Philippe Cluzel
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Philippe Cluzel has a total of 70 co-authors. Here are the top co-authors by number of shared publications:
Mayra Garcia-Alcala
Department of Molecular and Cellular Biology, Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University
6 shared publications- Mayra Garcia-Alcala
Department of Molecular and Cellular Biology, Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University
5 shared publications Calin C. Guet
‡FAS Center for Systems Biology and Department of Molecular and Cellular Biology, School of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts
3 shared publicationsEnrique Balleza
Harvard University
3 shared publicationsHeungwon Park
The James Franck Institute, University of Chicago, Chicago, Illinois, USA
3 shared publications- Kevin B. Wood
Department of Physics, University of Michigan
3 shared publications
Latest Publications
- Open Access
- 28 October 2026
- Published by eLife Sciences Publications, Ltd in eLife
- Vol. 15
Out-of-balance growth enables cost-free synthesis of the flagellum and other proteins in a single bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in Escherichia coli , is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- 28 October 2026
- Published by eLife Sciences Publications, Ltd in eLife
- Vol. 15
Open AccessOut-of-balance growth enables cost-free synthesis of the flagellum and other proteins in a single bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in Escherichia coli , is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- Open Access
- 8 October 2026
- Published by eLife Sciences Publications, Ltd in eLife preprints
- Vol. 15
Out-of-balance Growth Enables Cost-free Synthesis of the Flagellum and Other Proteins in a Single Bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in E. coli, is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- 8 October 2026
- Published by eLife Sciences Publications, Ltd in eLife preprints
- Vol. 15
Open AccessOut-of-balance Growth Enables Cost-free Synthesis of the Flagellum and Other Proteins in a Single Bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in E. coli, is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- Open Access
- 5 August 2026
- Published by eLife Sciences Publications, Ltd in eLife preprints
- Vol. 15
Out-of-balance Growth Enables Cost-free Synthesis of the Flagellum and Other Proteins in a Single Bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in E. coli, is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- 5 August 2026
- Published by eLife Sciences Publications, Ltd in eLife preprints
- Vol. 15
Open AccessOut-of-balance Growth Enables Cost-free Synthesis of the Flagellum and Other Proteins in a Single Bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in E. coli, is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- Open Access
Out-of-balance Growth Enables Cost-free Synthesis of the Flagellum and Other Proteins in a Single Bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in E. coli, is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
Open AccessOut-of-balance Growth Enables Cost-free Synthesis of the Flagellum and Other Proteins in a Single Bacterium
The cost of expressing unnecessary genes on growth has been explained at the population level through balanced growth assumptions. However, due to the high degree of stochasticity in growth and gene regulation, it remains unclear how classical growth laws derived from bulk measurements manifest at the level of individual cells. For example, flagellar gene expression, one of the most energy-intensive processes in E. coli, is typically associated with reduced growth rates in bulk measurements. By contrast, we find at the single-cell level that a strikingly opposite behavior coexists: flagellar gene activity is associated with an increase in growth rate. We show that this apparent contradiction, reminiscent of Simpson’s paradox, results from examining in each cell instantaneous growth and gene expression without indiscriminately aggregating the single-cell data like in bulk measurements. We attribute this unrecognized behavior to out-of-balance growth boosts that temporarily offset the burden of flagellum production, likely driven by a surplus of growth factors inherited across cell division. Furthermore, we show that this non-intuitive relationship between growth and gene expression at the single-cell level is general, as we observe similar effects with synthetic systems driving the expression of fluorescent proteins downstream of constitutive promoters. Finally, we use a computational model to show that the inheritance of parental growth factors can explain the apparent contradiction and constitutes a general mechanism for mitigating the short-term cost of nonessential and biosynthetically demanding genes in the single cell.
- Open Access
- 21 February 2023
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 19 (2) , e1010894
Resonant learning in scale-free networks
Large networks of interconnected components, such as genes or machines, can coordinate complex behavioral dynamics. One outstanding question has been to identify the design principles that allow such networks to learn new behaviors. Here, we use Boolean networks as prototypes to demonstrate how periodic activation of network hubs provides a network-level advantage in evolutionary learning. Surprisingly, we find that a network can simultaneously learn distinct target functions upon distinct hub oscillations. We term this emergent property resonant learning, as the new selected dynamical behaviors depend on the choice of the period of the hub oscillations. Furthermore, this procedure accelerates the learning of new behaviors by an order of magnitude faster than without oscillations. While it is well-established that modular network architecture can be selected through evolutionary learning to produce different network behaviors, forced hub oscillations emerge as an alternative evolutionary learning strategy for which network modularity is not necessarily required.
- 21 February 2023
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 19 (2) , e1010894
Open AccessResonant learning in scale-free networks
Large networks of interconnected components, such as genes or machines, can coordinate complex behavioral dynamics. One outstanding question has been to identify the design principles that allow such networks to learn new behaviors. Here, we use Boolean networks as prototypes to demonstrate how periodic activation of network hubs provides a network-level advantage in evolutionary learning. Surprisingly, we find that a network can simultaneously learn distinct target functions upon distinct hub oscillations. We term this emergent property resonant learning, as the new selected dynamical behaviors depend on the choice of the period of the hub oscillations. Furthermore, this procedure accelerates the learning of new behaviors by an order of magnitude faster than without oscillations. While it is well-established that modular network architecture can be selected through evolutionary learning to produce different network behaviors, forced hub oscillations emerge as an alternative evolutionary learning strategy for which network modularity is not necessarily required.
- 21 December 2022
- Published by Oxford University Press (OUP) in European Heart Journal
- Vol. 44 (26) , 2376-2385
Prevalence, patterns and outcomes of cardiac involvement in Erdheim–Chester disease
Cardiac involvement of Erdheim–Chester disease (ECD), a rare L group histiocytosis, has been reported to be associated with poor outcomes, but systematic studies are lacking. The present study aimed to investigate the prevalence, clinical features, imaging features, and prognosis of cardiac involvement in ECD in a large series. All patients with ECD who underwent cardiac magnetic resonance (CMR) imaging between 2003 and 2019 at a French tertiary center were retrospectively included. Primary outcome was all-cause mortality. Secondary outcomes were pericarditis, cardiac tamponade, conduction disorders, device implantation and coronary artery disease (CAD). A total of 200 patients were included [63 (54–71) years, 30% female, 58% BRAFV600E mutated]. Median follow-up was 5.5 years (3.3–9 years). On CMR, right atrioventricular sulcus infiltration was observed in 37% of patients, and pericardial effusion was seen in 24% of patients. In total, 8 patients (4%) had pericarditis (7 acute, 1 constrictive), 10 patients (5%) had cardiac tamponade, 5 patients (2.5%) had ECD-related high-degree conduction disorders, and 45 patients (23%) had CAD. Overall, cardiac involvement was present in 96 patients (48%) and was associated with BRAFV600E mutation [Odds ratio (OR) = 7.4, 95% confidence interval (CI) (3.5–16.8), P < 0.001] and ECD-related clinical events [OR = 5, 95%CI (1.5–21.2), P = 0.004] but not with lower survival in multivariate analysis [adjusted hazard ratio (HR) = 1.4, 95% CI (0.8–2.5), P = 0.2]. Cardiac involvement is present in nearly half of ECD patients and is associated with BRAFV600E mutation and complications (pericarditis, cardiac tamponade, and conduction disorders) but not with lower survival. More than half of ECD patients had evidence of cardiac involvement on CMR, which was associated with a higher risk of clinical events. Abbreviations: CMR cardiac magnetic resonance; ECD, Erdheim–Chester disease; HR, hazard ratio; OR, odds ratio.
- 21 December 2022
- Published by Oxford University Press (OUP) in European Heart Journal
- Vol. 44 (26) , 2376-2385
Prevalence, patterns and outcomes of cardiac involvement in Erdheim–Chester disease
Cardiac involvement of Erdheim–Chester disease (ECD), a rare L group histiocytosis, has been reported to be associated with poor outcomes, but systematic studies are lacking. The present study aimed to investigate the prevalence, clinical features, imaging features, and prognosis of cardiac involvement in ECD in a large series. All patients with ECD who underwent cardiac magnetic resonance (CMR) imaging between 2003 and 2019 at a French tertiary center were retrospectively included. Primary outcome was all-cause mortality. Secondary outcomes were pericarditis, cardiac tamponade, conduction disorders, device implantation and coronary artery disease (CAD). A total of 200 patients were included [63 (54–71) years, 30% female, 58% BRAFV600E mutated]. Median follow-up was 5.5 years (3.3–9 years). On CMR, right atrioventricular sulcus infiltration was observed in 37% of patients, and pericardial effusion was seen in 24% of patients. In total, 8 patients (4%) had pericarditis (7 acute, 1 constrictive), 10 patients (5%) had cardiac tamponade, 5 patients (2.5%) had ECD-related high-degree conduction disorders, and 45 patients (23%) had CAD. Overall, cardiac involvement was present in 96 patients (48%) and was associated with BRAFV600E mutation [Odds ratio (OR) = 7.4, 95% confidence interval (CI) (3.5–16.8), P < 0.001] and ECD-related clinical events [OR = 5, 95%CI (1.5–21.2), P = 0.004] but not with lower survival in multivariate analysis [adjusted hazard ratio (HR) = 1.4, 95% CI (0.8–2.5), P = 0.2]. Cardiac involvement is present in nearly half of ECD patients and is associated with BRAFV600E mutation and complications (pericarditis, cardiac tamponade, and conduction disorders) but not with lower survival. More than half of ECD patients had evidence of cardiac involvement on CMR, which was associated with a higher risk of clinical events. Abbreviations: CMR cardiac magnetic resonance; ECD, Erdheim–Chester disease; HR, hazard ratio; OR, odds ratio.
- Open Access
Monitoring lineages of growing and dividing bacteria reveals an inducible memory of mar operon expression
In Gram negative bacteria, the m ultiple a ntibiotic r esistance or mar operon, is known to control the expression of multi-drug efflux genes that protect bacteria from a wide range of drugs. Since different drugs induce this response, identifying the parameters that govern the dynamics of its induction is crucial to better characterize the process of tolerance and resistance. Most experiments have assumed that the properties of the mar transcriptional network can be inferred from population measurements. However, measurements from an asynchronous population of cells can mask underlying phenotypic variations of single cells. We monitored the activity of the mar promoter in single Escherichia coli cells in linear micro-colonies and established that the response to a steady level of inducer was heterogeneous within individual colonies. Specifically, sub-lineages defined by contiguous daughter-cells exhibited similar promoter activity, whereas activity was greatly variable between different sub-lineages. Specific sub-trees of uniform promoter activity persisted over several generations. Statistical analyses of the lineages suggest that the presence of these sub-trees is the signature of an inducible memory of the promoter state that is transmitted from mother to daughter cells. This single-cell study reveals that the degree of epigenetic inheritance changes as a function of inducer concentration, suggesting that phenotypic inheritance may be an inducible phenotype.
Open AccessMonitoring lineages of growing and dividing bacteria reveals an inducible memory of mar operon expression
In Gram negative bacteria, the m ultiple a ntibiotic r esistance or mar operon, is known to control the expression of multi-drug efflux genes that protect bacteria from a wide range of drugs. Since different drugs induce this response, identifying the parameters that govern the dynamics of its induction is crucial to better characterize the process of tolerance and resistance. Most experiments have assumed that the properties of the mar transcriptional network can be inferred from population measurements. However, measurements from an asynchronous population of cells can mask underlying phenotypic variations of single cells. We monitored the activity of the mar promoter in single Escherichia coli cells in linear micro-colonies and established that the response to a steady level of inducer was heterogeneous within individual colonies. Specifically, sub-lineages defined by contiguous daughter-cells exhibited similar promoter activity, whereas activity was greatly variable between different sub-lineages. Specific sub-trees of uniform promoter activity persisted over several generations. Statistical analyses of the lineages suggest that the presence of these sub-trees is the signature of an inducible memory of the promoter state that is transmitted from mother to daughter cells. This single-cell study reveals that the degree of epigenetic inheritance changes as a function of inducer concentration, suggesting that phenotypic inheritance may be an inducible phenotype.
- Open Access
- 12 August 2022
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 119 (33)
Robustness of DNA looping across multiple cell divisions in individual bacteria
DNA looping has emerged as a central paradigm of transcriptional regulation, as it is shared across many living systems. One core property of DNA looping–based regulation is its ability to greatly enhance repression or activation of genes with only a few copies of transcriptional regulators. However, this property based on a small number of proteins raises the question of the robustness of such a mechanism with respect to the large intracellular perturbations taking place during growth and division of the cell. Here we address the issue of sensitivity to variations of intracellular parameters of gene regulation by DNA looping. We use the lac system as a prototype to experimentally identify the key features of the robustness of DNA looping in growing Escherichia coli cells. Surprisingly, we observe time intervals of tight repression spanning across division events, which can sometimes exceed 10 generations. Remarkably, the distribution of such long time intervals exhibits memoryless statistics that is mostly insensitive to repressor concentration, cell division events, and the number of distinct loops accessible to the system. By contrast, gene regulation becomes highly sensitive to these perturbations when DNA looping is absent. Using stochastic simulations, we propose that the observed robustness to division emerges from the competition between fast, multiple rebinding events of repressors and slow initiation rate of the RNA polymerase. We argue that fast rebinding events are a direct consequence of DNA looping that ensures robust gene repression across a range of intracellular perturbations.
- 12 August 2022
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 119 (33)
Open AccessRobustness of DNA looping across multiple cell divisions in individual bacteria
DNA looping has emerged as a central paradigm of transcriptional regulation, as it is shared across many living systems. One core property of DNA looping–based regulation is its ability to greatly enhance repression or activation of genes with only a few copies of transcriptional regulators. However, this property based on a small number of proteins raises the question of the robustness of such a mechanism with respect to the large intracellular perturbations taking place during growth and division of the cell. Here we address the issue of sensitivity to variations of intracellular parameters of gene regulation by DNA looping. We use the lac system as a prototype to experimentally identify the key features of the robustness of DNA looping in growing Escherichia coli cells. Surprisingly, we observe time intervals of tight repression spanning across division events, which can sometimes exceed 10 generations. Remarkably, the distribution of such long time intervals exhibits memoryless statistics that is mostly insensitive to repressor concentration, cell division events, and the number of distinct loops accessible to the system. By contrast, gene regulation becomes highly sensitive to these perturbations when DNA looping is absent. Using stochastic simulations, we propose that the observed robustness to division emerges from the competition between fast, multiple rebinding events of repressors and slow initiation rate of the RNA polymerase. We argue that fast rebinding events are a direct consequence of DNA looping that ensures robust gene repression across a range of intracellular perturbations.
- Open Access
Robustness of DNA Looping Across Multiple Divisions in Individual Bacteria
DNA looping has emerged as a central paradigm of transcriptional regulation as it is shared across many living systems. One core property of DNA looping-based regulation is its ability to greatly enhance repression or activation of genes with only a few copies of transcriptional regulators. However, this property based on small number of proteins raises the question of the robustness of such a mechanism with respect to the large intracellular perturbations taking place during growth and division of the cell. Here we address the issue of sensitivity to variations of intracellular parameters of gene regulation by DNA looping. We use the lac system as a prototype to experimentally identify the key features of the robustness of DNA looping in growing E. coli cells. Surprisingly, we observe time intervals of tight repression spanning across division events, which can sometimes exceed ten generations. Remarkably, the distribution of such long time intervals exhibits memoryless statistics that is mostly insensitive to repressor concentration, cell division events, and the number of distinct loops accessible to the system. By contrast, gene regulation becomes highly sensitive to these perturbations when DNA looping is absent. Using stochastic simulations, we propose that the robustness to division events of memoryless distributions emerges from the competition between fast, multiple re-binding events of repressors and slow initiation rate of the RNA-polymerase. We argue that fast re-binding events are a direct consequence of DNA looping that ensures robust gene repression across a range of intracellular perturbations. Significance statement: It is well-established that certain intracellular regulators can stabilize DNA loops to greatly enhance activation or repression of gene transcription. In vitro but also in vivo ensemble measurements have determined that only a few copies of regulators are in fact needed to stably form DNA loops. In view of such a small number, we address the issue of sensitivity of gene regulation by DNA looping to variations of intracellular parameters in individual growing E. coli bacteria. Surprisingly, we find that DNA looping from the lac system is robust to a range of perturbations including divisions during which cells can maintain tight repression over many generations. We propose a mechanism that governs the observed robustness across a range of intracellular perturbations.
Open AccessRobustness of DNA Looping Across Multiple Divisions in Individual Bacteria
DNA looping has emerged as a central paradigm of transcriptional regulation as it is shared across many living systems. One core property of DNA looping-based regulation is its ability to greatly enhance repression or activation of genes with only a few copies of transcriptional regulators. However, this property based on small number of proteins raises the question of the robustness of such a mechanism with respect to the large intracellular perturbations taking place during growth and division of the cell. Here we address the issue of sensitivity to variations of intracellular parameters of gene regulation by DNA looping. We use the lac system as a prototype to experimentally identify the key features of the robustness of DNA looping in growing E. coli cells. Surprisingly, we observe time intervals of tight repression spanning across division events, which can sometimes exceed ten generations. Remarkably, the distribution of such long time intervals exhibits memoryless statistics that is mostly insensitive to repressor concentration, cell division events, and the number of distinct loops accessible to the system. By contrast, gene regulation becomes highly sensitive to these perturbations when DNA looping is absent. Using stochastic simulations, we propose that the robustness to division events of memoryless distributions emerges from the competition between fast, multiple re-binding events of repressors and slow initiation rate of the RNA-polymerase. We argue that fast re-binding events are a direct consequence of DNA looping that ensures robust gene repression across a range of intracellular perturbations. Significance statement: It is well-established that certain intracellular regulators can stabilize DNA loops to greatly enhance activation or repression of gene transcription. In vitro but also in vivo ensemble measurements have determined that only a few copies of regulators are in fact needed to stably form DNA loops. In view of such a small number, we address the issue of sensitivity of gene regulation by DNA looping to variations of intracellular parameters in individual growing E. coli bacteria. Surprisingly, we find that DNA looping from the lac system is robust to a range of perturbations including divisions during which cells can maintain tight repression over many generations. We propose a mechanism that governs the observed robustness across a range of intracellular perturbations.
- Open Access
Resonant Learning in Scale-free Networks
Over the last decades, analyses of the connectivity of large biological and artificial networks have identified a common scale-free topology, where few of the network elements, called hubs, control many other network elements. In monitoring the dynamics of networks’ hubs, recent experiments have revealed that they can show behaviors oscillating between ‘on’ and ‘off’ states of activation. Prompted by these observations, we ask whether the existence of oscillatory hubs states could contribute to the emergence of specific network dynamical behaviors. Here, we use Boolean threshold networks with scale-free architecture as representative models to demonstrate how periodic activation of the network hub can provide a network-level advantage in learning specific new dynamical behaviors. First, we find that hub oscillations with distinct periods can induce robust and distinct attractors whose lengths depend upon the hub oscillation period. Second, we determine that a given network can exhibit series of different attractors when we sequentially change the period of hub pulses. Using rounds of evolution and selection, these different attractors could independently learn distinct target functions. We term this network-based learning strategy resonant learning , as the emergence of new learned dynamical behaviors depends on the choice of the period of the hub oscillations. Finally, we find that resonant learning leads to convergence towards target behaviors over an order of magnitude faster than standard learning procedures. While it is already known that modular network architecture contributes to learning separate tasks, our results reveal an alternative design principle based on forced oscillations of the network hub. SIGNIFICANCE: Large networks of interconnected components such as genes or machines can coordinate complex behavioral dyamics. One outstanding question has been to identify the design principles that allow networks to learn new behaviors. Here, in simulating learning cycles, we randomly modify the interactions between components and select networks that exhibit a desired behavior. Surprisingly, we find that networks can learn new behaviors faster when the state of the most connected network component is forced to oscillate during learning. Remarkably, using distinct periods of oscillations allows a given network to learn distinct behaviors. While it is known that modular network architecture contributes to learning separate tasks, our results reveal an alternative design principle for which modules are not needed.
Open AccessResonant Learning in Scale-free Networks
Over the last decades, analyses of the connectivity of large biological and artificial networks have identified a common scale-free topology, where few of the network elements, called hubs, control many other network elements. In monitoring the dynamics of networks’ hubs, recent experiments have revealed that they can show behaviors oscillating between ‘on’ and ‘off’ states of activation. Prompted by these observations, we ask whether the existence of oscillatory hubs states could contribute to the emergence of specific network dynamical behaviors. Here, we use Boolean threshold networks with scale-free architecture as representative models to demonstrate how periodic activation of the network hub can provide a network-level advantage in learning specific new dynamical behaviors. First, we find that hub oscillations with distinct periods can induce robust and distinct attractors whose lengths depend upon the hub oscillation period. Second, we determine that a given network can exhibit series of different attractors when we sequentially change the period of hub pulses. Using rounds of evolution and selection, these different attractors could independently learn distinct target functions. We term this network-based learning strategy resonant learning , as the emergence of new learned dynamical behaviors depends on the choice of the period of the hub oscillations. Finally, we find that resonant learning leads to convergence towards target behaviors over an order of magnitude faster than standard learning procedures. While it is already known that modular network architecture contributes to learning separate tasks, our results reveal an alternative design principle based on forced oscillations of the network hub. SIGNIFICANCE: Large networks of interconnected components such as genes or machines can coordinate complex behavioral dyamics. One outstanding question has been to identify the design principles that allow networks to learn new behaviors. Here, in simulating learning cycles, we randomly modify the interactions between components and select networks that exhibit a desired behavior. Surprisingly, we find that networks can learn new behaviors faster when the state of the most connected network component is forced to oscillate during learning. Remarkably, using distinct periods of oscillations allows a given network to learn distinct behaviors. While it is known that modular network architecture contributes to learning separate tasks, our results reveal an alternative design principle for which modules are not needed.