I’m using Biacore T200 using as a ligand a monoclonal antibody “A” anti-VP3 on a CM5 sensor chip (via anti-mouse antibodies immobilized with amine coupling) and as analyte whole viruses (900 kDa) in a complex mixture to quantify them. The samples have cell debris, BSA, and impurities and lots of non-specific binding in my reference flow cell.
Thanks to SPR pages I managed to lower the non-specific binding to 5 RU by adding +200 mM NaCl in my HBS-EP+ running buffer. The problem is that there are very few viruses in my samples. Having a high ligand capture around 1000 RU I only manage to get 20 RU max of analyte. Even if I managed to lower the non-specific binding to 5 RU this is significant when subtracted from my 20 RU value (FC 2-1) which keeps me form quantifying them accurately.
I have tested many ways to get rid of the non-specific binding and 5 RU is as low as I can get. Now I’m trying to find ways to get a stronger analyte signal with this antibody that works really well in ELISA tests but not in SPR.
It appears that the epitope in VP3 is not very accessible thus the antibody “needs more time” in order to capture the virus.
I was thinking of raising the Biacore temperature or adding a “stabilization time” in the method so the analyte is able to interact longer with the antibody. Using a higher antibody concentration doesn’t help (I always get around 1000 RU of capture). The ligand capture is stable all along the cycles and regeneration is ok.
Finally, I have compared 2 monoclonal antibodies A and B both are directed against the same virus, different viral protein (VP3 vs VP1 respectively).
With antibody A I get the low response (20 RU) and with the antibody B I’m able to get a higher response (100 RU) in the exact same conditions, (same capture level around 1000 RU). Despite this, I’m trying to find a way to make antibody A work as good as antibody B (or at least prove that even if the have different kinetic profiles they both should allow me to quantify the virus in my samples)
My first idea was that the virus is too big to be detected. The evanescent field extents only 300 - 400 nm above the sensor surface. The thickness of the dextran layer is ~100 nM and on top of that you have an antibody. So the virus has to bind to the antibody to generate response. But size most of the virus is outside the evanescent field it wil not generate any response.
Second, the dextran matrix can be imagined as a tangled heap of wool. On the matrix but also in the matrix the antibody is immobilized. And the size of the virus is probably too large to penetrate into the matrix.
I would suggest using a sensor chip without a dextran matrix to try to detect the virus. However, IgG immobilization efficiency will be lower.
I could not find an example publication for you, but will look out for it.