DEV033449 Supplementary Material

Files in this Data Supplement:

• Supplemental Figure S1 -

  Fig. S1. Early expression of Ash1 and of an Ash1 target in embryonic mouse retina. (A,B) At E11.5, Ash1 transcripts accumulate in a narrow and discrete domain encompassed by the expression domain of Ngn2. nt, neural tube. Scale bar: 55 µm. (C,D) Antibodies directed against Ash1 were used to immunoprecipitate cross-linked chromatin fragments prepared from E14.5 mouse retina (NR), E14.5 mouse telencephalon (Tel) (C), or transfected P19 cell line (D). Immunoprecipitates were analyzed by real-time PCR for the abundance of Delta4 and Ath5 regulatory sequences. (E) Real-time PCR analysis of the Delta4 transcripts during embryonic mouse retina and telencephalon development.

• Supplemental Figure S2 -

  Fig. S2. Distal and proximal regulatory regions of the Atoh7 gene. Genomic sequences were downloaded from the http://genome.ucsc.edu web site (human, NCBI35/UCSC hg17; dog, UCSC canFam2, sequenced and assembled by the Broad Institute of MIT, Harvard and Agencourt Bioscience; mouse, NCBI34/UCSC mm7 draft produced by the Mouse Genome Sequencing Consortium; rat, UCSC rn3 draft; opossum, UCSC monDom1 draft; chicken, UCSC galGal2). We used the cow genomic sequences from the EMBL database (http://www.ebi.ac.uk/embl), release 84, section 'wgs_cow'. Homologous genomic loci were identified based on BLAST searches (Altschul et al., 1997). The corresponding sequence stretches were extended by up to 20 kb upstream and 1 kb downstream of the locus, and were aligned using MLAGAN (Brudno et al., 2003). Transcription factor binding sites were predicted using the binding site matrices from TransFac 8.4 (Matys et al., 2003), processed following a procedure described by Rahmann et al. (Rahmann et al., 2003). Potential matches to the binding site matrices were then obtained individually for the genomic sequences analyzed, and mapped to each other through close proximity in the genomic sequence alignment. Well-conserved predicted transcription factor binding sites were extracted for further analysis. (A) Comparative genomics analysis of Atoh7 5′ non-coding sequences based on sequences from mouse, rat, human, dog, horse, cow, opossum and chicken. The analysis identifies a highly conserved proximal region (mouse −380 to −650), as well as a shorter distal region of conservation (mouse −1570 to −1700). The bp position 0 corresponds to the translation start site. Note that much of the coding region (CDS, delimited by vertical red lines in A) is less conserved than the regulatory regions. (A, bottom) Schematic representation of the proximal and distal upstream regions of the ATH5 gene. The filled squares are the conserved E-boxes. (B) Two regions from a dot plot (Sonnhammer and Durbin, 1995) comparing the genomic regions upstream of, and including, the Atoh7 coding region in mouse and chicken. Detail on the right shows the distal, proximal and CDS regions corresponding to the conservation peaks in A and alignments in C. The cross-hair is on the first ATG of both mouse and chicken CDS. Detail on the left shows two stretches of marked long-range conservation around 10 kb and 15 kb upstream of the mouse Atoh7 gene. An additional sequence stretch, around 6.7 kb upstream of the mouse Atoh7 gene, is conserved between human, mouse, rat, dog, cow and opossum. It is not shown here as it appears to be missing from the chicken genome. (C) Multi-species alignments of the Atoh7 proximal and distal promoter sequences. The E-boxes are numbered E1 to E9 according to numeration in the chicken (Hernandez et al., 2007). E-box E4 is highly conserved in birds and mammals but seems absent from both Xenopus laevis and X. tropicalis (Hutcheson et al., 2005). We did a comparative analysis of two fish species (stickleback and tetraodon) for which we found relevant sequence stretches. Although these two species are easily comparable to each other and reveal several potentially fish-specific regions of interest, only the proximal elements (E-boxes 1, 2 and 4) are unambiguously comparable to those of chicken and mouse. Although it is impossible to align the fish sequences to the distal elements, we identified, based on dot plots, another upstream E-box conserved in both fish species that might functionally replace E8 or E9.

• Supplemental Figure S3 -

  Fig. S3. Proteins isolated from Atoh7-expressing cells specifically bind the distal E-boxes E8 and/or E9 in the Atoh7 promoter. A CMV/hAtoh7 plasmid was transfected into HeLa cells and protein extracts were prepared 24 hours later. Protein extracts form stable electrophoretic mobility shift assay complexes with a radioactively labeled DNA probe encompassing E8 and E9. Arrow indicates the position of probes that bind protein complex isolated from Atoh7-expressing, but not from GFP-expressing, cells. Monoclonal (M01, M02) and polyclonal (poly) antibodies directed against the basic domain competed efficiently with the binding of the complex to the probe, whereas IgG did not.

• Supplemental Figure S4 -

  Fig. S4. The normal upregulation of Atoh7 correlates with the increase in both the number and fluorescence intensity of the RFP⁺ cells. (A-C) HH28 retinas were electroporated with a wt₂._2kb/GFP and wt₂._2kb/RFP plasmids and fluorescent cells were detected 24 hours later. Scale bars: 100 µm. (D) Cell numbers obtained with the wt₂._2kb/GFP plasmid at HH22-23 and HH28 are set at 100 and cell numbers obtained with the wt₂._2kb/RFP plasmid are given relative to this value. Data are presented as the mean ±s.d.; at least five electroporated retinas were analyzed per condition.

• Supplemental Figure S5 -

  Fig. S5. In chick, Atoh7 acts upon the distal promoter region at E9. (A-I) HH28 retinas were electroporated with plasmids wt₂._2kb/GFP (A), mE8₂._2kb/GFP (B), mE9₂._2kb /GFP (C) in combination with an Atoh7 expression vector. HH28 retinas were electroporated with plasmids as follows: wt₂._2kb/RFP plus wt₂._2kb/GFP (D-F), mE9₂._2kb/GFP (G-I) in combination with an Atoh7 expression vector. (D-I, insets) Magnification of double-labelled cells. Scale bars: 500 µm in A-C; 100 µm in D-I. (J) Cell number obtained with the wt₂._2kb/RFP plasmid is set at 25 and cell numbers obtained with the wild-type and mE9 promoter/GFP plasmids are given relative to this value. Data are presented as the mean ±s.d.; at least five electroporated retinas were analyzed per condition. Activities of the mE9 and wild-type promoters co-localized. Mutation of E9 led to a marked decrease in the number and fluorescence intensity of the GFP⁺ cells.

• Supplemental Figure S6 -

  Fig. S6. E9 requires the proximal E-boxes to mediate the effect of Atoh7. (A-F) HH28 retinas were electroporated with plasmids as follows: wt₂._2kb/RFP plus wt₂._2kb/GFP (A,B), mE1mE2mE4₂._2kb/GFP (C-F), in combination with an Atoh7 expression vector. (B,F insets) Magnification of the merge. In E,F inset, the open arrowheads indicate cells expressing the mutant but not the wild-type promoter. In F, the white arrowheads indicate cells expressing the wild-type but not the mutant promoter. Scale bars: 50 µm.

• Supplemental Figure S7 -

  Fig. S7. In chick, the mouse Atoh7 promoter regions display chick-like properties. (A-F) HH22-23 (A-C) and HH28 (D-F) retinas were electroporated with plasmids wt₂._2kb/RFP and pG1M5-2.3kb. HH28 retinas were electroporated with plasmid pG1M5-2.3kb in combination with an empty (G) or an Atoh7 expression (H) vector. (D-F, insets) Magnification of double-labelled cells. In G and H, the exposure time was the same. Scale bars: 100 µm in A-F; 300 µm in G,H. (I) Cell number obtained with the wt₂._2kb/RFP plasmid at HH22-23 is set at 10 and cell numbers obtained with the pG1M5-2.3kb plasmid are given relative to this value. Data are presented as the mean ±s.d.; at least three electroporated retinas were analyzed per condition. The mouse and chicken sequences display the same promoter specificity. Activity of the mouse promoter is upregulated by Atoh7 in the chick retina.

• Supplemental Figure S8 -

  Fig. S8. Atoh7 targets expressed in newborn RGCs. (A-E) Co-localization of Atoh7 promoter activity with Atoh7, stathmin 2 or pleiotrophin expression. Cells were transfected with a wt₂._2kb/lac plasmid at HH28. After 24 hours in culture, lacZ-expression was revealed and cells were processed for DISH with Atoh7- (A,B,E), stathmin 2 (STMN2)- (C,D,E) or pleiotrophin (PTN)-specific riboprobes (E). In A and C, the blue arrowheads indicate double-labelled cells (B and D). In A, white arrowheads indicate ³⁵S-labelled cells. (E) Cell number obtained with the wt₂._2kb/lac plasmid is set at 100 and the numbers of double-labelled cells are given relative to this value. (F) In vivo occupancy of the promoter regions of the Snap25, Stmn2 and β3nAChR genes by chicken Atoh7. Crosslinked chromatin fragments were prepared from HH22-23 and HH29-30 chick retinas.

• Supplemental Figure S9 -

  Fig. S9. Distribution of stathmin 2 and pleiotrophin transcripts in the developing chick retina. At HH41, stathmin 2 transcripts accumulate in the ganglion cell layer (GCL), in the inner plexiform layer (IPL) and in the inner aspect of the inner nuclear layer (INL). Pleiotrophin transcripts accumulate in the INL and the outer nuclear layer (ONL).