Much attention has focused on understanding microbial interactions leading to stable co-cultures. In this work, substrate pulsing was performed to promote better control of the metabolic niches (MNs) corresponding to each species, leading to the continuous co-cultivation of diverse microbial organisms. We used a cell-machine interface, which allows adjustment of the temporal profile of two MNs according to a rhythm, ensuring the successive growth of two species, in our case, a yeast and a bacterium. The resulting approach, called ‘automated adjustment of metabolic niches’ (AAMN), was effective for stabilizing both cooperative and competitive co-cultures. AAMN can be considered an enabling technology for the deployment of co-cultures in bioprocesses, demonstrated here based on the continuous bioproduction of p-coumaric acid. The data accumulated suggest that AAMN could be used not only for a wider range of biological systems, but also to gain fundamental insights into microbial interaction mechanisms.

Keywords: automated flow cytometry; continuous cultivation; cross-feeding; diauxic shift; microbial interactions; population control; yeast-bacteria co-cultures.

Copyright © 2024 Elsevier Ltd. All rights reserved.

PubMed Disclaimer

Cell-to-cell heterogeneity presents challenges across various fields, from biomedicine to bioproduction, where precise cellular responses are vital. While single cell technologies have significantly enhanced our understanding of population heterogeneity, the predominant focus has been on monitoring intracellular compounds. Recognizing the added complexity introduced by the secretion system, in this review, we first provide a systematic overview of the distinct steps necessary for driving protein secretion. We discuss the various sources of noise acting from the synthesized preprotein to the secretory protein released based on a Gram-positive cellular system as a model. We next explore the applicability of single cell technologies for monitoring protein secretion throughout these functional stages. We also emphasize the importance of applying these single cell technologies for monitoring protein secretion during bioproduction.

Giving the cells exactly what they need, when they need it” is the core idea behind the proposed bioprocess control strategy: operating bioprocess based on the physiological behavior of the microbial population rather than exclusive monitoring of environmental parameters. We are envisioning to achieve this through the use of genetically encoded biosensors combined with online flow cytometry (FCM) to obtain a time-dependent “physiological fingerprint” of the population. We developed a biosensor based on the glnA promoter (glnAp) and applied it for monitoring the nitrogen-related nutritional state of Escherichia coli. The functionality of the biosensor was demonstrated through multiple cultivation runs performed at various scales—from microplate to 20 L bioreactor. We also developed a fully automated bioreactor—FCM interface for on-line monitoring of the microbial population. Finally, we validated the proposed strategy by performing a fed-batch experiment where the biosensor signal is used as the actuator for a nitrogen feeding feedback control. This new generation of process control, —based on the specific needs of the cells, —opens the possibility of improving process development on a short timescale and therewith, the robustness and performance of fermentation processes.

Microbial populations undergo phenotypic switching as a response to environmental perturbations. For instance, some bacteria switch from a planktonic lifestyle to biofilm, resulting in altered physiological properties such as increased robustness depending on the conditions. However, the precise detection of phenotypic switching events during the bacterial life cycle is still a technical challenge. Propidium iodide (PI) is one of the most frequently used fluorescence indicators for assessing cell viability based on membrane permeability, yet PI-stained cells sometimes display a red-but-not-dead phenotype. In Escherichia coli, this phenomenon is connected to modulation of porins in the outer membrane (OM) to adapt OM permeability according to nutrient availability. In this study, we explored PI staining to assess phenotypic changes in Pseudomonas sp. during biofilm development. We show that this switch is linked to excretion of extracellular DNA (eDNA), rather than modification of OM permeability. Confocal laser scanning microscopy (CLSM) enabled direct visualization of red fluorescent clusters outside intact membranes of viable cells, suggesting that PI binds eDNA. Besides, the occurrence of PI-positive sub-populations was correlated with biofilm formation in the model bacterium Pseudomonas putida KT2440. Engineered derivatives thereof with altered biofilm-forming capabilities exposed a whole continuum of phenotypic states involving planktonic cells and aggregates, and were identified according to the dynamic change of PI-positive cells with flow cytometry analysis. Our results demonstrate that PI is a fast, convenient and versatile staining for eDNA to rapidly monitor the phenotypic switching of Pseudomonas sp. during transitions in the bacterial lifestyle.

Isogenic cell populations can cope with stress conditions by switching to alternative phenotypes. Even if it can lead to increased fitness in a natural context, this feature is typically unwanted for a range of applications (e.g., bioproduction, synthetic biology, and biomedicine) where it tends to make cellular response unpredictable. However, little is known about the diversification profiles that can be adopted by a cell population. Here, we characterize the diversification dynamics for various systems (bacteria and yeast) and for different phenotypes (utilization of alternative carbon sources, general stress response and more complex development patterns). Our results suggest that the diversification dynamics and the fitness cost associated with cell switching are coupled. To quantify the contribution of the switching cost on population dynamics, we design a stochastic model that let us reproduce the dynamics observed experimentally and identify three diversification regimes, i.e., constrained (at low switching cost), dispersed (at medium and high switching cost), and bursty (for very high switching cost). Furthermore, we use a cell-machine interface called Segregostat to demonstrate that different levels of control can be applied to these diversification regimes, enabling applications involving more precise cellular responses.

Different expression vectors are available for the effective production of recombinant proteins by bacterial populations. However, the productivity of such systems is limited by the inherent noise of the gene circuits used for the synthesis of recombinant products. An extreme case of cell-to-cell heterogeneity that has been previously reported for the ara- and lac-based expression systems in E. coli is the all-or-none response. According to this mode of response, two subpopulations of cells are generated, i.e., a “low-” subpopulation exhibiting a shallow expression level and a “high-” subpopulation exhibiting a high-expression level. The “low-” subpopulation can be considered as a cluster of non-producing cells contributing to the loss of productivity. Here we describe the setup, design, and operation of a continuous culture where inducer addition is operated based on microbial population dynamics. The determination of population dynamics is done based on an automated flow cytometry (FC) procedure previously denoted as segregostat. We illustrate how this setup can be used to control the activation of an ara-based expression system and avoid phenotypic diversification leading to an all-or-none response. Upon the determination of the natural frequency of the gene circuit used as an expression system, our current protocol can be set up without the requirement of a feedback controller.

Microbial consortia are an exciting alternative for increasing the performances of bioprocesses for the production of complex metabolic products. However, the functional properties of microbial communities remain challenging to control, considering the complex interaction mechanisms occurring between co-cultured microbial species. Indeed, microbial communities are highly dynamic and can adapt to changing environmental conditions through complex mechanisms, such as phenotypic diversification. We focused on stabilizing a co-culture of Saccharomyces cerevisiae and Escherichia coli in continuous cultures. Our preliminary data pointed out that transient diauxic shifts could lead to stable co-culture by providing periodic fitness advantages to the yeast. Based on a computational toolbox called MONCKS (for MONod-type Co-culture Kinetic Simulation), we were able to predict the dynamics of diauxic shift for both species based on a cybernetic approach. This toolbox was further used to predict the frequency of diauxic shift to be applied to reach co-culture stability. These simulations were successfully reproduced experimentally in continuous bioreactors with glucose pulsing. Finally, based on a bet-hedging reporter, we observed that the yeast population exhibited an increased phenotypic diversification process in co-culture compared with mono-culture, suggesting that this mechanism could be the basis of the metabolic fitness of the yeast.

Predicting the fate of individual cells among a microbial population (i.e., growth and gene expression) remains a challenge, especially when this population is exposed to very dynamic environmental conditions, such as those encountered during continuous cultivation. Indeed, the dynamic nature of a continuous cultivation process implies the potential diversification of the microbial population resulting in genotypic and phenotypic heterogeneity. The present work focused on the induction of the arabinose operon in Escherichia coli as a model system to study this diversification process in continuous cultivations. As a preliminary step, the green fluorescent protein (GFP) level triggered by an arabinose-inducible ParaBAD promoter was tracked by flow cytometry in chemostat cultivations with glucose-arabinose co-feeding. For a wide range of glucose-arabinose co-feeding concentrations in the chemostats, the simultaneous occurrence of GFP positive and negative subpopulation was observed. In the second set of experiments, continuous cultivation was performed by adding glucose continuously and arabinose based on the capability of individual cells to switch from low GFP to high GFP expression states, performed with a technology setup called segregostat. In the segregostat cultivation mode, on-line flow cytometry analysis was used for adjusting the arabinose/glucose transitions based on the phenotypic switching profiles of the microbial population. This strategy allowed finding an appropriate arabinose pulsing frequency, leading to prolonged maintenance of the induction level with a limited increase in the phenotypic diversity for more than 60 generations. The results suggest that the steady forcing of individual cells into a given phenotypic trajectory may not be the best strategy for controlling cell populations. Instead, allowing individual cells to switch periodically around a predefined threshold seems to be a more robust strategy leading to oscillations, but within a predictable cell population behavior range.

Microbial cells experience physiological changes due to environmental change, such as pH and temperature, the release of bactericidal agents, or nutrient limitation. This has been shown to affect community assembly and physiological processes (e.g., stress tolerance, virulence, or cellular metabolic activity). Metabolic stress is typically quantified by measuring community phenotypic properties such as biomass growth, reactive oxygen species, or cell permeability. However, bulk community measurements do not take into account single-cell phenotypic diversity, which is important for a better understanding and the subsequent management of microbial populations. Raman spectroscopy is a nondestructive alternative that provides detailed information on the biochemical makeup of each individual cell. Here, we introduce a method for describing single-cell phenotypic diversity using the Hill diversity framework of Raman spectra. Using the biomolecular profile of individual cells, we obtained a metric to compare cellular states and used it to study stress-induced changes. First, in two Escherichia coli populations either treated with ethanol or nontreated and then in two Saccharomyces cerevisiae subpopulations with either high or low expression of a stress reporter. In both cases, we were able to quantify single-cell phenotypic diversity and to discriminate metabolically stressed cells using a clustering algorithm. We also described how the lipid, protein, and nucleic acid compositions changed after the exposure to the stressor using information from the Raman spectra. Our results show that Raman spectroscopy delivers the necessary resolution to quantify phenotypic diversity within individual cells and that this information can be used to study stress-driven metabolic diversity in microbial populations.

IMPORTANCE Microbial cells that live in the same community can exist in different physiological and morphological states that change as a function of spatiotemporal variations in environmental conditions. This phenomenon is commonly known as phenotypic heterogeneity and/or diversity. Measuring this plethora of cellular expressions is needed to better understand and manage microbial processes. However, most tools to study phenotypic diversity only average the behavior of the sampled community. In this work, we present a way to quantify the phenotypic diversity of microbial samples by inferring the (bio)molecular profile of its constituent cells using Raman spectroscopy. We demonstrate how this tool can be used to quantify the phenotypic diversity that arises after the exposure of microbes to stress. Raman spectroscopy holds potential for the detection of stressed cells in bioproduction.

Controlling and managing the degree of phenotypic diversification of microbial populations is a challenging task. This task not only requires detailed knowledge regarding diversification mechanisms but also advanced technical set-ups for the real-time analyses and control of population behaviour on single-cell level. In this work, set-up, design and operation of the so called segregostat are described which, in contrast to a traditional chemostat, allows the control of phenotypic diversification of microbial populations over time. Two exemplary case studies will be discussed, i.e. phenotypic diversification dynamics of Eschericia coli and Pseudomonas putida based on outer membrane permeabilization, emphasizing the applicability and versatility of the proposed approach. Upon nutrient limitation, cell population tends to diversify into several subpopulations exhibiting distinct phenotypic features (non-permeabilized and permeabilized cells). Online analysis leads to the determination of the ratio between cells in these two states, which in turn triggers the addition of glucose pulses in order to maintain a predefined diversification ratio. These results prove that phenotypic diversification can be controlled by means of defined pulse-frequency modulation within continuously running bioreactor set-ups. This lays the foundation for systematic studies, not only of phenotypic diversification but also for all processes where dynamics single-cell approaches are required, such as synthetic co-culture processes.

Efficient optimization of microbial processes is a critical issue for achieving a number of sustainable development goals, considering the impact of microbial biotechnology in agrofood, environment, biopharmaceutical and chemical industries. Many of these applications require scale-up after proof of concept. However, the behaviour of microbial systems remains unpredictable (at least partially) when shifting from laboratory-scale to industrial conditions. The need for robust microbial systems is thus highly needed in this context, as well as a better understanding of the interactions between fluid mechanics and cell physiology. For that purpose, a full scale-up/down computational framework is already available. This framework links computational fluid dynamics (CFD), metabolic flux analysis and agent-based modelling (ABM) for a better understanding of the cell lifelines in a heterogeneous environment. Ultimately, this framework can be used for the design of scale-down simulators and/or metabolically engineered cells able to cope with environmental fluctuations typically found in large-scale bioreactors. However, this framework still needs some refinements, such as a better integration of gas–liquid flows in CFD, and taking into account intrinsic biological noise in ABM.

In natural habitats, microbes form multispecies communities that commonly face rapidly changing and highly competitive environments. Thus, phenotypic heterogeneity has evolved as an innate and important survival strategy to gain an overall fitness advantage over cohabiting competitors. However, in defined artificial environments such as monocultures in small- to large-scale bioreactors, cell-to-cell variations are presumed to cause reduced production yields as well as process instability. Hence, engineering microbial production toward phenotypic homogeneity is a highly promising approach for synthetic biology and bioprocess optimization.

In this review, we discuss recent studies that have unraveled the cell-to-cell heterogeneity observed during bacterial gene expression and metabolite production as well as the molecular mechanisms involved. In addition, current single-cell technologies are briefly reviewed with respect to their applicability in exploring cell-to-cell variations. We highlight emerging strategies and tools to reduce phenotypic heterogeneity in biotechnological expression setups. Here, strain or inducer modifications are combined with cell physiology manipulations to achieve the ultimate goal of equalizing bacterial populations. In this way, the majority of cells can be forced into high productivity, thus reducing less productive subpopulations that tend to consume valuable resources during production. Modifications in uptake systems, inducer molecules or nutrients represent valuable tools for diminishing heterogeneity.

Finally, we address the challenge of transferring homogeneously responding cells into large-scale bioprocesses. Environmental heterogeneity originating from extrinsic factors such as stirring speed and pH, oxygen, temperature or nutrient distribution can significantly influence cellular physiology. We conclude that engineering microbial populations toward phenotypic homogeneity is an increasingly important task to take biotechnological productions to the next level of control.

Noise in gene and protein expression is a major cause for bioprocess deviation. However, this phenomenon has been only scarcely considered in real bioprocessing conditions. In this work, a scaling-law derived from genome-scale studies based on GFP reporter systems has been calibrated to an on-line flow cytometry device, allowing thus to get an insight at the level of promoter activity and associated noise during a whole microbial culture carried out in bioreactor. We show that most of the GFP reporter systems investigated and thus corresponding genes could be included inside the area covered by the scaling-law. The experimental results suggest that this scaling-law could be used to predict the dynamics of promoter activity, as well as the associated noise, in bioprocessing conditions. The knowledge acquired throughout this work could be used for the design of more robust expression systems.

Phenotypic heterogeneity is a major issue in the context of industrial bioprocessing. Stochasticity of gene expression is usually considered to be the main source of heterogeneity among microbial population, but recent evidence demonstrates that metabolic reactions can also be subject to stochasticity without any intervention of gene expression. Although metabolic heterogeneity can be encountered in laboratory-scale cultivation devices, stochasticity at the level of metabolic reactions is perturbed directly by microenvironmental heterogeneities occurring in large-scale bioreactors. Accordingly, analytical tools are needed for the determination of metabolic variability in bioprocessing conditions and for the efficient design of metabolic engineering strategies. In this context, implementation of single cell technologies for bioprocess monitoring would benefit from knowledge acquired in more fundamental studies.