Noise in bacterial gene expression

Drug tolerance and persistence in bacteria, fungi and cancer cells: Role of non-genetic heterogeneity

A common feature of bacterial, fungal and cancer cell populations upon treatment is the presence of tolerant and persistent cells able to survive, and sometimes grow, even in the presence of usually inhibitory or lethal drug concentrations, driven by non-genetic differences among individual cells in a population. Here we review and compare data obtained on drug survival in bacteria, fungi and cancer cells to unravel common characteristics and cellular pathways, and to point their singularities. This comparative work also allows to cross-fertilize ideas across fields. We particularly focus on the role of gene expression variability in the emergence of cell-cell non-genetic heterogeneity because it represents a possible common basic molecular process at the origin of most persistence phenomena and could be monitored and tuned to help improve therapeutic interventions.

Gram-positive bacteria are primed for surviving lethal doses of antibiotics and chemical stress

Antibiotic resistance kills millions worldwide yearly. However, a major contributor to recurrent infections lies in a small fraction of bacterial cells, known as persisters. These cells are not inherently antibiotic-resistant, yet they lead to increased antibiotic usage, raising the risk of developing resistant progenies. In a bacterial population, individual cells exhibit considerable fluctuations in their gene expression levels despite being cultivated under identical, stable conditions. This variability in cell-to-cell characteristics (phenotypic diversity) within an isogenic population enables persister cells to withstand antibiotic exposure by entering a non-dividing state. We recently showed the existence of "primed cells" in E. coli. Primed cells are dividing cells prepared for antibiotic stress before encountering it and are more prone to form persisters. They also pass their "prepared state" down for several generations through epigenetic memory. Here, we show that primed cells are common among distant bacterial lineages, allowing for survival against antibiotics and other chemical stress, and form in different growth phases. They are also responsible for increased persister levels in transition and stationary phases compared to the log phase. We tested and showed that the Gram-positive bacterium Bacillus megaterium, evolutionarily very distant from E. coli, forms primed cells and has a transient epigenetic memory that is maintained for 7 generations or more. We showed this using ciprofloxacin and the non-antibiotic chemical stress fluoride. It is well established that persister levels are higher in the stationary phase than in the log phase, and B. megaterium persisters levels are nearly identical from the early to late-log phase but are ~2-fold and ~4-fold higher in the transition and stationary phase, respectively. It was previously proposed that there are two distinct types of persisters: Type II forms in the log phase, while Type I forms in the stationary phase. However, we show that primed cells lead to increased persisters in the transition and stationary phase and found no evidence of Type I or II persisters with distant phenotypes. Overall, we have provided substantial evidence of the importance of primed cells and their transitory epigenetic memories to surviving stress.

ProTInSeq: transposon insertion tracking by ultra-deep DNA sequencing to identify translated large and small ORFs

Identifying open reading frames (ORFs) being translated is not a trivial task. ProTInSeq is a technique designed to characterize proteomes by sequencing transposon insertions engineered to express a selection marker when they occur in-frame within a protein-coding gene. In the bacterium Mycoplasma pneumoniae, ProTInSeq identifies 83% of its annotated proteins, along with 5 proteins and 153 small ORF-encoded proteins (SEPs; ≤100 aa) that were not previously annotated. Moreover, ProTInSeq can be utilized for detecting translational noise, as well as for relative quantification and transmembrane topology estimation of fitness and non-essential proteins. By integrating various identification approaches, the number of initially annotated SEPs in this bacterium increases from 27 to 329, with a quarter of them predicted to possess antimicrobial potential. Herein, we describe a methodology complementary to Ribo-Seq and mass spectroscopy that can identify SEPs while providing other insights in a proteome with a flexible and cost-effective DNA ultra-deep sequencing approach.

Transcriptional heterogeneity of catabolic genes on the plasmid pCAR1 causes host-specific carbazole degradation

To elucidate why plasmid-borne catabolic ability differs among host bacteria, we assessed the expression dynamics of the Pant promoter on the carbazole-degradative conjugative plasmid pCAR1 in Pseudomonas putida KT2440(pCAR1) (hereafter, KTPC) and Pseudomonas resinovorans CA10. The Pant promoter regulates the transcription of both the car and ant operons, which are responsible for converting carbazole into anthranilate and anthranilate into catechol, respectively. In the presence of anthranilate, transcription of the Pant promoter is induced by the AraC/XylS family regulator AntR, encoded on pCAR1. A reporter cassette containing the Pant promoter followed by gfp was inserted into the chromosomes of KTPC and CA10. After adding anthranilate, GFP expression in the population of CA10 showed an unimodal distribution, whereas a small population with low GFP fluorescence intensity appeared for KTPC. CA10 has a gene, antRCA, that encodes an iso-functional homolog of AntR on its chromosome. When antRCA was disrupted, a small population with low GFP fluorescence intensity appeared. In contrast, overexpression of pCAR1-encoded AntR in KTPC resulted in unimodal expression under the Pant promoter. These results suggest that the expression of pCAR1-encoded AntR is insufficient to ameliorate the stochastic expression of the Pant promoter. Raman spectra of single cells collected using deuterium-labeled carbazole showed that the C-D Raman signal exhibited greater variability for KTPC than CA10. These results indicate that heterogeneity at the transcriptional level of the Pant promoter due to insufficient AntR availability causes fluctuations in the pCAR1-borne carbazole-degrading capacity of host bacterial cells.IMPORTANCEHorizontally acquired genes increase the competitiveness of host bacteria under selective conditions, although unregulated expression of foreign genes may impose fitness costs. The "appropriate" host for a plasmid is empirically known to maximize the expression of plasmid-borne traits. In the case of pCAR1-harboring Pseudomonas strains, P. resinovorans CA10 exhibits strong carbazole-degrading capacity, whereas P. putida KT2440 harboring pCAR1 exhibits low degradation capacity. Our results suggest that a chromosomally encoded transcription factor affects transcriptional and metabolic fluctuations in host cells, resulting in different carbazole-degrading capacities as a population. This study may provide a clue for determining appropriate hosts for a plasmid and for regulating the expression of plasmid-borne traits, such as the degradation of xenobiotics and antibiotic resistance.

Escherichia coli cells are primed for survival before lethal antibiotic stress

Non-genetic factors can cause significant fluctuations in gene expression levels. Regardless of growing in a stable environment, this fluctuation leads to cell-to-cell variability in an isogenic population. This phenotypic heterogeneity allows a tiny subset of bacterial cells in a population called persister cells to tolerate long-term lethal antibiotic effects by entering into a non-dividing, metabolically repressed state. We occasionally noticed a high variation in persister levels, and to explore this, we tested clonal populations starting from a single cell using a modified Luria-Delbrück fluctuation test. Although we kept the conditions same, the diversity in persistence level among clones was relatively consistent: varying from ~60- to 100- and ~40- to 70-fold for ampicillin and apramycin, respectively. Then, we divided and diluted each clone to observe whether the same clone had comparable persister levels for more than one generation. Replicates had similar persister levels even when clones were divided, diluted by 1:20, and allowed to grow for approximately five generations. This result explicitly shows a cellular memory passed on for generations and eventually lost when cells are diluted to 1:100 and regrown (>seven generations). Our result demonstrates (1) the existence of a small population prepared for stress ("primed cells") resulting in higher persister numbers; (2) the primed memory state is reproducible and transient, passed down for generations but eventually lost; and (3) a heterogeneous persister population is a result of a transiently primed reversible cell state and not due to a pre-existing genetic mutation. IMPORTANCE Antibiotics have been highly effective in treating lethal infectious diseases for almost a century. However, the increasing threat of antibiotic resistance is again causing these diseases to become life-threatening. The longer a bacteria can survive antibiotics, the more likely it is to develop resistance. Complicating matters is that non-genetic factors can allow bacterial cells with identical DNA to gain transient resistance (also known as persistence). Here, we show that a small fraction of the bacterial population called primed cells can pass down non-genetic information ("memory") to their offspring, enabling them to survive lethal antibiotics for a long time. However, this memory is eventually lost. These results demonstrate how bacteria can leverage differences among genetically identical cells formed through non-genetic factors to form primed cells with a selective advantage to survive antibiotics.

Intrinsically disordered proteins and conformational noise: The hypothesis a decade later

Phenotypic plasticity is the ability of individual genotypes to produce different phenotypes in response to environmental perturbations. We previously postulated how conformational noise emanating from conformational dynamics of intrinsically disordered proteins (IDPs) which is distinct from transcriptional noise, can contribute to phenotypic switching by rewiring the cellular protein interaction network. Since most transcription factors are IDPs, we posited that conformational noise is an integral component of transcriptional noise implying that IDPs may amplify total noise in the system either stochastically or in response to environmental changes. Here, we review progress in elucidating the details of the hypothesis. We highlight empirical evidence supporting the hypothesis, discuss conceptual advances that underscore its fundamental importance and implications, and identify areas for future investigations.

Making use of noise in biological systems

Disorder and noise are inherent in biological systems. They are required to provide systems with the advantages required for proper functioning. Noise is a part of the flexibility and plasticity of biological systems. It provides systems with increased routes, improves information transfer, and assists in response triggers. This paper reviews recent studies on noise at the genome, cellular, and whole organ levels. We focus on the need to use noise in system engineering. We present some of the challenges faced in studying noise. Optimizing the efficiency of complex systems requires a degree of variability in their functions within certain limits. Constrained noise can be considered a method for improving system robustness by regulating noise levels in continuously dynamic settings. The digital pill-based artificial intelligence (AI)-based platform is the first to implement second-generation AI comprising variability-based signatures. This platform enhances the efficacy of the therapeutic regimens. Systems requiring variability and mechanisms regulating noise are mandatory for understanding biological functions.

Molecular Origins of Transcriptional Heterogeneity in Diazotrophic Klebsiella oxytoca

Phenotypic heterogeneity in clonal bacterial batch cultures has been shown for a range of bacterial systems; however, the molecular origins of such heterogeneity and its magnitude are not well understood. Under conditions of extreme low-nitrogen stress in the model diazotroph Klebsiella oxytoca, we found remarkably high heterogeneity of nifHDK gene expression, which codes for the structural genes of nitrogenase, one key enzyme of the global nitrogen cycle. This heterogeneity limited the bulk observed nitrogen-fixing capacity of the population. Using dual-probe, single-cell RNA fluorescent in situ hybridization, we correlated nifHDK expression with that of nifLA and glnK-amtB, which code for the main upstream regulatory components. Through stochastic transcription models and mutual information analysis, we revealed likely molecular origins for heterogeneity in nitrogenase expression. In the wild type and regulatory variants, we found that nifHDK transcription was inherently bursty, but we established that noise propagation through signaling was also significant. The regulatory gene glnK had the highest discernible effect on nifHDK variance, while noise from factors outside the regulatory pathway were negligible. Understanding the basis of inherent heterogeneity of nitrogenase expression and its origins can inform biotechnology strategies seeking to enhance biological nitrogen fixation. Finally, we speculate on potential benefits of diazotrophic heterogeneity in natural soil environments. IMPORTANCE Nitrogen is an essential micronutrient for both plant and animal life and naturally exists in both reactive and inert chemical forms. Modern agriculture is heavily reliant on nitrogen that has been "fixed" into a reactive form via the energetically expensive Haber-Bosch process, with significant environmental consequences. Nitrogen-fixing bacteria provide an alternative source of fixed nitrogen for use in both biotechnological and agricultural settings, but this relies on a firm understanding of how the fixation process is regulated within individual bacterial cells. We examined the cell-to-cell variability in the nitrogen-fixing behavior of Klebsiella oxytoca, a free-living bacterium. The significance of our research is in identifying not only the presence of marked variability but also the specific mechanisms that give rise to it. This understanding gives insight into both the evolutionary advantages of variable behavior as well as strategies for biotechnological applications.

The transcription factor network of E. coli steers global responses to shifts in RNAP concentration

The robustness and sensitivity of gene networks to environmental changes is critical for cell survival. How gene networks produce specific, chronologically ordered responses to genome-wide perturbations, while robustly maintaining homeostasis, remains an open question. We analysed if short- and mid-term genome-wide responses to shifts in RNA polymerase (RNAP) concentration are influenced by the known topology and logic of the transcription factor network (TFN) of Escherichia coli. We found that, at the gene cohort level, the magnitude of the single-gene, mid-term transcriptional responses to changes in RNAP concentration can be explained by the absolute difference between the gene's numbers of activating and repressing input transcription factors (TFs). Interestingly, this difference is strongly positively correlated with the number of input TFs of the gene. Meanwhile, short-term responses showed only weak influence from the TFN. Our results suggest that the global topological traits of the TFN of E. coli shape which gene cohorts respond to genome-wide stresses.

Cis-Regulatory Logic Produces Gene-Expression Noise Describing Phenotypic Heterogeneity in Bacteria

Gene transcriptional process is random. It occurs in bursts and follows single-molecular kinetics. Intermittent bursts are measured based on their frequency and size. They influence temporal fluctuations in the abundance of total mRNA and proteins by generating distinct transcriptional variations referred to as “noise”. Noisy expression induces uncertainty because the association between transcriptional variation and the extent of gene expression fluctuation is ambiguous. The promoter architecture and remote interference of different cis-regulatory elements are the crucial determinants of noise, which is reflected in phenotypic heterogeneity. An alternative perspective considers that cellular parameters dictating genome-wide transcriptional kinetics follow a universal pattern. Research on noise and systematic perturbations of promoter sequences reinforces that both gene-specific and genome-wide regulation occur across species ranging from bacteria and yeast to animal cells. Thus, deciphering gene-expression noise is essential across different genomics applications. Amidst the mounting conflict, it is imperative to reconsider the scope, progression, and rational construction of diversified viewpoints underlying the origin of the noise. Here, we have established an indication connecting noise, gene expression variations, and bacterial phenotypic variability. This review will enhance the understanding of gene-expression noise in various scientific contexts and applications.

Visualization of Gene Reciprocity among Lactic Acid Bacteria in Yogurt by RNase H-Assisted Rolling Circle Amplification-Fluorescence In Situ Hybridization

Recently, we developed an in situ mRNA detection method termed RNase H-assisted rolling circle amplification-fluorescence in situ hybridization (RHa-RCA-FISH), which can detect even short mRNA in a bacterial cell. However, because this FISH method is sensitive to the sample condition, it is necessary to find a suitable cell permeabilization and collection protocol. Here, we demonstrate its further applicability for detecting intrinsic mRNA expression using lactic acid bacteria (LAB) as a model consortium. Our results show that this method can visualize functional gene expression in LAB cells and can be used for monitoring the temporal transition of gene expression. In addition, we also confirmed that data obtained from bulk analyses such as RNA-seq or microarray do not always correspond to gene expression in individual cells. RHa-RCA-FISH will be a powerful tool to compensate for insufficient data from metatranscriptome analyses while clarifying the carriers of function in microbial consortia. By extending this technique to capture spatiotemporal microbial gene expression at the single-cell level, it will be able to characterize microbial interactions in phytoplankton-bacteria interactions.

Functional roles of microbial cell-to-cell heterogeneity and emerging technologies for analysis and control

Clonal cell populations often display significant cell-to-cell phenotypic heterogeneity, even when maintained under constant external conditions. This variability can result from the inherently stochastic nature of transcription and translation processes, which leads to varying numbers of transcripts and proteins per cell. Here, we showcase studies that reveal links between stochastic cellular events and biological functions in isogenic microbial populations. Then, we highlight emerging tools from engineering, computation, and synthetic and molecular biology that enable precise measurement, control, and analysis of gene expression noise in microorganisms. The capabilities offered by this sophisticated toolbox will shape future directions in the field and generate insight into the behavior of living systems at the single-cell level.

A Non-genetic Mechanism Involving the Integrin β4/Paxillin Axis Contributes to Chemoresistance in Lung Cancer

Tumor heterogeneity and cisplatin resistance are major causes of tumor relapse and poor survival. Here, we show that in lung cancer, interaction between paxillin (PXN) and integrin β4 (ITGB4), components of the focal adhesion (FA) complex, contributes to cisplatin resistance. Knocking down PXN and ITGB4 attenuated cell growth and improved cisplatin sensitivity, both in 2D and 3D cultures. PXN and ITGB4 independently regulated expression of several genes. In addition, they also regulated expression of common genes including USP1 and VDAC1, which are required for maintaining genomic stability and mitochondrial function, respectively. Mathematical modeling suggested that bistability could lead to stochastic phenotypic switching between cisplatin-sensitive and resistant states in these cells. Consistently, purified subpopulations of sensitive and resistant cells re-created the mixed parental population when cultured separately. Altogether, these data point to an unexpected role of the FA complex in cisplatin resistance and highlight a novel non-genetic mechanism.

Glycolytic and immunological alterations in human U937 monocytes in response to H1N1 infection

We monitored changes that took place in glycolytic enzymes, the pyruvate end product of glycolysis, tumor necrosis factor α (TNFα), and toll-like receptors (TLRs) both at the transcriptional and translational levels upon direct interaction between PR8-H1N1 and the human monocytes U937 in vitro system. U937 were first treated with H1N1 infectious viral particles or phorbol-12-myristate-13-acetate (PMA) or left untreated and later infected with the H1N1 virus. Levels of phosphofructokinase 1 (PFK1) and pyruvate were biochemically quantified. In addition, levels of TNFα, TLR3, and TLR7 were measured by ELISA. The transcriptional profiles of PFKs, inflammatory cytokines, TLR3 and TLR7 were relatively quantified by qRT-PCR. The results generally revealed significant changes in both the transcriptional and translational profiles of the studied biochemical and immunological parameters upon influenza infection in a time-dependent manner. In conclusion, H1N1 infection triggers transcriptional and translational changes in immortalized human monocytes, which might serve as markers for infection subject for further validation for their specificities.

Prochlorococcus Cells Rely on Microbial Interactions Rather than on Chlorotic Resting Stages To Survive Long-Term Nutrient Starvation

The ability of microorganisms to withstand long periods of nutrient starvation is key to their survival and success under highly fluctuating conditions that are common in nature. Therefore, one would expect this trait to be prevalent among organisms in the nutrient-poor open ocean. Here, we show that this is not the case for Prochlorococcus, a globally abundant and ecologically important marine cyanobacterium. Instead, Prochlorococcus relies on co-occurring heterotrophic bacteria to survive extended phases of nutrient and light starvation. Our results highlight the power of microbial interactions to drive major biogeochemical cycles in the ocean and elsewhere with consequences at the global scale.

Many microorganisms produce resting cells with very low metabolic activity that allow them to survive phases of prolonged nutrient or energy stress. In cyanobacteria and some eukaryotic phytoplankton, the production of resting stages is accompanied by a loss of photosynthetic pigments, a process termed chlorosis. Here, we show that a chlorosis-like process occurs under multiple stress conditions in axenic laboratory cultures of Prochlorococcus, the dominant phytoplankton linage in large regions of the oligotrophic ocean and a global key player in ocean biogeochemical cycles. In Prochlorococcus strain MIT9313, chlorotic cells show reduced metabolic activity, measured as C and N uptake by Nanoscale secondary ion mass spectrometry (NanoSIMS). However, unlike many other cyanobacteria, chlorotic Prochlorococcus cells are not viable and do not regrow under axenic conditions when transferred to new media. Nevertheless, cocultures with a heterotrophic bacterium, Alteromonas macleodii HOT1A3, allowed Prochlorococcus to survive nutrient starvation for months. We propose that reliance on co-occurring heterotrophic bacteria, rather than the ability to survive extended starvation as resting cells, underlies the ecological success of Prochlorococcus.

The route to transcription initiation determines the mode of transcriptional bursting in E. coli

Transcription is fundamentally noisy, leading to significant heterogeneity across bacterial populations. Noise is often attributed to burstiness, but the underlying mechanisms and their dependence on the mode of promotor regulation remain unclear. Here, we measure E. coli single cell mRNA levels for two stress responses that depend on bacterial sigma factors with different mode of transcription initiation (σ⁷⁰ and σ⁵⁴). By fitting a stochastic model to the observed mRNA distributions, we show that the transition from low to high expression of the σ⁷⁰-controlled stress response is regulated via the burst size, while that of the σ⁵⁴-controlled stress response is regulated via the burst frequency. Therefore, transcription initiation involving σ⁵⁴ differs from other bacterial systems, and yields bursting kinetics characteristic of eukaryotic systems.

Transcription noise in bacteria is often attributed to burstiness, but the mechanisms are unclear. Here, the authors show that the transition from low to high expression can be regulated via burst size or burst frequency, depending on the mode of transcription initiation determined by different sigma factors.

Gene expression noise in a complex artificial toxin expression system

Gene expression is an intrinsically stochastic process. Fluctuations in transcription and translation lead to cell-to-cell variations in mRNA and protein levels affecting cellular function and cell fate. Here, using fluorescence time-lapse microscopy, we quantify noise dynamics in an artificial operon in Escherichia coli, which is based on the native operon of ColicinE2, a toxin. In the natural system, toxin expression is controlled by a complex regulatory network; upon induction of the bacterial SOS response, ColicinE2 is produced (cea gene) and released (cel gene) by cell lysis. Using this ColicinE2-based operon, we demonstrate that upon induction of the SOS response noise of cells expressing the operon is significantly lower for the (mainly) transcriptionally regulated gene cea compared to the additionally post-transcriptionally regulated gene cel. Likewise, we find that mutations affecting the transcriptional regulation by the repressor LexA do not significantly alter the population noise, whereas specific mutations to post-transcriptionally regulating units, strongly influence noise levels of both genes. Furthermore, our data indicate that global factors, such as the plasmid copy number of the operon encoding plasmid, affect gene expression noise of the entire operon. Taken together, our results provide insights on how noise in a native toxin-producing operon is controlled and underline the importance of post-transcriptional regulation for noise control in this system.

Controlling cell-to-cell variability with synthetic gene circuits

Cell-to-cell variability originating, for example, from the intrinsic stochasticity of gene expression, presents challenges for designing synthetic gene circuits that perform robustly. Conversely, synthetic biology approaches are instrumental in uncovering mechanisms underlying variability in natural systems. With a focus on reducing noise in individual genes, the field has established a broad synthetic toolset. This includes noise control by engineering of transcription and translation mechanisms either individually, or in combination to achieve independent regulation of mean expression and its variability. Synthetic feedback circuits use these components to establish more robust operation in closed-loop, either by drawing on, but also by extending traditional engineering concepts. In this perspective, we argue that major conceptual advances will require new theory of control adapted to biology, extensions from single genes to networks, more systematic considerations of origins of variability other than intrinsic noise, and an exploration of how noise shaping, instead of noise reduction, could establish new synthetic functions or help understanding natural functions.

Estimating RNA numbers in single cells by RNA fluorescent tagging and flow cytometry

Estimating the statistics of single-cell RNA numbers has become a key source of information on gene expression dynamics. One of the most informative methods of in vivo single-RNA detection is MS2d-GFP tagging. So far, it requires microscopy and laborious semi-manual image analysis, which hampers the amount of collectable data. To overcome this limitation, we present a new methodology for quantifying the mean, standard deviation, and skewness of single-cell distributions of RNA numbers, from flow cytometry data on cells expressing RNA tagged with MS2d-GFP. The quantification method, based on scaling flow-cytometry data from microscopy single-cell data on integer-valued RNA numbers, is shown to readily produce precise, big data on in vivo single-cell distributions of RNA numbers and, thus, can assist in studies of transcription dynamics.