Q&A questions
1. With an ATP depletion assay, how do you ensure that you are truly measuring your kinase of interest and not a contaminating kinase or other ATPase?
2. Neither can be an active site probe. Screening natural proteins sounds great, but I have not seen many examples of substrate-specific, small molecule inhibitors. If all kinases see 3 mM ATP in cells, why do ATP Km values range from 1 to 100 uM?
3. What error rates are typical when following up single point HTS with IC50s? Do you routinely run compound solubility checks? If so, how?
4. What are the key issues that you focus on when adopting a new assay (both in vivo/in vitro), or new format for your assays? What are the key challenges you perceive facing kinase (PK, LK, CHO kinases) screening going forward?
5. Do you screen with the same kinase construct as used for co-crystallization?
6. How many big pharma companies use this HTS and is it their own or do they outsource?
7. It seems you are concerned with physiological relevance or similarity in your biochemical assay. What about ions and other buffer agents?
8. Do any of the panelists screen other enzymes with ATP-binding sites as part of their kinase screen (or as a follow-up screen) to evaluate off-target effects?
Answers from Kirk McMillan from Exelixis:
1. We evaluate kinases that are typically purified by at least two purifications for enzyme concentration dependency (typically 1-20 nM), ability to inhibit the ATPase activity using known or general inhibitors, substrate-dependent ATP consumption, et cetera.
2. We do typically screen at Km and evaluate/improve cell activity during the course of lead optimization.
3. We routinely have primary hit rates for kinases of 0.1% screening orthogonal pools in duplicate at 1-5 uM. Compounds are cherry-picked for re-testing and dose-response for positives.
4. I believe kinases are well served by existing assay technologies with respect to screening recombinant enzymes in biochemical assays. Cell-based assays targeting effects on specific substrates or autophosphorylation are important for lead validation and optimization and are not always amenable to microtiter plate format and thus have limited throughput.
5. Generally not. We make 10 to 30 different constructs as part of initial reagent development and direct those exhibiting greatest stability (calorimetry) and solubility to use in crystallization screening trials. HTS is focused on specific activity.
6. No answer.
7. We optimize assay conditions to maximize specific activity and kinase stability and minimize kinase concentration. Our assays are typically conducted in 20 mM Hepes, pH 7.5, 5 mM MgCl2 and low concentrations of detergents, such as Tween 20, many assays also contain low concentrations of BSA.
8. We do not, but screen ~100 kinases.
Answers from Kim Arndt from Wyeth:
1. We do not use ATP depletion assays.
2. You are right. There are not more examples of substrate or non-ATP competitive kinase inhibitors. However, if we are to find them, my prediction is that using natural protein substrates, kinases that are more than just the kinase domain, and higher ATP concentrations will favor finding the non-ATP competitive inhibitors.
3. Greater than 90% of the hits re-test. We only do solubility checks later in the process—once we are down to about 10 classes of hits.
4. A concern for me is finding new, patentable, chemical scaffolds that are active for inhibiting kinases.
5., Yes …
6. I do not know what is being asked by this question, e.g., “this HTS”
7. We try to use acetate buffers rather than chloride.
8. We currently screen only kinase panels for specificity. There are binding assays for ATP-binding proteins (beads with ATP-like compounds covalently bound that bind to many ATP-binding proteins—the proteins bound to the beads are analyzed by mass spec to determine if a given compound displaces one or more of the ATP-binding proteins) but we have not yet fully adapted this technology.
