KDM3A uncovers an actin-mediated control on IFT that precludes IFT81 overexpression. (A) Immunoblots with DDK antibody of total cell extracts harvested 24 h after transfection with IFT81-DDK plasmid. IFT81^null cells transfected with 5HT6-EGFP used as negative control for DDK signal (red arrow). Ponceau red is a loading control. (B) Line graph showing the number of 5HT6-EGFP–positive cells with a cilium at the indicated time points. Error bars represent SEM between fields of view. Linear regression and p-values from comparing the individual slopes by paired t test. See also Fig. S3 (E and F). (C) Median and range of the percentage of 5HT6-EGFP–positive cells that contain a cilium under each experimental condition. Each dot is the percentage of one independent experiment (N = 4 transfections, n > 100 GFP+ cells scored/condition). **, P < 0.01 (paired t test). (D) Mean ± 10th–90th percentile of lengths of 5HT6-EGFP–positive ciliafrom the two last cotransfections with IFT81-DDK presented in C. ***, P < 0.001 (paired t test). (E) Plates fixed after live imaging confirm equal cell densities in RPE1^WT and KDM3A^null cultures. (F) Immunoblots of total cell extracts from the indicated wild-type and mutant lines after 3 h −FCS (ciliary growth) indicate comparable total levels of endogenous IFT components. (G) Mean ± 10th–90th percentile of RPE1^WT cells cotransfected with 5HT6-EGFP and vector control (−) or IFT81-DDK plasmid that contain a cilium measured 24 h after transfection in the presence or absence of 0.5 µM CytoD. Each dot is the percentage of one field of view taken (N = 2 transfections, n = 4, 28, 12, and 20 fields scored/condition). Analysis of variance: P < 0.004; t test for the indicated paired comparisons: *, P < 0.05; ***, P < 0.0005. (H) Graphical summary illustrates the insensitivity of RPE1^WT to IFT81 overexpression unless actin is depolymerized or KDM3A is absent. ns, not significant.