I am working on a kinase inhibitor design project and I realised that there are some key questions that we need to get answered from the biologists before we can rationalize the selectivity of various kinase inhibitors for a given binding site. I also realised that these questions need to be answered for many other kinds of protein-inhibitor interactions.
1. Whenever we get different IC50 data for two inhibitors, we immediately try to look at binding interactions that may be different for the two moelcules to rationalize this observation. But as I have alluded before, it's not the IC50 but the Ki that's really to do with different binding interactions. The Ki and IC50 are related by an equation that includes both the Km value of ATP and the concentration of ATP in the two experiments, or in general, these two parameters for the natural binding substrate for the protein. Only if these two are the same for both inhibitor experiments is the IC50=Ki. So make sure you confirm this. Otherwise, extrapolate and calculate the new IC50s based on identical values for these parameters. Then rationalize the IC50s based on binding interactions.
2. For many kinases, three events are absolutely essential for activation:
a. Phosphorylation of one Ser, Thr or Tyr residue,
b. Binding of ATP (duh), and
c. Dephosphorylation of another Ser, Thr or Tyr residue.
Think of it like a logic gate. IF the answers to all a. b. and c. are YES, THEN the kinase will be activated and proceed to perform its function. (I got this from Alberts et al.'s Molecular Biology of the Cell)
In the assays that are run, it is important to know (and not very easy to always determine as I have been told) whether the necessary residue is phosphorylated or not. For one thing, inclusion of this knowledge in your docking and modeling can naturally make a big difference. And secondly, depending on the state of phosphorylation, you can think of different modes of inhibition for your inhibitor (eg. ATP blocking + substrate blocking).
As usual, it's important to know what the biologists are doing. They don't know the nuances of modeling/crystallography and you don't know the nuances of their assays. But it's important for both camps to think of questions which the other camp should answer that will affect their own work.
I'm not so sure about your three conditions for an active kinase. I might be misunderstanding, but this is what i think. Not all kinases are phosphorylated to make them active, and not all (or many) have activating dephosphorylations either. Personally i would break it down as saying you need
ReplyDelete1) the kinase to be in the active state (ie no inhibitors/pseudosubstrates in the active site, kinase is phosphorylated)
2) the substrate to get into the active site (consensus sequences in the substrate must match the kinase's requirements, the proteins must be close enough to interact). There are also a few kinases which have secondary non-active site docking interactions which are needed for efficient phosphorylation. Feel free to email me if you want to chat about this more, or would like some references
Hi. Thanks for the comment. I am pretty sure this is the case with the CDKs, but yes, I don't know if it exists for the majority of kinases. As you said, probably not. Do you have some references which catalog such characteristics?
ReplyDeleteLets see... here are a few that i like:
ReplyDeleteActivation Loop Phosphorylation:
Mora A, Komander D, van Aalten DM, Alessi DR.
PDK1, the master regulator of AGC kinase signal transduction.
Semin Cell Dev Biol. 2004 Apr;15(2):161-70.
Pseudosubstrate Regulation:
Kemp BE, Pearson RB, House C, Robinson PJ, Means AR. Regulation of protein kinases by pseudosubstrate prototopes.
Cell Signal. 1989;1(4):303-11.
Hubbard SR. Autoinhibitory mechanisms in receptor tyrosine kinases. Front Biosci. 2002 Feb 1;7:d330-40.
I think i might try to blog on this in the next little while too on my page.
Docking Interactions:
Biondi RM, Nebreda AR. Related Articles, Links
Abstract Signalling specificity of Ser/Thr protein kinases through docking-site-mediated interactions.
Biochem J. 2003 May 15;372(Pt 1):1-13.
I'm glad you're paying attention to Cheng-Prusoff :-). You should always be careful when dealing with IC50's (which are reported more often than Ki's). I love that equation.....used that extensively in my first paper in grad school :-)
ReplyDeleteIf you're working on designing kinase inhibitors (i don't like them, for various reasons...starting from the difficulty of getting specificity, which I don't want to get into here..) you could also have a decent affinity inhibitor, but when you make your models, you need to always include ~1 mM (or more) intracellular ATP concentrations, to then estimate if you have a bat's chance in sunlight of getting something that'll have decent enough affinity as well as specificity to become a drug candidate.
Sunil, yes, ATP is wildly rampant in cells. That is one reason why you just cannot get selectivity if you don't have some minimum potency. And why don't you like them? They don't bite if you trust them, feed them with robust functional groups, and put them in a secure, hydrogen bonding or hydrophobic, environment :)
ReplyDeleteThe thing is, the selectivity issue is getting more and more circumventable through exploiting subtle differences between kinases. But yes, it's far from a done deal.