Fig 2: Modulating tubulin glutamylation affects ER distribution.a, Representative images of U2OS and COS7 cells sequentially immunolabeled for polyE (polyglutamylation), polyE plus GT335 (glutamylation) and Alexa 647 conjugated a-tubulin. GT335 signal indicates monoglutamylation in this case because polyglutamylation is pre-saturated with polyE antibody. b, Quantifications of signal distributions for cells as in (a). Data points from the same cell are linked by solid lines. n = 20 cells for U2OS, 35 cells for COS7. c, Western blot analysis of U2OS, COS7, HeLa and RPE1 cells. d, MDR ratios of monoE (indicated by GT335 labeling) to polyE, or microtubules (indicated by a-tubulin labeling) to polyE. n = 20 cells for U2OS, 35 cells for COS7. e, Representative images of U2OS cells overexpressing the indicated TTLLs or CCPs. f, g, Quantifications of ER and microtubule MDRs of cells as in e; n = 37, 40, 34, 40, 40, 41, 39 cells (left to right) for (f); n = 29, 38, 37, 51, 38, 46, 39 cells for (g). h, Quantifications of ER MDR of cells in p180/KTN1 double knockout cells overexpressing the indicated TTLLs or CCPs; n = 63, 61, 64, 63, 61 cells (left to right). i, Quantifications of ER MDR of wild-type or KTN1 knockout cells overexpressing TTLL7; n = 40, 64 cells (left to right). j, U2OS cells transfected with the indicated siRNAs for 72 h were analyzed by western blotting. k, U2OS cells transfected with the indicated siRNAs for 48 h were analyzed by real time PCR. n = 4 repeats. l, m, Representative image and quantification of ER in control or CCP5 siRNA transfected cells. Scale bars, 10 µm. n = 64, 70, 69, 39, 39, 39 cells (left to right). Data are mean ± s.d. P values shown on top, two-tailed t-test. See Supplementary Information for uncropped western blots. Source data

Fig 3: Microtubule binding of CLIMP63, p180 and KTN1 are differently affected by microtubule glutamylation.a, Western blots of wild-type or the indicated knockout U2OS cells. b, c, Western blots and quantifications of U2OS cells either treated with CNB to deplete the centrosome or else transfected with control siRNA or siAKAP450 (to deplete Golgi-derived microtubules). n = 6 experiments. d, Representative images of U2OS cells transfected with CLIMP63-mEmerald, p180-mEmerald, or KTN1-mEmerald, with or without co-expression of TTLL7. Note that CLIMP63 overexpression leads to dramatic ER-microtubule alignment (~80% of cells), while p180 and KTN1 require co-expression of TTLL7 (which polyglutamylates microtubules) for robust ER-microtubule alignment. Scale bar, 10 µm. e, Quantifications of microtubule-ER alignments in cells transfected with the indicated expression plasmids; n = 4 experiments, with at least 100 cells counted in each experiment. f, Quantifications of microtubule-ER alignments in cells treated with or without CNB, and overexpressing CLIMP63-mEmerald. n = 4 experiments, with 100 cells counted per experiment. g, Schematic diagram depicting how microtubule glutamylation levels are modulated by actions of tubulin glutamylases (TTLLs) and deglutamylases (CCPs). h, Western blots of U2OS cells overexpressing TTLL4 or TTLL7. polyE detects microtubule polyglutamylation (at least 2 glutamates in the side chain); GT335 reacts with branch points of microtubule glutamylation and thus detects both mono- and polyglutamylation. i, Western blots of U2OS cells overexpressing CCP1 or CCP5. Cells were subjected to microtubule sedimentation, and pellets (P, microtubule fraction) and supernatants (S, soluble fraction) at 37 °C were analyzed. j–l, Relative PLA intensities for CLIMP63, p180, or KTN1 with a-tubulin (indicative of microtubule binding) in U2OS cells overexpressing the indicated TTLLs or CCPs. n = 138, 103, 112, 99, 102 cells (left to right) for (j); n = 112, 113, 107, 131, 121 cells for (k); n = 188, 108, 113, 123, 109 cells for (l). m, Ponceau S staining of p180 and KTN1 proteins purified from HEK293T cells. n, Coomassie Brilliant Blue staining of purified CLIMP63 from HEK293T cells. o, p, Mass spectra of microtubules glutamylated by TTLL6 for different times, or by else TTLL4 or TTLL7, and used for in vitro microtubule-binding assays. Spectra display characteristic distributions of masses with peaks separated by 129 Da (corresponding to one glutamate). Peak labels show number of glutamates on a- and ß-tubulin. Glutamate numbers are indicated in green, dark green, orange, and red for a1B, a1C, ßI and ßIVb isoforms, respectively. q, Representative micrographs of CLIMP63-mNeonGreen-2×Strep (CLIMP63, cyan) showing binding to unmodified microtubules or else microtubules (magenta) glutamylated in vitro by TTLL7. Interference reflection microscopy images for the microtubule channel were background subtracted and inverted. Average numbers of glutamates added to microtubules as quantified from mass spectrometry data for each group are shown on top. Scale bar, 5 µm. r, Affinities of CLIMP63 for unmodified microtubules and microtubules glutamylated by TTLL7. X-axis indicates the weighted average of glutamate residues attached to a- and ß-tubulin. n =107, 61, 127, and 136 microtubules with unmodified, a+0E/ß+1.4E, a+0E/ß+2.7E, and a+0.1E/ß+3.8E microtubules, respectively. Data are mean ± s.d. P values shown on top, Mann-Whitney test for panel r, two-tailed t-tests for others. See Supplementary Information for uncropped western blots. Source data

Fig 4: PLA of CLIMP63, p180 and KTN1 with a-tubulin, and depleting centrosome or Golgi-derived microtubules affects ER distribution.a, N-terminal mEmerald tag was knocked into endogenous CLIMP63 using CRISPR/Cas9 to facilitate PLA, as no anti-CLIMP63 antibody targeting its cytosolic region was available. Western blots of wild-type or knock-in cells are shown. b, Representative PLA images of p180 or KTN1 with a-tubulin. Note that all PLA dots (a dot indicates one binding event) are localized adjacent to both ER and microtubules. c, Representative PLA images of endogenous mEmerald-CLIMP63 with a-tubulin. d, Western blot analysis of cells transfected with control or AKAP450 siRNAs. e–h, Representative data and quantifications for PLA of CLIMP63, p180, or KTN1 with a-tubulin, indicative of microtubule binding. Cells are either untreated, treated with centrinone B (CNB) to deplete the centrosome, or transfected with control siRNA or siAKAP450 (to deplete Golgi-derived microtubules). n = 130, 134, 142, 143 cells (left to right) for (f), n = 116, 139, 140, 150 cells for (g), n = 119, 117, 139, 166 cells for (h). i, j, Asymmetry and MDR quantifications for PLA signals between CLIMP63, KTN1, or p180 with a-tubulin. n = 83, 74, 72, 40, 38 cells (left to right) for (i), n = 81, 79, 82, 43, 44 cells for (j). k, Representative images of U2OS cells with or without CNB treatment. Cells stably expressing mEmerald-Sec61ß (green, ER) are stained with DAPI (blue, DNA) and immunolabeled with anti-pericentrin (red) and anti-a-tubulin (magenta) antibodies. Perinuclear and peripheral regions in the cells (boxed) are enlarged at the right. l, Quantifications of ER MDR in wild-type or p180 knockout (KO1 and KO2) cells, with or without CNB treatment. n = 31, 30, 31, 43, 31, 39 cells (left to right). m, n, Representative images and quantifications of ER asymmetry for wild-type or CLIMP63/KTN1 double-knockout cells. Cells were transfected with control siRNA or AKAP450 siRNA to deplete Golgi-derived microtubules. n = 31, 31, 31, 32, 31, 32 cells (left to right). Scale bars, 10 µm. Data are mean ± s.d. P values shown on top, two-tailed t-tests. See Supplementary Information for uncropped western blots. Source data

Fig 5: p180 and KTN1 bind glutamylated and polyglutamylated microtubules, respectively.a, Representative micrographs of p180 and KTN1 fragments (cyan) binding to unmodified microtubules or microtubules glutamylated in vitro by TTLL6 (magenta). Internal reflection microscopy images for the microtubule channel were background subtracted and inverted. Average numbers of glutamate molecules (E) added to microtubules as quantified from mass spectroscopy data in each group are indicated. Scale bar, 5 µm. b, Binding of p180 and KTN1 fragments to unmodified and TTLL6-glutamylated microtubules. n = 110 (unmodified), 165 (a + 3.5 E, ß + 1.3 E) and 112 (a + 8.3 E, ß + 2.6 E) microtubules for p180; n = 141 (unmodified), 186 (a + 3.5 E, ß + 1.3 E) and 156 (a + 8.3 E, ß + 2.6 E) microtubules for KTN1. c, Representative micrographs of p180 and KTN1 microtubule interacting fragments (cyan) showing binding to unmodified microtubules or microtubules glutamylated by TTLL4 or TTLL7 (magenta). KTN1 and p180 are shown with different brightness/contrast settings for TTLL4- and TTLL7-modified microtubules, reflecting large differences in binding between mono- and polyglutamylated microtubules. Scale bar, 5 µm. d, e, Affinities of p180 and KTN1 for unmodified microtubules and microtubules glutamylated by TTLL7 (d) or TTLL 4 (e). The x-axis shows weighted averages of glutamate residues attached to a- and ß-tubulin. n = 185 (unmodified) and 117 (a + 0 E, ß + 1.2 E) microtubules for p180; n = 128 (unmodified) and 225 (a + 0 E, ß + 1.2 E) microtubules for KTN1; n = 179 (unmodified) and 237 (a + 0 E, ß + 1.5 E) microtubules for p180; n = 163 (unmodified) and 224 (a + 0 E, ß + 1.5 E) microtubules for KTN1. f, CLIMP63 binds centrosomal microtubules, KTN1 binds perinuclear polyglutamylated microtubules, and p180 can bind peripheral microtubules with less glutamylation. Together, these proteins maintain proper asymmetric ER distribution, which regulates organelle distributions. MT, microtubules. Data are mean ± s.d. with individual data points shown. Kruskal–Wallis (b) or Mann–Whitney tests (d, e); P values are shown.Source data