Chromatography Forum

Hi,

I have a problem with SEC of a protein product that also contains polysorbate. The polysorbate give a peak (micells?) exactly where the protein aggregates elute...

I was thinking about using a post-column derivatisation step and fluorescence detection? To test the idea I am thinking I could to use a t-union and then a small volume active mixer (Knauer) before the detector.

But what reagent should I use? Something that reacts fast and give good fluorescence response I guess. Or is the idea completely insane?

This would be run on a Waters Alliance with 474 detector.

Would there be any issues with a pre-column derivitization? Post-column can get tricky. Also, could you use a UV-Vis family detector instead of the RID? Your protein may have a chromaphore that absorbs in a different are than the polysorbate.

I have considered to use pre column derivatisation, but I have not tried it. I am a little bit scared of doing this since it seems to be performed at high pH and the protein will possibly dissociate (it contains two subunits). The MW of the protein will also change if a attach a derivatisation agent. So I change my sample too much I think. Doing the derivatisation after the column will not change the separation itself (and will be easier to sell to QA...)

The problem would be solved if I could use 270 nm (where I have no interferences). But the sample concentration is very low, just 2-3 µg/ml and I do not have enough sensitivity at 270 nm.

Here is what I will test: (when the FL-detector comes back from the dead)

1. After the column I will attach a 50 µm mobile phase mixer (passive, Waters)
2. The mobile phase from the column will be mixed with a solution of 1 mM naphthalene-2,3-dicarboxaldehyde (NDA), 8 mM of DL-Dithiothreitol in 0.1M of sodium borate buffer pH 9.5 (not sure of the ratio between mobile phase and derivatisation solution)
3. The detection will be using excitation at 420 nm and emission at 550 nm

Just a follow up:

The procedure works fine, I get very similar chromatogram in UV and in fluorescence (FL) (surprisingly little band-broadening). Placebo components are not visible in FL, only proteins.

The only thing that puzzles me is the amount of noise in FL (which should be more or less noiseless). I used 420 nm as excitation and 550 nm as emission wavelength. It is possible that the reagent is slightly fluorescent in the unreacted form, giving a signal? Or maybe the SEC column is leaking out proteins all the time (it is a very old column)?

I haven't done a lot with fluorescence. I would try running a sample with no derivitization to see what the fluorescing background is.

This is almost embarrassing...

The mobile phase contains tris-buffer. No wonder that I get some noise, since I am derivatising the mobile phase as well (tris contains a primary amino group).

I am amazed that I still could get peaks for the protein!!