Tau Becomes a More Favorable Substrate for GSK-3 When It Is Prephosphorylated by PKA in Rat Brain*

Microtubule-associated protein tau is abnormally hyperphosphorylated in Alzheimer's disease (AD) and other tauopathies and is believed to lead to neurodegeneration in this family of diseases. Here we show that infusion of forskolin, a specific cAMP-dependent protein kinase A (PKA) activator, into the lateral ventricle of brain in adult rats induced activation of PKA by severalfold and concurrently enhanced the phosphorylation of tau at Ser-214, Ser-198, Ser-199, and or Ser-202 (Tau-1 site) and Ser-396 and or Ser-404 (PHF-1 site), which are among the major abnormally hyperphosphorylated sites seen in AD. PKA activation positively correlated to the extent of tau phosphorylation at these sites. Infusion of forskolin together with PKA inhibitor or glycogen synthase kinase-3 (GSK-3) inhibitor revealed that the phosphorylation of tau at Ser-214 was catalyzed by PKA and that the phosphorylation at both the Tau-1 and the PHF-1 sites is induced by basal level of GSK-3, because forskolin activated PKA and not GSK-3 and inhibition of the latter inhibited the phosphorylation at Tau-1 and PHF-1 sites. Inhibition of cdc2, cdk5, or MAPK had no significant effect on the forskolin-induced hyperphosphorylation of tau. Forskolin inhibited spatial memory in a dose-dependent manner in the absence but not in the presence of Rp-adenosine 3′,5′-cyclic monophosphorothioate triethyl ammonium salt, a PKA inhibitor. These results demonstrate for the first time that phosphorylation of tau by PKA primes it for phosphorylation by GSK-3 at the Tau-1 and the PHF-1 sites and that an associated loss in spatial memory is inhibited by inhibition of the hyperphosphorylation of tau. These data provide a novel mechanism of the hyperphosphorylation of tau and identify both PKA and GSK-3 as promising therapeutic targets for AD and other tauopathies.

Tau is a major neuronal microtubule-associated protein. Its normal function is to promote microtubule assembly from tubulin subunits and to stabilize microtubules. The biological activity of tau is regulated by the level of its phosphorylation (1)(2)(3)(4)(5)(6)(7)(8). In normal adult brain, 2-3 mol of phosphates per mol of tau are present, whereas this protein is 3-to 4-fold more hyperphosphorylated in Alzheimer's disease (AD) 1 (9,10). This abnormally hyperphosphorylated tau becomes microtubule assembly incompetent and, consequently, aggregated into tangles of paired helical filaments (PHFs) (6,9,(11)(12)(13). The neurofibrillary pathology of AD type is also seen in a number of related neurodegenerative disorders called tauopathies, which are characterized clinically by dementia (14). Several studies have demonstrated that the abnormal hyperphosphorylation of tau is responsible for the loss of its biological activity, gain of its toxicity, and its aggregation into tangles of PHFs (6,8,11,13,15,16). Hence, elucidating the regulation of tau phosphorylation by protein kinases and phosphatases is critical for understanding the molecular mechanism of neurodegenerative diseases.
To date, at least 29 phosphorylation sites have been identified in PHF-tau (17)(18)(19)(20)(21). It is believed that several protein kinases are involved in the phosphorylation of tau, and no single kinase can phosphorylate all of these sites. Among more than ten protein kinases that have been shown to phosphorylate tau in vitro, glycogen synthase kinase-3 (GSK-3), cyclin-dependent kinases 5 (cdk5), cell division cycle 2 kinase (cdc2), mitogen-activated protein kinases (MAPKs), calcium-and calmodulin-dependent protein kinase II, and cAMP-dependent protein kinase (PKA) are some of the most implicated in the regulation of tau phosphorylation and in the abnormal hyperphosphorylation of tau in AD brain (for review, see Refs. 17, 22, and 23). Recent studies have suggested that several protein kinases probably work coordinately to phosphorylate tau. Phosphorylation of tau by one kinase may facilitate or inhibit the phosphorylation by other protein kinases at the same or other phosphorylation sites of tau in vitro (24 -28). Previously we found that phosphorylation of tau by PKA significantly promotes subsequent tau phosphorylation by GSK-3 at multiple phosphorylation sites in vitro (24,26). However, it was not known whether this type of regulation of tau phosphorylation occurred in vivo.
In the present study, we show that activation of PKA by * This work was supported by grants from the National Natural Science Foundation of China (Grants 39925012, 30170221, and 30100213), Science and Technology Committee of China (Grant G1999054007), National Educational Committee of China (Grant 2001-171), the Li Foundation, Inc., USA, and National Institutes of Health Grant AG19158. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Both authors contributed equally to this work. infusion of forskolin in the lateral ventricle of the brain in adult rats enhanced phosphorylation of tau in hippocampus not only at Ser-214, a PKA site, but also at GSK-3 sites, Ser-198, Ser-199, and/or Ser-202 (Tau-1 site) and Ser-396 and/or Ser-404 (PHF-1 site). The increase in the phosphorylation of tau at the Tau-1 site and the PHF-1 site by GSK-3 occurred due to the  priming of tau by PKA and not by any increase in GSK-3 activity. LiCl, which specifically inhibited GSK-3 activity but not other protein kinases such as PKA, cdc2, cdk5, and MAPKs, completely abolished the increase in the phosphorylation of tau at Tau-1 and PHF-1 sites but not at Ser-214. Furthermore, inhibition of cdc2, cdk5, and MAPK by specific respective inhibitors did not abolish the hyperphosphorylation of tau at Ser-214, Tau-1, and PHF-1 sites. In contrast, specific inhibition of PKA completely abolished the increase in the phosphorylation of tau at Ser-214, Tau-1, and PHF-1 sites. Moreover, the forskolin-induced hyperphosphorylation of tau inhibited spatial memory in rats. This study provides the first in vivo evidence that PKA-catalyzed phosphorylation facilitates tau phosphorylation by GSK-3 and associated loss in spatial memory, suggesting a role of PKA upstream to GSK-3 in the abnormal hyperphosphorylation and associated neurofibrillary degeneration and memory loss.
Surgical Procedures for Drug Infusion-Rats were first anesthetized by chloral hydrate (30 mg/kg, intraperitoneally) and placed on a stereotactic instrument with the incisor bar set 2 mm below the ear bars (i.e. flat skull). After the scalp was incised and retracted, a 50-l syringe (Hamilton) was stereotactically placed into the lateral ventricle of cerebrum at the co-ordinates from bregma and dura of anterior/posterior 0.8, lateral 1.5, and ventral 4 (in mm). Different concentrations of chemical compounds (see below, "Results"), separately or in combination, were injected (40 l) into the left ventricle of the brain. The compounds were dissolved in artificial cerebrospinal fluid (aCSF) composed of 140 mM NaCl, 3.0 mM KCl, 2.5 mM CaCl 2 , 1.0 mM MgCl 2 , and 1.2 mM Na 2 HPO 4 , pH 7.4. The same volume of aCSF was injected into the left ventricle for control animals. All surgical procedures were completed under sterile conditions and penicillin (200,000 units, intramuscularly) was injected to prevent infection (29,30).
Preparation of Rat Hippocampal Extracts-Rats were killed 24 h after injection, following measurement of spatial memory. The hippocampus was immediately removed and homogenized at 4°C using a Teflon glass homogenizer in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM NaF, 1 mM Na 3 VO 4 , 10 mM ␤-mercaptoethanol, 5 mM EDTA, 2 mM benzamidine, 1.0 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml aprotinin, and 2 g/ml pepstatin. The tissue homogenates were then divided into two portions. One portion of each homogenate was centrifuged at 12,000 ϫ g for 20 min at 4°C, and the resulting supernatant was stored at Ϫ80°C for assaying activities of protein kinases. The other portion was mixed in 2:1 (v/v) ratio with lysis buffer containing 200 mM Tris-HCl, pH 7.6, 8% SDS, 40% glycerol, boiled for 10 min in a water bath, then centrifuged at 12,000 ϫ g for 30 min, and the supernatant was stored at Ϫ80°C for Western blot analysis. The concentration of protein in the hippocampal extracts was measured by BCA kit (30) according to manufacturer's instruction (Pierce).
Western Blot Analysis-The phosphorylation of tau at various sites was determined by Western blots as described previously (31). Blots were developed with PS214 (1:1,000), Tau-1 (1:30,000), PHF-1 (1:500), or Tau-5 (1:200) antibodies and visualized by enhanced chemiluminescent substrate kit (Pierce) and exposure to CL-XPosure film (Pierce). The immunoreactivity of tau bands was quantitatively analyzed by Kodak Digital Science 1D software (Eastman Kodak Co., New Haven, CT) and expressed as sum optical density. The levels of total tau and phosphorylated tau at various sites were expressed as relative level of the sum optical density against control.
Protein Kinase Activity Assay-The GSK-3 activity in rat hippocampal extracts was measured using phospho-GS peptide 2 as described previously (32)(33)(34). Briefly, tissue extract, 7.5 g of protein was incubated for 30 min at 30°C with 20 M peptide substrate and 200 M [␥-32 P]ATP (1500 cpm/pmol of ATP) in 30 mM Tris, pH 7.4, 10 mM MgCl 2 , 10 mM NaF, 1 mM Na 3 VO 4 , 2 mM EGTA, and 10 mM ␤-mercaptoethanol in a total volume of 25 l. The reaction was stopped by addition of 25 l of 300 mM O-phosphoric acid. The reaction mixture was applied in triplicates on phosphocellulose paper (Pierce). The filters were washed three times with 75 mM O-phosphoric acid, dried, and counted by liquid scintillation counter. The GSK-3 activity was expressed as picomoles of phosphate incorporated/mg of protein/min at 30°C.
The PKA activity was measured using Kemptide as a substrate, as described previously (35,36). Briefly, tissue extract, 7.5 g of protein was incubated for 10 min at 30°C with 100 M Kemptide, 5 M cAMP, and 100 M [␥-32 P]ATP (1500 cpm/pmol of ATP) in 40 mM Tris-HCl (pH 7.4), 20 mM MgCl 2 , and 0.1 mg/ml bovine serum albumin. The reaction was stopped, and the kinase activity was determined as described above for GSK-3 activity assay.
The cdc2 kinase activity was measured using synthetic peptide (PK-TPKKAKKL) corresponding to amino acids 9 -18 of histone H1 (37). Briefly, tissue extract, 10 g of protein was incubated for 10 min at 30°C with 50 M peptide substrate and 200 M [␥-32 P]ATP (2000 cpm/pmol of ATP) in 30 mM Tris, pH 7.4, 10 mM MgCl 2 , 10 mM NaF, 1 mM Na 3 VO 4 , 2 mM EGTA, and 10 mM ␤-mercaptoethanol in a total volume of 25 l. The reaction was stopped, and the kinase activity was determined as described above for GSK-3 activity assay.
The MAPK activity was measured using synthetic peptide (APRT-PGGRR) corresponding to amino acids 95-98 of bovine myelin basic protein (38 -40). Briefly, tissue extract, 10 g of protein was incubated for 10 min at 30°C with 250 M peptide substrate in assay buffer: 20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, and 1 mM dithiothreitol. The reaction was stopped, and the kinase activity was determined as described above for the GSK-3 activity assay.
The cdk5 activity was measured using immunoprecipitation of cdk5 (30). Briefly, 50 g of protein was mixed with 2 g of antibody to cdk5. After incubation at 4°C overnight, 30 l of immobilized protein G suspension was added to the reaction mixture, and it was constantly mixed for 2-4 h. The sedimented beads were suspended in 25-l assay buffer: 30 mM Tris, pH 7.4, 10 mM MgCl 2 , 10 mM NaF, 1 mM Na 3 VO 4 , 2 mM EGTA, and 10 mM ␤-mercaptoethanol, and reacted with 200 M [␥-32 P]ATP for 10 min at 30°C. The reaction was stopped, and the kinase activity was determined as described above for GSK-3 activity assay. The levels of all protein kinase activities were expressed as relative to the enzymatic activity in control animals.
Immunocytochemistry-Twenty-four hours after injection, rats were fixed in situ by perfusion for 20 min at 4°C by Zamboni's solution containing 2% paraformaldehyde, 15% saturated picric acid, and 24 mM NaH 2 PO 4 /126 mM Na 2 HPO 4 (pH 7.2). The brain was removed from the skull of the fixed animals and sliced coronally into blocks that contained hippocampus. These tissue blocks were further fixed in the same Zamboni's solution for another 12 h at 4°C, paraffin-embedded, and cut into 5-m-thick sections. The immunocytochemical staining was performed as described previously (32,41). Briefly, the tissue sections were first treated with 100 mM NaOH at room temperature for 30 min, followed by incubation at 4°C for 48 h with one of the following primary antibodies: PS214 (1:500), Tau-1 (1:30,000), or PHF-1 (1:500). The bound primary antibodies were detected using Histostain TM -SP kits (Zymed Laboratories Inc., South San Francisco, CA) and visualized with diaminobenzidine (brown color). All sections were counterstained lightly with hematoxylin to show cell nuclei (blue).
Animals and Morris Water Maze-Sprague-Dawley rats (Grade II, male, weight 200 -250 g, 4 months old) were supplied by Experimental Animal Central of Tongji Medical College. Rats were allowed free access to food and water and maintained at constant temperature (25°C). Spatial memory was measured by Morris water maze, according to classic Morris protocol (42). This protocol enables the measurement of the acquisition and as well as the retention of the spatial memory. Rats tested in the water maze were extensively handled (2 min every day for 7 days). Before each experiment (2 h), the rats were brought to the site to allow them to be acclimatized. The rats were kept in cages on shelves in an outer room to eliminate directional olfactory and auditory cues. The temperature of the room and water was kept at 26 Ϯ 2°C. The water in the pool was made opaque with milk (1 kg of milk/m 3 of water) to hide the escape platform. The Plexiglas platform was 40 cm high, 10 cm in diameter, and its surface was scarred to help the rats climb on it. The water surface was 18 cm from the rim of the pool, and the inner wall was always carefully wiped to eliminate any local cues. The rim of the pool was 1.0 m from the nearest visual cue of red and blue marks. For spatial learning, rats were trained by successive 20 trials with a 30-s interval (4 trials/day) in a water pool to find a platform hidden in milky water. On each trial, the rat started from one of the middle of the four quadrants facing the wall of the pool and ended when the animal climbed the platform. The rats were not allowed to search for the platform more than 60 s, after which they were guided to the platform. Through these training sessions, rats acquired spatial memory about location of the safe platform. These rats were then injected with test drugs or vehicle only (see below). After 24 h of drug injection the animals were retested for spatial memory retention in the Morris water maze. In this case, we removed the safe platform from the water pool and allowed rats to freely swim in the water pool for 60 s to assay the influence of various drug treatments on spatial memory retention. The time rats stayed in the previous platform quadrant (quadrant time) and the pathway of rats swimming were recorded by a video camera fixed to the ceiling of the room, 1.5 m from the water surface. The camera was connected to a digital-tracking device attached to an IBM computer loaded with the water maze software. The longer a rat stayed in the previous platform-located quadrant, the better it scored spatial memory. We expressed spatial memory of a rat as quadrant time (%).
Statistical Analysis-Data were analyzed using SPSS 10.0 statistical software. The one-way analysis of variance procedure followed by least significant difference post hoc tests was used to determine the statistical significance of differences of the means. To analyze the correlations among variables, Pearson correlations were computed with bivariate correlations procedure.

Forskolin Induces tau Phosphorylation at Ser-214, Tau-1, and PHF-1 Sites in Vivo-
To study the effect of PKA activation on tau phosphorylation in vivo, we injected different doses of forskolin, a specific PKA activator (43,44), into left lateral ventricle of rats for 24 h. Changes in the phosphorylation of tau at several sites were then examined by Western blots using phosphorylation-dependent and site-specific tau antibodies. We found that the phosphorylation of tau at Ser-214 in rat hippocampus increased to ϳ4-, 7-, and 15-fold of the control level 24 h after injection of 20, 40, and 80 M forskolin, respectively (Fig. 1A). Furthermore, under the same conditions we found that the phosphorylation of tau at Ser-396 and/or Ser-404 (PHF-1 site) increased to 1.4-, 2-, and 3-fold (Fig. 1B), and the level of tau where neither serines 198, 199, nor 202 are phosphorylated (Tau-1 site) was decreased to 77%, 50, and 38% of the control level (Fig. 1C), respectively. The total level of tau measured by Tau-5 was not changed significantly by forskolin treatment (Fig. 1D).
We also studied the immunohistochemical distribution of tau phosphorylation induced by infusion of forskolin in the lateral ventricle. With antibodies PS214 (Fig. 2, A and B) and PHF-1 (Fig. 2, C and D), we observed a marked increase in the immunostaining of hippocampi of forskolin-treated animals as compared with that of vehicle-injected controls. The most dramatic increase in the staining was seen in the mossy fibers of CA3 sector. When stained with Tau-1, a decreased staining was observed in sections from forskolin-treated animals (Fig. 2, E  and F). The topology of the change of staining with Tau-1 was similar to those of staining with PS214 and PHF-1. These results indicated that the infusion of forskolin induced phosphorylation of tau at various sites in the same region of the brain. The tau phosphorylation occurred most dramatically in the mossy fibers of CA3 sector of hippocampus.
Forskolin Increases PKA but Not GSK-3, cdc2, cdk5, and MAPK Activities in Rat Hippocampus-To confirm the activation of PKA in situ in the hippocampus by injection of forskolin into the lateral ventricle, we measured PKA activity in the hippocampal extracts and found a dose-dependent activation of PKA (Fig. 3A). Up to 6-fold activation of PKA was observed when injected with 80 M forskolin. The PKA activation was highly correlated to the phosphorylation of tau observed. This is evidenced by a positive correlation between PKA activity and  H) and PKA (B and G) of the hippocampal extracts were measured using respective specific immunoreactivities of PS214 (Fig. 3F) and PHF-1 (Fig. 3H), both of which recognize only the phosphorylated tau, and a negative correlation between PKA activity and immunoreactivity of Tau-1, which recognizes unphosphorylated tau (Fig. 3G). However, among the phosphorylation sites of tau studied above, only Ser-214 is a PKA site. Several studies have demonstrated that Tau-1 and PHF-1 sites cannot be phosphorylated by PKA (26,45,46). Instead these sites can be phosphorylated by GSK-3, cdc2, cdk5, and MAPK in vitro (47)(48)(49)(50)(51). Hence, we studied whether infusion of forskolin could indirectly activate GSK-3, cdc2, cdk5, or MAPK activity in the brain, which might have induced phosphorylation of tau at Tau-1 and PHF-1 sites. Determination of GSK-3, cdc2, cdk5, and MAPK activities of the hippocampal extracts indicated no increase in these protein kinase activities upon forskolin treatment (Fig. 3, B-E). These results suggested that the increased phosphorylation of tau at Tau-1 and PHF-1 sites was not due to an increase of GSK-3, cdc2, cdk5, or MAPK activity in forskolin-injected rat brain.

Increase in the Phosphorylation of tau at Tau-1 and PHF-1 Sites Is by GSK-3 Basal Activity When tau Is Prephosphorylated by PKA in Rat
Hippocampus-We previously found that prephosphorylation of tau by PKA promotes subsequent phosphorylation of tau at Tau-1 and PHF-1 sites by GSK-3 in vitro (26,52). Hence, we investigated whether, in the forskolininjected brain, phosphorylation of tau by PKA at Ser-214 enhanced phosphorylation at PHF-1 and Tau-1 sites by the basal activity of GSK-3. We injected LiCl, which is commonly used to inhibit GSK-3 activity (53), together with forskolin. We found that the injection with 100 mM LiCl induced a reduction of ϳ40% in GSK-3 activity in rat brain hippocampus (Fig. 4A), but it had no effect on PKA activity (Fig. 4B). This inhibition of GSK-3 did not affect forskolin-induced tau phosphorylation at Ser-214 ( Fig. 4C) but dramatically decreased forskolin-induced tau phosphorylation at PHF-1 (Fig. 4D) and Tau-1 sites (Fig. 4E). These changes in tau immunoreactivities were not due to any changes in total tau level, because neither forskolin alone nor forskolin plus LiCl altered total tau level in the rat hippocampus (Fig. 4F). These results suggested that the forskolin-induced phosphorylation of tau at PHF-1 and Tau-1 sites was catalyzed by basal activity of GSK-3.
To confirm whether the GSK-3-catalyzed phosphorylation of tau at PHF-1 and Tau-1 sites is dependent on PKA-catalyzed tau phosphorylation in rat hippocampus, we injected R p -cAMPS, a specific PKA inhibitor (54,55), together with forskolin. This combination blocks forskolin-induced PKA activation but not other potential forskolin-induced effects. Kinase assays confirmed the blockage of forskolin-induced PKA activation by R p -cAMPS (Fig. 4G). Injection of R p -cAMPS did not affect GSK-3 activity (Fig. 4H). We found that the blockage of PKA activation also blocked the forskolin-induced phosphorylation of tau at Ser-214 (Fig. 4I), PHF-1 (Fig. 4J), and Tau-1 (Fig. 4K) sites. These changes in tau immunoreactivities were not due to FIG. 5. Effects of R p -cAMPS and LiCl on cdc2, cdk5, and MAPK activities. Rats were injected into lateral ventricle with aCSF as vehicle, 80 M forskolin alone, or forskolin combined with either 100 mM LiCl or 100 M R p -cAMPS. The activities of cdc2, cdk5, and MAPK were determined by using specific peptide substrates. Immunoprecipitation with specific antibody (for cdk5) was employed to assay the cdk5 activity in the tissue extract. All data are expressed as mean Ϯ S.D. of eight experiments. peptide substrates. The phosphorylation levels of tau at various sites were determined by Western blots using phosphorylation-dependent and site-specific tau antibodies PS214 (for Ser-214; C and I), PHF-1 (for Ser-396 and or Ser-404; D and J), and Tau-1 (for Ser-198, Ser-199, and or Ser-202; E and K). The total tau level was determined by Western blots using a phosphorylation-independent tau antibody Tau-5 (F and L). The immunoreactivities (IR) of the tau staining were quantitated. All data are expressed as mean Ϯ S.D. of eight experiments. **, p Ͻ 0.01 as compared with control injection; ##, p Ͻ 0.01 as compared with injection with forskolin alone. any changes in total tau level, because neither forskolin alone nor forskolin plus R p -cAMPS altered total tau level in the rat hippocampus (Fig. 4L). These results suggested that the increase in the phosphorylation of tau at PHF-1 and Tau-1 sites by GSK-3 probably resulted from the prephosphorylation of tau at other sites by activation of PKA in rat brain.
Tau Is Not Phosphorylated at Tau-1 and PHF-1 Sites by cdc2, cdk5, or MAPK When It Is Prephosphorylated by PKA in Rat Hippocampus-In vitro, tau is known to be phosphorylated at Tau-1 and PHF-1 sites by cdc2, cdk5, or MAPK as prolinedirected protein kinases (17,22,23,47,50,56,57). To exclude any possible activation of these kinases by forskolin and inhibition of these kinases by LiCl and or R p -cAMPs in the present study, we measured the activities of these protein kinases in rats injected with these compounds. We found that forskolin did not increase the activities of cdc2, cdk5, or MAPK and that 100 mM LiCl and 80 M R p -cAMPS did not affect the activities of cdc2, cdk5, or MAPK (Fig. 5).
Additionally, 200 M PD 98059 and 100 M SB 203580, which decreased MAPK activity to ϳ54% and ϳ43% of the controls (Fig. 7, E and F), respectively, had no significant effect on PKA (Fig. 7, A and B) or GSK-3 (Fig. 7, C and D) activity. Neither of these two MAPK inhibitors had any significant effect on the forskolin-induced hyperphosphorylation of tau at PHF-1 site (Fig. 7, G and H), Tau-1 site (Fig. 7, I and J), or Ser-214 (Fig. 7, K and L), nor did they alter the levels of total tau (Fig. 7, M and N). These results further confirmed that cdc2, cdk5, and MAPK were not responsible for the increase in phosphorylation of tau at Tau-1 site, PHF-1 site, and Ser-214 when PKA was activated by forskolin.
Induction of an Impairment of Spatial Memory by Forskolin Is Due to an Increase in Phosphorylation of tau by PKA and GSK-3-The impairment of spatial memory is a typical symptom of AD in early stages. Therefore, we studied the influence of forskolin on spatial memory using Morris water maze. The rats treated by aCSF spent ϳ55% of total swimming time (quadrant time percent) in the previous platform-located quadrant, whereas 20, 40, and 80 M forskolin significantly decreased this quadrant time in a dose-dependent manner (Fig.  8A). The forskolin-induced decrease in the quadrant time was abolished by 100 M R p -cAMPS (Fig. 8B). Under these conditions hyperphosphorylation of tau at Tau-1 site, PHF-1 site, and Ser-214 was also practically completely inhibited (see Fig.  4, I, J, and K). On the other hand, we found that 100 mM LiCl only partially abolished the decrease in quadrant time induced by 80 M forskolin (Fig. 8C). Co-injection of 80 M forskolin and 100 mM LiCl completely abolished hyperphosphorylation of tau at Tau-1 and PHF-1 sites but not at Ser-214 (see Fig. 4, C-E). These results revealed that hyperphosphorylation of tau both by GSK-3 and PKA impairs spatial memory and that the phosphorylation of tau by the two kinases has additive effect on the impairment of this memory. DISCUSSION Abnormal hyperphosphorylation of tau is a pivotal lesion in the pathogenesis of AD and other tauopathies. Among the major goals of research on this family of diseases are the identification of the protein phosphatases and protein kinases that are involved in the abnormal hyperphosphorylation of tau, and associated neurodegeneration and memory impairment, and the molecular mechanism by which tau is abnormally hyperphosphorylated. Here we show a novel mechanism by which PKA and GSK-3 produce abnormal hyperphosphorylation of tau where the phosphorylation of this protein by the former enzyme primes it for phosphorylation by the latter kinase without any requirement for stimulation of its activity over and above the basal activity. The present study also shows that the activation of PKA produces accumulation of the abnormally hyperphosphorylated tau and an impairment in spatial memory and that this memory deficit involves the abnormal hyperphosphorylation both by PKA and GSK-3. These data identify both PKA and GSK-3 as potential therapeutic targets for AD and other tauopathies that are characterized by neurofibrillary degeneration associated with the abnormal hyperphosphorylation of tau. It is interesting to note that 14-3-3 and ␣-synuclein, both of which stimulate PKA activity (61,62), are colocalized with neurofibrillary tangles in AD brain (63).
Forskolin is a well known PKA-specific activator (43,44). In brains of rats injected with forskolin we observed in a dose-dependent manner severalfold activation of PKA and a concurrent enhancement of the phosphorylation of tau not only at Ser-214, a known PKA site, but also at the GSK-3 sites, Ser-198, Ser-199, and/or Ser-202 (Tau-1 site), and Ser-396 and/or Ser-404 (PHF-1 site). Previous studies have shown that tau at Tau-1 and PHF-1 sites is readily phosphorylated in vitro by proline-directed protein kinases, GSK-3, cdc2, cdk5, and MAPK (for review see Ref. 21) but not by PKA (45,46,56,64). We found that infusion of brain with both forskolin and R p -cAMPS, a specific inhibitor of PKA, did not produce activation of PKA and phosphorylation of tau at Ser-214, Tau-1 site, or PHF-1 site, suggesting that the hyperphosphorylation of tau observed at these sites in animals injected with forskolin alone were due to the activation of PKA. Infusion of the brain with forskolin along with LiCl, which inhibited GSK-3 activity but not cdc2, cdk5, and MAPK, resulted in the activation of PKA and phosphorylation of tau at Ser-214 and not at Tau-1 and PHF-1 sites, and furthermore inhibition of cdc2, cdk5, and MAPK activities did not abolish the increase in the phosphorylation of tau at Ser-214, Tau-1 site, and PHF-1 site by forskolin, demonstrating that the phosphorylation of tau at the latter two sites was catalyzed by GSK-3 but not other proline-directed protein kinases such as cdc2, cdk5, and MAPK. Because forskolin activated PKA and not GSK-3, cdc2, cdk5, and MAPK in the animals injected with this drug, the phosphorylation of tau at Tau-1 and PHF-1 sites, which could be inhibited by LiCl, was by the basal activity of GSK-3. Furthermore, in the forskolininjected animals we observed the hyperphosphorylation of tau at the Ser-214, Tau-1, and PHF-1 sites immunohistochemically. All these findings of the present study taken together demonstrate that, when it is prephosphorylated by PKA in vivo, tau becomes a more favorable substrate for GSK-3 but not FIG. 8. Effects of forskolin, LiCl, and R p -cAMPS on spatial memory. To acquire spatial memory, rats were trained for 20 trials (4 trials/day) successively in Morris water maze over a period of 5 days. On day 6 these rats were injected into a lateral ventricle with aCSF as vehicle, 80 M forskolin alone, or forskolin combined with either 100 mM LiCl or 100 M R p -cAMPS for 24 h. The effects of these treatments on spatial memory were measured by Morris water maze, in which the platform hidden in milky water was removed and the quadrant time and pathway of swimming were recorded for 60 s by a computer. The spatial memory of rats was expressed as quadrant time (%), i.e. the longer the quadrant time, the better the spatial memory. Forskolin impaired the spatial memory of rats (A), and R p -cAMPS completely abolished these effects (B). LiCl only partial abolished the impairment of spatial memory induced by forskolin (C). All data are expressed as mean Ϯ S.D. of 12 animals. **, p Ͻ 0.01 as compared with control injection; ##, p Ͻ 0.01 as compared with injection with forskolin alone.
other protein kinases such as cdc2, cdk5, and MAPK, and thus activation of PKA alone is sufficient to produce abnormal hyperphosphorylation of tau not only at PKA sites but also at GSK-3 sites primed by PKA.
The changes in the phosphorylation of tau were seen in the mossy fibers of CA3 but not other sectors of the hippocampus and the associated cortex, suggesting that this area of the brain is one of the earliest and most vulnerable to neurofibrillary degeneration in rat brain. It is also possible that in areas of the hippocampus other than the CA3, the tau might be in a hyperphosphorylated but not aggregated state and thus immunohistochemically negative. Normal tau is immunohistochemically negative under most standard conditions, whereas on abnormal hyperphosphorylation and aggregation this protein is readily immunolabeled in tissue sections (see Ref. 9). An immunohistochemical staining pattern of the hyperphosphorylation of tau similar to the present study was observed previously in rats in which PP-2A activity was inhibited by okadaic acid (30).
In AD and other tauopathies the abnormal hyperphosphorylation of tau is associated with dementia. In the present study, the hyperphosphorylation of tau induced by forskolin impaired the spatial memory of adult rats. Furthermore, R p -cAMPS, which abolished the hyperphosphorylation of tau at Tau-1 site, PHF-1 site, and Ser-214, practically completely abolished the impairment of spatial memory. On the other hand, LiCl, which abolished the hyperphosphorylation of tau at Tau-1 and PHF-1 sites but not at Ser-214, partially blocked the impairment of spatial memory. These findings are consistent with the observations in AD and other tauopathies and the hypothesis that the abnormal hyperphosphorylation of tau impairs memory and phosphorylation of more than one site contributes to this deficit.
The hyperphosphorylated tau and spatial memory deficit demonstrated in this study can be produced in a short time, and because these effects are similar to early pathological changes occurring in AD, the rat model established here might serve as a useful AD model for testing new and selective PKA and GSK-3 inhibitors. R p -cAMPS and LiCl might hold potential as therapeutic drugs for AD and other tauopathies based on their selective inhibition of PKA and GSK-3, respectively. The therapeutic relevance targeting GSK-3 activity has been recently strengthened by the demonstration that inhibition of this enzyme activity (by LiCl) significantly reduces A␤ production in both neuronal cells and AD animal models (53).
In conclusion, phosphorylation of tau by PKA primes it for subsequent hyperphosphorylation by GSK-3 in vivo. Activation of PKA alone produces abnormal hyperphosphorylation of tau by the basal GSK-3 activity, suggesting that PKA might be involved in the abnormal hyperphosphorylation of tau in an upstream position to GSK-3 and possibly other proline-directed protein kinases. PKA is a potential therapeutic target for AD and other tauopathies that are characterized by neurofibrillary degeneration associated with the abnormal hyperphosphorylation of tau.