Cutting an antibody down to its binding end removes two things that cause trouble: the part that immune receptors bind, and half the size. That is why fragments appear in the awkward experiments, the ones with high background on tissue, a secondary that will not behave, or a target too crowded for a whole molecule to reach. The cost is avidity, half-life and the loss of anything that depended on the constant region.
What a Fab fragment antibody is, and what removing the tail changes
A single binding arm with its light chain and the first constant domain, and nothing that receptors on cells recognise. On tissue that removes a whole class of background at once, because the reagent can no longer be captured by the receptors macrophages and other cells carry. It also becomes monovalent, so the apparent affinity drops relative to the two-armed parent, and it clears faster in anything living. Use it where the background or the size is the problem, not as a general improvement.
Where an antibody fragment format is the right choice
Three situations recur. Dense or partially hidden epitopes, where a smaller reagent reaches what a whole molecule cannot. Multiplexing with two reagents of the same species, where a monovalent fragment blocks the second primary from being seen by the first secondary. And anything where signal from receptor binding has to be excluded before a result can be believed. Outside those, the whole molecule is usually the better reagent because it is more avid and more stable.
Making a Fab fragment of antibody stocks by digestion, or buying it recombinant
Digestion with a thiol protease cuts above the disulphides and gives one binding arm per cut; a different protease cuts below them and leaves the two arms joined, which is bivalent and behaves much more like the parent. Digestion is cheap, works on a stock you already have, and gives a heterogeneous product that needs purifying. A recombinant fragment expressed from a sequence is homogeneous, reproducible between lots and the right answer when the reagent is going into a process rather than an experiment.
An anti human Fab antibody, and using a fragment as the detection reagent
The commonest use of these reagents is not as a primary at all. A fragment raised against another species' binding end blocks that species' immunoglobulin, which is how a mouse primary is used on mouse tissue without the secondary lighting up the whole section. The same logic makes fragment secondaries the standard answer for double labelling. Check what the reagent was raised against and what it was adsorbed against, because those two lines on the datasheet decide whether it will work.
Fab purification, and what to check on the product
After digestion the mixture holds the fragment, the cut constant region and undigested parent, and the fragment is the part you want. Size separation resolves it and an affinity step against the constant region removes the rest. Confirm on a reducing and a non-reducing gel that the size is what it should be and that no parent remains, and record the degree of digestion, because an incompletely cut preparation behaves like a mixture of two reagents and is the commonest reason a fragment appears not to help.
Questions people ask about fab antibody
Is a fragment always better on tissue?
No. It removes receptor-mediated background and reaches crowded epitopes, and it loses avidity. If the whole molecule gives clean staining, keep it.
Which digestion should I use?
A thiol protease above the disulphides gives one binding arm per cut and a monovalent product; a different protease below them leaves the arms joined and stays bivalent. Choose by whether you need monovalency or only the loss of the tail.
Digested or recombinant?
Digested for an experiment on a stock you hold, because it is cheap and immediate. Recombinant where lot-to-lot reproducibility matters or the reagent is going into a process, because a digest is heterogeneous by construction.