From a conservative perspective, the sialotranscriptome of C. tarsalis confirms the presence of ubiquitous salivary mosquito protein families, such as the D7, 41 kDa, hyp37, 30 kDa antigen/Aegyptin, 23.4 kDa, mucins, cysteine rich peptide, antigen 5, amylase/maltase, and the immunity related proteins lysozyme, pathogen recognition molecules, and serine proteases possibly associated with immunity or with host matrix/fibrin degradation. Adenosine deaminase has also been found, but not the 5'nucleotidase/apyrase that has been abundantly found in mammalian feeding mosquitoes, but possibly absent or reduced in the bird feeding C. tarsalis. From another stand point, the C. tarsalis sialotranscriptome confirmed the presence of proteins so far known exclusively in culicine mosquitoes, such as endonucleases and the 62 kDa, 30.5 kDa, 7.8 kDa and 4.2 kDa families, all without a known function. Additionally, further confirmation of unique proteins of the Culex genus were found for members of the highly expanded CWRP, 9.7 kDa and the GPQ repeat families.
It has been previously indicated that many unique mosquito protein families appear to be related to bacterial proteins, (when using the Psiblast tool against the NR database) suggesting their acquisition by horizontal transfer. Interestingly, the genes coding for such proteins, now available for An. gambiae, C. quinquefasciatus and Ae. aegypti are often single exonic, as is the case of bacterial genes. In the case of C. tarsalis, this appears to be the case with the 41 kDa (which have a single short intron) and the CWRP (mostly uniexonic in C. quinquefasciatus) families. However, other possibilities for single exonic genes are their acquisition as gene duplication resulting from retrotransposition of an mRNA precursor deriving from an endogenous, multi exonic gene. At any rate, the frequency of the relatively unusual processes of horizontal transfer and/or retrotransposition in the acquisition of new genes associated with blood feeding appear to occur at a high rate in the formation of mosquito sialomes.
Since our transcriptome was based on ~2,000 ESTs from a non-normalized library, the question arises as to what extent novel putative secreted proteins can be discovered with a more extensive sequencing, or the use of a normalized library. As indicated in a recent review [68
], it has been our experience that, perhaps due to the relatively low complexity of the salivary gland proteome (when compared to organs like the mammalian liver or brain, for example), 1000-2000 sequenced clones are enough to display the majority of the sialome. Indeed, ~2,000 clones sequenced from an Ae. aegypti
salivary gland cDNA library [4
] discovered virtually all those found in another effort that obtained ~20,000 sequences [69
]. A similar situation was encountered with the An. gambiae
sialome, where ~4,000 sequences identified basically all those in a large sequencing effort, also ~20,000 sequences done by the Pasteur Institute [3
]. Sequencing of normalized libraries, or more intensive sequencing of existing libraries, perhaps with newer generation of pyrosequencing methodology, may increase the number of salivary secreted peptides, but possibly to no more than 10% of the number found with random sequencing of ~2,000 clones. It should be indicated, however, that these additional low abundance transcripts may account for pharmacologically potent peptides.
Finally, the fast divergence of salivary proteins allows the possibility of using C. tarsalis
proteins as specific markers of vector exposure, as is being attempted for ticks [70
], mosquitoes [26
] and sand flies [73