Identification and characterization of the bombykal receptor in the hawkmoth Manduca sexta

Crepuscular and nocturnal moths have developed exquisitely sensitive olfaction to locate their mates (reviewed in Kaissling, 2014). Females of the nocturnal hawkmoth Manduca sexta (Linnaeus 1763) release pulses of a species-specific sex-pheromone blend of eight aliphatic aldehydes while calling for their mates. Bombykal (E,Z)-10,12-hexadecadienal (BAL) and (E,E,Z)-10,12,14-hexadecatrienal are the two major blend components necessary to evoke male reproductive behavior (Starrat et al., 1979; Tumlinson et al., 1989, 1994). Conspecific males detect their species-specific sex-pheromone blend with olfactory receptor neurons (ORNs) on their antennae (for reviews, see Homberg et al., 1989; Kaissling et al., 1989). The ORNs express odorant receptors (ORs) in their dendritic cilia that innervate long, hair-like, trichoid sensilla (Altner and Prillinger, 1980; Keil and Steinbrecht, 1984). Odorant-binding ORs (OrX) in insects belong to a new family of 7TM proteins with inverted membrane topology, with an intracellular N terminus. They heteromerize with a larger, ubiquitous, inverted 7TM protein termed olfactory receptor coreceptor (Orco) that locates and maintains ORs in the ciliary membranes of ORNs (Clyne et al., 1999; Vosshall et al., 1999; Larsson et al., 2004; Benton et al., 2006; Vosshall and Hansson, 2011). In addition to this ‘chaperone-like function’, Orco forms a non-specific, spontaneously opening Ca²⁺-permeable cation channel in heterologous expression systems (Sato et al., 2008; Wicher et al., 2008; Jones et al., 2011; Sargsyan et al., 2011; Nolte et al., 2013). Because ORNs in Orco-mutant flies showed strongly diminished spontaneous action potential activity, Orco controls the membrane potential and, thus, the spontaneous activity of ORNs as a prominent leak channel (Larsson et al., 2004; Benton et al., 2007; Deng et al., 2011). Orco's property as a leaky ion channel controlling spontaneous membrane potential oscillations was previously termed the ‘pacemaker channel function’ to distinguish it from the ‘chaperone function’ of Orco (Stengl, 2010). Furthermore, experiments with heterologously expessed OrX/Orco complexes suggested that heteromers of OrX and Orco are able to form ligand-gated ion channels (German et al., 2013; Nakagawa et al., 2012; Sato et al., 2008; Wicher et al., 2008) that promote ionotropic primary transduction (Sato et al., 2008; Wicher et al., 2008), while there is slower metabotropic transduction at a later stage of the fruitfly's odor response (Wicher et al., 2008). Whether and how Orco contributes to odor transduction in vivo in different insect species is still under debate (Nakagawa and Vosshall, 2009; Nolte et al., 2013, 2016).

Sensitization of odor responses in intact fruitflies relied on proper Orco function, which was impaired with insufficient Orco phosphorylation by protein kinase C (PKC) (Sargsyan et al., 2011). Thus, for sensitization to occur, first, Orco needed to be phosphorylated via PKC, before it could be opened directly via cAMP binding. Furthermore, Drosophila melanogaster Orco can already be gated via phospholipase C (PLC) activation, even in absence of cAMP or cGMP (Sargsyan et al., 2011). It remained unknown whether odor stimulation first activates PLC for proper Orco phosphorylation or whether background activity of PLC keeps Orco phosphorylated. Patch clamp studies on primary cell cultures of M. sexta ORNs as well as tip recordings of pheromone-sensitive sensilla in intact hawkmoths indicated that different PLC- and PKC-dependent ion channels in ORNs play important roles in pheromone transduction, but it remained undetermined whether any of these many ion channels is Orco (Stengl, 1993, 1994, 2010; Stengl et al., 1992). In addition, it has not yet been rigorously examined whether in heterologous expression systems, OrX/Orco-dependent odor responses are mediated via Ca²⁺-dependent enzymes such as PLC and PKC.

As a prerequisite to answering open questions concerning hawkmoth pheromone transduction, it first needs to be determined whether either of the two putative hawkmoth pheromone receptor proteins, MsexOr1 and MsexOr4, responds to the main sex pheromone component BAL (Grosse-Wilde et al., 2011). It was previously noticed by different research groups that it was very difficult to obtain functional membrane insertion of MsexOrco and MsexOrX proteins, even more difficult than in other insect species. Therefore, the present study was dedicated to identifying the BAL receptor in M. sexta while employing different expression systems. Furthermore, because MsexOrco is Ca²⁺-permeable and increases intracellular Ca²⁺ levels in the cells expressing this protein, and because MsexOrX and MsexOrco proteins possess putative PKC phosphorylation sites, we asked whether manipulation of the Ca²⁺-dependent enzymes PLC and PKC may modify BAL response in different heterologous expression systems.

For the de-orphanization experiments, (E,Z)-10,12-hexadecadienal (BAL) was employed at concentrations of 100 nmol l⁻¹ to 300 µmol l⁻¹. To test the specificity of the hawkmoth pheromone receptor, the female silkworm sex pheromone bombykol (E,Z)-10,12-hexadecadienol (BOL) was used at a concentration of 100 μmol l⁻¹ for comparison. Because the trienal (E,E,Z)-10,12,14-hexadecatrienal is a relatively unstable substance, a more stable chemical mimic, (E,Z)-11,13-pentadecadienal (E11,Z13-15:AL) (C15), was employed (Christensen and Hildebrand, 1997).

Human embryonic kidney (HEK293) cells or Chinese hamster ovary (CHO) cells (Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany) were grown in a 1:1 mixture of Dulbecco's modified Eagle medium and Ham's F-12 medium with l-glutamine (PAA, Cölbe, Germany) containing 10% fetal bovine serum (PAA) at 37°C and 5% CO₂ content. HEK293 cells or CHO cells were plated on poly-l-lysine (0.01%, Sigma-Aldrich, Steinheim, Germany) coated coverslips in 24-well plates and cultured at a density of ∼10⁵ per well. The M. sexta OR constructs MsexOrco and MsexOr1 (Grosse-Wilde et al., 2010) were transfected at a 0.5–0.7 µg well⁻¹ concentration using Transficient (MBL International Corporation, Woburn, MA, USA).

Fluorescence images were acquired using a cooled CCD camera controlled by TILLVision 4.0 software (T.I.L.L. Photonics). The resolution was 640×480 pixels in a frame of 175×130 µm (40×/IR/0.8 objective). Image pairs were obtained by excitation for 150 ms at 340 and 380 nm; background fluorescence was subtracted. Cells were continuously perfused with bath solution in the recording/perfusion chamber (RC-27, Warner Instruments Inc., Hamden, CT, USA). The bath solution contained (in mmol l⁻¹): 135 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose; pH 7.4; osmolarity=295 mOsmol l⁻¹. BAL and the modulators were applied using either a bath perfusion system (BPS4 from ALA, New York, NY, USA) or a rapid solution changer (RSC160 from Biologic, Claix, France). BOL was obtained from Pheropbank (Wijk bij Duurstede, The Netherlands); the modulators U73122, U73343 and Gö6976 were obtained from Calbiochem (Darmstadt, Germany). BAL was synthesized by the working group ‘Mass Spectrometry/Proteomics’ of the Max-Planck Institute for Chemical Ecology (Jena, Germany). BAL, BOL and C15 were dissolved and applied in 0.01% dimethylsulfoxide (DMSO). For the analysis, only responding cells were used. Responses are characterized by a clear BAL-stimulation-related increase in [Ca²⁺]_i.

cRNA synthesis, oocyte microinjection and two-electrode voltage-clamp recording were performed as previously described (Hopf et al., 2015). Briefly, oocytes were microinjected with 6.25 ng cRNA encoding a ligand-specific OrX protein (MsexOr1, MsexOr4 or BmOr3) and 6.25 ng cRNA encoding Orco cRNA (MsexOrco). Currents were recorded and amplified using an OC-725C amplifier (Warner Instruments, Hamden, CT, USA). Before electrophysiological recording, BAL was prepared in DMSO as 10 to 300 mmol l⁻¹ stock solutions and C15 as 1 to 10 mmol l⁻¹ stock solutions, which were then diluted into the bath solution at 0.1% to give the final desired concentration.

To identify M. sexta's BAL receptor, the pheromone receptor candidates MsexOr1 and MsexOr4 (Grosse-Wilde et al., 2010) were coexpressed with MsexOrco first in X. laevis oocytes and then in HEK293 and CHO cells. Responses to different pheromone components were examined with two-electrode voltage-clamp recordings or with Ca²⁺ imaging.

In X. laevis oocytes expressing MsexOrco and MsexOr1, BAL stimulation activated inward currents in a concentration-dependent manner, with a threshold concentration of 30 μmol l⁻¹ (Fig. 1A). In addition, coexpression of MsexOrco with the BAL receptor BmOr3 from the silkmoth Bombyx mori also allowed for BAL-induced inward currents (Fig. 1B). The maxima of BmOr3-mediated BAL responses were higher than the MsexOr1-dependent responses, hinting at better membrane insertion of the silkmoth BAL receptor. In contrast to MsexOr1, coexpression of the second pheromone receptor candidate MsexOr4 with MsexOrco did not elicit BAL-induced inward currents (Fig. 1C). Because only the highest BAL concentration tested elicited very small responses, it is likely that MsexOr4 is indeed expressed in the plasma membrane, but that it is not specific for BAL. To examine whether MsexOR4 expressed in oocytes responds to minor pheromone components such as (E,E,Z)-10,12,14-hexadecatrienal, it was stimulated with the more stable mimic (E,Z)-11,13-pentadecadienal at concentrations of 1 and 10 μmol l⁻¹. However, no specific responses were obtained at any of the concentrations tested (Fig. 1D).

Thus, further investigations focused on MsexOr1 expression. To further examine the sensitivity of the hawkmoth BAL receptor, MsexOrco and MsexOr1 were expressed in HEK293 cells and Ca²⁺ imaging was employed to monitor receptor activation. While for femtomolar to nanomolar BAL concentrations the percentage of cells showing a response to stimulation did not significantly differ from the response to solvent control (0.1% DMSO, median: 0.0%), for 1 μmol l⁻¹ BAL there was a significantly higher percentage of responding cells (median: 6.2%; Fig. 2A,B). To examine the selectivity of MsexOrco and MsexOr1 to the hawkmoth pheromone, the receptor complex was stimulated with BOL, the sex pheromone of female B. mori. Application of 100 μmol l⁻¹ BOL produced only weak responses in comparison to 100 μmol l⁻¹ BAL (Fig. 2C).

As only coexpression of MsexOrco with MsexOr1 but not with MsexOr4 conferred BAL sensitivity, MsexOr1 is the BAL-ligand binding subunit. Because it was very difficult to obtain any plasma membrane expression of any of the subunits tested, it was not examined whether MsexOr1 expression alone is sufficient to elicit BAL responses, as was found for OrX proteins from other species (Wetzel et al., 2001; Sakurai et al., 2004; Nakagawa et al., 2005; Neuhaus et al., 2005; Grosse-Wilde et al., 2006; Smart et al., 2008; Deng et al., 2011). It still remains unknown whether MsexOr4 would detect minor components of the M. sexta pheromone blend. Because application of (E,Z)-11,13-pentadecadienal (up to 10 μmol l⁻¹), which mimics (E,E,Z)-10,12,14-hexadecatrienal, did not activate a current in oocytes expressing MsexOr4+MsexOrco constructs, the respective pheromone receptor of the second behaviorally relevant main pheromone compound still needs to be identified in the hawkmoth (Tumlinson et al., 1989, 1994; Kaissling et al., 1989). Alternatively, the lack of odor responsiveness could also indicate that MsexOr4 was not inserted into the plasma membrane. However, the weak response to high BAL concentration (Fig. 1C) indicates at least a minimum of membrane expression of MsexOr4.

As compared with heterologous expression of Orco and OrX proteins from other species, it was considerably more difficult to express M. sexta receptors in the plasma membrane of the different heterologous expression systems employed (Fig. 1A,B). Apparently, more molecules next to the chaperone Orco are necessary to insert MsexOrX proteins in the plasma membrane. In D. melanogaster, pheromone receptors are composed of the Orco/OrX construct and a sensory neuron membrane protein (SNMP) (Benton et al., 2007; Jin et al., 2008; Li et al., 2014). Furthermore, interactions between SNMP-1 and ligand-binding Ors, but not with Orco, were found in FRET experiments (German et al., 2013). The antennal transcriptome of M. sexta contains the message of SNMP-1 (Grosse-Wilde et al., 2011), which is also known to be expressed in pheromone-receptor-expressing sensory neurons in Antheraea polyphemus and Heliothis virescens (Forstner et al., 2008). This could indicate that SNMP-1 might be part of the M. sexta BAL receptor complex, necessary for plasma membrane insertion. Coexpression of MsexOr1+MsexOrco with SNMP-1 in HEK293 cells allowed for stronger Ca²⁺ responses upon stimulation with the synthetic Orco agonist VUAA1 (Jones et al., 2011) as compared with cells expressing only MsexOrco (Nolte et al., 2013). However, in the presence of SNMP together with MsexOrco and MsexOr1, only a few HEK293 cells appeared to express these molecules in the plasma membrane, allowing for Orco activation via its agonist VUAA1. Therefore, there are still more unknown molecular components necessary for stable plasma membrane insertion and maintenance.

Next, a possible dependence of the BAL response on PLC or PKC activity was examined, as reported previously for hawkmoth pheromone transduction (Stengl, 1993, 1994, 2010; Dolzer et al., 2008) and for fruitfly Orco (Sargsyan et al., 2011; Getahun et al., 2013). Thus, Ca²⁺ imaging experiments were performed in HEK293 cells expressing MsexOrco and MsexOr1. Application of 100 μmol l⁻¹ BAL induced an immediate, transient rise of [Ca²⁺]_i from 100 to 500 nmol l⁻¹ (Fig. 3A). It should be noted that the response to BAL in HEK293 is much slower than that in oocytes.

The rise in [Ca²⁺]_i in response to BAL application was strongly reduced and delayed by the PLC inhibitor U73122 (Fig. 3A,B). Within 250 s of BAL stimulation, there was a slow rise in [Ca²⁺]_i without reaching a clear peak in the presence of U73122. By contrast, the inactive analog U73343 had only minor effects (Fig. 3A,B). Also, PKC inhibition with Gö6976 significantly reduced and delayed BAL-induced [Ca²⁺]_i responses (Fig. 3A,B). Thus, BAL responses depended on PLC activity, as well as on PKC activity.

To exclude the possibility that the effect of PLC/PKC inhibition on BAL-induced Ca²⁺ responses was a specific property of the expression system, comparative experiments were performed in CHO cells. As shown for HEK293 cells, BAL application elicited a robust Ca²⁺ response in CHO cells expressing MsexOrco and MsexOr1. Preincubation of cells with U73122 or Gö6976 nearly abolished the BAL-dependent rise in [Ca²⁺]_i (Fig. 3C,D).

Thus, PLC/PKC activity is a prerequisite to BAL-receptor-mediated Ca²⁺ signals in HEK293 and CHO cells. This finding could be explained by different hypotheses. It could be hypothesized that MsexOrs and MsexOrco need to be phosphorylated before they can be gated by BAL as an ionotropic odor receptor–ion channel complex, as suggested previously in D. melanogaster (Sargsyan et al., 2011; Getahun et al., 2013). Alternatively, it could be hypothesized that MsexOrs couple to G_αq proteins, requiring activation of PLC for pheromone transduction, as suggested by previous findings in moths and cockroaches (Breer et al., 1990; Boekhoff et al., 1993, 1994; for a review, see Stengl, 2010). More experiments in different insect species are necessary to resolve the still unsolved puzzle of insect odor/pheromone transduction.

The authors thank S. Kaltofen and S. Bucks for assistance, and Drs J. Weißflog and A. Svatos for providing BAL.