Abstract
The position of centrosome, the main microtubule-organizing center (MTOC), is instrumental in the definition of cell polarity. It is defined by the balance of tension and pressure forces in the network of microtubules (MTs). As MTs polymerize against the cell periphery, pressure increases and produces pushing forces on the MTOC. How the mechanical interplay between MTs and the actin network is involved in the regulation of these forces remains poorly understood, in particular because its investigation is technically limited by the structural and biochemical complexity of the cell cytoplasm. Here, in a cell-free assay, we used purified proteins to reconstitute the interaction of an aster of dynamic MTs with actin networks of various compositions and architectures in cell-sized microwells. In the absence of actin filaments, the positioning of the MTOC was highly sensitive to variations in MT length. The presence of a bulk actin filament network limited MTs deformation and displacement, and MTOCs were hold in place. In contrast, the assembly of a dense and branched actin network along the edges of the wells centered the MTOCs by preventing MT slippage and thus maintaining an isotropic balance of pushing forces. In agreement with this, an asymmetric peripheral actin network caused the MTOC to decenter by creating an asymmetry in the pushing forces. Numerical simulations demonstrated that steric hindrance by actin networks, at the tip or along the entire length of MTs, can modulate MTOC positioning, as observed in the experiments. Overall, our results show that actin networks can limit the sensitivity of MTOC positioning to MT length and enforce robust MTOC centering or decentering depending on its architecture.
Introduction
The network of microtubules (MTs) supports the construction of cell body plan and directs its symmetry axes 1–3. The asymmetric organization of MTs directs the preferential orientation of vesicle transport and the position of key sensory organelles and thereby orients the main functions of polarized cells 2,4–6. The centrosome is the main microtubule-organizing center (MTOC), so its position is a key determinant of cell polarity 1,7. The positioning of centrosome, either at the cell center or at the cell periphery, in contact with the plasma membrane, is important for ciliogenesis, immune reactions, cell division, epithelial-to-mesenchymal transition or neuronal development 8–12.
Centrosome position is mainly controlled by combinations of pushing and pulling forces produced in the MT network 13–18. Homogeneous distribution of minus-end directed molecular motors, pulling on MTs in the cytoplasm or at the cell cortex, can enforce centrosome centering 19–22. Heterogeneous distribution of motors, due to local accumulations, can locally increase pulling forces and enforce MTOC off-centering up to the contact with the plasma membrane 23,24. Regulation of centrosome positioning by such pulling forces is robust: centering, in the case of homogeneous distribution of motors, or peripheral positioning, in the case of heterogeneous distribution, are both poorly sensitive to variations of MT length 13. On the opposite, production of pushing forces is much more sensitive to variations of MT length and hence appears as a less reliable positioning mechanism. MT-based pushing is ineffective if MTs are too short to reach the spatial boundaries. Longer MTs may allow centering if their length corresponds precisely to the length of its confining region. However, MTs longer than this critical length will induce an abrupt transition from centering to off-centering 20,25–27.
Actin networks are involved in centrosome positioning and have been proposed to modulate the pushing forces produced by MTs but the underlying mechanisms are obscure 14,18,28–31. There are numerous examples of physical interactions between MTs and actin filaments 32–37. Acting as obstacles, capturing sites or stabilizing sheaths, actin architectures may regulate and organize the spatial distribution of forces in the MT network, by either amplifying or buffering local asymmetries. However, how specific actin architectures, such as dense cortical networks, bundles of linear filaments or cytoplasmic mesh, specifically impact forces production and propagation along MTs, remains poorly understood.
Here, we show that the presence and architecture of actin networks affect the spatial distribution of MTs and the displacements of MTOC either toward or away from the geometrical center of the well. Our results revealed how various actin networks have distinct and specific impact on the distribution of pushing forces and confer some robustness to the mechanism of MTOC positioning by making it less sensitive to variations of MT length.
Results
aMTOC positioning in microwells
It is difficult to directly assess the mechanical effects of actin networks on the production of pushing forces by MTs in living cells because of the biochemical and structural complexities of cell interior, and the presence of motors exerting pulling forces on MTs. Therefore, to reconstitute the interaction of an aster of dynamic MTs with actin networks, we designed an in vitro reconstitution assay using purified proteins and microfabrication techniques.
We investigated how an astral array of MTs self-organizes in a cell-sized compartment. 3D microwell can be used to impose physical barrier to MT growth 20,38. As compared to lipid droplets 25,39, microwells offer the possibility to control the size and shape of the container 40. Actual boundaries are made of lipids in living systems. Therefore, we started by setting up the coating of microwell with a lipid bilayer 41 and further closed the upper surface of the container with a layer of mineral oil (Figure 1A-B, Supplementary Figure 1A-B). Lipids were properly diffusing in the bottom plane and the vertical edges of the microwell (Supplementary Figure 1C). However, there was a problem of tubulin precipitation, as reported previously 42,43. After 30 to 60 minutes of incubation of tubulin in the TicTac or the BRB80 buffer, the tubulin precipitated (Supplementary Figure 2A) and microtubule dynamics stopped. A screening of the biochemical conditions (Supplementary Figure 2B-F) suggested that at lower temperature (22°C) and with high concentration of BSA and GTP, tubulin precipitation could be delayed, up to two hours (Figure 1C). As purified centrosomes generate a quite variable number of MTs 33, we chose to work with artificial MTOCs (aMTOCs) made of short stabilized pieces of MTs grafted on a polystyrene bead (Figure 1D). They efficiently generated 15 to 20 dynamic MTs per aMTOC (Supplementary Figure 3A-D).
We first analyzed the sensitivity of aMTOC positioning to the ratio of MT length over container length by varying tubulin concentrations in microwells of controlled size. As tubulin concentration has been increased from 14 to 26 μM, the average length of MTs varied from 4 to 25 μm (Figure 1D, E). The radius of the microwell was close to 19 μm. Below 18 μM, MTs were shorter than 10 μm and aMTOC adopted a random position (Figure 1F and Supplementary Figure 4A, B). At 18 μM of tubulin, MTs were longer and could reach the microwell boundaries (Figure 1D, E). In about 60 min (Figure 1G, H and Supplementary Movie 1), most aMTOC reached the center of the microwell and remained there (Figure 1F and Supplementary Figure 4B). Centering was also efficient at 22 μM of tubulin (Figure 1F). At 26 μM of tubulin, most MTs were longer than 20 μm (Figure 1D). As they grew, they first ensured a proper centering but after an hour, MT elongation and slippage along microwell edges broke the network symmetry and MTs pushed aMTOC away from the center (Figure 1I, J and Supplementary Movie 2). Taken together, in our experimental system, aMTOC positioning appeared highly sensitive to the tubulin concentration and the MT length. Above a critical concentration of 22 μM, MT elongation and reorientation could bias the distribution of pushing forces and promote aMTOC off-centering, as previously described in water-in-oil droplets 25.
Previous works based on numerical simulations suggested that the friction along the boundaries of the container might prevent the symmetry break by enforcing a vortex-like structure in the network that preserves MTOC centering 13. In cells, the actin filaments form distinct cortical and cytoplasmic networks that might restrict MT lateral translocation and aster displacement 44,45. Therefore, we tested how these various architectures might impact either centering or off-centering mechanisms.
Assembly of various actin architectures in microwells
Rather than working with preassembled and stabilized actin filaments, we chose to grow actin filaments in the microwells in order to control their position and architecture by controlling the mechanism of actin network assembly. Unbranched actin network could be formed in the bulk of the microwell simply by spontaneously assembling 4 μM of actin monomers (Figure 2A, B and Supplementary Figure 5A). Alternatively, to limit the cytoplasmic pool, and favor the assembly of a dense cortical layer, actin filaments were nucleated near the lipid layer by a Nucleation Promoting Factor (NPF) attached to the lipid in the presence of the Arp2/3 complex and actin monomers in the solution (Figure 2C, D and Supplementary Figure 5A, B).
We first tested the impact of these two actin architectures on aMTOC positioning independently of MTs (Figure 2E). Unbranched actin filaments in the bulk had no visible impact on bead position as compared to similar conditions without them (Figure 2E and Supplementary Figure 5C). Position of the beads was a bit centered in the presence of cortical actin, as the thickness of the cortical layer restricted the available space for beads (Figure 2E and Supplementary Figure 5C, D).
These results showed that actin filaments in the bulk and branched cortical network can be reconstituted in 3D microwells and have distinct impact on aMTOC position independently of MTs. The two networks may also have specific effects on MT slippage or deformation and as such distinct impacts on the force distribution in the MT network and therefore on the positioning of the MTOC.
Bulk actin network impairs aMTOC displacement and aster self-centering
aMTOC displacements depend on the production of pushing forces by MT polymerization against container boundaries. However, the morphology of aster confers them a large effective cross-section that limits their displacements by viscous drag. We first tested whether the density of actin network in the bulk could impact the production of pushing forces against effective boundaries, reorganize the spatial distribution of MTs and thus affect the inner balance of force production by MT polymerization.
To test whether bulk actin network could impair the centering process, we worked in conditions where MT length was comparable to the microwell radius (ie at 18μM tubulin). aMTOC position was not dramatically affected by 1 μM of actin but appeared off-centered with 4 μM of actin (Figure 3A, B-left and Supplementary Figure 6A). Indeed, time-lapse imaging revealed that aMTOC remained stuck at their initial position in the presence of 4 μM of actin (Figure 3C). Importantly, at this concentration, unbranched actin network had no impact on MT elongation (Figure 3D). In addition, higher concentrations of tubulin, 26 μM, although capable of promoting MT elongation, could not overcome aster immobilization by the bulk actin meshwork (Figure 3B-right and Supplementary Figure 6B). This suggested that the defective centering was due to friction resisting aster displacement rather than steric effects blocking MT polymerization 32,33. Indeed, implementing physical hindrance to aster displacement by taking into account steric effect of actin filaments along MT aster in numerical simulations was sufficient to account for the immobilization of MTOC by linear bulk actin filaments (Figure 3E-G, Supplementary Figure 6C, D and Supplementary Movie 3). Therefore, we conclude that the presence of a dense network of actin filaments in the bulk can affect MT aster centering by resisting aster translocation rather than impairing MT elongation.
These data also suggested that restricting the actin network to the periphery might specifically limit MT slippage without impairing aster translocation.
Cortical branched actin meshwork favors aster centering
To test whether cortical actin could counteract MT slippage and MTOC off-centering, experiments were performed in the presence of long MTs (ie at 26μM tubulin) (Figure 4A). As described above, the cortical network clustered aMTOC in a smaller volume and thus induced a partial centering (Figure 2E-right, 4B-left). Interestingly, this centering was significantly improved by the growth of long MTs (Figure 4B-left and Supplementary Figure 7A). MT length was not significantly changed even in the absence or presence of cortical actin, suggesting that this effect is not due to interference of actin with MT elongation (Figure 4C). Instead, MTs appeared longer than the radius of the microwell and the network adopted a vortex-like structure (Figure 4D). Time-lapse imaging showed that in the absence of cortical actin, MTs were pivoting around the aMTOC, whereas they maintained their orientation and grew along the edge in the presence of cortical actin (Figure 4E, F and Supplementary Movie 4). MT pivoting in the absence of cortical actin appeared associated with aMTOC off-centering, whereas MT sneaking into the cortical actin was associated with aMTOC stable centering and maintenance at the center (Figure 4G and H). Numerical simulations in which friction was restricted to the cell periphery displayed similar aMTOC displacements and final positions (Figure 4I-K, Supplementary Figure 7B and Supplementary Movie 5), demonstrating that local steric interactions between cortical actin and MTs are indeed sufficient to prevent MTOC off-centering. From these results, we concluded that cortical actin network can counteract the effect of MT elongation and aster off-centering by restricting MT slippage and thus maintaining a regular distribution of force application sites along the cortex and around the aMTOC. In addition, even at lower concentrations of tubulin (18 μM), MT asters displayed a proper centering mechanism in the presence of cortical actin (Figure 4B-right, and Supplementary Figure 7C). This showed that cortical actin can enforce a robust centering that is less sensitive to MT length.
These results also suggested that a heterogeneous pattern of cortical friction might create an asymmetry in the angular distributions of MTs, and an alignment of pushing forces leading to aMTOC off-centering.
Asymmetric cortical actin meshwork induces aster off-centering
We reasoned that with lower actin filament density, which would be crosslinked to each other, we could enforce the asymmetry of the actin network growing from the walls of the microwells 46. Indeed, we found that with 0.5 μM instead of 2 μM of actin and 100 nM of α-actinin, the cortical network grew from all edges but formed an asymmetrically positioned ring (Figure 5A and Supplementary Figure 8A, B). The growth of MTs from the asters did not seem to have an impact on the asymmetric architecture of the actin network (Figure 5B). To analyze aster positioning with respect to this asymmetry, images were reoriented in order to align horizontally the center of the actin inner zone and the center of the microwell (Figure 5C, D and Supplementary Figure 8C, D). To test the potential guiding effect of asymmetric cortical actin networks, we worked in off-centering conditions, ie 26 μM tubulin, in which aMTOCs are randomly distributed in the microwell in the absence of actin (Figure 5E and Supplementary Figure 8E, F). By its thickness, the cortical actin network constrained asters positioning and limits aMTOC dispersion even in the absence of MTs (Figure 5F and Supplementary Figure 8E, F). However, as anticipated, the reorientation of the microwells with respect to the asymmetry of the actin network revealed that as MTs grew from the asters they shifted the aMTOCs toward the center of the actin inner zone (Figure 5G, H, I, Supplementary Figure 8E, F and Supplementary Movie 6). Interestingly, in conditions imposing shorter MTs and an efficient centering of the aster, ie 18μM of tubulin, the asters in the presence of the asymmetric actin network appeared off-centered toward the center of the actin inner zone as well (Figure 5J and Supplementary Figure 8G). Numerical simulations confirmed that a heterogeneous friction pattern due to variable thickness in the cortical actin network along microwell boundary was sufficient to push the MTOC away from the thicker actin layer and thus promote aster off-centering (Figure 5K-M, Supplementary Figure 8H and Supplementary Movie 7). Overall, these results showed that the cortical actin network architecture can direct the position of MTOCs, either at the center or away from it, depending on its heterogeneity. It controls the force balance at the MTOC by modulating the pattern of friction resisting MT slippage and thus directing the localization of the sites of application of pushing forces.
Discussion
Our results suggest that actin networks make aster positioning both more robust and more versatile. Indeed, in the absence of actin filaments, asters displayed abrupt transitions from centering to off-centering depending on MTs length. By contrast, bulk actin filaments resisted asters displacements and MTOC were hold in place, independently of the presence of MTs. Moreover, cortical actin networks specifically favored aster centering over a broad range of MT lengths. The presence of asymmetric actin resulted in off-centering of the aster. From these observations, we propose that actin networks can modulate the sensitivity of MT aster positioning to variation of MT length.
It has been reported that the presence of cytoplasmic actin network can affect MT organization 45,47. The immobilization of MTOC by bulk actin filaments in our system is reminiscent of the regulation of MT aster in Xenopus extract and sea urchin embryo 32,48. The robust polarization of MT network and off-centering of MTOC by asymmetric organization of the cortical actin is reminiscent of the mechanism supporting centrosome displacement to cell periphery during ciliogenesis and immune synapse formation 8,11,49. Thus, our reconstitution assay based on a minimal set of components might properly account for the actual mechanism regulating MT aster positioning by pushing forces in living cells. However, our experimental conditions and the use of short pieces of MTs attached to a bead did not offer us the possibility to control MT pivoting around the aMTOC, a property that could have promoted symmetry break and amplified MTOC off-centering 13,50,51.
Building a synthetic cell from scratch is a powerful strategy to improve our understanding of cell biology, and pave the way toward new living materials 52. Here, we established a way to combine dynamic MTs and actin filaments in cell-sized confinement. Our findings and techniques could be used as a step toward the reconstitution of the polarization process in synthetic cells.
Materials and Methods
Protein expression and purification
Tubulin was purified from fresh bovine brain by three cycles of temperature-dependent assembly/disassembly in Brinkley Buffer 80 (BRB80: 80 mM Pipes pH 6.8, 1 mM EGTA and 1 mM MgCl2) 53. MAP-free neurotubulin was purified by cation-exchange chromatography (EMD SO, 650 M, Merck) in 50 mM Pipes, pH 6.8, supplemented with 0.2 mM MgC12, and 1 mM EGTA. Fluorescently labelled tubulin (ATTO-488- or ATTO-647-labelled) and biotinylated tubulin were prepared by following previously published method 54. Actin was purified from rabbit skeletal-muscle acetone powder. Monomeric Ca-ATP-actin was purified by gel-filtration chromatography on Sephacryl S-300 at 4°C in G buffer (2 mM Tris–HCl, pH 8.0, 0.2 mM ATP, 0.1 mM CaCl2, 1 mM NaN3 and 0.5 mM dithiothreitol (DTT)). Actin was labelled on lysines with Alexa-568. The Arp2/3 complex, recombinant GST-α-actinin 4 and GST-WA (a truncated version of human WASP) were purified in accordance with previous methods 55,56. Snap-Streptavidin-WA (pETplasmid) was expressed in Rosettas 2 (DE3) pLysS (Merck, 71403). Culture was grown in TB medium supplemented with 30 μg/mL kanamycine and 34 μg/mL chloramphenicol, then 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was added and protein was expressed overnight at 16 °C. Pelleted cells were resuspended in Lysis buffer (20 mM Tris pH8, 500 mM NaCl, 1 mM EDTA, 15 mM Imidazole, 0,1% TritonX100, 5% Glycerol, 1 mM DTT). Following sonication and centrifugation, the clarified extract was loaded on a Ni Sepharose high performance column (GE Healthcare Life Sciences, ref 17526802). Resin was washed with Wash buffer (20 mM Tris pH8, 500 mM NaCl, 1 mM EDTA, 30 mM Imidazole, 1 mM DTT). Protein was eluted with Elution buffer (20 mM Tris pH8, 500 mM NaCl, 1 mM EDTA, 300 mM Imidazole, 1 mM DTT). Purified protein was dialyzed overnight 4°C with storage buffer (20 mM Tris pH8, 150 mM NaCl, 1 mM EDTA, 1 mM DTT), concentrated with Amicon 3KD (Merck, ref UFC900324). Aliquots were flash frozen in liquid nitrogen and stored at −80 °C.
Preparation of an artificial MTOC
To prepare microtubule seeds, the mixture containing 3 μM of fluorescent-labeled tubulin, 7 μM of biotinylated tubulin and 0.5 mM GMPCPP (Jena Bioscience, NU-405S) in BRB80 buffer was incubated at 37°C for 40 min. After the incubation, 10 μM of Taxol was added and the mixture was incubated at room temperature for 10 min. The microtubule seeds were then pelleted by centrifugation at 20,238 x g for 10 min and were resuspended in the BRB80 supplemented with 0.5 mM GMPCPP and 10 μM Taxol. The seeds were flash frozen and stored in liquid nitrogen. To prepare the Neutravidin-coated beads, the polystyrene beads containing surface primary amino groups (PolySciences, 17145-5, Diameter 3 μm) were incubated with 10 mM of Sulfo-NHS-LC-LC-Biotin (ThermoFisher, 21338) at room temperature for 40 min to modify their surface with biotin. The beads were washed with phosphate buffered saline (PBS) and then with HKEM buffer (10 mM HEPES (pH 7.5), 50 mM KCl, 5 mM MgCl2, 1 mM EGTA) supplemented with 0.1% bovine serum albumin (BSA). The beads were incubated with 1 mg/mL of Neutravidin (ThermoFisher, 31000) at 15°C for 30 min or at 4°C for 2 hours. After washing the beads with HKEM buffer supplemented with 0.1% BSA, the beads were resuspended in 200 μL of HKEM buffer supplemented with 0.1% BSA. The beads solution was then mixed with 10 μL of the microtubule seeds. The mixture was incubated under rotation at room temperature overnight. Before mixing the aMTOCs (microtubule seeds + beads) with the reaction mixture containing free tubulin, the solution containing aMTOCs was washed with HKEM supplemented with 0.1% BSA to remove excess seeds and Taxol.
Preparation of small unilamellar vesicles (SUV)
L-α-phosphatidylcholine (EggPC) (Avanti, 840051C), 1,2-distearoyl-sn-glycero-3 phosphoethanolamine-N-[biotinyl(polyethylene glycol)-2000] (DSPE-PEG(2000)-Biotin) (Avanti, 880129C) and ATTO 647N labeled DOPE (ATTO-TEC, AD 647N-161 dehydrated) were used. Lipids were mixed in glass tubes as follows: Type 1 (99% EggPC (10 mg/mL) and 1% DOPE-ATTO390 (1 mg/mL)), Type 2 (98.75% EggPC (10 mg/mL) and 0.25% DSPE-PEG-Biotin (10 mg/mL) and 1% DOPE-ATTO390 (1 mg/mL)), Type 3 (99.5% EggPC (10 mg/mL), 0.25% DSPE-PEG-Biotin (10 mg/mL) and 0.25% DOPE-ATTO647N (1 mg/mL)). The mixture was dried with nitrogen gas. The dried lipids were incubated in a vacuum overnight. After that, the lipids were hydrated in the SUV buffer (10 mM Tris (pH 7.4), 150 mM NaCl, 2 mM CaCl2). The mixture was sonicated on ice. The mixture was then centrifuged for 10 min at 20,238 x g to remove large structures. The supernatants were collected and stored at 4°C. The final concentration of lipids was adjusted to 0.5 mg/mL. The Type 2 SUV was used to bind snap-streptavidin-WA onto the lipid layer. The Type 3 SUV was used to visualize lipids on microwells. In other experiments, the Type 1 SUV was used.
Construction of microwells
The master mold (approximately 20 μm of thickness) was fabricated through photolithography using SU8 3025 (MicroChem) and then it was silanized by incubating with Trichloro (1H,1H,2H,2H-perfluoro-octyl) silane (Sigma, 448931). To make 1st PDMS, the mixture of prepolymer and curing agent (Dow, SYLGARD 184 silicone elastomer kit) was poured onto the master mold. It was baked at 70°C for 2 hours. The 1st PDMS was silanized by incubating with Trichloro (1H,1H,2H,2H-perfluoro-octyl) silane. The 2nd PDMS was made using the silanized 1st PDMS as a template. The 2nd PDMS was cut into small pieces and used as PDMS stamps.
Glasses were cleaned by successive chemical treatments: 30 min in acetone with sonication, 15 min in ethanol (96%), washing ultrapure water, 2 hours in HellmanexIII (2% in water, Hellma), washing in ultrapure water. The glasses were then dried. The slide glasses were oxidized in plasma cleaner (diener) for 2 min at power of 80% and then incubated overnight in a solution of 1 mg/mL of mPEG-Silane (30kDa, PSB-2014, Creative PEG works), 96% ethanol and 0.1%(v/v) HCl. The slide glasses were then dried and stored at 4°C.
To make microwell chip on a cover glass, the PDMS stamp was placed on the cleaned cover glass (20 mm × 20 mm, No.1), facing the pillar surface of the stamp onto the glass. NOA81(Norland Products) drop was put at the side of the PDMS stamp to fill the space between the PDMS pillars with NOA81. The NOA81 was cured by exposing it to UV light (UVKUB2, 100%, 12 min). The PDMS stamp and excess NOA81 were then removed.
Sample preparation
To make a reaction chamber, a cover glass with microwell chip was first oxidized in plasma cleaner (diener) for 2 min at power of 80%. The cover glass with microwell chip was attached onto the silane-PEG coated slide glass with two double-sided tapes (70 μm thickness), facing the side of microwell chip to the slide glass. The SUV solution (0.5 mg/mL) was introduced into the chamber and incubated for 10 min to make a supported lipid bilayer on the surface of microwell chip. It was washed with the SUV buffer to remove excess SUV and then washed with HKEM buffer supplemented with 0.1% BSA. Unless otherwise noted, microtubule and actin assembly were induced by diluting tubulin dimers (20% labelled) and/or actin monomers (10% labelled) in the reaction mixture containing the TicTac buffer (10 mM Hepes, 16 mM Pipes (pH 6.8), 50 mM KCl, 1 mM EGTA, 5 mM MgCl2) supplemented with 5% BSA, 4.4 mM GTP, 2.7 mM ATP, 20 mM DTT, 20 μg/mL catalase, 3 mg/mL glucose, 100 μg/mL glucose oxidase. Microtubule aster formation was induced by adding microtubule seeds coated beads (aMTOCs) into the mixture.
To induce branched actin assembly from the edge of microwells, the SUV solution containing DSPE-PEG-Biotin (Type 2 SUV) was used for lipid coating. Before introducing the reaction mixture into the chamber, HKEM buffer containing 200 nM of snap-streptavidin-WA (used as an NPF) and 0.1% BSA was loaded into the chamber and incubated for 5 min. The excess WA was then removed by perfusing HKEM buffer supplemented with 0.1% BSA. 80 nM of the Arp2/3 complex was added in the reaction mixture. To make asymmetric cortical actin structures, 100 nM of GST-α-actinin 4 was also added.
The reaction mixture was introduced into the chamber immediately after the preparation of the mixture. Mineral oil (Paragon Scientific, RTM13) was then loaded into the chamber in order to close the wells. Unless otherwise noted, the chamber was incubated at room temperature (22-23°C) in order to prevent tubulin precipitation. The final position of the aMTOCs was analyzed at 2 hours after sample preparation. Experiments were repeated to confirm the reproducibility.
For initial screening of biochemical conditions (in Figure 1C and Supplementary Figure 2), TicTac buffer supplemented with 0.1% BSA, 1 mM GTP, 2.7 mM ATP, 20 mM DTT, 20 μg/mL catalase, 3 mg/mL glucose, 100 μg/mL glucose oxidase was used as a control buffer solution. BSA (Sigma, A7030), Polyethylene Glycol (PEG) 3k (Sigma, P3640), PEG 20k (Sigma, 95172), Glycerol (CARLO ERBA), Ficoll400 (Sigma, F9378) and Dextran 40k (Sigma, 31389) were added in the solution at the indicated concentrations. The samples were incubated at the indicated temperatures.
Microscopy
Microtubule asters and actin filaments in microwells were visualized using a confocal spinning-disc system (EclipseTi-E Nikon inverted microscope equipped with a CSUX1-A1 Yokogawa confocal head, an Evolve EMCCD camera (Photometrics), a Nikon CFI Plan-Apo ×60 NA1.4 oil immersion objective, a Nikon CFI S-Fluor ×100 NA1.30 oil immersion objective and a Nikon x20 NA0.75 dry objective). Time-lapse imaging was performed using Metamorph software (Universal Imaging). For time-lapse imaging of microtubule asters in microwells, images were taken every 5 or 10 min to avoid photo-damages of microtubules. Photo-bleaching experiments of lipids were also performed on this system using an iLas2 device.
To measure the position of aMTOC in microwells, the images were acquired using an upright Axioimager M2 Zeiss microscope equipped with an EC Plan-Neofluar dry ×20 NA0.5 dry objective and CoolSNAP EZ (Photometrics) or using a confocal spinning-disk system.
To visualize individual microtubules and actin filaments, an objective-based total internal reflection fluorescence (TIRF) microscopy instrument composed of a Nikon Eclipse Ti, an azimuthal iLas2 TIRF illuminator (Roper Scientific), a ×60 NA1.49 TIRF objective lens, a x100 NA1.49 TIRF objective lens and an Evolve EMCCD camera (Photometrics) was used. This system was also used to visualize tubulin precipitation in microwells. Excitation was achieved using 491, 561 and 642 nm lasers.
Image processing and measurements
To visualize microtubule asters in microwells, the images were processed to improve the signal-to-noise ratio. Background subtraction was performed using Fiji (NIH). To further improve the signal-to-noise ratio, deconvolution was also performed for some of the images (Figure 1E, 3D and 4D) using DeconvolutionLab2 in Fiji before background subtraction. Background subtraction was not performed to show free tubulin signals in Figure 1C and Supplementary Figure 2A-E. Maximum projection was performed to show microtubule asters using Fiji. To visualize actin networks along the vertical edges, maximum projection was performed excluding the bottom and the top images of microwells (partial maximum projection). Microtubule length was manually measured using a 3D distance measurement tool in Fiji. Kymograph and temporal-color coded images were generated using plugins in Fiji.
Measurement of the center (centroid) of wells and the center of aMTOCs was performed using Fiji with bright-field images. Microwells containing only a single aMTOC were analyzed. Distance from the aMTOC to center of the well was measured from each XY coordinates. The actin inner zone (AIZ) was determined by setting thresholds of the fluorescence signals of actin. Then, the center of the AIZ was measured. The angles of the aMTOC and the center of the AIZ relative to the well center were calculated from each XY coordinate. To analyze aMTOC positioning relative to the center of the AIZ, wells were reoriented based on each XY coordinates in order to align the angles from the well center to center of the AIZ (See also Supplementary Figure 8C-D).
Numerical simulations
Simulations were performed using Cytosim software 57. The motion of elastic filaments and solids surrounded by a viscous fluid was calculated using Langevin dynamics 57. The main parameters used in the simulation were presented in Supplementary Table 1.
In the simulations, repulsive steric effects between actin filaments and microtubule aster were considered. As a limitation in the simulation, steric repulsion also occurs between actin filaments and between microtubules at the same steric force. Attractive steric forces between filaments were not included in the simulations. The steric parameters were adapted from the range in previous studies testing steric interactions between microtubules or actin filaments using Cytosim 58,59. Because of excessive computational costs, it was difficult to perform simulations of centrosome positioning with dense actin filaments. Therefore, to reduce computational costs, the effective diameter of microtubules and actin filaments was set to 100 nm 58,59 and the cell size was set to 10 μm in radius. In addition, the total time simulated was set to 150 to 250 seconds. The simulations were performed in the two-dimensional mode.
Bulk actin network was made by adding actin filaments in the cell without fixation of their position. To make actin meshwork near the cell periphery, the actin nucleation factors and branching factors were positioned near the cell periphery (within 7 to 9.2 μm from the cell center). The position of actin nucleation factors was fixed, so that the position of one of the ends of actin filaments was fixed at the initial position. When the actin branching factor binds to an existing actin filament, it nucleates a new actin filament from the existing filament. Asymmetric actin network was made by asymmetrically localizing the actin nucleation factors and the actin branching factors.
Statistics
Statistical tests were performed using R statistical software. Statistical test, sample sizes and P values are described in each figure legend.
Supplementary Movie 1. Centering of MT aster in the absence of actin.
Time-lapse imaging of MT aster formation at 18 μM of tubulin in the absence of actin. Data is also shown in Figure 1G. Scale bar, 10 μm.
Supplementary Movie 2. Off-centering of MT aster in the absence of actin.
Time-lapse imaging of MT aster formation at 26 μM of tubulin in the absence of actin. Data is also shown in Figure 1I. Scale bar, 10 μm.
Supplementary Movie 3. Simulation of MT aster positioning in the absence or presence of bulk actin filaments.
Simulations in the absence (left) or presence (right) of bulk actin filaments. From 25 to 150 sec. MTOC, gray, MT, black, Actin, pink. Data is also shown in Figure 3E.
Supplementary Movie 4. MT aster centering in the presence of cortical actin network. Time-lapse imaging of MT aster formation at 26 μM of tubulin in the absence (left) or presence (right) of cortical actin. Data is also shown in Figure 4E. Scale bar, 10 μm.
Supplementary Movie 5. Simulation of MT aster positioning in the absence or presence of actin filaments near cell periphery.
Simulations in the absence (left) or presence (right) of actin filaments near cell periphery. From 25 to 200 sec. MTOC, gray, MT, black, Actin, pink. Data is also shown in Figure 4I.
Supplementary Movie 6. MT aster positioning in the presence of asymmetric actin network. Time-lapse imaging of MT aster formation at 26 μM of tubulin in the presence of asymmetric actin network. Data is also shown in Figure 5G. Scale bar, 10 μm.
Supplementary Movie 7. Simulation of MT aster positioning in the absence or presence of asymmetric actin network.
Simulations in the absence (left) or presence (right) of asymmetric actin network near cell periphery. From 25 to 250 sec. MTOC, gray, MT, black, Actin, pink. Data is also shown in Figure 5K.
Acknowledgements
This work was supported by the European Research Council (Consolidator Grant 771599 (ICEBERG) to MT and Advanced Grant 741773 (AAA) to LB). S.Y. was supported by fellowships from the EMBO, Astellas Foundation for research on metabolic disorders and Mochida memorial foundation for medical and pharmaceutical research. This project was supported by the MuLife imaging facility, which is funded by GRAL, a program from the Chemistry Biology Health Graduate School of University Grenoble Alpes (ANR-17-EURE-0003).