Researchers at the Wellcome Trust Sanger Institute have devised a method they say enables the large-scale detection of low-affinity extracellular protein interactions, about which little is currently known.
The method, which they call AVEXIS for avidity-based extracellular interaction screen, is a high-throughput assay that its creators say overcomes the technical difficulties that have to date made it exceedingly difficult, if not impossible, to detect extracellular protein-protein interactions.
The authors say the method showed for the first time that a membrane-tethered receptor-ligand pair could mediate the initial contact between motor neuron axons and muscles to form the neuromuscular junction.
In their study
describing the method in the Feb. 22 online edition of Genome Research
, the authors say that though extracellular protein interactions such as those made by secreted and membrane-tethered proteins are crucial for “diverse cellular behaviors,” and about 20 percent of human genes encode such proteins, protein interaction datasets are bereft of them, “making current interaction networks biased and incomplete.”
Several reasons account for this, they say, including the difficulty with which cell surface proteins can be biochemically manipulated and the fact that functionally important post-translation modifications are not usually added in “commonly used expression systems such as bacteria and cell-free systems.”
Also, cell-surface protein interactions are highly transient, with half-lives lasting fractions of a second, making purification protocols involving wash steps “impractical,” the authors write. As a result, the most common methods for high-throughput protein-protein interaction detection, yeast 2-hybrid and biochemical purification, have been rendered “unsuitable” as identification tools for extracellular protein interactions.
“And because these proteins are a major part of the human genome in terms of the protein-coding genes — I reckon there’s about two-and-a-half thousand of them — we were missing a whole chunk of protein interactions because the techniques we were using just weren’t appropriate to detect them,” Gavin Wright, a junior investigator at the UK-based institute who directed the research, told ProteoMonitor.
Such interactions, he and his fellow authors say, have particular importance in the development of new drugs, which often target cell surface proteins. Most data on protein-protein interactions are about interactions within the cell, which have limited applications for new-drug discovery.
“Since extracellular proteins are readily accessible to systemically delivered drugs, AVEXIS can therefore be used to identify novel therapeutic opportunities to target both genetic and infectious diseases,” they say.
To create AVEXIS, the authors first produced the ectodomains of cell-surface proteins in rat cells as soluble recombinant proteins in order to remove the insoluble transmembrane region of the proteins while keeping their extracellular binding function.
The ectodomains were then expressed in two forms, a monomeric biotinylated “bait,” which could be captured on streptavidin-coated microtiter plates, and a pentamerized “prey” tagged with beta-lactamase to allow detection.
The prey pentamers were made by C-terminally tagging proteins with “a coiled-coil sequence from the rat cartilage oligomeric matrix protein,” the authors say. While the assay had low monomeric interaction affinity, interactions between the rat protein Cd200 and its receptor were “robustly detected [regardless] of bait-prey orientation.”
Further, by increasing the avidity of prey proteins through pentamerization, the sensitivity of detection over monomeric prey proteins could be increased by at least 250-fold.
“Since extracellular proteins are readily accessible to systemically delivered drugs, AVEXIS can therefore be used to identify novel therapeutic opportunities to target both genetic and infectious diseases.”
“Interaction specificity was demonstrated by preventing interactions in either orientation with blocking anti-bait monoclonal antibodies,” the authors say, and activities of both the bait and prey was normalized before being used in the assay though purification was not necessary for either.
To determine false-positive and false-negative rates, the team then performed a small screen using eight proteins from the SLAM/CD2 subfamily of human immunoglobulin superfamily “within which there are quantified positive hetero- and homophilic low-affinity interactions and … importantly, published negative interactions,” they say in the study.
They selected a prey activity threshold to detect the 8 micromolar CD244-CD48 interaction in both bait and prey orientations and confirmed all published negative interactions. At this prey activity, the weaker 9-22 micromolar CD2-CD58 interaction and all homophilic interactions could not be detected. At a higher prey activity all expected interactions except the SLAMF6 homophilic were detected, although the false-positive rate increased.
To eliminate false positives during screening, they selected a “stringent” prey threshold, which increased the false negative rate as a result. Overall, 0.28 percent of the interactions that were screened were positive, and because his group is confident that most of those interactions are true positives, the false-positive rate is “probably much lower than that,” Wright said.
He could not put a figure on the false-negative rate because so little information exists about interactions with zebrafish proteins, which the researchers used to test AVEXIS.
To identify novel extracellular receptor-ligand pairs, Wright and his colleagues expressed a protein library containing the entire ectodomains of 110 proteins mainly from the zebrafish immunoglobulin superfamily, in mammalian cells as both prey and bait molecules. They also normalized their activities, a step he said contributed to the low false-positive rate.
The group performed 9,751 interaction tests, of which 7,292 were screened in a reciprocal manner. In total, 6,105 unique possible interactions were tested. Of that, they identified 17 interactions between 19 proteins, of which 16 were cell surface proteins and three were secreted proteins.
Novel interactions that were identified included proteins for which extracellular ligands are not known, including a subnetwork involving the Mpzl2 protein, a gene expressed during mouse thymus development. They also identified new ligands for well-characterized proteins: for example, they found that neural cell adhesion molecule interacted with muscle-specific kinase receptor.
While functional data exists suggesting a correlation between the two proteins for correct neuromuscular junction formation and maintenance, the authors say their study is the first to show that a membrane-tethered receptor-ligand pair “could mediate the initial contact between motor neuron axons and muscles to form [the neuromuscular junction].”
To validate their results and determine the affinity detection threshold of their method, they used surface plasmon resonance to measure “off-rate constants, which have the advantage of being independent of active protein concentration estimates,” in 11 of the 17 interactions they identified. Their original findings were confirmed, they report, supporting a low false-positive rate.
Four of the interactions had half-lives at or below the limit of SPR sensitivity of one-tenth of a second, “demonstrating that AVEXIS can detect extremely low-affinity interactions,” the authors say.
In an e-mail, Ola Söderberg, a researcher at Uppsala University who was part of a team that developed an in situ application for proximity ligation assays for measuring protein-protein interactions, generally praised the study. He said, however, that AVEXIS’s ability to detect low-affinity and transient interactions will need to be further validated by other methods before its utility can be assessed.
“As there is always an equilibrium between the concentration of interacting proteins and free ones, higher concentrations will give more interactions,” Söderberg said. “Do the proteins interact in physiological concentrations in vivo? How will weak/transient interactions give a biological effect?
“So I think this method is useful to identify possible interaction partners, but the interactions will have to be tested with other methods [such as bimolecular fluorescence complementation or in situ PLA] to see if they have any biological relevance,” he said.
In follow-up work, Wright said, he and his colleagues have done additional screening with AVEXIS and expanded their research to include leucine-rich repeat proteins. They are preparing a paper for publication on the project, but Wright said that they are trying to create a network between cell surface receptor proteins expressed on neurons.
“The idea is that we’re trying to determine a neuron receptor code … with the idea that we can understand how the nervous system wires up correctly,” Wright said.
They would also like to miniaturize the method onto a microarray though they have not started work on that yet, Wright said.
He and his colleagues say in their Genome Research article that AVEXIS may not be useful for detecting homophilic extracellular interactions “possibly due to highly avid prey-prey association, which may then prevent prey-bait interactions,” although they reproducibly detected some homophilic interactions in their large screen.
In addition, Wright said, the method may have trouble detecting 2-chain co-receptors.
The issue, he said, is “the fact that you would have to express both proteins and get them tagged together in a way where they would remain functional in terms of their binding properties.”
Nevertheless, he and his fellow authors say that because AVEXIS can screen for extracellular protein interactions “from most metazoans,” is suitable for large-scale screens, has a low false-positive rate, and can detect “extremely low-affinity interactions, the method is a “versatile technique to detect and study low-affinity extracellular interactions in a high-throughput format.”