Send to printer »

Feature Articles : Feb 15, 2014 (Vol. 34, No. 4)

Mining Protein Interactions for Smarter Drug Design

  • Nsikan Akpan, Ph.D.

Protein-protein interactions (PPIs) hold humans together. A single protein or enzyme by itself is useless, but synchronize it with the right partners, and suddenly life forces appear—from the absorption of nutrients to the contraction of muscles.

Untapped targets for drug design lurk in these essential protein bonds. Researchers are employing many different approaches—such as signal transduction, quantum chemistry, and bioinformatics—to unearth the molecular crevices where a small molecule or peptide inhibitor could knock a disease off course.

To review disease arenas where PPI manipulation is showing much promise, scientists will convene at CHI’s “Protein-Protein Interactions” conference in April.

Often the first barrier to targeting a PPI is the cell membrane. For instance, antibodies are titans when it comes to recognizing PPIs, making them essential tools for epitope mapping and a bevy of research assays, such as immunofluorescence, ELISA, and chromatography. Yet antibodies have limited clinical scope because they do not easily access the intracellular compartment.

Cyclotides, a large family of plant-derived peptides, may offer a substitute, according to Julio Camarero, Ph.D., a pharmacologist at the University of Southern California. His upcoming presentation will cover how cyclotides interfere with the small binding clefts or large surfaces that define protein interactions. These peptides not only possess certain antibody-like features, they also present a bevy of advantages, including the ability to cross cell membranes.

“PPIs are a difficult target—some call them undruggable—because large surface interactions are hard to antagonize with small molecules,” says Dr. Camarero. Nonetheless, Dr. Camarero’s team intends to develop peptide-based therapeutics. In his presentation, Dr. Camarero will discuss how cyclotides are potent drug inhibitors both inside and outside the cell.

Cyclotides were initially discovered in a tea extract that pregnant African women used to induce labor, illustrating their cell-penetrating prowess, oral bioavailability, and clinical promise.

Cyclotides are medium-sized peptides (30–40 amino acids) that have a cyclic backbone, which consists of six “loop” sections interspersed with cysteine residues. With these residues connected via disulfide bonds, the loops converge on a cysteine-based “knot” at the peptide’s core. This conformation makes cyclotides exceptionally impervious to thermal, chemical, and enzymatic degradation.

Not only are cyclotides ultrastable, they are highly diverse. In fact, five of their loops are hypervariable. These properties, according to Dr. Camarero’s team, make cyclotides suitable as scaffolds for peptide-based therapeutics.

“We have immobilized the peptides onto glass to make microarrays, and even after washing them with organic solvents, they are still functional, says Dr. Camarero. “You can boil them or change pH, and they still fold because their state is kept by covalent interactions of the cysteine knot’s disulfides.”

In a recent cancer study, they took advantage of the cyclotide’s hypervariable loops to create a drug that reactivated p53 pathways in tumors. Many cancers switch off p53, a protein trigger for cell death, by overexpressing two proteins: Hdm2 or HdmX. Hdm2 is a ubquitin ligase that marks p53 for proteasomal degradation, whereas HdmX binds and sequesters p53 via protein interaction-mediated sequestration.

Using protein engineering, Dr. Camarero and his colleagues attached a p53 peptide domain to a cyclotide that could then compete for the p53 interaction with HdmX and Hdm2. The engineered cyclotide efficiently killed four cancer cell lines; its efficiency equaled that of the cancer drug Nutlin-3. The cyclotide outperformed Nutlin-3 in a mouse xenograft tumor model with HCT116-p53+/+ human colon cancer cells.

Fluorescently tagging cyclotide hypervariable regions yield nascent biomarkers, akin to fluorescently labeled antibodies, except some cyclotide-based molecules can enter cells with the same efficiency as the popular cell-penetrating protein domain Tat.

Dr. Camarero’s medical chemists have engineered other cyclotides that block HIV-1 viral entry as well as protease inhibitors. Some cyclotides naturally function as insect repellent, so team members are also working to create green pesticides from the peptide family.

“We envision this scaffold potentially playing an important role in the future development of peptide-based therapeutics,” Dr. Camarero adds.

The Other Side Of Proteasome

The proteasome is a cell’s trash compactor, degrading the majority of eukaryotic proteins—close to 80%—in order to maintain cellular homeostasis. Though the exact mechanisms are unclear, by highjacking the proteasome, cancer cells boost their proliferative capacity and stave off apoptosis, which are two keys factors involved with malignancy.

Cancer cells are more sensitive to proteasome inhibition than normal cells. Yet targeting the complex’ catalytic site has only yielded two FDA-approved drugs, and they are suitable only for blood cancers (multiple myeloma and Mantle cell lymphoma).

One person moving past classic inhibition and taking aim at allosteric targets—sites away from the catalytic cleft that regulate proteasome activity—is Maria Gaczynska, Ph.D., an associate professor of molecular medicine at the University of Texas Health Science Center at San Antonio.

She will present a novel small molecule that may extend proteasome inhibition to the realm of solid tumors. “Our idea is to expand the significance of proteasome inhibitors to as many types of cancer as possible,” says Dr. Gaczynska, whose team used molecular modeling, rather than library screening, to rationally design the small molecule.

The proteasome is a modular factory composed of two subunits, whose interactions dictate protein degradation. Shaped like a cylindrical garbage can, the 20S subunit opens to receive and enzymatically degrade ubiquitin-tagged proteins, but access to the 20S subunit is controlled by interacting with the 19S subunit.

The nascent inhibitor attacks the interface between the 20S and 19S subunits, silencing the proteasome and showing efficacy in carcinoma as well as in triple-negative breast cancer cell lines. Moreover, her team has witnessed some synergy with the FDA-approved drugs that target the catalytic site.

Cancer cells do not have as many mechanisms for avoiding allosteric inhibition, according to Dr. Gaczynska. With a standard competitive inhibitor, the first round of proteasomes is blocked, and endogenous substrates accumulate. Yet as the process snowballs, a tipping point is eventually reached where the substrate can outcompete the drug, rendering it null and void. If new rounds of proteasomes are produced, they won’t be blocked.

“With our concept, the cancer cell cannot do that. It will produce more of the 19S and 20S subunits, but our inhibitor will still be blocking [activity],” says Dr. Gaczynska, adding that some cancers develop small mutations in the active core, generating resistance to catalytic inhibitors. Allosteric sites between the 19S and 20S subunits are more conserved, so cancer adaptability “won’t be so fast or so easy.”

Heat Shock Proteins

Allosteric targeting with the cell’s protein recycling system—heat shock proteins—will feature in the drug design talk by Jason Gestwicki, Ph.D., associate professor of pharmaceutical chemistry at the University of California at San Francisco.

“Heat shock protein 70 (Hsp70) is a molecular chaperone that regulates protein quality control, including aspects of protein folding, trafficking, and turnover,” he explains. “To engage in these activities, Hsp70 is thought to work in concert with co-chaperones that help direct Hsp70-bound clients to their appropriate cellular fate.”

Protein misfolding and aggregation, which kill cells, are hallmarks of a plethora of neurodegenerative diseases, such as Alzheimer’s disease and Parkinson’s disease. Hsp70 directs folding, disaggregation, refolding, and degradation of substrate proteins through the assistance of co-chaperones.

“Our workhorse method involves combining purified Hsp70 with its co-chaperones and then using high-throughput screening for molecules that selectively inhibit these combinations,” says Dr. Gestwicki. “We like this method because it recapitulates aspects of Hsp70 biology, namely, the ability of the chaperone to work side-by-side with its co-chaperones in the cytosol.”

His team’s system is dubbed the “grey box,” because it lacks the inherent challenges of screening in cells or organisms—“black boxes”—but offers a versatile realm for examining interactions between purified proteins. Using this in vitro screening method, they identified co-chaperone interactions that governed the shuttling of proteins through the cell’s quality control system. Some small molecules identified by the team inhibit only certain combinations of co-chaperones, whereas others never directly target Hsp70.

Earlier this year, Dr. Gestwicki collaborated with Andrew Lieberman at the University of Michigan to show that an allosteric Hsp70 inhibitor reduces accumulation of an androgen receptor in models of Kennedy’s disease, a progressive neuromuscular disorder. He is also exploring Hsp70 co-chaperone interaction in tauopathies, such as frontotemporal dementia.

“We think that the protein quality control systems are rich in underexplored drug targets, including Hsp70, and that allosteric inhibitors and inhibitors of protein-protein interactions will be particularly useful in these settings,” concludes Dr. Gestwicki.

Water and Protein-Ligand Binding

Few places exist where water is more important than with protein-protein and protein-ligand interactions. José Duca, Ph.D., head of computer-aided drug discovery at Novartis, will present on how water factors into a new theory of binding that can be applied to any intermolecular interacting system.

“The energetic costs and gains of moving water determine how fast biomolecules bind and how long they stay together, factors that determine the occupancy of a binding site,” says Dr. Duca. His team’s research explores the solvation structure of biomolecules—the attractive forces that allow an individual compound to associate with solvents. The solvent, in this case, is water.

“Every biomolecule forms a unique solvation structure. Macromolecules form more complex, inhomogeneous solvation structures than small molecules,” Dr. Duca continues. Their theory is that solvation structure drive the energy dynamics that govern biomolecular state, and by default, protein-protein interactions.

“Knowledge of the solvation structure of a biomolecule of interest can be used to design ligands, that is, reverse-engineer the solvation structure into ligands and interpret experimental kinetics data.”

In a 2010 paper, he and his colleagues used the theory to estimate the rates of association and dissociation between Pcsk9 and the low-density lipoprotein (LDL) receptors. LDL receptors coat the surface of liver cells, where they bind to LDL cholesterol and remove it from the bloodstream. Higher low-density lipoprotein cholesterol levels in the blood causes both atherosclerosis and ischemic cardiovascular disease. LDL receptors are destroyed upon binding to Pcsk9 proteins, making the latter a crucial regulator of LDL cholesterol.

“Pcsk9 and the LDL receptor associate and dissociate relatively fast. The observed kinetics profile is well explained by the calculated solvation structure of Pcsk9,” notes Dr. Duca.

According to Dr. Duca, examining PPIs with these techniques will impact chemistry and drug discovery.

“Biomolecular structure is the puppet, and water the puppet master. Enter the ‘solvationome’ as another pillar,” remarks Dr. Duca. “PPIs are part of the ecosystem of targets we work with, and the more we learn about pathway biology, the more relevant we suspect that PPIs will become.”