Biophysical Journal, Volume 113 Supplemental Information Acetylated Microtubules Are Preferentially Bundled Leading to Enhanced Kinesin-1 Motility Linda Balabanian, Christopher L. Berger, and Adam G. Hendricks
!1 Acetylated microtubules are preferentially bundled leading to enhanced kinesin-1 motility Linda Balabanian 1, Christopher L. Berger 2, Adam G. Hendricks 1 1. Dept. of Bioengineering, McGill University, Montreal, QC 2. Dept. of Molecular Physiology and Biophysics, University of Vermont, Burlington, VT Correspondence: adam.hendricks@mcgill.ca Supporting Material Movies S1, S2, S3, S4: Time-lapse of kinesin-1 GFP motility assay on Taxol-stabilized extracted microtubule networks (see Fig. 1). 10x real-time. Movies S5 (see Fig. 3A): Time-lapse of tau-3rs Alexa 568 on Taxol-stabilized extracted microtubules (Fig. 3). 10x real-time. Figure S1: Extracted microtubule networks maintain the organization of the native cytoskeleton. A) CellMask plasma membrane staining (according to manufacturer s instructions, Thermo Fisher Scientific) of nonextracted live cell (left) compared to extracted cell (right) shows the breakdown of the plasma membrane of the extracted cell. B) SiR-tubulin (Spirochrome) staining of cells before and after extraction show that the organization of the microtubule (MT) network is similar in living and extracted cells, although the cytoskeleton contracts slightly during extraction. C) There is a measurable increase in the number of acetylated MTs in extracted cells compared to non-extracted cells. D) After setting the threshold for intensity (lower panels in C), the signal area is measured to determine the levels of acetylation and bundling. The proportion of acetylated tubulin over total tubulin in cells is
!2 estimated for non-extracted (n = 6 cells) compared to extracted cells with 20 µm Taxol (n = 8 cells) or 200 µm GMPCPP (n = 5 cells) as the MT-stabilizing agent, followed by fixation and immunostaining. The levels of acetylation in non-extracted cells agree with previous studies (1). The percentage of acetylated MTs increases in extracted cells, likely due to MT stabilization (ANOVA, post-hoc test, p < 0.05). On the left, the proportion of bundled MTs over total tubulin is estimated for the same cells. There are no significant differences in the extent of MT bundling between nonextracted and extracted cells (p > 0.50). Figure S2: Motor protein tracking generates high-resolution reconstruction of the microtubule network. A) Single-molecule localizations of kinesin-1 GFP were used to generate a super-resolution reconstruction of the microtubule (MT) network (left). Higher magnification images of the boxed region are shown below. The individual MTs can be clearly observed in the super-resolution reconstruction (left) compared to α-tubulin antibody staining (right). B) The intensity peaks of the two single filaments are apparent in the reconstruction, whereas the point spread function of the signals from these filaments is overlapping for the immunostaining. Note that the spacing between MTs varies along the length of the bundle. C-D) the full-width at half maximum (FWHM) was measured along the length of MTs at several points in (C-D). C) MT bundles are > 2x brighter (signal intensity) and 2x wider (FWHM) than single MTs on average in super-resolution reconstructions. D) Tubulin immunostaining of the bundles
!3 identified in (C) indicates bundles contain 2-3 MTs as estimated from signal intensity. The measured widths of single and bundled MTs are similar due to limited resolution in TIRF images. Error bars represent standard error of the mean. E,F) The apparent FWHM measured with the super-resolution reconstruction images corresponds to average spacing of ~40 nm between MTs in bundles. E) Theoretical estimates of signal width depending of number of MT filaments, based on approximations of reported sizes of kinesin-gfp430, MAPs and MT diameter in literature (2-6). F) The measured MT widths using super-resolution are similar to theoretical values (for 1, 2 and 3 MTs filaments respectively) and approximatively 4 times lower than the measured FWHM of filaments imaged using diffractionlimited immunofluorescence. Note that distance between MTs along the z-axis cannot be estimated from measurements, but the TIRF illumination is restricted to a depth of ~100 nm. Figure S3: Extracted microtubule cytoskeletons stabilized with GMPCPP. A) Detergent extraction is performed in the presence of GMPCPP (a slowly hydrolyzable GTP analog). Purified, exogenous tubulin conjugated to Alexa 647 (in magenta) added during extraction (0.25 µm) localizes to microtubule plus ends (yellow arrows), suggesting that GMPCPP acts to form stable caps at the plus ends of microtubules. B) Kinesin-1 trajectories are
!4 generated for analysis and the MT network is stained retrospectively. The area in the yellow box is zoomed for C) showing standard deviation (SD) map of kinesin-1 binding (upper left) and single-molecule trajectories (in different colours) (upper right). The arrows on acetylated tubulin and total tubulin images (lower left and right) show the regions of the numbered yellow lines shown on SD map, each line representing one of the 4 MT types of interest and used to generate kymographs shown in (D). D) Kinesin-1 run lengths and number of runs are increased on bundled MTs. E-F-G) Kinesin-1 behaviour on GMPCPP-stabilized MTs is mostly similar to observations on Taxol-MTs (Fig. 2), although kinesin run lengths are generally lower on GMPCPP-stabilized MTs compared to Taxol-stabilized MTs, which results in smaller differences in kinesin behaviour on single vs. bundled MTs. The raw means for each cell for all plots are shown in Fig. S4C. Acetylated single: n = 108, acetylated bundled: n = 1056, non-acetylated single: n= 1452, non-acetylated bundled: n = 833 total runs (5 cells over 3 experiments, runs > 50 nm). H) Off-axis displacement of kinesin-1 runs does not differ on single vs. bundled GMPCPP-stabilized MTs. Figure S4: Kinesin-1 motility on extracted microtubule networks stabilized with Taxol and GMPCPP. A) Taxol-stabilized and C) GMPCPP-stabilized extracted microtubules (MTs), for each cell (represented by different colours). See Fig. 1 and Fig. S3 respectively for normalized means. The means for run length and binding time are the inverse of the exponential decay of the exponential fit. B) Kinesin-1 run length (> 50 nm) distributions and fitted lines on single and bundled MTs for one Taxol-stabilized and D) for one GMPCPP-capped extracted cytoskeleton. For (B), single MTs: n = 678, bundled MTs: n = 676 runs and for (D), single MTs: n = 249, bundled MTs: n = 336 runs. E) Raw means of kinesin-1 in the absence (-tau) and presence (+tau) of 1 nm tau. F) Kinesin-1 run length (> 50 nm) distributions and fitted lines on single and bundled MTs with and without tau for one Taxol-stabilized extracted cytoskeleton. See Fig. 3 for normalized means. No tau: singe MTs: n = 121, bundled MTs: n = 562. With tau: single MTs: n = 116, bundled MTs: n = 384 runs.
!5 Figure S5: Kinesin-1 motility on single Taxol-stabilized microtubules polymerized from purified bovine tubulin. A) Kinesin-1 GFP430 kymograph on a reconstituted microtubule (MT). MTs were polymerized from purified bovine tubulin and bound to the coverslip with β-tubulin antibody that was absorbed to the coverslip surface. The motility assay, single-molecule tracking and analysis were conducted similarly as on Taxol-stabilized extracted MTs (see Materials & Methods). Direct comparison of kinesin motility on these reconstituted MTs with that on extracted MTs is limited as the reconstituted MTs are completely free from MAPs and do not intersect with other MTs or actin filaments. Also, blocking of non-specific interactions is more readily achieved in these purified systems. B) Kinesin run length and binding time distributions (runs > 50 nm) were exponentially fit and the mean was computed as the inverse of the decay constant. Error bars represent the standard deviation by bootstrapping using maximum likehood estimates (7). Kinesin-1 GFP run length mean is similar to previous in vitro studies using the same or similar kinesin constructs on reconstituted MTs (8, 9). C) The kinesin velocity distribution (for runs > 500 nm) indicates that maximal velocities are approximately 1 µm/s. Supporting References 1. Chesta, M. E., A. Carbajal, C. G. Bisig, and C. A. Arce. 2013. Quantification of acetylated tubulin. Cytoskeleton (Hoboken) 70:297-303. 2. Hirokawa, N., and R. Takemura. 2005. Molecular motors and mechanisms of directional transport in neurons. Nat Rev Neurosci 6:201-214. 3. Ross, J. L., M. Y. Ali, and D. M. Warshaw. 2008. Cargo transport: molecular motors navigate a complex cytoskeleton. Curr Opin Cell Biol 20:41-47. 4. Mikhaylova, M., B. M. Cloin, K. Finan, R. van den Berg, J. Teeuw, M. M. Kijanka, M. Sokolowski, E. A. Katrukha, M. Maidorn, F. Opazo, S. Moutel, M. Vantard, F. Perez, P. M. van Bergen en Henegouwen, C. C. Hoogenraad, H. Ewers, and L. C. Kapitein. 2015. Resolving bundled microtubules using anti-tubulin nanobodies. Nat Commun 6:7933. 5. Schneider, R., T. Korten, W. J. Walter, and S. Diez. 2015. Kinesin-1 motors can circumvent permanent roadblocks by side-shifting to neighboring protofilaments. Biophys J 108:2249-2257. 6. Chung, P. J., C. Song, J. Deek, H. P. Miller, Y. Li, M. C. Choi, L. Wilson, S. C. Feinstein, and C. R. Safinya. 2016. Tau mediates microtubule bundle architectures mimicking fascicles of microtubules found in the axon initial segment. Nat Commun 7:12278. 7. Woody, M. S., J. H. Lewis, M. J. Greenberg, Y. E. Goldman, and E. M. Ostap. 2016. MEMLET: An Easy-to-Use Tool for Data Fitting and Model Comparison Using Maximum-Likelihood Estimation. Biophys J 111:273-282. 8. Walter, W. J., V. Beranek, E. Fischermeier, and S. Diez. 2012. Tubulin acetylation alone does not affect kinesin-1 velocity and run length in vitro. PLoS One 7:e42218. 9. Kaul, N., V. Soppina, and K. J. Verhey. 2014. Effects of alpha-tubulin K40 acetylation and detyrosination on kinesin-1 motility in a purified system. Biophys J 106:2636-2643.