Figure S2.

Effect of depletion or over-expression of Class I ARF GTPases on 2D phenotype and proliferation of prostate cancer cells. (A and B) PC3 cells expressing mNG and Scr, (A) ARF1, or (B) ARF3 shRNA were plated at low density and imaged. Data is mean confluence ± SEM, normalized to time 0. n = 3 experimental replicates with 4 technical replicates/condition. P values (one-way ANOVA). (C and D) PC3 acini expressing mNG and Scr, (C) ARF1, or (D) ARF3 shRNA were plated for 4–72 h. CellTiter-Glow was added and luminescence measured to assess ATP-based cell viability. Data is mean luminescence ± SEM, normalized to 4-h time point. n = 4 experimental replicates with 3 technical replicates/condition. P values (one-way ANOVA). (E) Phase images of PC3 cells described in A and B. Scale bars, 100 μm. Representative of n = 3 experimental replicates with 4 technical replicates/condition. (F) Schema, machine learning applied to classify and quantify 2D PC3 cells into three phenotypic categories. Upper panel, mem:Venus Scr shRNA (green), whole cell stain (WCS, red), and Hoechst (nuclei, blue). Lower panels, Round (blue), Spindle (green), and Spread (red). Scale bar, 100 μm. Mean proportion of PC3 cells, expressing Scr shRNA, with each phenotype shown for n = 3 experimental replicates each with 18 technical replicates. Total of 290,830 cells quantified. (G and H) PC3 cells in 2D expressing Scr and (G) ARF1 or (H) ARF3 shRNA were classified into Round, Spindle, and Spread. Heatmaps, Log2 fold change over Scr. P values, one-way ANOVA, grayscale values as indicated. n = 4 and 2 experimental replicates with 3–4 technical replicates/condition for ARF1 and ARF3, respectively. (G) 15,392 (Scr), 26,486 (ARF1_KD2), 20,480 (ARF1_KD4), (H) 3,610 (Scr), 2,692 (ARF3_KD1), 2,882 (ARF3_KD2) cells were quantified in total. (I) Phase images of PC3 cells expressing mNG, ARF1-mNG or ARF3-mNG and Scr shRNA (upper panels). Scale bars, 100 μm. n = 3 experimental replicates with 4 technical replicates/condition. Confocal images (middle panels) show localization of mNG constructs (black, inverted images) in 2D cells. Scale bars, 20 μm. Magnified images of boxed regions shown (a–f). White arrows, ARF mNG in puncta. Scale bars, 10 μm. Images are representative of observations made in 3 experimental replicates. Also shown (lower panels) are phase images of PC3 acini, Round (red), Spindle (green), and Spread (blue). Scale bar, 100 μm. n = 6 and 4 experimental replicates for ARF1-mNG and ARF3-mNG, respectively, each with 2–4 technical replicates/condition. Quantitation shown in Fig. 4, E and F. (J and K) Confluence quantified in cells expressing mNG, (J) ARF1-mNG or (K) ARF3-mNG and Scr shRNA using phase images. Data is mean ± SEM, normalized to time 0. n = 3 experimental replicates with 4 technical replicates/condition. P values (one-way ANOVA). (L and M) PC3 acini expressing mNG, (L) ARF1-mNG or (M) ARF3-mNG and Scr shRNA were plated for 4–72 h. CellTiter-Glow was added and luminescence measured to assess ATP-based cell viability. Data is mean ± SEM, normalized to 4-h time point. n = 4 experimental replicates with 3 technical replicates/condition. P values (one-way ANOVA). (N) PC3 cells expressing mNG, ARF1-mNG, or ARF3-mNG and Scr shRNA were classified into Round, Spindle, and Spread. Heatmaps, Log2 fold change over mNG. P values, one-way ANOVA, grayscale values as indicated. n = 3 experimental replicates with 3–4 technical replicates/condition. 4,309 (mNG), 4,766 (ARF1), and 5,261 (mNG), 6,508 (ARF3) mNG-positive cells quantified in total.

or Create an Account

Close Modal
Close Modal