DEV015990 Supplementary Material

Files in this Data Supplement:

• Supplemental Figure S1 -

  Fig. S1. Generation of a conditional Gli3 allele and verifying the generation of a null allele. (A) Schematic of the gene targeting strategy. A 5.8 kb genomic fragment of Gli3 containing intron 8 was inserted into a targeting vector upstream of a neo cassette flanked with Frt sites (orange triangles), with one loxP site (red triangles) 3´ to the neo cassette and a TK cassette. Another loxP site was inserted into intron 9 in a 4.8 kb genomic fragment of Gli3. This fragment was then inserted 3΄ to the loxP site in the neo-TK vector (Fig. 1A). The targeting vector was inserted into the Gli3 locus by homologous recombination. (B) Targeted ES cell clones were identified by Southern blot analysis using SphI and a 1 kb external probe 5΄ of the SphI site (blue bar in A). (C) PCR primers S1 and AS3 (blue arrows in A) were used to detect the floxed allele (500 bp); S1 and AS2 (blue arrows in A) to detect the recombined allele (300 bp). (D-F) To verify that recombination between the loxP sites produces a Gli3-null allele, we intercrossed Gli3^rec/+ mice. Gli3^rec/rec P0 mice displayed the same brain phenotype as Gli3^Xt/Xt embryos (Fig. 1D-F). The brain lacked olfactory bulbs (Ob) and a cerebral cortex (Ctx) and had an enlarged midbrain (Mb). Cb, Cerebellum; Hb, hindbrain. Scale bar: 700 μm.

• Supplemental Figure S2 -

  Fig. S2. Variable tectal, isthmic and cerebellar phenotypes in Gli3^–/– and En1-Gli3 cko mutants. (A-L) Hematoxylin and Eosin (H+E) staining of (A,C,E,G,I) midline and (B,D,F,H,J) lateral sagittal sections of E18.5 and P0 Gli3^–/– mutant brains and (K,L) midline sagittal sections of P0 En1-Gli3 cko mutant brains. Note the variable morphology of the cerebellum (Cb), isthmus (Is), superior (Sc) and inferior colliculus (Ic). This is particularly obvious in medial sections. (M,N) Quantitative assessment of Is (M) and Ic (N) size in WT and Nes-Gli3 cko brains. The areas of the IC and Is were determined based on morphological and cytoarchitectural criteria. The IC was measured in sections ranging from medial (M) to more lateral (L). Data points represent means of samples from three different animals ±s.e.m. Student’s t-test was performed for statistical significance. Scale bar: 500 μm

• Supplemental Figure S3 -

  Fig. S3. Increased proliferation in Gli3 mutants. Proliferation, cell death and cell cycle exit were analyzed in En1-Gli3 cko or Nes-Gli3 cko mutants 1 or 2 days after gene inactivation to analyze morphologically normal tissue. To ensure that cells were counted specifically in the mes/r1 and to quantify ventral and dorsal regions separately we performed in situ hybridization and/or immunostaining for region-specific markers on sections adjacent to the ones used for PH3/BrdU/caspase 3 immunostaining (see Material and methods). The results are means of samples from at least three different animals ±s.e.m. Student’s t-test was performed; P-values are indicated. Note that although the differences in proliferation or cell death between WT and cko mutants are not statistically significant (P>0.05), we observed a strong trend (P≤0.1, highlighted in red) to increased proliferation in the dorsal, but not ventral mes/r1 or mes in the cko mutants. (A) Cells in the DNA replication phase (S phase) were quantified in E9.5 WT and En1-Gli3 cko mutants. BrdU-positive nuclei (1-hour BrdU pulse) and total number of nuclei (DAPI⁺) were counted. The number of BrdU⁺ cells was divided by the number of DAPI⁺ cells. There is no obvious difference in proliferation between WT and cko mutant in the ventral mes/r1 (wt-v and cko-v) or between WT ventral (wt-v) and dorsal mes/r1 (wt-d). By contrast, proliferation is slightly increased in the dorsal mes/r1 of En1-Gli3 cko mutants (cko-d) as compared with WT dorsal mes/r1 (P=0.08). (B) Cells undergoing G2/mitosis were quantified in E10.5 WT and En1-Gli3 cko mutants by counting the number of PH3-positive nuclei and subsequent normalization for the length of the ventricle. There is no obvious difference in proliferation between WT and cko mutant in the ventral mes/r1 (wt-v and cko-v). By contrast, proliferation appears to be slightly increased in the dorsal mes/r1 of En1-Gli3 cko compared with WT dorsal mes/r1 (wt-d and cko-d). (C) Cells undergoing G2/mitosis were quantified in E12.5 WT and Nes-Gli3 cko mutants by counting the number of PH3-positive nuclei and subsequent normalization for the length of the ventricle. There is no obvious difference in proliferation between WT and cko mutant in r1 (wt-r1 and cko-r1) or the ventral mes (wt-v and cko-v). By contrast, proliferation is slightly increased in the dorsal mes/r1 of Nes-Gli3 cko compared with WT dorsal mes/r1 (wt-d and cko-d). (D,E) Cells undergoing cell death were quantified in E9.5 (D) or E10.5 (E) WT and En1-Gli3 cko mutants by counting cells positive for cleaved caspase 3 and subsequent normalization for the length of the ventricle. Note that cell death is slightly decreased in the dorsal mes/r1 of E10.5 En1-Gli3 cko mutants (E, P≤0.1, wt-d and cko-d). (F) Cell cycle exit was quantified by analyzing the fraction of cells that exit the cell cycle in E12.5 WT and Nes-Gli3 cko mutants. Embryos were exposed to a single-pulse label of BrdU 24 hours before being sacrificed. Coronal sections of E12.5 WT and Nes-Gli3 cko mutants were immunostained for Ki67/BrdU. The fraction of cells that left the cell cycle (BrdU⁺ Ki67^– divided by the total number of BrdU⁺ cells) was determined (Chenn and Walsh, 2002). No significant difference was detected between WT and cko mutants in ventral mes (wt-v and cko-v), dorsal mes (wt-d and cko-d) or r1 (wt-r1 and cko-r1).

• Supplemental Figure S4 -

  Fig. S4. Rescue of cerebellar phenotype is apparent in direct comparison of Gli3^–/–;Fgf8^+/– and Gli3^–/–;Fgf8^+/+ mutant littermates. (A-L) Comparison of E18.5 Gli3^–/–;Fgf8^+/– (A,B,E,F,I,J) and Gli3^–/–;Fgf8^+/+ (C,D,G,H,K,L) littermates shows the partial rescue of the cerebellar, but not the midbrain, phenotype after removal of one copy of Fgf8. (A-K) In the histological analysis (H+E staining) the rescue is particular apparent medially. (I-L) RORa immunostaining shows that although Purkinje cells are still found in clusters in Gli3^–/–;Fgf8^+/– mutants, the Purkinje cell layer in the medial and lateral cerebellum is more comparable to WT (Fig. 3I). Note that the images in this figure are also used in Fig. 2J, Fig. 7A,G, and Fig. S2A,C. (M) Quantitative assessment of tectum (Tc), isthmus (Is) and cerebellar (Cb) size in WT and Gli3^–/–;Fgf8^+/– brains. Note that only the Tc and Is are increased in size in Gli3^–/–;Fgf8^+/– mutants.