How cells process information: quantification of spatiotemporal signaling dynamics

Abstract

Arguably, one of the foremost distinctions between life and non-living matter is the ability to sense environmental changes and respond appropriately--an ability that is invested in every living cell. Within a single cell, this function is largely carried out by networks of signaling molecules. However, the details of how signaling networks help cells make complicated decisions are still not clear. For instance, how do cells read graded, analog stress signals but convert them into digital live-or-die responses? The answer to such questions may originate from the fact that signaling molecules are not static but dynamic entities, changing in numbers and activity over time and space. In the past two decades, researchers have been able to experimentally monitor signaling dynamics and use mathematical techniques to quantify and abstract general principles of how cells process information. In this review, the authors first introduce and discuss various experimental and computational methodologies that have been used to study signaling dynamics. The authors then discuss the different types of temporal dynamics such as oscillations and bistability that can be exhibited by signaling systems and highlight studies that have investigated such dynamics in physiological settings. Finally, the authors illustrate the role of spatial compartmentalization in regulating cellular responses with examples of second-messenger signaling in cardiac myocytes.

Figure 1

Fluorescence-based biosensors. (A) A PIP₃ biosensor contains the PH domain of Akt tagged to EGFP. An increase in the synthesis of PIP₃ leads to recruitment of the biosensor to the membrane and hence increased fluorescence in the membrane. (B) The calcium biosensor, GCaMP contains a circularly permuted EGFP (cpEGFP) sandwiched by a Calmodulin (CaM) peptide and the CaM-binding M13 fragment of myosin light chain kinase. Increase in calcium concentration leads to binding of calcium-CaM to M13 thereby filling up a cavity in the cpEGFP and hence increasing fluorescence intensity. (C) A generic genetically encoded FRET-based kinase activity biosensor contains a target substrate peptide and phospho-amino acid binding peptide flanked by a cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP). Increase in kinase activity leads to phosphorylation of the target substrate domain. Subsequent conformational change upon binding of kinase substrate domain to the phospho-binding domain brings the two fluorescent proteins together resulting in increased FRET.

Figure 2

Modeling-driven dissection of signaling networks. Known non-linear connections (denoted by dark lines) in a network comprising three components A, B, and C are represented by a mathematical model. In this network, A activates B, B activates C, and C inhibits A in a negative feedback manner. Simulations of the model can then be matched with the experimental data. Any discrepancies between experiments and model can indicate the need for further improvements to the model, typically involving a revision of parameter values. Simulations can also reveal the presence of new connections (shown in dotted lines) leading to the formulation of a more complete network which can better explain experimental results.

Figure 3

A representative compartmentalized model of spatial and temporal dynamics of the activity of a species, A. The cell is divided into two homogenous compartments: the cytosol and the nucleus. A is generated and degraded in the cytosol at rates k_fc and k_rc, respectively, and is generated and degraded in the nucleus at rates k_fn and k_rn, respectively. Further, A is assumed to shuttle between the two compartments at effective rates of k_cn and k_nc. Separate ODEs for each compartment are written by summing up the reaction and diffusion rates.

Figure 4

Temporal dynamics of signaling circuits with negative feedback loops in a representative system comprising two molecules, X and Y. (A) The presence of a negative feedback loop alone can produce a transient adaptative output signal in response to a “step-input” signal. (B) Negative feedback loop coupled with time-delay [represented explicitly as Y(t) = X(t − τ)] produces a damping oscillation in response to a “step-input” signal. (C) Negative feedback with time-delay (represented by the presence of an extra species, Z in the circuit) coupled with additional positive feedback loops results in sustained oscillations in response to a “step-input” signal. Positive feedback loops are denoted in green. Negative feedback loops are denoted in red.

Figure 5

Bistability in a circuit with a positive feedback. In a circuit with positive feedback loops, multiple steady state responses are possible. The simplest case with two steady states is referred to as a bistable system. In this schematic, the system can exist in either of two steady states, the “ON” state or the “OFF” state, for a particular range of values of a certain kinetic parameter, k (say). A hallmark of bistable systems is the phenomenon of hysteresis, in which, the system response depends not only on current conditions but also the path taken to reach the current conditions. In this case, the system transitions from the “OFF” state to the “ON” state at the parameter value k₁, as k is increased. However, the system transitions from the “ON” state to the “OFF” state at the parameter value, k₂ different from k₁, as k is decreased.