Dβ3, an atypical nicotinic acetylcholine receptor subunit from Drosophila: molecular cloning, heterologous expression and coassembly


Address correspondence and reprints requests to Dr Neil S. Millar, Department of Pharmacology, University College London, Gower Street, London WC1E 6BT, UK. E-mail: n.millar@ucl.ac.uk


Insect nicotinic acetylcholine receptors (nAChRs) play a central role in mediating neuronal synaptic transmission and are the target sites for the increasingly important group of neonicotinoid insecticides. Six nicotinic acetylcholine receptor (nAChR) subunits (four α-type and two β-type) have been cloned previously from the model insect species Drosophila melanogaster. Despite extensive efforts, it has not been possible to generate functional recombinant nAChRs by heterologous expression of any combination of these six subunits. It has, however, been possible to express functional hybrid receptors when Drosophilaα subunits are co-expressed with vertebrate β subunits. This has led to the assumption that successful heterologous expression might require an, as yet, uncloned β-type insect subunit. Examination of the recently completed Drosophila genomic sequence data has identified a novel putative nAChR β-type subunit. Here we report the molecular cloning, heterologous expression and characterization of this putative Drosophila nAChR subunit (Dβ3). Phylogenetic comparisons with other ligand-gated ion channel subunit sequences support its classification as a nAChR subunit but show it to be a distantly related member of this neurotransmitter receptor subunit family. Evidence that the Dβ3 subunit is able to coassemble with other Drosophila nAChR subunits and contribute to recombinant nAChRs has been obtained by both radioligand binding and coimmunoprecipitation studies in transfected Drosophila S2 cells.


mAb, monoclonal antibody

nAChR, nicotinic acetylcholine receptor. Abbreviations used for Drosophila nAChR subunits (gene nomenclature suggested by FlyBase is given in parenthesis) are as follows

ALS, α-like subunit (nAcRα-96Aa)


acetylcholine receptor of Drosophila (nAcRβ-64B)


Drosophilaα2 subunit/second α subunit (nAcRα-96Ab)


Drosophilaα3 subunit (nAcRα-7E)


Drosophilaα4 subunit (nAcRα-80B)


Drosophilaβ3 subunit (nAcRβ-21C)


second β-like subunit of Drosophila (nAcRβ-96A).

Nicotinic acetylcholine receptors (nAChRs) play a central role in mediating fast synaptic transmission in the insect brain (reviewed by Gundelfinger and Schulz 2000). They are also the target site for the economically important class of neonicotinoid insecticides (Narahashi 1996; Tomizawa et al. 1999). These factors have prompted an extensive investigation of the molecular and functional properties of insect nAChRs. Recent studies of both native nAChRs and cloned nAChR subunits have been aimed at identifying the subunit composition of insect nAChRs. Substantial progress has been made in identifying the subunit composition of insect nAChRs using immunoprecipitation studies with native nAChRs from Drosophila (Schloss et al. 1991, 1992; Chamaon et al. 2000; Schulz et al. 2000). In contrast, a detailed study of cloned nAChR subunits has been hindered by difficulties which have been encountered in generating functional insect nAChRs in heterologous expression systems (Millar 1999; Sivilotti et al. 2000).

Six nAChR subunits have been identified and cloned from the fruit fly Drosophila melanogaster (Hermans-Borgmeyer et al. 1986; Bossy et al. 1988; Baumann et al. 1990; Sawruk et al. 1990a,b; Schulz et al. 1998; Lansdell and Millar 2000a). Four have been classified as α subunits (ALS, Dα2/SAD, Dα3 and Dα4) and two as non-α- or β subunits (ARD and SBD). To date, no combination of these Drosophila nAChR subunits has enabled the successful expression of recombinant nAChRs. Despite these problems, however, it has been possible to detect functional nAChRs and/or high-affinity binding of nicotinic radioligands when any of the four Drosophilaα subunits are co-expressed with vertebrate non-α subunits (Bertrand et al. 1994; Lansdell et al. 1997; Schulz et al. 1998; Lansdell and Millar 2000a,b). Similar results have been reported with nAChR α subunits cloned from other insect species (Huang et al. 1999, 2000). Such studies have led to the assumption that problems encountered in heterologous expression of insect nAChRs may be due to a requirement for an, as yet unidentified, insect β-type subunit (Millar 1999). The recently completed sequencing of the Drosophila genome (Adams et al. 2000) has provided an opportunity to identify additional genes encoding potential nAChR subunits. Recent analysis of the Drosophila genome sequence data identified a single novel putative nAChR β subunit gene which encodes a protein which is distantly related to other nAChR subunits (Littleton and Ganetzky 2000). In this study we have characterized this putative nAChR β subunit gene by isolation of a full-length cDNA clone and its heterologous expression in a Drosophila cell line. Direct evidence that the Dβ3 subunit is able to coassemble with other Drosophila nAChR subunits and contribute to recombinant nAChRs has been obtained by both radioligand binding and coimmunoprecipitation studies.

Materials and methods

CDNAs, antibodies and cell lines

Rat nAChR subunit cDNAs (α4 and β2) were provided by Dr Jim Patrick, Baylor College of Medicine. Drosophila nAChR subunit cDNAs (ALS, ARD, Dα2/SAD, Dα3 and SBD) were provided by Drs Heinrich Betz and Bertram Schmitt, Frankfurt. The Dα4 cDNA, isolated in this laboratory, has been described previously (Lansdell and Millar 2000a). Subcloning of α4, β2, ALS, ARD, Dα2/SAD, Dα3, Dα4 and SBD subunit cDNAs into the Drosophila expression vector pRmHa3 (provided by Dr Thomas Bunch, University of Arizona) has been described previously (Lansdell et al. 1997; Lansdell and Millar 2000a,b). The Drosophila EST clone SD09326, isolated by the Berkeley Drosophila Genome Project (BDGP), was obtained from Research Genetics Inc. (Huntsville, AL, USA). Monoclonal antibody mAbFLAG-M2, raised against an eight amino acid epitope-tag (Hopp et al. 1988), was obtained from Sigma. Monclonal antibody mAb290, raised against the rat β2 subunit (Whiting and Lindstrom 1988), was provided by Dr Jon Lindstrom (University of Pennsylvania). Schneider's Drosophila S2 cells (Schneider 1972) were provided by Dr Thomas Bunch, University of Arizona.

Molecular cloning

EST clone SD09326 (described above) in plasmid pOT2 was digested with EcoRI and XhoI and the cDNA insert subcloned into EcoRI and SalI sites within the Drosophila expression vector␣pRmHa3 (Bunch et al. 1988) to create plasmid pRmHa3-Dβ3+intron3. PCR primers were designed to sequences within exons 3 and 4 of the GadFly predicted gene sequence CG11822 (http://www.fruitfly.org/annot/index.html) and were used to amplify a 225-bp fragment from two different Drosophila cDNA libraries: a Drosophila 0–5 day head λZAPII cDNA library (provided by Dr␣Ron Davis, Baylor College of Medicine, TX) and a Drosophila adult λgt10 cDNA library (Clontech). The fragments were subcloned into pCRII (Invitrogen) and sequenced. The cDNA fragment derived from the λZAPII cDNA library was digested with BclI and EcoNI and subcloned into the same sites within pRmHa3-Dβ3+intron3 to create pRmHa3-Dβ3. A 3′ Dβ3 cDNA fragment was amplified from the λZAPII cDNA library with a forward primer located within the Dβ3 coding sequence and a reverse primer to the T3 promoter site of the λZAPII vector.

Construction of epitope-tagged subunits

Drosophila nAChR subunits were modified by introducing the eight amino acid (D-Y-K-D-D-D-D-K) FLAG epitope (Hopp et al. 1988) into a unique restriction enzyme site within the predicted large intracellular loop. Oligonucleotides encoding the FLAG epitope were introduced into an NcoI site in the ALS and ARD subunit cDNAs to (create plasmids pRmHa3-ALSFLAG and pRmHa3-ARDFLAG), into an SphI site in SAD and SBD (to create pRmHa3-SADFLAG and pRmHa3-SBDFLAG), into a BstEII site in Dα3 (to create pRmHa3-Dα3FLAG), into a Bsu36I site in Dα4 (to create pRmHa3-Dα4FLAG) and into a BsiWI site in Dβ3 (to create pRmHa3-Dβ3FLAG).

Heterologous expression

Schneider's Drosophila S2 cells (Schneider 1972) were grown in Shields and Sang M3 medium (Sigma) containing 12.5% heat-inactivated fetal calf serum, 100 units/mL penicillin and 100 µg/mL streptomycin (Life Technologies) at 25°C. Exponentially growing S2 cells were transfected by a modified calcium phosphate method as described previously (Millar et al. 1994; Lansdell et al. 1997). Cells were transiently transfected with plasmid pRmHa3 containing appropriate nAChR subunit cDNAs. Expression of nAChR subunit cDNAs from the metallothionein promoter of pRmHa3 was induced by the addition of CuSO4 (0.6 mm) for 24 h.

Radioligand binding

Radioligand binding to transfected S2 cells was performed as described previously (Lansdell and Millar 2000b). Care was taken to ensure that the number of receptor binding sites (either in cell homogenates or whole cell preparations) used for binding studies was low enough to avoid significant (> 10%) ligand depletion at low concentrations of radioligand. Preliminary experiments were conducted to ensure that incubation times were long enough to enable radioligand binding to reach equilibrium. Equilibrium radioligand binding data were analysed by the cvfit program (David Colquhoun, University College London; http://www.ucl.ac.uk/Pharmacology/dc). [3H]Epibatidine (48 Ci/mmol) was purchased from NEN Life Science Products, [125I]α-bungarotoxin (200 Ci/mmol) was purchased from␣Amersham Pharmacia Biotech, [3H]methyl-carbamylcholine (80 Ci/mol) was purchased from American Radiolabeled Chemicals.

Metabolic labelling and immunoprecipitation

Drosophila S2 cells were metabolically labelled essentially as described previously (Millar et al. 1995). After growth in methionine-free medium for 15 min, cells were labelled with 250 µCi ‘Pro-mix’ (Amersham Phamacia Biotech), a mixture of [35S]methionine and [35S]cysteine, in 1.5 mL methionine-free medium for 3 h. Medium containing 30 mg/L methionine and 10% heat-inactivated fetal calf serum was then added and the cells incubated for a further 90 min. Cells were washed three times with 10 mL phosphate-buffered saline and harvested into 300 µL ice-cold lysis buffer (150 mm NaCl, 50 mm Tris−Cl pH 8.0, 5 mm EDTA and 1% Triton X-100) containing protease inhibitors (0.2 mm phenylmethylsulfonyl fluoride, 2 mmN-ethylmaleimide and 10 µg/mL, each, of leupeptin, apoprotinin and pepstatin). Solubilization and all subsequent steps were performed at 4°C. After 1 h solubilization the cell lysate was pre-cleared by incubation overnight with 30 µL pre-washed protein G−Sepharose (Amersham Pharmacia Biotech) in a 1 : 1 mixture with lysis buffer. Non-solubilized material was pelleted by centrifugation at 14 000 g for 15 min. Cell lysates were incubated with mAbFLAG-M2 or mAb290 (specific for rat β2) for 3.5 h. The antibody−receptor complex was immunoprecipitated by the addition of 35 µL protein G−Sepharose, incubated for a further 3.5 h and isolated by centrifugation. Samples were washed with 4 × 1 mL lysis buffer. Samples were examined by sodium dodecyl sulfate−polyacrylamide gel electrophoresis followed by autoradiography as described previously (Lansdell et al. 1997).


Molecular cloning of Dβ3

The Genome Annotation Database of Drosophila (GadFly) project (see http://www.fruitfly.org/annot/index.html) has identified a putative gene (CG11822) and predicted transcript (CT33131) which exhibits sequence similarity to previously identified Drosophila nAChR subunits (Littleton and Ganetzky 2000). Comparison of the predicted amino acid sequence of CT33131 to the Swiss-Prot database using the␣blast algorithm (http://www.ncbi.nlm.nih.gov/blast) revealed closest sequence similarity with Ce13/UNC-29, a nAChR β subunit cloned from Caenorhabditis elegans (Ballivet et al. 1996) which has been shown to be capable of generating functional nAChRs after heterologous expression (Fleming et al. 1997). An EST nucleotide sequence was identified in the GenBank/EMBL database (Accession no. AI542842) with a sequence identical to that of CG11822/CT33131. The corresponding EST cDNA clone (SD09326), isolated as part of the BDGP EST project (http://www.fruitfly.org) was obtained from Research Genetics Inc.. Complete nucleotide sequencing of the cDNA insert in EST clone SD09326 confirmed that it contained sequence corresponding to the entire predicted ORF of CT33131 but was interrupted by an additional 121 bp not present in the transcript predicted by GadFly. Comparison of the EST sequence with Drosophila genomic sequence data (database Accession nos AE003589 and AC004573) suggested that EST clone SD09326 contained a single unspliced intron (between exons 3 and 4), preceding the putative third transmembrane domain Fig. 1(a). The consequence of the unspliced intron is to disrupt the predicted ORF of CG11822/CT33131 after amino acid 312, introducing a premature termination codon within the intronic sequence. Two Drosophila cDNA libraries (a 0–5 day head λZAPII library and an ‘adult’λgt10 library, see Materials and methods) were screened using PCR to establish whether cDNAs could be identified corresponding to that of the predicted transcript CT33131, i.e. lacking the presumed unspliced intron present in EST clone SD09326. Using oligonucleotide primers which hybridized to sequences within exons 3 and 4 of CG11822, a single PCR band was amplified from both libraries which corresponded to the size expected of the predicted (fully spliced) transcript. No evidence was obtained for expression of transcripts containing the presumed unspliced intron in either cDNA library. Comparison of genomic sequence data (GenBank/EMBL database Accession nos AE003589 and AC004573) with that of the cDNA sequence reveals that the entire coding region of this transcript spans only ≈ 1800 bp of genomic sequence and is interrupted by only three introns, none longer than ≈ 250 bp (Fig. 1a).

Figure 1.

Structure of predicted Drosophila gene CG11822 and Dβ3 cDNA. (a) Comparison of cDNA and genomic sequence data has enabled the identification of introns and exons in the predicted Drosophila gene CG11822 (http://www.fruitfly.org/annot/index.html). The position of exons within the Drosophila genomic sequence (AC004573) is indicated by open boxes. Exons identified in BDGP EST clone SD09326 and in cDNA clone Dβ3 are also shown. The position of the Dβ3 ORF in relation to the cDNA sequence is shown (filled box). The Dβ3 nucleotide sequence has been submitted to the EMBL/GenBank databases (Accession no. AJ218761). (b) A region of Drosophila genomic sequence (Accession no. AC004573) corresponding to the 3′ region of Dβ3. The amino acid sequence (single letter code) of the predicted C-terminal region of the Dβ3 subunit is shown above the genomic sequence. The predicted termination codon is indicated by an asterisk. The 3′-end of both the EST clone SD09326 and of an independent clone isolated from a Drosophila adult (0–5 day) head cDNA library is indicated by a vertical arrow above the␣nucleotide sequence. The 3′-end of the cDNA clone is preceded by␣a consensus 3′ polyadenylation signal sequence (aataaa), underlined.

The position of the 3′-end of EST clone SD09326 in relation to the end of the Dβ3 predicted ORF is illustrated in Fig. 1(b). A highly conserved consensus polyadenylation signal sequence is present 10 nucleotides upstream of the 3′-end of the EST cDNA clone. The distance of the consensus polyadenylation signal sequence from the 3′-end of the cDNA is typical of polyadenylation sites in higher eukaryotes (Zhao et al. 1999) and is consistent with the assumption that the 3′-end of the cDNA clone corresponds to the transcription termination site of the Dβ3 mRNA. This supports the conclusion that the end of the Dβ3 ORF identified in Fig. 1(b) corresponds to the true C-terminus of the Dβ3 subunit. There remains, however, a possibility that if EST clone SD09326 contains an unspliced intron at its extreme 3′-end, then the termination codon might not represent the true translational stop site. To examine this possibility, Dβ3 3′ cDNA sequence was amplified from a Drosophila (05 day) head cDNA library with a forward primer located within the Dβ3 coding sequence and a reverse primer to the T3 promoter site of the λZAPII vector. Nucleotide sequencing revealed that the transcriptional termination site, which was followed by a poly(A) sequence, occurred at an identical position to that of EST cDNA clone SD09326 (Fig. 1). This strongly supports the conclusion that the Dβ3 ORF terminates at the position identified in EST clone SD09326 (Fig. 1).

Features of the Dβ3 subunit

The predicted transcript CT33131 encodes an ORF of 441 amino acids (Fig. 2) with has features typical of a nAChR subunit (which we will refer to as ‘Dβ3’, i.e. the third nAChR β subunit of Drosophila). In common with other nAChR subunits, Dβ3 contains four hydrophobic putative transmembrane domains and a putative N-terminal signal sequence which is predicted to be cleaved after amino acid 39 (algorithm of von Heijne 1986). If the Dβ3 subunit is cleaved at this predicted position, a mature protein of 402 amino acids would be generated with a calculated of molecular mass of 45 kDa (excluding any contribution to the molecular mass from glycosylation). This is significantly smaller than that of many other previously cloned insect or mammalian nAChR subunits (Claudio 1989), largely a consequence of this subunit containing a very short putative intracellular loop region between the third and fourth predicted transmembrane domains (Fig. 2). In common with all other nAChR subunits identified to date, Dβ3 contains the 15 amino acid Cys loop motif in its putative extracellular domain (Fig. 2), but lacks the two adjacent Cys residues analogous to Cys192 and Cys193 in the Torpedo nAChR α subunit which are considered to be characteristic of nAChR α-type subunits. The predicted N-terminal extracellular domain contains three potential N-linked glycosylation sites (N-X-S/T; Fig. 2). Therefore, if Dβ3 is glycosylated, as are other nAChR subunits, the actual molecular mass of the mature subunit protein would be expected to be > 45 kDa. Although the presence of potential N-linked glycosylation sites is a feature of all nAChR subunits, the three predicted sites in Dβ3 are not conserved in any of the other previously identified Drosophila nAChR subunits.

Figure 2.

Alignment of the deduced amino acid sequence of Dβ3 with that of other Drosophila nAChR subunits. The deduced amino acid sequence of Dβ3 is compared with that of other Drosophila nAChR subunits. The alignment was produced using the GCG program pileup (Genetics Computer Group Inc.). Conserved amino acids have been identified by shading by means of the program MacBoxshade (http://www.netaxs.com/jayfar/mops.html). Identical amino acids are shown as white text on a black background and␣similar␣amino acids as black text on a shaded background. The predicted site of cleavage of the N-terminal signal sequence (after␣amino acid 39) is indicated by #. The position of potential N-linked glycosylation sites (N-X-S/T) in Dβ3 are indicated by ***. Putative transmembrane domains (M1−M4) and the 15 amino acid␣Cys␣loop␣motif are also indicated. Where gaps have been introduced to␣optimize sequence alignment, this is indicated by dots. The␣absence␣of additional amino acid sequence at the N- or C-terminus is indicated by ∼. For convenience in formatting the figure,␣the final six␣C-terminal amino acids of the ALS subunit (GSENTL) and the final␣seven amino acids of SBD (IVRQVLT) have been omitted.

The predicted amino acid sequence of Dβ3 was compared with that of other Drosophila ligand-gated ion channel subunits by use of the GCG program pileup (Genetics Computer Group Inc., Madison, WI, USA). As illustrated in Fig. 3, the Dβ3 sequence shows closest sequence similarity to the two previously identified Drosophila nAChR β-type subunits (ARD and SBD). In comparisons which exclude the highly divergent M3−M4 intracellular loop region, other previously cloned Drosophila nAChR subunits exhibit between 45 and 65% amino acid sequence identity (Gundelfinger and Hess 1992). In contrast, Dβ3 is considerably more distantly related to other nAChR subunits (≈ 20% amino acid sequence identity).

Figure 3.

Evolutionary relationship of Drosophila ligand-gated ion channel subunits. The deduced amino acid sequence of Dβ3 is compared with that of other Drosophila ligand-gated ion channel subunits with the GCG program pileup (Genetics Computer Group Inc.). A comparison has been made both with previously identified nAChR subunits (ALS, ARD, Dα2-Dα4 and SBD), GABA-gated Cl channel subunits (GRD, LCCH3 and RDL) and a glutamate-gated Cl channel subunit (GluCl). By convention, nAChR subunits have been identified as either ‘α’ or ‘β’ on the basis of the presence or absence, respectively, of adjacent cysteine residues analogous to cysteine 192 and 193 of the Torpedo nAChR α subunit. EMBL/GenBank Accession nos for these subunits are: ALS, X07194 (Bossy et al. 1988); ARD, X04016 (Hermans-Borgmeyer et al. 1986); Dα2/SAD X52274/X53583 (Baumann et al. 1990; Sawruk et al. 1990a); Dα3, Y15593 (Schulz et al. 1998); Dα4, AJ272159 (Lansdell and Millar 2000a); Dβ3, AJ218761 (this study); GluCl, U58776 (Cully et al. 1996); GRD, X78349 (Harvey et al. 1994); LCCH3, S62717 (Henderson et al. 1993); RDL, M69057 (ffrench-Constant et al. 1991); SBD, X55676 and Y14678 (Sawruk et al. 1990b; Lansdell et al. 1997).

Subunit coassembly examined by immunoprecipitation

A full-length cDNA encoding the putative nAChR subunit Dβ3, was constructed which lacked the unspliced intron present in EST clone SD09326 (see Materials and methods for details) and subcloned into Drosophila expression vector pRmHa3 (Bunch et al. 1988) to enable heterologous expression in Drosophila S2 cells. In order to facilitate coprecipitation studies, an eight amino acid ‘FLAG’ epitope-tag (Hopp et al. 1988) was introduced into all previously cloned Drosophila nAChR subunit cDNAs within the putative M3−M4 intracellular loop region (see Materials and methods for details), an approach which has been used previously to study coassembly of vertebrate nAChR subunits (Cooper and Millar 1997). Drosophila S2 cells were cotransfected with various subunit cDNA combinations and metabolically labelled by incorporation radioactive amino acids. After detergent solubilization of transfected cells, subunit coassembly was examined by immunoprecipitation with an antibody specific for the FLAG epitope tag (mAbFLAG-M2).

After electrophoresis, a specific radiolabelled band of the expected molecular mass was observed for all epitope-tagged subunits with the exception of ALSFLAG which could not be detected as a radiolabelled protein. Despite this inability to detect ALSFLAG by immunoprecipitation, the ALSFLAG subunit (and all of the other epitope-tagged Drosophilaα subunits) formed a high-affinity binding site for nicotinic radioligands when co-expressed with the rat β2 subunit (data not shown). This would suggest that, although ALSFLAG forms a viable nAChR subunit, it may be particularly prone to proteolysis during detergent solubilization and immunoprecipitation. Although it was not possible to examine coassembly of Dβ3 with ALSFLAG, clear evidence was obtained for the coprecipitation of Dβ3 with Dα2FLAG, Dα4FLAG and SBDFLAG subunits (Fig. 4). Evidence for weak coassembly of Dβ3 with Dα3FLAG was obtained but was only visible after long periods of␣autoradiography (not shown). Evidence for coassembly of Dβ3 with ARDFLAG was difficult to interpret because of comigration of these subunits during electrophoresis (data not shown). The apparent molecular mass of the Dβ3 subunit coprecipitated with other FLAG-tagged Drosophila nAChRs subunits was estimated to be ≈ 49 kDa. This is close to what might be expected for a glycosylated form of Dβ3 for which the predicted molecular mass of the mature unglycosylated protein is 45 kDa (see above). It is unlikely that the 49 kDa band corresponds to a protein other than the Dβ3 subunit as this band is present only in cells transfected with Dβ3 and is absent from the corresponding control cells which were not transfected with Dβ3 (compare lanes 2 and 3, 4 and 5, 7 and 8). Further evidence that the 49 kDa band corresponds to the Dβ3 subunit comes from immunoprecipitation experiments performed with FLAG-tagged Dβ3 subunit. The apparent molecular mass of the FLAG-tagged Dβ3 subunit is slightly larger than that of Dβ3 (compare lanes 8 and 9). An increase in molecular mass (≈ 0.9 kDa) would be expected because of the addition of the FLAG epitope tag but this difference in migration on the gel may also reflect changes in net charge or protein conformation after solubilization in SDS due to the introduction of the FLAG epitope. A similar sized increase in apparent molecular mass is seen consistently in other nAChR subunits after introduction of a FLAG epitope tag. The difference in apparent molecular masses of the Dα3 and Dα3FLAG subunit is illustrated in Fig. 4 (lanes 11 and 12). Disappointingly, we have been unable to detect coprecipitation of other insect nAChR subunits with Dβ3FLAG. A␣possible explanation for this might be that introduction of the FLAG epitope into Dβ3 causes greater disruption of subunit structure than when introduced into other subunits. As with the other nAChR subunits, the FLAG epitope was introduced into the presumed M3−M4 intracellular loop region of Dβ3. However, as discussed elsewhere, the Dβ3 subunit contains an atypically short M3−M4 intracellular loop domain.

Figure 4.

Co-immunoprecipitation of Dβ3. Dβ3 was co-expressed in Drosophila S2 cells together with FLAG epitope-tagged nAChR subunits. After metabolic labelling of transfected cells, the ability of Dβ3 to coassemble with other nAChR subunits was examined by immunoprecipitation with mAbFLAG-M2 (an antibody raised against the FLAG epitope). Dβ3 was coprecipitated when co-expressed with either Dα4FLAG, SBDFLAG or Dα2FLAG (lanes 3, 5 and 8). Absence of cross-reactivity of mAbFLAG-M2 with Dβ3 was confirmed (lane 6). With the exception of ALSFLAG (not shown), all epitope-tagged nAChR subunits examined could be detected as specific immunoprecipitated bands when expressed alone. Immunoprecipitation of Dα4FLAG, SBDFLAG and Dα2FLAG is shown (lanes 2, 4 and 7). Because ARDFLAG migrated at a position close to that of Dβ3 (not shown), it was␣not possible to␣assess unambiguously whether Dβ3 was coprecipitated by this subunit. Direct immunoprecipitation of Dβ3FLAG is␣shown (lane 9). Comparison of Dβ3FLAG with the coprecipitated Dβ3 subunit (lanes 3, 5 and 8) indicates that introduction of the epitope tag causes a small increase in apparent molecular mass, a phenomenon which has been observed consistently with other FLAG-tagged nAChR subunits. A similar difference in the apparent molecular masses of the Dα3 and Dα3FLAG. subunits (when coprecipitated with an antibody specific for the rat β2 subunit) is illustrated (compare lanes 11 and 12). The position of protein molecular mass markers are shown.

Subunit coassembly examined by radioligand binding

As reported previously, a high-affinity nicotinic radioligand binding site is formed when any of the four cloned Drosophila nAChR α subunits (ALS, Dα2, Dα3 and Dα4) are co-expressed with the vertebrate (e.g. rat) β2 subunit but not when they are co-expressed with either of the two previously cloned Drosophilaβ subunits (Lansdell et al. 1997; Lansdell and Millar 2000a, b). The Dβ3 subunit was co-expressed in Drosophila S2 cells with a variety of other Drosophila nAChR subunits and membrane preparations from transfected cells examined by radioligand binding. In addition to subunit pairs, combinations of 3–7 different Drosophila nAChR subunits were also co-expressed. No specific binding of any nicotinic radioligands ([α-125I]bungarotoxin, [3H]epibatidine or [3H]methyl-carbamylcholine) was detected with any of the Drosophila subunit combinations examined in the absence of vertebrate nAChR β subunits. We were also unable to detect specific nicotinic radioligand binding when Dβ3 was co-expressed with vertebrate nAChR α-type subunits, e.g. rat α4, despite coprecipitation data showing that these subunits are able to coassemble (data not shown). Thus it appears that Dβ3, like the two previously cloned Drosophilaβ subunits (ARD and SBD), is able to coassemble with vertebrate and insect α subunits but coassembly is unable to generate a high-affinity ligand binding site.

Evidence indicating subunit coassembly into complexes which fail to generate a nicotinic binding site raises the question of whether such subunit associations represent ‘dead-end’ assembly complexes. We have examined this question by co-expressing Drosophilaβ subunits with subunit pairs which are themselves able to coassemble productively (i.e. nAChR α subunits paired with the vertebrate β2 subunit). As has been reported previously, high levels of specific radioligand binding are detected when rat α4 and β2 subunits are co-expressed in Drosophila S2 cells (Lansdell et al. 1997). We observed a significant reduction in the level of total radioligand binding when α4+β2 subunits were co-expressed with either of the Drosophilaβ subunits Dβ3 or SBD (Fig. 5a). Such a ‘dominant-negative’ effect, which has been reported previously after co-expression of truncated nAChR subunits (Verrall and Hall 1992), would be expected if a consequence of the coassembly of Drosophilaβ subunits with vertebrate subunits is to prevent the productive coassembly of α4 and β2 with each other.

Figure 5.

Modulation of nAChR expression levels by co-expression of Drosophila nAChR β subunits. Drosophila S2 cells were transiently transfected with various nAChR subunit cDNA constructs. Membrane preparations from transfected cells were assayed for total [3H]epibatidine binding with a saturating concentration of radioligand (3 nm). (a) A dominant-negative effect was observed when Drosophilaβ subunits (SBD or Dβ3) were co-expressed with rat α4 + β2 subunits, illustrated by a reduction in total radioligand binding. (b) A reduction in total radioligand binding was detected when either ARD or SBD were co-expressed with the Dα3 + β2 subunit combination. In contrast, co-expression of Dβ3 resulted in an elevation in the level of radioligand binding. Data are means of 3–11 independent experiments (as indicated), each performed in triplicate. Very low levels of nonspecific binding were detected (typically, specific binding was ≈ 1000-fold higher than nonspecific). In all cases absolute levels of radioligand␣binding were determined as specific binding/mg protein (2.9 ± 0.3 pmol/mg, n = 6, for α4 + β2; 0.2 ± 0.03 pmol/mg, n = 11, for Dα3 + β2). Data have been normalized to that detected in the absence of Drosophilaβ subunits. Significant differences from controls, determined by two-tailed Student's t-test, are indicated (*p < 0.02; **p < 0.002; ***p < 0.001).

A similar series of experiments was performed with other subunit combinations which are capable of generating functional nAChRs but which contain a Drosophilaα subunit (e.g. Dα3 + β2). As illustrated in Fig. 5(b), when either of the two previously cloned Drosophilaβ subunits (ARD or SBD) were co-expressed with Dα3 + β2 a dominant-negative effect is observed, similar to that seen when they are co-expressed with α4 + β2. However, when Dβ3 was co-expressed with Dα3 + β2, a significant increase in the level of specific radioligand binding was observed (Fig. 5b). Thus, in contrast to its dominant-negative effect when co-expressed with the vertebrate α4 subunit, Dβ3 enhances the level of nicotinic radioligand binding sites when co-expressed with a subunit combination containing the Drosophila Dα3 subunit. A small increase in the level of total radioligand binding was observed when Dβ3 was co-expressed with subunit pairs containing other Drosophilaα subunits (i.e. ALS + β2, Dα2 + β2 and Dα4 + β2) but these effects were not statistically significant.

Because the coassembly of Dβ3 with the Dα3 + β2 subunit combination resulted in a significant enhancement in the total steady-state levels of radioligand binding (and SBD caused a significant decrease), we examined whether the coassembly of these subunits influences the affinity of radioligand binding. No significant difference in the affinity of radioligand binding was detected when either Dβ3 or SBD were co-expressed with Dα3 + β2. Kd values determined from at least three independent experiments were: 0.031 ± 0.009 nm, Dα3 + β2; 0.028 ± 0.005 nm, Dα3 + β2 + Dβ3; 0.029 ± 0.01 nm, Dα3 + β2 + SBD. Because the Kd values did not change and because saturating concentrations of radioligand were used, it seems reasonable to conclude that the increase in radioligand binding observed after coassembly of Dβ3 with Dα3 + β2 corresponds to an increase in the number of nAChR radioligand binding sites (and, conversely, that coassembly with SBD causes a reduction in nAChR binding sites).


A novel putative nAChR subunit, Dβ3, has been cloned from Drosophila and shown, by heterologous expression studies to coassemble with previously cloned nAChR subunits and contribute to the formation of recombinant nAChRs. Although the Dβ3 subunit displays features which are characteristic of nAChR subunits (as is described in Results), we consider it appropriate to describe Dβ3 as an atypical nAChR subunit. This is illustrated by the fact that, whereas previously cloned nAChRs subunits typically display 45–65% amino acid sequence identity in pair-wise comparisons (Gundelfinger and Hess 1992), Dβ3 shows only ≈ 20% identity with other nAChR subunits (even when comparisons ignore the highly divergent M3−M4 loop region). The length of the predicted M3−M4 intracellular loop in Dβ3 is shorter than that of other previously cloned nAChR subunits and, consequently, the size of the predicted mature Dβ3 subunit (402 amino acids) is significantly smaller than that of other insect or mammalian nAChR subunits, the size of which range, typically, from ≈ 440 to 600 amino acids (Claudio 1989). The recently characterized vertebrate α10 nAChR subunit (Elgoyhen et al. 2001) also has a relatively small ORF (predicted mature protein of 423 amino acids; mature unglycosylated molecular mass of 47 kDa). Despite this, the α10 subunit, like other vertebrate nAChR subunits, contains a predicted M3−M4 intracellular loop region which is considerably longer than that of Dβ3. The Dβ3 coding sequence ends at a position immediately after the predicted M4 domain, in contrast to other Drosophila nAChR subunits which extend for a further 20–40 amino acids (Fig. 2). There are, however, several examples of nAChR subunits which terminate at a position similar to that of Dβ3. These include the mammalian α9 and α10 subunits (Elgoyhen et al. 1994, 2001) and several nAChR subunits identified in C. elegans (Mongan et al. 1998). Although Dβ3, contains potential N-linked glycosylation sites within its predicted extracellular domain, as do other nAChR subunits, these do not occur at positions which are conserved in other subunits. Although nAChRs can differ widely in their exon−intron structure (Le Novére and Changeux 1995), the structure of the Dβ3 gene is somewhat unusual, containing only three introns, none of which are larger than ≈ 250 bp. As it contains very little intronic DNA, the Dβ3 gene occupies < 1800 bp of genomic sequence. This is in stark contrast to, for example, the recently cloned Drosophila nAChR Dα4 subunit which is encoded by 11 exons spanning ≈73 000 bp of genomic sequence (Lansdell and Millar 2000a).

The nomenclature for genes encoding Drosophila nAChR subunits, as proposed by FlyBase (http://flybase.bio.indiana.edu), includes information derived from the gene's chromosomal localization. The predicted gene CG11822 has been cytogenetically mapped to chromosomal location 21C5-21C6 (see http://flybase.bio.indiana.edu/.bin/fbidq.html?FBgn0031261). Consequently, although we have suggested the trivial name Dβ3 for this putative Drosophila nAChR subunit, in accordance with the suggested FlyBase nomenclature, we propose that the formal name for this gene should be nAcRβ-21C.

To date, reconstitution of functional nAChRs by heterologous expression of nAChR subunits cloned from Drosophila or from other insect species has been largely unsuccessful (Hermsen et al. 1998; Millar 1999; Gundelfinger and Schulz 2000). It has, however, been possible to detect functional nAChRs and/or high-affinity binding of nicotinic radioligands when any of the four Drosophilaα subunits is co-expressed with any one of several vertebrate non-α subunits (Bertrand et al. 1994; Lansdell et al. 1997; Schulz et al. 1998; Lansdell and Millar 2000a,b). Such findings have led to the assumption that problems encountered in heterologous expression of insect nAChRs may be due to a requirement for an, as yet unidentified, insect β-type subunit (Millar 1999). Vertebrate subunits which have been used successfully in co-expression studies with Drosophilaα subunits include the neuronal nAChR β2 and β4 subunits and the vertebrate muscle nAChR γ- and δ subunits (Bertrand et al. 1994; Lansdell and Millar 2000b). The recently completed Drosophila genome sequence has enabled a detailed search for putative novel genes but has not identified any novel putative Drosophila nAChR subunits which show particularly close sequence similarity to vertebrate subunits such as β2, β4, γ or δ subunits (see, for example, Littleton and Ganetzky 2000). Three further putative nAChR α-type subunits have been identified, all of which appear to be most closely related to the vertebrate neuronal nAChR α7 subunit (Littleton and Ganetzky 2000). The only putative novel Drosophila nAChR β subunit (here called Dβ3) which has emerged from analysis of the Drosophila genome shows closest sequence similarity to a C. elegans nAChR β subunit ‘Ce13/UNC-29’ (Ballivet et al. 1996) a nAChR subunit␣which has been shown by heterologous expression to be capable of generating functional nAChRs (Fleming et al.␣1997).

Although the cloning and heterologous expression of Dβ3 has not enabled the generation of functional recombinant nAChRs in the absence of a co-expressed vertebrate β subunit, we have obtained evidence to indicate that Dβ3 can coassemble with other Drosophila nAChR subunits and can contribute to␣the formation of hybrid (insect/vertebrate) recombinant nAChRs. This has been demonstrated by an increase in radioligand binding seen when Dβ3 is co-expressed as a third subunit with combinations of Drosophilaα subunit and␣vertebrate β subunits (e.g. Dα3 + rat β2). Interestingly, we also observe a dominant-negative effect when Dβ3 is co-expressed with the vertebrate nAChR α4 + β2 subunit combination.

Because no difference in the affinity of radioligand binding is observed when Dβ3 is co-expressed with subunit combinations such as Dα3 + β2, we can conclude that the increase in the level of specific radioligand binding represents an increase in the steady-state level of assembled Dα3-containing nAChRs (Bmax). This may be a consequence of Dβ3 increasing the efficiency of coassembly of Dα3 with the β2 subunit. The decrease in total radioligand binding detected when Dβ3 is co-expressed with rat α4 and β2 subunits suggests that coassembly of Dβ3 reduces the efficiency with which α4 and β2 subunits can coassemble with one another. This may be a consequence of α4 and Dβ3 coassembling into ‘dead-end’ subunit complexes which preclude the subsequent association of α4 with β2.

As discussed previously (Lansdell et al. 1997), one explanation for difficulties encountered in heterologous expression of Drosophila and other insect nAChRs is the unsuitability of commonly used heterologous expression systems. However, considering the success that has been achieved in heterologous expression studies when insect α subunits have been co-expressed with vertebrate β subunits, a more plausible explanation is that additional (presently uncloned) insect nAChR subunits are required. We have obtained evidence to suggest that Dβ3 is able to coassemble with other Drosophila nAChR subunits and contribute to the formation of recombinant nAChRs. However, as these receptors still require the co-expression of a nAChR β subunit the search must continue (either for the elusive ‘missing’ subunit or for a more optimal expression system).


We thank Heinrich Betz, Thomas Bunch, Ron Davis, Jon Lindstrom, Jim Patrick and Bertram Schmitt for generously providing some of the antibodies, plasmids, and cDNA libraries used in this study. This work was supported by grants to NSM from the Biotechnology and Biological Sciences Research Council and the Wellcome Trust.