Probing PIEZO1 Localization upon Activation Using High-Resolution
Atomic Force and Confocal Microscopy
Andra C. Dumitru,§ Amaury Stommen,§ Melanie Koehler, Anne-Sophie Cloos, Jinsung Yang,
Arnaud Leclercqz, Donatienne Tyteca,* and David Alsteens*
Cite This: Nano Lett. 2021, 21, 4950−4958 Read Online
ACCESS Metrics & More Article Recommendations *sı Supporting Information
ABSTRACT: PIEZO1 ion channels are activated by mechanical
stimuli, triggering intracellular chemical signals. Recent structural
studies suggest that plasma membrane tension or local curvature
changes modulate PIEZO1 channel gating and activation.
However, whether PIEZO1 localization is governed by tension
gradients or long-range mechanical perturbations across the cells is
still unclear. Here, we probe the nanoscale localization of PIEZO1
on red blood cells (RBCs) at high resolution (∼30 nm), and we
report for the first time the existence of submicrometric PIEZO1
clusters in native conditions. Upon interaction with Yoda1, an allosteric modulator, PIEZO1 clusters increase in abundance in
regions of higher membrane tension and lower curvature. We further show that PIEZO1 ion channels interact with the spectrin
cytoskeleton in both resting and activated states. Our results point toward a strong interplay between plasma membrane tension
gradients, curvature, and cytoskeleton association of PIEZO1.
KEYWORDS: PIEZO1 clusters, atomic force microscopy, single-molecule force spectroscopy, ion channel, red blood cells,
laser scanning confocal microscopy (CLSM), mechanotransduction
■ INTRODUCTION
PIEZO1 is a mechanosensitive ion channel associated with
numerous physiological functions such as erythrocyte volume
regulation, epithelial cell adhesion, blood vessel formation, and
development of vascular architecture.1−3 Impaired PIEZO1
function has been linked to erythrocyte osmotic fragility and
impairment of vascular development.4,5 Conversely, gain-of￾function PIEZO1 mutations lead to decreased red blood cell
(RBC) volume and hemolysis.6,7 Recently published cryo￾electron microscopy structures of purified PIEZO1 provide
detailed information on its homotrimeric structure consisting
of three propeller blades extending away from a central cap
region8−11 (Figure 1A). Recent studies suggest that the blades
and the helical beams attaching them to the cap region may
efficiently transfer mechanical forces from lipids to the pore via
a lever mechanism (Figure 1A).12,13 So far, experiments on
lipid bilayers,14 proteoliposomes,15 membrane patches,16 and
bleb membranes17 provided proof that PIEZO1 is sensitive to
lateral membrane tension and can sense forces directly
transmitted through the membrane. In this force-from-lipids
model, the ion channel is gated by mechanical perturbations in
the surrounding lipid bilayer.18,19 While the force-from-lipid
paradigm has been firmly established for bacterial mechano￾sensitive channels (MscL and MscS) and eukaryotic TREK,
TRAAK, or two-pore domain K+ (K2P) channels,20,21 it
contrasts with the force-from-filament tether model in which
mechanogating occurs via connections to the cytoskeleton or
extracellular tethers.22−24 Multiple studies used heterologous
systems and showed that the cytoskeleton is not essential for
PIEZO1 gating,17,25,26 but there is conflicting evidence that
PIEZO1 can sense and respond to localized or long-range
mechanical perturbations either exogenously applied or
endogenously originated.27−29 Furthermore, it also remains
unclear why PIEZO1 is recruited to cell-junction or focal
adhesion sites30,31 and whether it forms domains that could
facilitate recruitment at appropriate plasma membrane
locations.
Here we use confocal imaging and force−distance based
atomic force microscopy (FD-AFM) to determine the
distribution of PIEZO1 prior and after activation by Yoda1,
a high potency allosteric modulator.12,15 We first identify the
presence of submicrometric PIEZO1 clusters in native
conditions at the surface of RBCs. Those cells indeed represent
an ideal cell model to study PIEZO1 localization and activation
as the RBC surface is featureless with different curvature and
tension areas, and RBCs constitutively express PIEZO1
allowing for cations including Ca2+ entry. We further observe
Received: February 10, 2021
Revised: June 9, 2021
Published: June 14, 2021
pubs.acs.org/NanoLett Letter
© 2021 American Chemical Society 4950

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
Downloaded via HIMACHAL PRADESH UNIV on August 11, 2021 at 17:05:20 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
that PIEZO1 localization, accessibility and local density are
regulated by membrane tension and curvature. PIEZO1 cluster
abundance is increased in areas of high membrane tension in
response to Yoda1. FD-AFM and confocal imaging show that
the lateral organization of PIEZO1 follows a spectrin meshlike
distribution. Taken together, PIEZO1 cluster regulation by
membrane tension and cytoskeleton anchoring supports the
coexistence of force-from-lipid and force-from-filament gating
models upon PIEZO1 ion channel activation.
■ RESULTS AND DISCUSSION
Probing PIEZO1 Ion Channels by FD-AFM. In order to
evaluate the distribution of PIEZO1 on RBCs, we used a
polyclonal PIEZO1-antibody, which demonstrated a high
specificity toward PIEZO1 on RBC ghosts, as shown in
Western blot experiments (Figure S1A). We then validated the
use of this antibody to specifically bind the extracellular
domain of PIEZO1 ion channels on model surfaces and RBCs
using force-volume (FV) and FD-AFM (Figure 1B,C and
Figure S2). Thus, we covalently grafted single PIEZO1-
antibodies to the free end of a heterobifunctional polyethylene
glycol (PEG) linker attached to the AFM tip. In a first
approach, to mimic cell-surface PIEZO1 ion channels in vitro a
fragment of PIEZO1 was covalently immobilized on gold
surfaces using carbodiimide cross-linker chemistry (Figure 1B).
PIEZO1-antibody AFM tips were used to probe the binding
properties of the complex formed with PIEZO1 ion channels
fragments grafted on model surfaces by FV-AFM (see
Supporting Information (SI) Methods and Figure S2A−C).
Specific adhesion events were observed in ∼6% of the
retraction FD curves at rupture distances >10 nm, correspond￾ing to the extension of the PEG linker-antibody complex
(Figure 1D curves 1, 2). FD curves displaying no adhesion
events or unspecific adhesion events were discarded from
Figure 1. Probing PIEZO1 binding using a PIEZO1-antibody functionalized AFM tip. (A) Cartoon representation of PIEZO1 side and top views,
where each subunit is colored differently (PDB: 6B3R). PIEZO1 is a homotrimer protein with three propeller blades extending away from the
central cap region. (B) Interactions between recombinant PIEZO1 protein fragments and the PIEZO1-antibody are monitored by AFM on model
surfaces, where a PIEZO1 fragment is covalently coupled to a gold surface with NHS/EDC chemistry. (C) AFM tips functionalized with PIEZO1-
antibody were used to probe PIEZO1 ion channels on RBCs naturally expressing the protein. (D) Examples of FD curves recorded on model
surfaces coated with recombinant PIEZO1 protein fragments and (E) the corresponding unbinding forces histogram at 1 μm s−1
. (F) FD curves
obtained while probing PIEZO1 receptors on the surface of RBCs and (G) the corresponding unbinding forces histogram at 125 μm s−1
. (H) Free
energy landscape of PIEZO1-PIEZO1-antibody complexes unbinding. A single potential barrier separates the bound and the unbound states of a
two-state unbinding process (blue line). The activation free energy of unbinding is ΔG* and xβ represents the transition between the bound and
the unbound states along the reaction coordinate. In equilibrium conditions (F = 0), the energy barrier is spontaneously crossed at a transition rate
koff(0). The application of an external force to the bond (F > 0) tilts the energy landscape (dotted gray line), lowering the energy barrier. Probing
the interaction forces of a (bio)molecular bond over a wide range of loading rates allows determination of the parameters describing the free-energy
barrier stabilizing the bond, such as xβ and the off-rate koff. (I) Dynamic force spectroscopy plot showing the dependence of the rupture force on the
loading rate when probing the interaction between PIEZO1-antibody and either model surfaces coated with recombinant PIEZO1 protein
fragments (empty black dots, N = 3006) or PIEZO1 ion channels on the surface of RBCs (empty red dots N = 816). The most probable rupture
forces calculated from the Gaussian fits in Figure S3 are represented as filled black and red dots, mean ± SD. The solid line represents the fit of the
data using the Bell-Evans model and the dotted line is the Williams-Evans prediction for two parallel bonds. In panels E and G, BP represents the
binding probability and bin width is 25 pN. The data are representative of at least five independent experiments. Panels A−C created with
BioRender.com and used with permission.
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4951
further analysis (Figure 1D curves 3, 4). For specific binding
interactions, rupture forces of 71 ± 21 pN (mean ± standard
deviation (S.D.); N = 588 data points) at 1 μm s−1 retraction
speed were detected (Figure 1E). To further validate the
specificity, the following two independent control experiments
were performed: (i) a competition assay by injecting 1 μM free
PIEZO1-antibody in the solution and (ii) a negative control
using AFM tips functionalized with a nonspecific IgG isotype
similar in size to PIEZO1-antibody (Figure S1B,C). We
observed a significantly lower binding frequency, below 2%
versus 6% compared to our standard condition (PIEZO1-
antibody tips probing recombinant PIEZO1 fragments￾Figure 2. Probing resting state and chemically activated PIEZO1 on the surface of RBCs. (A) In resting state, the PIEZO1 pore has a closed
conformation. Yoda1 stabilizes the open conformation of the PIEZO1 channel, allowing for ion exchange. The three propeller blades are thought to
induce a local curvature around the central pore of PIEZO1 and could be used as levers to sense tension-induced flattening of the plasma
membrane and open the pore. (B,C) Intracellular Ca2+ content of Yoda1 treated RBCs (orange) was assessed by fluorimetry (B) or by flow
cytometry (C). Fluorimetry data were normalized to hemoglobin (Hb) content and median fluorescence intensity (MFI) was determined from
flow cytometry data. (D) Extent of RBC hemolysis upon treatment with Yoda1. (E,F) PIEZO1 distribution observed on fixed and permeabilized
RBCs by confocal microscopy in control (E) and Yoda1 treated conditions. (G) Quantification of mean fluorescence intensity per immunolabeled
RBC. (H) Quantification of PIEZO1 clusters abundance per immunolabeled RBC. (I,K) FD-based AFM topography images and (J,L)
corresponding adhesion maps of control and RBCs treated with Yoda1. (M) Binding frequencies measured when probing a PIEZO1-antibody
functionalized AFM tip against control (gray) or Yoda1 treated cells (yellow). (N) Adhesion forces measured on control (gray) and Yoda1 treated
cells (orange). Results in (B,C,D,G) are expressed as percentage of the control condition. Results are presented as means ± SEM of 3−12
experiments in B, means ± SD of two experiments in C, and means of three experiments in D. In E-N, Yoda1-treated RBCs represent cells treated
with 50 nM Yoda1. Each dot represents one independent experiment in G. Each dot represents a single RBC from at least three independent
experiments in H,M,N. %CTL is defined as percent control. ns: not significant. **: p < 0.01 and ***: p < 0.001. Panel A was created with
BioRender.com and used with permission.
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4952
functionalized surfaces). In agreement with the above data, the
average rupture force is 69 ± 20 pN in our standard conditions
and similar values of 66 ± 24 pN are obtained following an
injection of PIEZO1-antibodies (Welch’s t test: p = 0.101, ns).
Rupture forces quantified for anti-gH/gL tips probed on the
same surface coated with recombinant PIEZO1 fragments are
of 42 ± 13 pN, which is 30% lower than in normal conditions
and statistically relevant (Welch’s t test: p = 2.88 × 10−15),
confirming the specificity of the setup to detect PIEZO1
fragments (Figure S1C).
Similarly, we probed RBCs that naturally express PIEZO1
ion channels and have the great advantage of displaying
featureless surfaces, making them a good model for membrane
studies. To preserve the cellular biconcavity, RBCs were
instantaneously fixed in suspension, and then immobilized on
poly-L-lysine (PLL) coated glass Petri dishes and imaged with
the PIEZO1-antibody functionalized tips (Figure 1C) using
FD-based AFM32 (Figure S2D). Height and adhesion maps
were simultaneously recorded (see SI Methods and Figure
S2E,F) and force versus time (FT) curves were analyzed in
order to extract the rupture forces of specific adhesion events
for each pixel in the images (Figure S2A). Around 2% of the
FD curves showed specific adhesion events between PIEZO1-
antibody and PIEZO1 ion channels (Figure 1F curves 1, 2),
with rupture forces of 115 ± 26 pN (mean ± S.D, N = 268
data points) at ∼125 μm s−1 retraction speed (Figure 1G).
Other FD curves showing either no adhesion or unspecific
adhesion events were discarded for further quantification
(Figure 1F curves 3, 4).
PIEZO1-Antibody Binding Free-Energy Landscape.
The rupture forces probed on model surfaces and on RBCs
strongly differ due to the different tip velocities applied during
the experiments. To unambiguously demonstrate that these
rupture forces originated from the same binding partners, we
characterized the kinetic parameters underlying the interaction
between PIEZO1 and PIEZO1-antibody. In AFM-based
single-molecule force spectroscopy studies, the application of
an external force to a ligand−receptor bond adds a mechanical
potential that tilts the free-energy landscape and lowers the
free-energy barrier toward dissociation, thereby increasing the
likelihood of unbinding by reducing the lifetime of the bond
(Figure 1H).33 The Bell-Evans (BE) model predicts that the
most probable rupture force in the thermally activated regime
scales linearly with the logarithm of the loading rate (LR).34
We extracted the rupture forces and LRs from FT curves
recorded on both model surfaces and RBCs (Figure S3) and
we overlaid them in a dynamic force spectroscopy plot (Figure
1I, see SI Methods). The extracted rupture forces show a linear
dependency with the logarithm of the LR, spanning a broad
range between 102 to 107 pN s−1
. We observed a very good
alignment between the data obtained on purified ion channel
fragments and RBCs, confirming the physiological relevance of
our results obtained on model surfaces. The BE model was
used to fit the data (Figure 1I, solid black line I), assuming a
simple two-state model where the bound and the unbound
state are separated by a single energy barrier. From the slope of
the fit, we estimated the length scale of the energy barrier xu =
0.46 ± 0.04 nm and the kinetic off-rate (koff) or dissociation
rate is obtained from the intercept of the fit (at LR = 0)
yielding koff values of 0.62 ± 0.41 s−1
. In addition, we also
observed bivalent interactions that appear as uncorrelated
bonds loaded in parallel, where the load is equally distributed
among the existing bonds in an attachment, as predicted by the
Williams-Evans (WE) model (Figure 1I, dashed black line).35
While grafted at low density, these double interactions could
result from the fact that PIEZO1-antibodies have two antigen￾binding sites.
PIEZO1 Increased Accessibility upon Activation by
Yoda1. The small molecule Yoda1 has been shown to activate
PIEZO1 ion channels, acting as a gating modifier that
potentiates its mechanosensitivity.12,36 The exact mechanism
underlying this interaction is unknown, but PIEZO1 is thought
to harbor a specific binding site for Yoda1 in the region
between the anchor domain and the piezo repeat A983,36
(Figure 2A). Recent studies evidenced that Yoda1 causes the
ion channel to open up, which enables exchange of cations
including calcium (Ca2+)
37 (Figure 2A). We therefore treated
RBCs with Yoda1 at different concentrations to modulate
PIEZO1 channel activity and followed Ca2+ accumulation
using spectrofluorimetry and flow cytometry. In the 30−200
nM range, we observed a linear increase of intracellular Ca2+
content as a function of Yoda1 concentration (Figure 2B,C)
and excluded any effect related to RBC hemolysis (Figure 2D).
Higher Yoda1 concentrations induce a stronger Ca2+ intra￾cellular accumulation (Figure S4K), but in RBCs this is
accompanied by a loss of transversal membrane asymmetry
(Figure S4L), required for PIEZO1 activation.38 Moreover, a
careful examination of RBC morphology showed that even at
100 nM, loss of RBC biconcavity is clearly visible in fixed
RBCs (Figure S4A−J). We therefore decided to work at 50 nM
Yoda1 for subsequent experiments.
We first studied the nanoscale distribution of PIEZO1 on
RBCs in both resting and activated state using confocal laser
scanning microscopy (CLSM, Figure 2E,F) and FD-AFM
(Figure 2I−L). Using CLSM, we started by checking the
specificity of the PIEZO1-antibody (Figure S5A) and
optimized the immunostaining procedure by instantaneous
fixation in suspension using a mix of glutaraldehyde/
paraformaldehyde to preserve the RBC biconcavity (see SI
Methods) and by permeabilization with a Triton X-100
concentration that does not affect PIEZO1 organization at the
RBC surface (Figure S5B). We observed that PIEZO1 is
distributed into clusters of ∼275 nm in diameter (Figure
2E,F). Yoda1 treatment leads to a significant 1.7-fold increase
of the mean fluorescence intensity per RBC (Figure 2G) and a
significantly higher number of PIEZO1 clusters compared to
the control condition (Figure 2H). Previous electrophysiology
experiments suggested that PIEZO1 could exist in discrete
physical domains whose global properties can modify channel
gating39 and a very recent STORM superresolution imaging
study using a PIEZO1-GFP fusion protein indicates that
plasma membrane cholesterol domains coordinate the activity
of clustered PIEZO1 channels.40 However, to the best of our
knowledge, this is the first report of PIEZO1 clustering in
submicrometric domains observed in a native system.
AFM high-resolution height images and corresponding
adhesion maps recorded with PIEZO1-antibody derivatized
tips provide further insights into the localization of PIEZO1 at
the RBC surface (Figure 2I−L). Notably, Yoda1 treatment of
RBCs results in a significant 2-fold increase of the binding
frequency from 1.8 ± 0.3% to 3.5 ± 0.4% (Figure 2M).
However, the rise in binding frequency is not accompanied by
a change of the mean rupture binding force, which remains
constant in both control and Yoda1 conditions (Fcontrol ∼ 157
± 36 pN, FYoda1 ∼ 159 ± 21 pN, Figure 2N). Together, these
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4953
results support the fact that Yoda1 treatment induces an
increase of the PIEZO1 accessibility.
Plasma Membrane Distribution of PIEZO1 Clusters
upon Activation by Yoda1. While some previous studies
have shown that PIEZO1 channels diffuse easily within the
plasma membrane and are widely distributed throughout the
cell,28 PIEZO1 activation is stress and stimuli dependent.40−42
To further investigate this process, RBCs appear as a tool of
choice thanks to their well-defined shape with various
curvature and tension areas. It has been demonstrated that
membrane forces are nonuniform along the RBC membrane
and that the force density and membrane tension are larger in
the center/dimple than in the edge/rim.43 Moreover, the
presence of nanometer-scale heterogeneities32 could indeed
serve as reservoir for the generation of positive or negative
curvature at the local level, which could induce channel gating.
Therefore, we anticipated that PIEZO1 conformation and
cluster localization could be controlled by local or global
membrane curvature and tension. To evaluate this hypothesis,
we analyzed by CLSM and FD-AFM the PIEZO1 localization
at the RBC edges versus the center before and after activation
by Yoda1 (Figure 3). For CLSM experiments, RBCs were
immunolabeled with PIEZO1-antibody enabling to quantify
the PIEZO1 distribution in top and side view. The rim was
associated with high curvature (HC) areas, while the dimple
corresponded to low curvature (LC) regions (Figure 3A).
Representative CLSM images of single RBCs before and after
Yoda1 treatment are shown in Figure 3B,C.
Figure 3. Quantification of PIEZO1 receptors clustering on the surface of RBCs. (A) Schematic representation of a RBC displaying a low curvature
(LC) area in the central region and a high curvature (HC) area on the rim. The 2D horizontal projection of the RBC surface was divided by three
concentric circles, representing the inner, median, and outer areas. (B,C) Representative CLSM images of single RBCs showing PIEZO1
distribution in LC areas for control and Yoda1-treated RBCs. 1, RBC in upper view; 2, RBC in side view. (D) Quantification of fluorescence
intensity in HC (filled circles) and LC (empty circles) areas of CTL (gray) and Yoda1-treated (orange) RBCs. (E,F) Abundance of PIEZO1
clusters in (E) LC and in (F) HC areas for control and Yoda1-treated RBCs. (G−J) FD-based AFM height images and corresponding adhesion
maps of (G,H) control and (I,J) Yoda1-treated RBCs. (K,L) Spatial distribution of specific PIEZO1 unbinding events in the adhesion maps (H)
and (J), as a function of the distance from the center of the cell. (M) Adhesion maps were used to calculate normalized integrated intensity of
PIEZO1-antibody adhesion events in the inner, median, or outer regions of control and Yoda1-treated RBCs. Bar range represents data mean ±
SEM of 3−10 representative experiments. Ns: not significant; **: p < 0.01; ****: p < 0.0001; and *****: p < 0.00001. Yoda1-treated RBCs
represent cells treated with 50 nM Yoda1. Panel A created with BioRender.com and used with permission.
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4954
We observed that Yoda1 modulation of PIEZO1 ion
channels induces a significant increase of fluorescence intensity
(Figure 3D) and abundance of PIEZO1 clusters in LC but not
in HC areas (Figure 3E,F). PIEZO1 is distributed into clusters
of ∼315 nm diameter in HC, compared with 260 nm in LC
regions. Yoda1 treatment does not affect cluster size, which
remains stable at ∼275 nm diameter in HC versus 265 nm in
LC regions (Figure S6A,B).
We used an analogous approach to quantify FD-AFM
results. Specifically, based on the RBC round shape we
analyzed the distribution of PIEZO1 ion channels and
classified their location into three concentric regions (inner,
median, and outer; Figure 3A and SI Methods). Adhesion
maps of control and Yoda1-treated RBCs were analyzed and
we extracted profile plots of normalized integrated intensities
around inner, median, and outer regions (Figure 3G−L).
Before Yoda1 treatment, PIEZO1 are predominantly localized
at the RBC outer rim, which is in line with the higher number
of PIEZO1 clusters observed in the HC areas by CLSM (see
Figure 3 panel F versus E). After treatment by Yoda1, the
presence of PIEZO1 in the inner and median regions is
markedly increased, compared to a very slight increase in the
outer region (Figure 3M).
The increased abundance of PIEZO1 in the dimple (LC or
inner region) confirmed by AFM and CLSM suggests that, as
an allosteric modulator, Yoda1 could induce conformational
changes of the ion channel that increase the affinity toward the
PIEZO1-antibody and the intrinsic curvature of PIEZO1 upon
activation favors a stronger interaction with LC areas.
Moreover, we hypothesize that the open pore conformation
of PIEZO1 stabilized by Yoda1 treatment leads to more
clusters of sufficient size to be resolved by confocal microscopy
due to higher PIEZO1 membrane lateral diffusion. According
to this hypothesis, PIEZO1-GFP shows a restricted lateral
diffusion in the HEK293T plasma membrane while depletion
of cholesterol, which is known to control membrane tension,
causes a faster channel membrane diffusion.40
The increased number of PIEZO1 clusters upon Yoda1
treatment correlates with previous findings pointing toward a
higher force density and membrane tension in the dimple
region of RBCs.43 Our data are in good agreement with recent
computational studies showing that myosin-mediated forces
are not homogeneously distributed along the RBC plasma
Figure 4. PIEZO1 association with the spectrin skeleton. (A) Schematic representation of the RBC cytoskeleton. The two-dimensional cytoskeletal
meshwork has a quasi-hexagonal symmetry and has as main components (α1β2)2 spectrin tetramers connected by junctional complexes. The RBC
cytoskeleton is tethered to the plasma membrane through peripheral and integral proteins. (B−I) Representative confocal images of PIEZO1-
spectrin colocalization in control and Yoda1-treated RBCs after fixation and permeabilization . Green and red arrowheads correspond to PIEZO1
or spectrin only; yellow arrowheads indicate colabeling. Panels B,F are provided to highlight the preservation of the RBC morphology. (J)
Quantification of the extent of colocalization between PIEZO1 and spectrin from the confocal images shown in B−I. (K,L) Representative FD￾AFM height and adhesion maps of a single RBC probed with a PIEZO1-antibody AFM tip. (M) Overlay of a hexagonal lattice model and the
boxed region in the adhesion map in (L). Each unit cell in the hexagonal network consists of 6 spectrin-like rods 200 nm in length. Red dots
evidence overlap of PIEZO1 adhesion events and lattice connection points. (N,O) Histograms of nearest neighbor distribution for adhesion events
extracted from representative control and Yoda1 treated adhesion maps. A multimodal distribution of the distances is observed for both conditions.
Bin width is 20 nm, blue lines represent individual Gaussian fits, and dark gray lines global Gaussian fits. Results are from at least four independent
experiments. Ns: not significant. Panel A was created with BioRender.com.
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4955
membrane and that the force density is higher in the central
region compared to the rim. Another in silico study also
speculated that PIEZO1 channels could localize preferentially
in the dimple regions of RBCs as a result of surface tractions
and the control exerted by membrane curvature.42 This in
agreement with the force-from-lipids model, which assumes
that the channels respond to changes in membrane tension
directly.
Deciphering PIEZO1-Spectrin Skeleton Interplay.
Cryo-electron tomography studies revealed that the RBC
cytoskeleton is a densely packed heterogeneous network of
filaments.44 This two-dimensional cytoskeletal network has a
quasi-hexagonal symmetry and consists of (α1β2)2 spectrin
tetramers interconnected at short junctional complexes (JC)
made of actin filaments, protein 4.1, tropomodulin, and other
associated proteins44,45 (Figure 4A). The RBC cytoskeleton is
tethered to cell membrane by integral proteins such as Band3
or Glycophorin C (GPC) (Figure 4A). Band3 is actually the
most abundant protein in the RBC plasma membrane and half
of Band3 proteins in RBCs associate either with ankyrin or
protein 4.1, forming either ankyrin complexes or JC.46 Spectrin
tetramers are found in a compressed state in the cytoskeletal
network, and it is still a challenge to experimentally determine
the actual ultrastructure of the erythrocyte cytoskeleton.
Average lengths of spectrin tetramers estimated by different
techniques range between 35 and 200 nm, whereas JC are just
a fraction of around 30−50 nm.44,47−50
Besides the force-from-lipid model, the force-from-filament
gating model suggests that mechanosensitive ion channels
couple to extracellular structures or intracellular tethers (the
cytoskeleton), which transmit macroscopic forces that induce
mechanogating.24 To evaluate if PIEZO1 channels might
indeed use a complementary force-from-filament gating model
to enable long-distance sensing of mechanical perturbation
across an intact cell via the cytoskeleton, we tried to determine
whether the location of PIEZO1 on RBCs could be governed
by the underlying cytoskeleton.
We used CLSM to analyze the extent of PIEZO1
colocalization with spectrin and Band3, as main components
involved in RBC mechanical homeostasis (Figure 4B−J). For
the control RBCs, the extent of colocalization of PIEZO1 with
spectrin is of ∼0.45 and with Band3 of ∼0.40 (Figure 4J,
Figure S7A). Importantly, spectrin or PIEZO1, each decorated
by one primary antibody and recognized by two different
secondary antibodies, did not reach a coefficient higher than
∼0.6 (Figure S7B). Moreover, and quite excitingly, Band3-
spectrin colocalization coefficients are similar as for PIEZO1-
spectrin and PIEZO1-Band3 (Figure S7B). The extent of
PIEZO1 colocalization with spectrin or Band3 is not
significantly different in Yoda1 treated RBCs (Figures 4J;
S7A). Our hypothesis is that PIEZO1 ion channels interact
with spectrin, which would enable them to be anchored in the
RBC LC areas, where force density and tension are larger.43
This would allow PIEZO1 to sense and respond to mechanical
stimuli transduced as forces through the cytoskeleton.51 To
gain further insight about the PIEZO1-spectrin interaction
hypothesis, we analyzed multiparametric FD-AFM maps of
control and Yoda1 treated RBCs (Figure 4K,L shows an
example of control RBCs maps), focusing on the spatial
distribution of PIEZO1 channels probed in AFM experiments.
Figure 4M depicts a hexagonal lattice model, where each unit
cell is made of six spectrin-like rods of 200 nm, overlaid with a
representative region of an adhesion map (boxed region in
Figure 4L). An overlay of the hexagonal lattice model with the
localization of PIEZO1 (white dots correspond to adhesion
events) evidence a very good correlation between lattice
connection regions and PIEZO1 localization. Adhesive pixels
localized at a distance <20 nm, representing the size of the
PEG−PIEZO1-antibody complex tethered to the AFM tip
apex, were marked with a red dot (Figure 4M). Our analysis
shows that among the total number of PIEZO1 ion channels
detected in our FD-AFM experiments, around 50% overlap
with connection regions within the hexagonal lattice
representing the cytoskeletal network.
We further examined unbinding events in adhesion maps of
control and Yoda1 treated RBCs in terms of nearest neighbor
distributions (see SI Methods). Histograms of nearest
neighbor distances (Figure 4N,O) depict a multimodal
distribution with peaks centered at 69 ± 7, 119 ± 17, 178 ±
47, and 254 ± 38 nm for control RBCs. Similar distances
between probed PIEZO1 are found on Yoda1 treated RBCs at
70 ± 16, 123 ± 18, 175 ± 46, and 255 ± 34 nm. Remarkably,
our measured distances match well recent 3D-STORM
experiments, where the calculated edge lengths for the
cytoskeletal meshwork in RBCs is of ∼80 nm.50 Given the
distances between the probed ion channels, we confirm that
PIEZO1 follows a meshwork distribution similar to the
spectrin skeleton. Altogether, our data could suggest that
PIEZO1 is tethered to spectrin-Band3 complexes.
For the first time, we spatially solved the distribution of
PIEZO1 with high lateral resolution in native RBCs’ plasma
membrane. We observe an increased abundance of PIEZO1
clusters in areas of higher membrane tension upon Yoda1
activation. High-resolution FD-AFM, supported by colocaliza￾tion experiments using CLSM, evidence a cytoskeleton-like
organization of PIEZO1, adopting a pattern similar to the
spectrin mesh, and colocalizing with Band3 integral proteins.
These results indicate that membrane curvature and tension
contribute to the regulation of PIEZO1 cluster localization,
dynamics, and accessibility, while cytoskeleton anchoring could
have a role in reducing diffusion and stabilizing cluster
formation. Together, the interplay between PIEZO1 local￾ization, membrane tension, and cytoskeleton organization
support the coexistence of the force-from-lipid and the force￾from-filament gating models, even if the mechanism behind the
role of the cytoskeleton remains to be demonstrated. Our
study opens new avenues for the understanding of PIEZO1-
mediated mechanotransduction processes in living cells.
■ ASSOCIATED CONTENT
*sı Supporting Information
The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.nanolett.1c00599.

Additional details on the experimental methods, data
analysis, and Figures S1−S6 (PDF)
■ AUTHOR INFORMATION
Corresponding Authors
Donatienne Tyteca − de Duve Institute, Université Catholique
de Louvain, Brussels 1200, Belgium;
Email: [email protected]
David Alsteens − Louvain Institute of Biomolecular Science
and Technology, Université Catholique de Louvain, Louvain￾la-Neuve 1348, Belgium; orcid.org/0000-0001-9229-
113X; Email: [email protected]
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4956
Authors
Andra C. Dumitru − Louvain Institute of Biomolecular
Science and Technology, Université Catholique de Louvain,
Louvain-la-Neuve 1348, Belgium; orcid.org/0000-0003-
3574-0992
Amaury Stommen − de Duve Institute, Université Catholique
de Louvain, Brussels 1200, Belgium; orcid.org/0000-
0003-4761-8521
Melanie Koehler − Louvain Institute of Biomolecular Science
and Technology, Université Catholique de Louvain, Louvain￾la-Neuve 1348, Belgium; orcid.org/0000-0003-3042-
1749
Anne-Sophie Cloos − de Duve Institute, Université Catholique
de Louvain, Brussels 1200, Belgium
Jinsung Yang − Louvain Institute of Biomolecular Science and
Technology, Université Catholique de Louvain, Louvain-la￾Neuve 1348, Belgium
Arnaud Leclercqz − Louvain Institute of Biomolecular Science
and Technology, Université Catholique de Louvain, Louvain￾la-Neuve 1348, Belgium
Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.nanolett.1c00599

Author Contributions
§
A.C.D. and A.S. contributed equally.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Funding
This work was supported by the Université Catholique de
Louvain, the Fonds National de la Recherche Scientifique
(F.R.S-FNRS; Grant PDR T.0070.16 to D.A.) and the
Research Department of the Communauté Française de
Belgique (Concerted Research Action). D.A and D.T. are
Research Associate of the F.R.S.-FNRS. A.C.D and M.K are
postdoctoral fellows of the F.R.S-FNRS.
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
We thank Dr. Martin Delguste (Université Catholique de
Louvain, Belgium) for providing anti-gH/gL antibodies.
A were created in
Adobe Illustrator with support of BioRender.com.
■ REFERENCES
(1) Cahalan, S. M.; Lukacs, V.; Ranade, S. S.; Chien, S.; Bandell, M.;
Patapoutian, A. Piezo1 links mechanical forces to red blood cell
volume. eLife 2015, 4, e07370.
(2) Li, J.; Hou, B.; Tumova, S.; Muraki, K.; Bruns, A.; Ludlow, M. J.;
Sedo, A.; Hyman, A. J.; McKeown, L.; Young, R. S.; Yuldasheva, N.
Y.; Majeed, Y.; Wilson, L. A.; Rode, B.; Bailey, M. A.; Kim, H. R.; Fu,
Z.; Carter, D. A. L.; Bilton, J.; Imrie, H.; Ajuh, P.; Dear, T. N.;
Cubbon, R. M.; Kearney, M. T.; Prasad, K. R.; Evans, P. C.;
Ainscough, J. F. X.; Beech, D. J. Piezo1 integration of vascular
architecture with physiological force. Nature 2014, 515 (7526), 279−
282.
(3) Kefauver, J. M.; Ward, A. B.; Patapoutian, A. Discoveries in
structure and physiology of mechanically activated ion channels.
Nature 2020, 587 (7835), 567−576.
(4) Lukacs, V.; Mathur, J.; Mao, R.; Bayrak-Toydemir, P.; Procter,
M.; Cahalan, S. M.; Kim, H. J.; Bandell, M.; Longo, N.; Day, R. W.;
Stevenson, D. A.; Patapoutian, A.; Krock, B. L. Impaired PIEZO1
function in patients with a novel autosomal recessive congenital
lymphatic dysplasia. Nat. Commun. 2015, 6 (1), 8329.
(5) Zarychanski, R.; Schulz, V. P.; Houston, B. L.; Maksimova, Y.;
Houston, D. S.; Smith, B.; Rinehart, J.; Gallagher, P. G. Mutations in
the mechanotransduction protein PIEZO1 are associated with
hereditary xerocytosis. Blood 2012, 120 (9), 1908−1915.
(6) Ma, S.; Cahalan, S.; LaMonte, G.; Grubaugh, N. D.; Zeng, W.;
Murthy, S. E.; Paytas, E.; Gamini, R.; Lukacs, V.; Whitwam, T.; Loud,
M.; Lohia, R.; Berry, L.; Khan, S. M.; Janse, C. J.; Bandell, M.;
Schmedt, C.; Wengelnik, K.; Su, A. I.; Honore, E.; Winzeler, E. A.;
Andersen, K. G.; Patapoutian, A. Common PIEZO1 allele in African
populations causes RBC dehydration and attenuates Plasmodium
infection. Cell 2018, 173 (2), 443.
(7) Albuisson, J.; Murthy, S. E.; Bandell, M.; Coste, B.; Louis-dit￾Picard, H.; Mathur, J.; Fénéant-Thibault, M.; Tertian, G.; de
Jaureguiberry, J.-P.; Syfuss, P.-Y.; Cahalan, S.; Garcon, L.; Toutain, ̧
F.; Simon Rohrlich, P.; Delaunay, J.; Picard, V.; Jeunemaitre, X.;
Patapoutian, A. Dehydrated hereditary stomatocytosis linked to gain￾of-function mutations in mechanically activated PIEZO1 ion
channels. Nat. Commun. 2013, 4 (1), 1884.
(8) Saotome, K.; Murthy, S. E.; Kefauver, J. M.; Whitwam, T.;
Patapoutian, A.; Ward, A. B. Structure of the mechanically activated
ion channel Piezo1. Nature 2018, 554 (7693), 481−486.
(9) Zhao, Q.; Zhou, H.; Chi, S.; Wang, Y.; Wang, J.; Geng, J.; Wu,
K.; Liu, W.; Zhang, T.; Dong, M.-Q.; Wang, J.; Li, X.; Xiao, B.
Structure and mechanogating mechanism of the Piezo1 channel.
Nature 2018, 554 (7693), 487−492.
(10) Ge, J.; Li, W.; Zhao, Q.; Li, N.; Chen, M.; Zhi, P.; Li, R.; Gao,
N.; Xiao, B.; Yang, M. Architecture of the mammalian mechano￾sensitive Piezo1 channel. Nature 2015, 527 (7576), 64−69.
(11) Guo, Y. R.; MacKinnon, R. Structure-based membrane dome
mechanism for Piezo mechanosensitivity. eLife 2017, 6, e33660.
(12) Botello-Smith, W. M.; Jiang, W.; Zhang, H.; Ozkan, A. D.; Lin,
Y.-C.; Pham, C. N.; Lacroix, J. J.; Luo, Y. A mechanism for the
activation of the mechanosensitive Piezo1 channel by the small
molecule Yoda1. Nat. Commun. 2019, 10 (1), 4503.
(13) Wang, Y.; Chi, S.; Guo, H.; Li, G.; Wang, L.; Zhao, Q.; Rao, Y.;
Zu, L.; He, W.; Xiao, B. A lever-like transduction pathway for long￾distance chemical- and mechano-gating of the mechanosensitive
Piezo1 channel. Nat. Commun. 2018, 9 (1), 1300.
(14) Coste, B.; Xiao, B.; Santos, J. S.; Syeda, R.; Grandl, J.; Spencer,
K. S.; Kim, S. E.; Schmidt, M.; Mathur, J.; Dubin, A. E.; Montal, M.;
Patapoutian, A. Piezo proteins are pore-forming subunits of
mechanically activated channels. Nature 2012, 483 (7388), 176−181.
(15) Syeda, R.; Xu, J.; Dubin, A. E.; Coste, B.; Mathur, J.; Huynh, T.;
Matzen, J.; Lao, J.; Tully, D. C.; Engels, I. H.; Petrassi, H. M.;
Schumacher, A. M.; Montal, M.; Bandell, M.; Patapoutian, A.
Chemical activation of the mechanotransduction channel Piezo1.
eLife 2015, 4, e07369.
(16) Lewis, A. H.; Grandl, J. Mechanical sensitivity of Piezo1 ion
channels can be tuned by cellular membrane tension. eLife 2015, 4,
e12088.
(17) Cox, C. D.; Bae, C.; Ziegler, L.; Hartley, S.; Nikolova-Krstevski,
V.; Rohde, P. R.; Ng, C.-A.; Sachs, F.; Gottlieb, P. A.; Martinac, B.
Removal of the mechanoprotective influence of the cytoskeleton
reveals PIEZO1 is gated by bilayer tension. Nat. Commun. 2016, 7
(1), 10366.
(18) Kung, C. A possible unifying principle for mechanosensation.
Nature 2005, 436 (7051), 647−654.
(19) Teng, J.; Loukin, S.; Anishkin, A.; Kung, C. The force-from￾lipid (FFL) principle of mechanosensitivity, at large and in elements.
Pfluegers Arch. 2015, 467 (1), 27−37.
(20) Martinac, B. The ion channels to cytoskeleton connection as
potential mechanism of mechanosensitivity. Biochim. Biophys. Acta,
Biomembr. 2014, 1838 (2), 682−691.
(21) Cox, C. D.; Bavi, N.; Martinac, B. Origin of the Force: The
Force-From-Lipids Principle Applied to Piezo Channels. In Curr. Top.
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4957
Membr.; Gottlieb, P. A., Ed.; Current Topics in Membranes: Piezo
Channels, Academic Press: 2017; Vol. 79, Chapter 3, pp 59−96.
(22) Gillespie, P. G.; Walker, R. G. Molecular basis of
mechanosensory transduction. Nature 2001, 413 (6852), 194−202.
(23) Markin, V. S.; Hudspeth, A. J. Gating-Spring Models of
Mechanoelectrical Transduction by Hair Cells of the Internal Ear.
Annu. Rev. Biophys. Biomol. Struct. 1995, 24 (1), 59−83.
(24) Nourse, J. L.; Pathak, M. M. How cells channel their stress:
Interplay between Piezo1 and the cytoskeleton. Semin. Cell Dev. Biol.
2017, 71, 3−12.
(25) Bavi, N.; Richardson, J.; Heu, C.; Martinac, B.; Poole, K.
PIEZO1-Mediated Currents Are Modulated by Substrate Mechanics.
ACS Nano 2019, 13 (11), 13545−13559.
(26) Romero, L. O.; Caires, R.; Nickolls, A. R.; Chesler, A. T.;
Cordero-Morales, J. F.; Vásquez, V. A dietary fatty acid counteracts
neuronal mechanical sensitization. Nat. Commun. 2020, 11 (1), 2997.
(27) Coste, B.; Mathur, J.; Schmidt, M.; Earley, T. J.; Ranade, S.;
Petrus, M. J.; Dubin, A. E.; Patapoutian, A. Piezo1 and Piezo2 Are
Essential Components of Distinct Mechanically Activated Cation
Channels. Science 2010, 330 (6000), 55−60.
(28) Ellefsen, K. L.; Holt, J. R.; Chang, A. C.; Nourse, J. L.; Arulmoli,
J.; Mekhdjian, A. H.; Abuwarda, H.; Tombola, F.; Flanagan, L. A.;
Dunn, A. R.; Parker, I.; Pathak, M. M. Myosin-II mediated traction
forces evoke localized Piezo1-dependent Ca2+ flickers. Commun. Biol.
2019, 2 (1), 298.
(29) Pathak, M. M.; Nourse, J. L.; Tran, T.; Hwe, J.; Arulmoli, J.; Le,
D. T. T.; Bernardis, E.; Flanagan, L. A.; Tombola, F. Stretch-activated
ion channel Piezo1 directs lineage choice in human neural stem cells.
Proc. Natl. Acad. Sci. U. S. A. 2014, 111 (45), 16148−16153.
(30) Gudipaty, S. A.; Lindblom, J.; Loftus, P. D.; Redd, M. J.; Edes,
K.; Davey, C. F.; Krishnegowda, V.; Rosenblatt, J. Mechanical stretch
triggers rapid epithelial cell division through Piezo1. Nature 2017, 543
(7643), 118−121.
(31) Yao, M.; Tijore, A.; Cox, C. D.; Hariharan, A.; Van Nhieu, G.
T.; Martinac, B.; Sheetz, M. Force-dependent Piezo1 recruitment to
focal adhesions regulates adhesion maturation and turnover
specifically in non-transformed cells. bioRxiv 2020,
2020.03.09.972307 (accessed on May 14, 2021.).
(32) Dumitru, A. C.; Poncin, M. A.; Conrard, L.; Dufrene, Y. F.; ̂
Tyteca, D.; Alsteens, D. Nanoscale membrane architecture of healthy
and pathological red blood cells. Nanoscale Horiz. 2018, 3 (3), 293−
304.
(33) Evans, E.; Ritchie, K. Dynamic strength of molecular adhesion
bonds. Biophys. J. 1997, 72 (4), 1541−1555.
(34) Müller, D. J.; Dumitru, A. C.; Lo Giudice, C.; Gaub, H. E.;
Hinterdorfer, P.; Hummer, G.; De Yoreo, J. J.; Dufrene, Y. F.; ̂
Alsteens, D. Atomic Force Microscopy-Based Force Spectroscopy and
Multiparametric Imaging of Biomolecular and Cellular Systems.
Chem. Rev. 2020.
(35) Williams, P. M. Analytical descriptions of dynamic force
spectroscopy: behaviour of multiple connections. Anal. Chim. Acta
2003, 479 (1), 107−115.
(36) Lacroix, J. J.; Botello-Smith, W. M.; Luo, Y. Probing the gating
mechanism of the mechanosensitive channel Piezo1 with the small
molecule Yoda1. Nat. Commun. 2018, 9 (1), 2029.
(37) Xiao, B. Levering Mechanically Activated Piezo Channels for
Potential Pharmacological Intervention. Annu. Rev. Pharmacol.
Toxicol. 2020, 60 (1), 195−218.
(38) Tsuchiya, M.; Hara, Y.; Okuda, M.; Itoh, K.; Nishioka, R.;
Shiomi, A.; Nagao, K.; Mori, M.; Mori, Y.; Ikenouchi, J.; Suzuki, R.;
Tanaka, M.; Ohwada, T.; Aoki, J.; Kanagawa, M.; Toda, T.; Nagata,
Y.; Matsuda, R.; Takayama, Y.; Tominaga, M.; Umeda, M. Cell
surface flip-flop of phosphatidylserine is critical for PIEZO1-mediated
myotube formation. Nat. Commun. 2018, 9 (1), 2049.
(39) Bae, C.; Gnanasambandam, R.; Nicolai, C.; Sachs, F.; Gottlieb,
P. A. Xerocytosis is caused by mutations that alter the kinetics of the
mechanosensitive channel PIEZO1. Proc. Natl. Acad. Sci. U. S. A.
2013, 110 (12), E1162−E1168.
(40) Ridone, P.; Pandzic, E.; Vassalli, M.; Cox, C. D.; Macmillan, A.;
Gottlieb, P. A.; Martinac, B. Disruption of membrane cholesterol
organization impairs the activity of PIEZO1 channel clusters. J. Gen.
Physiol. 2020, 152 (8), e201912515.
(41) Cox, C. D.; Gottlieb, P. A. Amphipathic molecules modulate
PIEZO1 activity. Biochem. Soc. Trans. 2019, 47 (6), 1833−1842.
(42) Svetina, S.; Š
velc Kebe, T.; Bozič , B. A Model of Piezo1-Based ̌
Regulation of Red Blood Cell Volume. Biophys. J. 2019, 116 (1),
151−164.
(43) Alimohamadi, H.; Smith, A. S.; Nowak, R. B.; Fowler, V. M.;
Rangamani, P. Non-uniform distribution of myosin-mediated forces
governs red blood cell membrane curvature through tension
modulation. PLoS Comput. Biol. 2020, 16 (5), e1007890.
(44) Nans, A.; Mohandas, N.; Stokes; David, L. Native Ultra￾structure of the Red Cell Cytoskeleton by Cryo-Electron Tomog￾raphy. Biophys. J. 2011, 101 (10), 2341−2350.
(45) Fowler, V. M. The Human Erythrocyte Plasma Membrane: A
Rosetta Stone for Decoding Membrane−Cytoskeleton Structure. In
Curr. Top. Membr.; Bennett, V., Ed.; Current Topics in Membranes:
Functional Organization of Vertebrate Plasma Membrane, Academic
Press, 2013; Vol. 72, Chapter 2, pp 39−88.
(46) Alenghat, F. J.; Golan, D. E. Membrane Protein Dynamics and
Functional Implications in Mammalian Cells. In Curr. Top. Membr.;
Bennett, V., Ed.; Current Topics in Membranes: Functional Organ￾ization of Vertebrate Plasma Membrane, Academic Press, 2013; Vol. 72,
Chapter 3, pp 89−120.
(47) Byers, T. J.; Branton, D. Visualization of the protein
associations in the erythrocyte membrane skeleton. Proc. Natl. Acad.
Sci. U. S. A. 1985, 82 (18), 6153−6157.
(48) Liu, S. C.; Derick, L. H.; Palek, J. Visualization of the hexagonal
lattice in the erythrocyte membrane skeleton. J. Cell Biol. 1987, 104
(3), 527−536.
(49) Lux, S. E., IV Anatomy of the red cell membrane skeleton:
unanswered questions. Blood 2016, 127 (2), 187−199.
(50) Pan, L.; Yan, R.; Li, W.; Xu, K. Super-Resolution Microscopy
Reveals the Native Ultrastructure of the Erythrocyte Cytoskeleton.
Cell Rep. 2018, 22 (5), 1151−1158.
(51) Wang, J.; Jiang, J.; Yang, X.; Wang, L.; Xiao, B., Tethering Piezo
channels to the actin cytoskeleton for mechanogating via the E￾cadherin-β-catenin mechanotransduction complex. bioRxiv 2020,
2020.05.12.092148.
Nano Letters pubs.acs.org/NanoLett Letter

https://doi.org/10.1021/acs.nanolett.1c00599

Nano Lett. 2021, 21, 4950−4958
4958