In a landmark advancement for protein engineering and molecular therapeutics, researchers have unveiled BindCraft, a revolutionary pipeline capable of designing highly specific protein binders targeting traditionally “undruggable” interfaces on nucleic acid-guided multi-domain nucleases. This innovative approach not only challenges long-standing assumptions about the druggability of nucleic acid interaction sites but also opens new avenues for gene editing modulation and therapeutic intervention.
Nucleic acid interaction surfaces, characterized by expansive, highly charged, and often convex topographies, have long perplexed drug designers. Their intricate nature renders them resistant to targeting by conventional small molecules, which typically require well-defined binding pockets. The team focused their efforts on SpCas9, the Streptococcus pyogenes CRISPR-associated nuclease, a multi-domain enzyme central to the gene editing revolution. While SpCas9’s programmable nature has transformed genome engineering, it originally functions as a bacterial immune effector against viral invaders.
Interestingly, bacteriophages—the viruses that infect bacteria—have evolved small proteins known as anti-CRISPRs (Acrs) that inhibit the CRISPR-Cas system by blocking nucleic acid binding. Inspired by this natural defense, the researchers aimed to computationally design artificial binders that mimic these inhibitory mechanisms, thereby offering unprecedented control over CRISPR activity.
The design strategy targeted the bipartite REC1 domain of SpCas9, notable for its highly charged RNA-binding pocket essential for guide RNA interaction. Experimental validation revealed that all six designed binders effectively engaged the full-length apo (RNA-free) SpCas9 enzyme, with top candidates binder3 and binder10 displaying apparent dissociation constants in the submicromolar range, as determined by surface plasmon resonance measurements. These affinities, while modest, are significant given the challenging physicochemical nature of the target surface.
To elucidate the structural basis of these interactions, the research team attempted cryo-electron microscopy (cryo-EM) of the complexes. Despite high-quality datasets and distinct density for the binders, the dynamic conformations of the apo SpCas9 hindered achieving atomic-resolution models in the target region. Nevertheless, unambiguous density localized at the REC1 domain allowed confident docking of both binder3 and binder10, effectively validating the designed binding modes and confirming that these artificial proteins occupy the anticipated nucleic acid-binding pocket.
Functionally, the designed binders demonstrated potent inhibition of SpCas9-mediated gene editing in human embryonic kidney (HEK293T) cells. When co-transfected with CRISPR-Cas9 machinery, the synthetic binders curtailed gene editing efficiency significantly, exceeding the inhibitory capacity of natural anti-CRISPR AcrIIC2, which obstructs guide RNA loading through a distinct modality. While natural inhibitors such as AcrIIA2 and AcrIIA4 nearly abolished editing by directly obstructing target DNA binding, the artificial binders presented a novel mechanism of action, underscoring the versatility of BindCraft in crafting diverse protein interaction modalities.
Extending beyond Cas9, the team applied BindCraft to another formidable nucleic acid-guided enzyme, the Argonaute nuclease from Clostridium butyricum (CbAgo). Analogous to Cas9, CbAgo serves as a prokaryotic defense factor, employing small DNA guides to target complementary DNA sequences for cleavage. Notably, no natural protein inhibitors of Argonaute nucleases have been characterized to date, making CbAgo an ideal candidate to showcase the broad applicability of the design platform.
Targeting distinct functional domains within CbAgo, specifically the N-PIWI channel and the PAZ domain, the researchers created twelve binders and assessed their influence on CbAgo activity. Among these, two stand out by robustly inhibiting DNA cleavage, reducing catalytic turnover rates by 40- to 80-fold at nanomolar binder concentrations. Kinetic assays revealed a profound attenuation of enzyme function, with binder2 achieving a dissociation constant of approximately 5 nM, verified by bio-layer interferometry, and supported by size exclusion chromatography evidence indicating the formation of a stable CbAgo-binder complex.
Interestingly, the presence of guide DNA destabilized the CbAgo-binder2 complex, confirming that binder2 occupies the guide DNA binding channel, thereby interfering with essential nucleic acid engagement critical for cleavage activity. This mechanistic insight solidifies the notion that BindCraft-generated proteins can target nucleic acid interfaces with precision, potentially rivaling or complementing small molecule inhibitors.
The success of BindCraft in designing functional inhibitors against these challenging and dynamic nucleic acid binding surfaces challenges the prevailing dogma that such interfaces are beyond reach for protein-based therapeutics. By harnessing computational design and experimental validation synergistically, this technology paves the way for custom-tailored molecular tools capable of modulating gene editing outcomes or providing novel avenues for antibacterial or antiviral strategies.
Moreover, these findings herald a significant shift in therapeutic design paradigms, where synthetic protein inhibitors engineered in a single design iteration can surpass natural counterparts in specificity and efficacy. The ability to design de novo binders against large, flexible nucleic acid-binding surfaces could revolutionize drug discovery targeting transcription factors, RNA-binding proteins, and other critical biomolecules traditionally considered “undruggable.”
In practical applications, this approach could yield refined modulators of CRISPR systems that enable more precise gene editing control, reducing off-target effects or offering temporally regulated editing activities. The expansion to Argonaute nucleases also suggests potential interventions in prokaryotic immunity or microbiome engineering, domains currently limited by a lack of protein-based regulators.
Future directions will undoubtedly involve tightening affinity parameters, improving stability and cell permeability, and expanding the design repertoire to encompass broader classes of nucleic acid-protein interfaces. The integration of cryo-EM and other structural modalities with computational design further enhances the feedback loop necessary to refine and optimize such binders.
Altogether, BindCraft exemplifies the power of state-of-the-art protein engineering to overcome previously insurmountable challenges in drug targeting, opening new vistas in molecular medicine, biotechnology, and synthetic biology. This pioneering work epitomizes a new frontier where the line between natural biomolecular functions and engineered precision tools continues to blur, unlocking therapeutic potential once relegated to aspiration.
Subject of Research: Engineering of protein binders targeting nucleic acid interaction domains in multi-domain nucleases, specifically Cas9 and Argonaute enzymes.
Article Title: One-shot design of functional protein binders with BindCraft.
Article References:
Pacesa, M., Nickel, L., Schellhaas, C. et al. One-shot design of functional protein binders with BindCraft. Nature (2025). https://doi.org/10.1038/s41586-025-09429-6
Image Credits: AI Generated
