Scientists from the UK’s Wellcome Trust Sanger Institute have developed a large-scale method for identifying the interactions between cellular proteins that are major drug-discovery targets.
The ELISA-based assay, which they call the avidity-based extracellular interaction screen, or AVEXIS, is a high-throughput assay that overcomes the biochemical intractability of membrane proteins and their interactions.
Cell-surface proteins contain insoluble, hydrophobic transmembrane regions, and their extracellular interactions are often very transient, with half-lives of less than a second. According to the researchers, the AVEXIS assay can detect these short-lived interactions with a low false-positive rate.
The scientists used the AVEXIS assay to systematically screen for receptor-ligand pairs within the zebrafish immunoglobin superfamily. They were able to identify novel ligands for both well-known and orphan receptors.
The investigators found that the genes encoding the receptor-ligand pairs were often phylogenetically clustered and expressed in the same or adjacent tissues.
This work was published online on Feb. 22 in the journal Genome Research.
Gavin Wright, a junior investigator at the Wellcome Trust Sanger Institute, spoke with CBA News this week about AVEXIS and its applicability to compound screening and drug discovery.
Can you provide a little background on what led to the development of the AVEXIS platform?
From a personal perspective, my background was in looking at the cell surface protein interactions on the surface of leukocytes. I got a little frustrated that I could not do any kind of functional analysis on some of the interactions that we were identifying. We were mainly using the rat as an experimental model.
I then went off and did a postdoc and learned how to use the zebrafish, which is genetically a much more tractable organism. The idea then was to set up the laboratory where I could put the two pieces together and do the detailed biochemistry, but also do functional analysis relatively easily using the zebrafish as a model organism.
Specifically for this research, the motivation was the fact that identifying protein-protein interactions are of fundamental importance, because the interactions between proteins underlie biology — if things go wrong, we get disease.
After the genome sequencing project, there was a huge number of these genes which had no known function, and we wanted to identify the interactions between these new proteins. That challenged scientists to come up with large-scale protein interaction technologies to identify the protein-protein interactions that, in a sense, underlie basic biology.
Two methods were very successful: the yeast-two-hybrid methodology and a biochemical purification followed by protein identification by mass spectrometry. Neither of these is suitable for looking at extracellular protein-protein interactions, however.
Yeast-two-hybrid is not suitable, because the interaction has to occur within the nucleus of the yeast, and that goes to transcriptional output. That is a reducing environment, so these proteins that occupy an extracellular oxidizing environment would not fold correctly. They would not contain the correct posttranslational modifications, such as sugars and disulfide bonds.
What is not widely appreciated is the fact that the interactions between these extracellular receptor proteins are generally extremely weak. They have half-lives of half a second or even less. A biochemical purification procedure where you must incorporate quite stringent wash steps for several hours means that these very transient interactions would simply get washed away. So we were not seeing them.
That means our current ability to detect protein-protein interactions was very biased toward things that would work in the yeast-two-hybrid assay or form very stable complexes using this biochemical purification procedure.
As it turns out, these receptor proteins comprise a large percentage of the genes encoded within the human genome — around 20 percent of all genes encode a protein that is associated with the cellular membrane.
Furthermore, many of these proteins are particularly good drug targets because they are accessible to systemically-delivered therapeutics. Until this research, there was no scalable technique that we could use to get at these extracellular low-affinity interactions.
That was the goal of this research.
How does the AVEXIS technology work?
It is essentially an ELISA-based assay. The trick that we used to get around this low-affinity problem was that we multimerized one form of the protein. We did this using a tag that forms spontaneous pentamers in solution. This increases the local concentration of the receptor, compared to the main fragments that we were using, so that we can increase the duration of the interaction just long enough to detect it after a brief wash.
We can then produce our proteins, the extracellular receptor proteins in mammalian cells, so that they contain all of the correct posttranslational modifications.
So essentially, we came up with an assay that can detect very low-affinity interactions and once we developed this assay, we needed to build a library of recombinant proteins containing the ectodomain fragments of receptor proteins. Then we systematically screened for interactions within that library.
How is this relevant to the use of cell-based assays in drug discovery?
You could add the multimerized proteins that we used in our screening procedure to cellular assays in 96-well plates, to see if you can ectopically activate the signaling pathways that would be of particular value to whatever you are screening for.
For example, if you are screening for signaling pathways that may induce cell death, you can add in these soluble multimerized fragments and see if you induce cell death in your cellular assays.
What is the next step in this research?
We have basically done this as a proof-of-principle study using zebrafish proteins, which has allowed us to do a distribution analysis that showed that the interactions we were identifying did have an in vivo function. I was anticipating a lot of skepticism from measuring these incredibly weak protein-protein interactions. Some of these interactions have half lives of one tenth of a second, and many people wondered how they could be relevant in vivo. We have tried to demonstrate that they are relevant by using the zebrafish as a model organism, and determining when and where in the animal the genes encoding these receptor proteins are actually expressed.
From there, we could take this work into humans by using human proteins. This would allow us to identify novel signaling pathways that are involved in our own physiology. The disruption of these pathways could lead to new therapeutic avenues.
These receptor proteins are very accessible to systemically delivered therapeutics, so they make tractable drug targets, particularly for biotherapeutics, such as recombinant proteins or antibodies.
Is the AVEXIS platform something that you may wish to commercialize?
No, we currently have no plans to do so.