Protein aggregation
Aggregation is what happens when a protein finds an intermolecular arrangement more stable than the one it is supposed to hold on its own. That covers several different outcomes, and calling all of them "aggregation" is the source of most confusion in the field.
The states worth separating
Reversible self-association is concentration-dependent clustering that comes apart on dilution. It matters enormously for formulation and viscosity, and it is invisible to assays that only look for insoluble material.
Amorphous aggregation is disordered precipitation, often driven by partial unfolding at an interface, in heat, or at a pH near the isoelectric point. It usually shows up as turbidity and as material lost on a filter.
Amyloid fibrils are ordered, beta-sheet-rich polymers with a repeating structure. They form through nucleation and growth, they are self-propagating, and they are the state behind ThT kinetics and the seeding assays used in neurodegeneration.
Oligomers sit in between: soluble, often transient, structurally heterogeneous, and frequently the species people care most about biologically. They are also the hardest to measure honestly, because most assays either miss them or destroy them.
Why the kinetics look the way they do
Amyloid formation is a nucleated polymerization. Making the first nucleus is slow and improbable, and once fibrils exist they grow quickly at their ends. Fibrils also generate new nuclei, either by fragmenting or by catalyzing nucleation on their surfaces, which makes the whole process autocatalytic.
Three practical consequences follow. First, the reaction has a lag phase that shortens dramatically on seeding. Second, small differences in starting conditions produce large differences in timing, so replicates scatter more than people expect. Third, an intervention can act at several distinct points, and the shape of the kinetic curve is what tells you where.
Choosing the measurement
- ThT fluorescence reports on cross-beta structure over time. It is the workhorse for kinetics, and it says nothing about morphology or biological activity.
- Turbidity or light scattering reports on large particles, including amorphous material that ThT ignores.
- Sedimentation or filtration with a protein assay separates soluble from insoluble and gives a mass balance, which is the defensible way to check that something labeled an inhibitor did not simply keep material in a different insoluble form.
- Size exclusion chromatography resolves monomer, dimer and larger soluble species, and it is the standard readout for developability work. Run it with care, because dilution on the column can pull weak oligomers apart.
- Electron microscopy shows you morphology. It is the fastest way to find out that the reaction produced something other than fibrils.
- Cell-based seeding assays ask whether the aggregate is biologically active, which none of the above do.
What changes aggregation, in practice
Concentration, temperature, pH, ionic strength, agitation and interfaces all move the reaction, and interfaces matter more than most protocols admit. Air-liquid and plastic-liquid surfaces nucleate aggregation, which is why plate material, shaking, seals and even the number of freeze-thaw cycles change the result.
Sequence features matter too. Hydrophobic and beta-prone segments drive fibril formation, charge distribution sets the barrier for association, and in antibodies the same properties that create an aggregation liability often live in the CDRs, where they cannot be edited freely.
What this means in practice
Aggregation is not a single number. A candidate that survives a thermal ramp can still fail on interface stress, and a molecule that passes SEC on day one can grow a shoulder after a month at four degrees. Measuring the kind of aggregation that threatens your program, under the stress that program will actually see, beats collecting a general purpose score.