MBE Advance Access originally published online on October 12, 2005
Molecular Biology and Evolution 2006 23(2):317-326; doi:10.1093/molbev/msj037
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Variation in Positive Selection in Termite GNBPs and Relish
School of Tropical Biology, James Cook University, Douglas, Australia
E-mail: mark.bulmer{at}jcu.edu.au.
 |
Abstract |
|---|
 TOP Abstract IntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Social insects are model organisms for investigating molecular evolution in the innate immune system. Their diversity affords comparative analysis among closely related species, and group living is likely to contribute to the pathogen stress imposed on the immune system. We used different models of nucleotide substitution at nonsynonymous (amino acid altering) and synonymous (silent) sites to compare the different levels and type of selection among three immunity genes in 13 Australian termite species (Nasutitermes). The immunity genes include two encoding pathogen recognition proteins (gram-negative bacterial-binding proteins) that duplicated and diverged before or soon after the evolution of the termites and a transcription factor (Relish), which induces the production of antimicrobial peptides. A comparison of evolutionary models that assign four unrestricted classes of dN/dS (the ratio of the nonsynonymous to synonymous substitution rate) to different Nasutitermes lineages revealed that the occurrence of positive selection (dN/dS > 1) varies among lineages and the three genes. Positive selection appears to have driven the evolution of all three genes in an ancestral lineage of three subterranean termites. It had previously been suggested that there was a transition along this ancestral lineage to termite morphology and ecology associated with a diet of decayed wood, a diet that may expose termites to elevated levels of fungal and bacterial pathogens. Relish appears to have experienced the highest levels of selective pressure for change among all three genes. Positively selected sites in the molecule are located in regions that are important for its activation, which suggests that amino acid substitutions at these sites are a counter response to pathogen mechanisms that disrupt the activation of Relish.
Key Words: Relish • innate immunity • GNBP • termites • positive selection • evolutionary immunology
|
Introduction |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Microbial pathogens can avoid an inducible immune response by evading detection, disrupting the signaling apparatus of the immune response, or blocking the effector molecules that can harm them. If a pathogen can simply evade detection, then it does not need to evolve complex mechanisms that disrupt or block the immune system. Pattern recognition proteins (PRPs) such as gram-negative bacterial-binding proteins, GNBPs (Pili-Floury et al. 2004
), and peptidoglycan recognition proteins, PGRPs (Leulier et al. 2003), of the innate immune system might therefore be expected to be a focal point of conflict between host and pathogen that drives their molecular evolution. However, there is evidence of purifying selection in PRPs from studies that have focused on GNBPs and PGRPs in Daphnia and Drosophila, respectively (Jiggins and Hurst 2003; Little, Colbourne, and Crease 2004), which indicates that PRPs are constrained not to change. This may be attributable to the conserved nature of recognition epitopes such as lipopolysaccharides and peptidoglycans, which do not have much evolutionary room for maneuver without disruption of their biochemical properties (Girardin, Sansonetti, and Philpott 2002; Janeway and Medzhitov 2002; Loker et al. 2004). This argument, nevertheless, neglects the possibility that pathogens can interfere with the ability of the PRPs to transmit a signal. There is evidence for adaptive evolution in the signaling protein Relish (Begun and Whitley 2000) as well as several antimicrobial peptides (Hughes and Yeager 1997; Boniotto et al. 2003; Lazzaro and Clark 2003; Morrison et al. 2003; Semple, Rolfe, and Dorin 2003; Bulmer and Crozier 2004; Lynn et al. 2004).
Begun and Whitley (2000) found evidence of adaptive evolution in Relish from Drosophila simulans, and these authors speculated that microbial pathogens can interfere with its activation. Activation of Relish, a nuclear factor (NF) 
B–like transcription factor, results in the production of antibacterial peptides (Dushay, Åsling, and Hultmark 1996). Relish is composed of REL homology domain, a nuclear localization sequence (NLS), ankyrin repeats, and a PEST domain. Phosphorylation in or around the PEST domain by the I-B kinase complex (IKK) results in the cleavage of Relish by the caspase Dredd (Salmeron et al. 2001; Stoven et al. 2003). This cleavage is believed to prevent the ankyrin repeats (I-B–like domain) from inhibiting the activation of Relish and exposes the nuclear localization signal, allowing the REL homology domain to target genes. Bacteria such as Yersinia have systems that can interfere with the signaling response, and studies are now revealing specific components of NF-B activation that might be targeted (Orth et al. 1999; Neish et al. 2000; Lindmark et al. 2001). Bacteria may also target molecules much further upstream of Relish in the signaling pathway.
Social insects are model organisms for investigating adaptive evolution in the innate immune system (Schmid-Hempel 1998). Group living in social insects increases their vulnerability to disease. In addition, the male haploid genetic system in the Hymenoptera and inbreeding in the Isoptera can yield high relatedness among colony members, which potentially facilitates the spread of parasites and disease (Schmid-Hempel and Crozier 1999). In this study, we focused on a group of Australian termites in the genus Nasutitermes. These termites are closely related but vary in their foraging ecology, nest type, and location, and we have found evidence of diversifying selection in a group of their defensin-like antimicrobial peptides (Bulmer and Crozier 2004). Previous studies of molecular evolution in GNBP and Relish used population genetic approaches to detecting positive selection with two or three species. In this study, we use a comparative analysis with codon-based substitution models and 13 species to investigate molecular evolution in GNBP and Relish.
|
Methods |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Termites
Collection and identification of the different Australian nasutes and the outgroup Drepanotermes rubriceps has previously been described (Bulmer and Crozier 2004
). Mastotermes darwiniensis workers were collected from a foraging group in a log in Townsville, Queensland (19°20.063'S, 146°45.343'E).
Isolation and Characterization of Relish
We designed degenerate primers with the program CODEHOP (Rose et al. 1998) using protein alignments generated with the program BlockMaker (Henikoff et al. 1995) from Drosophila melanogaster (GenBank accession number NP_477094), Drosophila yakuba (AF204290), and Aedes aegypti (AAM97895) Relish sequences. The degenerate primers were used for one-step reverse transcriptase polymerase chain reaction (PCR) (Invitrogen, San Diego, Calif.) of mRNA. The mRNA was prepared from 20 untreated large Nasutitermes graveolus workers, ground in liquid nitrogen, homogenized by centrifugation through a Qiagen shredder for 2 min at 16,000xg, and purified with a Quick Prep mRNA kit (Amersham Biosciences, Castle Hill, Australia). A primer pair (forward: GCC TCA ACT GAG AAT CGT GGA RCA RCC NGT and reverse: GAA GAT GTA CTT TTT AGC AGT GTG GAT DAT NCC CAT) produced a band of the expected size (
300 bp). PCR cycle conditions were 55°C, 20 min; 94°C, 2 min; 40 times (94°C, 15 s; 55°C, 30 s; 72°C, 2 min); 72°C, 7 min. The 300-bp band was gel purified and amplified a second time (94°C, 2 min; 35 times [94°C, 2 min; 55°C, 30 s; 72°C, 30 s]; 72°C, 7 min) and gel purified and cloned with the PGEM-T Easy Vector system (Promega, Annandale, Australia). A Blast search revealed that the sequence of PCR-amplified clones had significant similarity to D. yakuba Relish (AF204290). The full sequence of the mRNA transcript for N. graveolus Relish was obtained by 5' and 3' rapid amplification of cDNA ends (RACE), performed as described previously (Bulmer and Crozier 2004). The majority of the transcript (18 nt downstream from the start codon to 12 nt upstream of the stop codon) was amplified by one-step reverse transcriptase PCR from different nasutes with five pairs of specific primers (table 1) that amplified overlapping regions of the gene. Sequences were assembled with the program Sequencher, version 4.2.
|
Isolation and Characterization of GNBP
For isolating GNBP from termites, degenerate primers were designed from Manduca sexta (AF177982), Bombyx mori (BAA92243), D. melanogaster (AAF50349), and Anopholes gambiae (EAA00167) using CODEHOP and BlockMaker. GNBP was initially isolated and characterized in M. darwiniensis with mRNA purified from five workers with the Quick Prep mRNA kit. A primer pair (forward: GGA TAC AAG GTA CCA GCA TGG CNC CNT TYG A and reverse: GGG GGT ACC ATT GGT TTC GTK CNT YCC ARA A) produced a band of the expected size (
180 bp). PCR cycle conditions, band purification, and cloning followed the same procedure described for Relish. The full sequence of the mRNA transcript for M. darwiniensis GNBP was obtained by 5'and 3' RACE (see above). This sequence and the sequence from A. gambiae (EAA07705 and EAA09916) were used to design degenerate primers that had greater specificity for termites. A primer pair (forward: GAATGC CAA CTG GTG ACT GYT NTG GCC and reverse: TTG GTC TCG ACC GTT CCA RAA RTC) produced a band of the expected size (600 bp) in both M. darwiniensis and N. graveolus. A screen of clones from the N. graveolus reverse transcriptase PCR product revealed two GNBPs (GNBP1 and GNBP2) that shared only 71% of their amino acids. A subsequent screen of M. darwiniensis clones also revealed two GNBPs that shared 75% of their amino acids. The full sequence of the mRNA transcript was obtained for these nasute GNBPs and M. darwiniensis GNBP2 with 5' and 3' RACE. The full transcript was then amplified by one-step reverse transcriptase PCR from different nasutes with two pairs of specific primers for each GNBP (table 1) that amplified overlapping regions of the gene. Sequences were assembled with the program Sequencher, version 4.2.
Statistical Analysis
Relish and GNBP cDNA sequences were aligned by inspecting the translated sequence with Se-Al, version 2.0a11 (Rambaut 2002). The phylogenetic trees were constructed with Bayesian analysis using the program MrBayes, version 3 (Huelsenbeck and Ronquist 2001). The optimum substitution models were determined with the programs MrModeltest, version 2.2 (Nylander 2004), and PAUP*, version 4.0b10 (Swofford 1998). All sequences, including partitioned sequences (see below), passed a test for stationarity with the program Tree-Puzzle (Schmidt et al. 2002), which uses 
2 tests that compare the nucleotide composition of each sequence to the frequency distribution assumed in the maximum likelihood model. A phylogenetic tree for the PAML and HYPHY analysis (see below) was generated with MrBayes using partitioned data, which included the first, second, and third codon positions of GNBP1, GNBP2, and Relish (fig. 2). The data were partitioned so that a separate nucleotide substitution model could be chosen for each codon position in each gene. The Akaike index results from MrModeltest provided substitution models for each of the nine partitions. The MrBayes analysis was run with four chains, and a burn-in of 104 generations followed by a search of 106 generations for the best tree. The tree was rooted with the cDNA sequence from D. rubriceps. The analysis was repeated four times to ensure that the best trees had not been found at local optima. The amino acid sequence of N. graveolus Relish (DQ058898) was aligned with D. melanogaster Relish (NP_996189) with ClustalX, version 1.81 (Higgins, Thompson, and Gibson 1996). Relish domains were identified with a search of the Conserved Domain Database of the National Center for Biotechnology Information (Marchler-Bauer and Bryant 2004).
|
The ratio of the nonsynonymous (dN) to synonymous nucleotide substitution rate (dS) provides a signature for detecting selection that results in amino acid change (positive selection) or maintenance (negative or purifying selection). We identified positively (dN/dS > 1) and negatively selected sites (dN/dS < 1) with the program package PAML (Yang 1997
; Yang et al. 2000; Yang, Wong, and Nielsen 2005) using the M8 model (Bayes Empirical Bayes analysis) and the program package HYPHY (Kosakovsky Pond, Frost, and Muse 2005) using the fixed effects (two-rate FEL) and random effects (REL) models of molecular evolution. In order to make comparisons between Relish and GNBPs, the M. darwiniensis sequence was not included in the PAML and HYPHY analysis. Mastotermes darwiniensis is very distantly related (Thompson et al. 2000) to the termitids (nasutes and Drepanotermes), and its inclusion could erroneously increase estimates of the synonymous substitution rate among the GNBPs. For the HYPHY analysis, we employed codon substitution models that combined the HKY85 (Hasegawa, Kishino, and Yano 1985) nucleotide substitution model with the Muse and Gaut model and that considered approximate branch lengths (Kosakovsky Pond and Frost 2005b; Kosakovsky Pond, Frost, and Muse 2005). We used a nominal alpha level of 0.1 for FEL and 0.05 for REL and a posterior probability cutoff for the Bayesian method of 0.95 for M8. The fixed and random effects models lack the power to detect positively selected sites for small data sets (<18), so these are conservative nominal alpha values (Kosakovsky Pond and Frost 2005b). To further reduce the chance of falsely accepting positively selected sites, we considered sites to be under positive selection if there was consensus between all three models.
We also used PAML and HYPHY models that allow dN/dS to vary among lineages to investigate whether selective pressure on all three genes varies among lineages. The PAML lineage-specific models require a priori hypotheses when testing whether dN/dS is greater than one in a specific lineage or a subset of lineages (Yang 1998). The genetic algorithm in HYPHY assigns four classes of dN/dS to lineages in a search for "the best model" of lineage-specific evolution (Kosakovsky Pond and Frost 2005a). This latter approach can identify lineages under positive selection without an a priori hypothesis for lineage-specific evolution.
We used the methods described by Huelsenbeck, Nielsen, and Bollback (2003), as implemented in the program SIMMAP (Bollback 2004), to infer the distribution of life history traits on the phylogenetic tree, using the set of postburn-in trees derived from our MrBayes runs using all sequences combined. We studied two characters with three character states each, namely, nest site (subterranean, epigeal, arboreal) and food (rotten wood, grass, sound wood). We investigated the effects of the character states being unordered or ordered (in the orders given) and of the trees being unrooted or rooted with D. rubriceps.
|
Results |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The gene tree of the GNBPs in Nasutitermes (see below for species ID), D. rubriceps (Dr), and M. darwiniensis (Md) indicates that GNBPs duplicated once (fig. 1). This duplication event occurred prior to the divergence of Mastotermes, the most basal lineage of the Isoptera (Krishna 1970
; Miura et al. 1998; Thompson et al. 2000). The phylogenetic tree generated with MrBayes and the partitioned data set has high credibility support (fig. 2) and appears to group termites according to their ecological attributes with the possible exception of Nasutitermes triodiae. Nasutitermes pluvialis (Np), Nasutitermes dixoni (Nd), and Nasutitermes fumigatus (Nf) are subterranean termites that forage in rotten wood. Nasutitermes exitiosus (Ne), N. graveolus (Ng), and Nasutitermes walkeri (Nw) build distinct nests (N. exitiosus constructs mounds and N. graveolus and N. walkeri construct arboreal carton nests) and feed on relatively sound wood. Nasutitermes comatus (Nc), N. triodiae (Nt), Tumilitermes pastinator (Tp), Nasutitermes longipennis (Nl), and Nasutitermes magnus (Nm) are mound builders that eat grass, which they store in their nest (Miller 1997). This tree also shows close congruence with the phylogeny of Miller (1997), which used morphological characters. The MrBayes phylogeny was used in the PAML and HYPHY analysis. The large majority of informative amino acid sites were under neutral or significant negative selection, which indicates that most nucleotide change is occurring at synonymous sites. For example, the REL models identified 62, 39, and 79 negatively selected sites for GNBP1, GNBP2, and Relish, respectively (alpha = 0.05). The synonymous substitutions at these sites are likely to mask any phylogenetic bias attributable to directional selection of a small number of nonsynonymous sites.
|
We used PAML (Yang et al. 2000
) and HYPHY models (Kosakovsky Pond and Frost 2005b) that allow dN/dS to vary at different sites in the molecules (381 amino acid sites in the aligned GNBPs and 972 in Relish). The M8 model indicates that GNBP1 is under positive selection. Model M8, which uses a beta distribution between the dN/dS interval from zero to one plus an extra category allowing for positive selection (dN/dS > 1), is significantly better than model M8A (Swanson, Nielsen, and Yang 2003), in which the extra dN/dS category is restricted to one (table 2). The Bayesian approach in PAML identifies one site, 34, as positively selected. The REL model also identifies site 34 as positively selected;, however, the FEL model does not identify any sites as positively selected at an alpha value of 0.1 (table 3). For GNBP2, model M8 is significantly better than M8A (table 2). The Bayesian approach indicates that only sites, 232 and 369 are under significant positive selection, and both the HYPHY models only identify site 369 as positively selected (Table 3). For Relish, model M8 is significantly better than M8A (table 2). The Bayesian approach identified seven sites under significant positive selection. There is an agreement between the HYPHY models and M8 for five of these sites (table 3). In summary, the strength of selection in Relish appears to have been sufficient for identifying five specific sites as positively selected with statistical support and consensus among the models. The GNBPs appear to be influenced by positive selection, but the models fail to agree on the identity of positively selected sites with the exception of site 369 in GNBP2 at the nominal alpha values we have selected.
|
|
|
Insect Relish has previously been characterized in fruit flies and mosquitoes, and the inferred domain structure of termite Relish closely resembles that of Relish in Drosophila (Dushay, Åsling, and Hultmark 1996
) and Aedes (Shin et al. 2002). Figure 3 shows an inferred protein alignment of Relish from D. melanogaster (NP_996189) with Relish from N. graveolus (DQ058898). The protein can be broadly divided into an NF-B domain (The N terminal REL homology domain and the immunoglobulin fold and plexin-like transcription factor domain [IPT domain]) with a putative NLS at its C terminal end, a spacer region joining the NF-B domain to an ankyrin repeat domain, and a C terminal region, which includes a PEST-like domain (Dushay, Åsling, and Hultmark 1996). Nasutitermes graveolus Relish has six ankyrin repeats that align with the ankyrin repeats in Drosophila Relish. Termite Relish also has the N terminal serine-rich region (SRR), but in contrast to that of the flies, termite Relish has a C terminal SRR and lacks the SRR immediately following the NF-B domain. Four of five positively selected sites (494, 511, 522, 568) identified by all three site selection models are located within the spacer region toward its 5' end, and the fifth site (921) is located in the PEST-like domain. The alignment of termite Relish sequences (see Supplementary Material online) reveals that deletions and/or insertions have occurred in two regions in the molecule in more than one sequence. In other Nasutitermes, these regions correspond with insertions of two and eight to nine amino acids in the N. graveolus sequence shown in figure 3 between positions 494–495 and 571–572, respectively. Both deletion/insertion sites lie in the spacer region, adjacent to sites identified as positively selected (494, 568), and one of these deletion/insertion sites (fig. 3, site 571–572) lies adjacent to a putative caspase cleavage site (see below). In addition, there is a deletion of a single amino acid of N. triodiae Relish (882–884). Interestingly, the M3 model in PAML identified 882 as positively selected, and it appears to be located in a caspase cleavage site (see below). We did not use the M3 model extensively because it tends to pick a larger number of positively selected sites than the M8, FEL, or REL model and therefore may be more likely to falsely identify positively selected sites.
A caspase cleavage site has been identified in Drosophila (Stoven et al. 2003), which aligns with two potential cleavage sites in N. graveolus (fig. 3; site 557, L-L-R-D-G, and site 572, I-S-T-D-G). We used the program GrabCas (Meese 2005) and N. fumigatus sequence (sequence with the smallest number of deletions) to identify potential caspase cleavage sites. Cleavage sites are identified according to their similarity with consensus recognition motifs, which are based on experimental data (Thornberry et al. 1997). The caspase cleavage site in Drosophila, L-Q-H-D-G, resembles the consensus target site for group III caspases that includes human caspases 6, 8, 9 (Stoven et al. 2003). The fly caspase, Dredd, which cleaves Relish in Drosophila (Stoven et al. 2003), is most similar to caspase 8 (Inohara et al. 1997). Site 572, I-S-T-D-G, and site 557, L-L-R-D-G, are identified by GrabCas as termite Relish sites that share similarity with the caspase 8 consensus target site (3rd and 15th choice, respectively, of 72 sites ranked according to similarity with the consensus sequence). We also identified a site at 880, L-E-H-D-G, which is striking because it matches the consensus target site for caspase 9 and is the most similar site (highest rank among 72 sites) to the caspase 8 consensus sequence. The third amino acid in this putative cleavage site corresponds to a site identified by the M3 model in PAML as positively selected (site 882). In N. triodiae, sites 558, L-R-G-G-D, and 880, L-E-G-G-D, have had mutations resulting in tandem glycines, which abolish any match to sites that resemble the caspase consensus sequences.
In addition to using models that identify positively selected sites, we used PAML models that identify positive selection in specific lineages. Miller (1997) suggests that the subclade (Np,(Nd,Nf)) and the other Australian Nasutitermes should be reclassified as two new genera, based on a cladistic analysis of morphological and habit characters of Australian Nasutitermitinae. There appear to have been important changes in the structure of the mandibles and gut along the lineage rooting (Np,(Nd,Nf)). We compared models with one-ratio of dN/dS along branches to two-ratio models with a separate unconstrained dN/dS ratio along the branch rooting (Np,(Nd,Nf)). In the two-ratio model, the dN/dS for the branch rooting (Np,(Nd,Nf)) was greater than one for Relish (7.5 nonsynonymous substitutions and 0 synonymous substitutions) and GNBP2 (7.6 nonsynonymous substitutions and 0 synonymous substitutions) but not for GNBP1 (6.9 nonsynonymous substitutions and 3.0 synonymous substitutions), and for Relish and GNBP2, the two-ratio model was significantly better than the one-ratio model (table 4). We then compared models with a separate unconstrained dN/dS ratio along the branch rooting (Np,(Nd,Nf)) to models with a constrained dN/dS ratio of one to test if the dN/dS ratio for this internal lineage was significantly greater than one. There was a significant difference for both Relish and GNBP2 (table 4).
|
We used a genetic algorithm approach that searches for an optimal model of lineage-specific evolution by assigning four unrestricted classes of dN/dS to lineages (Kosakovsky Pond and Frost 2005a
). This approach allows an averaged model probability that dN/dS is greater than one along a specific lineage. Ninety-five percent confidence intervals (CIs) for the Akaike index criterion, AICc (Kosakovsky Pond and Frost 2005a), for the best model (c-AIC for Relish = 1,2068.0, GNBP1 = 5,674.5, and GNBP2 = 5,336.3) did not overlap with the AICc measure for the single-rate model (c-AIC for Relish = 1,2092.8, GNBP1 = 5,693.8, and GNBP2 = 5,352.4) for all three loci (c-AIC threshold for inclusion for Relish = 1,2072.5, GNBP1 = 5,687.1, and GNBP2 = 5,342.2). For all three loci, the lineage rooting (Np,(Nd,Nf)) was placed into a dN/dS category greater than one, with a model-averaged probability of 99.8% for Relish, 95.8% for GNBP2, and 89.6% for GNBP1. Several other branches were categorized as positively selected (fig. 4). However, the 95% CIs for individual branch estimates of dN/dS were significantly different from one in only three cases. The GNBP2 and Relish lineage rooting (Np,(Nd,Nf)) had median values of dN/dS = 2.47, 95% CI of 1.42–5.54 and dN/dS = 4.58, 95% CI of 2.87–6.35, respectively, and the apical Relish branch leading to N. fumigatus, which had a median value of dN/dS = 4.57, 95% CI of 2.89–6.35. We estimated the number of nonsynonymous and synonymous substitutions along each branch with the free-ratio model in PAML (fig. 5). The PAML free-ratio model and HYPHY genetic algorithm approach identified the same lineages as positively selected, and the lineages with higher frequencies of nonsynonymous substitutions had higher model-averaged probabilities (compare figs. 4 and 5).
|
|
The SIMMAP results (not shown) did not differ appreciably between the sets of conditions used and indicated that, with P
1, the common ancestor of the group Np, Nd, and Nf was both subterranean and fed on rotten wood. The results are equivocal for the common ancestor of this group with Nc, with about equal chances of epigeal and subterranean nest site and rotten wood or grass feeding. Across the tree, there is a very significant association between the traits of nest site and food, with highly significant associations of subterranean nests and rotten wood feeding, epigeal nests and grass feeding, and arboreal nests and feeding on sound wood. In contrast to Miura, Roisin, and Matsumoto (2000), the results indicate that arboreal nesting in the Australian Nasutitermes is a derived trait. |
Discussion |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
A comparison between termite Relish and GNBPs indicates that Relish has experienced greater selective pressure to change its amino acid composition than have the GNBPs. The models identified a larger number of positively selected sites in Relish, and there was consensus between models for five of these sites. We have previously found evidence of strong positive selection in defensin-like antimicrobial peptides among the Australian Nasutitermes (Bulmer and Crozier 2004
). Positive selection therefore appears to be driving the molecular evolution of a range of innate immunity genes in Nasutitermes, although our results suggest that this selection is relatively weak among the GNBPs. The cell-wall moieties that PRPs recognize, molecules such as peptidoglycans that are not found in the host, may not be able to readily change properties such as molecular charge and shape in order to evade detection. The selective pressure on PRPs may therefore be relaxed relative to other components of the innate immune system in which coevolutionary interaction between host and pathogen occurs between molecules that have more room for maneuver. In other invertebrates, PRPs have been shown to be constrained by purifying selection (Jiggins and Hurst 2003; Little, Colbourne, and Crease 2004). In contrast to these studies, we may have been able to detect positive selection in the GNBPs because we compared a larger number of ecologically diverse species, and/or termites are exposed to relatively strong selective pressure acting on their immunity genes because group living increases their susceptibility to disease.
Positive selection appears to have occurred along a subset of lineages among all three genes (fig. 4). Moreover, the selective pressure for change differs between immunity genes along the majority of Nasutitermes lineages. We suggest that this variation is mediated by the diversity in the different loads and types of pathogens that the nasutes were exposed to as they radiated (Bulmer and Crozier 2004). For example, if the two GNBPs are responsible for the recognition of different pathogens, which appears to be the case for the PGRPs in Drosophila (Lemaitre, Reichhart, and Hoffmann 1997; Stenbak et al. 2004), then the selective pressure could vary between them as the termites radiated into new niches and were exposed to novel microbial environments.
Positive selection of Relish and GNBP2 and possibly GNBP1 coincides in a single lineage rooting a subset of subterranean termites, which share a common ancestor with N. comatus. Nasutitermes comatus is found in drier savannah-like regions of interior and Western Australia, whereas the others are found in moister habitat along a narrow eastern coastal strip (Watson and Abbey 1993; Miller 1997). The evolution of this subclade is also associated with a transition in the morphology of regions of the gut and mandibles that corresponds with a shift in diet from grass and cellulosic material in the soil litter to decayed wood (Miller 1997). During this transition, termites are likely to have faced a large suite of novel pathogens in decaying wood (Rosengaus et al. 2003), which could drive the adaptive evolution of all three immune genes. The coincidence in selective pressure on three immune genes suggests that there may have been sweeping changes in the innate immune system and provides support for Miller's (1997) proposal that N. pluvialis, N. dixoni, and N. fumigatus should be elevated to a separate genus. However, a new genus of this composition would render Nasutitermes paraphyletic and might be seen to require additional nomenclatural changes. To preserve the genus, we suggest that it is more useful not only to retain these three species as members of Nasutitermes but also to synonymize T. pastinator under Nasutitermes. Positive selection also coincides between Relish and GNBP2 in an ancestral lineage of two arboreal nesting termites that feed on wood (fig. 4, N. graveolus and N. walkeri). However, this selection results in only a single nonsynonymous substitution in each locus (fig. 5).
Positively selected sites and insertions/deletions in Relish are located in regions that are important in its activation. Four of five positively selected sites identified by all three models of site-specific selection are clustered in a "spacer" region (Dushay, Åsling, and Hultmark 1996) that in Drosophila is cleaved by the caspase Dredd (Stoven et al. 2003). The fifth site is located in a region that corresponds to a PEST-like domain that may be an important docking point for the kinase complex IKK. In D. simulans Relish, sites in the spacer region and a region including the PEST domain do not appear to be evolving neutrally, in contrast to the rest of the molecule (Begun and Whitley 2000). Begun and Whitley speculated that microbial pathogens might be interfering with Relish activation to prevent an innate immune response. In support of this hypothesis, bacteria have systems that can disrupt signaling in the innate immune system. For example, Yersinia can inhibit the phosphorylation of NF-B with a type III secretion system (Neish et al. 2000), and enteric bacteria inhibit Relish activation with a mechanism that is independent of the type III secretion system and may instead block caspase cleavage of Relish (Lindmark et al. 2001). The location of the positively selected sites and the insertions and deletions in termite Relish suggest that the molecule is under pressure to evolve counter responses to bacterial mechanisms that block the caspase cleavage or IKK phosphorylation. An adaptive counter response between Relish and bacterial secretory components that enter the cell may be a long-term arms race, given the apparent similarity in site-specific selective pressure between two very distantly related insects, termites and fruit flies. Relish does not appear to be under pressure to change in D. melanogaster (Begun and Whitley 2000) or the majority of Australian Nasutitermes lineages (fig. 4), which indicates that a molecular arms race involving specific components of the innate immune system has not been a continuous process with the radiation of these insects. Several factors may account for this, such as adaptive change that severely limits a pathogen's ability to evolve a specific molecular counter response, selection pressure for the loss of pathogen virulence, or periodic relaxation in pathogen pressure with the occupation of novel ecological niches.
As indicated above, the findings of positive selection in the evolution of PRPs such as the GNBPs (this paper) and in effector molecules such as termicin (Bulmer and Crozier 2004) are counterintuitive in the light of the supposed limited room for change in these molecules. The findings then support expectations of both great selection pressure acting on social insects with their populous colonies (Schmid-Hempel 1998; Schmid-Hempel and Crozier 1999) and significant changes in intensity and direction (kinds of pathogens) resulting from changes in microhabitat and food.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary materials are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/). Alignment of the protein sequence of Relish for 11 species of Nasutitermes and the outgroup D. rubriceps. The position of different domains is indicated above the alignments. ANK corresponds with an ankyrin repeat. Positively selected sites are identified with an asterisk. The positions of insertions and deletions are indicated with an arrow.
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
This work was supported by a grant from the Australian Research Council to R.H.C. We thank A. J. Shuetrim and J. A. Holt for assistance and advice with insect collections and Y. C. Crozier for assistance and guidance in the laboratory.
|
Footnotes |
|---|
Laura Katz, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Begun, D. J., and P. Whitley. 2000. Adaptive evolution of Relish, a Drosophila NF- B/IB protein. Genetics 154:1231–1238.
Bollback, J. P. 2005. SIMMAP. Stochastic character mapping of discrete traits on phylogenies. Version 1.0 Beta 2.0. Distributed by the University of California at San Diego. (http://brahms.ucsd.edu/simmap.html).
Boniotto, M., A. Tossi, M. DelPero, S. Sgubin, N. Antcheva, D. Santon, J. Masters, and S. Crovella. 2003. Evolution of the beta defensin 2 gene in primates. Genes Immun. 4:251–257.[CrossRef][Web of Science][Medline]
Bulmer, M. S., and R. H. Crozier. 2004. Duplication and diversifying selection among termite antifungal peptides. Mol. Biol. Evol. 21:2256–2264.
Dushay, M. S., B. Åsling, and D. Hultmark. 1996. Origins of immunity: Relish, a compound Rel-like gene in the antibacterial defense of Drosophila. Proc. Natl. Acad. Sci. USA 93:10343–10347.
Girardin, S. E., P. J. Sansonetti, and D. J. Philpott. 2002. Intracellular vs extracellular recognition of pathogens—common concepts in mammals and flies. Trends Microbiol. 10:193–199.[CrossRef][Web of Science][Medline]
Hasegawa, M., H. Kishino, and T. A. Yano. 1955. Dating of the human ape splitting by a molecular clock of mitochondrial-DNA. J. Mol. Evol. 22:160–174.
Henikoff, S., J. G. Henikoff, W. J. Alford, and S. Pietrokovski. 1995. Automated construction and graphical presentation of protein blocks from unaligned. Gene 163:GC17–GC26sequences [reprinted from Gene-Combis 163:Gc17–Gc26, 1995].[CrossRef][Medline]
Higgins, D. G., J. D. Thompson, and T. J. Gibson. 1996. Using CLUSTAL for multiple sequence alignments. Pp. 383–402 in R. F. Doolittle, ed. Computer methods for macromolecular sequence analysis. Academic Press Inc., San Diego, Calif.
Huelsenbeck, J. P., R. Nielsen, and J. P. Bollback. 2003. Stochastic mapping of morphological characters. Syst. Biol. 52:131–158.[CrossRef][Web of Science][Medline]
Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754–755.
Hughes, A. L., and M. Yeager. 1997. Coordinated amino acid changes in the evolution of mammalian defensins. J. Mol. Evol. 44:675–682.[CrossRef][Web of Science][Medline]
Inohara, N., T. Koseki, Y. M. Hu, S. Chen, and G. Nunez. 1997. CLARP, a death effector domain-containing protein interacts with caspase-8 and regulates apoptosis. Proc. Natl. Acad. Sci. USA 94:10717–10722.
Janeway, C. A., and R. Medzhitov. 2002. Innate immune recognition. Annu. Rev. Immunol. 20:197–216.[CrossRef][Web of Science][Medline]
Jiggins, F. M., and G. D. D. Hurst. 2003. The evolution of parasite recognition genes in the innate immune system: purifying selection on Drosophila melanogaster peptidoglycan recognition proteins. J. Mol. Evol. 57:598–605.[CrossRef][Web of Science][Medline]
Kosakovsky Pond, S. L., and S. D. W. Frost. 2005a. A genetic algorithm approach to detecting lineages-specific variation in selection pressure. Mol. Biol. Evol. 22:478–485.
———. 2005b. Not so different after all: a comparison of methods for detecting amino-acid sites under selection. Mol. Biol. Evol. 22:1208–1222.
Kosakovsky Pond, S. L., S. D. W. Frost, and S. V. Muse. 2005. HyPhy: hypothesis testing using phylogenies. Bioinformatics 21:676–679.
Krishna, K. 1970. Taxonomy, phylogeny, and distribution of termites. Pp. 127–152 in K. Krishna and F. M. Weesner, eds. Biology of termites. Academic Press, New York.
Lazzaro, B. P., and A. G. Clark. 2003. Molecular population genetics of inducible antibacterial peptide genes in Drosophila melanogaster. Mol. Biol. Evol. 20:914–923.
Lemaitre, B., J. M. Reichhart, and J. A. Hoffmann. 1997. Drosophila host defense: differential induction of antimicrobial peptide genes after infection by various classes of microorganisms. Proc. Natl. Acad. Sci. USA 94:14614–14619.
Leulier, F., C. Parquet, S. Pili-Floury, J. H. Ryu, M. Caroff, W. J. Lee, D. Mengin-Lecreulx, and B. Lemaitre. 2003. The Drosophila immune system detects bacteria through specific peptidoglycan recognition. Nat. Immunol. 4:478–484.[CrossRef][Web of Science][Medline]
Lindmark, H., K. C. Johansson, S. Stoven, D. Hultmark, Y. Engstrom, and K. Soderhall. 2001. Enteric bacteria counteract lipopolysaccharide induction of antimicrobial peptide genes. J. Immunol. 167:6920–6923.
Little, T. J., J. K. Colbourne, and T. J. Crease. 2004. Molecular evolution of Daphnia immunity genes: polymorphism in a gram-negative binding protein gene and an alpha-2-macroglobulin gene. J. Mol. Evol. 59:498–506.[CrossRef][Web of Science][Medline]
Loker, E. S., C. M. Adema, S.-M. Zhang, and T. B. Kepler. 2004. Invertebrate immune systems—not homogeneous, not simple, not well understood. Immunol. Rev. 198:10–24.[CrossRef][Web of Science][Medline]
Lynn, D. J., A. T. Lloyd, M. A. Fares, and C. O'Farrelly. 2004. Evidence of positively selected sites in mammalian 
-defensins. Mol. Biol. Evol. 21:819–827.
Marchler-Bauer, A., and S. H. Bryant. 2004. CD-Search: protein domain annotations on the fly. Nucleic Acids Res. 32:W327–W331.
Meese, E. 2005. GraBCas. (http://www.med-rz.uni-sb.de/med_fak/humangenetik/software/index.html).
Miller, L. R. 1997. Systematics of the Australian Nasutiterminae with reference to evolution within the Termitidae (Isoptera). Australian National University, Canberra, Australia.
Miura, T., K. Maekawa, O. Kitade, T. Abe, and T. Matsumoto. 1998. Phylogenetic relationships among subfamilies in higher termites (Isoptera, Termitidae) based on mitochondrial COII gene sequences. Ann. Entomol. Soc. Am. 91:515–523.
Miura, T., Y. Roisin, and T. Matsumoto. 2000. Molecular phylogeny and biogeography of the nasute termite Genus Nasutitermes (Isoptera: Termitidae) in the Pacific tropics. Mol. Phylogenet. Evol. 17:1–10.[CrossRef][Web of Science][Medline]
Morrison, G. M., C. A. M. Semple, F. M. Kilanowski, R. E. Hill, and J. R. Dorin. 2003. Signal sequence conservation and mature peptide divergence within subgroups of the murine beta-defensin gene family. Mol. Biol. Evol. 20:460–470.
Neish, A. S., A. T. Gewirtz, H. Zeng, A. N. Young, M. E. Hobert, V. Karmali, A. S. Rao, and J. L. Madara. 2000. Prokaryotic regulation of epithelial responses by inhibition of I kappa B-alpha ubiquitination. Science 289:1560–1563.
Nylander, J. A. A. 2004. MrModelTest v2. Distributed by the author, Evolutionary Biology Center, University of Uppsala, Uppsala, Sweden.
Orth, K., L. E. Palmer, Z. Q. Bao, S. Stewart, A. E. Rudolph, J. B. Bliska, and J. E. Dixon. 1999. Inhibition of the mitogen-activated protein kinase kinase superfamily by a Yersinia effector. Science 285:1920–1923.
Pili-Floury, S., F. Leulier, K. Takahashi, K. Saigo, E. Samain, R. Ueda, and B. Lemaitre. 2004. In vivo RNA interference analysis reveals an unexpected role for GNBP1 in the defense against gram-positive bacterial infection in Drosophila adults. J. Biol. Chem. 279:12848–12853.
Rambaut, A. 2002. Sequence Alignment Editor. Department of Zoology, University of Oxford, Oxford. (http://evolve.zoo.ox.ac.uk).
Rose, T. M., E. R. Schultz, J. G. Henikoff, S. Pietrokovski, C. M. McCallum, and S. Henikoff. 1998. Consensus-degenerate hybrid oligonucleotide primers for amplification of distantly related sequences. Nucleic Acids Res. 26:1628–1635.
Rosengaus, R. B., J. E. Moustakas, D. V. Calleri, and J. F. A. Traniello. 2003. Nesting ecology and cuticular microbial loads in dampwood (Zootermopsis angusticollis) and drywood termites (Incisitermes minor, I. schwarzi, Cryptotermes cavifrons). J. Insect Sci. 3:1–6.
Salmeron, A., J. Janzen, Y. Soneji, N. Bump, J. Kamens, H. Allen, and S. C. Ley. 2001. Direct phosphorylation of NF-kappa B1 p105 by the I kappa B kinase complex on serine 927 is essential for signal-induced p105 proteolysis. J. Biol. Chem. 276:22215–22222.
Schmid-Hempel, P. 1998. Parasites in social insects. Princeton University Press, Princeton, N.J.
Schmid-Hempel, P., and R. H. Crozier. 1999. Polyandry vs. polygyny vs. parasites. Phil. Trans. R. Soc. B 354:507–515.[CrossRef]
Schmidt, H. A., K. Strimmer, M. Vingron, and A. von Haeseler. 2002. TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18:502–504.
Semple, C. A., M. Rolfe, and J. R. Dorin. 2003. Duplication and selection in the evolution of primate ß-defensin genes. Genome Biol. 4:R31.[CrossRef][Medline]
Shin, S. W., V. Kokoza, A. Ahmed, and A. S. Raikhel. 2002. Characterization of three alternatively spliced isoforms of the Rel/NF-kappa B transcription factor Relish from the mosquito Aedes aegypti. Proc. Natl. Acad. Sci. USA 99:9978–9983.
Stenbak, C. R., J. H. Ryu, F. Leulier et al. (11 co-authors). 2004. Peptidoglycan molecular requirements allowing detection by the Drosophila immune deficiency pathway. J. Immunol. 173:7339–7348.
Stoven, S., N. Silverman, A. Junell, M. Hedengren-Olcott, D. Erturk, Y. Engstrom, T. Maniatis, and D. Hultmark. 2003. Caspase-mediated processing of the Drosophila NF-kappa B factor Relish. Proc. Natl. Acad. Sci. USA 100:5991–5996.
Swanson, W. J., R. Nielsen, and Q. Yang. 2003. Pervasive adaptive evolution in mammalian fertilization proteins. Mol Biol Evol 20:18–20.
Swofford, D. L. 1998. PAUP*: phylogenetic analysis using parsimony (* and other methods). Sinauer Associates, Sunderland, Mass.
Thompson, G. J., O. Kitade, N. Lo, and R. H. Crozier. 2000. Phylogenetic evidence for a single, ancestral origin of a ‘true’ worker caste in termites. J. Evol. Biol. 13:869–881.[CrossRef][Web of Science]
Thornberry, N. A., T. A. Ranon, E. P. Pieterson et al. (12 co-authors). 1997. A combinatorial approach defines specificities of members of the caspase family and granzyme B—functional relationships established for key mediators of apoptosis. J. Biol. Chem. 272:17907–17911.
Watson, J. A. L., and H. M. Abbey. 1993. Atlas of Australian termites. Commonwealth Scientific and Industrial Research Organisation, Canberra, Australia.
Yang, Z. 1997. PAML: a program package for phylogenetic analysis by maximum likelihood. CABIOS 13:555–556.
Yang, Z., W. S. Wong, and R. Nielsen. 2005. Bayes empirical Bayes inference of amino acid sites under positive selection. Mol. Biol. Evol. 22:1107–1118.
Yang, Z. H. 1998. Likelihood ratio tests for detecting positive selection and application to primate lysozyme evolution. Mol. Biol. Evol. 15:568–573.[Abstract]
Yang, Z. H., R. Nielsen, N. Goldman, and A. M. K. Pedersen. 2000. Codon-substitution models for heterogeneous selection pressure at amino acid sites. Genetics 155:431–449.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Y. Pauchet, D. Freitak, H. M. Heidel-Fischer, D. G. Heckel, and H. Vogel Immunity or Digestion: GLUCANASE ACTIVITY IN A GLUCAN-BINDING PROTEIN FAMILY FROM LEPIDOPTERA J. Biol. Chem., January 23, 2009; 284(4): 2214 - 2224. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. V Calleri II, E. McGrail Reid, R. B Rosengaus, E. L Vargo, and J. F.A Traniello Inbreeding and disease resistance in a social insect: effects of heterozygosity on immunocompetence in the termite Zootermopsis angusticollis Proc R Soc B, October 22, 2006; 273(1601): 2633 - 2640. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||