Bombyx mori Homologs of STIM1 and Orai1 Are Essential Components of the Signal Transduction Cascade That Regulates Sex Pheromone Production*

Sex pheromone production in the pheromone gland (PG) of the silkmoth, Bombyx mori, is mediated by store-operated channels (SOCs) acting downstream of pheromone biosynthesis activating neuropeptide (PBAN) binding. Although recent studies have implicated STIM1 and Orai1 as essential components of SOCs, little is known about the molecular nature of the SOCs involved in sex pheromone production. In this study we cloned silkmoth homologs of STIM1 and Orai1 and sought to determine whether they comprise the PG SOC pathway. BmSTIM1 is expressed in multiple tissues and, in the PG, is encoded by two transcripts of differing size. BmOrai1A and BmOrai1B, which are identical except for a 37-residue N-terminal truncation in BmOrai1B, arise from alternative splicing of the bmorai1 locus and are expressed as independent transcripts in various tissues. In the PG, only BmOrai1B is actively transcribed. Fluorescent chimeras demonstrated that BmSTIM1 expression is restricted to the endoplasmic reticulum, whereas both BmOrai1A and BmOrai1B localize to the cell surface. In Ca2+-free medium, thapsigargin-mediated depletion of endoplasmic reticulum Ca2+ stores resulted in redistribution of BmSTIM1 to the plasma membrane, but only when the BmOrai1 homologs were also overexpressed. Translocation was dependent on the BmSTIM1 C terminus “CRAC activation domain.” Ala mutation of Lys380, Lys383, Lys384, Arg382, and Arg385 suggests that translocation involves electrostatic interactions. Translocation was also seen following PBAN stimulation in cells co-expressing BmSTIM1, BmOrai1B, and the PBAN receptor. In vivo RNA interference-mediated knockdown of BmSTIM1 and BmOrai1 significantly reduced sex pheromone production without affecting cell viability.

The initiation of sexual activity for most species of moths is dependent on the receptiveness of the female, a state that is signaled to conspecific males through the environment via low molecular weight volatile compounds released by the female. In general, pheromones are de novo synthesized in a specialized tissue located within the terminal female abdominal segment known as the pheromone gland (PG). 3 Pheromones are predominantly hydrocarbon chains 10 -18 carbons in length with various sites of unsaturation and a reductively modified carbonyl carbon consisting of a primary alcohol, aldehyde, or acetate ester (1). In most species, sex pheromone production is mediated by pheromone biosynthesis activating neuropeptide (PBAN), a 33-34-neuropeptide present in insects from a variety of orders that is characterized by a C-terminal amidated pentapeptide FXPRL motif that is essential for biological activity (2,3). The molecular mechanisms underlying how the extracellular signal of PBAN is transmitted into the biological response of sex pheromone production has been the focus of numerous studies. It has been well established that regardless of species, pheromonotropic activity is dependent on extracellular Ca 2ϩ (2), suggesting that PBAN signaling is associated with the influx of extracellular Ca 2ϩ ; we directly demonstrated this crucial event in Bombyx mori using fluorescent Ca 2ϩ imaging techniques (4). Early pharmacological studies in heliothine species suggested that PBAN triggers the opening of receptor-operated Ca 2ϩ channels (ROCs) (5,6). The involvement of ROCs was strengthened when the PBAN receptor (PBANR) was identified as a G protein-coupled receptor that could function upstream of Ca 2ϩ influx when heterologously expressed in cultured insect cells (7,8). We recently expanded on the ROC studies and reported that in B. mori the PBAN pathway proceeds via store-operated channels (SOCs): sex pheromone production is blocked by classical SOC inhibitors, activators of diacyl glycerol-dependent channels are ineffective, and the pheromonotropic effects of PBAN can be mimicked by thapsigargin (TG) (4). The pheromonotropic effects of TG observed in other moth species suggest that the involvement of SOCs is likely conserved (9). The cellular mechanisms underlying PBAN-mediated activation of the SOC pathway and the molec-ular nature of the SOC itself, however, have yet to be determined.
The influx of extracellular Ca 2ϩ through SOCs occurs in response to depletion of endoplasmic reticulum (ER) Ca 2ϩ stores, which is triggered by an elevation in inositol 1,4,5triphosphate levels following receptor-linked phospholipase C activation (10,11). Recent discoveries have identified two proteins essential to the SOC pathway: stromal interaction molecule 1 (STIM1) and Orai1 (12,13). STIM1 is a single transmembrane (TM) domain protein that resides in the ER membrane. A change in the conformational state of the STIM1 intramolecular EF-hand motif (i.e. the Ca 2ϩ -sensing region) in response to diminished luminal Ca 2ϩ levels promotes the translocation of STIM1 from/within the ER to regions of the plasma membrane where it interacts with Orai dimers to activate store-operated Ca 2ϩ entry (SOCE) (14). Orai (or Ca 2ϩ release-activated Ca 2ϩ channel modulator 1, CRACM1) is a 4-TM domain protein localized at the plasma membrane that is predominantly a dimer under resting conditions. Interaction with the C terminus of STIM1 induces tetramerization of the Orai dimers to form the Ca 2ϩ -selective pore (15).
To further expand our understanding of the molecular mechanisms underlying sex pheromone production, we sought to determine whether B. mori homologs of STIM1 (BmSTIM1) and Orai1 (BmOrai1) are involved in the PBAN-mediated Ca 2ϩ influx. Herein we report on the cloning of the two homologs as well as their genomic organization, expression profile, functional expression in cultured insect cells, identification of functionally essential domains/residues, and knockdown effects on sex pheromone production. Taken together, we conclude that BmSTIM1 and BmOrai1 are integral components of the PBAN signaling cascade.

EXPERIMENTAL PROCEDURES
Insects-Larvae of the inbred p50 strain of B. mori and the B. mori racial hybrid (Shuko x Ryuhaku) were reared on an artificial diet and maintained under a 16L:8D photoperiod at 25°C (16).
Cloning B. mori Homologs of STIM and Orai-PGs from adult female B. mori (p50 strain) were dissected into a modified insect Ringer's buffer (RB) consisting of 35 mM NaCl, 36 mM KCl, 12 mM CaCl 2 , 16 mM MgCl 2 , 274 mM glucose, and 5 mM Tris-HCl (pH 7.5) and mechanically trimmed as described (17). Total RNA was isolated with Isogen (Nippongene Corp., Tokyo, Japan) according to the manufacturer's instructions and treated with DNase I (Invitrogen) prior to cDNA synthesis using SuperScript III (Invitrogen) and random hexamers. From this cDNA, a BmSTIM1 fragment was amplified using thermalcycler conditions consisting of: 95°C for 3 min, 33 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min, followed by a final extension of 72°C for 10 min with LA Taq (Takara Bio, Inc., Otsu, Japan) and the following oligonucleotide primers: sense 5Ј-TTCCTCCGCGAAGAGTTG (FLREEL), antisense 5Ј-TGTA-GCTCTTGAGTGACC (VTQEL). The 5Ј and 3Ј ends were obtained using vector-specific primers with a B. mori p50 PG cDNA library prepared previously (18). The complete BmSTIM1 ORF was amplified from PG cDNA using the sense primer, 5Ј-ATGCGTATCGGTTTCATTTTG (MRIGFIL), with one of the following antisense primers, 5Ј-TCACTTGTC-TCCTCCCAG (LGGDKstop) or 5Ј-CTTGTCTCCTCCCAGC (LGGDK). PG cDNA was used to amplify a BmOrai1 fragment with LA Taq as described above and oligonucleotide primers: sense 5Ј-ATGTCGGTTTGGTCAGCC (MSVWSA), and antisense 5Ј-TCTTTGATACCTGTCAC (VTGIK). As before, the 5Ј and 3Ј ends were obtained using vector-specific primers with a B. mori p50 PG cDNA library. The complete ORF for BmOrai1B was amplified from PG cDNA using the sense primer, 5Ј-ATGTCGGGTGAGACGCCG (MSGETP), and the antisense primer, 5Ј-CAGACGACCCGACTTTGG (QSRVV). To amplify the BmOrai1A ORF, the initial sense primer was used with the antisense primer shown above, and cDNA was prepared from various adult p50 tissues. In all cases, PCR products were separated on 1.3% agarose gels in TAE buffer and stained with ethidium bromide; products of the expected sizes were subcloned into pGEM-T Easy (Promega) and sequenced.
Northern Blot Analysis-PGs of newly emerged adult female moths (p50 strain) were dissected and trimmed as above. Total RNA was isolated from the trimmed PGs using a Micro-Fast Track kit (Invitrogen). After heat denaturation, 10 g of total RNA was electrophoresed on a 1.2% agarose gel and then transferred to a Hybond-XL nylon membrane (Amersham Biosciences). Prehybridization was performed for 1 h at 68°C in Perfect Hybridization Plus buffer (Sigma). The [␣-32 P]dCTPlabeled probes corresponding to nucleotides 510 -1452 of BmSTIM1 and the BmOrai1B ORF were prepared with a Random Primer DNA Labeling Kit version 2 (Takara Bio, Inc.) using PCR products amplified from plasmid DNAs and purified using Probe Quant G50 Microcolumns (Amersham Biosciences). Probe hybridization was performed at 68°C for 18 h at which point the blot was washed in an initial solution of 0.1% SDS, 1ϫ SSC for 8 min at 68°C and then transferred to 0.1% SDS, 0.2ϫ SSC for two 5-min washes at 68°C. The blot was exposed overnight at Ϫ80°C to BAS-2500 film (Fuji Film).
Reverse Transcriptase-PCR Expression Analysis-PGs and other tissues (head, flight muscle, egg, fat body, and Malpighian tubule) were dissected from adult female B. mori (p50 strain) into RB, and total RNA was isolated with Isogen and treated with DNase I prior to cDNA synthesis as before. BmSTIM1 was amplified using oligonucleotide primers to the ORF described above, whereas both BmOrai1A and -B transcripts were amplified with the "MSVWSA" sense primer and the ORF antisense primer described above. For control purposes, the low expression gene, glyceraldehyde-3-phosphate dehydrogenase (gapdh), was amplified using the sense primer, 5Ј-ATGTCAA-AAATTGGAATC, and the antisense primer, 5Ј-AGTCATCA-AGCCCTCAAC. PCR was performed using Ex Taq (Takara Bio, Inc.) and thermalcycler conditions consisting of: 95°C for 3 min, 33 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min, followed by a final extension of 72°C for 10 min. PCR products were separated on either 1.3 or 2% agarose gels in TAE buffer and stained with ethidium bromide. To confirm amplification of the desired transcripts, products of expected sizes were subcloned and sequenced as before.
Construction and Transfection of Expression Plasmids-Expression plasmids containing the respective BmSTIM1 and BmOrai1A or BmOrai1B ORFs C-or N-terminal fused to a fluorescent reporter protein were generated using the pIB/V5-His-TOPO TA Expression kit (Invitrogen). Chimeric genes were constructed via overlap extension PCR using plasmid DNAs with KOD-Plus-(Toyobo, Osaka, Japan). The gene-specific ORF sense primers described above for BmSTIM1 and BmOrai1A/B were used with chimeric antisense primers (Orai-CFP/Venus, 5Ј-GCCCTTGCTCACCATGACGACCCGAC-TTTGG; STIM-CFP/Venus, 5Ј-GCCCTTGCTCACCAT-TAAGATATTATTATATAAT). Chimeric sense primers (BmOrai1-CFP/Venus, 5Ј-CCAAAGTCGGGTCGTCATGG-TGAGCAAGGGC; BmSTIM1-CFP/Venus, 5Ј-ATTATATA-ATAATATCTTAATGGTGAGAAGGGC) were used with gene-specific antisense primers for CFP (5Ј-CAGCTCGTCC-ATGCCGAG) and Venus (5Ј-GGCGGCGGTCACGCGTTC). Truncations, deletions, and point mutations were likewise generated via overlap extension PCR using the following chimeric sense primers: BmSTIM1- In addition, the gene-specific antisense primer (5Ј-CTACGTGACTTCGTACTTGT) was used to generate CFP-BmOrai1B-(1-186). Plasmid DNAs for CFP and Venus were kindly provided by Dr. Akihiko Nakano (Molecular Membrane Biology Laboratory, RIKEN, Wako, Japan). The resulting PCR products were used as templates to generate the full-length chimeric genes using the respective gene-specific sense and antisense primers. The chimeric genes were cloned into pIB/V5-His-TOPO and sequenced to confirm the presence and orientation of the inserts. Transfections were performed using cultured insect cells (BmN and Sf9) attached to 27-mm glass-bottom dishes (Matsunami, Tokyo, Japan) with 2 g of plasmid DNA and 8 l of Cellfectin (Invitrogen) transfection reagent according to the manufacturer's recommended protocol. Cells were maintained overnight at 28°C. Transfection media was replaced with fresh IPL41 or TC100 insect media (Invitrogen) containing streptomycin/kanamycin (50 g/ml each) and the cells were returned to 28°C for an additional 24 h.
Fluorescent Confocal Microscopy-To examine the cellular localizations of the fluorescent BmSTIM1 and BmOrai1 reporter constructs, confocal microscopy of both live and fixed cells was performed using a Leica TCS NT scanning laser confocal system (Bachem, Budendorf, Switzerland). For live cell imaging, cells were washed with fresh media and CFP fluorescence was imaged at 442 nm with a helium-cadmium laser and Venus fluorescence imaged at 514 nm with an argon laser. To minimize cross-talk in co-localization experiments, channel recordings were made using individual laser lines. For fixed cell imaging, cells were fixed for 30 min with 4% formalin at 4°C, washed three times with PBS, and then imaged as above.
For co-localization experiments involving BmSTIM1-Venus and the ER membrane, a monoclonal rabbit anti-KDEL antibody (Stressgen) was used. Cells transiently expressing BmSTIM1-Venus were fixed for 30 min with 4% formalin at 4°C, washed twice with PBS, and then blocked for 1 h using PBS, 10% fetal bovine serum, 0.1% Triton X-100. Cells were then incubated for 1 h at room temperature with a 1:100 dilution of monoclonal rabbit anti-KDEL antibody in PBS, 10% fetal bovine serum. Cells were washed 4 times with PBS and then incubated for 1 h at room temperature with a 1:200 dilution of anti-rabbit IgG conjugated to rhodamine red-X. Cells were washed 4 times with PBS and imaged as above at 514 nm with an argon laser for Venus fluorescence and at 568 nm with an argon-krypton laser for rhodamine red-X fluorescence.
For co-localization experiments involving BmSTIM1-CFP and Ca 2ϩ hot spots, the fluorescent Ca 2ϩ indicator, Fluo4-AM (Molecular Probes) was used. Briefly, BmN cells transfected with the BmSTIM1-CFP expression plasmid were incubated for 30 min at 28°C in the dark with RB containing 6 M Fluo4-AM, 0.6% pluronic F127 (Molecular Probes). Cells were then washed with RB and imaged as above at 442 nm with a heliumcadmium laser for CFP fluorescence and at 488 nm with an argon laser for Fluo4-AM fluorescence.
The effects of TG-induced ER Ca 2ϩ depletion on the localization of BmSTIM1-Venus constructs with either BmOrai1A-CFP or BmOrai1B-CFP were examined in cells transiently expressing both constructs. Normal growth media (TC100) was replaced with RB, 3 mM EGTA, 5 M TG, 25 M ionomycin, and the cells were left at room temperature for 15 min. For negative controls, TC100 was replaced with RB alone. At the end of the 15-min period, the cells were fixed as before and imaged accordingly.
Co-localization experiments involving rhodamine red-PBAN (RR-PBAN) were performed using Sf9 cells transiently expressing BmOrai1B-CFP, BmSTIM1-Venus, and PBANR-EGFP (8). Internalization of RR-PBAN and subsequent imaging was carried out as described previously (8). To minimize channel cross-talk in the multicolor labeling experiments, sequential multiple channel fluorescence scanning was done with emission filter settings adjusted for each fluorescent protein. CFP excitation was carried out at 442 nm (CFP max ϭ 433 nm) with a helium-cadmium laser, EGFP excitation at 488 nm (EGFP max ϭ 488 nm) with an argon laser, Venus excitation at 514 nm (Venus max ϭ 515 nm) with an argon laser, and rhodamine red excitation at 568 nm (rhodamine red max ϭ 560 nm) with an argon-krypton laser. All images were processed and merged using Photoshop 7.0 (Adobe Systems Inc., San Jose, CA).
Synthesis and Injection of dsRNAs-The templates for the synthesis of dsRNAs corresponding to BmSTIM1 and BmOrai1B were prepared using KOD-Plus-with gene-specific primers containing T7 polymerase sites. Portions of each gene were amplified from plasmid DNAs using the BmSTIM1 sense primer (5Ј-TAATACGACTCACTATAGGGCGTTCCTCC-GCGAAGAGTTG) with the antisense primer (5Ј-TAATACG-ACTCACTATAGGGCGTGTAGCTCTTGAGTGACC) and the BmOrai1B sense primer (5Ј-TAATACGACTCACTATA-GGGCGATGTCGGGTGAGACGCCG) with the antisense primer (5Ј-TAATACGACTCACTATAGGGCGGACGACC-CGACTTTGGTC). PCR products were purified using a SV PCR purification kit (Promega) and used as templates for in vitro transcription using the AmpliScribe T7 High Yield Transcription kit (Epicenter) according to the manufacturer's recommended protocol. After synthesis, dsRNAs were diluted with diethyl pyrocarbonate-treated H 2 O, the RNA concentrations were measured (A 260 ), and the products analyzed by gel electrophoresis to confirm annealing. Samples were diluted to the desired concentration and 2 l (5 g/l) injected directly into the PG of newly emerged adult females (Shuko x Ryuhaku) using a 10-l microsyringe. Control females were injected with dsRNA corresponding to the EGFP ORF (Clontech). After injection, adults were maintained overnight under normal conditions.
In Vivo Bombykol Assay-Females were decapitated 24 h after injection and maintained at 25°C for another 24 h. The moths were then injected with either 5 pmol of B. mori PBAN in PBS or PBS alone. Abdominal tips were dissected 90 min after injection and bombykol production was measured by high pressure liquid chromatography using a Senshu-Pac NO2 column (Senshu Scientific Co., Tokyo, Japan) as described (19).
Metabolic Enzyme Assays-To assess the effect of BmSTIM1 and BmOrai1 dsRNAs on normal metabolic processes, PGs Ϯ dsRNA injections were assayed for GAPDH and catalase activities using the KDalert GAPDH Assay kit (Ambion) and Amplex Red Catalase Assay kit (Molecular Probes), respectively. PGs were dissected 48 h after injections and homogenized in the lysis buffer provided in the KDalert GAPDH Assay kit. GAPDH and catalase activities were determined according to the respective manufacturer's recommended protocols. Single PG equivalents were used per reaction with five replicates of each reaction.
Statistical Analyses-Quantitative data are expressed as mean Ϯ S.E. Data were analyzed in GraphPad Prism 3.0 (GraphPad Software Inc.) and statistical significance determined using analysis of variance and Tukey's multiple comparison test.

RESULTS
Molecular Identification of BmSTIM1-Sex pheromone production in B. mori is dependent on PBAN-mediated opening of SOCs (4). To determine whether STIM1 and Orai1 are involved in the PBAN signal transduction cascade, we searched the publicly available B. mori genome data base (KAIKOBLAST) for regions predicted to be homologous with the Drosophila melanogaster proteins and constructed primers accordingly. Using this approach, we amplified a 966-bp fragment from PG cDNAs with 69% identity to D. melanogaster STIM (DmSTIM). A B. mori p50 PG cDNA library prepared previously (18) was used to obtain the 5Ј and 3Ј ends of the putative ORF. Using PG-derived cDNA, additional primers were used to amplify a single 1734nucleotide ORF (GenBank accession AB425069) that was predicted to yield a 577-amino acid protein with 64% sequence identity to DmSTIM but only marginal (39 -44%) identity to human STIM1 and STIM2. However, BmSTIM1 has a number of conserved domain characteristics of STIM proteins including an N-terminal signal peptide (residues 1-19), an EF-hand Ca 2ϩ binding domain (residues 86 -98), a sterile ␣-motif domain (residues 144 -202), a Cys pair spaced 7-8 residues apart N-terminal to the EH-hand domain (Cys 62 and Cys 68 ), and a single TM domain (residues 222-242). In addition, the BmSTIM1 C terminus contains a highly conserved region (residues 340 -446) shown by multiple groups to be sufficient for Orai1 binding and activation, referred to as the "CRAC activation domain" (CAD) (20), the "STIM Orai activating region" (21), the "small Orai1-activating fragment" (22), or "coiled-coil domain containing region" b9 (23). BmSTIM1 also contains a hidden EF-hand motif (residues 120 -131) believed to be critical for association of the canonical EH-hand with the sterile ␣ motif domain (i.e. the oligomerization event that signals decreased luminal Ca 2ϩ ) (14). BmSTIM1, however, lacks the C-terminal Ser/Pro-rich and polybasic Lys-rich regions observed in mammalian STIM proteins (Fig. 1).
In a Northern blot of PG total RNA probed with a portion of BmSTIM1, we detected the presence of two STIM1-hybridizing transcripts of ϳ4.0 and 4.5 kb ( Fig. 2A). Repeated attempts to identify a second ORF failed, suggesting that the two transcripts likely encode identical ORFs but have different UTRs. Williams et al. (24) likewise reported the presence of two STIM1 transcripts (4.1 and 4.5 kb) in Northern blots of human tissues but only a single 90-kDa protein in Western blots.
To assess the comparative genomic organization of the bms-tim1 ORF with those of dmstim and human STIM1, the intronexon boundaries of bmstim1 were determined by aligning the ORF cDNA and predicted amino acid sequences with genomic sequences in the B. mori genome (25,26). Surprisingly, the genomic organization of the bmstim1 ORF more closely resembles human STIM1 than dmstim (Fig. 2B), with the bmstim1 locus encompassing 13 coding exons compared with the 7 of dmstim. The exon structure of the intraluminal portion of STIM appears to be fairly well conserved across 6 exons among the three species. In contrast, the C-terminal coding region of bmstim1 and human STIM1 encompasses 7 and 6 exons, respectively, whereas that of dmstim is limited to 2 exons.
Transcripts encoding proteins essential to silkmoth sex pheromone production have been shown to be PG specific and undergo up-regulation in the day preceding adult emergence from the pupal stage (8,27,28). To examine the expression profile of BmSTIM1, reverse transcriptase-PCR analyses were performed using primer pairs designed to amplify the entire ORF. We found that BmSTIM1 is expressed at varying levels in a number of adult female tissues (Fig. 4). Similar analyses using PG cDNA prepared from various stages of PG development indicated that BmSTIM1 expression is constant and does not undergo up-regulation (data not shown). NOVEMBER 6, 2009 • VOLUME 284 • NUMBER 45

Silkmoth Homologs of STIM1 and Orai1
Molecular Identification of BmOrai1 Homologs-Using genomic-based primers designed to the 351-amino acid D. melanogaster Orai (DmOrai) in conjunction with B. mori p50 PG cDNA, we amplified a transcript containing a putative 667-bp ORF. The product of the ORF was predicted to yield a 221-amino acid protein with 76% sequence identity to DmOrai but that had a truncated N terminus and, based on the B. mori genomic data, an unexpected putative translation initiation site. Interestingly, translation from a methionine located in the 5Ј UTR would generate a 34-amino acid peptide with an N-terminal sequence (MSVW) identical to DmOrai and that would reconcile the differences observed previously. Repeated rapid amplification of cDNA ends using a PG-derived cDNA library failed to identify transcripts encoding a larger ORF with the MSVW N terminus, suggesting that the 221-amino acid transcript was the only Orai1-like transcript present in the silkmoth PG. This is supported by Northern blot analysis of PG total RNA that showed a single Orai1-hybridizing transcript (Fig.  3A). A search of the publicly available B. mori EST data base (SILKBASE) (29) with the PG-derived product indicated transcripts in other tissues with identical 5Ј UTR and coding sequences. We also identified two clones (wds2053, wing disc spinning stage day Ϫ2; and fepMP07_F_E17, epidermis 4th molting stage) that differed from the PG-derived product via an N-terminal extension with an MSVW N terminus. The genomic organization of the B. mori orai1 locus (Fig. 3B) suggests that alternative splicing could yield two Orai1 transcripts: one with a start methionine localized in exon 1 that splices out exon 2 to generate an ORF encoding a 258-amino acid protein (BmOrai1A; GenBank accession AB425232). The other transcript utilizes an alternative splicing mechanism that includes exons 1 and 2 in the 5Ј UTR and has an initiation methionine localized in exon 3 with an ORF encoding a 221-amino acid protein with a N terminus that is truncated 37 amino acids (BmOrai1B; GenBank accession AB425231). Because the initi-ation codon of BmOrai1A resides in the 5Ј UTR of the BmOrai1B transcript, it is possible to amplify both products using the same set of primers such that BmOrai1A would generate a 777-bp amplimer and BmOrai1B would generate an 861-bp amplimer. Using this primer set for reverse transcriptase-PCR, we found that BmOrai1A and -B are expressed at varying levels in a number of adult tissues (Fig. 4). However, BmOrai1B was the only transcript expressed in the PG, whereas BmOrai1A was the only transcript amplified from Malpighian tubules. Similar to BmSTIM1, BmOrai1B expression in the PG does not undergo up-regulation in the days preceding adult emergence (data not shown).
The amino acid sequences of BmOrai1A and -B were aligned with those of DmOrai, Caenorhabditis elegans Orai, and human Orai1 (Fig. 5). All five proteins are predicted to have similar topology with four TM domains and both termini intra-    NOVEMBER 6, 2009 • VOLUME 284 • NUMBER 45 cellular. The TM domains comprise the regions of most homology with TM1 and TM2 exhibiting the highest degree of conservation across the four species. The glutamic acid residues essential for ion selectivity (30 -33) in TM1 and TM3 (Glu 88 and Glu 173 in BmOrai1A; Glu 51 and Glu 136 in BmOrai1B) are absolutely conserved, whereas the arginine linked to severe immunodeficiency in human patients (34) has been conservatively substituted (Lys 73 in BmOrai1A; Lys 36 in BmOrai1B). Alanine substitution studies of the acidic residues located in the first extracellular loop of DmOrai and human Orai1 suggest that these residues may play a role in ion pore selectivity (30,31). Interestingly, none of these residues are conserved in BmOrai1A or BmOrai1B suggesting that the ion selectivity properties of the silkmoth channels may be distinct from the other homologs.

Silkmoth Homologs of STIM1 and Orai1
Cellular Localization of Fluorescent-tagged BmSTIM1 and BmOrai1A/B-To examine the cellular localization of BmSTIM1 and BmOrai1A/B, we constructed fluorescent chimeras by tagging the respective C termini with fluorescent proteins (Venus and/or CFP) and transiently expressed the chimeras in cultured Sf9 and BmN cells (cell lines established from Spodoptera frugiperda and B. mori ovarian tissues, respectively). The cellular distributions of the fluorescent chimeras were examined under non-stimulated and store-depleted conditions via fluorescent confocal microscopy. Under non-stimulated conditions, BmSTIM1-Venus was limited to the cytosol as discrete vesicle-like formations as well as a defined structure reminiscent of ER localization (Fig. 6A), in contrast to Venus alone, which was observed throughout the cytosol (Fig. 6B). ER localization of BmSTIM1-Venus was confirmed via colocaliza-tion (Fig. 6, C-E) with an ER marker (anti-KDEL antibody). No STIM-associated fluorescence was observed at the cell surface.
Consistent with its putative role as an ER Ca 2ϩ sensor, B. mori STIM1-CFP co-localized with the fluorescent Ca 2ϩ -sensitive dye, Fluo4-AM (Fig. 6, F-H) suggesting that at resting state the EF-hand motif in BmSTIM1 may be complexed with Ca 2ϩ . Despite differences in the length of the N termini, both BmOrai1A-CFP and BmOrai1B-Venus were expressed at the plasma membrane (Fig. 6, I and J), an indication that the extended N terminus in BmOrai1A is not crucial for expression or cell surface targeting. Furthermore, co-expression of BmOrai1A-CFP and BmOrai1B-Venus resulted in co-localization of the fluorescent signals (Fig. 6K) opening the possibility that the two BmOrai1 variants may form heteromultimer complexes similar to those described for human Orai1-3 (33).
TG-induced Translocation of BmSTIM1-We next examined the effect of pharmacological-based depletion of ER Ca 2ϩ stores. When co-expressed with either BmOrai1A-CFP or BmOrai1B-CFP in the absence of TG stimulation in buffer replete with Ca 2ϩ , BmSTIM1-Venus distribution was limited to the cytosol (Fig. 7, A and C). In contrast, following a 15-min incubation in Ca 2ϩ -free RB supplemented with 3 mM EGTA, 25 M ionomycin, and 5 M TG, BmSTIM1-Venus co-localized at the cell surface with both BmOrai1A-CFP (Fig. 7B) and BmOrai1B-CFP (Fig. 7D). Because of resolution limitations, we were unable to determine whether the translocation of BmSTIM1-Venus resulted in insertion into the plasma membrane as has been suggested (35) or if it merely resulted in close juxtaposition of ER-associated BmSTIM1-Venus with the plasma membrane (36, 37). Regardless, our results clearly dem- Black and gray shading indicate sequence identity and conserved amino acid substitutions, respectively. Lines over the sequences indicate the TM domains. Conserved glutamic acid residues in TM1 and TM3 that have been shown to be necessary for ion selectivity are indicated by the asterisks ( * ), whereas the arrowhead indicates the arginine residue linked to severe immunodeficiency in human patients. Residues in the first extracellular loop believed to be important for regulating pore selectivity are boxed. Alignment was performed using the online CLUSTALW server at the Network Protein Sequence Analysis website (46). onstrate that TG-mediated depletion of ER Ca 2ϩ stores results in a redistribution of BmSTIM1-Venus from a cytosolic setting to regions closely associated with plasma membrane-bound BmOrai1 molecules. We also found that redistribution of BmSTIM1-Venus to the plasma membrane following TG-mediated store depletion is dependent on the co-expression of BmOrai1A or BmOrai1B. Using the same experimental conditions as described above, but in the absence of exogenous BmOrai1, there was no difference in the BmSTIM1-Venus localization pattern between non-stimulated cells (Fig. 7E) and TG-stimulated cells (Fig. 7F). Transcripts for both BmOrai1 variants, as well as BmSTIM1, however, can be amplified from these cells (data not shown) suggesting that there is a stoichastic limitation on the amount of endogenous BmOrai1 needed to complement the overexpressed BmSTIM1. Furthermore, these findings are consistent with those of Park et al. (20) who reported that plasma membrane targeting of STIM1 molecules lacking the polybasic Lys-rich region, such as BmSTIM1, requires the presence of exogenous Orai1.
Role of the BmSTIM1 CAD Region in TG-induced Translocation-A highly conserved region of the STIM1 C terminus, selectively known as CAD, STIM Orai activating region, small Orai1-activating fragment, and/or coiled-coil domain containing region b9, has recently been shown to be the minimal STIM1 requirement for Orai1 binding and activation (20 -23). The homologous region in BmSTIM1, corresponding to residues 340 -446 is 55% identical with human STIM1. To assess the role of this region, referred to as BmSTIM1-CAD, in TG-induced translocation, we transiently co-expressed C-terminal Venus chimeras of BmSTIM1 truncated at Asp 478 , Ala 377 , and Ala 277 with BmOrai1B-CFP and depleted ER Ca 2ϩ stores as before. BmSTIM1-(1-478)-Venus localization before and after TG addition (Fig. 8B) was indistinguishable from the full-length construct (Fig. 8A), indicating that residues 479 -577 likely do not play a role in BmSTIM1 translocation from the ER to the plasma membrane. Surprisingly, in the absence of ER Ca 2ϩ store depletion, BmSTIM1-(1-377)-Venus, which has half of the CAD region, was predominately localized at the plasma membrane (Fig. 8C). Depletion of the ER stores appeared to further enhance the plasma membrane localization of BmSTIM1-(1-377)-Venus and concomitantly reducing the diffuse cytosolic/ER distribution (Fig. 8C). These results suggest that residues 378 -477 of the BmSTIM1 C terminus may comprise a modulatory domain that either exerts a dominant inhibitory effect on translocation via interactions with some unidentified protein or its conformational state masks the region involved in translocation such that when it is missing translocation takes place in the absence of ER Ca 2ϩ store depletion. BmSTIM1-(1-277)-Venus, which lacks the putative CAD region, likewise localized at the plasma membrane (Fig. 8D, inset), albeit to a lesser degree and with a more pronounced cytosolic/ER distribution (Fig. 8D). In contrast to BmSTIM1-(1-377)-Venus, TG addition appeared to have no effect on the degree of cytosolically retained protein (Fig. 8D). Because these results confound analysis of the putative BmSTIM1 CAD region, we generated a BmSTIM1 CAD deletion mutant (BmSTIM1-⌬CAD-Venus) and examined its effects on TG-induced translocation. In the absence of ER Ca 2ϩ store depletion, BmSTIM1-⌬CAD-Venus exhibited a cytosolic distribution reminiscent of a diffuse ER localization (Fig. 9A). No change in the cytosolic localization was seen following TG treatment (Fig. 9A), suggesting that the putative BmSTIM1 CAD region is essential.
Intriguingly, the CAD region of all identified and putative STIM1 proteins contains a highly conserved cluster of basic amino acids (KIKKKR, residues 382-387 of human STIM1; KLRKKR, residues 380 -385 of BmSTIM1) that potentially could function in mediating STIM1 and Orai1 interactions. To determine whether these residues are important for BmSTIM1 translocation, we made multiple point mutations (K380A, K383A, K384A, R382A, and R385A) and assessed their effect on function as before. In the absence of ER Ca 2ϩ store depletion, the mutant BmSTIM1 (BmSTIM1-ALAAAA-Venus) exhibited an intracellular distribution (Fig. 9B) similar to the non-mutated full-length BmSTIM1 (Fig. 8A). TG-induced depletion of ER Ca 2ϩ stores, however, failed to promote redistribution of the mutant to the plasma membrane (Fig. 9B). These results, in conjunction with previous reports demonstrating the role of the CAD region, suggest that this cluster of basic residues is important for STIM1 recruitment to the plasma membrane.
Because BmOrai1A or -B co-expression is necessary for BmSTIM1 translocation (Fig. 7F), we sought to determine what role the BmOrai1B N and C termini have on TG-mediated BmSTIM1 translocation. Muik et al. (38) reported that the Orai1 C terminus is essential for STIM1 interactions. To examine the function of the BmOrai1B C terminus, we truncated N-terminal CFP-tagged BmOrai1B at Ser 186 (CFP-BmOrai1B-(1-186)) and transiently co-expressed it with full-length BmSTIM1-Venus. Similar to previous studies (21), removal of the Orai1 C terminus resulted in significant ER retention; however, CFP-BmOrai1B-(1-186) could be observed at the cell surface (Fig. 9C). Depletion of ER Ca 2ϩ stores with TG resulted in partial translocation of BmSTIM1-Venus (Fig. 9C, inset) with a significant portion of the protein diffused throughout the cytosol (Fig. 9C). These results indicate that, despite the C-terminal truncation, BmSTIM1-Venus is still capable of interacting with BmOrai1B, suggesting that interactions may occur within the N terminus. Recently, Park et al. (20) reported direct binding between STIM1 and the Orai1 N terminus. To examine the function of the BmOrai1B N terminus in TG-induced BmSTIM1 translocation, we truncated N-terminal CFP-tagged BmOrai1B at Lys 36 (CFP-(37-221)-BmOrai1B), generating a BmOrai1B protein devoid of its N terminus. This protein was almost entirely retained within the cytosol with little to no indication of cell surface localization (Fig. 9D). Unexpectedly, despite the lack of detectable plasma membrane-bound BmOrai1B, BmSTIM1-Venus exhibited partial cell surface localization following TG addition (Fig. 9D, inset). As with CFP- BmOrai1B-(1-186), however, a significant portion of BmSTIM1-Venus exhibited a diffuse cytosolic localization (Fig.  9D). Given that BmSTIM1 translocation is not observed in the absence of exogenous BmOrai1, these results suggest that even though the levels of the N-terminal-truncated BmOrai1B are below our detection threshold, they are sufficient for partial BmSTIM1 recruitment to the plasma membrane. Furthermore, the results of the BmOrai1B N-and C-terminal truncations suggest that efficient translocation of BmSTIM1 is dependent on the presence of both termini.
PBAN-induced Translocation of BmSTIM1-Most studies on the functions of STIM1 and Orai1 have been conducted only in the pharmacological context of TG-mediated store depletion. The role of these proteins in ROC activation, of which SOCs are a subset, has until recently largely been neglected. We have shown in dissected PGs that PBAN triggers the opening of SOCs (4). PBAN activation of PBANR heterologously expressed in cultured Sf9 cells likewise mobilizes extracellular Ca 2ϩ (7,8). To explore the role of BmSTIM1 and BmOrai1B downstream of PBANR activation, we co-expressed BmOrai1B-CFP (the BmOrai1 splice variant specific to the PG), BmSTIM1-Venus, and PBANR-EGFP in Sf9 cells, and assessed BmSTIM1-Venus localization in relation to a fluorescent rhodamine red analog of PBAN, RR-PBAN (8). We have previously shown that internalization of PBANR-associated RR-PBAN is dependent on an influx of extracellular Ca 2ϩ (39). PBANR-EGFP was used to facilitate initial identification under UV light of PBANR positive cells via fluorescence localized at the plasma membrane. Insect cells transiently expressing the proteins of interest were incubated for 30 min with 50 nM RR-PBAN in the presence or absence of external Ca 2ϩ prior to examining the intracellular localization of BmSTIM1-Venus via fluorescent confocal microscopy. To minimize overlap in the fluorescent emission profiles, in particular between PBANR-EGFP and BmSTIM1-Venus, cells were excited at 442 (CFP, max ϭ 433 nm), 514 (Venus, max ϭ 515 nm), and 568 nm (rhodamine red, max ϭ 560 nm). In the presence of external Ca 2ϩ , BmSTIM1-Venus was limited to large punctate structures in the cytosol that did not co-localize with BmOrai1B-CFP or the internalized RR-PBAN (Fig. 10A). These results demonstrate that overexpression of BmOrai1B and BmSTIM1 had no effect on RR-PBAN internalization and, given the cytosolic distribution of BmSTIM1-Venus, suggest that the PBAN-stimulated Ca 2ϩ influx refills ER stores. In the absence of external Ca 2ϩ , however, RR-PBAN failed to internalize and co-localized with BmSTIM1-Venus and BmOrai1B-CFP at the cell surface (Fig.  10B). These data provide a clear role for the STIM1-Orai1 pathway in ROC Ca 2ϩ influx and suggest that under physiological conditions the PBAN-PBANR complex in the PG triggers the opening of SOCs via translocation of endogenous ER-bound BmSTIM1 to regions of the plasma membrane replete with BmOrai1B molecules. In Vivo Knockdown of BmSTIM1 and BmOrai1-We previously demonstrated the effectiveness of dsRNA-mediated RNA interference in elucidating the downstream cellular targets of PBAN binding (40). We expected that a similar strategy would allow us to determine whether BmSTIM1 and BmOrai1B function within the biological context of sex pheromone produc-  tion. In that study, we injected B. mori pupae 1-2 days removed from the larval-pupal molt. However, because of the near ubiquitous expression pattern of BmSTIM1 and BmOrai1 the potential for confounding problems during adult development precluded this methodology. As an alternative approach, we directly injected dsRNAs (10 g in 2 l) into the PG of newly emerged female moths and assayed for production of the principal silkmoth sex pheromone component, (E,Z)-10,12-hexadecadien-1-ol (i.e. bombykol), 2 days later. Injection of dsRNAs corresponding to a 1-kb fragment of BmSTIM1 or to the entire BmOrai1B ORF reduced bombykol production to 46 and 43% of controls, respectively (Fig. 11). No decrease in bombykol production was observed in females injected with dsRNA corresponding to EGFP (Fig. 11). Because the observed decrease in bombykol production could have been the result of a global defect affecting PG cell viability instead of a disrupted PBAN signaling cascade, we sought to determine whether STIM1 and Orai1 knockdown affected the normal cellular function of GAPDH and catalase, enzymes that, respectively, function in glycolysis and the breakdown of H 2 O 2 . It is reasonable to assume that decreased cell viability would impact both GAPDH and catalase. We found that both enzymatic activities in dsRNA-injected PGs were comparable with non-injected PGs (data not shown), indicating that decreased bombykol production following BmSTIM1 and BmOrai1B knockdown is specific to disruption of the PBAN signaling cascade and not the result of a global defect. This is consistent with previous studies that reported dsRNA-mediated knockdown of STIM1 in cultured cells had no apparent affect on cell viability (41,42).

DISCUSSION
For most female moths, sex pheromone production is dependent on an influx of extracellular Ca 2ϩ (2,4), presumably via ROCs (5,6). We have recently shown in B. mori that PBAN binding triggers the opening of SOCs (4), suggesting the possible involvement of STIM1 and Orai1; two proteins that have been identified as essential components of SOC pathways (12)(13)(14)(15). Consequently, the present study was designed to identify silkmoth homologs of STIM1 and Orai1 and to assess their role, if any, in the B. mori PBAN signal transduction cascade.
Using a homology-based cloning approach, we identified B. mori homologs of STIM1 and Orai1. BmSTIM1 is 64% identical at the protein level to DmSTIM, but the genomic organization of the bmstim1 locus appears to more closely resemble that of the human locus. In humans and B. mori, the cytoplasmic portions of STIM are encoded on 7 and 6 exons, respectively, whereas that of DmSTIM is limited to 2 exons (Fig. 2B). The similarities with human STIM1 also appear to extend to the transcript level, as Northern blots ( Fig. 2A) of PG total RNA showed two STIM1-hybridizing bands of similar sizes (ϳ4 and 4.5 kb) to those reported for human STIM1 (24). In addition, BmSTIM1 transcripts are broadly expressed, albeit with varying abundance depending on the cell background (Fig. 4). In contrast to proteins previously identified as essential to the B. mori sex pheromone biosynthetic pathway (8,27,28), BmSTIM1 does not undergo up-regulation within the PG in relation to adult emergence.
In our efforts to clone the B. mori homolog of Orai1, we unexpectedly identified two isoforms: BmOrai1A, a 777-bp transcript encoding a 258-amino acid protein with 64% sequence identity to DmOrai, and BmOrai1B, a 666-bp transcript encoding a 221-amino acid protein with 76% sequence identity to DmOrai (Fig. 5). Both forms are broadly expressed, often with co-expression of the two as independent transcripts but with the BmOrai1A transcript frequently predominating (Fig. 4). Surprisingly, the BmOrai1B variant was the only transcript amplified in PGs. In addition to varying expression profiles, the two forms differ in the length of their N termini (BmOrai1B has a 37-amino acid N-terminal truncation). The extended BmOrai1A N terminus, however, is not necessary for expression and targeting to the plasma membrane as heterologously expressed BmOrai1A and BmOrai1B co-localized at the cell surface (Fig. 6, I-K). The presence of only the N-terminaltruncated BmOrai1B isoform in the PG, as opposed to both isoforms observed in most of the other tissues examined, opens the possibility that regulation of SOC function in the context of sex pheromone production is distinct. Alternative splicing of the bmorai1 gene likely gives rise to the two isoforms (Fig. 3B). The BmOrai1A transcript is generated from exons 1, 3, and 4 with a start methionine localized in exon 1. In contrast, the BmOrai1B transcript is generated from exons 1, 2, 3, and 4 but with exons 1 and 2 giving rise to the 5Ј UTR with the start methionine in exon 3. Two mouse Orai2 variants that differ only in the length of their N termini are generated from a similar alternative splicing mechanism (43).
Depending on the state of the ER Ca 2ϩ stores, STIM1 and Orai1 localize to specific cellular regions. At resting state, STIM1 is localized to the ER membrane, whereas Orai1 is at the cell surface. Depletion of ER Ca 2ϩ triggers a redistribution of STIM1 to form subplasma membrane clusters in close proximity to the plasma membrane-embedded Orai1. We found that under non-stimulated conditions, heterologously expressed BmSTIM1-Venus was restricted to the cytosol (Fig. 6A) and co-localized with an ER membrane marker (Fig. 6, C-E). No BmSTIM1-Venus was observed at the cell surface. In contrast, TG-mediated depletion of ER Ca 2ϩ stores in Ca 2ϩ -free RB triggered a redistribution of BmSTIM1-Venus to the plasma membrane where it co-localized with BmOrai1A-CFP as well as FIGURE 11. Effects of dsRNA injections on in vivo bombykol production. PGs of newly emerged female moths were injected with 10 g of dsRNAs corresponding to EGFP, BmSTIM1, or BmOrai1B. Results are expressed relative to bombykol levels in non-RNA interference-treated females following injections of 5 pmol of PBAN. Bars represent means Ϯ S.E. of replicates from multiple independent experiments. ***, statistically significant differences from PBAN alone (p Ͻ 0.001) and determined via analysis of variance with Tukey's multiple comparison test.
BmOrai1B-CFP (Fig. 7, B and D). These molecular dynamics of BmSTIM1 and BmOrai1 are consistent with recent conclusions on the molecular basis for STIM1-mediated SOCE initiation in mammals: i.e. whereas STIM is localized to the ER membrane at resting state, depletion of ER Ca 2ϩ stores triggers STIM1 translocation from the ER to the plasma membrane where it interacts with Orai1 oligomers and thereby in activation of SOCE (14,15). We further found that translocation from a cytosolic distribution to a cell surface distribution was dependent on coexpression of the BmOrai1 isoforms. No redistribution was seen in cells expressing only BmSTIM1-Venus (Fig. 7F) despite the presence of endogenous BmOrai1A and BmOrai1B. This observation suggests that a functional stoichiometry similar to that recently reported by Ji et al. (44) may need to be achieved for BmSTIM1 redistribution to occur. Our observations are also consistent with a recent report demonstrating that mammalian STIM1 proteins lacking the polybasic cluster of residues at their extreme C terminus, which BmSTIM1 lacks, requires the presence of exogenous Orai1 for plasma membrane clustering (20).
A number of recent reports (20 -23) have identified a region of the STIM1 C terminus as essential for Orai1 binding and activation. Based on sequence homology we identified a homologous region (residues 340 -446) in BmSTIM1, referred to as BmSTIM1-CAD. Truncation studies aimed at elucidating the role of this region in TG-induced BmSTIM1 translocation unexpectedly uncovered a potential modulatory domain. In the absence of ER Ca 2ϩ depletion, BmSTIM1-(1-477)-Venus localized to the ER, whereas BmSTIM1-(1-377)-Venus and BmSTIM1-(1-277)-Venus were partially localized at the plasma membrane (Fig. 8, B-D), suggesting that residues 378 -477 are involved, either directly or indirectly, in modulating BmSTIM1 recruitment to the plasma membrane. Similar observations have recently been reported for C-terminal fragments of human STIM1: residues 475-485 (22) and residues 445-475 (23). The overlapping region of BmSTIM1, however, is only 33% similar to the putative modulatory domain of human STIM1, suggesting that the inhibitory effect on translocation may arise from interactions with some unidentified protein, it is more likely though that the conformational state of this region masks the residues involved in translocation, such that when it is missing inhibition of translocation in the absence of ER Ca 2ϩ store depletion is removed. Because the truncation studies failed to address the role of the BmSTIM1 CAD region in translocation, we generated a CAD deletion mutant. Consistent with recent reports, we found that the ⌬CAD mutant failed to translocate to the plasma membrane (Fig. 9A). This observation suggests that the recruitment of BmSTIM1-(1-277)-Venus, which lacks the BmSTIM1 CAD region, to the cell surface occurs by a CAD-independent mechanism. Intriguingly, we found that a cluster of basic residues (Lys 380 -Arg 385 ) are essential for translocation and possibly explains how CAD functions (Fig. 9B). We speculate that impaired translocation resulted from disrupted electrostatic interactions between the negative charges in this region and corresponding positive charges in the BmOrai1B C terminus. In support of this hypothesis, Calloway et al. (45) reported that mutation of acidic residues in the cytoplasmic tail of Orai1 reduced STIM1 interactions, suggesting that electrostatic forces are involved in STIM1/Orai1 binding. In addition, we also showed that both the N and C termini of BmOrai1B are necessary for complete translocation of BmSTIM1 to the plasma membrane (Fig. 9, C and D). Unexpectedly, we found that loss of the BmOrai1B N terminus severely limited cell surface expression (Fig. 9D). This is in contrast to a previous report showing that that N-terminal-tagged CFP-Orai1 molecules lacking the native N terminus retained targeting to the plasma membrane (38), suggesting that the targeting mechanisms between mammalian and insect cells may be different.
To examine the physiological role for BmSTIM1 and BmOrai1B in moth sex pheromone production, we co-expressed both proteins with the B. mori PBANR, which had previously been shown to trigger Ca 2ϩ influx following PBAN activation (8). In the presence of extracellular Ca 2ϩ , BmSTIM1 remained localized within the cytosol (Fig. 10A), whereas removal of the divalent cation resulted in redistribution of BmSTIM1 to the cell periphery (Fig. 10B). These results provide a clear indication that the STIM1/Orai1 pathway is functional downstream of PBANR activation. We further examined the role of BmSTIM1 and Orai1 within the physiological context of sex pheromone production by using RNA interference silencing techniques to specifically knockdown expression of the two proteins within the PG. We found that direct injection of dsRNAs corresponding to portions of either gene significantly reduced sex pheromone production (Fig. 11) without affecting PG cell viability. The knockdown efficacy, however, was not complete, as bombykol production was not abolished. This partial reduction may have been the result of inefficient uptake of the dsRNAs by the pheromone producing cells, which were injected on the day of emergence and not during PG development as previously described (40). Alternatively, the partial reduction may be an indication that more than one type of channel is opened following PBAN binding. Regardless, these results clearly indicate that BmSTIM1 and BmOrai1B function, most likely in a stoichiometrically defined complex, in the B. mori PBAN signaling cascade, and by extension within the biological framework of bombykol production. Furthermore, given the high degree of similarity observed in the initial steps of the PBAN signal transduction cascade (i.e. dependence on extracellular Ca 2ϩ and involvement of ROCs), it is likely that STIM1 and Orai1 are essential to SOCE in most female moth sex pheromone biosynthetic pathways.