Sp22D: a multidomain serine protease with a putative role in insect immunity

Gene 251 (2000) 9–17 www.elsevier.com/locate/gene

Sp22D: a multidomain serine protease with a putative role in insect immunity M.J. Gorman *, O.V. Andreeva, S.M. Paskewitz Department of Entomology, University of Wisconsin, Madison, WI 53706, USA Received 18 February 2000; accepted 12 April 2000

Abstract Serine proteases play critical roles in a variety of insect immune responses; however, few of the genes that code for these enzymes have been cloned. Here, we describe the molecular characterization of a serine protease gene from the mosquito Anopheles gambiae. Sp22D codes for a 1322 amino acid polypeptide with a complex domain organization. In addition to the carboxy terminal serine protease catalytic domain, Sp22D contains two putative chitin binding domains, a mucin-like domain, two low density lipoprotein receptor class A domains, and two scavenger receptor cysteine rich domains. A typical signal peptide sequence and a lack of potential transmembrane helices suggest that Sp22D is secreted. Sp22D is expressed constitutively in three immunerelated cell types: adult hemocytes, fat body cells, and midgut epithelial cells. Wounding induces no changes in transcript abundance, but within 1 h after injection of bacteria, Sp22D mRNA increases 1.5-fold. Based on domain organization, tissue distribution, and transcriptional up-regulation in response to immune challenge, we suggest that Sp22D has an immune function. In addition, we predict that Sp22D is secreted into the hemolymph where it may interact with pathogen surfaces and initiate an immune response. © 2000 Published by Elsevier Science B.V. All rights reserved. Keywords: Anopheles gambiae; Chitin binding domain; LDLRA; Mosquito; Mucin domain; SRCR

1. Introduction Serine proteases play critical roles in a variety of invertebrate immune processes, mainly in the activation of other proteins by site-specific cleavage. Examples of serine protease mediated defense responses include hemolymph (blood ) coagulation, activation of antimicrobial peptide synthesis, and melanin synthesis. Unfortunately, despite the importance of these enzymes in invertebrate immunity, few of their genes have been cloned. Decidedly the best understood invertebrate defense response is hemolymph coagulation in the horseshoe crab (Iwanaga et al., 1998). This immune response is initiated in the presence of two types of pathogen surface molecules, the bacterial lipopolysaccharides (LPS) and Abbreviations: aa, amino acid; CBD, chitin binding domain; kb, kilobase; LDLRA, low density lipoprotein receptor class A; LPS, lipopolysaccharide; PBS, phosphate-buffered saline; PCE, proclotting enzyme; RT–PCR, reverse transcriptase–polymerase chain reaction; SRCR, scavenger receptor cysteine rich. * Corresponding author. Tel: +1-608-265-2874; fax: +1-608-262-3322. E-mail address: [email protected] (M. Gorman)

the fungal b-1,3-glucans. LPS is bound by Factor C, a protein with both an LPS binding region and a serine protease catalytic domain. LPS binding causes a conformational change in Factor C, which undergoes autoactivation. Activated Factor C cleaves a second serine protease, Factor B, which then cleaves a third serine protease, proclotting enzyme (PCE). PCE cleaves coagulogen to form coagulin, which self-aggregates to create a gel-like clot that can trap pathogens. PCE is also activated by Factor G, whose serine protease subunit becomes activated after another subunit binds to b-1,3-glucan. An insect defense response that has been under intense study recently is the activation of antimicrobial peptide synthesis in Drosophila. The best characterized of the antimicrobial peptide pathways is one that controls synthesis of drosomycin, an antifungal agent (Imler and Hoffmann, 2000). The presence of certain pathogens somehow activates an unknown, extracellular serine protease. This protease activates Spa¨tzle, a ligand for the transmembrane receptor, Toll. A signal transduction cascade follows Toll activation, and ends with the up-regulation of the drosomycin gene. Surprisingly, this

0378-1119/00/$ - see front matter © 2000 Published by Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 0 0 ) 0 0 18 1 - 5

10

M.J. Gorman et al. / Gene 251 (2000) 9–17

pathway is related to the embryonic dorsal–ventral patterning pathway, in which a protease cascade ends with the activation of Spa¨tzle by Easter. The drosomycin pathway and the horseshoe crab coagulation pathways share some interesting similarities: Spa¨tzle is structurally similar to coagulogen, and Factor B, PCE, and Easter are members of the clip domain family of serine proteases (Jiang and Kanost, 2000). A third example of an insect immune pathway involving serine proteases is the synthesis of melanin (Ashida and Brey, 1997; Jiang and Kanost, 2000). Melanin formation is typically associated with cellular encapsulation of large pathogens, but melanization is its own form of encapsulation in insects, including mosquitoes, that have small numbers of circulating hemocytes (blood cells). Pathogen recognition molecules that can induce melanization have been discovered in the hemolymph of two large moth larvae, Bombyx mori and Manduca sexta, but unlike the recognition molecules of the horseshoe crab coagulation pathways, they are not serine proteases. Recognition is followed by the activation of a protease cascade, and the last protease in the cascade (the only one to be cloned) is a clip domain serine protease that activates prophenoloxidase, a principal enzyme in the melanin synthesis process. Immune processes in Anopheles gambiae are of great interest because this mosquito is one of the major vectors of human malaria. The protozoan parasites that cause malaria (Plasmodium species) must survive for longer than a week in the body of the mosquito as they go through three developmental stages, and during this time, they must endure the various immune responses that the mosquito mounts against foreign invaders (Paskewitz and Gorman, 1999). Whether the presence of malaria parasites induces hemolymph coagulation is unknown, but antimicrobial peptide synthesis and melanotic encapsulation have been observed. Also of interest is how the mosquito responds to natural bacterial and fungal pathogens as well as those that may be used in biocontrol strategies to reduce mosquito-borne transmission of disease. To identify immune-related serine proteases in Anopheles gambiae, we cloned serine protease cDNAs from adult hemolymph and further characterized the genes that appeared most interesting. Three are candidate activators of prophenoloxidase or Spa¨tzle-like ligands, and a fourth is similar to an immune-related protease from M. sexta hemocytes (Paskewitz et al., 1999; Gorman et al., 2000). A much larger serine protease designated Sp22D was not fully characterized (Gorman et al., 2000). The Sp22D transcript is about 4.6 kb. It is expressed at all developmental stages but is much more abundant in adults. Based on semi-quantitative RT–PCR, Sp22D mRNA appears to be slightly up-regulated at about 1 h after immune challenge. The partial translated sequence of Sp22D (158 amino acids

(aa)) was most similar to the sequence of Tequila, a gene with unknown function in Drosophila. This was an interesting finding, however, because the Tequila sequence contains a protein domain found in a vertebrate pathogen recognition molecule, the scavenger receptor cysteine rich (SRCR) domain. Since Sp22D is expressed constitutively in hemolymph, is up-regulated after immune challenge, and resembles Tequila, we considered it an interesting candidate immune gene meriting further study. In this report, we describe the complex domain organization of the complete Sp22D protein sequence, quantitative northern blot analysis of the effect of immune challenge, and the tissue specificity of Sp22D mRNA as determined by in situ hybridization.

2. Materials and methods 2.1. Mosquito and bacteria stocks and immune challenge A Plasmodium-susceptible strain (4arr) of A. gambiae was used for all experiments, and mosquitoes were reared as described previously (Paskewitz et al., 1999). Bacteria stocks used for immune challenge were the XL1-Blue strain of Escherichia coli (Stratagene) and the 2001 strain of Micrococcus luteus (gift from John Lindquist, University of Wisconsin). Liquid cultures of the bacteria were grown overnight, combined and centrifuged. Adult females (4 days post-eclosion) were immobilized by chilling on ice, pricked with a minuten pin dipped in the pellet of bacteria, and allowed to recover for 1 or 3 h. Clean wounding was performed with a minuten pin that had been sterilized in a flame and then dipped in sterile saline. Uninjected controls were chilled on ice and allowed to recover for 1 h. All treatments were done in triplicate. 2.2. Quantitative northern blot analysis Changes in Sp22D transcript abundance in response to immune challenge were assayed by northern blot analysis as described (Gorman et al., 2000). Signal intensity was detected with a PhosphoImager SI (Molecular Dynamics) and quantitated by the volume quantitation function of ImageQuaNT (Molecular Dynamics). Sp22D transcript abundance was normalized against the ribosomal protein gene, rpS7 (Salazar et al., 1993). A t-test was used to detect differences between treatments. 2.3. cDNA library screening and sequencing A partial cDNA of Sp22D was used to screen a LambdaZAPII cDNA library made from larvae and pupae of the G3 strain of A. gambiae as described previously (Gorman et al., 2000). Hybridizing phages

M.J. Gorman et al. / Gene 251 (2000) 9–17

were converted into plasmids by the Rapid Excision Kit (Stratagene), and the largest clone was cycle sequenced in both directions. 2.4. Amino acid sequence analysis The predicted amino acid sequence of Sp22D was scanned for various domains and motifs. Signal peptide and transmembrane domain analyses were done with SignalP v. 2.0 (Nielsen et al., 1997; Nielsen and Krogh, 1998). Putative N-linked glycosylation sites, and low density lipoprotein receptor class A (LDLRA) and SRCR domains were detected with the PROSCAN and PROFILESCAN functions of PROSITE (Bucher and Bairoch, 1994; Hofmann et al., 1999). NetOGyc 2.0 was used to predict O-linked glycosylation sites. The blastp program of the BLAST 2.0 service was used for sequence similarity searches. Sequence alignments were done with the Clustal X program (Jeanmougin et al., 1998). 2.5. In situ hybridization A DIG-labeled riboprobe of Sp22D and a sense strand negative control probe were synthesized as described in Ingham and Jowett (1997) using the DIG RNA Labeling Kit (Boehringer Mannheim). The riboprobes were hybridized at 68°C in DIG Easy Hyb (Boehringer Mannheim) to blots of total RNA from adult female mosquitoes, and high stringency washes were done. Immunological detection was done according the manufacturer’s instructions for the DIG Wash and Block Buffer Set (Boehringer Mannheim). No bands were detected by the sense strand negative control. The Sp22D riboprobe detected a single band of ~4.6 kb, as expected. In situ hybridization to tissue sections of 4 day old adult females was done as described by Ingham and Jowett (1997) with only minor modifications. Mosquitoes were fixed for several hours or overnight in 4% paraformaldehyde in 0.1 M sodium phosphate (pH 7.2)/0.25% Triton X-100. Fixed mosquitoes were embedded in paraffin by the Histopathology Lab ( University of Wisconsin, Madison). Sections of 6 mm were placed on TESPA coated slides, dried at room temperature, and left at 40°C overnight. Dewaxing, paraformaldehyde treatment, proteinase K digestion

11

(5 mg/ml for 20 min), and dehydration of sections were done without modification. Following the published protocol, very high stringency conditions were used for hybridization. The hybridization solution was 50% formamide, 5×standard saline citrate, 100 mg/ml tRNA, 50 mg/ml heparin, and 0.1% Tween-20. Hybridization was done at 65°C overnight. Washes and immunological detection were performed as described previously ( Ingham and Jowett, 1997). Some sections were stained in the presence of 10% polyvinyl alcohol in the staining solution to enhance signal. Sections were observed by bright field and DIC optics. In situ hybridization to hemocytes and fat body cells was also done. Hemolymph was collected by cutting off the tip of the proboscis of a newly eclosed female, pressing on the thorax with a cover slip, and collecting the extruded hemolymph on to the tip of a sealed pulled glass needle. Hemolymph from four mosquitoes was added to 3 ml of medium in one well of a 12-well slide (Medium 199 with Earle’s salts, -glutamine, and 2.2 g/l sodium bicarbonate supplemented with 10% lamb serum; Gibco BRL). Slides were kept in a humid chamber during collection to avoid evaporation and cell lysis. Cells were allowed to settle for 5–10 min and fixed for 20 min in 10 ml 4% paraformaldehyde in phosphatebuffered saline (PBS). Cells were washed twice with PBS for 5 min and permeabilized with a 15 min incubation in PBS/0.1% Tween-20. Prehybridization was for 1 h at 40°C. Hybridization and immunological detection methods were identical to those used for tissue sections.

3. Results and discussion 3.1. Sp22D coding sequence analysis A partial Sp22D cDNA was used to screen a cDNA library, and the largest hybridizing clone was chosen for DNA sequencing (accession no. AF117751). An open reading frame coding for a 1322 aa polypeptide was identified (data not shown). The Sp22D protein has a complex domain organization (Fig. 1). A putative signal sequence at the N-terminus suggests that Sp22D is secreted, and a lack of transmembrane helices suggests that it is not an integral membrane protein. The signal sequence is followed by a region that lacks sequence

Fig. 1. Sp22D codes for a multidomain serine protease. Domain designations and amino acid numbering are indicated. The histidine (H ), glutamine (Q), and proline (P) rich stretches and two additional regions (white boxes) lack similarity to known protein sequences.

12

M.J. Gorman et al. / Gene 251 (2000) 9–17

Fig. 2. Two putative chitin binding domains align with mosquito CBDs. The two regions of Sp22D that resemble chitin binding domains have been aligned with CBDs from Ag-Aper1, A. gambiae chitinase, two A. aegypti chitinases, and ICHIT (Dimopoulos et al., 1998; Shen and JacobsLorena, 1999). Disulfide bonding is predicted to be the same as that in Tachychitin, an antimicrobial protein in the horseshoe crab, and numbering is based on the consensus sequence constructed by Shen and Jacobs-Lorena (1999).

Fig. 3. A mucin-like region is expected to be highly glycosylated. A threonine/proline rich region is predicted to contain 42 O-glycosylated threonines and serines (in bold italics).

Fig. 4. Two regions fit the LDLRA domain signature. An alignment with the LDLRA domain signature is shown.

similarity to known protein sequences but which has histidine rich (8/22 aa), glutamine rich (23/75 aa), and proline rich (11/42 aa) stretches. Sp22D contains two regions that resemble chitin binding domains (CBDs). These were aligned with the CBDs of several mosquito proteins, including Ag-Aper1 (A. gambiae adult peritrophin 1), ICHIT, and three chitinases (Fig. 2). The most obvious amino acid conservation was for six cysteine residues, which are predicted to form disulfide bonds as shown. Shen and JacobsLorena (1999) have demonstrated a tendency for charged amino acids at consensus sequence positions 24 and 64, and for aromatic amino acid at positions 28,

29, 53, and 65. As indicated in Fig. 2, Sp22D has aspartic acid residues at the positions corresponding to 24 and 64, and phenylalanine residues at 28 and 53; however, positions 29 and 65 are occupied by neutral residues ( lysine and proline). Each putative CBD is followed by a glutamine rich region in Sp22D, a feature not present in the other CBD-containing mosquito proteins. Chitin is found in the insect exoskeleton, which is degraded by chitinases during molting, and the peritrophic matrix, of which Ag-Aper1 is a component; additionally, it is a surface component of fungi and nematode eggs (reviewed in Shen and Jacobs-Lorena, 1999). Therefore, possible candidates for Sp22D binding

Fig. 5. Two SRCR domains align with a group A consensus sequence. In the consensus sequence a=aliphatic, h=hydrophobic, –=negatively charged, p=aromatic, and · =any amino acid. Amino acids that fit the consensus are in bold.

M.J. Gorman et al. / Gene 251 (2000) 9–17

13

Fig. 6. The C-terminal region contains a serine protease domain and is similar to Drosophila GRAAL. The carboxy part of Sp22D (aa 729–1322) is aligned with the carboxy part of GRAAL (aa 848–1449). Completely conserved amino acids are indicated by asterisks (1), conservative substitutions by colons (:), and semiconservative substitutions by periods ( · ). Sequence similarity is evident for both LDLRA and SRCR domains (in bold type) as well as the serine protease domain. The predicted cleavage site (U ), the residues of the catalytic triad (n), and amino acids that suggest trypsin-like substrate specificity (8) are indicated.

14

M.J. Gorman et al. / Gene 251 (2000) 9–17

include the mosquito’s exoskeleton or peritrophic matrix but also the surface of various insect pathogens. A 250 aa segment that contains threonine repeats and multiple prolines is likely to be highly O-glycosylated and thus resembles a mucin domain (Fig. 3). One function of O-glycosylation is to induce a rigid, extended conformation in the glycosylated domain (Jentoft, 1990). Mucin domains are often found in proteins associated with the cell surface, and they may act to position part of a protein beyond the glycocalyx to allow for interactions with other proteins, or mucin domains may themselves interact with other proteins (Shimizu and Shaw, 1993). The Drosophila scavenger receptor, dSR-CI, and insect hemomucin are both cell surface proteins with mucin domains and putative roles in insect immunity (Pearson et al., 1995; Theopold et al., 1996). Insect mucins are also associated with the midgut epithelial cell membranes and the peritrophic matrix ( Wang and Granados, 1997; Shen et al., 1999). Two domains correspond well to the LDLRA domain signature ( Fig. 4). LDLRA domains have been found in many proteins with diverse functions, including some involved in vertebrate immune responses. In many cases, the function of these domains is unknown, but for the LDLR family of receptors, they are the ligand binding domains. Known ligands for this family are structurally diverse, and include apolipoproteins, vitellogenin, a-2 macroglobulin complexes, and circumsporozoite protein, the most abundant protein on the surface of malaria sporozoites (Hussain et al., 1999). Interspersed with the LDLRA domains are two domains that align well with the scavenger receptor cysteine rich (SRCR) group A consensus sequence (Fig. 5). SRCR domains are present in a variety of cell surface and secreted proteins, most with an immune function (Resnick et al., 1994). The two domains from Sp22D have six cysteines, and thus belong to group A. The second domain is more typical and is similar to domains in SpSRCR1, a protein found in sea urchin coelomocytes, and SRCR-SCR-Car, a putative aggregation receptor from a marine sponge (Blumbach et al., 1998; Pancer et al., 1999). The functions of SRCR domains are unknown, but they are thought to mediate ligand binding. The C-terminal domain appears to be a typical serine protease catalytic domain (Fig. 6; Gorman et al., 2000). The sequence of a consensus cleavage site (R/VV ) suggests that Sp22D is activated by a serine protease with trypsin-like substrate specificity (i.e., with a preference for cleaving after the basic amino acids arginine and lysine). Three conserved amino acids (Asp1259, Gly1287, and Gly1297) suggest that Sp22D itself has trypsin-like substrate specificity. The putative catalytic domain is most similar to two unpublished Drosophila protein sequences, GRAAL (accession no. AJ251803.1) and Tequila (accession no. AF044207), which probably

correspond to the same gene. From the cleavage site to the C-terminus, Sp22D and GRAAL are 76% (189/249 aa) similar and 51% (128/249 aa) identical; notable sequence similarity extends through the SRCR and LDLRA domains (Fig. 6), and includes the two CBDs but not the remaining Sp22D domains (not shown). The extensive similarity between Sp22D and GRAAL/ Tequila suggest that they may be homologs. 3.2. Sp22D transcript abundance increases 1.5-fold after immune challenge Previously, we used semiquantitative RT–PCR to demonstrate a slight induction in Sp22D transcript abundance in response to injection of bacteria (Gorman et al., 2000). To quantify the increase, we performed quantitative northern blot analysis. These results demonstrate that Sp22D mRNA is approximately 1.5 times more abundant in mosquitoes 1 h after injection than in uninfected controls (Fig. 7). By 3 h post-injection, there was no significant difference between injected and control mosquitoes. Unlike the response of several other serine proteases and other immune related genes, wound-

Fig. 7. Sp22D transcript abundance increases 1.5-fold after injection of bacteria. RNA was isolated from uninjected (C ), saline injected (S ), and bacteria injected (B) mosquitoes at 1 and 3 h post-injection. Northern blot results are shown in the top panel, and a graph of the normalized transcript abundance is presented in the lower panel. Error bars indicate standard deviations.

M.J. Gorman et al. / Gene 251 (2000) 9–17

15

Fig. 8. Sp22D mRNA is present in the midgut, hemocytes, and fat body cells. In situ hybridization to whole body tissue sections of adults and hemocyte/fat body cell preparations detected Sp22D in the midgut (A), hemocytes (C–E), and fat body cells (F ). Negative control probes generated no signal (e.g., B). All hemocyte types had detectable levels of staining, including plasmatocytes (C ), granular cells (D), and prohemocytes ( E ).

16

M.J. Gorman et al. / Gene 251 (2000) 9–17

ing in the absence of bacteria elicited no increase. The peak of expression corresponds in time with the brief presence of M. luteus in the injected mosquito but not with the presence of E. coli, which persist in the mosquito for well over a week (Gorman and Paskewitz, 2000). Because Sp22D contains putative CBDs and chitin oligomers are known immune elicitors (Furukawa et al., 1999), we looked for but did not detect changes in signal intensity after injection of chitin hexamers (unpublished observations).

secrete the protein into the hemolymph, and by the midgut, which may secrete the protein into the gut lumen or into the space between the midgut and its basal lamina. As a putative pathogen recognition molecule, it may bind to a pathogen surface, resulting in autoactivation and the induction of a defense response. The use of recently produced antisera to Sp22D (Danielli et al., 2000) will facilitate analysis of the biochemistry and function of this molecule.

3.3. Tissue localization of Sp22D mRNA

Acknowledgements

In situ hybridization was done to determine the tissue specificity of Sp22D expression in adult mosquitoes (Fig. 8). The strongest signal was observed in the posterior midgut ( Fig. 8A), but the entire midgut was stained (not shown). The only other cells with detectable levels of staining were hemocytes and fat body cells (Fig. 8C– F ). The hemocyte types of A. gambiae have not been described; however, comprehensive descriptions of the hemocyte morphology of other mosquito species (Aedes aegypti, Culex quinquefasciatus, and Anopheles albimanus) have been published ( Kaaya and Ratcliffe, 1982; Hernandez et al., 1999). Consistent with previous descriptions, we observed cells with morphology similar to plasmatocytes, granular cells, and prohemocytes, but not thrombocytoids, spindle cells, or oenocytoids. Sp22D signal was clearly visible in all unspread hemocytes (data not shown), but was faint in spread cells, especially granular cells (Fig. 8C, D).

We are grateful to Peter Bryant, Tatsuhiro Shibata, Jules Hoffmann and Fotis Kafatos for sharing unpublished results regarding Tequila, GRAAL, and Sp22D. We thank John Lindquist for the stock of M. luteus. We regret that we were unable to cite all of the relevant primary literature due to a strict citation limitation. This work was supported by NIH Tropical Disease Research Unit grant AI28781.

4. Conclusions Sp22D codes for a putative trypsin-like serine protease with a complex domain organization. Three characteristics suggest that it may play a role in immunity. First, most of its domains are similar to regions of immune-related proteins. Second, Sp22D transcript abundance increases after bacterial infection but not wounding. Third, Sp22D mRNA is present in hemocytes, fat body, and midgut epithelium, which are all tissues with important immune functions. The domain organization of Sp22D is unlike any protease with a known function. However, like the LPS and b-1,3-glucan recognition proteases in horseshoe crab hemolymph, Sp22D resembles a pathogen recognition protein. The putative CBDs, which could bind to fungal or nematode surfaces, are the most obvious candidates for pathogen binding domains. In addition, because the vertebrate LDL receptor-related protein interacts with circumsporozoite protein, an exciting possibility is that Sp22D may interact with this protein on the surface of malaria parasites. Sp22D mRNA is constitutively expressed by hemocytes and fat body, which likely

References Ashida, M., Brey, P.T., 1997. Recent advances in research on the insect prophenoloxidase cascade. In: Brey, P.T., Hultmark, D. ( Eds.), Molecular Mechanisms of Immune Responses in Insects. Chapman & Hall, London, pp. 133–172. Blumbach, B., Pancer, Z., Diehl-Seifert, B., Steffen, R., Mu¨nkner, J., Mu¨ller, I., Mu¨ller, W.E.G., 1998. The putative sponge aggregation receptor: isolation and characterization of a molecule composed of scavenger receptor cysteine-rich domains and short consensus repeats. J. Cell Sci. 111, 2635–2644. Bucher, P., Bairoch, A., 1994. A generalized profile syntax for biomolecular sequences motifs and its function in automatic sequence interpretation. Proceedings of the Second International Conference on Intelligent Systems for Molecular Biology. AAAI Press, Menlo Park, California. Danielli, A., Loukeris, T.G., Laguexy, M., Mu¨ller, H.-M., Richman, A., Kafatos, F.C., 2000. A modular chitin-binding protease associated with hemocytes and hemolymph in the mosquito Anopheles gambiae. Proc. Natl. Acad. Sci. USA. in press. Dimopoulos, G., Seeley, D., Wolf, A., Kafatos, F.C., 1998. Malaria infection of the mosquito Anopheles gambiae activates immuneresponsive genes during critical transition stages of the parasite life cycle. EMBO J. 17, 6115–6123. Furukawa, S., Taniai, K., Yang, J., Shono, T., Yamakawa, M., 1999. Induction of gene expression of antibacterial proteins by chitin oligomers in the silkworm, Bombyx mori. Insect Mol. Biol. 8, 145–148. Gorman, M.J., Paskewitz, S.M., 2000. Persistence of infection in mosquitoes injected with bacteria. J. Invertebr. Pathol. In press. Gorman, M.J., Andreeva, O.V., Paskewitz, S.M., 2000. Molecular characterization of five serine protease genes cloned from Anopheles gambiae hemolymph. Insect Biochem. Mol. Biol. 30, 35–46. Hernandez, S., Humbert, L., Rodriquez, M.H., Torres, J.A., MartinezPalomo, A., Tsutsumi, V., 1999. Morphological and cytochemical characterization of female Anopheles albimanus (Diptera: Culicidae) hemocytes. J. Med. Entomol. 36, 426–434. Hofmann, K., Bucher, P., Falquet, L., Bairoch, A., 1999. The PROSITE database, its status in 1999. Nucleic Acids Res. 27, 215–219.

M.J. Gorman et al. / Gene 251 (2000) 9–17 Hussain, M.M., Strickland, D.K., Bakillah, A., 1999. The mammalian low-density lipoprotein receptor family. Annu. Rev. Nutr. 19, 141–172. Imler, J.L., Hoffmann, J.A., 2000. Signaling mechanisms in the antimicrobial host defense of Drosophila. Curr. Opin. Microbiol. 3, 16–22. Ingham, P.W., Jowett, T., 1997. Localization of RNA transcripts by in situ hybridization. C. In: Crampton, J.M., Beard, C.B., Louis, C. (Eds.), The Molecular Biology of Disease Vectors. Chapman & Hall, New York, pp. 268–282. Iwanaga, S., Kawabata, S., Muta, T., 1998. New types of clotting factors and defense molecules found in horseshoe crab hemolymph: their structures and functions. J. Biochem. 123, 1–15. Jeanmougin, F., Thompson, J.D., Gouy, M., Higgins, D.H., Gibson, T.J., 1998. Multiple sequence alignment with Clustal X. Trends Biochem. Sci. 23, 403–405. Jentoft, N., 1990. Why are proteins O-glycosylated? Trends Biochem. Sci. 15, 291–294. Jiang, H., Kanost, M.R., 2000. The clip-domain family of serine proteinases in arthropods. Insect Biochem. Mol. Biol. 30, 95–105. Kaaya, G.P., Ratcliffe, N.A., 1982. Comparative study of hemocytes and associated cells of some medically important dipterans. J. Morphol. 173, 351–365. Nielsen, H., Engelbrecht, J., Brunak, S., von Heijne, G., 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10, 1–6. Nielsen, H., Krogh, A., 1998. Prediction of signal peptides and signal anchors by a hidden Markov model. Proceedings of the Sixth International Conference on Intelligent Systems for Molecular Biology. AAAI Press, Menlo Park, California. Pancer, Z., Rast, J.P., Davidson, E.H., 1999. Origins of immunity: transcription factors and homologues of effector genes of the verte-

17

brate immune system expressed in sea urchin coelomocytes. Immunogenetics 49, 773–786. Paskewitz, S.M., Gorman, M.J., 1999. Mosquito immunity and malaria parasites. Am. Entomol. 45, 80–94. Paskewitz, S.M., Reese-Stardy, S.R., Gorman, M.J., 1999. An easterlike serine protease from Anopheles gambiae exhibits changes in transcript abundance following immune challenge. Insect Mol. Biol. 8, 329–337. Pearson, A., Lux, A., Krieger, M., 1995. Expression cloning of dSR-CI, a class C macrophage-specific scavenger receptor from Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 92, 4056–4060. Resnick, D., Pearson, A., Krieger, M., 1994. The SRCR superfamily: a family reminiscent of the Ig superfamily. Trends Biochem. Sci. 19, 5–8. Salazar, C.E., Mills-Hamm, D., Kumar, V., Collins, F.H., 1993. Sequence of a cDNA from the mosquito Anopheles gambiae encoding a homologue of human ribosomal protein S7. Nucleic Acids Res. 21, 4147. Shen, Z., Jacobs-Lorena, M., 1999. Evolution of chitin-binding proteins in invertebrates. J. Mol. Evol. 48, 341–347. Shen, Z., Dimopoulos, G., Kafatos, F.C., Jacobs-Lorena, M., 1999. A cell surface mucin specifically expressed in the midgut of the malaria mosquito Anopheles gambiae. Proc. Natl. Acad. Sci. USA 96, 5610–5615. Shimizu, Y., Shaw, S., 1993. Mucins in the mainstream. Nature 366, 630–631. Theopold, U., Samakovlis, C., Bromage, H.E., Dillon, N., Axelsson, B., Schmidt, O., Tempst, P., Hultmark, D., 1996. Helix pomatia lectin, an inducer of Drosophila immune response binds to hemomucin, a novel surface mucin. J. Biol. Chem. 271, 12708–12715. Wang, P., Granados, R.R., 1997. Molecular cloning and sequencing of a novel invertebrate intestinal mucin cDNA. J. Biol. Chem. 272, 16663–16669.