De Novo Design of a Protein Binder to Probe Gas Channel and Enhance the Oxygen Tolerance of [NiFe]-Hydrogenase.

[NiFe]-Hydrogenases display remarkable catalytic efficiency for H2 production and oxidation but are sensitive to O2, which severely restricts their biotechnological applications. The identities and functional relevance of gas channels leading to the buried [NiFe] active site remain elusive, hindering rational enzyme engineering. Here, we introduce an artificial intelligence-guided de novo protein binder design strategy to map hydrophobic O2 diffusion pathways in Escherichia coli hydrogenase-2 (Hyd-2). By integrating RFdiffusion, ProteinMPNN, and AlphaFold, ∼100,000 candidate binders were computationally screened, yielding two high-affinity binders, L1 and L2. Biophysical and electrochemical analyses show that L1 selectively occludes the primary O2 ingress channel, enhancing the enzyme's oxygen tolerance by more than threefold, whereas L2, which targets another putative channel, has a negligible effect, indicating that this pathway contributes minimally to O2 ingress under the tested conditions. Furthermore, we determined the L1-Hyd-2 complex structure by cryo-electron microscopy and revealed the key interaction residues and interface conformation. Integrated structural and computational analyses provide mechanistic insights into the system's oxygen tolerance and high-affinity L1-Hyd-2 interaction. All findings provide the first direct experimental evidence for hierarchical O2 diffusion channels in [NiFe]-hydrogenases and establish binder-mediated channel occlusion as a generalizable, mutation-free strategy for elucidating and modulating gas transport in metalloenzymes.