Copyright Locasale, Chakraborty. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Regulation of Signal Duration and the Statistical Dynamics of Kinase Activation by Scaffold Proteins Ruth Nussinov, Editor National Cancer Institute, United States of America, Tel Aviv University, Israel * E-mail: arupc/at/MIT.edu Conceived and designed the experiments: JL. Performed the experiments: JL. Analyzed the data: JL. Contributed reagents/materials/analysis tools: JL. Wrote the paper: JL AC. Received December 14, 2007; Accepted May 20, 2008. | |||||||||||||||||||||||||
Abstract Scaffolding proteins that direct the assembly of multiple kinases into a spatially localized signaling complex are often essential for the maintenance of an appropriate biological response. Although scaffolds are widely believed to have dramatic effects on the dynamics of signal propagation, the mechanisms that underlie these consequences are not well understood. Here, Monte Carlo simulations of a model kinase cascade are used to investigate how the temporal characteristics of signaling cascades can be influenced by the presence of scaffold proteins. Specifically, we examine the effects of spatially localizing kinase components on a scaffold on signaling dynamics. The simulations indicate that a major effect that scaffolds exert on the dynamics of cell signaling is to control how the activation of protein kinases is distributed over time. Scaffolds can influence the timing of kinase activation by allowing for kinases to become activated over a broad range of times, thus allowing for signaling at both early and late times. Scaffold concentrations that result in optimal signal amplitude also result in the broadest distributions of times over which kinases are activated. These calculations provide insights into one mechanism that describes how the duration of a signal can potentially be regulated in a scaffold mediated protein kinase cascade. Our results illustrate another complexity in the broad array of control properties that emerge from the physical effects of spatially localizing components of kinase cascades on scaffold proteins. | |||||||||||||||||||||||||
Author Summary Signal transduction is the science of cellular communication. Cells detect signals from their environment and use them to make decisions such as whether or when to proliferate. Tight regulation of signal transduction is required for all healthy cells, and aberrant signaling leads to countless diseases such as cancer and diabetes. For example, in higher organisms such as mammals, signal transduction that leads to cell proliferation is often guided by a scaffold protein. Scaffolding proteins direct the assembly of multiple proteins involved in cell signaling by providing a platform for these proteins to carry out efficient signal transmission. Although scaffolds are widely believed to have dramatic effects on how signal transduction is carried out, the mechanisms that underlie these consequences are not well understood. Therefore, we used a computational approach that simulates the behavior of a model signal transduction module comprising a set of proteins in the presence of a scaffold. The simulations reveal mechanisms for how scaffolds can dynamically regulate the timing of cell signaling. Scaffolds allow for controlled levels of signal that are delivered inside the cell at appropriate times. Our findings support the possibility that these signaling dynamics regulated by scaffolds affect cell decision-making in many medically important intracellular processes. | |||||||||||||||||||||||||
In the context of signal transduction, cells integrate signals derived from membrane proximal events and convert them into the appropriate cell decision. Within the complex networks that integrate these signals lies a highly conserved motif involving the sequential activation of multiple protein kinases. Signal propagation through these kinase cascades is often guided by a scaffolding protein that assembles protein kinases into a multi-protein complex. Signaling complexes maintained by scaffolds are intensely studied and have been shown to affect myriad cell decisions [1]–[7]. Despite numerous advances in the understanding of the signaling function of scaffold proteins [8]–[15], many questions remain. For instance, although scaffolds are believed to have profound effects on the dynamics of signal propagation [6],[9],[10],[16], the mechanisms that underlie how scaffolds regulate signaling dynamics are not well understood. One key factor in specifying a cellular decision is the duration of a signal (i.e. the time over which a kinase remains active) [17],[18]. Differences in signal duration have been implicated as the basis of differential decisions in myriad cell processes. For example, it has been suggested that decisions on growth factor induced cell proliferation, positive and negative selection of T cells, apoptotic programs, cell cycle progression, among many others, are regulated by the duration of signaling [19]–[24]. Therefore, the issue of how a signal output, such as the activity of extracellular regulatory kinase (ERK) in a MAPK pathway, is distributed over time, is of considerable interest. There are many ways in which the duration of the output of a kinase cascade can be controlled. Regulation of signaling dynamics can arise from processes upstream of the cascade [25]. For example, degradation of upstream signaling components such as the surface receptors [26] and differential kinetics of GTPase regulators [27],[28] can be essential in regulating MAPK signaling dynamics [25]. Also, multisite phosphorylation is predicted to influence signal duration [29]. It has been also been shown that differential modes of feedback regulation that are manifested under different conditions within the same cascade can regulate signal duration [30]. Scaffold proteins have also been implicated as key determinants in the regulation of signal duration [9],[10],[31]. Because the many factors that control scaffold mediated signaling are difficult to systematically control in a laboratory setting, a precise understanding of how scaffold proteins affect the dynamics of signal transduction has proven elusive. Computational models have been useful in understanding some of the many complex ways in which scaffolds influence signal transduction [16], [32]–[34]. However, it is currently impossible to model theoretically all aspects of any biological signaling process—computational models ultimately require that many gross simplifications be made. Our aim is, therefore, not to attempt to simulate every detail of a specific biochemical pathway but rather investigate the consequences that emerge from a simple scenario of scaffold mediated signaling whereby a model cascade assembles onto a scaffold. In modeling this scenario in itself, we hope to learn more about the functional and mechanistic consequences that these specific physical constraints, imposed by assembling components of a biochemical cascade onto a scaffold, confer to signaling pathways. In parsing these effects from the myriad others that are undoubtedly important, our hope is that our results can serve as a framework for understanding the extent to which these effects are important in specific biological contexts such as the Mitogen Activated Protein Kinase (MAPK) pathway. One theoretical analysis of scaffold mediated cell signaling revealed the presence of non mononotic behavior in signal output as a function of scaffold concentration [34]. If scaffolds are required for signaling, then too few scaffolds will be detrimental to signaling. On the other hand, if scaffolds are present in excess, signaling complexes become incompletely assembled and the signal output is attenuated. As a consequence of this “prozone” effect, scaffolds were shown to also differentially affect the kinetics of signaling. The observation that scaffolds can differentially affect signaling dynamics leads to many questions. How do scaffold proteins control the time scales involved in signal propagation? An important metric of cell signaling is the time it takes for a downstream kinase to become active [35],[36]. As signal transduction is stochastic in nature, the more precise question is: what is the distribution of times characterizing the activation of a downstream kinase? How do scaffolds affect this distribution, and what might be the biological consequences of changes in this distribution as a result of signaling on a scaffold? We compute first passage time distributions [37] using a stochastic computer simulation method to investigate these questions. Specifically, we use a kinetic Monte Carlo algorithm. We have previously used such methods to study a different question concerning the regulation of signal amplitude by scaffold proteins [33]. It is also possible that a differential equation model that considers mean-field kinetics could be used to study the first passage time distribution [37]. However, such an approach would require the imposition of absorbing boundary conditions that can make the numerical analysis difficult. Our simulation results suggest that, depending on physiological conditions, scaffold proteins can allow kinase cascades to operate in different dynamical regimes that allow for large increases and decreases in the speed and characteristic time scale of signal propagation. Furthermore, and perhaps more importantly, scaffolds are shown to influence the statistical properties of the times at which kinases are activated in complex ways. Scaffolding protein kinases cascades can allow for broadly distributed waiting times of kinase activation, whereas in the absence of a scaffold, the time it takes for a kinase to be activated is effectively characterized by a single time scale. These stochastic characteristics of scaffold-mediated kinase cascades are, to our knowledge, elucidated for the first time and may have diverse biological consequences that pertain to how signal duration is regulated. It is also our hope that our results provide a framework for achieving a deeper qualitative understanding of how scaffolding proteins can regulate the dynamics of cell signaling and the statistical properties of signal transduction. | |||||||||||||||||||||||||
Model of a Protein Kinase Cascade For our study, we considered a model three tiered protein kinase cascade such as the MAPK pathway [38]. Since our aim is to study the effects of spatially localizing protein kinases on signaling dynamics, we considered a minimal description of a model kinase cascade. Many factors that are undoubtedly important in regulating signaling dynamics were not considered. These factors include feedback regulation within the cascade, allosteric and or catalytic functions provided by the scaffold, and the effects of multiple phosphorylations of each kinase [11],[25],[26],[30],[39],[40].In our model, signal propagation occurs in a three step hierarchical fashion: an initial stimulus (S) activates a MAP3K (A) that in turn, activates a MAP2K (B), that subsequently can activate its MAPK (C) substrate—phosphatases can deactivate each activated species and this deactivation occurs regardless of whether or not the active kinase is bound to a scaffold. A schematic is presented in Figure 1A that illustrates the basic processes that are allowed in our model. A steady-state ensemble is considered. That is, simulations are allowed to first reach a dynamic steady-state and once this state is reached, dynamics are studied. We do not consider dynamics from the starting time that requires propagation through a hierarchical cascade. Recent work has studied the statistical dynamics of kinase activation that result from the hierarchical organization of a kinase cascade; in that study, it was shown that the hierarchical structure of the cascade gives rise to broad waiting time distributions of cascade activation. In the regime that we study here, these effects are absent since activation of the cascade requires that an inactive C protein encounter an active B protein; our motivation is thus to investigate how the dynamics of kinase activation can be affected by assembling components of the cascade onto a scaffolding protein that localizes single complexes. Therefore, we do not emphasize how the hierarchical structure of a signaling cascade effects signal propagation and instead focus on how assembly of the cascade onto a scaffold affects signaling dynamics. We also underscore the notion that in our approach, many undoubtedly important effects such as the hierarchical structure of protein kinase cascades, the influence of feedback loops, differential enzymatic mechanisms and allosteric control by scaffolds are neglected. Again, by excising these effects, we restrict our attention to a hypothetical scenario that aims only to investigate the consequences of assembling components of a cascade onto a scaffold protein. The key quantities computed and parameters used are discussed below in Table 1 and Figure 1B. Additional details are provided in the Methods section. The Concentration of Scaffold Proteins Sets Time Scales for Signal Propagation Through a Kinase Cascade To set the context, consider the consequences of signaling in two limiting cases in our model. When the binding affinity of the kinases to the scaffold, E, is low (defined here to be close to the thermal energy, E~kBT; kB is Boltzman's constant and T is the temperature) and kinases disassociate rapidly from the scaffold, few proteins on average are bound to a scaffold. Therefore, signaling dynamics corresponds to that of a kinase cascade in solution. For a very strong affinity, EkBT all available binding sites to scaffold proteins are occupied by kinases (on average). In this case, signaling dynamics are controlled by the time required for initial stimuli to encounter and interact with each fully assembled complex.Therefore, we consider cases in which kinases can disassociate from their scaffolds and exchange with unbound kinases on time scales pertinent to cell signaling processes. Such time scales correspond to disassociation constants (Kd) on the order of 1–10 µM and off rates, koff~1s−1. Such Kd values correspond to free energies of binding of roughly 7–9 kcal/mol, an energy scale typical of protein-protein interactions in kinase cascades [41]. We have used 12 kBT as the binding energy in our simulations which corresponds to ~7.2kcal/mol. We also discuss the robustness of our results with respect to changes in this value. Scaffold concentration has been identified as a key variable that can regulate the efficiency of signal propagation through a kinase cascade [2],[5],[34]. For the set of parameters used in the simulations (Table 1), signal output (defined as the average steady state value of the final kinase in the cascade) has a non-monotonic (biphasic) dependence on the relative concentration of scaffolds ζ (, where [Scaffold] is the concentration of the scaffold and [MAP3K]0 is the concentration of the first kinase in the cascade) and peaks at an optimal value of ζ=1 [33],[34]. To quantify signaling dynamics, we consider a survival probability S(t) (methods) that, as mentioned, can be viewed as a type of autocorrelation function. In Figure 2A, S(t) is computed for different values of the relative scaffold concentration, ζ. The intrinsic time of signal propagation, τ, is the value at which S(t) decays to e−1 of its original value (S(t=τ)=e−1). Upon increasing scaffold concentration, τ increases. At very high scaffold expression levels, signals propagate so slowly that cell signaling is not observed on experimentally measurable time scales which we take to be in our simulations 106 Monte Carlo (MC) steps; 1 MC step~1 µs assuming a lattice spacing of 10 nm and a diffusion coefficient of 10 µm2/s [42]. The increase in τ spans several orders of magnitude as is observed in Figure 2B. Distinct stages are also observed in the behavior of τ, and are separated by an inflection point occurring shortly past the optimal value of scaffold concentration (ζ~1). This phenomenon suggests that different physical processes are determining the signaling dynamics at different ranges of scaffold concentration. These results also suggest that the concentration of scaffold proteins can in principle set an intrinsic time scale that determines the speed of signal propagation. Such an intrinsic time scale arises solely from changes in the concentration of scaffold proteins. This time scale can span several orders of magnitude for biologically relevant affinities and diffusion coefficients and increases monotonically with increasing scaffold concentration. Note that these calculations consider only the speed of signaling and do not necessarily imply that signaling is more efficient when τ is small. To observe the total amount of integrated signal flux, the survival probability is conditioned with the probability that a kinase in the pool of signaling molecules is active in the steady state. We compute R(t) defined as S(t) multiplied by the average number of (the final downstream) kinases active at steady state, Scaffold Proteins Influence the Duration of Signaling by Controlling How Kinase Activation Is Distributed Over Time Figure 2 emphasizes how the characteristic time for signal propagation is influenced by changes in the relative scaffold concentration. It also appears that the qualitative features of S(t) change as scaffold concentration is varied. The decay of some distributions appears highly concentrated at a particular time while the decay of other distributions appears more broadly distributed.To further investigate this observation, we plotted the survival probability as a function of the dimensionless time, t/τ. If the decay of S(t) is purely exponential, then S(t/τ) will have the form e−t/τ. Figure 3 shows S(t/τ) for different values of scaffold concentration and a decaying exponential function is given as a reference. One notices that S(t/τ) is exponential at negligible scaffold concentrations. As scaffold concentration increases, the behavior of S(t/τ) deviates from a single exponential decay. Near ζ=1, S(t/τ) shows maximal deviation from purely exponential kinetics. As scaffold expression increases past this point, the shape of S(t/τ) reverts back to an exponential form. A deviation from exponential behavior can be quantified by considering a stretched exponential function, In the limits of small and large scaffold concentrations, the presence of a single exponential decay, β~1 indicates that signal propagation, or the relaxation of S(t/τ), occurs at one characteristic time scale. In the intermediate regime, β shows significant deviations from one, thus allowing for a broadly distributed signal. When β is significantly less than one, signals can steadily propagate over several decades. In this regime, the waiting time distribution f(t), A Multistate Kinetic Mechanism Illustrates the Competition Between the Many Time Scales Involved in Scaffold-Mediated Cell Signaling Why do we observe exponential and non-exponential behavior under different conditions? Signal transduction in our model occurs on a time scale that is much slower than the microscopic time scales associated with diffusion, binding/unbinding, and enzyme catalysis. We might therefore expect that some coarse-graining exists whereby events at these fast, “microscopic” time scales interact with other relevant biophysical parameters (e.g. scaffold concentration) to give rise to emergent properties that evolve on slower times scales. These processes are a manifestation of the collective dynamics of the many processes that occur on faster time scales. Understanding the factors that govern these emergent time scales would then provide insight into the origin of the different temporal characteristics that are revealed by our simulations.In order for a signal to propagate (i.e. for the last kinase in the cascade to become active), a hierarchical sequence of phosphorylation reactions among kinases must occur that leads to the final kinase in the cascade being activated by its upstream kinases. The activation process may occur either in solution or on a scaffold. Also, in the course of signaling, kinases can exchange from a scaffold. Some kinases are bound to a scaffold that contains an incomplete assembly of the necessary signaling molecules, and are not signaling competent. Ultimately, an inactive kinase can exist in one of three states: in solution, bound to a complete complex, or bound to an incomplete complex. Figure 4A contains a diagram of such a minimal picture and arrows denote transitions between the four states. This minimalist description clarifies the behavior in Figures 3A and 3B. For low scaffold concentrations (ζ1), kinases predominately exist in solution and signal transduction is dominated by the time it takes for an upstream kinase to encounter its downstream enzyme. Since a steady-state ensemble is used, the rate limiting step for signal propagation is the diffusion limited collision between an active B* molecule with an inactive C molecule. For high scaffold concentrations (ζ1), kinases predominately exist in incomplete signaling complexes and signal transduction is limited by a time scale that characterizes the turnover of a signaling incompetent complex to one that is able to signal. For intermediate concentrations, inactive kinases can exist in each of three states and transitions between these states also occur. Thus, the source of the nonexponential relaxation (i.e. β<1) arises from the mixing of many time scales that are relevant for intermediate scaffold concentrations. Figure 4B illustrates this minimal picture of the kinetics of signal propagation derived from these physical considerations. Also note that the sensitivity of our results to changes in model parameters can be understood from this simple picture of scaffold mediated signaling dynamics. For instance, changes in kinase and scaffold concentrations result in changes in the relative amount of kinases existing in the three states in ways that have been previously characterized [33],[34]. Changes to other parameters such as the rates of activation and deactivation and the concentration of phosphatases alter the rates of transitions between these different states. For instance, if phosphatase concentrations are very large, then activation in solution is very slow and occurs predominantly on a scaffold. Also, slower rates of activation (and larger rates of deactivation) result in a larger portion of signaling originating from kinases that are bound to scaffolds. In general, when the activation of kinases originates more (less) predominantly from a particular state in the minimal model, β increases (decreases). When multiple pathways to kinase activation contribute with comparable time scales, β is small, and signaling is broadly distributed over many time scales. We have performed many simulations with varying parameters to test the robustness and parameter sensitivity of our findings and find that that the qualitative behavior of our results follow this simple, qualitative, physical picture. Additional insight can be gleaned from consideration of the power spectrum of S(t). The power spectrum, In Figure 5, (τopt)−2 P(ωτopt) is plotted versus ωτopt where the time τopt is the characteristic time scale τ for relaxation at the optimal ζ=1scaffold concentration. That is, time is rescaled to units of τopt. For each curve, at lowωτopt1 frequencies P(ωτopt) is constant (P(ωτ→0)→τ2) signifying that kinase activation has become uncorrelated. At high ωτopt1 frequencies, kinase activation is correlated and a power law decay is observed for each curve P(ωτopt)~ω−2. As a reference, note that for an exponential decay, S(t)=e−t/τ, the transition between these two regimes occurs at ωτ~1 and is determined by the Lorentzian: In Figure 5, for high (ζ=3.5, blue) and low (ζ=0.001, green) scaffold concentrations power spectra closely resemble the Lorentzian with the transition to P(ωτopt)~ω−2 behavior occurring at different frequencies. At low ζ=0.001 concentrations, the inverse time scale or corner frequency at which kinase activation decorrelates is determined by the diffusion limited rates of activation and deactivation of the final kinase C*. The corner frequency can be estimated from At high ζ=3.5 concentrations, the corner frequency is determined by rates of formation and disassociation of an intact signaling complex. Furthermore, because of these many process that comprise the relaxation rate in this case, a numerical estimate of the corner frequency is difficult. In the case of the optimal (ζ=1.0, red) concentration, the transition from constant to P(ωτopt)~ω−2 behavior occurs smoothly over many decades from ωτopt~0.1 to ωτopt~10.0. The plot in Figure 5 also resolves different frequency dependent processes occurring in signal transduction. At high frequencies or short times, ωτopt<10.0, kinase activation is limited by the diffusive motion of the kinases in the cascade. At intermediate frequencies, 0.1<ωτopt<10.0, activation is dominated by transitions between kinases assembled in competent, incompetent, and solution based kinases. For low frequencies ωτopt<0.1 or long times, kinase activation decorrelates for each scaffold concentration. Relationship Between Signal Duration and Computed First Passage Time Statistics To illustrate how computed values of S(t) and the distribution of waiting times for kinase activation relate to conventional means of defining signal duration, we consider a differential equation for the time evolution of the activated form of the final kinase within the cascade. In this picture, species become activated at rates derived from the functional form that was fitted to the survival probabilities that were computed from the simulations. The waiting time or first-passage time distribution f(t) is used as a forward rate and the activated final kinase then can be degraded with a kinetics of degradation characterized by a rate constant, k. A kinetic equation describing this process is written as:The equation for x(t) can be solved and using the initial condition, x(0)=0: This concept of signal duration can be made more precise by considering a threshold amount of signal, T, that is required for the pathway to be considered active. With a chosen value of T, the signal duration, υ, is defined as the time it takes for the signal to decay to some threshold value, T. That is, the equation | |||||||||||||||||||||||||
We first showed that scaffold concentration is a key variable in regulating the speed of signal transduction. Moreover, we showed that the concentration of a scaffold protein can influence signaling dynamics by controlling the distribution of times over which kinases become active. This type of regulation may have many important consequences that are related to the influence of signal duration on cell decisions. Controlling the times over which kinases are activated may also be useful in directing a specific, robust response in a number of ways. Thus, the scaffold concentration itself provides another variable for maintaining signal specificity by controlling signal duration. This is consistent with data from genetic studies involving KSR1 [9],[10], where the authors reported that the concentration of KSR1 can control a cell decision involving commitment to adipogenesis. Our study focused solely on aspects of scaffold mediated regulation of signal transduction and we only considered the times at which kinases are active in the course of signal transduction. Many other factors also control signal duration. For example, our study does not consider the negative feedback loops that are often associated with the upregulation of phosphatases [18],[32]) or the role of receptor downregulation in controlling signal duration. Also we did not explicitly consider the role of positive versus negative feedback loops in shaping signal duration which is undoubtedly important [30]. It was our focus to study how spatially localizing kinases on a scaffold protein influences signal duration. We aimed to untangle this effect of scaffold proteins from other essential features of kinase cascades such as allostery and feedback regulation. Also, other theoretical studies have investigated the first passage time statistics in signal transduction cascades and have found interesting dynamics that result from, in part, the sequential activation of multiple steps in a kinase cascade [35],[36]. Our studies of signaling through scaffold proteins supplement these findings and, to our knowledge, provide the first study that shows how scaffolds affect the statistics of signal transduction. Several predictions from our model of how scaffolds regulate signaling dynamics can be tested. Measurements that monitor the time course of signaling for different scaffold concentrations could potentially resolve the differences in signaling dynamics that are predicted. Also, single molecule or fluorescence correlation based spectroscopic methods [46]–[48] could potentially probe the statistics of signaling dynamics inherent in kinase cascades and study how such statistics are related to reliable cell decisions. Such techniques can monitor the propagation of a signal, at the level of an individual molecule and thus directly measure how kinase activation within a single cell is distributed over time. | |||||||||||||||||||||||||
Kinetic Monte Carlo Simulations We simulate a model protein kinase cascade such as the mitogen-activated protein kinase (MAPK) cascade (Figure 1A) in the presence and absence of a scaffold with a kinetic Monte Carlo algorithm [49],[50], which allows us to monitor the relevant stochastic dynamics. Since we are investigating phenomena that occurs on the time scales of signal transduction, we course-grain the system so that proteins are represented as discrete objects, occupying a site on a lattice of dimensions 100×100×100 lattice spacings. Scaffold proteins are modeled as rigid, immobile objects containing three binding sites that are each specific for a particular kinase. When bound to a scaffold, kinases are tethered in nearest neighbor positions that are proximal to their downstream substrates. Allowing the scaffold and scaffold-bound species to move does not affect the qualitative results. Reflecting, no flux (i.e.Neumann) boundary conditions exist at each of the faces of the cubic lattice. The system is not periodically replicated since our simulation box is a size on the order a cell. Proteins can diffuse (i.e. translate on the lattice in random directions), bind and unbind, and undergo state transformations according to the prescribed reaction network involving a three staged cascade of activation and deactivation events (Figures 1A and 1B). Protein motion is subject to excluded volume (steric) constraints in that no two proteins can occupy the same site on the lattice. Chemical (state) transformations and binding events are modeled as thermally activated processes with energy barriers for activation, inactivation, binding and unbinding reactions. Parameters used are given in Table 1.We simulate the dynamics with a fixed time step Monte Carlo algorithm. In a Monte Carlo step, n trials are attempted, where n is the number of proteins in the simulation. For a given trial, a protein is first chosen at random with uniform probability. A displacement move in a uniformly random direction is attempted with probability, Upon considering all possible nearest neighbor interactions, reaction moves, as determined by the network topology, are tried with probabililty, Parameters Used The parameters used in the simulation were first constrained to typical literature values. Energies of disassociation were taken to be 12kbT corresponding of a disassociation constant Kd of roughly 1 µM. 200 stimulatory molecules, S, 200 molecules of kinase A and a 115 ratio of A, B, and C kinases was used. If we assume a lattice spacing of 10nm, a typical diameter of a protein, the concentration of kinases in our simulation box is roughly 1 µM for kinase A and kinase B and ~5 µM for kinase C. In a physiological context, assuming the radius of the cell is about 10 µm, this approximately corresponds to ~105 molecules of kinases A and B and a copy number of ~5×105 for kinase C in our simulation. 600 generic phosphatases are also present. These relative numbers are commensurate with reported kinase concentrations in Yeast and other systems [51],[52]. Chemical kinetics were modeled in the simplest possible way by considering a single elementary reactive collision; i.e.,As discussed in a previous study [33], an important variable that determines the role of scaffolding a kinase cascade is the amount of time required (τec) for an active kinase to encounter its downstream target. For simple diffusion, in three dimensions, where D is the diffusion constant and C is a typical concentration of kinases. Experiments indicate that τec is on the order of 10−4s–100 s [42]. Our studies focused on these experimentally relevant conditions. Steady-state values are reported. The system is first placed in a random configuration and simulations are allowed to “equilibrate” by letting the dynamics evolve to a time much larger than the time it takes for a kinase to diffuse the length of the simulation box. Kinases that are inactive at time t′ are tagged and waiting times are observed at time t+t′ (i.e. statistics are collected for the times at which the kinases become activated), and t′ is chosen to be a time longer than the time required for equilibration of the Monte Carlo trajectory. Calculation of Statistical Quantities Signaling dynamics can be defined microscopically as the distribution of times at which an individual kinase among of pool of available kinases becomes activated. Therefore, we quantify signaling dynamics by first considering the survival probability S(t). S(t) gives the probability that a particular kinase among the pool of signaling molecules has not been activated at time t provided that it was inactive at time t=0. S(t) is a two time point autocorrelation function:The survival probability is related to other dynamical properties; for instance, it can be related to a waiting time probability density function or first passage time distribution, f(t), in the following way: S(t) is the complement of the cumulative probability distribution of the first passage time. S(t) is computed from the simulations by integrating f(t). Such a calculation is analogous to the data obtained from a single molecule experiment that measures the statistics of enzyme dynamics [46]. This distribution of waiting times underlies the intrinsic duration of signal propagation in a protein kinase cascade—the decay of such a quantity is a measure of how fast the signaling cascade responds to stimuli. Important to note is that this quantity gives information only on the timing of the signal and not on its final magnitude. We also consider the product of the survival probability with the probability that a kinase in the pool of signaling molecules is active in the steady state, Power spectra were computed numerically. Real and imaginary parts of the Fourier transform were obtained from numerical integration using the trapezoidal rule with a step size Δt=0.001. P(ω) is calculated by squaring the real and imaginary parts of X(ω) P(ω) was sampled at N=100 logarithmically spaced (i.e., ωmax=ω0 (10δ(n−1)); n[1,100] so that ) angular frequencies beginning at: , where T is the total length of the autocorrelation function. | |||||||||||||||||||||||||
Footnotes The authors have declared that no competing interests exist. This work was funded through an NIH Director's Pioneer Award and NIH grant PO1 AI071195-01 to AKC. | |||||||||||||||||||||||||
References 1. Pawson, T; Scott, JD. Signaling through scaffold, anchoring, and adaptor proteins. Science. 1997;278:2075–2080. [PubMed] 2. Burack, WR; Shaw, AS. Signal transduction: hanging on a scaffold. Curr Opin Cell Biol. 2000;12:211–216. [PubMed] 3. Morrison, DK; Davis, RJ. Regulation of map kinase signaling modules by scaffold proteins in mammals. Ann Rev Cell Dev Biol. 2003;19:91–118. [PubMed] 4. Dard, N; Peter, M. Scaffold proteins in MAP kinase signaling: more than simple passive activating platforms. Bioessays. 2006;28:146–156. [PubMed] 5. Kolch, W. Coordinating ERK/MAPK signalling through scaffolds and inhibitors. Nat Rev Mol Cell Biol. 2005;6:827–837. [PubMed] 6. Pawson, T. Dynamic control of signaling by modular adaptor proteins. Curr Opin Cell Biol. 2007;19:112–116. [PubMed] 7. Dodge-Kafka, KL; Soughayer, J; Pare, GC; Michel, JJC; Langeberg, LK, et al. The protein kinase A anchoring protein mAKAP coordinates two integrated cAMP effector pathways. Nature. 2005;437:574–578. [PubMed] 8. Nguyen, A; Burack, WR; Stock, JL; Kortum, R; Chaika, OV, et al. Kinase suppressor of Ras (KSR) is a scaffold which facilitates mitogen-activated protein kinase activation in vivo. Mol Cell Biol. 2002;22:3035–3045. [PubMed] 9. Kortum, RL; Lewis, RE. The molecular scaffold KSR1 regulates the proliferative and oncogenic potential of cells. Mol Cell Biol. 2004;24:4407–4416. [PubMed] 10. Kortum, RL; Costanzo, DL; Haferbier, J; Schreiner, SJ; Razidlo, GL, et al. The molecular scaffold kinase suppressor of Ras 1 (KSR1) regulates adipogenesis. Mol Cell Biol. 2005;25:7592–7604. [PubMed] 11. Bhattacharyya, RP; Remenyi, A; Good, MC; Bashor, CJ; Falick, AM, et al. The Ste5 scaffold allosterically modulates signaling output of the yeast mating pathway. Science. 2006;311:822–826. [PubMed] 12. Park, SH; Zarrinpar, A; Lim, WA. Rewiring MAP kinase pathways using alternative scaffold assembly mechanisms. Science. 2003;299:1061–1064. [PubMed] 13. Scott, K; Zuker, CS. Assembly of the Drosophila phototransduction cascade into a signalling complex shapes elementary responses. Nature. 1998;395:805–808. [PubMed] 14. Smith, FD; Langeberg, LK; Scott, JD. The where's and when's of kinase anchoring. Trends Biochem Sci. 2006;31:316–323. [PubMed] 15. Bashor, CJ; Helman, NC; Yan, S; Lim, WA. Using Engineered Scaffold Interactions to Reshape MAP Kinase Pathway Signaling Dynamics. Science. 2008;319:1539–1543. [PubMed] 16. Heinrich, R; Neel, BG; Rapoport, TA. Mathematical models of protein kinase signal transduction. Mol Cell. 2002;9:957–970. [PubMed] 17. Marshall, CJ. Specificity of Receptor Tyrosine Kinase Signaling - Transient Versus Sustained Extracellular Signal-Regulated Kinase Activation. Cell. 1995;80:179–185. [PubMed] 18. Murphy, LO; Blenis, J. MAPK signal specificity: the right place at the right time. Trends Biochem Sci. 2006;31:268–275. [PubMed] 19. Murphy, LO; Smith, S; Chen, RH; Fingar, DC; Blenis, J. Molecular interpretation of ERK signal duration by immediate early gene products. Nat Cell Biol. 2002;4:556–564. [PubMed] 20. Dolmetsch, RE; Lewis, RS; Goodnow, CC; Healy, JI. Differential activation of transcription factors induced by Ca2+ response amplitude and duration. Nature. 1997;386:855–858. [PubMed] 21. Chen, LF; Fischle, W; Verdin, E; Greene, WC. Duration of nuclear NF-kappa B action regulated by reversible acetylation. Science. 2001;293:1653–1657. [PubMed] 22. Paul, S; Nairn, AC; Wang, P; Lombroso, PJ. NMDA-mediated activation of the tyrosine phosphatase STEP regulates the duration of ERK signaling. Nature Neuroscience. 2003;6:34–42. 23. Chen, YR; Wang, XP; Templeton, D; Davis, RJ; Tan, TH. The role of c-Jun N-terminal kinase (JNK) in apoptosis induced by ultraviolet C and gamma radiation - Duration of JNK activation may determine cell death and proliferation. J Biol Chem. 1996;271:31929–31936. [PubMed] 24. Daniels, MA; Teixeiro, E; Gill, J; Hausmann, B; Roubaty, D, et al. Thymic selection threshold defined by compartmentalization of Ras/MAPK signalling. Nature. 2006;444:724–729. [PubMed] 25. Sasagawa, S; Ozaki, Y; Fujita, K; Kuroda, S. Prediction and validation of the distinct dynamics of transient and sustained ERK activation. Nat Cell Biol. 2005;7:365-U331. [PubMed] 26. Schoeberl, B; Eichler-Jonsson, C; Gilles, ED; Muller, G. Computational modeling of the dynamics of the MAP kinase cascade activated by surface and internalized EGF receptors. Nat Biotechnol. 2002;20:370–375. [PubMed] 27. Kao, SC; Jaiswal, RK; Kolch, W; Landreth, GE. Identification of the mechanisms regulating the differential activation of the MAPK cascade by epidermal growth factor and nerve growth factor in PC12 cells. J Biol Chem. 2001;276:18169–18177. [PubMed] 28. Boykevisch, S; Zhao, C; Sondermann, H; Philippidou, P; Halegoua, S, et al. Regulation of Ras signaling dynamics by Sos-mediated positive feedback. Curr Biol. 2006;16:2173–2179. [PubMed] 29. Locasale, JW. Allovalency revisited: An analysis of multisite phosphorylation and substrate rebinding. J Chem Phys. 2008;128 30. Santos, SDM; Verveer, PJ; Bastiaens, PIH. Growth factor-induced MAPK network topology shapes Erk response determining PC-12 cell fate. Nat Cell Biol. 2007;9:324-U139. [PubMed] 31. Kiyatkin, A; Aksamitiene, E; Markevich, NI; Borisov, NM; Hoek, JB, et al. Scaffolding protein Grb2-associated binder 1 sustains epidermal growth factor-induced mitogenic and survival signaling by multiple positive feedback loops. J Biol Chem. 2006;281:19925–19938. [PubMed] 32. Kholodenko, BN. Cell-signalling dynamics in time and space. Nat Rev Mol Cell Biol. 2006;7:165–176. [PubMed] 33. Locasale, JW; Shaw, AS; Chakraborty, AK. Scaffold Proteins Confer Diverse Regulatory Properties to Protein Kinase Cascades. Proc Natl Acad Sci U S A. 2007;104:13307–13312. [PubMed] 34. Levchenko, A; Bruck, J; Sternberg, PW. Scaffold proteins may biphasically affect the levels of mitogen-activated protein kinase signaling and reduce its threshold properties. Proc Natl Acad Sci U S A. 2000;97:5818–5823. [PubMed] 35. Lu, T; Shen, TY; Zong, CH; Hasty, J; Wolynes, PG. Statistics of cellular signal transduction as a race to the nucleus by multiple random walkers in compartment/phosphorylation space. Proc Natl Acad Sci U S A. 2006;103:16752–16757. [PubMed] 36. Lan, YH; Papoian, GA. Stochastic resonant signaling in enzyme cascades. Phys Rev Lett. 2007;98 37. Redner, S. A Guide To First Passage Processes: Cambridge University Press. 2001 38. Chang, LF; Karin, M. Mammalian MAP kinase signalling cascades. Nature. 2001;410:37–40. [PubMed] 39. Burack, WR; Sturgill, TW. The activating dual phosphorylation of MAPK by MEK is nonprocessive. Biochemistry. 1997;36:5929–5933. [PubMed] 40. Kholodenko, BN. Negative feedback and ultrasensitivity can bring about oscillations in the mitogen-activated protein kinase cascades. Eur J Biochem. 2000;267:1583–1588. [PubMed] 41. Lauffenburger, DA; Linderman, JA. Receptors: Models for Binding, Trafficking, and Signaling. Oxford University Press. 1993 42. Arrio-Dupont, M; Foucault, G; Vacher, M; Devaux, PF; Cribier, S. Translational diffusion of globular proteins in the cytoplasm of cultured muscle cells. Biophys J. 2000;78:901–907. [PubMed] 43. Tanase-Nicola, S; Warren, PB; ten Wolde, PR. Signal detection, modularity, and the correlation between extrinsic and intrinsic noise in biochemical networks. Phys Rev Lett. 2006;97 44. van Zon, JS; Morelli, MJ; Tanase-Nicola, S; ten Wolde, PR. Diffusion of transcription factors can drastically enhance the noise in gene expression. Biophys J. 2006;91:4350–4367. [PubMed] 45. Warren, PB; Tanase-Nicola, S; ten Wolde, PR. Exact results for noise power spectra in linear biochemical reaction networks. J Chem Phys. 2006;125:144904. [PubMed] 46. Lu, HP; Xun, LY; Xie, XS. Single-molecule enzymatic dynamics. Science. 1998;282:1877–1882. [PubMed] 47. Yu, J; Xiao, J; Ren, XJ; Lao, KQ; Xie, XS. Probing gene expression in live cells, one protein molecule at a time. Science. 2006;311:1600–1603. [PubMed] 48. Slaughter, BD; Schwartz, JW; Li, R. Mapping dynamic protein interactions in MAP kinase signaling using live-cell fluorescence fluctuation spectroscopy and imaging. Proc Natl Acad Sci U S A. 2007;104:20320–20325. [PubMed] 49. Chandler, D. New York, NY: Oxford University Press; 1987. Introduction to Modern Statistical Mechanics. 50. Frenkel, DSB. Boston: Academic Press; 2002. Understanding Molecular Simulation. From Algorithms to Applications. 51. Ghaemmaghami, S; Huh, W; Bower, K; Howson, RW; Belle, A, et al. Global analysis of protein expression in yeast. Nature. 2003;425:737–741. [PubMed] 52. Ferrell, JE. Tripping the switch fantastic: How a protein kinase cascade can convert graded inputs into switch-like outputs. Trends Biochem Sci. 1996;21:460–466. [PubMed] | |||||||||||||||||||||||||