Cyclase-associated protein 2 (CAP2) controls MRTF-A localization and SRF activity in mouse embryonic fibroblasts

Recent studies identified cyclase-associated proteins (CAPs) as important regulators of actin dynamics that control assembly and disassembly of actin filaments (F-actin). While these studies significantly advanced our knowledge of their molecular functions, the physiological relevance of CAPs largely remained elusive. Gene targeting in mice implicated CAP2 in heart physiology and skeletal muscle development. Heart defects in CAP2 mutant mice were associated with altered activity of serum response factor (SRF), a transcription factor involved in multiple biological processes including heart function, but also skeletal muscle development. By exploiting mouse embryonic fibroblasts (MEFs) from CAP2 mutant mice, we aimed at deciphering the CAP2-dependent mechanism relevant for SRF activity. Reporter assays and mRNA quantification by qPCR revealed reduced SRF-dependent gene expression in mutant MEFs. Reduced SRF activity in CAP2 mutant MEFs was associated with altered actin turnover, a shift in the actin equilibrium towards monomeric actin (G-actin) as well as and reduced nuclear levels of myocardin-related transcription factor A (MRTF-A), a transcriptional SRF coactivator that is shuttled out of the nucleus and, hence, inhibited upon G-actin binding. Moreover, pharmacological actin manipulation with jasplakinolide restored MRTF-A distribution in mutant MEFs. Our data are in line with a model in which CAP2 controls the MRTF-SRF pathway in an actin-dependent manner. While MRTF-A localization and SRF activity was impaired under basal conditions, serum stimulation induced nuclear MRTF-A translocation and SRF activity in mutant MEFs similar to controls. In summary, our data revealed that in MEFs CAP2 controls basal MRTF-A localization and SRF activity, while it was dispensable for serum-induced nuclear MRTF-A translocation and SRF stimulation.


Scientific Reports
| (2021) 11:4789 | https://doi.org/10.1038/s41598-021-84213-w www.nature.com/scientificreports/ acquisition 3-7% power intensity of AOTF 488 nm (FRAP-wizard). Imaging/bleaching program: pre-bleaching 5 × 2 s, bleaching 3 × 1.5 s (ROI 5 µM diameter), post-bleaching 10 × 2 s, 15 × 5 s. The image series were analyzed using FIJI software according to previous studies 27 . In brief, the background and bleaching correction was applied, and then normalized fluorescence intensity for each time point was calculated. Nonlinear curve fitting (one phase exponential association) of the fluorescence intensity was performed with GraphPad Prism, where the net recovery after photobleaching is provided by the following equation: Y = Y0 + (Plateau − Y0) × (1 − exp(− K × x)), where Y0 is the Y value when time is zero directly after the bleaching impulse, Plateau is the Y value at infinite times, expressed as a fraction of the fluorescence before bleaching and was used to determine the dynamic actin pool (F-actin dynamic). The stable pool (F-actin stable) is the fraction of fluorescence that does not recover within the imaging period of 95 s calculated as 1 − (F-actin dynamic), K is the rate constant, and τ is the time constant, expressed in seconds, it is computed as the reciprocal of K. SRF-luciferase reporter gene assay. To assess SRF activity, we generated MEF cell lines (control and CAP2 mutant) expressing the firefly luciferase reporter where MRTF-SRF promoter 3 Da.luc was linked to GFP. To generate the MEF cell lines stably expressing the MRTF-SRF luciferase reporter, first HEK293T cells were transfected using the calcium phosphate method. For lentivirus production, HEKT293T cells were cotransfected with the lentiviral packaging vectors psPAX and pMDG.2 together with the lentiviral vector FUGW expressing MRTF-SRF promoter 3 Da.luc linked to GFP. Generation of the lentiviral luciferase reporter construct has been described before 26 . After 48 h, supernatants containing viral particles were harvested, filtered, and used to transduce MEF cells. Transduced MEF cells were selected by FACS-based cell sorting.
MEF cells expressing the MRTF-SRF luciferase reporter were serum-deprived overnight and stimulated either with or without 20% serum for 24 h or 48 h. Then, cells were lysed with 200 μL Triton lysis buffer on ice and collected in 1.5 mL Eppendorf tube, followed by 10 min centrifugation at 13,000 rpm at 4 °C. The amount of firefly luciferase was measured luminometrically for each condition using a Glomax 96 Microplate Luminometer (Promega) 28 .

Fixation and immunohistochemistry.
MEFs were plated on coverslips coated with 0.01% calf skin collagen (Sigma Aldrich) in 0.1 M acetic acid. Coverslips were fixed with 4% PFA in PBS for 10 min and afterwards washed three times with PBS for 5 min each.
For immunohistochemistry, coverslips were treated for 1 h with blocking solution and afterwards incubated over night at 4 °C in carrier solution containing the primary antibody. After three washing steps of 5 min in PBS, coverslips were incubated for 2 h at room temperature (RT) in carrier solution containing Alexa-488 conjugated secondary antibodies (1:200, Life Technologies). Afterwards, coverslips were counterstained for 10 min at RT with the intercalating dye Hoechst 33342 (Invitrogen) diluted 1:1000 in PBS. After two washing steps in PBS, coverslips were fixed on microscopic slides with Aqua-Poly/Mount (Polyscience Inc.). Images were acquired with a Leica SP5 confocal microscope using 20 × and 40 × objectives (N.A. 0.7, N.A. 1.3, respectively). Images were processed with Fiji software (ImageJ 1.51w) and analyzed by exploiting the cell counter plug-in. For determination of MRTF-A localization, cells were categorized in three categories as described before 26 : nuclear, cytoplasmic and both. For the category "both", only cells with the same amount of MRTF-A in nucleus and cytoplasm were counted. Primary antibodies used for immunohistochemistry: mouse anti-MRTF-A (G8, Santa Cruz) and chicken anti-GFP (ab13970, Abcam. Treatment with LATB and JASP. Previous to treatment with LATB (Abcam, #ab1442091) or JASP (Abcam, #ab141409), pIND20-MRTF-A-GFP control and CAP2 mutant cells were activated 24 h before with 333 ng/mL doxycycline. Cells were treated for 4 h with either 25 nM LATB, 25 nM JASP or an equal volume of DMSO as control. For FRAP assay, MEFs were pretreated for 1 h with 200 nM JASP. Immunoblot analysis. Immunoblots were performed as described before 17 . Briefly, generation of total protein from MEFs was conducted by homogenization of cells in lysis buffer containing protease inhibitor (Complete, Roche). G/F-actin ratio of MEF cells was determined with 'G-Actin/F-actin In Vivo Assay Biochem Kit' (#BK037, Cytoskeleton, Inc.) according to manufacturer's instructions. In brief, 2 × T75 flasks with 70-80% confluent cells were washed with 1 × PBS and detached from the surface using 2 mL of 0.25% Trypsin-EDTA (#25200-056, Gibco). Pelleted cells (5 min at 100×g, RT) were resuspended in 1.2 mL of LAS2 buffer at RT and lysed with 20 strokes in Dounce homogenizer. The lysate was incubated at 37 °C for 10 min, centrifuged at 350×g for 5 min at RT. G-and F-actin fractions were separated by centrifugation in Beckman SW 60 Ti rotor at 100,000×g for 60 min at 35 °C. F-actin fraction (pellet) was resuspended in equal volume of depolymerization buffer on ice. For SDS-PAGE, equal volumes of G-and F-actin was loaded.
Protein extracts were denatured at 95 °C in Laemmli buffer, separated by SDS-page and blotted onto a polyvinylidene difluoride membrane (Merck) by using a Wet/Tank Blotting System (Biorad). Membranes were blocked for 1 h and afterwards incubated with primary antibodies in blocking solution over night at 4 °C. As secondary antibodies, horseradish peroxidase (HRP)-conjugated antibodies (1:20,000, Thermo Fisher Scientific) were used and detected by chemiluminescence with ECL Plus Western Blot Detection System (GE Healthcare). Quantitative PCR. Total RNA from MEFs was isolated using peqGold Trifast (VWR) according to the manufacturer's instructions. To exclude DNA contamination, samples were treated with the TURBO DNA-free kit (Invitrogen), followed by reverse transcription using iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's protocol. Quantitative PCR (qPCR) was performed in the STEP-One Light cycler (ABI Systems) using the iTaq SYBR Green Supermix (Bio-Rad) for detection of target genes. Three technical replicates were averaged and normalized to GAPDH in order to determine mRNA levels. Relative changes were calculated using the ΔΔCt method.

Mitochondrial morphology. CTR and KO MEF cells were seeded in 8-well ibidi slides (Ibidi GmbH) at
a density of 7000 cells per well, stained with MitoTracker Deep Red FM (200 nM for 30 min at 37 °C; Invitrogen) and fixed with 4% paraformaldehyde for 20 min at room temperature. Images were acquired using a Leica DM6000 epi-fluorescence microscope (63 × objective), by using an excitation wavelength of 620 nm and detecting emission using a 670 nm filter (red). Mitochondrial shape was classified as described before 29 : (i) category I comprises cells with healthy, elongated and equally distributed mitochondria, which are organized in a tubular network; (ii) category II comprises cells with partially fragmented mitochondria, which are still distributed throughout the cytosol, (iii) category III comprises cells with completely fragmented mitochondria, accumulating around the nucleus. At least 500 cells were counted by an experimenter blinded to the genotype. Mitochondrial membrane potential. Mitochondrial membrane potential was analyzed by using the MitoPT TMRE Kit (ImmunoChemistry Technologies). CTR and KO MEF cells were seeded in 24-well plate (10,000 cells/well), treated or not with erastin (0.7 µM) for 16 h. Cells were stained with TMRE (0.2 µM) for 30 min at 37 °C and harvested for subsequent FACS analysis (excitation 488 nm, emission 690/50 nm). A decrease of TMRE fluorescence was representative of a loss of mitochondrial membrane potential. Data were collected from at least 5000 cells and four replicates per condition.

Mitochondrial superoxide formation. CTR and KO MEF cells, seeded in
Statistical analysis. Statistical analysis was performed using the Prism statistical analysis package (Graph-Pad Software). Data are expressed as mean ± standard error (SE) for all main figures and mean ± standard deviation (SD) for the supplementary figure. All experiments have been conducted in at least three independent experiments. For supplementary figure eight replicates per condition were used if not specified. Data sets followed a normal distribution and differences between groups were evaluated by either one-or two-way ANOVA followed by post-hoc Dunns' or Scheffé's test (Fig. 4C, Fig. S3A

CAP2 controls actin turnover and equilibrium in mouse fibroblasts.
To study the cellular function of CAP2 in mammalian cells, we generated immortalized mouse embryonic fibroblasts (MEFs) from two CAP2 −/− mice (termed KO) and two CAP2 +/+ control littermates (CTR) at embryonic day (E) 12.5. Immunoblots confirmed absence of CAP2 from both KO MEF lines (Fig. 1A). Due to the lack of specific antibodies suitable for immunocytochemistry, we examined subcellular CAP2 localization in MEFs by expressing a green fluorescent protein (GFP)-tagged CAP2 construct. GFP-CAP2 was homogenously distributed within the cytosol and largely absent from the nucleus (Fig. 1B). www.nature.com/scientificreports/ Since CAP2 has been implicated in actin regulation 12 , we first tested whether KO MEFs displayed actin defects. We therefore performed fluorescence recovery after photobleaching (FRAP) experiments in GFP-actinexpressing CTR and KO MEFs to determine actin turnover (Fig. 1C, Fig. S1; Movies S1-S2). In CTR MEFs, GFPactin rapidly recovered with a mean half-recovery time (t 1/2 ) of 14.73 ± 1.51 s (n = 20 cells from 3 independent experiments; Fig. 1D). Compared to CTR MEFs, fluorescence recovery was faster in KO MEFs (8.95 ± 0.95 s, n = 17/3, P < 0.01). Instead, the stable actin fraction that did not recover within 95 s was not different between CTR and KO MEFs ( Fig. 1E; CTR: 0.15 ± 0.02, KO: 0.14 ± 0.02 n = 20/3 and 17/3, respectively; P = 0.819). Hence, inactivation of CAP2 increased actin turnover, but did not alter the stable actin fraction. Treatment of CTR MEFs with jasplakinolide (JASP), a potent inducer of actin polymerization, drastically increased both t 1/2 (30.44 ± 7.34 s, n = 10/2, P < 0.001) as well as stable actin fraction (0.89 ± 0.02, n = 10/2, P < 0.001), thereby confirming suitability of the approach for determining actin turnover (Fig. 1D,E, Fig. S1, Movie S3). Next, we tested whether CAP2 was relevant for the equilibrium between G-actin and F-actin. To do so, we lysed MEFs in a buffer optimized to stabilize and maintain G-and F-forms of cellular actin, similar to previous studies 30 . This approach revealed an increase in the G-actin fraction in KO MFEs, suggesting a shift in the equilibrium towards G-actin (Fig. 1F,  Fig. S2). The G-actin increase in KO MEFs was not due to changes in total actin levels (Fig. 1G, Fig. S2). Together, CAP2 inactivation in MEFs caused faster actin turnover and a shift in the actin equilibrium towards G-actin, while it did alter total actin levels.
Next, we tested whether these actin changes were associated with cellular defects. We found that KO MEFs normally adhered to cell culture dishes and did not differ from CTR MEFs in size or solidity index that we calculated to assess cellular morphology (Fig. 1H  Further, real-time impedance measurements revealed no differences between CTR and KO MEFs in cell proliferation (Fig. S3A). Since previous studies revealed defects in mitochondrial morphology and function upon   [8][9][10]12,32 , we tested whether CAP2 inactivation affected mitochondrial morphology or function. We found that CAP2 inactivation did not affect metabolic activity, mitochondrial morphology, mitochondrial ROS production, lipid peroxidation or mitochondrial membrane potential ( Fig. S3B-G). Furthermore, oxidative stress induced by erastin treatment similarly changed metabolic and mitochondrial parameters in CTR and KO MEFs 33,34 . Together, morphology, proliferation, metabolic activity, mitochondrial function and response to oxidative stress were unchanged in KO MEFs. Hence, actin defects in KO MEFs were not associated with any obvious cellular defects.

CAP2 inactivation alters subcellular distribution of MRTF-A in mouse fibroblasts. MRTF-A
is a transcriptional coactivator that shuttles between the cytosol and the nucleus. This shuttling depends on G-actin, because G-actin binding is necessary for nuclear MRTF-A export and interferes with accessibility of its nuclear localization sequence 35,36 . Hence, the shift towards G-actin in KO MEFs might be associated with altered subcellular MRTF-A distribution. To test this, we generated CTR and KO MEF lines that stably expressed GFP-tagged MRTF-A (MRTF-A-GFP) and grouped MEFs into three categories: (1) MEFs with mainly nuclear MRTF-A-GFP (nuclear), (2) MEFs with mainly cytosolic MRTF-A-GFP (cytosolic) and (3) MEFs with equal MRTF-A-GFP levels in both compartments (equal), similar to previous studies 26 . In KO MEFs, MRTF-A-GFP distribution was different from CTR MEFs, as indicated by an almost sevenfold increase in the cytosolic MRTF-A-GFP fraction and a sevenfold decrease in the nuclear MRTF-A-GFP fraction ( Fig. 2A  www.nature.com/scientificreports/ were associated with altered MRTF-A localization, and we therefore hypothesized that CAP2 controls MRTF-A localization in an actin-dependent manner. 20 . Indeed, when compared to dimethyl sulfoxide (DMSO)-treated CTR MEFs, LATB increased the CTR MEF fraction with cytosolic MRTF-A-GFP localization and it substantially decreased the fraction with nuclear localization (Fig. 2D Apart from LATB, we tested JASP that stabilizes F-actin, reduces G-actin levels and that reportedly induced nuclear import of MRTF-A 20,38 . As expected, JASP doubled the fraction of CTR MEFs with nuclear MRTF-A-GFP localization and reduced the fraction with cytosolic localization (Fig. 2D,E; JASP: nuclear: 83.69 ± 2.61, cytosolic: 6.41 ± 1.05, equal: 9.90 ± 2.16, n = 9/3, P < 0.01). Similarly, JASP changed the subcellular MRTF-A-GFP distribution in KO MEF as indicated by a ninefold increased fraction with nuclear MRTF-A-GFP localization concomitant with a fourfold decreased cytosolic fraction (JASP: nuclear: 62.21 ± 6.32, cytosolic: 22.28 ± 5.95, equal: 15.50 ± 2.19, n = 9/3, P < 0.01). Notably, the subcellular MRTF-A distribution did not differ between CTR and KO MEFs upon JASP treatment (P = 0.237). Hence, pharmacologically induced reduction of G-actin levels restored MRTF-A localization in KO MEFs. These data suggested that CAP2 controlled MRTF-A localization in an actin-dependent manner.

CAP2 is dispensable for serum induced nuclear MRTF-A translocation. Serum stimulation
reportedly induced nuclear MRTF-A translocation in fibroblasts 21,23,39 . By exploiting MRTF-A-GFP expressing MEF lines, we next tested whether CAP2 was relevant for serum-induced nuclear import of MRTF-A. First, we starved MEFs in 0.3% fetal calf serum (FCS) for 48 h and, thereafter, determined nuclear translocation during the first six min of stimulation with 20% FCS by live-cell imaging. We restricted this analysis to CTR and KO MEFs of the 'cytosolic MRTF-A fraction' to avoid any false interpretation due to different MRTF-A localization before stimulation. As obvious from the movies and image sequences (Fig. 3A, Movies S4-S5), MRTF-A rapidly translocated into the nucleus in CTR and KO MEFs upon serum stimulation. Quantification of the latency of nuclear translocation revealed no difference between both groups ( Fig. 3B; (in s) CTR: 248.75 ± 31.25, KO: 185.00 ± 14.18, n = 12/3, P = 0.089). Next, we determined subcellular MRTF-A localization in all CTR and KO MEFs both upon 48 h of serum starvation and upon 10 min of serum stimulation. Compared to basal conditions (Fig. 2C), the CTR MEF fraction with nuclear MRTF-A localization was reduced by 40%, and the fraction with cytosolic MRTF-A was increased threefold upon starvation (Fig. 3C,D; nuclear: 40.90 ± 3.78, cytosolic: 34.73 ± 5.02, equal: 24.38 ± 3.63, n = 25/3, P < 0.001). As expected, serum stimulation induced a nuclear translocation of all MRTF-A in CTR MEFs, and we did not note CTR MEFs with predominantly cytosolic MRTF-A or equal localization in cytosol and nucleus (Fig. 3C,D; n = 9/3, P < 0.001). In contrast to CTR MEFs, serum starvation did not alter MRTF-A localization in KO MEFs (nuclear: 1.67 ± 1.07, cytosolic: 92.34 ± 2.39, equal: 6.00 ± 2.29, n = 15/3, P < 0.0543), while serum stimulation induced nuclear MRTF-A translocation in the majority of KO MEFs. However, unlike in FCS-stimulated CTR MEFs, we noted a fraction of 25% KO MEFs with equal localization of MRTF-A in cytosol and nucleus upon FCS stimulation (nuclear: 75.02 ± 3.79, equal: 24.98 ± 3.79, n = 9/3, P < 0.001). Upon serum stimulation, subcellular MRTF-A distribution was still different between CTR and KO MEFs (P < 0.05). In order to determine whether starvation and subsequent serum stimulation affects subcellular localization of CAP2, we overexpressed GFP-CAP2 in CTR MEF cells. Unlike MRTF-A, the cytoplasmic localization of GFP-CAP2 was not changed under those conditions (Fig. S4). Together, serum stimulation induced nuclear MRTF-A translocation in KO MEFs with a latency similar to CTR MEFs. However, different from CTR MEFs, in a quarter of serum-stimulated KO MEFs MRTF-A was still present in the cytosol.

Discussion
The present study aimed at deciphering the CAP2-dependent mechanism relevant for the control of the transcription factor SRF. We chose mouse embryonic fibroblasts (MEFs) as a cellular model system for our study, because in these cells CAP2 is expressed at substantial levels and SRF-dependent gene regulation and upstream regulatory mechanisms have been intensively studied 21 . We found a shift in the actin equilibrium towards G-actin in CAP2 mutant MEFs, which was associated with a reduction in nuclear MRTF-A, reduced SRF activity and decreased expression of established MRTF-SRF target genes. While drug-induced increase in G-actin levels altered MRTF-A localization in control MEFs, it did not affect MRTF-A localization in mutant MEFs, which was normalized upon drug-induced decrease in G-actin levels. These data let us propose a model in which CAP2 controls SRFdependent gene expression via regulating G-actin levels and nuclear MRTF-A localization.
This model is in good agreement with recent studies that identified important functions for CAPs in controlling F-actin dynamics [5][6][7][8][9][10][11][12] . Specifically, these studies showed that CAPs can (i) facilitate F-actin severing and actin subunit dissociation in synergy with ADF/cofilin and twinfilin [5][6][7][8][9]11 , (ii) catalyze ATP-for-ADP exchange on G-actin that is relevant for F-actin assembly 6,[40][41][42] , and (iii) inhibit activity of INF2 that promotes F-actin assembly 10 . Hence, by regulating various aspects of F-actin assembly and disassembly, CAPs control G-actin levels and, hence, interaction of G-actin with MRTF-A. Our finding of increased G-actin levels and reduced nuclear MRTF-A localization in CAP2 mutant MEFs, together with normalization of MRTF-A localization upon treatment with the F-actin stabilizing drug excluded that CAP2 acts primarily as a F-actin disassembly factor in MEFs.
We found reduced SRF activity in reporter assays as well as reduced expression of SRF target genes in CAP2 mutant MEFs. Supportively, reduced expression of MRTF-SRF targets in CAP2 mutant MEFs have been noted recently by others 19 . Although SRF activity has not been systematically analyzed in skeletal muscle from CAP2 mutant mice, decreased mRNA levels of established MRTF-SRF targets such as Acta1, Acta2 and Actc1 point towards reduced SRF activity in skeletal muscles from CAP2 mutant embryos, too 17 , in which SRF dysregulation may contribute to retarded skeletal muscle development. Indeed, a crucial function for the MRTF-SRF pathway during late embryonic skeletal muscle development is evident from gene-targeted mice 43,44 . Opposite to our findings in CAP2 mutant MEFs and to reduced expression of MRTF-SRF target genes in skeletal muscles from CAP2 mutant embryos 17 , a recent study reported upregulation of several MRTF-SRF target genes including Acta1 and Acta2 in heart tissue and isolated cardiomyocytes from CAP2 mutant mice, which was associated with increased nuclear MRTF levels. Interestingly, heart defects in CAP2 mutant mice including dilated cardiomyopathy and impaired cardiac conductance were partially restored upon pharmacological inhibition of MRTF-SRF activity, demonstrating that CAP2-dependent regulation of the MRTF-SRF pathway is physiologically relevant and that its dysregulation due to CAP2 inactivation contributes to pathological conditions 19 . The opposite effects of CAP2 inactivation on MRTF-SRF activity suggest different CAP2 activities towards F-actin dynamics in MEFs versus cardiomyocytes. Unlike in MEFs and presumably in skeletal muscle, CAP2 may primarily act as a F-actin disassembly factor in cardiomyocytes.
By chromatin immunoprecipitation combined with deep sequencing, previous studies convincingly demonstrated that serum-induced, SRF-mediated transcriptional response largely depends on the MRTF-SRF pathway 21 . In line with these data, we showed efficient MRTF-A translocation into the nucleus as well as elevated SRF activity upon serum stimulation in control MEFs. Interestingly, serum-stimulated nuclear MRTF-A translocation as well as SRF activation was similar to controls in CAP2 mutant MEFs. Hence, while we found a role for CAP2 in MRTF-A localization and SRF activity under basal conditions, in unstimulated MEFs, CAP2 was dispensable for serum stimulation of the MRTF-SRF pathway.
Via acting on transmembrane receptors including G-protein-coupled receptors, receptor tyrosine kinases or serine-threonine receptor kinases, extracellular signals are translated into intracellular signaling cascades that include Rho family small guanosine triphosphatases (GTPases) [45][46][47] . Effectors of Rho GTPases include, among others, formins, Wiskott-Aldrich syndrome protein (WASP), WASP-family verprolin homologues (WAVEs) and actin-related protein 2/3 (ARP2/3) complex, which orchestrate actin polymerization 48 . Hence, Rho GTPase signaling promotes incorporation of G-actin into filaments that releases MRTF from G-actin complexes and stimulates MRTF-SRF-dependent gene expression 20 . Rho GTPase signaling further shifts the F/G-actin equilibrium towards F-actin via activation of Rho-associated kinases (ROCKs) that in turn inhibits actin depolymerizing proteins of the ADF/cofilin family 49 . ADF/cofilin cooperates with CAPs in actin dynamics 5,6,8,9,12,50 , and inhibition of ADF/cofilin activity upon serum simulation may explain why CAP2 was dispensable for serum-induced stimulation of the MRTF-SRF pathway. In summary, our data revealed that CAP2 controls the subcellular localization of MRTF and thereby SRF activity in unstimulated MEFs, while CAP2 was dispensable for serum-induced nuclear MRTF translocation and MRTF-SRF stimulation.