An AAV purification protocol has to do two things that antibody processes do not: separate a 25 nm particle from everything else in a lysate, and then separate that particle from a nearly identical one that happens to be empty. The first is a capture problem, the second is the reason the route has a step most other processes do not need. This page walks the whole route and names the decision at each step.
The route at a glance
| Step | What it does | The decision to make |
|---|---|---|
| Harvest | Collects vector from the cells, the supernatant, or both | Whether the serotype and construct release into the medium or need lysis |
| Lysis and nuclease treatment | Frees intracellular vector and digests host DNA and plasmid residue | Detergent and nuclease choice; nuclease clearance has to be demonstrated later |
| Clarification | Removes cells, debris and precipitate before any column | Depth filtration train and final filter size — this is where load-ability is decided |
| Affinity capture | Binds the capsid specifically; removes most host-cell protein and DNA in one step | Serotype coverage of the ligand; elution is usually acidic |
| Full/empty separation | Separates genome-containing from empty capsids on an anion exchanger | Gradient shape and salt system — the hardest separation in the process |
| Polishing and concentration | Final impurity removal, buffer exchange, concentration | Tangential flow filtration, and whether a further chromatography step is needed |
Harvest: lysate or supernatant?
This is the first branch and it is set by the serotype and the construct rather than by preference. Some serotypes release a substantial fraction of vector into the medium; others stay largely cell-associated. Processes that harvest both streams exist and pay for it with a more complex clarification.
The practical consequence is downstream: a lysate carries far more host-cell DNA, protein and membrane fragments than a supernatant, and every one of those has to come out before the capsid is pure. Nuclease treatment — typically an endonuclease under magnesium — is what makes the DNA burden manageable, and the nuclease itself then becomes an impurity to clear.
Clarification: where the column step is won or lost
Clarification rarely gets the attention it deserves, because it is filtration rather than chromatography. But it decides whether the capture column runs or blocks. A lysate after nuclease treatment still carries debris across a wide size range, and the usual answer is a depth filtration train stepping down to a final filter before the column.
Two things are worth deciding deliberately here. How far down the filter train has to go — every stage costs vector yield through adsorption. And whether the clarified material is stable enough to hold before loading, because AAV preparations are not indefinitely stable at room temperature.
Affinity capture: the step that does the heavy lifting
Affinity capture for AAV runs on camelid VHH-derived ligands that bind the assembled capsid. A single step removes the large majority of host-cell protein, residual DNA and medium components, in the same way Protein A does for antibodies — and for the same reason: the ligand binds something only the product has.
Two constraints shape the choice. Serotype coverage is not universal; a ligand that binds one serotype well may bind another poorly, and that has to be checked rather than assumed. And elution is usually acidic, which for a capsid is not trivial — AAV capsids show conformational changes at low pH, and the eluate is normally neutralised immediately.
We do not make an AAV affinity ligand. MonoCore™ covers the anion exchange step that follows, and a blank matrix for immobilising a ligand of your own is available as custom affinity.
Full/empty separation: the step that makes AAV different
A capsid that contains a genome and one that does not are nearly identical in size and surface chemistry. What differs is the charge contributed by the packaged DNA, and that difference — small as it is — is what an anion exchange gradient separates. It is the hardest chromatographic step in the process, and the one where method development spends most of its time.
What helps: a shallow gradient, a well-controlled salt system, tight pH control, and a format whose flow distribution does not blur the separation further. What does not help: a capsid that has been stressed by the steps before it, because aggregate and damaged particles smear the peak.
On full/empty performance we have no data of our own — so we do not claim any. What we can say is the geometry: a 25 nm capsid is large relative to the pores of a porous bead, and how that size relation plays out in practice is the subject of which chromatography media are suitable for AAV polishing.
Polishing, concentration and formulation
What remains after the anion exchange step is usually residual host-cell protein, nuclease, and aggregate. A further chromatography step is sometimes used; more often the remaining work is done by tangential flow filtration, which concentrates the vector and exchanges it into the formulation buffer in one operation.
Two practical notes. AAV adsorbs to surfaces, so recovery losses in filtration and hold steps are real and worth measuring rather than assuming. And the final formulation usually carries a surfactant for exactly that reason.
Where the route differs by serotype
Serotype affects almost every step: whether the vector is released into the medium, how well an affinity ligand binds it, where the full and empty peaks sit in a salt gradient, and how much the capsid tolerates the acid of an affinity elution. A protocol developed for one serotype is a starting point for another, not a transfer.
The same applies to the production system. Transfection-derived material, baculovirus systems and producer cell lines differ in their impurity profile, and the clarification and polishing strategy follows that profile rather than the vector.
Frequently asked questions
Do I need to filter the feed before loading?
Yes, and it is the step that decides whether the column runs at all. A nuclease-treated lysate carries debris across a wide size range, which is why clarification uses a depth filtration train rather than a single filter. How far down the train has to go is a real decision: it has to protect the column, and every stage costs vector through adsorption. The mechanisms behind fouling are described in why chromatography media foul.
What is the standard AAV purification protocol?
Harvest — from lysate, supernatant, or both — followed by nuclease treatment, clarification by depth filtration, affinity capture on a capsid-binding ligand, separation of full from empty capsids on an anion exchanger, and finally polishing and concentration by tangential flow filtration into the formulation buffer. Every step is serotype-dependent to some degree.
How do you separate full and empty AAV capsids?
On an anion exchanger, with a shallow salt or pH gradient. Full and empty capsids differ only in the charge contributed by the packaged genome, so the separation depends on a small difference and is correspondingly sensitive to gradient shape, pH control and the condition of the capsids arriving from the steps before. It is the most demanding chromatographic step in the process.
Do you need affinity capture for AAV?
Not strictly, but without it the burden on the following steps is much larger. Affinity capture removes the majority of host-cell protein and DNA in one operation. Processes built on ion exchange capture exist, particularly at small scale or where no ligand covers the serotype, and they need more polishing to reach the same purity.
Why is clarification so important before an AAV column?
Because a nuclease-treated lysate carries debris across a wide size range, and that debris is what blocks a column. The filter train has to go far enough to protect the column and no further, since every filtration stage costs vector through adsorption. It is the step where the load-ability of the whole process is decided.
Does AAV survive the acid elution of an affinity step?
Usually, but it is not free. AAV capsids show conformational changes at low pH, so the eluate is normally neutralised immediately and the hold time kept short. Where a capsid is particularly sensitive, the alternatives to acid elution are the same ones used for antibodies — set out in how to elute from Protein A without acid, which covers the mechanisms rather than the molecule.
What limits throughput in an AAV process?
Usually clarification and the anion exchange step, for different reasons. Clarification is limited by filter area and the debris load. The anion exchange step is limited by how shallow the gradient has to be to resolve full from empty — and a gradient that has to be shallow cannot simply be run faster.
Still an open question?
Tell us the molecule, the feed volume and the system you run it on, and we will say whether a MonoCore™ capsule is the right starting point — or whether it is not.