Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Cellular differentiation wikipedia , lookup
Cell growth wikipedia , lookup
Cytokinesis wikipedia , lookup
G protein–coupled receptor wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
NMDA receptor wikipedia , lookup
List of types of proteins wikipedia , lookup
P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 Annu. Rev. Physiol. 2001. 63:847–69 c 2001 by Annual Reviews. All rights reserved Copyright MAINTAINING THE STABILITY OF NEURAL FUNCTION: A Homeostatic Hypothesis Graeme W Davis1 and Ilya Bezprozvanny2 1Department of Biochemistry, University of California, San Francisco, San Francisco, California 94143-0448; e-mail: [email protected]; 2Department of Physiology, University of Texas Southwestern Medical Center, Dallas, Texas 75390-9040; e-mail: [email protected] Key Words learning, synaptic plasticity, calcium, synaptogenesis, growth ■ Abstract The precise regulation of neural excitability is essential for proper nerve cell, neural circuit, and nervous system function. During postembryonic development and throughout life, neurons are challenged with perturbations that can alter excitability, including changes in cell size, innervation, and synaptic input. Numerous experiments demonstrate that neurons are able to compensate for these types of perturbation and maintain appropriate levels of excitation. The mechanisms of compensation are diverse, including regulated changes to synaptic size, synaptic strength, and ion channel function in the plasma membrane. These data are evidence for homeostatic regulatory systems that control neural excitability. A model of neural homeostasis suggests that information about cell activity, cell size, and innervation is fed into a system of cellular monitors. Intracellular- and intercellular-signaling systems transduce this information into regulated changes in synaptic and ion channel function. This review discusses evidence for such a model of homeostatic regulation in the nervous system. INTRODUCTION An increasing body of experimental evidence demonstrates that neurons have the capacity to respond to perturbations such as inappropriate synaptic function or altered innervation and thereby maintain their function within a normal physiological range (1). These experiments suggest that neurons have the capacity to monitor their own activity levels and transduce this information into regulated changes in excitability. For example, a neuron or muscle that is challenged by excessive innervation can regulate the function of each impinging synaptic terminal, decreasing the strength of each synapse such that the sum total excitation is normal despite the presence of abnormal innervation (1–3). The mechanisms that monitor excitation and maintain the functional properties of neurons and muscle are by definition homeostatic. These experiments have led to an emerging theory 0066-4278/01/0315-0847$14.00 847 P1: FXZ January 20, 2001 848 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY of neural homeostasis that may provide a basis for understanding how the stability of neural circuitry is maintained. Homeostatic regulation of neural activity could maintain the robust function of the nervous system during the restructuring and refinement of neural circuits that occur during postembryonic development. Homeostatic mechanisms may also ensure that neurons in all areas of the brain continue to function appropriately throughout life. Most neural circuits remain robust throughout life. During this time, most cellular components are recycled many times. One implication is that synaptic connections, once established, must be continually maintained (1, 4). Homeostasis is one of the most basic cellular processes by which a cell responds to a change in the intracellular or extracellular environment and maintains a constant physiology. Homeostatic processes have been revealed in systems ranging from the cellular response to changing extracellular osmolarity to the system-level maintenance of blood pressure (5). In the nervous system, the maintenance of nerve cell excitation within reasonable physiological limits is a form of electrical homeostasis ensuring that an individual neuron is neither induced to fire action potentials excessively nor silenced. Our ability to learn and adapt to our environment is thought to require changes in synaptic connectivity and neural excitation (6). It is essential, therefore, that the homeostatic regulation of neural activity does not preclude activity-dependent modification of neural circuitry. Rather, homeostatic mechanisms might establish limits beyond which activity-dependent changes may not reasonably modify cellular activity. As such, homeostatic regulation will ensure the stability of neural function without impairing our capacity for neural change. Thus whereas synaptic competition, activity-dependent synaptic plasticity, and neurotrophism encompass mechanisms that can generate precise change in synaptic structure and function (7–9), homeostasis is thought to constrain this change within reasonable physiological limits. The mechanisms of homeostatic regulation in the nervous system remain to be characterized in any detail. The experimental data suggest that there are many different types of regulation. Regulatory changes can be cell autonomous or can involve intercellular-signaling events. Homeostatic changes in ion channel density and neurotransmitter receptor function have been observed that allow neurons to modulate excitability cell autonomously. In other examples, trans-synaptic signals have been implicated in the homeostatic regulation of synaptic function (1, 4). It is also possible that homeostasis may be expressed at a circuit or system level. The types of perturbation to which neurons have been demonstrated to respond in a homeostatic manner are also diverse. Homeostatic compensation has been demonstrated in response to altered innervation, altered synaptic function, and altered cellular depolarization (1, 4). In addition, it is hypothesized that homeostatic mechanisms allow neurons and muscle to remain functional during the dramatic changes in cell shape and size that occur in normal development. The presence of multiple, redundant types of homeostatic signaling may be a consequence of the many different perturbations to which a cell must respond. Alternatively, such P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 849 mechanistic diversity may ensure a robust regulatory system. The implicit importance of homeostasis to appropriate neural function suggests that there are links to neural disease. Homeostatic regulation has been observed as a response to disease and damage in the nervous system. Ultimately, a thorough understanding of homeostatic signaling at a cellular and molecular level will be required before impaired homeostatic systems can be investigated as a cause of neural disease. EVIDENCE FOR HOMEOSTATIC REGULATION IN THE NERVOUS SYSTEM In systems ranging from the invertebrate neuromuscular junction (NMJ) to synapses in the vertebrate cortex, increasing evidence supports the hypothesis that homeostatic processes maintain stable levels of neuronal activity. In each experiment, synaptic or neuronal functions are examined before and after experimental perturbations that alter innervation, excitation, or cell growth. The data demonstrate that neurons and muscle have a remarkable capacity to respond to these perturbations and return to function within their normal physiological range—an indication of homeostatic regulation. Evidence for homeostatic regulation in the nervous system has been the focus of two recent reviews (1, 4). In this section we highlight several additional experiments and add further discussion. Homeostatic Response to Altered Post-Synaptic Excitability at the Vertebrate and Invertebrate Neuromuscular Junction Two independent genetic manipulations have been used to impair post-synaptic excitation at the Drosophila NMJ. The Drosophila NMJ is glutamatergic. Electrophysiological analysis of glutamate receptor knockouts reveals a reduction in quantal size (most likely owing to the altered conductance of the channels formed by the remaining glutamate receptor subunits). Remarkably, impaired post-synaptic receptor function is precisely compensated for by increased presynaptic transmitter release, resulting in normal muscle depolarization (10). Identical results were obtained by muscle-specific overexpression of protein kinase A. Increased protein kinase A activity decreases quantal size (11). Again, presynaptic release precisely compensates for reduced quantal size, and the evoked depolarization of the muscle remains normal. Thus two independent genetic manipulations that alter post-synaptic glutamate receptor function demonstrate that presynaptic transmitter release is modulated in response to changes in post-synaptic excitability. In each example, it is hypothesized that homeostatic-signaling systems modulate presynaptic release via a retrograde signal from muscle to nerve. Either this retrograde signal may enhance release, or these effects may be achieved by the repression of a negative regulator of presynaptic release (12). A similar homeostatic change has been observed at the vertebrate NMJ in experiments examining neuromuscular function in neuregulin-knockout mice. In P1: FXZ January 20, 2001 850 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY heterozygous knockout mice, post-synaptic acetylcholine receptor clustering is reduced, and, as a result, there is a decrease in the post-synaptic sensitivity to acetylcholine (13). As in Drosophila, there is a compensatory increase in presynaptic transmitter release at these synapses. Thus despite the large safety factor at the vertebrate NMJ, homeostatic regulatory mechanisms can be revealed. This raises the possibility that such homeostatic compensation may play a significant role in the maintenance of the vertebrate neuromuscular synapse during normal growth and function. The observed compensatory increase in synaptic efficacy, at both the Drosophila and vertebrate NMJs, is achieved during development. Therefore, the changes could be mediated by an increased rate of active-zone insertion. Alternatively, a portion of active zones at the wild-type synapse might normally be silent during development (14). Unveiling silent synapses could act as a compensatory response. Finally, a change in presynaptic release at the wild-type complement of active zones is also consistent with the observed data. Homeostatic Response to Altered Post-Synaptic Excitability at Central Synapses Neurons of the lobster stomatogastric ganglion show a remarkable ability to regenerate their stereotypic bursting properties after acute isolation from their normal synaptic circuitry (15, 16). In these experiments, single neurons are isolated from their surrounding synaptic circuitry and placed in single-cell culture. Remarkably, these single cells regenerate their endogenous bursting properties in the absence of patterned synaptic input by modifying the balance of their voltage-gated channel densities. This homeostatic modification of channel density can be blocked by supplying patterned activity, and it can also be blocked by preventing calcium entry to the neuron (15–17). The implication is that calcium entry is utilized as a means to monitor altered activity levels in the neuron. In this case, since cells are isolated from their surrounding synaptic connections, the source of calcium is likely to be calcium channels. The ability of calcium to act as an activity sensor is discussed at length in later sections. An alternate form of compensation has been observed in vitro at vertebrate central synapses. Cortical synapses in culture have the capacity to modulate postsynaptic sensitivity to neurotransmitters in response to either increased or decreased excitation (18, 19). In these experiments, increased excitation is achieved by chronically blocking inhibitory synaptic input. Reduced excitation is achieved by bathing cultures in tetrodotoxin (TTX). A remarkable feature of the observed compensation is the demonstration that quantal size not only compensates for altered excitation (increased quantal size compensates for TTX treatment and vice versa) but that compensation is achieved by factorial scaling of the size of the quantal current (18). These results implicate mechanisms of receptor insertion as a form of compensation (20). More recently, the activity of brain-derived neurotrophic factors has been implicated in this homeostatic response (21). P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 851 Homeostatic Response to Altered Innervation at Central and Neuromuscular Synapses The homeostatic control of neuronal or muscle function implies that the summed strength of all synaptic inputs should remain within certain bounds. According to this hypothesis, increasing the number of synaptic contacts to a single target should result in reduced synaptic efficacy at all (or a portion) of the synaptic terminals on a given target, and reducing the number of synaptic contacts should have the opposite effect. These phenomena have been observed at both central and neuromuscular synapses (1–3). At a vertebrate central synapse in vitro, focal recordings from individual boutons (a single active zone) demonstrate that total target innervation is inversely correlated with the efficacy of transmission at each individual active zone (3). Cells that receive increased innervation show decreased efficacy at each individual active zone and vice versa. The apparent compensatory response is achieved by an altered post-synaptic response to neurotransmitter (3). One possible interpretation is that the post-synaptic cell senses increased innervation, potentially through altered excitation, and this cell generates a compensatory change in synaptic efficacy. Alternatively, if quantal size is determined by post-synaptic receptor density, then the distribution of a fixed number of receptors across various numbers of synaptic inputs could, in theory, account for this result. Another test of the homeostatic response to altered innervation was recently achieved at the Drosophila NMJ (2). These experiments exploited a natural pattern of muscle innervation in which a pair of motoneurons each contacts two adjacent muscle targets. Innervation from these two motoneurons was biased onto one muscle target and away from the adjacent muscle target by overexpression of the cell adhesion molecule Fasciclin II (Fas II) exclusively in one of the two target muscles. The total bouton number was conserved such that one muscle received increased innervation (the muscle overexpressing Fas II) and the adjacent target received decreased innervation (having normal levels of Fas II expression) (22). Despite experimentally induced altered innervation (either an increase or decrease in bouton number), the muscles were normally depolarized, demonstrating a homeostatic response to altered innervation (Figure 1). At the target with fewer boutons, there was an increase in quantal size that produced a compensatory increase in muscle excitability, thereby generating normal depolarization. At the muscle target receiving increased innervation, there was a decrease in presynaptic release from each synaptic bouton, which allowed, once again, normal muscle depolarization. Taken together, these results demonstrate that homeostatic regulation of synaptic function is not only target specific, but it can be achieved through independent mechanisms at the terminals of a single neuron (2; Figure 1). Synaptic competition drives many of the rewiring events that occur during postembryonic development (8, 9). A remarkable feature of synaptic competition is the ability of the post-synaptic cell to remain functional despite the pruning P1: FXZ January 20, 2001 852 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY Figure 1 Homeostatic compensation for altered innervation at the Drosophila neuromuscular junction. (Top) A diagram of the motoneuron innervation to muscle 6 and muscle 7 in Drosophila. A single synaptic bouton is diagrammatically enlarged. Single boutons at the mature synapse contain multiple active zones (∼12 on average). (Bottom) Two images of the neuromuscular synapse are shown, the wild-type synapse at muscles 6 and 7 and an example of altered innervation after the overexpression of the cell adhesion molecule Fasciclin II only on muscle 6. The neuromuscular synapse is biased toward growth on muscle 6 and is biased away from muscle 7. At right, the results of quantal analysis are summarized for each muscle. Despite increased innervation at muscle 6, muscle depolarization is wild type. There is a compensatory decrease in per-bouton transmitter release resulting in normal total transmitter release. Despite reduced innervation at muscle 7, the muscle depolarization is wild type. There is a compensatory increase in quantal size that allows normal muscle depolarization despite fewer synaptic boutons. Thus two independent forms of compensation occur at the terminals of a single motoneuron contacting neighboring muscle targets. P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 853 and sprouting of presynaptic input(s). An interesting possibility is that synaptic competition is played out in a background of homeostatic regulation. Homeostatic Response to Cellular Growth The physiological changes caused by cellular growth are some of the most drastic perturbations that an individual cell will confront. During postembryonic development, a muscle cell can achieve a tremendous increase in size and mass (>50-fold growth in many vertebrate and invertebrate systems). Muscle has a requirement for homeostatic regulation of depolarization during this growth. As muscle size increases, input resistance decreases (23, 24). Decreased input resistance causes decreased depolarization for a given amount of synaptic current. Yet muscle must receive nearly constant levels of depolarization so that appropriate contractile properties are maintained. Developmental observation of neuromuscular transmission during a period of 40-fold growth at the crayfish NMJ demonstrates that constant levels of muscle depolarization are maintained by an increase in presynaptic transmitter release (achieved through growth of the motoneuron terminal) and an increase in the quantal size (23, 24). Thus compensatory changes are observed both pre- and post-synaptically to compensate for growth-related changes in muscle excitability. It has been proposed that a signal from muscle to nerve regulates presynaptic growth and function to compensate for the drop in muscle input resistance that occurs during growth (23–25). Recent experiments at the Drosophila NMJ provide experimental evidence supporting the presence of a homeostatic, retrograde-signaling system that couples synaptic development to muscle growth (1, 10). Several experiments demonstrate that neurons are able to preserve their unique response properties during profound cellular growth—a form of homeostatic regulation. An insect giant interneuron, for example, preserves its response properties to sensory input during an ∼50-fold increase in cell size by precisely coordinating the growth of dendrite diameter and dendrite length. This type of growth (termed ISO-electronic growth) alters the physical dimensions of the cell to preserve cellular response properties (26). There appear to be other solutions to this problem. As the dendrites of a central neuron grow, pre-existing sites of synaptic input become physically and electrically more distant from the soma. The presence of voltage-gated channels in the dendrite plasma membrane of vertebrate central neurons has been proposed as a mechanism to help propagate synaptic inputs made in distal regions of the dendritic tree to the site of synaptic integration (27, 28). It is not known whether these dendritic voltage-gated currents change in a coordinated manner during dendrite growth. In other instances, neurons appear to have evolved to take advantage of the changing response properties caused by growth to generate new functionality (29). Homeostasis During Disease and Injury The hypothesis that homeostatic regulation maintains the stability of neural function in the nervous system implies that this process should be revealed in the P1: FXZ January 20, 2001 854 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY etiology of neural disease. Myasthenia gravis affects the neuromuscular synapse, causing a progressive loss of post-synaptic acetylcholine receptors. Decreased receptor function is compensated for by increased presynaptic release at human myasthenic muscle (30, 31), a result that parallels experimental observations at the Drosophila NMJ (10, 11) and at the NMJ of neuregulin-knockout mice (13). Postpolio syndrome describes the gradual loss of neuromuscular function in a current population of people who survived polio as children (32). The neuromuscular synapses of these survivors are regenerated synapses that can be smaller than normal but are able to function at wild-type levels. In some cases, there is evidence of a compensatory enhancement of presynaptic release at these terminals. The cause of postpolio (the regression of these regenerated synapses after many years of nearly wild-type function) is not well understood, but it may include a gradual loss of compensatory mechanisms that are normally required to maintain wild-type function at regenerated, postpolio, neuromuscular synapses (32). The disruption of homeostatic regulation might adversely alter the function of neural circuits, resulting in seizure-like activity or degeneration. Moreover, homeostatic mechanisms that normally ensure a balance of synaptic efficacy in the central nervous system might lead to various clinical syndromes of the injured central nervous system (1). For example, spinal cord injury leads to the removal of major classes of descending and ascending inputs. As a result, homeostatic mechanisms controlling synaptic strength might lead to the observed, inappropriate strengthening of persisting inputs, resulting in debilitating post-traumatic disorders. As another example, lesions of mature axons in a hippocampal slice induce numerous changes, including the formation of new axon collaterals, an increase in synaptic connectivity between neurons, and hyperexcitability (33). These results implicate a homeostatic response to injury that could be deleterious and may be related to the cause of post-traumatic epilepsy (33). We understand very little about the underlying mechanisms of the diverse homeostatic regulatory mechanisms that have been observed. We can only speculate on the relationship between homeostatic signaling and the basis for a broad spectrum of other neurological disorders and diseases. Model for Homeostatic Regulation The types of perturbations to which neurons and muscle are known to respond in a homeostatic manner include changes in cell growth, innervation, and excitability. Each of these perturbations ultimately impacts post-synaptic depolarization. Thus it seems likely that neurons and muscle are endowed with the capacity to monitor levels of activity or depolarization to initiate a compensatory response. It is also possible that neurons and muscle can monitor cell growth and innervation independently of the effects of these parameters on cellular depolarization. This is suggested by the observation that perturbations producing the same effects on postsynaptic depolarization do not always initiate the same compensatory response. For example, there are different responses to decreased innervation vs decreased P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 855 post-synaptic receptor sensitivity at the Drosophila NMJ. Furthermore, increased innervation and increased post-synaptic receptor sensitivity will both increase postsynaptic depolarization, but increased receptor sensitivity is not compensated for, and the muscles become super excited (1, 2, 11). The experiments and developmental observations made to date suggest a model of homeostatic regulation in the nervous system (Figure 2). Information about cell activity, cell size, and innervation is fed into a system of cellular monitors. Cellular monitors are likely to reside in the soma so that summated activity or size can be monitored. Important components may also be localized at synaptic sites to achieve specificity. Information regarding activity, size, or growth is then transduced from various monitors to intracellular signals (potentially requiring transcriptional regulation) that act on effectors such as post-synaptic transmitter receptors or ion channels in the plasma membrane. Modulation of presynaptic release via retrograde signaling, as demonstrated at the Drosophila NMJ, will require trans-synaptic signaling to receptors and effectors at presynaptic terminals. The mechanism by which various monitors are coupled to compensatory signaling may be continuous, or a cell may establish thresholds beyond which changes Figure 2 A model for homeostatic regulation at the Drosophila neuromuscular junction. A homeostatic regulatory system that assures proper muscle depolarization will include a system of post-synaptic monitors of muscle depolarization. It is possible that muscle size and synapse sizes are monitored simultaneously through their respective effects on muscle depolarization. It is equally likely that there will be independent systems that can monitor muscle size and synapse sizes independent of each other and independent of muscle depolarization (thin versus thick arrows). The assimilation of this information into meaningful trans-synaptic signals will likely include both positive and negative regulators. P1: FXZ January 20, 2001 856 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY in activity, size, or innervation are not tolerated. The next sections describe evidence for cellular monitors that may transduce activity or growth into meaningful chemical signaling within or between cells. CELLULAR MONITORS OF CELL MASS AND GROWTH The connection between cellular growth (increased cell mass) and other cellular processes is a well-established feature of the developmental program. Experiments at the intersection of cell cycle control and embryonic development demonstrate that cells have the capacity to monitor cell mass and control both cellular growth and cellular proliferation (34). One example is the presence of a cell mass checkpoint during progression through the cell cycle. Cells must attain a certain mass to progress through the cell cycle (35). If cells in culture are limited in their growth, they stop proliferating, and synchronous proliferation begins upon the addition of nutrients or growth factors. Progress through the cell cycle can be driven experimentally by activation of cell cycle regulators (34). However, when division occurs without cellular growth, the cells continue to decrease in size until catastrophe. The systems that monitor cell size can be coupled to intercellular signaling and regulatory events, which demonstrates that monitors of cell mass can be coupled to other cellular regulatory systems, including intercellular signaling (34). Thus it seems plausible that monitors of cell mass may also be coupled to homeostatic systems that regulate neuronal and muscle excitability during growth. The presence of cell mass monitors raises the issue of the extent to which differences in neuronal size and complexity are impacted by cell-autonomous monitors of growth vs extrinsic factors that influence, for example, dendrite outgrowth and branching. It is almost certain that both cell-autonomous and exogenous factors are involved (36). CELLULAR MONITORS OF ACTIVITY Intracellular calcium levels can provide a sensitive readout of cellular activity in both neurons and muscle (37–39). Activity-induced calcium oscillations can be generated by calcium entry through neurotransmitter receptors, voltage-gated calcium channels, and calcium release from intracellular stores. Each source of calcium can provide temporal information about changes in activity and spatial information from different subcellular localization and diffusion within the cell. It has been demonstrated in a number of systems that information can be encoded in the frequency, amplitude, duration, and number of calcium oscillations (39–42). There are many questions related to calcium-dependent monitors of neural activity. How are calcium transients reliably encoded in the intracellular-signaling machinery (what are the calcium sensors)? If homeostatic regulatory machinery acts downstream of calcium sensors within the cell, how is flexibility achieved P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 857 such that each cell can establish different calcium-sensitive responses or a similar homeostatic response to a very different calcium signal? For example, a tonicfiring neuron has different requirements for homeostatic regulation compared with a phasic-firing neuron. Furthermore, over time the properties of a cell may change and the homeostatic system that monitors activity must, therefore, be adaptable. The hypothesis that calcium-dependent signaling could be the basis for a homeostatic monitor of neuronal activity was originally proposed in experiments that modeled the changes in ion channel density observed in stomatogastric neurons of the lobster after acute isolation in cell culture (15, 43). A series of calciumsensitive frequency filters were modeled that were sensitive to different frequencies of calcium oscillation. By coupling information from the calcium filters to ion channel densities, the authors were able to effectively model the homeostatic regulation of ion channel density that was observed experimentally. The use of calcium frequency filters takes advantage of information encoded in the timing of calcium oscillations and is therefore more elaborate than invoking calcium set points beyond which changes in calcium concentration are not tolerated. CamKII Calcium-calmodulin-dependent protein kinase II (CamKII) has many of the required features of a calcium sensor of neural activity. CamKII can act as a frequency monitor for calcium spikes (38, 39). The frequency response of this enzyme is dependent on the properties of calmodulin (CaM) trapping and autophosphorylation (39, 44). The properties of autophosphorylation also allow this enzyme to retain information of previous activity. For example, if the threshold for autophosphorylation is met during an initial stimulus and only a portion of the holoenzyme subunits are autophosphorylated, then the response to subsequent low-frequency stimuli may be increased (39, 45). Furthermore, CamKII is demonstrated, in vitro and in vivo, to have functional output that is relevant to the modulation of neuronal excitability, synaptic function, gene expression, and cell shape (46–48). There are several ways that the frequency response properties of CamKII might be modulated to accommodate different cellular set points as a homeostatic-activity sensor. CamKII activation is sensitive to a number of cellular variables including CaM concentration, phosphatase activity, enzyme localization, and the subunit composition of the holoenzyme (37, 44, 46). Modulation of any of these variables will potentially modulate CamKII function. The subunit composition of the holoenzyme may be a particularly potent parameter for altering the CamKII activity as an activity sensor. Neuronal CamKII consists primarily of two subunits, alpha and beta. These subunits can form homomeric and heteromeric enzymes composed of 6–12 subunits. The subunit composition influences the frequency response properties of the enzyme (39, 44). Several experiments indicate that subunit composition may be a modifiable feature of the kinase. Increased expression of the alpha subunit but not the beta subunit has been observed after long-term potentiation in the freely moving rat (49). Altered subunit composition may also P1: FXZ January 20, 2001 858 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY change the subcellular targeting of CamKII, and this could be a potent means of modulating kinase sensitivity and function (50). Calcineurin The serine/threonine phosphatase calcineurin is another example of a calciumsignaling molecule that has recently been implicated in signaling systems that can generate oscillation-specific changes in gene transcription (40, 51). This is most clearly demonstrated in experiments examining the signal transduction cascade of calcium, calcineurin, and NF-AT in lymphocytes and more recently in hippocampal neurons (40, 41, 52). Several important features of this signaling system are worth highlighting. It can mediate calcium oscillation-specific changes in gene expression that include both changes in transcriptional selectivity and changes in the magnitude of the transcriptional response. Signaling via calcineurin and NF-AT is cell type specific, depending on the NF-ATc family members expressed within each cell. Finally, this signaling system is responsive to cellular context because the transcriptional response can be strongly influenced by convergent signaling systems including ras and protein kinase C (40). This last point is particularly important because it means that a calcium signal can be interpreted by the cellular context in which the signal is delivered. Calcium signaling and CamKII signaling have essential functions during synaptic and neuronal development. Two recent studies are particularly relevant to the topic of homeostatic regulation. In Caenorhabditis elegans, synapse number at central neurons remains constant during normal larval development. A genetic analysis demonstrates that the maintenance of synapse number during development requires voltage-gated calcium channels and CamKII (47). These results implicate a system for maintaining synapse number and thereby maintaining post-synaptic function that is sensitive to both voltage and calcium signaling through voltagegated channels. Calcineurin, likewise, has functional output that is relevant to the regulation of neuronal excitability (53, 54). Post-synaptic calcineurin has been implicated in muscle-to-nerve retrograde-signaling systems that modulate synaptic strength at the vertebrate NMJ, in vitro (55). More recently, calcineurin modulation of inhibitory synaptic function has been implicated in the altered neuronal excitability after long-term potentiation (56). Calcium Channels The decoding of neural activity via calcium signaling is not achieved solely in the properties of the second messenger and transcriptional regulatory systems. Some of the selective properties of a calcium sensor for neural activity may be established by the properties of the voltage-gated channels themselves. L-type calcium channels provide a source of calcium influx capable of triggering calmodulin mobilization to the nucleus and activation of CREB-mediated transcription (48, 57). L-type calcium channels may be preferentially activated by synaptic P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 859 potentials [excitatory post-synaptic potentials (EPSPs)] as opposed to action potentials. This is suggested by experiments monitoring the L-type calcium current in response to simulated EPSP or action potential waveforms (57a). The differential activation of L-type calcium channels is linked to the observation that L-type channels are activated at relatively negative potentials and have slow activation kinetics. Both features favor activation by slow, low-amplitude EPSPs as opposed to action potentials. One can imagine that information regarding both synaptic and action potential activity might be necessary for a neuron to accurately achieve homeostatic regulation. Action potentials provide information about the summed total cellular depolarization. EPSPs can provide information about the localized sources of depolarization within the dendritic tree. In conclusion, calcium entry into a cell via calcium channels or transmitter receptors can be transduced into changes in calcium and second-messenger signaling that allow a cell to decode the voltage activity of a neuron into chemical signaling. There is the potential for this signaling to distinguish information about summed neuronal depolarization from information about local regions of synaptic activity in the dendrites. The responses downstream of calcium sensors can be cell type specific and context dependent within a given cell type. These are essential features of a system capable of monitoring neuronal activity and generating a homeostatic, compensatory response to altered activity or innervation. Many questions remain. How does a cell establish its own particular activity set points? Is this a cell-autonomous property linked to the genetic phenotype of a cell or is it influenced by early activity-dependent events? If early activity-dependent events participate in establishing the set points for the homeostasis of neuronal activity, is there a critical period for determining the calcium sensitivity (or set points) that will maintain cellular activity within a particular range? Can these properties be modulated? Specificity of Homeostatic Synaptic Regulation The specificity of homeostatic regulatory mechanisms has yet to be explored in any depth. A series of experiments at the Drosophila NMJ demonstrated that homeostatic regulation of synaptic function can be target specific at the terminals of a single neuron (2). Every other experiment has used cell-wide perturbation of activity or innervation, and, as a result, our current understanding of homeostatic signaling points to cell-wide regulatory change (4, 19). In theory, however, some forms of synaptic regulation must be synapse specific if homeostatic regulation is going to maintain the stability of neural circuitry as opposed to generating change (thereby functioning as another expression of synaptic plasticity). For a cell to regulate its own excitability, it requires a mechanism to monitor summed, total excitation. One possibility is that a cell can monitor some aspect of action-potential-induced depolarization, potentially via calcium-dependent signaling (15, 57). If a cell monitors summed depolarization in this manner, how might a homeostatic regulatory system be established such that specificity could P1: FXZ January 20, 2001 860 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY be achieved? Two candidate solutions to this problem draw on models for synapsespecific activity-dependent plasticity (Figure 3). One possibility is that activity monitors are present in the cell soma, at a central location capable of integrating action-potential-induced activity. This would allow homeostatic signaling to include a transcriptional response at the nucleus. A homeostatic signal could be sent throughout the entire dendritic tree. If this homeostatic signal required a synaptic cofactor that was sensitive to local synaptic activity such as CaMII or calcineurin, then the a cell-wide homeostatic signal would be active only at sites of inappropriate activity (Figure 3A). Synaptic microdomains of elevated calcium have been observed, mediated by signaling via metabotropic glutamate receptors and inositol-1,4,5-triphosphate receptors (58, 59). Because homeostatic regulation is expected to respond to persistent and inappropriate synaptic activity, these local sites of elevated calcium signaling will not require synaptic “memory” of previous activity or “synaptic tagging,” both of which have been proposed as mechanisms for synapse-specific long-term synaptic plasticity (60, 61). A second possibility is that the entire homeostatic-signaling mechanism is localized post-synaptically at each synapse. Information about sum total excitation could be conveyed to every synapse via action potential back propagation. In this model, each synapse would be endowed with a mechanism for comparing the activity of back propagating action potentials with local synaptic activity (Figure 3B). There is evidence that modulation of synaptic strength via correlated activity might be highly sensitive to the timing of synaptic and action-potential-mediated depolarization (62), and this could conceivably be utilized to mediate homeostatic regulation of synapses in this manner. However, purely Hebbian mechanisms may not be sufficient to account for homeostatic synaptic regulation because synapses that participate in firing a post-synaptic cell, but do so inappropriately, should be downregulated according to a homeostatic hypothesis. SIGNALING AT THE SYNAPSE The evidence for homeostatic regulation of synaptic and neuronal function implies the presence of elaborate intracellular- and intercellular-signaling mechanisms. Signaling from the site of activity integration to sites of synaptic input is essential. At the synapse, there is evidence for modulation of transmitter receptor function or insertion. There is also evidence for trans-synaptic signaling systems that can modulate presynaptic release. Trans-Synaptic Signaling Nerve growth factors and neurotrophins are among the most intensively studied trans-synaptic signaling molecules. Neurotrophins in particular have been implicated as both survival factors and as mediators of activity-dependent synapse modification (7). Alternatively, biochemical and cell-biological evidence is P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 861 Figure 3 Synapse-specific homeostatic signaling. A neuron with two sources of synaptic input is diagrammed. (A) Monitors of summed excitation may be located in the soma. Depolarization is transmitted to the soma via passive or active dendritic propagation. Homeostatic regulatory signals are then sent back to the dendrites in a uniform manner, but they only generate regulatory changes at synapses with inappropriate synaptic activity [increased synaptic excitation (right)]. Sites of inappropriate synaptic activity may be recognized by the homeostatic signals owing to localized calcium-dependent signaling, potentially involving the activity of Cam kinase or calcineurin. (B) An alternate model proposes that information of summed excitation is monitored via back propagating action potentials. Activity sensors at each synapse (synapse-specific monitors) could achieve homeostatic change through the comparison of synaptic vs action potential activity at each individual synapse. P1: FXZ January 20, 2001 862 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY accumulating to suggest a model of trans-synaptic signaling that is achieved via a trans-synaptic scaffold of proteins that spans the synaptic cleft from the postsynaptic density to the presynaptic cytoskeleton. Biochemical data have linked proteins of the pre- and post-synaptic densities via a trans-synaptic, asymmetric neuroligin-neurexin junction (63, 64). Neurexins have a single transmembrane domain and a short carboxy tail with protein 4.1 and CASK-PDZ domain-binding sequence motifs (65). Proteins with similarities to neurexins have been shown to mediate axon-glia interactions (66, 67). The subcellular localization of neurexins within central neurons remains to be determined. However, neuroligin-1 is localized postsynaptically at central synapses and has been shown to be a beta-neurexin ligand (68–70). Three neuroligins have been identified in the rat, and two neuroligins have been recently identified in mice (71). All three neuroligins possess an extracellular domain that shares homology with acetylcholinesterase, although this domain does not have enzymatic activity (68). All of the neuroligins have a cytoplasmic domain with a PDZ-binding motif capable of binding synaptic adapter proteins including post-synaptic density protein (PSD)-95 (72). Thus the neuroligin-neurexin asymmetric junction is linked to a large network of synaptic proteins through the binding of modular adapter proteins such as CASK and PSD-95 via PDZ-mediated interactions. Biochemical data have been assembled that describe a core trans-synaptic protein scaffold around the neurexin-neuroligin asymmetric junction that is hypothesized to participate in the organization of the synapse (Figure 4A). This model assumes an asymmetric junction with presynaptic neurexin and post-synaptic neuroligin. Presynaptic neurexin interacts with an evolutionarily conserved protein complex consisting of CASK, MINT1, and Veli (63). CASK, MINT1, and Veli are modular adapter proteins that contain multiple protein interaction domains including PDZ and SH3 domains. Significantly, both CASK and MINT1 are linked biochemically to the presynaptic calcium channel (64). The CASK SH3 domain has been shown to interact with a proline-rich sequence in the presynaptic calcium channel (64). The carboxy terminus of the apha1a and alpha1b calcium channel subunits also binds to the first PDZ domain of MINT1. Post-synaptically, neuroligin interacts with PSD-95 (72). Many other constituents of the post-synaptic density interact with PSD-95 including CamKII, nitric oxide synthase, and potassium channels (73). Thus an extensive trans-synaptic protein complex is linked biochemically to presynaptic calcium channels (64). This extensive, trans-synaptic protein complex could serve to regulate the localization or function of presynaptic calcium channels. Because transmitter release is exquisitely sensitive to the function and most likely the localization of the presynaptic calcium channel, this trans-synaptic protein complex may have a potent influence on synapse organization and function. This model is supported by genetic analysis of an analogous Lin-10/Lin-2/Lin-7 protein complex in C. elegans. The Lin-10/Lin-2/Lin-7 complex is necessary and sufficient for basolateral localization of the epidermal growth factor receptor tyrosine kinase Let-23 and is P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 863 necessary for post-synaptic glutamate receptor localization in C. elegans (74, 75). Until recently, however, there has not been experimental evidence supporting a role for the neurexin-neuroligin trans-synaptic complex in either synapse formation or modulation. An elegant series of recent experiments demonstrated a role for the neuroliginneurexin interaction in synapse formation (71). In these experiments, neuroligin expression was demonstrated to be sufficient to induce presynaptic differentiation in vitro. Neuroligin-induced synapse formation was analyzed both at the light level and ultrastructurally. This synaptogenic activity is mediated by specific sequences in the esterase domain of neuroligin, and evidence suggests that this signaling is mediated through interaction with presynaptic beta-neurexin (71). Among the synaptic proteins that are assembled by the neuroligin-neurexin interaction are synapsin and the modular adapter protein CASK/lin2. These data provide the first evidence of a cell-contact-mediated signaling event that is sufficient to drive synapse formation in the central nervous system. It remains to be determined whether similar signaling may occur at a mature synapse. The function of this complex, in vivo, also remains to be clarified. A neuroligin knockout is remarkably healthy (68). Neuroligins have widespread expression, however, and it is possible that different neuroligins have partially overlapping functions (68, 71). A recent result by Hsueh et al (76) offers a further glimpse of possible signaling from this extensive trans-synaptic protein complex. The modular adapter protein CASK/Lin2 is concentrated at central synapses and binds to neurexin and syndecan proteins at the plasma membrane as well as binding the cytoplasmic proteins Mint and Veli. Recent work demonstrated that CASK can enter the cell nucleus and bind to specific DNA sequences (called T elements) in a transcriptional coactivation complex with Tbr-1 (76). Among the T-element-containing genes that may be regulated by the CASK/Tbr-1 complex is reelin, a gene implicated in cerebrocortical development. Thus an integral element of the trans-synaptic protein complex can signal to the nucleus and modulate transcription. Parallels can be drawn to cadherin signaling via beta-catenin from the cell plasma membrane to the nucleus (77), suggesting that this scenario may become a more common aspect of synaptic signaling and regulation. The mechanism by which CASK nuclear translocation is initiated and the role of this signaling in synaptic regulation remain unclear. The generality of such a trans-synaptic signaling scaffold is suggested by the recent demonstration of a similar system at the vertebrate NMJ (78). In this system, biochemical data link synaptic alpha 4-beta 2-gamma 1 laminin to a presynaptic protein complex that includes a calcium channel subunit (Figure 4B). A potential trans-synaptic protein complex can be achieved through the synaptic basal lamina that traverses the neuromuscular synaptic cleft. Thus biochemical data at both central and neuromuscular synapses implicate a network of protein-protein interactions that may unify the pre- and postsynaptic densities. A role for trans-synaptic adhesion in synaptic regulation is not unique to the neurexin-neuroligin junction. Homophilic cell adhesion molecules with the potential to signal across the synapse have been implicated in the mechanisms of P1: FXZ January 20, 2001 864 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY synapse stabilization, growth, and activity-dependent plasticity (22, 79–81). Cell adhesion-dependent changes in synaptic morphology are demonstrated to be necessary, in some systems, for activity-dependent changes in synaptic function (82). However, signaling via cell adhesion molecules has not been demonstrated to directly influence the mechanisms of either transmitter release or post-synaptic transmitter detection (Figure 4C). The complexity and extent of the synaptic signaling complex are yet to be completely realized. Forward genetic studies have begun to elucidate new elements of the synapse that may have escaped traditional biochemical techniques owing to their extreme size. These proteins are among a growing number of proteins that define a perijunctional region of the synapse, just adjacent to the active zone (Figure 4D). Perijunctional proteins include Piccolo, Bassoon, Still-life, Highwire, and Futsch (83). The Highwire protein (5233 amino acids) has recently been demonstrated to control presynaptic growth (83, 84). Mutations in highwire show a dramatic synaptic overgrowth phenotype, suggesting that highwire may represent a repressor of synapse elaboration (84). The Futsch protein is also extremely large (5327 amino acids). It is associated with the synaptic microtubule cytoskeleton and is necessary for synaptic microtubule organization and normal synaptic growth (85, 86). Among the constituents of the periactive zone are molecules implicated in synaptic vesicle endocytosis. Dynamin and dynamin-associated protein (DAP160) are localized to the periactive zone in a honeycomblike lattice that surrounds active zones in the Drosophila neuromuscular bouton (87, 88). The function of a potential periactive-zone protein complex remains to be determined. However, since highwire and futsch are potent regulators of synaptic growth and function, signaling via proteins in the periactive zone may be important during homeostatic regulation. Regulation of Post-Synaptic Neurotransmitter Receptors An altered post-synaptic response to neurotransmitters is implicated in many of the known examples of homeostatic regulation. A variety of different mechanisms could account for altered post-synaptic sensitivity to neurotransmitters, including altered receptor insertion, localization, or altered function of existing receptors. Rapid, activity-dependent cycling of α-AMPA-type glutamate receptors appears to be a potent mechanism for altering synaptic function during acute synaptic plasticity (20). Green fluorescent protein-tagged receptors are delivered to dendritic spines and clusters after tetanic stimulation, and this receptor trafficking requires activation of the NMDA receptors (89). Agents that block processes of exocytosis and endocytosis have effects on synaptic function that are consistent with a mechanism by which receptors are trafficked to and from the synaptic membrane (90, 91). Most recently, it has been shown that the insertion of α-AMPA-type receptors requires CamKII activity and a PDZ protein interaction (92). In principle, intracellular-signaling systems that are coupled to the mechanisms of receptor cycling could represent a potent mechanism for homeostatic regulation of synaptic strength. P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 865 CONCLUSION The mechanisms of neural homeostasis are complex and may comprise diverse, interrelated signaling systems. Neurons and muscle are capable of responding in a homeostatic manner to diverse types of perturbation that include altered cell size, innervation, and excitation. The mechanisms of compensation are equally diverse, including regulated changes to synaptic size, presynaptic release, postsynaptic transmitter sensitivity, and the balance of voltage-gated channels. The complexity and diverse nature of homeostatic regulation may assure that neurons and muscle have a robust system with which to regulate excitability. Ultimately, the role of homeostatic signaling during neuronal development, the maintenance of neural function, and the etiology of neural disease will require an in-depth understanding of these homeostatic regulatory events at a molecular and cellular level. Given a limited molecular understanding of neural homeostasis, a forward genetic approach in model systems such as C. elegans and Drosophila may provide initial insight into these processes. ACKNOWLEDGMENTS We thank Kurt Marek, Suzanne Paradis, and Peter Clyne for critical comments on earlier versions of this manuscript. The writing of this review was supported in part by the BWF and National Institutes of Health grant NS39313-01 to GWD and grant NS39552 to IB. Visit the Annual Reviews home page at www.AnnualReviews.org LITERATURE CITED 1. Davis GW, Goodman CS. 1998. Genetic analysis of synaptic development and plasticity: homeostatic regulation of synaptic efficacy. Curr. Opin. Neurobiol. 8:149–56 2. Davis GW, Goodman CS. 1998. Synapsespecific control of synaptic efficacy at the terminals of a single neuron. Nature 392: 82–86 3. Liu G, Tsien RW. 1995. Properties of synaptic transmission at single hippocampal synaptic boutons. Nature 375:404–8 4. Turrigiano GG. 1999. Homeostatic plasticity in neuronal networks: the more things change, the more they stay the same. Trends Neurosci. 22:221–27 5. Philips J, Herskowitz I. 1997. Osmotic balance regulates cell fusion during mating 6. 7. 8. 9. 10. in Saccharomyces cerevisiae. J. Cell. Biol. 138:961–74 Bliss TV, Collingridge GL. 1993. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361:31–39 McAllister AK, Katz LC, Lo DC. 1999. Neurotrophins and synaptic plasticity. Annu. Rev. Neurosci. 22:295–318 Sanes JR, Lichtman JW. 1999. Development of the vertebrate neuromuscular junction. Annu. Rev. Neurosci. 22:389–442 Katz LC, Shatz CJ. 1996. Synaptic activity and the construction of cortical circuits. Science 274:1133–38 Petersen SA, Fetter RD, Noordermeer JN, Goodman CS, DiAntonio A. 1997. Genetic analysis of glutamate receptors in P1: FXZ January 20, 2001 866 11. 12. 13. 14. 15. 16. 17. 18. 19. 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY Drosophila reveals a retrograde signal regulating presynaptic transmitter release. Neuron 19:1237–48 Davis GW, DiAntonio A, Petersen SA, Goodman CS. 1998. Postsynaptic PKA controls quantal size and reveals a retrograde signal that regulates presynaptic transmitter release in Drosophila. Neuron 20:305–15 Marek KW, Ng N, Fetter R, Smolik S, Goodman CS, Davis GW. 2000. A genetic analysis of synaptic development: preand postsynaptic dCBP control transmitter release at the Drosophila NMJ. Neuron 25:537–47 Sandrock A, Dryer S, Rosen K, Gozani S, Kramer R, et al. 1997. Maintenance of acetylcholine receptor number by neuregulins at the neuromuscular junction in vivo. Science 276:599–604 Liao D, Zhang X, O’Brien R, Ehlers MD, Huganir RL. 1999. Regulation of morphological postsynaptic silent synapses in developing hippocampal neurons. Nat. Neurosci. 2:37–43 Marder E, Abbott LF, Turrigiano GG, Liu Z, Golowasch J. 1996. Memory from the dynamics of intrinsic membrane currents. Proc. Natl. Acad. Sci. USA 93:13481–86 Turrigiano G, Abbott LF, Marder E. 1994. Activity-dependent changes in the intrinsic properties of cultured neurons. Science 264:974–77 Turrigiano G, LeMasson G, Marder E. 1995. Selective regulation of current densities underlies spontaneous changes in the activity of cultured neurons. J. Neurosci. 15:3640–52 Turrigiano GG, Leslie KR, Desai NS, Rutherford LC, Nelson SB. 1998. Activitydependent scaling of quantal amplitude in neocortical neurons. Nature 391:892–96 O’Brien RJ, Kamboj S, Ehlers MD, Rosen KR, Fischbach GD, Huganir RL. 1998. Activity-dependent modulation of synaptic AMPA receptor accumulation. Neuron 21:1067–78 20. Turrigiano GG. 2000. AMPA receptors unbound: membrane cycling and synaptic plasticity. Neuron 26:5–8 21. Desai NS, Rutherford LC, Turrigiano GG. 1999. BDNF regulates the intrinsic excitability of cortical neurons. Learn. Mem. 6:284–91 22. Davis GW, Schuster CM, Goodman CS. 1997. Genetic analysis of the mechanisms controlling target selection: target-derived Fasciclin II regulates the pattern of synapse formation. Neuron 19:561–73 23. Lnenicka GA, Mellon D, Jr. 1983. Changes in electrical properties and quantal current during growth of identified muscle fibres in the crayfish. J. Physiol. 345:261–84 24. Lnenicka GA, Mellon D, Jr. 1983. Transmitter release during normal and altered growth of identified muscle fibres in the crayfish. J. Physiol. 345:285–96 25. Frank E. 1973. Matching of facilitation at the neuromuscular junction of the lobster: a possible case for influence of muscle on nerve. J. Physiol. 233:635–58 26. Hill AA, Edwards DH, Murphey RK. 1994. The effect of neuronal growth on synaptic integration. J. Comput. Neurosci. 1:239–54 27. Stuart G, Sakmann B. 1995. Amplification of EPSPs by axosomatic sodium channels in neocortical pyramidal neurons. Neuron 15:1065–76 28. Schwindt PC, Crill WE. 1995. Amplification of synaptic current by persistent sodium conductance in apical dendrite of neocortical neurons. J. Neurophysiol. 74:2220–24 29. Edwards DH, Yeh SR, Barnett LD, Nagappan PR. 1994. Changes in synaptic integration during the growth of the lateral giant neuron of crayfish. J. Neurophysiol. 72:899–908 30. Cull-Candy SG, Miledi R, Trautmann A, Uchitel OD. 1980. On the release of transmitter at normal, myasthenia gravis and myasthenic syndrome affected human endplates. J. Physiol. 299:621–38 31. Plomp JJ, van Kempen GT, Molenaar P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. PC. 1992. Adaptation of quantal content to decreased postsynaptic sensitivity at single endplates in alpha-bungarotoxintreated rats. J. Physiol. 458:487–99 Agre JC, Rodriquez AA, Tafel JA. 1991. Late effects of polio: critical review of the literature on neuromuscular function. Arch. Phys. Med. Rehabil. 72:923–31 McKinney RA, Debanne D, Gahwiler BH, Thompson SM. 1997. Lesion-induced axonal sprouting and hyperexcitability in the hippocampus in vitro: implications for the genesis of posttraumatic epilepsy. Nat. Med. 3:990–96 Su TT, O’Farrell PH. 1998. Size control: cell proliferation does not equal growth. Curr. Biol. 8:R687–89 Murray A. 1994. Cell cycle checkpoints. Curr. Opin. Cell. Biol. 6:872–76 Gao FB, Brenman JE, Jan LY, Jan YN. 1999. Genes regulating dendritic outgrowth, branching, and routing in Drosophila. Genes Dev. 13:2549–61 Schulman H, Heist K, Srinivasan M. 1995. Decoding Ca2+ signals to the nucleus by multifunctional CaM kinase. Prog. Brain Res. 105:95–104 Schulman H, Hanson PI, Meyer T. 1992. Decoding calcium signals by multifunctional CaM kinase. Cell Calcium 13:401– 11 De Koninck P, Schulman H. 1998. Sensitivity of CaM kinase II to the frequency of Ca2+ oscillations. Science 279:227–30 Crabtree GR. 1999. Generic signals and specific outcomes: signaling through Ca2+, calcineurin, and NF-AT. Cell 96:611– 14 Graef IA, Mermelstein PG, Stankunas K, Neilson JR, Deisseroth K, et al. 1999. Ltype calcium channels and GSK-3 regulate the activity of NF-ATc4 in hippocampal neurons. Nature 401:703–8 Buonanno A, Fields RD. 1999. Gene regulation by patterned electrical activity during neural and skeletal muscle development. Curr. Opin. Neurobiol. 9:110–20 867 43. LeMasson G, Marder E, Abbott LF. 1993. Activity-dependent regulation of conductances in model neurons. Science 259:1915–17 44. Brocke L, Srinivasan M, Schulman H. 1995. Developmental and regional expression of multifunctional Ca2+/calmodulindependent protein kinase isoforms in rat brain. J. Neurosci. 15:6797–808 45. Meyer T, Hanson PI, Stryer L, Schulman H. 1992. Calmodulin trapping by calciumcalmodulin-dependent protein kinase. Science 256:1199–202 46. Braun AP, Schulman H. 1995. The multifunctional calcium/calmodulin-dependent protein kinase: from form to function. Annu. Rev. Physiol. 57:417–45 47. Rongo C, Kaplan JM. 1999. CaMKII regulates the density of central glutamatergic synapses in vivo. Nature 402:195–99 48. Deisseroth K, Heist EK, Tsien RW. 1998. Translocation of calmodulin to the nucleus supports CREB phosphorylation in hippocampal neurons. Nature 392:198–202 49. Thomas KL, Laroche S, Errington ML, Bliss TV, Hunt SP. 1994. Spatial and temporal changes in signal transduction pathways during LTP. Neuron 13:737–45 50. Srinivasan M, Edman CF, Schulman H. 1994. Alternative splicing introduces a nuclear localization signal that targets multifunctional CaM kinase to the nucleus. J. Cell. Biol. 126:839–52 51. Naya FJ, Mercer B, Shelton J, Richardson JA, Williams RS, Olson EN. 2000. Stimulation of slow skeletal muscle fiber gene expression by calcineurin in vivo. J. Biol. Chem. 275:4545–48 52. Meldolesi J. 1998. Calcium signalling: oscillation, activation, expression. Nature 392:863–66 53. Lai MM, Hong JJ, Ruggiero AM, Burnett PE, Slepnev VI, et al. 1999. The calcineurin-dynamin 1 complex as a calcium sensor for synaptic vesicle endocytosis. J. Biol. Chem. 274:25963–66 54. Mulkey RM, Endo S, Shenolikar S, P1: FXZ January 20, 2001 868 55. 56. 57. 57a. 58. 59. 60. 61. 62. 17:30 DAVIS ¥ Annual Reviews AR114-33 BEZPROZVANNY Malenka RC. 1994. Involvement of a calcineurin/inhibitor-1 phosphatase cascade in hippocampal long-term depression. Nature 369:486–88 Wan J, Poo M. 1999. Activity-induced potentiation of developing neuromuscular synapses. Science 285:1725–28 Lu YM, Mansuy IM, Kandel ER, Roder J. 2000. Calcineurin-mediated LTD of GABAergic inhibition underlies the increased excitability of CA1 neurons associated with LTP. Neuron 26:197–205 Deisseroth K, Bito H, Tsien RW. 1996. Signaling from synapse to nucleus: postsynaptic CREB phosphorylation during multiple forms of hippocampal synaptic plasticity. Neuron 16:89–101 Mermelstein PG, Bito H, Deisseroth K, Tsien RW. 2000. Critical dependence of cAMP response element-binding protein phosphorylation on L-type channels supports a selective response to EPSPs in preference to action potentials. J. Neurosci. 20:266–73 Finch EA, Augustine GJ. 1998. Local calcium signalling by inositol-1,4,5trisphosphate in Purkinje cell dendrites. Nature 396:753–56 Takechi H, Eilers J, Konnerth A. 1998. A new class of synaptic response involving calcium release in dendritic spines. Nature 396:757–60 Frey U, Morris RG. 1998. Synaptic tagging: implications for late maintenance of hippocampal long- term potentiation. Trends Neurosci. 21:181–88 Casadio A, Martin KC, Giustetto M, Zhu H, Chen M, et al. 1999. A transient, neuron-wide form of CREB-mediated long-term facilitation can be stabilized at specific synapses by local protein synthesis. Cell 99:221–37 Bi GQ, Poo MM. 1998. Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type. J. Neurosci. 18:10464–72 63. Butz S, Okamoto M, Sudhof TC. 1998. A tripartite protein complex with the potential to couple synaptic vesicle exocytosis to cell adhesion in brain. Cell 94:773–82 64. Maximov A, Sudhof TC, Bezprozvanny I. 1999. Association of neuronal calcium channels with modular adaptor proteins. J. Biol. Chem. 274:24453–56 65. Hata Y, Butz S, Sudhof TC. 1996. CASK: a novel dlg/PSD95 homolog with an N-terminal calmodulin-dependent protein kinase domain identified by interaction with neurexins. J. Neurosci. 16:2488–94 66. Bellen HJ, Lu Y, Beckstead R, Bhat MA. 1998. Neurexin IV, caspr and paranodin— novel members of the neurexin family: encounters of axons and glia. Trends Neurosci. 21:444–49 67. Missler M, Sudhof TC. 1998. Neurexins: three genes and 1001 products. Trends Genet. 14:20–26 68. Song JY, Ichtchenko K, Sudhof TC, Brose N. 1999. Neuroligin 1 is a postsynaptic cell-adhesion molecule of excitatory synapses. Proc. Natl. Acad. Sci. USA 96: 1100–5 69. Ichtchenko K, Hata Y, Nguyen T, Ullrich B, Missler M, et al. 1995. Neuroligin 1: a splice site-specific ligand for betaneurexins. Cell 81:435–43 70. Nguyen T, Sudhof TC. 1997. Binding properties of neuroligin 1 and neurexin 1beta reveal function as heterophilic cell adhesion molecules. J. Biol. Chem. 272:26032–39 71. Scheiffele P, Jinhong F, Choih J, Fetter R, Serafini T. 2000. Neuroligin expressed in non-neuronal cells triggers presynaptic development in contacting axons. Cell 101: 1–10 72. Irie M, Hata Y, Takeuchi M, Ichtchenko K, Toyoda A, et al. 1997. Binding of neuroligins to PSD-95. Science 277:1511–15 73. Kennedy MB. 1997. The postsynaptic density at glutamatergic synapses. Trends Neurosci. 20:264–68 74. Kaech SM, Whitfield CW, Kim SK. 1998. The LIN-2/LIN-7/LIN-10 complex P1: FXZ January 20, 2001 17:30 Annual Reviews AR114-33 HOMEOSTATIC NEURAL FUNCTION CONTROL 75. 76. 77. 78. 79. 80. 81. 82. 83. mediates basolateral membrane localization of the C. elegans EGF receptor LET-23 in vulval epithelial cells. Cell 94:761–71 Rongo C, Whitfield CW, Rodal A, Kim SK, Kaplan JM. 1998. LIN-10 is a shared component of the polarized protein localization pathways in neurons and epithelia. Cell 94:751–59 Hsueh YP, Wang TF, Yang FC, Sheng M. 2000. Nuclear translocation and transcription regulation by the membrane- associated guanylate kinase CASK/LIN-2. Nature 404:298–302 Bredt DS. 2000. Cell biology. Reeling CASK into the nucleus. Nature 404:241– 42 Sunderland WJ, Son YJ, Miner JH, Sanes JR, Carlson SS. 2000. The presynaptic calcium channel is part of a transmembrane complex linking a synaptic laminin (alpha4beta2gamma1) with non-erythroid spectrin. J. Neurosci. 20:1009–19 Schuster CM, Davis GW, Fetter RD, Goodman CS. 1996. Genetic dissection of structural and functional components of synaptic plasticity. I. Fasciclin II controls synaptic stabilization and growth. Neuron 17:641–54 Martin KC, Kandel ER. 1996. Cell adhesion molecules, CREB, and the formation of new synaptic connections. Neuron 17:567–70 Tanaka H, Shan W, Phillips GR, Arndt K, Bozdagi O, et al. 2000. Molecular modification of N-cadherin in response to synaptic activity. Neuron 25:93–107 Davis GW, Schuster CM, Goodman CS. 1996. Genetic dissection of structural and functional components of synaptic plasticity. III. CREB is necessary for presynaptic functional plasticity. Neuron 17:669–79 Chang Q, Balice-Gordon RJ. 2000. Highwire, rpm-1, and futsch: balancing 84. 85. 86. 87. 88. 89. 90. 91. 92. 869 synaptic growth and stability. Neuron 26:287–90 Wan HI, DiAntonio A, Fetter RD, Bergstrom K, Strauss R, Goodman CS. 2000. Highwire regulates synaptic growth in Drosophila. Neuron 26:313–29 Hummel T, Krukkert K, Roos J, Davis G, Klambt C. 2000. Drosophila Futsch/22C10 is a MAP1B-like protein required for dendritic and axonal development. Neuron 26: 357–70 Roos J, Hummel T, Ng N, Klambt C, Davis GW. 2000. Drosophila Futsch regulates synaptic microtubule organization and is necessary for synaptic growth. Neuron 26:371–82 Roos J, Kelly RB. 1998. Dap160, a neuralspecific Eps15 homology and multiple SH3 domain-containing protein that interacts with Drosophila dynamin. J. Biol. Chem. 273:19108–19 Roos J, Kelly RB. 1999. The endocytic machinery in nerve terminals surrounds sites of exocytosis. Curr. Biol. 9:1411–14 Shi SH, Hayashi Y, Petralia RS, Zaman SH, Wenthold RJ, et al. 1999. Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation. Science 284:1811–16 Carroll RC, Beattie EC, Xia H, Luscher C, Altschuler Y, et al. 1999. Dynamindependent endocytosis of ionotropic glutamate receptors. Proc. Natl. Acad. Sci. USA 96:14112–17 Luscher C, Xia H, Beattie EC, Carroll RC, von Zastrow M, et al. 1999. Role of AMPA receptor cycling in synaptic transmission and plasticity. Neuron 24:649–58 Hayashi Y, Shi SH, Esteban JA, Piccini A, Poncer JC, Malinow R. 2000. Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domain interaction. Science 287:2262–67 P1: FQP January 31, 2001 18:29 Annual Reviews AR114-COLOR Figure 4 Synaptic organization and trans-synaptic signaling. (A) The constituents of a neuroliginneurexin trans-synaptic complex are diagrammed. Presynaptic neurexin interacts with the evolutionarily conserved protein complex consisting of CASK, MINT, and Veli (adapter complex) that, in turn, binds to the presynaptic calcium channel. (B) A trans-synaptic complex that parallels the neuroligin-neurexin complex has been identified at the vertebrate NMJ. Synaptic laminin interacts with a complex that includes the presynaptic calcium channel alpha subunit. (C ) A periactive zone (perijunctional) complex is diagrammed for the Drosophila NMJ. Within this region directly adjacent to the active zone (gray shading) are cell adhesion molecules, including integrins and Fasciclin II, capable of signaling across the synapse. PDZ-containing molecules in the post-synaptic density (PSD), including discs-large, may organize the post-synaptic density in a manner analogous to vertebrate central synapses. (D) Additional molecules have been localized to the perijunctional region, including proteins involved in endocytosis [Dynamin and DAP160 (87)]. Recently a number of very large proteins have been identified that are present within the perisynaptic region, including Highwire and Futsch.