Mechanical forces stimulate Golgi export

Exocytosis enables cells to deliver molecules to their surface or secrete them into the extracellular space, thereby maintaining plasma membrane (PM) and ECM homeostasis and organization and facilitating interactions with the environment. Newly synthesized secretory and membrane proteins are transported along intracellular routes from the ER through the Golgi complex and to the TGN. At the TGN, cargos are sorted and packaged into transport carriers, which are rapidly directed to the PM along microtubule (MT) tracks. Upon carrier fusion with the cell surface, transmembrane proteins are incorporated into the PM, while soluble cargos are released into the extracellular space through exocytosis.

Similar to the highly heterogeneous organization of the PM, exocytic events are spatially regulated (Grigoriev et al., 2007; Grigoriev et al., 2011; Yuan et al., 2015; Robinson et al., 1995; Huet-Calderwood et al., 2023; Fourriere et al., 2019; Stehbens et al., 2014), ensuring that secretion is precisely coordinated with cellular architecture and function. Previous studies have demonstrated that many exocytic events occur preferentially near focal adhesions (FAs) (Stehbens et al., 2014; Huet-Calderwood et al., 2017; Eisler et al., 2018; Lachuer et al., 2023; Huet-Calderwood et al., 2023; Fourriere et al., 2019). In particular, by combining the Retention Using Selective Hooks (RUSH) assay (Boncompain et al., 2012)—which synchronizes anterograde cargo transport—with a Selective Protein Immobilization (SPI) assay, we previously mapped the precise spatial organization of cargo arrival sites at the PM. Interestingly, our findings revealed that exocytosis does not occur randomly across the cell surface but rather at hotspots juxtaposed to FAs (Fourriere et al., 2019). Among the carriers involved in these targeted exocytic events are Golgi-derived, RAB6-positive carriers (Fourriere et al., 2019). RAB6 mediates multiple Golgi-to-PM export routes, including the trafficking of CARTS (carriers of the TGN to the cell surface), a class of RAB6- and protein kinase D (PKD)-dependent Golgi-to-PM transport carriers (Wakana et al., 2012).

FAs are dynamic, force-sensitive molecular platforms localized at the cell membrane. The core components of FAs are integrins, heterodimeric transmembrane receptors that connect the ECM with cytoplasmic components, enabling the transmission of extracellular signals into the cell. These outside-in signals control essential processes such as cell migration, tissue morphogenesis, and PM homeostasis (Kanchanawong and Calderwood, 2023). While substantial progress has been made in understanding how FAs (Kechagia et al., 2019) and the nucleus (Kirby and Lammerding, 2018; Cho et al., 2017) sense and transduce extracellular mechanical stimuli, it is far less understood whether and how intracellular organelles such as the Golgi reciprocally adapt their activity to these cues and thereby contribute to the overall cellular mechanoresponse. The aforementioned observation that membrane regions proximal to FAs act as hotspots for exocytosis suggests a spatial coordination of membrane trafficking and adhesion sites. Exocytosis delivers integrins, ECM components (such as collagens, fibronectin [FN], and fibrillin), and signaling molecules to the PM (Moreno-Layseca et al., 2019), thereby sustaining adhesion dynamics. While MTs have long been recognized to target FAs (Schmidt and Stehbens, 2024; Krylyshkina et al., 2003; Seetharaman and Etienne-Manneville, 2019; Kaverina et al., 1998), emerging evidence indicates that exocytosis near FAs is spatially orchestrated by MT guidance, RAB GTPase-dependent Golgi-to-PM trafficking, and, potentially, mechanochemical feedback loops. This spatial regulation may provide a mechanism for localizing membrane expansion, modulating membrane tension, and remodeling the ECM and PM proteome—processes critical for adhesion homeostasis during cell spreading and migration. Understanding these dynamics may also offer insights into how dysregulation of these processes contributes to pathological conditions such as cancer invasion.

While several mechanosensitive modules, such as FAs, fibrillar adhesions, and adherens junctions at the PM (De Pascalis and Etienne-Manneville, 2017; Elosegui-Artola et al., 2018), and the nuclear envelope (Thorpe and Lee, 2017; Amar et al., 2021; Jahed and Mofrad, 2019) have been well characterized, recent studies indicate that other organelles, such as the ER and lysosomes, can respond to mechanical forces (Phuyal and Baschieri, 2020; Rawal et al., 2025; Merta et al., 2025; Aceitón et al., 2025; Ren et al., 2025; Naughton et al., 2025). Moreover, mechanical forces applied to the PM can trigger Ca²⁺ release from the ER via an actomyosin-dependent mechanism (Kim et al., 2015). In the context of intracellular trafficking, neurite mechanical tension enhances active vesicular transport in neurons (Ahmed and Saif, 2014), while COPII-mediated transport from the ER to the Golgi apparatus is regulated by mechanical strain through interactions between the small G proteins Rac1 and Sar1 (Phuyal et al., 2022) and is sensitive to ER membrane viscosity (Jiménez-Rojo et al., 2025). The Golgi apparatus itself is mechanoresponsive: external forces generated by adhesion to the ECM influence lipid metabolism and Golgi mechanics (Romani et al., 2019), and cell–matrix adhesion controls Golgi organization and secretory output via an ARF1- and AXL-dependent pathway (Singh et al., 2018; Saha et al., 2026; Chakraborty et al., 2026; Joshi et al., 2026). Because the actomyosin cytoskeleton (Zilberman et al., 2011; Miserey-Lenkei et al., 2010) and MT networks (Fourriere et al., 2020) are closely associated with Golgi membranes, forces transmitted through the cytoskeleton can directly alter Golgi dynamics (Miserey-Lenkei et al., 2017). In addition, using optical-tweezers based microrheology, we showed that directly applying force to the Golgi apparatus delays actin-mediated membrane fission and thereby perturbs post-Golgi trafficking, indicating that Golgi membranes are mechanosensitive (Guet et al., 2014).

In this study, we address the impact of extracellular mechanical forces on post-Golgi carrier biogenesis. Our results show that conditions associated with impaired FA formation are accompanied by reduced Golgi-derived carrier biogenesis, whereas both substrate stiffening and mechanical stretching enhance Golgi export capabilities. Using Halo-Flipper probes, we further show that extracellular mechanical forces are accompanied by changes in Golgi membrane mechanical properties. This is paralleled with an increase in MT acetylation, changes in Golgi diacylglycerol (DAG) content, and PKD activity. In turn, inhibition of Golgi export suppresses this mechanoresponse, resulting in impaired cell spreading. Collectively, our data identify the Golgi apparatus as a mechanoresponsive organelle, revealing that its export activity is modulated by extracellular mechanical stimuli and is crucial for efficient cell spreading and mechanoadaptation.

In nonpolarized 2D cell cultures, we and others have shown that exocytic sites are not randomly distributed across the PM but preferentially localize close to FAs in a RAB6-dependent manner (Fourriere et al., 2019). RAB6 is a general regulator of post-Golgi carrier biogenesis, including CARTS (Wakana et al., 2012). However, whether CARTS exocytosis itself is spatially regulated with respect to FAs has not been directly tested. To address this, we mapped the delivery of the CARTS-specific secretory cargo pancreatic adenocarcinoma upregulated factor (PAUF). HeLa cells transiently co-expressing paxillin-GFP (FA marker) and mKate2-FM4-PAUF (CARTS cargo) were imaged by live-cell total internal reflection fluorescence (TIRF) microscopy. After exiting the Golgi apparatus, PAUF-positive carriers moved quasi-directionally through the cytoplasm (Video 1) and docked preferentially near FA-rich regions of the PM (Fig. S1 A). At these sites, CARTS often paused for several seconds before a sudden loss of fluorescence, consistent with fusion with the PM (Fig. S1 A, panel [i]). By contrast, outside FA areas, CARTS were absent or observed only transiently, without apparent pausing, suggesting that these carriers were likely in transit toward FA sites for exocytosis (Fig. S1 A, panel [ii]). To further support this spatial relationship, we examined whether CARTS distribution is affected by depletion of ELKS, a cortical RAB6 effector previously shown to promote docking and fusion of RAB6-positive secretory carriers at exocytic hotspots near FAs (Fourriere et al., 2019). ELKS knockdown by siRNA (Fig. S1 B) led to an accumulation of PAUF-positive CARTS in cytoplasmic regions proximal to FAs (marked by paxillin) (Fig. S1, C and D). This is consistent with previous observations for general RAB6 carriers and supports the idea that CARTS are preferentially targeted to FA-proximal exocytic hotspots (Fourriere et al., 2019). Finally, using an SPI assay at 40 min after cargo release from the ER, secreted PAUF accumulated at the periphery of the cell in proximity to FAs (Fig. S1, E and F), similar to previous findings combining SPI and RUSH of other secretory cargoes (Fourriere et al., 2019).

To test whether integrin-mediated adhesion modulates Golgi-to-PM trafficking, we devised two complementary assays: (1) a combined spreading-secretion assay and (2) modulation of FAs by substrate stiffness. First, for the combined spreading-secretion assay (Fig. 1 A), HeLa cells in suspension were plated on glass coated with FN or poly-L-lysine (PLL). As previously described (Fourriere et al., 2019), PLL-seeded cells failed to assemble FAs, contrary to FN, which promoted FA formation (Fig. S2 A). Cells were co-transfected with GFP-mem (a general PM marker, see Materials and methods) and mKate2-FM4-PAUF to measure by confocal imaging both cell spreading (cell adhesion area) and CARTS biogenesis (number of CARTS per cell at 30 min after cargo release from the ER) (Fig. 1 A, see Materials and methods). On FN, cells exhibited larger adhesion areas and produced significantly more CARTS than on PLL (Fig. 1, B–D). We note that adhesion area is distinct from total cell surface area and cell volume and that spreading can be accompanied by changes in cell height and volume (Guo et al., 2017; Venkova et al., 2022). CARTS numbers in cells seeded on FN increased over time, peaking at 4 h after seeding, whereas PLL-seeded cells showed persistently low CARTS counts even after 24 h (Fig. 1 C). Likewise, the adhesion area of cells seeded on FN steadily expanded over 4 h, while it remained minimal and unaltered over 24 h in cells seeded on PLL (Fig. 1 D). Similar results were obtained when cells were gently lifted with EDTA prior to the spreading-secretion assay, as compared with trypsinization, indicating that the observed differences were not attributable to trypsin-mediated cleavage of cell surface proteins or trypsin-induced signaling (Fig. S2 B). Importantly, CARTS number scaled linearly with adhesion area under all conditions, resulting in a constant number of CARTS per adhesion area (Fig. 1, E and F). These results suggest that adhesion-dependent cell states characterized by increased adhesion area associate with enhanced Golgi export capacity in a commensurate manner.

Second, we performed a substrate stiffness assay to further manipulate FA assembly, as the number of FAs decreases with decreasing substrate stiffness, resulting in minimal FA formation on soft substrates (Pelham and Wang, 1997; Barber-Pérez et al., 2020). HeLa cells were plated for 4 h on FN-coated polyacrylamide (PAA) gels of defined stiffness (2 kPa for soft gels and 30 kPa for stiff gels) or on glass, and FAs were visualized by immunofluorescence microscopy using vinculin staining (Fig. S2 C). Cells on soft substrates spread poorly and formed almost no vinculin-positive FAs, whereas cells on stiff substrates displayed robust spreading and abundant FAs (Fig. S2 A, left panels and Fig. S2 C). We then assessed Golgi export using the RUSH assay, which enables synchronized trafficking of selected cargos (Boncompain et al., 2012). HeLa cells stably expressing RUSH-EGFP-CD59 (a RAB6-dependent GPI-anchored protein) were plated on soft (2 kPa PAA gels) and stiff (30 kPa PAA gels or glass) FN-coated substrates for 4 h. Post-Golgi carriers were quantified at 45 min after biotin addition (Fig. 1 G). Cells on stiff substrates exhibited ∼150–200 carriers per cell, compared with ∼50 carriers per cell on soft substrates (Fig. 1 H). To directly assess whether active β₁ integrin levels at the basal membrane scale with this response, we stained cells using the active conformation-specific 9EG7 antibody (Bazzoni et al., 1995). While active β₁ integrin levels at the basal cell surface tended to be higher on FN-coated glass than on 2 or 30 kPa substrates, this difference was not statistically significant (Fig. S2, D and E). By contrast, FA area was strongly reduced on 2 or 30 kPa substrates as compared with glass, consistent with reduced spreading and lower carrier abundance under these conditions (Fig. S2 F). This is in line with previous work showing that substrate stiffness regulates adhesion organization, including the elongation of active α₅β₁-integrin–positive fibrillar adhesions (Barber-Pérez et al., 2020). Overall, our data support the hypothesis that adhesion-dependent cell states characterized by robust spreading and FA formation trigger an adaptive response in the secretory pathway.

Cell adhesion and spreading on integrin-activating substrates generate mechanical forces via integrin engagement and actomyosin contractility (Geiger et al., 2009). We have found that this scales with increased Golgi export capacity (Fig. 1). To test whether a distinct external mechanical cue produces similar effects, we applied sustained isotropic mechanical strain on cells using an equibiaxial stretching device (Fig. 2 A) (see Materials and methods [Le Roux et al., 2025]), a mechanical stimulus that mimics ECM stretching. HeLa cells expressing PAUF-mRFP were seeded on FN-coated stretchable PDMS membranes and allowed to adhere for 1 h at 37°C. To synchronize cargo at the TGN, cells were then shifted to 20°C for 2 h in the presence of cycloheximide (CHX). After shifting back to 37°C to resume Golgi export, cells were either left unstretched or subjected to a 15% isotropic strain for 30 min before fixation (Fig. 2 B). Widefield imaging of the PAUF-mRFP signal revealed that mechanical stretch increased the number of cytoplasmic CARTS by ∼40% compared with unstretched controls (Fig. 2, C and D). Measurements of cell adhesion area showed a variable increase (∼12% on average) upon stretch, which did not reach statistical significance (Fig. 2 E). This may reflect dynamic remodeling and partial relaxation of the adhesion area during the 30-min sustained stretch. In addition, the number of CARTS per adhesion area showed a slight but not statistically significant increase (Fig. 2 F). These data suggest that although cells rapidly remodel their adhesion footprint under sustained stretch, Golgi-derived carriers remain elevated at the end of the 30-min sustained stretch period. Stretch-induced changes in FA tension and/or mechanotransduction signaling may also contribute to the observed phenotype.

Our previous work using optical tweezers demonstrated that forces applied directly on Golgi membranes alter their mechanical properties and that actin depolymerization decreases Golgi rigidity (Guet et al., 2014). More recently, fluorescence lifetime imaging microscopy (FLIM)-based sensors of the Flipper family, such as Halo-Flipper, have been developed as lipid packing and membrane tension reporters (Strakova et al., 2019; Colom et al., 2018). Flippers respond to lipid packing and, thereby, membrane tension by shifting between planar and twisted conformations, which can be quantified by FLIM. We therefore tested whether extracellular mechanical forces alter Golgi tension using Halo-Flipper. First, to validate the use of Halo-Flipper at Golgi membranes, we exploited the effect of Latrunculin A on actin depolymerization, which is expected to reduce Golgi tension. RPE1-ManII-Halo stably expressing cells treated with low doses of Latrunculin A displayed compaction of the Golgi apparatus, as previously shown (Egea et al., 2006), and a significant decrease in Halo-Flipper fluorescence lifetime at the Golgi (Fig. 3, A and B), consistent with reduced membrane tension. These results align with our previous optical tweezers measurements (Guet et al., 2014) and confirm that Halo-Flipper can report changes in Golgi membrane mechanical properties.

We then examined the effect of two extracellular mechanical perturbations on Golgi tension. First, RPE1-ManII-Halo cells were seeded on FN-coated glass slides and allowed to spread for 0.5, 1, 2, 3, or 4 h. At each time point, cells were post-labeled with JF-646 Halo ligand to mark the Golgi area and Halo-Flipper, which was imaged by FLIM (Fig. 3 C). Over the spreading time course, Halo-Flipper fluorescence lifetimes at the Golgi increased, suggesting a progressive rise in Golgi membrane tension that parallels FA assembly and cell spreading (Fig. 3, C and D). Second, RPE1-ManII-Halo–expressing cells were seeded for 4 h on FN-coated PAA gels of different stiffness or on glass (Fig. 3 E). FLIM measurements showed reduced lifetimes on soft substrates compared with glass (Fig. 3, E and F), suggesting reduced Golgi membrane tension under low mechanical load. However, while gradual differences between soft and intermediate-stiffness substrates and between intermediate-stiffness substrates and glass were observed, they were more subtle and did not reach statistical significance in our analysis. Together, these data suggest that Golgi membrane mechanics is sensitive to extracellular mechanical cues, thereby linking adhesion-dependent cell states, cell spreading area, and post-Golgi carrier formation: robust spreading on FN-coated glass correlates with elevated Golgi tension and increased number of post-Golgi carriers, whereas limited adhesion on soft substrates corresponds to lower Golgi tension and reduced post-Golgi carrier abundance. These findings suggest that Golgi membrane mechanics responds to external mechanical forces, pointing to a potential mechanoadaptive regulation of Golgi function.

We next investigated how Golgi membranes respond to mechanical cues at the cell surface and what signaling pathways may relay forces to the Golgi apparatus. It is known that extracellular mechanical signals tune tubulin acetylation and GEF-H1/RhoA activation (Seetharaman et al., 2022; Krendel et al., 2002). GEF-H1 is a MT-associated RhoA exchange factor, which, upon MT acetylation, is released from MTs and is able to activate RhoA, providing a potential link between MT acetylation and downstream RhoA signaling (Krendel et al., 2002; Deb Roy et al., 2026). GEF-H1/RhoA activation can in turn lead to PLCε activation and DAG production at the TGN (Eisler et al., 2018). DAG is necessary for CARTS biogenesis by recruiting and activating PKD, a kinase required for the fission of Golgi-to-PM carriers such as CARTS (Wakana and Campelo, 2021). Based on this body of evidence, we hypothesized that MT acetylation could serve as one of the mediators transmitting mechanical signals from the cell surface to the Golgi membranes. Because astrocytes plated on stiffer substrates show higher levels of acetylated MTs (Seetharaman et al., 2022), we tested whether other mechanical stimuli—cell spreading on different ligand coatings and equibiaxial cell stretch—similarly induce MT acetylation.

First, to test whether MT acetylation can mediate mechanotransduction from FAs to the Golgi membranes, we compared MT acetylation levels in HeLa cells seeded on FN with those in cells seeded on PLL (Fig. 4 A). By 1 h on FN, acetylated tubulin appears to be mostly localized in the perinuclear area (Fig. 4 A). By 4 h, long protruding acetylated MTs extended from the perinuclear region toward the cell periphery (Fig. 4 A). This pattern was absent in PLL-plated cells at all time points (Fig. 4 A). Quantification of the acetylated-to-total tubulin ratio revealed a time-dependent increase on FN that plateaued at 4 h, whereas cells on PLL displayed no consistent trend, possibly reflecting large cell-to-cell variability and lack of integrin engagement on PLL substrates (Fig. 4 B). Because changes in MT acetylation can modulate GEF-H1/Rho signaling, we next asked whether this spreading-dependent increase in MT acetylation was accompanied by activation of the downstream Rho/ROCK-myosin II contractility axis. Rho activation leads to ROCK activation, which promotes myosin light chain (MLC) phosphorylation on Ser19 (Totsukawa et al., 2000). In our previous work (Miserey-Lenkei et al., 2010; Miserey-Lenkei et al., 2017), we showed that phosphorylated MLC (pMLC) associates with RAB6-positive carriers at the Golgi and that myosin II activity is required for the biogenesis of RAB6-positive post-Golgi carriers at the TGN membrane. We therefore assessed MLC phosphorylation at Ser19 as a reporter of myosin II activation downstream of Rho/ROCK signaling. Quantification at the basal cell–substrate interface revealed that pMLC levels increased at the earliest time point examined (1 h) after seeding on FN, but not on PLL, and then returned toward steady-state levels at 4 and 24 h (Fig. S3, A and B). In parallel, F-actin staining using phalloidin confirmed the formation of stress fibers, another hallmark of Rho/ROCK signaling, in cells seeded on FN, in contrast to cells seeded on PLL (Fig. S3 A). Although additional experiments will be needed to fully establish the significance of the transient increase in pMLC levels, these results are consistent with an enhanced Rho-dependent signaling at early stages of cell spreading on FN.

Next, to emulate externally applied stretch, we subjected HeLa cells to a 15% equibiaxial strain for 30 min using our stretching device (Fig. 2 A). External stretch suggested a trend toward higher tubulin acetylation relative to unstretched controls (Fig. 4 C), but the effect was heterogeneous across experiments and, overall, not statistically significant, indicating that additional factors may modulate this response (Fig. 4 D). This is consistent with recent work showing that cyclic mechanical stimulation can reorganize acetylated MT architecture without necessarily increasing total acetylated MT levels, highlighting that different mechanical cues may affect MT acetylation and organization in distinct ways (Li et al., 2023).

To determine whether MT acetylation alone is sufficient to increase the number of Golgi-derived carriers independently of external mechanical forces, we treated HeLa cells with tubacin, a histone deacetylase family member 6 (HDAC6) inhibitor that increases MT acetylation. As reported previously (Haggarty et al., 2003a; Haggarty et al., 2003b), tubacin elevated acetylated tubulin levels (Fig. S4, A and B) and significantly increased the number of PAUF-mRFP–positive carriers released from the TGN, without any significant increase in adhesion area (Fig. 5, A–D and Fig. S4 C). Tubacin treatment in RPE1-ManII-Halo–expressing cells did not lead to any significant increase in Golgi membrane tension as measured by Halo-Flipper FLIM (Fig. 5, E and F), suggesting that signaling pathways other than those downstream of tubulin acetylation may control Golgi membrane tension and function. Nevertheless, because HDAC6 has substrates beyond tubulin, we cannot exclude additional, indirect effects of tubacin on adhesion- or cytoskeleton-dependent signaling. Collectively, these results indicate that physiological spreading on FN enhances MT acetylation, whereas acute, sustained stretch produces only a variable trend in the same direction. In addition, pharmacological induction of MT hyperacetylation with tubacin is sufficient to promote Golgi export. Thus, MT acetylation emerges as an important component of the mechanotransduction pathway from the cell surface to the Golgi apparatus, while our data suggest that additional routes likely contribute to the regulation of Golgi membrane tension and function. Future experiments using hypo- and hyper-acetylated MT expression constructs (Joo and Yamada, 2014) may provide a more direct means of defining the precise role of MT acetylation and how this is integrated with parallel pathways to coordinate Golgi mechanoresponse and mechanosensitivity.

Having shown that mechanical cues and MT acetylation enhance CARTS formation, we next asked whether these inputs also regulate Golgi DAG levels and PKD activity, since DAG recruits and activates PKD to promote carrier biogenesis downstream of RhoA/PLCε signaling (Eisler et al., 2018). To monitor DAG at the Golgi area, HeLa cells were transfected with the GST-C1a-PKD DAG biosensor (Maeda et al., 2001) and analyzed by immunofluorescence microscopy. We note that, as with other lipid probes based on functional lipid-binding domains, this sensor has limitations, including the possibility that it may not fully capture transient DAG dynamics and could potentially interfere with endogenous protein–lipid interactions. First, we compared cells seeded on FN with cells seeded on PLL (Fig. 6 A). Cells on FN accumulated higher Golgi DAG levels than cells on PLL, based on the GST-C1a-PKD DAG biosensor (Fig. 6, A and B). In FN-seeded cells, Golgi DAG showed an increase at early time points, followed by a return toward steady-state levels at later time points. This may reflect rapid DAG turnover during transport carrier biogenesis, including incorporation into nascent carriers or conversion into other lipids involved in membrane fission (Fugmann et al., 2007; Campelo and Malhotra, 2012). By contrast, cells on PLL showed reduced levels of Golgi DAG at earlier time points as compared with 24 h. Second, we compared DAG levels at the Golgi area in cells on soft (2 kPa PAA gels) versus stiff (30 kPa PAA gels or glass) FN-coated substrates (Fig. 6, C and D). Consistent with our spreading data, and paralleling increased MT acetylation and Golgi-derived carrier biogenesis under these conditions, cells on stiff substrates tended to show higher DAG content at the Golgi area than cells on soft substrates, although only the difference between 2 kPa and glass was statistically significant (Fig. 6 D). Finally, to test whether MT acetylation alone modulates Golgi DAG levels, we treated HeLa cells with tubacin (Fig. 6, E and F). Whereas tubacin-induced MT hyperacetylation increased the number of cytoplasmic CARTS (Fig. 5B), it caused only a modest, nonsignificant increase in Golgi DAG content (Fig. 6, E and F). Consistently, Golgi membrane tension remained unchanged upon tubacin treatment (Fig. 5, E and F), suggesting that full DAG accumulation and membrane tension increase at the Golgi membranes likely require additional force-dependent activation of other DAG-synthesizing pathways (Romani et al., 2019).

To complement these measurements, we directly monitored PKD activity at the Golgi/TGN using a genetically encoded reporter (G-PKDrep, see Materials and methods [Fuchs et al., 2009; Eisler et al., 2018]). PKD activity was higher at early time points after cell seeding on FN (Fig. 7, A and B), consistent with the dynamics of DAG levels detected with the C1a-PKD–based biosensor (Fig. 6 B) as well as of pMLC signal (Fig. S3 B). Interestingly, higher Golgi PKD activity was also observed at early time points in cells seeded on PLL, indicating that the relationship between DAG/PKD signaling and Golgi export is more complex than captured by DAG measurements alone. While new tools capable of locally and quantitatively reporting lipid levels with sufficient sensitivity will be invaluable to validate and further dissect the dynamics of this lipid-mediated regulation of Golgi export, our findings suggest that DAG and PKD activity at the Golgi may be modulated by mechanical inputs, potentially as components of a mechanotransduction pathway linking cell surface mechanical forces to Golgi function.

Our data so far point to a critical feedback loop between the cell surface—particularly FAs—and the Golgi apparatus, through which extracellular mechanical cues reshape Golgi lipid composition, signaling activity, mechanics, and secretory output. We therefore tested whether this feedback loop is functionally required for efficient cell adhesion and spreading. To investigate this, we performed spreading assays in HeLa cells expressing the membrane marker GFP-mem, while selectively impairing Golgi export using different strategies. First, we treated cells with Golgicide A, a general inhibitor of Golgi-to-PM trafficking (Sáenz et al., 2009). Although initial cell spreading within the first hour was unaffected, a significant reduction in cell adhesion area was evident by 4 h in Golgicide A–treated cells (Fig. S5, A and B), suggesting that Golgi-derived export becomes an important contributor to continued spreading. We next inhibited PKD using CRT0066101 (Harikumar et al., 2010). PKD inhibition decreased the number of cytoplasmic CARTS (Fig. 8, A and B) (Sugiura et al., 2026, Preprint) and, similarly to Golgicide A treatment, led to a reduced expansion of the adhesion area at 4 h (Fig. 8, C and D). In parallel, FLIM measurements in RPE1-ManII-Halo cells showed that CRT0066101 treatment also reduced Golgi membrane tension, as reported by Flipper lifetime (Fig. 8, E and F). To begin addressing how PKD inhibition affects FA organization, we examined active β1 integrin and FA area in cells treated with CRT0066101 during spreading on FN (Fig. S5 C). PKD inhibition did not substantially change active β1 integrin levels at the basal cell surface at either 1 h or 4 h after seeding (Fig. S5 D), whereas FA area was reduced after 4 h (Fig. S5 E), in parallel with the decrease in overall cell spreading area (Fig. 8 D). These results suggest that inhibition of Golgi exocytic activity limits sustained cell spreading and is accompanied by changes in FA organization. Consistent with previous work showing that PKD2/PKD3 depletion does not significantly alter FA number per cell (Eisler et al., 2018), our data suggest that PKD-dependent Golgi export primarily supports sustained spreading downstream of adhesion-mediated signaling. This interpretation is also consistent with previous work showing that PKD activity supports directional migration and invasion in breast cancer models (Borges et al., 2015). Together, these findings indicate that disruption of PKD-dependent carrier biogenesis reduces both Golgi membrane tension and cell spreading efficiency. Under the same conditions, PKD inhibition did not significantly alter global MT acetylation levels after 4 h of spreading on FN (Fig. S5, F and G), consistent with the idea that MT acetylation acts upstream of PKD activation. This also suggests that any feedback from PKD-dependent CARTS biogenesis onto MT acetylation is limited, at least at the level of global acetylated tubulin detected in our assay.

Because PKD also regulates actin dynamics and transcription (Fu and Rubin, 2011), we took a more targeted approach by selectively impairing CARTS-mediated export. HeLa TGN46 KO cells, which lack the transmembrane cargo adaptor TGN46 that is required for cargo sorting into CARTS and show impaired CARTS-mediated export (Lujan et al., 2024), exhibited significantly reduced spreading compared with WT cells, a phenotype that persisted for at least 24 h after seeding (Fig. 8, G and H). Collectively, our results underscore that the Golgi apparatus actively responds and adapts to extracellular mechanical cues (Fig. 8 I). The ability of the Golgi apparatus to sense, respond, and secrete in response to extracellular mechanical forces supports proper cell adhesion, spreading, and homeostasis. Our work unveils new concepts in organelle mechanobiology, highlighting key mechanotransduction steps at the Golgi apparatus, a secretory mechanoresponse, and the importance of feedback communication between the PM and internal membranes in maintaining cellular homeostasis.

Our work reveals a mechanotransduction axis linking extracellular forces to the Golgi complex. These forces are transduced to remodel Golgi membranes by tuning their biophysical and biochemical characteristics and lead to an increase in the number of cytoplasmic Golgi-derived carriers (CARTS as well as CD59-positive carriers) to support adhesion and efficient spreading. We show that three distinct mechanical stimuli (spreading forces on integrin-activating substrates, substrate stiffness, and equibiaxial stretch) associate with an increase in the number of Golgi-derived carriers. In parallel, several of these cues are accompanied with changes in Golgi membrane tension (measured by Halo-Flipper FLIM) and Golgi DAG/PKD signaling (measured by a fluorescent DAG biosensor and a PKD activity reporter). We provide evidence that MT acetylation acts as a critical downstream effector of mechanical cues and is sufficient to boost post-Golgi carrier formation even in the absence of external forces. Finally, we show that Golgi export—particularly via PKD- and CARTS-dependent routes—is both upregulated by mechanical stimuli and required for sustained cell spreading, suggesting a reciprocal communication between FA establishment, cell spreading, and directed Golgi-to-PM trafficking to exocytic hotspots.

Ample evidence supports the idea that exocytosis in adherent cells is not uniformly distributed across the cell surface but occurs at secretion hotspots, typically juxtaposed to FAs (Stehbens et al., 2014; Eisler et al., 2018; Huet-Calderwood et al., 2017; Fourriere et al., 2019). FAs are molecular platforms required for adhesion and for sensing and transducing extracellular mechanical forces (Kechagia et al., 2019; Chastney et al., 2025). Integrin β1 itself is delivered from the TGN to the PM through a RAB6-mediated (Huet-Calderwood et al., 2023) and PKD-dependent pathway (Yeaman et al., 2004), directly linking Golgi export routes to the reinforcement of adhesion complexes. Notably, recent work further showed that a subset of newly synthesized α₅ integrin integrins can be delivered rapidly and locally to adhesion-proximal PM regions through a Golgi-bypass route (Lerche et al., 2026). Together, these findings highlight that biosynthetic integrin trafficking is spatially regulated and can contribute directly to adhesion dynamics. However, whether—and how—these mechanical cues drive adaptation of the secretory machinery has remained largely unexplored. Notably, recent studies show that ER exit sites (Farhan et al., 2025) and ER export can be upregulated upon extracellular mechanical cues by the GTPase Rac1 (Phuyal and Baschieri, 2020; Phuyal et al., 2022), and optogenetic and other mechanostimulatory perturbations alter ER export dynamics (Song et al., 2024; Chen et al., 2025). In addition, FAs and ECM can regulate the COPII machinery involved in ER to Golgi transport (Jung et al., 2022). Collectively, these reports point to coordinated, adhesion-dependent regulation across the secretory pathway. In line with this emerging view, a study published during revision of this manuscript showed that substrate stiffness regulates post-Golgi cargo sorting and secretion through a Src–FAK–AMPK–GBF1 pathway, further supporting the idea that the Golgi functions as a mechanically regulated hub within the secretory pathway (Serafino et al., 2026). Our findings therefore position the Golgi as an active mechanoresponsive organelle, extending mechanotransduction beyond conventional mechanosensor organelles, such as the PM and nucleus.

Our data place MT acetylation downstream of extracellular mechanical inputs and upstream of Golgi functional adaptation, consistent with its role as a tension-sensitive mediator (Seetharaman et al., 2022) and a regulator of PKD-dependent Golgi trafficking (Eisler et al., 2018). By showing that inhibiting tubulin deacetylation (using the HDAC6 inhibitor tubacin) partially phenocopies mechanical stimulation, we identify MT acetylation as both a sensor and an amplifier of force signals, upstream to the GEF-H1/RhoA/PLCε/DAG/PKD axis previously implicated in Golgi export (Eisler et al., 2018). We propose a model (Fig. 8 I) in which mechanical forces—transduced via a yet-unidentified mechanoreceptor—may trigger increased MT acetylation, release of GEF-H1, and RhoA signaling (Seetharaman et al., 2022). Active RhoA reinforces actomyosin contractility through ROCK and also promotes DAG production at the TGN via PLCε, recruiting and activating PKD to drive carrier biogenesis (Eisler et al., 2018) (Fig. 8 I). Direct perturbation of GEF-H1 or α-tubulin acetylatransferase 1, as well as the use of hypo- and hyper-acetylated tubulin constructs, will be important to further dissect the role of MT acetylation in this pathway. Too low Golgi membrane tension may render the TGN inefficient for carrier fission or alternatively may be a consequence of an altered TGN lipidome, so the full extent of this mechanical adaptation warrants further study. Of note, elevated membrane tension is known to work against membrane curvature generation (Dai and Sheetz, 1995; Carlsson, 2018; Le Roux et al., 2021), highlighting the importance of a well-balanced, dynamic control of Golgi membrane tension for secretory function. Importantly, our observations can be placed in the context of the opposing effects that membrane tension plays on endocytic versus exocytic pathways. High PM tension is well-documented to inhibit endocytosis (Apodaca, 2002; Kosmalska et al., 2015; Wu et al., 2017), a block that can be overcome locally by polymerization of actin to provide the force required for vesicle invagination and scission (Boulant et al., 2011; Gauthier et al., 2011). By analogy, actomyosin contractility may not be purely inhibitory but also contribute to promote the biogenesis and fission of post-Golgi carriers (Miserey-Lenkei et al., 2010) by shaping Golgi membrane mechanics and promoting DAG/PKD-dependent scission; thus, actin dynamics can act as a context-dependent permissive or driving force for membrane trafficking depending on the compartment and the direction of membrane flow.

Notably, we found that tubacin-induced MT hyperacetylation induced only a modest increase in Golgi DAG and no measurable change in Golgi tension, suggesting that additional regulatory layers and other effectors, such as Lipin-1 (Romani et al., 2019), may contribute to full DAG accumulation, PKD activation, and/or tension modulation. We note, however, that Flipper probes report membrane lipid packing/ordering as a proxy for membrane tension, and perturbations that alter Golgi lipid composition (e.g., PKD inhibition or DAG modulation) can alter Flipper lifetimes independently of tension reporters (Colom et al., 2018). Thus, while our FLIM data are consistent with tension changes, interpretation of lifetime shifts as pure mechanical readouts should be cautious and will require complementary validation in future work.

Taken together, our data suggest that early cell spreading on FN triggers a transient Rho/ROCK-dependent program that couples actomyosin contractility to Golgi secretory activity, as reflected by the early increase in pMLC, DAG, and PKD activity and consistent with a role in rapid membrane remodeling and Golgi-derived carrier biogenesis. Previous work has shown that human fibroblasts display a faster and stronger increase in pMLC levels during spreading on FN as compared with PLL (Ren et al., 2004), consistent with differential activation of contractility pathways under these conditions. Using more specific activity reporters for Rho or GEF-H1, as well as tools such as optoTAT (Deb Roy et al., 2026), an optogenetic technology that allows for local induction of MT acetylation, will be instrumental to fully dissect the precise molecular links between MT acetylation, RhoA activation, Golgi membrane tension, and PKD activation—and, more broadly, the Golgi machinery that tunes secretory output (Fig. 8 I). Nonetheless, by uncovering these connections, our work establishes Golgi-based mechanoadaptation as a key mechanism of cellular regulation and opens important avenues for future investigation.

Reciprocal feedback between exocytosis and membrane tension also provides a physiological rationale for the coupling we observed: exocytosis lowers PM tension by adding membrane facilitating further spreading and migration (Gauthier et al., 2011; Pontes et al., 2017). Thus, upregulation of Golgi-to-PM trafficking and secretion at adhesion-proximal hotspots can both respond to and modulate PM mechanics, promoting processes such as migration, tissue morphogenesis, and membrane homeostasis. In pathological contexts, such as cancer invasion and fibrosis, altered ECM mechanics could dysregulate this Golgi mechanoresponse, driving aberrant secretion of matrix components or signaling factors. Notably, delivery of newly synthesized integrin β1 relies on RAB6- and PKD-dependent Golgi export (Huet-Calderwood et al., 2017; Yeaman et al., 2004), suggesting that misregulation of this pathway could directly impact adhesion turnover and invasive behavior. Targeting elements of this axis (e.g., HDAC6, PLCε, PKD, TGN46, or other intermediates yet to be identified) may offer novel therapeutic strategies to modulate mechanoresponsive secretion. Future experiments using micropatterned substrates of defined size and geometry (Théry, 2010), ideally combined with substrates of tunable stiffness, will be important to separate the relative contributions of cell spreading area, FA organization, and mechanical load to Golgi carrier biogenesis and membrane tension regulation. Finally, extending these studies to polarized cells, 3D matrices, and in vivo models will be essential to understand Golgi-mediated mechanoregulation at the tissue level.

In summary, we uncovered a two-way dialog between the PM and the Golgi apparatus that is mediated by mechanical forces, MT acetylation, and TGN signaling (DAG and PKD). This feedback loop may allow cells to adapt their adhesive and migratory behavior to the mechanical properties of their environment, thus positioning the Golgi as a new mode of cellular mechanoregulation.

The Paxillin-pEGFP plasmid was a gift from Rick Horwitz (University of Virginia, Charlottesville, VA, USA; #15233; Addgene plasmid) (Laukaitis et al., 2001). The mKate2-FM4-PAUF plasmid was described earlier (Wakana et al., 2021). The double palmitoylated euromodulin (N-terminal)-GFP (GFP-mem) plasmid is used as a general PM marker as it is a fusion protein with a signal for posttranslational palmitoylation targeting the fusion protein to the cell membrane and was a gift from Francesc Tebar’s lab (Universitat de Barcelona, Spain) (Vidal-Quadras et al., 2011). The plasmids encoding for PAUF-mRFP (Wakana et al., 2012) and GST-C1a-PKD (Maeda et al., 2001) were donated by Vivek Malhotra (CRG, Barcelona, Spain), and the plasmid encoding for PAUF-MycHis (Kim et al., 2009) by S. S. Koh (Korea Research Institute of Bioscience and Biotechnology, Daejeon, Korea). Commercial antibodies used in this study were as follows: sheep anti-TGN46 (AHP500GT; Bio-Rad); mouse anti-GM130 (610822; BD Biosciences); mouse anti-paxillin (610568; BD Biosciences); mouse anti-paxillin, clone 349 (610051; BD Transduction Laboratories); rat anti-active β1 integrin, clone 9EG7 (553715; BD Biosciences); rabbit anti-α-tubulin (ab18251; Abcam); rabbit anti-tRFP (AB233; Evrogen); mouse anti-acetylated α-tubulin (T6793; Sigma-Aldrich); rabbit anti-glutathione-S-transferase (GST) (G7781; Sigma-Aldrich); rabbit anti-ERC1/ELKS (22211-1-AP; Proteintech); mouse anti-α-tubulin, clone B-5-1-2 (T6074; Sigma-Aldrich); mouse anti-Myc, clone 9E10 (sc-40; Santa Cruz Biotechnology); rabbit anti-Myc (2272; Cell Signaling Technology); and mouse anti-pMLC Ser19 (3675; Cell Signaling Technology). The pSer294-specific rabbit polyclonal antibody used for detection of G-PKDrep phosphorylation was kindly provided by Angelika Hausser (University of Stuttgart, Germany) and previously described (Hausser et al., 2005; Fuchs et al., 2009). Phalloidin-Alexa Fluor 647 was from Invitrogen and used at a 533-fold dilution from the stock (0.00375 U/μl final concentration) for 1 h. Secondary antibodies used were Alexa Fluor 488 (A32766)-, Alexa Fluor 555 (A32794)-, or Alexa Fluor 647 (A21448)-coupled donkey anti-mouse, anti-rabbit, and anti-sheep immunoglobulin G, as well as goat anti-mouse Alexa Fluor 488 (A11029), anti-rabbit Alexa Fluor 488 (A11034), and anti-mouse Alexa Fluor 594 (A11032) (all purchased from Invitrogen). Janelia Fluor probes (JF-646) were from Promega. The following reagents were purchased from Sigma-Aldrich: Fibronectin (11051407001), PLL (P1524), Tubacin (SML0065), Golgicide A (345862), and CHX (C4859). D/D solubilizer (635054) was obtained from Takara Bio. Fibronectin (354008) was also obtained from Corning. CRT0066101 was from Tocris Bioscience (4975). SYLGARD 184 Elastomer KIT was from Dow Inc. (24001673921).

HeLa cells were cultured in DMEM (Capricorn Scientific GmbH) supplemented with 10% FBS (Invitrogen), 1% penicillin-streptomycin (Gibco), and L-Glutamine (Labclinics) under 5% CO₂ at 37°C. HeLa Cells were transiently transfected using X-tremeGENE 9 DNA transfection reagent (Sigma-Aldrich) following the manufacturer’s instructions or with calcium phosphate (Fourriere et al., 2019). Cells were used for the designed experiments ∼16 h after transfection. In experiments where cells were placed at 20°C for 2 h to synchronize cargo release from the Golgi apparatus, cell culture medium was supplemented with HEPES 25 mM (Sigma-Aldrich). Additionally, cell culture medium was supplemented with CHX 100 µM to inhibit new protein synthesis before cargo release. TGN46 KO HeLa cells were generated as described previously (Lujan et al., 2024).

RPE1 cells were cultured in DMEM/F12 supplemented with 10% FBS under 5% CO₂ at 37°C. RPE1 cells were transfected using Lipofectamine 3000 (Invitrogen) following the manufacturer’s instructions. RPE-ManII-Halo stable cell line was generated as follows: RPE1 cells were transfected with pIRES-ManII-Halo (ManII-Halo sequence from prHom-ManII-Halo, kind gift from D. Toomre (Yale School of Medicine, New Haven, CT, USA), was inserted into pIRES-Neo3), and single clones were then selected following ampicillin selection. HeLa Flp-IN T-Rex GPKDrep cell line (Hela-PKDrep) (Fuchs et al., 2009; Eisler et al., 2018) was kindly provided by Angelika Hausser (University of Stuttgart, Germany). HeLa cells stably expressing PAUF-MycHis were generated and described by Sugiura et al. (2026, Preprint).

HeLa cells stably expressing PAUF-MycHis were transfected with control siRNA or an siRNA oligo targeting ELKS. The targeting sequences of siRNA were as follows: Control (GL2 luciferase): 5′-AAC​GTA​CGC​GGA​ATA​CTT​CGA-3′; ELKS: 5′-AAG​GAA​GTA​TTA​AGA​GAA​AAT-3′. For immunofluorescence, at 48 h after siRNA transfection, the cells were trypsinized and plated on FN-coated coverslips; 24 h later, the cells were fixed with 4% PFA and analyzed by fluorescence microscopy. For testing knockdown efficiency, at 72 h after siRNA transfection, the cells were lysed with 0.5% SDS and 0.025 U/μl benzonase nuclease (Sigma-Aldrich) in PBS. The cell lysates were analyzed by western blotting with anti-ELKS and anti–a-tubulin antibodies.

For MT acetylation and DAG production experiments, cells were pre-seeded for 1 h on FN-coated coverslips (10 µg/ml final concentration, overnight [o/n] incubation) and then treated with 10 µM Tubacin or DMSO for 3 h continuously. For CARTS formation experiments, cells pre-seeded on FN-coated coverslips were subjected to a pre-treatment of 30 min with Tubacin or DMSO, followed by a 2-h incubation at 20°C in presence of Tubacin or DMSO, and a 30-min CARTS release at 37°C with Tubacin or DMSO still present in the samples. For PKD inhibition experiments, cells were pre-seeded over FN-coated coverslips for 30 min and then incubated with CRT0066101 (5 µM) or DMSO for 30 min or 210 min. For actin depolymerization, cells were labeled with Halo flipper and JF-646 as explained and then further treated with 50 nM latrunculin-A for 10 min and immediately imaged for the following 10–15 min. Imaging was discontinued as soon as Golgi fragmentation was observed in the cells, which was normally occurring at 25–30 min after treatment. For Golgi export inhibition experiments, 30-min pre-seeded cells were treated with Golgicide A (10 µM) for 30 min or 210 min. The mKate2-FM4-PAUF construct contains the FM4 aggregation domain, retaining the cargoes in the ER after synthesis. By adding to the cell medium D/D solubilizer, a small molecule that solubilizes FM4-induced aggregates, it is possible to synchronize a wave of cargo release from the ER and follow their intracellular localization in time, all the way from the ER to the Golgi apparatus and to CARTS for secretion. For those cells transiently transfected with mKate2-FM4-PAUF plasmid, D/D solubilizer was added at 1 µM final concentration to allow the cargoes to be exported out of the ER.

For cell spreading assays (e.g., Fig. 1, A and B), HeLa cells transiently transfected with GFP-mem and mKate2-FM4-PAUF or RPE1-ManII-Halo stable cells were split and seeded over a ligand-coated coverslip (FN or PLL). Unless otherwise stated, cells were lifted by incubation with trypsin (0.05% trypsin, 0.53 mM EDTA) for 3–5 min in a 37°C incubator. When lifted using EDTA, cells were incubated with 10 mM EDTA in Ca²⁺- and Mg²⁺-free PBS for 10 min in a 37°C incubator. Cells were kept at 37°C, and 30 min before the end of the spreading assay, cell culture medium was supplemented with CHX (100 µM) and D/D solubilizer (1 µM) to allow cargoes to be released out of the ER. Cells were then fixed and visualized by fluorescence microscopy.

Cells were fixed with 4% (vol/vol) PFA in PBS for 15 min at RT. After three washes with PBS, cells were permeabilized with 0.2% Triton X-100 (TX-100) in PBS for 30 min at RT. After washing with PBS, blocking was performed with 4% BSA for 30 min at RT or o/n at 4°C. After washing the samples, the BSA solution at 2% was used to dilute the primary and secondary antibodies for incubation 1 h at RT. Samples were then mounted on glass slides using either ProLong Gold Antifade Reagent (Thermo Fisher Scientific) or Mowiol. Widefield microscopy images were acquired on an upright Nikon Eclipse Ni-U microscope equipped with a 60× water-dipping objective (NIR Apo 60×/WD 2.8, Nikon), an ORCA-Flash 4.0 camera (Hamamatsu), and controlled by Metamorph software. Confocal microscopy images were acquired on a TCS SP8 microscope (Leica Microsystems GmbH, Germany) equipped with an HC PL APO CS2 100×/1.40 oil objective, a pulsed white light laser operating at 20 MHz repetition rate, 70% master power and hybrid detectors in photon-counting mode and 8-bit depth, whereas spinning disk confocal microscopy images were acquired on a Nikon Inverted Eclipse Ti-E (Nikon) microscope equipped with a Spinning disk CSU-X1 (Yokogawa), an iXon EMCCD camera (Andor), or sCMOS Kinetx 22 camera (Photometrics) integrated in Metamorph software by Gataca Systems, using 60× or 100× CFI plan apochromat VC 1.4 NA oil immersion objectives (Nikon). The spinning disk microscope was in a cage incubator from Life Imaging Services for temperature control. GFP, mKate2, RFP, Alexa488, and Alexa647 were excited using 489, 589, 554, 499, or 653 nm lines, respectively, and the emission was detected between 499 and 580, 599–770, 564–750, 509–600, or 663–781 nm, respectively. For dual and triple-color imaging, sequential scan mode between lines was used to minimize cross talk and mechanical drift. For z-stack acquisitions, the distance between confocal planes was set as system optimized by the microscope acquisition software, the planes were reaching across the entire cell volume, and maximum intensity (for visual representation) or sum slices (for total volumetric intensity measurements) projections were computed. Images were typically acquired with 2× line accumulation. Laser power was adjusted to avoid pixel saturation and then kept constant throughout each experiment. Image analysis was performed using ImageJ software.

The assay was firstly described by Fourriere et al. (2019). Briefly, coverslips were incubated with bicarbonate buffer 100 mM for 1 h at 37°C. Then, coverslips were washed three times in PBS and dried before incubation with the anti-tRFP antibody for 3 h at 37°C or o/n at 4°C, using a 1:250 dilution from the stock concentration. For the detection of the coated antibody to validate uniformity of coverslip coating, an anti-rabbit Star Orange secondary antibody (Abberior) was used.

Images were acquired with a commercial Nikon Eclipse Ti System, equipped with a 100× oil objective with NA 1.49 using TIRF illumination and enclosed in an incubator chamber set to 37°C and 5% CO₂. The detection was carried out using an ANDOR technology EMCCD iXon 897 camera. The equipment presents an Agilent technologies laser box with wavelengths of 488 and 560 nm. Laser power was adjusted for each channel to prevent saturation. For the experiments tracking post-Golgi carriers transport to FAs, cells were seeded on FN-coated 35-mm glass-bottom dishes (MatTek), and cells were subjected to a cargo release synchronization before imaging. Before cargo channel acquisition, an image of the FAs channel was taken as a reference. Then, cargo channel was acquired live at 1 fps for 45 min. For SPI live-cell imaging, both channels were acquired quasi simultaneously every 30 s for 1 h. All live-cell imaging was acquired at 37°C.

The RUSH assay was performed as described by Boncompain et al. (2012).

To measure PKD activity, we used a HeLa Flp-IN T-Rex GPKDrep cell line (Fuchs et al., 2009; Eisler et al., 2018) kindly provided by Angelika Hausser (University of Stuttgart, Germany), cultured under selection with 500 µg/ml hygromycin B and 10 µg/ml blasticidin. Prior to the experiment, antibiotics were removed, and cells were induced using 10 ng/ml for 24 h. The day after, cells in suspension were seeded on FN-coated coverslips (10 µg/ml) for the indicated times and processed for immunofluorescence. Image analysis pipeline consisted of segmenting the Golgi regions using the EGFP mask and measuring the total intensity of pS294 signal and EGFP signal for normalization. The ratio between pS294 and EGFP is indicative of Golgi PKD activity.

A FLIM module from PicoQuant (Kit LSM Upgrade pour Nikon A1 PicoQuant, Opton Laser International) was added to the A1R Nikon confocal microscope with both resonant and galvano mode scanners. Images were acquired at 37°C and by using 485 nm laser excitation and a 27.5-ns dwell time per pixel with 5-frame acquisition. The emission bandpass filter wavelength was 506–619 nm. Single-stack images were captured and used for FLIM using Time-Correlated Single Photon Counting. Images were analyzed using the SymPhoTime 64 software (PicoQuant, Opton Laser International). The image acquisition time was typically 48 s for a 512 × 512 image. A region of interest was drawn manually based on a Golgi mask (Janelia Fluor 646 HaloTag channel, see below for details) using the Screen Dragon software. The photon arrival time histograms were fitted with a double exponential model, after deconvolution with the calculated impulse response function in SymPhoTime. Upon fitting, the intensity weighted average fluorescence lifetime was computed and used for subsequent analysis. All FLIM measurements were performed in living cells at 37°C.

ManII-Halo RPE1 cells were labeled with Halo-Flipper to report lipid packing defects in Golgi membranes due to changes in membrane tension or lipid composition. Cells were plated on Fluorodishes with glass or PAA gel substrates from 30 min to 4 h (depending on the experiment). Halo-Flipper (kindly provided by S. Matile Lab, University of Geneva, Switzerland) of 100 nM concentration was prepared in 200 μl serum-free medium and added to the cells by replacing the old medium. After a 15-min incubation, Halo-Flipper was replaced with 160 μl complete medium (containing 10% FBS) with 40 μl of 1 µM Janelia Fluor 646 HaloTag (JF-646, 200 nM final working concentration) and further incubated for 15 min, before FLIM and fluorescence images of the Golgi membrane were acquired. Halo-Flippers were gently imaged avoiding as much as possible light exposure of the samples before FLIM recording. To minimize possible Flipper-induced phototoxicity and singlet oxygen photosensitization (Torra et al., 2024), individual cells were measured only once. Halo-Flippers were made commercially available in 2025 and are now supplied by Spirochrome (Halo-Flipper, Spirochrome).

Janelia Fluor 646 HaloTag (JF-646) was used at 0.5 µM (1:2,000 dilution from the 1 mM stock) in culture media. Cells were labeled by incubation for 15 min at 37°C, followed by extensive washing.

PAA gels were prepared as described by Pérez-González et al. (2021).

The detailed protocol is described by Le Roux et al. (2025). Briefly, the PDMS mixture was prepared in a 1:10 ratio and degassed in a vacuum chamber for 1 h. PDMS was spread over PMMA plates using a spin coater in two steps: a first step during 5 s at 500 rpm with a 100 rpm/s acceleration, and a second step during 1 min at 500 rpm with a 300 rpm/s acceleration. PDMS was polymerized o/n at 65°C in an oven. PDMS membranes were peeled off from the PMMA plates and mounted on the stretch-rings. PDMS rings were sterilized by exposing them to UV light for 15 min and then were coated with FN o/n. The following day, FN excess was rinsed off, and cells were seeded on top of the FN-coated PDMS membranes for 1 h prior the experiment. For experiments carried out with non-transiently transfected cells, cells were subjected to 15% substrate stretch for 30 min at 37°C before fixation. In experiments characterizing the formation of CARTS upon mechanical strain stimulation, an intermediate step at 20°C for 2 h to synchronize cargo release was set before stretching. Control experiments were performed under the same conditions but in the absence of mechanical strain. Cells subjected to mechanical forces were fixed and imaged under stretch to prevent visual aberrations.

Analysis of the cell area was performed using a custom-designed macro in ImageJ. In brief, for each image cell area was determined by thresholding the GFP-mem channel, converting it to a mask and creating the selection. To determine number of CARTS, the ImageJ “Detect particles (ComDet)” plugin developed by Eugene A. Katrukha (Utrecht University, the Netherlands) was used, avoiding the perinuclear Golgi-positive area, using a maximum intensity z-projection of the entire confocal fluorescence microscopy z-stack of the cell or the epifluorescence image of the cell (when imaging the substrate stretching experiments). The normalized total intensity levels of acetylated tubulin were calculated according to the equation: Ratio of acetylated/total tubulin level = Acetylated tubulin intensity/Total tubulin intensity. DAG production (area and intensity) and Golgi area were calculated by manually drawing an ROI around the fluorescent-positive mass present in the perinuclear area of the cells. In the bulk SPI analysis, FAs and secreted cargo objects were segmented and the localization and fluorescence intensity of each pixel was calculated. The nearest neighbor distance between each pixel of secreted cargo to the closest pixel of FAs was measured using MATLAB. The intensity of each pixel of secreted cargo was also included to obtain the 2D plots. CD59-positive vesicle count analysis was carried out using the cells expressing RUSH cargo of interest, which were fixed at 45 min post-biotin addition to record post-Golgi carriers. The vesicles/carriers were manually counted from maximum intensity z-projection of the entire z-stack confocal fluorescence microscopy image of the cell using ImageJ and used for further analysis. To quantify integrin signal at the basal membrane, a single plane with maximum basal signal was selected from a z-stack and used for downstream analyses. A total cell mask was obtained from a sum projection high contrast and used to analyze total intensity of the active integrin signal (Integrated Density). To measure FA area per cell, a total cell mask was obtained from a sum projection with high contrast. A binary image from the FA marker was used to define FA area in the total cell mask. To quantify pMLC signal at basal cell–substrate interface, a single plane with maximum basal signal was selected from a z-stack and used for downstream analyses. A total cell mask was obtained from phalloidin staining and used to analyze total intensity of pMLC signal (Integrated Density).

Data were visualized and represented graphically using SuperPlots (Lord et al., 2020), as indicated in the corresponding figure panels. In brief, each small, light-colored symbol represents one individual cell measurement (N individual measurements per replicate), whereas the larger, black-outlined colored symbols represent the mean of all cells measured within one independent biological replicate (n replicates). Different colors denote distinct biological replicates. Horizontal black lines and error bars represent the mean and standard error of the mean (SEM), respectively, across biological replicates, thereby capturing variability between independent experiments. Individual cell measurements (N) are displayed to illustrate within-experiment variability but were not treated as independent n (number of biological replicates) for statistical testing. Statistical analyses were performed in GraphPad Prism (version 11.0). P values are reported as exact numerical values, and the statistical tests used are indicated in the figure legends. All quantification data underlying SuperPlot-based figures are available in Zenodo at Campelo (2026) https://doi.org/10.5281/zenodo.21073559. The repository includes GraphPad Prism project files used for plotting and statistical analysis.

Fig. S1 shows experiments related to spatial exocytosis of CARTS close to FAs. Fig. S2 is related to Fig. 1 and shows additional results indicating that adhesion-dependent cell states associate with CARTS formation. Fig. S3 shows experiments indicating that early cell spreading on FN transiently increases pMLC levels. Fig. S4 is related to Fig. 5 and shows additional experiments and information on MT hyperacetylation upon tubacin treatment. Fig. S5 is related to Fig. 8 and provides additional experiments on how Golgi-derived export is necessary for cell spreading and mechanoadaptation. Video 1 shows a live-cell TIRF microscopy time-lapse movie of CARTS dynamics related to Fig. S1 A.

The data underlying all figures are available in the published article and its online supplemental material. All quantification data underlying SuperPlot-based figures are available in Zenodo at Campelo (2026) https://doi.org/10.5281/zenodo.21073559. The repository includes GraphPad Prism project files used for plotting and statistical analysis.

We thank members of the Single Molecule Biophotonics lab at Institut de Ciències Fotòniques (ICFO), the Intracellular Transport: Engineering and Mechanisms Team at Institut Curie, the Intracellular Mechanics group and the Physics of Living Systems team at MSC, as well as Neus Sanfeliu-Cerdán, Santosh Phuyal, Hesso Farhan, Aurélien Roux, Gaëlle Boncompain, Pierre Sens, and Vivek Malhotra for valuable discussions. We thank Merche Rivas, Marina Perez, Angel Sandoval, and Maria Marsal for technical support at ICFO. We thank Angelika Hausser (University of Stuttgart, Germany) for kindly sharing the HeLa Flp-IN T-Rex GPKDrep cell line and the phospho-serine 294 (pS294)-specific antibody. We thank Rick Horwitz, Francesc Tebar, Vivek Malhotra, and Stefan Matile for kindly sharing reagents.

We acknowledge support from the Government of Spain (RYC-2017-22227, PID2020-113068RB-I00/10.13039/501100011033, and PID2023-147711NB-100; PID2022-138282NB-I00 project funded by the Ministerio de Ciencia e Innovación (MCIN)/Agencia Estatal de Investigación (AEI)/10.13039/501100011033/FEDER, UE; PID2022-142672NB-I00; “Unidad de Excelencia María de Maeztu” CEX2024-001431-M, funded by MICIU/AEI/10.13039/501100011033 to MELIS-UPF, and Severo Ochoa CEX2019-000910-S toICFO and CEX2023-001282-S to Institute for Bioengineering of Catalonia, Fundació Privada Cellex, Fundació Privada Mir-Puig, and Generalitat de Catalunya (CERCA, AGAUR), ERC Advanced Grants NANO-MEMEC (GA 788546) and MechanoSynth (GA 101097753), as well as LaserLab 4 Europe (GA 654148). E. Almacellas acknowledges the support of the Beatriu de Pinós postdoctoral fellowship program of the Department of Research and Universities of the Generalitat de Catalunya (2023 BP 00210). N. Mateos acknowledges funding from the European Union H2020 under Marie Sklodowska-Curie grant 754558-PREBIST. J. Angulo-Capel acknowledges funding from the European Union H2020 under the Marie Sklodowska-Curie grant agreement No 847517. A. Wolowczyk was supported by joint funding from an ICFO Student Research Fellowship (Fall 2024) and a grant from Homerton College, Cambridge, UK. Y. Wakana acknowledges support from Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (grant number 25K09568); and AMED Multidisciplinary Frontier Brain and Neuroscience Discoveries (Brain/MINDS 2.0) (grant number JP24wm0625506). P. Roca-Cusachs acknowledges support from the prize “ICREA Academia” for excellence in research. This work was supported by Centre National de la Recherche Scientifique, the Labex Cell(n)Scale -  grants ANR-11-LABX-0038, ANR-10-IDEX-0001-02, the French Agence Nationale de la Recherche (ANR), grant number ANR-22-CE13-0044 “MECHANGOLGI” project. C. Bhaskar Naidu was supported by the international EuReCa PhD program of Institut Curie, H2020-MSCA-COFUND-2018-EuReCa-Grant Agreement number: 847718. The authors greatly acknowledge the Cell and Tissue Imaging (PICT-IBiSA), Institut Curie, member of the national infrastructure France-BioImaging (https://ror.org/01y7vt929) supported by the French National Research Agency (ANR-24-INBS-0005 FBI BIOGEN). The authors also thank the recombinant antibody platform of the Institut Curie. Open Access funding provided by the Universitat Pompeu Fabra.

Author contributions: Chandini Bhaskar Naidu: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, resources, visualization, and writing—original draft. Javier Vera Lillo: conceptualization, data curation, formal analysis, investigation, methodology, project administration, supervision, validation, visualization, and writing—original draft, review, and editing. Eugenia Almacellas: conceptualization, formal analysis, investigation, methodology, validation, visualization, and writing—review and editing. Anabel-Lise Le Roux: investigation, methodology, resources, and writing—review and editing. Sabine Bardin: methodology. Nicolas Mateos: formal analysis and software. Jessica Angulo-Capel: formal analysis. Adam Wolowczyk: formal analysis, investigation, validation, visualization, and writing—review and editing. Yugo Terashima: investigation. Yuichi Wakana: investigation, resources, supervision, and visualization. Pere Roca-Cusachs: funding acquisition, methodology, resources, and supervision. Maria F. Garcia-Parajo: funding acquisition and supervision. Franck Perez: conceptualization, funding acquisition, resources, supervision, and writing—review and editing. Bruno Goud: conceptualization, funding acquisition, and writing—review and editing. Jean-Baptiste Manneville: conceptualization, data curation, formal analysis, funding acquisition, methodology, project administration, resources, software, supervision, validation, and writing—review and editing. Stéphanie Miserey: conceptualization, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing. Felix Campelo: conceptualization, data curation, formal analysis, funding acquisition, methodology, project administration, resources, software, supervision, validation, visualization, and writing—original draft, review, and editing.