NEW YORK (GenomeWeb) – A team led by researchers at Sanford Burnham Prebys Medical Discovery Institute has developed a platform for proteome-wide analysis of protein degradation.
In a paper published this week in Cell Reports, the researchers used the system to identify host protein targets of the HIV-1 accessory protein Vpu. According to Sumit Chanda, a Sanford Burnham professor and senior author on the study, it could be useful for studying protein interactions more generally as well as in drug development.
Called the Global Arrayed Protein Stability Analysis (GAPSA) platform, the system uses a cDNA matrix to express arrays of proteins which can then be screened against a protein of interest or other stimuli. The proteins are then monitored for changes in abundance reflecting destabilization due to exposure to the agents being tested.
It is well known that protein degradation is an important aspect of cellular regulation, Chanda said. "If you get rid of [a protein] then it can't function, and that changes the phenotype of the cell. And it's known that certain proteins can be degraded by other proteins."
However, he added, what proteins degrade what other proteins has not been explored in a systematic way.
For instance, "there are about 500 E3 ligases [which are involved in protein degradation], and they all have their own special targets," he said. "What we wanted to figure out is what is a platform where we can systematically ask across the entire human genome, what set of proteins can be degraded by X, or the other way around — what proteins degrade my protein of interest."
Chanda noted that while mass spectrometry has similarly been used to look at protein degradation in various biological systems, mass spec approaches suffer from their bias towards high-abundance proteins and interactions.
With the GAPSA platform, on the other hand, "we are actually artificially putting together these binary pairs, so we don't have to rely on endogenous expression levels [being detectable]," Chanda said. "That's one major advantage when you go from that kind of pooled format where you're looking at everything and you're trying to pick that needle out of the haystack. Essentially here, we're spreading out the haystack, which makes it a lot easier to find those needles."
Of course, the artificial nature of the experiment raises the possibility of identifying interactions that don't actually occur in vivo, and Chanda said, confirming interactions as real requires additional knockout experiments using techniques like CRISPR to query whether they occur endogenously.
In the Cell Reports work, the researchers found a false positive rate of around 20 percent for the interactions they validated, which Chanda called "pretty respectable for a genome-wide platform."
That said, the GAPSA analysis "is the beginning of a study," he noted. "It gives you an overview of the propensity of a given molecule to degrade. Then you have to go back and see in the biological context [you are interested in], how is this working? Can we knock it out in an animal or in a cell and show that you enhance the stability or decrease the stability of the protein?"
In the HIV-1 work, Chanda and his colleagues co-expressed Vpu with 433 antiviral interferon-stimulated genes to look at how the protein helps HIV-1 evade the host immune system. They identified a number of known Vpu degradation targets as well as several new targets that Vpu degrades to facilitate viral replication.
HIV research is an area of focus for his lab, Chanda said, but noted that the Cell Reports paper was mainly a proof-of-concept. "There was existing biology that showed that this protein does indeed target proteins for degradation, so if we could find additional proteins that it targets, then that would give us proof of concept for the platform," he said.
Ultimately, though, Chanda said he is most interested in using the platform for drug development work.
"Our particular interest in to find therapeutically relevant interactions and then to show that using a small molecule we can modulation these interactions in a relevant disease system," he said.
Specifically, he hopes to use the platform in the development of bi-functional small molecule drugs that use proteins like E3 ligases to degrade otherwise undruggable protein targets like the cancer-associated protein Ras.
In the case of these targets "there aren't good binding pockets to block their enzymatic activity," Chanda said, which has stymied efforts to develop small molecule inhibitors. "But there are good binding pockets for non-functional binders," he said. And this raises the possibility of using a bi-functional molecule that could bind a protein like Ras and an E3 ligase known to degrade it.
"My hypothesis is that there are small molecule binding pockets [on Ras and similarly undruggable proteins], they just won't inactivate the protein," he said. "But that's okay, because we're recruiting a second protein to do that for us. Essentially what we're hoping to do is convert these non-druggable targets into druggable targets."
To do that, drug developers need to know what proteins will degrade what other protein, which, Chanda said, has not been studied in a systematic, high-throughput manner.
"We're right now really relying on kind of serendipitous, one-gene-at-a-time discovery models, where, okay, we know this E3 ligase targets this protein because some grad student spent five years trying to figure that out," he said. With the GAPSA platform, Chanda and his colleagues can take a target and "in the course of 48 hours tell you what E3 ligases target it, so you can go and start building small molecules for these therapies."
Chanda said that the platform is fully automated and can screen entire genomes in around a day. The goal, he said, is to build up a large collection of data from multiple screens that will allow researchers to both look up interactions they are interested in without having to do a new screen and better understand the structure of protein degradation interactions.
Such a database would let researchers "do meta-analyses to see if we can really understand what the logic is behind these kinds of proteostasis events," he said. "Right now it's thought to be fairly random what protein picks what protein for degradation. Nothing is really random in the cell, but no one has really been looking at this from a 20,000-foot view. This gives us an opportunity to take that snapshot and really start pulling out higher-order structure in what are seemingly kind of serendipitous events."
The researchers have IP protection on aspects of the platform, but Chanda said he does not have any plans to spin it out into a commercial entity.
"This is, I think, a platform that will find you discoveries that can then be exploited for commercialization, but we don't intend to commercialize this platform," he said. "We want to make this platform open access. We want to build it up and work with as many people who are interested in working to identify new biology, and then that can be exploited for therapeutic and commercial value."