How can protein folding be used for computational drug design?

Protein folding is not just about predicting structures from sequences, but also identifying different states explored by a protein, or understanding the protein landscape. I can think of two/three major additions to the drug design community (apart from designing peptide drugs  and modifying or designing proteins/enzymes), that have originated from protein folding studies:   

Incorporating protein flexibility: 

• We  know for a fact that *a* single conformation is not an adequate representation of a protein molecule. We also know that for enzymes, the lock-and-key hypothesis fails to explain the transition state and hence considered obsolete. Therefore, understanding protein dynamics and flexibility is important as it plays a role in protein function and catalysis. This means, docking or screening ligands with one single protein molecule is not sufficient.

Figure (from L to R) : cartoon images of representative folding funnels of flexible proteins, rigid proteins and all-or-none folding transitions of a single conformation.  [Accommodating protein flexibility in computational drug design.]

• Accounting for conformational selection [where a protein is assumed to exist in a number of energetically similar conformations, and the ligand selectively binds one of those conformations, thereby increasing the number of that particular conformation] has lead to newer strategies in the design community, especially for designing protease inhibitors.

Conformational selection: P$P$ and P∗$P^{\ast }$ are two states, L is the ligand, although P∗$P^{\ast }$ has a higher energy, P∗L$P^{\ast }L$ is more stable. 

Identifying allosteric sites:

• Protein folding studies have helped in locating allosteric sites (which can't be identified using experiments) [Equilibrium fluctuations of a single folded protein reveal a multitude of potential cryptic allosteric sites] 

Newer targets:

• Studies on molecular chaperones, protein misfolding/aggregation have helped in identifying newer targets. For example: heat shock proteins (Hsp27/Hsp90 for cancer), and  hyperphosphorylated tau for CNS disorders.