Figure 1.
Domain Organization and Alignment of Spastin and Katanin.
The figure shows a sequence alignment of domains of Drosophila and human spastin, as well as human katanin. The black contour highlights the position of the human spastin linker, residues with a high degree of conservation are red, residues with low degree are blue. The location of the highly basic patch in human spastin is boxed in blue. Above the alignment, the location of the sequence in the context of human spastin is indicated, below the location in human katanin. For details, see text.
Figure 2.
Diffusion of Spastin along Microtubules.
Panel A shows an image sequence of GFP-Δ227 HsSpastin (green) on a microtubule (red). The scale bar is 1 µm, the time between two frames was 1 s (for analysis, the full time resolution was used; see Supporting Movie S1). Panel B shows the distribution of diffusion coefficients of 30 molecules. Each film sequence was analyzed frame by frame, and the x-y position of the molecule of interest was localized in each frame (see Material and Methods). The mean-squared displacements were calculated and plotted against time. These traces were used to calculate the 1-dimensional diffusion coefficient over a sliding window (size = 5 frames). The distribution of all local diffusion coefficients of all molecules is plotted in the histogram. Panel C plots a histogram of distances from frame to frame with a Gaussian fit.
Figure 3.
Diffusion of Katanin along Microtubules.
Panel A shows an image sequence of GFP-FLAG HsKatanin (green) on a microtubule (red). The scale bar is 5 µm, the time between two frames was approximately 1 s (for analysis, the full time resolution was used; see Supporting Movie S2). Panel B shows the distribution of diffusion coefficients of 30 molecules, analyzed as in Figure 2. The distribution of all local diffusion coefficients of all molecules is plotted in the histogram. Panel C plots a histogram of distances from frame to frame with a Gaussian fit.
Table 1.
Nucleotide dependence of microtubule interaction.
Table 2.
Salt dependence of microtubule interaction.
Figure 4.
The figure shows dwell-time distributions of GFP Δ227 HsSpastin on microtubules at different ionic strengths. The histograms contain insets showing example kymographs of binding events. The dotted lines are mono-exponential curve fits that were used to calculate the lifetime, τ.
Figure 5.
The figure shows dwell-time distributions of GFP-HsKatanin on microtubules at different ionic strengths. The histograms contain insets showing example kymographs of binding events. The dotted lines are mono-exponential curve fits that were used to calculate the lifetime, τ.
Figure 6.
Spastin Interaction with Subtilisin-treated Microtubules.
Panel A shows negative stain electron micrographs of microtubules and the spastin E442Q mutant in the presence of 1 mM ATP. In the top part, subtilisin-treated microtubules were used, the lower part shows native microtubules. Binding and bundling occurs only with native microtubules. Panel B shows a SDS-gel of tubulin before and after subtilisin-proteolysis. Panel C shows TIRF microscopy images of Alexa Fluor 555 microtubules (red) and GFP Δ227 HsSpastin E442Q (green). The GFP spastin construct decorates and bundles only untreated microtubules.
Figure 7.
Katanin Interaction with Subtilisin-treated Microtubules.
Panel A shows fluorescence microscopy images of GFP HsKatanin (green) and microtubules. The fluorescent, red parts of microtubules are subtilisin-treated, native stretches are unlabeled. Panel B shows a comparison of binding curves obtained by co-sedimentation of HsKatanin and microtubules. The fraction of bound enzyme was determined by SDS-PAGE of supernatants and pellets. The lines are connecting the data points without a specific model.
Figure 8.
Microtubule Binding Properties of the Spastin Linker (MTBD) Domain.
Panel A shows the results of a co-sedimentation assay of microtubules and GFP-linker domain constructs. From left to right: GFP-HsSpastin wild type linker domain (approximately 34 kDa) was mixed with microtubules (50 kDa) at a stoichiometric ratio of 1∶1. The microtubule-bound fraction appears in the pellet (P) of ultracentrifugation, the unbound fraction in the supernatant (S). Two amounts were analyzed by SDS-PAGE. Lanes 7–10 show the same experiment with the triple K310KK>Q310QQ mutation. The right SDS-gel shows a co-sedimentation assay with the analogous region from D. melanogaster spastin. Weight markers (lanes 1, 6, 11 left gel, and lanes 1 and 6 right gel) are given in kDa. Panel B: Binding of the GFP-linker constructs to microtubules in a TIRF microscopy assay.The top row shows Alexa Fluor 555 microtubules in red, the middle row GFP-linker constructs in green, the bottom row an overlay of red and green channels. All images were taken at the same gain and exposure time. The images show that the GFP HsSpastin wild type linker (left column) coats all microtubules densely, while the mutant (center column) and the Drosophila (right column) constructs fail to bind to a significant degree.
Figure 9.
Domain-Mapping of the Katanin-Microtubule Interaction.
Panel A displays the constructs used for katanin binding experiments. Panel B shows a quantitative SDS-gel of supernatants (unbound) and pellets (microtubule-bound) of an in vitro binding assay of truncated katanin constructs and microtubules. Increasing microtubule concentrations (0 to 10 µM) were incubated with a fixed katanin construct concentration (1 µM). The density of the katanin construct band was plotted against the microtubule concentration and fitted to a Hill function (panel C). The half-maximal saturation was reached at 0.34 µM (Kat12) and 0.40 µM (Kat2). Only constructs containing domain 2 were able to bind to microtubules.
Figure 10.
Nucleotide-Dependence of the Spastin-Microtubule Interaction.
Panel A shows a SDS-gel of co-sedimentation assays with spastin (E442Q mutant, 1 mM ATP) and a constant concentration of microtubules (1 µM; indicated by a dotted line). With increasing spastin concentrations, an increasing amount of protein is co-sedimented with microtubules. Panel B: Densitometric analysis allowed plotting of the spastin concentrations bound to microtubules against free spastin concentrations. The horizontal dotted line indicates the concentration of tubulin used in the assays. The diagonal dashed lines connect points with the same total concentration of spastin. The concentrations increase towards the upper right. Steeper curves reflect a higher affinity. Panel C shows the same experiment for wild type spastin.
Figure 11.
Nucleotide-Dependence of the Katanin-Microtubule Interaction.
Panel A shows a SDS-gel of co-sedimentation assays with katanin (E309Q mutant, 1 mM ATP) and a constant concentration of microtubules (2 µM; indicated by a dotted line). With increasing katanin concentrations, an increasing amount of protein is co-sedimented with microtubules. Panel B: Plot of the densitometric analysis as in Figure 10. Panel C shows the same experiment for wild type katanin.
Table 3.
Binding stoichiometry.
Figure 12.
Electron Micrographs and SDS-PAGE of the Coiled Coil Spastin Construct.
Panel A shows a SDS-gel with the analysis of an in vitro severing assay (left part) and an oligomerization assay. Lanes 1 and 11 are weight markers, lanes 2 to 9 results of the severing assay. The coiled coil spastin wild type construct was incubated with microtubules, and centrifuged after 15 min. Active spastin disassembles microtubules completely under these conditions, such that tubulin appears in the supernatant. Otherwise, microtubules are found in the pellet. Only the wild type coiled coil construct is active in the presence of ATP. The right part of the gel (lanes 12 and 13) shows the coiled coil wild type construct under reducing conditions (lane 12, 1 mM mercapto-ethanol) and oxidizing conditions (lane 13; omission of reducing agents). The coiled coil construct migrates as a dimer, due to a disulphide bridge introduced in the coiled coil domain. Panels B and C show electron micrographs of the E442Q mutant coiled coil construct without ATP, and with 2 mM ATP. Insets show particle averages of the predominant species. The averages were obtained with a semi-automated procedure implemented in EMAN2.
Figure 13.
Microtubule-Dependence of the ATPase Activity of the Coiled Coil Spastin Construct.
The figure shows the steady state ATPase turnover of wild type Δ227 HsSpastin (data from [8]) and the coiled coil construct. The activity of the reference construct is stimulated by microtubules according to a Hill curve (black line), the coiled coil construct lacks this feature. The blue curve is an interpolation of the data points and does not reflect a specific kinetic model.