In many cases, even if we have a relatively purified protein, we would want to add an additional polishing step in the form of a size exclusion chromatography (SEC) or gel filtration. In this post I will discuss some considerations and tips in regard to maximizing the use of SEC.
SEC: the good and the bad
Unlike the other chromatography techniques I have mentioned before, SEC power comes from its simple non-interacting separation according to shape/size. This aspect makes SEC kill two birds with one stone: size differentiation and buffer exchange (mostly much faster than an overnight dialysis).
So you might ask: where’s the bad part?
Well, because SEC differentiates according to size AND shape, two proteins which differentiate with respect to their molecular might elute together or very close one to another if their shape counteracts the size difference, making the separation ineffective. In most cases, however, a size difference of approximately three times usually suffices to get a reasonable separation.
Choosing the right SEC column for your needs
When approaching SEC methodology, one should evaluate which column arsenal is available at hand. Since SEC is a non-interacting separation methodology, the column’s physical properties has as much as impact on the separation quality as the matrix quality that is packed within. The problem, in such a case, arises from the fact that while separating protein B requires a certain column, protein B’s separation requires a different SEC column. Needless to say, SEC columns are terribly expensive so, the questions at hand are a) which minimal column set should I have and b) what SEC column criteria are the most critical when choosing a column?
SEC columns can be divided into preparative and analytical columns in respect to their load capacity and to their resolving power (which take form in their physical size). Preparative columns are larger in size\volume (length and width) so they can accommodate larger protein load (tens of mg/run) on the account of a relatively larger particles size (~30 micron) and lower length/width ratio.
Analytical columns are usually smaller in size/volume and can accommodate small amounts of proteins (up to 10 mg/run) without risking overload. Usually analytical column have a relative smaller particle size (~10-15 micron) which affects both the flow rate and the back pressure. The particles size has a key effect on the resolving power of the column and has a direct effect on the developing back pressure and is inversely proportional to the speed of separation. Thus, preparative column can be run at higher flow rate but achieve separation qualities which are usually inferior to the analytical columns. Notice that you should verify your FPLC’s pump can handle the specified working pressure. Even though the purpose of analytical column are for characterization rather for purification, they can be used for the latter purpose as their resolving power is bigger even though they require much more injections. Recently I have purified ~50mg of protein by performing no less than 13 injections on an analytical column – it took me a whole day but I got the best resolution possible with the columns I had in possession.
On top of the above considerations you should also evaluate what size range of protein you’re involved with. The pore size within the beads have a key effect on the separation of differently sized protein or complexes and should be chosen wisely. A good example for a relatively wide spectrum of columns capabilities can be seen in the EMBL Protein Expression and Purification Core Facility:
• Superose 12 HR 10/30 – this analytical column (designated with the “HR” mark for “high resolution”) has the capacity to resolve small to large proteins and complexes in the range of 1kDa-300kDa. Unlike its sister, Superose 6 HR 10/30, this column has more subtle size separation range but is equipped with smaller beads diameter (10 micron in average) making it’s separation superior to the Superose 6.
• Superdex 75 HR 10/30 – Another analytical column which has the capability to separate small to medium size proteins in the range of 3kDa-70kDa, complementing the Superose 12 for better separation in the medium range.
• HiLoad 16/60 & 26/60 Superdex 75 prep grade columns– The mark “prep grade” denotes the preparative nature of these sister columns. Both have similar separation range but differ in the packaging column cylinder, the 16/60 is thinner thus has ~ half the load capacity but better resolving power.
From the above paragraph you can clearly evaluate your technical demands: are you purifying proteins in masses (>20mg/purification) or you’re using SEC for the analytical part or purifying small amounts? Don’t forget that SEC columns can be borrowed or shared among severa laboratories (as long as these are maintained according to the manufacturers recommendations).
Settings that improve purification by Size exclusion chromatography
Having the best column for your needs doesn’t necessarily mean that your purification step will be ideal with ~100% resolution (the measure of separating one specie from another). In many cases some contaminates will be evident, some even in a cryptic shoulder that will be very difficult to observe. Here are the main settings you should give attention before running a separation:
• Load – avoid overloading the column with your precious protein unless you know there will be no close species to your protein’s elution volume. Seek advise with the column’s data sheet in regard to acceptable injection volume and protein mass. Measure the sample’s concentration and if the sample is not pure add 10-20% mass to your final protein concentration (according to the amount of contamination. As a rule of thumb your protein sample volume and mass for injection should be in the range of 0.5-3% of the column volume and no more than one tenth of column volume in terms of mass. This means that if you’re using a 24ml column, use 120-500 microliters of protein sample with a maximum protein mass of 2.4mg. Again, these estimates are for your first run and if you find you can inject more protein mass, then go ahead. However, if you find a cryptic shoulder (a sign of a closely eluting specie) then you might want to lower the amount of protein you inject. It will be easier for you to cut out a small peak of several injections than trail than a larger trail in a single overloaded injection.
• Flow rate – chromatography is based on the dynamic transition between equilibrium and disequilibrium between the mobile and the stationary phases (“theoretical plates”). For this equilibrium to occur, sufficient time should be supplied. With respect to our modern FPLC systems, this means a reasonable flow rate, which depends on the specific column (usually no lower than 0.2ml/min or higher than 4ml/min).
• Mid-run injection – In cases you have to do many injections you might want to explore mid-run injection. As the name implies, after performing the first separation analyze whether you can inject an additional sample before or after the expected elution volume. For example, if you have an elution profile of peak A eluting at 7ml with peak width of 1ml, peak B starting eluting at 12ml with a peak width of 1ml and peak C eluting at 20ml with peak width of 0.5ml, you can safely inject two injections 2ml apart. This technique can save a lot of time when the chromatogram is not crowded with many peaks.
• Consecutives SEC columns – In case you have a combination of complexes (80kDa) and monomers (40kDa) you might want to explore the use of two complementary columns, such as Superdex 75 and Superdex 200. Assuming you’re aiming at isolating the complex from the monomers and the other small contaminants, the first separation should be performed with Superdex 75, collecting the complexes eluting at void volume and then reinject the elution into Supderdex 200 to separate them from the other contaminates. This technique is especially effective when it is hard to separate complexes from monomers in a single column.
And one last tip – sometimes references or protocols use linear flow units (cm/hr) which is not compatible with the modern FPLC volumetric flow units used today (ml/min). Following is a formula for the conversion from cm/hr to ml/min (VF):
VF=0.013*X*d^2 where X is the linear flow value and d is the internal diameter of the column in centimeters.
Have any other tip or suggestion? Let me know!