Specificity Determinants of the Silkworm Moth Sex Pheromone

The insect olfactory system, particularly the peripheral sensory system for sex pheromone reception in male moths, is highly selective, but specificity determinants at the receptor level are hitherto unknown. Using the Xenopus oocyte recording system, we conducted a thorough structure-activity relationship study with the sex pheromone receptor of the silkworm moth, Bombyx mori, BmorOR1. When co-expressed with the obligatory odorant receptor co-receptor (BmorOrco), BmorOR1 responded in a dose-dependent fashion to both bombykol and its related aldehyde, bombykal, but the threshold of the latter was about one order of magnitude higher. Solubilizing these ligands with a pheromone-binding protein (BmorPBP1) did not enhance selectivity. By contrast, both ligands were trapped by BmorPBP1 leading to dramatically reduced responses. The silkworm moth pheromone receptor was highly selective towards the stereochemistry of the conjugated diene, with robust response to the natural (10E,12Z)-isomer and very little or no response to the other three isomers. Shifting the conjugated diene towards the functional group or elongating the carbon chain rendered these molecules completely inactive. In contrast, an analogue shortened by two omega carbons elicited the same or slightly higher responses than bombykol. Flexibility of the saturated C1–C9 moiety is important for function as addition of a double or triple bond in position 4 led to reduced responses. The ligand is hypothesized to be accommodated by a large hydrophobic cavity within the helical bundle of transmembrane domains.


Introduction
The identification of bombykol, (10E,12Z)-hexadecadien-1-ol (1), the sex pheromone for the silkworm moth, Bombyx mori [1], more than five decades ago triggered physiologists' interest in insect olfaction, and paved the way for current molecular studies. Probing the system with earlier techniques such as electroantennogram (EAG) and single-sensillum recordings (SSR), pioneers in the field unraveled an inordinate sensitivity and selectivity of the insect's olfactory system [2]. These earlier studies clearly demonstrated that structural modifications dramatically reduce neuronal responses or render the molecules completely inactive [3], but it remains mostly unknown how pheromone molecules interact with odorant receptors (ORs) housed in these neurons, although various moth sex pheromone receptors have been deorphanized to date [4][5][6][7][8][9][10][11]. To identify pheromone specificity determinants, we challenged with a panel of bombykol analogs the silkworm moth sex pheromone receptor, BmorOR1, co-expressed with its obligatory co-receptor, BmorOrco [4] in the Xenopus oocyte system. As the BmorOR1NBmorOrco-expressing oocytes showed robust and moderate responses to bombykol and bombykal, respectively, we investigated whether a functional recombinant pheromone-binding protein, BmorPBP1 [12], would enhance selectivity. Here, we provide strong evidence that bombykol does not require BmorPBP1 to activate BmorOR1. Additionally, we show that the stereochemistry of the double bonds, flexibility of saturated moiety, the functional group, and the number of carbons atoms after the unsaturations are specificity determinants of the pheromone molecule.

Selectivity of the Functional Group
First, we examined the response of BmorOR1NBmorOrcoexpressing oocytes to bombykol. The silkworm moth receptor responded to the sex pheromone in a dose-dependent fashion (EC 50 4.54610 28 M) and with a remarkable low threshold (,0.1 nM) ( Figure 1). Then, we compared the OR responses elicited by bombykol and bombykal. The literature is dichotomous regarding the selectivity of BmorOR1 towards these two components of the silkworm moth's sex pheromone system [3]. Using the Xenopus oocyte recording system, it has been shown that BmorOR1NBmorOrco is narrowly tuned to bombykol [4]. By contrast, it has been reported that BmorOR1-expressing HEK 293 cells responded almost equally to bombykol and bombykal [5]. In our hands, BmorOR1NBmorOrco-expressing oocytes were indeed more sensitive to bombykol, but responded to bombykal with about one order of magnitude higher threshold ( Figure 2). After activation stimulus was applied, oocytes were thoroughly washed until a steady baseline was reached. To save odorant samples and expedite these recovery times, all comparative studies were made by injecting test odorants rather than by perfusion, and comparative EC 50 s were calculated on the basis of source doses. Therefore, they are underestimation of the actual EC 50 s. The comparative EC 50 for bombykol and bombykal were 9.9610 27 M and 9.6610 26 M, respectively ( Figure 2). We analyzed our synthetic samples just prior to electrophysiological recordings to avoid possible misinterpretation derived from sample quality. There are two potential problems to consider, i.e., aldehydes are prone to degradation through auto-oxidation leading to lower than nominal concentrations and the bombykal sample may contain considerable amounts of unreacted bombykol (used as starting material). Our chemical analysis indicated that the two samples had the same concentration and that bombykol contamination in bombykal samples is very low (,0.9%) (Figure 3). If the response would be elicited by residues of bombykol in the bombykal samples, one would expect at least 2 orders of magnitude differences. Interestingly, the responses of the ''naked receptor'' differ from the neuronal activity of the olfactory system of the silkworm, which showed no cross-over whatsoever, with the bombykol and bombykal neurons responding specifically to the alcohol and aldehyde, respectively [2,3]. It has been suggested that addition of a pheromone-binding protein, BmorPBP1, to the HEK 293 cell system restores selectivity [5].

Bombykol and Bombykal are ''Trapped'' by BmorPBP1
In an attempt to reconcile the data in the literature we investigated whether addition of PBP would enhance selectivity of the BmorOR1NBmorOrco receptor complex when expressed in Xenopus oocytes. We compared the receptor responses to bombykol and bombykal solubilized either by DMSO or BmorPBP1. Interestingly, receptor activity was dramatically reduced when the ligands were solubilized by BmorPBP1 ( Figure 4). Bombykol (1 mM) dissolved in DMSO elicited robust receptor response, but very weak response when solubilized by BmorPBP1. Here, the ratio of BmorPBP1 to bombykol was 10:1. Bombykal (10 mM) elicited strong response when dissolved in DMSO and weak response when solubilized by BmorPBP1. The receptor response to bombykal solubilized by BmorPBP1 was on average ca. 34% of the response to the same ligand in DMSO, whereas for bombykol the ratio was 13%. This relatively higher response to bombykal in PBP might be merely because of the ratio of PBP:ligand. Given that bombykal requires a 10x higher dose, we prepared samples at a 1:1 ratio, whereas bombykol samples had a 10:1 protein/ligand ratio. These findings suggest that in Xenopus oocyte there are no negatively-charged surfaces in the vicinity of the receptors or the vitelline membrane surrounding the oocytes prevents the PBPodorant complexes from interacting with regions of localized low pH, which are necessary to trigger a conformational change that ''ejects'' ligands from PBPNpheromone complexes [12][13][14]. Regardless, the robust responses recorded without PBPs (Figures 1 and 2) strongly suggest that, unlike what has been  demonstrated for Obp76a = LUSH [15,16] in D. melanogaster, PBPpheromone complexes are not necessary for activation of moth ORs.

Pheromone Stereochemistry
It is well-known that position and configuration of unsaturation plays a crucial role in pheromone chemistry, but it is unknown if specificity is determined by pheromone receptors alone or in combination with other olfactory proteins. We tested the four possible isomers of bombykol (compounds 1, 3-5, Figure 5) and found that BmorOR1NBmorOrco-expressing oocytes respond with high intensity only to the natural stereoisomer of bombykol, (10E,12Z)-hexadecadien-1-ol, with very low responses to the (10Z,12E)-and (10Z,12Z)-isomers, and no response to the (10E,12E)-isomer ( Figure 6). These findings suggest that stereochemistry selectivity is mediated entirely by the receptor. This is in line with the experimental observation that, albeit with different affinities, all four geometric isomers of bombykol bind to the pheromone-binding protein, BmorPBP1 [17]. We also tested whether these double bonds could be replaced by triple bonds, but the receptor was not activated by 10,12-hexadecadiyn-1-ol (6) ( Figure 7). Next, we compared the effect of the alkyl moiety distal to the unsaturation. Elongating the bombykol molecule by adding two omega carbons renders (10E,12Z)-octadecadien-1-ol (7) completely inactive ( Figure 8). However, truncating two omega carbons led to a molecule with apparent higher affinity for the odorant receptor. Indeed, BmorOR1NBmorOrco receptor complex responded to (10E,12Z)-tetradecadien-1-ol (8) with nearly the same or even slightly higher intensity than that elicited by the native ligand, bombykol ( Figure 8). Contrary to the stringent requirement for unsaturation with the proper stereochemistry, our findings suggest that the binding pocket in BmorOR1NBmorOrco can accommodate a shorter ligand thus begging questions about the length and flexibility of the moiety between the functional group and unsaturation.

Flexibility and Length of the C1-C9 Saturated Moiety
To evaluate the positional effect of the unsaturation, we tested another ligand with the double bonds shifted towards the functional group, i.e., (8E,10Z)-hexadecadien-1-ol (9). This ligand showed minimal activation of the BmorOR1NBmorOrco receptor complex (Figure 7) thus implying that the length between the unsaturation and functional group is critical for receptor activation. To determine if the flexibility generated by an unsaturated moiety is important, we tested two bombykolrelated compounds each with an additional unsaturation between the functional group and the conjugated double bond moiety. The moderate and low responses elicited by (10E,12Z)hexadecadien-4-yn-1-ol (10) and (4Z,10E,12Z)-hexadecatrien-1ol (11), respectively ( Figure 9), strongly suggest that flexibility of the unsaturated moiety is essential for fitting into the binding pocket, particularly given the stronger effect of the double than the triple bond.

Conclusions
Structure activity analysis showed that the most important features of the sex pheromone of the silkworm moth are the stereochemistry of a conjugated diene, and the length and flexibility of the hydrocarbon moiety between the diene and the hydroxyl functional group. The length of the hydrocarbon chain distal from the diene moiety is limited to two carbons as in the natural pheromone, but a shorter version elicited as high activity in the receptor as bombykol. BmorOR1NBmorOrcoexpressing oocytes responded not only to bombykol, but also to bombykal. Addition of BmorPBP1 did not enhance selectivity, but dramatically reduced current responses thus suggesting that ligands are trapped. The requirements for a large hydrophobic cavity strongly suggest that the yet-to-be-identified binding site in BmorOR1 might be buried in the transmembrane domain.

Chemicals
Bombykol and bombykal were purchased from Plant Research International (Wageningen, The Netherlands) and kept sealed under helium at 280uC until use. For synthesis, solvents were dried by distillation over CaH 2 (benzene, dichloromethane) or sodium wire (tetrahydrofuran) or over dry potassium hydroxide (piperidine, pyrrolidine).

Chemical Analysis
Nuclear magnetic resonance spectroscopy was performed using a Bruker Avance 500 MHz instrument and deuteriochloroform as solvent. Mass spectra were recorded on a Mat95 XP magnetic sector mass spectrometer (Thermo Finnigan). Ionization was by electron impact at 70eV in positive ion mode with a source temperature of 220uC. Column chromatography was performed on silica gel (220-400 mesh, Fluka) and silica gel Merck 60 F 254 plates were used for TLC.

Receptor Cloning
Full-length BmorOR1 and BmorOrco gene sequences were amplified from constructs available from previous works in our laboratory [18,19]. They were transferred into pBlueScript by standard procedures and then subcloned into pGEMHE [20], and their sequences were confirmed by DNA sequencing (Davis Sequencing Center, Davis, CA).

In vitro Transcription Oocyte and Microinjection
In vitro transcription of cRNAs (BmorOR1 and BmorOrco) was performed by using a mMESSAGE mMACHINE T7 Kit (Ambion) according to the manufacturer's protocol. Plasmids were linearized with Nhe I, and capped cRNA was transcribed using T7 RNA polymerase. The cRNAs were purified with LiCl precipitation solution and re-suspended in nuclease-free water at a concentration of 200 ug/ml and stored at 280uC in aliquots. RNA concentrations were determined by UV spectrophotometry. cRNA were microinjected (2 ng of a receptor cRNA and 2 ng of an Orco cRNA) into Xenopus laevis oocytes on stage V or VI (EcoCyte Bioscience, Austin TX). The oocytes were then incubated at 18uC for 3-7 days in modified Barth's solution [in mM: 88 NaCl, 1 KCl, 2.4 NaHCO 3 , 0.82 MgSO 4 , 0.33 Ca(NO 3 ) 2 , 0.41 CaCl 2 , 10 HEPES, pH 7.4] supplemented with 10 mg/ml of gentamycin, 10 mg/ml of streptomycin and 1.8 mM sodium pyruvate.

Protein Expression and Purification
BmorPBP1 was prepared and purified, as previously described [14]. Lyophilized protein was dissolved in 1X Ringer's solution (see below) to make 2 mg/ml samples.

Sample Preparations and Electrophysiological Recordings
Stock solutions were prepared in dimethyl sulfoxide (DMSO) and stored at 220uC, if they could not be used immediately. An aliquot of each solution was taken, diluted with hexane, and analyzed by gas chromatography-mass spectrometry using analytical instrumentation, column and conditions previously described [14]. The oven was operated at 70uC, held at this initial temperature for 1 min, increased to 290uC at 10uC/min, and held at this final temperature for 10 min. Prior to electrophysiological measurements, stock solutions were brought to room temperature and diluted in 1X Ringer's solution [in mM: NaCl 96, KCl 2, CaCl 2 1.8, MgCl 2 1, HEPES 5, pH 7.6] containing 0.1% DMSO, except for preparation with BmorPBP1, which were diluted with the same buffer without DMSO. Two equal aliquots from the same initial solution (either bombykol or bombykal) were transferred to different vials from which decadic dilutions were made. One of the samples was diluted with Ringer-DMSO and the other was similarly diluted with Ringer-PBP. Thus, comparisons were made with samples derived from the same mother solution at the same concentration, but differing only in the solubilizer (solvent vs. PBP). All ligand solutions were freshly prepared and discharged if not used within 20 min. Chemical-induced currents were recorded with the two-electrode voltage-clamp technique at holding potential of 280mV. Signals were amplified with an OC-725C amplifier (Warner Instruments, Hamden, CT), lowpass filtered at 50 Hz and digitized at 1 kHz. Data acquisition and analysis were carried out with Digidata 1440A and software pCLAMP 10 (Molecular Devices, LLC, Sunnyvale, CA).