How do you purify an antibody from hybridoma supernatant?

The ligand decides this, not the protocol. Hybridoma culture is not automatically mouse — rat and hamster lines are common, and veterinary work brings its own species — but most hybridoma monoclonals are mouse IgG1, and IgG1 binds Protein A weakly. That is why Protein G is the standard capture ligand for mouse hybridoma work, and why a Protein A column that works beautifully for a human IgG1 can return almost nothing from a mouse line. Veterinary antibodies invert the rule again: for dog and cat IgG the vendor data put Protein A ahead of Protein G. The rest is a normal affinity cycle: equilibrate, load the clarified supernatant, wash, elute at low pH, neutralise immediately. MonoCore™ Protein A and Protein G run that cycle as drop-in capsules for standard FPLC systems, with recombinant ligands on a modified cellulose monolith, 15 µm channels, in 1.8 mL and 5.4 mL.

Which ligand binds which antibody

The pattern behind the table is well documented in the vendor literature — see the binding comparison for immunoglobulin-binding proteins; for the canine subclasses, see Bergeron et al. on comparative canine IgG characterisation. Three practical consequences follow. First, if you do not know the subclass of your line, determine it before you buy a column: an isotyping strip or ELISA on 50 µL of supernatant takes about an hour and decides the whole step. Second, “IgG binds Protein A” is a statement about human IgG1 that has quietly become a rule of thumb; in hybridoma and veterinary work it is wrong often enough to cost a batch. Third, where neither Fc ligand is convincing — rat IgG2b is the clearest case — a kappa-specific Protein L is the honest next option, even though we do not sell it; the Protein A/G chimera does not rescue that case, because it inherits both weak bindings.

The serum problem

Hybridoma supernatant usually carries serum, and serum carries its own IgG. Protein G does not distinguish between the antibody you made and the bovine IgG that came with the medium — and the asymmetry runs the wrong way: bovine IgG binds Protein G strongly, mouse IgG1 only moderately. At 10 % serum the feed carries roughly 5–30 mg of bovine IgG per litre, the same order of magnitude as a typical hybridoma titre. Plan the load on total IgG, not on your monoclonal alone.

• Use low-IgG or IgG-depleted serum from the start, or adapt the line to a serum-free medium. This is the only clean solution.
• If the antibody carries a kappa light chain — about 95 % of mouse monoclonals do — Protein L is the orthogonal option that actually solves this: it binds kappa and does not bind bovine immunoglobulins at all. Mouse binding is restricted to certain kappa families, so verify it on a small format before committing.
• Ion exchange can work where the pI difference is real, but bovine IgG is polyclonal with a broad pI range that overlaps most monoclonals — do not assume a clean split. And note that native Protein G also binds serum albumin, which is why recombinant ligands are the norm for antibody work; MonoCore™ Protein G carries a recombinant Protein G ligand. If albumin carry-over is critical for your feed, ask us for the ligand details before you plan the step.

Capture protocol for a hybridoma supernatant

The steps below explain the decisions inside that sequence.

1. Clarify. Centrifuge or depth-filter the supernatant. Whether you also filter through 0.22 µm immediately before the column depends on the medium: it matters most for a packed bed, whose interstitial gaps are the narrowest part of the flow path, and least for the 15 µm channels of a capsule. Filter in any case if the material will sit before it is loaded.
2. Equilibrate with 5–10 column volumes of TBS, or 20 mM sodium phosphate with 150 mM NaCl at pH 7.4.
3. Load the clarified supernatant at pH 7.0–7.4. Spent hybridoma supernatant is usually acidic from lactate — the phenol red turning yellow is exactly that — so check the pH and adjust with a tenth of a volume of 1 M Tris-HCl at pH 8.0, or dilute 1:1 with a 2× binding buffer. Protein G binds the Fc rather than charge, so conductivity does not normally need adjusting. Collect the flow-through: it is the only way to tell a ligand that did not bind from an elution that failed.
4. Wash with 5–10 column volumes of equilibration buffer, until UV returns to baseline.
5. Elute with 0.1 M glycine-HCl at pH 2.5–3.0 over 3–5 column volumes. Protein G binds the Fc with higher affinity than Protein A and over a more extensive contact area, so it needs the more acidic end: vendor protocols centre on about pH 2.7 for Protein G against pH 3.0 for Protein A. That half a pH unit is also why the neutralisation has to be immediate rather than merely prompt.
6. Neutralise in the collection tube: 1 M Tris-HCl at pH 8.5–9.0 — about 30–50 µL per mL of fraction for a 0.1 M glycine pool, 120–250 µL for 50–100 mM citrate or acetate, aiming for a pooled pH of 6.5–7.5.
7. Strip below the elution pH — 0.1 M glycine-HCl at pH 2.0 is the usual choice, because a strip at the elution pH removes nothing the elution did not already take. Then clean per the medium's specification, re-equilibrate with 10 column volumes, and store in the bacteriostatic solution the datasheet specifies: an affinity column parked in plain buffer grows.

The buffer choices and the aggregation trade-offs behind step 5 and 6 are set out in Protein A elution buffer: glycine, citrate or acetate — the chemistry of the elution is the same for Protein G, only the working pH tends to sit lower.

Scale: how many cycles?

Hybridoma work is small-volume work, and the arithmetic often favours a single cycle — but whether it does is a calculation, not a given. Titres vary by two orders of magnitude: a static flask with serum typically runs at 1–20 mg/L, an adapted serum-free or fed-batch culture at 30–200 mg/L, so a litre of supernatant carries anywhere from a few to a few hundred milligrams — plus the bovine IgG from the medium, which loads on the same ligand and competes for it. Take the conditions from the MonoCore™ Protein G datasheet, measure it on your own supernatant where those conditions do not match (how to read a capacity figure), and work out the number of cycles from those two numbers rather than from the format.

The high flow rate belongs to equilibration, wash and elution. For the load step, establish the residence time your antibody needs on your own feed — affinity binding is kinetically limited whatever the format, and breakthrough at very short contact times is a load-flow problem, not a capacity problem. Where the volume is larger, the 5.4 mL format runs the same method at 45 mL/min instead of 15 mL/min — about 8 column volumes per minute either way, so the residence time and therefore the method stay unchanged.

When the yield is disappointing

This page assumes culture supernatant. Ascites is a different feed: it needs delipidation before loading, and its host IgG is murine — no polishing step separates that from your monoclonal, which is the main argument for working from culture in the first place. Whatever the feed, check the subclass first: a mouse IgG1 on a Protein A column is the commonest cause of a hybridoma purification that returns almost nothing, and no amount of protocol tuning fixes a ligand that does not bind.

If the ligand is right and the yield is still low, work through the checks in Protein A column troubleshooting — they apply unchanged to Protein G: find out whether the product is in the flow-through or still on the column before you change anything.

Frequently asked questions

Does mouse IgG1 bind Protein A?

Weakly. Mouse IgG1 is the subclass most hybridoma lines produce, and it is the classic case for Protein G. IgG2a, IgG2b and IgG3 bind Protein A well in the vendor tables, with some older sources rating IgG3 lower — so the answer depends on the subclass of your line, and it is worth determining before choosing a column.

Can I purify canine or feline antibodies on Protein A?

For dog and cat IgG, Protein A is the better starting point — the vendor data rank it above Protein G for both species, which reverses the rule that holds for mouse hybridomas, and the same inversion applies to pig and guinea pig. One caveat that matters for monoclonal work: that rating describes serum IgG, which in dogs is dominated by the IgG-B subclass. Published work with Protein G resin has captured all four canine subclasses, so for a line that is not IgG-B, screen both rather than trusting either rating.

How do I deal with bovine IgG from the serum?

Best: switch to low-IgG or IgG-depleted serum, or a serum-free medium, before the culture runs. Bovine IgG binds Protein G more strongly than mouse IgG1 does, so it competes for the ligand rather than merely riding along. If the culture is already running, Protein L is the orthogonal route for kappa antibodies, because it does not bind bovine immunoglobulins at all.

What elution pH does Protein G need?

Usually pH 2.5–3.0, with vendor protocols centring on about 2.7 — lower than a typical Protein A elution, because Protein G binds Fc more tightly. Neutralise immediately in the collection tube and verify that the pooled pH lands between 6.5 and 7.5.

Can Protein G purify IgM or IgA?

Not as a capture route. Neither ligand binds mouse IgM or IgA meaningfully; for human IgM and IgA, Protein A is weak rather than absent, which is not enough to build a step on. The usual route is Protein L, which binds the kappa light chain and is indifferent to the heavy-chain class; failing that, thiophilic adsorption, ion exchange or size exclusion.

Is a monolith capsule worth it at hybridoma scale?

It is worth it when time per cycle is your constraint rather than capacity. A MonoCore™ Protein G capsule runs at about 8 column volumes per minute, which puts the residence time in seconds, and it fits a standard FPLC system without re-plumbing. If your current column runs once a week and nobody waits for it, the format changes little.