Characterization of clock‐related proteins and neuropeptides in Drosophila littoralis and their putative role in diapause

Insects from high latitudes spend the winter in a state of overwintering diapause, which is characterized by arrested reproduction, reduced food intake and metabolism, and increased life span. The main trigger to enter diapause is the decreasing day length in summer–autumn. It is thus assumed that the circadian clock acts as an internal sensor for measuring photoperiod and orchestrates appropriate seasonal changes in physiology and metabolism through various neurohormones. However, little is known about the neuronal organization of the circadian clock network and the neurosecretory system that controls diapause in high‐latitude insects. We addressed this here by mapping the expression of clock proteins and neuropeptides/neurohormones in the high‐latitude fly Drosophila littoralis. We found that the principal organization of both systems is similar to that in Drosophila melanogaster, but with some striking differences in neuropeptide expression levels and patterns. The small ventrolateral clock neurons that express pigment‐dispersing factor (PDF) and short neuropeptide F (sNPF) and are most important for robust circadian rhythmicity in D. melanogaster virtually lack PDF and sNPF expression in D. littoralis. In contrast, dorsolateral clock neurons that express ion transport peptide in D. melanogaster additionally express allatostatin‐C and appear suited to transfer day‐length information to the neurosecretory system of D. littoralis. The lateral neurosecretory cells of D. littoralis contain more neuropeptides than D. melanogaster. Among them, the cells that coexpress corazonin, PDF, and diuretic hormone 44 appear most suited to control diapause. Our work sets the stage to investigate the roles of these diverse neuropeptides in regulating insect diapause.

F I G U R E 1 Phylogeny of the virilis group, and neuropeptide expression in clock-and diapause-related cells in D. melanogaster brain. (a) Simplified phylogenetic tree demonstrating the relationship of the virilis group to D. melanogaster after Yusuf et al. (2022). (b) Expression of selected neuropeptides in the brain of D. melanogaster. Neuropeptides of the clock neurons are shown on the left and neuropeptides of the neurohormonal center (PI-pars intercerebralis; PL-pars lateralis) in the right brain hemisphere. As schematically indicated, most clock neurons get light input from the eyes via the accessory medulla (AME, curved yellow arrow) and project to the dorsal protocerebrum, where some are connected to the PL and PI (clock neurons: s-LN v -small ventrolateral neurons; l-LN v -large ventrolateral neurons; 5th LN-fifth lateral neuron; LN d -dorsolateral neurons; LPN-lateral posterior neurons; DN 1 , DN 2 , DN 3 -dorsal neurons). The neurosecretory PI and PL cells project to the corpora cardiaca/allata (CC/CA) complex. The brain is shown from a posterior view and all fibers running in the median bundle (MBL) of the anterior brain are stippled. The Drosophila insulin-like peptide 2 (DILP2)-and diuretic hormone 44 (DH44)-expressing cells of the PI run via the MBL to the CC and CA, while the corazonin (CRZ)-expressing cells in the PL innervate only the CC and do so via the MBL and a lateral tract (black arrow). Among the CRZ cells, one large CRZ CN cell and six smaller CRZ 1-6 cells can be distinguished. The CRZ 1-6 cells show additional arborizations in the dorsolateral protocerebrum (light green arrow), while the large CRZ CN cell runs only to the CC. The ion transport peptide (ITP) cells in the PL join the lateral tract of the CRZ neurons and like the DILP2 neurons, they innervate the CC and CA. All CRZ and ITP neurosecretory cells additionally express small neuropeptide (Continues) F I G U R E 1 (Continued) F (sNPF). The diuretic hormone 31 (DH31) cells are also joining this tract but innervate the CA only. stress resistance, and prolonged life span (Hahn & Denlinger, 2011). In most of the virilis group species, day length (also called photoperiod) is the main trigger of diapause, while temperature is less important (Aspi et al., 1993;Lankinen, 1986b;Lumme & Keränen, 1978;Vaze & Helfrich-Förster, 2016;Vesala & Hoikkala, 2011). Therefore, their diapause is also defined as photoperiodic diapause. As such, some species from this group enter adult diapause in late August, when day length is reduced and can maintain this state until May of the following year in natural conditions (Lumme et al., 1974). In comparison, flies living at lower latitudes, such as Drosophila melanogaster, enter a state of reproductive dormancy in response to a combination of adverse conditions such as low temperatures combined with food shortage and short photoperiods (Kubrak et al., 2014;Nagy et al., 2018Nagy et al., , 2019Ojima et al., 2018;Saunders et al., 1989). Further, dormancy in these species is not strongly dependent on day length and can be terminated as soon as the environmental conditions improve (Koštál, 2006;Saunders et al., 1989).
The timing of photoperiodic diapause of the virilis group requires an internal timing system. As hypothesized by Bünning (1936), the circadian clock is well suited for this role. Indeed, the importance of the circadian clock for photoperiodic responses has been demonstrated for several insects living at high latitudes (reviewed in Denlinger, 2022;Goto, 2022;Saunders, 2020), but such evidence is so far lacking for Drosophilids. One reason for this lack of knowledge is the fact that most virilis group species possess weak circadian clocks that are unable to maintain rhythmicity under constant darkness (Bahn et al., 2009;Beauchamp et al., 2018;Bertolini et al., 2019;Kauranen et al., 2012;Lankinen, 1986aLankinen, , 1986bMenegazzi et al., 2017). In all Drosophila species studied so far, the circadian clock in the brain consists of lateral and dorsal clock neurons (LNs and DNs; Figure 1b). In both LNs and DNs, the molecular clock keeps time via at least two feedback loops, which are best described in D. melanogaster (reviewed in Helfrich-Förster et al., 2017). In the first loop, the clock proteins PERIOD (PER) and TIMELESS (TIM) block their own transcription by inhibiting their transcriptional activators CLOCK (CLK) and CYCLE (CYC), while in the second loop the clock proteins VRILLE (VRI) and the PAR Domain Protein 1ε (PDP1ε) lead to a rhythmic transcription of the clock gene.
Another important component of the clock is the light-sensitive CRYP-TOCHROME (CRY), which is expressed in about half of the LNs and DNs of all Drosophila species investigated so far (Hermann et al., 2013).
CRY has an important role in transmitting light information directly to the molecular clock network, and in addition to that, most of the clock neurons get light input from the eyes via the accessory medulla (Figure 1b;Helfrich-Förster, 2020;Li et al., 2018). The clock neurons predominantly use neuropeptides for neural communication (Table 1) and once more, their expression is best described in D. melanogaster (Figure 1b;Ma et al., 2021;Reinhard, Schubert, et al., 2022). Among the neuropeptides expressed TA B L E 1 Selected neuropeptides in circadian clock neurons.
In all Drosophila species studied, the CRY-expressing circadian clock neurons project to the dorsal protocerebrum (Hermann et al., 2013), where they may signal to the neuroendocrine system in the pars intercerebralis (PI) and pars lateralis (PL) (indicated as a single curved gray arrow in Figure 1b). The PI and PL of D. melanogaster comprise several neurosecretory cells, which produce multiple neuropeptides ( Figure 1c; Table 2) and regulate various aspects of dormancy/diapause (Nässel & Zandawala, 2020). The PL consists of three sets of neurosecretory cells that produce (1) DH31, (2) ITP and sNPF, or (3) corazonin (CRZ) and sNPF. The PI, on the other hand, contains several sets of neurosecretory cells, including those producing Drosophila insulin-like peptides (DILPs) or diuretic hormone 44 (DH44) (Figure 1b). These neurosecretory cells project to the corpora cardiaca (CC), corpora allata (CA), or the anterior gut and release neuropeptides into the hemolymph ( Figure 1b; Table 2). The secreted neuropeptides regulate different aspects of diapause/dormancy such as prolongation of life span, feeding, fecundity, egg production/laying, metabolism, osmotic homeostasis, and stress resistance (Table 2). Recent work has also iden-  (Meiselman et al., 2022). Lastly, multiple clock neurons provide inputs to DH31-positive neurosecretory cells, which in turn regulate reproductive dormancy by modulating juvenile hormone (JH) biosynthesis in the CA (Kurogi et al., 2023). Hence, most of our current knowledge on the link between the circadian clock and photoperiodic responses stems from work done in D. melanogaster despite them exhibiting a shallow diapause that is highly variable and often does not involve the entire population (Saunders et al., 1989). Therefore, it is imperative to study this phenomenon in species, such as those from the virilis group, that exhibit more pronounced overwintering responses.
To close this gap, here we characterized the circadian clock network and part of the neurohormonal system that might control diapause in the virilis group species Drosophila littoralis. Drosophila littoralis is a wellinvestigated model for diapause timing and photoperiodic responses (Lankinen, 1986a(Lankinen, , 1986bLankinen & Forsman, 2006;Lumme & Oikarinen, 1977;Lumme et al., 1974;Salminen at al. et al., 2015). We mapped the distribution of various neuropeptides in D. littoralis that were shown to be expressed in either the clock neurons or neurosecretory cells of D. melanogaster. Among these peptides, we laid a special focus on CRZ, because it belongs to the superfamily of gonadotropin-releasing hormones that control reproductive maturation in vertebrates (Andreatta et al., 2020;Zandawala et al., 2018) and because CRZ was shown to be involved in diapause and reproduction control of different insects (Gospocic et al., 2017;Shiga et al., 2003;Tsuchiya et al., 2021). Our results reveal that the overall anatomy of the circadian clock network and the neurohormonal center in the PI/PL is well conserved between D. melanogaster and D. littoralis. TA B L E 2 Selected neuropeptides in neurosecretory cells of the pars intercerebralis (PI) and pars lateralis (PL).  Bergland et al., 2012;Choi et al., 2005;Gospocic et al., 2017;Kapan et al., 2012;Kubrak et al., 2016;Nässel et al., 2013;Shiga et al., 2003;Tsuchiya et al., 2021;Zandawala et al., 2021 Ion transport peptide (ITP) 4 PL cells Prevents the fly from water loss by promoting thirst and repressing excretion; suppresses feeding and regulates water and energy balance, which is most important during desiccating winter conditions during which the flies are in dormancy Dircksen, 2009;Gáliková et al., 2018 Diuretic hormone 31 (DH31)

Expression in PI/PL (per hemisphere) Function of neuropeptide/ neurosecretory cell References
3 PL cells Works antagonistically to DILP2 by repressing juvenile hormone biosynthesis; inhibits vitellogenesis leading to reproductive dormancy Kurogi et al., 2022 However, there are clear differences in neuropeptide expression between corresponding neurons in these two species, which could explain the differences in their dormancy response.

Diapause induction
To induce diapause in D. littoralis, the flies were transferred to 10 • C and LD 08:16. For comparing PDF and CRZ staining intensity in reproductive and diapausing flies, a group of 60 female virgin flies was collected and entrained either to diapause-inducing conditions (LD 08:16, 10 • C) or reproductive conditions (LD 20:04, 23 • C). To assess only the effect of the photoperiod, we exposed another group of flies to diapausing photoperiods (LD 08:16) or reproductive photoperiods (LD 20:04) at the same temperature of 16 • C. After 3 weeks of entrainment, flies were collected at ZT1 to evaluate the diapausing state. Each fly was singly dissected in phosphate-buffered saline (PBS) to verify the status of the ovaries and fat bodies. The flies were considered as diapausing when no oocyte at stage 14 was present (Kurogi et al., 2023). This state can also be defined as stage II ovaries following the description of Lankinen et al. (2022). The brains of the reproductive and diapausing flies were used for fluorescent immunocytochemistry and quantitative polymerase chain reaction (qPCR) analysis as described in Sections 2.5 and 2.7, respectively.

In silico identification of neuropeptide and clock genes
We identified the neuropeptide and clock genes of interest in the D. littoralis genome  using the following procedure.
First, we obtained the amino acid sequences of these proteins in D. melanogaster using FlyBase (v FB2022_05; Gramates et al., 2022) and used them to perform protein BLAST (Altschul et al., 1990) against the D. virilis reference sequence protein database. This step was necessary since D. littoralis protein sequences have not yet been predicted. Next, we used D. virilis protein sequences as a query to mine the D. littoralis whole-genome shotgun database via tBLASTn . This search was more accurate than using D. melanogaster sequences as a query since D. littoralis is more closely related to D. virilis than D.

Primary antibodies
All primary antibodies, their source, and the used dilutions are listed in Table 3. Except for the newly generated ITP-antibody (see Section 2.5), all primary antibodies are well characterized and have been extensively used in different insects including D. melanogaster and other fly species.
The amino acid sequences they are directed against are indicated in Table 3.
To investigate the degree to which the amino acid sequences of the neuropeptides AstC, CRZ, DH31, DH44, PDF, ITP, sNPF, and DILPs and the clock proteins PER and CRY1 are conserved in D. littoralis, we identified them in silico as described in Section 2.2 and compared the sequences with those of D. melanogaster and D. virilis. As shown in Figure 3, the sequences of most neuropeptides turned out to be highly conserved if not identical.
To assess the specificity of the antibodies that we used, we searched the genome of D. littoralis using tblastn to find sequences other than the addressed neuropeptides that could be recognized by the antibodies used. We blasted both the antigen and whole-neuropeptide sequences using algorithm parameters of low stringency: a word size of two and an expected threshold of one. Even with these parameters, we were not able to find any further hits in the genome. Only the neuropeptide sequences against which the antibodies had been generated were identified. This suggests that our antibodies are specific.
Regarding the specificity of the anti-PDH antibody, a recent study by Veenstra (2021) showed that this antibody cross-reacts with calcitonin-A in Locusta migratoria. To rule out this possibility in D.
littoralis, we searched the D. littoralis genome for calcitonin-A-and calcitonin-B-like sequences. Our genome searches failed to identify calcitonin-like sequences in D. littoralis. This suggests that our anti-PDF immunostaining represents authentic PDF expression.

Generation and characterization of an antibody against ITP
The antibody against ITP was generated in guinea pig by Scrum Inc.
The part of the C-terminal of ITP, EMDKYNEWRDTL, was chemically synthesized and the N-terminal was coupled to keyhole limpet hemocyanin. Three guinea pigs were immunized using a conventional method. The specificity of the three different sera was tested on brains of D. melanogaster by immunocytochemistry as described in Section 2.6. All immunostaining with the three different sera showed exactly the same staining pattern as reported in previous studies (Hermann-Luibl et al., 2014). Finally, we selected the serum with the lowest background signals for further analysis. We tested the specificity of the antibody by costaining the brain with anti-ITP rabbit (Table 3)

Fluorescent immunocytochemistry
Immunocytochemistry was performed on young female and male flies according to the protocol of Schubert et al. (2018). All images are based on brains from female flies, although we found no differences between the two sexes. If not otherwise stated, flies were collected 3-10 days posteclosion, fixed for 3.5 h (3 h for D. melanogaster) in 4% paraformaldehyde dissolved in PBS containing 0.5% Triton ×100 (PBST), and dissected after rinsing three times for 10 min each in PBS.
Brains were blocked in 5% normal goat serum in PBST (normal donkey serum in case of anti-PER-goat) overnight at 4 • C. The brains were then incubated in the primary antibody mix for 2 days at 4 • C and an additional day at 20 • C. After washing them in PBST, the brains were incubated in the secondary antibody mix (  The images were processed using Fiji (v1.53c; Schindelin et al., 2012). Only changes in brightness and contrast were performed on the images. At least 10 brains were scanned for each staining performed.

Quantification of immunostaining intensity
For CRZ and PDF immunostaining in the PL, staining intensity and size of the neurosecretory cell bodies were quantitatively compared between sister groups of D. littoralis female flies entrained to reproductive or diapausing conditions (as described in Section 2.2). The entrainment and immunostaining of the two groups were conducted simultaneously and confocal pictures were acquired maintaining the same laser intensity and general settings. The raw images were quantified in Fiji as follows: For each cell body, the focal plane showing the largest diameter was determined, and the cell body outlines were manually defined as region of interest (ROI) using the polygon selection tool. The cell area and its mean pixel intensity were measured. The mean pixel intensities were corrected by the mean background intensity that was measured in the closest nonstained area to the neurons using an ROI with a comparable size to the cell bodies.

quantitative PCR
The relative transcript level of Crz, Pdf, Kr-h1 (Kruppel homolog 1), and Dilp2 in reproductive and diapausing D. littoralis females was assessed by qPCR. Each gene sequence was manually retrieved as described in Section 2.3. The used primers are listed in Table 5

Gene
Sequence 5t'to 3t' Primer Melting Temperature (Tm) little-conserved [the darker, the more conserved]; white: not conserved), and the consensus sequences are shown below in bold letters. Regions that correspond to the antigen sequences used to generate the antibodies are circled in green in the D. melanogaster sequences. Amino acids that do not match the D. melanogaster reference are highlighted in pink.

Statistical analysis and figures
The statistical analysis and plot designing were performed with R snapgene.com]).

Sequence similarities of clock proteins and neuropeptides between D. melanogaster and D. littoralis
To assess the sequence similarity between the D. littoralis and D.
melanogaster clock proteins and neuropeptides, we performed an in silico analysis using the NCBI Reference Sequences indicated in Section 2. Subsequently, we compared the predicted D. littoralis neuropeptide and clock protein sequences with those of D. virilis and D.
melanogaster. As expected from the phylogeny (Figure 1), the identity percentage was always highest between D. littoralis and D. virilis (Table 6; Figure 3).
Although PER is a key component of the molecular clock, it was found to be not well conserved between D. melanogaster and the virilis group. However, its N-terminal domain was highly conserved across species. Since this is the sequence against which the D. melanogaster PER antibody was generated, it allowed us to use this antibody in D. lit-TA B L E 6 Percent identity of different proteins and mature neuropeptides.

PER and CRY expression in the clock neurons of D. littoralis show a rather similar pattern to D. melanogaster, but PDF expression differs
To date, the clock neurons of the virilis group have been identified with antibodies against PDP1 and CRY (Hermann et al., 2013) as the commonly used Drosophila PER antibody (Stanewsky et al., 1997) did not recognize the virilis group PER (Hermann et al., 2013). Here, we used a commercially available PER antibody directed against the wellconserved N-terminus of D. melanogaster PER (Table 3). This antibody labeled putative PER-expressing cells in D. littoralis, allowing us to detect the different clock neuron clusters ( Figure 4). As previously reported by Hermann et al. (2013), the virilis group possesses most, if not all, of the clock neurons described in D. melanogaster (Figure 4a-d).

F I G U R E 6
Pigment-dispersing factor (PDF) and short neuropeptide F (sNPF) immunostaining in the brain of Drosophila littoralis. (a) Z stack of 37 confocal planes. sNPF-like immunoreactivity is widely present in the dorsal brain. sNPF colocalizes with PDF in the six smaller CRZ 1-6 neurons (white arrow heads) but not in the large CRZ CN neuron (yellow arrow) and in the tract running to the pars intercerebralis (PI) and the corpora cardiaca/allata (CC/CA) complex (yellow arrowheads). Here, we identified the same clusters, and we could subdivide the dorsal neurons (DN 1 s) into posterior and anterior ones, the DN 1p s and DN 1a s (Figure 4c). However, we could not detect PER in the lateral group of large ventrolateral neurons (l-LN v s) (Figure 4b). The PER staining in the s-LN v s was also quite weak compared to the other clock clusters (Figure 4a,b). Furthermore, we could not detect PER staining in the dorsal neurons 2 (DN 2 s) (Figure 4a-d). Since all these clock neurons have been detected by the PDP1 antibody in other virilis species (Hermann et al., 2013), we assume that they are also present in D. littoralis. However, they might contain either no or very little PER, below the detection limit of the antibody. This could mean that the molecular clock in these clock neurons is very weak, which can affect circadian rhythms. In particular, the s-LN v s are very important for maintaining circadian rhythmicity in constant darkness (Grima et al., 2004;Helfrich-Förster, 1998;Shafer & Taghert, 2009;Stoleru et al., 2004), and a weak molecular clock in them may lead to weak rhythmic F I G U R E 7 Allatostatin-C (AstC), PERIOD (PER), pigment-dispersing factor (PDF), and ion transport peptide (ITP) localization in Drosophila littoralis brain.   (Klarsfeld et al., 2004) and in adults kept for several days under constant conditions (Yoshii et al., 2009). Perhaps this is the case in D. littoralis already under LD cycles. Additional stainings at timepoints other than ZT1 will clarify this hypothesis in the future.
CRY was found in half of the clock neurons of D. melanogaster (Benito et al., 2008;Yoshii et al., 2008). Similarly, we detected CRY-like immunoreactivity in four s-LN v s, the 5th LN, three LN d s, six DN 1p s, and two DN 1a s in D. littoralis (Figure 4e), but not in the l-LN v s (Figure 4f

CRZ colocalizes with PDF in neurosecretory cells of the PL
As CRZ is involved in the diapause of different insects as well as in their stress responses (Shiga et al., 2003;Tsuchiya et al., 2021;Zandawala et al., 2021), we investigated its expression in the brain of D.
littoralis. Similar to the expression pattern of D. melanogaster, we identified CRZ in seven "dorsolateral peptidergic neurons" per hemisphere in the PL that project into the PI and further, via median and lateral nerves, to the CC/CA complex (Figures 1b and 5a). While the six CRZ neurons had smaller somata, the seventh CRZ neuron had a larger cell body, which in D. melanogaster has been renamed CN neuron (Oh et al., 2019). We will maintain this nomenclature for D. littoralis and, henceforth, call the large CRZ neuron CRZ CN and the six smaller ones CRZ 1-6 neurons.
Surprisingly, double-labeling experiments with CRZ and PDF C7 antibodies (Table 3) revealed that the PDF neurons in the dorsal brain described above are in fact CRZ neurosecretory cells (Figure 5a). A similar colocalization of PDF in the CRZ cells of the PL was previously shown in the blow fly Protophormia terraenovae (Hamanaka et al., 2007), which is also strongly photoperiodic. Colocalization of CRZ and PDF was found not only in the somata of the neurons, but also in their central brain branches and their projections to the CC/CA complex (Figure 5a for D. littoralis). As the CC/CA complex is a neurohemal organ, PDF could be released together with CRZ into the hemolymph and target peripheral tissues.
Double labeling with CRZ and PDF antibodies allowed us to distinguish between the PDF fibers originating from the CRZ neurons and the faint PDF fibers stemming from the s-LN v s. This enabled us to reconstruct the s-LN v projections to the dorsal brain that terminate near the dorsal clock neurons (Figure 5b). In contrast to D. melanogaster, the s-LN v projections to the dorsal brain in D. littoralis show more pronounced fibers forming a "dorsal horn" (Figure 5b,c). A similar projection pattern of the s-LN v terminals was found in Drosophila yakuba, Drosophila willistoni, and Chymomyza costata (Bertolini et al., 2019;Hermann et al., 2013). In all these species, the s-LN v axons terminate in the PL close to the CRZ neurons, while those of D. melanogaster extend medially and terminate closer to the DILP neurons in the PI and even signal to them (Nagy et al., 2019). This observation suggests that the s-LN v s of D. littoralis are most likely not synaptically connected to the neurosecretory cells of the PI, but rather to those of the PL, notwithstanding the fact that they could also signal paracrine without forming any synapses (Shafer et al., 2022).

sNPF shows a different expression pattern in the clock neurons and the neurosecretory cells of D. littoralis compared to D. melanogaster
In D. melanogaster, the neuropeptide sNPF is widely expressed in the brain  and appears to be involved in the circadian clock as well as in the control of dormancy (Chen et al., 2013;Johard et al., 2009;Nagy et al., 2019). It is expressed in all four s-LN v s, as well as in all four ITP cells (Kahsai et al., 2010) (Oh et al., 2019). Assuming that D. littoralis CRZ CN neuron is also glucose sensing, the function of sNPF is probably served by other peptides expressed in these cells (i.e., CRZ and/or PDF).
Furthermore, the absence of sNPF from the s-LN v s of D. littoralis suggests that it does not transfer time information from the clock to the DILP neurons in the PI as found for D. melanogaster (Nagy et al., 2019).
Whether sNPF is involved in the control of diapause in D. littoralis needs to be investigated in future studies.

AstC colocalizes with ITP in two clock neurons of D. littoralis
Similar to sNPF, the neuropeptides AstC and ITP are expressed in several D. melanogaster clock neurons (Diaz et al., 2019;Hermann-Luibl et al., 2014;Johard et al., 2009;Zhang et al., 2021). AstC was described in the lateral posterior neurons (LPNs) and some DN 1p s and DN 3 s of D. melanogaster, while ITP expression was shown in the 5th LN and one CRY-positive LN d (Figure 1b).
Notably, AstC was recently shown to be crucial for temperature sensing during the induction of dormancy in D. melanogaster (Meiselman et al., 2022). Therefore, we investigated its presence in the brain of D. littoralis (Figure 7). We found a rather dense network of AstC-positive neurites in the central brain of D. littoralis including the protocerebral bridge and the subesophageal ganglion, with the highest density laterally to the PI (Figure 7a, white arrow). The latter resembles the AstC staining pattern in D. melanogaster (Zhang et al., 2021) and might partly originate from the LPNs . Indeed, we found AstC staining in the three LPNs of D. littoralis In addition, ITP expression is also well conserved between D. melanogaster and D. littoralis in the four neurosecretory cells in the PL that project to the CC/CA complex (Figures 1b and 7g ) (Hermann et al., 2013;Kahsai et al., 2010).
In summary, we did not find differences in ITP expression between D. melanogaster and D. littoralis, but a different expression pattern of AstC within the clock neurons. While AstC is absent from the DN 1p s of D. littoralis, it is strongly expressed in the LPNs and is present in more DN 3 s as compared to D. melanogaster. Moreover, unlike in D.
melanogaster, AstC is also present in the ITP-positive 5th LN and LN d of D. littoralis. This suggests that AstC might play an important role in the circadian system of D. littoralis. Since AstC in D. melanogaster is involved in the photoperiodic entrainment (Diaz et al., 2019) and in the induction of dormancy in response to low temperatures (Meiselman et al., 2022), it is tempting to speculate that AstC conveys diapause-inducing photoperiodic information to the PI and PL neurosecretory cells of D. littoralis.

DH31 is present in some neurosecretory cells of the PL and DN 1p clock neurons of D. littoralis
The neuropeptide DH31 is present in the LPNs and in a few DN 1p s clock neurons of D. melanogaster (Goda et al., 2016;Kunst et al., 2014;. However, within the clock network of D. littoralis, we could only find it in the DN 1p s (Figure 8b) and LPNs.
DH31 was recently described in three neurosecretory cells in the PL that project to the CA ("DH31CA neurons"; Kurogi et al., 2023) (Figure 1b) and were linked to reproductive dormancy in D.
melanogaster (Kurogi et al., 2023). This prompted us to investigate DH31 immunostaining in the brain of D. littoralis. We found that DH31 is widely expressed in D. littoralis brain (Figure 8a,b), including the three neurosecretory cells in the PL that project to the CA (Figure 8c). This suggests that DH31 might play a similar role in the induction of diapause in D. littoralis as it was found for dormancy in D. melanogaster (Kurogi et al., 2023).

DH44 is present in two CRZ 1-6 neurons of D. littoralis
Although it is not expressed in any of the clock neurons, DH44 plays a role in stress tolerance and was shown to be involved in the locomotor rhythmicity of D. melanogaster (Cabrero et al., 2002;Cavanaugh et al., 2014;King et al., 2017;Zandawala et al., 2018). Instead, it is expressed in six PI neurons that receive inputs from the clock neurons (King et al., 2017). We used the antibody against DH44 to check whether this is similar in D. littoralis. As true for D. melanogaster, in D.
littoralis, DH44 was present in six cells of the PI (Figure 8d). DH44positive fibers were also present in the CC/CA complex (Figure 9d) indicating that DH44 could be released into the circulation as also expected for D. melanogaster (Cabrero et al., 2002;Ohhara et al., 2018).
The DH44-positive PI cells of D. melanogaster have also been shown to signal to the crop (Ohhara et al., 2018) where they are possibly involved in crop extension. The crop extension is often seen in dormant flies since they use it to store food during unfavorable conditions (Kubrak et al., 2014). The conserved expression pattern of DH44 suggests that these functions might be conserved in D. littoralis.
Besides the PI, DH44 signal was present in the protocerebral bridge and in two of the small CRZ 1-6 neurons ( Figure 8e). These two cells are in most cases the ones that are the closest to the CRZ CN neurons, suggesting that the CRZ 1-6 neurons could be further subclassified. Future studies are required to reveal the function of the different subclasses of CRZ neurons.
In summary, we found several striking differences in neuropeptide expression between D. littoralis and D. melanogaster that are summarized in Figure 9. Specifically, sNPF and AstC were present in more ; D. littoralis flies lack DH31 in the LPN and we did not test for AstA). However, AstC is absent from the posterior dorsal neurons (DN 1p s) in D. littoralis, in which DH31 appears to be more prominently expressed as compared to D. melanogaster. (b) Neuropeptides in the neurosecretory cells. In the pars intercerebralis (PI), we did not see any difference between the two species, but we could only stain for diuretic hormone 44 (DH44) since our Drosophila insulin-like peptide 2 (DILP2) antibody did not work in D. littoralis. The most striking differences between the two species occur in the corazonin (CRZ)-positive neurons (large CRZ CN and six smaller CRZ 1-6 ) of the pars lateralis (PL), which all coexpress PDF in D. littoralis, while they coexpress sNPF in D. melanogaster. The six smaller CRZ 1-6 of D. littoralis express sNPF as a third neuropeptide and two of these cells contain DH44 as a fourth peptide. the PL, more neuropeptides were coexpressed in D. littoralis than in D. melanogaster (Figure 9b). Perhaps the most striking difference between the two species was the coexpression of PDF in all CRZ-positive neurons of D. littoralis.

CRZ and PDF levels as well as the morphology of the CRZ CN neuron change in diapausing D. littoralis
The PL has already been shown to be important for diapause regulation in the highly photoperiodic blow fly Protophormia terraenovae (Shiga & Numata, 2000), and the major neuropeptides in the PL of D.
littoralis, PDF and CRZ, are involved in diapause or reproduction in various insects (Gospocic et al., 2017;Kotwica-Rolinska et al., 2022;Nagy et al., 2019;Shiga et al., 2003;Tsuchiya et al., 2021). Therefore, we hypothesized that the CRZ CN and CRZ 1-6 neurons are important for diapause induction in D. littoralis and that we should see differences in the CRZ/PDF staining intensity between reproducing and diapausing flies.
To test our hypothesis, we measured CRZ/PDF immunostaining intensity in the somata of CRZ CN and CRZ 1-6 neurons and the size of the somata in diapausing and reproductive flies ( Figure 10). To induce diapause, virgin females were kept under LD 08:16 at 10 • C for 3 weeks, while the reproductive controls were maintained at LD 20:04 at 23 • C.
To assess the effect of photoperiod alone, CRZ intensity was further analyzed in diapausing (LD 08:16) and reproductive (LD 20:04) flies exposed to the same temperature of 16 • C (Figure 10f). We found the size of the CRZ CN soma to be very different in diapausing and reproductive flies. In diapausing flies, it had almost the same size as the small CRZ 1-6 neurons, while under reproductive conditions, the CRZ CN neuron was twice as big compared to the CRZ 1-6 neurons and had an irregular shape (Figure 10a-c). Because of these differences, the quantification of CRZ/PDF immunostaining was performed separately in CRZ CN and CRZ 1-6 neurons ( Figure 10a). Under diapausing conditions, the CRZ staining intensity was significantly higher in both neuronal types than under reproductive conditions, with the highest CRZ levels occurring in the CRZ CN neuron under diapause (Figure 10d). This was true as well for the flies entrained under the same temperature but in different photoperiods (Figure 10f). In contrast, PDF staining intensity was significantly lower in diapausing flies as compared to reproductive ones and this was true for CRZ CN and CRZ 1-6 neurons ( Figure 10d).
In addition, the size of the CRZ CN and CRZ 1-6 somata was significantly smaller under diapausing than under reproductive conditions with a higher significance in the CRZ CN neuron (Figure 10a-c). For comparison, we measured also the PDF staining intensity in the somata of l-LN v s under the two conditions but did not see any differences ( Figure 10e).
The observed differences in staining intensity could be due to a changed neuropeptide production (transcription/ translation) and/or a changed neuropeptide transport/release. In case of CRZ, the higher staining intensity under diapausing conditions might also be due to the small size of the neuron somata, because then the same amount of CRZ would be concentrated in the smaller volume of the cells. To investigate whether the differences in staining intensity were due to changes in neuropeptide production, we quantified Crz and Pdf transcripts under diapausing and reproductive conditions. As further controls, we checked the transcript levels of Dilp2 and Kr-h1. Kr-h1 transcription is controlled by the JH and, as such, serves as an indicator of JH activity, which promotes reproduction and inhibits diapause (Meiselman et al., 2017;Minakuchi et al., 2009;Zhang et al., 2021;Saunders et al., 1990).
While Dilp2 is upregulated in D. melanogaster under diapausing conditions (Kubrak et al., 2014), Kr-h1 should be downregulated in the absence of JH (Zhang et al., 2021). In accordance with D. melanogaster, we found an upregulation of Dilp2 and a downregulation of Kr-h1 in diapausing D. littoralis females (Figure 10g). With regard to Crz and Pdf, we found that both transcripts are significantly reduced under diapausing conditions as compared to reproductive conditions (Figure 10g). This shows that CRZ/PDF neuropeptide production is reduced during diapause and therefore the size of the cell somata may shrink. In case of PDF, clearly less peptide is stored in the somata of the cells. In case of CRZ, the same amount of the peptide might be stored in the somata of the neurons than under summer conditions, perhaps to enable a quick release once it is needed (see further discussion in Section 3.9).
Since the mRNA was extracted from the whole head, we cannot be sure that the changes seen in the mRNA levels are exclusively coming from the CRZ neurons. Indeed, the l-LN v s and partially the s-LN v s are the additional PDF-positive neurons in the brain of D. littoralis ( Figure 4e). It is, however, unlikely that they could influence the Pdf transcripts under different conditions, as no significant difference in the PDF staining intensity was detected between diapausing and reproductive flies in the l-LN v s (Figure 10e), and the s-LN v s showed an extremely low level of PDF (see Section 4.1). In the brain of D. melanogaster and D. littoralis, no additional neurons were found to be CRZ positive via immunohistochemistry. However, in the brain of D.
melanogaster, Crz transcripts were detected in the optic lobe via in situ hybridization (Choi et al., 2005). Interestingly, the same study showed the absence of these transcripts in the optic lobe of D. virilis. Being part of the virilis group, it is likely that the same is true in D. littoralis; however, we cannot completely exclude that these optic lobe neurons might present Crz transcripts and play a role in the differences seen at the mRNA level in this species. Whether these results are specific to CRZ neurons or affect other neurons in the brain, we show here that photoperiod and temperature clearly play a role in the transcription of these genes (Figure 10g). The same is true for the soma size of the CRZ CN neuron (Figure 10b,c).

3.9
CRZ and PDF may play a crucial role in diapause regulation in D. littoralis PDF appears to promote the reproductive state since its mRNA and peptide levels are significantly higher in reproductive than in diapausing flies. Most interestingly, a similar role of PDF was also proposed for D. melanogaster. Here, PDF from the s-LN v terminals signals to the DILP neurons in the PI increasing their cAMP levels (Nagy et al., 2019).
F I G U R E 1 0 Pigment-dispersing factor (PDF) and corazonin (CRZ) staining intensity, CRZ neurons area, and reproduction-related genes transcripts in Drosophila littoralis under different conditions. (a) Under diapausing conditions, the PDF staining intensity is lower in CRZ neurons, but CRZ intensity is higher. Additionally, the size of CN neuron (yellow arrow) is visibly reduced compared to reproductive conditions. Z stack of 17 confocal planes for diapausing and 22 for reproductive conditions. (b) Under different temperature and photoperiodic conditions, the cell size of (Continues) F I G U R E 1 0 (Continued) diapausing flies is significantly smaller in the large CRZ CN neurons as well as in the six smaller CRZ 1-6 , although with a smaller difference. (c) Under different photoperiodic conditions but same temperature, only the CRZ CN size is significantly reduced in diapausing flies. (d) CRZ (blue) intensity is significantly higher in the CRZ 1-6 and CRZ CN of diapausing flies. PDF intensity (magenta) is significantly lower in the CRZ 1-6 and CRZ CN of diapausing flies. (e) There is no significant difference between the PDF immunostaining quantification in large ventrolateral neurons (l-LN v s) under diapausing and reproductive conditions. (f) Under diapausing condition at temperature 16 • C, the CRZ intensity is higher than the one of reproductive condition at the same temperature. *p < .05, **p < .01, and ***p < .001 after Wilcoxon rank-sum test. (g) Relative mRNA expression (normalized to the summer condition) of genes related to reproduction of Dilp2, Kr-h1, Crz, and Pdf. There is a significant reduction in mRNA expression for Crz, Kr-h1, and Pdf in diapausing flies. On the contrary, there is a significant increase in the expression of Dilp2 under diapausing conditions. **p < .01 and ***p < .001 after Wilcoxon rank-sum test. Scale bars: 15 μm in panel (b). n = 20 flies in panels (b), (d), and (e); n = 10 flies in panels (c) and (f).
Subsequently, PKA is activated, which phosphorylates the transcription factor Eyes absent 3 (EYA3) and leads to its degradation (Hidalgo et al., 2023). EYA3 accumulates at high levels under winter conditions and promotes dormancy in D. melanogaster (Abrieux et al., 2020). Since PDF levels in the s-LN v terminals are higher under summer conditions than under winter conditions, PDF promotes EYA3 degradation in the summer, thereby keeping the flies in the reproductive state (Hidalgo et al., 2023). Future studies are needed to show whether a similar molecular mechanism works in D. littoralis.
In D. littoralis, a stronger PDF signaling to the DILP neurons may occur via the fibers of the CRZ CN neurons that pass by the PI and even follow the DILP neurons toward the CC (Figures 5 and 9). Furthermore, PDF may directly work on glucagon-like adipokinetic hormone signaling in D. littoralis. In D. melanogaster, PDF stemming most likely from PDF neurons in the abdominal ganglia activates adipokinetic hormone-producing cells in the CC by increasing their cAMP levels (Braco et al., 2022). The knockdown of the PDF receptor in these cells significantly extends life span of the flies under starvation conditions (Braco et al., 2022), a situation that is comparable to reproductive dormancy. Assuming that PDF receptor is also expressed in the CC of D. littoralis, PDF from the CRZ CN neurons might regulate both DILP and adipokinetic hormone metabolic signaling pathways. CRZ appears to have the opposite role of PDF and provokes diapause by inhibiting egg production as was shown in the ant Harpegnathos saltator and the fly D. melanogaster (Gospocic et al., 2017). This fits to the still high levels of CRZ immunostaining in the CRZ CN neurons under diapausing conditions, as egg production should be inhibited.
However, CRZ is not produced and probably just stored without any release when the flies are in diapause. Moreover, Crz mRNA levels are high and some CRZ is present under summer conditions, especially in the CRZ 1-6 neurons (Figure 10d), indicating that CRZ is needed even in the reproductive active state. Indeed, CRZ plays several crucial roles in fly physiology and metabolism (Kubrak et al., 2016;Zandawala et al., 2021). In the ant species Cataglyphis nodus, the CRZ neurons show size differences depending on developmental and behavioral states (Habenstein et al. 2021). They are smaller in freshly eclosed ants and interior workers compared to the bigger cells of foragers, supporting the idea that these neurons are important for metabolic and behavioral changes during life span. Thus, we hypothesize that CRZ is strongly elevated and released in autumn before the flies enter diapause to inhibit egg production. During diapause, CRZ is stored in the cell somata because it is not needed in flies that are barely metabolically active.
However, the cells are ready to release it in spring as soon as the flies become active again, and during summer conditions, it is produced and released at levels that are essential to control metabolic activity but are not high enough to inhibit egg production.

CONCLUSIONS
Here, we describe for the first time the neuropeptide composition in the circadian clock neurons and neurosecretory cells of D. littoralis and compare it with the expression in D. melanogaster. We found striking differences in the two fly species that might explain differences in their circadian rhythmic behavior and seasonal adaptation (Lankinen, 1986a;Menegazzi et al., 2017).

The s-LN v circadian clock neurons are less important in D. littoralis
The PDF-and sNPF-expressing s-LN v s, which are important for circadian rhythmicity and keeping D. melanogaster flies in the reproductive summer state (Grima et al., 2004;Helfrich-Förster, 1998;Hidalgo et al., 2023;Nagy et al., 2019;Shafer & Taghert, 2009), appear to be less important in D. littoralis. These neurons in D. littoralis express only low amounts of PER and very little PDF and lack sNPF completely, which altogether suggests that their circadian clock is impaired. This can explain the weak circadian rhythmicity of the animals. In addition, it appears unlikely that the little PDF released by the s-LN v s could activate the DILP neurons in the PI and influence the reproductive state of the animals. Thus, the s-LN v s might neither play an important role in rhythmic behavior nor in photoperiodism. On the other hand, CRY is well conserved between the two species, suggesting that it is crucial for circadian entrainment and the detection of changing photoperiods.

4.2
The neuropeptides PDF, CRZ, DH31, and DH44 might contribute to the control of seasonal reproduction in D. littoralis In D. littoralis, PDF stemming from the CRZ-positive neurosecretory cells in the PL may have overtaken the role of s-LN v -derived PDF in keeping the flies reproductively active. This PDF may act on PDF receptors on the adipokinetic hormone-expressing neurons in the CC and increase metabolic activity (Braco et al., 2022). PDF may even be released via the CC and could target peripheral tissues that express the PDF receptor (Krupp et al., 2013;Talsma et al., 2012). PDF might also affect DILP neurons in the PI and block EYA3 action and thereby keep the flies in the summer state as was shown by Hidalgo et al. (2023) in D. melanogaster. Other neuropeptides expressed in the PL may then induce reproductive diapause as soon as night length exceeds a critical value in D. littoralis. As discussed, CRZ is a good candidate for such a role, but DH31 may also serve this function. DH31 has recently been shown to inhibit vitellogenesis by repressing JH biosynthesis in the CA of D. melanogaster (Kurogi et al., 2023). As we show here, DH31 expression in the PL is well conserved between D. melanogaster and D.
littoralis, making it likely that it serves the same function in both species.
Lastly, DH44 may contribute to diapause preparation by controlling feeding and initiating crop filling for food storage, which is essential before entering diapause. In D. melanogaster, DH44-positive neurons are nutrient sensing, innervate the crop, express the mechanosensory channel Piezo, and stimulate feeding (Dreyer et al., 2019;Dus et al., 2015;Oh et al., 2021). In D. littoralis, DH44 is not only expressed in the neurosecretory cells of the PI but additionally in two PL cells, suggesting that it may play a prominent role in diapause preparation.

4.3
The AstC-positive clock neurons are good candidates for transferring information about night length and temperature to the neurosecretory system of D. littoralis So far, it is unclear which circadian clock neurons transfer the information about day/night length to the neurosecretory cells in the PI and PL of D. littoralis, but the AstC-positive clock neurons are good candidates.
Notably, AstC is expressed in more clock neurons of D. littoralis than of D. melanogaster. Here, we show that the AstC-and ITP-expressing 5th LN and LN d have exactly the same morphology as the corresponding neurons in D. melanogaster that only express ITP. As discussed, these two neurons get light input via the accessory medulla (Li et al., 2018) and project to the PL and PI (Hermann-Luibl et al., 2014), where they may form synapses with neurosecretory neurons. Furthermore, the same ITP-positive clock neurons were shown to be postsynaptic partners of most of the clock clusters (Shafer et al., 2022). Thus, the AstC-and ITP-expressing 5th LNs are perfectly suited to transfer information about day/night length to the neurosecretory system of D. littoralis. In addition, the AstC-positive LPNs and DN 3 s may transfer information about environmental temperature to the neurosecretory system of D. littoralis. The LPNs form a dense fiber network in the superior protocerebrum close to the PI in D. melanogaster  and, as we show here, this is also true in D.
littoralis (see Figure 8). In D. melanogaster, the LPNs are involved in synchronizing the daily activity of flies to temperature cycles (Reinhard, . The AstC-positive DN 3 s are even more promising candidates for transferring temperature information to the neurosecretory system because they have recently been shown to be temperature sensitive and to stimulate egg production via a still unknown pathway in D. melanogaster (Meiselmann et al., 2022). The same study showed that the reduction in AstC under lower temperatures is conserved in D. virilis as well. Since our study showed that more DN 3 s express AstC in D. littoralis compared to D. melanogaster, this pathway is likely valid also in D. littoralis.
In summary, our study lays the foundation for future investigations into the functional connection between the circadian clock and the neurosecretory system in a high-latitude fly. Future studies will have to demonstrate whether the hypotheses formulated here are true. Establishing CRISPR/Cas9 mutagenesis in D. littoralis will allow to unravel the proposed importance of certain clock and neuropeptide genes in the photoperiodic control of diapause.

ACKNOWLEDGMENTS
We thank Jan Veenstra for providing the antibodies against AstC, CRZ, DH31, DH44, DILP2, DILP3, and sNPF; Takeshi Todo for the CRY antiserum; Heinrich Dircksen for the antibodies against PDH and ITP; Kenji Tomioka for the antibody against PDF cricket; and Justin Blau for the antibodies against PDP1 and PDF C7. Finally, we thank Pingkalai Senthilan for advice on the qPCR experiments, Pekka Lankinen for providing the D. littoralis strain, and Anneli Hoikkala for providing meaningful information on this species and on the manuscript.
Open access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data supporting the findings of this study are available upon reasonable request from the corresponding author.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/cne.25522.