ChemDiv’s Small Molecule Tools for Protein-Protein Interactions
Drug discovery’s greatest successes have been unevenly distributed as far as targets are concerned. It is an often-quoted statistic, for example, that more than one third of all currently marketed drugs target G-protein coupled receptors.
The reason for this focus is that G-protein coupled receptors – as well as another area of small-molecule drug discovery’s big success, enzyme inhibition – naturally interact with small molecules, and tend to do so in deep pockets that also make good binding sites for drugs.
Protein-protein interactions (PPI), by comparison, can look a lot like maps of the world from Christopher Columbus’ time: not particularly detailed, but definitely covering an area that’s big. And flat.
But most of proteins’ interactions in cells are not with small molecules in deep pockets. They are with other proteins. And so, also like in Christopher Columbus’ time, it’s clear that there are riches to be discovered in that flat expanse, and researchers have been working on ways to target protein-protein interactions with small molecules.
One way to target such flat interactions is to catch them when they are not flat. Many proteins habitually switch between different conformations, some of which can offer binding opportunities in the form of temporary pockets. And small molecules can induce such conformational changes when they bind to protein surfaces.
Another approach has been to focus on so-called hotspots. Even though the overall interaction surface between two proteins is often large, more detailed analysis of how binding occurs has sometimes shown that it is possible to disrupt overall binding by focusing in on relatively small areas of interaction where a disproportional amount of binding energy is concentrated. Computational methods exist to predict such hotspots, and by focusing on them it is possible to bring the target surface down to a manageable size.
A recent article by Thangudu et al on the importance of binding hotspots described the chemistries that have been found to be useful for targeting such hotspots, and protein-protein interactions more generally. In that paper, the authors wrote “It was found that PPI inhibitors usually represent relatively large rigid small molecules, containing hydrophobic and aromatic groups… A recent study showed that SVM (support vector machine) kernels can be successfully used to select molecular descriptors for PPI inhibitors, which characterize specific molecular shapes and the presence of a privileged number of aromatic groups.” It’s worth noting here that these are precisely the organizing ideas behind ChemDiv’s new Escape from FlatLandSM chemistry design strategies, complemented by our Targeted DiversitySM framework; in this context incorporating known peptidomimetic recognition motifs, as well as PPI modalities based on tertiary and quarternary structures.
Another type of chemistry that has shown promise in targeting protein-protein interactions are macrocycles, chemical structures containing rings of at least 12 atoms. Such macrocycles tend to be fairly large, an advantage in targeting large protein-protein interaction surfaces, but still small enough to penetrate cells in the first place. ChemDiv has developed a number of macrocycle scaffolds, assembly-ready building blocks, and validation sets of macrocyclic screening probes to enable you to exploit these emerging insights.
Protein-protein interactions can be classified in different ways; in a 2007 review, Matthew Smith and Jason Gestwicki suggested that such interactions can be categorized into four different types, depending on whether binding is tight or loose, and whether the interaction is wide or narrow.
It is perhaps not surprising that some notable successes of targeting protein-protein interactions have come where those protein-protein interactions are structurally somewhat similar to medicinal chemistry’s traditional areas of success. In Smith and Gestwicki’s classification scheme, such interactions are “tight and narrow.”
One example is the interaction of an alpha-helix of one protein with a corresponding groove of another. The tumor suppressor p53 and its binding partner MDM2 interact in such a way. The nutlins, which activate p53 by disrupting that interaction, were among the earliest small molecule protein-protein interaction inhibitors reported when they were first described in 2004 by Vassilev et al. (see p53-MDM2 library).
But Smith and Gestwicki describe successful examples of targeting all four types of interactions. Even the most challenging category, “loose and wide,” has produced some small molecules that are capable of prying apart two protein partners.
One interesting subgroup of the protein-protein interactions that are tight and wide in Smith and Gestwicki’s classification system are the amyloids. The most famous, or infamous, example of this group is the amyloid plaque, aggregates of amyloid-beta that are anatomical calling cards of Alzheimer’s disease. But the more general definition of an amyloid is any protein that can polymerize and forms beta-pleated sheets. Amyloids meeting that definition have been implicated in more than a dozen diseases. Those include Alzheimer’s disease and other neurodegenerative disorders. But amyloids also play roles in other types of diseases, such as type 2 diabetes, atherosclerosis, and rheumatoid arthritis.
Collectively the pathologies resulting from beta sheet assemblies due to abnormal quarternary structure and resulting poly-molecular interactions are classified as protein misfolding diseases.
Clearly, Amyloids bind to each other via multiple complex interactions, and disrupting their cumulative interaction enough to prevent or reverse the formation of pathological/toxic aggregates is a formidable task. In addition, when amyloids are due to misfolded proteins – as is the case, for example, in prion diseases – such misfolding can occur in several different ways, and small molecules that can prevent one form of misfolding may have little effect against another.
In a 2009 paper, Robert et al. reported they could eradicate diverse prion strain structures using a combination of two small molecules, the polyphenol epigallocatechin-3-gallate (EGCG) and the 4,5-Dianilinophthalimide analog DAPH-12 broadly inhibited different amyloid formations in yeast, and reversed established amyloids. The paper, as well as some other small molecules that target amyloids, show that even protein-protein interactions with large contact surfaces that lack clear hotspots can, in principle, be targeted by small molecules.
Roberts BE, Duennwald ML, Wang H, Chung C, Lopreiato NP, Sweeny EA, Knight MN, Shorte J (2009): A Synergistic Small-Molecule Combination Directly Eradicates Diverse Prion Strain Structures. doi: 10.1038/nchembio.246
Smith, MC, Gestwicki, JE (2007): Features of Protein-Protein Interactions that Translate into Potent Inhibitors: Topology, Surface Area and Affinity. doi: 10.1017/erm.2012.10
Thangudu RR, Bryant SH, Panchenko AR, Madej T (2011): Modulating Protein–Protein Interactions with Small Molecules: The Importance of Binding Hotspots. doi: 10.1016/j.jmb.2011.12.026
Vassilev LT , Vu BT, Graves B, Carvajal D, Podlaski F, Filipovic Z, Kong N, Kammlott U, Lukacs C, Klein C, Fotouhi N, Liu EA (2004): In Vivo Activation of the p53 Pathway by Small-Molecule Antagonists of MDM2. doi: 10.1126/science.1092472
ChemDiv design strategies, libraries, and tools for PPI-driven discovery and development