Jeff W. Lichtman

Imaging Neuronal Subsets in Transgenic Mice Expressing Multiple Spectral Variants of GFP

We generated transgenic mice in which red, green, yellow, or cyan fluorescent proteins (together termed XFPs) were selectively expressed in neurons. All four XFPs labeled neurons in their entirety, including axons, nerve terminals, dendrites, and dendritic spines. Remarkably, each of 25 independently generated transgenic lines expressed XFP in a unique pattern, even though all incorporated identical regulatory elements (from the thyl gene). For example, all retinal ganglion cells or many cortical neurons were XFP positive in some lines, whereas only a few ganglion cells or only layer 5 cortical pyramids were labeled in others. In some lines, intense labeling of small neuronal subsets provided a Golgi-like vital stain. In double transgenic mice expressing two different XFPs, it was possible to differentially label 3 neuronal subsets in a single animal.

INDUCTION, ASSEMBLY, MATURATION AND MAINTENANCE OF A POSTSYNAPTIC APPARATUS

Key Points The neuromuscular junction (NMJ) is one of the best models to study the induction, assembly, maturation and maintenance of a postsynaptic apparatus. The three most widely appreciated experimental advantages of the NMJ are its size, its simplicity and its accessibility. Acetylcholine receptors (AChRs) are inserted at multiple points into the embryonic myotube membrane. As the NMJ develops, receptor density underneath the nerve increases and the number of extrasynaptic receptors is very small. Several mechanisms account for this phenomenon: some AChRs redistribute in the plane of the membrane, the metabolic stability of AChRs increases after clustering, myonuclei associated with the postsynaptic membrane become transcriptionally specialized to express AChRs, and AChR transcription is suppressed in non-synaptic nuclei. This reorganization depends on chemical influences from the motor nerve. Several molecules that are responsible for postsynaptic reorganization have been identified. Their identification has led to a working model of how AChRs are clustered at the developing NMJ. In this model, the nerve releases a protein called agrin, which signals through a muscle-specific tyrosine kinase known as MuSK. MuSK then acts through an effector protein, rapsyn, to promote AChR clustering. Although this basic model has received significant experimental support, there are other factors that affect NMJ development. Neuregulin and its ErbB receptor (another tyrosine kinase) also affect AChR clustering, but the interaction of these proteins with the agrin–MuSK–rapsyn pathway is unclear. The embryonic NMJs are very different from the adult NMJs. The junction changes from a simple oval plaque to a pretzel-like set of branches, and the junctional membrane changes from a flat sheet to an invaginated surface with gutters and folds. Moreover, the composition of the basal lamina and the cytoskeletal apparatus change as the NMJ matures. Finally, a shift in AChR subunit composition leads to a change in their Ca2+ permeability, and ion- and ligand-gated channels segregate into discrete alternating domains. The mechanisms that underlie each of these transformations have only begun to be uncovered. AbstractThe postsynaptic apparatus of the skeletal neuromuscular junction, like that of other synapses, contains a high-density patch of neurotransmitter receptors that is closely associated with a variety of extracellular, transmembrane and cytoplasmic proteins that have adhesive, structural and signalling roles. The postsynaptic apparatus is organized by signals from the presynaptic nerve terminal. It changes in shape, size and molecular architecture as it matures. Once mature, it can be maintained for the life of the organism, but has the capacity for remodelling in response to altered input. The molecular and cellular mechanisms that govern each of these stages are now being elucidated by a combination of microscopic and genetic methods, allowing the neuromuscular junction to serve as a model for smaller and less-accessible central synapses.The prefrontal cortex has a vital role in effective, organized behaviour. Both functional neuroimaging in humans and electrophysiology in awake monkeys indicate that a fundamental principle of prefrontal function might be adaptive neural coding — in large regions of the prefrontal cortex, neurons adapt their properties to carry specifically information that is relevant to current concerns, producing a dense, distributed representation of related inputs, actions, rewards and other information. A model based on such adaptive coding integrates the role of the prefrontal cortex in working memory, attention and control. Adaptive coding points to new perspectives on several basic questions, including mapping of cognitive to neurophysiological functions, the influences of task content and difficulty, and the nature of frontal lobe specializations.

Can molecules explain long-term potentiation?

Although over 100 molecules have been implicated in long-term potentiation and depression, no consensus on their underlying molecular mechanisms has emerged. Here we discuss the difficulties of providing molecular explanations for cellular neurobiological phenomena.

Synapse Elimination and Indelible Memory

where one, or at most a few, identifiable motor neurons innervate each identifiable muscle fiber, forming a stereotyped circuit exclusively dedicated to a particular Jeff W. Lichtman*‡ and Howard Colman† *Department of Anatomy and Neurobiology †Department of Neurology function (Figure 1). The greater degree of redundancy Washington University School of Medicine is perhaps the most salient difference between the large St. Louis, Missouri 63110 nervous systems of typical vertebrates and the much smaller nervous systems of invertebrates. If the principle difference in brain size between inverDetermining the way in which memories are stored in tebrates and terrestrial vertebrates were due to neuronal the brain is one of neurobiology’s great remaining chalredundancy, then one consequence would be a fundalenges. One approach to this question has been to ask mental change in the organization of synaptic circuitry. what kinds of changes are taking place during postnatal Taking neuromuscular connectivity as an example, one development when the brain is changing in response to might expect that if muscles were comprised of many environmental cues. This was the approach that led hundreds or thousands of identical fibers, then each Hubel and Wiesel to discover a developmentally remotor neuron projecting to a muscle should be able to stricted competition between the eyes for control of innervate many of the identical fibers, causing extensive cortical space (Wiesel, 1982) and, subsequently, led othaxonal divergence (Figure 2, top). At the same time, ers to discover evidence that seemingly analogous commultiple duplicated motor axons should be appropriate petitive events occur in other parts of the developing partners for each muscle fiber, causing axonal convernervous system (reviewed by Purves and Lichtman, gence (Figure 2, middle). In a nervous system with re1980). dundant neurons and redundant postsynaptic targets, One hallmark of these developmental refinements in the result of simultaneous divergence and convergence synaptic circuitry is the elimination of axonal connecwould be a set of highly overlapping circuits, in which tions. In the visual system, thalamocortical axons diseach target cell could receive convergent innervation connect from cortical layer IV cells (Hubel et al., 1977); in from many neurons and each input could diverge to the cerebellum, climbing fibers disconnect from Purkinje many target cells (Figure 2, bottom). Interestingly, such cells (Crepel et al., 1976; Lohof et al., 1996); in autonomic a pattern is the basic circuit plan of the neuromuscular ganglia, preganglionic inputs disconnect from ganglion system in developing terrestrial vertebrates. It is also cells (Lichtman, 1977); and at the neuromuscular juncthe case that throughout the central nervous system tion, motor axons disconnect from muscle fibers. In each vast overlapping webs of converging and diverging inof these areas, elaboration of synapses by the remaining puts between populations of similar neurons make it axon or axons also occurs. Thus, while some inputs are difficult to describe the circuit diagram. being eliminated, others are becoming stronger. The In vertebrate neuromuscular systems, this kind of exrationale for this selective synapse loss as a means of cessive fan-out and fan-in is only transitory because refining circuitry is the subject of this review. synapse loss reduces the number of inputs innervating What follows is a peripheral view of synaptic refineeach muscle fiber (Figure 3, upper panels). In invertements based mostly on what is known about synapse brates, such a reductive mechanism seems unneceselimination at the mammalian neuromuscular junction. sary because, as far as is known, no early stage of Owing to the greater accessibility and simplicity of deextensive overlap exists. Thus, in a peculiar way, the veloping neuromuscular connections when compared ontogeny of the vertebrate neuromuscular system leads to connections in the brain (Sanes and Lichtman, 1999), to a pattern that is ultimately more invertebrate-like. it has been possible to describe the process in greater After the phase of synapse elimination, each axon has detail than elsewhere. One of the challenges for the its own unique circuit. Thus, synapse elimination can future will be to see whether the conclusions reached be viewed as a mechanism that creates large numbers from studies of the peripheral nervous system apply to of specific circuits out of initially more diffuse and reduncentral synapses. dant connections (Figure 3, lower panels). It is important to emphasize that the kind of elimination Phylogenetic Origins of Synapse Elimination described above (in which the number of axons that At the vertebrate neuromuscular junction, synapse elimiinnervate a target cell is permanently reduced) is only nation refines connections between populations of preone form of synapse elimination. For example, in an and postsynaptic partners. In vertebrates (except fish), invertebrate (Aplysia), long-term habituation of a behaveach motor neuron is always part of a pool of tens ior or pharmacologically induced depression is associor hundreds of similar neurons that serve a common ated with decreases in the number, size and vesicle purpose. Likewise, in vertebrates each identifiable muscomplement of sensory neuron active zones (Bailey and cle is composed of hundreds or thousands of similar Chen, 1988, 1989; Bailey et al., 1992). These structural muscle fibers. This reduplicated arrangement contrasts changes reduce synapse number but may not have any with the pattern of connections in invertebrate muscle, effect on input number. Indeed, the rapid reversibility of behaviors like habituation (i.e., dishabituation) would seem to require that axons retain some kind of synaptic ‡ To whom correspondence should be addressed (e-mail: jeff@

Genetic evidence that relative synaptic efficacy biases the outcome of synaptic competition

Synaptic activity drives synaptic rearrangement in the vertebrate nervous system; indeed, this appears to be a main way in which experience shapes neural connectivity. One rearrangement that occurs in many parts of the nervous system during early postnatal life is a competitive process called ‘synapse elimination’. At the neuromuscular junction, where synapse elimination has been analysed in detail, muscle fibres are initially innervated by multiple axons, then all but one are withdrawn and the ‘winner’ enlarges. In support of the idea that synapse elimination is activity dependent, it is slowed or speeded when total neuromuscular activity is decreased or increased, respectively. However, most hypotheses about synaptic rearrangement postulate that change depends less on total activity than on the relative activity of the competitors. Intuitively, it seems that the input best able to excite its postsynaptic target would be most likely to win the competition, but some theories and results make other predictions. Here we use a genetic method to selectively inhibit neurotransmission from one of two inputs to a single target cell. We show that more powerful inputs are strongly favoured competitors during synapse elimination.

Roles of Neurotransmitter in Synapse Formation: Development of Neuromuscular Junctions Lacking Choline Acetyltransferase

Activity-dependent and -independent signals collaborate to regulate synaptogenesis, but their relative contributions are unclear. Here, we describe the formation of neuromuscular synapses at which neurotransmission is completely and specifically blocked by mutation of the neurotransmitter-synthesizing enzyme choline acetyltransferase. Nerve terminals differentiate extensively in the absence of neurotransmitter, but neurotransmission plays multiple roles in synaptic differentiation. These include influences on the numbers of pre- and postsynaptic partners, the distribution of synapses in the target field, the number of synaptic sites per target cell, and the number of axons per synaptic site. Neurotransmission also regulates the formation or stability of transient acetylcholine receptor-rich processes (myopodia) that may initiate nerve-muscle contact. At subsequent stages, neurotransmission delays some steps in synaptic maturation but accelerates others. Thus, neurotransmission affects synaptogenesis from early stages and coordinates rather than drives synaptic maturation.

Pre-existing pathways promote precise projection patterns.

A large body of evidence shows that molecular cues promote specific synapse formation by guiding axons and by mediating their association with targets, but much less is known about the contribution of physical cues (such as mechanical constraints) to these processes. Here we used the peripheral motor system to investigate the latter issue. In living mice, we viewed individual motor axons bearing a fluorescent reporter, and mapped the cohort of muscle fibers that they innervated both before and after nerve damage. When gross trauma was minimized (by a nerve-crushing rather than nerve-cutting procedure), regenerating axons retraced their former pathways, bifurcated at original branch points, and formed neuromuscular junctions on the same fibers that they originally innervated. Axonal growth through tubes of non-neural cells seemed to account for this specificity, and specificity degraded when the tubes were cut. These results suggest that nonspecific guidance cues can be sufficient to generate specific synaptic circuitry.

In vivo time-lapse imaging of synaptic takeover associated with naturally occurring synapse elimination.

During development, competition between axons causes permanent removal of synaptic connections, but the dynamics have not been directly observed. Using transgenic mice that express two spectral variants of fluorescent proteins in motor axons, we imaged competing axons at developing neuromuscular junctions in vivo. Typically, one axon withdrew progressively from postsynaptic sites and the competing axon extended axonal processes to occupy those sites. In rare instances when the remaining axon did not reoccupy a site, the postsynaptic receptors rapidly disappeared. Interestingly, the progress and outcome of competition was unpredictable. Moreover, the relative areas occupied by the competitors shifted in favor of one axon and then the other. These results show synaptic competition is not always monotonic and that one axons contraction in synaptic area is associated with another axons expansion.

Alternatively Spliced Isoforms of Nerve- and Muscle-Derived Agrin: Their Roles at the Neuromuscular Junction

Agrin induces synaptic differentiation at the skeletal neuromuscular junction (NMJ); both pre- and postsynaptic differentiation are drastically impaired in its absence. Multiple alternatively spliced forms of agrin that differ in binding characteristics and bioactivity are synthesized by nerve and muscle cells. We used surgical chimeras, isoform-specific mutant mice, and nerve-muscle cocultures to determine the origins and nature of the agrin required for synaptogenesis. We show that agrin containing Z exons (Z+) is a critical nerve-derived inducer of postsynaptic differentiation, whereas neural isoforms containing a heparin binding site (Y+) and all muscle-derived isoforms are dispensable for major steps in synaptogenesis. Our results also suggest that the requirement of agrin for presynaptic differentiation is mediated indirectly by its ability to promote postsynaptic production or localization of appropriate retrograde signals.

Asynchronous Synapse Elimination in Neonatal Motor Units: Studies Using GFP Transgenic Mice

In developing muscle, synapse elimination reduces the number of motor axons that innervate each postsynaptic cell. This loss of connections is thought to be a consequence of axon branch trimming. However, branch retraction has not been observed directly, and many questions remain, such as: do all motor axons retract branches, are eliminated branches withdrawn synchronously, and are withdrawing branches localized to particular regions? To address these questions, we used transgenic mice that express fluorescent proteins in small subsets of motor axons, providing a unique opportunity to reconstruct complete axonal arbors and identify all the postsynaptic targets. We found that, during early postnatal development, each motor axon loses terminal branches, but retracting branches withdraw asynchronously and without obvious spatial bias, suggesting that local interactions at each neuromuscular junction regulate synapse elimination.