Three repeat isoforms of tau inhibit assembly of four repeat tau filaments

Abstract

Tauopathies are defined by assembly of the microtubule associated protein tau into filamentous tangles and classified by the predominant tau isoform within these aggregates. The major isoforms are determined by alternative mRNA splicing of exon 10 generating tau with three (3R) or four (4R) approximately 32 amino acid imperfect repeats in the microtubule binding domain. In normal adult brains there is an approximately equimolar ratio of 3R and 4R tau which is altered by several disease-causing mutations in the tau gene. We hypothesized that when 4R and 3R tau isoforms are not at equimolar ratios aggregation is favored. Here we provide evidence for the first time that the combination of 3R and 4R tau isoforms results in less in vitro heparin induced polymerization than with 4R preparations alone. This effect was independent of reducing conditions and the presence of alternatively spliced exons 2 and 3 N-terminal inserts. The addition of even small amounts of 3R to 4R tau assembly reactions significantly decreased 4R assembly. Together these findings suggest that co-expression of 3R and 4R tau isoforms reduce tau filament assembly and that 3R tau isoforms inhibit 4R tau assembly. Expression of equimolar amounts of 3R and 4R tau in adult humans may be necessary to maintain proper neuronal microtubule dynamics and to prevent abnormal tau filament assembly. Importantly, these findings indicate that disruption of the normal equimolar 3R to 4R ratio may be sufficient to drive tau aggregation and that restoration of the tau isoform balance may have important therapeutic implications in tauopathies.

Figure 1. Schematic representation of human tau and Coomassie blue-stained gel of recombinant tau.

(A) Full-length 3-repeat tau (3R2N) and 4-repeat tau (4R2N) containing N-terminal inserts encoded by exons 2 and 3, and microtubule (MT)-binding repeats encoded by exons 9–12. The 6 different tau isoforms are generated by alternative splicing of exons 2, 3, and 10 shown in white. The tau isoforms 3R0N and 4R0N would not include exons 2 and 3. Exclusion of alternatively spliced exon 10 generates 3-repeat tau isoforms. Inclusion or exclusion of alternatively spliced exons 2 and 3 generates 3R1N, 4R1N, 3R2N, or 4R2N tau isoforms. Potential heparin binding sites (HEPBS) are indicated by bars. (B) Purified recombinant tau proteins used in the study were loaded heavy at 2 ug per well and separated on 10% SDS-PAGE gels before staining with Coomassie brilliant blue R-250 to demonstrate purity.

Figure 2. Tau Isoform Assembly Kinetics.

The kinetics of 3R (A) and 4R (B) tau isoform assembly was measured using thioflavin S binding fluorescence in the presence of 0.04 mg/ml heparin with or without DTT. Each time point represents 2–5 experiments performed on separate days with background fluorescence (no tau present in reaction) subtracted from each experiment. (C) The thioflavin S data presented in A and B is the integrated value from thioflavin S binding curves as shown for 4R, 3R, and no tau reactions in the presence of DTT at day 1. (D–E) Electron micrographs of 4R0N tau assembly reactions containing 0.04 mg/ml heparin with DTT can be used to confirm the presence of tau filaments.

Figure 3. Ratio of 3R to 4R tau isoforms determines the extent of tau assembly.

Thioflavin S binding fluorescence of different ratios of 3R and 4R tau isoforms in the presence of 0.04 mg/ml heparin with or without DTT was measured at 48 hours or steady state. 3R0N tau isoforms mixed with increasing molar fractions of 4R0N tau under non-reducing (A) and reducing (B) conditions. Ratios of 3R0N to 4R0N were as follows: 8 µM 3R0N (0% 4R0N), 6 µM 3R0N to 2 µM 4R0N (25% 4R0N), 4 µM 3R0N to 4 µM 4R0N (50% 4R0N), 2 µM 3R0N to 6 µM 4R0N (75% 4R0N) and 8 µM 4R0N (100% 4R0N). 3R2N tau isoforms mixed with increasing molar fractions of 4R2N tau under non-reducing (C) and reducing (D) conditions. Ratios of 3R2N to 4R2N were the same as for 3R0N to 4R0N. Statistical analysis was performed using a one way ANOVA (for 30/40 mixes with and without DTT and for 32/42 mixes with DTT) or by Kruskal-Wallis One Way Analysis of Variance on Ranks (for 32/42 without DTT) and all pairwise multiple comparisons were done by the Student-Newman- Keuls Method, with p<0.05 considered significant. An * denotes significance relative to 3R tau assembly, # denotes significance relative to 4R tau assembly, and ∧ denotes significance relative to the equimolar ratio of 4R:3R tau assembly.

Figure 4. Tau aggregation decreased in mixed isoform reactions.

Assembly reactions of single tau isoforms and different ratios of 3R0N/4R0N tau isoforms were centrifuged at 100,000×g to separate free from aggregated or polymerized tau. Coomassie stained gels of the pellets containing polymerized tau from A) 3R0N mixed with 4R0N with the following samples for each gel: 8 µM 3R0N (0% 4R0N), 6 µM 3R0N to 2 µM 4R0N (25% 4R0N), 4 µM 3R0N to 4 µM 4R0N (50% 4R0N), 2 µM 3R0N to 6 µM 4R0N (75% 4R0N) and 8 µM 4R0N (100% 4R0N). Densitometric analysis of pelleting gels B) 3R0N mixed with 4R0N was performed using Image J software . Results were normalized to 4R values and relative % tau aggregation is shown. Statistical analysis was performed using a one way ANOVA and all pairwise multiple comparisons were done by the Student-Newman-Keuls Method, with p<0.05 considered significant. An * denotes significance relative to 4R tau assembly and a ∧ denotes significance relative to the equimolar ratio of 4R:3R tau assembly.

Figure 5. 3R Tau Inhibits 4R Tau Assembly.

Increasing molar concentrations (0, 1, 2, or 3 µM) of 3R or 4R tau were added to 8 µM 4R tau assembly reactions containing 0.04 mg/ml heparin and DTT, and Thioflavin S binding fluorescence was measured after 24 and 48 hours. 3R0N or 4R0N tau spiked into 4R0N tau reactions after 24 hours (A) and 48 hours (B). 3R1N and 4R1N tau spiked into 4R1N tau reactions after 24 hours (C) and 48 hours (D). 3R2N and 4R2N tau spiked into 4R2N tau reactions after 24 hours (E) and 48 hours (F). Fluorescence was normalized to 4R0N (A and B), 4R1N (C and D), and 4R2N (E and F) assembly and the relative fluorescence is shown. Statistical analysis was performed using a one way ANOVA and all pairwise multiple comparisons were done by the Student-Newman- Keuls Method, with p<0.05 considered significant. An * denotes significance relative to unspiked 4R tau assembly reactions. Addition of 3R tau isoforms to 4R tau assembly reactions showed a dose dependent inhibition of tau assembly, in spite of an overall increase in tau protein levels, relative to unspiked 4R tau assembly reactions (p = 0.018, p = 0.001, and p = 0.001 for 1 µM, 2 µM, and 3 µM 3R0N, respectively, for day 1 and p = 0.046 and p = 0.026 for addition of 2 µM and 3 µM 3R0N, respectively, for day 2). Additionally, 4R1N and 4R2N tau assembly reactions spiked with 3R1N and 3R2N tau isoforms, respectively, showed negative correlations between 4R tau assembly and additions of increasing concentrations of 3R tau isoforms (p<0.05 for both 4R1N and 4R2N) using the Spearman Rank Order Correlation test in Sigmaplot. Assembly of 4R tau reactions were either unchanged or showed a dose dependent increase in tau assembly following addition of more 4R tau isoforms.

Figure 6. 3R tau isoforms inhibit 4R tau aggregation.

Different 4R tau isoform assembly reactions alone or spiked with increasing molar concentrations of 3R tau isoforms, with or without N-terminal inserts, were centrifuged at 100,000×g to separate free and polymerized tau. Coomassie stained gels of the pellets containing polymerized tau from A) 8 µM 4R0N spiked with 3R0N, C) 8 µM 4R0N spiked with 4R0N, with the following samples for each gel: 4R (0 µM 3R), 4R spiked with 1 µM 3R, 4R spiked with 2 µM 3R, and 4R spiked with 3 µM 3R. Densitometric analysis of pelleting gels B) 4R0N spiked with 3R0N and D) 4R0N spiked with 4R0N, was performed using Image J software . Statistical analysis performed using the Spearman Rank Correlation test in Sigmaplot showed a negative correlation between 4R tau assembly and additions of increasing concentrations of 3R tau isoforms (p<0.05), but a positive correlation between 4R tau assembly and additions of increasing concentrations of 4R tau isoforms (p = 0.05).

Figure 7. 3R tau isoforms inhibit 4R tau filament formation.

A) Electron micrograph of tau assembly reaction containing 8 µM 4R0N, B) Electron micrograph of tau assembly reaction containing 8 µM 4R0N spiked with 2 µM 3R0N, C–D) Quantification of average tau filament number and average tau filament length per field was performed on 6–8 images, collected randomly at 10,000× magnification, using Image J freeware . E) Electron micrograph of tau assembly reaction containing 8 µM 4R2N, F) Electron micrograph of tau assembly reaction containing 8 µM 4R2N spiked with 2 µM 3R2N, and G–H) Quantification of tau filament number per field performed as above. Statistical analysis performed by t-test in SigmaPlot showed a significant decrease (p<0.0001) in tau filament formation and tau filament length (p<0.0001) in both 4R spiked with 3R and 4R2N spiked with 3R2N assembly reactions compared with unspiked 4R or unspiked 4R2N tau assembly reactions, despite an overall increase in total tau protein.

Figure 8. Changes in tau filament morphology in mixed isoform reactions.

Filaments from 48 hour single isoform or mixed isoform assembly reactions were examined by electron microscopy. Representative filaments from reactions containing only 3R0N tau were observed as A) loosely twisted paired helical filaments or B) straight filaments while those formed in single 4R0N reactions contained predominantly C–D) straight filaments. The filament ends in these pure reactions were mostly observed to be well defined and cleanly stained. Filaments from the mixed isoform reactions were also able to form E) loose paired helical filament and F) straight filaments though the filament ends G–J) were often not well defined and demonstrated splaying or amorphous aggregation. The scale bar is 100 nm.