Complex interactions by surface plasmon resonance - How to calculate them... (Jan/04/2006 )

Hi,

has somebody some good experience with BIAcore technology?

i'd like to know what is the way to interprete complex sensorgrams which do not fit with usual models (from Langmuir to trivalent or tetravalent models) with all parameters fitting global at all concentration of analyte... Is this natural to put association and dissociation constants to local fitting mode in order to obtain nice curves with low chi square? But how to interpret this?
What about the Rmax paramater since the mass of ligand on the chip is not varying... Could this parameter be different at different concentration?

i'd like also to discuss with some people who have interest in analysis of cooperative interactions by surface plasmon resonance, to compare our results and discuss about the calculations used

i'm sure this would be usefull for everybody...

thanks a lot!

__Sébastien___

Hi Trypto,
I was doing some Biacore measuring a cytokine binding to a receptor and vice verse. The kinetics weren't a simple binding. The best fit seemed to be a conformational changes model and we did the suggested linked reaction and mass tranfer controls. I've tried analysing them with both local and global Rmax values. Since then Biacore has taken a back seat in the work I've been doing, although a colleague published some work his interaction was less complex and was a very stable interaction.

My advice would be to contact your local Biacore support. Our support in the UK were very helpful and have been involved with running local training courses at various levels. I can give you their details if you want.

Hope that's helpful,
Ceri

Hi Ceri,

thank you for your advice, i have never though that i can obtain such informations from support...
and i will be very grateful for some more informations obtained from your BIAcore regional support

Sebastien___

Hi Trypto,
The guy who helped me analyse some data is Phil Buckle. He's a Senior Application Specialist
at Biacore and his e-mail address is phil.buckle@biacore.com. Try e-mailing your data/questions to him I'm sure he could answers some of your questions or put you in touch with someone who could help.

As I said before we haven't really done anything more with the Biacore recently. Would you expect your interaction of your molecules to be more complex than a simple 1:1 binding ?

All the best,
Ceri

thank you for the e-mail, i have already sent my questions...

our project is to synthetize small trimeric mimics of non-covalent trimeric ligands
we have a lot of informations other than SPR indicating that interaction are non 1:1
so we have try to apply some more complex model (to tetravalent) with some parameters in the local fitting and found some interesting results, but difficult to interpret
in particular all sensorgrams are fitting well with a trivalent model using ka1 and kd1 local (we would like to fit all parameters local but it's not possible using BIAeval...)
as we can not reach equilibrium in our conditions, we extrapolized Req from ka1 and kd1 and get some Hill curve indicating a cooperative interaction

don't you think some informations are missing in such a study?
is our interpretation right since Req are obtained only from kinetics of the first binding event?

Sébastien_

Hi Sebastien,
Not sure I can help you with this. It's been a while since I thought about the Biacore and your data sounds beyond my experience. Are you working with TNF or TNFR family members by any chance? I used to work on them and I know they're trimeric. I think you're right about using the first association constant because without it the other reaction can't take place. Our reaction was linked so we had two sets of constants. On a basic level I used to look to which fit gave me the closest match to the curve (probably not technically correct) and checked that the Chi squared values were low hopefully and the variance of one of the plots wasn't too high. I guess it stands to reason that a ligand receptor interaction would be complex if there are several contact points between the molecules and that ligand:receptor interactions might cause a conformational change or be multivalent.

Biacore run training courses in kinetics. With the complexity of your data it sounds like the basic courses wouldn't be good enough though.

Good luck,
Ceri

Hi Ceri,

we are effectively working on CD40...
depending on the mass of receptor on the chip, we obtain better fit with trivalent model when more receptor is captured, but for some ligands (analyte) the fits are not the best at only some concentrations that why we analyze with some parameters local and observed that KA1 of the 1st interaction is varying with concentration of analyte
for my part i think that kinetics of all the 3 interactions are necessary for extrapolating final equilibrium parameters (Req and Rmax); parameters of the 2nd and 3rd interactions should reflect the cooperativity but there is no equation that combined all these parameters to Req!

i will give here some more informations later if some people are interesting with such an experiment

we stay in touch Ceri
thank you!

Sébastien_