some cases^[70]26,[71]27. Upon further examination of these
   NaN[3]-treated embryos, most of them formed micromeres, albeit in a
   delayed fashion under this treatment condition.

Fig. 2. Inhibitor treatment during the micromere formation compromises
gastrulation at Day 2.

   [72]Fig. 2
   [73]Open in a new tab

   a A summary diagram of inhibitor experiments conducted in this study.
   Metabolites are in blue, enzymes are in green, inhibitors are in brown,
   and the resultant phenotypes are in red. b An experimental flow of the
   inhibitor experiments. Each inhibitor was added prior to micromere
   formation and washed five times with seawater after 30 minutes of
   incubation. c and d The representative images of 16-cell stage embryos
   for each treatment group are shown in (c). The proportion of the
   embryos with micromeres (arrow) is shown in graph (d). A total of 120
   embryos were counted for each group. The adjusted p-value between the
   control (DMSO) and each treatment group is NaN[3], p = 0.038; 2-DG,
   p = 0.32; Cerulenin, p = 0.213. e–g The completed gastrulation (stars)
   and skeleton formation (arrows) are seen in the control and
   NaN[3]-treated embryos, while they are compromised in the 2-DG and
   cerulenin groups (d). The % of embryos fully developed after treatment
   with various doses of each inhibitor (g) is shown in the graph (f). The
   total number of embryos scored (n) in each group is DSMO L, n = 92;
   DMSO M, n = 101; DMSO H, n = 199; NaN[3] L, n = 124; NaN[3] M, n = 115;
   NaN[3] H, n = 167; 2-DG L, n = 103; 2-DG M, n = 147; 2-DG H, n = 239;
   Cerulenin L, n = 147; Cerulenin M, n = 213; Cerulenin H, n = 207. The
   adjusted p-value between the control (DMSO L) and each treatment group
   is NaN[3] L, p = >0.99; NaN[3] M, p = 0.12; NaN[3] H, p = 0.005; 2-DG
   L, p = >0.99; 2-DG M, p = <0.0001; 2-DG H, p = <0.0001; Cerulenin L,
   p = >0.99; Cerulenin M, p = 0.14; Cerulenin H, p = <0.0001. (h) An
   experimental flow of the rescue experiments. i, j The addition of
   glucose and palmitic acid (rescue reagents) to 2-DG- and
   cerulenin-treated embryos, respectively, rescued the developmental
   failure characterized by the completion of gastrulation (stars) and
   skeleton formation (arrows) (i). The embryos were pre-treated with each
   rescue reagent at 1×, 2.5×, or 5× the dose of the corresponding
   inhibitor, and the percentage of fully gastrulated embryos is shown in
   the graph (j). DSMO No Res, n = 61; DSMO Res 1×, n = 53; DSMO Res 5×,
   n = 79; 2-DG No Res n = 119; 2-DG Res 1×, n = 112; 2-DG Res 5×, n = 91;
   Cerulenin No Res, n = 88; Cerulenin Res 1×, n = 72; Cerulenin Res 5×,
   n = 54. The adjusted p-value between the control (No Res) and each Res
   group is Res 1×, p = 0.38; Res 2.5×, p = 0.27; Res 5×, p = 0.34 for the
   DMSO group, Res 1×, p = <0.0001; Res 2.5×, p = 0.0005; Res 5×,
   p = 0.0118 for the 2-DG group, and Res 1×, p = <0.0001; Res 2.5×,
   p = 0.0003; Res 5×, p = 0.0002 for the Cerulenin group. Across the
   experiments in this figure, the dose of each inhibitor is directly
   indicated in the image or the graph. Individual dots in the column of
   graphs suggest an independent cycle of experiments. All graph data are
   presented as mean values ± SD. Significant results from One-way ANOVA
   (d, f) or Mixed effects analysis (j) with Dunnett’s multiple
   comparisons test are shown in each graph. All experiments were
   performed at least three independent times. All scale bars = 50 μm.

   Importantly, 2-DG and Cer, yet not NaN[3], caused gastrulation failure
   in the majority of embryos at 2 days post-fertilization (Day 2) in a
   dose-dependent manner, suggesting a progressive inhibition
   (Fig. [74]2e–g). To test the specificity of these inhibitors, we
   performed a rescue assay. We pre-treated 2-DG embryos with excess
   glucose to outcompete the 2-DG and analyzed if the phenotypes were
   specific to 2-DG’s inhibition of glucose/sugar. To validate that Cer
   explicitly inhibits fatty acid synthesis, we pre-treated embryos with
   palmitic acid, the direct product of fatty acid synthase, prior to the
   Cer treatment (Fig. [75]2h). In both cases, pre-treatment with the
   specific metabolite was able to rescue the gastrulation phenotypes seen
   in embryos treated with the inhibitor alone (Fig. [76]2i, j). Further,
   the control group embryos that were treated with glucose or palmitic
   acid with the same method, yet without 2-DG or Cer, respectively,
   showed no developmental abnormality (Fig. [77]S2a). These results
   demonstrate the specific efficacy of 2-DG and Cer in this embryo.

   Since 2-DG could inhibit both glycolysis and other sugar metabolism, we
   explored multiple glycolysis inhibitors. These include inhibitors for
   glycolytic intermediates, such as 6-AN and T045849, that block the
   pentose phosphate pathways and hexosamine biosynthesis, respectively. A
   recent study examining the influence of glycolysis on cell fate has
   found that much of the phenotype of interest involves these side
   pathways more than the core glycolytic pathway^[78]28. Under the same
   treatment condition, however, the resultant embryos completed
   gastrulation in proportions similar to untreated embryos, showing
   little impact on development (Fig. [79]S2b).

   Further, other inhibitors that block the core glycolysis pathway, such
   as YZ-9, shikonin, PKM2-IN-1, PGM1-004A, and YN1, obstructed or delayed
   micromere formation at the 16-cell stage. However, gastrulation was
   little affected at Day 2 in most treatment groups, except the shikonin
   group (Fig. [80]S2c, d). The proportion of embryos that failed to form
   micromeres at the 16-cell stage roughly matched that of embryos with
   failed gastrulation at Day 2 in all treatment groups. Since micromere
   signaling requires micromere formation, the resultant phenotype at Day
   2 might be due to the lack of micromere formation rather than the
   micromere signaling^[81]29. In contrast, another inhibitor for fatty
   acid synthesis, GSK2194069, showed no phenotype at the 16-cell stage
   yet compromised gastrulation at Day 2, which is similar to Cer
   (Fig. [82]S2e).

   In this study, we tested multiple inhibitors for each metabolic
   pathway, as it is challenging to determine the efficacy of drugs when
   they show no effect. The resultant phenotypes were broadly consistent
   among inhibitors for the same path as summarized in Fig. [83]2a.
   Although we still cannot exclude the possibility that all glycolytic
   inhibitors failed to function in these embryos, the above observations
   largely suggest that neither the core nor the side chain of glycolysis
   is likely active during micromere formation at the 8–16-cell stage. In
   contrast, fatty acid synthesis appears to be critical for the inductive
   function of micromeres.

   To further test this possibility, we measured metabolic rate using a
   Seahorse XF Analyzer at the 8- to 16-cell stage. The extracellular
   acidification rate (ECAR), which serves as an indicator of glycolysis,
   showed little change by 2-DG or any other inhibitors. This result
   suggests that glycolysis may be inactive in this embryo at this
   developmental stage (Fig. [84]S2F, ECAR). In contrast, the oxygen
   consumption rate (OCR), which indicates the level of mitochondrial
   respiration, was reduced by NaN[3] treatment (Fig. [85]S2F). Therefore,
   NaN[3] appears to specifically block Oxphos, even though it shows no
   phenotype at Day 2.

   Additionally, we tested whether these inhibitors cause energy collapse
   in the cell by biochemical assays. The results found no significant
   change in the levels of major energy and electron carriers, such as
   ATP, NAD, and NADP, in the embryos treated with 2-DG or Cer during
   micromere formation (Fig. [86]S2g). Therefore, the observed phenotypes
   caused by these inhibitors were not due to the general collapse of
   energy in embryos. Collectively, fatty acid synthesis at the 16-cell
   stage appears to be essential for successful gastrulation and embryonic
   patterning at Day 2, while oxidative phosphorylation is not necessary.
   Further, glucose is likely shunted to another metabolic pathway, such
   as sugar metabolism, and thus, glycolysis might be less active in this
   early stage of the embryo.

   Lastly, to gain a deeper understanding of the phenotypes caused by 2-DG
   and Cer, we performed in situ hybridization (ISH) using representative
   germ layer markers selected from gene regulatory network maps available
   for sea urchin early embryogenesis (Fig. [87]S3a)^[88]30,[89]31. These
   include ets, tbr, and sm50 for the skeletogenic mesoderm, pks for the
   non-skeletogenic mesoderm, blimp1b, foxA, hox11/13b, and endo16 for the
   endoderm, and foxQ2 for the ectoderm. Embryos were treated with each
   inhibitor during micromere formation for 30 min, processed for ISH, and
   phenotyped for the signal patterns at the blastula (Day 1) or gastrula
   (Day 2) stages. In the Cer group, a substantial number of embryos
   showed increased mesoderm marker expressions while decreasing endoderm
   markers at both the blastula (Day 1) and gastrula (Day 2) stages
   (Fig. [90]S3b–e). Similarly, the 2-DG group showed decreased endoderm
   markers at Day 1, which then recovered by Day 2, thus exhibiting a
   delayed phenotype rather than a complete inhibition (Fig. [91]S3b–e).
   The 2-DG group also increased ectoderm marker expression at Day 2
   (Fig. [92]S3b–e).

   Consistent with these observations, the quantitative fluorescent ISH
   (HCR-FISH) indicates increased expression of mesoderm markers over
   endoderm markers in both treatment groups (Fig. [93]S3f, g). These
   results suggest that both Cer and 2-DG treatments at the 16-cell stage
   weaken the gene expression necessary for initial endoderm specification
   at Day 1, leading to decreased or delayed endoderm formation at Day 2.
   However, 2-DG generally exhibited milder effects, which might have
   resulted in some differences in molecular phenotypes at Day 2 compared
   to Cer, as gastrulation requires multiple regulatory steps after the
   16-cell stage^[94]32.

Metabolic activity is required for micromeres’ signaling function

   Next, to test whether the Cer and 2-DG specifically block micromeres’
   function in inductive signaling, we performed a series of
   micromere-transplant experiments, which are the conventional methods to
   test micromeres’ signaling capability. For these experiments, another
   sea urchin species, Lytechinus variegatus (Lv), was used because of its
   suitability for embryology. The embryos were stained with CytoPainter
   (a red fluorescent dye) and treated with DMSO, 2-DG, or Cer for 20 min
   during the 4th cell division. Micromeres from treated donor embryos
   were then used to replace micromeres from the untreated host embryos
   (Forward transplants; Fig. [95]3a). The resulting chimeras were allowed
   to develop to the prism stage. 100% of embryos that received micromeres
   from DMSO-treated embryos developed as expected. This percentage is in
   sharp contrast to those for embryos that received micromeres treated
   with either 2-DG or Cer (8% for the 2-DG group; 0% for the Cer group;
   Fig. [96]3b, c). Because the transplanted micromeres were stained with
   a red fluorescent dye, we were able to confirm transplant success.
   Further, the cell fate of the micromere lineage as skeletogenic cells
   was not impacted, suggesting that metabolic inhibition affects only the
   micromere signaling, not the fate of micromeres (Fig. [97]3b). Of note,
   the PGC formation, another cell fate of micromeres, was not confirmed
   in this setting due to many embryos of the experimental groups showing
   significantly delayed or compromised development.

Fig. 3. Inhibitor-treated micromeres compromise gastrulation and development.

   Fig. 3
   [98]Open in a new tab

   a An experimental flow of the forward micromere transplant.
   Transplanted micromeres colored by a red fluorescent dye contribute to
   skeletogenic cells and the PGC in the late-stage embryo. b–d The
   micromere-transplant embryos developed normally in the control group
   while being delayed and compromised in the 2-DG and cerulenin groups,
   respectively, as summarized in (d). The transplanted micromeres (red)
   contribute to skeletogenic cells (arrows) in all groups. Yet, their
   contribution to the PGC (arrowheads) is apparent only in the control
   group due to the severe developmental delay/failure in the experimental
   groups. Over 90% of the transplanted embryos survived in all
   groups (b). Both no-transplant and micromere-transplant embryos treated
   with inhibitors showed similar developmental delay or failure with the
   reduced alkaline phosphatase signal (brown) in the endoderm
   (stars) (d). e An experimental flow of the reverse micromere
   transplant. The functional micromeres, colored red, were transplanted
   to the inhibitor-treated non-micromere host embryos. f–h The functional
   micromere rescued the gastrulation defects to some extent, both in the
   2-DG and Cer groups, and the results are summarized in (g). Over 90% of
   the transplanted embryos survived in all groups (f).
   Micromere-transplant embryos showed less compromised phenotypes with
   alkaline phosphatase signal (brown) in the endoderm (stars) compared to
   no-transplant embryos (h). n in (c) and (h) indicates the total number
   of embryos scored, and % indicates the proportion of embryos showing
   the AP signal. Each inhibitor dose used in this study is 30 mM 2-DG and
   0.2 mM Cer. These experiments were performed at least two independent
   times. L. variegatus embryos were used in this figure’s experiments.
   All scale bars = 100 μm.

   Notably, the phenotypes found in the chimeric embryos were similar to
   those found in the treated whole embryos with no transplant, which show
   apparent gastrulation defect or delay (Fig. [99]3d). These
   defects/delays are the typical phenotypes of micromere-less embryos
   caused by a lack of the signaling center in the embryo^[100]21,[101]22.
   To examine if the differentiation of endodermal tissues is impaired in
   these embryos, we used alkaline phosphatase (AP) staining, a
   conventional endoderm marker in the sea urchin embryo^[102]33. The
   proportion of embryos with strong AP-staining was higher in
   DMSO-treated whole embryos or chimeras compared to 2-DG or Cer-treated
   embryos (Fig. [103]3d), suggesting that endoderm differentiation is
   impaired in these chimeric embryos treated with metabolic inhibitors.
   This phenotype is also consistent with the ISH results described above,
   which showed decreased endoderm gene expression in the
   inhibitor-treated whole embryos (Fig. [104]S3). Next, to test if
   functional micromeres can rescue the effect of each inhibitor, we
   performed reverse transplants where normal micromeres were transplanted
   to the inhibitor-treated host embryos (Fig. [105]3e). In these reverse
   transplant chimeras, the functional micromeres were primarily able to
   induce gastrulation, yet slightly less so in the Cer-treatment group
   (Fig. [106]3f–h).

   These results suggest that 2-DG and Cer treatments at the 16-cell stage
   influence the signaling activity of micromeres but do not significantly
   affect their skeletogenic cell fate or the fate and function of other
   blastomeres. This indicates that sugar metabolism and fatty acid
   synthesis may be more crucial, specifically for the signaling activity
   of micromeres. It is important to note that in this study, embryos were
   exposed to the inhibitors only during micromere formation for 30 min to
   minimize unwanted side effects. Furthermore, the impact of the
   inhibitor was rescued by the corresponding rescue reagent, suggesting
   the specificity of each inhibitor’s effect (Fig. [107]2). Nevertheless,
   inhibitors could cause general toxicity in some cases, which may
   explain the incomplete rescue of the reverse transplants. Additional
   approaches involving genetic perturbations with spatial and temporal
   precision will be awaited in the future.

2-DG and Cer impact similar metabolic pathways in micromeres and later
development

   The results so far all imply that 2-DG and Cer, not NaN[3], impact a
   similar downstream mechanism in micromeres.

   To identify a potential downstream mechanism, we processed metabolome
   analyses for embryos treated with each inhibitor (Cer, 2-DG, NaN[3])
   for 30 min during micromere formation. We assessed the effects of
   inhibitor treatment on micromere and non-micromere samples separately
   (Fig. [108]4a; Data [109]S2), along with whole embryos collected at the
   16-cell stage immediately after treatment and two days after treatment
   (Day 2) (Fig. [110]4b; Data [111]S3).

Fig. 4. 2-DG and cerulenin-treated embryos show similar metabolomic profiles
in micromeres and later development.

   [112]Fig. 4
   [113]Open in a new tab

   a Schema depicting the experimental procedure for micromere and
   non-micromere samples. The embryos were exposed to each inhibitor for
   30 min at the 8–16-cell stage, respectively, and separated into the
   micromere and non-micromere fractions by sucrose gradient, followed by
   metabolomics analysis. b Schema depicting the experimental procedure
   for whole embryo (WE) samples. The embryos were exposed to each
   inhibitor (3 mM 2-DG, 0.2 mM Cer, and 1 mM NaN[3]) for 30 min at the
   8–16-cell stage, respectively. The treated whole embryos were collected
   for metabolomics analysis at the 16-cell stage or Day 2. c Heatmap
   visualizing correlated effects of different inhibitor treatments across
   lineage-specific and WE samples using Spearman correlation of fold
   changes (relative to DMSO control group). d Heatmap visualizing shared
   pathway responses of inhibitors in the micromere samples. Pathways with
   nominally significant Spearman correlations (p-value < 0.05) in at
   least one treatment comparison are shown. Clusters were obtained via
   hierarchical clustering. e Regression plots showing the correlation of
   selected pathways between 2-DG and Cer-treated micromeres. Each point
   is a metabolite with the respective fold change in 2-DG and Cer
   treatment. Significantly enriched pathways via GSEA identified in each
   treatment group, with their respective adjusted p-value and
   directionality, are summarized in the 16-cell (f) and Day 2 (g) embryo
   groups.

   To address similar phenotypic changes across inhibitor treatment,
   especially at Day 2, we analyzed how treatment effects covary with each
   other across the metabolome (Fig. [114]4a, c). Cer and NaN[3]
   treatments have the strongest correlated effects in the micromere
   lineage (r = 0.72), non-micromere lineage (r = 0.65), and the 16-cell
   whole embryo (r = 0.66), but not the Day 2 whole embryo (r = 0.10). In
   contrast, 2-DG and Cer treatments displayed a strong correlation in the
   micromere lineage (r = 0.52) and at Day 2 (r = 0.72), while having
   weaker correlations in the non-micromere lineage (r = 0.31) and the
   16-cell whole embryo (r = 0.43). 2-DG and NaN[3] treatments displayed a
   moderate correlation of r = 0.3 to 0.4 across lineages and timepoints.

   To gain insight into potential pathways explicitly altered in
   micromeres, we analyzed pathway-specific effects that covary across
   inhibitor treatments in the micromere fraction (Fig. [115]4d). Pathways
   involving amino acid and nucleotide metabolism were impacted by all
   treatments (2-DG, Cer, and NaN[3]). Pathways such as fructose and
   mannose metabolism, glycolysis, TCA, and glycerolipid metabolism were
   unique signatures shared by 2-DG and Cer inhibitor treatments. Pathways
   involving fatty acid metabolism and ABC transporters were unique
   signatures shared by Cer and NaN[3] treatments. We observe that
   affected pathways of 2-DG and Cer treatment had significant
   correlations and included metabolites with the highest fold change
   differences (Fig. [116]4e). Additionally, 2-DG and Cer do not affect
   the same pathways uniquely in the non-micromere lineage, suggesting
   that micromeres are sensitive to 2-DG and Cer treatments (Fig. [117]S4a
   and b).

   To highlight pathways altered in whole embryos, we performed gene set
   enrichment analysis for each treatment group compared to the DMSO
   control group at the 16-cell and Day 2 stages. At the 16-cell stage,
   2-DG and Cer increased metabolites related to amino acids/ tRNA
   biosynthesis. 2-DG decreased lipid/fatty acid metabolism, and NaN[3]
   showed no significant change (Fig. [118]4f). In Day 2 embryos, 2-DG and
   Cer caused similar metabolic profiles with increased amino acids /tRNA
   biosynthesis and decreased lipid/ fatty acid metabolism (Fig. [119]4g).
   In contrast, NaN[3] decreased amino acids and related pathways
   (Fig. [120]4g). The stark metabolome differences observed between
   2-DG/Cer vs. NaN[3] treatment at Day 2 are consistent with the little
   phenotypic change seen in NaN[3] treatment. These results suggest that
   2-DG and Cer impact similar metabolic pathways, specifically in
   micromeres at the 16-cell stage, causing a similar phenotype at Day 2
   in the resultant embryos.

N-glycosylation is required for gastrulation

   Common pathways that 2-DG and Cer target in micromeres include fructose
   and mannose metabolism along with other sugar metabolites
   (Fig. [121]4e). A key metabolite, mannitol, is a sugar alcohol, and its
   oxidation results in mannose, which 2-DG also targets
   (Fig. [122]5a)^[123]34–[124]37. The mannose pathway regulates N-linked
   glycosylation (N-glycosylation) necessary for secreted proteins’
   modifications in the endoplasmic reticulum (ER). Its inhibition results
   in failed protein secretion and misfolded protein
   accumulation^[125]38–[126]40. Fatty acid synthesis is also essential
   for N-glycosylation and fatty acylation of secreted proteins in the
   ER^[127]41,[128]42.

Fig. 5. Inhibition of N-glycosylation during micromere formation compromises
gastrulation at Day 2.

   [129]Fig. 5
   [130]Open in a new tab

   a A summary diagram of the proposed pathways impacted by metabolic
   inhibitors used in this study. Inhibitors are shown in magenta. Rescue
   reagents, glucose (Glc) and palmitic acid (PA), are shown in dark blue.
   b–d Schema depicting the experimental procedure for the inhibitor
   treatment (b). The representative images of Day 2 embryos treated with
   0.5 ng/μL tunicamycin (Tun), 0.15 mM cerulenin (Cer), or 15 mM 2-DG
   during micromere formation (c). An arrow and arrowheads indicate
   successful gastrulation and failed gastrulation, respectively. The
   proportion of embryos with successful gastrulation was scored for each
   treatment group (d). In the graph, the dots in each column indicate
   biological replicates. The total number of embryos scored (n) in each
   group is DMSO, n = 554; Tun, n = 408; Cer, n = 247; 2-DG, n = 244. The
   adjusted p-value between the control (DMSO) and each treatment group is
   Tun, p = 0.0005; Cer, p = 0.0014; 2-DG, p = 0.014. All graph data are
   presented as mean values ± SD. Significant results from One-way ANOVA
   with Dunnett’s multiple comparisons test are shown with asterisks. All
   experiments were performed at least three independent times. All scale
   bars in this figure = 50 μm.

   To test whether the N-glycosylation pathway is involved in micromeres’
   signaling activity at the 16-cell stage, we first treated embryos with
   tunicamycin (Tun), which explicitly blocks N-glycosylation
   (Fig. [131]5a)^[132]39,[133]43. Its pulse treatment during micromere
   formation caused gastrulation failure similar to 2-DG and Cer
   treatments (Fig. [134]5b–d). Since Tun-treated embryos formed
   micromeres at the 16-cell stage with no delay (Fig. [135]S5a, b), we
   next performed forward transplants using micromeres treated with Tun at
   the 8–16-cell stage (Fig. [136]S5c). The resulting chimeras exhibited
   compromised gastrulation and reduced staining of the endoderm marker at
   the gastrula stage in the treatment group (Fig. [137]S5d–f). Therefore,
   N-glycosylation in micromeres appears to play a critical role in
   micromere signaling at the 16-cell stage.

ER and Wnt proteins are enriched at the vegetal pole through an asymmetric
cell division

   Many secreted proteins undergo glycosylation and/or fatty acylation in
   the ER, which is critical for protein secretion. These secreted
   proteins include Wnt proteins. To explore whether Wnt proteins are a
   part of the micromere signaling at the 16-cell stage, we first sought
   to analyze the ER localization dynamics during micromere formation. The
   KDEL/RDEL is a conserved four-amino-acid signal sequence to the ER
   found across organisms, serving as a general ER marker^[138]44–[139]46.
   In the sea urchin embryo, KDEL immunofluorescence suggests that the ER
   is evenly distributed on the spindle at the M-phase during symmetric
   cell divisions. In contrast, during the asymmetric cell division at the
   16-cell stage, the KDEL signal became asymmetric toward the micromere
   side of the spindle, resulting in asymmetric enrichment at the vegetal
   pole (Fig. [140]6a, arrows). KDEL/RDEL has also been previously used to
   visualize the ER in live sea urchin embryos^[141]47. Embryos introduced
   with mCherry-RDEL exhibited localization patterns similar to the KDEL
   immunofluorescence with no notable cellular toxicity, suggesting that
   mCherry-RDEL reflects the endogenous ER distribution, as reported
   previously (Fig. [142]6b).

Fig. 6. ER and Wnt localize to the vegetal pole through an asymmetric cell
division at the 16-cell stage.

   [143]Fig. 6
   [144]Open in a new tab

   a Immunofluorescence of KDEL, the ER signal sequence (magenta),
   counterstained with tubulin (green) and DNA (blue). Negative control
   (Nega. Cont.) was processed without the KDEL primary antibody. Arrows
   indicate the asymmetric cell division, while an arrowhead indicates the
   symmetric cell division during the 8–16-cell stages. b Live embryo
   images of GFP-Wnt5 or -Wnt8 (green; 500–1000 ng/μL stock),
   counterstained with mCherry-RDEL (magenta) at the 16-cell stage. Arrows
   indicate the vegetal pole. c, d Live embryo images of the GFP-Wnt5
   embryos treated with C59 during imaging (5–6 hpf) or with DMSO, Cer, or
   2-DG during micromere formation (4–4.5 hpf). Embryos were
   counterstained with membrane-mCherry (magenta; 200 ng/μL stock) to
   outline the cell. The squared regions are enlarged in the bottom row of
   each panel. GFP-Wnt5 particles (arrows) outside the cell, yet within
   the hyalin layer or fertilization envelope, were counted for each group
   and normalized to that of the DMSO control group in graph (d). In the
   graph, the dots in each column indicate biological replicates. The
   total number of embryos scored (n) in each group is DMSO, n = 124; Tun,
   Cer, n = 106; 2-DG, n = 75; C59, n = 50. The adjusted p-value between
   the control (DMSO) and each treatment group is Cer, p = 0.004; 2-DG,
   p = 0.007; C59, p = 0.0001. The graph data is presented as mean
   values ± SD. Significant results from One-way ANOVA with Dunnett’s
   multiple comparisons test are shown with asterisks. All experiments
   were performed at least three independent times. All scale bars in this
   figure = 50 μm.

   Although few antibodies that cross-react with sea urchin Wnt proteins
   are available, the Wnt5a antibody tested in this study displayed a
   specific band at the expected size, suggesting its presence in early
   embryos (Fig. [145]S6a). Wnt5a immunofluorescence indicates signal
   localization at the spindle and in micromeres during the 8–16-cell
   stage (Fig. [146]S6b). To analyze the localization dynamics in live
   embryos, we cloned the Wnt5a and Wnt8 ORFs from the 16-cell stage cDNA
   library of S. purpuratus to create GFP-fusion constructs. Similar to
   the immunofluorescence results, GFP-Wnt exhibited localization on the
   spindle and enrichment at the vegetal pole in live embryos at the
   16-cell stage, primarily following ER dynamics (Fig. [147]6b;
   Fig. [148]S6c).

   In live imaging of GFP-Wnt5a, we observed the GFP particles emerging
   from the ER area and, in some cases, in the extracellular space within
   the hyaline layer or fertilization envelope at the 16-cell stage
   (Fig. [149]6c). In this study, we were unable to track individual Wnt
   particle dynamics using conventional imaging methods. Therefore, we
   relied on the particles that appeared on the cell surface or in the
   extracellular space as a reflective measure of Wnt proteins emerging
   from the ER. Exposing the embryos to C59, a known Wnt secretion
   inhibitor^[150]48, decreased these surface/extracellular protein
   particles by nearly half compared to the DMSO control at the 16–32-cell
   stage (Fig. [151]6d), suggesting that these particles may reflect a
   part of the Wnt proteins.

   To test whether metabolic inhibition alters the Wnt dynamics, we
   treated the embryos with Cer or 2-DG for 30 min during micromere
   formation, washed off the inhibitors, and then imaged them at the
   16–32-cell stages. Both Cer and 2-DG reduced the surface/extracellular
   Wnt particles by nearly half compared to the DMSO control group
   (Fig. [152]6c, d). We observed similar patterns with GFP-Wnt8
   (Fig. [153]S6d, e). These results suggest that 2-DG and Cer treatments
   impact Wnt protein dynamics at the 16–32-cell stage, implying these
   metabolic regulations in micromeres are critical for their ER function
   in regulating signaling molecules at the 16-cell stage.

Discussion

   In this study, we demonstrate that micromeres have a unique metabolic
   profile as well as asymmetrically enriched ER at the 16-cell stage. A
   metabolome comparison of micromeres and non-micromeres suggests that
   micromeres are enriched with lipids and carbohydrates while reduced
   with amino acids and ABC transporters (Fig. [154]7a). Further, these
   results indicate that the initial metabolic asymmetry is established as
   early as the 16-cell stage. How this unique metabolic profile is
   established through a single asymmetric cell division is yet to be
   identified. However, we previously reported that the Activator of
   G-protein Signaling (AGS), a polarity factor, plays a critical role in
   micromere formation^[155]29,[156]49. AGS appears to regulate asymmetric
   distributions of polarity factors and cell fate determinants directly
   or indirectly in this process. If AGS also regulates the asymmetric ER
   distribution, the associated metabolites and secreted proteins
   processed in the ER might accompany it altogether, which will be
   important to test in the future.

Fig. 7. Unique metabolic regulation in micromeres contributes to embryonic
patterning.

   [157]Fig. 7
   [158]Open in a new tab

   a Schema depicting the metabolic asymmetry at the 16-cell stage and
   future germ layer territories, which is altered by each inhibitor
   treatment. The deeper color indicates more enrichment in the micromere
   (Mic) or non-micromere (Non-Mic) of the 16-cell stage embryo or
   expansion of the future germ layer territories. b Proposed Cartoon
   models summarizing altered embryonic patterning by each inhibitor
   treatment. In normal embryos, the micromeres signal to adjacent
   macromeres to facilitate the initial endoderm induction at the 16-cell
   stage. This initial signaling balances a competition between the future
   endoderm and the non-skeletogenic mesoderm lineages since cells may be
   intrinsically inclined to the mesoderm lineage. The micromere
   descendants then induce the non-skeletogenic mesoderm in the macromere
   descendants, defining the endoderm territory prior to gastrulation at
   Day 2. In contrast, the skeletogenic lineage is derived solely from the
   micromeres under normal conditions. Both 2-DG and Cer reduced the
   inductive signaling of micromeres, resulting in delayed or compromised
   endoderm initiation on Day 1 and increased ectoderm or mesoderm at Day
   2, respectively. Endo, endoderm. Meso mesoderm. Ecto ectoderm. NSM
   non-skeletogenic mesoderm. SM skeletogenic mesoderm.

   Metabolome and inhibitor analyses also suggest that conventional
   glycolysis may be less active during early embryogenesis of the sea
   urchin. Glucose appears to be partly shunted to another sugar
   metabolism, N-linked glycosylation, at the 16-cell stage. Considering
   that rapid development is critical for early embryogenesis, these
   embryos may benefit from taking the metabolic bypass using maternal
   resources such as sugar and lipids to control cellular activities
   immediately. In contrast, conventional glycolysis appears to be
   critical for presomitic mesoderm differentiation, which occurs at the
   later stage of embryogenesis in vertebrates^[159]50,[160]51. Therefore,
   conventional glycolysis may kick in later in development. Future
   studies in multiple model systems are awaited to test this hypothesis.

   Inhibitor treatments followed by metabolome analysis suggest that
   micromeres have a distinct response to each inhibitor compared to
   non-micromeres. 2-DG and Cer appear to alter similar metabolic
   pathways, such as fructose and mannose metabolism, but not in
   non-micromeres. This observation suggests that lineage-specific
   sensitivity is present for each metabolic inhibition. Additionally, in
   micromere-transplant experiments, micromeres treated with 2-DG and Cer
   at the 16-cell stage failed to induce gastrulation in the untreated
   host embryo at Day 2. In contrast, the untreated micromeres mostly
   rescued the phenotype in the inhibitor-treated host embryos at Day 2.
   These results consistently suggest that the 2-DG and Cer treatments
   specifically yet similarly impact micromeres’ signaling activity at the
   16-cell stage. It is crucial to test these observations further through
   genetic approaches by controlling the metabolites, ER, or signaling
   molecules directly during asymmetric cell division in the future.

   The micromeres’ inductive signaling begins as soon as this lineage
   forms at the 16-cell stage, which occurs autonomously even at an
   ectopic place^[161]21, suggesting its initial signaling at the 16-cell
   stage might be regulated by the maternal factors. In contrast, the
   downstream signaling components following the 16-cell stage have been
   extensively studied. These include Wnt/β-catenin, Activin/TGF-β, and
   Delta/Notch pathways, which are conserved elements of organizers in
   various organisms^[162]9,[163]10,[164]48,[165]52,[166]53. In this
   study, treatments with 2-DG and Cer during micromere formation reduced
   the surface/extracellular Wnt particles at the 16-28-cell stage and
   diminished endoderm marker expressions that are primarily regulated by
   the Wnt pathway at Day 1. These observations suggest that maternally
   regulated Wnt signaling may be involved in facilitating the initial
   endoderm signaling at the 16-cell stage. This hypothetical model could
   also explain the earlier report that nuclear β-catenin localization
   occurs first in macromeres at the 28-cell stage (Logan et al.^[167]54).
   However, in this study, GFP-Wnt proteins were overexpressed; therefore,
   we did not investigate the functional significance of these Wnt
   particles. Controlling Wnt and ER dynamics with spatial and temporal
   precision will be critical for understanding the functional importance
   of their enrichment at the vegetal pole in future studies.

   Furthermore, previous research has shown that disruptions to micromere
   signaling can alter endomesoderm specification through the signaling
   components Delta and Activin B^[168]55,[169]56. The transfate of
   endomesoderm cells has also been characterized previously in response
   to the surgical removal of the primary mesenchyme cells^[170]57.
   Moreover, the gene regulatory network that maps gene activations in
   transfating endoderm describes an expression profile in which
   mesodermal markers such as alx1 and ets1 are upregulated in a delayed
   manner in presumptive endoderm following depletion of the primary
   mesenchyme cells and the non-skeletogenic mesodermal region^[171]58. In
   this study, gene expressions in response to 2-DG and Cer treatments
   showed a similar shift in germ layer specification, with increased
   mesoderm and decreased endoderm markers, akin to the disruptions in
   micromere signaling (Fig. [172]7b). However, the steps towards
   gastrulation involve multiple phases and are dynamic in these
   regulative embryos^[173]32,[174]54,[175]56,[176]59–[177]63. Identifying
   the molecular details of micromere signaling will be crucial to connect
   the direct contributions of micromeres’ metabolic regulation at the
   16-cell stage to the observed phenotypes at Day 2.

   Although more future studies are awaited to uncover the molecular
   details and functional significance of micromere signaling and ER
   asymmetry, the observations in this study consistently suggest that
   micromeres’ unique metabolic regulation contributes to their signaling
   function as organizers. Perhaps increased ER accelerates maternal
   protein modifications in micromeres, providing immediate signaling
   production during rapid embryogenesis. Inductive signaling by
   organizers or organizer-like cells is highly conserved across
   organisms, and similar asymmetric regulation of metabolites and ER may
   be employed in inductive signaling of other organisms.

Methods

Outlines

     * Contact for reagent and resource sharing
     * Experimental model and subject details
     * Methods details

Contact for reagent and resource sharing

   For further information, requests should be directed to and will be
   fulfilled by the Lead Contact, M.Y. (Mamiko_Yajima@brown.edu).

Experimental model and subject details

   All sea urchins were wild-type, collected from the ocean, and
   maintained in the aquarium.

Methods details

Sea urchins and embryo culture

   Strongylocentrotus purpuratus (Sp) and Lytechinus variegatus (Lv) were
   obtained from Pete Halmay (peterhalmay@gmail.com), Kerchoff Marine
   Laboratories, California Institute of Technology, or from the Duke
   University Marine Lab in Beaufort, North Carolina, USA. Eggs and sperm
   were collected by injection of 0.5 M KCl. For sample collection and
   microinjection, eggs were fertilized in 1 mM 3-amino triazole (Sigma,
   St. Louis, MO, USA) to prevent the cross-linking of fertilization
   envelopes. These fertilized eggs were cultured in seawater at 16 °C for
   Sp or at room temperature (25 °C) for Lv until they reached the desired
   stage.

In situ hybridization (ISH) and HCR-FISH probe construction and signal
detection

   For ISH, the Open Reading Frame (ORF) of each gene of interest was
   PCR-amplified from the cDNA libraries of S. purpuratus using the
   primers in Table [178]1. The amplified PCR product was then inserted
   into the TOPO vector (Promega) by following the manufacturer’s
   protocol. DIG-UTP-labeled RNA probes for genes of interest were
   prepared from plasmid templates using Invitrogen MEGAscript SP6 or T7
   in vitro transcription kits (Thermo Fisher) according to the
   manufacturer’s instructions. Sp embryos were fixed using 4%
   paraformaldehyde at 24 hpf and 48 hpf. Fixed embryos were washed with
   MOPS buffer and stored in 70% EtOH at −20 °C until needed. Each RNA
   probe (0.3–0.5 ng/uL) was then added to the fixed embryo samples at
   60 °C overnight. Colorimetric detection of hybridized probes was
   accomplished by incubating embryos with anti-Digoxigenin-AP Fab
   fragments (Roche), followed by the substrates NBT and BCIP.

Table 1.

   Key resources table
   Reagent or resource Source Identifier
   Antibody
   Anti-Digoxigenin-AP, Fab fragments Roche # 11093274910
   WNT5A/B rabbit polyclonal antibody ProteinTech # 55184-1-AP
   KDEL rabbit polyclonal antibody Thermo Fisher Scientific # PA1-013
   FITC-conjugated Anti-β-tubulin Sigma-Aldrich # F2043
   Mouse Anti-β-Actin monoclonal antibody Cell Signaling Technology #
   8H10D10
   Cy3-conjugated goat anti-rabbit IgG Thermo Fisher Scientific # A10520
   Alexa 488-conjugated goat anti-mouse IgG Thermo Fisher Scientific #
   A32723
   HRP-conjugated goat anti-rabbit IgG Thermo Fisher Scientific # A16096
   Chemicals, peptides, and recombinant proteins
   Hoechst 33342 Thermo Fisher Scientific # 62249
   Cytopainter Abcam # 138893
   Cy5 dextran AlexaFluor™ 647 Thermo Fisher Scientific #[179]D22914
   Click-iT Plus OPP Protein Synthesis Assay Kits Life Technologies #
   [180]C10457
   2-Deoxy-D-glucose (2-DG) Calbiochem # 25972-1GM
   Cerulenin Millipore-Sigma # [181]D44101
   NaN[3] Millipore-Sigma # S2002
   YZ-9 Cayman Chemical # 15352
   Shikonin Millipore-Sigma # S7576
   GSK2194069 TOCRIS # 5303
   ST045849 Millipore-Sigma # SML2702-5MG
   6-Aminonicotinamide (6-AN) Millipore-Sigma # A68203
   PKM2-IN-1 MedChemExpress #HY-103617
   PGM1-004A MedChemExpress # HY-101143
   YN1 Millipore-Sigma # 5.31050
   D-(+)-Glucose Millipore-Sigma # G5400
   Palmitic Acid Cayman Chemical # 10006627
   Tunicamycin Cayman Chemical #11445
   C59 Cayman Chemical 16644
   H[2]O[2] Fisher Science # H323-500
   DTT Millipore-Sigma # 10197777001
   DMSO Millipore-Sigma # D8418
   ATP Assay Kit Abcam # ab83355
   NAD/NADH Assay Kit-WST DOJINDO (Kumamoto, Japan) # N509
   NADP/NADPH Assay Kit-WST DOJINDO (Kumamoto, Japan) # N510
   Hexokinase Assay Kit Abcam # ab136957
   Pierce BCA Protein Assay Kit Thermo Fisher Scientific #23227
   Methanol for GC-MS FUJIFILM Wako (Osaka, Japan) #138-14521
   Hexane for GC-MS FUJIFILM Wako (Osaka, Japan) #084-03421
   N-alkane, Qualitative Retention Time Index Std for GC-MS SHIMADZU
   (Kyoto, Japan) #31080
   Chloroform for GC-MS FUJIFILM Wako (Osaka, Japan) #038-02606
   2-isopropylmalic acid for GC-MS Millipore-Sigma #333115-100MG
   Critical commercial assays
   mMESSAGE mMACHINE SP6 Transcription Kit Ambion # AM1340
   Invitrogen MEGAscript SP6 Transcription Kit Ambion # AM1330
   Invitrogen MEGAscript T7 Transcription Kit Ambion # AM1334
   In-Fusion HD Cloning Clonetech # 639648
   Presh-SPE ACXs AiSTI SCIENCE (Wakayama, Japan) #SA-5589-003
   BPX5 capillary column SHIMADZU GLC (Tokyo, Japan) 054101
          mRNA construct primers
                 Construct               Primers
   SP64-GFP-Wnt5a ([182]XM_774853.5)

   F: GAGGGGATCGGTGGAGCTCCACCGGTATGGAAACGTGTACGAACTC

   R: TAACCAGATCCTAGTCAGTCACTAGTTTTACAAACATGTACATCAA
   SP64-GFP-Wnt8 ([183]NP_999832.1)

   F: GAGGGGATCGGTGGAGCTCCACCGGTATGGATGTCTTTACGAAATTTGTTCGTCATCTTCTTCTGC

   R: TAACCAGATCCTAGTCAGTCACTAGTCTATCCAACGGGCTGGCACG
   SP64-mCherry-Wnt5a ([184]XM_774853.5)

   F: GCATGGACGAGCTGTACAAGGCGGCCGCAATGGAAACGTGTACGAACTC

   R: TAACCAGATCCTAGTCAGTCACTAGTTTTACAAACATGTACATCAA
   SP64-mCherry-Wnt8 ([185]NP_999832.1)

   F: GCATGGACGAGCTGTACAAGGCGGCCGCAATGGATGTCTTTACGAAATT

   R: TAACCAGATCCTAGTCAGTCACTAGTCTATCCAACGGGCTGGCACG
   SP64-eCast/PDI-GFP-RDEL

   F: aataaacgctcaactttggcagaTCTATGAAGTATTTGGCTCTTTG

   R: tcctcgcccttgctcaccatgggcccAGCGACATCTTCTTCGATTTCGAC
   SP64-eCast/PDI-mCherry-RDEL

   F; aataaacgctcaactttggcagaTCTATGAAGTATTTGGCTCTTTG

   R: tcctcgcccttgctcaccatgggcccAGCGACATCTTCTTCGATTTCGAC
   membrane-mCherry                      Uchida & Yajima 2018
       ISH primers

   Sp-Delta

   NCBI #LOC115921237; Probe length: 899 bp

   F: GCAGGGACATTCGAACTTCG

   R: GCTGTCTCACAGGTGTGTCC

   Sp-Tbrain

   NCBI #LOC586389; Probe length: 1052 bp

   F: CCACCGCTGCACCAGACGAC

   R: CTGCCGGCTGGCGCCAATTGCG

   Sp-Ets1

   NCBI #LOC73306; Probe length: 1322 bp

   F: TCAATCATGGCGTCTATGCACTG

   R: ACAGCTGCAGGGATAACAGG

   Sp-SM50

   NCBI #373464; Probe length: 1342 bp

   F: ATGAAGGGAGTTTTGTTTATTGTGGCTAGTC

   R: GTTATGCCAACGCGTCTGCCTCTTGAAGC

   Sp-Endo16

   NCBI #LOC373279; Probe length: 1368 bp

   F: GGTTAAATATTTTGCTGTTCGCGG

   R: GAGTATTCGGTACTGGTGCTC

   Sp-FoxA

   NCBI #LOC578584; Probe length: 1323 bp

   F: ATGGCCAATAGTGCCATGATCTCG

   R: TCACATTGCATGGTTTGCTTG

   Sp-FoxQ2

   NCBI #LOC578584; Probe length: 1323 bp

   F: ATGACTTTATTCAGCATTGACAAC

   R:TAGCAGGATCCTACAGAAGACCAG

   Sp-PKS1

   NCBI #LOC588806; Probe length: 1952 bp

   F: ATGGGAAGCAATAAAACCAGCTGGGG

   R: GCAGCCGATACCTCACCAAGACTGTG

   Sp-Blimp1b

   NCBI #LOC751833; Probe length: 1057 bp

   F: ATGGGGTGCAACGACAACGCCGTG

   R: TGGGCTGTATGTGGCGATTCTTGG

   Sp-Hox11/13b

   NCBI #373462; Probe length: 1020 bp

   F: ATGCAGATCGGCATGGAACAGGCTTGG

   R: TCATCGAAGCTGTGGTTGATGACCAAC
   HCR-FISH probes

   Sp-FoxA

   NCBI #LOC578584;
   Molecular Instruments

   Sp-PKS1

   NCBI #LOC588806
   Molecular Instruments
   Software and algorithms
   Echinoderm gene/protein sequences EchinoBase
   [186]http://www.echinobase.org/Echinobase/
   Imaging software Nikon NIS Elements
   Quantitative analysis Image J [187]https://imagej.nih.gov/ij/
   Statistical analysis PRISM
   [188]https://www.graphpad.com/scientific-software/prism/
   Bioinformatics analysis Please see Table [189]2
   [190]Open in a new tab

   For HCR-FISH, all probes were designed and constructed by Molecular
   Instruments. Embryos were fixed using the normal ISH methods described
   above, and the hybridization and signal visualization were performed
   following the manufacturer’s protocol
   ([191]https://www.molecularinstruments.com/hcr-rnafish-protocols) with
   a probe concentration of 8 μM. Please see Table [192]1 for further
   details on reagents.

   Imaging was performed with the Nikon CSW Nikon Eclipse Ts2R microscope
   for brightfield imaging and the Nikon CSU-W1 Spinning disk laser
   microscope for fluorescent imaging. Embryos were scored visually by
   signal intensity and territory, using NIS Elements software (version
   5.21.03 for Windows).

Inhibitor treatments and seahorse analysis

   Approximately 250 embryos per well were treated with each inhibitor for
   30~60 minutes during the 8–16-cell stage and washed five times with
   seawater prior to the downstream culture up to Day 2 or experiments in
   the 24-well plate. Phenotype counting and brightfield imaging were
   performed using the NIKON Ts2R microscope and NIS Elements software.
   For Day 1 or 2 embryos, embryos were temporarily immobilized by
   treating them with 2x seawater for 5 min, followed by a wash with
   seawater. Oxygen consumption rate (OCR) and Extracellular Acidification
   Rate (ECAR) were detected by the Seahorse XF by following the
   manufacturer’s protocol (Agilent Technologies, MA, USA). The
   triplicates per group were processed, and each sample was assessed
   simultaneously in independent wells within the same assay plate per
   cycle. The resultant value of each sample group was normalized to that
   of the control group with no treatment. For information on individual
   inhibitors, please refer to Table [193]1.

Enzymatic assays

   The cell lysates of the embryo were prepared with 1.2 ml of the cell
   lysis buffer of the ATP Assay Kit (Abcam, ab833355). The 500 ml of cell
   lysates was transferred into a microfuge tube and centrifuged at
   2300 × g and 4 °C for 5 min. Next, the upper aqueous layer was
   centrifugally filtered through a Millipore 5-kDa cutoff filter (Human
   Metabolome Technologies; HMT, Inc., Tsuruoka, Japan) at 9100 × g and
   4 °C for 120 min to remove proteins. ATP, total NAD/NADH, and total
   NADP/NADPH in the filtrate were quantified with assay kits (ATP Assay
   Kit, Abcam ab83355, NAD/NADH Assay Kit-WST DOJINDO N509, NADP/NADPH
   Assay Kit-WST DOJINDO N510), respectively. The fluorescence intensity
   for ATP and optical density for NAD and NADP of the mixture in each
   well were measured at Ex/Em = 535/587 nm or 450 nm on a microplate
   reader (SpectraMax Paradigm, Molecular Devices, USA). The protein
   content in cell lysates was measured using bicinchoninic assays with
   the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, #23227).
   Hexokinase activity in cell lysates was measured with the Hexokinase
   Assay Kit (Abcam, ab136957). The concentration of ATP, total NAD/NADH,
   total NADP/NADPH, and hexokinase activity were calculated, and
   statistical analysis was performed using GraphPad Prism 9 software
   (GraphPad Software, Inc., La Jolla, CA, United States).

Plasmid construction, mRNA microinjection, and imaging

   The Open Reading Frame (ORF) of SpWnt5a and SpWnt8 was PCR-amplified
   from the egg or 16-cell stage embryo cDNA library of S. purpuratus. The
   amplified PCR fragment was inserted into the SP64 vector using the
   In-Fusion cloning kit (Clonetech), by following the manufacturer’s
   instructions. For GFP or mCherry-RDEL constructs, following the
   previous publication by Terasaki et al.^[194]47, the nucleotide
   fragment corresponding to the signal sequence (first 28 amino acids) of
   the S.purpuratus luminal protein eCast/PDI protein was cloned from the
   cDNA library indicated above. The cloned fragment was then inserted
   before the GFP/mCherry ORF sequence, followed by the RDEL
   (Arg-Asp-Glu-Leu) signal sequence, into the SP64 vector. mRNA was
   constructed using the Message Machine (Ambion) kit by following the
   manufacturer’s protocol and microinjected into the zygote at a
   stock concentration of 200–1000 ng/μL as indicated individually in the
   relevant figure legends. All embryos were imaged in live using the
   Nikon CSU-W1 Spinning disk laser microscope. For analysis, embryos were
   assessed visually and scored for the signals using NIS Elements and
   Image J software.

Alkaline phosphatase staining

   Alkaline phosphatase staining^[195]33 was performed as follows: All
   embryos were fixed with ice-cold 90% Methanol for one hour, followed by
   a PBS wash. Embryos were then rinsed with the pH 9.5 Tris reaction
   buffer (100 mm NaCl, 5 mm MgCl[2], and 100 mm Tris-HCl) and exposed to
   the 1 mL of the staining solution that contains 3.3 µL of 50 mg/mL
   nitroblue tetrazolium (NBT) and 3.3 µL of 25 mg/mL
   5-Bromo-4-Chloro-3-Indolyl Phosphate (BCIP) for 15 min at room
   temperature. All groups were exposed to the staining solution for the
   same amount of time to allow the comparison of relative staining
   between groups. All images were taken on the NIKON Ts2R epifluorescent
   microscope.

Immunofluorescence

   Embryos were fixed with 90% methanol for at least 1 hour at −20 °C,
   washed with PBS 3–5 times, and incubated with primary antibody at a
   dilution of 1:100–1:200 overnight at 4 °C. After washing the primary
   antibody with PBS 10 times, embryos were incubated with the secondary
   antibody at a dilution of 1:300–1:500 for 3–5 h at room temperature,
   washed with PBS 10 times, counterstained with Hoechst, and mounted on
   the slides. For individual antibody information, please refer to
   Table [196]1. All embryos were imaged on the Nikon CSU-W1 Spinning disk
   laser microscope.

Transplants

   The donor embryos were treated with each inhibitor or DMSO (control)
   together with 1 µL of 500× Cytopainter in 500 µL seawater for 20–30 min
   prior to micromere transplant. The transplant was performed in
   Calcium-free seawater under the stereomicroscope. The resultant embryos
   were cultured to the desired stage and imaged in live or fixed for
   alkaline phosphatase staining^[197]64. All images were taken on the
   NIKON Ts2R epifluorescent microscope.

Statistical analyses for imaging and biochemical experiments

   All images were analyzed by FIJI (NIH) or by Nikon NIS Elements, and
   statistical significance was performed by PRISM (GraphPad) using
   one-way or two-way ANOVA. P-values less than 0.05 were considered
   significant. Asterisks were used to indicate significance corresponding
   with * is p < 0.05, ** is p < 0.01, *** is p < 0.001, **** is
   p < 0.0001. Columns represent means ± SD or SEM.

Metabolome sample collections

   Each batch of embryos (~5 μL pack of embryos per sample) was collected
   from a single dish for all sample groups and control groups at once to
   minimize technical variations. Up to six batches of the samples were
   prepared for each sample group, and three batches of the best samples
   that showed the expected phenotypes were collected and processed for
   mass spectrometry. For Sample collections for micromere vs
   non-micromere fractions, 16-cell embryos were dissociated and separated
   into micromere and non-micromere fractions based on cell size by a
   sucrose gradient^[198]12,[199]65,[200]66, which was each confirmed
   under the microscope prior to sample collection.

Metabolome sample preparation and Mass Spectrometry

   Each sample was weighed, snap-frozen in liquid nitrogen, and stored at
   −80 °C until use in metabolomics analysis by gas chromatography coupled
   to triple quadrupole tandem mass spectrometry (GC-MS/MS). Three
   biologically independent samples (N = 3) were prepared, and each was
   analyzed. Raw metabolomic data were obtained through GC-MS analysis,
   following well-established methods by Ogiwara et al. and Osu et
   al.^[201]67,[202]68. Metabolites were extracted with 260 μL of a
   solvent mixture consisting of methanol (#138-14521, FUJIFILM Wako,
   Osaka, Japan), chloroform (#038-02606, FUJIFILM Wako, Osaka, Japan),
   and Milli-Q water (2.5:1:1), containing 5 μg of 2-isopropylmalic acid
   (#333115-100MG, Sigma-Aldrich, MO) as an internal standard.

   The metabolite extract was transferred to a microfuge tube and dried
   using a spin dryer (TAITEC, Koshigaya, Japan). Derivatization of the
   solid phase on the solid-phase cartridge, Presh-SPE ACXs, was performed
   following the manufacturer’s protocol (#SA-5589-003, AiSTI SCIENCE,
   Wakayama, Japan). Derivatized samples were eluted with 100 μL of hexane
   (#084-03421, FUJIFILM Wako, Osaka, Japan), and 1.0 μL of the
   derivatized solution was injected into a GC-MS/MS system (GC-MS-TQ8050,
   SHIMADZU, Kyoto, Japan). Metabolome analysis was conducted using the
   GC-MS-TQ8050 equipped with a BPX5 capillary column (#054101, internal
   diameter: 30 m × 0.25 mm; film thickness: 0.25 μm; SHIMADZU GLC, Tokyo,
   Japan). For GC-MS-TQ8050 analysis, the inlet temperature was maintained
   at 250 °C, and helium served as the carrier gas at a steady flow rate
   of 39.0 cm/s. Argon gas was used for collision-induced dissociation.
   The injector split ratio was set to 1:10. The total GC run time was
   23 min. The GC column temperature was programmed to stay at 60 °C for
   2 min, then increase from 60 °C to 330 °C at 15 °C per minute, and
   finally hold at 330 °C for 3 min. The transfer line and ion-source
   temperatures were 280 °C and 200 °C, respectively. The ionization
   voltage was 70 eV. The Automatic Adjustment of Retention Time (AART)
   function of the GC-MS solution software (SHIMADZU, Kyoto, Japan) and a
   standard alkane series mixture (C7 to C33, n-alkane, #31080, SHIMADZU,
   Kyoto, Japan) were used to correct retention times. Metabolite
   detection employed the Smart Metabolites Database and GC-MS Solution
   software (SHIMADZU, Kyoto, Japan), which contains the relevant Multiple
   Reaction Monitoring (MRM) method file and data on GC analytical
   conditions, MRM parameters, and retention index used for metabolite
   measurements. Peaks were identified automatically and manually verified
   based on specific precursor and product ions, as well as retention
   times. The area under the curve for each metabolite was normalized to
   that of the internal standard and defined as the metabolite level per
   milligram of sample weight.

Metabolome data analyses

   The metabolome data were analyzed using the online tool
   MetaboAnalyst^[203]69 ([204]https://www.metaboanalyst.ca/) and custom R
   scripts ([205]https://github.com/yajima-lab/SeaUrchinMicMetabolomics).
   Approximately, ~100 metabolites were identified per sample. Samples
   with missing values were imputed by 1/5 of the minimum value of each
   metabolite. For quality control, partial least squares discriminant
   analysis (PLS-DA) was performed to determine if the samples clustered
   into experimental groups. Differential abundance of each metabolite was
   assessed using a paired t-test for datasets involving comparisons
   across two conditions: Micromere vs. Non-Micromere (Mic/NM), Inhibitor
   Micromere vs Non-Micromere, and Inhibitor Whole Embryo (16-cell and Day
   2).

   To identify potential pathways of interest, we used Genes Set
   Enrichment Analysis (GSEA; Subramanian et al.^[206]70) for metabolomics
   data. We calculated the GSEA statistic, −log10(p-value) *
   sign(log2(FC)), using the nominal p-value from the paired t-test and
   fold change to rank metabolites. Fold changes were calculated as the
   average log2 fold difference between groups. For inhibitor-treated
   embryos, comparisons were between each inhibitor (2-DG, Cer, and
   NaN[3]) and the DMSO control group. For Mic/NM, comparisons were
   between micromere and non-micromere samples. Adjusted p-values are
   reported using the Benjamini–Hochberg procedure to control for false
   discovery rate.

   To identify shared treatment effects of 2-DG, Cer, and NaN[3] across
   the metabolome, we used Spearman rank correlation of fold changes for
   each pairwise comparison. These correlations were calculated for the
   entire metabolome as well as for pathways of interest.

   Gene sets included KEGG pathways^[207]71
   ([208]https://www.genome.jp/kegg-bin/show_organism?org=spu) and curated
   pathway sets capturing broad classifications (Carbohydrate Metabolism,
   Amino Acid Metabolism, Lipid Metabolism, etc.) based on groupings from
   individual KEGG pathways (Table [209]2). Metabolites were paired with
   corresponding KEGG IDs via the MetaboAnalyst conversion ID tool.

Table 2.

   Curated pathways created from a set of individual KEGG pathways with
   SPU IDs, which are present in the KEGG database:
   [210]https://www.genome.jp/kegg-bin/show_organism?menu_type=pathway_map
   s&org=spu
   Curated pathway KEGG list (spu codes)
   Glycolysis and TCA 00010; 00020
   Sugar metabolism 00030; 00040; 00051; 00052; 00053; 00500; 00520
   Carbohydrate metabolism 00010; 00020; 00030; 00040; 00051; 00052;
   00053; 00500; 00520
   Other carbohydrate metabolism 00620; 00630; 00640; 00650; 00562
   Main amino acids 00250; 00260; 00270; 00280; 00290; 00310; 00220;
   00330; 00340; 00350; 00360; 00380; 00400
   Other amino acids 00410; 00430; 00440; 00450; 00470; 00480
   Fatty acid metabolism 00061; 00062; 00071
   Lipid metabolism 00061; 00062; 00071; 00100; 00561; 00564; 00565;
   00600; 00590; 00592; 01040
   Other lipid metabolism 00100; 00561; 00564; 00565; 00600; 00590; 00592;
   01040
   Nucleotide metabolism 00230; 00240
   Energy metabolism 00190; 00910; 00920
   [211]Open in a new tab

Data visualization

   For generating visualizations, we used the R/Bioconductor ecosystem.
   PLS-DA plots were generated by centering and scaling using the plsda
   function in the mixOmics R package^[212]72. Individual metabolite
   dotplots and corresponding features were implemented using ggplot2 and
   ggpubr^[213]73,[214]74. Heatmaps were generated using the
   ComplexHeatmap R package^[215]75. Table [216]2 was generated using the
   kableExtra package^[217]76. GSEA was performed using the
   clusterProfiler package^[218]77.

Reporting summary

   Further information on research design is available in the [219]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [220]Supplementary Information^ (6.8MB, pdf)
   [221]Peer Review file^ (14.4MB, pdf)
   [222]41467_2025_62697_MOESM3_ESM.pdf^ (79KB, pdf)

   Description of Additional Supplementary Files
   [223]Supplementary Data 1^ (52.4KB, xlsx)
   [224]Supplementary Data 2^ (88.6KB, xlsx)
   [225]Supplementary Data 3^ (102.4KB, xlsx)
   [226]Reporting Summary^ (98.1KB, pdf)

Source data

   [227]Source Data^ (2.8MB, xlsx)

Acknowledgements