Back

A shark variable new antigen receptor recognizes an occluded epitope of fibroblast activation protein

Broadberry, R.; Ott, K.; Lake, E.; Ravi, T.; Hemme, C.; Chi, S.; West, J.; Gunaratne, G.; Ong, I. M.; LeBeau, A.; Grant, T.

2026-08-12 biochemistry
10.64898/2026.08.11.744296 bioRxiv
Show abstract

Variable new antigen receptors (VNARs) are the smallest naturally occurring antibody binding domains. Their size allows VNARs to access sterically restricted epitopes that are inaccessible to conventional antibodies. We recently identified a suite of VNARs that target fibroblast activation protein (FAP), a stromal serine protease indicative of extracellular matrix remodeling. The presence of FAP on the surface of cancer-associated fibroblasts (CAFs) that promote immunosuppression has made FAP a compelling therapeutic target for cancer therapy. Although antibodies targeting FAP have been developed, there is a paucity of information on how biologics engage FAP. Here, we used single-particle cryogenic electron microscopy (cryo-EM) to compare FAP recognition of three antibody architectures: a shark-derived VNAR, variable heavy (VH) and light domains (VL) of a humanized Immunoglobulin G (IgG), and a camelid-derived VHH. The humanized VH-VL domains and camelid VHH both target a solvent-exposed {beta}-propeller domain, whereas the VNAR binds a highly conserved, topologically recessed epitope at the FAP dimer interface. Radical-footprinting mass spectrometry (MS) further mapped two additional immune-derived VNARs to distinct FAP surfaces outside the shared {beta}-propeller epitope. These findings demonstrate how unique VNAR architecture can expand access to underexplored FAP surfaces and establish a structural framework for rational multiepitope targeting strategies.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.