On the dance floor that is the living cell, proteins execute all sorts of moves, facilitating processes such as transport, folding, and signaling. Many of the most interesting and consequential moves occur when proteins interact with other proteins—whether they stick with one partner or coordinate their steps with a sequence of partners. To appreciate the cell’s choreography as well as the roles played by individual proteins, scientists try to keep up with all the moves, that is, all the protein–protein interactions (PPIs). But doing so can be a challenge. The cell’s choreography is more complicated than two-step line dancing.
For decades, scientists interested in PPIs have relied on screening technologies such as yeast two-hybrid systems. These systems measure PPIs one at a time, that is, one two-partner interaction at a time, where a “bait” protein pairs with (or binds to) a “prey” protein. Other commonly used methods include co-immunoprecipitation, pull-down assays, and cross-linking.
All these methods share the same limitation: By taking in just a pair of dancers at a time, they treat all the other proteins in the dance as if they were wallflowers. Fortunately, new methods are being developed that capture wider views, tracking multiple pairs over time.
These methods owe much to researchers who have added to our knowledge of the “interactome,” which refers to the full set of molecular interactions within a cell. By mapping the interactome, researchers make it easier to put individual PPIs in context.
Contextualizing technologies include luminescence-based mammalian interactome mapping (LUMIER) and mammalian protein–protein interaction trap (MAPPIT) assays. Besides obtaining comprehensive views of PPIs, these technologies allow complex, overlapping pathways to be discerned.
The knowledge surrounding PPIs is steadily increasing in scale. For example, a map of ~14,000 human binary PPIs was published in Cell in 2014. Now, researchers are expanding the role of PPIs in drug discovery. Like choreographers, the researchers who use PPIs in drug discovery develop themes and variations and even allow a degree of improvisation.
Taking in large ensembles
A Seattle start-up company called A-Alpha Bio is applying genetic engineering and next-generation sequencing to one of the oldest systems in PPI research—the yeast two-hybrid system—to analyze millions of PPIs simultaneously. Co-founded in 2017 by David Younger, PhD, and Randolph Lopez, PhD, the company has developed a platform that could disrupt drug discovery.
The platform, which is the heart of A-Alpha Bio, grew out of work Younger carried out while earning his PhD at the University of Washington, in a laboratory led by David Baker, PhD, professor of biochemistry and director of the Institute for Protein Design. Younger, now CEO of A-Alpha Bio, asserts that AlphaSeq has become a “multiplexed, quantitative, and versatile platform for characterizing entire protein interaction networks in a single test tube.” The platform is built upon the connection between protein interaction strength and yeast mating.
The technology behind AlphaSeq was introduced in 2017, in a paper that was published in the Proceedings of the National Academy of Sciences. In this paper, which was entitled “High-throughput characterization of protein-protein interactions by reprogramming yeast mating,” Younter et al. described how to reprogram yeast mating to simultaneously characterize thousands of PPIs. The article’s authors also explained how the AlphaSeq platform could be used to observe changes in interaction strengths in a changing extracellular environment.
AlphaSeq starts by building libraries of wild-type Sacchromyces cerevisiae that are “neutered,” meaning that the yeast cells cannot bind together, as they normally would, to mate and form a diploid cell. Then, A-Alpha Bio displays proteins on the outside of the neutered yeast. When the cells are mixed together and grown in a test tube, PPIs result in the formation of diploid cells. The number of diploid cells that are formed can be counted by next-generation sequencing, giving a readout of the strengths of the interactions.
The most exciting aspect of AlphaSeq, according to Younger, is “the unprecedented amount of data that we can generate—binding strengths for millions of PPIs simultaneously from a single experiment.”
A-Alpha Bio’s approach, according to Younger, addresses an existing bottleneck in protein binder design. Younger tells GEN that, unlike other genetically encoded technologies for measuring PPIs for biologics discovery, optimization, and characterization—such as yeast surface display, phage display, and mammalian display—AlphaSeq “enables the characterization of protein interaction networks at a library-on-library scale.” In other words, AlphaSeq isn’t confined to mapping a large library of biologics candidates against a single target. Instead, AlphaSeq can enable the characterization of a biologics library against many targets—tens, hundreds, even thousands—simultaneously.
Optimizing new applications
A-Alpha Bio’s current focuses are in oncology and infectious disease. Younger tells GEN that “infectious disease provides a very clear articulation of the value of AlphaSeq.” The challenge with developing infectious disease therapeutics, he explains, is the broad diversity of surface proteins that microbes express, which allow them to evade the immune system. A therapeutic designed against one of these protein variants has several limitations. For example, the therapeutic would be effective only for a limited subset of patients affected by a particular strain. Also, resistance would be a large consideration.
When broadening a drug’s profile, developers may follow the traditional route. They start with optimization against a single target, and then they conduct additional tests against other variants, optimizing for each new target without reducing binding to the original target. This iterative process is very slow and expensive, and it often fails to converge on a final drug that has an adequately broad binding profile. With AlphaSeq, Younger asserts, “We can optimize candidate drugs against many different versions of a pathogen target simultaneously with our library-on-library capabilities.”
Although young and small, A-Alpha Bio has recently moved into Fluke Hall—an incubator space designed to foster and commercialize the University of Washington’s spin-outs. Also, the company recently closed out a $2.8 million seed round.
In a press release about the company’s recent funding success, Younger’s former PhD advisor David Baker, a co-developer of AlphaSeq and a scientific advisor to A-Alpha Bio, explained that although there are many methods available for screening large biomolecular libraries for a particular binding activity, there are “few approaches for assessing in parallel the very large number of possible interactions between biomolecules in two large libraries.” Baker continued that A-Alpha Bio’s technology “provides a way to not only quantitatively measure the interactions between all pairs of molecules in two libraries, but also the effect of small molecules and other perturbations on these interactions.”