The role of transcription in the development of electrical properties of neuronal membranes has been largely unexplored. To study the molecular events which result in the expression of these properties it is useful to describe the timing of the underlying RNA synthetic events. For example, the timing of the transcription involved in denervation-induced action potentials in frog slow muscle fibres and brain extract-induced sodium channels in chick muscle cells has been investigated. Previous studies of Xenopus laevis spinal neurones have established that the timing of the development of the neuronal action potential ionic dependence in dissociated cell cultures parallels that seen in vivo. This culture system, therefore, allows the determination of transcription-dependent periods necessary for the development of membrane properties known to have in vivo relevance. In the study described here, actinomycin D was used to examine the timing of the RNA synthetic events necessary for (1) neurite outgrowth and (2) development of the ionic dependence of the action potential. I report that inhibition of transcription at an early stage specifically blocks the appearance of the mature sodium-dependent action potential without affecting either neurite outgrowth or the development of delayed rectification.

Alternative splicing results in production of four agrin proteins (agrin0, agrin8, agrin11, and agrin19) with different AChR aggregating activities. However, the cellular origin of mRNAs encoding each agrin isoform remains unknown. Using single-cell PCR, we demonstrate that in the chick ciliary ganglion, nonneuronal cells express only mRNA encoding agrin0, whereas neurons express one or any combination of agrin mRNAs. Moreover, significant differences were observed between the agrin mRNA profiles of ciliary and choroid neurons in the ganglion. The abundance of each agrin mRNA, the fraction of neurons expressing each transcript, and the combinations of transcripts expressed by neurons also change during development. Our results demonstrate that transcripts encoding agrin proteins with high AChR aggregating activity are expressed exclusively by neurons in the ciliary ganglion and that alternative splicing of agrin mRNA is regulated during development and in a cell-specific manner.

The characteristic functions of tissues and organs result from the integrated activity of individual cells. Nowhere is this more evident than in the nervous system, where the activities of single neurons communicating via electrical and chemical signals mediate complex functions, such as learning and memory. The past decade has seen an explosion in the identification of genes encoding proteins, such as voltage-gated channels and neurotransmitter receptors, responsible for neuronal excitability. These studies have highlighted the fact that even within a neuroanatomically defined region, the coexistence of multiple cell types makes it difficult, if not impossible, to correlate patterns of gene expression with function. The recent development of techniques sensitive enough to study gene expression at the single-cell level promises to break this bottleneck to our further understanding. Using examples taken from our own laboratories and the work of others, we review these techniques, their application, and discuss some of the difficulties associated with the interpretation of the data.

Mutations in the para gene specifically affect the expression of sodium currents in Drosophila. While 65% of wild-type embryonic neurons in culture express sodium currents, three distinct mutations in the para locus resulted in a decrease in the fraction of cells from which sodium currents could be recorded. This reduction was allele-dependent: macroscopic sodium currents were exhibited in 49% of the neurons in parats1 cultures, 35% in parats2, and only 2% in paraST76. Voltage-clamp experiments demonstrated that the parats2 mutation also affected the gating properties of sodium channels. These results provide convincing evidence that para, a gene recently shown to exhibit sequence similarity to vertebrate sodium channels alpha subunits, encodes functional sodium channels in Drosophila. The finding that one para allele (paraST76) can virtually eliminate the expression of sodium currents strongly argues that the para gene codes for the majority of sodium channels in cultured embryonic neurons.

Agrin is an extracellular matrix protein involved in the formation of the postsynaptic apparatus of the neuromuscular junction. In addition to spinal motor neurons, agrin is expressed by many other neuronal populations throughout the nervous system. Agrin's role outside of the neuromuscular junction, however, is poorly understood. Here we use the polymerase chain reaction to examine expression and alternative splicing of agrin in mouse somatosensory cortex during early postnatal development in vivo and in dissociated cell culture. Peak levels of agrin gene expression in developing cortex coincide with ingrowth of thalamic afferent fibres and formation of thalamocortical and intracortical synapses. Analysis of alternatively spliced agrin messenger RNA variants shows that greater than 95% of all agrin in developing and adult somatosensory cortex originates in neurons, including isoforms that have little or no activity in acetylcholine receptor aggregation assays. The levels of expression of "active" and "inactive" isoforms, however, are regulated during development. A similar pattern of agrin gene expression is also observed during a period when new synapses are being formed between somatosensory neurons growing in dissociated cell culture. Changes in agrin gene expression, observed both in vivo and in vitro, are consistent with a role for agrin in synapse formation in the central nervous system.

Previous studies in postnatal mouse demonstrating high levels of alpha7 nicotinic acetylcholine receptors on layer IV somatosensory cortical neurons coincident with the onset of functional synaptic transmission led us to investigate whether the number and/or the localization of these receptors could be regulated by activity. Accordingly, we examined alpha-bungarotoxin binding in mouse somatosensory cortex following removal of all of the vibrissae on one side of the face, either by vibrissal follicle cauterization or daily plucking beginning on the day of birth. Following vibrissa plucking, the levels of [125I]alpha-bungarotoxin binding on postnatal day 6 were significantly higher (23 +/- 7%) in the denervated cortex (contralateral to the peripheral manipulation) than the intact cortex. Cauterization also resulted in significantly higher (14 +/- 3%) [125I]alpha-bungarotoxin binding in the contralateral vs. the ipsilateral cortex. In contrast, there was no difference in [125I]alpha-bungarotoxin binding in the left and right cortices of unoperated control animals. At postnatal day 14, levels of [125I]alpha-bungarotoxin binding in layer IV were very low in control animals as well as in animals subjected to whisker plucking or cautery. These findings suggest that reducing activity in the somatosensory pathway regulates the density of alpha7 nicotinic acetylcholine receptors during the first postnatal week. However, the normal decrease in receptor density that is seen during the second postnatal week of development proceeds despite altered sensory activity.

Agrin is a synaptic basal lamina protein that has been proposed to mediate motor neuron-induced clustering of acetylcholine receptors during development of the neuro-muscular junction. The chick ciliary ganglion is a parasympathetic ganglion that contains motor neurons that project to striated and smooth muscle targets in the eye. We have examined agrin gene expression in the chick ciliary ganglion during normal embryonic development using in situ hybridization and quantitative PCR techniques. Ganglia were specifically labeled by antisense agrin cRNA probes and the density of labeling changed during development. Hybridization was most intense in sections of ganglia obtained from embryos before embryonic Day 15 (E15), declining to relatively low levels by hatching at E20. Throughout embryonic development labeling was associated with glial cells, in addition to both ciliary and choroid neurons. Measurement of agrin mRNA levels by competitive PCR showed that agrin gene expression in the ganglion increased dramatically between E8 and E10, was sustained at high levels from E10 to E14, and declined thereafter. This time course is coincident with the period of synapse formation between ganglionic neurons and their peripheral targets. Previous studies in chick CNS have shown that alternative RNA splicing of a single exon encoding 11 amino acids gives rise to an active and an inactive agrin protein. Our analysis of RNA isolated from chick ciliary ganglia demonstrated that a second, previously uncharacterized exon encoding 8 amino acids can also be spliced into the same region. Alternative splicing of both the 8- and the 11-amino-acid exons results in expression of four distinct agrin transcripts in the ganglion. Changes in the level of total agrin mRNA in the ganglion reflect developmentally regulated changes in the levels of these alternatively spliced agrin isoforms. These results demonstrate that agrin is expressed in autonomic motor neurons of the peripheral nervous system and support a wider role for agrin as a synaptogenic protein, not limited to spinal motor neurons.

Previous studies in rat, showing a transient pattern of expression of the alpha 7 nicotinic acetylcholine receptor in the ventrobasal thalamus and barrel cortex during the first 2 postnatal weeks, suggest that these receptors may play a role in development of the thalamocortical system. In the present study, in situ hybridization and radiolabeled ligand binding were employed to examine the spatiotemporal distribution of alpha 7 mRNA and alpha-bungarotoxin binding sites in the thalamocortical pathway of mouse during early postnatal development. As in the rat, high levels of alpha 7 mRNA and alpha-bungarotoxin binding sites are present in the barrel cortex of mouse during the first postnatal week. Both alpha 7 mRNA and its receptor protein are observed in all cortical laminae, with the highest levels seen in the compact cortical plate, layer IV, and layer VI. When viewed in a tangential plane, alpha 7 mRNA and alpha-bungarotoxin binding sites delineate a whisker-related barrel pattern in layer IV by P3-5. Quantitative analysis reveals a dramatic decrease in the levels of expression of alpha 7 mRNA and alpha-bungarotoxin binding sites in the cortex by the end of the second postnatal week. Unlike in the rat, only low levels of alpha 7 mRNA or alpha-bungarotoxin binding sites are present in the ventrobasal complex of the mouse thalamus. The broad similarities between the thalamocortical development of rat and mouse taken together with the present results suggest that alpha 7 receptors located on cortical neurons, rather than on thalamic neurons, play a role in mediating aspects of thalamocortical development.