Cyclic AMP-dependent Protein Kinase Regulates the Alternative Splicing of Tau Exon 10

Hyperphosphorylation and deposition of tau into neurofibrillary tangles is a hallmark of Alzheimer disease (AD). Alternative splicing of tau exon 10 generates tau isoforms containing three or four microtubule binding repeats (3R-tau and 4R-tau), which are equally expressed in adult human brain. Dysregulation of exon 10 causes neurofibrillary degeneration. Here, we report that cyclic AMP-dependent protein kinase, PKA, phosphorylates splicing factor SRSF1, modulates its binding to tau pre-mRNA, and promotes tau exon 10 inclusion in cultured cells and in vivo in rat brain. PKA-Cα, but not PKA-Cβ, interacts with SRSF1 and elevates SRSF1-mediated tau exon 10 inclusion. In AD brain, the decreased level of PKA-Cα correlates with the increased level of 3R-tau. These findings suggest that a down-regulation of PKA dysregulates the alternative splicing of tau exon 10 and contributes to neurofibrillary degeneration in AD by causing an imbalance in 3R-tau and 4R-tau expression.

Tau is a neuronal microtubule-associated protein, the function of which is to stimulate microtubule assembly and stabilize microtubules. Hyperphosphorylation of tau leads to its aggregation into neurofibrillary tangles, a hallmark of Alzheimer disease (AD) 2 and related neurodegenerative diseases called tauopathies (1)(2)(3)(4). Adult human brain expresses six different tau isoforms from a single gene by alternative splicing of exons 2, 3, and 10 of its pre-mRNA (5). The exon 10 encodes the second microtubule binding repeat (6). Alternative splicing of exon 10 generates tau with three or four microtubule binding repeats (3R-tau or 4R-tau), which is under developmental and cell type-specific regulation. Only 3R-tau is expressed during embryogenesis, whereas 3R-tau and 4R-tau are expressed in approximately equal amounts in adult human brain (6,7). Sev-eral mutations in tau gene result in either an increase or a decrease in 4R-tau expression and cause frontotemporal dementia with parkinsonism linked to chromosome 17 , one of the tauopathies (8). Thus, alteration in the 3R-tau/ 4R-tau ratio is sufficient to trigger neurodegeneration in frontotemporal dementia and might also play a role in other neurodegenerative disorders such as Pick's disease, progressive nuclear palsy, or corticobasal degeneration in which the 3R-tau/4R-tau ratio is markedly altered (9 -12). Thus, the regulation of alternative splicing of human tau exon 10 has been of critical interest. However, results of studies of the alternative splicing of tau exon 10 in AD brain have been contradictory (13)(14)(15). Recent studies have shown that aggregation and deposition of 3R-tau may be associated with more advanced stages (16,17).
Cyclic AMP (cAMP)-dependent protein kinase, PKA, has emerged as a key kinase that is able to interact with many of the proteins involved in the etiology of AD as well as other tauopathies. It has been shown that PKA phosphorylates tau at several sites and primes phosphorylation of tau by glycogen synthase kinase-3␤ (37). PKA is a tetrameric holoenzyme consisting of two catalytic (C) subunits and two regulatory (R) subunits in the absence of cAMP. Stimulation by cAMP dissociates the holoenzyme and causes translocation to the nucleus of a fraction of the C subunit. Apart from regulation of transcription, little is known about the function of the C subunit in the nucleus. It is known that PKA phosphorylates several splicing factors and is involved in the pre-mRNA splicing (30,38,39). We recently found that in AD brain, the activity of PKA is down-regulated as a result of proteolysis of the regulatory subunit by over-activated calpain I (40). However, the role of PKA in the alternative splicing of tau exon 10 was unclear. In the present study we demonstrate that PKA phosphorylates SRSF1 and thereby enhances the inclusion of tau exon 10 and that down-regulation of PKA in AD brain correlates with increase in 3R-tau expression. These results suggest that PKA is involved in the tau pathology in AD via regulation of tau exon 10 splicing.

EXPERIMENTAL PROCEDURES
Human Brain Tissue-Medial frontal cortices from 15 AD and 15 age-matched normal cases used for this study (Table 1) were obtained from the Sun Health Research Institute Donation Program (Sun City, AZ); all cases were histopathologically confirmed. The tissue was stored at Ϫ70°C until used. The use of frozen human brain tissue was in accordance with the National Institutes of Health guidelines and was approved by our institutional review committee.
In Vitro Phosphorylation of SRSF1 by PKA-GST-SRSF1, GST-SRSF1 mutants, or as a control, GST (0.2 mg/ml) was incubated with various concentrations of PKA catalytic subunit in a reaction buffer consisting of 50 mM HEPES, pH 6.8, 10 mM ␤-mercaptoethanol, 10 mM MgCl 2 , 1.0 mM EGTA, and 0.2 mM [␥-32 P]ATP (500 cpm/pmol). After incubation at 30°C for 30 min, the reaction was stopped by boiling with an equal volume of 2ϫ Laemmli sample buffer. The reaction products were separated by SDS-PAGE. Incorporation of 32 P was detected by exposure of the dried gel to phosphor-image system.
GST Pulldown of PKA by SRSF1-GST, GST-SRSF1, and GST-SRSF1 deletion mutants were purified by affinity purification with glutathione-Sepharose but without elution from the beads. These beads coupled with GST, GST-SRSF1, and GST SRSF1 deletion mutants were incubated with crude extract from rat brain homogenate in buffer (50 mM Tris-HCl, pH 7.4, 8.5% sucrose, 50 mM NaF, 1 mM Na 3 VO 4 , 0.1% Triton X-100, 2 mM EDTA, 1 mM PMSF, 10 g/ml aprotinin, 10 g/ml leupeptin, and 10 g/ml pepstatin). After 4 h of incubation at 4°C, the beads were washed with washing buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 1 mM DTT) 6 times, the bound proteins were eluted by boiling in Laemmli sample buffer, and the samples were subjected to Western blot analysis.
Co-immunoprecipitation of PKA by SRSF1-HEK-293FT cells were transfected with pCEP4-SRSF1-HA for 40 h as described above and treated with 10 M forskolin for 8 h, and then the cells were washed twice with PBS and lysed by sonication in lysate buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 50 mM NaF, 1 mM Na 3 VO 4 , 2 mM EDTA, 1 mM PMSF, 2 g/ml aprotinin, 2 g/ml leupeptin, and 2 g/ml pepstatin). The cell lysate was centrifuged at 16,000 ϫ g for 10 min and incubated with anti-HA overnight at 4°C, and then protein G beads were added. After 4 h of incubation at 4°C, the beads were washed with lysate buffer twice and with TBS twice, and bound proteins were eluted by boiling in Laemmli sample buffer. The samples were subjected to Western blot analysis with the indicated primary antibodies.
Co-localization of PKA with SRSF1-HeLa cells were plated in 24-well plates onto coverslips 1 day before transfection at 50 -60% confluence. The cells were then transfected with pEGFP-N1-SRSF1 as described above. After 40 h of transfection, the cells were treated with 10 M forskolin for 30 min to activate PKA, and then the cells were washed with PBS and fixed with 4% paraformaldehyde in PBS for 30 min at room temperature. After washing with PBS, the cells were blocked with 10% goat serum in 0.2% Triton X-100, PBS for 2 h at 37°C and incubated with mouse anti-PKA-C␣ (1:50) overnight at 4°C. The cells were then washed and incubated for 1 h with secondary antibody (Alexa 488-conjugated goat anti-mouse IgG, 1:1000) plus TO-PRO-3 iodide at room temperature. The cells were washed with PBS, mounted with Fluoromount-G, and observed with a Nikon TCS-SP2 laser-scanning confocal microscope.
Knockdown of SRSF1 or PKA Catalytic Subunits, C␣ and C␤, with RNA Interference-For inhibition of SRSF1 expression, HEK-293T cells cultured in 12-well plates were transfected with various amounts of short interfering RNA (siRNA) using Lipofectamine 2000. After 48 h transfection, cells were lysed, and protein and RNA were extracted as described above. siRNA is a pool of 3 target-specific 20 -25-nucleotide siRNAs to knock down target gene expression (Santa Cruz Biotechnology). Both strands of siRNAs had 3Ј-dTdT tails. The same amount of scramble siRNA was used for controls.
Quantitation of tau Exon 10 Splicing by Reverse Transcription-PCR (RT-PCR)-Total cellular RNA was isolated from cultured cells by using the RNeasy Mini Kit (Qiagen, GmbH, Germany). One microgram of total RNA was used for first-strand cDNA synthesis with oligo-(dT) 15-18 by using the Omniscript Reverse Transcription kit (Qiagen). PCR was performed by using PrimeSTAR TM HS DNA Polymerase (Takara Bio Inc., Otsu, Shiga, Japan) with forward primer 5Ј-GGTGTCCACTC-CCAGTTCAA-3Ј and reverse primer 5Ј-CCCTGGTTTA-TGATGGATGTTGCCTAATGAG-3Ј to measure alternative splicing of tau exon 10. The PCR conditions were: 98°C for 3 min, 98°C for 10 s, and 68°C for 40 s for 30 cycles and then 68°C 10 min for extension. The PCR products were resolved on 1.5% agarose gels and quantitated using the Molecular Imager system (Bio-Rad).
Intracerebroventricular Injection-Sprague-Dawley rats at postnatal day 10 (P10) were first anesthetized by wrapping in ice pack for 5 min, and then 2 l of 1 mM forskolin in artificial CSF was injected into the left lateral ventricle of the brain at around 2.5 mm depth. The control animals were treated identically but with vehicle only. Rats were killed 3 days after injec-tion. The brains were immediately removed and processed for measuring the tau exon 10 splicing by RT-PCR.
Statistical Analysis-Where appropriate, the data are presented as the means Ϯ S.D. Data points were compared by the unpaired two-tailed Student's t test, and the calculated p values are indicated in the figures. For analysis of the correlation between levels of PKA-C␣ and 3R-tau or 4R-tau in human brain homogenates, the Pearson product-moment correlation coefficient r was calculated.

Activation of PKA Promotes Tau Exon 10 Inclusion in HEK-293T Cells and in Rat
Brain-To elucidate the regulation of the alternative splicing of tau exon 10 by PKA, we transfected minitau gene pCI/SI9-LI10 (comprising tau exons 9, 10, and 11 and part of intron 9 and the full-length of intron 10) into HEK-293T cells and treated them with 10 M forskolin to activate PKA for different times or with various concentrations of this activator for 8 h. Within 48 h of transfection, the total RNA was extracted and subjected to RT-PCR to measure tau exon 10 splicing. PKA activation with forskolin treatment activated PKA, which was demonstrated by phosphorylation of CREB (cAMP-responsive element-binding protein) and by increased luciferase activity (Fig. 1a). Interestingly, we observed that activation of PKA by forskolin promoted tau exon 10 inclusion in a time-dependent and dose-dependent manner (Fig. 1, b and c).
Mammalian brain expresses both C␣ and C␤ of PKA (40,43). PKA-C␥ is expressed in testis only (44). To understand which isoform of PKA regulates tau exon 10 splicing, we co-transfected PKA-C␣ or -C␤ with mini-tau gene pCI/SI9-LI10 into HEK-293FT cells and analyzed the tau exon 10 splicing. We found that overexpression of PKA-C␣ significantly promoted tau exon 10 inclusion, but overexpression of PKA-C␤ slightly inhibited tau exon 10 inclusion (Fig. 1d). Knock-down of PKA-C␣ expression (supplemental Fig. 1) did not affect the alternative splicing at basal level but significantly inhibited forskolin-promoted tau exon 10 inclusion (Fig. 1e). Down-regulation of PKA-C␤ by siRNA (supplemental Fig. 1) increased the tau exon 10 inclusion at basal level but did not influence forskolin promoted tau exon 10 inclusion (Fig. 1e). To determine that overexpression or knockdown of PKA catalytic subunit affects PKA activity, we measured CREB phosphorylation at Ser-133 in the cells transfected with PKA-C␣ or PKA-C␤ or their siRNAs. We observed that knockdown of PKA-C␣ or PKA-C␤ significantly decreased PKA activity, and overexpression of PKA-C␣ or PKA-C␤ increased PKA activity, as determined by phosphorylation of CREB (supplemental Fig. 2). Taken together, these results suggest that PKA-C␣ promotes and PKA-C␤ suppresses tau exon 10 inclusion and that promotion of forskolin on tau exon 10 inclusion is dependent on PKA-C␣ but not PKA-C␤.
Adult murine brain only expresses 4R-tau, and both 3R-tau and 4R-tau are expressed during P5 to P30 (supplemental Fig.  3). To determine the correlation of developmental expressions of PKA-C␣ and tau isoforms, we measured the level of PKA-C␣ in the rat brain, and we found that PKA-C␣ expression coincided with the expression of 3R-tau and 4R-tau (supplemental Fig. 3). To test whether PKA regulates the alternative splicing of tau exon 10 in vivo, we injected forskolin (2 l of 1 mM) into the rat left lateral ventricle at p10 and then measured the tau exon 10 splicing by RT-PCR 3 days after the injection. We detected a significant increase of 4R-tau expression in forskolin-treated rats (Fig. 1f), which provides evidence that PKA regulates the alternative splicing of tau exon 10 in vivo and activation of PKA enhances tau exon 10 inclusion.
PKA Enhances SRSF1-mediated Tau Exon 10 Inclusion-Splicing factor SRSF1 plays a critical role in the alternative splicing of tau exon 10. Overexpression of SRSF1 promoted tau exon 10 inclusion in a time-dependent manner (supplemental Fig. 4). To determine whether PKA modulates SRSF1-promoted tau exon 10 inclusion, we treated SRSF1-expressing HEK-293T cells with 10 M forskolin for 8 h and then measured exon 10 splicing. We observed that treatment with forskolin as well as SRSF1 overexpression promoted tau exon 10 inclusion, and forskolin treatment further enhanced SRSF1-promoted tau exon 10 inclusion (Fig. 2a). In contrast, knockdown of the expression of SRSF1 (supplemental Fig. 5) by siRNA suppressed tau exon 10 inclusion and eliminated the effect of forskolin on tau exon 10 inclusion (Fig. 2b). These findings suggest that PKA modulates SRSF1-promoted tau exon 10 inclusion.
To determine the isoform-specific role of PKA catalytic subunits on SRSF1-mediated tau exon inclusion, we overexpressed PKA-C␣ and -C␤ in pCEP4/SRSF1-transfected HEK-293FT cells and determined tau exon 10 splicing. Consistent with the findings shown in Fig. 1e, we found that PKA-C␣, but not PKA-C␤, further enhanced the SRSF1-promoted tau exon 10 inclusion (Fig. 2c).
To confirm the isoform-specific promotion of PKA-C on SRSF1-mediated tau exon 10 inclusion, we knocked down the expression of PKA-C␣ or PKA-C␤ in SRSF1-expression cells by RNA interference and measured the splicing products of tau exon 10. Overexpression of SRSF1 promoted tau exon 10 inclusion. Similar to the results in Fig. 1e, down-regulation of PKA-C␣ did not affect the alternative splicing of tau exon, but down-regulation of PKA-C␤ increased tau exon 10 inclusion. Knockdown of the PKA-C␣, but not PKA-C␤, reduced the tau exon 10 inclusion by SRSF1 (supplemental Fig. 6a).
To determine whether the promotion of tau exon 10 inclusion by forskolin is also isoform-specific, we knocked down the expression of PKA catalytic subunits (supplemental Fig. 1) and then treated cells with forskolin. We found that an increase of tau exon 10 inclusion by forskolin treatment was diminished by siPKA-C␣ but not by siPKA-C␤ (supplemental Fig. 6b). Further enhancement of SRSF1-promoted tau exon 10 inclusion by forskolin treatment also suppressed by siPKA-C␣ but not siPKA-C␤ (supplemental Fig. 6b). These results further support that PKA-C␣, but not PKA-C␤, promotes SRSF1 mediated tau exon 10 inclusion.
Alternative splicing is regulated by splicing factors and their phosphorylation, which results from the relative activity of the kinases and phosphatases. Cells from different sources and spe- Total RNA was extracted and subjected to RT-PCR to measure tau exon 10 splicing. c, SRSF1 was co-transfected with PKA-C␣ or C␤ into HEK-293FT cells for 48 h. The splicing products of tau exon 10 were determined by RT-PCR. d, two fractions of SRSF1 tagged with HA, elution 1 (E1) and E2, were immunopurified with anti-HA cross-linked onto protein G-agarose and subjected to Western blots by using anti-HA (left). Immunopurified (IP) SRSF1 in E1 and E2 was incubated with 32 P-labeled RNA oligonucleotides of tau including PPE at 37°C for 40 min. The incubation mixtures were separated by nondenaturing PAGE. The gel was dried and autoradiographed with a Phosphor-Imager. e, immunopurified SRSF1 was incubated with 32 P labeled RNA oligonucleotides of tau cover PPE (tau-RNA) or deleted PPE (tau-RNA ⌬PPE ) or deleted SC35-like element (tau-RNA ⌬SC35-like ) at 37°C for 40 min. The incubation mixture was separated by non-denaturing PAGE. The gel was dried and autoradiographed with a PhosphorImager. f, HEK-293FT cells were co-transfected with pCI/SI9-LI10 and pCEP4/HA or pCEP4/SRSF1-HA for 48 h, and then cells were fixed with formaldehyde, quenched with glycine, and lysed with lysis buffer. The nuclear fraction was collected, lysed, and sonicated. The extract of the nuclear fraction was used for immunoprecipitation with anti-HA. The proteins were digested with Proteinase K in the immunoprecipitated complex. The total RNA was extracted and reverse-transcribed to cDNA by using Oligo-dT primer. Tau exon 10 of cDNA was amplified with the two sets of primers, which are described under "Experimental Procedures." Results represent the mean Ϯ S.D.; *, p Ͻ 0.05; **, p Ͻ 0.01. cies and at different culture conditions have different basal activity of splicing factors or their regulators, resulting in the different splicing patterns in basal condition ( Fig. 1, a, b, and c, and Fig. 2, a, b, and c) (27). Nevertheless, SRSF1 or PKA promotes tau exon 10 inclusion consistently. SRSF1 promotes tau exon 10 inclusion by acting on the PPE at exon 10 (24). To confirm SRSF1 acting on PPE, we overexpressed SRSF1-HA in the HEK-293FT cells and immunopurified SRSF1 with anti-HA cross-linked to protein G-agarose. We got two elution fractions, E1 and E2, by elution buffer. We found that E1 had slower mobility than E2 on the SDS-PAGE, suggesting that phosphorylation of SRSF1 in E1 was higher than in E2 (Fig. 2d, left). Then, we incubated the purified SRSF1 with 32 P-labeled partial tau exon 10 RNA oligonucleotides and carried out EMSA. We found that SRSF1 in E1, but not in E2, made tau RNA oligonucleotides shift up (Fig. 2d, right), suggesting that binding of SRSF1 requires certain phosphorylation.
To determine whether SRSF binds onto PPE, immunopurified SRSF1 was incubated with a 32 P-labeled part of tau exon 10 RNA oligonucleotides (tau-RNA) containing SC35-likes and PPE elements. We found that SRSF1 slowed the mobility shift of tau-RNA (Fig. 2e). However, deletion of PPE, but not SC35like element, abolished the mobility shift (Fig. 2e), suggesting SRSF1 binds to the PPE of tau exon 10 mRNA.
To elucidate whether PKA modulates the binding of SRSF1 to tau exon 10, we performed, using anti-HA, RNA immuno-precipitation from pCEP4-SRSF1 and pCI/SI9-LI10 co-transfected HEK-293FT cells and amplified the precipitated RNA with RT-PCR by using two sets of primers (Fig. 2f). We found that treatment with forskolin increased the level of tau pre-mRNA co-immunoprecipitated with HA-SRSF1 by anti-HA (Fig. 2f), suggesting that PKA activation enhances the binding of SRSF1 to exon 10 of tau.
PKA Phosphorylates SRSF1 in Vitro and in Cultured Cells-The biological function of SRSF1 is highly regulated by its phosphorylation (23,27,34). To test whether PKA phosphorylates SRSF1, we incubated GST-SRSF1 with PKA in vitro. We observed that GST-SRSF1, but not GST, was phosphorylated by PKA in an enzyme concentration-dependent manner (Fig. 3a).
To determine whether PKA phosphorylates SRSF1 in living cells, we transfected CHO cells with pCEP4-SRSF1-HA and metabolically labeled cells with [ 32 P]orthophosphate. After 48 h of transfection and 3 h of labeling, cells were lysed and subjected to immunoprecipitation with anti-HA antibody. Immunoprecipitated products were separated by SDS-PAGE and visualized with phosphor-image analyzer. The results showed that forskolin treatment dramatically increased phosphorylation of SRSF1 (Fig. 3b), suggesting that PKA also phosphorylated SRSF1 in living cells.
SRSF1 has many putative phosphorylation sites. The majority of Ser/Thr residues are located at the RS domain with Arg-Ser repeats, which leads to difficulty in mapping The immunoprecipitated SRSF1-HA as well as whole cell lysates was analyzed by autoradiography. c, GST fused with different deletion mutants of SRSF1 were incubated with PKA in vitro for 30 min at 30°C. 32 P incorporation into GST-SRSF1 mutants was measured by autoradiography after the separation of the phosphorylation products by SDS-PAGE. Quantitation of the 32 P incorporation after being normalized by the protein level is shown in the bar graph. Results represent the mean Ϯ S.D.; *, p Ͻ 0.05; **, p Ͻ 0.01. the phosphorylation sites. However, we generated several deletion mutants of SRSF1 fused with GST and phosphorylated them in vitro with PKA to locate the regions of SRSF1 that are phosphorylated by PKA. We observed that deletion of any domains of SRSF1, including RS (RS1ϩRS2) and RS2, decreased the phosphorylation by PKA but could not abolish the phosphorylation (Fig. 3c). These results suggest that all three regions of SRSF1, RRM, RS1, and RS2, could be phosphorylated by PKA in vitro.
PKA Interacts with SRSF1-To investigate the interaction between PKA and various regions of SRSF1 (Fig. 4a), we performed a GST pulldown assay and co-immunoprecipitation. We found that PKA-C␣, but not PKA-C␤, was pulled down from rat brain extract by GST-SRSF1, but not GST (Fig. 4b), suggesting that SRSF1 interacts with PKA-C␣ but not PKA-C␤.
To identify the regions of SRSF1 that are involved in its interaction with PKA-C␣, we used the deletion mutants of GST-SRSF1 to pull down PKA from rat brain extract. We observed that all SRSF1 mutants pulled down PKA-C␣, but not PKA-C␤. Deletion of either RS domains or RRM reduced the interaction (Fig. 4b). Similar results were obtained from co-immunoprecipitation studies in the cells treated with forskolin, but RS domain did not co-immunoprecipitate a detectable level of PKA-C␣ (Fig. 4c), suggesting that SRSF1 interacts with PKA-C␣ mainly through the RRM.
To elucidate the interaction of SRSF1 with PKA-C␣ in intact cells, we transfected pEGFP-N1-SRSF1 into HeLa cells and then determined their subcellular localization by confocal microscopy. We observed that SRSF1 was localized extensively in the nucleus (Fig. 4d), whereas PKA-C␣ was mainly located in the cytoplasm. Forskolin treatment appeared to promote PKA-C␣ translocation into the nucleus and partially co-localized with SRSF1 (Fig. 4d).
Down-regulation of PKA Is Related to an Increase in 3R-tau Expression in AD Brain-Previously we have shown that PKA is down-regulated in AD brain as a result of increased degradation by over-activated calpain I (40). To investigate whether the down-regulation of PKA in AD brain causes a dysregulation in the alternative splicing of tau exon 10, we measured the levels of 3R-tau, 4R-tau, and PKA-C␣ in AD brains and age-and postmortem delay-matched normal human brains (Table 1) by Western blot analysis. We observed that in AD brain, the total tau level was increased by 3-fold, and 3R-tau was increased about 4-fold, but 4R-tau level was not significantly changed (Fig. 5, a and b), leading to an increase in the ratios of 3R-tau/ total tau and 3R-tau/4R-tau (Fig. 5c). These results suggest a change in 3R-tau/4R-tau ratio and probably alternative splicing of tau exon 10 in AD brain.
To study whether down-regulation of PKA-C␣ (Fig. 5, a and  d) correlated with the dysregulation of tau exon 10, we determined levels of PKA-C␣ and the ratio of 3R-tau to 4R-tau in AD brain and analyzed their correlation. We found a strong inverse correlation between the ratio of 3R-tau/4R-tau and PKA-C␣ levels (Fig. 5e), suggesting that the decreased PKA-C␣ in AD brain might contribute to the change in the ratio of 3R-tau and 4R-tau. . SRSF1 interacts with the PKA ␣ catalytic subunit. a, various SRSF1 deletion mutants used in this experiment are shown. b, GST-SRSF1, its deletion mutants, or GST coupled onto glutathione-Sepharose was incubated with rat brain extract, and the bound proteins were analyzed by Western blots developed with anti-GST, anti-PKA-C␣, or anti-PKA-C␤. c, SRSF1 and its deletion mutants tagged with HA were co-expressed in HEK-293FT cells for 48 h. The cell extracts were immunoprecipitated with anti-HA, and the immunoprecipitates (IP) were subjected to Western blots developed with antibodies indicated at the right of each blot. d, HeLa cells were transfected with GFP-SRSF1 and treated with forskolin (Fors) for 30 min followed by triple immunofluorescence staining. Scale bar, 10 m.  Braak and Braak (60). c Tangle score was a density estimate and was designated as none, sparse, moderate, or frequent (0, 1, 2, or 3 for statistics), as defined according to CERAD Alzheimer disease criteria (61). Five areas (frontal, temporal, parietal, hippocampal, and entorhinal) were examined, and the scores were combined for a maximum of 15.

DISCUSSION
The present study shows for the first time that the alternative splicing of tau exon 10 is probably dysregulated in AD brain, resulting in an imbalance between the levels of 3R-tau and 4R-tau. The increase in the level of 3R-tau was negatively correlated with PKA-C␣ level, which is down-regulated in AD brain due to activation of calpain I (40). Activation of PKA enhanced the binding of SRSF1 to tau pre-mRNA and promoted SRSF1-mediated exon 10 inclusion in cultured cells and in vivo in rat brain. PKA-C␣, but not PKA-C␤, interacted with and phosphorylated both the RS and RRM domains of SRSF1. Overexpression PKA-C␣, but not PKA-C␤, enhanced SRSF1mediated tau exon 10 inclusion. Knockdown of PKA-C␣ inhibited SRSF1 function in promotion of tau exon 10 inclusion. Thus, down-regulation of PKA-C␣ in AD brain (40) might be responsible for the increase of 3R-tau expression, resulting in an increase in 3R-tau/4R-tau ratio, which may contribute to neurofibrillary degeneration (Fig. 6).
The finding of increased 3R-tau level in AD brain is inconsistent with previous studies, which showed either an increase in 4R-tau expression in brain regions affected by sporadic AD (15,45) or no change in tau isoforms in AD brain (13,46). Most of these studies detected mRNA levels of 3R-tau and 4R-tau. However, RNA can be degraded very easily. Many studies have shown that the postmortem interval is related to the quality of RNA, but it is pH that affects RNA quality the most strongly. In postmortem tissue, the protein was found to be much more stable and its level unchanged, even when RNA was degraded (47,48). By immunohistochemical analysis, Espinoza et al. observed that tangles appear with both 3R-tau and 4R-tau in the FIGURE 5. Increased ratio of 3R-tau/4R-tau in AD brain correlates with PKA-C␣ level. a, the levels of 3R-tau, 4R-tau, total tau, PKA-C␣, and PP5 (a loading control) in the frontal cortex from six AD and six control cases were determined by Western blots developed with anti-3R-tau (RD3), anti-4R-tau (RD4), R134d, anti-PKA-C␣, and anti-PP5, respectively. b, the relative levels of 3R-tau, 4R-tau, and total tau in the frontal cortex from 15 AD and 15 control cases were detected by Western blots as described in panel a. The blots were densitometrically quantified and normalized with GAPDH. c, the relative ratios of tau proteins in these samples were calculated. d, the relative levels of PKA-C␣ in these samples were determined by Western blots developed with anti-PKA-C␣ and normalized with GAPDH. e, correlation of the ratio of 3R-tau/4R-tau (x axis) with PKA-C␣ level (y axis) is shown. The levels of PKA-C␣ were plotted against the ratio of 3R-tau/ 4R-tau in the frontal cortex from 15 AD cases. Results represent the mean Ϯ S.D.; *, p Ͻ 0.05; **, p Ͻ 0.01. FIGURE 6. Proposed mechanism by which down-regulation of PKA contributes to increase in tau exon 10 exclusion and neurodegeneration via phosphorylation of SRSF1 in AD. PKA appears to regulate tau exon 10 splicing by phosphorylating SRSF1 and enhancing its binding to polypurine enhancer cis element of exonic splicing enhancer at tau exon 10 and to promote exon 10 inclusion. In AD, decreased levels of PKA due to calpain I activation suppress its role in tau exon 10 inclusion. As a result, levels of 3R-tau increase, which disrupts the balance of the 3R-/4R-tau ratio that is required for the normal function of the adult human brain and leads to aggregation of tau and the formation of neurofibrillary tangles in the affected neurons. hippocampus in AD, and some advanced cases had large amounts of thioflavin-S-positive neurofibrillary tangles only detected by anti-3R-tau antibody but not anti-4R-tau antibody (16). In addition, they found that the pathology appeared to be more severe and displayed more abundant 3R-tau-positive tangles in the anterior as compared with the posterior hippocampus from the same cases (16). These results suggest that aggregation and deposition of 3R-tau may be associated with more advanced stages. In the present study we measured the protein levels of 3R-tau and 4R-tau in the temporal cortices of AD and control brains with a very short postmortem interval (Ͻ3 h) and found an increase in 3R-tau level. Previously, we have demonstrated that in AD brain, overactivation of calpain I due to calcium dysregulation causes degradation of the regulatory subunit of PKA, PKA-RII (40). A decrease in PKA-RII at basal conditions provides less protection to PKA-C from degradation. The C subunits of PKA are also decreased in AD brain, including C␣ (40), which could lead to tau exon 10 exclusion, resulting in an increase in 3R-tau expression.
PKA is a Ser/Thr protein kinase and is involved in many biological pathways. Under non-stimulated conditions, PKA is present as an inactive heterotetramer consisting of two C subunits and two R subunits. There are three isoforms of C subunits, C␣, C␤, and C␥, and four isoforms of R subunits, RI␣, RI␤, RII␣, and RII␤. The C␣ isoform is ubiquitously expressed, whereas the C␤ isoform is expressed only in brain (43). PKA-C␥ is expressed only in the testis (44). Although the R␣ isoforms are ubiquitously expressed, the R␤ isoforms are predominant in the nervous and adipose tissues. When a signal arrives at the cell surface, it activates the corresponding receptor, which in turn leads to the transient elevation of intracellular cAMP and consequently activates PKA by dissociating the C subunits and the R subunits. The free C subunits catalyze phosphorylation of the substrate proteins and then become vulnerable to degradation. The present study showed differential effect of PKA-C␣ and PKA-C␤ on tau exon 10 splicing. PKA-C␣ promoted tau exon 10 inclusion and PKA-C␤ suppressed tau exon 10 inclusion. Further study found that PKA-C␣, but not PKA-C␤, interacted with SRSF1 and promoted SRSF1-mediated tau exon 10 inclusion. The difference in sequences between PKA-C␣ and -C␤ is located in the first 50 amino acids of the N terminus, suggesting that PKA-C␣ via N terminus interacts with SRSF1, and PKA-C␤ may act on other splicing factors to inhibit tau exon 10. SRSF1 harbors 2 RRM motifs and 2 RS domains and is heavily phosphorylated in cells. There are 26 Ser, 2 Thr, and 7 Tyr putative phosphorylation sites, as predicted by using NetPhos 2.0. However, the majority of Ser/Thr residues are located at the RS domain; only eight Ser/Thr residues are at the RRM domain. Site/regional phosphorylation impacts SRSF1 function and subcellular localization differentially (27,34). It has been reported that several kinases phosphorylate SRSF1 in vitro and in cultured cells. Clk and SRPK mainly phosphorylate the RS domain and drive SRSF1 from the cytoplasm into the nucleus and from speckles into nascent transcripts, respectively (34). We recently reported that Dyrk1A phosphorylates SRSF1 at Ser-227, -234, and -238 at the RS domain and leads it into speckles (27). In the present study, by deletion mutations, we found that PKA not only phosphorylated the RS domain but also phosphorylated the RRM motif. The RS domain has been shown to mediate protein-protein (49) and protein-RNA interactions (50), to function in nuclear import (51)(52)(53), and to play a role in the targeting of proteins such as SC35 to nuclear speckles (54), whereas RRM determines their RNA binding specificity. It is known that SRSF1 binds to the PPE at exon 10 and promotes tau exon 10 inclusion. Tau deletion mutation ⌬K280 significantly decreases the SRSF1 binding and leads to tau exon 10 exclusion (24). In the present study we found that with EMSA, the hyperphosphorylated SRSF1, but not the hypophosphorylated SRSF1, bound to oligonucleotides of tau exon 10 and that with RNA immunoprecipitation assay, activation of PKA enhanced the binding of SRSF1 binding to PPE at exon 10. These data suggest that phosphorylation of SRSF1 by PKA might promote its binding to RNA.
Although hyperphosphorylation of tau plays a fundamental role in the development of Alzheimer-type neurofibrillary degeneration, imbalance in the cellular levels of 3R-and 4R-tau is emerging as an important concept in this pathology. Several lines of evidence, from transgenic mouse models to human tauopathies, emphasize the importance of a critical 3R-tau/4Rtau ratio in neurons. Disturbances of the 3R-tau/4R-tau ratio may lead to the characteristic neurofibrillary pathology. Close to 50% of all mutations in the tau gene causing human FTDP-17 affect tau exon exon 10 splicing and alter 3R/4R tau ratio (55). In addition to FTDP-17, dysregulation of exon 10 splicing may also contribute to other human tauopathies, such as Pick's disease (with a predominant increase in 3R-tau), progressive supranuclear palsy (4R-tau up-regulation), corticobasal degeneration (4R-tau up-regulation) (12), and Down syndrome (3Rtau up-regulation) (27). The neuronal tau expression levels and isoform content is highly cell-and region-specific both during development and in the mature brain (56). 3R-tau and 4R-tau are not functionally equivalent with respect to interactions with microtubules. In vitro, 4R has an ϳ3-fold higher binding affinity for microtubules than that of 3R-tau (57,58). In addition, 4R-tau is better at initiating and promoting microtubule assembly than 3R-tau (6,59). We observed an ϳ4-fold increase in 3R-tau and an insignificant increase in 4R-tau in the AD brain with overactivation of calpain I, resulting in an increase of 3R-tau/4R-tau. Imbalanced tau apparently offers a good substrate for the phosphorylation and aggregation in neurons, leading to neurofibrillary degeneration.