Synthetic Modules Research Articles

Modular Tunable α‐Helical Peptide Hydrogels for Neuronal Cells

AbstractThe assembly of synthetic modules into user‐defined hydrogels for cell culture and clinical applications carries significant advantages over naturally derived materials. Modularity is exploited in synthetic‐polymer‐based hydrogels to alter their physical properties, and spatiotemporally incorporate signaling peptides and growth factors. By contrast, the design of self‐assembling peptide hydrogels has focused largely on assessing cellular responses to unmodified hydrogels and introducing small cell‐adhesive ligands. In short, the modularity of peptide hydrogels is yet to be fully exploited. Previously, hydrogelating self‐assembling fibers (hSAFs) are reported in which two de novo designed α‐helical peptides form gels when mixed. Here, rational peptide design and engineering are used to alter the elastic properties of these hydrogels, to introduce adhesion peptides and growth factors, and to spatially pattern proteins within them. Analysis of the viability and morphology of primary rat neurons cultured on these hydrogels indicates that the modularity of the hSAF system provides a flexible platform for manipulating cell behavior, with potential future applications for modeling natural systems in vitro and regenerative medicine in vivo.

Genomic Comparisons Revealed the Key Genotypes of Streptomyces sp. CB03234-GS26 to Optimize Its Growth and Relevant Production of Tiancimycins.

Strain robustness and titer improvement are major challenges faced in the industrial development of natural products from Streptomyces. Tiancimycins (TNMs) produced by Streptomyces sp. CB03234 are promising anticancer payloads for antibody-drug conjugates, but further development is severely limited by the low titer of TNMs. Despite many efforts to generate various TNMs overproducers, the mechanisms underlying high TNMs production remain to be explored. Herein, genome resequencing and genomic comparisons of different TNMs overproducers were conducted to explore the unique genotypes in CB03234-GS26. Four target genes were selected for further bioinformatic analyses and genetic validations. The results indicated that the inactivation of histidine ammonia-lyase (HAL) showed the most significant effect by blocking the intracellular degradation of histidine to facilitate relevant enzymatic catalysis and thus improve the production of TNMs. Additionally, the potassium/proton antiporter (P/PA) was crucial for intracellular pH homeostasis, and its deficiency severely impaired the alkaline tolerance of the cells. Subsequent pan-genomic analysis suggested that HAL and P/PA are core enzymes that are highly conserved in Streptomyces. Therefore, HAL and P/PA represented novel targets to regulate secondary metabolism and enhance strain robustness and could become potential synthetic biological modules to facilitate development of natural products and strain improvement in Streptomyces.

Engineering Plant Metabolism for Synthesizing Amino Acid Derivatives of Animal Origin Using a Synthetic Modular Approach.

The biosynthesis of amino acid derivatives of animal origin in plants represents a promising frontier in synthetic biology, offering a sustainable and eco-friendly approach to enhancing the nutritional value of plant-based diets. This study leverages the versatile capabilities of Nicotiana benthamiana as a transient expression system to test a synthetic modular framework for the production of creatine, carnosine, and taurine-compounds typically absent in plants but essential for human health. By designing and stacking specialized synthetic modules, we successfully redirected the plant metabolic flux toward the synthesis of these amino acid derivatives of animal origin. Our results revealed the expression of a standalone creatine module resulted in the production of 2.3 μg/g fresh weight of creatine in N. benthamiana leaves. Integrating two modules significantly carnosine yield increased by 3.8-fold and minimized the impact on plant amino acid metabolism compared to individual module application. Unexpectedly, introducing the taurine module caused a feedback-like inhibition of plant cysteine biosynthesis, revealing complex metabolic adjustments that can occur when introducing foreign pathways. Our findings underline the potential for employing plants as biofactories for the sustainable production of essential nutrients of animal origin.

SARDIMM: High-Speed Near-Memory Processing Architecture for Synthetic Aperture Radar Imaging

The range-Doppler algorithm (RDA), a key technique for generating synthetic aperture radar (SAR) images, offers high-resolution images but requires significant memory resources and involves complex signal processing. Moreover, the multitude of fast Fourier transform (FFT) and inverse fast Fourier transform (IFFT) operations in RDA necessitates high bandwidth and lacks data reuse, leading to bottlenecks. This paper introduces a synthetic aperture radar dual in-line memory module (SARDIMM), which executes RDA operations near memory via near-memory processing (NMP), thereby effectively reducing memory accesses, execution time, and energy consumption. The embedded NMP module in SARDIMM optionally supports a combination of FFT, IFFT, and matched filter operations of the RDA for range and azimuth compression. The operator within the NMP module accelerates the FFT by performing two radix-2 single butterfly operations in parallel. The NMP module was implemented and validated on a Xilinx UltraScale+ field-programmable gate array (FPGA) using Verilog-HDL. The acceleration performance of RDA for images of various sizes was evaluated through a simulator modified with gem5 and DRAMSim3 and achieved a 6.34–6.93× speedup and 41.9–48.2% energy savings.

Characterization of Mild Acid Stress Response in an Engineered Acid-Tolerant Escherichia coli Strain.

Engineering acid-tolerant microbial strains is a cost-effective approach to overcoming acid stress during industrial fermentation. We previously constructed an acid-tolerant strain (Escherichia coli SC3124) with enhanced growth robustness and productivity under mildly acidic conditions by fine-tuning the expression of synthetic acid-tolerance module genes consisting of a proton-consuming acid resistance system (gadE), a periplasmic chaperone (hdeB), and ROS scavengers (sodB, katE). However, the precise acid-tolerance mechanism of E. coli SC3124 remained unclear. In this study, the growth of E. coli SC3124 under mild acid stress (pH 6.0) was determined. The final OD600 of E. coli SC3124 at pH 6.0 was 131% and 124% of that of the parent E. coli MG1655 at pH 6.8 and pH 6.0, respectively. Transcriptome analysis revealed the significant upregulation of the genes involved in oxidative phosphorylation, the tricarboxylic acid (TCA) cycle, and lysine-dependent acid-resistance system in E. coli SC3124 at pH 6.0. Subsequently, a weighted gene coexpression network analysis was performed to systematically determine the metabolic perturbations of E. coli SC3124 with mild acid treatment, and we extracted the gene modules highly associated with different acid traits. The results showed two biologically significant coexpression modules, and 263 hub genes were identified. Specifically, the genes involved in ATP-binding cassette (ABC) transporters, oxidative phosphorylation, the TCA cycle, amino acid metabolism, and purine metabolism were highly positively associated with mild acid stress responses. We propose that the overexpression of synthetic acid-tolerance genes leads to metabolic changes that confer mild acid stress resistance in E. coli. Integrated omics platforms provide valuable information for understanding the regulatory mechanisms of mild acid tolerance in E. coli and highlight the important roles of oxidative phosphorylation and ABC transporters in mild acid stress regulation. These findings offer novel insights to better the design of acid-tolerant chasses to synthesize value-added chemicals in a green and sustainable manner.

Synthetic Immunology—Building Immunity from the Bottom‐Up with Synthetic Cells

Synthetic cells can advance immunotherapy, offering innovative approaches to understanding and enhancing immune responses. This review article delves into the advancements and potential of synthetic cell technologies in immunology, emphasizing their role in understanding and manipulating immune functions. Recent progress in understanding vertebrate immune systems and the challenges posed by diseases highlight the need for innovative research methods, complementing the analysis of multidimensional datasets and genetic engineering. Synthetic immune cell engineering aims to simplify the complexity of immunological systems by reconstructing them in a controlled setting. This approach, alongside high‐throughput strategies, facilitates systematic investigations into immunity and the development of novel treatments. The article reviews synthetic cell technologies, focusing on their alignment with the three laws of immunity: universality, tolerance, and appropriateness. It explores the integration of synthetic cell modules to mimic processes such as controlled T‐cell activation, bacteria engulfment and elimination, or cellular maturation into desirable phenotypes. Together, such advancements expand the toolbox for understanding and manipulating immune functions. Synthetic cell technologies stand at the innovation crossroads in immunology, promising to illuminate fundamental immune system principles and open new avenues for research and therapy.

Classification of lung cancer subtypes on CT images with synthetic pathological priors

The accurate diagnosis on pathological subtypes for lung cancer is of significant importance for the follow-up treatments and prognosis managements. In this paper, we propose self-generating hybrid feature network (SGHF-Net) for accurately classifying lung cancer subtypes on computed tomography (CT) images. Inspired by studies stating that cross-scale associations exist in the image patterns between the same case’s CT images and its pathological images, we innovatively developed a pathological feature synthetic module (PFSM), which quantitatively maps cross-modality associations through deep neural networks, to derive the “gold standard” information contained in the corresponding pathological images from CT images. Additionally, we designed a radiological feature extraction module (RFEM) to directly acquire CT image information and integrated it with the pathological priors under an effective feature fusion framework, enabling the entire classification model to generate more indicative and specific pathologically related features and eventually output more accurate predictions. The superiority of the proposed model lies in its ability to self-generate hybrid features that contain multi-modality image information based on a single-modality input. To evaluate the effectiveness, adaptability, and generalization ability of our model, we performed extensive experiments on a large-scale multi-center dataset (i.e., 829 cases from three hospitals) to compare our model and a series of state-of-the-art (SOTA) classification models. The experimental results demonstrated the superiority of our model for lung cancer subtypes classification with significant accuracy improvements in terms of accuracy (ACC), area under the curve (AUC), positive predictive value (PPV) and F1-score.

A Modular Cloning Toolkit for the production of recombinant proteins in Leishmania tarentolae.

Modular Cloning (MoClo) is based on libraries of standardized genetic parts that can be directionally assembled via Golden Gate cloning in one-pot reactions into transcription units and multigene constructs. Here, a team of bachelor students established a MoClo toolkit for the protist Leishmania tarentolae in the frame of the international Genetically Engineered Machine (iGEM) competition. Our modular toolkit is based on a domesticated version of a commercial LEXSY expression vector and comprises 34 genetic parts encoding various affinity tags, targeting signals as well as fluorescent and luminescent proteins. We demonstrated the utility of our kit by the successful production of 16 different tagged versions of the receptor binding domain (RBD) of the SARS-CoV-2 spike protein in L. tarentolae liquid cultures. While highest yields of secreted recombinant RBD were obtained for GST-tagged fusion proteins 48 h post induction, C-terminal peptide tags were often degraded and resulted in lower yields of secreted RBD. Fusing secreted RBD to a synthetic O-glycosylation SP20 module resulted in an apparent molecular mass shift around 10 kDa. No disadvantage regarding the production of RBD was detected when the three antibiotics of the LEXSY system were omitted during the 48-h induction phase. Furthermore, the successful purification of secreted RBD from the supernatant of L. tarentolae liquid cultures was demonstrated in pilot experiments. In summary, we established a MoClo toolkit and exemplified its application for the production of recombinant proteins in L. tarentolae.

Thermally-resilient, phase-invertible, ultra-stable all-aqueous compartments by pH-modulated protein colloidal particles

The essence of compartmentalization in cells is the inspiration behind the engineering of synthetic counterparts, which has emerged as a significant engineering theme. Here, we report the formation of ultra-stable water-in-water (W/W) emulsion droplets. These W/W droplets demonstrate previously unattained stability across a broad pH spectrum and exhibit resilience at temperatures up to 80℃, overcoming the challenge of insufficient robustness in dispersed droplets of aqueous two-phase systems (ATPS). The exceptional robustness is attributed to the strong anchoring of micelle-like casein colloidal particles at the PEO/DEX interface, which maintains stability under varying environmental conditions. The increased surface hydrophobicity of these particles at high temperatures contributes to the formation of thermally-stable droplets, enduring temperatures as high as 80℃. Furthermore, our study illustrates the adaptable affinity of micelle-like casein colloidal particles towards the PEO/DEX-rich phase, enabling the formation of stable DEX-in-PEO emulsions at lower pH levels, and PEO-in-DEX emulsions as the pH rises above the isoelectric point. The robust nature of these W/W emulsions unlocks new possibilities for exploring various biochemical reactions within synthetic subcellular modules and lays a solid foundation for the development of novel biomimetic materials.

Multi-modular metabolic engineering of heme synthesis in Corynebacterium glutamicum

Heme, an iron-containing porphyrin derivative, holds great promise in fields like medicine, food production and chemicals. Here, we developed an engineered Corynebacterium glutamicum strain for efficient heme production by combining modular engineering and RBS engineering. The whole heme biosynthetic pathway was methodically divided into 5-ALA synthetic module, uroporphyrinogen III (UPG III) synthetic module and heme synthetic module for further construction and optimization. Three heme synthetic modules were compared and the siroheme-dependent (SHD) pathway was identified to be optimal in C. glutamicum for the first time. To further improve heme production, the expression of genes in UPG III synthetic module and heme synthetic module was coordinated optimized through RBS engineering, respectively. Subsequently, heme oxygenase was knocked out to reduce heme degradation. The engineered strain HS12 showed a maximum iron-containing porphyrin derivatives titer of 1592 mg/L with the extracellular secretion rate of 45.5% in fed-batch fermentation. Our study constructed a C. glutamicum chassis strain for efficient heme accumulation, which was beneficial for the advancement of efficient heme and other porphyrins production.

Designing a Synthetic 3D-Printed Knee Cartilage: FEA Model, Micro-Structure and Mechanical Characteristics

Articular cartilage morphology and composition are essential factors in joint biomechanics, and their alteration is a crucial aspect of osteoarthritis (OA), a prevalent disease that causes pain and functional loss. This research focuses on developing patient-specific synthetic cartilage using innovative Digital Anatomy polymers. The objectives include investigating the morphology, characterizing the mechanical properties, and replicating the architecture of natural cartilage. This approach offers potential alternatives to traditional manufacturing methods and reduces the need for expensive in vivo experiments. Finite Element Analysis (FEA) validates a novel patient-specific measurement setup. It provides insights into the role of morphology in the distribution of stress and strain within cartilage. CAD design is also utilized to create standardized fiber-reinforced samples that mimic the layered micro-architecture of natural cartilage, allowing for the study of their contribution to the overall mechanical properties. The results demonstrate that 3D-printed polymers can effectively replicate the elastic properties of cartilage. The proposed patient-specific simulator produces reliable results, which have been validated through FEM analysis. While the recreated microstructure closely resembles biological cartilage samples, the elastic properties are slightly underestimated. In conclusion, designing an in silico knee joint is a feasible approach that offers numerous advantages for further development. The Young’s modulus values of our synthetic cartilage modules range from 2.43 MPa to 7.24 MPa, within the range reported in the literature. Moreover, Young´s modulus at the micro level shows the differences between surface 1.74 MPa and internal substrate 1.83 MPa depending on the fiber orientation. Finally, our model proves to be mechanically and morphologically accurate at both the macro and micro levels.

A synthetic cell-free 36-enzyme reaction system for vitamin B12 production

Adenosylcobalamin (AdoCbl), a biologically active form of vitamin B12 (coenzyme B12), is one of the most complex metal-containing natural compounds and an essential vitamin for animals. However, AdoCbl can only be de novo synthesized by prokaryotes, and its industrial manufacturing to date was limited to bacterial fermentation. Here, we report a method for the synthesis of AdoCbl based on a cell-free reaction system performing a cascade of catalytic reactions from 5-aminolevulinic acid (5-ALA), an inexpensive compound. More than 30 biocatalytic reactions are integrated and optimized to achieve the complete cell-free synthesis of AdoCbl, after overcoming feedback inhibition, the complicated detection, instability of intermediate products, as well as imbalance and competition of cofactors. In the end, this cell-free system produces 417.41 μg/L and 5.78 mg/L of AdoCbl using 5-ALA and the purified intermediate product hydrogenobyrate as substrates, respectively. The strategies of coordinating synthetic modules of complex cell-free system describe here will be generally useful for developing cell-free platforms to produce complex natural compounds with long and complicated biosynthetic pathways.

Multivariate modular metabolic engineering for enhanced l-methionine biosynthesis in Escherichia coli

Backgroundl-Methionine is the only bulk amino acid that has not been industrially produced by the fermentation method. Due to highly complex and strictly regulated biosynthesis, the development of microbial strains for high-level l-methionine production has remained challenging in recent years.ResultsBy strengthening the l-methionine terminal synthetic module via site-directed mutation of l-homoserine O-succinyltransferase (MetA) and overexpression of metAfbr, metC, and yjeH, l-methionine production was increased to 1.93 g/L in shake flask fermentation. Deletion of the pykA and pykF genes further improved l-methionine production to 2.51 g/L in shake flask fermentation. Computer simulation and auxotrophic experiments verified that during the synthesis of l-methionine, equimolar amounts of l-isoleucine were accumulated via the elimination reaction of cystathionine γ-synthetase MetB due to the insufficient supply of l-cysteine. To increase the supply of l-cysteine, the l-cysteine synthetic module was strengthened by overexpression of cysEfbr, serAfbr, and cysDN, which further increased the production of l-methionine by 52.9% and significantly reduced the accumulation of the byproduct l-isoleucine by 29.1%. After optimizing the addition of ammonium thiosulfate, the final metabolically engineered strain MET17 produced 21.28 g/L l-methionine in 64 h with glucose as the carbon source in a 5 L fermenter, representing the highest l-methionine titer reported to date.ConclusionsIn this study, a high-efficiency strain for l-methionine production was derived from wild-type Escherichia coli W3110 by rational metabolic engineering strategies, providing an efficient platform for the industrial production of l-methionine.

Multi-step biosynthesis of the biodegradable polyester monomer 2-pyrone-4,6-dicarboxylic acid from glucose

Background2-Pyrone-4,6-dicarboxylic acid (PDC), a chemically stable pseudoaromatic dicarboxylic acid, represents a promising building block for the manufacture of biodegradable polyesters. Microbial production of PDC has been extensively investigated, but low titers and yields have limited industrial applications.ResultsIn this study, a multi-step biosynthesis strategy for the microbial production of PDC was demonstrated using engineered Escherichia coli whole-cell biocatalysts. The PDC biosynthetic pathway was first divided into three synthetic modules, namely the 3-dehydroshikimic acid (DHS) module, the protocatechuic acid (PCA) module and the PDC module. Several effective enzymes, including 3-dehydroshikimate dehydratase for the PCA module as well as protocatechuate 4,5-dioxygenase and 4-carboxy-2-hydroxymuconate-6-semialdehyde dehydrogenase for the PDC module were isolated and characterized. Then, the highly efficient whole-cell bioconversion systems for producing PCA and PDC were constructed and optimized, respectively. Finally, the efficient multi-step biosynthesis of PDC from glucose was achieved by smoothly integrating the above three biosynthetic modules, resulting in a final titer of 49.18 g/L with an overall 27.2% molar yield, which represented the highest titer for PDC production from glucose reported to date.ConclusionsThis study lays the foundation for the microbial production of PDC, including one-step de novo biosynthesis from glucose as well as the microbial transformation of monoaromatics.

Synthesis of Piperazic Acid‐Containing Cyclodepsipeptide Core of Verucopeptin

Comprehensive SummaryChemically, N1 nitrogen of piperazic acid is more nucleophilic than N2 nitrogen, but amide bonds predominantly formed at N2 nitrogen are prevalent in piperazic acid‐containing natural products, with only one exception of sanglifehrin. Thus two orthogonal protecting groups of nitrogen are often employed to realize selective coupling of N2 nitrogen, resulting in increased synthetic steps and low synthetic efficiency. However, we develop selective deprotection of N2‐Cbz from the N1,N2‐diCbz‐piperazic acid‐containing peptide to form the N2 amide exclusively, avoiding the tedious orthogonal protection strategy which is commonly applied when the easily‐accessible N1,N2‐diCbz‐piperazic acid is used as the synthetic module. We employ this selective‐deprotection strategy to achieve an efficient synthesis of piperazic acid‐containing cyclodepsipeptide core of verucopeptin with an overall yield of 21% from commercially/readily available building blocks. The key steps include late stage coupling of piperazic acid with 3‐hydroxyleucine derivatives, and HATU‐mediated macrolactamization of 19‐membered macrocycle at N9 and C10. The selective deprotection of N2‐Cbz from the N1,N2‐diCbz‐piperazic acid at late‐stage would greatly facilitate the total syntheses of piperazic acid‐containing cyclodepsipeptides of biological interest.

Design of diversified chimeric antigen receptors through rational module recombination

Chimeric antigen receptor (CAR)-T cells have shown great promise in cancer therapy. However, the anti-tumor efficiency is limited due to the CAR-induced Tcell apoptosis or exhaustion. The intracellular domain of CAR comprised of various signaling modules orchestrates CAR-T cell behaviors. The modularity of CAR signaling domain functions as the "mainboard" to assemble diversified downstream signaling components. Here, we implemented the modular recombination strategy to construct a library of CARs with synthetic co-signaling modules adopted from immunoglobin-like superfamily (IgSF) and tumor necrosis factor receptor superfamily (TNFRSF). We quantitatively characterized the signaling behaviors of these recombinants by both NFAT and NF-κB reporter, and identified a set of new CARs with diverse signaling behaviors. Specifically, the 28(NM)-BB(MC) CAR-T cells exhibited improved cytotoxicity and Tcell persistence. The synthetic approach can promote our understanding of the signaling principles of CAR molecule, and provide a powerful tool box for CAR-T cell engineering.

Instructional materials that control cellular activity through synthetic Notch receptors

The field of regenerative engineering relies primarily on the dual technical platforms of cell selection/conditioning and biomaterial fabrication to support directed cell differentiation. As the field has matured, an appreciation for the influence of biomaterials on cell behaviors has resulted in engineered matrices that meet biomechanical and biochemical demands of target pathologies. Yet, despite advances in methods to produce designer matrices, regenerative engineers remain unable to reliably orchestrate behaviors of therapeutic cells in situ. Here, we present a platform named MATRIX whereby cellular responses to biomaterials can be custom defined by combining engineered materials with cells expressing cognate synthetic biology control modules. Such privileged channels of material-to-cell communication can activate synthetic Notch receptors and govern activities as diverse as transcriptome engineering, inflammation attenuation, and pluripotent stem cell differentiation, all in response to materials decorated with otherwise bioinert ligands. Further, we show that engineered cellular behaviors are confined to programmed biomaterial surfaces, highlighting the potential to use this platform to spatially organize cellular responses to bulk, soluble factors. This integrated approach of co-engineering cells and biomaterials for orthogonal interactions opens new avenues for reproducible control of cell-based therapies and tissue replacements.

The cyclic peptide G4CP2 enables the modulation of galactose metabolism in yeast by interfering with GAL4 transcriptional activity.

Genetically-encoded combinatorial peptide libraries are convenient tools to identify peptides to be used as therapeutics, antimicrobials and functional synthetic biology modules. Here, we report the identification and characterization of a cyclic peptide, G4CP2, that interferes with the GAL4 protein, a transcription factor responsible for the activation of galactose catabolism in yeast and widely exploited in molecular biology. G4CP2 was identified by screening CYCLIC, a Yeast Two-Hybrid-based combinatorial library of cyclic peptides developed in our laboratory. G4CP2 interferes with GAL4-mediated activation of galactose metabolic enzymes both when expressed intracellularly, as a recombinant peptide, and when provided exogenously, as a chemically-synthesized cyclic peptide. Our results support the application of G4CP2 in microbial biotechnology and, additionally, demonstrate that CYCLIC can be used as a tool for the rapid identification of peptides, virtually without any limitations with respect to the target protein. The possible biotechnological applications of cyclic peptides are also discussed.

TUTOR: Training Neural Networks Using Decision Rules as Model Priors

The human brain has the ability to carry out new tasks with limited experience. It utilizes prior learning experiences to adapt the solution strategy to new domains. On the other hand, deep neural networks (DNNs) generally need large amounts of data and computational resources for training. However, this requirement is not met in many settings. To address these challenges, we propose the TUTOR (training neural networks using decision rules as model priors) DNN synthesis framework. TUTOR targets tabular datasets. It synthesizes accurate DNN models with limited available data and reduced memory/computational requirements. It consists of three sequential steps. The first step involves generation, verification, and labeling of synthetic data. The synthetic data generation module targets both categorical and continuous features. TUTOR generates synthetic data from the same probability distribution as the real data. It then verifies the integrity of the generated synthetic data using a semantic integrity classifier module. It labels synthetic data based on a set of rules extracted from the real dataset. Next, TUTOR uses two training schemes that combine synthetic and training data to learn the DNN model parameters. These two schemes focus on two different ways in which synthetic data can be used to derive a prior on the model parameters and, hence, provide a better DNN initialization for training with real data. In the third step, TUTOR employs a grow-and-prune synthesis paradigm to learn both the weights and the architecture of the DNN to reduce model size while ensuring its accuracy. We evaluate the performance of TUTOR on nine datasets of various sizes. We show that in comparison to fully connected DNNs, TUTOR, on an average, reduces the need for data by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$5.9\times $ </tex-math></inline-formula> (geometric mean), improves accuracy by 3.4%, and reduces the number of parameters (floating-point operations) by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$4.7\times $ </tex-math></inline-formula> ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$4.3\times $ </tex-math></inline-formula> ) (geometric mean). Thus, TUTOR enables less data-hungry, more accurate, and more compact DNN synthesis.

Systems metabolic engineering upgrades Corynebacterium glutamicum to high-efficiency cis, cis-muconic acid production from lignin-based aromatics

Lignin displays a highly challenging renewable. To date, massive amounts of lignin, generated in lignocellulosic processing facilities, are for the most part merely burned due to lacking value-added alternatives. Aromatic lignin monomers of recognized relevance are in particular vanillin, and to a lesser extent vanillate, because they are accessible at high yield from softwood-lignin using industrially operated alkaline oxidative depolymerization. Here, we metabolically engineered C. glutamicum towards cis, cis-muconate (MA) production from these key aromatics. Starting from the previously created catechol-based producer C. glutamicum MA-2, systems metabolic engineering first discovered an unspecific aromatic aldehyde reductase that formed aromatic alcohols from vanillin, protocatechualdehyde, and p- hydroxybenzaldehyde, and was responsible for the conversion up to 57% of vanillin into vanillyl alcohol. The alcohol was not re-consumed by the microbe later, posing a strong drawback on the producer. The identification and subsequent elimination of the encoding fudC gene completely abolished vanillyl alcohol formation. Second, the initially weak flux through the native vanillin and vanillate metabolism was enhanced up to 2.9-fold by implementing synthetic pathway modules. Third, the most efficient protocatechuate decarboxylase AroY for conversion of the midstream pathway intermediate protocatechuate into catechol was identified out of several variants in native and codon optimized form and expressed together with the respective helper proteins. Fourth, the streamlined modules were all genomically combined which yielded the final strain MA-9. MA-9 produced bio-based MA from vanillin, vanillate, and seven structurally related aromatics at maximum selectivity. In addition, MA production from softwood-based vanillin, obtained through alkaline depolymerization, was demonstrated.