Nucleation of Huntingtin Aggregation Proceeds via Conformational Conversion of Pre-Formed, Sparsely-Populated Tetramers

Abstract

Pathogenic huntingtin exon-1 protein (htt^ex1), characterized by an expanded polyglutamine tract located between the N-terminal amphiphilic region and a C-terminal polyproline-rich domain, forms fibrils that accumulate in neuronal inclusion bodies, and is associated with a fatal, autosomal dominant neurodegenerative condition known as Huntington's disease. Here a complete kinetic model is described for aggregation/fibril formation of a htt^ex1 construct with a 35-residue polyglutamine repeat, htt^ex1Q₃₅. Using exchange NMR spectroscopy, it is previously shown that the reversible formation of a sparsely-populated tetramer of the N-terminal amphiphilic domain of htt^ex1Q₃₅, comprising a D₂ symmetric four-helix bundle, occurs on the microsecond time-scale and is a prerequisite for subsequent nucleation and fibril formation on a time scale that is many orders of magnitude slower (hours). Here a unified kinetic model of htt^ex1Q₃₅ aggregation is developed in which fast, reversible tetramerization is directly linked to slow irreversible fibril formation via conversion of pre-equilibrated tetrameric species to "active", chain elongation-capable nuclei by conformational re-arrangement with a finite, monomer-independent rate. The unified model permits global quantitative analysis of reversible tetramerization and irreversible fibril formation from a time series of ¹H-¹⁵N correlation spectra recorded during the course of htt^ex1Q₃₅ aggregation.

Keywords: NMR; elongation competent nuclei; huntingtin exon‐1 protein; pre‐nucleation tetramers; unified kinetic model of aggregation.

Figure 1

Schematic of the unified kinetic model of fast pre‐nucleation tetramerization coupled with slow monomer‐independent conversion, chain elongation and fibril surface‐mediated secondary nucleation steps of htt^ex1Q₃₅ fibrillization. k ₁ and k ₂ are association rate constants from monomer to dimer and from dimer to tetramer, respectively; k ₋₁ and k ₋₂ are the respective corresponding dissociation arte constants;    K _eq1 and K _eq2 are the dimerization and tetramerization equilibrium constants, respectively; k _c is the first order rate constant for the unimolecular conversion of pre‐nucleation transient tetramer T to nuclei P; and k ₊ and k _s are the rate constants for fibril elongation and surface‐mediated secondary nucleation, respectively. The equilibrium m⇌D⇌T is established “instantaneously”: K _eq,1 = [D]/[m]² = k ₁/k _‐1; K _eq,2 = [T]/[D]² = k ₂/k _‐2 and [T] = K _eq,2[D]² = K _eq,2(K _eq,1)² = (k ₂/k ₋₂)(k ₁/k _‐1]²[m]⁴.

Figure 2

Pre‐nucleation tetramerization on the microsecond timescale. A) Concentration‐dependent ¹H‐δ_ex (top) and ¹⁵N‐δ_ex (bottom) exchange‐induced chemical shifts. B) Time course of ¹H‐¹⁵N cross‐peak volumes and intensities during aggregation (top) and ¹H‐¹⁵N cross‐peak volume/intensity (V/I) ratios (bottom) as a function of the concentration of monomeric htt^ex1Q₃₅. The total sample concentration was 600 µm. The experimental data are shown as circles and the best fits to the unified kinetic scheme in Figure 1, under the assumption of equal changes in chemical shifts (Δω) between monomers and dimers and monomers and tetramers,^{[ 14} , ^{17 ]} are shown as continuous solid lines. All experiments were recorded at 5 °C and 800 MHz (see the Experimental Section) and the data for other residues used in the analysis are provided in Figure S1 (Supporting Information). The optimized values of Δω are given in Table S1 (Supporting Information).

Figure 3

Quantitative analysis of htt^ex1Q₃₅ aggregation decay profiles. A) Simulated time‐dependence of the concentration of htt^ex1Q₃₅ tetramers T in monomer units for htt^ex1Q₃₅ samples with total monomer concentrations m _tot of 160, 250, and 420 µm calculated using the values of K _eq1 and K _eq2 determined from the data in Figure 1. B) Time‐dependence of the average ¹H‐¹⁵N cross‐peak intensities for residues in the PRD domain of htt^ex1Q₃₅. The experimental data (shown as circles) were recorded at 5 °C and 800 MHz and normalized to the first time point (at t ≈ 0 h). The best‐fit curves are shown as black continuous lines and were obtained from a global fit to the kinetic scheme in Figure 1 and the model described by Equations (1) and (2) (see Experimental Section for details of the fitting procedure).

Figure 4

Simulation of the time dependence of mature fibrils M, nuclei P, and the M/P ratios during the course of htt^ex1Q₃₅ fibrillization at 5 °C using the optimized values of the rate constants (k _c = 0.07 ± 0.01 h⁻¹; k _s = 0.3 ± 0.04 M⁻¹h⁻¹; and k ₊ = 6.4 (±0.6) x 10⁵ M⁻¹h⁻¹) obtained from the global fits to the experimental data shown in Figures 2 and 3.