De novo enzyme design seeks to build proteins containing ideal active sites with catalytic residues surrounding and stabilizing the transition state(s) of the target chemical reaction^{1,2,3,4,5,6,7}. The generative artificial intelligence method RFdiffusion^8,9 solves this problem, but requires specifying both the sequence position and backbone coordinates for each catalytic residue, limiting sampling. Here we introduce RFdiffusion2, which eliminates these requirements, and use it to design zinc metallohydrolases starting from quantum chemistry-derived active site geometries. From an initial set of 96 designs tested experimentally, the most active has a catalytic efficiency (k_cat/K_M) of 16,000 M⁻¹ s⁻¹, orders of magnitude higher than previously designed metallohydrolases^6,7,10,11. A second round of 96 designs yielded 3 additional highly active enzymes, with k_cat/K_M values of up to 53,000 M⁻¹ s⁻¹ and a catalytic rate constant (k_cat) of up to 1.5 s⁻¹. The design models of the four most active designs differ from known structures and from each other, and the crystal structure of the most active design is very close to the design model, demonstrating the accuracy of the design method. The most active enzymes are predicted by PLACER¹² and Chai-1 (ref. ¹³) to have preorganized active sites that effectively position the substrate for nucleophilic attack by a water molecule activated by the bound metal. The ability to generate highly active enzymes directly from the computer, without experimental optimization, should enable a new generation of potent designer catalysts^14,15.

Tissue engineering aims to repair, restore, and/or replace tissues in the human body as an alternative to grafts and prostheses. Biomaterial scaffolds can be utilized to provide a three-dimensional microenvironment to facilitate tissue regeneration. Previously, we reported that scaffold pore size influences vascularization and extracellular matrix composition both in vivo and in vitro, to ultimately influence tissue phenotype for regenerating cranial suture and bone tissues, which have markedly different tissue properties despite similar multipotent stem cell populations. To rationally design biomaterials for specific cell and tissue fate specification, it is critical to understand the molecular processes governed by cell-biomaterial interactions, which guide cell fate specification. Building on our previous work, in this report we investigated the hypothesis that scaffold pore curvature, the direct consequence of pore size, modulates the differentiation trajectory of mesenchymal stem cells (MSCs) through alterations in the cytoskeleton. First, we demonstrated that sufficiently small pores facilitate cell clustering in subcutaneous explants cultured in vivo, which we previously reported to demonstrate stem tissue phenotype both in vivo and in vitro. Based on this observation, we cultured cell-scaffold constructs in vitro to assess early time point interactions between cells and the matrix as a function of pore size. We demonstrate that principle curvature directly influences nuclear aspect and cell aggregation in vitro. Scaffold pores with a sufficiently low degree of principle curvature enables cell differentiation; pharmacologic inhibition of actin cytoskeleton polymerization in these scaffolds decreased differentiation, indicating a critical role of the cytoskeleton in transducing cues from the scaffold pore microenvironment to the cell nucleus. We fabricated a macropore model, which allows for three-dimensional confocal imaging and demonstrates that a higher principle curvature facilitates cell aggregation and the formation of a potentially protective niche within scaffold macropores which prevents MSC differentiation and retains their stemness. Sufficiently high principle curvature upregulates yes-associated protein (YAP) phosphorylation while decreased principle curvature downregulates YAP phosphorylation and increases YAP nuclear translocation with subsequent transcriptional activation towards an osteogenic differentiation fate. Finally, we demonstrate that the inhibition of the YAP/TAZ pathway causes a defect in differentiation, while YAP/TAZ activation causes premature differentiation in a curvature-dependent way when modulated by verteporfin (VP) and 1-oleyl-lysophosphatidic acid (LPA), respectively, confirming the critical role of biomaterials-mediated YAP/TAZ signaling in cell differentiation and fate specification. Our data support that the principle curvature of scaffold macropores is a critical design criterion which guides the differentiation trajectory of mesenchymal stem cells’ scaffolds. Biomaterial-mediated regulation of YAP/TAZ may significantly contribute to influencing the regenerative outcomes of biomaterials-based tissue engineering strategies through their specific pore design.

De novo enzyme design starts from ideal active site descriptions consisting of constellations of catalytic residue functional groups around reaction transition state(s), and seeks to generate protein structures that can accurately hold the site in place. Highly active enzymes have been designed starting from such descriptions using the generative AI method RFdiffusion [1–3], but there are two current methodological limitations. First, the geometry of the active site can only be specified at the residue level, so for each catalytic residue functional group placed around the reaction transition state, the possible locations of the residue backbone must be enumerated by building side chain rotamers back from the functional group. Second, the location of the catalytic residues along the sequence must be specified in advance, which considerably limits the space of solutions which can be sampled. Here we describe a new deep generative method, Rosetta Fold diffusion 2 (RFdiffusion2), that solves both problems, enabling enzymes to be designed from sequence agnostic descriptions of functional group locations without inverse rotamer generation. We first evaluate RFdiffusion2 on an in silico enzyme design benchmark of 41 diverse active sites and find that it is able to successfully build proteins scaffolding all 41 sites, compared to 16/41 with prior state-of-the-art deep learning methods. Next, we design enzymes around three diverse catalytic sites and characterize the designs experimentally; in each case we identify active catalysts in testing less than 96 sequences. RFdiffusion2 demonstrates the potential of atomic resolution generative models for the design of de novo enzymes directly from their reaction mechanisms.

Acidic pH is critical to the function of the gastrointestinal system, bone-resorbing osteoclasts, and the endolysosomal compartment of nearly every cell in the body. Non-invasive, real-time fluorescence imaging of acidic microenvironments represents a powerful tool for understanding normal cellular biology, defining mechanisms of disease, and monitoring for therapeutic response. While commercially available pH-sensitive fluorescent probes exist, several limitations hinder their widespread use and potential for biologic application. To address this need, we developed a novel library of pH-sensitive probes based on the highly photostable and water-soluble fluorescent molecule, Rhodamine 6G. We demonstrate versatility in terms of both pH sensitivity (i.e., pK_a) and chemical functionality, allowing conjugation to small molecules, proteins, nanoparticles, and regenerative biomaterial scaffold matrices. Furthermore, we show preserved pH-sensitive fluorescence following a variety of forms of covalent functionalization and demonstrate three potential applications, both in vitro and in vivo, for intracellular and extracellular pH sensing. Finally, we develop a computation approach for predicting the pH sensitivity of R6G derivatives, which could be used to expand our library and generate probes with novel properties.

De novo enzyme design starts from a description of an ideal active site composed of catalytic residues surrounding the reaction transition state(s), and builds a protein structure that contains this site ^1–7 . Generative AI methods such as RFdiffusion ^11,12 now enable the direct generation of proteins around active sites, but to date, such scaffolding has required specification of both the position in the sequence and the backbone coordinates of each catalytic residue, which complicates sampling. Here we introduce a generative AI method called RFdiffusion2 that overcomes these limitations and use it to design zinc metallohydrolases starting from a density functional theory description of the active site geometry. Of an initial set of 96 designs tested experimentally, the most active has a k _cat / K _M of 16,000 M ⁻¹ s ⁻¹ , orders of magnitude higher than previously designed metallohydrolases. ^6,7,13,14 A second round of 96 designs yielded 3 additional highly active enzymes, with k _cat / K _M up to 53,000 M ⁻¹ s ⁻¹ and k _cat up to 1.5 s ⁻¹ . The structures of the four enzymes are very different from each other and from the structures in the PDB. Each enzyme positions the reaction substrate almost perfectly for nucleophilic attack by a water molecule activated by the bound metal, and are predicted by PLACER ¹⁵ and Chai-1 ⁴⁴ to have highly preorganized active sites. The ability to generate highly active catalysts straight out of the computer, without experimental optimization, using quantum chemistry calculated active site geometries should open the door to a new generation of potent designer enzymes. ^16,17

De novo design of enzymes that cleave amide bonds represents a major challenge ^1–2 . Here we use RFdiffusion methods ^3–4 to de novo design proteins with water activating zinc based active sites adjacent to a peptide substrate binding pocket. Of 135 designs tested, 36% catalyze cleavage of their cognate peptide substrates at the targeted scissile bond. The best design accelerates peptide-bond hydrolysis by more than 10 ⁸ -fold relative to the uncatalyzed reaction; four substitutions increase this acceleration to over 10 ¹⁰ -fold by improving active-site preorganization ⁵ . We use this approach to design metalloproteases that selectively cleave the human disease associated proteins TDP-43, amyloid-β, and serum amyloid A, with rate enhancements of up to 10 ⁹ -fold. We further design caged cytokines and receptor antagonists that are selectively activated by the designed proteases, enabling bioorthogonal control of cell state. Thus, de novo design can yield highly active, programmable proteases with applications in therapeutics, biotechnology, and bioremediation.