Molecular Basis of Pheromonogenesis Regulation in Moths

Abstract

Sexual communication among the vast majority of moths typically involves the synthesis and release of species-specific, multicomponent blends of sex pheromones (types of insect semiochemicals) by females. These compounds are then interpreted by conspecific males as olfactory cues regarding female reproductive readiness and assist in pinpointing the spatial location of emitting females. Studies by multiple groups using different model systems have shown that most sex pheromones are synthesized de novo from acetyl-CoA by functionally specialized cells that comprise the pheromone gland. Although significant progress was made in identifying pheromone components and elucidating their biosynthetic pathways, it wasn’t until the advent of modern molecular approaches and the increased availability of genetic resources that a more complete understanding of the molecular basis underlying pheromonogenesis was developed. Pheromonogenesis is regulated by a neuropeptide termed Pheromone Biosynthesis Activating Neuropeptide (PBAN) that acts on a G protein-coupled receptor expressed at the surface of pheromone gland cells. Activation of the PBAN receptor (PBANR) triggers a signal transduction cascade that utilizes an influx of extracellular Ca²⁺ to drive the concerted action of multiple enzymatic steps (i.e. chain-shortening, desaturation, and fatty acyl reduction) that generate the multicomponent pheromone blends specific to each species.

In this chapter, we provide a brief overview of moth sex pheromones before expanding on the molecular mechanisms regulating pheromonogenesis, and conclude by highlighting recent developments in the literature that disrupt/exploit this critical pathway.

1 Introduction

Sexual communication in most moths is dependent on the female’s ability to relay information regarding conspecificity, reproductive status, and location to receptive males. Research in earnest into the underpinnings of this chemical-based sexual communication originated with the seminal structure elucidation study published more than 50 years ago by Butenandt and co-workers (Butenandt et al. 1959). In that study, the authors reported the first chemical identification of an insect sex pheromone, (E,Z)-10,12-hexadecadien-1-ol (i.e. bombykol), extracted from 500,000 female silkworm moth (Bombyx mori) abdominal glands. Similar herculean efforts lead to the structural identification of sex pheromones from the cabbage looper Trichoplusia ni (Berger 1966) and the gypsy moth Lymantria dispar (Bierl et al. 1970). Since then, advances in analytical methodologies have facilitated elucidation of sex pheromones from several hundred lepidopteran species (El-Sayed 2014).

Sex pheromones are frequently de novo synthesized as multicomponent blends from acetyl-CoA (a process termed pheromonogenesis) in a specialized organ commonly referred to as the pheromone gland (PG) that is comprised of a single layer of modified epidermal cells between the eighth and ninth abdominal segments (Tillman et al. 1999; Jurenka 2003). Most moths produce Type I sex pheromones, which consist of long, straight chain hydrocarbons (C_10–18) with varying double bonds and functional modifications (alcohol, aldehyde, or acetate ester) of the carbonyl carbon (Tillman et al. 1999; Jurenka 2003, 2004; Ando et al. 2004). In contrast, Type II sex pheromones account for a small percentage (~15%) of the known lepidopteran compounds and are characterized by unmodified carbonyl carbons that consist of longer polyunsaturated hydrocarbons (C_17–23) and their epoxide derivatives (Ando et al. 2004; also see Chap. 11 volume 2). Early research on sex pheromone biosynthetic pathways clearly established that fatty acid metabolism intermediates (e.g. palmitic acid/hexadecanoic acid) provided the framework for downstream modifications. Using radiolabeled precursors, researchers were further able to elucidate specific biochemical steps to determine that pheromonogenesis, at least of the Type I pheromones, was derived from the dynamic interplay of selective β-oxidation reactions (i.e. chain-shortening), unique desaturases, and diverse reductive modifications (Bjöstad et al. 1987).

Despite years of foundational biochemical/chemical research, continued interest in the sex pheromone field has been fueled by its clear potential in integrated pest management strategies (Witzgall et al. 2008, 2010) and the ability to offer intriguing evolutionary insights (Roelofs et al. 2002; Lassance et al. 2010; Albre et al. 2013). Recent advances in genome/transcriptome sequencing, expansion of available molecular databases, and the advent of gene knockdown/knockout methodologies (e.g. RNA interference, CRISPR and TALENs) have greatly facilitated our understanding of moth pheromonogenesis at the cellular and molecular levels. This review will focus on the molecular mechanisms governing initiation and propagation of the signal that drives moth pheromonogenesis with a final section highlighting studies that describe approaches to disrupt and/or exploit this critical pathway.

2 Regulation of Pheromonogenesis

2.1 Hormonal and Neuroendocrine Regulation

2.1.1 Hormonal Regulation

Early observations that the production and release of pheromones in some insect species coincided with female reproductive cycles lead to the hypothesis that pheromone production was hormonally regulated (Barth 1965). The two predominant hormones in insects, juvenile hormone (JH) and 20-hydroxyecdysone (20E), are now recognized as critical regulators of pheromone production in cockroaches (Schal et al. 2003), beetles (Seybold and Vanderwel 2003; Haberer et al. 2010), flies (Wicker-Thomas et al. 2009; Bilen et al. 2013), ants (Cuvillier-Hot et al. 2004; Holman 2012), and wasps (Kelstrup et al. 2014). In moths, the role of JH varies. For relatively long-lived moths, in which sex pheromone production is delayed and activities related to migration and reproduction are asynchronous (i.e. noctuid species such as the armyworm Pseudaletia unipuncta, the black cutworm moth Agrotis ipsilon, and the cotton bollworm Heliothis armigera), JH functions in the control of pheromone production (Cusson and McNeil 1989; Picimbon et al. 1995; Fan et al. 1999; Zhou et al. 2000). In A. ipsilon, JH stimulates the release of a peptidergic factor (see Sect. 2.1.2) from production sites in the brain to trigger pheromone production in 4-day old sexually mature females (Picimbon et al. 1995). In species with shorter lifespans, such as H. armigera, in which females initiate pheromone production at an earlier stage, JH (JH-II) primes the female PG to respond to the peptidergic factor (Fan et al. 1999). Conversely, JH has also been implicated in pheromonostasis, i.e. suppression of pheromone production after mating (Webster and Cardé 1984). Exogenous JH has been shown to suppress pheromone production in some moth species (Rafaeli and Bober 2005; Bober et al. 2010; Zhang et al. 2014b), and the male-derived sex peptide that mediates the post-mating behavioral switch in Drosophila has both allatotropic (triggering JH biosynthesis) and pheromonostatic effects in H. armigera (Fan et al. 1999, 2000; Hanin et al. 2012).

Non-JH hormonal factors from the bursa copulatrix have also been reported to be required for pheromone production in the redbanded leafroller (Argyrotaenia velutinana), the eastern spruce budworm (Choristoneura fumiferana) and the oblique banded leaf roller (C. rosaceana) (Fabriàs et al. 1992; Delisle et al. 1999). It has been postulated that the relative importance of the bursa copulatrix in the hormonal regulation of pheromone production may be related to the evolution of enzyme desaturation systems in specific pheromone biosynthetic pathways, as found for instance in tortricid moths (Delisle et al. 1999).

2.1.2 Neuroendocrine and Neural Regulation

Pheromonogenic control in yet other moth species has been shown to proceed by a non-hormonal mechanism, as surgical removal of the corpora allata (CA; site of JH synthesis) had no discernible effect on the calling behavior of female saturnid moths (Riddiford and Williams 1971) and injection of CA homogenates also failed to stimulate pheromone production in Helicoverpa (Heliothis) zea (Raina and Klun 1984). Furthermore, circadian oscillations in pheromone production and emission coinciding with specific points of the day:night cycle (Raina and Klun 1984; Hunt and Haynes 1990; Delisle and Royer 1994; Kamimura and Tatsuki 1994; Gemeno and Haynes 2000; Foster 2000; Rosén 2002; Mazor and Dunkelblum 2005; Fónagy et al. 2011; Bloch et al. 2013; see Chap. 7) and the presence of a circulating pheromonogenic factor in the hemolymph of moths during scotophase (Ichikawa 1998; Jacquin et al. 1994; Ramaswamy et al. 1995) suggested a neuroendocrine component to pheromonogenic regulation. Biochemical analyses using adult H. zea females revealed that the factor was a peptide hormone, subsequently purified to homogeneity (see Sect. 2.2.1) and designated Pheromone Biosynthesis Activating Neuropeptide (PBAN), that was present in the brains and subesophageal ganglion (SOG) (Raina and Klun 1984; Raina et al. 1989). Accumulating evidence has supported circadian regulated release of PBAN from the corpora cardiaca into the hemolymph for direct pheromonotropic activity on PGs. However, reports describing pheromonotropic activity of a PBAN-like immunoreactive factor in the ventral nerve cord (VNC) and terminal abdominal ganglion, along with impaired pheromone production after severing the VNC suggest regulation may involve a neural component as well (Marco et al. 1996; Iglesias et al. 1998; Teal et al. 1999; Rosén 2002).

Neural signals from the VNC and depletion of sperm in the spermatheca are also important post-copulatory factors that regulate post-mating inhibition of pheromone production in polyandrous moths (Delisle et al. 2000; Delisle and Simard 2002). Mated females of polyandrous (multiple matings) species usually display a refractory period to reproduction after mating, which is largely due to the transfer of male humoral factors (sperm and seminal fluid) during copulation. Some of these male factors have short-term effects, whereas others can induce long-term suppression of female receptivity, as described in both butterflies and moths (Wedell 2005).

2.2 Purification and Characterization of the Pheromone Biosynthesis Activating Neuropeptide (PBAN)

2.2.1 HPLC-Based Identification of PBAN

Determination that the moth phermonotropic factor (i.e. PBAN) was a peptide hormone present in the brains and SOG of adult H. zea females facilitated HPLC (high-performance liquid chromatography) purification of the 33-amino acid PBAN from 2500 H. zea female brain-SOG complexes (Raina et al. 1989). Neuropeptides with similar functionalities and moderate overall sequence homology were likewise purified to homogeneity and sequenced from B. mori (Kitamura et al. 1989, 1990) and L. dispar (Masler et al. 1994). Consistent with its presumed role as the cue driving circadian oscillations in pheromone production, PBAN levels in both the brain and hemolymph fluctuate in accordance with photoperiod (Rafaeli et al. 1991, 1993; Rafaeli 1994; Ramaswamy et al. 1995; Iglesias et al. 2002; Nagalakshmi et al. 2007; Závodská et al. 2009). All PBANs have a conserved FxPRL-NH₂ (Phe-Xxx-Pro-Arg-Leu-amide) C-terminal pentapeptide motif that is critical for pheromonotropic activity (Raina and Kempe 1990; Kitamura et al. 1989). In addition, these pheromonotropic peptides exhibited species cross-reactivity as well as functional cross-reactivity with locust myotropins and tachykinins (Kuniyoshi et al. 1992; Fónagy et al. 1992a, b; Nachman et al. 1993a, b), suggesting that the cognate FxPRL-NH₂ peptide receptors were also similar.

2.2.2 Structure-Function Analysis of PBAN

Initial structure-function analyses of PBAN examined the pheromonotropic efficacy of peptide fragments generated either as a series of N-terminally truncated synthetic peptides (Raina and Kempe 1990) or endoproteinase Glu-C fragments (Kitamura et al. 1989). In both studies, the minimal sequence needed to stimulate pheromone production consisted of the C-terminal pentapeptide core (i.e. FxPRL). Comparison of amidated, hydroxylated, and methyl ester versions of the pentapeptide revealed the critical importance of the C-terminal amide (Kitamura et al. 1989; Kuniyoshi et al. 1992; Nagasawa et al. 1994). Sequential amino acid substitution of the core pentapeptide motif in B. mori (FSPRL-NH₂) revealed that Phe and Ser could be replaced with similar residues with little disruption of pheromonotropic activity, whereas Pro, Arg, and Leu could not (Kuniyoshi et al. 1991). Comparison of the pheromonotropic efficacies of FxPRL-NH₂ peptides from diverse species provided further insights into the structure-function relationships and suggested that the variable “x” position had greater pheromonotropic properties if occupied by Thr compared to Val, Ser, or Gly (Abernathy et al. 1995). More recent structure-function analyses revealed that the positively charged basic Arg (R; two positions from the C terminus) is the most critical residue within the hexapeptide motif (Kim et al. 2008; Kawai et al. 2012). It is followed in importance by the branched chain Leu, aromatic Phe, and then to a lesser extent by the other residues (Kim et al. 2008).

To provide greater insights into the role the individual residues in the C-terminal pentapeptide motif might play in receptor activation, Nachman and co-workers used nuclear magnetic resonance (NMR) spectroscopy, circular dichroism, and molecular dynamics simulations to determine that a cyclic analog of the pentapeptide adopts a C-terminal β turn in solution (Nachman et al. 1991). The analog, which introduced significant conformation constraints and increased the overall rigidity of the pentapeptide, retained biological activity, indicating that this conformation is crucial for receptor activation. Molecular simulations using the linear pentapeptide active core suggested the conformation was not specific to the cyclization process. Subsequent NMR analyses of a hexapeptide (TFSPRL-NH₂) analog and the full-length H. zea PBAN confirmed that the peptide assumes a C-terminal type I′ β turn in solution (Wang et al. 1994; Clark and Prestwich 1996). A more recent NMR study of an 18-amino acid pheromonotropin from Pseudaletia separata characterized by a C-terminal FxPRL-NH₂ revealed an extended β sheet structure devoid of the previously identified β turn (Bhattacharya et al. 2015). However, that study was performed in water as opposed to a more polar solvent (e.g. trifluoroethanol/water or dimethyl sulfoxide/water) that would presumably more accurately mimic the lipid bilayer environment in which the cell surface receptors are embedded.

2.3 Molecular-Based identification of PBAN

2.3.1 PBAN Transcripts

Following purification of the respective PBANs, cloning methods employing sequence information provided by the isolated peptides facilitated molecular elucidation of the B. mori and H. zea PBAN gene products (Davis et al. 1992; Kawano et al. 1992; Sato et al. 1993; Ma et al. 1994). In both instances, post-translational proteolytic processing of the encoded open reading frames was predicted to yield the respective PBANs and four additional peptides with C-terminal FxPRL-NH₂ motifs identified as diapause hormone (DH) and α, β, and γ subesophageal neuropeptides (i.e. SGNPs). Among the four additional peptides, DH had previously been isolated to homogeneity and shown to function in embryonic diapause (Imai et al. 1991). Synthetic α, β, and γ SGNPs were reported to have pheromonotropic activity in H. zea (Ma et al. 1994), but in B. mori the α and γ SGNPs were less effective than PBAN (β SGNP was comparable) at stimulating pheromone production and all three were less potent than DH in diapause induction (Sato et al. 1993). Later studies using PBANR receptors heterologously expressed in Xenopus oocytes, however, reported that the three SGNPs were more potent than PBAN in generating chloride currents (Watanabe et al. 2007).

Organization of the FxPRL-NH₂ open reading frames is conserved in both the B. mori and H. zea transcripts with the DH sequence downstream of the signal peptide followed by the α and β SGNPs, PBAN, and then γ SGNP. Since initial cloning, PBAN-encoding cDNAs with similar sequence architecture have been published for 22 lepidopterans with additional sequences deposited in GenBank or the Transcriptome Shotgun Assembly (TSA) sequence databases (Table 8.1) with most of the peptides composed of 33 residues (Fig. 8.1). Outliers include the Ascotis selenaria cretacea (Japanese giant looper) PBAN, which is 27 amino acids, and the 37 amino acid Omphisa fuscidentalis PBAN. A second 37 amino acid PBAN recently identified in Ostrinia nubilalis suggests close conservation of PBAN gene architecture between the closely related crambid subfamilies Pyraustinae and Spilomelinae (Fodor et al. 2017). The A. s. cretacea PBAN transcript is also unique in that it generates a fused β SGNP/PBAN with a double FxPRL motif (Kawai et al. 2007). Alignment of multiple lepidopteran PBAN sequences revealed that the variable position in the FxPRL-NH₂ motif reported to have an effect on pheromonotropic activity (Abernathy et al. 1995) is a conserved Ser (Fig. 8.1). The exception is Maruca vitrata (legume pod borer), which has Asn, an uncharged polar residue with a bulkier sidechain than Ser (Chang and Ramasamy 2014). Furthermore, all of the published PBAN cDNAs to date contain a dibasic KK motif upstream of the α SGNP sequence. While KK cleavage has been reported to be infrequent (Veenstra 2000), proteolytic processing of the PBAN prepropeptides was confirmed via HPLC-based fractionation of B. mori SOGs (Sato et al. 1993) and MALDI (matrix-assisted laser desorption/ionization) mass spectrometry of individual H. zea SOG neuronal clusters (Ma et al. 2000).

Table 8.1 Accession numbers for PBAN and PBANR sequences identified in lepidopteran species

Fig. 8.1

Multiple sequence alignment of PBAN coding sequences from diverse lepidopteran species. The alignment was made using MUSCLE implemented in Geneious 7. Sequences correspond to the portion of the DH/PBAN transcript from the FxPRL portion of the β SGNP through the predicted proteolytic cleavage site at the C terminus of the PBAN sequence. Predicted cleavage and amidation sites are indicated. The essential FxPRL active core of PBAN is indicated by the dashed grey lines. Genome-based butterfly sequences are clustered at the bottom of the alignment. Protein accession numbers are indicated in parentheses

The presence of PBAN sequence (and/or prepropeptide) variants was initially described in B. mori following HPLC-based purification of two peptides (PBAN-I and PBAN-II) with pheromonotropic activity that differed from one another by a single N-terminal Arg residue (Kitamura et al. 1989, 1990). Since then potential sequence variants have been deposited with the NCBI database for a number of species including B. mori (three point mutations between AAB24327 and BAA05954-K109 N, M139I, and E146V), Spodoptera litura (one point mutation between AJT60314 and AKT95050-E53G), and Helicoverpa armigera (three point mutations between AAL05596 and AAM43840-deletion of N3, insertion of G before G30, and M179I).

At this point, it is uncertain if these variants represent population differences, are differentially expressed variants, or are merely the result of sequence errors introduced during cloning. However, differentially expressed PBAN prepropeptide transcript variants have been reported in the sand fly Phlebotomus papatasi (Choi et al. 2015) and are suggested based on a band doublet observed on an RT-PCR gel of fire ant thoraces (Choi et al. 2011). More recently, transcripts that vary in the length and composition of their 3′UTRs (untranslated regions) have been identified in O. nubilalis (Fodor et al. 2017).

2.3.2 PBAN Gene Structure

The lepidopteran PBAN genomic structure is conserved with PBAN genes in B. mori (Xu et al. 1995), H. armigera (Zhang et al. 2005), M. vitrata (Chang and Ramasamy 2014), and Clostera anastomosis (Jing et al. 2007) encompassing six exons with identical exon coding (Fig. 8.2). Exon one encodes the signal peptide and a portion of DH, exon two the remaining portion of DH, exon three an uncharacterized peptidergic sequence, exon four the α and β SGNPs and a portion of PBAN, and exon five the remaining portion of PBAN and γ SGNP. The stop codon is located in exon six. Splicing of all four genes follows the GT-AG rule and utilizes 0, 2, 1, 2, 1 phasing; however, despite the similarities, the overall sizes of the genes differ with varying intron lengths (Fig. 8.2). The O. nubilalis PBAN was recently reported to have the same genomic structure (Fodor et al. 2017).

Fig. 8.2

Genomic organization of the DH-PBAN gene in four moth species. Schematic comparison of genomic DNA and the translated peptides for the DH-PBAN gene in Bombyx mori (Xu et al. 1995), Helicoverpa zea (Zhang et al. 2005), Clostera anastomosis (Jing et al. 2007), and Maruca vitrata (Chang and Ramasamy 2014). Darker shaded boxes indicate exons, whereas lighter shaded boxes indicate the encoded peptides. Horizontal solid lines represent introns with the corresponding intron phase in parentheses. GKR, KK, GR, and GRR indicate probable endoproteolytic cleavage sites. SP - signal peptide; DH-diapause hormone; α-α SGNP; β-β SGNP; PBAN - pheromone biosynthesis activating neuropeptide; γ-γ SGNP. Note, while the sizes of exons and introns are indicated, the models are not drawn to scale

Limited promoter analyses, which focused on elucidating how DH expression was regulated in relation to embryonic diapause as opposed to pheromonogenesis, identified potential differences in transcription between the B. mori and H. armigera genes. POU-M2, a eukaryotic transcription factor with a bipartite DNA binding domain implicated in neuroendocrine function, activated expression from the B. mori PBAN promoter in vitro but failed to do so with a conserved region of the H. armigera promoter (Zhang et al. 2004a, 2005). In contrast, an E-box element (CAGCTG) present in the H. armigera promoter was reported to be critical for transcriptional activation (Hong et al. 2006), which was dependent on co-ordinate interactions with upstream activating and inhibitory regions. Taken together, the findings suggest that the two species utilize variations in transcriptional regulation to drive the respective differences in diapause programs. Additionally, an ecdysone response element was identified in the promoter region of the B. mori PBAN gene (Xu et al. 1995). While ecdysteroids have not been associated with diapause control, they are critical regulators of lepidopteran reproduction (Van Wielendaele et al. 2013; De Loof et al. 2016). Consequently, the response element may link PBAN transcription with reproductive competence; however, the role it has in pheromonogenesis remains to be revealed.

2.4 Other FxPRL-NH₂ Peptides

The critical C-terminal pentapeptide is now recognized as a defining feature of the PBAN/pyrokinin (FxPRL) family of pleiotropic neuropeptides present throughout Insecta and includes pyrokinins, PBANs, myotropins, DH, and the α, β, and γ SGNPs (Predel and Nachman 2006; Jurenka and Nusawardani 2011; Altstein et al. 2013; Jurenka 2015; Yaginuma and Niimi 2015). In addition to the pheromonotropic effects in moths, FxPRL-NH₂ peptides also regulate the induction of cuticular melanization in moth larvae (Matsumoto et al. 1992; Altstein et al. 1996), the induction of embryonic diapause and seasonal polyphenism in moths (Imai et al. 1991; Uehara et al. 2011), the termination of pupal diapause in heliothine moths (Xu and Denlinger 2003; Zhang et al. 2004b, c; Zhao et al. 2004), prothoracic gland ecdysteroidogenesis (Zhang et al. 2004c; Watanabe et al. 2007), visceral muscle contraction in cockroaches (Holman et al. 1986; Nachman et al. 1986; Predel et al. 2001), acceleration of puparium formation in flies (Zdárek and Nachman 1997; Zdárek et al. 1998, 2002, 2004), production of fatty acid components in male H. armigera hair-pencil aedeagus complexes (Bober and Rafaeli 2010), and the biosynthesis of trail pheromones in Solenopsis invicta (Choi and Vander Meer 2012). This multi-functionality is similar to the structural variation described for the chemosensory protein (CSP) family of multi-function transporters, which are widely expressed in diverse tissues including the PG (see Chaps. 6, 9, and 10, volume 2; Xuan et al. 2014, 2016; Picimbon 2017).

PBAN control of pheromonogenesis, however, is not ubiquitous throughout the moths (Tang et al. 1989; Subchev and Jurenka 2001; Fujii et al. 2010) nor is it specific to moths that produce Type I pheromone components (albeit our knowledge of this system is more complete) as it has been reported to regulate production of Type II pheromones in the giant looper A. s. cretacea (Wei et al. 2004; Fujii et al. 2007). Furthermore, some Lepidopteran species such as T. ni do not exhibit diel periodicity in pheromone production (Hunt and Haynes 1990; also see references in Rafaeli and Jurenka 2003; Altstein 2004a), and as such would be expected to have little need for PBAN-mediated regulation. However, T. ni brain extracts were found to have pheromonotropic activities in other moth species (Tang et al. 1989). Since then it has become apparent that PBAN is a pleiotropic regulator of diverse activities (see above). Indeed, the elucidated primary structure of the HPLC-purified B. mori peptide responsible for larval cuticular melanization (i.e. melanization and reddish coloration hormone) was identical to the PBAN sequence (Matsumoto et al. 1990).

2.5 Identification of the PBAN Receptor (PBANR)

2.5.1 PBANR: Early Studies

The involvement of a cell surface receptor that mediates the pheromonotropic effects of PBAN was demonstrated early on following direct stimulation of dissected PGs by PBAN (Soroker and Rafaeli 1989; Jurenka et al. 1991b; Fónagy et al. 1992a, c). Pharmacological profiling with NaF (sodium fluoride), a potent G protein activator that had pheromonotropic effects (Rafaeli and Gileadi 1996a, b) further pointed to the involvement of a G protein-coupled receptor (GPCR). The photoaffinity labeling of a ~50 kDa membrane protein in H. armigera PG cells with a biotinylated PBAN analog provided incontrovertible evidence of a PG-derived cell surface protein (Rafaeli and Gileadi 1999; Rafaeli et al. 2003, 2007). However, molecular identification of the moth PBAN receptor (PBANR) ultimately depended on publication of the Drosophila melanogaster genome (Adams et al. 2000).

2.5.2 Homology-Based Cloning of PBANR

Sequence homologies between mammalian receptors and putative GPCRs in the Drosophila genome led researchers to propose that co-evolution of receptors and their ligands would yield closely aligned receptor families (Hewes and Taghert 2001). Based on this hypothesis, similarities in the active core of FxPRL-NH₂ peptides and neuromedin U (FRPRN-NH₂) suggested that the respective receptors are evolutionarily related. Functional analyses demonstrated that three Drosophila GPCRs (CG8784, CG8795, and CG9918) that clustered in phylogenetic analyses with the neuromedin U receptor (NmUR) clade were activated to varying degrees by FxPRL-NH₂ peptides (Park et al. 2002). A subsequent study reported pheromonotropic effects of mammalian NmU in H. zea, which further bolstered the receptor co-evolution hypothesis and showed that homology-based methods could be used to clone receptors from the NmUR clade (Choi et al. 2003). The H. zea GPCR identified in that study was amplified from PG cDNAs and, when heterologously expressed in cultured Sf9 cells, dose-dependently triggered an influx of extracellular Ca²⁺ in response to synthetic H. zea PBAN. This was interpreted as evidence that the authors had identified the first PBANR (i.e. HelzePBANR). Using a similar approach, the B. mori PBANR (BommoPBANR) was likewise cloned from PG cDNAs. BommoPBANR mobilized extracellular Ca²⁺ in response to PBAN stimulation, had significant sequence similarity with NmUR homologs, and was up-regulated on the day preceding adult eclosion (Hull et al. 2004), a time period that coincides with B. mori pheromonogenesis (Matsumoto et al. 2007, 2010).

2.5.3 The Complexity of PBANR

2.5.3.1 Identification of PBANR Variants

Perplexingly, the ~50 kDa protein labeled with the biotinylated PBAN analog in the intersegmental membranes that comprise the H. armigera PG (Rafaeli and Gileadi 1999; Rafaeli et al. 2003, 2007) was closer in size to BommoPBANR (45.9 kDa) than the smaller HelzePBANR (38.6 kDa). Despite presumably mediating similar biological responses and significant sequence identity through the seventh transmembrane domain (TM7), BommoPBANR was differentiated by the presence of a 67-aa C-terminal extension critical for ligand-induced internalization (Hull et al. 2004, 2005), an endocytotic mechanism associated with GPCR feedback regulation and desensitization (Moore et al. 2007; Marchese et al. 2008). Further confounding the issue, PBANRs subsequently identified in H. armigera (Rafaeli et al. 2007), S. littoralis (Zheng et al. 2007), Spodoptera exigua (Cheng et al. 2010), and Plutella xylostella (Lee et al. 2011) also lacked the C-terminal extension, suggesting feedback regulation of these receptors differed from BommoPBANR. The prevalence of the “short” PBANRs raised questions concerning the evolutionary significance of the BommoPBANR extension. Initially, comparisons were made with type I gonadotropin-releasing hormone receptors in which non-mammalian receptors have a C-terminal tail and undergo rapid ligand-induced internalization, whereas mammalian receptors lack the extended C terminus and have significantly different internalization kinetics (Pawson et al. 1998; McArdle et al. 2002). The potential biological significance of the “short” and “long” PBANRs also led to speculation that the varied C-terminal lengths reflected differences in the importance of the second messenger 3′,5′-cyclic adenosine monophosphate (cAMP) in the respective species. The identification of three PBANR variants concomitantly expressed in Helicoverpa virescens (referred to as HelviPBANR A-C) with a conserved N-terminal sequence, but with differing C-terminal lengths (Kim et al. 2008), further underscored the complexity of the PBAN signaling system. Similar to BommoPBANR, the HelviPBANR-C variant has an extended C terminus and contains a defined internalization motif (see 2.7.7), whereas the C-terminal end of the HelviPBANR-A variant resembles HelzePBANR. Moreover, HelviPBANR-C was preferentially amplified from PGs and generated robust Ca²⁺ mobilization responses following stimulation with H. zea PBAN. In contrast, the other two variants were amplified from larval tissues and failed to respond to the concentration of the synthetic PBAN assayed (Kim et al. 2008). These results initiated a re-evaluation of the species-specific “short” and “long” PBANR paradigm.

Using modified cloning methods, multiple PBANR variants (PBANR-As, -A, -B, and -C) were amplified from PGs of B. mori, H. zea, H. armigera, and P. separata (also referred to as Mythimna separata) that differed only in the length of their respective C-terminal ends (Fig. 8.3a). Similar to H. virescens, the most abundant PG transcripts were the “long” PBANR-C variants (Fig. 8.3b), all of which underwent ligand-induced internalization (Lee et al. 2012a). In contrast, the “short” PBANR-A variants were less abundant, mobilized extracellular Ca²⁺ poorly in response to a range of PBAN concentrations, and exhibited different internalization kinetics (Lee et al. 2012a, b). Previous preferential amplification of the shorter variants (Choi et al. 2003; Rafaeli et al. 2007; Cheng et al. 2010; Lee et al. 2011) was attributed to the high GC content (55–80%) of the extended C-terminal ends (Lee et al. 2012a), which can reduce PCR amplification efficiencies by serving as pause or termination sites (McDowell et al. 1998). Thus it is now apparent that the ~50 kDa protein labeled with the biotinylated PBAN analog (Rafaeli and Gileadi 1999; Rafaeli et al. 2003, 2007) was most likely the HelarPBANR-C variant (51.1 kDa) rather than a glycosylated HelarPBANR-A variant (38.7 kDa) as first proposed.

Fig. 8.3

Identification of multiple PBANR variants in Bombyx mori pheromone gland. (a) Schematic diagram depicting the sizes and structures of the various BommoPBANR variants cloned. (b) RT-PCR based expression profile of BommoPBANR variants in various tissues and at varying developmental time points relative to adult emergence (day 0). Abbreviations: PG pheromone gland, Br brain, FM flight muscle, Eg unfertilized egg, MT Malpighian tubule, FB fat body, MG midgut. This research was originally published in Frontiers of Endocrinology. Lee et al. (2012a). (c) Genomic organization and alternative splicing of the Bombyx mori PBANR gene. The four BommoPBANR variants (As, A, B, and C) are depicted. Light grey shading corresponds to untranslated exons, medium grey to translated exons, and dark grey to a non-translated exon that is unique to the As variant. Non-shaded boxes indicate non-spliced intronic sequences. Initiation (ATG) and stop sites (TGA or TAG) are indicated by their respective codons. (Figure adapted from Lee et al. 2012a)

2.5.3.2 PBANR Variants Arise from Alternative Splicing

Alternative splicing has been extensively documented for GPCRs (Minneman 2001; Markovic and Challiss 2009) and is one of the principal means by which organisms generate functional protein diversity in a temporal- and/or tissue-dependent manner. The modular aspect of the PBANR variants (i.e. variation specific to the C terminus) is consistent with alternative splicing. The availability of the B. mori genome (Mita et al. 2004; Duan et al. 2010) allowed further exploration of that hypothesis. BommoPBANR localizes to a >50 kb segment of chromosome 12 and encompasses six exons and five introns (Fig. 8.3c). The N terminus through the last transmembrane domain (i.e. TM7) are encoded on exons 2–4, the C terminus on exons 5–6, and the 5′ untranslated region on exon 1. The introduction of premature stop codons following retention of introns 3 or 4 yields the BommoPBANR-As and BommoPBANR-A variants, respectively. BommoPBANR-C arises from a five-nucleotide frame shift insertion at the 3′ end of exon 5 that changes codons for the remaining ten amino acids (residues 404–413) and introduces a stop codon that generates a 67 amino acid C terminus. In contrast, BommoPBANR-B is generated from conventional splicing of exons 2–6 (Lee et al. 2012a). As more lepidopteran genomes become available, it will be interesting to see if the splicing mechanisms that generate the BommoPBANR variants are conserved in other species and what cellular/transcriptional factors trigger those splicing events.

2.5.3.3 PBANR Variants: Fine-Tuning the PBAN Signal?

To date, PBANRs have been reported or annotated in 15 species (Table 8.1) with multiple variants present in O. nubilalis (Nusawardani et al. 2013), Manduca sexta (FJ240221-FJ240224), Chilo suppressalis (KT031039-KT031040), Mamestra brassicae (Fodor et al. 2018), and based on genomic sequencing data, three Papilio species.

While the biological significance of concomitant expression of multiple PBANRs in PGs remains to be determined, one possibility is that they provide a mechanism for fine-tuning cellular responsiveness to the respective PBAN signals. In one model, nominally non-responsive PBANR-A receptors expressed at the cell surface could potentially function as ligand sinks that compete with PBANR-C for ligand binding. The net result would be less bioactive peptide available to trigger the cellular response thus decreasing overall sensitivity. In a second model, heterodimerization of the shorter variants with the longer variants could impede trafficking to the cell surface, thereby decreasing the pool of available receptors for ligand binding, which would likewise decrease overall cellular sensitivity. When co-expressed in cultured cells with their predominant full-length receptor forms, truncated variants of some mammalian receptors have been reported to exert dominant negative effects on signaling (Seck et al. 2005; Zmijewski and Slominski 2009; Chow et al. 2012). Alternatively, because many receptor variants exhibit distinct spatial and temporal expression profiles as well as altered ligand binding, atypical feedback regulation, and differential activation of downstream effector pathways (Markovic and Challiss 2009), the multiple PBANR transcripts may reflect a spatio-temporal dependence of functionality. This hypothesis is especially attractive given the pleiotropic complexity of PBAN, the multiplicity of reports detailing PBANR activation by multiple FxPRL-NH₂ peptides (Choi et al. 2003; Watanabe et al. 2007; Kim et al. 2008; Hariton-Shalev et al. 2013; Shalev and Altstein 2015), and the varied expression profile of PBANR transcripts, which have been amplified from diverse tissues including the PG, brain, SOG, ventral nerve cord, thoracic ganglion, ovary, and male abdominal tip (Rafaeli et al. 2007; Watanabe et al. 2007; Bober and Rafaeli 2010; Cheng et al. 2010). Indeed, PBANR expression in larval tissues (Zheng et al. 2007; Kim et al. 2008) suggested possible roles in melanization and/or pupal diapause. Recent studies seem to support this hypothesis with larval-derived and PG-derived PBANRs differing markedly in their three-dimensional conformations, regions/degrees of electrostatic potential, and ligand binding properties (Hariton-Shalev et al. 2013; Shalev and Altstein 2015). While suggestive, these findings require further validation using alternative expression systems, the inclusion of more PBANRs, and the use of various potential endogenous ligands.

2.6 Other FxPRL-NH₂ Receptors

Although significant progress has been made in molecular characterization of PBANRs, the presence of transcripts in diverse tissues, pleiotropic activation (i.e. DH, PBAN, and SGNPs), and the concomitant expression of multiple variants have collectively raised questions regarding the spatio-temporal interactions between the receptor and the FxPRL-NH₂ peptides that regulate pheromonogenesis. These questions were both clarified (and further obscured) following identification of the B. mori DH receptor (BommoDHR) (Homma et al. 2006). DH is one of the five FxPRL-NH₂ peptides encoded on the PBAN prepropeptide gene and functions in induction of embryonic diapause and seasonal polyphenism (Imai et al. 1991; Uehara et al. 2011), the termination of pupal diapause in heliothine moths (Xu and Denlinger 2003; Zhang et al. 2004b; Zhao et al. 2004), and prothoracic gland ecdysteroidogenesis (Zhang et al. 2004c; Watanabe et al. 2007). Although BommoDHR was cloned from developing ovaries using a homology-based approach similar to that used for the PBANRs, sequence identity between BommoDHR and the BommoPBANR variants is only ~40% (Homma et al. 2006). DHRs have since been either cloned or identified based on sequence homology from a number of lepidopterans (Jurenka and Nusawardani 2011). The two receptor types, along with homologs in other insect orders referred to as pyrokinin 1 receptor (PKR1; DHR-like) and pyrokinin 2 receptor (PKR2; PBANR-like), are phylogenetically distinct (Jurenka and Nusawardani 2011; Nusawardani et al. 2013; Jiang et al. 2014). Despite these differences, the activities of DH and PBAN on HelzePBANR and BommoPBANR were reported to be comparable (Choi et al. 2003, 2007; Watanabe et al. 2007). Conversely, PBAN had >20-fold lower activity on BommoDHR (Homma et al. 2006) and no activity on OstnuDHR (Nusawardani et al. 2013), suggesting that greater ligand discrimination occurs with DHR than PBANR. However, functional analyses performed by other groups using different expression systems and assays, came to different conclusions as DH had 15-fold lower activity than PBAN on HelviPBANR-C (Kim et al. 2008) and PBAN activity on HelzeDHR was virtually indistinguishable from DH (Jiang et al. 2014). While these discrepancies likely reflect methodological variances and/or complications associated with heterologous expression (Zhang et al. 2014a), in vitro differences in the efficacy of the two peptides (Stern et al. 2007; Watanabe et al. 2007; Hariton-Shalev et al. 2013; Shalev and Altstein 2015) support regulation of distinct functionalities by the respective ligand-receptor pairs. However, the reduction in pheromonogenesis observed in response to RNA interference (RNAi)-mediated knockdown of PBANRs in B. mori (Ohnishi et al. 2006), P. xylostella (Lee et al. 2011), and male H. armigera (Bober and Rafaeli 2010) have provided unequivocal demonstration of PBANR involvement in mediating the biological effects of PBAN. In those studies, pheromonogenesis was only partially inhibited (~50% reduction) not abolished, suggesting limited penetrance of the dsRNA into the PG cells or that receptor levels, while reduced, were still sufficient to propagate the pheromonogenic signal. Alternatively, those findings may indicate that a full pheromonogenic effect depends on additional endocrine signals and/or other FxPRL-NH₂ receptor/ligand pairs. Despite increasing the complexity of our model for pheromone regulation, the latter hypothesis is attractive as transcripts for both PBANR and DHR have been amplified from PG cDNAs (Watanabe et al. 2007; Nusawardani et al. 2013).

2.7 Structure-Function Analysis of PBANR

2.7.1 Elucidating GPCR Structural Requirements Critical to Ligand Binding and Activation

Targeted disruption of insect neuropeptide signaling, which modulates virtually all aspects of insect biology, physiology and behavior, has been proposed as a novel pest control strategy with great potential for development by the agro-chemical industry (Altstein and Nässel 2010; Audsley and Down 2015). Successful exploitation of this strategy, however, requires a comprehensive understanding of the molecular mechanisms underlying ligand binding and receptor activation. Efforts to determine the atomic structures of GPCRs by standard NMR and X-ray crystallography methods were initially hampered by the necessity of a lipid bilayer suspension. Consequently, researchers turned to in silico methods using structurally related templates and/or structure-function analyses of GPCR mutants to gain insights into GPCR functionality. Chimeric receptors that incorporate domains from distant, but related GPCRs have also provided insights into the molecular determinants that govern ligand-receptor interactions (Yin et al. 2004) and revealed roles for the N terminus and extracellular loops (ECL) in ligand binding/discrimination (Peeters et al. 2011; also see Chap. 4 volume 2).

2.7.2 PBANR Extracellular Domains

To elucidate the structural determinants governing PBANR activation, Choi et al. (2007) generated a series of chimeric GPCRs that swapped the extracellular domains of HelzePBANR and the D. melanogaster pyrokinin receptor 1 (DromePKR1; analogous to DHR), which is ~100-fold less responsive to PBAN. Ligand discrimination was found to largely reside in ECL3, and to a lesser extent the N terminus (Choi et al. 2007), two domains that have been implicated in peptide ligand-GPCR interactions (Gether 2000; Gether et al. 2002; Peeters et al. 2011). Impaired activity following a swap of the respective ECL2 domains was attributed to disruption of the disulfide linkage connecting ECL2 and TM3 that is critical for GPCR folding and ligand binding (de Graaf et al. 2008). To further explore the role of HelzePBANR ECL3 in ligand discrimination, three separate point mutations were later made to residues (G297, S300, and F303) with functional groups that could potentially interact with a peptide ligand (Fig. 8.4a). Alanine substitution of S300 and F303 reduced the efficiency of Ca²⁺ mobilization compared to non-mutated controls in response to PBAN stimulation, suggesting that both residues may comprise potential contact points or contribute to the overall stabilization of the ligand binding pocket (Choi and Jurenka 2010). The role of N-glycosylation, which has been linked with efficient cell surface trafficking (Duvernay et al. 2005), was also examined within the context of HelzePBANR-mediated Ca²⁺ influx (Choi et al. 2007). Glutamine substitution of two consensus N-glycosylation sites (N19 and N22) in the HelzePBANR N terminus (Fig. 8.4a) negatively impacted PBAN-stimulated Ca²⁺ influx, an effect that was attributed to disruption of forces stabilizing the overall HelzePBANR structure (Choi et al. 2007). However, it is unclear what kind of effect, if any, the substitutions had on receptor trafficking. Deletion of the first 27 residues from the BommoPBANR N terminus, which likewise has two consensus N-glycosylation sites (N18 and N21), had no effect on receptor trafficking, ligand binding, or ligand-induced internalization (Hull et al. 2011). This variation in responses may be an artifact of the different assays used to assess functionality, or could reflect intrinsic differences between the respective receptors as N-glycosylation effects on GPCR trafficking and activity have been reported to be receptor-dependent (Duvernay et al. 2005).

Fig. 8.4

Schematic illustration of sites in Helicoverpa zea and Bombyx mori PBANRs. (a) Residues predicted to comprise the PBAN ligand-binding pocket in HelzePBANR. (b) Residues predicted to comprise the PBAN ligand-binding pocket in BommoPBANR. (c) Schematic illustration of sites in BommoPBANR-C that have undergone functional analysis via site-directed mutagenesis

2.7.3 HelzePBANR Ligand Pocket

Although rhodopsin (a light sensitive GPCR) is an imperfect template for modeling peptide GPCRs (Sabio et al. 2008; Mobarec et al. 2009; Congreve et al. 2011) it can offer structural insights into potential regions of ligand contact (Congreve et al. 2011). Using molecular docking techniques with a PBAN analog (YFSPRL-NH₂) and a sequence optimized HelzePBANR conformation that utilized coordinates from the bovine rhodopsin crystal X-ray structure, Stern et al. (2007) identified twenty amino acid residues that potentially comprise the ligand binding pocket (Fig. 8.4a). This in silico HelzePBANR structure was also used to evaluate the conformational effects of the ECL swaps (see Sect. 2.7.2) between HelzePBANR and DromePKR1 (Choi et al. 2007). In that evaluation, each domain swap reduced the number of putative ligand contact points. The largest reduction was observed with the ECL2 swap, an effect that likely resulted from misorientation of the cysteines composing the ECL2-TM3 disulfide bridge (Choi et al. 2007). In a complementary approach, Jurenka and Nusawardani (2011) used an evolutionary trace method that mapped conserved residues to a three-dimensional model of HelzePBANR to identify sites critical for ligand selection and binding. The authors of that study further refined their predictions based on the presupposition that the spatial coordinates of GPCR binding domains are frequently evolutionarily conserved. Overall, they identified eleven TM residues potentially comprising a conserved FxPRL-NH₂ binding domain (Fig. 8.4a). They also suggested that the charged residues in HelzePBANR ECL3 (K294, E297, and D301) could potentially contribute to the ligand specificity revealed in the ECL3 domain swap between HelzePBANR and DromePKR1 (Choi et al. 2007), which is consistent with previous reports that ECL3 is involved in ligand specific conformational changes (Gether 2000; Gether et al. 2002; Peeters et al. 2011). However, mutagenesis analysis of E297 had little effect on receptor activity (Choi and Jurenka 2010). Because of the different in silico approaches, the two HelzePBANR models that were developed yielded different aspects of the potential binding pocket. The structural approach focused on the potential role of the ECLs, whereas the evolutionary trace approach focused on identifying the conserved GPCR binding pocket bounded by the TM helices. Taken together, the approaches identified a number of potential ligand interaction points that still await experimental verification.

2.7.4 BommoPBANR Ligand Pocket

In a separate in silico study (Kawai et al. 2014), coordinates based on crystal structures for two class A GPCRS (human β2 adrenergic receptor and human A_2A adrenergic receptor) facilitated identification of twenty-seven potential ligand interaction sites in BommoPBANR (Fig. 8.4b). Only three of the twenty potential residues implicated in the rhodopsin-based HelzePBANR structure (Stern et al. 2007) were identified in the BommoPBANR model. However, all of the contact points predicted by the evolutionary trace method (Jurenka and Nusawardani 2011) were present. Sequential Ala-substitution of the residues revealed roles in ligand binding, receptor activation (i.e. mobilization of extracellular Ca²⁺), and cell surface trafficking/protein stability. Given their interhelical localization, the four residues (S207, F211, F212, and H284) that affected cell surface expression are predicted to contribute to stabilization of the TM helical bundle. Consequently, the impaired expression observed by the authors was likely the result of receptor misfolding. Kawai and co-workers (2014) further reported a reduction in both ligand binding and receptor activation following Ala-substitution of eleven residues (E95, E120, N124, V195, F276, W280, F283, R287, Y307, T311 and F319), whereas three residues (F209, F303, G315) were implicated in ligand binding alone, and a single residue (Y318) in receptor activation. In this last case, Ala-substitution generated a mutant that exhibited normal ligand binding but impaired receptor activation, suggesting that it may be crucial in the PBAN-induced conformational change that converts the receptor from the non-activated to the activated state. Furthermore, the defects observed with five of the putative binding sites (F212, F276, W280, F283, and F319) may not be exclusively related to ligand binding as they are highly conserved in class A GPCRs and may function in the receptor conformational switch (Holst et al. 2010).

Molecular docking simulations using the BommoPBANR structure and a 5-aa FSPRL-NH₂ analog identified a number of receptor-ligand interactions largely localized to the TM bundle (Kawai et al. 2014). Similar simulations using a NmUR model and a 5-aa analog of NmU further revealed that points of contact between the critical Leu and amide in the respective ligands and the putative binding pockets were conserved: PBANR E95/NmUR E117 (TM2), PBANR E120/NmUR E142 (TM3), PBANR F283/NmUR F313 (TM6), PBANR Y318/NmUR F345 (TM7), and PBANR F319/NmUR Y346 (TM7). The Glu residues in TM2 and TM3 appear to be critically important for ligand binding among the NmUR clade of receptors, as conservation of those sites in other class A GPCRs is more limited (Kawai et al. 2014).

While the ligand-binding pocket described by Kawai and co-workers (2014) is sufficient to accommodate the C-terminal FSPRL-NH₂ active core, steric hindrance precludes it from accepting the full-length 33-aa peptide, suggesting that the non-essential N-terminal portion of PBAN interacts with the ECLs. These interactions could potentially contribute to the stabilization of ligand binding as well as serve as a selectivity filter for differentiating between ligands with similar active cores (i.e. PBAN vs DH). In support of this model, two ECL residues (V195 in ECL2 and F303 in ECL3) important in binding the 10-aa PBAN analog were not identified as contact points for the 5-aa analog (Kawai et al. 2014). Furthermore, FxPRL-NH₂ ligand discrimination has been demonstrated experimentally (Homma et al. 2006; Stern et al. 2007; Watanabe et al. 2007; Hariton-Shalev et al. 2013; Nusawardani et al. 2013; Shalev and Altstein 2015) and when functionally important residues in BommoPBANR and BommoDHR are compared, all are conserved with the exception of V195 (Glu in DHR) and F303 (Pro in DHR).

2.7.5 PBANR Intracellular Domains

In contrast to the ligand binding functions of the ECL domains, the C terminus and intracellular loops (ICLs) are critical for propagation and termination of the ligand signal. Ligand binding promotes G-protein dissociation, activation of downstream signal transduction cascades, and subsequent negative feedback regulation/desensitization of the activated GPCR, typically effectuated via endocytotic removal of the receptor from the cell surface (Ferguson 2001; Kristiansen 2004). Knowledge of the specific structural motifs within GPCRs (insect GPCRs in particular) that mediate these processes, however, is limited. Structure-function studies have begun to address this deficiency by providing insights into the mechanisms underlying propagation of the PBAN signal.

2.7.6 G Protein-Coupling

Pheromonogenesis is dependent on an influx of extracellular Ca²⁺ (Jurenka et al. 1991a; Choi and Jurenka, 2004, 2006; Rafaeli 2009; see Sect. 3.2.1). In B. mori, this event is mediated by receptor dissociation of a Gαq heterotrimeric G protein (Hull et al. 2010). Receptor-G protein coupling frequently involves ionic interactions between cationic residues near TM6 of the receptor and anionic residues in the C terminus of the G protein (Yang et al. 2002; Kleinau et al. 2010). Alignment of PBANRs with other NmU-clade GPCRs revealed a dibasic site (R263 and R264 in BommoPBANR, see Fig. 8.4c) at this junction (Hull et al. 2011). Ligand-induced internalization, a cellular event that occurs downstream of receptor activation, was significantly reduced following site-directed mutagenesis of these residues (Hull et al. 2011). The disruption in internalization suggests that PBANR signaling was impacted, providing indirect evidence for this region in PBANR-G protein coupling.

2.7.7 C-Terminal Motifs Critical for Ligand-Induced Internalization

A number of conserved C-terminal motifs play critical roles in GPCR desensitization and endocytosis (Ferguson 2001; Kristiansen 2004). The C-terminal region of BommoPBANR has two such motifs, NPxxY (residues 325–329) and YxxΦ (residues 360–363) (Fig. 8.4c). Although NPxxY has been reported to function in the internalization of multiple vertebrate GPCRs (Barak et al. 1995; Gripentrog et al. 2000; He et al. 2001; Bouley et al. 2003), its role in endocytosis is receptor dependent (Slice et al. 1994; Hunyady et al. 1995). The YxxΦ motif (Y = Tyr, x = any amino acid, and Φ = amino acid with a bulky hydrophobic sidechain) has also been implicated in ligand-induced internalization (Paing et al. 2004; Pandey 2009) and is present in the C terminus of numerous peptide GPCRs. Using a series of C-terminal truncations, the BommoPBANR internalization motif was mapped to a 10-aa region spanning residues 357–367, which contain the YxxΦ motif (Hull et al. 2005). Ala-substitution of the critical residues in the motif likewise impaired internalization (Hull et al. 2005), albeit not to the same extent as the C-terminal truncation, which suggests that, similar to other receptors (Johnson et al. 1990; Nussenzveig et al. 1993; Thomas et al. 1995), the PBANR endocytotic mechanism utilizes multiple signals. The C-terminal YxxΦ motif, YSAL, is highly conserved among the lepidopteran PBANRs and a number of related receptors (i.e. PKR2) in other species, but has diverged in DHRs (YTAM/V), and is not readily apparent in PKR1. This variance suggests that regulation of those receptors either utilizes a different internalization signal or proceeds via a non-endocytotic pathway. Whether or not this sequence is sufficient in and of itself to promote internalization of PBANRs from other species has yet to be experimentally determined.

2.7.8 Phosphorylation-Dependent Internalization of BommoPBANR

Desensitization and internalization of GPCRs are triggered in response to ligand-induced phosphorylation of sites in the ICLs and/or C terminus by G protein-coupled receptor kinases (GRKs) and/or second messenger-dependent kinases, such as protein kinase C (PKC) (Ferguson 2001; Kristiansen 2004). Consistent with this paradigm, BommoPBANR internalization was blocked by the general kinase inhibitor staurosporine (Hull et al. 2005) and significantly impaired following double Ala-substitution of two consensus PKC sites in the BommoPBANR C terminus, S333 and S366 (Hull et al. 2011). In support of PKC-mediated phosphorylation as an internalization trigger, RNAi knockdown of endogenous PKC in Sf9 insect cells also blocked PBANR internalization (Hull et al. 2011). Furthermore, localization of S366 within the 10-aa region (i.e. residues 357–367) critical for ligand-induced internalization (Fig. 8.4c) and the incomplete blockage of internalization following Ala-substitution of the YxxΦ motif are consistent with S366 functioning as a pivotal site for PBANR internalization. Sequence alignments have shown that both the S333 and S366 PKC sites are highly conserved in other PBANRs, which may indicate that feedback regulation of this class of receptors is evolutionarily conserved. Although PKC sites are predicted in the C terminus of most DHRs, the S366 site has not been conserved, providing additional evidence that DHR regulation may proceed via a different pathway.

3 PBAN Signal Transduction

The driving element of numerous studies over the years has been to elucidate the molecular basis underlying conversion of the external PBAN signal into the biological response of pheromone production and release. Initial studies sought to unravel the complex signaling interconnections by examining the effects of various pharmacological compounds (both inhibitors and activators) on pheromonogenesis. While data generated using these compounds can be ambiguous given the possibility of non-specific pharmacological effects and target specificity that varies with concentration (e.g. NaF at 10 mM acts as a phosphatase inhibitor but at 1–2 mM acts as a G protein activator), it can provide insights into potential mechanisms. Advances in molecular techniques, in particular the applicability of RNAi, have provided additional tools to decipher the molecular components underlying the PBAN signaling cascade. This cascade, which has been most extensively elucidated in heliothine moths (H. zea, H. virescens, and H. armigera) as well as B. mori, is now thought to diverge depending on the step in the biosynthetic pathway that is ultimately activated (i.e. early step vs late step).

3.1 G Protein Activation

The initial step in most extracellular signal transduction cascades requires dissociation of heterotrimeric guanine nucleotide binding proteins (i.e. G proteins α, β, and γ) from cell surface receptors and subsequent activation of downstream effectors (Cabrera-Vera et al. 2003). Receptor association/dissociation is dependent on the guanine nucleotide binding and hydrolysis activity of Gα subunits, which have been classified based on sequence variation and effector pathways activated into five sub-types: Gαs (stimulate cAMP production), Gαi/o (inhibit cAMP production), Gαq (stimulate Ca²⁺ influx), Gα_t (phototransduction), and Gα_12/13 (actin cytoskeletal remodeling) (Cabrera-Vera et al. 2003; Meigs and Lyakhovich 2012). Prior to PBANR identification, PBAN-induced elevation of cAMP levels (Rafaeli and Soroker 1989; also see Sect. 3.3.1) and the pheromonotropic effects of NaF (1–2 mM) on isolated PGs (Rafaeli and Gileadi 1996a) suggested the involvement of G proteins in the PBAN signal transduction cascade. Using homology-based cloning and genomic mining methods, transcripts for four Gα subunits (two Gαs, a Gαo, and a Gαq) were amplified from B. mori PGs (Hull et al. 2007a, 2010). Sequential RNAi knockdown of the four Gα subunits revealed that only Gαq had a role in transmitting the PBAN pheromonotropic signal (Hull et al. 2010).

3.2 PBAN-Induced Influx of Ca²⁺

3.2.1 Essential Role of Extracellular Ca²⁺

Initial studies using isolated PGs from diverse moth species demonstrated that the pheromonotropic effects of PBAN require extracellular Ca²⁺ (Jurenka et al. 1991a; Fónagy et al. 1992d; Jurenka et al. 1994; Ma and Roelofs 1995; Matsumoto et al. 1995b; Soroker and Rafaeli 1995; Zhao et al. 2002; Choi and Jurenka 2004, 2006; Hull et al. 2007a). Moreover, pharmacological manipulation (e.g. ionomycin, A23187, or thapsigargin) of intracellular Ca²⁺ levels could trigger pheromone production (Jurenka et al. 1991a; Fónagy et al. 1992c, d; Rafaeli 1994; Jurenka et al. 1994; Ma and Roelofs 1995; Matsumoto et al. 1995a, b; Soroker and Rafaeli 1995; Rafaeli and Gileadi 1996a; Zhao et al. 2002; Hull et al. 2007a), whereas inorganic Ca²⁺ channel blockers inhibited pheromone production (Jurenka et al. 1991a; Fónagy et al. 1992d; Ma and Roelofs 1995; Matsumoto et al. 1995b; Choi and Jurenka 2004). Taken together, these findings provided indirect evidence for PBAN-dependent opening of cell surface ion channels and the concomitant influx of Ca²⁺. Subsequent advances in fluorescent Ca²⁺ imaging techniques provided direct evidence for the rise in intracellular Ca²⁺ in response to PBAN binding in isolated H. zea and B. mori PGs (Choi and Jurenka 2006; Hull et al. 2007a).

3.2.2 Identification of the PBAN-Activated Ca²⁺ Channels

The most pervasive Ca²⁺-permeable ion channels in cells are voltage-operated channels (VOCs) (Lacinova 2005) and receptor-activated Ca²⁺ channels (RACCs) (Prakriya and Lewis 2015; Redondo and Rosado 2015), which include diacylglycerol (DAG)-dependent channels and store-operated channels (SOCs). Consistent with the early prediction of receptor involvement, VOC blockers had no effect on pheromone production in H. zea (Jurenka et al. 1991a; Choi and Jurenka 2006) or B. mori (Hull et al. 2007a), whereas SKF-96365, an inhibitor of both VOC and RACC, had pronounced pheromonostatic effects in H. virescens and B. mori (Jurenka 1996; Hull et al. 2007a). Further pharmacological manipulation of channel activity using inhibitors/activators of SOCs suggested that PBAN signals through an SOC pathway rather than a DAG-dependent channel (Hull et al. 2007a).

For many systems, the SOC pathway consists of stromal interaction molecule 1 (STIM1) functioning as a Ca²⁺ sensor and Orai1 as the pore-forming unit of the channel (López et al. 2016). Consistent with a role in the PBAN-activated SOC pathway, targeted knockdown of B. mori homologs of STIM1 and Orai1 negatively affected pheromone production without affecting non-pheromonotropic enzyme activities (Hull et al. 2009). The dependence on extracellular Ca²⁺ in PBAN-regulated pheromone pathways and the presence of STIM1 and Orai1 transcripts in moth PG transcriptomes (Ding and Löfstedt 2015) suggests that the STIM1-Orai1 SOC pathway is likely conserved in moths.

3.3 Role of Other Second Messengers

3.3.1 cAMP

While extracellular Ca²⁺ has been shown to be an absolute requirement for pheromonotropic activity in every moth species studied to date, the role of cAMP in the PBAN signal cascade appears to be species-dependent. Early cAMP radioimmunoassays demonstrated a PBAN-mediated increase of cAMP levels in isolated H. armigera PGs (Rafaeli and Soroker 1989; Rafaeli 1994; Soroker and Rafaeli 1995; Rafaeli and Gileadi 1996a). Furthermore, pharmacological manipulation (e.g. cAMP analogs, phosphodiesterase inhibition, or adenylate cyclase activation) of PG cAMP levels promoted pheromone production in H. armigera (Rafaeli and Soroker 1989; Soroker and Rafaeli 1995; Rafaeli and Gileadi 1996a), H. zea (Jurenka et al. 1991a), H. virescens (Jurenka 1996), and Argyrotaenia velutinana (Jurenka et al. 1994). In contrast, similar studies failed to find cAMP-linked pheromonotropic effects in B. mori (Fónagy et al. 1992d), S. litura (Matsumoto et al. 1995b), or O. nubilalis (Ma and Roelofs 1995) and no evidence was found of PBAN-mediated cAMP elevation in B. mori PGs (Hull et al. 2007b). There is, however, a strong correlation between this second messenger event and the pheromone biosynthetic activity under PBAN control. In species that utilize cAMP, the pheromonotropic control point resides in fatty acid biosynthesis, most likely the acetyl-CoA carboxylase (Tang et al. 1989; Jurenka et al. 1991b; Tsfadia et al. 2008). However, in species that do not undergo cAMP elevation, PBAN regulates a step(s) further along in the biosynthetic pathway, usually fatty acyl reduction (Fabriàs et al. 1994; Ma and Roelofs 1995; Ozawa et al. 1995; Ozawa and Matsumoto 1996; Moto et al. 2003; Eltahlawy et al. 2007) and, in B. mori, a second step involving cytoplasmic lipid droplet lipolysis (Fónagy et al. 2000; Ohnishi et al. 2006). While the evidence is currently too limited to draw broad conclusions regarding the relationship between cAMP signaling and PBAN regulation, the predictable associations suggest an avenue of potential research, in particular within species (Thaumetopoea pityocampa, M. sexta, Sesamia nonagrioides) in which the pheromonotropic control point is known be a step late in biosynthesis (Fabriàs et al. 1995; Fang et al. 1995; Mas et al. 2000) or species (Ostrinia furnacalis, M. brassicae, Dendrolimus punctatus, P. separata) where PBAN regulates a step in the fatty acid pathway (Jacquin et al. 1994; Zhao and Li 1996; Zhao et al. 2002; Fónagy et al. 2011; Köblös et al. 2015). It would likewise be interesting to examine the role of the PBANR variants in the contrasting signal transduction cascades. Jurenka and Rafaeli (2011) proposed that structural variations in the C-terminal lengths of the PBANR variants may contribute to the differing downstream responses with shorter C-terminal tail PBANRs linked to cAMP dependent pathways and the longer C-terminal PBANRs linked to Ca²⁺ influx alone.

3.3.2 IP₃

Similar to Ca²⁺ and cAMP, the phosphoinositide IP₃ (inositol 1, 4, 5-triphosphate) is a signal transduction messenger. IP₃ is generated from phospholipase C (PLC)-mediated hydrolysis of PIP₂ (phosphatidylinositol-4,5-bisphosphate) in response to receptor activation and typically functions in the propagation of receptor-mediated Ca²⁺ signaling by mobilizing intracellular Ca²⁺ stores (Balakrishnan et al. 2015). An early study on the PBAN mode of action reported that pheromonotropic activity of H. armigera PGs was reduced following pharmacological depletion of IP₃ (Rafaeli 1994). A later study in B. mori reported that total inositol phosphate levels in isolated PGs rose in response to PBAN and that RNAi knockdown of a putative IP₃ receptor suppressed pheromone production (Hull et al. 2010). These findings implicated PBANR-mediated activation of PLC. In support of this, pharmacological inhibition of PLC activity with either U73122 or compound 48/80 negatively impacted pheromone production in B. mori, whereas the inactive analog of U73122 had no effect (Hull et al. 2010). The pheromonostatic effects of compound 48/80, however, differed from a previous study that found no effect on B. mori pheromone production (Matsumoto et al. 1995a). Given that the preponderance of evidence available with the more recent study strongly pointed to PLC activity, the contrasting result was attributed to methodological differences. Separate studies demonstrating the critical importance of SOC components STIM1 and Orai1 (see Sect. 3.2.2 and Hull et al. 2009) in pheromone production likewise implicated PLC activity.

3.4 PBAN-Mediated PLC Activity

PCL-dependent activation of SOCs is predominantly driven by PLCβ and PLCγ (Drin and Scarlata 2007). PLCβ is generally activated downstream of GPCRs (Drin and Scarlata 2007), whereas PLCγ functions downstream of tyrosine kinase and non-receptor tyrosine kinases (Patterson et al. 2005). Using genomic mining methods, PLCβ1, PLCβ4, and PLCγ transcripts were amplified from B. mori PGs (Hull et al. 2010). Consistent with the expected signaling paradigm, RNAi-mediated knockdown of PLCβ1 significantly reduced pheromone production. PLCγ knockdown likewise mitigated the pheromonotropic effects of PBAN (Hull et al. 2010). Based on findings in other systems (Patterson et al. 2005), PLCγ was postulated to function in PBAN signaling as a molecular scaffold that stabilizes the protein-protein interactions essential for formation of the SOC complex rather than catalyzing PIP₂ hydrolysis.

3.5 Signal Transduction Post-PBAN-Mediated Ca²⁺ Influx

3.5.1 Calmodulin

As discussed above, the role of cAMP in PBAN signaling appears to differentiate the enzymatic step in the respective sex pheromone biosynthetic pathways under PBAN control. The GPCR-mediated generation of cAMP can be an indication that the receptor couples through Gαs, which stimulates adenylate cyclase activity following receptor dissociation. However, cAMP production in H. armigera reportedly occurred downstream of Ca²⁺ influx (Soroker and Rafaeli 1995), suggesting the involvement of a Ca²⁺-dependent adenylate cyclase. Additional pharmacological profiling of the PBAN cascade revealed that inhibition of calmodulin, a multi-functional Ca²⁺ binding protein that interacts with diverse proteins, blocked the PBAN-mediated increase of cAMP in H. armigera (Rafaeli and Gileadi 1996a) and mitigated the pheromonotropic effects of PBAN in H. armigera (Soroker and Rafaeli 1995) as well as S. litura and B. mori (Matsumoto et al. 1995a, b; Ozawa and Matsumoto 1996). In support of these results, a calmodulin homolog identical to the D. melanogaster protein was purified from B. mori PGs (Iwanaga et al. 1998). Among the enzymatic activities reportedly mediated by Ca²⁺-bound calmodulin are adenylate cyclases (Halls and Cooper 2011), suggesting that the Ca²⁺-dependent increase in cAMP observed in heliothine moths is likely driven by one of these cyclases. Because many calmodulin interacting proteins are directly or indirectly involved in protein phosphorylation, the results observed in S. litura and B. mori, neither of which utilizes cAMP in PBAN signaling, may be attributable to impaired phosphorylation cascades.

3.5.2 Kinase Activity

GPCR-mediated activation of biosynthetic pathway enzymes typically involves a phosphorylation cascade driven by diverse kinase (phosphorylation) and phosphatase (dephosphorylation) steps. The generation of cAMP, the critical role of calmodulin, and the importance of PKC in feedback regulation of BommoPBANR in vitro (see Sect. 2.7.8) strongly suggested kinase activity in PBAN signaling. While early studies assessing the effect of both broad spectrum and specific kinase inhibitors found no effect on pheromone production in either B. mori (Matsumoto et al. 1995a) or H. armigera (Soroker and Rafaeli 1995), the PKC activator, phorbol 12-myrstate 13-acetate (PMA), was found to have pheromonotropic activity in H. armigera (Soroker and Rafaeli 1995). This effect, however, did not extend to B. mori or S. litura (Matsumoto et al. 1995b; Ozawa et al. 1995). A more recent study using anti-phosphoamino acid antibodies found clear evidence of PBAN-mediated phosphorylation in B. mori (Ohnishi et al. 2011). Furthermore, RNAi-mediated knockdown of a Ca²⁺-bound calmodulin dependent kinase II (CaMKII) in B. mori PGs reduced PBAN-induced pheromone production and diminished phosphorylation of a critical lipid droplet-associated protein, whereas knockdown of putative protein kinase A (PKA) and PKC transcripts had no effect (Ohnishi et al. 2011).

3.5.3 Phosphatase Activity

In contrast to the early kinase inhibitor studies, pharmacological inhibition of phosphatase activity had pronounced pheromonostatic effects in B. mori (Matsumoto et al. 1995a, b; Ozawa and Matsumoto 1996; Fónagy et al. 1999) as well as H. zea and H. virescens (Jurenka 1996). Inhibition of ionophore-induced pheromone production in H. zea suggested that phosphatase activity occurs downstream of Ca²⁺ influx (Jurenka 1996), thus ruling out an effect similar to LiCl, which inhibits IP₃ generation. The effectiveness of inhibitors specific for calcineurin (Fónagy et al. 1999), a protein phosphatase b activated by Ca²⁺-bound calmodulin, was consistent with previous studies demonstrating calmodulin activity. In support of this role, both calcineurin subunits were amplified from B. mori PGs (Yoshiga et al. 2002). Determination of the rate-limiting steps in heliothine moths and B. mori suggest that calcineurin or calcineurin-like phosphatase activity comprises the penultimate control point in PBAN signaling. In heliothine moths, PBAN activates acetyl-CoA carboxylase, the critical point in fatty acid biosynthesis that catalyzes carboxylation of acetyl-CoA to yield malonyl-CoA. In B. mori (and other moths), PBAN regulates a fatty acyl reductase that shares biochemical characteristics with HMG-CoA reductase (Ozawa et al. 1995). In both cases (i.e. acetyl-CoA carboxylase and HMG-CoA reductase), enzymatic activity is phosphorylation-dependent (Zammit and Easom 1987; Brownsey et al. 2006).

3.6 Model of Pheromone Regulation by PBAN Signaling

Based on diverse studies spanning more than 20 years (many of which were briefly described above), a model for the molecular signaling cascade underlying PBAN-mediated regulation of pheromone production has emerged (Fig. 8.5). Circadian activation of extero-receptors and brain hormones such as allatotropins/allatostatins that influence JH biosynthesis (Cusson and McNeil 1989; Woodhead et al. 1989; Picimbon et al. 1995; Stay and Tobe 2007) may have a role in PBAN release into the hemolymph where it interacts with PBANRs localized at the plasma membrane of PG cells. The ensuing conformational change in PBANR results in dissociation of the heterotrimeric G protein complex with subsequent Gαq activation of PLCβ1-mediated hydrolysis of PIP2 into DAG and IP₃. The soluble IP₃ diffuses through the cytosol to activate IP₃ receptors in the endoplasmic reticulum (ER) membrane, which promotes release of stored Ca²⁺. The drop in luminal Ca²⁺ levels results in translocation of STIM1 to the plasma membrane where it triggers an influx of extracellular Ca²⁺ through Orai1 channels, presumably via interactions with a scaffolding complex that includes PLCγ. The concomitant rise in intracellular Ca²⁺ allows for formation of Ca²⁺-calmodulin complexes, at which point the pathway exhibits species-dependent divergence. In heliothines and species that utilize cAMP, the Ca²⁺-calmodulin complexes stimulate adenylyl cyclase activity. The rise in cAMP then drives a cascade culminating in activation of the fatty acid biosynthetic pathway enzyme, acetyl CoA-carboxylase. In B. mori, and presumably species in which PBAN regulates a step late in pheromonogenesis, the Ca²⁺-calmodulin complexes activate both calcineurin (a protein phosphatase) and calmodulin-dependent kinase II (CamKII). Calcineurin in turn activates fatty acyl reductase, the terminal step in pheromone biosynthesis, while CamKII-dependent phosphorylation of lipid storage droplet protein-1 promotes lipolytic release of stored pheromone precursors (Fig. 8.5).

Fig. 8.5

Proposed PBAN signal transduction cascade. (1) PBAN circulating in the hemolymph binds to PBANR in the plasma membrane of PG cells. (2) PBAN binding promotes dissociation of Gαq from PBANR with subsequent activation of PLCβ1. (3) PLC-mediated hydrolysis of PIP₂ yields DAG and soluble IP₃. (4) Cytosolic IP₃ interacts with IP₃ receptors in the ER membrane. (5) Activation of IP₃ receptors promotes release of stored Ca²⁺. (6) The reduction in ER luminal Ca²⁺ levels promotes interactions between STIM1 and Orai1 channels in the plasma membrane. The resulting complex formation may be stabilized by protein-protein interactions with SH3 domains in PLCγ. (7) The activated Orai1 channels open allowing an influx of extracellular Ca²⁺. (8) Free calmodulin complexes with the intracellular Ca²⁺. (9a) In heliothines, the Ca²⁺-calmodulin complex stimulates adenylate cyclase activity and the production of cAMP, which subsequently initiates a protein kinase A/C phosphorylation cascade. PBAN signaling culminates in activation of acetyl-CoA carboxylase, the limiting step in fatty acid biosynthesis. Given evidence in the literature that this enzyme is activated in response to dephosphorylation and that pharmacological inhibition of phosphatase activity in H. zea and H. virescens has pheromonostatic effects, it is likely that a protein phosphatase, possibly calcineurin, may function in acetyl-CoA carboxylase activation. (9b) In B. mori, calcineurin is activated by the Ca²⁺-calmodulin complex, which also activates (9c) a calmodulin-dependent protein kinase II (CamKII). (10a) CamKII phosphorylates a lipid droplet storage protein critical for lipolytic release of pheromone precursors stored in cytosolic lipid droplets. (10b) Calcineurin dephosphorylates fatty acyl reductase, the terminal enzymatic reaction in the B. mori pheromone biosynthetic pathway. Abbreviations: cAMP cyclic adenosine 3′, 5′-monophosphate, DAG diacylglycerol, ER endoplasmic reticulum, GDP guanosine diphosphate, Gq G protein α subunit q, GTP guanosine-5′-triphosphate, IP ₃ inositol 1,4,5-trisphosphate, PIP2 phosphatidylinositol (4,5)-bisphosphate, PLC phospholipase C, STIM1 stromal interaction molecule 1

4 Targeted Disruption of PBAN Pathway

Current integrated pest management strategies that focus on mating disruption frequently exploit synthetic pheromone blends (Witzgall et al. 2008; El-Sayed et al. 2009). However, for species that utilize multi-component pheromone blends with cost prohibitive chemistries, targeted disruption of pheromone biosynthetic pathways has significant potential as an alternative control measure. This is the case for the black cutworm moth, A. ipsilon, a polyphagous, polyandrous pest with multi-continental populations and intra-specific genetic variations (Wakamura et al. 1986; Picimbon et al. 1995, 1997; Gadenne et al. 1997; Duportets et al. 1998; Gemeno and Haynes 1998; Gemeno et al. 2000; Du et al. 2015). Insect GPCRs in particular have been proposed as promising targets for the next generation of insecticides (Scherkenbeck and Zdobinsky 2009; Van Hiel et al. 2010; Bai and Palli 2013; Grimmelikhuijzen and Hauser 2013; Audsley and Down 2015). This interest has driven significant efforts in developing peptidomimetics that overcome limitations (i.e. environmental instability, poor cuticular penetrance, and susceptibility to proteolytic degradation in the hemolymph) inherent to peptides that make them unsuitable for pest management. Because this topic has been extensively reviewed elsewhere (Altstein 2001, 2004b; Nachman et al. 2009a; Scherkenbeck and Zdobinsky 2009), we provide only a brief overview of some of the most intriguing developments.

4.1 Peptidomimetics

4.1.1 PBAN Agonists

PBAN agonists, small molecules that activate the receptor in the absence of the endogenous ligand, provide valuable insights into the structural requirements and chemistries crucial for ligand binding and cuticular penetration. In addition, they offer possibilities in pest management as continuous pheromonogenic stimulation via a bound agonist could lead to pheromone release asynchronous with male mating behaviors and/or depleted pheromone. Early peptide engineering studies revealed that modification of the terminal Phe in the pentapeptide FTPRL-NH₂ with a hydrophobic cage-like o-carborane moiety (a cluster composed of boron, carbon, and hydrogen), 1-pyrenebutyric acid, 9-fluoreneacetic acid, or 2-amino7-bromofluorene yielded topically active pheromonogenic analogs with enhanced cuticular penetrance and greater hemolymph persistence (Nachman et al. 1996; Teal and Nachman 1997, 2002). Additional studies incorporating β-amino acids further highlight the importance of the Phe residue for pheromonotropic activity (Nachman et al. 2009a).

4.1.2 PBAN Antagonists

The structural, conformational and dynamic features of agonists can serve as the basis for rational design of antagonists, which require the compound to bind the receptor without activating the signal transduction cascade. Replacing the Thr in the pheromonogenic septapeptide RYFTPRL-NH₂ with D-Phe yielded a linear peptide antagonist that significantly inhibited pheromone production following injection (Zeltser et al. 2000). Backbone cyclization techniques have also yielded antagonists with pheromonostatic effects that can persist for several hours (Altstein et al. 2000). A linear RYF[dF]PRL-NH₂ analog that incorporated an aliphatic amine exhibited enhanced cuticular penetration while retaining pheromonostatic properties (Nachman et al. 2009b).

4.1.3 Receptor Selective Analogs

FxPRL-NH₂ analogs have been reported to have differing receptor effects depending on the activity assayed (e.g. melanotropic vs pheromonotropic) despite mediation of both activities by the same peptidergic sequence (Matsumoto et al. 1992; Altstein et al. 1996) and receptor (Zheng et al. 2007; Kim et al. 2008). Sequential D-Phe scan of a modified PBAN sequence (YFSPRL-NH₂) generated a selective antagonist that significantly reduced pheromone production with no effect on pupal melanization (Ben-Aziz et al. 2005). An amphiphilic version of the antagonist, that incorporated an aliphatic amine via succinic acid at the N terminus of the pentapeptide, retained selective antagonist properties while exhibiting enhanced cuticular penetrance (Nachman et al. 2009b). Similarly, replacement of the critical Phe with a β-homo-amino acid yielded an analog that affected melanization but had no effect on pheromone production (Nachman et al. 2009a). Incorporation of a dihydroimidazoline moiety into the FxPRL-NH₂ hexapeptide sequence likewise generated a selective melanotropic antagonist devoid of either pheromonotropic or pheromonostatic activities (Nachman et al. 2010). The selectivity observed in these peptidomimetic studies suggests that the melanotropic receptor tolerates greater conformational deviations in the ligand than the pheromonotropic receptor. This ligand selectivity is corroborated by both in vitro and in silico studies of FxPRL-NH₂ receptors that show dissimilar three-dimensional conformations, electrostatic potentials, and ligand preferences (Hariton-Shalev et al. 2013; Shalev and Altstein 2015). While the development of selective antagonists will undoubtedly provide additional insights into the development of novel pest management agents, it is apparent that despite years of study, our understanding of FxPRL-NH₂ pleiotropism at the molecular level will remain a fertile area of research.

4.2 RNAi: The New Frontier?

As a biorational approach that can be specifically tailored to individual pest species, RNAi holds great promise for the future of insect pest management (Price and Gatehouse 2008; Burand and Hunter 2013). Though still in its infancy, the viability of using transgenic plants that trigger RNAi-mediated suppression of select pest genes has been effectively demonstrated (Baum et al. 2007; Mao et al. 2007, 2011; Pitino et al. 2011). While those studies focused on the control potential associated with knockdown of diverse enzymes, current studies assessing the effects of neuropeptide/GPCR RNAi knockdown on peptidergic regulation of insect biology (e.g. Terhzaz et al. 2007; Arakane et al. 2008; Badisco et al. 2011; Bai et al. 2011; Terhzaz et al. 2015; Zandawala et al. 2015) may provide an additional biorational set of tools for the development of next generation pest management strategies.

4.2.1 RNAi-Knockdown: PBAN

To date, RNAi-mediated knockdown of PBAN has only been reported for two species, H. zea (Choi et al. 2012) and S. litura (Lu et al. 2015). In both species, injection of double-stranded RNAs (dsRNAs) corresponding to a fragment of the respective DH-PBAN gene markedly reduced sex pheromone production. In H. zea, however, the PBAN dsRNA injections, which were performed using 4–5 day old female pupae, also affected adult emergence with a significantly higher percentage of injected pupae unable to eclose (Choi et al. 2012). A similar phenotype was reported in another heliothine moth following knockdown of PBAN, but not PBANR, suggesting that the failure to eclose properly may be linked to DH, which functions in termination of pupal diapause in heliothine moths (Xu and Denlinger 2003; Sun et al. 2003).

4.2.2 Genome Editing: PBAN

Advances in genome editing methodologies have extended targeted gene mutagenesis capabilities. One such approach utilizes Transcription Activator-Like Effector Nucleases (TALENs) to introduce small deletions or insertions at the gene level that cause frameshift mutations/truncations. Recently, Shiomi et al. (2015) used this method to make targeted deletions in the B. mori DH-PBAN gene yielding prepropeptides severely truncated within the signal peptide region precluding generation of the PBAN sequence. While the mutations clearly affected the induction of embryonic diapause, the pheromonogenic effects, which were not the focus of the study and were thus only assessed superficially, appeared to be muted with a slight reduction in the male behavioral response.

4.2.3 RNAi-Knockdown: PBANR

PBANR transcripts have been knocked-down in B. mori (Ohnishi et al. 2006), P. xylostella (Lee et al. 2011), and H. armigera (Bober and Rafaeli 2010). Injection of dsRNAs corresponding to a 417-nt fragment of BommoPBANR into 1-day-old pupae triggered receptor knockdown and significantly impaired sex pheromone production and disrupted lipolysis of cytoplasmic lipid droplets (Ohnishi et al. 2006). Similarly, knockdown of PluxyPBANR in pupae 1 day prior to adult emergence with dsRNAs corresponding to a 549-nt fragment resulted in a ~50% reduction in sex pheromone production and a 20–40% reduction in female mating (Lee et al. 2011). That group also reported decreased expression of two desaturases thought to be involved in the P. xylostella sex pheromone biosynthetic pathway following PluxyPBANR knockdown (Lee and Kim 2011). Unlike B. mori and P. xylostella, the effects of HelarPBANR knockdown were evaluated in adult male moths. An earlier study reported expression of HelarPBANR in the male aedeagus, a reproductive organ adjacent to the male abdomen through which sperm from the testis is transferred during copulation and which is usually associated with male-derived sex pheromone-like compounds (Rafaeli et al. 2007). Injection of dsRNAs corresponding to a 880-nt fragment of HelarPBANR in 1-day-old adult male H. armigera significantly reduced PBAN-stimulated production of male volatile compounds (Bober and Rafaeli 2010). While the relevance of these compounds in H. armigera mating behavior remains to be demonstrated, similar compounds have been linked to stimulation of female receptivity and inhibition of male competition (Teal and Tumlinson 1984; Kehat and Dunkelblum 1990; Huang et al. 1997; Hillier and Vickers 2004; Hillier et al. 2006). In the European corn borer, O. nubilalis, the male scent odor is crucial for the acceptance of the male by the female (Royer and McNeil 1992; Picimbon 1996; Farrell and Andow 2017). Regardless, the results demonstrate that in a wide variety of moths the role of PBANR functionality in pheromone biosynthesis is certainly not restricted to females and further underscores the pleiotropic nature of the receptor and its multifunctional ligand.

5 Concluding Remarks

The past 30 years have witnessed significant progress in our understanding of pheromonogenesis in moths and its neuroendocrine regulation. Interestingly, rather than clarifying our understanding of pheromonotropic control, elucidation of the “black box” has illuminated yet another layer of complexity and provided new puzzles for us to unravel.

Some of the questions raised with this new framework of entomology, chemical ecology, physiology and molecular biology research that we find the most intriguing include:

• What is the molecular basis for regulation of the pleiotropic FxPRL-NH₂ peptide/receptor system?

• How is ligand selectivity of PBANRs/DHRs achieved?

• What is the evolutionary significance of the different control points (fatty acid biosynthesis vs terminal modification) in the PBAN pathway, and how did this divergence arise?

• What biological role do the concomitantly expressed PBANR variants play in PBAN signaling?

• How are transcription and alternative splicing of PBANR regulated?

Undoubtedly, rapid developments in mRNA sequencing, bioinformatics, molecular engineering, and proteomics will play a significant role in resolving these new questions. In addition, advances such as CRISPR in insect genome editing (Taning et al. 2017), and RNAi (see Chap. 5), despite the current limitations of this technology in lepidopterans (Terenius et al. 2011), can provide unequivocal demonstration of the roles calmodulin, calcineurin, and acetyl-CoA carboxylase have in heliothine pheromonogenesis and finally reveal how conserved PBAN signaling pathways function across species. Similar application of these technologies can also provide insights into the role of antagonistic peptidomimetics in receptor regulation.

Continued research into the mechanisms underlying PBANR function in moths, as well as related receptors in other species, will help answer questions regarding the biological significance of the FxPRL-NH₂ family and how alternative splicing plays a role in mediating that biology. This knowledge will provide insights into the complexities of GPCRs, and can potentially be applied towards the development of novel biorationally designed insect control agents. These fundamental studies will also continue to provide insights into mammalian endocrinology, lipid biology, and the molecular interactions underlying peptidergic binding/activation of pleiotropic GPCRs.