The organization of putative octopaminergic pathways in the brain and subesophageal ganglion of the honeybee was investigated with a well-defined polyclonal antiserum against octopamine. Five prominent groups of just over 100 immunoreactive (IR) somata were found in the cerebral ganglion: Neurosecretory cells in the pars intercerebralis innervating the corpora cardiaca via NCC I, one cluster mediodorsal to the antennal lobe, one scattered on both sides of the midline of the protocerebrum, one between the lateral protocerebral lobes and the dorsal lobes, and a single soma on either side of the central body. With the exception of the pedunculi and beta-lobes of the mushroom bodies, varicose immunoreactive fibers penetrate all parts of the cerebral ganglion. Strong labelling was found in the central complex and the protocerebral bridge. Fine networks of labelled processes invade the antennal lobes, the calyces and a small part of the alpha-lobes of the mushroom bodies, the protocerebrum, and all three optic ganglia. In the subesophageal ganglion, one labelled cell body was found in the lateral soma layer of the mandibular segment. Each of the three neuromeres contains a group of six to ten somata in the ventral median parts. Most of the ventral median cells send their neurites dorsally through the midline tracts, whereas the neurites of a few cells follow the ventral cell body neurite tracts. Octopamine-IR was demonstrated in all neuropils that contain pathways for proboscis extension learning in honeybees. Because octopaminergic mechanisms seem to be involved in the behavioral plasticity of the proboscis extension reflex, our study provides anatomical data on the neurochemical organization of an appetitive learning paradigm.

A histochemical staining method for acetylcholinesterase (AChE) and an antiserum raised against nicotinic acetylcholine receptors (AChR) of locust nervous tissue were applied in order to reveal certain candidates of cholinergic pathways in the brain of the honeybee. The AChE staining marked layers in the optic lobes, fibers connecting the two brain hemispheres, and fiber tracts as well as soma clusters within the protocerebrum. The calycal input regions of the mushroom bodies were labelled, whereas the intrinsic Kenyon cells showed no staining. Although the antennal afferents projecting into the dorsal lobe showed strong AChE activity, projections into the antennal lobe showed rather weak staining. Application of the antiserum against the AChR showed immunoreactivity in neuropiles, tracts, somata, and the antennal nerve. The immunoreactivity of the optic lobes coincided with the banding pattern of the AChE staining. A particularly striking overlap of AChR immunoreactivity and AChE staining was found in the lip neuropile of the mushroom bodies, which would suggest a cholinergic input into this neuropile via fibers of the median antennoglomerular tract. Because the antiserum against locust AChR binds in neuropiles displaying AChE activity, we conclude that this antiserum also cross-reacts with the bee's receptor. This interpretation is supported by experiments showing alpha-bungarotoxin (alpha-BTX) binding sites in some areas of strong immunoreactivity.

Information transmission and processing in the brain is achieved through a small family of chemical neurotransmitters and neuromodulators and a very large family of neuropeptides. In order to understand neural networks in the brain it will be necessary, therefore, to understand the connectivity, morphology, and distribution of peptidergic neurons, and to elucidate their function in the brain. In this study we characterize the distribution of substances related to Dip-allatostatin I in the honeybee brain, which belongs to the allatostatin-A (AST) peptide family sharing the conserved c-terminal sequence -YXFGL-NH(2). We found about 500 AST-immunoreactive (ASTir) neurons in the brain, scattered in 18 groups that varied in their precise location across individuals. Almost all areas of the brain were innervated by ASTir fibers. Most ASTir neurites formed networks within functionally distinct areas, e.g., the antennal lobes, the mushroom bodies, or the optic lobes, indicating local functions of the peptide. A small number of very large neurons had widespread arborizations and neurites were found in the corpora cardiaca and in the cervical connectives, suggesting that AST also has global functions. We double-stained AST and GABA and found that a subset of ASTir neurons were GABA-immunoreactive (GABAir). Double staining AST with backfills of olfactory receptor neurons or mass fills of neurons in the antennal lobes and in the mushroom bodies allowed a more fine-grained description of ASTir networks. Together, this first comprehensive description of AST in the bee brain suggests a diverse functional role of AST, including local and global computational tasks.

The olfactory system is a classical model for studying sensory processing. The first olfactory brain center [the olfactory bulb of vertebrates and the antennal lobe (AL) of insects] contains spherical neuropiles called glomeruli. Each glomerulus receives the information from one olfactory receptor type. Interglomerular computation is accomplished by lateral connectivity via interneurons. However, the spatial and functional organization of these lateral connections is not completely understood. Here we studied the spatial logic in the AL of the honeybee. We combined topical application of neurotransmitters, olfactory stimulations, and in vivo calcium imaging to visualize the arrangement of lateral connections. Suppression of activity in a single glomerulus with γ-aminobutyric acid (GABA) while presenting an odor reveals the existence of inhibitory interactions. Stimulating a glomerulus with acetylcholine (ACh) activates inhibitory interglomerular connections that can reduce odor-evoked responses. We show that this lateral network is patchy, in that individual glomeruli inhibit other glomeruli with graded strength, but in a spatially discontinuous manner. These results suggest that processing of olfactory information requires combinatorial activity patterns with complex topologies across the AL.

Primary cell cultures were prepared from specific regions of the pupal honeybee brain which are involved in proboscis extension learning. Defined areas could be dissociated purely by mechanical treatment. We show that cultured neurons regenerate new neurites and remain viable for up to three weeks in a serum-free, chemically-defined medium. Several labelling techniques were employed to identify subpopulations of cultured neurons. For example, acetylcholinesterase staining; fluorescent beads to distinguish identified cell populations of co-cultured brain areas; various markers for surface antigens such as a monoclonal antibody to olfactory projection neurons of the antennoglomerular tracts and monopolar cells of the optic lobes, as well as anti-HRP immunoreactivity and alpha-bungarotoxin binding; and various antisera for detecting transmitter phenotype. The appearance of transmitter-immunoreactive cells agreed closely with that expected from their known distribution in situ. Our results suggest that cultured cells retain surface properties and transmitter phenotype of their in vivo counterparts, despite differences in basic morphology. Thus our culture system provides the important initial step for future in vitro investigations of the cellular and electrophysiological properties of neurons mediating proboscis extension learning.