RACK1 plays a critical role in mast cell secretion and Ca²⁺ mobilization by modulating F-actin dynamics

Allergic reactions and associated inflammatory processes are triggered by multivalent antigens crosslinking antigen-specific IgE previously bound to the high affinity IgE receptor (FcεRI) on the surface of mast cells (MCs) and basophils (da Silva et al., 2014; Galli et al., 2020; Karasuyama et al., 2018). Following FcεRI crosslinking, MCs immediately degranulate and release preformed mediators stored in secretory granules (SGs). This is followed by the release of newly formed lipid mediators, and 8 to 24 h later, newly synthesized mediators (cytokines/chemokines) are released (Blank et al., 2014; Mukai et al., 2018; Wernersson and Pejler, 2014).

Crosslinking FcεRI activates a signaling cascade that is accompanied by membrane–cytoskeleton rearrangement (Dráber et al., 2012; Wollman and Meyer, 2012) and changes in intracellular Ca²⁺ concentration ([Ca²⁺]_i) (Hamawy et al., 1995; Siraganian et al., 2010). In the absence of stimulation, the cortical actin filaments (F-actin) act as a barrier preventing access of SGs to the plasma membrane, while stimulation induces disassembly of the cortical F-actin allowing for degranulation (Deng et al., 2009; Nishida et al., 2005; Wilson et al., 2016). Additionally, FcεRI crosslinking induces Ca²⁺ release from the lumen of the rough endoplasmic reticulum (ER) (Hamawy et al., 1995; Siraganian et al., 2010). Subsequently, the reduction in ER Ca²⁺ concentration ([Ca²⁺]_ER) triggers Ca²⁺ influx across the plasma membrane through Ca²⁺ channels, primarily the Ca²⁺ release-activated channel (CRAC) Orai1 (Ambudkar et al., 2017; Baba and Kurosaki, 2008; Vig et al., 2008). This store-operated Ca²⁺ entry (SOCE)-dependent increase in cytosolic [Ca²⁺] accompanies the actin cytoskeleton reorganization and facilitates the fusion of SGs to the plasma membrane (Cohen et al., 2012; Wollman and Meyer, 2012).

The molecular components involved in regulated secretion have not yet been fully elucidated (Bugajev et al., 2020; Klein and Sagi-Eisenberg, 2019). The receptor for activated C kinase 1 (RACK1) acts as a hub that facilitates assembly of key signaling complexes (Adams et al., 2011; Ron et al., 2013) and is known to play a role in immune cell activation (Corsini et al., 2021). RACK1 is a member of the tryptophan-aspartate repeat (WD repeat) family of proteins and has a seven-bladed β-propeller structure, which permits the binding of multiple partners (Ron et al., 1994). The ability of RACK1 to integrate various signaling pathways is important in immune cell response, since this process requires precise regulation of signaling between various receptors, changes in [Ca²⁺]_i, and cytoskeleton rearrangement (Ballek et al., 2016; Negoro et al., 2014; Ron et al., 2013; Yao et al., 2014). Modifications in the expression level of RACK1 have been associated with abnormal release of inflammatory mediators by immune cells (Corsini et al., 2021, 2015, 2014; Yao et al., 2014). Additionally, the gene RACK1 (also known as GNB2L1), which encodes RACK1, appears to be involved in asthma (Hwang et al., 2008). In asthmatic patients, aberrant Ca²⁺ influx compromises the function of T regulatory cells which can be correlated with a high level of RACK1 (Negoro et al., 2014) and RACK1 expression is also increased in lungs of mice with ovalbumin (OVA)-induced asthma (Pu et al., 2018). It is known that acute asthma is triggered by MC activation (Bradding and Arthur, 2016). Although RACK1 is a component of MC lipid rafts (Freitas Filho et al., 2019), its function in MCs has not been previously investigated. In the present study, the involvement of RACK1 in MC mediator release is shown. These findings reveal previously undescribed roles for RACK1 in actin cytoskeleton dynamics as well as in Ca²⁺ signaling during regulated secretion.

Initial experiments determined the presence and distribution of RACK1 in MCs. By immunofluorescence, RACK1 was identified in resident MCs in lung from a mouse with OVA-induced asthma, which increases lung MCs (Fig. 1A; Fig. S1), mouse bone marrow-derived mast cells (BMMCs) (Fig. 1B), the human MC line ROSA (KIT D816V) (Fig. 1C), and the rat MC line RBL-2H3 (Fig. 1D). In lung, RACK1 co-localized with the MC specific marker mouse MC tryptase 6 (Pejler et al., 2007). In BMMCs, ROSA MCs and RBL-2H3 MCs, RACK1 displayed a punctate staining pattern that was dispersed throughout the cytoplasm. These results demonstrate that RACK1 is present in MCs both in vivo and in vitro.

The functional role of RACK1 was investigated primarily using RBL-2H3 MCs, a widely used model to study MC secretion (Falcone et al., 2018). The cells were stimulated via FcεRI and the distribution of RACK1 in non-stimulated and stimulated cells was analyzed. In non-stimulated cells, RACK1 was distributed throughout the cytoplasm in a punctate manner. However, after antigen stimulation for 1 and 5 min, the intensity of RACK1 immunostaining increased close to the plasma membrane as well as in the juxtanuclear region (Fig. 1E). 3D image projections show that following stimulation, there was an increase in RACK1 over the nucleus (inset). Therefore, the distribution of RACK1 is related to the physiological state of the MCs.

To further investigate the involvement of RACK1 in MC function, RBL-2H3 MCs were transduced with lentiviral particles encoding shRNAs against RACK1 to generate RACK1 knockdown (KD) MCs (clones ShRACK1 cl.12 and ShRACK1 cl.29). As assessed by RT-PCR there was a 67.6±8.22% (mean±s.e.m.) decrease in RACK1 mRNA expression compared to control cells (Fig. 2A). Furthermore, as shown by western blotting, RACK1 KD led to a 62.5±9.48% reduction in RACK1 protein levels (Fig. 2B,C). RACK1 expression was lower in ShRACK1 cl.29 cells than in ShRACK1 cl.12 cells, but there was no significant difference between them. Therefore, the ShRACK1 cl.29 MCs were used for all further studies. Additionally, after transfection with control shRNA (ShCtrl), mRNA and RACK1 expression were similar to untransduced (UT) cells, confirming that transduction with non-targeting shRNA did not affect RACK1 expression (Fig. 2A–C).

Since RACK1 has been correlated with maintenance of morphology in diverse cell types, it was important to examine the role of RACK1 in regulating MC morphology. As shown by scanning electron microscopy (Fig. 2D), control cells (UT and ShCtrl) had the typical fusiform shape of RBL-2H3 MCs, and their surface was covered with small, short microvilli. In contrast, the majority of RACK1 KD MCs (ShRACK1 cl.29) were round or oval, and their surface was altered with a few ruffles. After staining the actin cytoskeleton with Alexa Fluor 488-conjugated phalloidin (phalloidin–Alexa 488) (Fig. 2E), a black-and-white representation of each cell was created (Fig. 2F) to accentuate the shape of the individual cell, and was used to determine the roundness index (RI) (Schober et al., 2009) and cell area (Fig. 2G,H). There was a significant increase in the RI of RACK1 KD cells (Fig. 2G), but there was no difference in the area of the analyzed cells (Fig. 2H).

Because the shape of the RACK1 KD MCs was altered and actin microfilaments are essential to maintain cell morphology, the actin distribution in these cells (Fig. 3) was investigated. After staining with phalloidin–Alexa 488, the fluorescence intensity of actin filaments was analyzed by confocal microscopy. In non-stimulated UT cells and ShCtrl cells, the F-actin filaments were continuous, especially in the cortical regions. In contrast, in non-stimulated RACK1 KD MCs (ShRACK1 cl.29) the cortical F-actin was fragmented with spaces in the subcortical areas characterized by the reduction or absence of fluorescence in those regions. A similar phenotype was observed when ShCtrl cells were pretreated with a low concentration (0.5 µM) of latrunculin B (LatB), which inhibits actin polymerization. Following 5 min of antigen stimulation, in control cells (UT and ShCtrl), F-actin was primarily localized to continuous filaments adjacent to the plasma membrane and associated with surface ruffles and microvilli. In stimulated RACK1 KD cells, cortical F-actin was interrupted by numerous spaces and there was a significant decrease in the fluorescence intensity of F-actin, especially in cortical regions. Additionally, when ShCtrl cells were pretreated with LatB and then stimulated, the F-actin distribution was similar to that seen in the stimulated RACK1 KD cells. These results indicate that RACK1 is involved in maintaining the integrity of cortical F-actin.

In order to further analyze the impact of RACK1 on actin cytoskeleton arrangement, the ratio of depolymerized monomer G-actin to the polymerized F-actin in the non-stimulated and antigen-stimulated cells was determined by dividing the G-actin levels by the F-actin levels (Fig. 4A,B). In resting cells, although there was a shift in the G-actin/F-actin ratio in RACK1 KD cells and ShCtrl plus LatB cells when compared to ShCtrl cells, this was not significant. After stimulation for 5 min, the G-actin/F-actin ratio decreased slightly in control cells (UT and ShCtrl cells). However, there was a significant increase in the G-actin/F-actin ratio in stimulated RACK1 KD cells when compared with stimulated ShCtrl cells. A similar result was seen in stimulated LatB-treated cells. It is important to note that RACK1 KD had no effect on the expression of β-actin, either at the mRNA or protein level (Fig. 4C–E).

Since RACK1 appears to be involved in actin cytoskeleton organization and influences F-actin assembly, the impact of RACK1 KD on F-actin dynamics during MC activation in living cells was investigated. Transduced ShCtrl cells and ShRACK1 cl.29 cells were transiently transfected to express LifeAct–RFP and the actin rearrangement following antigen stimulation was examined in living cells by total internal reflection fluorescence microscopy (TIRFM). A rapid decrease in cortical F-actin intensity was observed 50 s after antigen stimulation in ShCtrl cells. Subsequently, F-actin returned as fluorescent puncta at 400 s, and at 600 s appeared as a highly organized network of filaments in the cell cortex (Fig. 4F; Movie 1). In contrast, in non-stimulated RACK1 KD MCs, at 0 s cortical F-actin was already disorganized (Fig. 4F). With time after stimulation, cortical F-actin intensity appeared to rise, but in a completely disorganized manner, which persisted even after 600 s. The F-actin formed large aggregates throughout the cell cortex, instead the organized bundles seen in the ShCtrl cells (Fig. 4F; Movie 2). These results indicate that RACK1 has an essential role in actin cytoskeleton assembly and antigen stimulus-induced F-actin dynamics in MCs.

Since cortical F-actin modulates the secretion of preformed mediators, the release of β-hexosaminidase was evaluated. In non-stimulated RACK1 KD MCs (ShRACK1 cl.12 and ShRACK1 cl.29) the basal level of β-hexosaminidase release was higher than that seen in non-stimulated UT cells or ShCtrl cells (Fig. 5A, inset). LatB treatment also resulted in a significant increase in β-hexosaminidase release in non-stimulated MCs. Following antigen stimulation, the average percentage of β-hexosaminidase activity released was increased 1.77±0.32-fold (mean±s.e.m.) in RACK1 KD MCs in comparison with ShCtrl cells (Fig. 5A). β-hexosaminidase release was also more rapid in RACK1 KD MCs. The amount of β-hexosaminidase activity released by RACK1 KD MCs 5 min after antigen stimulation was similar to that released by ShCtrl cells 60 min after stimulation. ShCtrl plus LatB cells showed the highest average β-hexosaminidase release when antigen stimulated (2.18±0.51-fold higher in comparison with ShCtrl cells) (Fig. 5A). Mouse BMMCs were transduced with ShRACK1 cl.12 or ShRACK1 cl.29 (Fig. S2) and stimulated for 60 min via FcεRI. In the BMMC RACK1 KD cells (ShRACK1 cl.12 and ShRACK1 cl.29), the released β-hexosaminidase activity was 1.4±0.08- and 1.49±0.18-fold higher (mean±s.d.), respectively, than in untransduced BMMCs or in BMMCs transduced with ShCtrl (Fig. 5B).

In order to determine whether stimulation via FcεRI was critical for the increased release of β-hexosaminidase observed in the RACK1 KD cells, the cells were also stimulated independently of FcεRI with thapsigargin (TG) (Fig. 5C) or with 0.1 µg/ml Ca²⁺ ionophore A23187 (Fig. 5D). RACK1 KD MCs showed increased β-hexosaminidase release with both stimuli when compared with ShCtrl cells. Additionally, ShCtrl plus LatB cells demonstrated responses similar to those of ShRACK1 KD MCs (Fig. 5C,D). Thus, these results show that there is also an increase in degranulation when the cells are activated independently of FcεRI. The increased release by RACK KD cells could also be explained by the presence of more total β-hexosaminidase or a larger number of FcεRI in the RACK1 KD cells. However, the total β-hexosaminidase activity (Fig. 5E) and the expression of FcεRI (Fig. 5F,G) were unaltered in RACK1 KD MCs.

In addition to increasing the release of preformed mediators, RACK1 also influenced the release of newly synthetized mediators, but not newly formed lipid mediators (Fig. S3). Following antigen stimulation, the release of prostaglandin D2 (PGD2) and leukotriene C4 (LTC4) was not significantly affected by RACK1 KD (Fig. S3A,B). However, the antigen-stimulated secretion of several cytokines was substantially reduced in RACK1 KD MCs when compared to ShCtrl cells (Fig. S3C). The reduced release of interleukin (IL)-4 and IL-13 in RACK1 KD cells was also confirmed and quantified by ELISA assay (Fig. S3D,E). The production and release of newly synthetized cytokines in MCs stimulated via FcεRI requires the activation of transcription factors. Therefore, RBL-2H3-derived GFP reporter cell lines were used to assess the activation of the transcription factors nuclear factor κB (NFκB) and nuclear factor of activated T cells (NFAT). The impact of RACK1 KD on activation of NFκB (Fig. S3F) was different from that on NFAT (Fig. S3G). These results indicate that RACK1 is involved in the synthesis and release of newly synthetized mediators.

Protein tyrosine phosphorylation is one of the initial steps during MC activation via FcεRI. Non-stimulated RACK1 KD cells had significantly reduced levels of tyrosine phosphorylation when compared to control cells. After antigen stimulation for 1 and 5 min, the level of phosphorylation increased in both control and RACK1 KD cells, but it was not significantly different between the cell lines (Fig. 5H,I). The ratio between tyrosine phosphorylation and α-tubulin in stimulated/non-stimulated ShCtrl cells increased from 1.24±0.06 (mean±s.e.m.) at 1 min and 1.33±0.04 at 5 min, which was similar to UT cells (1.21±0.13 at 1 min and 1.38±0.07 at 5 min). However, this ratio was higher for RACK1 KD cells. For ShRACK1 cl.12 cells, the ratio was 1.34±0.22 at 1 min and 1.59±0.23 at 5 min. The ratios for the ShRACK1 cl.29 cells were 1.34±0.10 at 1 min and 1.81±0.09 at 5 min (Fig. 5I). Thus, RACK1 appears to play a role in antigen-induced tyrosine phosphorylation in MCs.

In resting MCs, cortical F-actin acts as a barrier for the fusion of SGs with the plasma membrane, thus preventing preformed mediator release. Therefore, it was of interest to determine whether SGs were present in the cortical F-actin-free regions seen in RACK1 KD MCs. Cells were immunolabeled for CD63 to identify SGs and stained with phalloidin–Alexa 488 to label F-actin. In non-stimulated control cells (UT and ShCtrl), CD63⁺ SGs were dispersed throughout the cytoplasm and in the juxtanuclear region, while cortical F-actin formed a continuous layer under the plasma membrane (Fig. 6A). In contrast, in non-stimulated RACK1 KD MCs (ShRACK1 cl.29) or ShCtrl plus LatB cells, SGs were aggregated and closer to the surface and the cortical F-actin was discontinuous (Fig. 6A). When the control cells (UT and ShCtrl) were antigen stimulated for 5 min the SGs moved closer to the cell surface and cortical F-actin was now partially disrupted (Fig. 6B). However, the stimulated RACK1 KD MCs and ShCtrl plus LatB cells had fewer SGs, and some of them were intercalated with the cortical F-actin (Fig. 6B). 3D reconstruction of Z-stacks from the confocal images demonstrated the relationship of the SGs with the cortical F-actin (Fig. 6C). Non-stimulated control cells (UT and ShCtrl) showed a barrier of F-actin at the cell surface, and very few SGs could be seen. In contrast, in the non-stimulated RACK1 KD MCs and ShCtrl plus LatB cells there were cortical F-actin-free regions with SGs present in these actin-free areas. Therefore, the higher basal levels of released β-hexosaminidase activity in resting RACK1 KD MCs and ShCtrl plus LatB cells can partially be explained by increased access of SGs to the plasma membrane.

The possible interaction of RACK1 with β-actin, one of the major isoforms of actin in non-muscle cells, was then examined. RACK1 was immunoprecipitated from RBL-2H3 MC lysates before and after antigen stimulation (Fig. 7A). No difference was seen in the expression of RACK1 following MC activation (Fig. 7A,B). Furthermore, β-actin co-precipitated with RACK1 in non-stimulated and antigen-stimulated cells (Fig. 7A). The amount of β-actin co-precipitating with RACK1 was increased 3.96±1.45 fold at 1 min and 3.94±1.57 at 5 min (mean±s.d.) after cell activation (Fig. 7C).

Through high-resolution confocal microscopy (Fig. 7D), a space was observed between RACK1 and β-actin, suggesting that RACK1 was not binding directly to actin, but rather to a linker. Therefore, a mass spectrometry analysis (MS/MS) of proteins that co-immunoprecipitated with endogenous RACK1 was performed to identify possible actin-binding proteins (ABPs) and cytoskeleton regulatory proteins (Fig. 7E). The proteomic analysis revealed diverse possible ABPs that co-immunoprecipitated with RACK1 (Table S1). These ABPs had significant differences in intensity between the non-stimulated and stimulated cells (Fig. 7E). The ABPs belong to different families of actin binding proteins (Pollard, 2016), such as filament severing/capping proteins (gelsolin, flightless I homolog, adseverin, tropomodulin-3 and tropomyosin α3), focal adhesion proteins (vinculin and talin), cross-linking proteins (spectrin-α non-erythrocytic 1, α-adducin, members of the ARP2/3 complex, HCLS1 and α-actinin-4) and motor proteins (myosin Va). Additionally, proteins that indirectly regulate the actin cytoskeleton were also identified by MS/MS (Fig. 7E). These proteins included calmodulin-3, RAC1, RhoG, RAB27a, PP2A-α and pleckstrin.

Two of the ABPs that co-immunoprecipitated with RACK1, vinculin and MyoVa (also known as MYO5A), were selected to validate the proteomic analysis. The samples that were immunoprecipitated with RACK1 were immunoblotted for vinculin and MyoVa. In MCs, the focal adhesion ABP vinculin appears to be involved in FcεRI activation (Torres et al., 2008). The levels of vinculin were significantly higher after antigen stimulation, 3.07±0.13 fold for 1 min and 3.5±1.07 fold for 5 min (mean±s.d.), relative to non-stimulated cells (Fig. 7F,G). MyoVa is a motor protein that might act as a regulator of MC cortical F-actin (Singh et al., 2013). MyoVa co-immunoprecipitated with RACK1 in resting cells and this interaction increased after antigen stimulation by 1.57±0.20 fold for 1 min and 2.13±0.26 fold for 5 min relative to non-stimulated cells (Fig. 7F,H). These results demonstrate that RACK1 interacts with ABPs in MCs.

Since actin polymerization/depolymerization, as well as MC mediator release via FcεRI stimulation, are Ca²⁺-dependent processes, the effect of RACK1 KD on the Ca²⁺ response following activation was examined. Cells were antigen stimulated in Ca²⁺-containing medium. Using Fura-2-based fluorimetry and single-cell Ca²⁺ imaging analysis the individual cellular response for each cell line was averaged. Following stimulation, the levels of [Ca²⁺]_i differed between control cells (UT and ShCtrl) and RACK1 KD MCs (ShRACK1 cl.12 and ShRACK1 cl.29). The antigen-induced increase in [Ca²⁺]_i was significantly higher in the RACK1 KD MCs, especially at later times after stimulation (Fig. 8A).

The involvement of RACK1 in the two main events of antigen-stimulated SOCE, the depletion of the ER-Ca²⁺ stores and the influx of Ca²⁺ across the plasma membrane, triggered by FcεRI activation, was then investigated. Transient intracellular Ca²⁺ store depletion after stimulation was assessed by stimulating the cell lines in the absence of extracellular Ca²⁺. When stimulated without external Ca²⁺, control cells (UT and ShCtrl) showed a transient peak in [Ca²⁺]_i that, over time, returned to basal levels. In contrast, in RACK1 KD MCs this transient peak was lower (Fig. 8B; −Ca²⁺). After replenishment of the extracellular Ca²⁺ at ∼200 s post stimulation, the Ca²⁺ influx in RACK1 KD MCs was also lower in comparison to control cells (UT and ShCtrl) (Fig. 8B; +Ca²⁺). These results suggest that RACK1 plays an important role in antigen-stimulated SOCE.

Since Ca²⁺ mobilization is altered in antigen-stimulated RACK1 KD MCs and the actin cytoskeleton is disorganized in these cell lines, the possible correlation between these events was investigated. ShCtrl cells were pretreated or not with 0.5 mM LatB and the Ca²⁺ response induced by antigen stimulation was analyzed. Initially, cells were stimulated in the presence of extracellular Ca²⁺. The increase in [Ca²⁺]_i was also significantly higher in LatB-treated cells (Fig. 8C). When cells were antigen stimulated in a Ca²⁺-free solution, the initial [Ca²⁺]_i response was not significantly different between the non-treated ShCtrl and LatB-treated cells (Fig. 8D; −Ca²⁺). The replenishment of extracellular Ca²⁺ at ∼200 s resulted in an increased influx of Ca²⁺ into LatB treated cells (Fig. 8D; +Ca²⁺). LatB treatment affected the antigen-stimulated Ca²⁺ mobilization. However, when stimulated in the absence of external Ca²⁺, the LatB-treated cells did not demonstrate the defective depletion of intracellular ER-Ca²⁺ stores and consequent Ca²⁺ entry shown by RACK1 KD cells.

To better evaluate the function of RACK1 in Ca²⁺ mobilization in MCs, cells were activated with TG to bypass FcεRI signaling. TG raises [Ca²⁺]_i by inhibiting Ca²⁺ ATPase pumps located in the ER membrane, thus blocking Ca²⁺ reuptake into the ER (Christensen et al., 1993). ER-Ca²⁺ store depletion induced by this treatment can then lead to activation of plasma membrane Ca²⁺ channels and extracellular Ca²⁺ influx. The response induced by TG stimulation in the presence of extracellular Ca²⁺ was less in the RACK1 KD cells than in the control cells (UT and ShCtrl cells), although the initial rise was similar (Fig. 8E). This suggests that the TG-dependent intracellular ER store depletion was normal in the RACK1 KD cells, but that Ca²⁺ influx was defective. To evaluate this possibility, a Ca²⁺ addback assay was performed where TG was added in the absence of extracellular Ca²⁺, followed by the addition of 1 mM CaCl₂. The initial intracellular Ca²⁺ release from intracellular ER stores induced by TG stimulation was the same for both ShCtrl cells and RACK1 KD cells (Fig. 8F; −Ca²⁺). However, when the extracellular Ca²⁺ was replenished at ∼200 s post stimulation the RACK1 KD cells showed a diminished Ca²⁺ influx (Fig. 8F; +Ca²⁺). These findings confirm that RACK1 is involved in TG-induced SOCE.

The possible correlation between the disorganized actin cytoskeleton and Ca²⁺ mobilization in RACK1 KD MCs was examined by treating ShCtrl cells or not with LatB before activation of SOCE with TG. The response induced by TG in the presence of extracellular Ca²⁺ was less in LatB-treated cells than in non-treated ShCtrl cells (Fig. 8G). In the absence of extracellular Ca²⁺, the intracellular Ca²⁺ response was the same for non-treated and LatB treated cells after stimulation with TG (Fig. 8H; −Ca²⁺). In contrast, after the replenishment of extracellular Ca²⁺ at ∼200 s post stimulation there was a decreased influx of Ca²⁺ in LatB-treated cells (Fig. 8H; +Ca²⁺). Therefore, LatB treatment, which results in actin cytoskeleton disarrangement, reduces Ca²⁺ entry trigged by TG and is similar to that seen in the RACK1 KD MCs.

Because RACK1 appears to be a multifunctional regulator of Ca²⁺ mobilization in MCs, it was also of interest to investigate whether the main CRAC channel subunit Orai1, which is involved in Ca²⁺ influx during SOCE in MCs, binds to RACK1. In non-stimulated cells Orai1 co-precipitated with RACK1 and the amount of this interaction increased ∼3.8±1.28 fold (mean±s.d.) after antigen stimulation for 1 min and was further increased (7.02±0.96 fold) at 5 min (Fig. 8I,J). These data suggest a possible direct effect of RACK1 on Orai1 function.

Changes in the distribution and morphology of the ER following actin rearrangement were also investigated by confocal microscopy and transmission electron microscopy (Figs S4–S6). By confocal microscopy, the differences in the distribution of the ER in the non-stimulated or stimulated UT, ShCtrl, ShRACK1 cl.29, ShCtrl plus LatB cells was related to the morphological changes induced by rearrangement of the actin cytoskeleton (see Fig. 3). By transmission electron microscopy of non-stimulated UT, ShCtrl, ShRACK1 cl.29 and ShCtrl plus LatB cells, and of stimulated UT and ShCtrl cells, the ER was observed to be composed of flattened cisternae with ribosomes attached to the cytoplasmic side of the cisternal membrane. However, when the cells were antigen stimulated for 5 min, differences could be seen in the ER in the ShRACK1 cl.29, ShCtrl plus LatB cells. The ER cisternae from the stimulated ShRACK1 cl.29 and stimulated ShCtrl plus LatB cells were dilated. This alteration in the ER could influence the formation and/or function of the CRAC channels as well as their coupling to STIM1 and consequently Ca²⁺ mobilization in the stimulated ShRACK1 cl.29, ShCtrl plus LatB cells.

The present study showed, for the first time, that the scaffold protein RACK1 is expressed in MCs and that RACK1 plays a critical role in the secretion of MC mediators due to its role in maintaining the actin cytoskeleton and in regulating intracellular Ca²⁺. In RBL-2H3 MCs, RACK1 was localized throughout the cytoplasm, but there was an increased concentration under the plasma membrane and in the perinuclear region after stimulation. This is consistent with the distribution and function of RACK1 in other cell types (Ballek et al., 2016; Buensuceso et al., 2001; Neasta et al., 2012).

RACK1 knockdown affected RBL-2H3 MC morphology, in agreement with previous reports on other cell types, which have shown that RACK1 has important roles in maintaining cell shape and in actin cytoskeletal organization (Buensuceso et al., 2001; Cox et al., 2003; O'Donovan et al., 2007). In cortical neurons, RACK1 binds to both the G and F forms of β-actin, possibly regulating cell morphology (Kershner and Welshhans, 2017; Neasta et al., 2016). In RAT2 fibroblasts, RACK1 knockdown leads to a failure in the antigen-induced rearrangement of F-actin while RACK1 depletion adversely affects F-actin assembly when the cells spread over a fibronectin substrate (Klímová et al., 2016). In contrast, overexpression of RACK1 results in an increase in actin stress fibers in Chinese hamster ovary cells (Buensuceso et al., 2001). Taken together with our findings, these results demonstrate that RACK1 is intimately associated with the organization of the actin cytoskeleton.

The present study also showed that RACK1 influences secretion of preformed mediators from MCs. SGs were localized in cortical actin-free regions of non-stimulated RACK1 KD cells or LatB pretreated cells and the increased secretion could be correlated with a marked disruption of the cortical F-actin barrier. This increased access of the SGs to the plasma membrane may facilitate granule fusion and increase mediator release from these cells since the cortical F-actin network acts as a physical barrier preventing docking and fusion of SGs with the plasma membrane (Aunis and Bader, 1988; Deng et al., 2009; Wilson et al., 2016; Wollman and Meyer, 2012). Therefore, it appears that RACK1 is an essential negative regulator of MC degranulation due to its participation in F-actin arrangement. It has been suggested that RACK1 participates in cortical granule exocytosis in rat eggs by binding to the actin cytoskeleton and tyrosine kinases, which facilitates granule translocation towards the egg membrane (Haberman et al., 2011). Additionally, when the ABPs coronin 1a and 1b, from the same WD40 repeat family as RACK1, bind to F-actin they also exert a negative regulatory effect on MC secretion (Föger et al., 2011).

High-resolution microscopy indicated that RACK1 does not bind directly to cortical F-actin in MCs. Several candidates that could serve as a linker between RACK1 and actin, including vinculin and MyoVa, were identified by proteomic analysis of RACK1 co-immunoprecipitates. Vinculin was previously described as a possible candidate for RACK1 interaction in neurons (Kershner and Welshhans, 2017). In MCs, it has been demonstrated that vinculin and other focal adhesion proteins localize to the same regions as IgE-clustered FcεRI, which are distinct from focal adhesion plaques, thus suggesting that vinculin participates in antigen-induced F-actin rearrangement (Torres et al., 2008). Additionally, in resting MCs, MyoVa is highly associated with RACK1 and this association increases following stimulation. This is in agreement with the role of MyoVa in SG release in MCs and other secretory cells (Desnos et al., 2007; Singh et al., 2013; Wollman and Meyer, 2012). MyoVa regulates the integrity of cortical F-actin in non-stimulated MCs (Singh et al., 2013). Following FcεRI activation, MyoVa is recruited to the MC cortex overlapping the region where most of the SG fusion occurs (Wollman and Meyer, 2012). Moreover, in MyoVa-deficient BMMCs degranulation is reduced (Singh et al., 2013).

In contrast to increased release of preformed mediators in RACK1 KD cells, RACK1 depletion impairs the release of newly synthesized cytokines. This can also be explained by F-actin disruption. The actin cytoskeleton may exhibit a functional dichotomy in regulating SGs release versus cytokine secretion in MCs (Föger et al., 2011). Actin-dependent processes have been implicated in the regulation of membrane trafficking from the ER and Golgi to the plasma membrane (Müsch and Rodriguez-Boulan, 1997). Additionally, NFκB transcription factor activation, necessary for cytokine production, was reduced in RACK1 KD MCs. The direct involvement of RACK1 in the intracellular pathways responsible for the synthesis of cytokines has been demonstrated in other immune cells (Corsini et al., 2014, 2009; Yao et al., 2014).

Following FcεRI activation, RACK1 KD MCs showed an increased Ca²⁺ mobilization. Our results agree with previous observations in HEK293 cells where the reduction in RACK1 expression had an impact on SOCE by increasing carbachol-induced Ca²⁺ mobilization. Carbachol stimulation, as with antigen activation, generates inositol 1,4,5-trisphosphate (IP₃) and consequently triggers Ca²⁺ influx across the plasma membrane (Bandyopadhyay et al., 2008; Patterson et al., 2004). ShCtrl cells treated with LatB as well as RACK1 KD cells showed a similar response following antigen stimulation in the presence of extracellular Ca²⁺. Therefore, the functional role of RACK1 in Ca²⁺ mobilization during FcεRI activation can be partially explained by the impact of RACK1 on actin cytoskeleton arrangement. In T cells, Ca²⁺-dependent retrograde actin flow directs ER tubule extensions and STIM1–Orai1 complexes to the synapse center, creating a self-organizing process for CRAC channel localization (Hartzell et al., 2016). Although there is substantial evidence that F-actin organization is important in FcεRI signaling, there is little experimental data showing the molecular mechanism of F-actin regulation of Ca²⁺ influx during MC activation (Tolarová et al., 2004; Wollman and Meyer, 2012).

RACK1 also appears to have an actin-independent role in ER Ca²⁺-store depletion in MCs. As demonstrated here and by Tolarová et al. (2004) when RACK1 KD cells are activated by FcεRI in Ca²⁺-free medium, in contrast to what is seen in cells treated with LatB, Ca²⁺ release from ER stores is defective. After replenishment of extracellular Ca²⁺, the [Ca²⁺]_i in RACK1 KD cells remains low, possibly reflecting reduced Ca²⁺ release from ER stores. After antigen stimulation in the presence of extracellular Ca²⁺, Ca²⁺ release from the ER occurs virtually simultaneously with the influx of extracellular Ca²⁺. In the RACK1 KD cells, although the Ca²⁺ release from the ER is deficient, this deficiency may be compensated for by increased Ca²⁺ influx resulting in an increased [Ca²⁺]_i in RACK1 KD cells_. Although CRAC channels are activated by the loss of Ca²⁺ from ER stores, the relationship between store depletion and subsequent channel activation is not linear (Parekh et al., 1997). Our data support the hypothesis that RACK1, in association with F-actin arrangement, orchestrates FcεRI-induced Ca²⁺ mobilization in MCs.

The present results demonstrated a role for RACK1 in SOCE induced by TG activation. RACK1 KD cells and LatB-treated cells showed a decrease in Ca²⁺ entry after stimulation by TG, in contrast to what is seen with antigen activation. TG blocks Ca²⁺ reuptake into the ER by inhibiting sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) pumps (Dar and Pecht, 1992), inducing an increase in [Ca²⁺]_i due to a spontaneous loss of Ca²⁺ from the IP₃-sensitive pool in the ER without stimulating IP₃-generation (Christensen et al., 1993). Our results indicate that RACK1 and/or actin cytoskeleton organization has an impact on Ca²⁺ influx through the plasma membrane independently of IP₃-induced ER Ca²⁺ release.

RACK1 also appears to have an intrinsic role in ER Ca²⁺-store depletion in MCs. RACK1 KD, in contrast to LatB treatment, resulted in defective FcεRI-activated Ca²⁺ release from ER stores. In carbachol-activated HEK293 cells, RACK1 binds to IP₃ receptors (IP₃Rs) and this association enhances IP₃-binding affinity to its receptors regulating Ca²⁺ release from ER stores (Bandyopadhyay et al., 2008).

Our study demonstrated that RACK1 binds to Orai1 in an antigen-stimulus-dependent manner. RACK1 acts as a scaffold protein in the Orai1–STIM1–IP₃R–TRPC3 complex induced by agonist stimulation in HEK293 cells (Bandyopadhyay et al., 2008; Woodard et al., 2010). In MCs stimulated via FcεRI, SOCE is induced by the interaction of ER STIM1 clusters with Orai1 and TRPC1 channels (Baba et al., 2008; Vig et al., 2008). Our results suggest that RACK1, in addition to its association with Orai1, is a novel regulator of the stimulus-induced Ca²⁺ signaling complex in MCs.

The morphological changes in the ER seen in the stimulated ShRACK1 cl.29 and ShCtrl plus LatB cells could influence the formation and/or function of the CRAC channels as well as their coupling to STIM1 and consequently Ca²⁺ mobilization in these cells. The effect of RACK1 KD on Ca²⁺ signaling appears to be mediated by the actin rearrangement seen after RACK1 KD, since treatment of ShCtrl cells with LatB produces the same results. It is known that the actin cytoskeleton is intimately associated with the ER (Gurel et al., 2014; Lynch et al., 2011) and that disruption in the cytoskeleton affects organization of the ER. Rearrangement of the cytoskeleton might also affect the interaction of the CRAC channel subunit Orai1 with STIM1, which would also affect Ca²⁺ mobilization and subsequent mediator release. SOCE is initiated by Ca²⁺ depletion from the ER, which causes conformational changes and clustering of STIM1 (Di Capite et al., 2011; Vig et al., 2008).

In conclusion, this study shows that knockdown of RACK1 in MCs results in a depolymerization and redistribution of the actin cytoskeleton that leads to gross morphological changes, increased release of preformed mediators and alterations in Ca²⁺ signaling.

C57BL/6 mice (7–8 weeks old, male) were used in this study. Animals were housed in the Animal Research Facilities of the Ribeirão Preto Medical School, University of São Paulo. Some mice were sensitized by subcutaneous injection of 4 µg chicken ovalbumin (OVA grade V, Millipore Sigma, St Louis, MO, USA) and alum 1.6 mg (Millipore Sigma) in PBS at days 0 and 7 to produce experimental asthma as previously described (Prado et al., 2015). The research was conducted in accordance with Ethical Principles in Animal Experimentation adopted by the National Council for Animal Experimentation Control. Experimental protocols were approved by the Ethics Committee on Animal Use of the Ribeirão Preto Medical School (protocols 140/2014, 043/2016).

Animals were euthanized in a CO₂ chamber at a flow rate of 20–30% CO₂ per minute according to Ribeirão Preto Medical School, Ribeirão Preto, Brazil guidelines, and fragments of the right and left lobes of the lung were washed with PBS and fixed in 4% formaldehyde (EM Sciences, Hatfield, PA, USA) in PBS for 4 h. Fragments from the right lobe were embedded in paraffin, and 5 µm sections were cut and mounted on glass slides. Serial sections were stained either with Toluidine Blue (0.1%, pH 2.8) or Hematoxylin and Eosin (H&E) and coverslips mounted with Permount (Thermo Fisher Scientific Inc., Waltham, MA, USA). Images were collected with an Olympus BX61 Motorized Slide Scanner Microscope Pred VS120 (Olympus, Hamburg, Germany). Fragments from the left lung were embedded in Tissue Freezing Medium (EM Sciences). The samples were then frozen in acetone cooled with dry ice and stored at −80°C. 12 µm sections were cut with a Microm cryostat (Thermo Fisher Scientific, Microm International GmbH, Waldorf, Germany) and placed on silane-coated Unifrost Microscope Slides (Azer Scientific, Morgantown, PA, USA). Sections were used for fluorescence microscopy as described below.

In order to obtain BMMCs, mice were euthanized as described above and bone marrow cells were obtained by flushing the marrow from the femurs with Dulbecco's PBS containing heparin (1000 IU/ml; Produtos Roche Químicos e Farmacêuticos, Rio de Janeiro, Brazil) and DNase (1000 IU/ml, Type I; Millipore Sigma). The cells were dissociated by aspiration with a Pasteur pipette. Cells were then rinsed twice by centrifugation at 27 g in PBS and cultured according to Jamur et al. (2005). The cultures were maintained in suspension for up to 3 weeks.

The human MC line ROSA (KIT D816V) was generously provided by Dr Olivier Hermine (Laboratory of Molecular Mechanisms of Hematologic Disorders and Therapeutic Implications. Imagine Institute, Paris, France) and maintained in suspension according to Arock et al. (2008) in serum-free MC culture medium [Iscove's modified Dulbecco's medium supplemented with 2×10⁻³ M L-glutamine, 100 IU/l penicillin, 100 μg/l streptomycin, 7.5×10⁻⁵ M β-mercaptoethanol, 2×10⁻⁴ mol/l bovine serum albumin (BSA), 5×10⁻⁷ M iron-saturated human transferrin, 1.7×10⁻⁶ M insulin, 20 ng/ml recombinant human stem cell factor].

RBL-2H3 rat MCs (Barsumian et al., 1981) and the stable RBL-2H3 cell lines expressing NFκB-GFP reporter and the NFAT-GFP reporter (kindly provided by Dr Reuben P. Siraganian, NIH, NIDCR, Bethesda, MD) (de Castro et al., 2010; Grodzki et al., 2009), were grown as monolayers in Dulbecco's minimum essential medium (DMEM) supplemented with 15% fetal calf serum and an antibiotic-antimycotic mixture (100 IU/ml penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml amphotericin B). Transfected cells were selected with Geneticin (0.5 mg/ml). All cell lines were periodically checked for the presence of specific markers such as FcεRI and mast cell-specific proteases. They were also routinely monitored for contamination.

All cells were grown at 37°C in a humidified incubator with 5% CO₂ in air. All the reagents used for cell culture were purchased from Thermo Fisher Scientific (Thermo Fisher Scientific, Invitrogen, Carlsbad, CA, USA).

The following primary antibodies and their conjugates were used: rabbit polyclonal anti-mouse mast cell tryptase 6 (mMCP-6) (1:1000) kindly provided by the late Dr Michael F. Gurish. Mouse monoclonal antibody (mAb) anti-FcεRI α subunit (15 μg/ml; clone BC4) generously provided by Dr Reuben Siraganian (NIH – NIDCR) was conjugated to FITC (BC4-FITC) using the FluoReporter™ FITC Protein Labeling Kit according to the manufacturer's directions (Thermo Fisher Scientific, Invitrogen). Rabbit polyclonal antibody anti-human Orai1 was produced against a C-terminal epitope ELAEFARLQDQLDHRGD and affinity purified by Lofstrand Labs Limited (1:100; Gaithersburg, MD, USA). Rabbit monoclonal antibody (mAb) anti-human RACK1 (1:2000 for western blotting and 1:100 for immunofluorescence; clone ERP7388; ab129084), anti-GRP78 BiP (1:400, clone 21685), goat antibody anti-human RACK1 (1:75; ab166796), and mouse mAb anti-chick α-tubulin (1:1000; clone DM1A; ab80779) were purchased from Abcam (Cambridge, MA, USA). Rabbit mAb anti-human RACK1 (1:1000; #5432), rabbit polyclonal antibody anti-human α/β-tubulin (1:1000; #2148), mouse mAb anti-human β-actin (1:5000 for western blotting and 1:1000 for immunofluorescence; clone 8H10D10; #3700), and anti-phospho-tyrosine (p-Tyr-1000) MultiMab™ rabbit mAb mix (1:1000; #8954) were purchased from Cell Signaling Technology Inc. (Beverly, MA, USA). Mouse mAb anti-rat CD63 (6 µg/ml; clone AD1; 551458) was purchased from BD Biosciences Pharmingen (San Jose, CA, USA). Rabbit mAb anti-mouse GAPDH (1:25,000; G9545) and mouse mAb anti-human vinculin (1:300; clone hVIN-1; V9131) were purchased from Millipore Sigma. Affinity purified rabbit polyclonal anti-chicken medial tail of MyosinVa (1:300) (Espindola et al., 1992) was kindly provided by Dr Enilza Maria Espreafico (Department of Cell and Molecular Biology and Pathogenic Bioagents, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil).

The secondary antibodies donkey anti-rabbit IgG conjugated to horseradish peroxidase (HRP) and donkey anti-mouse IgG conjugated to HRP were used for immunoblotting (1:10,000; Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA). The following secondary antibodies were used for immunofluorescence: donkey anti-rabbit IgG F(ab)′₂-Alexa 488 or 594; donkey anti-goat IgG F(ab)′₂-Alexa 594; donkey anti-mouse IgG F(ab)^′₂-Alexa 594 (1:1000; Thermo Fisher Scientific).

For antigen stimulation, RBL-2H3 MCs and BMMCs were sensitized overnight (ON) with mouse IgE anti-TNP ascites fluid (1:5000, generously provided by Dr Reuben Siraganian, NIH – NIDCR) and then stimulated via FcεRI with 50 ng/ml of DNP₅₄-HSA. For FcεRI independent stimulation, cells were incubated with 1 µM TG or 0.1 µg/ml Ca²⁺ ionophore A23187. For some experiments, the cells were treated with 0.5 µM latrunculin B (LatB) for 15 min at 37°C prior to stimulation and the same concentration of LatB was maintained in the medium during stimulation. All reagents were purchased from Millipore Sigma.

Cryosections from lungs were rinsed in PBS and incubated with Image-iT FX Signal Enhancer (Thermo Fisher Scientific) for 30 min at room temperature (RT). Next, sections were rinsed in PBS and incubated for 45 min at RT in PBS containing 0.5% BSA and normal donkey IgG (5 μg/ml; Jackson ImmunoResearch). Lung sections were labeled with primary antibodies diluted in PBS ON at 4°C. Then, sections were rinsed thoroughly in PBS and incubated for 45 min at RT with the secondary antibodies diluted in PBS. Sections were then rinsed, and coverslips mounted with Fluoromount-G (EM Sciences).

RBL-2H3 MCs were plated (4.0×10⁴ cells/coverslip) on 13 mm round coverslips, sensitized or not with IgE anti-TNP and cultured for 16 h. BMMCs were plated (5.0×10⁴ cells/coverslip) on 13 mm round coverslips treated with Cell-Tak Cell and Tissue Adhesive (Corning Life Sciences, Tewksbury, MA, USA). The cells were stimulated or not via FcεRI, rinsed in PBS and fixed for 20 min with 2 or 4% paraformaldehyde (EM Sciences) in PBS at RT. Cells were permeabilized with 0.01% saponin (Millipore Sigma) in PBS for 30 min, or with 0.3% Triton X-100 (Millipore Sigma) for 10 min at RT. Next, cells were rinsed twice in PBS and incubated for 45 min at RT in PBS containing 1% BSA and 5 μg/ml normal donkey IgG. Cells were labeled with primary antibodies diluted in PBS containing 1% BSA for 1 h at RT. Next, cells were rinsed thoroughly in PBS and incubated for 30 min at RT with the secondary antibodies diluted in PBS. For F-actin staining, phalloidin conjugated with Alexa Fluor 488 (2.6 IU/ml; Thermo Fisher Scientific) in PBS was used. Cells were then rinsed in PBS and mounted on glass slides with Fluoromount-G (EM Sciences). For nuclear staining, after incubation with secondary antibodies, the cells were incubated for 15 min at RT with DAPI (Thermo Fisher Scientific) at a concentration of 0.2 μg/ml in PBS. Cells incubated without primary antibody served as controls and were all negative. Samples were analyzed using a Zeiss LSM 780 laser scanning confocal microscope (Carl Zeiss, Heidelberg, Germany) or a LEICA TCS-NT SP5 laser scanning confocal microscope (Leica Microsystems, Heidelberg, Germany) or an Olympus Fluoview 1200 scanning confocal microscope (Olympus Scientific Solutions, Waltham, MA). High-resolution images were acquired using a Nikon Eclipse Ti2-E A1 high-resolution microscope (Nikon Instruments Inc., Melville, NY). Some images were deconvoluted.

The images obtained with the SP5 confocal microscope were analyzed using the Leica Application Suite X software (LAS X; Leica Microsystems). ImageJ (Schindelin et al., 2012) was used to determine roundness index and cell area (Schober et al., 2009), and to apply the FIRE look-up-table and create the surface plot using the 3D Surface Plot Plug-In.

MISSION^® lentiviral transduction particles encoding RACK1 shRNA were designed against the following target sequences: clone 12 (ShRACK1 cl.12), 5′-TCTGGCTAACTGCAAGCTAAA-3′ (Catalog No. TRCN0000012702) and clone 29 (ShRACK1 cl.29), 5′-TCTGGCTAACTGCAAGCTAAA-3′ (Catalog No. TRCN0000294547). A control shRNA (ShCtrl) was also used, the MISSION TRC2 pLKO.5-puro non-mammalian with the insert sequence, 5′-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTT-3′ (Catalog No. SHC202). All of these contained a puromycin resistance gene and were purchased from Millipore Sigma.

RBL-2H3 MCs and BMMCs were transduced with control or the two premade lentivirus particles encoding RACK1 shRNAs for 16 h or 48 h, respectively, at a multiplicity of infection (MOI) of 6. The medium containing the virus was removed and replaced with fresh medium and the cells cultured for 24 h before addition of puromycin (1–3 µg/ml; Millipore Sigma) in order to select for cells in which the shRNA was integrated. Real-time PCR and immunoblotting were employed to monitor RACK1 mRNA and protein expression, respectively.

For real-time PCR experiments, total RNA was purified from 5.0×10⁶ cells using the Illustra™ RNAspin Mini Isolation Kit (GE Healthcare Life Sciences, Chicago, IL, USA) according to the manufacturer's instructions. For cDNA synthesis, 5 μg of total RNA was reverse transcribed using the GoScript™ Reverse Transcription System according to the manufacturer's instructions (Promega Co., Madison, WI, USA). Gene-specific primers were used for quantitative PCR analysis. Power SYBR Green PCR Master Mix (Thermo Fisher Scientific) was used with 10 ng of RNA/well of the cDNA product in an ABI 7500 Real Time PCR System (Thermo Fisher Scientific). For all RT-PCR analysis, GAPDH mRNA was used to normalize RNA inputs. Primer sequences are as follows: rat RACK1 forward, 5′-GTGCTCTTCGAGGTCACTCC-3′; rat RACK1 reverse, 5′-CGGTTGTCAGAGGAGAAAGC-3′; rat β-actin forward, 5′-GCATTGCTGACAGGATGCAG-3′; rat β-actin reverse, 5′-GTAACAGTCCGCCTAGAAGCA-3′; rat GAPDH forward, 5′-ATGACTCTACCCACGGCAAG-3′; and rat GAPDH reverse, 5′-CTGGAAGATGGTGATGGGTT-3′.

RBL-2H3 MCs (7×10⁵ cells) were lysed as previously described (Freitas-Filho et al., 2016). Briefly, cells were washed twice with ice cold PBS and immediately lysed with triple detergent lysis buffer (Jamur and Oliver, 1996) [50 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.1% SDS, 1% Nonidet-P40, 0.5% sodium deoxycholate, containing 15 µl/ml of protease inhibitor (Thermo Fisher Scientific) and 10 µl/ml of phosphatase inhibitor cocktail (Thermo Fisher Scientific)], and removed from the flasks by scraping. For immunoblotting, the lysates were sonicated, centrifuged at 18,000 g, and the supernatants collected. The protein content was quantified using the Bradford reagent (Millipore Sigma), with BSA as the standard (Bradford, 1976). Lysate containing 20 µg of protein was mixed with NuPAGE^® LDS Sample Buffer (Thermo Fisher Scientific), boiled, and the proteins were separated electrophoretically on NuPAGE^® 4-12% Bis-Tris Protein Gels (Thermo Fisher Scientific) and transferred to PVDF membranes using Trans-Blot^® TurboTM Transfer System RTA Transfer Kits (Bio-Rad Laboratories, Hercules, CA). After transfer, the membranes were blocked for 1 h at RT in 0.05 M Tris-HCl, 0.15 M NaCl, pH 7.5, and 0.05% Tween 20 (TTBS) containing 4% BSA or 5% nonfat dry milk (Bio-Rad). After blocking, the membranes were incubated overnight at 4°C or 1 h at RT with the individual primary antibodies diluted in TTBS. The membranes were then washed, incubated for 30 min at RT with the appropriate anti-IgG conjugated to HRP, washed and developed using SuperSignal™ West Pico Plus Chemiluminescent Substrate (Thermo Fisher Scientific). The images were obtained by exposing the membranes to X-ray film (Carestream^® Kodak^® BioMax^® MR film) or with ImageQuant LAS 4000 (GE Healthcare Life Sciences) and ChemiDoc XRS+ (Bio-Rad Laboratories, Inc., Hercules, CA, USA). After exposure, the films were digitized and the mean optical density of the target protein was determined using ImageJ (Schindelin et al., 2012) or ImageQuant LAS™ 4000 (GE Healthcare Life Sciences) software. For stripping, after the first immunoblotting, the membrane was immersed in Restore™ Western Blot Stripping Buffer (Thermo Fisher Scientific) for 15 min at RT, washed, and the immunobloting was repeated as described above.

For scanning electron microscopy, RBL-2H3 MCs were plated on 13 mm round coverslips (5.0×10⁴ cells/coverslip). Cells were rinsed in warm PBS (37°C) and fixed with 2% glutaraldehyde (EM Sciences) in warm PBS for 2 h at RT. Cells were post fixed in 1% OsO₄ (EM Sciences) for 2 h, rinsed in Milli-Q water, incubated with a saturated solution of thiocarbohydrazide (EM Sciences), followed by 1% OsO₄. This step was repeated once. The cells were dehydrated in a graded series of ethanol and critically point-dried with liquid CO₂ in a Tousimis Autosandri-810 (Tousimis Research Co., Rockville, MD, USA), mounted on aluminum stubs with silver paint (EM Sciences), and coated with gold in a BAL-TEC SCD 050 Sputter Coater (Bal-Tec AG, Balzers, Lichtenstein). Samples were examined with a JEOL JSM-6610 LV scanning electron microscope (JEOL Ltd., Tokyo, Japan).

For transmission electron microscopy, cells were plated (4.5×10⁴ cells/well) in six-well tissue culture plates (Corning Life Sciences) and media were changed daily for 2 days. The cells were then rinsed in PBS and fixed in 2% glutaraldehyde (Electron Microscopy Sciences) and 2% formaldehyde (Electron Microscopy Sciences) in 0.1 M cacodylate buffer (pH 7.4), containing 0.025% CaCl₂ for 2 h at RT. The cells were rinsed in 0.1 M cacodylate buffer (pH 7.4) and postfixed in 2% OsO₄ (Electron Microscopy Sciences) in 0.1 M cacodylate buffer (pH 7.4) for 1 h, rinsed in Milli-Q water, and dehydrated in a graded ethanol series. Cells were removed from the tissue culture plates with propylene oxide and embedded in EMBED 812 (Electron Microscopy Sciences). Thin sections were cut with a diamond knife, mounted on copper grids, and stained for 10 min each in Reynolds's lead citrate (Reynolds, 1963) and 0.5% aqueous uranyl acetate, and examined with a JEOL JEM-100CXII (JEOL Ltd., Tokyo, Japan) transmission electron microscope equipped with a Megaview III camera (EMSIS GmbH Muenster, Germany).

The width of the ER lumen was measured at six different points in a minimum of six cells for each condition using the line function in Adobe Photoshop.

The G-actin/F-actin ratio in RBL-2H3 cells transduced or not with shRNAs was determined using the G-actin/F-actin in vivo assay kit (Cytoskeleton, Inc., Denver, CO, USA), according to the manufacturer's instructions. Briefly, 10⁶ cells of each cell line were plated, sensitized or not with IgE anti-TNP and cultured for 16 h. Then, cells were stimulated or not via FcεRI. Cells were washed in RT PBS and processed according to the kit instructions for the detailed method. Non-stimulated or antigen-stimulated samples were always run on the same gel and membranes were developed and exposed at the same time.

To evaluate the influence of RACK1 on cortical F-actin dynamics, transduced shRNA RBL-2H3 MCs were electroporated to transiently express LifeAct-RFP (pCMV-LifeAct^®-TagRFP; IBIDI, Martinsried, Germany) using an Amaxa^® Nucleofector^® (Lonza, Allendale, NJ, USA) following the manufacturer's instructions (kit T, program X001) and plated in 35 mm glass bottom (No. 1.0 collagen coated cover slip) culture dishes (MatTek Corporation, Ashland, MA, USA). After 24 h, cells were sensitized or not with IgE anti-TNP for 16 h and then stimulated via FcεRI. Live-cell TIRF images were acquired with an Olympus IX81 motorized inverted microscope (Olympus, Center Valley, PA, USA) using a 568 nm laser for excitation of RFP, a TIRF-optimized Olympus Plan APO 60× (1.45 NA) oil immersion objective and Lambda 10-3 filter wheel (Sutter Instruments, Novato, CA, USA) containing the 575lp filter for emission. Images were collected using a Hamamatsu ORCA-Flash4.0 camera (Hamamatsu Photonics K.K. Systems Division, Higashi-ku, Hamamatsu City, Japan) and MetaMorph imaging software (Molecular Devices, San Jose, CA, USA).

MC degranulation was assessed by measuring the activity of β-hexosaminidase released. RBL-2H3 MCs (3.0×10⁴ cells/well) were plated in 96-well tissue culture plates (Corning Life Sciences). BMMCs (8.0×10⁴ cells/well) were plated in 96-well plates treated with Cell-Tak Cell and Tissue Adhesive (Corning Life Sciences). Cells were sensitized or not with IgE anti-TNP for 16 h and then stimulated or not via FcεRI, TG or Ca²⁺ ionophore for 5, 15, 30, 45 and 60 min. β-hexosaminidase activity was quantified in the supernatants and cell lysates by spectrophotometric analysis of the hydrolysis of 4-nitrophenyl N-acetyl-β-D-glucosaminide (Millipore Sigma) as previously described (Jamur et al., 2005; Silveira e Souza et al., 2008). β-hexosaminidase release was calculated as the percentage of β-hexosaminidase activity measured in the supernatants relative to the total amount of β-hexosaminidase activity measured in the supernatant and cell lysates. Total β-hexosaminidase activity was determined by adding the activity measured in the supernatant to that measured in the cells.

To assess lipid mediator release, RBL-2H3 MCs were plated (10⁵ cells/well) in 24-well tissue culture plates (Corning Life Sciences) sensitized or not with IgE anti-TNP and FcεRI activated or not for 30 min. Release of PGD2 and LTC4 into culture supernatants was analyzed using EIA kits (Cayman Chemical, Ann Arbor, MI, USA).

To measure newly synthesized cytokine release, RBL-2H3 MCs were plated (10⁵ cells/well) in 24-well tissue culture plates (Corning Life Sciences) sensitized or not with IgE anti-TNP and stimulated via FcεRI or not for 1 h. Culture supernatants were discarded and the cells were washed and incubated with fresh medium for an additional 23 h. The supernatants were collected and released cytokines were analyzed using the Proteome Profiler Rat Cytokine Array Kit, Panel A (R&D Systems, Inc. Minneapolis, MN, USA). IL-4 and IL-13 in the culture supernatants were measured using murine IL-4 ELISA Kit (PeproTech, Inc., Rocky Hill, NJ, USA) and rat IL-13 ELISA kit (eBioscience, Inc., San Diego, CA, USA). All experiments were performed according to the manufacturer's instructions.

Tyrosine phosphorylation was analyzed as previously described (de Castro et al., 2012). Briefly, 2×10⁵ RBL-2H3 cells/well were plated in 24-well plates, sensitized or not with IgE anti-TNP and cultured for 16 h. Then, cells were stimulated or not via FcεRI. At 1 or 5 min after stimulation, cell monolayers were washed twice with ice cold PBS containing 1 mM Na₃VO₄ (Millipore Sigma) and immediately lysed with hot sample buffer (160 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol, 0.005% bromophenol blue, 2 mM Na₃VO₄, 1% protease inhibitor cocktail, 1% phosphatase inhibitor cocktail; Thermo Fisher Scientific). The cell lysates were boiled for 15 min, and the proteins were separated electrophoretically and the western blotting was performed as described above.

Cells were cultivated (5.0×10⁵ cells) for 16 h, harvested with trypsin-EDTA, and rinsed by centrifugation at 27× g in PBS. To evaluate FcεRI surface expression, non-permeabilized cells were incubated at 4°C for 1 h with anti-FcεRI α subunit (mAb BC4) conjugated to FITC in PBS containing 1% BSA and 5 µg/ml normal donkey IgG (Jackson ImmunoResearch). To evaluate NFκB and NFAT activation, NFκB2 cells and VB9 cells were stimulated or not via FcεRI, rinsed, and cultured for an additional 5 h (NFκB activation) or 15 h (NFAT activation), washed in PBS, and fixed for 20 min with 2% paraformaldehyde before analysis with a Guava EasyCyte Mini System using Cytosoft Blue software (Guava Technologies Inc., Hayward, CA, USA).

For RACK1 immunoprecipitation, 4.5×10⁶ RBL-2H3 MCs were sensitized or not with IgE anti-TNP, and stimulated or not via FcεRI for 1 and 5 min. Cells were lysed using Pierce IP Lysis Buffer (Thermo Fisher Scientific), containing 10 µl/ml of protease inhibitor (Thermo Fisher Scientific), and 10 µl/ml of phosphatase inhibitor cocktail (Thermo Fisher Scientific). Cells were then removed from the flasks by scraping. The samples were placed on ice for 10 min, vortexed, centrifuged at 18,000 g and the supernatants collected. The protein content was quantified using the Bradford reagent (Millipore Sigma), with BSA as the standard (Bradford, 1976). Protein (1 mg) was precleared by incubation with an aliquot of protein A-Sepharose^® CL-4B beads (GE Healthcare Life Sciences) for 3 h, centrifuged at 950 g and the supernatant collected. Then, the precleared samples were incubated with rabbit mAb anti-human RACK1 (Cell Signaling Technology Inc.) or as a control CromoPure rabbit-IgG, whole molecule (Jackson ImmunoResearch Laboratories Inc.) for 1 h at 4°C with rotation. Next, the immunocomplexes were incubated overnight with protein A beads. The samples were then washed by centrifugation at 150 g in cold TTBS and boiled in NuPAGE^® LDS Sample Buffer with 0.1 M DTT (Thermo Fisher Scientific). The RACK1 immunoprecipitates were subjected to electrophoresis, transfer, and immunoblotted as previously described.

Three independent samples of the RACK1 co-immunoprecipitated from each condition, 0 min, 1 min and rabbit IgG (IP Ctrl), were pooled and loaded on NuPAGE^® 4-12% Bis-Tris Protein Gels (Thermo Fisher Scientific). After electrophoresis, the gels were stained with Coomassie Brilliant Blue R-250 (Thermo Fisher Scientific), cut into lanes and each lane was cut into 1 mm cubes. The proteomic analysis was performed by Creative Proteomics (Shirley, NY, USA). The gel cubes were submitted to in-gel digestion using trypsin (Promega, Madison, WI, USA). The resulting tryptic peptides were analyzed using nanoliquid chromatography coupled to tandem mass spectrometry (nanoLC-MS/MS Ultimate 3000 nano UHPLC system; Thermo Fisher Scientific) equipped with a trapping column (PepMap C18, 100 Å, 100 μm×2 cm, 5 μm) and an analytical reversed-phase column (PepMap C18, 100 Å, 75 μm×50 cm, 2 μm). The full scan was performed between 300–1650 m/z at the resolution 60,000 at 200 m/z, the automatic gain control target for the full scan was set to 3e6. The MS/MS scan was operated in Top 20 mode using the following settings: resolution 15,000 at 200 m/z; automatic gain control target 1e5; maximum injection time 19 ms; normalized collision energy at 28%; isolation window of 1.4 Th; charge sate exclusion: unassigned, 1, >6; dynamic exclusion 30 s. The acquired MS raw data were analyzed and searched against Rattus norvegicus entries (April 2020; 29,952 sequences) from the UniProt database (http://www.uniprot.org/) based on the species of the samples using Maxquant (1.6.2.6). The parameters were set as follows: the protein modifications were carbamidomethylation (C) (fixed), oxidation (M) (variable); the enzyme specificity was set to trypsin; the precursor ion mass tolerance was set to 10 ppm, and MS/MS tolerance was 0.5 Da. In this study, a protein was considered as a potential binding partner if it was identified enriched at least 1.5 times in RACK1 co-immunoprecipitated samples compared to rabbit IgG control. The fold-change cutoff between 0 min and 1 min samples was set as proteins with quantitative ratios above 1.5 or below 1/1.5 and had more than a 95% confidence.

RBL-2H3 MCs (2.5×10⁴–4×10⁴ cells) were plated in 35 mm glass bottom (No. 1.0 collagen coated cover slip) culture dishes (MatTek Corporation). Cells were then sensitized or not with IgE anti-TNP. After 16 h, the media was discarded and the cells were washed twice with a standard extracellular solution (SES; 145 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 10 mM glucose, 1 mM CaCl₂, and 10 mM HEPES; pH 7.4 adjusted with NaOH) containing 0.1% BSA. Cells were incubated with the Ca²⁺ indicator Fura-2 acetoxymethyl ester (2 µM; Calibiochem, San Diego, CA, USA) diluted in SES for 30 min at 37°C in a 5% CO₂ incubator. Then, the cells were washed twice with SES with or without CaCl₂ and analyzed, before and after stimulation. All the experiments were performed using an Olympus IX50 microscope (Olympus) coupled to a Polychrome V spectrofluorimeter (Till Photonics LLC, Pleasanton, CA, USA). Fluorescence images were collected every 3 s for at least 600 s with a CoolSnap HQ camera (Photometrics, Tucson, AZ, USA), and the data were analyzed with MetaFluor imaging software (Molecular Devices) and plotted using Origin software (OriginLab, Northampton, MA, USA).

Data were plotted and analyzed using GraphPad Prism 6.0 software (GraphPad Software, Inc., La Jolla, CA). In all shRNA experiments and co-immunoprecipitation experiments, differences between the groups were assessed by one-way ANOVA followed by a post-hoc multiple Bonferroni's or Dunnet's test. In other experiments, the significance of intergroup differences was evaluated by a two-tailed paired t-test. The P values are as labeled as follows: *P<0.05; **P<0.005; ***P<0.0005; ns, not significant. Differences were considered statistically significant when at least P<0.05.

Part of the text and some of the figures in this paper are from Edismauro Garcia Freitas Filho's PhD thesis (2019) from the Cell and Molecular Biology Program at the Ribeirão Preto Medical School, University of São Paulo. The authors thank Elizabete Rosa Milani and Dr Roberta Rosales from the Multiuser Confocal and Multi-photon Microscopy Laboratory for their assistance with the confocal microscopy; Maria Dolores Seabra Ferreira, Maria Teresa Picinoto Maglia and Jose Augusto Maulin from the Multiuser Electron Microscopy Laboratory for their assistance with the electron microscopy; and Tânia A. Defina for assistance with the RT-PCR experiments, all from the Department of Cell and Molecular Biology and Pathogenic Bioagents, FMRP-USP, Ribeirão Preto, SP, Brazil. We also thank Dr Vania Luiza Deperon Bonato (Biochemistry and Immunology Department, FMRP-USP) for the mouse lung tissue.