Kinetic Transition in Amyloid Assembly as a Screening Assay for Oligomer-Selective Dyes

Abstract

Assembly of amyloid fibrils and small globular oligomers is associated with a significant number of human disorders that include Alzheimer's disease, senile systemic amyloidosis, and type II diabetes. Recent findings implicate small amyloid oligomers as the dominant aggregate species mediating the toxic effects in these disorders. However, validation of this hypothesis has been hampered by the dearth of experimental techniques to detect, quantify, and discriminate oligomeric intermediates from late-stage fibrils, in vitro and in vivo. We have shown that the onset of significant oligomer formation is associated with a transition in thioflavin T kinetics from sigmoidal to biphasic kinetics. Here we showed that this transition can be exploited for screening fluorophores for preferential responses to oligomer over fibril formation. This assay identified crystal violet as a strongly selective oligomer-indicator dye for lysozyme. Simultaneous recordings of amyloid kinetics with thioflavin T and crystal violet enabled us to separate the combined signals into their underlying oligomeric and fibrillar components. We provided further evidence that this screening assay could be extended to amyloid-β peptides under physiological conditions. Identification of oligomer-selective dyes not only holds the promise of biomedical applications but provides new approaches for unraveling the mechanisms underlying oligomer versus fibril formation in amyloid assembly.

Keywords: amyloidosis; assembly pathway; dye fluorescence; oligomer formation; thioflavin T.

Figure 1

Sigmoidal versus biphasic amyloid kinetics and corresponding aggregate morphologies. Amyloid growth kinetics of hen egg-white lysozyme (hewL), recorded with thioflavin T (ThT), displaying either (a) sigmoidal (21 μM hewL) or (b) biphasic (350 μM hewL) growth kinetics (pH 2, 52 °C, 400 mM NaCl). Second panel in (a) is the same data on a semi-log scale to highlight the flat baseline. Spikes in the traces mark the times when aliquots were collected for imaging. Atomic force microscopy (AFM) images of the aggregate morphologies observed during (c) sigmoidal versus (d) biphasic growth of hewL, at the indicated time points. False color height scale: 3 nm.

Figure 2

FTIR spectra of hewL monomers, curvilinear fibrils (CFs) and Rigid fibrils (RFs). (a) Peak-matched FTIR absorption spectra of the Amide-I band for hewL monomers and isolated CFs and RFs at pH 2. CFs and RFs show a prominent peak in the characteristic amyloid band between 1630 and 1610 cm⁻¹. Yet their peak wavenumbers are slightly but reproducibly shifted with respect to each other. (b) The CF and RF differences spectra obtained after subtraction of the monomer spectrum in (a).

Figure 3

Summary of dye structures. Chemical structure of dyes in this study with indications of oligomer selectivity, together with the reference amyloid indicator dye Thioflavin T.

Figure 4

Fluorescence kinetics during sigmoidal and biphasic hewL amyloid assembly for various fluorescent dyes. Kinetics responses of (a) thioflavin T, (b) the amyloid dye X-34 (λ_ex = 370 nm, λ_em = 482 nm), (c) acridine orange (λ_ex = 495 nm, λ_em = 532 nm), and (d) acid fuchsin (λ_ex = 560 nm, λ_em = 590 nm) to amyloid growth of hewL at the indicated concentrations (in mg/mL) and under fixed solution conditions (pH 2, 52 °C, 450 mM NaCl). Orange (0.3 & 0.6 mg/mL) versus blue traces (2 & 4 mg/mL) represent sigmoidal versus biphasic growth conditions. Black traces are dye controls recorded from the buffer solution. For overall comparison, ThT traces are displayed on a linear scale, which obscures its weak biphasic response. Dye concentrations were 15 μM. Sharp initial transients resulted from the temperature dependence of dye fluorescence upon heating to 52 °C.

Figure 5

Crystal violet as the oligomer-selective dye (OSD). (a) Fractional change of 15 μM ThT (open symbols) versus 5 μM crystal violet (filled symbols) fluorescence during sigmoidal versus biphasic amyloid assembly (350 μM hewL, pH 2, 52 °C) in the presence of either 150 or 400 mM NaCl). (b) ThT (15 μM) kinetics of hewL undergoing sigmoidal fibril growth in the presence of increasing concentrations of CV. (c) Correlation of fractional CV versus ThT augmentation during biphasic versus sigmoidal growth. During the oligomer-dominated phase of biphasic kinetics, fractional changes of CV and ThT fluorescence are in a lock step and are of comparable magnitude. In contrast, CV becomes essentially unresponsive to the second, fibril-dominated phase. During sigmoidal fibril growth, CV barely increases 1.5-fold, compared to the nearly 100-fold fluorescence increase of ThT. Upon aligning the (noise-limited) CV traces with the upswing in ThT, their responses are strictly linearly correlated in time.

Figure 6

Evolution of amyloid aggregate morphology during biphasic kinetics (a) Staggered incubation of 350 μM hewL undergoing biphasic amyloid growth at pH 2, 52 °C, and 450 mM NaCl, recorded simultaneously with 15 μM ThT (●) and 5 μM CV (●). Fresh solutions were added to the 96-well plates at the moments indicated by the arrows. Corresponding total incubation periods are shown next to each arrow. (b) AFM images of aliquots imaged from wells incubated for the indicated time periods. While the initial phase of biphasic growth indicates the presence of gOs and increasing numbers of CFs, the late phase shows the simultaneous presence of RFs and CFs, often in direct contact with each other. The false color scale indicates aggregate heights. All images are 3 μm on a side.

Figure 7

Correlation of CV and ThT kinetics. (a,d) HewL amyloid formation (pH 2, 450 mM NaCl, T = 52 °C) at concentrations below (orange) and above (blue) the COC, simultaneously monitored with (a) ThT (15 μM) and (d) CV (5 μM). The traces are the average of the three recordings from separate wells. (b,e) Superposition of the CV (solid lines) and ThT (dashed lines) responses, after matching the CV data using the Γ-factor determined in (c) and (f), respectively. Subtracting the matched CV from the ThT trace yields the orange sigmoidal trace in (e). (c,f) Correlation of the fractional changes of CV versus ThT responses during (c) sigmoidal versus (f) biphasic growth kinetics. The Γ-coefficient indicates the ratio of the CV versus ThT response amplitude in the linear regime of each plot.

Figure 8

Transition from sigmoidal to biphasic kinetics and associated gO/CF formation for Aβ40 and Aβ42. (a) Transition in Aβ40 growth kinetics from pure sigmoidal (orange) to biphasic (blue) kinetics (pH 7.4, no salt). (b) Same as (a) but for Aβ42. Semilog plot emphasizes weak ThT response during gO/CF phase. (c) ThT fractional change during Aβ40 growth in physiological saline. Notice the significant increase in gO/CF amplitude relative to panel (a). (d–f) TEM images of samples of (a)—Aβ40 RFs following sigmoidal growth at 50 μM (d) versus biphasic growth at 150 μM, with gO/CFs formed within 1.5 days, and (e) mixtures of gO/CF and RFs after 6 days (f).