Craig Nicholson

Synaptogenesis: Altering your inhibitions

development is regulated by the excitatory input that neurons receive. The molecular mechanisms that lead from excitation to the establishment of excitatory synapses are quite well understood, but less is known about the mechanisms that modulate the development of inhibitory synapses. Greenberg and colleagues have now demonstrated the involvement of the activityregulated transcription factor NPAS4 in the modification of both types of synapse. The authors used DNA microarrays and other means to identify transcription factors that are regulated specifically by Ca2+ influx into neurons during the period of development in which inhibitory synapses are maturing. They found just one, NPAS4, and went on to investigate whether this protein has a role in the regulation of synapses. As a marker of inhibitory synapses, the authors used the presence of opposed GAD65 (a GABA-producing enzyme) puncta at the presynaptic membrane and GABAA-γ2 (a GABAA receptor subunit) puncta at the postsynaptic membrane. Immunostaining revealed that knocking down NPAS4 with RNA interference (RNAi) in dissociated, immature rat hippocampal cultures significantly reduced the colocalization of the two types of puncta (and, by proxy, the number of inhibitory synapses), whereas overexpressing NPAS4 had the opposite effect. The density of the GABAA-γ2 puncta was most affected, and so the authors inferred that NPAS4 regulates inhibitory synapses through effects on postsynaptic specializations. By contrast, manipulation of NPAS4 levels had no effect on the development of excitatory synapses. The authors also investigated the effects of NPAS4 on synapses in functional circuits in rat organotypic hippocampal slices. Acute removal of Npas4 from these cells using a Cre recombinase system increased the interval between miniature inhibitory postsynaptic currents, indicating that there was a decrease in inhibitory inputs onto these neurons; conversely, it decreased the interval between miniature excitatory postsynaptic currents, indicating that there was an increase in the number and/or strength of excitatory synapses. Overexpression of NPAS4 had the opposite effect on both types of current. These results indicate that NPAS4 regulates the number of inhibitory synapses in the developing brain and has an important role in modulating the strength and number of both inhibitory and excitatory synapses in response to excitatory input in the mature nervous system. Further characterization of the genes that are regulated by NPAS4 will improve our understanding of how neuronal activity modulates synaptic function. Craig Nicholson

Transport: Keeping motors running smoothly

containing secretory vesicles are transported along microtubules from the proximal golgi apparatus to the distal pre-synaptic membrane, and various signalling molecules are transported from the pre-synaptic membrane to the nucleus. The structurally and functionally diverse molecular motor proteins dynein and kinesin mediate this transport, which is modulated by microtubuleassociated proteins (MAPs). A new study by Holzbaur and colleagues further characterizes the role of tau, a MAP that has been implicated in the pathology of several neurodegenerative diseases, in anterograde and retrograde vesicle transport. Anterograde transport is directed towards the pre-synaptic membrane and is carried out by kinesin, which can move only unidirectionally along microtubules. Conversely, dynein can move bi-directionally in vitro but carries out only retrograde transport in the cell. The binding of these motor proteins to microtubules, and their motility once they are bound, is affected by MAPs. Tau is one such MAP, and has been shown to inhibit kinesin activity in vitro and in vivo; however, the specifics of this inhibition and the effects of tau on dynein were largely unknown. To gain further insight, the authors sought to visualize individual interactions between microtubule-bound tau and kinesin and dynein in cell-free studies in vitro. First they used total internal reflection fluorescence microscopy to visualize the effect of fluorescent Alexa-546-tagged recombinant tau on green fluorescent protein (GFP)tagged dynein and kinesin. They found that tau affected the binding of kinesin to microtubules, but not that of dynein: kinesin binding decreased as local microtubule-bound tau concentrations increased. Tau also affected kinesin and dynein motility differently: when kinesin motors encountered a high concentration of microtubule-bound tau, they detached from the microtubule. By contrast, dynein motors reversed direction at tau clusters. There are many neuronal isoforms of tau, all of which are comprised of two domains: a projection domain that is thought to have a role in recruiting dynein activators; and a microtubule-binding domain, which can contain different numbers of repeated elements. The authors wanted to quantify the concentration-dependent effects of tau, and investigate whether the different isoforms produced different effects. When they incubated microtubules with different concentrations of the shortest and longest isoforms — tau23 and tau40, respectively — they found that the binding and motility of both kinesin and dynein decreased in a tau-concentrationdependent manner. However, only kinesin was affected at physiological concentrations of tau23, and neither kinesin nor dynein was affected by physiological concentrations of tau40. To further characterize these effects, the authors used recombinant tau23 proteins with truncated projection domains but intact microtubulebinding domains. They found that these truncated proteins were stonger inhibitors of kinesin and dynein binding and motility than wild-type tau23, and thus that the microtubulebinding domain of tau is sufficient for its inhibitory effects. The authors posit that the projection domain of tau recruits motor proteins to the microtubule; however, this hypothesis requires further investigation. This study shows that tau has a modulatory role in axonal transport. A proximal-to-distal gradient of tau, which is known to exist in healthy neurons, would allow kinesin to bind to microtubules at the cell body and then translocate distally, at which point the higher concentration of tau would cause the kinesin to detach. Conversely, dynein would still be able to bind to the microtubules distally and mediate retrograde transport. An additional level of control that might be afforded by modulatory phosphorylation of tau remains to be explored. Craig Nicholson

Neuron–glia interactions: Parting the waves

Glia can discriminate patterns of synaptic activity and modulate synaptic plasticity through balanced A1 and A2 adenosine receptor activation.

Neurodegenerative disorders: ASIC mind

for neuronal degeneration in multiple sclerosis (MS) are poorly understood. A study by Fugger and colleagues now shows that acid-sensitive ion channel 1 (ASIC1), a neuronally expressed, proton-gated cation channel, might contribute to the pathophysiology of this disease. MS is a neuroinflammatory disease that leads to the demyelination and axonal degeneration of CNS neurons. The degeneration is thought to be caused by inflammation-induced mitochondrial dysfunction and by increased influx of Na+ and Ca2+ ions into the neurons. This led the authors to investigate the role of ASIC1 in a mouse model of MS. The authors induced experimental autoimmune encephalomyelitis (EAE; a mouse model of MS) in wild-type mice and mice that lacked either one or both copies of Accn2, the gene that encodes ASIC1. They found that the Accn2–/– mice displayed significantly reduced clinical symptoms compared with both the Accn2+/– mice and the wild-type mice, implicating ASIC1 in the neurodegeneration. As ASIC1 is proton-gated, the authors used micro pH meters to measure the acidity of mouse spinal cords, and found that the extracellular pH of EAE-induced mice, both wild-type and Accn2–/–, was significantly lower than that of control mice. Furthermore, by contrast with the control mice, the spinal pH levels in the EAE-induced mice were lower than the pH that is required for the opening of the proton-gated pores of ASIC1. This acidosis in the neural tissue of EAE-induced mice has been proposed to be the result of axonal hypoxia and inflammation-induced mitochondrial dysfunction. Indeed, the levels of hypoxia-inducible factor 1α were elevated in the spinal cords of the EAE-induced mice. To further examine ASIC1’s role in the development of the neurodegenerative pathophysiology, the authors measured Accn2 mRNA levels in retinal and cerebellar neurons throughout the course of EAE progression in wild-type mice. ASIC1 expression has previously been shown to be upregulated by proinflammatory mediators and, in accordance, the authors detected significantly elevated Accn2 mRNA levels in both types of neurons 15 and 30 days after EAE induction. Finally they investigated whether inhibition of ASIC1 after EAE induction might exert a neuroprotective effect in vivo. They showed that the clinical symptoms of wild-type, EAEinduced mice treated with amiloride, a non-specific ASIC blocker that is licensed for the treatment of hypertension, were reduced to a similar extent as in the Accn2–/–, EAE-induced mice. Treatment of Accn2–/–, EAE-induced mice with amiloride caused no further clinical improvement, suggesting that the effect of the drug is mediated through ASIC1. Together these results indicate that ASIC1 might contribute to the pathophysiology of MS, and that compounds that block the channel, such as amiloride, might be valuable therapeutics for slowing the progression of this disease. Craig Nicholson

Neural circuits: Sing a song of “sex please”

females with a courtship ‘song’ created by distinctive wing vibrations. The females themselves do not normally sing but, by inducing male-like singing in genetically manipulated females, a new study has shown that part of the neural circuitry that underlies this male-specific behaviour is also functional in the opposite sex. The male-specific fly courtship song is thought to be generated by neurons in the thoracic ganglion, which are in turn controlled by descending neurons from the brain. However, the neural circuitry of these regions is anatomically very similar in males and females, so it was puzzling why females do not also sing for sex. Much of the reproductive behaviour of D. melanogaster stems from the actions of the various transcription-factor products of the fruitless (fru) gene, which is expressed only in neurons. Owing to alternative mRNA splicing, some of these proteins are gender-specific — the male-specific fru products are collectively known as FruM. Miesenböck and Clyne therefore investigated whether the lack of singing in females might be due to the gender-specific differences in Fru-protein-containing neurons. They engineered flies that expressed a light-activatable cation channel under the control of the fru gene, and then exposed these flies’ thoracic neurons to light. They found that photoactivation of these fruexpressing neurons readily induced singing in males. Remarkably, singing could also be induced in this manner in females, but only when a fourfold higher light intensity was used. This showed that the motor programme for singing can be turned on artificially in females, albeit at a higher activation threshold than is required for males. Why, then, do females not normally sing? To investigate this matter, the authors generated FruM-expressing female flies. These FruM females, which have a largely male nervous system but an otherwise female physiology, could be induced to sing by the same light intensity as male flies. Furthermore, an analysis of the songs themselves revealed that, whereas the songs of the FruF female flies were off key, those of the FruM females were very similar to the true courtship song of normal male flies. Thus, neural components or properties that are specifically imparted by the actions of FruM are needed to activate the thoracic motor programme and control the execution of the courtship song. This study shows that female D. melanogaster possess the motor programme that allows males to generate the courtship song, but that Fru-expression-dependent differences from the male neural circuitry prevent the programme being activated in females. Craig Nicholson

Sing a song of “sex please”: Neural circuits

females with a courtship ‘song’ created by distinctive wing vibrations. The females themselves do not normally sing but, by inducing male-like singing in genetically manipulated females, a new study has shown that part of the neural circuitry that underlies this male-specific behaviour is also functional in the opposite sex. The male-specific fly courtship song is thought to be generated by neurons in the thoracic ganglion, which are in turn controlled by descending neurons from the brain. However, the neural circuitry of these regions is anatomically very similar in males and females, so it was puzzling why females do not also sing for sex. Much of the reproductive behaviour of D. melanogaster stems from the actions of the various transcription-factor products of the fruitless (fru) gene, which is expressed only in neurons. Owing to alternative mRNA splicing, some of these proteins are gender-specific — the male-specific fru products are collectively known as FruM. Miesenböck and Clyne therefore investigated whether the lack of singing in females might be due to the gender-specific differences in Fru-protein-containing neurons. They engineered flies that expressed a light-activatable cation channel under the control of the fru gene, and then exposed these flies’ thoracic neurons to light. They found that photoactivation of these fruexpressing neurons readily induced singing in males. Remarkably, singing could also be induced in this manner in females, but only when a fourfold higher light intensity was used. This showed that the motor programme for singing can be turned on artificially in females, albeit at a higher activation threshold than is required for males. Why, then, do females not normally sing? To investigate this matter, the authors generated FruM-expressing female flies. These FruM females, which have a largely male nervous system but an otherwise female physiology, could be induced to sing by the same light intensity as male flies. Furthermore, an analysis of the songs themselves revealed that, whereas the songs of the FruF female flies were off key, those of the FruM females were very similar to the true courtship song of normal male flies. Thus, neural components or properties that are specifically imparted by the actions of FruM are needed to activate the thoracic motor programme and control the execution of the courtship song. This study shows that female D. melanogaster possess the motor programme that allows males to generate the courtship song, but that Fru-expression-dependent differences from the male neural circuitry prevent the programme being activated in females. Craig Nicholson

Neural circuits: Sing a song of |[ldquo]|sex please|[rdquo]|

females with a courtship ‘song’ created by distinctive wing vibrations. The females themselves do not normally sing but, by inducing male-like singing in genetically manipulated females, a new study has shown that part of the neural circuitry that underlies this male-specific behaviour is also functional in the opposite sex. The male-specific fly courtship song is thought to be generated by neurons in the thoracic ganglion, which are in turn controlled by descending neurons from the brain. However, the neural circuitry of these regions is anatomically very similar in males and females, so it was puzzling why females do not also sing for sex. Much of the reproductive behaviour of D. melanogaster stems from the actions of the various transcription-factor products of the fruitless (fru) gene, which is expressed only in neurons. Owing to alternative mRNA splicing, some of these proteins are gender-specific — the male-specific fru products are collectively known as FruM. Miesenböck and Clyne therefore investigated whether the lack of singing in females might be due to the gender-specific differences in Fru-protein-containing neurons. They engineered flies that expressed a light-activatable cation channel under the control of the fru gene, and then exposed these flies’ thoracic neurons to light. They found that photoactivation of these fruexpressing neurons readily induced singing in males. Remarkably, singing could also be induced in this manner in females, but only when a fourfold higher light intensity was used. This showed that the motor programme for singing can be turned on artificially in females, albeit at a higher activation threshold than is required for males. Why, then, do females not normally sing? To investigate this matter, the authors generated FruM-expressing female flies. These FruM females, which have a largely male nervous system but an otherwise female physiology, could be induced to sing by the same light intensity as male flies. Furthermore, an analysis of the songs themselves revealed that, whereas the songs of the FruF female flies were off key, those of the FruM females were very similar to the true courtship song of normal male flies. Thus, neural components or properties that are specifically imparted by the actions of FruM are needed to activate the thoracic motor programme and control the execution of the courtship song. This study shows that female D. melanogaster possess the motor programme that allows males to generate the courtship song, but that Fru-expression-dependent differences from the male neural circuitry prevent the programme being activated in females. Craig Nicholson

Neurogenesis: Par for the course

between proliferation and differentiation of progenitor cells ultimately defines the size of the brain. Progenitor cells of the ventricular zone (VZ) show an enrichment of partitioning-defective (Par) proteins — well-established polarity markers — at their apical surface. Götz and colleagues now demonstrate that Parcomplex proteins have a crucial role in maintaining the potential of VZ progenitor cells to proliferate. Progenitor cells in the VZ can either continue to proliferate or start differentiating. The molecular basis for the switch from proliferation to differentiation is the subject of intense investigation. Par-complex proteins have a well-characterized role in neuroblast division in Drosophila and have previously been implicated in VZ-progenitor division in vertebrates. The authors used immunostaining to examine Par-complex-protein expression at various stages of mouse cortical development. Several Parcomplex proteins, including Par3 and Par6, were strongly expressed at embryonal day (E) 12 but significantly lower levels were detected at E14 and E16 — when neurogenesis is reaching its midpoint and proliferation is trailing off. Next the authors used shorthairpin RNA (shRNA) to knockdown Par3 expression in E12 cortical progenitors in vitro. They found that this decreased the number of cells arising from each progenitor and increased the proportion of differentiated versus non-differentiated daughter cells, indicating that a lack of Par3 results in premature differentiation of progenitor cells. In subsequent in vivo studies in which Par3 expression was reduced by shRNA-containing lentiviral vectors, the transduced progenitor cells began to differentiate soon after transduction and also remained resident in the lower cortical layers, rather than migrating upward with the non-transduced progenitors. To further characterize the role of Par-complex proteins in VZprogenitors, the authors used retroviral vectors to overexpress Par3 and Par6 both in vitro and in vivo. This resulted in a significant increase in progenitor proliferation and a coincident decrease in differentiation. The authors also showed that these manipulations had no influence on the length of the cell cycle. Together these results indicate that Par-complex proteins are key for the proliferation potential of VZ progenitor cells, and thus for determining the size and thickness of the cortex. Craig Nicholson

Learning and memory: Day to remember

be influenced by the time of day, but the nature and mechanism of this modulation has been elusive. Now, a new study shows that melatonin, a hormone released in a circadian fashion, affects memory consolidation in zebrafish. Cahill and colleagues assessed memory formation in diurnal zebrafish using an active-avoidance conditioning paradigm, in which the zebrafish learned to associate a particular compartment of their tank with mild electric shocks. They found that zebrafish that were conditioned during the day learned more quickly than those that were conditioned at night and had significantly better retention when tested 24 hours after conditioning. Furthermore, when the zebrafish were kept in constant darkness for several days, those that were conditioned during the period of their circadian cycle that corresponded to daytime (their subjective daytime; SD) outperformed those that were conditioned during their subjective night (SN). This implicated the endogenous circadian system in the effect of time-of-day on learning. To ascertain whether the modulation of learning ability was due to altered memory formation or memory retrieval, the authors again used zebrafish that had been kept in constant darkness for several days. They found that those that were conditioned during their SD and then tested during their SN 36 hours later had significantly higher memory retention than those that were conditioned during their SN and tested during their SD. Thus, the circadian system was affecting memory formation rather than retrieval. Melatonin release peaks during the night and falls during the day, and melatonin has been shown to affect neuronal firing in the hippocampus. The authors therefore decided to investigate whether melatonin mediates the effects of the circadian system on memory formation. They found that bathing the zebrafish in 50 μM melatonin prior to SD conditioning significantly suppressed memory formation, whereas administration after conditioning or prior to testing had no effect. Furthermore, administration of a melatonin-receptor antagonist prior to SN conditioning significantly improved memory retention, as did removal of the pineal gland, the site of melatonin release. Taken together, these results show that memory formation in zebrafish is inhibited during the night relative to the day, and that this modulation is mediated at least in part by circadian melatonin release. This might direct future research into improving mental performance in humans. Craig Nicholson

Learning and memory: Cease or persist?

The establishment of persistent long-term memory (LTM) requires an initial consolidation phase and a subsequent persistence-bestowing phase. It has previously been shown that hippocampal protein synthesis and brain-derived neurotrophic factor (BDNF) expression are necessary for LTM consolidation. Here, Medina and colleagues show that BDNF is sufficient to confer persistence to long-term memories. Infusion of the protein-synthesis inhibitor anisomycin into the dorsal hippocampus of rats 12 hours after training is known to selectively abrogate the persistence phase of LTM formation: rats that undergo this treatment exhibit consolidated memory of the training after 2 days but the memories do not persist for 7 days. The authors used an inhibitory–avoidance protocol to determine the effect of hippocampal delivery of human recombinant BDNF (hrBDNF) on this deficit. They found that delivery of hrBDNF 15 minutes after anisomycin infusion completely rescued the memory impairment: whereas rats treated with anisomycin alone exhibited reduced latency in encroaching onto a region which they had been trained to associate with an electrical footshock, rats that were also treated with hrBDNF exhibited normal latency. To reinforce their findings, the authors investigated the interplay between BDNF and the strength of the applied footshock. Stronger footshocks are known to induce longer-lasting memories. The authors found that a strong footshock that induced LTM formation increased BDNF expression in the dorsal hippocampus, whereas a footshock too weak to induce LTM formation did not. Significantly, infusion of hrBDNF 12 hours after delivery of a weak footshock resulted in the formation of a persistent LTM. How does BNDF exert its effects on LTM persistence? BDNF is known to phosphorylate and activate ERK protein, and the authors showed that a footshock that increased BDNF levels also induced ERK phosphorylation 12 hours post-shock. Furthermore, infusion of an ERK inhibitor 12 hours post-shock impaired 7-day LTM retention. Together, these results show that BDNF is necessary and sufficient to induce LTM persistence through its effects on the activity of ERK. These findings might have implications for the treatment of age-related memory impairments, some of which are thought to result from faulty persistence. Craig Nicholson