Neuroligin 1 expression is linked to plasticity of behavioral and neuronal responses to sex pheromone in the male moth Agrotis ipsilon

In most animal species, including insects, the sense of smell plays a crucial role in many aspects of life, such as mate attraction and recognition, navigation, and location of predators and food sources (Brennan, 2010; Galizia and Sachse, 2010; Gadenne et al., 2016). The detection and processing of odors, at the neuronal and behavioral levels, are heavily influenced by extrinsic factors, including photoperiod, temperature or previous chemosensory experiences, but also by intrinsic factors such as the physiological status of the individual (Palmer and Kristan, 2011; Gadenne et al., 2016; Ross and Fletcher, 2019). The neuronal plasticity occurs through the olfactory pathway, from olfactory receptor receptors (ORNs) to the olfactory centers, and relies, in part, on short-term and long-term structural and/or functional alterations in synaptic architecture (Wilson et al., 2004).

Synapses contain cell-adhesion molecules (CAMs) that orchestrate the bidirectional organization of their pre-synaptic and post-synaptic compartments (Dalva et al., 2007). Among CAMs, neuroligins (NLGs) constitute a multigene family encoding post-synaptic transmembrane proteins, which are highly conserved across a wide range of species and even between vertebrates and insects (Knight et al., 2011). The NLG family is composed of four members (NLGN1, NLGN2, NLGN3 and NLGN4) in mammals, and an additional NLGN5 gene (also known as NLGN4Y) has been characterized on the Y chromosome in humans (Ichtchenko et al., 1995, 1996; Bolliger et al., 2001). Automated annotation of sequenced insect genomes led to the identification of six NLG genes (Nlg1-5), including two paralogous Nlg-4 genes in the silkworm Bombyx mori (Tsubota and Shiotsuki, 2010), five NLG genes (NLG1-5) in the honey bee Apis mellifera (Biswas et al., 2008) and four NLG genes (dnlg1-4) in the fruit fly Drosophila melanogaster (Banovic et al., 2010; Sun et al., 2011). NLGs are type I transmembrane proteins composed of a large N-terminal extracellular acetylcholinesterase (AChE)-like domain, a single transmembrane domain and a short cytoplasmic tail that contains a post-synaptic density 95/disc-large/zona occludens (PDZ) domain-binding motif (Ichtchenko et al., 1995, 1996) and is exposed to a variety of intracellular regulatory signals (Kim and Sheng, 2004; Meyer et al., 2004). Through their AChE-like domain, NLGs may also interact as ligands with either α-neurexins or β-neurexins, which are CAMs located pre-synaptically, in order to form trans-synaptic bridges that mediate the precise apposition of pre-synaptic and post-synaptic membranes (Varoqueaux et al., 2004; Boucard et al., 2005). A spatiotemporal recognition code between distinct NLGs and neurexins isoforms exists, which is presumed to govern the specification of synaptic connectivity and, thus, to control the balance between excitatory and inhibitory synapses (Prange et al., 2004; Levinson and El-Husseini, 2005; Chih et al., 2006; Siddiqui et al., 2010). The emphasis of NLGs in the development and functioning of neuronal networks in brain has been highlighted in humans, as specific NLG mutations were found to be linked to several neurobiological disorders, such as autism and schizophrenia (Chih et al., 2004; Talebizadeh et al., 2006; Yu et al., 2017). In mammals, the tissular and subcellular expression patterns, as well as the synaptic functions of NLGs, differ according to the protein type (Varoqueaux et al., 2004; Chih et al., 2005; Budreck and Scheiffele, 2007). NLGN1 was reported to bind tightly to β-neurexins, but not to α-neurexins (Ichtchenko et al., 1995), and was also shown to be expressed in both the peripheral and central nervous system, with a preferential localization to excitatory glutamatergic synapses (Song et al., 1999; Chih et al., 2005), as well as in a variety of glial cells, including astrocytes and oligodendrocytes (Scheiffele et al., 2000; Sakers and Eroglu, 2019). Accumulating in vitro and in vivo evidence indicates that NLGN1 plays a role in the formation, maturation and activity of synaptic structures (Hu et al., 2015). NLGN1 overexpression in primary hippocampal neuronal cultures increased the number of pre-synaptic and post-synaptic contacts, while NLGN1 silencing by RNA interference led to an inhibition of synaptogenesis (Dean et al., 2003; Chih et al., 2005; Shipman et al., 2011; Burton et al., 2012). In NLGN1 transgenic mice, NLGN1 overexpression provoked an elevation in synapse density, excitatory/inhibitory ratio and synaptic transmission within hippocampus as well as impairments in memory acquisition (Dahlhaus et al., 2010). Concerning insects, it has been demonstrated in A. mellifera that sensory experience- and olfactory learning-induced brain plasticity was accompanied by changes in neuropil volume and synaptic density correlated with increased NLG1 expression (Biswas et al., 2010; Reinhard and Claudianos, 2012). In addition, in D. melanogaster, the loss of dnlg1 function has been associated with neuromuscular defects characterized by a marked reduction in bouton number, synaptic transmission and post-synaptic apparatus, while increased dnlg1 expression in mutants promoted ectopic post-synaptic differentiation (Banovic et al., 2010).

Over the past decade, there has been increasing evidence that the noctuid moths are excellent model organisms for studying the plasticity of olfactory system (Anton et al., 2007; Gadenne et al., 2016). In these species, the sexual communication relies on the attraction of males by sex pheromones produced and emitted by conspecific females. The pheromone signal is recognized by ORNs located in the antennae and is integrated first in the primary olfactory centers of brain, the antennal lobes (ALs), which are compartmentalized into spherical neuropil structures, the olfactory glomeruli (Haupt et al., 2010). The ALs consist of two complexes of glomeruli: the macroglomerular complex (MCG), which specializes in the processing of sex pheromone input from ORNs, and the ordinary glomeruli, which are involved in the processing of all the other odors that the male encounters (Haupt et al., 2010). The processed information is then conveyed via projection neurons (PNs) to higher-order brain centers, the mushroom bodies and the lateral horn, ultimately leading to a characteristic orientation behavior of the male towards the pheromone source that is supported by optomotor anemotaxis (Haupt et al., 2010). ALs also contain local interneurons (LNs), which connect many glomeruli and form a network providing transfer of information between glomeruli. LNs have been characterized as GABAergic, cholinergic and glutamatergic, and they help to structure the odor representation in the AL, which ultimately shapes the tuning profiles of PNs (Haupt et al., 2010).

Despite its stereotypy, the orientated flight behavior and the underlying processing steps exhibit high degrees of flexibility. In addition to modulation by circadian rhythm (Groot, 2014), age-, odor experience- and reproductive status-related plasticities in the male response to sex pheromone have been unveiled in noctuid moths, Spodoptera littoralis and Agrotis ipsilon. Indeed, a brief pre-exposure of S. littoralis males to sex pheromone induced a lowering of the threshold for the behavioral response 24 h later. This long-term effect is correlated with a higher sensitivity of ORNs and AL neurons to pheromone signal, accompanied by a volumetric increase in MGC glomeruli and in the calyces of the mushroom bodies (Anderson et al., 2007; Guerrieri et al., 2012; Anton et al., 2016). It is also known that A. ipsilon males are apt to elicit the sex pheromone-guided orientation flight only when they reach their full sexual maturity within 3–4 days after emergence (Gadenne et al., 1993; Duportets et al., 1998; Gassias et al., 2018). This increase in behavioral responsiveness is paralleled by a raise in the sensitivity of sex pheromone-responding AL neurons (Gadenne and Anton, 2000). By contrast, newly mated males quickly cease to behaviorally respond to sex pheromone from 15 min after the onset of copulation (Vitecek et al., 2013), and this inhibition is concomitant with a decrease in AL neuron sensitivity (Gadenne et al., 2001). This post-mating refractory period lasts until the end of the scotophase, and the males are able to re-mate only during the following scotophase, when they have fully recovered their sensitivity to sex pheromone and also replenished their sex accessory glands to produce a new spermatophore (Vitecek et al., 2013; Gassias et al., 2018).

In order to decipher the molecular mechanisms underlying the modulation of olfactory system in A. ipsilon males, we tested the hypothesis that NLG1 is involved in mediating the plasticity of behavioral and neuronal responses to sex pheromone. We first cloned a full-length cDNA encoding A. ipsilon NLG1 (AiNLG1) and determined its tissue expression profile. We then quantified the level of the expression of AiNLG1 in ALs as a function of age (immature versus mature), of odor experience (naive versus pre-exposed to sex pheromone) and of reproductive status (virgin versus mated) in relation to the pheromone signal responsiveness. Our results demonstrated that age- and odor experience-related increase in sex pheromone responsiveness is timely correlated with AiNLG1 overexpression in ALs, thus suggesting a potential role of NLG1 as central organizer of the olfactory plasticity in A. ipsilon.

We prepared a synthetic sex pheromone blend at 1 ng µl⁻¹ containing three components identified in natural extracts of the pheromone glands of the female Agrotis ipsilon (Hufnagel 1766) (Picimbon et al., 1997; Gemeno and Haynes, 1998): (Z)-7-dodecen-1-yl acetate (Z7-12:OAc), (Z)-9-tetradecen-1-yl acetate (Z9-14:OAc) and (Z)-11-hexadecen-1-yl acetate (Z11-16:OAc), mixed at a ratio of 4:1:4. This blend was proven to be the most attractive to males in field trapping experiments (Causse et al., 1988), and it elicits similar behavior in a wind tunnel as natural extracts of the pheromone gland (Barrozo et al., 2010; Vitecek et al., 2013). The three components were purchased from Sigma-Aldrich (Saint-Quentin Fallavier, France) and diluted in hexane (>98% purity, CAS 110-54-3, Carlo-Erba, Val-de-Reuil, France). A dose of 10 ng of the sex pheromone blend was used in the experiments, and this dose has previously been described as behaviorally and electrophysiologically active (Gadenne et al., 2001; Deisig et al., 2012; Gassias et al., 2018).

Adult A. ipsilon originated from a laboratory colony established in Bordeaux and transferred to Versailles, France. The colony is based on field catches in southern France, and insects were reared on an artificial diet in individual cups until pupation (Poitout and Bues, 1974). Pupae were sexed, and males and females were kept separately on an inversed light:dark cycle (16 h light:8 h dark photoperiod, with the scotophase starting at 10:00 h) at 22°C. Newly emerged adult males and females were removed from the hatching containers every day and reared in a separate room with access to a 20% sucrose solution ad libitum. The day of emergence was considered as day-0. The males were anesthetized with an exposure to carbon dioxide for 10 s, and tissue dissections were performed in the middle of the scotophase (between 13:00 h and 16:00 h), when males respond maximally to the sex pheromone (Barrozo et al., 2010). For the expression profiles of AiNLG1, antennae, ALs, brains with excised ALs, thoracic muscles, midguts, fat bodies and testes of males were dissected in Ringer's solution, then immediately flash-frozen in Eppendorf (Montesson, France) vials kept in liquid nitrogen and stored at −80°C until treatment.

Total RNA was extracted using TRIzol reagent (TRI Reagent^®, Euromedex, Souffelweyersheim, France), according to the manufacturer's instructions, and quantified by spectrophotometry at 260 nm. DNase treatment was performed with 2 U TURBO™ DNase 1 (Ambion, Villebon-sur-Yvette, France) for 30 min at 37°C, followed by a 10-min inactivation at 75°C. After DNase treatment, single-stranded cDNA was synthesized from total RNAs (1 µg) with SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions, with a deoxynucleotide triphosphate (dNTP) mix, RNase OUT, Oligo(dT) primer and sterile water to a final reaction volume of 20 µl. The mix was heated to 65°C for 5 min before the enzyme was added and then incubated for 1 h at 42°C.

The amino acid sequence of Helicoverpa armigera NLG1 was searched against the genome assembly of A. ipsilon in the NCBI database (accession number GCA_004193855.1), and a specific primer pair was designed based on the sequence of a putative AiNLG1 transcript. Polymerase chain reaction (PCR) was carried out on 200 ng brain cDNA with 2.5 U High Expand Fidelity DNA polymerase (Roche, Saint-Quentin Fallavier, France), 0.4 µmol l⁻¹ primer pair (NLG1dir, NLG1rev) and 0.25 mmol l⁻¹ of each dNTP. Following an initial 5-min denaturation at 94°C, the thermal amplification procedure consisted of 30 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, elongation at 72°C for 1 min and then final elongation at 72°C for 10 min. To obtain the full-length cDNA sequence of AiNLG1, 5′- and 3′-rapid amplification of cDNA ends (RACE) was conducted using a SMART RACE cDNA Amplification Kit (Clontech, Mountain View, CA, USA) following the manufacturer's instructions. For 5′-RACE, we used the specific reverse primers (NLG15′-RACE1, NLG15′-RACE2) and Universal Primer Mix (UPM; Clontech) as the forward anchor primer; for 3′-RACE, we used UPM as the reverse primer and the specific forward primers (NLG13′-RACE1, NLG13′-RACE2). Touchdown PCR was performed using hot start as follows: 1 min at 94°C, five cycles of 30 s at 94°C, 30 s at 64°C and 1 min at 72°C, then five cycles of 30 s at 94°C, 30 s at 62°C and 1 min at 72°C, then 25 cycles of 30 s at 94°C, 30 s at 60°C and 1 min at 72°C, then 15 min at 72°C. Specific primers used for reverse transcription (RT)-PCR and RACE reactions are listed in Table S1.

RT-PCR and RACE products were purified by agarose gel electrophoresis (NucleoSpin^® Extract II, Macherey-Nagel GmbH & Co. KG, Düren, Germany) and subcloned into PCRII-Topo plasmid (Invitrogen). After colony isolation, DNA minipreps were prepared (NucleoSpin^® Plasmid DNA Purification, Macherey-Nagel GmbH & Co. KG) and the DNA clone containing the proper insert was then sequenced (GATC Biotech SARL, Marseille, France).

Behavioral tests were performed using a Plexiglass wind tunnel measuring 190 cm length×75 cm width×75 cm height (VT Plastics, Genevilliers, France). Both ends of the tunnel were enclosed with white synthetic fabric, allowing air to pass through but preventing insect escape. Air movement was provided by an exhaust fan at the downwind end of the tunnel sucking the air at a speed of 0.3 m s⁻¹ and evacuating odorized air out of the building. The tunnel was maintained in the dark except for one red visible light source, allowing visual observation. Experiments were performed at 23°C, 40±10% relative humidity, during the middle of the scotophase (between 13:00 h and 16:00 h), when males respond maximally to sex pheromone (Barrozo et al., 2010). A single experimental male was introduced into the wind tunnel and placed on a 36 cm-high platform in the middle of the tunnel width and 160 cm downwind from the pheromone source. After 1 min, during which the male adjusted to the airflow, the pheromone blend was delivered for 3 min. During this time, the following behavioral items were observed: take-off, partial flight (flight half way between the release site and the pheromone source), complete flight (arrival within 10 cm of the source) and landing on the pheromone source. The partial and complete flights as well as landings were considered as an orientated response towards the sex pheromone. Orientated and random flights were counted together in order to quantify the general flight activity of males, and this activity was ∼95% for all behavioral tests. We also noted the delay in the orientated response of males by measuring the period of time between the onset of pheromone stimulation and the take-off. All experiments were performed double blind to avoid partial observations. On each experimental day, different groups of males were tested, including at least one group of males that were expected to show a high response level to avoid experimental bias.

The pheromone stimulus was delivered with a programmable olfactometer adapted from Party et al. (2009). Air from the building was charcoal filtered and re-humidified, then divided into several equal flows (300±10 ml min⁻¹) using a Y connector (model 1/8″ P514, Upchurch Scientific, Silsden, UK) and a 5-port manifold (model P-115, Upchurch Scientific). Each flow was connected to a miniature electrovalve (model LHDA1233115H, The Lee Company, Westbrook, ME, USA) driven by a Valve-Bank programmer (AutoMate Scientific, Berkeley, CA, USA). Activating the appropriate valve directed the flow to a 4 ml glass vial containing the pheromone source and closed by a septum cork. The inlet and outlet of the source pheromone were made of two hypodermic needles (18-g size) inserted through the septum and connected with polytetrafluorethylene tubing (1.32 mm i.d., 20 cm length). A 10 µl sample of the pheromone blend at 1 ng µl⁻¹ was deposited on a filter paper and placed inside the vial. Contaminated air was removed from the set-up by an exhaust fan after each behavioral assay.

Males of different ages (1, 2, 3, 4 and 5 days) were collected in the mid-scotophase and transferred to cylindrical plastic containers in a dark box, then transported from the rearing room to the wind tunnel room for quantification of the behavioral response to sex pheromone and response delay. The males were then killed using liquid nitrogen, and the ALs were dissected from frozen heads for analysis of AiNLG1 expression.

Pre-exposure treatments were performed in the wind tunnel at mid-scotophase. Newly emerged males (day-0) were transferred from the rearing room to the wind tunnel room, and placed in cages (maximum of three males per cage) before the end of the photophase. A cage was introduced into the wind tunnel and placed on a 36 cm-high platform in the middle of the tunnel width and 160 cm downwind from the pheromone source. After 1 min, each male, kept within the cage, was subjected to pheromone stimulation during 3 min, after which it was removed from the wind tunnel and transferred to the rearing room. The behavioral response and response delay of the pre-exposed males were analyzed at 2, 24, 48 and 72 h after pre-exposure to sex pheromone. The males were then killed using liquid nitrogen, and ALs were dissected from frozen heads for quantification of AiNLG1 expression. Control groups were composed of 0-, 1-, 2- and 3-day-old males exposed to hexane for 3 min.

Mating experiments were performed as previously described (Barrozo et al., 2010). Briefly, virgin 4-day-old sexually mature males and virgin 3-day-old sexually mature females were individually paired in cylindrical plastic containers before the onset of scotophase in a room under an inversed light:dark cycle (16 h light:8 h dark photoperiod) at 22°C. In our photoperiod conditions, the scotophase begins at 10:00 h and ends at 18:00 h and copulation occurs, on average, between 13:30 h and 16:00 h, with a mean duration of 1 h 30 min (Vitecek et al., 2013). In A. ipsilon males, the post-mating refractory period, which is characterized by a transient inhibition of behavioral and central nervous responses to sex pheromone, starts at 15 min after the onset of copulation and lasts up to the end of the scotophase (Gadenne et al., 2001; Vitecek et al., 2013; Gassias et al., 2018).

Visual observation of the pairs was done every 5 min during the copulation period in order to detect the onset and the end of the mating. Once copulation had ended, newly mated males were removed from the observation room. The behavioral response to sex pheromone of males and delay response were analyzed immediately (0 h) and at 1 h and 23 h after the end of the copulation. The males were then killed using liquid nitrogen, and ALs were dissected from frozen heads for quantification of AiNLG1 expression. Control groups consisted of virgin 4- and 5-day-old males, which were paired with virgin 3-day-old females but not allowing them to mate. In order to confirm that mating was successful, all mated females were also dissected to check for the presence of the spermatophore.

The protein sequences of insect NLG homologs were aligned using ClustalW in BioEdit 7.0. The signal peptide and the transmembrane domain were predicted with SignalP 4.1 (https://services.healthtech.dtu.dk/service.php?SignalP-4.1) and TMHMM 2.0 (https://services.healthtech.dtu.dk/service.php?TMHMM-2.0). The AChE-like domain was found using the NCBI Conserved Domain Architecture Retrieval Toll (http://www.ncbi.nlm.nih.gov/Structure/lexington/lexington.cgi). PDZ domain-binding motif was predicted with PDZPepInt (Kundu et al., 2014) and EF-hand metal-binding motifs were obtained from Biswas et al. (2008). The phylogenetic tree of insect NLGs was constructed by the neighbor-joining method with Poisson correction of distances using the MEGA 7 program (Kumar et al., 2016) with default settings.

Accession numbers for NLG protein sequences retrieved from GenBank are as follows: the lepidopteran H. armigera NLGs (HaNLG1, XP_021194794.1; HaNLG2, XP_021189212.1; HaNLG3, XP_021197425.1; HaNLG4-XL, XP_021196129.1; HaNLG4-YL, XP_021196885.1; HaNLG5, XP_021191347.1); the dipteran D. melanogaster NLGs (DmNLG1, NP_731172; DmNLG2, NP_523496; DmNLG3, NP_001036685; DmNLG4, NP_001163661); the coleopteran Tribolium castaneum NLGs (TcNLG1, XP_008192968.1; TcNLG2, XP_015835928.1; TcNLG3, XP_001810887.1; TcNLG4, EFA03840; TcNLG5, EFA03175); and the hymenopteran A. mellifera NLGs (AmNLG1, ACM48186; AmNLG2, XP_006571554.2; AmNLG3, NP_001139208; AmNLG4, NP_001139209; AmNLG5, NP_001139211).

The means of AiNLG1 expression level and of delay response were compared using one-way ANOVA followed by a Tukey's multiple comparison post hoc test. P<0.05 was accepted as statistically significant. For behavioral experiments, statistical differences between groups of males were evaluated using an R×C test of independence by means of a G-test and also by applying the Williams correction (Sokal and Rohlf, 1995).

Adequate measures were taken to minimize pain or discomfort of animals. Experiments were conducted according to international standards on animal welfare and were compliant with local and national regulations. All applicable international guidelines for the care and use of animals were followed.

By searching the nucleotidic sequence of H. armigera NLG1 against the genome database of A. ipsilon, we designed a pair of specific primers based on the sequence of a putative NLG1 transcript and succeeded in amplifying a partial fragment of 919 bp via RT-PCR on brain total RNA extracted from 5-day-old males. The remaining 5′ and 3′ ends of this fragment were then obtained by RACE-PCR using gene-specific primers. The nucleic acid sequences for the 5′-RACE and 3′-RACE reaction products were assembled with the original fragment to generate a full-length cDNA named AiNLG1, which was deposited in the GenBank database under the accession number MZ423824. The AiNLG1 cDNA extends over 5047 bp and contains a 5′-untranslated region (5′-UTR) of 180 bp, an open reading frame (ORF) of 4401 bp and a 3′-UTR of 466 bp, with a polyadenylation signal upstream of the poly(A) tail (Fig. S1). The AiNLG1 ORF was translated into a predicted protein sequence of 1467 amino acids (Fig. S1), which is composed of the typical functional regions of NLG protein family: an N-terminus cleaved signal sequence, a large extracellular AChE-like domain comprising two EF-hand metal-binding motifs, followed by a single transmembrane domain and a C-terminus PDZ domain-binding motif (Fig. 1). Comparison of the predicted protein of AiNLG1 with NLG1 characterized in other insect species showed a high percentage of amino acid identity with the lepidopteran H. armigera (HaNLG1; 89.5%), followed by the coleopteran T. castaneum (TcNLG1; 28.1%), the dipteran D. melanogaster (DmNLG1; 23.1%) and the hymenopteran A. mellifera (AmNLG1; 20.8%) (Fig. 1). A neighbor-joining phylogenetic tree of insect NLGs was also created, and, as expected, its analysis revealed that AiNLG1 belongs to the cluster of lepidopteran NLG1 (Fig. 2).

In order to gain some preliminary information about the functions of AiNLG1, a tissue distribution screen was performed by evaluating the expression level of AiNLG1 transcript by qPCR in the antenna, ALs, brain with excised ALs, thoracic muscle, midgut, fat body and testis in 5-day-old males. As shown in Fig. 3, AiNLG1 is expressed at higher levels in both ALs and the rest of the brain than in peripheral tissues including the thoracic muscle and midgut. No amplification of the AiNLG1 transcript was detected in the antenna, fat body and testis (Fig. 3).

In order to explore the possible role of AiNLG1 in age-dependent olfactory plasticity, we evaluated the proportion of sex pheromone-guided orientated flight and response delay as well as the level of AiNLG1 mRNA in ALs in 1- to 5-day-old males. The percentage of orientated behavioral response was only ∼14% and 21% in 1- and 2-day-old sexually immature males and significantly raised at the age of 3 days to reach ∼78% and 75% in 4- and 5-day-old sexually mature males, respectively (Fig. 4A). The response delay of 1- and 2-day-old males (97±11 s and 91±9 s, respectively) was higher than that of 4- and 5-day-old males (45±10 s and 39±12 s, respectively) (Fig. 4B). The expression of AiNLG1 mRNA in ALs was detected at day-1, and gradually increased from day-2 until day-4 and then decreased slightly at day-5 (Fig. 4C). The level of AiNLG1 mRNA in ALs of 5-day-old sexually mature males was approximatively threefold higher than that in ALs of 1-day-old sexually immature males (Fig. 4C).

In order to ascertain the possible involvement of AiNLG1 in odor experience-dependent olfactory plasticity, newly emerged males (day-0) were pre-exposed to sex pheromone blend for 3 min, and then the proportion of sex pheromone-guided orientated flight and response delay as well as the level of AiNLG1 mRNA in ALs were determined at 2, 24, 48 and 72 h after sex pheromone pre-exposure. The percentage of orientated behavioral response of males at 24 h (PE-24h), 48 h (PE-48h) and 72 h (PE-72h) following pre-exposure was significantly higher than that of day-1, day-2 and day-3 control males (C1, C2 and C3), respectively (Fig. 5A). This increase in the behavioral responsiveness to sex pheromone was accompanied by a decrease in response delay and a rise in the amount of AL AiNLG1 mRNA in PE-24h, PE-48h and PE-72h males compared with C1, C2 and C3 males, respectively (Fig. 5B,C). By contrast, there was no significant difference in the behavioral response, response delay and AL AiNLG1 mRNA level between PE-2h and CO males (Fig. 5A–C). The quantity of AiNLG1 mRNA in ALs of PE-72h males was ∼1.5-fold higher than that in ALs of C3 males (Fig. 5C).

In order to assess the possible function of AiNLG1 in mating-dependent olfactory plasticity, virgin 4-day-old sexually mature males were individually paired with virgin 3-day-old sexually mature females, and then we determined the proportion of sex pheromone-guided orientated flight and response delay as well as AiNLG1 mRNA expression level in ALs of mated males at the end of the copulation, and at 1 h and 23 h later. The percentage of orientated behavioral response of males at 0 h (PM-0h) and 1 h (PM-1h) post-mating was low in comparison with that of virgin 4-day-old sexually mature males (C4) and then increased sharply at 23 h (PM-23h) to reach that of virgin 5-day-old sexually mature males (C5) (Fig. 6A). PM-0h and PM-1h males responded behaviorally more slowly to sex pheromone than C4 males, whereas PM-23h males responded as quickly as C5 males (Fig. 6B). The AL AiNLG1 mRNA level was not significantly different in PM-0h and PM-1h males compared with C4 males, and in PM-23h compared with C5 males (Fig. 6C).

In this study, we examined whether NLG1 might be a mediator of the plasticity of the olfactory system in insects. In order to test this hypothesis, we turned to the long-living noctuid moth A. ipsilon, a model organism for which the behavioral and neuronal response to the female emitted-pheromone signal are known to be dependent on age, odor experience and reproductive status of the male.

We first succeeded in isolating a single 5047 bp cDNA from the nervous tissue of the adult male A. ipsilon, and this clone was found to encode a predicted AiNLG1 protein that exhibits all the signature motifs of the NLG family. The large extracellular AChE-like domain of AiNLG1 has sequence homology to the typical α/β-hydrolase fold structure of the carboxyl-cholinesterase superfamily, but it is devoid of the active site residues for enzymatic activity (Varoqueaux et al., 2004; Boucard et al., 2005) and contains two EF-hand metal-binding motifs, which are Ca²⁺-binding sites that are presumed to be essential for the interaction of NLG with its cognate receptor neurexin. The C-terminus PDZ domain-binding motif of AiNLG1 is known to be critical for synaptic protein recruitment through the binding of NLG to PDZ domain-containing scaffolding proteins, which, in turn, cluster together post-synaptic transmitter receptors, ion channels and signaling proteins (Irie et al., 1997; Kim and Sheng, 2004; Meyer et al., 2004). Finally, the overall protein sequence of AiNLG1 was found to exhibit a degree of similarity with NLG1 proteins previously identified in the lepidoptera H. armigera, with more than 89% identity, and also those identified in other insect orders including Diptera, Hymenoptera and Coleoptera. These homology sequence data were further supported by our phylogenetic analysis demonstrating that AiNLG1 belongs to the cluster of the lepidopteran NLG1s. In addition, our phylogenetic tree revealed that NLG sequences diversified independently during evolution in insects from a common ancestral NLG sequence as observed for vertebrate NGLs (Knight et al., 2011). It is also interesting to note that two Nlg4 forms have been characterized in H. armigera (Pearce et al., 2017) as well as in B. mori (Tsubota and Shiotsuki, 2010), Plutella xyllostella (You et al., 2013) and Danaus plexippus (Zhan and Reppert, 2013), indicating the existence of a duplication event of the Nlg4 gene that would have recently occurred in the Lepidoptera lineage.

As for all the members of the NGL1 subfamily, the transcriptional activity of AiNLG1 is tissue dependent (Chih et al., 2005; Biswas et al., 2008; Durand et al., 2016), measured here in the brain and peripheral organs at adult stage. Elevated levels of the AiNLG1 transcript were detected more specifically in ALs, the analogs of the vertebrate olfactory bulb, thus suggesting a possible involvement of the AiNGL1 protein in the proper assemblage and activity of neuronal networks within the primary olfactory neuropiles in A. ipsilon. This is consistent with previous reports of NLGN1 mRNA abundance in the mitral cells of the rat olfactory bulb, which are functional equivalents of insect PNs, during early post-natal development (Clarris et al., 2002). We cannot rule out the possibility that AiNLG1 also acts in other brain areas, as evidenced by the observation that its mRNA was detected in brain after excision of the ALs. This speculation is supported by in situ hybridization experiments performed in A. mellifera demonstrating that the expression of NLGs, and especially AmNLG3, is not restricted to ALs, but is also present at high levels within higher-order brain regions, and more precisely in the Kenyon cells of mushroom bodies (Biswas et al., 2008), which are known to be essential for multisensory integration as well as learning and memory. We plan in future studies to perform in situ mapping of AiNLG1 expression in the entire brain of adult A. ipsilon.

It is interesting to note that, in addition to its expression in the brain, AiNLG1 mRNA is present in the midgut. It is well established that the general organization and the roles of the enteric nervous system (ENS) of insects are clearly analogous to those of the vertebrate ENS. Thus, the insect ENS is formed by interconnected networks of ganglia and diffuse nerve plexuses that extend superficially on the layers of longitudinal and circular striated muscles surrounding the gut epithelium, and many of the vital functions of this organ, including rhythmic propulsive movements, ingestion, digestion, absorption and defecation, are controlled by the activity of the intrinsic neurons of the ENS (Copenhaver, 2007). In the human ENS, NLGN1 is enriched within the post-synaptic boutons of excitatory glutamatergic synapses that are inserted between the longitudinal and circular muscle layers of the intestine. Reduced abundance of NLGNI was associated with altered synaptic development, leading to intestinal contractile dysfunction characterized by obstructions and the toxic megacolon syndrome in patients with Hirschsprung's disease (Wang et al., 2015). Similar to the action of its human NLGN1 homolog, AiNLG1 might be an actor of the excitatory synaptic transmission within the ENS in the control of gut motility in A. ipsilon males.

We also observed minor expression of AiNLG1 in skeletal muscles of A. ipsilon. Unlike the cholinergic excitatory innervation of the vertebrate skeletal muscle, the insect skeletal muscle is innervated by glutaminergic excitatory motor neurons and may receive an additional nerve supply by GABAergic inhibitory neurons and axonal terminals of neuromodulatory neurons including octopaminergic neurons (Pflüger and Duch, 2011; Wolf, 2014). In D. melanogaster, it has been observed that the neuromuscular junctions establish on all muscles of dnlg1 mutants in a usual timing, but exhibit marked deficits in synapse density, synaptic transmisssion and post-synaptic apparatus, including scaffolding proteins and glutamate receptors, and that restoration of dlng1 expression in mutants induced an ectopic post-synaptic differentiation (Banovic et al., 2010). It seems reasonable to hypothesize that, like Drosophila dlgn1, AiNLG1 is probably not required for initiating the formation of the excitatory glutaminergic neuromuscular synapses but is rather indispensable for their functional maturation by promoting the post-synaptic specializations.

Finally, this tissue-specific gene expression analysis highlights that AiNLG1 may potentially serve as an organizer for central and peripheral excitatory trans-synaptic signaling involved in the functioning of the olfactory system, as well as intestinal and neuromuscular motor control, in A. ipsilon males.

We noticed that AiNLG1 mRNA in ALs is at higher amounts in old sexually mature males than in young sexually immatures males. This result points to the existence of an age-dependent upregulation of AiNLG1 expression within ALs, which is concomitant with the age-dependent increase in neuronal and behavioral responses to the pheromone signal. Such a correlation led us to assume that increased AiNLG1 expression reflects an anatomical and functional shaping of synaptic connections, notably through the maturation of excitatory glutaminergic synapses that takes place within AL glomeruli, and thus facilitates the processing of pheromone information and the emergence of orientated flight behavior. In concordance with this assumption, it has been demonstrated in male rats that the expression of the NLGN1 protein in neocortex, and especially within the primary olfactory cortex, is timely upregulated with the synaptogenesis accompanying the pubertal maturation of reproductive behavior (Song et al., 1999). In order to prove a functional involvement of AiNLG1 in age-related plasticity of sex pheromone responsiveness, we will evaluate the effects of its knockdown using RNA interference technology, which has been shown to be highly efficient for gene silencing in A. ipsilon brain, on both the activity of PNs in the MCG by intracellular recordings and the behavioral response to the pheromone signal. In line with this experimental perspective, nlg-1 mutants of Caenorhabditis elegans were found to have deficits in the processing of chemosensory cues (Hunter et al., 2010).

In mammals, it is known that steroid hormones play a major role in fetal and neonatal brain development by controlling many neurodevelopmental processes, including apoptosis, growth, neuronal differentiation and synaptogenesis, via the activation of GABAergic and glutaminergic systems (González et al., 2007; McCarthy, 2008; Konkle and McCarthy, 2011). A series of recent reports has shown that NLGs are steroid-regulated genes (Kang et al., 2004; Bethea and Reddy, 2012) and that altered prenatal steroidogenic activity is associated with a decline in the brain NLG expression levels, thus contributing to the appearance of autistic disturbances in childhood (Sellers et al., 2015; Baron-Cohen et al., 2020). In A. ipsilon, growing evidence has accumulated these past years about the involvement of a steroidal system, and in particular 20-hydroxyecdysone (20E), the main insect steroid hormone, in male sexual development. Indeed, it has been uncovered that 20E is necessary to enhance high sensitivity in sex pheromone responsiveness, and that this stimulatory action is at least part transduced by the binding of 20E to the Ecdysone Receptor (EcR)/Ultraspiracle (USP) complex in ALs and, ultimately, engenders the upregulation of synaptic plasticity-related proteins such as synaptotagmins (Duportets et al., 2013; Abrieux et al., 2014; Bozzolan et al., 2015). Other interesting data are that age-dependent increases in the circulating level of 20E match fluctuations in AiNLG1 expression amount in ALs (Vitecek et al., 2013). By taking into account the above-mentioned data, it seems reasonable to suggest the existence of a possible functional interplay between the 20E- and NLG1-initiated signaling pathways that operate in ALs. Thus, we propose that the activation of 20E–EcR–USP signaling in response to the elevation of circulating 20E level is responsible for the increased AiNGL1 expression, which is probably linked to the synaptic plasticity occurring in ALs during male sexual maturation in A. ipsilon. To verify whether AiNLG1 is a potential molecular target of the 20E genomic signaling pathway, we will examine the impact of 20E on the transcriptional activity of AiNGL1 in ALs by injecting 20E into young sexually immature males, in which the circulating hormone level is low, or by removing the testes, which are recognized as the primary organs for steroidogenesis in male moths (Loeb et al., 1988; Yamamoto et al., 2017), in old sexually mature males.

We observed that young sexually immature males displayed an increase in their behavioral response to sex pheromone at 24, 48 and 72 h after a brief exposure to this odor, and this long-term effect was associated with a pronounced expression of the AiNLG1 mRNA in ALs compared with that observed in naive males. Histological and neurophysiological experiments conducted on S. littoralis have unveiled that the increased behavioral response of males to sex pheromone after odor pre-exposure coincided with an increase in both the sensitivity of AL neurons and the size of MCG glomeruli (Anton et al., 2016). It has been uncovered that the expansion of AL glomeruli in young Drosophila females through olfactory pre-exposure is in part due to an increase in the number of neuronal branching and synaptic contacts (Devaud et al., 2001, 2003). Taken together, all the above data lead us to assume that the short pre-exposure of young Agrotis males to the female sex pheromone induces early AiNLG1 overexpression within AL glomeruli, thus triggering NLG1-mediated signaling that participates in the shaping of synaptic architecture and consequently promoting pheromone information processing and the expression of orientated flight behavior. This assumption is strengthened by earlier data in A. mellifera demonstrating that sensory experience- and odor associate learning-driven brain plasticity is accompanied by upregulation of AmNLG1 expression (Biswas et al., 2010; Reinhard and Claudianos, 2012).

A recent study reported that, in young worker bees exposed to queen mandibular pheromone (QMP), QMP-evoked physiological and behavioral effects are in part initiated by augmentation in the circulating level of ecdysteroids, including 20E, via regulation of biosynthesis rate (Trawinski and Fahrbach, 2018). An interesting observation in A. ipsilon is that young males pre-exposed to sex pheromone exhibit higher amounts of 20E in both hemolymph and brain than do naive males (E. Gassias, unpublished data). This raises the possibility that sex pheromone exposure evokes an elevation of circulating 20E that might result in the activation of AL NLG1-mediated signaling via 20E-induced NLG1 overexpression, thus contributing to the emergence of sex pheromone-guided flight behavior in young A. ipsilon males. In order to ascertain the existence of such an interaction between 20E and NLG1 signals through early olfactory experience, in future investigations we will examine whether 20E deprivation, through chirurgical castration, prevents the sex pheromone-induced effects on the level of AiNLG1 expression in ALs as well as on the sex pheromone behavioral response in pre-exposed males.

As in many other species, A. ipsilon males have a post-ejaculatory refractory period, preventing them from re-mating until the next scotophase, by which time they replenish their sex accessory glands for a potential new ejaculate (Gadenne et al., 2001; Vitecek et al., 2013). This transient inhibition of male sexual behavior results from a decline in the sensitivity of sex pheromone-responding AL neurons: most PNs have much higher pheromone response thresholds after mating (Barrozo et al., 2010). Our results showed that the transcriptional activity of AiNLG1 in ALs of mated males is unchanged at 0, 1 and 23 h after copulation compared with that of virgin males. Thus, there is a high probability that mating-triggered switch off of the central processing of pheromone information during the refractory period, as well as the restoration of sex pheromone responsiveness on the following night, are not driven by alterations in the AL NLG1-initiated trans-synaptic signaling through modulation of AiNLG1 expression. In concordance with this, we observed no significant variation in the amounts of transcripts encoding other synaptic plasticity-related proteins, including NLG3, which is known to be present at both glutamatergic and GABAergic synapses, and synaptotagmin I, in ALs of A. ipsilon mated males (E. Gassias, unpublished data). All these data are also supported by the fact that this mating-related functional plasticity of the central olfactory system occurs rapidly, within 15 min after copulation (Vitecek et al., 2013), and probably without requiring any long-lasting structural modifications (i.e. glomerular size, synapse number) compared with the other forms of olfactory plasticity in relation to sexual development and odor experience in A. ipsilon adult males.

Finally, the present study reveals that, in long-living male moths, the increased behavioral and neuronal sensitivity to sex pheromone through sexual maturation and odor experience is linked to the upregulation of the expression of NGL1 within the primary olfactory neuropiles and thus highlights a role played by this post-synaptic CAM in mediating the plasticity of the central olfactory system, for which the underlying molecular mechanisms appear to be highly conserved between insects and vertebrates.