I am fairly new to SPR and have been encountering some issues regarding negative response units in experiments where I know from other types of experiments, such as western blotting (and in previous SPR experiments), binding should occur.
I am using a CM5 chip that has been immobilized with anti-IgG Fc mAbs and a Biacore T200.
The ligand that I capture on the chip is a sialic acid binding lectin called CD22 (130 kDa). The analyte that I use as a model glycoprotein is Haptoglobin (100-200 kDa), a highly sialylated serum glycoprotein.
In the first graph from my first attempt at this experiment we can see that the CD22-Fc capturing is successful, resulting in a delta RU of approximately 1200. However, if we look at the blank substracted (Fc 2-1 minus blanks) response for Haptoglobin in HBS-EP+ buffer, we see that there is a concentration-dependent decrease in RU.
Initially, I thought the buffer might be the issue so I switched to a buffer that is reported to be more suitable for glycan-lectin interactions (TSM-Tween). This initially resulted in 1 experiment with the exact same settings and concentrations where Haptoglobin seemed to bind to CD22-Fc. However, trying to repeat this experiment at ligand RUs in TSM-Tween resulted in negative analyte binding again. As far as I know all the batches of samples were the same and the concentrations were also the same. Does anybody have pointers to fix this issue?
1) How is the response on the reference? When the response on the reference is higher than on the active channel you get negative curves after reference subtraction.
2) It is possible that the haptoglobin is displacing/dissociating the CD22-Fc from the antibody. When the Fc-part of the CD22 has a high(er) affinity for the sialic acids on the haptoglobin than the affinity of the anti-Fc mab this is a possibility. Because at the end of the analyte injection the baseline is lower and stable hence some of the CD22-Fc must be dissociated during the haptoglobin injection.
Why did you use the capturing method and not a direct immobilization of the CDD-Fc?
It does seem like in Fc 1 (top), the Haptoglobin signal increases over time whereas it stays relatively constant in Fc 2 (below). I also added a blank (just buffer only) for comparison for both Fc 1 and Fc 2. The lowest (3rd) screenshot is the Fc 2-1 of the blank. The blank also seems to increase over time. Is this normal/what you would expect? As for point 2, it might be possible that the CD22-Fc coating is lower at the end than at the beginning, but I would not expect the Fc-tail to bind to sialic acids on haptoglobin. I used this method of indirect capture as it was easier; We had already made anti-IgG Fc CM5 chips for other experiments and I figured I could use it for this purpose. I hope my explanation is clear so far!
Not sure I can follow you; maybe some confusion on what a reference channel and a blank is.
What do you use as a negative reference channel? You should have two with anti-IgG Fc channels and capture on one the CD22-Fc. The one without CD22-Fc will show you the non-specific binding of the haptoglobin to the IgG. A blank injection (buffer only) should in theory give no binding curve and should rise to a certain response level and stay horizontal until the end of the injection although some positive or negative drift is common and yours is not that high (www.sprpages.nl/troubleshooting/baseline-drift
For evaluation of the results, first subtract the reference channel and then the blank (double referencing: www.sprpages.nl/experiments/data-processing
By using double referencing, you compensate for both reference and blank injection behaviour.
I would expect the third sensorgram flatter but this is what you get. Maybe you can improve the sensorgrams by longer equilibration (after regeneration) or by washing the system (desorb) to see if the curves get better.
Hi Arnoud. Thanks again. I'll read the sources you linked! Maybe I did not explain it well. In the first image, with the 2 sensorgrams, the top image is the Fc1 sensorgram, which shows a blank run and a run in which Haptoglobin the analyte. The bottom image is the Fc2 sensorgram from the same samples. It looks like the haptoglobin is binding to the Fc1 (which only has anti-IgG but no CD22-Fc) whereas the haptoglobin does not seem to bind to the Fc2. Is this plausible, looking at the image of the Fc1 and Fc2 I showed?
Yes, it is plausible. Sometimes the non-specific interaction is higher than the specific. I have two proteins with the same behaviour. It is impossible to do SPR with them on a CM5 sensor chip. Other surfaces we did not try. We were unable to pinpoint why the non-specific was higher than the specific because the interaction between both proteins can be measured in functional assays.
Sometimes SPR is not the right technique.