G-protein-coupled receptors (GPCRs) are the largest and most versatile of cell surface receptors, with a broad repertoire of ligands and functions. Around one-third of approved medicines act on these membrane proteins, targeting a wide range of diseases from allergic rhinitis to pain, hypertension, and schizophrenia.
The human genome project has identified approximately 720 GPCR genes, but the majority of GPCR drugs in current clinical practice only target around 120 of them, leaving the majority of the GPCRome therapeutically unexplored.
This is largely due to a combination of historical, technical, and biological challenges. Many GPCRs remain poorly characterized, with unknown endogenous ligands, unclear physiological roles, or expression patterns that complicate target validation. Others have been difficult to study using traditional screening approaches because of issues such as low expression, poor solubility, or the lack of robust assays.
In a recent paper published in Nature, researchers describe a fundamentally new approach to overcoming many of these longstanding barriers. The team showed that it is possible to computationally design compact, highly stable miniproteins (proteins with fewer than 100 amino acids) from scratch that can directly engage GPCR binding pockets and modulate receptor activity. Using this strategy, they generated functional miniproteins against 11 GPCRs spanning multiple receptor classes implicated in itch and pain, cancer, metabolic disease, and migraine.
“Creating agonistic biologics against GPCRs has been near-impossible,” Christoffer Norn, CEO and cofounder of Skape Bio, told DDN. “We think this is a way to do it.”
Why GPCR agonists are so hard
Although small molecules remain the major GPCR drug type, biologic therapeutics such as antibodies, nanobodies, and peptides are steadily gaining ground due to their superior specificity and versatility.
However, designing biologics that actively modulate GPCR signaling has remained exceptionally difficult. These receptors are complex, flexible membrane proteins which constantly shift between multiple active and inactive states, making it hard to design molecules that bind predictably. Additionally, short extracellular loops limit accessibility and reduce the effectiveness of conventional screening and structural approaches.
“Even when designs are highly specific to a given receptor, tiny structural differences — sometimes on the scale of a single atom — can determine whether a receptor is active or inactive,” explained Norn. As a result, many designs typically need to be tested to identify those that truly function.
Compounding the problem is how GPCRs are typically studied. Many screening approaches require receptors to be removed from their native membrane environment, either by solubilization or by mutating membrane-facing residues to improve stability. Both strategies risk distorting the receptor’s natural conformation, making it difficult to identify biologics that will function correctly in cells.
Screening GPCRs in their native environment
In this study, Skape Bio and researchers from the Institute for Protein Design at the University of Washington tackled both problems simultaneously. They used AI-driven computational design to generate novel miniproteins capable of activating or blocking GPCRs, then functionally screened millions of candidates directly in cell-based assays to identify designs that modulate signaling in a native-like context.
Because the proteins are designed de novo, the researchers have complete control over both structure and sequence. “We start with a structure of the target. When we pick that structure, we’re already thinking about what the desired product profile is,” said Norn. “If we want an agonist, we start from an active-state structure — and make sure we’re targeting a region where binding locks the receptor into the active conformation. That way, we bias the GPCR toward the state we actually want.”
Once they have the structural backbones, the researchers use AI tools to design the sequences. “We then predict which binders are most likely to work using AlphaFold-type methods, and we layer in additional biophysical constraints to further boost the hit rate. That leaves us with a library of up to 100,000 candidate designs,” said Norn.
Those designs then go into a high-throughput, cell-based screen. Each cell produces just one candidate binder alongside the same GPCR target, using fluorescent proteins to create a direct visual readout of binding. When a design binds its GPCR target, the receptor is retained in the endoplasmic reticulum (ER), and the cell appears yellow under the microscope. Designs that fail to bind allow the receptor to traffic to the cell surface, producing a green signal instead.
“By retaining GPCRs within the ER and detecting binding-induced trafficking changes using microscopy, we can screen millions of cells and efficiently identify the designs that genuinely work,” described Norn.
From target selection to hit identification, the entire workflow can take as little as three months for a new GPCR. In one example from the study, a designed miniprotein antagonist matched the efficacy of an approved clinical drug in mobilizing hematopoietic stem and progenitor cells, while showing a reduced side-effect profile, including reduced leukocyte mobilization and no detectable systemic cytokine elevation.
Engineering drug-like properties
One of the key advantages of miniproteins is their small size, which allows them to slip into narrow binding pockets that are often inaccessible to antibodies and other larger biologics. This compact form factor enables high shape complementarity while maintaining strong and selective receptor engagement.
Miniproteins are also highly customizable. Standard protein engineering strategies — such as PEGylation, fusion to Fc domains, or attachment to albumin-binding motifs — can be readily applied to tune their pharmacological properties, including half-life and tissue exposure.
In the study, the team modified a candidate miniprotein to investigate the pharmacokinetics of unmodified and Fc-fused forms. When administered intravenously in mice, the unmodified miniprotein was cleared from the bloodstream in less than an hour. By contrast, fusion to an Fc domain extended its circulation time to roughly 25 hours.
“These are modular proteins,” Norn emphasized. “If you want a long half-life, you can engineer that in.”
Building a GPCR discovery platform
By combining de novo protein design with native, cell-based screening, Skape Bio is beginning to turn one of the most complex target classes in biology into a more systematic and programmable space. Miniproteins provide a rare combination of precision, flexibility, and speed — small enough to reach buried binding pockets, yet robust enough to be engineered into drug-like molecules.
The immediate implications are therapeutic, opening new routes to GPCR targets that have long resisted conventional approaches. But the longer-term impact may be even broader. As the platform scales, it offers a path toward mapping functional interactions across much of the GPCRome, generating not just drug candidates but a deeper understanding of receptor biology itself.
