Learning protein fitness landscapes with deep mutational scanning data from multiple sources

Ancestral reconstruction of polyethylene terephthalate degrading cutinases reveals a rugged and unexplored sequence-fitness landscape

The use of protein engineering to generate enzymes for the degradation of polyethylene terephthalate (PET) is a promising route for plastic recycling, yet traditional engineering approaches often fail to explore protein sequence space for optimal enzymes. In this work, we use multiplexed ancestral sequence reconstruction (mASR) to address this, exploring the evolutionary sequence space of PET-degrading cutinases. Using 20 statistically equivalent phylogenies of the bacterial cutinase family, we generated 48 ancestral sequences revealing a wide range of PETase activities, highlighting the value of mASR in uncovering functional variants. Our findings show PETase activity can evolve through multiple pathways involving mutations remote from the active site. Moreover, analyzing the PETase fitness landscape with local ancestral sequence embedding (LASE) revealed that LASE can capture sequence features linked to PETase activity. This work highlights mASR's potential in exploration of sequence space and underscores the use of LASE in readily mapping the protein fitness landscapes.

Technological advancement spurs Komagataella phaffii as a next-generation platform for sustainable biomanufacturing

Biomanufacturing stands as a cornerstone of sustainable industrial development, necessitating a shift toward non-food carbon feedstocks to alleviate agricultural resource competition and advance a circular bioeconomy. Methanol, a renewable one‑carbon substrate, has emerged as a pivotal candidate due to its abundance, cost-effectiveness, and high reduction potential, further bolstered by breakthroughs in CO₂ hydrogenation-based synthesis. Capitalizing on this momentum, the methylotrophic yeast Komagataella phaffii has undergone transformative technological upgrades, evolving from a conventional protein expression workhorse into an intelligent bioproduction chassis. This paradigm shift is fundamentally driven by converging innovations across CRISPR-empowered advancement in genome editing and AI-powered metabolic pathway design in K. phaffii. The integration of CRISPR systems with droplet microfluidics high-throughput screening has redefined strain engineering efficiency, achieving much higher editing precision than traditional homologous recombination while compressing the "design-build-test-learn" cycle. Concurrently, machine learning-enhanced genome-scale metabolic models facilitate dynamic flux balancing, enabling simultaneous improvements in product titers, carbon yields, and volumetric productivity. Finally, technological advancement promotes the application of K. phaffii, including directing more efficiently metabolic flux toward nutrient products, and strengthening efficient synthesis of excreted proteins. As DNA synthesis automation and robotic experimentation platforms mature, next-generation breakthroughs in genome modification, cofactor engineering, and AI-guided autonomous evolution will further cement K. phaffii as a next-generation platform for decarbonizing global manufacturing paradigms. This technological trajectory positions methanol-based biomanufacturing as a cornerstone of the low-carbon circular economy.

Integrating protein contact networks for the engineering of thermostable lipase A

In the field of industrial biocatalysis, the rapid advancement of enzyme functional evolution necessitates new theories and computational methods to achieve target functions with fewer iterations. This study identified key residues affecting enzyme stability by constructing the protein contact network (PCN) of Lipase A. Comparing the PCNs of the wild-type (WT) and the 6B variant revealed that changes in residue interactions and node properties (e.g., degree and betweenness centrality (BC)) positively impacted stability. Using thresholds for degree and BC, 25 candidate sites were screened, and 11 out of 18 single-point mutation designs improved thermal stability. Mutations were divided into three groups (M1, M2, M3) based on network communities and contributions, followed by iterative combinations. M1, containing five mutations distributed across four communities, increased the melting temperature (Tm) by 14.61 °C, close to the predicted 13.97 °C, demonstrating a linear additive effect. In M2, three new mutations resulted in a non-linear additive effect, with a ΔTm of 17.58 °C (Expected ΔTm = 18.93 °C). In contrast, the three new mutations in M3 destabilized the enzyme (Observed ΔTm = 15.94 °C vs Expected ΔTm = 19.92 °C). Molecular dynamics simulations showed that polar edge nodes enhanced network connectivity, while proline mutations rigidified flexible regions, improving stability. Conversely, M3 mutations disrupted α-helix stability by increasing the dihedral angle fluctuations of residue Y161, might to a stability-activity trade-off. The PCN provides valuable insights for developing efficient and precise design strategies.

Biophysics-based protein language models for protein engineering

Protein language models trained on evolutionary data have emerged as powerful tools for predictive problems involving protein sequence, structure, and function. However, these models overlook decades of research into biophysical factors governing protein function. We propose Mutational Effect Transfer Learning (METL), a protein language model framework that unites advanced machine learning and biophysical modeling. Using the METL framework, we pretrain transformer-based neural networks on biophysical simulation data to capture fundamental relationships between protein sequence, structure, and energetics. We finetune METL on experimental sequence-function data to harness these biophysical signals and apply them when predicting protein properties like thermostability, catalytic activity, and fluorescence. METL excels in challenging protein engineering tasks like generalizing from small training sets and position extrapolation, although existing methods that train on evolutionary signals remain powerful for many types of experimental assays. We demonstrate METL's ability to design functional green fluorescent protein variants when trained on only 64 examples, showcasing the potential of biophysics-based protein language models for protein engineering.

Advances in stress-tolerance elements for microbial cell factories

Microorganisms, particularly extremophiles, have evolved multiple adaptation mechanisms to address diverse stress conditions during survival in unique environments. Their responses to environmental coercion decide not only survival in severe conditions but are also an essential factor determining bioproduction performance. The design of robust cell factories should take the balance of their growing and bioproduction into account. Thus, mining and redesigning stress-tolerance elements to optimize the performance of cell factories under various extreme conditions is necessary. Here, we reviewed several stress-tolerance elements, including acid-tolerant elements, saline-alkali-resistant elements, thermotolerant elements, antioxidant elements, and so on, providing potential materials for the construction of cell factories and the development of synthetic biology. Strategies for mining and redesigning stress-tolerance elements were also discussed. Moreover, several applications of stress-tolerance elements were provided, and perspectives and discussions for potential strategies for screening stress-tolerance elements were made.

A protein fitness predictive framework based on feature combination and intelligent searching

Machine learning (ML) constructs predictive models by understanding the relationship between protein sequences and their functions, enabling efficient identification of protein sequences with high fitness values without falling into local optima, like directional evolution. However, how to extract the most pertinent functional feature information from a limited number of protein sequences is vital for optimizing the performance of ML models. Here, we propose scut_ProFP (Protein Fitness Predictor), a predictive framework that integrates feature combination and feature selection techniques. Feature combination offers comprehensive sequence information, while feature selection searches for the most beneficial features to enhance model performance, enabling accurate sequence-to-function mapping. Compared to similar frameworks, scut_ProFP demonstrates superior performance and is also competitive with more complex deep learning models-ECNet, EVmutation, and UniRep. In addition, scut_ProFP enables generalization from low-order mutants to high-order mutants. Finally, we utilized scut_ProFP to simulate the engineering of the fluorescent protein CreiLOV and highly enriched mutants with high fluorescence based on only a small number of low-fluorescence mutants. Essentially, the developed method is advantageous for ML in protein engineering, providing an effective approach to data-driven protein engineering. The code and datasets for scut_ProFP are available at https://github.com/Zhang66-star/scut_ProFP.

Comprehensive Understanding of Laboratory Evolution for Food Enzymes: From Design to Screening Innovations

In the field of food processing, enzymes play a pivotal role in improving product quality and flavor, and extending shelf life. However, the exposure of traditional food enzymes to high temperatures during processing often leads to a decrease in activity or even inactivation, limiting the effectiveness of their application under high-temperature conditions. Therefore, the modification of thermostability and activity of enzymes to adapt to extreme conditions through protein engineering has become a key way to improve the efficiency and economic benefits of industrial production. Directed evolution and semirational design strategies in the laboratory have proven to be broadly applicable frameworks for biochemical researchers in the food field, including those who are beginners. In this review, we systematically summarize semirational design strategies and high-throughput screening strategies, and introduce some intuitive computer simulation software to improve the thermostability and enzyme activity of food enzymes. The application of these strategies and techniques provides a comprehensive guide for the optimization of food enzymes. In addition, the latest hot topics of genetically engineered food enzymes in the field of application are discussed.

Deep learning in template-free de novo biosynthetic pathway design of natural products

Natural products (NPs) are indispensable in drug development, particularly in combating infections, cancer, and neurodegenerative diseases. However, their limited availability poses significant challenges. Template-free de novo biosynthetic pathway design provides a strategic solution for NP production, with deep learning standing out as a powerful tool in this domain. This review delves into state-of-the-art deep learning algorithms in NP biosynthesis pathway design. It provides an in-depth discussion of databases like Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, and UniProt, which are essential for model training, along with chemical databases such as Reaxys, SciFinder, and PubChem for transfer learning to expand models' understanding of the broader chemical space. It evaluates the potential and challenges of sequence-to-sequence and graph-to-graph translation models for accurate single-step prediction. Additionally, it discusses search algorithms for multistep prediction and deep learning algorithms for predicting enzyme function. The review also highlights the pivotal role of deep learning in improving catalytic efficiency through enzyme engineering, which is essential for enhancing NP production. Moreover, it examines the application of large language models in pathway design, enzyme discovery, and enzyme engineering. Finally, it addresses the challenges and prospects associated with template-free approaches, offering insights into potential advancements in NP biosynthesis pathway design.

Zero-shot prediction of mutation effects with multimodal deep representation learning guides protein engineering

Mutations in amino acid sequences can provoke changes in protein function. Accurate and unsupervised prediction of mutation effects is critical in biotechnology and biomedicine, but remains a fundamental challenge. To resolve this challenge, here we present Protein Mutational Effect Predictor (ProMEP), a general and multiple sequence alignment-free method that enables zero-shot prediction of mutation effects. A multimodal deep representation learning model embedded in ProMEP was developed to comprehensively learn both sequence and structure contexts from ~160 million proteins. ProMEP achieves state-of-the-art performance in mutational effect prediction and accomplishes a tremendous improvement in speed, enabling efficient and intelligent protein engineering. Specifically, ProMEP accurately forecasts mutational consequences on the gene-editing enzymes TnpB and TadA, and successfully guides the development of high-performance gene-editing tools with their engineered variants. The gene-editing efficiency of a 5-site mutant of TnpB reaches up to 74.04% (vs 24.66% for the wild type); and the base editing tool developed on the basis of a TadA 15-site mutant (in addition to the A106V/D108N double mutation that renders deoxyadenosine deaminase activity to TadA) exhibits an A-to-G conversion frequency of up to 77.27% (vs 69.80% for ABE8e, a previous TadA-based adenine base editor) with significantly reduced bystander and off-target effects compared to ABE8e. ProMEP not only showcases superior performance in predicting mutational effects on proteins but also demonstrates a great capability to guide protein engineering. Therefore, ProMEP enables efficient exploration of the gigantic protein space and facilitates practical design of proteins, thereby advancing studies in biomedicine and synthetic biology.

Addressing epistasis in the design of protein function

Mutations in protein active sites can dramatically improve function. The active site, however, is densely packed and extremely sensitive to mutations. Therefore, some mutations may only be tolerated in combination with others in a phenomenon known as epistasis. Epistasis reduces the likelihood of obtaining improved functional variants and dramatically slows natural and lab evolutionary processes. Research has shed light on the molecular origins of epistasis and its role in shaping evolutionary trajectories and outcomes. In addition, sequence- and AI-based strategies that infer epistatic relationships from mutational patterns in natural or experimental evolution data have been used to design functional protein variants. In recent years, combinations of such approaches and atomistic design calculations have successfully predicted highly functional combinatorial mutations in active sites. These were used to design thousands of functional active-site variants, demonstrating that, while our understanding of epistasis remains incomplete, some of the determinants that are critical for accurate design are now sufficiently understood. We conclude that the space of active-site variants that has been explored by evolution may be expanded dramatically to enhance natural activities or discover new ones. Furthermore, design opens the way to systematically exploring sequence and structure space and mutational impacts on function, deepening our understanding and control over protein activity.

Neural network extrapolation to distant regions of the protein fitness landscape

Machine learning (ML) has transformed protein engineering by constructing models of the underlying sequence-function landscape to accelerate the discovery of new biomolecules. ML-guided protein design requires models, trained on local sequence-function information, to accurately predict distant fitness peaks. In this work, we evaluate neural networks' capacity to extrapolate beyond their training data. We perform model-guided design using a panel of neural network architectures trained on protein G (GB1)-Immunoglobulin G (IgG) binding data and experimentally test thousands of GB1 designs to systematically evaluate the models' extrapolation. We find each model architecture infers markedly different landscapes from the same data, which give rise to unique design preferences. We find simpler models excel in local extrapolation to design high fitness proteins, while more sophisticated convolutional models can venture deep into sequence space to design proteins that fold but are no longer functional. We also find that implementing a simple ensemble of convolutional neural networks enables robust design of high-performing variants in the local landscape. Our findings highlight how each architecture's inductive biases prime them to learn different aspects of the protein fitness landscape and how a simple ensembling approach makes protein engineering more robust.

Through virtual saturation mutagenesis and rational design for superior substrate conversion in engineered d-amino acid oxidase

The d-amino acid oxidase (DAAO) is pivotal in obtaining optically pure l-glufosinate (l-PPT) by converting d-glufosinate (d-PPT) to its deamination product. We screened and designed a Rasamsonia emersonii DAAO (ReDAAO), making it more suitable for oxidizing d-PPT. Using Caver 3.0, we delineated three substrate binding pockets and, via alanine scanning, identified nearby key residues. Pinpointing key residues influencing activity, we applied virtual saturation mutagenesis (VSM), and experimentally validated mutants which reduced substrate binding energy. Analysis of positive mutants revealed elongated side-chain prevalence in substrate binding pocket periphery. Although computer-aided approaches can rapidly identify advantageous mutants and guide further design, the mutations obtained in the first round may not be suitable for combination with other advantageous mutations. Therefore, each round of combination requires reasonable iteration. Employing VSM-assisted screening multiple times and after four rounds of combining mutations, we ultimately obtained a mutant, N53V/F57Q/V94R/V242R, resulting in a mutant with a 5097% increase in enzyme activity compared to the wild type. It provides valuable insights into the structural determinants of enzyme activity and introduces a novel rational design procedure.