Evolutionary modification of AGS protein contributes to formation of micromeres in sea urchins

Abstract

Evolution is proposed to result, in part, from acquisition of new developmental programs. One such example is the appearance of the micromeres in a sea urchin that form by an asymmetric cell division at the 4^th embryonic cleavage and function as a major signaling center in the embryo. Micromeres are not present in other echinoderms and thus are considered as a derived feature, yet its acquisition mechanism is unknown. Here, we report that the polarity factor AGS and its associated proteins are responsible for micromere formation. Evolutionary modifications of AGS protein seem to have provided the cortical recruitment and binding of AGS to the vegetal cortex, contributing to formation of micromeres in the sea urchins. Indeed, introduction of sea urchin AGS into the sea star embryo induces asymmetric cell divisions, suggesting that the molecular evolution of AGS protein is key in the transition of echinoderms to micromere formation and the current developmental style of sea urchins not seen in other echinoderms.

Fig. 1

AGS and Gαi are necessary for asymmetry in cell division and Vasa localization. a Diagrammatic representation of the transition from the 8–16 cell stage. Nucleus, Blue; Vasa, Red. The major organizing center of this embryo is the bottom-most (vegetal) cells, the micromeres. b AGS-kd (upper panels) result: In the control (2 mM Nanos-MO stock), the vegetal side of the aster is near to the vegetal cortex (arrowheads), whereas AGS-kd (2 mM AGS-MO stock) shifted the mitotic apparatus to the center of a cell (arrowheads). White squared regions are enlarged in e. Gαi-kd (lower panels) result: Gαi-kd diminished Gαi signal substantially at the vegetal cortex (arrows). c–j, Representative images and graphs for each measurement are shown. Each measurement was performed on 10 individual embryos that showed no AGS or Gαi-signal at the cortex as described in ‘Data analysis’ of the Methods. c, d Migration of nucleus to the vegetal cortex was repressed by AGS or Gαi-kd (measured by the relative value of vd/td). e, f Spindle orientation (r) and centrosome positioning (vd/td) was altered by AGS-kd. g, h Unequal cell division (micromere formation) was inhibited by AGS or Gαi-kd (measured by the relative value of Micro/Macro). i, j Asymmetric distributions of Vasa to the micromere-side of the spindle was diminished by AGS or Gαi-kd (measured by the relative value of Micro-side/ Macro-side). Both AGS- and Gαi-kd as well as PTX that locks Gαi in inactive state diminished the flatten-shaped micromere-side asters as well as Vasa signal from the vegetal cortex (dashed-circles), resulting in symmetric localization of Vasa over the spindle. Scale bars = 20 μm. k Proportion of embryos that formed micromeres and showed enriched Vasa signal at the 16-cell stage. () indicates total number of embryos counted. The statistical significance obtained by one-way Nova between controls and experimental samples was indicated as p-value. Each experiment was performed at least three independent times. For a detailed procedure of each data analysis, please refer to the “Data analysis” section of the Methods. Scale bars = 20 μm

Fig. 2

Stability of the regulatory protein complex at the vegetal cortex is sensitive to PLK1 activity. a Gαi-IP results. suNuMA, AGS and dynein interact with Gαi in extracts from the 8–16 cell stage. b Gαi-IP was performed after treatment with Nocodazole to destabilize microtubules. c–e Embryos were treated with Nocodazole at 8–16 cell stage for 10 min to disturb microtubule polymerization. c Nocodazole diminished spindle but Vasa (red) remained asymmetrically localized at the vegetal cortex independent of microtubules. d Vasa (green) remained at the close proximity to Gαi (red) after the treatment (arrowheads). e suNuMA (green) also remained linked to the Gαi (red) signal at the vegetal cortex after the treatment with Nocodazole (arrowheads). Scale bars = 520 μm. f Vasa-IP results. Vasa interacts with dynein and suNuMA in the absence of PLK1. In a, b, f, IPs were conducted at least three independent times and two each of the IP-ed samples are shown. Each of the top and bottom groups of immunoblots (demarcated by the line) was processed from a single blot. () indicates the molecular weight of each protein. The representative phenotypes of 70% or larger in each population are shown (n = 30 or larger)

Fig. 3

PLK1 activity alters Vasa and complex localization at the vegetal cortex and a model for AGS-dependent micromere formation. a PLK1 was localized over every spindle yet was depleted from the vegetal side of the 16-cell spindle (arrows). b membrane-mCherry-PLK1 mRNA (1 μg/μl stock) was co-injected with Vasa-GFP mRNA (1 μg/μl stock) and images were taken in live embryos. These embryos either formed no micromeres, or resulted in the formation of aberrant size and number (1–3 cells) of micromere-like cells at the 16-cell stage (arrows), whereas embryos with membrane-mCherry-PLK1-Kinase-Dead (PLK1-dead) or Vasa-GFP mRNA only showed no knockdown phenotype and formed micromeres with Vasa enrichment (arrowheads). For a–b, the representative phenotypes of 85% or larger in each population are shown (n = 100 or larger). c A working model for the molecular mechanisms of micromere formation and Vasa distribution in normal or Gαi/AGS perturbed embryos. Diagram shows a vegetal tier of blastomeres at 8–16 cell stage that is about to form the micromeres in this embryo. The amount of AGS on the centrosomes is significantly lower at this stage and omitted from this diagram. d A hypothetical model of molecular organization at the vegetal cortex. e A summary diagram of AGS-dependent asymmetric cell divisions. AGS-dependent asymmetric localization of cell fate factors enables a rapid lineage segregation and transition to the micromere specification pathway by pre-localizing organizer factors to the vegetal cortex during asymmetric cell divisions

Fig. 4

AGS is important for micromere’s function as organizers. a An endoderm marker, Endo1 (red) was highly expressed in the archenteron of gastrula (Day 2 PF) in the normal embryo, whereas no archenteron or Endo1 expression was found in AGS-MO injected embryos even with lower dose (1 mM). DNA, blue. The representative phenotypes of 85% or larger in each population are shown (n = 50 or larger). b, c AGS-MO (1 mM stock) injection resulted in embryonic death after blastula stage (AGS-kd only) yet co-injection of SpAgs-mRNA (0.25 μg/μl) that is insensitive to AGS-MO rescued the knockdown phenotype (AGS-kd + mRNA) to the level of Control embryos (1 mM Nanos-MO stock). The graph in c demonstrates % of embryos that formed micromeres at 16-cell stage (blue bars) and completed gastrulation at Day 3 PF (red bars). () indicates the number of embryos examined. Each experiment was performed at least three independent times. The statistical significance obtained by one-way Nova is  indicated as p-value. Scale bars = 50 μm. d, e AGS is essential for specification of micromeres as a developmental organizer. Diagram indicates the experimental procedure of micromere-transplant. Over 86% of embryos injected with AGS-MO (2 mM stock) remained as blastula and failed in gastrulation (AGS-kd). On the other hand, when these host embryos received wild-type micromeres (AGS-kd + Organizer (micromere) transplant), the RITC-labeled micromeres rescued the developmental defects, and differentiated normally into skeletogenic and coelomic pouch cells (red, arrows). In some cases, however, the transplanted cells entered the blastocoel and the embryos failed in gastrulation (a dashed circle in Blastula). Table in e indicates the number of embryos that successfully developed to each developmental stage (Larva, Gastrula, or Blastula), or that showed no detection of RITC-labeled transplanted cells (a dashed circle in RITC-negative Blastula), or that were observed in total (Total #). Scale bars = 50 μm

Fig. 5

Overexpression of AGS changes spindle orientation and anchoring. a AGS-GFP overexpression (AGS-OE; 1 μg/μl) changed overall embryo morphology at the 16-cell stage. Images were obtained by immunofluorescence with methanol fixation that diminishes the live GFP signal. The AGS-GFP signal was thus visualized by anti-GFP antibody. At M-phase, spindle orientation was randomized and many of the spindle poles were anchored to the nearby cortex (arrowheads in the magnified view). ~50% of embryos showed phenotype, n = 14. b, c AGS-OE caused ectopic recruitment of microtubules labeled by mCherry-EMTB to the cortex (arrows) in over 80% of embryos, n = 25. Images taken are from live embryos. Each experiment was performed at least three independent times. Scale bars = 20 μm

Fig. 6

AGS, Gαi, and Vasa proteins are enriched in micromere-like cells of the pencil urchin embryo but not in the sea star embryo. a Immunofluorescence images of AGS (green) and Gαi (red) in sea star embryos. No specific signal was found in any region of the cortex during early embryogenesis in over 90% of embryos, n < 50. b, d Immunofluorescence images of pencil urchin embryos. In b, at 8 and 16 cell stages, embryos were enriched in AGS and Gαi at the vegetal cortex (arrows). These proteins were also enriched within the cytoplasm of the vegetal blastomeres or micromere-like cells in some embryos. Embryos lacking micromere-like cells also lacked AGS and Gαi enrichment on the cortex. n = 20. In c AGS was localized to the centrosomes in addition to the cortex, similar to that of sea urchin (S. purpuratus). n = 10. In d, Vasa (green) was enriched in the cytoplasm of the micromere-like cells, and enrichment was absent when micromere-like cells were not present. n = 15. The representative phenotypes of 80% or larger in each population are shown. Each experiment was performed at least two independent times. Scar bars = 10 μm

Fig. 7

Sea urchin AGS induces asymmetric cell divisions during early embryogenesis and extra invaginations after blastulation in sea star embryos. a A summary diagram that depicts Vasa (red) and AGS (green) localization patterns during 8–16 cell stage. Sea urchin embryos undergo asymmetric cell division to form micromeres (organizers) accompanied by Vasa and AGS enrichment at the vegetal pole at 16-cell stage, whereas sea star embryos undergo symmetric cell division with no enrichment of either molecule. b–f Brightfield images (b), Live imaging (d), or Immunofluorescence images (e) of sea star embryos expressing 1.5 μg/μl stock of SpAGS-GFP mRNA that underwent random asymmetric cell divisions to form micromere-like cells during early embryogenesis (arrowheads in (b) and (d)) and extra invaginations at the larval stage (arrowheads in (e)). On the other hand, the controls injected with 1.5 μg/μl stock of SpAGS-dGoLoco#1 or SpAGS-dC-term that lacks GoLoco#1 or the entire C-terminal domain, respectively, or with dye underwent symmetric cell divisions (c) and single invagination (f). The number of embryos that underwent asymmetric cell divisions from 2- to 16-cell stage (c) or extra invagination at Day 2 (f) were scored and shown in %. n indicates the total number of embryos scored. In d, AGS-GFP was enriched on the spindle and at the cortex (arrow). White squared regions are enlarged in bottom row images. A white squared region is enlarged in the bottom row images. In e, Vasa (red) was enriched in the germline of the control larva (arrow), which was not found in the experimental group. n in the graphs c and f, indicates the total number of embryos scored. Each experiment was performed at least three independent times. Each image shows the representative phenotypes scored in the corresponding graph. Scale bars = 50 μm

Fig. 8

SpAGS-overexpression induces endo-mesodermal lineages in sea star embryos. a–c 1.5 μg/μl stock of SpAGS-GFP mRNA was injected into unfertilized eggs and the micromere-like cells were labeled with red fluorescent dextran (dye) during 8–32 cell stage (a). At Day 2, embryos were scored for each category based on where the dye signal was found (b, c). Arrows indicate the signal enrichment at the bottom of the gut. At the occasion when the signal was scattered in various tissues, the embryo was counted for each category. n indicates the total number of embryos scored. d–g 0.75 μg/μl stock of SpAGS-GFP mRNA or dye was injected into one of the blastomeres at the 8-cell stage (d). At Day 2 (e), embryos that were injected with dye (f) or with dye and SpAGS-GFP mRNA (g) were scored for each category based on where the dye signal was found. When the signal was scattered in various tissues, the embryo was scored in  each lineage category. In e, arrow indicates AGS-GFP enrichment at the vegetal-most region of the gut and arrowheads indicate the signal enrichment of extra invaginations. n in the graphs f–g indicates the total number of embryos scored. Each experiment was performed at least three independent times. Each image shows the representative phenotypes scored in the corresponding graph. Scale bars = 50 μm

Fig. 9

The evolutionary modification of AGS proteins among echinoderms. a Predicted motifs of each AGS protein are depicted based on NCBI blast search results (See Supplementary Tables. 6 and 7 for details). Conserved TPR domains are highlighted in purple and GoLoco motifs in yellow (See Supplementary Table 7B for the specific sequence of each GoLoco motif). Less conserved or partial domains are colored in light purple or light yellow, respectively. () indicates species used to construct each cartoon. Vasa, red; AGS, green. b A model for introduction of micromeres during echinoid diversification. The timing of micromere formation coincides with introduction of GoLoco motif#1 in AGS protein during echinoid diversification