Functional annotation of enzyme-encoding genes using deep learning with transformer layers.

Functional annotation of open reading frames in microbial genomes remains substantially incomplete. Enzymes constitute the most prevalent functional gene class in microbial genomes and can be described by their specific catalytic functions using the Enzyme Commission (EC) number. Consequently, the ability to predict EC numbers could substantially reduce the number of un-annotated genes. Here we present a deep learning model, DeepECtransformer, which utilizes transformer layers as a neural network architecture to predict EC numbers. Using the extensively studied Escherichia coli K-12 MG1655 genome, DeepECtransformer predicted EC numbers for 464 un-annotated genes. We experimentally validated the enzymatic activities predicted for three proteins (YgfF, YciO, and YjdM). Further examination of the neural network's reasoning process revealed that the trained neural network relies on functional motifs of enzymes to predict EC numbers. Thus, DeepECtransformer is a method that facilitates the functional annotation of uncharacterized genes.

Deep learning for metabolic pathway design.

The establishment of a bio-based circular economy is imperative in tackling the climate crisis and advancing sustainable development. In this realm, the creation of microbial cell factories is central to generating a variety of chemicals and materials. The design of metabolic pathways is crucial in shaping these microbial cell factories, especially when it comes to producing chemicals with yet-to-be-discovered biosynthetic routes. To aid in navigating the complexities of chemical and metabolic domains, computer-supported tools for metabolic pathway design have emerged. In this paper, we evaluate how digital strategies can be employed for pathway prediction and enzyme discovery. Additionally, we touch upon the recent strides made in using deep learning techniques for metabolic pathway prediction. These computational tools and strategies streamline the design of metabolic pathways, facilitating the development of microbial cell factories. Leveraging the capabilities of deep learning in metabolic pathway design is profoundly promising, potentially hastening the advent of a bio-based circular economy.

Metabolic engineering for sustainability and health.

Bio-based production of chemicals and materials has attracted much attention due to the urgent need to establish sustainability and enhance human health. Metabolic engineering (ME) allows purposeful modification of cellular metabolic, regulatory, and signaling networks to achieve enhanced production of desired chemicals and degradation of environmentally harmful chemicals. ME has significantly progressed over the past 30 years through further integration of the strategies of synthetic biology, systems biology, evolutionary engineering, and data science aided by artificial intelligence. Here we review the field of ME from its emergence to the current state-of-the-art, highlighting its contribution to sustainable production of chemicals, health, and the environment through representative examples. Future challenges of ME and perspectives are also discussed.

Metabolic engineering of Mannheimia succiniciproducens for malic acid production using dimethylsulfoxide as an electron acceptor.

Microbial production of various TCA intermediates and related chemicals through the reductive TCA cycle has been of great interest. However, rumen bacteria that naturally possess strong reductive TCA cycle have been rarely studied to produce these chemicals, except for succinic acid, due to their dependence on fumarate reduction to transport electrons for ATP synthesis. In this study, malic acid (MA), a dicarboxylic acid of industrial importance, was selected as a target chemical for mass production using Mannheimia succiniciproducens, a rumen bacterium possessing a strong reductive branch of the TCA cycle. The metabolic pathway was reconstructed by eliminating fumarase to prevent MA conversion to fumarate. The respiration system of M. succiniciproducens was reconstructed by introducing the Actinobacillus succinogenes dimethylsulfoxide (DMSO) reductase to improve cell growth using DMSO as an electron acceptor. Also, the cell membrane was engineered by employing Pseudomonas aeruginosa cis-trans isomerase to enhance MA tolerance. High inoculum fed-batch fermentation of the final engineered strain produced 61 g/L of MA with an overall productivity of 2.27 g/L/h, which is the highest MA productivity reported to date. The systems metabolic engineering strategies reported in this study will be useful for developing anaerobic bioprocesses for the production of various industrially important chemicals.

An interactive metabolic map of bio-based chemicals.

Metabolic engineering for the bio-based production of chemicals requires thorough understanding of metabolic reactions including enzymes, cofactors, reactants, and products. Here we present an interactive bio-based chemicals map that visualizes compounds, enzymes, and reaction pathways together with strategies for the production of chemicals by biological, chemical, and combined methods.

Designing Microbial Cell Factories for the Production of Chemicals.

The sustainable production of chemicals from renewable, nonedible biomass has emerged as an essential alternative to address pressing environmental issues arising from our heavy dependence on fossil resources. Microbial cell factories are engineered microorganisms harboring biosynthetic pathways streamlined to produce chemicals of interests from renewable carbon sources. The biosynthetic pathways for the production of chemicals can be defined into three categories with reference to the microbial host selected for engineering: native-existing pathways, nonnative-existing pathways, and nonnative-created pathways. Recent trends in leveraging native-existing pathways, discovering nonnative-existing pathways, and designing de novo pathways (as nonnative-created pathways) are discussed in this Perspective. We highlight key approaches and successful case studies that exemplify these concepts. Once these pathways are designed and constructed in the microbial cell factory, systems metabolic engineering strategies can be used to improve the performance of the strain to meet industrial production standards. In the second part of the Perspective, current trends in design tools and strategies for systems metabolic engineering are discussed with an eye toward the future. Finally, we survey current and future challenges that need to be addressed to advance microbial cell factories for the sustainable production of chemicals.

Applications of artificial intelligence to enzyme and pathway design for metabolic engineering.

Metabolic engineering for developing industrial strains capable of overproducing bioproducts requires good understanding of cellular metabolism, including metabolic reactions and enzymes. However, metabolic pathways and enzymes involved are still unknown for many products of interest, which presents a key challenge in their biological production. This challenge can be partly overcome by constructing novel biosynthetic pathways through enzyme and pathway design approaches. With the increase in bio-big data, data-driven approaches using artificial intelligence (AI) techniques are allowing more advanced protein and pathway design. In this paper, we review recent studies on AI-aided protein engineering and design, focusing on directed evolution that uses AI approaches to efficiently construct mutant libraries. Also, recent works of AI-aided pathway design strategies, including template-based and template-free approaches, are discussed.

Synthetic Formatotrophs for One-Carbon Biorefinery.

The use of CO2 as a carbon source in biorefinery is of great interest, but the low solubility of CO2 in water and the lack of efficient CO2 assimilation pathways are challenges to overcome. Formic acid (FA), which can be easily produced from CO2 and more conveniently stored and transported than CO2, is an attractive CO2-equivalent carbon source as it can be assimilated more efficiently than CO2 by microorganisms and also provides reducing power. Although there are native formatotrophs, they grow slowly and are difficult to metabolically engineer due to the lack of genetic manipulation tools. Thus, much effort is exerted to develop efficient FA assimilation pathways and synthetic microorganisms capable of growing solely on FA (and CO2). Several innovative strategies are suggested to develop synthetic formatotrophs through rational metabolic engineering involving new enzymes and reconstructed FA assimilation pathways, and/or adaptive laboratory evolution (ALE). In this paper, recent advances in development of synthetic formatotrophs are reviewed, focusing on biological FA and CO2 utilization pathways, enzymes involved and newly developed, and metabolic engineering and ALE strategies employed. Also, future challenges in cultivating formatotrophs to higher cell densities and producing chemicals from FA and CO2 are discussed.

DeepTFactor: A deep learning-based tool for the prediction of transcription factors.

A transcription factor (TF) is a sequence-specific DNA-binding protein that modulates the transcription of a set of particular genes, and thus regulates gene expression in the cell. TFs have commonly been predicted by analyzing sequence homology with the DNA-binding domains of TFs already characterized. Thus, TFs that do not show homologies with the reported ones are difficult to predict. Here we report the development of a deep learning-based tool, DeepTFactor, that predicts whether a protein in question is a TF. DeepTFactor uses a convolutional neural network to extract features of a protein. It showed high performance in predicting TFs of both eukaryotic and prokaryotic origins, resulting in F1 scores of 0.8154 and 0.8000, respectively. Analysis of the gradients of prediction score with respect to input suggested that DeepTFactor detects DNA-binding domains and other latent features for TF prediction. DeepTFactor predicted 332 candidate TFs in Escherichia coli K-12 MG1655. Among them, 84 candidate TFs belong to the y-ome, which is a collection of genes that lack experimental evidence of function. We experimentally validated the results of DeepTFactor prediction by further characterizing genome-wide binding sites of three predicted TFs, YqhC, YiaU, and YahB. Furthermore, we made available the list of 4,674,808 TFs predicted from 73,873,012 protein sequences in 48,346 genomes. DeepTFactor will serve as a useful tool for predicting TFs, which is necessary for understanding the regulatory systems of organisms of interest. We provide DeepTFactor as a stand-alone program, available at https://bitbucket.org/kaistsystemsbiology/deeptfactor.

Glutaric acid production by systems metabolic engineering of an l-lysine-overproducing Corynebacterium glutamicum.

There is increasing industrial demand for five-carbon platform chemicals, particularly glutaric acid, a widely used building block chemical for the synthesis of polyesters and polyamides. Here we report the development of an efficient glutaric acid microbial producer by systems metabolic engineering of an l-lysine-overproducing Corynebacterium glutamicum BE strain. Based on our previous study, an optimal synthetic metabolic pathway comprising Pseudomonas putida l-lysine monooxygenase (davB) and 5-aminovaleramide amidohydrolase (davA) genes and C. glutamicum 4-aminobutyrate aminotransferase (gabT) and succinate-semialdehyde dehydrogenase (gabD) genes, was introduced into the C. glutamicum BE strain. Through system-wide analyses including genome-scale metabolic simulation, comparative transcriptome analysis, and flux response analysis, 11 target genes to be manipulated were identified and expressed at desired levels to increase the supply of direct precursor l-lysine and reduce precursor loss. A glutaric acid exporter encoded by ynfM was discovered and overexpressed to further enhance glutaric acid production. Fermentation conditions, including oxygen transfer rate, batch-phase glucose level, and nutrient feeding strategy, were optimized for the efficient production of glutaric acid. Fed-batch culture of the final engineered strain produced 105.3 g/L of glutaric acid in 69 h without any byproduct. The strategies of metabolic engineering and fermentation optimization described here will be useful for developing engineered microorganisms for the high-level bio-based production of other chemicals of interest to industry.

Tools and strategies of systems metabolic engineering for the development of microbial cell factories for chemical production.

Sustainable production of chemicals from renewable non-food biomass has become a promising alternative to overcome environmental issues caused by our heavy dependence on fossil resources. Systems metabolic engineering, which integrates traditional metabolic engineering with systems biology, synthetic biology, and evolutionary engineering, is enabling the development of microbial cell factories capable of efficiently producing a myriad of chemicals and materials including biofuels, bulk and fine chemicals, polymers, amino acids, natural products and drugs. In this paper, many tools and strategies of systems metabolic engineering, including in silico genome-scale metabolic simulation, sophisticated enzyme engineering, optimal gene expression modulation, in vivo biosensors, de novo pathway design, and genomic engineering, employed for developing microbial cell factories are reviewed. Also, detailed procedures of systems metabolic engineering used to develop microbial strains producing chemicals and materials are showcased. Finally, future challenges and perspectives in further advancing systems metabolic engineering and establishing biorefineries are discussed.

Enhanced succinic acid production by Mannheimia employing optimal malate dehydrogenase.

Succinic acid (SA), a dicarboxylic acid of industrial importance, can be efficiently produced by metabolically engineered Mannheimia succiniciproducens. Malate dehydrogenase (MDH) is one of the key enzymes for SA production, but has not been well characterized. Here we report biochemical and structural analyses of various MDHs and development of hyper-SA producing M. succiniciproducens by introducing the best MDH. Corynebacterium glutamicum MDH (CgMDH) shows the highest specific activity and least substrate inhibition, whereas M. succiniciproducens MDH (MsMDH) shows low specific activity at physiological pH and strong uncompetitive inhibition toward oxaloacetate (ki of 67.4 and 588.9 µM for MsMDH and CgMDH, respectively). Structural comparison of the two MDHs reveals a key residue influencing the specific activity and susceptibility to substrate inhibition. A high-inoculum fed-batch fermentation of the final strain expressing cgmdh produces 134.25 g L-1 of SA with the maximum productivity of 21.3 g L-1 h-1, demonstrating the importance of enzyme optimization in strain development.

Machine learning applications in systems metabolic engineering.

Systems metabolic engineering allows efficient development of high performing microbial strains for the sustainable production of chemicals and materials. In recent years, increasing availability of bio big data, for example, omics data, has led to active application of machine learning techniques across various stages of systems metabolic engineering, including host strain selection, metabolic pathway reconstruction, metabolic flux optimization, and fermentation. In this paper, recent contributions of machine learning approaches to each major step of systems metabolic engineering are discussed. As the use of machine learning in systems metabolic engineering will become more widespread in accordance with the ever-increasing volume of bio big data, future prospects are also provided for the successful applications of machine learning.

Current status and applications of genome-scale metabolic models.

Genome-scale metabolic models (GEMs) computationally describe gene-protein-reaction associations for entire metabolic genes in an organism, and can be simulated to predict metabolic fluxes for various systems-level metabolic studies. Since the first GEM for Haemophilus influenzae was reported in 1999, advances have been made to develop and simulate GEMs for an increasing number of organisms across bacteria, archaea, and eukarya. Here, we review current reconstructed GEMs and discuss their applications, including strain development for chemicals and materials production, drug targeting in pathogens, prediction of enzyme functions, pan-reactome analysis, modeling interactions among multiple cells or organisms, and understanding human diseases.

Enhanced production of poly­3­hydroxybutyrate (PHB) by expression of response regulator DR1558 in recombinant Escherichia coli.

During microbial production of target product, accumulation of by-products and target product itself may be toxic to host strain. Thus, development of abiotic stress tolerant strains are essential to achieve high productivity of target product with sustained metabolism. Expression of DR1558 from Deinococcus radiodurans, a response regulator in two-component signal transduction system, was reported to increase the tolerance against oxidative stress in Escherichia coli. In this study, the effect of overexpression of DR1558 was examined on poly­3­hydroxybutyrate (PHB) production in recombinant E. coli expressing Ralstonia eutropha PHB biosynthesis genes. It was found that dr1558 overexpressing E. coli produced 5.31 g PHB/L and 9.24 g dry cell weight/L, while control strain produced 1.52 g PHB/L and 4.47 g dry cell weight/L in 48 h shake-flask cultivation. Transcriptional analysis of E. coli suggested that DR1558 could improve the expression efficiency of the genes related to central carbon metabolism and threonine bypass pathway in PHB producing E. coli. When thrABC genes were overexpressed, PHB content was increased in recombinant E. coli, which suggests that stress-tolerant genes from extremophiles should be useful in the development of engineered strains for the production of bio-based products.

Metabolic engineering of Corynebacterium glutamicum for the production of glutaric acid, a C5 dicarboxylic acid platform chemical.

Corynebacterium glutamicum was metabolically engineered for the production of glutaric acid, a C5 dicarboxylic acid that can be used as platform building block chemical for nylons and plasticizers. C. glutamicum gabT and gabD genes and Pseudomonas putida davT and davD genes encoding 5-aminovalerate transaminase and glutarate semialdehyde dehydrogenase, respectively, were examined in C. glutamicum for the construction of a glutaric acid biosynthesis pathway along with P. putida davB and davA genes encoding lysine 2-monooxygenase and delta-aminovaleramidase, respectively. The glutaric acid biosynthesis pathway constructed in recombinant C. glutamicum was engineered by examining strong synthetic promoters PH30 and PH36, C. glutamicum codon-optimized davTDBA genes, and modification of davB gene with an N-terminal His6-tag to improve the production of glutaric acid. It was found that use of N-terminal His6-tagged DavB was most suitable for the production of glutaric acid from glucose. Fed-batch fermentation using the final engineered C. glutamicum H30_GAHis strain, expressing davTDA genes along with davB fused with His6-tag at N-terminus could produce 24.5 g/L of glutaric acid with low accumulation of l-lysine (1.7 g/L), wherein 5-AVA accumulation was not observed during fermentation.