Article https://doi.org/10.1038/s41467-023-41981-5
MYH10 activation rescues contractile defects
in arrhythmogenic cardiomyopathy (ACM)
Nieves García-Quintáns1, Silvia Sacristán1, Cristina Márquez-López1,
Cristina Sánchez-Ramos 1, FernandoMartinez-de-Benito 1,2, David Siniscalco3,
Andrés González-Guerra 1, Emilio Camafeita 1,2, Marta Roche-Molina1,
Mariya Lytvyn1, David Morera1, María I. Guillen1, María A. Sanguino 1,
David Sanz-Rosa 1,2,4, Daniel Martín-Pérez1, Ricardo Garcia 3 &
Juan A. Bernal 1,2
The most prevalent genetic form of inherited arrhythmogenic cardiomyo-
pathy (ACM) is caused by mutations in desmosomal plakophilin-2 (PKP2). By
studying pathogenic deletion mutations in the desmosomal protein PKP2,
here we identify a general mechanism by which PKP2 delocalization restricts
actomyosin network organization and cardiac sarcomeric contraction in this
untreatable disease. Computationalmodeling of PKP2 variants reveals that the
carboxy-terminal (CT) domain is required for N-terminal domain stabilization,
which determines PKP2 cortical localization and function. Inmutant PKP2 cells
the expression of the interacting protein MYH10 rescues actomyosin dis-
organization. Conversely, dominant-negative MYH10 mutant expression
mimics the pathogenic CT–deletion PKP2mutant causing actin network
abnormalities and right ventricle systolic dysfunction. A chemical activator of
non-muscle myosins, 4-hydroxyacetophenone (4-HAP), also restores normal
contractility. Our findings demonstrate that activation of MYH10 corrects the
deleterious effect of PKP2 mutant over systolic cardiac contraction, with
potential implications for ACM therapy.
The most frequent genetic alterations in arrhythmogenic cardiomyo-
pathy (ACM) affect the desmosomal gene PKP2, accounting for
40–60% of genotype-positive patients1–3. ACM is a genetic disease of
the heart muscle4,5 that predisposes to sudden cardiac death (SCD)6,
particularly in young patients and athletes7,8. To date all therapeutic
interventions offered to these patients are palliative and not curative.
The potential therapy for ACM will require new knowledge about the
molecularmechanisms that drive the pathological development of the
disease.
ACM, for many years known as arrhythmogenic right ventricular
cardiomyopathy (ARVC)9, shows autosomal-dominant inheritance of
desmosomal proteins and usually manifests as impaired function of
the right ventricle (RV)10. Desmosomes are dynamic intercellular
junctions thatmaintain the structural integrity of skin andheart tissues
by withstanding shear forces, linking the cellular edge to the internal
cytoskeleton11,12. Actin is one of the major cytoskeletal proteins in
eukaryotic cells andplays anessential role in several cellularprocesses,
including mechano-resistance and contractile force generation13.
Defective regulation of the organization of actin filaments in sarco-
meres, owing to genetic mutations or deregulated expression of
cytoskeletal proteins, is a hallmark of many heart and skeletal muscle
disorders14.
The ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/),
part of the NCBI Entrez system, attempts to establish relationships
between gene variants and phenotype. Remarkably, all nonsense PKP2
variants are classified as pathogenic or likely pathogenic. Two
Received: 2 March 2023
Accepted: 26 September 2023
Check for updates
1CentroNacional de InvestigacionesCardiovasculares (CNIC),Madrid, Spain. 2CIBERde EnfermedadesCardiovasculares (CIBERCV),Madrid, Spain. 3Materials
Science Factory, Instituto de Ciencia deMateriales deMadrid (ICMM), CSIC, Madrid, Spain. 4Universidad Europea, Madrid, Spain. e-mail: jabernal@cnic.es
Nature Communications |         (2023) 14:6461 1
12
34
56
78
9
0
()
:,;
12
34
56
78
9
0
()
:,;
mechanisms can drive pathogenesis, on one hand, mRNAs containing
premature stop codons may be eliminated by a surveillance pathway
called nonsense-mediatedmRNAdecay that results in the reduction of
mRNA levels and the haploinsufficient phenotype. Instead, translation
of these aberrant mRNAs may lead to deleterious gain-of-function
(GOF) or dominant negative activity of the resulting proteins. We
previously demonstrated that PKP2 (c.2203C >T), encoding the
p.R735* mutant, which creates a premature stop codon and a
C-terminal truncated PKP2 protein and is found in ACM patients from
different families4,15, can operate as a GOF disease mutant16, under-
lining the functional importance of the PKP2 C-terminal domain. In
general, pathogenic mutations in PKP2 have usually been associated
with the classical form of the disease that predominantly includes
anomalous electrocardiograms,with structural abnormalities that lead
to a progressive global RV contractile dysfunction1.
Here we identified MYH10 actomyosin component as key dif-
ferential interactor of PKP2 cardiac isoform using biochemical and
mass spectrometry approaches. Themain role ofMYH10 protein is to
orchestrate the mechanoenzymatic properties of stress fibers.
Myosins execute numerous mechanical tasks in cells, including spa-
tiotemporal organization of the actin cytoskeleton, adhesion,
migration, cytokinesis, tissue remodeling, and membrane
trafficking17–21. Using in vitro and in vivo models, we restore actin
network and cardiac contraction by stimulating MYH10 activity in an
ACM mouse model of disease. We show that localization of PKP2
C-terminal deletion mutants found in ACM patients from different
families4,15 alters actomyosin organization and function. We dis-
covered that PKP2 mutants´ expression in cardiac cells led to func-
tional alterations in MYH10 what is key to developing an ACM
contractile phenotype. In these PKP2mutants a rearrangement of the
F-actin cytoskeleton pairedwith sarcomere dysfunction. Our findings
illustrate how expression of mutant MYH10 disrupts actin network
distribution and compromises RV cardiac systolic contraction in
mice following a similar mechanism than p.R735* mutant when
inducing ACM. Finally, we demonstrate that activation of non-muscle
myosins with 4-hydroxyacetophenone (4-HAP) attenuates contrac-
tion disfunction observed in ACM animals.
Results
PKP2 interacts with actomyosin proteins
We have previously shown that C-terminal PKP2 deletion mutant
(c.2203C > T), encoding the PKP2-p.R735* protein operates as a GOF
protein in arrhythmogenic cardiomyopathy (ACM) compromising RV
function16, although functional mechanistic details remain unre-
solved. To gain insight into the molecular basis that induce the
contractile defects we generated and compared wild-type and
mutant PKP2 protein interactomes. We initially identified cardiac
specific PKP2a22 binding partners by pull-down of Halo-tagged PKP2
proteins followed bymass spectrometry (MS) analysis. PKP2-HaloTag
pull-down products included previously identified interacting part-
ners such as the desmosomal components, plakoglobin (JUP), des-
moplakin (DSP) and desmoglein (DGS) together with unknow
interacting partners like cytoskeletal actin proteins (ACTA2 and
ACTB), or myosin 9 (MYH9) and myosin 10 (MYH10), also known as
non-muscle myosins NMIIA and NMIIB (Fig. 1a). To validate the
results obtained using HaloTag-PKP2 variants we also performed an
orthogonal immunoprecipitation (EGFP-tag) assay followed by MS
analysis. In Fig. 1, we show a list of proteins identified by both inde-
pendent techniques, confirming desmosomal and cytoskeletal pro-
teins as strong PKP2 interactors. Finally, immunoprecipitations
followed by western blot assays confirmed a differential interaction
with actomyosin protein MYH10 after expressing EGFP-PKP2 or EGFP-
PKP2-p.R735* in cells (Fig. 1b, c) validating actomyosin components as
the major differential interactors between wild-type PKP2 and
mutant p.R735* proteins.
Mutant PKP2 modifies structure of actin filament network
Since active MYH10 binds actin filaments and bundles them to form
stressfibers, the crosslinkingof actinfilaments byMYH10 increases the
overall stiffness of actin networks.We hypothesized that PKP2-p.R735*
protein modifies actomyosins organization and function, altering
F-actin fiber organization and decreasing cellular resistance to defor-
mation (stiffness). We measured cell stiffness in intact HL-1 cardiac
cells by acquiring force-volume maps by atomic force microscopy
(AFM). Although HL-1 cells derived from atrial cardiomyocytes, they
are considered a valuable model to study basic biology of contractile
differentiated cardiac cells23. The force maps obtained were used to
represent spatial variation in Young’s modulus as a measure of local
cell elasticity24 in cell lines stably encoding PKP2 or mutant PKP2-
p.R735*. The nanomechanical maps show that Young’s modulus was
higher in regions with a high density of branched actin cytoskeleton
networks and stress fibers (Fig. 1d). These observations were quanti-
fied by statistical analysis of Young’s modulus maps using a bottom-
effect correction expression25 (Supplementary Fig. 1). The graphs show
that the median value of the Young’s modulus measured in PKP2 cells
was higher than that measured in PKP2-p.R735* cells (Fig. 1d). These
results support the observation that PKP2-p.R735* cells lack most
stress and F-actin fibers.
Biochemically, cells expressing the PKP2-p.R735* mutant had a
skewed F-actin:G-actin ratio compared with control PKP2 cells, indi-
cating a shift toward decreased actin polymerization (Fig. 2a). Images
also revealed a scarcity of organized F-actin filaments and sarcomere-
like structures in the cells encoding the PKP2-p.R735*mutant (Fig. 2b).
F-actin filament length quantification using images from cytoplasmic
area of HL-1 cells revealed a reduction in length of actin filaments in
cells expressing the PKP2-p.R735* mutant compared with the PKP2
controls (Supplementary Fig. 2). We also explored the effect of PKP2
and the PKP2-p.R735* mutant proteins on the actin organization of
mouse neonatal cardiomyocytes (MNC). The percentage of MNC dis-
playing an organized sarcomeric network varied between transduced
cardiomyocytes, with a larger proportion of PKP2 cells showing
regions with established sarcomeres compared to mutant PKP2-
p.R735* transduced MNC 69.97 ± 6.5% in PKP2 vs. 30.02 ± 6.1% in
PKP2-p.R735*; n = 54–79; Fig. 2c). To quantify sarcomeric structural
configuration in MNC, we used fast Fourier transform of cardiomyo-
cyte images after phalloidin staining26. The regularity of sarcomeric
striations within the cardiomyocytes resulted in the appearance of
peaks in the Fourier spectrum of the image. The frequency of the peak
was used to determine the sarcomere period, that was the same
(1,78μm) in both groups (Fig. 2d), and the amplitude of this peak, the
sarcomeric power, which can be considered the best indicator of the
sarcomere organization level (73.08 ± 12.69 Arb.Units in PKP2 vs.
61.92 ± 6.47 Arb.Units in PKP2-p.R735*; n = 10–19; Fig. 2e). The higher
the sarcomeric power value, the greater the sarcomere organization.
We also calculated the general organization level (coherency) of the
actin fiber network in these PKP2 and PKP2-p.R735* cardiomyocytes
(Fig. 2f). This parameter of the actin fibers was analyzed to verify the
dominant local orientationon the images fromPKP2 and PKP2-p.R735*
cells using the ImageJ plugin OrientationJ27 (Supplementary Fig. 3).
Images revealed that coherency in control cells was higher compared
to the organization found in the actin fibers in PKP2-p.R735* cells, that
showsamostly randomorganizationwith values close to0 (0.37 ± 0.23
Arb.Units in PKP2 vs. 0.13 ± 0.08Arb.Units in PKP2-p.R735*; n = 78–98).
Higher coherency values observed in control PKP2 MNC reflected that
the direction of the actin fibers in space were forming oriented
structures (Fig. 2b, f and Supplementary. 3).
Transverse maximal projections of z-stack images in cells
expressing PKP2 confirmed the organization of thick actin filament
bundles into a curved structure over the nucleus. In contrast, both in
vitro models expressing PKP2-p.R735* lacked this structure and the
nucleus was close to the external plasma membrane (Supplementary
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 2
Fig. 4). Thiswas reflected in the abnormal height ofmost PKP2-p.R735*
cells, which were shorter than PKP2 cells (5.41 ± 0.24 μm in PKP2 MNC
vs. 4.71 ± 0.21 μm in PKP2-p.R735* MNC; n = 22; 4.50 ±0.22 μm in PKP2
HL-1 cells vs. 3.73 ± 0.27 μm in PKP2-p.R735* HL-1 cells; n = 13) (Sup-
plementary Fig. 4). These data show that the higher absolute levels of
F-actin and stress-fiber formation in PKP2 cells are associated with
greater cell stiffness, whereas the lower cell stiffness inmutant p.R735*
correlates with a reduction in F-actin and stress fiber disassembly.
To test whether PKP2-p.R735* functions throughMyh10 to disturb
actin filament network, we coexpressed PKP2-p.R735* and functional
EGFP-MYH10 genes in HL-1 andMNC. Confocal imaging confirmed that
cells transfected with wild-type EGFP-MYH10 significantly improved
actin network, reestablishing orientated actin structures (coherency)
in cells expressing the deleterious PKP2-p.R735*mutant (Fig. 3, 3b and
Supplementary Fig. 5). We also detected that expression of MYH10
GOF mutant p.R577* fused to EGFP changed the actin cytoskeleton
organization and distribution (Fig. 3c, d). Compared to the diffused
staining observedwhenmutant EGFP-MYH10-p.R577* is present, EGFP-
Myh10 showed an organized distribution along the F-actin fibers
(Supplementary Fig. 6). Taken together these results support a model
where differential protein-protein interaction in PKP2-p.R735* with
MYH10 alters actomyosin network organization and function.
PKP2 C-terminal deletion modifies its subcellular localization
and defines actomyosin function
As subcellular localization is essential to protein function, we studied
the effect of C-terminal deletion of PKP2 in protein spatial distribution
in MNC and HL-1 cells. Confocal microscopy revealed that most of the
PKP2-p.R735* mutant signal was located in the cytosol, displaced from
the cell edge (Figs. 2b, 4a, 4b and Supplementary. 7). Images and
quantification analysis confirmed that PKP2-p.R735* mutant was pre-
sent at lower levels by the plasmamembrane relative to the cytoplasm
(Fig. 4a, b). After transfection with pCAG-PKP2, pCAG-PKP2-p.R735*, or
both plasmids, images and immunoblot analysis of cell fractions
demonstrated that wild-type PKP2 associates with the plasma mem-
brane, whereas the PKP2-p.R735* mutant is predominantly detected in
the cytoplasm and only occasionally at the membrane fraction
(Fig. 4c). Confocal images fromMNCencoding PKP2 and PKP2-p.R735*
tagged protein (fused to tdTomato or EGFP respectively) showed that
PKP2was localizedmainly in cell-cell border at the desmosomal plaque
c
a
d
EGFP IP
Halotag pulldown
JUP
DSG1
DSP
ACTB
ACTA2
MYH10
MYH9
MYH11
MYL12B
MYL6B
TPM1
POTEKP
PKP2 p.R735*
D
es
m
os
om
al
Pr
ot
ei
ns
C
yt
os
ke
le
ta
l
Pr
ot
ei
ns
0
5
10
- -
- -+
+
+
+
PKP2 p.R735*
0
5
Yo
un
g´
s 
m
od
ul
us
 (K
Pa
)
10
15 p<0.0001
Force volumen map
PKP2
p.R735*
5m
16
12.8
9.6
3.4
3.2
0
EG
FP-
PKP
2
EG
FP-
p.R
735
*
0.0
0.2
0.4
0.6
0.8
1.0
1.2
R
el
at
iv
e
M
YH
10
/P
KP
2
(A
rb
. U
ni
ts
)
Pr
es
su
re
 (K
Pa
)
b
IPInput
100 
75 
25
EGFP-PKP2
EGFP
250MYH10
EGFP-PKP2
EGFP-p.R735*
EGFP 
+ - - + - -
-
-
+
-
-
+
-
-
+
-
-
+
MW 
(kDa)
EGFP-p.R735*
p=0.0069
Fig. 1 | PKP2 interacts with actomyosin proteins and alters actin cytoskeleton.
a Unique peptide counting of most abundant proteins identified in the orthogonal
HaloTag pulldown and EGFP immunoprecipitation assays. The indicated EGFP- or
Halo-tagged proteins were stably expressed in HEK293T cells and subjected to pull-
down or immunoprecipitationwithHalo-Link resin orα-EGFP antibodies respectively,
prior to proteomic analysis. The enrichment of cytoskeletal proteins in mutant
p.R.735* relative to control PKP2 cells is shown, box displaying the greatest enrich-
ment of proteins known to form the cytoskeleton are indicated in black.
bRepresentativewestern blot of proteins co-immunoprecipitatedwith EGFP (28KDa),
EGFP-tagged wild-type PKP2 (120KDa), or EGFP-tagged PKP2-p.R735* mutant
(104KDa) proteins. Blots show PKP2 wild-type and mutant versions, MYH10 or EGFP.
Blots from input samples are also shown. c Quantification of MYH10 immunopreci-
pitation using PKP2 andPKP2-p.R735* fusion EGFPproteins as bait. Data arepresented
as mean± sem; n=3 and show relative amount of MYH10 to PKP2 or R735* in the
immunoprecipitation assays. d Illustrative atomic force microscopy (AFM) image of
HL1 cells stably expressing PKP2 or mutant p.R735* genes. Scale bars, 5 µm. Box plot
showing quantification of Young’s modulus values obtained by fitting force-distance
curves. Data are presented from three independent experiments as median with
minimum to maximum values (box limits); n=8080 in PKP2 cells and n=93134 in
PKP2-p.R735* cells, Significance was established as p<0.05, (Mann–Whitney test).
Arb. Units arbitrary units, kPa kiloPascals, MW molecular weight, kDa kiloDalton.
Source data are provided as a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 3
following a linear pattern. In contrast, PKP2-p.R735* was mainly loca-
lized at the cytoplasm, and exceptionally at the cell-cell contact area.
These data shows that PKP2 protein subcellular localization is not
altered when co-expressed with mutant PKP2-p.R735* protein
(Fig. 4a–c). Consistently, imaging analysis of heart sections from mice
transduced with adeno-associated viruses (AAV) encoding PKP2 or
p.R735* EGFP N-terminal fusion protein confirmed the differential
cytoplasmic localization of mutant protein (Fig. 4d). These results
demonstrate that the C-terminal region of PKP2 is important for
proper protein localization at the border contact region in vitro and
in vivo. Also, highlight that PKP2 subcellular localization at the des-
mosome is not altered when the wild-type PKP2 protein is co-
expressed with mutant PKP2-p.R735*. These data suggest that the
GOF mutant PKP2 has a additional activity in the cytosol, as such
function is not mediated by a deregulation of the levels or proper
localization of wild-type PKP2.
In AAV-PKP2-p.R735* animals left ventricle (LV) function is com-
parable to controls expressing PKP2 although right ventricle end sys-
tolic volume (RV-ESV) was significantly larger in AAV-PKP2-p.R735*mice
(28.6 ± 4.6 μL in AAV-PKP2-R735* vs. 20.0 ± 6.1 μL, and 20.6 ± 2.4 μL in
non-transduced control mice, and AAV-PKP2 mice respectively; n = 6),
whereas RV-end diastolic volume (RV-EDV) was similar in all mouse
groups. Likewise, RV-ejection fraction (RV-EF) was lower in AAV-PKP2-
p.R735* mice (57.8 ± 5.3% in AAV-PKP2-R735* mice vs. 71.1 ± 3.5%, and
69.5 ± 3.8%, in non-transduced control mice, and AAV-PKP2 mice
respectively; n =6) (Fig. 4e and Supplementary Table 1). These data
show that the cardiac expression of an ACM PKP2-related pathogenic
mutant causes PKP2 delocalization and a RV systolic dysfunction.
b
EGFP-p.R735*
5 m
F-actinEGFP-PKP2
20 m
F-actin
d e
0
50
100
150
Po
w
er
 (A
rb
. U
ni
ts
)
p<0.0001
0.6
1.2
1.8
2.4
Pe
rio
d
(
m
)
p=0.6837
EGF
P-PK
P2
EGF
P-p.R
735*
EGF
P-PK
P2
EGF
P-p.R
735*
f
0.0
0.2
0.4
0.6
0.8
1.0
C
oh
er
en
cy
 (
Ar
b.
 U
ni
ts
)
EGF
P-PK
P2
EGF
P-p.R
735*
p<0.0001
c
a
F/
G
 a
ct
in
 ra
tio
PKP2 p.R735*
3
2
1
0
G
50 
37 
F G G GF F F
PKP2 p.R735*
β-
ac
tin
MW
(kDa)
p<0.0001
0
20
40
60
80
100 p<0.0001
C
el
ls 
w
ith
 s
ar
co
m
er
es
 (%
)
EGF
P-PK
P2
EGF
P-p.R
735*
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 4
PKP2 N-terminal misfolding recovery ameliorates protein
delocalization and cardiac dysfunction
To gain insight into how the shortened mutant PKP2 protein affects
protein localization and function, we modeled and compared the
structures of wild-type PKP2 and the p.R735* mutant. The best models
with minimal energy and correct folding were selected after a final
refinement cycle to minimize clashes. Three-D views revealed a major
difference in topology and orientation of the N-terminal betweenwild-
type PKP2 andmutant p.R735* proteins, suggesting that theN-terminal
region is a disordered structure. This assumption is also supported by
analysis with disopred v3.16, which predicts secondary structure, dis-
ordered regions, and protein-binding profiles. Thus, PKP2 sequence
analysis predicts two disordered regions, corresponding to the
N-terminal and a previously described disordered inter-ARM domain.
However, potential misfolding of the PKP2 N-terminal domain is
restricted by the C-terminal region, which generates a stable structure.
Because of the lack of the C-terminal region in the p.R735*mutant, the
N-terminal domain is unable to fold properly along the central core of
ARM repeats to generate a stable structure. In this situation, the
N-terminal domain is a disordered region, able to adopt a variable
number of conformational states. The reorientation of the N-terminal
domain in the p.R735* mutant alters its topology and exposes an
internal section that remains buried in the wild-type PKP2 structure
due to the stabilizing action of the C-terminal domain (Fig. 5a). This
conformational adjustment also modifies the global electropotential
properties at the protein surface, which could explain the altered
interaction of p.R735* mutant PKP2 with partners.
Together, these findings speak to a model wherein the delocali-
zation of PKP2 is a consequence of an unstable conformation that
alters MYH10 activity to properly organize the actin filaments. To test
whetherN-terminal instability is a general featureof truncatedmutants
of PKP2 we analyzed additional C-terminal mutants reported in the
ACM ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/). We
then generated two additional PKP2 C-terminal deletion mutants
identified in familiar cases1, (2028G >A) p.W676*, and (2421C >A)
p.Y807* fused to an EGFP and expressed in MNC. The twomutants are
predicted to alter PKP2 stability and consequently its localization and
actin organization (Fig. 5b). In fact, in all mutants there are 2 dis-
ordered regions predicted from sequence corresponding to the
N-term and a previously described disordered inter-ARM region. The
short C-terminal is unable to stabilize the N-domain which is intrinsi-
cally unstable. Then, the N-terminus is unable to fold properly along
the central core ARM repeats to generate a stable structure, only
covering partially the cavity generated by the ARM-repeat domain.
Image analysis revealed that the p.W676*, p.Y807* and p.R735*
PKP2 mutants´ signal were in the cytosol, and not associated to
the plasma membrane (Fig. 5b). After cell transfection with
pCAG-EGFP-PKP2, pCAG-EGFP-PKP2-R735*, pCAG- EGFP-PKP2-p.W676*,
pCAG-EGFP-PKP2-p.Y807* confocal imaging also demonstrated that
wild-type PKP2 associatedwith the plasmamembranemaintained awell-
organized actin cytoskeleton, whereas C-terminal deletion mutants that
are predominantly detected in the cytoplasm showed a disorganized
cytoskeleton and a lack of organized F-actin filaments (Fig. 5c).
As all these C-terminal deletion mutants show a disorganized
N-terminal that exposes the internal region of PKP2, we hypothesized
that occlusion of the internal cavity with an unrelated tag could be
sufficient to partially rescue the phenotype associated with mutants´
delocalization. As a proof of concept we tested this hypothesis in the
PKP2-p.R735*mutant. We generated a PKP2-p.R735*-EGFP ‘closed’
fusion protein, inwhich the EGFP partially blocks the internal fragment
in PKP2-p.R735* and stabilizes the mutant structure (Fig. 6a). Confocal
z-stack images of transfected MNC detected PKP2-p.R735*-EGFP not
only in the cytoplasm but also at higher levels at the plasma mem-
brane, demonstrating partial rescue of proper PKP2 localization
(Fig. 6b). Quantitative analysis further proved that PKP2-p.R735*-EGFP
recovered membrane associated localization (Fig. 6c). Confocal ima-
ging of phalloidin staining in cells showed that cells encoding PKP2-
p.R735*-EGFP had a better-conformed actin cytoskeleton with more
orientated actin structures than cells expressing unmodified EGFP-
p.R735* (Fig. 6d).
We also tested whether AAV-mediated in vivo expression of the
PKP2-p.R735*-EGFP mutant in mouse cardiomyocytes would develop
the ACM phenotype associated with cardiac-specific p.R735* expres-
sion. Unlike AAV-p.R735*mutant, AAV-expressed p.R735*-EGFP did not
significantly increaseRV-ESV (24.5 ± 4.2μL in PKP2-R735*-EGFPmice vs.
21.4 ± 4.8μL in AAV-PKP2mice; n = 8–10) or significantly reduce RV-EF
(64.1 ± 5.6% in AAV-PKP2-R735*-EGFP mice vs. 68.8 ± 5.3% in AAV-PKP2
mice; n = 8–10) maintaining similar cardiac function when compared
to AAV-PKP2 (Fig. 6e, f and Supplementary Table 2) and confirming a
reduced proficiency to induce ACM in these mice. Western blot ana-
lysis of two hearts per group from EGFP tagged proteins (Fig. 6g)
revealed similar levels of EGFP signal between AAV-PKP2-R735*-EGFP
and AAV-EGFP-PKP2-R735* animals, confirming that functional differ-
ences observed in the MRI analysis are not mediated by differential
protein levels between EGFP-PKP2-R735* and PKP2-R735*-EGFP. Finally,
to examine whether systolic contractile differences observed between
these hearts are originated in the cardiac cells itself we tracked sar-
comere lengths in isolated cardiomyocytes across the contractile cycle
from AAV-EGFP, AAV- EGFP-PKP2, AAV-EGFP-PKP2-p.R735* and AAV-
PKP2-p.R735*-EGFP. At relaxed state therewere no differences between
sarcomere lengths in all tested groups. The analysis of isolated control
cardiomyocytes showed that sarcomeric shortening was comparable
betweenuntransduced control, andAAV-PKP2mice, demonstrating no
effect of AAV infection on28 (Fig. 6h). Isolated cardiomyocytes from
Fig. 2 | Mutant PKP2 modifies structure of actin filament network.
aRepresentative immunoblots showing thedistributionofβ-actin in pellet (F-actin)
and supernatant (G-actin) fractions of cells encoding PKP2, or p.R735*mutant.
Graph shows F-actin:G-actin ratios calculated from densitometry analysis. Data are
presented as mean± sd; n = 6. Significance was established as p <0.05 (unpaired
Student t test). b Confocal images showing phalloidin-staining of the F-actin net-
work (blue) in mouse neonatal cardiomyocytes (MNC) expressing PKP2 or p.R735*
proteins tagged with EGFP (yellow). Nuclei were stained with DAPI (gray). Magni-
fication shows the different structure of F-actin depending on the expression of
PKP2 or p.R735*. Scale bars, 20 µm and 5 µm. c Percentage of MNC with visually
apparent and organized sarcomeres; n = 54 EGFP-PKP2 cells, n = 79 EGFP-pR735*
cells examinedoverfive independent experiments.d Sarcomeric periodicity in cells
with organized sarcomere network in each group of cells. Red dot line defines
normal sarcomeric period (1,78 µm).Boxes depict the 25th−75th percentilewith a line
showing the median. Whiskers display minimum to maximum values. n = 19, ROIs
for EGFP-PKP2, n = 10ROIs for EGFP-PKP2-p.R735*. e TTorg analysis of the captured
images assessed the organization of the sarcomeres, and this is represented by the
sarcomeric powermetric. Calculated transverse organization level of sarcomeres is
based on the calculation of the peak amplitude in the Fourier spectrum of the
image at the sarcomeric frequency. Boxes depict the 25th−75th percentile with a line
showing the median. Whiskers display minimum to maximum values (n = 19, ROIs
for EGFP-PKP2, and n = 10 ROIs for EGFP-PKP2-p.R735*). f Analysis of cellular actin
orientation (coherency), using the ImageJ plugin OrientationJ, was performed with
an automated routine to quantitatively assess the differences among groups ana-
lyzing 20 × 20 µmROIs fromconfocal images. Orientation coherency increaseswith
the orientation anisotropy in the image (from zero for random orientations to one
for unique orientations). Boxes depict the 25th−75th percentile with a line showing
the median. Whiskers display minimum to maximum values; n = 78 ROIs for PKP2,
n = 98 ROIs for PKP2-p.R735*. Statistical significance was determined by unpaired t-
test with Welch’s correction (two-sided) in (c–f). Significance was established as
p <0.05. Data are presented asmean ± sd. MWmolecular weight, kDa kiloDalton, %
percentage,μmmicrometers, ArbUnits arbitrary units. Source data are provided as
a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 5
AAV-EGFP-PKP2-p.R735* mice had decreased maximal shortening
compared with controls what demonstrates an effect of PKP2
C-terminal mutant on contractility. When compared with AAV-EGFP-
PKP2, cardiomyocytes from AAV-EGFP-PKP2-p.R735* mice had 45%
reduced shortening, whilst shortening was partially recovered in car-
diomyocytes expressing AAV-PKP2-R735*-EGFP (Fig. 6h and Supple-
mentary Fig. 8). All together, these data validate that PKP2
delocalization is critical for protein dysfunction and the ability of
mutant PKP2 to alter sarcomeric contraction and compromise RV
cardiac systolic function.
Mice encoding PKP2-p.R735* or MYH10-p.R577* mutant present
sarcomere alterations and RV systolic dysfunction
It has been recently shown that coordinated functions of MYH9 and
MYH10, in time and space, are required for proper sarcomere assem-
bly in an in vitro model of human cardiomyocyte29. We questioned
tdTomato-p.R735*
tdTomato-p.R735*
tdTomato-PKP2 F-actin
F-actin
F-actin
F-actin
5 m10 m
tdTomato-PKP2
a
c
EGFP-MYH10
20 m
F-actin
F-actinEGFP-MYH10-p.R577*
5 m
C
oh
er
en
cy
 (
Ar
b.
 U
ni
ts
)
0.0
0.2
0.4
0.8
0.6
d
EGF
P-MY
H10
EGF
P-MY
H10-
p.R5
77*
EGFP-MYH10
EGFP-MYH10
b
p=0.0004
+ - -
+
+
-- +
- + - +
C
oh
er
en
cy
 (
Ar
b.
 U
ni
ts
)
0.0
0.2
0.4
0.1
0.5
0.3
p=0.9677
tdTomato-PKP2
tdTomato-p.R735*
EGFP-MYH10
p<0.0001
p=0.0018p=0.0008
p<0.0001
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 6
whether MYH10 disfunction could affect general sarcomere organi-
zation and be responsible for decreased performance in mice
observed by MRI (Fig. 4e). Actin filaments from AAV-EGFP-MYH10-
p.R577*, AAV-EGFP-PKP2, AAV-EGFP-PKP2-R735*, and non-transduced
sham control mice were visualized in isolated cardiomyocytes by
staining with phalloidin (Fig. 7a) to measure the angle of deviation
from the main direction (striation orientation) that the sarcomeres
follow. Image analysis revealed a significant increase in the AAV-EGFP-
MYH10-p.R577* and AAV-EGFP-PKP2-R735* angle variation compared
with the other two groups (5.57 ± 0.18° and 5.56 ±0.19° in PKP2-p.R735*
and MYH10-p.R577* cardiomyocytes respectively vs. 3.69 ± 0.13° and
3.84 ± 0.13° in non-transduced, AAV-PKP2; n = 28) (Fig. 7b). As less
variation in sarcomere orientation corresponds to optimal force
generation30, this data predicts a comparative and similar force loss
only in sarcomeres of AAV-p.R735* and AAV-EGFP-MYH10-p.R577*
hearts, in agreement with cardiac functional MRI results (Fig. 7c, d and
Supplementary Fig. 9). Consistently, the contraction analysis of AAV-
EGFP-PKP2-p.R735* and AAV-EGFP-MYH10-p.R577* isolated cardiomyo-
cytes also showed that sarcomeric shortening was compromised in
both mice lines (Fig. 7e), establishing a clear correlation between sar-
comeric misalignment and reduced systolic contraction in vivo. Alto-
gether, our data shows that PKP2 cytoplasmic delocalization and
MYH10 protein–protein interaction alters actomyosin component
relative distribution and activity, affecting sarcomere organization,
contraction capacity and RV cardiac systolic function inducing an
ACM-like phenotype.
Activation of MYH10 corrects PKP2-p.R735* defects
Asgenetic overdoseofMYH10normalized the actomyosin structureof
PKP2-p.R735* cells and dominant negative EGFP-MYH10-p.R577*
mutant mimics contractile defects observed in ACM, we hypothesized
that 4-hydroxyacetophenone (4-HAP), a selective non-muscle myosins
activator, would reduce deleterious effects of PKP2mutant. This small
molecule activatesMYH10 andMYH14 by promoting their assemblies31
and has been proposed for treating different types of metastatic
tumors32,33. Using in vitro and in vivo models of ACM described above
we studied the response to 4-HAP treatment. Cardiomyocytes
encoding the PKP2-p.R735* mutant had a disorganized actin cytoske-
leton (Fig. 2b). Images from cytoplasmic area of MNC expressing the
PKP2-p.R735* mutant after 4-HAP administration (50μM) revealed an
improved F-actin filaments network (Fig. 8a).We detected a significant
reorganization of PKP2-p.R735* mutant sarcomeres in treatedMNC, as
indicated by the increase of cells with evident sarcomeres from 30.02
to 70.83% after 4-HAP treatment (Fig. 8b), and the improvement in
sarcomeric power compared to PKP2 transduced control (61.9 ± 6.4
Arb. Units in untreated PKP2-p.R735* vs. 73.1 ± 12.6, 67.9 ± 12.3 and
70.2 ± 11.2 in non-treated PKP2, and 4-HAP treated PKP2, and mutant
PKP2-p.R735* MNC respectively; n = 16–22) (Fig. 8c). We also observed
a significant improvement in the PKP2-p.R735* mutant actin general
spatial disposition, reestablishing orientated actin structures of fibers
after 4-HAP treatment (Fig. 8d). Furthermore, in vivo, administrationof
4-HAP (1mg/kg) for one week improved cardiac RV systolic function
(Fig. 9a) (22.2 ± 6.3 μL in untreated mice vs. 15.9 ± 3.2 μL in 4-HAP
treated AAV-R735* mice; n = 8–12) and general RV performance
(49.1 ± 7.5% vs. 58.5 ± 4.0% in sham and treated AAV-PKP2-p.R735*-EGFP
mice respectively; n = 8–12) (Fig. 9b) in animals expressing PKP2-
p.R735* mutant. Unlike AAV-EGFP-PKP2-p.R735* mutant, 4-HAP admin-
istration didnot significantly reducedRV-ESVor significantly increased
RV-EF in AAV-EGFP-PKP2 (Fig. 9 and Supplementary Fig. 10). Con-
sistently, in isolated cardiomyocytes from AAV-EGFP-PKP2, and AAV-
EGFP-PKP2-p.R735* mice after treatment the actin filaments stained
withphalloidin showed that the angle of sarcomeric deviation from the
main direction was restored to normal (Fig. 9e). Sarcomeric realign-
ment in treated mice also correlated with improved contraction
measured in isolated cardiomyocytes. In these cardiomyocytes sar-
comeric shortening was indistinguishable between AAV-EGFP-PKP2,
AAV-EGFP-PKP2-p.R735* after 4-HAP treatment (Fig. 9f). In summary,
therapeutic activation of MYH10 corrects cellular contractile pheno-
type and may potentially limit RV adverse clinical complications in
patients with ACM.
Discussion
Herein, we reveal that PKP2 C-terminal deletion mutants act as a key
regulator of cardiac contractile disfunction in arrhythmogenic cardi-
omyopathy (ACM) and that their direct protein-protein interactor
MYH10 has a critical role in the pathological process. Four major
findings fromour in vitro and in vivo experimentalmodels support this
conclusion. First, expression of C-terminal deletion of PKP2 modifies
MYH10 activity, altering F-actin fiber organization and decreasing
cellular resistance to deformation (stiffness). Second, cardiac-specific
PKP2C-terminal GOFmutants, contributed to pathologicalmyocardial
ACM by altering the actomyosin network, what translates into sarco-
meric abnormal organization and right ventricular (RV) cardiac func-
tion. Third, mutant MYH10-p.R577* expression compromises
sarcomere fine tune organization and cardiac performance alike PKP2
C-terminal deletion mutant. Finally, therapeutic administration of
MYH10 activator 4-HAP to animals with reduced RV ejection fraction
(EF) improves systolic function and restoredmechanical performance.
Apart from extensive literature onMYH10 role in vitro and during
development17,34–37, little is known about its function in adult cardiac
cells. It has been shown that MYH10 dysregulation in iPS-derived car-
diomyocytes is associated to increased sarcomere misorientation and
sarcomeric dysfunction. In our in vivo setting, it is possible that
reduced force generated in C-terminal deletion mutants is transduced
into sarcomeres andderived from integratedfibermisorientationhas a
consequence on tissue performance. From a biophysical perspective,
vectorial nature of force predicts that an increase in the actin angle
deviation from the perpendicular sarcomere axis will produce a
reduction on the final magnitude of the generated force. Thus, mis-
alignment in a single PKP2 or MYH10 mutant cardiomyocytes can
partially justify a decline in the force the heart muscle can exert to
neighbor sarcomere. Consistently, the image analysis of angle varia-
tion from individual cardiomyocytes correlates with a significant dif-
ference between mutants PKP2-p.R735* and MYH10-p.R577*, with
Fig. 3 | Mutant PKP2 functions through MYH10 to disturb actin organization.
a Representative images of actin filaments (yellow) in MNC expressing tdTomato
fusion PKP2 or PKP2-p.R735* (blue) together with functional EGFP-MYH10
(Magenta). Right panels show a magnification of actin filaments in cells with or
without EGFP-MYH10. This experiment was repeated independently 3 times. Scale
bars, 10 µm and 5 µm. b Graphs show quantification from confocal images (a) of
actin filament coherency (organization). Boxes depict the 25th−75th percentilewith a
line showing the median. Whiskers display minimum to maximum values. (n = 36
ROIs for tdTomato-PKP2 transfection; n = 40 ROIs for tdTomato-PKP2 with EGFP-
MYH10 cotransfection; n = 17 ROIs for EGFP-PKP2-p.R735* transfection; n = 51 ROIs
for tdTomato-PKP2-p.R735* with EGFP-MYH10 cotransfection) Statistical sig-
nificance was determined by one-way ANOVA with the Tukey multiple comparison
post-test. Significancewas established as p <0.05. Data are presented asmean± sd.
c Representative confocal images showing merge signals of phalloidin-staining of
the F-actin network (blue), wild-type EGFP-MYH10 ormutant EGFP-MYH10-p.R577*
(yellow) and nuclei (gray) in MNC. A magnification of selected white box area
showing F-actin signal. d Analysis of cellular actin orientation (coherency), using
OrientationJ, to quantitatively assess the differences amongwild-type EGFP-MYH10
ormutant EGFP-MYH10-p.R577* groups. Boxes depict the 25th−75th percentilewith a
line showing the median. Whiskers display minimum to maximum values. n = 24
ROIs for MYH10; n = 37 ROIs for MYH10-p.R577*. Statistical significance was
determinedby unpaired t test withWelch’s correction (two-sided). Significancewas
stablished as p <0.05. Data are presented as mean± sd. Arb. Units arbitrary units,
μmmicrometers. Source data are provided as a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 7
transduced control animals, not observed either in wild-type mice or
animals treated with 4-HAP for one week. We speculate that such
variation in the sarcomeric organization of an ultrastructure-related
force is at least in part behind the deficit of heart muscle contraction
observed in functional MRI analysis (Figs. 4, 6, 7 and 9).
As mentioned above, dominant inherited cardiomyopathy could
be caused by loss-of-function (haploinsufficiency), a dominant
negative effect or GOF. We identified a possible mechanism through
which the PKP2-p.R735* variant could function as a GOF mutation and
not as a dominant negative or a simple loss-of-function allele induced
by a nonsense-mediatedmRNA decaymechanism. As proteins need to
be in the correct cellular compartment to perform their function, it is
not surprising disease mutations can lead to depletion of the protein
from its native location, or its accumulation in the wrong
100 kDa
75 kDa
 Membrane fraction
-
-+
+ +
+
Long exposure
-
+ +
+
150 kDa
-Pkp2
-N cadh
PKP2
p.R735*
100 kDa
75 kDa
37 kDa
PKP2
p.R735* -
-+
+ +
+
-Pkp2
-Gapdh
Cytosolic fraction
d
c
a
EGFP-p.R735*
tdTomato-PKP2
10 m
e
Diastole Systole
PK
P2
p.
R
73
5*
LVRV
Short Axis
btdTomato-PKP2
EGFP-p.R735*
EGFP-p.R735*EGFP-PKP2
10 μm
+ -
+
+
- +
+
+
0.0
0.5
1.0
1.5
2.0
2.5
In
te
ns
ity
 M
em
b/
C
yt
o 
(A
rb
. U
ni
ts
)
tdTomato-PKP2
EGFP-p.R735*
R
V-
ES
V 
(μ
l)
Sham PKP2 p.R735*
0
10
20
30
40
50
p=0.9663 p=0.0231
p=0.0141
R
V-
EF
 (%
)
Sham PKP2 p.R735*
40
50
60
70
80 p=0.7805 p=0.0007
p=0.0002
p<0.0001
p<0.0001
p<0.0001
p=0.0010
p=0.0431
p=0.1596
Membrane relative location
p<0.0001
0.0 0.5 1.0 1.5 2.0
EGFP
EGFP-p.R735*
EGFP-PKP2
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 8
compartment. We show that mutant PKP2-p.R735* is mislocalized,
although does not sequester or alter wild-type PKP2 localization and
dynamics at the cell-to-cell contact area (Fig. 4). This data suggests that
PKP2-p.R735* mutant effect is not dependent on wild-type PKP2 dys-
function or cellular mislocalization and must be a consequence of
acquired new activities. We determined that the pathogenic PKP2-
p.R735* variant functions by interrupting or rewiring highly connected
interaction networks to disturb F-actin homeostasis. We therefore
predict that GOF is themode of action of other C-terminally truncated
PKP2 variants stable to retain MYH10 interaction and to disrupt acto-
myosin organization. The conformational changes induced by
C-terminal deletion mutations in PKP2 alter mutant proteins localiza-
tion and function (Fig. 5). PKP2 C-terminal deletions exposes the nor-
mally buried internal interface to the solvent. This structural alteration
in PKP2may rewire highly connected cellular interaction networks in a
subtle but disease-specificmanner. Protein trafficking and localization
is regulated by interplay between diverse cis-acting targeting
sequences and trans-acting trafficking factors. Thus, PKP2 C-terminal
deletionsmay affect protein trafficking and localization either directly,
through the loss of a domain essential for directing PKP2 to the des-
mosome, or indirectly, by generating unknow or altered interactions
that keep the protein away from the cell-to-cell contact area. Our
results (Fig. 6) make a strong case for this indirect effect, since incor-
poration of the EGFP tag in the closed PKP2-p.R735*-EGFP mutant
partially restored proper protein localization compared with the open
PKP2-p.R735*mutant. This experiment also suggests that the PKP2
C-terminal domain lacks a specific motif required for desmosomal
localization and that the PKP2 mutants with open conformation gen-
erate or strengthen interactions that retain the protein in the soluble
fraction.
The ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/),
part of the NCBI Entrez system, attempts to establish relationships
between gene variants and phenotype. Among them, 34 are nonsense
mutations and 23 spandifferent portions of the 8 armadillo domains in
PKP2. Remarkably, all nonsense PKP2 variants are classified as patho-
genic or likely pathogenic, underlining the functional importance of
the PKP2 C-terminal domain. We predict that PKP2 mutants that lack
the ability to stably localize in the desmosome will share similarities in
their phenotypes when they maintain their interaction with MYH10.
The RV and the LV free myocytes are developed from different
progenitor cells, their walls possess distinct architecture, and response
mechanisms against stress. RV has a thinner free wall and support an
intense stress load compared with the LV. The most common subtype
of ACM is arrhythmogenic right ventricular cardiomyopathy (ARVC),
which is overrepresented in patients with mutations resulting in pre-
mature termination of the PKP2 protein38,39. Therefore, it is not sur-
prising that pathogenic mutations in PKP2 have been usually asso-
ciated with the classical form of the disease that predominantly shows
structural abnormalities that lead to a progressive global RV
dysfunction1. We observed a clear structural and functional phenotype
in the RVmeanwhile significant functional variations are not evident in
the LV of mutant animals. In patients, these phenotypic differences
between ventricles have been sometimes explained due to environ-
mental stressors like exercise16,40 that define the final disease outcome.
For example, under extreme exercise conditions theRVgoes through a
greater load increase compared with the left ventricle, where near-
linear increase in pulmonary artery pressures predominantly con-
tributes to a disproportionate increase in RV wall stress41. It is rea-
sonable that ACM develops preferentially in the RV in PKP2 mutant
carriers with compromised structural integrity of the myocardial
cytoskeleton and half the amount of functional PKP2 protein at des-
mosomal structures.
Sarcomere assembly, maturation and maintenance is a complex
process that depends on multiple components42,43 Adult cardiomyo-
cytes become isotropic, organizing their contractile cytoskeleton and
remodeling nuclei, and junctions44. Here we present compelling evi-
dence that mutations in desmosomal protein PKP2 related to ACM
disturb sarcomeric homeostasis by modifying MYH10 activity. Our
results highlight the role of MYH10 as a key component for sarcomere
maintenance in adult cardiomyocyte, including evidence of cardio-
myopathy development with their alteration, indicative of sarcomere
organization interference.
With the lack of treatments for patients with ACM, our data
identifies the actomyosin network in the context of cardiacmechanical
performance regulator in ACM disease and highlights the therapeutic
potential of MYH10 pharmacological activation to improve sarcomere
organization and to recover cardiac contractability. It is essential to
understand the pathological mechanism of the target mutation on
PKP2 to develop the potentially most beneficial treatment.
Methods
Animal
All animal procedures followed the guidelines from Directive
2010/63/EU of the European Parliament on the protection of ani-
mals used for scientific purposes. Animal experiments were car-
ried out in accordance with the CNIC Institutional Ethics
Committee recommendations and were approved by the Animal
Experimentation Committee (Scientific Procedures) of Comuni-
dad de Madrid (project number PROEX 019/17).
All mice used in this study were generated on a C57BL/6J back-
ground and were provided by Charles River Laboratories. Three to 5-
month-old male mice and neonatal mice of both sexes were analyzed.
The animals weremaintained in a specific pathogen-free animal facility
under controlled temperature (22° ± 0.8 °C), 55 ± 10% relative humidity
and a 12 h light/dark cycle. Mice had access to food (Teklad global rat/
mouse chow, Harlan Interfauna) andwater ad libitum. Adultmicewere
Fig. 4 | PKP2 C-terminal deletion modifies its subcellular localization and
definesactomyosin function. aRepresentative images of showing thedistribution
in mouse neonatal cardiomyocytes (MNC) of tdTomato-PKP2, EGFP-PKP2-p.R735*
mutant or both. Scale bars, 10 µm. b Dot-plot showing the distribution of PKP2
versions as ratio intensity of membrane-cytoplasm. Lanes one and two show PKP2
membrane-cytoplasm ratio intensity, last two lanes show PKP2-p.R735* ratio;
n = 41 cells for tdTomato-PKP2 tranfection, n = 27 cells for TtdTomato-PKP2 with
EGFP-PKP2-p.R735* cotransfection; n = 33 cells for EGFP-PKP2-p.R735* transfection;
n = 31 cells for EGFP-PKP2-p.R735* with tdTomato-PKP2 cotransfection from three
independent experiments. Significancewas stablished as p <0.05 (onewayANOVA
with Tukey multiple comparison post-test). Data are presented as mean± sd.
c Immunoblot analysis of membrane and cytosolic cell fractions. Symbols + and –
indicate transfection or not with vectors expressing PKP2, p.R735*, or both variants
together. Blots show PKP2 (wild-type and mutant versions), N-cadherin (load
control of membrane fraction) and GAPDH (load control of cytosolic fraction).
Black dotted lines delimit short and long exposures of the same blot to detect
p.R735* and PKP2 proteins. The results shown are representative of three inde-
pendent experiments. d, Images of heart sections immunostained with EGFP anti-
body. Animals were transducedwith AAV-EGFP, AAV-EGFP-PKP2 or AAV-EGFP-PKP2-
p.R735*. Red arrowheads indicate PKP2 or PKP2-p.R735* cellular location. Bottom
graph shows relative percentage of membrane localization; n = 9 mice (EGFP and
EGFP-PKP2), and n = 7 mice (EGFP-PKP2-p.R735*). Scale bars, 10 µm.
e Representative short-axis cardiac MRI images taken at the end of diastole and
systole in mice transduced with wild-type AAV for PKP2 or the p.R735*mutant.
Quantification of right ventricular (RV) end systolic volume (RV-ESV) and ejection
fraction (RV-EF), by MRI in anesthetized non-transduced sham control mice, AAV-
PKP2 and AAV-PKP2-p.R735* transduced mice. Data are presented as mean± sd;
n = 6mice. Statistical significance was determined in (d, e) by one-way ANOVAwith
Tukeymultiple comparison post-test. Significancewas established asp <0.05. Data
are presented as mean ± sem. Arb. Units arbitrary units, KDa kiloDalton, μm
micrometers, μL microliters, % percentage, RV right ventricle, ESV end systolic
volume, EF ejection fraction. Source data are provided as a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 9
euthanized by carbon dioxide inhalation, and neonates P1 by decap-
itation with sharp well‐maintained scissors.
Adeno-associated virus (AAV) animal models
To develop all animal models, AAV vectors were produced with the
double transfectionmethod by using HEK293T (ATCC, CRL-3216) cells
as described previously16,45. AAV plasmidswere cloned andpropagated
in the Stbl3 E. coli strain (Life Technologies). Shuttle plasmids pAAV-
PKP2, pAAV-PKP2-p.R735*46, pAAV-EGFP-PKP2, pAAV-EGFP, pAAV-EGFP-
PKP2-p.R735*, pAAV-PKP2-p.R735*-EGFP, and pAAV-EGFPMYH10-p.R577*
were packaged into AAV-9 capsids with the use of the helper plasmid
pDG9 (providing the three adenoviral helper genes, and rep and cap
a
PKP2-p.R735*PKP2 PKP2-p.W676* PKP2-p.Y807*
N-Ter ARM C-Ter N-Ter ARM N-Ter ARM N-Ter ARM
b c
EGFP-p.R735*
EGFP-p.W676*
EGFP-p.Y807*
F-actin
F-actin
F-actin
5 m
EGFP-PKP2
20 m
F-actin
EGF
P-
PKP
2 EG
FP-
p.R7
35* EG
FP-
p.W6
76* EG
FP-
p.Y8
07*
0.0
0.2
0.4
0.6
0.8
1.0
C
oh
er
en
cy
 (A
rb
. U
ni
ts
.)
p<0.0001
p<0.0001
p<0.0001
p=0.0307p=0.0753
p=0.7992
Fig. 5 | PKP2 C-terminal deletion mutants show a disorganized N-terminal and
altered subcellular localization. a 3D view of protein PKP2, PKP2-p.R735*, PKP2-
p.W676*, and PKP2-p.Y807*. Below the schematic diagram of the domains within
PKP2 proteins. Different domains are shown: N-domain (N-Ter, blue), armadillo
domain (ARM, yellow), C-domain (C-Ter, Purple). b Representative images of
mouse neonatal cardiomyocytes (MNC) transfected with EGFP-PKP2, EGFP-PKP2-
p.R735*, EGFP-PKP2-p.W676* or EGFP-PKP2-p.Y807* (yellow) showing its location.
Cells are also stained with phalloidin (blue) and DAPI (gray). Scale bars, 20 µm and
5 µm. c Analysis of cellular actin orientation, using the ImageJ plugin OrientationJ.
Boxes depict the 25th−75th percentile with a line showing the median. Whiskers
display minimum to maximum values. n = 78 ROIs for EGFP-PKP2 n = 98 ROIs for
EGFP-PKP2-p.R735*;n = 49ROIs for EGFP-PKP2-p.W676* n = 38 ROIs for EGFP-PKP2-
p.Y807*. Statistical significance was determined by one-way ANOVA with Tukey’s
multiple comparison post-test. Significance was stablished as p <0.05. Data are
presented asmean± sd.Arb.Units arbitraryunits,μmmicrometers. Sourcedata are
provided as a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 10
viral genes), obtained from PlasmidFactory. The AAV shuttle and
helper plasmids were transfected into HEK293T cells by calcium-
phosphate co-precipitation. A total of 690 µg of equimolar-mixed
plasmid DNA was used per Hyperflask (Corning) seeded with 1.2 × 108
cells the day before. Seventy-two hours after transfection, the cells
were collected by centrifugation, and the cell pellet was resuspended
in TMS (50mM Tris-HCl, 150mM NaCl, 2mM MgCl2) on ice before
digestion with Benzonase nuclease (25 KU/mL; Millipore) at 37 °C for
30min. Clarified supernatant containing the viral particles was
obtained by iodixanol gradient centrifugation47. Virus-containing gra-
dient fractions were concentrated through Amicon UltraCel columns
(Millipore) and stored at −80 °C. For AAV injection mice were anes-
thetized with a 100 µL intraperitoneal injection containing ketamine
(60mg/kg), xylazine (20mg/kg), and atropine (9mg/kg). Once anes-
thetized, animalswere placed on a heated pad at 37 ± 0.5 °C to prevent
hypothermia. A small incision (4mm) was made to expose the right
g
a b
d
e
EGF
P-
PKP
2 EG
FP-
p.R7
35* p.R
735*
-
EGF
P
0.0
0.2
0.4
0.6
0.8
1.0
C
oh
er
en
cy
(A
rb
. U
ni
ts
)
EGFP-PKP2
EGFP-p.R735*
p.R735*-EGFP
EGFP
GAPDH
100 
 75 
25 
 37
-
-
-
-
+
-
-
-
+
-
-
-
-
+
-
-
-
+
-
-
-
-
+-
-
-
+
-
-
-
-
+
-
-
-
+
1 2 3 4 5 6 7 8 9AAV
EGFP-
fusion 
protein
[
Free EGFP
Mice
MW
(kDa)
EGFP-PKP2
EGFP-p.R735*
p.R735*-EGFP F-actin
F-actin
F-actin
10 m 5 m
PKP2-p.R735*-EGFP
N-Ter ARM C-EGFP
c
-
-
+
+-
-
EGFP-PKP2
EGFP-p.R735*
p.R735*-EGFP
+ - -
0.0
0.5
1.0
1.5
2.0
2.5
In
te
ns
ity
 M
em
b/
C
yt
o 
(A
rb
. U
ni
ts
)
h
f
0
10
20
30
40
50
R
V-
ES
V 
(μ
L)
p=0.6000
EGFP-PKP2
EGFP-p.R735*
p.R735*-EGFP
- + --
- -- +
- +- -
p=0.9803 p=0.2840
p=0.0132
0
20
40
60
80
100
R
V-
EF
 (%
)
p=0.8605
- + --
- -- +
- +- -
p=0.0747
p=0.0002
p=0.2376
p<0.0001
p<0.0001 p<0.0001
p<0.0001 p<0.0001
p=0.1978
Sa
rc
om
er
e 
co
nt
ra
ct
io
n 
le
ng
th
 (μ
m
)
EGFP-PKP2 
EGFP-p.R735* 
p.R735*-EGFP 
-
+
-
-
+
-
-
-
-
-- +
p=0.9834 p<0.0001 p=0.0210
0.00
0.05
0.10
0.15 p=0.0210
p<0.0001
p=0.0874
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 11
femoral vein. To increase vessel diameter and facilitate infusion, blood
flow was interrupted with a cotton bud for a couple of seconds. Once
the vein was dilated, an insulin syringe vessel was introduced into the
vein, and 1 × 1011 virus particles were injected in a volume of 50 µL,
taking care to prevent introduction of air bubbles. Animals were then
analgesized with buprenorphine (subcutaneous 0.1mg/kg) and main-
tainedon the heated paduntil recovery.Micewere individually housed
in wire-bottomed cages and paracetamol was administered orally
for 1 week.
Cardiomyocyte isolation
Mice were anesthetized by intraperitoneal injection of ketamine and
xylazine, and hearts were rapidly excised, cannulated with a 21-gauge
gavage feeding needle, connected to a Langendorff perfusion appa-
ratus, and perfused at 37 °C for 5min at 0.8mL/min with a perfusion
buffer containing (in mM) 113 NaCl, 4.7 KCl, 0.6 KH2PO4, 12 NaHCO3,
0.6 Na2HPO4, 0.032 phenol red, 10 KHCO3, 10 HEPES, 30 taurine, 1.2
MgSO4·7H2O, 5.5 glucose, and 10 2,3-butanedione monoxime (BDM),
pH 7.46. After 5min, the hearts were perfused for 20min at
0.8mL/min with digestion buffer, consisting of the perfusion buffer
plus 12.5μM CaCl2, 0.14mg/mL trypsin (Invitrogen, Carlsbad, CA),
0.005 U/ml DNAse (Sigma) and 0.2mg/mL Liberase Blendzyme 4
(Roche Diagnostics). Hearts were then cut from the cannula below the
atrium and placed in a dish containing 2.5mL digestion buffer. The
tissue was gently dissected into small pieces with fine forceps and
dissociated by passing through a sterile plastic transfer pipette. The
cell suspension was filtered through a 200μM nylon filter in a 15mL
falcon tube, and digestion was inactivated by addition of 10mL of first
cardiomyocyte stopping digestion buffer (perfusion buffer containing
10% fetal bovine serum and 12.5μM CaCl2). Cardiomyocytes were
allowed to settle to the bottom of the tube for 30min in darkness. The
supernatant was discarded, and the pellet was resuspended in 8mL of
a second cardiomyocyte stopping buffer (perfusion buffer containing
5% fetal bovine serum and 12.5μMCaCl2). The cells were again allowed
to settle to the bottom of the tube for 30min in darkness, and the
supernatantwasdiscarded. The cellswere thenpassed through a series
of 20min incubations in 8mL of modified second cardiomyocyte
stopping buffer with incremental increases in CaCl2 to a final con-
centration of 1mM.
Cardiomyocytes immunostaining and analysis
For phalloidin staining, cardiomyocytes were fixed for 20min in 4%
paraformaldehyde (PFA). Cells were blocked and permeabilized for 1 h
in TSABlockingReagent (FP1012, PerkinElmer) plus0.1% triton (T9284,
Sigma) at room temperature. Then Phalloidin-iFluor 594 Reagent
(ab176757, Abcam) was added at 1:1000 dilution and incubated for
90min. Finally, after wash with PBS, samples were mounted in Mowiol
mounting medium (Mowiol 4-88, Glycerol, 200mM Tris-HCl pH 8.5
and 2.5% 1,4-diazabicyclo-[2,2,2]-octane). Fluorescence images were
obtained with a Leica SP8 confocal microscope with HC PL APO
100x/1.4 oil objective. Actin filament dispersion in adult cardiomyo-
cytes were analyzed using Directionality plugin of Fiji.
Cardiomyocytes contraction analysis
Cardiomyocyte contractionwas recorded using LeicaDMi8withHCPL
APO 40× objective and DFC3000G-0047012016 camera. Myocytes
were superfused with Tyrode’s solution containing the following (in
mM): 136 NaCl, 5.4 KCl, 0.9 CaCl2, 1.2 KH2PO4, 5 HEPES, 10 glucose, at
pH 7.4, adjusted with NaOH. Myocytes were electrically field stimu-
lated at 1 Hz with square-wave pulses (9,9mA, 3ms) delivered through
a pair of platinum electrodes using an Axopatch 200B amplifier.
Screen was recorded with OBS program at 80fps. Then contraction
was analyzed with Sarcoptim plugin from Fiji software. Ten beats from
each cardiomyocyte were analyzed. The difference between the aver-
age of maximum and minimum is represented. At least ten cardio-
myocytes from three different animals were analyze.
Cardiac magnetic resonance imaging (MRI)
Cardiac magnetic resonance imaging (MRI) images were obtained
from isofluorane- anesthetized animals (induction: 2 vol.% inhaled
isoflurane in 1000 cm3/min O2, during measurement: 1.8–1.4 vol.%
inhaled isoflurane in 500 cm3/min O2) while monitoring for core body
temperature, cardiac rhythm, and respiration rate. In vivo cardiac
images were acquired using an Agilent VNMRS DD1 7T MRI system
(Santa Clara, California, USA) and a k-space segmented ECG-triggered
cine gradient-echo sequence. After shimming optimization, cardiac
four-chamber and left two-chamber views were acquired and used to
plan the short axis sequence. Mice were imaged with the following
parameter settings: number of slices, 14; slice thickness, 0.8mm; slice
gap, 0,2mm;matrix size, 256 × 256; field of view, 30 × 30mm2; gating,
ECG and respiratory triggered; cardiac phases, 20; averages, 4; Echo-
Time ~1.25ms, minimum TR, 6ms; flip angle, 15°; trigger delay, 0ms;
trigger window, 5ms; dummy scans, 2. Images from short axis cine
gradient-echo MRI sequences were quantitatively analyzed by manual
detection of endocardial borders in end-diastole and end-systole, with
exclusion of papillary muscles.
Cell lines
The HEK293T (ATCC, CRL-3216) cell line was maintained in DMEM
(GIBCO) supplementedwith 10%FBS, 1%penicillin/streptomycin, and2
mM L-glutamine. The atrial cardiomyocyte cell line HL-1 (Sigma,
Aldrich) was maintained in Claycomb medium (Sigma Aldrich) sup-
plemented with 10% FBS, 1% penicillin/streptomycin, and 2 mM
L-glutamine, as previously described48. HL-1 cells were seeded on
plates coated with 0.02% gelatin/fibronectin (Sigma Aldrich). Cell lines
were maintained at 37 °C with 5% of CO2.
Mouse neonatal cardiomyocyte (MNC) isolation
Cardiac cells were isolated from day-old neonatal mouse hearts, pre-
viously segregated by sex49, in sterile conditions as described in the
Fig. 6 | PKP2 N-terminal misfolding recovery ameliorates protein delocaliza-
tion and cardiac dysfunction. a 3D view of protein PKP2-p.R735*-EGFP with
schematic diagram of the domains: N-terminal domain (N-Ter, blue), armadillo
domain (ARM, yellow), C-terminal EGFP (C-EGFP, green). b Representative images
of EGFP-PKP2, EGFP-PKP2-p.R735* and PKP2-p.R735*-EGFP protein localization and
F-actin structure in MNC. A magnification of selected white box area showing
F-actin signal. Scale bars, 10 µmand 5 µm. c Plot shows the quantification of plasma
membrane intensity levels normalized to cytoplasm intensity (n = 41 cells for EGFP-
PKP2 n = 32 cells for EGFP PKP2-p.R735* n = 48 cells for PKP2-p.R735*-EGFP) from
three independent experiments, d Analysis of cellular actin coherency from (b)
(n = 78 ROIs for EGFP-PKP2; n = 98 ROIs for EGFP-PKP2-p.R735*; n = 31 ROIs for
PKP2-p.R735*-EGFP). e, f Right ventricular (RV) end systolic volume (ESV) and
ejection fraction (EF) determined by MRI in hearts of non-transduced (sham con-
trol) and AAV-EGFP-PKP2, AAV-EGFP-PKP2-p.R735*, or AAV-PKP2-p.R735*-EGFP
transduced mice; (n = 10 mice for all the groups but PKP2-p.R735*-EGFP n = 8).
gMice were transduced as in (e), prior to Western blot analysis. Heart lysates were
immunoblotted to analyze free EGFP and fusion protein levels. GADPHwas used as
loading control. h Graph showing the difference between the maximum and
minimum sarcomere length in ten consecutive beats from isolate adult cardio-
myocytes of sham control non-transduced, AAV-EGFP, AAV-EGFP-PKP2, AAV-EGFP-
PKP2-p.R735* and AAV-PKP2-p.R735*-EGFPmice; (n = 8, n = 11, n = 10 and n = 11 car-
diomyocytes respectively, from three independent mice). Statistical significance
was determined by one-way ANOVA with Tukey’s multiple comparison post-test.
Significance was stablished as p <0.05. Data are presented as mean ± sd (c, d) or
mean ± sem (e, f,h).μmmicrometers, Arb. Units arbitrary units,μLmicroliters,MW
molecular weight, kDa kiloDalton, RV right ventricle, ESV end systolic volume, EF
ejection fraction. Source data are provided as a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 12
mouse and rat neonatal heart dissociation kit protocol (Miltenyi Biotec
130-098-373). Pipetting approximately 1.6 × 105 cells per well of glass
bottom Ibidi plates covered with a layer of 1% gelatin. Cells were
maintained at 37 °C with 5% of CO2.
Transfection and stable cell lines
Transient transfection with cDNAs encoding EGFP, EGFP-tagged PKP2,
PKP2-p.R735*, PKP2-p.W676*, PKP2-p.Y807*, MYH10 and MYH10-
p.R577* proteins were performed in 8-well cell culture dishes (Ibidi)
a b
e
EGFP-p.R735* EGFP-MYH10-p.R577*EGFP-PKP2
10 μm
2.5 μm
c d
0
10
20
30
40
R
V-
ES
V
(μ
L)
p=0.2979
EG
FP-
MY
H10
-
p.R
577
*
EG
FPSha
m
p=0.0065
p=0.0206
45
60
75
90
R
V-
EF
(%
)
EG
FP-
MY
H10
-
p.R
577
*
EG
FPSha
m
p=0.1825 p<0.0001
p<0.0001
0.05
0.10
0.15
0.20
Sa
rc
om
er
e 
co
nt
ra
ct
io
n 
le
ng
th
 (μ
m
)
0.00
EG
FP-
PKP
2
EG
FP-
p.R
735
*
Sha
m
EG
FP-
MY
H10
-
p.R
577
*
p=0.9864 p<0.0001 p=0.6423
p=0.0035
p=0.0001
D
is
pe
rs
io
n
(º
)
2
4
6
8
10
0
EG
FP-
PKP
2
EG
FP-
p.R
735
*
Sha
m
EG
FP-
MY
H10
-
p.R
577
*
p=0.9007 p<0.0001 p=0.9999
p<0.0001
Fig. 7 | Mice encoding PKP2-p.R735* or MYH10-p.R577* mutant present sarco-
mere alterations and right ventricle systolic dysfunction. a Representative
images showing cardiac actin network in phalloidin-stained isolated cardiomyo-
cytes from in vivo transduced AAV-EGFP-PKP2, AAV-EGFP-PKP2-p.R735* or AAV-
EGFP-MYH10-p.R577*mice. Red arrows indicate the normal actin fibers direction.
Lower panels show a magnification from yellow boxes. Scale bars, 10 µm and
2.5 µm b Graph quantifies the dispersion from main direction of the actin fila-
ments in phalloidin-stained adult cardiomyocytes isolated from mice non-
transduced (sham control) and transduced with AAV-EGFP-PKP2, AAV-EGFP-PKP2-
p.R735* or AAV- EGFP-MYH10-p.R577*. Data are presented asmean ± sd; n = 28 cells
from two independent experiments. Significance was stablished as p < 0.05. (one-
way ANOVA with the Tukey multiple comparison post-test). c, d Right ventricular
(RV) end systolic volume (ESV) and ejection fraction (EF) determined by MRI in
hearts of non-transduced (sham control) and AAV-EGFP or AAV-EGFP-MYH10-
p.R577* transduced mice. Data are presented as mean ± sd; (n = 10 mice for all the
groups but EGFP n = 5), repeated twice, e Graph showing the sarcomere con-
traction length as the difference between the maximum and minimum length in
ten consecutive beats from isolated cardiomyocytes of non-transduced control,
AAV-EGFP-PKP2, AAV-EGFP-PKP2-p.R735* and AAV-EGFP-MYH10-p.R577* mice
(n = 11 cells for all the groups except for sham which was n = 6). Statistical sig-
nificance was determined for b–e by one-way ANOVA with the Tukey’s multiple
comparison post-test; for (c, d) by Brown-Forsythe and Welch ANOVA tests,
Significance was established as p < 0.05. Data are presented as mean ± sem.
° degrees, μmmicrometers, μL microliters, % percentage, RV right ventricle, ESV
end systolic volume, EF ejection fraction. Source data are provided as a source
data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 13
using the jetPRIME® reagent according to the manufacturer’s protocol
(Polyplus transfection®). Stable HL-1 cell lines were generated using
the PiggyBac transposon system. In brief, cells were transfected with
plasmids expressing PKP2, PKP2-p.R735*, EGFP, EGFP-tagged PKP2, and
PKP2-p.R735*, together with the pPB-transposase50. Cells were then
selected for geneticin resistance (G418, ThermoFisher Scientific) and
expanded for further experiments.
Atomic force microscopy (AFM)
The AFM experiments were performed with a commercial instrument
(JPK NanoWizard 3, JPK Instruments AG, Germany) mounted on an
Axio Vert A1 inverted microscope (Carl Zeiss, Germany). For these
experiments, HL-1 cells stably expressing PKP2 or PKP2-p.R735* were
maintained in Claycomb cell culture medium. To ensure the cells
stayed alive and adherent during the force spectroscopy experiments,
4-HAP
EGFP-p.R735*
F-actin
F-actin
5 μm
F-actin
F-actin
EGFP-PKP2
EGFP-PKP2
EGFP-p.R735*
4-HAP
10 μm
Untreated
Untreated
EGFP-p.R735*
EGFP-PKP2
+ -+ -
-- + +
+ +--
0
50
100
150
Po
w
er
 (A
rb
. U
ni
ts
)
EGFP-PKP2
EGFP-p.R735*
4-HAP
C
el
ls
 w
ith
 s
ar
co
m
er
es
 (%
)
+ -+ -
-- + +
+ +--
EGFP-PKP2
EGFP-p.R735*
4-HAP
a
b
0.0
0.2
0.4
0.6
0.8
1.0
C
oh
er
en
cy
(A
rb
. U
ni
ts
)
+ -+ -
-- + +
+ +--
EGFP-PKP2
EGFP-p.R735*
4-HAP
c d
p<0.0001
p=0.0001p=0.4425
p=0.9974
p=0.6137
p=0.0141
p=0.9203
0
50
100
150
p=0.2419
p<0.0001
p<0.0001
p=0.9941
p=0.3319
p=0.0102
p=0.9963
p=0.8564
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 14
measurements were made at a constant temperature of 37 °C. Cell
studies were conducted using specially adapted Biotool Cell XXL
cantilevers (NanoandMore, Germany) with a nominal spring constant
of 0.1 N/m and a height of 15 µm; the use of extra-long tips of ∼15 µm
minimizes accidental contact of the cellwith the cantilever beambody.
The half-cone angle was ∼12°, and the nominal radius at the tip apex
was 25 nm. The upper cantilever surface was gold-coated to improve
the signal-to-noise ratio in the deflection signal. To control the force
applied on the cell, the deflection sensitivity was calibrated on a Petri
dish. The spring constant (0.092–1.1 N/m) was calculated using the
thermal noise method51.
Force-volumemaps24 were generated for whole cells by acquiring
force-distance curves (128 × 128 pixels2) over a 30 µm² area. The tip
sample distance was modulated by applying a triangular waveform.
Each individual force-distance curve was acquired at a velocity of
100 µm/s (8Hz) and a range of ∼6 µm. To prevent sample damage, the
maximum force applied to cells was 2 nN.
Maps were analyzed with in-house software written in Python. We
analyzed 8 HL-1 cells expressing PKP2 and 5 cells each expressing the
PKP2-p.R735* variant from three independent experiments. The pro-
gram includes bottom-effect corrections for a conical tip to correct for
finite cell thickness25.
HaloTag pull-down and EGFP co-immunoprecipitation
PKP2 isoformswere expressed inHEK293T cells as N-terminal Halo tag
fusion proteins. Cell pellets were lysed using Mammalian Lysis Buffer
(G9381, Promega). The bait-prey complexes, containing the PKP2-
Halo-tagged fusion protein (bait) and the potential binding partners
(prey), were pulled down using HaloLink resin (Promega Madison, WI)
and extensively washed in buffer containing 100mM Tris (pH 7.6),
150mM NaCl, 1mg/mL BSA, and 0.05% IGEPAL® CA-630 (octylphe-
noxypolyethoxyethanol, I3021, Sigma-Aldrich, Oakville, ON). Purified
bait-prey protein complexes were digested overnight with TEV pro-
tease at 4 °C to release Halo-linked PKP2 protein, and the tag-free
protein complexes were isolated with a His-Trap-Spin column.
For EGFP co-immunoprecipitation HEK293T cells were trans-
fected with plasmid expressing pEGFP-PKP2 or pEGFP-PKP2-p.R735*.
After 48 hours, cells were washed with ice-cold PBS and dislodged by
scraping. Proteins were extracted in NP-40 buffer (150mM NaCl,
50mM Tris-HCl pH 8.0, 1% NP-40, and protease and phosphatase
inhibitors). Samples were incubated with shaking for 2 h at 4 °C and
then centrifuged for 30min at 20,000 × g at 4 °C. All protein samples
were quantified by the Lowry method (BioRad), and 1mg of each
sample were used for co-immunoprecipitation (Co-IP) and input,
respectively. Co-IP was performed with Dynabeads® A (10008D,
ThermoFisher Scientific). For each sample, 50 μl Dynabeads were
washed four times with NP-40 buffer. After this, rat IgG1 anti-EGFP
(kindly provided by the Monoclonal Antibody facility at the CNIO,
Spain) were added to the Dynabeads at 1:100 dilution and incubated
with shaking for 45min at 4 °C. The antibody was then removed, and
1mg of protein per sample was added to the Dynabeads, followed by
incubation overnight with shaking at 4 °C. Unbound proteins were
removed bywashing the Dynabeads four times inNP-40 buffer. Bound
proteins were eluted with 30 μL loading buffer (10% SDS, 10mM β-
mercapto-ethanol, 20% glycerol, 200mM Tris-HCl pH 6.8, 0.05% Bro-
mophenol blue), heated for 15min at 95 °C. Proteins were western
blotted with antibodies against, MYH10 (3404 S, Cell Signaling Tech-
nology, 1:1000), and EGFP (632381, Clontech, 1:1000). Secondary
antibodies were HRP anti-mouse (ABIN6699027, Antibodies online,
1:4000) and anti-rabbit (ABIN5563398, Antibodies online,1:4000), as
appropriate. Immunoblots were developed in the iBright 1500 system.
Protein digestion, mass spectrometry, and peptide
identification
The eluted protein complexes were in-gel digested with trypsin as
described previously52, and the resulting peptides were analyzed by
liquid chromatography coupled to tandem mass spectrometry (LC-
MS/MS), using an Easy nLC-1000 nano-HPLC apparatus (Thermo Sci-
entific, San Jose, CA, USA) coupled to a hybrid quadrupole-Orbitrap
mass spectrometer (Q Exactive HF, Thermo Scientific). The dried
peptideswere takenup in0.1% (v/v) formic acid and then loadedonto a
PepMap100 C18 LC pre-column (75 μm I.D., 2 cm, Thermo Scientific)
and eluted on line onto an analytical NanoViper PepMap 100 C18 LC
column (75 μm I.D., 50 cm, Thermo Scientific) with a continuous gra-
dient consisting of 10–35% B (80% acetonitrile, 0.1% formic acid) for
60min at 200 nL/min. Peptides were ionized using a Picotip emitter
nanospray needle (New Objective). Each mass spectrometry (MS) run
consisted of enhanced FT-resolution spectra (120,000 resolution) in
the 400–1500m/z range followed by data-dependent MS/MS spectra
of the 20 most intense parent ions acquired during the chromato-
graphic run. For the survey scan, the AGC target value in the Orbitrap
was set to 1,000,000. Fragmentation in the linear ion trap was per-
formed at 27% normalized collision energy, with a target value of
100,000 ions. The full target was set to 30,000, with 1 microscan and
50ms injection time, and the dynamic exclusion was set to 0.5min.
The MS/MS spectra were searched with the Sequest algorithm in
Proteome Discoverer 1.4 (Thermo Scientific). The UniProt human
protein database (March 2017, 158,382 entries) was searched with the
following parameters: trypsin digestion with 2 maximum missed
cleavage sites; precursor and fragment mass tolerances of 800 ppm
and 0.02Da, respectively; Cys carbamidomethylation as a static
modification; and Met oxidation as a dynamic modification. The
results were analyzed using the probability ratio method53, and a false
discovery rate (FDR) for peptide identificationwas calculated based on
search results against a decoy database using the refined method54.
In-silico model of human PKP2, PKP2-p.R735*, PKP2-p.W676*,
PKP2-p.Y807*, and PKP2-p.R735*-EGFP proteins
FASTA sequence of mature human PKP2 (Uniprot Id: Q99959-2)
(https://www.uniprot.org/uniprotkb/Q99959/entry#Q99959-2) and its
mutant formsp.W646*, p.R735*, and p.Y807* were submitted to a local
implementation of I-Tasser software suite v5.144 for threading
Fig. 8 | Activation of MYH10 corrects cytoskeletal defects observed in PKP2
mutant cardiomyocytes. a Confocal representative images of F-actin (blue) in
MNCexpressing PKP2 or PKP2-p.R735* with orwithout 4-HAP treatment. Nuclei are
shown inwhite.White boxes showamagnificationof F-actinfilaments organization.
Scale bars, 10 µm and 5 µm. b Percentage of MNC with visually apparent and
organized sarcomeres. Boxes depict the 25th−75th percentile with a line showing the
median. Whiskers display minimum to maximum values; n = 54 cells for untreated
EGFP-PKP2; n = 62 cell for EGFP-PKP2 treated with 4-HAP; n = 79 cells for EGFP-
PKP2-p.R735*;n = 26 cells for EGFP-PKP2-p.R735* treatedwith4-HAPexaminedover
three independent experiments. c TTorg analysis of the captured images assessed
the organization of the sarcomeres, and this is represented by the sarcomeric
power metric. Calculated transverse organization level of sarcomeres is based on
the calculation of the peak amplitude in the Fourier spectrum of the image at the
sarcomeric frequency, Boxes depict the 25th−75th percentile with a line showing
the median. Whiskers display minimum to maximum values; n = 19 ROIs for
untreated EGFP-PKP2;n = 16 ROIs for EGFP-PKP2 treatedwith 4-HAP; n = 16ROIs for
EGFP-PKP2-p.R735*; n = 22 ROIs for EGFP-PKP2-p.R735* treated with 4-HAP.
d Analysis of cellular actin coherency, using the ImageJ plugin OrientationJ. Boxes
depict the 25th−75th percentile with a line showing the median. Whiskers display
minimum to maximum values; n = 28 ROIs for untreated EGFP- PKP2; n = 26 ROIs
for EGFP-PKP2 treated with 4-HAP; n = 23 ROIs for EGFP-PKP2-p.R735*; n = 31 ROIs
for EGFP-PKP2-p.R735* treated with 4-HAP. Statistical significance was determined
by two-way ANOVA with the Tukey multiple comparison post-test (b–d). Sig-
nificance was established as p <0.05. Data are presented as mean ± sd. μmmicro-
meters, Arb. Units arbitrary units, % percentage. Source data are provided as a
Source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 15
modeling. In each case, the best model with minimal energy and cor-
rect folding (best structural alignment to template pdb ID: 3tt9 cor-
responding to the ARM repeat; (https://www.rcsb.org/structure/3TT9)
was selected. A final cycle of refinement to minimize clashes and
energy was made with the relax tool55,56 of Rosetta suite v3.8 (www.
rosettacommons.org). The model with correct topology and minimal
score was selected as final model.
PKP2-p.R735*-EGFP model was performed in two steps, FASTA
sequence of the linker-EGFP (KPVATMVSKGEELFTGVVPILVELDGDVNG
HKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVTTLTYGVQCFSRYP
DHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIEL
KGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSV
QLADHYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVT
AAGITLGMDELYK)was submitted to a local implementationof I-Tasser
a b
c
fe
R
V
-E
F
(%
)
30
40
50
60
70
80
4-HAP          -           +              -           +
d
4-HAP          -           +              -           +
0
10
20
30
40
R
V-
ES
V 
(μ
L)
EGFP-PKP2 EGFP-PKP2 EGFP-p.R735*EGFP-p.R735*
0
10
20
30
40
50
- + - - + -
- - + - - +
- - - + + +
R
V-
ES
V 
(
L)
EGFP-PKP2
EGFP-p.R735*
4-HAP
p=0.1315 p=0.0050
p<0.0001
p=0.0767p=0.0960
p=0.0003
p=0.8954
p=0.6195
p=0.0072
20
40
60
80
100
R
V
-E
F
(%
)
p=0.1815
-
-
-
+
-
-
-
+
- + ++
+
+
- -
- -EGFP-PKP2
EGFP-p.R735*
4-HAP
p=0.0067
p=0.0003
p=0.5820 p=0.0893
p=0.0927
p=0.3824
p=0.2795
p=0.0014
0
2
4
6
8
10
D
is
pe
rs
io
n
(º
)
+ - + -
- + - +
- - + +
EGFP-PKP2
EGFP-p.R735*
4-HAP
p<0.0001 p=0.003
p=0.1321 p=0.9782
p=0.3165
0.00
0.05
0.10
0.15
0.20
Sa
rc
om
er
e 
co
nt
ra
ct
io
n 
le
ng
th
 (μ
m
)
+ - + -
-
-
+
-
-
+ +
+
EGFP-PKP2
EGFP-p.R735*
4-HAP
p=0.0062 p=0.0048
p=0.9842
p=0.9978
p=0.9482
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 16
software suite v5.1 for threading modeling as described above44. The
best model with minimal energy and correct folding (best structural
alignment to templates previously published) was selected. For mod-
eling the full fusion protein PKP2-p.R735*-EGFP a comparative mod-
eling was made using de RosettaCM tool of the of Rosetta suite v3.8
(www.rosettacommons.org) with the model before, the model of
PKP2-p.R735* modeled previously, the FASTA sequence of each one
and the alignment with the complete sequence. The model with best
structural alignment to the templates andminimal energywas selected
as final candidate. A final cycle of refinement to minimize clashes and
energy was made with the relax tool as before. The conformer with
correct topology and minimal score was selected as final model.
Immunostaining and imaging analysis
Mouse neonatal cardiomyocytes (MNC) and HL-1 cells were fixed with
4% formaldehyde for 5minutes in 0.1% Triton X-100 PBS and then
blocked and then blocked and permeabilized for 30min at room
temperaturewith 2%BSA. After that, cellswere stainedwith Phalloidin-
iFluor 594Reagent (ab176757, Abcam, 1:1000)orAlexa Fluor™Plus 647
Phalloidin (A30107, Invitrogen, 1:1000) for two hours. Cells were
washed with phosphate buffered saline before addition of DAPI
(62248, ThermoFisher Scientific, 1:1000) and were mounted in
Vectashield-mounting medium). Images were acquired using the
LIGHTNING module of a Leica TCS SP8 Scan head WLL confocal
microscope (Leica Microsystems GmbH, Germany) equipped with an
HC PL Apo CS2 63x/1.4 OIL objective. Z-stack images were captured
and presented as maximal projections. Images were processed with
IMARIS software (version 9.1.2, Oxford Instruments, UK). Cell height
wasmeasured frommaximal projections of transverse Z-stack images.
Actin and myosin filaments were analyze using ridge detection plug-
ging in Fiji software. To estimate the local orientation of the actin
fibers, we used OrientationJ (http://bigwww.epfl.ch/demo/orientation/
), which is an ImageJ plugin27. Of every pixel of the image OrientationJ
evaluates the local orientation and isotropic properties (coherency).
The coherency values reflect the direction of the actin fibers in space
regions of interest (ROI) of 20 × 20μm in size; thus, values vary from0
(completely anisotropic areas, zero directionality) to 1 (highly oriented
structures). Cell images were also subjected to Fast Fourier Transfor-
mation (FFT) for analysis of sarcomere organization (ROIs size
20 × 20μm). Sarcomeric power (the peak amplitude) in the Fourier
spectrum of cardiomyocyte images was determined using ImageJ
software (version 1.46) with a plugin TTorg (http://mirror.imagej.net/
plugins/ttorg). The higher the sarcomeric power value, the greater the
sarcomeric organization26.
To calculate ratio membrane-cytoplasm fluorescence, z-stacks
images fromMNC or HL-1 cells encoding TdTomato-PKP2, EGFP-PKP2,
EGFP-PKP2-p.R735*, PKP2-p.R735*-EGFP, EGFP-PKP2-p.W676* or EGFP-
PKP2-p.Y807* were acquired with a Leica SP8 Navigator confocal
microscope with HC PL Apo CS2 63x/1.4 OIL objective. Regions of
interest (ROIs) were drawn over maxima projections images to define
the plasma membrane and the cytoplasm, excluding the nucleus and/
or vacuoles. The ratio plasma membrane-cytoplasm intensity was
calculated to normalize the intensity of the plasma membrane to the
level of expression on each single cell.
Heart immunodetection
Hearts were collected and fixed in 4%paraformaldehyde in phosphate-
buffered saline (PBS) for overnight at 4 °C. After, these hearts were
incubated in 70% ethanol overnight at 4 °C. Samples were included in
paraffin and cross-sections (5μm) were then prepared. Tissues were
prepared and stained following the indications of theDAB substrate kit
(Vector Laboratories). Anti GFP (R1091P, ORIGENE, 1:500) was used as
the primary antibody. Images were acquired using an Olympus BX51
microscope using the fitted 10x or 20x UPlanSApo objectives and Cell
Sens Entry Ink acquisition software. Images were analyzed using Ima-
geJ and were processed for presentation with GraphPad.
Protein extraction from ventricular tissue samples of untrans-
duced, EGFP, EGFP-PKP2, EGFP-PKP2-p.R735*, and PKP2-p.R735*-EGFP
mice was performed in 125mM NaCl, 50mM Tris-HCl pH 8.0,1mM
EDTA, 1% SDS 1%NP-40, supplementedwith protease and phosphatase
inhibitors using a TissueLyser. Protein samples were quantified by the
Lowry method (BioRad), separated on 8% SDS-PAGE gels, and western
blotted with antibodies against, EGFP Living Colors (632381, Clontech,
1:1000) and GAPDH (sc-32233, Santa Cruz Biotechnology;1:2000).
Secondary antibodies were HRP anti-mouse (ABIN6699027, Anti-
bodies online, 1:4000). Immunoblots were developed in an iBright
1500 system.
Cell fractionation
After 24 h, transfected cells werewashed oncewith ice-cold PBS 1× and
after dislodged by scraping using Buffer NP40 (50mMTris-HCl pH7,5;
150mM NaCl; 1% Nonidet P 40 substitute). Cells were lysed for 30min
at 4 °C on a rotator and the lysates were cleared by centrifugation
(17,949× g for 30min at 4 °C).
Plasma membrane proteins were extracted using the Plasma
Membrane Protein Extraction Kit from Abcam. Proteins from mem-
brane and cytoplasmic extracts were western blotted with antibodies
against PKP2 (EB10841, Everest Biotech, 1:1000), N-cadherin (sc-59987,
Santa Cruz Biotechnology, 1:1000), and GAPDH (sc-32233, Santa Cruz
Biotechnology, 1:1000). Secondary antibodies were anti-goat
(ABIN2169607, Antibodies online, 1:4000) and anti-mouse
(ABIN6699027, Antibodies online, 1:4000) as appropriate. Immuno-
blots were developed in an iBright 1500 system.
4-hydroxyacetophenone (4-HAP) treatment
50 μM 4-hydroxyacetophenone was added to culture medium 48 h
before performing any experiment. Animals were intraperitoneal
treated with 1mg/kg of 4-hydroxyacetophenone for seven
consecutive days.
Statistics
Statistical analysis was performed using the GraphPad Prism 9 soft-
ware. All the experimental outcomes have been analyzed blindly.
Mouse experiments were designed tominimize the number of animals
Fig. 9 | Activation of MYH10 corrects PKP2-p.R735* contractil dysfunction.
a Right ventricular (RV) end systolic volume (ESV) and b ejection fraction (EF)
determinedbyMRI in hearts of non-transduced (sham) andAAV-EGFP-PKP2orAAV-
EGFP-PKP2-p.R735* transduced mice and treated or not with 4-HAP. Data are pre-
sented as mean ± sd; n = 6 mice (sham), n = 8 mice (EGFP-PKP2), n = 12 mice (EGFP-
PKP2-p.R735*). Statistical significance was determined by Brown-Forsythe and
Welch ANOVA tests with p <0.05 considered statistically significant. c, d Graphs
illustrate global RV-ESV and RV-EF in all animals analyzed before and after 4-HAP
administration. Red line represents mean, and dots represent individual animals
with a n = 7 (EGFP-PKP2) n = 12 (EGFP-PKP2-p.R735*). Pre-and post-4-HAP treatment
values from each mouse are connected by solid lines. e Graph showing the dis-
persion from main direction of the actin filaments in phalloidin-stained
cardiomyocytes isolated frommice transduced with AAV-EGFP-PKP2 or AAV-EGFP-
PKP2-p.R735* and treated or not with 4-HAP (n = 12 cells from mice treated with
4-HAP and n = 28 cells from untreated mice). f Graph showing the sarcomere
contraction length as the difference between themaximumandminimum length in
ten consecutive beats from isolated cardiomyocytes of AAV-EGFP-PKP2 and AAV-
EGFP-PKP2-p.R735* transducedmice, treated or not with 4-HAP. Data are presented
asmean± sem; n = 9 cells frommice treated with 4-HAP; n = 11 cells from untreated
mice from three independent experiments. Statistical significance was determined
by one-way ANOVA with Tukey’s multiple comparison post-test with p <0.05
considered statistically significant. Data are presented as mean ± sem. RV right
ventricle, ESV end systolic volume, EF ejection fraction, μL microliters, % percen-
tage, μmmicrometers, ° degrees. Source data are provided as a source data file.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 17
needed to give sufficient statistical power. Nodatawereexcluded from
the analysis. When sterically possible, data are shown in dot plots to
demonstrate data distribution and represent individual data points.
Data are shown as mean and standard deviation (sd) or standard error
of the mean (sem). Comparisons of the groups were performed by
Student t test, or Student t test witch Welch´s correction, or unpaired
two-tailed Mann–Whitney, or one-way Brown–Forsythe and Welch
ANOVA tests, one-way ANOVA or two-way ANOVA with the Tukey
multiple comparison post-test. Significance was established
as p < 0.05.
Reporting summary
Further information on research design is available in the Nature
Portfolio Reporting Summary linked to this article.
Data availability
The authors declare that the data supporting the findings of this study
are available within the paper and in the Supplementary Information.
Should any raw data files be needed in another format they are avail-
able from the corresponding author upon request. Source data are
provided with this paper.
References
1. van Tintelen, J. P. et al. Plakophilin-2 mutations are the major
determinant of familial arrhythmogenic right ventricular dysplasia/
cardiomyopathy. Circulation 113, 1650–1658 (2006).
2. Groeneweg, J. A. et al. Clinical presentation, long-term follow-up,
and outcomes of 1001 arrhythmogenic right ventricular dysplasia/
cardiomyopathy patients and family members. Circ. Cardiovasc
Genet. 8, 437–446 (2015).
3. den Haan, A. D. et al. Comprehensive desmosome mutation analysis
in north americans with arrhythmogenic right ventricular dysplasia/
cardiomyopathy. Circ. Cardiovasc. Genet. 2, 428–435 (2009).
4. Alcalde, M. et al. Stop-gain mutations in PKP2 are associated with a
later age of onset of arrhythmogenic right ventricular cardiomyo-
pathy. PLoS ONE 9, e100560 (2014).
5. Austin, K. M. et al. Molecular mechanisms of arrhythmogenic car-
diomyopathy. Nat. Rev. Cardiol. 16, 519–537 (2019).
6. Goff, Z. D. & Calkins, H. Sudden death related cardiomyopathies -
Arrhythmogenic right ventricular cardiomyopathy, arrhythmogenic
cardiomyopathy, and exercise-induced cardiomyopathy. Prog.
Cardiovasc. Dis. 62, 217–226 (2019).
7. Coelho, S. A., Silva, F., Silva, J. & Antonio, N. Athletic training and
arrhythmogenic right ventricular cardiomyopathy. Int. J. Sports
Med. 40, 295–304 (2019).
8. Maron, B. J. et al. Recommendations for physical activity and
recreational sports participation for young patients with genetic
cardiovascular diseases. Circulation 109, 2807–2816 (2004).
9. Corrado, D., Basso, C. & Judge, D. P. Arrhythmogenic Cardiomyo-
pathy. Circ. Res. 121, 784–802 (2017).
10. McKoy, G. et al. Identification of a deletion in plakoglobin in
arrhythmogenic right ventricular cardiomyopathy with palmo-
plantar keratoderma and woolly hair (Naxos disease). Lancet 355,
2119–2124 (2000).
11. Patel, D. M. & Green, K. J. Desmosomes in the heart: a review of
clinical and mechanistic analyses. Cell Commun. Adhes. 21,
109–128 (2014).
12. Al-Jassar, C., Bikker, H., Overduin, M. & Chidgey, M. Mechanistic
basis of desmosome-targeted diseases. J. Mol. Biol. 425,
4006–4022 (2013).
13. Koenderink, G. H. & Paluch, E. K. Architecture shapes contractility in
actomyosin networks. Curr. Opin. Cell Biol. 50, 79–85 (2018).
14. Grimes, K. M., Prasad, V. & McNamara, J. W. Supporting the heart:
functions of the cardiomyocyte’s non-sarcomeric cytoskeleton. J.
Mol. Cell. Cardiol. 131, 187–196 (2019).
15. Gerull, B. et al. Mutations in the desmosomal protein plakophilin-2
are common in arrhythmogenic right ventricular cardiomyopathy.
Nat. Genet. 36, 1162–1164 (2004).
16. Cruz, F. M. et al. Exercise triggers ARVC phenotype in mice
expressing a disease-causing mutated version of human
plakophilin-2. J. Am. Coll. Cardiol. 65, 1438–1450 (2015).
17. Pecci, A., Ma, X., Savoia, A. & Adelstein, R. S. MYH9: structure,
functions and role of non-muscle myosin IIA in human disease.
Gene 664, 152–167 (2018).
18. Vicente-Manzanares, M., Ma, X., Adelstein, R. S. & Horwitz, A. R.
Non-muscle myosin II takes centre stage in cell adhesion and
migration. Nat. Rev. Mol. Cell Biol. 10, 778–790 (2009).
19. Vicente-Manzanares, M., Zareno, J., Whitmore, L., Choi, C. K. &
Horwitz, A. F. Regulation of protrusion, adhesion dynamics, and
polarity by myosins IIA and IIB in migrating cells. J. Cell Biol. 176,
573–580 (2007).
20. Vicente-Manzanares, M., Newell-Litwa, K., Bachir, A. I., Whitmore, L.
A. & Horwitz, A. R. Myosin IIA/IIB restrict adhesive and protrusive
signaling to generate front-back polarity in migrating cells. J. Cell
Biol. 193, 381–396 (2011).
21. Juanes-Garcia, A., Llorente-Gonzalez, C. & Vicente-Manzanares, M.
Molecular control of non-musclemyosin II assembly.Oncotarget 7,
5092–5093 (2016).
22. Gandjbakhch, E. et al. Plakophilin 2A is the dominant isoform in
human heart tissue: consequences for the genetic screening of
arrhythmogenic right ventricular cardiomyopathy. Heart 97,
844–849 (2011).
23. White, S. M., Constantin, P. E. & Claycomb, W. C. Cardiac physiol-
ogy at the cellular level: use of cultured HL-1 cardiomyocytes for
studies of cardiacmuscle cell structure and function.Am. J. Physiol.
Heart Circ. Physiol. 286, H823–829, (2004).
24. Dufrene, Y. F. et al. Imaging modes of atomic force microscopy for
application in molecular and cell biology. Nat. Nanotechnol. 12,
295–307 (2017).
25. Garcia, P. D. & Garcia, R. Determination of the elastic moduli of a
single cell cultured on a rigid support by force microscopy. Bio-
phys. J. 114, 2923–2932 (2018).
26. Pasqualin, C., Gannier, F., Malecot, C. O., Bredeloux, P. & Maupoil,
V. Automatic quantitative analysis of t-tubule organization in car-
diac myocytes using ImageJ. Am. J. Physiol. Cell Physiol. 308,
C237–245, (2015).
27. Rezakhaniha, R. et al. Experimental investigation of collagen wavi-
ness and orientation in the arterial adventitia using confocal laser
scanning microscopy. Biomech. Model Mechanobiol. 11,
461–473 (2012).
28. Schuldt, M. et al. Proteomic and functional studies reveal detyr-
osinated tubulin as treatment target in sarcomere mutation-induced
hypertrophic cardiomyopathy. Circ. Heart Fail 14, e007022 (2021).
29. Fenix, A. M. et al. Muscle-specific stress fibers give rise to sarco-
meres in cardiomyocytes. Elife 7, e42144 (2018).
30. Friedrich, O. et al. Microarchitecture is severely compromised but
motor protein function is preserved in dystrophic mdx skeletal
muscle. Biophys. J. 98, 606–616 (2010).
31. Surcel, A. et al. Pharmacological activation of myosin II paralogs to
correct cell mechanics defects. Proc. Natl Acad. Sci. USA 112,
1428–1433 (2015).
32. Bryan, D. S. et al. 4-Hydroxyacetophenone modulates the acto-
myosin cytoskeleton to reducemetastasis.Proc.Natl Acad. Sci. USA
117, 22423–22429 (2020).
33. Surcel, A. et al. Targeting mechanoresponsive proteins in pan-
creatic cancer: 4-hydroxyacetophenone blocks dissemination and
invasion by activating MYH14. Cancer Res. 79, 4665–4678 (2019).
34. Ma, X. et al. Ablationof nonmusclemyosin II-B and II-C reveals a role
for nonmuscle myosin II in cardiac myocyte karyokinesis.Mol. Biol.
Cell 21, 3952–3962 (2010).
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 18
35. Juanes-Garcia, A. et al. A regulatorymotif in nonmuscle myosin II-B
regulates its role in migratory front-back polarity. J. Cell Biol. 209,
23–32 (2015).
36. Hodges, J. L., Newell-Litwa, K., Asmussen, H., Vicente-Manzanares,
M. & Horwitz, A. R. Myosin IIb activity and phosphorylation status
determines dendritic spine and post-synaptic density morphology.
PLoS ONE 6, e24149 (2011).
37. Tullio, A. N. et al. Nonmuscle myosin II-B is required for normal
development of the mouse heart. Proc. Natl Acad. Sci. USA 94,
12407–12412 (1997).
38. Lazzarini, E. et al. The ARVD/C genetic variants database: 2014
update. Hum. Mutat. 36, 403–410 (2015).
39. Bhonsale, A. et al. Impact of genotype on clinical course in
arrhythmogenic right ventricular dysplasia/cardiomyopathy-asso-
ciated mutation carriers. Eur. Heart J. 36, 847–855 (2015).
40. James, C. A. et al. Exercise increases age-related penetrance and
arrhythmic risk in arrhythmogenic right ventricular dysplasia/car-
diomyopathy-associated desmosomal mutation carriers. J. Am.
Coll. Cardiol. 62, 1290–1297 (2013).
41. La Gerche, A. & Claessen, G. Increased flow, dam walls, and
upstream pressure: the physiological challenges and atrial con-
sequences of intense exercise. JACC Cardiovasc. Imaging 9,
1389–1391 (2016).
42. Guo, Y. & Pu, W. T. Cardiomyocyte maturation: new phase in
development. Circ. Res. 126, 1086–1106 (2020).
43. Ahmed, R. E., Tokuyama, T., Anzai, T., Chanthra, N. & Uosaki, H.
Sarcomere maturation: function acquisition, molecular mechan-
ism, and interplay with other organelles. Philos. Trans. R. Soc. Lond.
B Biol. Sci. 377, 20210325 (2022).
44. Yang, J. et al. The I-TASSER Suite: protein structure and function
prediction. Nat. Methods 12, 7–8 (2015).
45. Roche-Molina, M. et al. Induction of sustained hypercholester-
olemia by single adeno-associated virus-mediated gene transfer
of mutant hPCSK9. Arterioscler. Thromb. Vasc. Biol. 35,
50–59 (2015).
46. Cruz, F. M. et al. Exercise triggers arrhythmogenic right ventricular
cardiomyopathy phenotype in mice expressing a disease-causing
mutated version of human plakophilin-2 after single adeno-
associated virus-mediated gene transfer. J. Am. Coll. Cardiol. 65,
1438–1450 (2015).
47. Hauswirth, W. W., Lewin, A. S., Zolotukhin, S. & Muzyczka, N. Pro-
duction and purification of recombinant adeno-associated virus.
Methods Enzymol. 316, 743–761 (2000).
48. Claycomb, W. C. et al. HL-1 cells: a cardiac muscle cell line that
contracts and retains phenotypic characteristics of the adult car-
diomyocyte. Proc. Natl Acad. Sci. USA 95, 2979–2984 (1998).
49. Wolterink-Donselaar, I. G., Meerding, J. M. & Fernandes, C. A
method for gender determination in newborn dark pigmented
mice. Lab Anim. 38, 35–38 (2009).
50. Cadinanos, J. & Bradley, A. Generation of an inducible and opti-
mized piggyBac transposon system. Nucleic Acids Res. 35,
e87 (2007).
51. Lozano, J. R., Kiracofe, D., Melcher, J., Garcia, R. & Raman, A. Cali-
bration of higher eigenmode spring constants of atomic force
microscope cantilevers. Nanotechnology 21, 465502 (2010).
52. Bonzon-Kulichenko, E. et al. A robust method for quantitative high-
throughput analysis of proteomes by 18O labeling. Mol. Cell Pro-
teom. 10, M110 003335 (2011).
53. Martinez-Bartolome, S. et al. Properties of average score distribu-
tions of SEQUEST: the probability ratio method.Mol. Cell Proteom.
7, 1135–1145 (2008).
54. Navarro, P. & Vazquez, J. A refined method to calculate false dis-
covery rates for peptide identification using decoy databases. J.
Proteome Res. 8, 1792–1796 (2009).
55. Nivon, L. G., Moretti, R. & Baker, D. A Pareto-optimal refine-
ment method for protein design scaffolds. PLoS ONE 8,
e59004 (2013).
56. Conway, P., Tyka, M. D., DiMaio, F., Konerding, D. E. & Baker, D.
Relaxation of backbone bond geometry improves protein energy
landscape modeling. Protein Sci. 23, 47–55 (2014).
Acknowledgements
This studywas supportedbyMCIUgrant BFU2016-75144-R andPID2020-
116935RB-I00, and by a “la Caixa” Banking Foundation grant under the
project code HR18-00304” to J.A.B.; The study was also supported by
the “Ayudas a la Investigación Cátedra Real Madrid-Universidad Eur-
opea” (2017/RM01). C.M.-L. and S.S. hold MCIU predoctoral contracts
BES-2017-079715, and BES-2017-079707 respectively. R.G. acknowl-
edges funding from the European Research Council under grant ERC-
AG-340177 (3DNanoMech) and from the MCIU under grant MAT2016-
76507-R. The CNIC is supported by the Instituto de Salud Carlos III
(ISCIII), the Ministerio de Ciencia e Innovación (MCIN) and the Pro CNIC
Foundation and is a Severo Ochoa Center of Excellence, grant
CEX2020-001041-S funded by MICIN/AEI/10.13039/501100011033. The
microscopy experiments were carried out at the Dynamic Microscopy
and Image Unit, CNIC, ICTS-ReDib, co-financed byMCIN/AEI /10.13039/
501100011033 and FEDER “A way of making Europe” (#ICTS-2018-04-
CNIC-16). Imaris full analysis were carried out at the Microscopy &
Dynamic Imaging, CNIC, ICTS-ReDib, co-funded by MCIN/AEI
/10.13039/501100011033. Biomedical Imaging has been conducted at
the Advanced Imaging Unit of the CNIC (Centro Nacional de Investiga-
ciones Cardiovasculares Carlos III), Madrid, Spain. This project used the
ReDIB ICTS infrastructure TRIMA@CNIC, Ministerio de Ciencia e Inno-
vación (MCIN).
Author contributions
N.G.-Q., S.S., C.M.-L., and J.A.B. designed and interpreted the experi-
mental work to characterized the PKP2 mutant models, and cells with
the help of C.S.-R., A.G.-G., M.L., M.A.-S., and M.I.-G. C.M.-L., D.M.-P.,
D.S.-R., M.R.-M., and A.G.-G. constructed the plasmids and generated
the cell lines. F.MdB. generated the in silico protein models. E.C. was
responsible for mass spectrometry analyses. A.G.-G., D.M., M.L., and
M.I.-G. helped with the animal work. D.S. performed the AFM experi-
ment and analyzed the data. R.G. designed, analyzed, and interpreted
the AFM experiment. J.A.B. conceived the project and wrote the
manuscript. All authors discussed the results and commented on the
manuscript.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s41467-023-41981-5.
Correspondence and requests for materials should be addressed to
Juan A. Bernal.
Peer review information Nature Communications thanks the anon-
ymous reviewer(s) for their contribution to the peer review of this work.
A peer review file is available.
Reprints and permissions information is available at
http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jur-
isdictional claims in published maps and institutional affiliations.
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 19
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons license, and indicate if
changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons license and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2023
Article https://doi.org/10.1038/s41467-023-41981-5
Nature Communications |         (2023) 14:6461 20