Functional and informatics analysis enables glycosyltransferase activity prediction

GASP: A Pan-Specific Predictor of Family 1 Glycosyltransferase Acceptor Specificity Enabled by a Pipeline for Substrate Feature Generation and Large-Scale Experimental Screening

Glycosylation represents a major chemical challenge; while it is one of the most common reactions in Nature, conventional chemistry struggles with stereochemistry, regioselectivity, and solubility issues. In contrast, family 1 glycosyltransferase (GT1) enzymes can glycosylate virtually any given nucleophilic group with perfect control over stereochemistry and regioselectivity. However, the appropriate catalyst for a given reaction needs to be identified among the tens of thousands of available sequences. Here, we present the glycosyltransferase acceptor specificity predictor (GASP) model, a data-driven approach to the identification of reactive GT1:acceptor pairs. We trained a random forest-based acceptor predictor on literature data and validated it on independent in-house generated data on 1001 GT1:acceptor pairs, obtaining an AUROC of 0.79 and a balanced accuracy of 72%. The performance was stable even in the case of completely new GT1s and acceptors not present in the training data set, highlighting the pan-specificity of GASP. Moreover, the model is capable of parsing all known GT1 sequences, as well as all chemicals, the latter through a pipeline for the generation of 153 chemical features for a given molecule taking the CID or SMILES as input (freely available at https://github.com/degnbol/GASP). To investigate the power of GASP, the model prediction probability scores were compared to GT1 substrate conversion yields from a newly published data set, with the top 50% of GASP predictions corresponding to reactions with >50% synthetic yields. The model was also tested in two comparative case studies: glycosylation of the antihelminth drug niclosamide and the plant defensive compound DIBOA. In the first study, the model achieved an 83% hit rate, outperforming a hit rate of 53% from a random selection assay. In the second case study, the hit rate of GASP was 50%, and while being lower than the hit rate of 83% using expert-selected enzymes, it provides a reasonable performance for the cases when an expert opinion is unavailable. The hierarchal importance of the generated chemical features was investigated by negative feature selection, revealing properties related to cyclization and atom hybridization status to be the most important characteristics for accurate prediction. Our study provides a GT1:acceptor predictor which can be trained on other data sets enabled by the automated feature generation pipelines. We also release the new in-house generated data set used for testing of GASP to facilitate the future development of GT1 activity predictors and their robust benchmarking.

Descriptor-augmented machine learning for enzyme-chemical interaction predictions

Descriptors play a pivotal role in enzyme design for the greener synthesis of biochemicals, as they could characterize enzymes and chemicals from the physicochemical and evolutionary perspective. This study examined the effects of various descriptors on the performance of Random Forest model used for enzyme-chemical relationships prediction. We curated activity data of seven specific enzyme families from the literature and developed the pipeline for evaluation the machine learning model performance using 10-fold cross-validation. The influence of protein and chemical descriptors was assessed in three scenarios, which were predicting the activity of unknown relations between known enzymes and known chemicals (new relationship evaluation), predicting the activity of novel enzymes on known chemicals (new enzyme evaluation), and predicting the activity of new chemicals on known enzymes (new chemical evaluation). The results showed that protein descriptors significantly enhanced the classification performance of model on new enzyme evaluation in three out of the seven datasets with the greatest number of enzymes, whereas chemical descriptors appear no effect. A variety of sequence-based and structure-based protein descriptors were constructed, among which the esm-2 descriptor achieved the best results. Using enzyme families as labels showed that descriptors could cluster proteins well, which could explain the contributions of descriptors to the machine learning model. As a counterpart, in the new chemical evaluation, chemical descriptors made significant improvement in four out of the seven datasets, while protein descriptors appear no effect. We attempted to evaluate the generalization ability of the model by correlating the statistics of the datasets with the performance of the models. The results showed that datasets with higher sequence similarity were more likely to get better results in the new enzyme evaluation and datasets with more enzymes were more likely beneficial from the protein descriptor strategy. This work provides guidance for the development of machine learning models for specific enzyme families.

A multimodal Transformer Network for protein-small molecule interactions enhances predictions of kinase inhibition and enzyme-substrate relationships

The activities of most enzymes and drugs depend on interactions between proteins and small molecules. Accurate prediction of these interactions could greatly accelerate pharmaceutical and biotechnological research. Current machine learning models designed for this task have a limited ability to generalize beyond the proteins used for training. This limitation is likely due to a lack of information exchange between the protein and the small molecule during the generation of the required numerical representations. Here, we introduce ProSmith, a machine learning framework that employs a multimodal Transformer Network to simultaneously process protein amino acid sequences and small molecule strings in the same input. This approach facilitates the exchange of all relevant information between the two molecule types during the computation of their numerical representations, allowing the model to account for their structural and functional interactions. Our final model combines gradient boosting predictions based on the resulting multimodal Transformer Network with independent predictions based on separate deep learning representations of the proteins and small molecules. The resulting predictions outperform recently published state-of-the-art models for predicting protein-small molecule interactions across three diverse tasks: predicting kinase inhibitions; inferring potential substrates for enzymes; and predicting Michaelis constants KM. The Python code provided can be used to easily implement and improve machine learning predictions involving arbitrary protein-small molecule interactions.

ALDELE: All-Purpose Deep Learning Toolkits for Predicting the Biocatalytic Activities of Enzymes

Rapidly predicting enzyme properties for catalyzing specific substrates is essential for identifying potential enzymes for industrial transformations. The demand for sustainable production of valuable industry chemicals utilizing biological resources raised a pressing need to speed up biocatalyst screening using machine learning techniques. In this research, we developed an all-purpose deep-learning-based multiple-toolkit (ALDELE) workflow for screening enzyme catalysts. ALDELE incorporates both structural and sequence representations of proteins, alongside representations of ligands by subgraphs and overall physicochemical properties. Comprehensive evaluation demonstrated that ALDELE can predict the catalytic activities of enzymes, and particularly, it identifies residue-based hotspots to guide enzyme engineering and generates substrate heat maps to explore the substrate scope for a given biocatalyst. Moreover, our models notably match empirical data, reinforcing the practicality and reliability of our approach through the alignment with confirmed mutation sites. ALDELE offers a facile and comprehensive solution by integrating different toolkits tailored for different purposes at affordable computational cost and therefore would be valuable to speed up the discovery of new functional enzymes for their exploitation by the industry.

"Mix and match" auto-assembly of glycosyltransferase domains delivers biocatalysts with improved substrate promiscuity

Glycosyltransferases (GT) catalyze the glycosylation of bioactive natural products, including peptides and proteins, flavonoids, and sterols, and have been extensively used as biocatalysts to generate glycosides. However, the often narrow substrate specificity of wild-type GTs requires engineering strategies to expand it. The GT-B structural family is constituted by GTs that share a highly conserved tertiary structure in which the sugar donor and acceptor substrates bind in dedicated domains. Here, we have used this selective binding feature to design an engineering process to generate chimeric glycosyltransferases that combine auto-assembled domains from two different GT-B enzymes. Our approach enabled the generation of a stable dimer with broader substrate promiscuity than the parent enzymes that were related to relaxed interactions between domains in the dimeric GT-B. Our findings provide a basis for the development of a novel class of heterodimeric GTs with improved substrate promiscuity for applications in biotechnology and natural product synthesis.

Late-stage modification of bioactive compounds: Improving druggability through efficient molecular editing

Synthetic chemistry plays an indispensable role in drug discovery, contributing to hit compounds identification, lead compounds optimization, candidate drugs preparation, and so on. As Nobel Prize laureate James Black emphasized, "the most fruitful basis for the discovery of a new drug is to start with an old drug"1. Late-stage modification or functionalization of drugs, natural products and bioactive compounds have garnered significant interest due to its ability to introduce diverse elements into bioactive compounds promptly. Such modifications alter the chemical space and physiochemical properties of these compounds, ultimately influencing their potency and druggability. To enrich a toolbox of chemical modification methods for drug discovery, this review focuses on the incorporation of halogen, oxygen, and nitrogen-the ubiquitous elements in pharmacophore components of the marketed drugs-through late-stage modification in recent two decades, and discusses the state and challenges faced in these fields. We also emphasize that increasing cooperation between chemists and pharmacists may be conducive to the rapid discovery of new activities of the functionalized molecules. Ultimately, we hope this review would serve as a valuable resource, facilitating the application of late-stage modification in the construction of novel molecules and inspiring innovative concepts for designing and building new drugs.

Data-Driven Protein Engineering for Improving Catalytic Activity and Selectivity

Protein engineering is essential for altering the substrate scope, catalytic activity and selectivity of enzymes for applications in biocatalysis. However, traditional approaches, such as directed evolution and rational design, encounter the challenge in dealing with the experimental screening process of a large protein mutation space. Machine learning methods allow the approximation of protein fitness landscapes and the identification of catalytic patterns using limited experimental data, thus providing a new avenue to guide protein engineering campaigns. In this concept article, we review machine learning models that have been developed to assess enzyme-substrate-catalysis performance relationships aiming to improve enzymes through data-driven protein engineering. Furthermore, we prospect the future development of this field to provide additional strategies and tools for achieving desired activities and selectivities.

Exploring the Evolvability of Plant Specialized Metabolism: Uniqueness Out Of Uniformity and Uniqueness Behind Uniformity

The huge structural diversity exhibited by plant specialized metabolites has primarily been considered to result from the catalytic specificity of their biosynthetic enzymes. Accordingly, enzyme gene multiplication and functional differentiation through spontaneous mutations have been established as the molecular mechanisms that drive metabolic evolution. Nevertheless, how plants have assembled and maintained such metabolic enzyme genes and the typical clusters that are observed in plant genomes, as well as why identical specialized metabolites often exist in phylogenetically remote lineages, is currently only poorly explained by a concept known as convergent evolution. Here, we compile recent knowledge on the co-presence of metabolic modules that are common in the plant kingdom but have evolved under specific historical and contextual constraints defined by the physicochemical properties of each plant specialized metabolite and the genetic presets of the biosynthetic genes. Furthermore, we discuss a common manner to generate uncommon metabolites (uniqueness out of uniformity) and an uncommon manner to generate common metabolites (uniqueness behind uniformity). This review describes the emerging aspects of the evolvability of plant specialized metabolism that underlie the vast structural diversity of plant specialized metabolites in nature.

Harnessing Plant Sugar Metabolism for Glycoengineering

Plants possess an innate ability to generate vast amounts of sugar and produce a range of sugar-derived compounds that can be utilized for applications in industry, health, and agriculture. Nucleotide sugars lie at the unique intersection of primary and specialized metabolism, enabling the biosynthesis of numerous molecules ranging from small glycosides to complex polysaccharides. Plants are tolerant to perturbations to their balance of nucleotide sugars, allowing for the overproduction of endogenous nucleotide sugars to push flux towards a particular product without necessitating the re-engineering of upstream pathways. Pathways to produce even non-native nucleotide sugars may be introduced to synthesize entirely novel products. Heterologously expressed glycosyltransferases capable of unique sugar chemistries can further widen the synthetic repertoire of a plant, and transporters can increase the amount of nucleotide sugars available to glycosyltransferases. In this opinion piece, we examine recent successes and potential future uses of engineered nucleotide sugar biosynthetic, transport, and utilization pathways to improve the production of target compounds. Additionally, we highlight current efforts to engineer glycosyltransferases. Ultimately, the robust nature of plant sugar biochemistry renders plants a powerful chassis for the production of target glycoconjugates and glycans.

Pour some sugar on me: The diverse functions of phenylpropanoid glycosylation

The phenylpropanoid metabolism is the source of a vast array of specialized metabolites that play diverse functions in plant growth and development and contribute to all aspects of plant interactions with their surrounding environment. These compounds protect plants from damaging ultraviolet radiation and reactive oxygen species, provide mechanical support for the plants to stand upright, and mediate plant-plant and plant-microorganism communications. The enormous metabolic diversity of phenylpropanoids is further expanded by chemical modifications known as "decorative reactions", including hydroxylation, methylation, glycosylation, and acylation. Among these modifications, glycosylation is the major driving force of phenylpropanoid structural diversification, also contributing to the expansion of their properties. Phenylpropanoid glycosylation is catalyzed by regioselective uridine diphosphate (UDP)-dependent glycosyltransferases (UGTs), whereas glycosyl hydrolases known as β-glucosidases are the major players in deglycosylation. In this article, we review how the glycosylation process affects key physicochemical properties of phenylpropanoids, such as molecular stability and solubility, as well as metabolite compartmentalization/storage and biological activity/toxicity. We also summarize the recent knowledge on the functional implications of glycosylation of different classes of phenylpropanoid compounds. A balance of glycosylation/deglycosylation might represent an essential molecular mechanism to regulate phenylpropanoid homeostasis, allowing plants to dynamically respond to diverse environmental signals.

Machine Learning Enables Prediction of Pyrrolysyl-tRNA Synthetase Substrate Specificity

Knowledge about the substrate scope for a given enzyme is informative for elucidating biochemical pathways and also for expanding applications of the enzyme. However, no general methods are available to accurately predict the substrate specificity of an enzyme. Pyrrolysyl-tRNA synthetase (PylRS) is a powerful tool for incorporating various noncanonical amino acids (NCAAs) into proteins, which enabled us to probe, image, rationally engineer, and evolve protein structure and function. However, the incorporation of a new NCAA typically requires the selection of large libraries of PylRS with randomized mutations at active sites, and this process requires multiple rounds of selection for each new substrate. Therefore, a single aminoacyl-tRNA synthetase with broad substrate promiscuity is ideal to facilitate widespread applications of the genetic NCAA incorporation technique. Herein, machine learning models were developed to predict the substrate specificity of PylRS to accept novel NCAAs that could be incorporated into proteins by three PylRS mutants. The models were built from a training set of 285 unique enzyme-substrate pairs of three PylRS mutants including IFRS, BtaRS, and MFRS against 95 NCAAs. The best BaggingTree (BT) model was then used for virtually screening a NCAAs library containing 1474 phenylalanine, tyrosine, tryptophan, and alanine analogues, and 156 NCAAs were predicted to be accepted by at least one of the three PylRS mutants. Then, 27 NCAAs including 24 positive and 3 negative substrates were experimentally tested for their activities, and 20 of the 24 positive substrates showed weak or strong activity and were accepted by at least one PylRS mutant, among which 11 NCAAs were never reported to be incorporated into proteins before. Three negative substrates did not show any activity. Experimental results suggested that the BT model provides a three-class classification accuracy of 0.69 and a binary classification accuracy of 0.86. This study expanded the substrate scope of three PylRS variants and provided a framework for developing machine learning models to predict substrate specificity of other PylRS variants.

Multi-omics reveal key enzymes involved in the formation of phenylpropanoid glucosides in Artemisia annua

Although mainly known for producing artemisinin, Artemisia annua is enriched in phenylpropanoid glucosides (PGs) with significant bioactivities. However, the biosynthesis of A. annua PGs is insufficiently investigated. Different A. annua ecotypes from distinct growing environments accumulate varying amounts of metabolites, including artemisinin and PGs such as scopolin. UDP-glucose:phenylpropanoid glucosyltransferases (UGTs) transfers glucose from UDP-glucose in PG biosynthesis. Here, we found that the low-artemisinin ecotype GS produces a higher amount of scopolin, compared to the high-artemisinin ecotype HN. By combining transcriptome and proteome analyses, we selected 28 candidate AaUGTs from 177 annotated AaUGTs. Using AlphaFold structural prediction and molecular docking, we determined the binding affinities of 16 AaUGTs. Seven of the AaUGTs enzymatically glycosylated phenylpropanoids. AaUGT25 converted scopoletin to scopolin and esculetin to esculin. The lack of accumulation of esculin in the leaf and the high catalytic efficiency of AaUGT25 on esculetin suggest that esculetin is methylated to scopoletin, the precursor of scopolin. We also discovered that AaOMT1, a previously uncharacterized O-methyltransferase, converts esculetin to scopoletin, suggesting an alternative route for producing scopoletin, which contributes to the high-level accumulation of scopolin in A. annua leaves. AaUGT1 and AaUGT25 responded to induction of stress-related phytohormones, implying the involvement of PGs in stress responses.

Contrastive learning in protein language space predicts interactions between drugs and protein targets

Sequence-based prediction of drug-target interactions has the potential to accelerate drug discovery by complementing experimental screens. Such computational prediction needs to be generalizable and scalable while remaining sensitive to subtle variations in the inputs. However, current computational techniques fail to simultaneously meet these goals, often sacrificing performance of one to achieve the others. We develop a deep learning model, ConPLex, successfully leveraging the advances in pretrained protein language models ("PLex") and employing a protein-anchored contrastive coembedding ("Con") to outperform state-of-the-art approaches. ConPLex achieves high accuracy, broad adaptivity to unseen data, and specificity against decoy compounds. It makes predictions of binding based on the distance between learned representations, enabling predictions at the scale of massive compound libraries and the human proteome. Experimental testing of 19 kinase-drug interaction predictions validated 12 interactions, including four with subnanomolar affinity, plus a strongly binding EPHB1 inhibitor (KD = 1.3 nM). Furthermore, ConPLex embeddings are interpretable, which enables us to visualize the drug-target embedding space and use embeddings to characterize the function of human cell-surface proteins. We anticipate that ConPLex will facilitate efficient drug discovery by making highly sensitive in silico drug screening feasible at the genome scale. ConPLex is available open source at https://ConPLex.csail.mit.edu.

Second-Shell Residues Contribute to Catalysis by Predominately Preorganizing the Apo State in PafA

Residues beyond the first coordination shell are often observed to make considerable cumulative contributions in enzymes. Due to typically indirect perturbations of multiple physicochemical properties of the active site, however, their individual and specific roles in enzyme catalysis and disease-causing mutations remain difficult to predict and understand at the molecular level. Here we analyze the contributions of several second-shell residues in phosphate-irrepressible alkaline phosphatase of flavobacterium (PafA), a representative system as one of the most efficient enzymes. By adopting a multifaceted approach that integrates quantum-mechanical/molecular-mechanical free energy computations, molecular-mechanical molecular dynamics simulations, and density functional theory cluster model calculations, we probe the rate-limiting phosphoryl transfer step and structural properties of all relevant enzyme states. In combination with available experimental data, our computational results show that mutations of the studied second-shell residues impact catalytic efficiency mainly by perturbation of the apo state and therefore substrate binding, while they do not affect the ground state or alter the nature of phosphoryl transfer transition state significantly. Several second-shell mutations also modulate the active site hydration level, which in turn influences the energetics of phosphoryl transfer. These mechanistic insights also help inform strategies that may improve the efficiency of enzyme design and engineering by going beyond the current focus on the first coordination shell.

A general model to predict small molecule substrates of enzymes based on machine and deep learning

For most proteins annotated as enzymes, it is unknown which primary and/or secondary reactions they catalyze. Experimental characterizations of potential substrates are time-consuming and costly. Machine learning predictions could provide an efficient alternative, but are hampered by a lack of information regarding enzyme non-substrates, as available training data comprises mainly positive examples. Here, we present ESP, a general machine-learning model for the prediction of enzyme-substrate pairs with an accuracy of over 91% on independent and diverse test data. ESP can be applied successfully across widely different enzymes and a broad range of metabolites included in the training data, outperforming models designed for individual, well-studied enzyme families. ESP represents enzymes through a modified transformer model, and is trained on data augmented with randomly sampled small molecules assigned as non-substrates. By facilitating easy in silico testing of potential substrates, the ESP web server may support both basic and applied science.

In-depth analysis of biocatalysts by microfluidics: An emerging source of data for machine learning

Nowadays, the vastly increasing demand for novel biotechnological products is supported by the continuous development of biocatalytic applications which provide sustainable green alternatives to chemical processes. The success of a biocatalytic application is critically dependent on how quickly we can identify and characterize enzyme variants fitting the conditions of industrial processes. While miniaturization and parallelization have dramatically increased the throughput of next-generation sequencing systems, the subsequent characterization of the obtained candidates is still a limiting process in identifying the desired biocatalysts. Only a few commercial microfluidic systems for enzyme analysis are currently available, and the transformation of numerous published prototypes into commercial platforms is still to be streamlined. This review presents the state-of-the-art, recent trends, and perspectives in applying microfluidic tools in the functional and structural analysis of biocatalysts. We discuss the advantages and disadvantages of available technologies, their reproducibility and robustness, and readiness for routine laboratory use. We also highlight the unexplored potential of microfluidics to leverage the power of machine learning for biocatalyst development.

Plant glycosyltransferases for expanding bioactive glycoside diversity

Glycosylation is a successful strategy to alter the pharmacological properties of small molecules, and it has emerged as a unique approach to expand the chemical space of natural products that can be explored in drug discovery. Traditionally, most glycosylation events have been carried out chemically, often requiring many protection and deprotection steps to achieve a target molecule. Enzymatic glycosylation by glycosyltransferases could provide an alternative strategy for producing new glycosides. In particular, the glycosyltransferase family has greatly expanded in plants, representing a rich enzymatic resource to mine and expand the diversity of glycosides with novel bioactive properties. This article highlights previous and prospective uses for plant glycosyltransferases in generating bioactive glycosides and altering their pharmacological properties.

Rational design of enzyme activity and enantioselectivity

The strategy of rational design to engineer enzymes is to predict the potential mutants based on the understanding of the relationships between protein structure and function, and subsequently introduce the mutations using the site-directed mutagenesis. Rational design methods are universal, relatively fast and have the potential to be developed into algorithms that can quantitatively predict the performance of the designed sequences. Compared to the protein stability, it was more challenging to design an enzyme with improved activity or selectivity, due to the complexity of enzyme molecular structure and inadequate understanding of the relationships between enzyme structures and functions. However, with the development of computational force, advanced algorithm and a deeper understanding of enzyme catalytic mechanisms, rational design could significantly simplify the process of engineering enzyme functions and the number of studies applying rational design strategy has been increasing. Here, we reviewed the recent advances of applying the rational design strategy to engineer enzyme functions including activity and enantioselectivity. Five strategies including multiple sequence alignment, strategy based on steric hindrance, strategy based on remodeling interaction network, strategy based on dynamics modification and computational protein design are discussed and the successful cases using these strategies are introduced.

Data-driven enzyme engineering to identify function-enhancing enzymes

Identifying function-enhancing enzyme variants is a 'holy grail' challenge in protein science because it will allow researchers to expand the biocatalytic toolbox for late-stage functionalization of drug-like molecules, environmental degradation of plastics and other pollutants, and medical treatment of food allergies. Data-driven strategies, including statistical modeling, machine learning, and deep learning, have largely advanced the understanding of the sequence-structure-function relationships for enzymes. They have also enhanced the capability of predicting and designing new enzymes and enzyme variants for catalyzing the transformation of new-to-nature reactions. Here, we reviewed the recent progresses of data-driven models that were applied in identifying efficiency-enhancing mutants for catalytic reactions. We also discussed existing challenges and obstacles faced by the community. Although the review is by no means comprehensive, we hope that the discussion can inform the readers about the state-of-the-art in data-driven enzyme engineering, inspiring more joint experimental-computational efforts to develop and apply data-driven modeling to innovate biocatalysts for synthetic and pharmaceutical applications.

Recent advances in machine learning applications in metabolic engineering

Metabolic engineering encompasses several widely-used strategies, which currently hold a high seat in the field of biotechnology when its potential is manifesting through a plethora of research and commercial products with a strong societal impact. The genomic revolution that occurred almost three decades ago has initiated the generation of large omics-datasets which has helped in gaining a better understanding of cellular behavior. The itinerary of metabolic engineering that has occurred based on these large datasets has allowed researchers to gain detailed insights and a reasonable understanding of the intricacies of biosystems. However, the existing trail-and-error approaches for metabolic engineering are laborious and time-intensive when it comes to the production of target compounds with high yields through genetic manipulations in host organisms. Machine learning (ML) coupled with the available metabolic engineering test instances and omics data brings a comprehensive and multidisciplinary approach that enables scientists to evaluate various parameters for effective strain design. This vast amount of biological data should be standardized through knowledge engineering to train different ML models for providing accurate predictions in gene circuits designing, modification of proteins, optimization of bioprocess parameters for scaling up, and screening of hyper-producing robust cell factories. This review briefs on the premise of ML, followed by mentioning various ML methods and algorithms alongside the numerous omics datasets available to train ML models for predicting metabolic outcomes with high-accuracy. The combinative interplay between the ML algorithms and biological datasets through knowledge engineering have guided the recent advancements in applications such as CRISPR/Cas systems, gene circuits, protein engineering, metabolic pathway reconstruction, and bioprocess engineering. Finally, this review addresses the probable challenges of applying ML in metabolic engineering which will guide the researchers toward novel techniques to overcome the limitations.

The Sabatier principle as a tool for discovery and engineering of industrial enzymes

The recent breakthrough in all-atom, protein structure prediction opens new avenues for a range of computational approaches in enzyme design. These new approaches could become instrumental for the development of technical biocatalysts, and hence our transition toward more sustainable industries. Here, we discuss one approach, which is well-known within inorganic catalysis, but essentially unexploited in biotechnology. Specifically, we review examples of linear free-energy relationships (LFERs) for enzyme reactions and discuss how LFERs and the associated Sabatier Principle may be implemented in algorithms that estimate kinetic parameters and enzyme performance based on model structures.

Glycoside-specific metabolomics combined with precursor isotopic labeling for characterizing plant glycosyltransferases

Glycosylation by uridine diphosphate-dependent glycosyltransferases (UGTs) in plants contributes to the complexity and diversity of secondary metabolites. UGTs are generally promiscuous in their use of acceptors, making it challenging to reveal the function of UGTs in vivo. Here, we described an approach that combined glycoside-specific metabolomics and precursor isotopic labeling analysis to characterize UGTs in Arabidopsis. We revisited the UGT72E cluster, which has been reported to catalyze the glycosylation of monolignols. Glycoside-specific metabolomics analysis reduced the number of differentially accumulated metabolites in the ugt72e1e2e3 mutant by at least 90% compared with that from traditional untargeted metabolomics analysis. In addition to the two previously reported monolignol glycosides, a total of 62 glycosides showed reduced accumulation in the ugt72e1e2e3 mutant, 22 of which were phenylalanine-derived glycosides, including 5-OH coniferyl alcohol-derived and lignan-derived glycosides, as confirmed by isotopic tracing of [¹³C₆]-phenylalanine precursor. Our method revealed that UGT72Es could use coumarins as substrates, and genetic evidence showed that UGT72Es endowed plants with enhanced tolerance to low iron availability under alkaline conditions. Using the newly developed method, the function of UGT78D2 was also evaluated. These case studies suggest that this method can substantially contribute to the characterization of UGTs and efficiently investigate glycosylation processes, the complexity of which have been highly underestimated.

Reconstitution of monoterpene indole alkaloid biosynthesis in genome engineered Nicotiana benthamiana

Monoterpene indole alkaloids (MIAs) are a diverse class of plant natural products that include a number of medicinally important compounds. We set out to reconstitute the pathway for strictosidine, a key intermediate of all MIAs, from central metabolism in Nicotiana benthamiana. A disadvantage of this host is that its rich background metabolism results in the derivatization of some heterologously produced molecules. Here we use transcriptomic analysis to identify glycosyltransferases that are upregulated in response to biosynthetic intermediates and produce plant lines with targeted mutations in the genes encoding them. Expression of the early MIA pathway in these lines produces a more favorable product profile. Strictosidine biosynthesis was successfully reconstituted, with the best yields obtained by the co-expression of 14 enzymes, of which a major latex protein-like enzyme (MLPL) from Nepeta (catmint) is critical for improving flux through the iridoid pathway. The removal of endogenous glycosyltransferases does not impact the yields of strictosidine, highlighting that the metabolic flux of the pathway enzymes to a stable biosynthetic intermediate minimizes the need to engineer the endogenous metabolism of the host. The production of strictosidine in planta expands the range of MIA products amenable to biological synthesis.

The biosynthesis of strictosidine, a key intermediate of monoterpene indole alkaloids, was successfully reconstructed in Nicotiana benthamiana, demonstrating the potential of Nicotiana benthamiana as a bioproduction chassis for small molecules.

Microbial Lipases and Their Potential in the Production of Pharmaceutical Building Blocks

Processes involving lipases in obtaining active pharmaceutical ingredients (APIs) are crucial to increase the sustainability of the industry. Despite their lower production cost, microbial lipases are striking for their versatile catalyzing reactions beyond their physiological role. In the context of taking advantage of microbial lipases in reactions for the synthesis of API building blocks, this review focuses on: (i) the structural origins of the catalytic properties of microbial lipases, including the results of techniques such as single particle monitoring (SPT) and the description of its selectivity beyond the Kazlauskas rule as the “Mirror-Image Packing” or the “Key Region(s) rule influencing enantioselectivity” (KRIE); (ii) immobilization methods given the conferred operative advantages in industrial applications and their modulating capacity of lipase properties; and (iii) a comprehensive description of microbial lipases use as a conventional or promiscuous catalyst in key reactions in the organic synthesis (Knoevenagel condensation, Morita–Baylis–Hillman (MBH) reactions, Markovnikov additions, Baeyer–Villiger oxidation, racemization, among others). Finally, this review will also focus on a research perspective necessary to increase microbial lipases application development towards a greener industry.

Family 1 glycosyltransferases (GT1, UGTs) are subject to dilution-induced inactivation and low chemo stability toward their own acceptor substrates

Glycosylation reactions are essential but challenging from a conventional chemistry standpoint. Conversely, they are biotechnologically feasible as glycosyltransferases can transfer sugar to an acceptor with perfect regio- and stereo-selectivity, quantitative yields, in a single reaction and under mild conditions. Low stability is often alleged to be a limitation to the biotechnological application of glycosyltransferases. Here we show that these enzymes are not necessarily intrinsically unstable, but that they present both dilution-induced inactivation and low chemostability towards their own acceptor substrates, and that these two phenomena are synergistic. We assessed 18 distinct GT1 enzymes against three unrelated acceptors (apigenin, resveratrol, and scopoletin—respectively a flavone, a stilbene, and a coumarin), resulting in a total of 54 enzymes: substrate pairs. For each pair, we varied catalyst and acceptor concentrations to obtain 16 different reaction conditions. Fifteen of the assayed enzymes (83%) displayed both low chemostability against at least one of the assayed acceptors at submillimolar concentrations, and dilution-induced inactivation. Furthermore, sensitivity to reaction conditions seems to be related to the thermal stability of the enzymes, the three unaffected enzymes having melting temperatures above 55°C, whereas the full enzyme panel ranged from 37.4 to 61.7°C. These results are important for GT1 understanding and engineering, as well as for discovery efforts and biotechnological use.

Data-Driven Synthetic Cell Factories Development for Industrial Biomanufacturing

Revolutionary breakthroughs in artificial intelligence (AI) and machine learning (ML) have had a profound impact on a wide range of scientific disciplines, including the development of artificial cell factories for biomanufacturing. In this paper, we review the latest studies on the application of data-driven methods for the design of new proteins, pathways, and strains. We first briefly introduce the various types of data and databases relevant to industrial biomanufacturing, which are the basis for data-driven research. Different types of algorithms, including traditional ML and more recent deep learning methods, are also presented. We then demonstrate how these data-based approaches can be applied to address various issues in cell factory development using examples from recent studies, including the prediction of protein function, improvement of metabolic models, and estimation of missing kinetic parameters, design of non-natural biosynthesis pathways, and pathway optimization. In the last section, we discuss the current limitations of these data-driven approaches and propose that data-driven methods should be integrated with mechanistic models to complement each other and facilitate the development of synthetic strains for industrial biomanufacturing.

Leaf necrosis resulting from downregulation of poplar glycosyltransferase UGT72A2

Reactive species (RS) causing oxidative stress are unavoidable by-products of various plant metabolic processes, such as photosynthesis, respiration or photorespiration. In leaves, flavonoids scavenge RS produced during photosynthesis and protect plant cells against deleterious oxidative damages. Their biosynthesis and accumulation are therefore under tight regulation at the cellular level. Glycosylation has emerged as an essential biochemical reaction in the homeostasis of various specialized metabolites such as flavonoids. This article provides a functional characterization of the Populus tremula x P. alba (poplar) UGT72A2 coding for a UDP-glycosyltransferase that is localized in the chloroplasts. Compared with the wild type, transgenic poplar lines with decreased expression of UGT72A2 are characterized by reduced growth and oxidative damages in leaves, as evidenced by necrosis, higher content of glutathione and lipid peroxidation products as well as diminished soluble peroxidase activity and NADPH to NADP+ ratio under standard growing conditions. They furthermore display lower pools of phenolics, anthocyanins and total flavonoids but higher proanthocyanidins content. Promoter analysis revealed the presence of cis-elements involved in photomorphogenesis, chloroplast biogenesis and flavonoid biosynthesis. The UGT72A2 is regulated by the poplar MYB119, a transcription factor known to regulate the flavonoid biosynthesis pathway. Phylogenetic analysis and molecular docking suggest that UGT72A2 could glycosylate flavonoids; however, the actual substrate(s) was not consistently evidenced with either in vitro assays nor analyses of glycosylated products in leaves of transgenic poplar overexpressing or downregulated for UGT72A2. This article provides elements highlighting the importance of flavonoid glycosylation regarding protection against oxidative stress in poplar leaves and raises new questions about the link between this biochemical reaction and regulation of the redox homeostasis system.

Advances and Challenges in Enzymatic C-glycosylation of Flavonoids in Plants

Flavonoid glycosides play required determinant roles in plants and have considerable potential for applications in medicine and biotechnology. Glycosyltransferases transfer a sugar moiety from uridine diphosphate-activated sugar molecules to an acceptor flavonoid via C-O and C-C linkages. Compared with O-glycosylflavonoids, C-glycosylflavonoids are more stable, are resistant to glycosidase or acid hydrolysis, exhibit better pharmacological properties, and have received more attention. Herein, we discuss the mining of C-glycosylflavones and the corresponding C-glycosyltransferases and evaluate the differences in structure and catalytic mechanisms between C-glycosyltransferase and O-glycosyltransferase. We conclude that promiscuity and specificity are key determinants for general flavonoid C-glycosyltransferase engineering and summarize the C-glycosyltransferase engineering strategy. A thorough understanding of the properties, catalytic mechanisms, and engineering of C-glycosyltransferases will be critical for any future biotechnological applications in areas such as the production of desired C-glycosylflavonoids for nutritional or medicinal use.

UGT72, a Major Glycosyltransferase Family for Flavonoid and Monolignol Homeostasis in Plants

Simple Summary

Phenylpropanoids are specialized metabolites playing crucial roles in plant developmental processes and in plant defense towards pathogens. The attachment of sugar moieties to these small hydrophobic molecules renders them more hydrophilic and increases their solubility. The UDP-glycosyltransferase 72 family (UGT72) of plants has been shown to glycosylate mainly two classes of phenylpropanoids, (i) the monolignols that are the building blocks of lignin, the second most abundant polymer after cellulose, and (ii) the flavonoids, which play determinant roles in plant interactions with other organisms and in response to stress. The purpose of this review is to bring an overview of the current knowledge of the UGT72 family and to highlight its role in the homeostasis of these molecules. Potential applications in pharmacology and in wood, paper pulp, and bioethanol production are given within the perspectives.

Abstract

Plants have developed the capacity to produce a diversified range of specialized metabolites. The glycosylation of those metabolites potentially decreases their toxicity while increasing their stability and their solubility, modifying their transport and their storage. The UGT, forming the largest glycosyltransferase superfamily in plants, combine enzymes that glycosylate mainly hormones and phenylpropanoids by using UDP-sugar as a sugar donor. Particularly, members of the UGT72 family have been shown to glycosylate the monolignols and the flavonoids, thereby being involved in their homeostasis. First, we explore primitive UGTs in algae and liverworts that are related to the angiosperm UGT72 family and their role in flavonoid homeostasis. Second, we describe the role of several UGT72s glycosylating monolignols, some of which have been associated with lignification. In addition, the role of other UGT72 members that glycosylate flavonoids and are involved in the development and/or stress response is depicted. Finally, the importance to explore the subcellular localization of UGTs to study their roles in planta is discussed.

Glycosyltransferases: Mining, engineering and applications in biosynthesis of glycosylated plant natural products

UDP-Glycosyltransferases (UGTs) catalyze the transfer of nucleotide-activated sugars to specific acceptors, among which the GT1 family enzymes are well-known for their function in biosynthesis of natural product glycosides. Elucidating GT function represents necessary step in metabolic engineering of aglycone glycosylation to produce drug leads, cosmetics, nutrients and sweeteners. In this review, we systematically summarize the phylogenetic distribution and catalytic diversity of plant GTs. We also discuss recent progress in the identification of novel GT candidates for synthesis of plant natural products (PNPs) using multi-omics technology and deep learning predicted models. We also highlight recent advances in rational design and directed evolution engineering strategies for new or improved GT functions. Finally, we cover recent breakthroughs in the application of GTs for microbial biosynthesis of some representative glycosylated PNPs, including flavonoid glycosides (fisetin 3-O-glycosides, astragalin, scutellarein 7-O-glucoside), terpenoid glycosides (rebaudioside A, ginsenosides) and polyketide glycosides (salidroside, polydatin).

Characterization of a Group of UDP-Glycosyltransferases Involved in the Biosynthesis of Triterpenoid Saponins of Panax notoginseng

UDP-glycosyltransferase (UGT)-mediated glycosylation is a common modification in triterpene saponins, which exhibit a wide range of bioactivities and important pharmacological effects. However, few UGTs involved in saponin biosynthesis have been identified, limiting the biosynthesis of saponins. In this study, an efficient heterologous expression system was established for evaluating the UGT-mediated glycosylation process of triterpene saponins. Six UGTs (UGTPn17, UGTPn42, UGTPn35, UGTPn87, UGTPn19, and UGTPn12) from Panax notoginseng were predicted and found to be responsible for efficient and direct enzymatic biotransformation of 21 triterpenoid saponins via 26 various glycosylation reactions. Among them, UGTPn87 exhibited promiscuous sugar-donor specificity of UDP-glucose (UDP-Glc) and UDP-xylose (UDP-Xyl) by catalyzing the elongation of the second sugar chain at the C3 or/and C20 sites of protopanaxadiol-type saponins with a UDP-Glc or UDP-Xyl donor, as well as at the C20 site of protopanaxadiol-type saponins with a UDP-Glc donor. Two new saponins, Fd-Xyl and Fe-Xyl, were generated by catalyzing the C3-O-Glc xylosylations of notoginsenoside Fd and notoginsenoside Fe when incubated with UGTPn87. Moreover, the complete biosynthetic pathways of 17 saponins were elucidated, among which notoginsenoside L, vinaginsenoside R16, gypenoside LXXV, and gypenoside XVII were revealed in Panax for the first time. A yeast cell factory was constructed with a yield of Rh2 at 354.69 mg/L and a glycosylation ratio of 60.40% in flasks. Our results reveal the biosynthetic pathway of a group of saponins in P. notoginseng and provide a theoretical basis for producing rare and valuable saponins, promoting their industrial application in medicine and functional foods.

Glycan processing in the Golgi: optimal information coding and constraints on cisternal number and enzyme specificity

Many proteins that undergo sequential enzymatic modification in the Golgi cisternae are displayed at the plasma membrane as cell identity markers. The modified proteins, called glycans, represent a molecular code. The fidelity of this glycan code is measured by how accurately the glycan synthesis machinery realises the desired target glycan distribution for a particular cell type and niche. In this paper, we construct a simplified chemical synthesis model to quantitatively analyse the tradeoffs between the number of cisternae, and the number and specificity of enzymes, required to synthesize a prescribed target glycan distribution of a certain complexity to within a given fidelity. We find that to synthesize complex distributions, such as those observed in real cells, one needs to have multiple cisternae and precise enzyme partitioning in the Golgi. Additionally, for fixed number of enzymes and cisternae, there is an optimal level of specificity (promiscuity) of enzymes that achieves the target distribution with high fidelity. The geometry of the fidelity landscape in the multidimensional space of the number and specificity of enzymes, inter-cisternal transfer rates, and number of cisternae, provides a measure for robustness and identifies stiff and sloppy directions. Our results show how the complexity of the target glycan distribution and number of glycosylation enzymes places functional constraints on the Golgi cisternal number and enzyme specificity.

Machine learning modeling of family wide enzyme-substrate specificity screens

Biocatalysis is a promising approach to sustainably synthesize pharmaceuticals, complex natural products, and commodity chemicals at scale. However, the adoption of biocatalysis is limited by our ability to select enzymes that will catalyze their natural chemical transformation on non-natural substrates. While machine learning and in silico directed evolution are well-posed for this predictive modeling challenge, efforts to date have primarily aimed to increase activity against a single known substrate, rather than to identify enzymes capable of acting on new substrates of interest. To address this need, we curate 6 different high-quality enzyme family screens from the literature that each measure multiple enzymes against multiple substrates. We compare machine learning-based compound-protein interaction (CPI) modeling approaches from the literature used for predicting drug-target interactions. Surprisingly, comparing these interaction-based models against collections of independent (single task) enzyme-only or substrate-only models reveals that current CPI approaches are incapable of learning interactions between compounds and proteins in the current family level data regime. We further validate this observation by demonstrating that our no-interaction baseline can outperform CPI-based models from the literature used to guide the discovery of kinase inhibitors. Given the high performance of non-interaction based models, we introduce a new structure-based strategy for pooling residue representations across a protein sequence. Altogether, this work motivates a principled path forward in order to build and evaluate meaningful predictive models for biocatalysis and other drug discovery applications.

Predicting interactions between compounds and proteins represents a long-standing dream of drug discovery and protein engineering. Robust models of enzyme-substrate scope would dramatically advance our ability to design synthetic routes involving enzymatic catalysis. However, the lack of standardization between compound-protein interaction studies makes it difficult to evaluate the generalizability of such models. In this work we take a critical step forward by standardizing high-quality datasets measuring enzyme-substrate interactions, outlining rigorous evaluations, and proposing a new way to integrate structural information into protein representations. In testing previous modeling approaches, we highlight a surprising inability of existing models to effectively leverage compound-protein interactions to improve generalization, which challenges a perception in the literature. This establishes future opportunities for model development and integration of enzyme-substrate scope models into computer-aided synthesis planning software.

Unraveling the Glucosylation of Astringency Compounds of Horse Chestnut via Integrative Sensory Evaluation, Flavonoid Metabolism, Differential Transcriptome, and Phylogenetic Analysis

The seeds of Chinese horse chestnut are used as a source of starch and escin, whereas the potential use of whole plant has been ignored. The astringency and bitterness of tea produced from the leaves and flowers were found to be significantly better than those of green tea, suggesting that the enriched flavonoids maybe sensory determinates. During 47 flavonoids identified in leaves and flowers, seven flavonol glycosides in the top 10 including astragalin and isoquercitrin were significantly higher content in flowers than in leaves. The crude proteins of flowers could catalyze flavonol glucosides' formation, in which three glycosyltransferases contributed to the flavonol glucosylation were screened out by multi-dimensional integration of transcriptome, evolutionary analyses, recombinant enzymatic analysis and molecular docking. The deep exploration for flavonol profile and glycosylation provides theoretical and experimental basis for utilization of flowers and leaves of Aesculus chinensis as additives and dietary supplements.

Computer Simulation to Rationalize “Rational” Engineering of Glycoside Hydrolases and Glycosyltransferases

Glycoside hydrolases and glycosyltransferases are the main classes of enzymes that synthesize and degrade carbohydrates, molecules essential to life that are a challenge for classical chemistry. As such, considerable efforts have been made to engineer these enzymes and make them pliable to human needs, ranging from directed evolution to rational design, including mechanism engineering. Such endeavors fall short and are unreported in numerous cases, while even success is a necessary but not sufficient proof that the chemical rationale behind the design is correct. Here we review some of the recent work in CAZyme mechanism engineering, showing that computational simulations are instrumental to rationalize experimental data, providing mechanistic insight into how native and engineered CAZymes catalyze chemical reactions. We illustrate this with two recent studies in which (i) a glycoside hydrolase is converted into a glycoside phosphorylase and (ii) substrate specificity of a glycosyltransferase is engineered toward forming O-, N-, or S-glycosidic bonds.

Catalytic flexibility of rice glycosyltransferase OsUGT91C1 for the production of palatable steviol glycosides

Steviol glycosides are the intensely sweet components of extracts from Stevia rebaudiana. These molecules comprise an invariant steviol aglycone decorated with variable glycans and could widely serve as a low-calorie sweetener. However, the most desirable steviol glycosides Reb D and Reb M, devoid of unpleasant aftertaste, are naturally produced only in trace amounts due to low levels of specific β (1-2) glucosylation in Stevia. Here, we report the biochemical and structural characterization of OsUGT91C1, a glycosyltransferase from Oryza sativa, which is efficient at catalyzing β (1-2) glucosylation. The enzyme's ability to bind steviol glycoside substrate in three modes underlies its flexibility to catalyze β (1-2) glucosylation in two distinct orientations as well as β (1-6) glucosylation. Guided by the structural insights, we engineer this enzyme to enhance the desirable β (1-2) glucosylation, eliminate β (1-6) glucosylation, and obtain a promising catalyst for the industrial production of naturally rare but palatable steviol glycosides.

Structure-function relationship of terpenoid glycosyltransferases from plants

Covering: up to 2021Terpenoids are physiologically active substances that are of great importance to humans. Their physicochemical properties are modified by glycosylation, in terms of polarity, volatility, solubility and reactivity, and their bioactivities are altered accordingly. Significant scientific progress has been made in the functional study of glycosylated terpenes and numerous plant enzymes involved in regio- and enantioselective glycosylation have been characterized, a reaction that remains chemically challenging. Crucial clues to the mechanism of terpenoid glycosylation were recently provided by the first crystal structures of a diterpene glycosyltransferase UGT76G1. Here, we review biochemically characterized terpenoid glycosyltransferases, compare their functions and primary structures, discuss their acceptor and donor substrate tolerance and product specificity, and elaborate features of the 3D structures of the first terpenoid glycosyltransferases from plants.

Structural modeling of two plant UDP-dependent sugar-sugar glycosyltransferases reveals a conserved glutamic acid residue that is a hallmark for sugar acceptor recognition

Glycosylation is one of the common modifications of plant metabolites, playing a major role in the chemical/biological diversity of a wide range of compounds. Plant metabolite glycosylation is catalyzed almost exclusively by glycosyltransferases, mainly by Uridine-diphosphate dependent Glycosyltransferases (UGTs). Several X-ray structures have been determined for primary glycosyltransferases, however, little is known regarding structure-function aspects of sugar-sugar/branch-forming O-linked UGTs (SBGTs) that catalyze the transfer of a sugar from the UDP-sugar donor to an acceptor sugar moiety of a previously glycosylated metabolite substrate. In this study we developed novel insights into the structural basis for SBGT catalytic activity by modelling the 3d-structures of two enzymes; a rhamnosyl-transferase Cs1,6RhaT - that catalyzes rhamnosylation of flavonoid-3-glucosides and flavonoid-7-glucosides and a UGT94D1 - that catalyzes glucosylation of (+)-Sesaminol 2-O-β-d-glucoside at the C6 of the primary sugar moiety. Based on these structural models and docking studies a glutamate (E290 or E268 in Cs1,6RhaT or UGT94D1, respectively) and a tryptophan (W28 or W15 in Cs1,6RhaT or UGT94D1, respectively) appear to interact with the sugar acceptor and are suggested to be important for the recognition of the sugar-moiety of the acceptor-substrate. Functional analysis of substitution mutants for the glutamate and tryptophan residues in Cs1,6RhaT further support their role in determining sugar-sugar/branch-forming GT specificity. Phylogenetic analysis of the UGT family in plants demonstrates that the glutamic-acid residue is a hallmark of SBGTs that is entirely absent from the corresponding position in primary UGTs.

Enzymkatalysierte späte Modifizierungen: Besser spät als nie

Die Enzymkatalyse gewinnt zunehmend an Bedeutung in der Synthesechemie. Die durch Bioinformatik und Enzym‐Engineering stetig wachsende Zahl von Biokatalysatoren eröffnet eine große Vielfalt selektiver Reaktionen. Insbesondere für späte Funktionalisierungsreaktionen ist die Biokatalyse ein geeignetes Werkzeug, das oftmals der konventionellen De‐novo‐Synthese überlegen ist. Enzyme haben sich als nützlich erwiesen, um funktionelle Gruppen direkt in komplexe Molekülgerüste einzuführen sowie für die rasche Diversifizierung von Substanzbibliotheken. Biokatalytische Oxyfunktionalisierungen, Halogenierungen, Methylierungen, Reduktionen und Amidierungen sind von besonderem Interesse, da diese Strukturmotive häufig in Pharmazeutika vertreten sind. Dieser Aufsatz gibt einen Überblick über die Stärken und Schwächen der enzymkatalysierten späten Modifizierungen durch native und optimierte Enzyme in der Synthesechemie. Ebenso werden wichtige Beispiele in der Wirkstoffentwicklung hervorgehoben.

Enzymatic Late-Stage Modifications: Better Late Than Never

Enzyme catalysis is gaining increasing importance in synthetic chemistry. Nowadays, the growing number of biocatalysts accessible by means of bioinformatics and enzyme engineering opens up an immense variety of selective reactions. Biocatalysis especially provides excellent opportunities for late-stage modification often superior to conventional de novo synthesis. Enzymes have proven to be useful for direct introduction of functional groups into complex scaffolds, as well as for rapid diversification of compound libraries. Particularly important and highly topical are enzyme-catalysed oxyfunctionalisations, halogenations, methylations, reductions, and amide bond formations due to the high prevalence of these motifs in pharmaceuticals. This Review gives an overview of the strengths and limitations of enzymatic late-stage modifications using native and engineered enzymes in synthesis while focusing on important examples in drug development.

Highly Efficient Biosynthesis of Glycyrrhetinic Acid Glucosides by Coupling of Microbial Glycosyltransferase to Plant Sucrose Synthase

Glycyrrhetinic acid (GA) is a principal bioactive pentacyclic triterpenoid from Glycyrrhiza uralensis. Uridine diphosphate-dependent glycosyltransferases (UGTs) have been widely used to catalyze glycosylation of diverse nature products for the development of potential therapeutic compounds. In this study, we have characterized a UGT109A3 from Bacillus subtilis, which can glycosylate both the free C3 hydroxyl and C30 carboxyl groups of GA to yield a unique 3, 30-O-β-D-diglucoside-GA. By coupling the microbial UGT109A3 to plant sucrose synthase (SUS), GA-diglucoside could be biosynthesized in an efficient and economical way. With a fed-batch glycosylation, a large scale of GA-diglucoside (6.26 mM, 4.98 g/L in 8 h) could be enzymatically transformed from GA. The obtained GA-diglucoside showed a significant water solubility improvement of around 3.4 × 103 fold compared with that of the parent GA (29 μM). Moreover, it also exhibited dose-dependent cytotoxicity toward human colon carcinoma Caco-2 cell line according to MTT assay, having an IC50 at 160 μM. This study not only establishes efficient platform for producing GA-glucosides, but is also valuable for developing further the biosynthesis of other complex glycosylated natural products.

Glycosyl Oxocarbenium Ions: Structure, Conformation, Reactivity, and Interactions

Carbohydrates (glycans, saccharides, and sugars) are essential molecules in all domains of life. Research on glycoscience spans from chemistry to biomedicine, including material science and biotechnology. Access to pure and well-defined complex glycans using synthetic methods depends on the success of the employed glycosylation reaction. In most cases, the mechanism of the glycosylation reaction is believed to involve the oxocarbenium ion. Understanding the structure, conformation, reactivity, and interactions of this glycosyl cation is essential to predict the outcome of the reaction. In this Account, building on our contributions on this topic, we discuss the theoretical and experimental approaches that have been employed to decipher the key features of glycosyl cations, from their structures to their interactions and reactivity.We also highlight that, from a chemical perspective, the glycosylation reaction can be described as a continuum, from unimolecular SN1 with naked oxocarbenium cations as intermediates to bimolecular SN2-type mechanisms, which involve the key role of counterions and donors. All these factors should be considered and are discussed herein. The importance of dissociative mechanisms (involving contact ion pairs, solvent-separated ion pairs, solvent-equilibrated ion pairs) with bimolecular features in most reactions is also highlighted.The role of theoretical calculations to predict the conformation, dynamics, and reactivity of the oxocarbenium ion is also discussed, highlighting the advances in this field that now allow access to the conformational preferences of a variety of oxocarbenium ions and their reactivities under SN1-like conditions.Specifically, the ground-breaking use of superacids to generate these cations is emphasized, since it has permitted characterization of the structure and conformation of a variety of glycosyl oxocarbenium ions in superacid solution by NMR spectroscopy.We also pay special attention to the reactivity of these glycosyl ions, which depends on the conditions, including the counterions, the possible intra- or intermolecular participation of functional groups that may stabilize the cation and the chemical nature of the acceptor, either weak or strong nucleophile. We discuss recent investigations from different experimental perspectives, which identified the involved ionic intermediates, estimating their lifetimes and reactivities and studying their interactions with other molecules. In this context, we also emphasize the relationship between the chemical methods that can be employed to modulate the sensitivity of glycosyl cations and the way in which glycosyl modifying enzymes (glycosyl hydrolases and transferases) build and cleave glycosidic linkages in nature. This comparison provides inspiration on the use of molecules that regulate the stability and reactivity of glycosyl cations.

Dissection of the Glycosylation in the Biosynthesis of the Heptadecaglycoside Antibiotic Saccharomicin A

Oligosaccharide natural products have diverse biological activities and represent a potentially important source for drug development. In this study, we focus on the glycosylation pathway in the biosynthesis of saccharomicin A (SA-A), an oligosaccharide antibiotic containing 17 sugar moieties. By extensive gene-knockout studies with comparative metabolic profile analysis, we established a complete pathway in assembling the heptadecasaccharide chain of SA-A, the longest saccharide chain found in natural products.

Promiscuity and specificity of eukaryotic glycosyltransferases

Glycosyltransferases are a large family of enzymes responsible for covalently linking sugar monosaccharides to a variety of organic substrates. These enzymes drive the synthesis of complex oligosaccharides known as glycans, which play key roles in inter-cellular interactions across all the kingdoms of life; they also catalyze sugar attachment during the synthesis of small-molecule metabolites such as plant flavonoids. A given glycosyltransferase enzyme is typically responsible for attaching a specific donor monosaccharide, via a specific glycosidic linkage, to a specific moiety on the acceptor substrate. However these enzymes are often promiscuous, able catalyze linkages between a variety of donors and acceptors. In this review we discuss distinct classes of glycosyltransferase promiscuity, each illustrated by enzymatic examples from small-molecule or glycan synthesis. We highlight the physical causes of promiscuity, and its biochemical consequences. Structural studies of glycosyltransferases involved in glycan synthesis show that they make specific contacts with ‘recognition motifs’ that are much smaller than the full oligosaccharide substrate. There is a wide range in the sizes of glycosyltransferase recognition motifs: highly promiscuous enzymes recognize monosaccharide or disaccharide motifs across multiple oligosaccharides, while highly specific enzymes recognize large, complex motifs found on few oligosaccharides. In eukaryotes, the localization of glycosyltransferases within compartments of the Golgi apparatus may play a role in mitigating the glycan variability caused by enzyme promiscuity.

Machine learning for enzyme engineering, selection and design

Machine learning is a useful computational tool for large and complex tasks such as those in the field of enzyme engineering, selection and design. In this review, we examine enzyme-related applications of machine learning. We start by comparing tools that can identify the function of an enzyme and the site responsible for that function. Then we detail methods for optimizing important experimental properties, such as the enzyme environment and enzyme reactants. We describe recent advances in enzyme systems design and enzyme design itself. Throughout we compare and contrast the data and algorithms used for these tasks to illustrate how the algorithms and data can be best used by future designers.

Enzymes in biotechnology: Critical platform technologies for bioprocess development

Enzymes are core elements of biosynthetic pathways employed in the synthesis of numerous bioproducts. Here, we review enzyme promiscuity, enzyme engineering, enzyme immobilization, and cell-free systems as fundamental strategies of bioprocess development. Initially, promiscuous enzymes are the first candidates in the quest for new activities to power new, artificial, or bypass pathways that expand substrate range and catalyze the production of new products. If the activity or regulation of available enzymes is unsuitable for a process, protein engineering can be applied to improve them to the required level. When cell toxicity and low productivity cannot be engineered away, cell-free systems are an attractive option, especially in combination with enzyme immobilization that allows extended enzyme use. Overall, the above methods support powerful platforms for bioprocess development and optimization.

Natural and engineered transglycosylases: Green tools for the enzyme-based synthesis of glycoproducts

An increasing number of transglycosylase-based processes provide access to oligosaccharides or glycoconjugates, some of them reaching performance levels compatible with industrial developments. Nevertheless, the full potential of transglycosylases has not been explored because of the challenges in transforming a glycoside hydrolase into an efficient transglycosylase. Advances in studying enzyme structure/function relationships, screening enzyme activity, and generating synthetic libraries guided by computational protein design or machine learning methods should considerably accelerate the development of these catalysts. The time has now come for researchers to uncover their possibilities and learn how to design and precisely refine their activity to respond more rapidly to the growing demand for well-defined glycosidic structures.

Recent progress in synthesis of carbohydrates with sugar nucleotide-dependent glycosyltransferases

Sugar nucleotide-dependent glycosyltransferases (GTs) are key enzymes that catalyze the formation of glycosidic bonds in nature. They have been increasingly applied in the synthesis of complex carbohydrates and glycoconjugates with or without in situ generation of sugar nucleotides. Human GTs are becoming more accessible and new bacterial GTs have been identified and characterized. An increasing number of crystal structures elucidated for GTs from mammalian and bacterial sources facilitate structure-based design of mutants as improved catalysts for synthesis. Automated platforms have also been developed for chemoenzymatic synthesis of carbohydrates. Recent progress in applying sugar nucleotide-dependent GTs in enzymatic and chemoenzymatic synthesis of mammalian glycans and glycoconjugates, bacterial surface glycans, and glycosylated natural products from bacteria and plants are reviewed.