In insects, immune functions play pivotal role in controlling pathogenic and parasitic attack. The immune mechanism comprised of activation of humoral immune system, epithelial responses, cellular immunity, non-specific melanisation reactions, cytokine changes, inflammation , proteolysis and stress responses [1,2]. Though host responses after microbial infection are well analyzed [3,4], those after eukaryotic parasite attack in insects are seldom reported [5,6]. Parasitic infection is known to be recognized by circulating hemocytes causing transient activation of immune genes to synthesize effector molecules as well as to induce encapsulation, phagocytosis, melanisation etc [7,8].

During large scale rearing of commercially important silkproducing lepidopteran insect, Bombyx mori, the worms are often attacked by bacteria, virus and fungus. After the infection, innate immunity components including peptidoglycan recognition proteins (PGRP), C-type lectins, fibrinogen-related proteins (FRP) and scavenger receptors were activated [4,9]. In addition to pathogenic attacks, B. mori larvae are infested by a dipteran fly, Exorista bombycis (Tachnidae). The fly lays eggs on larval cuticle. On hatching, maggot invade the body and live as an endoparasitoid that induced organismal variations like loss of biomass, less production of silk protein, biochemical changes etc., [10]. Our recent studies showed differential production of host-response proteins, upregulation of associated genes and activation of immunocompetence in B. mori induced by the infestation [11]. The host-response proteins comprised of stress proteins, cytokines, Toll-pathway components, proteolytic enzymes, apoptosis – associated proteins and melanisation pathway components [12,13]. Genome-wide analysis showed that some of the immune proteins are paralogous [4,14] however structural differences indicated that the genes encoding them might have originated from a common ancestor by duplication or as splice variants [4] under different types of selection pressure [15]. Some genes like hemolin evolved as lineage-specific in different lepidopterans, Hyalophora cecropia, Samia cynthia ricini, Antheraea pernyi and B. mori [4,16]. Altogether, immune components of different insects belong to orthologous and paralogus proteins as well as independent lineage proteins indicating its evolutionary conservation [17,18].

Phylogenetic analsis involving various proteins of different families and functions point to its origin and divergence. Analysis of major intrinsic proteins (MIPs), plasma membrane intrinsic proteins (PIPs) and nodulin 26 like intrinsic proteins (NIPs) revealed protein phylogeny in eukaryotes [19,20]. Besides, bootstrap analysis allowed potential comparison of proteins that are conserved across different species and to divide them among different groups based on divergence [21].

In this report we examined conservation of host response proteins among different insect groups as well as divergence among the proteins through phylogenetic analysis using multiple host – response proteins identified from B. mori larvae and its similar sequences from few representative insect species, belong to different orders. The new approach using phylogenetic analysis using multiple host – response proteins revealed temporal divergence of the host – response proteins which is essential in vulnerable organisms for adaptation against parasitic attack.

Rearing of the silkworm, B. mori, infestation by E. bombycis and collection of integument, haemolymph and hemocyte samples were carried out as described earlier [11-13] and briefed here. Protein samples were extracted from integument using total protein extraction reagent (T-PER; Pierce). For protein from hemocytes, the cells were pelletized from haemolymph by centrifugation at 3000 rpm at 4°C and plasma was separated. The hemocytes were treated with lysis buffer containing Triton-X 100 (Sigma), protease inhibitor (Sigma; 100 x) and phosphatase inhibitor (Sigma; 1 x) for 10 minutes on ice and the supernatant was separated by centrifugation at 10000 RPM for 10 minutes at 4°C and quantified [22]. The protein samples were collected at different time points from control and E. bombycis – infested final instar larvae of B. mori. The protein samples were denatured by 2M urea and reduced by 1M DTT before resolving on 10% SDS-PAGE. For 2D electrophoresis analysis, the protein was extracted in 2D buffer [23,24], followed by focussing on 11 cm strip (pH 3-10) on Protean IEF cell (Biorad), reduced in DTT and alkylated in iodoacetamide as described earlier [11]. The second dimension was resolved on a 12% precasted PAGE gel (Biorad). Exclusive bands/ spots were identified from the infested tissues, cut and digested by trypsin (In-gel digestion) before analysing by mass spectrometry (MALDI-TOF-MS) and denovo sequencing [25,26]. List of the peptides of different host response proteins identified from all the three tissues is presented in Supplementary Table 1 A-D.

Phylogenetic analysis

Amino acid sequences of 19 host-response proteins of B. mori were obtained from NCBI database. The sequences were used for homology search by BLASTp program (NCBI). In order to identify similarities and evolutionary relationships, amino acid sequences of the homologous proteins from representative insect species belonging to different orders, Lepidoptera, Diptera, Coleoptera and Hymenoptera collected from NCBI website were used for multiple sequence alignment with the B. mori sequences using Clustal W2 [27] (www.ebi.ac.uk/tools/msa/ clustalw2). Fully conserved amino acids (*), conserved substitutions (:), semi conserved substitutions (.) and amino acid replacements were identified from each protein set. Phylograms based on similarity were realized for each of the proteins and the branch strength was evaluated by Bootstrap values of 1000 replicates.

In order to analyze phylogenetic position of the host-response proteins, amino acid sequences of all the 19 proteins from B. mori and its similar sequences from the representative insects from different insect orders were aligned. The aligned sequences were used to construct phylogenetic tree based on maximum likelihood method using molecular evolutionary genetic analysis (MEGA version 5.05) program [28]. Bootstrap values for Tree building method were obtained from 1000 replicates. The branch length information was obtained based on genetic distances derived from Jones-Taylor-Thornton (JTT) model [29]. Branch lengths are proportional to sequence divergence. The bar represents 0.2 or 0.5 substitutions per site.

Bioinformatics analysis

Transmembrane regions of each protein were predicted by DAS (Dense Alignment Surface method [30] (http://www.sbc.su.se/~miklos/DAS/ ) and Signal peptides were predicted by SignalP 4.0 [31] (http://www.cbs.dtu.dk/services/SignalP/ ). Map position of each gene encoding the respective host- response protein in chromosomes was identified from KAIKOBASE (http://sgp.dna.affrc.go.jp/KAIKObase/ ). Conserved domains of each protein were identified from “CDD: NCBI’s conserved domain database” [32].

In invertebrates, immune system recognizes and gets rid of invading micro- and macroparasites through activation of humoral immunity and cell-mediated events. Invertebrates lack adaptive immunity represented by antibody production and key cellular players like T-cell and B-cell system present in mammals [33]. However humoral immune mechanism in invertebrates and vertebrates showed significant conservation in immune pathways indicated common evolutionary origin [34]. In the lepidopteran model and commercially important silkworm, B. mori, infestation by the eukaryotic parasitoid, E. bombycis induced enhanced expression of immune proteins and other host-response proteins. Protein extracts of integument, hemocytes and haemolymph of B. mori larvae infested by E. bombycis and control larvae were resolved on 10% SDS-PAGE or by 2D electrophoresis. The gels showed exclusive bands and spots in the extracts of infested larvae and were absent in the tissue protein extracts of non-parasitized controls injected with 0.9% NaCl and in non-parasitized non-injected controls (Supplementary Figure 1). Mass spectrometry of these exclusive bands/spots from all the tissues showed specific peptides of proteins that perform different functions associated with humoral immunity, signalling, stress responses, cell homeostasis, autophagy, apoptosis, and proteolysis as well as melanisation reactions (Table 1). Observed and calculated mass, amino acid sequences of the peptides and intensity are given in Supplementary Tables 1A-1D. This indicated activation of heterogeneous group of host-response proteins which exhibit co-regulation, cross talk and synergetic activation after the infestation. Similar activation of Toll- and IMD pathways was reported from the honey bee Apis mellifera [3] and Drosophila [35,36]. Genes encoding these proteins (except paralytic peptide) have already been mapped on different chromosomes of B. mori (KAIKOBASE; Table 2).

Table 1: Biological function of the host-response and immune-response proteins identified by mass spectrometry and three candidate proteins from integument, haemolymph and hemocytes of the final instar larvae of the silkworm, B. mori infested with maggot of E. bombycis.

Table 2: Details of chromosome mapping of the genes of immune response/host response proteins induced in fifth instar larva of B. mori after infestation by E. bombycis (source:KAIKOBASE).

Fourteen peptides of the stress protein, heat shock protein (Hsp70) were identified from tissues of infested larvae of B. mori indicated its abundance (Supplementary Table 1A-D) as the measured mass spectrometry signals showed linear dependence on the amount of materials present in the entire sample analysed [37]. On the other hand, some of the proteins like DDC, paralytic peptide, Autophagy – 5 like (Atg5) and apoptosis-inducing factor (AIF) were represented by few peptides indicating low abundance. However few peptides might not represent all isoforms of a specific protein [37].

Amino acid length of the host – response proteins varied from 131 (paralytic peptide) to 2463 (Notch) (Table 3). Two cytokines viz., paralytic peptide and spatzle were with shorter sequence length whereas signalling proteins (Toll and Notch) were with longest sequence (Figure 1). Average length of all except the signalling proteins was 501 ± 195 amino acids. Many of the major intrinsic proteins of insects (e.g., aquaporins) are over 600 amino acids length due to extended C-terminal ends [19] which support greater receptor-dependent signalling [38]. Out of the 19 proteins identified from B. mori, 18 had one or more transmembrane (TM) regions. Notch showed eight TM regions, AIF with four, Toll with three and trehalase with two TM regions. The TM domains prevent proteins moving out of the locale [39] and are essential for transmembrane interactions and stability [40] thereby enhances immune reactions at specific infected locations. Divergence and mutations in these regions affect functional variability of the immune proteins [41]. Seven proteins were with signal peptides and presumed to be transmembranous and/ or secretory in nature (Table 4). However, TM region/ signal peptide was not detected in phenol oxidase (PPO 2) sequence corroborate the view that PPO 2 may be released to haemolymph by cell rupture [4,42].

Table 3: Key to amino acid substitutions and conservations in the host- response proteins of B. mori when aligned with the similar proteins from representative species of other insect orders presented in supplementary Table 2.

Table 4: Potential transmembrane region and Signal peptides from immune and host response proteins induced in fifth instar larva of B. mori through infection by E. bombycis.

Figure 1: Distribution of number of amino acids of host-response proteins that are activated in the final instar larva of B. mori after infestation by the dipteran parasitoid, E. bombycis. Note that signal proteins possess maximum number of amino acids and cytokines possess minimum number.

The similar/orthologous sequences possess highest amino acid conservation and low interferences due to long indels and larger divergence [19]. The amino acid substitutions in them could provide differences within a particular insect order and phylogenetic information among the orders as well. Therefore similarity and divergence of the host – response proteins of B. mori was compared with similar proteins of other insects, identified by BLASTp (NCBI). Representative species of insects that showed highest level of similarity were selected from the insect orders, Lepidoptera, Diptera, Hymenoptera and Coleoptera (Supplementary Table 2). Out of them, search for immune genes that were activated after microbial infection, were performed in Drosophila melanogaster, Anopheles gambiae, Aedes aegypti, Apis mellifera and Tribolium castaneum in addition to B. mori [3,4,43-47]. In D. melanogaster, a similar search for immune genes had been made after infection by the eukaryotic parasitoid, Asobara tabida (Hymenoptera: Braconidae) [5]. In the present study, the amino acid similarity was analysed by multiple sequence alignment of each host response protein from B. mori with its similar sequences and the divergence was assessed through unrooted phylogram (Figures 2A- 2S) which showed independent groupings with high bootstrap values revealing adaptive evolution in the immune proteins, depending on its functional differences [48].

Figure 2A-2S: Phylogram realised after performing multiple sequence alignment by CLUSTAL W using total amino acid sequence of host-response proteins that are activated in the final instar larva of B. mori and its homologues from representatives of different insect orders: The proteins analysed were (A) – Hsp70, (B) – Chaperonin subunit 4 Delta, (C) – SERPIN – 4, (D) – paralytic peptide (PP), (E) – prophenol oxidase activating enzyme (PPAE), (F) – DOPA decarboxylase (DDC), (G) – prophenol oxidase 2 (PPO2), (H) – Chitinase, (I) – β-N-Acetylglucosaminidase, (J) – Spatzle, (K) – BmToll (18 wheeler), (L) – Nedd2 like Caspase, (M) – Autophagy 5- like (Atg5), (N) – Apoptosis - inducing factor (AIF), (O) – Trehalase, (P) – Notch, (Q) Dorsal, (R) – Relish and (S) – Cactus. Details of accession numbers of proteins, name of the insects and similarity exhibited are given in the Supplementary Table 2. Numbers in the nodes represent bootstrap values computed from 1000 replications.

B. mori Hsp70 is a secretory protein that is grouped with homologues of other lepidopterans with high bootstrap value (91-95%) (Figure 2A). The B. mori Hsp70 showed early divergence whereas that of other three lepidopteran species (Danaus plexippus, Helicoverpa armigera and Spodoptera frugiperda) showed orthology based on high bootstrap value (91–96%). In the coleopteran, T. castaneum, Hsp70 showed an independent genetic differentiation from other species. Similarly, chaperonin of B. mori and D. plexippus grouped independently with large bootstrap value (100%) (Figure 2B). However, the group showed delayed divergence from ancestor.

The serine protease inhibitor, Serpin-4 of B. mori showed close proximity with that of D. plexippus and grouped separately with high bootstrap value (98%) (Figure 2C). SERPIN-4 of dipterans formed a separate cluster.

Paralytic peptide (PP) is an insect cytokine belongs to a group of growth inhibitors. Multiple sequence alignment of BmPP showed similarity with that of Samia cynthia pryeri, but with a low bootstrap value (23%). PP of Spodoptera litura and Mythimna separata formed separate grouping with high bootstrap value of 81% (Figure 2D).

Prophenol oxidase activating enzyme (PPAE), a serine protease of B. mori is clustered with that of Manduca sexta with 98% bootstrap value (Figure 2E). PPAE of M. sexta and B. mori appeared as orthologs however PPAE of D. plexippus seems originated at an early stage from a common ancestor of all the lepidopterans. PPAE of all dipterans were clustered separately with 76% bootstrap value.

DDC phylogram showed three major groups; a cluster of B. mori and other lepidopterans, a group of dipterans and an isolate of hymenopteran representative with high bootstrap value (Figure 2F). Among the lepidopterans, DDC of B. mori and Manduca mucosa were orthologous to others with higher bootstrap value (78%) and showed phylogenetic differentiation at a later stage.

In B. mori, two subunits of prophenol oxidase (PPO) were reported viz., PPO1 and PPO2 [42]. However in our study, only PPO2 was identified by mass spectrometry indicating its larger expression after the infestation. The PPO2 phylogram consisted of two independent groupings, one of lepidopterans with 100% bootstrap values and another with all other insect species. Among the lepidopterans, PPO2 of Helicoverpa armigera and Spodoptera frugiperda were diverged at a later stage from a common ancestor and formed orthologs of B. mori (Figure 2G). PPO2 of coleopterans and dipterans diverged earlier than lepidopterans from the ancestor.

An endochitinase, chitinase of all the lepidopteran representatives, were grouped together with high bootstrap value (100 %) and showed an independent lineage descended from a common ancestor (Figure 2H). Chitinase of Danaus plexippus showed earlier divergence followed by B. mori and M. sexta. Chitinase of Helicoverpa armigera and Spodoptera litura showed an orthologous development. Lepidopteran chitinases showed later divergence than other insect orders. An exochitinase, β-N-acetylglucosaminidase (βGAm) from B. mori, D. plexippus and Spodoptera frugiperda were grouped separately from other lepidopteran, dipteran, coleopteran and hymenopteran species (Figure 2I). βGAm of Danaus plexippus and Spodoptera frugiperda were orthologs of B. mori with high bootstrap value (100%).

Two components of Toll pathway viz., a cytokine, spatzle and BmToll were identified from the parasitized larvae. Spatzle phylogram showed two major groups, one for all lepidopterans together and a second formed of dipteran, hymenopteran and coleopterans insects together (Figure 2J). Within the lepidopteran cluster, B. mori and M. sexta Spatzle were differentiated into orthologs with high bootstrap value (75%). Notably, Drosophila melanogaster spatzle showed its presence as outgroup to all other insect representatives. BmToll showed two major clusters, one for dipterans alone and another for lepidopterans and hymenopterans together (Figure 2K). B. mori and D. plexippus Toll were orthologus to that of S. frugiperda with high bootstrap value (97%).

The apoptosis-inducing protein, B. mori caspase (Nedd2-like caspase) was clustered with D. plexippus and H. armigera (Figure 2L). Caspase of Dipterans showed separate cluster with high bootstrap value whereas hymenopterans formed another subgroup. Atg5 of B. mori and D. plexippus showed separate subgroup from other lepidopterans, dipterans, coleopterans and hymenopterans (Figure 2M). B. mori Atg5 showed orthology to that of other insects with high bootstrap value (100%). AIF of B. mori and D. plexippus formed a group (Figure 2N) and were orthologous to other insects with higher bootstrap value (100%).

Membrane-bound trehalase-2 precursor of dipterans formed a major cluster. All other insect orders formed a second group including a lepidopteran subgroup. BmTrehalase showed close proximity with that of Spodoptera litura (Figure 2O) and is orthologous to other insects.

The signalling protein, Notch from B. mori and Tribolium castaneum formed a group with a bootstrap value of 69% whereas that of dipterans formed a separate group (Figure 2P).

Though Toll and Spatzle were detected by mass spectrometry, no other components of Toll or IMD pathway were reported from any tissues of B. mori larva after infestation by E. bombycis. Hence two NF κB transcription factors (TF), Dorsal (Toll pathway component) and Relish (BmRel; IMD pathway component) and a NF κB inhibitor, Cactus were considered as candidate proteins for analysis. Phylogram showed that Dorsal and Rel of B. mori as orthologous to other species (Figure 2Q-2R). Dorsal of Lepidopterans grouped separately from other orders whereas Rel of B. mori and H. armigera formed separate group from all other representative insect species. Cactus of B. mori and D. plexippus clustered separately whereas that of dipterans and hymenopterans were subgrouped separately (Figure 2S).

Cluster formation in the phylogram of each host-response protein of B. mori is influenced by the presence of amino acid substitutions and insertions. The conserved, semiconserved and substituted amino acids as well as amino acid replacements in each protein are given in Table 3. The conserved substitutions were higher in the stress proteins Hsp70 and chaperonin (72 to 80%) as well as in melanisation pathway components, Cactus, chitinase and Atg5 (>50%). The signalling proteins such as BmToll and Notch as well as SERPIN and PPAE showed low level (~30%) of amino acid conservation. In PPO2 and Toll, six amino acid replacements were observed in B. mori (Table 3). The cytokines, paralytic peptide and spatzle did not show amino acid variations whereas other proteins showed one to five amino acid replacements. Accumulation of conserved amino acids were reported in proteins that are crucial to reduce host – parasite interactions in the Plasmodium - mosquito system [49,50], whereas, absence of amino acid variation indicates a role of purifying selection against rare allelic variability [51,52]. Marked replacements of amino acids in different immune proteins can also occur due to action of negative selection on the encoding genes [50,51,53,54]. Such proteins may be under selective constraint and no change happens in the protein due to either functional constraints of the protein itself or in their targets [15].

However substitution of amino acids within similar proteins can alter response of the protein towards infection [55,56]. Notably, in most of the proteins, the amino acid substitutions were occurred within the active conserved domains (Table 3) indicative of significant implications on interaction of the respective immune proteins.

Amino acid diversity among identical proteins was observed in the dipteran insects, A. gambiae and several species of Drosophila [43-45], in the coleopteran, T. castaneum [47], hymenopteran, A. mellifera [3] and in the lepidopteran, B. mori [4]. Present study showed amino acid substitutions in specific conserved domains of different immune proteins indicating that the divergences are necessary to defend individual species from species - specific parasites/pathogens and for adaptation to its ecological niche under selection pressure [19,44,57-59].

Though the role of selection is undisputable, how infection interacts with evolutionary forces is not clear. However it is known that linkage and recombination influences gene interaction [60]. Interaction between multiple loci or genes could contribute to phenotypic variability such as occurrence of disease [61,62] which may, in turn, be influenced by linkage or recombination. Genes encoding the host – response proteins of B. mori (present study) were mapped on 13 different chromosomes (Table 2). Three genes each were located in Chromosome numbers 12 and 15 whereas Chromosome 9 had two and other chromosomes had one each. Genes found at larger distance as well as those located in different chromosomes may be evolutionarily independent however those located closer may undergo recombination together. However, “rates of recombination are not a uniform function of physical distance along the DNA” [63,64]. In B. mori, genes encoding the Toll pathway components, NF κB transcription factor Dorsal and its inhibitor Cactus are located on chromosome 12. It is interesting to investigate whether these genes undergo recombination together or not. In B. mori, upregulation of Dorsal is accompanied by down regulation of Cactus at 48 h after the infestation by E. bombycis whereas the reverse trend is noticed at 72 h after infestation indicating a negative correlation in expression pattern of these genes [11,12]. Under normal conditions, Cactus binds with Dorsal to prevent the transcription of antimicrobial proteins (AMP). Following infection, Cactus undergoes degradation and releases Dorsal to induce AMP gene transcription. Moreover, down regulation of Cactus gene was observed after infection [12,65,66] demonstrating that Cactus regulates the functioning of Dorsal through a gene – mediated feedback mechanism also [67]. Presence of Cactus and Dorsal genes together on the chromosome 12 seemed to promote an effective regulation of the Toll pathway which is developed through adaptation and balanced selection [68] to meet antiparasitic responses [69,70]. It has been shown that another NF κB transcription factor, Rel gained function by recombination with a locus encoding cytoplasmic retention domains [71]. However the probable recombinant hotspots [64] of Cactus – Dorsal pair in chromosome 12 of B. mori need to be confirmed.

Amino acid insertions and divergence

Sequence alignment revealed few short insertion sequences in hostresponse proteins of B. mori (Table 2). In chaperonin delta subunit, proline, isoleucine and valine were inserted at 190-192^nd position in B. mori and at corresponding region in D. plexippus. Insertions of new amino acids in the stress protein at a later evolutionary stage indicated its requirement to adapt under changed cellular milieu induced by infestation [48]. Such adaptive variations happen at a larger speed in immune proteins than that occurred in non-immune proteins [44,58,72]. Similarly, the amino acid sequence, APNLDDFSMVIQV at position from 461 – 474 in AIF of B. mori, was absent in other insect species compared here (Figure 3). Notably, this C-terminal dimerization domain (Pfam 14721) is a part of the conserved domain of mitochondrial AIF which is critical for energy metabolism and induction of caspase-independent - apoptosis by interacting with DNA [73]. Similar ‘indels’ (insertions and deletions) have been reported in active sites of Relish in fruit flies [44,74] which indicates that both AIF of B. mori and Relish of Drosophila are under selective pressure [48,58] induced by respective parasitic interactions.

Figure 3: Multiple sequence alignment of portion of the amino acid sequence of apoptosis – inducing factor (AIF) of B. mori with homologues from other insect species showed an insertion of APNLDDFSMVIQV in B. mori which was absent in all other insects analysed. Symbol details are given in text.

Phylogenetic analysis

In order to understand functional diversity that evolved among various immune and host –response proteins to resist parasitic attack, a phylogenetic network among the proteins of representative species belonging to different insect orders is obligatory. A phylogram was realized through MEGA analysis with 1000 bootstrap replications using the aligned amino acid sequences of the 19 immune – and host - response proteins of B. mori as well as its similar sequences and orthologues from other insect species (Figure 4). The whole set of hostresponse proteins were grouped into three clades (Clusters A – C) of heterogeneous proteins which were supported by high bootstrap value (60-95%). Though the proteins were components of melanization pathway, Toll pathway, proteolytic enzymes, apoptosis-associated proteins, cytokines and chaperones from different insect orders, it showed uniform divergence pattern in the phylogeny tree indicating similar tendency of adaptation to develop immunity in different insect orders.

Figure 4: Phylogenetic analysis showing evolutionary relationship among host – response proteins that are activated in the final instar larva of B. mori after infestation by the dipteran parasitoid, E. bombycis: Unrooted phylogram revealed three clusters (A, B, C) which showed protein divergence at different timings of evolution. Oval circle represents central node. Note the early divergence of caspase and late divergence of BmToll in Cluster A, fork formation of Dorsal and BmRel from a common ancestor in Cluster B and co-differentiation of Atg5, AIF and PPO2 in Cluster C. Bold arrow heads indicate the bootstrap values greater than 50% computed from 1000 replications. Accession numbers in red fonts in each group represent protein from B. mori. Details of the proteins are as given in the legend of Figure 2 and Supplementary Table 2.

In B. mori, genes encoding the host-response proteins showed differential expression after the parasitic invasion [11-13]. Few immune genes showed regulated expression in control larvae which was deregulated due to the parasitic maggot infestation (Table 5). This is possibly to initiate new physiological roles for the gene products indicating functional divergence under the influence of infestation [75,76]. Thus a co-ordinated net-working of proteins of different functions is essential to elicit host defence mechanism against specific parasite attacks [4,77], which necessitate synchronized evolution of immune-associated proteins.

Table 5: Genes that showed pleiotropic regulation* of expression of different genes in the normal larvae of B. mori and deregulated after infestation with E. bombycis.

All proteins of Cluster A were with TM regions indicating functional significance of TM region in the immune proteins by enhancing the transmembrane interactions [40,56,77]. An exception in this cluster was apoptotic enzyme Nedd2-like caspase, which possess TM region however without signal peptide. Phylogram showed large genetic divergence in this cluster. Caspase showed earlier divergence, whereas, BmToll showed slow divergence revealing that functionally different proteins respond differentially to selection [48]. Multiple sequence alignment of caspase showed merely 18.48% conserved sequences, leading to rapid diversity and evolution [78] possibly to protect the cells through apoptosis induced by infection [78,79].

BmToll is a member of the Toll receptor family that signals attack of different pathogens/parasites in vertebrates and invertebrates. The phylogram showed that in B. mori, Toll had evolved at a later stage of evolution presumably due to slow but accumulated demand (77% varied amino acids and six replacements; Table 3) by pathogen invasion to activate immune responses which is a recent adaptation in holometabolous insects [56,80,81]. Thus the BmToll is under strong evolutionary pressure by infections in different higher organisms including human beings [56,82,83].

Presence of the cytokines, Spatzle and paralytic peptide in the cluster A and its divergence before BmToll was also key to adaptive evolution of immune system in B. mori. Functionally, spätzle binds with Toll to activate the pathway [84], whereas, paralytic peptide activates the DOPA decarboxylase (DDC) of the melanisation pathway [85,86] indicating that both specific (Toll) and non-specific (melanisation) immune reactions have diverged synchronously in the B. mori immune system.

Cluster B showed common ancestry of NF κB transcription factors, Relish and Dorsal. Relish and Dorsal showed divergence by forking from a single ancestor with higher bootstrap value (83%) supporting that it is originated from a generalized ancient signalling system [87] that possesses significant DNA binding preferences [88]. Sequence alignment of Dorsal and Relish showed 58% conservation in the amino acid sequence positioned from 1-344. However, from 345 – 470^th amino acid position, most of the amino acids present in Relish were absent in Dorsal showing 11 long deletions that seems to be the reason for functional diversity among the sister genes. Presence of ‘indel’ in Relish had been reported earlier in insects [15]. Similar deletions induced lack of enhancer activity in the desatF gene of D. melanogaster [89].

In cluster B, melanization initiation factor and prophenol oxidase activating enzyme (PPAE; a serine protease) showed early divergence through an independent lineage whereas serine protease inhibitor (SERPIN 4) was evolved at a later stage. Even though PPAE activates prophenol oxidase (PPO) to initiate melanisation [90], PPAE gene expression showed pleiotropic interaction with expression of more than one immune gene (Table 5). Association of PPAE with early and late evolved proteins of different immune functions displayed initiation of diverse immune processes by PPAE in B. mori larvae at different time points in the evolutionary tree.

An earlier activation of melanisation events is a prior requirement against specific or non-specific infection whereas continued melanisation could lead to quinone toxicity which could be prevented by the serine protease inhibitor (SERPIN). Thirty four SERPIN genes were identified from B. mori which might have originated by successive gene duplications at multiple genomic locations and formed gene clusters [91]. SERPIN 4 inhibits PPAE in the PPO activation pathway [92] and regulates quinone production. Moreover a compromised PO activity under the influence of selection was reported from the beetle T. castaneum [93]. This revealed that negative selection acts against immune system component after optimising an adapted trait (melanisation). Further a combination of negative (purifying) and positive selection could not only lead to low genetic variability [57] but interact to induce rapid functional and genetic diversity in immune proteins/genes [51,52,58].

Cluster C proteins did not possess signal peptide indicating that absence of signal peptides affect phylogenetic grouping of proteins. In this cluster, Atg5 and PPO2 formed a subgroup with 66% bootstrap value. Functionally, Atg5 is associated with autophagy [94,95]. Autophagy was induced in the integumental epithelium at 24 h after the E. bombycis maggot infestation in B. mori which was followed by lysis of the integument. The lysis showed a curvilinear relation with phenol oxidase (PO) activity indicating higher PO activity in mid – infestation which was preceded by activation of PPO2 gene [11,12]. Melanisation of the integument and autophagy in integumental epithelium are involved in protecting host larval integument from the early parasitic influence. This supports functional divergence and synchronized evolution of Atg5 and PPO2 from a common ancestor in the phylogram.

In the commercially important lepidopteran silkworm, Bombyx mori, heterogeneous group of exclusive host-response proteins showed its larger presence after infestation by the dipteran parasitoid, E. bombycis. Multiple sequence alignment of the host-response proteins with its similar sequences showed orthology and accumulation of conserved amino acids which are crucial in reducing the hostparasite interaction to suppress parasitic influence. Few amino acids were replaced from most of the proteins, presumably due to action of negative selection on the encoding genes.

However under the phylogenetic analysis, host-response proteins from B. mori and its homologues from other insect orders showed uniform diversification pattern. The phylogram revealed three clusters of host-response proteins with functional diversity and temporal evolution. Caspase showed earlier divergence and rapid evolution whereas BmToll evolved at a later stage indicating that functionally different proteins respond differentially to selection. Similarly, serine protease, PPAE showed early independent divergence whereas serine protease inhibitor (SERPIN) diverged later and had diverse functions. The NF κB transcription factors, Dorsal and Relish originated early from a common ancestor but diverged in a forked fashion to perform transcription, probably through two (Toll and Immuno deficiency) pathways. Prophenol oxidase, Atg5 and apoptosis inducing factor were originated from a common ancestor. Expression pattern of the genes encoding these proteins showed significant co-regulation which indicated that regulation of gene expression originated along with functional divergence. The phylogram showed temporal divergence of the cell death-associated proteins indicating its requirement in a temporal fashion to protect the cells from parasitism in the initial stages or to induce cellular death at later stages of infection. Thus the phylogenetic analysis of multiple host-response proteins of B. mori induced after the parasitic attack showed functional divergence and temporal evolution according to the functional requirement induced by parasitism in insects.

The authors are thankful to anonymous reviewers for critical reading and suggestions to improve the manuscript and to Central Silk Board for facilities. The authors acknowledge financial support received from Department of Biotechnology (DBT), Government of India, New Delhi, in the form of research projects to ARP (BT/ PR11722/PBD/19/197/2008 dated 11/6/2009 and BT/PR6355/PBD/19/236/2012 dated 08/01/2013). AJ and PM are supported by research fellowship from the projects.