These simulations and experiments show a tightly coordinated relationship between four factors in the origins and dynamics of spatial domains. These factors are cell shape, subcellular localization of components, network topology, and the kinetic parameters for biochemical reactions. Intracellular signals most often arise from the activity of plasma membrane-bound components. The local shape of the region involved in the production of the signal predisposes how these signals spread. This effect of cell shape implies that mechanical forces at the plasma membrane will control the flow of spatial information through intracellular networks. In the case of a cell with projections like a neuron, the dendritic structure, with its high surface-to-volume ratio, favors the local accumulation of membrane-generated signals. Thus, the physical constraints for microdomain formation comes from cell shape (and presumably cell mechanics) and the resulting surface to volume effect (). These physical constraints are, however, insufficient to permit the transmission of spatial information to downstream components. For transmission of spatial information, both the topology of the network and the kinetic parameters for the biochemical reactions need to be appropriately tuned. In the system we have studied, a feed-foward loop is necessary to transmit both the information regarding the extent of the activity as well as the spatial range (microdomain characteristics). Network topology by itself is also insufficient, since the appropriate kinetic parameters for key reac-tants are essential for the propagation of spatial information. Thus, the dynamics of the microdomains of signaling components is truly a systems property arising from the interplay of the physical and chemical characteristics of a cell ().
Integration of the Physical Determinants and the Biochemical Characteristics Required for the Transmittal of Spatial Information through Signaling Networks
The development of live-cell imaging techniques has demonstrated the realtime existence of microdomains of components within the cell. Are such microdomains relevant to the specificity of physiological effects? In hippocampal neurons, the answer is yes. For example, MAPK phosphorylates A-type potassium channels Kv4.2 (Morozov et al., 2003
), which are predominantly expressed in the distal dendrites as compared to the cell bodies (Maletic-Savatic et al., 1995
; Sheng et al., 1992
). Local increases in dendritic MAPK, even when the entire cell is stimulated, may allow for inhibition of Kv4.2 in a spatially defined manner. This process could facilitate depolarization at particular dendrites through dendritic action potentials of local origin or by routing back-propagating action potentials, both of which have been implicated in long-term potentiation of synaptic transmission (Remy and Spruston, 2007
; Watanabe et al., 2002
; Losonczy et al., 2008
). Moreover, cAMP microdomains could influence the scaling of distal and proximal synaptic contributions to the integrated response at the cell body, and in this way help to determine the pattern of firing generated by inputs to different regions of the dendritic arbor. For example, local increases in the concentration of cAMP would shift the activation range of HCN-type cation channels to less negative membrane potentials and thus be expected to maintain the inward current through these channels during an excitatory postsynaptic potential (Magee, 1999
). Thus, the MAPK- and cAMP-mediated effects on Kv4.2 and HCN channels, respectively, could have important consequences for the spatial specification of synaptic plasticity.
From these simulations, it becomes apparent that defining the underlying causes for the formation and dissipation of spatial domains of signaling components can be very useful in understanding how the interplay between structural changes at the cellular level and biochemical changes can lead to altered physiological and pathophysiological responses. For spatially restricted systems such as the plasma membrane-bound β-adrenergic receptor-Gs-adenylyl cyclase, surface-to-volume ratios will play a role in the formation of microdomains. Our simulations show that in real geometries of cells like neurons, diffusion appears not to be a major contributor of microdomain dynamics. In contrast, negative regulators such as phosphodiesterases and phosphatases as part of regulatory loops are major determinants of microdomain dynamics. Our results show that these negative regulators not only modulate dynamics of microdomains of upstream components but also propagate spatial information through signaling networks. The ability to transmit spatial information arises from the modulation of activity that the negative regulators are subjected to as part of a regulatory loop. Changing the net reaction balance between production and degradation of signaling components at any location affects the overall size of the microdomain. Thus, reaction balance at the various parts of the signaling network within the overall reaction diffusion system can be a critical driver of spatial patterns within cells. Here, the importance of kinetic parameters for key reactions within the network should also be emphasized. Evolution may have selected these biochemical parameters so as to achieve dynamic systems level spatial behaviors. The role of diffusible negative regulators in defining the boundaries of microdomains adds to the list of important systems-level properties that can be attributed to negative regulators. Previous studies from our laboratory had shown the critical role of negative regulators such as PP1 in the gating of the CaMKII signal in LTP (Blitzer et al., 1998
) and MKP in the design of a flexible MAPK switch (Bhalla et al., 2002
In conclusion, these studies show that the dynamics of microdomains of signaling components in cells arises from shape-constrained net-reaction balance within regulatory loops. The microdomains of signaling components provide compelling evidence for the inhomogeneous nature of coupled biochemical reactions within cells. Since such coupled biochemical reactions can store information (Bhalla and Iyengar, 1999
), the existence of microdomains indicates that such information may be stored in specific locations within the cell.