The product of the period (per) gene of Drosophila melanogaster is continuously required for the functioning of the circadian pacemaker of locomotor activity. We have used internally marked mosaics to determine the anatomical locations at which per expression is required for adult rhythmicity, and thus where the fly's circadian pacemaker is likely located in this holometabolous insect. We first provide a detailed description of the distribution and nature of per-expressing cells in the fly's CNS. Using an antibody to the per gene product, or to that of a reporter of per expression, in conjunction with an antibody to the embryonic lethal-abnormal visual system (elav) gene product--which is used as a marker of neuronal identity--we have experimentally confirmed previously proposed assignments of per-expressing cells to the neuronal and glial classes. Thus, we found that per expression and elav immunoreactivity colocalized in large cells located in the lateral cortex of the central brain, as well as in more dorsally located cells in the posterior central brain. In contrast, we found that cells located at the margins of the cortex and the neuropil, and within the neuropil, as well as smaller cortical cells found throughout the brain's cortex, were elav negative, supporting the notion that they are glial in nature. Using internally marked mosaics, we find that the pacemaker is located in brain but is not exclusive to the eyes, the ocelli, or the optic lobes, which is consistent with previous reports obtained in this and other insects of this class. Although the pacemaker may be a paired structure, we show that the functioning of one of them is sufficient for rhythmicity. Finally, we report that glial expression is sufficient for some behavioral rhythmicity to be manifest. However, the rhythmicities of animals for which per expression was confined to glia were weak, suggesting that neuronal per expression as well may be required for normal pacemaker function.

No abstract

Rhythmic oscillations of the PER protein, the product of the Drosophila period (per) gene, in brain neurons of the adult fly are strongly involved in the control of circadian rhythms. We analyzed temporal and spatial expression patterns of three per-reporter fusion genes, which share the same 4 kb regulatory upstream region but contain increasing amounts of per's coding region fused in frame to the bacterial lacZ gene. The fusion proteins contained either the N-terminal half (SG), the N-terminal-two-thirds (BG), or nearly all (XLG) of the PER protein. All constructs led to reporter signals only in the known per-expressing cell types within the anterior CNS and PNS. Whereas the staining intensity of SG files was constantly high at different Zeitgeber times, the in situ signals in BG and XLG files cycled with approximately 24 hr periodicity in the PER-expressing brain cells in wild-type and per01 loss of function files. Despite the rhythmic fusion-gene expression within the relevant neurons of per01 BG files, their locomotor activity in light/dark cycling conditions and in constant darkness was identical to that of per01 controls, uncoupling protein cycling from rhythmic behavior. The XLG construct restored weak behavioral rhythmicity to (otherwise) per01 files, indicating that the C-terminal third of PER (missing in BG) is necessary to fulfill the biological function of this clock protein.

Central nervous system ganglia within the head of the beetle Pachymorpha sexguttata were labeled using an antibody that recognizes an evolutionarily conserved region of the period (per) gene product of Drosophila melanogaster. per and the protein it encodes (PER) are believed to play a central role in the generation of endogenous circadian rhythms in flies; therefore anti-PER-mediated immunoreactivity may help to uncover cellular components of the circadian clock system in that insect and in others. In the beetle, application of this antibody led to the staining of a distinct set of neurons located in the optic lobes and the central brain, plus small numbers of putative glial cells in the optic lobes. Neuronal perikarya (including their nuclei in a few cases), the axons, and terminal regions of the neurons were stained. The network formed by these labeled cells and processes are candidates for the neuronal basis of the beetle's circadian clock system: the pacemaker region situated next to the medulla neuropil, its connection to the apparent site of Zeitgeber input, and putative efferent pathways projecting to control centers of various effector systems. Anti-PER-mediated labeling and that resulting from application to beetle specimens of an antiserum against pigment-dispersing hormone (PDH) were compared; in the Drosophila brain all "PDH cells" express the per gene as well. In the beetle, however, the set of "PER cells" and PDH ones is at least in part nonoverlapping. The hypothesis that neurons stained by application of anti-PER participate in the control of the beetle's circadian rhythms is discussed in the context of previous electrophysiological and immunohistochemical studies. Also considered are analogies to, and differences from, labeling of the PER protein in fruit flies and PER-like immunoreactivity in other animals.

Colonies of the South African honeybee race Apis mellifera capensis (Escholtz) were maintained under constant conditions of illumination (200 lux), temperature (25 k l°C) and relative humidity (65*3%). Activity was measured at the hive entrance. After ad libitum feeding for at least 5 days, food was presented for only 2 h/day either for 1 week (series 1) or for 2 weeks (series 2). In the last part of each experiment, food was again available all the time. Colonies which showed free-running circadian activity rhythms (with periods ranging from 22.6 to 24.8 h) during ad libitum feeding were submitted to feeding cycles with interfeeding intervals (T) of 22,23,24 and 25 h. In most of these experiments the rhythms were synchronized by the feeding schedule, resulting in a stable phase-angle difference between onset of activity and onset of food availability. The duration of this anticipatory activity was positively correlated with T. When ad libitum feeding was resumed, the period of the rhythm induced by the feeding schedule persisted for a few days. Thereafter, the rhythm was free-running again with a period close to that observed in the first part of the experiment. The conclusion is drawn that, under the influence of periodic feeding, the activity of honeybee colonies has the characteristics of an entrained circadian system.