The Dα5, Dα6 and Dα7 subunits differ from other Drosophila nAChR subunits in their close sequence similarity to the vertebrate α7 nAChR subunit [41, 48], a subunit that is notable for its ability to form both homomeric and heteromeric nAChRs [49–52]. In addition to being one of the best characterised homomeric nAChRs, the vertebrate α7 subunit can co-assemble into heteromeric nAChRs by co-assembly with the α8 subunit (in avian speices) . Although an α8 subunit is not present in mammals, recent evidence indicates that the mammalian α7 subunit can also form functional heteromeric nAChRs by co-assembly with β2 [51, 52].
Relatively limited information is available about the physiological roles of the Dα5, Dα6 and Dα7 subunits in Drosophila, or about the role of analogous subunits in other insect species. There is, however, evidence from studies of native nAChRs in Drosophila that Dα5 forms part of a nAChR that is sensitive to α-bungarotoxin , Dα6 forms part of the spinosad-sensitive nAChR  and that Dα7 is required for the visually-mediated cholinergic escape response .
As has been discussed elsewhere [10, 11], difficulties have been encountered in the efficient functional expression of insect nAChRs. Here we report the cloning of a full-length cDNA of the Drosophila Dα5 subunit corresponding to a previously described isoform B . Other isoforms of Dα5 described previously (isoforms A and C)  are a consequence of alternative splicing and have fewer exons than isoform B. Isoform A lacks exon 7, which codes for part of the second transmembrane domain, whilst isoform C lacks exon 5, which codes for the region containing the extracellular Cys-loop. The cloning of the Dα5 subunit was first reported in 2002  but no expression studies were described at that time. More recently, it has been reported that Dα5 does not generate functional homomeric nAChRs when expressed in Xenopus oocytes, even when co-expressed with RIC-3 . Functional expression was, however, reported in the same study when Dα5 was co-expressed with Dα6 and RIC-3 . In the present study, we have detected functional responses when Dα5 is co-expressed with Dα6 but, in contrast to the previous study , we have also obtained evidence for the functional expression of homomeric Dα5 nAChRs. Similarly, we have demonstrated that Dα7 can form both homomeric and heteromeric nAChRs. As far as we are aware, there have been no previous reports of the successful functional expression of Dα7, as either a homomeric or a heteromeric nAChR. Given the difficulties encountered in obtaining reproducible functional expression of insect recombinant nAChRs, it is not surprising that there may be some apparent differences in subunit combinations found to generate functional receptors in this and previous studies, particularly since the focus of the most detailed previous study was the identification of a spinosyn-sensitive nAChRs .
Our studies conducted in cell lines provided evidence that the pairwise combination Dα5 + Dα6 generates a high affinity radioligand binding site, a finding that agrees with previous studies demonstrating functional expression of Dα5 + Dα6 nAChRs in oocytes . Interestingly, we have found no evidence of specific binding when Dα7 was co-expressed with Dα5 and Dα6 in the same cell lines. This lack of specific binding would seem to suggest that, in the two cell lines examined, co-assembly of Dα7 with the Dα5 and Dα6 subunit interferes with the formation of correctly assembled complexes. We observed a somewhat similar situation in oocytes, where expression of Dα7 alone generates functional nAChRs but it fails to do so when co-expressed with Dα6. This may reflect a tendency for Dα6 and Dα7 to assemble into non-functional complexes. The one situation where this tendency is not dominant is when all three subunits (Dα5 + Dα6 + Dα7) are co-expressed with RIC-3 in oocytes, where they are able to form a functional ‘triplet’ nAChR with high apparent affinity for acetylcholine.
The present findings suggest that the environment provided by the host cell exerts a substantial effect on the assembly of these nAChR subtypes, a phenomena that has been reported previously for other nAChRs [47, 55, 56]. Previous studies by another research group  support the conclusion that co-assembly of Dα5 + Dα6 nAChRs is somewhat inefficient. Not only was functional expression of the Dα5 + Dα6 subunit combination found to be inconsistent in the previous study, but it also appeared to be dependent on the ratio of cRNAs injected . Perhaps this inconsistent functional expression reflects a tendency for some subunit combinations to assemble into non-functional complexes and that this may be more prevalent in certain subunit stoichiometries. It is possible that, in the native cellular environment, factors determining efficiency of subunit assembly and maturation may differ, perhaps as a consequence of a different array of endogenous chaperone proteins. This conclusion is supported by previous studies that have indicated that influence of RIC-3 on maturation of nAChRs is influenced by the host cell  and may help to explain the differences that we have observed in the ability of some subunit combinations to assemble into nAChRs in different expression systems.
The data obtained from expression studies in Drosophila and human cell lines is broadly similar. However, successful expression in human cells required incubation at a temperature lower than they would normally be maintained at (25°C, rather than 37°C) [note: Drosophila S2 cells are routinely maintained at 25°C]. Previous studies have demonstrated that the folding and assembly of the nAChRs from insects  and from some other non-insect species, such as the cold-water ray Torpedo, can be influenced by temperature. This temperature dependence appears to be a consequence of inefficient protein folding and/or subunit assembly at higher temperatures. Previously, due to difficulties in expression of Dα6 and Dα7 nAChR subunits, we examined the ability of subunit chimeras to assemble into complexes capable of binding 125I]-α-bungarotoxin . From such studies, it was possible to conclude that the Dα6 and Dα7 subunits were capable of heterometic co-assembly. In the present study the data from subunit chimeras is less clear cut. Although higher levels of 125I]-α-bungarotoxin were seen consistently when the Dα5 chimera was co-expressed with either the Dα6 and Dα7 chimeras, it was not clear in all cases whether this was greater than an additive effect. Nevertheless these findings are consistent with the conclusion that Dα5 is able to co-assemble into heteromeric complexes. For all subunit combinations examined, responses to acetylcholine were completely blocked by α-bungarotoxin, a finding that is consistent with previous studies conducted with native nAChRs purified from Drosophila which demonstrated that Dα5 is part of an α-bungarotoxin binding nAChR .
As mentioned above, a previous study has reported the functional expression of heteromeric Dα5 + Dα6 nAChRs (co-expressed with RIC-3) in Xenopus oocytes and also the inability of either Dα5 or Dα6 to form functional homomeric nAChRs . Significantly, the authors of this earlier study describe substantial difficulties in achieving reliable functional expression. In the present study, despite demonstrating the functional expression of several combinations of the Dα5, Dα6 and Dα7 subunits, we have also encountered a much lower success rate than is typically achieved with other nAChRs. In both transfected cell lines and in Xenopus oocytes, we occasionally failed to detect evidence of radioligand binding or functional expression, despite success with other nAChRs that were expressed as positive controls (for example the mammalian α7 nAChR). The difficulties that we and others have encountered may be associated with a tendency for these subunits to co-assemble into non-functional complexes. It is possible that this may reflect a requirement for additional chaperone proteins. Indeed, a study conducted with a C. elegans nAChR has demonstrated a requirement for three different chaperone proteins for efficient functional heterologous expression .