BMP, Wnt and FGF signals are integrated through evolutionarily conserved enhancers to achieve robust expression of Pax3 and Zic genes at the zebrafish neural plate border

Abstract

Neural crest cells generate a range of cells and tissues in the vertebrate head and trunk, including peripheral neurons, pigment cells, and cartilage. Neural crest cells arise from the edges of the nascent central nervous system, a domain called the neural plate border (NPB). NPB induction is known to involve the BMP, Wnt and FGF signaling pathways. However, little is known about how these signals are integrated to achieve temporally and spatially specific expression of genes in NPB cells. Furthermore, the timing and relative importance of these signals in NPB formation appears to differ between vertebrate species. Here, we use heat-shock overexpression and chemical inhibitors to determine whether, and when, BMP, Wnt and FGF signaling are needed for expression of the NPB specifiers pax3a and zic3 in zebrafish. We then identify four evolutionarily conserved enhancers from the pax3a and zic3 loci and test their response to BMP, Wnt and FGF perturbations. We find that all three signaling pathways are required during gastrulation for the proper expression of pax3a and zic3 in the zebrafish NPB. We also find that, although the expression patterns driven by the pax3a and zic3 enhancers largely overlap, they respond to different combinations of BMP, Wnt and FGF signals. Finally, we show that the combination of the two pax3a enhancers is less susceptible to signaling perturbations than either enhancer alone. Taken together, our results reveal how BMPs, FGFs and Wnts act cooperatively and redundantly through partially redundant enhancers to achieve robust, specific gene expression in the zebrafish NPB.

Fig. 1.

BMP and Wnt signaling regulate pax3a and zic3, and FGF signaling regulates pax3a during late gastrulation. (A-C,G-I) pax3a and zic3 expression decreases upon heat shock-mediated dkk1 overexpression. pax3a and zic3 expression decreases slightly in hsp70l:dkk1 embryos heat shocked at 75% epiboly [pax3a, B versus A (81%, n=27); zic3, H versus G (93%, n=28)] and a shield stage heat shock leads to a greater decrease in expression [pax3a, C versus A (92%, n=49); zic3, I versus G (100%, n=33)]. (D-F,J-L) pax3a and zic3 expression in the NPB decreases and shifts medially upon heat shock-mediated bmp2b overexpression. A 75% epiboly heat shock leads to a loss of posterior NPB expression (E versus D and K versus J, insets) and a slight medial shift of the anterior NPB domain (D,E,J,K, bars) (pax3a, 100%, n=22; zic3, 94%, n=18). A shield stage heat shock leads to a drastic decrease in pax3a and zic3 expression throughout the anterior-posterior axis, and NPB expression shifts medially into a teardrop shape [pax3a, F versus D (100%, n=28); zic3, L versus J (100%, n=39)]. (M,N) Morpholino-mediated knockdown of wnt8a leads to a greater decrease in pax3a expression than does dkk1 overexpression [N versus M (90%, n=20)]. (O-R) Inhibition of FGF signaling by SU5402 treatment beginning at 7 hpf leads to a decrease in pax3a expression by 12 hpf (84%, n=44) (P versus O). The same SU5402 treatment does not decrease NPB zic3 expression (100%, n=27) (R versus Q). (S,T) Optical cross-sections allow the distinction between NPB and mesodermal pax3a expression. An optical cross-section of the trunk at ~12 hpf reveals pax3a expression in the ectoderm only (black arrowheads, S), whereas an optical cross-section slightly later reveals pax3a mRNA in the ectoderm (black arrowheads, T) and paraxial mesoderm (white arrowheads, T). The insets in D-F,J,K are posterior dorsal views, all other pictures are dorsal trunk views with anterior upwards.

Fig. 2.

Wnt and FGF signaling play partially redundant roles in regulating NPB specifiers. (A-H) Embryos treated with SU5402 starting at 7 hpf exhibited decreased pax3a expression (100%, n=28) (B versus A), but no decrease in zic3 expression (100%, n=16) (F versus E). dkk1 overexpression induced by 7 hpf heat shock decreases expression of pax3a (100%, n=24) (C versus A) and zic3 (100%, n=16) (G versus E). When the Wnt and FGF signaling pathways are both attenuated with heat shock-induced dkk1 overexpression and SU5402 treatment starting at 7 hpf, pax3a and zic3 mRNA levels decrease more than when either pathway is knocked down alone [pax3a, D versus B,C (100%, n=24); zic3, H versus F,G (100%, n=17)]. All pictures are dorsal trunk views with anterior upwards.

Fig. 3.

Two enhancers in pax3a intron 4 drive gene expression in the NPB and dorsal neural tube. (A) The Fugu pax3a locus is shown along with a Vista plot (Frazer et al., 2004; Mayor et al., 2000) of sequence identity between Fugu and zebrafish, Fugu and human, and Fugu pax3a and its paralog pax3b. Enhancer regions are shown in red and pax3a exons are shown in yellow. (B-O) GFP expression driven by IR1 (B-H) and IR2 (I-O) is shown as detected by in situ hybridization for embryos at 12 hpf (B-G,I-N) and by fluorescence at 24 hpf (H,O). Optical cross-sections through the trunk region confirm that IR1 and IR2 drive gene expression in the ectoderm (E,L) and double in situ hybridization for GFP (brown) and pax3a (purple) demonstrate that pax3a expression overlaps with the activity of IR1 and IR2 (F,G,M,N). G and N are higher magnification views of the areas outlined in F,M, respectively. B,H,I,O are lateral views; C,J are dorsal trunk views; D,K are posterior dorsal views.

Fig. 4.

pax3a IR1 requires Wnt and FGF signaling, and pax3a IR2 requires BMP and FGF signaling for full activity. (A-F) dkk1 was overexpressed by heat shock in embryos with pax3a IR1:GFP or pax3a IR2:GFP. pax3a IR1 activity decreased with dkk1 overexpression induced with a 75% epiboly heatshock (100%, n=25) (B versus A) and was almost completely lost with a shield stage heat shock (100%, n=12) (C versus A). pax3a IR2:GFP activity did not decrease when dkk1-mCherry overexpression was induced by heat shock at 75% epiboly (92%, n=12) (E versus D) or shield stage (89%, n=18) (F versus D). (G-L) bmp2b was overexpressed by heat shock. pax3a IR1activity intensifies and expands into the neural plate and epidermis when bmp2b overexpression is induced with a 75% epiboly heat shock (93%, n=41) (H versus G) and heat shock at shield stage leads to an even greater activity increase and causes expansion into the entire neural plate (84%, n=19) (I versus G). pax3a IR2 activity decreases and shifts medially (J-L, bars) upon bmp2b induction by 75% epiboly heat shock (100%, n=33) (K versus J) and further decreases and shifts medially into a tear-drop shape with a shield-stage heat shock (100%, n=14) (L versus J). (M-P) SU5402 treatment (90 μM) beginning at 7 hpf causes a decrease in the activity of pax3a IR1 (94%, n=17) (N versus M) and pax3a IR2 (100%, n=25) (P versus O). All pictures are dorsal trunk views with anterior upwards.

Fig. 5.

The combination of pax3a IR1 and IR2 is less susceptible to signaling perturbations than either enhancer alone. (A-G) GFP expression driven by a pax3a IR2+IR1 composite enhancer is shown under various treatments. dkk1-mCherry overexpression beginning with a shield stage heat shock mildly decreases enhancer activity (60%, n=42) (B versus A). bmp2b overexpression beginning with a shield stage heat shock causes increased enhancer activity, a shift medially in the trunk region and expansion in the posterior (87%, n=31) (C versus A). SU5402 treatment starting at 7 hpf drastically decreases the activity of the composite enhancer (100%, n=14) (E versus D). wnt8a-MO injection causes a more severe decrease in activity than dkk1 overexpression (80%, n=10) (G versus F). (H) A model for pax3a IR1 and IR2 activity. IR1 is activated by Wnt and FGF signaling in a wide band surrounding the NPB (green). IR2 activity is precisely positioned at the NPB by BMP signaling and a repressive neural plate factor and is amplified by FGF signaling (yellow). All pictures are dorsal trunk views with anterior upwards.

Fig. 6.

Two zic3 enhancers drive gene expression in the NPB and dorsal neural tube. (A) The Fugu zic3/zic6 locus is shown along with a Vista plot of the level of sequence identity between Fugu, zebrafish and human zic3 loci. Enhancers are shown in red and exons are shown in yellow. (B-M) GFP expression driven by E1 (B-G) and E2 (H-M) is shown as detected by in situ hybridization for embryos at 12 hpf (B-F,H-L) and by GFP fluorescence at 24 hpf (G,M). Double in situ hybridization for GFP (brown) and zic3 (purple) demonstrate that zic3 expression overlaps with the activity of E1 and E2 (E,F,K,L). E,K are higher magnification views of the areas outlined in F,L, respectively. B,H,G,M are lateral views; D,J are posterior dorsal views; C is a trunk dorsal view; I is an anterior dorsal view.

Fig. 7.

zic3 E1 is regulated by Wnt, BMP and FGF signaling, and zic3 E2 is regulated by Wnt and BMP signaling. (A-F) dkk1 was overexpressed by heat shock in embryos with zic3 E1:GFP or zic3 E2:GFP. dkk1 overexpression beginning with a 75% epiboly heat shock decreases zic3 E1 activity (100%, n=23) (B versus A) and slightly decreases zic3 E2 activity (88%, n=50) (E versus D). The NPB activity of both enhancers is almost completely lost, with dkk1 overexpression induced with a shield stage heat shock [E1, C versus A (100%, n=21); E2, F versus D (100%, n=55)]. (G-L) bmp2b was overexpressed by heat shock. zic3 E1 loses activity in the posterior NPB and shifts medially in the anterior (H, bar) upon bmp2b induction with 75% epiboly heat shock (93%, n=29) (H versus G). Shield-stage heat shock causes a dramatic reduction of E1 activity throughout and causes a drastic medial shift (I, bar) (96%, n=23) (I versus G). zic3 E2 activity does not change when bmp2b overexpression is induced with a 75% epiboly heat shock (79% with wild-type GFP level, n=38) (K versus J), but decreases with a shield-stage heat shock (100%, n=24) (L versus J). (M-P) SU5402 treatment beginning at 65% epiboly dramatically decreases zic3 E1 activity (100%, n=27) (N versus M), but not zic3 E2 activity (81% with wild-type GFP level, n=22) (P versus O). All E1 pictures are dorsal views of the trunk with anterior upwards; all E2 pictures are anterior dorsal views with anterior upwards.

Fig. 8.

pax3a IR1 and zic3 E1 and E2 are probably direct targets of canonical Wnt signaling. (A-E) Mutating six putative Tcf/Lef-binding sites from pax3a IR1 (rectangles in A) decreases enhancer activity at 12 hpf (92%, n=36) (D versus B) and 24 hpf (100%, n=32) (E versus C). (F-J) Mutating three putative Tcf/Lef-binding sites (rectangles, F) does not affect zic3 E1 activity at 12 hpf (89%, 16 out of 18 with wild-type GFP levels) (I versus G) and 24 hpf (61%, 11 of 18 with wild-type GFP levels) (J versus H). (K-O) Mutating five putative Tcf/Lef-binding sites (rectangles, K) reduces zic3 E2 activity in the NPB at 12 hpf (87%, n=20) (N versus L) and the dorsal neural tube at 24 hpf (100%, n=40) (O versus M). (P-AA) Embryos containing pax3a IR1:GFP, zic3 E1:GFP or zic3 E2:GFP were injected with GR-Lef1-βcat mRNA. The activity of all three enhancers significantly increases with dexamethasone and cycloheximide treatment relative to ethanol and cycloheximide treatment [pax3a IR1, Q,S versus P,R (69%, P=7x10⁻⁷, n=16); zic3 E1, U,W versus T,V (57%, P=0.007, n=14); zic3 E2, Y,AA versus X,Z (73%, P=5×10⁻⁹, n=37)]. (BB) The distribution of GFP staining levels in embryos injected with GR-Lef1-βcat mRNA and treated with cycloheximide and dexamethasone or ethanol is shown for each enhancer. B-E,G-J,M,O,P,Q,T,U,X,Y are lateral views with dorsal towards the right; L,N,Z,AA are anterior-dorsal views with anterior upwards; R,S,V,W are dorsal trunk views with anterior upwards.

Fig. 9.

pax3a and zic3 are regulated by integrating inputs from multiple enhancers and signaling pathways. (A) pax3a IR1 requires Wnt and FGF signaling for its full activity and is activated by BMP signaling. The effect of Wnt signaling is probably direct. FGF and BMP signaling could be acting directly on IR1 or indirectly through Wnts (broken lines) as Wnt ligands are upregulated upon bmp2b overexpression (supplementary material Fig. S3) and Wnts are upregulated upon FGF overexpression in Xenopus (Hong et al., 2008). pax3a IR2 is repressed by BMP overexpression, requires FGF signaling for full activity and is probably repressed by a neural plate factor (N.P. repressor). (B) NPB pax3a expression has a strong requirement for FGF signaling and a weaker requirement for Wnt signaling. pax3a expression is repressed by BMP overexpression and probably by a repressor in the neural plate. (C) zic3 E1 requires FGF and Wnt signaling for full activity and is repressed by BMPs. FGF could be acting indirectly by inducing expression of Wnt ligands. zic3 E2 requires Wnt signaling for NPB activity and is repressed by BMP overexpression. (D) zic3 NPB expression has a strong requirement for Wnt signaling and a cryptic requirement for FGF signals that only becomes apparent when Wnt signaling is attenuated. zic3 in the NPB is repressed by BMP overexpression and probably also repressed by a factor in the neural plate.