Increasing the binding affinity of an antibody to its target antigen is a crucial task in antibody therapeutics development. This paper presents a pretrainable geometric graph neural network, GearBind, and explores its potential in in silico affinity maturation. Leveraging multi-relational graph construction, multi-level geometric message passing and contrastive pretraining on mass-scale, unlabeled protein structural data, GearBind outperforms previous state-of-the-art approaches on SKEMPI and an independent test set. A powerful ensemble model based on GearBind is then derived and used to successfully enhance the binding of two antibodies with distinct formats and target antigens. ELISA EC50 values of the designed antibody mutants are decreased by up to 17 fold, and KD values by up to 6.1 fold. These promising results underscore the utility of geometric deep learning and effective pretraining in macromolecule interaction modeling tasks.
Plastic waste, particularly polyethylene terephthalate (PET), presents significant environmental challenges, driving extensive research into enzymatic biodegradation. However, existing PET hydrolases (PETases) are limited by narrow sequence diversity and suboptimal performance. This study introduces VenusMine, a protein discovery pipeline that integrates protein language models (PLMs) with a representation tree to identify PETases based on structural similarity using sequence information. Using the crystal structure of IsPETase as a template, VenusMine identifies and clusters target proteins. Candidates are further screened using PLM-based assessments of solubility and thermostability, leading to the selection of 34 proteins for biochemical validation. Results reveal that 14 candidates exhibit PET degradation activity across 30–60 °C. Notably, a PET hydrolase from Kibdelosporangium banguiense (KbPETase) demonstrates a melting temperature (Tm) 32 °C higher than IsPETase and exhibits the highest PET degradation activity within 30 – 65 °C among wild-type PETases. KbPETase also surpasses FastPETase and LCC in catalytic efficiency. X-ray crystallography and molecular dynamics simulations show that KbPETase possesses a conserved catalytic domain and enhanced intramolecular interactions, underpinning its improved functionality and thermostability. This work demonstrates a novel deep learning approach for discovering natural PETases with enhanced properties.
Deep learning-based methods for generating functional proteins address the growing need for novel biocatalysts, allowing for precise tailoring of functionalities to meet specific requirements. This advancement leads to the development of highly efficient and specialized proteins with diverse applications across scientific, technological, and biomedical fields. This study establishes a pipeline for protein sequence generation with a conditional protein diffusion model, namely CPDiffusion, to create diverse sequences of proteins with enhanced functions. CPDiffusion accommodates protein-specific conditions, such as secondary structures and highly conserved amino acids. Without relying on extensive training data, CPDiffusion effectively captures highly conserved residues and sequence features for specific protein families. We applied CPDiffusion to generate artificial sequences of Argonaute (Ago) proteins based on the backbone structures of wild-type (WT) Kurthia massiliensis Ago (KmAgo) and Pyrococcus furiosus Ago (PfAgo), which are complex multi-domain programmable endonucleases. The generated sequences deviate by up to nearly 400 amino acids from their WT templates. Experimental tests demonstrated that the majority of the generated proteins for both KmAgo and PfAgo show unambiguous activity in DNA cleavage, with many of them exhibiting superior activity as compared to the WT. These findings underscore CPDiffusion’s remarkable success rate in generating novel sequences for proteins with complex structures and functions in a single step, leading to enhanced activity. This approach facilitates the design of enzymes with multi-domain molecular structures and intricate functions through in silico generation and screening, all accomplished without the need for supervision from labeled data.
Protein language models (PLMs) have shown remarkable capabilities in various protein function prediction tasks. However, while protein function is intricately tied to structure, most existing PLMs do not incorporate protein structure information. To address this issue, we introduce ProSST, a Transformer-based protein language model that seamlessly integrates both protein sequences and structures. ProSST incorporates a structure quantization module and a Transformer architecture with disentangled attention. The structure quantization module translates a 3D protein structure into a sequence of discrete tokens by first serializing the protein structure into residue-level local structures and then embeds them into dense vector space. These vectors are then quantized into discrete structure tokens by a pre-trained clustering model. These tokens serve as an effective protein structure representation. Furthermore, ProSST explicitly learns the relationship between protein residue token sequences and structure token sequences through the sequence-structure disentangled attention. We pre-train ProSST on millions of protein structures using a masked language model objective, enabling it to learn comprehensive contextual representations of proteins. To evaluate the proposed ProSST, we conduct extensive experiments on the zero-shot mutation effect prediction and several supervised downstream tasks, where ProSST achieves the state-of-the-art performance among all baselines. Our code and pretrained models are publicly available 2 .
Ancestral metabolism has remained controversial due to a lack of evidence beyond sequence-based reconstructions. Although prebiotic chemists have provided hints that metabolism might originate from non-enzymatic protometabolic pathways, gaps between ancestral reconstruction and prebiotic processes mean there is much that is still unknown. Here, we apply proteome-wide 3D structure predictions and comparisons to investigate ancestorial metabolism of ancient bacteria and archaea, to provide information beyond sequence as a bridge to the prebiotic processes. We compare representative bacterial and archaeal strains, which reveal surprisingly similar physiological and metabolic characteristics via microbiological and biophysical experiments. Pairwise comparison of protein structures identify the conserved metabolic modules in bacteria and archaea, despite interference from overly variable sequences. The conserved modules (for example, middle of glycolysis, partial TCA, proton/sulfur respiration, building block biosynthesis) constitute the basic functions that possibly existed in the archaeal-bacterial common ancestor, which are remarkably consistent with the experimentally confirmed protometabolic pathways. These structure-based findings provide a new perspective to reconstructing the ancestral metabolism and understanding its origin, which suggests high-throughput protein 3D structure prediction is a promising approach, deserving broader application in future ancestral exploration.
Prokaryotic Argonaute (pAgo) proteins, a class of DNA/RNA-guided programmable endonucleases, have been extensively utilized in nucleic acid-based biosensors. The specific binding and cleavage of nucleic acids by pAgo proteins, which are crucial processes for their applications, are dependent on the presence of Mn2+ bound in the pockets, as verified through X-ray crystallography. However, a comprehensive understanding of how dissociated Mn2+ in the solvent affects the catalytic cycle, and its underlying regulatory role in this structure–function relationship, remains underdetermined. By combining experimental and computational methods, this study reveals that unbound Mn2+ in solution enhances the flexibility of diverse pAgo proteins. This increase in flexibility through decreasing the number of hydrogen bonds, induced by Mn2+, leads to higher affinity for substrates, thus facilitating cleavage. More importantly, Mn2+-induced structural flexibility increases the mismatch tolerance between guide-target pairs by increasing the conformational states, thereby enhancing the cleavage of mismatches. Further simulations indicate that the enhanced flexibility in linkers triggers conformational changes in the PAZ domain for recognizing various lengths of nucleic acids. Additionally, Mn2+-induced dynamic alterations of the protein cause a conformational shift in the N domain and catalytic sites towards their functional form, resulting in a decreased energy penalty for target release and cleavage. These findings demonstrate that the dynamic conformations of pAgo proteins, resulting from the presence of the unbound Mn2+ in solution, significantly promote the catalytic cycle of endonucleases and the tolerance of cleavage to mismatches. This flexibility enhancement mechanism serves as a general strategy employed by Ago proteins from diverse prokaryotes to accomplish their catalytic functions and provide useful information for Ago-based precise molecular diagnostics.