Circadian plasticity in photoreceptor cells controls visual coding efficiency in Drosophila melanogaster

Abstract

In the fly Drosophila melanogaster, neuronal plasticity of synaptic terminals in the first optic neuropil, or lamina, depends on early visual experience within a critical period after eclosion. The current study revealed two additional and parallel mechanisms involved in this type of synaptic terminal plasticity. First, an endogenous circadian rhythm causes daily oscillations in the volume of photoreceptor cell terminals. Second, daily visual experience precisely modulates the circadian time course and amplitude of the volume oscillations that the photoreceptor-cell terminals undergo. Both mechanisms are separable in their molecular basis. We suggest that the described neuronal plasticity in Drosophila ensures continuous optimal performance of the visual system over the course of a 24 h-day. Moreover, the sensory system of Drosophila cannot only account for predictable, but also for acute, environmental changes. The volumetric changes in the synaptic terminals of photoreceptor cells are accompanied by circadian and light-induced changes of presynaptic ribbons as well as extensions of epithelial glial cells into the photoreceptor terminals, suggesting that the architecture of the lamina is altered by both visual exposure and the circadian clock. Clock-mutant analysis and the rescue of PER protein rhythmicity exclusively in all R1-6 cells revealed that photoreceptor-cell plasticity is autonomous and sufficient to control visual behavior. The strength of a visually guided behavior, the optomotor turning response, co-varies with synaptic-terminal volume oscillations of photoreceptor cells when elicited at low light levels. Our results show that behaviorally relevant adaptive processing of visual information is performed, in part, at the level of visual input level.

Figure 1. Experience-dependent plasticity in the lamina of wild-type Drosophila melanogaster.

a, Autofluorescence profiles of the lamina neuropil (La) from adult flies reared for four days in constant darkness (DD) or constant light (LL; flies were fixed at ZT 6). Note the gross morphological size changes of the laminae. b, Electron-micrographs of distal lamina sections of flies sectioned at day and night. Six synaptic terminals of photoreceptor cells (R) converge on to two Large Monopolar Cells (L), forming the so called cartridge. Scale bar, 1 µm. c, Significant differences between LL- (white columns) and DD-flies (black columns) were found for lamina volume, cross-sectional area of the photoreceptor terminals (1/6 hatched in b), circumference of the membrane surrounding these terminals (1/6 dotted line in b) and presynaptic T-bars residing on this membrane. Spacing between neighboring synapses was statistically indistinguishable. Finally, in LL-flies more shallow capitate projections were found, whereas in DD-flies more deep capitate projections invaded the photoreceptor terminals. Number of presynaptic T-bars and capitate projections were counted per section. Error bars denote s.e.m. values.

Figure 2. Sensitivity and dynamic range of optomotor behavior depends on early visual experience.

a, The optomotor response: A tethered female fly attempts to follow the rotation of a periodical pattern of 27.7° width at 2.6 Hz contrast frequency and thereby deflects its abdomen. Flies were stimulated by both counter-clockwise (CCW) and clockwise (CW) rotation of the cylinder and the behavioral net response was calculated . b, Optomotor responses were maximal at about 10 cd m⁻² ( = 100%) for all tested fly groups. At low light intensities behavioral responses of flies reared in DD₄- or DD₂LL₂-conditions (two days darkness followed by two days light) were statistically indistinguishable, with both displaying a higher optomotor sensitivity than LL₄-reared flies. Flies were kept under LL-conditions before eclosion. c, The stimulus was restricted in the azimuth direction to various angular ranges of the flies' frontal visual field. With increasing area of frontal stimulation the optomotor responses of LL-reared flies increased continuously in a fine-graded manner, whereas the optomotor response of DD-reared flies saturated already at 45° of stimulated azimuth (n = 8 to 11 flies per condition). The dynamic range (D.R.), computed as the quotient of the optomotor response at 360° stimulation versus that at 30°–60°, was therefore higher in LL-flies than in DD-flies.

Figure 3. Day/night structural changes of photoreceptor cells and capitate projections.

a, Typical cross-sectional profiles of cartridges at the distal lamina. At ZT 7 the axonal profiles reach their largest cross-sectional area, in the middle of the night (ZT 17) they are smallest. Scale bar, 1.0 µm. b, Examples of presynaptic ribbons of tetrad synapses (arrows; arrowheads, capitate projections). Scale bar, 0.5 µm. c, The analysis of photoreceptor cell area, membrane circumference and residing presynaptic T-bars shows robust changes during the 24-hour cycle (mean size ± s.e.m. of 40–55 photoreceptor cells; about 10 cartridges per fly; n_ZT1 = 5, n_ZT7 = 5, n_ZT12 = 5, n_ZT17 = 6, n_ZT0 = 5 flies were analyzed for each point of time). d, Capitate projections of epithelial glial cells (arrowheads) were distinguished based on their position relative to the photoreceptor terminal. Shallow capitate projections (above) embrace the axon on its surface, deep capitate projections (below) extrude deep into the photoreceptor cell terminal. Small arrows indicate synaptic vesicles inside the photoreceptor cell axon. Scale bar, 0.1 µm. e, Shallow capitate projections were found to be most abundant during day times, deep ones during night times. Fitting the data to 24-hour sine-waves revealed that the two curves were in anti-phase by about 12 h.

Figure 4. Structural plasticity of the lamina is controlled by the parallel action of a circadian clock and a phototransduction-dependent mechanism.

a, Lamina volumes of adult wild-type CS (WT CS) flies kept under LL, LD, or DD conditions during adult life. All flies had experienced 12:12 LD cycles throughout larval and pupal development. The lamina volume oscillated in a circadian manner around a mean value which was larger in light-experienced flies than in dark-reared flies (F(2,1699)  = 156.7, p<0.0001, ANOVA). Sinusoidal-function fitting revealed a best fitting period of 23.7 h for LL (F(1,563) = 17.5, p<0.0001) and DD (F(1,515) = 5.6, p<0.05) lamina volume oscillations and 24.0 h for LD (F(1,621) = 34.2, p<0.0001). ‘Circadian Time’ (for LL and DD) ZT 0  =  light on, ZT 12  =  light off; for LL- and DD-flies read ZT as ‘Circadian Time’. b, Calculated amplitude deviations from the respective mean lamina volumes of all data points in a. Light experience significantly modified the time course of the circadian lamina-volume oscillations. This is most apparent at ZT 9, where LL reared flies showed the highest volume and DD reared flies the lowest volume. c, Lamina volumes of 10 to 15 adult mutant male flies per point of time kept under LL, LD or DD conditions. Circadian volume oscillations of the lamina were absent in the two clock mutants per⁰¹ and tim⁰¹ reared under LL- or DD-conditions. LD-reared animals displayed an increase in lamina volume during the day and a rapid decrease at night. The blind norpA^P24 mutant displayed a robust circadian rhythm of lamina volumes. Under all rearing conditions the mean lamina volume rested at the DD-level of wild-type flies. d, Rescue of circadian lamina-volume oscillations in per⁰¹-mutants by the transgenic expression of PER-protein exclusively in photoreceptor cells R1-6 using per⁰¹; Rh1(−180)-per−1/+ flies. LD-reared R1-6-rescue flies displayed lamina-volume oscillations which were statistically indistinguishable in amplitude and time course from those of WT CS flies. Constant LL - or DD-conditions delayed the subjective day peaks by several hours in R1-6-rescued animals compared to wild-type flies. These observations are consistent with a previously described delay of the PER-protein cycle under DD-conditions in animals of the same genotype. For genetic reasons, in this experiment female flies were used. They were raised under LD-conditions up to their second day of adulthood before being reared for two more days in LL, LD or DD (n = 8 to 10 per point of time). Note that the overall lamina volume in females is higher than in males.

Figure 5. Sensitivity and dynamic range of optomotor behavior is controlled by experience-dependent and circadian mechanisms.

a, b, LD-reared WT flies were behaviorally tested in the paradigms described in Fig. 2. Both, the sensitivity of the optomotor response at low light intensities (a, n>16 per point of time) and the dynamic range of optomotor behavior (b, n>30 per point of time) oscillated in a circadian, anti-phasic manner: highest behavioral sensitivity at night was accompanied by lowest dynamic range, and vice versa. c, Rescue of circadian oscillations in optomotor sensitivity by the restoration of the circadian clock in photoreceptor cells (Fig. 4d). Adult flies of the indicated genotypes were reared as described above and were behaviorally tested during their night- or day-phases (ZT 18–21 [set to 100%] and ZT 6–9, respectively). Irrespective of the genotype all LD-reared flies displayed a higher optomotor sensitivity during the night (F(1,74) = 14.6, p<0.0005). DD-reared wild-type flies as well as clock-rescued flies (per⁰¹; Rh1(−180)-per−1/+) showed a similar subjective day/night difference in behavioral sensitivity, which was absent in dark-reared per⁰¹-mutants. Error bars indicate s.e.m. values.

Figure 6. Model of parallel pathways.

Light serves at least two functions in the visual system of Drosophila, it entrains and keeps the autonomous circadian clock of photoreceptors in phase and it triggers the phototransduction cascade. Both cellular mechanisms are active in parallel in photoreceptor cells and both converge in the volume control of their synaptic terminals in the lamina. The neuronal readout of this peripherally controlled morphological and functional plasticity is further computed downstream of the photoreceptor terminals, within the lamina and/or e.g. in the lobula plate, to instruct the appropriate behavior.