The drug Dyngo-4a constrains both lateral mobility and internalization of caveolae independently of dynamin.

(A) Immunoblots of dynamin triple knock out cells showing dynamin 2 expression after 5 days of Tamoxifen-treatment (KO) or no treatment (Ctrl) as indicated. (B) Representative immunofluorescent staining of Caveolin1 in dyn TKO cells treated as in (A) as indicated, Scale bar, 5 μm. (C) Quantification of the number of caveolae per square micrometer counted in the basal PM in immunofluorescent labeled dyn TKO cells treated as in (A) as indicated. Mean ± SD from at least 20 cells per condition. Significance was assessed using t test, p =0.7421. (D) Color-coded trajectories showing caveolae movements in the PM. Cav1-GFP was transiently expressed in dyn TKO cells after 5 days of Tamoxifen treatment. Cells were serum-starved 1 h prior to 30 min treatment with DMSO (Ctrl) or 30 μM Dyngo-4a before imaged on TIRF over 5 min. Red trajectories represent caveolae with long duration times and purple represents caveolae with short duration times as indicated. Scale bar, 10 μm. (E-F) Quantification of Cav1-GFP track duration time (E) and track displacement length (F). Numbers were related to ctrl-treated cells, and track mean from at least 23 cells per condition are shown ± SD (E) and track length mean from at least 15 cells per condition are shown ± SD (F). Significance was assessed using t test, ****p ≤ 0.0001. (G) Visualization of the track displacement of color-coded trajectories in (D) by alignment of the starting position of all tracks, representing lateral mobility. (H) Kymographic visualization of the track duration of color-coded trajectories in (D) that were present at time zero, representing axial stability.

Dyngo-4a binds and inserts into membranes.

(A) Top, Illustration of Langmuir-Blodget experiment. Bottom, representative Dyngo-4a adsorption to POPC monolayer. Dyngo-4a (30 μM) (black line) or DMSO (gray line) was injected underneath the lipid film with the starting surface pressure of 20 mN m−1and the surface pressure shift (Δπ) was recorded over time. (B) Top, illustration of the QCM-D set up. Bottom, QCM-D measurements of SLB formation and Dyngo-4a (30 μM) (black line) or DMSO (gray line) adsorption to a POPC SLB, followed by rinsing. Arrows depict time of injection of DMSO (equivalent to the concentration that Dyngo-4a is dissolved in), sample injection (Dyngo-4a or DMSO), DMSO injection and saline buffer wash (150 mM NaCl, pH 7.4) injection respectively.

Dyngo-4a inserts underneath the phospholipid headgroups at a similar position as cholesterol.

(A) Molecular dynamic simulation of Dyngo-4a in a membrane. Snapshot of a single Dyngo-4A molecule (cyan stick and red volume) in POPC bilayer (phospholipid headgroups in dark gray and lipid chains as light gray sticks, dark blue represents the aqueous phase) (B-C) AWH simulations of Dyngo-4a (B) or Chol (C) in POPC membrane. PMF profile show the energy required to move a Dyngo-4a molecule or a Chol molecule from the membrane center to the aqueous phase, respectively. (D) Molecular dynamic simulation of Dyngo-4a in a membrane. Snapshot of a single Dyngo-4A molecule (cyan stick and red volume) in POPC:Chol (70:30%) bilayer (phospholipid headgroups in dark gray and lipid chains as light gray sticks, chol as orange sticks, dark blue represents water). (E) AWH simulations of Dyngo-4a in POPC:Chol (70:30%) membrane. The PMF profile shows the energy required to move a Dyngo-4a molecule from the membrane center to the aqueous phase. In the PMF profiles the lines represent the mean and the shaded regions represent the SD from triplicate simulations. (F) Molecular dynamic simulation snapshot of multiple Dyngo-4a molecules in POPC:Chol (70:30%) bilayer (color coding as in (D)). Dyngo-4a molecules did not cross the membrane throughout the simulation so they were placed randomly at the beginning of the simulations and could translocate from one leaflet to another via periodic boundary condition. (G) Deuterium order parameter profiles for POPC tails in POPC:Chol (70:30%) membrane. Left show sn1 tail (saturated, 16 carbons) and right shows sn2 tail (unsaturated, 18 carbons). No Dyngo-4a (black), single Dyngo-4a (gray), multiple Dyngo-4a (red), dashed lines depict the outer leaflet and solid line the inner leaflet. (H) Contour plot based on heatmaps in Fig. S3E highlighting the predominant configuration whit Dyngo-4a (red) localized adjacent to a chol (orange) cluster in the outer leaflet.

Dyngo-4a decreases the lipid packing in the outer leaflet of the PM which is counteracted by increased levels of cholesterol.

(A) Visualization of the PM mobility of a ctrl-treated cell (top) and a Dyngo-4a-treated cell (bottom) from 5 minute TIRF movies. The outline of the PM at the start (0 s, red), middle (60, 120, 180 and 240 s, gray) and end (300 s, black) are depicted. Insets to the right show magnifications of the indicated areas. Scale bar, 10 μm. (B) Quantification of the PM mobility of cells following treatment with DMSO (ctrl) or Dyngo-4a at different concentrations as indicated. The means is shown from five cells per condition ± SD. (C) Representative electron micrographs of caveolae in DMSO treated (ctrl) and Dyngo-4a treated cells. Scale bar, 100 nm. (D) Young moduli calculated from force-indentation curves from the representative curves in (Fig S4) over 31 and 33 cells treated with DMSO and Dyngo-4a, respectively. Each data point is an average value calculated from at least 40 individual force curves on each cell. Statistical analysis: parametric student t-test with Welch’s correction. ns = non significant. (E) Representative deconvolved fluorescent micrographs of HeLa cells stained with C-Laurdan captured at 442 nm (left panel) and 483 nm (middle panel). Yellow arrows depict analyzed stretch of PM with no neighboring fluorescent internal membrane as seen in the right panel. Red area corresponds to membrane with high membrane order and purple corresponds to membrane with low membrane order. Scale bar, 5 μm. (F) Quantification of the mean GP-values for DMSO (ctrl) or Dyngo-4a treated cells as indicated. All data points from at least 26 cells per condition are shown, n = 3, mean ± SD. Significance was assessed using t test, ** p ≤ 0.0065. (G) Quantification of Cav1-mCh track duration time following chol:MβCD treatment as indicated. Numbers were related to ctrl-treated cells. All data points from at least 26 cells per condition are shown, n = 3, mean ± SD. Significance was assessed using t test, ****p ≤ 0.0001.