Using massive cDNA sequencing, proteomics and customized computational biology approaches, we have isolated and identified the most abundant secreted proteins from the salivary glands of the sand fly Lutzomyia longipalpis. Out of 550 randomly isolated clones from a full-length salivary gland cDNA library, we found 143 clusters or families of related proteins. Out of these 143 families, 35 were predicted to be secreted proteins. We confirmed, by Edman degradation of Lu. longipalpis salivary proteins, the presence of 17 proteins from this group. Full-length sequence for 35 cDNA messages for secretory proteins is reported, including an RGD-containing peptide, three members of the yellow-related family of proteins, maxadilan, a PpSP15-related protein, six members of a family of putative anticoagulants, an antigen 5-related protein, a D7-related protein, a cDNA belonging to the Cimex apyrase family of proteins, a protein homologous to a silk protein with amino acid repeats resembling extracellular matrix proteins, a 5′-nucleotidase, a peptidase, a palmitoyl-hydrolase, an endonuclease, nine novel peptides and four different groups of proteins with no homologies to any protein deposited in accessible databases. Sixteen of these proteins appear to be unique to sand flies. With this approach, we have tripled the number of isolated secretory proteins from this sand fly. Because of the relationship between the vertebrate host immune response to salivary proteins and protection to parasite infection, these proteins are promising markers for vector exposure and attractive targets for vaccine development to control Leishmania chagasi infection.
The New World sand fly, Lutzomyia longipalpis, is an important vector of Leishmania chagasi, causal agent of the visceral form of leishmaniasis in humans. These sand flies, as well as many other blood feeders, contain potent pharmacological components in their saliva that assist them in obtaining their blood meals and evading host inflammatory and immune responses. Components such as anticoagulants, inhibitors of platelet aggregation and vasodilators have been described in blood feeders (Ribeiro, 1987, 1995). In the saliva of Lu. longipalpis, apyrase, anticoagulant, vasodilatory and immunomodulatory activities have been reported; some of these molecules were characterized at the molecular level (Lerner et al., 1991; Charlab et al., 1999; Gillespie et al., 2000).
Arthropod saliva modifies the physiology of the host at the site of the bite, making it more permissive for pathogen invasion (Titus and Ribeiro, 1988; Ribeiro, 1989; Kamhawi, 2000; Nuttall et al., 2000). In some vector/parasite systems, immune response to arthropod feeding or arthropod bites (probably to salivary proteins) precludes the establishment of the pathogen in the vertebrate host (Bell et al., 1979; Jones and Nuttall, 1990; Wikel et al., 1997; Nazario et al., 1998; Kamhawi et al., 2000). There are two hypotheses (not mutually exclusive) to explain this protection. First, vertebrate immune response to insect salivary proteins, particularly antibodies, may neutralize the effect of the salivary component(s) responsible for pathogen establishment. There is evidence supporting this hypothesis; Morris et al. (2001) demonstrated that animals vaccinated with maxadilan, a potent vasodilator and immunomodulator from Lu. longipalpis, raised antibodies against this salivary protein, which resulted in animals protected against Le. major infection. Second, vertebrate host immune response to salivary proteins creates an inhospitable environment to the pathogen, which is either killed or its future development negatively affected at the site of the bite. In this case, the anti-salivary immune response may be mainly a cellular rather than humoral response. Evidence to support this hypothesis comes from the work of Kamhawi et al. (2000), where animals pre-exposed to Phlebotomus papatasi sand fly bites generated a strong delayed-type hypersensitivity response at the site of the bite; this response was related to protection against Le. major infection. Mice vaccinated with a 15 kDa salivary protein (PpSP15) produced a strong delayed-type hypersensitive response (DTH) and antibodies in black/6 mice, resulting in protection against Le. major infection when parasites were co-inoculates with salivary gland homogenate (SGH). Further evidence suggested that DTH was responsible for the observed anti-Leishmania protection and that antibodies were not necessary (Valenzuela et al., 2001a).
In humans, we have reported a correlation between immune response to Lu. longipalpis salivary proteins and cellular response to Leishmania chagasi, a protective mechanism against leishmaniasis (Barral et al., 2000). Additionally, we have reported a number of salivary antigens recognized by sera of humans, which may be related to protection against leishmaniasis (Gomes et al., 2002); however, the identity of these proteins remains to be solved.
We have hypothesized that the protective effect of sand fly salivary proteins is related to a cellular response to these molecules, probably from CD4 T cells (Valenzuela et al., 2001b). Accordingly, these responses depend on the immunogenetic background of the vertebrate host; thus, targeting a single protein may not be an adequate approach to identify the right vaccine for a population. A broader approach or selection of multiple candidates may be required. High-throughput approaches based on massive cDNA sequencing, proteomics and customized computational biology approaches are helping us to reveal the proteins present in the salivary glands of different blood-feeding arthropods, including the mosquitoes Anopheles gambiae (Francischetti et al., 2002), Anopheles stephensi (Valenzuela et al., 2003), Anopheles darlingi (Calvo et al., 2004) and Aedes aegypti (Valenzuela et al., 2002a), the tick Ixodes scapularis (Valenzuela et al., 2002b) and the bug Rhodnius prolixus (Ribeiro et al., 2004). In the present work, we explored the proteins and transcripts encoded in the salivary gland of Lu. longipalpis, targeting the isolation and full sequencing of the secreted and putative secreted proteins, which are potential markers for vector exposure and vaccine candidates to control Le. chagasi infection.
Materials and methods
Sand fly rearing
Lutzomyia longipalpis (Lutz and Neiva 1912), Jacobina strain, were reared at the Walter Reed Army Medical Research Institute and at the Laboratory of Parasitic Diseases at the NIH, using as larval food a mixture of fermented rabbit feces and rabbit food. Adult sand flies were offered a cotton swab containing 20% sucrose and were used for dissection of salivary glands at 0–2 days following emergence (for cDNA library construction) and 4–7 days following emergence for proteomic analysis. Salivary glands were stored in groups of 20 pairs in 20 μl NaCl (150 mmol l–1):Hepes buffer (10 mmol l–1, pH 7.4) at– 70°C until needed.
Salivary gland cDNA library
Lu. longipalpis salivary gland mRNA was isolated from 80 salivary gland pairs using the Micro-FastTrack mRNA isolation kit (Invitrogen, San Diego, CA, USA). The PCR-based cDNA library was made following the instructions for the SMART cDNA library construction kit (BD-Clontech, Palo Alto, CA, USA) with some modifications (Valenzuela et al., 2001a, 2002c). The obtained cDNA libraries (large, medium and small sizes) were plated by infecting log phase XL1-blue cells (BD-Clontech), and the amount of recombinants was determined by PCR using vector primers flanking the inserted cDNA and visualized on a 1.1% agarose gel with ethidium bromide (1.5 μg ml–1).
Massive sequencing of cDNA library
Lu. longipalpis salivary gland cDNA libraries were plated to a number of ∼200 plaques per plate (150 mm Petri dish). The plaques were randomly picked and transferred to a 96-well polypropylene plate (Novagen, Madison, WI, USA) containing 75 μl of water per well. Four microliters of the phage sample was used as a template for a PCR reaction to amplify random cDNAs. The primers used for this reaction were sequences from the triplEX2 vector. PT2F1 (5′-AAG TAC TCT AGC AAT TGT GAG C-3′) is positioned upstream of the cDNA of interest (5′ end), and PT2R1 (5′-CTC TTC GCT ATT ACG CCA GCT G-3′) is positioned downstream of the cDNA of interest (3′ end). Platinum Taq polymerase (Invitrogen) was used for these reactions. Amplification conditions were: 1 hold of 75°C for 3 min, 1 hold of 94°C for 2 min and 30 cycles of 94°C for 1 min, 49°C for 1 min and 72°C for 1 min 20 s. Amplified products were visualized on a 1.1% agarose gel with ethidium bromide. PCR products were cleaned using the PCR multiscreen filtration system (Millipore, Bedford, MA, USA). Three microliters of the cleaned PCR product was used as a template for a cycle-sequencing reaction using the DTCS labeling kit from Beckman Coulter (Fullerton, CA, USA). The primer used for sequencing, PT2F3 (5′-TCT CGG GAA GCG CGC CAT TGT-3′), is upstream of the inserted cDNA and downstream of the primer PT2F1. Sequencing reaction was performed on a Perkin Elmer 9700 thermacycler (Foster City, CA, USA). Conditions were 75°C for 2 min, 94°C for 2 min and 30 cycles of 96°C for 20 s, 50°C for 10 s and 60°C for 4 min. After cycle-sequencing the samples, a cleaning step was done using the multiscreen 96-well plate cleaning system from Millipore. Samples were sequenced immediately on a CEQ 2000XL DNA sequencing instrument (Beckman Coulter) or stored at –30°C.
Detailed description of the bioinformatic treatment of the data can be found elsewhere (Valenzuela et al., 2002c). Briefly, primer and vector sequences were removed from raw sequences, compared against the GenBank non-redundant (NR) protein database using the stand-alone BlastX program found in the executable package at ftp://ftp.ncbi.nlm.nih.gov/blast/executables/ (Altschul et al., 1997) and searched against the Conserved Domains Database (CDD) (ftp://ftp.ncbi.nlm.nih.gov/pub/mmdb/cdd/), which includes all Pfam (Bateman et al., 2000) and SMART (Schultz et al., 1998, 2000) protein domains. The predicted translated proteins were searched for a secretory signal through the SignalP server (Nielsen et al., 1997). Sequences were clustered using the BlastN program (Altschul and Lipman, 1990) as detailed before (Valenzuela et al., 2002c), and the data presented in the format of Table 1. The electronic version of the table (available on request from ) has additional hyperlinks to ClustalX alignments (Jeanmougin et al., 1998) as well as FASTA-formatted sequences for all clusters.
Full-length sequencing of selected cDNA clones
An aliquot (4 μl) of the λ-phage containing the cDNA of interest was amplified using the PT2F1 and PT2R1 primers (same conditions as described above). The PCR samples were cleaned using the multiscreen PCR 96-well filtration system (Millipore). Cleaned samples were sequenced first with PT2F3 primer and subsequently with custom primers. Full-length sequences were again compared with databases as indicated for the nucleotide sequences above, and the data displayed as in Table 2.
Tris-glycine gels (4–20%), 1 mm thick (Invitrogen), were used. Gels were run with Tris-glycine SDS buffer according to manufacturer's instructions. To estimate the molecular mass of the samples, SeeBlue™ markers from Invitrogen (myosin, bovine serum albumin, glutamic dehydrogenase, alcohol dehydrogenase, carbonic anhydrase, myoglobin, lysozyme, aprotinin, and insulin, chain B) were used. SGH were treated with equal parts of 2×SDS sample buffer (8% SDS in Tris-HCl buffer, 0.5 mol l–1, pH 6.8, 10% glycerol and 1% bromophenol blue dye). Thirty pairs of homogenized salivary glands per lane (approximately 30 μg protein) were applied when visualization of the protein bands stained with Coomassie blue was desired. For amino-terminal sequencing of the salivary proteins, 40 homogenized pairs of salivary glands were electrophoresed and transferred to polyvinylidene difluoride (PVDF) membrane using 10 mmol l–1 CAPS, pH 11.0, 10% methanol as the transfer buffer on a Blot-Module for the XCell II Mini-Cell (Invitrogen). The membrane was stained with Coomassie blue in the absence of acetic acid. Stained bands were cut from the PVDF membrane and subjected to Edman degradation using a Procise sequencer (Perkin-Elmer Corp.). To find the cDNA sequences corresponding to the amino acid sequence obtained by Edman degradation, we used a search program (written in Visual Basic by J. M. C. Ribeiro) that checked these amino acid sequences against the three possible protein translations of each cDNA sequence obtained in the DNA sequencing project. A more detailed account of this program is found elsewhere (Valenzuela et al., 2002c).
Massive cDNA sequencing
An unamplified cDNA library from the salivary glands of the female sand fly Lu. longipalpis was plated, and 550 plaques were randomly picked and sequenced. The resulting sequences were clustered using BlastN with a cutoff of 10E-20, obtaining 143 unique clusters of related sequences. All sequences within each cluster were compared with the non-redundant (NR) protein database using the BlastX program (Altschul et al., 1997) and with the CDD database, containing all Pfam and SMART motifs (Bateman et al., 2000; Schultz et al., 2000), using the RPS-BLAST program (Altschul et al., 1997). The three possible reading frames of each sequence were inspected for long-reading frames with an initial methionine residue followed by at least 40 residues; these were submitted to the SignalP server (http://www.cbs.dtu.dk/services/SignalP-2.0/) for verification of secretory signal peptide. This resulted in the identification of 35 clusters coding for proteins containing a secretory signal peptide. These analyses are summarized in Table 1. Because of space limitations, we show only the first 40 clusters; the remaining clusters can be obtained from the electronic version of this table (available on request from ).
Table 1 shows the most abundant clusters in descending order, from the most abundant sequences to the least. From these 40 clusters, 28 contain sequences with predicted secretory proteins, and those remaining represent clusters for housekeeping genes or unknown sequences without a clear secretory signal peptide. We found 65 sequences (out of 550) of probable housekeeping genes, arranged in 61 clusters, an average of 1.07 sequences per cluster. Examples of these housekeeping genes are: protein disulphide isomerase (cluster 35); 6-phosphogluconate dehydrogenase (cluster 36); ribosomal protein (cluster 39); actin (cluster 54); adenylate kinase (cluster 92); ATP-synthase (cluster 65) and cytochrome oxidase (cluster 57). These clusters represent only the clusters from the middle of Table 1 downwards, with only a few sequences per cluster (most are only one sequence per cluster) representing the low abundant messages.
Another set of cDNA found in this library contains clones that do not have similarities to other genes in the NCBI databank, do not have an assigned function (by CDD analysis) and do not have a secretory signal peptide. We found 34 sequences (out of 550) of these types of genes arranged in 23 clusters (1.48 sequences per cluster).
The most abundant clones found in this cDNA library are the ones coding for secretory proteins. We found 451 cDNA (out of 550 sequences) of potentially secreted proteins arranged in 40 clusters (an average of 11.27 sequences per cluster). Although four of these clusters (cluster 33 with two sequences, cluster 84 with one sequence, cluster 88 with one sequence and cluster 110 with one sequence) resulted in no signal peptide in our analysis, we included these sequences as secretory proteins because they are truncated versions of the salivary adenosine deaminase cDNA and the salivary hyaluronidase cDNA, previously analyzed and reported to be secretory proteins (Charlab et al., 1999). Therefore, the number of cDNAs coding for secretory proteins is 10.53 times greater than the cDNA coding for housekeeping genes and 7.61 times greater than the cDNA coding for proteins with unknown function; overall, cDNA coding for secretory proteins is 4.4 times greater than cDNA coding for non-secreted proteins. The cDNAs coding for secretory proteins represent 82% of the cDNAs sequenced in this sand fly library. In fact, out of the first 40 clusters, 28 belong to cDNA coding for secretory proteins (Table 1).
Proteome analysis of Lu. longipalpis salivary proteins
The supernatant of Lu. longipalpis SGH was separated by one-dimension SDS-PAGE, the proteins transferred to PVDF membrane, and Edman degradation performed on the Coomassie blue-stained bands. Fig. 1 (left side) shows the resulting N-terminal sequence of 17 proteins. By searching our database using an in-house program (written by J. M. C. Ribeiro) similar to the BLOCK program, we were able to identify all the coding cDNA, which correspond to the 14 N-terminal sequences (right side of Fig. 1). We identify by Edman degradation the N-terminus of the cDNA corresponding to clusters 2, 3, 5, 7, 8, 8b, 9, 11, 12, 13, 14, 17, 18, 19, 22, 27 and 115. With the exception of cluster 115, all clusters were within the first 40 clusters. Not all the attempted Edman degradation experiments resulted in a sequence, probably because of the small amount of protein or because these proteins were blocked at the N-terminal end.
Full-length sequence of clones coding for secreted proteins
Because we are primarily interested in the identification and characterization of secreted proteins from this cDNA library, we selected the 35 clusters containing sequences with a clear signal peptide and obtained full sequence on all of them. Table 2 shows the analysis of these sequences, including the signal peptide cleavage site, molecular mass, isoelectric point and best match to NCBI database. The last column of Table 2 indicates whether the clone has been found in the proteome analysis. Extensive analysis is presented in the electronic version of Table 2 (available upon request from ).
The selected cDNAs in Table 2 are arranged in descending order from the cluster containing the largest number of cDNA to the clusters containing the least number (from the most abundant transcripts to the least abundant transcripts in this cDNA library).
Description of cDNAs coding for secreted proteins
Cluster 1 (LJL35; GenBank acc. no. AF132516) – RGD peptide
The cDNA from this cluster codes for a secreted 6 kDa peptide with no similarities to other proteins in databanks. We previously isolated this transcript by PCR subtraction technique (Charlab et al., 1999). The function of this protein remains unknown, although it may inhibit platelet aggregation by interfering with the binding of platelet receptor to fibrinogen by the RGD motif present on it. We did not find this protein in our proteomic analysis, probably because of its small size or because the N-terminal amino acid is blocked.
Clusters 2, 9 and 22 – family of yellow-related proteins
Proteins from this family were found in cluster 2 (GenBank acc. no. AF132518), cluster 9 (GenBank acc. no. AY445935) and cluster 22 on this cDNA library.
Cluster 2 codes for a secreted protein of 45 kDa previously characterized by PCR subtraction (Charlab et al., 1999). The gene coding for this protein was first described in Drosophila melanogaster (Geyer et al., 1986). When comparing with other Diptera, this protein was not found in the salivary glands of Ae. aegypti (Valenzuela et al., 2002a), An. gambiae (Francischetti et al., 2002) or An. stephensi (Valenzuela et al., 2003). It was, however, described in Ae. aegypti whole-larvae extract and purified as a dopachrome-converting enzyme (Johnson et al., 2001). The homologue of Lu. longipalpis yellow protein was also described in the salivary glands of the sand fly P. papatasi (Valenzuela et al., 2001b). The function of this protein in the saliva remains to be elucidated. In the proteomic analysis (Fig. 1), this cluster codes for one of the most abundant proteins (AYVEIGYSLRNIT) in the saliva of this sand fly. Interestingly, a protein with similar molecular mass was the most recognized antigen by the sera of individuals who were exposed to sand flies and developed a cellular immunity to Le. chagasi (Gomes et al., 2002). Therefore, this protein is a candidate marker for epidemiological studies on vector exposure and may be a potential vaccine candidate to control Le. chagasi infection.
This yellow-related protein is highly similar to proteins found in cluster 9 (GenBank acc. no. AY445935) and cluster 22. Fig. 2 shows the ClustalW alignment of this protein with related proteins, including yellow-related salivary proteins from P. papatasi (Valenzuela et al., 2001b), Ae. aegypti midgut dopachrome conversion enzyme (Johnson et al., 2001) and D. melanogaster yellow protein. Similarities among the phlebotomine's yellow proteins are greater than non-salivary yellow from other Diptera and are marked as gray-shaded amino acids (Fig. 2A). Bootstrap neighbor-joining analysis (Fig. 2B) resulted in grouping insect yellow proteins in three distinct clades: the first group contains the Aedes, Anopheles and Drosophila yellow-related proteins (non-salivary); the second group contains the two P. papatasi yellow-related salivary proteins and the Lu. longipalpis yellow salivary protein from cluster 2; and the third group contains the Lu. longipalpis yellow salivary proteins from clusters 9 and 22.
Cluster 3 (LJL08; GenBank acc. no. M77090) – maxadilan
This cluster codes for the most studied salivary protein from a sand fly. It was discovered to be the most potent vasodilator from any organism (Ribeiro et al., 1989) and later as an immunomodulatory molecule (Soares et al., 1998). Recently, Morris et al. reported that mice vaccinated with this 6 kDa peptide produced antibodies against this molecule, and these animals were protected against Le. major infection (Morris et al., 2001). The N-terminal of this protein, XDATXQFRKAIEDDK, was found in the proteomic analysis. Surprisingly, the cDNA we found in this library is only 69% identical to the previously reported maxadilan. Differences may be due to strain differences or the polymorphism reported for this molecule (Lanzaro et al., 1999).
Clusters 4, 6, 7, 18, 25, 26, 34, 67 and 115 – novel peptides from the saliva of Lu. longipalpis
With our approach, we identified a number of small peptides in this cDNA library. Only two peptides were previously reported from the salivary glands of Lu. longipalpis: maxadilan (Ribeiro et al., 1989) and an RGD-containing peptide (Charlab et al., 1999). Following is the description of the nine clusters containing these novel peptides. Sequences of these peptides are shown in Fig. 3.
Cluster 4 (LJL38; GenBank acc. no. AY438269). This cDNA codes for a small secreted peptide of only 2.5 kDa, with no similarities to other proteins in the searched databases. This peptide was previously isolated and fully sequenced from the saliva of Lu. longipalpis by molecular sieving and reverse-phase HPLC (J. M. C. Ribeiro, unpublished results). The function of this protein remains unknown.
Cluster 6 (LJS192; GenBank acc. no. AY438270). This cDNA codes for a secreted protein of 10 kDa, with no similarities to other proteins in the searched databases. Searching a motif database, this protein matched pfam02414 from Borrelia orfA. This cDNA appears to be unique to sand flies and has not been found in any mosquito, tick or triatomine salivary gland cDNA libraries.
Cluster 7 (LJM19; GenBank acc. no. AY438271). This cDNA codes for a secreted protein of 11 kDa, with no similarities to other proteins in the databank. The N-terminus of the predicted salivary protein, SEDXENIFHDNAY, was found in the proteomic analysis.
Cluster 18 (LJL17; GenBank acc. no. AY452695). This cluster codes for a secreted protein of 10 kDa, with no similarities to other proteins in the databank. The N-terminus sequence, NEDYEKQFGDIVD, found in the proteomic analysis, is not exactly in the position predicted by the SignalP program (Fig. 3). The observed N-terminus sequence suggests that a cleavage occurred at position P49–N50, resulting in a peptide of 7 kDa. The cleavage was probably due to either a protease in the saliva or the fact that this peptide bond is unstable at the conditions where the protein was isolated.
Cluster 25 (LJS238; GenBank acc. no. AY455909). This cluster codes for a small secreted protein of 5 kDa, with no similarities to other proteins in databases searched.
Cluster 26 (LJS169; GenBank acc. no. AY455912). This cluster codes for a secreted protein of 11.5 kDa, with no significant similarities to other proteins in databases.
Cluster 34 (LJS105; GenBank acc. no. AY455910). This cDNA codes for a protein of 7 kDa, with no significant similarities to other proteins in the databases searched.
Cluster 67 (LJL124; GenBank acc. no. AY455915). This cDNA codes for a protein of 6 kDa, with no significant similarities to other protein in the searched databases.
Cluster 115 (LJS201; GenBank acc. no. AY455919). This cDNA codes for a protein of 9 kDa, with no significant similarities to other proteins in the databases searched. The N-terminus sequence, GLKDAMEHFKNGKKELTTKDF, found in the proteomic analysis, is not exactly in the position predicted by the SignalP program (Fig. 3). The observed N-terminus sequence suggests that a cleavage occurred at position R36–G37, resulting in a peptide of 7.1 kDa. The cleavage was probably due to a protease in the saliva or the fact that this peptide bond is unstable at the conditions where the protein was isolated.
Cluster 5 (LJM04; GenBank acc. no. AF132517) – PpSP15-like proteins
This cDNA was previously isolated by PCR subtraction and named SL1. This cDNA codes for a protein of 14 kDa and is similar to the protein of 15 kDa from P. papatasi (PpSP15) that conferred protection against Le. major infection. The N-terminus predicted from this cDNA, EHPEEKXIRELAR, was present in our proteomic analysis (Fig. 1). Previously, we reported three similar proteins (three different clusters) on the salivary glands of P. papatasi (PpSP12, PpSP14 and PpSP15). We found only one cluster of this family of proteins in the Lu. longipalpis cDNA library.
Clusters 8, 8b, 11, 14, 17 and 19 – family of putative anticoagulants (C-type lectin)
One of the most abundant families of protein found in this cDNA library is the family of putative anticoagulants with homology to C-type lectins. Six clusters (8, 8b, 11, 14, 17 and 19) belong to this family.
Cluster 8 (LJL91; GenBank acc. no. AY445934). This cDNA codes for a secreted protein of 16 kDa and is similar to the previously described anticoagulant from Lu. longipalpis (Charlab et al., 1999). These proteins contain a C-type lectin or C-type lectin-like domain. This domain functions as a calcium-dependent carbohydrate-binding pocket involved in extracellular matrix organization, pathogen recognition and cell-to-cell interactions. Factor IX/X and Von Willebrand factor binding proteins contain these domains, suggesting that the anticoagulant(s) of Lu. longipalpis may be binding some of the targets of these coagulation factors and inhibiting the blood coagulation cascade. An alignment showing all the C-type lectin proteins found in this cDNA library is shown in Fig. 4A. A ClustalW alignment of this protein and other proteins containing lectin-like domains is shown in Fig. 4B. Phylogenetic tree analysis of C-type lectin-like proteins from Lu. longipalpis and other organisms is shown in Fig. 4C. C-type lectin-like salivary proteins have only been described in Lu. longipalpis sand flies.
Cluster 10 (LJL34; GenBank acc. no. AF132511) – antigen 5-related proteins
This cluster codes for a secreted protein of 29 kDa similar to antigen 5-related protein from vespid venom (Lu et al., 1993). Similar proteins have been isolated from the salivary glands of Ae. aegypti (Valenzuela et al., 2002a) and An. gambiae (Francischetti et al., 2002). We previously isolated this cDNA from the salivary glands of Lu. longipalpis by PCR subtraction (Charlab et al., 1999).
Cluster 12 (LJL13; GenBank acc. no. AF420274) – D7-related protein
This cluster codes for a secreted protein of 26 kDa and belongs to the D7 family of proteins (Valenzuela et al., 2002c). One salivary D7 protein from An. stephensi was shown to be an inhibitor of the blood coagulation Factor XII (Isawa et al., 2002). The N-terminus of the predicted salivary protein, WQDVRNADQTL, was found in the proteomic analysis (Fig. 1).
Cluster 13 (LJL23; GenBank acc. no. AF131933) – apyrase
This cluster codes for a secreted protein of 35 kDapreviously characterized by PCR subtraction (Charlab et al., 1999). This protein belongs to a family first described in Cimex lectularius and shown to be a salivary apyrase (Valenzuela et al., 1998). Later, the clone coding for an homologous cDNA from P. papatasi was expressed and demonstrated to be a calcium-dependent apyrase (Valenzuela et al., 2001b). Recently, a Cimex homologue from humans was demonstrated to have ATPase and low ADPase activity (Murphy et al., 2003), while the insect apyrases have high ADPase activities. The N-terminus of the predicted salivary protein, APPGVEWYHFGL, was found in the proteomic analysis (Fig. 1). A protein with similar molecular mass to the cDNA prediction for this cluster (35 kDa) was recognized with high frequency among serum of individuals exposed naturally to sand flies who developed a cellular immunity to Le. chagasi (Gomes et al., 2002). Therefore, this protein may be a good marker for vector exposure and potential vaccine candidate to control Le. chagasi infection.
Clusters 16, 21, 23, 58 – novel sequences
Cluster 16 (LJL143; GenBank acc. no. AY445936). This cluster codes for a secreted protein of 32 kDa. It has no significant homology to known proteins from GenBank. This protein may only be present in phlebotomines, because we did not find similar sequences in either the Drosophila or Anopheles genome or any other databases.
Cluster 21 (LJM114; GenBank acc. no. AY455907). This cluster codes for a secreted protein of 14 kDa, with no similarities to other proteins in the databank. It is a relatively abundant cDNA in this library and is therefore a novel protein in sand fly saliva.
Cluster 23 (LJM78; GenBank acc. no. AY455908). This cluster codes for a secreted protein of 37 kDa, with no similarities to other proteins on the databases searched. This cDNA represents a novel protein from the saliva of a sand fly.
Cluster 58 (LJS03; GenBank acc. no. AY455914). This cDNA codes for a secreted protein of 15 kDa, with no significant similarities to other proteins in the searched databases.
Cluster 20 (LJL04; GenBank acc. no. AY455906) – collagen binding-like proteins
This cluster codes for a secreted protein of 29 kDa, with similarities to PpSP32 salivary protein from P. papatasi and to the collagen adhesion protein from Bacillus cereus (acc. no. NP_830673). This 29 kDa protein is rich in glycine (51 amino acids), arginine (25 amino acids), proline (25 amino acids) and lysine (23 amino acids). These four amino acids represent 44% of the amino acids in this protein. Additionally, the pI of this protein is 10.2, making it a very basic protein. Different short repeats were identified in this protein, particularly the repeats GQG and GTRP (Fig. 5). Because of its similarities to the collagen-binding protein from B. cereus, the high pI, the richness in glycine and the repeated amino acid motifs, this salivary protein may bind to the extracellular matrix proteins of the vertebrate host.
Cluster 27 (LJL11; GenBank acc. no. AF132510) – 5′-nucleotidase
This cluster codes for a secreted protein of 61 kDa. It is the secreted 5′-nucleotidase from Lu. longipalpis, previously isolated by PCR subtraction (Charlab et al., 1999). The N-terminus of the predicted salivary protein, EDGSYEIIILHTN, was found in the proteomic analysis (Fig. 1), suggesting that this protein is very abundant in the saliva of this insect.
Cluster 38 (LJL09; GenBank acc. no. AY455911) – angiotensin converting enzyme
This cDNA codes for a secreted protein of 71 kDa, with high similarities to angiotensin converting enzyme (ACE) from An. gambiae, D. melanogaster, chicken and human (Fig. 6). The function as a peptidase remains to be elucidated. The activity may be similar to the kininase activity observed in the tick I. scapularis (Ribeiro and Mather, 1998).
Cluster 56 (LJM26; GenBank acc. no. AY455913) – serine protease inhibitor
This cDNA codes for a secreted protein of 49 kDa, with high similarities to the serpin family of protease inhibitors (Fig. 7). This protein may function as an anticoagulant in the saliva or may be responsible for regulation of the insect immune response. Anticomplement activity in Lu. longipalpis has been recently described (Cavalcante et al., 2003); because of its putative anti-protease activity, this cDNA may code for the salivary anti-complement activity.
Cluster 71 (LJL138; GenBank acc. no. AY455916) – endonuclease
This cDNA codes for a protein of 44 kDa, with high similarities to non-specific RNA/DNA endonucleases from different organisms. A similar cDNA was isolated from the salivary glands of the tsetse fly (Glossina morsitans) and Culex quinquefasciatus (Ribeiro et al., 2004) and the protein was named TsaI (Li et al., 2001). The function as an endonuclease remains to be elucidated. The presence of endonuclease in the saliva of an insect is puzzling and it is interesting to note that enzymes directed to nucleotide metabolism, such as 5′-nucleotidase and adenosine deaminase, are present in this sand fly (Charlab et al., 2000).
Cluster 97 (LJS138; GenBank acc. no. AY455917) – translocon-associated protein
This cDNA codes for a secreted protein of 16 kDa, with high similarities to translocon-associated protein from different organisms and may be associated with an ER protein export function.
Cluster 113 (LJS193; GenBank acc. no. AY455918) – palmitoyl thioesterase
This cDNA codes for a protein of 32 kDa, with high similarities to palmitoyl-(protein) thioesterase from different organisms.
In this work, we have isolated and characterized the most abundant secretory proteins and transcripts from the salivary glands of the sand fly Lutzomyia longipalpis. When comparing this PCR-based cDNA library construction/massive sequencing approach to the PCR subtraction technology we have used before (Charlab et al., 1999), we have tripled the finding of cDNA coding for secreted proteins. Additionally, the isolated cDNA are, in the majority, full-length cDNA, making the process of isolation and sequencing less laborious than in the PCR subtraction method, which cuts the full-length cDNA into smaller pieces.
It is interesting to note that we needed to sequence only 550 cDNA to identify the majority of secretory salivary proteins for this sand fly. The most abundant transcripts in this Lu. longipalpis salivary gland cDNA library correspond to putative secreted proteins, which indicates that most transcripts in this organ are directed to secretion. From these 550 cDNA, we obtained 143 different clusters of related sequences, suggesting that the cDNA library was diverse enough to have a good representation of the transcripts present in the salivary gland of this sand fly.
From the 143 different clusters or families of proteins, we identified 35 clusters of proteins containing a secretory signal peptide. Interestingly, the majority of the cDNAs sequenced (451 out of 550) were in these 35 clusters. We confirmed by Edman degradation the presence of 17 secreted proteins in the salivary gland of this sand fly. These 17 proteins are included in these 35 clusters. In fact, all the proteins that resulted with N-terminus data had a corresponding cDNA in these 35 clusters, and the predicted signal cleavage site (SignalP program for these 35 clusters) matched perfectly with the N-terminus obtained, with the exception of two proteins. These two proteins probably underwent further processing by a protease.
When comparing these 35 clusters with sequences deposited in existing databases, we found that nine of these cDNAs correspond with already described Lu. longipalpis proteins. We found eight cDNAs with low homology to already described Lu. longipalpis proteins, suggesting that they belong to related families of proteins but are different enough to be clustered in different groups. We found 10 cDNAs with homology to other proteins in databases, including an angiotensin converting enzyme, a protease and a palmitoyl-hydrolase; the rest of the cDNAs (seven) did not match to any protein in the existing databases. These are probably novel proteins only present thus far in this sand fly.
The sand fly Lu. longipalpis is the main vector of Le. chagasi, the causal agent of visceral leishmaniasis. The relationship between Lu. longipalpis saliva and human visceral leishmaniasis was recently undertaken. Barral et al. (2000) studied serum from children living in an endemic area of visceral leishmaniasis and found a positive correlation between children producing antibodies against Lu. longipalpis and delayed-type hypersensitive response (DTH) to Le. chagasi. DTH to Leishmania is a marker for protection against this parasite, while positive serology is a sign of poor prognosis. On the other hand, no correlation was found between children producing antibodies to Lu. longipalpis and positive serology to Le. chagasi. In an extension of this work, we recently reported that children from an endemic area of leishmaniasis, who in a 6-month period developed antibodies to Le. chagasi, did not produce detectable antibodies to Lu. longipalpis salivary proteins. On the other hand, individuals that developed a cellular response to Le. chagasi (DTH) produced IgG, IgG1 and IgE antibodies to Lu. longipalpis salivary proteins (Gomes et al., 2002). These data support the hypothesis that induction of an immune response to salivary proteins from Lu. longipalpis may facilitate a protective immune response to Le. chagasi.
The salivary proteins, as well as the cDNA coding for secreted proteins identified in this work, are good candidates as markers for vector exposure to continue performing epidemiological studies with single recombinant proteins instead of whole SGH. In previous work, we identified five salivary antigens by western blot that resulted in molecular masses of 45, 44, 43, 35 and 17 kDa (Gomes et al., 2002). In the present study, we have identified a limited number of cDNAs that code for a molecular mass similar to these antigens; these are the yellow-related proteins (clusters 2, 9 and 22), an apyrase (cluster 13) and an anticoagulant protein (clusters 8, 14, 19). Recombinant expression of these proteins will determine whether these are the antigens recognized by individuals exposed to sand flies that showed protection to Le. chagasi infection.
Ultimately, these salivary proteins represent good vaccine candidates to control Leishmania infection. We reported that a delayed-type hypersensitivity response to a sand fly salivary protein was responsible for the protective effect against Le. major infection in mice (Valenzuela et al., 2001a). Because this type of protection is T cell dependent and the response may depend on the immunogenetic background of the host, focusing on a single protein may not be a proper vaccine strategy in an outbred population. This high-throughput approach is providing a larger candidate repertoire to select the best vaccine candidate or the best cocktail of protective salivary components.
The authors thank Dr José M. C. Ribeiro for help in bioinformatic analysis and critical review and comments on this manuscript, and Nancy Shulman for editorial assistance.
- © The Company of Biologists Limited 2004