Struggling to model nice-looking data

Hi everyone, 

Has anyone had any experiences where they've obtained nice-looking SPR data but fitting on the machine doesn't work at all? I've recently been using an NTA chip to look at the interaction between a peptide dimer (dissolved in running buffer - pH 8.0 10 mM HEPES, 500 mM NaC, 0.05 % tween-20) and a histagged protein and while my data looks nice, the Biacore is really struggling to model the dissociation. Could this be from secondary binding effects or potentially a bulk / RI issue?

One suggestion I've had is to try using 5mM imidazole in the running buffer which might prevent some unspecific interaction between the peptide and the Ni2+ ions at the chip surface (but would also surely remove the immobilised protein which is immobilised via the histag)?


Thanks for any advice!!!!!!

Hi,
What looks like a nice sensorgram to the eye can be one which has not an 1:1 interaction. I think this is here the case. The bulk jump but especially the non 1:1 interaction is causing the deviation for the fitting. You can try a hetrogeen ligand model and model two interactions to see if the fitting gets better. Then also try a steady state fitting and compare the kinetic values (KD).
Adding imidazole can minimize non-specific binding but as you say also wash away the ligand. You have to try.
In addition: the standard protocol uses running buffer with EDTA  ( www.sprpages.nl/sensor-chips-intro/biacore-sensor-chips/nta ).

Kind regards
Arnoud

Hi Arnoud, 

Thank you so much for the advice. I realise that I forgot to mention that I did have EDTA in the running buffer so that isn't an issue. 

I tried a heterogeen ligand model and it looked much better, and the kinetic values were similar ish to a steady state fit (kD = 130 nM and 80 nM in heterogeen vs kD = 170 nM in steady state). I'm not sure if this is close enough to be acceptable, although these values are around what I would expect them to be and perhaps would be helped with multiple repeats.

Best wishes, 

Josh

Hi Josh,
In heterogeen model also check if the 130 nM has the most response, that is an additional indication that it is a plausible value.
You can also capture less ligand to minimize mass transport etc. With 90 RU now you can easily go down to 30 RU. This can make the kinetics also more 1:1.

kind regards
Arnoud

Hi Arnoud, 

Unfortunately, the 130 nM Kd has a lesser RU than the 80 nM one, but I will try again with reduced ligand immobilisation to try and get the data to be more 1:1.

I was also wondering in what situations it is acceptable to use a non-1:1 binding model and whether non 1:1 binding models would be accepted for publication. 

Best wishes, 

Josh

Hi Josh,It is difficult to defend the heterogenic model in a publication. You have to explain why you think it is the right model to report. Often essential details are left out and the reader must guess if the experiments and data analysis is done properly ( www.sprpages.nl/experiments/reporting-results ). Just reporting one set of kinetic values in a table is in my opinion not enough. You have to be honest about your data. If it doesn't fit then don't report the values or explain the doubt you have. And yes, we have a lot of kinetic experiments we cannot report bacause we don't get good fits. But they serve as a starting point for new experiments and still can generate new interaction theories.

I always encourage to read all of the papers of David Myszka ( www.sprpages.nl/reviews/surveys ). Then you know what is expected from you when reporting kinetics.

I hope you get good results with lowering the ligand concentration.

Regards
Arnoud