Complementary mesoscale dynamics of spectrin and acto-myosin shape membrane territories during mechanoresponse

The spectrin-based membrane skeleton is a major component of the cell cortex. While expressed by all metazoans, its dynamic interactions with the other cortex components, including the plasma membrane or the acto-myosin cytoskeleton, are poorly understood. Here, we investigate how spectrin re-organizes spatially and dynamically under the membrane during changes in cell mechanics. We find spectrin and acto-myosin to be spatially distinct but cooperating during mechanical challenges, such as cell adhesion and contraction, or compression, stretch and osmolarity fluctuations, creating a cohesive cortex supporting the plasma membrane. Actin territories control protrusions and contractile structures while spectrin territories concentrate in retractile zones and low-actin density/inter-contractile regions, acting as a fence that organize membrane trafficking events. We unveil here the existence of a dynamic interplay between acto-myosin and spectrin necessary to support a mesoscale organization of the lipid bilayer into spatially-confined cortical territories during cell mechanoresponse.


Supplementary Figure 3. βII-Spectrin organization in P1; βII-Spectrin/area during spreading; βII-Spectrin and Actin recovery after Latrunculin A washout in cortex-mimicry zones
A-B) MEFs fixed during P1, immunolabelled for endogenous F-actin (magenta) and βII-spectrin (green, endogenous and GFP-tagged), and analyzed by 3D confocal microscopy. Optical sectioning is optimized to resolve the cortex on the cell dorsal plane during P1, as shown in the cartoon (scale bar: 20 µm). C-D) Cells fixed at different time points after seeding (between 5-20 minutes) and immunolabelled for endogenous βIIspectrin and F-actin. Projected cell area and fluorescence integrated intensities in TIRFM for the two proteins are reported, displaying linear correlation (n=376 cells, see Table 1). E-F) Total data point distribution of the graphs shown in Figure 2E and 2G, outliers were excluded from the analysis (threshold 0.0007; untreated: n=9 cells; blebbistatin: n=5 cells). Both analysis showed normal gaussian distribution between the physiological speed range of -0.1 µm sec -1 and +200 µm sec -1 . F) MEFs seeded on micropatterned fibronectincoated lines. TIRFM images of cells fixed during the washout phase after latrunculin A treatment are shown (green: βII-spectrin, magenta: F-actin, red: fibronectin. Scale bar: 20 µm). DAPI (blue) is visualized in EPI mode to discriminate intact cells from debris. The white dashed box (1) is zoomed to highlight peculiar actin nodes formation in non-adhesive cell cortex. All images are representative of many cells immunolabelled in n=2 or more independent experiments.

Supplementary Figure 4. βII-Spectrin perturbation by cytoskeletal impairing drugs
A) GFP-βII-spectrin expressing MEFs imaged by live TIRFM during administration of cytoskeletal impairing drugs are shown before (-) and during (+) the treatments (scale bar: 20 µm). Whole-cell mean fluorescence intensities are normalized to the pre-treatment frames (blue circles), and plotted in B at 5 minutes (red circles) and 30 minutes (green circles) of treatment (n=12 cells, data are presented as mean values ± SD, oneway Anova with multiple comparisons: ****p<0.0001, **p=0.0095). C) Four ROIs (20x20 μm) related to drug treatments with no apparent fluctuations in GFP-βII-spectrin intensities are presented (ROIs related to cells in panel A treated with latrunculin A and blebbistatin). This ROIs showed differential GFP-βII-spectrin reaction depending on different location. The normalized intensities across the entire projected cell area underestimate the mesoscale differential behavior of the meshwork, mean intensities normalized to the untreated frame are reported on the top-left corner of each image. D) FRAP single cell analysis of mobile fractions: the same graph presented in Figure 4 F is reported here highlighting the difference in mobile fraction within the same cell (connected by black lines) before (blue) and 5' after treatment (red) (Statistical analysis: paired Student's t-test two-tailed, data are presented as mean values ± SD, * p=0.0413, ** p=0.0021, **** p<0.0001, n=12 cells).

Supplementary Figure 5. βII-Spectrin FL, ∆PE/ANKbs and ∆ABD accumulate at negative curvatures that spontaneously form during cell spreading/polarization
A-C-E) Representative images of spontaneous retractile events observed in MEFs expressing GFP-βII-spectrin variants during cell spreading by live TIRFM (GFP-βII-spectrin variants in green and RFP-Actin in magenta, scale bar: 20 μm). Relevant events are highlighted by the dashed boxes and zoomed in the lower panels: protruding zones are indicated by white arrowheads, retracting zones by white arrows. Line scan analysis of arrows with circular ends (1-2-3) are reported in B-D-F for both proteins, directionality is reported above the graphs. G) Cartoon model of Actin/βII-spectrin opposite polarity exploited by protrusion/retraction events during the polarization phase of cell spreading. Images are representative of many cells in n=3 or more independent experiments.

Supplementary Figure 7. βII-Spectrin reactions to osmotic changes: global versus local behavior
A) Representative images of GFP-βII-spectrin (green) and PM-marker (magenta) transfected MEFs observed by live TIRFM during osmotic shocks. Five isotonic (1x)-to-hypotonic (0.5x) cycles were applied (B). Initial fluorescent signals were normalized to obtain the non-stoichiometric ratio βII-spectrin/PM (LUT fire, scale bar: 10 μm). Average ratio is plotted in B (black line, n=3 cells, mean ± SD), while the ratio of the cell in A is shown by the orange line. C-D) As positive control, the same protocol was applied to MEFs transfected with soluble GFP (green) and PM-marker (magenta), while GFP/PM ratio (LUT fire, scale bar: 10 μm) is plotted in D (n=3 cells, data are presented as mean values ± SD). Zonal ratio analysis at two extreme cases is reported in E: ROI1 presented high βII-spectrin/PM ratio, while ROI2 displayed lower ratio. As show in the graph (F), the two ROIs behave differently: while ROI1 reacted similarly to the whole-cell analysis presented in B, ROI2 showed an initial decrease of the ratio sustained during the first two iso-to-hypotonic cycles, followed by a compensatory effect that restored the initial ratio during the last four cycles. A similar effect in a different PM zone is presented in G: lamellipodia (dashed box and zoomed in the bottom panel) characterized by high actin and low βII-spectrin content were blocked during hypotonic shocks. Kymograph generated across the dashed yellow line (3): βII-spectrin/PM ratio was low in lamellipodia (* asterisk) compared to the adjacent cell body (** asterisks), and lamellipodia blockage during the hypotonic shock was observed. H) The same cell presented in Figure 7 A is shown: endogenous βII-spectrin (green), clathrin heavy chain (CHC, magenta) and F-actin (blue), imaged by TIRFM (scale bar: 10 µm, images are representative of many cells immunolabelled in n=2 or more independent experiments).

Supplementary Figure 8. Schemes of the patented devices implemented in this study
Exploded-view drawings of the cell stretching dish (A) and the cell compression device (B), where the main components and assembly order are highlighted.