Probing Cadherin Interactions in Zebrafish with E- and N-Cadherin Missense Mutants

Abstract

Cadherins are cell adhesion molecules that regulate numerous adhesive interactions during embryonic development and adult life. Consistent with these functions, when their expression goes astray cells lose their normal adhesive properties resulting in defective morphogenesis, disease, and even metastatic cancer. In general, classical cadherins exert their effect by homophilic interactions via their five characteristic extracellular (EC) repeats. The EC1 repeat provides the mechanism for cadherins to dimerize with each other whereas the EC2 repeat may facilitate dimerization. Less is known about the other EC repeats. Here, we show that a zebrafish missense mutation in the EC5 repeat of N-cadherin is a dominant gain-of-function mutation and demonstrate that this mutation alters cell adhesion almost to the same degree as a zebrafish missense mutation in the EC1 repeat of N-cadherin. We also show that zebrafish E- and N-cadherin dominant gain-of-function missense mutations genetically interact. Perturbation of cell adhesion in embryos that are heterozygous mutant at both loci is similar to that observed in single homozygous mutants. Introducing an E-cadherin EC5 missense allele into the homozygous N-cadherin EC1 missense mutant more radically affects morphogenesis, causing synergistic phenotypes consistent with interdependent functions being disrupted. Our studies indicate that a functional EC5 repeat is critical for cadherin-mediated cell affinity, suggesting that its role may be more important than previously thought. These results also suggest the possibility that E- and N-cadherin have heterophilic interactions during early morphogenesis of the embryo; interactions that might help balance the variety of cell affinities needed during embryonic development.

Keywords: adhesion; affinity; cadherin; extracellular repeat; heterodimer; heterophilic; homodimer; homophilic.

Figure 1

E- and N-cadherin tend to be expressed in a complementary pattern. Expression of: E-cadherin (A–F, M–O, and T and U) and N-cadherin (G–L, P–R, and S). (A–L) Low-magnification view of wild-type embryos at similar stages of development; shown in side view, except (B, F, H, and L) shown in dorsal view and (D) shown in an oblique view. Prechordal mesendoderm (white arrowheads) and the dorsal forerunner cluster (black arrows) both express E- and N-cadherin. (M–R) Higher-magnification view of the dorsal forerunner cluster in wild-type embryos at similar stages of development, (M and P) are higher-magnification views of embryos shown in (B and H). Note that the forerunner cluster is tightly opposed to the blastoderm margin. Bar, 50 μm. (S–U) Low-magnification view of homozygous mutant embryos and their siblings at similar stages of development shown in dorsal view: (S) the E-cadherin nonsense allele, (T) the N-cadherin nonsense allele, and (U) the N-cadherin EC1 missense allele. RNA transcripts are comparable between homozygous mutants and wild-type siblings, allowing for differences in the embryos shape. Arrow in S indicates the forerunner cluster, which in all the epiboly mutants continues to migrate vegatalward. Arrows in (T) show the wider anterior axis typical of all N-cadherin mutants due to the impaired convergence of cells to the dorsal midline, and arrows in (U) show that migration of the epidermis over the neural keel is defective in the N-cadherin missense mutant compared to its wild-type sibling. e-cad, E-cadherin; n-cad, N-cadherin; WT, wild-type.

Figure 2

Forerunner cell development and left–right asymmetry are perturbed in E- and N-cadherin missense mutants. (A) Location of E- and N-cadherin zebrafish mutants as determined from reported sequences (Lele et al. 2002; Malicki et al. 2003; Masai et al. 2003; Birely et al. 2005; Kane et al. 2005; von der Hardt et al. 2007; Warga and Kane 2007; Yamaguchi et al. 2010). Only the EC repeat alleles are shown. At the E-cadherin locus, all homozygous mutants display a common phenotype where epiboly of the deep cells slows relative to the enveloping layer and dorsal forerunner cells (see Figure 1S), followed by retraction of the blastoderm and disassociation of the embryo, usually before onset of somitogenesis. At the N-cadherin locus, all homozygous mutants display a common phenotype where convergence of the deep cells is defective (see Figure 1, T and U). Embryos eventually display pronounced defects in brain morphology, including cells detaching from the neuroepithelial surface and, where examined, abnormalities in muscle, eye, and tail development. Solid arrows denote a nonsense mutation, open arrows denote a missense mutation, asterisk denotes a mutation reported to have dominant effects, and red denotes the alleles used in this study. For further information see Table 1. Abbreviations: Cyt, cytoplasmic domain; EC1–5, extracellular repeats; Pro, prodomain; S, signal sequence; TM, transmembrane domain. (B and C) High-magnification view of the dorsal forerunner cells (arrows), showing localization of β-catenin. (B) A representative embryo from a tb08 (ncad; EC1) clutch, exhibiting a phenotypic wild-type appearance (39/39). (C) Representative embryos from a dtv43 (ecad; EC5) clutch showing more intense staining in the forerunner cells of the homozygous mutant (8/8) compared to the heterozygous zygotic–maternal dominant (ZMD) mutant (27/28), even in isolated cells (arrowheads). Wild-type embryos had less intense staining in the forerunner cells (12/12) and resembled embryo shown in (B). Note also that the forerunner cluster in the homozygous mutant is at the leading edge of the enveloping layer, but already far from the deep cell margin (17/17). In the heterozygous ZMD mutant the forerunner cluster is closer to the deep cell margin, but not completely touching (47/47). In the wild-type sibling, the forerunner cluster abuts the blastoderm margin (23/23), as seen in (B). Bar, 25 μm. (D and E) High-magnification view of the forerunner cluster once cells have begun to aggregate into rosettes, visualized with sox32. (D) Showing reduced cell number in the homozygous tb08 (ncad; EC1) mutant (11/11) compared to phenotypic wild-type (WT) (30/30). (E) Showing range of phenotypes for the forerunner cluster in E-cadherin mutants. A homozygous tx230 (ecad; 0) mutant, its wild-type sibling, and a homozygous dtv43 (ecad; EC5) mutant are shown. For the tx230 (ecad; 0) allele, this phenotype is strictly a recessive trait, but for the dtv43 (ecad; EC5) allele, this phenotype is a maternal trait (Table 1); homozygous wild-type and heterozygous ZMD mutant siblings (not shown) also frequently have more than one cluster. For quantification of the range of phenotypes in the different alleles and genotypes see graphs in (I) below. Bar, 40 μm. (F and G) The morphology of Kupffer’s vesicle in live embryos, Kupffer’s vesicle is the zebrafish organ of left–right asymmetry, and the derivative of dorsal forerunner cells. (F) Showing smaller vesicle in the homozygous tb08 (ncad; EC1) mutant (9/9) compared to phenotypic wild-type (13/13), and (G) showing multiple vesicles in a single surviving homozygous dtv43 (ecad; EC5) mutant and its ZMD mutant sibling (9/34). Bar, 50 μm. (H) Quantification of left–right asymmetry using looping of the heart. A minimum of five clutches was examined for each category; number of embryos is shown below. (I) Quantification of the number of forerunner clusters in the different alleles of E-cadherin (left), and number of forerunner rosettes for the different genotypes of dtv43 (ecad; EC5) and tx230 (ecad; 0) (right). A minimum of three clutches was examined for each category; number of embryos is shown below. (J) Quantification of left–right asymmetry using looping of the gut or left-sided expression in the diencephalon. A minimum of two clutches was examined for each category; number of embryos is shown below.

Figure 3

The N-cadherin EC5 missense allele has a dominant phenotype. (A–C) Phenotype of embryos from a p79 (ncad; EC5) mutant clutch, shown in: side view (A–C), optical cross section (A’–C’), and outline (A”–C”) traced from optical cross section; br, brain; n, notochord; nk, neural keel; and s, somites. Note in optical cross section how the brain goes from oval in the wild-type (55/55) to triangular in the heterozygote (113/113) to mushroom-shaped in the homozygous mutant (55/55). In addition, the neural tube and somites are wider posteriorly in both the heterozygous and homozygous mutant. (D) Genotype of embryos sorted by phenotype using microsatellite marker z6425 closely linked to N-cadherin (Warga and Kane 2007). Arrow indicates product segregating with the wild-type allele; asterisk denotes an individual sorted as phenotypically wild-type, but possessing both products indicating either a recombination event or sorting error.

Figure 4

N-cadherin missense mutants are more severe than N-cadherin nonsense mutants. Only homozygous mutants and homozygous wild-type (WT) are shown. (A–D) High-magnification view of the eye showing N-cadherin protein (red) and DAPI nuclear staining (blue) in: tm101 (ncad; 0) mutants (12/12), p79 (ncad; EC5) mutants (11/11), and tb08 (ncad; EC1) mutants (12/12). Bar, 30 μm. (E–H) Embryo morphology of the different N-cadherin alleles. (I–N) High-magnification view of: (I–K) the enveloping layer showing E-cadherin protein (red) and DAPI nuclear staining (blue); and (L–N) the presumptive mesoderm showing N-cadherin protein (red) and DAPI nuclear staining (blue). Bar, 20 μm. (O–Q) Expression of dlA in neuroblasts in: (O, O’) tm101 (ncad; 0) mutants (16/16), (P, P’) p79 (ncad; EC5) mutants (16/16), and (Q, Q’) tb08 (ncad; EC1) mutants (14/14); arrowheads denote width of the neural keel. (R–T) Expression of dlC in the somitic mesoderm of: (R, R’) tm101 (ncad; 0) mutants (12/12), (S, S’) p79 (ncad; EC5) mutants (41/41), and (T, T’) tb08 (ncad; EC1) mutants (15/15). (U–W) Expression of fgf8a in the somitic mesoderm of: (U, U’) tm101 (ncad; 0) mutants (22/22), (V, V’) p79 (ncad; EC5) mutants (26/26), and (W, W’) tb08 (ncad; EC1) mutants (23/23). Both dlC and fgf8a are also expressed in the nervous system. Embryos are shown in side view except (O–Q’) shown in dorsal view.

Figure 5

β-catenin mislocalizes to the cytoplasm in E- and N-cadherin missense mutants. (A–G) High-magnification view of the blastoderm, showing localization of β-catenin at 40% epiboly in: (A) WT, (B) tx230 (ecad; 0), (C) tm94 (ecad; EC1), (D–D”) dtv43 (ecad; EC5), (E) tm101 (ncad; 0), (F) p79 (ncad; EC5), and (G) tb08 (ncad; EC1) alleles. In wild-type or homozygous nonsense mutants (A, B, D, and E) β-catenin is membrane-bound, but in heterozygous or homozygous missense mutants (C, D’, D”, F, and G), β-catenin is also cytoplasmic. Note that for the homozygous missense mutants in (C, D”, and G), cells are also more disorganized and staining appears more intense where cells aggregate together, but absent where there is extracellular space. (D–D”) are siblings. Bar, 20 μm. For quantification of phenotypes please refer to Table 2. WT, wild-type; ZMD, zygotic–maternal dominant.

Figure 6

Missense alleles of E- and N-cadherin genetically interact. (A–C) Morphology of the heterozygous dtv43 (ecad; EC5) ZMD mutant, and the homozygous p79 (ncad; EC5) and tb08 (ncad; EC1) mutant; black arrows indicate cells detaching from the neural keel, and white arrows indicate disaggregated cells in the brain and tail. (D–D”) Embryos derived from a female heterozygous for the p79 (ncad; EC5) allele and a male heterozygous for the dtv43 (ecad; EC5) allele. (E–E”) Embryos derived from a female heterozygous for the tb08 (ncad; EC1) allele and a male heterozygous for the dtv43 (ecad; EC5) allele. (F–F”) Embryos derived from a female heterozygous for the tm94 (ecad; EC1) allele and a male heterozygous for the tb08 (ncad; EC1) allele. ZMD, zygotic–maternal dominant.

Figure 7

The EC5 missense allele of E-cadherin exacerbates the homozygous phenotype of the EC1 missense allele of N-cadherin. (A–F) embryos obtained from a female heterozygous for dtv43 (ecad; EC5) and tb08 (ncad; EC1), and a male heterozygous for tb08 (ncad; EC1), in these crosses the expected ratio for +/+, ZMDdtv43/+, tb08/tb08, and ZMDdtv43/+; tb08/tb08 is 1/8. For comparison (B’–F’) embryos were obtained from a male and female heterozygous for tb08 (ncad; EC1), in these crosses the expected ratio for +/+ and tb08/tb08 is 1/4. The proportion of embryos exhibiting each phenotype is indicated for each genotype. Expression of: (A) notch1b in the neural keel and notochord; (B, B’) notch5 in the neural keel, notochord, and somitic mesoderm; (C, C’) deltaD in the neural keel and somitic mesoderm; (D, D’) myoD in the somitic mesoderm; (E, E’) no tail in the notochord; and (F, F’) foxa2 in the endoderm and notochord. Embryos are shown in dorsal view; designations: open circles, notochord and arrowheads, extent of neural keel. ZMD, zygotic–maternal dominant.

Figure 8

The EC5 missense allele of E-cadherin does not affect the homozygous phenotype of the N-cadherin nonsense allele. (A–D) embryos obtained from a female heterozygous for dtv43 (ecad; EC5) and tm101 (ncad; 0), and a male heterozygous for tm101 (ncad; 0); in these crosses the expected ratio for +/+, ZMDdtv43/+, tm101/tm101, and ZMDdtv43/+; tm101/tm101 is 1/8. For comparison (A’–D’) embryos were obtained from a male and female heterozygous for tm101 (ncad; 0); in these crosses the expected ratio for +/+ and tm101/tm101 is 1/4. The proportion of embryos exhibiting each phenotype is indicated for each genotype. Expression of: (A, A’) notch5 in the neural keel, notochord, and somitic mesoderm; (B, B’) deltaD in the neural keel and somitic mesoderm; (C, C’) myoD in the somitic mesoderm, and (D, D’) no tail in the notochord. Embryos are shown in dorsal view; designations: open circles, notochord and arrowheads, extent of neural keel. ZMD, zygotic–maternal dominant.

Figure 9

Double homozygous E- and N-cadherin missense mutants radically alter forerunner cell patterning. (A–D) Homozygous dtv43 (ecad; EC5) mutants from a double-mutant clutch before the tb08 (ncad; EC1) mutant phenotype can be detected. They are visualized with no tail to show dorsal forerunner cell (arrows) phenotypes. For quantification of range of phenotypes see graph (M) below. Note that no tail is also expressed in the notochord, which is also perturbed in homozygous dtv43 (ecad; EC5) mutants. (E–H) Homozygous dtv43 (ecad; EC5) single mutants and their (I–L) homozygous dtv43 (ecad; EC5); tb08 (ncad; EC1) double-mutant siblings visualized with no tail and anti β-catenin staining; (E and I) are low-magnification view and (E’–H and I’–L) are high-magnification view of Kupffer’s vesicle. Bar in (E’–H and I’–L), 60 μm. (E–G and I–K) Expression of no tail in Kupffer’s vesicle in: dtv43 (ecad; EC5) single mutants (39/39) and dtv43 (ecad; EC5); tb08 (ncad; EC1) double mutants (17/17). For quantification of range of phenotypes see graph (O) below. Black arrows indicate vesicles and white arrows indicate cells disaggregating from the kinked notochord, a distinguishing feature of the tb08 (ncad; EC1) allele. (H and L) Expression of β-catenin in Kupffer’s vesicle in: dtv43 (ecad; EC5) single mutants (29/29) and dtv43 (ecad; EC5); tb08 (ncad; EC1) double mutants (9/9). Dots indicate enveloping layer nuclei showing that β-catenin staining is comparable between siblings. (M) Quantification of range of forerunner cell phenotype at 75% epiboly shown in (A–D); and (N) at the one-somite stage, 3 hr later after the tb08 (ncad; EC1) notochord phenotype can be seen. (O) Quantification of range of vesicle phenotypes at the five-somite stage shown in (E–G and I–K). For (M) a total of three crosses was examined, for (N and O) a total of four crosses was examined for each category; number of embryos is shown below. dfc, dorsal forerunner cell; KV, Kupffer’s vesicle.