Super-resolution imaging reveals stretch-induced architectural rearrangement of desmoplakin in desmosomes

To the Editor:

Desmosomes are intercellular junctions that confer mechanical resilience to tissues including the heart, epidermis, and other epithelial tissues. Central to this function is the plakin protein desmoplakin (DP), which anchors the desmosomal plaque to the keratin intermediate filament cytoskeleton in skin. Pathogenic DP impairs keratinocyte adhesion and structural integrity, resulting in cutaneous disorders such as striate palmoplantar keratoderma and lethal acantholytic epidermolysis bullosa, and systemic disease exemplified by Carvajal syndrome (Perl et al., 2024). Affected individuals typically present with increased skin fragility, hyperkeratosis, and blistering under mechanical stress, underscoring the importance of investigating desmosome biomechanics. Structurally, desmosomes have a tripartite organization: transmembrane desmosomal cadherins mediate binding in the extracellular domain (ECD); the outer dense plaque (ODP) contains plakoglobin, plakophilins, and the N-terminal DP head domain; and the inner dense plaque (IDP) includes the DP tail domain, which binds keratin to form a cytoskeletal linkage (Fig. 1a). This IDP connection supports a tensile filament network that distributes mechanical load and resists forces from stretching, friction, and pressure (Broussard et al., 2020). However, the structural response of desmosomes to mechanical forces and the mechanisms through which force is propagated across the desmosome remain poorly understood.

Figure 1. Desmoplakin architecture undergoes nanoscale remodeling in human keratinocytes in response to uniaxial mechanical stretch.

(a) Schematic of desmosome organization highlighting desmoplakin (DP) domains, plaque proteins (plakoglobin (PG), plakophilin (PKP)), and cadherins (desmoglein (DSG), desmocollin (DSC)). (b) Diagram of experimental setup showing NHEKs cultured on flexible membranes under static or stretch conditions using a vacuum-based Flexcell system. (c) Finite element modeling of uniaxial stretch across the cell culture area indicates a uniform applied stress of 55 kPa. (d) Widefield images of DPI (magenta) and DAPI (cyan) in static vs stretched cells, scale bar = 20 μm. (e) Corresponding dSTORM super-resolution images reveal resolved desmosomal plaques, with zoomed insets highlighting single desmosomes, scale bars = 5 μm, 500 nm. (f) Representative linescans and plaque-to-plaque (PtP) distances for a single desmosome from static and stretched cells, scale bar = 250 nm. (g) Quantification reveals a significant increase of PtP distance following stretch (static: 246 nm (211-264), n=44; stretch: 264 nm (232-286), n=27; p=0.03). Data is from three independent experiments, and each n represents a desmosome. Values reported as median (IQR) with a Mann-Whitney U test for significance. p < 0.05 (*).

Recent work has revealed DP serves not only as a structural anchor but also as a regulator of desmosome adhesive function, reflecting dynamic adaptations to biochemical cues. DP changes architecture during maturation and hyperadhesive transitions and in response to actomyosin generated forces and is mechanically sensitive (Beggs et al., 2022, Dong et al., 2025, Sadhanasatish et al., 2023, Stahley et al., 2016). Super-resolution direct stochastic optical reconstruction microscopy (dSTORM) enables 20 nm resolution imaging of individual desmosomal plaques not resolvable by conventional microscopy thereby allowing quantitative mapping of individual desmosome architecture. Plaque-to-plaque (PtP) distances measured perpendicular to the plasma membrane consistently place the DP head domain near the plasma membrane, with the tail domain extending into the IDP variably based on rod length and desmosome maturation, supporting a model where DP is arranged at an angle within the plaque (Ainslie et al., 2025, Beggs et al., 2022, Stahley et al., 2016).

Here we examine how mechanical stretch affects desmosome architecture in keratinocytes. By integrating biomechanical perturbation with dSTORM, we investigate whether external forces drive the nanoscale reorganization of DP. We hypothesized that physiological stretch modulates DP tail domain position within the IDP, and that this change may enable desmosomes to absorb mechanical strain and maintain cell adhesion.

To test this hypothesis, normal human epithelial keratinocytes (NHEKs) were grown on flexible membranes and subjected to 13% uniaxial stretch for 30 minutes or unstretched to serve as control (Fig. 1b-c). The direction and magnitude of stretch were selected to ensure uniform strain across keratinocytes and to approximate the mid-range of physiological stretch experienced by the epidermis (Gudipaty et al., 2017, Lutz et al., 2020). While held in the stretched position, cells were fixed and immunolabeled with an antibody targeting the DPI rod domain (AA 1900-1950). Widefield fluorescence microscopy revealed punctate DP staining at cell borders, corresponding to desmosomes, in both groups (Fig. 1d). dSTORM resolved distinct desmosome plaque pairs, representing DPI rod domains contributed by adjacent cells (Fig. 1e-f). To determine whether stretch altered DP architecture, we quantified the distances between the two DPI plaques in individual desmosomes. We observed a significant 18 nm increase in PtP distance in stretched NHEKs compared to static controls (Fig. 1g). Because the response may vary with orientation, only desmosomes with midlines perpendicular to the direction of stretch were included in the analysis.

To further investigate how mechanical stretch influences desmosome architecture, we next examined the spatial arrangement of DP isoforms: DPI, DPIa, and DPII. These isoforms differ in length of their central rod domain, and are expressed in the epidermis, with DPI and DPII being the predominant forms. Although their precise functional differences remain unclear, loss of a single isoform has been associated with weakened intracellular adhesion and skin fragility (Cabral et al., 2012).

Because all three isoforms are identical in sequence outside of their rod domain length, current antibodies cannot distinguish between them. Therefore, we stably expressed each isoform with a C-terminal mEGFP tag in DP knockout HaCaT cells (Fig. 2a) (Ainslie et al., 2025). This enabled us to independently assess the isoform specific structural response to mechanical stretch. Cells were seeded on the flexible membranes and subjected to 13% uniaxial strain for 30 minutes or kept under static conditions. After fixation, cells were labeled with an anti-mEGFP and imaged by dSTORM. In all three cell lines, mechanical stretch resulted in a significant increase in PtP distance, indicating that each isoform undergoes an architectural change following stretch (Fig. 2b,c). These findings suggest all DP isoforms undergo stretch-induced remodeling and that mechanosensitivity is an inherent property of DP, regardless of isoform.

Figure 2. Mechanical stretch induces isoform-specific conformational changes in desmoplakin.

(a) Schematics of DP isoforms (DPI, DPIa, and DPII) fused to mEGFP and labeled with an anti-GFP nanobody above confocal images of DP KO HaCaT cells stably expressing each isoform showing DP-mEGFP localization at cell-cell borders, scale bar = 25 μm. (b) dSTORM images of individual desmosomes from cells expressing DPI, DPIa, or DPII-mEGFP under static or 13% uniaxial stretch conditions, scale bar = 200 nm. (c) Quantification of plaque-to-plaque (PtP) distances revealed a significant increase in the spacing between DP C-termini following stretch across all isoforms (DPI static: 192 nm (160-220), n = 19; DPI stretch: 216 nm (196-230), n = 25; p=0.03; DPIa static: 156 nm (140-175), n = 20; DPIa stretch: 172 nm (164-180), n = 23; p=0.01; and DPII static: 144 nm (106-162), n = 29; DPII stretch: 160 nm (144-181), n = 38; p=0.002). (d) dSTORM images of individual desmosomes from DPII N-term under static or 13% uniaxial stretch, scale bar = 200 nm. (e) No significant change was observed in PtP distance measured from the DP N-terminus in DPII cells (static: 86 nm (80-92), n=20; stretch: 80 nm (76-92), n=20; p=0.55). Data is from three independent experiments, and each n represents a desmosome. Values reported as median (IQR) with a Mann-Whitney U test for significance for each isoform. p ≥ 0.05 (not significant (ns), p < 0.05 (*), p < 0.01(**)).

Finally, to test whether the increase in C-terminal PtP distance reflects a displacement of DP within the plaque, we labeled stretched and control DPII-mEGFP HaCaT cells with an antibody targeting the DP head domain (Fig. 2d). We found no significant difference between stretched and static conditions (Fig. 2e).

We found that mechanical stretch significantly increased the distance between the tail domains of DP compared to static controls. This suggests DP tail domain localization may serve as a primary read-out of mechanical adaptation with the desmosome. The increased PtP distance is unlikely to result from overall displacement of DP away from the plasma membrane, which would be expected to produce a comparable shift in the DP head domain, although this cannot be excluded due to limited data.

The magnitude of the observed change in tail domain position corresponds to approximately 5-8% of the total DP length, consistent with a modest conformational change rather than large-scale molecular extension. Several structural mechanisms may account for this. One possibility is a change in the angular orientation of DP, with the long axis becoming more perpendicular to the plasma membrane, similar to changes proposed in maturation and hyperadhesion (Beggs et al., 2022, Stahley et al., 2016). Assuming the observed differences arise primarily from such a change in molecular orientation, a simple geometric estimate (Ainslie et al., 2025) suggests a shift in DP orientation of approximately 4–11° (Supplementary Fig. 1). Alternatively, force could induce secondary or tertiary structural changes to the coiled-coil rod domain, or alter head domain conformation, as reported in response to actomyosin-generated tension in cancer cells and cardiomyocytes (Dong et al., 2025).

The mechanical responsiveness of DP to external strain across all isoforms underscores a conserved role in adhesion under stress. We observed a slight difference in the magnitude of PtP changes between isoforms, which may reflect isoform-dependent mechanical compensation arising from differences in rod length. More broadly, changes in DP architecture may serve as a molecular buffer that preserves epidermal cohesion and barrier integrity during mechanical challenge. Understanding these nanoscale responses provides critical insight into how biomechanical dysfunction of DP contributes to skin-fragility and blistering disorders and may inform future strategies for the treatment of skin diseases driven by mechanical stress.

Abbreviations:

AA

(amino acid)

DP

(desmoplakin)

ECD

(extracellular domain)

IDP

(inner dense plaque)

NHEK

(normal human epithelial keratinocyte)

ODP

(outer dense plaque)

PtP

(plaque-to-plaque)

Data availability statement:

Data is available at https://doi.org/10.6084/m9.figshare.32060250