Characterization of an Invertebrate-Type Dopamine Receptor of the American Cockroach, Periplaneta americana

We have isolated a cDNA coding for a putative invertebrate-type dopamine receptor (Peadop2) from P. americana brain by using a PCR-based strategy. The mRNA is present in samples from brain and salivary glands. We analyzed the distribution of the PeaDOP2 receptor protein with specific affinity-purified polyclonal antibodies. On Western blots, PeaDOP2 was detected in protein samples from brain, subesophageal ganglion, thoracic ganglia, and salivary glands. In immunocytochemical experiments, we detected PeaDOP2 in neurons with their somata being located at the anterior edge of the medulla bilaterally innervating the optic lobes and projecting to the ventro-lateral protocerebrum. In order to determine the functional and pharmacological properties of the cloned receptor, we generated a cell line constitutively expressing PeaDOP2. Activation of PeaDOP2-expressing cells with dopamine induced an increase in intracellular cAMP. In contrast, a C-terminally truncated splice variant of this receptor did not exhibit any functional property by itself. The molecular and pharmacological characterization of the first dopamine receptor from P. americana provides the basis for forthcoming studies focusing on the significance of the dopaminergic system in cockroach behavior and physiology.

(or octopamine/tyramine) receptors was introduced based on their similarities in structure and in signaling properties with mammalian adrenergic receptors [57]. INDRs and α-adrenergic-like octopamine receptors also seem to couple not only to cAMP, but also to Ca 2+ signaling, and to share pharmacological properties [58].
In cockroaches, information has been accumulated on the pharmacological properties of dopamine receptors in salivary glands and other tissues (for reviews, see: [12,21]). In contrast, comparatively little is known about the exact repertoire and molecular properties of aminergic receptors in P. americana [59][60][61], and, until this study, no molecular data on dopamine receptors have been available. Here, we show that the mRNA encoding an INDR, which we term PeaDOP2, is expressed in the brain and salivary glands of P. americana. Immunohistochemical analysis has shown that the receptor protein is present in specific brain neurons with their somas being located at the anterior edge of the medulla. When stably expressed in human embryonic kidney (HEK) 293 cells, PeaDOP2 promotes the formation of cAMP with an EC 50 of 160 nM for dopamine. The effect of dopamine is mimicked by (±)-2-amino-6,7-dihydroxy-1,2,3,4-tetrahydronaphthalene (6,7-ADTN) and (±)-6-Chloro-7,8-dihydroxy-3-allyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine (6-chloro-APB) and abolished by co-incubation with known dopamine receptor antagonists, e.g., chlorpromazine and cis(Z)-flupentixol. This study marks the first comprehensive molecular, pharmacological, and functional characterization of a dopamine receptor in the cockroach and furthers the understanding of biogenic amine signaling in this model insect.

Molecular Cloning and Sequence Analysis of a Dopamine Receptor from P. americana
Using degenerate oligonucleotide primers for highly conserved GPCR transmembrane (TM) regions 6 and 7, we amplified a cDNA fragment of 109 bp from P. americana brain cDNA coding for a putative dopamine receptor. Rapid amplification of cDNA ends (RACE) was undertaken with gene-specific primers in order to obtain the missing 5' and 3' parts of the putative Peadop2 cDNA (see Experimental Section). The full-length cDNA consists of 1815 nucleotides (Peadop2A; Accession No.: HG794355) and was independently amplified as a complete fragment by using two gene-specific primers. The longest open reading frame encodes a protein of 512 amino acids (PeaDOP2A) with a predicted molecular mass of 57.7 kDa. In addition, we identified a truncated splice variant of Peadop2A, which we named Peadop2B and which possessed an alternative C-terminal region. The nucleotide sequence of Peadop2B is 1470 bp in length (Accession No.: HG794356), and the longest open reading frame encodes a protein of 456 amino acids (51.1 kDa).
Analyses of the deduced amino acid sequences of both putative receptor variants with the topology predictor Phobius [62] revealed the expected hallmarks of GPCRs including structural features such as seven hydrophobic transmembrane domains, an extracellular N-terminus, and a cytoplasmic C-terminal region ( Figure 1). Several highly conserved sequence motifs and amino acid residues among the aminergic GPCRs are found in the PeaDOP2 receptor variants. The aspartate residue D 3.32 according to the nomenclature of Ballesteros and Weinstein [63] in TM3 (D 151 in PeaDOP2) is predicted to be relevant for binding the amine group of catecholamines such as dopamine, whereas serine residues in TM5 form hydrogen bonds with the hydroxyl groups of dopamine [64,65]. The characteristic C 6.47 -W 6.48 -x-P 6.50 -F 6.51 -F 6.52 motif in TM6 (C 404 WLPFF), which interacts with the aromatic ring of dopamine, is also present. The PeaDOP2A/B receptors possess the D 3.49 -R 3.50 -Y 3.51 (D 168 RY) sequence located at the end of TM3 involved in receptor activation (for a review, see: [19]). Furthermore, the conserved N 7.49 -P 7.50 -x-x-Y 7.53 (N 443 PxxY) motif in TM7, which is crucial for stabilizing the inactive conformation of the receptor, is present in both variants. Biogenic amine receptors are known to have numerous post-translational modifications. The N-terminal region of PeaDOP2A/B contains four consensus sites for N-linked glycosylation (N 2 GS, N 32 WS, N 49 FT, N 55 AS; Figure 1). One consensus site for phosphorylation by protein kinase C ([S/T]-x-[R/K]) occurs in CPL2 (S 184 VR) and four additional sites in CPL3 (S 264 LK, T 282 LR, S 357 TR and S 378 RK). The C-terminus of PeaDOP2A harbors three cysteine residues (C 463 VCC), which are highly conserved among INDRs and serve as potential targets for palmitoylation. In contrast, the short C-terminal region of PeaDOP2B does not contain any cysteine residues.

Figure 1.
Amino acid sequence alignment of PeaDOP2A and orthologous receptors from Drosophila melanogaster (DmDOP2A, NP_733299), Papilio xuthus (PxDOP2, BAD72870), and Apis mellifera (AmDOP2, AAM19330). Identical residues (≥75%) are shown as white letters against black, whereas conservatively substituted residues are shaded. Putative transmembrane domains (TM1-TM7) are indicated by gray bars. Potential N-glycosylation sites (▼), PKC phosphorylation sites (•), and putative palmitoylation sites (*) of PeaDOP2A are indicated. Underlined letters represent the region within the 3rd cytoplasmic loop from which the PeaDOP2-specific antigen was derived. R 456 ( † ) is the last amino acid residue present in the truncated variant PeaDOP2B. The amino acid position is given on the right.
A multiple amino acid sequence comparison within the conserved transmembrane domains and short linker regions of PeaDOP2A in invertebrate and human dopamine receptors was used to calculate a phylogenetic tree ( Figure 2). The PeaDOP2A receptor clustered with orthologous receptors of several insects and was incorporated into the branch of the INDRs.

Tissue Distribution of Peadop2A and Peadop2B mRNA
The expression pattern of Peadop2A and Peadop2B mRNA in various tissues of P. americana was investigated by RT-PCR with specific primers corresponding to sequences within the differing C-termini. Transcripts of Peadop2A and Peadop2B were detected in samples of the brain and salivary glands ( Figure 3). Conversely, no receptor mRNA expression was detected in samples of Malpighian tubules, midgut, and leg muscle ( Figure 3). To ensure that the fragments were not amplified from genomic DNA, samples were treated with DNase I. When tissue samples were treated additionally with an RNase cocktail, no PCR product could be amplified (data not shown).

Generation of an Anti-PeaDOP2 Antibody and Immunohistochemical Localization of PeaDOP2A/B Receptors
We generated a polyclonal antiserum directed against a part of the third cytoplasmic loop of PeaDOP2A/B receptors. Western blots of membrane proteins isolated from P. americana brain, subesophageal ganglion, thoracic ganglia, and salivary glands showed that the antibody recognized two bands of ~55 and ~46 kDa ( Figure 4). These molecular weights are in accordance with the predicted molecular weights of 57.7 and 51.1 kDa for PeaDOP2A and PeaDOP2B, respectively. No bands were detected in protein samples isolated from the abdominal ganglia and the terminal ganglion ( Figure 4). After pre-absorption with the peptide (15 µg/mL) used for immunization, the signal was completely lost (data not shown) supporting the specificity of the antibody for the PeaDOP2A and PeaDOP2B proteins.
To investigate the cellular distribution of the PeaDOP2 receptors, cryosections of P. americana brains were examined with these antibodies ( Figure 5). In frontal sections of anterior brain regions, specific labeling was found in a network of fibers in the ventral protocerebrum and in commissures running ventrally to the mushroom bodies ( Figure 5A). In sections of central brain parts, prominent labeling also occurred in large parts of the optic lobes. A group of PeaDOP2-immunoreactive somata was located ventral to the medulla. These neurons sent projections radially via the medulla (not shown) to the lamina ( Figure 5B). Additional fibers projected to the ventral protocerebrum and the optic commissure ( Figure 5A,B). Alignments were performed with BioEdit (version 7.0.5) by using the core amino acid sequences lacking the variable regions of the amino and carboxy terminus and the 3rd cytoplasmic loop. The genetic distance was calculated with MEGA4. The receptor sequences, followed by their accession numbers, are listed in the order illustrated:  shown on the left. Detection of PCR products amplified on total RNA isolated from brain, salivary glands, Malpighian tubules, midgut, and leg muscle. Amplification failed when samples were digested with an RNase cocktail prior to RT-PCR (data not shown). The lower panel shows RT-PCR products amplified with actin-specific (Accession No. AY116670) primers as a control.

Functional Analyses of PeaDOP2A/B Receptors in HEK 293 Cells
HEK 293 cell lines constitutively expressing PeaDOP2A or PeaDOP2B receptors were generated in order to examine their second messenger coupling and pharmacological properties. Expression of PeaDOP2A/B was confirmed by Western blotting and immunocytochemistry. The ligand specificity of the receptors was tested by the application of various biogenic amines (serotonin, dopamine, tyramine, octopamine, and histamine; 10 µM). Only dopamine significantly stimulated cAMP production in PeaDOP2A-expressing cells (300 pmol cAMP/mg protein; Figure 6A). None of the biogenic amines induced a change of the intracellular cAMP level ([cAMP] i ) in the cell line stably expressing the PeaDOP2B receptor ( Figure 6A). In PeaDOP2A-expressing cells, the dose-dependent effect of dopamine on [cAMP] i was analyzed with concentrations ranging from 1 nM to 10 µM ( Figure 6B). The dopamine effect was concentration-dependent and saturable, resulting in a sigmoidal dose-response curve ( Figure 6B). Half-maximal stimulation of cAMP production (EC 50 ) was achieved at a dopamine concentration of 160 nM (logEC 50 = −6.78 ± 0.04, mean ± SEM). Maximal stimulation of cAMP synthesis was achieved at dopamine concentrations of ≥1 µM. None of the dopamine concentrations showed an effect on [cAMP] i in PeaDOP2B-expressing cells ( Figure 6B) or in non-transfected cells.

Discussion
Aminergic receptors constitute a subfamily of rhodopsin-like GPCRs. The biogenic amine dopamine is found in deuterostomes and in protostomes and regulates and modulates numerous physiological functions and behaviors. Until now, three aminergic receptors have been cloned and experimentally examined from P. americana: an octopamine receptor [59], a tyramine receptor [60], and a serotonin receptor [61]. No dopamine receptor has been molecularly characterized to date. Here, we have cloned the cDNA encoding a dopamine receptor from P. americana, pharmacologically characterized the receptor, and investigated its expression profile.

Receptor Variants Occur by Alternative Splicing of the Peadop2 Gene
Splice variants have been described for a series of aminergic GPCRs, including dopamine receptors from Caenorhabditis elegans [68], the crustacean Panulirus interruptus [69], and the insect D. melanogaster [50]. Splice variants differ in their tissue distribution, ligand-binding properties, G-protein coupling, and activation of signaling pathways (for a review, see: [70]). Ono and Yoshikawa [66] have shown that the genomic structure of INDRs is highly conserved within the coding region. A large part of the receptor protein ranging from the N-terminus, up to and including TM7, is encoded by exon 1, whereas the C-terminal region is encoded by exons 2 and 3. The PeaDOP2 variants identified in our study also belong to the INDR group of dopamine receptors (see below). Interestingly, a potential splice site is present at the site of divergence of the two isoforms. The Peadop2B variant most likely arises by retention of the intron following exon 1; this leads to an in-frame stop codon. Likewise, variants that possess truncated C-termini have been described in the orthologous receptor of the cat flea C. felis [67]. The two variants of the DOP2 receptor from D. melanogaster [53,54] also differ in their C-terminal sequence. Here, the difference arises in exon 3. Unfortunately, functional effects of the truncated C-terminus in C. felis or the differing C-termini in D. melanogaster DOP2 receptors have not been comparatively analyzed. Nevertheless, the high conservation of splice sites is in favor of a common ancestor of the INDR genes and also argues for high selection pressure retaining the functional variability by alternatively spliced INDR proteins.

Structural Characteristics of the PeaDOP2 Receptor
Based on bioinformatics, PeaDOP2 clearly belongs to the class of INDRs according to the nomenclature of Mustard et al. [19] (Figure 2). This dopamine receptor class is more closely related to α-adrenergic-like octopamine receptors than to mammalian D1-like receptors [19,58]. Hauser et al. [71] have postulated that, during the evolution of aminergic GPCRs, new receptors evolved that needed new ligands. Because of structural constraints, one way in which to gain such ligands was to "borrow" them from related systems. This caused so-called "ligand-hops" between receptor families. Since INDRs are not only present in hemimetabolous (this work) and holometabolous [53][54][55][56]66,67,72] insects, but also in crustaceans [73], arachnids [74], and nematodes [68], this receptor class seems to be common for the Ecdysozoa group of protostome animals. Interestingly, the cephalochordate amphioxus (Branchiostoma floridae), which represents the most basal of the chordates, also expresses a receptor, AmphiAmR2, showing close structural and pharmacological similarities to the INDRs [75].
The amino acid sequence of PeaDOP2 displays characteristic properties of amine-activated GPCRs in general and INDRs in particular [19]. The N-terminal region of the PeaDOP2 receptor contains four consensus sites for N-linked glycosylation (Figure 1). At least one of these sites is glycosylated when the protein is heterologously expressed in HEK 293 cells as monitored by PNGase F treatment and Western blotting (data not shown).
PeaDOP2 contains four consensus sites for phosphorylation by protein kinase C. Thus, phosphorylation of PeaDOP2 can potentially lead to the uncoupling of the receptor from G-protein-dependent signaling pathways and/or the activation of G-protein-independent pathways [76]. The C-terminal region of PeaDOP2A contains three cysteine residues that allow palmitoylation and, thus, formation of a fourth CPL. Palmitoylation has previously been demonstrated for several human dopamine receptors [77][78][79]. The high degree of conservation of these cysteine residues in INDR sequences ( Figure 1) argues in favor of their functional significance for, e.g., transport to the cell membrane and localization in lipid microdomains [77,80], optimal G-protein interaction [81,82], desensitization, and internalization processes [83], and possibly also receptor oligomerization [84]. In PeaDOP2B, these cysteine residues are lacking.
Amino acid residues that are essential for dopamine binding are conserved in PeaDOP2. An aspartic acid in TM3 (D 3.32 ; D 151 in PeaDOP2) acts as a counter-ion for the protonated amino group of dopamine. Serine in TM5 forms hydrogen bonds with the catechol hydroxyl groups [64]. A tryptophan (W 405 ) and a phenylalanine (F 409 ) residue within the sequence motif C 6.47 -W 6.48 -x-P 6.50 -F 6.51 -F 6.52 in TM6 interact with the aromatic ring of the ligand. This interaction modulates the bend angle of TM6 around the highly conserved proline kink at Pro 6.50 , leading to a rotation or tilting movement of the cytoplasmic end of TM6 upon activation (for a review, see: [85]). The D 3.49 -R 3.50 -Y 3.51 sequence at the start of CPL2, which is believed to play a key role in receptor activation (for a review, see: [85]), is also present in PeaDOP2 (D 168 RY). The signature motif N 7.49 -P 7.50 -x-x-Y 7.53 in TM7 of GPCRs is additionally present in PeaDOP2 (N 443 PVIY). This motif is crucial for internalization, sequestration, ligand affinity, stabilization of the active conformation and thus receptor activation, and receptor signaling as revealed by numerous mutagenesis studies [86][87][88][89].

PeaDOP2A but Not PeaDOP2B Is a Functional Dopamine Receptor
Activation of PeaDOP2A with dopamine leads to an increase in [cAMP] i , whereas tyramine, octopamine, serotonin, and histamine have no effect on [cAMP] i . The EC 50 value of 160 nM for the dopamine-induced cAMP increase is similar to that of the D. melanogaster DOP2 receptor (350 nM, [54] [53,58,91]. In summary, PeaDOP2A seems to couple exclusively to G s proteins, whereas PeaDOP2B does not couple to any of the classical G-protein-mediated signaling pathways and, therefore, might be considered as a non-functional receptor. However, our immunocytochemical analyses have shown that PeaDOP2B is translated and transported to the plasma membrane in HEK 293 cells (data not shown). Studies of various aminergic receptors suggest that truncated receptor variants might associate with full-length receptors and thereby modulate the binding and/or signal transduction properties in such heteromeric complexes. This has been shown, for example, for the D2-like receptor DOP-3 and its truncated splice variant DOP-3nf in C. elegans [68]. Because of an in-frame stop codon in the third intracellular loop, DOP-3nf lacks TM6 and TM7 of the full-length DOP-3 receptor [68]. DOP-3 attenuates forskolin-stimulated cAMP formation in response to dopamine stimulation, whereas DOP-3nf does not [68]. When DOP-3 was co-expressed with DOP-3nf, the ability to inhibit forskolin-stimulated cAMP formation was reduced [68]. Similar observations have been reported for human dopamine receptors and have been implicated in disease states such as schizophrenia [92][93][94][95][96]. Whether PeaDOP2B assembles and impairs PeaDOP2A function will be investigated in a forthcoming investigation and is beyond the scope of this study.

Pharmacological Properties of the PeaDOP2A Receptor
Early studies have shown that the pharmacological properties of dopamine receptors in the insect brain differ remarkably from those of their vertebrate counterparts [97,98]. This observation has been confirmed in detail for heterologously expressed dopamine receptors (for a review, see: [19]). In PeaDOP2A-expressing cells, treatments with non-selective dopamine receptor agonists (6,7-ADTN and apomorphine) and the D1-like receptor agonist 6-chloro-APB lead to an increase in [cAMP] i with a rank order of potency: 6,7-ADTN > dopamine > 6-chloro-APB > apomorphine. These agonists act similarly on INDRs from various insect species such as D. melanogaster (dopamine > 6-chloro-APB), A. mellifera (6,7-ADTN > 6-chloro-APB > dopamine > apomorphine), and B. mori (6,7-ADTN > dopamine > apomorphine) [53,56,90]. In contrast, the prototypical D1-like dopamine receptor partial agonist SKF 38393 does not show any effect on [cAMP] i in PeaDOP2A-expressing cells. This corresponds well to the lack of activity of SKF 38393 on all INDRs in insects [53,56,90]. The selective D2-like dopamine receptor agonist bromocriptine fails to increase [cAMP] i in PeaDOP2A-expressing cells. Since bromocriptine has been shown to be a potent agonist at insect D2-like receptors [37,50,52], it can be used pharmacologically to distinguish D2-like receptors from D1-like receptors and INDRs in insects.

Distribution of the PeaDOP2 Receptor in the Nervous System
Peadop2 expression in the brain was examined by RT-PCR, Western blot analysis, and immunohistochemistry. At the mRNA level, both Peadop2A and Peadop2B were detected in the brain (Figure 3). An anti-PeaDOP2-antibody was raised against a region of the CPL3 and recognized both receptor variants. On Western blots containing membrane proteins from brain, subesophageal ganglion, and thoracic ganglia, two protein bands were immunostained (Figure 4) arguing in favor of the translation of both receptor proteins, i.e., PeaDOP2A/B. However, as long as no variant specific antibodies are available, it cannot be unequivocally ruled out that the occurrence of two protein bands is due to protein modification (e.g., glycosylation, degradation) or dimerization.
In P. americana brain sections, the anti-PeaDOP2 antibody labels a cluster of neurons in the anterior border area of the medulla, a network of fibers in the optic lobes, and fibers in the ventrolateral protocerebrum and the optic commissures ( Figure 5). In the locust Schistocerca gregaria, the somata of 18-20 dopamine-immunoreactive tangential neurons (DTan) are clustered at the anterior edge of the medulla [28]. The somata of these neurons are clustered at the anterior edge of the medulla. Their axons project via the first optic chiasma to the posterior edge of the lamina and give rise to dense arborizations in the proximal layer of the lamina with some processes extending into the distal layer [28]. Since the DTan neurons in S. gregaria and the PeaDOP2-immunoreactive neurons in P. americana share similar distribution patterns, future double-labeling experiments with the anti-dopamine and anti-PeaDOP2 antibodies would be worthwhile. A co-distribution of the two antigens would suggest an autoreceptor function for PeaDOP2, as has been described for mammalian dopamine receptors.
The location of the PeaDOP2-immunoreactive cells between the central brain and the optic lobe is also reminiscent of the location of PDF (peptide dispersing factor)-expressing ventral lateral neurons (LNvs) described by Stengl and Homberg [99] in the cockroach Rhyparobia maderae. These neurons are members of the circadian clock circuitry. However, to determine whether PeaDOP2 immunoreactive cells are indeed PDF-expressing clock neurons, double-labeling experiments with anti-PDF antibodies need to be performed. Co-expression of PeaDOP2 and PDF would suggest that dopamine acting via PeaDOP2 modulates PDF-positive LNvs and thus participates in the circadian behavior of cockroaches.
The Peadop2 expression pattern in the brain is clearly distinct from that of orthologous INDRs in the holometabolous insects D. melanogaster, A. mellifera, and P. xuthus, as their INDRs are predominantly expressed in the mushroom bodies of the adult brain [54,55,72].

Possible Function of the PeaDOP2A Receptor in Saliva Generation and Secretion
The salivary gland of P. americana is an established organotypic model for the investigation of the cellular actions of biogenic amines, since its secretory activity is regulated and modulated by dopamine and serotonin (for a recent review, see: [12]). Cockroaches have innervated acinar-type salivary glands with the secretory acini consisting of three cell types: peripheral cells (p-cells), central cells (c-cells), and centroacinar cells. Dopaminergic fibers project to the peripheral cells and salivary duct cells [30,100]. In isolated glands, dopamine leads to the secretion of protein-free saliva by activating the peripheral cells [7]. In addition, dopamine affects the modification of primary saliva by stimulating salivary duct cells [101]. RT-PCR and Western blotting have revealed high expression levels of PeaDOP2 in the salivary glands. The pharmacological profile of dopamine-induced salivary secretion has been investigated in detail by Marg et al. [8]. Notably, the order of potency for agonists (6,7-ADTN > dopamine > 6-chloro-APB >> SKF 38393 = ineffective) and antagonists (flupentixol > chlorpromazine ≥ butaclamol) is the same as for the expressed PeaDOP2A receptor. Thus, the INDR, PeaDOP2, is a likely candidate in the mediation of the effects of dopamine on salivary secretion in cockroaches, although whether a true D1-like receptor is also involved in this process remains to be examined.
In the migratory locust Locusta migratoria, dopamine induces hyperpolarization of salivary gland acinar cells [43] and salivary secretion [44]. The authors postulated that these effects are mediated via D1-like receptors [43,44]. Based on pharmacological experiments alone, however, it is difficult to decide whether a true D1-like receptor or an INDR mediates the effect of dopamine in the locust salivary gland. Dopamine also potently induces salivary secretion in isolated tick (Ixodes scapularis) salivary glands (for a recent review, see: [102]). Transcripts of a DOP1 receptor were found in salivary glands of I. scapularis [103]. Immunohistochemistry for the tick DOP1 receptor revealed scattered patches of reactivity in the junctions between cells on the luminal surface of specific secretory acini [102,103]. Recently, also an INDR was identified in I. scapularis [73] and found to be expressed in salivary glands, too [102].

Multiple Sequence Alignment and Phylogenetic Analysis
Amino acid sequences used for phylogenetic analyses were identified by protein-protein BLAST searches of the NCBI database with the deduced amino acid sequence of Peadop2A (PeaDOP2A) as "bait". Multiple sequence alignments of the complete amino acid sequences were performed with ClustalW. Values for identity (ID) and similarity (S) were calculated by using the BLOSUM62 substitution matrix in BioEdit 7.0.5 [105]. MEGA 4 [106] was used to calculate the genetic distances between the core sequences and to construct neighbor-joining trees with 2000-fold bootstrap re-sampling. The D. melanogaster ninaE-encoded rhodopsin 1 and FMRFamide receptor were used as outgroups.

Antibody Production and Purification
Antibodies were raised against a maltose-binding protein (MPB) containing a specific region of the 3rd cytoplasmic loop (CPL3) of PeaDOP2 (see Figure 1), which was constructed as follows. The CPL3-specific fragment was amplified by PCR with AB-F (5'-AAAGAATTCGCAGTAAT CCAGACG-3') and AB-R (5'-TTTGTCGACTCAGTGATGTGGAGGAGACAT-3') as the sense and antisense primer, respectively. The fragment was cloned into the pMAL-c2X vector (New England Biolabs, Frankfurt, Germany). The fusion protein was over-expressed in E. coli BL21 and purified by amylose affinity-chromatography. Anti-PeaDOP2 polyclonal antiserum was raised in rabbits (Pineda-Antikörper-Service, Berlin, Germany). For purification of crude anti-PeaDOP2-antiserum, the CPL3-specific fragment of PeaDOP2 was cloned into the pET-30a vector (Novagen, Darmstadt, Germany) and over-expressed. The purified protein was coupled to a HiTrap NHS-activated High Performance column (GE Healthcare, Freiburg, Germany). Antibodies from 50 mL crude antiserum were affinity purified as described previously [60].

Western Blot Analysis
Cockroach brains were homogenized in 150 µL Roti-load sample buffer (Roth, Karlsruhe, Germany) and incubated for 5 min at 95 °C. Alternatively, membrane proteins were isolated as described previously [61] and incubated for 5 min at 60 °C. Proteins were separated by 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Approximately 10 µg protein, as determined by a modified Bradford assay (Roti-Nanoquant, Roth, Karlsruhe, Germany), was run per lane and then transferred to polyvinylidene fluoride membranes (PVDF, Roth, Karlsruhe, Germany). PVDF membranes were labeled with antibodies as described previously [61] by applying anti-PeaDOP2 (1:8000) followed by incubation with a secondary antibody conjugated to horseradish peroxidase (1:20,000; American Qualex, La Mirada, CA, USA). For controls, antibodies were pre-absorbed with the fusion protein (15 µg/mL) used for immunization. Immunoreactivity was detected by enhanced chemifluorescence.

Immunofluorescence Staining of Brain Sections
Dissected brains were fixed for 2 h in 3% paraformaldehyde, 0.075 M lysine HCl, 0.01 M sodium periodate, 0.2 M sucrose in 0.1 M phosphate buffer (pH 7.0). For immunofluorescence staining, preparations were rinsed in phosphate-buffered saline (PBS), transferred to 10% (w/v) sucrose in PBS for 1 h, incubated overnight in 25% sucrose in PBS at 4 °C, and frozen in melting isopentane (−155 °C). Sections (20 µm) were cut at −30 °C on a cryostat and collected on poly-L-lysine-coated cover-slips. The sections were successively incubated in 0.01% Tween 20 in PBS, 50 mM NH 4 Cl in PBS, PBS, and blocking solution containing 1% normal goat serum, 0.8% bovine serum albumin (BSA), and 0.5% Triton X-100 in PBS, with each step lasting 5 min at room temperature. Sections were then incubated overnight at 4 °C with anti-PeaDOP2 (1:800) diluted in blocking solution, washed for 3 × 10 min in PBS, and incubated with secondary antibodies for 1 h at room temperature. Sections were washed again for 3 × 10 min in PBS and mounted in Mowiol 4.88 (Hoechst, Frankfurt, Germany) containing 2% n-propyl gallate. Fluorescent images were recorded with a Zeiss LSM 510 confocal laser-scanning microscope (Carl Zeiss, Jena, Germany).

Construction of pcPeadop2A/B-HA Expression Vector
Expression-ready constructs of Peadop2A and Peadop2B cDNA were generated by PCR. To monitor transfection efficiency and receptor protein expression, a hemagglutinin (HA) epitope tag was engineered to the 3' end of each cDNA. First-round PCR was performed with a sense primer 5'-GATTAAGCTTCCACCATGAACGGAAGCCTAGCAG-3' and the antisense primers 5'-TGGGACGTCGTATGGGTACATCGAGTAGAGTTCGTGTTG-3' (for Peadop2A) or 5'-TGGGACGTCGTATGGGTACCTTCTGAAATCTCGACTCC-3' (for Peadop2B). In second-round PCR experiments, the same sense primer was used in combination with the antisense primer 5'-TTTGGATCCTTAAGCGTAGTCTGGGACGTCGTATGGG-3'. PCR products were digested with Bam HI and Hind III and sub-cloned into pcDNA3.1(+) vector (Invitrogen, Karlsruhe, Germany) yielding pcPeadop2A-HA and pcPeadop2B-HA. The correct insertion was confirmed by DNA sequencing.

Functional Expression of the PeaDOP2A/B-HA Receptor
Approximately 8 µg pcPeadop2A/B-HA vector was separately introduced into exponentially growing HEK 293 cells (~4 × 10 5 cells per 5 cm Petri dish; Greiner, Frickenhausen, Germany) by a modified calcium phosphate method [107]. Stably transfected cells were selected in the presence of the antibiotic G418 at 0.8 mg/mL. Isolated foci were propagated and analyzed for the expression of PeaDOP2A/B-HA by immunocytochemistry and Western blot analysis with a commercial anti-HA antibody (anti-HA high affinity; Roche, Penzberg, Germany).

Functional Characterization of PeaDOP2A and PeaDOP2B Receptors
Assays to determine the ability of PeaDOP2A/B-HA receptors to stimulate adenylyl cyclase activity were performed as described earlier [108]. PeaDOP2A/B-expressing cells were grown in minimal essential medium with GlutaMAX™ (Invitrogen, Karlsruhe, Germany), 10% (v/v) fetal bovine serum, 1% (v/v) non-essential amino acids, and antibiotics (all from Invitrogen, Karlsruhe, Germany). Cells were incubated with ligands for 30 min at 37 °C in the presence of the phosphodiesterase inhibitor isobutylmethylxanthine (IBMX; Sigma, Taufkirchen, Germany; final concentration 100 µM) and lysed by adding 0.5 mL ice-cold ethanol. After 1 h at 4 °C, the lysate was transferred to a reaction tube and lyophilized. The amount of cAMP produced was determined by using the TRK 432 cyclic AMP assay kit (GE Healthcare, Freiburg, Germany). Data were analyzed and displayed by using PRISM 4.01 software (GraphPad, San Diego, CA, USA).

Conclusions
Cockroaches are established model organisms for studying circadian rhythms [4], insect learning [16], and the aminergic control and modulation of salivary secretion [12]. The pharmacological characterization and tissue localization of the first dopamine receptor provides the basis for forthcoming studies examining its role in cockroach behavior and physiology. Interference with PeaDOP2 expression by applying the RNAi technique or receptor activation with specific PeaDOP2 ligands can be used to unravel its contribution to rhythmic behavioral patterns and memory formation in this insect.