Periodic perturbation of genetic oscillations
Highlights
► The model includes gene transcription negatively regulated by protein formed with a delay. ► Universal and specific features of the periodically forced mean-field kinetics are shown. ► Monte Carlo simulations exhibit, however, only some of these features.
Introduction
Periodic perturbation or, in other words, forcing of oscillatory kinetics is a very basic interdisciplinary subject attracting attention already a few decades (see, e.g., reviews [1], [2], [3], recent articles [4], [5], [6], [7], and references therein). Customarily, the analysis of this effect is based on the mean-field (MF) kinetic equations. The corresponding kinetics exhibit many interesting universal and specific features. To show and classify such kinetics, one usually uses the perturbation frequency and amplitude, ω and α, as governing parameters. This parameter plane can be divided up into regions where the dynamic behavior is qualitatively the same. Boundaries of the regions correspond to bifurcations. Typically, one can observe stable periodic oscillations (limit cycle) and quasi-periodic oscillations associated usually with a doubly periodic attractor (2-torus). Quasi-periodic oscillations usually occur provided that the perturbation is not too strong (i.e., α is not too large). In the sea of such oscillations, there are a multitude of narrow regions (Arnol’d tongues) corresponding to entrainment by frequencies kω/l (k and l are integers). With decreasing α, these regions touch the ω/ω∘ axis at ω = kω∘/l, where ω∘ is the natural frequency. In practice, one can observe only the main tongues. If α is large, oscillations occur, as a rule, with the imposed frequency, ω. In addition, one can sometimes find relatively narrow regions corresponding to fully chaotic oscillations associated with strange attractors. The difference between quasi-periodic and fully chaotic oscillations can be seen in the distribution of the frequencies obtained by expanding the kinetics into Fourier series. Quasi-periodic oscillations exhibit primarily the natural frequency, imposed frequency, and multiple frequencies. Chaos associated with strange attractors is characterized by a broader distribution of frequencies. With decreasing reactant population, due to fluctuations, the natural oscillations exhibit stochastic features and eventually may disappear. In this limit, the perturbation-related kinetic features, predicted by the MF theory, may be partly or completely smeared.
To motivate our present study in this area, we recall that the gene transcription in cells is usually governed by regulatory proteins. Due to the abundance of the protein-mediated feedbacks, the kinetics of the formation of RNAs (mRNAs and non-coding RNAs) and proteins in cells are often complex, e.g., bistable or oscillatory even in the simplest genetic networks (reviewed in Refs. [8], [9], [10], [11], [12]). The kinetic models of oscillatory genetic networks are now numerous (see, e.g., reviews [10], [11], [12] and articles [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48]). Many of them are generic and focused on the simplest schemes. There are also more complex and specific models scrutinizing, e.g., the cell cycle [16], [19], [37]. The design of relatively complex schemes of genetic oscillators is discussed by Kobayashi et al. [44], [45]. Oscillations in complex distributed genetic networks containing many genes are analyzed in Refs. [47], [48]. The bulk of these models is based on the MF kinetic equations ignoring fluctuations of the RNA and protein populations.
In cells, the numbers of copies of each RNA and protein are often relatively low [49], [50] (lower or often appreciably lower than a few hundreds and thousands, respectively) and accordingly the gene expression frequently exhibits stochastic features (reviewed in Refs. [8], [9], [12], [51], [52], [53]). The corresponding theoretical studies are now numerous and focused primarily on the non-oscillatory kinetics. The stochastic oscillatory kinetics of gene expression were analyzed, e.g., in Refs. [17], [18], [20], [26], [27], [38].
The effect of periodic perturbations on bistable kinetics of gene expression has recently been scrutinized in Refs. [54], [55], [56], [57]. In this study, we analyze the situation when a gene forming the simplest oscillatory network is regulated by protein which itself belongs to another oscillatory network (the latter oscillations may be spontaneous and include only RNAs and proteins or induced, e.g., by the external carbon source [58]). In other words, this means that the former oscillations are periodically perturbed. Such perturbations are of considerable intrinsic interest. The potential biological significance of this effect for entraining and/or amplifying genetic oscillations has just been recognized [59]. In fact, this area is now open for studies and debate. To our knowledge, the first corresponding theoretical results were obtained by Hasty et al. [14]. They used a two-variable MF oscillatory model (it was later employed also in Ref. [60]) with mutually-regulated synthesis and degradation of proteins. The periodic perturbation, αsin (ωt), was added to the protein-regulated rate of formation of one of the proteins, i.e., the regulation and perturbation were assumed to be independent. For this model, Hasty et al. have constructed a course-grained kinetic phase diagram and illustrated the ability of amplification of perturbations. Battogtokh et al. [61], [62] have analyzed periodically perturbed genetic oscillations with emphasis on the eukaryotic cell cycle. The perturbation was added to the MF equations for the growth of the cell mass and the cell-cycle phase [61] or to the rate of synthesis of one of the enzymes [62]. Tsumoto et al. [63] have presented a MF model showing how light adaptation may result in the entrainment of circadian oscillations. Herzel et al. [64], [65] have recently used simple MF amplitude-phase oscillator models to illustrate how coupling of oscillators may influence the entrainment of circadian clocks. In particular, Herzel et al. [64] have explained their experimental finding that the lung clocks entrain to extreme zeitgeber cycles, whereas the suprachiasmatic nucleus clocks do not. Concerning the experiments, we refer to a recent interesting study by Mondragón-Palomino et al. [66] where the phases of hundreds of synthetic genetic oscillators have been tracked relative to a common external stimulus (the transcriptional inducer, arabinose) to map the entrainment regions predicted by a detailed model of the clock. Synthetic oscillators were observed to be frequency-locked in wide intervals of the external period and showed higher-order resonance.
As already noticed, the kinetic models of oscillatory kinetics of gene expression are now numerous. Many simplest generic models include the delay between gene transcription and formation of regulatory protein(s) (for the classification of the mechanisms of genetic oscillations, see, e.g., Refs. [11], [38]). In our work, we study the effect of periodic perturbations on one of the simplest generic oscillators including a gene with negative regulation of the mRNA production by protein formed with a delay [this model belongs to Goodwin’s class (reviewed in Ref. [11])]. Using the MF kinetic equations, we first show how the oscillatory kinetics can be perturbed in this case (Section 2). Complementing the MF treatment, we present Monte Carlo (MC) simulations illustrating the role of fluctuations in such kinetics (Section 3).
Compared to previous studies (Refs. [14], [60], [61], [62], [63], [64], [65]), our treatment has four novel and/or complementary ingredients. First, the model we adopt is different. Secondly, we employ a different scheme of perturbation (compared to Refs. [14], [60], [61], [62]). Specifically, the perturbation is assumed to modulate the mRNA synthesis rate, and the effects of the inherent regulation and perturbation are factorized (in Refs. [14], [60], [61], [62], these effects are additive). Thirdly, we exhibit typical MF kinetics in much more detail. In addition, we present MC simulations illustrating smearing of some of the MF kinetic features (in Refs. [14], [60], [61], [62], [63], [64], [65], the stochastic effects are not discussed). In particular, we illustrate evolution of the kinetics with decreasing mRNA and protein populations. Our MF and MC results are expected to be of interest for general readership.
In our model, the delay between gene transcription and formation of the regulatory protein is related to protein conversion steps [see steps (3) in the scheme described below in Section 2.1]. For a simple negative loop, the delay is a necessary condition for oscillations [11]. It can be introduced by using ordinary differential equations (as in this work) or, slightly more formally, by employing delay differential equations [9], [11]. In the deterministic case in the absence of periodic perturbation, these two approaches are nearly equivalent [11]. In the deterministic case with periodic perturbation, the results are also expected to be similar although it should be proved. In the stochastic regime, the type of fluctuations depends on the mechanism of the delay. For example, the type of fluctuations related to protein conversion is different compared to that related to the delay occurring due to mRNA translation. Under such circumstances, one cannot exclude tiny differences in the results predicted by the two approaches. With this reservation, we believe that our general conclusions are valid irrespective of the mechanism of the delay.
Section snippets
Model
As already noted in the Introduction, our model includes the delay between gene transcription,and formation of the regulatory protein. The latter process is considered to occur via mRNA translation into protein P1,and conversion of P1 to protein P2 and then to protein P3,Step (1) is assumed to be negatively regulated by P3, and accordingly this regulation is delayed by steps (3). In addition, the model includes degradation of mRNA, P1 and P3,
Monte Carlo simulations
With fluctuations of the reactant populations, kinetic oscillations become stochastic and may disappear if the reactant populations are small. The MF kinetic features related to periodic perturbation may be smeared partly or completely. To include fluctuations of the reactant populations into the kinetics under consideration and to scrutinize their role, we have used the standard Gillespie MC algorithm [67] (see also Ref. [57] for the details related to gene expression) based on calculation of
Conclusion
In summary, we have performed a MF analysis and MC simulations of periodic perturbation of the simplest genetic oscillatory network including a gene with negative regulation of the mRNA production by protein, obtained via mRNA translation and two steps of conversion. The main results of our study are as follows:
- (i)
As expected, the forced kinetics are primarily either periodic or quasi-periodic.
- (ii)
The kinetic phase diagram for the model under consideration exhibits regions similar to those predicted
Note added in the proof
To introduce into the subject, we have mentioned Refs. [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48] in Section 1. In addition, we may mention that Suzuki et al. [68] have recently suggested an oscillatory scenario of cell differentiation.
References (68)
Chaotic response of nonlinear oscillators
Phys Rep
(1982)Periodic perturbation of the kinetics of heterogeneous catalytic reactions
Surf Sci Rep
(2004)- et al.
Mixed-mode oscillations and chaos from a simple second-order oscillator under weak periodic perturbation
Phys Lett A
(2011) Stochastic and delayed stochastic models of gene expression and regulation
Math Biosci
(2010)Kinetic models of gene expression including non-coding RNAs
Phys Rep
(2011)- et al.
A model for restriction point control of the mammalian cell cycle
J Theor Biol
(2004) - et al.
Analysis of a generic model of eukaryotic cell-cycle regulation
Biophys J
(2006) - et al.
Intracellular delay limits cyclic changes in gene expression
Math Biosci
(2007) - et al.
The role of microRNA in the delayed negative feedback regulation of gene expression
Biochem Biophys Res Commun
(2007) - et al.
A simple model of genetic oscillations through regulated degradation
Chaos Solit Fract
(2008)
Dissecting the dynamics of the Hes1 genetic oscillator
J Theor Biol
Oscillations and bistability in the stochastic model of p53 regulation
J Theor Biol
Kinetic oscillations in the expression of messenger RNA, regulatory protein, and nonprotein coding RNA
Chem Phys Lett
Oscillatory dynamics in a simple gene regulatory network mediated by small RNAs
Physica A
Oscillation, cooperativity, and intermediates in the self-repressing gene
Chem Phys Lett
Comparison of stationary and oscillatory dynamics described by differential equations and Boolean maps in transcriptional regulatory circuits
Phys Lett A
Effects of nonlinear degradation of protein on the oscillatory dynamics in a simple gene regulatory network
Physica A
Nested feedback loops in gene regulation
Physica A
Models of stochastic gene expression
Phys Life Rev
Signatures of gene expression noise in cellular systems
Progr Biophys Mol Biol
Bursts and pulses: insights from single cell studies into transcriptional mechanisms
Curr Opin Genet Develop
Coupling oscillations and switches in genetic networks
Biosystems
Periodic perturbation of the bistable kinetics of gene expression
Physica A
Oscillatory signaling processes: the how, the why and the where
Curr Opin Genet Develop
Periodically perturbed chemical systems
Localized mode-locking structure in the stimulus parameter space for a sinusoidally forced FitzHugh–Nagumo neuron
J Phys Soc Jpn
Effects of periodic perturbations on the oscillatory behavior in the NO+H2 reaction on Pt(100)
Phys Rev E
Periodic perturbation of one of two identical chemical oscillators coupled via inhibition
Phys Rev E
Stochasticity in gene expression: from theories to phenotypes
Nature Rev Genet
Oscillations and temporal signalling in cells
Phys Biol
A comparative analysis of synthetic genetic oscillators
J R Soc Interf
A synthetic oscillatory network of transcriptional regulators
Nature
Synthetic gene network for entraining and amplifying cellular oscillations
Phys Rev Lett
A detailed predictive model of the mammalian circadian clock
Proc Natl Acad Sci USA
Cited by (11)
mRNA function after intracellular delivery and release
2018, BioSystemsBounded noise induced first-order phase transitions in a baseline non-spatial model of gene transcription
2018, Physica A: Statistical Mechanics and its ApplicationsCitation Excerpt :In this manuscript we mainly investigate these phenomena from a theoretical perspective. However, the above described simple cell circuital module exhibits some essential features for complex heterogeneous cellular dynamics: nonlinearity [1], multi-stability [2–9], feedbacks [10–12] and extrinsic forms of stochasticity [13–19]. The concept of noise-induced-transitions (NITs) [20] is important in modern biology [16,17]: Gaussian extrinsic noises perturbing monostable nonlinear systems may induce multimodal stationary probability densities.
Pacemaker dynamics in the full Morris-Lecar model
2014, Communications in Nonlinear Science and Numerical SimulationCitation Excerpt :The results in this article show that the Morris–Lecar model can be used a as a low dimensional model for describing the dynamics of pacemaker muscle cells. Such kind of models, because of its relative simplicity and biophysical foundation, have proven to be useful to study many aspects of biological rhythm of different types; e.g., circadian oscillations [38–40], neuronal dynamics [35–37,30,31], biochemical oscillators [41–43], among others. In particular, the Morris–Lecar model is based in a well-established formalism for the electrophysics of the membrane potential, and the ranges of the values of the model parameters have been determined from experiments.
Patterning, dynamics and evolution in the ocellar complex of the fruit fly
2018, Understanding Complex SystemsKinetics of lipid-nanoparticle-mediated intracellular mRNA delivery and function
2017, Physical Review EMicroRNA intercellular transfer and bioelectrical regulation of model multicellular ensembles by the gap junction connectivity
2017, Journal of Physical Chemistry B