Spectrin coordinates cell shape and signaling essential for epidermal differentiation

Cells in our body exhibit a wide range of shapes, sizes, and functions, from giant, multinucleated osteoclasts and contractile skeletal muscle cells to small, disc-shaped red blood cells (Luxenburg et al., 2007; Calderón et al., 2014; Elgsaeter et al., 1986). Defects in cell shape often correlate with tissue malfunction and disease phenotypes (Lee et al., 2016; Kalluri and Weinberg, 2009; Delaporte et al., 1990; Emery, 2002; PAULING and ITANO, 1949). For example, in common skin diseases such as psoriasis, atopic dermatitis, or squamous cell carcinoma skin, keratinocytes exhibit abnormal cell shape associated with impaired differentiation and compromised epidermal barrier function (Koegel et al., 2009; Van Der Kammen et al., 2017). However, whether cell shape directly regulates cell differentiation and tissue function remains an open question.

At the body’s surface, the interfollicular epidermis (hereafter, epidermis) is a stratified, multilayered epithelium that provides a barrier to protect organisms from water loss and external challenges. The epidermis exhibits progressive cell shape changes that are linked to the differentiation status of keratinocytes. Basal cuboidal stem cells, when initiating differentiation, move upward into the suprabasal spinous layer (SS) to begin flattening and increase in size. Upon entering the first of the three granular layers, SG3, keratinocyte flattening is substantially increased to take on the shape of Kelvin’s flattened tetrakaidecahedrons in SG2, which is essential to maintain the tight junctional (TJ) barrier in this layer (Kubo et al., 2009; Yoshida et al., 2013; Yokouchi et al., 2016; Rübsam et al., 2017). While passing through the SG1 cells take on their final squamous shape that coincides with terminal differentiation. This process involves enucleation and transglutaminase (TGM)-mediated cross-linking of structural proteins and barrier lipids necessary to form a functional stratum corneum barrier that physically separates the organism from the environment (Simpson et al., 2011; Fuchs, 2007). The mammalian epidermis is thus an ideal model to address the relationship between cell position, shape, and differentiation.

A key determinant of cell shape is the actomyosin cytoskeleton and its associated adhesion receptors. The actomyosin cytoskeleton forms a contractile network made up of actin fibers (F-actin), myosin II motor proteins, and dozens of actin-binding proteins that regulate its organization and dynamics to control the contractile mechanical state of cells (Pollard, 2016). By engaging integrin-based cell–matrix adhesions and cadherin-based adherens junctions (AJs), actomyosin-generated forces are transmitted across cells and tissues (Noordstra et al., 2023; Saraswathibhatla et al., 2023). Moreover, actomyosin activity also modifies signaling cascades involved in fate regulation (Luxenburg and Zaidel-Bar, 2019; Pollard, 2016; Zaidel-Bar et al., 2015). In the epidermis, an increase in cortical F-actin levels accompanies differentiation and the gradual formation of squamous shapes, a process regulated by E-cadherin (Rübsam et al., 2017). Yet, how cortical actomyosin networks are spatially patterned across layers and how these changes are linked to differentiation remain poorly understood.

Spectrins are tetramers consisting of two α- and two β-spectrins. Vertebrates have two α-spectrin (αI and αII) and five β-spectrin (βI–βV) variants encoded by different genes. While αI-spectrin is expressed only in erythrocytes, αII-spectrin (encoded by Sptan1) is the only α-spectrin in non-erythrocyte cells that can interact with each of the five β-spectrins (Bennett and Healy, 2009; Bennett and Lorenzo, 2016; Teliska and Rasband, 2021) to form antiparallel tetramers that integrate into the cortical actomyosin network in diverse cell types, including erythrocytes, neurons, and fibroblasts (Leterrier and Pullarkat, 2022). These tetramers display tensile, spring-like behavior that in combination with a dynamic interplay with myosin enable spectrin–actin networks to absorb shocks, sense and generate forces, and regulate their elastic properties that together control cell shape (Leterrier and Pullarkat, 2022; Lorenzo et al., 2023; Ghisleni et al., 2020). Moreover, the organization of spectrin-actomyosin networks promotes microdomain formation of transmembrane proteins, e.g., ion channels, in the plasma membrane (Lorenzo et al., 2023).

The function of spectrin in epithelia remains less well defined. As in erythrocytes, spectrin controls cortical actomyosin organization and cell shape in Drosophila follicular epithelium and in Caenorhabditis elegans epidermis (Praitis et al., 2005; Ng et al., 2016). In mice, Sptan1 knockout is embryonic lethal, characterized by neural tube, cardiac, and craniofacial defects as well as abnormal growth (Stankewich et al., 2011). In cultured keratinocytes, spectrin was implicated in regulating early keratinocyte differentiation (Zhao et al., 2011, 2013; Wu et al., 2015), but whether spectrins control in vivo epithelial cell shapes and regulate epithelial differentiation and barrier formation is not known.

In the present study, we identify spectrin as an essential component of the epidermal actomyosin cytoskeleton. We show that distinct actin and spectrin gradients drive changes in the conformation, mechanics, and signaling properties of this submembranous network to guide transitions in cell shape and fate when cells move into a new layer. E-cadherin orchestrates this layer-specific organization of the spectrin-actomyosin cortex to control tension states and membrane nanoscale organization, resulting in the spatial activation of EGF receptor (EGFR) and the calcium channel TRPV3 in the uppermost epidermal layers to promote terminal differentiation and thus epidermal barrier formation. These observations provide novel insights into how dynamic changes in cortical organization integrate mechanics and signaling to coordinate cell fate and cell shape in a physiologically relevant mammalian system.

Although it is well established that keratinocytes adopt an increasingly squamous morphology during stratification, their exact in vivo 3D shapes and volumes have not been defined. To this end, we used a K14-Cre–driven rainbow transgene (Snippert et al., 2010) to fluorescently label single cells in vivo, followed by 3D rendering of the cytoplasmic YFP signal to determine cell shape, size, and volume (Fig. 1, a and b; and Fig. S1 a). This analysis showed that initially cuboidal basal keratinocytes double in both size and volume twice, first when transiting into the spinous layer and again when moving into the granular layer during stratification. (Fig. 1, a–c and Fig. S1 b). Only upon transitioning into fully squamous corneocytes, the size-to-volume ratio is increased due to a further expansion of surface area (1.41-fold). Thus, keratinocytes undergo highly robust and highly stereotypic changes in cell size, shape, and volume.

We next asked whether epidermal cell shape depends on intercellular contacts, as AJs are key determinants of cell shape (Luxenburg and Zaidel-Bar, 2019; Niessen et al., 2024). To this end, we first separated the SS and the SG3 layer from the late-differentiated (SG2/SG1) layers (Matsui et al., 2021) to then isolate individual cells. While SS/SG3 cells rounded up, SG2/SG1 cells maintained their characteristic flattened tetrakaidecahedron shape also upon loss of cell–cell contacts (Fig. 1, d and e; and Fig. S1 a) (Yokouchi et al., 2016).

The ability of single SG1/SG2 cells to maintain their in vivo shape is reminiscent of the characteristic red blood cell shape for which the spectrin cortical cytoskeleton is essential (Leterrier and Pullarkat, 2022). Immunostaining for non-erythrocyte αII-spectrin and F-actin in epidermal whole mounts and sections from newborn mice revealed that spectrin partially colocalized with F-actin to form cortical micro honeycomb-like lattices in suprabasal layers (Fig. 1, f and g; and Fig. S1 c). Whereas αII-spectrin cortical enrichment was highest in the SG3 layer (Fig. 1 h) to then drop again, F-actin localization at the cortex became increasingly more intense in upper layers (Rübsam et al., 2017), peaking in SG1 cells (Fig. 1 i), suggesting that distinct spectrin–actin conformations determine layer-specific cell shapes. In agreement, first suprabasal αII-spectrin enrichment correlated with the initial flattening of suprabasal cells at embryonic (E) day 15 (Fig. S1 d, arrow) and precedes stratum corneum formation (Rübsam et al., 2017). Thus, spectrin is an integral part of the keratinocyte actin network, and stratification-induced cell shape transitions go along with changes in cortical F-actin spectrin ratios.

We thus asked directly whether αII-spectrin regulates epidermal cell shape and depleted αII-spectrin during epidermal development using lentiviruses encoding two different Sptan1 short hairpins (shSptan1-0595 or shSptan1-9753) or a control shRNA (shScr) together with a GFP-tagged histone 2B reporter (H2B-GFP) to identify transduced cells (Fig. S1, e–g) (Beronja et al., 2010). As a readout for cell shape, we then quantified the cell area and the cell shape index (perimeter [P]/√area [A]) (Sahu et al., 2020) of each layer. Depletion of αII-spectrin resulted in less-flattened suprabasal cells, as indicated by an increased cell area and a decrease in shape index (Fig. 1, h and i). Basal spectrin-depleted cells also increased in size and became even more spherical. Inactivation of αII-spectrin in all epidermal cells using keratin14-Cre (Sptan1^epi−/−) (Fig. S1, h and i) confirmed changes in cell shape in newborn epidermis (Fig. S1, j and k). Thus, αII-spectrin regulates keratinocyte shape in all layers of the epidermis.

We next asked how spectrin is recruited to the cortex in keratinocytes. One candidate is ankyrin, a spectrin-binding partner that enables its assembly into networks (Bennett and Healy, 2009; Bennett and Lorenzo, 2016). Ankyrin also interacts with E-cadherin (Kong et al., 2023; Kizhatil et al., 2007; Bennett and Lorenzo, 2013). Immunostaining of E17.5 embryos for ankyrin G (encoded by Ank3) that is expressed in the developing epidermis (Sennett et al., 2015) revealed cortical enrichment in the granular layer, which is lost upon spectrin depletion (Fig. 2, a, c, and d). In contrast, Ank3 depletion in E17.5 embryos (Fig. S2, a–c) did not alter spectrin localization or levels (Fig. 2, a and b), with no changes in F-actin localization and cell shape (Fig. S2, d–g).

E-cadherin–based AJ can recruit spectrin to the cortex (Pradhan et al., 2001). Epidermal inactivation of E-cadherin (Ecad^epi−/−) (Tunggal et al., 2005) resulted in loss of αll-spectrin from the cortex of suprabasal layers (Fig. 2 e), whereas F-actin remained cortical but with increased intensity in the spinous layer (Rübsam et al., 2017). In contrast, E-cadherin maintained its localization in αII-spectrin–depleted epidermis (Fig. 1 j). Consistently, αII-spectrin was also mislocalized in cultured E-cadherin depleted keratinocytes (Fig. S2 h), whereas depletion of αII-spectrin did not affect AJ assembly in vitro (Fig. S2 i).

We then asked whether cortical recruitment of spectrin is necessary to control cell shape. Like spectrin depletion, epidermal loss of E-cadherin altered cell shapes in all layers with a reduced flattening of stratum granular (SG) cells (Fig. 2, f and g). Isolated E-cad^−/− SG1/SG2 cells were also smaller but maintained their flattened tetrakaidecahedron cell shape, even when treated with latrunculin B to depolymerize F-actin (Fig. 2, h and i; and Fig. S2, j and k), indicating that once the shape of SG2 cells is set this shape becomes independent of AJ or F-actin. In contrast to spectrin depletion, Ecad^−/− spinous layer cells showed increased flattening, perhaps due to the increased cortical actin in this layer. Together, these data indicate that E-cadherin, through polarized organization of the spectrin-actomyosin cortex, regulates transitions in cell shape that guide terminal differentiation when cells move up through the epidermis.

We next asked whether F-actin and spectrin coordinate the organization of the epidermal cortex into a honeycomb lattice. Treatment of E17.5 embryos with low levels of latrunculin B to reduce F-actin but not disturb E-cadherin adhesion (Fig. S3 a) reduced cortical αII-spectrin levels in all layers of the epidermis (Fig. 3, a and b). Conversely, embryonic KD of αII-spectrin reduced suprabasal cortical F-actin levels in E17.5 embryos (Fig. 3, c and d). High-resolution imaging of newborn epidermal whole mounts revealed a striking loss of F-actin honeycomb lattices in Sptan1^epi−/− mice, with over 60% of suprabasal cells now showing a more streak- or spot-like organization (Fig. 3, e and f, arrow), indicating that spectrin not only recruits but is essential to properly organize F-actin at the cortex.

We then asked how AJ assembly initiates the organization of the actin-spectrin network by switching primary keratinocytes to high Ca²⁺ (1.8 mM) to induce contact formation and stratification. Upon initial E-cadherin engagement (6 h Ca²⁺), only F-actin but not spectrin was recruited to early AJ, so-called AJ zippers (Fig. 3, g and h), to form peri-junctional rings (Vasioukhin et al., 2000) (Fig. 3, g and h; and Fig. S3 b). Consistently, αII-spectrin knockdown did not affect initial E-cadherin engagement or actin recruitment (6 h Ca²⁺, Fig. S2 i).

Stratification of keratinocytes into a multilayer (48 h Ca²⁺) upregulated spectrin protein levels (Fig. S3 c), in agreement with the in vivo suprabasal increase in cortical spectrin (Fig. 1, g and h). As seen in vivo, spectrin co-localized with F-actin and E-cadherin to form micro honeycomb-like cortical F-actin-spectrin lattices (Fig. 3 g, arrow, h; Fig. S3, b and d). Spectrin was also localized to the F-actin cortical ring that supported barrier-forming TJs in the uppermost apical cell layer, albeit to a much lesser extent than F-actin (Fig. 3, g and h, arrowhead; Fig. S3 b) (Rübsam et al., 2017).

Low levels of latrunculin B reduced cortical αII-spectrin recruitment, similar to in vivo, and disturbed the honeycomb organization (Fig. 3 i and Fig. S3 e). Strikingly, depletion of αII-spectrin also resulted in a loss of cortical honeycomb lattices, with F-actin now showing a linear fiber organization reminiscent of F-actin stress fibers (Fig. 3 j). In contrast, αII-spectrin KD did not prevent formation of the apical TJ-associated F-actin cortical ring even though F-actin levels were reduced (Fig. S3, h and i).

Thus, F-actin and spectrin spatiotemporally coordinate the organization of the cortex during stratification. Upon initial adhesion, AJs first recruit F-actin and then reorganize the F-actin cortex (Vaezi et al., 2002) to allow recruitment of αII-spectrin necessary to further restructure the cortex into F-actin-spectrin lattices in suprabasal layers.

The observed reorganization of F-actin into stress fiber-like structures upon loss of αII-spectrin suggested a change in myosin recruitment and activity. The cortex of keratinocytes is highly enriched for the actin-binding protein myosin II that regulates contractile states of the actin cytoskeleton (Dor-On et al., 2017; Sumigray et al., 2012). In Drosophila, loss of either β- or α-spectrin increased junctional myosin levels and activity (Deng et al., 2015; Forest et al., 2018; Ibar et al., 2023). We thus examined where at the cortex myosin II was recruited. Whole-mount imaging for non-muscle myosin heavy chain IIa (myosin-IIa) revealed an enrichment of myosin-IIa mostly at the lateral junctions in the spinous layer (Fig. 4 a, SS). Staining increased in the granular layer with myosin-IIa, now mostly located to spots that decorated the actin-spectrin honeycomb lattices (Fig. 4 a, SG3/2), thus indicating that cells increase their tensile state when moving from the spinous to the granular layer. Loss of spectrin increased myosin intensity and spot size, suggesting an increase in the contractile state of F-actin networks (Fig. 4 a), similar to Drosophila. Vice versa, inhibition of myosin contractility using the Rho-associated protein kinase inhibitor, Y27632, decreased cortical αII-spectrin levels in E17.5 embryos (Fig. 4, b and c), thus indicating that spectrin and myosin II jointly coordinate organization and tensile states of the F-actin cortex.

We next used primary keratinocytes to assess whether spectrin-dependent recruitment of myosin II altered the tensile state of the cortex. Co-staining of F-actin, αII-spectrin, and myosin-IIa in stratified control keratinocytes also revealed a spot-like integration of myosin IIa into the spectrin-F-actin lattices. Upon depletion of αII-spectrin, the now linear F-actin fibers (Fig. 3 j) were intensely decorated with myosin-IIa in a highly periodic pattern, indicating that these fibers are highly contractile (Fig. 4 d and Fig. S4 a). Lowering myosin II motor activity (blebbistatin, 5 µM) in control and spectrin-depleted keratinocytes was sufficient in both to reorganize honeycomb spectrin-F-actin lattices into ultrafine, almost diffuse F-actin structures (Fig. 4 e and Fig. S4 b), except upon spectrin depletion, their appearance was more anisotropic and streak-like (Fig. 4, e and f, arrow; Fig. S4 b), thus resembling the in vivo F-actin defects observed in the αII-spectrin–deficient epidermis (Fig. 3 e). Together, these data show that tension is necessary to properly assemble spectrin and actin into a honeycomb network, with spectrin controlling its tension state, albeit to different extents in vivo versus in vitro.

We then assessed the mechanical properties of the cortex by performing linear laser ablation experiments to measure strain-dependent recoil in stratified keratinocytes. Prior to the formation of cortical spectrin-actomyosin lattices, the ablation of F-actin–positive intercellular junctions elicited a viscoelastic recoil response of junctional vertices as described for simple epithelial cells (Fig. S4, c and d) (Wu et al., 2014). In contrast, virtually no recoil of junctional vertices was observed once spectrin-actomyosin lattices were formed (Fig. S4, c and d). Either subsequent ablation of the lattices themselves adjacent to the already-ablated junctions or ablation of only these lattices induced a strong recoil response with elliptical openings that increased in size over time (Fig. 4, g–i; and Fig. S4, e and f). These data indicate that the cortical spectrin-actomyosin lattices exhibit viscoelastic properties as shown previously in zebrafish and C. elegans (Saha et al., 2016; Thi Kim Vuong-Brender et al., 2017). Interestingly, curved intercellular borders straightened upon ablation, showing that the spectrin-actomyosin lattice also regulated the tension state of AJs to control cell shape (Fig. S4 g). Importantly, depleting αII-spectrin resulted in a faster, two-fold larger recoil with highly irregular openings (Fig. 4, j and k), which was reversed upon blebbistatin treatment (Fig. 4, l and m). Further, E-cadherin loss induced similar recoil behavior with highly irregular, larger openings (Fig. 4, n and o). Taken together, these data suggest a model in which E-cadherin, through recruitment of spectrin, orchestrates distinct cortical networks in each layer to dissipate the myosin-driven increase in tension and increase the stability of the cortex to prevent damage.

Although suprabasal flattening is a hallmark of differentiation, whether these changes in cell shape promote epidermal differentiation is still unclear (Simpson et al., 2011; Luxenburg and Zaidel-Bar, 2019). We thus asked whether depletion of αII-spectrin also changes epidermal differentiation. In E17.5 control embryos, staining for keratin (K)14 only marked the basal layer, whereas in Sptan1^KD epidermis, suprabasal cells were also K14⁺ (Fig. 5 a arrow; Fig. S5, a and b). In agreement, EdU incorporation assays showed that suprabasal proliferation was increased threefold compared with control (Fig. S5, d and e). In contrast, the suprabasal marker K10 was not obviously altered, indicating at least a partial induction of differentiation (Fig. 5 a and Fig. S5 a). Basal spindle orientation that can regulate cell fate (Williams et al., 2011) was also not altered (Fig. S5 c). Thus, in addition to cell shape, spectrin is critical for initiating basal cell differentiation upon translocating suprabasally. We next asked whether the ability to induce differentiation is linked to changes in cell shape. Although both K14⁺- and K14⁻ suprabasal cells showed an increase in cell area compared with control suprabasal cells (Fig. 5 b), only K14⁺ but not K14⁻ suprabasal cells exhibited a defect in flattening, suggesting that flattening but not cell shape changes per se are linked to initiation of differentiation. Further, in contrast to control, where loricrin⁺ granules filled the entire cytoplasm, loricrin was detected predominantly at the cell periphery in the granular layer cells of Sptan1^KD embryos (Fig. 5 a and Fig. S5 b), suggesting impaired terminal differentiation and barrier formation.

TJ formation is a key feature of the SG2 layer and is necessary for epidermal barrier function (Kubo et al., 2009; Rübsam et al., 2017). To follow the formation of a functional TJ barrier and determine whether spectrin is required, we stratified primary keratinocytes and measured transepithelial electrical resistance (TEER). Although depleting spectrin did not obviously alter the recruitment of the TJ marker occludin to the apical cell–cell contacts (Fig. 5 c), TEER remained low over time (Fig. 5 d), thus indicating the formation of leaky TJs that could not be explained by lower cell numbers (Fig. S5 f). In the mouse epidermis, the tetrakaidecahedron-shaped SG2 cells that transition out of the SG2 layer not only have mature apical TJs but also form a new nascent TJ ring at basal tricellular contacts, which is smaller but aligned with and mirrors the shape of the upper, mature TJ ring (Fig. 5 e, Ctr) due to their geometric stacking arrangement (Yokouchi et al., 2016). This shape correlation and alignment were lost upon loss of αII-spectrin (Fig. 5, e and f), likely due to the alterations in cell shape, which may explain dysfunctional TJs.

To further explore changes in terminal differentiation, we assessed the activity of the cross-linker enzymes TGMs in the granular layer, necessary for terminal differentiation and proper formation of the outermost stratum corneum barrier (Eckert et al., 2005; Simpson et al., 2011). Incubation with a fluorescent substrate peptide to localize TGM1 activity (Sugimura et al., 2008) showed high enrichment of this peptide at the cell cortex of control granular layer cells, which was reduced in Sptan1^KD E17.5 embryos and Sptan1^epi−/− newborns (Fig. 5, g–i; and Fig. S5, g and h). Similarly, Ecad^epi−/− epidermis also showed decreased TGM activity (Fig. 5, l and m). TGM1 protein levels were not changed upon epidermal loss of either αII-spectrin or E-cadherin (Fig. S5, i–k). Thus, E-cadherin–dependent cortical organization of spectrin-actomyosin networks directs TGM activation necessary for terminal differentiation (Fig. S5, k and l). Toluidine blue exclusion assays (Hardman et al., 1998) showed no dye penetration in control E17.5 embryos except most ventrally, indicating the formation of an intact barrier (Hardman et al., 1998). In contrast, E17.5 embryos from Sptan1^KD and Sptan1^epi−/− showed dye penetration in the head and larger ventral parts (Fig. 5, j and k; and Fig. S5 l), and E18.5 Sptan1^KD embryos still showed staining of the head region (Fig. S5 m), indicating a delay in stratum corneum barrier formation. However, transepidermal water loss was not obviously increased in Sptan1^epi−/− newborns (Fig. 5 n), suggesting a transient defect in barrier formation during development. Taken together, spectrin regulates cell shape and promotes differentiation to enable functional TJ organization and cortical TGM activation downstream of E-cadherin in the upper suprabasal layers, necessary for proper barrier formation.

We previously showed that E-cadherin controls TJ formation by regulating the EGFR (Rübsam et al., 2017). Furthermore, EGFR activation promotes the opening of the calcium channel TRPV3 necessary for TGM1 activation and subsequent terminal epidermal differentiation (Cheng et al., 2010). As spectrin functions as a cortical platform that organizes membrane domains and membrane protein activities (Machnicka et al., 2014), we hypothesized that αII-spectrin promotes terminal differentiation through the EGFR–TRPV3 pathway. Immunofluorescence analysis showed EGFR staining in all epidermal layers in both control and Sptan1^KD E17.5 embryos (Fig. S6 a). Whereas phosphorylated EGFR (pEGFR) was highly enriched at the cortex of Ctr granular layer cells, many Sptan1^KD granular layer cells showed diffuse or no pEGFR staining, with now many cells showing ectopic pEGFR staining, especially in the spinous layer (Fig. 6, a–c).

To determine whether αII-spectrin controls TRPV3 localization, we infected E9 embryos with lentivirus encoding TRPV3-GFP (Xiao et al., 2008). In control E17.5 epidermis, TRPV3^GFP was recruited to the cortex in the granular layer (Fig. 6 d), where it co-localized with pEGFR (Fig. S6 b). In contrast, TRPV3^GFP remained more diffuse in the Sptan1^KD granular layer (Fig. 6 d and Fig. S6 c). Furthermore, colocalization of endogenous TRPV3 with E-cadherin at the cell membrane was significantly reduced in Sptan1^KD primary keratinocytes (Fig. 6, f and g). Thus, spectrin spatially controls activation of EGFR and localization of TRPV3 at the cortex.

Given that spectrin directs the layer-specific organization and mechanics of the cortical cytoskeleton, we asked whether αII-spectrin regulates the activation of the EGFR–TRPV3–TGM axis through actomyosin. To this end, we treated shCtr E17.5 embryos with latrunculin and Y27632 to decrease F-actin levels and myosin II motor activity, respectively. Similar to depletion of αII-spectrin, latrunculin or Y27632 treatment decreased pEGFR levels in the granular layer and increased spinous cortical pEGFR levels relative to DMSO-treated embryos (Fig. 6, h–j). Importantly, latrunculin or Y27632 also decreased the enrichment of TRPV3^GFP in vivo (Fig. 6, k and l) and the recruitment of endogenous TRPV3 to intercellular contacts in cultured keratinocytes (Fig. 6, m and n; and Fig. S6, d and e). Moreover, these treatments further decreased cortical enrichment of the TGM1 substrate peptide in E17.5 embryos (Fig. 6, o–q).

Together, the results demonstrate that the layer-specific spectrin-actomyosin organization and mechanics shaped by E-cadherin and distinctive spectrin and F-actin gradients control the spatial confinement and activation of EGFR and TRPV3 at the cortex to only activate TGM in the granular layer, thereby promoting epidermal barrier formation.

To further explore the relationship between EGFR activity, known to alter TJs and cortical mechanics (Muhamed et al., 2016; Rübsam et al., 2017), and spectrin-actomyosin cortical organization and TRPV3 localization, we inhibited EGFR using gefitinib in either E17.5 TRPV3^GFP-expressing embryos or in cultured keratinocytes. Gefitinib treatment significantly reduced cortical enrichment of both αII-spectrin and TRPV3 (Fig. 7, a–g). Interestingly, activation of EGFR by TGFα treatment also decreased αII-spectrin and TRPV3 localization at intercellular junctions (Fig. 7, d–g).

Finally, we asked whether EGFR activity regulates TGM1 activity in the epidermis. As in Sptan1^KD embryos (Fig. 5, g–i), cortical enrichment of the TGM1 substrate peptide was markedly decreased upon gefitinib treatment, indicating reduced cortical TGM1 activity (Fig. 7, h–j). Thus, in the granular layer, it spatially confines EGFR/TRPV3 activity at the cell membrane to promote terminal differentiation and establish a functional epidermis. Taken together, these and previous findings (Rübsam et al., 2017; Cheng et al., 2010) highlight a critical interdependency between the specific organization and tensile state of the spectrin-actomyosin cortex and signaling to coordinate the characteristic cell shape with the proper differentiation state in each epidermal layer.

In 1858, physician Rudolf Virchow established a correlation between cell shape and function, describing how abnormalities in cell shape can lead to disease development (Virchow, 1871). While the regulation of cell shape is a well-studied process dependent on adhesion and actomyosin, the mechanisms by which these processes also regulate differentiation and function require further elucidation. The layered structure of the epidermis provides a unique example of the correlation between cell position, shape, and function, i.e., differentiation and barrier function (Peskoller et al., 2022; Luxenburg and Zaidel-Bar, 2019). This study demonstrates that E-cadherin–dependent integration of αII-spectrin into submembranous actomyosin networks is crucial for the differential organization and tension state of the cortex to coordinate cell shape and differentiation state in each layer, essential to form a functional skin barrier.

Previously, we demonstrated that E-cadherin, a master regulator of cell–cell adhesion, is a key regulator of actomyosin-dependent tension to spatially control the localization and activation of EGFR, essential for TJ barrier function in the epidermis (Rübsam et al., 2017). However, how E-cadherin controls polarized localization of actomyosin tension and EGFR activity across layers remained unclear. Our new data now show that differential gradients of F-actin, myosin II, and spectrin drive a layer-specific organization of honeycomb-shaped spectrin-actomyosin networks necessary to dissipate myosin-dependent tension. Further, spectrin’s elastic properties likely allow the cortex to not only accommodate the change in cell shape when moving into a new layer but also to absorb forces experienced during this movement. Finally, this layer-specific organization of the spectrin-actomyosin cortical network regulates the localization and activity of EGFR/TRPV3 Ca²⁺ channel complexes necessary to activate TGM only in the granular layer to promote terminal differentiation and epidermal barrier function. Our data thus link cell adhesion with position-dependent cytoskeletal configuration to direct spatial control of cell shape, cell fate, and barrier function.

Previous studies on cell shape-dependent control of differentiation in the epidermis focused on early differentiation of basal progenitor cells and involved regulation of spindle orientation as well as compression-induced delamination (Le et al., 2016; Luxenburg et al., 2011; Miroshnikova et al., 2018). In contrast, other studies in which different actin-binding proteins, such as cofilin, WDR1, thymosin β4, and others, were deleted in the epidermis did not show any obvious changes in global differentiation markers (Luxenburg et al., 2015; Padmanabhan et al., 2020; Mahly et al., 2022). Similarly, K10 was not obviously altered upon loss of αII-spectrin, even if the basal marker K14 was upregulated. Our data now show that changes in cortical spectrin-actomyosin organization not only control basal cell shape and initial differentiation but also drive suprabasal cell shape transitions to regulate terminal differentiation, thus coordinating cell shape and fate across the tissue.

How spectrin regulates early differentiation when transitioning in the spinous layer is less clear. One potential candidate is YAP, based on previous studies that implicated spectrin and actomyosin tension in controlling YAP signaling in other epithelia (Deng et al., 2015; Fletcher et al., 2015). However, YAP nuclear localization was unchanged in Sptan1-depleted epidermis, suggesting that spectrin controls differentiation through a YAP-independent mechanism.

In the epidermis, mechanical regulation of EGFR contributes to barrier function by promoting terminal differentiation through TRPV3 and TGM1 (Cheng et al., 2010) and regulating TJs by reinforcing high-tension states in the SG2 layer (Rübsam et al., 2017), but whether these two functions are facets of the same mechanical regulation or regulated through distinct mechanisms remains unclear. Previous studies have shown that loss of myosin IIa/b in the epidermis impaired TJ barrier function but did not obviously affect stratum corneum barrier function (Sumigray et al., 2012), whereas loss of the Arp2/3 complex that enhances branched actin network assembly affects both barriers (Zhou et al., 2013). These observations further demonstrate the profound complexity of the actomyosin cytoskeleton and the importance of its fine-tuning for the execution of distinct biological processes.

Our previous and current data indicate that layer-specific organization of E-cadherin–directed spectrin-actomyosin lattices also controls the activity and localization of EGFR, with active EGFR being internalized in lower layers, whereas in the granular layer, active EGFR is retained at the membrane. Either loss of spectrin or reducing actomyosin tension reverses this localization of active EGFR, with membrane localization now being confined to lower layers. The organization of the spectrin-actomyosin cortex can inhibit diffusion of plasma membrane receptors to induce molecular crowding, known as fencing (Fujiwara et al., 2016), as well as organize membrane trafficking events such as endocytosis (Ghisleni et al., 2020). We thus propose a model in which the spectrin-actomyosin organization in the granular layers limits the mobility of EGFR and TRPV3, while the high-tension state of this network also inhibits internalization of activated EGFR, thus bringing these proteins in close proximity to activate TRPV3 and, subsequently, TGM1 (Fig. 7 k). In agreement, both ligand-binding and actin-dependent membrane organization regulate EGFR distribution at the plasma membrane (Sankaran et al., 2021; Bag et al., 2015). A potential alternative model involves a mechano-sensitive unfolding of spectrin (Leterrier and Pullarkat, 2022; Renn et al., 2019) or other, unknown players within the granular layer cortex that then interact with and cage the EGFR/TRPV3 to the cortex. Thus far, however, there is no evidence for tension-dependent binding partners that may explain EGFR or TRPV3 retention.

In conclusion, our data provide evidence for a model in which E-cadherin shapes the distinct organization and mechanics of the spectrin-actomyosin cortex in each layer to determine cell shape and differentiation status and control localization of pEGFR–TRPV3 complexes at the plasma membrane, resulting in TGM-dependent cross-linking only in the granular layer, thereby fixing cell shapes and promoting stratum corneum barrier function.

To generate epidermal E-cadherin knockout (Ecad^epi−/−) BL/6N mice, female Cdh1^flox/flox mice were crossed to male Ecad^fl/wt;K14-Cre mice. Transgenic Ecad^flox/flox mice and K14-Cre mice and their genotyping were described previously (Hafner et al., 2004; Boussadia et al., 2002). Hsd:ICR (CD1) mice (Envigo) were used for all in vivo knockdown experiments. To generate epidermal αII-Spectrin knockout (Sptan1^epi−/−) mice, Sptan1^flox mice were ordered from The Jackson Laboratory (Strain ID- B6;129S-Sptan1^tm1.1Mnr/J, Strain #033392). Female Sptan1^flox/flox mice were crossed to male Sptan1^fl/wt;K14-Cre mice. Sptan1^flox mice were described previously (Huang et al., 2017). To generate epidermal confetti BL/6N mice, female R26R-Confetti^tg/tg mice were crossed with male K14-Cre mice for stochastic activation and recombination of the brainbow2.1 cassette in the epidermis, resulting in four possible outcomes: Expression of nuclear GFP, membrane-associated CFP, cytoplasmic YFP, or cytoplasmic RFP.

Epidermal confetti newborn mice were sacrificed and fixed in 4% PFA overnight at 4°C. Backskin was then taken and mounted in Mowiol. As a basis for rendering the cytoplasmic YFP signal of the brainbow2.1, it was used due to homogenous intracellular expression across epidermal layers and a good signal/noise ratio. Confocal stacks were acquired using a Leica TCS SP8 (PlanApo 63×, 1.4 NA CS2) in Leica lightning mode, pinhole 0.97 AU, voxel size 0.046 × 0.046 × 0.259, followed by the built-in lightning deconvolution. Imaris was then used for surface rendering using a surface detail of 0.2 µm and smoothing. Cell volume, surface area, and sphericity were extracted from the rendered cells in Imaris. Cell position was determined from XZ projections of the same stack using the measure tool in Fiji. Cell position was determined relative to the stratum corneum by measuring the distance of each cell to the upper side of the stratum corneum. The stratum corneum border was readily detectable due to preservation of the fluorescent signal and high cellularity.

Primary keratinocytes were isolated and cultured as described before (Rübsam et al., 2017): newborn mice were decapitated and incubated in 50% betaisodona/PBS for 30 min at 4°C, 1 min PBS, 1 min 70% EtOH, 1 min PBS, and 1 min antibiotic/antimycotic solution. Skin was incubated in 2 ml dispase (5 mg ml⁻¹ in culture medium) solution. After overnight incubation at 4°C, the skin was transferred onto 500 μl FAD medium on a 6-cm dish, and the epidermis was separated from the dermis as a sheet. Epidermis was transferred dermal side down onto 500 μl of TrypLE (Thermo Fisher Scientific) and incubated for 20 min at RT. Keratinocytes were washed out of the epidermal sheet using 3 ml of 10% FCS/PBS. After centrifugation, keratinocytes were resuspended in FAD medium and seeded onto collagen type-1– (0.04 mg ml⁻¹) (L7213; Biochrom) coated cell culture plates. Primary murine keratinocytes were kept at 32°C and 5% CO₂.

Primary keratinocytes were cultured in DMEM/HAM’s F12 (FAD) medium with low Ca²⁺ (50 μM) (Biochrom) supplemented with 10% FCS (chelated), penicillin (100 U ml⁻¹), streptomycin (100 μg ml^-1, A2212; Biochrom), adenine (1.8 × 10⁻⁴ M, A3159; SIGMA), L-glutamine (2 mM, K0282; Biochrom), hydrocortisone (0.5 μg ml⁻¹, H4001; Sigma-Aldrich), EGF (10 ng ml⁻¹, E9644; Sigma-Aldrich), cholera enterotoxin (0.1 nM, C-8052; Sigma-Aldrich), insulin (5 μg ml⁻¹, I1882; Sigma-Aldrich), and ascorbic acid (0.05 mg ml⁻¹, A4034; Sigma-Aldrich).

Overexpression: Keratinocytes were transfected at 100% confluency with ViromerRED. 1.5 µg DNA was diluted in 100 μl Buffer E, added to 1.25 μl ViromerRED, and incubated for 15 min at RT (Rübsam et al., 2017). 33 μl transfection mix was used per well with 0.5 ml FAD medium (24-well plate). Knockdown: Keratinocytes were transfected at 100% confluency with ViromerBLUE (lipocalyx; Halle Germany). The transfection mix was prepared according to the manufacturer’s protocol. 100 μl transfection mix was used per well with 1 ml FAD medium (12-well plate, 5 nM siRNA f.c.). siRNA (siPOOLs) against αII-Spectrin were obtained from siPOOLs from siTOOLs.

Lentiviral plasmids were generated by cloning oligonucleotides into pLKO.1-TRC (gift from David Root, Broad Institute, Cambridge, MA, USA; #10878; Addgene plasmid), lentivirus-GFP, or lentivirus-RFP (gift from Elaine Fuchs, Rockefeller University, New York, NY, USA; #25999; Addgene plasmid) by digestion with EcoRI and AgeI, as described in the Genetic Perturbation Platform (GPP) website (http://portals.broadinstitute.org/gpp/public/resources/protocols). See also (Beronja et al., 2010; Soffer et al., 2022) for generation of lentiviruses. shRNA sequences were obtained from GPP (https://portals.broadinstitute.org/gpp/public/): Sptan1(0595) construct #TRCN0000090595, target sequence 5′-GCC​ACC​GAT​GAA​GCT​TAT​AAA-3′. Sptan1(9753) construct #TRCN0000089753, target sequence 5′-CCA​GTC​ACA​ATC​ACC​AAT​GTT-3′. Ank3(6780) construct #TRCN0000236780, target sequence 5′-TGC​CGT​GGT​TTC​CCG​GAT​TAA-3′. The TRPV3-GFP construct was a gift from Michael X. Zhu (UTHealth Houston, Houston, Texas, USA).

Female mice at E8.5 were anesthetized with isoflurane, injected with the painkiller, Rheumocam Veterinary 5 mg/ml according to the manufacturer’s instructions (Chanelle Pharma), and each embryo (up to six per litter) was injected with 0.4–1 μl of ∼2 × 10⁹ colony-forming units (CFUs) of the appropriate lentivirus. See also Beronja et al. (2010) for further details. Controls were both uninfected littermates of shSptan1-0595/9753;H2B-GFP/shSptan1-0595;H2B-RFP/shAnk3-6780;H2B-GFP lentivirus-injected embryos and shScr;H2B-GFP lentivirus-injected embryos. For shSptan1-0595/9753; TRPV3-GFP lentivirus-injected embryos, shScr; TRPV3-GFP were controls. In utero lentivirus injection infects ∼60–70% of the dorsal skin epidermis (Beronja et al., 2010).

Immunofluorescence stainings of keratinocytes were performed as described before (Sahu et al., 2020): cells were seeded on collagen-coated glass coverslips in a 24-well plate and switched to high Ca²⁺ medium as indicated in the results section. Cells were fixed using 4% PFA for 10 min at RT, washed three times for 5 min using PBS, permeabilized using 0.5% Triton X-100/PBS, and blocked using 5% NGS/1% BSA/PBS for 1 h at RT. Primary antibodies were diluted as indicated in the antibody section in Background Reducing Antibody Diluent Solution (ADS) (DAKO). Coverslips were placed growth surface down onto a 50-µl drop of staining solution on parafilm in a humidified chamber and incubated overnight at 4°C. Coverslips were washed again with PBS three times for 10 min. Secondary antibodies and DAPI (Sigma-Aldrich) were diluted 1:500 in ADS, and coverslips were incubated for 1 h at RT. Secondary antibodies were washed off via three wash steps using PBS for 10 min. Coverslips were mounted using Mowiol (Calbiochem).

Embryos or newborn mice were embedded in OCT (scigen), frozen, sectioned at 10 μM, and fixed in 4% formaldehyde for 10 min. Sections were then blocked with 0.1% Triton X-100, 1% BSA, 5% normal donkey serum in PBS, or in MOM Basic kit reagent (Vector Laboratories). Sections were incubated with primary antibodies overnight at 4°C and with secondary antibodies for 1 h at RT.

Ankyrin3 (IF 1:500, WB 1:1,000, #27766-1-AP; Proteintech), E-cadherin (IF 1:200, #610182; BD Transduction Laboratories, clone number 36 or IF 1:500, #3195; Cell Signaling), EGFR (IF 1:500, #ab52894; Abcam), pEGFR (Y1068) (IF 1:500, #ab40815; Abcam), GAPDH (WB 1:10,000, #AM4300; Ambion or WB 1:1,000, #5174; Cell Signaling), GFP (1:3,000, #ab13970; Abcam), Keratin10 (IF 1:1,000, #PRB-159P; BioLegend), Keratin14 (IF 1:2,000, #PRB 155P; Covance), Loricrin (1:1,000, #Poly19051; BioLegend), Myosin heavy chain IIa (IF 1:500, PRB-440; BioLegend), phospho-Myosin Light Chain 2 (Thr18/Ser19) (IF 1:100, #3674; Cell Signaling), occludin (IF 1:400, #33-1,500; Invitrogen), TRPV3 (IF 1:1,000, #b94582; Abcam), αII-Spectrin (IF 1:500, WB 1:1,000, ab11755; Abcam), TGM1 (IF 1:500, #12912-3-AP; Proteintech), α-catenin (IF 1:2,000, #C2081; Sigma-Aldrich), β-catenin (IF 1:1,000, #ab32572; Abcam), and YAP (IF 1:500, #14074; Cell Signaling). Phalloidin was used to stain F-actin (IF 1:500, #P1951; Sigma-Aldrich, TRITC conjugated). Secondary antibodies were species-specific antibodies conjugated with either Alexa Fluor 488, 594, or 647, used at a dilution of 1:500 for immunofluorescence (Molecular Probes, Life Technologies), or with horseradish peroxidase antibodies used at 1:5,000 for immunoblotting (Bio-Rad Laboratories).

Inhibitors used in this study, if not indicated otherwise: Blebbistatin myosin inhibitor, 5 μM (#B0560; Sigma-Aldrich); Latrunculin B, 0.1 μM (L5288; Sigma-Aldrich).

Epidermal whole mounts were prepared as has been described previously (Rübsam et al., 2017). Backskin was prepared from newborn mice, and subcutaneous fat was removed with curved tweezers. The epidermis was mechanically and carefully peeled off from the dermis with ultrafine curved tweezers, thereby separating the basal from the suprabasal layers. During the whole procedure, the basal side of the sheet was kept floating on PBS2+ (PBS supplemented with 0.5 mM MgCl2 and 0.1 mM CaCl2). Subsequently, the epidermal sheet was fixed, floating on 4% PFA on ice for 10 min, washed in PBS for 5 min, and permeabilized with 0.5% Triton X-100/PBS for 1 h at RT. The permeabilized sheet was washed for 5 min in PBS and blocked with 10% FCS/PBS for 30 min/RT. For staining, epidermal sheets were cut into ∼5 × 5-mm pieces. The SC of the epidermis is prone to unspecific binding of ABs. Additionally, it cannot be permeabilized to allow AB permeation. Thus, all following steps were performed by incubating the sheet from the basal side, leaving the SC side dry. Primary ABs were diluted either in AB diluent solution (S3022; Dako) and incubated overnight at 4°C. Secondary Abs, including DAPI and phalloidin, were incubated for 2 h/RT. After each AB incubation, the sheet was rinsed three times with PBS and washed three times for 10 min. Finally, the stained sheet was mounted in 50 μl Mowiol.

Dorsal back skin of newborn mice was excised (∼10 mm × 10 mm), and fat tissue was removed with fine curved forceps. The epidermis was mechanically and carefully peeled off from the dermis with ultrafine curved tweezers, thereby separating the basal from the suprabasal layers. Subsequently, the suprabasal epidermis was floated on a 400-μl droplet of 44 μg/ml rETA (Staphylococcus aureus Exfoliative Toxin A)/1 mM CaCl₂ and incubated in the humidifying chamber at 37°C for 35 min. ETA diffuses intercellularly up to the tight junctions in the SG2 and cuts desmogleins-1, thereby loosening intercellular connections below the SG2 (Matsui et al., 2021). Layers below the SG2 layer (SS-SG3) were then peeled from the SG-1/2-corneum layers. Both parts were then floated onto 500-μl droplets of 0.05% trypsin/0.48 mM EDTA and incubated in the humidifying chamber at 37°C for 30 min. After the trypsinization, SS-SG3 cells were dissociated, and SG-1/2 cells were washed off from the stratum corneum by manual pipetting and transferred to a 1.5-ml Eppendorf tube. To neutralize, 1 ml of PBS++ was added to the trypsin solution containing SG and SS cells and centrifuged at 3,000 rpm for 5 min at RT. For treatment, cells were resuspended in DMEM (Gibco).

After isolation, cells were treated with 1 ml DMEM with or without latrunculin B, 0.1 μM, shaking at 300 rpm for 1 h at 37°C. After the treatment, cells were washed with 1 ml of PBS++, centrifuged at 3,000 rpm for 5 min at RT, and fixed with 4% PFA for 10 min at RT. Cells were permeabilized using 0.5% Triton X-100/PBS for 10 min and washed with 500 μl of PBS++. Consecutively, cells were treated with phalloidin-TRITC and DAPI for 30 min at RT. After the incubation, cells were centrifuged and mounted on coverslips with 30 μl of Mowiol. Cells were then imaged, and the circumference of the cells was delineated using the freehand tool in Fiji. The area and circularity were extracted using the measure function in Fiji.

For in vivo latrunculin and Y27632 treatments, wild-type embryos or shScr; TRPV3-GFP–transduced embryos were collected on E17.5 with 2.5 μM latrunculin (Sigma-Aldrich) or with 40 μM Y27632 (Sigma-Aldrich) or DMSO in serum-free DMEM (Biological Industries) at 37°C for 1 h before embedding in OCT.

Cryosections were washed with PBS++ until the cryo tissue embedding compound Tissue-Tek O.C.T Compound (Sakura Finetek 4583) was removed. Then, sections were circled with a hydrophobic barrier marker pen (ReadyProbes - Thermo Fisher Scientific) and blocked with 10% NGS/PBS for 1 h at RT. Subsequently, sections were treated with and without TGM peptide substrate (pepK5: 5/6-FITC-Doa-YEQHKLPSSWPF-NH2/OH and K5pepQN: 5/6-FITC-Doa-YENHKLPSSWPF-NH2/OH, Intavis Peptide Services) for 30 min at RT in a humidified chamber and washed three times with PBS to remove the unspecific binding. Next, sections were fixed with 4% PFA for 5 min at RT and washed with PBS three times for 5 min. Sections were mounted using Mowiol.

E17.5 embryos were collected and immersed in an ice-cold methanol gradient in water, taking 2 min per step (1–25%, 2–50%, 3–75%, and 4–100% methanol), and then rehydrated using the reverse procedure. Embryos were immersed in 0.2% toluidine blue solution. Embryos were washed in PBS before image capture. See also Hardman et al. (1998) for further details.

Confocal images were obtained with a Leica TCS SP8, equipped with gateable hybrid detectors using LAS X software. Objectives used with this microscope: Leica PlanApo 63×, 1.4 NA CS2. Images to be used for deconvolution were obtained at optimal resolution according to Nyquist and deconvolved using Huygens Deconvolution. Alternatively, images were acquired with a Nikon C2+ laser-scanning confocal microscope using NIS Elements C software. Objectives: 60×, 1.4 NA oil or a 20×, 0.75 NA air objective (Nikon). Imaging was performed at RT, and samples were mounted in Mowiol.

Keratinocytes were seeded on glass-bottom dishes (# 81158; Ibidi) in FAD low Ca²⁺ medium. At confluency, the medium was switched to a high Ca²⁺ medium for 48 h. 2 h prior to imaging, the medium was changed to a high Ca²⁺ medium containing 1 μM SiR-actin (#CY-SC001; Spirochrome). For laser ablation, a spinning Disk microscope (UltaView VoX, Perkin Elmer) equipped with a 355 nM pulsed YAG laser (Rapp Opto Electronic) and a Hamamatsu EMCCD C9100-50 camera was used with Volocity software. Experiments were performed at 37°C and 5% CO₂ using a water immersion objective 60×, 1.2 NA (Nikon). The apical area of apical stratified cells was ablated by drawing a line of consistent length with 30–50% laser power transmission and 10 iterations using UGA-40 Software (Biomedical Instruments).

A total of 5 × 10⁵ keratinocytes were seeded on transwell filters (Corning [#3460], 0.4-μm pore size). Cells were allowed to settle and then switched to 1.8 mM high Ca²⁺ medium. Formation of TER was measured over time using an automated cell monitoring system (cellZscope, nanoAnalytics).

RNA was extracted from samples using a Direct-zol RNA extraction kit (R2060; Zymo Research), and equal amounts of RNA were reverse transcribed using the ProtoScript First Strand cDNA Synthesis Kit (New England Biolabs). Semiquantitative PCR was conducted using a StepOnePlus System (Thermo Fisher Scientific). Data are presented as mRNA levels of the gene of interest normalized to peptidylprolyl isomerase B (Ppib) mRNA levels. Primers used in this study: Sptan1 forward 5′-TCG​ACA​AGG​ACA​AGT​CTG​GC-3′; Sptan1 reverse 5′-AAC​AGG​GCA​AGC​AGT​GTA​GG-3′; Ank3 forward 5′-CTG​ACG​TTC​ACG​AGG​GAG​TT-3′; Ank3 reverse 5′-TAT​CTA​ACG​TGT​CCG​CTG​CC-3′; PPIB forward 5′-GTG​AGC​GCT​TCC​CAG​ATG​AGA-3′; PPIB reverse 5′-TGC​CGG​AGT​CGA​CAA​TGA​TG-3′.

Intensities from layers as indicated in the figures and legends were measured using the freehand selection tool (Image J) with a width of five pixels. All intensity measurements were determined by corrected total cell F=florescence = integrated density- (area X mean of fluorescence of background readings). Intensity levels in each layer were normalized according to controls.

Embryos that were injected with lentiviruses encoding shScr;H2B-GFP, shsptan1;H2B-GFP, and shAnk3-6780;H2B-GFP on E8.5 and littermates were harvested at E17.5 frozen in OCT, sectioned (10 μm), fixed, and stained for E-cadherin (3195; Cell Signaling Technology, 1:500). For the analysis of Ctr, Cdh1^epi−/−, and Sptan1^epi−/− newborn epidermis, cryosections were PFA fixed and stained with a combination of Dsg1/2 antibody (61002; Progen, 1:200) and Dsg3 antibody (D218-3; MBL, 1:2,000). Samples were imaged using confocal microscopy. The cell borders were delineated by the Freehand tool (Fiji) and cell area and perimeter using the “measure” function of Fiji. The cell shape index was calculated by dividing the perimeter by the square root of the area (Sahu et al., 2020).

The number of GFP+ and YAP+ cells was counted manually. The percentage of YAP+ cells was calculated as (number of YAP+GFP+ double-positive cells/total number of GFP+ cells) × 100 for the basal and suprabasal layers.

The mean gray value of the granular layer cells’ cortical and cytoplasmic intensity levels were measured with the “wand tool” (ImageJ) with a width of five pixels. The ratio between the two means was calculated and referred to as the fold change.

The number of independent experiments and biological replicates performed for all experiments, P values, and the statistical tests that were used are indicated in the figure legends.

Fig. S1 provides extended information on 3D cell rendering using epidermal confetti mice, αII-spectrin expression in epidermal development, αII-spectrin knockdown and knockout efficiency, and epidermal cell shapes upon epidermis specific knockout of αII-spectrin, related to Fig. 1. Fig. S2 shows knockdown efficiency of Ankyrin-3, Ankyrin-3–dependent organization of F-actin and E-cadherin in the developing epidermis, the interdependency of αII-spectrin and E-cadherin in junctional recruitment in vitro, and extended information on latrunculin B treatment of isolated SG cells, related to Fig. 2. Fig. S3 shows actomyosin-dependent levels of E-cadherin in the developing epidermis, extended data on the spatial organization and expression levels of αII-spectrin in stratified keratinocytes in vitro, and the interdependency of cortical F-actin and αII-spectrin in vitro, related to Fig. 3. Fig. S4 provides extended information on the cortical interdependency of αII-spectrin and myosin IIa in vitro and laser ablation experiments showing force generation and dissipation of cortical F-actin networks in keratinocytes in vitro, related to Fig. 4. Fig. S5 shows extended information on αII-spectrin–dependent regulation of epidermal differentiation, suprabasal proliferation and normal spindle orientation upon epidermal knockdown of αII-spectrin, reduced TGM1 activity in epidermal αII-spectrin knockouts, and impaired outside-in barrier function of αII-spectrin knockdown embryos, related to Fig. 5. Fig. S6 shows normal distribution of total EGFR in the epidermis upon αII-spectrin knockdown, extended information on αII-spectrin–dependent cortical localization of TRPV3 in the epidermis, and normal junctional E-cadherin localization upon low level interference with actomyosin in vitro, related to Fig. 6.

All animal protocols in this study have been approved by either the animal experiment committee of LANUV, North Rhine-Westphalia, Germany, or University Animal Care and Use Committee, confirmation number TAU-MD-IL-2206-162-4. All methods were carried out in accordance with relevant guidelines and regulations. The animal protocols and the reporting in this manuscript follow the recommendations in the ARRIVE guidelines.

The authors declare that the data supporting the findings of this study are available within the paper and its supplementary information files. Additional data are available from the corresponding author upon reasonable request.

We would like to thank the CECAD imaging facility under the supervision of Astrid Schauss.

This work is supported by the Deutsche Forschungsgemeinschaft, German Research Foundation: Germany’s Excellence Strategy – EXC 2030/CECAD – 390661388 (to C.M. Niessen); CECAD project: CECAD Career-Promoting Grant for Senior Postdoctoral Scientists (to M. Rübsam); DFG-Schwerpunktprogramm (SPP) 1782 NI 1234/6-2, Project-ID 388932620 (to C.M. Niessen); DFG-Forschungsgruppe (FOR) 2743 NI 1234/7-1, Project-ID 678823 (to C.M. Niessen); ANR/DFG NI 1234/9, Project-ID 505673300 (to C.M. Niessen); European Union NETSKINMODELS, COST Action no. CA21108 (to C.M. Niessen); Israel Science Foundation (ISF) grant number 145/23 (to C. Luxenburg). This work was carried out in partial fulfilment of the requirements for a Ph.D. degree for AS. from the Faculty of Medical & Health Sciences, Tel-Aviv University Open Access funding provided by University of Cologne Medical Department; Department of Biology.

Author contributions: Arad Soffer: conceptualization, data curation, formal analysis, investigation, methodology, visualization, and writing—original draft, review, and editing. Aishwarya Bhosale: conceptualization, data curation, formal analysis, investigation, methodology, and project administration. Roohallah Ghodrat: formal analysis and investigation. Marc Peskoller: data curation, investigation, and project administration. Takeshi Matsui: methodology. Carien M. Niessen: conceptualization, funding acquisition, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing. Chen Luxenburg: conceptualization, funding acquisition, project administration, supervision, and writing—original draft, review, and editing. Matthias Rübsam: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing.