Actin-microtubule coordination at growing microtubule ends

Abstract

To power dynamic processes in cells, the actin and microtubule cytoskeletons organize into complex structures. Although it is known that cytoskeletal coordination is vital for cell function, the mechanisms by which cross-linking proteins coordinate actin and microtubule activities remain poorly understood. In particular, it is unknown how the distinct mechanical properties of different actin architectures modulate the outcome of actin-microtubule interactions. To address this question, we engineered the protein TipAct, which links growing microtubule ends via end-binding proteins to actin filaments. We show that growing microtubules can be captured and guided by stiff actin bundles, leading to global actin-microtubule alignment. Conversely, growing microtubule ends can transport, stretch and bundle individual actin filaments, thereby globally defining actin filament organization. Our results provide a physical basis to understand actin-microtubule cross-talk, and reveal that a simple cross-linker can enable a mechanical feedback between actin and microtubule organization that is relevant to diverse biological contexts.

Figure 1. GFP–TipAct localizes to growing microtubule ends via EB3 where it interacts with actin filaments, and also guides microtubule growth along actin bundles.

(a) Domain structures of full-length MACF and GFP–TipAct. (b) Kymograph of microtubule growth with mCherry–EB3 and GFP–TipAct (top, Supplementary Movie 1). Fluorescence intensity along the dashed line on the kymograph (bottom). Experimental setups where growing microtubules interact with actin filaments (c) (top), or actin bundles (f) (top), via EB3 and TipAct. Montages showing the representative localization of GFP–TipAct in these experiments (c,f) (bottom). (d) Time series of an experiment as in c (top). Arrowheads show a microtubule growing plus-end that interacts with an actin filament (Supplementary Movie 2). (e) (top), kymograph of the microtubule in d. (e) (bottom), fluorescence intensity along the dashed line on the kymograph. (g) Time series of an experiment as in f (top). (h) (top), kymograph of the microtubule growing along an actin bundle, indicated by arrowheads in g (Supplementary Movie 3). (h) (bottom), fluorescence intensity along the dashed line on the kymograph. In b, e and h, the location of the microtubule seed is indicated above the merge pane. The GFP–TipAct and mCherry–EB3 intensities were normalized by their mean values at the microtubule lattice, and the actin intensity by its mean value (when applicable). Composition in b, c and d: 16 μM tubulin, 100 nM EB3 and 50 nM GFP–TipAct for n=5 experiments; in f and g: 27 μM tubulin, 100 nM EB3, 50 nM GFP–TipAct and 500 nM fascin, for n=8 experiments. Scale bars, 5 μm; time, min:s. ABD, actin-binding domain; MT, microtubule; MtLS, microtubule tip localization signal; MTBD, microtubule-binding domain; CC, coiled-coil.

Figure 2. GFP–TipAct captures and redirects microtubule growth along actin bundles.

(a) Classification of the interaction outcome between growing microtubule plus-ends and actin bundles, and how the intersection angle θ and microtubule length L were defined. (b) Time series of a microtubule grown without GFP–TipAct whose plus-end is mechanically deflected by an actin bundle (Supplementary Movie 5). Analysis of the probability of each interaction outcome as defined in a, as a function of θ, both without (c) and with increasing concentrations of GFP–TipAct (d,e). (f) Analysis of the probability of each interaction outcome as a function of L with (bottom) or without (top) 50 nM GFP–TipAct. (g) Comparison of the compound probability of each outcome for all the concentrations of GFP–TipAct used. For each of the following conditions: 0, 6.25, 12.5, 25 and 50 nM GFP–TipAct, n=708, 459, 421, 443 and 914 interactions were analysed, respectively. (h) Time series of a microtubule grown with GFP–TipAct that zippers onto an actin bundle as it grows (Supplementary Movie 6). Composition: 25 μM tubulin, 100 nM EB3 and 500 nM fascin for n=6 experiments with (h) and n=9 experiments without (b) 50 nM GFP–TipAct. In b and h, white arrowheads indicate the growing plus-ends of the microtubules. Scale bars, 10 μm; time, min:s.

Figure 3. GFP–TipAct allows linear arrays of actin bundles to stably and efficiently template microtubule organization.

Early (t=0) and late (t~40–50 min) time points of microtubule growth within linear arrays of actin bundles with (a), and without (b) GFP–TipAct. (c) Evolution of the distributions of microtubule orientation angle θ_MT, for the experiments in a and b, respectively. Colour gradients indicate the time evolution of the distribution. Each curve is shifted 0.17 a.u. relative to the previous one. Black curves show the time-averaged distribution of actin-bundle orientation angle θ_ACTIN. Evolution of microtubule alignment (d), and microtubule–actin overlap (e), as functions of 〈L_MT〉. Error bars are s.d. The gray-dashed line at 〈L_MT〉=19 μm in d indicates where the trends diverge. Normalized histograms of the difference between microtubule and mean F-actin orientation angle 〈θ_ACTIN〉, for short (f), and long (g), microtubules as defined in d. Data are the average of n=6/9 experiments, comprising 17/9 histograms for long microtubules, and 18/11 histograms for short microtubules, with/without GFP–TipAct, respectively. Error bars are s.d. Dark green and purple lines and insets show fits to equation (1) (see Methods). Data and fits without GFP–TipAct are shifted up 0.8 a.u. Composition of the experiments: 27 μM tubulin, 100 nM EB3 and 500 nM fascin; with (a) and without (b) 50 nM GFP–TipAct. Scale bars, 10 μm; time, h:min:s. MT, microtubule.

Figure 4. GFP–TipAct allows growing microtubules to transport, pull and bundle actin filaments, globally dictating F-actin organization.

(a) Time series of a growing microtubule end that transports a short actin filament (Supplementary Movie 8). Schematic of this effect (bottom). (b) Time series of a growing microtubule that aligns and then pulls on an actin filament (Supplementary Movie 9). The dashed yellow line in the actin pane shows the contour changes of the actin filament. Arrowheads in the merge pane show the microtubule plus-end. Schematic of this effect (bottom). (c) Time series of a growing microtubule that captures actin filaments and forms a bundle, which then guides the growth of another microtubule (Supplementary Movie 11). The dashed yellow line in the actin pane shows the formation of the bundle. White arrowheads and arrows in the merge pane show the growing plus-ends of these two microtubules. Schematic of this effect (bottom). Steady-state F-actin organization in the vicinity of a radial array of dynamic microtubules, with (e), and without (d) GFP–TipAct. Composition in a and b: 16 μM tubulin, 100 nM EB3 and 50 nM GFP–TipAct for n=8 and 9 experiments, respectively; in c: 20 μM tubulin, 100 nM EB3 and 25 nM GFP–TipAct for n=10 experiments; in d and e: 20 μM tubulin, 100 nM EB3, with (e) and without (d) 200 nM GFP–TipAct for n=20 and 5 experiments, respectively. Scale bars, 5 μm (a–c); 10 μm (d,e); time, min:s. MT, microtubule.