In this study we have identified 13 mesonivirus isolates and characterized their genome organization and phylogenetic relationships. Based on species demarcation criteria employed previously for mesoniviruses , five of these new isolates would be assigned to the same species as NDiV and CavV (Alphamesonivirus-1), other previously described mesoniviruses - HanaV MenoV, and NseV - would be assigned to three new species (Alphamesonivirus-2, Alphamesonivirus-3 and Alphamesonivirus-4, respectively), and eight of the new isolates would represent three new species (Alphamesonivirus-5, Alphamesonivirus-6 and Alphamesonivirus-7) . However, we consider this basis for species demarcation, which employs only a genetic standard of pairwise sequence divergence to assign viruses to a species, should be re-evaluated following further assessment of the ecology of these viruses and their potential for genetic recombination which may provide a more informed analysis of suitable species demarcation criteria.
The isolates of viruses assigned to the species Alphamesonivirus-1 illustrate the wide geographic distribution and mosquito host range of some mesoniviruses. The original four isolates of NDiV were made in northern and central Vietnam from Culex vishnui and Cx. tritaeniorhynchus mosquitoes collected indoors during a surveillance program for Japanese encephalitis virus (JEV) . Our NDiV isolate from Java, Indonesia (Ngewontan strain) was also obtained from a pool of Cx. vishnui mosquitoes in 1981. The four NDiV isolates from Houston, Texas (Houston strain) were made from Cx. quinquefasciatus and Aedes albopictus collected outdoors within the Houston metropolitan area during West Nile virus (WNV) surveillance in 2004 and 2010, respectively. Interestingly, all of these isolates were from mosquitoes captured in or near human dwellings and from species that feed on humans. The close similarity of isolates from Houston and Vietnam may suggest a recent translocation, possibly during the Vietnam conflict when Houston hosted a major air base for embarkation/disembarkation. Isolates of Kampaeng Phet, Botang Baru and Karang Sari viruses, which may be considered three new mesonivirus species, have a geographic distribution extending at least from central Thailand to Kalimantan and Java in Indonesia, and have been isolated from at least two species of Culex mosquitoes.
Another consideration is the potential effect of mesonivirus infection on the susceptibility and vector competence of mosquitoes for viral pathogens of vertebrates. For example, both Cx. vishnui and Cx. tritaeniorhynchus are important vectors of JEV in Asia, and Cx. quinquefasciatus is the major vector of WNV in Houston. Recent experimental studies with Ae. aegypti mosquitoes infected with certain strains of the Wolbachia indicate that the presence of the symbiont bacterium interferes with dengue virus replication and decreases vector competence, possibly by upregulating or priming the mosquito’s innate immune system [11, 12]. Similar results have been reported for Wolbachia-infected Ae. aegypti and chikungunya virus  and with Wolbachia-infected Cx. quinquefaciatus and WNV . If a bacterial endosymboint can alter a mosquito’s vector competence for arboviruses, it seems plausible that a viral symbiont could have a similar effect . This is an important area for future investigation.
Due to the continuous efforts of the virus discovery program of World Reference Center for Emerging Viruses and Arboviruses (WRCEVA), we have continued to isolate mesoniviruses from various insect vectors collected from widespread geographic locations (e.g., Nepal, Colombia and South Florida) suggesting that these viruses are more common than previously thought. Zirkel et al.  suggested that these viruses may have their origins in pristine rainforests and emergence may have been facilitated through anthropogenic-induced modifications (e.g., altered land use, deforestation).
The detailed analysis and comparison of mesonivirus genome architecture conducted here has revealed some unexpected characteristics. Firstly, the presence of block insertions of up to 588 nt in the 5’ terminal quadrant of ORF1a of several mesoniviruses has not been reported previously. The function of this region is presently unknown in mesoniviruses and other nidoviruses and so the structural and functional consequences of these insertions, which contain various imperfect repeats is unclear. Although sharing similar genome architecture, nidoviruses vary greatly in genome size. A previous analysis of the evolution of nidovirus genomes concluded that genome expansion has occurred in a wave-like fashion in which the three major coding regions (ORF1b, ORF1a and the 3’ORFs) expanded consecutively in a hierarchy that reflects the roles of their encoded proteins in the virus replication cycle . This implies that nidoviruses have an inherent capacity for genome expansion, most likely associated with the transitional retention of sequences that serve as a resource for the evolution of new functions. The block insertions detected in ORF1a appear to be functionally redundant and their potential role in such evolutionary processes is presently unclear.
The comparative analysis also revealed that the stem-loop structure which had previously been identified at the RFS site of NDiV is not conserved in the other mesoniviruses and so may not be responsible for activating the -1 ribosomal frame shift. Secondary structure predictions using IPknot on the aligned sequences downstream of the conserved ‘slippery’ sequence site revealed a conserved pseudoknot structure that conformed to all nucleotide substitutions. However, the predicted structure featured only four relatively short regions of complementarity and no estimations of minimum free energy for the represented structures are available through this algorithm. A possible ‘slippery’ sequence (CACUUUU) was also detected in the ORF3a/ORF3b overlap region but no conserved stem-loop or pseudoknot structure was predicted by IPknot in the downstream sequence. It is, therefore, unclear whether ORF3b is expressed by internal initiation or as a read-through extension of ORF3a, which would generate a double-membrane spanning protein. Functional analysis of each of the regions corresponding to the RFS in ORF1a/ORF1b and the putative RFS in ORF3a/ORF3b would help resolve the mechanisms of mesonivirus gene expression.
In conclusion, we have identified and characterized several new mesoniviruses from mosquitoes of human medical importance sampled over time from widespread geographic regions. Several important questions related to their transmission, maintenance in insect hosts in nature, their potential impact of infection on the insect’s behavior, fertility, fecundity and survival, their evolution, their mechanisms of gene expression and their potential to be developed as biological control agents, warrant further investigation.