SEARCH

SEARCH BY CITATION

Keywords:

  • Aphid;
  • ion channel;
  • genome

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Aphids are major pests of crops, causing hundreds of millions of dollars worth of damage annually. Ion channel proteins are often the targets of modern insecticides and mutations in ion channel genes can lead to resistance to many leading classes of insecticides. The sequencing of the pea aphid, Acyrthosiphon pisum, genome has now allowed detailed in silico analysis of the aphid ion channels. The study has revealed significant differences in the composition of the ion channel families between the aphid and other insects. For example A. pisum does not appear to contain a homologue of the nACh receptor alpha 5 gene whilst the calcium channel beta subunit has been duplicated. These variations could result in differences in function or sensitivity to insecticides. The genome sequence will allow the study of aphid ion channels to be accelerated, leading to a better understanding of the function of these economically important channels. The potential for identifying novel insecticide targets within the aphid is now a step closer.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Many aphid species are important crop pests causing estimated losses worth hundreds of millions of dollars annually. For example the peach-potato aphid, Myzus persicae, infests a large number of commercially important food plants including sugar beet, potatoes, lettuce, sweet pepper and tomato, whereas cereal crops are attacked by other species such as the Russian wheat aphid, Diuraphis noxia. Aphids cause both direct feeding damage and transmit pathogenic plant viruses.

Ion channels are responsible for the movement of ions across cell membranes and are essential components of the nervous system. They are also involved in numerous critical functions within the insect such as muscle contraction and ion homeostasis. There are several families of ion channels that have different specificities for the ions transported. Ion channels can be activated by voltage-sensing or by ligand binding, allowing the essential fine control of ion channel activity. In insects, the ion channel proteins are the target of many natural venoms and toxins used by predators to render the insect immobile. Many insecticides act by targeting ion channels (Raymond-Delpech et al., 2005), thereby interfering with nervous system function, leading to the death of the insect. There are relatively few chemical classes of insecticide available for use against aphids, the main ones being neonicotinoids, which target nicotinic acetylcholine (nACh) receptors, and pyrethroids, which target voltage-gated sodium channels. Aphids resistant to these classes of chemical are increasing, as seen with the spread of kdr (knockdown resistance) to pyrethroids (Martinez-Torres et al., 1999), leading to an urgent need to develop new insecticides with differing modes of action. Recently, a new class of insecticides was launched, targeting ryanodine receptors of lepidopteran and sucking pests. This class includes Cyazypyr™ and Rynaxypyr® (Sattelle et al., 2008) and these compounds will enable aphid control in previously resistant populations.

The sequencing of the pea aphid, Acyrthosiphon pisum, genome has provided an opportunity to study the complete set of genes encoding aphid ion channels. Here we report the genes present and compare them with their counterparts in other insects. It is hoped that the identification of these genes will lead to a better understanding of aphid ion channels, and highlighting the differences between insect channels will lead to an improvement in knowledge of the differences in efficacy of current insecticides among aphids and other insects. In addition, the genome sequence may facilitate the design of aphid-specific insecticides to better control these important pests.

Herein we follow the tradition of distinguishing between voltage-gated and ligand-gated channels, although this is of course an oversimplification.

Results and discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Voltage-gated sodium and calcium channels

Voltage-gated sodium channel (para). Voltage-gated sodium channel (VGSC) α-subunits are integral membrane proteins that conduct sodium ions through a cell's plasma membrane. In electrically excitable cells, they are responsible for the rising phase of action potentials. The genomes of insects for which sequence data is currently available contain a single identifiable VGSC gene named ‘para’ (Davies et al., 2007a). A putative para homologue has been identified in the A. pisum genome (Supplementary Table S1). The predicted protein has: the expected four repeat domains, each containing 6 transmembrane (TM)-spanning segments (S1–S6); positively charged S4 regions which act as the channel's voltage sensor; and extracellular/cytoplasmic pore regions formed by the ‘P-loops’ (located between S5 and S6 of the four domains and the S5 and S6 transmembrane regions, respectively). Surprisingly, the channels sodium ‘selectivity filter’ is made up of amino acids D, E, N and S (located on the P-loops of each domain) rather than the classic DEKA filter (Catterall, 2000), which normally defines a VGSC. Point mutations within VGSCs have been implicated in kdr and super-kdr-like resistance to the pyrethroid insecticides used in aphid control (reviewed in Davies et al., 2007b).

Insect VGSCs are presumed to gain heterogeneity and functional diversity by extensive alternative splicing (Olson et al., 2008). Nine such splicing sites have been identified in para and its orthologues to date; seven exons, a, b, i, j, e, f, h, are optional and form parts of the intracellular linkers of mature channel proteins, whereas exons c/d and l/k are mutually exclusive and are situated in the transmembrane regions. Exons l/k have been implicated in differential sensitivity to pyrethroid insecticides (Du et al., 2006). Exons l/k are evolutionarily highly conserved, and are present in the A. pisum sequence, as are exons a, b. Given the absence of some splice sites in the aphid gene, it seems that fewer variants of this channel will be present in the pea aphid compared with, for example, Drosophila melanogaster, although there may be other unidentified splice sites.

Voltage-gated sodium channel auxillary subunits. High level expression of insect VGSCs, with maximal current and optimal kinetics, requires a small auxillary subunit, TipE (Warmke et al., 1997). Four additional TipE-like proteins (TEH 1-4) have been identified in D. melanogaster (Derst et al., 2006). We have identified genes encoding all five of these auxiliary subunits in A. pisum on a ∼123 kb genome tract (Supplementary Table S1).

Sodium channel 60E. Like the VGSC α-subunit genes, the D. melanogaster sodium channel 1 (DSC1) gene (Salkoff et al., 1987) and the orthologous Blatella germanica BSC1 gene (Liu et al., 2001) belong to the 4 × 6 TM domain family. Although originally classified as sodium channels based on their similarity to para, they actually encode calcium-selective channels (Zhou et al., 2004). The partially reduced expression of DSC1 in a smell-impaired D. melanogaster mutant smi60E correlates with certain olfactory defects. (Kulkarni et al., 2002), suggesting a role in olfactory avoidance response. A putative homologue has been identified in the A. pisum genome (Supplementary Table S1).

Pickpocket (ENaC/DEG) channels. The epithelial sodium channel (ENaC)/degenerin (DEG) gene family encode membrane-bound, voltage-insensitive ion channels (Kellenberger & Schild, 2002). Individual subunits assemble as homotrimers or heterotrimers to form functional channels (Jasti et al., 2007). There are more than thirty predicted ENaC/DEG subunits in D. melanogaster (Littleton & Ganetzky, 2000) but only a handful have been characterized (Supplementary Table S1). Pickpocket (ppk), the first subunit to be identified, is involved in locomotive behaviour in fruit fly larvae. Other ppk genes have been implicated in fluid re-distribution (Adams et al., 1998; Liu et al., 2003a), gustatory salt perception (Liu et al., 2003b), and male response to female pheromones (Lin et al., 2005). Preliminary analysis of the pea aphid genome has indicated that there are between 23 and 26 ppk genes present with conserved structural features, but low overall sequence similarity.

Calcium channels. Calcium channels are formed from heteromeric units containing a pore-forming α subunit, a β subunit and α2δ subunits, combinations of which allow distinct channel types to be formed. D. melanogaster has four α subunits, three of which correspond to the mammalian L, N and T channels and one which is related to a calcium channel found in Caenorhabditis elegans (Littleton & Ganetzky, 2000). In A. pisum we have identified a single gene for the L-, N- and T-type α subunits, however, no match to the C. elegans-related CG13762 gene was found (see Supplementary Fig. S1).

The auxiliary subunits of calcium channels (β and α2δ subunits) have a role in targeting the α subunit to the plasma membrane and also in aligning the α subunit into the correct conformation. We identified two orthologues of the β subunit in A. pisum using the D. melanogasterβ subunit gene as the target. These proteins, ACYPI005770 and ACYPI008838, show 71% and 47% similarity to the D. melanogaster orthologue, respectively. D. melanogaster also has three α2δ subunits, which are post-translationally cleaved into two peptides (α2 and δ) (Dickman et al., 2008), and the analysis of the A. pisum genome reveals four orthologues of these genes with the α2δ 1 subunit being duplicated (Supplementary Table S2).

Narrow abdomen or Dma1U. Dma1U in D. melanogaster (Nash et al., 2002) belongs to a unique family of putative 4 × 6 TM ion channels related to voltage-gated sodium and calcium channels. These channels diverge from voltage-gated sodium/calcium channels by virtue of a unique selectivity filter sequence and a reduced number of positively charged amino acids (arginine or lysine) in the voltage-sensing fourth transmembrane domains, suggesting that they may form channels with distinctive properties. Dma1U corresponds to the narrow abdomen (na) gene, a locus for point mutations associated with changes in abdominal morphology, locomotor behaviour and day/night circadian rhythms (Lear et al., 2005). We have identified a gene corresponding to na in A. pisum (Supplementary Table S1); its function is likely to be similar to that of the D. melanogaster channel since the sequence is highly conserved (∼80% identity).

Potassium channels

Potassium channels are the largest family of ion channels in invertebrates and have a wide range of roles, including the maintenance of the membrane potential in cells, maintenance of the circadian rhythm and potassium homeostasis. There are several distinct families of potassium channels, both voltage-gated and non-voltage-gated and each family is responsible for various different currents. In this study, we have identified thirty-two potassium channel genes in A. pisum (Supplementary Fig. S2 and Supplementary Table S3).

Voltage-gated potassium channels:

  • Shaker potassium channels The Shaker family of potassium channels control membrane potential in a variety of cells and have six transmembrane domains, S1-S6. Mutations in the potassium channel Shaker in D. melanogaster give rise to disturbances of the transient outward potassium (iA) current. Other members of the Shaker family include ShaI, ShaB, ShaW and ShaWl which contribute to the transient and delayed rectifier currents in neurons and muscles. D. melanogaster has a single gene for ShaB and ShaI, whilst ShaW is encoded by two genes, ShaW and ShaWl (Littleton & Ganetzky, 2000). We have identified an orthologue of each of these genes in A. pisum. The aphid ShaB protein model is highly divergent from D. melanogaster at the N and C terminals and is much shorter at 558 amino acids compared with 985aa for D. melanogaster. ShaW and Shaker are extremely well conserved throughout, whereas ShaI and ShaWl are more divergent in the C-terminal regions and are again shorter in length than the D. melanogaster protein.

  • KCNQ channels The KCNQ potassium channel is encoded by a single gene in D. melanogaster and is responsible for the M-current (Littleton & Ganetzky, 2000). P-element insertion lines in the KCNQ gene have shown that the KCNQ channel is involved in control of heart rhythm and the knockout line showed increased arrhythmias compared with wild-type controls (Ocorr et al., 2007). A. pisum contains a single KCNQ gene, hmm111514, which is currently modelled as a pseudogene as it contains a premature stop codon. The aphid protein model shows low homology, 24%, to the D. melanogaster protein and is especially divergent at the C-terminal end.

  • Eag potassium channels There are three sub-families of the Eag potassium channels; Eag, Egl and Erg. Although their basic structure is similar to the voltage-gated potassium channels with six transmembrane domains and a pore domain, the Eag channels also encode a C-terminal sequence with a cyclic-nucleotide binding domain (Guy et al., 1991). D. melanogaster Eag mutants display a phenotype of leg-shaking, caused by repetitive firing and enhanced neurotransmitter release in the motor neurons, whereas homozygous mutations in the D. melanogaster erg (Seizure) gene result in paralysis (Warmke et al., 1994; Wang et al., 1997a). We have identified three Eag-type channels in A. pisum. The aphid Eag and Egl proteins are shorter than the D. melanogaster orthologues, having much shorter N- and C-terminal sequence. The aphid Erg protein, however, has a longer C terminus but shorter N terminus.

Non-voltage gated potassium channels:

  • Calcium-activated potassium channels This group of potassium channels are heavily involved in the control of cell excitability (Derst et al., 2003). The group is divided into large (BK), intermediate (IK) and small-conductance (SK) channels that are regulated by both voltage gating and calcium ions. A C-terminal calcium-binding motif known as the ‘calcium bowl’ is encoded by the amino acid sequence DDDDD (Bian et al., 2001). In D. melanogaster a single gene, Slowpoke (pSlo), is responsible for the large-conductance current and associates with a second gene, Slack, to form the IK channel. The SK channel is also encoded by a single gene. We have a single homologue of each of these genes in A. pisum. The pSlo protein is especially well conserved with 82% identity between the aphid and D. melanogaster sequences. The current aphid Slack protein model shows good conservation with D. melanogaster, with the exception of two large deletions (250aa and 200aa) in the central section of the protein.

  • Inward-rectifying potassium channels Three inward-rectifying potassium (irk) channel genes have been identified in D. melanogaster, dIr, dIrk2 and dIrk3.These are involved in neuronal excitability and maintenance of potassium levels. It has been shown that renal tissues have high expression of all three irk genes, indicating a role in fluid secretion (Evans et al., 2005). We have identified two genes in A. pisum, which show close homology to dIr and dIrk2.

  • Two-pore, weakly inward rectifying potassium channels (TWIK) TWIK channels, characterized by four transmembrane domains and two potassium selective pores, are involved in a diverse range of processes including mediating resting potential and mechanosensation (Buckingham et al., 2005). We have identified 16 putative TWIK genes in A. pisum, all containing the potassium ion selectivity motif GYGD. One of these genes, ACYPIG684244, is currently modelled as a pseudogene.

Chloride channels

Voltage-gated chloride channels  Voltage-gated chloride channels are involved in several biological and physiological functions such as pH regulation, volume homeostasis and organic solute transport. The channels are characterized by the presence of twelve transmembrane domains. Mammals have ten genes encoding voltage-gated chloride channels whereas only three were found in D. melanogaster (Littleton & Ganetzky, 2000). We have found five in A. pisum (Supplementary Table S4), with three of them being similar to D. melanogaster CG8594 gene.

Ionotropic glycine and γ-aminobutyric acid and glycine receptors. Ionotropic glycine and γ-aminobutyric acid (GABA) receptors are chloride channels mediating rapid, mostly inhibitory, synaptic transmission. Both GABA and glycine receptors belong to the dicysteine-loop (‘Cys-loop’) superfamily of neurotransmitter receptors (Grenningloh et al., 1987; Schofield et al., 1987). These chloride channels are pentameric (Nayeem et al., 1994) and they are the target of cyclodiene and phenylpyrazole insecticides.

In vertebrates, these GABA receptors have been divided into two families: GABAA, characterized by bicuculline sensitivity (Sieghart, 1995), and bicuculline-insensitive GABAC receptors (Woodward et al., 1993). Several GABAA (α 1-6, β 1-4, γ1-4, δ, ε, θ and π 1-3) and five GABAC (ρ1-5) genes have been identified with the GABAA receptors having two α, two β and one of the other types of subunits (δ, ε, θ or π) (reviewed in Kash et al., 2004). The glycine receptors are composed of three α and two β subunits and in humans, fourα (α1–α4) and one β genes have been identified (Bowery & Smart, 2006).

Both ionotropic GABA and glycine receptor subunits have large N-terminal and short C-terminal extracellular domains and four hydrophobic transmembrane domains (M1–M4) with a large intracellular loop of variable length between M3 and M4. In D. melanogaster, ten different genes encoding proteins with similarity to GABA/glycine subunits have been identified (Littleton & Ganetzky, 2000). The GABA receptors are bicuculline-insensitive but are regulated by some though not all of the allosteric modulators of GABAA receptors. Splice variants have been identified in some of these genes (Hosie et al., 2001). In A. pisum, we have identified ten genes encoding proteins with similarity to GABA/Glycine subunits (Supplementary Table S4) and interestingly, two of the genes show similarity to the D. melanogaster Resistant to dieldrin (Rdl) gene and two to D. melanogaster pH-sensitive chloride Channel (pHCl).

There are five genes in D. melanogaster, present in other insect species, with no identifiable orthologues in A. pisum (Supplementary Table S4). Two of the genes identified in A. pisum encode subunits not present in D. melanogaster and are most similar to human Glycine receptor α3.

Two other putative chloride receptors genes have been identified in A. pisum, a histamine-gated chloride channel subunit 1 (HisCl1) and glutamate-gated chloride channel (GluClα) (Supplementary Table S4).

Cyclic nucleotide-gated ion channels

Cyclic nucleotide-gated (CNG) ion channels are activated by cyclic nucleotides, cAMP or cGMP, and play an important role in vision and olfaction. They are tetramers, with each subunit having an intracellular N- and C-terminal region and six transmembrane spanning domains with a pore-loop domain between the fifth and sixth transmembrane domain. The cyclic nucleotides bind the C-terminal domain. Four CNG genes have been identified in D. melanogaster (Littleton & Ganetzky, 2000). The search of orthologous genes in the A. pisum genome yielded five putative genes: one putative orthologue for each one of the D. melanogaster genes (Supplementary Table S5). There are two putative genes encoding the D. melanogaster CG7779, but these are incomplete and probably represent two halves of the same gene.

Glutamate receptors

The ionotropic glutamate receptor family mediates most forms of excitatory synaptic transmission in the mammalian CNS and they play an important role in learning and memory. In mammals, four categories have been identified according to the agonist response: alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, kainate, N-Methyl-D-Aspartate (NMDA) and δ. All of the receptor subunits are thought to adopt the same membrane topology and each subunit is composed of an extracellular N-terminal domain, three transmembrane domains, a single ‘P-loop’ membrane domain and an intracellular C-terminal domain that interacts with scaffolding proteins (Hollmann et al., 1994; Roche et al., 1994; Bennett & Dingledine, 1995). Approximately 30 ionotropic glutamate receptor genes were identified in D. melanogaster (Littleton & Ganetzky, 2000). Invertebrates have additional subtypes, absent in mammals, which have low similarity to the Arabidopsis thaliana Glur5 glutamate receptor. Some glutamate receptors are the target of the avermectin family of insecticides.

In mammals, NMDA receptors are heterotetramers and seven subunits have been identified encoded by the genes NR1, NR2A, NR2B, NR2C, NR2D, NR3A and NR3B. A functioning receptor requires at least two subunits: NR1 and NR2 (Stephenson et al., 2008) and diversity is generated by the association of the obligatory NR1 subunit with different NR2 subunits.

In D. melanogaster, two NMDA receptor genes have been identified (NMDAR1 and NMDAR2) and our analysis of the A. pisum genome shows the presence of one copy of each one of these genes. One copy of NMDAR3 was also identified (Supplementary Table S6 and Supplementary Fig. S3), a gene which is absent from D. melanogaster but present in other insects such as Anopheles gambiae, Tribolium castaneum, Apis mellifera and Nasonia vitripennis. Two other genes identified in D. melanogaster and described as showing NMDA selective glutamate activity in Flybase, CG3814 and CG30379, could not be identified in A. pisum; these genes have only been identified in other Drosophila species.

AMPA and kainate receptors mediate fast synaptic transmission in the CNS and both are described as ionotropic, although there is evidence that they can be metabotropic, inducing the activation of G proteins and signalling cascades (Wang et al., 1997b; Huettner, 2003; Lerma, 2003). Kainate receptors are tetramers. Five different subunits have been identified, GLUK5, GLUK6, GLUK7, GLUK1 and GLUK2 (Hollmann & Heinemann, 1994), divided into two families based on sequence similarity.

Ten genes encoding kainate selective glutamate receptors were identified in D. melanogaster. We have identified seven putative genes in the A. pisum genome (Supplementary Fig. S4 and Supplementary Table S6) but were unable to identify genes homologous to KaiRIA, GluRIIa, GluRIIB, GluRIIC and GluRIIE, present in members of the Drosophilidae family. The gene CG5621 also seems to be absent from A. pisum despite being present in Aedes aegypti, An. gambiae and Culex quinquefasciatus. Interestingly in A. pisum, there is a cluster of kainate receptor genes located in the same scaffold, with three genes with similarity to CG3822 and one gene similar to clumsy. There is another gene similar to CG3822 located in another cluster.

The AMPA receptors, have four types of subunits, GluR1, GluR2, GluR3, and GluR4 (also called GluRA-D), and are usually heterotetrameric, consisting of symmetric ‘dimer of dimers’ of GluR2 and either GluR1, GluR3 or GluR4. Only two AMPA receptors have been described in D. melanogaster, GluRI and GluRIB. We have found two genes in the A. pisum genome, both similar to GluRIB that might be two halves of the same gene (Supplementary Fig. S4 and Supplementary Table S6). In addition to these receptors, several novel subtypes have been identified in D. melanogaster with no mammalian orthologues: CG32704, CG15627, CG7385, CG6185, CG2657 and CG42315. One orthologue of each has been identified in A. pisum (Supplementary Fig. S5 and Supplementary Table S6). Another ionotropic channel, CG14076, has six copies in A. pisum, two located in the same scaffold (Supplementary Fig. S5 and Supplementary Table S6). Several other ionotropic receptors present in D. melanogaster are not found in A. pisum.

Nicotinic acetylcholine receptors

Nicotinic acetylcholine (nACh) receptors are prototypical members of the Cys-loop ligand-gated ion channel (cysLGIC) superfamily, consisting of five subunits arranged around a central ion channel. In insects, nACh receptors mediate the fast actions of acetylcholine in the nervous system and are of particular interest as they are targets of a major group of insecticides, the neonicotinoids, which have been the fastest-growing class of insecticides in modern crop protection (Jeschke & Nauen, 2008). The A. pisum genome contains 11 genes encoding putative nACh receptor subunits (Fig. 1), a similar number to that found in An. gambiae, Ap. mellifera, Bombyx mori, D. melanogaster and T. castaneum, all with between 10 and 12 subunits. Eight of the aphid subunits fall into core groups (Dα1 Group, Dα2 Group, etc.), which are highly conserved (over 60% amino acid identity) between different insect species (Jones et al., 2007). As in other insects, A. pisum also has divergent subunits, which show particularly low sequence identity (20% or less) with other nACh receptors and may perform species-specific roles. The complement of A. pisum divergent subunits, comprising one β (Apisumβ2) and two α (Apisum α9 and Apisum α10) subunits, is currently unique to the aphid. Apisumα4 and Apisumα6 have alternative splicing of exons 4 and 8, respectively, which is also seen in orthologous subunits of other insect species (Supplementary Fig. S6). Such splicing effectively diversifies receptor function and generates species-specific isoforms by introducing variations in (1) the Cys-loop that is involved in receptor assembly (Lansdell & Millar, 2000), (2) loops E and B that contribute to ligand binding, and (3) the second transmembrane domain that forms the ion channel (Grauso et al., 2002).

image

Figure 1. Tree showing relationships of Acyrthosiphon pisum, Apis mellifera and Drosophila melanogaster nAChR subunit protein sequences. Numbers at each node signify bootstrap values with 1000 replicates and the scale bar represents substitutions per site. The nAChR subunits shown in the tree (as well as GenBank accession numbers) are as follows: Da1 (CAA30172), Da2 (CAA36517), Da3 (CAA75688), Da4 (CAB77445), Da5 (AAM13390), Da6 (AAM13392), Da7 (AAK67257), Db1 (CAA27641), Db2 (CAA39211), Db3 (CAC48166), Amela1 (DQ026031), Amela2 (AY540846), Amela3 (DQ026032), Amela4 (DQ026033), Amela5 (AY569781), Amela6 (DQ026035), Amela7 (AY500239), Amela8 (AF514804), Amela9 (DQ026037), Amelb1 (DQ026038), Amelb2 (DQ026039). The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987) available with the CLUSTALX program (Thompson et al., 1997) and displayed using TreeView (Page, 1996).

Download figure to PowerPoint

We have found features particular to the A. pisum nACh receptor gene family. For example, the aphid is the first insect that appears not to possess a Dα5 orthologue (Fig. 1). In this regard, it is worth noting that the closely related Apisumα7 subunit has unusual tandem duplications of exons 6 and 7, which likely represent alternatively spliced exons (Supplementary Fig. S6). It is interesting to speculate that the genome duplication events seen in Apisumα7 may represent the early steps of diversification leading to the generation of a separate gene identified as α5 in more highly evolved insects. This finding would imply that α5 is the newest member of the insect core group of subunits which is perhaps subjected to fewer functional constraints, because Amelα5, Tcasα5 and Bmα5 show considerably more divergence in amino acid sequence than any other core subunit orthologues (Jones et al., 2006; Jones & Sattelle, 2007; Shao et al., 2007). The two aphid divergent subunits, Apisumα9 and Apisumα10, are the first examples of insect nACh receptor subunits without both cysteine residues of the characteristic Cys-loop motif (Table 1). Instead, both subunits have the first cysteine while the second is replaced by a glycine residue. The Cys-loop plays a key role in receptor assembly (Green & Wanamaker, 1997) and removal of one cysteine from the Torpedo californicaα subunit altered the response to ACh (Sumikawa & Gehle, 1992). Thus, Apisumα9 and Apisumα10 may show distinct functional properties. As neonicotinoid insecticides are frequently used to combat outbreaks of aphid infestation of crops (Jeschke & Nauen, 2008), knowledge of the receptor family targeted by such chemicals will accelerate our understanding of the molecular basis of aphid control.

Table 1.  Cys-loop sequences of divergent insect nAChR subunits Thumbnail image of

Transient Receptor Potential channels

Transient Receptor Potential (TRP) proteins have six transmembrane domains and form calcium-permeable cation channels. The first description of a TRP gene was in D. melanogaster but there have been relatively few comparative studies on the TRP genes in insects. D. melanogaster has 13 well described genes, in seven sub-families (Supplementary Table S7) (Montell, 2005). Most were identified from mutant flies which display aberrant sensory perception phenotypes and many TRP channels do seem to be involved in various forms of sensation. In contrast, A. pisum seems to have only eight or nine trp genes, with no representatives of the TRPML and TRPP sub-families (Supplementary Fig. S7 and Supplementary Table S7). Interestingly, these two sub-families are the most distantly related members of the whole gene super-family and have been referred to as the group-2 subfamilies. As aphids are primarily parthenogenetic, it is perhaps not surprising that a gene, the TRPP Pkd-2, with a role in sperm motility, is absent. The aphid genome, then, only has representatives of the group 1 trp sub-families.

The best characterized of these subfamilies is TRPC, or classical TRP, which is involved in photoreception (Hardie & Minke, 1992). TRPL (TRP-like) and TRP-γ form heteromultimeric complexes with TRP (Phillips et al., 1992; Xu et al., 2000). Together, these three genes make up the TRPC subfamily. In D. melanogaster, these channels are held in a signalling complex involving the scaffolding protein InaD (Georgiev et al., 2005). Only one member of the TRPC sub-family was found in the aphid genome and InaD is not obviously present, suggesting that the physiology of visual signalling is markedly different in the aphid.

As in D. melanogaster, there are two members of the TRPV subfamily, one each related to nanchung and inactive, though the aphid equivalent of inactive is more similar to that of A. mellifera than to D. melanogaster. In flies, these genes are involved in sound perception (Kim et al., 2003) and the responses to humidity (Liu et al., 2007) and heat (Chentsova et al., 2002). The single insect member of the TRPM family is also conserved in the aphid. The function of this gene has not been determined, although it is known not to play a role in larval thermotaxis (Rosenzweig et al., 2005). Nompc is the sole member of the TRPN subfamily and mutations in this gene produce a phenotype of defective mechanoreception (Kernan et al., 1994; Walker et al., 2000). There is a clear orthologue of NompC in the aphid genome (ACYPIT949557).

The TRPA subfamily is the largest and most diverse in both D. melanogaster and A. pisum. Though the genomes of both species seem to contain four members of this subfamily, it has proved very difficult, based on sequence data, to deduce a gene-for-gene relationship between the two TRPA subfamilies. The one exception is TrpA1, which has a clear aphid equivalent (ACYPIT433926). In flies, this gene is involved in larval thermotaxis (Rosenzweig et al., 2005). The other three Drosophila TRPA genes, pyrexia, waterwitch (CG31284) and painless (painless is somewhat divergent) are involved in responses to heat (Lee et al., 2005), humidity (Liu et al., 2007) and in mechanosensation/nociception (Goodman, 2003; Tracey et al., 2003), respectively. The best candidate for an orthologue to any of these genes in the aphid is for waterwitch where a putative orthologue seems to have been duplicated, with two copies adjacent to each other in the inverse orientation. There is no obvious aphid equivalent of pyrexia. There is a gene distantly related to painless, ACYPIT907328, but the sequence divergence is sufficient to suggest that this may be a novel member of the TRP family, rather than an orthologue.

Ryanodine receptors

Ryanodine receptors (so named due to their interaction with an insecticidal alkaloid ryanodine) are ligand-gated calcium channels that mediate the release of calcium ions from intracellular stores. They are located in the sarcoplasmic reticulum of skeletal muscle, where calcium release is important in muscle contraction, and also in the endoplasmic reticulum of neurons and other cell types, where they are heavily involved in calcium signalling (Sattelle et al., 2008). Each receptor is composed of a large cytoplasmic domain at the N terminus and a small transmembrane domain at the C terminus (Hamilton, 2005). Functional channels present as large homomeric tetramers. The A. pisum receptor (Supplementary Table S2) is highly conserved with 77.7% identity/85.8% similarity with the D. melanogaster receptor. The sequence conservation within aphids is even more striking, with > 98% amino acid identity between A. pisum, M. persicae and Aphis gossypii channels. Compared with the D. melanogaster gene which is composed of twenty-six exons, with exons 9a/b being mutually exclusive and exon 13 having an optional splice site, the A. pisum receptor has an estimated 98 exons and no equivalent splice sites. The recently commercialized synthetic insecticide Cyazypyr™, which targets the ryanodine receptors of aphids, acts as a channel activator and binds at a site distinct from the ryanodine binding site (Sattelle et al., 2008).

Inositol 1,4,5-triphosphate (IP3) receptor

A second receptor that also releases calcium from intracellular stores is the inositol 1,4,5-triphosphate (IP3) receptor. The receptor is activated by binding of the IP3 ligand and is involved in numerous cellular response pathways and sensory mechanisms (Hasan and Rosbash, 1992). The IP3 receptor protein is highly conserved between D. melanogaster and A. pisum with amino acid identity of 62% and similarity of 72% (Supplementary Table S2).

General discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The sequencing of the A. pisum genome is an important milestone in basic insect and insecticide research. This sequencing effort and the uncovering of ion channel genes will play an important part in the development of the next generation of chemicals to control these pests in three ways. Firstly, the genome will allow an understanding of the broad architecture of aphid ion channels and will reveal the role of these genes in aphid biology. Important differences in ion channel composition between aphids and other insect species have already been identified in this study. A major finding is that aphids are the only insect to date without an orthologue of the Drosophila Dα5 subunit. Interestingly, it does have orthologues of insect subunits shown to be targeted by imidacloprid (Apisumα2) (Matsuda et al., 1998) and spinosad (Apisumα6) (Perry et al., 2007). Also, of considerable interest is the finding that two of the divergent nACh receptor subunits Apisumα9 and Apisumα10 are the only insect nAChRs known to have incomplete signature Cys-loop sequences, likely to profoundly modify their assembly and functional properties. Another example of the role of ion channels in differences in aphid biology is seen with the TRP genes. The TRP channels that are involved in vision are reduced in A. pisum when compared with D. melanogaster. In terms of sensory process, the best conserved genes between the two species appear to be those involved in detecting humidity and sound/mechanical stimuli. The genes involved in larval thermotaxis and heat avoidance are less well conserved. It is tempting to speculate that these differences reflect the relative importance of the senses in the two species.

This study has shown that the pea aphid is divergent from other insects in the number of auxiliary calcium channel subunits it encodes. The aphid has a second beta subunit and a duplicated α2δ subunit 1. These genes may act in an aphid-specific manner or be expressed in alternative tissues to give rise to novel calcium currents (Grabner et al., 1994). It is also seen that the pea aphid contains a duplicated GABA rdl subunit and the duplicated gene contains the A302S allele as is the case with Myzus persicae. The presence of this allele, however, does not confer resistance to insecticides that act at GABA chloride channels, such as cyclodienes (Anthony et al., 1998). The presence of the second gene in A. pisum would indicate that the duplication is likely to be present throughout aphid species.

The second important role of the genomic information will be in the design and development of new insecticides to control aphid infestations. The genome data is likely to include novel target proteins that insecticides can act upon either as agonists or antagonists. The ion channel amino acid sequences will also be of use to researchers trying to identify how chemicals bind to these receptors and what domains are involved in the binding. The genome data will allow the development of better tools such as specific cell lines for binding studies and in this way accelerate research into new chemistries. An important example of this will be in the better understanding of how the new ryanodine receptor-inhibiting chemicals bind to the target and what motifs in the protein sequence are responsible for the aphid selectivity.

The third use of the genomic data will be in the area of insecticide resistance research. Through knowledge of the sequences of the target genes of insecticides, gene sequencing of suspected resistant aphids can proceed rapidly without the complication of lengthy gene cloning steps. This will allow novel mutations causing insecticide resistance to be identified and quickly characterized, and diagnostic tests to be developed to enable early screening of aphid populations. In this way, the problem of resistance will be easier to contain and manage in resistance management programs.

Some of the current gene models produced through the in silico analysis of the genome will need to be confirmed through expressed sequence tag (EST) data. Currently, this level of EST coverage is not available for A. pisum but it is expected to be generated within 12 months. Once this information is available then the current models will no doubt be refined and extra detail added such as splice variation. Future research on the aphid ion channels may focus on understanding which subunits combine together to form the channels, and in which tissues or developmental stages these are expressed. The roles of the novel aphid ion channel proteins will also need to be addressed.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

To identify putative nAChR subunits, the A. pisum genome was screened with every member of the Ap. mellifera, D. melanogaster and T. castaneum nAChR gene families, using the TBLASTN algorithm. Candidate aphid nAChR subunits were identified based on their considerable sequence homology with previously characterized subunits, particularly in the N-terminal ligand-binding domain and the four transmembrane regions.

Sequences of the TRP gene family from D. melanogaster were used to perform reciprocal BLAST searches of the genome sequence database of A. pisum. Translated Drosophila sequences were used to search the aphid genome sequences using the TBLASTN algorithm. Sequences which showed a significant match to the Drosophila sequence used in the initial query were then used to search the NCBI sequence database, using the TBLASTX algorithm, in order to ascertain whether they were likely to be true orthologues of the Drosophila gene in question. In cases where the reciprocal BLAST criteria were not fully met, further searches of the aphid genome were performed using the most similar sequence found in the database, in order to assemble fragments of any aphid TRP gene sequences with no clear Drosophila orthologue.

For sodium channels, potassium channels, calcium channels and the ryanodine receptor, TBLASTN and BlastP searches of the A. pisum genome were carried out using D. melanogaster orthologues. The Genome Informatics Laboratory of Indiana University Biology Department, which provides computed annotations for the first draft genome sequence release (Acyr_1.0) of A. pisum (http://insects.eugenes.org/aphid/), was also queried. Softberry protein-based gene predictions software FGENESH+ (http://linux1.softberry.com/berry.phtml) was used to refine gene sequence data.

Previously identified members of the Glutamate receptor, Cyclic nucleotide-gated ion channel and chloride receptor families were retrieved for D. melanogaster and Homo sapiens. These sequences were uploaded into TBLASTN with A. pisum specified as the search set. High probability hits were compared with the Gnomon prediction in AphidBase (http://www.aphidbase.com, 2008 release). Where available, EST data were used to verify the coding sequence of each candidate. To control for possible erroneous annotation, multiple sequence alignments were performed and clustering patterns on individual phylogenetic trees were noted. Gene names were assigned following the closest Drosophila orthologue.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank the International Aphid Genomics Consortium and the Baylor College of Medicine Human Genome Sequencing Centre for making the A. pisum genome sequences publicly available prior to publication. Rothamsted Research receives grant-aided support from the BBSRC of the UK. We would also like to acknowledge the USDA grant (number 2005-35604-15446).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results and discussion
  5. General discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Figure S1. Phylogenetic tree showing relationships of Apis pisum and Drosophila melanogaster calcium channel protein sequences. Numbers at each node signify bootstrap values with 1000 replicates and the scale bar represents substitutions per site. The proteins shown in the tree (as well as GenBank accession numbers) are as follows: DmeI Alpha 1T (NP_572296), Dmel Alpha 1D (NP_602305), Dmel Alpha Cacophony (NP_511133), Dmel Ca Beta 1 (NP_523546), Dmel CG12295 (NP_610902), Dmel CG4587 (NP_001097163), Dmel CG12455 (NP_609779) and Dmel CG13762 (NP_570013).The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987) available with the CLUSTALX program (Thompson et al., 1997) and displayed using TreeView (Page, 1996).

Figure S2. Phylogenetic tree showing relationships of Acythosiphon pisum and Drosophila melanogaster potassium channel protein sequences. Numbers at each node signify bootstrap values with 1000 replicates and the scale bar represents substitutions per site. The proteins shown in the tree (as well as GenBank accession numbers) are as follows: DmeI Slo (NP_524486), Dmel Slack (NP_001097259), Dmel KCNQ (NP_788300), Dmel ShaWl (NP_001097131), Dmel ShaW (NP_476721), Dmel ShaB (NP_728783), Dmel Shaker (NP_728124), Dmel ShaI (NP_524159), Dmel Ih (NP_001033947),Dmel Seizure (NP_476713),Dmel Eag (NP_511158),Dmel Egl (NP_477009),Dmel SK (NP_726988),Dmel Irk1 (NP_651131), Dmel Irk2(NP_651149) and Dmel Irk3 (NP_609903). The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987) available with the CLUSTALX program (Thompson et al., 1997) and displayed using TreeView (Page, 1996).

Figure S3. Phylogenetic tree of the NMDA receptors in Drosophila melanogaster and Acyrthosiphon pisum. Tree showing relationships of A. pisum, Apis mellifera and D. melanogaster NMDA receptors protein sequences. Numbers at each node signify bootstrap values with 1000 replicates and the scale bar represents substitutions per site. The proteins shown in the tree (as well as GenBank accession numbers) are as follows: DmelNmdar1 (NP_730940), DmelNmdar2 (NP_001014716), DmelCG3814 (NP_610824), DmelCG30379 (NP_724626), AmelNmdar1 (NP_001011573), AmelNmdar2 (XP_396271), Amel CG3814 (XP_391854), AmelNmdar3 (XP_624931). The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987) available with the CLUSTALX program (Thompson et al., 1997) and displayed using MEGA.

Figure S4. Phylogenetic tree of the kainate receptors in Drosophila melanogaster and Acyrthosiphon pisum. Tree showing relationships of A. pisum, Apis mellifera and D. melanogaster kainate receptors protein sequences. Numbers at each node signify bootstrap values with 1000 replicates and the scale bar represents substitutions per site. The proteins shown in the tree (as well as GenBank accession numbers) are as follows: DmeCG5621 (NP_650927), Dmel KaiRIA (NP_651982), Dmel GluRIIE (NP_001036733), DmelCG3822 (NP_650925), Dmel CG9935 (NP_651931), Dmel CG11155 (NP_651941), Dmel Clumsy (NM_078886), Dmel GluRIIC (NP_608557), Dmel GluRIIA (NP_523484), Dmel GluRIIB (NP_523485), AmelCG3822-3 (XP_392185), AmelCG3822-1(XP_394265), AmelCG3822-2 (XP_394264). The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987) available with the CLUSTALX program (Thompson et al., 1997) and displayed using MEGA.

Figure S5. Phylogenetic tree of the other ionotropic receptors in Drosophila melanogaster and Acyrthosiphon pisum. Tree showing relationships of A. pisum, Apis mellifera and D. melanogaster or other forms of ionotropic glutamate receptors protein sequences. Numbers at each node signify bootstrap values with 1000 replicates and the scale bar represents substitutions per site. The proteins shown in the tree (as well as GenBank accession numbers) are as follows: DmelCG32704 (NP_727328), DmelCG15627 (NP_608863), DmelCG42315 (NP_650924), DmelCG7385 (NP_649176), DmelCG2657 (NP_001097043), DmelCG6185 (NP_648455), DmelCG11775 (NP_649833), DmelCG17152 (NP_648548), DmelCG8533 (NP_649148), DmelCG33492 (NP_995744), DmelCG15732 (NP_572795), DmeCG15324 (NP_572411), DmelCG14076 (NP_649074), DmelCG10101 (NP_649720), DmelCG10633 (NP_647962), DmelCG31718 (NP_723585), DmelCG14586 (NP_649013), DmelCG14585 (NP_649012), AmelCG32704 (XP_393270), AmelCG15627 (XP_396400), AmelCG17274 (XP_624089), AmelCG7385 (XP_624096), AmelCG6185 (XP_001123055), AmelCG7385 (XP_624096), AmelCG14076 (XP_001121834), AmelPlr2 (XP_395079). The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987) available with the CLUSTALX program (Thompson et al., 1997) and displayed using MEGA.

Figure S6. Alternative splicing of exons in Acyrthosiphon pisum nAChR subunits. Comparison of alternative exons of Drosophila melanogaster and A. pisum α4 (a) and α6 (b) subunits. Aphid residues that differ from those of the orthologous Drosophila exon are underlined. The Cys-loop is marked by asterisks and loops involved in ligand binding (LpB and LpE) are indicated. (c) Putative splice variants of Apisumα7 which generate diversity in loop C (LpC) potentially affecting ACh binding and in transmembrane regions 1 and 2 (TM1 and TM2) potentially affecting ion channel properties.

Figure S7. Tree of sequence similarity showing the translated transient receptor potential sequences of Acyrthosiphon pisum and Drosophila melanogaster. A neighbour-joining tree was produced from a ClustalX alignment of the translated transient receptor potential sequences from A. pisum and D. melanogaster. D. melanogaster sequences were obtained from FlyBase (http://flybase.org), and are labelled according to their FlyBase gene symbol. Sequences from A. pisum were retrieved from the genome sequence reference assembly and putatively identified by a reciprocal TBLASTX search, using the D. melanogaster transient receptor potential sequences as the initial search queries. A. pisum sequences are labelled with the appropriate gene model reference number (with ACYPIT prefix) where available. Two additional A. pisum transient receptor potential sequences were found, and are labelled with the prefix Ap and an appropriate identity based on their similarity to D. melanogaster sequences. The partial translated sequences from the aphid genome were then aligned against the Transient Receptor Potential sequences from Drosophila using ClustalX (version 1.83). The ClustalX alignments were used to generate bootstrapped neighbour-joined trees, which were viewed using Treeview (version1.6.6) software.

Table S1. Sodium channels

Table S2. Calcium Channels

Table S3. Potassium channels

Table S4. Chloride channels

Table S5. Cyclic-gated nucleotide Channels

Table S6. Glutamate Receptors

Table S7. The Transient Receptor Potential channels of Acyrthosiphon pisum and Drosophila melanogaster.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
IMB_975_sm_Suppl_info.doc583KSupporting info item

Please note: Neither the Editors nor Wiley Blackwell are responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.