Antagonism of cell adhesion by an alpha-catenin mutant, and of the Wnt-signaling pathway by alpha-catenin in Xenopus embryos

Abstract

In Xenopus laevis development, beta-catenin plays an important role in the Wnt-signaling pathway by establishing the Nieuwkoop center, which in turn leads to specification of the dorsoventral axis. Cadherins are essential for embryonic morphogenesis since they mediate calcium-dependent cell-cell adhesion and can modulate beta-catenin signaling. alpha-catenin links beta-catenin to the actin-based cytoskeleton. To study the role of endogenous alpha-catenin in early development, we have made deletion mutants of alphaN-catenin. The binding domain of beta-catenin has been mapped to the NH2-terminal 210 amino acids of alphaN-catenin. Overexpression of mutants lacking the COOH-terminal 230 amino acids causes severe developmental defects that reflect impaired calcium-dependent blastomere adhesion. Lack of normal adhesive interactions results in a loss of the blastocoel in early embryos and ripping of the ectodermal layer during gastrulation. The phenotypes of the dominant-negative mutants can be rescued by coexpressing full-length alphaN-catenin or a mutant of beta-catenin that lacks the internal armadillo repeats. We next show that coexpression of alphaN-catenin antagonizes the dorsalizing effects of beta-catenin and Xwnt-8. This can be seen phenotypically, or by studying the effects of expression on the downstream homeobox gene Siamois. Thus, alpha-catenin is essential for proper morphogenesis of the embryo and may act as a regulator of the intracellular beta-catenin signaling pathway in vivo.

Figure 1

α-catenin–GFP and β-catenin constructs used in this work. GFP was added to either the NH₂ terminus of full-length, or to the NH₂ or COOH termini of deletion mutants of chick αN-catenin. 1 represents the start methionine of αN-catenin, and 906 is the terminal amino acid residue. See Materials and Methods for details on the construction of these mutants. Full-length β-catenin has a COOH-terminal hemagglutinin (HA) tag, and ΔArm has a COOH-terminal myc tag.

Figure 2

Expression of α-catenin–GFP fusion proteins in Xenopus embryos and the localization of the β-catenin binding site. (A) Expression of proteins detected by anti-GFP. (B) After immunoprecipitations with anti–β-catenin, associated α-catenin was detected using anti-GFP. In both A and B, the numbering is as follows: lane 1, GFP injected; lane 2, uninjected; lane 3, αNcatNtermGFP + GFP; lane 4, αNcatNtermGFP + ΔArm; lane 5, GFPαNcatCterm; lane 6, GFPαNcat; lane 7, GFPαNcatNterm. Note that GFPαNcatCterm, the protein lacking the NH₂ terminus of αN-catenin, does not bind to β-catenin (B, lane 5). Results show that the presence of GFP does not prevent α-catenin binding to β-catenin, provided that the NH₂-terminal binding domain is present. In addition, coexpression of ΔArm does not affect expression levels. In each case, 1.5 ng total RNA was injected into one blastomere at the two-cell stage.

Figure 3

Mutants lacking the COOH terminus of αN-catenin result in severe gastrulation defects. (A) GFPαNcat-injected embryos develop normally. (B) GFPαNcatCterm-expressing embryos develop normally. (C and D) Both αNcatNtermGFP- and GFPαNcatNterm-injected mRNAs induce defects beginning at stage 10, here shown by rips (arrows) occurring at the dorsal lips of the embryos. Embryos are shown with the ventral side toward the top. mRNA was injected into the animal pole of one cell of two-cell embryos.

Figure 4

ΔArm binds α-catenin– GFP fusion proteins. The 9E10 monoclonal antibody that recognizes the myc tag on ΔArm was used to immunoprecipitate, and anti-GFP was used with subsequent immunoblotting to detect α-catenin–GFP fusion proteins in the precipitates. Embryos were coinjected with 0.3 ng of the first mRNA and 1.2 ng of the second. First lane, GFP + ΔArm; second lane, αNcatNtermGFP + ΔArm; third lane, uninjected.

Figure 5

ΔArm and GFPαNcat rescue the defects caused by dominant-negative α-catenin mutants. (A, B, and C) αNcatNtermGFP + GFP, ΔArm, or GFPαNcat, respectively. (D–F) GFPαNcatNterm + GFP, ΔArm, or GFPαNcat. In all cases, 0.3 ng of αNcatNtermGFP or GFPαNcatNterm mRNA was coinjected with 1.2 ng of the second mRNA. Mutants in A and D do not develop further than gastrulation. Rescued mutants gastrulate (B and C, and E and F) and continue to develop normally. mRNA was injected into the animal pole of one cell of two-cell embryos.

Figure 6

Histology of gastrulating embryos expressing α-catenin– GFP fusion proteins. Stage 10–11 embryos were fixed in MEMFA and imbedded in paraplast (A and C) or plastic (B, D, E, and F) for sectioning. Arrowheads point to the blastocoel (bc). The abbreviation dl denotes the presumptive dorsal side of the embryos. (A) An uninjected normal embryo. (B) An embryo injected with GFPαNcatCterm mRNA, (C) αNcatNtermGFP, and (D) GFPαNcatNterm. Embryos expressing αNcatNtermGFP completely lack the blastocoel and are disorganized compared to normal embryos. Also noticeable is the lack of integrity of the ectodermal layers in the two mutants. Higher magnification emphasizes the disorganization of cells in an embryo expressing αNcatNtermGFP compared to a normal embryo expressing GFPαNcatCterm. mRNA was injected into the animal pole of both cells of two cell embryos. Bars: (A–D) 200 μm; (E and F) 100 μm.

Figure 7

Redistribution of GFPαNcatNterm and αNcatNtermGFP from the glycoprotein fraction to the soluble fraction by ΔArm. (A) Soluble fractions immunoblotted with the anti-GFP polyclonal antibody. (B) Glycoprotein fractions immunoblotted with the anti-GFP polyclonal antibody. Lanes in both cases are 1, uninjected; 2, GFP injected; 3, GFPαNcatCterm injected; 4, αNcatNtermGFP + GFP; 5, αNcatNtermGFP + ΔArm; 6, GFPαNcatNterm + GFP; 7, GFPαNcatNterm + ΔArm; 8, GFPαNcat. Note the decrease in B between lanes 4 and 5 and lanes 6 and 7, implying that ΔArm acts by binding to the α-catenin mutants, keeping them from binding to the cadherin complex. When coinjections were done, 0.3 ng of the first mRNA was coninjected with 1.2 ng of the second mRNA.

Figure 8

Levels of endogenous Xenopus α-catenin bound to cadherins decrease in embryos expressing the dominant-negative mutants αNcatNtermGFP or GFPαNcatNterm. Embryos were lysed in NP-40 buffer and immunoprecipitated using the monoclonal antibody to C-cadherin 6B6. The immunoprecipitates were immunoblotted with the polyclonal anti–αN-catenin antibody (CME). The lanes correspond to embryos injected with mRNA for 1, uninjected; 2, GFP; 3, αNcatNtermGFP; 4, GFPαNcatNterm; 5, uninjected. Lane 5 represents uninjected embryos subjected to immunoprecipitation with an anti-myc epitope antibody (9E10) as a negative control. Embryos were injected into both cells at the two-cell stage to obtain protein expression in the highest possible number of cells. Notice that the levels of endogenous α-catenin decrease significantly in lanes 3 and 4 compared to lanes 1 and 2.

Figure 9

Animal cap assays reveal the dissociated nature of blastomeres expressing dominant-negative α-catenin. (A) Animal caps in 1× MMR normally heal and involute after 1 h as shown in GFPαNcat-injected embryos. (B) αNcatNtermGFP-expressing animal caps fail to involute, and dissociated blastomeres disperse after 1 h. (C) Dissociated GFP-expressing animal caps reaggregate in 2 mM Ca²⁺, whereas in D animal caps expressing αNcatNtermGFP do not. mRNA was injected into the animal pole of both cells of two-cell embryos.

Figure 10

Coexpression of α-catenin diminishes the dorsalizing effects of Xwnt-8 and β-catenin. (A) Noninjected normal tadpoles. (B) Embryos expressing GFP + GFPαNcat. (C) Embryos expressing Xwnt-8 + GFP. (D) Xwnt-8 + GFPαNcat. (E) β-catenin + GFPαNcatCterm. (F) β-catenin + GFPαNcat. Embryos coexpressing α-catenin with Xwnt-8 develop longer trunks and are less dorsalized than embryos coexpressing GFP. Embryos ventrally expressing GFPαNcat + GFP develop normally (B). GFPαNcat antagonizes the induction of a secondary dorsoanterior axis by β-catenin (F). In B–D, embryos were coinjected with 0.125 ng of the first mRNA and 3 ng of the second. In E and F, 0.2 ng of β-catenin mRNA was coinjected with 3.0 ng of the αN-catenin constructs. One ventral blastomere was injected in each four-cell embryo. 1-mm glass beads hold the embryos upright.

Figure 11

Full-length α-catenin can antagonize the endogenous induction of a dorsoanterior axis when expressed in dorsal blastomeres. 3.0 ng of either GFPαNcatCterm, which does not bind to β-catenin, or GFPαNcat, the full-length construct, were injected into both dorsal blastomeres of four-cell embryos. Embryos expressing GFPαNcat often develop the ventralized phenotypes illustrated here (B and D), with no anterior head structures present. Embryos expressing high levels of GFPαNcatCterm develop normally (A and C). Two stages are shown.

Figure 12

(A) α-Catenin inhibits the induction of the homeobox-containing gene, Siamois, by full-length β-catenin or Xwnt-8 in animal caps. One cell of two-cell embryos was injected in the animal hemisphere, and animal cap explants were taken at stage 9 and allowed to develop to stage 10.5, when total RNA was prepared. In lanes 2–6, coinjections were with 0.125 ng of the first mRNA and 3.0 ng of the second. Lane 1, uninjected whole embryos; lane 2, full-length β-catenin + GFP; lane 3, full-length β-catenin + αNcatNtermGFP; lane 4, full-length β-catenin + GFPαNcat; lane 5, GFP + αNcatNtermGFP; lane 6, GFP + GFPαNcat; lane 7, 0.125 ng Xwnt-8 + 3.0 ng GFPαNcatCterm; lane 8, 0.125 ng Xwnt-8 + 3.0 ng GFPαNcat; lane 9, 0.125 ng Xwnt-8 + 1.5 ng GFP; lane 10, 0.125 ng Xwnt-8 + 3 ng GFP; lane 11, 0.125 ng Xwnt-8 + 0.75 ng GFPαNcat; lane 12, 0.125 ng Xwnt-8 + 1.5 ng GFPαNcat; lane 13, 0.125 ng Xwnt-8 + 3 ng GFPαNcat; 14, 0.125 ng GFP + 3 ng GFPαNcat. The ubiquitously expressed elongation factor 1α (EF1α) was included as a loading control. RNAse protection assays show that either full-length αN-catenin (GFPαNcat) or the dominant-negative αN-catenin (αNcatNtermGFP) decrease the levels of Siamois mRNA induced in animal caps by β-catenin or Xwnt-8. Moreoever, increasing concentrations of GFPαNcat result in proportionate decreases in Siamois induction by Xwnt-8. (B) α-Catenin inhibits the endogenous expression of Siamois when injected into dorsal blastomeres at the four-cell stage. Both dorsal blastomeres were injected with mRNA, and embryos were extracted for total RNA at stage 10.5. Lane 1, 1.5 ng GFPαNcat; lane 2, 3.0 ng GFPαNcat; lane 3, 3.0 ng GFPαNcatCterm; lane 4, uninjected embryos. Embryos injected with GFPαNcat mRNA (which develop into ventralized embryos) express lower levels of Siamois than normal embryos or embryos injected with GFPαNcatCterm (which develop normally).

Figure 12

(A) α-Catenin inhibits the induction of the homeobox-containing gene, Siamois, by full-length β-catenin or Xwnt-8 in animal caps. One cell of two-cell embryos was injected in the animal hemisphere, and animal cap explants were taken at stage 9 and allowed to develop to stage 10.5, when total RNA was prepared. In lanes 2–6, coinjections were with 0.125 ng of the first mRNA and 3.0 ng of the second. Lane 1, uninjected whole embryos; lane 2, full-length β-catenin + GFP; lane 3, full-length β-catenin + αNcatNtermGFP; lane 4, full-length β-catenin + GFPαNcat; lane 5, GFP + αNcatNtermGFP; lane 6, GFP + GFPαNcat; lane 7, 0.125 ng Xwnt-8 + 3.0 ng GFPαNcatCterm; lane 8, 0.125 ng Xwnt-8 + 3.0 ng GFPαNcat; lane 9, 0.125 ng Xwnt-8 + 1.5 ng GFP; lane 10, 0.125 ng Xwnt-8 + 3 ng GFP; lane 11, 0.125 ng Xwnt-8 + 0.75 ng GFPαNcat; lane 12, 0.125 ng Xwnt-8 + 1.5 ng GFPαNcat; lane 13, 0.125 ng Xwnt-8 + 3 ng GFPαNcat; 14, 0.125 ng GFP + 3 ng GFPαNcat. The ubiquitously expressed elongation factor 1α (EF1α) was included as a loading control. RNAse protection assays show that either full-length αN-catenin (GFPαNcat) or the dominant-negative αN-catenin (αNcatNtermGFP) decrease the levels of Siamois mRNA induced in animal caps by β-catenin or Xwnt-8. Moreoever, increasing concentrations of GFPαNcat result in proportionate decreases in Siamois induction by Xwnt-8. (B) α-Catenin inhibits the endogenous expression of Siamois when injected into dorsal blastomeres at the four-cell stage. Both dorsal blastomeres were injected with mRNA, and embryos were extracted for total RNA at stage 10.5. Lane 1, 1.5 ng GFPαNcat; lane 2, 3.0 ng GFPαNcat; lane 3, 3.0 ng GFPαNcatCterm; lane 4, uninjected embryos. Embryos injected with GFPαNcat mRNA (which develop into ventralized embryos) express lower levels of Siamois than normal embryos or embryos injected with GFPαNcatCterm (which develop normally).