cAMP-dependent protein kinase 1

PKA as a target of Hedgehog acting through the G-protein Smoothened

The simplest possible mechanism by which HH relieves inhibition by Patched takes into consideration the fact that both SMO and PTC are integral membrane proteins and assumes that PTC is inactivated by its association with the HH-SMO complex. It is also possible that PTC in its active form is already associated with SMO in the absence of bound HH. In this case, when HH binds to the SMO-PTC complex, PTC is released from its association with SMO thus abolishing PTC's inhibitory activity on Fused. Additionally, PKA becomes inhibited by G protein-coupled signaling (Alcedo, 1996).

smo is required for hedgehog-dependent expression of decapentaplegic. In the case of the wing disc, the principal target of hh is dpp, whose expression is restricted to a thin strip of cells running along the anterior side of the anterior-posterior compartment boundary. Mosaic imaginal discs, in which small clones of cells within the normal dpp domain lack wild-type smo activity, fail to express dpp in the mutant region. Clones that lack smo and the cAMP-dependent protein kinase (PKA) catalytic subunit, do however express dpp, indicating that smo is not absolutely required for dpp transcription; rather, it acts upstream of PKA to mediate activation of dpp by hh (van den Heuval, 1996).

Modification of PKA activity during learning in the honeybee

To investigate the function cAMP-dependent protein kinase (PKA) exerts in the induction of long-term memory, changes in PKA activity induced by associative learning in vivo were measured in the antennal lobes (ALs) of honeybees. The temporal dynamics of PKA activation depend on both the sequence of conditioned and unconditioned stimuli and the number of conditioning trials. Only multiple-trial conditioning, which induces long-term memory (LTM), leads to a profound prolongation of PKA activation mediated by the NO/cGMP system (see Drosophila Nos). Imitation of this prolonged PKA activation in the ALs in combination with single-trial conditioning is sufficient to induce LTM. These findings not only demonstrate the close connection between conditioning procedure and temporal dynamics in PKA activation but also reveal that already during conditioning a distinct temporal pattern of PKA activation is critical for LTM induction in intact animals (Müller, 2000).

Associative olfactory conditioning of the proboscis extension response (PER) in the honeybee induces different forms of memory, depending on the number of conditioning trials. The memory induced by a single conditioning trial decays over several days and is sensitive to amnestic treatments. This memory is independent of NO synthase (NOS) blockers and of protein synthesis. In contrast, multiple conditioning trials induce a stable, long-lasting memory. This memory is dissectable into two independent, parallel phases. The first phase is a medium-term memory (MTM) in the hours range, which requires a constitutive PKC activity, and the second phase is an LTM (1 day or more), which requires PKA- and NO-dependent processes. Interestingly, LTM can be divided into an early phase (eLTM, 1-2 days) and a protein synthesis-dependent late phase (lLTM, 3 days or more)(Müller, 2000 and references therein).

Even though it has been demonstrated that the cAMP cascade is important for the induction of long-lasting neuronal and behavioral changes, the findings presented here reveal evidence for a direct connection between conditioning procedure, temporal dynamics in PKA activation, and their contribution to formation of LTM in intact animals. Direct measurement of changes in PKA activity in the ALs induced by in vivo stimulation reveals that multiple conditioning trials that induce LTM also induce an extremely prolonged PKA activation. The latter contributes to LTM formation processes, since imitation of the extended PKA activation in the AL in conjunction with a single conditioning trial induces LTM (Müller, 2000).

Recent findings suggest that a distinct temporal activation of the cAMP cascade, dependent on distinct stimulation parameters, is required for the induction of long-lasting neuronal and behavioral changes. A close connection has been demonstrated between stimulation parameters and the temporal dynamics of changing cAMP levels, adenylate cyclase activity, PKA activity, and CREB phosphorylation (Müller, 2000 and references therein).

During associative conditioning in the honeybee, the temporal dynamics of PKA activation in the ALs depend on both the sequence of CS and US stimulation and the number of conditioning trials. The mechanism underlying the sequence-dependent PKA activation is distinguishable and independent from that underlying multiple trial-induced PKA activation, as demonstrated by selective impairment of the latter by blocking NOS activity. Regardless of the number of trials, US/CS backward pairing and US stimulation induce the same transient PKA activity in the ALs. The US-induced PKA activation in the ALs is mediated by octopamine (see Tyramine β hydroxylase). The octopamine in the ALs is most likely released by the VUMmx1 neuron, which has been shown to substitute for the US function in associative olfactory conditioning. The extensive aborizations of the VUMmx1 neuron in the ALs suggest that the US-mediated PKA activation occurs within all AL glomeruli. In contrast to this, CS stimulation induces odor-specific changes in Ca2+ levels in distinct subsets of glomeruli. Thus, it is conceivable that the prolonged PKA activation induced by CS/US forward pairing is due to a sequence-dependent interaction between an odor-specific Ca2+-mediated process in distinct glomeruli and a general US/octopamine-mediated process (Müller, 2000 and references therein).

Dually regulated enzymes, like the Ca2+/calmodulin-dependent adenylate cyclase, have been suggested as molecular convergence sites of different inputs important for neuronal plasticity and learning. In membrane fractions of Aplysia neurons, the maximal in vitro activation of the Ca2+/calmodulin-dependent adenylate cyclase is achieved when the Ca2+ stimulus precedes the transmitter stimulus. Although direct evidence is lacking, the dually-regulated adenylate cyclase may be implicated in the sequence-specific prolongation of PKA activity induced by a CS/US forward pairing in the honeybee (Müller, 2000).

Findings from Aplysia and Drosophila assign a critical role for induction of long-lasting changes to the balance of activator and repressor isoforms of CREB. The results from the honeybee, however, show that already during the short conditioning time window a distinct temporal pattern of PKA activation is critical for LTM induction. Assuming a similar connection between the PKA pathway and CREB in the induction of LTM in honeybees, future investigations demand a characterization of whether and how the multiple trial-induced prolonged PKA activation acts on CREB function. But since regulation of CREB isoforms and their function in long-term neuronal and behavioral changes are results of a complex interaction of different second messenger systems, it is very likely that different signaling cascades contribute to LTM formation (Müller, 2000 and references therein).

It is conspicuous that both the formation of multiple trial-induced LTM and the prolonged PKA activation in the ALs require NO-mediated mechanisms. The finding that photorelease of cGMP in the ALs in combination with single-trial conditioning induces LTM supports the idea that the NO/cGMP system within the ALs mediates the prolongation of PKA activation during conditioning. Although the neurons containing the NO-activated guanylate cyclase have not been described in the honeybee, it is most likely that the NO-releasing neurons that modulate cGMP levels in the target cells are located within the ALs. Uncaging NO in the entire AL in combination with single-trial conditioning, however, leads to a significant reduction in conditioned PER as tested at 3 hr and 3 days. Although the reason for this learning impairment is unknown, photolyzing NO in the entire AL probably interferes with a specific function of NO in signal processing during olfactory learning. The latter is very likely, since an odor induces changes in Ca2+ concentrations in a subset of glomeruli only. This in turn results in the activation of the Ca2+-dependent NOS and thus in release of NO in a characteristic subset of glomeruli only. Possibly such a CS-specific release of NO within a subset of glomeruli contributes to aspects of olfactory signal processing required for learning (Müller, 2000 and references therein).

In this context it is interesting to note that multiple conditioning trials lead to more specific responses and thus may be based on more specific synaptic plasticity. The latter may be due to a Hebbian mechanism of pre/postsynaptic activity detection. It has been proposed that such a mechanism may also be essential for invertebrates and that NO may play a central function as a retrograde signaling molecule (Müller, 2000 and references therein).

A series of studies in Drosophila convincingly demonstrate that the mushroom bodies (MBs) are essential for olfactory learning, and that they support context generalization in visual learning, and are required for memory formation of courtship conditioning. These findings not only support the important role of the MBs as multisensory processing centers but also demonstrate that the contribution of the MBs differs, depending on the learning paradigm and the sensory modality used. In contrast to the considerable knowledge with regard to the function of the MBs in Drosophila learning, it was only recently proposed that the ALs contribute to short-term memory in Drosophila courtship conditioning (Müller, 2000 and references therein).

In honeybees, it has been demonstrated that initial olfactory memory (tested 20 min after conditioning) can be induced independently in either the MBs or the ALs. In contrast to Drosophila, however, the majority of studies focused on the function of the ALs in olfactory learning. It has been demonstrated that differential olfactory conditioning causes changes in the neural representation of the rewarded and the unrewarded odor in the ALs for up to at least 30 min after conditioning. Moreover, the requirement of a constitutively active PKC in the ALs for multiple trial-induced MTM suggests that processes located in the ALs contribute to memory maintenance in the range of hours. The results presented here now demonstrate that a prolonged PKA activation in the ALs induced by multiple-trial conditioning is implicated in induction of LTM. However, since imitation of prolonged PKA activation in conjunction with single-trial conditioning does not reach the level of conditioned PER after multiple-trial conditioning, a contribution by other brain areas must be proposed. Collectively, all these findings provide evidence that the ALs are sites that contribute to processes of associative olfactory learning during the conditioning procedure itself and in early phases of memory formation for up to several hours. Moreover, the ALs are possibly also sites of long-lasting structural changes. Activity-dependent changes described for the glomerular volume in the ALs may well be the result of structural plasticity underlying long-term memory (Müller, 2000 and references therein).

Interestingly, in mice and sheep the accessory olfactory system has also been implicated in the formation of olfactory memory. While female mice form a memory of the pheromones of the mating male, sheep learn to recognize the odors of their lambs in the first hours after birth. In both cases, NO has been demonstrated to mediate the formation of this memory. In mice the coincident activation of pheromonal inputs and exogenous administration of NO in the accessory olfactory system can induce a pheromone-specific olfactory memory without mating. Blocking of NOS activity in the olfactory system of sheep prevents the formation of olfactory memory. Although the targets of the NO/cGMP system in the olfactory systems of honeybees, mice, and sheep differ, the conspicuous parallels suggest a conserved function of NO-mediated signaling in the olfactory systems with respect to olfactory memory formation (Müller, 2000 and references therein).

The role of cAMP-dependent protein kinase (PKA) in associative olfactory learning of the honeybee, Apis mellifera, has been examined. In the bee, specific interference with molecules to clarify their role in a certain behavior is difficult, because genetic approaches, such as mutants or transgenic animals, are not feasible at the moment. As a new approach in insects in vivo, use of short antisense oligonucleotides is reported here. Phosphorothioate-modified oligodeoxynucleotides complementary to the mRNA of a catalytic subunit of PKA directly injected into the bee brain cause a reversible and specific downregulation of both the amount of the catalytic subunit and of PKA activity by 10%-15%. The amounts of the regulatory subunit of PKA, as well as PKC, are not affected. The slight 'knockdown' of PKA activity during the training procedure, a classical olfactory conditioning of the proboscis extension reflex, neither affects acquisition nor memory retention at 3 or at 6 hr after training. However, it causes an impairment of long-term memory retention 24 hr after training. Downregulation of PKA 3 hr after training has no detectable effect on memory formation. It is concluded that PKA contributes to the induction of a long-term memory 24 hr after training when activated during learning. The antisense technique is feasible in honeybees in vivo and provides a new and powerful tool for the study of the molecular basis of learning and memory formation in insects (Fiala, 1999).

Changes in sleep are achieved by spatial and temporal enhancement of cyclic-AMP-dependent protein kinase (PKA) activity specifically in the adult mushroom bodies

Sleep is one of the few major whole-organ phenomena for which no function and no underlying mechanism have been conclusively demonstrated. Sleep could result from global changes in the brain during wakefulness or it could be regulated by specific loci that recruit the rest of the brain into the electrical and metabolic states characteristic of sleep. This study addresses this issue by exploiting the genetic tractability Drosophila, which exhibits the hallmarks of vertebrate sleep. Large changes in sleep are achieved by spatial and temporal enhancement of cyclic-AMP-dependent protein kinase (PKA) activity specifically in the adult mushroom bodies of Drosophila. Other manipulations of the mushroom bodies, such as electrical silencing, increasing excitation or ablation, also alter sleep. These results link sleep regulation to an anatomical locus known to be involved in learning and memory (Joiner, 2006).

To determine whether specific brain loci regulate sleep, the GAL4/UAS (upstream activating sequence) system was used to express a catalytic subunit of PKA in various regions of the fly brain. PKA was first expressed under the control of the RU486-inducible pan-neuronal driver elavGeneSwitch. Restricting the expression of PKA to adult neurons decreased daily sleep, supporting earlier studies with mutants such as dunce that increase PKA levels, and showing that PKA directly regulates sleep rather than a developmental process that might affect sleep. PKA was expressed under the control of different GAL4 drivers, and the changes in total daily sleep were examined in the different driver/transgene combinations relative to driver/background and background/transgene controls. When both controls were taken into account, the expression of PKA by only two drivers led to changes in sleep that exceeded 2 s.d. These were 201Y, which increased sleep by 75 +/- 3% and 93 +/- 4% respectively, and c309, which decreased sleep by 46 +/- 11% and 43 +/- 14% per day compared with the two sets of controls. Changes in sleep caused by all other GAL4 drivers remained within 1 s.d. of the mean (Joiner, 2006).

Next, whether activity levels during wake periods were affected by the 201Y and c309 drivers was examined. Many GAL4 driver/UAS-PKA lines were hypoactive, but line 201Y had normal waking activity. Similarly, activity normalized to waking time in c309 was not significantly higher in PKA-driven animals than in either control, indicating that c309 was not hyperactive. It is concluded that the sleep phenotypes of animals expressing PKA under control of the 201Y and c309 drivers are not associated with abnormal waking activity. Interestingly, both these drivers are known to be expressed in the mushroom bodies (MBs), a brain region implicated in associative learning (Joiner, 2006).

Given the strong, yet opposite, effects that 201Y and c309 had on sleep, their expression patterns in the fly brain were further characterized by crossing them into animals bearing a UAS transgene for green fluorescent protein (GFP). It was found that 201Y is expressed largely in the γ lobes and the core region of the α/β lobes of the MBs, whereas c309 is expressed in the α/β and γ lobes but not in the core region of the α/β; lobes. This differential expression pattern within the MBs indicates that PKA might affect the regulation of sleep by the MBs in both a positive and a negative fashion by using anatomically distinct classes of neurons. Consistent with this notion of heterogeneous cell types within the MBs, some MB drivers, such as 30Y and 238Y, promoted sleep during the day but inhibited sleep during the night, leading to only marginal overall changes in daily sleep. This effect was not observed with any driver that was expressed exclusively outside the MBs. A small increase in daytime sleep was also frequently produced by the pan-neuronal elavGeneSwitch driver, which decreased overall sleep. The expression patterns of 238Y and 30Y overlap those of 201Y and c309, supporting the idea that 238Y and 30Y are expressed in both sleep-promoting and sleep-inhibiting areas (Joiner, 2006).

To test the hypothesis that PKA expression in MBs regulates adult sleep, the PKA transgene was expressed under the control of an RU486-activatable MB GAL4 driver, P{MB-Switch}. It was confirmed selective expression of this driver in the MBs by coupling it to a GFP reporter, and inducible expression was found in the MBs. Sleep was significantly reduced in response to RU486 in MB-Switch/PKA animals but was unaffected by the drug in control animals harbouring either the driver or the transgene alone. Thus, PKA overexpressed preferentially in specific neurons of adult MBs is sufficient to reduce sleep (Joiner, 2006).

Next sleep structure in the hyposomnolent animals was compared with that of controls. In both MB-Switch/PKA animals and c309/PKA animals, loss of sleep was caused by a shortened sleep bout duration without a concomitant increase in the sleep bout number. The underlying cause of reduced sleep in both sets of animals therefore seems to be impaired sleep need, because the alternative-normal sleep need, but an inability to maintain the sleep state-would be expected to produce an increase in sleep bout number. In contrast, in 201Y sleep bout duration remained unchanged (Joiner, 2006).

It was then asked whether the reduction of sleep in MB-Switch/PKA animals was due to an impaired accrual of a sleep-inducing signal. If this were so, then a hallmark of sleep, homeostatic rebound-sleep that exceeds baseline to compensate for lost sleep-should not occur on relief of induced PKA expression. However, when RU486 was withdrawn after about three days of sleep deprivation, an average rebound of 156 +/- 38 min was observed. This is a robust rebound, comparable to that produced when genetically identical but uninduced flies were submitted to a standard 12 h of mechanical deprivation (137 +/- 26 min). Behavioural rebound was also observed in animals expressing elavGeneSwitch-driven PKA, after withdrawal of RU486, and was accompanied by a decrease in PKA activity in fly heads. Rebound after withdrawal of RU486 indicates that PKA might not prevent the accrual of sleep-promoting signals but might suppress homeostatic output (Joiner, 2006).

To determine whether PKA affects sleep by regulating synaptic output in MB neurons, either of two K⁺ channels, Kir2.1 or EKO, were inducibly expressed under the control of the MB-Switch driver. Such transgenic expression should suppress action-potential firing by hyperpolarizing neurons and decreasing membrane resistance, thus leading to reduced synaptic transmission. It was found that induction of either Kir2.1 or EKO caused a significant increase in sleep. Because the opposite was observed with PKA expression in the same neurons, it indicates that PKA might decrease sleep by increasing either excitability or synaptic transmission. To address this issue further, a sodium channel (NaChBac), which depolarizes neurons and increases excitability, was inducibly expressed. When expressed under the control of the MB-Switch driver, the sodium channel caused a decrease in sleep, similar to that produced by PKA, confirming that PKA increases the output of these neurons (Joiner, 2006).

The MB-Switch driver is expressed in a subpopulation of MB neurons similar to those labelled by c309, and both drivers had sleep-inhibiting effects. As noted above, this pattern of expression differed from that of other drivers, which had sleep-promoting or bidirectional effects on sleep, thus leading to a proposal that the MBs contain sleep-inhibiting and sleep-promoting neurons. To determine the overall effect of MBs on sleep, they were ablated with hydroxyurea, and sleep and activity were examined in adult flies. An overall increase in activity was observed. However, normalization of this activity to waking time indicates that the phenotype derives less from hyperactivity than from a reduction in sleep. Even so, the reduction in sleep was much less than that seen with other manipulations of the MBs or in short-sleep mutants such as minisleep. This supports the conclusion that MBs exert both positive and negative influences on sleep that are integrated to produce the overt behavioural state. A model takes into account these results; notably the integrator downstream of the MBs promotes activity in the default state. Thus, when MBs are ablated the overall effect is increased wakefulness (Joiner, 2006).

Opposing effects of the c309 and 201Y drivers are also observed in a different behavioural model. They parallel MB-dependent changes in brain activity during the sleep/wake cycle that are associated with salience, a behavioural trait that may correspond to arousal. Consistent with the data was the observation that reducing synaptic transmission using the c309 driver inhibited salience, whereas the 201Y driver in the same type of experiment yielded no change. Increased arousal wouldbe predicted with 201Y, but in those experiments the animals were already awake (Joiner, 2006).

Because MBs receive and transduce considerable sensory, particularly olfactory, input to the fly brain, it is speculate that they promote arousal or sleep by allowing or inhibiting the throughput of sensory information. In addition, given the major function that MBs have in regulating plasticity in the fly brain, it is likely that this is linked to their role in sleep. In mammals, sleep deprivation suppresses the performance of learned tasks, and sleep permits memory consolidation. Sleep and sleep deprivation also differentially affect cortical synaptic plasticity. In Drosophila, MBs participate in the consolidation or retrieval of memories involving olfactory cues, courtship conditioning and context-dependent visual cues by mechanisms that include cAMP signalling. Distinct anatomical regions of the MBs have been shown to be important for at least some forms of memory, as has now also been shown for sleep. Thus, memory and sleep may involve similar molecular pathways (cAMP signalling) and anatomical regulatory loci (MBs) (Joiner, 2006).

Modulation of L-type calcium channels in Drosophila via a pituitary adenylyl cyclase-activating polypeptide (PACAP)-mediated pathway that includes via cAMP and PKA

Modulation of calcium channels plays an important role in many cellular processes. Previous studies have shown that the L-type Ca(2+) channels in Drosophila larval muscles are modulated via a cAMP-protein kinase A (PKA)-mediated pathway. This raises questions on the identity of the steps prior to cAMP, particularly the endogenous signal that may initiate this modulatory cascade. Data is presented suggesting the possible role of a neuropeptide, pituitary adenylyl cyclase-activating polypeptide (PACAP), in this modulation. Mutations in the amnesiac (amn) gene, which encodes a polypeptide homologous to human PACAP-38, reduced the L-type current in larval muscles. Conditional expression of a wild-type copy of the amn gene rescued the current from this reduction. Bath application of human PACAP-38 also rescued the current. PACAP-38 did not rescue the mutant current in the presence of PACAP-6-38, an antagonist at type-I PACAP receptor. 2',5'-dideoxyadenosine, an inhibitor of adenylyl cyclase, prevented PACAP-38 from rescuing the amn current. In addition, 2',5'-dideoxyadenosine reduced the wild-type current to the level seen in amn, whereas it failed to further reduce the current observed in amn muscles. H-89, an inhibitor of PKA, suppressed the effect of PACAP-38 on the current. The above data suggest that PACAP, the type-I PACAP receptors, and adenylyl cyclase play a role in the modulation of L-type Ca(2+) channels via cAMP-PKA pathway. The data also provide support for functional homology between human PACAP-38 and the amn gene product in Drosophila (Bhattacharya, 2004; full text of article).

Octopamine regulates sleep in Drosophila through protein kinase A-dependent mechanisms

Sleep is a fundamental process, but its regulation and function are still not well understood. The Drosophila model for sleep provides a powerful system to address the genetic and molecular mechanisms underlying sleep and wakefulness. This study shows that a Drosophila biogenic amine, octopamine, is a potent wake-promoting signal. Mutations in the octopamine biosynthesis pathway produced a phenotype of increased sleep, which was restored to wild-type levels by pharmacological treatment with octopamine. Moreover, electrical silencing of octopamine-producing cells decreased wakefulness, whereas excitation of these neurons promoted wakefulness. Because protein kinase A (PKA) is a putative target of octopamine signaling and is also implicated in Drosophila sleep, its role in the effects of octopamine on sleep was investigated. Decreased PKA activity in neurons rendered flies insensitive to the wake-promoting effects of octopamine. However, this effect of PKA was not exerted in the mushroom bodies, a site previously associated with PKA action on sleep. These studies identify a novel pathway that regulates sleep in Drosophila (Crocker, 2008).

By modulating the excitability of octopamine-producing cells, the output of these cells was manipulated. In mammals, one can record from specific cell populations to determine when the cells fire action potentials. Although this assay is difficult to do in flies, it was possible to electrically modulate the cells through expression of K⁺ and Na⁺ ion channels. When octopamine-producing cells were more depolarized (expression of a Na⁺ channel), the animal was awake more and unable to stay asleep, whereas when the cells were hyperpolarized (expression of a K⁺ channel), the animals slept more (Crocker, 2008).

Based primarily on larval crawling assays, octopamine and tyramine were implicated previously in locomotor behavior (see Tyramine β hydroxylase). Specifically, larvae move slower through quadrants when they have decreased octopamine levels (the TbetaH^nm18 and Tdc2^RO54 mutants). More recent work showed that adult Tdc2^RO54 flies also have a decrease in locomotor activity attributable to the lack of tyramine. The data showing differences in activity in the Tdc2^RO54 and the T?H^nm18 mutants support the claim that tyramine plays an important role in locomotion. Thus, whereas increased levels of tyramine in Tbh mutants increase activity, decreased levels in Tdc mutants decrease locomotor activity. However, both mutations increase sleep, which is most likely attributable to the loss of octopamine. In addition to overall sleep, it was found that other sleep parameters such as latency to sleep and arousal threshold are affected in flies carrying these mutations. It is inferred that tyramine plays a role in locomotion, but octopamine specifically affects arousal states (Crocker, 2008).

Studies of other invertebrate species support a role for octopamine in arousal (Corbet, 1991). In fact, octopamine agonists are potential natural pesticides because they cause insect species to 'walk off' the leaves. As in Drosophila, changes in octopamine levels affect behavior in honey bees, as demonstrated through feeding and injection of octopamine as well as through analysis of endogenous levels of octopamine. Injections of octopamine promote flying in honeybees. In addition, octopamine and tyramine regulate other behaviors in honeybees such as hive maintenance and foraging. Octopamine and tyramine also modulate sensory input in honeybees. In the locust, octopamine mediates heightened arousal in response to new visual stimuli. A specific subset of octopamine-producing neurons in the brain of the locust fires during the presentation of new visual stimuli, causing dishabituation of the descending contralateral movement detector interneuron. Interestingly, application of endogenous octopamine can mimic this state of heightened arousal. This study suggests that octopamine serves to promote arousal in Drosophila. It is possible that the increased arousal seen with too much octopamine, or decreased arousal with too little, is a result of improper gating of sensory stimuli, but without electrophysiological data it is not possible to draw any conclusions. Note also that the Tdc2 cells important for sleep and arousal in the fly brain have not been identified yet (Crocker, 2008).

Previous studies, octopamine was fed to flies to rescue or verify a phenotype of the TβH^nm18 flies. The ability of octopamine to rescue egg laying in TβH^nm18 mutants was assayed in this manner, because TβH^nm18 flies are unable to release eggs. Animals were placed on different levels of octopamine, and 10 mg/ml octopamine over a period of 6 d provided maximal rescue (Monastirioti, 1996). Using the same concentration, this study found that a steady increase in octopamine levels led to a decrease in nighttime sleep. Based on the specific effect on nighttime sleep, it is speculated that octopamine levels are already high during the daytime, thereby precluding any effects of an increase. This analysis is supported by the Na⁺ channel data in which a significant decrease in total sleep was found only during the nighttime sleep periods. It is speculated that, normally, activity of these cells is low at night, and so expression of the Na⁺ channel causes them to fire more and release octopamine at an abnormal time, thereby producing a decrease in sleep. Similar results, indicating nighttime-specific effects, were obtained with overexpression of Tdc2. Work in other insects also supports the idea of modulated octopamine release. Pribbenow (1996) demonstrated that honeybees who are already in a heightened arousal state of antennae scanning do not change scanning frequency in response to octopamine administration, but, in animals scanning at a low frequency, injections of octopamine significantly increase scanning (Crocker, 2008).

Thse data suggest that the effects of octopamine are mediated through PKA-dependent signaling. In mammals, there are nine different adrenergic receptors, some of which signal through PKA. The α1 adrenergic receptor is the only receptor associated with a wake-promoting effect in that the agonist methoxamine causes an increase in waking . However, the antagonist has no effect on total sleep. It is important to note that the α1 receptor in mammals is thought to be coupled to phospholipase C and G_q. The β adrenergic receptors (which are coupled to cAMP and PKA) probably do not have specific effects on sleep in mammals because, contrary to known effects of norepinephrine, the agonist increases sleep and the antagonist decreases sleep. Studies in Drosophila may be better able to identify biogenic amine receptors relevant for sleep because of the ease of genetic manipulation. Many G-protein-coupled receptors in Drosophila display activity that allows their bona fide classification as octopamine receptors. The current data suggest that receptors sensitive to mianserin are likely to be involved in regulating fly sleep. Because mianserin inhibits cAMP signaling, these data not only further support a role for PKA but also implicate β receptors in octopamine action. It is noted that none of these receptors is known to display a circadian cycling profile (Crocker, 2008).

Given that PKA has been shown to regulate sleep in Drosophila, a link between the various molecules that affect Drosophila sleep is starting to be apparent. Interestingly, however, octopamine does not appear to act through the mushroom bodies, a structure known to mediate effects of PKA on sleep and also to express a class of octopamine receptors. Because flies lacking mushroob bodies still have substantial amounts of sleep, it is clear that other parts of the fly brain can drive sleep. The current study shows that even PKA can affect sleep in regions outside the mushroom body. Defining the site of action of sleep-regulating molecules such as octopamine should help to identify these other brain regions (Crocker, 2008).

Perturbing dynamin reveals potent effects on the Drosophila circadian clock

Transcriptional feedback loops are central to circadian clock function. However, the role of neural activity and membrane events in molecular rhythms in the fruit fly Drosophila is unclear. To address this question, a temperature-sensitive, dominant negative allele was expressed of the fly homolog of dynamin called shibire^ts1 (shi^ts1), an active component in membrane vesicle scission. Broad expression in clock cells resulted in unexpectedly long, robust periods (>28 hours) comparable to perturbation of core clock components, suggesting an unappreciated role of membrane dynamics in setting period. Expression in the pacemaker lateral ventral neurons (LNv) was necessary and sufficient for this effect. Manipulation of other endocytic components exacerbated shi^ts1's behavioral effects, suggesting its mechanism is specific to endocytic regulation. PKA overexpression rescued period effects suggesting shi^ts1 may downregulate PKA pathways. Levels of the clock component Period were reduced in the shi^ts1-expressing pacemaker small LNv of flies held at a fully restrictive temperature (29°C). Less restrictive conditions (25 degrees C) delayed cycling proportional to observed behavioral changes. Levels of the neuropeptide Pigment-dispersing factor (PDF), the only known LNv neurotransmitter, were also reduced, but Period cycling was still delayed in flies lacking PDF, implicating a PDF-independent process. Further, shi^ts1 expression in the eye also results in reduced Per protein and per and vri transcript levels, suggesting that shibire-dependent signaling extends to peripheral clocks. The level of nuclear Clk, transcriptional activator of many core clock genes, is also reduced in shi^ts1 flies, and Clk overexpression suppresses the period-altering effects of shi^ts1. It is proposed that membrane protein turnover through endocytic regulation of PKA pathways modulates the core clock by altering Clk levels and/or activity. These results suggest an important role for membrane scission in setting circadian period (Kilman, 2009).

Daily rhythms of sleep and wake are driven by transcriptional feedback loops in pacemaker neurons. In Drosophila, the transcription factor Clock (Clk) heterodimerizes with cycle (cyc) to directly activate key components of a principal feedback loop, period (per) and timeless (tim), and of a secondary feedback loop, par domain protein 1 (pdp-1) and vrille (vri). Per and perhaps Tim feed back and repress Clk/Cyc DNA binding leading to molecular oscillations in clock components. Vri feeds back to repress transcription of Clk, while Ppd may regulate clock output. Clk also activates clockwork orange (cwo), which represses Clk-activated transcription of its targets. These molecular feedback loops are thought to operate in a cell-autonomous manner. Several components of these feedback loops are conserved with mammals (Kilman, 2009).

Molecular clocks are evident in many peripheral tissues, such as the eye, as well as the central brain. Brain clocks are divided into 7 anatomical clusters: small and large ventral lateral neurons (sLNv, lLNv), dorsal lateral neurons (LNd), three groups of dorsal neurons (DN1, DN2, DN3), and the lateral posterior neurons (LPN). The neuropeptide Pigment Dispersing Factor (PDF) is expressed uniquely by and is the only known transmitter of the LNv. Mutants of PDF or its receptor display short period damping rhythms. pdf⁰¹ pacemaker molecular oscillations are eventually low amplitude or phase-dispersed, indicating PDF feeds back to maintain synchrony. Mammalian rhythms are also lost in mutants of the Vasoactive Intestinal Peptide (VIP) system, indicating a conserved role for neuropeptidergic signaling in clocks. Under light-dark conditions (LD), the PDF+ sLNv mediate behavioral anticipation of the transition from dark to light ('morning') while 'evening' anticipation is mediated by PDF- clocks: the DN1, LNd, and one sLNv. Under constant darkness (DD), the LNv dominate behavioral period determination and reset non-PDF clocks. PDF neurons may also receive a number of other neurotransmitter inputs. In addition, electrical silencing of PDF neurons suppresses core clock function. A number of intracellular signaling pathways have been identified as contributing to core circadian function. However, the mechanisms of feedback between receptor and/or ion channel signaling and transcriptional feedback rhythms remain unclear (Kilman, 2009).

To explore the role of the network in circadian function, vesicle traffic was perturbed as a way of disrupting intercellular communication. shibire (shi), the Drosophila homolog of dynamin, is a GTPase necessary for vesicle scission. The dominant negative shi^ts1 allele has been used at the restrictive temperature (29^oC) to inhibit synaptic transmission. However shi is also involved in other endocytic pathways that may affect intercellular signaling including receptor-mediated endocytosis and recycling of membrane proteins, such as ion channels. This study shows shi^ts1 expression in clock cells at 25^oC results in robust long behavioral rhythms. Period effects are exacerbated by perturbing endocytic/endosomal pathways and suppressed by overexpressing arrestin2 or a catalytic subunit of Protein Kinase A (PKA-C1). Long period results from PDF-independent delays in the molecular clock of the sLNv. With further impairment at 29^oC, shi^ts1 expression in either the LNv or in peripheral eye clocks also drastically reduces Clk target gene levels. Clk itself is reduced in the sLNv and the long period is suppressed by Clk overexpression. These results suggest that modulation of cell membrane processes such as receptor signaling pathways may powerfully affect the molecular clock (Kilman, 2009).

These data suggest an important function for membrane events, specifically endocytosis, in circadian timing. While previous studies have demonstrated roles for neural activity in circadian output, in sustaining molecular rhythms, and in synchrony, this work strongly suggests a substantial role in circadian timing. Expression of shi^ts1 in pacemaker neurons results in strikingly long periods, suggesting potent effects on circadian timing through perturbing vesicle scission. These effects are enhanced by co-expression of other components of endocytic pathways leading to early endosomes, consistent with shi function in traffic, recycling, and turnover of cell membrane components. PKA expression rescues period defects induced by shi^ts1, suggesting a functional link between the membrane, PKA, and behavioral period. The LNv-expressed shi^ts1 results in delays in the phase of Per molecular rhythms in the sLNv sufficient to account for the delay in behavior. While shi^ts1 effects on behavior require Pdf, those on the molecular clock of the sLNv are Pdf-independent, implicating a novel pathway. In fact these perturbations of the molecular clock are not specific to locomotor pacemakers, but appear in peripheral clocks as well, suggesting membrane-clock interactions are a general property of clock cells. Reductions in the levels of Clk and Clk-activated transcripts are consistent with the hypothesis that membrane events regulate the molecular clock through Clk (Kilman, 2009).

Several lines of evidence indicate that shi^ts1 effects are not operating principally by blocking pacemaker neural output. Expression of tetanus toxin in PDF neurons blocks responses to arousing effects of cocaine, indicating that PDF neurons use a classical neurotransmitter and that tetanus toxin is expressed at functional levels capable of blocking this process. Yet tetanus toxin expression in PDF+ cells does not significantly alter period or rhythmicity. In shi^ts1 expressing flies, delayed PDF neuronal clocks still delay the offset of evening behavior, implying PDF cells can still reset evening clocks. In addition, no desynchronization was observed of molecular rhythms among the sLNv as might be expected if communication were disrupted. Period altered shi^ts1-expressing flies also largely preserve rhythmicity at 25^oC suggesting a primary clock effect rather than an output effect. Likewise PDF, the only known sLNv output, is also not necessary for shi^ts1 molecular effects. In pdf⁰¹ mutants, shi^ts1 expression blocks the effects on behavioral period but does not block the effect of shi^ts1 on PER LNv rhythms. The uncoupling of sLNv molecular rhythms from behavioral rhythms clearly demonstrates an output function for PDF in pacemaker neuron function. This also implies that other neural clusters drive behavior in pdf⁰¹. Moreover, these results demonstrate that shi^ts1's effects on sLNv PER do not operate through PDF. Taken together these data suggest the period differences that were seen do not result primarily from alterations of sLNv transmitter output. Instead it seems likely shi^ts perturbs another target or pathway regulating sLNv activity (Kilman, 2009).

While effects of shi^ts1 are typically tested at 29°C or above, shi^ts1 effects noted in this study have been observed at just 25°C, below the reported paralytic temperature for shi^ts1 (Masur, 1991). However, ultrastructural shi^ts1 effects have been observed even at the nominal permissive temperature (18°C) for behavioral paralysis (Masur, 1991). Thus, shi^ts1 is likely modestly defective at 18°C and this impairment grows with increasing temperature until a threshold is reached at which paralysis is evident when driven in motorneurons. However under conditions of overexpression, the temperature threshold for various phenotypes may differ from paralysis. The finding of slight period lengthening relative to controls even at 18°C is consistent with a modest defect, with core clock effects getting stronger gradually as the temperature increases. The evidence that shi^ts1 is not perturbing outputs (at least at 25°C) raises that possibility that other membrane scission-sensitive processes, such as receptor endocytosis, may have a lower threshold for disruption than synaptic transmission (Kilman, 2009).

What might be the nature of the membrane perturbation evoked by shi^ts1? More broadly, endocytosis regulates membrane protein turnover, and a variety of targets could influence neuronal activity, including ion channels, pumps, and transporters, which in turn could feedback to regulate the core clock. Endocytosis has a well-established role in down-regulation of metabotropic or ionotropic receptors. In the sLNv, potential receptors include (but are not limited to) acetylcholine, GABA, serotonin, dopamine, histamine, and neuropeptides such as IPNamide. Ion channel density may also be modulated by endocytosis and could influence core clock rhythms. In contrast, the finding that PKA overexpression can suppress shi^ts1 effects on period provide evidence that down regulation of G-protein coupled receptors that stimulate cAMP and PKA may be a mechanism for shi action. The identification and functional analysis of the relevant membrane targets of shi^ts1 will be critical to understanding the role of the membrane in circadian function (Kilman, 2009).

Targets of Activity

Reduced PKA activity in anterior imaginal disc cells leads to cell-autonomous induction of decapentaplegic, wingless, and patched transcription, independent of hedgehog gene activity. Expression of a mutant regulatory subunit to anterior cells at the AP border of the wing imaginal disc results in localized PKA inactivation and can substitute for hh in promoting disc growth. PKA inhibition in anterior cells of the AP border can induce patched expression and can thus substitute for the growth-promoting activity of hh during larval life (Li, 1995).

Cyclic AMP (cAMP)-dependent protein kinase A is essential during limb development to prevent inappropriate decapentaplegic and wingless expression. A constitutively active form of PKA can prevent inappropriate dpp and wg expression, but does not interfere with their normal induction by hh. It seems that the basal activity of PKA imposes a block on the transcription of dpp and wg, and that hh exerts its organizing influence by alleviating this block (Jiang, 1995).

In the anterior compartment of appendage discs and anterior to the morphogenetic furrow in the eye disc, cells that lack cAMP-dependent protein kinase activity ectopically express decapentaplegic. Pka-C1- cells can influence the fate of neighboring cells to reorganize anterior patterns in appendages and trigger ectopic morphogenetic furrows in the developing retina. This organizing activity of Pka mutant cells is dependent on dpp activity. These findings suggest that PKA is a component of a signaling pathway that represses dpp expression and that HH antagonizes this pathway to maintain dpp expression at the anterior-posterior compartment border and in the morphogenetic furrow (Pan, 1995).

Removing activity of the gene encoding Pka-C1 is functionally equivalent to removing ptc activity or to providing cells with the HH signal. These findings suggest that cyclic AMP-dependent protein kinase A is a component of the signal transduction pathway through which HH and PTC direct localized expression of dpp (or wg) and establish the compartment boundary organizer (Lepage, 1995).

dpp is a target of the hh signal acting through Fused. fu mutations rescue the phenotype due to ectopic expression of hh or to the lack of patched activity. fu is also required for the activation of engrailed caused when hh is ectopically activated in the wing disk. Although fu, cos-2 and cubitus interruptus probably form part of the same pathway that controls dpp expression, Protein kinase A probably controls dpp expression by a different pathway (Sánchez-Herrero, 1996).

Phosphorylation of Fused occurs in response to Hedgehog and cannot be blocked by activation of Protein kinase A, which is thought to be an antagonist of signaling from hedgehog. This suggests that Fused and Protein kinase A function downstream of Hedgehog but in parallel pathways that eventually converge downstream of Fused (Thérond, 1996).

Two putative light-sensitive ion channels have been isolated from Drosophila, encoded by the transient-receptor-potential (trp) and transient-receptor-potential-like (trpl) genes. Two calmodulin-binding sites are present in the C-terminal domain of the Trpl protein: CBS-1 and CBS-2. CBS-1 binds calmodulin in a Ca2+-dependent fashion. In contrast, CBS-2 binds the Ca2+-free form of calmodulin, with dissociation occurring at Ca2+ concentrations between 5 and 25 microM. Phosphorylation of a serine residue within a peptide encompassing CBS-1 by cyclic AMP-dependent protein kinase (PKA) abolishes calmodulin binding, and phosphorylation of the adjacent serine by protein kinase C appears to modulate this phosphorylation by PKA. Interpretation of these findings provides a novel model for ion-channel gating and modulation in response to changing levels of intracellular Ca2+ (Warr, 1996).

Phototransduction in Drosophila is mediated by a G-protein-coupled phospholipase C transduction cascade in which each absorbed photon generates a discrete electrical event, the quantum bump. In whole-cell voltage-clamp recordings, cAMP, as well as its nonhydrolyzable and membrane-permeant analogs 8-bromo-cAMP (8-Br-cAMP) and dibutyryl-cAMP, slow down the macroscopic light response by increasing quantum bump latency, without changes in bump amplitude or duration. In contrast, cGMP or 8-Br-cGMP has no effect on light response amplitude or kinetics. None of the cyclic nucleotides activate any channels in the plasma membrane. The effects of cAMP are mimicked by application of the non-specific phosphodiesterase inhibitor IBMX and the adenylyl cyclase activator forskolin; zaprinast, a specific cGMP-phosphodiesterase inhibitor, is ineffective. Bump latency is also increased by targeted expression of either an activated G_salpha subunit, which increases endogenous adenylyl cyclase activity, or an activated catalytic protein kinase A (PKA) subunit. The action of IBMX is blocked by pretreatment with the PKA inhibitor H-89. The effects of cAMP are abolished in mutants of the ninaC gene, suggesting this nonconventional myosin as a possible target for PKA-mediated phosphorylation. Dopamine (10 µM) and octopamine (100 µM) mimic the effects of cAMP. These results indicate the existence of a G-protein-coupled adenylyl cyclase pathway in Drosophila photoreceptors that modulates the phospholipase C-based phototransduction cascade (Chyb, 1999).

Altered gene regulation and synaptic morphology in Drosophila learning and memory mutants

Genetic studies in Drosophila have revealed two separable long-term memory pathways defined as anesthesia-resistant memory (ARM) and long-lasting long-term memory (LLTM). ARM is disrupted in

radish (rsh) mutants, whereas LLTM requires CREB-dependent protein synthesis. Although the downstream effectors of ARM and LLTM are distinct, pathways leading to these forms of memory may share the cAMP cascade critical for associative learning. Dunce, which encodes a cAMP-specific phosphodiesterase, and rutabaga, which encodes an adenylyl cyclase, both disrupt short-term memory. Amnesiac encodes a pituitary adenylyl cyclase-activating peptide homolog and is required for middle-term memory. This study demonstrates that the Radish protein localizes to the cytoplasm and nucleus and is a PKA phosphorylation target in vitro. To characterize how these plasticity pathways may manifest at the synaptic level, synaptic connectivity was assayed and an expression analysis was performed to detect altered transcriptional networks in rutabaga, dunce, amnesiac, and radish mutants. All four mutants disrupt specific aspects of synaptic connectivity at larval neuromuscular junctions (NMJs). Genome-wide DNA microarray analysis revealed approximately 375 transcripts that are altered in these mutants, suggesting defects in multiple neuronal signaling pathways. In particular, the transcriptional target Lapsyn, which encodes a leucine-rich repeat cell adhesion protein, localizes to synapses and regulates synaptic growth. This analysis provides insights into the Radish-dependent ARM pathway and novel transcriptional targets that may contribute to memory processing in Drosophila (Guan, 2011).

Drosophila has proven to be a powerful model for identifying gene products involved in learning and memory based on olfactory, visual, and courtship behavioral assays. How proteins identified in these studies regulate neuronal function or physiology to specifically alter behavioral plasticity is an ongoing area of investigation. Using the well-characterized 3rd instar larval NMJ as a model glutamatergic synapse, the effects on synaptic connectivity were compared of several learning mutants that alter cAMP signaling (dnc¹, rut¹, amn¹) with the poorly characterized ARM mutant rsh¹. Each mutant altered synaptic connectivity at NMJs in a specific manner, suggesting that changes in neuronal connectivity in the CNS might contribute to the behavioral defects found in these strains. The observations in dnc¹ and rut¹ are similar to previous studies of synaptic morphology in these mutants. Gene expression was assayed in the mutants using microarray analysis, which revealed many neuronal transcripts that were transcriptionally altered. A long-term goal is to link transcriptional changes in specific loci to the behavioral and morphological defects found in learning and memory mutants (Guan, 2011).

Experimental approaches to define the biochemical transition from short-term plasticity to long-term memory storage have suggested a key role for cAMP signaling. At the molecular level, one of the best-characterized pathways for STM has been described for gill withdrawal reflex facilitation in Aplysia. In this system, conditioned stimuli act through a serotonergic G protein-coupled receptor pathway to activate adenylyl cyclase in the presynaptic sensory neuron, resulting in the synthesis of cAMP. cAMP activates PKA, which phosphorylates a presynaptic potassium channe, leading to prolonged calcium influx and enhanced neurotransmitter release from the sensory neuron. Insights into the LLTM pathway in Aplysia have implicated CREB function. Robust training or stimulation with serotonin induces translocation of the catalytic subunit of PKA into the nucleus, where it activates the transcription factor CREB-1 and inhibits the transcriptional suppressor CREB-2. CREB-1 acts on additional transcription factors to produce specific mRNAs that are transported to dendrites and captured by activated synapses. Local synthesis of new proteins and subsequent growth of synaptic connections is predicted to underlie long-term memory in the system. It is likely that similar molecular pathways exist in other species. Transgenic Drosophila with inducible inhibition of PKA show memory impairment. PKA is also activated during hippocampal LTP induction in mammals, and transgenic mice that express an inhibitor of PKA have defective LTP and hippocampal-dependent memory, suggesting a general role for cAMP/PKA in the transition from learning to memory storage (Guan, 2011).

In addition to CREB-dependent LLTM, which requires transcription and translation for its formation, the Radish-dependent ARM pathway represents a distinct long-term memory storage mechanism. These various memory pathways partially overlap in time. Three hours after training ~50% of memory is stored as STM, with the rest present as ARM, which is formed immediately after training in flies and can last for days depending on training intensity. ARM is not blocked by agents that disrupt electrical activity in the brain, suggesting that a biochemical pathway for ARM is likely initiated by learning stimuli, but does not require continued neuronal excitation for its expression. ARM is also not as sensitive to translation inhibition, as a 50% reduction of protein synthesis by cycloheximide does not affect ARM, but blocks LLTM (Guan, 2011).

Similar to the role of CREB in LLTM, Radish appears to be a key regulator of the ARM phase of memory. In contrast to the molecular pathways underlying STM (cAMP/PKA cascade) and LLTM (PKA/CREB), the signaling mechanisms mediating ARM are unknown. Unfortunately, the amino acid sequence of the radish locus gives little insight into its function, as it lacks known structural motifs or domains. Radish contains a serine/arginine-rich sequence with very limited homology to splicing factors, hinting that it may be involved in RNA processing. The Radish protein also contains PKA phosphorylation sites and multiple NLS sites within its sequence. Consistent with these sequence features, This study found that Radish is phosphorylated by PKA in vitro, linking ARM to the cAMP/PKA pathway. By generating a GFP-tagged Radish transgenic animal, it was possible to characterize Radish localization. Radish was prominently localized to cell bodies of neurons in the CNS, but was enriched in the nucleus in other cell types such as salivary gland and muscle cells. Given the overlap between several of the NLS and PKA sites in Radish, it will be interesting to explore whether the phosphorylation state of Radish regulates its subcellular distribution. An attractive hypothesis is that activated PKA phosphorylates Radish at synapses, resulting in transport to the nucleus with accompanying effects on transcription or RNA processing that would modify long-term synaptic function. Given that ARM can last for days, a change in nuclear function is an attractive biological underpinning, even though ARM has been suggested to be a translation-independent form of memory. Given that general protein synthesis was reduced by only 50% in the previous studies, it is quite possible that ARM and LLTM have different thresholds for translational inhibition (Guan, 2011).

In terms of synaptic modifications in rsh¹ mutants, this study found that larval NMJ synapses were altered compared with controls. Specifically, rsh¹ mutants had shorter axonal projections onto target muscles and displayed more synaptic boutons within the innervated region. These alterations gave rise to a more compact innervation pattern than observed in controls. Overgrowth of synapses at larval NMJs was also observed in dnc¹ mutants, whereas reduced innervation length was found in rut¹ mutants. As such, rsh¹ mutant NMJs display a unique phenotype compared with mutants that increase or decrease cAMP levels. The molecular mechanisms by which Radish regulates synaptic growth are unclear. Radish could directly interface with growth regulators at the synapse in a PKA-dependent fashion. Indeed, an interaction between Radish and Rac1 was found in a high-throughput yeast two-hybrid screen for interacting Drosophila proteins. Rac1 is a Rho family GTPase that regulates neuronal and synaptic morphology via reorganization of the cytoskeleton. Rac1 function has also been linked to PAK1 and the Fragile-X Mental Retardation protein (FMRP), which alter synaptic and behavioral plasticity in mammals. Recently, Rac activity has been linked to memory decay in Drosophila (Shuai. 2010), indicating that a Radish-Rac link might control memory processing via alterations in cytoskeletal modulation of synaptic function or stability. Although it is possible that Radish regulates synaptic properties through a Rac1 interaction, no robust Rac1-Radish interaction was observed in either yeast-two hybrid or GST pull-down experiments. No Radish-GFP enrichment was observed at larval synapses where the synaptic growth defect was quantified, although the protein was present in larval axons. As such, it may be that NMJ defects in rsh¹ arise through downstream effects secondary to the loss of Radish function in a neuronal compartment besides the synapse (Guan, 2011).

To further explore this possibility and examine links between rsh and the STM pathway, genome-wide microarray studies were performed on several learning and memory mutants. Although there were some shared transcriptional changes between rsh¹ and the other mutants (dnc¹, rut¹, amn¹), most of the changes in rsh¹ were unique. Although linking these changes to a direct effect on the underlying biology will require more work, several interesting loci were identified that could contribute to synaptic plasticity defects. The Drosophila NFAT homolog, a transcription factor that binds to the activity-regulated AP-1 (Fos/Jun) dimer, was robustly up-regulated by sevenfold in rsh¹ mutants. The RNA-binding protein smooth (sm) was also up-regulated in rsh¹ mutants. Mutations in sm have been shown to alter axonal pathfinding. Other genes that were transcriptionally altered in rsh¹ mutants and that would be predicted to influence synaptic connectivity were the Sh potassium channel, the adapter protein Disabled, and the Lapsyn cell adhesion protein. The potential role of Lapsyn was intriguing, as LRR-containing proteins have been implicated in the regulation of neurite outgrowth and synapse formation. In particular, netrin-G ligand and synaptic-like adhesion molecule (SALM) are known LRR proteins that regulate neuronal connectivity and synapse formation. In Drosophila, LRR repeat proteins have been implicated in motor neuron target selection. Given the roles of other LRR-containing proteins in the regulation of neuronal connectivity, this study explored whether Lapsyn might also function in this pathway. Lapsyn was up-regulated by neuronal activity in addition to being up-regulated in rsh¹, making it an interesting transcriptional target to assay for a role in synaptic modification (Guan, 2011).

Lapsyn mRNA expression was broadly up-regulated in the brain by neuronal activity, suggesting a potential widespread effect on neuronal function. Lapsyn-GFP transgenic protein targeted to the presynaptic terminal, partially overlapping with the periactive zone, a region of the nerve terminal enriched in proteins that regulate synaptic vesicle endocytosis and synaptic connectivity. Animals lacking Lapsyn died at the end of embryogenesis, although the early stages of nervous system formation appeared normal. It was possible to partially rescue Lapsyn mutants with neuronal expression of a Lapsyn transgene, indicating an essential function for the protein in the nervous system. Rescue to adulthood required expression outside the nervous system, suggesting Lapysn is likely to have functions in other tissue types as well. Manipulations of Lapsyn expression in the nervous system resulted in distinct defects in synaptic connectivity at the NMJ. Heterozygotes expressing only a single copy of the Lapsyn gene displayed supernumerary satellite bouton formation, a phenotype commonly associated with mutants that disrupt synaptic endocytosis or that alter the transmission or trafficking of synaptic growth factors through the endosomal system. This increase in satellite boutons in Lapsyn heterozygotes suggests that the protein plays a role in the regulation of synaptic growth signaling. Overexpression of Lapsyn, as induced by activity or observed in rsh¹ mutants, also elicited a change in synaptic growth, resulting in an increase in overall bouton number at larval NMJs. Thus, regulation of Lapsyn levels modulate synaptic growth mechanisms at NMJs. Lapsyn mutant heterozygotes also display defects in larval associative learning, although this phenotype could not be rescued with pan-neuronal overexpression. The lack of a specific rescue makes it unclear whether the learning defects are linked to a non-Lapsyn function, or if a more specific spatial and temporal expression of Lapsyn is required for functional rescue (Guan, 2011).

How Lapsyn participates in synaptic signaling is currently unclear. The closest mammalian homologs of Lapsyn are the NGL family of synaptic adhesion molecules. Three isoforms are found in mammals, NGL-1, NGL-2, and NGL-3, which interact with netrin-G1, netrin-G2, and the receptor tyrosine phosphatase LAR, respectively. NGL-1 promotes axonal outgrowth, whereas NGL-2 is capable of triggering synapse formation. The interaction of NGL-3 with LAR is intriguing, as the Drosophila LAR homolog has been shown to bind the heparan sulfate proteoglycans Syndecan and Dallylike to regulate synaptic growth at the NMJ. The homology between Lapsyn and the mammalian NLG family is restricted to the extracellular LRR domain, with no homology observed in the intracellular C terminus. The three mammalian NLGs also lack homology to each other at the C terminus, except for the presence of a PDZ-binding domain at the end of the intracellular domain. It will be important to identify binding partners for Lapsyn at the synapse to define how it may regulate synaptic adhesion or signaling between the pre- and postsynaptic compartments to regulate synaptic growth. Likewise, additional studies into the Radish-dependent ARM phase of memory may reveal how rsh-dependent changes in Lapsyn levels contribute to the synaptic and behavioral defects of this memory mutant (Guan, 2011).

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.