XB-ART-49947Development December 1, 2014; 141 (24): 4794-805.
Fezf2 promotes neuronal differentiation through localised activation of Wnt/β-catenin signalling during forebrain development.
Brain regionalisation, neuronal subtype diversification and circuit connectivity are crucial events in the establishment of higher cognitive functions. Here we report the requirement for the transcriptional repressor Fezf2 for proper differentiation of neural progenitor cells during the development of the Xenopus forebrain. Depletion of Fezf2 induces apoptosis in postmitotic neural progenitors, with concomitant reduction in forebrain size and neuronal differentiation. Mechanistically, we found that Fezf2 stimulates neuronal differentiation by promoting Wnt/β-catenin signalling in the developing forebrain. In addition, we show that Fezf2 promotes activation of Wnt/β-catenin signalling by repressing the expression of two negative regulators of Wnt signalling, namely lhx2 and lhx9. Our findings suggest that Fezf2 plays an essential role in controlling when and where neuronal differentiation occurs within the developing forebrain and that it does so by promoting local Wnt/β-catenin signalling via a double-repressor model.
PubMed ID: 25468942
PMC ID: PMC4299278
Article link: Development
Genes referenced: arx barhl2 bmp4 chrd.1 ctnnb1 fezf2 foxh1.2 gnl3 gsc irx3 lhx2 lhx9 myt1 neurog1 nodal3.1 nodal3.2 otx2 pax6 rax rpl8 sia1 smad1 sox3 tle1 tle2 tle4 tle5 tuba4b ventx1.1 ventx1.2 wnt1 wnt3a wnt8a wnt8b
Antibodies: FLAG Ab1 Myt1 Ab1 Sox3 Ab2 Tuba4b Ab4
Morpholinos: fezf2 MO2 lhx2 MO1 lhx9 MO1
Article Images: [+] show captions
|Fig. 1. fezf2 knockdown leads to defects in forebrain neuronal differentiation. (A) Whole-mount in situ hybridisation for arx in control MO (20/20) or fezf2 MO (15/18) injected Xenopus embryos. Arrowhead indicates the forebrain. (B-J) One blastomere at the 2-cell stage was injected with fezf2 MO and embryos were sectioned at stage 30 transversely across the forebrain, and stained for Sox3 (B,C), MyT1 (E,F) or TUNEL (H,I). FITC staining identifies the injected side (B,E,H). Arrowheads indicate MyT1+ (E,F) or TUNEL+ (H,I) cells. (D,G,J) Statistical analysis of Sox3+ (n=4 embryos), MyT1+ (n=6 embryos) and TUNEL+ (n=4 embryos) cells. All control sides have been normalised to 100%. Error bars represent s.e.m. *P<0.05; ***P<0.001; ns, not significant. Scale bar: 25 µm.|
|Fig. 2. fezf2 promotes Wnt/β-catenin signalling and induces neuronal differentiation through Wnt/β-catenin in vitro and in vivo. (A) fezf2 misexpression in early Xenopus embryos leads to strong dorsoanteriorisation (31/35 embryos examined showed the illustrated phenotype) compared with lacZ (β-gal) controls (39/39). (B) fezf2 misexpression enhances Smad2/3 phosphorylation and inhibits Smad1/5/8 phosphorylation as assessed in western blots. Blastula stage (st. 8) indicates the pre-activation state. Elongation factor 4E (elF4E) was used as a loading control. (C,D) qPCR shows that fezf2 promotes the expression of xnr3 (C) and sia (D) in early embryos (n=3 replicates). (E) TOPFlash assay shows that fezf2 promotes Wnt/β-catenin signalling (n=4 replicates). (F-H) fezf2 expression colocalises with active Wnt signalling in the forebrain. (F) The transgenic construct. (G) Dorsal and lateral views of stage 30 embryos; GFP signal for Wnt activity (green); Katushka signal for fezf2 expression (red); +bf, merged image with bright-field. (H) Knockdown of fezf2 reduces Wnt activity in the forebrain as assessed by expression of the 7LEF-dEGFP F1.1 Wnt reporter line. Arrowhead indicates the diencephalon. (I) The Wnt inhibitor δNTcf3 antagonises Fezf2-induced neuronal differentiation in mouse neuronal progenitors, as assessed by the induction of axonogenesis. (J) Statistics of I (n=4 replicates). (K,L) Electroporation experiments show that the Wnt inhibitor δNTcf3 antagonises Fezf2-induced neuronal differentiation in the tadpole forebrain. (K) Transverse sections of the forebrain area of stage 30 embryos electroporated correspondingly and stained for MyT1 (red), GFP (green) and with DAPI (blue). Left images, merge; right images, MyT1 alone. (L) Statistics of K (n=5 embryos). Control side is normalised to 100%. (M) qPCR analysis shows that the Wnt inhibitor δNTcf3 antagonises Fezf2-induced ngn1 expression in stage 20 animal cap explants (n=3 replicates). In all qPCR analyses, ribosomal protein L8 (rpl8) was used as an internal control. *P<0.05, **P<0.01, ***P<0.001. Error bars represent s.e.m. Scale bars: 100 µm in I; 50 µm in K.|
|Fig. 3. The Fezf2-regulated endogenous level of Wnt/β-catenin signalling governs forebrain neurogenesis. (A) Different Fezf2 constructs. Different N-terminal domains (Eh1-repressor, VP16 activator or Eve repressor) are shown in different colours. The zinc-finger DNA-binding domain is shown in blue. (B) Western blot of gastrula stage Xenopus embryos injected with nuclear lacZ (control), eve-fezf2, VP16-fezf2 and wt-fezf2 and assayed for phosphorylated Smad1 or Smad2 and α-Tubulin (loading control). (C) pTransgenesis system transgenic constructs to assess the impact of Fezf2 and/or Wnt activities on forebrain development. (D) Expression of NβT-GFP (a-e, stage 40 embryo) and arx (a′-e′, stage 30 embryo) in the forebrain of transgenic embryos harbouring the transgenes shown in C. Inset (a) shows the fluorescence from Katushka (red). (E) Quantification of neural tissue growth phenotypes from D.|
|Fig. 4. Fezf2 functions through interaction with members of Groucho family. (A) Tle1, Tle2, Tle4 and Aes constructs. Note that Aes lacks the protein-interaction WD domain. (B) Wild-type Fezf2 and ΔEh1-Fezf2 with a mutated Eh1 domain. (C) Immunoprecipitation of extracts from Xenopus embryos injected with different combinations of the indicated mRNAs, showing that Fezf2 interacts with Tle1, Tle2 and Tle4 (lanes 6, 8 and 12) but not Aes (lanes 14, 15). The Eh1 domain is required for the proper interaction between Fezf2 and Tle1/2/4 (lanes 6 and 7, 8 and 9, 12 and 13).|
|Fig. 5. Fezf2 represses the activity of lhx2 and lhx9 in the forebrain. (A) ChIP-qPCR analysis of Fezf2 binding to the lhx2 promoter. Region 1 showed very high enrichment (n=3 replicates). (B,C) qPCR analysis of lhx2 and lhx9 expression in p3hGR-VP16-Fezf2-injected animal cap explants aged to stage 12 and treated with CHX alone or CHX+DEX (n=3 replicates). (D,E) qPCR analysis of lhx2 and lhx9 expression in neuralised animal cap explants aged to stage 20 (n=3 replicates). (F,G) In situ hybridisation analysis shows that mild knockdown of fezf2 leads to expansion of the lhx2 (F) and lhx9 (G) expression area. Arrowhead indicates the epithalamus; bracket indicates the ventral diencephalon. (H,I) In situ hybridisation analysis of arx (H) or ngn1 (I) in stage 28 morphants (lateral views). Arrowheads indicate arx or ngn1 expression. (J) qPCR analysis shows that lhx2 and lhx9 antagonise expression of the fezf2-induced Wnt-responsive gene xnr3 in stage 14 animal cap explants (n=3 replicates). (K) qPCR analysis shows lhx2 and lhx9 antagonise fezf2-induced ngn1 expression in stage 20 animal cap explants (n=3 replicates). In all qPCR analyses, rpl8 was used as internal control. Error bars represent s.e.m. *P<0.05, ***P<0.001; ns, not significant.|
|Fig. 6. Mechanistic model of Fezf2 function in the forebrain. (A) In the presence of Fezf2. Fezf2 interacts with Groucho-family repressors and inhibits the expression of lhx2/lhx9. Consequently, β-catenin binds to the Tcf complex and Wnt signalling is activated, promoting the expression of neurogenin 1 and thus stimulating neuronal differentiation. (B) In the absence of Fezf2. Lhx2/Lhx9 inhibits Wnt signalling, resulting in the degradation of β-catenin. In the absence of β-catenin, the Tcf complex is maintained in a repressive state. This repressive Tcf complex inhibits neurogenin 1 expression, thus inhibiting neurogenesis. Progenitor cells that have exited the proliferation state cannot differentiate and thus enter apoptosis.|
|Figure S1. Temporal expression pattern of X. tropicalis fezf2 revealed by qPCR analysis at different stages, including fertilised egg (FE), blastula (stage 8), gastrula (stage 10.5), neurula (stage 15-20), and tailbud (stage 22-24) (n=3 replicates). Error bars represent ±s. e. m.|
|Figure S2. Comparison of expression pattern of fezf2 expression relative to various nerual markers at different stages, related to Figure 1. Anterior is to the bottom in stage 15 and 20 embryos (frontal view), and to the left in stage 28 and 35 embryos (lateral view).|
|Figure S3. (A) Schematic of fezf2 MO design and knockdown efficiency. The MO targets the exon3-intron3 junction (above) and semi-quantitative PCR analysis shows mature fezf2 mRNA levels in stage 24 embryos (below), injected with control MO (1) or fezf2 MO (2). Black arrow pair indicates the position of validation primers for RTPCR analysis, the forward primer overlaps the exon3-exon4 junction. (B) qPCR results, n=3 replicates. In both cases, the ribosomal protein L8 (rpl8) gene was used as an internal control.|
|Figure S4 (related to Figure 1). fezf2 knockdown, but not control MO knockdown, leads to defects in forebrain neuronal differentiation. (A) fezf2 knockdown led to defects in forebrain development after initial patterning stage. Whole-mount in situ hybridisation analyses on control versus fezf2 MO injected stage 28 embryos, stained using the forebrain-specific markers, otx2 and pax6. (B) fezf2 knockdown did not cause defects in early forebrain patterning, as revealed by whole-mount in situ hybridisation on control versus fezf2 MO injected stage 17 embryos, stained using the forebrain-specific markers arx, otx2, and pax6. Anterior is to the bottom. (C-H) 1 of 2 blastomeres at 2-cell stage was injected with control MO, cultured to stage 30, fixed sectioned transversely across the forebrain and stained via immunofluorescence for the transcription factor Sox3 (marker for neural progenitors) (C-D) or MyT1 (marker for differentiated primary neurons) (F-G). (C) Merged image. (D) Sox3 staining alone. (E). Statistical analysis of Sox3+ cells in control MO injected side relative to the uninjected side of the forebrain (n=3 embryos, no significant difference). The uninjected side has been normalised to 100%. (F) Merged image. (G) MyT1 staining alone. (H) Statistical analysis of MyT1+ cells in control MO injected side relative to the uninjected side of the forebrain (n=3 embryos, no significant difference). The uninjected side has been normalised to 100%. (I-K) Transverse sections were prepared as for panels C-D and then processed for TUNEL staining. (I) Merged image. (J) TUNEL-positive cells alone. (K) Percentage of the TUNEL-positive cells in the control MO injected side compared to the uninjected side (n=3 embryos). In all cases, the FITC tag on the control MO was used to identify the injected side; DAPI was used to stain nuclei. Error bars represent ±s. e. m. Scale bar: 25 μM. ns: not significant. (L) (a-b) Immunofluorescence staining using the neuronal differentiation marker, Ntubulin, on stage 32 embryos injected with 10 ng of control or fezf2 MO. (a) control MO; (b) fezf2 MO, showing a strong reduction of N-tubulin staining in the rostral area. (c) Sections of stage 32 embryos at the prethalamus level and stained with N-tubulin antibody. Embryos were injected with 5 ng of fezf2 MO into one of the two blastomeres at the 2-cell stage. A strong reduction of N-tubulin staining was observed in the injected side. Arrowhead: injected side. DAPI was used to provide global visualisation on the head structure. In all cases, error bars represent ±s. e. m.|
|Figure S5 (related to Figure 2). fezf2 promotes Wnt/β-catenin signalling. (A) wnt8 mis-expression inhibited BMP signalling while activating TGF-β/Nodal signalling. 50 pg of either nuclear β-gal mRNA or wnt8a mRNA was injected into X. laevis embryos at 1-2 cell stage and collected at the gastrula stage (10.5) for Western blot analysis. Phospho-Smad 2/3 antibody (TGF-β/Nodal signalling ) or phospho-Smad 1/5/8 (BMP signalling) were used assay for the activation states of TGF-β/Nodal or BMP signalling. A significant increase on the phosphorylation level of Smad 2/3 was observed together with significant decrease on the phosphorylation level of Smad 1/5/8 was seen in wnt8 mRNA injected embryos. (B-G) fezf2 mis-expression promotes Wnt-responsive dorsal markers goosecoid and chordin expression. In all cases ribosomal protein L8 (rpl8) gene was used as internal control for qPCR analysis. (B-D) goosecoid expression. (B) Uninjected control embryo. (C) fezf2 mRNA overexpressing embryo. (D) qPCR analysis, showing an increase of goosecoid expression (n=3 replicates, P<0.05). (E-G) chordin expression. (E) Uninjected control embryo. (F) fezf2 mRNA overexpressing embryo. (G) qPCR analysis, showing an increase of chordin expression (n=3 replicates, P<0.001). (H-M) fezf2 mis-expression inhibits ventral markers vent1 and bmp4 expression. Embryo treatments are same as in (B-G). (H-J) vent1 expression. (H) Uninjected control embryo. (I) fezf2 mRNA overexpressing embryo. (J) qPCR analysis, showing a decrease of vent1 expression (n=3 replicates, P<0.001). (K-M) bmp4 expression. (K) Uninjected control embryo. (L) fezf2 mRNA overexpressing embryo. (M) qPCR analysis, showing a decrease in vent1 expression (n=3 replicates, P<0.01). (N) fezf2 mis-expression leads to nuclear accumulation of β-catenin. 250 pg of either nuclear β-gal (control) or fezf2 mRNA was injected at 1-2 cell stage and section made at stage 10, DAPI mask used to reveal nuclear content and the localisation of nuclear β-catenin after masking, showing increased nuclear enrichment of β-catenin in fezf2 mis-expressed embryos. Both grayscale and pseudo-colour images have been used to show the level of β-catenin presence in the nucleus. (a and a’) control; (b and b’) fezf2 injected embryos. Arrowhead: blastopore lip. (O-P) Axis duplication assay. (O) Schematic representation of ventral blastomere injection. 250 pg of mRNA, together with nuclear β-gal mRNA as tracer, was injected into one of the ventral blastomere (arrowhead) at the 4-cell stage. Embryos were fixed at approximately stage 38 and stained with Red-gal. D: dorsal side. (P) Results of axis duplication assay. (a) Control embryo, nuclear β-gal mRNA injection only. (b) 250 pg fezf2 mRNA. (c) 250 pg β- catenin mRNA. (Q) Dual-luciferase assay on control FOPFlash plasmid containing mutated TCF-binding sites. pTK-renilla was used as endogenous control. (R) qPCR analysis on wnt1, wnt3a, and wnt8b in nuclear β-gal (control) or fezf2-injected neuralised animal cap explants aged to stage 15. n=3 replicates. * P<0.05. In all cases, error bars represent ±s. e. m. (K) Schematic representation of animal cap explant assay. mRNA at different combinations were injected at the 1-2 cell stage and allowed to develop until the mid-blastula stage (stage 8). Animal cap explants were excised at stage 8 and allowed to develop to stage 20, at which point they were collected for qPCR analysis.|
|Figure S6 (related to Figure 3). fezf2-regulated endogenous level of Wnt/β-catenin signalling governs forebrain neurogenesis in transgenic embryos. (A-C) Phenotypic effect of antimorphic Fezf2. (A) nuclear β-gal (control) mRNA, (B) fezf2 mRNA, showing strongly dorsal-anteriorized (DAI 8-9) embryos. (C) VP16-fezf2 mRNA, showing strongly ventralised (DAI 1-2) embryos. (D) GFP expression in the forebrain of embryos from different transgenic lines at day 7, top view. (a) NβTGFP ·fezf2·Katushka. Green: GFP; red: Katushka (inset). (b) NβT-GFP·fezf2·VP16- Fezf2. (c) NβT-GFP·fezf2· N90βcatenin. (d) NβT-GFP·fezf2· N51Tcf3. (e) NβTGFP ·fezf2·GSK3βS9A.|
|Figure S7 (related to Figure 5). Demonstration of ChIP-qPCR efficiency and Fezf2 binding site. (A) ChIP-qPCR result of FLAG-tagged FoxH1 using IgG (control) or anti-FLAG antibody, showing enrichment in the proximity of branchyury (xbra) promoter. (B) Schematics of the three highly conserved regions screened by ChIPqPCR, highlighted areas have been screened using primers designed against each of the different regions.|
|Figure S8 (related to Figure 6). Fezf2 represses the activities of lhx2 and lhx9. (A-B) Control experiments showing the expression of lhx2 and lhx9 in stage 12 animal cap explants injected with p3hGR-VP16-Fezf2, either left untreated, or treated with either CHX or DEX alone. qPCR results of (A) lhx2 and (B) lhx9 are shown (n=3 replicates, P<0.001). rpl8 was used as an internal control. (C-D) Stage 20 animal cap explants neuralised by chordin mRNA expressed high level of lhx2 and lhx9. qPCR results of (C) lhx2 and (D) lhx9 are shown (n=3 replicates, P<0.001 and P<0.05). rpl8 was used as an internal control. (E) Bar graph of the quantification and statistics of phenotypes from control or fezf2 morphant groups regarding the expression of lhx2 and lhx9. (F-G) Validation of lhx2 i2e3 MO. (F) Validation by RT-PCR. M: marker. Lane 1-2: rpl8; 1: control MO; 2: lhx2 MO. Lane 3-4: lhx2; Lane 3: control MO, Lane 4: lhx2 MO. (G) Validation by qPCR. rpl8 was used as an internal control in both cases. n=3 replicates, P<0.001. (H-I) Validation of lhx9 e1i1 MO. (H) Validation by RT-PCR. M: marker. Lane 1-2: rpl8; 1: control MO; 2: lhx9 MO. Lane 3-4: lhx9; Lane 3: control MO, Lane 4: lhx9 MO. (I) Validation by qPCR. rpl8 was used as an internal control in both cases. n=3 replicates, P<0.01. (J) Bar graph of the quantification and statistics of phenotypes from different morphant groups regarding arx expression. (K) Bar graph of the quantification and statistics of phenotypes from different morphant groups regarding ngn1 expression. In all bar graphs, numbers above each bar represent counts of embryos examined in each group. Error bars represent ±s. e. m.|