The phylogenetic analysis of glucose transporters in An. stephensi
Four genes are annotated as glucose transporters (ASTE005839, ASTE003001, ASTE006385 and ASTE008063) in the Vectorbase of An. stephensi (AsteS1.6). To investigate the relationships of these genes between An. stephensi and other organisms, a phylogenetic tree was constructed based on the amino acid sequence of An. gambiae, An. stephensi, Ae. aegypti, D. melanogaster and H. sapiens using the Maximum Likelihood and Bayesian inference phylogenetic analysis (Fig. 1). In H. sapiens, Glut transporters can be divided into three classes: class 1 (GLUT1, GLUT2, GLUT3, GLUT4 and GLUT14); class 2 (GLUT5, GLUT7, GLUT9 and GLUT11); and class 3 (GLUT6, GLUT8, GLUT10 and GLUT12) [31–34]. Due to the high similarity between An. stephensi ASTE005839, D. melanogaster GLUT1 (FBpp0305693) and H. sapiens GLUT1 (NP 006507.2), we named ASTE005839 Asteglut1. ASTE008063 and ASTE006385 were categorized into GLUT- class 3, and therefore, we named them Asteglut3 and Asteglut4, respectively. ASTE003001 was not phylogenetically related to any GLUT class, therefore, we named this Asteglutx (Fig 1).
Expression of Asteglut genes in An. stephensi
To determine the expression pattern of Asteglut genes in An. stephensi. We analyzed the expression levels of these genes by qPCR in the head, salivary glands, midgut, ovary and carcass 24 h before a blood meal, respectively. Asteglut1, Asteglut3 and Asteglut4 were mainly localized in the midgut tissue of An. stephensi (Fig. 2a, c, d). In addition to the midgut, Asteglut1 and Asteglut4 were also expressed in the head and salivary glands (Fig. 2a, d). Asteglutx was distributed in all five tissues (Fig. 2b). We next investigated the influence of parasite infection on the expression in the midgut of the four Asteglut genes. Asteglut genes were differentially regulated by P. berghei 24 h post-infection. (Fig. 2e). Plasmodium berghei infection significantly decreased the expression of Asteglut1 and Asteglut4 (t(14) = 2.585, P = 0.0216; t(14) = 3.001, P = 0.0095), while increased the expression of Asteglutx compared to those in normal blood feeding mosquitoes. No influence on Asteglut3 expression was observed during parasite infection (Fig. 2e, t(14) = 0.343, P = 0.7369).
Knockdown of Asteglut1 facilitates P. berghei infection in An. stephensi
To investigate the role of Asteglut1, Asteglutx, Asteglut3 and Asteglut4 in the capability of An. stephensi to transmit P. berghei, double-stranded RNA (dsRNA)-mediated silencing strategy was employed. The expression levels of Asteglut1, Asteglutx, Asteglut3 and Asteglut4 were examined two days post-dsRNA treatment. The expression levels of these genes were significantly decreased by 57.8%, 40%, 65% and 80% compared to the dsGFP control, respectively (Fig. 3a-d, t(14) = 2.529, P = 0.02; t(14) = 7.024, P < 0.0001; t(14) = 3.184, P = 0.0002; t(14)= 3.997, P = 0.0013). However, only knockdown of Asteglut1 significantly increased oocyst number of P. berghei (Fig. 3e, U = 597, P = 0.0067). The dsAsteglutx, dsAsteglut3 and dsAsteglut4 treatments had no apparent effect on the intensity of P. berghei infection (Fig. 3f-h, U = 746, P = 0.3778; U = 762, P = 0.4748; U = 685, P = 0.3542). No significant difference of infection prevalence was observed between dsGFP and any dsAsteglut treated mosquitoes (Fig 3e-h). We next analyzed the knockdown specificity of Asteglut1 and found this gene was specifically knocked-down (Fig. 3i). Thus, the increasing susceptibility of An. stephensi to P. berghei infection was due to the knockdown of Asteglut1, instead of the compensatory effect of other Astegluts (Fig 3i).
Knockdown of Asteglut1 significantly elevates the glucose level in the mosquito midgut
We next analyzed the influence of Asteglut1 on sugar transportation in An. stephensi. The glucose and trehalose levels in the midgut and hemolymph of dsRNA-treated mosquitoes were examined. The glucose level of the Asteglut1-knockdown group was significantly higher than that in dsGFP controls 24 h prior to blood-feeding (Fig. 4a, t(8) = 4.374, P = 0.0024). However, its level in hemolymph is comparable to that in the dsGFP control (Fig. 4c). There was no significant difference between sugar levels in the midgut or hemolymph either just before (0 h) or 24 h post-blood-feeding (Fig. 4, for statistics details, see Additional file 4: Text S1). In addition, knockdown of Asteglut1 did not change the level of trehalose in the midgut or in hemolymph (Fig. 4b, t(8) = 1.299, P = 0.2302; t(8) = 0.146, P = 0.8875; t(8) = 1.752, P = 0.1180; Fig. 4d, t(8) = 0.3585, P = 0.7292; t(8) = 0.1686, P = 0.8703; t(8) = 0.4252, P = 0.6820). Thus, Asteglut1 might play a role in transportation of glucose, but not trehalose in the mosquito midgut.
Transcriptional analysis of Asteglut1-knockdown mosquitoes
To explore how Asteglut1 regulated P. berghei infection, we performed a transcriptome analysis of the mosquito’s midgut treated with dsAsteglut1 and dsGFP 24 h post-blood meal, respectively. A total of 6 G PE clean sequences were generated by the Illumina HiSeq ´10 (Additional file 1: Table S1). Principal components analysis (PCA) showed a clear separation between dsAsteglut1 and dsGFP treatments (Additional file 2: Figure S1). The Venn diagram shows that the expression of 10,240 genes was overlapped in the two groups (Fig. 5a). A total of 46 genes were differentially expressed (Fig. 5b, Additional file 3: Table S2) with 26 upregulated and 20 downregulated genes. These differentially expressed genes belong to multiple functional clusters, including cytoskeletal and structural, immunity, metabolism, proteolysis, redox, transport and those of unknown function (Fig. 5c).
Among the ‘redox’ functional cluster, five genes encoding cytochrome P450 (CYP450) were upregulated, indicating that detoxification was activated in mosquitoes [35]. The gene encoding peroxiredoxin that controls cytokine-induced peroxide levels in mammalian cells was also significantly upregulated, but the role of this gene in parasite control in mosquitoes is still unknown [36]. We also observed that DUOX (dual oxidase), which is involved in Plasmodium elimination, was significantly downregulated in dsAsteglut1-treated mosquitoes (P < 0.0001) [37]. It is highly possible that the reduction of DUOX expression might render mosquitoes more permissive to P. berghei infection.
The CLIP (class of serine proteases) family are involved in the melanization of P. berghei in An. gambiae [38]. Two CLIP genes, clip2 and clip19, were significantly downregulated in dsAsteglut1 treated mosquitoes, while clipb3 was upregulated compared to dsGFP mosquitoes (P = 0.0087, P = 0.0034) [38, 39]. Next, we examined whether the increasing parasite infection could be due to the dysregulation in mosquito melanization. Midguts of mosquitoes treated with dsRNA 8 days post-infection were collected and melanization was visualized microscopically. We found that the number of melanized ookinetes increased with the number of oocysts (Fig. 5d, t(68) = 0.707, P = 0.482). Thus, there was no significant difference in the melanization rate between dsAsteglut1-treated and the dsGFP control group.
Five immune related genes were differentially regulated. Caudal, the negative regulator of Imd pathway was significantly upregulated (P = 0.0337) [40], while the peptidoglycan recognition proteins, pgrp-la, -lc, -ld, and the antimicrobial peptides, defensin were significantly downregulated (P = 0.0111, P = 0.0378, P = 0.0022) [25, 40–43]. These results indicate that Asteglut1 might control parasite infection by regulating mosquito immune responses.