RPEL Proteins Are the Molecular Targets for CCG-1423, an Inhibitor of Rho Signaling

Epithelial–msenchymal transition (EMT) is closely associated with cancer and tissue fibrosis. The nuclear accumulation of myocardin-related transcription factor A (MRTF-A/MAL/MKL1) plays a vital role in EMT. In various cells treated with CCG-1423, a novel inhibitor of Rho signaling, the nuclear accumulation of MRTF-A is inhibited. However, the molecular target of this inhibitor has not yet been identified. In this study, we investigated the mechanism of this effect of CCG-1423. The interaction between MRTF-A and importin α/β1 was inhibited by CCG-1423, but monomeric G-actin binding to MRTF-A was not inhibited. We coupled Sepharose with CCG-1423 (CCG-1423 Sepharose) to investigate this mechanism. A pull-down assay using CCG-1423 Sepharose revealed the direct binding of CCG-1423 to MRTF-A. Furthermore, we found that the N-terminal basic domain (NB) of MRTF-A, which acts as a functional nuclear localization signal (NLS) of MRTF-A, was the binding site for CCG-1423. G-actin did not bind to CCG-1423 Sepharose, but the interaction between MRTF-A and CCG-1423 Sepharose was reduced in the presence of G-actin. We attribute this result to the high binding affinity of MRTF-A for G-actin and the proximity of NB to G-actin-binding sites (RPEL motifs). Therefore, when MRTF-A forms a complex with G-actin, the binding of CCG-1423 to NB is expected to be blocked. NF-E2 related factor 2, which contains three distinct basic amino acid-rich NLSs, did not bind to CCG-1423 Sepharose, but other RPEL-containing proteins such as MRTF-B, myocardin, and Phactr1 bound to CCG-1423 Sepharose. These results suggest that the specific binding of CCG-1423 to the NLSs of RPEL-containing proteins. Our proposal to explain the inhibitory action of CCG-1423 is as follows: When the G-actin pool is depleted, CCG-1423 binds specifically to the NLS of MRTF-A/B and prevents the interaction between MRTF-A/B and importin α/β1, resulting in inhibition of the nuclear import of MRTF-A/B.

We have recently shown that importin a/b1 and CRM1 (exportin1/XPO1) regulate the nuclear import and export, respectively, of Mycd family members [17], [18]. The N-terminal basic domain (NB) of Mycd family members (also known as B2 [2]), which is located between the second and third G-actinbinding RPEL motifs and is conserved among all members of Mycd family, is a binding site for importin a/b1 and functions as a nuclear localization signal (NLS). Actin dynamics do not affect the interaction between Mycd and importin a/b1 and the nuclear localization of Mycd, whereas G-actin significantly suppresses the interaction between MRTF-A/B and importin a/b1 and affects the nuclear import of MRTF-A/B [17], [19]. In the presence of G-actin, MRTF-A/B preferentially form a complex with G-actin owing to their high binding affinity for G-actin, resulting in blocking access of importin a/b1 to NB. The nuclear import of Phactr1, one of the other RPEL-containing proteins, is similarly regulated [20]. Phactr1 contains four RPEL motifs and two basic amino acid-rich NLSs, which are in close proximity to the RPEL motifs. Phactr1 nuclear accumulation is mediated by importin a/ b1. In resting cells, actin binding to the three C-terminal RPEL motifs inhibits the nuclear accumulation of Phactr1. This inhibition is due to the competitive binding of G-actin and importin a/b1 to NLSs associated with the N-and C-terminal RPEL motifs. CCG-1423 was originally identified as an inhibitor of RhoA signaling [21]. Although CCG-1423 has been reported to block the nuclear accumulation of MRTF-A [22], [23], the molecular mechanism is yet to be determined. Based on the cumulative evidence, we speculated that CCG-1423 directly inhibits MRTF-A binding to importin a/b1. In this study, we addressed this hypothesis and identified the inhibitory mechanism of this small molecule. These findings suggest a possible strategy(s) for anti-EMT drug discovery.

Plasmids
The construction of the plasmids used in this study, except for the mouse Nrf2 and rat Phactr1 expression plasmids, is described elsewhere [17], [18], [24]. In brief, each of cDNAs of mouse Mycd family members, mouse Nrf2 (NCBI Reference Sequence: NM_010902.3), and rat Phactr1 (NCBI Reference Sequence: NM_214457.2) were amplified by reverse transcription PCR and inserted into a mammalian expression plasmid, pCS2+, with a Flag tag at the N-termini. The sequences were confirmed.
Cell Culture and Immunocytochemistry NIH3T3 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Transfection of the indicated plasmids was performed using Trans IT-LT1 (PanVera Corporation, Madison, WI). The transfected cells were then cultured under the indicated conditions for 24 h. Immunocytochemistry was performed according to previously reported procedures [18]. The cells were incubated with phalloidin conjugated to Alexa Fluor 568 or the indicated primary antibodies followed by the specified secondary antibodies with Hoechst 33258. Fluorescent images were collected with the aid of a Biorevo BZ-9000 fluorescence microscope (Keyence, Osaka, Japan). The expression patterns of MRTF-A were categorized into three groups: nuclear-specific localization (N); diffuse distribution in the nucleus and the cytoplasm (NC), defined as equivalent immunostaining intensities of the target molecules in the cytoplasm and nucleus; and cytoplasmic localization (C). In each experiment, 100-200 cells were examined. The proportion of cells exhibiting the respective expression patterns is presented.

Promoter Assay
NIH3T3 cells were transfected with the indicated plasmids and cultured for 24 h. For further 20 h, the cells were cultured under serum-starved conditions and were re-stimulated with serum for 4 h. Cell extracts prepared using a passive lysis buffer (Promega) were subjected to luciferase assay with a luciferase assay kit (Promega). Relative promoter activity was expressed in luminescence units normalized to the b-galactosidase activity of pSVb-gal in the cell extracts. These assays were performed in triplicate and were repeated three times.

Protein-protein Interaction in vitro
All of the proteins used in this analysis were prepared using the TNT SP6 High-Yield Expression System based on an optimized wheat germ extract (Promega). We preliminarily confirmed that there was no significant protein cross-reaction between any of the antibodies against the indicated tag peptides, nuclear import proteins, and wheat germ extract for in vitro-translation and evaluated the expression levels of the respective in vitro-translated proteins by immunoblotting (IB) using previously specified antibodies [17], [18]. The composition of the immunoprecipitation (IP) buffer in this study was as follows: 20 mM Tris-HCl (pH 7.5), 0.5% Nonidet P-40, 150 mM NaCl, 1 mM EDTA, 50 mM NaF, 10 mM b-glycerophosphate, and proteinase inhibitors [complete Mini (Roche Applied Science)]). The IP buffer mixtures (total 500 ml) containing Flag-tagged MRTF-A proteins (15 or 30 ml), defined amounts of the indicated proteins [HAtagged importin a1 protein, 10 ml; importin b1 protein 10 ml; bactin R62D (unpolymerized mutant) protein, 20 ml], and CCG-1423 (10 mM) or vehicle [dimethyl sulfoxide (DMSO)] were subjected to IP analyses as previously described [17]. Proteins in the immunoprecipitates were detected by IB. Target proteins were detected with a SuperSignal chemiluminescence detection kit (Pierce, Rockford, IL). For IB analysis, 3.3% of the input proteins and 22.2% of the IP proteins were loaded on the input and IP lanes, respectively. Quantification of the respective IB signals' intensities was performed with the NIH ImageJ software. These interaction analyses were repeated three times.

Protein-protein Interaction in Cultured Cells
NIH3T3 cells were transfected with the expression plasmids for Flag-MRTF-A, HA-importin a1, and importin b1. The transfected cells were cultured under serum-stimulated conditions for 30 h. For the final 16 h, the cells were cultured in the presence of either 10 mM CCG-1423 or vehicle (DMSO), and then were restimulated with fresh serum for 15 min. Whole cell extracts were prepared by incubation with the IP buffer, followed by sonication. The whole cell extracts thus obtained were subjected to IP/IB analyses as described above.

Confirmation of Direct Binding of CCG-1423 to RPEL Proteins
Direct binding of CCG-1423 to each of Mycd family members and Phactr1 was examined by pull-down assay using CCG-1423 Sepharose. The indicated in vitro-translated proteins were purified using anti-Flag M2 affinity gel or anti-HA affinity matrix and were used as inputs. The IP buffer (total 400 ml) containing the indicated protein(s) (300 ng), 0.005% bovine serum albumin, CCG-1423 Sepharose or control Sepharose without CCG-1423 coupling (bed volume 25 ml), and free CCG-1423 (10 mM) or DMSO were incubated at 4uC for 2 h with gentle shaking. After washing with the IP buffer, followed by washing with phosphatebuffered saline, the pull-down proteins were detected by IB. Detection and quantification of target proteins were performed as described earlier. For IB analysis, 20% of the input proteins and 22% of the pull-down proteins were loaded on the input and the pull-down lanes, respectively. These interaction analyses were repeated three times.
Binding Assay of MRTF-A/B and Phactr1 to CCG-1423 Sepharose Using NIH3T3 Cell Whole Extracts NIH3T3 cells were transfected with each of the expression plasmids for Flag-MRTF-A, Flag-MRTF-B, and Flag-Phactr1 and were cultured under serum-stimulated conditions for 30 h. For the final 16 h, they were cultured under either serum-stimulated or serum-starved conditions. The serum-starved cells were further cultured for 10 min in the presence of 2 mM of LatB (an inhibitor of actin polymerization). The serum-stimulated cells were further re-stimulated with fresh serum for 15 min. Whole cell extracts were prepared by incubation with the IP buffer containing either 1 mM of Jasp (a stabilizer of F-actin) for serum-stimulated cells or 2 mM LatB for serum-starved cells followed by sonication. The resulting whole cell extracts were subjected to the binding assay using CCG-1423 Sepharose as described earlier.

Statistical Analysis
All graphs show means and standard errors. Statistical analysis was performed using Student's t-test.

CCG-1423 Treatment Inhibits Serum-induced Nuclear Accumulation of MRTF-A
We examined the effect of CCG-1423 on the subcellular localization of exogenously expressed Flag-MRTF-A in NIH3T3 cells under serum-starved and serum-stimulated conditions ( Figure  S1A, B). In majority (51.967.7%) of the cells expressing Flag-MRTF-A under serum-starved conditions, the protein was primarily observed in the cytoplasm. In contrast, in a large proportion (66.460.7%) of serum-stimulated cells, Flag-MRTF-A protein accumulated primarily in the nucleus. CCG-1423 treatment markedly reduced (17.561.6%) the proportion of cells showing the nuclear accumulation of the protein. In majority (53.265.0%) of the cells treated with CCG-1423, the protein was evenly distributed in the cytoplasm and nucleus. In correspondence with these changes, the protein's transactivation ability for the SM22a promoter also reduced in CCG-1423-treated cells ( Figure S1C). Similar to that of exogenously expressed Flag-MRTF-A, serum-induced nuclear accumulation of endogenous MRTF-A was inhibited by CCG-1423 ( Figure S1D, E). These results suggest that CCG-1423 treatment inhibits serum-induced nuclear import of MRTF-A.

CCG-1423 Inhibits MRTF-A Binding to Importin a/b1
We speculated that CCG-1423 directly inhibits the nuclear import of MRTF-A. To test this hypothesis, we examined the effect of CCG-1423 on the interaction between MRTF-A and importin a/b1 in vitro. In the presence of CCG-1423, the binding of MRTF-A to importin a/b1 markedly reduced, but the formation of importin ab1 heterodimer did not reduce ( Figure 1A, lane IPB). G-actin binding to MRTF-A was also unaffected by CCG-1423 ( Figure 1B, lane IPB). These results suggest a possibility that CCG-1423 binds directly to MRTF-A and prevents the interaction between MRTF-A and importin a/ b1 but does not impair the function of RPEL motifs as the Gactin-binding sites. To confirm the in vitro results, we examined the inhibitory effect of CCG-1423 on the interaction between MRTF-A and importin a/b1 in culture cells. This interaction was detected in DMSO-treated cells, but was inhibited in CCG-1423treated cells ( Figure 1C).

Direct Binding of CCG-1423 to MRTF-A Mediated by the NB
We covalently coupled Sepharose with CCG-1423 (CCG-1423 Sepharose) using a photo-crosslinking agent ( Figure 2) and performed a pull-down assay using the CCG-1423 Sepharose to examine MRTF-A binding to CCG-1423 ( Figure 3A, B). Because MRTF-A is associated with various proteins including G-actin, SRF, Smad, and other protein factors in cells, these protein factors may affect the interaction between MRTF-A and CCG-1423 Sepharose. In these assays, to rule out this possibility, in vitrotranslated Flag-tagged proteins were purified using an anti-Flag M2 affinity gel and were used as inputs ( Figure 3A, B, left columns). Wild-type MRTF-A protein clearly bound to CCG-1423 Sepharose, but such binding was severely reduced by free CCG-1423 (20.666.3% of the binding level in the absence of free CCG-1423) ( Figure 3A, middle and right columns). However, this protein did not bind to a control Sepharose without CCG-1423 coupling ( Figure 3B, middle and right columns). An MRTF-A protein with mutation in NB (MRTF-A NBmut), in which the NB sequence KLKRAR was mutated to ALAAAR, exhibited a low binding level (11.268.5% of the wild-type protein level) ( Figure 3B, middle and right columns). These results strongly suggest that CCG-1423 binds specifically and directly to MRTF-A under mediation by NB. The basic amino acids in the NB sequence ( Figure 3B) play a critical role in CCG-1423 binding to NB.
Because NB is in close proximity to the G-actin-binding RPEL motifs (Figure 7), we predicted that CCG-1423 binding to NB competes with G-actin binding to RPEL motifs. We addressed this possibility using purified MRTF-A and b-actin R62D (G-actin) proteins. In the presence of G-actin, MRTF-A binding to CCG-1423 Sepharose reduced, but G-actin was clearly detected in the bound fraction ( Figure 3C, left column). G-actin did not bind solely to CCG-1423 Sepharose ( Figure 3C, right column). Taken together with the data shown in Figure 1B, CCG-1423 does not inhibit G-actin binding to MRTF-A, suggesting that in the anti-HA-affinity matrix, or anti-Flag M2 affinity gel in the presence of either CCG-1423 (+) or vehicle (DMSO; 2), and the resulting immunoprecipitates were analyzed by immunoblotting (IB) with the indicated antibodies. Positions of molecular weight markers are indicated on the side of IB panels in kilodaltons (left columns). Control experiments with the control gel (cntl) showed no significant signals on IB (lanes IPC). The respective immunoprecipitation (IP)/IB signal intensities were quantified as described in Materials and Methods (right columns). The percentage values indicate the relative levels of MRTF-A binding to importin a1/b1 (A) or G-actin binding to MRTF-A (B) normalized by the binding of MRTF-A or G-actin in the absence of CCG-1423, which were set at 100% (mean 6 s.e.m of the results from three independent experiments). (C) Inhibitory effect of CCG-1423 on the interaction between MRTF-A and importin a/b1 in cultured cells. NIH3T3 cells were transfected with the expression plasmids as described in Materials and Methods. Whole cell extracts from the cells re-stimulated with serum were subjected to IP/IB analysis as described earlier.
Representative data are shown (n = 3). doi:10.1371/journal.pone.0089016.g001  [26]. Activated CH Sepharose 4B beads were coupled with the photoaffinity linker, and the beads were treated with 1 M ethanolamine (pH 11) to block the remaining reactive groups. The Sepharose beads with photoaffinity linker were agitated with 50 mM Tris-HCl (pH 7.4) buffer containing 0.1 mM CCG-1423, and then were exposed to UV light for 1 h. CCG-1423 was randomly coupled with the photoaffinity linkers on Sepharose by UV irradiation. The CCG-1423 Sepharose was washed with methanol and dried. doi:10.1371/journal.pone.0089016.g002 presence of G-actin, MRTF-A preferentially forms a complex with G-actin because of its high binding affinity for G-actin and results in inhibition of CCG-1423 binding to MRTF-A. Thus, G-actinfree MRTF-A rather than MRTF-A associated with G-actin is the more likely CCG-1423 target protein.
To further address the effects of actin dynamics on MRTF-A binding to CCG-1423 Sepharose, we performed the CCG-1423 binding assay using whole cell extracts from NIH3T3 cells expressing Flag-MRTF-A cultured under different conditions where either cellular F-actin or G-actin levels increased (Figure 4). The effects of Jasp and LatB on cellular F-actin levels in NIH3T3 cells cultured under serum-starved or serum-stimulated conditions were shown ( Figure 4A). Treatment with Jasp increased F-actin staining. In contrast, treatment with LatB markedly decreased F-actin staining. Significant binding of Flag-MRTF-A to CCG-1423 Sepharose was detected only in the cell extracts from F-actin-rich culture conditions ( Figure 4B, middle panel). These binding properties coincided well with the results of in vitro binding assays shown in Figure 3C. The competitive inhibitory effect of free CCG-1423 was also observed in the CCG-1423 binding assay using whole cell extracts ( Figure 4B, lower panel).

Binding Specificity of CCG-1423
We then investigated whether CCG-1423 binds specifically to NLS of MRTF-A. It has been reported that the nuclear import of Nrf2, a transcription factor essential for antioxidant response element-mediated gene expression, is mediated by three distinct basic amino acid-rich NLSs ( Figure 5A) and importin a/b1 [28]. We confirmed that Nrf2 forms a complex with importin a/b1 ( Figure 5B). Although the sequences of Nrf2 NLSs are rich in basic amino acids ( Figure 5A), significant binding of Nrf2 to CCG-1423 Sepharose was not observed ( Figure 5C). Furthermore, the pulldown assay showed that importin a/b1 did not bind to CCG-1423 Sepharose ( Figure 5D). These results suggest that CCG-1423 does not bind to any protein with a basic amino acid-rich NLS.
We addressed the binding specificity of CCG-1423 to other Mycd family members (MRTF-B and Mycd) and one of other RPEL containing proteins (Phactr1). Figure 6A shows the sequences of NLSs of Mycd family members and Phactr1. The sequence of NLS of the Mycd family (NB) is conserved among all members from different species and is located between the second and third RPEL motifs [17]. The conserved amino acids of NLS across Mycd family members and Phactr1 were highlighted. Similarly, Phactr1 C-terminal NLS is located between the third and fourth RPEL motifs [20]. We performed a pull-down assay using CCG-1423 Sepharose to examine the binding of respective RPEL-containing proteins to CCG-1423 ( Figure 6B). In these assays, in vitro-translated Flag-tagged proteins were purified using an anti-Flag M2 affinity gel and were used as inputs. These analyses revealed that MRTF-B, Mycd, and Phactr1 bound to CCG-1423 Sepharose. Bindings of Flag-MRTF-B and Phactr1 to CCG-1423 Sepharose were also observed in the binding assay using whole cell extracts ( Figure S2). The binding of mutant MRTF-B protein with mutation in NB (MRTF-B NBmut) to CCG-1423 Sepharose severely reduced, suggesting that CCG-1423 also binds to MRTF-B under mediation by NB ( Figure S3A). We then

Discussion
CCG-1423, which was originally identified as an inhibitor of RhoA signaling [21], is thought to be the MRTF-A inhibitor because CCG-1423 reduces cell growth and migration and blocks the nuclear accumulation of MRTF-A [22], [23]. However, the mode of inhibitory action is yet to be determined. In this study, we addressed our hypothesis that CCG-1423 directly inhibits MRTF-A binding to importin a/b1. Our novel findings are as follows: (1) CCG-1423 described in Materials and Methods: nuclear-specific localization (N), diffuse distribution in the nucleus and the cytoplasm (NC), and cytoplasmic localization (C) (lower panel). Asterisks indicate differences from the values under serum re-stimulated conditions without CCG-1423 in the respective localization categories (*P = 0.0002 and **P = 0.0002). doi:10.1371/journal.pone.0089016.g006 inhibits the interaction between MRTF-A and importin a/b1 but not G-actin binding to MRTF-A, (2) A pull-down assay using CCG-1423 Sepharose revealed direct and specific binding of CCG-1423 to MRTF-A. Furthermore, the functional NLS of MRTF-A (NB) is the binding site for CCG-1423, (3) In the presence of G-actin, MRTF-A preferentially forms a complex with G-actin rather than CCG-1423 because of its high binding affinity for G-actin, indicating competitive binding of G-actin and CCG-1423 to the N-terminal region of MRTF-A containing three RPEL motifs and NB, but it remains elusive whether or not all basic amino acid rich NLS bind to CCG-1423, and (4) CCG-1423 is expected to specifically bind to the NLSs of RPEL-containing proteins such as Mycd family members and Phactr1. These results suggest that CCG-1423 prevents the interaction between MRTF-A and importin a/b1 by masking NB, resulting in inhibition of the nuclear import of MRTF-A and that Gactin-free MRTF-A is the more likely CCG-1423 target protein.
These molecular mechanisms are schematically summarized in Figure 7. A similar inhibitory action is expected to be applicable to the interaction between MRTF-B or Phactr1 and importin a/b1.
CCG-1423 inhibits the interaction between MRTF-A and importin a/b1 ( Figure 1A). We demonstrated that CCG-1423 binds directly and specifically to MRTF-A under mediation by NB ( Figure 3B) and that the basic amino acids in the NB sequence ( Figure 7) play a critical role in CCG-1423 binding to NB. Because the sequences of NBs of Mycd family members are identical ( Figure 6A), CCG-1423 is expected to bind to each of the NB sequences of MRTF-B and Mycd. Actually, we demonstrated that CCG-1423 binds to MRTF-B under mediation by NB ( Figure  S3A). CCG-1423 also binds to Phactr1. Although we have not identified the binding site, CCG-1423 is expected to bind to Phactr1 C-terminal NLS (KREIKRR) because this NLS is also located between two RPEL motifs. However, CCG-1423 does not simply recognize a cluster of basic amino acids because CCG-1423 scarcely binds to Nrf2, in which three distinct basic amino acidrich NLSs are present ( Figure 5). CCG-1423 has a strict affinity for a specific sequence and/or tertiary protein structure. Further study is necessary to reveal the binding specificity of CCG-1423. Another possibility is that CCG-1423 inhibits the function of importin a/b1 in the nuclear import machinery. However, this possibility is less likely because importin a/b1 does not bind to CCG-1423 Sepharose ( Figure 5D).
We demonstrated that G-actin-free MRTF-A is the more likely CCG-1423 target protein ( Figures 3C and 4). These results suggest that CCG-1423 immediately binds to MRTF-A under conditions where Rho-activation induces rapid depletion of the G-actin pool and prevents the interaction between MRTF-A and importin a/ b1 in living cells. In resting cells, MRTF-A forms a stable complex with G-actin, and this complex formation significantly suppresses the interaction between MRTF-A/B and importin a/b1 [17], [19]. Thus, CCG-1423 is effective only under conditions where the G-actin pool is depleted.
The Larsen group has most recently reported that CCG-1423 binds specifically to an unknown 24-kD protein in PC-3 cell lysates using tag-free photoaffinity probes [29], suggesting that another target of CCG-1423 exists. Scarce information is currently available about this protein; therefore, future study is required to clarify its function. In their study, high molecular weight proteins (.140 kD) were not detected. This result would be explained by the complex formation between MRTF-A and G-actin; CCG-1423 is less likely to bind to MRTF-A associated with G-actin. The nuclear accumulation of MRTF-A occurs transiently just after serum stimulation and thereafter nuclear MRTF-A is gradually exported to the cytoplasm. Re-stimulation with fresh serum induces the nuclear accumulation of MRTF-A again ( Figure S4).
In the cytoplasm, MRTF-A forms a stable complex with G-actin. The Larsen group probably used the proliferating PC-3 cell lysates. However, for the reasons stated above, they could not detect MRTF-A/B.
Our present findings provide a new strategy for anti-EMT drug discovery by focusing on the nuclear import of MRTF-A. Immobilization of small molecules on Sepharose or microplates using a photoaffinity reaction is an effective method for detection of small molecule-protein interactions. This system using CCG-1423 as the leading compound would be a useful tool for anti-EMT drug screening because non-specific binding to CCG-1423 Sepharose was not detected in our study (Figures 3 and 5). Furthermore, we are currently working to determine whether a high-throughput screening system could be established using a series of CCG-1423-related compounds immobilized on microarrays and purified MRTF-A protein with fluorescent tag.
In conclusion, CCG-1423 binds specifically to MRTF-A under mediation by the NB, resulting in inhibition of the interaction between MRTF-A and importin a/b1. However, this inhibitory action of CCG-1423 is restricted to the conditions where the Gactin pool is depleted. A similar inhibitory action is expected be applicable to the interaction between MRTF-B or Phactr1 and importin a/b1. Sepharose. An MRTF-B NBmut protein carries a mutation in NB, in which the NB sequence KLKRAR was mutated to ALAAAR. The pull-down assays were performed as described in the legend for . Bar = 20 mm. The images were quantified as described in Materials and Methods: nuclear-specific localization (N), diffuse distribution in the nucleus and the cytoplasm (NC), and cytoplasmic localization (C) (lower panel). Asterisks indicate differences from the values under serum re-stimulated conditions without CCG-1423 in the respective localization categories (*P = 4.024610 26 and **P = 0.0015). (TIF) Figure S4 Gradual nuclear export of MRTF-A under serum-stimulated conditions. NIH3T3 cells were cultured under serum-starved conditions for 20 h (serum2) and were then re-stimulated with 10% serum (serum+) for 15 min and 24 h, respectively. Twenty-four hours later, the cells were re-stimulated with fresh serum for 15 min (serum+24 h/serum+15 min). The cells were stained with anti-MRTF-A antibody (red). Bar = 20 mm.