|Home | About | Journals | Submit | Contact Us | Français|
The STAT family of transcription factors activate expression of immune system genes in vertebrates. The ancestral STAT gene (AgSTAT-A) appears to have duplicated in the mosquito Anopheles gambiae, giving rise to a second intronless STAT gene (AgSTAT-B), which we show regulates AgSTAT-A expression in adult females. AgSTAT-A participates in the transcriptional activation of nitric oxide synthase (NOS) in response to bacterial and plasmodial infection. Activation of this pathway, however, is not essential for mosquitoes to survive a bacterial challenge. AgSTAT-A silencing reduces the number of early Plasmodium oocysts in the midgut, but nevertheless enhances the overall infection by increasing oocyst survival. Silencing of SOCS, a STAT suppressor, has the opposite effect, reducing Plasmodium infection by increasing NOS expression. Chemical inhibition of mosquito NOS activity after oocyte formation increases oocyte survival. Thus, the AgSTAT-A pathway mediates a late phase anti-plasmodial response that reduces oocyst survival in An. gambiae.
Cytokines such as interleukines and interferons play a central role in regulating and coordinating the immune response. Seven members of the STAT (signal transducers and activators of transcription) family of transcription factors that activate the expression of immune genes in response to cytokine signaling have been characterized in vertebrates (reviewed by Schindler et al., 2007). STAT-1 regulates the transcriptional activation of nitric oxide synthase (NOS) in response to interferon-gamma (IFN-γ). IFN-γ interacts with its membrane receptors and activates a Janus kinase (JAK), which in turn activates STAT-1 by phosphorylating specific residues (Wormald and Hilton, 2004). To prevent toxicity to the host, this pathway is carefully regulated by suppressors such as SOCS-1 (suppressor of cytokine signaling) and PIAS-1 (protein inhibitor of activated STAT). SOCS-1 is part of a negative feedback regulatory loop, since transcription of SOCS-1 mRNA is also regulated by the STAT-1 pathway (Wormald and Hilton, 2004).
In Drosophila, a single member of the STAT family (STAT92E) and several components of the STAT pathway have been identified using a genetic approach. Unpaired (Upd) and two other homologous (Upd2 and Upd3) act as cytokines that activate the membrane receptor Domeless (Dome), which signals through the JAK kinase Hopscotch (Hop) to activate STAT92E (reviewed by Agaisse and Perrimon, 2004). Homologues of the vertebrate STAT repressors SOCS (SOCS36E) (Callus et al., 2002) and PIAS (dPIAS) (Betz et al., 2001) have also been characterized. In Drosophila, the STAT pathway regulates several developmental processes such as stem cell maintenance, eye development and the expression of pair-rule genes involved in embryonic segmentation (reviewed by Arbouzova & Zeidler, 2006). This pathway is also activated in response to bacterial and viral infections and participates in cellular immune responses by regulating prohemocyte differentiation (Krzemien et al, 2007) and hemocyte proliferation (Hanratty and Dearolf, 1993 and Harrison et al., 1995).
Disruption of Hop increases the viral load in flies infected with Drosophila C virus (DCV), and activation of Hop is necessary, but not sufficient, for the expression of virus-induced genes (Dostert et al., 2005). In Drosophila, the STAT pathway regulates the expression of a member of the thioester-containing protein (TEP1), SOCS36E and members of the tot family (totA, totC and totM) in response to bacterial challenge (reviewed by Agaisse and Perrimon, 2004). The protein product of the TotA gene is expressed in the fat body and is secreted into the larval hemolymph in response to several stress conditions, including bacterial challenge, but its function remains unknown (Ekengren et al., 2001). Bioinformatic analysis revealed that there are no orthologues of the Tot genes in An. gambiae or in other insect species outside the Drosophila genus (C. Barillas-Mury unpublished). TEP1 is a member of the α-2-macroglobulin family of thioester-containing proteins. The function of this gene has been explored in the mosquito Anopheles gambiae (Levashina et al., 2001). TEP1 is produced by hemocytes, and is secreted into the hemolymph in response to bacterial infection. It binds to the surface of bacteria and has been shown to promote phagocytosis in an immuno-responsive cell line from the mosquito Anopheles gambiae (Levashina et al., 2001). TEP1 is also secreted into the mosquito hemolymph in response to Plasmodium berghei infection and binds to the surface of ookinetes as they emerge from midgut epithelial cells. Knockdown of TEP1 in An. gambiae by dsRNA-mediated silencing increases the number of oocysts that develop, indicating that, when present, TEP1 promotes parasite killing through a lytic mechanism (Blandin, S., et al., 2004; Garver et al., 2008).
A member of the STAT family (annotated as STAT1), that translocates to the nucleus and binds DNA in response to bacterial challenge was previously characterized in the mosquito An. gambiae (Barillas-Mury et al., 1999). Completion of the An. gambiae genome revealed the existence of a second STAT gene (annotated as STAT2) (Christophides et al., 2002). The annotated names could be misleading because these genes are not orthologues of the vertebrate STAT1 and STAT 2 genes. STAT2 is the ancestral gene, thus we will refer to it as AgSTAT-A. A spliced mRNA of the AgSTAT-A gene appears to have been reverse-transcribed and integrated back into the An. gambiae genome, giving rise to a second “intronless” STAT gene (annotated as STAT1, but we will refer to it as AgSTAT-B).
In this manuscript we examine the participation of the STAT pathway in the mosquito immune response to bacteria and Plasmodium. We find that AgSTAT-B regulates AgSTAT-A mRNA expression and that AgSTAT-A mediates the transcriptional activation of NOS and SOCS in response to infection. These findings provide direct evidence that, in insects, NOS expression is also regulated by the STAT pathway and suggests that the organization of this signaling cascade precedes the divergence of insects and vertebrates. AgSTAT-A silencing does not affect An. gambiae survival following bacterial injection or feeding, but significantly increases the number of P. berghei and P. falciparum mature oocysts that develop. As expected, knockdown of AgSOCS, a STAT repressor, has the opposite effect, decreasing Plasmodium infection. A detailed functional analysis revealed that the AgSTAT-A pathway limits Plasmodium infection by inducing NOS expression and drastically reducing the number of early oocysts that complete maturation, defining a distinct late phase in the mosquito’s anti-plasmodial responses.
AgSTAT-A has multiple introns and is located in Chr. X-4B, while AgSTAT-B is an intronless gene located in Chr. 3L-38B (Fig. 1A). Sequencing of the AgSTAT-A cDNA from adult females revealed that the gene consists of 9 exons separated by 8 introns (Fig 1A) (GenBank accession No. FJ792607). The second intron, located in the 5′ UTR, is very large (21.1 Kb). The current annotation of the AgSTAT-A gene in the An. gambiae genome predicts a small intron that would split exon 4 into two exons. However, it is clear from our cDNA sequence that this predicted intronic sequence is part of a continuous open reading frame in exon 4 (Fig. S1). The alignment of the deduced amino acid sequences of members of the STAT family from different species (Fig. S2) was used to build a phylogenetic tree (Fig 1B). Insect STATs have higher homology to the vertebrate STAT5 & STAT6 genes (Fig 1B). AgSTAT-A clearly clusters with the STAT genes from other mosquito species (Culex tritaeniorhynchus and Aedes aegypti) indicating that this is most likely the ancestral form of the gene, while AgSTAT-B is diverging fast. The duplication of the STAT gene appears to have taken place after anopheline and culicine mosquitoes diverged, because only one STAT gene has been identified in the Ae. aegypti and D. melanogaster genome sequences. AgSTAT-A mRNA is not expressed in pupae, but is abundant in other developmental stages (Fig 1C). For example, AgSTAT-A is expressed at much higher levels (about 1,000 fold) than AgSTAT-B in adult females (Fig 1C). In contrast, AgSTAT-B mRNA is expressed at high levels in the pupal stage (Fig 1C), indicating that AgSTAT-B is also a functional gene.
Several immuno-responsive cell lines from An. gambiae were screened, and the Sua5.1 cell line (Muller et al., 1999) was found to express both AgSTAT genes, AgNOS and AgSOCS. Furthermore, the mRNA levels of NOS and SOCS increase significantly when these cells are challenged with heat-killed bacteria (Fig. 2A). Double stranded RNAs (dsRNAs) synthesized in vitro from the most divergent regions (N-terminal) of AgSTAT-A and AgSTAT–B were used to silence gene expression. These probes were designed so that they do not share any stretch of identical nucleotide sequence longer than 11 bp. Endogenous AgSTAT-A and AgSTAT-B mRNA levels are significantly reduced (84% and 68%, respectively) when cells were transfected with their respective dsRNA (Fig. 2B–C). Unexpectedly, AgSTAT-B silencing also reduced AgSTAT-A mRNA levels by about 80%. This effect was confirmed and further investigated in experiments described in the following section. The induction of AgNOS and AgSOCS mRNAs in response to infection is significantly reduced when either AgSTAT-A or AgSTAT-B are silenced (Fig. 2B–C), indicating that, as in vertebrates, the STAT pathway also regulates SOCS and NOS expression in An. gambiae. Silencing both STAT genes, by transfecting cells simultaneously with both dsRNAs, does not further reduce NOS or SOCS expression (data not shown).
AgSTAT-A silencing was effective and did not reduce AgSTAT-B expresssion in An. gambiae G3 females in which the gene was silenced by systemic injection of dsRNAs (Fig. 3A). However, AgSTAT-B silencing also decreased AgSTAT-A mRNA in adult females (data not shown). To rule out the possibility of unspecific cross-silencing, an independent dsRNA was designed from the 5′ UTR region of AgSTAT-B, which bears no sequence homology to AgSTAT-A. This AgSTAT-B 5′-UTR dsRNA, also silenced AgSTAT-A expression in adult females (Fig. 3B), indicating that AgSTAT-B regulates the basal levels of AgSTAT-A mRNA. These findings support a model for the STAT pathway in An. gambiae in which both STAT genes are in the same signaling cascade, with AgSTAT-B acting upstream and regulating AgSTAT-A expression (Fig. 3C). Based on this model, we proceeded to evaluate the participation of AgSTAT-A on antibacterial and antiplasmodial responses in An. gambiae females.
The Toll and Imd pathways play a central role in Drosophila antibacterial responses and disruption of these cascades severely compromises the ability of mutant flies to survive a bacterial challenge (reviewed by Tanji and Ip, 2005; Hoffmann and Reichhart, 2002). Although the STAT92E pathway is clearly activated in response to bacteria, disruption of this cascade does not have the same dramatic effect on fly survival to bacterial infection (Agaisse and Perrimon, 2004). Furthermore, the Relish-dependent transcriptional activation of Attacin-A was recently found to be hyperactivated in the absence of Stat92E (Kim et al., 2007). Stat92E is involved in the formation of a repressosome complex that replaces the transcription factor Relish at the promoters of immune effector genes. This is thought to be an important mechanism to turn off transcriptional activation of genes activated by the Imd pathway and to prevent the chronic toxicity that could result from prolonged immune activation (Kim et al., 2007). Systemic injection of E. coli and M. lueteus in An. gambiae adult females induces the expression of NOS and SOCS mRNAs (Fig 3D) and silencing of AgSTAT-A significantly reduces NOS and SOCS mRNA levels in challenged females 4 h post-challenge (Fig 3E). The effect of AgSTAT-A silencing on mosquito survival to injection of a mixture of Escherichia coli and Micrococcus luteus or of two other bacterial species isolated from the gut flora of our An. gambiae laboratory colony, Asaia siamensis and Pseudomonas sp., was investigated. STAT-A silencing did not have a significant effect on mosquito survival to systemic challenge with these three bacterial species (Fig. 3F). Survival to chronic bacterial feeding with either E. coli or M. luteus was also not affected by STAT-A silencing (Fig. 3G).
AgSTAT-A silencing significantly reduces the basal pre-invasion mRNA levels of AgNOS and AgSOCS, while TEP1 expression is not affected (Fig. 4A). Plasmodium berghei infection induces the expression of AgNOS, AgSOCS and AgTEP1 24 h PI (Fig. 4B). AgSTAT-A silencing was very effective (~99%) in the carcass and dramatically reduced AgNOS, AgSOCS and AgTEP1 expression (~99%) 24 h PI (Fig. 4C). Furthermore, AgSTAT-A silencing exacerbated infection, increasing the median number of P. berghei oocysts 4.4 fold at 7 days PI (KS test, p <0.001) (Fig. 4D and E, left panels). AgSTAT-B silencing has a similar effect as AgSTAT-A, also enhancing infection 7 days PI (data not shown). As expected, silencing of AgSOCS, a suppressor of STAT signaling, has the opposite effect and reduces the median number of oocysts by 7 fold (Fig. 4D right panel and 4E middle panel) (KS test p <0.001). AgSTAT-A silencing also enhances P. falciparum infection by 1.7 fold. (Fig. 4E, right panel) (KS test p <0.001).
Because AgSTAT-A silencing prevents the transcriptional activation of TEP1 in response to Plasmodium infection (Fig. 4C), and silencing either of these genes enhances infection in the G3 strain 7–8 days PI, we decided to investigate whether the effect of AgSTAT-A silencing on Plasmodium infection is mediated by TEP1. In the refractory (R) L35 strain, ookinetes invade the midgut but are killed and melanized as soon as they come in contact with mosquito hemolymph, and melanized parasites remain visible on the midgut surface for the lifetime of infected females. TEP1 is a critical mediator of this early response, and silencing of this gene prevents parasite death and allows oocyst development (Blandin et al., 2004).
We confirmed that, in the R strain, TEP1 silencing prevents parasite lysis and melanization. No melanizations were observed when TEP1 was silenced and the number of parasites present 7 days PI was also significantly higher (p < 0.005, KS test) (Fig S3). In contrast, AgSTAT-A silencing did not revert Plasmodium melanization, but instead, resulted in a significant decrease in the number of melanized parasites (Fig. 5A), suggesting that less parasites completed midgut invasion. To further explore this possibility, the effect of silencing AgSTAT-A in G3 females on the number of early oocyst was evaluated by analyzing infected midguts 2 days PI. AgSTAT-A silencing significantly reduces the number of oocysts that form in G3 females (Fig. 5B). To rule out the possibility that the reduction in oocyst formation is due to some unspecific off-target effect of the AgSTAT-A dsRNA used, AgSTAT-A was silenced indirectly by silencing AgSTAT-B, an upstream gene that regulates AgSTAT-A expression (Fig. 3C). The dsRNA from the 5′ UTR region of AgSTAT-B reduces AgSTAT-A expression very efficiently (Fig. 3B) and also decreases the number of oocysts that form 2 days PI in G3 females (Fig. 5C).
To better understand how AgSTAT-A silencing can reduce the number of early oocysts that form 2 days PI, but at the same time increase the number of late oocysts present 7–8 days PI, the effect of silencing was evaluated at these two time points in G3 strain mosquitoes that were infected with the same mouse (Fig. 5D). In two independent experiments, we confirmed that AgSTAT-A silencing significantly decreases the number of early oocyst present 2 days PI (2.5 fold) relative to the control. However, when midguts from the same experiment are analyzed 8 days PI, the median number of oocysts is 5 fold higher in the AgSTAT-A silenced group. These two opposite effects are possible because the number of parasites present in the control group decreases dramatically (10–12 fold) between day 2 and 8 PI (Fig. 5D). In contrast, there is no significant decrease in oocyst numbers between these two time points when AgSTAT-A is silenced. We conclude that AgSTAT-A is required for early parasite survival, but also limits Plasmodium infection by mediating oocyst lysis and decreasing the number of early oocysts that complete development. This phenotype defines a distinct late phase in the anti-plasmodial immune responses of An. gambiae and is illustrated in Fig. 5E.
In An. stephensi, NOS mRNA levels increase in response to Plasmodium infection and this induction is still observed 2, 3 and 9 days PI. Hemolymph levels of nitrite/nitrate are significantly higher 7 days PI and NOS specific activity is 5.3 fold higher 9 days PI (Luckhart et al., 1998). The hypothesis that NOS is an important effector of the AgSTAT pathway was investigated by testing whether reducing NOS expression in SOCS-silenced females could rescue parasite survival.
As expected, SOCS silencing reduced the number of oocysts present 8 days PI by 4.3 fold (p<0.001, KS test) (Fig. 6A), and NOS mRNA levels 4 days PI were 6.7 fold higher than the control group (Fig. S4). In the double silenced group (dsNOS + dsSOCS), NOS mRNA levels 4 days PI are 2.9 fold lower than in the SOCS-silenced group (Fig. S4). Furthermore, this decrease in NOS expression in the double-silenced group completely reverts the decrease in infection observed when SOCS alone is silenced (Fig. 6A).
When NOS alone is silenced, NOS mRNA levels 4 days PI are 3.65 fold lower than in the double-silenced (dsNOS + dsSOCS) group (Fig. S4). Interestingly, this further reduction in NOS mRNA expression does not enhance infection, but has the opposite effect and drastically reduces the number of oocysts present 8 days PI (p<0.001) (Fig. 6A). The possibility that this could be due to a reduction in oocyst formation, similar to what is observed when either AgSTAT-A or AgSTAT-B are silenced (Fig 5B–C) was explored by analyzing infected midguts 2 days PI. When NOS expression is very low, the number of oocysts that form by 2 days PI is, indeed, greatly reduced (6 fold) (Fig. 6B). Taken together, our data indicate that, when NOS expression is reduced by silencing either AgSTAT-B, AgSTAT-A, or NOS directly, the number of parasites that complete midgut invasion and transform into oocysts is reduced (Fig. 5B–C, ,6B),6B), and suggests that a minimum level of NOS activity is required by early stages of Plasmodium in the mosquito. When NOS is directly silenced, the reduction in oocyst formation is more drastic and oocyst numbers are so low by 2 days PI, that it is not possible to evaluate the effect of NOS silencing in oocyst growth and maturation.
To investigate the effect of NOS on oocyst survival, without affecting oocyst formation, we took an alternative approach. Mosquitoes were infected and oocysts were allowed to form. Oral administration of the NOS inhibitor L-NAME or its inactive enantiomer D-NAME was initiated 36 h PI. L-NAME administration significantly increased the number of oocysts present 8 days PI (Fig. 6C). Furthermore, oral administration of L-NAME increased the prevalence of infection in SOCS-silenced females from 43% to 75% (P<0.02, χ2), as well as the intensity of infection (Fig 6D). These findings indicate that, once oocysts form, reducing NOS activity enhances infection by increasing oocyst survival.
The cellular localization of NOS was determined by immunofluorescence staining of infected midguts collected 8 days PI. NOS protein is present in the cytoplasm of midgut epithelial cells (Fig. 6E) and cannot be detected in hemocytes. SOCS-silenced midguts in which the infection level was low had a tendency to stain stronger for NOS compared to midguts of females injected with dsLacZ or to double-silenced (SOCS+NOS) midguts (Fig. 6E).
To determine which organ(s) could be the source of nitric oxide (NO) production and to quantitate NOS expression, NOS mRNA levels were analyzed in the midgut, hemocytes and carcass of SOCS-silenced females 4 days PI. NOS mRNA levels are significantly increased in the midgut (4–10 fold) and carcass (4–6 fold) (p<0.01, t-test) but not in circulating hemocytes (Fig. 6F).
Hemocytes attach to the basal surface of the midgut in response to Plasmodium infection, around 24 h PI (Blandin et al., 2004). Although TEP1 is only expressed in hemocytes, higher levels of TEP1 mRNA are detected in dissected midguts due to increased hemocyte attachment (Vlachou et al., 2005). This adhesion is transient, as TEP1 levels are no longer higher on infected midguts collected 40–48 h PI (Vlachou et al., 2005). To further explore whether hemocyte adhesion to the midgut or carcass also increases as part of the late responses mediated by the AgSTAT-A pathway, TEP1 expression was also determined in midgut, hemocytes and carcass of SOCS-silenced females 4 days PI. TEP1 mRNA levels are significantly higher in hemocytes, but remained unchanged in the midgut and carcass (Fig. 6F) (p<0.01, t-test).
The duplication of the STAT gene in An. gambiae is exceptional among insects. The fact that AgSTAT-B is located in a different chromosome and lacks introns strongly suggests that an mRNA from the ancestral AgSTAT-A gene was reverse-transcribed and inserted into a different chromosome. Although AgSTAT-B mRNA is much less abundant, it regulates the basal levels of AgSTAT-A mRNA, the predominant form expressed in adults. Co-silencing of both genes has the same effect on NOS and SOCS induction in response to bacterial challenge as silencing AgSTAT-A alone, suggesting that they are part of the same signaling pathway (Fig. 3C). The addition of a second transcription factor to the same pathway could be important to amplify the response when this cascade is activated. AgSTAT-A regulates the basal levels as well as the induction of NOS and SOCS in response to bacteria in the Sua5.1 cell line and in adult females, as well as the transcriptional activation of these genes in response to Plasmodium infection. Taken together, these findings provide clear experimental evidence that NOS expression is regulated by the STAT pathway in An. gambiae; indicating that this is probably an ancient signaling cascade that precedes the divergence of insects and vertebrates. The previous observation that there are multiple consensus DNA-binding sites for transcription factors in the upstream regulatory region of the An. stephensi NOS gene, including STAT-binding sites, is in agreement with our findings (Luckhart and Rosenberg, 1999). The fact that all the STAT genes from insects cluster with the vertebrate STAT5 and STAT6 genes indicates that these two family members are closer to the common ancestral gene between insects and vertebrates. In vertebrates, the STAT family has undergone extensive gene duplication and divergence, resulting in seven functional STAT genes. Presumably, although the STAT1 gene is very divergent from the common ancestral gene, it has retained the regulation of inducible NOS (iNOS) expression.
Although AgSTAT-A participates in the transcriptional induction of NOS, activation of this signaling cascade is not necessary for adult females to survive a systemic or an oral challenge with E. coli and/or M. luteus. These bacteria have very low virulence in An. gambiae, and very large numbers of bacterial are required to cause any lethality (in the experiments shown, adult female were injected with 1 × 105 CFU from each bacterial species). Asaia siamensis and Pseudomonas sp. are much more virulent, and significant mortality is observed after injecting only 6 CFU, but AgSTAT-A silencing also does not have a significant effect on survival. An. gambiae may require the induction of NOS through the STAT pathway to effectively control microorganisms with different biology, such as intracellular pathogens.
In Drosophila, increased production of nitric oxide (NO) is observed during hemocyte-mediated melanotic encapsulation responses, and injection of exogenous NO induces diptericin expression in larvae (Nappi et al., 2000). NO induction of diptericin reporters in the larval fat body requires imd, suggesting that NO acts upstream of the Imd pathway (Foley and O’Farrell, 2003). Inhibition of nitric oxide synthase (NOS) activity prevents diptericin induction in response to Gram-negative bacteria and increases larval sensitivity to infection (Foley and O’Farrell, 2003). In Drosophila, however, whether NOS expression is regulated by the STAT92E pathway remains to be established.
AgSTAT-A silencing does not affect the pre-invasion levels of TEP1, but mediates the transcriptional activation of TEP1 24 h PI, suggesting that AgSTAT-A regulates the transcriptional activation of TEP1 in hemocytes to replenish the TEP1 protein that is secreted in response to infection. In contrast to TEP1 silencing, AgSTAT-A silencing does not revert the refractory phenotype in the R strain, but instead, decreases the number of parasites that complete invasion. TEP1 binds to the surface of ookinetes as they emerge from the midgut epithelial cell surface (around 24 h PI) and promotes lysis (Blandin et al., 2004). AgSTAT-A silencing, however, enhances infection at a later time by preventing oocyst lysis.
Several lines of evidence indicate that there are large parasite losses when mosquitoes are infected with Plasmodium. For example, it has been estimated that only one in every 50 P. berghei ookinetes present in the midgut lumen of An. gambiae succeed in forming a mature oocyst 10 days PI (Alavi et al., 2003). Furthermore, silencing of either TEP1 or Leucine rich-repeat immune protein 1 (LRIM1) increases P. berghei infection 4–5 fold (Blandin et al., 2004; Osta et al., 2004). Based on these observations, it is estimated that at least 80% of parasites are lysed soon after they invade the midgut, as they come in contact with the mosquito immune system (Blandin et al., 2004). However, there is very limited information regarding parasite losses during the oocyst stage. It has been documented that the number of P. berghei parasites present in the midgut 24 h PI (ookinete stage) decreases 2.8 fold by 48 h PI (mostly early oocysts) in adult females injected with dsGFP (control group), and at 3 days PI a further 35 % decrease in the number of oocysts is observed (Abraham et. al., 2005). Here we provide evidence that this decrease in oocysts numbers is part of an active anti-plasmodial response.
The significant increase in P. berghei and P. falciparum oocysts present 7 days PI when AgSTAT-A is silenced, together with the decrease in infection following AgSOCS silencing, indicate that activation of this signaling cascade is an important component of the immune responses of An. gambiae against the parasite. The effects of AgSTAT-A signaling on Plasmodium survival are complex, due to the dual effect of NOS in parasite survival in the mosquito. When NOS expression is reduced during the early stages of infection, less parasites complete midgut invasion and less oocysts form (Fig. 6B). It is very unlikely that this effect is due to an unspecific off-target effect of the NOS dsRNA used to silence expression, as it is also observed when two other upstream genes, AgSTAT-B or AgSTAT-A are silenced (Fig. 5B, C), and co-silencing of SOCS and NOS increases NOS mRNA levels and rescues this effect (Fig. 6A). Taken together, our data suggest that some minimal level of NOS activity is required for early stages of P. berghei to survive in the mosquito. Once oocysts form, reducing NOS expression enhances infection (Fig. 6C) and rescues the effect of SOCS silencing (Fig 6A, D), indicating that NOS is a major effector molecule limiting oocyst survival in response to STAT activation.
SOCS silencing overactivates the STAT pathway and significantly increases NOS mRNA expression in the midgut and carcass, but not in hemocytes. NOS protein is expressed in the cytoplasm of midgut epithelial cells. The staining is homogeneous and NOS expression is not higher in cells in close proximity to developing oocysts, suggesting a generalized rather than a local epithelial response to infection. Although hemocytes do not appear to be the main source of NO production, activation of the STAT pathway in these cells may release a signaling molecule(s) that induce NOS expression in other tissues.
Other immune pathways described so far in An. gambiae affect Plasmodium soon after invasion is completed, at the ookinete-to-oocysts transition (Fig. 5E). The broadly accepted view is that once oocysts mature and modify their surface, they are “hidden” and unaffected by the mosquito’s immune system. Our findings establish that this is not the case, and define a distinct late phase in the mosquito antiplasmodial responses (Fig. 5E). The elucidation of killing mechanisms that target the oocyst stage is particularly attractive to control disease transmission, as Plasmodium parasites require several days to complete this stage in their life cycle. We conclude that activation of the STAT pathway in An. gambiae limits Plasmodium infection by decreasing parasite survival at the oocyst stage and that NOS is a key effector of this response.
Two different An. gambiae strains were used. The G3 strain is naturally susceptible to P. berghei and P. falciparum infection. The L3-5 strain was genetically selected from G3 to be refractory and melanizes many different Plasmodium species including P. berghei (Collins et al., 1986). Mosquitoes were reared at 27 °C and 80% humidity on a 12-h light-dark cycle under standard laboratory conditions.
The A. gambiae Sua 5.1 cell line was established from neonatal larvae of the Suakoko mosquito strain, as described (Muller et al., 1999). Cells were grown in 6 well-plates and kept continuously at 26 °C in Schneider medium supplemented with 10% FCS and Penicillin/Streptomycin. Cells were grown to 70% confluency and transfected with 1 μg of dsRNA using the Effectene Transfection Reagent (Quiagen) according to the manufacturer instructions. Briefly, dsRNA was mixed with the enhancer solution and incubated for 5 min. The Effectene Reagent was added and the mixture incubated for 10 min, mixed with growth media and added directly to cells. Six hours later the cells were washed with PBS, resuspended in fresh media and harvested 5 days later. Cell cultures were challenged with E. coli and M. luteus 5 days after they were transfected with dsRNA and harvested 6 hours post-challenge. Bacteria from a single colony were grown in LB broth at 37 °C with vigorous shaking. Saturated overnight cultures of E. coli (250 μl) and M. luteus (1 ml) were used to seed 25 ml of LB broth and both cultures grown until they reach an OD of 0.5. One milliliter from each culture was mixed and centrifuged, the bacterial pellet was washed twice with PBS and was resuspended in 100 μl of cell culture media. Bacteria were heat-killed by placing them in a boiling bath for 5 min. and 10 μl of this suspension was added to each well containing 2.5 ml of culture media. The silencing efficiency, relative to the expression levels in control cells transfected with dsLacZ (100%) ranged from 84–90% for AgSTAT-A and from 64–77% for AgSTAT-B.
A 218-bp fragment of the LacZ gene was amplified using the primers (5′ to 3′) F-GAGTCAGTGAGCGAGGAAGC and R-TATCCGCTCACAATTCCACA and cloned into the pCR®II-TOPO® vector. T7 promoters were incorporated onto the ends of this fragment by amplifying the cloned insert using the following primers M13F-GTAAAACGACGGCCAGT and M13R-CTCGAGTAATACGACTCACTATAGGGCAGGAAACAGCTATGAC. The PCR product was used as template to synthesize dsRNA in vitro using the MEGAscript RNAi kit (Ambion, Austin, TX). dsRNA was further purified with water and concentrated to 3μg/μl using a Microcon YM-100 filter (Millipore). A similar cloning strategy was used for all other genes. A cDNA fragment from the N-terminal region of AgSTAT-A and AgSTAT-B, with the lowest homology between them, was used to generate dsRNA templates, using the following primer sets: dsAgSTAT-A (5′ to 3′), F-CCGGAGAGCAACTTCACGAT and R-GATGAACGTGTTGGTAATGAGC; dsAgSTAT-B, F-GATTTGAAATATCCCGTGTTA and R-TTCTCGTAACTGGCTTTTCATT. A second dsRNA was synthesized based on the 5′ UTR sequence of AgSTAT-B, which bears no sequence homology to AgSTAT-A, using the following primers: UTR-dsAgSTAT-B, F-GCTTTACAGCGGCACCGATAGTAA and R-TTTAGGAAATGGAACTGCTCTAAA. NOS primers to generate dsRNA: F-GGTGTTCTCGATCGCGTGTTCTT and R- CGCAGCGTCAGCATGTATTTCTC; SOCS primers to generate dsRNA: F-GCCGGACCTGCAGAAGATTAC; R-CTTCGGTAGCGTCAGCTCGTTGAT.
Female mosquitoes were injected with 69 nl of a 3μg/μl solution of dsRNA (207 ng dsRNA) from the gene of interest at 1–2 days post-emergence. Control mosquitoes were injected with dsLacZ. Four days later, females were either fed on a mouse infected with P. berghei, or were infected by feeding on a P. falciparum gametocyte culture provided in an artificial feeding system. Four double silencing experiments dsSOCS was mixed 1:1 with either dsLacZ or dsNOS and 207 ng of dsRNA injected in a 69 nl volume. All phenotypes were confirmed in at least two independent experiments. The silencing efficiency in sugar-fed females, relative to controls injected with dsLacZ (100% expression) was: 81–92% for AgSTAT-A, 76–83% for AgSTAT-B, 61–98% for AgSOCS and 85–90% for AgNOS. Silencing of AgSTAT-A in the carcass (midgut removed) of Plasmodium-infected females ranged from 80–99.7%.
Adult females were injected with dsRNA 1–2 days post-emergence, and four days later they were injected with a mixture of Escherichia coli and Micrococcus luteus. Bacterial cultures were grown to an OD of 0.5 in LB broth, 200 μl from each culture were mixed and centrifuged for five minutes at maximum speed. The supernatant was discarded, the pellet was washed twice with PBS, and resuspended in 125 μl of PBS. Mosquito survival was monitored daily for 8 days after intrathoracic injection of 0.138 μl of either PBS or bacterial suspension. Asaia siamensis and Pseudomonas sp. cultures were grown at 27 °C. Bacteria were suspended in PBS and serially diluted so that 6 CFU were injected in a volume of 0.138 μl of PBS. Three groups of 30 mosquitoes were used for each treatment and survival curves analyzed by the Log Rank survival test (SigmaStat). The effect of AgSTAT-A silencing on NOS and SOCS expression was determined in samples collected four hours after bacterial challenge.
For the oral challenge, bacterial cultures were grown to OD 0.6, 4 mls of culture were centrifuged and the pellet resuspended in 2 ml of 10% sucrose solution. This bacterial suspension was placed in a plastic tube plugged with cotton and changed daily. The inverted tube was placed in the mosquito cage and mosquitoes fed ad libitum.
Mosquito females were infected with P. berghei by feeding on anesthetized infected Balb/C mice. The infectivity of the mice was established by determining the parasitemia and by performing an exflagellation assay as described previously (Billker et al., 1997). In all the studies, mice having 2–3 exflagellations/field under 40× objective were used to infect mosquitoes. Blood-fed mosquitoes were kept at 21 °C and 80% humidity. P. berghei infections were performed using a transgenic GFP-P. berghei strain (GFP-CON transgenic 259cl2 strain; Franke-Fayard et al., 2004) infection phenotypes were determined 7–8 days PI. The distribution of parasite numbers in individual mosquitos between the control and experimental groups were compared using the Kolmogorov-Smirnov test (KS test). Al phenotypes were confirmed in 2–3 independent experiments.
P. berghei midgut infections 2 days PI were quantified by immunofluorescence using mouse anti-Pbs21 antibody and NOS expression was detected with a commercial Universal anti-NOS rabbit polyclonal antibody (Affinity Bioreagents, Inc., catalog no. PA1-039) as previously described (Han et al., 2000).
SOCS expression was silenced by injection of dsSOCS in 2 day old females and were fed on a P. berghei-infected mouse 3 days later. Nω-nitro-L-arginine methyl ester (L-NAME) (Sigma) or the inactive enantiomer D-NAME (Sigma) were administered as 1 mg/ml solutions in water as previously described (Luckhart et al., 1998). Sugar cubes were provided and the treatments were started 36 h PI and continued until samples were collected at 8 days PI.
An. gambiae (G3) female mosquitoes were infected artificially by membrane feeding with a P. falciparum gametocyte culture. P. falciparum 3D7 strain was maintained in O+ human erythrocytes using RPMI 1640 medium supplemented with 25 mM HEPES, 50 mg/L hypoxanthine, 25 mM NaHCO3 and 10%(v/v) heat-inactivated type O+ human serum (Trager and Jensen, 1976; Zolg, et al., 1982). Gametocytogenesis was induced as previously described (Ifediba and Vanderberg, 1981). Mature gametocyte cultures (stages IV and V) that were 14–16 days-old were used to feed mosquitoes in 37°C warmed membrane feeders for 30 min. Midguts were dissected 8 days PI and oocysts stained with 0.05% (w/v) mercurochrome in water and counted by light microscopy. The distribution of parasite numbers in individual mosquitos between control and experimental groups were compared using the Kolmogorov-Smirnov test (KS test).
Hemocytes were collected using a slight modification of a procedure previously described (Castillo et al., 2006). Briefly, 2–3μl of anticoagulant solution consisting of 60% Schneider’s medium, 10% fetal bovine serum and 30% citrate buffer (98mM NaOH, 186mM NaCl, 1.7mM EDTA and 41mM citric acid, buffer pH 4.5) vol/vol were injected gradually into the thorax. Diluted hemolymph was collected 5 min. after the injection through an incision in the abdominal wall using a capillary tube and immediately placed in TRIzol solution (invitrogen) for RNA extraction.
We thank André Laughinghouse, Kevin Lee, Tovi Lehman, and Robert Gwadz for insectary support and Juraj Kabat from the NIH Biological Imaging Section for his assistance with confocal imaging. We are also grateful to José Ribeiro and Jesus Valenzuela for their comments and insight. This research was supported by the Division of Intramural Research of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.