Postmortem changes in the phosphorylation state of tau-protein in the rat brain

Neurobiology of Aging, Vol. 19, No. 6, pp. 535–543, 1998 Copyright © 1999 Elsevier Science Inc. Printed in the USA. All rights reserved 0197-4580/98 $–see front matter

PII:S0197-4580(98)00094-3

Postmortem Changes in the Phosphorylation State of Tau-Protein in the Rat Brain ¨ RTNER,* CARSTEN JANKE,* MAX HOLZER,* EUGEEN VANMECHELEN,† AND ULRICH GA THOMAS ARENDT*1 *Paul Flechsig Institute of Brain Research, Department of Neuroanatomy, University of Leipzig, Jahnallee 59, 04109 Leipzig, Germany and †Innogenetics N.V., Industriepark Zwijnaarde 7, 9052 Zwijnaarde (Ghent), Belgium Received 24 December 1997; Accepted 26 October 1998 ¨ RTNER, U., C. JANKE, M. HOLZER, E. VANMECHELEN, AND TH. ARENDT. Postmortem changes in the phosphorylation GA state of tau-protein in the rat brain. NEUROBIOL AGING 19(6) 535–543, 1998.—The phosphorylation state of tau-protein is crucial for the regulation of neuronal microtubule organization. Functional conclusions on tau-protein require an accurate assessment of phosphorylated sites. Therefore, the in vivo distribution and postmortem preservation of some phospho-epitopes on tau-protein were examined in the rat brain under different fixation and preparation conditions. Detection of tau-protein with a phosphorylationindependent antiserum revealed both axonal and somatodendritic localizations, which were not influenced by a postmortem interval of 30 min. The phospho-epitopes recognized by 12E8, AT8, and PHF-1 were mainly localized in the somatodendritic compartment. The binding sites of AT8 and PHF-1 were rapidly dephosphorylated postmortem, whereas the Tau-1 epitope was unmasked in the somatodendritic region. The axonally located phospho-epitope of AT270 and the nuclear epitope of AT100 were still detectable after a postmortem interval of 30 min. Postmortem dephosphorylation and inhibition of this process by PP1 and/or PP2A was further demonstrated on Western blot. In conclusion, rapid processing of tau-protein is essential for the correct assessment of investigations on phospho-isoforms. © 1999 Elsevier Science Inc. 12E8 AT8 Perfusion-fixation electrophoresis

AT100 PHF-1

AT180 AT270 Postmortem delay

BR134 Rat

Immersion-fixation Tau-1 Tau-protein

Immunohistochemistry Two-dimensional gel

abnormally highly phosphorylated (PHF-tau). Therefore, it may be prevented from efficient binding to microtubules, which could result in the destabilization of microtubules, an impairment of axonal transport, and eventually in degeneration and death of the affected nerve cells (9). Dephosphorylation of normal tau is regulated by numerous protein phosphatases, including phosphatase 1 (PP1), phosphatase 2A (PP2A), and the calcium/calmodulindependent phosphatase 2B (PP2B, also referred to as calcineurin) (15,20,21,56). This mechanism of dephosphorylation seems to not be effective for PHF-tau (21). However, the accurate assessment of experimentally obtained data on tau-protein, especially of phospho-isoforms, largely depends on preparation conditions. It has been shown that many phospho-epitopes originally thought to be specific for PHF-tau are, if only rapidly processed, also present in normal brain tau in humans and rats (24,36). Therefore, in the present study, we examined the influence of fixation conditions and postmortem delay on the phosphorylation state of tau in the rat brain. The subcellular localization of tau and the distribution of different phospho-isoforms were immunohistochemically examined using the phosphorylation-independent tau antiserum BR134 as well as the phosphorylation-dependent antitau antibodies 12E8, AT8, AT100, AT180, AT270, PHF-1, and

MICROTUBULES are one of the main cytoskeletal elements in eukaryotic cells and are especially prominent in nerve cells. They are composed primarily of tubulin dimers and associated proteins. Among these latter proteins, high-molecular-weight proteins (MAP 1 and MAP 2) and a group of proteins, collectively referred to as tau (approximate molecular weight 42,000 – 68,000) are major species in neuronal tissue (11,12,55). The isoform complexity of tau arises both from alternative splicing of a single gene as well as from posttranslational modifications, such as phosphorylation at different sites. Alternative splicing of tau transcripts theoretically gives rise to the expression of six distinct isoforms (19) containing either three or four sequence repeats in the carboxy-terminal region of the molecule that bind to microtubules (25). This variety is enlarged, furthermore, by the insertion of one or two amino terminal sequences (19). However, the highest complexity of isoform pattern is caused by its phosphorylation state, which is involved in the regulation of tau-protein function and metabolism (22,33). Phosphorylation of specific sites of tau modulates its ability to interact with and to stabilize microtubules. Increased phosphorylation of these sites is thought to decrease the ability of tau to promote microtubule polymerization (27,31). In Alzheimer’s disease, autopsy-derived human tau was shown to be

1 Address correspondence to: Thomas Arendt, Paul Flechsig Institute of Brain Research, Department of Neuroanatomy, University of Leipzig, Jahnallee 59, 04109 Leipzig, Germany. E-mail: [email protected]

535

¨ RTNER ET AL. GA

536 TABLE 1

SYNOPSIS OF THE IMMUNOHISTOCHEMICAL STAINING PATTERN AS REVEALED BY DIFFERENT TAU-SPECIFIC ANTIBODIES

Antibodies

BR134

Perfusion-Fixation

30-min Postmortem Delay of Fixation

Immersion-Fixation

No change

No change

AT100 AT270 12E8

Neuropil, neuronal perikarya, proximal dendrites, mossy fibers Neuronal nuclei Axonal structures Neuronal somata, mossy fibers

No change No change Staining of mossy fibers becomes less intense

PHF-1

Neuropil, neuronal somata, mossy fibers

AT8

Neuronal somata, apical dendrites, mossy fibers, Schaffer collaterals Neuropil, mossy fibers

Decreased staining of neuronal somata, mossy fibers labelling persists Only neuronal perikarya are still labelled in the neocortex Additional slight labelling of neuronal perikarya

No change No change Further decreased staining of mossy fibers, labelling of neuronal somata remains unchanged Complete loss of immunoreactivity

Tau-1

Tau-1. Phosphorylation-dependent changes in the isoform pattern of tau, furthermore, were analyzed by two-dimensional (2D) gel electrophoresis. METHODS

Tissue Preparation For immunohistochemical studies, adult Wistar rats (6 months) were deeply anaesthetized with ether and perfused through the ascending aorta with 100 mL of ice-cold phosphate-buffered saline (PBS, 10 mM Na2HPO4/NaH2PO4, 150 mM NaCl, pH 7.4) followed by 150 mL of ice-cold 4% (w/v) paraformaldehyde (PA) in PBS. The brains were postfixed in the same solution for 2 days and cryoprotected in 30% (w/v) sucrose. To introduce an artificial postmortem interval, some brains were removed after perfusion with PBS. After sagittal bisection, one half was immediately immersed in PA, the other half was stored at 37°C for 30 min before placing in PA. To investigate the influence of phosphatases 1 and 2A activity on the postmortem phosphorylation state of tau, selected brains were perfused with 5 mL of 20 mM okadaic acid ammonium salt (OA, Biotrend) after preperfusion with PBS. The brains were bisected thereafter and prepared as mentioned above. For electrophoretic analysis, the brains were rapidly removed and homogenized, without the cerebellum, in ice-cold Tris-buffered saline (10 mM Tris-HCl, 150 mM NaCl, pH 7.4) containing a cocktail of protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mg/mL pepstatin, 1 mg/mL leupeptin, 1 mg/mL aprotinin, 1 mg/mL Na-p-tosyl-L-lysine chloromethyl ketone (TLCK). A rapid processing was essential to maintain tau in its most native in situ phosphorylation state. Crude brain homogenates were employed for tau-protein purification steps either immediately after preparation or after subsequent experimental procedures. To study the effect of the endogenous phosphatase activity, freshly prepared homogenates were incubated at 37°C for 30 min prior to further preparation steps. Phosphatases 1 and 2A were inhibited by the addition of 1.2 mM OA to the homogenate; control incubations were performed omitting OA. Immunohistochemistry The immunoreactivity of tau was examined in the hippocampal formation and neocortex. Frozen coronal sections (30 mm thick) were cut and collected in PBS. Subsequent incubations were performed in PBS. Endogenous peroxidase activity was quenched

Complete loss of immunoreactivity Striking labelling of neuronal perikarya

with 0.3% (v/v) H2O2, and nonspecific binding sites were blocked with 1% (w/v) bovine serum albumin (Sigma), 0.3% (w/v) fat-free dried milk (Sigma), and 0.1% (w/v) gelatin (Sigma). Free-floating sections were reacted (4°C, overnight) with the polyclonal antiserum BR134 (1:1000, M. Goedert) or with one of the following monoclonal antibodies (mab) detecting phosphorylation-dependent tau-epitopes: 12E8 (1:1000, courtesy of P. Seubert, Athena Neurosciences, Inc.); AT8, AT100, AT180, and AT270 (1:500, E. Vanmechelen; Innogenetics); PHF-1 (1:1000, courtesy of P. Davies); or Tau-1 (1:1000, Boehringer Mannheim). Primary antibodies were omitted in control incubations. Immunoreaction was visualized using the biotin-avidin system (biotinylated sheep anti-mouse and donkey anti-rabbit antibodies, 1:1000, Amersham; avidin-peroxidase conjugate, 1:1000, ExtrAvidin®, Sigma) and 0.05% (w/v) 3,39-diamino benzidine (Sigma) as chromogen. Some sections were treated with alkaline phosphatase (AP) prior to incubation with the primary antibody. AP treatment was carried out in AP-buffer (100 mM Tris-HCl, pH 8.4, 150 mM NaCl, 2 mM ZnCl2, 2 mM MgCl2) with 6 U/mL alkaline phosphatase from Escherichi coli (Type III, Sigma) for 2 h at 67°C. Control incubations were performed in the presence of 200 mM sodium pyrophosphate. Electrophoresis and Western Blotting Preparation of Soluble Tau-Protein. High speed (50,000 3 g, 4°C, 30 min) supernatants of the tissue homogenates containing 1.0 M NaCl were supplemented with b-mercaptoethanol to a final concentration of 0.5% (v/v) and heat denatured (100°C, 5 min). After centrifugation (50,000 3 g, 4°C, 30 min), the supernatant was precipitated with 2.5% (w/v) perchloric acid and centrifuged (50,000 3 g, 4°C, 30 min). Tau was precipitated from the supernatants with 6% (w/v) trichloroacetic acid, pelleted (50,000 3 g, 4°C, 20 min) and washed twice with ice-cold methanol. To obtain dephosphorylated tau-protein, purified tau fractions were dissolved in AP-buffer (see Immunohistochemistry) containing protease inhibitors and treated with E. coli alkaline phosphatase (Sigma) at 67°C for 4 h. Control incubations were performed in the presence of 200 mM sodium pyrophosphate. 2D Gel Electrophoresis. Separation on 2D-gels was performed as described previously (28). The first dimension was run as nonequilibrium, pH-gradient electrophoresis with 4% (w/v) Phar-

POSTMORTEM CHANGES OF TAU PHOSPHORYLATION

537

malyte, pH 3–10 (Pharmacia). The second dimension was run as sodium dodecyl sulfate polyacrylamide gel electrophoresis (30). Western Blotting. Proteins were transferred from the sodium dodecyl sulfate polyacrylamide gel electrophoresis gels onto polyvinyl-difluoride membranes (DuPont/NEN) in a tank blotting system. For immunodetection, antibodies were diluted in TBS containing 0.05% (v/v) Tween-20; washing steps were performed in TBS containing 0.1% (v/v) Tween-20. Membranes were incubated in blocking reagent (gelatin, 1% w/v) for 1 h. After two washing steps, blots were incubated for 1 h in a primary antibody (12E8, 1:2000; AT8, 1:1000; AT180, 1:500; AT270, 1:2000; BR134, 1:4000; PHF-1, 1:1000; or Tau-1, 1:7500). Bound antibodies were detected with biotinylated sheep anti-mouse or donkey anti-rabbit antibodies (1:2000, Amersham), an avidin-peroxidase conjugate (ExtrAvidin®, 1:2000, Sigma) and 3,39diamino benzidine/NiCl2. Primary antibodies were omitted in control incubations. RESULTS

Immunohistochemistry The in vivo distribution of tau-protein and postmortem preservation of some of its phospho-epitopes were examined in the rat brain hippocampus and neocortex under different fixation and preparation conditions (for a synoptic overview, see Table 1). For this purpose, the phosphorylation-independent tau antiserum BR134 and the phosphorylation-dependent anti-tau antibodies 12E8, AT8, AT100, AT180, AT270, PHF-1, and Tau-1 were used. Depending on the specific antibody used, tau-immunoreactivity could be localized over the axonal, somatodendritic, or even the nuclear compartment. In perfusion-fixed brains, the pantropic BR134 stained mainly neuropil of the hippocampal formation and neocortex. In addition, a stronger immunoreactivity was observed in neuronal perikarya, apical dendrites (Fig. 1A) and mossy fibers. The immunoreactivity of the mab AT100 was localized in nuclei (Fig. 1C). The mab AT270 stained exclusively axonal structures (Fig. 1E). The mab 12E8 stained preferentially neuronal somata in both the neocortex and hippocampus. Mossy fibers were immunoreactive as well (Fig. 2A). PHF-1 staining was distributed over neuropil, with somewhat intenser staining of neuronal somata and mossy fibers (Fig. 2D). The mab AT8 (recognizing a phosphorylated epitope largely complementary to Tau-1) prominently labeled somata and apical dendrites of pyramidal neurons, both in the neocortex and hippocampal formation, as well as hippocampal mossy fibers and Schaffer collaterals (Fig. 3A, B, and E). Tau-1 immunoreactivity was found in mossy fibers of the hippocampal formation of perfusion-fixed brains (Fig. 4A and D). Furthermore, staining of neuropil was noticed. The mab Tau-1, however, stained no neuronal somata in perfusion-fixed tissue. The mab AT180 failed to detect tau-protein at tissue sections under any fixation condition. After immersion-fixation, the overall immunoreactivity obtained with BR134 and the mabs AT100 and AT270 was unchanged. The distribution of the 12E8 immunoreactivity was similar to that in perfusion-fixed tissue. However, staining of hippocampal mossy fibers was less intense (Fig. 2B). The immunoreactivity of PHF-1 was drastically decreased in neuronal somata, whereas mossy fiber staining was preserved (Fig. 2E). AT8 staining of apical dendrites disappeared after immersionfixation. Neuronal perikarya were still labeled in the neocortex, but immunoreactivity was nearly lost in the hippocampal formation (Fig. 3C and F). The neuropil staining with the mab Tau-1 persisted in the neocortex as well as in the hippocampal formation. Additionally, a slight reactivity of neuronal perikarya was observed (Fig. 4B and E).

FIG. 1. Effects of postmortem delay on the distribution of tau-protein detected with the phosphorylation independent antiserum BR134 (A, B) or with the phosphorylation dependent mabs AT100 (C, D) and AT270 (E, F). (A) The distribution of tau in the neocortex as stained with the pantropic antiserum BR134 after perfusion-fixation is not altered after a postmortem delay of fixation (B). Reactivity of AT100 is mainly confined to nuclei both after perfusion-fixation (C) and after a 30-min postmortem interval (D) (CA3 region of hippocampus). The pattern of AT270 staining is localized to axons in perfusion-fixed tissue (E) and is not altered after a postmortem delay of fixation (F) (dentate gyrus). mf: Mossy fibers; pyr: pyramidal cell bodies; Scale bars: (A/B) 30 mm; (C/D) 50 mm; (E/F) 200 mm.

A postmortem interval of 30 min did not influence the intensity and distribution of reactivity detected by BR134 and the mabs AT100 and AT270 (Fig. 1B, D, and F). The staining pattern of 12E8 was the same as in immersion-fixed tissue. A preferential labeling of neuronal somata in both the neocortex and hippocampus was observed. Compared to perfusion-fixed brains, mossy fibers were less immunoreactive (Fig. 2C). Reactivity with PHF-1 (Fig. 2F) or AT 8 (Fig. 3D and G) were now completely abolished. However, a stronger perikaryal staining was observed with mab Tau-1 (Fig. 4C and F). Perfusion with OA after preperfusion with physiological saline did not influence the stability of phosphorylated tau-epitopes within the observed postmortem interval. The pattern and intensity of staining showed no differences between brains with and without OA-perfusion. Treatment of slices with alkaline phosphatase did not alter the immunoreactivity of the phosphorylation-independent tau antiserum BR134. But, alkaline phosphatase completely abolished the immunoreaction of some antibodies that require phosphorylated epitopes, such as AT8, AT100, and AT270. However, 12E8 and PHF-1, which also recognize phosphorylated residues on tauprotein, showed no markedly modified immunoreactivity. Tau-1’s

¨ RTNER ET AL. GA

538

also reduced AT180 immunoreactivity and completely abolished AT8-immunoreactivity. These effects were sensitive to OA. Dephosporylation with alkaline phosphatase reduced the pattern of tau isoforms as detected by BR134 to three major bands. Each of these bands was resolved along the pH-gradient into a chain of single spots of equivalent molecular weights. A similar pattern was detected with Tau-1. Staining with 12E8 after alkaline phosphatase treatment was very weak and could only be detected for the higher-molecular-weight isoforms. Immunoreactivity of AT270, PHF-1, AT180, and AT8 were completely abolished by treatment with alkaline phosphatase. DISCUSSION

The results of the present study show that within the detection limits of the methods applied, the subcellular localization of tau-protein in the rat brain is not altered up to a postmortem delay of 30 min. However, the phosphorylation state of tau-protein depends on postmortem fixation and preparation conditions. The immunohistochemical results were confirmed using 2D-gel-electrophoretic analysis, where phosphorylation was detected according to the pI, and changes in the apparent molecular weight were caused by increased stiffness of the proteins (23). Previous studies have described tau-protein in the normal brain FIG. 2. Effects of postmortem delay and fixation conditions on the distribution of phosphorylated tau detected with the mabs 12E8 (A-C) and PHF-1 (D-F) in the hippocampus (CA3 region). (A) Immunoreactivity of mab 12E8 after perfusion-fixation. Mossy fibers and neuronal perikarya are stained. Labeling of Mossy fibers is reduced after immersion-fixation (B), and is hardly detectable after a 30-min postmortem interval (C) while staining of neuronal somate remains unchanged. (D) PHF-1 immunoreactivity in Mossy fibers and neuronal perikarya of perfusion-fixed brains (CA3). Reactivity is lost over somata after immersion-fixation (E) and completely abolished after a 30-min postmortem delay (F). mf: Mossy fibers; pyr: pyramidal cell bodies. Scale bar: (A) 50 mm.

staining pattern was similar to that seen after a 30-min postmortem delay of fixation. In addition to neuropil and axonal structures, neuronal somata were stained. Western Blot. Tau-protein, immediately prepared after decapitation and separated on 2D-electrophoresis, displayed a complex pattern of isoforms detected by the pantropic antiserum BR134 (Fig. 5). The phosphorylation-dependent monoclonal antibodies labeled only part of this pattern on Western blots. The mab Tau-1 stained a pattern similar to that of BR134, but those tau-isoforms with the most acidic pI and highest molecular weight were omitted from Tau-1 detection. The mab 12E8 detected preferentially the acidic isoforms of tau-protein with a relatively high molecular weight. AT270 and PHF-1 also detected that fraction of tau. AT180 and AT8 labeled only a minor subfraction of the most acidic tau isoforms. The mab AT100 failed to detect tau-protein on Western blot in any of the preparations. After an artificially introduced postmortem interval of 30 min, the staining pattern of BR134 was more distinctly resolved. Furthermore, reactivity was shifted toward the lower molecular weights and more basic pI. These alterations, however, could be prevented by incubating with OA. Changes in the staining pattern of Tau-1 induced by a postmortem delay were similar to those on BR134. Immunoreactivity of the mab 12E8 was clearly decreased after the postmortem interval, an effect that was reduced in the presence of OA. Staining intensity of both AT270 and PHF-1 was dramatically reduced after a 30-min postmortem delay. There was hardly any effect of the presence of OA. The postmortem interval

FIG. 3. Effects of postmortem delay and fixation conditions on the distribution of phosphorylated tau detected with the mabs AT8 in the hippocampus (A-D) and neocortex (E-G). After perfusion-fixation, the mab AT8 labels neuronal perikarya and apical dendrites as well as mossy fibers and Schaffer collaterals (A, B, and E). Immersion-fixation results in a loss of reactivity in the hippocampus (C). In the neocortex, only neuronal perikarya remain labeled (F). After a postmortem delay of fixation immunoreactivity is completely lost (D and G). mf: Mossy fibers; pyr: pyramidal cell bodies. Scale bars: (B) 50 mm; (A/C/D) 500 mm; (E/F/G) 100 mm.

POSTMORTEM CHANGES OF TAU PHOSPHORYLATION

FIG. 4. Effects of postmortem delay and fixation conditions on the distribution of phosphorylated tau detected with the mab Tau-1 in the hippocampus. Compared to staining obtained after perfusion-fixation (A), an increasing labeling of neuronal perikarya is observed after immersionfixation (B) and a postmortem delay of 30 min (C) (dentate gyrus). Similar effects in Tau-1 reactivity are noticed in the CA3 region: (D) perfusionfixation, (E) immersion-fixation without delay, (F) 30-min postmortem delay. mf: Mossy fibers; pyr: pyramidal cell bodies. Scale bars: (A-C) 200 mm; (D-F) 50 mm.

as mainly localized in the axon-neuropil compartment of neurons (6,7,45,52). The present data, however, show that tau immunoreactivity is also seen over neuronal somata as revealed by the phosphorylation-independent anti-tau antiserum BR134 (18). Staining of glial elements as previously reported (38,41,42) can not be ruled out, but was not obvious in the present study. The overall staining pattern was not influenced by a postmortem delay of fixation up to 30 min or treatment of slices with alkaline phosphatase. This finding is not consistent with previous observations (45) that have demonstrated a perikaryal accumulation of tau in immersion-fixed brains, as compared to perfused tissue that was increased along with an increase of the postmortem interval. This discrepancy may result from the different antibodies that were used. However, with postmortem intervals longer than 30 min, loss of tau immunoreactivity in axons and accumulation in neuronal somata might become apparent during autolysis and cytoskeletal disruption. Western blot analyses of freshly prepared tissue revealed a highly complex pattern of tau-isoforms as detected by BR134. This pattern was more distinctly resolved after the postmortem delay of preparation, which might reflect changes caused by the loss of phosphates. The pattern of tau treated with alkaline phosphatase differed from the pattern obtained after the postmortem delay. Whereas the latter was still highly complex after 2D resolution, enzymatic-dephosphorylated tau was separated into three major bands. These bands correspond to the three splicing products of tau in adult rats. Nevertheless, each of the bands was resolved into certain spots with different pI. Similar observations have been made on human tau-protein (28). Posttranslational modifications, such as deamination of glutamine and

539 asparagine residues, may be considered as cause for the pIheterogeneity. The present study suggests that a postmortem delay of fixation determines the staining pattern of some phosphorylation-dependent anti-tau antibodies. Several antibodies have been found to react with phospho-epitopes of tau-protein. Most of these phosphorylation sites were also found in adult rat brain tau (54). The mab 12E8 requires tau-phosphorylation at Ser-262 and Ser-356 and was shown to bind PHF-tau, rapidly processed adult rat and biopsy derived human brain tau (32,46). The present immunohistochemical data illustrate the preferential localization of this phospho-isoform in neuronal somata in addition to its occurrence in hippocampal mossy fibers. Modified fixation conditions caused no alterations in staining of somata, whereas mossy fibers were labeled slightly weaker after immersion fixation or a postmortem delay. This phospho-epitope of tau seems, therefore, to resist dephosphorylation in neuronal somata within the observed postmortem interval. However, these phosphates are less stable in axonal structures. Furthermore, treatment of sections with alkaline phosphatase did not remove phosphates from the 12E8 epitope. This could be caused by formaline stabilization of these phosphorylated sites or a general resistance to dephosphorylation by alkaline phosphatase. In contrast to immunohistochemical staining, Western blot analyses showed a clear decrease of the 12E8 immunoreactivity after incubation of homogenates at 37°C for 30 min. However, even the incubation of preparations with alkaline phosphatase could not completely dephosphorylate the 12E8epitope, which argues in favor of a resistance of these phosphorylation sites to the used phosphatase. The failure of alkaline phosphatase to remove phosphate from Ser-262 was reported recently (34). The mab AT8 recognizes tau when Ser-202 and Thr-205 are phosphorylated (17). AT8 stained mainly the somatodendritic compartment of neurons after perfusion-fixation. Immersion-fixation caused a strong decrease in AT8 immunoreactivity that was completely abolished by a 30-min delay of fixation or treatment with alkaline phosphatase. Tau-1 requires a similar but dephosphorylated sequence as AT8 (recognizes tau when Ser-199 and 202 are dephosphorylated) (5,49). Accordingly, staining in perfused brains was preferentially seen over the neuropil which may mostly represent axonal structures as was previously described in hippocampal cultures (35). Immersion-fixation caused a modified staining with additional immunoreactivity of cell bodies. This alteration was even more prominent after a 30-min postmortem interval, while axonal staining persisted. The same staining pattern resulted from treatment with alkaline phosphatase. This finding is in agreement with previous studies (41,45,50). The strong representation of the AT8 epitope accompanied by the missing Tau-1 immunoreactivity in the somatodendritic compartment in perfused tissue clearly indicates the localization of this specific tau phospho-isoform in neuronal perikarya and proximal dendrites in vivo, as was suggested previously (30,41,50). The immunohistochemical data on AT8 immunoreactivity were paralleled by Western blot analysis, where reactivity of tau-protein was completely lost after a 30-min postmortem interval or enzymatic dephosphorylation of homogenates. Effects on Tau-1 staining on Western blots, however, were less prominent, supporting earlier observations that changes in AT8 and Tau-1 reactivity do not behave exactly complementarily (10). The mab AT100 (also referred to as AT10) requires the phosphorylation of Thr-212 and Ser-214 on tau-protein (26). In formaline fixed tissue, this antibody stained exclusively neuronal nuclei. There was no effect of postmortem delay on staining intensity. The localization of immunoreactivity most likely reflects

¨ RTNER ET AL. GA

540

FIG. 5. Comparative study of the effects of postmortem delay, an inhibition of PP1 and 2A, and alkaline phosphatase on the 2D pattern of tau-protein on Western blot. Aliquotes of soluble tau, prepared immediately after decapitation (first lane), after 30-min incubation in the presence of OA (second lane), after 30-min postmortem delay (third lane), and after treatment with alkaline phosphatase (forth lane) were resolved on 2D electrophoresis and detected on Western blot with the antibodies BR134, Tau-1, 12E8, AT270, PHF-1, AT180, and AT8.

the presence of tau-proteins in the nuclei as reported previously in human brains (8,40). Staining intensity was not influenced by the postmortem delay of fixation, whereas treatment with alkaline phosphatase eliminated immunoreactivity. On Western blot, the AT100 immunoreactive tau-protein fraction was not detected, which could be explained by an insufficient solubilization of the nuclear tau-protein pool because of the preparation conditions in the present study. This observation parallels previous findings by Matsuo et al. (36), who report that AT100 immunoreactivity is restricted to the PHF-tau fraction. Nuclear tau has solubility properties similar to PHF-tau and requires formic acid extraction for solubilization (51). The present data suggest that phosphorylation of Thr-212 and Ser-214 seem to be specific for nuclear tau. Recognition of tau-protein by the mab AT180 depends on phosphorylation of Thr-231 (16). This antibody failed to detect tau-proteins in the tissue sections, which most likely results from

formaline fixation. Thus, no indication on cellular localization of this phospho-isoform could be obtained in vivo. A weak immunoreactivity, however, was seen on Western blot of freshly prepared tissue. This result confirms Western blot studies on tissue of adult rat cerebral cortex (43) and human biopsy-derived material (36). Immunoreactivity was less intense after a postmortem delay, which reflects the effects of endogenous phosphatase activity. AT180 staining disappeared completely after treatment with alkaline phosphatase. The mab AT270 recognizes tau when Thr-181 is phosphorylated (16). Immunoreactivity was restricted to axonal structures. The differences of fixation and the introduced postmortem interval did not alter the staining pattern. Treatment of sections with alkaline phosphatase, however, abolished AT270 staining. In contrast, the Western blot preparation indicated a drastic decrease of immunoreactivity after the postmortem interval of 30 min. The

POSTMORTEM CHANGES OF TAU PHOSPHORYLATION diminished immunoreactivity in homogenates contrasts with the immunohistochemical results. The fact, that AT270 may crossreact with phosphorylated neurofilaments (unpublished observation) could explain this disagreement between histological and Western blot data. The mab PHF-1 requires tau-protein phosphorylation at Ser396 and Ser-404 (39). The present immunohistochemical data suggest the presence of these phosphorylation sites at tau-proteins in the somatodendritic as well as axonal compartment of neurons. The same observation was made previously in rat cerebral cortex (50). A postmortem delay of fixation caused a strong loss of somatodendritic staining intensity, whereas labeling of axonal structures persisted for a longer period. This difference could reflect a weaker turnover of these specific phospho-residues in axons caused by less activity of phosphatases. Western blot analyses confirmed the immunohistochemical data. PHF-1 immunoreactivity was dramatically decreased after incubation of homogenates for 30 min. The postmortem instability of the PHF-1 epitope was reported previously (36). In a recent paper, an increasing postmortem phosphorylation of Ser-396/-404 up to 30 min post excision was found in biopsy-derived human tissue (47). These data could not be verified in the present study, perhaps because of the different processing paradigms. Treatment with alkaline phosphatase completely abolished PHF-1 immunoreactivity on Western blot. Immunocytochemical reactivity, however, did not disappear after enzymatic dephosphorylation. Formaline fixation possibly protects the epitope from dephosphorylation. We conclude that postmortem dephosphorylation differs with respect to specific phosphorylation sites. The loss of AT8 and PHF-1 immunoreactivity after a short fixation delay is due to a very rapid epitope dephosphorylation postmortem, whereas the binding sites of the mabs 12E8, AT100, and AT270 are preserved even after a 30-min postmortem fixation delay. The results are in agreement with previous observations that the dephosphorylation of tau was much slower at the PHF-1 and AT270 binding sites than that of the AT8 epitope (36). Furthermore, postmortem dephosphorylation of the PHF-1 binding site seems to occur to a different extent in separate compartments of neurons. Alkaline phosphatase treatment of sections did not remove the phosphates from all the binding sites of the antibodies used. On the other hand, incubation of homogenates with alkaline phosphatase completely abolished immunoreactivity of at least the mabs AT8, AT180, AT270, and PHF-1 on Western blot. It is, therefore, assumed that formaline fixation might protect some phosphorylated sites from artificial enzymatic dephosphorylation. Recent studies have demonstrated that normal tau is dephosphorylated in situ by at least PP1, PP2A, and PP2B (2,13,53). The sequences of the AT8/Tau-1 and PHF-1 epitopes are likely to be dephosphorylated by PP2A and/or PP2B in vitro (1,3,10,37,44,48)

541 as well as in vivo (unpublished data). To test the involvement of PP1 and PP2A in the postmortem dephosphorylation of tau, brains were perfused with OA (4). To compensate for dilution effects during brain passage, and considering that PP2A concentration in neurons is about 1 mM, a 20 mM OA solution was used for perfusion (14). However, OA perfusion could not arrest the postmortem dephosphorylation processes. Immunoreactivities of tau, irrespective of its phosphorylation state, as revealed by all antibodies used, did not change compared to those of controls. This might be explained by insufficient diffusion through blood vessels because the incubation of brain homogenates with 1.2 mM OA preserved at least the phosphates at Ser-202/Thr-205 (AT8), Thr-231 (AT180), and Ser-262/-356 (12E8). The effect was specific for OA because control incubations contained abundantly dephosphorylated tau-protein. The results of this study indicate that postmortem delay of fixation and/or preparation of tissue must be considered in the assessment of immunohistochemical and biochemical data on tau-proteins. The in vivo localization of tau-protein was found in the axonal as well as somatodendritic compartment at least up to a 30-min postmortem interval. However, the subcellular distribution seems to be specific for phospho-isoforms as was shown with phosphorylation-dependent antibodies, which confirm previous studies. 12E8, AT8, and PHF-1 immunoreactive tau-proteins were present in the somatodendritic part of neurons and in hippocampal mossy fibers, but was more or less absent in other axonal structures. Tau-1 immunoreactive tau was localized in neuropil, but not in neuronal somata. Immunoreactivity using the mab AT270 was restricted to axonal structures and that of AT100 in nuclei. Furthermore, it was shown that postmortem dephosphorylation occurs site specific. The binding sites of AT8 and PHF-1 are rapidly dephosphorylated, whereas epitopes of the mabs 12E8, AT100, and AT270 survive longer postmortem intervals. Western blot analysis supported these findings (extended also for the AT180-epitope) and documented the involvement of PP1 and/or PP2A in the dephosphorylation of at least the 12E8, AT8, and AT180 binding sites. In conclusion, a correct interpretation of studies on tau phospho-isoforms depends on a rapid postmortem fixation and processing. ACKNOWLEDGEMENTS

We are grateful to M. Goedert (Cambridge), P. Seubert (Athena Neurosciences), and P. Davies (New York) for providing us with the antibodies BR134, 12E8, and PHF-1, respectively. We thank Dr. S. Taylor for linguistic revisions. This study was supported by the Bundesministerium fu¨r Bildung, Forschung und Technologie (BMBF), Interdisciplinary Center for Clinical Research at the University of Leipzig (01 KS 9504, Project C1), and by the Deutsche Forschungsgemeinschaft (Ar 200/2–2).

REFERENCES 1. Anderton, B. H.; Brion, J.-P.; Couck, A.-M.; Davies, D. R.; Gallo, J.-M.; Hanger, D. P.; Ladhani, K.; Latimer, D. A.; Lewis, C.; Lovestone, S.; Marquardt, B.; Miller, C. C. J.; Mulot, S. F. C.; Reynolds, C. H.; Rupniak, T.; Smith, C. J.; Stabel, S.; Woodgett, J. Modulation of PHF-like tau phosphorylation in cultured neurons and transfected cells. Neurobiol. Aging 16:389 – 402; 1995. 2. Arias, C.; Sharma, N.; Davies, P.; Shafit–Zagardo, B. Okadaic acid induces early changes in microtubule-associated protein 2 and tau phosphorylation prior to neurodegeneration in cultured cortical neurons. J Neurochem. 61:673– 682; 1993. 3. Baum, L.; Seger, R.; Woodgett, J. R.; Kawabata, S.; Maruyama, K.; Koyama, M.; Silver, J.; Saitoh, T. Overexpressed tau protein in

cultured cells is phosphorylated without formation of PHF: Implication of phosphoprotein phosphatase involvement. Brain Res. Mol. Brain Res. 34:1–17; 1995. 4. Bialojan, C.; Takai, A. Inhibitory effect of a marine-sponge toxin, okadaic acid, on protein phosphatases. Specificity and kinetics. Biochem. J. 256:283–290; 1988. 5. Biernat, J.; Mandelkow, E. M.; Schroter, C.; Lichtenberg-Kraag, B.; Steiner, B.; Berling, B.; Meyer, H.; Mercken, M.; Vandermeeren, A.; Goedert, M.; Mandelkow, E. The switch of tau protein to an Alzheimer-like state includes the phosphorylation of two serine-proline motifs upstream of the microtubule binding region. EMBO J. 11:1593– 1597; 1992.

¨ RTNER ET AL. GA

542 6. Binder, L. I.; Frankfurter, A.; Rebhun, L. I. The distribution of tau in the mammalian central nervous system. J. Cell Biol. 101:1371–1378; 1985. 7. Binder, L. I.; Frankfurter, A.; Rebhun, L. I. Differential localization of MAP2 and tau in mammalian neurons in situ. Ann. N.Y. Acad. Sci. 466:145–166; 1986. 8. Brady, R. M.; Zinkowski, R. P.; Binder, L. I. Presence of tau in isolated nuclei from human brain. Neurobiol. Aging 16:479 – 486; 1995. 9. Bramblett, G. T.; Goedert, M.; Jakes, R.; Merrick, S. E.; Trojanowski, J. Q.; Lee, V. M. Abnormal tau phosphorylation at Ser396 in Alzheimer’s disease recapitulates development and contributes to reduced microtubule binding. Neuron 10:1089 –1099; 1993. 10. Burack, M. A.; Halpain, S. Site-specific regulation of Alzheimer-like tau phosphorylation in living neurons. Neuroscience 72:167–184; 1996. 11. Cleveland, D. W.; Hwo, S. Y.; Kirschner, M. W. Purification of tau, a microtubule-associated protein that induces assembly of microtubules from purified tubulin. J. Mol. Biol. 116:207–225; 1977a. 12. Cleveland, D. W.; Hwo, S. Y.; Kirschner, M. W. Physical and chemical properties of purified tau factor and the role of tau in microtubule assembly. J. Mol. Biol. 116:227–247; 1977b. 13. Fleming, L. M.; Johnson, G. V. Modulation of the phosphorylation state of tau in situ: the roles of calcium and cyclic AMP. Biochem. J. 309:41– 47; 1995. 14. Giasson, B. J.; Mushynski, W. E. Okadaic acid reversibly inhibits neurite outgrowth in embryonic dorsal root ganglion neurons. J. Neurobiol. 32:193–201; 1997. 15. Goedert, M.; Cohen, E. S.; Jakes, R.; Cohen, P. p42 MAP kinase phosphorylation sites in microtubule-associated protein tau are dephosphorylated by protein phosphatase 2A1. Implications for Alzheimer’s disease. FEBS Lett. 312:95–99; 1992. 16. Goedert, M.; Jakes, R.; Crowther, R. A.; Cohen, P.; Vanmechelen, E.; Vandermeeren, M.; Cras, P. Epitope mapping of monoclonal antibodies to the paired helical filaments of Alzheimer’s disease: Identification of phosphorylation sites in tau protein. Biochem. J. 301:871–877; 1994. 17. Goedert, M.; Jakes, R.; Vanmechelen, E. Monoclonal antibody AT 8 recognises tau protein phosphorylated at both serine 202 and threonine 205. Neurosci. Lett. 189:167–170; 1995. 18. Goedert M.; Spillantini M. G.; Jakes R.; Rutherford D.; Crowther R. A. Multiple isoforms of human microtubule-associated protein tau: Sequences and localization in neurofibrillary tangles of Alzheimer’s disease. Neuron 3:519 –526; 1989. 19. Goedert, M.; Spillantini, M. G.; Potier, M. C.; Ulrich, J.; Crowther, R. A. Cloning and sequencing of the cDNA encoding an isoform of microtubule-associated protein tau containing four tandem repeats: Differential expression of tau protein mRNAs in human brain. EMBO J. 8:393–399; 1989. 20. Gong, C. X.; Grundke–Iqbal, I.; Damuni, Z.; Iqbal, K. Dephosphorylation of microtubule-associated protein tau by protein phosphatase-1 and -2C and its implication in Alzheimer’s disease. FEBS Lett. 341:94 –98; 1994. 21. Gong, C. X.; Sing, T. J.; Grundke–Iqbal, I.; Iqbal, K. Phosphoprotein phosphatase activities in Alzheimer’s disease brain. J. Neurochem. 61:912–927; 1993. 22. Guttmann, R. P.; Erickson, A. C.; Johnson, G. V. Tau self-association: Stabilization with a chemical cross-linker and modulation by phosphorylation and oxidation state. J. Neurochem. 64:1209 –1215; 1995. 23. Hagestedt, T.; Lichtenberg, B.; Wille, H.; Mandelkow, E. M.; Mandelkow, E. Tau protein becomes long and stiff upon phosphorylation: correlation between paracrystalline structure and degree of phosphorylation. J. Cell Biol. 109:1643–1651; 1989. 24. Hasegawa, M.; Morishima–Kawashima, M.; Takio, K.; Suzuki, M.; Titani, K.; Ihara, Y. Protein sequence and mass spectrometric analyses of tau in the Alzheimer’s disease brain. J. Biol. Chem. 267:17047– 17054; 1992. 25. Himmler, A.; Drechsel, D.; Kirschner, M. W.; Martin, D. W., Jr. Tau consists of a set of proteins with repeated C-terminal microtubulebinding domains and variable N-terminal domains. Mol. Cell Biol. 9:138 –1388; 1989. 26. Hoffmann, R.; Lee, V. M. Y.; Leight, S.; Varga, I.; Otvos, L., Jr.

27. 28.

30. 31. 33. 32.

34. 35. 36.

37.

38. 39.

40. 41. 42. 43.

44.

45.

46.

47.

Unique Alzheimer’s disease paired helical filament specific epitopes involve double phosphorylation at specific sites. Biochemistry 36: 8114 – 8124; 1997. Jameson, L.; Frey, T.; Zeeberg, B.; Dalldorf, F.; Caplow, M. Inhibition of microtubule assembly by phosphorylation of microtubule-associated proteins. Biochemistry 19:2472–2479; 1980. Janke, C.; Holzer, M.; Klose, J.; Arendt, T. Distribution of isoforms of the microtubule-associated protein tau in gray and white matter areas of human brain: A 2D-gelelectrophoretic analysis. FEBS Lett. 379: 222–226; 1996. Laemmli, U. K. Cleavage of structural proteins during the assembly of bacteriophage T4. Nature 227:680 – 685; 1970. Lindwall, G.; Cole, R. D. Phosphorylation affects the ability of tau protein to promote microtubule assembly. J. Biol. Chem. 259:5301– 5305; 1984. Litersky, J. M.; Johnson, G. V. Phosphorylation of tau in situ: Inhibition of calcium-dependent proteolysis. J. Neurochem. 65:903– 911; 1995. Litersky, J. M.; Johnson, G. V. W.; Jakes, R.; Goedert, M.; Lee, M.; Seubert, P. Tau-protein is phosphorylated by cyclic AMP-dependent protein kinase and calcium/calmodulin-dependent protein kinase II within its microtubule-binding domains at Ser-262 and Ser-356. Biochem. J. 316:655– 660; 1996. Liu, W. K.; Yen, S. H. The state of phosphorylation of normal adult brain tau, fetal tau, and tau from Alzheimer paired helical filaments at amino acid residue Ser262. J. Neurochem. 66:1131–1139; 1996. Mandell, J. W.; Banker, G. A. A spatial gradient of tau-protein phosphorylation in nascent axons. J. Neurosci. 16:5727–5740; 1996. Matsuo, E. S.; Shin, R. W.; Billingsley, M. L.; Van deVoorde, A.; O9Connor, M.; Trojanowski, J. Q.; Lee, V. M. Biopsy-derived adult human brain tau is phosphorylated at many of the same sites as Alzheimer’s disease paired helical filament tau. Neuron. 13:989 – 1002; 1994. Merrick, S. E.; Demoise, D. C.; Lee, V. M. Site-specific dephosphorylation of tau-protein at Ser202/Thr205 in response to microtubule depolymerization in cultured human neurons involves protein phosphatase 2A. J. Biol. Chem. 271:5589 –5594; 1996. Migheli, A.; Butler, M.; Brown, K.; Shelanski, M. L. Light and electron microscope localization of the microtubule-associated tau protein in rat brain. J. Neurosci. 8:1846 –1851; 1988. Otvos, L., Jr.; Feiner, L.; Lang, E.; Szendrei, G. I.; Goedert, M.; Lee V. M. Monoclonal antibody PHF-1 recognizes tau protein phosphorylated at serine residues 396 and 404. J. Neurosci. Res. 39:669 – 673; 1994. Papasozomenos, S. C. Nuclear tau immunoreactivity in presenile dementia with motor neuron disease: A case report. Clin. Neuropathol. 14:100 –104; 1995. Papasozomenos, S. C.; Binder, L. I. Phosphorylation determines two distinct species of tau in the central nervous system. Cell Motil. Cytoskeleton 8:210 –226; 1987. Riederer, B. M.; Innocenti, G. M. Differential distribution of tau proteins in developing cat cerebral cortex and corpus callosum. Eur. J. Neurosci. 3:1134 –1145; 1991. Rosner, H.; Rebhahn, M.; Vacun, G.; Vanmechelen, E. Developmental expression of tau proteins in the chicken and rat brain: Rapid down-regulation of a paired helical filament epitope in the rat cerebral cortex coincides with tha transition from immature to adult tau isoforms. Int. J. Dev. Neurosci. 13:607– 617; 1995. Saito, T.; Ishiguro, K.; Uchida, T.; Miyamoto, E.; Kishimoto, T.; Hisanaga, S. In situ dephosphorylation of tau by protein phosphatase 2A and 2B in fetal rat primary cultured neurons. FEBS Lett. 376:238 – 242; 1995. Schwab, C.; Bondada, V.; Sparks, D. L.; Cahan, L. D.; Geddes, J. W. Postmortem changes in the levels and localization of microtubuleassociated proteins (tau, MAP2 a, MAP2 and MAP1B) in the rat and human hippocampus. Hippocampus 4:210 –225; 1994. Seubert, P.; Mawal–Dewan, M.; Barbour, R.; Jakes, R.; Goedert, M.; Johnson, G. V.; Litersky, J. M.; Schenk, D.; Lieberburg, I.; Trojanowski, J. Q.; Lee, V. M.-Y. Detection of phosphorylated Ser262 in fetal tau, adult tau, and paired helical filament tau. J. Biol. Chem. 270:18917–18922; 1995. Song, J.; Combs, C. K.; Pilcher, W. H.; Song, L. Y.; Utal, A. K.;

POSTMORTEM CHANGES OF TAU PHOSPHORYLATION

543

Coleman, P. D. Low initial tau phosphorylation in human brain biopsy samples. Neurobiol. Aging 18:475– 481; 1997. Sontag, E.; Nunbhakdi–Craig, V.; Lee, G.; Bloom, G. S.; Mumby, M. C. Regulation of the phosphorylation state and microtubulebinding activity of Tau by protein phosphatase 2A. Neuron 17:1201– 1207; 1996. Szendrei, G. I.; Lee, V. M.; Otvos, L., Jr. Recognition of the minimal epitope of monoclonal antibody Tau-1 depends upon the presence of a phosphate group but not its location. J. Neurosci. Res. 34:243–249; 1993. Tashiro, K.; Hasegawa, M.; Ihara, Y.; Iwatsubo, T. Somatodendritic localization of phosphorylated tau in neonatal and adult rat cerebral cortex. Neuroreport 8:2797–27801; 1997. Thurston, V. C.; Zinkowski, R. P.; Binder, L. I. Tau as a nucleolar protein in human nonneuronal cells in vitro and in vivo. Chromosoma. 105:20 –30; 1996. Trojanowski, J. Q.; Schuck, T.; Schmidt, M. L.; Lee, V. M. Distribu-

tion of tau proteins in the normal human central and peripheral nervous system. J. Histochem. Cytochem. 37:209 –215; 1989. Vincent, I.; Rosado, M.; Kim, E.; Davies, P. Increased production of paired helical filament epitopes in a cell culture system reduces the turnover of tau. J. Neurochem. 62:715–723; 1994. Watanabe, A.; Hasegawa, M.; Suzuki, M.; Takio, K.; Morishima– Kawashima, M.; Titani, K.; Arai, T.; Kosik, K. S.; Ihara, Y. In vivo phosphorylation sites in fetal and adult rat tau. J. Biol. Chem. 268:25712–25717; 1993. Weingarten, M. D.; Lockwood, A. H.; Hwo, S. Y.; Kirschner, M. W. A protein factor essential for microtubule assembly. Proc. Natl. Acad. Sci. USA 72:1858 –1862; 1975. Yamanoto, H.; Saitoh, Y.; Fukunaga, K.; Nishimura, H.; Miyamoto, E. Dephosphorylation of microtubule proteins by brain protein phosphatases 1 and 2A, and its effect on microtubule assembly. J. Neurochem. 50:1614 –1623; 1988.

48.

49. 50. 51. 52.

53. 54.

55. 56.