A comprehensive understanding of ligand-receptor interactions would be tremendously useful in many areas of basic research and drug discovery, but identifying these interactions remains challenging because of the well-known difficulties of working with membrane proteins. In this issue, Wollscheid and colleagues1 describe ligand-based receptor capture (LRC), a new method for discovering the cell-surface targets of known ligands. They show the utility of LRC in several different and important contexts—such as receptors used by viruses to enter cells—suggesting that this is a truly general strategy for solving the problem of receptor identification.

The most common approaches for analyzing endogenous ligand-receptor interactions are affinity purification2 and screening3. The galanin receptor, for example, was purified from tissue lysate using affinity chromatography2. Of course these methods work best when the binding between the molecules is extremely strong (Kd in the μM range). But methods based solely on affinity purification are less effective when the binding is weak because the complex is more prone to dissociate. Receptors have also been identified by coupling receptor binding and activation to a cell-based readout that can be used to screen for ligands. This approach was successful in identifying the ligand of an orphan G protein–coupled receptor (GPCR) and in validating the receptor of the psychoactive drug salvinorin A (ref. 3). Although all of the above-mentioned strategies are powerful and have proven their value in defining ligand-receptor interactions, they are designed for receptors in classic signaling pathways (e.g., GPCRs and receptor tyrosine kinases) and would miss other interactions, such as the binding of the HIV envelope protein GP120 to CD4 (ref. 4). Thus, there is a need for a completely general, unbiased approach for identifying ligand-receptor interactions, which prompted the development of LRC by Wollscheid and colleagues1.

The field of chemoproteomics began with the use of affinity enrichment methods to identify targets of bioactive natural products. For example, the chemical synthesis and modification of FK506, a known immunosuppressive drug, enabled the discovery of the FK506-binding protein through affinity purification of the target, which revealed the mechanism of action of this drug5. The past decade has seen the emergence of a wide array of chemoproteomics methods that use chemical probes to profile active enzymes6,7, to identify sites of protein post-translational modifications8 and to image proteins in cells and tissues9.

Among these studies was a previous report by Wollscheid and collaborators that provided a method for cataloging cell-surface proteins that are glycosylated10. Aldehyde and ketone functional groups are generally absent from the surface of cells but can be generated on glycoproteins by mild oxidation. The authors treated cells with sodium metaperiodate to produce aldehydes, captured labeled proteins with a biotin probe and identified the enriched proteins by mass spectrometry. A fraction of the proteins identified were glycosylated cell-surface protein receptors, validating the approach.

In their new study, Wollscheid and colleagues1 build on this previous work to provide a method for identifying specific interactions between known ligands and unknown cell-surface receptors. LRC relies on a novel trifunctional chemoproteomics probe called TRICEPS (Fig. 1). Cells are treated with sodium metaperiodate and the TRICEPS probe conjugated to a ligand of interest, the proteome is trypsin digested, and TRICEPS-modified peptides are captured with streptavidin-coated beads. The peptides are eluted from the beads using the glycosidase PNGaseF to selectively release glycosylated peptides, which are then compared to negative control samples with quantitative proteomics. Enrichment of a receptor in the presence of a ligand, relative to the control, indicates an interaction.

Figure 1: LRC relies on a novel chemical probe TRICEPS, which contains a biotin for peptide enrichment, a hydrazine for covalent crosslinking and an N-hydroxysuccinimidyl ester for ligand attachment.
figure 1

A TRICEPS-modified ligand (TRICEPS + ligand) is added to cells treated with the mild oxidant sodium metaperiodate, which generates aldehydes (shown in red) from carbohydrates linked to cell-surface proteins. After binding of the ligand to the receptor, the hydrazine group forms a covalent bond with the aldehyde to label the protein. Subsequent enrichment and quantitative mass spectrometry analysis identifies peptides belonging to the ligand's receptor as being enriched (shown in blue) compared with the control sample (no ligand). Wollscheid and colleagues1 use LRC to identify known and novel ligand-receptor interactions.

Importantly, the authors show that TRICEPS labels a wide variety of ligands; the only requirement is that the ligand contain free amino groups. Ligand-receptor binding increases the local concentration of the probe's hydrazine group around oxidized sites on the receptor, enhancing the reaction rate. Thus, unlike the earlier method10, which labels all cell-surface glycoproteins, LRC primarily labels only glycoprotein receptors that bind the ligand.

To validate LRC and demonstrate its scope, Wollscheid and colleagues1 study some well-characterized receptor ligands, including insulin, transferrin, apelin-17, trastuzumab (Herceptin), epidermal growth factor (EGF) and designed ankyrin repeat proteins (DARPins) that target ErbB2. They find that each ligand selectively enriches only its known receptor, with very little or no off-target enrichment. Importantly, the authors also show in the case of DARPins that the ErbB2 peptides identified by LRC can be used to pinpoint specific DARPin-binding domains in ErbB2. In sum, LRC can be applied to many types of protein and peptide ligands, can identify diverse classes of protein-receptor interactions and can be used to map binding interfaces. The number of different ligands analyzed provides confidence that it is a robust approach.

After testing the method on known ligand-receptor interactions, the authors attempt to discover novel interactions, choosing the example of vaccinia virus infection of HeLa cells. Intact, mature vaccinia virions are labeled with TRICEPS and applied to HeLa cells. The experiment yields seven virus-binding proteins, four of which are novel. Knocking down five of these seven receptors individually sharply reduces vaccinia virus infectivity. Thus, LRC can discover new ligand-receptor interactions even in complex cases that involve multiple receptors.

Wollscheid and colleagues1 have shown that LRC works with both protein and peptide ligands, but an interesting avenue for future research would be to test the method with small-molecule ligands. Another important direction is in vivo applications. Although the requirement for mild oxidation limits LRC to cell or tissue culture, a new generation of TRICEPS probes that are safe and effective in animals is conceivable. The development of a ligand that can covalently capture glycosylated cell-surface receptors provides an elegant example of the power of chemoproteomics in enabling biological discovery. Given that most cell-surface proteins are glycosylated and that the authors have previously shown that this chemistry can survey all major classes of surface proteins10, LRC is general and likely to find utility in basic research on receptor biology and in pharmaceutical development.