AMP-activated protein kinase (AMPK) regulates cellular energy homeostasis by inhibiting anabolic processes and activating catabolic processes. Recent studies have demonstrated that metformin, which is an AMPK activator, modifies alternative precursor mRNA (pre-mRNA) splicing. However, no direct substrate of AMPK for alternative pre-mRNA splicing has been reported. In the present study, we identified the splicing factor serine/arginine-rich splicing factor 1 (SRSF1) as a novel AMPK substrate. AMPK directly phosphorylated SRSF1 at Ser133 in an RNA recognition motif. Ser133 phosphorylation suppressed the interaction between SRSF1 and specific RNA sequences without altering the subcellular localization of SRSF1. Moreover, AMPK regulated the SRSF1-mediated alternative pre-mRNA splicing of Ron, which is a macrophage-stimulating protein receptor, by suppressing its interaction with exon 12 of Ron pre-mRNA. The findings of this study revealed that the AMPK-dependent phosphorylation of SRSF1 at Ser133 inhibited the ability of SRSF1 to bind RNA and regulated alternative pre-mRNA splicing.
AMP-activated protein kinase (AMPK) is a serine/threonine kinase that regulates metabolic processes . AMPK is a heterotrimeric complex comprising a catalytic α subunit and regulatory β and γ subunits . Its various isoforms (α1, α2, β1, β2, γ1, γ2, γ3) can form multiple heterotrimeric combinations . Regarding AMPK activity, the phosphorylation of the α subunit at Thr172 in the activation loop of the kinase domain by the upstream kinases LKB1 [4–6] and CaMKKβ is essential [7–9]. Furthermore, the binding of AMP/ADP to the γ subunit of AMPK changes the structure of the AMPK complex, thereby suppressing Thr172 dephosphorylation by phosphatases and maintaining AMPK activation . AMPK responds to low energy levels by up-regulating catabolic pathways such as fatty acid oxidation via the phosphorylation of the ACC2 isoform of acetyl-CoA carboxylase (ACC) and inhibiting anabolic pathways such as fatty acid synthesis via the phosphorylation and inactivation of the ACC1 isoform .
The term alternative splicing describes the splicing of precursor mRNA (pre-mRNA) in more than one pattern for generating multiple mature mRNAs . It is a key mechanism that dramatically increases the proteomic diversity of the human genome . One of the regulators of alternative splicing is a conserved serine–arginine (SR) protein . The 12 human SR proteins feature a structure with one or two RNA recognition motifs (RRMs) at the N-terminus, followed by a C-terminal RS domain comprising multiple arginine–serine dipeptide repeats . Serine/arginine-rich splicing factor 1 (SRSF1) is one of the best-characterized SR proteins containing two RRMs and one RS domain. The RRMs recognize the consensus sequence of pre-mRNA and regulate alternative splicing [16,17]. In addition, the RS domain is required for the nuclear-cytoplasmic shuttling of SRSF1 .
SRSF1 is phosphorylated by several kinases, which regulate its subcellular location and functions . The RS domain of SRSF1 is phosphorylated by SRPK1/2 [20,21] in the cytoplasm and Clk/Sty [22,23] in the nucleus. Phosphorylation by these kinases modulates the shuttling of SRSF1 between the cytoplasm and nucleus. In addition to the phosphorylation of the RS domain, SRSF1 is phosphorylated at Ser119 in the inter-linker between RRM1 and RRM2 by protein kinase A (PKA) . PKA-dependent phosphorylation of SRSF1 enhances its RNA-binding capacity and modulates its splicing activity . Interestingly, a previous study reported that metformin, an activator of AMPK, modifies alternative splicing . Therefore, AMPK is a potential regulator of alternative pre-mRNA splicing. However, no direct substrate of AMPK for alternative pre-mRNA splicing has been reported.
In this study, we identified SRSF1 as a novel substrate of AMPK. AMPK directly phosphorylated SRSF1 at Ser133 in an RRM2. This phosphorylation suppressed the ability of SRSF1 to interact with target pre-mRNA sequences and regulated alternative splicing of the target gene. We proposed that AMPK is also important for the regulation of alternative splicing through SRSF1 phosphorylation.
Materials and methods
Antibodies against HA, Flag, SRSF1 (all from Sigma–Aldrich, St Louis, MO, U.S.A.), Myc (Medical and Biological Laboratories, Nagoya, Japan), AMPK α1, AMPK γ1, phospho-Ser79 ACC, ACC and β-actin (Cell Signaling Technology, Inc., Danvers, MA, U.S.A.) were purchased. The phospho-Ser133 SRSF1 antibody was generated by immunizing rabbits with the phosphopeptide LPPSG-pS-WQDLK at SCRUM (Tokyo, Japan).
The plasmid expressing HA-AMPK α1 was previously described , and Flag-AMPK β1 and γ1 were also previously described . The DNA fragment corresponding to human SRSF1 (full length) was cloned into pRK7-Myc and pRK7-Flag vectors. To generate GST-tagged human SRSF1, SRSF1 fragments (full length and amino acids 109–158) were cloned in a pGEX-KG vector. The SRSF1 S133A mutants (full length and amino acids 109–158) were generated using a PrimeSTAR Mutagenesis Basal Kit (TaKaRa Bio, Shiga, Japan). To generate the Ron mini-gene expression plasmid, a fragment of Ron (nucleotides 2507–2991) was amplified from human genomic DNA (Promega, Madison, WI, U.S.A.) as a template using primers in exons 10 (5′-ataagcttttcctgaatatgtggtccgag-3′ and 12 (5′-atggatccctagctgcttcctccgcc-3′). The PCR fragment was cloned into the HindIII and BamHI restriction sites of the pcDNA3.1 vector.
To generate shRNA plasmids of SRSF1, the following oligonucleotides were cloned into pLKO.1 (Addgene, Watertown, MA, U.S.A.): SRSF1 #1 sense, 5′-ccgggcaaccacgaaacctgtaatactcgagtattacaggtttcgtggttgctttttc-3′; SRSF1 #1 antisense: 5′-aattgaaaaagcaaccacgaaacctgtaatactcgagtattacaggtttcgtggttgc-3′. For cell treatments, Compound C (Enzo Biochem, New York, NY, U.S.A.) and A769662 (ChemScene LLC, Monmouth Junction, NJ, U.S.A.) were diluted in DMSO. For dephosphorylation assays, lambda protein phosphatase (New England Biolabs, Ipswich, MA, U.S.A.) was used.
Cell culture and transfection
HEK293T and MCF7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Wako, Osaka, Japan) containing 10% fetal bovine serum (FBS; Gibco, Invitrogen, Carlsbad, CA, U.S.A.) for HEK293T and 5% FBS for MCF7 cells under 5% CO2 at 37°C. Transfection with Lipofectamine (Invitrogen) was performed according to the manufacturer's instructions.
Cells were lysed in lysis buffer (10 mM Tris–HCl [pH 7.5], 100 mM NaCl, 1% Igepal, 2 mM EDTA, 10 mM sodium pyrophosphate, 50 mM NaF, 1 mM leupeptin, 1 mM aprotinin, 1 mM PMSF). For dephosphorylation assays, protein phosphatase lysis buffer (10 mM Tris–HCl [pH 7.5], 120 mM NaCl, 1% Igepal, 2 mM EDTA, 1 mM leupeptin, 1 mM aprotinin, 1 mM PMSF) was used.
For lentiviral infection, pLKO.1 shRNA vectors expressing shRNAs were transfected into HEK293T cells with the lentiviral packaging plasmids psPAX2 and pMD2. Viruses were collected 48 h after transfection and concentrated via ultracentrifugation for 1.5 h at 23 000 rpm and 4°C. Cells were infected, selected for puromycin resistance and analyzed 7 days after infection.
Tandem affinity purification of AMPK-binding proteins via LC–MS/MS
Tandem affinity purification was performed using InterPlay Mammalian TAP system according to the manufacturer's instructions (Agilent Technologies, Santa Clara, CA, U.S.A.). Briefly, HEK293T cells expressed CBP/SBP-tagged AMPKβ1 and collected with lysis buffer. Then, washed streptavidin resin was added to the lysates, which were rotated at 4°C for 2 h. Each resin was washed three times with wash buffer (InterPlay TAP purification kit, Agilent Technologies). Next, elution buffer (InterPlay TAP purification kit, Agilent Technologies) was added to each resin, which was rotated at 4°C for 30 min. Following rotation, the supernatant was transferred to a fresh tube and concentrated using an Amicon Ultra-15 PLBC Ultracel membrane (Merck Millipore, Billerica, MA, U.S.A.). AMPK-binding proteins were analyzed using LC–MS/MS by Fred Hutchinson Cancer Research Center (Seattle, WA, U.S.A.).
In vitro kinase assay and kinetic analysis
The GST-fused SRSF1 fragment (amino acids 109–158) was purified from bacteria as a substrate. GST-fused SRSF1 was phosphorylated for 30 min at 30°C by AMPK α1/β1/γ1 recombinant protein (Carna Biosciences, Inc., Kobe, Japan) in a kinase reaction mixture (18 mM Tris–HCl [pH 7.5], 10 mM MgCl2, 50 μM cold ATP, 1 mM DTT, 1 µCi [γ³²P] ATP). For the AMPK kinase assays, samples were subjected to SDS–PAGE and autoradiography. For the kinetic analysis, samples were spotted on a disc of Whatman Regenerated Cellulose Membrane (GE Healthcare, Chicago, IL, U.S.A.), washed six times with 100 mM Phosphate Buffer Solution, pH 7.0 (Wako) and subjected to a liquid scintillation system.
Cells were lysed in Triton-X100 buffer (10 mM Tris–HCl [pH 7.5], 100 mM NaCl, 1% Triton-X100, 2 mM EDTA, 10 mM pyrophosphate, 50 mM NaF, 1 mM leupeptin, 1 mM aprotinin, 1 mM PMSF) and immunoprecipitated with antibody and Protein G Sepharose beads (GE Healthcare) or anti-FLAG Affinity Gel (Sigma–Aldrich).
MCF7 cells grown in four-well plates were washed with PBS and fixed/permeabilized with 4% paraformaldehyde and 0.1% Triton-X100 in PBS. After being washed with PBS, the cells were blocked with PBS containing 2% bovine albumin (PBS-BSA) for 30 min, followed by incubation with an antibody for 90 min. The cells were washed and incubated for 90 min with the secondary antibody (Alexa 594-conjugated goat anti-mouse IgG, 1:500). The cells were washed with PBS-BSA, and DAPI (Invitrogen) was added to stain DNA. Photographs were acquired on a computer using FV1200 IX83 (Olympus, Tokyo, Japan).
In vitro RNA–protein pull-down assays
Biotinylated RNA oligonucleotides were synthesized by Sigma–Aldrich and used for RNA pull-down assays. Streptavidin resin beads (Agilent Technologies) and biotinylated RNA oligonucleotides were incubated together for 15 min at room temperature in RNA capture buffer (20 mM Tris–HCl [pH 7.5], 1 M NaCl, 1 mM EDTA) and washed twice 20 mM Tris–HCl (pH 7.5). HEK293T cells were lysed using a lysis buffer and mixed with protein–RNA-binding buffer (0.2 M Tris–HCl [pH 7.5], 0.5 M NaCl, 1 mM EDTA) and 15% glycerol. The mixture was added to RNA-bound beads and incubated for 60 min at 4°C with rotation. The RNA-binding protein complexes were washed five times with wash buffer (20 mM Tris–HCl [pH 7.5], 10 mM NaCl, 0.1% Igepal). Next, 1× SDS sample buffer was added to the samples, which were heated for 5 min at 95°C. Samples were examined via Western blotting.
Reverse transcription PCR (RT-PCR)
For A769662 treatment, MCF7 cells were cultured in high-glucose DMEM (4500 mg/l) without FBS. For compound C treatment, MCF7 cells were cultured in low-glucose DMEM (1000 mg/l). After treatment, total RNA was isolated using RNA isoPlus (TaKaRa Bio) according to the manufacturer's protocol. Total RNA was converted to cDNA using qPCR RT Master Mix with gDNA Remover (Toyobo, Osaka, Japan). PCR was performed using EmeraldAmp PCR Master Mix (TaKaRa Bio) with the forward primer 5′-acctagttccactgaagcct-3′ and reverse primer 5′-accagtagctgaagaccagt-3′ to measure alternative splicing of Ron exon 11.
RNA-cross-linking immunoprecipitation assays
HEK293T cells were transfected with Flag-SRSF1 or an empty vector (PRK7-Flag) and the Ron mini-gene. UV cross-linking with 480 mJ/cm2 was performed 36 h after transfection, and the cross-liked cells were then collected with RIP buffer (150 mM KCl, 25 mM Tris–HCl [pH 7.5], 5 mM EDTA, 1 mM DTT, 0.05% Igepal, 10 mM pyrophosphate, 50 mM NaF, 1 mM leupeptin, 1 mM aprotinin, 1 mM PMSF, 100 U/ml recombinant RNase inhibitor [TaKaRa Bio]). Then, immunoprecipitation was performed with Anti-FLAG M2 Affinity Agarose Gel, and immunoprecipitation products were treated with 0.2 mg/ml proteinase K (New England Biolabs) for 30 min at 37°C. Immunoprecipitated RNA and input RNA was purified and used for real-time PCR. For quantitative real-time PCR, total RNA was isolated using NucleoSpin RNA (TaKaRa Bio) according to the manufacturer's protocol. Total RNA was converted to cDNA using a PrimeScript RT-PCR Kit (TaKaRa Bio). Real-time PCR was conducted using FastStart Universal SYBR Green Master Mix (Roche, Basel, Switzerland) and primers (forward, (5′-ctactggtggcggaggaag-3′; reverse, 5′-ggccgccactgtgctggatatctgc-3′).
Results are presented as the mean of at three or four independent experiments, and error bars correspond to SD. Significant differences between groups were assessed using a two-tailed Student's t-test or the Tukey–Kramer test.
SRSF1 interacts with AMPK
To discover unidentified AMPK substrates involved in alternative pre-mRNA splicing, we purified proteins bound to AMPK using a two-step pull-down assay with an AMPK β1 subunit that was fused to two different peptide tags. AMPK-interacting proteins were identified using LC–MS/MS. We identified SRSF1 as a putative interacting protein of AMPK. To test whether SRSF1 associates with AMPK, we performed in vitro pull-down assays using GST-fused SRSF1 and lysate from HEK293T cells transiently transfected with the AMPK complex. We found that GST-fused SRSF1, but not GST alone, associated with the AMPK complex (Figure 1A). We additionally confirmed the association between SRSF1 and AMPK in both exogenous and endogenous condition by co-immunoprecipitation. Myc-SRSF1 was co-immunoprecipitated with HA-AMPKα1, Flag-AMPK β1 and AMPK γ1 subunits (Figure 1B). AMPKα1 subunit was co-immunoprecipitated with SRSF1 (Figure 1C). These results suggest that SRSF1 interacts with AMPK.
AMP-activated protein kinase (AMPK) interacts with serine/arginine-rich splicing factor 1 (SRSF1).
AMPK phosphorylates SRSF1
We examined the amino acid sequence of SRSF1 to identify a potential AMPK phosphorylation site in SRSF1. Only the amino acid sequence surrounding Ser133 matched the consensus sequence of an AMPK phosphorylation motif, which was determined by the frequency of amino acids in the motif around the well-validated AMPK phosphorylation sites  (Figure 2A). To examine whether AMPK directly phosphorylates Ser133, we performed in vitro kinase assays using a polypeptide containing Ser133 (amino acids 109–158) purified from bacteria as a substrate. AMPK phosphorylated the SRSF1 wild-type polypeptide, whereas mutation of Ser133 to alanine (S133A) completely abrogated phosphorylation by AMPK (Figure 2B,C), indicating that AMPK directly phosphorylates SRSF1 at Ser133 in vitro. To confirm whether AMPK phosphorylates SRSF1 in cultured cells, we generated an anti-phospho-Ser133 SRSF1 antibody (pS133-SRSF1). We transiently transfected HEK293T cells with Myc-SRSF1 and the AMPK complex. AMPK overexpression enhanced Ser133 phosphorylation in wild-type SRSF1, whereas this phosphorylation was completely abolished by treatment with lambda protein phosphatase (Figure 2D). Moreover, phosphorylation of Ser133 was not detected in a non-phosphorylatable mutant, S133A in AMPK-overexpressing cells (Figure 2E). These data confirmed the specificity of anti-pS133-SRSF1 antibody and illustrated that AMPK phosphorylates exogenous SRSF1. We also examined whether AMPK phosphorylates endogenous SRSF1 at Ser133 in cultured cells. Compound C, an AMPK inhibitor, suppressed SRSF1 phosphorylation at Ser133 (Figure 2F), whereas A769662, an AMPK activator, increased the phosphorylation level (Figure 2G). These results suggest that AMPK is a major upstream kinase for phosphorylation of SRSF1 at Ser133.
AMP-activated protein kinase (AMPK) phosphorylates serine/arginine-rich splicing factor 1 (SRSF1).
AMPK does not affect SRSF1 localization
SRSF1 undergoes various post-translational modifications that alter its intracellular localization, thereby changing its function [29–31]. Based on these observations, we hypothesized that the phosphorylation of SRSF1 at Ser133 affects its localization. To examine whether endogenous SRSF1 localization is regulated by AMPK activity, we performed immunofluorescence staining using MCF7 cells. Contrary to our expectation, the subcellular localization of SRSF1 was not changed by treatment with the AMPK activator A769662 in MCF7 cells (Figure 3A). To further confirm whether SRSF1 phosphorylation changes its localization, we observed the localization of the Ser133 phospho-defective and phospho-mimetic mutants of SRSF1. Similar to the findings presented in Figure 3A, the subcellular localization of SRSF1 was not affected by Ser133 mutation (Figure 3B). This result suggests that the phosphorylation of SRSF1 by AMPK does not modulate its subcellular localization.
AMP-activated protein kinase (AMPK) does not affect serine/arginine-rich splicing factor 1 (SRSF1) localization.
AMPK suppresses the interaction of SRSF1 with GGA motif RNA via phosphorylation of SRSF1
SRSF1 is a member of the SR protein family, which is important for regulating both constitutive and alternative splicing through sequence-specific binding to pre-mRNA. A previous structural study suggested that RRM2 of SRSF1 directly interacts with 5′-GGA-3′ motif RNA . Indeed, we also confirmed that the 5′-biotinylated RNA probe CUGAAGGACA interacted with SRSF1, whereas a mutant probe, in which the interacting motif was changed to CUGAAUUACA, did not interact with SRSF1 (Figure 4A). Because Ser133 is located in RRM2, we considered whether phosphorylation of SRSF1 at Ser133 affects the interaction between SRSF1 and the motif RNA. Similarly, we compared the binding of wild-type or SRSF1 mutants to 5′-biotinylated RNA probes. The interaction between SRSF1 and 5′-biotinylated RNA probes was reduced by both the S133D and S133A mutations (Figure 4B). This result suggests that the side chain of SRSF1 Ser133 may be directly involved in the regulation of RNA binding. Thus, we next investigated whether AMPK-induced Ser133 phosphorylation affects the RNA-binding ability of SRSF1. The result indicated that AMPK-phosphorylated SRSF1 exhibited decreased binding to 5′-biotinylated RNA probes (Figure 4C). These data suggest that AMPK suppresses the interaction of SRSF1 with 5′-GGA-3′ motif RNA via phosphorylation of Ser133 in vitro.
AMP-activated protein kinase (AMPK)-induced phosphorylation of serine/arginine-rich splicing factor 1 (SRSF1) affects the affinity of SRSF1 for the GGA motif.
AMPK regulates SRSF1-mediated Ron exon 11 splicing
SRSF1 is known to regulate alternative splicing of the pre-mRNA of Ron, a macrophage-stimulating protein (MSP) receptor. SRSF1 interacts with Ron pre-mRNA and promotes the production of ΔRon through the skipping of Ron exon 11 (Figure 5A), which enhances cell motility and invasion . To elucidate the regulation of SRSF1-induced alternative splicing of Ron exon 11 by AMPK, we treated MCF7 cells with Compound C or A769662 to alter AMPK activity. The total RNA was extracted and subjected to RT-PCR using primer pairs that could detect Ron mRNA containing or lacking exon 11. We found that inhibition of AMPK by Compound C treatment promoted Ron exon 11 skipping under a low-glucose condition that can be expected to increase AMPK activity, whereas AMPK activation by A769662 treatment under a high-glucose condition decreased Ron exon 11 skipping (Figure 5B,C). To confirm whether AMPK affects Ron splicing via SRSF1, we generated SRSF1-depleted MCF7 cells using shRNA (Figure 5D). Similar to the findings in Figure 5C, treatment with A769662 in cells infected with scrambled shRNA suppressed Ron exon 11 skipping, and as expected, SRSF1 depletion also decreased exon 11 skipping (Figure 5E). However, A769662 treatment did not further suppress Ron exon 11 skipping in SRSF1-depleted MCF7 cells (Figure 5E). These results suggest that AMPK alters Ron exon 11 splicing via SRSF1.
AMP-activated protein kinase (AMPK) regulates serine/arginine-rich splicing factor 1 (SRSF1)-mediated Ron exon 11 splicing.
AMPK suppresses the interaction with Ron via SRSF1 phosphorylation
As phosphorylation of SRSF1 affected the alternative splicing of Ron, we next focused on the binding of SRSF1 to Ron pre-mRNA. A recent study found that SRSF1 binds directly to the 5′-CGGAGGAAG-3′ sequence in exon 12, which leads to Ron exon 11 skipping and this interaction is abrogated by mutation of the 5′-CGGAGGAAG-3′ sequence to 5′-CGGUUGUUG-3′ . As this study, we also confirmed that SRSF1 bound to the 5′-biotinylated RNA probe CGGAGGAAG (Ron exon 12 probe) in vitro and the interaction was decreased by its mutation (Figure 6A). We also confirmed whether the interaction with Ron is regulated by SRSF1 phosphorylation. Similar to the results in Figure 4B, the interaction of SRSF1 with the Ron exon 12 probe was decreased by S133D and S133A mutation (Figure 6B). This result suggests that the side chain of Ser133 is important for the interaction of SRSF1 with Ron, and also that SRSF1 interacts with Ron exon 12 via RRM2. We next examined whether AMPK regulates the binding of SRSF1 to Ron exon 12. The data indicated that SRSF1 phosphorylated by AMPK exhibited reduced binding to the Ron exon 12 probe in vitro (Figure 6C). Furthermore, to confirm whether AMPK regulates the binding between SRSF1 and Ron in cultured cells, expression of the exogenous Ron mini-gene that binds to SRSF1 was measured by qPCR. We found that the co-expression of AMPK with A769662 treatment reduced the level of SRSF1-interacting Ron mRNA derived from the Ron mini-gene (Figure 6D). These results suggest that AMPK regulates the alternative pre-mRNA splicing of Ron by suppressing the interaction between SRSF1 and Ron pre-mRNA.
Phosphorylated serine/arginine-rich splicing factor 1 (SRSF1) exhibited decreased Ron mRNA binding
in vitro and in cultured cells.
In this report, we identified SRSF1 as the first AMPK substrate that regulates alternative splicing. AMPK phosphorylated the RRM2 domain of SRSF1 and regulated the interaction of SRSF1 with RNA recognized by RRM2 without changing its localization. We also uncovered that the alternative splicing of Ron exon 11 is altered by AMPK, indicating that AMPK regulates SRSF1-dependent alternative splicing via SRSF1 phosphorylation.
SRSF1 has two RRMs, one of which is the canonical RRM on the N-terminal side, followed by RRM2, which is known as pseudo-RRM . Efficient RNA binding and splicing require both RRM1 and RRM2 of SRSF1 [16,17]. However, RRM2 plays a major role in determining substrate specificity in vivo[18,35]. RRM2 of SRSF1 contains SWQDLKD, a phylogenetically conserved amino acid sequence, and this motif directly interacts with RNA . The results of crystal structure analysis illustrated that SRSF1 binds to purine-rich exonic splicing enhancer (ESE) containing the GGA motif via the SWQDLKD motif of RRM2 . In fact, we observed that the phosphorylation of SRSF1 at Ser133 inhibited its binding to RNA containing the GGA motif. Moreover, crystal structure analysis further revealed that Ser133, Trp134, and Gln135 form an ideal surface for binding the GGA motif via confirmation of hydrogen bonding between Ser133 and Gln135 . Therefore, it is suggested that the phosphorylation of Ser133 inhibits the interaction of SRSF1 with RNA by suppressing the formation of hydrogen bonds with Gln135. In fact, Ser133 mutations including both the S133A phospho-defective mutation and S133D phospho-mimetic mutation decreased binding to RNA containing the GGA motif. These data also suggest the importance of Ser133 side chains forming hydrogen bonds with Gln135 in the interaction of SRSF1 with RNA.
Post-translational modification of the SRSF1 RS domain is also an important factor that regulates SRSF1 function. RS domain phosphorylation by SRPKs promotes the transport of SRSF1 into the nucleus [20,21], whereas dephosphorylation by protein phosphatases promotes its export to the cytoplasm . The SWQDLKD motif directly interacts with the small lobe of SRPK1 . Moreover, the disruption of the interaction between RRM2 and the N-terminal region of the RS domain by mutations in RRM2 impairs the nuclear translocation of SRSF1 . Thus, we predicted that Ser133 phosphorylation affects the intracellular localization of SRSF1. Although it was not confirmed whether the binding of SRSF1 to SRPK1 is regulated by Ser133 phosphorylation, the subcellular localization of SRSF1 was not affected by Ser133 phosphorylation. In addition, Myc-SRSF1 was detected in a hyper-phosphorylated state, whether or not Ser133 was phosphorylated (Figure 2D). This indicates that phosphorylation of SRSF1 by other kinases such as SRPK1 that regulates subcellular localization may be unaffected by Ser133 phosphorylation. Thus, the major alternative splicing changes resulting from the phosphorylation of SRSF1 at Ser133 appear to inhibit the interaction between SRSF1 and RNA. AMPK is known to phosphorylate substrates in both the nucleus and cytoplasm . Therefore, it is speculated that AMPK phosphorylates SRSF1 in the nucleus and then regulates alternative splicing. In Figure 2C, we showed the extent of Ser133 phosphorylation by AMPK using a recombinant fragment of SRSF1. However, to confirm that SRSF1 is phosphorylated at a rate similar to known AMPK substrates, it is necessary to repeat the experiment with full-length SRSF1 and to compare it with well-characterized AMPK substrates.
SRSF1 is highly expressed in many different cancers, and it is identified as a proto-oncogene . In addition to increased cell proliferation , SRSF1 overexpression is associated with cell motility and invasion via the promotion of Ron exon 11 skipping . Ron, a tyrosine kinase receptor for MSP, consists of α and β subunits . MSP stimulates the kinase activity of Ron and activates the downstream pathway . In addition to promoting cell growth, Ron promotes tumor progression by regulating epithelial-to-mesenchymal transition (EMT) . The ΔRon isoform is constitutively active without its ligand . Interestingly, ΔRon transcripts are increased in tumors, and they promote tumor metastasis . A previous study reported that AMPK activation suppresses EMT by modulating the Akt-MDM2 Foxo3a axis . In this study, we found that AMPK inhibits the SRSF1-dependent production of the ΔRon isoform. Thus, we suggest that AMPK also suppresses EMT by regulating the alternative splicing of SRSF1-target genes. Moreover, SRSF1 is overexpressed in many different cancer types, and it modulates cancer-related splicing events . Thus, AMPK is a therapeutic target of SRSF1-dependent tumor progression.
AMPK is known to suppress mTORC1 activity by phosphorylating Raptor, a component of mTORC1 that controls mRNA translation . Moreover, the well-known AMPK activator metformin inhibits the translation of a subset of mRNAs via the mTORC1/4E-BP pathway . A recent study indicated that metformin alters alternative splicing by reducing the protein expression of RNA-binding motif protein 3 (RBM3) . Presumably, AMPK regulates RBM3 protein expression levels by inhibiting mTORC1 activity. In this study, we demonstrated that AMPK regulates alternative splicing via the direct phosphorylation of SRSF1, which suppresses the SRSF1 RNA-binding ability, suggesting that AMPK regulates alternative splicing through multiple pathways.
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by JSPS KAKENHI [grant no. 17K15273] and Tokyo University of Agriculture (TUA) graduate student research supporting grants [grant no. 46404789F].
E.M. and T.S. managed the project, designed the experiments, and wrote the manuscript. E.M. performed most of the experiments, analyzed the data, and prepared the figure. K.A. and T.S. performed TAP purification and in vitro kinase assay. K.K., Y.M., J.I., and Y.Y. contributed to the experimental design and data analysis.
The authors would like to thank Enago (www.enago.jp) for English language review.
AMP-activated protein kinase
Cdc2-like kinase 1
Dulbecco's modified Eagle's medium
exonic splicing enhancer
fetal bovine serum
mammalian target of rapamycin
mTOR complex 1
RNA-binding motif protein 3
RNA recognition motif
- RS domain
domain rich in arginine–serine dipeptide repeats
- SR protein
splicing factor containing arginine–serine dipeptide repeats
SR-specific protein kinase 1
SR protein splicing factor 1 (aka ASF/SF2)
tandem affinity purification