LA JOLLA, Calif.—A team led by scientists at The Scripps Research Institute (TSRI) has developed a versatile new method that should enhance the discovery of new drugs and the study of proteins. The new method enables researchers to quickly find ligands that bind to hundreds of thousands of proteins in their native cellular environment. Ligands can be developed into important tools for studying how proteins work in cells, which may lead to the development of new drugs. The method can be used even without prior knowledge of protein targets to discover ligand molecules that disrupt a biological process of interest—and to quickly identify the proteins to which they bind. This research was published ahead of print Jan. 19 in Cell.

 

“This new platform should be useful not only for discovering new drugs, but also for discovering new biology,” said co-lead author Christopher G. Parker, a research associate in the laboratory of TSRI professor Benjamin F. Cravatt, chairman of the Department of Chemical Biology.

 

About 25,000 proteins are encoded in the human genome, but public databases list known ligands for only about 10 percent of them. The new method for exploring this field involves the development of a set of small, but structurally varied, candidate ligand molecules known as “fragments.” Each candidate ligand is modified with a special chemical compound so that, when it binds with moderate affinity to a protein partner, it can be made to stick permanently to that partner by a brief exposure to UV light. A further modification provides a molecular handle by which scientists can grab and isolate these ligand-protein pairs for analysis.

 

“Chemical probes are powerful tools to study protein function, and can serve as leads for potential drugs; however, the vast majority of the human proteome lacks chemical probes,” Parker tells DDNews. “To address this, we sought to develop a general method to broadly assess the small molecule ligandability of proteins in their native cellular environment. In order to achieve this, we chose to use small molecule fragments, which are low molecular weight, structurally simple cores that enable more efficient exploration of chemical space. Fragment screening is a well-established approach in drug discovery, but this method has traditionally been limited to biophysical screening techniques (NMR, crystallography) on purified protein targets.”

 

“We sought to develop an approach that extends fragment-based screening for exploring biologically relevant chemical space in human cells. We did this by appending a photo-activatable ‘capture’ tag to fragments, which would allow us to transform weak, reversible fragment-protein interactions into covalent ones and subsequently identify them using quantitative mass spectrometry. We call these chemical probes “fully functionalized fragment” probes, or FFFs. We were then able to advance these interactions, in several instances, into more developed chemical probes that affect the protein target’s function in cells,” explains Parker.

 

For an initial demonstration, the team assembled a small “library” of candidate ligands whose structural features include many that are found in existing drugs. By applying just 11 of them to human cells, the researchers identified more than 2,000 distinct proteins that had bound to one or more of the ligands.

 

These ligand-bound proteins include many from categories—such as transcription factors—that previously had been considered “un-ligandable” and therefore un-targetable with drugs. In fact, only 17 percent of these proteins have known ligands, according to the widely used DrugBank database.

 

“Proteins can be considered un-ligandable or un-druggable for a number of reasons,” says Parker, such as lacking traditional small-molecule binding pockets or “[requiring] their native environment to maintain their structure and function.”

 

“In our approach, by examining protein ligandability in a more global and unbiased fashion in human cells, we were able to overcome some of these challenges,” he notes. “However, the fragment ligands in our platform inherently have lower binding affinities and are not very selective, so although they may provide initial leads to start with, a substantial amount of work may lie ahead to transform the annotated fragment-protein interactions into ones of higher selectivity and affinity.”

 

The researchers used further methods to identify, for many ligand-protein interactions, the site on the protein where the coupling occurred. The candidate ligands initially used to screen for protein binding partners are generally too small to bind to their partners tightly enough to disrupt their functions in cells. But the team showed that, in multiple cases, these initial small (“fragment”) ligands could be developed into larger, more complex molecules that display higher-affinity interactions and disrupt their protein partner’s functions.

 

“Having more fragment probes should, in theory, expand the number of potential targets as well as provide more information about fragment preferences among proteins, which could aide in efforts to develop more potent and selective ligands down the road. In our report, we actually made a much larger library of approximately 450 fragment-based probes and demonstrated their utility in phenotypic screens. Further expansion of our current fragment probe library is ongoing,” Parker comments.

 

Collaborating chemists at Bristol-Myers Squibb helped create a library of several hundred slightly more complex candidate ligands. With TSRI colleagues Prof. Enrique Saez and co-first author Andrea Galmozzi, the team then tested these ligands to find any that could promote the maturation of fat cells (adipocytes)—a process that in principle can alleviate the insulin resistance that leads to type 2 diabetes.

 

Traditional functional screens of this type do not pinpoint the proteins or other molecules through which the effect on the cell occurs. But with this new discovery method, the researchers quickly found not only a ligand that strongly promotes adipocyte maturation but also its binding partner, PGRMC2, a protein about which little was known.

 

“We feel that a general framework has been laid out in this paper to not only broadly map out protein ligandability in cells but also how to use this information to develop ligands that may perturb protein function. Moving forward, we will continue use this platform to develop chemical tools to study interesting proteins as well as broaden our understanding of what is the ligandable, and, by extension, druggable proteome,” Parker concludes.