DEV031989 Supplementary Material

Files in this Data Supplement:

• Supplemental Figure S1 -

  Fig. S1. Efficacy of MITF shRNA in human and mouse melanoma cells. An shRNA to knock down MITF expression was constructed in the psiStrike-hMGFP vector, which expresses hMGFP to mark transfected cells. The sequence of the shRNA was chosen from a location in which the chick, quail, human and mouse genes have identical sequence so that it could be used in all species. The construct was transfected into B16-F10 mouse melanoma cells (A) and G361 human melanoma cells (B). After 24 hours, the cells were immunolabeled with anti-MITF (red). No hMGFP-positive cells were also positive for MITF. In addition, a few B16-F10 cells were MITF-negative without the shRNA.

• Supplemental Figure S2 -

  Fig. S2. Melanoblasts transdifferentiate into glial cells upon loss of MITF. (A) Similar to Fig. 2, quail melanoblast clusters were transfected with either the MITF shRNA construct or the empty vector. After 72 hours, the cells were fixed and stained for 7B3 and with anti-MITF antibodies. Elimination of MITF by shRNA in these cells causes them to express the 7B3 antigen. (B) Apoptosis was measured in melanoblast cluster cells 24 hours after transfection using Magic Red apoptosis reagent, which detects caspase 3/7 activity. No difference in apoptosis was observed between MITF shRNA-transfected and scrambled control-transfected cultures. (C) Cell proliferation was monitored in melanoblast cluster cells by cotransfecting RFP-ligase with the MITF shRNA or scrambled control constructs. RFP-ligase, in which DsRed is fused to DNA ligase I, is localized to sites of DNA replication during S phase (visible as puncta in the nucleus), and is visible throughout the nucleus during the rest of the cell cycle (Easwaran et al., 2005). The percentage of cells in S phase was calculated and compared between MITF shRNA-transfected and scrambled control-transfected cultures. No difference was observed.

• Supplemental Figure S3 -

  Fig. S3. Identification and confirmation of efficacy of the chick MITF proximal promoter. (A) The transcription initiation site of chick MITF was determined by 5′ RACE using the GeneRacer Kit (Invitrogen). 5′ RACE identified three transcription initiation sites, indicated in red. The 5′- and 3′-most were identified by several clones each, and the middle site was identified by only one clone. The 5′-most site is nearest that reported for human MITF (Fuse et al., 1996), so we refer to it as +1. (B) pMITF-EGFP drives expression of EGFP in migrating melanoblasts (arrows) in the dorsolateral migratory pathway of a stage 21 embryo. Scale bar: 10 µm. (C) Dorsal view of a stage 23 embryo electroporated with pMITF-EGFP. EGFP-positive cells are visible in a pattern characteristic of dorsolaterally migrating neural crest cells.

• Supplemental Figure S4 -

  Fig. S4. Repression of MITF by mutant FOXD3. For this study, we investigated the use of several luciferase vectors and various normalizing methods and vectors. We found that FOXD3 has some effect on the activity of some normalizing vectors, so the reliability of pCS2-βGal as a normalizing vector was confirmed by comparing β-galactosidase activity with counts of transfected cells. The vectors used here are those found to have the least sensitivity to FOXD3 repression. To verify the function of the mutant FOXD3 proteins, they were first tested in transfections in B16-F10 cells with the following: (1) empty pGL4.70; (2) CMV-BF-Rluc, which contains eight brain factor (a forkhead box transcription factor related to FOXD3) binding sites that are known to bind FOXD3 plus a minimal CMV promoter; and (3) CMV-Rluc, which contains a strong CMV promoter driving Renilla luciferase expression. In all cases, cotransfection with normal pFOXD3 was found to reduce expression, although the reduction was small. pFOXD3-VP16 increased expression from CMV-BF-Rluc dramatically, indicating that the fusion protein can function as a transcriptional activator. pFOXD3-VP16 also increased expression from CMV-Rluc and pGL4.70. pFOXD3-ΔC did not alter activity compared with the FOXD3-negative samples. Based on these data, we conclude that pGL4.70 contains at least one functional FOXD3 binding site. When transfected with pMITF-Rluc, all three FOXD3 variants repressed transcription, although FOXD3-VP16 and FOXD3-ΔC showed somewhat less repression than wild-type FOXD3. This is consistent with FOXD3 repressing transcription of MITF without directly binding to the DNA. Owing to the likely FOXD3 binding site in the pGL4.70 backbone, we conclude that some of the repression observed with pFOXD3 is probably due to residual binding to the backbone, pGL4.70. Thus, the actual level of repression due to the natural action of FOXD3 on the MITF-M promoter is likely to be more accurately reflected by that seen with pFOXD3-ΔC. The increase in expression seen with pFOXD3-VP16 compared with pFOXD3 is most likely to be due to binding to the backbone, rather than interaction with the MITF-M promoter. Data shown are representative of at least three separate trials. Error bars indicate the 95% confidence interval.