Gqα-linked phospholipase Cβ1 and phospholipase Cγ are essential components of the pheromone biosynthesis activating neuropeptide (PBAN) signal transduction cascade

Abstract

Sex pheromone production for most moths is regulated by pheromone biosynthesis activating neuropeptide (PBAN). In Bombyx mori, PBAN binding triggers the opening of store-operated Ca²⁺ channels, suggesting the involvement of a receptor-activated phospholipase C (PLC). In this study, we found that PLC inhibitors U73122 and compound 48/80 reduced sex pheromone production and that intracellular levels of ³H-inositol phosphate species increased following PBAN stimulation. In addition, we amplified cDNAs from pheromone glands corresponding to PLCβ1, PLCβ4, PLCγ and two G protein α subunits, Go and Gq. In vivo RNA interference-mediated knockdown analyses revealed that BmPLCβ1, BmGq1, and unexpectedly, BmPLCγ, are part of the PBAN signal transduction cascade.

Introduction

Reproductive behaviour for most moth species is dependent on the ability of mature females to attract conspecific males. This crucial event is facilitated by sex pheromones, which are species-specific multi-component blends of relatively simple C₁₀–C₁₈ aliphatic compounds that are de novo synthesized and released, most usually, by female moths (Jurenka, 2003). Sex pheromone fatty acid precursors are produced in the pheromone gland (PG) from acetyl-CoA with intermediate compounds frequently modified by unique desaturases prior to reductive modification of the carbonyl carbon to generate bioactive end products consisting of aldehydes, alcohols or acetate esters (Jurenka, 2003). In most moth species, sex pheromone biosynthesis is regulated by pheromone biosynthesis activating neuropeptide (PBAN) (Kitamura et al., 1989; Raina et al., 1989). PBAN originates in the suboesophageal ganglion and acts directly on the PG following adult emergence (Rafaeli, 2009) to stimulate sex pheromone production by regulating, in a species dependent manner, either the initial or the terminal steps in sex pheromone biosynthetic pathways. In Argyrotaenia velutinana (Tang et al., 1989), Helicoverpa zea (Jurenka et al., 1991a), Cadra cautella (Jurenka, 1997), Spodoptera exigua (Jurenka, 1997) and Mamestra brassicae (Jacquin et al., 1994), PBAN regulates a step (or steps) in fatty acid biosynthesis, most likely acetyl-CoA carboxylase as shown in Helicoverpa armigera (Tsfadia et al., 2008). In Thaumetopoea pityocampa (Arsequell et al., 1990), Spodoptera littoralis (Martinez et al., 1990) and Manduca sexta (Fang et al., 1995) the pheromonotropic control point is the terminal reductase, while in Bombyx mori, PBAN regulates lipolysis of the stored sex pheromone precursor fatty acid as well as the terminal reductase (Matsumoto et al., 2007). Interestingly, the differing points of PBAN regulation appear to be linked to adenylyl cyclase activity as those species in which fatty acid biosynthesis is modified by PBAN are dependent on cyclic adenosine 3′, 5′-monophosphate (cAMP) production whereas those species, such as B. mori, that rely on PBAN activation of the terminal reductase are not (Hull et al., 2007b; Rafaeli, 2009). Irrespective of species, however, the pheromonotropic effects of PBAN have been shown to be dependent on the presence of extracellular Ca²⁺ (Rafaeli & Jurenka, 2003), suggesting that the opening of cell surface cation channels and the concomitant influx of extracellular Ca²⁺ represent the most conserved, and crucial, point of PBAN control.

The influx of extracellular Ca²⁺ is tightly regulated by various Ca²⁺ permeable cation channels. Receptor-activated Ca²⁺ channels (RACCs) are an important class of diverse Ca²⁺ channels that open downstream of ligand-mediated receptor activation (Berridge, 2001). Store-operated channels (SOCs) represent a specific subset of RACCs that are dependent on the depletion of endoplasmic reticulum (ER) Ca²⁺ stores, an event mediated by soluble inositol 1,4,5-trisphosphate (IP₃) interaction with specific IP₃ receptors (IP₃R) in the ER membrane (Parekh & Penner, 1997). It has recently been shown that depletion of ER Ca²⁺ stores results in translocation of stromal interaction molecule 1 (STIM1), an ER membrane protein, to the cell surface (Cahalan, 2009). Once at the cell surface, STIM1 interacts with Orai1, a four transmembrane cell surface protein shown to be the principal molecular constituent of most SOCs. These STIM1-Orai1 interactions promote the tetramerization of Orai1 and the concomitant influx of extracellular Ca²⁺ (reviewed in Cahalan, 2009).

The IP₃ signal that initiates SOC activation is generated by a receptor-activated phospholipase C (PLC). PLCs are Ca²⁺ dependent enzymes found throughout evolution that catalyse the hydrolysis of phosphatidylinositol (4,5)-bisphosphate into IP₃ and diacylglycerol. In mammals, PLCs comprise six major families (PLCβ, γ, δ, ε, ζ and η) consisting of multiple isoforms (Drin & Scarlata, 2007), with PLCβ and PLCγ the predominant isoforms involved in mediating SOCs. PLCβ activation is G protein-coupled receptor (GPCR) dependent (Drin & Scarlata, 2007), whereas PLCγ is activated downstream of tyrosine kinase receptors and nonreceptor tyrosine kinases (Patterson et al., 2005). Despite varying modes of activation the principal domain architecture of PLCs is conserved: an amino terminal plekstrin homology (PH) domain, multiple EF-hand motifs, a catalytic core composed of conserved X and Y domains and a C-terminal C2 Ca²⁺ dependent membrane-targeting domain. PLCβs are distinguished by a C-terminal extension, and PLCγ by a catalytic domain split by Src homology (SH)2 and SH3 protein interaction domains (Oude Weernink et al., 2007).

We have recently demonstrated that the channels opened downstream of PBAN binding are SOCs (Hull et al., 2007a) and that B. mori homologues of STIM1 and Orai1 are integral components of PBAN signalling (Hull et al., 2009). These results strongly implicate PLC involvement. In this paper, we sought to expand our understanding of the mechanisms underlying the PBAN cascade by defining the role of PLCs in transmitting the pheromonotropic signal.

Results

Effect of PLC inhibitors on B. mori sex pheromone production

As B. mori sex pheromone production is dependent on SOC activation, we sought to examine the effect of known PLC inhibitors on production of the principal B. mori sex pheromone component, bombykol (ie E,Z-10, 12-hexadecadien-1-ol). Bombykol levels were determined by high performance liquid chromatography (HPLC) using n-hexane extracts of isolated PGs stimulated with PBAN (Matsumoto et al., 1990) in the presence and absence of inhibitors. Using PGs pre-incubated for 30 min ± inhibitor, we found that U73122, a common PLC inhibitor shown to be highly effective in lepidopterans at 50 µM (Fellner et al., 2005), dose-dependently reduced bombykol production with a maximum inhibition of 40% at 50 µM (Fig. 1). No inhibition was observed in vehicle [dimethyl sulfoxide (DMSO)] treated PGs (Fig. 1) or with 50 µM U73343, the inactive analogue of U73122 (Fig. 1). Contrary to a previous study (Matsumoto et al., 1995), we also found that compound 48/80, a different PLC inhibitor (Miao et al., 1997), also significantly reduced bombykol production (∼50% reduction, P < 0.001; Fig. 1).

Figure 1. Effects of phospholipase C (PLC) inhibition on pheromone biosynthesis activating neuropeptide (PBAN)-stimulated bombykol production. Pheromone glands (PGs) were pre-incubated for 30 min in the presence of 10, 25 or 50 µM U73122 (PLC inhibitor), 50 µM U73343 (inactive analogue of U73122), or 5 mg/mL compound 48/80 (PLC inhibitor) and then incubated 90 min in the presence of inhibitors + 120 nM PBAN. U73122 and U73343 were dissolved in dimethyl sulfoxide (DMSO). Data are expressed as the percentage bombykol production relative to control PGs stimulated with 120 nM PBAN and represent mean values ± SEM of eight replicates from three independent experiments. Statistically significant differences from PBAN (***; P < 0.001) were determined by one-way anova with Tukey's multiple comparison test.

Effect of PBAN on total inositol phosphate levels

To assess more directly the role of PLC, we measured the effect of PBAN on the PLC-catalysed generation of inositol phosphate (IP) species. To inhibit the intracellular inositol phosphatases that contribute to short IP half-lives, we performed assays in the presence of LiCl (Berridge et al., 1982) and measured the level of total ³H-IPs in PGs pre-loaded with myo-[2-³H]inositol. PGs were pre-incubated for 15 min in a modified insect Ringer's buffer (RB) containing 10 mM LiCl and then stimulated for 30 min in RB/10 mM LiCl ± PBAN (120 nM). PGs were homogenized under acidic conditions and then neutralized, and total IPs were isolated using an anion exchange resin. Consistent with the above inhibition studies, we observed a significant increase in ³H-IPs levels in PGs incubated with PBAN compared to untreated PGs (Fig. 2), indicating that PBAN binding stimulates phosphatidylinositol (4,5)-bisphosphate hydrolysis. Taken together, these results are a clear indication that the PBAN signal cascade proceeds via PLC activation.

Figure 2. Level of intracellular ³H-inositol phosphates (³H-IPs) following pheromone biosynthesis activating neuropeptide (PBAN) stimulation. Pheromone glands (PGs) were pre-loaded en masse for 2 h with myo-[2-³H]inositol (3 µM final concentration). PGs were then incubated for 15 min in Ringer's buffer (RB) + 10 mM LiCl (cellular IPase inhibitor) prior to stimulation with RB ± 120 nM PBAN for 30 min. Results shown are the mean ± SEM from a representative experiment of three performed in duplicate. Statistical significance (**; P= 0.0026) determined via unpaired Student's t-test.

Molecular identification of B. mori PLC homologues

To facilitate molecular identification of the specific PLC activated downstream of PBAN binding, we searched the B. mori genome (Xia et al., 2004) for sequences predicted to be similar to the catalytic domains of known PLCs and identified portions of five non-overlapping contiguous sequences (contigs 159970, 457194, 503327, 600350 and 481400) as potential PLC catalytic domains. Surprisingly, four of the five putative PLC encoding regions (contigs 155970, 457194, 503327 and 600350) could be amplified from B. mori PG cDNA. To identify the respective 5′ and 3′ ends of the putative open reading frames (ORFs), we performed rapid amplification of cDNA ends PCR (RACE-PCR) with a B. mori p50 (inbred strain) PG-specific cDNA library (Yoshiga et al., 2000). Sequence analysis of the resulting products indicated that contigs 600350 and 457194 encoded different portions of the same transcript. No sequence information was obtained for contigs 155970 and 503327. Primers designed to amplify the putative ORF containing the contig 600350 sequence generated a 3210 nt amplimer predicted to encode a protein of 1070 amino acids with a molecular mass of 123 kDa and domains characteristic of PLCβ including a PH domain (residues 21–136), an EF hand-like Ca²⁺ binding domain (residues 221–312), phosphatidylinositol-specific phospholipase X-box (residues 313–463) and Y-box (residues 533–649) domains and a C-terminal C2 domain (residues 670–769). The predicted protein has highest sequence identity (78%) with a putative PLCβ4 from Aedes aegypti (XP_001653804) and significant homology (80%) with a characterized Drosophila melanogaster PLCβ4 (norpA, AAA28724.1; Bloomquist et al., 1988) and human PLCβ4 (CAI43092, 69% similarity). Based on these results we designated the cloned ORF as BmPLCβ4 (accession no. GU266211).

As the p50 PG-specific cDNA library proved ineffective at amplifying the terminal ends of the putative PLC encoding regions of contigs 155970 and 503327, we used homology to known and putative insect PLCs as a guide to facilitate in silico gene walking through the B. mori genome database. For contig 159970, a putative ORF was identified based on the high sequence identity between the contig 159970 coding sequence and other portions of the B. mori genome with a putative Ae. aegypti PLCγ. Specific primers designed to amplify the putative ORF generated an amplimer of 3603 nt predicted to yield a 1201 amino acid protein with a molecular mass of 138 kDa and domain architecture most similar to PLCγ: a PH domain (residues 24–140), an EF hand Ca²⁺ binding domain (residues 187–215), an EF hand-like domain (residues 224–308), a split phosphatidylinositol-specific phospholipase X and Y domain (X domain: residues 309–453; Y domain: residues 938–1055) separated by tandem SH2 domains (residues 543–634 and 655–737) upstream of a SH3 domain (residues 779–835) and a C2 domain (residues 1075–1178) (Fig. 3). The predicted protein has highest sequence identity (64%) with a putative PLCγ from Apis mellifera (XP_624101) and significant homology (64%) with a characterized D. melanogaster PLCγ (small wing; AAF48595.3; Thackeray et al., 1998) and human PLCγ (AAA36452.1, 66% similarity; Burgess et al., 1990). Based on these results we designated the cloned ORF as BmPLCγ (accession no. GU266212).

Figure 3. Bombyx mori phospholipase C γ (PLCγ) sequence alignment with other PLCγ. Protein sequence alignment was performed using the Network Protein Sequence Analysis CLUSTALW server (Combet et al., 2000). Species abbreviations for insect PLCγ are as per National Center for Biotechnology Information/Swiss Prot: B. mori, Bommo; Apis mellifera, Apime (XP_624101) and Drosophila melanogaster; Drome (AAF48595) (Thackeray et al., 1998). The alignment also included the human PLCγ homologue (AAA36452) (Burgess et al., 1990) with the highest similarity. Black and grey shading indicate sequence identity and conserved amino acid substitutions, respectively. Domains characteristic of PLCγ are boxed.

Using an analogous approach, we identified a putative ORF encompassing the coding sequence of contig 159970 based on sequence similarity between D. melanogaster PLCβ1 and a putative ORF present in B. mori cDNA clone fefu-PO9_F_017, which was obtained from a full-length B. mori (p50T strain) unfertilized egg cDNA library that was prepared via the V-capping method (Ohtake et al., 2004). Using specific primers based on contig 159970 and the fefu-PO9_F_017 clone sequence, we amplified a 3583 nt transcript containing a 3339 nt ORF predicted to encode a 1113 amino acid protein with a molecular mass of 126 kDa. The predicted protein domain architecture is most consistent with PLCβ: a PH domain (residues 4–140), an EF hand-like Ca²⁺ binding domain (residues 226–316), phosphatidylinositol-specific phospholipase X-box (residues 317–465) and Y-box (residues 548–664) domains and a C2 domain (residues 685–784) (Fig. 4). The protein has highest sequence identity (59%) with a recently cloned Spodoptera litoralis PLCβ1 (ACD69415) and significant homology (66%) with a characterized D. melanogaster PLCβ1 (plc21C; AAA28820; Shortridge et al., 1991) and human PLCβ1 (AAF86613, 55% similarity; Peruzzi et al., 2000). Consequently, this gene product was designated BmPLCβ1 (accession no. GU266210).

Figure 4. Bombyx mori phospholipase C β1 (PLCβ1) sequence alignment with other PLCβ1. Protein sequence alignment was performed using the Network Protein Sequence Analysis CLUSTALW server (Combet et al., 2000). Species abbreviations for insect PLCβ1 are as per National Center for Biotechnology Information/Swiss Prot: B. mori, Bommo; Spodoptera littoralis, Spoli (ACD69415) and Drosophila melanogaster, Drome (AAA28820) (Shortridge et al., 1991). The alignment also included the human PLCβ1 homologue (AAF86613) (Peruzzi et al., 2000) with the highest similarity. Black and grey shading indicate sequence identity and conserved amino acid substitutions, respectively. Domains characteristic of PLCβ are boxed.

Phylogenetic analysis provided further evidence for gene classification as the putative B. mori gene products sorted accordingly with various bona fide and putative mammalian and insect PLC isoforms (Fig. 5).

Figure 5. Phylogenetic analysis of vertebrate and insect phospholipase C β (PLCβ) and PLCγ. Phylogenetic analysis was performed with MacVector v. 11.0 using the neighbor-joining method (Saitou & Nei, 1987) with 1000 random bootstrapped replicates. Sequence accession numbers are listed to the right of either the common vertebrate species name or the abbreviated insect species name. Species abbreviations are as per National Center for Biotechnology Information/Swiss Prot: Drosophila melanogaster, Drome; Culex quinquefasciatus, Culqu; Aedes aegypti, Aedae; Apis mellifera, Apime, Tribolium castaneum, Trica; Nasonia vitripennis, Nasvi; Pediculus humanus corporis, Pedhu; Spodoptera littoralis, Spoli; Anopheles gambiae, Anoga. The Bombyx mori PLCs are highlighted in grey boxes.

In B. mori, expression of a number of genes involved in sex pheromone biosynthesis is restricted to the PG where they are up-regulated in response to adult emergence from the pupae (Matsumoto et al., 2007). Transcripts for BmPLCβ4, BmPLCβ1 and BmPLCγ, however, are not PG-specific as they could be amplified from multiple tissues (Fig. 6). Furthermore, we found no evidence of transcript up-regulation in relation to adult emergence (data not shown).

Figure 6. Reverse transcription PCR expression analyses of Bombyx mori phospholipase C β1 (BmPLCβ1), BmPLCβ4 and BmPLCγ. Tissues were dissected from newly eclosed adult female moths: FB, fat body; PG, pheromone gland; FM, flight muscle; Ant, antennae; L, leg; MT, Malpighian tubule; E, unfertilized egg. PCR products were separated on 1.3% agarose gels in Tris-acetate-EDTA (TAE) buffer and stained with ethidium bromide. The negative image of the stained gel is shown. Controls included glyceraldehyde 3-phosphate dehydrogenase and the PG-specific B. mori Bmpgdesat1. cDNA levels were not normalized for gene expression.

Molecular identification of B. mori Gα subunits

In the canonical GPCR-mediated signal transduction cascade, all four isoforms of PLCβ (ie PLCβ 1–4) can be activated by GPCR-dissociated guanosine-5′-triphosphate (GTP)-bound Gq α subunits but to differing degrees, with PLCβ1 exhibiting the strongest activation (Smrcka & Sternweis, 1993). To further assess the role of canonical GPCR signalling, we used an approach that previously facilitated our identification of two Gs α subunits (P50Gs1 and P50Gs2) from B. mori PG cDNA (Hull et al., 2007b). Using degenerate primers designed to the conserved regions of Gα subunits and PG cDNA, we amplified fragments of transcripts identical to B. mori antennal Gq and Go α subunits (Miura et al., 2005). Primers designed to amplify across the ORF of both subunits resulted in products with predicted amino acid sequences 99% identical to those published previously; the sequence incongruities observed are likely to reflect the differences in the strains of B. mori used rather than distinct isoforms. Similar differences were observed between the inbred p50 strain and hybrid strain (Kinsyu × Showa) Gα s subunit sequences (Hull et al., 2007b) and are likely to represent single-nucleotide polymorphisms present in the varying strains of Bombyx (Xia et al., 2009). Based on sequence identity, the PG-derived gene products were designated BmGq1 and BmGo1 (accession nos. AB425234 and GU266213). As multiple Gα subunit isoforms are present in insect genomes (Rützler et al., 2006), we also scanned the B. mori genome for regions of similarity with known insect Gα subunits. Using this approach, we identified two additional BmGq and BmGo genes, designated BmGq2 and BmGo2 (accession nos. AB425233 and GU266214, respectively). BmGq2 is 88% identical to BmGq1 and 94% identical to a Gq α subunit from Spodoptera frugiperda (ACJ06653), whereas BmGo2 is 98% identical to BmGo1 and 96% identical to a Go α subunit from Locusta migratoria (P38404) (Raming et al., 1990). Tissue expression analysis indicated that BmGq1 and BmGo1 were expressed to varying degrees in multiple tissues in addition to the PG and that BmGq2 and BmGo2 expression excluded the PG (data not shown). These results are supported by Western blot analysis as single anti-Gq and anti-Go immunoreactive bands corresponding to the predicted molecular weights of BmGq1 and BmGo1 were observed in the cytosolic and membrane bound fractions of PG lysates (Fig. 7). As multiple processes (ie subunit assembly, protein–membrane interactions and post-translational modifications including myristolation and/or palmitoylation) contribute to the association of G α proteins with the cytosolic face of plasma membranes (Vögler et al., 2008), the presence of anti-Gq and anti-Go immunoreactive bands in the membrane fraction is not unexpected. The immunoreactive bands in the cytosol are likely to correspond to G α subunits that have translocated to the cytosol following stimulation (Vögler et al., 2008). Hughes et al. (2001) similarly demonstrated the presence of anti-Gq immunoreactive bands in cytosolic and membrane fractions of mammalian HEK-293 cells.

Figure 7. Western blot of G protein α q and o subunit expression in pheromone gland (PG) lysates. Immunoblot of 10 Bombyx mori PG (Shuko × Ryuhaku) equivalents after 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The membrane was probed with commercially available polyclonal Gq α and Go α antibodies. Immunoreactive bands are indicated by arrows and the cytosolic and membrane bound fractions are indicated by C and M, respectively.

Effects of injected double-stranded RNAs on bombykol production

To assess the in vivo role of the genes identified above, we sought to knockdown the corresponding transcripts by stimulating the endogenous RNA interference (RNAi) mechanism via injected double-stranded RNAs (dsRNA). As our genes of interest are not PG-specific, we modified our previous protocol (Ohnishi et al., 2006) by injecting dsRNAs (10 µg in 2 µL) directly into the PG of newly emerged female moths, rather than pupae, and then determined the levels of bombykol 2 days later. No decrease in bombykol production was observed in females injected with diethylpyrocarbonate (DEPC)-treated H₂O alone or with dsRNA corresponding to enhanced green fluorescent protein (EGFP; Fig. 8). Injection of dsRNA corresponding to a 1022 bp fragment of BmPLCβ4 (nt 22–1043) likewise had no statistically significant effect on bombykol production. In contrast, dsRNA corresponding to a 634 bp fragment of BmPLCβ1 (nt 1–634) reduced bombykol production by ∼60% (Fig. 8). Unexpectedly, we found that dsRNA corresponding to an 887 bp fragment of BmPLCγ (nt 1–887) likewise reduced bombykol production by 60% (Fig. 8). Furthermore, of the four Gα subunits (BmGq1, BmGo1, P50Gs1 and P50Gs2) expressed in the PG, injection of dsRNAs corresponding only to BmGq1 significantly reduced bombykol production (Fig. 8), strongly suggesting that BmGq1 is coupled to PBANR and thus is the G protein α subunit utilized in transmitting the PBAN signal to the downstream PLC effector.

Figure 8. Effects of injected double-stranded RNAs (dsRNAs) on in vivo bombykol production. Pheromone glands (PGs) of newly emerged female moths were injected with 10 µg dsRNAs corresponding to Bombyx mori phospholipase C β4 (BmPLCβ4), BmPLCβ1, BmPLCγ, P50Gs1, P50Gs2, BmGq1, BmGo1, BmIP3R or EGFP. Results are expressed relative to bombykol levels in non-dsRNA injected females following intra-abdominal injections of 5 pmol pheromone biosynthesis activating neuropeptide (PBAN). Control values correspond to approximately 41 ng bombykol/PG. Bars represent mean values ± SEM of replicates from multiple independent experiments. Statistical significance (***; P < 0.001) determined by one-way anova with Tukey's multiple comparison test. Abbreviations: DEPC-H₂O, diethylpyrocarbonate-treated H₂O; EGFP, enhanced green fluorescent protein; IP3R, IP3 receptor; Go, G protein α subunit o; Gq, G protein α subunit q; Gs, G protein α subunit s.

As SOC activation is dependent on the interaction between IP₃ and IP₃R, we sought to assess the effects of BmIP₃R knockdown on bombykol production. Based on homology (74% similarity) with the D. melanogaster IP₃R, we identified a 1 kb portion of a putative IP₃R in the B. mori genome. This sequence was amplified from PG cDNA and used as a template for in vitro transcription with the resulting dsRNAs injected as above. The ∼60% reduction in bombykol (Fig. 8) observed in females injected with dsRNAs corresponding to the BmIP₃R fragment is consistent with PLC involvement in the PBAN signalling cascade and is also strikingly similar to the results obtained following PLC knockdown.

Discussion

Early pharmacological studies in H. armigera hinted at the possibility that the pheromonotropic effects of PBAN were transmitted by a G protein-linked transduction cascade (Rafaeli & Gileadi, 1996). This hypothesis was bolstered by inhibition studies in heliothine species that suggested the involvement of RACCs (Jurenka et al., 1991b; Jurenka, 1996). However, it was not until identification of PBANR as a GPCR (Choi et al., 2003; Hull et al., 2004) that the hypothesis was verified. Since then we have expanded on the molecular mechanisms underlying moth sex pheromone production by characterizing the PBAN-activated Ca²⁺ channels as SOCs (Hull et al., 2007a) and demonstrating that STIM1 and Orai1 are essential components of the pathway (Hull et al., 2009). Although our understanding of the beginnings (ie PBANR activation) and ends (ie influx of Ca²⁺ through SOCs) of the cascade has been enhanced in recent years, our knowledge of the intermediate steps has largely been limited to indirect evidence. To address this deficiency, we sought to examine the molecular role of canonical receptor-mediated PLC signalling in transmitting the pheromonotropic signal.

The reduction in bombykol production following pharmacological inhibition of PLC activity (Fig. 1) is consistent with our previous results demonstrating the involvement of SOCs (Hull et al., 2007a) and supports the presence of a canonical signal transduction cascade in PBAN signalling. The less than robust inhibition (∼40%) observed is likely to be the result of inefficient uptake of the inhibitor by individual PG cells. Despite the incomplete inhibition, the results are comparable for both inhibitors used. The pheromonostatic effect of compound 48/80, however, contrasts with a previous study that reported compound 48/80 was ineffective at inhibiting PBAN pheromonotropic activity (Matsumoto et al., 1995). We speculate that our inclusion of an inhibitor pre-incubation period, which was omitted in the previous study, is the reason for the differing results. Regardless, both the U73122 effects and, more importantly, the significant increase in ³H-IPs levels following PBAN stimulation (Fig. 2), substantiate the validity of the compound 48/80 results and demonstrate that phosphatidylinositol (4,5)-bisphosphate hydrolysis occurs downstream of PBANR activation.

As PBANR has been characterized as a GPCR, our initial expectation was that one of the PLCβ isoforms would be activated in response to PBAN stimulation. In support of this, we amplified a putative PLCβ1 from PG-specific cDNA predicted to encode a protein with significant homology (77% similarity) to the PLCβ1 encoded by the D. melanogaster plc21C gene (Shortridge et al., 1991), which has been shown to genetically interact with D. melanogaster Gq α to affect larval viability and adult flight (Banerjee et al., 2006) as well as normal olfactory transduction (Kain et al., 2008). We also amplified a putative PLCβ4 predicted to encode a protein with 80% sequence similarity to the PLCβ4 encoded by the D. melanogaster norpA gene (Bloomquist et al., 1988). Interestingly, norpA is essential to D. melanogaster phototransduction, which proceeds via SOCs composed of cell surface transient receptor potential (TRP) Ca²⁺ channels (Minke & Parnas, 2006). We also amplified a PG-derived transcript with high homology to the PLCγ encoded by the D. melanogaster small wing gene (Thackeray et al., 1998). Unexpectedly, we found that dsRNA-mediated RNAi knockdown of BmPLCβ4 had no statistically significant effect on bombykol production, whereas knockdown of both BmPLCβ1 and BmPLCγ significantly reduced bombykol production (Fig. 8). These results strongly suggest that BmPLCβ1 is activated downstream of PBAN binding, most likely via GTP-bound BmGq1, knockdown of which also resulted in reduced bombykol production (Fig. 8). We speculate that the partial reduction in bombykol production following direct injection of dsRNAs into the PG may have been the result of inefficient uptake of the dsRNAs by the pheromone producing cells. Alternatively, the lack of complete abrogation of bombykol production could be an indication that the PBAN signal cascade bifurcates at some point or that another protein compensates for the gene targeted for knockdown. Regardless, these results clearly indicate that BmPLCβ1, BmPLCγ, BmGq1 and BmIP₃R function in B. mori PBAN signalling and by extension within the biological framework of bombykol production.

Even though PLCγ is normally associated with receptor tyrosine kinases, recent studies have shown that knockdown/knockout of PLCγ adversely affects both receptor and SOC-mediated Ca²⁺ influx but has little effect on the depletion of ER Ca²⁺ (reviewed in Patterson et al., 2005). The defect is absent in cells over-expressing a phosphatidylinositol (4,5)-bisphosphate hydrolysis-deficient PLCγ mutant but is manifested following deletion of the SH3 domain, suggesting that PLCγ may function as a molecular scaffold via its protein–protein interaction domains (Patterson et al., 2002). Intriguingly, PLCγ has also been shown to interact with various TRP channel isoforms (Patterson et al., 2002; Tong et al., 2004; Tu et al., 2005). We speculate that BmPLCγ likewise functions as a molecular scaffold in B. mori PGs, perhaps stabilizing the protein–protein interactions essential to maintaining the BmSTIM1-BmOrai1 complex. Although no interaction was seen with the mammalian STIM1 (Litjens et al., 2007), it is possible given the differences between the C-terminal ends of mammalian and B. mori STIM1 that BmPLCγ interacts directly with BmSTIM1, or alternatively with BmOrai1.

The data in this study, taken together with pharmacological studies in various other moth species, suggest that the PBAN signalling mechanism leading up to and including Ca²⁺ influx is transmitted via a canonical signal transduction cascade that is likely conserved (Fig. 9): PBAN activation of PBANR results in G protein dissociation and GTP-bound Gq α activation of PLCβ1, which hydrolyses phosphatidylinositol (4,5)-bisphosphate to generate IP₃. Soluble IP₃ acts on ER-membrane bound IP₃Rs to induce the release of ER Ca²⁺ stores. The drop in ER Ca²⁺ concentration subsequently triggers translocation of STIM1 to the cell surface where it interacts with Orai1 to trigger an influx of extracellular Ca²⁺. It is after this last step that the species-specificity of the pheromonotropic control point manifests with cAMP production and activation of acetyl-CoA carboxylase in heliothine species or activation of reductases in M. sexta, T. pityocampa, S. littoralis and B. mori.

Figure 9. Schematic diagram of the proposed pheromone biosynthesis activating neuropeptide (PBAN) signal transduction cascade. 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; IP₃, inositol 1,4,5-trisphosphate; PIP₂, phosphatidylinositol (4,5)-bisphosphate; PLC, phospholipase C; STIM1, stromal interaction molecule 1.

In conclusion, we have shown via pharmacological, biochemical and molecular methods that PBAN signalling in B. mori proceeds via a canonical pathway involving Gq-mediated PLC activation, and that BmGq1, BmPLCβ1, BmIP₃R and BmPLCγ are necessary components. Furthermore, our finding that PLCγ is involved in mediating the pheromonotropic effects of PBAN is consistent with reports from mammalian systems in which PLCγ has been shown to be a required molecular component of GPCR-mediated Ca²⁺ influx and suggests that that this function has been evolutionarily conserved. The precise role of PLCγ in the PBAN signal transduction cascade, however, remains to be established.

Experimental procedures

Insect rearing

Larvae of the inbred p50 strain of B. mori and the B. mori racial hybrid (Shuko × Ryuhaku) were reared on an artificial diet and maintained as described (Fónagy et al., 1992).

In vitro bombykol assay

Adult females (Shuko × Ryuhaku) were decapitated within 3 h of emergence and maintained at 25 °C for 24 h. Abdominal tips were dissected into a modified insect RB [35 mM NaCl, 36 mM KCl, 12 mM CaCl₂, 16 mM MgCl₂, 274 mM glucose and 5 mM Tris-HCl (pH 7.5)] and incubated ± 120 nM synthetic B. mori PBAN for 90 min. Nonpolar constituents were then extracted in 100 µL n-hexane and bombykol production was measured by HPLC using a Senshu-PacNO₂ column (Senshu Scientific Co., Ltd., Tokyo, Japan) as described (Matsumoto et al., 1990). For inhibition studies, abdominal tips were pre-incubated with inhibitor for 30 min prior to PBAN stimulation. U73122 and U73343 were purchased from Calbiochem (San Diego, CA, USA); all other reagents were purchased from Sigma (St Louis, MO, USA).

Measurement of inositol phosphates

Total ³H-IP levels were determined using B. mori (Shuko × Ryuhaku) PGs dissected and mechanically trimmed as described previously (Ozawa & Matsumoto, 1996). Trimmed PGs (30) were rinsed in RB and then incubated en masse in 1500 µL RB containing myo-[2-³H]inositol [specific activity; 16 curie (Ci)/mmol] (Amersham Biosciences, Piscataway, NJ, USA) at a final concentration of 3 µM for 2 h. PGs were then washed three times with 1 mL RB and pre-incubated for 15 min in RB + 10 mM LiCl prior to stimulation with 120 nM PBAN. ³H-IPs were isolated and radioactivity measured as described (Yang et al., 2002) with minor modifications. Briefly, stimulated PGs were homogenized in 20 mM formic acid followed by neutralization with 150 mM NaOH. Cellular debris was pelleted by centrifugation and the supernatant applied to analytical grade (AG) 1 × 8 anion exchange resin (Bio-Rad Laboratories, Richmond, VA, USA) slurry in 1.7 mL tubes. Tubes were centrifuged at high speed for 30 s and the resin washed with 1 mL deionized H₂O (dH₂O). The tubes were then re-centrifuged and the resin washed with 60 mM ammonium formate/5 mM sodium borate. Total inositol phosphates were eluted using 500 µL 1 M ammonium formate/0.1 M formic acid and radioactivity determined using 200 µL aliquots in UltimaGold scintillation fluid (Perkin Elmer, Boston, MA, USA).

Reverse transcriptase-PCR

Bombyx mori (p50) PGs were dissected and trimmed from adult moths as above. Total RNA was isolated with Isogen (Nippongene Corp., Tokyo, Japan) according to the manufacturer's instructions and treated with DNase I (Invitrogen, Carlsbad, CA, USA) prior to cDNA synthesis using SuperScript III (Invitrogen) and random hexamers. Genes of interest were amplified with oligonucleotide primer pairs (see Table 1) corresponding to B. mori genomic sequences using rTaq (Takara Bio., Otsu, Japan). Thermacylcer conditions consisted of: 95 °C for 2 min, 33 cycles of 94 °C for 30 s, 55 °C for 30 s, 72 °C for 60 s, with a final extension at 72 °C for 5 min. PCR products were separated on 2% agarose gels in Tris-acetate-EDTA (TAE) buffer and stained with ethidium bromide. Products of expected sizes were sub-cloned into pGEM-T Easy (Promega, Madison, WI, USA) and sequenced. A p50 PG specific library (Yoshiga et al., 2000) was used for 5′ and 3′ RACE with vector specific primers (T7 and T3) and gene specific primers. RACE PCR was performed using LA Taq (Takara Bio.) with thermacycler conditions consisting of 95 °C for 2 min, 33 cycles of 94 °C for 30 s, 55 °C for 30 s and 72 °C for 2 min, with a final extension at 72 °C for 10 min. Specific amplimers were sub-cloned and sequenced.

Table 1.  Oligonucleotide primers used in this study

Primer	Orientation	Sequence*
* Sequence corresponding to the T7 polymerase promoter site is underlined.
PLCx 481400 F	Sense	CAGGTGCTATTAACGG
PLCx 481400 R	Antisense	GTTTCAACAACTCGACGG
PLCx 457194 F	Sense	GACGGTCTGATCCGCTAC
PLCx 457194 R	Antisense	CAACCAGCCAACAAGACC
PLCx 159970 F	Sense	TATCTGACAGGTGACCAG
PLCx 159970 R	Antisense	TCTTGAAATGCGCTCGCC
PLCx 600350 F	Sense	AGATGCGTAGAGCTAGAC
PLCx 600350 R	Antisense	CAGCGGTTCCTTAAGCAG
PLCx 503327F	Sense	AGATGCGTGGAGTTGGAC
PLCx 503327 R	Antisense	GTCCAGAGGCTTATCGAG
BmPLCb4 ORF F	Sense	TACAACACCATGACCAAG
BmPLCb4 ORF R	Antisense	TGCAAAAAATTCCGTTTTAC
BmPLCg ORF F	Sense	ATGAAGTCGGTCTGCTTC
BmPLCg ORF R	Antisense	TCAACCGTCTGTTATAATTGAC
BmPLCb1-F	Sense	GATGAGTGCTAGCACGAGATCAGCCA
BmPLCb1-R	Antisense	CGGTCTAAGACGTTCCTGTCCCTTG
P50 GAPDH F	Sense	GCGTACGTGGTGGTGGTCAG
P50 GAPDH R	Antisense	TGGTGCAAGAAGCATTTGAG
BmGq1 F	Sense	ATGGAGTGCTGCATGTCA
BmGq1 R	Antisense	AACGAGATTGTACTCTTT
BmGq2 R	Antisense	AGCCAAATTAAATTCTTTGAG
BmGo1 F	Sense	AGGGAGAACTGCCGATATG
BmGo1 R	Antisense	GTACAGGCCGCAGCCGCG
BmGo2 F	Sense	GTGCGCGCTATAGCCATG
Desat1 F	Sense	AAAGAATTCAAAAAATGCCTCCTAATTCAG
Desat1 R	Antisense	AAAGGTACCTCAATTCATTCCATTTGGTCG
EGFP RNAi F	Sense	TAATACGACTCACTATAGGGCGAATGGTGAGCAAGGGCG
EGFP RNAi R	Antisense	TAATACGACTCACTATAGGGCGACTTGTACAGCTCGTCC
PLCb4 RNAi F	Sense	CCGGATCCTAATACGACTCACTATAGGGCGAACTGGCAGATACCAGTC
PLCb4 RNAi R	Antisense	CCGGATCCTAATACGACTCACTATAGGGCGTGTCGGTACATCTCCACG
PLCg RNAi F	Sense	TAATACGACTCACTATAGGGCGATGAAGTCGGTCTGCTTC
PLCg RNAi R	Antisense	TAATACGACTCACTATAGGGCGTGCTTGGAGTACAGCATC
PLCb1 RNAi F	Sense	TAATACGACTCACTATAGGGCGATGAGTGCTAGCACGAGATC
PLCb1 RNAi R	Antisense	TAATACGACTCACTATAGGGCGCGAACTGGAACTTGGAC
IP3R RNAI F	Sense	TAATACGACTCACTATAGGGCGTCCTACAGCGCATGATC
IP3R RNAi R	Antisense	TAATACGACTCACTATAGGGCGACGTACGCTTCTTTCACC
Gs1 RNAi F	Sense	CCGGATCCTAATACGACTCACTATAGGGCATGGGGTGCGCCCAGTCG
Gs1 RNAi R	Antisense	CCGGATCCTAATACGACTCACTATAGGGCGCGCAGGTTGTTGGCGATG
Gs2 RNAi F	Sense	CCGGATCCTAATACGACTCACTATAGGGCGATGGGCTGTTTCAAGTCG
Gs2 RNAi R	Antisense	CCGGATCCTAATACGACTCACTATAGGGCGTCGTATTGCTTCAAATGC
Go RNAi F	Sense	CCGGATCCTAATACGACTCACTATAGGGCATGGGTTGCTTCGGGTC
Go RNAi R	Antisense	CCGGATCCTAATACGACTCACTATAGGGCAGCAGCTCGTACTGGCGC
Gq RNAi F	Sense	CCGGATCCTAATACGACTCACTATAGGGCGATGGAGTGCTGCATGTCAG
Gq RNAi R	Antisense	CCGGATCCTAATACGACTCACTATAGGGCGTACTCTTTCAGGTTGGAC

The inbred p50 strain of B. mori was used for tissue expression analyses with various tissues dissected from adult females. Total RNA and first-strand cDNAs were prepared as above. PCR was performed using Ex Taq (Takara Bio.) with gene specific primers (see Table 1) and thermacycler conditions consisting of 95 °C for 2 min, 33 cycles at 94 °C for 30 s, 55 °C for 30 s and 72 °C for 3 min, followed by a final extension of 72 °C for 10 min. PCR products were separated on 1.3% agarose gels in TAE buffer and stained with ethidium bromide. To confirm amplification of the desired transcripts, products of expected sizes were sub-cloned and sequenced as before. Amplification of glyceraldehyde 3-phosphate dehydrogenase was used a positive control.

Degenerate PCR cloning

Degenerate PCR was performed as described previously (Hull et al., 2007b). Briefly, total RNA was isolated from 10 B. mori (p50) PGs and first strand cDNA synthesized using SuperScript III (Invitrogen) and random hexamers. Fragments of Gα transcripts were amplified with LA Taq (Takara Bio.) using multiple combinations of degenerate oligonucleotide primers designed to conserved regions of Gα (Hull et al., 2007b). PCR products were sub-cloned and sequenced.

Protein immunoblotting

PGs from 10 adult female B. mori (Shuko × Ryuhaku) were mechanically trimmed and homogenized in lysis buffer (25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 50 mM NaCl, 50 mM NaF, 5 mM ethylenediamine tetraacetic acid, 1 mM p-amidiniophenyl methylsulphonyl fluoride and 10 µg/mL each of aprotinin and leupeptin). Cell lysates were centrifuged at 15 000 g for 5 min at 4 °C with the resulting supernatant incubated with 4× Laemmli buffer in boiling water for 5 min; the pellet was re-suspended in lysis buffer containing 1% TritonX-100 and likewise incubated with 4× Laemmli in boiling water. Proteins were resolved using 10% polyacrylamide gel electrophoresis and transferred to an Immobilon-P membrane (Millipore, Bedford, MA, USA). Membranes were blocked for 60 min in Tris-buffered saline containing 5% powdered milk, washed and then probed for 60 min using polyclonal antibodies to Go or Gq α subunits (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) diluted 1:1000 in Can Get Signal 1 (Toyobo, Osaka, Japan). After washing, the membranes were then probed with a 1:10 000 dilution of goat anti-rabbit IgG-horse radish peroxidase (Zymed Laboratories Inc., San Francisco, CA, USA) in Can Get Signal 2 (Toyobo). Chemiluminescent detection was carried out using an ECL Western blotting system (Amersham, Piscataway, NJ, USA) and a LAS-3000 detector (Fuji Film, Tokyo, Japan).

Synthesis and injection of dsRNA

The DNA templates for in vitro transcription were prepared from plasmid DNAs using gene specific primers containing T7 polymerase promoter sites on the 5′ ends of the sense and antisense primers (see Table 1). PCR was performed using KOD-Plus- (Toyobo) with the resulting products purified with a SV PCR Purification kit (Promega) and used as templates for in vitro transcription using an AmpliScribe T7 High Yield Transcription kit (Epicentre, Madison, WI, USA) according to the manufacturer's instructions. Transcription products were purified using Isogen as described above and re-suspended in DEPC-H₂O. Products were analysed by gel electrophoresis to confirm annealing and the RNA concentrations determined spectrophotometrically. Samples were diluted (5 µg/µL) and 10 µg dsRNAs directly injected in the PG of newly eclosed adult females using a 10 µL microsyringe (Hamilton, Reno, NV, USA). Control pupae were injected with a similar volume of DEPC-H₂O alone. After injection, females were maintained under normal conditions for 48 h and then analysed for bombykol production via in vitro assays as described above.

Statistical analyses

Quantitative data are expressed as mean ± SEM. Data were analysed in GraphPad Prism v. 4.0 (Graphpad Software Inc., La Jolla, CA, USA) and statistical significance defined as P < 0.001 via one-way anova analysis with Tukey's multiple comparison test.

Acknowledgements

We are grateful to Dr Kazuei Mita (National Institute of Agrobiological Sciences, Tsukuba) for providing us with the fefu-PO9-017 cDNA clone. We wish to thank Shinji Atsusawa and Masaaki Kurihara for rearing the B. mori colony and Ryosuke Kajigaya for his work on BmIP₃R. This work was supported by the Lipid Dynamics Research Program from RIKEN and the Targeted Proteins Research Program (TPRP) and Grants-in-aid for Scientific Research (B) 20380040 from the Ministry of Education, Culture, Sports, Science and Technology of Japan.