Transcription factors modulate c-Fos transcriptional bursts

Stochastic transcription in the p53-mediated response to DNA damage is modulated by burst frequency

Discontinuous transcription has been described for different mammalian cell lines and numerous promoters. However, our knowledge of how the activity of individual promoters is adjusted by dynamic signaling inputs from transcription factors is limited. To address this question, we characterized the activity of selected target genes that are regulated by pulsatile accumulation of the tumor suppressor p53 in response to ionizing radiation. We performed time-resolved measurements of gene expression at the single-cell level by smFISH and used the resulting data to inform a mathematical model of promoter activity. We found that p53 target promoters are regulated by frequency modulation of stochastic bursting and can be grouped along three archetypes of gene expression. The occurrence of these archetypes cannot solely be explained by nuclear p53 abundance or promoter binding of total p53. Instead, we provide evidence that the time-varying acetylation state of p53's C-terminal lysine residues is critical for gene-specific regulation of stochastic bursting.

Kinetics of HTLV-1 reactivation from latency quantified by single-molecule RNA FISH and stochastic modelling

The human T cell leukemia virus HTLV-1 establishes a persistent infection in vivo in which the viral sense-strand transcription is usually silent at a given time in each cell. However, cellular stress responses trigger the reactivation of HTLV-1, enabling the virus to transmit to a new host cell. Using single-molecule RNA FISH, we measured the kinetics of the HTLV-1 transcriptional reactivation in peripheral blood mononuclear cells (PBMCs) isolated from HTLV-1+ individuals. The abundance of the HTLV-1 sense and antisense transcripts was quantified hourly during incubation of the HTLV-1-infected PBMCs ex vivo. We found that, in each cell, the sense-strand transcription occurs in two distinct phases: the initial low-rate transcription is followed by a phase of rapid transcription. The onset of transcription peaked between 1 and 3 hours after the start of in vitro incubation. The variance in the transcription intensity was similar in polyclonal HTLV-1+ PBMCs (with tens of thousands of distinct provirus insertion sites), and in samples with a single dominant HTLV-1+ clone. A stochastic simulation model was developed to estimate the parameters of HTLV-1 proviral transcription kinetics. In PBMCs from a leukemic subject with one dominant T-cell clone, the model indicated that the average duration of HTLV-1 sense-strand activation by Tax (i.e. the rapid transcription) was less than one hour. HTLV-1 antisense transcription was stable during reactivation of the sense-strand. The antisense transcript HBZ was produced at an average rate of ~0.1 molecules per hour per HTLV-1+ cell; however, between 20% and 70% of HTLV-1-infected cells were HBZ-negative at a given time, the percentage depending on the individual subject. HTLV-1-infected cells are exposed to a range of stresses when they are drawn from the host, which initiate the viral reactivation. We conclude that whereas antisense-strand transcription is stable throughout the stress response, the HTLV-1 sense-strand reactivation is highly heterogeneous and occurs in short, self-terminating bursts.

Constraining the complexity of promoter dynamics using fluctuations in gene expression

Gene expression is an inherently stochastic process with transcription of mRNAs often occurring in bursts: short periods of activity followed by typically longer periods of inactivity. While a simple model involving switching between two promoter states has been widely used to analyze transcription dynamics, recent experimental observations have provided evidence for more complex kinetic schemes underlying bursting. Specifically, experiments provide evidence for complexity in promoter dynamics during the switch from the transcriptionally inactive to the transcriptionally active state. An open question in the field is: what is the minimal complexity needed to model promoter dynamics and how can we determine this? Here, we show that measurements of mRNA fluctuations can be used to set fundamental bounds on the complexity of promoter dynamics. We study models wherein the switching time distribution from transcriptionally inactive to active states is described by a general waiting-time distribution. Using approaches from renewal theory and queueing theory, we derive analytical expressions which connect the Fano factor of mRNA distributions to the waiting-time distribution for promoter switching between inactive and active states. The results derived lead to bounds on the minimal number of promoter states and thus allow us to derive bounds on the minimal complexity of promoter dynamics based on single-cell measurements of mRNA levels.

Decoding the grammar of transcriptional regulation from RNA polymerase measurements: models and their applications

The genomic revolution has indubitably brought about a paradigm shift in the field of molecular biology, wherein we can sequence, write and re-write genomes. In spite of achieving such feats, we still lack a quantitative understanding of how cells integrate environmental and intra-cellular signals at the promoter and accordingly regulate the production of messenger RNAs. This current state of affairs is being redressed by recent experimental breakthroughs which enable the counting of RNA polymerase molecules (or the corresponding nascent RNAs) engaged in the process of transcribing a gene at the single-cell level. Theorists, in conjunction, have sought to unravel the grammar of transcriptional regulation by harnessing the various statistical properties of these measurements. In this review, we focus on the recent progress in developing falsifiable models of transcription that aim to connect the molecular mechanisms of transcription to single-cell polymerase measurements. We discuss studies where the application of such models to the experimental data have led to novel mechanistic insights into the process of transcriptional regulation. Such interplay between theory and experiments will likely contribute towards the exciting journey of unfurling the governing principles of transcriptional regulation ranging from bacteria to higher organisms.

Discontinuous transcription

Numerous studies based on new single-cell and single-gene techniques show that individual genes can be transcribed in short bursts or pulses accompanied by changes in pulsing frequencies. Since so many examples of such discontinuous or fluctuating transcription have been found from prokaryotes to mammals, it now seems to be a common mode of gene expression. In this review we discuss the occurrence of the transcriptional fluctuations, the techniques used for their detection, their putative causes, kinetic characteristics, and probable physiological significance.

Transcriptional Bursting and Co-bursting Regulation by Steroid Hormone Release Pattern and Transcription Factor Mobility

Genes are transcribed in a discontinuous pattern referred to as RNA bursting, but the mechanisms regulating this process are unclear. Although many physiological signals, including glucocorticoid hormones, are pulsatile, the effects of transient stimulation on bursting are unknown. Here we characterize RNA synthesis from single-copy glucocorticoid receptor (GR)-regulated transcription sites (TSs) under pulsed (ultradian) and constant hormone stimulation. In contrast to constant stimulation, pulsed stimulation induces restricted bursting centered around the hormonal pulse. Moreover, we demonstrate that transcription factor (TF) nuclear mobility determines burst duration, whereas its bound fraction determines burst frequency. Using 3D tracking of TSs, we directly correlate TF binding and RNA synthesis at a specific promoter. Finally, we uncover a striking co-bursting pattern between TSs located at proximal and distal positions in the nucleus. Together, our data reveal a dynamic interplay between TF mobility and RNA bursting that is responsive to stimuli strength, type, modality, and duration.

Measuring transcription at a single gene copy reveals hidden drivers of bacterial individuality

Single-cell measurements of mRNA copy-number inform our understanding of stochastic gene expression [1-3], but these measurements coarse-grain over the individual copies of the gene, where transcription and its regulation stochastically take plasce [4, 5]. Here we combine single-molecule quantification of mRNA and gene loci to measure the transcriptional activity of an endogenous gene in individual Escherichia coli bacteria. Interpreted using a theoretical model for mRNA dynamics, the single-cell data allows us to obtain the probabilistic rates of promoter switching, transcription initiation and elongation, mRNA release and degradation. Unexpectedly, we find that gene activity can be strongly coupled to the transcriptional state of another copy of the same gene present in the cell, and to the event of gene replication during the bacterial cell cycle. These gene-copy and cell-cycle correlations demonstrate the limits of mapping whole-cell mRNA numbers to the underlying stochastic gene activity, and instead highlight the contribution of previously hidden variables to the observed population heterogeneity.

Transcriptional processes: Models and inference

Many biochemical events involve multistep reactions. One of the most important biological processes that involve multistep reaction is the transcriptional process. Models for multistep reaction necessarily need multiple states and it is a challenge to compute model parameters that best agree with experimental data. Therefore, the aim of this work is to design a multistep promoter model which accurately characterizes transcriptional bursting and is consistent with observed data. To address this issue, we develop a model for promoters with several OFF states and a single ON state using Erlang distribution. To explore the combined effects of model and data, we combine Monte Carlo extension of Expectation Maximization (MCEM) and delay Stochastic Simulation Algorithm (DSSA) and call the resultant algorithm as delay Bursty MCEM. We apply this algorithm to time-series data of endogenous mouse glutaminase promoter to validate the model assumptions and infer the kinetic parameters. Our results show that with multiple OFF states, we are able to infer and produce a model which is more consistent with experimental data. Our results also show that delay Bursty MCEM inference is more efficient.

Dynamics of Notch-Dependent Transcriptional Bursting in Its Native Context

Transcription is well known to be inherently stochastic and episodic, but the regulation of transcriptional dynamics is not well understood. Here, we analyze how Notch signaling modulates transcriptional bursting during animal development. Our focus is Notch regulation of transcription in germline stem cells of the nematode C. elegans. Using the MS2 system to visualize nascent transcripts and live imaging to record dynamics, we analyze bursting as a function of position within the intact animal. We find that Notch-dependent transcriptional activation is indeed "bursty"; that wild-type Notch modulates burst duration (ON-time) rather than duration of pauses between bursts (OFF-time) or mean burst intensity; and that a mutant Notch receptor, which is compromised for assembly into the Notch transcription factor complex, primarily modifies burst size (duration × intensity). These analyses thus visualize the effect of a canonical signaling pathway on metazoan transcriptional bursting in its native context.

Enhancer Priming Enables Fast and Sustained Transcriptional Responses to Notch Signaling

Information from developmental signaling pathways must be accurately decoded to generate transcriptional outcomes. In the case of Notch, the intracellular domain (NICD) transduces the signal directly to the nucleus. How enhancers decipher NICD in the real time of developmental decisions is not known. Using the MS2-MCP system to visualize nascent transcripts in single cells in Drosophila embryos, we reveal how two target enhancers read Notch activity to produce synchronized and sustained profiles of transcription. By manipulating the levels of NICD and altering specific motifs within the enhancers, we uncover two key principles. First, increased NICD levels alter transcription by increasing duration rather than frequency of transcriptional bursts. Second, priming of enhancers by tissue-specific transcription factors is required for NICD to confer synchronized and sustained activity; in their absence, transcription is stochastic and bursty. The dynamic response of an individual enhancer to NICD thus differs depending on the cellular context.

Exact and efficient hybrid Monte Carlo algorithm for accelerated Bayesian inference of gene expression models from snapshots of single-cell transcripts

Single cells exhibit a significant amount of variability in transcript levels, which arises from slow, stochastic transitions between gene expression states. Elucidating the nature of these states and understanding how transition rates are affected by different regulatory mechanisms require state-of-the-art methods to infer underlying models of gene expression from single cell data. A Bayesian approach to statistical inference is the most suitable method for model selection and uncertainty quantification of kinetic parameters using small data sets. However, this approach is impractical because current algorithms are too slow to handle typical models of gene expression. To solve this problem, we first show that time-dependent mRNA distributions of discrete-state models of gene expression are dynamic Poisson mixtures, whose mixing kernels are characterized by a piecewise deterministic Markov process. We combined this analytical result with a kinetic Monte Carlo algorithm to create a hybrid numerical method that accelerates the calculation of time-dependent mRNA distributions by 1000-fold compared to current methods. We then integrated the hybrid algorithm into an existing Monte Carlo sampler to estimate the Bayesian posterior distribution of many different, competing models in a reasonable amount of time. We demonstrate that kinetic parameters can be reasonably constrained for modestly sampled data sets if the model is known a priori. If there are many competing models, Bayesian evidence can rigorously quantify the likelihood of a model relative to other models from the data. We demonstrate that Bayesian evidence selects the true model and outperforms approximate metrics typically used for model selection.

Mechanisms regulating zygotic genome activation

Following fertilization, the two specified gametes must unite to create an entirely new organism. The genome is initially transcriptionally quiescent, allowing the zygote to be reprogrammed into a totipotent state. Gradually, the genome is activated through a process known as the maternal-to-zygotic transition, which enables zygotic gene products to replace the maternal supply that initiated development. This essential transition has been broadly characterized through decades of research in several model organisms. However, we still lack a full mechanistic understanding of how genome activation is executed and how this activation relates to the reprogramming of the zygotic chromatin architecture. Recent work highlights the central role of transcriptional activators and suggests that these factors may coordinate transcriptional activation with other developmental changes.

NF-κB-Chromatin Interactions Drive Diverse Phenotypes by Modulating Transcriptional Noise

Noisy gene expression generates diverse phenotypes, but little is known about mechanisms that modulate noise. Combining experiments and modeling, we studied how tumor necrosis factor (TNF) initiates noisy expression of latent HIV via the transcription factor nuclear factor κB (NF-κB) and how the HIV genomic integration site modulates noise to generate divergent (low-versus-high) phenotypes of viral activation. We show that TNF-induced transcriptional noise varies more than mean transcript number and that amplification of this noise explains low-versus-high viral activation. For a given integration site, live-cell imaging shows that NF-κB activation correlates with viral activation, but across integration sites, NF-κB activation cannot account for differences in transcriptional noise and phenotypes. Instead, differences in transcriptional noise are associated with differences in chromatin state and RNA polymerase II regulation. We conclude that, whereas NF-κB regulates transcript abundance in each cell, the chromatin environment modulates noise in the population to support diverse HIV activation in response to TNF.

Multiplex RNA single molecule FISH of inducible mRNAs in single yeast cells

Transcript levels powerfully influence cell behavior and phenotype and are carefully regulated at several steps. Recently developed single cell approaches such as RNA single molecule fluorescence in-situ hybridization (smFISH) have produced advances in our understanding of how these steps work within the cell. In comparison to single-cell sequencing, smFISH provides more accurate quantification of RNA levels. Additionally, transcript subcellular localization is directly visualized, enabling the analysis of transcription (initiation and elongation), RNA export and degradation. As part of our efforts to investigate how this type of analysis can generate improved models of gene expression, we used smFISH to quantify the kinetic expression of STL1 and CTT1 mRNAs in single Saccharomyces cerevisiae cells upon 0.2 and 0.4 M NaCl osmotic stress. In this Data Descriptor, we outline our procedure along with our data in the form of raw images and processed mRNA counts. We discuss how these data can be used to develop single cell modelling approaches, to study fundamental processes in transcription regulation and develop single cell image processing approaches.

Live-cell imaging reveals the interplay between transcription factors, nucleosomes, and bursting

Transcription factors show rapid and reversible binding to chromatin in living cells, and transcription occurs in sporadic bursts, but how these phenomena are related is unknown. Using a combination of in vitro and in vivo single-molecule imaging approaches, we directly correlated binding of the Gal4 transcription factor with the transcriptional bursting kinetics of the Gal4 target genes GAL3 and GAL10 in living yeast cells. We find that Gal4 dwell time sets the transcriptional burst size. Gal4 dwell time depends on the affinity of the binding site and is reduced by orders of magnitude by nucleosomes. Using a novel imaging platform called orbital tracking, we simultaneously tracked transcription factor binding and transcription at one locus, revealing the timing and correlation between Gal4 binding and transcription. Collectively, our data support a model in which multiple RNA polymerases initiate transcription during one burst as long as the transcription factor is bound to DNA, and bursts terminate upon transcription factor dissociation.

A stochastic model for post-transcriptional regulation of rare events in gene expression

Post-transcriptional regulation of gene expression is often a critical component of cellular processes involved in cell-fate decisions. Correspondingly, considerable efforts have focused on modeling post-transcriptional regulation of stochastic gene expression and on quantifying its impact on the mean and variance of protein levels. However, the impact of post-transcriptional regulation on rare events corresponding to large deviations in gene expression is less well understood. Here, we study a simple model involving post-transcriptional control of gene expression and characterize the impact of regulation on large deviations in activity fluctuations. We derive analytical results for the large deviation function for protein production rate and for the corresponding driven process which characterizes system fluctuations conditional on the rare event. Our results suggest that fluctuations in burst size rather than frequency play a dominant role in controlling rare events. The results derived also provide insight into how post-transcriptional regulation can be used to fine-tune the probability of rare fluctuations and to thereby control phenotypic variation in a population of cells.

An Optogenetic Platform for Real-Time, Single-Cell Interrogation of Stochastic Transcriptional Regulation

Transcription is a highly regulated and inherently stochastic process. The complexity of signal transduction and gene regulation makes it challenging to analyze how the dynamic activity of transcriptional regulators affects stochastic transcription. By combining a fast-acting, photo-regulatable transcription factor with nascent RNA quantification in live cells and an experimental setup for precise spatiotemporal delivery of light inputs, we constructed a platform for the real-time, single-cell interrogation of transcription in Saccharomyces cerevisiae. We show that transcriptional activation and deactivation are fast and memoryless. By analyzing the temporal activity of individual cells, we found that transcription occurs in bursts, whose duration and timing are modulated by transcription factor activity. Using our platform, we regulated transcription via light-driven feedback loops at the single-cell level. Feedback markedly reduced cell-to-cell variability and led to qualitative differences in cellular transcriptional dynamics. Our platform establishes a flexible method for studying transcriptional dynamics in single cells.

Transcriptional Burst Initiation and Polymerase Pause Release Are Key Control Points of Transcriptional Regulation

Transcriptional regulation occurs via changes to rates of different biochemical steps of transcription, but it remains unclear which rates are subject to change upon biological perturbation. Biochemical studies have suggested that stimuli predominantly affect the rates of RNA polymerase II (Pol II) recruitment and polymerase release from promoter-proximal pausing. Single-cell studies revealed that transcription occurs in discontinuous bursts, suggesting that features of such bursts like frequency and intensity could also be regulated. We combined Pol II chromatin immunoprecipitation sequencing (ChIP-seq) and single-cell transcriptional measurements to show that an independently regulated burst initiation step is required before polymerase recruitment can occur. Using a number of global and targeted transcriptional regulatory perturbations, we showed that biological perturbations regulated both burst initiation and polymerase pause release rates but seemed not to regulate polymerase recruitment rate. Our results suggest that transcriptional regulation primarily acts by changing the rates of burst initiation and polymerase pause release.

Gene transcription in bursting: a unified mode for realizing accuracy and stochasticity

There is accumulating evidence that, from bacteria to mammalian cells, messenger RNAs (mRNAs) are produced in intermittent bursts - a much 'noisier' process than traditionally thought. Based on quantitative measurements at individual promoters, diverse phenomenological models have been proposed for transcriptional bursting. Nevertheless, the underlying molecular mechanisms and significance for cellular signalling remain elusive. Here, we review recent progress, address the above issues and illuminate our viewpoints with simulation results. Despite being widely used in modelling and in interpreting experimental data, the traditional two-state model is far from adequate to describe or infer the molecular basis and stochastic principles of transcription. In bacteria, DNA supercoiling contributes to the bursting of those genes that express at high levels and are topologically constrained in short loops; moreover, low-affinity cis-regulatory elements and unstable protein complexes can play a key role in transcriptional regulation. Integrating data on the architecture, kinetics, and transcriptional input-output function is a promising approach to uncovering the underlying dynamic mechanism. For eukaryotes, distinct bursting features described by the multi-scale and continuum models coincide with those predicted by four theoretically derived principles that govern how the transcription apparatus operates dynamically. This consistency suggests a unified framework for comprehending bursting dynamics at the level of the structural and kinetic basis of transcription. Moreover, the existing models can be unified by a generic model. Remarkably, transcriptional bursting enables regulatory information to be transmitted in a digital manner, with the burst frequency representing the strength of regulatory signals. Such a mode guarantees high fidelity for precise transcriptional regulation and also provides sufficient randomness for realizing cellular heterogeneity.

Enhancer Histone Acetylation Modulates Transcriptional Bursting Dynamics of Neuronal Activity-Inducible Genes

Neuronal activity-inducible gene transcription correlates with rapid and transient increases in histone acetylation at promoters and enhancers of activity-regulated genes. Exactly how histone acetylation modulates transcription of these genes has remained unknown. We used single-cell in situ transcriptional analysis to show that Fos and Npas4 are transcribed in stochastic bursts in mouse neurons and that membrane depolarization increases mRNA expression by increasing burst frequency. We then expressed dCas9-p300 or dCas9-HDAC8 fusion proteins to mimic or block activity-induced histone acetylation locally at enhancers. Adding histone acetylation increased Fos transcription by prolonging burst duration and resulted in higher Fos protein levels and an elevation of resting membrane potential. Inhibiting histone acetylation reduced Fos transcription by reducing burst frequency and impaired experience-dependent Fos protein induction in the hippocampus in vivo. Thus, activity-inducible histone acetylation tunes the transcriptional dynamics of experience-regulated genes to affect selective changes in neuronal gene expression and cellular function.

The finite state projection based Fisher information matrix approach to estimate information and optimize single-cell experiments

Modern optical imaging experiments not only measure single-cell and single-molecule dynamics with high precision, but they can also perturb the cellular environment in myriad controlled and novel settings. Techniques, such as single-molecule fluorescence in-situ hybridization, microfluidics, and optogenetics, have opened the door to a large number of potential experiments, which begs the question of how to choose the best possible experiment. The Fisher information matrix (FIM) estimates how well potential experiments will constrain model parameters and can be used to design optimal experiments. Here, we introduce the finite state projection (FSP) based FIM, which uses the formalism of the chemical master equation to derive and compute the FIM. The FSP-FIM makes no assumptions about the distribution shapes of single-cell data, and it does not require precise measurements of higher order moments of such distributions. We validate the FSP-FIM against well-known Fisher information results for the simple case of constitutive gene expression. We then use numerical simulations to demonstrate the use of the FSP-FIM to optimize the timing of single-cell experiments with more complex, non-Gaussian fluctuations. We validate optimal simulated experiments determined using the FSP-FIM with Monte-Carlo approaches and contrast these to experiment designs chosen by traditional analyses that assume Gaussian fluctuations or use the central limit theorem. By systematically designing experiments to use all of the measurable fluctuations, our method enables a key step to improve co-design of experiments and quantitative models.

Buffering and Amplifying Transcriptional Noise During Cell Fate Specification

The molecular processes that drive gene transcription are inherently noisy. This noise often manifests in the form of transcriptional bursts, producing fluctuations in gene activity over time. During cell fate specification, this noise is often buffered to ensure reproducible developmental outcomes. However, sometimes noise is utilized as a “bet-hedging” mechanism to diversify functional roles across a population of cells. Studies of bacteria, yeast, and cultured cells have provided insights into the nature and roles of noise in transcription, yet we are only beginning to understand the mechanisms by which noise influences the development of multicellular organisms. Here we discuss the sources of transcriptional noise and the mechanisms that either buffer noise to drive reproducible fate choices or amplify noise to randomly specify fates.

Diverse Spatial Expression Patterns Emerge from Unified Kinetics of Transcriptional Bursting

How transcriptional bursting relates to gene regulation is a central question that has persisted for more than a decade. Here, we measure nascent transcriptional activity in early Drosophila embryos and characterize the variability in absolute activity levels across expression boundaries. We demonstrate that boundary formation follows a common transcription principle: a single control parameter determines the distribution of transcriptional activity, regardless of gene identity, boundary position, or enhancer-promoter architecture. We infer the underlying bursting kinetics and identify the key regulatory parameter as the fraction of time a gene is in a transcriptionally active state. Unexpectedly, both the rate of polymerase initiation and the switching rates are tightly constrained across all expression levels, predicting synchronous patterning outcomes at all positions in the embryo. These results point to a shared simplicity underlying the apparently complex transcriptional processes of early embryonic patterning and indicate a path to general rules in transcriptional regulation.

Single-Molecule Analysis Reveals Linked Cycles of RSC Chromatin Remodeling and Ace1p Transcription Factor Binding in Yeast

It is unknown how the dynamic binding of transcription factors (TFs) is molecularly linked to chromatin remodeling and transcription. Using single-molecule tracking (SMT), we show that the chromatin remodeler RSC speeds up the search process of the TF Ace1p for its response elements (REs) at the CUP1 promoter. We quantified smFISH mRNA data using a gene bursting model and demonstrated that RSC regulates transcription bursts of CUP1 only by modulating TF occupancy but does not affect initiation and elongation rates. We show by SMT that RSC binds to activated promoters transiently, and based on MNase-seq data, that RSC does not affect the nucleosomal occupancy at CUP1. Therefore, transient binding of Ace1p and rapid bursts of transcription at CUP1 may be dependent on short repetitive cycles of nucleosome mobilization. This type of regulation reduces the transcriptional noise and ensures a homogeneous response of the cell population to heavy metal stress.

Nascent RNA kinetics: Transient and steady state behavior of models of transcription

Regulation of transcription is a vital process in cells, but mechanistic details of this regulation still remain elusive. The dominant approach to unravel the dynamics of transcriptional regulation is to first develop mathematical models of transcription and then experimentally test the predictions these models make for the distribution of mRNA and protein molecules at the individual cell level. However, these measurements are affected by a multitude of downstream processes which make it difficult to interpret the measurements. Recent experimental advancements allow for counting the nascent mRNA number of a gene as a function of time at the single-cell level. These measurements closely reflect the dynamics of transcription. In this paper, we consider a general mechanism of transcription with stochastic initiation and deterministic elongation and probe its impact on the temporal behavior of nascent RNA levels. Using techniques from queueing theory, we derive exact analytical expressions for the mean and variance of the nascent RNA distribution as functions of time. We apply these analytical results to obtain the mean and variance of nascent RNA distribution for specific models of transcription. These models of initiation exhibit qualitatively distinct transient behaviors for both the mean and variance which further allows us to discriminate between them. Stochastic simulations confirm these results. Overall the analytical results presented here provide the necessary tools to connect mechanisms of transcription initiation to single-cell measurements of nascent RNA.

Rhythms of the Genome: Circadian Dynamics from Chromatin Topology, Tissue-Specific Gene Expression, to Behavior

Circadian rhythms in physiology and behavior evolved to resonate with daily cycles in the external environment. In mammals, organs orchestrate temporal physiology over the 24-h day, which requires extensive gene expression rhythms targeted to the right tissue. Although a core set of gene products oscillates across virtually all cell types, gene expression profiling across tissues over the 24-h day showed that rhythmic gene expression programs are tissue specific. We highlight recent progress in uncovering how the circadian clock interweaves with tissue-specific gene regulatory networks involving functions such as xenobiotic metabolism, glucose homeostasis, and sleep. This progress hinges on not only comprehensive experimental approaches but also computational methods for multivariate analysis of periodic functional genomics data. We emphasize dynamic chromatin interactions as a novel regulatory layer underlying circadian gene transcription, core clock functions, and ultimately behavior. Finally, we discuss perspectives on extending the knowledge of the circadian clock in animals to human chronobiology.

S-phase transcriptional buffering quantified on two different promoters

Transcriptional buffering enforced during DNA replication shows that histone acetylation governs the homeostasis process and can also restrict promoters from reaching maximum transcriptional potential

Imaging of transcription by quantitative fluorescence-based techniques allows the examination of gene expression kinetics in single cells. Using a cell system for the in vivo visualization of mammalian mRNA transcriptional kinetics at single-gene resolution during the cell cycle, we previously demonstrated a reduction in transcription levels after replication. This phenomenon has been described as a homeostasis mechanism that buffers mRNA transcription levels with respect to the cell cycle stage and the number of transcribing alleles. Here, we examined how transcriptional buffering enforced during S phase affects two different promoters, the cytomegalovirus promoter versus the cyclin D1 promoter, that drive the same gene body. We found that global modulation of histone modifications could completely revert the transcription down-regulation imposed during replication. Furthermore, measuring these levels of transcriptional activity in fixed and living cells showed that the transcriptional potential of the genes was significantly higher than actual transcription levels, suggesting that promoters might normally be limited from reaching their full transcriptional potential.

Modulation of transcriptional burst frequency by histone acetylation

Many mammalian genes are transcribed during short bursts of variable frequencies and sizes that substantially contribute to cell-to-cell variability. However, which molecular mechanisms determine bursting properties remains unclear. To probe putative mechanisms, we combined temporal analysis of transcription along the circadian cycle with multiple genomic reporter integrations, using both short-lived luciferase live microscopy and single-molecule RNA-FISH. Using the Bmal1 circadian promoter as our model, we observed that rhythmic transcription resulted predominantly from variations in burst frequency, while the genomic position changed the burst size. Thus, burst frequency and size independently modulated Bmal1 transcription. We then found that promoter histone-acetylation level covaried with burst frequency, being greatest at peak expression and lowest at trough expression, while remaining unaffected by the genomic location. In addition, specific deletions of ROR-responsive elements led to constitutively elevated histone acetylation and burst frequency. We then investigated the suggested link between histone acetylation and burst frequency by dCas9p300-targeted modulation of histone acetylation, revealing that acetylation levels influence burst frequency more than burst size. The correlation between acetylation levels at the promoter and burst frequency was also observed in endogenous circadian genes and in embryonic stem cell fate genes. Thus, our data suggest that histone acetylation-mediated control of transcription burst frequency is a common mechanism to control mammalian gene expression.

Distribution shapes govern the discovery of predictive models for gene regulation

Systems biology seeks to combine experiments with computation to predict biological behaviors. However, despite tremendous data and knowledge, biological models make less-accurate predictions compared with other fields. By analyzing single-cell, single-molecule measurements of mRNA during yeast stress response, we explore why and how the shapes of experimental distributions control prediction accuracy. We show how asymmetric data distributions with long tails cause standard modeling approaches to yield excellent fits but make meaningless predictions. We show how these biases arise from the violation of fundamental assumptions in standard modeling approaches. We demonstrate how advanced computational tools solve this dilemma and achieve predictive understanding of spatiotemporal mechanisms of transcription control including RNA polymerase initiation and elongation and mRNA accumulation, transport, and decay.

Despite substantial experimental and computational efforts, mechanistic modeling remains more predictive in engineering than in systems biology. The reason for this discrepancy is not fully understood. One might argue that the randomness and complexity of biological systems are the main barriers to predictive understanding, but these issues are not unique to biology. Instead, we hypothesize that the specific shapes of rare single-molecule event distributions produce substantial yet overlooked challenges for biological models. We demonstrate why modern statistical tools to disentangle complexity and stochasticity, which assume normally distributed fluctuations or enormous datasets, do not apply to the discrete, positive, and nonsymmetric distributions that characterize mRNA fluctuations in single cells. As an example, we integrate single-molecule measurements and advanced computational analyses to explore mitogen-activated protein kinase induction of multiple stress response genes. Through systematic analyses of different metrics to compare the same model to the same data, we elucidate why standard modeling approaches yield nonpredictive models for single-cell gene regulation. We further explain how advanced tools recover precise, reproducible, and predictive understanding of transcription regulation mechanisms, including gene activation, polymerase initiation, elongation, mRNA accumulation, spatial transport, and decay.

Transcriptional burst fraction and size dynamics during lens fiber cell differentiation and detailed insights into the denucleation process

Genes are transcribed in irregular pulses of activity termed transcriptional bursts. Cellular differentiation requires coordinated gene expression; however, it is unknown whether the burst fraction (i.e. the number of active phases of transcription) or size/intensity (the number of RNA molecules produced within a burst) changes during cell differentiation. In the ocular lens, the positions of lens fiber cells correlate precisely with their differentiation status, and the most advanced cells degrade their nuclei. Here, we examined the transcriptional parameters of the β-actin and lens differentiation-specific α-, β-, and γ-crystallin genes by RNA fluorescent in situ hybridization (FISH) in the lenses of embryonic day (E) E12.5, E14.5, and E16.5 mouse embryos and newborns. We found that cellular differentiation dramatically alters the burst fraction in synchronized waves across the lens fiber cell compartment with less dramatic changes in burst intensity. Surprisingly, we observed nascent transcription of multiple genes in nuclei just before nuclear destruction. Nuclear condensation was accompanied by transfer of nuclear proteins, including histone and nonhistone proteins, to the cytoplasm. Although lens-specific deletion of the chromatin remodeler SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A member 5 (Smarca5/Snf2h) interfered with denucleation, persisting nuclei remained transcriptionally competent and exhibited changes in both burst intensity and fraction depending on the gene examined. Our results uncover the mechanisms of nascent transcriptional control during differentiation and chromatin remodeling, confirm the burst fraction as the major factor adjusting gene expression levels, and reveal transcriptional competence of fiber cell nuclei even as they approach disintegration.

A transgenic mouse for imaging activity-dependent dynamics of endogenous Arc mRNA in live neurons

Localized translation plays a crucial role in synaptic plasticity and memory consolidation. However, it has not been possible to follow the dynamics of memory-associated mRNAs in living neurons in response to neuronal activity in real time. We have generated a novel mouse model where the endogenous Arc/Arg3.1 gene is tagged in its 3' untranslated region with stem-loops that bind a bacteriophage PP7 coat protein (PCP), allowing visualization of individual mRNAs in real time. The physiological response of the tagged gene to neuronal activity is identical to endogenous Arc and reports the true dynamics of Arc mRNA from transcription to degradation. The transcription dynamics of Arc in cultured hippocampal neurons revealed two novel results: (i) A robust transcriptional burst with prolonged ON state occurs after stimulation, and (ii) transcription cycles continue even after initial stimulation is removed. The correlation of stimulation with Arc transcription and mRNA transport in individual neurons revealed that stimulus-induced Ca²⁺ activity was necessary but not sufficient for triggering Arc transcription and that blocking neuronal activity did not affect the dendritic transport of newly synthesized Arc mRNAs. This mouse will provide an important reagent to investigate how individual neurons transduce activity into spatiotemporal regulation of gene expression at the synapse.

Single Molecule Fluorescence In Situ Hybridization (smFISH) Analysis in Budding Yeast Vegetative Growth and Meiosis

Single molecule fluorescence in situ hybridization (smFISH) is a powerful technique to study gene expression in single cells due to its ability to detect and count individual RNA molecules. Complementary to deep sequencing-based methods, smFISH provides information about the cell-to-cell variation in transcript abundance and the subcellular localization of a given RNA. Recently, we have used smFISH to study the expression of the gene NDC80 during meiosis in budding yeast, in which two transcript isoforms exist and the short transcript isoform has its entire sequence shared with the long isoform. To confidently identify each transcript isoform, we optimized known smFISH protocols and obtained high consistency and quality of smFISH data for the samples acquired during budding yeast meiosis. Here, we describe this optimized protocol, the criteria that we use to determine whether high quality of smFISH data is obtained, and some tips for implementing this protocol in other yeast strains and growth conditions.

Imaging mRNA In Vivo, from Birth to Death

RNA is the fundamental information transfer system in the cell. The ability to follow single messenger RNAs (mRNAs) from transcription to degradation with fluorescent probes gives quantitative information about how the information is transferred from DNA to proteins. This review focuses on the latest technological developments in the field of single-mRNA detection and their usage to study gene expression in both fixed and live cells. By describing the application of these imaging tools, we follow the journey of mRNA from transcription to decay in single cells, with single-molecule resolution. We review current theoretical models for describing transcription and translation that were generated by single-molecule and single-cell studies. These methods provide a basis to study how single-molecule interactions generate phenotypes, fundamentally changing our understating of gene expression regulation.

Identification of gene regulation models from single-cell data

In quantitative analyses of biological processes, one may use many different scales of models (e.g. spatial or non-spatial, deterministic or stochastic, time-varying or at steady-state) or many different approaches to match models to experimental data (e.g. model fitting or parameter uncertainty/sloppiness quantification with different experiment designs). These different analyses can lead to surprisingly different results, even when applied to the same data and the same model. We use a simplified gene regulation model to illustrate many of these concerns, especially for ODE analyses of deterministic processes, chemical master equation and finite state projection analyses of heterogeneous processes, and stochastic simulations. For each analysis, we employ Matlab and Python software to consider a time-dependent input signal (e.g. a kinase nuclear translocation) and several model hypotheses, along with simulated single-cell data. We illustrate different approaches (e.g. deterministic and stochastic) to identify the mechanisms and parameters of the same model from the same simulated data. For each approach, we explore how uncertainty in parameter space varies with respect to the chosen analysis approach or specific experiment design. We conclude with a discussion of how our simulated results relate to the integration of experimental and computational investigations to explore signal-activated gene expression models in yeast (Neuert et al 2013 Science 339 584-7) and human cells (Senecal et al 2014 Cell Rep. 8 75-83)5.

What shapes eukaryotic transcriptional bursting?

Isogenic cells in a common environment present a large degree of heterogeneity in gene expression. Part of this variability is attributed to transcriptional bursting: the stochastic activation and inactivation of promoters that leads to the discontinuous production of mRNA. The diversity in bursting patterns displayed by different genes suggests the existence of a connection between bursting and gene regulation. Experimental strategies such as single-molecule RNA FISH, MS2-GFP or short-lived protein reporters allow the quantification of transcriptional bursting and the comparison of bursting kinetics between conditions, allowing therefore the identification of molecular mechanisms modulating transcriptional bursting. In this review we recapitulate the impact on transcriptional bursting of different molecular aspects of transcription such as the chromatin environment, nucleosome occupancy, histone modifications, the number and affinity of regulatory elements, DNA looping and transcription factor availability. More specifically, we examine their role in tuning the burst size or the burst frequency. While some molecular mechanisms involved in transcription such as histone marks can affect every aspect of bursting, others predominantly influence the burst size (e.g. the number and affinity of cis-regulatory elements) or frequency (e.g. transcription factor availability).

Imaging transcriptional dynamics

Recent advances in imaging techniques have enabled visualizations of nascent transcripts or individual protein molecules at high spatiotemporal resolution, revealing the complex nature of transcriptional regulation. Here, we highlight recent studies that have provided comprehensive insights to transcriptional dynamics using such quantitative imaging techniques. Specifically, they demonstrated that transcriptional activity is stochastic, and such transcriptional bursting is modulated by multiple components like chromatin environments, concentration of transcription factors, and enhancer-promoter interactions. Moreover, recent studies suggested that regulation of transcriptional activity is more complex than previously thought, by showing that transcription factors and RNA polymerases also move within the cell with distinct kinetics and sometimes form dynamic clusters to mediate transcriptional initiation.

Estrogen-dependent control and cell-to-cell variability of transcriptional bursting

Cellular decision-making and environmental adaptation are dependent upon a heterogeneous response of gene expression to external cues. Heterogeneity arises in transcription from random switching between transcriptionally active and inactive states, resulting in bursts of RNA synthesis. Furthermore, the cellular state influences the competency of transcription, thereby globally affecting gene expression in a cell-specific manner. We determined how external stimuli interplay with cellular state to modulate the kinetics of bursting. To this end, single-cell dynamics of nascent transcripts were monitored at the endogenous estrogen-responsive GREB1 locus. Stochastic modeling of gene expression implicated a two-state promoter model in which the estrogen stimulus modulates the frequency of transcriptional bursting. The cellular state affects transcriptional dynamics by altering initiation and elongation kinetics and acts globally, as GREB1 alleles in the same cell correlate in their transcriptional output. Our results suggest that cellular state strongly affects the first step of the central dogma of gene expression, to promote heterogeneity in the transcriptional output of isogenic cells.

Frequency Modulation of Transcriptional Bursting Enables Sensitive and Rapid Gene Regulation

Gene regulation is a complex non-equilibrium process. Here, we show that quantitating the temporal regulation of key gene states (transcriptionally inactive, active, and refractory) provides a parsimonious framework for analyzing gene regulation. Our theory makes two non-intuitive predictions. First, for transcription factors (TFs) that regulate transcription burst frequency, as opposed to amplitude or duration, weak TF binding is sufficient to elicit strong transcriptional responses. Second, refractoriness of a gene after a transcription burst enables rapid responses to stimuli. We validate both predictions experimentally by exploiting the natural, optogenetic-like responsiveness of the Neurospora GATA-type TF White Collar Complex (WCC) to blue light. Further, we demonstrate that differential regulation of WCC target genes is caused by different gene activation rates, not different TF occupancy, and that these rates are tuned by both the core promoter and the distance between TF-binding site and core promoter. In total, our work demonstrates the relevance of a kinetic, non-equilibrium framework for understanding transcriptional regulation.

DNA residence time is a regulatory factor of transcription repression

Transcription comprises a highly regulated sequence of intrinsically stochastic processes, resulting in bursts of transcription intermitted by quiescence. In transcription activation or repression, a transcription factor binds dynamically to DNA, with a residence time unique to each factor. Whether the DNA residence time is important in the transcription process is unclear. Here, we designed a series of transcription repressors differing in their DNA residence time by utilizing the modular DNA binding domain of transcription activator-like effectors (TALEs) and varying the number of nucleotide-recognizing repeat domains. We characterized the DNA residence times of our repressors in living cells using single molecule tracking. The residence times depended non-linearly on the number of repeat domains and differed by more than a factor of six. The factors provoked a residence time-dependent decrease in transcript level of the glucocorticoid receptor-activated gene SGK1. Down regulation of transcription was due to a lower burst frequency in the presence of long binding repressors and is in accordance with a model of competitive inhibition of endogenous activator binding. Our single molecule experiments reveal transcription factor DNA residence time as a regulatory factor controlling transcription repression and establish TALE-DNA binding domains as tools for the temporal dissection of transcription regulation.

Cell Fate Decisions During Preimplantation Mammalian Development

The early mouse embryo offers a phenomenal system to dissect how changes in the mechanisms controlling cell fate are integrated with morphogenetic events at the single-cell level. New technologies based on live imaging have enabled the discovery of dynamic changes in the regulation of single genes, transcription factors, and epigenetic mechanisms directing early cell fate decision in the early embryo. Here, we review recent progress in linking molecular dynamic events occurring at the level of the single cell in vivo, to some of the key morphogenetic changes regulating early mouse development.

Inferring gene regulatory networks from single-cell data: a mechanistic approach

Background

The recent development of single-cell transcriptomics has enabled gene expression to be measured in individual cells instead of being population-averaged. Despite this considerable precision improvement, inferring regulatory networks remains challenging because stochasticity now proves to play a fundamental role in gene expression. In particular, mRNA synthesis is now acknowledged to occur in a highly bursty manner.

Results

We propose to view the inference problem as a fitting procedure for a mechanistic gene network model that is inherently stochastic and takes not only protein, but also mRNA levels into account. We first explain how to build and simulate this network model based upon the coupling of genes that are described as piecewise-deterministic Markov processes. Our model is modular and can be used to implement various biochemical hypotheses including causal interactions between genes. However, a naive fitting procedure would be intractable. By performing a relevant approximation of the stationary distribution, we derive a tractable procedure that corresponds to a statistical hidden Markov model with interpretable parameters. This approximation turns out to be extremely close to the theoretical distribution in the case of a simple toggle-switch, and we show that it can indeed fit real single-cell data. As a first step toward inference, our approach was applied to a number of simple two-gene networks simulated in silico from the mechanistic model and satisfactorily recovered the original networks.

Conclusions

Our results demonstrate that functional interactions between genes can be inferred from the distribution of a mechanistic, dynamical stochastic model that is able to describe gene expression in individual cells. This approach seems promising in relation to the current explosion of single-cell expression data.

Electronic supplementary material

The online version of this article (doi:10.1186/s12918-017-0487-0) contains supplementary material, which is available to authorized users.

Stochastic models of gene transcription with upstream drives: exact solution and sample path characterization

Gene transcription is a highly stochastic and dynamic process. As a result, the mRNA copy number of a given gene is heterogeneous both between cells and across time. We present a framework to model gene transcription in populations of cells with time-varying (stochastic or deterministic) transcription and degradation rates. Such rates can be understood as upstream cellular drives representing the effect of different aspects of the cellular environment. We show that the full solution of the master equation contains two components: a model-specific, upstream effective drive, which encapsulates the effect of cellular drives (e.g. entrainment, periodicity or promoter randomness) and a downstream transcriptional Poissonian part, which is common to all models. Our analytical framework treats cell-to-cell and dynamic variability consistently, unifying several approaches in the literature. We apply the obtained solution to characterize different models of experimental relevance, and to explain the influence on gene transcription of synchrony, stationarity, ergodicity, as well as the effect of time scales and other dynamic characteristics of drives. We also show how the solution can be applied to the analysis of noise sources in single-cell data, and to reduce the computational cost of stochastic simulations.

Nonlatching positive feedback enables robust bimodality by decoupling expression noise from the mean

Fundamental to biological decision-making is the ability to generate bimodal expression patterns where 2 alternate expression states simultaneously exist. Here, we use a combination of single-cell analysis and mathematical modeling to examine the sources of bimodality in the transcriptional program controlling HIV's fate decision between active replication and viral latency. We find that the HIV transactivator of transcription (Tat) protein manipulates the intrinsic toggling of HIV's promoter, the long terminal repeat (LTR), to generate bimodal ON-OFF expression and that transcriptional positive feedback from Tat shifts and expands the regime of LTR bimodality. This result holds for both minimal synthetic viral circuits and full-length virus. Strikingly, computational analysis indicates that the Tat circuit's noncooperative "nonlatching" feedback architecture is optimized to slow the promoter's toggling and generate bimodality by stochastic extinction of Tat. In contrast to the standard Poisson model, theory and experiment show that nonlatching positive feedback substantially dampens the inverse noise-mean relationship to maintain stochastic bimodality despite increasing mean expression levels. Given the rapid evolution of HIV, the presence of a circuit optimized to robustly generate bimodal expression appears consistent with the hypothesis that HIV's decision between active replication and latency provides a viral fitness advantage. More broadly, the results suggest that positive-feedback circuits may have evolved not only for signal amplification but also for robustly generating bimodality by decoupling expression fluctuations (noise) from mean expression levels.