Large scale active-learning-guided exploration for in vitro protein production optimization

Chloroplast Cell-Free Systems from Different Plant Species as a Rapid Prototyping Platform

Climate change poses a significant threat to global agriculture, necessitating innovative solutions. Plant synthetic biology, particularly chloroplast engineering, holds promise as a viable approach to this challenge. Chloroplasts present a variety of advantageous traits for genetic engineering, but the development of genetic tools and genetic part characterization in these organelles is hindered by the lengthy time scales required to generate transplastomic organisms. To address these challenges, we have established a versatile protocol for generating highly active chloroplast-based cell-free gene expression (CFE) systems derived from a diverse range of plant species, including wheat (monocot), spinach, and poplar trees (dicots). We show that these systems work with conventionally used T7 RNA polymerase as well as the endogenous chloroplast polymerases, allowing for detailed characterization and prototyping of regulatory sequences at both transcription and translation levels. To demonstrate the platform for characterization of promoters and 5' and 3' untranslated regions (UTRs) in higher plant chloroplast gene expression, we analyze a collection of 23 5'UTRs, 10 3'UTRs, and 6 chloroplast promoters, assessed their expression in spinach and wheat extracts, and found consistency in expression patterns, suggesting cross-species compatibility. Looking forward, our chloroplast CFE systems open new avenues for plant synthetic biology, offering prototyping tools for both understanding gene expression and developing engineered plants, which could help meet the demands of a changing global climate.

Developing, characterizing and modeling CRISPR-based point-of-use pathogen diagnostics

Recent years have seen intense interest in the development of point-of-care nucleic acid diagnostic technologies to address the scaling limitations of laboratory-based approaches. Chief among these are combinations of isothermal amplification approaches with CRISPR-based detection and readouts of target products. Here, we contribute to the growing body of rapid, programmable point-of-care pathogen tests by developing and optimizing a one-pot NASBA-Cas13a nucleic acid detection assay. This test uses the isothermal amplification technique NASBA to amplify target viral nucleic acids, followed by Cas13a-based detection of amplified sequences. We first demonstrate an in-house formulation of NASBA that enables optimization of individual NASBA components. We then present design rules for NASBA primer sets and LbuCas13a guide RNAs for fast and sensitive detection of SARS-CoV-2 viral RNA fragments, resulting in 20 - 200 aM sensitivity without any specialized equipment. Finally, we explore the combination of high-throughput assay condition screening with mechanistic ordinary differential equation modeling of the reaction scheme to gain a deeper understanding of the NASBA-Cas13a system. This work presents a framework for developing a mechanistic understanding of reaction performance and optimization that uses both experiments and modeling, which we anticipate will be useful in developing future nucleic acid detection technologies.

Accelerating Genetic Sensor Development, Scale-up, and Deployment Using Synthetic Biology

Living cells are exquisitely tuned to sense and respond to changes in their environment. Repurposing these systems to create engineered biosensors has seen growing interest in the field of synthetic biology and provides a foundation for many innovative applications spanning environmental monitoring to improved biobased production. In this review, we present a detailed overview of currently available biosensors and the methods that have supported their development, scale-up, and deployment. We focus on genetic sensors in living cells whose outputs affect gene expression. We find that emerging high-throughput experimental assays and evolutionary approaches combined with advanced bioinformatics and machine learning are establishing pipelines to produce genetic sensors for virtually any small molecule, protein, or nucleic acid. However, more complex sensing tasks based on classifying compositions of many stimuli and the reliable deployment of these systems into real-world settings remain challenges. We suggest that recent advances in our ability to precisely modify nonmodel organisms and the integration of proven control engineering principles (e.g., feedback) into the broader design of genetic sensing systems will be necessary to overcome these hurdles and realize the immense potential of the field.

Mechanism-based and data-driven modeling in cell-free synthetic biology

Cell-free systems have emerged as a versatile platform in synthetic biology, finding applications in various areas such as prototyping synthetic circuits, biosensor development, and biomanufacturing. To streamline the prototyping process, cell-free systems often incorporate a modeling step that predicts the outcomes of various experimental scenarios, providing a deeper insight into the underlying mechanisms and functions. There are two recognized approaches for modeling these systems: mechanism-based modeling, which models the underlying reaction mechanisms; and data-driven modeling, which makes predictions based on data without preconceived interactions between system components. In this highlight, we focus on the latest advancements in both modeling approaches for cell-free systems, exploring their potential for the design and optimization of synthetic genetic circuits.

Leveraging Active Learning to Establish Efficient In Vitro Transcription and Translation from Bacterial Chromosomal DNA

Gene expression is a fundamental aspect in the construction of a minimal synthetic cell, and the use of chromosomes will be crucial for the integration and regulation of complex modules. Expression from chromosomes in vitro transcription and translation (IVTT) systems presents limitations, as their large size and low concentration make them far less suitable for standard IVTT reactions. Here, we addressed these challenges by optimizing lysate-based IVTT systems at low template concentrations. We then applied an active learning tool to adapt IVTT to chromosomes as template DNA. Further insights into the dynamic data set led us to adjust the previous protocol for chromosome isolation and revealed unforeseen trends pointing at limiting transcription kinetics in our system. The resulting IVTT conditions allowed a high template DNA efficiency for the chromosomes. In conclusion, our system shows a protein-to-chromosome ratio that moves closer to in vivo biology and represents an advancement toward chromosome-based synthetic cells.

Extremophiles in a changing world

Extremophiles and their products have been a major focus of research interest for over 40 years. Through this period, studies of these organisms have contributed hugely to many aspects of the fundamental and applied sciences, and to wider and more philosophical issues such as the origins of life and astrobiology. Our understanding of the cellular adaptations to extreme conditions (such as acid, temperature, pressure and more), of the mechanisms underpinning the stability of macromolecules, and of the subtleties, complexities and limits of fundamental biochemical processes has been informed by research on extremophiles. Extremophiles have also contributed numerous products and processes to the many fields of biotechnology, from diagnostics to bioremediation. Yet, after 40 years of dedicated research, there remains much to be discovered in this field. Fortunately, extremophiles remain an active and vibrant area of research. In the third decade of the twenty-first century, with decreasing global resources and a steadily increasing human population, the world's attention has turned with increasing urgency to issues of sustainability. These global concerns were encapsulated and formalized by the United Nations with the adoption of the 2030 Agenda for Sustainable Development and the presentation of the seventeen Sustainable Development Goals (SDGs) in 2015. In the run-up to 2030, we consider the contributions that extremophiles have made, and will in the future make, to the SDGs.

Challenges in developing cell culture media using machine learning

Microbial and mammalian cells are widely used in the food, pharmaceutical, and medical industries. Developing or optimizing culture media is essential to improve cell culture performance as a critical technology in cell culture engineering. Methodologies for media optimization have been developed to a great extent, such as the approaches of one-factor-at-a-time (OFAT) and response surface methodology (RSM). The present review introduces the emerging machine learning (ML) technology in cell culture engineering by combining high-throughput experimental technologies to develop highly efficient and effective culture media. The commonly used ML algorithms and the successful applications of employing ML in medium optimization are summarized. This review highlights the benefits of ML-assisted medium development and guides the selection of the media optimization method appropriate for various cell culture purposes.

Physical Laboratory Automation in Synthetic Biology

Synthetic Biology has overcome many of the early challenges facing the field and is entering a systems era characterized by adoption of Design-Build-Test-Learn (DBTL) approaches. The need for automation and standardization to enable reproducible, scalable, and translatable research has become increasingly accepted in recent years, and many of the hardware and software tools needed to address these challenges are now in place or under development. However, the lack of connectivity between DBTL modules and barriers to access and adoption remain significant challenges to realizing the full potential of lab automation. In this review, we characterize and classify the state of automation in synthetic biology with a focus on the physical automation of experimental workflows. Though fully autonomous scientific discovery is likely a long way off, impressive progress has been made toward automating critical elements of experimentation by combining intelligent hardware and software tools. It is worth questioning whether total automation that removes humans entirely from the loop should be the ultimate goal, and considerations for appropriate automation versus total automation are discussed in this light while emphasizing areas where further development is needed in both contexts.

RECOVER identifies synergistic drug combinations in vitro through sequential model optimization

For large libraries of small molecules, exhaustive combinatorial chemical screens become infeasible to perform when considering a range of disease models, assay conditions, and dose ranges. Deep learning models have achieved state-of-the-art results in silico for the prediction of synergy scores. However, databases of drug combinations are biased toward synergistic agents and results do not generalize out of distribution. During 5 rounds of experimentation, we employ sequential model optimization with a deep learning model to select drug combinations increasingly enriched for synergism and active against a cancer cell line-evaluating only ∼5% of the total search space. Moreover, we find that learned drug embeddings (using structural information) begin to reflect biological mechanisms. In silico benchmarking suggests search queries are ∼5-10× enriched for highly synergistic drug combinations by using sequential rounds of evaluation when compared with random selection or ∼3× when using a pretrained model.

Optimizing Ethanol Production in Saccharomyces cerevisiae at Ambient and Elevated Temperatures through Machine Learning-Guided Combinatorial Promoter Modifications

Bioethanol has gained popularity in recent decades as an ecofriendly alternative to fossil fuels due to increasing concerns about global climate change. However, economically viable ethanol fermentation remains a challenge. High-temperature fermentation can reduce production costs, but Saccharomyces cerevisiae yeast strains normally ferment poorly under high temperatures. In this study, we present a machine learning (ML) approach to optimize bioethanol production in S. cerevisiae by fine-tuning the promoter activities of three endogenous genes. We created 216 combinatorial strains of S. cerevisiae by replacing native promoters with five promoters of varying strengths to regulate ethanol production. Promoter replacement resulted in a 63% improvement in ethanol production at 30 °C. We created an ML-guided workflow by utilizing XGBoost to train high-performance models based on promoter strengths and cellular metabolite concentrations obtained from ethanol production of 216 combinatorial strains at 30 °C. This strategy was then applied to optimize ethanol production at 40 °C, where we selected 31 strains for experimental fermentation. This reduced experimental load led to a 7.4% increase in ethanol production in the second round of the ML-guided workflow. Our study offers a comprehensive library of promoter strength modifications for key ethanol production enzymes, showcasing how machine learning can guide yeast strain optimization and make bioethanol production more cost-effective and efficient. Furthermore, we demonstrate that metabolic engineering processes can be accelerated and optimized through this approach.

Efforts to Minimise the Bacterial Genome as a Free-Living Growing System

Exploring the minimal genetic requirements for cells to maintain free living is an exciting topic in biology. Multiple approaches are employed to address the question of the minimal genome. In addition to constructing the synthetic genome in the test tube, reducing the size of the wild-type genome is a practical approach for obtaining the essential genomic sequence for living cells. The well-studied Escherichia coli has been used as a model organism for genome reduction owing to its fast growth and easy manipulation. Extensive studies have reported how to reduce the bacterial genome and the collections of genomic disturbed strains acquired, which were sufficiently reviewed previously. However, the common issue of growth decrease caused by genetic disturbance remains largely unaddressed. This mini-review discusses the considerable efforts made to improve growth fitness, which was decreased due to genome reduction. The proposal and perspective are clarified for further accumulated genetic deletion to minimise the Escherichia coli genome in terms of genome reduction, experimental evolution, medium optimization, and machine learning.

Enabling technology and core theory of synthetic biology

Synthetic biology provides a new paradigm for life science research ("build to learn") and opens the future journey of biotechnology ("build to use"). Here, we discuss advances of various principles and technologies in the mainstream of the enabling technology of synthetic biology, including synthesis and assembly of a genome, DNA storage, gene editing, molecular evolution and de novo design of function proteins, cell and gene circuit engineering, cell-free synthetic biology, artificial intelligence (AI)-aided synthetic biology, as well as biofoundries. We also introduce the concept of quantitative synthetic biology, which is guiding synthetic biology towards increased accuracy and predictability or the real rational design. We conclude that synthetic biology will establish its disciplinary system with the iterative development of enabling technologies and the maturity of the core theory.

Scientific discovery in the age of artificial intelligence

Artificial intelligence (AI) is being increasingly integrated into scientific discovery to augment and accelerate research, helping scientists to generate hypotheses, design experiments, collect and interpret large datasets, and gain insights that might not have been possible using traditional scientific methods alone. Here we examine breakthroughs over the past decade that include self-supervised learning, which allows models to be trained on vast amounts of unlabelled data, and geometric deep learning, which leverages knowledge about the structure of scientific data to enhance model accuracy and efficiency. Generative AI methods can create designs, such as small-molecule drugs and proteins, by analysing diverse data modalities, including images and sequences. We discuss how these methods can help scientists throughout the scientific process and the central issues that remain despite such advances. Both developers and users of AI toolsneed a better understanding of when such approaches need improvement, and challenges posed by poor data quality and stewardship remain. These issues cut across scientific disciplines and require developing foundational algorithmic approaches that can contribute to scientific understanding or acquire it autonomously, making them critical areas of focus for AI innovation.

Activated alkyne-enabled turn-on click bioconjugation with cascade signal amplification for ultrafast and high-throughput antibiotic screening

Antimicrobial susceptibility testing plays a pivotal role in the discovery of new antibiotics. However, the development of simple, sensitive, and rapid assessment approaches remains challenging. Herein, we report an activated alkyne-based cascade signal amplification strategy for ultrafast and high-throughput antibiotic screening. First of all, a novel water-soluble aggregation-induced emission (AIE) luminogen is synthesized, which contains an activated alkyne group to enable fluorescence turn-on and metal-free click bioconjugation under physiological conditions. Taking advantage of the in-house established method for bacterial lysis, a number of clickable biological substances (i.e., bacterial solutes and debris) are released from the bacterial bodies, which remarkably increases the quantity of analytes. By means of the activated alkyne-mediated turn-on click bioconjugation, the system fluorescence signal is significantly amplified due to the increased labeling sites as well as the AIE effect. Such a cascade signal amplification strategy efficiently improves the detection sensitivity and thus enables ultrafast antimicrobial susceptibility assessment. By integration with a microplate reader, this approach is further applied to high-throughput antibiotic screening.

Cell-Free Expression System Derived from a Near-Minimal Synthetic Bacterium

Cell-free expression (CFE) systems are fundamental to reconstituting metabolic pathways in vitro toward the construction of a synthetic cell. Although an Escherichia coli-based CFE system is well-established, simpler model organisms are necessary to understand the principles behind life-like behavior. Here, we report the successful creation of a CFE system derived from JCVI-syn3A (Syn3A), the minimal synthetic bacterium. Previously, high ribonuclease activity in Syn3A lysates impeded the establishment of functional CFE systems. Now, we describe how an unusual cell lysis method (nitrogen decompression) yielded Syn3A lysates with reduced ribonuclease activity that supported in vitro expression. To improve the protein yields in the Syn3A CFE system, we optimized the Syn3A CFE reaction mixture using an active machine learning tool. The optimized reaction mixture improved the CFE 3.2-fold compared to the preoptimized condition. This is the first report of a functional CFE system derived from a minimal synthetic bacterium, enabling further advances in bottom-up synthetic biology.

Employing active learning in the optimization of culture medium for mammalian cells

Medium optimization is a crucial step during cell culture for biopharmaceutics and regenerative medicine; however, this step remains challenging, as both media and cells are highly complex systems. Here, we addressed this issue by employing active learning. Specifically, we introduced machine learning to cell culture experiments to optimize culture medium. The cell line HeLa-S3 and the gradient-boosting decision tree algorithm were used to find optimized media as pilot studies. To acquire the training data, cell culture was performed in a large variety of medium combinations. The cellular NAD(P)H abundance, represented as A450, was used to indicate the goodness of culture media. In active learning, regular and time-saving modes were developed using culture data at 168 h and 96 h, respectively. Both modes successfully fine-tuned 29 components to generate a medium for improved cell culture. Intriguingly, the two modes provided different predictions for the concentrations of vitamins and amino acids, and a significant decrease was commonly predicted for fetal bovine serum (FBS) compared to the commercial medium. In addition, active learning-assisted medium optimization significantly increased the cellular concentration of NAD(P)H, an active chemical with a constant abundance in living cells. Our study demonstrated the efficiency and practicality of active learning for medium optimization and provided valuable information for employing machine learning technology in cell biology experiments.

Automating the design-build-test-learn cycle towards next-generation bacterial cell factories

Automation is playing an increasingly significant role in synthetic biology. Groundbreaking technologies, developed over the past 20 years, have enormously accelerated the construction of efficient microbial cell factories. Integrating state-of-the-art tools (e.g. for genome engineering and analytical techniques) into the design-build-test-learn cycle (DBTLc) will shift the metabolic engineering paradigm from an almost artisanal labor towards a fully automated workflow. Here, we provide a perspective on how a fully automated DBTLc could be harnessed to construct the next-generation bacterial cell factories in a fast, high-throughput fashion. Innovative toolsets and approaches that pushed the boundaries in each segment of the cycle are reviewed to this end. We also present the most recent efforts on automation of the DBTLc, which heralds a fully autonomous pipeline for synthetic biology in the near future.

What remains from living cells in bacterial lysate-based cell-free systems

Because they mimic cells while offering an accessible and controllable environment, lysate-based cell-free systems (CFS) have emerged as valuable biotechnology tools for synthetic biology. Historically used to uncover fundamental mechanisms of life, CFS are nowadays used for a multitude of purposes, including protein production and prototyping of synthetic circuits. Despite the conservation of fundamental functions in CFS like transcription and translation, RNAs and certain membrane-embedded or membrane-bound proteins of the host cell are lost when preparing the lysate. As a result, CFS largely lack some essential properties of living cells, such as the ability to adapt to changing conditions, to maintain homeostasis and spatial organization. Regardless of the application, shedding light on the black-box of the bacterial lysate is necessary to fully exploit the potential of CFS. Most measurements of the activity of synthetic circuits in CFS and in vivo show significant correlations because these only require processes that are preserved in CFS, like transcription and translation. However, prototyping circuits of higher complexity that require functions that are lost in CFS (cell adaptation, homeostasis, spatial organization) will not show such a good correlation with in vivo conditions. Both for prototyping circuits of higher complexity and for building artificial cells, the cell-free community has developed devices to reconstruct cellular functions. This mini-review compares bacterial CFS to living cells, focusing on functional and cellular process differences and the latest developments in restoring lost functions through complementation of the lysate or device engineering.

Deep learning for optimization of protein expression

Recent progress in high-throughput DNA synthesis and sequencing has enabled the development of massively parallel reporter assays for strain characterization. These datasets map a large number of DNA sequences to protein expression levels, sparking increased interest in data-driven methods for sequence-to-expression modeling. Here, we highlight advances in deep learning models of protein expression and their potential for optimizing strains engineered to produce recombinant proteins. We review recent works that built highly accurate models and discuss challenges that hinder adoption by end users. There is a need to better align this technology with the constraints encountered in strain engineering, particularly the cost of acquiring large amounts of data and the requirement for interpretable models that generalize beyond the training data. Overcoming these barriers will help to incentivize academic and industrial laboratories to tap into a new era of data-centric strain engineering.

Machine learning for metabolic pathway optimization: A review

Optimizing the metabolic pathways of microbial cell factories is essential for establishing viable biotechnological production processes. However, due to the limited understanding of the complex setup of cellular machinery, building efficient microbial cell factories remains tedious and time-consuming. Machine learning (ML), a powerful tool capable of identifying patterns within large datasets, has been used to analyze biological datasets generated using various high-throughput technologies to build data-driven models for complex bioprocesses. In addition, ML can also be integrated with Design-Build-Test-Learn to accelerate development. This review focuses on recent ML applications in genome-scale metabolic model construction, multistep pathway optimization, rate-limiting enzyme engineering, and gene regulatory element designing. In addition, we have discussed some limitations of these methods as well as potential solutions.

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 automated Galaxy-SynBioCAD pipeline for synthetic biology design and engineering

Here we introduce the Galaxy-SynBioCAD portal, a toolshed for synthetic biology, metabolic engineering, and industrial biotechnology. The tools and workflows currently shared on the portal enables one to build libraries of strains producing desired chemical targets covering an end-to-end metabolic pathway design and engineering process from the selection of strains and targets, the design of DNA parts to be assembled, to the generation of scripts driving liquid handlers for plasmid assembly and strain transformations. Standard formats like SBML and SBOL are used throughout to enforce the compatibility of the tools. In a study carried out at four different sites, we illustrate the link between pathway design and engineering with the building of a library of E. coli lycopene-producing strains. We also benchmark our workflows on literature and expert validated pathways. Overall, we find an 83% success rate in retrieving the validated pathways among the top 10 pathways generated by the workflows.

Variability in cell-free expression reactions can impact qualitative genetic circuit characterization

Cell-free expression systems provide a suite of tools that are used in applications from sensing to biomanufacturing. One of these applications is genetic circuit prototyping, where the lack of cloning is required and a high degree of control over reaction components and conditions enables rapid testing of design candidates. Many studies have shown utility in the approach for characterizing genetic regulation elements, simple genetic circuit motifs, protein variants or metabolic pathways. However, variability in cell-free expression systems is a known challenge, whether between individuals, laboratories, instruments, or batches of materials. While the issue of variability has begun to be quantified and explored, little effort has been put into understanding the implications of this variability. For genetic circuit prototyping, it is unclear when and how significantly variability in reaction activity will impact qualitative assessments of genetic components, e.g. relative activity between promoters. Here, we explore this question by assessing DNA titrations of seven genetic circuits of increasing complexity using reaction conditions that ostensibly follow the same protocol but vary by person, instrument and material batch. Although the raw activities vary widely between the conditions, by normalizing within each circuit across conditions, reasonably consistent qualitative performance emerges for the simpler circuits. For the most complex case involving expression of three proteins, we observe a departure from this qualitative consistency, offering a provisional cautionary line where normal variability may disrupt reliable reuse of prototyping results. Our results also suggest that a previously described closed loop controller circuit may help to mitigate such variability, encouraging further work to design systems that are robust to variability. Graphical Abstract.

A versatile active learning workflow for optimization of genetic and metabolic networks

Optimization of biological networks is often limited by wet lab labor and cost, and the lack of convenient computational tools. Here, we describe METIS, a versatile active machine learning workflow with a simple online interface for the data-driven optimization of biological targets with minimal experiments. We demonstrate our workflow for various applications, including cell-free transcription and translation, genetic circuits, and a 27-variable synthetic CO₂-fixation cycle (CETCH cycle), improving these systems between one and two orders of magnitude. For the CETCH cycle, we explore 10²⁵ conditions with only 1,000 experiments to yield the most efficient CO₂-fixation cascade described to date. Beyond optimization, our workflow also quantifies the relative importance of individual factors to the performance of a system identifying unknown interactions and bottlenecks. Overall, our workflow opens the way for convenient optimization and prototyping of genetic and metabolic networks with customizable adjustments according to user experience, experimental setup, and laboratory facilities.

Optimization of biological networks is often limited by wet lab labor and cost, and the lack of convenient computational tools. Here, aimed at democratization and standardization, the authors describe METIS, a modular and versatile active machine learning workflow with a simple online interface for the optimization of biological target functions with minimal experimental datasets.

Cell-Free Protein Synthesis for High-Throughput Biosynthetic Pathway Prototyping

Biological systems provide a sustainable and complimentary approach to synthesizing useful chemical products. Metabolic engineers seeking to establish economically viable biosynthesis platforms strive to increase product titers, rates, and yields. Despite continued advances in genetic tools and metabolic engineering techniques, cellular workflows remain limited in throughput. It may take months to test dozens of unique pathway designs even in a robust model organism, such as Escherichia coli. In contrast, cell-free protein synthesis enables the rapid generation of enzyme libraries that can be combined to reconstitute metabolic pathways in vitro for biochemical synthesis in days rather than weeks. Cell-free reactions thereby enable comparison of hundreds to thousands of unique combinations of enzyme homologs and concentrations, which can quickly identify the most productive pathway variants to test in vivo or further characterize in vitro. This cell-free pathway prototyping strategy provides a complementary approach to accelerate cellular metabolic engineering efforts toward highly productive strains for metabolite production.

Differentially Optimized Cell-Free Buffer Enables Robust Expression from Unprotected Linear DNA in Exonuclease-Deficient Extracts

The use of linear DNA templates in cell-free systems promises to accelerate the prototyping and engineering of synthetic gene circuits. A key challenge is that linear templates are rapidly degraded by exonucleases present in cell extracts. Current approaches tackle the problem by adding exonuclease inhibitors and DNA-binding proteins to protect the linear DNA, requiring additional time- and resource-intensive steps. Here, we delete the recBCD exonuclease gene cluster from the Escherichia coli BL21 genome. We show that the resulting cell-free systems, with buffers optimized specifically for linear DNA, enable near-plasmid levels of expression from σ70 promoters in linear DNA templates without employing additional protection strategies. When using linear or plasmid DNA templates at the buffer calibration step, the optimal potassium glutamate concentrations obtained when using linear DNA were consistently lower than those obtained when using plasmid DNA for the same extract. We demonstrate the robustness of the exonuclease deficient extracts across seven different batches and a wide range of experimental conditions across two different laboratories. Finally, we illustrate the use of the ΔrecBCD extracts for two applications: toehold switch characterization and enzyme screening. Our work provides a simple, efficient, and cost-effective solution for using linear DNA templates in cell-free systems and highlights the importance of specifically tailoring buffer composition for the final experimental setup. Our data also suggest that similar exonuclease deletion strategies can be applied to other species suitable for cell-free synthetic biology.

Cell-Free Protein Synthesis with Fungal Lysates for the Rapid Production of Unspecific Peroxygenases

Unspecific peroxygenases (UPOs, EC 1.11.2.1) are fungal biocatalysts that have attracted considerable interest for application in chemical syntheses due to their ability to selectively incorporate peroxide-oxygen into non-activated hydrocarbons. However, the number of available and characterized UPOs is limited, as it is difficult to produce these enzymes in homologous or hetero-logous expression systems. In the present study, we introduce a third approach for the expression of UPOs: cell-free protein synthesis using lysates from filamentous fungi. Biomass of Neurospora crassa and Aspergillus niger, respectively, was lysed by French press and tested for translational activity with a luciferase reporter enzyme. The upo1 gene from Cyclocybe (Agrocybe) aegerita (encoding the main peroxygenase, AaeUPO) was cell-free expressed with both lysates, reaching activities of up to 105 U L⁻¹ within 24 h (measured with veratryl alcohol as substrate). The cell-free expressed enzyme (cfAaeUPO) was successfully tested in a substrate screening that included prototypical UPO substrates, as well as several pharmaceuticals. The determined activities and catalytic performance were comparable to that of the wild-type enzyme (wtAaeUPO). The results presented here suggest that cell-free expression could become a valuable tool to gain easier access to the immense pool of putative UPO genes and to expand the spectrum of these sought-after biocatalysts.

High-Throughput Experimentation Using Cell-Free Protein Synthesis Systems

Cell-free protein synthesis can enable the combinatorial screening of many different components and concentrations. However, manual pipetting methods are unfit to handle many cell-free reactions. Here, we describe a microfluidic method that can generate hundreds of unique submicroliter scale reactions. The method is coupled with a high yield cell-free system that can be applied for broad protein screening assays.

Key reaction components affect the kinetics and performance robustness of cell-free protein synthesis reactions

Cell-free protein synthesis (CFPS) reactions have grown in popularity with particular interest in applications such as gene construct prototyping, biosensor technologies and the production of proteins with novel chemistry. Work has frequently focussed on optimising CFPS protocols for improving protein yield, reducing cost, or developing streamlined production protocols. Here we describe a statistical Design of Experiments analysis of 20 components of a popular CFPS reaction buffer. We simultaneously identify factors and factor interactions that impact on protein yield, rate of reaction, lag time and reaction longevity. This systematic experimental approach enables the creation of a statistical model capturing multiple behaviours of CFPS reactions in response to components and their interactions. We show that a novel reaction buffer outperforms the reference reaction by 400% and importantly reduces failures in CFPS across batches of cell lysates, strains of E. coli, and in the synthesis of different proteins. Detailed and quantitative understanding of how reaction components affect kinetic responses and robustness is imperative for future deployment of cell-free technologies.

Biotechnology Applications of Cell-Free Expression Systems

Cell-free systems are a rapidly expanding platform technology with an important role in the engineering of biological systems. The key advantages that drive their broad adoption are increased efficiency, versatility, and low cost compared to in vivo systems. Traditionally, in vivo platforms have been used to synthesize novel and industrially relevant proteins and serve as a testbed for prototyping numerous biotechnologies such as genetic circuits and biosensors. Although in vivo platforms currently have many applications within biotechnology, they are hindered by time-constraining growth cycles, homeostatic considerations, and limited adaptability in production. Conversely, cell-free platforms are not hindered by constraints for supporting life and are therefore highly adaptable to a broad range of production and testing schemes. The advantages of cell-free platforms are being leveraged more commonly by the biotechnology community, and cell-free applications are expected to grow exponentially in the next decade. In this study, new and emerging applications of cell-free platforms, with a specific focus on cell-free protein synthesis (CFPS), will be examined. The current and near-future role of CFPS within metabolic engineering, prototyping, and biomanufacturing will be investigated as well as how the integration of machine learning is beneficial to these applications.

DeepReac+: deep active learning for quantitative modeling of organic chemical reactions

Various computational methods have been developed for quantitative modeling of organic chemical reactions; however, the lack of universality as well as the requirement of large amounts of experimental data limit their broad applications. Here, we present DeepReac+, an efficient and universal computational framework for prediction of chemical reaction outcomes and identification of optimal reaction conditions based on deep active learning. Under this framework, DeepReac is designed as a graph-neural-network-based model, which directly takes 2D molecular structures as inputs and automatically adapts to different prediction tasks. In addition, carefully-designed active learning strategies are incorporated to substantially reduce the number of necessary experiments for model training. We demonstrate the universality and high efficiency of DeepReac+ by achieving the state-of-the-art results with a minimum of labeled data on three diverse chemical reaction datasets in several scenarios. Collectively, DeepReac+ has great potential and utility in the development of AI-aided chemical synthesis. DeepReac+ is freely accessible at https://github.com/bm2-lab/DeepReac.

Tuning Cell-Free Composition Controls the Time Delay, Dynamics, and Productivity of TX-TL Expression

The composition of cell-free expression systems (TX-TL) is adjusted by adding macromolecular crowding agents and salts. However, the effects of these cosolutes on the dynamics of individual gene expression processes have not been quantified. Here, we carry out kinetic mRNA and protein level measurements on libraries of genetic constructs using the common cosolutes PEG-8000, Ficoll-400, and magnesium glutamate. By combining these measurements with biophysical modeling, we show that cosolutes have differing effects on transcription initiation, translation initiation, and translation elongation rates with trade-offs between time delays, expression tunability, and maximum expression productivity. We also confirm that biophysical models can predict translation initiation rates in TX-TL using Escherichia coli lysate. We discuss how cosolute composition can be tuned to maximize performance across different cell-free applications, including biosensing, diagnostics, and biomanufacturing.

Evaluation of an E. coli Cell Extract Prepared by Lysozyme-Assisted Sonication via Gene Expression, Phage Assembly and Proteomics

Over the past decades, starting from crude cell extracts, a variety of successful preparation protocols and optimized reaction conditions have been established for the production of cell-free gene expression systems. One of the crucial steps during the preparation of cell extract-based expression systems is the cell lysis procedure itself, which largely determines the quality of the active components of the extract. Here we evaluate the utility of an E. coli cell extract, which was prepared using a combination of lysozyme incubation and a gentle sonication step. As quality measure, we demonstrate the cell-free expression of YFP at concentrations up to 0.6 mg/mL. In addition, we produced and assembled T7 bacteriophages up to a titer of 108  PFU/mL. State-of-the-art quantitative proteomics was used to compare the produced extracts with each other and with a commercial extract. The differences in protein composition were surprisingly small between lysozyme-assisted sonication (LAS) extracts, but we observed an increase in the release of DNA-binding proteins for increasing numbers of sonication cycles. Proteins taking part in carbohydrate metabolism, glycolysis, amino acid and nucleotide related pathways were found to be more abundant in the LAS extract, while proteins related to RNA modification and processing, DNA modification and replication, transcription regulation, initiation, termination and the TCA cycle were found enriched in the commercial extract.

Biofoundry-assisted expression and characterization of plant proteins

Many goals in synthetic biology, including the elucidation and refactoring of biosynthetic pathways and the engineering of regulatory circuits and networks, require knowledge of protein function. In plants, the prevalence of large gene families means it can be particularly challenging to link specific functions to individual proteins. However, protein characterization has remained a technical bottleneck, often requiring significant effort to optimize expression and purification protocols. To leverage the ability of biofoundries to accelerate design-built-test-learn cycles, we present a workflow for automated DNA assembly and cell-free expression of plant proteins that accelerates optimization and enables rapid screening of enzyme activity. First, we developed a phytobrick-compatible Golden Gate DNA assembly toolbox containing plasmid acceptors for cell-free expression using Escherichia coli or wheat germ lysates as well as a set of N- and C-terminal tag parts for detection, purification and improved expression/folding. We next optimized automated assembly of miniaturized cell-free reactions using an acoustic liquid handling platform and then compared tag configurations to identify those that increase expression. We additionally developed a luciferase-based system for rapid quantification that requires a minimal 11-amino acid tag and demonstrate facile removal of tags following synthesis. Finally, we show that several functional assays can be performed with cell-free protein synthesis reactions without the need for protein purification. Together, the combination of automated assembly of DNA parts and cell-free expression reactions should significantly increase the throughput of experiments to test and understand plant protein function and enable the direct reuse of DNA parts in downstream plant engineering workflows.

The all-E. coliTXTL toolbox 3.0: new capabilities of a cell-free synthetic biology platform

The new generation of cell-free gene expression systems enables the prototyping and engineering of biological systems in vitro over a remarkable scope of applications and physical scales. As the utilization of DNA-directed in vitro protein synthesis expands in scope, developing more powerful cell-free transcription-translation (TXTL) platforms remains a major goal to either execute larger DNA programs or improve cell-free biomanufacturing capabilities. In this work, we report the capabilities of the all-E. coli TXTL toolbox 3.0, a multipurpose cell-free expression system specifically developed for synthetic biology. In non-fed batch-mode reactions, the synthesis of the fluorescent reporter protein eGFP (enhanced green fluorescent protein) reaches 4 mg/ml. In synthetic cells, consisting of liposomes loaded with a TXTL reaction, eGFP is produced at concentrations of >8 mg/ml when the chemical building blocks feeding the reaction diffuse through membrane channels to facilitate exchanges with the outer solution. The bacteriophage T7, encoded by a genome of 40 kb and ∼60 genes, is produced at a concentration of 1013 PFU/ml (plaque forming unit/ml). This TXTL system extends the current cell-free expression capabilities by offering unique strength and properties, for testing regulatory elements and circuits, biomanufacturing biologics or building synthetic cells.

In silico, in vitro, and in vivo machine learning in synthetic biology and metabolic engineering

Among the main learning methods reviewed in this study and used in synthetic biology and metabolic engineering are supervised learning, reinforcement and active learning, and in vitro or in vivo learning. In the context of biosynthesis, supervised machine learning is being exploited to predict biological sequence activities, predict structures and engineer sequences, and optimize culture conditions. Active and reinforcement learning methods use training sets acquired through an iterative process generally involving experimental measurements. They are applied to design, engineer, and optimize metabolic pathways and bioprocesses. The nascent but promising developments with in vitro and in vivo learning comprise molecular circuits performing simple tasks such as pattern recognition and classification.

Analysis of the Innovation Trend in Cell-Free Synthetic Biology

Cell-free synthetic biology is a maturing field that aims to assemble biomolecular reactions outside cells for compelling applications in drug discovery, metabolic engineering, biomanufacturing, diagnostics, and education. Cell-free systems have several key features. They circumvent mechanisms that have evolved to facilitate species survival, bypass limitations on molecular transport across the cell wall, enable high-yielding and rapid synthesis of proteins without creating recombinant cells, and provide high tolerance towards toxic substrates or products. Here, we analyze ~750 published patents and ~2000 peer-reviewed manuscripts in the field of cell-free systems. Three hallmarks emerged. First, we found that both patent filings and manuscript publications per year are significantly increasing (five-fold and 1.5-fold over the last decade, respectively). Second, we observed that the innovation landscape has changed. Patent applications were dominated by Japan in the early 2000s before shifting to China and the USA in recent years. Finally, we discovered an increasing prevalence of biotechnology companies using cell-free systems. Our analysis has broad implications on the future development of cell-free synthetic biology for commercial and industrial applications.

Impact of Porous Matrices and Concentration by Lyophilization on Cell-Free Expression

Cell-free expression systems have drawn increasing attention as a tool to achieve complex biological functions outside of the cell. Several applications of the technology involve the delivery of functionality to challenging environments, such as field-forward diagnostics or point-of-need manufacturing of pharmaceuticals. To achieve these goals, cell-free reaction components are preserved using encapsulation or lyophilization methods, both of which often involve an embedding of components in porous matrices like paper or hydrogels. Previous work has shown a range of impacts of porous materials on cell-free expression reactions. Here, we explored a panel of 32 paperlike materials and 5 hydrogel materials for the impact on reaction performance. The screen included a tolerance to lyophilization for reaction systems based on both cell lysates and purified expression components. For paperlike materials, we found that (1) materials based on synthetic polymers were mostly incompatible with cell-free expression, (2) lysate-based reactions were largely insensitive to the matrix for cellulosic and microfiber materials, and (3) purified systems had an improved performance when lyophilized in cellulosic but not microfiber matrices. The impact of hydrogel materials ranged from completely inhibitory to a slight enhancement. The exploration of modulating the rehydration volume of lyophilized reactions yielded reaction speed increases using an enzymatic colorimetric reporter of up to twofold with an optimal ratio of 2:1 lyophilized reaction to rehydration volume for the lysate system and 1.5:1 for the purified system. The effect was independent of the matrices assessed. Testing with a fluorescent nonenzymatic reporter and no matrix showed similar improvements in both yields and reaction speeds for the lysate system and yields but not reaction speeds for the purified system. We finally used these observations to show an improved performance of two sensors that span reaction types, matrix, and reporters. In total, these results should enhance efforts to develop field-forward applications of cell-free expression systems.

Improving cell-free glycoprotein synthesis by characterizing and enriching native membrane vesicles

Cell-free gene expression (CFE) systems from crude cellular extracts have attracted much attention for biomanufacturing and synthetic biology. However, activating membrane-dependent functionality of cell-derived vesicles in bacterial CFE systems has been limited. Here, we address this limitation by characterizing native membrane vesicles in Escherichia coli-based CFE extracts and describing methods to enrich vesicles with heterologous, membrane-bound machinery. As a model, we focus on bacterial glycoengineering. We first use multiple, orthogonal techniques to characterize vesicles and show how extract processing methods can be used to increase concentrations of membrane vesicles in CFE systems. Then, we show that extracts enriched in vesicle number also display enhanced concentrations of heterologous membrane protein cargo. Finally, we apply our methods to enrich membrane-bound oligosaccharyltransferases and lipid-linked oligosaccharides for improving cell-free N-linked and O-linked glycoprotein synthesis. We anticipate that these methods will facilitate on-demand glycoprotein production and enable new CFE systems with membrane-associated activities.

Easy Synthesis of Complex Biomolecular Assemblies: Wheat Germ Cell-Free Protein Expression in Structural Biology

Cell-free protein synthesis (CFPS) systems are gaining more importance as universal tools for basic research, applied sciences, and product development with new technologies emerging for their application. Huge progress was made in the field of synthetic biology using CFPS to develop new proteins for technical applications and therapy. Out of the available CFPS systems, wheat germ cell-free protein synthesis (WG-CFPS) merges the highest yields with the use of a eukaryotic ribosome, making it an excellent approach for the synthesis of complex eukaryotic proteins including, for example, protein complexes and membrane proteins. Separating the translation reaction from other cellular processes, CFPS offers a flexible means to adapt translation reactions to protein needs. There is a large demand for such potent, easy-to-use, rapid protein expression systems, which are optimally serving protein requirements to drive biochemical and structural biology research. We summarize here a general workflow for a wheat germ system providing examples from the literature, as well as applications used for our own studies in structural biology. With this review, we want to highlight the tremendous potential of the rapidly evolving and highly versatile CFPS systems, making them more widely used as common tools to recombinantly prepare particularly challenging recombinant eukaryotic proteins.

Optimising protein synthesis in cell‐free systems, a review

Over the last decades, cell-free systems have been extensively used for in vitro protein expression. A vast range of protocols and cellular sources varying from prokaryotes and eukaryotes are now available for cell-free technology. However, exploiting the maximum capacity of cell free systems is not achieved by using traditional protocols. Here, what are the strategies and choices one can apply to optimise cell-free protein synthesis have been reviewed. These strategies provide robust and informative improvements regarding transcription, translation and protein folding which can later be used for the establishment of individual best cell-free reactions per lysate batch.

A Streptomyces venezuelae Cell-Free Toolkit for Synthetic Biology

Prokaryotic cell-free coupled transcription–translation (TX-TL) systems are emerging as a powerful tool to examine natural product biosynthetic pathways in a test tube. The key advantages of this approach are the reduced experimental time scales and controlled reaction conditions. To realize this potential, it is essential to develop specialized cell-free systems in organisms enriched for biosynthetic gene clusters. This requires strong protein production and well-characterized synthetic biology tools. The Streptomyces genus is a major source of natural products. To study enzymes and pathways from Streptomyces, we originally developed a homologous Streptomyces cell-free system to provide a native protein folding environment, a high G+C (%) tRNA pool, and an active background metabolism. However, our initial yields were low (36 μg/mL) and showed a high level of batch-to-batch variation. Here, we present an updated high-yield and robust Streptomyces TX-TL protocol, reaching up to yields of 266 μg/mL of expressed recombinant protein. To complement this, we rapidly characterize a range of DNA parts with different reporters, express high G+C (%) biosynthetic genes, and demonstrate an initial proof of concept for combined transcription, translation, and biosynthesis of Streptomyces metabolic pathways in a single “one-pot” reaction.

Flapjack: Data Management and Analysis for Genetic Circuit Characterization

Characterization is fundamental to the design, build, test, learn (DBTL) cycle for engineering synthetic genetic circuits. Components must be described in such a way as to account for their behavior in a range of contexts. Measurements and associated metadata, including part composition, constitute the test phase of the DBTL cycle. These data may consist of measurements of thousands of circuits, measured in hundreds of conditions, in multiple assays potentially performed in different laboratories and using different techniques. In order to inform the learn phase this large volume of data must be filtered, collated, and analyzed. Characterization consists of using this data to parametrize models of component function in different contexts, and combining them to predict behaviors of novel circuits. Tools to store, organize, share, and analyze large volumes of measurement and metadata are therefore essential to linking the test phase to the build and learn phases, closing the loop of the DBTL cycle. Here we present such a system, implemented as a web app with a backend data registry and analysis engine. An interactive frontend provides powerful querying, plotting, and analysis tools, and we provide a REST API and Python package for full integration with external build and learn software. All measurements are associated with circuit part composition via SBOL (Synthetic Biology Open Language). We demonstrate our tool by characterizing a range of genetic components and circuits according to composition and context.

Advanced machine-learning techniques in drug discovery

The popularity of machine learning (ML) across drug discovery continues to grow, yielding impressive results. As their use increases, so do their limitations become apparent. Such limitations include their need for big data, sparsity in data, and their lack of interpretability. It has also become apparent that the techniques are not truly autonomous, requiring retraining even post deployment. In this review, we detail the use of advanced techniques to circumvent these challenges, with examples drawn from drug discovery and allied disciplines. In addition, we present emerging techniques and their potential role in drug discovery. The techniques presented herein are anticipated to expand the applicability of ML in drug discovery.

Robust Cell-Free Expression of Sub-Pathological and Pathological Huntingtin Exon-1 for NMR Studies. General Approaches for the Isotopic Labeling of Low-Complexity Proteins

The high-resolution structural study of huntingtin exon-1 (HttEx1) has long been hampered by its intrinsic properties. In addition to being prone to aggregate, HttEx1 contains low-complexity regions (LCRs) and is intrinsically disordered, ruling out several standard structural biology approaches. Here, we use a cell-free (CF) protein expression system to robustly and rapidly synthesize (sub-) pathological HttEx1. The open nature of the CF reaction allows the application of different isotopic labeling schemes, making HttEx1 amenable for nuclear magnetic resonance studies. While uniform and selective labeling facilitate the sequential assignment of HttEx1, combining CF expression with nonsense suppression allows the site-specific incorporation of a single labeled residue, making possible the detailed investigation of the LCRs. To optimize CF suppression yields, we analyze the expression and suppression kinetics, revealing that high concentrations of loaded suppressor tRNA have a negative impact on the final reaction yield. The optimized CF protein expression and suppression system is very versatile and well suited to produce challenging proteins with LCRs in order to enable the characterization of their structure and dynamics.

Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism

Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.

Artificial cells drive neural differentiation

We report the construction of artificial cells that chemically communicate with mammalian cells under physiological conditions. The artificial cells respond to the presence of a small molecule in the environment by synthesizing and releasing a potent protein signal, brain-derived neurotrophic factor. Genetically controlled artificial cells communicate with engineered human embryonic kidney cells and murine neural stem cells. The data suggest that artificial cells are a versatile chassis for the in situ synthesis and on-demand release of chemical signals that elicit desired phenotypic changes of eukaryotic cells, including neuronal differentiation. In the future, artificial cells could be engineered to go beyond the capabilities of typical smart drug delivery vehicles by synthesizing and delivering specific therapeutic molecules tailored to distinct physiological conditions.

Improving the reaction mix of a Pichia pastoris cell-free system using a design of experiments approach to minimise experimental effort

A renaissance in cell-free protein synthesis (CFPS) is underway, enabled by the acceleration and adoption of synthetic biology methods. CFPS has emerged as a powerful platform technology for synthetic gene network design, biosensing and on-demand biomanufacturing. Whilst primarily of bacterial origin, cell-free extracts derived from a variety of host organisms have been explored, aiming to capitalise on cellular diversity and the advantageous properties associated with those organisms. However, cell-free extracts produced from eukaryotes are often overlooked due to their relatively low yields, despite the potential for improved protein folding and posttranslational modifications. Here we describe further development of a Pichia pastoris cell-free platform, a widely used expression host in both academia and the biopharmaceutical industry. Using a minimised Design of Experiments (DOE) approach, we were able to increase the productivity of the system by improving the composition of the complex reaction mixture. This was achieved in a minimal number of experimental runs, within the constraints of the design and without the need for liquid-handling robots. In doing so, we were able to estimate the main effects impacting productivity in the system and increased the protein synthesis of firefly luciferase and the biopharmaceutical HSA by 4.8-fold and 3.5-fold, respectively. This study highlights the P. pastoris-based cell-free system as a highly productive eukaryotic platform and displays the value of minimised DOE designs.