Protein Design by Directed Evolution

Deep-learning-based design of synthetic orthologs of SH3 signaling domains

Evolution-based deep generative models represent an exciting direction in understanding and designing proteins. An open question is whether such models can learn specialized functional constraints that control fitness in specific biological contexts. Here, we examine the ability of generative models to produce synthetic versions of Src-homology 3 (SH3) domains that mediate signaling in the Sho1 osmotic stress response pathway of yeast. We show that a variational autoencoder (VAE) model produces artificial sequences that experimentally recapitulate the function of natural SH3 domains. More generally, the model organizes all fungal SH3 domains such that locality in the model latent space (but not simply locality in sequence space) enriches the design of synthetic orthologs and exposes non-obvious amino acid constraints distributed near and far from the SH3 ligand-binding site. The ability of generative models to design ortholog-like functions in vivo opens new avenues for engineering protein function in specific cellular contexts and environments.

Droplet-based microfluidics and enzyme evolution

Enzymes are widely used as catalysts in the chemical and pharmaceutical industries. While successful in many situations, they must usually be adapted to operate efficiently under nonnatural conditions. Enzyme engineering allows the creation of novel enzymes that are stable at elevated temperatures or have higher activities and selectivities. Current enzyme engineering techniques require the production and testing of enzyme variant libraries to identify members with desired attributes. Unfortunately, traditional screening methods cannot screen such large mutagenesis libraries in a robust and timely manner. Droplet-based microfluidic systems can produce, process, and sort picoliter droplets at kilohertz rates and have emerged as powerful tools for library screening and thus enzyme engineering. We describe how droplet-based microfluidics has been used to advance directed evolution.

Computational scoring and experimental evaluation of enzymes generated by neural networks

In recent years, generative protein sequence models have been developed to sample novel sequences. However, predicting whether generated proteins will fold and function remains challenging. We evaluate a set of 20 diverse computational metrics to assess the quality of enzyme sequences produced by three contrasting generative models: ancestral sequence reconstruction, a generative adversarial network and a protein language model. Focusing on two enzyme families, we expressed and purified over 500 natural and generated sequences with 70-90% identity to the most similar natural sequences to benchmark computational metrics for predicting in vitro enzyme activity. Over three rounds of experiments, we developed a computational filter that improved the rate of experimental success by 50-150%. The proposed metrics and models will drive protein engineering research by serving as a benchmark for generative protein sequence models and helping to select active variants for experimental testing.

Perspectives on Computational Enzyme Modeling: From Mechanisms to Design and Drug Development

Understanding enzyme mechanisms is essential for unraveling the complex molecular machinery of life. In this review, we survey the field of computational enzymology, highlighting key principles governing enzyme mechanisms and discussing ongoing challenges and promising advances. Over the years, computer simulations have become indispensable in the study of enzyme mechanisms, with the integration of experimental and computational exploration now established as a holistic approach to gain deep insights into enzymatic catalysis. Numerous studies have demonstrated the power of computer simulations in characterizing reaction pathways, transition states, substrate selectivity, product distribution, and dynamic conformational changes for various enzymes. Nevertheless, significant challenges remain in investigating the mechanisms of complex multistep reactions, large-scale conformational changes, and allosteric regulation. Beyond mechanistic studies, computational enzyme modeling has emerged as an essential tool for computer-aided enzyme design and the rational discovery of covalent drugs for targeted therapies. Overall, enzyme design/engineering and covalent drug development can greatly benefit from our understanding of the detailed mechanisms of enzymes, such as protein dynamics, entropy contributions, and allostery, as revealed by computational studies. Such a convergence of different research approaches is expected to continue, creating synergies in enzyme research. This review, by outlining the ever-expanding field of enzyme research, aims to provide guidance for future research directions and facilitate new developments in this important and evolving field.

The shortest path method (SPM) webserver for computational enzyme design

SPMweb is the online webserver of the Shortest Path Map (SPM) tool for identifying the key conformationally-relevant positions of a given enzyme structure and dynamics. The server is built on top of the DynaComm.py code and enables the calculation and visualization of the SPM pathways. SPMweb is easy-to-use as it only requires three input files: the three-dimensional structure of the protein of interest, and the two matrices (distance and correlation) previously computed from a Molecular Dynamics simulation. We provide in this publication information on how to generate the files for SPM construction even for non-expert users and discuss the most relevant parameters that can be modified. The tool is extremely fast (it takes less than one minute per job), thus allowing the rapid identification of distal positions connected to the active site pocket of the enzyme. SPM applications expand from computational enzyme design, especially if combined with other tools to identify the preferred substitution at the identified position, but also to rationalizing allosteric regulation, and even cryptic pocket identification for drug discovery. The simple user interface and setup make the SPM tool accessible to the whole scientific community. SPMweb is freely available for academia at http://spmosuna.com/.

Computational Protein Design - Where it goes?

Proteins have been playing a critical role in the regulation of diverse biological processes related to human life. With the increasing demand, functional proteins are sparse in this immense sequence space. Therefore, protein design has become an important task in various fields, including medicine, food, energy, materials, etc. Directed evolution has recently led to significant achievements. Molecular modification of proteins through directed evolution technology has significantly advanced the fields of enzyme engineering, metabolic engineering, medicine, and beyond. However, it is impossible to identify desirable sequences from a large number of synthetic sequences alone. As a result, computational methods, including data-driven machine learning and physics-based molecular modeling, have been introduced to protein engineering to produce more functional proteins. This review focuses on recent advances in computational protein design, highlighting the applicability of different approaches as well as their limitations.

A transfer-learning approach to predict antigen immunogenicity and T-cell receptor specificity

Antigen immunogenicity and the specificity of binding of T-cell receptors to antigens are key properties underlying effective immune responses. Here we propose diffRBM, an approach based on transfer learning and Restricted Boltzmann Machines, to build sequence-based predictive models of these properties. DiffRBM is designed to learn the distinctive patterns in amino-acid composition that, on the one hand, underlie the antigen's probability of triggering a response, and on the other hand the T-cell receptor's ability to bind to a given antigen. We show that the patterns learnt by diffRBM allow us to predict putative contact sites of the antigen-receptor complex. We also discriminate immunogenic and non-immunogenic antigens, antigen-specific and generic receptors, reaching performances that compare favorably to existing sequence-based predictors of antigen immunogenicity and T-cell receptor specificity.

Advances in cold-adapted enzymes derived from microorganisms

Cold-adapted enzymes, produced in cold-adapted organisms, are a class of enzyme with catalytic activity at low temperatures, high temperature sensitivity, and the ability to adapt to cold stimulation. These enzymes are largely derived from animals, plants, and microorganisms in polar areas, mountains, and the deep sea. With the rapid development of modern biotechnology, cold-adapted enzymes have been implemented in human and other animal food production, the protection and restoration of environments, and fundamental biological research, among other areas. Cold-adapted enzymes derived from microorganisms have attracted much attention because of their short production cycles, high yield, and simple separation and purification, compared with cold-adapted enzymes derived from plants and animals. In this review we discuss various types of cold-adapted enzyme from cold-adapted microorganisms, along with associated applications, catalytic mechanisms, and molecular modification methods, to establish foundation for the theoretical research and application of cold-adapted enzymes.

Protein Fitness Prediction Is Impacted by the Interplay of Language Models, Ensemble Learning, and Sampling Methods

Advances in machine learning (ML) and the availability of protein sequences via high-throughput sequencing techniques have transformed the ability to design novel diagnostic and therapeutic proteins. ML allows protein engineers to capture complex trends hidden within protein sequences that would otherwise be difficult to identify in the context of the immense and rugged protein fitness landscape. Despite this potential, there persists a need for guidance during the training and evaluation of ML methods over sequencing data. Two key challenges for training discriminative models and evaluating their performance include handling severely imbalanced datasets (e.g., few high-fitness proteins among an abundance of non-functional proteins) and selecting appropriate protein sequence representations (numerical encodings). Here, we present a framework for applying ML over assay-labeled datasets to elucidate the capacity of sampling techniques and protein encoding methods to improve binding affinity and thermal stability prediction tasks. For protein sequence representations, we incorporate two widely used methods (One-Hot encoding and physiochemical encoding) and two language-based methods (next-token prediction, UniRep; masked-token prediction, ESM). Elaboration on performance is provided over protein fitness, protein size, and sampling techniques. In addition, an ensemble of protein representation methods is generated to discover the contribution of distinct representations and improve the final prediction score. We then implement multiple criteria decision analysis (MCDA; TOPSIS with entropy weighting), using multiple metrics well-suited for imbalanced data, to ensure statistical rigor in ranking our methods. Within the context of these datasets, the synthetic minority oversampling technique (SMOTE) outperformed undersampling while encoding sequences with One-Hot, UniRep, and ESM representations. Moreover, ensemble learning increased the predictive performance of the affinity-based dataset by 4% compared to the best single-encoding candidate (F1-score = 97%), while ESM alone was rigorous enough in stability prediction (F1-score = 92%).

In the business of base editors: Evolution from bench to bedside

With the advent of recombinant DNA technology in the 1970s, the idea of using gene therapies to treat human genetic diseases captured the interest and imagination of scientists around the world. Years later, enabled largely by the development of CRISPR-based genome editing tools, the field has exploded, with academic labs, startup biotechnology companies, and large pharmaceutical corporations working in concert to develop life-changing therapeutics. In this Essay, we highlight base editing technologies and their development from bench to bedside. Base editing, first reported in 2016, is capable of installing C•G to T•A and A•T to G•C point mutations, while largely circumventing some of the pitfalls of traditional CRISPR/Cas9 gene editing. Despite their youth, these technologies have been widely used by both academic labs and therapeutics-based companies. Here, we provide an overview of the mechanics of base editing and its use in clinical trials.

Computational protein design repurposed to explore enzyme vitality and help predict antibiotic resistance

In response to antibiotics that inhibit a bacterial enzyme, resistance mutations inevitably arise. Predicting them ahead of time would aid target selection and drug design. The simplest resistance mechanism would be to reduce antibiotic binding without sacrificing too much substrate binding. The property that reflects this is the enzyme "vitality", defined here as the difference between the inhibitor and substrate binding free energies. To predict such mutations, we borrow methodology from computational protein design. We use a Monte Carlo exploration of mutation space and vitality changes, allowing us to rank thousands of mutations and identify ones that might provide resistance through the simple mechanism considered. As an illustration, we chose dihydrofolate reductase, an essential enzyme targeted by several antibiotics. We simulated its complexes with the inhibitor trimethoprim and the substrate dihydrofolate. 20 active site positions were mutated, or "redesigned" individually, then in pairs or quartets. We computed the resulting binding free energy and vitality changes. Out of seven known resistance mutations involving active site positions, five were correctly recovered. Ten positions exhibited mutations with significant predicted vitality gains. Direct couplings between designed positions were predicted to be small, which reduces the combinatorial complexity of the mutation space to be explored. It also suggests that over the course of evolution, resistance mutations involving several positions do not need the underlying point mutations to arise all at once: they can appear and become fixed one after the other.

Theory of Protein Charge Transfer: Electron Transfer between Tryptophan Residue and Active Site of Azurin

One reaction step in the conductivity relay of azurin, electron transfer between the Cu-based active site and the tryptophan residue, is studied theoretically and by classical molecular dynamics simulations. Oxidation of tryptophan results in electrowetting of this residue. This structural change makes the free energy surfaces of electron transfer nonparabolic as described by the Q-model of electron transfer. We analyze the medium dynamical effect on protein electron transfer produced by coupled Stokes-shift dynamics and the dynamics of the donor-acceptor distance modulating electron tunneling. The equilibrium donor-acceptor distance falls in the plateau region of the rate constant, where it is determined by the protein-water dynamics, and the probability of electron tunneling does not affect the rate. The crossover distance found here puts most intraprotein electron-transfer reactions under the umbrella of dynamical control. The crossover between the medium-controlled and tunneling-controlled kinetics is combined with the effect of the protein-water medium on the activation barrier to formulate principles of tunability of protein-based charge-transfer chains. The main principle in optimizing the activation barrier is the departure from the Gaussian-Gibbsian statistics of fluctuations promoting activated transitions. This is achieved either by incomplete (nonergodic) sampling, breaking the link between the Stokes-shift and variance reorganization energies, or through wetting-induced structural changes of the enzyme's active site.

Novel Production Methods of Polyhydroxyalkanoates and Their Innovative Uses in Biomedicine and Industry

Polyhydroxyalkanoate (PHA), a biodegradable polymer obtained from microorganisms and plants, have been widely used in biomedical applications and devices, such as sutures, cardiac valves, bone scaffold, and drug delivery of compounds with pharmaceutical interests, as well as in food packaging. This review focuses on the use of polyhydroxyalkanoates beyond the most common uses, aiming to inform about the potential uses of the biopolymer as a biosensor, cosmetics, drug delivery, flame retardancy, and electrospinning, among other interesting uses. The novel applications are based on the production and composition of the polymer, which can be modified by genetic engineering, a semi-synthetic approach, by changing feeding carbon sources and/or supplement addition, among others. The future of PHA is promising, and despite its production costs being higher than petroleum-based plastics, tools given by synthetic biology, bioinformatics, and machine learning, among others, have allowed for great production yields, monomer and polymer functionalization, stability, and versatility, a key feature to increase the uses of this interesting family of polymers.

Protein Function Analysis through Machine Learning

Machine learning (ML) has been an important arsenal in computational biology used to elucidate protein function for decades. With the recent burgeoning of novel ML methods and applications, new ML approaches have been incorporated into many areas of computational biology dealing with protein function. We examine how ML has been integrated into a wide range of computational models to improve prediction accuracy and gain a better understanding of protein function. The applications discussed are protein structure prediction, protein engineering using sequence modifications to achieve stability and druggability characteristics, molecular docking in terms of protein–ligand binding, including allosteric effects, protein–protein interactions and protein-centric drug discovery. To quantify the mechanisms underlying protein function, a holistic approach that takes structure, flexibility, stability, and dynamics into account is required, as these aspects become inseparable through their interdependence. Another key component of protein function is conformational dynamics, which often manifest as protein kinetics. Computational methods that use ML to generate representative conformational ensembles and quantify differences in conformational ensembles important for function are included in this review. Future opportunities are highlighted for each of these topics.

Machine Learning Guided Batched Design of a Bacterial Ribosome Binding Site

Optimization of gene expression levels is an essential part of the organism design process. Fine control of this process can be achieved by engineering transcription and translation control elements, including the ribosome binding site (RBS). Unfortunately, the design of specific genetic parts remains challenging because of the lack of reliable design methods. To address this problem, we have created a machine learning guided Design-Build-Test-Learn (DBTL) cycle for the experimental design of bacterial RBSs to demonstrate how small genetic parts can be reliably designed using relatively small, high-quality data sets. We used Gaussian Process Regression for the Learn phase of the cycle and the Upper Confidence Bound multiarmed bandit algorithm for the Design of genetic variants to be tested in vivo. We have integrated these machine learning algorithms with laboratory automation and high-throughput processes for reliable data generation. Notably, by Testing a total of 450 RBS variants in four DBTL cycles, we have experimentally validated RBSs with high translation initiation rates equaling or exceeding our benchmark RBS by up to 34%. Overall, our results show that machine learning is a powerful tool for designing RBSs, and they pave the way toward more complicated genetic devices.

Evolution of tunnels in α/β-hydrolase fold proteins—What can we learn from studying epoxide hydrolases?

The evolutionary variability of a protein’s residues is highly dependent on protein region and function. Solvent-exposed residues, excluding those at interaction interfaces, are more variable than buried residues whereas active site residues are considered to be conserved. The abovementioned rules apply also to α/β-hydrolase fold proteins—one of the oldest and the biggest superfamily of enzymes with buried active sites equipped with tunnels linking the reaction site with the exterior. We selected soluble epoxide hydrolases as representative of this family to conduct the first systematic study on the evolution of tunnels. We hypothesised that tunnels are lined by mostly conserved residues, and are equipped with a number of specific variable residues that are able to respond to evolutionary pressure. The hypothesis was confirmed, and we suggested a general and detailed way of the tunnels’ evolution analysis based on entropy values calculated for tunnels’ residues. We also found three different cases of entropy distribution among tunnel-lining residues. These observations can be applied for protein reengineering mimicking the natural evolution process. We propose a ‘perforation’ mechanism for new tunnels design via the merging of internal cavities or protein surface perforation. Based on the literature data, such a strategy of new tunnel design could significantly improve the enzyme’s performance and can be applied widely for enzymes with buried active sites.

So far very little is known about proteins tunnels evolution. The goal of this study is to evaluate the evolution of tunnels in the family of soluble epoxide hydrolases—representatives of numerous α/β-hydrolase fold enzymes. As a result two types of tunnels evolution analysis were proposed (a general and a detailed approach), as well as a ‘perforation’ mechanism which can mimic native evolution in proteins and can be used as an additional strategy for enzymes redesign.

Exploring the Effect of Enhanced Sampling on Protein Stability Prediction

Changes in protein stability due to side-chain mutations are evaluated by alchemical free-energy calculations based on classical molecular dynamics (MD) simulations in explicit solvent using the GROMOS force field. Three proteins of different complexity with a total number of 93 single-point mutations are analyzed, and the relative free-energy differences are discussed with respect to configurational sampling and (dis)agreement with experimental data. For the smallest protein studied, a 34-residue WW domain, the starting structure dependence of the alchemical free-energy changes, is discussed in detail. Deviations from previous simulations for the two other proteins are shown to result from insufficient sampling in the earlier studies. Hamiltonian replica exchange in combination with multiple starting structures and sufficient sampling time of more than 100 ns per intermediate alchemical state is required in some cases to reach convergence.

Nondegenerate Saturation Mutagenesis: Library Construction and Analysis via MAX and ProxiMAX Randomization

Protein engineering can enhance desirable features and improve performance outside of the natural context. Several strategies have been adopted over the years for gene diversification, and engineering of modular proteins in particular is most effective when a high-throughput, library-based approach is employed. Nondegenerate saturation mutagenesis plays a dynamic role in engineering proteins by targeting multiple codons to generate massively diverse gene libraries. Herein, we describe the nondegenerate saturation mutagenesis techniques that we have developed for contiguous (ProxiMAX) and noncontiguous (MAX) randomized codon generation to create precisely defined, diverse gene libraries, in the context of other fully nondegenerate strategies. ProxiMAX randomization comprises saturation cycling with repeated cycles of blunt-ended ligation, type IIS restriction, and PCR amplification, and is now a commercially automated process predominantly used for antibody library generation. MAX randomization encompasses a manual process of selective hybridisation between individual custom oligonucleotide mixes and a conventionally randomized template and is principally employed in the research laboratory setting, to engineer alpha helical proteins and active sites of enzymes. DNA libraries generated using either technology create high-throughput amino acid substitutions via codon randomization, to generate genetically diverse clones.

Efficient generative modeling of protein sequences using simple autoregressive models

Generative models emerge as promising candidates for novel sequence-data driven approaches to protein design, and for the extraction of structural and functional information about proteins deeply hidden in rapidly growing sequence databases. Here we propose simple autoregressive models as highly accurate but computationally efficient generative sequence models. We show that they perform similarly to existing approaches based on Boltzmann machines or deep generative models, but at a substantially lower computational cost (by a factor between 10² and 10³). Furthermore, the simple structure of our models has distinctive mathematical advantages, which translate into an improved applicability in sequence generation and evaluation. Within these models, we can easily estimate both the probability of a given sequence, and, using the model's entropy, the size of the functional sequence space related to a specific protein family. In the example of response regulators, we find a huge number of ca. 10⁶⁸ possible sequences, which nevertheless constitute only the astronomically small fraction 10^-80 of all amino-acid sequences of the same length. These findings illustrate the potential and the difficulty in exploring sequence space via generative sequence models.

Directed Evolution: Methodologies and Applications

Directed evolution aims to expedite the natural evolution process of biological molecules and systems in a test tube through iterative rounds of gene diversifications and library screening/selection. It has become one of the most powerful and widespread tools for engineering improved or novel functions in proteins, metabolic pathways, and even whole genomes. This review describes the commonly used gene diversification strategies, screening/selection methods, and recently developed continuous evolution strategies for directed evolution. Moreover, we highlight some representative applications of directed evolution in engineering nucleic acids, proteins, pathways, genetic circuits, viruses, and whole cells. Finally, we discuss the challenges and future perspectives in directed evolution.

An in vivo selection system with tightly regulated gene expression enables directed evolution of highly efficient enzymes

In vivo selection systems are powerful tools for directed evolution of enzymes. The selection pressure of the systems can be tuned by regulating the expression levels of the catalysts. In this work, we engineered a selection system for laboratory evolution of highly active enzymes by incorporating a translationally suppressing cis repressor as well as an inducible promoter to impart stringent and tunable selection pressure. We demonstrated the utility of our selection system by performing directed evolution experiments using TEM β-lactamase as the model enzyme. Five evolutionary rounds afforded a highly active variant exhibiting 440-fold improvement in catalytic efficiency. We also showed that, without the cis repressor, the selection system cannot provide sufficient selection pressure required for evolving highly efficient TEM β-lactamase. The selection system should be applicable for the exploration of catalytic perfection of a wide range of enzymes.

CRISPR/Cas9-based directed evolution in mammalian cells

An increasingly powerful set of new CRISPR/Cas-based methods is becoming available for directed evolution of proteins in mammalian cells. Although in vitro techniques or microbial expression systems have been dominating directed evolution, there are now promising approaches to diversify proteins in mammalian cells in situ. This can be achieved by simple indel mutagenesis or more sophisticated homology repair mechanisms for cassette mutagenesis of coding sequences. Cas9 variant fusions to base editors and other effectors pose another promising way to introduce diversity into proteins. CRISPR/Cas9-based directed evolution in mammalian cells opens a new exciting era of discovery for the many classes of proteins for which a mammalian cellular context is preferable.

100th Anniversary of Macromolecular Science Viewpoint: Data-Driven Protein Design

The design of synthetic proteins with the desired function is a long-standing goal in biomolecular science, with broad applications in biochemical engineering, agriculture, medicine, and public health. Rational de novo design and experimental directed evolution have achieved remarkable successes but are challenged by the requirement to find functional "needles" in the vast "haystack" of protein sequence space. Data-driven models for fitness landscapes provide a predictive map between protein sequence and function and can prospectively identify functional candidates for experimental testing to greatly improve the efficiency of this search. This Viewpoint reviews the applications of machine learning and, in particular, deep learning as part of data-driven protein engineering platforms. We highlight recent successes, review promising computational methodologies, and provide an outlook on future challenges and opportunities. The article is written for a broad audience comprising both polymer and protein scientists and computer and data scientists interested in an up-to-date review of recent innovations and opportunities in this rapidly evolving field.

Genetically Encodable Scaffolds for Optimizing Enzyme Function

Enzyme engineering is an indispensable tool in the field of synthetic biology, where enzymes are challenged to carry out novel or improved functions. Achieving these goals sometimes goes beyond modifying the primary sequence of the enzyme itself. The use of protein or nucleic acid scaffolds to enhance enzyme properties has been reported for applications such as microbial production of chemicals, biosensor development and bioremediation. Key advantages of using these assemblies include optimizing reaction conditions, improving metabolic flux and increasing enzyme stability. This review summarizes recent trends in utilizing genetically encodable scaffolds, developed in line with synthetic biology methodologies, to complement the purposeful deployment of enzymes. Current molecular tools for constructing these synthetic enzyme-scaffold systems are also highlighted.

Evolving the naturally compromised chorismate mutase from Mycobacterium tuberculosis to top performance

Chorismate mutase (CM), an essential enzyme at the branch-point of the shikimate pathway, is required for the biosynthesis of phenylalanine and tyrosine in bacteria, archaea, plants, and fungi. MtCM, the CM from Mycobacterium tuberculosis, has less than 1% of the catalytic efficiency of a typical natural CM and requires complex formation with 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase for high activity. To explore the full potential of MtCM for catalyzing its native reaction, we applied diverse iterative cycles of mutagenesis and selection, thereby raising k_cat/K_m 270-fold to 5 × 10⁵m⁻¹s⁻¹, which is even higher than for the complex. Moreover, the evolutionarily optimized autonomous MtCM, which had 11 of its 90 amino acids exchanged, was stabilized compared with its progenitor, as indicated by a 9 °C increase in melting temperature. The 1.5 Å crystal structure of the top-evolved MtCM variant reveals the molecular underpinnings of this activity boost. Some acquired residues (e.g. Pro⁵² and Asp⁵⁵) are conserved in naturally efficient CMs, but most of them lie beyond the active site. Our evolutionary trajectories reached a plateau at the level of the best natural enzymes, suggesting that we have exhausted the potential of MtCM. Taken together, these findings show that the scaffold of MtCM, which naturally evolved for mediocrity to enable inter-enzyme allosteric regulation of the shikimate pathway, is inherently capable of high activity.

Uncovering key residues responsible for the thermostability of a thermophilic 1,3(4)-β-d-glucanase from Nong flavor Daqu by rational design

Fungal 1,3(4)-β-D-glucanases were usually applied in brewing and feedstuff industries, however, the thermostability limits the most their application. The characterized 1,3(4)-β-D-glucanase (NFEg16A) from Chinese Nong-flavor (NF) Daqu showed the highest thermostability among GH16 fungal 1,3(4)-β-D-glucanases, with half-lives of thermal inactivation (t_1/2) of 44.9 min at 90 °C, so multiple rational designs were used to identify the key residues for its thermostability. Based on protein sequence and 3D structure analyses around the catalytic regions. Nine site-mutants were constructed, among which N173Y and S187A were identified as the most thermotolerant and thermolabile ones, with t_1/2 values of 61 min and 14.0 min at 90 °C, respectively. Therefore, N173 and S187 were then selected as "hotspots" for site-saturation mutagenesis. Interestingly, most of the N173 and S187 variants exhibited a similar thermostability to that of N173Y and S187A, respectively, confirming their different roles in the thermostability of NFEg16A. In addition, each S187A and its surrounding substitutions (D144 N and T164 N) was independently detrimental to the thermostability of NFEg16A, since the t_1/2 (90 °C) of S187A, D144 N and T164 N were 14.0 min, 20.6 min and 27.2 min, respectively. Surprisingly, combinatorial substitution of S187A with D144 N or T164 N showed positive effects on the thermostability, with the increase of t_1/2 (90 °C) to 30.9 min and 63.5 min for S187A-D144 N and S187A-T164 N, respectively. More importantly, S187A-T164 N showed higher thermostability than that of wild type. In short, we successfully identified two key sites and their surrounding residues in response to the thermostability of NFEg16A and further improved its thermostability by several rational designs. These findings could be used for the protein engineering of homologous 1,3(4)-β-D-glucanases, as well as other enzyme family members with high similarities.

Arming Yourself for The In Silico Protein Design Revolution

Proteins mediate many essential processes of life to a degree of functional precision unmatched by any synthetic device. While engineered proteins are currently used in biotech, food, biomedicine, and material technology-based industries, the true potential of proteins is practically untapped. The emerging field of in silico protein design is predicted to provide the next quantum leap in the biotech industry. Having predictive control over protein function and the ability to redefine these functions have driven the field of protein engineering into an era of unprecedented development. This article provides a holistic analysis of protein design R&D (current state-of-the-art tools and knowhow) and commercial landscape, as well as a one-stop-shop profile of in silico protein design technology for biotechnology stakeholders.

Designing peptide nanoparticles for efficient brain delivery

The targeted delivery of therapeutic compounds to the brain is arguably the most significant open problem in drug delivery today. Nanoparticles (NPs) based on peptides and designed using the emerging principles of molecular engineering show enormous promise in overcoming many of the barriers to brain delivery faced by NPs made of more traditional materials. However, shortcomings in our understanding of peptide self-assembly and blood-brain barrier (BBB) transport mechanisms pose significant obstacles to progress in this area. In this review, we discuss recent work in engineering peptide nanocarriers for the delivery of therapeutic compounds to the brain: from synthesis, to self-assembly, to in vivo studies, as well as discussing in detail the biological hurdles that a nanoparticle must overcome to reach the brain.

Deep Dive into Machine Learning Models for Protein Engineering

Protein redesign and engineering has become an important task in pharmaceutical research and development. Recent advances in technology have enabled efficient protein redesign by mimicking natural evolutionary mutation, selection, and amplification steps in the laboratory environment. For any given protein, the number of possible mutations is astronomical. It is impractical to synthesize all sequences or even to investigate all functionally interesting variants. Recently, there has been an increased interest in using machine learning to assist protein redesign, since prediction models can be used to virtually screen a large number of novel sequences. However, many state-of-the-art machine learning models, especially deep learning models, have not been extensively explored. Moreover, only a small selection of protein sequence descriptors has been considered. In this work, the performance of prediction models built using an array of machine learning methods and protein descriptor types, including two novel, single amino acid descriptors and one structure-based three-dimensional descriptor, is benchmarked. The predictions were evaluated on a diverse collection of public and proprietary data sets, using a variety of evaluation metrics. The results of this comparison suggest that Convolution Neural Network models built with amino acid property descriptors are the most widely applicable to the types of protein redesign problems faced in the pharmaceutical industry.

Computer-Assisted Recombination (CompassR) Teaches us How to Recombine Beneficial Substitutions from Directed Evolution Campaigns

A main remaining challenge in protein engineering is how to recombine beneficial substitutions. Systematic recombination studies show that poorly performing variants are usually obtained after recombination of 3 to 4 beneficial substitutions. This limits researchers in exploiting nature's potential in generating better enzymes. The Computer-assisted Recombination (CompassR) strategy provides a selection guide for beneficial substitutions that can be recombined to gradually improve enzyme performance by analysis of the relative free energy of folding (ΔΔGfold ). The performance of CompassR was evaluated by analysis of 84 recombinants located on 13 positions of Bacillus subtilis lipase A. The finally obtained variant F17S/V54K/D64N/D91E had a 2.7-fold improved specific activity in 18.3 % (v/v) 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]). In essence, the deducted CompassR rule allows recombination of beneficial substitutions in an iterative manner and empowers researchers to generate better enzymes in a time-efficient manner.

Protein Thermal Stability Engineering Using HoTMuSiC

The rational design of enzymes is a challenging research field, which plays an important role in the optimization of a wide series of biotechnological processes. Computational approaches allow screening all possible amino acid substitutions in a target protein and to identify a subset likely to have the desired properties. They can thus be used to guide and restrict the huge, time-consuming search in sequence space to reach protein optimality. Here we present HoTMuSiC, a tool that predicts the impact of point mutations on the protein melting temperature, which uses the experimental or modeled protein structure as sole input and is available at the dezyme.com website. Its main advantages include accuracy and speed, which makes it a perfect instrument for thermal stability engineering projects aiming at designing new proteins that feature increased heat resistance or remain active and stable in nonphysiological conditions. We set up a HoTMuSiC-based pipeline, which uses additional information to avoid mutations of functionally important residues, identified as being too well conserved among homologous proteins or too close to annotated functional sites. The efficiency of this pipeline is successfully demonstrated on Rhizomucor miehei lipase.

High-throughput droplet-based microfluidics for directed evolution of enzymes

Natural enzymes have evolved over millions of years to allow for their effective operation within specific environments. However, it is significant to note that despite their wide structural and chemical diversity, relatively few natural enzymes have been successfully applied to industrial processes. To address this limitation, directed evolution (DE) (a method that mimics the process of natural selection to evolve proteins toward a user‐defined goal) coupled with droplet‐based microfluidics allows the detailed analysis of millions of enzyme variants on ultra‐short timescales, and thus the design of novel enzymes with bespoke properties. In this review, we aim at presenting the development of DE over the last years and highlighting the most important advancements in droplet‐based microfluidics, made in this context towards the high‐throughput demands of enzyme optimization. Specifically, an overview of the range of microfluidic unit operations available for the construction of DE platforms is provided, focusing on their suitability and benefits for cell‐based assays, as in the case of directed evolution experimentations.

GeneORator: An Effective Strategy for Navigating Protein Sequence Space More Efficiently through Boolean OR-Type DNA Libraries

Directed evolution requires the creation of genetic diversity and subsequent screening or selection for improved variants. For DNA mutagenesis, conventional site-directed methods implicitly utilize the Boolean AND operator (creating all mutations simultaneously), producing a combinatorial explosion in the number of genetic variants as the number of mutations increases. We introduce GeneORator, a novel strategy for creating DNA libraries based on the Boolean logical OR operator. Here, a single library is divided into many subsets, each containing different combinations of the desired mutations. Consequently, the effect of adding more mutations on the number of genetic combinations is additive (Boolean OR logic) and not exponential (AND logic). We demonstrate this strategy with large-scale mutagenesis studies, using monoamine oxidase-N ( Aspergillus niger) as the exemplar target. First, we mutated every residue in the secondary structure-containing regions (276 out of a total 495 amino acids) to screen for improvements in k_cat. Second, combinatorial OR-type libraries permitted screening of diverse mutation combinations in the enzyme active site to detect activity toward novel substrates. In both examples, OR-type libraries effectively reduced the number of variants searched up to 10¹⁰-fold, dramatically reducing the screening effort required to discover variants with improved and/or novel activity. Importantly, this approach enables the screening of a greater diversity of mutation combinations, accessing a larger area of a protein's sequence space. OR-type libraries can be applied to any biological engineering objective requiring DNA mutagenesis, and the approach has wide ranging applications in, for example, enzyme engineering, antibody engineering, and synthetic biology.

Virus-Inspired Function in Engineered Protein Cages

The structural and functional diversity of proteins combined with their genetic programmability has made them indispensable modern materials. Well-defined, hollow protein capsules have proven to be particularly useful due to their ability to compartmentalize macromolecules and chemical processes. To this end, viral capsids are common scaffolds and have been successfully repurposed to produce a suite of practical protein-based nanotechnologies. Recently, the recapitulation of viromimetic function in protein cages of nonviral origin has emerged as a strategy to both complement physical studies of natural viruses and produce useful scaffolds for diverse applications. In this perspective, we review recent progress toward generation of virus-like behavior in nonviral protein cages through rational engineering and directed evolution. These artificial systems can aid our understanding of the emergence of viruses from existing cellular components, as well as provide alternative approaches to tackle current problems, and open up new opportunities, in medicine and biotechnology.

Consensus sequence design as a general strategy to create hyperstable, biologically active proteins

Consensus sequence design offers a promising strategy for designing proteins of high stability while retaining biological activity since it draws upon an evolutionary history in which residues important for both stability and function are likely to be conserved. Although there have been several reports of successful consensus design of individual targets, it is unclear from these anecdotal studies how often this approach succeeds and how often it fails. Here, we attempt to assess generality by designing consensus sequences for a set of six protein families with a range of chain lengths, structures, and activities. We characterize the resulting consensus proteins for stability, structure, and biological activities in an unbiased way. We find that all six consensus proteins adopt cooperatively folded structures in solution. Strikingly, four of six of these consensus proteins show increased thermodynamic stability over naturally occurring homologs. Each consensus protein tested for function maintained at least partial biological activity. Although peptide binding affinity by a consensus-designed SH3 is rather low, K _m values for consensus enzymes are similar to values from extant homologs. Although consensus enzymes are slower than extant homologs at low temperature, they are faster than some thermophilic enzymes at high temperature. An analysis of sequence properties shows consensus proteins to be enriched in charged residues, and rarified in uncharged polar residues. Sequence differences between consensus and extant homologs are predominantly located at weakly conserved surface residues, highlighting the importance of these residues in the success of the consensus strategy.

Design Methodologies and the Limits of the Engineering-Dominated Conception of Synthetic Biology

Synthetic biology is described as a new field of biotechnology that models itself on engineering sciences. However, this view of synthetic biology as an engineering field has received criticism, and both biologists and philosophers have argued for a more nuanced and heterogeneous understanding of the field. This paper elaborates the heterogeneity of synthetic biology by clarifying the role of design and the variability of design methodologies in synthetic biology. I focus on two prominent design methodologies: rational design and directed evolution. Rational design resembles the design methodology of traditional engineering sciences. However, it is often replaced and complemented by the more biologically-inspired method of directed evolution, which models itself on natural evolution. These two approaches take philosophically different stances to the design of biological systems. Rational design aims to make biological systems more machine-like, whereas directed evolution utilizes variation and emergent features of living systems. I provide an analysis of the methodological basis of these design approaches, and highlight important methodological differences between them. By analyzing the respective benefits and limitations of these approaches, I argue against the engineering-dominated conception of synthetic biology and its "methodological monism", where the rational design approach is taken as the default design methodology. Alternative design methodologies, like directed evolution, should be considered as complementary, not competitive, to rational design.

In Silico-Directed Evolution Using CADEE

Recent years have seen an explosion of interest in both sequence- and structure-based approaches toward in silico-directed evolution. We recently developed a novel computational toolkit, CADEE, which facilitates the computer-aided directed evolution of enzymes. Our initial work (Amrein et al., IUCrJ 4:50-64, 2017) presented a pedagogical example of the application of CADEE to triosephosphate isomerase, to illustrate the CADEE workflow. In this contribution, we describe this workflow in detail, including code input/output snippets, in order to allow users to set up and execute CADEE simulations on any system of interest.

Affinity Improvement of a Cancer-Targeted Antibody through Alanine-Induced Adjustment of Antigen-Antibody Interface

To investigate favorable single amino acid substitutions that improve antigen-antibody interactions, alanine (Ala) mutagenesis scanning of the interfacial residues of a cancer-targeted antibody, B5209B, was performed based on X-ray crystallography analysis. Two substitutions were shown to significantly enhance the binding affinity for the antigen, by up to 30-fold. One substitution improved the affinity by a gain of binding enthalpy, whereas the other substitution improved the affinity by a gain of binding entropy. Molecular dynamics simulations showed that the enthalpic improvement could be attributed to the stabilization of distant salt bridges located at the periphery of the antigen-antibody interface. The entropic improvement was due to the release of water molecules that were stably trapped in the antigen-antibody interface of the wild-type antibody. Importantly, these effects of the Ala substitutions were caused by subtle adjustments of the binding interface. These results will be helpful to design high-affinity antibodies with avoiding entropy-enthalpy compensation.

Biomolecular Assemblies: Moving from Observation to Predictive Design

Biomolecular assembly is a key driving force in nearly all life processes, providing structure, information storage, and communication within cells and at the whole organism level. These assembly processes rely on precise interactions between functional groups on nucleic acids, proteins, carbohydrates, and small molecules, and can be fine-tuned to span a range of time, length, and complexity scales. Recognizing the power of these motifs, researchers have sought to emulate and engineer biomolecular assemblies in the laboratory, with goals ranging from modulating cellular function to the creation of new polymeric materials. In most cases, engineering efforts are inspired or informed by understanding the structure and properties of naturally occurring assemblies, which has in turn fueled the development of predictive models that enable computational design of novel assemblies. This Review will focus on selected examples of protein assemblies, highlighting the story arc from initial discovery of an assembly, through initial engineering attempts, toward the ultimate goal of predictive design. The aim of this Review is to highlight areas where significant progress has been made, as well as to outline remaining challenges, as solving these challenges will be the key that unlocks the full power of biomolecules for advances in technology and medicine.

Imaging-Based Screening Platform Assists Protein Engineering

Protein engineering involves generating and screening large numbers of variants for desired properties. While modern DNA technology has made it easy to create protein diversity on the DNA level, the selection and validation of candidate proteins from large libraries remains a challenge. We built a screening platform that integrates high-quality fluorescence-based image analysis and robotic picking of bacterial colonies. It allows tracking each individual colony in a large population and collecting quantitative information on library composition during the protein evolution process. We demonstrate the power of the screening platform by optimizing a dim far-red-emitting fluorescent protein whose brightness increased several fold using iterative cycles of mutagenesis and platform-based screening. The resulting protein variant mCarmine is useful for imaging cells and structures within live tissue as well as for molecular tagging. Overall, the platform presented provides powerful, flexible, and low-cost instrumentation to accelerate many fluorescence-based protein optimization projects.

Mutational Basin-Hopping: Combined Structure and Sequence Optimization for Biomolecules

The study of energy landscapes has led to a good understanding of how and why proteins and nucleic acids adopt their native structure. Through evolution, sequences have adapted until they exhibit a strongly funneled energy landscape, stabilizing the native fold. Design of artificial biomolecules faces the challenge of creating similar stable, minimally frustrated, and functional sequences. Here we present a biminimization approach, mutational basin-hopping, in which we simultaneously use global optimization to optimize the energy and a target function describing a desired property of the system. This optimization of structure and sequence is a generalized basin-hopping method and produces an efficient design process, which can target properties such as binding affinity or solubility.

Engineering enzyme microenvironments for enhanced biocatalysis

Protein engineering provides a means to alter protein structure leading to new functions. Much work has focused on the engineering of enzyme active sites to enhance catalytic activity, however there is an increasing trend towards engineering other aspects of biocatalysts as these efforts can also lead to useful improvements. This tutorial discusses recent advances in engineering an enzyme's local chemical and physical environment, with the goal of enhancing enzyme reaction kinetics, substrate selectivity, and activity in harsh conditions (e.g., low or high pH). By introducing stimuli-responsiveness to these enzyme modifications, dynamic control of activity also becomes possible. These new biomolecular and protein engineering techniques are separate and independent from traditional active site engineering and can therefore be applied synergistically to create new biocatalyst technologies with novel functions.

Engineered Metalloenzymes with Non-Canonical Coordination Environments

Nature employs a limited number of genetically encoded, metal-coordinating residues to create metalloenzymes with diverse structures and functions. Engineered components of the cellular translation machinery can now be exploited to encode non-canonical ligands with user-defined electronic and structural properties. This ability to install "chemically programmed" ligands into proteins can provide powerful chemical probes of metalloenzyme mechanism and presents excellent opportunities to create metalloprotein catalysts with augmented properties and novel activities. In this Concept article, we provide an overview of several recent studies describing the creation of engineered metalloenzymes with interesting catalytic properties, and reveal how characterization of these systems has advanced our understanding of nature's bioinorganic mechanisms. We also highlight how powerful laboratory evolution protocols can be readily adapted to allow optimization of metalloenzymes with non-canonical ligands. This approach combines beneficial features of small molecule and protein catalysis by allowing the installation of a greater variety of local metal coordination environments into evolvable protein scaffolds, and holds great promise for the future creation of powerful metalloprotein catalysts for a host of synthetically valuable transformations.

Blue-Light Receptors for Optogenetics

Sensory photoreceptors underpin light-dependent adaptations of organismal physiology, development, and behavior in nature. Adapted for optogenetics, sensory photoreceptors become genetically encoded actuators and reporters to enable the noninvasive, spatiotemporally accurate and reversible control by light of cellular processes. Rooted in a mechanistic understanding of natural photoreceptors, artificial photoreceptors with customized light-gated function have been engineered that greatly expand the scope of optogenetics beyond the original application of light-controlled ion flow. As we survey presently, UV/blue-light-sensitive photoreceptors have particularly allowed optogenetics to transcend its initial neuroscience applications by unlocking numerous additional cellular processes and parameters for optogenetic intervention, including gene expression, DNA recombination, subcellular localization, cytoskeleton dynamics, intracellular protein stability, signal transduction cascades, apoptosis, and enzyme activity. The engineering of novel photoreceptors benefits from powerful and reusable design strategies, most importantly light-dependent protein association and (un)folding reactions. Additionally, modified versions of these same sensory photoreceptors serve as fluorescent proteins and generators of singlet oxygen, thereby further enriching the optogenetic toolkit. The available and upcoming UV/blue-light-sensitive actuators and reporters enable the detailed and quantitative interrogation of cellular signal networks and processes in increasingly more precise and illuminating manners.