Flexible goal-driven orientation requires that the position of a target be stored, especially in case the target moves out of sight. The capability to retain, recall and integrate such positional information into guiding behaviour has been summarized under the term spatial working memory. This kind of memory contains specific details of the presence that are not necessarily part of a long-term memory. Neurophysiological studies in primates indicate that sustained activity of neurons encodes the sensory information even though the object is no longer present. Furthermore they suggest that dopamine transmits the respective input to the prefrontal cortex, and simultaneous suppression by GABA spatially restricts this neuronal activity. Here we show that Drosophila melanogaster possesses a similar spatial memory during locomotion. Using a new detour setup, we show that flies can remember the position of an object for several seconds after it has been removed from their environment. In this setup, flies are temporarily lured away from the direction towards their hidden target, yet they are thereafter able to aim for their former target. Furthermore, we find that the GABAergic (stainable with antibodies against GABA) ring neurons of the ellipsoid body in the central brain are necessary and their plasticity is sufficient for a functional spatial orientation memory in flies. We also find that the protein kinase S6KII (ignorant) is required in a distinct subset of ring neurons to display this memory. Conditional expression of S6KII in these neurons only in adults can restore the loss of the orientation memory of the ignorant mutant. The S6KII signalling pathway therefore seems to be acutely required in the ring neurons for spatial orientation memory in flies.

The accumulation of amyloid-β (Aβ) into plaques is a hallmark feature of Alzheimer's disease (AD). While amyloid precursor protein (APP)-related proteins are found in most organisms, only Aβ fragments from human APP have been shown to induce amyloid deposits and progressive neurodegeneration. Therefore, it was suggested that neurotoxic effects are a specific property of human Aβ. Here we show that Aβ fragments derived from the Drosophila orthologue APPL aggregate into intracellular fibrils, amyloid deposits, and cause age-dependent behavioral deficits and neurodegeneration. We also show that APPL can be cleaved by a novel fly β-secretase-like enzyme. This suggests that Aβ-induced neurotoxicity is a conserved function of APP proteins whereby the lack of conservation in the primary sequence indicates that secondary structural aspects determine their pathogenesis. In addition, we found that the behavioral phenotypes precede extracellular amyloid deposit formation, supporting results that intracellular Aβ plays a key role in AD.

Drosophila melanogaster flies cross surmountable gaps in their walkway of widths exceeding their body length with an astounding maneuver but avoid attempts at insurmountable gaps by visual width estimation. Different mutant lines affect specific aspects of this maneuver, indicating a high complexity and modularity of the underlying motor control. Here we report on two mutants, ocelliless(1) and tay bridge(1), that, although making a correct decision to climb, fail dramatically in aiming at the right direction. Both mutants show structural defects in the protocerebral bridge, a central complex neuropil formed like a handlebar spanning the brain hemispheres. The bridge has been implicated in step-length control in walking flies and celestial E-vector orientation in locusts. In rescue experiments using tay bridge(1) flies, the integrity of the bridge was reestablished, concomitantly leading to a significant improvement of their orientation at the gap. Although producing directional scatter, their attempts were clearly aimed at the landing site. However, this partial rescue was lost in these flies at a reduced-visibility landing site. We therefore conclude that the protocerebral bridge is an essential part of a visual targeting network that transmits directional clues to the motor output via a known projection system.

Several aspects of locomotor control have been ascribed to the central complex of the insect brain; however, the role of distinct substructures of this complex is not well known. The tay bridge1 (tay1) mutant of Drosophila melanogaster was originally isolated on the basis of reduced walking speed and activity. In addition, tay1 is defective in the compensation of rotatory stimuli during walking and histologically, tay1 causes a mid-sagittal constriction of the protocerebral bridge, a constituent of the central complex. Cloning of the tay gene revealed that it encodes a novel protein with no significant homology to any known protein. To associate the behavioral phenotypes with the anatomical defect in the protocerebral bridge, we used different driver lines to express the tay cDNA in various neuronal subpopulations of the central brain in tay1-mutant flies. These experiments showed an association of the aberrant walking speed and activity with the structural defect in the protocerebral bridge. In contrast, the compensation of rotatory stimuli during walking was rescued without a restoration of the protocerebral bridge. The results of our differential rescue approach are supported by neuronal silencing experiments using conditional tetanus toxin expression in the same subset of neurons. These findings show for the first time that the walking speed and activity is controlled by different substructures of the central brain than the compensatory locomotion for rotatory stimuli.

Cellular ultrastructures for signal integration are unknown in any nervous system. The ellipsoid body (EB) of the Drosophila brain is thought to control locomotion upon integration of various modalities of sensory signals with the animal internal status. However, the expected excitatory and inhibitory input convergence that virtually all brain centres exhibit is not yet described in the EB. Based on the EB expression domains of genetic constructs from the choline acetyl transferase (Cha), glutamic acid decarboxylase (GAD) and tyrosine hydroxylase (TH) genes, we identified a new set of neurons with the characteristic ring-shaped morphology (R neurons) which are presumably cholinergic, in addition to the existing GABA-expressing neurons. The R1 morphological subtype is represented in the Cha- and TH-expressing classes. In addition, using transmission electron microscopy, we identified a novel type of synapse in the EB, which exhibits the precise array of two independent active zones over the same postsynaptic dendritic domain, that we named 'agora'. This array is compatible with a coincidence detector role, and represents ~8% of all EB synapses in Drosophila. Presumably excitatory R neurons contribute to coincident synapses. Functional silencing of EB neurons by driving genetically tetanus toxin expression either reduces walking speed or alters movement orientation depending on the targeted R neuron subset, thus revealing functional specialisations in the EB for locomotion control.

SUMMARYWe tested whether Drosophila larvae can associate odours with a mechanosensory disturbance as a punishment, using substrate vibration conveyed by a loudspeaker (buzz: ). One odour (A) was presented with the buzz, while another odour (B) was presented without the buzz (A/B training). Then, animals were offered the choice between A and B. After reciprocal training (A/B), a second experimental group was tested in the same way. We found that larvae show conditioned escape from the previously punished odour. We further report an increase of associative performance scores with the number of punishments, and an increase according to the number of training cycles. Within the range tested (between 50 and 200Hz), however, the pitch of the buzz does not apparently impact associative success. Last, but not least, we characterized odour-buzz memories with regard to the conditions under which they are behaviourally expressed -or not. In accordance with what has previously been found for associative learning between odours and bad taste (such as high concentration salt or quinine), we report that conditioned escape after odour-buzz learning is disabled if escape is not warranted, i.e. if no punishment to escape from is present during testing. Together with the already established paradigms for the association of odour and bad taste, the present assay offers the prospect of analysing how a relatively simple brain orchestrates memory and behaviour with regard to different kinds of ʻbadʼ events. Supplementary material available online at

β-secretase (or BACE1) is the key enzyme in the production of β-amyloid (Aβ), which accumulates in the senile plaques characteristic for Alzheimer's Disease (AD). Consequently, the lack of BACE1 prevents β-processing of the Amyloid Precursor Protein (APP) and Aβ production, which made it a promising target for drug development. However, the loss of BACE1 is also detrimental, leading to myelination defects and altered neuronal activity, functions that have been associated with the cleavage of Neuregulin and a voltage-gated sodium channel subunit. Here we show that the Drosophila orthologue of BACE, dBACE, is required for glial survival. Cell-specific knockdown experiments reveal that this is a non-cell autonomous function, as a knockdown of dBACE in photoreceptor neurons leads to progressive degeneration of glia in their target zone, the lamina. Interestingly, this phenotype is suppressed by the loss of the fly Amyloid Precursor Protein (APPL), whereas a secretion-deficient form of APPL enhances the degeneration. This shows that full-length APPL in neurons promotes the death of neighboring glial cells and that β-processing of APPL is needed to prevent glial death. These results therefore not only demonstrate a novel function for an APP protein in glia, but they also show this function specifically requires regulation by β-cleavage.

Summary Odorants of behaviorally relevant objects (e.g., food sources) intermingle with those from other sources. Therefore to determine whether an odor source is good or bad—without actually visiting it—animals first need to segregate the odorants from different sources. To do so, animals could use temporal stimulus cues, because odorants from one source exhibit correlated fluctuations, whereas odorants from different sources are less correlated. However, the behaviorally relevant timescales of temporal stimulus cues for odor source segregation remain unclear. Using behavioral experiments with free-flying flies, we show that (1) odorant onset asynchrony increases flies' attraction to a mixture of two odorants with opposing innate or learned valence and (2) attraction does not increase when the attractive odorant arrives first. These data suggest that flies can use stimulus onset asynchrony for odor source segregation and imply temporally precise neural mechanisms for encoding odors and for segregating them into distinct objects.