Mutation-Specific Functional Impairments in Distinct Tau Isoforms of Hereditary FTDP-17

See allHide authors and affiliations

Science  04 Dec 1998:
Vol. 282, Issue 5395, pp. 1914-1917
DOI: 10.1126/science.282.5395.1914


Tau proteins aggregate as cytoplasmic inclusions in a number of neurodegenerative diseases, including Alzheimer's disease and hereditary frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17). Over 10 exonic and intronic mutations in thetau gene have been identified in about 20 FTDP-17 families. Analyses of soluble and insoluble tau proteins from brains of FTDP-17 patients indicated that different pathogenic mutations differentially altered distinct biochemical properties and stoichiometry of brain tau isoforms. Functional assays of recombinant tau proteins with different FTDP-17 missense mutations implicated all but one of these mutations in disease pathogenesis by reducing the ability of tau to bind microtubules and promote microtubule assembly.

FTDP-17 comprises a group of hereditary neurodegenerative syndromes with diverse but overlapping clinical and neuropathological features (1, 2). The signature lesions of FTDP-17 brains are insoluble filamentous aggregates of hyperphosphorylated tau proteins similar to brains of patients with Alzheimer's disease (AD) (3). The absence of senile plaques and Lewy bodies indicates that neuronal loss could be a consequence of the tau pathology. Genetic linkage analysis demonstrated association of the tau gene with FTDP-17 (4–6). Missense mutations in exons 9 [Gly272 to Val (G272V)], 10 [Asn279 to Lys, Pro301 to Leu] (N279K, P301L), 12 [Val337 to Met] (V337M), and 13 [Arg406 to Trp] (R406W) (7) [numbered according to the longest tau isoform (8)] as well as three intronic mutations close to the 5′ splice site of exon 10 have been reported. Here we describe how five different FTDP-17 mutations differentially impair specific properties of tau (Fig. 1A).

Figure 1

(A) The tau mutations in FTDP-17 studied here are illustrated on a schematic of the longest central nervous system tau isoform (7). The V337M mutation is located in the interrepeat region between repeat 3 and 4, R406W near the COOH-terminus, and N279K and P301L within exon 10. The proposed RNA stem-loop structure at the exon 10 splice site is also shown. The C-to-U intronic mutation is located within the stem-loop at exon 10 splice site +14. (B and C) Cortical biopsy samples were immediately extracted in high-salt (0.75 M NaCl) reassembly (RA) buffer [0.1 M MES, 0.5 mM MgSO4, 1 mM EGTA, 2 mM dithiothreitol (pH 6.8), and a mixture of protease inhibitors (2 mM phenylmethysulfonyl fluoride andN-tosyl-l-phenylalanylchloromethyl ketone,N-tosyl-l-leucylchloromethyl ketone, leupeptin, pepstatin, and soybean trypsin inhibitor, each at 1 μg/ml)] as described (11). The heat-stable fractions were dephosphorylated with E. coli alkaline phosphatase overnight at 37°C (10). The nondephosphorylated (−) and dephosphorylated (+) samples along with a mixture of six isoforms of recombinant tau proteins were resolved on 7.5% SDS-polyacrylamide gel and transferred to nitrocellulose replica for immunoblotting with a rabbit polyclonal antiserum to recombinant tau (17026). 125I-labeled protein A was used for immunoblot quantitation (B). After dephosphorylation, each of the biopsy-derived tau isoforms aligned exactly with the corresponding recombinant human brain tau isoform. Quantitative data are shown in (C) where columns represent mean ± SEM values of biopsies (n = 4). The 2N, 1N, and 0N tau isoforms compose 9.5% ± 0.4%, 53.7% ± 0.4%, and 36.8% ± 0.8% of total tau.

In the adult human brain, six different tau isoforms are generated by alternative mRNA splicing of 11 exons in the tau gene (8, 9). These isoforms differ by the presence or absence of one or two (29 or 58 amino acids long, respectively) NH2inserts and by a COOH region with three (3Rtau) or four (4Rtau) imperfect 18–amino acid microtubule (MT)-binding repeats, each of which is separated by an interrepeat domain of 13 or 14 amino acids (8). We first determined the relative abundance of the six tau isoforms in brain by analyzing dephosphorylated tau (10) from biopsy samples of normal human cortex (11) (Fig. 1B). Quantitation of dephosphorylated 3Rtau isoforms with two, one, or no NH2-terminal inserts and the corresponding 4Rtau isoforms from four different brain biopsies showed that the ratio of 4Rtau to 3Rtau isoforms (4R/3R) was about 1 (Fig. 1C).

Consistent with the findings of Clark et al. (6), dephosphorylated insoluble tau from frontal cortex of affected family members with N279K or P301L mutations or with an intronic mutation close to the 5′ exon 10 splice site comigrated exclusively with 4Rtau isoforms (Fig. 2A). In contrast, dephosphorylated insoluble tau from affected members of families with a V337M or a R406W mutation comigrated with all six tau isoforms (Fig. 2A). We also compared the 4R/3R tau ratio in the soluble fractions of the same FTDP-17 samples with the ratios observed in similar samples from AD patients and age-matched controls (Fig. 2, B and C). A twofold increase was observed in the disinhibition-dementia-parkinsonism-amytrophy complex (DDPAC) sample (exon 10 5′ splice mutation). A similar twofold increase also was observed in samples from pallido-ponto-nigral degeneration (PPND) subjects with the N279K mutation. In sharp contrast, the brain from an affected member of the Montreal kindred with the exon 10 P301L mutation showed a 25% reduction in 4R/3R of soluble tau that may reflect the selective incorporation of 4Rtau isoforms into insoluble aggregates. These data suggest that the N279K and P301L mutations result in selective aggregation of 4Rtau into insoluble filamentous lesions by different mechanisms and that the N279K mutation may increase 4R/3R by enhanced splicing of exon 10.

Figure 2

Biochemical profile of tau in FTDP-17 brains. (A) Insoluble tau in FTDP-17 brains contains either 4Rtau or all six tau isoforms. Sarkosyl-insoluble tau was extracted as described (10) from autopsy-derived frontal gray matter of AD, PPND (N279K), DDPAC (exon 10 splice +14 mutation), Montreal (P301L), Seattle BK (V337M), or FTD004 (R406W) patients. A portion of each sample was dephosphorylated by incubation with E. colialkaline phosphatase (18). The dephosphorylated and nondephosphorylated samples plus a mixture of six recombinant human brain tau isoforms were resolved on 7.5% SDS polyacrylamide gels and transferred to a nitrocellulose membrane for immunoblot analysis with T14, a phosphorylation-independent monoclonal antibody to tau (17). (B to E) Soluble tau from FTDP brains with different 4R/3R tau ratios. To determine the isoform composition of soluble tau in affected frontal cortex (B and C) and unaffected cerebellum (D and E), the frontal gray or cerebellar white matter was extracted in high-salt RA buffer. The heat-stable supernatant was then dephosphorylated with E. coli alkaline phosphatase and resolved on SDS polyacrylamide gels. A polyclonal rabbit antiserum to tau was used for125I-labeled quantitative immunoblot analysis shown on the left (B and D). The 4R and 3R tau isoforms were quantified (C and E). *, P < 0.05; **, P < 0.01.

Because the cerebellum is nearly devoid of tau pathology in FTDP-17 patients (3), 4R/3R in soluble fractions of cerebellum should not be perturbed by an insoluble pool of aggregated tau. The finding of a 1.5- to 2-fold increase in 4R/3R in the PPND and DDPAC cerebellum but a 4R/3R of about 1 in the Montreal cerebellum (Fig. 2, D and E) is consistent with the idea that the N279K but not the P301L mutation increases splicing of exon 10. Analysis of affected brains from both Seattle BK (V337M) and FTD004 (R406W) kindreds showed no change in the 4R/3R ratio compared with the ratios in the soluble frontal cortex and cerebellum fractions of AD patients and controls.

To determine how different mutations in the tau gene alter specific tau functions, we compared the MT-binding properties of wild type (WT) and mutant recombinant tau proteins (12) with a fixed amount of taxol-stabilized MTs by measuring the amounts of bound and unbound tau (Fig. 3). Both WT and N279K mutant tau proteins bound equally well to MTs, with maximum binding (B max) of about 1.2 μM, whereas P301L, V337M, and R406W mutant tau proteins exhibited much lowerB max values of 0.88, 0.91, and 0.63 μM, respectively (Fig. 3 and Table 1). Surprisingly, the tau isoform with the R406W mutation had the lowest MT-binding capacity (about half that of WT tau) despite the fact that it is not near the MT binding domains. None of the mutant tau proteins was able to bind MTs as avidly as WT tau (Fig. 3 and Table 1), and the affinity of the R406W tau mutant was the lowest. Thus, our data suggest that one of the dysfunctional consequences of the FTDP-17 missense mutations (except for N279K) in tau is a reduction in their binding capacity as well as their binding affinity for MTs.

Figure 3

Microtubule-binding properties of tau are differentially altered by tau mutations. MTs were assembled from phosphocellulose purified bovine tubulin (Cytoskeleton, Inc.) as described (17). Taxol-stabilized MT dimer (3 μM) was incubated with 0 to 2 μM recombinant tau at 37°C for 20 min. A sucrose cushion was underlaid and the bound and free tau were separated by centrifugation (50,000g) at 25°C for 20 min. 125I-labeled quantitative immunoblot analysis with the polyclonal antiserum to tau (17026) was performed to determine bound and free tau. Each experiment was repeated at least three times with at least two recombinant tau preparations. Binding curves were generated by plotting bound versus free tau and fitting data by nonlinear regression with the standard binding equation (bound tau) =B max × (free tau)/[K d + (free tau)].

Table 1

Summary of MT binding and MT assembly promoting properties of WT and mutant tau. AU, absorbance unit.

View this table:

We next assessed the ability of WT and mutant tau proteins to promote MT assembly by the light-scattering technique (13, 14). This assay (Table 1) measures the lag time required to nucleate MTs, the ability of tau to initiate MT assembly, and the maximum amount of MT polymers that form in the presence of tau. The N279K mutation did not affect the ability of tau to promote MT assembly. However, P301L, R406W, and V337M mutant tau proteins nucleated MTs less efficiently than WT tau. The R406W tau mutant had the slowest initiation rate of MT assembly followed by P301L and V337M tau mutants (Fig. 4 and Table 1). The R406W tau mutant was the most defective in promoting MT assembly because the maximum polymer mass formed in the presence of this mutant was only about 75% that formed by WT tau. Despite a longer lag time for MT nucleation, the P301L tau mutant promoted MT assembly to the same extent as WT tau and N279K tau mutants.

Figure 4

Missense mutations reduce the ability of tau to promote MT assembly. To assess the abilities of WT and mutant tau proteins to promote MT assembly, light-scattering assays were performed (13, 14) with some modifications. Briefly, 35 μM bovine tubulin monomer was mixed at 4°C with 15 μM WT or mutant tau in RA buffer supplemented with 1 mM guanosine triphosphate. Turbidity was measured in quartz cuvettes by reading the absorbance at 350 nm (A 350) with a Beckman DU640 spectrophotometer. Samples were gradually warmed to 37°C with a defined rate and turbidity continuously measured at 1-min intervals. Duplicate samples were examined in each experiment and each experiment was repeated at least three times with at least two different recombinant tau preparations. A representative assay result is shown. In the absence of tau, turbidity remains at baseline, indicating minimal MT assembly. In the presence of WT tau, turbidity increased at an initiation rate of 0.15 absorbance unit per minute after a lag time of about 2 min, and plateaus within about 7 min when maximal MT assembly is reached.

This study demonstrates that some FTDP-17 mutations alter the MT-binding properties of tau, and others alter the ratio of 4R/3R tau isoforms. Intronic mutations (for example, in the DDPAC kindred) presumably alter 4R/3R tau ratios by weakening a stem-loop structure at the 5′ splice site of exon 10, which normally inhibits the inclusion of exon 10 into tau mRNA (5). How the N279K mutation increases 4Rtau production is less clear. Because the N279K mutation does not alter the binding of tau to MTs or decrease the ability of tau to promote MT assembly, the selective aggregation of 4Rtau into filamentous lesions probably results from overproduction of 4Rtau proteins by increasing the inclusion of exon 10 into more tau mRNAs. Presumably, this mutation could act by altering an exon-splicing enhancer (ESE) sequence in exon 10 (15). At the nucleotide level, the N279K mutation changes the normal sequence of TAAGAA to GAAGAA, a potential GAR (where R is a purine) ESE motif (6). These ESE sequences have been shown to bind to the SR family of proteins (16), which are essential factors for splice site selection in the alternative splicing of cellular and viral mRNAs.

In contrast, the missense mutations P301L, V337M, and R406W alter the biochemical properties of tau. The P301L mutation in exon 10 reduces the MT-binding affinity of 4Rtau and diminishes the number of 4Rtau proteins that could bind to MTs. As P301 is part of a highly evolutionarily conserved PGGG motif in all MT-binding repeats and is located in the MT-binding repeat specific for 4Rtau, it is not surprising that this missense mutation affects only 4Rtau. Similarly, the V337M mutation also interfered with the interaction of tau with MTs, but all six tau isoforms were affected because the V337M mutation is located in exon 12, which is included in all tau splice variants. Finally, despite the fact that R406W is not located in or near a MT-binding repeat, it had the most profound effects on the ability of tau to interact with MTs, and it is possible that a R to W substitution alters the conformation of tau, thereby reducing its binding to MTs, as is thought to occur after phosphorylation at Ser396 and Ser404 (17). The P301L, V337M, and R406W missense mutations may cause FTDP-17 by a partial loss of function mechanism. A slight decrease in the binding of mutant tau to MTs could destabilize MTs over time and have deleterious effects, including a disruption of axonal transport. However, these mutations also could result in a toxic gain of function because reduced binding of mutant tau leads to increased free cytoplasmic tau and an increased propensity to form insoluble aggregates. Similarly, the overproduction of 4Rtau due to an intronic mutation close to the exon 10 5′ splice site and to the N279K mutation also could result in a partial loss of function and a toxic gain of function. There are no data currently on the toxicity of 4R versus 3R tau protein in vitro. Indeed, tauopathies containing either 4R aggregates (some of the FTDP-17 kindreds described here), 3R aggregates (as observed in Pick's disease), or both 4R and 3R aggregates (as in AD and other FTDP-17) demonstrate that aggregation of both or either variety can cause disease.

  • * To whom correspondence should be addressed: Email: vmylee{at}


View Abstract

Stay Connected to Science

Navigate This Article