Identification of a Polymorphic Mucin-like Gene Expressed in the Midgut of the Mosquito, Aedes aegypti, Using an Integrated Bulked Segregant and Differential Display Analysis

Corresponding author: David W. Severson, Department of Biological Sciences, University of Notre Dame, Notre Dame, IN 46556., david.w.severson.1{at}nd.edu (E-mail)

Communicating editor: Z-B. ZENG

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The identification of putative differentially expressed genes within genome regions containing QTL determining susceptibility of the mosquito, Aedes aegypti, to the malarial parasite, Plasmodium gallinaceum, was investigated using an integrated, targeted approach based on bulked segregant and differential display analysis. A mosquito F₂ population was obtained from pairwise matings between the parasite-susceptible RED strain and the resistant MOYO-R substrain. DNA from female carcasses was used to genotype individuals at RFLP markers of known chromosomal position around the major QTL (pgs 1). Midguts, dissected 48 hr after an infected blood meal, were used to prepare two RNA bulks, each representing one of the parental genotypes at the QTL interval. The RNA bulks were compared by differential display PCR. A mucin-like protein gene (AeIMUC1) was isolated and characterized. The gene maps within the pgs 1 QTL interval and is expressed in the adult female midgut. AeIMUC1 RNA abundance decreased with time after blood meal ingestion. No differential expression was observed between the two mosquito strains but three different alleles with inter- and intrastrain allelic polymorphisms including indels and SNPs were characterized. The AeIMUC1 gene chromosome location and allelic polymorphisms raise the possibility that the protein might be involved in parasite-mosquito interactions.

MALARIA occurs in more than 90 countries and the World Health Organization estimates that the disease affects 300–500 million people and kills over 1 million young children each year, mostly in sub-Saharan Africa. Transmission of the Plasmodium parasite to a human host is totally dependent upon the availability of a competent mosquito vector. Strategies for malaria control are numerous, largely based on preventive and/or therapeutic drug administration and on vector control measures such as insecticide applications. However, the increase of both parasite drug resistance and mosquito insecticide resistance has reduced control efficiency. New malaria disease control strategies are needed. Vector-based control could be achieved by genetic approaches if, for example, transmission-blocking proteins were identified (CRAMPTON et al. 1994 ; COLLINS and PASKEWITZ 1995 ).

The mosquito vector acquires Plasmodium parasites while blood-feeding on an infected vertebrate host. In the mosquito midgut lumen, male and female gametes escape from the red blood cells, undergo fertilization and the resulting zygotes evolve into motile ookinetes. The ookinetes must then pass through the peritrophic matrix, penetrate the midgut epithelial cells, and lodge on the basal lamina where they transform into oocysts. Maturation of the oocysts leads to sporozoite release into the hemocoel. The sporozoites migrate to the salivary glands and are inoculated into a new vertebrate host when the mosquito feeds.

Plasmodium development in the mosquito midgut progresses in concert with blood meal digestion and therefore the parasite has to confront both physical and biochemical barriers. The peritrophic matrix (PM) is an extracellular layer that forms around the food bolus in the guts of most arthropods (PETERS 1992 ; TELLAM 1996 ). Although its function is not clearly defined, it likely protects the gut epithelium from damage by food particles, facilitates digestion, and also acts as a protective barrier against pathogens (PETERS 1992 ; TELLAM 1996 ). The PM is composed of chitins, proteins, and proteoglycans integrated in a sheet-like structure that Plasmodium ookinetes must cross to reach the midgut epithelium. The mosquito midgut is a monolayer of epithelial cells standing on a continuous basal lamina. The apical side of the epithelial cells is folded into microvilli and covered with glycocalyx (LEHANE 1996). Interactions between the parasite and the mosquito midgut determine the developmental fate of ookinetes and subsequently the successful transmission of malaria (SHAHABUDDIN 1998 ). Understanding the genetic mechanisms involved in susceptibility and refractoriness of mosquito species to Plasmodium will help to identify candidate genes suitable for genetic manipulation of mosquito vector competence.

Aedes aegypti and Plasmodium gallinaceum constitute a vector-parasite model of choice for studies investigating the genetic basis for vector competence for human malaria (KILAMA and CRAIG 1969 ; THATHY et al. 1994 ). A linkage map based on restriction fragment length polymorphisms (RFLP) has been developed for A. aegypti (SEVERSON et al. 1993 ) and these markers have been used to identify quantitative trait loci (QTL) that determine P. gallinaceum susceptibility (pgs). Two QTL affecting pgs have been identified: a major QTL, pgs 1, located on chromosome 2 and a minor QTL, pgs 2, on chromosome 3 (SEVERSON et al. 1995 ). Targeting the identification of differentially transcribed sequences within the genome regions carrying QTL could provide an effective approach for isolating genes associated with vector competence.

Bulked segregant analysis (BSA) provides a method to focus on regions containing QTL of interest. BSA involves screening for differences between pooled DNA samples (MICHELMORE et al. 1991 ). Individuals are selected from a population segregating for a trait of interest and bulked pools are prepared with DNA from individuals containing identical parental genotypes across the target region. The resultant bulks will reflect DNA polymorphisms linked to the target region. BSA has already been used successfully to identify specific random markers linked with traits of interest (GIOVANNONI et al. 1991 ; MICHELMORE et al. 1991 ; SEVERSON et al. 1999 ).

Differential display PCR (DD-PCR) of RNA has been described as a powerful technique to identify differentially expressed genes (LIANG and PARDEE 1992 ). This technique is based on the random amplification of fragments from first-strand cDNA using a combination of an anchored oligo(dT) primer and a random oligonucleotide primer. The method has been used successfully to identify differentially expressed genes (DIMOPOULOS et al. 1996 ; KANG et al. 1996 ; SCHROEDER et al. 1999 ). Nevertheless, drawbacks such as the large number of false positives generated are often reported and different methods have been proposed to facilitate the identification of putative genes (LIANG et al. 1993 ; LUCE and BURROWS 1998 ; FROST and GUGGENHEIM 1999 ; KOHROKI et al. 1999 ).

In this study, we report an integrated, targeted approach to identifying putative differentially expressed genes within genome regions containing QTL. We initially used the bulked segregant approach to genotype A. aegypti F₁ intercross progeny at RFLP marker loci across the pgs 1 QTL. These progeny had been dissected to remove their midguts 48 hr following a P. gallinaceum-infected blood meal, and the midguts were stored at -80°. To target differential RNA expression, we pooled total RNA extracted from dissected mosquito midguts, with individual midguts selected on the basis of RFLP marker genotypes as determined from the mosquito carcass following midgut removal (e.g., susceptible vs. refractory genotypes at the QTL). We then used RT-PCR and the differential display technique to identify several putative differentially expressed cDNA sequences. We have used these procedures to identify and characterize a mucin-like gene. The gene has been recently cloned independently by others and has been named AeIMUC1 for A. aegypti intestinal mucin (RAYMS-KELLER et al. 2000 ). We have adopted its prior name. Here we report that the gene maps within the pgs 1 QTL interval and is expressed only in the A. aegypti adult female midgut. Interestingly, this gene does not show differential expression between the susceptible and refractory mosquito strains, but does reflect several strain-specific single nucleotide substitutions and indels that result in putative polypeptide product polymorphisms.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mosquitoes and parasites:
Mosquito strains were reared in an environmental chamber at 26°, 84% relative humidity, under a 16-hr light/8-hr dark cycle with a 1-hr crepuscular period at the beginning and the end of each light cycle. Adults were maintained on a 5% sucrose solution. Female mosquitoes were blood-fed on anesthetized rats or on infective white Leghorn chicks.

The P. gallinaceum strain is routinely maintained in the laboratory by natural sporozoite transmission between the A. aegypti RED strain and chicks. One-week-old birds are infected by exposure to infected mosquitoes. The parasitemia is monitored daily on thin Giemsa-stained blood smears from 1 week after the infection until a gametocytemia range of 1–3% is reached.

Genetic crosses were performed to obtain F₁ intercross progeny between the P. gallinaceum-susceptible A. aegypti RED strain and the P. gallinaceum-resistant A. aegypti MOYO-R substrain. MOYO-R substrain derives from the Moyo-In-Dry strain. Selection of the substrain and relative susceptibilities of the MOYO-R substrain and the RED strain are described by THATHY et al. 1994 .

F₂ female mosquitoes, 5–10 days old, were infected by feeding on infective chicks for 15 min. Completely engorged females were then separated and maintained in groups of 15 individuals until dissection. For each dissection experiment, a group of engorged females was retained for midgut dissections 7 days after blood-feeding, to check both prevalence and intensities of infection as indicated by the number of developing oocysts. Some females were also routinely fed on uninfected chicks as parasite negative controls.

F₁ intercross females used for the DD-PCR analysis were dissected 48 hr after the infective blood meal. We chose 48 hr after feeding because by this time, in susceptible individuals, the ookinetes should have crossed the midgut epithelium and should actively be developing into oocysts. Midguts were carefully removed and individually homogenized in 20 µl of Tri Reagent (Molecular Research Center, Inc., Cincinnati, OH) and 0.2 µl of Polyacryl Carrier (Molecular Research Center, Inc.). Midgut homogenates were stored at -80° until RNA extractions. The remaining mosquito carcass was homogenized in 200 µl of DNA extraction buffer using a Tissue Tearor (Biospec, Bartlesville, OK) and DNA extractions were performed as previously described (SEVERSON 1997 ).

With mosquitoes used for semiquantitative RT-PCR (see below), midguts were dissected at several hours postinfection (hpi), i.e., 24, 48, and 72 hpi and 7 days postinfection, and RNA extractions were performed on bulked samples of 15 midguts.

BSA-DD assay by selective genotyping using RFLP markers:
For BSA, DNA extraction from individual F₁ intercross females (n = 155), digestion by EcoRI, Southern blotting, and hybridizations were performed as previously described (SEVERSON 1997 ). Southern blots were exposed to phosphorimaging screens (Molecular Dynamics, Sunnyvale, CA) and scanned using a Storm 840 phosphorimaging system (Molecular Dynamics).

All individuals were genotyped at the LF338, LF181, LF282, and LF272 loci on chromosome 2. The RFLP probes were chosen on the basis of their known chromosomal positions around the major QTL (pgs 1) determining P. gallinaceum susceptibility in A. aegypti (SEVERSON et al. 1995 ). The RFLP genotypes were then used for identifying individuals suitable for preparing two RNA bulks for differential display analysis (Fig 1). Procedures for the selective genotyping are described elsewhere (SEVERSON et al. 1999 ).

View larger version (31K): In this window In a new window Download PPT slide   Figure 1. Schematic description of the BSA-DD approach. Genetic crosses were performed between the P. gallinaceum-susceptible A. aegypti RED strain and the P. gallinaceum-resistant A. aegypti MOYO-R substrain and the F1 intercross progeny were fed on infective chicks. Females were dissected 48 hr after the blood meal. Midguts were removed and stored separately until the RNA extraction. Genotyping was performed at RFLP marker loci across the pgs 1 QTL on DNA extracted from the mosquito carcass. Two RNA pools were prepared with individual midguts from female mosquitoes that showed susceptible vs. refractory genotypes at the QTL interval. The two RNA bulks were compared by DD-PCR and the putative differentially expressed cDNAs were identified.

RNA extraction and differential display PCR:
Midgut homogenates preserved in Tri Reagent were bulked as described above and total RNA was isolated according to the manufacturer's instructions.

DD-PCR was performed using the Delta differential display kit (Clontech, Palo Alto, CA). Two to five hundred nanograms of total RNA were reverse transcribed using Superscript II RNaseH^- reverse transcriptase (GIBCO-BRL, Gaithersburg, MD) and an anchored oligo(dT) primer. The resulting single-stranded cDNA was amplified by PCR with 1x Advantage KlenTaq polymerase mix (Clontech) using a pairwise combination of arbitrary primers and oligo(dT) primers in the presence of [-³³P]dATP (3000 Ci/mmol; ICN, Costa Mesa, CA). Long-distance PCR was performed using a thermocycler (Hybaid) according to the DD-PCR kit recommendations except that the cycle number was increased: 1 cycle of 94° for 5 min, 40° for 5 min, and 68° for 5 min; 5 cycles of 94° for 2 min, 40° for 5 min, and 68° for 5 min; and 30 cycles of 94° for 1 min, 60° for 1 min, and 68° for 2 min followed by a final extension at 68° for 10 min. The labeled amplified fragments were separated on 5% denaturing polyacrylamide gels at 50 W for 4 hr. Gels were dried and exposed to X-ray film for 24 hr. The film and original gel were aligned using Identi-kit stickers (Sigma, St. Louis) and bands of interest, e.g., differentially displayed, were then excised using a razorblade. The DNA was eluted from the gel by boiling in pyrocarbonic acid diethyl ester-treated water for 5 min. An aliquot of 7 µl of the eluted cDNA was reamplified by PCR with the same primer set used to generate the original band. The reamplified PCR product was separated on a 1.2% LMT agarose gel, excised using a razorblade and the DNA was purified using Ultrafree-DA (Millipore, Bedford, MA).

cDNA cloning and sequencing:
The gel-purified cDNA was ligated to pCR-2.1 vector (TOPO TA cloning kit, Invitrogen, San Diego) and the recombinant plasmids were transformed into competent Escherichia coli TOP10 cells (Invitrogen). Recombinant plasmids were purified from overnight cultures using the alkaline lysis technique (SAMBROOK et al. 1989 ). Plasmid templates were subjected to cycle sequencing using the ABI Prism big dye terminator kit (PE Applied Biosystems, Foster City, CA) and the insert sequence was obtained on both strands using the ABI Prism 310 genetic analyzer. The DNA sequences were submitted to the BLAST program (ALTSCHUL et al. 1997 ) for similarity searches with the sequence databases. Sequence alignment and primer selection were performed using the GCG software (DEVEREUX et al. 1984 ).

Virtual Northern analysis:
Virtual Northern analysis was carried out to confirm the expression of the DD-PCR products. Full-length cDNAs were generated from 1 µg total RNA of each of the two midgut bulks used in the DD-PCR and of bulks from both infected and noninfected midguts for the two A. aegypti strains, RED and MOYO-R. The RNA was reverse-transcribed with the SMART PCR cDNA synthesis kit (Clontech) according to the manufacturer's instructions. The cDNAs were amplified from 1.5 µl of the first-strand cDNAs in a 100-µl reaction mixture with the PCR primer supplied with the kit. The PCR cycling profile was a first denaturation at 95° for 2 min followed by 18 cycles at 95° for 15 sec, 65° for 30 sec, and 68° for 6 min. Four hundred nanograms of the amplified cDNA products were resolved on a 1.2% TAE/agarose gel and subjected to Southern blotting and probe hybridizations as previously described (SEVERSON 1997 ). The blot was successively hybridized with probes prepared from (1) a cloned DD-PCR product and (2) a RFLP marker clone (LF272) that represents partial sequence for the ribosomal protein S17 gene (RpS17; D. W. SEVERSON, unpublished data).

Northern blotting:
Tissue and/or developmental stage-specific expression was determined by standard Northern blotting. Ten micrograms of total RNA extracted from female guts, female carcasses following midgut removal, larvae, pupae, and males were separated by electrophoresis in 1.2% formaldehyde/agarose gels and transferred to Nylon membranes (Hybond, Amersham, Arlington Heights, IL). Hybridizations were performed at 42° in 50% formamide, 5x Denhardt's solution, 5x SSPE, 0.1% SDS, and 15 µg/ml denatured salmon sperm DNA. Membranes were washed to high stringency in 0.1x SSPE/0.5% SDS at 65° and visualized using a phosphorimaging system.

Semiquantitative RT-PCR:
Because of the small amount of RNA recovered from midgut extractions, we did semiquantitative RT-PCR assays to assess differential expression. Total RNAs extracted from midguts dissected at different time points after an infective blood meal were used for expression analysis. One microgram of total RNA was reverse transcribed in a 20-µl reaction mixture containing 50 mM Tris-HCl pH 8.3, 3 mM MgCl₂, 75 mM KCl, 10 mM DTT, 500 µM of each dNTP, 20 pmol of oligo(dT)₁₈ (Clontech) and 200 units of Superscript II RNaseH^- (GIBCO-BRL).

Quantification assays were investigated by multiplex RT-PCR, using the RpS17 gene sequence as the internal control for normalization. For each primer set, preliminary PCR amplifications were carried out under our standard conditions to determine the number of cycles necessary to attain visualized PCR products while avoiding saturation. Optimal cycle number was set at 22. Multiplex RT-PCRs were then performed with 1 µl of the RT products in a total volume of 25 µl in the presence of 50 mM Tris-HCl pH 8.3, 3 mM MgCl₂, 50 mM KCl, 400 µM of each dNTP, 0.25 units of Taq polymerase, 0.5 µCi [-³²P]dCTP, 10 pmol of the DD-PCR product-specific primers and 20 pmol of the RpS17 primers. PCR products (3 µl) were separated by electrophoresis on 8% acrylamide gels. The gels were dried, exposed to phosphorimaging screens and visualized using a Storm 840 phosphorimaging system. The RT-PCR products were quantified using the ImageQuant 5.0 software (Molecular Dynamics).

Linkage analysis and genetic mapping:
The cloned DD-PCR product was radiolabeled and used as a probe to the Southern blots of the F₁ intercross progeny as described above. Multipoint linkage analysis was conducted using the MAPMAKER computer package (LANDER et al. 1987 ; LINCOLN and LANDER 1990 ). Recombination frequencies were converted into map distances (cM) using the Kosambi function (KOSAMBI 1944 ).

Competitive PCR:
Strain-specific substitutions were confirmed by a competitive PCR assay. This assay was performed by the amplification of 1 ng of individual genomic DNA in a final volume of 25 µl using 50 mM Tris-HCl pH 8.3, 3 mM MgCl₂, 50 mM KCl, 400 µM of each dNTP, 0.25 units of Taq polymerase, 10 pmol of a reverse primer, 3 pmol of a forward primer and 20 pmol of an internal forward primer. PCRs were carried out at 95° for 2 min in a first denaturation, at 95° for 15 min, 65° for 1 min, and 72° for 2 min for 30 cycles, and then at 72° for 6 min for a final extension. PCR products were resolved on a 6% acrylamide gel, stained with ethidium bromide and visualized under UV light.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BSA-DD and identification of differentially expressed genes:
Selective genotyping of F₁ intercross females was used to identify individuals homozygous at two RFLP loci closely flanking the major QTL pgs 1, LF181 and LF282, and heterozygous at two more distant RFLP loci, LF338 and LF272 (Fig 2). Two RNA bulked pools were prepared, each with midgut RNA from six individuals homozygous for the parental RED and MOYO-R genotypes at the loci flanking the QTL interval. The cDNAs from the RNA bulks were compared by DD-PCR to identify genes activated by the infected blood meal. Forty-six combinations of an arbitrary primer and an oligo(dT) primer were examined and 21 candidate differentially expressed cDNA products were isolated.

View larger version (6K): In this window In a new window Download PPT slide   Figure 2. Genetic mapping of AeIMUC1. Map distances of MUC1 from the RFLP markers used to genotype the F1 intercross progeny (n = 155 females) are indicated in centimorgans (cM). The pgs 1 QTL interval is shown by the shaded area.

Two cDNA clones 456 and 508 bp in length were obtained with the same primer combination (AP8 and DT5) from the RED type bulk and the MOYO-R type bulk, respectively, that showed high similarity (97 and 96% identities, respectively) with an A. aegypti mucin-like (AeIMUC1) gene sequence (accession no. AF125984). The RED type sequence contains a deletion of a 51-bp motif at the 3' end of the sequence. Virtual Northern blot analysis confirmed the AeIMUC1 length polymorphism between the two A. aegypti strains (Fig 3A).

View larger version (51K): In this window In a new window Download PPT slide   Figure 3. AeIMUC1 length polymorphism. (A) Virtual Northern blot of reverse-transcribed RNAs (400 ng of cDNAs per lane) isolated from female midguts dissected 48 hr after a blood meal. Midguts were from the following sources: bulk, females used to prepare the RNA bulks; inf, females fed on infected chicks; ni, females fed on noninfected chicks. (Top) The blot was probed with 32P-labeled AeIMUC1 DD-PCR insert. The autoradiogram shows the length polymorphism between the two strains, RED and MOYO-R. (Bottom) The blot was stripped and rehybridized with the RpS17 ribosomal probe used as loading control. (B) Ethidium bromide-stained agarose gel showing left, AeIMUC1 allelic polymorphism among several A. aegypti laboratory strains and right, AeIMUC1 species specificity. A. aegypti strains: Lvp, Liverpool; Bz, Bronze; For, Formosus; Sil, Silver; Ham, Hamburg; MD, Moyo-in-Dry; MR, MOYO-R; Red, RED-EYE; C-, PCR negative control; M, 174 Hae III marker.

AeIMUC1 polymorphism:
Primers (MLPF and MLPR, see Table 1) flanking the deletion polymorphism were designed and used to verify allele identities for the individuals bulked for the differential display analysis. All the mosquitoes pooled in the MOYO-R type bulk carried the MOYO-R allele (AeIMUC^M1, 242 bp), while those pooled in the RED type bulk carried the RED allele (AeIMUC^R1, 190 bp; data not shown).

PCR amplifications with the MLP primers were then performed on bulked DNA from several A. aegypti laboratory strains. Of interest, a third allele was identified in two strains, the RED strain and the Formosus strain (Fig 3B). The new allele was intermediate in size (212 bp). A screening of random individuals from the RED strain was undertaken and indeed two alleles were segregating within the strain population. Subsequently, the third allele (AeIMUC^R2) was cloned and sequenced. Fig 4C shows the alignment of the three alleles. The AeIMUC^R2 allele has, when compared to the AeIMUC^M1 allele, a 30-bp deletion and the gap occurs completely within the same deleted region found in the AeIMUC^R1 allele. Within this polymorphic region, in the AeIMUC^M1 allele, a 9-bp repeat motif (AGTGGTTTC) flanks a 42-bp sequence and in the two other alleles the 3' deletions involve this region. Whereas all the alleles have the 3' repeat, the 5' motif is absent in the AeIMUC^R1 allele. Thus the AeIMUC^R2 deletion occurs in the sequence flanked by the repeat motifs whereas the AeIMUC^R1 deletion consists of the full 42-bp sequence plus one repeat motif. Further, an adenosine residue is deleted in the AeIMUC^R1 allele 23 bases upstream from the deleted sequence. Also, as shown in Fig 4C, a 3-bp sequence reflects allele-specific substitutions. We also noted that the AeIMUC1 gene from A. aegypti Rexville strain (RAYMS-KELLER et al. 2000 ) has the AeIMUC^R1 type allele.

View larger version (49K): In this window In a new window Download PPT slide   Figure 4. AeIMUC1 allelic polymorphisms. (A) Amino acid sequence alignment of the three AeIMUC1 alleles. Asterisks mark allele-specific substitutions. The different protein domains are shaded and indicated. (B) Nucleotide sequence comparison at the deletion within the mucin domain. The repeat motif is underlined and the single nucleotide substitutions are marked with asterisks. Numbers indicate the position in the cDNA nucleotide sequence. (C) Nucleotide sequence comparison showing the 3' end deletion polymorphism. The adenosine deletion in the AeIMUCR1 allele is indicated by an arrow. The termination codon for each allele is in boldface type. The trinucleotide substitution is boxed. The repeat motif flanking the deleted area is underlined. (D) Nucleotide sequence comparison of the 65-bp intron. Single nucleotide substitutions are in boldface type. The splice junctions are italicized.

Allele-specific single base substitutions were confirmed using a competitive PCR assay with primers (MSXF, X = M1, R1, or R2, see Table 1) designed with 3' single base substitution anchors. PCR was performed on genomic DNA from individuals previously genotyped relative to the deletion polymorphism by adding one of the MSXF primers to the primers (MLPF, MLPR) flanking the length polymorphism (see MATERIALS AND METHODS). The MLPF primer was used at a limited amount (3 pmol) as a PCR control, to allow amplification even if the MSXF primer did not anneal. There was complete correlation between deletion polymorphisms and the expected amplicon with the corresponding MSXF primer (data not shown).

Cloning of AeIMUC1 gene:
The size of the AeIMUC1 gene mRNA as detected on virtual Northern blots was estimated to be 1200 bp. To facilitate cloning of the complete putative coding sequence of each allele, primers (MUCF, MUCR, see Table 1) were designed on the basis of the DD-PCR insert's similarity to the A. aegypti mucin-like protein published sequence. Full gene sequences for each allele were obtained by PCR amplifications using the full-length cDNAs generated in the virtual Northern procedure from both A. aegypti strains. The 3' and 5' ends were obtained by using a combination of an internal primer and one of the two primers used in the SMART PCR cDNA synthesis kit, respectively MUCIPF/CDS and MUCIPR/SMO (see Table 1).

Sequences of 1020, 1018, and 1072 bp in length were obtained for the AeIMUC^R1, AeIMUC^R2, and AeIMUC^M1 alleles, respectively. Each sequence includes a complete open reading frame encoding a putative protein, of 275 amino acids (aa) for the AeIMUC^R1 allele, 273 aa for the AeIMUC^R2 allele, and 271 aa for the AeIMUC^M1 allele (Fig 4A). The AeIMUC^M1 sequence terminates earlier when compared to the two other alleles because of the trinucleotide substitution which results in a TAA stop codon at peptide position 272 (Fig 4C). The adenosine residue deletion observed in the AeIMUC^R1 allele results in a frameshift that produces a stop codon (TAG) just downstream from this single base deletion. With the AeIMUC^R2 allele, a TAG stop codon occurs upstream of the 30-bp deletion. A putative transcription initiator sequence (TCAAT) and a putative polyadenylation signal (AATAAA) appear in each of the three alleles, located respectively at 22 bp of the 5' end and at 19 bp upstream from the poly(A) tail (Fig 5).

View larger version (51K): In this window In a new window Download PPT slide   Figure 5. Nucleotide and deduced amino acid sequences of AeIMUCM1 cDNA. The 5' and 3' noncoding regions are represented by lowercase type. A putative transcription initiator sequence (tcaat) and a potential polyadenylation signal (aataaa) are in boldface and italic. The translation initiation codon and the termination codon are in boldface type and underlined. The predicted peptide signal sequence is underlined with a broken line. The regions showing deletion polymorphisms in the two other alleles are boxed and the repeat motifs are italicized. Primer set (MLPF, MLPR) used to determine allele identity and primer designed for the competitive assay (MSM1F) are indicated above the sequence. The SNPs specific to the AeIMUCM1 allele are in boldface type. The location of the intron in the corresponding genomic sequence is indicated by an arrow.

The complete nucleotide sequences of AeIMUC1 for the three alleles revealed that the AeIMUC^R2 allele has two deletions: the 30-bp deletion at the 3' end of the sequence (see above) and a 24-bp deletion in the mucin domain. As shown in Fig 4B, if compared to the AeIMUC^M1 allele, the 24-bp deletion involves a 13-bp repeat motif and the 11-bp sequence located between the two repeats. The corresponding region in the AeIMUC^R1 allele has four single nucleotide substitutions that change the first repeat motif.

Comparison of the nucleotide sequences for the three alleles revealed multiple point mutations. A total of 46 single nucleotide polymorphisms (SNPs) were found. The SNP distribution among the AeIMUC1 alleles is as follows: 10 for the MUC^M1 allele, 21 for the AeIMUC^R1 allele, and 15 for AeIMUC^R2 allele (Fig 5; GenBank accession nos. AF308863 and AF308864). Some, but not all of them, are nonsynonymous cSNPs and thus, as shown in Fig 4A, 22 amino acid substitutions are found.

Domain organization of AeIMUC1:
BLAST search with the full-length AeIMUC1 amino acid sequence revealed high similarities to peritrophic matrix proteins and insect chitinases. The deduced amino acid sequence of AeIMUC1 contains several conserved domains. The N-terminal amino acid sequence contains a signal peptide of 18 amino acids, typical of a secretory protein. A central core presents characteristic features of a mucin-like domain (GENDLER and SPICER 1995 ). The amino acid sequence is rich in threonine, proline, alanine, valine, and serine (32.5, 20.9, 16.3, 14.0, 14.0%, respectively). Three imperfect tandem repeats (TTTASPAV) are present in the AeIMUC^M1 and AeIMUC^R1 alleles while only two repeats are present in the AeIMUC^R2 allele (Fig 4A). For the AeIMUC^R2 allele, the 24-bp deleted sequence represents the first amino acid tandem repeat. As predicted by a search on the NetOglyc WWW server (HANSEN et al. 1998 ), all threonine and serine residues of this mucin domain are potential O-glycosylation sites. The predicted molecular mass of the peptide is 29.2 kD, but the likely presence of numerous O-linked carbohydrates suggests that the glycoprotein mass might be higher.

AeIMUC1 sequence also contains three putative chitin-binding domains (CBDs). Chitin-binding domains are characterized by positional conservation of six cysteines and aromatic residues (SHEN and JACOBS-LORENA 1999 ). The six cysteine residues form disulfide bonds and the conserved aromatic amino acids could be involved in binding specificity. The first putative CBD of AeIMUC1 peptide is located at the N terminus, upstream to the mucin domain and the two others are at the C terminus, downstream to the mucin domain (Fig 4A). The three CBDs share significant similarity with Anopheles gambiae peritrophin 1 (GenBank accession no. AAC39127), with A. aegypti chitinase (GenBank accession no. T14075), and with AgICHIT protein, which is an A. gambiae intestinal protein composed of two CBDs separated by a mucin domain (GenBank accession no. CAA09389). Fig 6 represents the sequence alignment of CBDs found in A. gambiae and A. aegypti proteins. AeIMUC1 CBD motifs share highest homology with Ag-Aper1, a major protein of the A. gambiae peritrophic matrix, and it is likely that the AeIMUC1 protein is also a constituent of the peritrophic matrix in A. aegypti.

View larger version (51K): In this window In a new window Download PPT slide   Figure 6. Alignment of the AeIMUCM1 putative chitin-binding domains with homologous sequence motifs from A. gambiae ICHIT protein (AgICHIT-1 and AgICHIT-2), A. gambiae peritrophin 1 (AgAper1-1 and AgAper1-2), A. aegypti chitinase 1 (AeChit1), A. aegypti chitinase 2 (AeChit2-1, AeChit2-2, and AeChit2-3), A. gambiae chitinase (AgChit), and A. gambiae serine protease Sp22D (AgSP22D-1 and AgSP22D-2). Conserved cysteines are in boldface type and the disulfide bonds are illustrated at the bottom, according to SHEN and JACOBS-LORENA 1999 . Conserved aromatic amino acids are marked with asterisks.

Genomic sequence:
Genomic DNA fragments specific for MUC1 were produced by amplification of gDNA from individuals previously genotyped relative to the deletion polymorphism with primers (MUCF, MUCR) flanking the open reading frame (ORF). As compared to cDNA amplification products, a larger amplicon was obtained for each of the three alleles, suggesting the presence of introns in the genomic DNA. Genomic fragments when cloned and sequenced revealed a 65-bp intron with the GU-AG feature characteristic of eukaryotic nuclear pre-mRNA. The 5' splice site is located four amino acids downstream to the start codon (Fig 5). Sequences for the three alleles are similar; Fig 4D shows the substitutions between the three alleles.

Expression of AeIMUC1:
Developmental stage and tissue-specific expression of AeIMUC1 were investigated by Northern blot analysis using total RNA from larvae, pupae, female carcasses (whole body minus midgut), female midguts, and males. AeIMUC1 was detected only in RNA from female midguts (Fig 7), suggesting that AeIMUC1 gene expression is female midgut specific.

View larger version (67K): In this window In a new window Download PPT slide   Figure 7. Developmental and tissue-specific expression of MUC1. (Top) Northern analysis of 32P-labeled AeIMUC1 DD-PCR insert on total RNAs (10 µg per lane) isolated from different developmental stages and tissues (L, larvae; P, pupae; C, female carcasses; G, female midguts; M, males). (Bottom) The blot was stripped and rehybridized with the RpS17 ribosomal probe used as loading control.

AeIMUC1 gene expression following P. gallinaceum infection was examined by semiquantitative RT-PCR analyses on total RNAs from midguts dissected 24, 48, and 72 hr and 7 days after an infective blood meal. RNAs extracted from midguts 48 hr after a noninfective feeding and from midguts of nonfed females were also screened, as AeIMUC1 expression could be induced by parasite ingestion or by the blood meal. The ratio of MLP PCR products/RpS17 PCR products was plotted against time after the blood meal for three independent experiments. Results were analyzed initially by a two-way ANOVA. AeIMUC1 expression levels were similar for both A. aegypti strains (P > 0.05) while significant differences in expression were observed for time after an infected blood meal (P < 0.05). No significant interaction effects were observed. We therefore pooled data across strains and the results were analyzed by a one-way ANOVA. Significant differences in time after blood meal ingestion were evident (P < 0.05); gene expression decreased with time after the blood meal ingestion (Fig 8A).

View larger version (31K): In this window In a new window Download PPT slide   Figure 8. Expression analysis of AeIMUC1 following exposure to a parasite-infected blood meal. Midgut RNAs from the susceptible RED and the resistant MOYO-R A. aegypti strains at different times following blood-feeding were used for semiquantitative RT-PCR standardized against RpS17. (A) Female midguts at 24, 48, and 72 hr and 7 days postfeeding on an infected blood meal (pi). (B) Female midguts at 48 hr following a noninfected (pf) and an infected (pi) blood meal. Values are means and standard deviations for three independent experiments (n = 15 pooled midguts/experiment).

We also compared AeIMUC1 expression levels at 48 hr following an uninfective vs. infective blood meal (Fig 8B). Because we observed no significant differences in expression between strains, data for the two strains were pooled and subjected to a t-test. We observed a significant difference (P = 0.008) in expression relative to blood meal status; AeIMUC1 expression levels remained higher at 48 hr postfeeding among individuals that received an infective blood meal.

Genetic mapping:
Southern blots of the F₁ intercross progeny used for the BSA-DD procedure were probed with the AeIMUC1 fragment obtained by DD-PCR to genetically determine its map position. Recombination frequencies of AeIMUC1 and the RFLP markers previously mapped in the F₁ intercross population localized the gene between two markers (LF181 and LF282) flanking the major QTL (pgs 1) determining P. gallinaceum susceptibility in A. aegypti (Fig 2).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The development of detailed genetic linkage maps using DNA-based markers has facilitated the dissection of complex traits into discrete components referred to as QTL, but the actual isolation of the associated genes remains a significant challenge. The integrated BSA-DD approach described in this article provides a rapid and accurate method to directly isolate candidate genes associated with QTL influencing a phenotype of interest such as pathogen/vector interactions. The procedure requires a population segregating for a genome region of interest and the ability to isolate both DNA and RNA from individuals within this population. As in classical BSA analysis, molecular markers defining a target region are used to identify individuals homozygous at loci flanking the target interval and heterozygous at more distant loci. However, with the BSA-DD approach, RNA from the selected individuals is then used to prepare two pools, each representing one of the parental genotypes across the target interval. In theory the two bulks should contain identical RNA species except for those RNAs differentially expressed by genes within the genome region of interest. Comparisons of the RNA bulks should identify polymorphisms associated with genes located in the targeted region. Additionally, RNA can be extracted from specific tissues considered to be likely sites for differential expression associated with QTL.

We demonstrated the utility of the BSA-DD approach for the isolation of differentially expressed RNAs within the genome region containing the major QTL, pgs 1, determining P. gallinaceum susceptibility in the mosquito A. aegypti. We characterized a mucin-like protein (AeIMUC1) that is adult female midgut specific and shares similarities with PM proteins. The structure of AeIMUC1, with three cysteine-rich domains and a mucin domain, strongly suggests that this protein is a mucin-like peritrophin of the adult female A. aegypti mosquito. The gene maps within the pgs 1 QTL interval, shows significant differences in expression following uninfected vs. parasite infected blood meals, and reflects allelic length polymorphism between our susceptible and resistant A. aegypti mosquito strains, suggesting that it could be directly involved in P. gallinaceum susceptibility. The mucin-like protein, AeIMUC1, recently identified by others in another A. aegypti strain was shown to be induced in the larval midgut after heavy metal exposure as well as in the adult female midgut within the first 6 hr after blood meal ingestion (RAYMS-KELLER et al. 2000 ). These data are in accordance with our results. The fact that in our experiments no overexpression of AeIMUC1 was detected at 24 hr after an infected blood meal suggests that the mRNA is rapidly degraded.

Two types of PMs are defined in insects, relating to their site of synthesis. Type 1 PM is secreted by midgut epithelial cells in many adult blood-feeding insects, whereas type 2 is synthesized by specialized cells within the cardia, an organ located in the anterior part of the midgut. The distinction between the two PM types is not clear at the molecular level. Typically, mosquitoes produce a different PM type at different life stages, e.g., type 2 PM during larval life and type 1 PM during adult life. Most insect PM proteins characterized to date have been isolated from larval stages (ELVIN et al. 1996 ; CASU et al. 1997 ; WANG and GRANADOS 1997 ). Most of these PM proteins (called peritrophins) also share common properties with mammalian mucins (TELLAM et al. 1999 ) and they are characterized by the presence of both cysteine-rich or chitin-binding domains and proline/threonine stretches. Similarities in their structural organization and location (covering epithelial cells) suggest that both proteins have analogous functions, particularly to protect the midgut epithelium. None of the PM proteins from adult-stage insects described so far show all characteristic features of mucin-like peritrophins. The only peritrophin characterized previously from blood-fed adult mosquitoes is Ag-Aper1 from A. gambiae, which has two CBDs but no mucin domain (SHEN and JACOBS-LORENA 1998 ). Other adult midgut proteins have been described but their association with the PM has not been demonstrated.

The AeIMUC1 protein contains three CBDs that show high sequence similarity to each other. Presence of CBDs suggests that the protein has the capacity to bind chitin within the PM. CBDs are commonly found in both insect PM proteins and invertebrate chitinases and share amino acid sequence similarities (SHEN and JACOBS-LORENA 1999 ). They are cysteine rich and have highly conserved aromatic residues. In CBDs of all related genes six cysteine residues can be aligned and likely allow the formation of disulfide bonds. Disulfide bonds help to maintain a digestive protease-resistant structure while the aromatic residues are presumed to determine binding specificity (ELVIN et al. 1996 ; WANG and GRANADOS 1997 ; SHEN and JACOBS-LORENA 1999 ). AeIMUC1 CBDs show high similarity with Ag-Aper1 CBDs, although MUC1 contains three CBDs while Ag-Aper1 contains only two CBDs (SHEN and JACOBS-LORENA 1998 ). A. aegypti chitinase also has three CBDs with significant amino acid sequence similarity to AeIMUC1 CBDs. AeIMUC1 CBDs show greater homology to Ag-Aper1 domains than to A. aegypti chitinase domains, and it is unlikely that AeIMUC1 protein and A. aegypti chitinase interact. They probably have synergistic functions: AeIMUC1 could promote the formation of PM structure by connecting chitin fibers, whereas A. aegypti chitinase is likely involved in PM degradation. Multiple CBDs are common in insect PM proteins and invertebrate chitinases and it has been argued that they result from gene duplication (SHEN and JACOBS-LORENA 1999 ).

It has been suggested previously that a carbohydrate might be involved in vector-parasite specificity by enabling the ookinete to recognize the peritrophic matrix and/or the midgut epithelium (BILLINGSLEY and RUDIN 1992 ; RAMASAMY et al. 1997 ; ZIELER et al. 1999 ). In adult female A. aegypti, the PM forms after the blood meal and is complete after 12 hr (RICHARDS and RICHARDS 1977 ). The ookinetes begin to penetrate the PM 24 hr after feeding, after the membrane is completely formed (SIEBER et al. 1991 ). Several studies suggest that while the PM is a physical barrier for parasite transmission (see review in SHAHABUDDIN 1998 ) it is not an absolute barrier and instead it reduces the intensity of infection (BILLINGSLEY and RUDIN 1992 ). The PM may act as a recognition site for the penetrating ookinetes, by exposing specific carbohydrate ligands to the parasites (BILLINGSLEY 1994 ). It is known that P. gallinaceum ookinetes produce and secrete a chitinase (HUBER et al. 1991 ) that is secreted as a proenzyme and is activated by a mosquito midgut trypsin (SHAHABUDDIN et al. 1993 , SHAHABUDDIN et al. 1995 ). During ookinete penetration through the PM the membrane is disrupted only around the invading point (SIEBER et al. 1991 ) and inhibitors of chitinase block parasite transmission (SHAHABUDDIN et al. 1993 , SHAHABUDDIN et al. 1996 ). Therefore once the PM is encountered and recognized, the parasites use their own chitinase to cross it. The MUC1 glycoprotein with high O-glycosylated sites may be a likely candidate for a PM recognition molecule; the CBDs allow the protein to crosslink with chitin fibrils and the mucin domain exposes carbohydrates to the endoperitrophic space.

The inter- and intrastrain allelic polymorphism observed with the AeIMUC1 gene and its genome location within the pgs 1 QTL interval are suggestive that it may play a significant role in A. aegypti infection by P. gallinaceum and, ultimately, vector competence to transmit the parasite to vertebrates. However, these results are simply correlative and do not in themselves confirm this hypothesis. Obviously, this needs to be tested directly using a gene expression system such as the Sindbis virus-transducing vector (OLSON et al. 2000 ). SNP-based association studies should also be done to determine the potential influence of individual sequence variants on the susceptible/resistant phenotype.

Identification of differentially expressed genes is of great importance, as they constitute candidate genes for use in developing effective control strategies for vector-borne diseases. Although several techniques that produce differential transcripts between two populations of RNAs exist (LEE et al. 1991 ; HUBANK and SCHATZ 1994 ; PRASHAR and WEISSMAN 1996 ; SUZUKI et al. 1996 ), RNA-based methods are often limited for mosquito studies by the availability of sufficient material. The RNA differential display method overcomes this limitation, as only 1 µg of total RNA is sufficient for use as starting material. The main drawbacks reported with this technique are the high levels of false positives and biases for high copy number genes (LIANG et al. 1993 ; BERTIOLI et al. 1995 ; LUCE and BURROWS 1998 ). In addition, this method is likely to miss mutations that affect phenotype expression but not transcription levels. The more recent method of suppression subtractive hybridization (SSH; DIATCHENKO et al. 1996 ) offers an alternative for identification of differentially expressed genes. The SSH method incorporates normalization and subtraction in a single procedure that facilitates isolation of cDNAs representing rare transcripts and is also appropriate for an integrated BSA-SSH approach.

Current proposed genetic-control strategies for mosquitoes are based on modification of the capacity of the natural vector to support parasite development (COLLINS and PASKEWITZ 1995 ). Germline transformation of insect vectors has already been achieved (JASINSKIENE et al. 1998 ; CATTERUCCIA et al. 2000 ), so the identification of transmission-blocking genes remains a primary goal in efforts to engineer malaria-refractory mosquitoes. At present, no gene has been shown to be directly correlated with genetic refractoriness. Targeting differential gene expression in genome regions associated with parasite susceptibility could allow for the rapid and efficient identification of the associated genes.

We have shown that the BSA-DD approach provides an efficient method for the identification of putative differentially expressed genes within a diploid organism for which molecular markers delimiting QTL exist. The technique should be useful for the identification of genes that play key roles in a wide range of systems, including plants and animals.

	ACKNOWLEDGMENTS

We are grateful to Jeanne Romero-Severson for helpful discussions. The sequences reported in this article have been deposited in the GenBank database and assigned the following accession numbers: AeIMUC^M1, AF308862; AeIMUC^R1, AF308863; AeIMUC^R2, AF308864. This work was supported by a grant (AI 33127) from the National Institutes of Health.

Manuscript received October 2, 2000; Accepted for publication April 11, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALTSCHUL, S. F., T. L. MADDEN, A. A. SCHAFFER, J. ZHANG, and Z. ZHANG et al., 1997  Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402[Abstract/Free Full Text].

BERTIOLI, D. J., U. H. SCHLICHTER, M. J. ADAMS, P. R. BURROWS, and H. H. STEINBISS et al., 1995  An analysis of differential display shows a strong bias towards high copy number mRNAs. Nucleic Acids Res. 23:4520-4523[Free Full Text].

BILLINGSLEY, P. F., 1994  Vector-parasite interactions for vaccine development. Int. J. Parasitol. 24:53-58[Medline].

BILLINGSLEY, P. F. and W. RUDIN, 1992  The role of the mosquito peritrophic membrane in bloodmeal digestion and infectivity of Plasmodium species. J. Parasitol. 78:430-440[Medline].

CASU, R., C. EISEMANN, R. PEARSON, G. RIDING, and I. EAST et al., 1997  Antibody-mediated inhibition of the growth of larvae from an insect causing cutaneous myiasis in a mammalian host. Proc. Natl. Acad. Sci. USA 94:8939-8944[Abstract/Free Full Text].

CATTERUCCIA, F., T. NOLAN, C. BLASS, H. M. MULLER, and A. CRISANTI et al., 2000  Toward Anopheles transformation: Minos element activity in anopheline cells and embryos. Proc. Natl. Acad. Sci. USA 97:2157-2162[Abstract/Free Full Text].

COLLINS, F. H. and S. M. PASKEWITZ, 1995  Malaria: current and future prospects for control. Annu. Rev. Entomol. 40:195-219[Medline].

CRAMPTON, J. M., A. WARREN, G. J. LYCETT, M. A. HUGHES, and I. P. COMLEY et al., 1994  Genetic manipulation of insect vectors as a strategy for the control of vector-borne disease. Ann. Trop. Med. Parasitol. 88:3-12[Medline].

DEVEREUX, J., P. HAEBERLI, and O. SMITHIES, 1984  A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395.

DIATCHENKO, L., Y. F. LAU, A. P. CAMPBELL, A. CHENCHIK, and F. MOQADAM et al., 1996  Suppression subtractive hybridization: a method for generating differentially regulated or tissue-specific cDNA probes and libraries. Proc. Natl. Acad. Sci. USA 93:6025-6030[Abstract/Free Full Text].

DIMOPOULOS, G., A. RICHMAN, A. DELLA TORRE, F. C. KAFATOS, and C. LOUIS, 1996  Identification and characterization of differentially expressed cDNAs of the vector mosquito, Anopheles gambiae.. Proc. Natl. Acad. Sci. USA 93:13066-13071[Abstract/Free Full Text].

ELVIN, C. M., T. VUOCOLO, R. D. PEARSON, I. J. EAST, and G. A. RIDING et al., 1996  Characterization of a major peritrophic membrane protein, peritrophin-44, from the larvae of Lucilia cuprina. cDNA and deduced amino acid sequences. J. Biol. Chem. 271:8925-8935[Abstract/Free Full Text].

FROST, M. R. and J. A. GUGGENHEIM, 1999  Prevention of depurination during elution facilitates the reamplification of DNA from differential display gels. Nucleic Acids Res. 27:e6[Abstract/Free Full Text].

GENDLER, S. J. and A. P. SPICER, 1995  Epithelial mucin genes. Annu. Rev. Physiol. 57:607-634[Medline].

GIOVANNONI, J. J., R. A. WING, M. W. GANAL, and S. D. TANKSLEY, 1991  Isolation of molecular markers from specific chromosomal intervals using DNA pools from existing mapping populations. Nucleic Acids Res. 19:6553-6558[Abstract/Free Full Text].

HANSEN, J. E., O. LUND, N. TOLSTRUP, A. A. GOOLEY, and K. L. WILLIAMS et al., 1998  NetOglyc: prediction of mucin type O-glycosylation sites based on sequence context and surface accessibility. Glycoconj. J. 15:115-130[Medline].

HUBANK, M. and D. G. SCHATZ, 1994  Identifying differences in mRNA expression by representational difference analysis of cDNA. Nucleic Acids Res. 22:5640-5648[Abstract/Free Full Text].

HUBER, M., E. CABIB, and L. H. MILLER, 1991  Malaria parasite chitinase and penetration of the mosquito peritrophic membrane. Proc. Natl. Acad. Sci. USA 88:2807-2810[Abstract/Free Full Text].

JASINSKIENE, N., C. J. COATES, M. Q. BENEDICT, A. J. CORNEL, and C. S. RAFFERTY et al., 1998  Stable transformation of the yellow fever mosquito, Aedes aegypti, with the Hermes element from the housefly. Proc. Natl. Acad. Sci. USA 95:3743-3747[Abstract/Free Full Text].

KANG, D., G. LIU, H. GUNNE, and H. STEINER, 1996  PCR differential display of immune gene expression in Trichoplusia ni.. Insect Biochem. Mol. Biol. 26:177-184[Medline].

KILAMA, W. L. and G. B. CRAIG, JR., 1969  Monofactorial inheritance of susceptibility to Plasmodium gallinaceum in Aedes aegypti.. Ann. Trop. Med. Parasitol. 63:419-432[Medline].

KOHROKI, J., M. TSUCHIYA, S. FUJITA, T. NAKANISHI, and N. ITOH et al., 1999  A novel strategy for identifying differential gene expression: an improved method of differential display analysis. Biochem. Biophys. Res. Commun. 262:365-367[Medline].

KOSAMBI, D. D., 1944  The estimation of map distance from recombination values. Ann. Eugen. 12:172-175.

LANDER, E. S., P. GREEN, J. ABRAHAMSON, A. BARLOW, and M. J. DALY et al., 1987  MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181[Medline].

LEE, S. W., C. TOMASETTO, and R. SAGER, 1991  Positive selection of candidate tumor-suppressor genes by subtractive hybridization. Proc. Natl. Acad. Sci. USA 88:2825-2829[Abstract/Free Full Text].

LEHANE, M. J., A. B., and P. F. BILLINGSLEY, (Editors), 1996 Biology of the Insect Midgut. Chapman & Hall, New York.

LIANG, P. and A. B. PARDEE, 1992  Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science 257:967-971[Abstract/Free Full Text].

LIANG, P., L. AVERBOUKH, and A. B. PARDEE, 1993  Distribution and cloning of eukaryotic mRNAs by means of differential display: refinements and optimization. Nucleic Acids Res. 21:3269-3275[Abstract/Free Full Text].

LINCOLN, S. E., and E. S. LANDER, 1990 Mapping genes controlling quantitative traits using MAPMAKER/QTL. Whitehead Institute for Biomedical Research, Cambridge, MA.

LUCE, M. J. and P. D. BURROWS, 1998  Minimizing false positives in differential display. Biotechniques 24:766-768. 770.[Medline].

MICHELMORE, R. W., I. PARAN, and R. V. KESSELI, 1991  Identification of markers linked to disease-resistance genes by bulked segregant analysis: a rapid method to detect markers in specific genomic regions by using segregating populations. Proc. Natl. Acad. Sci. USA 88:9828-9832[Abstract/Free Full Text].

OLSON, K. E., K. M. MYLES, R. C. SEABAUGH, S. HIGGS, and J. O. CARLSON et al., 2000  Development of a Sindbis virus expression system that efficiently expresses green fluorescent protein in midguts of Aedes aegypti following per os infection. Insect Mol. Biol. 9:57-65[Medline].

PETERS, W., (Editor), 1992 Peritrophic membranes. Springer-Verlag, Berlin.

PRASHAR, Y. and S. M. WEISSMAN, 1996  Analysis of differential gene expression by display of 3' end restriction fragments of cDNAs. Proc. Natl. Acad. Sci. USA 93:659-663[Abstract/Free Full Text].

RAMASAMY, R., I. C. WANNIARACHCHI, K. A. SRIKRISHNARAJ, and M. S. RAMASAMY, 1997  Mosquito midgut glycoproteins and recognition sites for malaria parasites. Biochim. Biophys. Acta 1361:114-122[Medline].

RAYMS-KELLER, A., M. MCGAW, C. ORAY, J. O. CARLSON, and B. J. BEATY, 2000  Molecular cloning and characterization of a metal responsive Aedes aegypti intestinal mucin cDNA. Insect Mol. Biol. 9:419-426[Medline].

RICHARDS, A. G. and P. A. RICHARDS, 1977  The peritrophic membranes of insects. Annu. Rev. Entomol. 22:219-240[Medline].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SCHROEDER, A. A., C. E. LAWRENCE, and M. S. ABRAHAMSEN, 1999  Differential mRNA display cloning and characterization of a Cryptosporidium parvum gene expressed during intracellular development. J. Parasitol. 85:213-220[Medline].

SEVERSON, D. W., 1997 RFLP analysis of insect genomes, pp. 309–320 in The Molecular Biology of Insect Disease Vectors, edited by J. M. CRAMPTON, C. B. BEARD and C. LOUIS. Chapman & Hall, London.

SEVERSON, D. W., A. MORI, Y. ZHANG, and B. M. CHRISTENSEN, 1993  Linkage map for Aedes aegypti using restriction fragment length polymorphisms. J. Hered. 84:241-247[Abstract/Free Full Text].

SEVERSON, D. W., V. THATHY, A. MORI, Y. ZHANG, and B. M. CHRISTENSEN, 1995  Restriction fragment length polymorphism mapping of quantitative trait loci for malaria parasite susceptibility in the mosquito Aedes aegypti.. Genetics 139:1711-1717[Abstract].

SEVERSON, D. W., D. ZAITLIN, and V. A. KASSNER, 1999  Targeted identification of markers linked to malaria and filarioid nematode parasite resistance genes in the mosquito Aedes aegypti.. Genet. Res. 73:217-224[Medline].

SHAHABUDDIN, M., 1998  Plasmodium ookinete development in the mosquito midgut: a case of reciprocal manipulation. Parasitology 116:S83-S93.

SHAHABUDDIN, M., T. TOYOSHIMA, M. AIKAWA, and D. C. KASLOW, 1993  Transmission-blocking activity of a chitinase inhibitor and activation of malarial parasite chitinase by mosquito protease. Proc. Natl. Acad. Sci. USA 90:4266-4270[Abstract/Free Full Text].

SHAHABUDDIN, M., M. CRISCIO, and D. C. KASLOW, 1995  Unique specificity of in vitro inhibition of mosquito midgut trypsin-like activity correlates with in vivo inhibition of malaria parasite infectivity. Exp. Parasitol. 80:212-219[Medline].

SHAHABUDDIN, M., F. J. LEMOS, D. C. KASLOW, and M. JACOBS-LORENA, 1996  Antibody-mediated inhibition of Aedes aegypti midgut trypsins blocks sporogonic development of Plasmodium gallinaceum.. Infect. Immun. 64:739-743[Abstract].

SHEN, Z. and M. JACOBS-LORENA, 1998  A type I peritrophic matrix protein from the malaria vector Anopheles gambiae binds to chitin. Cloning, expression, and characterization. J. Biol. Chem. 273:17665-17670[Abstract/Free Full Text].

SHEN, Z. and M. JACOBS-LORENA, 1999  Evolution of chitin-binding proteins in invertebrates. J. Mol. Evol. 48:341-347[Medline].

SIEBER, K. P., M. HUBER, D. KASLOW, S. M. BANKS, and M. TORII et al., 1991  The peritrophic membrane as a barrier: its penetration by Plasmodium gallinaceum and the effect of a monoclonal antibody to ookinetes. Exp. Parasitol. 72:145-156[Medline].

SUZUKI, H., T. YAOI, J. KAWAI, A. HARA, and G. KUWAJIMA et al., 1996  Restriction landmark cDNA scanning (RLCS): a novel cDNA display system using two-dimensional gel electrophoresis. Nucleic Acids Res. 24:289-294[Abstract/Free Full Text].

TELLAM, R. L., (Editor), 1996 The peritrophic matrix, in Biology of the Insect Midgut. Chapman & Hall, New York.

TELLAM, R. L., G. WIJFFELS, and P. WILLADSEN, 1999  Peritrophic matrix proteins. Insect Biochem. Mol. Biol. 29:87-101[Medline].

THATHY, V., D. W. SEVERSON, and B. M. CHRISTENSEN, 1994  Reinterpretation of the genetics of susceptibility of Aedes aegypti to Plasmodium gallinaceum.. J. Parasitol. 80:705-712[Medline].

WANG, P. and R. R. GRANADOS, 1997  Molecular cloning and sequencing of a novel invertebrate intestinal mucin cDNA. J. Biol. Chem. 272:16663-16669[Abstract/Free Full Text].