An insight into the sialome of Simulium guianense (DIPTERA:SIMulIIDAE), the main vector of River Blindness Disease in Brazil

Glycosidases

"Amylases and maltases are ubiquitous enzymes that help digestion of carbohydrates and are commonly found in sialotrancriptomes from Nematocera including mosquitoes, biting midges, sand flies, and black flies [67]. These enzymes can be recognized by the KOG motif 0471, named Alpha-amylase." The proteome of the mosquito An. gambiae has 17 members of this family, one of which (AGAP002102) is expressed in the SGs [64]. The proteome of Ae. aegypti contains 24 such enzymes, at least two of which are expressed in their SGs [25] while Culex quinquefasciatus has 35 such enzymes, with two also expressed in their SGs [68]. Additional File 2 presents two truncated gene products coding for glycosidases (Sg-214 and Sg-296). Glycosidases of S. guianense have 79% sequence identity to other described black fly enzymes (blastp comparisons can be seen in Additional File 2). Phylogenetic analysis of the S. guianense protein sequences together with their closest BLAST matches against the NR database indicates that the two S. guianense proteins group into different clades with strong bootstrap support (Figure 3). Sg-296 groups to other Simulium enzymes, to a salivary sand fly enzyme, and to drosophilids, as indicated by clade I (Figure 3). Sg-214, on the other hand, groups with a second set of Simulium enzymes and, with 76% bootstrap support, to Culicine mosquitoes and salivary biting midge enzymes,[34, 69] as shown by clade II (Figure 3). Notice that the mosquitoes, black flies, and Culicoides sequences each group within subclades having strong bootstrap support, as expected. A third clade of mosquito-only enzymes (including anophelines and culicines) is also obtained, which merges without strong bootstrap support to Clade II. Interestingly, the mosquito enzymes in both clades II and III have all been previously described in salivary transcriptomes, suggesting a common origin of these sugar-hydrolyzing enzymes in the ancestral fly originating mosquitoes, black flies, and biting midges. These results indicate that the two S. guianense sequences appear to be a product of ancient gene duplication, Sg-296 from Clade I being the most ancient, as it groups with enzymes of Brachycera, while the two salivary gene products from Culicine mosquitoes appear to derive from a gene duplication after the split of the Culicidae. The sequences of the glycosidases Sg-214 and Sg-296 found in the sialotranscriptome of S. guianense were confirmed by proteome analysis within the fractions 16 and 17, respectively, just above the 49-kDa standard (Figure 2 and Table 4).

Serine proteases

"Serine proteases are commonly found within hematophagous insect sialomes [70] except in sand flies, where it was only found in the Phlebotomus ariasi sialome [30]. This family has an important role in the immune system, acting as prophenoloxidase activators or in digesting skin matrix components such as in an elastase function, or hydrolyzing host blood-clotting components such as fibrinogen/fibrin, or activating plasminogen [71, 72]. In Ae. aegypti, transcripts coding for serine protease with the CUB domain were reported, indicating specialized substrate recognition [73]." The sialotrancriptome of S. guianense allowed the identification of transcripts coding for three secreted serine proteases varying with predicted mol wt between 25.4 and 27 kDa, which may derive from three polymorphic genes. Alignment of representative members of these three gene products from S. guianense (Sg-416, Sg-138, and Sg-244) with their best BLAST matches (only Diptera sequences were included) produces a phylogram indicative of one clade that is related to mosquito and fruit fly sequences with strong bootstrap support (Figure 4) and one additional Simulium-specific clade, the latter containing two sub clades. This Simulium-specific clade is quite divergent, having only 27% or less identity to their best Diptera match, indicating fast evolution of this clade. The sub-clades (marked as Simulium I, II, and III in Figure 4) each contain one enzyme from each of the Simulidae thus far analyzed for their sialotranscriptome, indicating conservation of these threee salivary expressed genes in black flies. An additional S. nigrimanum sequence is also found in this Simulium-specific clade, indicating that a fourth gene may be expressed in this fly. The serine protease proteins found in the sialotranscriptome of S. guianense were confirmed by proteome analysis within the fractions F24 and F25, located near the 28-kDa marker, consistent with its predicted (25 to 27 kDa) mature weight of these proteins (Figure 2 and Table 4).

Hyaluronidases

Hyaluronidases are enzymes that cleave hyaluronic acid, which is a main component of the extracellular matrix in vertebrates. This enzyme was first described in saliva of New World Lutzomyia longipalpis[74] and thereafter in the SGs of several other Old and New sand fly species [75, 76]. It was also reported in S. vittatum[77]. Hyaluronidases also have been described in the sialotranscriptome of C. quinquesfaciatus[68] and Glossina morsitans morsitans[51]. Interestingly, although Phlebotomus papatasi and Phlebotomus dubosqui SGHs displayed hyaluronidase activity, no such transcripts were found in their cDNA libraries [75]. Hyaluronidase transcripts were also absent from S. vittatum and S. nigrimanum sialotranscripomes [14, 15]. Here, we found one full-length sequence (with five ESTs) coding for a protein with 37.8 mol wt and pI 9.2 matching the pfam01630 domain named "Glyco_hydro_56, Hyaluronidase" with an e value of 1e^-61. The NR database of the NCBI revealed identities above 43% to hyaluronidases from Lu. longipalpis and Phlebotomus arabicus in addition to matching other insect enzymes from Pediculus humanus and some vespids; however, these non-dipteran sequences were only 34% identical at the AA sequence level. Fourteen tryptic peptides obtained by MS/MS had matches to hyaluronidase protein within fraction 20, just above the 38-kDa standard and consistent with the predicted 37-kDa mature mol wt of this protein (Figure 2 and Table 4).

Apyrase

"This enzyme hydrolyzes ATP and ADP to AMP and orthophosphates and has been commonly found in blood feeding arthropods, where it has been suggested as a typical case of convergent evolution [67]. Because ADP and ATP are important activators of platelet and neutrophils, apyrase activity removes these agonists of hemostasis and inflammation [78]. Different genes have been described for this activity such as members of the 5'-nucleotidase family in mosquitoes and triatomines [79–82], the Cimex-type apyrase family in bed bugs and sand flies [83, 84] and the type CD-39 protein family in fleas [52]. Expression of this enzyme in mosquitoes has helped to understand the feeding preference in Anopheles, Aedes, and Culex genus [85]. As Culex has birds as the primary source of blood and does not face the platelet barrier, members of this genus reveal little or absent expression of this enzyme [72]." In black flies, this enzyme activity was previously described in SGHs from several species with different degrees of anthropophy or zoophilic, gonotrophic cycle and vector or non-vector status, revealing dependence on Ca⁺² or Mg⁺² ions for activation and with positive association to species with confirmed vector status for O. volvulus[8, 86]. While we do not know the origin of black fly salivary apyrases, transcripts coding for members of the 5'-nucleotidase family have been previously described in S. vitattum and S. nigrimanum[14, 15]. Studies in black fly sialotrancriptomes also revealed an increase in the expression of putative apyrase transcripts in S. nigrimanum when compared with S. vittatum, with statistically significant difference (p = 0.00337) [15]. The 5'-nucleotidases are ubiquitous enzymes usually found bound to the extracellular face of biologic membranes through a glycophosphatidyl-inositol phosphate anchor [87, 88]. However, salivary secreted enzymes of mosquitoes [80] and triatomine bugs [81] lack the carboxyterminal domain where the glycolypid is anchored, allowing their secretion. Here, we found 14 transcripts coding for the putative salivary apyrase of S. guianense (Additional File 1). Alignment of the putative apyrase of S. guianense with their simulid homologs plus vertebrate sequences known to be membrane anchored reveals the lack of the carboxyterminal site for the glycolipid anchor in S. guianense (Figure 5), as was also found for other Simulium putative apyrases [14, 15], indicating the S. guianense enzyme to be secreted. Fifty-eight tryptic peptides were deducted by MS/MS with matches to apyrase protein (Sg-354) originated from fraction 16, located just below the 62-kDa standard (Figure 2 and Table 4).

Adenosine deaminase (ADA)

ADA transcripts, although previously found in sialotranscriptome of mosquitoes and sand flies,[16, 26, 29, 73, 74, 89, 90] here appear for the first time in black fly sialotranscriptomes. Ae. aegypti salivary homogenates hydrolyze adenosine (Ado) to inosine, and then to hypoxantine plus ribose, with enzymatic activities in saliva and SGHs [91, 92]. Recombinant ADA from P. dubosqi was shown to have potent activity [93]. Here, we found a singleton EST producing one truncated sequence with 68% identity with Ae. aegypti ADA. Puzzlingly, Ado is a powerful antiplatelet and vasodilator, and the presence of a salivary ADA should be considered non-adaptative; however, Ado is also a potent inducer of mast cell degranulation, and for this reason it may be in the insect's interest to remove this product. Interestingly, P. papatasi does not codify transcripts to ADA but contains Ado and AMP in its saliva, which acts as the main salivary vasodilator [94].

Destabilase

"This enzyme is an endo-ε-(γ-Glu)-Lys isopeptidase, which cleaves isopeptide bonds formed by transglutaminase (Factor XIIIa) between Gln glutamine γ-carboxamide and ε-amino groups of lysine and was first described in the saliva of leeches. Its activity leads to dissolution of stabilized fibrin [67]. Destabilases are members of the lysozyme superfamily of proteins [95, 96]." A 3' truncated singleton EST is 86% identical to S. nigrimanum destabilase (Additional File 1). It is still unknown whether salivary homogenates of black flies have destabilase activity.

Serine protease inhibitors (serpins)

Serpins were previously reported in sialomes of Ae. aegypti, Ae. albopictus, Ochlerotatus triseriatus and Lu. longipalpis. The gene coding for the protein gi|3411116 in Ae. aegypti represents the main salivary anticlotting protein in this mosquito with specificity to fXa [97, 98], having as homologue the protein gi|56417456 in Ae. albopictus. Targets of others serpins found in mosquito sialotranscriptomes are unknown [16]. Here, we found one singleton EST (Sg-500) coding for a 3' truncated serpin with 60% identity to a homologous serine protease inhibitor from Ae. aegypti and An. gambiae. Serpins have not been found in black fly sialotranscriptomes, possibly because these flies use Kunitz-domain proteins as anticlotting agents.

Kunitz-domain protease inhibitors

"Kunitz domain-containing proteins are associated with protease inhibitors and so far have been found in sialotranscriptomes of Nematocera black flies and biting midges but not in mosquitoes, sand flies, or bloodsucking Hemiptera. Kunitz domain-containing proteins, however, are abundant in tick sialotranscriptomes." Hematophagous arthropods secrete protease inhibitors that can act in specific points of the coagulation cascade, mainly against thrombin or factor Xa (or both). This activity has been previously described in SGHs of several black fly species such as S. vittatum, S. ochraceum, S. argus, and S. metallicum[5–7, 99] and Culicoides midges [100]. The sialotranscriptome of S. guianense contains a typical single Kunitz protein deducted from three ESTs (Sg-395). This protein has its best blastp match to its homologous S. nigrimanum (72% of identity) and S. vittatum (58%) proteins. Previously, S. vittatum salivary homogenates were shown to have potent fXa inhibitory activity, but its molecular nature remains unknown [5]. Salivary homogenates of S. guianense also inhibited the same target of the coagulation cascade (data not published). It is possible that the fXa inhibitor of Simulium resides in a Kunitz domain-containing protein. Two tryptic peptides obtained by MS/MS matching Kunitz-domain protein (Sg-395) were found within fraction 30, located just below the 6-kDa standard (Figure 2 and Table 4).

Immunity-related products

In this group, we found full coding sequences to two ubiquitous antimicrobial peptides: lysozyme and cepropin. These families are commonly found in hematophagous arthropods, and their presence was previously reported in black fly sialotrancriptomes [14, 15].

The S. guianense sialotranscriptome revealed 25 ESTs coding for members of the lysozyme family (Additional File 1), where several possible alleles of the same gene were identified. S. guianense salivary lysozyme is 79% identical to S. nigrimanum salivary lysozyme and 52% identical to its closest mosquito relative (Additional File 2). Lysozyme activity in S. guianense was confirmed to exist in SGHs (unpublished, Chagas). This activity was described in both male and female mosquitoes [101, 102]. Seven tryptic peptides deducted by MS/MS had matches to lysozyme protein (Sg-263) within fraction 30, located just below the 6-kDa standard (Figure 2 and Table 4).

Cecropins are small secreted basic proteins of 3 kDa mol wt, rich in aliphatic AAs, mainly Val, with highest conservation in its carboxy terminal region. Ten ESTs from the S. guianense sialotranscriptome code for two closely related, possibly allelic, cecropins. These cecropins, as expected, have their best matches to other Simulium and mosquito cecropins. The cecropin peptides (Sg-368 and Sg-369) found in the sialotranscriptome of S. guianense were confirmed by proteome analysis within the fraction 31, located just above the 3-kDa marker, consistent with the predicted 3-kDa mature mol wt of this protein (Figure 2 and Table 4).

Antigen-5 family

" This ubiquitous family belongs to the wider CAP superfamily [103]. Most members have no known function, but a few have been related to pathogen defenses in plants, as toxins in snake and lizard venoms [104, 105], as a platelet aggregation inhibitor in a tabanid fly [106], and as a possible inhibitor of the classical pathway of complement activation in the stable fly [107, 108]." Members of this family are found in all nematoceram sialotranscriptomes [67]. The annotated An. gambiae proteome reveals 21 proteins for this family, 2 of which are expressed in the SGs [64]. Similarly, the Ae. aegypti and C. quinquefasciatus proteome have over 30 members of the family, of which at least 2 are expressed in their SGs [25, 68]. The S. guianense sialotranscriptome reveals two clusters coding for CAP family members. The phylogram resulting of the alignment of the two S. guianense proteins with their 25 best blastp matches from the NR protein database (excluding drosophilid proteins) reveals one clade of bloodsucking Nematocera with relatively strong bootstrap support (70%), and strong bootstrap support for the sub-clades containing each of the four families. Interestingly, all members of this clade (marked as "Salivary" in Figure 6) were found in sialotrancriptomes, suggesting a common salivary ancestor for this particular CAP-coding gene within Nematocera. The S. guianense protein Sg-453 is outside this clade and may represent an additional Simulium gene member of the CAP family that has been recruited for a salivary function. Thirty tryptic peptides deducted by MS/MS had matches for an Antigen-5 protein (Sg-457) within fraction 24, located just above the 28-kDa marked, near the predicted 30-kDa mature mol wt of this protein (Figure 2 and Table 4).

Yellow family

This family is insect-specific in eukaryotes and received this name due to mutation of a gene that induces a yellow phenotype in Drosophila, resulting from the disruption of melanin formation. In Nematocera, this family is abundantly expressed in sand flies and has been suggested as important markers of vector exposure [109]. S. guianense reveals only one EST to Yellow protein with a match to Yellow of C. quinquesfaciatus (64% identity). Interestly, S. vittatum also reveals one transcript to the Yellow family [14]. The function of this protein in Simulium remains unknown.

ML domain family

S. guianense contains one transcript coding for a protein containing the ML domain, which was not previously reported in black fly sialomes. This domain is implicated in lipid recognition, particularly in the recognition of pathogen-related products, but could also have a lysosomal function [110] and thus have a housekeeping function. It has an immunoglobulin-like β-sandwich fold similar to that of E-set. The blast to NR database suggests similarity to Niemann-Pick Type C-2 putative from An. aegypti (also known as Epipidymal secretory protein E1) and similarities to MPA2 allergen from Nasonia vitripennis (Hymenoptera). The function of this protein in S. guianense remains unknown.

Lipocalins

Lipocalins are proteins widely distributed in animals and plants. This protein family is highly expressed in triatomines, such as Rhodnius prolixus, Triatoma infestans[47], Triatoma brasiliensis[111] and Dipetalogaster maxima[112]. In triatomines, lipocalins are reported as carriers of nitric oxide, kratagonists (binders of agonists) of biogenic amines, Ado nucleotides, and thromboxane A₂, as well as inhibitors of collagen-induced platelet aggregation and thrombin and as allergens [70, 113–116]. We found two ESTs coding for different lipocalins (Sg-671 and Sg-568). The deducted AA sequence of both transcripts matches human apolipoprotein in the Swissprot database. If secreted in saliva, these proteins are candidates for agonist lipid binding.

Aegyptin family

This protein family, commonly found in sialotranscriptomes of mosquitoes, was first named as 30-kDa Aedes allergen [117] and as GE-rich protein [89]. It has revealed high levels of expression in the sialotranscriptome of Anopheles funestus[18]. Functional analyses of an Ae. aegypti family member, named Aegyptin, as well as a member from An. stephensi, have demonstrated its function as an antagonist of collagen-induced platelet aggregation and as a useful tool for inhibiting platelet-collagen interaction in vitro and in vivo[18, 72, 118]. Previous black fly sialotranscriptomes have revealed proteins similar to Aegyptin, supporting the common origin of hematophagy in mosquitoes and black flies as proposed by Grimaldi and Engel [53]. Here, we found 23 ESTs of the S class (Additional File 1) coding for what appear to be alleles of a single gene similar to Aegyptin-like proteins, showing 60% identity to Aegyptins from sialotranscriptomes of black flies and 35% to mosquito homologs. The alignment revealed GE-rich regions mainly in the middle of the sequences (not shown). Thirty-seven tryptic peptides obtained by MS/MS had matches to Aegyptin protein (Sg-276) within fraction 24, located just above the 28-kDa marker, consistent with predicted (28 kDa) mature weight of this protein (Figure 2 and Table 4).

Diptera Secreted Protein from Conserved Insect Family

Five proteins found in sialotranscriptome of S. guianense were similar to secreted protein from insects. This family was previously described in S. nigrimanum sialotranscriptome. They vary between 37 and 57 kDa mol wt with pI 6.1 to 9.6. The best matches to the NR database showed similarities to several families of Diptera (Culicidae, Ceratopogonidae, Drosophilidae) and Hymenoptera (Pteromalidae and Formicidae) maintaining a low degree of conserved AA. The phylogram suggests at least four different genes to exist in S. guianense. Pattern-initiated PSI-BLAST (PHI-BLAST) using the initiation pattern G-x-[MLI]-x(6)-[WF]-x(7,12)-[KNE]-x-[IMFL]-x(37,40)-[VIL]-x-[YF]-x(3)-[QKR]-x(14)-[IL]-x(5,6)-[NDE]-x(5)-[ILV]-[AS] shows the diversity of this protein family within insects (Additional File 3).

Laminin-like Secreted Salivary Protein Found in Simulium and Culicoides

Laminin-like proteins have been suggested to be extracellular matrix proteins [34]. The S. guianense sialotranscriptome revealed one truncated protein (Sg-431) with four ESTs coding to a S. vittatum homolog (76% identity) and to a Culicoides nubeculosus protein with 32% identity. Two tryptic peptides obtained by MS/MS had matches for laminin-like protein within fraction 14, just above the 62-kDa standard (Figure 2 and Table 4).

D7/OBP superfamily

"The odorant-binding protein family is ubiquitous in insects. The D7 protein family, specific to bloodsucking Nematocera, is recognized as a member of the OBP superfamily [119, 120] but it contains two additional α-helices [121, 122]. Short and long forms of the D7 family exist in which one or two D7 domains exist in the same protein, producing proteins with mature mol wt of ~18 or ~28 kDa. In Simulium, an extra-short family with ~12 kDa is also found, reminiscent of sand fly salivary proteins, which also have an ultra-short form but bear no similarities to the black fly proteins at the AA level [67]. "The S. guianense OBP/D7 sequences were grouped in the three subfamilies described below.

Long D7 family: Two proteins with two OBP domains are recognized in the S. guianense sialotranscriptome. When searched against the NR database using blastp, these proteins only produce significant matches to other Simulium proteins, but all three produce two matches each to the PFAM PBP_GOBP domain when using the tool rpsblast, one in the first half and the other in the second half of the protein. A third truncated protein has only one OBP domain but matches long D7 proteins of S. nigrimanum. Long D7 proteins had tryptic peptides deducted by MS/MS within fractions F23 and F26 near the 28-kDa standard, consistent with the predicted (28 kDa and 24 kDa) mature weight of these proteins (Figure 2 and Table 4).

D7 16-kDa family: Two S. guianense proteins containing one OBP domain were found. Sg-331 produces significant similarities only to other Simulium proteins, but Sg-350 additionally retrieves OBP from C. quinquefasciatus. Tryptic peptides were deducted by MS/MS matches to D7 16-kDa proteins within the fractions 29 and 30, just above the 6-kDa marker (Figure 2 and Table 4).

Ultra-short D7 proteins (10-12 kDa mature weight): This was the most expressed family within the D7/OBP superfamily, encompassing 80 ESTs (Additional File 1). All clusters contain signal peptide, suggesting secreted proteins. These data suggest the existence of at least four genes coding for ultra-short D7 proteins (Sg-75, Sg-190, Sg-363, and Sg-383) and several possible alleles. Only Sg-383 was deducted by MS/MS within fraction 31, just below the 6-kDa marker (Figure 2 and Table 4).

SVEP vasodilator family

" This family is specific to black flies and was originally described in SGHs of S. vittatum, when it was named Simulium vasodilator erythema protein (SVEP) because it produced a prolonged vasodilation when tested in rabbit skin [10]. A recombinant protein (rSVEP) was expressed and functionally characterized as a potent vasodilator, possibly activating ATP-dependent K+ channels [123]. This property has an important role during blood feeding and was suggested as one key compound of the competence vector of these flies in the transmission of Onchocerca parasites [124]." Sialotranscriptomes of two black fly species identified SVEP to belong to a diverse multigene family with at least five genes for each species [14, 15]. The sialotranscriptome of S. guianense also revealed proteins homologous to SVEP, totaling 190 ESTs with identities to other Simulium SVEPs varying from 50 to 70%. Alignment of members of this family showed sequences with similar sizes but with few conserved AAs. Comparative phylogenetic analysis of all SVEP proteins, after 10 000 bootstraps grouped the majority of the members of S. vittatum in a specific clade (I) with 80% bootstrap support (Figure 7). The phylogram indicates at least three genes and several either recent gene duplications and/or alleles coding to members of this family (clade IV on Figure 7). The protein Sg-13 shares clade II, with 92% of bootstrap support, with its homologous S. nigrimanum proteins. ESTs coding for Sg-13 or very closely related proteins represent more than 50% of the sequences coding for SVEP members in this sialotranscriptome. Clade III groups only SVEPs from S. vittatum and S. nigrimanum without bootstrap support. Clade IV reveals a possible case of gene duplication or expression (or both) of a very polymorphic gene from S. guianense, and the last clade (V) groups two clusters (Sg-344 and Sg-343) of S. guianense with its homologous S. nigrimanum (with 100% bootstrap support), which appear completely distinct from other SVEP proteins. This scenario indicates that at least two genes (Sg-13 and Sg-344) have common ancestors with S. nigrimanum and a third gene could have given rise to the increased of expression of this protein family, shown in clade IV, possibly with many recent gene duplications. Interestingly, S. guianense has 190 ESTs coding for SVEP (14.4% of the transcripts of the S class), more than double those of S. nigrimanum (6.8% of its ESTs from the S class) and more than three times those of S. vittatum (which codes 4.5% of its ESTs from the S class to SVEP). These increases are highly significant, having scores of χ² = 16.72 (p = 2.3E^-8) and χ² = 72.5 (p = 1.6E^-17) to S. nigrimanum and S. vittatum, respectively. All sequences of SVEP proteins found in the sialotranscriptome of S. guianense were confirmed by proteome analysis within fraction 30, located just above the 14-kDa marker, consistent with the predicted (14 kDa) mature weight of SVEP protein (Figure 2 and Table 4).

H-rich, acid proteins of Simulium

This protein family is known by its repeats of histidine, proline, glutamine, and glutamic acid residues. The repeat nature of these proteins had been suggested to interact with matrix proteins--possibly collagen--and function in a manner analogous to mosquito Aegyptins, which inhibit collagen-induced platelet aggregation [88]. It is also possible that the His repeats act as antimicrobials by chelating Zn or other trace element ions [125–127]. The black fly S. vittatum revealed this family to be the most abundant protein family expressed in its sialotranscriptome, with four repeat regions in its sequences (arginine-rich, HG repeat, HPH repeat, and QPE repeat) [14]. Similarly, mosquito and Culicoides sialotrancriptomes also contain proteins with Pro-His and Gly-His repeats, but no other sequence similarities. The S. guianense sialotranscriptome has 9.6% of all its secretory ESTs coding for members of this family, having above 70% identity to their homologous S. nigrimanum proteins. Alignment (Figure 8) shows that the S. guianense sequences, together with their homologous S. nigrimanum proteins, contain one repeat region coding for Pro-Lys-Pro residues, whereas in S. vittatum, the Lys residue is substituted by Gln. The phylogram of this protein family (Figure 8), when added to mosquito and Culicoides sequences, reveals that all Simulium sequences indicate (as expected) a common ancestor with 93% bootstrap support, with S. guianense sharing the same branch with S. nigrimanum (also as expected).

Mucins

" Mucins are low-complexity proteins rich in serine and threonine residues. They are frequently found in sialotranscriptomes of bloodsucking arthropods such as mosquitoes [72, 73, 128], biting midges [69], bed bugs [46] and black flies [14, 15]. While these proteins' biologic function is not completely known, they have been suggested to provide protection to internal parts of the salivary ducts and also to have antimicrobial functions. They are commonly expressed in moist epithelia, where they offer protection [129, 130] In addition, mucins are modified post-translationally, and their mature forms have N-acetyl-galactosamine residues [130]." Two types of Simulium-specific mucins are found in the sialotranscriptome of S. guianense, as follows.

Simulium mucin family

Nine proteins (with 40 ESTs) in the S. guianense sialotranscriptome code for Simulium mucin. Their coding sequences have high amounts of Ser + Thr residues, varying from 34.6 to 42.6%, and from 40-144 galactosylation sites are predicted by the NetOglyc server [131]. Although similar sequences were found in the sialotranscriptome of S. nigrimanum, members of this family were absent in S. vittatum. Alignment of the S. guianense and S. nigrimanum sequences reveals extensive similarities and identities along the whole sequence, but the phylogram clearly distinguishes S. guianense and S. nigrimanum specific clades (Figure 9). The variation among the S. guianense sequences may result from splice variants, polymerase slippage on nucleotide repeats, or multiple genes. Several tryptic peptides were deduced by MS/MS with matches for Simulium mucin within fraction 16 (just below the 62-kDa marker) and to fraction 9 (just above the 96-kDa marker) (Figure 2 and Table 4).

Acid mucins proteins similar to Basic 7-13 Simulium family

The sialotranscriptome of S. nigrimanum reported small basic proteins (pI 8.1-10.6) with mature weight varying from 7 to 13 kDa and above 65% identity to orphan proteins found in the sialotranscriptome of S. vittatum[15]. The sialotranscriptome of S. guianense revealed 13 proteins (with 39 ESTs) with best matches to members of this family--from both S. vittatum and S. nigrimanum--at a 50% identity level; however, these sequences code for acid, not basic, proteins (pI 4.1-4.4) (Figure 10), with 0 to 19 potential galactosylation sites. Notably, the S. guianense proteins have an extended central domain containing Gly-Ser repeats that vary in size among the proteins, which may reflect polymerase slippage among closely related genes (Figure 10). The phylogenetic tree cluster produces monospecific branches indicative either of single polymorphic genes or, alternatively, of multiple genes that possibly interact, producing gene conversions (Figure 10).

Simulium collagen-like family

Previous sialotranscriptomes of black flies reported specific proteins named as Simulium collagen-like that are rich in Pro-Gly residues [14, 15]. Homologs to this family were found in the sialotranscriptome of S. guianense, with 17 clusters containing 48 ESTs (Table 2). These 17 clusters are variations of only three sequences, which were aligned with their Simulium homologs (having ~ 40% identity) to produce the phylogenetic tree shown in Figure 11. Alignment revealed relatively conserved AAs in the length of the sequence, with some gaps due to the longer sequences from S. vittatum. The phylogram maintains monospecific clades, as seen before for the mucin families (Figure 11). Several tryptic peptides were deducted by MS/MS within fraction 23, just above the 28-kDa standard, above the predicted (22 kDa) mature weight of these proteins (Figure 2 and Table 4).

Sv 7.8 kDa family

Members of this family were first found in the sialotranscriptome of S. vittatum coding for proteins with 7.8 kDa mol wt [14]. Later, sialotranscriptome of S. nigrimanum added six more transcripts to this family, suggesting it to be a divergent multifamily gene from Simulium[15]. Sequences from S. guianense maintain 60% identity to its homologous S. nigrimanum protein and 50% to the S. vittatum protein, coding for basic proteins (pI 9-10.4) with mature weight varying from 7.1 to 13.4 kDa. Alignment revealed low levels of conserved AAs and at least four genes to S. guianense proteins of this family, marked as clades I-IV on Figure 12. Several tryptic peptides were deducted by MS/MS in the fractions F27, F30, F32, and F31. These fractions are located in the gel just above the 14-kDa marker and just above the 3-kDa marker. These results are consistent with the predicted (7 to 13 kDa) mature weight of these proteins (Figure 2 and Table 4).

Basic 7-13 Simulium family

The S. guianense sialotranscriptome added two more proteins (Sg-420 and Sg-403) with six ESTs to this family coding to basic proteins (pI 11) and 8 kDa, with more than 59% similarities to their homologous S. nigrimanum and S. vittatum proteins. Tryptic peptides were found by MS/MS within fractions 30 and 31, just below the 6-kDa standard (Figure 2 and Table 4). Their function remains unknown.

Simulium 4.8-kDa family

Five more transcripts were added to this family, which appears highly conserved in sialotranscriptomes of Simulium. Their sequences code to acidic proteins (pI 4.4) with 5 kDa of mature weight and are devoid of cysteines. These peptides have unknown function.

Simulium Basic 7.4-kDa family

The cluster Sg-422 (with four ESTs) codes to a basic peptide of 7 kDa mol wt and above 50% identity to their homologous proteins from S. vittatum and S. nigrimanum. This protein family also does not contain any Cys residues on the mature peptide. Four tryptic peptides originated from Sg-422 were deduced by MS/MS within fraction 31, just below the 6-kDa standard, consistent with the predicted (6.8 kDa) mature weight of this protein (Figure 2 and Table 4).

Simulium Basic 13-kDa

Sg-446 added three more ESTs to this family, and has ~ 40% identity to other Simulium proteins. They do not match other known proteins in any of the NR, GO, KOG, CDD, PFAM, or SMART databases. Two tryptic peptides were deduced by MS/MS within fraction 29, just below the 14-kDa standard and consistent with the predicted (13.7 kDa) mature weight of this protein (Figure 2 and Table 4).

5-Cys Simulium family

This family received this name because it contains five Cys in their sequences. One protein with nine ESTs is here reported in the S. guianense sialotranscriptome coding to an acid protein (pI 5.7) with 14-kDa mol wt and above 60% of identity to homologs found in Simulium sialotranscriptomes. Three tryptic peptides were deduced by MS/MS within fraction F30, just below the 6-kDa standard (Figure 2 and Table 4).

Families deorphanized from S. nigrimanum. Six additional protein families were characterized in common between S. nigrimanum and S. guianense, and no other known protein. They do not produce significant matches to others proteins in the NR database, and have thus deorphanized these S. nigrimanum proteins.

Deorphanized S. nigrimanum 8-10 Cys W family. This family is so named because their members contain from 8 to 10 conserved Cys and Trp in their mature sequences. The sialotranscriptome of S. nigrimanum revealed two distinct subfamilies, one containing 10 Cys and 5 conserved Trp and other containing 8 Cys and 6 Trp. This last group was suggested as a candidate protein in the etiology of pemphigus foliaceus due its similarity to proteins annotated as junctional adhesion molecules [15]. The S. guianense sialotranscriptome added two more proteins to this family (11 ESTs), which contain 9 Cys and 5 or 6 Trp. These proteins were confirmed by MS/MS within fraction 29, just below the 14-kDa standard, near their predicted (16.9 kDa) mature weights (Figure 2 and Table 4). The function of this protein family remains unknown.

The sialotrancriptome of S. guianense added three more proteins (Sg-319, Sg-320, and Sg-321) with 12 ESTs to the Acid 28-kDa family. PAGE MS/MS results reported many tryptic peptides for these proteins within fraction F23, just above the 28-kDa marker (Figure 2 and Table 4), in accordance with the predicted (22 to 27 kDa) mature mol wt of these proteins. One protein (Sg-136) with nine ESTs was added to Simulium Basic 28-kDa family. This protein (Sg-136) had several tryptic peptides deducted by MS/MS within fraction 24, consistent with a mass near 28 kDa (Figure 2 and Table 4). The protein family named as 19-kDa family, first seen in S. nigrimanum, was deorphanized with two proteins (Sg-303 and Sg-309) with 10 ESTs coding for basic proteins of 16.8 MW and signal peptide in their sequences. Tryptic peptides were found by MS/MS within the fraction 27, located just above the 14-kDa standard and consistent with the predicted (16.8 kDa) mature weight of this protein (Figure 2 and Table 4). Other putative secreted peptides were also deducted from the S. guianense sialotranscriptome, such as the cluster Sg-258 (22 ETS) coding to basic protein of 8-kDa mol wt that has 70% identity to the orphan protein of the S. nigrimanum sialotranscriptome previously named 8-kDa basic protein family. Five tryptic peptides were deducted by MS/MS within fraction 31, just below the 6-kDa marker (Figure 2 and Table 4). The smaller peptide found in this cDNA library also represents one case of deorphanization with two ESTs in cluster Sg-438 matching members of the Sn basic 4.4-kDa family.

Novel peptide similar to kunitoxin

The S. guianense has two clusters (Sg-375 and Sg-409) coding to novel peptide distantly similar (32% identity) to the snake peptide kunitoxin [132]. They are Cys-rich and were suggested as protease and serine protease inhibitors in snake venom glands [133]. Although the snake peptides have a typical Kunitz domain, this domain is not identified in the black fly protein. Kunitoxin inhibits plasmin and thrombin, blocks L-type calcium channels, and forms part of the neurotoxic complexes with PLA₂ molecules [133]. No similar sequences have been found so far in any previously described sialotranscriptome. Together, the black fly family grouped eight ESTs coding for this secreted basic peptide with 8-9 kDa. The PAGE/MS/MS run reported four tryptic peptides for the Kunitoxin-like protein at fraction 31, coincident with a well-stained band between 3 and 6-kDa standards (Figure 2 and Table 4).