In this study we demonstrated that sequential infection of C6/36 cells or Cx. quinquefasiatus mosquitoes with CxFV Izabal and WNV did not interfere with either growth or transmission of WNV. This finding is not surprising given that Culex flaviviruses are being discovered in mosquito populations around the world in locations where WNV and other flaviviruses circulate sympatrically. Therefore, the prevalence of CxFV Izabal in Cx. quinquefasciatus in Guatemala does not explain the lack of human disease attributable to WNV in this region.
Growth of WNV in C6/36 cells was not inhibited by prior infection of CxFV Izabal. The WNV titer in CxFV Izabal (+) C6/36 cells did not reach the maximum titer observed in CxFV Izabal (−) cells due to death of cells caused by CxFV Izabal (). As suggested by Hoshino et al. 
, CPE observed in C6/36 cells may be the result of an unnatural association between this Culex
-derived virus and Aedes
-derived cell line since CxFV apparently replicates avirulently in its mosquito host. Therefore, future studies should include utilization of Culex
-derived cell lines. The natural host range of CxFV Izabal across mosquito species and genera is not known.
Data regarding the establishment of superinfection by homologous viruses in cell culture have been variable. C6/36 cells persistently infected with Aedes aegypti
densonucleosis virus remained permissive to infection with Haemagogus equinus
densovirus (HeDNV), arguing against the induction of an anti-viral or immune state in the cells that would otherwise inhibit superinfection by this a similar virus 
. However, interference between superinfecting alphaviruses in mosquito cell culture has been documented multiple times 
. The cellular and molecular mechanisms that support replication of WNV in CxFV (+) cells are not known and require further study.
Overall, neither growth nor transmission of WNV in Cx. quinquefasciatus
was significantly affected by CxFV Izabal when viruses were administed sequentially. These findings are in contrast to what has been found previously for flavivirus - flavivirus superinfections involving WNV in Culex
mosquitoes, however insect-only flaviviruses are fairly divergent from other vector-borne flavivirues such as WNV 
. Previous studies have demonstrated that transmission of a superinfecting flavivirus was blocked if the secondary flavivirus was antigenically-similar to the primary infecting flavivirus 
. Interference to arbovirus superinfection in mosquitoes or mosquito cells by homologous viruses could be the result of RNA interference (RNAi). RNAi is a mechanism by which invertebrates respond to viral infection through the specific recognition and degradation of viral mRNA sequences by virus-derived small interfering RNAs (viRNA) 
. However, our data demonstrate that prior infection with CxFV Izabal does not interfere with WNV replication when the viruses are inoculated simultaneously, or when mosquitoes are exposed to WNV one week following inoculation with CxFV Izabal. If an RNAi pathway was induced in Cx. quinquefasciatus
by CxFV Izabal it would most likely still be effective seven days post-inoculation when mosquitoes were exposed to WNV as it has been previously reported that viRNAs targeting Sindbis virus in C6/36 cells were first detected 48h following infection and were still abundant 7 DPI 
, and viRNAs targeting WNV in Cx. quinquefasciatus
midguts were detected 7 and 14 days post exposure to WNV 
There are numerous potential reasons why mosquitoes would be permissive to co-infecting flaviviruses. First, if CxFV Izabal induced an RNAi response in Cx. quinquefasciatus
, the viRNAs generated might not be sufficiently homologous to WNV to interfere with the establishment of WNV infection. RNAi is a highly sequence-specific mechanism with little tolerance to mismatches between the viRNA trigger and mRNA target sequences 
, and the nucleotide sequence identity between CxFV and other vertebrate-infecting flaviviruses is relatively low. Kim et al. 
reported between 25 and 52% nucleotide sequence homology between CxFV (TX24518) and WNV among structural and non-structural genes. Similarly, Hoshino et al. 
reported 17–25% sequence identity for structural proteins and 17–40% identity among non-structural proteins between their Japanese isolate of CxFV and other flaviviruses. Therefore, any interference of CxFV Izabal viRNAs with WNV was probably minimal. Secondly, it is possible that over a history of co-evolution between insect-only flaviviruses and their mosquito hosts, these viruses have evolved a way to either evade or suppress an immune mechanism that would otherwise interfere with their own replication, or replication of a subsequently-infecting virus. Flock house virus encodes an RNAi suppressor protein, B2, that is necessary for establishment of viral infection in Drosophila S2 cells 
. Virus-encoded suppressors of RNAi have also been found in plant viruses such as tobacco etch potyvirus 
. The molecular mechanisms that permit co-existence of both WNV and CxFV Izabal, potentially even within the same tissues and cells, requires further study. Thirdly, CxFV replication in mosquito cells is presumably similar to that of other flaviviruses due to similar genome organization 
. Zebovitz and Brown 
determined that interference of superinfecting alphaviruses in cell culture was due to competition for replication sites or metabolites, and that viral RNA synthesis was necessary for inhibition of alphavirus superinfection. The non-structural proteins encoded by flaviviruses play important roles in virus replication and maturation 
, and the 5′ and 3′ untranslated regions contain conserved nucleotide sequences and RNA secondary structures involved in virus replication and translation 
. Camissa-Parks et al. 
discovered that the 3′ stem loop structure of CFAV differed from that of other vertebrate-infecting flavivirus RNAs, and the 3′ pentanucleotide sequence, which is completely conserved among mosquito- and tick-borne flaviviruses contained a point mutation in cell fusing agent virus. The function of this pentanucleotide sequence element is thought to be involved in the binding of cellular or viral proteins to the 3′ stem loop structure during RNA replication 
. The 3′ UTR of CxFV also was found to contain four tandem repeats, hypothesized to be specially adapted for replication in the mosquito host 
since deletion of conserved tandem repeat sequences alters virus growth properties 
. Finally, CxFV may target and replicate in different mosquito tissues than WNV or other flaviviruses. In Culex quinquefasciatus
inoculated simultaneously with CxFV Izabal and WNV, midgut and head tissues became infected with both viruses, demonstrating a potential for physical interaction between CxFV Izabal and WNV (,). However it is unclear if these tissue tropisms would be the same for mosquitoes naturally-infected with CxFV or if infection of these tissues is an artifact of inoculation, or inoculation simultaneously with WNV. More work is needed on characterizing the tissue tropisms of CxFV in naturally-infected mosquitoes and the mechanism by which this virus propagates and is transmitted.
One limitation of this study is that the natural mechanism by which mosquitoes become infected with CxFV has not yet been elucidated, so mosquitoes in this study were infected with CxFV Izabal by intrathoracic inoculation. It is unclear how or if the results of this study would be different using naturally-infected mosquitoes. Route of infection has been shown to affect the outcome of arbovirus superinfection studies. Most notably, Aedes triseriatus
mosquitoes that were infected transovarially with LaCrosse virus remained permissive to secondary infection with a homologous or heterologous bunyavirus 
, whereas mosquitoes exposed to the primary infection by intrathoracic inoculation became refractory to superinfection after seven days 
. Ideally, these studies should be repeated using mosquitoes naturally-infected with CxFV to fully understand the dynamics of interaction, or lack thereof, between these two flaviviruses within the mosquito vector. However the advantage of inoculations is that experiments can be standardized by infecting mosquitoes of the same age with approximately the same amount of virus, and 100% infection rates are assured.
Interestingly, both CxFV Izabal and WNV were found in saliva of co-infected specimens when mosquitoes were exposed to both viruses simultaneously by intrathoracic inoculation, but no CxFV Izabal was found in the saliva of singly-infected specimens. This observation suggests that CxFV Izabal may be infecting the salivary glands by “piggybacking”on WNV. This phenomenon has been suggested for expansion of cellular tropism by human immunodeficiency virus (HIV), whereby Epstein-Barr virus, cytomegalovirus, human t-lymphotrophic virus, and sperm proteins share large regions of similarity with the CD4 protein of T-helper lymphocytes, a cellular receptor used by HIV 
. Because HIV binds to CD4, binding of HIV to CD4 homologues on other co-infecting viruses or sperm may allow HIV to “piggyback” into additional cell types which it normally would not infect 
. The molecular basis for this interaction between CxFV Izabal and WNV and how these results compare to natural infection is unknown. Our intrathoracic inoculation data suggest that CxFV Izabal may have the potential to enhance WNV transmission in some mosquito populations; however WNV transmission was not enhanced in Honduras colony when mosquitoes were exposed per os
In summary, this is the first study to address the potential effect of an insect-specific flavivirus on transmission of WNV. We have demonstrated that CxFV Izabal does not interfere with growth of WNV in C6/36 cells or in Cx. quinquefasciatus, nor does it inhibit infection, dissemination, or transmission of WNV. These findings are in contrast to what would be expected based on previous studies following flavivirus – flavivirus superinfections. We hypothesize that both CxFV Izabal and WNV have evolved mechanisms for persistence and transmission by a common mosquito vector, Cx. quinquefasciatus, despite the presence of mosquito immune defenses and the prevalence of co-circulating flaviviruses. Future studies should address the effect of CxFV and WNV co-infection in mosquitoes naturally infected with CxFV, as well as the tissue tropisms and molecular mechanisms of CxFV replication and transmission in mosquitoes.