Quantifying free thiols
Free thiol content matters in three places: on an antibody before conjugation, where it sets the payload stoichiometry; on a designed binder, where an unpaired cysteine is a developability liability; and on a purified protein, where it reports how completely the disulfides formed. The methods differ in what they resolve, and the sample handling matters more than the method.
Handling comes first
Thiols oxidize in air, catalyzed by trace copper and iron. A sample that sat on the bench for an hour before assay reads low, and the error is one-directional, so it looks like good news.
Prepare in degassed buffer with EDTA at one to five millimolar. Keep the sample cold and assay promptly. Where a delay is unavoidable, alkylate immediately with iodoacetamide or N-ethylmaleimide to freeze the state, and quantify the adduct later.
Run a reduced control alongside every set. Fully reducing an aliquot and assaying it gives the total cysteine content, and the ratio of native to reduced tells you the fraction unpaired, which is more informative than an absolute number and far less sensitive to how the assay drifts.
Ellman assay
DTNB reacts with free thiol to release TNB, which absorbs at 412 nanometers. It takes fifteen minutes, needs a plate reader and costs almost nothing.
Its limits are worth knowing. Sensitivity bottoms out around ten micromolar thiol, so a dilute protein or a low occupancy needs concentration first. The reaction is pH dependent and wants around pH 8, since the thiolate is the reactive species. Buried thiols react slowly or not at all, so a native protein can read near zero while a denatured aliquot of the same material reads a full equivalent. That difference is information, and running both native and denatured is the cheapest way to learn whether an unpaired cysteine is exposed.
Any thiol-containing buffer component interferes completely. DTT, beta-mercaptoethanol and even low millimolar cysteine in a formulation will swamp the signal.
PEG-maleimide shift
Alkylate with a five or ten kilodalton PEG-maleimide and run reducing SDS-PAGE. Each modified thiol adds a discrete mass shift, and the gel shows the distribution of zero, one and two modifications per chain rather than an average.
This is the method when you want to know how many sites are occupied and on which chain, and it needs no specialized instrument. It is semi-quantitative by densitometry, which is usually enough, since the question is typically whether a population exists rather than its exact size.
It is also the practical readout for confirming a partial reduction landed where intended before committing payload to the batch.
Differential alkylation with mass spectrometry
The definitive method. Alkylate free thiols with light iodoacetamide, then reduce and alkylate the previously paired cysteines with a heavy isotopologue. Digest and run LC-MS.
Every cysteine in the sequence is then reported individually as a ratio of light to heavy label, so you learn which specific residues were unpaired and in what fraction of molecules. Nothing else gives site-level resolution.
It is a day of instrument time and a competent proteomics workflow, which is why it is reserved for characterization and for troubleshooting rather than routine control.
Reading the result
Expect a therapeutic-grade IgG to carry some free thiol, often a few percent of total cysteine, and sometimes considerably more. Treating the starting material as fully oxidized is the most common reason a conjugation overshoots its target DAR.
For a designed binder, an unpaired cysteine in a variable domain is a liability rather than a curiosity. It drives disulfide-linked dimerization on storage, it scrambles with the native bonds, and it changes behavior between lots. Where the design does not need it, removing it is usually the right call, and where it appears unexpectedly in a panel, it is worth checking before the panel is ranked on anything else.