Nineteen cytosolic GST genes have been identified in N. vitripennis, which is more than twice the number in the honeybee but a considerable reduction in the size of this family compared with Drosophila melanogaster, Anopheles gambiae and T. castaneum (Table 1, Fig. 1). Three of the N. vitripennis GST genes are incomplete in the genome assembly. NvGSTS2 and NvGSTS6 contain sequence gaps and NvGSTD2 contains numerous frameshifts and is almost certainly a pseudogene. A single putative microsomal GST gene was identified in N. vitripennis compared with the two found in A. mellifera and three found in An. gambiae. The microsomal GSTs are structurally unrelated to the cytosolic GSTs, although both classes play a similar role in protection against oxidative stress and detoxification of xenobiotics (Hayes et al., 2005).
Table 1. Numbers of validated glutathione-S-transferases, cytochrome P450s and carboxylesterases annotated in the insect genomes and their distribution across GST classes, P450 clans and CCE classes
|Species||Drosophila melanogaster||Anopheles gambiae||Aedes aegypti||Apis mellifera||Nasonia vitripennis||Tribolium castaneum||Bombyx mori|
|Cytosolic GSTs|| || || || || || || |
| Cytosolic GST total||37||28||27||8||19||35||23|
|P450s|| || || || || || || |
| Mitochondrial CYPs||11||9||10||6||7||9||8|
| P450 total||85||106||164||46||92||131||86|
|CCEs|| || || || || || || |
|Dietary class*|| || || || || || || |
| A* clade||0||0||0||5||8||0|| |
| B* clade||13||16||22||3||5||14|| |
| C* clade||0||0||0||0||0||12|| |
|Hormone/semiochemical processing|| || || || || || || |
| D clade||3||0||0||1||4||2|| |
| E clade||2||4||2||2||11||7|| |
| F clade||3||6||6||2||2||2|| |
| G clade||0||4||6||0||0||0|| |
|Neuro/developmental|| || || || || || || |
| H clade||5||10||7||1||1||1|| |
| I clade||1||1||1||1||1||1|| |
| J clade||1||2||2||2||2||2|| |
| K clade||1||1||1||1||1||1|| |
| L clade||4||5||5||5||5||5|| |
| M clade||2||2||2||1||1||2|| |
| CCEs total||35||51||54||24||41||49|| |
Figure 1. Unrooted distance Neighbour-joining tree showing the phylogenetic relationships of the predicted glutathione-S-transferase (GST) proteins of Nasonia vitripennis (shown in magenta) in relation to GST proteins from Drosophila melanogaster (green), Anopheles gambiae (blue), Apis mellifera (red) and Tribolium castaneum (orange). Protein names show family (GST), class (D = Delta, E = Epsilon, O = Omega, S = Sigma, T = Theta, Z = Zeta, U = unassigned) and gene number. Alternative splice variants of the An. gambiae GSTS1 and GSTD1 genes are designated by a hyphenated Arabic number. Distance bootstrap values of >70% (500 replicates) are indicated at the corresponding nodes (*). The putative pseudogene N. vitripennis GSTD2 has been omitted as the large number of frameshifts in the genome sequence precluded accurate prediction of protein sequence.
Download figure to PowerPoint
All but one (GSTZ1) of the cytosolic GSTs in the jewel wasp genome was predicted by the automatic annotation (Werren et al., 2010) and locus identifiers are provided in Table S1. Two GSTs, GSTO1 and GSTT1, were re-annotated manually to improve the alignment with other members of these GST classes. There is expressed sequence tag (EST) support for 10 of the 19 cytosolic GSTs (Table S1), with transcripts identified in libraries from different developmental stages.
As in the honeybee, the Sigma class is the largest GST class in the jewel wasp. Eight of its 19 GSTs belong to this class. They are located on four scaffolds and two, GSTS6 and GSTS5, are products of a recent duplication event. The reason for the expansion of the Sigma class of GSTs in Hymenoptera is unclear. Structural and catalytic roles have been assigned to the Sigma GSTs (Clayton et al., 1998; Singh et al., 2001). In addition, some members of this class have high levels of activity against 4-hydroxynonenal, which is a by-product of lipid peroxidation, and hence this class of GSTs is thought to play an important role in protection from oxidative stress (Singh et al., 2001). Oxidative stress could be a particular challenge to N. vitripennis larvae as the progressive death of host tissues releases a range of reactive oxygen species (Schafer & Buettner, 2001). ESTs for five of the eight Sigma class GSTs have been recovered from N. vitripennis larvae, two of them at relatively high levels (Table S1).
The Delta and Epsilon class GSTs, which are the largest classes of GSTs in the non-hymenopteran insect orders (Table 1), are sparsely represented in N. vitripennis. A cluster of three Delta GSTs is found on scaffold 7 but the internal sequence, GSTD2, is a probable pseudogene. Two further Delta class GSTs are present in the N. vitripennis genome but, as in the A. mellifera genome, no Epsilon class GSTs were identified. The Delta and Epsilon class GSTs are unique to insects and contain the majority of the GSTs associated with detoxification of insecticides. Parasitoid wasps may be protected from exposure to many synthetic chemicals by the detoxification capabilities of the host and thus their genomes may not have undergone the extensive expansion of the insect-specific GST classes observed in many other free-living insects.
The remaining GST classes represented in the jewel wasp, the Omega, Theta and Zeta classes, are ubiquitously distributed in nature, suggesting that they play key roles in endogenous metabolic processes. Secure 1:1:1:1 orthologues of GSTZ1 are found in all insect species analysed to date. This enzyme has a defined role in the catalysis of tyrosine and phenylalanine degradation (Board et al., 1997). Two Omega GSTs are found in N. vitripennis, twice the number found in A. mellifera or An. gambiae but fewer than the four or five Omega GSTs found in T. castaneum or D. melanogaster, respectively. An expansion in the Theta class GST is also observed in N. vitripennis compared with A. mellifera. A paralogous cluster of three Theta GSTs on scaffold 8 of N. vitripennis contrasts with the single orthologue found in the honeybee.
As has been found in the other sequenced insect genomes, the genomic locations of GST genes show some clustering in N. vitripennis (Fig. 2). All three from the Theta class sit within scaffold 8 on chromosome 2 while two from the Sigma class and two from the Delta class are located within single scaffolds (16 and 17, respectively) on chromosomes 3 and 5, respectively. At a higher level of organization, there is also some non-random association of GSTs from different classes to particular chromosomes; for example both genes from the Omega class are on chromosome 1, six of the eight from the Sigma class are on chromosome 3 and four from the Delta class are on chromosome 5. Interestingly, only four of the GSTs (GSTD4, GSTD5, GSTS7 and GSTS8) are located in the regions of low recombination, high gene density and high retrotransposon abundance that may contain the centromeres of the N. vitripennis chromosomes (Werren et al., 2010). This contrasts strongly with the distribution of P450 and CCE clades associated with environmental responsiveness (see below).
Figure 2. Locations of glutathione-S-transferase, cytochrome P450 and carboxylesterase (CCE) genes on Nasonia vitripennis chromosomes (in part adapted from fig. 1 of Niehuis et al., 2010; and from fig. S24 of Werren et al., 2010). Putative centromeric regions are indicated as shaded regions within the chromosomes, coordinates (in cM) for the corresponding genetic maps are shown by the scale bars and the clade assignments of the CCEs are shown in brackets.
Download figure to PowerPoint
The N. vitripennis genome data currently include 92 full-length P450 genes, plus nine pseudogenes and seven fragments of genes. Its complement of active, full-length P450s is thus twice the size of the honeybee's and similar to those reported for An. gambiae, D. melanogaster and B. mori, although somewhat smaller than those reported for T. castaneum and Aedes aegypti (Table 1). These differences are not due to the CYP2 and mitochondrial P450 clans; the numbers of N. vitripennis members of these two clans sit within the range of 6–11 found for each of these clans in the other species. Instead the differences can be explained largely by the CYP4 and CYP3 clans. These latter have proven much more variable among the other genomes and the honeybee had the lowest numbers previously reported for both (4 and 28, respectively), but particularly for the CYP4s. The number of N. vitripennis CYP4s (30) is still lower than those in any of the non-hymenopteran genomes, although its complement of CYP3s (48) lies within the range seen in the non-hymenopteran genomes (36–84).
As might be expected from the tight conservation of gene numbers in this clan, the mitochondrial P450s of N. vitripennis show high levels of 1:1 orthology with mitochondrial P450s from the other insect genomes. Specifically, five of its seven genes in this clan show 1:1:1:1:1 orthologies with P450s from the four other insect genomes included in our phylogenetic analyses (A. mellifera, D. melanogaster, An. gambiae and T. castaneum; Fig. 3). Three of these involve orthologues of the terminal hydroxylases (shade, disembodied and shadow) involved in the biosynthesis of 20-hydroxyecdysone which have been thoroughly characterized in D. melanogaster (Rewitz et al., 2007). Another one of the N. vitripennis mitochondrial clan P450s also fits this overall pattern of precise orthologies, except that T. castaneum has two members of this clade. However, the seventh N. vitripennis mitochondrial clan P450 sits in a clade (denoted CYP12) with much weaker orthologies, including multiple representatives within each of the Diptera and no A. mellifera representative at all. Notably, members of this CYP12 clade have been implicated in insecticide resistance in both the housefly M. domestica (Guzov et al., 1998) and D. melanogaster (Brandt et al., 2002; Bogwitz et al., 2005).
Figure 3. Neighbour-joining trees of the four P450 clans: (A) CYP2; (B) mitochondrial; (C) CYP3; and (D) CYP4. All putatively functional P450 genes from the genomes of Nasonia vitripennis (Nv; magenta), Apis mellifera (Am; red), Tribolium castaneum (Tc; orange), Drosophila melanogaster (Dm; green) and Anopheles gambiae (Ag; blue) are included, with non-hymenopteran clades collapsed where necessary (black). Two-letter species designations replace ‘CYP’ in sequence names. Greater than 50% support in 1000 bootstrap replications is indicated at the corresponding nodes (*). Putative P450 orthologues are indicated with the CYP family or subfamily name, D. melanogaster gene name and, if known, function in the ecdysteroid synthesis pathway. Trees were rooted using a clade of apparent orthologues (CYP4AA for CYP4 clan) or a P450 sequence from a non-insect: Homo sapiens CYP2J2 (CYP2 and mitochondrial clans) and CYP3A4 (CYP3 clan).
Download figure to PowerPoint
The seven N. vitripennis members of the CYP2 clan also show strong orthologies with P450s from the other four genomes analysed here, although in this case there is only one set of precise 1:1:1:1:1 orthologies. This is the CYP303 clade, the D. melanogaster representative of which is encoded by the nompH gene and affects mechano-sensory structures in sensory bristles (Willingham & Keil, 2004). There are also three other clades within this clan for which functions are known, the phantom and spook/spookier P450s which are again involved in 20-hydroxyecdysone biosynthesis (Rewitz et al., 2007) and another key controller of metamorphosis, juvenile hormone epoxidase (Helvig et al., 2004). Both N. vitripennis and A. mellifera have single representatives of each of these clades.
The 48 P450s found in the N. vitripennis genome that sit in the CYP3 clan include 17 in the CYP9 family, with the remainder sitting in a set of related clades broadly classified as the CYP6 family. The larger numbers of CYP3s in the jewel wasp as compared with honeybee are reflected in the numbers of both the CYP6s and CYP9s (31 and 17 cf 23 and 5, respectively). Characteristically for the clan, the N. vitripennis CYP3s show no precise orthologies with P450s from the other sequenced genomes, even A. mellifera, but instead fall into a few, quite large and presumptively relatively recent radiations. Thus there are two radiations of CYP6s (CYP6CK and CYP6AS) and one of CYP9s (CYP9AG/H), containing 10, 9 and 14 genes, respectively, which are currently specific to Nasonia. Amino acid identities within these clades lie in the ranges of 35–87, 43–70 and 31–93%, respectively. The corresponding figures for synonymous site nucleotide differences (dS) are 0.12–6.12, 0.5–3.0 and 0.21–6.93. Moreover, over half of the pairwise comparisons of genes within each of these radiations shows saturation for silent site differences (analyses not shown) which, under a wide range of estimates for the rate of the molecular clock in Hymenoptera (see e.g. Bromham & Leys, 2005), implies that they diverged several tens of millions of years ago.
Several members of dipteran and lepidopteran CYP6 radiations are involved in resistance to a broad range of insecticides (OPs, synthetic pyrethroids, DDT and neonicotinoids; Carino et al., 1994; Liu & Scott, 1996; Daborn et al., 2002; Li et al., 2007; Muller et al., 2008) and/or the detoxification of host plant allelochemicals in the gut (Danielson et al., 1997; Mittapalli et al., 2005; Wen et al., 2005; Mao et al., 2006, 2007; Li et al., 2007; Rupasinghe et al., 2007). Less is known of the functions of the CYP9s but the limited evidence available also implicates them in the detoxification of insecticides and other allelochemicals (Stevens et al., 2000; Poupardin et al., 2008), with some evidence also for involvement in semiochemical metabolism (Maibeche-Coisne et al., 2005). The Nasonia CYP6s and to a lesser extent CYP9s are thus good candidates for involvement in its detoxification of insecticides and host metabolites. ESTs have been recovered for 17 of the jewel wasp's CYP6s and six of its CYP9s (Table S2). In aggregate, the CYP6 ESTs are roughly evenly split between larval and pupal/adult samples, whereas most of the CYP9 ESTs derive from pupal/adult material.
The incorporation of the N. vitripennis data into the P450 phylogenies also throws new light on the evolution of some of the A. mellifera P450s. Specifically, in the case of the CYP6s, Fig. 3 shows that the 16 honeybee CYP6ASs (which is one third of its total complement of P450s) all derived from duplications that occurred after the bee and wasp lineages diverged (∼180 Mya, seeIntroduction).
As noted above, the CYP4 clan showed the greatest difference in P450 numbers between N. vitripennis and A. mellifera (30 vs. 4, respectively). Three of the N. vitripennis CYP4s in fact sit in clades of CYP4AA and CYP4G genes that (except for one missing A. mellifera gene) show 1:1:1:1:1 orthologies across the five genomes compared. Members of these clades from other species have been implicated in functions as diverse as 20-hydroxyecdysone biosynthesis (Maibeche-Coisne et al., 2000), pheromone metabolism (Maibeche-Coisne et al., 2005) and pyrethroid insecticide resistance (Pridgeon et al., 2003). Sex-selective expression has also been demonstrated for one of the A. mellifera orthologues (Evans & Wheeler, 2001). The other 27 N. vitripennis CYP4s are distributed across clades showing variable levels of taxon-specific radiations, the major N. vitripennis radiation being its 19 members of the CYP4J+ clade. This clade has been linked to lipid metabolism in some other species (Simpson, 1997; Feyereisen, 2005) and, notably, in another hymenopteran, the fire ant Solenopsis invicta, to caste determination (Liu & Zhang, 2004). Members of the sister group of the CYP4J+s, containing the CYP4AVs of N. vitripennis and A. mellifera (specifically the CYP4L4 and CYP4S4 of Mamestra brassicae), have been implicated in pheromone metabolism (Maibeche-Coisne et al., 2002, 2005) while suggested functions for other closely related CYP4s, including the CYP4Cs, CYP4Ds, CYP4Hs, CYP4Hs and CYP4Js, range from juvenile hormone and pheromone metabolism to DDT, pyrethroid and carbamate insecticide resistance (Sutherland et al., 1998; Scharf et al., 2001; Maibeche-Coisne et al., 2002, 2004; Shen et al., 2003; David et al., 2005). ESTs have been found for 13 of the N. vitripennis CYP4s, generally more frequently from pupal/adult than larval material (Table S2).
While the GST genes of N. vitripennis were well scattered across the chromosome, with less than a third in the putative centromeric regions (see above), as many as half of the 92 P450s that have been mapped to contigs lie in or very closely adjacent to such regions (Fig. 2). About half of them are also located in physical clusters of between two and eight related P450 sequences (Fig. S1). With very few exceptions, P450s in the same physical cluster are also from the same phylogenetic radiation. However, there is no obvious trend for the P450s in the physical clusters to be in the putative centromeric regions or elsewhere on the chromosomes.
We have found 41 CCE sequences in the N. vitripennis genome (Table S3), almost twice as many as the 24 found in the genome of A. mellifera and similar to the numbers found in the other (dipteran and coleopteran) insect genomes so far sequenced and characterized for CCEs (Table 1). The CCEs of the coleopteran, T. castaneum, which had not been thoroughly annotated previously, are catalogued in Table S4, with some summary comments also added below.
As in the other genomes, the N. vitripennis CCEs fall into three main phylogenetic classes. Functional work on other species suggests that these classes broadly represent dietary/detoxification, hormone/semiochemical processing, and neuro/developmental functions (Fig. 4). Numbers of N. vitripennis CCEs in the three classes are 13, 17 and 11, respectively, which in most respects conforms with the distribution found for the dipteran and coleopteran CCEs. The one respect in which N. vitripennis clearly differs from these precedents is its relatively high numbers of CCEs in the hormone/semiochemical processing class: 17 as compared with 8–14 in the other genomes. This is the same class that is most under-represented in the A. mellifera genome, where there are only five.
Figure 4. Unrooted distance Neighbour-joining tree showing a phylogeny of the carboxylesterases (CCEs) from the genomes of Nasonia vitripennis, Drosophila melanogaster, Anopheles gambiae, Apis mellifera and Tribolium castaneum, together with other previously characterized CCE sequences. CCEs are colour coded as in Figs 1 and 3 except that black is used for CCEs from non-sequenced genomes (see Oakeshott et al., 2005 and Claudianos et al., 2006 for details) and bold black is used for clades which have been collapsed. The optimal tree (shown as a cladogram) was constructed using pairwise deletion of gaps/missing data and a PAM001 matrix substitution model. Percentage bootstrap confidence values greater than 50% (1000 replicates) are shown at nodes by an asterisk (*). Also shown preceding the taxon name are the predictions for the subcellular localization (S = secreted) and catalytic status (+/−) of the CCEs as deduced from the presence/absence of relevant sequence motifs (as per Oakeshott et al., 2005 and Claudianos et al., 2006). The functional assignment of clades follows the system proposed by Oakeshott et al. (2005) and Claudianos et al. (2006) except that the previous functional groups B and C have been merged into a new group denoted group B* to reflect the closer relationships now evident with the α-esterase radiation. The inclusion of the T. castaneum data also reveals a new clade denoted C* which is largely comprised of coleopteran CCEs of unknown function.
Download figure to PowerPoint
The previous work on the genomics of the insect CCEs has identified two to six distinct clades within each of the three major classes, with each species having representatives of all the clades of the neuro/developmental class and most of the clades within the other two classes. The same pattern occurs for N. vitripennis. In the case of the neuro/developmental CCEs, the distribution is conserved to the extent that exactly the same numbers of CCEs are found in each of six clades of neuro/developental CCEs as were found in the A. mellifera genome. This reflects the relatively ancient origins of all the major lineages within this class; indeed all 11 of the N. vitripennis members of the class have 1:1 orthologues in A. mellifera and most also have their own orthologues in the dipteran and coleopteran genomes. As in the other genomes, essentially all of the CCEs in the neuro/developmental class except for the acetylcholinesterases (AChEs) are catalytically incompetent and have additional N- or C terminal domains. Work on D. melanogaster and A. mellifera members of some of these clades show that they act as specific cell adhesion molecules during the development of the nervous system (Oakeshott et al., 2005; Claudianos et al., 2006; Biswas et al., 2008).
The so-called dietary/detoxification class also shows reasonably good conservation of numbers within its two clades in comparison with A. mellifera. However, only one of 13 Nasonia CCEs in this class (Nv1603542.1) has a 1:1 othologue in the honeybee. And the largest clade in this class in N. vitripennis (clade A, with eight N. vitripennis CCEs; Fig. 5) may be specific to the Hymenoptera as it has no representatives among the dipteran and coleopteran sequences. ESTs for six of the clade A CCEs have been recovered from pupal and/or adult N. vitripennis but only for two of them have ESTs been reported from larvae (Table S3), suggesting that their presumptive dietary/detoxification function is not primarily associated with larval feeding.
Figure 5. Expanded view of the hymenopteran β-esterase clade from Fig. 4, showing 11 JHE-like enzymes in Nasonia vitripennis (magenta) compared with the two in Apis mellifera (JHE (GB15327) and GB10820). Also shown is a putative JHE from the sawfly (BAD91552.1). Most sequences in this radiation have an aliphatic residue (L/I/Y/F/V) at the second position (red) of the catalytic motif (brackets), which is generally unusual in carboxylesterases but often associated with JHE function (see text).
Download figure to PowerPoint
The other clade in this dietary/detoxification class represented in the N. vitripennis genome (clade B*) has members from all of the insect orders for which genome sequences are available. This includes the α-esterases implicated in insecticide resistance in the Diptera (Oakeshott et al., 2005; Claudianos et al., 2006). There are five N. vitripennis CCEs in clade B*, four of them in a single, recent radiation. Interestingly, all four of these have a distinctive GQSAG signature in a key active site motif called the nuceophilic elbow; the GQSAG has been associated, albeit not exclusively, with juvenile hormone esterases (JHEs) (Crone et al., 2007a,b and see below). There is no other evidence to suggest that any of these function as JHE in Nasonia and there are other candidates for JHE function in the hormone/semiochemical processing clades below. Only one of the four CCEs in the clade B* radiation appears in the EST database (Nv1602306.1, in larvae; see Table S3) but proteomic analysis has identified another one of them (Nv1602279.1) in N. vitripennis venom (de Graf et al., 2010). Its function in venom is unclear, but it is interesting to speculate as to whether it plays some role in the modulation of host hormonal processes. This could involve juvenile hormone metabolism, which is known to be disrupted by parasitoid venoms (Beckage & Gelman, 2004) .
On the face of it, the tight conservation of numbers in the so-called dietary/detoxification class of CCEs across the sequenced genomes might suggest that this class has little relevance to differential needs for detoxification functions across species. It is important to note, however, that there are also a few members of the so-called hormone/semiochemical processing class in other species that have well-described detoxification functions. Examples include the E4 and FE4 esterases that confer OP insecticide resistance on the aphid Myzus persicae and another esterase that confers a similar phenotype on the planthopper Nilaparvata lugens (Vontas et al., 2000; Oakeshott et al., 2005; Li et al., 2007). One complicating factor in interpreting CCE gene diversity is a striking paucity of studies ascribing esterase activity against naturally occurring dietary constituents (e.g. Lindroth, 1989) to specific esterase genes.
The large difference in the numbers of the hormone/semiochemical processing class of CCEs across the species that was noted above is for the most part explained by one specific clade (E), which has 11 representatives in N. vitripennis but only two in A. mellifera, and 2–7 in the other species. Figure 4 shows that there has been a series of duplication events within this clade since the divergence from the A. mellifera lineage (estimated to be about 180 Mya; Werren et al., 2010). As with the CYP6 and CYP9 radiations above, however, most of the pairwise comparisons among the clade E CCEs of N. vitripennis are essentially saturated for synonymous site differences (analyses not shown), suggesting that they diverged from one another relatively early in the evolution of the wasps. Even outside this clade, the level of orthology of the N. vitripennis CCEs in the hormone/semiochemical processing class is relatively low; only two of the six N. vitripennis CCEs in the respective clades (D and F) have direct orthologues in A. mellifera and just one of them does so among the dipteran and coleopteran sequences.
The data available for members of clade E from other species indicate a diversity of functions. The clade includes the E4 and FE4 esterases conferring OP resistance on M. persicae (E4 and FE4 are very closely related and only FE4 is shown in Fig. 4) (Li et al., 2007), the antennal Apo1PDE esterase of the silkmoth Antheraea polyphemus that degrades that species' sex pheromone (Ishida & Leal, 2005) and the β-esterases implicated in the hormonal control of reproductive behaviour in Drosophila (Robin et al., 2009). The one hymenopteran member of clade E for which there are empirical functional data is GB15327 from the honeybee (Fig. 5). Recent expression data and RNAi studies indicate that GB15327 affects juvenile hormone metabolism, and it has therefore been proposed to be a JHE (Mackert et al., 2008). Interestingly this enzyme has a GLSAG motif rather than the distinctive GQSAG nucleophilic elbow motif found in all of the functionally validated JHEs from other species (Crone et al., 2007a,b). N. vitripennis has three orthologues of this putative JHE, one with GLSAG and two with GISAG, and ESTs have been found for all three of them (Table S3). There is another esterase in this clade which does have a GQSAG, from another hymenopteran, the sawfly Athalia rosae (Fig. 5), but there are no functional data to determine its status as a JHE. Evidence from other species indicates that the JHE function has arisen in several different clades (Crone et al., 2007a,b) so the identity of the N. vitripennis JHE(s) remains uncertain at this stage.
Outside clade E, the other N. vitripennis members of the putative hormone/semiochemical processing class of CCEs are two 1:1 orthologues of A. mellifera esterases in the clade (F) containing the known dipteran JHEs and a set of 4:1 orthologues of an A. mellifera esterase in a clade (D) containing an Antheracea CCE (Apo1IE) which is located in the integument and implicated in the processing of olfactory signals (Vogt & Riddiford, 1981). There are no empirical data on the functions of the A. mellifera CCEs in these clades. However, ESTs have been recovered from adults for three of the N. vitripennis esterases in clade D (Nv1601317, Nv1601350, Nv1601375; Table S3) and these would therefore seem to be the best esterase candidates for pheromone or kairomone processing in the jewel wasp.
The three classes of CCE genes differ qualitatively in their locations on the N. vitripennis chromosomes (Fig. 2). A large proportion (77%) of its CCE genes in the dietary/detoxification and hormone/semiochemical processing classes are located in or very near the putative centromeric regions, but none of the neuro/developmental CCE genes are located in these regions. Thus the more recently diverged CCE genes are more likely to occur in the putative centromeric regions and the older genes are more likely to occur elsewhere. Unlike the dipteran genomes (Oakeshott et al., 2005), the jewel wasp genome contains no large clusters of CCE genes, although it does have seven smaller clusters, each containing 2–4 genes from the same clade (Fig. S2). Four of these clusters involve the radiation of JHE-like CCEs in clade E. Six of the seven clusters are also located in putative centromeric regions.
Finally we note that Tables 1 and S4 and Fig. 4 provide the first thorough annotation and the first phylogenetic treatment of the CCEs of T. castaneum. There are 49 CCEs in T. castaneum, which is very close to the highest number previously recorded for sequenced insect genomes, from the mosquito Ae. aegypti (54). This number is consistent with the relatively toxin-rich ecology of T. castaneum and its relatively high numbers of GSTs and P450s (Table 1 and Tribolium Genome Sequencing Consortium, 2008). Its complement of neuro/developmental CCEs is very similar to those of the other genomes and again deep orthologies are evident across the insect orders. The only variation is that T. castaneum, like the Diptera but unlike the hymenopterans, has two neurotactins rather than one. There are 11 T. castaneum CCEs in the hormone/semiochemical processing group of clades identified by Claudianos et al. (2006). Specifically, there are two in the clade (D) containing CCEs associated with pheromone degradation in the integument, seven in the clade (E) of secreted β-esterases which may also contain the hymenopteran JHE (see above), and two in the clade (F) containing dipteran JHEs, with none in the clade (G) containing the lepidopteran JHEs. One of the clade F CCEs, Tc13193, has a GQSAG and is therefore the best candidate for the T. castaneum JHE. The complement of CCEs in the jewel wasp from the dietary/detoxification clades of Claudianos et al. (2006) is entirely contained in a single radiation of 14 sequences that form a sister group to the α-esterases of the higher Diptera (clade B* in Fig. 4), which have been consistently implicated in OP insecticide resistance (Oakeshott et al., 2005; Hartley et al., 2006). However, there is also an additional major radiation containing 12 T. castaneum CCEs and a single lepidopteran CCE, none of which have known functions. This clade, which we have tentatively called C*, also sits most comfortably in the dietary/detoxification class of Claudianos et al. (2006). All of the CCEs in this clade have secretion signals and all the key residues required for catalytic competence, and none of them have additional trans-membrane or other domains characteristic of many neuro/developmental CCEs. Insights into their functions await empirical biochemical and physiological studies.