Vesicles clustering around Wdr35-/- cilia lack electron dense decorations although electron-dense clathrin coated vesicles are still observed budding from the mutant plasma membrane (Figure 7- source data 1)

TEM sample preparation: 24 h serum-starved MEFs were chemically fixed for flat embedding using the following protocol: (1) Cells were grown on 60 mm dishes, and ciliogenesis was induced by serum starvation for 24 h. (2) For prefixation under culture conditions, 25 % glutaraldehyde was added to the growth medium to a final concentration of 1 %, mixed gently, and incubated for a few min at 37 ℃. (3) The growth medium (containing the glutaraldehyde) was replaced with a sample buffer (0.1 M HEPES, 4 mM CaCl2, pH 7.2) containing 2 % glutaraldehyde and incubated 1 h at room temperature (replacing the fixation buffer with fresh one after 20 min). All prefixation solutions were pre-warmed to 37 ℃, and all steps were done at 37 ℃, to preserve the cytoskeleton. (4) The fixation buffer was replaced with fresh fixation buffer and incubated for 4 h at 4 ℃. (5) After that, the sample was washed once in sample buffer and 2–3 times in distilled water, each for 5–10 min, gently removing and replacing the buffer. (6) Samples were incubated in 1 % OsO4 (EMS) (in distilled water) for 1 h at 4 ℃, (7) washed 3–4 times for 10 min each in distilled water, and (8) incubated in 1 % uranyl acetate (EMS) in distilled water overnight at 4 ℃. (9) Then, samples were rinsed 3–4 times for 10 min each in distilled water and (10) dehydrated using a graduated series of ethanol: 30 %, 50 %, 70 %, 80 %, 90 %, 96 % ethanol, 5 min each step at 4 °C, followed by twice rinsed in anhydrous 100 % ethanol 10 min each at RT. (11) Infiltration was performed using a 1:1 mixture of LX112 (Ladd Research, USA; EMS) and ethanol 2 h, followed by pure LX112 overnight and another 2 h pure LX112, where all steps were performed at room temperature. (12) Flat embedding: For flat embedding, the caps of the BEEM embedding capsule (size, #3, EMS) were cut off and capsules filled with LX112. The capsules were inverted over a selected area of the cell monolayer in the dish, and the resin cured at 60 °C oven for 48 h. The capsule was then removed by breaking off from the dish, leaving the monolayer cells embedded in the surface of the block. (13) Sectioning and post- staining: For sectioning and post-staining, 300 nm thick serial sections were cut by Leica Ultracut UCT (Leica microsystem, Wetzlar, Germany) with a diamond knife and sections picked up with a Formvar (EMS) coated 1x2 mm slot copper grid (EMS). Sections were post-stained with 2 % uranyl acetate for 10 min, then with lead citrate for 5 min. Imaging: Sections were stained on the grid with fiducials (15 nm gold nanoparticles, Sigma-Aldrich). 70 nm thick sections were cut for regular TEM imaging, and 300 nm thick sections were prepared for tomographic acquisition. Tilt series were acquired on a Tecnai F30 (FEI) transmission electron microscope, operated at 300 kV, and equipped with 2048x2048 Gatan CCD camera and FEI Titan Halo transmission electron microscope operated at 300 kV equipped with a field emission gun (FEG) and a Gatan K2 direct detector. The SerialEM software (Mastronarde, 2005) was used for automatic acquisition of double tilt series. Tomographic tilt series were recorded with a pixel size of 1.235 nm on Titan Halo and 1.178 nm on F30, a maximum tilt range of about 60°, and tilt steps of 1°. Tomographic reconstruction, joining of tomograms from consecutive sections, segmentation, and visualization of the tomograms was done using the IMOD software package (Kremer et al., 1996). 24 h serum-starved WT, Wdr35-/-, and Dync2h1-/- cells were serially sectioned parallel to the adherent surface. Two to four 300nm parallel serial sections are required to get the whole 3D volume ultrastructural view covering full cilia and their cellular surroundings. We reconstructed 45 tomograms to get a minimum of 3-4 whole cell volumes for each genotype. We took micrographs of 30 WT, 20 Wdr35-/-, and 30 Dync2h1-/- cells for this study. Immunofluorescence: Cells were washed two times with warm PBS, then fixed in either 4 % PFA in 1X PHEM/PBS 15 min at room temperature, 2 % fresh glutaraldehyde in 1X PHEM for 15 min, or pre-extracted for 30 s on ice in PEM (0.1 M PIPES pH 6.8, 2 mM EGTA, 1 mM MgSO4) prior to fixing in ice cold methanol on ice for 10 minutes according to Table 1, then washed twice with PBS. 1X PHEM (pH 6.9) contains: 60 mM PIPES, 25 mM HEPES, 10 mM EGTA, and 4 mM MgSO4·7H20). The cells were treated twice with 50 mM NH3Cl for 15 min each for PFA fixed cells, or 0.01 mg of NaBH4 in 1X PBS for 7 min for glutaraldehyde fixed cells to quench autofluorescence. Cells were then washed twice with PBS. Cells were permeabilized with 0.25 % Triton-X 100/TBS for 10 min at room temperature. Cells were rinsed twice in 1X TBS for 5 min. Blocking for non-specific binding was done by incubating samples in 10 % donkey serum in 0.2 % Tween-20/TBS for 60 min at room temperature. Samples were washed twice with PBS. Primary antibodies (Table 1) were added to samples and incubated for 60 min at room temperature or 4 ℃ overnight, in dilutant made of 1 % donkey serum in 0.025 % Triton X-100/TBS. Samples were washed in 0.25 % Triton-X 100/TBS 4-6 times, 10 min each. Secondary antibodies diluted in 1 % donkey serum and 0.025 % Triton X-100/TBS were incubated on samples for 60 min room temperature. Samples were washed with 0.25 % Triton-X 100/TBS 4-6 times 10 min, stained with DAPI (1:1000) in PBS for 5 min at room temperature, again washed with PBS and directly imaged or coverslips were added on slides using ProLong Gold antifade (ThermoFisher Scientific), according to the manufacturer’s instructions. Confocal imaging for was done on a Leica SP5 using the LAS-AF software, 405 nm diode, Argon and 561 and 648 nm laser lines, three Photomultiplier tubes, and one HyD GaSP detector, as per the requirement of the experiment. Images were scanned using a 63X 1.4NA oil immersion objective and latter processed using ImageJ and Imaris software. Image analysis and measurements: All image processing was performed using FIJI (Schindelin et al., 2012). Etomo and IMOD (Kremer et al., 1996) were used to reconstruct tomograms and manually segment tomograms respectively. These segmentations were used to create objects using the 3D Image suite in FIJI. The 3D centroids were obtained and the manually segmented ROI on the 2D slice that the 3D centroid was on was selected to move forward with. A 20 nm width band around this ROI was measured using the “Make Band” function. The integrated density of this band ROI was quantified as an indication of how electron dense the region around the user segmented vesicle is. 3D objects were measured using the 3D Image Suite. To quantify clathrin intensity around the cilia base, a point was manually selected as the center of the basal point. The user was blinded to file name and condition while quantification took place. This point was expanded 1 µm in each direction to create a shall of 2 µm diameter in x,y, and z. This shell was then measured using the 3D image suite in ImageJ (Ollion et al., 2013). Statistical analyses were carried out in GraphPad Prism8.