Disease-related Modifications in Tau Affect the Interaction between Fyn and Tau
2005; Elsevier BV; Volume: 280; Issue: 42 Linguagem: Inglês
10.1074/jbc.m505895200
ISSN1083-351X
AutoresKiran Bhaskar, Shu-Hui Yen, Gloria Lee,
Tópico(s)14-3-3 protein interactions
ResumoMicrotubule-associated protein tau is the major component of the neurofibrillary tangles of Alzheimer disease (AD) and is genetically linked to frontotemporal dementias (FTDP-17). We have recently shown that tau interacts with the SH3 domain of Fyn, an Src family non-receptor tyrosine kinase, and is tyrosine-phosphorylated by Fyn on Tyr-18. Also, tyrosine-phosphorylated tau is present in the neuropathology of AD. To determine whether alterations in the tau-Fyn interaction might correlate with disease-related factors in AD and FTDP-17, we have performed real-time surface plasmon resonance studies on a panel of 21 tau constructs with Fyn SH3. We report that the interaction between Fyn SH3 and 3R-tau was 20-fold higher than that with 4R-tau. In addition, the affinity between 4R-tau and Fyn SH3 was increased 25–45-fold by phosphorylation-mimicking mutations or by FTDP-17 mutations. In vitro kinase reactions show that tau, with lower affinity SH3 interactions, exhibited a lower level of Tyr-18 phosphorylation under our reaction conditions. Lastly, we have demonstrated that tau is phosphorylated on Tyr-18 in the tau P301L mouse model for tauopathy (JNPL3). In summary, our results suggest that disease-related phosphorylation and missense mutations of tau increase association of tau with Fyn. Because these effects are mediated through the 4R component of the tau population, these results also have implications for the FTDP-17 diseases caused by increased expression of 4R-tau. Our data support a role for the Fyn-tau interaction in neurodegeneration. Microtubule-associated protein tau is the major component of the neurofibrillary tangles of Alzheimer disease (AD) and is genetically linked to frontotemporal dementias (FTDP-17). We have recently shown that tau interacts with the SH3 domain of Fyn, an Src family non-receptor tyrosine kinase, and is tyrosine-phosphorylated by Fyn on Tyr-18. Also, tyrosine-phosphorylated tau is present in the neuropathology of AD. To determine whether alterations in the tau-Fyn interaction might correlate with disease-related factors in AD and FTDP-17, we have performed real-time surface plasmon resonance studies on a panel of 21 tau constructs with Fyn SH3. We report that the interaction between Fyn SH3 and 3R-tau was 20-fold higher than that with 4R-tau. In addition, the affinity between 4R-tau and Fyn SH3 was increased 25–45-fold by phosphorylation-mimicking mutations or by FTDP-17 mutations. In vitro kinase reactions show that tau, with lower affinity SH3 interactions, exhibited a lower level of Tyr-18 phosphorylation under our reaction conditions. Lastly, we have demonstrated that tau is phosphorylated on Tyr-18 in the tau P301L mouse model for tauopathy (JNPL3). In summary, our results suggest that disease-related phosphorylation and missense mutations of tau increase association of tau with Fyn. Because these effects are mediated through the 4R component of the tau population, these results also have implications for the FTDP-17 diseases caused by increased expression of 4R-tau. Our data support a role for the Fyn-tau interaction in neurodegeneration. The ability of microtubule-associated protein tau to interact with microtubules is mediated by its carboxyl-terminal domain, which contains three or four copies of a microtubule-binding motif. In the middle portion of tau is a proline-rich domain that we have previously shown to contain a PXXP motif that interacts with the SH3 domain of Src family non-receptor tyrosine kinases (1Lee G. Newman S.T. Gard D.L. Band H. Panchamoorthy G. J. Cell Sci. 1998; 111: 3167-3177Crossref PubMed Google Scholar). In human neuroblastoma cells, tau co-immunoprecipitated with the non-receptor tyrosine kinase family member Fyn, and we have recently reported that Fyn phosphorylated tau at its amino terminus (Tyr-18) (2Lee G. Thangavel R. Sharma V.M. Litersky J.M. Bhaskar K. Fang S.M. Do L.H. Andreadis A. Van Hoesen G. Ksiezak-Reding H. J. Neurosci. 2004; 24: 2304-2312Crossref PubMed Scopus (321) Google Scholar). Although this phosphorylation does not affect the ability of tau to bind microtubules, this phosphorylated form of tau is present in the neurofibrillary tangles of Alzheimer disease (AD) 2The abbreviations used are: ADAlzheimer diseaseSPRsurface plasmon resonanceMALDI-TOFmatrix-assisted laser desorption ionization time-of-flightGSTglutathione S-transferase. as well as in preparations of paired helical filaments from AD brain (2Lee G. Thangavel R. Sharma V.M. Litersky J.M. Bhaskar K. Fang S.M. Do L.H. Andreadis A. Van Hoesen G. Ksiezak-Reding H. J. Neurosci. 2004; 24: 2304-2312Crossref PubMed Scopus (321) Google Scholar). We hypothesize that the interaction between Fyn and tau may have a role in the signaling events underlying neuropathology. A role for Fyn in AD has been suggested by immunocytochemical studies showing alterations in Fyn staining in AD brain (3Shirazi S.K. Wood J.G. Neuroreport. 1993; 4: 435-437Crossref PubMed Scopus (111) Google Scholar, 4Ho G.J. Hashimoto M. Adame A. Izu M. Alford M.F. Thal L.J. Hansen L.A. Masliah E. Neurobiol. Aging. 2005; 26: 625-635Crossref PubMed Scopus (75) Google Scholar). Moreover, brain slices from Fyn-deficient mice show protection from neurotoxicity induced by Aβ (5Lambert M.P. Barlow A.K. Chromy B.A. Edwards C. Freed R. Liosatos M. Morgan T.E. Rozovsky I. Trommer B. Viola K.L. Wals P. Zhang C. Finch C.E. Krafft G.A. Klein W.L. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6448-6453Crossref PubMed Scopus (3129) Google Scholar), the other prominent component of AD neuropathology, and Fyn depletion reduces the synaptotoxicity and neurotoxicity exhibited by an Aβ-based mouse model for AD (6Chin J. Palop J.J. Yu G.-Q. Kojima N. Masliah E. Mucke L. J. Neurosci. 2004; 24: 4692-4697Crossref PubMed Scopus (142) Google Scholar). Also, Fyn can activate GSK3β (7Lesort M. Jope R.S. Johnson G.V.W. J. Neurochem. 1999; 72: 576-584Crossref PubMed Scopus (219) Google Scholar) and cdk5 (8Sasaki Y. Cheng C. Uchida Y. Nakajima O. Ohshima T. Yagi T. Taniguchi M. Nakayama T. Kishida R. Kudo Y. Ohno S. Nakamura F. Goshima Y. Neuron. 2002; 35: 907-920Abstract Full Text Full Text PDF PubMed Scopus (291) Google Scholar), kinases known to phosphorylate tau at disease-associated sites. These findings suggest Fyn as a link between amyloid and tau pathology. Alzheimer disease surface plasmon resonance matrix-assisted laser desorption ionization time-of-flight glutathione S-transferase. Tau pathology is also found in frontotemporal dementia and Parkinsonism linked to chromosome 17 (FTDP-17), a group of age-related neurodegenerative diseases caused by autosomal dominant mutations in the tau gene (reviewed in Refs. 9Goedert M. Jakes R. Biochim. Biophys. Acta. 2005; 1739: 240-250Crossref PubMed Scopus (315) Google Scholar and 10Lee V.M. Goedert M. Trojanowski J.Q. Annu. Rev. Neurosci. 2001; 24: 1121-1159Crossref PubMed Scopus (2145) Google Scholar). The mutations, of which over 30 have been identified, can be divided into splice site and missense mutations (reviewed in Refs. 11D'Souza I. Schellenberg G.D. Biochim. Biophys. Acta. 2005; 1739: 104-115Crossref PubMed Scopus (113) Google Scholar and 12Hutton M. Ann. N. Y. Acad. Sci. 2000; 920: 63-73Crossref PubMed Scopus (82) Google Scholar). The splice site mutations predominantly result in the overexpression of mRNA encoding tau with four microtubule-binding repeats (4R-tau) relative to those with three repeats (3R-tau). As in AD, the tau lesions in FTDP-17 diseases comprise accumulations of abnormally phosphorylated tau, and many of the phosphorylated serine and threonine tau epitopes in AD have also been found in FTDP-17. In this study, we have investigated the impact of FTDP-17 mutations on the Fyn-tau interaction. Using the surface plasmon resonance (SPR) biosensor technique, we determined the kinetic parameters of the Fyn SH3-tau interaction for two alternatively spliced isoforms of tau (3R-tau and 4R-tau) and the effect of FTDP-17 and phosphorylation-mimicking mutations. Our results identify several functional differences between these proteins that were not present when similar assays were conducted with Src SH3. The significance of our SH3 interaction results were then tested by in vitro kinase assays; these assays indicated that the tyrosine phosphorylation of tau at Tyr-18 correlated with the SH3 domain interaction. Lastly, we determined that a mouse model of an FTDP-17 disease contained tyrosine-phosphorylated tau. Overall, our data provided evidence for the importance of the Fyn-tau interaction in the neurodegenerative process and led us to present a new hypothesis regarding the role of 4R-tau in disease. Expression of Recombinant Tau—Human tau with 3- or 4-microtubule-binding repeats (352- or 383-residue isoforms) were subcloned into either pET-3d or pET-17b vectors (Novagen, Madison, WI) for bacterial protein expression using standard techniques. The following mutations were introduced using site-directed mutagenesis (QuikChange mutagenesis kit, Stratagene, La Jolla, CA) (Fig. 1): 1) FTDP-17 missense mutations V337M, G272V, R406W, P301L, and R5H; and 2) phosphorylation-mimicking mutations S199D, S199D/S202D, S396D, S404D, S396D/S404D, and T231D/S235D. A construct with an internal deletion of the PXXP motif (ΔP) has been previously described (1Lee G. Newman S.T. Gard D.L. Band H. Panchamoorthy G. J. Cell Sci. 1998; 111: 3167-3177Crossref PubMed Google Scholar). All constructs were verified by DNA sequencing. tau proteins were expressed in Escherichia coli and purified as previously described (13Brandt R. Lee G. J. Biol. Chem. 1993; 268: 3414-3419Abstract Full Text PDF PubMed Google Scholar). The purity and size of wild type and mutant tau proteins were analyzed by gel electrophoresis and MALDI-TOF, showing homogeneous preparations of the expected molecular weights (supplemental Figs. S1 and S2). Prior to SPR analysis, protein was desalted, degassed, and equilibrated in HBS-EP buffer (10 mm Hepes, pH 7.4, 0.15 m NaCl, 3 mm EDTA, 0.005% (v/v) surfactant P20). Protein concentrations were determined by Bradford assay. Each protein was freshly diluted to obtain different concentrations using HBS-EP buffer containing 1 mm dithiothreitol prior to SPR analysis. GST-Fyn SH3 and GST-Src SH3 fusion proteins were affinity-purified on glutathione-Sepharose beads as described by the manufacturer (Amersham Biosciences). For SPR, GST-Fyn SH3 and GST-Src SH3 were used at 10 μg/ml in HBS-EP buffer containing 1 mm dithiothreitol. Surface Plasmon Resonance—SPR experiments were performed at 25 °C using the BIAcore 3000 (BIAcore Inc., Piscataway, NJ). Goat anti-GST antibody was immobilized on two flow cells (FC1 and FC2) of a CM4 sensor chip by amine coupling at pH 5.0 in 10 mm sodium acetate, according to BIAcore instructions. Approximately 6200 resonance units corresponding to 6.2 ng/mm2 goat anti-GST antibody was immobilized on each surface. A blank surface (FC3) was made by ethanolamine deactivation of the activated dextran surface. For SPR, GST fusion proteins were diluted to 10 μg/ml in HBS-EP buffer and injected into FC2. This resulted in 150–200 resonance units on the surface matrix, corresponding to 150–200pg/mm2 of GST fusion protein bound to the GST antibody in FC2. Next, purified tau protein was injected on all three flow cells (FC1, -2, and -3) at a flow rate of 40 μl/min and resonance units measured. After each binding cycle, the surfaces were regenerated with 10 mm glycine (pH 2.2) to remove the GST fusion protein + Tau complex from the antibody surface. The first flow cell (FC1) was the minus GST protein control channel and was subtracted from the sample channel (FC2) during the run. Bulk refractive index contributions were therefore expected to be 0 or not significant in reference-subtracted sensorgrams. To obtain the association constant (on-rate or ka) and the dissociation constant (off-rate or kd), the interaction between GST fusion protein and different tau constructs was measured at tau concentrations of 1250, 1000, 750, 500, and 250 nm. To correct for nonspecific binding, blank runs were performed with HBS-EP buffer on all three surfaces before and after each binding analysis; this value was subtracted prior to kinetic analysis. The resulting sensorgrams were analyzed using the 1:1 Langmuir binding model in the BIAevaluation version 3.1 software program (BIAcore Inc.). Quality of the kinetic fit was determined by the residuals and the χ2 values (supplemental Fig. S3). Kinetic constants ka and kd were calculated by local fit analysis. The apparent equilibrium association constants (KA = ka/kd) and the dissociation constants (KD = kd/ka) were calculated from the ka and kd values for each tau concentration. All interaction analyses were repeated three times. TABLES ONE and TWO show average KD values (mean ± S.D.); kinetic constants, ka and kd and KA are shown in the supplemental material (Tables S1 and S2). Differences in KD values for tau mutants were analyzed for significance by one-way analysis of variance followed by the Tukey-Kramer multiple comparisons test.TABLE ONEEquilibrium dissociation constants for tau and GST-Fyn SH3 interactionTau construct3R-tau4R-tauKD (μm)Wild type0.326 ± 0.086.77 ± 0.5ap < 0.001 versus wild type 3R-tau.ΔP4.0 ± 0.7ap < 0.001 versus wild type 3R-tau.NDT231D/S235D2.8 ± 0.4ap < 0.001 versus wild type 3R-tau.NDS202D31.0 ± 8.0ap < 0.001 versus wild type 3R-tau.0.26 ± 0.06bp < 0.001 versus wild type 4R-tau.S199D/S202D25.0 ± 1ap < 0.001 versus wild type 3R-tau.0.16 ± 0.01bp < 0.001 versus wild type 4R-tau.S396D0.2 ± 0.2NDcND, not determined.S404DNDcND, not determined.0.15 ± 0bp < 0.001 versus wild type 4R-tau., dS.D. is <0.01.S396D/S404D0.29 ± 0.030.16 ± 0.08bp < 0.001 versus wild type 4R-tau.V337M0.17 ± 0.040.18 ± 0.06bp < 0.001 versus wild type 4R-tau.G272V0.19 ± 0.010.18 ± 0.05bp < 0.001 versus wild type 4R-tau.R406W0.20 ± 0dS.D. is <0.01.0.15 ± 0.02bp < 0.001 versus wild type 4R-tau.R5H0.13 ± 0.020.17 ± 0.03bp < 0.001 versus wild type 4R-tau.P301LNDcND, not determined.0.16 ± 0.09bp < 0.001 versus wild type 4R-tau.a p < 0.001 versus wild type 3R-tau.b p < 0.001 versus wild type 4R-tau.c ND, not determined.d S.D. is <0.01. Open table in a new tab TABLE TWOEquilibrium dissociation constants for tau and GST-Src SH3 interactionTau construct3R-tau4R-tauKD (μm)Wild type0.17 ± 0.010.26 ± 0.05T231D/S235D5.36 ± 0.2ap < 0.001 versus wild type 3R-tau.NDbND, not determined.V337M0.19 ± 0.040.19 ± 0.06R406W0.23 ± 0.010.15 ± 0.05P301LNDbND, not determined.0.27 ± 0.09a p < 0.001 versus wild type 3R-tau.b ND, not determined. Open table in a new tab In Vitro Kinase Assay—In vitro tyrosine kinase reactions (40 μl) containing bacterially expressed wild type or mutant tau (2 μm), Src (11 units), or Fyn (2 units) (Upstate Biotechnology, Inc. Charlottesville, VA), 1.56 mm MnCl2, 12.5 mm MgCl2, 0.125 mm EGTA, 15.6 μm NaVO4, 0.125 mm dithiothreitol, 62.5 μm ATP, and 6.25 mm Tris-HCl (pH 7.2) were incubated for 5, 10, 20, or 60 min or 4 or6hat30°C. The samples were analyzed for phosphorylation at Tyr-18 by immunoblotting with anti-PY18 antibody (2Lee G. Thangavel R. Sharma V.M. Litersky J.M. Bhaskar K. Fang S.M. Do L.H. Andreadis A. Van Hoesen G. Ksiezak-Reding H. J. Neurosci. 2004; 24: 2304-2312Crossref PubMed Scopus (321) Google Scholar), whereas total tau was detected using Tau46.1 antibody. Quantitation of the ECL signal was performed on a Bio-Rad Fluor-S™ MultiImager system and Quantity One® image analysis software. Statistical significance was determined by unpaired t test. Immunocytochemistry—For immunocytochemistry, paraffin-embedded brain sections (4 μm thick) were prepared from 8-month-old P301L tau-transgenic mice as previously described (14Lewis J. McGowan E. Rockwood J. Melrose H. Nacharaju P. Van Slegtenhorst M. Gwinn-Hardy K. Murphy P.M. Baker M. Yu X. Duff K. Hardy J. Corral A. Lin W.L. Yen S.H. Dickson D.W. Davies P. Hutton M. Nat. Genet. 2000; 25: 402-405Crossref PubMed Scopus (1141) Google Scholar). Sections were sequentially immunostained, first with TG3 (1:10, generously provided by Dr. Peter Davis (15Jicha G.A. Lane E. Vincent I. Otvos Jr., L. Hoffmann R. Davies P. J. Neurochem. 1997; 69: 2087-2095Crossref PubMed Scopus (218) Google Scholar)) or AT8 (1:200, Pierce) followed by 9G3-biotin. Rhodamine-conjugated anti-mouse IgM or IgG secondary antibodies were used for TG3 or AT8 detection. Alexa-488-conjugated streptavidin (1:200, Molecular Probes, Eugene, OR) was used for 9G3 detection. Sections were visualized using the Nikon E800 microscope. 9G3-biotin was prepared using the EZ-Link NHS-PEO solid phase biotinylation kit (Pierce). 3R-tau Binds to Fyn SH3 with a Higher Affinity than 4R-tau—To investigate the interaction between Fyn and tau, we used SPR spectroscopy to quantitate the interaction between tau and the SH3 domain of Fyn. We started by comparing two wild type tau proteins that are generated by alternative splicing. Adult human brain expresses both 3R-tau and 4R-tau, whereas fetal human expresses only 3R-tau. In this study, we have used the shortest 3R-tau and 4R-tau isoforms. Neither contain inserts at the amino terminus and differ only in the number of microtubule-binding repeats. The real-time interaction between 3R-tau and GST-Fyn SH3 was measured at 25 °C by sequentially injecting proteins onto sensor chip surfaces with immobilized goat anti-GST antibody. SPR responses of injected tau were recorded with and without a prior injection of GST-Fyn SH3. Prior to kinetic analysis, the tau-alone control response was subtracted from the tau-plus GST-Fyn SH3 response. Our data indicated that 3R-tau interacted with GST-Fyn SH3 with an equilibrium dissociation constant (KD) of 0.32 μm (TABLE ONE, representative SPR sensorgram shown in Fig. 2). This value was the average of five KD values calculated from the measured kinetic rates of association (ka) and dissociation (kd) measured at five tau concentrations (average ka and kd values are shown in supplemental Tables S1 and S2). It should be noted that the value of KD is inversely proportional to the affinity between interacting proteins and that, therefore, a lower KD indicates a higher affinity between two interacting partners. Our previous data obtained by GST "pull-down" assay identified a 233PXXP motif in tau that interacted with the SH3 domain of Fyn (1Lee G. Newman S.T. Gard D.L. Band H. Panchamoorthy G. J. Cell Sci. 1998; 111: 3167-3177Crossref PubMed Google Scholar). SPR analysis confirmed that the 3R-tau mutant lacking the 233PXXP motif (ΔP) interacted very poorly with GST-Fyn SH3. The interaction displayed 12-fold lower affinity (KD = 4 μm) compared with that of 3R-tau (TABLE ONE). In addition, we measured the interaction between Fyn SH3 and a tau mutant, where both Thr-231 and Ser-235 had been replaced with aspartic acid, thus mimicking phosphorylation at the PXXP motif. Thr-231 in tau is phosphorylated by several kinases, including GSK3β and cdk5, and the presence of phospho-Thr-231, as well as phospho-Ser-235, in AD have been shown by phospho-specific antibody probes and mass spectrometry data (reviewed Refs. 16Sergeant N. Delacourte A. Buee L. Biochim. Biophys. Acta. 2005; 1739: 179-197Crossref PubMed Scopus (234) Google Scholar, 17Stoothoff W.H. Johnson G.V. Biochim. Biophys. Acta. 2005; 1739: 280-297Crossref PubMed Scopus (366) Google Scholar, 18Iqbal K. Alonso A del C. Chen S. Chohan M.O. El-Akkad E. Gong C.X. Khatoon S. Li B. Liu F. Rahman A. Tanimukai H. Grundke-Iqbal I. Biochim. Biophys. Acta. 2005; 1739: 198-210Crossref PubMed Scopus (742) Google Scholar). The interaction between Fyn SH3 and the 3R-tau-T231D/S235D mutant showed an 8-fold reduction in affinity (KD = 2.8 μm; TABLE ONE). These results confirm the importance of the PXXP motif in mediating the tau-Fyn SH3 interaction. Surprisingly, although 4R-tau interacted with GST-Fyn SH3, the affinity (KD = 6.77 μm) was 20-fold lower compared with 3R-tau (TABLE ONE, representative SPR sensorgram shown in Fig. 2). To confirm these results, we tested 4R-tau prepared from two different prokaryotic plasmids, each plasmid sequence having been confirmed by DNA sequencing and protein analyzed by MALDI-TOF. Our data indicated that the alternative splicing of tau dramatically alters the interaction of tau with Fyn. Previously, we had shown that tau also interacted with the SH3 domain of Src. Therefore, as a comparison, we also analyzed the interaction between tau and Src SH3. For 3R-tau, the interaction with Src SH3 resembled that with Fyn SH3, showing a comparable KD value of 0.17 μm (TABLE TWO). However, for 4R-tau, the interaction with the Src SH3 showed a much lower KD than for Fyn SH3 (KD = 0.26 μm for Src versus KD = 6.77 μm for Fyn), suggesting that for 4R-tau, binding to Src was preferred relative to Fyn. As measured by KD, there was no significant difference between the interaction of 3R-tau or 4R-tau with Src SH3. Therefore, the observed difference between 3R-tau and 4R-tau was specific for the tau-Fyn interaction. Amino Acid Replacements Far from the PXXP Impact on SH3 Interactions—As our data suggested that phospho-Thr-231 and phospho-Ser-235 would interfere with the association of tau with SH3, we tested two other disease-related phosphorylation sites (Fig. 1). Mass spectrometry, as well as immunocytochemical data, have also established phospho-Ser-199/202 and phospho-Ser-396/404 as abnormally phosphorylated tau sites in disease. To study the effects of phosphorylation at these sites, which are recognized by monoclonal antibodies AT8 and PHF1, respectively, we synthesized tau constructs that mimicked phosphorylation by converting Ser-199, -202, -396, and -404, either individually or in pairs, to aspartic acid. When 3R-tau was mutated to mimic phosphorylation at the AT8 site, either as a single S202D or double S199D/S202D replacement, the protein showed a significant reduction in affinity (95 and 76-fold, respectively) to the Fyn SH3 (TABLE ONE). Interestingly, an opposite effect was observed when 4R-tau carrying these same mutations were tested. 4R-S202D and 4R-S199D/S202D displayed a 26- and 42-fold increase, respectively, in affinity. Mutations in 3R-tau that mimic phosphorylation at the PHF1 site, either as a single S396D or double S396D/S404D mutant, resulted in modest differences in binding that were not statistically significant. In contrast, similar mutations introduced in 4R-tau (4R-S404D and 4R-S396D/S404D) displayed 45- and 42-fold increases in affinity, respectively (TABLE ONE). Therefore, although 3R-tau phosphorylation-mimicking mutants have either decreased or unchanged binding affinities to Fyn SH3, similar 4R mutants have enhanced binding affinities. To assess whether FTDP-17 missense mutations affect the tau-Fyn interaction, we first investigated four mutations (V337M, G272V, R406W, and R5H) in 3R-tau (Fig. 1). In comparing the binding kinetics of the mutants to GST-Fyn SH3, all four mutants tested showed modest 2–3-fold increases in binding affinity that were not statistically significant. In contrast, when we investigated the same four mutants and P301L in 4R-tau, all mutants showed a significant increase in binding affinity to Fyn SH3. The affinities for the interactions with the V337M, G272V, R406W, and R5H mutants were all at least 30-fold higher than wild type 4R-tau, with the P301L mutant showing the highest increase (42-fold) (TABLE ONE). Lastly, FTDP-17 mutations, either in 3R or 4R-tau, did not significantly alter their binding to Src SH3 (TABLE TWO). In summary, the SPR data has indicated that pseudophosphorylation and FTDP-17 mutations differentially affect 3R- and 4R-tau with regards to their interaction with the Fyn SH3 domain. Our results are summarized in Fig. 3, where we have graphed the reciprocal of the KD (1/KD) as an indication of binding affinity, because the KD value reflects the rate of the dissociation of the interaction. For the graph, 1/KD was normalized to the 3R-tau-Fyn SH3 value, which was set at 100. In addition, we used the calculated mean of the KD values gathered for each group of similar modifications, because the differences between the values within each group were not statistically significant. We noted that 4R-tau binds Fyn SH3 less well than 3R-tau and that pseudophosphorylation at AT8 or PHF1 or FTDP-17 mutation increases the affinity of 4R-tau to match that of unmodified 3R-tau (Fig. 3, left). In contrast, 3R-tau was affected only by pseudophosphorylation at AT8, which resulted in a decrease in affinity. The differences noted between the splicing isoforms and for the FTDP-17 mutations were not found for the Src SH3 interaction (Fig. 3, right). 233PXXP Motif in Tau Is Essential for Tyr-18 Phosphorylation by Fyn or Src Kinase—A functional consequence of SH3 domain interactions for Fyn or Src might be to direct substrates to the catalytic domain of the kinase (19Weng Z. Thomas S.M. Rickles R.J. Taylor J.A. Brauer A.W. Seidel-Dugan C. Michael W.M. Dreyfuss G. Brugge J.S. Mol. Cell. Biol. 1994; 14: 4509-4521Crossref PubMed Scopus (206) Google Scholar). Using co-transfected COS cells, we have recently determined that Tyr-18 in human tau was phosphorylated by Fyn and Src (2Lee G. Thangavel R. Sharma V.M. Litersky J.M. Bhaskar K. Fang S.M. Do L.H. Andreadis A. Van Hoesen G. Ksiezak-Reding H. J. Neurosci. 2004; 24: 2304-2312Crossref PubMed Scopus (321) Google Scholar). If the interaction between tau and the SH3 domain of Fyn or Src was critical for the phosphorylation of Tyr-18 of tau, in vitro phosphorylation of Tyr-18 might be expected to be consistent with the SPR results. Therefore, to extend our understanding of the functional significance of the tau-SH3 domain interaction, we performed in vitro kinase reactions with wild type tau and tau mutants using Fyn or Src kinases and a phospho-Tyr-18-specific antibody (anti-PY18) to detect the tyrosine phosphorylation of tau (2Lee G. Thangavel R. Sharma V.M. Litersky J.M. Bhaskar K. Fang S.M. Do L.H. Andreadis A. Van Hoesen G. Ksiezak-Reding H. J. Neurosci. 2004; 24: 2304-2312Crossref PubMed Scopus (321) Google Scholar). When the 3R-tau ΔP mutant was used as a substrate for Fyn, there was a decrease in PY18 (Fig. 4, lane 2). To quantitate the level of phosphorylation, we measured the ratio of PY18 to total tau. This parameter indicated that ΔP was phosphorylated 15-fold less than wild type tau (Fig. 4). Similarly, the T231D/S235D mutant was phosphorylated 12-fold less than wild type tau (Fig. 4, lane 3). Assays performed with Src on the ΔP and T231D/S235D proteins similarly showed 19- and 4-fold less PY18, respectively (Fig. 4, lanes 5 and 6). The correlation between these results and the SPR data suggest that the in vitro phosphorylation of tau at Tyr-18 by Fyn or Src kinase is strongly dependent on the interaction between the PXXP motif and the SH3 domain. Because 3R-ΔP and 3R-T231D/S235D proteins showed a positive correlation between SH3 binding and Tyr-18 phosphorylation, one might anticipate that the 4R-tau and FTDP-17 mutants would exhibit similar results. However, when kinase reactions were carried out with these proteins under the same conditions as above (reaction incubation time of 6 h), we found minimal differences between 3R-tau, 4R-tau, and FTDP-17 mutants (data not shown). We then considered the fact that, although the ΔP and T231D/S235D proteins had alterations that directly impacted on the site of interaction between Fyn and tau, the inclusion of exon 10 in 4R-tau and the FTDP-17 mutations represented changes that were all located some distance away from the PXXP motif (Fig. 1). In addition, because SPR data is derived from real-time binding measurements, the 6-h incubation condition might not reflect kinetic effects exerted by indirect conformational alterations. Moreover, we had noted that all of the changes in KD had resulted from changes in the measured rate of association (ka). Thus, we performed a time course of phosphorylation to determine whether the inclusion of exon 10 or the FTDP-17 mutations might affect the phosphorylation of Tyr-18 primarily during the initial stages of the reaction. By terminating the Fyn kinase reaction after 0.5, 1, 2, 5, 10, 20, 60, and 240 min of incubation, we found that the levels of phosphorylation, as determined by the ratio of PY18 to total tau, were higher for 3R-tau than 4R-tau (Fig. 5A). The measured differences were statistically significant, as determined by unpaired t test. In addition, although phosphorylation of the 3R-V337M mutant was similar to 3R-tau (Fig. 5B), the 4R-R406W and P301L mutants had significantly higher levels of phosphorylation when compared with 4R-tau (Fig. 5C). These findings support our SPR results. Tyr-18 Phosphorylated Tau in P301L Tau Transgenic Mouse Brain—Given that FTDP-17 mutations affected the tyrosine phosphorylation of tau and our previous finding that Tyr-18 of tau was phosphorylated in AD, we tested an FTDP-17 mouse model (14Lewis J. McGowan E. Rockwood J. Melrose H. Nacharaju P. Van Slegtenhorst M. Gwinn-Hardy K. Murphy P.M. Baker M. Yu X. Duff K. Hardy J. Corral A. Lin W.L. Yen S.H. Dickson D.W. Davies P. Hutton M. Nat. Genet. 2000; 25: 402-405Crossref PubMed Scopus (1141) Google Scholar) for Tyr-18 phosphorylation of tau. Brain sections from an 8-month-old P301L tau transgenic mouse were probed with an antibody specific for phosphorylated Tyr-18 (9G3), whereas AT8 or TG3 were each used as a control antibody to identify tau pathology. We found 9G3 labeling in the brain stem, basal ganglia, entorhinal cortex, and hippocampus, coinciding with AT8-positive neurons (Fig. 6); TG3 also labeled the same neurons as 9G3. At the subcellular level, we noted that the TG3 staining pattern was often punctate (Fig. 6, insets), as reported for the pre-tangle st
Referência(s)