Understanding how allosteric proteins respond to changes in their environment is a major goal of current biological research. We show that these responses can be quantified by analyzing protein energy networks using a method recently developed in our group. On the basis of this method, we introduce here a quantity named energetic coupling, which we show is able to discriminate allosterically active mutants of the lactose repressor (LacI) protein, and of the catabolite activator protein (CAP), a dynamically driven allosteric protein. Our method assumes that allostery and signal transmission can be more accurately described as efficient energy propagation, and not as the more widely used atomic motion correlations. We demonstrate the validity of this assumption by performing energy-propagation simulations. Finally, we present results from energy-propagation simulations performed on folded and fully extended conformations of the postsynaptic density protein 95 (PSD-95). They show that the protein backbone provides a more efficient route for energy transfer, when compared to secondary or tertiary contacts. On the basis of this, we propose energy propagation through the backbone as a possible explanation for the observation that intrinsically disordered proteins can efficiently transmit signals while lacking a well-defined tertiary structure.