Regulation of the c-Abl (ABL1) tyrosine kinase is important because of its role in cellular signaling, and its relevance in the leukemiogenic counterpart (BCR-ABL). Both auto-inhibition and full activation of c-Abl are regulated by the interaction of the catalytic domain with the Src Homology 2 (SH2) domain. The mechanism by which this interaction enhances catalysis is not known. We combined computational simulations with mutagenesis and functional analysis to find that the SH2 domain conveys both local and global effects on the dynamics of the catalytic domain. Locally, it regulates the flexibility of the αC helix in a fashion reminiscent of cyclins in cyclin-dependent kinases, reorienting catalytically important motifs. At a more global level, SH2 binding redirects the hinge motion of the N and C lobes and changes the conformational equilibrium of the activation loop. The complex network of subtle structural shifts that link the SH2 domain with the activation loop and the active site may be partially conserved with other SH2-domain containing kinases and therefore offer additional parameters for the design of conformation-specific inhibitors.

MOTIVATION: Recent developments of several analytical techniques and fast sampling procedures are making it possible to study non-linear dynamics on an automated basis. Approaches that suggest experimental setups and fully exploit the data are needed. To prevent unreliable fits of parameters to experimental data, model-independent methods would be advantageous. One such method consists in representing the response of a system by means of a transformation or functional, of an excitation, i.e. a flux of a metabolite. RESULTS: A functional of unknown form can be expanded in series if its functional derivatives are known. An algorithm for calculating such generalized derivatives from impulse-perturbation experiments was developed. The only assumption was that the implicit kinetics is time-independent, i.e. the system is time-invariant. The method is illustrated on a part of fructose catabolism. A reaction network that involves 14 metabolites and 11 enzymes and includes several branches and feedback loops is considered. It is shown that the method provides good approximations to the responses of this complex system. AVAILABILITY: FUNDER is available from http://bbm1.ucm.es/torralba/funder/down/ SUPPLEMENTARY INFORMATION: http://bbm1.ucm.es/torralba/funder/