SR Protein Kinases Regulate the Splicing of Cardiomyopathy-Relevant Genes via Phosphorylation of the RSRSP Stretch in RBM20

(1) Background: RNA binding motif 20 (RBM20) regulates mRNA splicing specifically in muscle tissues. Missense mutations in the arginine/serine (RS) domain of RBM20 lead to abnormal gene splicing and have been linked to severe dilated cardiomyopathy (DCM) in human patients and animal models. Interestingly, many of the reported DCM-linked missense mutations in RBM20 are in a highly conserved RSRSP stretch within the RS domain. Recently, it was found that the two Ser residues within this stretch are constitutively phosphorylated, yet the identity of the kinase(s) responsible for phosphorylating these residues, as well as the function of RSRSP phosphorylation, remains unknown. (2) Methods: The ability of three known SR protein kinases (SRPK1, CLK1, and AKT2) to phosphorylate the RBM20 RSRSP stretch and regulate target gene splicing was evaluated by using both in vitro and in vivo approaches. (3) Results: We found that all three kinases phosphorylated S638 and S640 in the RSRSP stretch and regulated RBM20 target gene splicing. While SRPK1 and CLK1 were both capable of directly phosphorylating the RS domain in RBM20, whether AKT2-mediated control of the RS domain phosphorylation is direct or indirect could not be determined. (4) Conclusions: Our results indicate that SR protein kinases regulate the splicing of a cardiomyopathy-relevant gene by modulating phosphorylation of the RSRSP stretch in RBM20. These findings suggest that SR protein kinases may be potential targets for the treatment of RBM20 cardiomyopathy.


Introduction
RNA binding motif 20 (RBM20) is a splicing factor that is highly expressed in heart muscles [1][2][3]. RBM20 is the major regulator of alternative splicing of the Ttn gene [1,2,4,5], which encodes the giant sarcomeric protein titin. Titin is a functionally pleiotropic protein that serves as a molecular blueprint for the maintenance of sarcomere integrity and force transduction [6,7], a molecular spring that defines muscle stiffness [8,9], and a molecular signaling mediator for muscle hypertrophy and protein quality control [10][11][12]. Beyond Ttn, RBM20 has been shown to regulate the splicing of over 30 genes, including the contractile gene myosin heavy chain 6 (Myh6), as well as calcium-handling genes such as ryanodine receptor 2 (Ryr2) and calcium/calmodulin-dependent protein kinase type II d (Camk2d) [2,5,[13][14][15]. Genetic ablation of Rbm20 results in aberrant gene splicing and dilated cardiomyopathy (DCM) in rodents [2,16]. Furthermore, recent studies have shown that

In Vitro Kinase Assays
In vitro profiling of the kinase panel was performed at Reaction Biology Corporation, using the "HotSpot" assay platform. Briefly, 5 µM purified RBM20 protein [24,26] was incubated with 10 nM AKT2, 12.5 nM CLK1, 0.8 nM SRPK1, 1.5 nM CLK2, or 0.1 nM SRPK2 in reaction buffer (20 mM HEPES pH 7.5, 10 mM MgCl 2 , 1 mM EGTA, 0.02% Brij35, 0.02 mg/mL BSA, 0.1 mM Na 3 VO 4 , 2 mM DTT, and 1% DMSO). To assess the activity of different kinases, the same amount of AKT2, CLK1, SRPK1, CLK2, or SRPK2 was incubated with or without a standard substrate (crosstide for AKT2, MBP for CLK1 and CLK2, and RS peptide for SRPK1 and SRPK2) in the same reaction buffer. Compounds were delivered into the reaction, followed 20 min later by the addition of a mixture of non-radioactive ATP (Sigma-Aldrich) and γ-33 P ATP (PerkinElmer, Waltham, MA, USA) to a final concentration of 10 µM. Reactions were carried out at 25 • C for 120 min, followed by spotting of the reactions onto P81 ion exchange filter paper (Sigma-Aldrich). Unbound phosphate was removed by extensive washing of filters in 0.75% phosphoric acid. The phosphorylated signal for each reaction was calculated in nM [45].

Middle-Down Mass Spectrometry (MS)
Following in vitro kinase assays, RBM20 phosphorylation was assessed by using middle-down MS as previously described [26]. Briefly, RBM20 was digested and subjected to online liquid chromatography (LC)-MS and tandem MS (MS/MS) analyses [46,47], as well as offline high-resolution MS/MS analysis of phosphorylated peptides [48]. The online LC-MS and MS/MS data were processed and analyzed by using DataAnalysis software from Bruker Daltonics (Billerica, MA, USA). Mass spectra were deconvoluted by using the Maximum Entropy algorithm in the DataAnalysis software. For online MS/MS data, the output from the DataAnalysis software was analyzed by using MS-Align + [48,49] to identify RBM20 peptides. The cutoff of E-and p-values was set to E-10 to ensure the confident identification of peptides. For offline MS/MS data, the mass and charge lists for fragment ions were output from the DataAnalysis software for peptide sequence identification, using MS-Align+. In-house developed MASH Suite Pro [50,51] was used for the manual validation of fragment ion assignment and the localization of phosphorylation sites. A minimum fit of 60% and S/N threshold of 3 were set for peak picking. Fragment ions, including c, c −1 , z • , and z •+1 ions, were validated within 10 ppm mass error. All reported masses are the monoisotopic masses.

Co-Immunoprecipitation (co-IP)
Co-IPs were carried out by using a Pierce Co-IP kit (Pierce, IL, USA), as per the manufacturer's instructions. Briefly, 100 µg of purified antibodies was coupled with resin. Protein samples (1 mg) were incubated with the antibody-coupled resin for 2 h. Proteinantibody complexes were eluted in 50 µL elution buffer after mixing and washing. The eluted protein samples were subjected to immunoblotting [52].

Data Analysis
GraphPad Prism software was used for statistical analysis. Results are presented as mean ± SEM. Statistical significance for each variable was estimated by the paired t-test (two-tailed). Significance was considered as probability values of p < 0.05 indicated by one asterisk, p < 0.01 indicated by two asterisks, and p < 0.001 indicated by three asterisks.

Co-Transfection with AKT2, CLK1, or SRPK1 Increases Phosphorylation of the RBM20 RSRSP Stretch in HeLa Cells
Kinases belonging to the AKT, CLK, and SRPK kinase families have previously been shown to phosphorylate the RS domain in splicing factors [31][32][33]. To test whether kinases belonging to these families can also phosphorylate the RSRSP stretch in the RS domain of RBM20, plasmids carrying AKT2, CLK1, and SRPK1 were co-transfected with either WT or mutant RBM20 constructs in HeLa cells. HeLa cells were chosen because these cells do not express RBM20, thus allowing for the avoidance of confounding signals from endogenous protein. Co-transfection of WT or mutant RBM20 with empty plasmid (no kinase) was used as a control. Cells were harvested, and lysates were prepared 48 h post-co-transfection. Proteins from each treatment were subjected to Western blot analysis, using anti-phospho-RBM20 (gift from Dr. Hidehito Kuroyanagi) [40], which recognizes phosphorylation of the RSRSP stretch located in the RS domain of RBM20, and anti-pan-RBM20 (home-made) antibodies [2]. The Western blot analysis confirmed the increased expression of AKT2, CLK1, and SRPK1 in HeLa cells transfected with these kinases relative to cells transfected with empty plasmid ( Figure 1A,C,E). Phosphorylation of the RSRSP stretch in WT RBM20 was significantly increased by co-transfection with all three kinases when compared to that in cells co-transfected with empty plasmid (Figure 1). Furthermore, co-transfection with each of the three kinases increased phosphorylation of the mutated RSRSP stretch in S638A and S640G RBM20, with the exception that phosphorylation of S638A RBM20 did not increase significantly in response to co-transfection with AKT2 Genes 2022, 13, 1526 5 of 15 ( Figure 1). As expected, phosphorylation of the double mutant (S638A/S640G), which lacks the two phosphorylatable Ser residues in the RSRSP stretch, was not detected ( Figure 1). Collectively, these data demonstrate that co-transfection with AKT2, CLK1, or SRPK1 increases the phosphorylation of the RSRSP stretch in RBM20-transfected HeLa cells. expression of AKT2, CLK1, and SRPK1 in HeLa cells transfected with these kinases relative to cells transfected with empty plasmid ( Figure 1A,C,E). Phosphorylation of the RSRSP stretch in WT RBM20 was significantly increased by co-transfection with all three kinases when compared to that in cells co-transfected with empty plasmid (Figure 1). Furthermore, co-transfection with each of the three kinases increased phosphorylation of the mutated RSRSP stretch in S638A and S640G RBM20, with the exception that phosphorylation of S638A RBM20 did not increase significantly in response to co-transfection with AKT2 ( Figure 1). As expected, phosphorylation of the double mutant (S638A/S640G), which lacks the two phosphorylatable Ser residues in the RSRSP stretch, was not detected ( Figure 1). Collectively, these data demonstrate that co-transfection with AKT2, CLK1, or SRPK1 increases the phosphorylation of the RSRSP stretch in RBM20-transfected HeLa cells. Figure 1. SR protein kinases phosphorylate serine residues in the RBM20 RSRSP stretch in HeLa cells. A, C, and E. HeLa cells were co-transfected with plasmids carrying the SR protein kinases AKT2 (A), CLK1 (C), or SRPK1 (E) and WT or mutant RBM20. Cell lysates were subject to immunoblotting with antibodies against pRBM20, RBM20, AKT2, CLK1, and SRPK1. Blots shown are representative of three independent experiments. Transfection of SR protein kinase plasmids increased the expression of the respective kinases significantly (B,D,F). Quantification of RBM20 phosphorylation in HeLa cells co-transfected with individual kinases or empty plasmid (control). Data are shown as mean ± SD (n = 3); * p < 0.05, ** p < 0.01. WT, wild type.

CLK1 and SRPK1 Directly Phosphorylate the RBM20 RS Domain In Vitro
While increased phosphorylation of RBM20 upon co-transfection with AKT2, CLK1, or SRPK1 in HeLa cells suggests a role for these kinases in the regulation of RBM20 phosphorylation, it is possible that the mode of regulation is indirect. Next, to determine whether AKT2, CLK1, and SRPK1 can directly phosphorylate RBM20, we performed in vitro kinase assays with γ-33 P-ATP and purified RBM20 ( Figure S2A). An analysis of 33 P incorporation confirmed that all three kinases phosphorylated RBM20 in vitro, with the lowest level of 33 P incorporation detected for RBM20 incubated with AKT2 (Figures 2A  and S2B). Although this result shows that all three kinases can directly phosphorylate Ratio of pRBM20 to total RBM20 Ratio of pRBM20 to total RBM20 135kDa 135kDa 92kDa Figure 1. SR protein kinases phosphorylate serine residues in the RBM20 RSRSP stretch in HeLa cells (A,C,E). HeLa cells were co-transfected with plasmids carrying the SR protein kinases AKT2 (A), CLK1 (C), or SRPK1 (E) and WT or mutant RBM20. Cell lysates were subject to immunoblotting with antibodies against pRBM20, RBM20, AKT2, CLK1, and SRPK1. Blots shown are representative of three independent experiments. Transfection of SR protein kinase plasmids increased the expression of the respective kinases significantly (B,D,F). Quantification of RBM20 phosphorylation in HeLa cells co-transfected with individual kinases or empty plasmid (control). Data are shown as mean ± SD (n = 3); * p < 0.05, ** p < 0.01. WT, wild type.

CLK1 and SRPK1 Directly Phosphorylate the RBM20 RS Domain In Vitro
While increased phosphorylation of RBM20 upon co-transfection with AKT2, CLK1, or SRPK1 in HeLa cells suggests a role for these kinases in the regulation of RBM20 phosphorylation, it is possible that the mode of regulation is indirect. Next, to determine whether AKT2, CLK1, and SRPK1 can directly phosphorylate RBM20, we performed in vitro kinase assays with γ-33 P-ATP and purified RBM20 ( Figure S2A). An analysis of 33 P incorporation confirmed that all three kinases phosphorylated RBM20 in vitro, with the lowest level of 33 P incorporation detected for RBM20 incubated with AKT2 (Figures 2A and S2B). Although this result shows that all three kinases can directly phosphorylate RBM20 in vitro, whether these kinases directly phosphorylate Ser residues within the RS domain could not be determined with this assay. To gain insight, phosphorylation of the RBM20 RS domain by the aforementioned kinases was assessed by using middle-down MS ( Figures 2B and S1). The MS analysis confirmed the phosphorylation of the peptide encompassing the RBM20 RS domain by CLK1 and SRPK1, but not AKT2 ( Figure 2B). Given that RBM20 phosphorylation by AKT2 was detected based on 33 P incorporation, this result may indicate that the AKT2-mediated phosphorylation of Ser residues within the RBM20 RS domain has low efficiency in vitro or, alternatively, that regulation of the RBM20 RS domain phosphorylation by AKT2 is indirect. Taken together, these results show that CLK1 and SRPK1 can directly phosphorylate the RBM20 RS domain in vitro. On the other hand, while AKT2 can directly phosphorylate other sites in RBM20, it appears that this kinase does not directly regulate the phosphorylation of Ser residues within the RBM20 RS domain, at least in vitro. RBM20 RS domain by the aforementioned kinases was assessed by using middle-d MS ( Figure 2B and Supplementary Figure S1). The MS analysis confirmed the phosp ylation of the peptide encompassing the RBM20 RS domain by CLK1 and SRPK1, bu AKT2 ( Figure 2B). Given that RBM20 phosphorylation by AKT2 was detected base 33 P incorporation, this result may indicate that the AKT2-mediated phosphorylation o residues within the RBM20 RS domain has low efficiency in vitro or, alternatively, regulation of the RBM20 RS domain phosphorylation by AKT2 is indirect. Taken toge these results show that CLK1 and SRPK1 can directly phosphorylate the RBM20 RS main in vitro. On the other hand, while AKT2 can directly phosphorylate other sit RBM20, it appears that this kinase does not directly regulate the phosphorylation o residues within the RBM20 RS domain, at least in vitro.

AKT2, CLK1, and SRPK1 Interact with RBM20 in Co-Transfected HeLa Cells and Reg Titin Pre-mRNA Splicing
Next, we sought to determine whether RBM20 interacts with AKT2, CLK1, SRPK1 in co-transfected HeLa cells, using co-immunoprecipitation (co-IP). Plasmid taining RBM20 with an 8xHis-tag was co-transfected with individual kinase construc HeLa cells. Cells were harvested 48 h after co-transfection, and protein lysates were pared and subjected to co-IP. An anti-His-tag antibody was used to capture RBM20 tein complexes, which were subsequently analyzed by Western blot, using anti-RB anti-SRPK1, anti-CLK1, and anti-HA-tag antibodies. The Western blot analysis reve that SRPK1 ( Figure 3A), CLK1 ( Figure 3B), and AKT2 ( Figure 3C) immunoprecipi with RBM20. In the reciprocal experiments, SRPK1, CLK1, and HA-tagged AKT2 used as bait to capture RBM20. Consistently, the IP of all three kinases also pulled d RBM20 ( Figure 3D-F). Notably, controls lacking antibody conjugated to the beads f to capture any of the target proteins ( Figure 3A-F).
To determine whether these three kinases can also regulate titin pre-mRNA spli WT or mutant RBM20 was co-transfected along with a titin minigene exon 64-70 cons and individual kinase constructs in HeLa cells. A schematic showing titin minigene s variants is displayed in Figure 3G, with the size of each variant indicated. Cells were vested 48 h after co-transfection, total RNA was isolated, and RT-PCR was performe detect changes in the splicing of the titin minigene construct. Three splice variants Next, we sought to determine whether RBM20 interacts with AKT2, CLK1, and SRPK1 in co-transfected HeLa cells, using co-immunoprecipitation (co-IP). Plasmid containing RBM20 with an 8xHis-tag was co-transfected with individual kinase constructs in HeLa cells. Cells were harvested 48 h after co-transfection, and protein lysates were prepared and subjected to co-IP. An anti-His-tag antibody was used to capture RBM20 protein complexes, which were subsequently analyzed by Western blot, using anti-RBM20, anti-SRPK1, anti-CLK1, and anti-HA-tag antibodies. The Western blot analysis revealed that SRPK1 ( Figure 3A), CLK1 ( Figure 3B), and AKT2 ( Figure 3C) immunoprecipitated with RBM20. In the reciprocal experiments, SRPK1, CLK1, and HA-tagged AKT2 were used as bait to capture RBM20. Consistently, the IP of all three kinases also pulled down RBM20 ( Figure 3D-F). Notably, controls lacking antibody conjugated to the beads failed to capture any of the target proteins ( Figure 3A-F).
To determine whether these three kinases can also regulate titin pre-mRNA splicing, WT or mutant RBM20 was co-transfected along with a titin minigene exon 64-70 construct and individual kinase constructs in HeLa cells. A schematic showing titin minigene splice variants is displayed in Figure 3G, with the size of each variant indicated. Cells were harvested 48 h after co-transfection, total RNA was isolated, and RT-PCR was performed to detect changes in the splicing of the titin minigene construct. Three splice variants were detected in HeLa cells transfected with WT RBM20 in the absence of kinases (NC), while only the largest variant was detected in cells co-transfected with individual kinases ( Figure 3H). In HeLa cells transfected with S638A-mutated RBM20 in the absence of kinases, only the smallest variant was expressed, whereas co-transfection with CLK1 or SRPK1, but not AKT2, shifted the expression pattern in such a way that expression of the two largest variants was favored; however, the smallest variant was still present, albeit at lower levels ( Figure 3H). Unlike in HeLa cells transfected with S638A RBM20, co-transfection of S640G or double S638A/S640G mutant RBM20 with CLK1, SRPK1, or AKT2 promoted expression of only the largest splice variant ( Figure 3H). These results demonstrate that CLK1, SRPK1, and AKT2 not only interact with RBM20 but can also regulate splicing of the RBM20 target gene, Ttn, in vitro. largest variants was favored; however, the smallest variant was still present, albeit at lower levels ( Figure 3H). Unlike in HeLa cells transfected with S638A RBM20, co-transfection of S640G or double S638A/S640G mutant RBM20 with CLK1, SRPK1, or AKT2 promoted expression of only the largest splice variant ( Figure 3H). These results demonstrate that CLK1, SRPK1, and AKT2 not only interact with RBM20 but can also regulate splicing of the RBM20 target gene, Ttn, in vitro.

(H). RT-PCR detection of
Ttn minigene splice variants after co-transfection with individual SR protein kinases and either WT or mutant RBM20. Input, total protein before co-IP; IP-His, elution from co-IP with anti-His-tag antibody; Control, elution from co-IP without antibody conjugation to the beads; NC, control without kinase transfection; WT, RBM20 wild type.

Ex66
Ex68 Ex67 RT-PCR detection of Ttn minigene splice variants after co-transfection with individual SR protein kinases and either WT or mutant RBM20. Input, total protein before co-IP; IP-His, elution from co-IP with anti-His-tag antibody; Control, elution from co-IP without antibody conjugation to the beads; NC, control without kinase transfection; WT, RBM20 wild type.

Inhibition of CLK and SRPK Family Kinases Reduces RBM20 RSRSP Phosphorylation and Leads to Aberrant Pre-mRNA Splicing in NRCMs
Next, we tested whether the inhibition of endogenous SRPK and CLK family kinases impacts RBM20 RSRSP phosphorylation in isolated NRCMs. NRCMs were isolated from 1-day-old rats and treated with either SRPIN340 or TG003-selective inhibitors of SRPK and CLK kinases, respectively. NRCMs were harvested at different time points (0 min, 5 min, 1 h, 6 h, 12 h, 24 h, or 48 h) following inhibitor treatment, and RBM20 phosphorylation was assessed by Western blot. In agreement with the results of our previous experiments, we detected a time-dependent decrease in the phosphorylation of the RSRSP stretch in RBM20 beginning 5 min after the initiation of inhibitor treatment ( Figure 4A,B,D). Conversely, phosphorylation of the RBM20 RSRSP stretch was not decreased following vehicle (DMSO) treatment ( Figure 4C,D). Ttn and Camk2d are well-established targets of RBM20 [3]. Inhibition of kinases SRPK1 and CLK1 increased the expression of larger variants of both Ttn (t-1) and Camk2d (d-1) when compared to the control DMSO treatment ( Figure 4E). Collectively, these results are in accordance with those of our previous experiments and verify that CLK and SRPK family kinases regulate phosphorylation of the RSRSP stretch within the RS domain of RBM20 and target splicing in vivo.
impacts RBM20 RSRSP phosphorylation in isolated NRCMs. NRCMs were isolated 1-day-old rats and treated with either SRPIN340 or TG003-selective inhibitors of S and CLK kinases, respectively. NRCMs were harvested at different time points (0 m min, 1 h, 6 h, 12 h, 24 h, or 48 h) following inhibitor treatment, and RBM20 phospho tion was assessed by Western blot. In agreement with the results of our previous ex ments, we detected a time-dependent decrease in the phosphorylation of the R stretch in RBM20 beginning 5 min after the initiation of inhibitor treatment (F 4A,B,D). Conversely, phosphorylation of the RBM20 RSRSP stretch was not decrease lowing vehicle (DMSO) treatment ( Figure 4C,D). Ttn and Camk2d are well-establ targets of RBM20 [3]. Inhibition of kinases SRPK1 and CLK1 increased the expressi larger variants of both Ttn (t-1) and Camk2d (d-1) when compared to the control D treatment ( Figure 4E). Collectively, these results are in accordance with those of our vious experiments and verify that CLK and SRPK family kinases regulate phospho tion of the RSRSP stretch within the RS domain of RBM20 and target splicing in vivo  Ttn variants, T1-T4; Camk2d variants, d1 and d2. GAPDH served as a loading control for Western blot and housekeeping gene for normalization of RT-PCR data. Graph shows mean ± SD (n = 3); * p < 0.05 compared between SRPIN340 and DMSO, and # p < 0.05 compared between TG003 and DMSO.

Phosphorylation of the RSRSP Stretch in RBM20 and Pre-mRNA Splicing of RBM20 Targets Are Altered in Akt2 KO and Transgenic Mice
Due to the lack of specific AKT2 inhibitors, we evaluated the role of AKT2 in RBM20 phosphorylation in Akt2 KO and overexpressing mice ( Figure 5A,B). Heart tissues collected from WT mice were used as a control. The Western blot analysis revealed that RBM20 phosphorylation was significantly reduced and increased in the hearts of Akt2 KO mice and Akt2 overexpressing mice, respectively, compared to WT (Figure 5A,C). Consistent with our previous finding that AKT regulates RBM20 expression via modulation of the PI3K/AKT/mTOR signaling pathway [31,32], the expression of RBM20 was reduced and increased in the hearts of Akt2 KO and overexpressing transgenic mice, respectively ( Figure 5A,D). The splicing of RBM20 target transcripts was also altered in the hearts of these mice ( Figure 5E). Specifically, the abundance of the larger d0 variant of Camk2d was increased in the hearts of Akt2 KO mice, while the abundance of Ttn variants was unchanged ( Figure 5E). Conversely, in the hearts of Akt2 overexpressing mice, the abundance of larger Ttn transcript variants (T1 and T2) was increased, whereas the Camk2d variant abundances did not differ ( Figure 5E). Collectively, these results provide additional evidence supporting a role for AKT2 in regulating phosphorylation of the RBM20 RSRSP stretch, either directly or indirectly, in vivo. These data also show that AKT2 regulates the splicing of RBM20 targets in the heart; however, whether this occurs through modulation of RBM20 phosphorylation, altered RBM20 expression, or both remains to be determined.

Overexpression of CLK1, SRPK1, or AKT2 Alone Does Not Facilitate RBM20 Nucleocytoplasmic Transport
It is well-established that certain SR proteins (e.g., SRSF1) move continuously between the nucleus and the cytoplasm, and that this shuttling is facilitated by hyperphosphorylation [53]. Since overexpression of CLK1, SRPK1, or AKT2 increases RBM20 RSRSP stretch phosphorylation in co-transfected HeLa cells (Figure 1), we sought to determine

Overexpression of CLK1, SRPK1, or AKT2 Alone Does Not Facilitate RBM20 Nucleocytoplasmic Transport
It is well-established that certain SR proteins (e.g., SRSF1) move continuously between the nucleus and the cytoplasm, and that this shuttling is facilitated by hyperphosphorylation [53]. Since overexpression of CLK1, SRPK1, or AKT2 increases RBM20 RSRSP stretch phosphorylation in co-transfected HeLa cells (Figure 1), we sought to determine whether increased phosphorylation of RBM20 promotes nucleocytoplasmic transport. HeLa cells were co-transfected with WT RBM20 and either CLK1, SRPK1, or AKT2, and the localization of RBM20 was assessed by immunocytochemistry 48 h post-co-transfection. Co-transfection of WT RBM20 with empty plasmid (no kinase) served as a control. As expected, RBM20 was localized to the nucleus of transfected HeLa cells without kinase overexpression. Interestingly, overexpression of CLK1, SRPK1, or AKT2 did not alter RBM20 localization ( Figure 6A-D), suggesting that hyperphosphorylation of RBM20 phosphorylation alone does not promote shuttling between the nucleus and cytoplasm.

Discussion
Herein, we demonstrated that phosphorylation of S638 (S635 in human) and S640 (S637 in human) in the RSRSP stretch of RBM20 is regulated by SRPK and CLK family kinases, as well as by AKT2. Specifically, we showed that these kinases interacted with and phosphorylated RBM20, using a combination of co-transfection experiments in HeLa

Discussion
Herein, we demonstrated that phosphorylation of S638 (S635 in human) and S640 (S637 in human) in the RSRSP stretch of RBM20 is regulated by SRPK and CLK family kinases, as well as by AKT2. Specifically, we showed that these kinases interacted with and phosphorylated RBM20, using a combination of co-transfection experiments in HeLa cells and in vitro kinase assays. Moreover, we showed that CLK1 and SRPK1 directly phosphorylated the RS domain in RBM20, and the inhibition of endogenous CLK and SRPK family kinases in NRCMs confirmed such regulation of RSRSP stretch phosphorylation in the heart. Conversely, altered RBM20 phosphorylation in Akt2 KO and overexpressing mice provided additional support for AKT2-mediated regulation of RBM20 RSRSP stretch phosphorylation in vivo; however, it remains unclear whether this regulation is direct or indirect. We also observed that SRPK1, CLK1, and AKT2 can regulate the pre-mRNA splicing of RBM20 target genes, which likely occurs via the modulation of RBM20 phosphorylation.
Alternative pre-mRNA splicing is a common phenomenon in higher eukaryotes that increases protein diversity and serves as an additional regulatory mechanism governing gene expression in different cell types and during development [32]. The SR protein family of splicing factors and SR protein kinases play important roles in alternative pre-mRNA splicing in different ways, such as through spliceosome assembly, splicing catalysis, and signaling transduction [32]. The RS domain in SR proteins mediates a protein-protein interaction network to facilitate cross-intron interactions, exon definition, splice site selection, and the eventual formation of the higher-order spliceosome [54,55]. Posttranslational modifications are critical regulators of the protein-protein interactions mediated by the RS domain(s) of SR proteins [56]. At least three well-known posttranslational modifications occur on SR proteins: methylation, acetylation, and phosphorylation [57][58][59][60]. Currently, the most well understood of these modifications is phosphorylation, which is regulated by multiple kinases belonging to the CMGC family of kinases [61], as well as AKT [62,63]. The regulation of SR protein phosphorylation in vivo by three families of kinases, namely AKTs, SRPKs, and CLKs, has been confirmed through either genetic ablation or chemical inhibition [63][64][65][66][67][68]. In this study, we determined whether members of each of these kinase families can also regulate the phosphorylation of RBM20. We showed that AKT2, SRPK1, and CLK1 regulate RBM20 phosphorylation both in vitro and in vivo. Furthermore, we provided evidence that these kinases can modulate titin pre-mRNA splicing, thus making them potential therapeutic targets for the treatment of diastolic dysfunction [69].
SR protein phosphorylation plays a myriad of roles in the splicing of target transcripts. Prior studies have shown that phosphorylation is required for spliceosome assembly, but that dephosphorylation is crucial for splicing catalysis [70][71][72][73]. A specific phosphorylation state in some SR proteins is important to function properly in splicing, and partial, rather than full, phosphorylation of certain SR proteins is required for splicing activity [74,75]. Whether RBM20 phosphorylation plays a role in assembly of the spliceosome or splicing catalysis will need to be investigated in future studies. Aside from splicing, SR protein phosphorylation is also important for nuclear import and nucleocytoplasmic shuttling [36,53,76]. Based on the data presented herein, it appears that increased phosphorylation of RBM20 through kinase overexpression does not promote nucleocytoplasmic transport of the protein as it does for the shuttling SR protein SRSF1 [53]. This finding is consistent with our recently published data showing that constitutive pseudo-phosphorylation of phosphorylation sites within the RS domain in RBM20 by Ser-to-Asp mutagenesis does not promote nucleocytoplasmic transport of the protein in H9c2 cells [26]. Nevertheless, a notable caveat is that there is evidence in the literature suggesting that SR protein shuttling is also dependent on RNA binding and HeLa and H9c2 cells to not express titin, the primary target of RBM20. Thus, confirmation of these results in cells that express titin (e.g., cardiomyocytes) is warranted. Additional studies will also be necessary to determine the role of RBM20 RS domain phosphorylation in nuclear import. Taken together, our results provide new information on RBM20 phosphorylation. Further studies deciphering the role of RBM20 phosphorylation in alternative splicing in muscle tissues may aid the development of new strategies for the treatment of cardiomyopathies.