Journal of Chemical Ecology

, Volume 36, Issue 12, pp 1293–1305

Binding of the General Odorant Binding Protein of Bombyx mori BmorGOBP2 to the Moth Sex Pheromone Components

Authors

  • Xiaoli He
    • Department of Biological Chemistry, Rothamsted Research
  • George Tzotzos
    • Department of Chemistry and Analytical SciencesThe Open University
  • Christine Woodcock
    • Department of Biological Chemistry, Rothamsted Research
  • John A. Pickett
    • Department of Biological Chemistry, Rothamsted Research
  • Tony Hooper
    • Department of Biological Chemistry, Rothamsted Research
  • Linda M. Field
    • Department of Biological Chemistry, Rothamsted Research
    • Department of Biological Chemistry, Rothamsted Research
Article

DOI: 10.1007/s10886-010-9870-7

Cite this article as:
He, X., Tzotzos, G., Woodcock, C. et al. J Chem Ecol (2010) 36: 1293. doi:10.1007/s10886-010-9870-7

Abstract

Insects use olfactory cues to locate hosts and mates. Pheromones and other semiochemicals are transported in the insect antenna by odorant-binding proteins (OBPs), which ferry the signals across the sensillum lymph to the olfactory receptors (ORs). In the silkworm, Bombyx mori (L.), two OBP subfamilies, the pheromone-binding proteins (PBPs) and the general odorant-binding proteins (GOBPs), are thought to be involved in both sensing and transporting the sex pheromone, bombykol [(10E,12Z)-hexadecadien-1-ol], and host volatiles, respectively. Quantitative examination of transcript levels showed that BmorPBP1 and BmorGOBP2 are expressed specifically at very high levels in the antennae, consistent with their involvement in olfaction. A partitioning binding assay, along with other established assays, showed that both BmorPBP1 and BmorGOBP2 bind to the main sex pheromone component, bombykol. BmorPBP1 also binds equally well to the other major pheromone component, bombykal [(10E,12Z)-hexadecadienal], whereas BmorGOBP2 discriminates between the two ligands. The pheromone analogs (10E,12Z)-hexadecadienyl acetate and (10E,12Z)-octadecadien-1-ol bind to both OBPs more strongly than does bombykol, suggesting that they could act as potential blockers of the response to sex pheromone by the male. These results are supported by further comparative studies of molecular docking, crystallographic structures, and EAG recording as a measure of biological response.

Key Words

Binding assayBombykolElectroantennogram (EAG) recordingOdorant-binding protein (OBP)OlfactionPheromone- binding protein (PBP)Real-time polymerase chain reaction (PCR)SemiochemicalsSex pheromoneSilkworm moth

Introduction

Insects use their antennae to sense chemical signals (semiochemicals) that guide their interactions within and between species, including interactions with hosts such as plants and vertebrates. Odorant-binding proteins (OBPs) provide the initial molecular interactions for semiochemicals by ferrying the hydrophobic semiochemical molecules across the antennal sensillum lymph to the olfactory receptors (ORs) (Vogt and Riddiford, 1981; Van den Berg and Ziegelberger, 1991). In Lepidoptera, two subfamilies of OBPs, pheromone-binding proteins (PBPs) and general odorant-binding proteins (GOBPs) are involved in recognizing and transporting pheromones and host odorants, respectively (Breer et al., 1990; Robertson et al., 1999; Leal, 2003; Krieger et al., 2005; Pelosi et al., 2006; Kaissling, 2009). This classification is based mainly on the sequence and expression patterns in male and female antennae (Vogt and Riddiford, 1981; Vogt et al., 1991a, b; Krieger et al., 1991; Steinbrecht, 1998; Robertson et al., 1999). Sequence comparisons of these two OBP subfamilies indicate that they have evolved along different lineages (Pelosi et al., 2006; Gong et al., 2009; Zhou et al., 2009). However, the functional roles that GOBPs play in insect olfaction are not fully elucidated, and their binding to host odorants has not been demonstrated. Liu et al. (2010) reported that GOBPs of the navel orangeworm, Amyelois transitella (Walker) (Lepidoptera: Pyralidae), bind with high affinity to the major component of the sex pheromone, (11Z, 13Z)-16Ald.

In the silkworm, Bombyx mori (L.) (Lepidoptera: Bombycidae), 44 genes that encode putative OBPs have been identified from the genome sequence (Gong et al., 2009). The genes that encode pheromone binding protein 1 (BmorPBP1) and general odorant binding protein 2 (BmorGOBP2) are clustered close together on chromosome 19 in the same orientation. They have the same intron insertion sites between codons and splice sites within codons. The transcripts for BmorPBP1 and BmorGOBP2 are expressed at an extremely high level in adult antennae compared to other tissues (Zhou et al., 2009). The BmorGOBP2 transcript is higher than BmorPBP1 in the brains of both sexes (Gong et al., 2009). Immunocytochemical studies showed that both proteins are expressed in the antennae of males and females (Steinbrecht, 1998; Maida et al., 2005).

Pheromones are blends of volatile chemicals that typically are specific to a species, and, for B. mori, the pheromone blend comprises bombykol, bombykal, and (10E,12E)-hexadecadien-1-ol (Kaissling et al., 1978). However, only bombykol is able to induce mating behavior at a physiological concentration, and bombykal acts as an antagonist to bombykol (Kaissling, 2009). BmorPBP1 and its binding to pheromone components has been the subject of intense study in recent years, as a model system for understanding the functions that OBPs may play in insect olfaction in vivo. Cells that express BmorPBP1 are closely associated with the cells that express BmorOR-1 and BmorOR-3 in the long trichoid sensilla (Krieger et al.,2005; Nakagawa et al.,2005; Forstner et al., 2006), and co-expression of BmorPBP1 with BmorOR-1 in an ‘empty’ neuron of Drosophila melanogaster increased the sensitivity of the receptor to bombykol (Syed et al., 2006). BmorPBP1 also has been shown to mediate a response to bombykol, but not to bombykal in cultured HEK293 cells expressing BmorOR-1 (Grosse-Wilde et al., 2006) and in moth sensilla (Pophof, 2004). Recently, a highly sensitive receptor to bombykol has been identified in D. melanogaster (Syed et al., 2010). This raises the question on the selectivity and specificity that olfactory receptors alone contribute in insect olfaction.

Binding of BmorPBP1 to bombykol has been demonstrated by X-ray crystallography (Sandler et al., 2000); NMR structural characterization (Damberger et al., 2000; Horst et al., 2001; Lee et al., 2002); electrospray ionization-mass spectrometry (ESI-MS) (Oldham et al., 2000; Hooper et al., 2009); and other biochemical methods (Wojtasek and Leal 1999; Leal et al., 2005a, b). The binding is both pH and ligand dependent. Thus, at an acidic pH, or in the absence of bombykol, the C-terminus of the PBP forms an α-helix and occupies the ligand binding pocket, and a conformational change, brought about by a change in pH to more acid near the ORs, is thought to be the ligand release mechanism when the BmorPBP1-bombykol complex reaches the ORs (Wojtasek and Leal 1999; Leal et al., 2005a, b). However, there is some evidence that BmorPBP1 can bind to both bombykol and bombykal (Gräter et al., 2006; Zhou et al., 2009). A recent binding study that used a high-throughput ESI-MS analysis showed that BmorPBP1 bound much more strongly to (10E,12Z)-hexadecadienoic acid and (10,12)-hexadecadiyn-1-ol than to bombykol (Hooper et al., 2009).

BmorGOBP2 is expressed in larval sensilla (Laue, 2000) and in the long trichoid sensilla of female moths (Steinbrecht, 1998; Maida et al., 2005), which respond specifically to linalool and benzoic acid (Kaissling, 2009). However, the binding of BmorGOBP2 to these two plant volatiles has not been elucidated, whereas the binding of BmorGOBP2 to both bombykol and bombykal has been demonstrated. Furthermore, X-ray crystallography has shown that, when bombykol is bound to BmorGOBP2, a different conformation is adopted from that found when bombykol binds to BmorPBP1. In the former, a hydrogen bond is formed with Arg110 rather than with Ser56, as is the case with BmorPBP1. Furthermore, a second hydrogen bond is formed with Glu98, which leads to an overall stronger affinity of BmorGOBP2 for bombykol (Zhou et al., 2009).

Ligand specificity and selectivity, and, in fact, the functional roles of insect OBPs in olfaction, have been challenged consistently, even more so since insect olfactory receptors have been characterized (Krieger et al., 2005). Zhou et al. (2009) showed differential binding of BmorGOBP2 to bombykol and bombykal by X-ray crystallography. In the present study, we provide further evidence of the possible involvement of the BmorGOBP2 gene in moth olfaction by examining the tissue expression patterns throughout the developmental stages of B. mori, and through a comparative binding study of recombinant BmorGOBP2 to plant volatiles, as well as to the sex pheromone components. We employed molecular ligand docking to predict the binding interactions of BmorPBP1 and BmorGOBP2 with the ligands. Two synthetic pheromone analogs were found to bind better than the sex pheromone components. Electroantennogram (EAG) recordings from the antennae of male moths were made with the analogs to investigate how the binding data and the molecular docking predictions might relate to biological significance.

Methods and Materials

Real-time PCR Bombyx mori cocoons were provided by Prof. Y. –P. Haung (Institute of Plant Physiology and Ecology, CAS, China), and adult moths were dissected, immediately after eclosion, into antennae, heads (without antennae), and bodies. Heads and bodies also were collected from 2nd, 3 rd, 4th, and 5th instar larvae. Detailed protocols for total RNA extraction with the RNAqueous kit (Ambion, Huntingdon, UK) and two step real-time PCR on each tissue with the ETROscript kit (Ambion) on an ABI 7500 (Applied Biosystems, Foster City, CA, USA) have been described previously (Zhou et al., 2009). The sequences of BmorGOBP2 (GenBank No. NP_001037498) and BmorPBP1 (GenBank No. NP_001037494) were downloaded from NCBI GenBank (http://www.ncbi.nlm.nih.gov/). The B. mori reference gene, actin A4 (GenBank No. U49644), was used in each real-time PCR experiment to check loading, reverse transcriptase efficiency, and the integrity of the transcripts. For each tissue, real-time PCR analyses were conducted with two tissue preparations. The reactions were hot-started for 2 min at 95°C, followed by 40 cycles of 95°C for 15 sec, 60°C for 30 sec, and 72°C for 15 sec. The PCR primers (Table 1) were designed with Primer3 (http://frodo.wi.mit.edu/).
Table 1

The PCR primers used in real-time PCR for quantification of transcript levels in the silkworm, Bombyx mori

OBP gene

Forward primer (5′-3′)

Reverse primer (5′-3′)

BmorGOBP2

ATCATATGACCGCCGAGGTGATGAGCCACG

GGAATTCTCAGTATTTTTCGATAACTGCTT

BmorPBP1

CAGTGGATGCGTCTCAAGAA

GTCTCATCGGCTCCATGTTT

Actin

CGTTCGTGACATCAAGGAGA

ACAGGTCCTTACGGATGTCG

Relative tissue expression was quantified as described in Zhou et al. (2009). Briefly, after PCR, Ct values were exported into the LinRegPCR program to correct the amplification efficiencies for each reaction. The relative expression levels (Pfaffl ratio) of each OBP gene to the reference gene then were calculated in each tissue from [(E,obpΔCt,obp]/[(E,refΔCt,ref], where E,obp and E,ref are the corrected amplification efficiencies for OBP and the reference gene, respectively. ΔCt,obp is calculated from [Ct,obp of heads or antennae—Ct,obp of body], and ΔCt,ref is calculated from [Ct,ref of heads or antennae—Ct,ref of body]. The results are presented as the mean fold change of two biological samples combined from both male and female tissues.

Protein Expression and Purification of Recombinant OBPs

Full-length cDNAs that encode mature BmorGOBP2 and BmorPBP1 were cloned into the bacterial expression vector pET17b (Novagen, Darmstadt, Germany) between the NdeI and EcoRI restriction sites, and verified by sequencing. Plasmids with the correct inserts were transformed into BL21(DE3)pLysS E. coli cells, and protein synthesis was induced at OD600 of 0.5–0.8 with IPTG (4 mM) for 3 h. All proteins were found to be expressed as inclusion bodies, and solubilization was performed by denaturation in urea/DTT, renaturation, and extensive dialysis, by using a protocol applied successfully to other OBPs. The recombinant proteins were purified by two rounds of anion-exchange chromatography with a HiPrep 16/40 column (GE Healthcare, Hatfield, UK) filled with DE-52 resin (Whatman, Kent, UK), followed by gel filtration on a Sephacryl S-200 HiPrep 26/60 column (GE Healthcare). A MonoQ column also was used at the final stage of purification. The purified proteins were stored at −20°C in 20 mM Tris-HCl pH 7.4 buffer.

Chemical Synthesis

The seven analogs of bombykol, including bombykal, were synthesized in-house at Rothamsted Research. Nuclear magnetic resonance spectroscopy (NMR) and gas chromatography with flame ionization detection (GC-FID) analyses confirmed the structure and purity of the synthesized compounds. 1 H- and 13 C-NMR spectroscopy were performed by using a Bruker 500 Advance NMR spectrometer with 1 H referenced to CDCl3 (7.25 ppm) and 13 C to CDCl3 (77.0 ppm). Purified compounds were analyzed on a Hewlett-Packard 5890 Series II gas chromatograph (GC), fitted with a nonpolar HP-1 capillary column (40 m, 0.32 mm i.d., 0.52 μm film thickness), a cool-on-column injector, and a flame ionization detector (FID) (Agilent Technologies, West Lothian, UK). The GC oven temperature was maintained at 40°C for 1 min and then raised by 10°C /min to 200°C, and the carrier gas was nitrogen at 80 cm/sec. Benzoic acid (99%), (±)-linalool (97%), and N-phenyl-1-naphthylamine (NPN) (98%) were purchased from Sigma-Aldrich (Dorset, UK).

Fluorescence Competitive Binding Assay

To measure the binding of the fluorescent probe NPN to OBPs, a 2 μM protein solution (1 ml) in 20 mM Tris-HCl, pH 7.4 and 4.5, was titrated with aliquots of 1 mM NPN dissolved in methanol to final concentrations of 0.05–16 μM. The protein/NPN complex was excited at 337 nm, and emission spectra were recorded between 300 and 450 nm on a luminescence spectrometer LS50B (Perkin-Elmer, Cambridge, UK) at 25°C in a right angle configuration with a 1 cm light path quartz cuvette and 5 nm slits for both excitation and emission. The competitive binding of ligands was measured by using NPN (4 μM) as the fluorescent reporter and 0.05–7 μM concentrations of each ligand dissolved in methanol, which gave molar ratios from 0.01 to 1.61 (bombykol:NPN). The total volume of methanol in the protein solution was maintained at 21 μl (2.1%) by adding methanol without ligands. Bound ligand was evaluated from the values of fluorescence intensity assuming that the protein was 100% active, with a stoichiometry of 1:1 protein:ligand at saturation.

Partitioning Binding Assay

In the partitioning binding assay (Danty et al., 1999), the protein concentration was calculated from the OD280 values and the extinction coefficient of the protein solution assuming all pairs of Cys residues form disulfide bridges. In practice, 20 μl of protein (0.5 mM) in 20 mM Tris buffer (pH 7.4) were added to the 100-μl v-shaped vial (Wheaton Scientific, Millville, NJ, USA), and then 20 μl of hexane containing a mixture of ligands (14 μM each), including bombykol, were layered on top [higher concentrations up to 10 mM of protein were used and similar results obtained (data not shown)]. The two phases were mixed gently, centrifuged at 13,000 × g for 5 min, and then incubated at room temperature for at least 1 h. After incubation, 2 μl of the top phase containing the ligands were injected into the GC, and the amount of ligand that had gone into the protein phase was determined from the GC trace by using Equations 1 and 2 (below), and through comparison with the results obtained before incubation and without protein.

Each experiment was performed with at least three replicates and repeated at least twice for each sample. The binding of ligands to the OBPs was quantified by GC-FID on an HP Agilent 6890 Series GC system (Hewlett Packard, Wokingham, Berkshire, UK) with a wax column (HP-5MS, 25 m, 0.25 mm i.d., 0.25 μm film thickness; Agilent Technologies) and a cool-on-column injector, and a flame ionization detector (FID). The system was operated under the following temperature program: 100°C for 1 min, increased to 250°C at a rate of 10°C/min, and held at the final temperature for 10 min. The carrier gas was nitrogen at a linear flow rate of 80 cm/sec. The amount of each ligand in the hexane fraction was calculated from the chromatograms by using a single point internal standard (IS). The internal standard used was tridecane (99% chemical purity) (Sigma-Aldrich, Dorset, UK). The first analysis contained a known amount of internal standard and the compounds of interest, and the binding of the ligand was calculated from:
$$ {\hbox{Internal}}\,{\hbox{Response}}\,{\hbox{Factor}}\left( {\hbox{IRF}} \right) = {{{\left[ {{\hbox{areaIS}} \times {\hbox{amountSC}}} \right]}} \left/ {{\left[ {{\hbox{amountIS}} \times {\hbox{areaSC}}} \right]}} \right.}, $$
(1)
$$ {\hbox{amount}}\,{\hbox{of}}\,{\hbox{specific}}\,{\hbox{compund}} = {{{\left[ {{\hbox{amountIS}} \times {\hbox{IRF}}} \right]}} \left/ {{{\hbox{areaIS}},}} \right.} $$
(2)
Where, IS is the internal standard and SC is the specific compound of interest.

Docking Experimental Procedure

Flexible docking of the ligands to BmorPBP1 and BmorGOBP2 was performed by using a genetic search algorithm and a semi-empirical force field of the AutoDock (v.4.2) program (Morris et al., 2009). The crystallographic structures of BmorPBP1 (1DQE) (Sandler et al., 2000) and of BmorGOBP2 (2WC6) (Zhou et al., 2009) were used. Water molecules were removed, and polar hydrogens were added to the proteins and ligands. For the ligands, all torsions were released except the ones around the conjugated double and triple bonds. The default AutoDock force field was applied (Huey et al., 2007). The whole proteins were covered by grid maps with a spacing of 0.374 Å. For the genetic algorithm, default parameters were used starting from random positions and orientation of the ligands. Each ligand was subjected to 100 Lamarkian genetic algorithm runs, with 25x106 evaluations in each and with the rest of the parameters set to the default values of the AutoDock GUI, AutoDock Tools. The root mean square deviation (rmsd) tolerance of the resulting docked structures was ≤ 2 Å. The binding pocket topologies (area and volume) were calculated by using the CASTp server (http://sts.bioengr.uic.edu/castp/calculation.php) (Dunda et al., 2006).

Electrophysiology

Insects for electrophysiological studies either were obtained as pupae from Prof. Y. –P. Haung (see above) or purchased from Warwick Insect Technologies (Coventry, UK). Pupae were sexed and kept at 20°C, 16:8 h L:D, until adults emerged. Electroantennogram (EAG) recordings were made by using Ag-AgCl glass electrodes filled with saline solution [composition as in Maddrell (1969), but without glucose]. An antenna of an adult male adult B. mori was excised and suspended between the two electrodes. The tip of the terminal process of the antenna was removed to ensure a good contact. Signals were passed through a high impedance amplifier (UN-06, Syntech, Hilversum, The Netherlands) and analyzed by using a customized software package (Syntech).

The stimulus delivery system, which employed a filter paper in a disposable glass Pasteur pipette cartridge, has been described (Wadhams et al., 1982). The stimulus (2 sec duration) was delivered into a purified airstream (1 l/min) flowing continuously over the preparation. Standard solutions (1 mg/ml) of test compounds were applied (10 μl) to filter paper strips, and the solvent was allowed to evaporate (30 sec) before the strip was placed in the cartridge. The control stimulus was hexane (10 μl). Fresh cartridges were prepared immediately prior to each stimulation (N = 5). Responses to control and test solutions were compared for significant differences by using Student’s t-test.

Results

Expression of Bombyx mori OBPs

The levels of transcripts of BmorGOBP2 and BmorPBP1 were very low in heads at all larval stages and in adult heads without antennae (Fig. 1). They were very high in the adult antennae, indicating that both genes are highly up-regulated in adult antennae and could be involved in olfaction and detection of semiochemicals. These results are consistent with high levels of the moth proteins found in antennae by Western blot analysis and immunocytochemical labeling experiments (Steinbrecht, 1998; Maida et al., 2005), and also with the developmental profile of OBP Aaeg-OBP10 mRNA isolated from the intact yellowfever mosquito, Aedes aegypti (L.) (Diptera: Culicidae) (Bohbot and Vogt, 2005).
https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig1_HTML.gif
Fig. 1

Transcript levels of Bombyx mori OBP genes BmorPBP1 and BmorGOBP2. The transcript levels in adult antennae (mixed sexes) and heads at different larval developmental stages were measured by real-time PCR (see Methods and Materials). The fold-changes are normalized with the actin gene and relative to the transcript level in the body of each stage. The values are the mean of two biological samples and three replicates of each sample

The B. mori transcripts also were observed in the earlier stages. We cannot exclude the possibility that two OBPs may have a different function in the larva than in the adult. Larvae are the stage that feeds on plant materials, and they may use this OBP to detect plant volatiles, whereas adults are the stage that reproduces and disperses, and no longer needs to feed. Thus, they may use this OBP to detect pheromone components.

Fluorescence Competitive Binding Assay

In the binding assay with BmorGOBP2, attempts to displace NPN by the sex pheromone component bombykol and the plant volatiles [(±)-linalool and benzoic acid] showed that bombykol gave the most displacement, whereas the other two ligands showed much less displacement (Fig. 2). The binding of bombykol to BmorGOBP2 also was pH-dependent, as demonstrated for BmorPBP1 (Leal et al., 2005b), with displacement of NPN greatest at pH 7.5, but displacement of NPN greatly decreased at pH 4.5 over a range of concentrations (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig2_HTML.gif
Fig. 2

Fluorescence displacement of N-phenyl-1-naphthylamine (NPN) from Bombyx mori BmorGOBP2 by bombykol (♦); benzoic acid (X); and (±)-linalool (▲). The protein and NPN were both at 2 μM. The protein/NPN complexes were excited at 337 nm and the fluorescence emission at 395 nm was measured before and upon titration with methanol solutions of the competitors to final concentrations of 0.5–7 μM. The reductions in NPN fluorescence emission at 395 nm were normalized to the NPN fluorescence before titration. The decrease in NPN fluorescence intensity at the emission maximum (395–400 nm), at increasing concentrations of competitors, is presented as (Fligand/Fnpn) with Fligand=the peak fluorescence intensity at a distinct competitor concentration. Fnpn=the peak fluorescence intensity of the OBP/NPN complex

https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig3_HTML.gif
Fig. 3

Fluorescence displacement of NPN (2 μM) by bombykol from Bombyx mori BmorGOBP2 (2 μM) at different pHs using a mean of three or four replicates with standard deviations

We next measured the displacement of NPN from recombinant BmorGOBP2 and BmorPBP1 by bombykol and bombykal (Fig. 4). For BmorPBP1, there was no difference in binding between bombykol and bombykal, with the reductions in NPN fluorescence of 41.1 ± 19.4% (N = 30) and 34.3 ± 14.2% (N = 12), respectively, at 6 μM concentration. Interestingly, BmorGOBP2 showed a similar binding to bombykol as BmorPBP1, but BmorGOBP2 unexpectedly showed a difference in binding of bombykol from bombykal 41.1 ± 15.1% (N = 8) and 15.0 ± 3.7% (N = 3) (P < 0.01, d.f.=10) at 6 μM concentration, respectively. However, the displacement at lower concentrations was low (10–20%). This could be due to non-specific binding and could be affected by other compounds that might bind during protein expression in E. coli cells (Oldham et al., 2000).
https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig4_HTML.gif
Fig. 4

Displacement of the fluorescent probe NPN by bombykol and bombykal. The protein and NPN were both incubated at a concentration of 2 μM. The OBP/NPN complexes were excited at 337 nm and the fluorescence emission at 395 nm was measured before and upon titration with bombykol or bombykal at a concentration of from 1–6 μM. The reductions in NPN fluorescence emission at 395 nm were normalized to the NPN fluorescence before titration. Each point is the average of at least two replicates (N) with the standard error (N = 30 for BmorPBP1/bombykol, N = 12 for BmorPBP1/bombykal, N = 8 for BmorGOBP2/bombykol, N = 3 for BmorGOBP2/bombkal)

Partitioning Binding Assays

We deployed a partitioning binding assay (Danty et al., 1999) to avoid using the fluorescent probe and the possibility of non-specific displacement. In this assay, the protein in buffer is incubated with a mixture of ligands dissolved in hexane in two phases, (see Methods and Materials), so that each compound competes for binding. The depletion of the compound from the hexane phase by the protein in the buffer can be measured by GC-FID analysis under equilibrium condition as compared with a control (with no protein). The advantage of this assay is that a mix of ligands (Fig. 5) can be used, and the high concentration of protein in the buffer resembles the high concentration of OBP within the antennal lymph. Both BmorGOBP2 and BmorPBP1 bound to bombykol and most of its analogs, except those with their double bonds replaced by triple bonds (Figs. 6, 7 and 8). Again, there was no difference between the binding of bombykol and bombykal to BmorPBP1, which is consistent with the fluorescence displacement binding results (Fig. 4) and with those of others (Gräter et al., 2006). For BmorGOBP2, bombykol binding was slightly better than bombykal, indicating the effect of changing of the functional group from an alcohol in bombykol to an aldehyde in bombykal to form two hydrogen bonds in BmorGOBP2 (Fig. 7) (Zhou et al., 2009). Interestingly, (10E,12Z)-hexadecadienyl acetate (3) bound best, and (10E,12Z)-octadecadien-1-ol (4) and bombykol (1) bound equally well. Moving the position of the double bonds of the pheromone (6) reduced the binding to the proteins, and changing from a double bond in bombykol to a triple bond in (10E)-hexadecen-12-yn-1-ol (7) and (10,12)-hexadecadiyn-1-ol (8) further reduced the binding to both BmorGOBP2 and BmorPBP1. Overall, BmorGOBP2 displayed better differential binding to the analogs than did BmorPBP1.
https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig5_HTML.gif
Fig. 5

Structures of the Bombyx mori sex pheromone components and their analogs used in this study

https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig6_HTML.gif
Fig. 6

Binding of bombykol and its analogs to Bombyx mori odorant-binding proteins (OBPs). The ligand binding (ng compound per μg protein) was measured with the partitioning binding assay (see Methods and Materials) as the depletion of compound from the top hexane phase into the bottom protein phase relative to that without protein. Each compound is represented with its code as shown in Fig. 5. Each column is the mean of at least three replicates (N) and the bar denotes the standard error (N = 3 for OBP/compound 3, N = 3 for OBP/compound 4, N = 7 for OBP/bombykol, N = 4 for OBP/bombykal, N = 4 for OBP/compound 5, N = 4 for OBP/compound 6, N = 4 for OBP/compound 7, N = 3 for OBP/compound 8)

https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig7_HTML.gif
Fig. 7

Molecular contacts of bombykol in the binding pocket of Bombyx mori BmorGOBP2 as predicted by molecular docking. The figures were drawn with AutoDockTools (ADT) Python (Sanner, 1999). The spheres and the cloud around the docked bombykol molecule (green stick) are the electrostatic potential (red=oxygen, blue=nitrogen, grey=carbon) for van der Waals (vdWs) contacts and vdW’s volume over the whole molecule, respectively. The hydroxyl group of the docked bombykol molecule, with the lowest binding energy, forms hydrogen bonds with Arg110 and Val66 (a) and Glu98 (b) as in crystal structures. Ser56 forms no vdWs interactions with bombykol

https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig8_HTML.gif
Fig. 8

Three-dimensional structure of Bombyx mori BmorPBP1 and BmorGOBP2 bound with the sex pheromone component (10E,12Z)-hexadecadien-1-ol (stick). The C-termini and the key amino acid residues, which form hydrogen bonds, are labeled. The BmorPBP1 structure 1DQE (Sandler et al., 2000) and BmorGOBP2 structure 2WC6 (Zhou et al., 2009) are downloaded from the protein data bank (http://www.rcsb.org/pdb/home/home.do) and displayed with Swiss-Pdb Viewer 3.7

Molecular Docking

The previously published X-ray crystallography structures of BmorPBP1 (Sandler et al., 2000) and BmorGOBP2 (Zhou et al., 2009) provide a unique opportunity for molecular modeling and ligand docking. AutoDock 4.2 was used to predict blind docking of bombykol and the analogs to BmorPBP1 (1DQE) and BmorGOBP2 (2WCK), and to estimate the free energy of binding of the protein-ligand complexes. The docking algorithm successfully placed the analogs into the same binding cavity as those of the models obtained from crystal structures. Each ligand was subjected to 100 genetic algorithm runs (see Docking Experimental Procedures in the Methods and Materials). AutoDock performs cluster analysis or “structure binning” based on all-atom mean square deviation (RMSD). The resulting families of docked conformations are ranked in the order of increasing energy (rank 1 is the lowest energy cluster) (Table 2). In very few cases, less populated clusters were observed with lower energy. However, these were not selected as representative of the predicted bound ligand structure as their mean energy difference was 2.5 kcal/mol less than that of the most populated cluster. This is within the range of error of the AutoDock force field (±2.5 kcal/mol). In these cases, the largest cluster was taken as the predicted bound ligand structure (Table 2). Molecular docking showed that in the conformation of bombykol with the lowest binding energy, there are two hydrogen bonds with Glu98 and Arg110 (Fig. 7). The energy minima reached for the BmorGOBP2/ligand complexes are considerably lower than the corresponding ones for BmorPBP1 (Table 2), with most of the lower binding energy conformations having their hydroxyl group oriented towards Arg110 and away from Ser56 without any van der Waals interaction between Ser56 and bombykol (Fig. 7), which is consistent with the crystal structures of BmorGOBP2/ligand complexes (Fig. 8) (Zhou et al., 2009). Furthermore: 1) The docked ligand conformations show tighter clustering (fewer clusters) for BmorGOBP2 than for BmorPBP1. The lowest energy clusters are invariably more populated for all ligands (higher N values) for BmorGOBP2 than for BmorPBP1. This may be explained by a smaller binding cavity in BmorGOBP2 (volume = 616.5 Å3) as compared to that of BmorPBP1 (volume = 888.8 Å3) that may restrict the ligand(s) from attaining multiple conformations within the binding pocket; 2) The ligands rotate more freely in BmorPBP1 than in BmorGOBP2, as shown by greater RMSD differences between the best binding conformers; 3) Both bombykol and bombykal reach the same energy minimum (−7.87) with BmorGOBP2, showing that they have a similar binding affinity. These data provide further evidence and support the binding results (Figs. 2, 4 and 6) in showing that BmorGOBP2 can bind sex pheromones and discriminate between the pheromone analogues better than can BmorPBP1.
Table 2

Ligand binding energies for the Bombyx mori odorant-binding proteins Bmorgobp2 and Bmorpbp1 calculated from molecular modeling

  

BmorGOBP2 (2wc6) (kcal/mol)

BmorPBP1 (1dqe) (kcal/mol)

Ligand No

Compound

Clustera

Rank

N

Emin

Emean

Cluster

Rank

N

Emin

Emean

4

(10E,12Z)-Octadecadien-1-ol

3

1

86

−8.48

−7.92

11

1

32

−7.10

−6.62

3

(10E,12Z)-Hexadecadienyl acetate

5

1

48

−8.24

−7.70

7

1

32

−7.33

−6.85

6

(8E,10Z)-Hexadecadien-1-ol

9

1

80

−7.90

−7.30

17

1

28

−6.69

−6.14

1

Bombykol

6

1

88

−7.87

−7.28

12

1

35

−6.83

−6.29

2

Bombykal

4

1

83

−7.87

−7.48

11

1b

20

−6.87

−6.55

7

(10E)-Hexadecen-12-yn-1-ol

6

1

67

−7.65

−7.29

15

1c

14

−6.62

−6.12

8

(10,12)-Hexadecadiyn-1-ol

9

1

44

−7.44

−7.01

17

1d

2

−6.54

−6.27

5

(10E,12Z)-Tetradecadien-1-ol

5

1

77

−7.22

−6.84

9

1

74

−6.43

6.09

aCluster is the number of distinct multi-member conformational clusters found out of 100 runs. N is the number of conformations in the cluster. Emin is the free energy of binding of the most favorable conformation within a given cluster; Emean is the cluster’s mean energy (See Methods and Materials)

bMost populated cluster is Rank 4 comprising of 33 conformations

cMost populated cluster is Rank 5 comprising of 17 conformations

dMost populated cluster is Rank 12 comprising of 18 conformations

The analog (10E,12Z)-hexadecadienyl acetate (3) has a lower Emin value (i.e., the free energy of binding of the most favorable conformation within a given cluster). Thus, the analog is predicted to bind better than bombykol or bombykal to both BmorGOBP2 and BmorPBP1, consistent with the results obtained with the partitioning binding assays (Fig. 6) and by the ESI-MS analysis (Hooper et al., 2009). This indicates that the functional group of the analog may interact with Glu98 and Arg110 to form two hydrogen bonds as observed in the crystallographic structures of BmorGOBP2/bombykol (Zhou et al., 2009), and that it, together with bombykol, are better ligands than bombykal. The molecular docking also predicts the lowest binding energy for (10E,12Z)-octadecadien-1-ol (4)/OBP complexes (Table 2).

Electrophysiological Responses

Electroantennogram (EAG) recordings showed that bombykol (1) and bombykal (2) are the most active in triggering an antennal response (Table 3, Fig. 9). (10E,12Z)-Hexadecadienyl acetate (3) and (10E,12Z)-octadecadien-1-ol (4) were shown, for the first time, to elicit significant responses from the male B. mori antenna (P < 0.01 and P < 0.05, respectively), whereas changing double bonds to triple bonds (7) and (8) reduced the EAG responses, respectively. This was consistent with the partitioning binding results (Fig. 6), as well as the ligand docking predictions (Table 2). The biological activity thus correlates with the binding data and molecular docking prediction for these compounds. Shortening the length of the alkyl carbon chain by two carbons (5) or moving two double bonds (6) further reduced the EAG activity as demonstrated by earlier work (Kaissling et al., 1978).
Table 3

Electroantennographic response (± standard error) of male Bombyx mori antennae to semiochemical or analog ligands

Ligand No.

Compounda

Response (%)b

Significance (P)c

2

Bombykal

307 ± 51

<0.01

1

Bombykol

234 ± 43

<0.01

3

(10E,12Z)-Hexadecadienyl acetate

198 ± 29

<0.01

4

(10E,12Z)-Octadecadien-1-ol

151 ± 11

<0.05

7

(10E)-Hexadecen-12-yn-1-ol

125 ± 8

<0.05

5

(10E,12Z)-Tetradecadien-1-ol

110 ± 6

Not sig.

6

(8E,10Z)-Hexadecadien-1-ol

105 ± 5

Not sig.

8

(10,12)-Hexadecadiyn-1-ol

67 ± 13

Not sig.

aLigands presented at 1 mg/ml

bResponses expressed as % of an artificial 0.1 mV signal (i.e., 0.01 mV = 100%)

cP = significance of difference from hexane control (N = 5)

https://static-content.springer.com/image/art%3A10.1007%2Fs10886-010-9870-7/MediaObjects/10886_2010_9870_Fig9_HTML.gif
Fig. 9

Electroantennogram (EAG) recordings from male Bombyx mori antennae. The stimulus (2 sec duration) was delivered into a purified airstream (1 l/min) flowing continuously over the preparation. The horizontal lines indicate a recording time of 15 sec in each case. Standard solutions (1 mg/ml) of test compounds were applied (10 μl) to filter paper strips and the solvent was allowed to evaporate (30 sec) before the strip was placed in a glass cartridge. The control stimulus was hexane (10 μl). Fresh glass cartridges were prepared immediately prior to each stimulation. Responses to control and test solutions were compared for significant differences by using Student's t-test. N = 5

Discussion

Our understanding of the molecular and biochemical mechanisms that mediate chemoreception in insects has been improved greatly by the discovery of olfactory and taste receptor proteins. However, 50 years after the discovery of the first insect sex pheromone from B. mori, it is still unclear how such hydrophobic compounds selectively reach the dendrites of sensory neurons in vivo across aqueous space to interact with the sensory receptors. The presence of soluble polypeptides in high concentration in the lymph of chemosensilla still poses unanswered questions. More than two decades after their discovery and despite the wealth of structural and biochemical information available, the physiological functions of odorant-binding proteins (OBPs) are still not well understood. It is possible that the combinatorial actions by different OBPs and ORs may provide the foundation for insects to discriminate various odorants (Pophof, 2004; Krieger et al., 2005; Xu et al., 2005; Syed et al., 2006; Laughlin et al., 2008). The transcripts of BmorGOBP2 and BmorPBP1 were low in the larval stages, but very high in adult antennae, suggesting a role in adult moth olfaction related to mating and reproduction (adult moths do not feed). When the corresponding recombinant proteins were studied in partitioning binding assays, BmorPBP1 bound to the sex pheromone bombykol as well as its analogs (see also Sandler et al., 2000; Horst et al., 2001; Oldham et al., 2000; Hooper et al., 2009; Zhou et al., 2009). However, there was evidence from the binding assays that BmorGOBP2 can also bind the sex pheromone components bombykol and bombykal (Fig. 6; Gräter et al., 2006; Zhou et al., 2009). It is interesting that both BmorPBP1 and BmorGOBP2 bind to the sex pheromone components, suggesting that the specificity of binding to insect OBPs does not determine the specificity of the receptor neuron response (Kaissling, 2009). This prompted further investigations into the specific binding of BmorGOBP2 to other semiochemical analogs.

Both the fluorescence competition and the partitioning binding assays showed that BmorGOBP2 bound to bombykol more strongly than to bombykal, which is consistent with the results obtained by a cold binding assay (Zhou et al., 2009). The structural evidence indicates that the lack of a hydrogen bond to Glu98 in the BmorGOBP2/bombykal complex may be responsible for this difference (Fig. 8) (Zhou et al., 2009). On the other hand, although BmorGOBP2/ligand complexes have lower energy minimums than BmorPBP1/ligand complexes, the docking experiments failed to provide evidence for the discriminatory binding of BmorGOBP2 between bombykol and bombykal (Table 2) as suggested previously (Zhou et al., 2009). It can be argued that BmorPBP1 could have higher affinity for bombykol than bombykal because of the difference in the energy minimum between BmorPBP1/bombykol and BmorPBP1/bombykal complexes (−6.83 kcal/mol vs −6.40 kcal/mol) (Table 2).

The specific expression of BmorGOBP2 in the female sensilla has been demonstrated (Steinbrecht, 1998; Maida et al., 2005), but neither binding to any semiochemicals nor the physiological role of BmorGOBP2 in B. mori has been reported. The specific and high up-regulation of the expression of BmorGOBP2 gene in adult antennae suggests that BmorGOBP2 may have a role in the perception of semiochemicals, most likely for plant volatiles. However, our data support the view that BmorGOBP2 can be used preferentially to capture the sex pheromone rather than common plant volatiles such as linalool and benzoic acid. Previous experiments that used an antibody raised against BmorGOBP2 have shown that only a small fraction of sensilla in male antennae contain BmorGOBP2, whereas only few sensilla of female antennae express BmorPBP1 (Steinbrecht, 1998; Maida et al., 2005). The perception of bombykol and bombykal occurs via male receptor neurons within sensilla containing only BmorPBP1. Furthermore, the female receptor neurons innervating the abundant long sensilla trichodea that contain only BmorGOBP2 are very sensitive to linalool and benzoic acid, and never have been found to respond to the pheromones (Kaissling, 2009). These results argue that there is a physiological role of BmorGOBP2 in sex pheromone perception in vivo. However, the moth GOBP2 also has been found expressed in male moth antennae (Vogt et al., 1991a) and co-expressed with PBPs (Nardi et al., 2003). The dramatic up-regulation of BmorGOBP2 gene in adult antennae and the binding to sex pheromone components of BmorGOBP2, as reported in this study and others (Gräter et al., 2006), warrant further investigation.

The binding data of BmorGOBP2 also suggest that the protonation of OBPs plays an important role in ligand binding and release, and support the mechanism of pH-dependent release of ligands as demonstrated previously for BmorPBP1 (Wojtasek and Leal 1999; Damberger et al., 2000; Horst et al., 2001; Kowcun et al., 2001; Lee et al., 2002; Leal, 2003) and other insect OBPs. For the honey bee, Apis mellifera L. (Hymenoptera: Apidae), PBP AmelASP1, ligand binding is activated at low pH and is dependent on protonation of Asp35, which locks down the C-terminus in a confirmation that closes the active site by a hydrogen bond to the main chain of Val118 (Pesenti et al., 2008). In the predicted BmorPBP1/bombykol structure, the C-terminus blocks the rear entry to the pocket, with Val135 coming within 4 Å of bombykol (Sandler et al., 2000; Gräter et al., 2006). However, with BmorGOBP2, the blocking of the rear pocket by the bulge formed by amino acid 33-35 means that the C-terminus cannot sense the ligand directly and plays a role in ligand release (Zhou et al., 2009). The involvement of the C-terminus in BmorGOBP2 binding remains to be determined.

We used synthetic pheromone analogs to test the specificity in ligand binding of the two moth OBPs. (10E,12Z)-Hexadecadienyl acetate (3) showed better binding to BmorGOBP2 than did either bombykol or bombykal in the partitioning binding assay, consistent with the ligand docking experiments (Table 2). The ligand docking experiments also predict that BmorGOBP2 forms more energetically favorable complexes with this analog than does BmorPBP1, and that (10E,12Z)-octadecadien-1-ol (4) should have a good binding to both OBPs as supported by our binding experiments (Fig. 6) and to BmorPBP1 by previous ESI-MS analysis (Hooper et al., 2009). Our binding data for (10E,12Z)-tetradecadien-1-ol (5) is contradictory to the prediction of the docking results, showing good binding to BmorPBP1 and BmorGOBP2 (Fig. 6), but consistent with the poor binding obtained with a cold binding assay (Zhou et al., 2009). Our binding data confirmed the reduced binding of (8E,10Z)-hexadecadien-1-ol (6) compared with the sex pheromones, as measured by the ESI-MS analysis, but contradicted the better binding predicted by ligand docking (Table 2) and the cold binding assay (Zhou et al., 2009). The replacement of double bonds by triple bonds (7, 8) significantly reduced the binding for both OBPs. The molecular docking also placed them as less favorable ligands for both OBPs. However, the ESI-MS analysis showed that (10,12)-hexadecadiyn-1-ol (8) bound twice as much as bombykol. In agreement with Hooper et al. (2009), bombykal did not bind to BmorPBP1. The negative values of the binding data to compound (7) and (8) are unexpected. Other compounds that bound to the OBPs during expression in E. coli (Oldham et al., 2000) were displaced by the ligands and measured by GC together with (7) and (8). Because of these surprising results, a measure of the biological activity was sought by recording overall EAG responses for the sex pheromones as well as the analogs (Table 3, Fig. 9). As expected, the EAG responses were significant for the pheromone components, bombykol (1) and bombykal (2). The monoyne analogue (7) elicited a small EAG response but the diyne analog (8) elicited no EAG response. The EAG activity relates well, not only to the results obtained by the partitioning binding assay, which showed significantly increased binding of (10E,12Z)-hexadecadienyl acetate (3) and (10E,12Z)-octadecadien-1ol (4), but also to the ligand docking results (Table 2) and to the ESI-MS analysis (Hooper et al., 2009). Unlike the data from in vitro binding and molecular docking, the higher EAG activity of sex pheromone components (1) and (2) over activity of the analogs (3) and (4) correlate better with the sexual behavior of adult male moths. Further behavioral studies are needed to establish any physiological relevance of the analogs (3) and (4) in moth olfaction. Single sensillum recordings from sensillum types with known protein content could investigate precisely the effects of their binding to protein in vivo.

We previously reported the X-ray crystallography structures of BmorGOBP2 complexed with bombykol and four of its analogs that showed that the mode of bombykol binding to BmorGOBP2 differed from that found for BmorPBP1. In the BmorGOBP2 complex, the hydroxyl hydrogen of bombykol forms a hydrogen bond to Arg110, and the hydroxyl group of bombykol also is involved in forming a second hydrogen bond with Glu98, whereas in BmorPBP1, one hydrogen bond is formed with Ser56 (Fig. 8). This could explain why (10E,12Z)-hexadecadienyl acetate (3) appears to bind more strongly to BmorGOBP2 than other analogs. This chemical may be utilizing different molecular interactions when it binds to an OBP, which may influence the conformational properties of the OBP/ligand complex and thus signal transduction. EAG recordings derived from a response of the whole antenna and, at realistic physiological concentrations, as used here, usually correlate with behavior, which could be a positive or negative response. The differences in EAG activities and ligand binding between the sex pheromone components and the analogs necessitate further investigations into the physiological roles of OBPs in vivo and interaction between OBPs and sensory receptors. Such research has begun with electrophysiological approaches (Pophof, 2004) and modern molecular technologies such as RNA interference (RNAi) to control OBP gene expression in mosquitoes (Biessmann et al., 2010; Pelletier et al., 2010). Similar studies in crop pests, such as the tobacco hornworm, Manduca sexta (L.) (Lepidoptera: Sphingidae), which shares sex pheromone components with B. mori, could lead to an alternative pest control strategy by preventing sex pheromone detection, thus reducing moth populations (Zhou et al., 2010).

Acknowledgments

Rothamsted Research receives grant-aided support from the Biotechnology and Biological Sciences Research Council (BBSRC) of the United Kingdom. XLH was funded by the BBSRC SCIBS initiative. The authors also like to thank the reviewers for their valuable comments, and Associate Editor Steven Seybold for his constructive editing of the manuscript.

Copyright information

© Springer Science+Business Media, LLC 2010