Machine learning-aided engineering of hydrolases for PET depolymerization.

Plastic waste poses an ecological challenge1-3 and enzymatic degradation offers one, potentially green and scalable, route for polyesters waste recycling4. Poly(ethylene terephthalate) (PET) accounts for 12% of global solid waste5, and a circular carbon economy for PET is theoretically attainable through rapid enzymatic depolymerization followed by repolymerization or conversion/valorization into other products6-10. Application of PET hydrolases, however, has been hampered by their lack of robustness to pH and temperature ranges, slow reaction rates and inability to directly use untreated postconsumer plastics11. Here, we use a structure-based, machine learning algorithm to engineer a robust and active PET hydrolase. Our mutant and scaffold combination (FAST-PETase: functional, active, stable and tolerant PETase) contains five mutations compared to wild-type PETase (N233K/R224Q/S121E from prediction and D186H/R280A from scaffold) and shows superior PET-hydrolytic activity relative to both wild-type and engineered alternatives12 between 30 and 50 °C and a range of pH levels. We demonstrate that untreated, postconsumer-PET from 51 different thermoformed products can all be almost completely degraded by FAST-PETase in 1 week. FAST-PETase can also depolymerize untreated, amorphous portions of a commercial water bottle and an entire thermally pretreated water bottle at 50 ºC. Finally, we demonstrate a closed-loop PET recycling process by using FAST-PETase and resynthesizing PET from the recovered monomers. Collectively, our results demonstrate a viable route for enzymatic plastic recycling at the industrial scale.

Asymmetric Synthesis of α-Chloroamides via Photoenzymatic Hydroalkylation of Olefins.

Photoenzymatic intermolecular hydroalkylations of olefins are highly enantioselective for chiral centers formed during radical termination but poorly selective for centers set in the C-C bond-forming event. Here, we report the evolution of a flavin-dependent "ene"-reductase to catalyze the coupling of α,α-dichloroamides with alkenes to afford α-chloroamides in good yield with excellent chemo- and stereoselectivity. These products can serve as linchpins in the synthesis of pharmaceutically valuable motifs. Mechanistic studies indicate that radical formation occurs by exciting a charge-transfer complex templated by the protein. Precise control over the orientation of molecules within the charge-transfer complex potentially accounts for the observed stereoselectivity. The work expands the types of motifs that can be prepared using photoenzymatic catalysis.

Improved Bst DNA Polymerase Variants Derived via a Machine Learning Approach.

The DNA polymerase I from Geobacillus stearothermophilus (also known as Bst DNAP) is widely used in isothermal amplification reactions, where its strand displacement ability is prized. More robust versions of this enzyme should be enabled for diagnostic applications, especially for carrying out higher temperature reactions that might proceed more quickly. To this end, we appended a short fusion domain from the actin-binding protein villin that improved both stability and purification of the enzyme. In parallel, we have developed a machine learning algorithm that assesses the relative fit of individual amino acids to their chemical microenvironments at any position in a protein and applied this algorithm to predict sequence substitutions in Bst DNAP. The top predicted variants had greatly improved thermotolerance (heating prior to assay), and upon combination, the mutations showed additive thermostability, with denaturation temperatures up to 2.5 °C higher than the parental enzyme. The increased thermostability of the enzyme allowed faster loop-mediated isothermal amplification assays to be carried out at 73 °C, where both Bst DNAP and its improved commercial counterpart Bst 2.0 are inactivated. Overall, this is one of the first examples of the application of machine learning approaches to the thermostabilization of an enzyme.

Learning the local landscape of protein structures with convolutional neural networks.

One fundamental problem of protein biochemistry is to predict protein structure from amino acid sequence. The inverse problem, predicting either entire sequences or individual mutations that are consistent with a given protein structure, has received much less attention even though it has important applications in both protein engineering and evolutionary biology. Here, we ask whether 3D convolutional neural networks (3D CNNs) can learn the local fitness landscape of protein structure to reliably predict either the wild-type amino acid or the consensus in a multiple sequence alignment from the local structural context surrounding site of interest. We find that the network can predict wild type with good accuracy, and that network confidence is a reliable measure of whether a given prediction is likely going to be correct or not. Predictions of consensus are less accurate and are primarily driven by whether or not the consensus matches the wild type. Our work suggests that high-confidence mis-predictions of the wild type may identify sites that are primed for mutation and likely targets for protein engineering.

Transient neonatal zinc deficiency due to a new autosomal dominant mutation in gene SLC30A2 (ZnT-2).

Transient neonatal zinc deficiency (TNZD) has a clinical presentation similar to that of acrodermatitis enteropathica but is caused by a low zinc concentration in maternal breast milk. TNZD becomes clinically evident during breastfeeding and is resolved by weaning and the introduction of complementary nutrition. We present a 4-month-old girl with TNZD due to a new autosomal dominant mutation (663delC) in the maternal SLC30A2 gene not previously described in the literature.

Stability Oracle: a structure-based graph-transformer framework for identifying stabilizing mutations.

Engineering stabilized proteins is a fundamental challenge in the development of industrial and pharmaceutical biotechnologies. We present Stability Oracle: a structure-based graph-transformer framework that achieves SOTA performance on accurately identifying thermodynamically stabilizing mutations. Our framework introduces several innovations to overcome well-known challenges in data scarcity and bias, generalization, and computation time, such as: Thermodynamic Permutations for data augmentation, structural amino acid embeddings to model a mutation with a single structure, a protein structure-specific attention-bias mechanism that makes transformers a viable alternative to graph neural networks. We provide training/test splits that mitigate data leakage and ensure proper model evaluation. Furthermore, to examine our data engineering contributions, we fine-tune ESM2 representations (Prostata-IFML) and achieve SOTA for sequence-based models. Notably, Stability Oracle outperforms Prostata-IFML even though it was pretrained on 2000X less proteins and has 548X less parameters. Our framework establishes a path for fine-tuning structure-based transformers to virtually any phenotype, a necessary task for accelerating the development of protein-based biotechnologies.

Biosensor and machine learning-aided engineering of an amaryllidaceae enzyme.

A major challenge to achieving industry-scale biomanufacturing of therapeutic alkaloids is the slow process of biocatalyst engineering. Amaryllidaceae alkaloids, such as the Alzheimer's medication galantamine, are complex plant secondary metabolites with recognized therapeutic value. Due to their difficult synthesis they are regularly sourced by extraction and purification from the low-yielding daffodil Narcissus pseudonarcissus. Here, we propose an efficient biosensor-machine learning technology stack for biocatalyst development, which we apply to engineer an Amaryllidaceae enzyme in Escherichia coli. Directed evolution is used to develop a highly sensitive (EC50 = 20 µM) and specific biosensor for the key Amaryllidaceae alkaloid branchpoint 4'-O-methylnorbelladine. A structure-based residual neural network (MutComputeX) is subsequently developed and used to generate activity-enriched variants of a plant methyltransferase, which are rapidly screened with the biosensor. Functional enzyme variants are identified that yield a 60% improvement in product titer, 2-fold higher catalytic activity, and 3-fold lower off-product regioisomer formation. A solved crystal structure elucidates the mechanism behind key beneficial mutations.

Using machine learning to predict the effects and consequences of mutations in proteins.

Machine and deep learning approaches can leverage the increasingly available massive datasets of protein sequences, structures, and mutational effects to predict variants with improved fitness. Many different approaches are being developed, but systematic benchmarking studies indicate that even though the specifics of the machine learning algorithms matter, the more important constraint comes from the data availability and quality utilized during training. In cases where little experimental data are available, unsupervised and self-supervised pre-training with generic protein datasets can still perform well after subsequent refinement via hybrid or transfer learning approaches. Overall, recent progress in this field has been staggering, and machine learning approaches will likely play a major role in future breakthroughs in protein biochemistry and engineering.

Two sequence- and two structure-based ML models have learned different aspects of protein biochemistry.

Deep learning models are seeing increased use as methods to predict mutational effects or allowed mutations in proteins. The models commonly used for these purposes include large language models (LLMs) and 3D Convolutional Neural Networks (CNNs). These two model types have very different architectures and are commonly trained on different representations of proteins. LLMs make use of the transformer architecture and are trained purely on protein sequences whereas 3D CNNs are trained on voxelized representations of local protein structure. While comparable overall prediction accuracies have been reported for both types of models, it is not known to what extent these models make comparable specific predictions and/or generalize protein biochemistry in similar ways. Here, we perform a systematic comparison of two LLMs and two structure-based models (CNNs) and show that the different model types have distinct strengths and weaknesses. The overall prediction accuracies are largely uncorrelated between the sequence- and structure-based models. Overall, the two structure-based models are better at predicting buried aliphatic and hydrophobic residues whereas the two LLMs are better at predicting solvent-exposed polar and charged amino acids. Finally, we find that a combined model that takes the individual model predictions as input can leverage these individual model strengths and results in significantly improved overall prediction accuracy.

Two sequence- and two structure-based ML models have learned different aspects of protein biochemistry.

Deep learning models are seeing increased use as methods to predict mutational effects or allowed mutations in proteins. The models commonly used for these purposes include large language models (LLMs) and 3D Convolutional Neural Networks (CNNs). These two model types have very different architectures and are commonly trained on different representations of proteins. LLMs make use of the transformer architecture and are trained purely on protein sequences whereas 3D CNNs are trained on voxelized representations of local protein structure. While comparable overall prediction accuracies have been reported for both types of models, it is not known to what extent these models make comparable specific predictions and/or generalize protein biochemistry in similar ways. Here, we perform a systematic comparison of two LLMs and two structure-based models (CNNs) and show that the different model types have distinct strengths and weaknesses. The overall prediction accuracies are largely uncorrelated between the sequence- and structure-based models. Overall, the two structure-based models are better at predicting buried aliphatic and hydrophobic residues whereas the two LLMs are better at predicting solvent-exposed polar and charged amino acids. Finally, we find that a combined model that takes the individual model predictions as input can leverage these individual model strengths and results in significantly improved overall prediction accuracy.

GroovDB: A Database of Ligand-Inducible Transcription Factors.

Genetic biosensors are integral to synthetic biology. In particular, ligand-inducible prokaryotic transcription factors are frequently used in high-throughput screening, for dynamic feedback regulation, as multilayer logic gates, and in diagnostic applications. In order to provide a curated source that users can rely on for engineering applications, we have developed GroovDB (available at https://groov.bio), a Web-accessible database of ligand-inducible transcription factors that contains all information necessary to build chemically responsive genetic circuits, including biosensor sequence, ligand, and operator data. Ligand and DNA interaction data have been verified against the literature, while an automated data curation pipeline is used to programmatically fetch metadata, structural information, and references for every entry. A custom tool to visualize the natural genetic context of biosensor entries provides potential insights into alternative ligands and systems biology.

Discovery of Novel Gain-of-Function Mutations Guided by Structure-Based Deep Learning.

Despite the promise of deep learning accelerated protein engineering, examples of such improved proteins are scarce. Here we report that a 3D convolutional neural network trained to associate amino acids with neighboring chemical microenvironments can guide identification of novel gain-of-function mutations that are not predicted by energetics-based approaches. Amalgamation of these mutations improved protein function in vivo across three diverse proteins by at least 5-fold. Furthermore, this model provides a means to interrogate the chemical space within protein microenvironments and identify specific chemical interactions that contribute to the gain-of-function phenotypes resulting from individual mutations.