Molecular and functional characterization of an octopamine receptor from honeybee (Apis mellifera) brain

Authors


Address correspondence and reprint requests to Dr Arnd Baumann, Institut für Biologische Informationsverarbeitung 1, Forschungszentrum Jülich, D-52425 Jülich, Germany. E-mail: a.baumann@fz-juelich.de

Abstract

Biogenic amines and their receptors regulate and modulate many physiological and behavioural processes in animals. In vertebrates, octopamine is only found in trace amounts and its function as a true neurotransmitter is unclear. In protostomes, however, octopamine can act as neurotransmitter, neuromodulator and neurohormone. In the honeybee, octopamine acts as a neuromodulator and is involved in learning and memory formation. The identification of potential octopamine receptors is decisive for an understanding of the cellular pathways involved in mediating the effects of octopamine. Here we report the cloning and functional characterization of the first octopamine receptor from the honeybee, Apis mellifera. The gene was isolated from a brain-specific cDNA library. It encodes a protein most closely related to octopamine receptors from Drosophila melanogaster and Lymnea stagnalis. Signalling properties of the cloned receptor were studied in transiently transfected human embryonic kidney (HEK) 293 cells. Nanomolar to micromolar concentrations of octopamine induced oscillatory increases in the intracellular Ca2+ concentration. In contrast to octopamine, tyramine only elicited Ca2+ responses at micromolar concentrations. The gene is abundantly expressed in many somata of the honeybee brain, suggesting that this octopamine receptor is involved in the processing of sensory inputs, antennal motor outputs and higher-order brain functions.

Abbreviations used
[Ca2+]i

intracellular calcium concentration

cAMP

cyclic AMP

ECFP

enhanced cyan fluorescent protein

ES

extracellular solution

GPCR

G protein-coupled receptor

HEK

human embryonic kidney

IBMX

isobutylmethylxanthine

PBS

phosphate-buffered saline

PER

proboscis extension response

PKC

protein kinase C

SDS

sodium dodecyl sulfate

TM

transmembrane

US

unconditioned stimulus

Octopamine is a monophenolic amine that belongs to a group of neuroactive compounds known as biogenic amines. Biochemical and pharmacological experiments suggest that octopamine exerts its effects by binding to membrane proteins that belong to the superfamily of G protein-coupled receptors (GPCRs). These receptor proteins share the structural motif of seven transmembrane (TM) domains (Baldwin et al. 1997; Okada et al. 2001). Activation of the receptors may lead to changes in the concentration of intracellular second messengers such as cyclic nucleotides [cyclic AMP (cAMP) and cyclic GMP], inositol-1,4,5-trisphosphate and Ca2+. Octopamine-mediated changes in the intracellular concentration of cAMP and/or Ca2+ ([Ca2+]i) have been reported for several protostomian species (for reviews, see Roeder 1999; Blenau and Baumann 2001).

Since its discovery in the salivary glands of the octopus (Erspamer and Boretti 1951), octopamine has been found in high concentrations in neuronal and non-neuronal tissues of many nematodes, annelids, arthropods and molluscs (David and Coulon 1985). Owing to its regulatory functions, octopamine is considered to be a neurotransmitter, neuromodulator and/or neurohormone (for reviews, see David and Coulon 1985; Roeder 1999). Many behavioural and physiological reactions have been attributed to the signalling action of octopamine (David and Coulon 1985; Orchard et al. 1993), particularly in a number of studies in the honeybee (Braun and Bicker 1992; Erber et al. 1993; Burrell and Smith 1995; Pribbenow and Erber 1996; Schulz and Robinson 2001; Scheiner et al. 2002). It has been shown that octopamine can modulate the responsiveness of sensory receptors, interneurones and motoneurones, and so affects complex behavioural responses.

Octopamine also plays a major role in olfactory learning and memory formation in the honeybee (Hammer 1993, 1997; Hammer and Menzel 1998; Menzel et al. 1999). In these studies, olfactory conditioning of the proboscis extension response (PER) is used as the learning paradigm (Bittermann et al. 1983; Menzel and Müller 1996). In a classical conditioning protocol bees learn to associate an odour (conditioned stimulus) with sucrose presentation (unconditioned stimulus; US). During conditioning an odour is presented shortly before stimulating the antenna with sucrose, which elicits the PER. The animal is then rewarded by applying sucrose to the proboscis, leading to an association of the odour with the reward. After a single learning trial up to 80% of the bees respond with a conditioned PER. An identified neurone (VUMmx1) can mediate the US during olfactory conditioning in the honeybee (Hammer 1993). Immunohistological studies suggest that the VUMmx1 neurone belongs to a group of octopaminergic cells. It has been shown that electrical stimulation of the VUMmx1 neurone (Hammer 1993) or the injection of its putative transmitter octopamine, either into the antennal lobes or the calyces of the mushroom bodies (Hammer and Menzel 1998), can substitute for the US during olfactory conditioning.

To understand the molecular mechanisms that are controlled by the octopaminergic system it is necessary to characterize the respective octopamine receptors in the bee. Here we describe the molecular cloning and functional characterization of the first octopamine receptor from honeybee brain. The encoded protein is 587 amino acid residues in length and shares ∼60% amino acid similarity with octopamine receptors from Drosophila (Han et al. 1998) and Lymnea (Gerhardt et al. 1997a). The functional characterization of this octopamine receptor should help to identify its role in honeybee neuromodulation and learning.

Materials and methods

Isolation of cDNA clones and sequencing

A cDNA library from Apis mellifera brains was constructed in λ ZAP II (Stratagene, La Jolla, CA, USA) and screened with a 900-bp EcoRI/XbaI restriction fragment of the AmBAR1-cDNA clone (Ebert et al. 1998; Kokay et al. 1999) as a probe. Hybridization was performed under high stringency in 5 × SET (20 × SET contains 3 m NaCl, 0.4 m Tris-HCl, pH 7.5, 0.02 m EDTA), 5 × Denhardt's solution (100 × Denhardt's contains 2% bovine serum albumin, 2% Ficoll 400, 2% polyvinylpyrrolidone), 100 µg/mL autoclaved herring testis DNA, 0.1% sodium dodecyl sulfate (SDS) and ∼1 × 106 cpm/mL labelled probe at 61°C overnight. Filters were rinsed in 1 × SET, 0.1% SDS for 5 min at 22°C and twice for 30 min at 61°C. Plasmid DNA of positive clones was isolated using the in vivo excision protocol (Stratagene). Subcloning of restriction fragments was done into pBluescript SK(-) vector (Stratagene) by standard cloning techniques (Sambrook et al. 1989). Sequencing of restriction fragments used the thermo sequenase fluorescent-labelled primer cycle sequencing kit (Amersham-Pharmacia, Piscataway, NJ, USA) and the LICOR electrophoresis system (MWG Biotech, Ebersberg, Germany). The nucleotide sequence of Amoa1 has been submitted to the EMBL database (accession number AJ 547798).

Multiple sequence alignment and phylogenetic analysis

Biogenic amine receptor sequences for phylogenetic analysis from both Drosophila and A. mellifera were deduced from cDNA sequences in GenBank release 133.0. The Drosophila genomic DNA sequence was also searched using known biogenic amine receptors as query sequences to identify additional receptor homologues. TBLASTn searches, not filtered for low complexity, employing the Blosum 45 scoring matrix with a gap penalty of 16 : 1 were used to ensure that all homologues were identified. The results of the homology searches were combined with the computer-generated hypothetical protein predictions in GenBank to enhance the accuracy of receptor sequence prediction. All single nucleotide discrepancies were resolved in favour of the genomic DNA sequence. Multiple sequence alignments were performed, and highly divergent sequences at the amino and carboxyl termini and between TM5 and TM6 were trimmed from the sequences. Genetic distance between sequences was then calculated with ClustalX (Thompson et al. 1997) using the Blosum scoring matrix option. Neighbour joining trees were constructed in ClustalX using 1000-fold bootstrap re-sampling and the resulting trees were displayed graphically by Treeview (Page 1996) using the divergent muscarinic receptors as an outgroup. For clarity, two branches of 15% and 18% bootstrap support were collapsed to the base of the tree. A second tree was created from all pharmacologically defined tyramine and octopamine receptors, and from two additional orthologous receptors deduced from cDNA sequences. These sequences were from a range of insect species, a nematode and several molluscs. Alignment and tree construction parameters were the same as above.

Construction of pcAmoa1 expression vector

A truncated version of the Amoa1 cDNA containing a unique HindIII restriction site and the Kozak consensus motif (Kozak 1984) immediately 5′ to the initiating ATG codon was constructed by PCR. The following oligonucleotides were used: 5′-GATAAGCTTCCACCATGCGATCCGTATTC and 5′-ATGGATCCTCAAGGTCAA. The PCR product was digested with HindIII and XbaI. The Amoa1 cDNA in pBluescript SK(-) was digested with XbaI and EcoRI. Restriction fragments were gel purified and ligated into HindIII and EcoRI-cut pcDNA1.amp-vector (Invitrogen, Carlsbad, CA, USA). To monitor transfection efficiency and receptor protein expression, a haemagglutinin epitope-Tag (HA-Tag) was engineered to the 3′ end of the cDNA clone by standard cloning techniques. The resulting recombinant was named pcAmoa1-HA.

In situ hybridization

Hybridization to cryosections of adult honeybee brain was performed with digoxigenin-labelled riboprobes. Antisense and sense probes were transcribed using T7 and T3 RNA polymerase respectively from a cDNA fragment (1974–2092bp) encoding part of the third intracellular loop of the AmOA1 receptor, which had been cloned into pBluescript. Pre-hybridization was performed in 50% formamide, 5 × SSC (20 × SSC contains 3 m NaCl, 0.3 m sodium citrate, pH 7.4), 100 µg/mL autoclaved herring testis DNA, 50 µg/mL heparin and 0.1% Tween 20 at 45°C for 30 min. Hybridization was done in pre-hybridization solution containing 0.5 µg/mL digoxigenin-labelled probe overnight in a humidified chamber at 45°C. Washing was carried out in 50% formamide and 2 × SSC at 37°C for 60 min followed by two washes for 90 min each. For detection of hybrids, the sections were incubated with anti-digoxigenin antibody conjugated with alkaline phosphatase and stained with Nitro-blue tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate (NBT/BCIP) solution following the DIG Nucleic Acid Detection Kit (Roche Applied Science, Penzberg, Germany).

Heterologous expression of pcAmoa1-HA

Exponentially growing human embryonic kidney (HEK) 293 cells (∼2 × 105 cells per 5 cm dish) were transfected with 10 µg pcAmoa1-HA by a modified calcium phosphate method (Chen and Okayama 1987). For Ca2+ fluorimetric experiments cells were co-transfected with pcAmoa1-HA and a gene encoding enhanced cyan fluorescent protein (ECFP; BD Biosciences, Heidelberg, Germany). Twenty hours after transfection the precipitate was washed off with phosphate-buffered saline (PBS) followed by PBS/1.34 mm EDTA. Cells were either transferred on to poly-l-lysine coated coverslips (for Ca2+ fluorimetry) or were left in the Petri dishes (for membrane preparations) and fed with fresh medium. Functional coupling of expressed receptors to intracellular signalling pathways was tested 24 h later.

Functional characterization of AmOA1 receptors

The ability of AmOA1 to trigger changes in [Ca2+]i was monitored with the Ca2+-sensitive fluorescence dye Fluo-4 (Molecular Probes, Eugene, OR, USA). Experiments were done on co-transfected HEK 293 cells which facilitated identification of AmOA1-expressing cells because they showed ECFP fluorescence (λexc 435 nm; λem 520–560 nm). Cells were incubated at 37°C in extracellular solution (ES; 150 mm NaCl, 5 mm KCl, 2 mm MgCl2, 2 mm CaCl2, 10 mm HEPES, 30 mm glucose, pH 7.4) containing 2 µm Fluo-4AM (Molecular Probes) and 0.02% Pluronic® F-127 (Molecular Probes). After 45 min cells were washed with dye-free ES. For receptor activation cells were superfused with ES containing different concentrations of octopamine or tyramine. A single-cell photon-counting system (PhoCal, Life Science Resources, Cambridge, UK) was used to measure [Ca2+]i-dependent changes in Fluo-4 fluorescence. Excitation wavelength was 480 nm (xenon lamp, 100 W; Nikon, Dusseldorf, Germany). Fluorescence emission was detected at 520–560 nm. The sampling rate of the photon-counting system was adjusted to 100 ms.

Assays to determine the ability of AmOA1 to activate adenylyl cyclase were performed after transient expression of pcAmoa1-HA in HEK 293 cells. Incubations with different ligands were performed at 37°C for 30 min in the presence of the phosphodiesterase inhibitor isobutylmethylxanthine (IBMX; final concentration 10 µm). Cells were lysed by adding ice-cold ethanol (2 mL/dish). After 2 h at 4°C the lysate was transferred into an eppendorf cup and lyophilized. The amount of cAMP produced was determined using the TRK 432 cAMP assay kit (Amersham). Mean values of cAMP per Petri dish were determined in duplicate on two independent transfections.

Results

Molecular and structural properties of the octopamine receptor from honeybee brain

A cDNA fragment encoding TM6–7 of a GPCR was used to isolate full-length cDNA clones from head and brain-specific cDNA libraries (Blenau et al. 1998; Ebert et al. 1998; Kokay et al. 1999). The longest recombinant (Amoa1) consists of 3023 bp. The major open reading frame contains an initiation codon (ATG) at position 1058–1060 and is terminated by a translational stop codon (TGA; nucleotides 2819–2821). Nonsense codons are found in all three reading frames preceding the ATG codon. Interestingly, there are eight upstream ATG sequences in the 5′ non-translated region of the transcript. The open reading frames headed by these ATG sequences range from 18 to 153 nucleotides. The 3′ non-coding region of Amoa1 consists of 202 nucleotides and is terminated by a poly(dA) tail of 40 residues. The deduced amino acid sequence of Amoa1 (AmOA1, Fig. 1) consists of 587 residues with a calculated molecular weight of 66.5 kDa.

Figure 1.

Deduced amino acid sequence of AmOA1. Amino acids are numbered beginning with the initiating methionine. The position of the last residue in each line is given in the right margin. The seven putative TMs are overlined. Three potential N-glycosylation sites (▾) as well as three cysteine residues for potential palmitoylation in the C-terminus (◊) are indicated. Amino acid residues that are implicated in ligand binding are shown as white letters on a black background. Potential phosphorylation sites for PKC (▪) and protein kinase A (□) are also indicated.

The deduced amino acid sequence of Amoa1 shows characteristic features of the GPCR superfamily (Strader et al. 1995; Valdenaire and Vernier 1997). The hydropathy profile (data not shown) reveals seven hydrophobic domains, which most probably serve as membrane-spanning segments (TM1–7; Fig. 1). The presence of highly conserved residues that contribute to ligand binding in biogenic amine receptors supports the hypothesis that AmOA1 belongs to the subfamily of biogenic amine receptors (Strader et al. 1995). The conserved residues include an Asp residue in TM3 (D145) and Ser residues in TM5 (S251ALGS255), which are also present in AmOA1 (Fig. 1). AmOA1 contains three consensus sites for N-linked glycosylation (N-X-S/T). One site is located in the N-terminus (N47AT) and two sites are present in the second extracellular loop (N214MT, N224TT). Five consensus sites for phosphorylation by protein kinase C (PKC) (S/T-X-R/K) are present within the intracellular loops IL1 and IL3 and in the C-terminus (Fig. 1). In addition, four consensus sites for phosphorylation by protein kinase A (RRPS405RR, RRNS409CES, KRRT534NTL and RRGS541) are present. The C-terminus of AmOA1 comprises 72 amino acids and harbours three cysteine residues (C526, C528, C530) which might be a target for post-translational palmitoylation (O'Dowd et al. 1989; Jin et al. 2000).

When the AmOA1 sequence was used to query the non-redundant translated GenBank databases (release 133.0), a high degree of amino acid similarity with other insect and molluscan octopamine receptors was revealed. The greatest similarity (identical and conservatively substituted amino acid residues) of 66% exists between AmOA1 and the splice variant 1B of the Drosophila octopamine receptor (DmOA1B, GenBank accession no. AJ007617). The amino acid similarity between AmOA1 and the Drosophila octopamine receptor OAMB (Han et al. 1998) is 59%. To further define biogenic amine receptor relationships, we compared all known honeybee sequences, including that of AmOA1, to all Drosophila biogenic amine receptor sequences, either deduced from cDNA or predicted from the Drosophila genome. Six honeybee and 11 Drosophila receptor sequences were deduced directly from cDNA sequences whereas an additional 10 were predicted from genomic DNA. Two of these predicted sequences were excluded from the analysis because they were either highly divergent, pseudogenes or contained errors of prediction that could not be resolved.

A multiple sequence alignment of these 25 biogenic amine receptors was performed and used to build a phylogenetic tree (Fig. 2a). This tree groups AmOA1 with the two Drosophila octopamine receptor splice variants, DmOA1B and OAMB (Han et al. 1998). These sequences are most closely related to a clade containing the DAMB (Han et al. 1996) and AmDOP2 (Humphries et al. 2003) sequences. Both the DAMB and Amdop2 genes encode dopamine receptors unique to protostomes. Notably, AmOA1 is quite distinct from tyramine receptors (i.e. DmTYR and AmTYR1; Fig. 2a).

Figure 2.

Dendrogram of AmOA1 and protostomian biogenic amine receptors. (a) Phylogenetic comparison of all known Apis and Drosophila biogenic amine receptor sequences. The amino acid sequences were deduced from cDNA sequences except for eight of the Drosophila sequences, which were predicted from genomic DNA. Two additional likely biogenic amine receptor sequences were excluded from the analysis owing to sequence divergence and difficulty predicting a reliable protein sequence. AmOA1 was aligned with dopamine receptors from Apis (AmDOP1, accession no. (#)Y13427; AmDOP2, #AF498306; AmBAR3) and Drosophila (DAMB, #U61264; DmDOP1, #X77234; DmDOP2, #AAN15955); octopamine receptors from Drosophila (DmOA1A/OAMB, #AF065443; DmOA1B, #AJ007617); serotonin receptors from Drosophila (Dm5HT7, #P20905; Dm5HT1a, #P28285; Dm5HT1b, #P28286; Dm5HT2a, #CAA57429); tyramine receptors from Apis (AmTYR1, #AJ245824) and Drosophila (DmTYR, #M60789); a muscarinic receptor from Drosophila (DmMusc, #S05661) and hypothetical Drosophila proteins. (b) Phylogenetic relationships between octopamine and tyramine receptors of insects, molluscs and nematodes. AmOA1 was aligned with octopamine receptors from Aplysia (AcOA, #AAF37686; AkOA, #AAF28802), Bombyx (BmOA, #Q17232), Drosophila (DmOA1A/OAMB, #AF065443; DmOA1B, #AJ007617), Heliothis (HvOA, #Q25188) and Lymnea (LsOA1, #O77408; LsOA2, #O01670); tyramine receptors from Apis (AmTYR, #AJ245824), Caenorhabditis (CeTYR, #NM171978), Drosophila (DmTYR, #M60789) and Locusta (LmTYR, #Q25321). The numbers at the nodes of the branches represent the percentage bootstrap support for each branch. The scale bars allow conversion of branch lengths in the dendrogram to genetic distance between clades (0.1 = 10% genetic distance). Am, Apis mellifera; Ac/Ak, Aplysia californica/kurodai; Bm, Bombyx mori; Ce, Caenorhabditis elegans; Dm, Drosophila melanogaster; Hv, Heliothis virescens; Lm, Locusta migratoria; Ls, Lymnea stagnalis.

Phylogenetic analysis of all pharmacologically characterized octopamine and tyramine receptors (Fig. 2b) showed that AmOA1 clusters with the Lymnea and Drosophila sequences, LsOA1 (Gerhardt et al. 1997a), OAMB (Han et al. 1998) and DmOA1B, which is in complete agreement with the sequence homology search results.

Functional expression of Amoa1 cDNA in HEK 293 cells

To investigate the functional properties of AmOA1, HEK 293 cells were transiently transfected with pcAmoa1-HA. Octopamine receptors are known to increase intracellular cAMP and/or [Ca2+]i concentrations (reviewed in Roeder 1999), so we analysed the effects of AmOA1 activation on these intracellular second messenger systems. To monitor Ca2+ signals, transfected HEK 293 cells were loaded with the calcium-sensitive dye Fluo-4. Superfusing the cells with low concentrations of octopamine (1 nm) did not change [Ca2+]i(Figs 3a and c). An octopamine concentration of 50 nm, however, reproducibly induced oscillations in [Ca2+]i (Figs 3a and b). In most of the cells analysed, the responses lasted as long as the ligand was present (Fig. 3a). In some experiments, cells responded to ligand application with a delay (Fig. 3b; 50 nm octopamine) and stopped signalling even in the continuous presence of the ligand (Fig. 3b; 100 nm octopamine). The presence of the ligand, however, was necessary for the cellular responses. After stimulating the cells with 50 nm octopamine (Fig. 3b), the perfusion medium was changed for ligand-free ES, which led to the complete disappearance of Ca2+ signals (Fig. 3b; ‘wash’). Adding 100 nm octopamine to the ES re-evoked the Ca2+ signals. The higher ligand concentration increased the frequency of the Ca2+ spikes (Fig. 3b) but did not change the signal amplitude. Similar observations were made with octopamine concentrations up to 300 nm (not shown). At high octopamine concentrations (1 µm) the cells responded with one large, slowly decaying Ca2+ signal (Fig. 3c). Once the cells had been treated with micromolar concentrations of octopamine they could not be stimulated with octopamine again (nanomolar to micromolar concentrations) even after long periods (≥ 15 min) of wash-out (not shown).

Figure 3.

Agonist modulation of [Ca2+]i in transiently Amoa1-HA-transfected HEK 293 cells. Octopamine induced transient increases in [Ca2+]i. Calcium responses of individual cells expressing the AmOA1 receptor during stimulation with octopamine or tyramine are shown. Cells were loaded with the Ca2+-sensitive dye Fluo-4. Relative fluorescence intensity is depicted as counts per 100 ms on the ordinate. Application of the ligands is indicated above the signals. In the following descriptions, the number of responding cells as a proportion of the total number of cells analysed is given in parentheses. (a) Superfusion with 1 nm and 50 nm octopamine (10 of 13). (b) Superfusion with 50 nm octopamine. The perfusion was changed for a ligand-free test solution and subsequently for a test solution containing 100 nm octopamine (five of five). (c) Superfusion with 1 nm and 1 µm octopamine. Concentrations ≥ 50 nm led to oscillations in [Ca2+]i whereas 1 µm octopamine caused a large increase in [Ca2+]i which decayed slowly (four of four). (d) Cells were superfused with 50 nm and 1 µm tyramine (five of seven).

Application of tyramine in the concentration range used for octopamine activated Ca2+ signals only at high ligand concentrations (1 µm). Compared with the effect of octopamine at the same concentration, the tyramine effect was always delayed (Figs 3c and d). Octopamine (1 µm) and tyramine (1 µm) did not induce any Ca2+ response in non-transfected cells (data not shown).

To examine the ability of AmOA1 to activate adenylyl cyclase, transfected cells were incubated with increasing concentrations of octopamine and tyramine (1 nm to 10 µm) and with 1 µm dopamine and serotonin in the presence of the phosphodiesterase inhibitor IBMX (10 µm). Octopamine stimulated cAMP production only moderately at high concentrations (≥ 1 µm) (Fig. 4). Tyramine, serotonin and dopamine did not change cAMP levels in transfected cells. As was described earlier, non-transfected HEK 293 cells did not show a cAMP response to any of these amines (Gotzes et al. 1994). However, because HEK 293 cells express endogenous β-adrenergic receptors (Gerhardt et al. 1997b), incubation of non-transfected and Amoa1-transfected cells with noradrenaline (10 µm) caused a significantly higher level of cAMP production (∼19 pmol/dish; not shown) than was observed with octopamine (∼2.5 pmol/dish; Fig. 4). From these results we conclude that activation of heterologously expressed AmOA1 with octopamine at physiological concentrations specifically causes [Ca2+]i oscillations. The increase in intracellular cAMP concentration observed at high octopamine concentrations was most probably a secondary effect, induced by massive Ca2+ release.

Figure 4.

Agonist modulation of intracellular cAMP in transiently Amoa1-HA-transfected HEK 293 cells. Agonist-induced increases in intracellular cAMP. Data are mean values of two independent experiments conducted in duplicate. Error bars indicate maximum and minimum values. Cells were incubated with IBMX only (10 µm) as a control.

Expression pattern of the Amoa1 gene

The distribution of Amoa1 mRNA was analysed by in situ hybridization to cryosections of adult worker bee brain. A series of frontal sections was hybridized with digoxigenin-labelled riboprobes of either the antisense or the sense transcript. Labelling of cell somata was observed using the antisense probe only. Hybridization signals were present in many different soma clusters of the brain. Prominently labelled soma clusters in the brain include mushroom body intrinsic neurones (Fig. 5), cells belonging to the optic lobes and the deutocerebrum. This general result is similar to the mRNA distribution patterns previously described for dopamine (Blenau et al. 1998) and tyramine receptors (Blenau et al. 2000) from honeybee. Specific characteristics of Amoa1 labelling compared with Amtyr1 labelling are apparent in the soma cell cluster of mushroom body intrinsic cells. Amoa1 expression is not uniformly distributed among the somata of this cluster (Fig. 5b). Some somata exhibit very strong signals whereas others are only weakly stained. The distribution of differentially stained somata, however, does not correlate with expression of Amoa1 in distinct cell types of intrinsic mushroom body neurones. A similar observation was described for the Drosophila octopamine receptor OAMB (Han et al. 1998).

Figure 5.

In situ hybridization of Amoa1 antisense riboprobes to frontal sections of the honeybee brain. (a) Specific labelling is seen in the somata of many brain areas. The arrows indicate labelling of soma clusters in the brain. No specific labelling was obtained when a sense probe was used (data not shown). (b) Enlargement of the mushroom body calyces of one hemisphere. The non-uniformity of the signals in the somata of mushroom body intrinsic cells is clearly visible. Signals differ between very strong (black arrows) and weak (white arrows). Scale bars in both figures 100 µm. al, Antennal lobe; lam, lamina; lc, lateral calyx of the mushroom body; mc, median calyx of the mushroom body; med, medulla; pe, pedunculus of the mushroom body.

Discussion

In the present study we have identified and functionally characterized the first octopamine receptor from the honeybee, A. mellifera. Phylogenetic analysis of the deduced amino acid sequence shows that AmOA1 is a member of the subfamily of protostomian octopamine receptors. Activation of the heterologously expressed receptor by nanomolar concentrations of octopamine leads to oscillations in [Ca2+]i. The receptor encoding mRNA is abundantly expressed in many soma clusters of the honeybee brain, suggesting that AmOA1 plays a role in the processing of sensory information, antennal motor output and in higher-order brain functions.

Molecular properties of the AmOA1 receptor

Seven octopamine receptors have been functionally characterized so far, two from Lymnea stagnalis (Gerhardt et al. 1997a, 1997b), one from Drosophila melanogaster (Han et al. 1998), one from Aplysia californica (Chang et al. 2000), one from Aplysia kurodai (Chang et al. 2000), one from Heliothis virescens and one from Bombyx mori (von Nickisch-Rosenegk et al. 1996). These receptors belong to three distinct clades, two of which contain octopamine receptors and one of which contains receptors that interact with tyramine and/or octopamine. Despite the fact that the AmOA1 receptor resides within a larger clade, which also contains the insect dopamine receptors DAMB and AmDOP2 (Han et al. 1996; Kokay et al. 1999; Humphries et al. 2003), we find no evidence that it is responsive to dopamine. The AmOA1 receptor shares the characteristic seven TM domain motif of GPCRs and possesses signature amino acid residues that are specifically implicated in ligand binding (Strader et al. 1995; Baldwin et al. 1997; Palczewski et al. 2000). The protonated amino group of octopamine probably pairs with the carboxyl group of Asp145 in TM3 of AmOA1. Serine residues (Ser251 and Ser255) located in TM5 of AmOA1 are potential candidates for interaction with the hydroxyl group of the benzoyl ring of octopamine.

Signal transmission from activated GPCRs to intracellular effector systems is usually mediated by binding to specific heterotrimeric G proteins. The physical interaction of the receptor with its associated G protein is determined by amino acid residues located in the third intracellular loop and the C terminus of the receptor protein (Bourne 1997; Wess 1997). However, the amino acid sequence homology between the intracellular loops of AmOA1 and published octopamine receptor sequences is too low to allow deduction of signalling capabilities of the AmOA1 receptor by sequence comparison alone.

Functional coupling to intracellular second messenger pathways

To characterize the signalling properties of AmOA1, cDNA encoding the receptor was transiently transfected into HEK 293 cells. In contrast to various insect cell lines (Orr et al. 1992; Hu et al. 1994; Van Poyer et al. 2001; Näsman et al. 2002), these cells do not express endogenous octopamine receptors (Gotzes et al. 1994; Han et al. 1996, 1998; Blenau et al. 2000). Activation of heterologously expressed AmOA1 by octopamine led to Ca2+ oscillations (Fig. 3). The effect was specific for octopamine, because nanomolar concentrations of the ligand were sufficient to evoke the signal, whereas micromolar concentrations of tyramine were necessary to cause a delayed, short-lasting rise in [Ca2+]i. In addition to these Ca2+ signals, transfected cells also displayed a very small increase in intracellular cAMP concentration. The cAMP response, however, was only observed when high concentrations of octopamine were applied. These results suggest that the increase in [Ca2+]i is the primary cellular response to AmOA1 activation. It has been shown that heterologously expressed GPCRs can activate different intracellular signalling systems, depending on the cell line used for expression and the agonist used for receptor stimulation (Robb et al. 1994; Reale et al. 1997; Sidhu and Niznik 2000). In the case of AmOA1, however, we hypothesize that the massive increase in [Ca2+]i caused by stimulation with high concentrations of octopamine activates adenylyl cyclase in a secondary reaction. Nevertheless, the cellular cAMP response is minute. Therefore, AmOA1 should be considered a member of the native OCT1 receptor gene family, which is known to mediate increases in cellular Ca2+ concentration (Evans and Robb 1993; Roeder 1999).

A distinctive feature of the calcium responses observed in pcAmoa1-transfected HEK 293 cells is the occurrence of Ca2+ oscillations. Such oscillations have been observed in numerous cell types and represent universal intracellular signals (Thomas et al. 1996; Berridge et al. 2000). We observed the first Ca2+ oscillations when the transfected cells were stimulated with 50 nm octopamine. Raising the octopamine concentration resulted in an increase in the oscillation frequency (Fig. 3). Similar Ca2+ oscillations have been reported for a rat metabotropic glutamate receptor, mGluR5 (Kawabata et al. 1996). A homologous glutamate receptor (mGluR1) generated a single transient Ca2+ signal. The Ca2+ signal depended on the phosphorylation status of the receptor proteins. The mGluR5 receptor possesses a consensus site for phosphorylation by PKC in the C-terminus. This site is absent in the mGluR1 receptor. A phosphorylation site was introduced into the mGluR1 receptor by site-directed mutagenesis. Activation of the ‘mutated’ mGluR1 receptor led to Ca2+ oscillations that were very similar to those generated by mGluR5 (Kawabata et al. 1996). A potential PKC phosphorylation site is also present in the C-terminus of AmOA1. It will be interesting to test by site-directed mutagenesis whether phosphorylation of AmOA1 is necessary for shaping the Ca2+ signals.

Functional implications of the AmOA1 receptor

Octopamine is an important neuromodulator in insects. In the honeybee brain, five clusters of approximately 100 octopamine-immunoreactive somata have been identified (Kreissl et al. 1994). These neurones project into all regions of the brain and suboesophageal ganglion except for the mushroom body pedunculi and parts of the α- and β-lobes (Kreissl et al. 1994).

We investigated the expression pattern of the AmOA1 encoding mRNA in the bee brain. The gene is abundantly expressed in all major regions of the brain. This expression pattern differs substantially from the pattern reported for the Drosophila OAMB receptor gene (Han et al. 1998), although it is possible that the difference is quantitative rather than qualitative. OAMB-encoding mRNA is highly enriched in the somata of intrinsic mushroom body neurones of Drosophila. Furthermore, only scattered expression was detected in the ellipsoid body of the central complex and some somata of the medulla (Han et al. 1998).

In the honeybee and in Drosophila, the mushroom bodies are involved in olfactory learning and memory formation (Erber et al. 1980; De Belle and Heisenberg 1994; Menzel et al. 1994; Meller and Davis 1996). Specificity of gene expression in this neuropil might therefore indicate involvement of the receptor in learning processes (Han et al. 1998). In the bee, the specific role of octopamine during learning was demonstrated in the mushroom bodies by microinjection of octopamine, which can substitute for the US during olfactory PER conditioning (Menzel et al. 1994; Hammer and Menzel 1998). It is important to note that similar effects were found after octopamine injections into the antennal lobes, which also participate in olfactory memory formation (Erber et al. 1980; Hammer and Menzel 1998).

A number of experiments with microinjections have shown that octopamine can specifically modify behavioural and neuronal responses in different neuropils of the bee brain. Octopamine injections into the optic lobes can enhance visual antennal reflexes (Erber et al. 1993; Erber and Kloppenburg 1995) and change the properties of motion-sensitive interneurones (Kloppenburg and Erber 1995). Octopamine injections into the dorsal lobe can enhance the activity of antennal motoneurones whose somata and dendrites are located in this brain area (Pribbenow and Erber 1996).

The Amoa1 gene is expressed in cells of the mushroom bodies, the antennal lobes and the optic lobes. Therefore, the AmOA1 receptor is probably involved in the processing of olfactory and visual information. Activation of the heterologously expressed AmOA1 receptor leads to an increase in [Ca2+]i. Because a low ligand concentration is required to evoke Ca2+ signals, we hypothesize that the increase in [Ca2+]i is the primary cellular response mediated by AmOA1. In summary, the distribution pattern of the Amoa1 transcript and the functional properties of the heterologously expressed AmOA1 receptor strongly suggest that this receptor is involved in processing sensory information and in mediating higher-order brain functions in the honeybee.

Acknowledgements

We thank Dr S. Frings (Heidelberg) for advice and support in Ca2+ fluorimetry. We gratefully acknowledge the technical assistance of S. Balfanz and M. Bruns (Jülich) and of J. Buchholz (Technische Universität Berlin). For helpful suggestions on the manuscript we thank Dr R. Scheiner (Technische Universität Berlin). This study was supported by grants Ba 1441/2-2, 2-3 (to AB) and Bl 469/1-2 (to WB) from the Deutsche Forschungsgemeinschaft. LG was a PhD fellow of the Graduiertenkolleg GRK120 (Signalketten in lebenden Systemen).

Ancillary