The CCHamide1 Neuropeptide Expressed in the Anterior Dorsal Neuron 1 Conveys a Circadian Signal to the Ventral Lateral Neurons in Drosophila melanogaster

(; Bloomington Drosophila Stock Center [BDSC] #5905) flies were used as the control strain in this study. Additionally, the following strains were used:() (),(),(),(BDSC #1522),(),;(Vienna Drosophila Resource Center [VDRC] #60012),(VDRC #104974),(VDRC #103055),(BDSC #51261),(BDSC #43865),(), and(; BDSC #55851). To exclude unwanted mutations,,,, andstrains were outcrossed at least six times withcontrol flies. All RNAi strains and corresponding control strains co-expressedto enhance the RNAi efficiency (). To generate theline, we employed the recombinase-mediated cassette exchange (RMCE) system to insert aconstruct into the 5′-untranslated region (UTR) of thegene (). Thegene trap cassette,, was injected into theline by BestGene (BestGene Inc., Chino Hills, CA, United States). The RMCE integration was verified by polymerase chain reaction (PCR), as described previously (). Flies were reared at 25°C under a 12-h light: 12-h dark cycle (LD) onmedium (0.7% agar, 8.0% glucose, 3.3% yeast, 4.0% cornmeal, 2.5% wheat embryo, and 0.25% propionic acid). Drosophila melanogaster w 1118 w w;tim(UAS)-Gal4 tim-Gal4 w;Pdf-Gal4 w;clk856-Gal4 y w;UAS-GFP S65T w;UAS-Epac1-camps w UAS-dicer2 w;UAS-CCHa1 RNAi w;UAS-CCHa1-R RNAi Mi{MIC}CCHa1 MI09190 Mi{MIC} MI03750 y w;CCHa1 SK8 daughterless-Gal4 da-Gal4 w;Pdf-Gal4, w;UAS-CCHa1 RNAi w;UAS-CCHa1-R RNAi w;da-Gal4 y w;CCHa1 SK8 w UAS-dicer2 CCHa1-R-Gal4 Gal4 CCHa1-R Gal4 pBS-KS-attB1-2-GT-SA-Gal4-Hsp70pA Mi{MIC} MI03750 Drosophila [Blau and Young, 1999] [Renn et al., 1999] [Gummadova et al., 2009] [Shafer et al., 2008] [Ren et al., 2015] [Dietzl et al., 2007] [Venken et al., 2011] [Venken et al., 2011]

Male flies aged 3–6 days were used to record locomotor activity rhythms. Flies were confined to recording tubes containing agar/sugar food (2% agar and 4% sucrose) for theActivity Monitor (DAM2, Trikinetics Inc., Waltham, MA, United States). The monitors were placed in an incubator (CN-40A, Mitsubishi Electric, Tokyo, Japan) and maintained at constant temperature of 20 (±0.25)°C. Standard cool white light-emitting diodes were set above the monitors in the incubator, and were controlled by an LC4 light controller (Trikinetics Inc.). The light intensity used in all experiments was 100 lux (3.2 μW⋅). We recorded the activity of flies in LD for 7 days, followed by 14 days of constant darkness (DD). Activity was recorded in 1-min bins using a conventional infrared light sensor that counted the number of beam crosses for each individual fly. Drosophila -2

For visual inspection, raw data were displayed as actograms using ActogramJ(). Free-running periods in DD were determined using a chi-square periodogram analysis. If a robust peak above the 0.05 confidence level appeared in the periodogram, the period was designated as statistically significant (). 1 [Schmid et al., 2011] [Sokolove and Bushell, 1978]

Average activity profiles and siesta durations were calculated on 3 consecutive days (days 5–7) in LD. The average activity profiles were smoothed using an 11-point moving average. To calculate siesta duration, daily sleep of individual flies was analyzed in 30-min intervals using a macro program written in Microsoft Excel (Microsoft, Redmond, WA, United States;). The times when flies began to sleep longer than 10 min in the morning and shorter than 10 min in the evening were assigned as the onset and offset of siesta, respectively. The average activity profiles in the 1st day of DD were smoothed using a 31-point moving average. [Gmeiner et al., 2013]

Whole flies were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) with 0.1% Triton X-100 for 2.5 h at room temperature (RT). Fixed flies were washed three times with PBS, and then their brains were dissected. The brains were then washed three times with PBS containing 0.5% Triton X-100 (PBS-T), after which they were blocked in PBS-T containing 5% normal donkey serum for 1 h at RT and subsequently incubated with primary antibodies at 4°C for 48 h. After washing six times with PBS-T, the brains were incubated with secondary antibodies at RT for 3 h, washed six more times in PBS-T, and mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA, United States). The primary antibodies used were rabbit anti-CCHa1 (), chicken anti-green fluorescent protein (GFP) (1:1000; Rockland, Limerick, PA, United States), mouse anti-PDF (1:1000; Developmental Studies Hybridoma Bank) (), rabbit anti-PDP1 (1:9000) (), rat anti-TIMELESS (TIM) (1:3000) (), and rabbit anti-PDF (1:15000) (). We used the following fluorescence-conjugated secondary antibodies at a 1:500 dilution: Alexa Fluor488 nm (goat anti-chicken, goat anti-mouse), 555 nm (goat anti-rabbit), 647 nm (goat anti-mouse), antibodies (Life Technologies, Carlsbad, CA, United States); and goat anti-rabbit and anti-rat Cy3 antibodies (Millipore, Billerica, MA, United States). [Veenstra and Ida, 2014] [Cyran et al., 2005] [Cyran et al., 2003] [Yoshii et al., 2008] [Abdelsalam et al., 2008] Gryllus ®

Staining was visualized using laser scanning confocal microscopes (Olympus FV1200, Olympus, Tokyo, Japan). For quantification of immunostaining signal, the confocal microscope settings were kept consistent throughout the experiments. Measurement of staining intensity was performed using Fiji software (), as described previously (). For quantification of PDP1 staining, the intensity values were obtained from all the stained clock neurons. In our confocal images, the size of the nuclei of clock neurons was approximately 8 pixels. Therefore, for each cell, a circle of 8 pixels marking the nucleus of the cell, where PDP1 is localized, was selected using the selection brush tool, and the mean intensity of these 8 pixels was measured. The background intensity was subtracted from the signals, and then the mean intensities per cell group and brain hemisphere were calculated. For quantification of CCHa1 staining, the polygon selection tool of Fiji was used to select each cell body of the DNneurons. For quantification of PDF staining and area of branching at the s-LNterminals, the polygon selection tool was used to select the area of the branching s-LNdorsal terminals that were visualized upon Z-projection of multiple confocal stacks (). To visualize projections of the DN, s-LN, and l-LNneurons, we traced and reconstituted their projections using the Fiji plugin Simple Neurite Tracer (). [Schindelin et al., 2012] [Yoshii et al., 2009a] [Kozlov et al., 2017] [Longair et al., 2011] 1a v v 1a v v

Male flies, aged 5–7 days, of the genotypeorwere anesthetized on ice, and their brains were dissected in cold hemolymph-like saline (HL3) (). The experiments were conducted between zeitgeber time (ZT) 6–9. The brains were then mounted at the bottom of a plastic petri dish in HL3 with their anterior surface facing upward and allowed to recover from dissection for 10–15 min prior to imaging. An epifluorescence imaging setup (VisiChrome High Speed Polychromator System, ZEISS Axioskop2 FS plus, Visitron Systems GmbH, Puchheim, Germany) with a 40× water-immersion objective (ZEISS 40×/1.0 DIC VIS-IR) was used to conduct the experiments. The LNclock neurons were located according to their characteristic positions, and regions of interest were defined on single cell bodies using the VisiView Software (version 2.1.1, Visitron Systems GmbH). Time-lapse frames were acquired at 0.2 Hz for 750 s, exciting the cyan fluorescent protein (CFP) fluorophore of the ratiometric cAMP sensor with 405 nm light. Emissions of CFP and YFP were detected separately with a charge-coupled device camera (Photometrics, CoolSNAP HQ, Visitron Systems GmbH) using a beam splitter. After measuring baseline CFP and YFP levels for ∼100 s, substances were applied drop-wise using a pipette. The CCHa1 peptide (synthesized by Peptide Institute Inc., Osaka, Japan) was diluted in 0.1% dimethyl sulfoxide (DMSO) in HL3 to a concentration of 10M; 10M of the forskolin derivative, NKHin HL3 with 0.1% DMSO was used as positive control, whereas HL3 alone with 0.1% DMSO served as negative control. Inverse Förster resonance energy transfer (iFRET) was calculated according to the following equation: iFRET = CFP/(YFP-CFP0.357) (). In this equation, CFP and YFP are background-corrected raw fluorescence data, and 0.357 is subtracted from YFP fluorescence as this is the fraction of CFP spillover into the YFP channel in our imaging setup. Finally, iFRET traces of individual neurons were normalized to baseline levels and averaged for each treatment. To quantify and statistically compare the response amplitudes of each treatment, maximum iFRET changes were determined for individual neurons. w;clk856-Gal4;UAS-Epac1-camps UAS-dicer2;clk856-Gal4/CCHa1-R;UAS-Epac1-camps/+ RNAi [Stewart et al., 1994] [Shafer et al., 2008] v -5 -5 477 ∗

Statistics were performed using Systat (v 11.00.01, Systat Software Inc., San Jose, CA, United States). Data were compared using a one-way analysis of variance (ANOVA) test followed bypairwise comparison with Bonferroni correction. The-values of data that were not normally distributed were corrected by multiplying by 2, according to. post hoc p [Glaser (1978)]

(control) andstrains were used to examine the efficiency ofRNAi;is ubiquitously expressed (). Male flies were anesthetized on ice at ZT2 in LD, and 15 heads were quickly sampled for each strain. Total RNA was extracted using a spin column-based RNA extraction kit (RNA Basic Kit, Nippon Genetics, Tokyo, Japan); 120 ng of total RNA was used for reverse transcription (ReverTra Ace with gDNA Remover, Toyobo, Osaka, Japan). Quantitative PCR was performed by Mx3000P Real-Time PCR System (Agilent, CA, United States) using Thunderbird SYBR qPCR Mix (Toyobo, Osaka, Japan). We used the following primers:forward primer, 5′-GTTTCCAGTCCCGTCACTCC-3′ and reverse primer, 5′-GTGCGTGTGCTTCATTGTGA-3′;(used as a house-keeping gene) forward primer, 5′-AGCTGAACCTCTCGGGACAC-3′ and reverse primer, 5′-TGCCTCGGACTGCCTTGTAG-3′ (). The delta-delta CT method was used to determine the relative amount of themRNA. UAS-dicer2;+/UAS-CCHa1-R RNAi UAS-dicer2;da-Gal4/UAS-CCHa1-R RNAi CCHa1-R da-Gal4 CCHa1-R RpL13A CCHa1-R [Cronmiller and Cummings, 1993] [Ling and Salvaterra, 2011]

The CCHa1-immunoreactive neurons are widely distributed in the adult brain (). Many small cell bodies were stained in the optic lobes as well as relatively large cell bodies along the lateral protocerebrum, subesophageal ganglion, and intermediate and superior medial protocerebrum. To identify clock neurons, we performed a double staining with the anti-TIM and anti-CCHa1 antibodies; TIM and CCHa1 were clearly co-localized in the DNneurons () but not in other clock neurons (). Figure 1A Figures 1A,B Supplementary Figure S1 1a

Circadian expressions of PDF in the s-LNand the Ion transport peptide (ITP) in the 5th s-LNand ITP-positive LNhave been previously reported (;). Therefore, we investigated whether the expression of CCHa1 in the DNneurons is also regulated by the circadian clock. While the rhythmic changes in PDF and ITP levels are observed in their axonal projections, obvious difference in CCHa1 staining intensity was observed not only in the DNterminals but also in the cell bodies (). Since the diffused and fine dendritic arborizations of the DNterminals made an objective quantification difficult, we quantified CCHa1 staining only in the cell bodies. We did so for two independent experiments in LD and three independent experiments in DD. CircWave () and one-way ANOVA revealed circadian changes in staining intensity in DNcell bodies in all experiments except for one in DD (). When we pooled the two experiments in LD and three experiments in DD, CircWave and ANOVA revealed circadian rhythms under both conditions (LD and DD) (). In LD, CCHa1 exhibits a trough at ZT8 and a peak at ZT17 (and). In DD, the rhythm maintains a similar phase, but with dampened amplitude. v v d 1a 1a 1a 1a [Park et al., 2000] [Hermann-Luibl et al., 2014] [Oster et al., 2006] Figure 2A Supplementary Figure S2 Figure 2B Figure 2B Supplementary Figure S2

To investigate the effect of the clock on the CCHa1 rhythm, the brains ofmutants together with control flies were immunostained at two time points, ZT8 and ZT17, in LD. The CCHa1 staining intensity was significantly different between the two time points in control flies, whereasmutants did not show any difference (), suggesting that the rhythmic expression of CCHa1 is controlled by the clock. per 1 per 1 Figure 2C

To investigate the CCHa1-R localization in thebrain, we generated aline using the RMCE method in theline. For this, the-mediated integration cassette was inserted using site-directed mutagenesis in the 5′-UTR intron of thegene () (). Theline showed GFP expression in the entire brain, including strong expression in the mushroom bodies (). Furthermore, triple staining with anti-GFP, anti-PDP1, and anti-PDF antibodies revealed thatis expressed in the l-LNand s-LNneurons but not in other clock neurons (). Moreover, GFP was strongly expressed in the l-LNneurons and weakly in the s-LNneurons, suggesting that the levels of theexpression are different between these cell groups. Drosophila CCHa1-R-Gal4 Mi{MIC} MI03750 Minos CCHa1-R CCHa1-R-Gal4/UAS-GFP CCHa1-R CCHa1-R Figure 3A Figure 3B Figures 3C,D [Venken et al., 2011] v v v v

Using theline, in which thegene trap cassette was inserted into the second exon of thegene, we were able to clearly visualize that the DNprojections run along the s-LNdorsal projections (), in agreement with previous studies (;). Mi{MIC}CCHa1 MI09190 gfp CCHa1 1a v Figure 3E [Shafer et al., 2006] [Helfrich-Förster et al., 2007]

The receptor of CCHa1 is a G protein-coupled receptor and belongs to the bombesin receptor subtype 3 (;). To investigate the physiological response to CCHa1 in the s-LNand l-LNneurons, we conducted live cAMP imaging in the adult brain using aline expressing a genetically encoded cAMP FRET sensor,(;). To evaluate our imaging method, we recorded inverse FRET values (CFP/YFP ratios) for single s-LNand l-LNcell bodies in living brains offlies. The flies were treated with HL3 saline or NKH(an adenylyl cyclase activator) which acted as negative and positive controls, respectively. Both the s-LNand l-LNneurons responded to bath application of 10M NKHby increasing cAMP, and did not respond to HL3 (). Interestingly, the s-LNneurons showed a significant increase in cAMP levels in response to 10M CCHa1, whereas the l-LNneurons did not. To determine whether the s-LNneurons directly respond to CCHa1 through CCHa1-R, we knocked downin the s-LNneurons (), and repeated cAMP imaging. The efficiency ofknockdown was verified by quantitative PCR using(). Since we used a single copy ofin this imaging experiment, the inversed FRET values in response to the application of NKHwere lower than in theflies. Nevertheless, the response to CCHa1 was severely reduced after knockdown of its receptor (). [Hewes and Taghert, 2001] [Ida et al., 2012] [Nikolaev et al., 2004] [Shafer et al., 2008] v v v v v v v v v v UAS UAS-Epac1-camps clk856-Gal4;UAS-Epac1-camps CCHa1-R dicer2;clk856-Gal4/UAS-CCHa1-R;UAS-Epac1-camps/+ RNAi CCHa1-R da-Gal4 UAS-Epac1-camps clk856-Gal4;UAS-Epac1-camps 477 -5 477 -5 477 Figure 4 Supplementary Figure S3A Figures 4A,B

To investigate the effect of CCHa1 on behavioral rhythms, we recorded locomotor activity rhythms ofmutants andknockdown flies under LD and DD. The efficiency ofknockdown was verified by immunostaining ().mutants andknockdown flies were significantly less active than the corresponding control flies (); this mainly concerned morning activity. In LD, the morning activity bout was narrower in flies without CCHa1, and in case of the mutants, the offset of morning activity was much earlier, which resulted in a longer siesta (). In theknockdown flies, this was less evident, but they still slept significantly longer during their siesta than the controls (). Moreover, under DD,mutants andknockdown flies were significantly less active in the morning than the corresponding control flies, resulting in a shift of most activities into the second half of the active phase (). Altogether, our results suggest that CCHa1 is particularly involved in the generation of robust morning activity and in its timing. Under LD, the diminished morning activity of flies lacking CCHa1 appeared advanced, whereas under DD it appeared delayed. CCHa1 CCHa1 CCHa1 CCHa1 CCHa1 CCHa1 CCHa1 CCHa1 Supplementary Figure S3B Figures 5A,B Figures 5C,D Figure 5C Figures 6A,B

Most interestingly, the absence of CCHa1 had virtually no influence on the free-running period in DD. Althoughandflies showed slightly but significantly shorter free-running periods than the correspondingandcontrol strains (), the free-running periods ofnull mutants andflies were not significantly different from their control strains. Thus, even if there is a difference in the free-running period, it would be small, and the genetic background may influence the period more than CCHa1. tim > CCHa1-R RNAi clk856 > CCHa1 RNAi UAS Gal4 CCHa1 SK8 tim > CCHa1 RNA Table 1 i

Since CCHa1 signaling from the DNneurons mainly affects morning activity of the flies, it should influence the molecular oscillations of the s-LNneurons that control it. To test this, we measured the abundance of the PDP1 clock protein in clock neurons ofknockdown flies during LD and DD (). We sampled brains at day 5 (starting from circadian time [CT] 1) in DD, because we expected an internal desynchronization among the clock neurons inknockdown flies after several cycles of a free-running condition but not merely after the 1st day of DD. As expected, the largest differences in PDP1 oscillations betweenand control flies were present in the s-LNneurons. In LD, theknockdown phase-advanced the decrease of PDP1 in the morning (arrow in), which fits the earlier decrease in morning activity that we observed. In DD, it phase-delayed the entire PDP1 oscillation (arrows in), which correlates with the delay in the activity of the mass center (green arrows in). 1a v v CCHa1 CCHa1 tim > CCHa1 RNAi CCHa1 Figure 7 Figure 7A Figure 7B Figure 6

Theknockdown also affected the other clock neurons significantly. It reduced PDP1 levels in most neurons. In LD, this reduction occurred predominantly during the time at which the PDP1 level of controls was already at its trough. Consequently, the knockdown enhanced the amplitudes of the PDP1 oscillations (). In DD, a higher-amplitude cycling of PDP1 afterknockdown occurred only in the 5th s-LNand LNneurons (). In the other neurons, the PDP1 oscillations were very weak or even absent already in the control flies. Especially, the l-LNneurons showed very low levels of PDP1 in both control andflies (), consistent with a previous report indicating that molecular oscillations in the l-LNneurons require light input via Cryptochrome (). The DN, DN, and DNneurons showed very weak PDP1 oscillations peaking at approximately CT22 in control flies, but they were almost arrhythmic inflies. Thus, these results suggest that CCHa1 signaling has an impact on molecular oscillations and particularly modulates the phase of the s-LNneurons. CCHa1 CCHa1 tim > CCHa1 RNAi tim > CCHa1 RNAi Figure 7A Figure 7B Figure 7B v d v v 1a 1p 2 v [Yoshii et al., 2015]

Since the small morning activity peak inmutants in LD is reminiscent ofmutants (), we investigated the effect ofon PDF levels. The s-LNterminals at the dorsal protocerebrum show circadian changes in PDF level and their axonal morphology, by which the time information appears to be conveyed from the s-LNneurons to other clock neurons in a time-dependent manner (;). Thecontrol flies showed significant differences between ZT2 and ZT11 in PDF staining intensity and the area of the axonal arborizations of the s-LNterminals (), in agreement with a previous study (). Whilenull mutants also showed a difference in PDF level, their PDF level at ZT11 was higher than that of control flies, leading to a lower amplitude of PDF cycling in themutants (). The same tendency was also observed in the area of the s-LNaxonal arborizations (). At both time-points, the arborization pattern of the s-LNterminals inmutants was more extensive than that in control flies. Therefore, the temporal morphological changes in the axonal arborizations were less clear in the mutants. This indicates an effect of CCHa1 on the PDF-dependent output from the s-LNneurons. CCHa1 Pdf 1 CCHa1 w CCHa1 CCHa1 CCHa1 [Renn et al., 1999] [Park et al., 2000] [Fernandez et al., 2008] [Fernandez et al., 2008] v v v v v v Figure 8 Figure 8B Figure 8C

Central circadian clocks in animals are composed of multiple pacemaker clock neurons that utilize different “neuromessengers” (neuropeptides and classical neurotransmitters) () (;;). Supposedly, these pacemaker neurons communicate with each other to generate coherent and robust circadian rhythms. However, little is known as to how these clock neurons are interconnected, even in well-studied model organisms such as the mouse and fruit fly. Here, we show that the DNand s-LNneurons of the adult fly are closely connected via neurites, and that they communicate with each other. We describe CCHa1 as novel circadian neuropeptide that is expressed in the DNand signals the s-LNneurons. So far, it was known that the DNneurons express IPNamide and glutamate (;;) whereas the s-LNneurons express PDF, short neuropeptide F (sNPF), and glycine (;;) (). Thus, together with CCHa1, both types of clock neurons utilize two neuropeptides and one classical transmitter—and they are so far the onlyclock neurons with three neuromessengers (). While it is unclear whether IPNamide and sNPF mediate signals within the clock network (at least sNPF appears to signal downstream neurons ()), glutamate, PDF, and glycine participate in communication between the s-LNand DNneurons (;;;). Glutamate signals originating from the DNneurons (and other DN neurons) are transmitted via inhibitory metabotropic glutamate receptors to the s-LNneurons, and reduce their cAMP and Calevels (;). Glycine also exerts inhibitory effects via ionotropic Clchannels but the signals are transmitted from the s-LNto the DNneurons, where it reduces spiking frequencies (). On the other hand, PDF increases cAMP levels in the DNneurons (). Similar to PDF, CCHa1 appears to exert excitatory effects, but in the opposite direction: it is secreted from the DNneurons and increases cAMP levels in the s-LNneurons. In summary, the DNand s-LNneurons appear to be mutually connected via excitatory and inhibitory neurotransmitters, respectively. Figure 9 Figure 9 Figure 9 [Muraro et al., 2013] [Beckwith and Ceriani, 2015] [Hermann-Luibl and Helfrich-Förster, 2015] [Shafer et al., 2006] [Hamasaka et al., 2007] [Collins et al., 2014] [Helfrich-Förster, 1995] [Johard et al., 2009] [Frenkel et al., 2017] [Selcho et al., 2017] [Hamasaka et al., 2007] [Shafer et al., 2008] [Collins et al., 2014] [Frenkel et al., 2017] [Hamasaka et al., 2007] [Collins et al., 2014] [Frenkel et al., 2017] [Shafer et al., 2008] 1a v 1a v 1a v v 1a 1a v v 1a 1a 1a v 1a v Drosophila 2+ -

Regarding CCHa1, we detected the expression of its receptor not only on the s-LN, but also on the l-LNneurons. Nevertheless, only the s-LNneurons responded to the application of CCHa1 with an increase in cAMP levels. This does not imply that CCHa1 does not exert any effect on the l-LNneurons. There are several possibilities that can explain our results. Either the l-LNneurons are less responsive to CCHa1 and would need a higher concentration to respond, or activation of the CCHa1 receptor induces a different signaling cascade that leads to an increase in Calevels instead of cAMP. Since we used a rather high concentration of CCHa1 (10M), we believe that the second possibility is more likely. For example, PDF signals not only through cAMP, but also weakly through Ca(). In the cockroach, PDF increases cAMP levels in some clock neurons and Calevels in others (;). Thus, it will be interesting to investigate whether the l-LNneurons respond to CCHa1 with an increase in Calevels (see also the paragraph “comparison with the mammalian system”). v v v v v v 2+ -5 2+ 2+ 2+ [Mertens et al., 2005] [Wei et al., 2014] [Gestrich et al., 2018]

The level of CCHa1 in the DNneurons is low in the evening and high in the night-morning (). Similarly, a recent study reported thatmRNA expression in the LNneurons is high in late night-morning and low in the evening (), suggesting that the s-LNneurons receive CCHa1 signals in a circadian manner. Perhaps most of the CCHa1 is secreted in the morning, a time at which the s-LNneurons are thought to secrete PDF (;). Thus, the DNneurons also receive PDF signals in the morning, implying that the DNand s-LNneurons are reciprocally and temporally coupled. This might explain why morning activity is not only strongly reduced when the s-LNlack PDF, but also when CCHa1 signaling is absent (). 1a v v v 1a 1a v v Figure 2 Figure 5 CCHa1-R [Abruzzi et al., 2017] [Collins et al., 2014] [Liang et al., 2017]

The CCHa1 signals from the DNneurons seem to be important for a normal rhythmic release of PDF, as can be judged from the flattened amplitude of PDF cycling and daily s-LNterminal remodeling inmutants (). As the latter finding is based on the observation of PDF immunostaining but not GFP staining, as in previous studies (;;), we might have overlooked smaller morphological changes. Nevertheless, the high PDF staining inmutants revealed a complex arborization pattern of the s-LNterminals at the two tested time-points. Furthermore, it has been shown that the level of PDF is responsible for the daily remodeling of the s-LNterminals (). Thus, CCHa1 signaling may modulate the level of PDF, resulting in suppression of the daily remodeling of the s-LNterminals inmutants. It is not clear whether the high level of PDF in the s-LNterminals indicates a high PDF release throughout day and night, but it is imaginable that a permanent PDF release exerts similar negative effects on the normal expression of morning activity as does an excessively low release of PDF. Additionally, the changed modulation of PDF signaling by CCHa1 would indirectly affect the other clock neurons that express the PDF receptor. Indeed the PDP1 oscillations inknockdown flies were changed: under LD, they showed a high-amplitude PDP1 cycling in all clock neurons, and under DD, PDP1 cycling was enhanced in the 5th s-LNand LNneurons, but dampened in the remaining clock neurons (). An earlier study showed that the amplitude of PDP1 cycling is an indicator of the siesta length in LD (). Mutants that lacked Cryptochrome (mutants) exhibited high-amplitude PDP1 cycling and a long siesta, identical to our findings in themutants andknockdown flies (). In DD, the situation is apparently more complicated since only the LNs showed high-amplitude cycling. Nevertheless, with the exception of the absent morning activity, free-running locomotor activity rhythms were astonishingly normal in these flies. This largely contrasts with the weak rhythmicity of themutants. 1a v v v v v v d CCHa1 SK8 CCHa1 CCHa1 CCHa1 cry 1 CCHa1 CCHa1 Pdf 1 Figure 8 Figure 7 Figure 7 [Fernandez et al., 2008] [Depetris-Chauvin et al., 2014] [Petsakou et al., 2015] [Depetris-Chauvin et al., 2014] [Kistenpfennig et al., 2018]

As already discussed above, PDF and CCHa1, the excitatory neuromessengers of the s-LNand DNneurons, respectively, are necessary for a pronounced morning activity of the flies. In contrast, the two inhibitory neuromessengers of these clock neurons, glutamate and glycine, are not needed for morning activity (;;). Evidently, glutamate and CCHa1 from the DNneurons affect activity differently. This also holds true for the activity level. Whereas the knockdown of the glutamate receptor increases general activity levels (), we found here that the knockdown ofdecreases activity levels. Thus, glutamate inhibits activity, whereas CCHa1 promotes activity, most likely by affecting PDF release. Moreover, CCHa1 does not only promote activity and consequently shorten the siesta; it also changes the phase of activity. Under LD, theknockdown advances the end of morning activity, whereas under DD conditions, it delays the median of activity. These phasic changes cannot only be detected in the activity pattern, but are even more prominent in the PDP1 oscillations in the s-LNneurons. The knockdown of the glutamate receptor again exerts different effects on the phase of activity than the knockdown of: it delays the end of evening activity under LD and DD conditions and consequently delays the onset of night sleep (;). In these studies, the molecular oscillations in the different neurons have not been determined. Therefore, we cannot judge their phase. Nevertheless, in summary, both neuromessengers may cooperate in adjusting the phase of the s-LNneurons, and consequently, the phase of the activity rhythm. v 1a 1a v v [Hamasaka et al., 2007] [Collins et al., 2014] [Frenkel et al., 2017] [Collins et al., 2014] [Hamasaka et al., 2007] [Collins et al., 2014] CCHa1 CCHa1 CCHa1

Interestingly, CCHa1 appears to exert only minor effects on the free-running period of the activity rhythm under DD. This is different for glutamate and glycine, which seem to shorten the period (;), and for PDF, which lengthens the period (). We found that theknockdown shortens the period, implying that CCHa1 might exert a lengthening effect on period, similar to PDF. However, the period of themutant was not significantly shorter than that of the controls, suggesting that the effects of CCHa1 on the period are rather small and depend on the genetic background. [Hamasaka et al., 2007] [Frenkel et al., 2017] [Renn et al., 1999] CCHa1 CCHa1

The mammalian SCN also uses several neurotransmitters for intercellular coupling, among which VIP and its receptor, VPAC2R, are proposed to have functional homology with thePDF/PDF-R ().transgenic mice, which lack VPAC2R, show arrhythmic or a shorter free-running rhythm in behavior under constant conditions, which is a phenocopy of theormutants (). The mammalian homolog ofis the receptor of GRP, bombesin 2 [Flybase, the DRSC Integrative Ortholog Prediction Tool ()], although there is no homolog of. The GRP-receptor deficient mice show normal free-running activity rhythms in DD but attenuated response to a light-pulse (). The activated GRP-receptor is coupled to Gq protein, which activates phospholipase C and increases Calevels but does not directly elevate cAMP levels (). Here, we found an increase in cAMP levels in the s-LN, indicating that CCHa1-R couples to a different G protein (). However, we cannot exclude the possibility that CCHa1-R in the l-LNfunctions via a phospholipase C-activating Gq protein that increases Calevels, perhaps in cooperation with light (see also discussion above). If true, this may explain why the l-LNdo not respond to CCHa1 application with an increase of cAMP levels, although they express CCHa1-R (). If the l-LNneurons need light and CCHa1 in order to respond—and, in fact, they are light-responsive (;)—this might also explain why the flies respond differently toknockdown in LD and DD. In LD, they phase advance activity, perhaps because the l-LNand s-LNneurons are affected by the absence of CCHa1 and the l-LNsignal on the s-LNneurons, as has been shown earlier (). In DD, the flies phase delay activity, perhaps because only the s-LNare affected. Certainly, this hypothesis is highly speculative and needs to be verified by future studies. Drosophila Vipr -/- Pdf Pdf-r Drosophila CCHa1-R CCHa1 CCHa1 [Mertens et al., 2005] [Aton et al., 2005] [Hu et al., 2011] [Aida et al., 2002] [Roesler and Schwartsmann, 2012] [Sheeba et al., 2008] [Fogle et al., 2011] [Yoshii et al., 2009b] 2+ 2+ v v v v v v v v v Figure 4 Figure 3

Recently,showed that the calcitonin receptor in mammals and itshomolog, DH31-receptor, share a similar function: controlling circadian body temperature rhythms. Thus, neuropeptide ligands responsible for circadian circuits are not conserved but their G-protein coupled receptors are well-conserved across animal clocks. [Goda et al. (2018)] Drosophila