Paraxial protocadherin mediates cell sorting and tissue morphogenesis by regulating C-cadherin adhesion activity

Abstract

Little is known about how protocadherins function in cell adhesion and tissue development. Paraxial protocadherin (PAPC) controls cell sorting and morphogenetic movements in the Xenopus laevis embryo. We find that PAPC mediates these functions by down-regulating the adhesion activity of C-cadherin. Expression of exogenous C-cadherin reverses PAPC-induced cell sorting and gastrulation defects. Moreover, loss of endogenous PAPC results in elevated C-cadherin adhesion activity in the dorsal mesoderm and interferes with the normal blastopore closure, a defect that can be rescued by a dominant-negative C-cadherin mutant. Importantly, activin induces PAPC expression, and PAPC is required for activin-induced regulation of C-cadherin adhesion activity and explant morphogenesis. Signaling through Frizzled-7 is not required for PAPC regulation of C-cadherin, suggesting that C-cadherin regulation and Frizzled-7 signaling are two distinct branches of the PAPC pathway that induce morphogenetic movements. Thus, spatial regulation of classical cadherin adhesive function by local expression of a protocadherin is a novel mechanism for controlling cell sorting and tissue morphogenesis.

Figure 1.

PAPC does not mediate homophilic cell adhesion. (A) Cell aggregation assay of CHO cells stably expressing FL-PAPC (FL-PAPC-CHO), M-PAPC (M-PAPC-CHO), C-cadherin (C-CHO, as positive control), or GFP (GFP-CHO, as negative control). (B) Cell attachment flow assay of stable FL-PAPC-CHO and M-PAPC-CHO cells on PAPC-EC.Fc substrate. Flow assay of C-CHO on C-cad-EC.Fc substrate was performed as positive control, and assay of FL-PAPC-CHO on C-cad-EC.Fc was used as negative control. (C) Cell attachment flow assay of stable FL-PAPC-A431 and M-PAPC-A431 cells on PAPC-EC.Fc substrate. Mock-transfected A431 (Vector-A431) cells were used as negative control, and adhesion of M-PAPC-A431 on human E-cad-EC.Fc was the positive control. (D) Cell attachment flow assay of stable PAPC-expressing XTC cells (PAPC-XTC) and parental XTC cells on PAPC-EC.FC substrate. The adhesion of XTC and PAPC-XTC on C-cadherin (C-cad-EC.Fc) substrate was positive control. (E) Blastomere adhesion assay with animal cap cells that ectopically express GFP, FL-PAPC (FL), or M-PAPC (M). 1.5 ng RNA was injected into embryos. Adhesion substrates were coated with 0.1 mg/ml PAPC-EC.Fc, 0.1 mg/ml PAPC-EC.His, or 10 μg/ml of C-cad-EC.Fc. (F) Blastomere adhesion assay with dorsal trunk mesodermal blastomeres from stage 12 embryos. Adhesion substrates were coated as in E. Expression of PAPC in cells used for adhesion assays was shown by anti-PAPC Western blot on the right of each graph. Error bars are the SEM.

Figure 2.

FL-PAPC and M-PAPC induce cell sorting and decreased adhesion with similar activity. (A–C) Cell dispersal assays. 500 pg of control GFP mRNA (A), FL-PAPC mRNA (B), or M-PAPC mRNA (C) was coinjected with 200 pg of NLS-GFP mRNA into one animal blastomere at the 32-cell stage. Pictures of the injected embryos were taken under fluorescence-microscope at stage 14. (D–I) Cell dissociation and reaggregation assays. Dissociated animal cap blastomeres from control GFP mRNA- (D and G), FL-PAPC mRNA- (E and H), or M-PAPC mRNA- (F and I) injected embryos were mixed with those from uninjected embryos and allowed to form coaggregates overnight. All injections were traced by coinjecting NLS-GFP mRNA. (D–F) Overview of the aggregates. (G–I) Bisectional view of the aggregates. (J–L) Blastomere aggregation assays. As in D–I, blastomeres expressing GFP (J), FL-PAPC (K), or M-PAPC (L) were allowed to aggregate in the presence of calcium for 1 h on a rocker. At the end of the assay, pictures of the aggregates were taken.

Figure 3.

Both FL-PAPC and M-PAPC down-regulate C-cadherin adhesion activity. (A) Blastomere adhesion assay of GFP mRNA- (control), FL-PAPC mRNA-, or M-PAPC mRNA-injected embryos (1.5 ng/embryo) on 4 μg/ml C-cad-EC.Fc–coated substrates. **, P < 0.001 (by t-test) compared with the control GFP-expressing blastomeres. (B) M-PAPC expression does not change either total or cell surface C-cadherin protein levels in blastomeres. Embryos were injected as in A. C-cadherin levels in total embryo lysates (lanes 1–2), stage 9 animal cap explants (lanes 3–4), and dissociated animal cap cells that were either mock treated (lanes 5–6) or trypsin/EDTA treated (lanes 7–8) were determined by Western blotting with anti– C-cadherin mAb (6B6). Expression of M-PAPC was confirmed by anti-PAPC blotting, and anti–α-tubulin blots served as loading control. Arrowhead, a fragment of M-PAPC. (C) C-cadherin–activating antibody AA5 (1 μg/ml Fab fragment) reverts the down-regulation of C-cadherin–mediated adhesion induced by either activin-treatment (left) or M-PAPC expression (right). **, P < 0.01. (D) Blastomere adhesion to fibronectin is not changed by PAPC-expression. (E) Blastomere adhesion to anti-IL2R mAb BB10 is not changed by PAPC-expression. FL, FL-PAPC; M, M-PAPC; IL2R, IL2Rα. Error bars are the SEM.

Figure 4.

Overexpression of C-cadherin reverts M-PAPC–induced cell sorting and gastrulation defects. (A) Coexpression of C-cadherin reverts M-PAPC–induced cell sorting. Cell dispersal assays were performed by coinjecting different doses of C-cadherin mRNA, along with 300 pg of M-PAPC mRNA and 180 pg of NLS-GFP mRNA, as tracer. As control, 200 pg of NLS-GFP mRNA alone was injected (GFP). The C-cadherin mRNA doses were 0 pg (M), 75 pg (M + 0.25C-cad), 150 pg (M + 0.5C-cad), 300 pg (M + 1C-cad), and 600 pg (M + 2C-cad). (B) Exogenous C-cadherin expression rescues M-PAPC–induced blastopore-closure defects. 1 ng of GFP mRNA (Control), 0.5 ng of M-PAPC mRNA alone (M-PAPC), or 0.5 ng M-PAPC mRNA plus 1 ng of C-cadherin mRNA (M-PAPC + C-cad) were injected at the 4-cell stage into the animal poles of all four blastomeres. At stage 12 (top row), control embryos showed normal, nearly closed blastopores (30/30); M-PAPC mRNA-injected embryos failed to close their blastopores and showed an exogastrula phenotype (0/10 normal); coinjection of C-cadherin mRNA rescued the blastopore-closure defect (30/31 normal). In another experiment, embryos were allowed to develop to stage 18 (bottom row). GFP mRNA-injected embryos appeared normal (20/20), and M-PAPC mRNA-injected embryos failed to close their blastopores and exhibited exogastrula phenotype (0/20 normal), whereas coinjection of C-cadherin mRNA significantly rescued the defect (15/20 normal).

Figure 5.

Loss of endogenous PAPC expression results in increased C-cadherin–mediated adhesion in DMZ cells and gastrulation defects that can be rescued by decreased C-cadherin adhesion. (A) Effects of loss of PAPC on C-cadherin–mediated adhesion in DMZ cells. 80 ng of control (COMO) or PAPC morpholinos (PAPCMO) were injected into the DMZ of 4–8–cell stage embryos. At stage 10.5, blastomeres were isolated from the VMZ or DMZ of five COMO- or PAPCMO-injected embryos and then tested for adhesion to purified C-cad-EC.Fc. Half of the blastomeres from the DMZ of control embryos were treated with 1 μg/ml AA5 Fab fragment before adhesion assay. **, P < 0.001 compared with the rest samples. (B and C) Rescue of PAPC morpholino-induced gastrulation defects by DN C-cadherin. 80 ng COMO, PAPCMO, or PAPCMO supplemented with either 25 pg C-cadherin cytoplasmic tail RNA (Ctail) or 50 pg FL-PAPC(-UTR) RNA (FL) was injected into the DMZ of 4-cell stage embryos. At stage 12.5, pictures of six randomly picked embryos from each group of embryos were taken (B). Meanwhile, the relative blastopore size (the ratio of the blastopore diameter versus the embryo diameter) of each injected embryo was measured and graphed (C). Error bars are the SEM. ***, P < 0.0001 compared with the rest groups of embryos. n, the number of embryos in each group.

Figure 6.

Requirement for PAPC in the activin-induced regulation of C-cadherin adhesion activity and animal cap morphogenesis. (A) Activin induces PAPC protein expression in animal cap explants. Stage 8 animal caps were treated with 5 ng/ml activin for 0, 30, 60, 120, and 180 min and immediately processed for anti-PAPC Western blot analysis. (B) PAPC morpholinos (PAPCMO) suppress activin-induced PAPC expression and endogenous PAPC expression. Lanes 1–4: stage 9 animal caps analyzed by anti-PAPC Western blotting. 80 ng COMO or PAPCMO were injected into the animal hemisphere at the 2–4–cell stage, and animal caps were excised, dissociated, and treated with or without 5 ng/ml activin for 1.5 h. Lanes 5–6: whole embryos analyzed by anti-PAPC Western blotting. 80 ng of COMO or PAPCMO were injected into the dorsal side at the 4-cell stage, and incubated to stage 12. The level of PAPC was normalized to the level of α-tubulin. (C) PAPCMO block activin-induced down-regulation of C-cadherin activity. Blastomere adhesion assays on 4 μg/ml C-cad-EC.Fc–coated substrates were performed with the blastomeres described in B (lanes 1–4). In addition, blastomeres from embryos injected with PAPCMO plus morpholino-resistant FL-PAPC mRNA or with FL-PAPC mRNA alone were also assayed in parallel. **, P < 0.01 compared with COMO-injected, untreated blastomeres. Error bars are the SEM. (D) PAPCMO blocked activin-induced animal cap elongation. 80 ng of COMO or PAPCMO were injected into the animal hemisphere of 2–4 cell stage embryos. (1) The control COMO-injected caps fully elongated (30/30). The PAPCMO-injected caps were divided into two groups according to their phenotypes: (2a) partial elongation (12/30) and (2b) no elongation (18/30). Five caps from each of the groups were processed for Western blot analyses with anti-PAPC and anti–α-tubulin antibodies.

Figure 7.

PAPC-regulation of C-cadherin adhesion and cell sorting is independent of Frizzled-7. (A) Frizzled-7 morpholinos (Xfz7MO) and DN Frizzled-7 (DN-Xfz7) cause sever gastrulation defects. 40 ng of control morpholino (COMO), 40 ng of Xfz7MO, or 2.4 ng of DN-Xfz7 mRNA was injected into 2-cell stage embryos, which were allowed to develop to stage 20. (B) PAPC down-regulates C-cadherin activity even in the presence of Xfz7-MO or DN-Xfz7. Embryos were injected as in A. At the 4-cell stage, half of the injected embryos were further injected with 1.2 ng of FL-PAPC mRNA. Blastomere adhesion assays were performed on 4 μg/ml C-cad-EC.Fc–coated substrates. Error bars are the SEM. (C) Xfz7-MO and DN-Xfz7 have no affects on M-PAPC–induced cell sorting. (top row) Cell dispersal assays with 10 ng of COMO, 10 ng of Xfz7MO, or 500 pg of DN-Xfz7mRNA (all with NLS-GFP mRNA as tracer) injected into one blastomere of 32-cell stage embryos. (bottom row) Cell dispersal assays with 40 ng of COMO, 40 ng of Xfz7MO, or 4 ng of DN-Xfz7mRNA injected into 2-cell stage embryos, followed by injection of 500 pg of M-PAPC mRNA plus 200 pg of NLS-GFP mRNA (as tracer) into one blastomere at the 32-cell stage.

Figure 8.

Models for the role of PAPC in activin-induced regulation of C-cadherin adhesion and tissue morphogenesis. (A) Relationship between PAPC, Frizzled-7 signaling, and regulation of C-cadherin–mediated adhesion. The membrane-bound PAPC extracellular domain, as well as wild-type PAPC, down-regulates C-cadherin adhesion activity either directly or indirectly, and the regulation of C-cadherin adhesion activity contributes to convergence and extension cell movements. PAPC also interacts with Xfz7 and participates in the activation of RhoA and JNK by Xfz7-mediated signaling to affect tissue separation and convergent extension. Full-length PAPC is required for Xfz7-mediated tissue separation. (B) A signaling cascade that mediates the activin-induced tissue morphogenesis. The findings from this study are shown in bold. Additional signaling steps shown in other studies to be important for activin-induced morphogenesis are indicated by dotted arrows.