Amino Acid Residues Contributing to Function of the Heteromeric Insect Olfactory Receptor Complex

Olfactory receptors (Ors) convert chemical signals—the binding of odors and pheromones—to electrical signals through the depolarization of olfactory sensory neurons. Vertebrates Ors are G-protein-coupled receptors, stimulated by odors to produce intracellular second messengers that gate ion channels. Insect Ors are a heteromultimeric complex of unknown stoichiometry of two seven transmembrane domain proteins with no sequence similarity to and the opposite membrane topology of G-protein-coupled receptors. The functional insect Or comprises an odor- or pheromone-specific Or subunit and the Orco co-receptor, which is highly conserved in all insect species. The insect Or-Orco complex has been proposed to function as a novel type of ligand-gated nonselective cation channel possibly modulated by G-proteins. However, the Or-Orco proteins lack homology to any known family of ion channel and lack known functional domains. Therefore, the mechanisms by which odors activate the Or-Orco complex and how ions permeate this complex remain unknown. To begin to address the relationship between Or-Orco structure and function, we performed site-directed mutagenesis of all 83 conserved Glu, Asp, or Tyr residues in the silkmoth BmOr-1-Orco pheromone receptor complex and measured functional properties of mutant channels expressed in Xenopus oocytes. 13 of 83 mutations in BmOr-1 and BmOrco altered the reversal potential and rectification index of the BmOr-1-Orco complex. Three of the 13 amino acids (D299 and E356 in BmOr-1 and Y464 in BmOrco) altered both current-voltage relationships and K+ selectivity. We introduced the homologous Orco Y464 residue into Drosophila Orco in vivo, and observed variable effects on spontaneous and evoked action potentials in olfactory neurons that depended on the particular Or-Orco complex examined. Our results provide evidence that a subset of conserved Glu, Asp and Tyr residues in both subunits are essential for channel activity of the heteromeric insect Or-Orco complex.


Introduction
The detection of odorants and pheromones is essential for insects to find food, avoid predators and noxious agents in the environment, and find appropriate mating partners. Insects sense odorants and pheromones via specialized olfactory sensory neurons (OSNs) located on two sensory appendages on the head, the antennae and maxillary palps. Both of these appendages are covered with specialized sensory hairs called sensilla that house one to four OSNs in the vinegar fly Drosophila melanogaster and the silkmoth Bombyx mori. The dendritic knobs of OSNs are enriched in membrane-bound odorant receptors (Ors) that play a primary role in recognizing an odorant or a pheromone [1,2,3,4]. The Or expressed in an OSN determines the sensitivity and specificity of the OSN [5], which in turn governs innate and learned olfactory behaviors, such as attraction to food and pheromones and avoidance of repellents [6].
Insects possess 60-400 members of the Or family, which can be divided into three distinct functional classes. The first two classes of Ors are ligand-selective-those that respond to general odorants and a smaller number specialized to detect pheromones [7,8,9,10,11]. General odorant-selective Ors have little homology between insect species, whereas pheromone receptors in the moth show some homology [9,12,13,14,15,16,17,18] ( Figure S1A). The third functional class is a single member of the Or family called Orco, which is highly conserved in all known insect species and functions as an obligate chaperoning co-receptor in complex with ligand-selective Ors [19,20,21,22]. Most OSNs co-express one of the canonical Ors and the Orco family, and these two types of receptors comprise a heteromultimeric complex of unknown stoichiometry [12,22,23] The role of Ors in all animals is to convert chemical signals to electric signals. In vertebrates, Ors are G-protein-coupled receptors (GPCRs), and the odorant ligand activates a signaling pathway that leads to the production of intracellular second messengers and subsequent opening of ion channels [24]. In contrast, how the insect Or-Orco complex converts odorant or pheromone binding to OSN depolarization is less well understood. While insect Ors were initially assumed to be seven transmembrane domain GPCRs, further analysis showed that they lack sequence similarity with known GPCRs [25]. In addition, the membrane topology of insect Ors is inverse to that of GPCRs, with the amino terminus located intracellular and an extracellular carboxy terminus [22,26] Despite the lack of homology and inverted topology relative to GPCRs, several groups have provided evidence that insect Ors signal through or are modulated by Gproteins [27,28,29,30]. Other studies have reported that Gproteins and cyclic nucleotides are not involved in odor responses in vivo or in vitro [31,32,33]. Therefore, the question of whether insect Ors function like GPCRs or are modulated by G proteins remains controversial. Several groups have recently proposed that the insect Or-Orco complex functions as a novel type of ligandgated nonselective cation channel [29,32,33,34,35,36], which may rely on Ga s and Ga q pathways for function [29,30].
While there is agreement among these groups that the Or-Orco complex can form a non-selective cation channel, there is not yet a clear consensus on how these membrane proteins function. Wicher et al. proposed that Orco can function as a cyclic nucleotide activated cation channel in the absence of a ligand-selective Or subunit [29]. Jones et al. provided support for this model in their discovery of VUAA1, an allosteric agonist that can activate Orco expressed in heterologous cells without an partner Or subunit [33]. Two other groups demonstrated that the sensitivity of the Or-Orco complex to the cation channel blocker ruthenium red depended on the Or-Orco subunit composition [36]. This suggested that both Or and Orco subunits may contribute to ion permeability of this membrane protein complex. Despite much recent interest in studying the structure and function of insect odorant receptors, a large number of important questions remain unsolved. How do odorants and pheromones activate the Or-Orco complex and how do ions permeate this protein complex? In the heteromeric complex, do both Or and Orco subunits contribute to ion permeability? If so, is it possible to narrow down the regions that contribute to ion permeability? Do such regions contain residues previously associated with ion-conducting pores in ion channels, or does this class of receptors have a completely novel structure?
In the present study, we began to answer these questions by carrying out a comprehensive mutational analysis of all conserved Glu, Asp, and Tyr residues in the Or-Orco complex using the silkmoth bombykol pheromone receptor complex as a model. Our aim was to identify regions in Or and/or Orco that are required for ion permeability and other biophysical properties of this protein complex. Of the 83 conserved residues in Or and Orco mutated, three altered both current-voltage relationships and ion selectivity of the Or-Orco complex. Our results suggest that both the ligand-selective Or subunit and the Orco co-receptor contribute to cation channel activity and that some amino acid residues near the carboxy terminus of both subunits are important for Or-Orco channel function.

Results
Site-directed mutagenesis of conserved residues in the insect Or-Orco complex We carried out a comprehensive mutational screen to identify amino acids in Ors and Orco that are important for odor-evoked function. Many cation channels have Glu, Asp, or Tyr residues in their ion selectivity filters, and these residues play important roles in selective ion permeation [37]. Because the insect Or-Orco complex also forms non-selective cation channels [29,32,33,34], we reasoned that among the conserved Glu, Asp, or Tyr residues, some may be essential for Or-Orco cation channel function. To identify such conserved residues, we first compared the amino acid sequence of the bombykol receptor, BmOr-1 (Bombyx mori olfactory receptor 1) [38], with pheromone receptors in other insects ( Figure S1A). We also compared BmOrco (Bombyx mori Orco) with Orco in other insect species ( Figure S1B). 29 Glu, Asp, or Tyr residues in BmOr-1 and 54 Glu, Asp, or Tyr residues in BmOrco are highly conserved ( Figure 1A). We used site-directed mutagenesis to mutate all 83 Glu, Asp, and Tyr residues to Gln, Asn and Ala, respectively. Each mutant BmOr-1 or BmOrco contained a single amino acid mutation for a total of 83 individual mutants analyzed in this study. We expressed either mutant BmOr-1 with wild type BmOrco or wild type BmOr-1 with mutant BmOrco in Xenopus oocytes and measured current-voltage relationships of responses elicited by bombykol. We quantified a rectification index, defined as the ratio of current amplitude at +50 mV versus 280 mV, and the reversal potential of each mutant Or-Orco combination and compared it to the same parameters in wild type Or-Orco ( Figure 1). To eliminate confounding effects of biological variation found in oocytes derived from different animals, responses of a given Or-Orco mutant combination were recorded from oocytes derived from the same individual.

Effects of Glu, Asp or Tyr mutation in BmOr-1 and BmOrco on ion selectivity
The alteration of amino acids involved in pore formation can alter ion channel selectivity. We therefore performed ion substitution experiments to ask if any of the eight BmOr-1 and five BmOrco mutations that affected the rectification index and reversal potential also affected ion permeability. To calculate permeability ratios, the reversal potential of oocytes expressing wild type and mutants was measured in Na + and K + extracellular solution. We tested the 13 candidate mutants (BmOr-1: Y170, D226, D299, E325, E356, D367, E375, D378; BmOrco: E171, E329, D343, E422, Y464) in combination with a paired wild type Or or Orco subunit.
The P K /P Na of two BmOr-1 mutants (D299N, E356Q) was slightly decreased compared with that of wild type (Figure 2A,B). The same decrease in P K /P Na compared to wild type was found for the homologous BmOr-3 E365Q mutation ( Figure S3B). In contrast, the P K /P Na of BmOrco Y464A was slightly increased compared to wild type (Figure 2A,B). The other 10 candidate mutants did not show altered P K /P Na (BmOr-1: Y170A, D226N, E325Q, D367N, E375N, D378N; BmOrco: E171Q, E329Q, D343N, E422Q) ( Figure S2C). Therefore only a small number of the mutations examined here-BmOr-1 D299, BmOr-1 E356, and BmOrco Y464-altered the ion selectivity of the BmOr-1-BmOrco complex ( Figure 2C), although the remaining 10 residues may be involved in selectivity of ions other than K + .

Analysis of accessibility of methane thiosulfonate reagents to BmOr-BmOrco Cys mutants
We next used the Cys-modifying reagent, 2-(trimethylammonium)ethyl methanethiosulfonate, bromide (MTSET), to examine the accessibility of Cys mutants of BmOr-1 D299, BmOr-1 E356, and BmOrco Y464 to this modifying reagent. MTSET is a reagent that covalently modifies Cys residues that can be used to probe the functional properties of an ion channel [39]. If the targeted Cys is located in the channel pore, MTSET modification can affect ion permeability [39].
We first determined the sensitivity of wild type BmOr-1-BmOrco to MTSET. The responsiveness of oocytes expressing wild type BmOr-1-BmOrco to bombykol was reduced 46% by MTSET ( Figure 3A), perhaps due to modification of native Cys residues (data not shown). We next compared the effect of MTSET on BmOr-1 D299C, BmOr-1 E356C, and 6 other BmOr-1 mutants (Y170C, D226C, E325C, E367C, E375C, D378C) in combination with wild type BmOrco ( Figure 3A,B). With the exception of BmOr-1 E356C, all other BmOr-1 Cys mutants showed a wild type response magnitude to bombykol in the presence of MTSET ( Figure 3A,B). In contrast, the response of BmOr-1 E356C was strongly and irreversibly inhibited by MTSET application (Figure 3A,B). To exclude the possibility that the BmOr-1 E356C mutation affects ion permeation properties indirectly by a global change in protein structure, we determined the dose-response curve of BmOr-1 E356C to wild type BmOr-1. The EC 50 value of the BmOr-1 E356C mutant was indistinguishable from wild type ( Figure S4A), suggesting that the BmOr-1 E356C mutation does not affect ligand binding.
We next examined the effect of MTSET modification on the BmOrco Y464C mutant. In contrast to oocytes expressing wild type BmOr-1-BmOrco, which showed no activation by MTSET ( Figure 3A), oocytes expressing wild type BmOr-1 with BmOrco Y464C showed a rapid inward current induced by MTSET ( Figure 3A). A similar response to cysteine-modifying reagents has been observed for cyclic nucleotide-gated channels when Cys mutant residues were located in the ion channel pore domain [40]. The EC 50 value of the BmOrco Y464C mutant was indistinguishable from wild type ( Figure S4B). Taken together, these results show that BmOr-1 E356 in transmembrane domain (TM) 6 and BmOrco Y464 in TM7 are the only residues examined in this study that are sensitive to the MTSET Cys modifying reagent.

Analysis of corresponding TM7 Tyr mutation in Drosophila Orco
To ask if the function of BmOrco Y464 is conserved in other insects, we mutated the corresponding Y478 residue in Drosophila Orco [19] ( Figure S1B). We expressed Drosophila Orco Y478A along with one of five wild type Drosophila Ors (Or22a, Or47a, Or59b, Or85a, or Or85b) in oocytes, and measured rectification index, reversal potential, and ion selectivity of these channels gated by their cognate ligands (ethyl butyrate, pentyl acetate, methyl acetate, ethyl-3-hydroxybutyrate, or 2-heptanone, respectively) [5] ( Figure 4A). Drosophila Orco Y478A combined with Or59b, Or85a, and Or22a did not respond to odor ligands, precluding further functional analysis ( Figure 4A). Or85b expressed with Drosophila Orco Y478A showed a decreased rectification index but no effect on reversal potential or ion selectivity when compared to Or85b expressed with wild type Drosophila Orco ( Figure 4A). Or47a expressed with Drosophila Orco Y478A showed a decreased rectification index, an increased reversal potential, and decreased P K /P Na compared to Or47a expressed with wild type Drosophila Orco ( Figure 4A). This suggests that mutating the TM7 Y478 residue in Drosophila Orco causes biophysical phenotypes that depend on the particular Or-Orco complex being studied.
Analysis of Drosophila Orco Y478A and BmOrco Y464A mutants in Drosophila sensory neurons Heterologous expression of Or-Orco in Xenopus oocytes allowed us to measure the effect of single amino acid mutations on several biophysical properties of the complex, such as rectification index, reversal potential, and ion selectivity. To ask if mutation of the conserved Y478 residue in Drosophila Orco (and the homologous BmOrco Y464A mutation) affected the function of the Or-Orco complex in vivo, we turned to extracellular recordings of spontaneous and odor-evoked action potentials of Drosophila OSNs expressing wild type Orco or Orco Y478A. This Orco mutation might impact spontaneous or odor-evoked action potentials, or both, in vivo.
To examine the effect of the Orco Y478A mutation in vivo, we generated transgenic flies that expressed wild type or mutant Drosophila Orco in all Orco-expressing OSNs. This was achieved using the Gal4/UAS system [41] to express wild type and Orco Y478A under the control of Orco-Gal4 in Orco null mutant flies [19]. For control experiments, we expressed wild type Orco using Gal4/UAS in the same Orco mutant genetic background. We first confirmed that Drosophila Orco Y478A is expressed at comparable levels to wild type Drosophila Orco and is appropriately trafficked to OSN dendrites in the antenna ( Figure 4B).
We next recorded from ab2A, ab2B, ab3A, ab3B and ab5B OSNs, which express endogenous wild type Or59b, Or85a, Or22a, Or85b and Or47a, respectively [3,4] (Figure 4C,D). Similar to what we observed for Orco Y478A in oocytes, we found a diversity of functional effects that depended on the specific Or-Orco complex. The ab2A Or59b neuron expressing Orco Y478A showed a large increase in spontaneous activity and a strong decrease in responses evoked by methyl acetate, an odorant that elicits an excitatory response in wild type ab2A neurons [5] ( Figure 4C,D). In contrast, in the ab2B neuron expressing Or85a, the Orco Y478A mutation caused a change in spontaneous activity with no effect on odor-evoked activity ( Figure 4C,D). The reverse phenotype was found for Orco Y478A expressed in ab3B OSNs expressing Or85b, such that the Or85b-Orco Y478A receptor had wild type spontaneous activity but a strong decrease in odor-evoked activity ( Figure 4C,D). The ab5B neuron expressing Or47a showed a decrease in both spontaneous and odor-evoked activity ( Figure 4C,D). Finally, there was no effect of the Orco Y478A mutation on spontaneous or evoked responses in the ab3A neuron expressing Or22a ( Figure 4C,D).
To ask if this effect of the Orco TM7 Tyr residue on the in vivo function of Drosophila Orco was conserved, we expressed wild type BmOrco or BmOrco Y464A in an Orco null mutant background ( Figure 4E,F). Wild type BmOrco in combination with native Drosophila Or59b, Or85a, Or22a, Or85b, and Or47a in ab2A, ab2B, ab3A, ab3B, and ab5B OSNs, respectively, showed normal responses to the cognate odor ligand of each OSN (data not shown), suggesting that silkmoth BmOrco can function with endogenous Drosophila Ors as previously shown for Orco from several other insect species [42]. We then turned to the functional analysis of BmOrco Y464A. This mutant combined with Or59b, Or85a, Or22a, and Or85b did not respond to odor ligands, precluding further functional analysis (data not shown). However, BmOrco Y464A expressed with Or47a showed normal spontaneous activity but a strong increase in odor-evoked activity ( Figure 4E,F).
Taken together, these results suggest that the TM7 Tyr residue in Orco (Y478 in Drosophila and Y464 in Bombyx) contributes to the function of the insect Or-Orco complex in vivo, although the function of this Tyr residue may not be identical between DmOrco and BmOrco. Because the effect of the Orco Y478A mutation differed depending on the Or-Orco complex, we suggest first, that each insect Or-Orco complex possesses different functional properties and second, that both the ligand selective Or subunit and the Orco co-receptor subunit contributes to channel activity.

Discussion
In this study, we carried out comprehensive site-directed mutagenesis of all conserved Glu, Asp, and Tyr residues in the silkmoth bombykol receptor to probe the structure-function relationships of the Or-Orco complex. 13 of the 83 residues caused functional alterations in odor-evoked cation channel activity. Furthermore, three of the 13 residues showed altered ion selectivity. Two of the residues were located in transmembrane domain (TM) TM5 and TM6 in a ligand-selective Or and a third was in TM7 in Orco. These three residues may contribute to ion permeability of the receptor complex, although our data cannot resolve whether these three residues are part of an ion-conducting pore or merely influence the function of a pore residing elsewhere in the protein complex.
Pore domains of cation channels are typically formed by the assembly of multiple subunits [37]. Insect Or-Orco functions as a heteromultimer [12,22,23], but the stoichiometry and subunit composition of the complex are unknown. Further, it is unclear whether the ion-conducting pore structure of Or-Orco complex is formed by Orco alone or whether both ligand-selective Ors and Orco contribute to the pore. There is evidence that the Orco subunit alone can form an ion channel [29,33], but there is also suggestive evidence that the ligand-selective Or contributes to the ionic properties of the Or-Orco complex. First, sensitivity of Or-Orco to ruthenium red varies with Or subunit composition   [34,36]. Second, the reversal potential of Anopheles Orco alone is slightly larger than that of the Anopheles Or10-Orco complex, indicating that ion selectivity is modulated by the ligand-selective Or subunit [33]. In this study, we provide confirmatory evidence that both subunits contribute to ionic permeability of the insect Or-Orco complex.
Our work suggests that the Or-Orco complex has two important characteristics. First, the biophysical properties of the channel vary according to subunit composition, even with highly similar proteins such as BmOr-1-Orco and BmOr-3-Orco. Second, because ligand-selective Or sequences within and between insect species are extremely divergent, the primary amino acid sequence of the ion-conducting pore is likely to differ according to the subunit composition of the Or-Orco complex. This is consistent with our model that the ion pore of the insect Or requires the participation of both Or and Orco subunits. Neither Orco nor any ligand-selective Or has been found to have any homology to known ion-conducting pore domains in other ion channels. This suggests that insect Ors will define a completely novel structural domain for ion selectivity and permeability not found in other ion channels. Further experiments will be needed to understand the precise pore structure of the Or-Orco complex. A comprehensive substituted cysteine accessibility study could reveal amino acid residues that reside in the pore lumen. High-resolution x-ray structural analysis of open and closed channel states would inform the stoichiometry of the complex and the mechanisms underlying ion conduction of this unusual family of ion channels.

Odor ligands and protein modification reagents for heterologous expression analysis
Bombykol and bombykal were synthesized as previously reported [38] and solutions of these two compounds were prepared in dimethylsulfoxide (DMSO). Other odorants were purchased from Tokyo Kasei (Tokyo, Japan) and were directly diluted into control bath solution (88 mM NaCl, 1 mM KCl, 0.

Gene expression in Xenopus laevis oocytes and twovoltage clamp recording
Stage V to VII oocytes were treated with 2 mg/ml of collagenase B (Roche Diagnostics, Tokyo, Japan) in Ca 2+ -free saline solution (82.5 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , and 5 mM HEPES, pH 7.5) for 1 to 2 h at room temperature. cRNA was synthesized from linearized modified pSPUTK vector. Oocytes were microinjected with 25 ng of BmOr-1 or BmOr-3 cRNA and 25 ng of cRNA BmOrco or Drosophila Orco. Injected oocytes were incubated for 3-4 days at 18uC in bath solution supplemented with 10 mg/ml of penicillin and streptomycin.
Whole-cell currents were recorded using the two-electrode voltage-clamp technique as previously described [12]. Intracellular glass electrodes were filled with 3 M KCl. Signals were amplified with an OC-725C amplifier (Warner Instruments, Hamden, CT, USA), low-pass filtered at 50 Hz and digitized at 1 kHz. The control bath solution contained 115 mM NaCl, 2.5 mM KCl, 1.8 mM BaCl 2 , and 10 mM HEPES, titrated to pH 7.2 with NaOH.
The rectification index was calculated as the ratio of current amplitude recorded at +50 mV versus amplitude at 280 mV. In ion substitution experiments, the following solutions were used: 115 mM XCl, 10 mM HEPES, titrated to pH 7.2 with XOH (X = Na + or K + ). To calculate the permeability ratio, the following extended form of the Goldman-Hodgkin-Katz flux equation was used: P K /P Na = [Na] o /[K] o ?exp(DE rev ?F/RT). In the measurement of E rev , the junction potential was corrected. In the experiment using MTSET reagent, each oocyte was pretreated with MTSET (2.5 mM) for 60 seconds before the application of bombykol. Ligands were delivered through the superfusing bath solution via a silicon tube connected to a computer-driven solenoid valve. Data acquisition and analysis were carried out with Digidata1322A (Axon instruments, Foster city, CA, USA) and pCLAMP software (Axon instruments, Foster city, CA, USA).

Fly strains and transgenic constructs
Drosophila melanogaster stocks were maintained on conventional cornmeal-agar-molasses medium under a 12 hour light:12 hour dark cycle at 25uC. Transgenic animals were generated in the w 1118 genetic background (Genetic Services Inc., Cambridge, MA, USA) using the phiC31-based integration system [44] targeted to the attP2 docking site on chromosome II [45]. UAS-Orco [22] and Orco-Gal4 [19] transgenes were randomly integrated via conventional P-element vectors.

Single sensillum electrophysiology and odorants
Female transgenic flies were recorded between 5 and 7 days after adult eclosion. Single sensillum recordings were performed as described [46,47].
30 ml of the desired odor dilution was pipetted onto a filter paper strip (3650 mm), which was then carefully inserted into a glass Pasteur pipette. Prior to any recordings, charcoal-filtered air was forced through the pipette for 1-3 sec to remove dead space in the odor delivery system. For actual recordings, charcoal-filtered air was continuously applied to the insect antenna, with odor delivered through the pipette to the fly antennae for 1 sec. Each pipette was used at most three times and no more than three sensilla were tested per animal. Sensilla types were identified by size, location on the antenna, and responsiveness to known preferred odorants [5].
Data were collected using Autospike software (Syntech, Kirchzarten, Germany) and analyzed by custom spike sorting algorithms [46]. Spikes from the ab5A and B neurons were not sorted because of the similarity in spike amplitudes. The data were analyzed by calculating the number of spikes/sec in 200 msec bins. The peak odor-evoked activity was calculated by subtracting the average spontaneous activity (expressed in spikes/sec) during the two seconds before odor application from peak activity during odor delivery. This value is referred to as Dspikes/sec. The onset of odor-evoked responses varied due to slight variations in the position of the odor delivery system relative to the sensillum being recorded. To correct for this, we calibrated the inferred odor onset for each sensillum recorded based on excitatory responses for each sensillum elicited by control stimuli (ab2: 10 25 methyl acetate; ab3: 10 25 2-heptanone; ab5: 10 22 geranyl acetate).