A Type I Peritrophic Matrix Protein from the Malaria VectorAnopheles gambiae Binds to Chitin

Upon feeding, mosquito midguts secrete the peritrophic matrix (PM), an extracellular chitin-containing envelope that completely surrounds the blood meal. Because the malaria parasite must cross the PM to complete its life cycle in the mosquito, the PM is a potential barrier for malaria transmission. By antibody screening of an expression library we have identified and partially characterized a cDNA encoding a putative PM protein, termed Anopheles gambiae adult peritrophin 1 (Ag-Aper1).Ag-Aper1 is the first cloned PM gene from a disease vector. Northern analysis detected an abundant Ag-Aper1 transcript only in the adult gut, and not in any other tissues or at any other stages of development. The predicted amino acid sequence indicates that it has two tandem chitin-binding domains that share high sequence similarity with each other and also with the chitin-binding domain of an adult gut-specific chitinase from the same organism. The presumed ability of Ag-Aper1 to bind chitin was verified by a functional assay with the baculovirus-expressed recombinant protein. Ag-Aper1 did bind to chitin but not to cellulose, indicating that Ag-Aper1 binds chitin specifically. The double chitin-binding domain organization of Ag-Aper1 suggests that each protein molecule is able to link two chitin polymer chains. Hence, this protein is likely to act as a molecular linker that connects PM chitin fibrils into a three-dimensional network.

The peritrophic matrix (PM) 1 is an extracellular layer that surrounds the food bolus in the guts of most arthropods (1)(2)(3)(4). All PMs are composed of chitin, proteins, and proteoglycans (1)(2)(3)(4). There are two types of PMs in insects. The type 1 PM is thick (usually between 2 and 20 m) and is secreted by most midgut epithelial cells. Secretion of the type 1 PM is triggered by the dramatic distention of the midgut epithelium that occurs during ingestion of a meal (5). Many adult blood-feeding insects, including mosquitoes, form type 1 PMs after blood feeding. Two-dimensional protein gel analysis indicated that black flies have only two major proteins (6), whereas mosquitoes have a considerably more complex mixture of proteins (7). So far no type I PM proteins have been cloned. The type 2 PM is thin (usually less than 1 m) and is synthesized continuously from a specialized organ, the cardia (or proventriculus), which is located at the junction of the foregut and midgut. Mosquitoes and many other hematophagous insects produce type 2 PM during larval life and type 1 PM during adult life. Several type 2 larval PM proteins have been cloned (8 -10). Little is known about the structural and biochemical differences between these two very different types of PM.
The function of the PM has not been clearly established. It is likely that the PM provides protection for midgut epithelial cells from damage from food particles, prevents infection from viruses, bacteria, and other pathogens, and also facilitates digestion by partitioning digestive enzymes and the ingested food between the endo-and the ecto-peritrophic compartments (1)(2)(3)(4). The PM may also function as a partial barrier to toxins and other macromolecules. For example, exogenous chitinase significantly increases the toxicity of Bacillus thuringiensis' delta toxin in Spodoptera littoralis, presumably by increasing PM permeability (11).
Human malaria is caused by protozoan parasites, Plasmodium sp. Malaria is one of the most serious global health problems, causing about 2 million deaths annually. Anopheline mosquitoes are the sole vectors for human malaria. Upon ingestion of an infected blood meal, Plasmodium gametes mate, develop into ookinetes, and about one day later, the ookinetes penetrate the PM and the gut epithelial cells. After traversing the gut epithelium and reaching the gut basal lamina, the ookinetes develop into oocysts. Each oocyst eventually releases thousands of sporozoites that invade the salivary glands and infect the next person when the mosquito bites again. The PM is a thick and dense layer that the Plasmodium parasite must traverse. Although it is known that the parasite secretes a chitinase to facilitate the penetration of the PM (12), it is unclear what role, if any, the PM plays in triggering chitinase secretion or whether secretion of other hydrolytic enzymes by the parasite is also required. The resolution of these issues will require the elucidation of the molecular composition and structure of the Anopheles PM. Here, we report for the first time, the cloning of a PM protein from a disease vector and propose that this PM protein plays a major role in the spatial organization of the PM.

MATERIALS AND METHODS
Preparation of the PM-An A. gambiae (G3 strain) colony was maintained as described previously (13). Adult mosquitoes (about 5 days old) were fed with a protein-free latex meal as described previously (7) to induce PM formation. PMs were dissected in PBS buffer 30 min after mosquitoes had a latex meal. Dissected PMs were transferred to a 1.5-ml centrifuge tube (kept on ice) with a fine glass pipette. After collecting PMs for about 30 min, the tube was centrifuged for 1 min at * This work was supported by grants from the John D. and Catherine T. MacArthur Foundation and from the NIAID, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF030431.
12,000 ϫ g, the residual dissection buffer was removed, and the PM pellet was stored at Ϫ80°C.
Preparation of PM Antibodies-Five BALB/cJ mice were used to produce polyclonal antisera. About 200 PMs dissected from latex-fed mosquitoes were homogenized in 250 l of PBS and mixed with an equal volume of adjuvant. Complete Freund's adjuvant was used for the initial immunization and incomplete Freund's adjuvant for all boosts. For each immunization, about 100 l of the antigen/adjuvant mixture (containing 40 PMs) per mouse were injected intraperitoneally. The antisera were collected after 6 injections at 2-week intervals.
SDS-PAGE Analysis-Dissected PMs were rinsed briefly with 50 mM Tris-HCl (pH 7.5) and centrifuged at 12,000 ϫ g for 10 min. The pellet was then homogenized in H 2 O containing 1 mM phenylmethylsulfonyl fluoride using a disposable plastic pestle. After vortexing for 2 min, the sample was centrifuged at 12,000 ϫ g for 10 min. The supernatant was the water-soluble fraction, and the pellet was the water-insoluble fraction. Both fractions were boiled for 10 min with SDS sample buffer containing mercaptoethanol (7). The proteins in both fractions were analyzed by SDS-PAGE (12% acrylamide) as described by Laemmli (14).
Western Blot Analysis of PM Proteins-Five guts from 4th-instar larvae or from adult mosquitoes 1, 3, 5, 7, and 9 d after eclosion, were dissected and kept frozen at Ϫ80°C. After adding 50 l of 2ϫ SDS sample buffer (14), each sample was boiled for 5 min. An aliquot from each sample (10 l) was analyzed by SDS-PAGE. After electrophoresis, the proteins were blotted onto a nitrocellulose membrane (Micron Separation, Inc., MA). PM proteins were detected with the anti-PM antiserum. An alkaline phosphatase-conjugated goat anti-mouse IgG was used as the secondary antibody, and a chromogenic reaction was used to detect the bound antibodies according to the manufacturer's recommendations (Schleicher & Schuell).
cDNA Library Screening-The anti-PM antiserum was used to screen an A. gambiae midgut cDNA expression library (13) with an immunoscreening kit (Stratagene, CA). A phagemid (pBluescript SK) was obtained from each purified bacteriophage by in vivo excision following the manufacturer's recommendations (Stratagene). Both strands of the cDNA insert were sequenced with an automated sequencer (Applied Biosystem).
Northern Analysis-Total RNA was isolated using the method described by Chomczynski and Sacchi (15). RNA samples were fractionated on a 1.2% agarose gel containing 1 M formaldehyde, and then transferred onto a GeneScreen nylon membrane. A probe was made by random labeling of the 774-nucleotide Ag-Aper1 cDNA insert with [␣-32 P]dCTP (16). Church's buffer was used for both prehybridization and hybridization. Hybridization was carried out at 65°C. The final wash was done in 0.1ϫ SSC at 60°C.
PCR Amplification of the Ag-Aper1 Genomic DNA Fragment-Two primers (Aper1N, 5Ј-ATGAAAGTTTCTGCCAGTTTGGTA; Aper1C, 5Ј-AATCTTTCCACCCTGAAGCATGCC), located in the 5Ј-and 3Ј-untranslated sequences, respectively, were designed to amplify the genomic DNA fragment of Ag-Aper1. The PCR was carried out for 32 cycles at 94°C for 40 s, 60°C for 40 s, and 72°C for 1.5 min. A. gambiae genomic DNA was used as a template. The PCR products were analyzed by agarose gel electrophoresis. Further identification of the PCR products as Ag-Aper1 was accomplished by Southern hybridization with the cloned cDNA.
Southern Analysis-Genomic DNA from whole adult mosquitoes was isolated according to Campbell et al. (17). DNA (5 g) was digested with either EcoRI or HindIII and fractionated by agarose gel electrophoresis. The DNA in the agarose gel was alkali-denatured, neutralized with HCl, and then transferred onto a GeneScreen nylon membrane. Two probes, one made from the cDNA insert and the other from a PCR genomic fragment, were used to hybridize to either of two identical blots. The hybridization and washing conditions were the same as for the Northern analysis described above.
Recombinant Protein Expression and Chitin-binding Assay-The Ag-Aper1 was expressed by recombinant baculovirus in insect cells using the Baculovirus BacPack expression system (CLONTECH, CA). The manufacturer's recommendations were followed. To obtain the transfer plasmid pBacPack9-Ag-Aper1, the Ag-Aper1 cDNA was excised from the plasmid by digestion with BamHI and XhoI, and cloned into the transfer plasmid vector pBacPak9 (CLONTECH). The recombinant baculovirus was generated by co-transfection of pBacPack9-Ag-Aper1 and baculovirus DNA (provided by CLONTECH) into Sf21 cells. To express Ag-Aper1, recombinant viruses were used to infect 10 7 log growth phase Sf21 cells. The cells were collected by centrifugation 60 h postinfection. The cells were lysed in 6 ml of PBS ϩ 0.1% Triton X-100, sonicated, and the insoluble debris were removed by centrifugation at 10,000 ϫ g for 5 min. About 0.5 ml (settled volume) of chitin (New England BioLabs) or cellulose (Sigma) was washed with PBS and mixed with 2.5 ml of cleared cell lysate. After rocking for 30 min, the matrices were packed into columns. The columns were washed with 2 ml of PBS and then eluted sequentially with 0.5 ml of 50 mM NaAc (pH 5.0) and 0.5 ml with 50 mM glycine-HCl (pH 2.0). The matrices were then stripped by boiling with 0.5 ml of 0.5% SDS. The protein composition of each fraction was analyzed by SDS-PAGE and detected either by Coomassie Blue staining or by Western blotting with the PM antiserum.

SDS-PAGE Analysis of PM Proteins-To obtain PMs without
contaminating food proteins, mosquitoes were fed with a protein-free latex meal. We found that the PM formed very quickly, within 30 min of feeding. This is much earlier than the previously reported 13-h lag for blood-fed mosquitoes (18). The protein composition of the water-insoluble and water-soluble PM fractions was analyzed by SDS-PAGE (Fig. 1). Proteins in the 60-kDa range are the main component of the insoluble fraction, whereas a 180-kDa protein and some smaller proteins appear to compose the soluble fraction. Because adult mosquito PM proteins are glycosylated (7), it is difficult to predict the actual size or number of protein species in the PM. If proteins are glycosylated to different extents, the PM protein complexity could be lower than the number of bands in the gel.
Western Analysis of Anopheles PM Proteins-Mice were immunized with total adult PM proteins, and the antiserum was used to detect PM proteins in gut extracts. Three major bands were detected (Fig. 2). One is approximately 180 kDa, which may correspond to the 180-kDa water-extractable protein (Fig.  1, lane 2). The two other major bands of around 60 kDa may correspond to the major insoluble proteins shown in lane 1 of Fig. 1. The antiserum cross-reacts with some proteins from the larval gut (Fig. 2, lane 1). Although one of the larval proteins co-migrates with the ϳ180 kDa adult PM protein, another major cross-reacting protein migrates with an apparent molecular mass of over 220 kDa. There were no cross-reacting larval PM proteins in the range of 60 kDa. The results of the experiment in Fig. 2 also indicate that PM proteins are synthesized within the first day of adult life. Although the abundance of the 180-kDa protein does not appear to increase with age, the group of ϳ60-kDa proteins appears to increase in amount until day 4 after eclosion, when mosquitoes reach maturity for blood feeding. The finding that PM proteins are stored in the gut epithelium before blood feeding is consistent with the observation that the PM forms within minutes of feeding.
Cloning of a cDNA That Encodes a PM Protein-A gut cDNA expression library was screened with the anti-PM serum. A cDNA clone, named Ag-Aper1 for Anopheles gambiae adult peritrophin 1, was obtained. It has an insert of 0.77 kilobase pairs and an open reading frame encoding 153 amino acids (Fig. 3). The cDNA has 27 base pairs of 5Ј-untranslated sequence and 254 base pairs of 3Ј-untranslated sequence. The polyadenylation signal (AATAAAA) is found 30 base pairs upstream of the poly(A) end. The deduced amino acid sequence has a highly hydrophobic N terminus, as expected for a secretory protein. According to von Heijne's rules (19), the signal peptide cleavage site is predicted to be after Ala-17. The conceptual translation of the Ag-Aper1 cDNA predicts a protein of 15 kDa, pI of 4.0 with a high content of proline (17.5%), cysteine (8.8%), and aspartic acid (9.5%). The high proline content suggests that Ag-Aper1 has little secondary structure.
Western analysis with the same antiserum used to clone Ag-Aper1 did not detect a major protein at ϳ15 kDa (Fig. 1). This could be because of protein glycosylation.
Domain Structure of Ag-Aper1-The Ag-Aper1 amino acid sequence is internally repeated. The amino acid residues of the first and second halves of the protein are 41% identical and 60% conserved (Fig. 4), indicating that the protein is composed of two similar domains. Interestingly, the two domains also share high sequence similarity with the chitin-binding domain of AgChi-1, the gut-specific chitinase from A. gambiae (20) (Fig.  4). Sequence comparisons with different data bases by BLAST searching (21) revealed that Ag-Aper1 also shares low but significant similarity with several groups of proteins that interact (or are presumed to interact) with chitin. These include chitinases from nematodes (Caenorhabditis elegans (Swiss-Prot accession no. Q11174, Nhan, N); Brugia malayi (22)), arthropods (Penaeus japonicus (23), Manduca sexta (24)), and insect larval type 2 PM proteins (Lucilia cuprina peritrophins 44 and 95 (8 -9), Trichoplusia ni mucin (10)). The larval type 2 PM proteins have 5 putative chitin-binding domains, whereas Ag-Aper1 has only two.

FIG. 2. Western analysis of PM proteins from mosquito midguts.
Midgut proteins were fractionated by SDS-PAGE, blotted onto a nitrocellulose membrane, and detected with a mouse polyclonal antiserum prepared against total PM proteins. Lane 1, guts from 4th-instar larvae; lanes 2-6, midguts from sugar-fed females aged for 1, 3, 5, 7, and 9 days, respectively. Migration of protein size markers (in kDa) is indicated on the left. Expression Pattern of Ag-Aper1 mRNA-Northern analysis was used to measure Ag-Aper1 mRNA expression at different developmental stages and in different tissues. The Ag-Aper1 mRNA was only detected in adult guts. RNA from adult carcass (whole body minus gut), pupae, or larvae yielded no detectable signal (Fig. 5). Therefore, Ag-Aper1 is an adult gut-specific gene. In the adult midgut, a strong signal was detected 1 day after eclosion (Fig. 6) even though the mRNA appears to be absent from pupae (Fig. 5). Thus, it is likely that Ag-Aper1 transcription is turned on soon after eclosion. Ag-Aper1 mRNA abundance does not change appreciably with age of the mosquito or as a function of time after feeding (Fig. 6). This expression pattern is in agreement with the profile of PM proteins detected by Western analysis (Fig. 2) but different from that of many digestive enzymes that are induced after blood feeding (25)(26)(27)(28).
Genomic Organization of Ag-Aper1-Ag-Aper1 gene copy number was estimated by Southern blotting analysis of genomic DNA digested with either EcoRI or HindIII (Fig. 7). Two bands in the 6-kilobase pair region were detected in the EcoRI digest, whereas only one band was detected with the HindIII digest. Although the Ag-Aper1 cDNA does not have an EcoRI site, it is possible that the Ag-Aper1 gene has an intron with an EcoRI site. To investigate this possibility, we performed PCR with primers located at the 3Ј-and 5Ј-ends of the cDNA sequence. The size of the PCR product was the same using either the genomic DNA or the cDNA clone as the template (result not shown), indicating that no introns are present within this interval. When the Southern blot of Fig. 7 was rehybridized with the PCR product of genomic DNA it gave the identical pattern as with the cDNA probe (result not shown). Therefore, the Southern blot analysis suggests that they may be two genes (similar or identical) located within 7 kilobase pairs. However, we cannot rule out the possibility that there is only one Ag-Aper1 gene per haploid genome and that because of sequence polymorphism, one allele has an internal EcoRI site although the other does not.
The Recombinant Ag-Aper1 Protein Specifically Binds to Chitin-The structure of Ag-Aper1 (Fig. 4) suggested that this is a chitin-binding protein. This premise was tested with a functional assay. Ag-Aper1 was expressed in an insect cell line from a recombinant baculovirus and detected with the antiperitrophic matrix serum as a protein of ϳ30 kDa (Fig. 8, lanes  2 and 9 in the lower panels). An extract from cells infected with a recombinant baculovirus expressing the green fluorescent protein was analyzed as a control (Fig. 8, lanes 1 and 8). The antiserum did not detect any proteins, even though more extract was loaded on the gels. This indicates that the antiserum detected specifically the recombinant Ag-Aper1 protein. The apparent molecular weight of the recombinant protein is higher than that calculated (15 kDa). This may be because of protein glycosylation.
The following experiments were to investigate whether Ag-Aper1 binds to chitin, and if so, whether binding is specific for this polymer. Total soluble proteins from cells expressing Ag-Aper1 were mixed with an insoluble matrix made of chitin (a polymer of N-acetylglucosamine subunits) or cellulose (a polymer of glucose subunits). The recombinant Ag-Aper1 protein appeared to bind to chitin efficiently, because no Ag-Aper1 was detected in the flow-through (Fig. 8, lane 3). The chitin matrix did not absorb proteins nonspecifically since most proteins were present in the flow-through (Fig. 8, upper panel, lanes 2  and 3). The bound proteins were eluted with a pH 2 buffer (Fig.  8, lane 6), but not with a pH 5 buffer (Fig. 8, lane 4). The interaction of Ag-Aper1 with chitin appears to be specific, because no Ag-Aper1 binding was detected to a cellulose matrix (Fig. 8, right panel). Virtually all of the Ag-Aper1 was detected in the flow-through from the cellulose matrix (Fig. 8, lower  panel, lane 10). Two bands were detected in the pH 2 eluate by Coomassie Blue staining (Fig. 8, upper panel, lane 6). One migrates at about 30 kDa, which coincides with the mobility of the band detected by Western analysis (Fig. 8, lower panels). Therefore, this band is likely Ag-Aper1. The upper band also FIG. 5. Tissue and stage specificity of Ag-Aper1 expression. About 5 g of total RNA isolated from different tissues and developmental stages were fractionated by denaturing agarose gel electrophoresis and transferred onto a nylon membrane. A 32 P-labeled Ag-Aper1 cDNA insert (Fig. 3) was used to probe the Northern blot. An autoradiograph is shown in the upper panel. The weaker upper band was not detected reproducibly. Lc, carcasses (whole body minus gut) from 4th-instar larvae; Lg, guts from 4th-instar larvae; P, whole pupae; Ac, adult carcasses; Ag, adult guts. A photograph of the ethidium bromide-stained gel before transfer to the membrane is shown in the lower panel.
FIG. 6. Temporal expression of Ag-Aper1. Total RNA was isolated from 30 midguts from sugar-fed mosquitoes of different ages (days after eclosion are indicated) or from 5-day-old mosquitoes at different times after feeding on a ␥-globulin meal (hours after feeding are indicated). One-fourth of each RNA sample was fractionated by denaturing agarose gel electrophoresis and transferred onto a nylon membrane. The blot was hybridized with a 32 P-labeled Ag-Aper1 probe. An autoradiograph is shown in the upper panel. A photograph of the ethidium bromidestained gel before transfer to the membrane is shown in the lower panel. appears to be a chitin-binding protein, and originates either from the Sf21 cells or from the baculovirus itself. Boiling with SDS did not release a significant amount of protein from either matrix (Fig. 8, lanes 7 and 13). We conclude that Ag-Aper1 binds to chitin but not to cellulose.

DISCUSSION
Significance of the Ag-Aper1 Two-domain Structure-Chitin is a linear polymer of N-acetylglucosamine and a key structural component of the PM. In the presence of inhibitors of chitin synthesis, such as Dimilin or Polyoxin D, PM formation is repressed or completely blocked (1). Moreover, feeding mosquitoes with a meal containing chitinase completely blocks PM formation (29). Thus, chitin is an essential component of the PM. However, it is unknown how chitin fibers interact with proteins to form the PM. The characterization of the Ag-Aper1 gene has shed some light on this issue. Ag-Aper1 has two chitin-binding domains connected by a short linker sequence. This double-domain structure suggests that each Ag-Aper1 molecule can bind two chitin fibers and thus form a threedimensional chitin-Ag-Aper1 network (Fig. 9). Growth of the PM would occur by addition of new chitin fibers and Ag-Aper1 to the existing network, as illustrated in Fig. 9.
The two chitin-binding domains of Ag-Aper1 are highly homologous to the chitin-binding domain of the Anopheles gutspecific chitinase (Fig. 4). In the case of chitinase, the chitinbinding domain may enhance enzyme function by facilitating the binding of the enzyme to the chitin fibers (Fig. 9). Although the Ag-Aper1 is presumed to be an essential component of PM synthesis and structure, the gut chitinase would play an inverse but equally important role of controlling PM turnover and porosity. Chitin hydrolysis is likely to serve two physiologically important roles: 1) control of PM thickness by balancing PM growth with degradation; and 2) increase PM porosity to support the heavy "molecular traffic" that occurs during digestion, digestive enzymes crossing the PM from the epithelium toward the food bolus and hydrolytic products traveling in the opposite direction.
The Chitin-binding Domain As a Basic Module in Evolution-The Ag-Aper1 protein has two chitin-binding domains that share significant amino acid sequence similarity. It is unlikely that these two domains have arisen independently by convergent evolution. Rather, they probably resulted from gene duplication. Furthermore, the high similarity among the chitin-binding domains of Ag-Aper1 and AgChi-1 suggests that they may be derived from a common ancestor. We propose that the chitin-binding domain is a basic module that combined with other protein sequences to generate new function or modify existing function.
Ag-Aper1 As a Candidate Antigen for a Transmission-blocking Vaccine-Elucidation of PM structure may help clarify how Plasmodium manages to pass through the PM. This knowledge could be used to develop novel strategies to control malaria transmission. For instance, the blood of sheep vaccinated with the L. cuprina larval PM (type 2) hinders larval development, probably because of the blockage of PM pores (32,33). Although the adult mosquito forms a different type of PM (type 1), it is conceivable that antibodies to adult Anopheles PM proteins will have a similar effect on the PM. This effect could impede Plasmodium to cross the PM, prolong digestion time, and perhaps shorten the life span of the mosquito. The combination of antibodies against PM proteins and gut chitinase may have even more pronounced effects. Inhibition of chitinase may result in a thicker PM, which in turn inhibits Plasmodium development (20,34). The cloning of genes encoding PM proteins may provide the tools to test these ideas.
Use of the Ag-Aper1 Promoter to Drive the Expression of Foreign Anti-malaria Genes in Transgenic Mosquitoes-With the advance of germ line transformation of mosquitoes (35), new approaches became possible for the control of malaria transmission. For instance, it may be possible to genetically engineer mosquitoes that express foreign genes that are toxic to Plasmodium (36). This approach requires the identification of a promoter capable of driving the expression of a foreign gene in the right tissue and at the right time. The gut is an excellent target, because Plasmodium must complete a complex developmental program while exposed to the relatively confined environment of the gut lumen. Most gut-specific genes that  have been characterized to date reach peak expression and secrete their protein products into gut lumen relatively late after a blood meal, at a time when the lumen (and any parasites therein) is shielded from the secretory epithelium by a thick PM. Ag-Aper1 is different. Large amounts of Ag-Aper1 mRNA accumulate prior to the blood meal (Fig. 6), and its protein is presumed to be stored and released immediately after feeding. Thus, the Ag-Aper1 promoter (and its signal peptide) should be ideal for driving the expression of antimalaria proteins in transgenic Anopheles.