S100A11 promotes focal adhesion disassembly via myosin II-driven contractility and Piezo1-mediated Ca²⁺ entry

S100 proteins form a family of small Ca²⁺-binding proteins with multifaceted roles in diverse intra- and extra-cellular processes, including Ca²⁺ homeostasis, signal transduction, proliferation, differentiation and migration (Donato et al., 2013; Gonzalez et al., 2020; Moore, 1965), as well as inflammation (Sreejit et al., 2020) and cancer (Bresnick et al., 2015). Central to the role of S100 proteins as signaling molecules is their capacity to become activated by Ca²⁺ binding (Schäfer and Heizmann, 1996). S100 proteins typically form symmetric homodimers (Yap et al., 1999) in which each subunit supplies two central EF-hand motifs connected by a short hinge region. Binding of Ca²⁺ via the EF hands induces a conformational change that exposes a normally inaccessible hydrophobic cleft for interaction with other target proteins (Otterbein et al., 2002; Santamaria-Kisiel et al., 2006). Ca²⁺ greatly increases the affinity of S100 proteins to their targets, and consequently S100 proteins have been described as intracellular Ca²⁺ sensors, able to detect and translate fluctuations in intracellular Ca²⁺ levels into specific cellular responses (Zimmer and Weber, 2010). In addition to the common EF hand motif arrangement, individual S100 members contain short C-terminal tails that vary in length and sequence and give each member additional specific functional roles.

S100A11, also called S100C or calgizzarin (Ohta et al., 1991; Todoroki et al., 1991), is widely expressed and, like other S100 family members, plays a role in numerous cellular functions under both physiological and pathological conditions (He et al., 2009). Among other functions, it is associated with cell cycle and cell growth regulation, and apoptosis (Sakaguchi and Huh, 2011; Sakaguchi et al., 2003; Xia et al., 2018). S100A11 also has a well-established role in a variety of diseases, including cancer metastasis, inflammation and neurological diseases (Zhang et al., 2021). S100A11 binds filamentous actin (F-actin) in a Ca²⁺-dependent manner (Sakaguchi et al., 2000; Zhao et al., 2000) and participates in different cellular processes involving actin cytoskeleton remodeling, including membrane wound healing (Jaiswal et al., 2014), pseudopodal protrusion formation (Shankar et al., 2010) and cell migration (Meng et al., 2019; Wang et al., 2013). Furthermore, S100A11 localizes along actin stress fibers (SFs), filopodia and lamellipodia in normal and immortalized human fibroblasts, implicating it in actin organization and regulation of dynamics (Sakaguchi et al., 2000). S100A11 also inhibits smooth muscle myosin ATPase activity in a Ca²⁺-dependent manner (Zhao et al., 2000), pointing towards a potential role in regulating actomyosin contractility. Furthermore, Ca²⁺ entry into migrating cancer cells as a result of plasma membrane injury recruits S100A11 together with annexin A2, another Ca²⁺-binding protein, to the site of injury, where the S100A11–annexin A2 complexes help resealing the membrane by stimulating cortical actin polymerization (Jaiswal et al., 2014). Similarly, S100A11 and annexin A2 also cooperate during actin-dependent wound healing in endothelial cells (Ashraf and Gerke, 2021). Complexes between S100A11 and annexin 6 on the other hand might stabilize connections between the plasma membrane and the cytoskeleton (Chang et al., 2007). S100A11 also drives the formation of actin-dependent pseudopodal protrusions in migrating tumor cells (Shankar et al., 2010).

Despite the ample evidence demonstrating S100A11–F-actin interaction and S100A11-dependent actin cytoskeleton regulation, previous research has focused more strongly on other S100 members, particularly S100A4, S100A6 and S100A10, in relation to actin cytoskeleton dynamics (Gómez-Contreras et al., 2017; Jurewicz et al., 2020; Sayeed et al., 2013). As a result, little is known about the molecular mechanisms by which S100A11 modulates actin remodeling. Here, we have investigated the intracellular localization of S100A11 in HeLa and U2OS osteosarcoma cells and detected a novel localization to stable peripheral focal adhesions (FAs). Moreover, S100A11 levels at FAs increase sharply, yet transiently, at the onset of FA disassembly, implicating S100A11 in FA turnover regulation. Recruitment of S100A11 to disassembling FAs requires non-muscle myosin II (NMII)-mediated contractility, opening of stretch-activatable Piezo1 Ca²⁺ channels at stressed FAs and activation by intracellular Ca²⁺. Furthermore, we demonstrate aberrant FA morphology and impaired FA disassembly in S100A11-knockout (KO) cells. Together, our results thus reveal a novel mechano-regulated role for S100A11 during NMII-, Piezo1- and Ca²⁺-dependent FA disassembly.

S100A11 is an actin-binding protein (Zhao et al., 2000) previously reported to localize to different elements of the actin cytoskeleton, including filopodia, lamellipodia and SFs in normal and transformed fibroblast (Sakaguchi et al., 2000). Although these localization patterns suggest a general role for S100A11 in actin cytoskeleton regulation, its precise function in these processes remains poorly understood. To further investigate potential roles of S100A11 in cytoskeleton regulation, we analyzed the intracellular localization of endogenous S100A11 in HeLa cells by immunofluorescence (IF). Co-staining for F-actin with phalloidin confirmed the previously reported localization of S100A11 along actin SFs (Fig. 1A). However, S100A11 was also strongly enriched at SF termini within the cell periphery. Co-staining for vinculin confirmed the prominent localization of S100A11 to FAs (Fig. 1B). This observation revealed a previously unknown additional localization of S100A11 to FAs. S100A11 has also been implicated in microtubule binding (Broome and Eckert, 2004), but co-staining for S100A11 and β-tubulin showed that S100A11 does not localize to microtubules in these cells (Fig. S1A). Nevertheless, the non-overlapping localization patterns of S100A11 and microtubules ruled out potential fluorescence channel crosstalk in our setup. In the IF experiments, we used an enhanced staining protocol (Flores-Maldonado et al., 2020), which increased the contrast of the obtained S100A11 images compared to that in standard IF protocols. S100A11 localization to FAs could also be independently confirmed by total internal reflection fluorescence (TIRF) microscopy of transfected HeLa cells expressing GFP–S100A11 (Fig. S1B). Given that the TIRF imaging z-range is restricted to ∼200 nm above the cell culture substrate, GFP–S100A11 was not detected in SFs, which typically extend higher into the cytoplasmic space, in these experiments. However, both SF and FA localization could be detected by epi-fluorescence imaging of GFP–S100A11-expressing cells, even against strong cytoplasmic GFP–S100A11 levels (Fig. S1B).

We also assessed S100A11 localization in U2OS cells, an osteosarcoma cell line featuring prominent actin SFs (Burnette et al., 2014; Hotulainen and Lappalainen, 2006). Again, S100A11 localized to SFs as well FAs (Fig. 1C) in most cells (89±13% of cells, n=192, mean±s.d., Fig. S2A–C), although it localized primarily to peripheral adhesions at the end of dorsal SFs. FA localization of S100A11 was further verified by live-cell TIRF microscopy imaging of U2OS cells co-expressing GFP–S100A11 and vinculin–mCherry (Fig. 1D). Again, the TIRF images showed good overall colocalization of S100A11 and vinculin in the majority of FAs. Compared to HeLa cells, S100A11 and vinculin signals overlapped less completely. Instead, S100A11 was often enriched at the distal end of FAs or its localization extended to areas in the FA vicinity. Compared to IF images, live-cell TIRF images also showed higher levels of diffuse, non-FA-associated S100A11 across the entire basal plasma membrane, indicating that a fraction of cytoplasmic S100A11 continuously associates with the membrane and/or the cortical actin cytoskeleton network. Endogenous S100A11 or transiently expressed GFP–S100A11 also localized to FA and SFs in PtK2, NIH3T3 and cancer-associated fibroblast (CAF) cells (Fig. S3).

In vitro, S100A11 binds F-actin in a Ca²⁺-dependent manner (Sakaguchi et al., 2000; Zhao et al., 2000). To investigate whether Ca²⁺-binding was also required for association with native actin SFs and FAs, U2OS cells were ‘de-roofed’ using a previously established microsonication protocol (Franz and Müller, 2005). The exposed SFs (Fig. S4A) were then incubated with recombinant GFP–S100A11 protein (Fig. S4B). In presence of 1 mM Ca²⁺, GFP–S100A11 localized along SFs, while a GFP control alone only showed weak non-specific association. Importantly, when free Ca²⁺ was removed from the incubation buffer by adding 5 mM EGTA, GFP–S100A11 also displayed only weak diffuse binding, confirming Ca²⁺ dependency for SF localization (Fig. S4B). GFP S100A11–protein also displayed specific binding to exposed SFs in HeLa cells (Fig. S4C).

Strong localization to SFs and FAs in different cell types pointed to a potential role of S100A11 in regulating these cytoskeletal structures. To obtain insight into dynamic aspects of such possible S100A11-dependent mechanisms, HeLa cells co-expressing GFP–S100A11 and Lifeact–mCherry were imaged by TIRF live-cell microscopy. Only the FA-associated F-actin signal is recorded in this imaging mode, which was largely identical to experiments using bona fide FA markers, such as vinculin or paxillin. Timelapse movies recorded at intervals of 10 s over a total duration of 20 min showed primarily diffuse S100A11 localization to the basal cell membrane and specific FA localization at steady intensity levels (Movie 1). However, we also consistently observed defined areas of vastly yet transiently enhanced S100A11 intensity (Fig. 2A; Movie 1). These striking S100A11 flashes typically occurred at or close to FAs at the cell periphery and typically lasted between 10 and 90 s for an average time of 48±31 s (mean±s.d., n=23). In 10 TIRF movies of individual cells recorded over 60 min each, we detected a total of 23 discrete S100A11 flashes, or 2.3±1.8 (mean±s.d.) flashes per cell and hour. From the recorded timelapse movies (see, for example, Movie 1) we noticed frequent loss of FAs in the areas experiencing an S100A11 flash (20 out of 23 cases, ∼87% of events) within the following 8 min (maximal observed post-flash FA disassembly time). In contrast, peripheral FAs experiencing no S100A11 flash only disassembled with a probability of ∼4% (1 out of 27 analyzed randomly selected peripheral FAs) within the same 8 min time window, pointing towards a functional correlation between S100A11 recruitment and FA disassembly. In the majority of cases (18 out of 23, ∼78% of events), S100A11 flashes were highly localized in small areas encompassing a single FA (Fig. 2A,C), whereas in the remaining cases (5 out of 23, ∼22% of events) areas of transient S100A11 recruitment were broader and covered from two to five adjacent FAs (Fig. 2B).

To investigate the relationship between transient S100A11 recruitment and FA disassembly further, kymographs were generated along trajectories of peripheral FAs targeted by S100A11 flashes (Fig. 2D,E; Fig. S5). The kymographs revealed that FA disassembled within 239±125 s (mean±s.d., n=20) after an S100A11 flash. Pre-flash FAs displayed a slow ‘sliding’ behavior typical for FAs in stationary cells (Smilenov et al., 1999) at a speed of 0.06±0.01 µm/min (mean±s.d., n=5). In contrast, after S100A11 flashing the FA translation speed increased to 0.45±0.23 µm/min, or ∼7-fold (Fig. 2E). A previously identified force-dependent FA disassembly mechanism involves increased actomyosin-mediated pulling forces and FA retraction speeds (Crowley and Horwitz, 1995; Webb et al., 2004). Likewise, the elevated translocation speeds of FAs after S100A11 flashing suggested that these FAs might disassemble under increased intracellular pulling forces.

Although S100A11 flashes typically occurred over areas covering a single or several FAs, in some cases S100A11 flashes were spatially restricted to a small region at the distal FA end (Fig. 3A,B; Movie 2). These highly localized S100A11 flashes also appeared to target specific FAs for disassembly, as a neighboring FA only several micrometers away displaying no flashing remained stable (Fig. 3B). Tracking the fluorescence intensities of GFP–S100A11 and vinculin–mCherry at the disassembling FA over time initially showed stable vinculin but steadily increasing S100A11 levels, followed by a sudden drop in both S100A11 and vinculin signals. These results suggest that prior to FA disassembly, S100A11 could progressively accumulate at FAs until reaching a critical level, after which it was quickly released, and rapid FA disassembly commences.

The spatial and temporal pattern of S100A11 flashing was reminiscent of intracellular Ca²⁺ flashes occurring in different cell types (Ellefsen et al., 2019; Varadarajan et al., 2022; Wei et al., 2009, 2010), including during FA disassembly (Giannone et al., 2002). Ca²⁺-binding activates S100A11 by shifting the homodimer into its active, F-actin-binding conformation, and local Ca²⁺ influx might therefore be a mechanism for S100A11 recruitment to peripheral FAs. We investigated whether artificially elevating intracellular Ca²⁺ levels would stimulate S100A11 recruitment and FA disassembly. For this, we co-transfected HeLa cells with GFP–S100A11 and the fluorescent Ca²⁺ sensor R-GECO1 (Zhao et al., 2011) for live-cell TIRF imaging. Adding the Ca²⁺ ionophore ionomycin to cells growing in Ca²⁺-containing medium strongly activated the Ca²⁺ sensor within <60 s, mirrored by a slight increase in the S100A11 signal (Fig. 4A,B; Movie 3), consistent with transient Ca²⁺-dependent membrane or actin cortex recruitment of S100A11. Initial S100A11 membrane recruitment was not specifically targeted to FAs but occurred evenly across the entire membrane. However, ionomycin treatment reliably induced strong secondary highly localized S100A11 flashes at the cell periphery within the next 5 to 10 min (Fig. 4A,B; Movie 3), and these were similar in size and duration to the random flashes observed under steady-state conditions in untreated cells (Fig. 2A). During the local secondary flashes, the R-GECO1 signal also showed mild transient elevation, suggesting a Ca²⁺-dependent recruitment mechanism of S100A11 to FAs. In a variation of the previous experiments, we also incubated cells in Ca²⁺-free medium containing ionomycin. This treatment never induced S100A11 flashes, but subsequent addition of Ca²⁺ again stimulated S100A11 flashes within a similar timeframe to that seen for ionomycin addition to Ca²⁺-containing medium. Using this protocol on HeLa cells co-expressing GFP–S100A11 and vinculin–mCherry corroborated that Ca²⁺-stimulated membrane recruitment of S100A11 coincided with rapid FA translocation and FA disassembly within these areas (Fig. 4C,D,E; Movie 4).

Although the ionomycin experiments showed that intracellular Ca²⁺ influx stimulates S100A11 recruitment to areas of dissembling FAs, the consistent delay between initial global membrane recruitment of S100A11 immediately after ionomycin addition and the targeted localization of S100A11 to FAs suggested the requirement for an additional mechanism besides Ca²⁺-activated F-actin binding by S100A11. Specifically, the enhanced FA translocation speeds just prior or during S100A11 recruitment pointed towards the possible involvement of actomyosin-driven SF contraction. In agreement, pre-incubation with the NMII inhibitor blebbistatin fully suppressed ionomycin-stimulated S100A11 recruitment, FA translocation and FA disassembly (Fig. 5A,B,D,E; Movie 5), while stimulating actomyosin contractility with low doses of calyculin A (5 nM) induced prominent FA localization of S100A11 (Fig. S6). Elevated intracellular Ca²⁺ levels have been previously shown to stimulate actomyosin contractility in different non-muscle cell types (Kong et al., 2019; Lee and Auersperg, 1980), possibly through Ca²⁺-dependent activation of myosin light chain kinase (MLCK) (Katoh et al., 2001). Holographic tomography imaging showed that ionomycin treatment strongly stimulates U2OS cell contraction and the formation of membrane blebs, cytoplasm-filled membrane extrusions driven by increased intracellular tension, whereas NMII inhibition by blebbistatin efficiently blocked both cell contraction and bleb formation (Fig. S7; Movie 7). Likewise, the MLCK inhibitor ML7 fully blocked ionomycin-induced contraction (Movie 8), whereas NMIIA (MYH9) knockout (NMIIA KO) cells were also unresponsive to ionomycin treatment (Movie 8). Finally, omitting extracellular Ca²⁺ from the imaging solution also prevented ionomycin-induced cell contraction (Movie 8). Importantly, imaging cells in Ca²⁺-free solution alone did not induce cell contraction or FA loss even at extended time periods (1 h; Movie 9), nor did blebbistatin preincubation disrupt FA and SF architecture (compare F-actin arrangement in Fig. 5B after 30 min blebbistatin preincubation). Together, these results suggest that ionomycin stimulates NMII-dependent contractility via Ca²⁺-mediated MLCK activation.

Recent findings demonstrate that NMII-dependent traction forces transmitted by actin SFs produce highly localized Ca²⁺ spikes near tensed FAs by recruiting the force-sensitive Ca²⁺-permeable channel Piezo1 (Ellefsen et al., 2019; Yao et al., 2022). Ca²⁺ influx after contractility-dependent Piezo1 activation near tensed FA sites also provided an intriguing possible explanation for Ca²⁺-activated S100A11 recruitment to tensed FAs. Indeed, pre-treatment with the Piezo1 inhibitor GsMT4x efficiently blocked ionomycin-induced S100A11 recruitment to FAs (Fig. 5C,D,E; Movie 6). Our results therefore suggest that ionomycin-mediated Ca²⁺ influx promotes S100A11 localization to FAs in two different ways: firstly, by stimulating NMII-dependent cellular contractility, leading to enhanced tensioning of FAs and local Piezo1 activation, and secondly by activating the F-actin binding capacity of S100A11 through a secondary Ca²⁺ signal generated by the stretch-opened Piezo1 channels.

We further investigated biomechanical aspects of S100A11 recruitment using a previously established micropipette-based pulling assay for exerting pulling forces onto individual FA sites (Riveline et al., 2001). In this assay, a glass micropipette is first lowered onto the perinuclear region of a target cell, and then gently pushed against the nuclear region (Fig. 6A), thereby applying tension forces onto peripheral FAs through interconnected SFs. First, we tested whether external pulling forces would recruit S100A11 to tensed FAs. Indeed, micropipette pulling induced S100A11 accumulation to stressed FAs within a similar time (3–5 min) to that seen for Ca²⁺-stimulated SF contraction (Fig. 6B,C; Movies 10 and 11). Interestingly, S100A11 again localized to the distal FA end, similar to what was observed during random S100A11 flashes in unstimulated cells (Fig. 3A,B). In contrast, micropipette pulling did not induce S100A11 recruitment in Ca²⁺-free medium, demonstrating a strict requirement for local S100A11 activation by Ca²⁺ (Fig. 6D). Likewise, no S100A11 recruitment occurred in Ca²⁺-containing medium in presence of the Piezo1 inhibitor GsMT4x (Fig. 6E). We furthermore tested whether external pulling could bypass the requirement for NMII-driven contractility, using cells incubated with blebbistatin (Fig. 6F) or NMIIA KO cells (Fig. 6G). In both cases external pulling still induced S100A11 recruitment to peripheral FAs, suggesting that the required role of NMII in S100A11 recruitment is limited to the establishment of SF contractility. However, recruitment in NMII-deficient cells appeared to be weaker compared to in control cells, raising the possibility that additional functions of NMIIA, or other NMII isoforms, also contribute to S100A11 translocation.

Although the above experiments demonstrate NMII- and Ca²⁺-dependent recruitment of S100A11 to peripheral FAs disassembling under force, it remained unclear whether S100A11 contributed to the FA disassembly process itself. To further investigate its role, we generated an S100A11-knockout U2OS cell line (S100A11 KO) and verified absence of expression by western blotting (Fig. 7A). We also analyzed FA morphology in wild-type and S100A11 KO cells (Fig. 7B–E). Compared to wild-type cells, S100A11 KO cells displayed a larger average individual FA area (Fig. 7C), caused primarily by a subpopulation of vastly enlarged FAs (>2.5 µm²) absent from wild-type cells (Fig. 7D). S100A11 KO cells also featured an increased total FA area per cell (Fig. 7E). S100A11 KO cells displayed a striking FA arrangement pattern characterized by dramatically elongated FAs extending along almost the entire length of thick, peripheral SFs, whereas wild-type cells rarely displayed this phenotype (Fig. S8). Re-expression of GFP–S100A11 on the other hand fully rescued the KO phenotype and reverted FA morphology and sizes to wild-type levels (Fig. 7B–E). FAs are perpetually remodeled, and cell type-specific FAs sizes reflect the balance between assembly and disassembly rates (Webb et al., 2002). Enlarged FAs in S100A11 KO cells could thus result from impaired FA disassembly. U2OS are comparatively stationary cells with some stable central FAs displaying little movement over several hours, but peripheral FAs disassemble rapidly during membrane retraction (Movie 12). Analyzing translocation velocities of peripheral FAs in retracting membrane regions in wild-type (Fig. 7F) and KO cells (Fig. 7G) over a 60 min observation period showed significantly reduced FA retraction speeds in S100A11 KO cells (Fig. 7I; Movie 13). Moreover, translocating FAs within membrane retractions remained large in S100A11 KO cells but shrank in wild-type cells (Fig. 7H,I). In agreement, FAs in retracting membrane regions of S100A11 KO cells displayed longer lifetimes (typically >60 min) compared to WT cell FAs, which turned over within 26.1±15.8 min (mean±s.d., n=20, Fig. 7J). Reduced FA translocation and disassembly rates in S100A11 KO cells support a role for S100A11 in NMII contractility-dependent FA disassembly during membrane retraction.

Here, we report a novel localization of S100A11 to FAs, both dynamically to disassembling peripheral FAs, and at comparatively steady levels at stable, stationary FAs. These two different localization patterns suggest a dual role for S100A11 in both FA maintenance and disassembly. Strikingly, S100A11 levels increased sharply at the onset of peripheral FA disassembly in a process requiring NMII-based contractility and Piezo1-dependent Ca²⁺ influx at stressed adhesion sites. Our results therefore reveal a mechano-sensitive role of S100A11 and implicate it in intracellular tension-driven FA disassembly mechanism (Fig. 8).

FAs are dynamic cell–matrix adhesion sites that undergo continuous rearrangement during cell spreading, shape changes and migration (Webb et al., 2002), and actomyosin contractility is a known modulator of FA turnover (Aureille et al., 2023 preprint). In migrating cells, FA disassembly typically occurs at the rear, facilitating cell membrane retraction and forward propulsion (Broussard et al., 2008; Kirfel et al., 2004). Cell membrane retraction requires actomyosin contractility (Chen, 1981), and there is abundant evidence that NMII-dependent contractile forces also contribute to peripheral FA translocation and disassembly (Crowley and Horwitz, 1995; Wolfenson et al., 2011). However, the role of NMII in regulating FA lifetime is complex, given that NMII activity is also required for FA growth and stabilization (Chrzanowska-Wodnicka and Burridge, 1996; Hirata et al., 2015; Oakes et al., 2012; Ridley and Hall, 1992). A resolution of this apparent discrepancy might reside in the fact that the molecular composition of FAs itself is force sensitive (Kuo et al., 2011; Wolfenson et al., 2011), and increasing NMII-mediated contraction forces above a certain threshold might shift the balance between FA stabilizers and destabilizers.

Interestingly, a previous proteomics screen identified S100A11 as part of the NMII-responsive proteome of FAs (Kuo et al., 2011). However, this study found an inverse relation between NMII-activity and S100A11 integration into FAs, in contrast to our findings. S100A11 was also detected in isolated FAs in a later proteomics study (Huang et al., 2014). The short-lived, transient recruitment regime of S100A11 suggests a function of S100A11 in initiating FA disassembly, rather than regulating the disassembly process itself. Similarly, during plasma membrane wound healing, Ca²⁺-activated S100A11 rapidly translocates to the injury site, but S100A11 levels then drop back to base levels within less than 30 s (Ashraf and Gerke, 2022), likely before completion of actin-driven membrane resealing. S100A11 might thus play functional roles in initiating dynamic actin remodeling processes in different cellular contexts.

In some cases, S100A11 localized predominantly to the distal tip of disassembling FAs. According to traction force experiments with high spatial resolution, peak traction in individual FAs is shifted towards the distal FA end (Plotnikov et al., 2012), further supporting a traction force-dependent recruitment mechanism of S100A11. In some cases, S100A11 also localized to broader areas surrounding individual or groups of disassembling FAs, suggesting that S100A11 might also regulate cellular structures in the FA vicinity. FAs are intricately linked to surrounding cortical F-actin (Vignaud et al., 2021), and FA disassembly might require severing or remodeling such cytoskeletal crosslinks. Alternatively, membrane ripping might occur during fast FA translocation and disassembly (Kirfel et al., 2004), which could trigger the well-established role of S100A11 in plasma membrane wound healing. However, we usually observed strongest S100A11 localization signals just prior to rapid FA translocation, when membrane tears are unlikely to have occurred. In any case, localization to FAs, SFs, and membrane areas adjacent to disassembling FAs points towards a multifunction role of S100A11 in FA regulation.

Although transient S100A11 localization to FAs reliably predicted subsequent FAs disassembly, and aberrant FA morphology and delayed FA disassembly during membrane retraction in S100A11 KO cells confirmed a functional role of S100A11 in FA disassembly and the underlying molecular mechanisms, as well as its binding partners in and around the FA plaque, remain to be elucidated. Surprisingly, global Ca²⁺ influx after ionomycin treatment initially provoked only weak, diffuse S100A11 membrane recruitment, whereas specific recruitment to FA sites required a secondary Ca²⁺ signal through tension-activated Piezo1 channels in the FA vicinity. Apparently, S100A11 recruitment to specific cellular compartments requires precise local and temporal control of free Ca²⁺ levels. Along similar lines, local rather than global Ca²⁺ increases are required for FA disassembly (Giannone et al., 2004). Successful isolation of S100A11–annexin A1 complexes also requires precisely controlled cellular Ca²⁺ thresholds (Gerke and Moss, 2002), further underlining the Ca²⁺-concentration dependency of S100A11 interactions with binding partners.

Besides Ca²⁺-activated F-actin binding, additional mechano-regulated mechanisms might target S100A11 specifically to tensed FAs. FAs contain a variety of tension-sensitive proteins and force-induced unfolding of these proteins could expose specific binding sites for S100A11 in tensed FAs. S100A11 might also serve as a docking site for the subsequent recruitment of additional FA regulators, or it might co-translocate with other proteins. Here, potential candidates are annexins, which cooperate with S100A11 in the context of plasma membrane wound healing (Jaiswal et al., 2014), membrane organization (Chang et al., 2007) and endocytosis (Seemann et al., 1997).

Additional FA disassembly mechanisms have been outlined in detail, with microtubule-dependent processes receiving particular attention. Disrupting microtubules by nocodazole reinforces SFs and stabilizes FAs (Bershadsky et al., 1996; Danowski, 1989; Liu et al., 1998), whereas microtubule repolymerization after nocodazole wash-out quickly induces global FA disassembly (Ezratty et al., 2005). Likewise, repeated microtubule contact targets specific FAs for disassembly (Kaverina et al., 1998; Small et al., 2002; Small and Kaverina, 2003), for instance by stimulating focal adhesion kinase, dynamin- and clathrin-mediated endocytosis of FA components (Chao and Kunz, 2009; Ezratty et al., 2009). Moreover, Ca²⁺ influx-dependent mechanisms have been identified, in particular calpain-mediated cleavage of FA components (Huttenlocher et al., 1997), which, however, can also act downstream of microtubule-induced FA disassembly pathways (Bhatt et al., 2002). Similar to what is seen with ionomycin stimulation, nocodazole washout also stimulates cell contraction, suggesting that both pathways might share common molecular mechanisms involving contractility regulation and membrane tension-dependent local Ca²⁺ influx. In agreement, recent findings have established a mechanistic link between microtubule- and NMII-dependent FA disassembly (Aureille et al., 2023 preprint).

Several recent studies have identified tension-dependent recruitment and activation of Piezo1 channels to FAs (Ellefsen et al., 2019) and implicated the resulting local Ca²⁺ signals in the modulation of FA function (Yao et al., 2022). Piezo1 activation at tensed FAs might either result from enhanced membrane tension or via direct linkage of Peizo1 to still unknown FA components (Yao et al., 2022). Asymmetric local Ca²⁺ flickers have previously been identified as a mechanism to steer cell migration (Wei et al., 2009), likely because they target specific FAs for disassembly. MLCK-dependent phosphorylation of myosin II has been identified as the force-producing mechanism driving Piezo1 activation (Ellefsen et al., 2019). Given that MLCK itself is activated by intracellular Ca²⁺, this raises the intriguing possibility of a feedback loop in which Piezo1-mediated Ca²⁺-dependent influx near tensed FAs further stimulates actomyosin contractility, leading to further Piezo activation and so on, as has been previously suggested (Ellefsen et al., 2019; Hirata et al., 2015).

Cancer cell migration and invasion crucially depend on actin cytoskeleton remodeling and adhesion modulation, and actin-binding proteins are frequently dysregulated in cancer and contribute to malignancy and unfavorable clinical prognosis (Izdebska et al., 2020; Suresh and Diaz, 2021). S100A11 also has a well-established role in cancer (Bresnick et al., 2015), promoting aggressive traits, such as increased pseudopodal protrusion, migration, invasion and metastasis (Niu et al., 2016; Shankar et al., 2010). In this context, the newly identified role of S100A11 in regulating spatially controlled FA disassembly could contribute to its pro-migratory and -invasive properties. In conclusion, our study identifies a novel NMII-, Piezo1- and Ca²⁺-dependent role for S100A11 during force-dependent FA disassembly. The novel identification of S100A11 as a mechano-responsive protein thus further expands its diverse functional repertoire.

Human osteosarcoma U2OS cells were kindly provided by Tsukasa Matsunaga, Kanazawa University, Japan; HeLa cells were obtained from the ATCC. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% L-glutamine, penicillin (10,000 U/ml) and streptomycin sulfate (10,000 µg/ml) (Gibco). For maintenance, cells were passaged two times a week and incubated in a humidified incubator at 37°C, 5% CO₂. Transient plasmid transfection was performed using Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer's instructions. Lifeact–mCherry was Addgene plasmid #67302 (deposited by Moritoshi Sato; Kawano et al., 2015), GFP-S100A11 was Addgene plasmid #107201 (deposited by from Volker Gerke and Ursula Rescher). The vinculin–mCherry construct has been published previously (Carisey et al., 2013) and R-GECO1 Addgene plasmid #32444 (deposited by Robert Campbell; Zhao et al., 2011).

To produce recombinant His-tagged GFP–S100A11, cDNA encoding S100A11 from a human cDNA library was amplified by PCR using the forward primer 5′-ACTATACGGTGGATCCatggcaaaaatctccagcc-3′ and the reverse primer 5′-GCAGGTCGACAAGCTTtcaggtccgcttct-3′ purchased from Macrogen, Japan (lowercase letters highlight the sequence specific to S100A11). The N-terminal His-tag (x6His) GFP–S100A11 construct was then generated by infusion cloning. S100A11 was inserted into the linear PcoldI His-tag GFP plasmid (Takara, Japan), which was cut using BamHI and HindIII sites. The nucleotide sequence of the PcoldI His-tagged GFP–S100A11 expression vector was confirmed by DNA sequencing and restriction enzyme digestion. Protein expression was performed in pG-KJE8 Escherichia coli BL21 (Takara, Japan) using the protocol provided by the manufacturer. Purification of the His-tagged GFP–S100A11 protein was performed as described in a previous study (Ngo et al., 2015).

The U2OS S100A11 KO cell line was generated by genome editing using a pCas9P plasmid expressing Cas9 [pSpCas9(BB)-2A-Puro (PX459) V2.0 deposited by Feng Zhang, Addgene plasmid #62988; Ran et al., 2013] and two sgRNA sequences (#57891 & #57892, both on exon2 of S100A11) selected from the Brunello human CRISPR knockout library (Doench et al., 2016). The sgRNA sequences were: 5′-AGAACTAGCTGCCTTCACAA-3′ and 5′-CCAGCCCTACAGAGACTGAG-3′. After cloning of the S100A11-KO plasmid, 1 µg of the constructed plasmid was transfected into U2OS cells using Lipofectamine 3000 transfection reagent. The following day, the transfection medium was removed, and cells were cultured in growth medium containing 10 µg/ml puromycin for 3 days to select for knockout cells. Finally, single clones were isolated by limited dilution into 96-well plates and then expanded. Absence of S100A11 expression was confirmed by western blotting using rabbit anti-S100A11 (ProteinTech Group). Overall cell shape and SF and FA morphology of three clones were characterized and found to be indistinguishable, and all subsequent experiments were conducted with a single clone.

For immunofluorescence staining cells were cultured in uncoated 35 mm glass bottom dishes (either FD35 Fluorodish, WPI, or 1130H, Matsunami Glass) for 24 h to 48 h and then fixed in 4% paraformaldehyde solution in PBS for 10 min at room temperature (RT). This was followed by simultaneous blocking and permeabilization in a solution composed of 50 mM glycine, 0.05% Tween 20, 0.1% Triton X-100 and 0.1% BSA in PBS for 30 min. Primary antibodies (rabbit anti-S100A11, GeneTex, 1:200; Mouse monoclonal anti-β-tubulin, Sigma-Aldrich, clone TUB 2.1, 1:500; and mouse anti-Vinculin Merck, V9131, 1:500) were then diluted in PBS containing 10 mM glycine, 0.05% Tween 20 and 0.1% Triton X-100, and cells were incubated for 2 h at RT or overnight at 4°C. After three washes with PBS to remove unbound primary antibodies, cells were stained with the appropriate secondary antibodies [goat anti-rabbit-IgG conjugated to Alexa Fluor (AF)568, goat anti-rabbit-IgG conjugated to AF488, goat anti-mouse-IgG conjugated to AF568, or goat anti-mouse-IgG conjugated to AF488, all Invitrogen) diluted 200× in 0.1% Tween 20 in PBS for 30 min. Phalloidin–AF568 or phalloidin–AF488 (Invitrogen) diluted at 1:200 were added along with secondary antibodies when cells were co-stained for F-actin. Finally, cells were washed three times with PBS and imaged using the 60× objective of an inverted fluorescence phase contrast microscope (BZ-X810 Keyence All-in-one Fluorescence Microscope).

Total internal reflection fluorescence (TIRF) microscopy was performed using a Nikon Ti-E inverted microscope equipped with a 100× CFI Apo TIRF objective lens and 488 nm and 561 nm lasers (Sapphire, Coherent). A high-sensitivity electron-multiplying charge-coupled-device camera (iXon3; Andor, DU897E-CS0) with electron multiplying and preamplifier gains of 296 and 2.4×, respectively, was used in combination with a 1.5× C-mount adapter (Nikon) to acquire images. The Z-position was adjusted for optimal focus and maintained using the Nikon Perfect Focus System during time-lapse imaging. Cells were imaged at 28°C under ambient air conditions in culture medium containing 20 mM HEPES (pH 7.4). Timelapse series were stabilized, and histograms were equalized in Fiji. Kymographs were generated using the ‘KymoResliceWide’ (https://github.com/UU-cellbiology/KymoResliceWide) or ‘Multi Kymograph’ (https://biii.eu/multi-kymograph) Fiji plugins. Fluorescence intensity plots were generated in Fiji/ImageJ after correcting for photobleaching. For this, first a region of interest (ROI) was selected in a feature-free area and the intensity signal was plotted versus time using the ‘Plot Z-axis profile’ command. The fluorescence decay curve was then fitted using the ‘Exponential with offset’ function and the offset value denoting background fluorescence intensity was stored. For quantification of fluorescence intensity in cellular subregions containing FAs, suitable ROIs were selected and photobleaching correction was performed using the ‘Simple ratio’ function and the background intensity value determined in the first step.

Cell lysates were prepared by washing cells once with ice-cold PBS and then scrapping the cells in 200 μl Triton X-100 lysis buffer containing 1% (v/v) Triton X-100, 150 mM NaCl, 50 mM Tris-HCl pH 7.4 and Complete Protease Inhibitor Cocktail (Thermo Fisher Scientific). The mixture was kept for 30 min on ice and then centrifuged at 13,800 g for 20 min at 4°C. 20 μl of protein sample was mixed with 4 μl of 6× SDS sample buffer, boiled at 95°C for 5 min, and then loaded onto 10–20% gradient Mini-PROTEAN TGX Precast gels (Bio-Rad) for SDS-PAGE. The electrophoresis was first run for ∼40 min at 80 V, followed by 150 V for∼60 min. After electrophoresis, proteins were transferred onto 0.2 µm PVDF membranes in a wet tank apparatus (Bio-Rad) run at 100 V for 1 h at room temperature while cooling with an ice pack in a Tris-Glycine transfer buffer containing 20 mM Tris, 200 mM glycine, and 20% (v/v) methanol. Subsequently, the membrane was blocked with 5% non-fat dried milk in TBS-T [20 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.05% (v/v) Tween 20] for 1 h and then incubated with primary antibodies [rabbit anti-S100A11 (ProteinTech), and mouse monoclonal TUB 2.1 anti-β-tubulin (Sigma-Aldrich) at 1:1000 dilution in blocking buffer overnight at 4°C]. After three washes in TBS-T for 10 min, blots were incubated with HRP-conjugated secondary antibodies [goat anti-rabbit-IgG conjugated to HRP (Invitrogen) or goat anti-mouse-IgG conjugated to HRP (SouthernBiotech)] at 2000× dilution in blocking buffer for 1 h. After three final washes in TBS-T, blots were developed using the Western Lightning Plus-ECL chemiluminescence substrate and imaged using an AE-9300 Ez capture MG (ATTO, Tokyo, Japan) imaging system. Full images of western blots shown in this paper are presented in Fig. S9.

For ionomycin-induced FA disassembly, cells were grown on glass bottom dishes for 16 h. Cells were then transferred to the TIRF microscopy platform and 3 µM ionomycin (Merck) was added directly during live-cell image acquisition. Typically, cells were imaged for 1 h until FA disassembly had completed. To assess the effect of NMII inhibition on ionomycin-induced GFPSA11 accumulation at FA, HeLa cells transiently transfected with S100A11–GFP were transferred onto the TIRF stage and incubated for 30 min in the presence of 50 µM blebbistatin (Merck). After addition of ionomycin (final concentration 3 µM), cells were imaged for 20–30 min. The effect of NMII inhibition on ionomycin induced GFP–S100A11 accumulation was then assessed by comparing the fluorescence intensity of blebbistatin-treated cells with that of untreated control cells (five cells per group). The effect of NMII inhibition on cell contraction was measured by treating untransfected U2OS cells with 50 µM of blebbistatin followed by immediate live-cell imaging by holographic tomography (3D Explorer, NanoLive). After 30 min of imaging in blebbistatin-containing medium, cells were treated with 3 µM ionomycin and imaged for a further 20–30 min. MLCK inhibition was achieved by incubating cells with 30 μM ML7 hydrochloride (MedChemExpress) for 30 min, after which cells were treated with 3 μM ionomycin. To examine the role of Piezo1 in mediating Ca²⁺ influx and S100A11 recruitment near stressed FAs, GFP–S100A11-transfected cells were treated with 3 μM of the Piezo1 inhibitor GsMTx4 (Peptide Institute) for 3 h. Subsequently, cells were stimulated with 3 μM ionomycin, and the dynamics of GFP–S100A11 was investigated by live-cell TIRF imaging. In other experiments, GsMTx4-treated cells were subjected to external force application using a micropipette assay. Cell contractility was stimulated by incubating cells in low doses of calyculin A (5 nM, Alomone labs) during live-cell imaging by holographic tomography (U2OS cells transfected with Lifeact–mCherry) and TIRF microscopy (HeLa cells transfected with GFP-S100A11).

U2OS wild-type and S100A11 KO cells were transfected with vinculin–mCherry using Lipofectamine 3000 (Thermo Fisher Scientific). After 24 h, cells were replated in 35-mm glass-bottom dishes coated with 20 μg/ml fibronectin for imaging. Time-lapse image sets were acquired at 1 frame per minute for a total of 60 min at 37°C in a chamber providing 5% CO₂ and an automated fluorescence microscope (KEYENCE BZ-X810). The acquired images were analyzed in ImageJ using background subtraction, frame stabilization and photobleaching correction plug-ins. To determine FA dynamics, peripheral cell regions displaying active membrane retraction (net cell edge movement >5 µm over 1 h) were identified. Within these areas, 15×15 μm² subregions were selected, which typically contained 10 to 25 FAs. In these regions, FAs were then tracked using the Cellpose TrackMate module in ImageJ (Stringer et al., 2021). Individual FA area and retraction speeds were obtained from at least 70 FAs from five different cells per cell type. Non-dynamic FA parameters (individual FA size and FA area per cell) were determined from immunofluorescence images of fixed U2OS wild-type, S100A11 KO and S100A11 KO/GFP-S100A11 rescue cells by adopting a previously described method (Horzum et al., 2014). Statistical tests for significance were performed in the GraphPad Prism software or Origin 2021.

Micropipette external force application experiments were performed based on a previous report (Riveline et al., 2001). Briefly, glass micropipettes were fabricated by pulling borosilicate glass capillaries (inner diameter 0.58 mm, outer diameter 1.00 mm) into nanopipettes with a tip diameter of ∼100 nm using a CO₂ laser puller (Model P-2000, Sutter Instruments Co., USA). For external force application experiments, U2OS cells transiently expressing GFP–S100A11 were plated into 35 mm glass bottom dish coated with 10 μg/ml human plasma fibronectin and grown overnight. After transferring cells into growth medium containing 20 mM HEPES, pH 7.4, the cell sample was placed onto an inverted Olympus CKX53 light microscope equipped with a 40× lens (LUCPLFLN40X), a camera (U3-3880CP-M-GL Rev.2.2, IDS) and a U-LGPS fluorescence light source (Olympus). A micropipette mounted on a manual x,y,z-micromanipulator system (Narishige, Japan) was then positioned over a target cell, while the intracellular localization of GFP–S100A11 was recorded by timelapse fluorescence microscopy. The micropipette was then lowered until contacting the plasma membrane of the recorded cell, and subsequently gently pushed against the protruding cell nucleus, thereby exerting pulling forces against peripheral FAs. GFP–S100A11 localization to stressed FA was typically observed within 30–180 s of force application. In some cases, cells were co-transfected with GFP–S100A11 and vinculin–mCherry to verify FA localization. The requirement for non-muscle NMIIA (NMIIA) was further investigated by using U2OS NMIIA knockout cells (Weißenbruch et al., 2021) or by preincubating U2OS wild-type cells with blebbistatin (50 μM) for 30 min before force application. Additional experiments were performed in Ca²⁺-free medium to investigate the requirement of Ca²⁺ for S100A11 translocation.

We thank Toshio Ando, Hiroki Konno, Satoshi Arai and Cong Quang Vu (WPI NanoLSI, Kanazawa University) for generous help with laboratory equipment and reagents and Volker Gerke (Universität Münster) for an S100A11–mApple expression plasmid. We also thank WPI NanoLSI for the use of core facilities (cold room, spectrophotometer and ultracentrifuges).

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.261492.reviewer-comments.pdf