Larval S. frugiperda was reared on an artificial diet (wheat bran 40 g, yeast powder 34 g, casein 25 g, sorbic acid 2 g, corn meal 100 g, vitamin composite powders 4 g, methylparaben 4 g, agar 18 g, and distilled water 900 ml) in the laboratory under 16/8 light/dark, at 27 ± 1°C and 75% relative humidity. Larvae at the second day of 5th instar were used for the immunohistochemistry experiments. Adults were fed on a 10% sucrose solution.

In order to examine the distribution of serotonin-immunoreactive neurons in the brain and the gnathal ganglion, immunohistochemistry with the antisynapsin for labeling the neuropil structure and antiserotonin for labeling the serotonergic neurons was performed. The detailed procedures were described in previous studies (Zhao et al., 2016; Tang et al., 2019). The preparations were dissected. The brain and the gnathal ganglion were isolated from the insect body in Ringer’s solution (150 mM NaCl, 3 mM CaCl₂, 3 mM KCl, 25 mM sucrose, and 10 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid, pH 6.9) and fixed in 4% paraformaldehyde (PFA) solution (4% PFA in 0.1 M phosphate buffer) for 2–4 h at room temperature. Followed the rinse with phosphate-buffered saline (PBS) (684 mM NaCl, 13 mM KCl, 50.7 mM Na₂HPO₄, and 5 mM KH₂PO₄, pH 7.4) for 6 times, each time 10 min, the brain and the gnathal ganglion were preincubated in 10% normal goat serum (NGS) (Sigma Aldrich, St Louis, Mosby, United States) in PBS containing 0.5% Triton X-100 (PBST) for 3 h at room temperature to minimize the non-specific staining. Next, the brain and the gnathal ganglion were incubated in the primary antibodies, anti-SYNORF1 (1:100, Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, Iowa, United States) and antiserotonin serum (1:4,000, Immunostar Incorporation, Hudson, Wisconsin, United States), in PBST containing 5% NGS for 5 days at 4°C. After that, the samples were rinsed with PBS for 6 × 20 min and then incubated in the secondary antibodies, Alexa Fluor 488 conjugated goat antimouse (1:400, Invitrogen, Eugene, Oregon, United States) and Alexa Fluor 633 conjugated goat antirabbit (1:400, Invitrogen, Eugene, Oregon, United States) in PBST containing 5% NGS for 3 days at 4°C. Finally, the samples were rinsed 6 × 20 min with PBS, dehydrated in a series ethanol (50, 70, 90, and 96% and 2 × 100%, 10 min each time), cleared in methyl salicylate, and mounted in Permount.

All the images were obtained by laser scanning confocal microscope (Nikon A1, Japan) with 10X/2.2 air objective. Fluorescent dyes of Alexa Fluor 488 and Alexa Fluor 633 were excited by a 488-nm Argon laser and a 633-nm HeNe laser, respectively. The resolution of image is 1,024 × 1,024 and the interval is 2–3 μm. The value of high voltage (HV) is set at 35–50 and the laser intensity and other parameters are adjusted during scanning.

Confocal image data format was converted into tag image file format by Image J software [version 1.53f51, National Institutes of Health (NIH), United States]. Software of Amira version 5.3 (Visage Imaging, Fürth, Germany) was then used to analyze the image stacks. Neuropil structures and nerves were reconstructed by using the tool of LabelField of Amira and serotonin-immunoreactive neurons were reconstructed by using the module of SkeletonTree. Volumes of neuropils and cell bodies were measured by using the tool of TissueStatistics. Adobe Photoshop was used to adjust color, brightness, and contrast of confocal image when necessary and the image panels were edited by Adobe Illustrator 2021 (Adobe System, San Jose, California, United States). ANOVA and the bar chart of the average diameter and volume of cell bodies were performed by using GraphPad Prism version 9.0 (GraphPad Software Incorporation, San Diego, California, United States). The nomenclature of neuroanatomical structures, serotonin-immunoreactive neurons, and abbreviations suggested by Ito et al. (2014) and Tang et al. (2019) were used for larval S. frugiperda.

The immunoreactivity of anti-SYNORF1 revealed the synapsin-enriched neuropil and that of antiserotonin revealed cell bodies and cell fibers of serotonergic neurons in the brain (Figures 1A1–A5, merged; B1-B5, anti-SYNORF1; C1-C5, antiserotonin). The brain of larval S. frugiperda contained three main neuromeres: the protocerebrum (PR), the deutocerebrum (DE), and the tritocerebrum (TR). Based upon the intensity of immunoreactivity, several prominent neuropils in the PR could be identified (Figures 1B1–B5). Three-dimensional reconstructions of the neuropils were also created and then their volumes and relative volumes to those of the whole brain were measured (Figure 2). The alpha lobe (α) was located anteriorly and pointed vertically to the dorsal surface and the beta lobe (β) was located medially and pointed horizontally to the middle line of the PR (Figures 1B1,B2, 2A). The pedunculus (PED) lay in the middle of each hemisphere, forming the lobe (LOB) of mushroom body together with the α and β lobes (Figures 1B3,B4, 2A). Volume of LOB is 25.44 × 10⁴ μm³, about 4.55% of the whole-brain neuropil (Figures 2B,C). The calyx (CA) was located posteriorly in the PR (Figures 1B3,B4, 2A). Volume of CA is 34.54 × 10⁴ μm³, about 6.17% of the whole-brain neuropil (Figures 2B,C). The central body (CB), an unpaired neuropil, was located horizontally in the center of the PR, crossed the midline, and linked both the hemispheres (Figures 1B3, 2A). Volume of CB is 2.34 × 10⁴ μm³, about 0.41% of the whole-brain neuropil (Figures 2B,C). The lateral accessory lobe (LAL) was also visible, located laterally to the mushroom body lobes (Figures 1B3, 2A). Volume of LAL is 13.84 × 10⁴ μm³, about 2.47% of the whole-brain neuropil (Figures 2B,C). The protocerebral bridge (PB) was located posteriorly, on either side of middle line of the PR (Figures 1B4, 2A). Volume of PB is 1.53 × 10⁴ μm³, about 0.27% of the whole-brain neuropil (Figures 2B,C). The optical lobe (OL) was located on most lateral side of the PR (Figures 1B5, 2A). Volume of OL is 18.60 × 10⁴ μm³, about 3.33% of the whole-brain neuropil (Figures 2B,C). In addition to these prominent neuropils mentioned above, the PR also contained a large neuropil, referred as midbrain (MBr), which has homogeneous intensity of immunoreactivity without obvious boundaries. Volume of MBr is 397.28 × 10⁴ μm³, about 70.57% of the whole-brain neuropil (Figures 2B,C). The antennal lobe (AL), a spherical structure of the DE, was located most anteriorly of the brain (Figures 1B1, 2A). The volume of AL is 14.38 × 10⁴ μm³, about 2.57% of the whole-brain neuropil (Figures 2B,C). The TR was located most ventral of the brain, at the root of circumesophageal connective, which linked the brain and the gnathal ganglion (Figures 1B1–B3, 2A). Volume of the TR is 54.74 × 10⁴ μm³, about 9.72% of the whole-brain neuropil (Figures 2B,C).

The serotonergic neurons revealed by the immunoreactivity to antiserotonin serum had their cell bodies in the cell body layer and neural fibers projected in wide regions of the neuropils and crossed the midline forming six commissures linking both the hemispheres (Figures 1C1–C5). All the identified cell bodies were counted and their diameters and volumes were measured (Figures 2D–F). There are about 40 serotonergic neurons in the brain (Table 1). The cell body cluster of PR-M was located in the medial region of the posterior PR and contained 18 cell bodies, nine in each hemisphere. Diameters of these cell bodies are in the range of 12.19–16.66 μm and volumes are in the range of 407–1,531 μm³. Serotonergic neurons of PR-M were bilateral and extended widespread projection to the neuropils of the both hemisphere protocerebra via the commissures 1–3 (Figures 2G,H, 3). Innervation regions of their terminals include posterior optic tubercle (POTU), anterior ventrolateral protocerebrum (AVLP), superior intermediate protocerebrum (SIP), superior medial protocerebrum (SMP), inferior medial protocerebrum (IMP), and CB (Figure 3). However, the neuropils of PB, CA, PED, α and β lobes, and lateral horn lack serotonergic neuron processes.

The cell body of PR-L was located in the lateral region of the lateral PR, only one in each hemisphere (Figure 2D). Diameters of these cell bodies are in the range of 17.31–20.01 μm and volumes are in the range of 2,199–3,089 μm³ (Figures 2E,F). Its axon projected to ipsilateral regions of optic lobe, posterior lateral protocerebrum (PLP), and superior lateral protocerebrum (SLP) (Figures 4A–C and Table 1). There were 10 labeled cell bodies in the cluster of PR-LD, located dorsally to the lateral PR (Figure 2D). Diameters of these cell bodies are in the range of 9.15–13.81 μm and volumes are in the range of 431–1,162 μm³ (Figures 2E,F). There were 4 labeled cell bodies in the cluster of PR-LA, located anteriorly to the lateral PR (Figure 2D and Table 1). Diameters of these cell bodies are in the range of 9.46–10.46 μm and volumes are in the range of 512–550 μm³ (Figures 2E,F). No visible processes were observed from PR-LD and PR-LA. Two LALs were also innervated by serotonin neurons and linked by commissure 4 (Figures 2G,H, 4D–F); however, the cell bodies for these serotonergic neurons were unable to be traced.

The cluster of DE-L contained only one cell body located laterally to the AL on each hemisphere (Figure 2D and Table 1). Diameters of these cell bodies are in the range of 15.37–16.91 μm and volumes are in the range of 1,213–1,728 μm³ (Figures 2E,F). The neurites of DE-L ran dorsoposteriorly into the ipsilateral PR via medial antennal lobe tract, crossed the midline via commissure 5, and then projected via contralateral medial antennal tract to the contralateral AL. In addition, this neuron also gave rise to arborizations in both the hemispheres of SLP (Figures 2G,H, 5A–D and Table 1).

There were four cell bodies in TR-A of both the hemispheres (Figure 2D). Diameters of these cell bodies are in the range of 10.92–12.57 μm and volumes are in the range of 440–715 μm³ (Figures 2E,F). No visible processes, however, were observed from TR-A. The TR, indeed, contained the processes of serotonergic neurons, which were from the neurons in the gnathal ganglion via the circumesophageal connective (Figure 5E). In addition, a single serotonergic neuron passing the TR, projected upward along the medial side, and then crossed the midline giving off arborizations in the contralateral superior medial PR (Figures 2G,H, 5E,F and Table 1).

The gnathal ganglion was composed of three neuromeres: the mandibular neuromere (MdNe), the maxillary neuromere (MxNe), and the labial neuromere (LbNe) (Figures 6A–E). The average volume of three neuromeres were about 87.92–105.17 × 104 μm³ (Figure 6F). There were about 24 cell bodies of serotonergic neurons distributed in several clusters of GNG-AD, GNG-AV, GNG-M, and GNG-L (Figures 6A–D,G).

GNG-AD contained seven cell bodies, situated in the medial region of anterodorsal gnathal ganglion. Diameters of these cell bodies are in the range of 8.87–26.27 μm and volumes are in the range of 965–7,847 μm³ (Figures 6H,I). These neurons sent their arborization into the TR. GNG-AV contained two cell bodies, located in the anterior region of ventral gnathal ganglion. Diameters of these cell bodies are in the range of 9.34–10.79 μm and volumes are in the range of 758–1,038 μm³ (Figures 6H,I). GNG-M contained three cell bodies, located in the median area of ventral gnathal ganglion. Diameters of these cell bodies are in the range of 3.53–4.17 μm and volumes are in the range of 305–381 μm³ (Figures 6H,I). No visible processes were observed from neurons of GNG-AV and GNG-AD.

Three clusters, GNG-L1, GNG-L2, and GNG-L3, were located in the lateral cell body layer to the MdNe, the MxNe, and the LbNe, respectively. Each cluster contained four cell bodies with two on each hemisphere. Diameters of these cell bodies are in the range of 8.56–14.41 μm and volumes are in the range of 607–1,799 μm³ (Figures 6H,I). These neurons sent their thick axons to the contralateral hemisphere via the commissure, gave off many fine arborizations, and then projected upward to the TR through the circumesophageal connective (Figures 6J,K and Table 1).

There are about 40 serotonergic neurons in the brain and about 24 serotonergic neurons in the gnathal ganglion of S. frugiperda larvae. The similar numbers were also reported in other larvae species, moths M. sexta (Granger et al., 1989; Griss, 1989), H. armigera (Tang et al., 2019), flies Drosophila melanogaster (D. melanogaster) (Vallés and White, 1988; Huser et al., 2012), Calliphora erythrocephala and Sarcophaga bullata (Nässel and Cantera, 1985), and the beetle Tenebrio molitor (Breidbach, 1990). In adults of M. sexta and D. melanogaster, the number of serotonergic neurons in the gnathal ganglion is also about 20 (Vallés and White, 1988; Homberg and Hildebrand, 1989a; Sitaraman et al., 2008). Similarly, in the central brain of adults of these species, e.g., brain neuropils excluding the optic lobe, the number of serotonergic neurons is about 40 in M. sexta, D. melanogaster, the honeybee Apis mellifera (A. mellifera), the wasp Trichogramma evanescens (T. evanescens), and the blood-feeding bug Triatoma infestans (Schürmann and Klemm, 1984; Lange et al., 1988; Vallés and White, 1988; Homberg and Hildebrand, 1989a; Sitaraman et al., 2008; van der Woude and Smid, 2017). No cell bodies of serotonergic neurons were found in the optic lobe of S. frugiperda. The similar results were also observed in H. armigera (Tang et al., 2019). However, in adults, there are about 600 serotonergic neurons located in the optic lobe of M. sexta, 80 in armyworm Mythimna separata, 100 in butterfly Mimathyma schrenckii (M. schrenckii), 40 in D. melanogaster, 40–60 in A. mellifera, and 120 in mantis Tenodera sinensis (Schürmann and Klemm, 1984; Vallés and White, 1988; Homberg and Hildebrand, 1989b; Leitinger et al., 1999; Niu et al., 2004; Sitaraman et al., 2008; Guo et al., 2018). The differentiation of serotonergic neurons in the optic lobe is dependent on the development of optic lobe neuropils during the metamorphosis (Nässel et al., 1987).

Cell body sizes of serotonergic neurons of S. frugiperda larvae were also measured. Cell bodies in diameters are about 9–20 μm in the brain and 3.5–26 μm in the gnathal ganglion. In T. evanescens, the diameters of cell body sizes of serotonergic neurons are about 2 μm, while in A. mellifera, the diameters of cell body sizes of serotonergic neurons are 8–30 μm (Schürmann and Klemm, 1984; van der Woude and Smid, 2017). Therefore, the size of serotonergic neurons in different species may be related to insect body sizes (van der Woude and Smid, 2017). Within a species, however, why the size of serotonergic neurons differ from different clusters is not clear. The size of cell body may be related to the size of arborization. For example, the cell bodies of PR-M, PR-L, and DE-L in S. frugiperda larvae are larger and their arborizations are spread wider. The size of cell body may also be related to the function of neurosecretion. Four serotonergic neurons of GNG-AD located in the medial gnathal ganglion are very large. Similar results were also found in locust Schistocerca gregaria, cockroach Periplaneta americana, larval H. armigera, and larval and adult M. sexta (Bishop and O’Shea, 1983; Tyrer et al., 1984; Griss, 1989; Homberg and Hildebrand, 1989a; Tang et al., 2019). Intracellular recordings from such neurons of M. sexta larvae revealed overshooting soma spikes of large amplitude and long duration, which suggest that these neurons are neurosecretory cells (Griss, 1989).

In the PR, the cell bodies of serotonergic neurons of larval S. frugiperda are distributed mainly in four clusters: PR-M, PR-L, PR-LD, and PR-LA. The processes of PR-LD and PR-LA are invisible; probably, they are not developed yet in the present stage. In contrast, PR-M has processes projecting to wide region in the PR, including the CB, bilateral areas of superior intermediate protocerebra, superior and inferior medial protocerebra, anterior ventrolateral protocerebra, and POTU. The arborizations in these areas are quite dense. The neurons in the cluster of PR-L have processes projecting mainly to the ipsilateral posterior and superior lateral protocerebra. A few processes of these neurons project to the inner part of the ipsilateral optic lobe. Genetic manipulations demonstrated that a subset of serotonergic neurons in the anterior, medial, and lateral PR evoked hunger (Albin et al., 2015) and several serotonergic neurons in the inferior PR and the lateral PR inhibited the attraction of ethanol in adult D. melanogaster (Xu et al., 2016). Many processes were also found in both the LALs of larval S. frugiperda; however, their cell bodies were unable to be traced. The serotonergic processes in the LAL of H. armigera originated from the cells in the cluster of PR-A, which located in anterior region of the PR (Tang et al., 2019). The patterns of serotonergic neuron arborizations of PR-M and PR-L are similar between S. frugiperda and H. armigera. The neuropils of the PR of larval S. frugiperda, including PB, CA, PED, α and β lobes, and lateral horn, lack serotonergic neuron processes. Similar findings have been reported in larvae of H. armigera, M. sexta, and D. melanogaster (Granger et al., 1989; Huser et al., 2012; Tang et al., 2019). In contrast, the mushroom bodies of adult M. sexta and D. melanogaster contain fine serotonergic neuron processes (Homberg and Hildebrand, 1989a; Sitaraman et al., 2008). These results suggest that some serotonergic neurons in the mushroom body are remodeled during the metamorphosis from the larva to the adult. The mushroom bodies of insects are related to learning and memory activities and serotonin have been demonstrated to be involved in olfaction and place learning and memory in D. melanogaster (Sitaraman et al., 2008, 2012). The lack of serotonergic neuron in the larval mushroom body, however, does not suggest that serotonin plays no role in learning and memory at larval stage. Recently, a serotonin receptor, 5-HT7, was found expressing in the mushroom body of Drosophila larvae, which was shown to mediate the associative olfactory appetitive learning and memory (Huser et al., 2017; Ganguly et al., 2020). Whether serotonin mediates the associative learning and memory for S. frugiperda in the same manner could be investigated by using molecular methods in future study.

A pair of deutocerebral serotonergic neurons DE-L of larval S. frugiperda has arborizations in the contralateral AL and bilateral superior lateral protocerebra, which was similar to the reports in the larvae of M. sexta and H. armigera (Kent et al., 1987; Tang et al., 2019). In adult of M. sexta, the branching pattern persists and expands in the AL with the development of glomeruli (Kent et al., 1987). The similar arborization patterns between larvae and adult may indicated DE-L neurons that play the same function in both the different life stages. Electrophysiological recordings demonstrated that the deutocerebral serotonergic neuron showing responses to odorants and mechanical stimuli in adult moths Bombyx mori (B. mori) and Helicoverpa assulta (Hill et al., 2002; Zhao and Berg, 2009). In D. melanogaster, these two deutocerebral serotonergic neurons could counteract the inhibition of the ethanol attraction from the serotonergic neurons of the PR (Xu et al., 2016). In B. mori, dye-filled DE-L neuron also gave off some arborizations in the lateral accessory lobe, but such innervation pattern was not found in larval S. frugiperda or other studied lepidopteran species (Dacks et al., 2006; Zhao and Berg, 2009; Tang et al., 2019).

The cells in the cluster of TR-A in the TR of larval S. frugiperda were also similar to that of H. armigera and other species (Nässel, 1988; Granger et al., 1989; Wegerhoff, 1999; Tang et al., 2019). The cell bodies were weakly stained with antiserotonin serum and their neuronal processes were not detected. Throughout the TR, however, serotonergic neuron processes are abundant and they may originate from the frontal ganglion and the gnathal ganglion. Two neurons, which linked the TR, form a commissure in the frontal of the medial PR and give off some arborizations in the SMP. The TR is the stomatogastric center. The findings of the serotonergic neuron network between the TR, the PR, and the gnathal ganglion could facilitate us to explore the roles of serotonin in feeding, for instance, food detection, food intake, and nutrient choice, and help in finding a potential approach to modulate the feeding behavior of the gluttonous pest and reduce its damage.

The immunoreactivity to antiserotonin serum in the cell cluster of GNG-AV and GNG-M of larval S. frugiperda was weak and their processes were unable to be traced. The cluster of GNG-AD gave off processes in the most anterior of the gnathal ganglion and projected to the TR. All the three neuromeres of the gnathal ganglion of larval S. frugiperda contain widespread processes of serotonergic neurons originated from the cells in the cluster GNG-L. The thick processes from the cell clusters on both the sides form a horseshoe pattern, cross the midline via a commissure, and project anteriorly to the contralateral TR. Such neurons and their branching patterns show high conservation across insect taxa, which were also found in larvae of H. armigera, M. sexta, Tenebrio molitor (T. molitor) and the flies D. melanogaster, Conistra erythrocephala (C. erythrocephala), and Sarcophaga bullata (S. bullata) (Nässel and Cantera, 1985; Griss, 1989; Breidbach, 1990; Huser et al., 2012; Tang et al., 2019).

In adults of insect species, serotonergic neurons and their processes in the gnathal ganglion were also found in the similar patterns (Bishop and O’Shea, 1983; Tyrer et al., 1984; Rehder et al., 1987; Vallés and White, 1988; Griss, 1989; Homberg and Hildebrand, 1989a; Breidbach, 1990). Previous studies demonstrated that the serotonergic neurons in the lateral side of the gnathal ganglion facilitated the food ingestion of D. melanogaster larvae (Schoofs et al., 2018). In addition to the internal regulation, the serotonergic neurons of the gnathal ganglion were also involved in mediating taste detection (Yao and Scott, 2021). One class of serotonergic neurons in the gnathal ganglion responds to gustatory detection of sugars and the other class to gustatory detection of bitter compounds (Yao and Scott, 2021). As in other species, larval S. frugiperda possesses taste sensilla on the maxilla, responding to the stimuli sugar and bitter substances to regulate the feeding preference (Hou et al., 2020). How the serotonergic neurons in the gnathal ganglion of larval S. frugiperda regulate the feeding preference would be an interesting issue in the future study.