# Max-Planck-Institut für Biologische Kybernetik, Spemannstr. 38, D.72076 Tübingen, % Botanisches Institut, Auf der Morgenstelle 1, D-72076 Tübingen,
The notion of a common cellular mechanism underlying rhythms and learning has been investigated previously1-3. The period (per) protein (PER) is an essential component of the Drosophila biological clock4,5. A subset of PER-expressing lateral neurons (LNs) project to the central brain and terminate near the calyces of the mushroom bodies (MBs)6,7. These LN arborizations seem to be critical for the transfer of circadian signals to the central brain (see abstract by Helfrich-Förster). Drosophila MBs are a paired neuropil each consisting mainly of ~ 2500 parallel Kenyon cell fibres derived from four neuroblasts. Feeding hydroxyurea (HU) to newly hatched larvae selectively deletes these cells, resulting in near complete MB ablation (see abstract by Armstrong et al.) and loss of odour learning capacity8. Mutants affecting MB gross anatomy have similar learning deficits9. We have conducted neurohistochemical and behavioural studies with MB-ablated flies8 and brain structure mutants9 to examine whether circadian rhythms and learning are related anatomically through the MBs.
In HU-treated MB-less flies, pacemaker LN arborizations appeared normal. Moreover, these flies had circadian activity rhythms which were not distinguishable in strength from the untreated control during both entrainment in L:D and freerun in D:D. We observed increased activity and period length (by ~ 10 min) in MB-less flies.
We also measured the influence of seven MB mutations on circadian locomotor activity rhythms. Histological preparations were made with every fly to determine whether LNs were affected. All mutations were placed on a common Canton-S (CS) genetic background to control for polygenic effects on brain anatomy and behaviour9 (see abstract by de Belle). Activity rhythms were normal for all mutants with one exception: 21% of mushroom body defect4 (mud4) flies were arrhythmic. This behaviour correlated perfectly with exceptionally enlarged MB calyces and deformities in other brain structures. LNs were present but often in greater number than in CS and with misrouted projections into the central brain. We suppose that mud4 arrhythmicity resulted from pleiotropic effects of the mutation on other central brain structures (i.e., displacement by the enormous calyces) rather than from aberrant MB anatomy specifically. Our results imply that MBs are not involved in circadian locomotor activity rhythms.