Salivary Gland Transcriptomes and Proteomes of Phlebotomus tobbi and Phlebotomus sergenti, Vectors of Leishmaniasis

Background Phlebotomus tobbi is a vector of Leishmania infantum, and P. sergenti is a vector of Leishmania tropica. Le. infantum and Le. tropica typically cause visceral or cutaneous leishmaniasis, respectively, but Le. infantum strains transmitted by P. tobbi can cause cutaneous disease. To better understand the components and possible implications of sand fly saliva in leishmaniasis, the transcriptomes of the salivary glands (SGs) of these two sand fly species were sequenced, characterized and compared. Methodology/Principal Findings cDNA libraries of P. tobbi and P. sergenti female SGs were constructed, sequenced, and analyzed. Clones (1,152) were randomly picked from each library, producing 1,142 high-quality sequences from P. tobbi and 1,090 from P. sergenti. The most abundant, secreted putative proteins were categorized as antigen 5-related proteins, apyrases, hyaluronidases, D7-related and PpSP15-like proteins, ParSP25-like proteins, PpSP32-like proteins, yellow-related proteins, the 33-kDa salivary proteins, and the 41.9-kDa superfamily of proteins. Phylogenetic analyses and multiple sequence alignments of putative proteins were used to elucidate molecular evolution and describe conserved domains, active sites, and catalytic residues. Proteomic analyses of P. tobbi and P. sergenti SGs were used to confirm the identification of 35 full-length sequences (18 in P. tobbi and 17 in P. sergenti). To bridge transcriptomics with biology P. tobbi antigens, glycoproteins, and hyaluronidase activity was characterized. Conclusions This analysis of P. sergenti is the first description of the subgenus Paraphlebotomus salivary components. The investigation of the subgenus Larroussius sand fly P. tobbi expands the repertoire of salivary proteins in vectors of Le. infantum. Although P. tobbi transmits a cutaneous form of leishmaniasis, its salivary proteins are most similar to other Larroussius subgenus species transmitting visceral leishmaniasis. These transcriptomic and proteomic analyses provide a better understanding of sand fly salivary proteins across species and subgenera that will be vital in vector-pathogen and vector-host research.


Introduction
Sand flies are bloodsucking nematoceran Diptera that transmit the protozoan parasites of the genus Leishmania. Similar to that of other bloodsucking arthropods, sand fly saliva comprises antihemostatic, immunomodulatory, and antigenic components. The saliva is deposited into the host skin every time the sand fly ingests a blood meal to facilitate feeding. Also during the bite by an infected sand fly, Leishmania parasites are egested into the wound with the saliva. Sand fly saliva can enhance Leishmania infection in naive mice [1,2]. Conversely, pre-exposure of mice to sand fly saliva conferred a protective effect against Leishmania infection [3,4]. Even single salivary proteins have been characterized as potential Leishmania vaccine candidates in mouse, hamster, and dog models of cutaneous or visceral leishmaniasis [5][6][7][8][9].
To broaden the repertoire of subgenus Larroussius salivary proteins and provide the first report from a subgenus Paraphlebotomus sand fly, we prepared and analyzed the transcriptomes and proteomes of P. tobbi and P. sergenti, both proven vectors in the Old World. Phlebotomus sergenti, subgenus Paraphlebotomus, is the main vector of Le. tropica, principally an agent of cutaneous leishmaniasis [29][30][31]. Phlebotomus tobbi, on the other hand, is an important vector of Le. infantum [32] together with the taxonomically related P. ariasi and P. perniciosus, sand flies of the subgenus Larroussius [29,33,34]. In contrast to other members of the subgenus, P. tobbi transmits the cutaneous form of the Le. infantum [32]. Additionaly, we characterized P. tobbi antigens, glycoproteins, and hyaluronidase activity; the later one compared with 6 sand fly species belonging to vectors of cutaneous or visceral leishmaniases.

Sand fly salivary glands
Colonies of P. tobbi (originating from Turkey), P. papatasi (Turkey), P. sergenti (Israel), P. argentipes (India), P. arabicus (Israel), P. perniciosus (Spain), and L. longipalpis (Brazil) were kept in the insectary of Charles University in Prague as described in [35]. The P. sergenti colony, originating from Turkey, was reared in similar conditions at the Laboratory of Malaria and Vector Research, National Institutes of Health (Rockville, MD, USA). For mRNA extraction, salivary glands (SGs) from non-bloodfed 1-to 2-day-old female sand flies were dissected and stored in RNA Later (Ambion, Inc., Austin, TX, USA). For other assays and analysis, SGs of non-bloodfed 5-to 7-day-old females were stored at 270uC; SGs were stored in NuPAGE LDS sample buffer (Invitrogen, Carlsbad, CA, USA) for proteome analysis and in Tris buffer (20 mM Tris, 150 mM NaCl, pH 7.8) for hyaluronidase assays, affinity blot, and immunoblot. Before use, samples were homogenized by three freeze-thaw cycles in liquid nitrogen. Protein concentration in resulting SG homogenate (SGH) was measured on Qubit Fluorometer (Invitrogen) following manufacturer's guidelines.

Salivary gland cDNA library construction and sequencing
An SG cDNA library was constructed from P. sergenti (Turkey) and P. tobbi. MicroFastTrack mRNA isolation kit (Invitrogen) was used to isolate SG mRNA from 40 SG pairs dissected into 20 ml of RNA Later (Ambion). A cDNA library was constructed using SMART TM cDNA Library Construction Kit (BD Clontech, Palo Alto, CA, USA) following the manufacturer's protocol, with some modifications as described in [36]. For each species, three cDNA libraries were constructed according to PCR product size -large, medium, and small. PCR amplicons were washed and concentrated to 4-7 ml on Microcon YM-100 columns (Millipore, Billerica, MA, USA). Concentrated samples (3 ml) were ligated into the lTripleEx2 vector and packed into the phage particles with Gigapack III Gold Packaging Extract (Stratagene, La Jolla, CA, USA). Phage libraries were used to infect the log-phase XL-1 Blue Escherichia coli (Clontech) plated onto four LB agar plates per each library size. Transfected plaques were randomly selected and transferred into 96-well V-shape plates with 75 ml of ultrapure water per well. Four 96-well plates of phage were picked per each library size, resulting in 12 plates (1,152 clones) per sand fly species. Phages (3 ml) were subjected to PCR using FastStart PCR Master Mix (Roche, Molecular Biochemicals, Indianopolis, IN, USA) and vector-specific primers (PT2F1 59-AAGTACTCTAG-CAATTGTGAGC-39 and PT2R1 59-CTCTTCGCTAT-TACGCCAGCTG-39). Amplification conditions were as follows: 1 hold of 75uC for 3 min, 1 hold of 94uC for 2 min, 34 cycles of 94uC for 1 min, 49uC for 1 min, and 72uC for 2 min. The final elongation step lasted for 10 min at 72uC. Products were cleaned using ExcelaPure 96-well UF PCR Purification Plates (Edge Biosystems, Gaithersburg, MD, USA) and cleaned PCR products were used as a template for cycle-sequencing reaction using BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Fullerton, CA, USA) and a vector-specific forward primer (PT2F3 59-CTCGGGAAGCGCGCCATTGT-39). Products of cycle-sequencing reaction were cleaned using Sephadex and MultiScreen HV Plates (Millipore), dried, resuspended in formamide, and stored at 220uC until sequenced on an ABI 3730XL 96-Capillary DNA Sequencer (Applied Biosystems).

Proteome analysis
For mass spectrometric (MS) analysis, SGH samples of P. sergenti (Turkey) and P. tobbi were dissolved in Laemmli sample buffer in parallel with or without 2-mercaptoethanol and electrophoretically separated on 12% polyacrylamide SDS minigel with initial voltage 80 V and 120 V upon entry of sample to the gel. Gels were stained for total proteins with Coomassie Brilliant Blue R-250. Individual bands were cut, destained and digested as was described in [37]. Samples (0.5 ml) were transferred to a 384 spot stainless steel MALDI target (AB Sciex, Framingham, MA, USA) and let to dry. Dried droplets were covered with a 0.5 ml drop of alpha-cyanohydroxycinnamic acid (Fluka, Switzerland) solution (2 mg/ml in 80% acetonitrile) and allowed to dry. Spectra were acquired with 4800 Plus MALDI TOF/TOF analyzer (AB Sciex) equipped with a Nd:YAG laser (355 nm; firing rate 200 Hz). Voltages were set as follows: source1 20 kV, grid1 16 kV, source1 lens 10 kV, lens1 5 kV, mirror1 14.085 kV, mirror2 20.3 kV and reflector detector 1.905 kV. Digitizer bin size was set to 0.5 ns, vertical scale 0.5 V,

Author Summary
Phlebotomine female sand flies require a blood meal for egg development, and it is during the blood feeding that pathogens can be transmitted to a host. Leishmania parasites are among these pathogens and can cause disfiguring cutaneous or even possibly fatal visceral disease. The Leishmania parasites are deposited into the bite wound along with the sand fly saliva. The components of the saliva have many pharmacologic and immune functions important in blood feeding and disease establishment. In this article, the authors identify and investigate the protein components of saliva of two important vectors of leishmaniasis, Phlebotomus tobbi and P. sergenti, by sequencing the transcriptomes of the salivary glands. We then compared the predicted protein sequences of these salivary proteins to those of other bloodsucking insects to elucidate the similarity in composition, structure, and enzymatic activity. Finally, this descriptive analysis of P. tobbi and P. sergenti transcriptomes can aid future research in identifying molecules for epidemiologic assays and in investigating sand fly-host interactions.
Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org vertical offset 0.0, input bandwith 500 MHz. Spectra were externally calibrated using ProteoMass peptide MALDI calibration kit (Sigma-Aldrich). Spectra were recorded in the range 700 to 4000 Da, focus mass 2100 Da. Spectra were summed from 40 positions per 50 shots, 2000 shots in total. Spectra were processed by 4000 Series Explorer version 3.5.3 (AB Sciex) without smoothing; baseline subtraction was performed with peak width set to 50. Spectra were deisotoped and peaks with a local signal-to-noise ratio greater than 5 were picked and searched by local Mascot v. 2.1 (Matrix Science, Boston, MA, USA) against a database of protein sequences derived from the cDNA library. Database search criteria were as follows: enzyme: trypsin; taxonomy: none; fixed modification: carbamidomethylation; variable modification: methionine oxidation; peptide mass tolerance: 80 ppm; one missed cleavage allowed.
Only hits that were scored as significant (P,0.05) were included.
The data associated with this manuscript may be downloaded from ProteomeCommons.org Tranche using the following hash: mCZfFsOLaBtSfR+Jh6o8OwEgjrqDp4m3VntpJdAPqPGFNzNp-TPry8IhEuGeLw9 TmpHcTRMSiiuiNNRL/6xP65TLvyNwAAAAAAAADGQ = = . The hash may be used to prove exactly what files were published as part of this manuscript's data set, and the hash may also be used to check that the data has not changed since publication.
Sequences with Phred quality scores lower than 25 were removed, as well as primers and vector sequences. Resulting sequences were grouped based on nucleotide homology of 90% identity over 100 residues and aligned into consensus transcript sequences (contigs) using the CAP3 sequence assembly program. BLAST programs were used to compare contigs and singletons (contigs with a single sequence) to the non-redundant protein database of the NCBI, the Gene Ontology database (GO) [43], to COG conserved domains database [44], Protein Family database (Pfam) [45], Simple Modular Architecture Tool database (SMART) [46], and to rRNA Nucleotide Sequences, and Mitochondrial and Plasmid Sequence (MIT-PLA) databases available from NCBI. The three frame translations of each dataset were submitted to the SignalP server [47] to detect signal peptides. The grouped and assembled sequences, BLAST results, and SignalP results combined by dCAS software in an Excel spreadsheet were manually verified and annotated. Additionally, glycosylation sites were determined in selected sequences using NetNGlyc prediction server [48]. For phylogenetic analysis, protein sequences without signal peptide were aligned using ClustalX (version 2.0) [49] with related sequences obtained from GenBank and manually refined in BioEdit 7.0 editing software. For each alignment, best substitution matrix was determined by ProtTest software, version 2.0 [50]. This matrix was then used by TREE-PUZZLE 5.2 [51] to reconstruct maximum likelihood phylogenetic trees from the protein alignments using quartet puzzling with 1000 puzzling steps. Resulting trees were visualized in MEGA 4 [52]. All protein and nucleotide accession numbers mentioned in the text, tables and figures are listed in Text S1.

Hyaluronidase activity
Hyaluronidase activity was compared between seven sand fly species: P. tobbi, P. sergenti (Israel), P. papatasi, P. argentipes, P. arabicus, P. perniciosus, and L. longipalpis. Hyaluronidase activity in SGs was quantified using a sensitive assay on microtitration plates coupled with biotinylated HA (bHA). bHA, prepared as described in [53], was immobilized onto Covalink NH microtiter plates (Nunc, Placerville, NJ, USA) using the method in [54] at a final concentration of 1 mg bHA per well. The plates were incubated overnight at 4uC and washed three times in PBS (containing 2 M NaCl and 50 mM MgSO 4 , pH 7.2). Plates with immobilized bHA were blocked with 1% BSA in PBS for 45 min, washed and equilibrated to pH 5.0 (0.1 M acetate, 0.1 M NaCl, 0.1% Triton X-100, pH 5.0), the pH optimum for sand fly salivary hyaluronidase [53]. SGHs were incubated for 45 min at 37uC in triplicate at a final concentration of 0.5 gland per well. As a standard, Putatively secreted salivary proteins from Phlebotomus sergenti with cluster name, GenBank accession number (GenBank Accn), presence in the proteome analysis as confirmed by mass spectrometry ( Figure 1B), putative mature protein features (pI, predicted isoelectric point; Mw, predicted molecular weight; AA, number of amino acid residues), and best match to non-redundant protein database. doi:10.1371/journal.pntd.0001660.t002 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org   sergenti (Ser), P. tobbi (Tob), and Lutzomyia longipalpis (Lon). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Conserved cysteine residues are indicated above the alignment by letter C and T cell epitopes predicted for P. duboscqi by Kato et al. [26] are indicated by asterisk (* For hyaluronidase zymography, 8% polyacrylamide gels (0.75 mm thick) were copolymerized with 0.002% hyaluronic acid (HA). As the hyaluronidase activities and band patterns varied among sand fly species, different loads were used per lane to obtain bands of equal intensity. The equivalent of 1/2 gland (L. longipalpis and P. sergenti) or 1/20 gland (other tested species) was loaded for zymography under non-reducing conditions, and the equivalent of 2.5 glands (L. longipalpis and P. sergenti) or 1/4 gland (other tested species) was loaded for zymography under reducing conditions. The total protein content per lane was as follows (non-reducing/ reducing conditions): L. longipalpis = 110/550 ng; P. papatasi = 12.5/62.5 ng; P. sergenti = 140/700 ng; P. argentipes = 14/ 70 ng; P. arabicus = 10.5/52.5 ng; P. tobbi = 10/50 ng; P. perniciosus = 10.5/52.5 ng. For reducing conditions, samples were treated with 3% 2-mercaptoethanol for 40 min at 45uC. SDS-PAGE electrophoresis was carried out using Mini-Protean II apparatus (Bio-Rad, Hercules, CA, USA) and constant voltage at 150 V. After electrophoresis, gels were rinsed 2620 min in 0.1 M Tris, pH 7.8, and 20 min in 0.1 M acetate buffer, pH 5.5 (both with 1% Triton X-100 to wash out SDS) and then incubated in 0.1 M acetate buffer (without detergent) for 120 min at 37uC. The gels were then washed in water, soaked in 50% formamide for 30 min and stained in Stains-all (Sigma, St. Louis, MO, USA) solution (100 mg/ml in 50% formamide) for 24 h in the dark. Hyaluronidase activity was visible as a pink band on a dark blue background.

Immunoblotting
Immunoblot was performed using P. tobbi SGH separated by SDS-PAGE on 10% polyacrylamide gel under non-reducing conditions using the Mini-Protean III apparatus (Bio-Rad). Separated proteins were electrotransferred onto nitrocellulose (NC) membrane by iBlot Dry Blotting System (Invitrogen). After transfer, the NC membrane was cut into strips with the equivalent of four glands/strip and free binding sites were blocked by 5% low-fat dried milk in 20 mM Tris buffer with 0.05% Tween (Tris-Tw) overnight at 4uC. The strips were then incubated with serum obtained from rabbit repeatedly exposed to P. tobbi females. Serum was diluted 1:250 in Tris-Tw and incubated with P. tobbi proteins for 1 h, followed by 1 h incubation with peroxidase-conjugated swine anti-rabbit IgG (Sevapharma, Prague, Czech Republic) diluted 1:1,000 in Tris-Tw. Substrate solution contained Tris buffer, diaminobenzidine, and H 2 O 2 .

Affinity blotting
Affinity blot was performed using P. tobbi SGH separated and electrotransferred as described for Immunoblot. After transfer, free binding sites on NC membrane were blocked by 5% BSA in 20 mM Tris-Tw overnight at 4uC. The strip was then incubated for 1 h at room temperature with biotinylated lectin from Canavalia ensiformis (Concanavalin A, Sigma) diluted 0.2 mg/ml in Tris-Tw. To control the reaction specificity, another strip was incubated with lectin preincubated for 30 min with the ligand, 0.5 M methyl-a-D-mannopyranoside. Avidin-peroxidase (Sigma) was added at a final concentration of 2.5 mg/ml and incubated for 1 h at room temperature. Substrate solution contained Tris buffer, diaminobenzidine, and H 2 O 2 .

Ethics statement
All animals used in this study were maintained and handled strictly in accordance with institutional guidelines and legislation for the care and use of animals for research purpose Czech Act No. 246/1992 coll. on Protection Animals against Cruelty in present statues at large that complies with all relevant European Union and international guidelines for experimental animals. The experiments were approved by the Committee on the Ethics of Animal Experiments of the Charles University in Prague (Permit Number: 24773/2008-10001) and were performed under the Certificate of Competency (Registration Number: CZU 934/05) in accordance with the Examination Order approved by Central Commission for Animal Welfare of the Czech Republic. All efforts were made to minimize suffering of experimental animals within the study.

Salivary gland transcripts analysis
Phlebotomus tobbi and P. sergenti cDNA libraries were constructed from SGs of female sand flies dissected one day after emergence. From each cDNA library, 1,152 randomly selected clones were sequenced. Obtained ESTs were deposited in the NCBI dbEST database under accession numbers GW814275-GW815416 (1,142 sequences) for P. tobbi and GW813185-GW814274 (1,090  Phlebotomus arabicus (Ara), P. argentipes (Arg), P. ariasi (Ari), P. duboscqi (Dub), P. papatasi (Pap), P. perniciosus (Per), P. sergenti (Ser), P. tobbi (Tob), Lutzomyia longipalpis (Lon), and related sequences from Cimex lectularius and Homo sapiens. Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Nucleotide binding sites (*) and Ca 2+ binding sites (+), as predicted for human apyrase by Dai et al. [70], are indicated. The position of E92Y point mutation of human apyrase described by Yang  sequences) for P. sergenti. High-quality sequences were grouped together based on sequence homology, and resulting assembled sequences were analyzed using the dCAS cDNA annotation software [38] and verified by manual annotation. In the P. tobbi cDNA library, 997 high-quality sequences were grouped into 68 contigs and 125 singletons (one sequence in cluster); in P. sergenti, 853 high-quality sequences were grouped into 56 contigs and 196 singletons.
Similar to other sand flies studied so far, the most abundant transcripts in both libraries were those coding for putative secretory proteins. BLAST comparison of translated nucleotide sequences with the non-redundant (NR) protein database showed high similarity with other sand fly secreted salivary proteins. In P. tobbi, 81 clusters containing 863 sequences (average 10.7 sequences per cluster) matched to sand fly salivary proteins. Of them, we found 62 clusters (796 sequences) with predicted signal peptide sequence. In P. sergenti, 50 clusters containing 553 sequences (average 11.1 sequences per cluster) matched to sand fly salivary proteins. Of them, 32 clusters (482 sequences) with predicted signal protein sequence were found. Tables 1 and 2 list representative secreted salivary proteins from P. tobbi and P. sergenti, respectively, deposited into NCBI GenBank database. The tables show GenBank accession numbers, putative mature protein features, best match to NR protein database, and presence in the proteome analysis as confirmed by MS ( Figure 1). Additionally, Figure 1A and 1B show detailed analysis of MS results for P. tobbi and P. sergenti, respectively, including cluster name, Gen Bank accession number, and molecular weight of mature proteins under reducing and non-reducing conditions.
The putative secreted salivary proteins of P. tobbi and P. sergenti could be divided into ten main protein families ( Figure S1): antigen 5-related protein, apyrase, hyaluronidase, D7-related and PpSP15like protein (odorant-binding proteins superfamily), ParSP25-like protein, PpSP32-like protein, yellow-related protein, the 33-kDa salivary proteins, and the 41.9-kDa superfamily. The following paragraphs describe these families in detail, focusing on protein family characteristics, possible function, biochemical, immuno-
All sand fly Ag5r proteins have similar predicted molecular mass (ranging from 28.8 to 31.2 kDa) and are alkaline (Table S1). In P. tobbi and P. sergenti, the predicted molecular mass corresponded well with the one measured in proteomic analysis ( Figure 1, Tables 1 and 2) suggesting single-domain protein and negligible post-translational modifications. We identified 14 highly conserved cysteine residues proportionally distributed through the whole sequence length (Figure 2), possibly involved in disulfide bonding.
Although the members of this family were described in sialotranscriptomes of all bloodsucking arthropods characterized [55,56], their role is mostly unknown with a few exceptions. In Stomoxys calcitrans, Ag5r protein possesses immunoglobulin Fc binding activity [58]. In Tabanus yao, members of the Ag5r protein family can probably serve as an inhibitors of angiogenesis (RTS disintegrin motif) [59] or a potent platelet inhibitor (RGD motif) [60]. The Ag5r proteins are not specific for salivary glands thus they may possess other functions not associated with feeding [23,56].
Several studies showed antigenic properties associated with Ag5r proteins. Plasmid coding for Ag5r protein from P. ariasi (ParSP05/AAX44092) induced a cell-mediated immune response in Swiss Webster mice [27], showing that sand fly Ag5r proteins might modulate cell-mediated host immune response. This presumption is also supported by several T cell epitopes predicted for P. duboscqi Ag5r proteins [26] that include regions highly conserved among sand flies (Figure 2). Antibody response to sand fly Ag5r proteins was demonstrated in P. perniciosus; Ag5r protein (PpeSP07/ABA43055) reacted with IgG antibodies from sera of P. perniciosus bitten dogs [21]. In other bloodsucking diptera, Ag5r proteins are mostly associated with IgE antibody response. Ag5r protein of Simulium vittatum seems to be the major allergen for insect bite hypersensitivity sharing common IgE-binding epitopes with Ag5r protein from Culicoides nubeculosus [61,62]. Specific anti-Ag5r IgE antibodies were also observed in Ugandan individuals bitten by Glossina morsitans [63].
Phylogenetic analysis of Ag5r proteins from sand flies and other insects showed a strongly supported distinct clade of sand fly Ag5r proteins (Figure 3) similar to a previous analysis by [28]. The relationship within the sand fly clade reflected phylogenetic relationship within phlebotomine sand flies [33], showing three distinct branches: clade I with species belonging to subgenera Euphlebotomus, Larroussius, and Adlerius; clade II with Phlebotomus and Paraphlebotomus species (P. papatasi, P. duboscqi, and P. sergenti); and Lutzomyia in clade III (Figure 3).

Apyrase
Apyrase (EC 3.6.1.5) appears to be a universal enzyme used to prevent blood coagulation by diverse hematophagous animals such as bloodsucking leeches, ticks, triatomine bugs, fleas, and mosquitoes. This enzyme hydrolyses both ATP and ADP to AMP, thus destroying an important physiologic stimulus of platelet aggregation released from damaged tissues and blood cells. Apyrases of bloodsucking insects are divided into three families: CD-39 (the actin/heat shock 70/sugar kinase superfamily); 59nucleotidase; and Cimex-type [55,56].
Orthologs of the Cimex apyrase family have also been identified in vertebrates and termed calcium-activated nucleotidases (CANs) [68]. In contrast to sand flies, human soluble CAN-1 (SCAN-1) preferentially hydrolyses UDP and GDP; however, the engineered SCAN-1 mutant Glu92Tyr shows five times and seven times higher hydrolysis activity for ADP and ATP, respectively [69]. This mutated tyrosine is conserved among species of the genus Phlebotomus (Figure 4), supporting its key role in substrate specificity for phlebotomine apyrases [69]. In human SCAN-1, other amino acid residues essential for binding nucleotide and Ca 2+ were identified [70], some of them being absolutely conserved among the analyzed apyrase proteins (Asp44, Ser100, Asp114, Glu216, Arg232, Ser277), while others were uniformly mutated within sand fly apyrases (Asp101Glu, Gly160Ser, Ile214Trp) ( Figure 4).
Besides hydrolyzing activity, sand fly apyrases also possess antigenic properties. Antibodies from dogs experimentally or naturally exposed to P. perniciosus strongly recognized PpeSP01 (ABB00906) and PpeSP01B (ABB00907) apyrases [21]. In humans naturally exposed to sand flies, anti-sand fly saliva IgG antibodies recognized a protein band corresponding, in molecular weight, to apyrase [11,12]. Moreover, antibodies elicited by P. duboscqi saliva refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Active site residues (*) and cysteine residues forming disulfide bridges (C) as predicted for Apis hyaluronidase by Markovic-Housley et al. [75] are indicated. Red residues (N) denote predicted N-glycosylation sites, including one (+) highly conserved among aligned sequences. doi:10.1371/journal.pntd.0001660.g007 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org also recognized bacterially expressed P. duboscqi apyrase [66], indicating that not all antibodies are specific for possible glycan modifications of sand fly apyrases.

Hyaluronidase
Hyaluronidase is an enzyme that catalyzes the hydrolysis of hyaluronic acid, a major component of the extracellular matrix in vertebrates. It is an ubiquitous enzyme found in mammals, bacteria and in the venom of bees, wasps, spiders, and snakes [71]. Figure 11. Multiple sequence alignment of the PpSP32-like family of salivary proteins. Multiple sequence alignment of sand fly PpSP32like proteins from Phlebotomus arabicus (Ara), P. argentipes (Arg), P. ariasi (Ari), P. duboscqi (Dub), P. papatasi (Pap), P. perniciosus (Per), P. sergenti (Ser), P. tobbi (Tob), and Lutzomyia longipalpis (Lon). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Red residues (N) denote predicted N-glycosylation sites. doi:10.1371/journal.pntd.0001660.g011 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org In bloodsucking Diptera, hyaluronidase activity has been found primarily in the saliva of telmophagic insects: horse flies, black flies, biting midges, and sand flies [72]. Thus, hyaluronidase is believed to decreases host skin tissue viscosity, assisting other salivary components to diffuse and create a pool of blood [60,72,73].
Sand fly hyaluronidase belongs to the same family as mammalian and Hymenopteran hyaluronidases (endo-b-N-acetyl-hexosaminidases, E.C. 3.2.1.35) and is different from that of bloodsucking leeches and nematodes (endo-b-glucuronidases, E.C. 3.2.1.36) [71,74]. Hyaluronidase activity has been detected in all eight sand fly species studied to date ( [23,28,53,73], Figure 6). Our zymographic analyses of P. tobbi ( Figure 6) and P. sergenti originating from Israel ( Figure 6) and Turkey [53] showed the potent activity of sand fly hyaluronidase. Based on the microplate method, P. tobbi hyaluronidase activity is one of the highest measured ( Figure 6A). In contrast, hyaluronidase of P. sergenti had the lowest activity among the species of the genus Phlebotomus ( Figure 6A). Under non-reducing conditions, P. tobbi and P. sergenti hyaluronidase revealed diffuse bands with the molecular weight of around 110 and 135 kDa, respectively ( Figure 6B). Hyaluronidase of P. sergenti is probably a homodimer, because under reducing conditions, the activity was observed at about half of the molecular weight, both in the Israeli ( Figure 6B) and Turkish strains [53], while hyaluronidase of P. tobbi was monomeric with similar molecular weight under non-reducing and reducing conditions and the activity reduced to minimum when denaturated and treated with b-mercaptoethanol ( Figure 6B). Similar features were observed for the hyaluronidase of P. perniciosus, the other Larroussius species ([53], Figure 6), which suggests common biochemical characteristics of this enzyme between closely related species. In general, the remarkably high activity of salivary hyaluronidase may aid the spread of other salivary components as well as transmitted pathogens. Indeed, hyaluronidase coinjected with Le. major promotes infection in BALB/c mice [72]; however, no association was found between hyaluronidase activity and the sand fly capacity to vector either cutaneous or visceral leishmaniasis ( Figure 6A).
Although sand fly hyaluronidase is a very potent enzyme, it is scarcely found in transcriptomic and proteomic approaches probably due to the low abundance of transcripts combined with the large size of the protein. Hyaluronidase transcripts have been reported in only two of seven salivary cDNA libraries, namely in L. longipalpis and P. arabicus [23,24,28]. In P. sergenti, no transcript was found, and in P. tobbi, only one 39-truncated transcript was identified (PtSP125/JN192442). Amino acid residues that constitute the catalytic site (Asp111, Glu113, and Glu247) and form disulfide bridges (Cys22-Cys313 and Cys189-Cys201) in bee hyaluronidase [75] are conserved among the sand fly hyaluronidase sequences (Figure 7). Based on the NetNGlyc prediction server [48], several putative glycosylation sites were predicted in sand fly hyaluronidases, including one highly conserved among aligned sequences (Figure 7).
Allergenic properties of sand fly hyaluronidase are not known, although it has been identified or suspected as the main allergen in the saliva of other bloodsucking Diptera, namely biting midges and horseflies [59,76]. However, there is no record of typical IgEmediated allergic reaction to sand fly saliva; only negligible amount of anti-saliva IgE was measured in hosts repeatedly bitten by sand flies [11,19,77].
D7-related (D7r) proteins are named after the D7 protein, originally described in Aedes aegypti as a major salivary protein exclusively synthesized in bloodsucking females [79,80]. Salivary proteins related to D7 have also been found in black flies and biting midges [56,78] and all sand fly species studied to date. In the P. tobbi SG cDNA library we found seven clusters homologous to D7r sequences (HM164145-HM164151) and three clusters in the P. sergenti cDNA library (PsSP4/HM560863, PsSP5/ HM569360, PsSP7/HM560864) (Tables 1 and 2). Within all sand flies, D7r proteins have similar predicted molecular mass (25.3-28.1 kDa) and wide range of pI (4.82-9.5) (Table S1). Based on the results from NetNGlyc prediction server [48], we found a mixture of putative glycosylated and non-glycosylated D7r sequences in most of the sand fly species studied with the exception of L. longipalpis, P. sergenti, and P. papatasi, where no Nglycosylation sites were found (Figure 8). All sand fly D7r predicted proteins contain nine highly conserved cysteine residues (Figure 8), implying there is a single non-disulphide-bond-forming cysteine.
The other family related to OBPs, PpSP15-like proteins, is closely related to larger D7-related proteins [25,56] and are named after 15-kDa salivary protein of P. papatasi (PpSP15/ AF335487) [6]. They have not been identified in any Diptera other than sand flies [25,56]. It is the most abundant family among sand fly salivary proteins, and P. tobbi and P. sergenti are not exceptions; six and seven members of this family were found in each cDNA library, respectively (Tables 1 and Table 2). Several members were also detected by proteomic analysis, having similar molecular mass as predicted based on the amino acid sequences  (Tables 1 and Table 2; Figure 1). Within the sand flies, PpSP15like proteins have a similar predicted molecular mass (12.2-17.1 kDa) and surprisingly wide range of pI (6.33-9.44) (Table  S1). In accordance with previous reports [25,28], all sand fly PpSP15-like proteins show high degree of variability of around six conserved cystine residues ( Figure S2).
In mosquitoes, some salivary D7 strongly bind biogenic amines and leukotrienes as well as components of the coagulation cascade, thus promptly antagonizing the host defense system [81][82][83] , found in mosquito D7 salivary proteins that bind cysteinyl leukotrienes [83], is also found in the sand fly D7r proteins (Figure 8).
Sand fly PpSP15-like proteins and D7r proteins possess antigenic properties. PpSP15-like proteins were reported as promising anti-Leishmania vaccine candidates [6,27,84]. Phlebotomus papatasi SP15 protein is able to protect mice against Le. major challenge, and a DNA vaccine containing the PpSP15 cDNA provided the same protection [6]. ParSP03 (AAX56359), a PpSP15-like protein from P. ariasi, elicited similar delayed-type hypersensitivity and humoral immune responses upon DNA vaccination [27].
D7r could serve as a marker of exposure to sand fly bites. In humans, all tested serum samples from individuals naturally exposed to P. papatasi strongly bound to a P. papatasi protein band with a molecular mass corresponding to PpSP30 D7r protein (AAL11049) [12,18]. As an ideal marker of exposure, this protein was recognized by both IgE and IgG antibodies, including all tested IgG subclasses [18]. D7r proteins seem to be applicable also for measurement of dog exposure, the main reservoir host for visceral leishmaniasis, since IgG antibodies from animals bitten by P. perniciosus [21] or L. longipalpis [16,85] recognized D7r proteins of the respective species (PpeSP4/DQ150623, PpeSP04B/ DQ150624, PpeSP10/DQ153104, LJL13/AF420274). Moreover, L. longipalpis-bitten dogs bind also to the LJL13 D7r recombinant form [16].
Phylogenetic analysis of D7r proteins showed several major clades (Figure 9). Phlebotomus sergenti sequences clustered together forming a distinct subclade within clade III that contains P. papatasi and P. duboscqi. In contrast, P. tobbi D7r protein sequences are divided among clades I and II, which contain sequences from P. arabicus, P. ariasi, P. argentipes, P. perniciosus, and L. longipalpis. Interestingly, clade II only contained sequences with predicted Nglycosylation sites, which may suggest a unique functional characteristic of D7 molecules within this clade that have arisen after gene duplication. Similarly, phylogenetic analysis of PpSP15like proteins ( Figure 10) revealed several separated groups, consistently clustering P. sergenti sequences with P. duboscqi and P. papatasi, and P. tobbi sequences with those from P. perniciosus and other sand flies studied to date, including a single member from L. longipalpis. PpSP15 could be a multicopy gene, as more than two alleles were found in several P. papatasi individuals, some of them unique to the population origin [86].

PpSP32-like proteins
This family is named PpSP32 from the original identification in P. papatasi (AAL11050) [24] and due to the lack of homology to a conserved protein domain. PpSP32-like proteins have been described solely in sand flies and are found in all species studied so far; we identified homologous sequences also in P. tobbi (PtSP27/HM173642, PtSP28/HM173643, PtSP29/HM173644) and P. sergenti (PsSP44/HM569368). The predicted molecular mass of P. tobbi PpSP32-like proteins (24.5 kDa) is slightly lower than what was measured in proteomic analysis (Figure 1, Tables 1  and 2). All sequences have a wide range of predicted molecular mass (ranging from 22.5 to 34.9 kDa), no protein domain match, Figure 13. Multiple sequence alignment of the yellow-related family of salivary proteins. Multiple sequence alignment of yellow-related proteins from Phlebotomus arabicus (Ara), P. argentipes (Arg), P. ariasi (Ari), P. duboscqi (Dub), P. papatasi (Pap), P. perniciosus (Per), P. sergenti (Ser), P. tobbi (Tob), Lutzomyia longipalpis (Lon), and related sequence from Drosophila simulans (XP_002103634). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Red residues (N) denote predicted N-glycosylation sites. The span of the MRJP protein domain (pfam03022) is marked by arrows. Based on the crystal structure of Lon AAS05318 [9], the cystine residues forming disulfide bonds are indicated by letter C and conserved amino acids contained in the ligand binding pocket by an asterisk (*). doi:10.1371/journal.pntd.0001660.g013 Figure 14. Phylogenetic analysis of the yellow-related family of sand fly salivary proteins. Phylogenetic analysis of yellow-related proteins from Phlebotomus arabicus (Pab), P. argentipes (Pag), P. ariasi (Par), P. duboscqi (Pdu), P. papatasi (Pp), P. perniciosus (Ppe), P. sergenti (Ps), P. tobbi (Pt), Lutzomyia longipalpis (Lulo or LJM), and related sequence from Drosophila simulans (XP_002103634). Phylogenetic analysis was conducted on amino acid sequences without signal peptide using Tree Puzzle (version 5.2 by) maximum likelihood (WAG model), quartet puzzling, and automatically estimated internal branch node support (10,000 replications). Sequence names, accession numbers, and branch node values are indicated. Underlined sequences possess predicted N-glycosylation sites. doi:10.1371/journal.pntd.0001660.g014 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org and are alkalic (pI ranging from 9.3 to 10.6) (Table S1). An interesting common feature of this protein family is that it possesses highly conserved N-and C-terminal regions with extremely variable internal sequence ( Figure 11). Within the genus Phlebotomus there are predicted N-glycosylation sites in the variable and C-terminal regions ( Figure 11).
To date, no function has been associated with sand fly PpSP32like proteins, although L. longipalpis and P. perniciosus proteins have been hypothesized to possess collagen binding activity [24,25] and in P. papatasi, PpSP32 transcripts are expressed independently of either diet or age [87], indicating a vital role for these molecules in feeding. Figure 15. Multiple sequence alignment of the ParSP25-like family of sand fly salivary proteins. Multiple sequence alignment of Phlebotomus tobbi PtSP73 (HM173639), PtSP75 (HM173640), and PtSP76 (HM173641) with related sequences from P. arabicus (Ara), P. ariasi (Ari), and P. perniciosus (Per). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. doi:10.1371/journal.pntd.0001660.g015 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org Phylogenetic analysis of PpSP32-like proteins reflects again the taxonomic relationship within Phlebotomine sand flies [33]. True to form, phylogenetic position of P. tobbi PpSP32-like proteins are within a subclade I with P. perniciosus and the P. sergenti PpSP32-like protein is within the Phlebotomus and Paraphlebotomus clade II ( Figure 12).

Yellow-related proteins
Phlebotomine yellow-related proteins are characterized by the presence of major royal jelly protein domain (MRJP; pfam03022). Originally, MRJP proteins were described from honeybee larval jelly, making up to 90% of the protein content [88]. Sequences related to MRJP proteins were described in Drosophila, where it is related to cuticle pigmentation and, when mutated, it produced a yellow phenotype and thus named Yellow proteins [89,90]. In bloodsucking Diptera, salivary yellow-related proteins have only been described in sand flies [55,56] and black flies [91].
Ribeiro and Arca [55] proposed that in Phlebotomines, salivary yellow-related proteins work as kratagonists, the binders of biogenic amines. Indeed, Xu et al. [9] proved that the bacterially expressed L. longipalpis yellow-related proteins (LJM11, LJM17/ AAD32198, and LJM111/ABB00904) bind biogenic amines, namely serotonin, catecholamines, and histamine. The proteins differed in affinity to the particular ligand, suggesting functional divergence within the family [9]. The midgut yellow protein in Aedes aegypti is involved in the melanization pathway as a dopachrome conversion enzyme [92]; however, in sand flies the yellow-related proteins found in the midgut lumen probably originating from swallowed saliva [93] and researchers failed to Figure 16. Multiple sequence alignment of the 33-kDa salivary protein family. Multiple sequence alignment of Phlebotomus sergenti PsSP49 (HM569369) and P. tobbi PtSP66 (HM173645) proteins with related sequences from P. arabicus (Ara), P. argentipes (Arg), P. ariasi (Ari), P. duboscqi (Dub), P. perniciosus (Per), and Lutzomyia longipalpis (Lon). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Red residues (N) denote predicted N-glycosylation sites. doi:10.1371/journal.pntd.0001660.g016 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org detect dopachrome convertase activity in salivary yellow-related proteins [28,56]. In Glossina morsitans, the ubiquitous tissue expression of the protein suggests also a housekeeping role for yellow-related proteins [91].
Yellow proteins are also under consideration for anti-Leishmania vector-based vaccines. LJM17 from L. longipalpis elicited leishmanicidal Th1 cytokines in immunized dogs [8], and LJM11 protected laboratory animals against both Le. major and Le. infantum [7,9]. In contrast, mice immunized with P. papatasi yellow-related Figure 17. Multiple sequence alignment of the sand fly members of 41.9-kDa salivary protein superfamily. Multiple sequence alignment of Phlebotomus sergenti PsSP82 (HM569371) and P. tobbi PtSP49 (HM173648) proteins with related sequences from P. arabicus (Ara), P. ariasi (Ari), P. duboscqi (Dub), P. perniciosus (Per), and Lutzomyia longipalpis (Lon). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. Red residues (N) denote predicted N-glycosylation sites. doi:10.1371/journal.pntd.0001660.g017 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org proteins PpSP42 or PpSP44 (AAL11052 and AAL11051, respectively) elicited Th2 cytokines and exacerbated Le. major infection [84]. It remains to be elucidated whether the protection induced by yellow-related proteins is related to particular protein immunogenicity, to sand fly species, or to the vector-Leishmaniahost combination, as all of these factors can contribute to vaccine efficacy. Recently, Xu et al. [9] showed that L. longipalpis LJM11 but not LJM111 produces a DTH response in mice challenged by SGH. The authors related this immunogenicity to electrostatic potential on the protein surface, which is positive in LJM11; thus the protein is probably more attractive to antigen-presenting cells [9].
Yellow-related proteins are highly conserved among sand flies. Phylogenetic analysis produced three major clades combining Larroussius, Adlerius and Euphlebotomus (clade I); Phlebotomus and Paraphlebotomus (clade II); and Lutzomyia (clade III), while subclades discerned each subgenus ( Figure 14). Interestingly, P. sergenti illustrates a gene duplication event that preceded speciation and was followed by a clear gene duplication expansion that is seen in one of the subclades. Gene duplication in bloodsucking arthropod salivary molecules is fundamental for the functional diversification of proteins, as can be seen with the range of substrates bound by the L. longipalpis yellow-related proteins [9]. Within clade II, two subclades can be seen distinguishable by the presence of putative N-glycosylation sites. Moreover, sequences in clade IIa have a slightly higher predicted isoelectric point than the glycosylated sequences in clade IIb ( Figure 14, Table S1), indicating another feature that might be responsible for functional diversification.

ParSP25-like proteins
ParSP25-like transcripts were found in P. tobbi but not in P. sergenti SG library. Phlebotomus tobbi ParSP25-like molecules (PtSP73/ HM173639, PtSP75/HM173640, and PtSP76/HM173641) have predicted molecular mass ranging from 27.8 to 38.8 kDa and contain a large proportion of acidic residues resulting in a pI of 4.560.1. The sequences share similarity with eight other sand fly salivary proteins from three sand fly species [25,27,28] (Figure 15), all of them with predicted pI between 4.4 and 5.0 (Table S1). Analysis of the putative protein sequences revealed highly conserved regions rich in amino acid residues such as Asp, Tyr, Glu, and Ser and no predicted N-glycosylation sites ( Figure 15). Though the function is not known, some members of this family were shown to be highly antigenic. Mice immunized with a plasmid coding for ParSP25 (AAX55664) elicited high levels of anti-P. ariasi IgG1 and a strong DTH reaction when challenged with P. ariasi saliva [27]. Moreover, dogs exposed to P. perniciosus bites strongly bind to protein band characterized as PpeSP08 (ABA43056) [21]. Sand fly ParSP25-like proteins are most likely genus-specific because, so far, the sequences have been found only in Adlerius (P. arabicus) and Larroussius species (P. ariasi, P. perniciosus, P. tobbi) and not in representatives of the other subgenera ( Figure 15).

The 33-kDa family
These proteins, named by Anderson et al. [25] as members of the 33-kDa family, have not yet been found in any Diptera other than sand flies. PsSP49 (HM569369) and PtSP66 (HM173645) share sequence similarity with seven other sand fly salivary proteins from six sand fly species both from both New and Old World sand flies [24][25][26][27][28] (Figure 16). All sand fly 33-kDa family proteins have similar predicted molecular weight (32.3-34.5 kDa) and alkalic pI (8.2-9.1) (Table S1). PsSP49 and PtSP66 were both identified in the proteomic analysis ( Figure 1). Two highly conserved N-glycosylation sites were predicted among all sand fly sequences ( Figure 16) and both PsSP49 and PtSP66 were found above the predicted molecular weight in the proteomic analysis ( Figure 1, Tables 1 and 2), indicating a post-translational modification. Indeed, the two proteins from P. arabicus (PabSP32/ACS93510 and PabSP34/ACS93511) showed glycosylation by ProQ Emerald staining [28]. The function is unknown; however, P. perniciosus PpeSP06 (ABA43054) and the L. longipalpis LJL143 (AAS05319) were identified as antigens for dogs living in endemic areas of Le. infantum [16,21], the later one also shown to be a candidate for vaccine against canine leishmaniasis [8]. Figure 18. Multiple sequence alignment of the PsSP28, PtSP8, and PtSP81 salivary proteins. Multiple sequence alignment of Phlebotomus sergenti PsSP28 (HM569370) and P. tobbi PtSP8 (HM173646) and PtSP81 (HM173647) proteins with related sequences from P. ariasi (Ari) and P. perniciosus (Per). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. doi:10.1371/journal.pntd.0001660.g018 Figure 19. Multiple sequence alignment of the Phlebotomus sergenti PsSP98 salivary protein. Multiple sequence alignment of Phlebotomus sergenti PsSP98 protein (HM569366) with related sequences from P. arabicus (Ara) and P. argentipes (Arg). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. doi:10.1371/journal.pntd.0001660.g019 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org 41.9-kDa superfamily 41.9-kDa protein superfamily is specific to bloodsucking Nematocera encompassing members of mosquitoes, biting midges, black flies, and sand flies [56]. The P. sergenti and P. tobbi members of this superfamily, PsSP82 (HM569371) and PtSP49 (HM173648), share sequence similarity with five other sand fly salivary proteins from five sand fly species (Figure 17). These sand fly proteins have a wide range of predicted molecular weight (27.5-56.6 kDa) and pI (4.3-8.5) (Table S1) but only one of them, P. perniciosus PpeSP19 (ABA43063), has been found by proteomic analysis [25]. All sequences are rich in putative N-glycosylation sites ( Figure 17) and the function is not known.

Other putative salivary proteins
Several other putative salivary proteins were identified in the transcriptomes of P. tobbi and P. sergenti SGs. They are smaller than 15 kDa, their function is not known, and are, thus far, unique to sand flies. Additionally, none of these small proteins have been found in the proteomic analysis (Figure 1, Tables 1 and 2).

Antigens and glycoproteins of Phlebotomus tobbi salivary proteins
To identify antigens and glycoproteins in P. tobbi SGH, electrophoretically separated proteins were incubated with anti-P. tobbi rabbit serum and a lectin Concanavalin A (ConA), respectively ( Figure 22). When compared with the proteome analysis in Figure 1, the protein bands visible by silver staining are most likely yellow-related proteins (PtSP37 and PtSP38), apyrases (PtSP4 and PtSP10), antigen 5-related proteins (PtSP77 and PtSP79), PpSP32-like proteins (PtSP28 and PtSP29), D7-related proteins (PtSP58 and PtSP60), and PpSP15-like proteins (PtSP9, PtSP23, and PtSP32). Anti-P. tobbi antibodies recognized all identified bands as well as other six high molecular weight proteins not visible by silver staining (Figure 22, lane 2).
Most of the P. tobbi proteins reacted with ConA, indicating they are N-glycosylated. The lectin binding was specific, as the reactivity was totally inhibited when ConA was preincubated with specific monosaccharide methyl-a-D-mannopyranoside. The most intense reaction was observed with the high molecular weight band not visible by silver staining, and with the bands of molecular weight similar to one yellow-related protein (PtSP38) and both apyrases. Among the nine silver-stained bands, three did not react with ConA, namely bands with molecular weight similar to D7related proteins, PpSP15-like proteins, and one yellow-related protein (PtSP37) (Figure 22, lane 4). The reactivity with ConA is in agreement with N-glycosylation as predicted by NetNGlyc server [48], with the exception of PtSP10 apyrase (Table 3).
We can speculate that the most glycosylated band with the highest molecular weight might be hyaluronidase. Although producing a minor unstainable band, it is predicted to be highly glycosylated ( Figure 7) and its activity is clearly visible around 135 kDa in zymography analyses ( Figure 6).
Within sand fly yellow-related proteins, it is common that glycosylated and non-glycosylated forms occur in the same species. As proved for P. papatasi [93] and predicted for protein sequences of Phlebotomus (P. papatasi and P. duboscqi) and Paraphlebotomus (P. sergenti) species, at least one form is glycosylated, forming a well supported subclade with glycosylated sequences from other species (Figure 14). Glycosylated and non-glycosylated forms are also Figure 20. Multiple sequence alignment of the Phlebotomus sergenti PsSP73 salivary protein. Multiple sequence alignment of Phlebotomus sergenti PsSP73 protein (HM569367) with related sequences from P. arabicus (Ara) and P. ariasi (Ari). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. doi:10.1371/journal.pntd.0001660.g020 Figure 21. Multiple sequence alignment of the Phlebotomus tobbi PtSP71 salivary protein. Multiple sequence alignment of Phlebotomus tobbi PtSP73 protein (HM173639) with related sequences from P. ariasi (Ari) and P. perniciosus (Per). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequence-editing software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. doi:10.1371/journal.pntd.0001660.g021 Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org present in P. tobbi, as proven by blot analysis (Figure 22), although the closely related P. perniciosus possesses only glycosylated forms. Interestingly, in sand fly species within the clades I and III (Figure 14), all published sequences are glycosylated with an exception of P. tobbi and P. ariasi, which has at least one nonglycosylated form. Further research is needed to investigate whether the presence of sugar side chains may contribute to the antigenicity of the yellow-related proteins.

Conclusions
With over 80 species of sand flies implicated in Leishmania transmission, it is vital to continue describing their salivary proteins in the search for vaccine candidates and markers of exposure. In this study, we prepared and analyzed the transcriptome and proteome data of P. tobbi and P. sergenti to broaden our knowledge on the repertoire of Larroussius salivary proteins and provide the first report from a Paraphlebotomus sand fly, respectively.
P. tobbi has been reported to transmit Le. infantum that causes cutaneous leishmaniasis [32]. Interestingly, the salivary proteins of P. tobbi are highly homologous to those of P. perniciosus, a vector of Le. infantum that causes visceral disease. It is likely that, in this instance, the salivary proteins of P. tobbi are not the determining factor for these different disease manifestations. However, in general, it is possible that the divergence, diversity or amount of sand fly salivary proteins or non proteinaceous components of the saliva correlate with different disease manifestations of the same species of Leishmania.
The transcriptome data can be utilized to prepare recombinant proteins that can be used to test their potential as anti-Leishmania vaccines or in epidemiologic studies to develop more specific and efficient methods for measurement of vector exposure. Finally, recombinant salivary proteins may also help us to understand the Figure 22. Phlebotomus tobbi salivary gland antigens and glycoproteins. Salivary gland homogenate of Phlebotomus tobbi was separated on 8% polyacrylamide gel under non-reducing conditions. Separated proteins were silver stained (lane 1) or electrotransferred to nitrocellulose membrane and incubated with (2) serum from rabbit repeatedly exposed to bites of P. tobbi females, (3) non-immune rabbit serum, (4) Concanavalin A lectin, or (5) ConA preincubated with methyla-D-mannopyranoside to control specificity of the ConA reaction. Probable cluster name as compared to mass spectrometry results in Figure 1 and molecular weight (kDa) are indicated on the left and right side, respectively. doi:10.1371/journal.pntd.0001660.g022 Table 3. N-glycosylation sites of Phlebotomus tobbi salivary proteins.  Phlebotomus sergenti and Phlebotomus tobbi Sialome www.plosntds.org mechanism of blood sucking or find biological activities of many of these novel sequences.

Supporting Information
Figure S1 Phlebotomus tobbi and P. sergenti protein families. Analysis of salivary proteins from Phlebotomus tobbi (Pt) and P. sergenti (Ps). Phylogenetic analysis was conducted on amino acid sequences with signal peptide using Tree Puzzle (version 5.2) by maximum likelihood (WAG model), quartet puzzling, and automatically estimated internal branch node support (10,000 replications). Sequence cluster names and branch node values are indicated. Protein families are listed on the right. (TIFF) Figure S2 Multiple sequence alignment of the PpSP15like family of salivary proteins. Multiple sequence alignment of the PpSP15-like proteins from Phlebotomus arabicus (Ara), P. argentipes (Arg), P. ariasi (Ari), P. duboscqi (Dub), P. papatasi (Pap), P. perniciosus (Per), P. sergenti (Ser), P. tobbi (Tob), and Lutzomyia longipalpis (Lon). Sequences without signal peptide were aligned using ClustalX and manually refined using BioEdit sequenceediting software. Accession numbers are indicated in the sequence name. Identical amino acid residues are highlighted black and similar residues grey. (TIFF) Table S1 List of sand fly salivary proteins with their identifiers and selected protein features. Published sand fly salivary proteins from Phlebotomus arabicus (Ara), P. argentipes (Arg), P. ariasi (Ari), P. duboscqi (Dub), P. papatasi (Pap), P. perniciosus (Per), P. sergenti (Ser), P. tobbi (Tob), and Lutzomyia longipalpis (Lon). The proteins are listed with their house name, GenBank accession numbers (Accn) for both nucleotide and protein sequences, protein name, presence in the proteome analysis as confirmed by mass spectrometry or Edman degradation, protein family, predicted signal peptide (SignalP), putative mature protein features (pI, predicted isoelectric point; Mw, predicted molecular weight; AA, number of amino acid residues), and reference. (XLS) Text S1 Accession numbers for genes and proteins mentioned in the text including tables, figure and supplemental files. (DOC)