Verexpression Represses the Title Loaded From File Zfp423 Intron 59 Enhancer in P19 CellsSince the predicted Zfp423 binding sites were not required for enhancer activity on a heterologous promoter and deletion 10781694 of these binding sites appears to increase enhancer strength, we next tested whether ZNF423 expression had any effect on reporter gene expression (Figure 5). Co-transfection of reporter constructs with a short hairpin RNA (shRNA) targeting Zfp423 mRNA that reduces Zfp423 to ,20 normal levels in P19 cells did not reproducibly alter reporter activity in triplicate measures from a single DNA preparation per construct (Figure 5A). In parallel, a construct including the conserved segment at the intron 3 site showed no enhancer activity either before or after knockdown of Zfp423 RNA, in comparison to pGL4-TAL. However, overexpression of human ZNF423 substantially attenuated expression of the reporter, compared to pcDNA expression vector control (Figure 5B), providing further evidence for a negative effect of Zfp423 at high expression levels. This effect was not seen in cotransfection when the Zfp423 binding sites were mutated, suggesting a direct effect of binding. Since the Zfp423 binding site overlaps a predicted binding site for EBF family members, we then tested whether reducing Ebf expression in P19 cells affects reporter activity (Figure 5C). From duplicate measurements for each of three independent DNA preparations per construct, we again saw no effect of shRNA targeting Zfp423, but a highly significant decrease on overexpression of human ZNF423 (p,1027, Tukey HSD pair-wise test after one-factor ANOVA). Targeting Ebf1 with shRNA resulted in a similar loss of reporter expression (p,1027), suggesting that Ebf1 contributes to activation of this enhancer. Targeting Ebf2, which appeared to be expressed at lower levels based on qRT-PCR data (Figure 2C) had only a modest effect. Simultaneously targeting both Ebf1 and Zfp423 by co-transfection of shRNA constructs was not significantly different from targeting Ebf1 alone. However, overexpression of ZNF423 together with Ebf1 shRNA further reduced reporter expression (p = 0.029) below the level achieved with EbfZfp423 Binds Autoregulatory SitesZfp423 Binds Autoregulatory SitesFigure 4. Zfp423 intron 5 binding site is an enhancer. (A) Reporter constructs are shown Title Loaded From File schematically. A pTAL minimal promoter was inserted upstream of the firefly luciferase reporter in pGL4. Fragments of the Zfp423 intron 5 putative enhancer were placed as indicated by the alignment of grey boxes. Position of the Zfp423 consensus match is indicated by XX in the “sites mutated” construct. Horizontal lines indicated constructs tested together, with pGL4-TAL repeated in each group. (B) Reporter gene expression levels, expressed as the ratio of firefly luciferase to co-transfected Renilla luciferase enzymatic activity measured by luminescence. Each measurement was normalized to the average of control vector (pGL4-TAL) samples run on the same day. For each construct, 2? DNA preparations were each used in 2? independent co-transfections for 9?2 (top eight constructs) or 6 (bottom three constructs) measurements. (C) Sequence changes in the sites mutated construct. Top line shows sequence of the mouse reference clone. The overlapping ROAZ (Zfp423), OLF1 site predicted by SynoR is boxed. Consensus Zfp423 (ROAZ) binding site and an adjoining consensus half-site (grey text) are indicated. Bottom line shows mutated sequence, with altered residues.Verexpression Represses the Zfp423 Intron 59 Enhancer in P19 CellsSince the predicted Zfp423 binding sites were not required for enhancer activity on a heterologous promoter and deletion 10781694 of these binding sites appears to increase enhancer strength, we next tested whether ZNF423 expression had any effect on reporter gene expression (Figure 5). Co-transfection of reporter constructs with a short hairpin RNA (shRNA) targeting Zfp423 mRNA that reduces Zfp423 to ,20 normal levels in P19 cells did not reproducibly alter reporter activity in triplicate measures from a single DNA preparation per construct (Figure 5A). In parallel, a construct including the conserved segment at the intron 3 site showed no enhancer activity either before or after knockdown of Zfp423 RNA, in comparison to pGL4-TAL. However, overexpression of human ZNF423 substantially attenuated expression of the reporter, compared to pcDNA expression vector control (Figure 5B), providing further evidence for a negative effect of Zfp423 at high expression levels. This effect was not seen in cotransfection when the Zfp423 binding sites were mutated, suggesting a direct effect of binding. Since the Zfp423 binding site overlaps a predicted binding site for EBF family members, we then tested whether reducing Ebf expression in P19 cells affects reporter activity (Figure 5C). From duplicate measurements for each of three independent DNA preparations per construct, we again saw no effect of shRNA targeting Zfp423, but a highly significant decrease on overexpression of human ZNF423 (p,1027, Tukey HSD pair-wise test after one-factor ANOVA). Targeting Ebf1 with shRNA resulted in a similar loss of reporter expression (p,1027), suggesting that Ebf1 contributes to activation of this enhancer. Targeting Ebf2, which appeared to be expressed at lower levels based on qRT-PCR data (Figure 2C) had only a modest effect. Simultaneously targeting both Ebf1 and Zfp423 by co-transfection of shRNA constructs was not significantly different from targeting Ebf1 alone. However, overexpression of ZNF423 together with Ebf1 shRNA further reduced reporter expression (p = 0.029) below the level achieved with EbfZfp423 Binds Autoregulatory SitesZfp423 Binds Autoregulatory SitesFigure 4. Zfp423 intron 5 binding site is an enhancer. (A) Reporter constructs are shown schematically. A pTAL minimal promoter was inserted upstream of the firefly luciferase reporter in pGL4. Fragments of the Zfp423 intron 5 putative enhancer were placed as indicated by the alignment of grey boxes. Position of the Zfp423 consensus match is indicated by XX in the “sites mutated” construct. Horizontal lines indicated constructs tested together, with pGL4-TAL repeated in each group. (B) Reporter gene expression levels, expressed as the ratio of firefly luciferase to co-transfected Renilla luciferase enzymatic activity measured by luminescence. Each measurement was normalized to the average of control vector (pGL4-TAL) samples run on the same day. For each construct, 2? DNA preparations were each used in 2? independent co-transfections for 9?2 (top eight constructs) or 6 (bottom three constructs) measurements. (C) Sequence changes in the sites mutated construct. Top line shows sequence of the mouse reference clone. The overlapping ROAZ (Zfp423), OLF1 site predicted by SynoR is boxed. Consensus Zfp423 (ROAZ) binding site and an adjoining consensus half-site (grey text) are indicated. Bottom line shows mutated sequence, with altered residues.