poky/chuk/ikk1 is required for differentiation of the zebrafish embryonic epidermis

Abstract

An epidermis surrounds all vertebrates, forming a water barrier between the external environment and the internal space of the organism. In the zebrafish, the embryonic epidermis consists of an outer enveloping layer (EVL) and an inner basal layer that have distinct embryonic origins. Differentiation of the EVL requires the maternal effect gene poky/ikk1 in EVL cells prior to establishment of the basal layer. This requirement is transient and maternal Ikk1 is sufficient to allow establishment of the EVL and formation of normal skin in adults. Similar to the requirement for Ikk1 in mouse epidermis, EVL cells in poky mutants fail to exit the cell cycle or express specific markers of differentiation. In spite of the similarity in phenotype, the molecular requirement for Ikk1 is different between mouse and zebrafish. Unlike the mouse, EVL differentiation requires functioning Poky/Ikk1 kinase activity but does not require the HLH domain. Previous work suggested that the EVL was a transient embryonic structure, and that maturation of the epidermis required replacement of the EVL with cells from the basal layer. We show here that the EVL is not lost during embryogenesis but persists to larval stages. Our results show that while the requirement for poky/ikk1 is conserved, the differences in molecular activity indicate that diversification of an epithelial differentiation program has allowed at least two developmental modes of establishing a multilayered epidermis in vertebrates.

Figure 1. Poky mutants are delayed in epiboly

(A–H) Stills from time lapse analysis of epiboly in wild type (A–D) and poky mutant embryos (E–H) timed from sphere stage (0 min). (I–P) Whole mount in situ expression in 50% epiboly wild type (I–L) and age matched poky mutant (M–P) embryos, (I,M) ntl, (J,N) chd, (K,O) eve1, gsc, (L, P) fbp1b. (A–H, L and P) lateral views. (I–K, M–O) animal pole views.

Figure 2. poky encodes chuk/ikk1/ikkα

(A) schematic of domains of Ikk1. Amino acid sequence around the T-loop activation domain in wild type; poky mutant and human sequence is aligned below. The site of the missense mutation in poky^p20ad is indicated by the asterisk (*). T-loop activating serines are indicated in red. Missense mutation in poky^p20ad substituting the conserved phenylalanine for serine is indicated in green. Nonconserved amino acids in the human sequence are indicated in blue. (B) percent rescue to 24 hpf by ikk1 mRNA. (C,D,E,H,I,J) Whole mount in situ hybridization for cyt1 (C,H), krt8 (D,I) and krt4 (E,J) at 50% epiboly. (F,G,K,L, M–V) whole mount in situ hybridization for krt4 at 50% epiboly following microinjection of the indicated mRNA. (W–Y) Cell autonomy of Ikk1. (W) Anti myc staining of myc-Ikk1 scatter labeled embryo. (X) krt4 expression of the same embryo. (Y) Merged image of W and X. Rescued EVL cells (arrowhead) and isolated myc positive DEL cells (arrow) can be observed.

Figure 3. poky mutant embryos display junction protein localization and apical basal polarity

Wild type embryos displayed tight localization of actin microfilaments at cell-cell boundaries (arrows) at sphere (A, magnified in C) and robust labeling of cell vertices at 50% epiboly (E magnified in G, arrowheads). poky mutant embryos displayed actin localization to cell cell boundaries (arrows) at sphere stage (B magnified in D). At 50% epiboly (F magnified in H) the localization to cell-cell boundaries was less robust. Gaps were observed between neighboring cells (arrows) and little labeling was observed at cell vertices (arrowheads). Wild type embryos displayed localization of ZO1 to the tight junction at EVL cell borders at sphere (I magnified in K) and 50% epiboly (M magnified in O) with strong localization to cell-cell boundaries (arrows) and cell vertices (arrowheads). poky mutants also displayed localization at sphere stage (J, magnified in L). By 50% epiboly poky mutant cells were smaller but still displayed localized ZO1 (N magnified in P, arrow), although they lacked cell vertex labeling (arrowhead, N,P). Wild type (Q,S,U) and poky mutant (R,T,V) embryos display apical localization of aPKC (green, Q–T), localization of ZO1 to sites of EVL cell-cell contact (red, Q,R), basolateral localization of β-catenin (red, S,T) and cadherin (U,V). (W–Z) TEM of wild type (W,X) and poky mutant (Y,Z) EVL cells at 30% epiboly (tight junctions, black arrows).

Figure 4. poky mutants fail to form an EVL barrier

Wild type (A), poky mutant (B) and wild type positive control (C) processed for TUNEL. DIC (D–F) and interstitial space labeled with fluorescent dextran (G–I) of wild type (D,G), intact poky mutants (E,H) and clearing poky mutants (F,I) reveals large spaces in the wild type embryo at 50% epiboly (arrowhead, G) but little interstitial space in poky mutants (arrowheads, H,I). White light (J,L) and fluorescent (K,M) of propidium iodide stained live wild type (J,K) and clearing poky mutants (L,M) shows labeling of multiple nuclei in poky mutants (arrowhead, M). Wild type control embryos at 100% epiboly develop normally in 1 × Ringer’s saline (N). poky mutant embryo cultures in Ringer’s saline did not lyse but failed to initiate epiboly (O). poky mutant embryos cultured in E3 displayed blastoderm lysis when controls reached 50% epiboly (P). Wild type embryos (Q–S) displayed little interstitial biotin labeling. poky mutant embryos showed little interstitial label at sphere stage (T). 30% epiboly poky mutant embryos had some interstitial signal (arrowhead, U). At 50% epiboly poky mutants had strong interstitial label throughout the blastoderm (arrowhead, V).

Figure 5. poky mutants fail to down regulate cell division in the EVL

Embryos were stained with phalloidin (green) to reveal actin localization at cell boundaries and DAPI (red) to reveal nuclear morphology. Dividing cells were identified by their nuclear morphology (arrowheads, A&B). Multinucleate EVL cells were observed in wild type and mutant embryos (arrow, B). Total number of EVL cells was calculated by determining the number of cells per unit area as C and D and then calculating the total EVL surface area by calculating the radius and height as in E. (F) Percent mitotic nuclei. (G) Total number of EVL cells.

Figure 6. poky mutant phenotype requires zygotic transcription

(A–H) EVL of phalloidin stained embryos (A–D) low magnification (E–H) high magnification. Untreated (A,E) and alpha-amanitin treated (B,F) wild type embryos display similar organization of actin microfilaments with robust localization to cell borders and cell vertices (arrowheads). Untreated poky mutants (C,G) display smaller EVL cells with less robust localization of actin to the EVL cell borders. Alpha-amanitin treated poky mutants (D,H) display a wild type EVL cell morphology, tight actin localization to the cell boundaries, and robust vertex labeling (arrowhead). (I–L) ZO1 localization at the same magnification as E–H. Untreated (I) and alpha-amanitin treated (J) wild type embryos display similar localization of ZO1 to tight junctions. Untreated poky mutants (K) display more EVL cells with less robust localization of ZO1 to cell borders and vertices. Alpha-amanitin treated poky mutants (L) display a wild type EVL cell morphology and tight ZO1 localization to the cell boundaries and robust localization to the vertices (arrowhead). (M,N) Wild type untreated and treated embryos exclude biotin from the DEL. (O) poky mutants do not exclude biotin and show strong interstitial labeling with fluorescent streptavidin. (P) alpha aminitin treated poky embryos exclude biotin from the DEL.

Figure 7. poky mutant adults have normal skin and wild type EVL cells persist into larval stages

(A,B) Optical section through wild type (A) and poky mutant (B) epidermis expressing nuclear p63 protein (green) at 24 hpf. (C,D) 2 year-old wild type and poky mutants have no obvious skin lesions. (E,F) Pigment, scales and fins are normal in wild type (E) and poky mutants (F). Histology of adult skin reveals the typical bilayered epidermis (bracket) in wild type (G) and poky mutants (H). Lineage tracing of embryonic and larval EVL (I–N). EVL from embryos scatter labeled with fluorescent dextran were observed up to 9 dpf. (I) 2 dpf low magnification image of the tail shows multiple cells labeled; location at higher magnification in J. (J) A clone of EVL cells is easily observable at 2 dpf (yellow arrowhead marks a single EVL cell). The location of these cells was traced relative to underlying muscle cells (red asterisk) at 2 dpf (J), 3 dpf (K), 5 dpf (L), 7 dpf (M) and 9 dpf (N).