Transparent Process

A new factor in the Aedes aegypti immune response: CLSP2 modulates melanization

Sang Woon Shin, Zhen Zou, Alexander S Raikhel

Author Affiliations

  1. Sang Woon Shin*,1,,
  2. Zhen Zou1 and
  3. Alexander S Raikhel*,1
  1. 1 Department of Entomology, University of California, 3401 Watkins Drive, Riverside, California, 92521, USA
  1. *Corresponding authors. Tel: +1 951 827 7072; Fax: +1 951 827 2130; E-mail: sang.shin{at} or Tel: +1 951 827 2129; Fax: +1 951 827 2130; E-mail: alexander.raikhel{at}
  1. These authors contributed equally to this work


Microbial infections in the mosquito Aedes aegypti activate the newly identified CLSP1 and CLSP2 genes, which encode modular proteins composed of elastase‐like serine protease and C‐type lectin domains. These genes are predominantly regulated by the immune deficiency pathway, but also by the Toll pathway. Silencing of CLSP2, but not CLSP1, results in the activation of prophenoloxidase (PPO), the terminal enzyme in the melanization cascade, suggesting that CLSP2 is a negative modulator of this reaction. Haemolymph PPO activation is normally inhibited in the presence of Plasmodium parasites, but in CLSP2‐depleted mosquitoes, the Plasmodium‐induced block of melanization is reverted, and these mosquitoes are refractory to the parasite. Thus, CLSP2 is a new component of the mosquito immune response.


Melanization is a unique defence mechanism in arthropods involved in wound healing and pathogen encapsulation. Phenoloxidases are key enzymes of melanization, which mediate the enzymatic conversion of tyrosine to eumelanin. A serine‐protease (SP) cascade, similar to the blood‐clotting cascade of vertebrates, proteolytically activates prophenoloxidases (PPOs) to phenoloxidases. The spontaneous activation of the melanization reaction is tightly controlled; otherwise, excessive amounts of quinones generated as a result of systemic hyperactivation of the phenoloxidase system would be deleterious to arthropods. Thus, arthropods have developed a series of negative modulators of melanization. Before infection or injury, the SP cascade involved in PPO activation is inhibited by several serpins specifically controlling SPs (Kanost, 1999; Salzet et al, 1999). Other melanization inhibitors from moth (Manduca sexta) and crayfish (Pacifastacus leniusculus) have been shown to prevent formation of melanin from phenoloxidase‐oxidized phenol substrates (Zhao et al, 2005; Söderhäll et al, 2009).

We study two related genes (CLSP1 and CLSP2) that encode modular proteins consisting of C‐type lectin (CTL) and elastase‐like SP (ESP) domains from Aedes aegypti. Both genes are induced by microbial infections and controlled by the nuclear factor (NF)‐κB factors of the immune deficiency (IMD) and Toll pathways. RNA interference (RNAi) depletion experiments showed that CLSP2 is essential for preventing spontaneous haemolymph melanization. The PPO was proteolytically processed in the haemolymph of mosquitoes 24 h post‐blood meal (PBM), but this response was not observed in mosquitoes infected with the avian malaria parasite, Plasmodium gallinaceum. The CLSP2 gene was activated in P. gallinaceum‐infected mosquitoes; however, RNAi depletion of CLSP2 in infected mosquitoes resulted in the appearance of proteolytically processed PPO. Moreover, these mosquitoes showed a high level of resistance to P. gallinaceum. This suggests that CLSP2 is involved in modulation of melanization.

Results And Discussion

Characterization of CLSP1 and CLSP2

We cloned full‐length complementary DNAs (cDNAs) corresponding to CLSP1 and CLSP2 open reading frames by combining the methods of cDNA library screening, PCR cloning and rapid amplification of cDNA ends, followed by sequencing (supplementary Fig S1,S2 online). Sequences of both genes can be found in the VectorBase—CLSP1 (AAEL011622) and CLSP2 (AAEL011616). CLSP1 and CLSP2 genes encode related proteins, each composed of an amino‐terminal ESP domain and a carboxy‐terminal CTL domain (Fig 1A). C‐termini of both CLSP1 and CLSP2 contain the three‐amino‐acid sequence QPD, which is a signature of the galactose‐specific CTL. The presence of a signal peptide indicates the secretory nature of these protein factors (Fig 1A). Proteins with a composite structure such as those in CLSP1 and CLSP2 have not been previously reported for mosquitoes. The horseshoe‐crab factor C contains SP and CTL domains, and is involved in the haemolymph clotting system (Nakamura et al, 1986). Factor C functions as a pattern‐recognition receptor (PRR); in the presence of diaminopimelic acid (DAP)‐positive peptidoglycan bacteria, it undergoes autoproteolytic cleavage, and the resulting activated protein, containing a SP domain, initiates a haemolymph‐coagulation cascade (Ariki et al, 2004). On the basis of the similarity with horseshoe‐crab factor C, we expected that mosquito CLSPs were immune proteins with dual functions, as pathogen sensors and signal transducers. Both CLSPs contain putative cleavage sites, further emphasizing similarities with factor C (supplementary Fig S1,S2 online).

Figure 1.

Characterization of CLSP1 and CLSP2 genes. (A) Predicted domain structure of CLSP factors. The black box indicates a putative signal peptide, SP and CTL are serine protease and CTL domains, respectively. (B) Immune‐inducible expression of CLSP genes after infection with either bacteria Enterobacter cloacae (Ec) or fungi Beauveria bassiana (Bb; left panel). CLSP genes were highly expressed in the fat body on infection (right panel). (C) Regulation of CLSP gene expression by the immune deficiency pathway. Induction of CLSP genes by E. cloacae infection was blocked in RNA interference REL2‐depleted mosquitoes. (D) Regulation of CLSP genes by Toll pathway. Left panel: Cactus depletion activates transcription of CLSP genes; this activation was compromised by simultaneous depletion of REL1. Right panel: Induction of CLSP genes by B. bassiana (2D Bb) was compromised in REL1‐depleted mosquitoes. Depletion of TOLL5A had no effect. BD are Northern blot analyses. 8 h, 1D or 2D Ec, UGAL mosquitoes 8 h, 1 day or 2 days after septic injury with E. cloacae; 1D or 2D Bb, UGAL mosquitoes 1 day or 2 days after septic injury with B. bassiana; FB, fat body; Naive, previtellogenic UGAL; OV, ovary; iLuc, iCactus, iREL1, iREL2 indicate luciferase‐, Cactus‐, REL1‐, REL2‐ and double‐stranded RNA‐treated mosquitoes, respectively.

The N‐terminal SP domains of Aedes CLSPs are similar to the SP domains of Drosophila melanogaster gastrulation defective (DGD), D. melanogaster ModSP and M. sexta HP14 (Han et al, 2000; Ji et al, 2004; Buchon et al, 2009). The primary specificity pockets of these SPs do not contain the three‐amino‐acid sequence DGG, which is a signature of the trypsin‐like SPs (supplementary Fig S3 online). This suggests that these SPs belong to a subgroup of chymotrypsin/ESPs (Jiang et al, 2005). Arthropod ESPs are involved in the initiation of SP cascades by proteolytically activating other SPs. The Toll pathway establishes the dorsal–ventral polarity during Drosophila embryonic development. Activation of DGD (Han et al, 2000) in the extracellular perivitelline space leads to sequential cleavage of two SPs: Snake (DeLotto & Spierer, 1986) and Easter (Chasan & Anderson, 1989). M. sexta HP14 is autoactivated by an interaction with a peptidoglycan to initiate HP21 CLIP SP in the melanization SP cascade (Ji et al, 2004). HP14 is a modular SP, which in addition to an ESP domain contains four LDLrA repeats, one Sushi domain and one unique wonton domain in the N‐terminal (Wang & Jiang, 2006). Drosophila ModSP (CG31217), an orthologue of M. sexta HP14, mediates signals from PRRs and leads to the activation of Grass CLIP SP in the Toll immune pathway (Buchon et al, 2009).

The C‐termini of CLSP1 and CLSP2 factors are related to CTL‐type molecules that recognize microbial pathogens through lectin–carbohydrate interaction and act as PRRs. A M. sexta CTL, IML‐2, agglutinates Escherichia coli in the presence of calcium and enhances melanization and encapsulation (Yu & Kanost, 2003). Two Anopheles CTLs, CTL4 and CTLMA2, have been shown to have a role in defence against DAP‐positive peptidoglycan bacteria, as well as an antagonist role in the melanization response to Plasmodium berghei ookinetes (Osta et al, 2004; Schnitger et al, 2009).

Regulation of expression of Aedes CLSP genes

We performed northern blot analyses after infections with DAP‐positive peptidoglycan bacteria Enterobacter cloacae and fungi Beauveria bassiana, which are known to activate the IMD and Toll pathways, respectively. Transcripts of approximately 1.4 Kb in size for each CLSP gene were highly elevated after infection with E. cloacae, but to a lesser degree after B. bassiana (Fig 1B). The transcripts of both CLSP genes were enriched in the fat body after infection (Fig 1B). These transcripts were not visible in the ovaries and barely detectable in haemocytes (Fig 1B).

To further investigate the regulatory network responsible for controlling expression of CLSP genes, we evaluated the transcript abundance of these genes after RNAi silencing of key factors of the IMD and Toll pathways. Depletion of the mosquito IMD pathway NF‐κB factor REL2, but not the mosquito Toll pathway NF‐κB factor REL1, blocked upregulation of CLSP genes elicited by E. cloacea, which is consistent with the role of the IMD pathway in mediating CLSP gene activation (Fig 1C). RNAi silencing of Cactus—a gene encoding an inhibitor of Toll NF‐κB factor REL1—resulted in upregulation of CLSP genes. The activation of CLSP genes by Cactus depletion was compromised by simultaneous depletion of REL1, suggesting involvement of the Toll pathway in regulation of CLSP genes (Fig 1D). In an additional experiment, upregulation of CLSP genes activated by infection with B. bassiana was compromised as a result of RNAi depletion of REL1, further supporting involvement of the Toll pathway in regulation of the CLSP genes (Fig 1D). Aedes TOLL5A has been proposed to be the receptor of the Toll immune pathway (Shin et al, 2006). However, depletion of TOLL5A did not yield a response similar to REL1 RNAi in compromising upregulation of CLSP gene expression after fungal infection (Fig 1D). The reason for this discrepancy is unclear. Overall, these experiments showed the dual regulation of CLSP genes by the IMD and Toll pathways.

CLSP2 as a modulator of melanization

Proteolytic cleavage of haemolymph PPO, which suggests its activation, was observed in mosquitoes after fungal infection with B. bassiana, as well as in Cactus‐ and Serpin‐1‐depleted mosquitoes (Zou et al, 2010). This was demonstrated by the appearance of a 20‐kDa protein band, detected by western blot analysis of mosquito haemolymph with antibodies against M. sexta PPO1/2 (Zou et al, 2010). By using the same approach, we observed PPO haemolymph cleavage 5 h after infections with either E. cloacae or Staphylococcus aureus bacteria (Fig 2A). Haemolymph PPO cleavage was not detectable 24 h later, suggesting that it is an acute response to bacterial infection. As the activation of CLSP genes was linked to bacterial infection, we investigated the abundance of the CLSP2 transcript over the time course of E. cloacae infection; the transcript continued to accumulate up to 24 h post‐infection, when acute activation of PPO cleavage was cleared (Fig 2B). Furthermore, we tested whether CLSP factors were involved in controlling PPO haemolymph activation; RNAi depletion of CLSP2, but not CLSP1, resulted in PPO cleavage of haemolymph (Fig 2C), suggesting that the CLSP2 factor functions as a negative modulator of haemolymph melanization.

Figure 2.

The effect of CLSP2 on haemolymph activation of prophenoloxidase. (A) Acute transient activation of haemolymph PPO after septic infections with either Enterobacter cloacae or Staphylococcus aureus. A 20‐kDa band is present in haemolymph samples at 5 h, but absent at 24 h, after infection. Immunoblot using Manduca PPO1/2 antibodies; the antibodies against Aedes lipophorin II (LPII) were used as the loading control. (B) Progression of CLSP2 gene expression over the course of infection with E. cloacae. Northern blot analysis. (C) Spontaneous PPO activation in CLSP2‐depleted mosquitoes. The 20‐kDa cleavage product is present in haemolymph samples from CLSP2‐depleted mosquitoes, but not in those treated with iLuc or iCLSP1. (D) The haemolymph PPO activation in CLSP2‐depleted mosquitoes is compromised by a simultaneous depletion of IMP2, showing that CLSP2 suppresses the PPO cleavage mediated by IMP2. Depletion of IMP1 had no effect. Immunoblot using Manduca PPO1/2 antibodies; the antibodies against Aedes LPII were used as the loading control. (E) Reverse‐transcription polymerase chain reaction analysis showing the specificity of CLSP‐gene knockdowns 5 days after double‐stranded RNA (dsRNA) treatment. 5 h, 24 h Ec, mosquitoes 5 or 24 h after infection with E. cloacae; 5, 24 h Sa, mosquitoes 5 or 24 h after infection with S. aureus; iCLSP1 and iCLSP2 indicate CLSP1 and CLSP2 dsRNA‐treated mosquitoes, respectively. Naive, previtellogenic UGAL.

In A. aegypti, two CLIP SPs, immune melanization protease (IMP)1 and IMP2, have been implicated in mediating haemolymph PPO cleavage (Zou et al, 2010). We investigated the relationship between CLSP2 and these two proteases in haemolymph melanization. Simultaneous depletion of CLSP2 with IMP2, but not with IMP1, stopped the PPO cleavage triggered by depletion of CLSP2 (Fig 2D). This suggests that CLSP2 is involved in the suppression of IMP2‐mediated immune melanization, functioning either upstream from or directly on this CLIP SP. In accordance with these data, we observed no melanotic tumours in CLSP2‐depleted mosquitoes (supplementary Fig S4 online). As a control, we used depletion of Serpin 2, which is involved in tissue melanization, and which depletion by RNAi results in the formation of melanotic tumours (Zou et al, 2010).

Taken together, our results indicate that A. aegypti CLSP2 factor is a negative modulator of haemolymph melanization that is probably responsible for its timely inhibition following an acute phase of infection, as well as for preventing its spontaneous activation. Moreover, we suggest that, similar to the horseshoe‐crab factor C, the mosquito CLSP2 factor is a modular protein containing the CTL domain—which probably acts as a PRR pathogen sensor—and the ESP domain, which is involved in control of melanization. It seems that, similar to D. melanogaster DGD and ModSP, as well as M. sexta HP14 (Han et al, 2000; Ji et al, 2004; Buchon et al, 2009), Aedes CLSP2 functions by interacting with a CLIP SP, in this case IMP2. Further studies are required to understand the mechanistic basis of CLSP2 action.

CLSP2 is required for Plasmodium development

We observed an increase in the haemolymph PPO activation in the mosquitoes at 24 h PBM. However, this PPO activation did not occur when the blood meal contained the malaria parasite P. gallinaceum (Fig 3A). Depletion of CLSP2, but not of CLSP1, reversed the effect of Plasmodium infection and resulted in haemolymph PPO cleavage in infected mosquitoes at 24 h PBM (Fig 3B). Expression of both CLSP genes was elevated in P. gallinaceum‐infected mosquitoes at 24 h PBM; however, the CLSP2 gene was upregulated to a higher level and its transcript was still present at 48 h PBM (Fig 3C; supplementary Fig S5 online). This indicated that the CLSP2 gene was responsive to an unknown stimuli linked to Plasmodium infection. Further studies are required to explain the mechanism of CLSP gene activation by Plasmodium infection.

Figure 3.

Interaction of CLSP2 with Plasmodium parasite in mediating PPO activation. (A) Haemolymph prophenoloxidase activation was observed in mosquitoes 24 h post‐blood meal (PBM); the 20‐kDa cleavage band was not present in haemolymph samples from mosquitoes blood‐fed on chickens infected with the avian malaria parasite, Plasmodium gallinaceum. (B) The 20‐kDa cleavage product was present in haemolymph samples from CLSP2‐depleted mosquitoes fed on P. gallinaceum‐infected mosquitoes, but not in those treated with iLuc control or iCLSP1. (A,B) Immunoblot using Manduca PPO1/2 antibodies; the antibodies against Aedes lipophorin II (LPII) were used as the loading control. (C) Northern blot analyses of activation of CLSP genes in mosquitoes fed on either uninfected or P. gallinaceum‐infected chickens. The CLSP2 gene was upregulated at 24 and 48 h PBM on P. gallinaceum‐infected chickens. Naive, previtellogenic UGAL mosquitoes; PBM, post‐blood meal; PPO, prophenoloxidase.

Importantly, RNAi depletion of CLSP2 resulted in a significantly reduced number of oocysts in midguts of P. gallinaceum‐infected mosquitoes, as compared with iLuc controls. (Fig 4; supplementary Fig S6 online). At the same time, several melanized ookinetes increased in these mosquitoes (Fig 4). RNAi depletion of CLSP1 had no effect on oocyst levels in midguts of P. gallinaceum‐infected mosquitoes (Fig 4). Thus, CLSP2 was required for P. gallinaceum survival in Aedes mosquitoes. It has been reported that RNAi‐mediated depletion of CTLs, CTL4 and CTLMA2 (mannose‐specific CTL2) functioned as protective agonists on the development of ookinete to oocysts in P. berghei in A. gambiae (Osta et al, 2004).

Figure 4.

Reduction of Plasmodium gallinaceum oocyst number in CLSP2‐depleted mosquitoes. Data were collected from three independent experiments with CLSP double‐stranded‐RNA‐treated mosquitoes (supplementary Fig S6 online) and pooled; sample sizes are shown in parentheses. The number of fully developed oocysts in each midgut is shown as a coloured circle. The mean number of parasite oocysts for each group is indicated with a black line and the number on the right. Mean values of melanized ookinetes are indicated by black columns.

Mosquitoes have an effective defence system that is finely tuned to detect various pathogens, including malaria parasites. An integral part of the anti‐Plasmodium response involves targeting of the parasite by complexes of leucine‐rich repeat proteins, LRIM1, APL1A and APL1C, and thioester‐containing protein 1 (Riehle et al, 2006; Povelones et al, 2009; Baxter et al, 2010). In turn, the parasite has evolved effective means to overcome mosquito defences. The ability for tremendous proliferation at the sporozoite developmental stage is an evolutionary adaptation that allows the parasite to overcome losses during the main bottleneck of its developmental cycle from an ookinete to an oocyst. However, Plasmodium evades mosquito immunity by other mechanisms that allow ookinetes to successfully develop to oocysts, which are poorly defined. Lambrechts et al (2007) showed that P. falciparum evades the melanization immune response of A. gambiae. Our data indicate that CLSP2 is required for parasite survival in the mosquito host and might be involved in the parasite‐mediated evasion of the haemolymph melanization response.


Experimental animals. The colony of A. aegypti mosquitoes was maintained in the laboratory and fed continuously on water and 10% sucrose solution. To initiate egg development, mosquitoes were blood‐fed on white rats. All procedures using vertebrate animals were approved by the University of California Riverside Institutional Animal Care and Use Committee (#A20100016; 05/27/2010). The avian malaria parasite P. gallinaceum was maintained in the laboratory by natural transmission between A. aegypti and chickens. Chickens were housed in cages in the vivarium, and all procedures were pre‐approved by the Institutional Animal Care and Use Committee. Midguts from mosquitoes 8 days after a parasite‐harbouring infectious blood‐meal were dissected and stained with 1% mercurochrome solution. The parasites were viewed and counted under a Nikon E400 light microscope.

Synthesis and microinjection of double‐stranded RNA, septic injury, reverse transcription–polymerase chain reaction and RNA blot analysis. The experiments were conducted as described previously (Zou et al, 2010). Primers are listed in the supplementary information online.

Immunoblot analysis. The experiments were conducted as described previously (Zou et al, 2010). The primary antibody against M. sexta PPO1 and PPO2 (provided by Dr H. Jiang) was used for the immunoblot analysis. Antibodies against Aedes lipophorin II were used as loading controls.

Statistical analysis. Data on Plasmodium refractoriness from three independent experiments were collected and analysed using the Kolmogorov–Smirnov goodness‐of‐fit test and pooled. Statistical significance between samples was evaluated using the Mann–Whitney U‐test (Graphpad 5.0).

Supplementary information is available at EMBO reports online (

Conflict of Interest

The authors declare that they have no conflict of interest.

Supplementary Information

Supplementary Information [embor2011130-sup-0001.pdf]


This work was supported by National Institutes of Health/National Institute of Allergy Infectious Diseases grant RO1 AI059492.