Fast protein liquid chromatography

Size exclusion, ion exchange, affinity, hydrophobic interaction and desalting, on ÄKTA Pure and ÄKTA Start systems.

For gel filtration the workhorses are Superdex 75 and Superdex 200 Increase 10/300 GL (24 ml bed, analytical to semi-preparative) and HiLoad 16/600 columns (120 ml bed, preparative). Ion exchange runs on HiTrap Q and SP in 1 ml and 5 ml formats, affinity on HiTrap IMAC, GST and Strep columns, and buffer exchange on HiTrap Desalting.

The choice depends on what you are trying to achieve: a polishing step and an aggregate check are a different job from a three-step purification from lysate. Describing the goal in the order form matters more than naming a column, since the column follows from the goal.

The limit on a SEC column is volume before it is mass. On a Superdex 200 Increase 10/300 GL with its 24 ml bed, 250-500 µl gives good resolution, and up to about 1.2 ml (5% of the column volume) is the absolute ceiling, with visibly broader peaks. A HiLoad 16/600 takes 1-5 ml on its 120 ml bed.

Mass is limited by viscosity rather than by capacity: above roughly 10-20 mg/ml of total protein the sample band becomes dense enough to finger through the bed and distort the peaks. Twenty milligrams in 500 µl is fine for a well-behaved protein; the same mass in 100 µl is not.

Ion exchange and affinity are the opposite case — there the mass matters. HiTrap columns bind on the order of tens of milligrams of protein per millilitre of resin, and the sample volume is almost unlimited because the protein concentrates on the column.

Centrifuge, then filter at 0.22 µm. A particle that would pass unnoticed in a microcentrifuge tube blocks the frit at the top of the column, and unblocking a Superdex costs more than the run. If the protein does not survive filtration, 15 minutes at 20 000 g and careful pipetting off the pellet is the fallback.

The sample should be in a buffer compatible with the first step. For SEC that means it can be almost anything, provided it does not interact with the matrix — very low salt promotes ionic interaction with Superdex, so 100-150 mM NaCl is the usual baseline. For ion exchange the conductivity of the sample has to be below the point at which the protein binds, which often means dilution or desalting first. Buffers we prepare are filtered and degassed; if you bring your own, filter it at 0.22 µm.

What we cannot take: unfixed pathogenic material or anything above biosafety level 1 without a prior arrangement, samples containing organic solvents at high percentage, and anything that would leave the column contaminated for the next user. Say what the sample is when you order, not when it arrives.

An apparent one, from a calibration with standard proteins, and it should be read with caution.

SEC separates by hydrodynamic radius, not by mass. A globular protein behaves like the globular standards and the estimate is reasonable. An elongated, disordered or heavily glycosylated protein migrates like something much larger — an intrinsically disordered protein of 20 kDa can elute where a 60 kDa globular standard elutes, and this is a routine source of wrong oligomeric-state claims in the literature.

Weak interaction with the matrix pushes in the other direction, delaying elution and making the protein look smaller.

If the mass is the actual question, SEC-MALS gives it independently of shape, and AUC gives mass together with the frictional ratio. The SEC elution volume is best used as a comparison between related samples run on the same column, not as an absolute measurement.

Sometimes, within limits that are worth knowing before you pay for another run.

On a single SEC column, two species need roughly a factor of two in mass — more precisely, a clear difference in hydrodynamic radius — before they resolve. A monomer and a dimer separate; a protein and the same protein missing twenty residues do not. Moving from a 24 ml Increase column to a 120 ml HiLoad, halving the flow rate and reducing the injection volume all help, and together they can turn a shoulder into two peaks. None of them will separate species that differ by 10% in size.

Where SEC runs out, the answer is usually a different mode rather than a better SEC run. Ion exchange separates on charge and will resolve a deamidated or truncated form that SEC cannot see; hydrophobic interaction picks up conformational differences. Running two orthogonal modes in sequence resolves most things worth resolving.

Yes, and it is one of the common orders: IMAC capture, then either desalting or SEC as a polishing step, with SDS-PAGE on the fractions.

Send the cleared lysate, not the cell pellet, unless lysis has been agreed as part of the work. The lysate must be centrifuged hard and filtered, since crude extract loaded unfiltered onto an IMAC column is the fastest way to ruin it. Include 10-20 mM imidazole in the loading buffer to suppress background binding, and tell us whether the construct carries a protease site and whether you want the tag removed.

Two things to check on your side first: that the tag is accessible in the folded protein, and that the lysate contains no chelators. EDTA in a lysis buffer strips the nickel off the column, and it happens more often than you would think.

ÄKTA Start is a compact system with a fixed-wavelength 280 nm UV monitor, a conductivity monitor, a sample pump and a fraction collector. Flow up to about 5 ml/min and a low pressure limit, around 0.5 MPa. It handles affinity capture, desalting and simple step gradients on HiTrap columns very well, and its limitation is that it will not drive a high-backpressure analytical column.

ÄKTA Pure runs from microlitres to tens of millilitres per minute at pressures up to 20 MPa, reads several UV wavelengths at once, supports fine gradients, column switching, loop and sample-pump injection, pH monitoring and temperature control. Anything involving a Superdex Increase column, a shallow ion-exchange gradient, a multi-step method or A260/A280 monitoring goes on it.

In practice: routine capture and desalting on Start, everything analytical or multi-step on Pure. Which one your sample runs on is decided by the method, and it does not change what you receive.

A Superdex 200 Increase 10/300 at 0.5-0.75 ml/min is 40-60 minutes for one column volume and a bit. A HiLoad 16/600 at 1 ml/min is closer to two and a half hours. An affinity capture with a step elution on a 5 ml HiTrap is 20-30 minutes, and a desalting run is under ten.

Equilibration is the part people forget: a SEC column needs one to two column volumes of buffer before the first injection, so the first run of the day starts an hour before the first sample.

A realistic day is six to eight analytical SEC runs, or two to three preparative HiLoad runs, or a full capture-plus-polish sequence on one sample. Turnaround on an order depends on the queue and on whether a method has to be developed first.

In fractions, in the elution buffer, and more dilute than it went in. SEC typically dilutes by a factor of two to four; an affinity elution concentrates instead, but hands the protein back in whatever the elution buffer was — imidazole, glutathione, low pH.

You get the chromatogram as a data file with the UV, conductivity and fraction marks, the fraction numbering, and an SDS-PAGE of the relevant fractions if that was part of the order. Pooling is done according to what you asked for: purity at the expense of yield, or the other way round. Say which you prefer, because the two choices give visibly different results.

Concentration by centrifugal ultrafiltration and buffer exchange can be added, though both cost material — concentration is where fragile proteins aggregate, and it is worth deciding whether you actually need 10 mg/ml or whether 2 mg/ml would do.

For people from partner laboratories who purify regularly, yes. Training covers the system, the software, column care and the pressure limits that protect the columns from an expensive mistake.

Trained users book their own time and run their own methods, with facility columns or with a column of their own that stays dedicated to their sample. A dedicated column is the sensible option for anything that might leave residue: nucleic acids, sticky membrane proteins, samples with a history of aggregation.

Sending samples as a service is the alternative and needs no training. Purifications from lysate, method development and anything requiring a preparative column are usually run this way in any case.