Fig. 1. Scheme for the formation and the action of the unconventional messenger nitric oxide (NO) in the nervous system. Elevation of cytosolic calcium by voltage-dependent Ca²⁺ channels (VDCC), receptor channels or by release of calcium from intracellular stores activates the nitric oxide synthase (NOS). The active NOS converts -arginine into NO and citrulline using NADPH as co-factor. Due to its properties, the formed NO can cross cell membranes and diffuses from its site of production to act on targets in neighbouring cells. The highly reactive NO is not restricted to synapses and does not bind to specific receptors. It can exert its biological function via a wide range of reactions, limited by half-life and diffusion barriers.

Fig. 2. Scheme of NOS localization in the brain of various insect species. This sketch summarizes the general properties of NOS distribution within the brain of various insects, which are shown in Fig. 3, Fig. 4, Fig. 5, Fig. 6 and Fig. 7 in detail. Neuropiles that exhibit high NOS activity in all tested species are darkly shaded, while neuropiles which exhibit different NOS activity in distinct species are lightly shaded. In all insects investigated yet, the primary site of olfactory information processing, the antennal lobes (al), exhibit high levels of NOS. In contrast, the visual neuropiles (la, lamina; me, medulla; lo, lobula), show very distinct levels of NOS in different insect species. The central complex with its fan-shaped central body (cb) is labelled in all tested species. Distinct neuropiles of the mushroom bodies (mb) show different labelling. While the input area, the calyx (cx) of the mushroom bodies, shows different labelling in distinct species, the peduncle (p) is generally labelled in various species.

Fig. 3. Distribution of Ca²⁺/calmodulin-dependent NOS in the brain of various insect species revealed by the histochemical NADPH diaphorase technique. Representative sections at the level of the antennal lobes of Acheta domesticus (A), Apis mellifera (B), Drosophila melanogaster (C), Gryllus bimaculatus (D) and Schistocerca gregaria (E) are shown. The antennal lobes are the primary centres of olfactory information processing in insects. In all species tested, the glomeruli (g) within the antennal lobes exhibit a strong but individual, inhomogeneous degree of NADPH diaphorase staining. While labelling in the antennal nerve (an) in Schistocerca is hardly visible, the antennal nerve which contains axons of the sensory neurons exhibits staining in the other species. Thus, the arborizations in the glomeruli of Schistocerca exclusively derives from strongly labelled interneurons (in) of the antennal lobe. In the other species, both the sensory neurons as well as the interneurons of the antennal lobe contribute to the staining of the glomeruli. All scales 50 μm.

Fig. 4. Distribution of Ca²⁺/calmodulin-dependent NOS in representative sections at the level of the visual lobes of Acheta domesticus (A), Apis mellifera (B), Drosophila melanogaster (C) and Schistocerca gregaria (D). The visual lobes, consisting of lamina (la), medulla (me) and lobula (lo) exhibit very different NADPH diaphorase labelling in distinct species. While, in Drosophila, labelling is hardly detectable, Apis (B) exhibits an intermediate and Acheta (A) and Schistocerca (D) show strong labelling in the visual lobes. Where labelling is visible, it appears in very distinct layers within the visual neuropiles, suggesting an innervation by defined types of NADPH diaphorase positive neurons. In the lamina of Schistocerca, a subpopulation of monopolar cells which receive input from photoreceptors and project to the medulla express strong staining. By far the strongest staining is observed in the lobula of Acheta and the anterior lobe (al) of Schistocerca. In the medulla of Acheta, a subset of exclusively unstained cells is visible (A, arrow). The retina (r) exhibits no labelling in all species. All scales 50 μm.

Fig. 5. Distribution of Ca²⁺/calmodulin-dependent NOS in representative sections at the level of the central complex of Acheta domesticus (A), Apis mellifera (B), Drosophila melanogaster (C), Gryllus bimaculatus (D) and Schistocerca gregaria (E). In insects, the unpaired central complex has been implicated in the regulation of behavioural activity. Compared with the surrounding neuropil, the central body (cb) with the typical fan-shaped pattern shows a strong, although distinct, staining of the columnar structure in all species. While in Schistocerca (E, arrow), fine arborizations of NADPH diaphorase labelled cells innervate the columns of the upper division of the central body, the other species show distinct staining patterns in the neuropil of the columns, which are probably due to innervation by distinct neurons. In Acheta (A, arrow) a subset of exclusively unstained cells is visible in the lower division of the central body. All scales 50 μm.

Fig. 6. Distribution of Ca²⁺/calmodulin-dependent NOS in representative sections at the level of the mushroom bodies of Acheta domesticus (A), Apis mellifera (B), Drosophila melanogaster (C), Gryllus bimaculatus (D) and Schistocerca gregaria (E). The mushroom bodies, two highly ordered structures have been implicated in higher control of motor programs and mechanisms of learning and memory. With exception of Drosophila (C), where only the α-, β- and γ-lobes (α, β/γ) show weak labelling, in all species the mushroom body neuropiles exhibit a strong compartmentalized staining pattern. While Acheta (A), Apis (B) and Gryllus (D) show labelling in the calyx (cx), the peduncle (p) and the lobes, staining in Schistocerca (E) is observed mainly in the peduncle and the lobes. In Apis (B), the olfactory input into the calyx, the lip (li) area exhibits strong labelling. Since in all species the somata of the mushroom body intrinsic Kenyon cells (kc) are not labelled, the staining in the neuropil is most likely due to extrinsic neurons. All scales 50 μm.

Fig. 7. Scheme of NO action on soluble guanylate cyclase and cGMP cascade. The soluble guanylate cyclase (sGC) is a major target of NO action, while receptor type guanylate cyclase (rGC) is activated by peptide hormones. In contrast to cAMP, which mainly acts on cAMP-dependent protein kinase, the cGMP formed upon activation of sGC activates at least three classes of enzymes. (i) Activation of the cGMP-dependent protein kinases which leads to phosphorylation of substrate proteins. (ii) Action on cGMP-regulated cyclic nucleotide phosphodiesterases (PDE) leads to an increase or a decrease of cAMP-levels. While members of the PDE 2 family are stimulated in their hydrolysing activity, cGMP inhibits enzymes of the PDE 3 family. (iii) The third class of cGMP targets are the cyclic nucleotide gated channels, which are highly conductive for calcium. Co-localization of different cGMP-regulated enzymes within a given cell results in a complex interaction of the NO response with different second messenger cascades.

Fig. 8. Scheme of the function of the diffusible messenger NO in the glomeruli of the antennal lobe, which are ideally suited to serve as diffusion compartments for NO action. Release of NO in a glomerulus theoretically can modulate signal transduction by acting on synaptic connections within this glomerulus. The NO synthase in the glomeruli mainly derives from local interneurons but also from sensory neurons, depending on insect species. While the interneurons innervate several glomeruli, a sensory neuron is usually restricted to a single glomeruli. Thus, release of NO by these two types of neurons can principally modulate distinct aspects in signal processing. (A) Synchronous release of NO in several glomeruli ,e.g. by depolarization of a local interneuron (ln), can modulate synaptic transmission in a subgroup of glomerular circuitries. This results in an integrative modulation of signal transduction. (B) In contrast, release of NO within a single glomerulus, e.g. by sensory neurons (sn), affects synaptic transmission within these glomeruli only.

Fig. 9. Diagram illustrating the specific effect of NO synthase inhibition on memory formation of the honeybee, Apis mellifera. By pairing an odour stimulus with a sucrose reward, the proboscis extension response can be conditioned. (A) In untreated honeybees, a single conditioning trial (pairing of an odour with sucrose reward) leads to a medium-term memory, while multiple trial conditioning (three trials, inter-trial interval 2 min) induces a long-term memory that lasts for days. (B) Blocking of NO synthase activity during associative learning has no effect on single trial memory but selectively impairs a component of the multiple trial memory. Blocking of NO synthase at any time before or after conditioning has no effect on memory formation or retrieval of memory.

Table 1. Properties of Neuronal Ca²⁺/Calmodulin-Dependent NO Synthase

*Data from Muller (1994).

†Data from Bredt and Snyder (1990).

‡Data rom Hope et al. (1991).