Thioflavin T (ThT) Assays: Measuring Peptide Fibril Formation

What thioflavin T reports

Thioflavin T (ThT) is a benzothiazole dye that fluoresces weakly in solution and strongly when it binds the cross-beta structure of amyloid-type fibrils. In early work with synthetic amyloid-beta peptides, the free dye had excitation and emission maxima near 385 and 445 nm; bound to fibrils, a new excitation maximum appeared at 450 nm with enhanced emission at 482 nm.^[1] Monomeric and dimeric peptide gave no signal, guanidine dissociation of the aggregates removed it, and high salt had no effect. Insulin fibrils also fluoresced intensely.

So ThT is a reporter of fibril mass, not of aggregation in general. Amorphous aggregates, small oligomers and particles of other shapes can be weakly stained or invisible. A flat ThT trace means "no ThT-positive fibrils under these conditions", which is a narrower statement than "no aggregation".

Choosing the dye concentration

The dye is not a neutral observer at every concentration. A systematic study using three amyloid-forming proteins found that ThT fluorescence correlated linearly with fibril concentration across dye concentrations from 0.2 to 500 µM, that the signal at a fixed fibril level traced a bell-shaped curve against dye concentration, and that maximal fluorescence appeared at 20 to 50 µM.^[2] In kinetic runs, 20 µM ThT or less had little effect on aggregation, while 50 µM or more could change the shape of the curves, depending on the protein.

For in situ kinetics, where dye is present throughout, keep it near 10 to 20 µM and check once that less dye gives the same lag time. End-point reads, where dye is added just before reading, can use more.

Worked arithmetic: stock, wells and normalization

Dye stock first. ThT is usually dissolved in water or buffer, passed through a 0.2 µm filter to remove particles, and its concentration checked by absorbance using the Beer-Lambert law. A commonly used extinction coefficient is about 36,000 M⁻¹ cm⁻¹ at 412 nm. If a 1:100 dilution reads A412 = 0.72 in a 1 cm cell, the diluted solution is 0.72 ÷ 36,000 = 20 µM and the stock is 2.0 mM. Adding 2 µL of that stock to a 200 µL well gives 20 µM, at the upper end of the range above. The dilution calculator does the C1V1 = C2V2 step.

Normalization next. With blank-subtracted readings, the fraction of final signal is (F − F0) ÷ (Fmax − F0). For F0 = 120, Fmax = 2,520 and F = 1,320, that is 1,200 ÷ 2,400 = 0.50, the curve midpoint. That estimates how far fibril formation has progressed, not how much peptide is still soluble; for that, spin down fibrils and measure the peptide left in the supernatant.

Reading the curve: lag phase, growth and plateau

A typical ThT trace is sigmoidal. In a classic insulin study the kinetics were described by the lag time for formation of stable nuclei (nucleation) and an apparent rate constant for fibril growth (elongation).^[3] Adding preformed fibrils as seeds eliminated the lag phase; higher insulin concentration shortened the lag and sped up growth; acidic pH gave shorter lags and faster growth than neutral pH; higher ionic strength shortened the lag but slowed growth; and agitation weakened the effects of concentration and ionic strength.

A sigmoid F(t) = F0 + (Fmax − F0) ÷ (1 + e^(−(t − t50)/τ)) is a common empirical fit. Under one widely used convention the apparent growth rate is 1/τ and the lag time is t50 − 2τ. With t50 = 6.0 h and τ = 0.8 h, the growth rate is 1.25 h⁻¹ and the lag time is 6.0 − 1.6 = 4.4 h; at that point the curve has reached 1 ÷ (1 + e²) = 11.9% of its amplitude. Quote the convention with the number, because a "lag time" read by eye or from a 5% threshold will differ.

For mechanism rather than description, curves recorded at several peptide concentrations can be fitted globally with nucleation and growth models; a published protocol and web software describe how the concentration dependence constrains the mechanism and set out data quality checks first.^[4] The concentration dependence can also be unusual. For human GLP-1, a 31-residue incretin hormone and the native ligand of one of the receptors studied for Tirzepatide, ThT kinetics at pH 8.2 fitted a nucleation-polymerization mechanism, while at pH 7.5 the lag time increased with increasing peptide concentration, linked to off-pathway species and N-terminal protonation.^[5] Small pH changes can reshape the experiment; see buffers and pH selection.

Making plate assays reproducible

Fibril formation is a stochastic, nucleation-limited process, so replicate wells often disagree. A methodical study of a notoriously erratic aggregating protein found that salts and polyanionic polymers did not significantly improve reproducibility, that orbital agitation in the plate reader was a crucial first step, and that glass beads improved it further.^[6] The tightest replicates came from seeding with preformed fibrils or adding submicellar sodium dodecyl sulfate, giving 3-4% standard deviation in the final ThT level. Those conditions can change the aggregation pathway, so record them.

• Start state. Remove preformed aggregates (filtration or centrifugation); hidden seeds shorten the lag. Record stock preparation and age.
• Plate and seal. Black plates with clear or non-binding bottoms, sealed against evaporation; dilute peptides can be lost to plastic, covered in peptide adsorption to labware.
• Reader settings. Temperature, shaking mode and interval, read interval, and excitation/emission wavelengths near 440-450 and 480-485 nm.
• Replicates. At least three wells per condition, with the spread of lag times reported, not only the mean curve.
• Blanks. Buffer plus dye, and peptide without dye, so background and any intrinsic signal can be subtracted.

Interference and orthogonal checks

ThT is easy to fool. In a study using three polyphenols, the absorbance and fluorescence of quercetin and curcumin significantly biased ThT readings in both real-time and single time-point assays, with curcumin interfering at concentrations as low as 0.01 µM against 0.5 µM fibrillar amyloid-beta.^[7] Resveratrol, which is not spectroscopically active at ThT wavelengths, still biased readings at higher concentrations, implying competition for fibril binding. Test every added compound for bias and confirm fibrils independently, for example by Congo red spectral shift or electron microscopy.

Excipients and cosolvents matter too: in the insulin study the dye ANS lengthened the lag and appeared to promote amorphous aggregates, urea shortened the lag, and TMAO and sucrose lengthened it.^[3] Vehicle-matched controls are therefore essential when stocks contain an organic solvent; see DMSO stock solutions. Circular dichroism can show the shift toward beta-sheet secondary structure that accompanies fibril formation.

Amylin analogs and a record checklist

Amylin is a classic fibril-forming peptide hormone. The paper describing the design of Cagrilintide calls amylin’s high propensity toward amyloid fibrils a hallmark that makes analog design challenging, and describes Cagrilintide as a stable, lipidated amylin analogue.^[8] ThT kinetics are a natural screen for whether solution conditions trigger fibril formation in such peptides, alongside the solubility and aggregation basics.

• Peptide. Identity, stock solvent, measured concentration, filtration or centrifugation step and age of the stock.
• Dye. ThT concentration and how it was checked, whether it was present throughout or added at the end.
• Conditions. Buffer, pH, ionic strength, additives, temperature, plate type, seal, agitation and any beads or seeds.
• Analysis. Blank subtraction, the fitting equation and lag-time convention, number of replicates and their spread.
• Confirmation. Interference controls for every additive and at least one orthogonal method.

This guide describes general laboratory methods and published literature only. It is not a statement about the aggregation behaviour of any particular material. Terms are defined in the glossary; questions can be sent to support.