2.1 Reagents

Purified actin and Arp2/3 complex were purchased from Cytoskeleton, Inc. ATTO 488 actin was purchased from Hypermol. General actin buffer (G-buffer) and actin polymerization buffer (F-buffer) were prepared as directed by Cytoskeleton, Inc. Fascin and hexa-histidine-VCA (His₆-tag VCA) were purified as detailed in Wubshet et al. (2021); 10 mg/mL stock concentration of actin was diluted to 1 mg/mL and further serially diluted as needed to working concentrations using G-buffer + 0.5 mM DTT + 0.2 mM ATP. 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn-glycero-3-[(N-(5-amino-1-carboxypentyl)iminodiacetic acid)succinyl] (nickel salt) (DGS-NTA(Ni)), and cholesterol were purchased from Avanti Polar Lipids. Optiprep, a density gradient medium, was purchased from Sigma Aldrich.

2.2 Micropipette aspiration

The micropipette aspiration setup consists of a pulled and cut micropipette tip mounted on a micropipette holder, a pipette micromanipulator, a liquid reservoir, a filling syringe, and a pressure transducer.

To prepare micropipettes, standard glass capillaries (WPI) were pulled using a Sutter P-87 pipette puller (Sutter Instruments). After pulling a parallel-walled pipette tip, pipettes were manually cut to an acceptable size in the range of 4–8 μm using a heated glass rod. A glass rod is heated using a Bunsen burner and brought to contact with the pulled capillary at the tip. Instantly after contact, the glass rod is pressed through the pipette in a swift motion to result in a clean pipette cut. This is principally similar to cutting pipettes using microforges. Next, pipettes were submerged in a 1% BSA solution to prevent GUV adhesion. Integration of the micropipette aspiration setup, shown in Figure 1a, is as follows: (1) Pipettes are mounted and secured into a pipette holder. (2) A three-way valve connects a fluidic line between the filling syringe, the micropipette, and the fluid reservoir. (3) The reservoir is pneumatically connected to a high-speed pressure clamp (HSPC, ALA Scientific Instruments). The filling syringe, containing a solution osmotically matched to the encapsulated solution in GUVs, is used to fill the micropipette line and reservoir. Elimination of all air bubbles is vital for performing micropipette aspiration experiments. Once the sample is ready for imaging and aspiration, a slight positive pressure should be maintained either by keeping the reservoir at a slightly higher altitude or applying a minimal positive pressure using the pressure transducer. This should be maintained until the GUV of interest is located under the microscope. Finally, using a micromanipulator, the pipette will be brought to close proximity. This should result in a slight pushing of the GUV from the positive pressure coming from the pipette tip. Once GUV is located, the pressure transducer is used to apply an appropriate negative pressure to induce aspiration of the GUV.

2.3 Actin encapsulation inside GUVs

First, inner solutions of the desired actin/actin-binding protein composition were prepared. To reconstitute fascin-bundled actin networks inside GUVs, fluorescent filamentous actin was reconstituted by incubating 5.3 μM actin, 0.53 μM of ATTO 488 actin in F-buffer supplemented with 3 mM ATP on ice for 15 min. Next, fluorescent filamentous actin was brought to room temperature, and fascin, at a 1:5 ratio to actin, was added to the solution to initiate the formation of bundles. To reconstitute actin cortex into GUVs, filamentous actin was prepared identical to the above followed by the addition of 500 nM Arp2/3 and 500 nM His₆-tag VCA; 7.5% Optiprep is included in the final volume of inner solutions to facilitate GUV sedimentation as a result of a density difference between inner and outer solutions. After the addition of crosslinker proteins, the inner solution is immediately emulsified in a lipid/oil mixture via pipetting the solution up and down to create volumetric confinement to actin networks before reaching full network assembly. Droplet emulsions are then dispensed in the rotating cDICE chamber to form GUVs. Prior to the addition of emulsified inner solution, in a rotating cDICE chamber, we add an outer aqueous solution (similar osmolarity to the inner solution of ~200 mOsm) and an oil/lipid mixture. These two solutions are planarly layered due to centrifugal forces. Oil phase in our oil/lipid mixture contains 20/80% (v/v) mineral oil/silicon oil, and our standard lipid composition uses DOPC/cholesterol at 70/30 mol/mol. Details of actin encapsulation inside GUVs are as described in Bashirzadeh et al. (2021).

2.4 Imaging and analysis

Micropipette aspiration of actin GUVs (of a diameter of ~15–40 μm) was observed using Olympus IX-81 inverted microscope equipped with a spinning disk confocal (Yokogawa CSU-X1), OBIS LS/LX lasers (Coherent), and an iXON3 EMCCD camera (Andor Technology). Metamorph was used to control the above components and acquire images; 40×/1.3 NA objective was used for imaging samples. Fluorescence images of actin networks were acquired to visualize networks by exciting ATTO 488 actin at 200 ms exposure, while low-light brightfield images were simultaneously taken to capture and track the micropipette tip as it was operated using a micromanipulator. To record actin dynamics in response to changes in aspiration pressure, timelapse images were taken every 300 ms. Images were processed using ImageJ. After acquiring line-scan data from ImageJ, using Python's matplotlib library, we plotted intensity profiles and bar graphs. Statistical analysis between peak means of each two groups, pre- and post-aspiration of bundles and branched bundles, was done using Student's t-test with p = 0.05 as the level of significance.

3.1 Actin filaments inside GUVs display resistance to aspiration

To validate our micropipette aspiration setup (Figure 1a), empty 70% DOPC/30% cholesterol GUVs with a trace amount of fluorescent lipid were aspirated at incremental pressure points (Figure 1b). After filling all fluidic tubings with glucose solution, a net positive pressure is applied at the tip of the micropipette by slightly raising the reservoir with respect to the tip to assure air bubbles are not drawn into the tip when the micropipette tip is not submerged into the sample solution containing GUVs (Figure 1a). Once the micropipette is positioned, using the micromanipulator, in the correct frame and plane where GUVs are, we use the motorized microscope stage to locate a GUV of interest with respect to the micropipette tip, which is in a fixed frame. Once aspirated, GUVs instantly seal the pipette and protrude. At each pressure increment, we observed aspirated protrusions elongating (Figure 1b). We can also see that there is not a visible decrease in the GUV size outside of the pipette despite an increase in elongation of protrusion. This is consistent with prior observations (Olbrich et al., 2000) and is attributed to area dilation due to the elastic stretching of the GUV lipid bilayer. Following validation of the micropipette aspiration setup, we investigated the mechanical response of GUVs encapsulating actin filaments.

As the major determinants of cellular mechanics, we became interested in how actin filaments encapsulated inside GUVs respond to loads by using micropipette aspiration. We reconstituted actin filaments by polymerizing 5.3 μM globular actin using F-buffer. Similar to empty GUVs, actin filaments were encapsulated in DOPC/cholesterol (70/30) GUVs. When unaspirated, fluorescence images of F-actin GUVs display uniform distribution of filaments throughout the GUV lumen (Figure 2 top). Interestingly, we saw that this symmetry/uniformity breaks when actin-GUVs are aspirated (Figure 2 bottom). The actin density is lower throughout the protrusion compared to the GUV lumen outside of the pipette. Figure 2b displays line scan profiles of crossecting lines similarly shown in Figure 2a pre- and post-aspiration, showing the diminished F-actin fluorescence in the micropipette post-aspiration. Actin filaments are flexible polymers with lengths ranging from submicron to a few microns in length. Our findings suggest that polymerized filaments at 5.3 μM behave in a connected manner, possibly through entanglement (Gardel et al., 2003), to resist aspiration and remain outside of pipette constriction. Transforming actin filaments into more complex networks, for example via branching, could potentially increase network resistance thus diminishing actin fluorescence in the micropipette.

3.2 Fascin-bundled actin aligns along the axis of the flow in response to aspiration

Following our observation that entangled actin filaments resist entering the micropipette upon aspiration, we continued to investigate how other actin architectures reorganize in response to aspiration-induced loading. First, we encapsulated fascin-bundled actin networks inside GUVs as illustrated in the schematic in Figure 3a. We reconstituted actin filaments by polymerizing 5.3 μM actin using F-buffer followed by adding fascin at 1:5 ratio to actin to bundle actin filaments. Encapsulation using the modified cDICE technique was done promptly before the complete formation of the bundled network. As has been shown previously (Bashirzadeh et al., 2020; Wubshet et al., 2021), fascin, at 1:5 ratio to actin, assembles stiff protrusive bundles in random orientations. These protrusive bundles do not reorient with respect to the GUV via diffusion but rather maintain a fixed architecture. Often, there is a dominant bundle that highly deforms the GUV resulting in the longest axis of the GUV, which we refer to as the dominant bundle. We observed, upon applying negative pressure via a micropipette of a GUV initially at a distance from the pipette tip, the GUV as a whole reoriented in a manner where the dominant actin bundle aligned with the axis of flow as it moved near the tip of the pipette. We show, in Figure S1, the reorientation of GUV with respect to the dominant bundle. Hydrodynamically, this alignment is energetically favorable to minimize drag against fluidic current in a manner where the smallest cross-sectional area is perpendicular to the current, akin to how ideal shapes of water transport systems such as boats and ships do not sail on their side.

For GUVs that have a network of fascin-bundled actin, we observed dynamic reorganization under micropipette aspiration (Figure 3b top). Once aspiration began, at the first instance of pressure increment, fascin-bundled networks almost instantly reorganized and started to align by converging at the pipette tip (Figure 3b), and the opposite ends of bundles that were in the unaspirated region of the GUV remained oriented in random directions. We call this the alignment initiation stage. Such an arrangement is reminiscent of a matcha whisker. Fascin GUVs equilibrate in this arrangement exhibiting resistance for further insertion into the micropipette. Further increasing aspiration pressure, however, begins to align the fascin bundles along the axis of aspiration, in part due to energetically unfavorable bent fascin-bundled filaments. At the final aspiration pressure of 533 Pa, initially randomly reoriented bundles, were fully aligned into one thick bundle fully constricted by the walls of the capillary. Line scans of bundle networks pre-aspiration show random reorientation of bundles as indicated by the multipeak profile (Figure 3c top left) and alignment of bundles at the final stage of aspiration is shown by the single-peak line profiles (Figure 3c top right). Figure 3d presents a statistical analysis of multiple GUVs where we compared peak-counts in the intensity profile plots which represent the number of resolved bundles pre- and post-aspiration. Note that these may not represent the true number of bundles in a strict sense as we are interested in thick bundles that protrude the GUVs or provide major resistance to alignment that can be easily identified as an intensity peak. Our initial mechanistic hypothesis for this observation was that these bundles were not crosslinked to one another, and the increase in membrane tension exerts forces on the protruded bundles may facilitate bundles to slide and rearrange.

To support this idea, we performed an experiment where we encapsulated bundles that are branched (i.e., the bundles are physically linked to one another). We achieved this architecture by co-encapsulating fascin with Arp2/3 complex along with its activator VCA. From our prior work, we have demonstrated that co-encapsulation of Arp2/3 with fascin suppresses protrusive bundles via branching and shortening fascin bundles. When aspirated, branched bundled actin GUVs (Figure 3b bottom) exhibited an entirely different response where there was no observed alignment of bundles. Since these bundles are not protrusive, membrane forces are not directly coupled to the bundles to rearrange them. These architectures continued to resist inserting into the micropipette until the aspiration exceeded the area dilation threshold resulting in GUV rupture. Line scans support our observation where multipeak profiles pre-aspiration (Figure 3c bottom left) remain at the final stage of aspiration (Figure 3c bottom right). A comparison of the number of bundles along the crossecting line shows that bundles converged from multi-peak pre-aspiration to a single-peak aligned bundle post-aspiration (Figure 3d). In contrast, branched bundles showed no difference between pre- and post-aspiration conditions. This suggests that physical crosslinking between bundles and their interaction with membranes have a strong influence on the reorganization in response to aspiration load.