Neuropeptides as potential modulators of behavioral transitions in the ant Cataglyphis nodus

Age‐related behavioral plasticity is a major prerequisite for the ecological success of insect societies. Although ecological aspects of behavioral flexibility have been targeted in many studies, the underlying intrinsic mechanisms controlling the diverse changes in behavior along the individual life history of social insects are not completely understood. Recently, the neuropeptides allatostatin‐A, corazonin, and tachykinin have been associated with the regulation of behavioral transitions in social insects. Here, we investigated changes in brain localization and expression of these neuropeptides following major behavioral transitions in Cataglyphis nodus ants. Our immunohistochemical analyses in the brain revealed that the overall branching pattern of neurons immunoreactive (ir) for the three neuropeptides is largely independent of the behavioral stages. Numerous allatostatin‐A‐ and tachykinin‐ir neurons innervate primary sensory neuropils and high‐order integration centers of the brain. In contrast, the number of corazonergic neurons is restricted to only four neurons per brain hemisphere with cell bodies located in the pars lateralis and axons extending to the medial protocerebrum and the retrocerebral complex. Most interestingly, the cell‐body volumes of these neurons are significantly increased in foragers compared to freshly eclosed ants and interior workers. Quantification of mRNA expression levels revealed a stage‐related change in the expression of allatostatin‐A and corazonin mRNA in the brain. Given the presence of the neuropeptides in major control centers of the brain and the neurohemal organs, these mRNA‐changes strongly suggest an important modulatory role of both neuropeptides in the behavioral maturation of Cataglyphis ants.


| INTRODUCTION
Insect societies often comprise large numbers of individuals (Hölldobler & Wilson, 1990;Wilson, 1971), and the organization within social-insect colonies is highly sophisticated and requires a species-specific and flexible task allocation. Some species exhibit distinct polymorphic worker castes devoted to specific tasks, for example, minor and major workers in leafcutting ants (Hughes et al., 2003;Wilson, 1980Wilson, , 1984. However, most social insect species lack task-specific polymorphic worker castes. Instead, individual workers show a high degree of behavioral flexibility and different behavioral phenotypes are rather associated with the age (temporal polyethism) than with a specific morph (Hölldobler & Wilson, 1990;Robinson, 1992;Robinson, 2002). Ants of the genus Cataglyphis (Foerster 1850) represent prominent examples for a marked temporal polyethism.
Cataglyphis workers undergo distinct behavioral transitions. Newly emerged ants (callows), recognizable by their pale and soft cuticle, behave very sluggishly. Then, workers go through several interior phases inside the darkness of the nest. During the first 2 weeks, interior workers (interior I) serve as food storage for the colony (repletes) before they become more active (interior II) and accomplish tasks such as brood care or nest maintenance. The most drastic change, the interior-forager transition, happens after about 4 weeks when the ants leave the nest for the first time (Schmid-Hempel & Schmid-Hempel, 1984; reviewed by Wehner & Rössler, 2013). In order to calibrate their visual navigational systems, the ants first perform learning walks for 2-3 days close to the nest entrance before they finally start to forage solitarily Fleischmann et al., 2018; for reviews, see Grob et al., 2019;Rössler, 2019). The interior-forager transition is accompanied by fundamental changes in behavior. Whereas interior workers predominantly rely on olfactory and tactile cues inside the dark nest, foragers navigate on far-ranging foraging trips under bright sunlight mostly using visual compass cues (Ronacher, 2008;reviewed by Wehner & Rössler, 2013;Buehlmann et al., 2014;Huber & Knaden, 2015). The interior-forager transition is associated with high levels of structural synaptic plasticity in two visual pathways of the ants' brains Schmitt et al., 2016;Stieb et al., 2010;Stieb et al., 2012). Recently, a detailed atlas of all brain neuropils and major connecting fiber systems in Cataglyphis nodus was published (Habenstein et al., 2020). This was complemented by a comprehensive peptidomic analysis of the C. nodus brain . Based on this profound knowledge on the behavioral transitions, brain anatomy, neuroplastic changes, and neuropeptide inventory, C. nodus ants are well-suited experimental models to address the potential role of neuropeptides in modulating intrinsic mechanisms underlying age-related behavioral plasticity.
Neuropeptides play crucial roles throughout an insect's life. They are involved in physiological and developmental processes and in the modulation of many essential behaviors (for reviews, see Fricker, 2012;Kastin, 2013;Nässel & Zandawala, 2019). Recent studies suggest that the neuropeptides allatostatin-A (Ast-A), tachykinin (TK), and corazonin (Crz) act as regulators of behavioral plasticity in social insects (Gospocic et al., 2017;Han et al., 2021;Pratavieira et al., 2014;Schmitt et al., 2017;Takeuchi et al., 2003). Age-related fluctuations of Ast-A and/or TK have been revealed in honeybees (Pratavieira et al., 2014;Takeuchi et al., 2003) and in the ant Cataglyphis fortis (Schmitt et al., 2017). Ast-A is a major regulator of the biosynthesis of juvenile hormone (JH) in many insects (for reviews, see Weaver & Audsley, 2009;Verlinden et al., 2015). However, a clear evidence for an allatoregulatory role is missing in Cataglyphis. In honeybees, Ast-A induces shifts in the stress reactivity and modulates appetitive olfactory memory (Urlacher et al., 2016;Urlacher et al., 2019). Ast-A as well as TK-related peptides (TRP) are suspected to further modulate aggressive behavior in bees (Pratavieira et al., 2018). TK signaling modulates locomotion and general activity levels (Kahsai et al., 2010;Winther et al., 2006), olfactory sensitivity Jung et al., 2013;Ko et al., 2015;Winther et al., 2006), and nociception (Im et al., 2015). Although most of these studies were conducted in Drosophila, the similarities in the location of tachykinergic neurons in the antennal lobes (ALs) and the central complex (CX) may indicate similar functions in Cataglyphis Schmitt et al., 2017). Whereas TK and Ast-A have been suggested to be involved in behavioral transitions in closely related C. fortis (Schmitt et al., 2017), Crz remained largely neglected in previous studies.
Crz is a highly conserved neuropeptide and omnipresent in most insect species (Predel et al., 2007). In Drosophila, the endocrine corazonergic pathway is known to alter stress responses and metabolism (Johnson, 2017;Kapan et al., 2012;Kubrak et al., 2016;Zandawala et al., 2021). In addition, Crz signaling is important for the development of moth larvae (Kim et al., 2004;Tanaka, Hua, et al., 2002) and induces morphological changes in locusts, which are associated with behavioral transitions (Maeno et al., 2004;Sugahara et al., 2018;Tanaka, Zhu, et al., 2002;Tawfik et al., 1999). The effect of Crz on developmental processes is not restricted to insects, as similar effects of Crz on growth and maturation could be found in bristle worms (Andreatta et al., 2020). Gospocic et al. (2017) suggested that Crz is an important regulator of the sexual and behavioral differentiation in the ponerine ant Harpegnathos saltator. Higher Crz levels were found in workers compared to queens (or pseudoqueens).
Moreover, worker-specific hunting behavior could be induced by experimental increase of Crz levels.
In the present study, we investigated the spatial expression patterns of Ast-A, TK, and Crz in callows, interior workers, and foragers of the ant C. nodus. We compared general innervation patterns and cell body volumes of peptidergic neurons across different behavioral stages using immunohistochemistry. In addition, we analyzed the mRNA levels of the three neuropeptides in the brain. The results show that expression levels and soma sizes particularly of Ast-A and Crz relate to the behavioral stage. This strongly suggests the involvement of these neuropeptides as modulators during behavioral maturation of Cataglyphis workers.

| Animals
C. nodus nests were located in the Strofylia National Park, Greece (38 15 0 N, 21 37 0 E). To be able to differentiate foragers from interior workers with high accuracy, foragers were marked outside the nest using acrylic paint (Motip Lackstift Acryl, MOTIP DUPLI GmbH, Haßmersheim, Germany) over at least three consecutive days before the nests were excavated. Afterward, all unmarked animals were considered to be either interior workers or callows. Callows were discriminated from interior workers by their pale cuticle. To ensure that callows and interior workers were never exposed to sunlight, the nest was excavated during night using red light. After excavation, the ants were kept in complete darkness in large plastic boxes with permanent access to water and honey water. Animals of different behavioral stages were collected arbitrarily out of the boxes and brains were dissected within 30 h after nest excavation. Dissections of the brains from ants of different behavioral stages were pseudo-randomized to avoid effects of potential daytime-dependent changes of neuropeptide expressions.
For the expression analyses, ant colonies were transferred to the University of Würzburg. In the laboratory, the ants were kept in a dark cabinet within a climate chamber at a constant temperature (28 C) and relative humidity (30%). The colonies had permanent access through a light trap connection into a foraging arena with a 12 h/12 h day/night cycle. During the 12-h light cycle, light was generated by UV-emitting fluorescent tubes (Repti Glo 2.0/10.0, Exo Terra, Holm, Germany). In the foraging arena, the animals had an infinite supply of water and were fed with honey water (1:2) and dead arthropods twice a week. The animals were allowed to acclimate for 48 h. Thereafter, all ants in the foraging arena were marked on three consecutive days and considered as foragers. In the following, all unmarked ants were classified as interior workers since no callows had remained in the colony after the transport to Würzburg.  Table 1).

| Primary antibodies
The polyclonal anti-Ast-A antibody was raised in rabbit and conjugated to thyroglobulin with glutaraldehyde. The specificity of the antiserum has been verified by a competitive enzyme-linked immunosorbent assay (ELISA) (Vitzthum et al., 1996) and its affinity was demonstrated in diverse insect species (e.g., Carlsson et al., 2013;Kahsai & Winther, 2011;Kreissl et al., 2010) including Cataglyphis ants (Schmitt et al., 2017).
The polyclonal Crz antiserum was raised in rabbit and coupled to bovine serum albumin. The specificity of the antibody was characterized by ELISA (Veenstra, 1991;Veenstra & Davis, 1993). The antibody was used to study the immunoreactivity to Crz in different arthropods (e.g., Hansen et al., 2001;Hou et al., 2017;Roller et al., 2003).

| Immunohistochemistry
Animals were anesthetized on ice and decapitated under red light.
The head capsule was opened between the compound eyes and the were removed and fixated in 4% formaldehyde in phosphate-buffered saline (PBS) overnight at 4 C. Thereafter, the brains were rinsed in PBS (3 Â 10 min), stored in PBS in the dark at 4 C for up to 3 days, and subsequently transferred in a cooler to Würzburg. In Würzburg, samples were stored at 4 C until further processing for up to two more days. To facilitate the penetration of the antibodies, the brains were treated with PBS containing Triton-X (PBST; 2% PBST for 10 min, 0.5% PBST for 3 Â 10 min). Subsequently, the brains were preincubated in 0.5% PBST with 2% normal goat serum (NGS) for 1 h at 4 C before the brains were incubated in the primary antibody solution (2% NGS in 0.5% PBST, either Ast-A, Crz or TK, dilution see Table 1) for 4 days at 4 C. After rinsing in PBS (3 Â 10 min), the brains were transferred into the anti-synapsin antibody solution (2% NGS in 0.2% PBST, Table 1)  were damaged or the quality of confocal images was not sufficient for analyzing the volumes. In those cases, we used only the volume of the corazonergic somata of one hemisphere.

| RNA extraction and cDNA synthesis
• Preparation in Greece: Ants were chosen arbitrarily from the colony according to their behavioral stage. Then, ants were anesthetized on ice, the head was removed, and a small window was cut into the head capsule to allow the penetration of RNAlater stabilization solution (Thermo Fisher Scientific) into the brain tissue. The samples were stored in the RNAlater solution in the dark at 4 C for up to 3 days before they were transferred in a cooler to Germany.
• Preparation in the laboratory: Ants were collected arbitrarily from the colony according to their behavioral stage and immediately snap-frozen in liquid nitrogen. Samples were stored in a freezer at À80 C until the brain was dissected.
All brains were dissected in RNAlater solution, transferred into a reaction tube and snap-frozen in liquid nitrogen. Samples were stored at were determined relative to transcripts of elongation factor-1 alpha (EF1 α) using the ΔΔCt method. EF1 α was shown to be a suitable reference gene in brains of honeybees (Lourenço et al., 2008;Reim et al., 2013) and several ant species (Moreira et al., 2017;Ratzka et al., 2011). Therefore, we decided to use this gene in C. nodus. We calculated and compared the absolute copy number of EF1a. Here, we observed no differences (p Greece = 0.59, χ 2 = 1.056, Kruskal-Wallis test; p lab = 0.432, As not all data sets were normally distributed, we used nonparametric tests for all statistical analyses. Data sets with more than two groups were tested by using the Kruskal-Wallis test (α = .05). To analyze significant differences between the groups, post hoc pairwise comparisons were performed using the Mann-Whitney U test with Bonferroni correction.
For all data sets with two behavioral groups, we used the Mann-Whitney U test (α = .05).

| Nomenclature
We followed the unified nomenclature rules for insect brains (Ito et al., 2014). To identify neuropils of the central brain, we used the three-dimensional atlas of the C. nodus brain as a reference (Habenstein et al., 2020) (see also https://www.insectbraindb.org to access 3D reconstructions of the C. nodus brain).

| RESULTS
3.1 | Ast-A, Crz, and TK immunoreactivity in Cataglyphis workers To compare the localization of neuropeptides in Cataglyphis brains across different behavioral stages, we used immunohistological stainings with antibodies specific for Ast-A, Crz, and TK. Close visual inspection did not reveal any obvious differences in the general distribution and branching patterns of Ast-A-, Crz-, and TK-like immunoreactivity (Ast-A-ir, Crz-ir, and TK-ir) among callows, interior workers, and foragers. Across all three behavioral stages, Ast-A-ir is localized in numerous neurons of the brain with arborizations in important primary sensory neuropils such as the AL and the optic lobes (OL, Figure 1(a,f)). We found Ast-A-ir scattered across the entire volumes of virtually all AL glomeruli (Figure 1(a,d)). In the lobula (LO) and medulla (ME), Ast-A-ir neurons are only present in very distinct neuropil layers. In contrast, branches of Ast-A-ir neurons were found across the entire lamina neuropil (Figure 1(f)). Ast-A-ir is also present in prominent high-order integration centers of the brain (Figure 1(a-c,e)).
Ast  Figure 1(b)). In honeybees, the neurons of the PCT provide GABAergic feedback to the MB calyces (Grünewald, 1999;Haehnel & Menzel, 2010). Our staining revealed Ast-A-ir neurons also beyond the major neuropils in brain regions with less T A B L E 2 Oligonucleotides used for quantitative real-time PCR   (Habenstein et al., 2020). Our detailed brain atlas (Habenstein et al., 2020) enabled us to assign branching patterns to neuropils within central brain regions that have not been accessible in a previous study on Ast-A neurons in the closely related ant C. fortis (Schmitt et al., 2017).
In particular, Ast-A-ir is present in the lateral accessory lobe, the superior neuropils (SNP), the inferior neuropils, the ventrolateral neuropils (VLNP), the lateral horn (LH), and less pronounced also in more posterior neuropils, such as in the ventromedial neuropils (VMNP) and the periesophageal neuropils (Figure 1(a,g-i)). We found no Ast-A-ir in the antennal mechanosensory and motor center (AMMC). Numerous cell bodies of the Ast-A-ir neurons are dispersed across the entire cell body rind (CBR). The distribution of TK-ir neurons did not reveal any obvious differences between the three behavioral stages. Similar to the Ast-A-ir, innervation by numerous TK-ir neurons was present in major sensory input regions, high-order processing centers and extensively in the CANP ( Figure 2). We found very dense TK-ir across the entire volume of virtually all AL glomeruli (Figure 2(a,c)). In the ME and LO, TK-ir is present only in distinct layers. However, in contrast to the Ast-A-ir pattern, we found no TK-ir in the LA (Figure 2(d)). Similar to the pattern found in Ast-A-ir neurons in the CX, TK-ir is restricted to the CB and absent in the protocerebral bridge and noduli. The PED is intensely innervated by TKir neurons, which is reflected in labeling of multiple layers in the MB VL ( Figure 2(a,b)). TK-ir was found in very prominent ventral and several distinct dorsal layers of the VL (Figure 2(b,f)). In the honeybee, the ventralmost layer is known as the gamma lobe receiving input from Type II In contrast to the labeling of Ast-A and TK, Crz-ir neurons show very different and distinct branching patterns. They are restricted to a very small number of neurons forming a tight cluster in each brain hemisphere. Each cluster consists of four cell bodies, located in the pars lateralis (Figure 3(a,d)), a region housing neurosecretory cells in insects (Veelaert et al., 1998;Siegmund & Korge, 2001; reviewed by Raabe, 2012). The Crz-ir neurons extend projections into the superior medial protocerebrum (SMP) and the flange (FLA, Figure 3(a-c)). In addition, one axon bundle per hemisphere projects into the ipsilateral retrocerebral complex (RCC, Figure 3(c,e,f)), a major neurohemal organ in insects (Gade et al., 1997;Predel, 2001). The rest of the central brain, including sensory input regions and high-order processing and integration sites, did not show any Crz-ir. Like the other two neuropeptides, Crz-ir neurons did not exhibit obvious differences in their branching patterns between callows, interior workers, and foragers.

Despite consistent general branching patterns of Ast-A-, TK-, and
Crz-ir neurons in ants of different behavioral stages, we analyzed potential differences in the volume of cell bodies. This was possible in the case of Crz-ir neurons due to the relatively large size of the cell bodies and their small overall number. We reconstructed the three-dimensional structure of the four Crz-ir somata in callows, interior workers, and foragers to compare their volumes. Foragers displayed significantly larger Crz-ir somata than callows (p < .001, Z = 4.424, χ 2 = 20.345) and interior workers (p < .001, Z = 4.422, χ 2 = 19.417, Figure 4). To exclude seasonal or colony related factors, the experiment was repeated for interior workers and foragers from another C. nodus colony at a different time of the year (June instead of August). Likewise, corazonergic cell-body volumes were significantly (p < .001, Z = 3.971, U = 151.0) larger in foragers (mean volume 9.22 ± 2.28 Â 10 3 μm, n = 13) compared to interior workers (mean volume = 4.97 ± 1.11 Â 10 3 μm, n = 12, data not shown).

| Quantitative real-time PCR comparison of neuropeptide mRNA expression
To investigate potential differences at the level of gene expression, we compared the mRNA expression levels of TK, Ast-A, and Crz in ants of the three behavioral stages. To measure mRNA levels as close to the natural conditions as possible, we conducted the experiments on ants collected in their natural habitat within the Strofylia National Park (Greece). This revealed neuropeptide-specific differences between mRNA expression levels in the three behavioral stages ( Figure 5). Ast-A mRNA expression was significantly lower in interior workers compared to both callows (p = .047, Z = 2.414, χ 2 = 11.00) and foragers (p = .005, Z = 3.12, χ 2 = 12.542, Figure 5(a)). In contrast, TK mRNA expression levels remained stable throughout the different behavioral stages (p = .788, χ 2 = 0.477, Figure 5 (b)). The largest change in relative mRNA expression levels was found for Crz with more or less similar expression levels in callows and interior workers (p = 1.00, Z = 0.864, χ 2 = 3.938), but a significant increase in foragers compared to callows (p = .002, Z = 3.433, χ 2 = 16.50) and interior workers (p = .005, Z = 3.124, χ 2 = 12.562). Interestingly, the median of the relative expression value for Crz in foragers is almost twice as high as in the other behavioral stages (Figure 5(c)).
To minimize artificially generated errors (e.g., changed mRNA production caused by stress during excavation of the nest), we repeated the experiment using the same colony after it was transferred to constant conditions in our climate chamber in Würzburg. Since no more callows Ast-A and TK belong to the most abundant neuropeptides in the Cataglyphis brain. Both have multiple paracopies within the respective prepropeptide  and, as shown in our present work, extensively innervate major portions of the ant's brain, with differences in individual neuropils as described in the results section.
They are present in high-order processing centers as well as in major visual, antennal mechanosensory, and olfactory neuropils (Figures 1 and 2) which is in line with previous studies in the closely related ant C. fortis (Schmitt et al., 2017) and with data from mass spectrometric analyses in C. nodus . hypothesis that TK is implicated in spatial orientation (Kahsai et al., 2010).
Using the three-dimensional atlas of the Cataglyphis brain as a reference (Habenstein et al., 2020), we were also able to map TK-and Ast-A-ir in less well-known CANP. In addition to the AL glomeruli and subregions of the MBs, we found TK-and Ast-A-ir in the LH, the , and foragers (n = 12). Crz expression is significantly increased in foragers. For all neuropeptides, expression level of callows was set to one to determine relative changes compared with other behavioral stages. Boxes indicate 25th and 75th percentiles and the solid line represents the median of the data. Whiskers extend to all data points within 1.5 times of the interquartile length from the median. Outliers are shown as circle. For statistics, the Kruskal-Wallis test with Bonferroni correction was used. Significant different data are indicated by an asterisk (α = .05) VLNP, and the SIP-neuropils involved in primary and high-order olfactory processing (Galizia & Rössler, 2010;Habenstein et al., 2020).
Interestingly, treatments with Ast-A lead to dose-dependent impairments of associative olfactory learning in honeybees (Urlacher et al., 2016). Ast-A and TK might exert this function on intrinsic MB neurons (KCs) leading to effects on olfactory memory formation.
Modulation at the level of the LH may provide the substrate for orchestrating the behavioral significance or sensitivity for olfactory stimuli, as it was demonstrated in fruit flies Winther et al., 2006;Winther & Ignell, 2010) and the American cockroach (Jung et al., 2013). In addition to the OL, we also found innervation by TK-and Ast-A-ir neurons of CANPs involved in processing of visual and antennal mechanosensory input. In particular, VLNP and VMNP are important target regions of visual fiber bundles from the OL in C.
The spatial distribution of Crz is largely conserved in insect brains (e.g., Predel et al., 2007;Roller et al., 2003;Settembrini et al., 2011;Veenstra & Davis, 1993;Wen & Lee, 2008). Corazonergic cell bodies are often located in the region of the pars lateralis and exhibit projections to the SMP and the RCC. This was also found in the honeybee and termites (Verleyen et al., 2006;Z avodsk a et al., 2009). Similarly, in C. nodus, we found a cluster of four Crz-ir neurons with cell bodies located in the pars lateralis and projections to the SMP and the RCC. In addition, neurites extend into the FLA, a small subregion within the CANP. Whether this is also the case in other insects analyzed so far cannot be conclusively judged due to the lack of detailed brain atlases of the central protocerebral neuropils. As an important neurohemal organ, the RCC produces and releases hormones such as the adipokinetic hormone or JH (Diederen et al., 2002;Goldsworthy et al., 1972;Noriega, 2014;Tobe & Pratt, 1974;Tobe & Stay, 1985). JH together with the egg yolk precurser protein vitellogenin (VG) have been implicated in orchestrating the timing of behavioral transitions in social Hymenoptera (reviewed by Hamilton et al., 2017).
The synthesis of JH in the corpora allata of the RCC is mediated via allatoregulatory neuropeptides from neurosecretory cells from the brain (Tobe & Stay, 1985). Although studies in other insects could demonstrate an inhibitory effect of allatostatins and stimulatory effects of allatotropins on JH biosynthesis, the underlying mechanisms seem to be complex and may vary between different species (reviewed by Stay, 2000;Bendena et al., 2020). Recent studies therefore started to speculate about the participation of further neuropeptides such as sNPF or Crz in regulating JH biosynthesis (Bendena et al., 2020;Johnson, 2017). However, so far no evidence exists in the case of Crz (Johnson, 2017). Interestingly, Crz acts antagonistically to VG in Harpegnathos ants and in Drosophila (Gospocic et al., 2017). An antagonistic relationship between JH and VG was shown to be involved in behavioral maturation of honeybee workers (reviewed by Harwood et al., 2017). Crz might act on similar pathways, which makes the potential role of an interplay between Crz, JH, and VG in behavioral maturation in social insects a highly interesting target for future studies.
F I G U R E 6 Changes of allatostatin-A (Ast-A), tachykinin (TK), and corazonin (Crz) expression in brains of interior workers and foragers (laboratory study). (a) Ast-A expression in interior workers and foragers. Ast-A expression is significantly increased in foragers. (b) TK expression in interior workers and foragers. No significant difference of the TK expression was found. (c) Crz expression in interior workers and foragers. Crz expression is significantly increased in foragers. For all neuropeptides, expression level of interior workers was set to one to determine the relative changes compared with foragers. Boxes indicate 25th and 75th percentiles and the solid line represents the median of the data. Whiskers extend to all data points within 1.5 times of the interquartile length from the median. Outliers are shown as circle. For statistics, the Mann-Whitney U test was used. Significant different data are indicated by an asterisk (α = .05)

| Changes in neuropeptide expression at different behavioral stages
We had focused on Ast-A, TK, and Crz as candidate modulators of the behavioral transitions in Cataglyphis ants. We hypothesized if these neuropeptides act as modulators of behavioral stage transitions in adult workers, this should be reflected in changes of neuropeptide expression patterns between respective stages. In the case of Ast-A and Crz, we found clearly increased expression levels in foragers suggesting a regulatory role associated with behavioral maturation. In contrast, TK expression levels remained unchanged in C. nodus brains.
This is different in the honeybee, where TK brain expression is upregulated in foragers (Han et al., 2021;Takeuchi et al., 2003).
Manipulation experiments further indicate that TK modulates response thresholds of the proboscis extension response (PER) in a stage specific manner: Whereas TK injection decreased PER responsiveness, TK knockdown had an opposite effect (Han et al., 2021).
Injection of TK and Ast-A in another study induced aggressive behavior in 7-day-old honeybee workers (Pratavieira et al., 2018). Typically, aggressive behavior is attributed to older honeybee workers compared to younger stages (Breed et al., 1990;Pearce et al., 2001;Rittschof et al., 2018). Similar observations were made in Cataglyphis cursor (Nowbahari & Lenoir, 1989). Schmitt et al. (2017) demonstrated slight alterations of TK innervation patterns in the CB during early stages (between Days 7 and 14) in C. fortis ants raised under controlled laboratory conditions. This may be related to an early transition between more or less motionless interior I-and highly mobile interior II stages in C. fortis (Schmid-Hempel & Schmid-Hempel, 1984). In our present work using freshly excavated natural colonies, we were not able to differentiate between these early interior substages and rather compared the three major behavioral stages-callows, interior workers, and foragers, in which we did not find significant changes in TK expression. Similarly, the TK distribution was not different between nurse and forager honeybees (Takeuchi et al., 2004). This suggests that the changes in TK innervation of the CB observed by Schmitt et al. (2017) may reflect an age-specific transition or sensitive phase during early adult development. Future work may need to focus on potential changes of TK expression at higher temporal resolution.
The changes seen in both Crz expression and expansion of Crz-ir cell body volumes suggest that this neuropeptide is a robust candidate involved in the orchestration of the interior-forager transition in Cataglyphis. In various insect species, Crz is associated with reproductive behavior and the control of developmental processes (e.g., Choi et al., 2005;Gospocic et al., 2017;Hou et al., 2018;Imura et al., 2020;Kim et al., 2004;Lee et al., 2008;Sugahara et al., 2018). The high concordance of findings in diverse insect taxa suggests a relatively early emergence of an ancient corazonergic pathway in the insect nervous system. Accordingly, Crz signaling could be a conserved trait that coopted the establishment of age-related polyethism in social insects, both in Hymenoptera and Isoptera (Gospocic et al., 2017). Our finding on changes in volume of Crz-ir cell bodies together with increased expression levels in foragers further support this idea. Together with the substantial volume increase of Crz-ir cell bodies, the expression results indicate an increase of Crz also at the peptide level. However, this still needs to be confirmed in future studies, for example, using quantitative mass spectrometric approaches. Our data suggest that shifts in Crz expression levels can be attributed to the transition from interior worker to forager (nurse-forager transition), as both the cell body volumes and mRNa expression levels do not differ between callows and interior workers. This indicates that Crz may potentially alter hormone release and physiological processes during the interiorforager transition, and thus might contribute substantially to the underlying behavioral switches. However, whether the increase in Crz expression is causal in triggering the behavioral switch or rather a consequence of changes in behavior between interior workers and foragers still requires future manipulation experiments, for example, neuropeptide injections in defined behavioral phenotypes. Although this remains speculative in Cataglyphis, a potentially important role of Crz in regulating behavioral transitions is supported by results from Crz injections inducing a forager-like behavior in the ponerine (basal) ant species H. saltator (Gospocic et al., 2017). However, due to the large phylogenetic distance between the two ant species future manipulation in Cataglyphis are necessary to confirm this.
Leaving the social environment and darkness of the nest for the first time is probably the most far-reaching behavioral decision in the life of social insects. Accordingly, the correct timing of this transition is most important for the ecological success of a colony. Furthermore, the transition needs to be flexible and adjustable to the specific needs of a colony.
This requires the control of a variety of external and intrinsic feedback mechanisms. Crz appears to be one of the key players in these complex species-specific systems, which is most likely orchestrated by further neuropeptide modulators, hormones, and biogenic amines.

| CONCLUSION
This study revealed the distribution of the neuropeptides Ast-A, Crz, and TK in major neuropils and less-well known central neuropils of the brain in C. nodus worker ants. Although differences are present between the distribution of Ast-A-and TK-ir, their wide distribution in major sensory neuropils and high-order integration centers suggests potential roles in modulating a diverse range of behavioral processes.

ACKNOWLEDGMENTS
The authors thank the Greek government and the management board of the Strofylia National Park for the permission to perform their research in the park and to transfer Cataglyphis ants to Germany. The authors are very grateful for the administrative help from Vasiliki