Measuring Protein-Protein Interactions

Comments

December 09, 2021

The genomic revolution unveiled a milieu of disease-linked genes. Interrogating a protein’s function is more difficult. Measuring how protein-protein interactions change in response to a cell’s perturbation, though—with chemicals or tweaked genes—can reveal insights into proteins that play roles in specific diseases.

Quantifying protein interactions

Proteins don’t act alone. They slice and dice and catalyze reactions, often in tandem with other proteins. Understanding such protein-protein interactions, or PPIs, is useful to interrogate how, say, a mutation in one gene affects how proteins ‘talk’ with one another to drive a disease.

Scientists have typically measured only one protein-protein interaction at a time. In a slow and laborious process, they grab hold of a protein, pull it from a cell, and map which proteins it interacts with. Methods like quantitative affinity purification mass spectrometry, or q-AP-MS, can measure how protein networks change within their native context, the cell, at a higher throughput.¹

Search Mass Spectrometry
Search Now Search our directory to find mass spec tools for your research needs.

In its basic form, q-AP-MS works like this: Cells (or entire animals) are ‘fed’ with heavy forms of amino acids that, over time, incorporate into proteins. Other cells are prepared, in parallel, to serve as negative controls. The cells are then transfected with a gene encoding the protein to be studied; the protein is typically modified with a tag, such as GFP, so that it can be easily segregated later on. Next, the proteins are slashed from the cell and run through a gel that has antibodies that bind to the protein tag. The proteins are cleaved into smaller fragments, and everything is run through a mass spectrometer. The result is a quantitative snapshot of which proteins bind to the chosen protein.

In the last several years, though, other methods have filled in some of the drawbacks associated with q-AP-MS. For one, fusing a tag to the protein of interest can fundamentally change how the protein behaves. The method also only provides a snapshot of PPIs in the cell at one timepoint, and the method is biased toward strongly interacting proteins; transient connections are harder to detect and measure.²

Solving PPI problems

To overcome the first issue, consider CRISPR.

“For many years, AP-MS experiments have relied on the tagging of proteins with the relatively big fluorescent protein GFP and/or their heterologous over-expression,” says Christian Linke-Winnebeck, Senior Scientist and Leader of the R&D Analytics Team at ChromoTek, which is part of the Proteintech Group. “While this approach has, of course, yielded much knowledge, the protein of interest may still be affected by bulky tags and higher than normal expression levels.”

That’s why, instead of tagging with GFP, labs are now using CRISPR to tag genes directly in the genome, to avoid perturbing the gene’s expression level.³ Other groups, says Linke-Winnebeck, are also using “split-fluorescent proteins” to tag proteins. In that case, “only a small fragment of a fluorescent protein is fused to your protein of interest, and the rest of it is expressed on a separate plasmid,” he says. “Using only a small fragment makes life much easier for the CRISPR-Cas endogenous tagging approach,” since it can be hard to insert large genes using CRISPR.

In a recent preprint, a German lab used precisely this approach to “tag more than 1,300 endogenous genes in human cell lines with a fragment of the fluorescent protein mNeonGreen,” says Linke-Winnebeck, and then split open the cells and ran the proteins through a special gel containing high-affinity antibody fragments, or nanobodies, for mNeonGreen, sold by ChromoTek as MNeonGreen-Trap. “This allowed the visualization and in-depth analysis of the 1,300 proteins in their natural setting.”

To track PPIs in real time, and collect data that extends beyond a temporal snapshot, scientists have turned to methods besides q-AP-MS. One, called BRET (bioluminescence resonance energy transfer), uses light to detect interacting proteins. One protein is fused to a luciferase (which acts like an electron donor), while the second protein is fused to a fluorescent protein, which serves as an acceptor. When the two come near one another, electrons are transferred from the donor to the acceptor, which fluoresces.⁴ This experiment can even be parallelized in 384-well plates, and fluorescence can be measured over about two hours.

“This enables real-time, kinetic analysis of PPIs, allowing for a more thorough understanding of PPI dynamics and how this may differ, for example, across compound treatments or between related proteins,” says Danette Daniels, Senior Research Scientist at Promega, which sells a NanoBRET^TM system. Many proteins can also be tagged in parallel to study multiple PPI networks at once.

“These PPI methods are commonly used in screening efforts to identify new compounds that modulate target PPIs and can easily be adapted to better understand how clinically relevant mutations impact cellular PPIs and their response to potential therapeutic intervention,” says Daniels. And “because these methods allow for protein interaction dynamics to be assayed in the native cellular environment, they are often used to confirm that findings from biochemical PPI studies”—such as q-AP-MS, which requires that proteins be removed from cells—“are reflective of what actually occurs within the cell.”

To measure transient, or weak, protein interactions, scientists can use surface plasmon resonance, or SPR, a label-free, optical-based biophysical approach in which a small amount of ligand is immobilized on a surface, and then a binding partner, such as a protein, is introduced via a microfluidic flow system.⁵ As the two bind, changes in reflected light intensity (due to changes in refractive index very close to the sensor surface) can be used to measure the proportion of molecules that bind together. By measuring over time, these data are used to calculate binding affinities and association/dissociation kinetics. Through improvements in sensitivity, automation, and parallel fluidics, modern SPR instruments can routinely screen or characterize 100s to 1000s of interactions a day.

“This information-rich technique allows us to obtain several types of information including affinity (how tight an interaction is or how strongly a therapeutic binds its target), kinetics (how fast an interaction is), concentration (how much of the functionally active product is present in the sample) and how specific that interaction is,” says Paul Belcher, Product Strategy Manager for Cytiva’s SPR products, called Biacore.

The technology is especially useful for drug discovery, says Belcher, because large libraries of antibodies can be screened in parallel to see how they bind to proteins. Once a drug is identified, SPR can also be used to track and measure its affinity and/or kinetics during maturation in exquisite detail.

Measuring PPIs for clinical discovery

These three methods—q-AP-MS, BRET, and SPR—are, together, mapping and measuring and dissecting PPIs in impressive detail. They have also ushered in a veritable revolution in drug discovery and basic as well as clinical research.

In one study, researchers at Scripps Research used tandem mass tags (chemical labels for more quantitative, and parallelized, mass spectrometry) and AP-MS to study how a transcription factor called ATF6 controls the endoplasmic reticulum and ‘orders’ it to decrease the amount of an amyloidogenic protein, called ALLC, produced within the cell. The finding bolsters our understanding of systemic amyloid diseases such as transthyretin-related amyloid disease.⁶

Another study used AP-MS to study more than 7,000 proteins at once. Researchers took HEK293T kidney cells, transfected them with 2,594 different ‘baits’ and used mass spectrometry to reveal 23,744 interactions across 7,668 proteins.⁷ Other groups have extended this work further; earlier this year, a team at the University of British Columbia mapped more than 125,000 PPIs in seven different mouse tissues, including heart and brain. Most of the interactions had not been previously reported.⁸ They used a technique somewhat similar to q-AP-MS, called protein correlation profiling-stable isotope labeling of mammals.

NanoBRET, too, has proven key to understanding how cancer-linked oncogenes tweak protein networks as a cancer develops. The technology was used to demonstrate that the NRAS mutant showed greater interaction affinity with the RAS effector protein RAF1 when compared to wild-type RAS protein.⁹Researchers confirmed, without splitting open cells, that the mutation changed PPI networks that are important for regulating cancer. The paper has “helped to define the molecular mechanisms underlying a case of aggressive cancer, providing another step forward in better predicting the biological behavior of RAS mutations,” Daniels says.

And when it comes to the story of our times, COVID-19, SPR has proven critical to measure the binding affinities and dynamics of how candidate drugs interact with the virus. Last year, for instance, Regeneron Pharmaceuticals published binding kinetics for two antibodies (casirivimab and imdevimab) against SARS-CoV-2¹⁰; they are part of a monoclonal therapy recommended by the WHO, according to Belcher. In that study, “Biacore SPR systems were used to determine binding kinetics and affinities for anti-spike mAbs binding to the SARS COV2 spike protein.”