Self-adaptive biosystems through tunable genetic parts and circuits

Transfer learning for cross-context prediction of protein expression from 5'UTR sequence

Model-guided DNA sequence design can accelerate the reprogramming of living cells. It allows us to engineer more complex biological systems by removing the need to physically assemble and test each potential design. While mechanistic models of gene expression have seen some success in supporting this goal, data-centric, deep learning-based approaches often provide more accurate predictions. This accuracy, however, comes at a cost - a lack of generalization across genetic and experimental contexts that has limited their wider use outside the context in which they were trained. Here, we address this issue by demonstrating how a simple transfer learning procedure can effectively tune a pre-trained deep learning model to predict protein translation rate from 5' untranslated region (5'UTR) sequence for diverse contexts in Escherichia coli using a small number of new measurements. This allows for important model features learnt from expensive massively parallel reporter assays to be easily transferred to new settings. By releasing our trained deep learning model and complementary calibration procedure, this study acts as a starting point for continually refined model-based sequence design that builds on previous knowledge and future experimental efforts.

Inducible plasmid copy number control for synthetic biology in commonly used E. coli strains

The ability to externally control gene expression has been paradigm shifting for all areas of biological research, especially for synthetic biology. Such control typically occurs at the transcriptional and translational level, while technologies enabling control at the DNA copy level are limited by either (i) relying on a handful of plasmids with fixed and arbitrary copy numbers; or (ii) require multiple plasmids for replication control; or (iii) are restricted to specialized strains. To overcome these limitations, we present TULIP (TUnable Ligand Inducible Plasmid): a self-contained plasmid with inducible copy number control, designed for portability across various Escherichia coli strains commonly used for cloning, protein expression, and metabolic engineering. Using TULIP, we demonstrate through multiple application examples that flexible plasmid copy number control accelerates the design and optimization of gene circuits, enables efficient probing of metabolic burden, and facilitates the prototyping and recycling of modules in different genetic contexts.

Engineering genetic circuits: advancements in genetic design automation tools and standards for synthetic biology

Synthetic biology (SynBio) is a field at the intersection of biology and engineering. Inspired by engineering principles, researchers use defined parts to build functionally defined biological circuits. Genetic design automation (GDA) allows scientists to design, model, and analyze their genetic circuits in silico before building them in the lab, saving time, and resources in the process. Establishing SynBio's future is dependent on GDA, since the computational approach opens the field to a broad, interdisciplinary community. However, challenges with part libraries, standards, and software tools are currently stalling progress in the field. This review first covers recent advancements in GDA, followed by an assessment of the challenges ahead, and a proposed automated genetic design workflow for the future.

Multi-layer CRISPRa/i circuits for dynamic genetic programs in cell-free and bacterial systems

CRISPR-Cas transcriptional circuits hold great promise as platforms for engineering metabolic networks and information processing circuits. Historically, prokaryotic CRISPR control systems have been limited to CRISPRi. Creating approaches to integrate CRISPRa for transcriptional activation with existing CRISPRi-based systems would greatly expand CRISPR circuit design space. Here, we develop design principles for engineering prokaryotic CRISPRa/i genetic circuits with network topologies specified by guide RNAs. We demonstrate that multi-layer CRISPRa/i cascades and feedforward loops can operate through the regulated expression of guide RNAs in cell-free expression systems and E. coli. We show that CRISPRa/i circuits can program complex functions by designing type 1 incoherent feedforward loops acting as fold-change detectors and tunable pulse-generators. By investigating how component characteristics relate to network properties such as depth, width, and speed, this work establishes a framework for building scalable CRISPRa/i circuits as regulatory programs in cell-free expression systems and bacterial hosts. A record of this paper's transparent peer review process is included in the supplemental information.

Precise Prediction of Promoter Strength Based on a De Novo Synthetic Promoter Library Coupled with Machine Learning

Promoters are one of the most critical regulatory elements controlling metabolic pathways. However, the fast and accurate prediction of promoter strength remains challenging, leading to time- and labor-consuming promoter construction and characterization processes. This dilemma is caused by the lack of a big promoter library that has gradient strengths, broad dynamic ranges, and clear sequence profiles that can be used to train an artificial intelligence model of promoter strength prediction. To overcome this challenge, we constructed and characterized a mutant library of Trc promoters (P_trc) using 83 rounds of mutation-construction-screening-characterization engineering cycles. After excluding invalid mutation sites, we established a synthetic promoter library that consisted of 3665 different variants, displaying an intensity range of more than two orders of magnitude. The strongest variant was ∼69-fold stronger than the original P_trc and 1.52-fold stronger than a 1 mM isopropyl-β-d-thiogalactoside-driven P_T7 promoter, with an ∼454-fold difference between the strongest and weakest expression levels. Using this synthetic promoter library, different machine learning models were built and optimized to explore the relationships between promoter sequences and transcriptional strength. Finally, our XgBoost model exhibited optimal performance, and we utilized this approach to precisely predict the strength of artificially designed promoter sequences (R² = 0.88, mean absolute error = 0.15, and Pearson correlation coefficient = 0.94). Our work provides a powerful platform that enables the predictable tuning of promoters to achieve optimal transcriptional strength.

Massively parallel characterization of engineered transcript isoforms using direct RNA sequencing

Transcriptional terminators signal where transcribing RNA polymerases (RNAPs) should halt and disassociate from DNA. However, because termination is stochastic, two different forms of transcript could be produced: one ending at the terminator and the other reading through. An ability to control the abundance of these transcript isoforms would offer bioengineers a mechanism to regulate multi-gene constructs at the level of transcription. Here, we explore this possibility by repurposing terminators as 'transcriptional valves' that can tune the proportion of RNAP read-through. Using one-pot combinatorial DNA assembly, we iteratively construct 1780 transcriptional valves for T7 RNAP and show how nanopore-based direct RNA sequencing (dRNA-seq) can be used to characterize entire libraries of valves simultaneously at a nucleotide resolution in vitro and unravel genetic design principles to tune and insulate termination. Finally, we engineer valves for multiplexed regulation of CRISPR guide RNAs. This work provides new avenues for controlling transcription and demonstrates the benefits of long-read sequencing for exploring complex sequence-function landscapes.

Bioinspired Networks of Communicating Synthetic Protocells

The bottom-up synthesis of cell-like entities or protocells from inanimate molecules and materials is one of the grand challenges of our time. In the past decade, researchers in the emerging field of bottom-up synthetic biology have developed different protocell models and engineered them to mimic one or more abilities of biological cells, such as information transcription and translation, adhesion, and enzyme-mediated metabolism. Whilst thus far efforts have focused on increasing the biochemical complexity of individual protocells, an emerging challenge in bottom-up synthetic biology is the development of networks of communicating synthetic protocells. The possibility of engineering multi-protocellular systems capable of sending and receiving chemical signals to trigger individual or collective programmed cell-like behaviours or for communicating with living cells and tissues would lead to major scientific breakthroughs with important applications in biotechnology, tissue engineering and regenerative medicine. This mini-review will discuss this new, emerging area of bottom-up synthetic biology and will introduce three types of bioinspired networks of communicating synthetic protocells that have recently emerged.

Attenuator LRR - a regulatory tool for modulating gene expression in Gram-positive bacteria

With the rapid development of synthetic biology in recent years, particular attention has been paid to RNA devices, especially riboswitches, because of their significant and diverse regulatory roles in prokaryotic and eukaryotic cells. Due to the limited performance and context-dependence of riboswitches, only a few of them (such as theophylline, tetracycline and ciprofloxacin riboswitches) have been utilized as regulatory tools in biotechnology. In the present study, we demonstrated that a ribosome-dependent ribo-regulator, LRR, discovered in our previous work, exhibits an attractive regulatory performance. Specifically, it offers a 60-fold change in expression in the presence of retapamulin and a low level of leaky expression of about 1-2% without antibiotics. Moreover, LRR can be combined with different promoters and performs well in Bacillus thuringiensis, B. cereus, B. amyloliquefaciens, and B. subtilis. Additionally, LRR also functions in the Gram-negative bacterium Escherichia coli. Furthermore, we demonstrate its ability to control melanin metabolism in B. thuringiensis BMB171. Our results show that LRR can be applied to regulate gene expression, construct genetic circuits and tune metabolic pathways, and has great potential for many applications in synthetic biology.

Rapid prototyping and design of cybergenetic single-cell controllers

The design and implementation of synthetic circuits that operate robustly in the cellular context is fundamental for the advancement of synthetic biology. However, their practical implementation presents challenges due to low predictability of synthetic circuit design and time-intensive troubleshooting. Here, we present the Cyberloop, a testing framework to accelerate the design process and implementation of biomolecular controllers. Cellular fluorescence measurements are sent in real-time to a computer simulating candidate stochastic controllers, which in turn compute the control inputs and feed them back to the controlled cells via light stimulation. Applying this framework to yeast cells engineered with optogenetic tools, we examine and characterize different biomolecular controllers, test the impact of non-ideal circuit behaviors such as dilution on their operation, and qualitatively demonstrate improvements in controller function with certain network modifications. From this analysis, we derive conditions for desirable biomolecular controller performance, thereby avoiding pitfalls during its biological implementation.

Are There Biomimetic Lessons from Genetic Regulatory Networks for Developing a Lunar Industrial Ecology?

We examine the prospect for employing a bio-inspired architecture for a lunar industrial ecology based on genetic regulatory networks. The lunar industrial ecology resembles a metabolic system in that it comprises multiple chemical processes interlinked through waste recycling. Initially, we examine lessons from factory organisation which have evolved into a bio-inspired concept, the reconfigurable holonic architecture. We then examine genetic regulatory networks and their application in the biological cell cycle. There are numerous subtleties that would be challenging to implement in a lunar industrial ecology but much of the essence of biological circuitry (as implemented in synthetic biology, for example) is captured by traditional electrical engineering design with emphasis on feedforward and feedback loops to implement robustness.

Cheetah: A Computational Toolkit for Cybergenetic Control

Advances in microscopy, microfluidics, and optogenetics enable single-cell monitoring and environmental regulation and offer the means to control cellular phenotypes. The development of such systems is challenging and often results in bespoke setups that hinder reproducibility. To address this, we introduce Cheetah, a flexible computational toolkit that simplifies the integration of real-time microscopy analysis with algorithms for cellular control. Central to the platform is an image segmentation system based on the versatile U-Net convolutional neural network. This is supplemented with functionality to robustly count, characterize, and control cells over time. We demonstrate Cheetah's core capabilities by analyzing long-term bacterial and mammalian cell growth and by dynamically controlling protein expression in mammalian cells. In all cases, Cheetah's segmentation accuracy exceeds that of a commonly used thresholding-based method, allowing for more accurate control signals to be generated. Availability of this easy-to-use platform will make control engineering techniques more accessible and offer new ways to probe and manipulate living cells.