The cys-loop ligand-gated ion channel superfamily of the honeybee, Apis mellifera
- First Online:
- Cite this article as:
- Jones, A.K. & Sattelle, D.B. Invert Neurosci (2006) 6: 123. doi:10.1007/s10158-006-0026-y
- 287 Views
Members of the cys-loop ligand-gated ion channel (cys-loop LGIC) superfamily mediate neurotransmission in insects and are targets of successful insecticides. We have described the cys-loop LGIC superfamily of the honeybee, Apis mellifera, which is an important crop pollinator and a key model for social interaction. The honeybee superfamily consists of 21 genes, which is slightly smaller than that of Drosophila melanogaster comprising 23 genes. As with Drosophila, the honeybee possesses ion channels gated by acetylcholine, γ-amino butyric acid, glutamate and histamine as well as orthologs of the Drosophila pH-sensitive chloride channel (pHCl), CG8916, CG12344 and CG6927. Similar to Drosophila, honeybee cys-loop LGIC diversity is broadened by differential splicing which may also serve to generate species-specific receptor isoforms. These findings on Apis mellifera enhance our understanding of cys-loop LGIC functional genomics and provide a useful basis for the development of improved insecticides that spare a major beneficial insect species.
KeywordsAlternative splicingApis melliferaDrosophila melanogasterInsecticide targetsIon channelsNeurotransmitter receptors
In insects, members of the cys-loop ligand-gated ion channel (cys-loop LGIC) superfamily mediate both fast excitatory and inhibitory synaptic transmission in the nervous system. The superfamily includes cation permeable nicotinic acetylcholine receptors (nAChRs) (Tomizawa and Casida 2001; Sattelle et al. 2005), γ-amino butyric acid (GABA)-gated anion channels (Buckingham et al. 2005), glutamate-gated chloride channels (GluCls) (Vassilatis et al. 1997) and histamine-gated chloride channels (HisCls) (Gisselmann et al. 2002; Zheng et al. 2002). The cys-loop superfamily of ionotropic receptors are of considerable interest as they are targets of widely used insecticides (Gepner et al. 1978; Raymond-Delpech et al. 2005). For example, nAChRs are targets of neonicotinoids (Matsuda et al. 2001), a class of insect control chemicals which include imidacloprid with worldwide annual sales of approximately one billion dollars (Tomizawa and Casida 2005). Also, GABA receptors and GluCls are targets of fipronil and avermectins respectively (Bloomquist 2003).
The honeybee, Apis mellifera, is an important beneficial insect in agriculture. Its contribution to crop pollination is valued at more than US $14 billion per year in the US alone (United States Department of Agriculture http://www.ars.usda.gov/main/main.htm). In France, the use of fipronil and imidacloprid has been suspended over concerns that they may be playing a role in the drastic decline of honeybee populations (http://www.moraybeekeepers.co.uk/insecticide.htm, http://www.pan-uk.org/press/pr140604.htm). Although links between the use of these insecticides and bee population decline have yet to be proven, the importance of developing compounds that distinguish between pest and beneficial insect species has been highlighted.
Studies using honeybees have shown that nAChRs play roles in various aspects of behaviour and that sub-lethal doses of imidacloprid can alter foraging and learning (Lozano et al. 1996; Guez et al. 2001; Lambin et al. 2001; Lozano et al. 2001; Decourtye et al. 2004; Dacher et al. 2005; Thany and Gauthier 2005; Gauthier et al. 2006). The possible existence of other members of the cys-loop LGIC superfamily in the honeybee has been indicated by immunocytochemical techniques that identified GABA (Bicker et al. 1985), glutamate (Bicker et al. 1988) and histamine (Bornhauser and Meyer 1997). More recently, whole-cell patch-clamp recordings have demonstrated the existence of nAChRs as well as GABA-gated and glutamate-gated chloride channels in honeybee antennal lobe neurons (Barbara et al. 2005). Clues to functional roles of these receptors in the honeybee have been identified based on pharmacological studies. For example, picrotoxin, which is a GABA receptor antagonist, impairs olfactory discrimination (Stopfer et al. 1997; Hosler et al. 2000). Also, fipronil, which blocks both honeybee GABA receptors and GluCls (Barbara et al. 2005), affects olfactory learning and sucrose perception (El Hassani et al. 2005). The role of HisCls in the honeybee has yet to be determined although studies on Drosophila melanogaster show that HisCls function in the visual system (Gengs et al. 2002; Hamasaka and Nassel 2006) and at least one HisCl subunit is transcribed in the testes (Iovchev et al. 2006).
Four A. mellifera nAChR subunits have been cloned and shown to be expressed in the honeybee nervous system (Thany et al. 2003, 2005) and with the sequencing of the honeybee genome (The Honey Bee Genome Sequencing Consortium 2006), the complete A. mellifera nAChR gene family has now been described (Jones et al. 2006). In addition, an A. mellifera HisCl sequence has been identified (Iovchev et al. 2006). Here, we use the A. mellifera genome sequence information to describe the honeybee cys-loop LGIC gene superfamily and compare it with that of D. melanogaster.
Materials and methods
Identification of cys-loop LGIC subunits in the A. mellifera genome
The characterisation of honeybee nAChR subunits has been previously described (Jones et al. 2006). To identify the remaining members of the honeybee cys-loop LGIC superfamily, the A. mellifera genome (database version 34.2b available at http://www.ensembl.org/Apis_mellifera/index.html and assembly version 3.0 available at http://www.ensembl.org/Apis_mellifera/) was screened with cDNA sequences of every member of the D. melanogaster cys-loop LGIC superfamily using the tBLASTn algorithm (Altschul et al. 1990). Candidate honeybee cys-loop LGIC subunits were identified based on their considerable sequence homology with previously characterized subunits (sequences with lowest similarity had E Value 1e-55), particularly in the N-terminal ligand-binding domain and the four transmembrane regions. The highly variable N-terminal signal peptides of several subunits were identified by using tBLASTn to screen EST data (http://www.flybase.bio.indiana.edu/blast/). RT-PCRs were performed to verify and correct the open-reading frame sequences of each subunit and to detect potential single nucleotide polymorphisms and splice variants.
The multiple protein sequence alignment was constructed with ClustalX (Thompson et al. 1997) using the slow-accurate mode with a gap opening penalty of 10 and a gap extension penalty of 0.1 as well as applying the Gonnet 250 protein weight matrix (Benner et al. 1994). The protein alignment was viewed using GeneDoc (http://www.psc.edu/biomed/genedoc). Identity values between subunit sequences were calculated using the GeneDoc program. The neighbour-joining method (Saitou and Nei 1987) and bootstrap resampling (Felsenstein 1985), available with the ClustalX program, were used to construct a phylogenetic tree, which was then displayed using the TreeView application (Page 1996). Signal peptide cleavage sites were predicted using the SignalP 3.0 server (Dyrlov Bendtsen et al. 2004) and membrane-spanning regions were predicted using the TMpred program (available at http://www.ch.embnet.org/software/TMPRED_form.html). The PROSITE database (Hulo et al. 2006) was used to identify potential phosphorylation and nuclear targeting sites.
Reverse transcription and polymerase chain reaction
The primers used in RT-PCR analysis are shown (5′–3′). Forward primers are on the left of each pair and reverse primers are on the right
1st PCR reaction
2nd PCR reaction
aggcagtatgctgaatgac + ttatttcgcctcctcgaga
caacatctccgcgatattg + gagaagcacgagatcgtcc
catagcaacataagcgagc + tcacgtgccatttgaattca
ggacaatttgcttcgcgg + gttgtaataattgatacgctc
atgcatcacaggatgtggt + cagaggacataaaatatcca
aggatgtggttgcagcag + atagatggcgttgaacagc
gaacgctcattgcctagg + cactcgttctgctcttcc
cagcttctccaggctgc + ccgaaacaggtacgtgga
tgtcgatggacgagacac + aatgtactcggcgaacatta
ggcatggtttccgcaatc + tccaataagttacgttcaac
cctctttgtcgctcaagg + aggaaggtgctccaataga
gctgtacgacaagcacag + acgttcaagatgaggaagg
tgggttgggtggcgcta + gaagtgcctttgtagtgga
cgctgtctctcacctcc + aacgtgaggaacatgaagta
tgcaaaaatcggaattacg + gagaaattgatccaagcgg
ttgatgtttcttcagcagc + taggtgacgtaaatgaccc
cagacctccgtcgaagc + atataggaaaacctattctgc
gtttgccggtgctcgtg + cgtatcattggatattcagg
tagttctatgacacaaactg + caaataaaagaccaataaag
gactagaccaccaggag + tctgctacgtctatcgatc
Primers used to amplify specific variants arising from alternative splicing
agtcaccatgtatgtcctc + gataggtgtatggttgaatg
tccgtgtccgaagtgttg + tgtgaccgagcacgcga
agtcaccatgtatgtcctc + agttagactaatctccatgg
tccgtgtccgaagtgttg + cggtgaccaaggaccttg
cgtcaccatgtacgtcttg + gataggtgtatggttgaatg
ctcgctatccgaagtgaaa + tgtgaccgagcacgcga
cgtcaccatgtacgtcttg + agttagactaatctccatgg
ctcgctatccgaagtgaaa + cggtgaccaaggaccttg
gaacgctcattgcctagg + tcgtaatcgttcatccagc
cagcttctccaggctgc + tgatgtcactaattgtcgca
gaacgctcattgcctagg + tcgtaatcgttcatccagc
cagcttctccaggctgc + gtaacgtcatctatttttgag
The A. mellifera cys-loop LGIC superfamily consists of 21 subunit members
Percentage identity/similarity between A. mellifera and D. melanogaster cys-loop LGIC protein sequences. Proposed orthologs are shown in bold
The Apis cys-loop LGICs and their Drosophila orthologs share the same motifs preceding TM2 which is important for ion charge selectivity (Keramidas et al. 2004). Thus Amel_Rdl, Amel_GluCl, Amel_HisCl1, Amel_HisCl2 and Amel_pHCl possess the PAR sequence (Fig. 1) which is characteristic of anion channels. The remaining subunits have alterations in this motif which may affect ion channel properties. Indeed, Drosophila GRD and LCCH3 subunits, which have ADR and SAR motifs respectively, form heteromultimeric GABA-gated cation channels when expressed in Xenopus laevis oocytes (Gisselmann et al. 2004). A subset of Drosophila cys-loop LGICs were found to share with vertebrate glycine receptor subunits a pair of cysteines in addition to the cys-loop (Dent 2006). The Apis orthologs of these subunits (Amel_GluCl, Amel_HisCl1, Amel_HisCl2), as well as Amel_pHCl, also possess this putative second cys-loop which is located in a domain important for ligand binding, loop C (Fig. 1). Unusually for cys-loop LGIC subunits, the TM3-TM4 intracellular loop of Amel_8916 contains putative bipartite nuclear targeting sequences (Dingwall and Laskey 1991) located from residues 440–456 and 441–457 which are not present in its Drosophila ortholog, CG8916.
Splice variants increase Apis cys-loop LGIC diversity
Drosophila pHCl possesses three splice variants generated either by an insertion of an exon in the N-terminal extracellular domain (Variant 1), use of an alternative exon that includes TM1 (Variant 2) or use of differential splice sites causing the intracellular loop between TM3 and TM4 to differ by a stretch of 17 amino acids (Variant 3) (Schnizler et al. 2005). Examination of two A. mellifera EST sequences (Accession Nos. BI507717 and BI508873) and RT-PCR analysis showed that Amel_pHCl also possesses Variant 1, although the inserted exon is longer and introduces an N-glycosylation site (Fig. 3c). As with Drosophila pHCl, differential splicing causes the TM3-TM4 loop in Amel_pHCl to differ by a stretch of 17 residues (Variant 3), which, when compared to the equivalent region in Drosophila pHCl, differs by one residue (a valine in Drosophila and an isoleucine in Apis, Fig. 3c). While use of an alternative exon equivalent to Variant 2 was not observed, the insertion of an additional exon in loop C (denoted Variant 4, Fig. 3c) was detected in Amel_pHCl transcripts lacking the Variant 1 insertion.
Additional variants of Amel_Rdl and Amel_HisCl1 were detected in RT-PCR products where use of different splice sites in the TM3-TM4 intracellular loop generates two transcripts denoted either ‘short’ or ‘long’ variants (Fig. 3d, e). The Amel_Rdl short and long variants differ only by one amino acid residue but interestingly splicing to generate the short form creates a putative protein kinase C phosphorylation site (Fig. 3d). The Amel_HisCl1 short variant possesses partially overlapping protein kinase C and cAMP-dependent protein kinase phosphorylation sites (Fig. 3e). Splicing to generate the long form eliminates both these phosphorylation sites, however the extra stretch of amino acids contain two adjacent protein kinase C phosphorylation sites.
Several Drosophila cys-loop LGIC subunits undergo pre-mRNA A-to-I editing, a process which effectively causes select adenosine residues in the genome to be read as guanosine in transcripts (Seeburg 2002). These include Dα5, Dα6, Dα7, Dβ1, Dβ2, Rdl and GluCl (Grauso et al. 2002; Hoopengardner et al. 2003; Fayyazuddin et al. 2006). Whereas conservation of A-to-I editing was observed in Dα6 and its honeybee ortholog, Amelα6, which results in changes of functionally significant residues (Jones et al. 2006), no potential RNA editing or single nucleotide polymorphisms leading to amino acid alterations were observed in transcripts of other Apis cys-loop LGIC subunits.
We have described the cys-loop LGIC superfamily of the honeybee, Apis mellifera, an agriculturally important insect that is also well studied as a model of behaviour. The superfamily consists of 21 subunit-encoding genes, which is slightly smaller than that of the 23 subunits of Drosophila melanogaster. This is in line with findings that the honeybee and Drosophila genomes show similarities in the repertoire of genes underlying functions that differ dramatically in the two species such as brain function and behaviour (The Honey Bee Genome Sequencing Consortium 2006). While Drosophila and Anopheles gambiae each possess 10 nAChR subunit genes (Jones et al. 2005; Sattelle et al. 2005), Apis has an extra nAChR subunit that is highly divergent (Jones et al. 2006). Interestingly, the analysis in this report indicates that the Drosophila cys-loop LGIC subunit, Ntr (CG6698), clusters more closely with nAChR subunits than with anion channels (Fig. 2). At the amino acid level, Ntr only marginally resembles nAChRs more closely than other members of the cys-loop LGIC superfamily, showing 11% identity as opposed to 9% with Grd and HisCls. It does not possess either the GEK or PAR motif characteristic of cation or anion channels respectively (Keramidas et al. 2004). It will be of interest to determine whether Ntr represents a highly divergent nAChR subunit or defines a novel ion channel clade. Analysis of cys-loop LGIC superfamilies indicates that Drosophila possesses a family of three subunits (CG7589, CG6927 and CG11340) not found in chordates or Caenorhabditis elegans (Dent 2006). Only one member of this family, Amel_6927, was found in the Apis genome, thus accounting for the smaller size of the honeybee cys-loop LGIC superfamily when compared to Drosophila.
Consistent with histochemical and electrophysiological data (Bicker 1999; Grunewald et al. 2004; Barbara et al. 2005), putative GABA and GluCl receptor forming subunits have been identified in the honeybee genome. These include orthologs to Drosophila Grd and LCCH3 which are able to form heteromultimeric GABA-gated cation channels in Xenopus oocytes (Gisselmann et al. 2004). In other invertebrates, a GABA-gated cation channel has been identified in C. elegans (Beg and Jorgensen 2003) and physiological experiments indicate that GABA mediates excitation by activating cation channels in snail (Zhang et al. 1997) and crab (Swensen et al. 2000). Electrophysiological studies on cultured honeybee antennal lobe neurons and Kenyon cells, however, show that GABA induces inhibitory chloride currents (Grunewald et al. 2004; Barbara et al. 2005). Thus, it remains to be established whether GABA-gated cation channels function in insects in vivo. Recently, an A. mellifera HisCl cys-loop LGIC sequence was reported (Iovchev et al. 2006) and we observe in the honeybee genome two putative HisCl subunits (Amel_HisCl1 and Amel_HisCl2) as well as the HisCl-like Amel_12344 (Fig. 2) (Dent 2006). This suggests the presence of histamine-gated channels in honeybee, although in light of electrophysiological data where histamine did not induce measurable currents in cultured antennal lobe neurons whereas glutamate and GABA did (Barbara et al. 2005), the range of their function may be restricted when compared with GluCl and GABA receptors. The honeybee cys-loop LGIC superfamily also contains uncharacterised subunits (Amel_8916, Amel_12344 and Amel_6927), all of which have orthologs in Drosophila. Whilst there are ionotropic serotonin-gated cation receptors in vertebrates (Reeves and Lummis 2002) and a serotonin-gated chloride channel (MOD-1) has been identified in C. elegans (Ranganathan et al. 2000), it remains to be determined whether there are serotonin-gated ion channels in insects. We performed phylogenetic analysis which showed that neither the rat 5HT3A subunit (Accession No. NP_077370) nor MOD-1 was closely related with any cys-loop LGICs of Drosophila or Apis (data not shown). Thus, no obvious insect serotonin-gated ion channel candidates were identified. It will be of interest to determine the neurotransmitters that act on the uncharacterised insect cys-loop LGICs as well as their functional roles.
As with Drosophila (ffrench-Constant and Rocheleau 1993; Semenov and Pak 1999; Lansdell and Millar 2000; Grauso et al. 2002; Schnizler et al. 2005), splice variants diversify the Apis cys-loop LGIC superfamily (Jones et al. 2006) (Fig. 3). For example, alternative splicing introduces amino acid substitutions in functionally significant regions such as loops F and C of Amel_Rdl (Fig. 3a). Indeed, studies of Drosophila Rdl expressed in Xenopus oocytes show that splice variants differ in their apparent agonist affinity (Hosie et al. 2001). Also, alternative splicing generates insect-specific variants. For example, Rdl exon 6c of Drosophila and Apis differs remarkably in the region of loop F as well as loop C, including the presence of an N-glycosylation site in Drosophila only (Fig. 3a). Since N-glycosylation can affect processing and assembly (Buller et al. 1994; Griffon et al. 1999) as well as channel desensitization and conductance of cys-loop LGICs (Nishizaki 2003), alternative splicing has considerable potential to alter receptor properties in different insect species. It is worth noting that a residue in loop C along with the loop F region contribute to the high selectivity of imidacloprid for nAChRs of insects rather than vertebrates (Shimomura et al. 2004), further highlighting the potential to exploit splice variants as targets for the control of particular insect species.
Species-specific splice variants have also been observed for Drosophila pHCl (Schnizler et al. 2005) and Amel_pHCl where the length of insertions differs considerably between the two species. Also, an N-glycosylation site is introduced only in Variant 1 of Apis (Fig. 3c). The use of an alternative exon that spans the TM1 region of Drosophila pHCl (Variant 3) was not observed in its Apis ortholog. However, unlike its Drosophila counterpart, Amel_pHCl was found to have an insertion that may lengthen considerably loop C (Variant 4), which is likely to have an impact on ligand binding. In Amel_Rdl and Amel_HisCl1, the use of differential splice sites disrupts and/or introduces potential phosphorylation sites in the intracellular region between TM3 and TM4 (Fig. 3d, e). Since phosphorylation of this region modulates receptor function (Smart 1997) such splice variants will likely diversify subunit activity. It will be of interest to see whether similar variants exist in other insects. In this regard it is interesting to note that a Drosophila HisCl1 variant has been cloned with a four amino acid deletion in its TM3-TM4 intracellular loop (Zheng et al. 2002). However, no phosphorylation sites were affected and no functional difference was observed between the long and short forms.
The TM3-TM4 intracellular loop of Amel_8916 contains putative bipartite nuclear targeting sequences (Dingwall and Laskey 1991) which were not observed in its Drosophila ortholog, CG8916. Interestingly, the putative N-terminus of Amel_8916, which was identified from an EST (Accession BI509369), does not possess a potential signal leader peptide sequence (Dyrlov Bendtsen et al. 2004), consistent with the inferred intracellular location of Amel_8916. To our knowledge, this is the first cys-loop LGIC reported to possess nuclear targeting sequences. If indeed Amel_8916 is located in the nucleus, its functional role is likely to differ dramatically from that of Drosophila CG8916. Other types of ion channels are present in the nucleus. Examples include: ryanodine and inositol 1,4,5-trisphosphate receptors which mediate intranuclear calcium signalling (Humbert et al. 1996; Marius et al. 2006); and chloride channel protein CLIC4 which increases the rate of apoptosis (Suh et al. 2004). It is interesting to speculate that Amel_8916 may also modulate processes within the nucleus.
The identification of the full complement of honeybee cys-loop LGIC subunits presents an important basis for understanding the diversity of cys-loop LGIC superfamilies in different organisms, studying key components of the insect nervous system, characterising pesticide targets as well as developing insecticides with improved selectivity. The comparison of Apis and Drosophila insect cys-loop LGIC superfamilies has highlighted species-specific proteome diversification arising from differential splicing and RNA editing. Genome projects have been either completed or are in progress for several insect pest species such as the malaria mosquito A. gambiae (Holt et al. 2002), the yellow and dengue fever mosquito Aedes aegypti (http://www.msc.tigr.org/aedes/aedes.shtml), the red flour beetle Tribolium castaneum (http://www.hgsc.bcm.tmc.edu/projects/tribolium/) and the pea aphid Acyrthosiphon pisum (http://www.hgsc.bcm.tmc.edu/projects/aphid/). Sequence information produced from such projects will provide further insights into the diversity of insect cys-loop LGIC superfamilies, all of which represents promising differences to target for future rational insecticide design. The use of heterologous expression systems such as Xenopus oocytes has allowed the functional characterisation of several Drosophila cys-loop LGICs such as Rdl (ffrench-Constant et al. 1993), GRD and LCCH3 (Gisselmann et al. 2004), GluCl (Cully et al. 1996), HisCl1 and HisCl2 (Gisselmann et al. 2002; Zheng et al. 2002), and pHCl (Schnizler et al. 2005). Similar studies of heterologously expressed Apis ion channels, in combination with the use of three-dimensional models based on the crystal structure of molluscan acetylcholine binding proteins (Ernst et al. 2003; Smit et al. 2003; Celie et al. 2005; Sattelle et al. 2005), will likely prove useful in screening for novel compounds that show low selectivity for honeybee receptors and in determining the mechanisms of insecticide interaction with cys-loop LGICs.
We are indebted to the A. mellifera Genome Project (Human Genome Sequencing Center at Baylor College of Medicine) which provided the starting point for this study. We wish to thank Valerie Raymond-Delpech for providing adult honeybees. The financial support of the Medical Research Council and Dupont is gratefully acknowledged.