DEV014142 Supplementary Material

Files in this Data Supplement:

• Supplemental Figure S1 -

  Fig. S1. Characterization of the Mitf^mi-rw (rw) deletion. Genomic DNA from wild-type and homozygous Mitf^mi-rw mice was first subjected to a series of PCR reactions using primer sets spaced about 10 kb apart to roughly map the deleted region. Fine mapping was then performed using sets of primers spaced 1 kb apart. Finally, after mapping the expected junction across the deletion to within 1 kb, the junction was amplified and sequenced. Confirmation of the deletion by Southern blot analysis is shown. (A) Schematic representation of the BamHI and BglII restriction sites in wild-type and rw genomic DNA and the position of the 5′ and 3′ hybridization probes. (B) The sizes of the hybridizing bands are consistent with a clean deletion in rw of 86,346 bp, as determined by PCR using primers spanning the deletion.

• Supplemental Figure S2 -

  Fig. S2. Exon 1B1b-deleted Mitf cDNAs initiate translation from internal start codons. (A) Immunoprecipitation/immunoblots of electrophoresed extracts from NIH3T3 cells transiently transfected with the V5-tagged Mitf cDNAs depicted on the left. Expression constructs lacking exon 1B1b are called AΔB, JΔB, etc. JΔB and EΔB contain novel open reading frames potentially initiating from AUGs in the normally non-coding 5′ exons. These AUGs, however, might not serve as efficient initiation signals as they do not conform to Kozak rules. Note that MITF pan-specific immunoprecipitation, followed by anti-V5 immunoblotting (upper panel, marked IP:anti-CTD/IB: anti-V5), shows the lower-molecular-weight products expected from an exon 1B1b deletion (compare lanes 2-4 with 6-8). When the same immunoprecipitates were immunoblotted with anti-1B1b MITF instead of anti-V5 antibodies (lower panel, marked IP: anti-CTD/IB: anti-1B1b), only the full-size exon 1B1b-containing products are detected (lanes 2-4). (B) Potential AUG start codons corresponding to methionine 62, 75, 77, 105, 114 and 116, were mutated individually or in combinations in an expression construct of JΔB. The positions of these methionines in the respective exons are shown in the schematic on the left. The AUG potentially linked to a novel open reading frame in JΔB was not specifically mutated because, as mentioned, it does not conform to Kozak rules and would lead to a product larger than the ones actually observed in in vitro translations. Coupled in vitro transcription/translation of the respective cDNAs shows a product differing in electrophoretic mobility only when the AUG corresponding to Met62 was mutated, suggesting that this AUG is the major site of translation initiation. When this methionine is mutated, however, initiation appears to shift to one of the downstream AUGs or, potentially, alternative start codons. (C) Exon 1B1b-deleted MITF isoforms accumulate in the nuclei of cells. NIH3T3 cells were transfected either with wt J-MITF or rw JΔB-MITF and labeled with the indicated antibodies. Note the absence of anti-1B1b staining on cells expressing JΔB MITF, confirming the specificity of this antibody. (D) Transcriptional activities of the novel MITF isoforms. Expression constructs encoding the indicated isoforms were co-transfected into NIH3T3 cells along with the tyrosinase-luciferase (tyr-luc) reporter construct and fold-increase in activity over baseline was measured 24 hours later. Cell extracts were also checked for MITF protein expression, using immunoprecipitation and immunoblotting as in the top panel shown in A. Note that all novel isoforms show activity, and that particularly J- and E-Mitf show multiple bands that are likely to reflect initiation from various start codons.

• Supplemental Figure S3 -

  Fig. S3. CHX10 directly binds to the A-, H- and D-promoters in chromatin immunoprecipitation (ChIP) assays. Eye primordia and surrounding tissues were dissected from E10.5-11 wild-type embryos, minced and fixed in 1% formaldehyde. The crosslinking reaction was stopped with 0.125 M glycine on ice. The tissue was then frozen, ground on dry ice, and DNA was extracted using ChIP IT Kit (Active Motif) with enzymatic shearing. Immunoprecipitation was performed separately with an anti-CHX10 C-terminal (C) or N-terminal (N) antibody as well as a control IgG. For each promoter, potential homeodomain binding sites were determined within the first 1100 bp preceding the respective transcriptional start sites. Primers were then designed to amplify PCR products of approximately 200 bp covering the potential binding sites. Red boxes indicate amplicons containing sites where CHX10 binds and light-gray boxes sites where CHX10 does not bind. Boxes above the horizontal lines correspond to the CHX10 consensus sequence 5′TAATtPuPu3′ and boxes below the horizontal lines to the consensus sequence 5′ATTAaPyPy3′.

• Supplemental Figure S4 -

  Fig. S4. Sequence conservation of homeodomain binding motifs in Mitf promoters. The respective binding motifs corresponding to CHX10 consensus binding sites of M. musculus (Mm), R. norvegicus (Rn), H. sapiens (Hs) and P. troglodytes (Pt) were aligned. Numbers (Hx-1, Hx-2, etc) correspond to the sites indicated in Fig. S3. Framed sequences correspond to the ones showing interaction with CHX10 in the ChIP assay in Fig. S3. Note that the choice of amplicons for the D-promoter (see Fig. S3) only indicates that at least two of the three framed sites interact with CHX10.