Integrating microfluidics and synthetic biology: advancements and diverse applications across organisms

Abstract

Synthetic biology is the design and modification of biological systems for specific functions, integrating several disciplines like engineering, genetics, and computer science. The field of synthetic biology is to understand biological processes within host organisms through the manipulation and regulation of their genetic pathways and the addition of biocontrol circuits to enhance their production capabilities. This pursuit serves to address global challenges spanning diverse domains that are difficult to tackle through conventional routes of production. Despite its impact, achieving precise, dynamic, and high-throughput manipulation of biological processes is still challenging. Microfluidics offers a solution to those challenges, enabling controlled fluid handling at the microscale, offering lower reagent consumption, faster analysis of biochemical reactions, automation, and high throughput screening. In this review, we diverge from conventional focus on automating the synthetic biology design-build-test-learn cycle, and instead, focus on microfluidic platforms and their role in advancing synthetic biology through its integration with host organisms – bacterial cells, yeast, fungi, animal cells – and cell-free systems. The review illustrates how microfluidic devices have been instrumental in understanding biological systems by showcasing microfluidics as an essential tool to create synthetic genetic circuits, pathways, and organisms within controlled environments. In conclusion, we show how microfluidics expedite synthetic biology applications across diverse domains including but not limited to personalized medicine, bioenergy, and agriculture.

---

1. Introduction

Synthetic biology is an interdisciplinary field of science and engineering that focuses on designing, constructing, and modifying biological systems or components to achieve specific functions, behaviors, or applications. It draws inspiration from principles of engineering, genetics, molecular biology, and computer science to manipulate genetic material, such as DNA, reorganize biological parts to create new organisms or modify existing ones. The goals of synthetic biology are to gain a deeper understanding of biological processes in a host organism and to engineer biological systems with enhanced or novel capabilities. This includes producing valuable chemicals and materials to developing innovative medical treatments and technologies and therefore has the potential to address a wide range of critical global challenges.^1,2

Although the impact of synthetic biology is far-reaching in many fields and industries, a continuous challenge with synthetic biology is mimicking a precise, dynamic, and high-throughput manipulation of biological processes. To address this challenge, microfluidics is becoming an essential tool since it provides the precise manipulation and fluidic control at the microscale, enabling a controlled environment and high throughput analysis of thousands and millions of samples. Furthermore, microfluidics enables precise compartmentalization within tiny structures like droplets and traps. With such structures in a device, the device operation is automated to simulate bench operations and can seamlessly be integrated with established technologies like microscopy. This integration facilitates observation, analysis, and screening of genetic constructs and engineered cells.

Therefore, in this review, rather than delving into specific areas like automating the design-build-test-learn cycle (which we³ and others⁴ have reviewed), microfluidic advancements in synthesizing droplets, vesicles, and artificial cells,⁵ or integrating microfluidics to develop a physiological organ biomimetic system for applications related to physiology and drug discovery.⁶ We describe the use of different microfluidic paradigms (e.g., droplet, channels, digital microfluidics, and paper-based) as the main tool to push the field of synthetic biology. In addition, we illustrate the use of microfluidic devices in synthetic biology, highlighting their existing applications and potential integration and advancement within the field. As shown in Fig. 1, this review categorizes the discussion into the integration of diverse microfluidic technologies: digital microfluidics (DMF),^7,8 channel microfluidics,⁹ droplet microfluidics,^10–12 and paper microfluidics.^13,14 We demonstrate how these technologies are applied to different organisms used in synthetic biology applications: bacterial cells, yeast, filamentous fungi, animal cells, and cell-free systems. For example, we detail the use of microfluidics as an essential tool for the creation of synthetic genetic circuits, pathways, and organisms, while providing a controlled environment for these synthetic constructs to operate. Finally, the review will encompass studies pertaining to cell-free systems and examine the ways in which microfluidics have advanced the field of synthetic biology applications, especially in relation to diagnostics.

Fig. 1 Microfluidic toolbox for synthetic biology applications. Four types of technologies are shown in the microfluidic toolbox: digital microfluidics (liquid droplets are manipulated using an electric field generated between the electrodes and the ground plate), channel-based microfluidics (controlled flow within confined micron-channels to culture cells or perform biochemical reactions), droplet microfluidics (pico- or nano-liter droplets are generated by two immiscible phases, aqueous and oil, and each droplet act as an individual bioreactor), and paper microfluidics (nitrocellulose paper to transport or to store fluid). Such devices are used to facilitate different experimental paradigms with different host organisms in synthetic biology, such as mammalian cells, bacterial cells, yeast cells, and filamentous fungi or using cell-free systems.

2. Applications of integrated synthetic biology and microfluidics in bacteria

The advancement in microfluidic technology (mother machine, channel-based microfluidics, digital microfluidics, and droplet-in-channel microfluidics) plays a pivotal role in the ability to study bacterial cells in spatiotemporal resolution with precise control and in high-throughput, gaining a deeper understanding of cellular phenotyping, genetic circuits and responses to stimuli and to environmental changes. Each of the different microfluidic paradigms have unique advantages that enhance the design and the control of synthetic biological systems as shown in Table 1, where we provide a summary of research advancements discussed in this section.

Table 1 Comparative analysis of microfluidic platforms for advancing synthetic biology in bacterial cells: applications, advantages, and limitations

Microfluidic platform	Organism name	Synbio applications	Advantages	Disadvantages	Ref. number
Mother machine	E. coli	Growth and division monitoring, gene expression dynamics, genetic circuits	Faster and more reliable analysis for single-cell studies – real time microscopy	Unable to detect secreted metabolites, redesign for varied bacterial strains, challenge analysis of mobile cells	15–17
Channel-based	Mycobacterial cells	Dynamic drug response screening	Real-time evaluation and heterogeneous analysis	Low throughput (5 chambers), not applicable to study single cell level	18
Channel-based	E. coli	Transcriptional dynamics, population-level phenotypes	Continuous growth monitoring, environmental control, and real-time analysis	Limited accessibility and integration of the technology, difficult setup, and operation	19, 20
Digital microfluidics	E. coli	DNA assembly, transformation, whole genome sequencing, enzyme screening	Automation, small volumes, parallelization of experiments, and high transformation efficiency	Low throughput and incomplete integration of the synbio ‘build’ pipeline	21–25
Droplet microfluidics	E. coli, Bacillus licheniformis	Gene expression optimization, enzyme variant screening, protein expression	High throughput screening (>200 Hz), genetic stability, enzyme variant identification	Use of lysis solution prior to screening – cell death, DNA recovery from droplets is challenging, and screening based on labeled markers (e.g., fluorescence or absorbance)	26, 27
Droplet microfluidics	E. coli, Bacillus licheniformis	Protein engineering and enzyme activity	High throughput screening (>200 Hz), protein activity based sorting, and secreted proteins (no lysis solution added)	Screening based on fluorescence or absorbance markers, and more complex circuits needed for protein secretion, lower concentration of protein outside the cell than inside – lower detection sensitivity	28, 29

---

2.1. Analyzing bacterial dynamics: channel-based microfluidic platforms for investigating single cell bacteria

Synthetic biology's unique ability to engineer cellular functions stems from genetic circuit design, creating functions that would not exist naturally.^30–32 These genetic circuits, including well-known examples like the toggle switch,³³ oscillators,³⁴ feedback loops,³⁵ and Boolean logic gates,³⁶ play a crucial role in engineering desired functions by manipulating and editing the cell's DNA to encode specific proteins or re-wiring metabolic pathways. This manipulation allows for dynamic control, enabling precise regulation to generate valuable products such as pharmaceuticals and food additives. However, regulatory connections within genetic circuits may sometimes disrupt cellular functions and hinder adaptation to environmental conditions.^30–32 Thus, achieving a quantitative understanding of these dynamics and functions are critical, yet challenging, since it involves the integration of predictive mathematical models with quantitative experiments in a high throughput manner. In this context, advances in microfluidic technologies like the mother machine and channel microfluidics provide the ideal platforms to study genetic circuits and bacterial dynamics with exceptional detail.

The mother machine is a microfluidic device created to monitor the growth and division of E. coli cells in a high throughput and temporal manner.³⁷ The general ‘mother machine’ device is designed with micron-sized channels (trenches) to trap bacterial cells that are oriented perpendicular to a continuous growth medium feeding channel which removes the progeny cells from the trenches. Using such a platform with microscopy allows for the temporal observation of single cells that can be captured in thousands of channels, which generates insightful data on the millions of individual cells rather than the averaged population data. The mother machine provides fast and reliable analysis of high-throughput single-cell bacteria, and their biological processes through the integration of an automated, time-lapse, image-based microscopy.⁵ Using such tools, we can understand their growth kinetics, gene expression dynamics, and heterogeneity in bacterial populations. This microfluidic platform has massively contributed to the recent advances in genetic circuits and understanding bacterial dynamics.^15,38 Zhang et al.¹⁵ have reported a genetic circuit resembling closely to the repressilator, with controlled amplitude and period. They use mVenus as the reporter gene regulated by a dual-input promoter and uses salicylate for inducing changes in amplitude and IPTG for period. Using such a circuit with the mother machine enabled precise oscillation control such that these cells could potentially be used in pulsatile drug delivery systems and other time-controlled stimuli. Another example is Choudhary et al.,¹⁶ who have characterized their mother machine device (Fig. 2A), to monitor a diverse phenotypic (oxidative stress) response of genetically identical bacterial cells when exposed to hydrogen peroxide (H₂O₂). Their novel approach has shown that phenotypic heterogeneity results from precise and rapid feedback between individual cells and their direct environment. The short-range cell–cell interaction reveals a collective defense mechanism within the bacterial population, opposing the traditional bet hedging explanation. This work not only contributes to our understanding of identical bacterial cell behavior but offers potential applications in studying the response of synthetic heterogeneous cell population under oxidative stress.

Fig. 2 Microfluidics and synthetic biology for bacterial cells. (A) Workflow to study heterogeneity in response via cell–cell interactions of E. coli at the single cell level due to oxidative stress using a mother-machine device integrated with microscopy imaging and machine learning (adapted from ref. 16 with permission from Elsevier) (Choudhary et al.¹⁶). (B) Schematic diagram of the microfluidic platform to study biofilm formation in 2D format showing the side view of the microfluidic chip, seeding zone before and during bacteria loading, top view of the seeding zone when loading and during biofilm growth and biofilm formation (adapted from ref. 39 with permission from John Wiley and Sons) (Zhang et al.³⁹). (C) Design of a digital microfluidic device for one-pot DNA assembly and bacterial transformation. As shown, a device design with an modular electrode pattern and with different reaction regions. A side view of the device layers containing an ITO top-plate connected to an impedance circuit for liquid volume detection and an aluminum block housing a negative temperature coefficient thermistor connected to a thermoelectric cooling module for temperature control (adapted from ref. 21 with permission from Royal Society of Chemistry) (Perry et al.²¹). (D) Workflow of functional metagenomic screening to identify β-glucuronidase enzymes. Step 1: droplet screening of a metagenomic plasmid population transformed into E. coli (single bacterial cells were encapsulated into pL droplets with lysis solution and fluorescent substrate followed by incubation and high throughput screening and sorting based on β-glucuronidase); step 2: The sorted genes were then recovered into single bacterial colonies and screened in an agarose-substrate solution and the positive colonies were screened repeatedly; step 3: analysis of enzyme activity in the well plate format followed by DNA sequencing (adapted from ref. 26 with permission from Springer Nature) (Neun et al.²⁶).

Historically, bacterial dynamics have traditionally been studied using continuous deterministic mathematical models such that representation of a biological system assumes all variables and parameters are precisely defined and predictable.^40,41 There is growing recognition that biochemical kinetics at the single-cell level display stochastic behavior, characterized by random variability or unpredictability, commonly observed in biological systems. This defines the need for stochastic models to accurately capture heterogeneity. However, these models are associated with challenges; they are computationally more intensive than deterministic models and significantly harder to calibrate with experimental data. In that context, the mother machine device has been used as a powerful tool to study the dynamics and stochasticity of biological processes. Yang et al.¹⁷ studied the stochastic nature of cell death in genetically identical bacterial cells during carbon-starvation of E. coli. They observed individual cells using a modified mother machine device and measured cell membrane damage over time. The study revealed that the variation in cell life cycles is not explained by initial conditions but by a stochastic mechanism. The mechanism suggests that as cells starve, they produce damaging units at a constant rate, and their removal saturates with higher starvation, leading to amplified noise and different death times in younger cells. Interestingly, as the cells age, the variation among them diminishes, with deterministic processes progressively influencing lifespan dynamics. Understanding how those models work is crucial for advancing synthetic biology, which can enable better prediction and control of engineered systems at the single cell level.

Aside from monitoring bacterial dynamics in response to a stimulus, the mother machine can also be used to study environmental conditions and how dynamic feedback within the environment impacts microbe growth and community. Daniels et al.,⁴² have adapted a mother machine device coupling it to a community batch culture to study the interaction of two different species at the single cell level, Vibrio natriegens (chitin degrader) and Alteromonas macleodii (consumer of degrader metabolic by-products). Through their approach, they were able to study the dynamics of their interaction and have observed a shift from mutualism (early stages show consumer promoting degrading growth) to competition (later stages show competition for metabolic by-products such as acetate). Their platform exemplifies a microfluidic device that holds promise for applications in synthetic cells and their interactions with pathogens, particularly for therapeutic purposes.

Beyond using mother-machine devices, channel-based microfluidics, the original conception of microfluidics, has been a tool to study bacterial dynamics for decades. Recently, there are studies using such a device to examine the dynamic drug response of antibiotics bacterial cells. Mistretta et al.,¹⁸ have used channel-based microfluidics for drug dosage screening of mycobacterial cells, they have designed a microfluidic system with independent microculture chambers (five different culture conditions) coupled with a dilution tree, where each microchamber used a hydro-pneumatic mechanism to trap and to culture bacterial cells as monolayers. The authors showed, with their platform, a novel understanding on how antibiotics (e.g., moxifloxacin) affect mycobacterial growth in a dose- and time-dependent manner, with clonal cells displaying heterogeneous responses at the same drug concentration. In addition, they have shown that in concentrations close to the minimum inhibitory concentration, the drug's target is upregulated which potentially can trigger adaptive mechanisms in a fraction of the population that could make the bacteria resistant to the antibiotic. Using such a device, there is potential to expand on this platform for real-time evaluation of key genes involved in regulating antibiotic bacterial resistance, applicable not only to monoclonal populations but also to heterogeneous populations.

The channel format can also be used to study the effect of spatial organization on microbial communities on gene expression and growth, Gupta et al.⁴³ have designed a microfluidic platform, based on a previously reported device.⁴⁴ Their design has two separated monolayer culturing chambers connected through 25, 50, 100, and 250 μm interaction channels. Controlling the distance between strains (different diffusion rates of molecules), biological responses such as growth rate, metabolic activity, and cell state decisions, can be mapped quantitatively revealing the degree of sensitivity (linear regime) or decoupling (saturated regime) of the strains interactions, which cannot be studied in real-time using traditional culturing methods. Another example is Zhang et al.,³⁹ who have designed a microfluidic device to study bacterial biofilm, their design as controllable bacterial seeding that allows semi-2D structure culturing (Fig. 2B). By coupling their device to time-lapse microscopy, they were able to study the dynamics of biofilm formation. Wang et al.,⁴⁵ further expanded the semi-2D structure to study spatial transcriptomics of E. coli biofilm, and in addition, developed a new method of growth-based encoding strategy to label biofilms in a spatially specific manner. By combining machine-learning-guided fluorescence-activated cell sorting (FACS) and RNA-seq with ultra-low input, they were able to profile the spatial transcriptome of E. coli biofilm. They have discovered that the nutrient-limited biofilm's interior was not dormant; instead, many genes were specifically activated in that region, sustained by community-level metabolic coordination. This finding was only possible due to the integration of a microfluidic device and time-lapse microscopy to machine learning, cell sorting and RNAseq.

Beyond investigating bacterial organization, such devices are also used for screening genetic circuit dynamics. Graham et al.¹⁹ have developed a channel-based microfluidic platform with colony traps called “Dynomics” platform. The platform integrates multiplexed microfluidics, fluorescence microscopy, deep neural networks, and explainable artificial intelligence algorithms. Their device was designed to enhance the understanding of transcriptional dynamics on a genome-wide scale. The Dynomics platform allows for continuous growth, precise environmental control, and optical monitoring of 2176 distinct GFP-reporter E. coli microcolonies for up to 14 days. The authors demonstrated the Dynomics platform's real-time biosensor capabilities by predicting heavy metal presence in water samples based on dynamic transcription profiles from 1807 E. coli promoters. Lezia et al.²⁰ expanded the platform to screening genetic circuits that can coordinate cellular behaviour using cell–cell interactions in large populations. This is of great importance for synthetic biology as in many areas such as “living” therapeutics where cell–cell communication has been used to engineer bacterial cells to control mechanisms of reducing inflammation at the population level. Their work shows how the use of a multiplexed microfluidic platform enhances the capacity to develop an arrayed-screening workflow for dynamic single cell circuits that can operate at the population level. They validated their approach by using a mutant library of the pre-existing synchronized oscillator circuit that is capable of functioning over long timescales.

2.2. Digital microfluidics for bacterial synthetic biology and biomedical applications

Engineering genetic constructs is one of the most important parts of synthetic biology, serving as the basis for engineering new biological functions.^46,47 These constructs are meticulously designed to encode specific genetic elements, such as proteins, regulatory sequences, or metabolic pathways, allowing researchers to introduce desired traits into living organisms. Built with modularity and standardization, constructs are easily assembled and manipulated, akin to building blocks in engineering. Systematic design, construction, and testing of constructs yield insights into fundamental biological principles, driving iterative optimization for practical applications. Nevertheless, the current pipeline for designing, constructing, and inserting constructs into microbial hosts predominantly involves time-consuming manual labor, or the use of centralized biofoundries, which is expensive and is not accessible to all the scientific community.⁴⁸ Microfluidic technologies, particularly digital microfluidics, have emerged as promising alternatives owing to their capacity to offer decentralized systems to automate the entire process of DNA construction.

Digital microfluidics (DMF) manipulates pL–μL liquid droplets on an electrically controlled surface.^7,8 The mechanism of droplet movement leverages electric fields, surface tension, and fluidic polarity through electrode-driven potential, allowing for precise and individual control for applications requiring liquid-handling tasks. DMF offers numerous advantages for bacterial synthetic biology, including automation and manipulation of small volumes of droplets and parallelization of experiments without any pumps or valves at the sacrifice of throughput. In recent years, DMF platforms have been used to automate construct assembly and transformation processes⁴⁹ and other synthetic biology processes.⁵⁰ For example, Perry et al.²¹ have developed a partial, miniaturized biofoundry, using a rapid-prototype digital microfluidic platform for “one-pot” Golden Gate DNA assembly and E. coli transformation (Fig. 2C). The authors fabricated their device using rapid prototype techniques to address challenges in cleanroom fabrication. Their device also includes a novel ‘slash’ electrode geometry and modular design for reproducible fabrication and usage and for easy programming to enable standardized methods for building constructs – an important process in synthetic biology. A key feature with their platform is they have integrated an impedance-based adaptive closed-loop water replenishment (using a 3D printed syringe pump) to maintain sample volumes and concentrations, enabling 25 thermocycling steps on-chip, which remarkably has yielded comparable transformation efficiency to those obtained in the bench scale experiments. In addition, a closed-loop temperature control system was included that further increased the on-chip heat shock transformation efficiency. For a synthetic biology application, the authors validated their automated system with a large and complex plasmid (six DNA fragments assembled into a 14 kb plasmid – the violacein biosynthesis pathway) and showed that their system has better or at least comparable performance to standard techniques.

Aside from DNA assembly and transformation, digital microfluidics (DMF) has been employed for automating the amplification and quantification of DNA. However, its capabilities extend further, offering potential integration into the broader biofoundry pipeline outlined by Perry et al.²¹ This integration could include various stages, such as DNA extraction, amplification, quantification, assembly, transformation, and culturing, thereby enhancing the efficiency and automation of complex biological workflows. Liu et al.,²² have introduced a digital microfluidic device for rapid whole genome sequencing of low-abundance bacterial DNA by automating the process of genome amplification. They have achieved 100× higher sensitivity than conventional in-tube amplification, enabling detection as little as 10 fg of DNA within 2 hours and identification of the target bacterium within 30 minutes. Hu et al.,²³ added the extraction (with amplification) and real-time product detection all on the same platform. Their platform uses magnetic beads, nucleic acid extraction, LAMP amplification, and real-time product detection by fluorescence. By comparing on-chip and off-chip DNA extraction, they were able to show equally efficient on- and off-chip extraction in an automated and controllable process that could be integrated with other construction platforms (like Perry et al.²¹) to fully integrate the process of engineering of microbial organisms (plasmid construction, amplification, extraction, and sequencing).

Cell-related applications, especially culturing bacterial cells as miniaturized bioreactors, have been an attractive target application on DMF,^51,52 and potentially be an elegant method for optimizing fermentation conditions prior to scale-up.^53–58 Culturing cells on a DMF platform enables integration with the construction pipeline by offering a method to induce gene expression precisely at the optimal time point⁵⁹ as well as a method to optimize culture and media fermentation conditions prior to scale-up given the control capabilities DMF offers over culture conditions.⁶⁰ In addition, DMF enables scale-down modeling down to the single cell⁶¹ with real-time monitoring⁶² providing insights into cellular behavior, facilitating dynamic control of single cell conditions, and efficient screening and optimization of a library of engineered strains. Generally, detecting a protein or metabolite product requires colorimetric or fluorescent substrate.⁵⁹ Instead, Qiu et al.²⁴ have developed DMF platform that integrates an optical oxygen sensor film to indirectly measure the cell's metabolic activity (high oxygen, low metabolic activity = low fluorescence; low oxygen, high metabolic activity = high fluorescence) of E. coli. The oxygen-sensitive probe film was made of platinum(II)-5,10,15,20-tetrakis-(2,3,4,5,6-pentafluorophenyl)-porphyrin (PtTFPP) and embedded on the DMF, results showed no negative effects on droplet manipulation or cell growth. E. coli culture droplets on the chip exhibited stable, bright, and homogeneous fluorescence emission after 6 hours of on-chip culturing, confirming cell viability. As proof-of-concept, they have tested the metabolic state of E. coli in different culture and antibiotic conditions and their results demonstrated similar results compared to standard 96-well plate methods. Their platform offers rapid and reliable online metabolic testing with minimal sample handling and smaller volumes compared to standard methods. In another method, Sklavounos et al.²⁵ developed a DMF platform for the analysis of bacterial growth and phenotype detection using growth metabolic markers. The custom DMF combines a heating module for bacterial cultivation and a machine-learning-enabled low-cost color camera for real-time monitoring of absorbance and fluorescence. The examples provided by Qiu and Sklavounos represent promising systems for optimizing culture conditions for metabolite production before scaling up.

2.3. Droplet-in-channel microfluidics: advancing bacterial synthetic biology through high-throughput

High-throughput screening in pico- to nanoliter of water-in-oil droplets has made droplet microfluidics a preferred platform for bacterial synthetic biology research. Droplets are serially generated, transported, and sorted in microfluidics platforms with high-performance throughput. These operations enable droplet-in-channel microfluidics the ability to culture cells in individual droplets, which act as microbioreactors to optimize gene expression levels for metabolite production. Saleski et al.⁶³ presented a novel approach to optimize gene expression levels by integrating a genetic pathway for isobutanol into E. coli's chromosome using a previous designed⁶⁴ droplet microfluidics with sorting throughput of 200 Hz. They have generated a mutant library of random chromosomal integration using Tn5 transposase, co-encapsulated the library as single cell with an auxotrophic mNeonGreen expressing K12 E. coli sensor strain. Their screening results found that their chromosomal integration approach resulted in higher isobutanol titers (10.0 ± 0.9 g L⁻¹) compared to transforming plasmids. This method of genetic engineering provided enhanced genetic stability in the mutants, which reduced variability in isobutanol production. The capability of efficiently assessing large libraries makes it possible to simultaneously re-engineer genes through multiple pathways in a multiplexed manner and conduct screening to achieve a desired phenotype.

Bacteria is an excellent host for alternative production of enzymes, which are key components in various biological applications; like the synthesis of pharmaceutical drugs and the breakdown of starches into fermentable sugars for the production of biorenewables as well as many foods and beverage-based products.^65–70 Detecting these variants efficiently through high-throughput methods will facilitate expedited and thorough investigations of enzyme functionality. Traditionally, accessing such enzymes entails a cell lysis procedure to measure the desired protein's activity. Wong et al.²⁷ have designed a lysis-on-demand (LoD) genetic system that controls cell lysis rate in E. coli based on the enterobacteria phage T4 holin–endolysin system, which is controllable by changing the inducer concentration. They used their LoD system in 7 pL droplets to assay phosphotriesterase (PTE) produced in the cytoplasm of E. coli, and they have shown titratable intracellular enzyme release upon lysis induction while preserving viable cells for genetic material retrieval. Neun et al.²⁶ introduced a lysis buffer into their droplets such that a functional metagenomics approach can be used to discover novel catalytic properties of enzymes. In their pipeline (Fig. 2D), a ∼million-membered metagenomic library (transformed E. coli) were encapsulated in 4 pL droplets with a fluorogenic substrate and lysis solution, and sorted at 1 kHz throughput based on β-glucuronidase activity, resulting in the identification of 1432 hits from a ∼7 million droplet population, reducing the library size by ∼5000 fold. After droplet sorting, the plasmids were recovered from the lysate, re-transformed, and screened for enzyme activity as single colonies on agar plates (134 hits) and well plates (28 hits) resulted in only two unique and genuine hits. The hits were also different from known enzymes and unpredicted by bioinformatics. Additionally, one of the hits exhibited promiscuous catalytic activity, highlighting its versatile function.

Although chemical lysis is typically used, there are preferences to observe protein expression in real-time and to test protein activity without destroying the cell walls. Karamitros et al.,²⁸ studied protein expression at different locations (cytoplasmic, surface display, and periplasmic) in the bacterial cells with L-asparaginase (L-ASNase) enzyme in E. coli by designing three different expression vectors that allowed the bacteria to produce the enzymes in different locations in the cell. They encapsulated single cells into 60 pL droplets and observed the enzyme reaction and expression for 12 minutes at the different locations. On their platform, they identified that periplasmic expression for L-ASNase is the optimal strategy, offering both high yield (fluorescence signal) and substrate accessibility without cell lysis. Some have moved from the typical E. coli bacterial host for protein expression to industrial relevant hosts for production. Yuan et al.²⁹ used a mutant library of Bacillus licheniformis strains to be screened based on secreted α-amylase yield. They have encapsulated single cells with fluorescent substrate in 8 pL droplets and incubated for 8 hours, followed by sorting in 300 Hz throughput. Using their approach, the authors were able to couple the enzyme activity with a fluorescence substrate to obtain Bacillus licheniformis strains with higher α-amylase productivity – over 50% improvement – compared to the wild type. Another example is Adolfsen, et al.,⁷¹ who have developed an improved treatment for phenylketonuria (PKU) by employing a biosensor-based high-throughput screening workflow to measure the enzymatic activity of the mutant library of phenylalanine ammonia lyase (PAL). Their aim was enhancing the PAL enzyme degradation of phenylalanine to the non-toxic product trans-cinnamate (TCA). Their biosensor was designed such that in the absence of TCA, an allosteric transcription factor (aTF) binds to the DNA operator site, suppressing GFP expression, however, when PAL converts phenylalanine into TCA, it binds to the aTF enabling GFP expression. Using Poisson distribution for cell encapsulation in droplets, researchers successfully screened ∼one million mutants and have identified a lead PAL variant with five mutations, resulting in approximately two-fold higher in vivo PAL activity compared to the original strain. The optimized PAL variant was used to create a second-generation PKU therapeutic strain with the potential for a more effective and lower dosing treatment, as demonstrated in preclinical testing, warranting further evaluation in clinical trials. These platforms and methods offer intriguing avenues for exploring protein activity and production with and without destroying the cell.

3. Applications of integrated synthetic biology and microfluidics in yeast

Yeast is an ideal organism to produce alternative proteins like antibodies and enzymes given their eukaryotic nature. Their ability to perform more complex post translational modifications and protein folding compared to bacteria, makes them a more desirable organism for these synthetic biology applications. Furthermore, yeast is also highly efficient in converting sugars into different chemicals given their robust protein expression system compared to bacteria. Integrating synthetic biology with microfluidics offers powerful tools for studying and manipulating complex cellular processes to produce valuable products in yeast. The precision capabilities of microfluidic devices researchers can engineer metabolic pathways in yeast to produce valuable chemicals and alternative proteins like antibodies and enzymes and study single cells to further understand complex yeast biology. The monitoring of single-cell behavior on a microfluidic device allows for an in-depth look at gene expression dynamics, and cellular responses to various environmental stimuli. Droplet microfluidics allows for high throughput screening of large libraries of microbes, which accelerates the discovery and optimization of novel enzymes for enhanced biofuels and recombinant antibody production. This integration opens exciting possibilities for synthetic biology relating to both biotechnological problems as well as health related applications as shown in Table 2.

Table 2 Comparative analysis of microfluidic platforms for advancing synthetic biology in yeast cells: applications, advantages, and limitations

Microfluidic platform	Organism name	Synbio applications	Advantages	Disadvantages	Ref. number
Channel microfluidics (including mother-machine)	Saccharomyces cerevisiae (budding yeast)	Real-time observation of cellular responses to stimuli, morphological changes, cell division, and gene expression	Cost-effective analysis of signaling pathways related to lifespan and aging-induced diseases	Specific applicability to yeast cells, unable to capture cellular behavior of more complex organisms	72–75
Droplet microfluidics	Saccharomyces cerevisiae, Yarrowia lipolytica	High throughput screening in yeast libraries, protein directed evolution applications for enzyme and antibody production	Precise control over individual cells, integration with various biosensors, analysis of thousands of samples – (>300 Hz)	Limitation in detection methods (primarily fluorescence and absorbance), may not capture all aspects of enzyme activity or protein binding	76–78
Droplet microfluidics	Saccharomyces cerevisiae	Mass spectrometry for detection of enzymatic activity and metabolites production	High selectivity, label-free detection, and single-cell level	Lower throughput (>1 Hz), limited to specific analytes, may require specific expertise for operation and interpretation	79
Droplet microfluidics	Saccharomyces cerevisiae	Detection of metabolites production using biosensors	High selectivity, detection of metabolites through fluorescence, single-cell level, and high throughput (>300 Hz)	Biosensor design is challenging, limitation in constructing a circuit for a broader range of metabolites	80, 81

---

3.1. Deeper insights into yeast biology: microfluidic monitoring of single-cell behavior and gene expression dynamics

Yeast has emerged as a favored microorganism host in synthetic biology for producing recombinant secreted proteins by integrating genetic networks into their genome. This preference arises from their similarity to higher eukaryotic organisms in protein secretion and in folding pathways.⁸² Additionally, yeasts offer other advantages: possess a wide array of genetic tools, can easily grow in cost-effective media, are generally considered biosafe across species, and are genetically tractable.⁸³ For these reasons, yeast is an ideal host for new breakthroughs in science, such as humanized yeast. Over the last two decades, humanized yeast has gained popularity as a powerful tool for engineering genetic human networks within yeast cells. This approach enables the study of human diseases, drug metabolism, and various aspects of human biology in a simplified and highly controllable experimental setting.^84,85 However, understanding the numerous interactions within human genetic networks in yeast is still incomplete, potentially compromising the robustness of the platform. Therefore, developing tools to help understanding yeast biology in early-stage research is crucial for advancing these fields.

In that context, microfluidics technologies have given unprecedented insights into yeast biology through the combination of channel microfluidics and time-lapse microscopy, enabling a controlled environment with real-time observation of cellular responses to stimuli in single cells. Single cell monitoring is particularly valuable in the context of humanized yeast, mainly observed in aging research, because controlled monitoring of morphological changes, cell division, and gene expression under various environmental stimuli provide a deeper understanding of the cell's behavior over time. Saccharomyces cerevisiae (budding yeast) serves as an ideal organism for these studies due to its rapid growth, short division time, and cost-effectiveness compared to mammalian cells, making it feasible to explore signaling pathways related to lifespan and aging-induced diseases like neurodegeneration and cancer.

The mother machine, commonly used for bacterial cells, has been re-designed particularly for Saccharomyces cerevisiae.^73,86,87 Mother cells are securely trapped in individual trenches to be precisely monitored while they continuously divide. This has led to the development of open-ended trenches where budding daughter cells can be flushed out of the trenches in both directions. Through use of the mother machine device, Li et al.,⁷² investigate the aging process of individual yeast cells by measuring their lifespan, dynamically tracking changes in gene expression, chromatin state, and organelle morphology. They found that isogenic wild-type cells exhibited two distinct phenotypic changes during aging: “mode 1” with elongated daughter cells and nucleolar changes, and “mode 2” with small round daughter cells and mitochondrial decline. Mode 1 cells had enlarged and fragmented nucleoli, while mode 2 cells exhibited aggregated mitochondria before cell death, indicating organellar dysfunction in each mode. These findings provide valuable insights into the distinct organellar failures associated with different aging modes, shedding light on the complexity of cellular aging, which could only be discovered with the mother machine platform.

Redesigning and optimizing microfluidics platforms could significantly advance the study of humanized yeast, extending its applications beyond humanized aging studies. These platforms enable the study of single humanized yeast cells, offering real-time insights into studying human neurodegeneration and cancer, with higher throughput than conventional experiments. One alternative device was designed by Ryley and Pereira-Smith,⁸⁷ where the new device was fabricated with physical traps (μm-sized structures made of PDMS posts) to isolate individual mother cells and remove excess daughter cells. Sarnoski et al.,⁷³ further expanded the device by adding a platform to accommodate diploid yeast cells (bigger cells), along with an optimized elliptical cell-trapping unit (Fig. 3A). Xu, Zhu and Wang et al.,⁷⁴ have used a four trap structure to compare hydrodynamic forces affecting mother cell retention and daughter dissection. They have optimized the microfluidic trapping structure which allows newborn buds to rotate in the downstream direction. Wang et al.,⁷⁵ have increased the number of traps and optimized the platform for long-term culturing for cell-aging analysis and studying replicative lifespan (RLS) of diploid yeast. The device was designed with 1100 traps arranged in an array to hold single yeast cells and remove daughter cells under a laminar-perfused medium. Experimental characterization ensured reliable retention of mother cells, adequate space for diploid cell growth, and continuous culturing and time-lapse monitoring for over 60 hours to determine RLS and budding time interval during yeast aging. These adaptations showcase innovative strategies for isolating mother cells, accommodating different cell types, optimizing trapping structures, and enabling long-term culturing and analysis, ultimately advancing our capabilities in studying humanized yeast.

Fig. 3 Microfluidics and synthetic biology for yeast cells. (A) Microfluidic device designed to measure diploid yeast single-cell and to examine phenotypes related to aging. As shown, the microscopy setup with the device for trapping single cells, an elliptical cell-trapping unit to trap only single mother cells, a graph showing survival rate of diploid cells, and a frequency histogram related to cell death (adapted from ref. 73 with permission from Elsevier) (Sarnoski et al.⁷³). (B) Workflow of screening yeast cells that are antibody producers using droplet microfluidics. Screening was performed by measuring red and green fluorescence based on cells that secreted the desired IgG which binds to the Alexa486-fused antigen which turns the droplet to a green fluorescence. Non desired IgG binds to anti-kappa magnetic beads and to the Alexa647-fused anti-FC (red fluorescence) (adapted from ref. 78 with permission from Royal Society of Chemistry) (Lebrun at al⁷⁸). (C) Screening and sorting workflow of the yeast mutant library that are engineered to produce triacetic acid lactone. Using picoinjection, an E. coli GFP-biosensor is used to detect the desired product (adapted from ref. 80 with permission from National Academy of Sciences of the United States of America) (Bowman et al.⁸⁰).

3.2. Speedy discoveries: high throughput screening in yeast libraries using microfluidic platforms

Droplet microfluidics has accelerated high throughput screening in yeast libraries by offering precise control over individual cells by providing a droplet microenvironment. This enables the analysis of thousands of samples based on microscopy imaging, fluorescence and absorbance detection, and integration with other biosensors. These tools when combined with droplet microfluidics are highly valuable for protein engineering applications, especially enzymes that catalyze specific reactions to control the host's pathway to produce valuable products. Examples are searching for thermo-tolerant enzymes for bioremediation, screening microbes with increased enzyme secretions for faster rates of biofuel production and screening for recombinant antibody production in microbes.

Starting with Su et al.,⁷⁶ they have reported a highly efficient microfluidics system for the directed evolution of the laccase enzyme. Using the pico-injection technique to add a substrate (containing ABTS and Amplex Red) and a heating step, they screened a random mutagenesis library of the thermotolerant laccase enzyme. This enzyme catalyzes the oxidation of a wide range of substrates using molecular oxygen, leading to the reduction of oxygen to water. After only two rounds of directed evolution and combinatorial mutagenesis, they selected a variant through fluorescence. The enhanced variant showed improved thermostability and solvent resistance due to three amino acid substitutions (Asp to Asn) on the surface of the enzyme. Structural analysis through homology modeling and molecular dynamics simulation revealed that the variant has more hydrogen bonds at these positions, enhancing the thermoresistance compared to the wild type. The enzyme's stability in a more extreme environment makes it suitable for bioremediation applications. In another example, Johansson et al.,⁷⁷ have developed a workflow to boost recombinant protein production of the secreted α-amylase, starch degrading enzyme, used for bioethanol production. By combining droplet microfluidic screening with a large-scale CRISPR library to their workflow, they were able to control gene expression (increase or decrease expression) of 345 genes in S. cerevisiae, where the resulted shift in expression led to an increase in enzyme production. Screening was performed by measuring droplet fluorescence and the selected ‘hits’ were genes related to α-amylase production, such as vesicle trafficking, endosome to Golgi transport, phagophore assembly, cell cycle, and energy supply. The cell encapsulation and sorting workflow can further be used for the development of recombinant proteins.

Lebrun et al.,⁷⁸ have developed a robust and engineerable alternative for antibody screening due to the integration of yeast secreting antibodies (e.g., complex molecules like full-length IgGs) and high throughput screening of droplet microfluidics (Fig. 3B). They have presented the screening of secreted human IgGs using picolitre-droplets reactors to encapsulate yeast cells. The yeast Yarrowia lipolytica was chosen as the chassis due to its strong bioproduction and secretion abilities, and they have engineered the strain for antibody secretion, higher strain production, and developed an in-drop high-throughput immunoassay, based on previous reported studies,^88,89 for efficient identification and enrichment of antibody-secreting yeasts. Moreover, their screening approach could be adapted for high-throughput screening of any yeast strain used for antibody research, making it a versatile and universal tool.

Fluorescence and absorbance have still been the most used method of detection of analytes in droplets, however, their versatility is constrained by certain limitations. Fluorescence markers correlate primarily with cell viability and protein activity or binding. While they offer high sensitivity, their applicability is not universal, which means fluorescence substrates are not always available for each type of protein. Absorbance, on the other hand, primarily reflects cell density, specific molecules, and protein activity, but is less sensitive due to its dependence on path length which can be very small in microfluidic platforms. Other techniques that have also garnered interest in recent years are mass spectrometry and biosensors. Mass spectrometry (MS) is a label-free detection technique that could offer high selectivity and application versatility. Wink et al.⁷⁹ developed a coupled microfluidic platform with nESI-FTMS (nano-electrospray ionization-Fourier transform ion cyclotron resonance MS) and cell-droplet imaging, allowing single-cell level quantification of biocatalytic products with high selectivity and correlation between product concentration and cell number. They successfully determined cell-specific product formation rates and quantified biocatalytic heterogeneity in an enantioselective reduction using Saccharomyces cerevisiae.

Biosensors combine biological elements with transducers to detect specific target analytes of interest. They come in various types, including optical, electrochemical, piezoelectric, and thermal biosensors, and when integrated to microfluidics, they are a powerful tool for high throughput screening. Bowman et al.,⁸⁰ have developed a microfluidic pipeline for single cell encapsulation of a producer cell followed by pico-injection of cell-based (S. cerevisae or E. coli) biosensors (Fig. 3C). With this approach, they were able to detect analytes (triacetic acid lactone, naringenin, and L-DOPA) that could not be measured using traditional, enzyme-substrate or protein binding methods. The cell biosensor was constructed by deliver a plasmid to E. coli, which consists of GFPmut2 placed downstream of a AraC promoter along with AraC-TAL transcription factor. This allows fluorescent signal to directly correlate to TAL and L-DOPA production. Using their approach, they were able to screen a chemical mutated cells library and successfully identify a S. cerevisae mutant with up to 3-fold improvement in producing triacetic acid lactone. Li et al.,⁸¹ have used a similar pipeline of droplet screening using cell-biosensor pico-injected to enhance erythritol (low calorie biological sweetener) production in Yarrowia lipolytica. Their biosensor cell was an E. coli strain with a transcription activation of fluorescence protein based on binding affinity of erythritol to the repressor protein. Using their approach, they were able to screen and sort a library of mutated Y. lipolytica and achieved up to 26% of improvement of erythritol production. Hence, the adaptability of droplet microfluidics enables a wide range of detection methods, making it valuable for screening active enzymes and monitoring metabolite production.

4. Applications of integrated synthetic biology and microfluidics in filamentous fungi

Fungal synthetic biology has lagged behind compared to bacteria and yeast, however, they are becoming more attractive hosts for cell factories or biomass products.^90,91 As a result, there is an increasing interest to develop fungal synthetic biology toolboxes to improve production of proteins and bioactive compounds.^92,93 In addition, fungi are becoming popular to study complex cellular processes, gene regulation, and metabolic pathways. Through genetic manipulation, like gene knockouts and insertions, their protein productivity can be enhanced, and be used in a wide range of bio-based products such as enzymes, biofuels, pharmaceuticals, and food ingredients.^93,94 By combining microfluidics with filamentous fungi, scientists can explore intricate cellular processes, optimize metabolic pathways, and (re-)engineer these organisms, and select high producers by microfluidic screening. Two main approaches in microfluidics have been explored for filamentous fungi: microchannels for hyphae and growth observation, and droplet-based microfluidics for screening high producers (Table 3).

Table 3 Comparative analysis of microfluidic platforms for advancing synthetic biology in filamentous fungi: applications, advantages, and limitations

Microfluidic platform	Organism name	Synbio applications	Advantages	Disadvantages	Ref. number
Channel microfluidics	Basidiomycete, ascomycete and zygomycete	Studying fungal navigation, behavior, and strategies	Replication of soil pore space, micro-spatial heterogeneity, insights into foraging strategies	Low throughput, not integrated with analysis of metabolites production and growth	95, 96
Channel microfluidics	Clonostachys rosea and Fusarium graminearum	Investigating interactions between two fungal strains	Live-cell imaging, dynamics of necrotrophic hyperparasitism	Low throughput, needs device design optimization to better mimic the “real” environment (plants)	97
Droplet microfluidics	Aspergillus niger	Screening mutant libraries for increased α-amylase production	Efficient sorting system, discovery of high α-amylase production mutant, and high throughput (>30 Hz)	Low incubation time of filamentous fungi (2 days) – not suitable for screening of post-germination enzymes	98
Droplet microfluidics	Trichoderma reesei	High-throughput workflow for protoplast transformants	Rapid isolation of targeted fungi, faster and more efficient than traditional methods	Platform developed for early protein detection, not suitable for long term incubation	99
Droplet-digital microfluidics	Clonostachys rosea	Screening cell wall degrading enzymes	Solid-state fermentation approach, novel droplet composition, improved enzyme production, throughput (7 Hz)	Colloidal chitin yet not optimized as solid support for other fungi strains. Not suitable for chitinase producers	100

---

4.1. Unraveling fungal hyphal growth and foraging strategies through microfluidics and maze-like devices

The exploration of filamentous fungi, particularly mycelial growth and branching has garnered significant interest across various industries due to their potential applications in sustainable material production and industrial processes. Pure mycelium materials exhibit versatile properties that can potentially replace petrochemical polymers and animal-based leather.¹⁰¹ Some fungal species are known to produce a large amount of mycelium when they are grown on lignocellulosic substrates but is not the case for other cultivation types (e.g., solid state fermentation). To expand large-scale mycelium growth to industrial-relevant scales, strain development and synthetic biology techniques need to be implemented. Biocontrol agent strains like Clonostachya rosea and Trichoderma asperellum often lack adaptation to large-scale fungal growth or agricultural climate conditions, but could be significantly improved through synthetic biology techniques.^102,103 Using microfluidics can help monitor, assess the growth, and study different cultivation conditions of the engineered fungal strains. Channel-based microfluidic devices can be patterned with microchannels for observing hyphal formation, tracking fungal growth in real-time, and investigating interactions among different fungal strains. These microchannels furnish a controlled and precisely defined environment, which can provide new strategies on improving metabolite or protein production in relation to their hyphal growth patterns and morphological changes.

Devices such as the one developed by Aleklett et al.,⁹⁵ called “Soil Chip” system, have been designed to study fungal navigation through maze-like soil structures that could mimic agricultural environments or industrial fermentation conditions (SSF). Their microfluidic system was designed to replicate soil pore space and micro-spatial heterogeneity, to study hyphal growth behavior, and strategies of seven Basidiomycota litter decomposing species. The structures designed in the device made the hyphae encounter micrometer constrictions, sharp turns, and obstacles, leading to distinct responses among the species in terms of foraging range, persistence, and spatial exploration. The use of microstructures within the chip provided unprecedented insights into fungal foraging strategies and behavior through multivariate trait analyses. They have observed distinct trade-offs in hyphal foraging, including growth speed versus branching. Notably, species exhibited different growth distances depending on the spatial structures, with some showing much further growth in certain structures. In a similar study by Hopke et al.,⁹⁶ a microchannel “maze-like” device was designed. The authors introduced a “Fungus Olympics” competition using a microfluidic device featuring four maze patterns to assess the growth velocity and branching frequency of fourteen filamentous fungi, including ascomycete, basidiomycete, and zygomycete species. Their findings showed that growth velocities varied from 1 to 4 μm min⁻¹ in straight channels, but the time taken to complete mazes did not correlate directly with linear growth velocity. Instead, fungi exhibited two distinct strategies for navigating mazes: high-frequency branching, exploring all possible paths, and low-frequency branching, exploring only a few paths. The selection of hyphae navigation relied on the maze complexity, with high-frequency branching observed in mazes featuring sharp turns, while low-frequency branching proved more efficient in mazes with shallower turns. These cultivation techniques hold potential for translation into industrial scale applications.

Using microchannels, culturing different fungi in a controlled environment can allow for fungal interactions to be analyzed. Gimeno et al.,⁹⁷ have developed a microfluidic platform to directly examine interactions between two strains of filamentous fungi (Fig. 4A). The device was fabricated with two opposite inlets for fungal inoculum and the growth behavior and interaction of the two strains were observed in microchannels by direct confrontation of the two fungal species, enabling live-cell imaging, and revealing the dynamics of necrotrophic hyper parasitism. Using this technology, the authors investigated Clonostachys rosea (biological control agent – potential biopesticide) against a GFP-tagged Fusarium graminearum (plant pathogen) uncovering dynamic interactions at the single-cell level. According to their observations, loss of fluorescence in parasitized hyphae containing GFP-tagged F. graminearum was correlated to the detection of GFP in the mycelium of C. rosea. GFP signals within the C. rosea hyphal network suggested not only the disruption of cell wall of F. graminearum, but the uptake of GFP protein could be based on cytoplasm transport from the prey to the predator.

Fig. 4 Microfluidics and synthetic biology for filamentous fungi. (A) Overview and design of a microfluidic device for measuring fungal-to-fungal interaction. As shown, the design of the PDMS device containing two inoculum inlets and six chamber channels (two for the interactions and four controls), a device setup – PDMS slab bonded to a glass-bottomed Petri dish, and a schematic of fungal interaction zone (adapted from ref. 97 with permission from Springer Nature) (Gimeno et al.⁹⁷). (B) A universal high-throughput system for optimizing protein production in filamentous fungi using core–shell droplets (reproduced from ref. 98 with permission from American Chemical Society) (Zhang et al.⁹⁸). (C) Screening workflow of filamentous fungi using solid-state droplet fermentation. As shown, the design of solid-state fermentation droplet generator and the design of a low voltage sorter, and the workflow to screen and sort filamentous fungi cultivated in colloidal chitin media followed by off-chip solid incubation and fluorescent-activated droplet sorting. (adapted from ref. 100 with permission from Springer Nature) (Samlali et al.¹⁰⁰).

4.2. Microfluidic approaches for high-throughput screening and sorting of filamentous fungi mutants and transformants

The ability to encapsulate single fungal spores in droplets has made droplet microfluidics a useful tool for screening mutant libraries of many kinds.^104,105 However, compared to yeast and bacteria, the growth morphology of filamentous fungi (hyphae growing) and the incubation time required for protein screening makes the screening of those microbes very challenging. The primary challenge in screening these organisms arises from the mycelia branches, which can pierce through the droplets and clog microfluidic channels. Zhang et al.⁹⁸ used droplet microfluidics to screen for increased α-amylase production in Aspergillus niger using fluorescence detection (Fig. 4B). GFP 11 (the 11th strand in GFP) was fused to the C-terminus of α-amylase to create a new strain of A. niger called AG11. A mutant library of the new strain was created through ARTP mutagenesis. The single spores of the mutants were encapsulated in core-shell droplets consisting of Gel-MA along with GFP 1–10 acting as a substrate. Once the α-amylase was produced from the mutants, the GFP-11 fused to it would re-assemble into the full GFP protein when secreted into the droplet with the substrate. This way, a fluorescent signal would increase with increased production of α-amylase. Using this efficient core-shell droplet sorting system, researchers were able to discover a mutant that produced more than double the wildtype production of α-amylase.

Similarly, Luu et al.⁹⁹ developed a droplet-based microfluidic system for high-throughput workflow of T. reesei protoplast transformants. Their workflow consisted of generating 1 nL droplets of single spores, followed by 24 hours of incubation, droplet screening and sorting and evaluation of transformants. Within the droplets, they were able to detect early expression of the fluorescent marker GFP during fungal growth before extensive mycelial branching and sort a GFP-expressing fungi from a large spore library, demonstrating the potential to rapidly isolate targeted fungi. Additionally, the regeneration of T. reesei protoplasts within droplets was much faster and more efficient than traditional plate-based methods.

Several approaches have been reported to address the issue of mycelium branching while analyzing enzyme activity. Samlali et al.,¹⁰⁰ developed a high throughput droplet screening of C. rosea based on secretion of cell wall degrading enzymes (β-glucanases, and β-nacetylgalactosaminidases) (Fig. 4C). Their proteins are known to be expressed in the later stages of spore germination, and those enzymes are to be screened after a longer incubation time, at least 96 hours. Thus, a novel droplet composition was developed to keep the hyphae contained (with minimal piercing) in the droplet. The authors have developed a solid-state in droplet fermentation approach by co-encapsulating single spores with colloidal chitin (which served as a solid support for the branching mycelia). They have also developed a low-voltage sorter using co-planar electrodes to sort the polydispersed droplets – normally seen after a long incubation period. Using their workflow, they were able to screen and sort a library of UV-mutated C. rosea, and achieved up to 100-fold improvement in the cell wall degrading enzyme production.

5. Applications of integrated synthetic biology and microfluidics in animal cells

Microfluidics has emerged as a transformative field offering novel insights into the intricate workings of mammalian cells. Due to their relevance to human biology, mammalian cells play important roles in disease modeling, antibody discoveries, and drug testing, all without the use of human patients. By leveraging precisely controlled microscale fluid flows, microfluidic platforms enable researchers to create tailored environments that mimic the in vivo conditions for cells. These systems provide a unique advantage in studying cellular behavior, interactions, and responses to various stimuli with unprecedented spatiotemporal resolution. Microfluidic devices also facilitate single-cell analyses, allowing the dissection of heterogeneity within cell populations, unveiling rare subpopulations, and deciphering dynamic cellular processes. Furthermore, these platforms offer the ability to manipulate and co-culture different cell types, fostering a deeper understanding of complex cellular interactions and tissue-level dynamics as summarized in Table 4. In the realm of mammalian cell biology, microfluidics has opened new avenues for studying fundamental processes, disease mechanisms, and potential therapeutic interventions, ushering in a new era of precision and depth in cellular exploration in biomedical research.

Table 4 Comparative analysis of microfluidic platforms for advancing synthetic biology in animal cell models: applications, advantages, and limitations

Microfluidic platform	Cell lineage	Synbio applications	Advantages	Disadvantages	Ref. number
Organ-on-a-chip	iPSC, derived dopaminergic neurons, primary human brain astrocytes, microglia, pericytes, endothelial cells	Parkinson's disease modeling	Accurate disease model displaying key cellular components of Parkinson's disease, organoid can be cultured on chip, controlled tissue environment, controlled tissue environment, flow in channel access to the model	Model specifically represents the substantia nigra region of the brain, requires five mammalian cell types, complex microfluidic chip requiring two channel layers and five separated compartments	106
Tumor-on-a-chip	Primary pancreatic cancer cells, pancreatic stellate cells, U937 cells	Pancreatic cancer modeling	Creation of patient-specific tumor model, controlled tissue environment, drug testing/delivery to the tumor model through flow in channel access	Requires patient samples as well as two additional cell types, patient cell samples needs to be cultured off chip for organoid formation	107
Organ-on-a-chip	Human pluripotent stem cells	Polycystic kidney disease modeling	Accurate disease model displaying key physical components of PKD, controlled tissue environment, drug testing/delivery to the tumor model through flow in channel access	Organoid formation needs to be done off chip, chip is limited to 6 cultured chambers on device	108
Digital microfluidics	Human CD4+T cells, HEK-293 cells, MCF-7 cells	Mammalian cell genetic engineering	Biological payload delivered to the cells through lentiviral transduction and electroporation on chip, stable cell line development, automated operations	Primary cells require off chip culturing prior to electroporation, electrodes on the microfluidic chip limit how many simultaneous reactions can be completed	109, 110
Droplet microfluidics	Human hybridoma cells, primary mouse B-cells	Platform for antibody screening and selection	High-throughput screening, successful sorting of mammalian cells based on antibody production (anti-ACE-1 and IgG)	Sorting of droplets is limited to fluorescence through antibody interaction with a fluorescent substrate or binding to an antigen	89
Droplet and digital microfluidics	HAP1 cells, NCI-H1299 cells, CD4 T cells	Platform for genetically engineered single cell isolation and gene patterns for latent HIV	Single cell isolation from a large genetically heterogenous population, cells per their genetic material can be recovered or cultured following isolation	Limited throughput for large-scale studies^111,112 Challenging to successfully design beads for the specific disease genetic marker – lack of database^113	111–113

---

5.1. Advancements in organ-on-a-chip technology: modeling human disease and personalized therapeutics

Disease modeling with synthetic biology holds promise for advancing our understanding of human health and disease. Using a model that displays key characteristics of a given disease provides a platform for modifying human genes related to disease progression. Synthetic biologists can use the ubiquitous CRISPR tools to knock out, knock in, knockdown important genes to better understand their involvement in the development of the disease as well as discover genetic predispositions. They can test new bio-circuits or therapeutics to evaluate the efficiency of such treatments and to understand how these new drugs could maintain or improve cellular function. Models often used to test such circuits or drugs are two-dimensional cell cultures and animal models.¹¹⁴ The challenges are the technical complexity in using animal models and the ethical concerns with modifying animals.^115,116 However, microfluidics provides an excellent platform that can simulate the structure and function of human organs, aligning with efforts to find alternatives to genetically engineering animals. An example is to culture organoids on such microfluidic platforms. They serve as an ideal platform for investigating the progression of human diseases due to their ability to offer a more accurate representation of human biology, in contrast to two-dimensional cell cultures. Additionally, they present the advantage of personalized patient models, a feat that cannot be achieved by animal models.¹¹⁶ An organoid is a miniature structure composed of ex vivo cells that mimics the structure and function of a real organ.¹¹⁷ Growing in a three-dimensional configuration, the cells can form a complex system to act as an organ like structure simulating the human body.¹¹⁶ Culturing patient primary cells on this platform creates a more accurate representation of each patients unique phenotypic behavior to various drug treatments and genetic modifications. Previously, genetically engineered mice have been used to study different genotypes of various diseases like pancreatic cancer,¹¹⁸ polycystic kidney disease (PKD),¹¹⁹ and Parkinson's disease,¹²⁰ to better understand the role of different genes in disease development. The combination of tissue engineering and precision of a microfluidic chip, organ-on-a-chip platforms are able to accomplish real time monitoring of disease models, more precise tissue environments through co-cultures of various cell types as well as controllable culture conditions.¹¹⁶ Microfluidics combines genetic engineering and three-dimensional organoids through the organ-on-a-chip technology, providing the precision and control of the microenvironment to better study each individual's disease.¹²¹

Current advances in organ-on-a-chip technology have emulated biologically accurate human organ systems. Beginning with the human brain-chip model, which was able to mimic key cellular mechanisms of Parkinson's disease, a neurodegenerative disease that affects the brain.¹²¹ Pediaditakis et al.,¹⁰⁶ have created a flow-in channel microfluidic device made from PDMS with two channel layers and a 50 μm PDMS membrane separation layer containing pores of 7 μm (Fig. 5A). In the top channel layer five types of cells were cultured: iPSC (induced pluripotent stem cells) derived dopaminergic neurons, primary human brain astrocytes, microglia, and pericytes, and in the bottom layer, endothelial cells were seeded and cultured. These dual layers within the three-dimensional cell culture were designed to emulate the substantia nigra region of the brain, primarily implicated in Parkinson's disease. In this condition, the accumulation of substantial quantities of α-synuclein leads to the deterioration of this region at the onset of the disease. When flowing α-synuclein fibrils through the channel, cultured cells displayed characteristics of α-synuclein toxicity which included: oxidative stress, abnormal mitochondrial functions, astrogliosis microglia activation, and neural loss with time. These are all characteristics found in the substantia nigra region in patients diagnosed with the disease. Using an organ-on-a-chip platform, the authors were able to successfully display key characteristics of Parkinson's disease making a new platform for studying the disease.

Fig. 5 Microfluidics and synthetic biology for mammalian cells. (A) Schematic design of a brain-on-a-chip microfluidic device mimicking the substantia nigra. The device was designed with two channels: The lower vascular channel – to culture iPSC-derived brain endothelial cells, and upper ‘brain’ channel – to culture iPSC-derived dopaminergic neurons, primary human brain astrocytes, microglia, and pericytes. A device side view is shown to describe the vascular channel and the brain channel (adapted from ref. 106 with permission from Springer Nature) (Pediaditakis et al.¹⁰⁶). (B) Schematic of a multi-chamber microfluidic device for seeding and growing patient-derived-organoids. The chambers are separated by a 0.4 μm porous membrane to allow perfusion of the cell culture medium through the lower chamber to maintain cell viability (reproduced from ref. 107 with permission from Springer Nature) (Haque et al.¹⁰⁷). (C) Workflow of the digital microfluidic triDrop system – a platform using a three-droplet arrangement for electroporation of mammalian cells (adapted from ref. 110 with permission from John Wiley and Sons) (Little and Leung et al.¹¹⁰). (D) Integrated digital microfluidic platform designed for automated viral generation, packaging, and transduction for standardizing viral-based genome engineering methods (adapted from ref. 109 with permission from American Chemical Society) (Quach et al.¹⁰⁹). (E) Device operation of a droplet-digital microfluidic device for on-demand single-cell encapsulation and analysis to be used to successfully sort and to recover single isoclones to establish monoclonal genetic engineered cell lines (adapted from ref. 104 with permission from John Wiley and Sons) (Samlali et al.¹¹²).

Organ-on-chip microfluidic devices can be used to create patient specific tumor models for numerous diseases. There is a need for developing more accurate in vitro platforms for tumors because it is essential to better represent human biology and the heterogeneity of cancer in many individuals. Haque et al.,¹⁰⁷ have developed an organ-on-chip device to study pancreatic ductal adenocarcinoma, pancreatic cancer (Fig. 5B). Their workflow consisted of 3D (Matrigel dome) culturing patient tumor sample biopsies to form tumor organoids, followed by culturing the organoids in a double layered PDMS microfluidics chip with a porous membrane in between. To mimic realistic tumor microenvironment, the primary cancer cell organoids were co-cultured with stromal cells (pancreatic stellate cells) and U937 monocytes in the top layer. The stromal cells allow for the formation of desmoplastic stroma which is a dense fibrous tissue that surrounds and infiltrates certain tumors, affecting their growth and behavior, and is usually caused from an increase of extracellular proteins produced by non-cancerous tissues as a reaction to the tumor. The increase of extracellular proteins promotes tumor growth and increases the possibility of resistance to chemotherapy treatment in the patient. Using such a platform, the authors tested the drug response of Gemcitabine and ATRA (all trans retinoic acid) by flowing through the system for 72 hours. When compared to the non-treated organoids, the treatment caused an increase in apoptosis of primary cancer cells. Thus, the pancreatic cancer on a chip model can further be used to test novel treatments as well as to study the behaviors of genetically modified primary cancer cells, as opposed to using in vivo models. Another example is Li et al.,¹⁰⁸ who modeled a kidney to investigate the mechanisms underlying polycystic kidney disease (PKD), which is a hereditary disorder caused by the formation of fluid-filled cysts in the kidneys, which disrupt kidney function and ultimately lead to kidney failure. Through various flow rate experiments researchers have observed a connection between glucose absorption and cyst formation. Excessive absorption triggered an increase in fluid secretion leading to cyst formation. To verify the connection between glucose absorption and cyst formation, a drug that inhibits the uptake of glucose, phloretin, was applied to the kidney on a chip system, where cyst size decreased by 77% when treated with the drug. This confirms the connection between glucose uptake and cystogenesis. Due to the precise fluid control ability the microfluidic chip can perform, they were able to successfully study cystogenesis and confirm the ability to use the kidney on a chip system for drug testing.

5.2. Digital microfluidics: genetic engineering and analysis of mammalian cells

The progress towards manipulating and editing mammalian cells has greatly progressed in the past few years such that we have the ability to program these cells to produce compounds like drugs, antibodies, and immunotherapies.¹²² Due to the controllable environment that DMF provides and the ability to automate fluidic operations, this makes it an ideal platform for gene editing of mammalian cells.^123,124 Through engineering mammalian cells, we can advance therapies by integrating circuits into cells to treat diseases like cancer or autoimmune diseases.¹²⁴ For example, to edit cells, like CAR-T cells (chimeric antigen receptor-T cell), for cell-based immunotherapies, human cells are taken out of the body, genetically engineered through physical and mechanical-poration techniques,¹²⁵ and these cells are expanded and re-introduced into the body. T-Cells are modified through genetic engineering to incorporate CARs, enhancing their ability to identify and eradicate harmful cells within the body, such as tumour cells. This personalized medicine technique involves tailoring T-cells to recognize the specific cancer cells present in individual patients and currently this technique has been observed in treating patients.^126,127

One of the main challenges of immunotherapy is that the cells used in this technique are difficult to recover and to grow in large numbers. Therefore, there is a requirement to work with fewer primary immune cells. However, transfecting these cells at lower densities has been found to have negative effects on their viability and effectiveness using the main technologies available.¹²⁸ To address this challenge, Little and Leung et al.,¹¹⁰ developed a system where they were able to electroporate 40 000 mammalian cells simultaneously on a DMF device (Fig. 5C). Their system called the triDrop functions by aligning three droplets in a continuous line on the device. The two outer droplets are composed of a high conductive liquid, and the middle containing the cells and foreign genetic material. A voltage is delivered to the outer droplets through the electrodes on the device which creates an electric field through all three droplets. This allows electroporation to happen without any damage to the very sensitive mammalian cells. The device consists of a glass bottom plate with chromium electrodes, SU-8 dielectric layer, and a Teflon hydrophobic coating. The top glass plate consists of gold electroporation electrodes. Using the triDrop system they were able to electroporate 2000 kDa FITC-tagged dextran molecules, 5 kB eGFP-plasmid, and eGFP mRNA into primary human CD4+ T cells. Additionally, to demonstrate the applicability for immunotherapy applications the triDrop system was able to perform a 5-plex arrayed electroporation assay. 4 different guide RNAs were designed for the knock-out of the TRAC locus, a site commonly used for T cell receptor therapies. A total of five combinations were created and 40 000 cells were used per reaction. The triDrop successfully used a significantly lower number of cells for electroporation, compared to commercial techniques which are important in immunotherapy applications when working with limited numbers of precious cells. This opens the door for further applications in immunotherapy and synthetic biology when using the triDrop system.

Another way to create genetic edits is by lentiviral-based systems, which are commonly used to transport external genetic material into mammalian cells. However, the process of preparing lentiviruses that carry the necessary tools for cell editing is time-consuming and demanding, involving steps that need careful fine-tuning and delicate handling. Therefore, Quach et al.¹⁰⁹ have developed a DMF platform called LENGEN which is a system able to culture cells, package, generate, and transduce lentiviruses in mammalian cells all on one chip (Fig. 5D). The current workflow of lentivirus transduction consists of generation of lentiviruses in a packaging cell line like HEK-293, where a packaging plasmid, envelope plasmid, and transfer plasmid are lipofected into the packaging cells. The lentivirus to be used for gene editing will be produced in the packaging cells and secreted into the cell culture media producing a viral titer. The virus is then added to the cell culture of the target cells for gene editing at a specific MOI (multiplicity of infection). Taking this process to the microscale will require less volumes of reagents and faster transduction rates. The DMF chip consists of chromium electrodes, SU-8 dielectric layer, and Teflon hydrophobic coating. Hydrophilic sites are present on the chip for the purpose of seeding and culturing the cells on device. Two tracks formed by the electrodes each lead to six target sites where the reactions can take place, giving the LENGEN system the ability to perform 12 simultaneous and automated lentiviral workflows (packaging, generation, and transduction). Beginning with the first step in the process HEK-293 cells are cultured on the device. Viral DNA needed for viral production and lipofectamine solution are delivered to the HEK-293 culturing site in droplets via electrode actuations. Following a 24 hour incubation, the lentiviral particles are secreted by the cells in the supernatant. Different viral concentrations are then delivered to the target cells on chip. Using this workflow Quach et al.,¹⁰⁹ were able to successfully generate lentiviruses to target the ESR1 gene in human breast cancer cells (MCF-7 cells); 70% of breast cancer tumors have expression for this gene. They successfully completed this using the LENGEN system through designing a virus with shRNA for knock-down effects on the gene and additionally a lentivirus with CRISPR-Cas9 for knock-out of the gene in MCF-7 cells.

With the design and construction of new or edited biological parts, genomic analysis is necessary to verify the engineered cells. DMF is particularly useful for integrating genomic analysis.^50,129,130 An example is Lamanna et al.,¹¹¹ created a pipeline called DISCO for isolating single cells and collecting the cell contents for either genome, transcriptome, or proteome analyses. The workflow on the DISCO consists of the cells being co-cultured and put through certain conditions for selection experiments. Cells can be individually addressed and selected through the custom control software or the AI the system that is paired with can be used to manually select cells in an automated fashion. Following selection, the laser delivers a high energy pulses to the selected cell(s) which creates a cavitation bubble that disrupts the cells membrane, releasing the cells contents into the media. The cell contents can then be collected and analyzed for genome, transcriptome, and proteome analysis. One -omic application they showed is with HAP1 cells that had the CD47 cell surface protein gene knocked out by CRISPR-Cas9 and were cultured on the device along with wild type cells. Both cell types were immunofluorescent labeled, imaged, and lysed, and a genome analysis was conducted. Differences in genome sequencing were found between wildtype and CRISPR edited cells, as well as among the CRISPR edited cells. This platform has opened the door for single cell analysis on a microfluidic platform that can further be used to study disease biology and the genetic heterogeneity that is found among patients.

5.3. Advancing cell analysis through droplet microfluidics: from antibody discovery to single-cell proteomics and HIV persistence insights

Droplet microfluidics' high-performance capability has broadened the scope of application for mammalian cells, paralleling its impact on other organisms like bacteria, yeast, and fungi. The use of droplets have opened the discovery of new therapeutics and has created a lens for visualizing mammalian cells at the single cell level.^{100,131–134} Two most popular applications that use such tools are antibody discovery for therapeutic purposes¹¹³ and single cell sequencing for disease pathology.^134,135

The extensive application of antibodies in therapeutic contexts has rendered them a significant focus for high-throughput screening. One of the pioneer works on antibody screening has been reported by El Debs et al.¹³³ The authors have reported a workflow using droplet microfluidics to screen and to sort hybridoma cells producing anti-ACE-1 (angiotensin converting enzyme 1) antibodies. These monoclonal antibodies could bind to drug targets specifically to treat congestive heart failure. In this work, hybridoma cells were encapsulated in 660 pL droplets, and incubated for 6 hours, then re-injected into a unique fusion device where the droplets were fused to smaller 25 pL droplets containing an ACE-1 fluorescent substrate, which creates a fluorescent signal upon the cleavage of ACE-1, the droplets were re-injected into a microfluidic binary sorter and sorted based on fluorescent signal. They validated their workflow by culturing sorted cells and have shown a 10-fold increase in anti-ACE-1 antibody production compared to non-sorted cells. Subsequently, Gérard et al.⁸⁹ have developed a workflow to sort antibodies based on their ability to bind to membrane-bound antigens. The authors have sorted primary mouse B-cells for their production of IgG antibodies and their ability to bind to soluble antigens TT (tetanus toxoid – a vaccine target) and antigen GPI (glycosylphosphatidylinositol – multifunctional enzyme), or for membrane bound antigen TSPAN8 (tetraspanin-8 – membrane-bound cancer target), assessing IgG repertoire diversity, clonal expansion, and somatic hypermutation. Cells were encapsulated into 40 pL droplets and binary sorted based on fluorescence. Upon being sorted, cells in the sorted droplets were re-encapsulated with lysis buffer and hydrogel beads carrying 10⁹ of the reverse transcriptase enzymes for VH and Vl (genes that code for the IgG antibody heavy chain and light chain sequences), in addition, each bead also contained its own unique barcode for distinguishing the separate cells. Using this workflow, the authors have yielded 100–1000 IgG sequences per mouse. Out of these, they were able to produce 77 recombinant antibodies, with 93% binding to the soluble antigen and 14% to the membrane-bound antigen. Additionally, the platform enabled recovery of ∼450–900 IgG sequences from ∼2200 activated human memory B cells, indicating its versatility when activated ex vivo. Thus, creating a platform using the unique ability to sequence individual cells and applying it to the discovery of antibodies for therapeutic purposes.

In the context of gene editing in mammalian cells, the ability to compartmentalize single cells made droplet microfluidics a great tool for addressing the challenge of developing a stable knockout cell line is a less time-consuming and intricate procedure. Samlali et al.,¹¹² have developed a platform combining droplet microfluidics and digital microfluidics for the isolation of individual cells in droplets (Fig. 5E), the identification of successfully edited clones, and the subsequent expansion of these isoclones. Their system can isolate single isoclones from a population of cells, showing that their platform can be used as an alternate method to picking out successful knock-out transfections. The hybrid microfluidic device is divided into two sections, one with a T-junction for on demand droplet generation and the second containing twelve traps used for cell trapping and encapsulation. The cell trapping functions by the cells being flowed through the device at a certain flow rate, due to the shape of the channels and the force from the aqueous flow the cells are trapped efficiently. Encapsulation of the trapped cells occurs during the phase change, when the aqueous flow containing cells is switched out for oil. Electrodes found under the trap when actuated create an electrostatic force, when the oil is flowed through the main channel cutting off the aqueous flow the electrostatic force from the electrode is greater further trapping the cell and surround media in the trap, forming an aqueous droplet. Three layers that make up the device are glass substrates with patterned chromium electrodes which forms the DMF layer. The second is a SU-8 5 dielectric layer and the third is a PDMS channel layer. The electrodes on the DMF layer are aligned to sit under the microfluidic channels and under the traps for increased droplet control when actuated. Additionally, on device manipulations can be automated through an in-house python program that combines the syringe pump system used for flow rate controls and the electrode actuation system into one. To show the device's ability to isolate single cells from a large population NCI-H1299 lung squamous cells were used. Cells were transfected using lipofectamine with eGFP plasmids as well as a plasmid with Cas9 and a sgRNA for the knock-out of the RAF1 gene. Cells are flowed into the droplet-digital microfluidics device, trapped, and fluorescent microscopy is used to determine which cells are producing eGFP confirming a successful transfection. The fluorescent cells are then encapsulated into droplets and can be released from the traps for sorting. Following sorting the isolated clones can then be recovered and cultured to a large scale. This droplet-digital microfluidics device shows how it can be integrated into the gene editing workflow, through effectively isolating successful transfections among the NCI-H1299 cells.

Droplet-based microfluidics constitute a pivotal instrument for the investigation of viral disease genomics, facilitating the unveiling of previously inaccessible information. One example is the work described by Clark et al.¹¹³ on rare CD4 T cells housing HIV, a significant obstacle to curing HIV. Those cells are a type of immune cell that contain latent or dormant HIV within their genetic material, they are present in individuals undergoing antiretroviral therapy (ART), which is a treatment regimen used to suppress HIV replication and reduce viral load. Existing studies have struggled to identify these cells due to their elusive nature, hindering the understanding of their contribution to HIV persistence under ART. The authors have developed a groundbreaking pipeline of droplet encapsulation of single cells with lysis buffer and molten agarose-poly(T) – which formed hydrogel beads with the genomic DNA and polyadenylated RNA, followed by whole-transcriptome amplification, hydrogel droplets screening and sorting. Applying FIND-seq to long-term ART-treated HIV individuals, they have analyzed gene patterns in memory CD4 T cells containing HIV gag DNA—a marker of the HIV latent or dormant. Their findings revealed unique transcriptomic signatures, shedding light on HIV-infected CD4 T cell persistence despite viral suppression, which could be essential insights to guide the pursuit of an HIV cure.

6. Applications of integrated synthetic biology and microfluidics in cell-free systems

Cell-free protein synthesis (CFPS) is a powerful alternative that uses in vitro protein synthesis machinery to achieve gene expression without living cells.^136,137 In a cell-free system, when exogenous mRNA or DNA is mixed with crude extract or lysate from a cell source, it creates a readily controllable transcription-translation machinery within an open environment. Most-used lysates are derived from bacteria, yeast, rabbit reticulocyte, wheat germ, or insects.¹³⁸ The CFPS system presents numerous advantages over cell-based systems. These advantages include mixing all the reaction components in a ‘one-pot’ reaction to enable rapid protein production, achieving high protein yields by introducing novel energy sources, can be performed without the requirement of continuously culturing and the need to passage cells,¹³⁹ and have showcased the potential for scaling-up cell-free reactions for massive production.¹⁴⁰ The fusion of a CFPS system with microfluidic chips holds significant potential for screening and therapeutic applications as well as diagnostic purposes (detecting toxins, antigens, and cells) as summarized in Table 5. Such an integration opens opportunities to develop rapid point-of-care (POC) diagnostics, cost-effective production of drugs,¹⁴⁰ antibodies,¹⁴⁰ vaccines,¹⁴¹ and enabling the realization of personalized medicine, which is a pressing need in the present context.¹⁴²

Table 5 Comparative analysis of microfluidic platforms for advancing synthetic biology in cell-free model: applications, advantages, and limitations

Microfluidic platform	Organism name	Synbio applications	Advantages	Disadvantages	Ref. number
Flow-based	Cell-free	Glycosylation of proteins	High glycosylation efficiency	Device fabrication is complicated	143
Microchemostat	Cell-free	Cell-free protein synthesis	Increased protein yield	Fabrication and control of the valve system are challenging	144
Droplet-based	Cell-free	Reduction of expression noise	Reduces gene expression noise	Three noise minimizing modules are added at the same time; it is difficult to control and investigate the impact of each module	145
Droplet microfluidics	Cell-free	Prototyping regulatory parts for cell-free protein synthesis	High throughput prototyping of regulatory parts	The protein yield decreased ∼30% than traditional bulk reaction	146
Droplet microfluidics	Cell-free	Synthesis of synthetic cells for studying cellular interactions	Mimics real cell behavior in various environments	Only applications use Mg^2+ as reaction mediator will be suitable for this protocol	147
Digital microfluidics	Cell-free	Portable Zika viral RNA extraction and amplification	Portable and automated diagnostic tool	The diagnostic process requires specialized equipment and expertise in digital microfluidics	148
Paper-based	Cell-free	Putrescine detection in food samples	Portable and affordable food quality detector	Instead of testing sample from PBS, the sample collection step should be more realistic to monitor the real-time food quality	149

---

6.1. Enhancing biosynthesis yields using microfluidics and cell-free systems

Cell-free systems have emerged as a powerful tool for biosynthesis of important molecules like proteins because it provides an alternative production for them whilst eliminating the use of live organisms for synthesis. However, the typical way of adding multiple reagents together in one step can cause unwanted competing reactions and side products. This is a result of lack of precise control over each step in the synthesis. On the other hand, using microfluidics to create a step-by-step workflow for each reaction in the synthesis eliminates these issues. It also provides advantages such as reaction compartmentalization, adjustable residence time, enzyme tethering for reuse, and the potential for screening conditions prior to scale-up manufacturing.

Microfluidics offers a unique opportunity for achieving spatiotemporal control over glycosylation reactions, which is challenging to accomplish using conventional cell-based and cell-free glycosylation systems. One of the main goals for CFPS is to ensure high yields of the protein of interest and eliminating side reactions that could lower such desired yields. An example of this is Aquino et al.,¹⁴³ presented a flow-based microfluidic device for cell-free glycoprotein synthesis (Fig. 6A). Their system was called the glycosylation-on-a-chip system, where three modules were connected. The first module was designed for mixing E. coli cell-free extract and plasmid DNA encoding the acceptor protein, the second module for protein glycosylation where glycosylation machinery derived from C. jejuni was coated inside the device channel wall. The product of the first device was streamed into the second module. After, the protein product will be transferred into the third component where proteins were isolated by immobilized metal affinity capture. The end product is the C. jejuni GalNAc5(Glc)Bac heptasaccharide with reducing end bacillosamine followed by five N-acetylgalactosamine residues (GalNAc) and a branching glucose (Glc). Using their device, they achieved 100% of glycosylation of the added acceptor protein within 2 hours, better than typical traditional methods. Most importantly, they demonstrated the pivotal glycosylation catalyst, Campylobacter jejuni OST enzyme PglB, can be immobilized within the device while maintaining high glycosylation efficiency. Similar to the above, groups like Lavickova et al.,¹⁴⁴ aim to further find conditions to obtain optimal protein production using cell-free recombinant elements (PURE) system and microfluidics (Fig. 6B). The authors employed a microchemostat chip with hydrogel membrane, a flow layer, a control layer, and a glass slide. In the flow layer, eight chemostat reactors and feeding channels are separated from the main reactors by hydrogel membranes. For the cell-free expression on chip, the chemostat reactors were filled with the PURExpress components in a ratio of 2 : 2 : 1 for solution A (2.5×, energy solution), solution B (2.5×, protein/ribosome), and DNA solution (5×), respectively. The feeding channel was loaded with a 1.5× feeding solution (1.5× solution A and PURE buffer), either once in the beginning of the experiment for the batch with static dialysis reaction, or every 10 min for the continuous dialysis experiments. Similarly, they performed steady-state and chemostat experiments with fresh PURE components injection and waste withdrawal. For standard batch reactions, augmenting the reaction chambers with semi-permeable membranes led to a prolongation of protein synthesis from approximately 2 to at least 30 h, increasing total protein eGFP yield by at least 7-fold. For chemostat operation, they showed that combining steady-state protein synthesis eGFP with continuous dialysis led to a 6-fold increase in protein levels at steady-state.

Fig. 6 Microfluidics and synthetic biology using cell-free systems. (A) Schematic of glycosylation-on-a-chip system. In the first module of the device, one stream containing E. coli cell-free extract and a second stream containing plasmid DNA encoding the acceptor protein are combined at the inlet and mixed by diffusion as they travel through the channels. The product of the first chip is then delivered to a second module where it is subjected to an environment enriched with glycosylation machinery. In the third module, protein product is isolated using immobilized metal affinity capture (IMAC) (adapted from ref. 143 with permission from Frontiers Media SA) (Aquino et al.¹⁴³). (B) Schematic of the microfluidic device featuring eight individual chemostat reactors and the features of a single reactor showing the feeding channel separated from the main reactor by hydrogel membranes. The flow layer is shown in black, the control layer in dark gray, and the anti-evaporation layer in light gray. A bright-field image of the hydrogel membranes between the reactor and feeding channel is also shown (adapted from ref. 144 with permission from American Chemical Society) (Lavickova et al.¹⁴⁴). (C) Digital microfluidic system showing a schematic containing an exploded view of the DMF cartridge, and cross-sectional view of the “Zed Box” control hardware (adapted from ref. 148 with permission from Royal Society of Chemistry) (Narahari et al.¹⁴⁸). (D) A schematic of cell-free putrescine biosensor stored on a paper-based device to detect spoiled beef (adapted from ref. 149 with permission from American Chemical Society) (Selim et al.¹⁴⁹).

Some groups recognize the problem of expression noise in cell-free applications, which was pointed out by Bartelds et al.,¹⁴⁵ and they reported a microfluidic system to reduce such noise. They used a droplet-based microfluidic device to synthesize MazF protein (endoribonuclease enzyme) to reduce the cell-free gene expression noise. The MazF protein is a sequence-specific ribonuclease that acts preferentially on ACA-sequence-containing single-stranded RNA, resulting in inhibition of translation through site-specific cleavage.¹⁵⁰ The device consisted of a PDMS channel layer and a glass slide. The channels for droplet production are 25 μm in width. The inner solution was composed of the complete gene expression mixture (in vitro transcription–translation reaction mixture: lysate, feeding buffer, and DNA templates) and the outer solution of 5% 008-FluoroSurfactant in FC40 oil. Their findings demonstrated the successful synthesis of MazF in a cell-free gene expression system, effectively inhibiting the expression of both destabilized eGFP and mCherry reporter genes by targeting mRNA with ACA sites. Notably, the presence of MazF led to a significant ∼18-fold increase in the degradation rate of destabilized eGFP mRNA. They have also established that the T7p10-MazF and T7p14-deGFP templates, along with the number of ACA sites, serve as independent factors to fine-tune protein levels in the cell-free gene expression system. Leveraging droplet-based microfluidics, we have shown that MazF synthesis reduces destabilized eGFP expression noise by more than 2-fold.

6.2. Microfluidic cell-free technology: designing multicellular systems

Microfluidic devices offer the potential to assemble synthetic cellular populations from the ground up. This capability establishes the groundwork for designing minimal living tissues, to determine which genetic parts are necessary for growth and communication. Through the fusion of engineered micron-sized droplet compartments and integrated reaction networks, we can enable the design and construction of complex and multiscale chemical systems, initiating a bottom-up approach to designing the minimal cell.¹⁵¹

Engineering and optimizing regulatory parts plays an important role in developing the minimal cell programming toolbox. With droplet microfluidics, Gan et al.,¹⁴⁶ proposed a high-throughput method to prototype regulatory parts and validate it by CFPS. The authors used a microfluidic device to perform a workflow: first, single plasmids from a user-designed library were encapsulated in droplets. Secondly, cell-free protein expression reagents were mixed with the plasmid-containing droplets by pico-injection. Lastly, fluorescent-based sorting was performed to separate droplets containing desired RBS mutant from waste droplets. Using this platform, they successfully prototyped a large randomized RBS mutant library (106 mutants) using E. coli CFPS system. Based on the next generation sequence (NGS) data, they characterized 48 beneficial RBS mutants that were significantly enriched. They identified 8–9 mutants that had same activities as the wild-type RBS after only one round of screening despite low similarity to the wild-type RBS in sequence.

Moving towards building a complex and fully functional synthetic cell is to first construct a phospholipid barrier to mimic the cell's membrane. A work by Weiss et al.,¹⁴⁷ constructed a high-throughput microfluidic device to produce synthetic cells called droplet-stabilized giant unilamellar vesicles (dsGUVs). To generate dsGUVs, they first encapsulated various lipid compositions in the form of GUVs. Then, using pico-injection, they introduced 10 mM of MgCl₂ into the droplet to form a continuous lipid bilayer inside. They also incorporated transmembrane proteins and cytoskeletal proteins into the dsGUVs by pico-injection. By combining all of these separate cellular components they were able to create droplets that mimicked real cells. Lastly, the success of the dsGUVs was shown by removing them from the oil phase to an aqueous phase, where they remained stable. This shows that the dsGUVs created using droplet microfluidics have potential for studying cells interactions in various environments. A further step was taken by Gonzales et al.,¹⁵¹ as they used double-emulsion microfluidics for generating monodisperse liposomes encapsulating cell-free expression systems (CFESs). They employed a standard microfluidic design that included one aqueous phase containing PURE CFES, one liquid-oil inlet, and one outer aqueous inlet. They used a plasmid consisting of a constitutive T7 RNA polymerase-mediated promoter to express a red fluorescent mCherry protein and two copies of a dimeric Broccoli RNA aptamer stabilized by the F30 stem-loop between the stop codon of mCherry and the terminator of the gene construct. Binding of a small-molecule dye, 3,5-difluoro-4-hydroxybenzylidene imidazolinone (DFHBI), to the Broccoli RNA aptamer results in a green fluorescence signal and allows simultaneous fluorescence monitoring of transcribed mRNA and reporter protein levels. The cell-free compartments and DNA were encapsulated into droplets, then those droplets were encapsulated again by a liquid layer forming the double-emulsion synthetic cells. By analyzing the gene expression in synthetic cells, the authors claimed that it is different from gene expression in bulk environments, emphasizing the importance of the physiological environment to compartmentalized biochemical reactions.

6.3. Using cell-free technology for diagnostics and biosensing in microfluidics

The power of synthetic circuits and its integration with cell-free systems and microfluidics has enabled a boon in diagnostics and biosensors.^152–158 This approach leverages the power of cell-free reactions to detect and analyze target molecules with high sensitivity and specificity due to the controlled and confined environment within microfluidic devices. By this integration, rapid and accurate diagnostic tools can be developed for various applications in health and industry. For example, the recent outbreak of SARS-CoV-2 pandemic has emphasized the need for portable, point-of-care platforms that can extract and quantify viral RNA from patient samples and help distribute the diagnostic burden on the healthcare system.¹⁵⁹ Another example is the increasing concern in public health due to food quality control. Any deviation from proper processing, storage, or distribution may result in serious biochemical contamination.¹⁶⁰

Being able to detect infectious diseases using low-volume patient samples is a key advantage of microscale technologies such as microfluidics. Narahari et al.,¹⁴⁸ described a digital microfluidic platform for portable, automated, and integrated Zika viral RNA extraction and amplification, their system included a four-layer DMF cartridge and DropBot automation stage (Fig. 6C). The workflow consisted of patient plasma lysis using DMF, RNA capture by magnetic particles and loading it to the DMF automation system, use of a motorized magnet to separate lysis debris and clean RNA, amplification of the RNA, and virus RNA detection using paper disc with cell-free reagents and DNA switch. The authors showed that their microfluidic method is the first to be employed for sample processing prior to a one-pot colorimetric assay aimed at detecting amplified RNA. Their findings reveal a detection limit of approximately 1000 copies per mL for free RNA in plasma and between 10 and 100 PFU mL⁻¹ for samples containing Zika virus particles. These values correspond to RT-qPCR cycle threshold values of 37.9 and 32.0, respectively, which fall within physiological levels observed in infected patient urine (and, in some cases, serum). This system also demonstrated the possibility of using a portable automation system to diagnose infectious disease outbreaks remotely. Besides the medical applications, microfluidic technologies also offer novel ways to monitor food security, one of those works was done by Selim et al.¹⁴⁹ They introduced a paper-based microfluidic system for putrescine (potential biomarker for microbial deterioration) detection in beef samples (Fig. 6D). A standard paper microfluidic device was employed in this study, it included four separate loading zones: negative control (+beef/−biosensor), positive control (+putrescine/+biosensor), biosensor with no sample (−beef/+biosensor), and biosensor with sample (+beef/+biosensor). CFPS reagents were immobilized on the paper disc for downstream detection. In their putrescine biosensing circuit, the T7 promoter regulates the expression of the PuuR repressor, which binds to a synthetic putrescine-inducible promoter containing the operator puuO on the reporter plasmid. In the absence or low levels of putrescine, the repressor, PuuR, binds to the synthetic promoter, obstructing RNA polymerase from binding to puuO and preventing eGFP production. On the other hand, when putrescine is present, it binds to PuuR, releasing it from puuO, thereby allowing RNA polymerase to bind to the promoter and enabling eGFP expression. The new hybrid promoter found in this study showed a faster response time (∼1 h), around a 33-fold increase in dynamic range and a lower detection limit of 5.34 mM. Paired with paper microfluidics, this system can be used as a portable, affordable, and biodegradable food quality detector.

7. Conclusion and future perspectives

In this review, we have summarized useful ways on how microfluidics have been used across synthetic biology applications. Bringing complex biological processes to the microscale combined with automation, enables microfluidics an ideal tool for solving universally laborious biotechnological problems. Highlighting the combination of microfluidics with key organisms like bacteria, yeast, filamentous fungi, mammalian cells, and cell-free systems provides a blueprint to incorporating microfluidics into the synthetic biology workflow.

Many challenges that hinder productivity across numerous fields include frequent manual intervention to control biological processes, reliable regulation of experimental conditions and lack of standardized automation protocols. Microfluidics provides a steady manipulation of small volumes of fluid, allowing continuous control over the microenvironment, whilst providing the benefits of a standardized automated platform. Combining microfluidics and synthetic biology has allowed the automated genetic engineering of bacteria, which are organisms used heavily for alternative biofuel production and environmental bioremediation. In addition, the combination has enabled successful high throughput screening of yeast and fungi to identify strains able to produce valuable proteins more efficiently like enzymes and antibodies for biomedical applications. Furthermore, culturing mammalian cells on highly controlled microfluidic chips have allowed the accurate development of organ on a chip models that have potential for personalized medical treatment applications as well as for the very promising gene-editing immunotherapeutic applications.¹⁶¹ Filamentous fungi have been sorted in a high throughput and automated way for the efficient production of relevant enzymes important in the production of biofuels and commodity chemicals. Lastly, using droplet microfluidics researchers were able to synthesize synthetic cells in droplets providing biologically relevant platforms for protein and cellular behaviors without the need for live cells. The intersection of microfluidics with important organisms used for synthetic biology has allowed new breakthroughs and the overcoming of technically challenging obstacles, challenges unable to be accomplished on the macroscale with a lack of precise liquid control and an unstable microenvironment.

However, the transition from controlled laboratory environments to real-world industrial applications and large-scale production for commercial use remains a significant challenge in microfluidics technology. The absence of standardized designs results in multiple iterations and optimization of different devices for similar organisms or purposes. The lack of standardizing devices significantly increases the cost associated with fabrication and the need for a cleanroom for prototyping microfluidic devices, often limit accessibility to these technologies, constraining their widespread use.¹⁶²

Addressing these challenges may necessitate the advancement of microfluidic chip technology by creating more adaptable platforms capable of accommodating a wide range of organisms, or applications within the same organism. The “mother-machine”, initially developed by Wang et al.³⁷ in 2010, serves as a foundational platform for bacterial time-lapse studies and exemplifies this adaptability. Recent innovations in droplet sorting devices by Isozaki et al.,^163,164 Caen et al.,¹⁶⁵ and Samlali et al.,¹⁰⁰ offer promising prospects for standardizing sorter devices in high-throughput screenings across various organisms. These devices possess the capacity to sort different droplet sizes, catering to bacterial, yeast cells, single mammalian cells in the picolitre range, as well as spheroids, organoids, and filamentous fungi in the nanoliter range.

Furthermore, advancements in image and data processing techniques are novel approaches that can be integrated with microfluidics to analyze complex biological or chemical processes at a microscopic level without relying on traditional chemical tagging methods like fluorescence or absorbance. This expansion could be increased by the progress of artificial intelligence algorithms, enabling integration with pattern recognition, cell or particle classification, and the prediction of dynamic behaviors within these systems. Such advancements could lead to the development of fully automated and autonomous systems, fostering image-based higher throughput analysis in microfluidics.¹⁶⁶

In summary, while there are challenges to overcome, the continuous development and integration of microfluidics as a tool for synthetic biology has revolutionized various scientific and industrial fields from biomedical applications to bioenergy. Therefore, the use of microfluidics technology will help to further accelerate the emergence of innovative solutions within synthetic biology, driving transformative developments in the field.

Conflicts of interest

The authors declare no competing interests.

Acknowledgements

The authors also acknowledge the funding support from the Natural Sciences and Engineering Research Council of Canada and the Fonds de Recherche Nature et technologies. CLA thanks the Fonds de Recherche for a graduate scholarship and SCCS thanks Concordia University for a Research Chair.

References

1. S. M. Brooks and H. S. Alper, Nat. Commun., 2021, 12, 1–16.
2. T. A. Dixon, P. S. Freemont, R. A. Johnson and I. S. Pretorius, Nat. Commun., 2022, 13, 1–5.
3. S. R. Little, J. M. Perry, K. Samlali and S. C. C. Shih, in Droplet Microfluidics, The Royal Society of Chemistry, 2020.
4. P. C. Gach, K. Iwai, P. W. Kim, N. J. Hillson and A. K. Singh, Lab Chip, 2017, 17, 3388–3400.
5. Y. Ai, R. Xie, J. Xiong and Q. Liang, Small, 2020, 16, 1903940.
6. L. A. Damiati, S. A. Damiati and S. Damiati, in New Frontiers and Applications of Synthetic Biology, ed. V. Singh, Academic Press, 2022, pp. 423–435.
7. M. Abdelgawad and A. R. Wheeler, Adv. Mater., 2009, 21, 920–925.
8. E. Samiei, M. Tabrizian and M. Hoorfar, Lab Chip, 2016, 16, 2376–2396.
9. Q. Zhao, D. Yuan, J. Zhang and W. Li, Micromachines, 2020, 11, 461.
10. T. Moragues, D. Arguijo, T. Beneyton, C. Modavi, K. Simutis, A. R. Abate, J.-C. Baret, A. J. deMello, D. Densmore and A. D. Griffiths, Nat. Rev. Methods Primers, 2023, 3, 32.
11. K. S. Elvira, F. Gielen, S. S. H. Tsai and A. M. Nightingale, Lab Chip, 2022, 22, 859–875.
12. T. S. Kaminski, O. Scheler and P. Garstecki, Lab Chip, 2016, 16, 2168–2187.
13. Anushka, A. Bandopadhyay and P. K. Das, Eur. Phys. J.: Spec. Top., 2023, 232, 781–815.
14. E. Noviana, T. Ozer, C. S. Carrell, J. S. Link, C. McMahon, I. Jang and C. S. Henry, Chem. Rev., 2021, 121, 11835–11885.
15. F. Zhang, Y. Sun, Y. Zhang, W. Shen, S. Wang, Q. Ouyang and C. Luo, Commun. Biol., 2022, 5, 1–8.
16. D. Choudhary, V. Lagage, K. R. Foster and S. Uphoff, Cell Rep., 2023, 42, 112168.
17. Y. Yang, O. Karin, A. Mayo, X. Song, P. Chen, A. L. Santos, A. B. Lindner and U. Alon, Nat. Commun., 2023, 14, 1–11.
18. M. Mistretta, N. Gangneux and G. Manina, Sci. Rep., 2022, 12, 1–18.
19. G. Graham, N. Csicsery, E. Stasiowski, G. Thouvenin, W. H. Mather, M. Ferry, S. Cookson and J. Hasty, Proc. Natl. Acad. Sci. U. S. A., 2020, 117, 3301–3306.
20. A. Lezia, N. Csicsery and J. Hasty, Cell Syst., 2022, 13, 365–375.e5.
21. J. M. Perry, G. Soffer, R. Jain and S. C. C. Shih, Lab Chip, 2021, 21, 3730–3741.
22. Y. Liu, P. Jeraldo, H. Mendes-Soares, T. Masters, A. E. Asangba, H. Nelson, R. Patel, N. Chia and M. Walther-Antonio, ACS Omega, 2021, 6, 25642–25651.
23. S. Hu, Y. Jie, K. Jin, Y. Zhang, T. Guo, Q. Huang, Q. Mei, F. Ma and H. Ma, Biosensors, 2022, 12, 324.
24. W. Qiu and S. Nagl, ACS Sens., 2021, 6, 1147–1156.
25. A. A. Sklavounos, C. R. Nemr, S. O. Kelley and A. R. Wheeler, Lab Chip, 2021, 21, 4208–4222.
26. S. Neun, P. Brear, E. Campbell, T. Tryfona, K. El Omari, A. Wagner, P. Dupree, M. Hyvönen and F. Hollfelder, Nat. Chem. Biol., 2022, 18, 1096–1103.
27. C. F. A. Wong, L. Van Vliet, S. V. Bhujbal, C. Guo, M. Sletmoen, B. T. Stokke, F. Hollfelder and R. Lale, ACS Synth. Biol., 2021, 10, 1882–1894.
28. C. S. Karamitros, M. Morvan, A. Vigne, J. Lim, P. Gruner, T. Beneyton, J. Vrignon and J. C. Baret, Anal. Chem., 2020, 92, 4908–4916.
29. H. Yuan, R. Tu, X. Tong, Y. Lin, Y. Zhang and Q. Wang, J. Ind. Microbiol. Biotechnol., 2022, 49, 3.
30. M. Xie and M. Fussenegger, Nat. Rev. Mol. Cell Biol., 2018, 19, 507–525.
31. A. Roy and S. Klumpp, Biophys. J., 2018, 114, 484–492.
32. P.-F. Xia, H. Ling, J. L. Foo and M. W. Chang, Biotechnol. Adv., 2019, 37, 107393.
33. T. S. Gardner, C. R. Cantor and J. J. Collins, Nature, 2000, 403, 339–342.
34. L. Potvin-Trottier, N. D. Lord, G. Vinnicombe and J. Paulsson, Nature, 2016, 538, 514–517.
35. C. L. Kelly, A. W. K. Harris, H. Steel, E. J. Hancock, J. T. Heap and A. Papachristodoulou, Nucleic Acids Res., 2018, 46, 9875–9889.
36. B. Wang, R. I. Kitney, N. Joly and M. Buck, Nat. Commun., 2011, 2, 508.
37. P. Wang, L. Robert, J. Pelletier, W. L. Dang, F. Taddei, A. Wright and S. Jun, Curr. Biol., 2010, 20, 1099–1103.
38. P. Allard, F. Papazotos and L. Potvin-Trottier, Front. Bioeng. Biotechnol., 2022, 10, 1–16.
39. Y. Zhang, Y. Cai, L. Zeng, P. Liu, L. Z. Ma and J. Liu, Small Sci., 2022, 2, 2200047.
40. H. H. McAdams and A. Arkin, Trends Genet., 1999, 15, 65–69.
41. D. J. Wilkinson, Nat. Rev. Genet., 2009, 10, 122–133.
42. M. Daniels, S. van Vliet and M. Ackermann, ISME J., 2023, 17, 406–416.
43. S. Gupta, T. D. Ross, M. M. Gomez, J. L. Grant, P. A. Romero and O. S. Venturelli, Nat. Commun., 2020, 11, 1–15.
44. J. Atencia, J. Morrow and L. E. Locascio, Lab Chip, 2009, 9, 2707–2714.
45. T. Wang, P. Shen, Y. He, Y. Zhang and J. Liu, Nat. Chem. Biol., 2023, 19, 940–950.
46. P. Fu, Biotechnol. J., 2006, 1, 690–699.
47. Z. Chen and M. B. Elowitz, Cell, 2021, 184, 2284–2301.
48. M. B. Holowko, E. K. Frow, J. C. Reid, M. Rourke and C. E. Vickers, Synth. Biol., 2020, 6, ysaa026.
49. E. Moazami, J. M. Perry, G. Soffer, M. C. Husser and S. C. C. Shih, Anal. Chem., 2019, 91, 5159–5168.
50. A. N. Davis, K. Samlali, J. B. Kapadia, J. Perreault, S. C. C. Shih and N. Kharma, ACS Omega, 2021, 6, 22514–22524.
51. H. M. Hegab, A. Elmekawy and T. Stakenborg, Biomicrofluidics, 2013, 7, 21502.
52. C. R. Nemr, A. A. Sklavounos, A. R. Wheeler and S. O. Kelley, SLAS Technol., 2023, 28, 2–15.
53. D. Yang, S. Y. Park and S. Y. Lee, Adv. Sci., 2021, 8, 2100743.
54. J. T. Boock, A. J. E. Freedman, G. A. Tompsett, S. K. Muse, A. J. Allen, L. A. Jackson, B. Castro-Dominguez, M. T. Timko, K. L. J. Prather and J. R. Thompson, Nat. Commun., 2019, 10, 487.
55. X. Liu, L. Li, J. Liu, J. Qiao and G. R. Zhao, Biotechnol. Biofuels, 2019, 12, 1–12.
56. V. J. J. Martin, D. J. Pitera, S. T. Withers, J. D. Newman and J. D. Keasling, Nat. Biotechnol., 2003, 21, 796–802.
57. F. Zhang, J. M. Carothers and J. D. Keasling, Nat. Biotechnol., 2012, 30, 354–359.
58. M. P. Robinson, N. Ke, J. Lobstein, C. Peterson, A. Szkodny, T. J. Mansell, C. Tuckey, P. D. Riggs, P. A. Colussi, C. J. Noren, C. H. Taron, M. P. Delisa and M. Berkmen, Nat. Commun., 2015, 8, 8072.
59. M. C. Husser, P. Q. N. Vo, H. Sinha, F. Ahmadi and S. C. C. Shih, ACS Synth. Biol., 2018, 7, 933–944.
60. S. H. Au, S. C. C. Shih and A. R. Wheeler, Biomed. Microdevices, 2011, 13, 41–50.
61. P. Tewari Kumar, D. Decrop, S. Safdar, I. Passaris, T. Kokalj, R. Puers, A. Aertsen, D. Spasic and J. Lammertyn, Micromachines, 2020, 11, 308.
62. J. Wu, M. Zhang, J. Huang, J. Guan, C. Hu, M. Shi, S. Hu, S. Wang and H. Ma, Analyst, 2023, 148, 4659–4667.
63. T. E. Saleski, M. T. Chung, D. N. Carruthers, A. Khasbaatar, K. Kurabayashi and X. N. Lin, Sci. Adv., 2021, 7, 1–12.
64. T. E. Saleski, A. R. Kerner, M. T. Chung, C. M. Jackman, A. Khasbaatar, K. Kurabayashi and X. N. Lin, Metab. Eng., 2019, 54, 232–243.
65. J. C. Way, P. A. Silver and R. J. Howard, Int. J. Astrobiol., 2011, 10, 359–364.
66. R. C. Rodrigues, H. S. Pereira, R. L. Senra, A. de Oliveira Barros Ribon and T. A. de Oliveira Mendes, Food Chemistry Advances, 2023, 3, 100476.
67. P. P. Peralta-Yahya, M. Ouellet, R. Chan, A. Mukhopadhyay, J. D. Keasling and T. S. Lee, Nat. Commun., 2011, 2, 483.
68. G. Bokinsky, P. P. Peralta-Yahya, A. George, B. M. Holmes, E. J. Steen, J. Dietrich, T. S. Lee, D. Tullman-Ercek, C. A. Voigt, B. A. Simmons and J. D. Keasling, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 19949–19954.
69. C. M. Denby, R. A. Li, V. T. Vu, Z. Costello, W. Lin, L. J. G. Chan, J. Williams, B. Donaldson, C. W. Bamforth, C. J. Petzold, H. V. Scheller, H. G. Martin and J. D. Keasling, Nat. Commun., 2018, 9, 1–10.
70. A. Kang, K. W. George, G. Wang, E. Baidoo, J. D. Keasling and T. S. Lee, Metab. Eng., 2016, 34, 25–35.
71. K. J. Adolfsen, I. Callihan, C. E. Monahan, P. J. Greisen, J. Spoonamore, M. Momin, L. E. Fitch, M. J. Castillo, L. Weng, L. Renaud, C. J. Weile, J. H. Konieczka, T. Mirabella, A. Abin-Fuentes, A. G. Lawrence and V. M. Isabella, Nat. Commun., 2021, 12, 1–13.
72. Y. Li, Y. Jiang, J. Paxman, R. O'Laughlin, S. Klepin, Y. Zhu, L. Pillus, L. S. Tsimring, J. Hasty and N. Hao, Science, 2020, 369, 325–329.
73. E. A. Sarnoski, R. Song, E. Ertekin, N. Koonce and M. Acar, iScience, 2018, 7, 96–109.
74. X. Xu, Z. Zhu, Y. Wang, Y. Geng, F. Xu, M. A. Marchisio, Z. Wang and D. Pan, Anal. Bioanal. Chem., 2021, 413, 2181–2193.
75. Y. Wang, Z. Zhu, K. Liu, Q. Xiao, Y. Geng, F. Xu, S. Ouyang, K. Zheng, Y. Fan, N. Jin, X. Zhao, M. A. Marchisio, D. Pan and Q. A. Huang, J. Nanobiotechnol., 2022, 20, 1–17.
76. X. Su, J. Yang, H. Yuan, C. Liu, R. Tu, P. Liu, Q. Wang and L. Zhu, J. Agric. Food Chem., 2022, 70, 13700–13708.
77. S. A. Johansson, T. Dulermo, C. Jann, J. D. Smith, A. Pryszlak, G. Pignede, D. Schraivogel, D. Colavizza, T. Desfougères, C. Rave, A. Farwick, C. A. Merten, K. R. Roy, W. Wei and L. M. Steinmetz, Lab Chip, 2023, 23, 3704–3715.
78. E. Lebrun, V. Shenshin, C. Plaire, V. Vigneres, T. Pizette, B. Dumas, J. M. Nicaud and G. Mottet, Lab Chip, 2023, 23, 3487–3500.
79. K. Wink, M. van der Loh, N. Hartner, M. Polack, C. Dusny, A. Schmid and D. Belder, Angew. Chem., Int. Ed., 2022, 61, e202204098.
80. E. K. Bowman, J. M. Wagner, S. F. Yuan, M. Deaner, C. M. Palmer, S. D'Oelsnitz, L. Cordova, X. Li, F. F. Craig and H. S. Alper, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e202204098.
81. S. Li, Y. Zhang, L. Li, Y. Yuan, H. Sun, X. H. Xing, X. Wang and C. Zhang, Biochem. Eng. J., 2023, 198, 109036.
82. E. J. Thak, S. J. Yoo, H. Y. Moon and H. A. Kang, FEMS Yeast Res., 2020, 20, foaa009.
83. R. S. K. Walker and I. S. Pretorius, Genes, 2018, 9, 340.
84. F. J. Boonekamp, E. Knibbe, M. A. Vieira-Lara, M. Wijsman, M. A. H. Luttik, K. van Eunen, M. den Ridder, R. Bron, A. M. Almonacid Suarez, P. van Rijn, J. C. Wolters, M. Pabst, J.-M. Daran, B. M. Bakker and P. Daran-Lapujade, Cell Rep., 2022, 39, 111010.
85. A. H. Kachroo, M. Vandeloo, B. M. Greco and M. Abdullah, Dis. Models Mech., 2022, 15, dmm049309.
86. Y. Li, M. Jin, R. O'Laughlin, P. Bittihn, L. S. Tsimring, L. Pillus, J. Hasty and N. Hao, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 11253–11258.
87. J. Ryley and O. M. Pereira-Smith, Yeast, 2006, 23, 1065–1073.
88. K. Eyer, R. C. L. Doineau, C. E. Castrillon, L. Briseño-Roa, V. Menrath, G. Mottet, P. England, A. Godina, E. Brient-Litzler, C. Nizak, A. Jensen, A. D. Griffiths, J. Bibette, P. Bruhns and J. Baudry, Nat. Biotechnol., 2017, 35, 977–982.
89. A. Gérard, A. Woolfe, G. Mottet, M. Reichen, C. Castrillon, V. Menrath, S. Ellouze, A. Poitou, R. Doineau, L. Briseno-Roa, P. Canales-Herrerias, P. Mary, G. Rose, C. Ortega, M. Delincé, S. Essono, B. Jia, B. Iannascoli, O. Richard-Le Goff, R. Kumar, S. N. Stewart, Y. Pousse, B. Shen, K. Grosselin, B. Saudemont, A. Sautel-Caillé, A. Godina, S. McNamara, K. Eyer, G. A. Millot, J. Baudry, P. England, C. Nizak, A. Jensen, A. D. Griffiths, P. Bruhns and C. Brenan, Nat. Biotechnol., 2020, 38, 715–721.
90. G. P. Monteiro, I. M. D. C. Tavares, M. C. F. de Carvalho, M. S. Carvalho, A. B. Pimentel, P. H. Santos, E. V. D. B. Vilas Boas, J. R. de Oliveira, V. R. Capelossi, M. Bilal and M. Franco, Chem. Eng. Commun., 2023, 210, 1207–1222.
91. D. T. Troiano, T. Hofmann, S. Brethauer and M. H. P. Studer, Curr. Opin. Biotechnol., 2023, 81, 102942.
92. S. Wang, H. Chen, X. Tang, H. Zhang, W. Chen and Y. Q. Chen, Appl. Microbiol. Biotechnol., 2017, 101, 8063–8075.
93. C. Jo, J. Zhang, J. M. Tam, G. M. Church, A. S. Khalil, D. Segrè and T.-C. Tang, Mater. Today Bio, 2023, 19, 100560.
94. E. Moreno-Giménez, M. Gandía, Z. Sáez, P. Manzanares, L. Yenush, D. Orzáez, J. F. Marcos and S. Garrigues, Front. Bioeng. Biotechnol., 2023, 11, 1–17.
95. K. Aleklett, P. Ohlsson, M. Bengtsson and E. C. Hammer, ISME J., 2021, 15, 1782–1793.
96. A. Hopke, A. Mela, F. Ellett, D. Carter-House, J. F. Peña, J. E. Stajich, S. Altamirano, B. Lovett, M. Egan, S. Kale, I. Kronholm, P. Guerette, E. Szewczyk, K. McCluskey, D. Breslauer, H. Shah, B. R. Coad, M. Momany and D. Irimia, PLoS One, 2021, 16, 1–20.
97. A. Gimeno, C. E. Stanley, Z. Ngamenie, M. H. Hsung, F. Walder, S. S. Schmieder, S. Bindschedler, P. Junier, B. Keller and S. Vogelgsang, Commun. Biol., 2021, 4, 1–10.
98. C. Zhang, X. Wu, F. Song, S. Liu, S. Yu and J. Zhou, ACS Sens., 2023, 8, 3468–3477.
99. X. C. Luu, Y. Shida, Y. Suzuki, N. Sato, A. Nakumura and W. Ogasawara, New Biotechnol., 2022, 72, 149–158.
100. K. Samlali, C. L. Alves, M. Jezernik and S. C. C. Shih, Microsyst. Nanoeng., 2022, 8, 1–16.
101. S. Vandelook, E. Elsacker, A. Van Wylick, L. De Laet and E. Peeters, Fungal Biol. Biotechnol., 2021, 8, 20.
102. A. Khairullina, N. Micic, H. J. L. Jørgensen, N. Bjarnholt, L. Bülow, D. B. Collinge and B. Jensen, Plants, 2023, 12, 500.
103. W. Rivera-Méndez, M. Obregón, M. E. Morán-Diez, R. Hermosa and E. Monte, Biol. Control, 2020, 141, 104145.
104. T. Beneyton, I. P. M. Wijaya, P. Postros, M. Najah, P. Leblond, A. Couvent, E. Mayot, A. D. Griffiths and A. Drevelle, Sci. Rep., 2016, 6, 1–10.
105. R. He, R. Ding, J. A. Heyman, D. Zhang and R. Tu, J. Ind. Microbiol. Biotechnol., 2019, 46, 1603–1610.
106. I. Pediaditakis, K. R. Kodella, D. V. Manatakis, C. Y. Le, C. D. Hinojosa, W. Tien-Street, E. S. Manolakos, K. Vekrellis, G. A. Hamilton, L. Ewart, L. L. Rubin and K. Karalis, Nat. Commun., 2021, 12, 1–17.
107. M. R. Haque, C. R. Wessel, D. D. Leary, C. Wang, A. Bhushan and F. Bishehsari, Microsyst. Nanoeng., 2022, 8, 36.
108. S. R. Li, R. E. Gulieva, L. Helms, N. M. Cruz, T. Vincent, H. Fu, J. Himmelfarb and B. S. Freedman, Nat. Commun., 2022, 13, 1–12.
109. A. B. V. Quach, S. R. Little and S. C. C. Shih, Anal. Chem., 2022, 94, 4039–4047.
110. S. R. Little, Z. Leung, A. B. V. Quach, A. Hirukawa, F. Gholizadeh, M. Hajiaghayi, P. J. Darlington and S. C. C. Shih, Adv. Mater. Technol., 2023, 8, 1–15.
111. J. Lamanna, E. Y. Scott, H. S. Edwards, M. D. Chamberlain, M. D. M. Dryden, J. Peng, B. Mair, A. Lee, C. Chan, A. A. Sklavounos, A. Heffernan, F. Abbas, C. Lam, M. E. Olson, J. Moffat and A. R. Wheeler, Nat. Commun., 2020, 11, 1–13.
112. K. Samlali, F. Ahmadi, A. B. V. Quach, G. Soffer and S. C. C. Shih, Small, 2020, 16, 2070190.
113. I. C. Clark, P. Mudvari, S. Thaploo, S. Smith, M. Abu-Laban, M. Hamouda, M. Theberge, S. Shah, S. H. Ko, L. Pérez, D. G. Bunis, J. S. Lee, D. Kilam, S. Zakaria, S. Choi, S. Darko, A. R. Henry, M. A. Wheeler, R. Hoh, S. Butrus, S. G. Deeks, F. J. Quintana, D. C. Douek, A. R. Abate and E. A. Boritz, Nature, 2023, 614, 318–325.
114. X. Yan, X. Liu, C. Zhao and G.-Q. Chen, Signal Transduction Targeted Ther., 2023, 8, 199.
115. A. Boussommier-Calleja, R. Li, M. B. Chen, S. C. Wong and R. D. Kamm, Trends Cancer, 2016, 2, 6–19.
116. C. M. Leung, P. de Haan, K. Ronaldson-Bouchard, G. A. Kim, J. Ko, H. S. Rho, Z. Chen, P. Habibovic, N. L. Jeon, S. Takayama, M. L. Shuler, G. Vunjak-Novakovic, O. Frey, E. Verpoorte and Y. C. Toh, Nat. Rev. Methods Primers, 2022, 2, 33.
117. Y. Hou, Y. Zheng, X. Zheng, Y. Sun, X. Yi, Z. Wu and J. M. Lin, Lab Chip, 2023, 23, 2654–2663.
118. N. Ideno, H. Yamaguchi, T. Okumura, J. Huang, M. J. Brun, M. L. Ho, J. Suh, S. Gupta, A. Maitra and B. Ghosh, Lab. Invest., 2019, 99, 1233–1244.
119. R. Lakhia, H. Ramalingam, C.-M. Chang, P. Cobo-Stark, L. Biggers, A. Flaten, J. Alvarez, T. Valencia, D. P. Wallace, E. C. Lee and V. Patel, Nat. Commun., 2022, 13, 4765.
120. O. Elabi, A. Gaceb, R. Carlsson, T. Padel, R. Soylu-Kucharz, I. Cortijo, W. Li, J.-Y. Li and G. Paul, Sci. Rep., 2021, 11, 1120.
121. S. Grebenyuk, A. R. Abdel Fattah, M. Kumar, B. Toprakhisar, G. Rustandi, A. Vananroye, I. Salmon, C. Verfaillie, M. Grillo and A. Ranga, Nat. Commun., 2023, 14, 193.
122. N. Zhao, Y. Song, X. Xie, Z. Zhu, C. Duan, C. Nong, H. Wang and R. Bao, Signal Transduction Targeted Ther., 2023, 8, 112.
123. D. V. Foss, J. J. Muldoon, D. N. Nguyen, D. Carr, S. U. Sahu, J. M. Hunsinger, S. K. Wyman, N. Krishnappa, R. Mendonsa, E. V. Schanzer, B. R. Shy, V. S. Vykunta, V. Allain, Z. Li, A. Marson, J. Eyquem and R. C. Wilson, Nat. Biomed. Eng., 2023, 7, 647–660.
124. G. M. Allen, N. W. Frankel, N. R. Reddy, H. K. Bhargava, M. A. Yoshida, S. R. Stark, M. Purl, J. Lee, J. L. Yee, W. Yu, A. W. Li, K. C. Garcia, H. El-Samad, K. T. Roybal, M. H. Spitzer and W. A. Lim, Science, 2022, 378, 1–25.
125. A. K. Fajrial, Q. Q. He, N. I. Wirusanti, J. E. Slansky and X. Ding, Theranostics, 2020, 10, 5532–5549.
126. R. C. Sterner and R. M. Sterner, Blood Cancer J., 2021, 11, 69.
127. A. J. Kansagra, N. V. Frey, M. Bar, T. W. Laetsch, P. A. Carpenter, B. N. Savani, H. E. Heslop, C. M. Bollard, K. V. Komanduri, D. A. Gastineau, C. Chabannon, M. A. Perales, M. Hudecek, M. Aljurf, L. Andritsos, J. A. Barrett, V. Bachanova, C. Bonini, A. Ghobadi, S. I. Gill, J. A. Hill, S. Kenderian, P. Kebriaei, A. Nagler, D. Maloney, H. D. Liu, N. N. Shah, M. A. Kharfan-Dabaja, E. J. Shpall, G. J. Mufti, L. Johnston, E. Jacoby, A. Bazarbachi, J. F. DiPersio, S. Z. Pavletic, D. L. Porter, S. A. Grupp, M. Sadelain, M. R. Litzow, M. Mohty and S. K. Hashmi, Bone Marrow Transplant., 2019, 54, 1868–1880.
128. Z. Zhang, S. Qiu, X. Zhang and W. Chen, BMC Biotechnol., 2018, 18, 1–9.
129. X. Yu, W. Ruan, F. Lin, W. Qian, Y. Zou, Y. Liu, R. Su, Q. Niu, Q. Ruan, W. Lin, Z. Zhu, H. Zhang and C. Yang, Proc. Natl. Acad. Sci., 2023, 120, e2221934120.
130. H. Sinha, A. B. V. Quach, P. Q. N. Vo and S. C. C. Shih, Lab Chip, 2018, 18, 2300–2312.
131. G. Ronteix, S. Jain, C. Angely, M. Cazaux, R. Khazen, P. Bousso and C. N. Baroud, Nat. Commun., 2022, 13, 25–28.
132. L. Mathur, B. Szalai, N. H. Du, R. Utharala, M. Ballinger, J. J. M. Landry, M. Ryckelynck, V. Benes, J. Saez-Rodriguez and C. A. Merten, Nat. Commun., 2022, 13, 1–15.
133. B. El Debs, R. Utharala, I. V. Balyasnikova, A. D. Griffiths and C. A. Merten, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 11570–11575.
134. T. Kamath, A. Abdulraouf, S. J. Burris, J. Langlieb, V. Gazestani, N. M. Nadaf, K. Balderrama, C. Vanderburg and E. Z. Macosko, Nat. Neurosci., 2022, 25, 588–595.
135. W. M. Zhou, Y. Y. Yan, Q. R. Guo, H. Ji, H. Wang, T. T. Xu, B. Makabel, C. Pilarsky, G. He, X. Y. Yu and J. Y. Zhang, J. Nanobiotechnol., 2021, 19, 1–21.
136. S. Damiati, R. Mhanna, R. Kodzius and E.-K. Ehmoser, Genes, 2018, 9, 144.
137. A. M. Jackson, J. Boutell, N. Cooley and M. He, Briefings Funct. Genomics Proteomics, 2004, 2, 308–319.
138. A. Tinafar, K. Jaenes and K. Pardee, BMC Biol., 2019, 17, 1–14.
139. R. W. Martin, B. J. Des Soye, Y. C. Kwon, J. Kay, R. G. Davis, P. M. Thomas, N. I. Majewska, C. X. Chen, R. D. Marcum, M. G. Weiss, A. E. Stoddart, M. Amiram, A. K. Ranji Charna, J. R. Patel, F. J. Isaacs, N. L. Kelleher, S. H. Hong and M. C. Jewett, Nat. Commun., 2018, 9, 1–9.
140. B. C. Bundy, J. P. Hunt, M. C. Jewett, J. R. Swartz, D. W. Wood, D. D. Frey and G. Rao, Curr. Opin. Chem. Eng., 2018, 22, 177–183.
141. K. Pardee, S. Slomovic, P. Q. Nguyen, J. W. Lee, N. Donghia, D. Burrill, T. Ferrante, F. R. McSorley, Y. Furuta, A. Vernet, M. Lewandowski, C. N. Boddy, N. S. Joshi and J. J. Collins, Cell, 2016, 167, 248–259.e12.
142. G. Kanter, J. Yang, A. Voloshin, S. Levy, J. R. Swartz and R. Levy, Blood, 2007, 109, 3393–3399.
143. A. K. Aquino, Z. A. Manzer, S. Daniel and M. P. DeLisa, Front. Mol. Biosci., 2021, 8 DOI:10.3389/fmolb.2021.782905.
144. B. Lavickova, L. Grasemann and S. J. Maerkl, ACS Synth. Biol., 2022, 11, 4134–4141.
145. M. W. Bartelds, Ó. García-Blay, P. G. A. Verhagen, E. J. Wubbolts, B. van Sluijs, H. A. Heus, T. F. A. de Greef, W. T. S. Huck and M. M. K. Hansen, ACS Synth. Biol., 2023, 12, 2217–2225.
146. R. Gan, M. D. Cabezas, M. Pan, H. Zhang, G. Hu, L. G. Clark, M. C. Jewett and R. Nicol, ACS Synth. Biol., 2022, 11, 2108–2120.
147. M. Weiss, J. P. Frohnmayer, L. T. Benk, B. Haller, J. W. Janiesch, T. Heitkamp, M. Börsch, R. B. Lira, R. Dimova, R. Lipowsky, E. Bodenschatz, J. C. Baret, T. Vidakovic-Koch, K. Sundmacher, I. Platzman and J. P. Spatz, Nat. Mater., 2018, 17, 89–95.
148. T. Narahari, J. Dahmer, A. Sklavounos, T. Kim, M. Satkauskas, I. Clotea, M. Ho, J. Lamanna, C. Dixon, D. G. Rackus, S. J. R. da Silva, L. Pena, K. Pardee and A. R. Wheeler, Lab Chip, 2022, 22, 1748–1763.
149. A. S. Selim, J. M. Perry, M. A. Nasr, J. M. Pimprikar and S. C. C. Shih, ACS Appl. Bio Mater., 2022, 5, 5487–5496.
150. Y. Zhang, J. Zhang, H. Hara, I. Kato and M. Inouye, J. Biol. Chem., 2005, 280, 3143–3150.
151. D. T. Gonzales, N. Yandrapalli, T. Robinson, C. Zechner and T. Y. D. Tang, ACS Synth. Biol., 2022, 11, 205–215.
152. J. Shin, P. Jardine and V. Noireaux, ACS Synth. Biol., 2012, 1, 408–413.
153. S. J. Moore, J. T. MacDonald, S. Wienecke, A. Ishwarbhai, A. Tsipa, R. Aw, N. Kylilis, D. J. Bell, D. W. McClymont, K. Jensen, K. M. Polizzi, R. Biedendieck and P. S. Freemont, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 4340–4349.
154. J. Garamella, R. Marshall, M. Rustad and V. Noireaux, ACS Synth. Biol., 2016, 5, 344–355.
155. F. Greiss, S. S. Daube, V. Noireaux and R. Bar-Ziv, Nat. Commun., 2020, 11, 1–9.
156. D. Foshag, E. Henrich, E. Hiller, M. Schäfer, C. Kerger, A. Burger-Kentischer, I. Diaz-Moreno, S. M. García-Mauriño, V. Dötsch, S. Rupp and F. Bernhard, New Biotechnol., 2018, 40, 245–260.
157. R. Marshall, C. S. Maxwell, S. P. Collins, T. Jacobsen, M. L. Luo, M. B. Begemann, B. N. Gray, E. January, A. Singer, Y. He, C. L. Beisel and V. Noireaux, Mol. Cell, 2018, 69, 146–157.e3.
158. M. Karlikow, S. J. R. da Silva, Y. Guo, S. Cicek, L. Krokovsky, P. Homme, Y. Xiong, T. Xu, M. A. Calderón-Peláez, S. Camacho-Ortega, D. Ma, J. J. F. de Magalhães, B. N. R. F. Souza, D. G. de Albuquerque Cabral, K. Jaenes, P. Sutyrina, T. Ferrante, A. D. Benitez, V. Nipaz, P. Ponce, D. G. Rackus, J. J. Collins, M. Paiva, J. E. Castellanos, V. Cevallos, A. A. Green, C. Ayres, L. Pena and K. Pardee, Nat. Biomed. Eng., 2022, 6, 246–256.
159. J. Bedford, D. Enria, J. Giesecke, D. L. Heymann, C. Ihekweazu, G. Kobinger, H. C. Lane, Z. Memish, M. D. Oh, A. A. Sall, A. Schuchat, K. Ungchusak and L. H. Wieler, Lancet, 2020, 395, 1015–1018.
160. L. Wunderlichová, L. Buňková, M. Koutný, P. Jančová and F. Buňka, Compr. Rev. Food Sci. Food Saf., 2014, 13, 1012–1030.
161. F. Ahmadi, A. B. V. Quach and S. C. C. Shih, Biomicrofluidics, 2020, 14, 61301.
162. S. Dekker, W. Buesink, M. Blom, M. Alessio, N. Verplanck, M. Hihoud, C. Dehan, W. César, A. Le Nel, A. van den Berg and M. Odijk, Sens. Actuators, B, 2018, 272, 468–478.
163. A. Isozaki, D. Huang, Y. Nakagawa and K. Goda, Microfluid. Nanofluid., 2021, 25, 1–10.
164. A. Isozaki, A. Isozaki, Y. Nakagawa, M. H. Loo, Y. Shibata, N. Tanaka, D. L. Setyaningrum, J. W. Park, Y. Shirasaki, H. Mikami, D. Huang, H. Tsoi, C. T. Riche, T. Ota, H. Miwa, Y. Kanda, T. Ito, T. Ito, K. Yamada, O. Iwata, K. Suzuki, S. Ohnuki, Y. Ohya, Y. Kato, T. Hasunuma, T. Hasunuma, S. Matsusaka, M. Yamagishi, M. Yazawa, S. Uemura, K. Nagasawa, H. Watarai, H. Watarai, D. Di Carlo, K. Goda, K. Goda, K. Goda and K. Goda, Sci. Adv., 2020, 6, 1–12.
165. O. Caen, S. Schütz, M. S. S. Jammalamadaka, J. Vrignon, P. Nizard, T. M. Schneider, J.-C. Baret and V. Taly, Microsyst. Nanoeng., 2018, 4, 33.
166. H. F. Tsai, S. Podder and P. Y. Chen, Micromachines, 2023, 14, 1–30.

---