Dear all,
I am using the following empirical equation to model substrate inhibition kinetics in a UGT-catalyzed reaction: v=vmax*S/(Km+S+S^2/Ki). In the special case when Ki is smaller than Km I get very high uncertainty in parameter estimation. Can someone elaborate on why this could be the case? The variability in measured reaction rates at given substrate concentration is actually very low so I assume it has something to do with the relatively lower Ki. Any advice is much appreciated!
Hi there,
The equation you mentioned is valid for the case where enzyme becomes inhibited by excess of substrate. This phenomenon occurs for the range of concentrations of substrate far above the Km, therefore Ki is always far bigger than Km. The case you are interested in can't be treated and solved with the equation you gave. I then suspect an other mechanism than the excess of substrate (product inhibition maybe?).
If you look at a graph of the data with the predicted curve overlaid, does the predicted curve follow the data points closely? If not, that wold indicate that the equation is not a correct description of the phenomenon, as Dominique Liger said. And the high uncertainty in the parameter values would be a consequence of the poor fit.
@Dominique, for the analyzed class of enzyme substrate inhibition has been often described and therefore been considered first. I will, however try a product inhibition model as well and report back if I get a better fit.
@Adam, please find the plot of substrate conc. vs reaction rate attached. The curve follows the observations well so I assume the high error is not a cause of poor fit.
Thanks for your valuable input!
What values were estimated for Km and Ki, and what were the standard errors? The fit of the curve to the data is very good, and the kinetic rate equation is the correct one for substrate inhibition, so it is odd that you are having trouble.
The estimates are:
Vmax (SE): 28.79 pmol/min/mg (13.29)
Km (SE): 155.4 uM (87.3)
Ki (SE): 39.8 uM (22.05)
Many thanks, David
Hi again,
Having a quick look at your v=f(S) plot, I'm surprised by the high estimate Vm. I would estimate personnally Vm to be around 8 and therefore Km being around 40µM and Ki around 200µM...
My experience with substrate inhibition has been that uncertainty is highly dependent on the fit at concentrations below max observable (roughly 100 uM on your graph) and high concentration. I would double the number of data points below 100 uM and also include data points at 0.5 and 1 mM if possible. It is also the case that it is assumed that all other subtrates/co-factors in the reaction are at saturation. I believe your S/N in the 50 -100 uM range is your biggest limiting factor for an accurate fit.
I agree with Dominique. Try using a different set of initial estimates for the fitting, especially a lower value of Vmax. Sometimes, nonlinear regression programs can get stuck in a local minimum that is not the global minimum if the starting estimates are not well-chosen.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling