The Mosquito Anopheles Gambiae

Published Date: 02 Nov 2017

ABSTRACT

Protein-protein interactions (PPIs) orchestrating fundamental biological processes can be interrogated in silico and experientially validated. Disruption of PPIs among male and female reproduction associated proteins in Anopheles gambiae can impair reproduction in vector and affect their population dynamics. Investigations were conducted to characterize PPIs in mating plug associated proteins and their influence on mating phenotype in the mosquito. Mating plug protein associated orthologs in Drosophila melanogaster were identified in STRING database confirmed using and Artemis Comparison Tool (ACT). Interacting networks were identified and interrogated using Cytoscape routine software and Reactome database, and maximum likelihood (ML) phylogeny among the proteins identified and mapped on the network using Geneious software. Expression profiles of the genes were established using qRT-PCR. Forty three orthologs were identified with 16 novel proteins predicted as putatively interacting with 27 known mating plug proteins. Within the network, 555 nodes with 2344 edges were identified among which 14 An. gambiae proteins mapped to 17 orthologs in D. melanogaster. Four main complexes were identified on the network. The complex from D. melanogaster was the most prominent. Metabolism of proteins, gene expression, 3’- UTR-mediated translational regulation, regulation of beta cell development, diabetes pathways, signal recognition, and membrane trafficking are gene ontology processes identified in the hub proteins . The 14 An. gambiae interologs identified two clusters for pre-mating and post-mating events. An. gambiae mating plug proteins, Cathepsin B (AGAP004533) and Trypsin-Like Serine protease (AGAP005195) observed in post-mating cluster were co-expressed. Prominent interactions in mating related proteins in An. gambiae observed are putative candidates for manipulation of reproduction phenotype in the mosquito.

Key words: Protein-protein interaction, Anopheles gambiae, Drosophilia melanogaster, vector control, biological network.

BACKGROUND

The mosquito Anopheles gambiae is the main vector of malaria, a disease that kills over a million people each year. The development of effective tools for controlling vector populations is of paramount importance [1]. The overarching goal of malaria vector control is to reduce the vectorial capacity of local vector populations below the critical threshold needed to achieve a malaria reproduction rate (R0, the expected number of human cases that arise from each human case in a population) of less than 1. Anopheles gambiae sensu stricto (A. gambiae), is the type species of the Anopheles gambiae sensu lato complex: a group of seven closely related African species morphologically indistinguishable as adults [2] and incompletely reproductively isolated from one another (hybrid females are fertile) [2,3]. As in many other mosquitoes, mating in Anopheles gambiae Giles occurs primarily in swarms [4, 5, 6, 7, 8, 9], but also indoors [10]. After sunset, males form swarms that typically consist of 20–300 males flying in a cloud of ≈1 m diameter [11] for 20–30 min. Females fly into swarms and often are observed leaving in copula. An. gambiae females usually mate once [12]. Females produce and lay egg batches of ≈150 eggs [13], in cycle of 3 d after every blood meal [14, 15].

Molecular studies done by Rogers et al [1] showed that, in the malaria mosquito Anopheles gambiae, seminal secretions produced by the male accessory glands (MAGs) are transferred to females in the form of a coagulated mass called the mating plug. Mass Spectrometric analysis identified 15 MAG-expressed proteins that are transferred to females as part of the mating plug. Two of these proteins, the MAG-specific transglutaminase AGAP009099 and its glutamine-rich substrate Plugin, are responsible for the coagulation of the liquid secretions of the MAGs into a solid mass. Some of the other MAG-proteins from the plug, particularly the three small Acp-like proteins described AGAP009362, AGAP009370, and AGAP012830, could represent important modulators of female behavioural responses to copulation, such as a reduced receptivity to further mating and induced oviposition [16]. The identification of a number of female proteins, mainly proteases, associated with the mating plug suggests a direct interaction between male and female proteins that may be important for plug processing. Indeed two of the female proteases identified on the plug (AGAP005194 and AGAP005195) were shown previously to be expressed exclusively in the atrium of virgin females and were considerably down-regulated at 24 h after mating [17].

Studies on the molecular basis of the female response to mating in insects have so far been mainly limited to Drosophila melanogaster, where postcopulatory changes in gene expression are generally of small scale (<2-fold) [18, 19, 20, 21]. In the fruit fly D. melanogaster, post mating responses are modulated by components of the seminal fluid, produced primarily by the male accessory glands (MAGs) and transferred to females during mating. Besides stimulating egg laying and triggering reduced sexual receptivity, MAG secretions have been shown to induce expression of immune peptides and reduction of female lifespan [22, 23, 24]. However, assigning specific functions to these seminal fluid proteins has proved to be difficult. Even in D. melanogaster, where genetic tools are well developed, only a handful of the >100 secreted Acps identified in the MAGs have been functionally characterized [25]. Work done on two MAGs proteins (transglutaminase and plugin) showed that 15% of females mated to males in which AGAP009099 was knocked out did not receive a mating plug, compared to 1.8% females mated to males injected with LacZ as controls [1]. These findings confirms the putative functional role of AGAP009099 (Transglutaminase) enzyme in plug formation given that some LacZ females equally never received the plug. Molecular analysis have proven to be very slow in determining MAGs protein functions given that only the putative function of AGAP009099 is known in mating plug formation in Anopheles gambiae though the vast number of plug proteins identified.

The past failures of releasing sterilized males [26, 27] and a growing awareness of the importance of understanding male reproductive success have led to repeated calls for addressing this neglected area in mosquito biology [28, 29]. Of particular concern is our lack of knowledge about factors and pathways ensuring male reproductive success, such as those that result in sperm storage, oviposition, and the inhibition of remating in females. The past decade has seen phenomenal advances in Anopheles genomics and proteomics [30]. These advances, coupled with the visionary but technically challenging development of mosquito transgenics and other genetic manipulation techniques, open up the possibility of developing novel technologies to suppress mosquito populations or to make parasite-refractory mosquitoes, and make mosquito-based transmission-blocking technologies possible. The availability of large amount of biological data mostly stored in publicly accessible databases has made it possible for researchers to begin constructing protein-protein interaction maps (interactomes) geared towards understanding organisms at the system level. Computational techniques, with the ability to combine the large and heterogeneous data sets, are critical to the elucidation of more biological insights [1]. The computational methods are meant to augment high throughput experimental methods that have successfully been used in determining several interactomes.

We established that In silico protein-protein binding interactions on An. gambiae mating plug protein orthologs in D. melanogaster can predict molecular interactions potentially favoring plug formation in the male mosquito. The hubbal proteins predicted main MAGs proteins (AGAP009368) playing putative important roles in plug formation and its putative downstream interaction with the main female Trypsin-Like Serine Protease (AGAP005195) known to digest the mating plug and expressed mainly in the female atria of An. gambiae. The clustering of proteins based on ML Phylogeny was tested using qRT-PCR which confirmed co-expression of these female proteases (AGAP005195) and Cathepsin B (AGAP004533) in the female atria suggesting their co-expression and co-functionality in post-mating processes within the female mosquito. These findings demonstrate that the use of protein-protein binding soft-wares in identifying favorable interactions which could be confirmed with in-vitro laboratory analysis. Defining the functions of these MAGs proteins will help in the better understanding of molecular pathways involving these proteins in mating plug formation and hence they could serve as targets for vector reproductive control of An. gambiae mosquito.

RESULTS

Orthologs and synteny of An. gambiae mating plug genes in D. melanogaster

Twenty seven D. melanogaster mating plug protein orthologs were identified in An. gambiae, among which 16 were present in the STRING database (Table 1). Among the orthologs, about 35, 30, 21, 12 and 2 % were present in 2L, 3R, 2R, 3L and X Chromosomal arms respectively of the mosquito, and about 22 and 56 % were female and male specific respectively (Table 1). Male specific proteases from MAGs were identified. AGAP013150 and NOVEL Zinc Carboxypeptidase (ZCP7) proteins were homologous to AGAP004671 and AGAP008071 proteins respectively. NOVEL accessory gland protein (ACP1) has no homolog. Prediction scores for the orthologs in the STRING database ranged from 40-1311(Table 1). Sequence comparison of identified orthologs from the STRING database showed about 24, 12 and 65 % hit matches (M), mismatches (MM) and no matches (NM) respectively of the orthologs (Table 1). The matches had one-to-one coverage of 79- 97% identity.

Network of An. gambiae mating plug proteins orthologs in D. melanogaster

Alignments and scores of the predicted protein orthologs of D. melanogaster from STRING database predicted a network in Cytoscape of 555 nodes with 2344 edges as summarised in Figure 1. Degree (number of interactions per node) of distribution were 1(150 nodes) to 52(25 nodes). The nodes clustered with coefficients 0.0 (150 nodes) to 1.0 (225 nodes). ClusterViz predicted 14 complexes with 238/400, 78/227, 53/1287, 38/160, 31/57, 30/72, 16/16, 15/21, 14/11, 10/41, 10/10, 10/14, 9/17 and 7/12 nodes/edges in descending order. These complexes observed in D. melanogaster, Campylobacter jejuni, Arabidopsis thaliana and Homo sapiens. Drosophila melanogaster exclusively comprised complexes 1, 8, 11 and 13, among which complex 1 most prominent. The complexes were associated with various biological and functional processes (supporting material, Table S1). Fifty hubs (nodes) were predicted in the network (Figure 2) as key controllers inside biochemical pathways or complexes. There were 3 - 30 degrees of interactions per node (Figure 2). The AGAP004533 (Q9VY87) interologs were identified on the mating plug wherein AGAP007120 (NDKA), AGAP009623 (G3P2), AGAP006818 (RIR2) were identified as string interactiong with AGAP009584 on the mating plug. The most prominent hub (Q8SX59) in the hubbal interactions (Figure 2) had no interolog in An. gambiae but shared similar structural protein properties with plugin (AGAP009368) protein in the mating plug (Table 1). Accessory gland protein (Acp29AB) (galactose binding, GO: 0005534) (234 amino acid) was attached to a hubbal proteins (Q8WS79). Q9VEM7 in D. melanogaster had interolog AGAP005195 in An. gambiae , and was the only female protein present in the network. The Q9VEM7 was not a hub protein but was terminally linked to Q9VIT3 a hub nodal protein (Figure 4).

Predicted secondary protein structure of predicted An. gambiae mating plug proteins

Results of D. melanogaster, Q8SX59 protein secondary structure properties analyzed comparatively to plugin (AGAP009368) are summarized in Table (4). The D. melanogaster protein had 317 residues, with about 20% of the residues, vs 10.41% in plugin buried while for residues exposed with more than 16% of their surface, Q8SX59 had 80.76% against 89.59% for plugin. The protein contains about 4, 3 and 93 % helices, strands and loops respectively. Most of the proteins had low complexity regions (LCRs) of 13-317 residues. Plugin LCRs were 1-260 residues. The Q8SX59 (CG9083) protein appeared to be non- globular, lacking DNA binding properties binding to other protein and localized is nuclear (Table 2). PROSITE predicted sequence motifs on this protein were Casein kinase II (CK2-phospho) and N-myristoylation site (Myristyl). The Gene Ontology terms (Table 3) indicate molecular functions; GO: 0016829, is associated with catalyses of C-C, C-O and C-N bonds cleavages and GO: 0005509 which interacts selectively and non-covalently with Calcium ions. The predicted GO: 0009987 was responsible for cell growth and/or maintenance and GO: 0008152 was involved in metabolic processes leading to cell growth. The protein Q9VEM7 (243 amino acids) in D. melanogaster in relation to AGAP005195 (243 amino acids) in the female An. gambiae (Figure 3) had about 37% identity and shared: trypsine like and Gamma-aminobutyric acid (GABA) similar domains.

Biological process of Anopheles gambiae mating plug protein orthologs in Drosophila melanogaster

Among four identified complexes involving D. melanogaster proteins in Reactome database, the first one (complex 1) identified 113 processes for D. melanogaster and 28 proteins in the network involved in these processes (supporting material, Table S1). The processes were associated with membrane trafficking (1.1e-05), metabolism of proteins (3.0e-04), metabolism of amino acids and derivatives (4.5e-01), signalling by insulin receptor (2.1e-01), axon guidance (7.1e-01), regulatory RNA pathways (3.8e-02). The second one (Complex 8) was associated with six processes involving eight genes. These six procceses were associated with hemostasis (2.5e-05) as the main process and this involved two D. melanogaster orthologs Q9U1I4 and Q9VFC2 mapped to AGAP003139 as interolog in An. gambiae. The orthologs were identified as Serine protease inhibitors involved in platelet plug formation and degradation (supporting material, Table 1S). AGAP003139 protein is expressed in both males and females but more significantly in the latter (VectorBase.org). Complex 11 and 13 showed no identified processes for further analysis. The 50 hubbal proteins also identified 49 biological processes out of the 2021 known processes identified. Twelve proteins were involved in these processes which were: metabolism of proteins (8.8e-13), gene expression 2.0e-06, 3’- UTR-mediated translational regulation (7.7e-08), regulation of beta cell development (1.3e-06), diabetes pathways (6.8e-06), signal recognition (Preprolactin) (5.0e-07), and membrane trafficking (1.3e-03) (supporting material, Table S1).

Gene expression and pathways associated with Anopheles gambiae mating plug orthologs in Drosophila melanogaster

The D. melanogaster proteins identified on a sub-network within the main network were putatively expressed in ovary, mated spermatheca, male accessory glands and testes reproductive tissues. Network indicating interactions between Q8SX59, Q9VEM7 and ACP29AB and their neighbours is summarized in Figure 4. Putative plugin protein (Q8SX59) was identified in the sub-network as a major protein which interacts with Ribonucleoside-diphosphate reductase subunit M2 (RIR2), an interolologous string to AGAP009584 (Table 1), is involved in oxidation reduction process in cells (GO: 0055114) and is expressed in the ovary. RIR2 showed interactions to Q9VCC9 and Q9U9A9 proteins identified in complex 1(supporting material, Table S1) which are both involved in the target of rapamycin pathway (GO: 0032007) which initiates lipid transport molecules for egg development. The D. melanogaster ortholog Q9VEM7 was present on the right arm of figure 4. This protein is linked to Q9VIT3 expressed both in the ovary and in the testes and has as known function protein binding (GO: 0005515).

Acp29AB, a MAG protein in D. melanogaster, was present on the left arm of Figure 4 beginning at RIR2 and was putatively associated with post mating behaviour in the female (GO: 0007617). The details of the predicted functional interactions can be seen in Table 3. The Acp29AB appear to interact with several proteins in STRING database including Acp26Aa (ovulation, GO: 0030728) and Acp76A (serpin, GO: 0004867) proteins. On the sub-network, Acp29AB interacts with Q8WS79 (ERR) expressed in the ovary and in the spermatheca. This is one of the hubbal proteins on the plug network and identified in the generic transcription pathway (supporting material, Table 1S). It is a steroid hormone receptor (GO: 0003707) and involved in transcription regulation (GO: 0045449). Q8WS79 combines with a steroid hormone to initiate a change in cell activity. Two proteins (Q9VTX7 and Q7YSG8) expressed in the MAGs were observed terminally on the left arm of Figure 4 and seen to be involved in imaginal disc development (GO: 00007444) and multicellular organismal process (GO: 0032501) respectively. Acp29AB probably interacts with Protein similar (SIMA) expressed in the MAGs and in the ovary via Q8WS79 in the spermatheca. It has predicted signal transducer activity (GO: 0004871) and transcription regulation (GO: 0045449). The signal transducer role could aid Q8WS79 in changes in cell function. SIMA and Q8WS79 interact with the Q9VAV6 which is expressed in the testes. Q9VAV6 has as function histone acetyl transferase activity GO: 0004402 possibly regulating gene expression. It is linked to two other proteins Q8IP72 and Q9VJY9 (Supporting material, Table S1) involved in pre-microRNA processing (GO: 0031054) and female germline-sterm cell division (GO: 0048132). This protein interacts with IF4A which is a eukaryotic initiation factor involved in protein biosynthesis (GO: 0006412) with helicase activity (GO; 0004386) for RNA transcription and splicing.

Molecular phylogeny of An. gambiae mating plug proteins with orthologs in D. melanogaster

Fourteen An. gambiae interologs were mapped to 17 D. melanogaster orthologs on the network. The Phylogenetic models and their predicted values used for comparative analysis between D. melanogaster and An. gambiae mating plug proteins interologs can be seen on Table 4.The WAG tree was selected as the best tree based on; ML ratio test for the two methods was X2 = 833.47 and P<0.001, small tree size, least evolutionary change based on its small branch support values, parsimony and gamma values (Figure 5). Based on this, MAFFT alignment on the An. gambiae interolog proteins were subjected to the WAG maximum likelihood model as well (Figure 6). The topologies indicate that most proteins were maintained in their branch groups though rooted differently (Figures 5, 6). The most prominent difference was related to that of putative plugin protein (Q8SX59) in the network which grouped with transglutaminase (Q9VLU2 and Q8IPH0) in the Drosophila tree but both are apart in the gambiae tree. In the gambiae tree transglutaminase is closer to AGAP001649 (Q8IMY3 and A4V4A3) and AGAP009673 (Q7KTY3). The former is a glucosylceramidase (GO: 0008152) involved in glucosidase activity which trims mannosidases involved in N-linked glycoprotein biosynthesis probably activating the transglutaminase enzyme meanwhile the latter is an acyltransferase (GO: 0008415) associated with acylation of proteins. Both protein functions are linked to prosite motifs present in transglutaminase. The Transglutaminase and AGAP001649 proteins were MAGs specific. Plugin clusterd with female protein AGAP005195 (Q9VEM7) in the gambiae tree. The proteins AGAP008276 and AGAP008277 were clustered with AGAP005195 in An. gambiae tree and shared the same GO terms (GO: 0004252). The weak aLRT branch support (1.24x10-4) separating plugin (blue star), AGAP004533, AGAP005195 (red star), AGAP008276 and AGAP008277 on the tree topology (Figure 6) splits the tree into two clusters.

Co-expression of AGAP004533 and AGAP005195 on the Anopheles gambiae postmating cluster

The qRT-PCR analyses revealed that the expression levels of AGAP005195 in virgin females of An. gambiae were down-regulated 24hr post mating, The down-regulation of the gene after 6hrs post-mating was statistically significant for both AGAP005195 (P<0.001) and for AGAP004533 (P<0.01) using Bonferonni posttests. The difference between the two transcripts was not statistically significant at these time points (P=0.0147). AGAP004533 and AGAP005195 had similar expression levels in the virgin females and both were down-regulated 48hrs post-mating and maintained up to 10 days post mating (Figure 7). The expression levels between the transcripts across the various time periods were not statistically significant using Bonferonni posttests.

DISCUSSION

From these results 43 plug proteins in total were obtained with 16 putative novel proteins as strings to the 27 known proteins? Which imply that some MAGs proteins are yet to be identified. Dottorini et al [16] identified 46 MAG proteins in An. gambiae with 25 putative orthologs in D. melanogaster. The number of putative Acps in D. melanogaster is 83 [31], with less than 20 with extensive experimental support [32, 33, 34, 35, 36]. Residues at interaction surfaces determine recognition specificity required to correctly route signals. As a consequence of gene amplification in the same cell, the unique function of each two-component system requires evolution to maintain strict monogamous interactions between protein partners [37] therefore the 16 proteins identified as strings to those established proteins on the mating plug probably play roles as reproductive proteins since they interact hence co-evolve. The majority of proteins identified by Rogers et al. [1] originated from Chromosome 3R but with the added strings the majority now comes from chromosome 3R and 2L. According to Stump et al [38], proteins in the genomic areas of chromosome 2L and X play important roles in reproductive isolations of the various mosquito forms at embryonic levels. MAGs proteins show higher rates of protein divergence [39, 40, 41, 42, 43, 44, 45] and protein polymorphism [46, 44] compared to "average" proteins in D. melanogaster and D. simulans [44]. This shows that proteins on chromosomes 3R and 2L could play important roles in reproductive variation between Anopheles mosquito species.

The results obtained here showed low proportion of matches to orthologs predicted from STRING database despite their importance in establishing close evolutionary relationship between An. gambiae and D. melanogaster species as evidenced by percentage identity used. Almost half of the genes in An. gambiae and D. melanogaster genomes (are interpreted as orthologs with an average sequence identity of about 56%, slightly lower than that observed between the orthologs of the pufferfish and human (diverged about 450 million years ago). This indicates that these two insects diverged considerably faster than vertebrates [47]. The number of proteins that matched validated STRING database predictions for subsequent application. Matches in ACT software are uniform throughout without gaps within the protein alignments. This implies that those proteins predicted by the STRING database which did not match had some few gaps within the alignment though with good score predictions. We could also rely on the protein-structure-function paradigm which requires a protein to fold into its 3D structure for proper function. The STRING database predicts orthologs based on different parameters including protein alignments which play a great role in homology modelling for proteins. Homology cut-off is a strongly varying function of alignment length up to a length of about 70-80 residues. For example, for alignment length 30, sequence similarity has to be at least 43% (gaps allowed with a gap opening penalty of three residue identities) to infer structural homology. For very long alignment lengths 25% sequence identity is sufficient. Note that below these values of sequence similarity structural homology cannot be asserted or excluded given that the region of weaker sequence similarity is a "don’t know" region [48]. A few matches were inversions which could be an adaptive evolutionary advantage in An. gambiae.

The PPI identified in the network revealed a vast network given the number of nodes and edges. Some proteins had just one interaction while others had multiple interactions and mostly seen among hubs. The nature of interactions observed showed that the An. gambiae orthologs in D. melanogaster will drive similar cellular processes hence confirming the close evolution between these two species. The network was built on a domain interaction landscape. Protein interactions are largely mediated by interactions between structural domains [49]. To achieve this, they usually search for (i) small distances between atoms in interacting domains, (ii) specific bonding patterns between domains, and (iii) conservation of interacting faces. Once a conserved 3D structural interface between two domains is determined and defined, the structural domain-domain interactions (sddi) databases can use it to predict all the protein pairs where such structural domain pair is present. In this way, the sddi datasets allow to infer all the putative protein interacting pairs that include structural domains that have been shown to interact in at least one Protein Data Bank (PDB). The 14 complexes identified different biological processes and the largest was that for D. melanogaster which was not surprising because the input proteins were all from this species. The other complexes were due to application of protein names common across different taxa, hence identifying PPIs taking place in these organisms Cytohubber using EPC predicted the top 50 hubs (nodes) in the network (Figure 2) which is a way to decipher key controllers inside biochemical pathways or complex networks [50]. We hypothesise that these hub interologs (An. gambiae mating plug proteins identified on the network) are equally key proteins in the reproductive pathway in An. gambiae because interologs like AGAP004533 (Q9VY87) were present on the mating plug with AGAP007120 (NDKA), AGAP009623 (G3P2), AGAP006818 (RIR2) as string interactions to AGAP009584 on the mating plug. The top most ranked hub (Q8SX59) in the hubbal interactions (Figure 2) had no interolog in An. gambiae but shared similar structural protein properties with plugin (AGAP009368) protein which is found on the mating plug (Table 1) and believed to be very important in the formation of this plug [1]. The presence of Acp29AB (234 amino acid) which is a MAG specific protein [36] in D. melanogster and secreted only during mating confirmed our network as a possible reproductive process. Q9VEM7 in D. melanogaster having as interolog AGAP005195 in An. gambiae was the only female protein found on the network. It was not a hub protein but was seen terminally linked to Q9VIT3 a hub nodal protein (Figure 4). AGAP005195 was terminal in the network which could confirm the fact that the plug is formed in the male and transferred to the female during mating [1].

Based on the comparative analysis of plugin and Q8SX59 from our results, they showed common secondary structure properties which shows they could perform similar molecular functions within the cell. PROSITE predicted sequence motifs on this protein were Casein kinase II (CK2-phospho), [51] and N-myristoylation site (Myristyl) [52] showed that they possess phospho kinase activities required for binding interactions and N-myristoylation which is an acylation process absolutely specific to the N-terminal amino acid glycine in proteins. The protein property predictions for Q8SX59 were important because they looked very close to those of Plugin in An. gambiae (Table 2). In Q8SX59 protein large clusters of Glycine residues were observed within the protein accounting for the LCRs. Such regions in proteins deprive them of a fixed 3D structure hence obeying the protein-trinity paradigm which states that, the native protein can exist in any of the three thermodynamic states-ordered, molten and globule [53] making them very flexible. The flexible nature of regions lacking well-defined folding structures is thought to be responsible for their versatile binding capabilities; this flexibility could allow these regions to bind several different targets [54]. Ekman and co-workers [55] noted that highly connected 'hub' proteins contain an increased fraction with LCRs compared to non-hub proteins. Both proteins share similar predicted molecular properties (Table 2) and also lack close orthologs in both species which can be seen by the spurious score for Plugin (Table 1) predicted by the STRING database. The comparative analysis also showed GABAA receptor domains common to AGAP005195 and Q9VEM7 and these domains have been found in the spermatozoon plasma membrane and GABA is known to stimulate hyperactivation and promotes the acrosome reaction [56]. The identified Interpro domain in both proteins makes them functionally similar hence putatively participate in similar roles on the predicted network.

The 50 hubs identified from the network as seen in the results identified more processes as a complex as compared to the other individual complexes found. This confirms the role of these hubs as key controllers of the network and could be used to globally understand the various processes taking place within reproductive proteins in An. gambiae and D. melanogaster. The 50 hub proteins put together included all the processes identified in the first complex with a wider coverage of pathways on the network. Protein metabolism and translational regulation which were identified processes are aspects of the two main proteins: transglutaminase (AGAP009099) and Plugin (AGAP009369) (Table 1), involved in mating plug formation in Anopheles gambiae [1]. Signal recognition and membrane trafficking are processes required for secreted proteins from male and female reproductive tissues. Gene expression process occurs before and after mating in male and female An. gambiae reproductive tissues and a wide variation of genes up and down-regulated in the female atria was identified by Rogers et al [17]. Hemostasis identified from complex 6 on the network is related to plug formation and degradation within the female hence confirming a plug formation related process on the network.

The sub-network based on key reproductive proteins in An. gambiae (plugin and AGAP005195) and D. melanogaster (Acp29AB) showed that all the proteins were expressed in reproductive tissues except the orthologs of plugin (Q8SX59) and AGAP005195 (Q9VEM7) which putatively plays specific roles in males and females of An. gambiae respectively. This could show points of reproductive divergence between both species. The structural homology between plugin and Q8SX59 could infer it as the main protein in the male mosquito’s reproductive process. The possible interaction of Q8SX59 and RIR2 would be a phosphorylation reaction involved in An. gambiae. The (GO: 0055114) function identified on RIR2 is similar to Pyridine nucleotide-disulphide oxidoreductase (PF00070) which is a PFam domain found on AGAP009099 (Transglutainase), a protein involved in the cross linking of Plugin for plug formation hence favoring its interaction with our putative plugin protein. Using CLC main work bench (V5.7.1), PFam domain predictions for plugin showed a GGDEF domain (PF00990) which is a sensory receptor domain and a Histidine Kinase domain (PF00512) involved in a two component transduction system known in bacteria. Both domains need a response regulator in order to function properly. This domain (PF00072) was found on AGAP005195 which is the only secreted female interolog to Q9VEM7 found on the network. This domain receives the signal from the sensor partner in a bacterial two-component system. The domain similarity between AGAP005195 and Q9VEM7 and also the proteins with which it clusters could mean it is an interaction between male and female reproductive process. The possible location of this interaction should be in the female atria because all the terminal proteins (Q7JY68, Q9V9M7, Q9VEM7, Q9W3L4, IF5A, and R10AB) carried by Q9VIT3 are expressed specifically in the testes, male accessory gland and the mated spermatheca. IF5A is a translation elongation factor (GO: 0003746) and a translation initiation factor (GO: 0003743). It is the only eukaryotic protein to have a hypusine residue which is a post translational modification of a lysine by the addition of butyl amino [57] group from spermidine found in semen. Q9V9M7 is a ribonucleoprotein (GO: 0030529) and involved in translation.

Presence of Acp29AB identified on the sub-network was critical because MAGs proteins are secreted only during mating in D. melanogaster implying they are reproductive specific hence it could play its function of sperm competition [58] due to its interaction between Q8WS79 expressed in the ovary, SIMA expressed in the MAGs and Q9VAV6 expressed in the testes. The presence of Acp29AB could imply that the predicted An. gambiae orthologs in D. melanogaster will interact with it a particular manner to drive similar reactions in both species therefore more work still needs to be done in identifying ACPs in An. gambiae. Acp29AB from our results was seen to interact with Acp26Aa and Acp76A in the STRING database. The former codes for a male accessory gland peptide that stimulates egg laying in mated D. melanogaster females during the first post-mating day meanwhile the later which contains a serpin signature plays a role in the observed regulation of Acp proteolysis and/or in the coagulation of seminal fluid to form a mating plug. Interologs AGAP003139 and AGAP009212 on the An. gambiae mating plug showed similar serpin signatures and they are specific to the MAGs. The small information on ACPs in An. gambiae remains a great center of concern.

The compared values from the WAG ML analysis for An. gambiae proteins and their interologs in D. melanogaster confirmed the closeness of these two species given that their diversion is a 250 million years ago which is recent and their related PPIs will perform similar biological processes within related cells. The substitution rates were similar and implying both species could have similar evolutionary patterns hence directing similar biological processes in both organisms. The difference observed was the tree topology which rooted branches differently. This difference resulted from difference in percentage identity in multiple sequence alignments (0.8% for gambiae and 0.0% for drosophila) and percentage pairwise identities from alignments (5.6% for gambiae and 7.6% for drosophila). Comparing the tree topologies in both species revealed that most proteins were maintained in their branch groups though rooted differently. The most interesting difference was related to that of plugin which clustered with AGAP005195 in the post-mating cluster. As seen above, both proteins are expected to be involved in a 2 response regulatory system of phosphorylation reactions. AGAP005195 possesses both Trypsin-Histidine and Trypsin-Serine bindind sites confirming it as a Serine Protease and account for the specificity of AGAP005195 on the plug given that it was the only female specific serine protease identified on the plug from the two known (AGAP005194 and AGAP005195) to be down-regulated after mating. The placement of plugin protein and AGAP004533 (Q9VY87) [a thiol protease (GO: 0008234)] on the tree shows their independence to the related proteins given their vast interactions on the network [plugin (Q8SX59) as top node (15 interactions) and AGAP004533 as 8th (30 interactions)]. AGAP004533 is a cysteine type endopeptidase (GO: 0004197) and this site is common to both transglutaminase and plugin proteins therefore it can be hypothesized that it could be involved in plug digestion within the female. The weak aLRT branch support (1.24x10-4) separating plugin (blue star), AGAP004533, AGAP005195 (red star) (Figure 6), AGAP008276 and AGAP008277 on the tree topology should imply a split of the tree in two parts, possibly those proteins involved in pre-mating grouped with Transglutaminase meanwhile Post-mating proteins are grouped with Plugin. The identified mating plug proteins in An. gambiae identified on the network showed no direct interaction between them implying more An. gambiae proteins involved in plug formation still need to be identified.

The similar expression patterns between AGAP005195 and AGAP004533 potentially imply co-expression and possibly co-regulation of the genes, given that both genes revealed relatively similar expression levels. Grigoriev [59] analyzed physical interactions in yeast and observed that proteins encoded by co-expressed genes interact with each other more frequently than with random pairs. We think one of the genes should affect the expression of the other given that both are proteases (AGAP004533 (Cathepsin B) and AGAP005195 (Trypsin-Serine Protease)) are expressed in the female atria [1] or they could be co-regulated together. However, investigations in yeast demonstrated poor correlation between co-expression and co-regulation [60, 61] suggesting a substantial number of co-expressed genes might not be co-regulated. Further investigations on the expression patterns on female expressed genes during mating will provide vital targets for functional analysis as a way of understanding these genes for use in reproductive control of the vector.

CONCLUSION

This work is the first one ever done to understand PPIs on Anopheles gambiae proteins since their identification by Rogers et al [1]. The 16 Strings to the already identified proteins found on the mating plug and some of these proteins with good alignment scores in the STRING database play important roles in protein function on the network. The main hubs identified so far remain the main focus for functional analyses as these proteins form backbones in networks. Q8SX59 identified as the main hub in the network having similar properties to Plugin protein (AGAP009368) makes plugin a putative key protein in the reproductive process in An. gambiae. Further to that its putative identified PPI with Trypsin-Like Serine Protease (AGAP005195) could imply a primary interation driving post-mating physiological processes in the female mosquito.The predicted network showed clusters of complexes driving important processes in D. melanogaster and this could just be similar in An. gambiae. An understanding of interaction pathways for plug formation and post-mating events in the female, we could unveil important interactors for functional analysis through RNAi. The identified co-expression between AGAP004533 and AGAP005195 provides insides to co-regulated genes which can be controlled at transcription promoter regions on the chromosome. The role played by Anopheles gambiae mosquito especially in Africa remains a great area of research and looking for more efficient reproductory vector control targets is of great interest for the control of malaria.

MATERIALS AND METHODS

Anopheles gambiae mating Plug proteins ortholog predictions in Drosophila melanogaster

Accession numbers for An. gambiae plug proteins (27 in number) were obtained from Rogers et al. [1]. For the proteins without accession numbers, the whole sequence of the Chromosome arm was downloaded from Vectorbase. The sequence was then imported into the Genomics Workbench version 5.5 (CLC Bio, Aarhus, Denmark) wherein the regions were extracted and blasted against the NCBI’s non redundant (nr) database (http://www.ncbi.nlm.nih.gov). The 27 An. gambiae mating plug proteins were used in the Search Tool for the Retrieval of Interacting Genes (STRING) [62] database in 2010 to predict their their orthologs in D. melanogaster.

Chromosome synteny network design and analysis of pathway in Anopheles gambiae mating plug proteins to Drosophila melanogaster

Direct orthologs of An. gambiae in D. melanogaster predicted with the STRING database were used in Artemis Comparison Tool (ACT) (V.9.0) with a filter > 50% [63] to confirm related proteins. The accession numbers for the identified direct D. melanogaster orthologs in STRINGdb and indirect orthologs from UniProtKB for the An. gambiae proteins were used in the Uniprot protein knowledgebase (UniProtKB) [64] database to obtain the Protein IDs . The IDs were applied in Agile Protein Interaction DataAnalyzer (APID2NET) plugin hosted in Cytoscape (V.2.8.0) [65] to build the PPI network. Various complexes within the network were identified using the ClusterViz plugin [66]. The main hubs (key nodes) within the network were detected using Cyto-Hubba plugin [67] through the Edge Percolation Component (EPC) algorithm [68]. The Pathways in which the complexes were involved were identified using Reactome database [69]. Proteins within a sub-pathway were probed for putative expressions in reproductive tissues using FlyATLAS http://flyatlas.org/ [70] and FlyBase [71] for D. melanogaster.

Protein Secondary Structure Prediction and maximum likelihood phylogeny

The orthologs of An. gambiae plug proteins in D. melanogaster via Genious version 5.5 [72] were subjected to a Multiple Sequence Alignment using Multiple Alignment Program for amino acid or nucleotide sequences (MAFFT) routine version 6.814b [73]. The alignments were subjected to a maximum likelihood phylogeny (PhyML) based on nearest neighbour interchanges (NNI) to explore space of tree topologies starting from a fast distance-based tree [74]. Reversible mitochondrial substitution model (mtREV) [75] and Whelan and Goldman (WAG) [76] amino acid substitution models on D. melanogaster orthologs were evaluated. The best identified model for the tree was used to build the tree with An. gambiae interologs.

qRT-PCR analysis on AGAP005195 and AGAP004533 and their putative effect on post-mating behavior in An gambiae

RNA extraction, cDNA synthesis, and SYBR-green based qRT-PCR was performed as described previously [17] using the primers AGAP005195 (Fwd: CGCATCGATCGTGCTATAGC and Rev: AAGTAGTCCAACATCGTCACGAAA) and AGAP004533 (Fwd: CCTGGAGCTATTG GGTCCGG and Rev: CCAAGTTGGAACCGAACGGG). The ribosomal protein gene RpL19 was used as an internal control to normalize in An. gambiae (AGAP004422), using previously described primers [17]. The expression analysis was conducted on cDNA from AGAP004533 and AGAP005195 obtained from the female atria of An. gambiae. These genes were specifically selected since AGAP004533 was among the 50 hub proteins identified and shown to be present in the female virgin atria while AGAP005195 was the only female Trypsin protease identified on the plug [1]. Mosquitoes from a laboratory colony of the G3 strain were separated by sex as pupae, raised supplied with sucrose ad libitum in cages. The mosquitoes were mated as previously described [17]. Atria of three days old virgin females and post mated females were dissected and RNA extracted as described previously [17].

Data Analysis

The sequences were subjected to multiple alignments using CLC workbench 6.6.1(CLC Bio, Aarhus, Denmark) and Genious version 5.5 created by Biomatters. A phylogenetic tree was constructed by Neighbor Joining method using p-distance estimates, and tested by using interior-branch test. Amino acid pair-wise identities and similarities scores were used to evaluate amino acid sequences during protein alignments. Evolutionary relationships for the various haplotypes were calculated with the maximum likely hood ratio test (LRT) using PHYML. Orthologs were identified using STRING database with alignment score confirming identity. Artemis Comparison tool was used with a >50% filter to confirm orthology from the STRING database. Gene ontology (GO) values were used to confirm functions of genes mapped on the interactome. Hyper geometric tests were performed in the Reactome database to map network proteins within pathways. Approximate likelihood ratio test (aLRT) was used to identify protein clusters of An. gambiae Plug proteins on a Whelan and Goldman (2001) maximum likelihood (ML) tree. The QRT-PCR data was analyzed using the comparative CT, conducted for three independent experiments. Analysis of Variance (ANOVA) was done using GraphPad Prism version 4.0 for Windows, (GraphPad Software, La Jolla California USA, www.graphpad.com) to evaluate relative differences between expressions levels of various transcripts used.