Enantiomeric Aβ peptides inhibit the fluid shear stress response of PIEZO1

Traumatic brain injury (TBI) elevates Abeta (Aβ) peptides in the brain and cerebral spinal fluid. Aβ peptides are amphipathic molecules that can modulate membrane mechanics. Because the mechanosensitive cation channel PIEZO1 is gated by membrane tension and curvature, it prompted us to test the effects of Aβ on PIEZO1. Using precision fluid shear stress as a stimulus, we found that Aβ monomers inhibit PIEZO1 at femtomolar to picomolar concentrations. The Aβ oligomers proved much less potent. The effect of Aβs on Piezo gating did not involve peptide-protein interactions since the D and L enantiomers had similar effects. Incubating a fluorescent derivative of Aβ and a fluorescently tagged PIEZO1, we showed that Aβ can colocalize with PIEZO1, suggesting that they both had an affinity for particular regions of the bilayer. To better understand the PIEZO1 inhibitory effects of Aβ, we examined their effect on wound healing. We observed that over-expression of PIEZO1 in HEK293 cells increased cell migration velocity ~10-fold, and both enantiomeric Aβ peptides and GsMTx4 independently inhibited migration, demonstrating involvement of PIEZO1 in cell motility. As part of the motility study we examined the correlation of PIEZO1 function with tension in the cytoskeleton using a genetically encoded fluorescent stress probe. Aβ peptides increased resting stress in F-actin, and is correlated with Aβ block of PIEZO1-mediated Ca2+ influx. Aβ inhibition of PIEZO1 in the absence of stereospecific peptide-protein interactions shows that Aβ peptides modulate both cell membrane and cytoskeletal mechanics to control PIEZO1-triggered Ca2+ influx.


Traumatic brain injury (TBI) elevates Abeta (Aβ) peptides in the brain and cerebral spinal fluid.
Aβ peptides are amphipathic molecules that can modulate membrane mechanics. Because the mechanosensitive cation channel PIEZO1 is gated by membrane tension and curvature, it prompted us to test the effects of Aβ on PIEZO1. Using precision fluid shear stress as a stimulus, we found that Aβ monomers inhibit PIEZO1 at femtomolar to picomolar concentrations. The Aβ oligomers proved much less potent. The effect of Aβs on Piezo gating did not involve peptide-protein interactions since the D and L enantiomers had similar effects. Incubating a fluorescent derivative of Aβ and a fluorescently tagged PIEZO1, we showed that Aβ can colocalize with PIEZO1, suggesting that they both had an affinity for particular regions of the bilayer. To better understand the PIEZO1 inhibitory effects of Aβ, we examined their effect on wound healing. We observed that over-expression of PIEZO1 in HEK293 cells increased cell migration velocity ~10-fold, and both enantiomeric Aβ peptides and GsMTx4 independently inhibited migration, demonstrating involvement of PIEZO1 in cell motility. As part of the motility study we examined the correlation of PIEZO1 function with tension in the cytoskeleton using a genetically encoded fluorescent stress probe. Aβ peptides increased resting stress in F-actin, and is correlated with Aβ block of PIEZO1-mediated Ca 2+ influx. Aβ inhibition of PIEZO1 in the absence of stereospecific peptide-protein interactions shows that Aβ peptides modulate both cell membrane and cytoskeletal mechanics to control PIEZO1-triggered Ca 2+ influx.
There is a correlation between traumatic brain injury (TBI), Amyloid β peptide levels (Aβ) and the onset of TBI related diseases 1 . The relationship between Aβ and clinical manifestations of TBI is not understood 1 . Since, the initiating mechanical forces ultimately resulting in TBI appear not to visibly damage cells, a multi-step force signaling mechanism is likely involved, including changes to the cytoskeleton. A potential link in this pathway is the activity of mechanosensitive (MSC) PIEZO channels 2,3 that respond to stress in the lipid bilayer [4][5][6] . Since amphipathic drugs including Aβ peptides alter membrane structure 7,8 and thus membrane mechanics, we examined the effects of Aβs on PIEZO1.
How amphipaths may alter mechanical channels is suggested from studies on bacterial MSCs that respond to global forces applied through bilayer lipids 9,10 or by lipid perturbations around the channels 11 . The recently published cryoEM structure of mouse Piezo1 in a bilayer dome underscores the connection between lipids and MSCs as an integrated system 12 . Modulation of Piezo channels by amphipaths can potentially influence a number of cellular responses. For example, the differentiation of mouse neural stem cells into either neurons or glia, requires PIEZO1; inhibition of PIEZO1 by the amphipathic peptide GsMTx4 influences this fate decision 13 . Homeostatic control of epithelial cell number relies on PIEZO1 sensing of cell crowding and cell division 14,15 . PIEZO1 channels are involved in neurite extension and inhibition of PIEZO1 can alter neuronal outgrowth 16 . All these processes are associated with sensing the local mechanical environment 17 .
In this work, we asked whether Aβ can affect PIEZO1. We tested the ability of Aβ peptides to modulate the cellular response to mechanical inputs generated by a precision fluid shear stress system, measuring PIEZO1-mediated Ca 2+ fluxes. A stable overexpressing PIEZO1 cell line made the response larger and more reliable. Titration of Aβ peptides showed that monomeric Aβs in the femtomolar to picomolar range were effective inhibitors, but Aβ oligomers were much less potent. PIEZO1 inhibition by D and L enantiomers of Aβs proved equally effective, showing a lack of peptide-protein or other stereospecific interactions. The peptide-channel interaction is likely mediated via bilayer lipids. The inhibitory effects of Aβs were similar to responses to GsMTx4, another amphipathic, stereo-nonspecific inhibitor of PIEZO1 18 but requiring much higher concentrations. Inhibition of PIEZO1 by Aβs or GsMTx4 almost completely blocked cell migration in a wound healing assay. Because migration requires a pliable cytoskeleton, we measured the effect of Aβs on cytoskeletal stresses using a genetically encoded optical force probes embedded in actin 19 . The inhibition of PIEZO1 by Aβ and the resulting block of Ca 2+ influx was accompanied by significantly increased actin tension. PIEZO1 activity is causally linked to cell motility and cytoskeletal reorganization.

Results
Fluid shear stress assay. We first created a cell line that expressed a fluorescently tagged version of human PIEZO1 called hPIEZO1-1591-EGFP 5 . The fusion protein's encoded internal GFP and N-terminal His tag did not significantly affect the channel properties 5 . The cDNA was integrated into the genome of HEK293T cells using a lentivirus vector. The single channel conductance of this stably overexpressing PIEZO1 cell line (hP1-CL) was indistinguishable from that of hPIEZO1 in transiently transfected HEK293T cells ( Supplementary Fig. 1).
Aβ peptide inhibition of PIEZO1. hP1-CL cells were grown on fibronectin-coated microfluidic chambers ( Supplementary Fig. 2) and loaded with the calcium indicator, Fluo4-AM and subjected to single square shear pulses 20 ; 10 ms duration at ~15 dynes/cm 2 . The resulting robust increase in [Ca 2+ ] i (Fig. 1A,B, blue triangles) reached a maximum about ∆t = 22 s, followed by a loss of fluorescence indicating adaptation or inactivation (∆t = 60 s). The time course for fluorescence decay followed a single exponential with τ = 132 ± 10 s (6 traces from 4 independent experiments with SEM). The long delay between the stimulus and the peak response probably reflects amplification by CICR (Ca 2+ induced Ca 2+ release). Control HEK293T cells showed a minimal  Fig. 3A, dark blue triangles) produced cells unresponsive to fluid shear, suggesting a requirement of the PIEZO1 shear response for cytoskeletal integrity coupled to the bilayer. This is similar to our observation that mouse PIEZO1 whole cell currents required an intact cytoskeleton to respond to cell indentation 21 .
We tested two Aβ peptides and found that both inhibited the shear response (Fig. 1B). Incubation of hP1-CL with 10 µM L-Aβ   As a control, we examined the responses to the known inhibitor, D-GsMTx4 18 . At 5 µM, we observed complete inhibition of the response in the shear assay ( Supplementary Fig. 3A, light blue diamonds, 3 independent experiments with SEM). Supplementary Fig. 3B shows that responses to shear stress were similar in the stably transduced cell line and in transiently transfected cells, whereas cells expressing only mCherry produced no response ( Supplementary Fig. 3B). Suppressing PIEZO1 expression with miRNAs also inhibited the response, but scrambled miRNA sequences had no effect ( Supplementary Fig. 3B). These data are summarized in Panel B ( Supplementary Fig. 3) and is an average of 3 independent experiments with SEM.
We examined whether the Aβ could inhibit PIEZO1 currents in the patch. We formed outside/out patches from the hP1 cell line and tested whether 10 µM of Aβ  or Aβ (1-42) (Peptide Method I) inhibited channel activity, a concentration that blocked the shear response. Surprisingly, neither peptide inhibited outside/out patch currents ( Supplementary Fig. 5, represents 3 independent experiments).
Peptide binding specificity to the channel was assessed using the L and D enantiomers of Aβ(1-40) and Aβ(1-42), (c.f. 22 ). For all these experiments, each data point is an average of 4 independent experiments with SEM. The ratio of inhibitory constants, K i , between of the L and D Aβ peptides was ~300-fold for Aβ(1-40) (Fig. 1C, blue double arrow) and ~70-fold for Aβ(1-42) (Fig. 1C, red double arrow). Using the scrambled peptide as an estimate of non-specific binding, at least a > 10,000-fold difference between L and D forms was expected because the D-form will be incapable of specific interactions. Since we observed only a 300-fold difference, it indicated a mechanism that does not involve peptide-channel interactions. The monomeric peptide AgS (AggreSure L-Aβ(1-40); Fig. 1C, black squares) inhibited PIEZO1 with a K i about two orders of magnitude lower (~30 fM), and indicated that the sensitivity of enantiomeric peptides is related to state of the peptide (i.e. aggregate versus monomer).

PIEZO1 inhibition by monomeric peptides.
We prepared monomers of four Aβ peptides (see Peptide-Method II) and titrated their inhibitory effects on the shear response of hP1-CL cells (Fig. 1D, 4 independent experiments with SEM indicated for each concentration). The difference between the K i values of AgS peptide (~30 fM) and monomerized L-Aβ(1-40) (~52 fM) was negligible ( Table 1). The K i s for all the monomeric peptides were in the fM to sub-pM range, with K i values for enantiomeric pairs of Aβ(1-40) and Aβ  differing by only 2-to-3-fold (Table 1). These results are consistent with the assumption that peptides prepared by method I differed in oligomeric state (or conformation) from AgS in a way that reduced PIEZO1 inhibitory potency by 3-6 orders of magnitude. The similar efficacy of the D and L peptides suggests that they only interact with PIEZO1 through non-contact (lipidic) mechanisms.
We were unable to washout the channel inhibition within 20 min or by increasing the shear stress to 25 dynes/cm 2 . In conventional solution kinetics, the upper limit for the association rate constant (k a ) is usually considered to be ~10 8 M −1 s −1 , a diffusion-controlled limit. The dissociation rate constant (k d ) therefore should be less than 10 −4 s −1 , and the time constant for achieving binding equilibrium at a peptide concentration = K i , will be on the order of 10 4 s (a few hours). If the dissociation constant is 100 fM, the time constant would be about 10 5 s (a day, or so). The fact that we achieved equilibrium in much shorter times suggests that the local membrane concentration was much higher than in solution. Titration of the Aβ response showed a negatively cooperative Hill coefficient <1 suggesting that higher concentrations buffer the monomers into less effective oligomers. In contrast to Aβ inhibition, GsMTx4, had a much higher K i of ~250 nM 23,24 with a Hill coefficient of ~ 1.9 ( Supplementary Fig. 3C), and channel activity was readily restored by GsMTx4 washout (data not shown). and Panel B is the PIEZO1 channel (green). Panel C is the overlay of the red and green fluorescent channels showing yellow puncta where there is overlap indicating co-localization (indicated by arrows). Note that not all channels are associated with Aβ (1-42) peptide. Panels D-F are images near the top of the cell which also show co-localization of the peptide and PIEZO1 (indicated by arrows). Piezo proteins are known to segregate into spatial domains 25,26 in the absence of Aβs, and the altered environment of the channels may favor Aβ association.

Co-localization of peptide and channel.
Aβ peptides affect PIEZO1-mediated cell migration. PIEZO1 channels are involved in neurite extension 16 and other types of cell motility. This led us to test the effects of Aβ peptides on hP1-CL motility in a wound healing assay. Cells were grown to confluence on glass coverslips half-coated with PDMS (Polydimethylsiloxane). The PDMS layer was then peeled off allowing us to accurately measure the collective cell migration rates across the newly exposed glass surface.
Surprisingly, hP1-CL cells migrated ~10-fold faster than native HEK293T cells (Fig. 3 Top panel) showing the involvement of PIEZO1 channels in cell sheet movement. Addition of either D-GsMTx4 (5 µM) or enantiomeric forms of monomeric Aβ(1-40) (10 pM) slowed cell migration by over 10-fold, consistent with PIEZO1 involvement in cell movement (Fig. 3 Bottom left panel). Figure 3 Bottom right panel shows the average velocity for 3 independent experiments with SEM.
Aβ peptides are involved in cytoskeletal remodeling. Cell movement in the presence of PIEZO1 channels appears to rely on cytoskeletal integrity and plasticity, consistent with the ability of cytoskeletal disrupting agents to prevent PIEZO1 activation (Supplementary Fig. 4A). We examined forces within the cytoskeleton during channel inhibition. By transient transfection, we introduced a genetically encoded optical probe of actin stress (cpst-FRET) into stable hP1-mCherry-CL cells 19 (Fig. 4A), resulting in probe expression into F-actin. The fluorescence spectrum of hP1-mCherry-1591 does not significantly overlap with the spectrum of the actin force sensor. For each experiments described below, we analyzed and averaged (with SD) 15 cells from 3 different experiments. After establishing baseline actin stress for 30 min, addition of the peptides resulted in an increase in actin stress (Fig. 4B). 10 pM of monomeric L-Aβ(1-40) (red circles) appeared slightly more effective than D-Aβ (1-40) (blue triangles) or 5 µM D-GsMTx4 (black squares) in increasing actin-associated tension. Figure 4C shows that the fluid shear stress associated with a simple bath exchange had no effect on actin stress (red circles). 10 pM of monomerized scrambled Aβ peptide (black squares) also had no effect on actin stress. However, removal of Ca 2+ from the extracellular bath greatly increased the resting actin stress (blue triangles). The steady state changes in actin stress (at t = 60 min) are summarized in Fig. 4D.
Interestingly, FRET images of hP1-mCherry-CL cells show lamellipodia containing the actin probe (Fig. 4E). Addition of D-GsMTx4 or Aβ (1-40) enantiomers led to a drop of F-actin tension and its retraction from the lamellipodia (Fig. 4E,F). Our results support a model where Ca 2+ influx through PIEZO1 channels maintains the cytoskeletal plasticity required for cell migration. Control experiments with native HEK293T cells revealed no effects on actin tension by bath Ca 2+ removal, D-GsMTx-4, or Aβ peptides (Fig. 5).

Discussion
An important advantage to the shear stress approach, relative to techniques such as patch clamp 27 , is the ability to measure PIEZO1 channel function without altering cellular integrity. We know that the channel's environment clearly affects the channel's response to mechanical stress 21 . Assays intended to mimic the in situ situation must leave the environment minimally altered. We have shown that monomeric Aβ peptides can block PIEZO1 activity at fM concentrations, and that other forms of the peptides (i.e. aggregated) are less potent. Using enantiomers we demonstrated that the Aβ effects are dominated by long range effects that are not stereospecific similar to the observations seen with GsMTx4 22 .
We examined the electrophysiological effects of Aβs using patch clamp. In control cells, channel activity in response to stretch was obvious 18,21 . We could not observe PIEZO1 inhibition by Aβ peptides at concentrations that inhibited the response to fluid shear stress. The inability to observe patch current correlations with the extreme sensitivity of the shear stress assay is a warning that extrapolation of patch results does not always apply How may Aβs affect channel function? Mouse PIEZO1 gate by local bilayer tension 5,6 . In neurons, PIEZO1 activity appears concentrated in cholesterol-rich lipid raft domains containing Stomatin-like 3 protein (STOML3) 26,29 . These membrane nanostructures can redirect the mechanical forces that modulate MSCs 26,30 . The insertion of Aβ peptides into a confined lipid domain compresses 31 the surrounding lipid and channels. Alternatively, Aβ disrupts the boundary lipids of PIEZO1 11 , as suggested by cryoEM 12 . Perhaps Aβs perturb these structural lipids rendering the channel inactive. Aβ peptides are known to concentrate in lipid rafts 32,33 . The extremely low concentration of monomeric peptide needed to inhibit channel function (fM) may reflect a partition coefficient that concentrates monomeric peptides in constrained domains containing PIEZO1.
Another possibility is that Aβ affects PIEZO1 activity indirectly. Activation of PIEZO1 channels may result from forces created through the cytoskeleton. Cells generate traction forces by interactions with the extracellular matrix (ECM) and can use integrin rich focal adhesion regions to transmit tension to the bilayer activating  -test). The scrambled Aβ peptide had no effect on the cells (*P = 0.813), whereas bath Ca 2+ removal increased actin stress (**P < 0.0001). Enantiomeric Aβ(1-40) peptides and D-GsMTx4 all increased actin stress (paired t-test *** and ****P < 0.0001; *****P < 0.0004). Panel E presents actin force probe images of hP1-mCherry-CL cells highlighting lamellipodia. Addition of D-GsMTx4 or the enantiomeric Aβ peptides reduced actin-force probe FRET emissions (as shown by loss of fluorescent signal in the circled regions). Panel F shows changes to fluorescence after retraction of actin from lamellipodia. Data normalized using the fluorescent intensity before the addition of peptide (F 0 ). The change in intensity is F/F 0 where F is the final measurement in the series. The data average 20 lamellipodia from 4 cells (with SEM). Before and after intensities were analyzed by paired t-test (*P < 0.04; **P < 0.0001; ***P < 0.01). Note that the control increased fluorescent intensity.
Scientific REPORtS | (2018) 8:14267 | DOI:10.1038/s41598-018-32572-2 PIEZO1 channels 34 . The connection between PIEZO1 (Fam38a) and integrins was demonstrated prior to the discovery of PIEZO1's mechanical channel properties 35 and activation of PIEZO1 channels through the ECM was shown by Poole et al. 26 . The disruption of PIEZO1 activity by Aβ would be attributed to inhibition of any step along a signaling pathway.
Aβ peptides vary considerably in their structure and oligomeric states 36 . Moreover, opinions vary regarding toxicity of the various peptide forms, ranging from soluble dimer/trimers to fibers 37 . Recent work suggests that Aβ monomers are random coils 38 .
Although Aβ peptides are often associated with Alzheimer's disease, monomeric Aβ peptides at low concentration can exert positive physiological effects on synaptic plasticity and neuronal survival 39,40 . Giuffrida et al. 41 reported the interesting observation that monomeric Aβ (1-42) is neuroprotective, preventing trophic deprivation in developing neurons and protecting mature neurons against excitotoxic death at low concentrations. We have shown the connection between PIEZO1-mediated Ca 2+ influx and cytoskeleton remodeling and cell mechanics (Fig. 4). The elevation in Aβ peptide concentration following head trauma might be intended to limit the PIEZO1 response to mechanical activation. Consistent with our results on the loss of efficacy with oligomer formation, Giuffrida et al. 41 suggested that oligomerization decreases the concentration of monomers, and may explain a negative Hill coefficient (Table 1). These results also suggest that other amphipathic molecules may also exhibit environmentally sensitive affinities as a function of bilayer tension 31 .
The coupling of PIEZO activity to motility is clear from the effects of over-expression in the HEK293T cells, and inhibition by Aβ peptides. Collective cell migration is related to tissue remodeling events involved in wound healing and cancer 42 , and is distinct from anchorage-independent ("amoeboid") cell migration observed in the setting of FAM38a (PIEZO1) depletion 35,43 . Linking the flux of Ca 2+ through PIEZO1 to cytoskeletal changes 34 allows for rapid and coordinated cell migration. The ubiquitous expression of Aβ peptide 44 raises the question of how it may influence mechanical responses throughout the organism.

PIEZO1 cell lines.
Vector hP1-1591-mCherry. The vector hP1-1591-EGFP-mCherry with an N-terminal his tag was amplified using a BamH1 forward primer and EcoR1 reverse primer. The gel purified PCR product was used in an InFusion reaction with pBabe puro vector DNA that had been cut with BamH1 and EcoR1. The resulting construct, pBabePuro-NHis-hP1-1591-mCherry was treated with restriction enzymes MluI and XhoI. A PIEZO1 cDNA fragment containing inserted EGFP at amino acid residue 1591 and with MluI and XhoI ends was inserted by ligation (pBabePuro-NHis-hP1-1591-EGFP). Integrity of changes was confirmed by DNA sequencing.

InfHP1BabeR1rev -CACCGGTACTGAATTCCTACTACTCCTTCTCACGAGTCC.
Generation of stable cell line. To generate viral particles, AmphoPak cells in 30 mm dishes were grown to 25% confluency and transfected with DNA (1 µg per dish) using Mirus transfection reagent according to manufacturer's recommendation. After 2 days, the media was collected and filtered through a 45 micron filter.
1 ml freshly harvested virus plus 2 µg/ml polybrene were added to adherent HEK293T cells grown to 25-50% confluency in 30 mm dishes. Fresh medium was added after 3 hrs, and the cells were allowed to incubate overnight. Lentiviral infection was repeated after 24 hours. The twice infected cells were allowed to incubate for an additional 24 hours, and subjected to puromycin selection (Sigma) at four concentrations between 0.25 to 4.0 ug/ml. Single colonies were generated from cells grown at 2.8 µg/ml puromycin. The resulting clonal hP1-CL cell line was used in subsequent experiments. Shear stress assay. The microfluidic chambers had glass coverslip bottoms coated with human fibronectin (Invitrogen) 20 and fluid flow guided by PDMS channels 20 . Cells were cultured in the chamber for 3 days, with daily medium changes. On the experimental day, chambers were rinsed with isotonic solution (75 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 1 mM CaCl 2 and 10 mM HEPES, adjusted to 320 mOsm with mannitol). A stock of Fluo-4 AM was diluted to 5 µM in isotonic saline and incubated with cells for 10-30 min at 37C°. Chambers were then rinsed twice with isotonic solution and incubated for 8 minutes to complete Fluo-4 AM de-esterification. A single fluid shear stress pulse of ~15 dynes/cm 2 for 8-10 ms was applied to the chamber using a high-speed pressure servo (HSPC-1 from ALA). Peptide stocks were kept at 4 °C until used, dissolved to desired concentrations, and immediately added to cells. Each chamber had a minimum of 15 cells.
Images of the cellular response to shear stress were captured by an EM-CCD camera (C9100 model, Hamamatsu) at a nominal rate of 1 Hz. To quantify the response, the average intensity in the field of view was background subtracted 45 . Image intensity was adjusted for the rate of bleaching. Each condition was repeated in four separate chambers, and the results averaged. The responses were normalized to the control experiments, and the inhibition was calculated as previously described 46 . The data points were fit by a Hill plot with the equation: where Vmax is the maximum response, n is the Hill coefficient, and k is related to the concentration at half-maximal response. The extracted parameters are shown in Table 1.
Electrophysiology. Cell-attached and outside-out patches were used as previously described 47 . All experiments were at room temperature. Pressure or suction steps were applied to the membranes using a high speed pressure clamp (ALA) 48  The two miRNAs were introduced into one vector (chaining) according to the manufacturer's specifications.
Cell Migration. Half of a glass coverslip was coated with catalyzed PDMS (polydimethylsiloxane) (9:1) and allowed to harden. Cells were cultured on the complete coverslip kept in tissue culture dishes. After achieving confluency, the PDMS was peeled from the glass. The location of the front edge of the confluent cells was monitored for 5 hr at 37 C° with 5% CO 2 -supplemented DMEM media containing 10% bovine serum. Test peptides were added to the culture media at the start of the 5 hr experiment. To analyze the average migration speed cell front, images were processed using Cell Profiler (open source software, http://cellprofiler.org/) to measure the area covered by the cells in each frame. Dividing by the width of the frame, an average (with SEM) speed of movement in a direction perpendicular to the cell front was calculated The datasets generated and/or analyzed during the current study are available from the corresponding author on request.
Super resolution microscopy. Samples were imaged on a Visitech VTI-iSIM attached to a Nikon TE2000 inverted microscope. Images of samples excited at 488 nm and 568 nm were captured at 60 z-intervals of 200 nm.
Cytoskeleton stress probe. To measure cytoskeletal stress, hP1-mCherry-CL cells were transiently transfected with cDNA encoding the cpst-actin FRET probe 49 . The FRET ratio of 528 nM and 475 nM was measured by simultaneous imaging using an image splitter. Baseline actin stress was measured at 37 °C in isotonic solution for 30 min before peptide addition (see above). Method II. (monomer procedure 41 ): One mg of peptide was dissolved in 0.5 ml TFA (trifluoro acetate) and sonicated for 10 min. TFA was removed with streaming argon after which 800 µl of 1,1,3,3,3-Hexafluoro-2-propanol (THIP) (Sigma-Aldrich) was added and incubated 1 hr at 37 °C. This solvent allows aggregated peptides to dissociate into monomers by interfering with interactions used to stabilize aggregate peptides. The solvent was removed by streaming argon, and 800 THIP was added and frozen to −80 °C. The material was lyophilized and dissolved in 5 mM DMSO (Sigma-Aldrich) and then diluted in buffer. All peptides were stored at −80 °C.

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.