Insect-Specific Flaviviruses from Culex Mosquitoes in Colorado, with Evidence of Vertical Transmission

Study area.

The five-county study area (Larimer, Weld, Morgan, Logan, and Washington counties) is located in northeastern Colorado and has been previously described.17 Briefly, mosquito sampling was conducted along three rivers: the Cache la Poudre River and the Big Thompson River, which both emerge from the Rocky Mountains in western Larimer County, and the South Platte River, into which the other two rivers merge in the eastern Colorado plains (Figure 1). Mosquito collections were conducted in plains, foothills, and montane areas along these riparian corridors in sampling sites that could be accessed by automobile. Site locations were mapped with a global positioning system (GPS) receiver (Trimble Geo XT; Trimble Corp., Sunnyvale, CA) and visualized using ArcGIS 9.3 (ESRI, Redlands, CA).

Location of mosquito sampling sites for 2006 (along the Cache la Poudre River) and 2007 (along the Big Thompson River and South Platte River). The location of the targeted five-county area in Colorado is shown in the inset map.

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Location of mosquito sampling sites for 2006 (along the Cache la Poudre River) and 2007 (along the Big Thompson River and South Platte River). The location of the targeted five-county area in Colorado is shown in the inset map.

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

Location of mosquito sampling sites for 2006 (along the Cache la Poudre River) and 2007 (along the Big Thompson River and South Platte River). The location of the targeted five-county area in Colorado is shown in the inset map.

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

Mosquito collection and identification.

Mosquitoes were collected using CO₂-baited Centers for Disease Control and Prevention (CDC) miniature light traps (John W. Hock Company, Gainesville, FL) that were suspended ∼1.5 m above the ground and operated from afternoon (1500–1700 hours) until morning (0800–1000 hours). Sampling sites contained two traps baited with ∼1 kg dry ice, and they were located directly along the aforementioned rivers. Sampling during the summer of 2006 included 10 sites located along the Cache la Poudre River and 1 site by the Dixon Reservoir in Fort Collins (Figure 1). This spanned an elevation gradient from below 1,600 m in Fort Collins up to 2,360 m in the Poudre Canyon. The sites were sampled every 2 weeks from April 10 to October 27, 2006.

Sampling in 2007 included 20 sites along the Big Thompson and South Platte rivers and two additional sites located south of the Big Thompson River in the Loveland area (Figure 1). This included an elevation gradient ranging from 1,215 m in the prairie landscape of eastern Colorado to 1,840 m in the montane habitat of the Big Thompson Canyon. The sites were sampled every 2 weeks beginning on June 20 and continuing through September 14, 2007. Collected mosquitoes were examined with a dissecting microscope and identified to species using morphological characters.18,19

Detection of flavivirus RNA and nucleotide sequencing.

Culex mosquitoes were examined for presence of viral RNA by RT-PCR following the method of Bolling and others,20 with the modifications outlined below. Mosquitoes were identified on a chill table and placed in pools of 1–50 by species, sex, site, trap location, and date. Mosquito pools were then stored at −70°C until processed for viral RNA detection.

Each pool was triturated for 45 seconds with a vortex mixer in a 5-mL round-bottom polypropylene tube (Becton Dickinson, Franklin Lakes, NJ) using 1.5 mL diluent (minimum essential medium [MEM] containing 2% fetal bovine serum [FBS] and 100 μg/mL penicillin/streptomycin supplemented with l-glutamine and non-essential amino acids) and four copper-coated steel shot (4.5-mm diameter and 0.177-in caliber). Suspensions were centrifuged at 3,000 rpm for 10 min at 4°C. Total RNA was extracted from 140 μL supernatant using the QIAamp viral RNA Mini kit (Qiagen Inc., Valencia, CA). RNA was eluted in 60 μL nuclease-free water (Ambion Inc., Austin, TX).

RT-PCR was carried out using SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA) and GoTaq DNA polymerase (Promega, Madison, WI). Mosquito pools were first tested using universal flavivirus primers (cFD2 and MAMD) targeting a portion of the NS5 gene.21 Pools testing positive for flavivirus RNA were tested with virus-specific primers (Calbertado virus [CLBOV] primer sequences available on request).5,22,23 PCR products were visualized after electrophoresis on a 1% agarose gel stained with ethidium bromide. Negative (no template) and positive controls were included in each RT-PCR.

Infection rates per 1,000 individuals were calculated as bias-corrected maximum likelihood estimates using the Excel Add-In PooledInfRate, version 3.0.24 Presented infection rates are based on males and females combined. The most common Culex spp. found in our light-trap collections were Cx. tarsalis Coquillett and Cx. pipiens. Because Cx. tarsalis are easily identifiable, even when badly damaged, mosquitoes identified only as Culex spp. were grouped with Cx. pipiens for analyses.

For nucleotide sequencing, RT-PCR was carried out as described above using primers FU1 or FU2 and cFD3.25 The PCR products were purified using a QIAquick Gel Extraction kit (Qiagen Inc., Valencia, CA) and sequenced on an ABI 3130 Genetic Analyzer. Sequences obtained for the NS5 genome region were compared with other flavivirus sequences using the BLASTn program (blast.ncbi.nlm.gov/Blast.cgi).26

Phylogenetic analysis.

Nucleotide sequences of insect-specific flaviviruses detected from Culex spp. in Colorado were compared with homologous flavivirus sequences as follows. The sequences were aligned using ClustalX 1.81.27 Phylogenetic trees were constructed using neighbor-joining, maximum parsimony, and maximum likelihood analyses in PAUP 4.028 as well as Bayesian analysis using MrBayes 3.1.29 Resulting trees were midpoint-rooted using RETREE from the Phylip package30 and were displayed using Treeview 1.6.6.31

Virus isolation.

To isolate viruses, 100 μL homogenate supernatants from infected pools were inoculated onto Vero cells (African green monkey kidney), DF-1 cells (chicken embryo fibroblast), and C6/36 cells (Ae. albopictus) in 25-cm² flasks. After addition of 1 mL medium, the flasks were rocked for 1 hour at room temperature; 4 mL medium were then added, and cells were monitored daily for cytopathic effects (CPE). The Vero and DF-1 cells were maintained at 37°C and 7% CO₂ with MEM supplemented with 10% FBS, penicillin/streptomycin, l-glutamine, and non-essential amino acids. The C6/36 cells were maintained at 28°C with L-15 medium supplemented with 7% FBS, penicillin/streptomycin, l-glutamine, and non-essential amino acids. Cell culture medium was harvested after each passage for four passages and tested by RT-PCR to assess virus propagation.

Immunofluorescence assay.

Indirect fluorescent antibody assays (IFA) were conducted on spot slides of C6/36 cells infected with either CxFV or CLBOV using polyclonal antibodies against Japanese encephalitis virus (JEV), St. Louis encephalitis virus (SLEV), WNV, CxFV, or CLBOV. Murine hyperimmune ascitic fluids for JEV (M30178ABY), SLEV (VS0102), and WNV (M30200ABY) were obtained from the Division of Vector-Borne Infectious Diseases, Centers for Disease Control and Prevention, Fort Collins, CO. Virus-specific antisera for the Colorado strains of CxFV and CLBOV were produced by subcutaneous immunization of imprinting control region (ICR) mice with infected C6/36 cell culture medium clarified by centrifugation. Mice were immunized three times at 2-week intervals with complete Freund's adjuvant for the first immunization and incomplete Freund's adjuvant for the remaining immunizations. The use of all animals was reviewed and approved by the Animal Care and Use Committee at Colorado State University (Animal Protocol 08-067A-03). An indirect enzyme-linked immunosorbent assay (ELISA) of mouse sera 2 week after the final immunization confirmed antibody production. Serum was cross-absorbed two times to sonicated, uninfected C6/36 cells to reduce non-specific binding. Secondary antibody for IFA was biotinylated sheep anti-mouse immunoglobulin G (IgG) followed by streptavidin-fluorescein (GE Healthcare Bio-Sciences Corp., Piscataway, NJ).

Persistently CxFV-infected mosquito colony.

A Cx. pipiens colony was established from egg rafts collected in Fort Collins during the summer of 2005. After the discovery of CxFV and CLBOV in local field populations, the colony was determined to be infected with CxFV by RT-PCR as described above and virus isolation in C6/36 cells. Mosquito homogenates taken from the colony in 2005 and stored at −80°C, tested positive for CxFV, indicating that the mosquitoes were infected at the time that the colony was established. The colony is maintained in a 60 × 60 × 60-cm cage at 25°C, 75% relative humidity, and a 16 hour to 8 hour, light to dark cycle. To investigate viral maintenance within the colony, RNA extractions were initially performed using a QIAamp viral RNA Mini kit as described above. To increase sensitivity of RNA detection in individual mosquitoes, Trizol (Invitrogen, Carlsbad, CA) was used to extract total RNA from individual egg rafts, individual fourth instar larvae, and individual adult mosquitoes. A quantitative RT-PCR assay was designed using a QuantiTect SYBR Green RT-PCR kit (Qiagen Inc., Valencia, CA) and CxFV-specific primers (available on request).

Discovery of insect-specific flaviviruses in Culex spp.

In 2006, ∼1,300 Culex mosquitoes from northeastern Colorado were tested for viral RNA by RT-PCR using universal flavivirus primers. Unexpectedly, numerous Culex pools testing positive for flavivirus RNA were negative for the flaviviruses that most commonly are found in Culex in Colorado: WNV and SLEV. To rule out contamination, 80 pools (89 mosquitoes) of Culiseta spp. mosquitoes and 64 pools (1,997 mosquitoes) of Ae. vexans Meigen were tested by RT-PCR with the universal flavivirus primers. All of these pools tested negative.

PCR products from flavivirus-positive Culex pools were then sequenced, revealing similarity in some cases to a previously described insect-specific flavivirus, CxFV. In addition, we discovered sequences from a novel insect-specific flavivirus referred to as Calbertado virus (CLBOV), because this virus has been found in California, Alberta, Canada,32,33 and Colorado. Specific primers were designed to detect these insect-specific flaviviruses in mosquito pools.

In 2007, we collected ∼43,000 Culex spp. mosquitoes. All Culex pools were first tested with the universal flavivirus primers. Thereafter, all flavivirus-positive pools were tested with WNV-specific, CxFV-specific, and CLBOV-specific primers.

Trends for insect-specific flaviviruses in Culex spp. in time and space.

Infection rates for WNV, CxFV, and CLBOV in Cx. tarsalis and Cx. pipiens were calculated by month (Table 1) and site (Table 2) for both 2006 and 2007. Infection rates include male and female pools of mosquitoes combined, because viral RNA was detected in both. The seasonal patterns of infection rates are shown for 2007 (Figure 2) when Culex abundance was higher and virus detection was based on larger numbers of mosquitoes than in 2006.

Monthly infection rates from June to September 2007 for Cx. tarsalis (A–C) and Cx. pipiens (D–F) with WNV (A and D), CxFV (B and E), and CLBOV (C and F). Error bars indicate 95% skewness-corrected confidence intervals. Estimates for infection rates per 1,000 mosquitoes were calculated with Excel Add-In PooledInfRate 3.0.24

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Monthly infection rates from June to September 2007 for Cx. tarsalis (A–C) and Cx. pipiens (D–F) with WNV (A and D), CxFV (B and E), and CLBOV (C and F). Error bars indicate 95% skewness-corrected confidence intervals. Estimates for infection rates per 1,000 mosquitoes were calculated with Excel Add-In PooledInfRate 3.0.24

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

Monthly infection rates from June to September 2007 for Cx. tarsalis (A–C) and Cx. pipiens (D–F) with WNV (A and D), CxFV (B and E), and CLBOV (C and F). Error bars indicate 95% skewness-corrected confidence intervals. Estimates for infection rates per 1,000 mosquitoes were calculated with Excel Add-In PooledInfRate 3.0.24

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

In the case of the 2006 collections, most CxFV-positive pools (36/37) came from Cx. pipiens, with only a single Cx. tarsalis pool positive for CxFV. Conversely, all (34) CLBOV-positive pools were from Cx. tarsalis (Table 1). This pattern was less clear in 2007, when both CxFV and CLBOV were found in Cx. tarsalis as well as Cx. pipiens. However, there still was a trend for most CxFV-positive pools being recorded from Cx. pipiens (contributing 125 of 178 CxFV-positive pools) and most CLBOV-positive pools coming from Cx. tarsalis (contributing 113 of 121 CLBOV-positive pools).

The overall infection rates per 1,000 Cx. tarsalis in 2006 were 5.47 for WNV, 0.83 for CxFV, and 40.13 for CLBOV. The corresponding infection rates for Cx. tarsalis in 2007 were 2.65 for WNV, 1.34 for CxFV, and 2.95 for CLBOV. Differences between years are likely related, in part, to the fact that different sites were sampled in 2006 and 2007. The overall infection rates per 1,000 Cx. pipiens in 2006 were 0 for WNV, 462.42 for CxFV, and 0 for CLBOV. The corresponding infection rates for Cx. pipiens in 2007 were 2.41 for WNV, 72.09 for CxFV, and 3.18 for CLBOV. Notably, CxFV infection rates in Cx. pipiens were very high compared with the other two viruses.

In 2007, the WNV and CxFV infection rates in Cx. tarsalis increased gradually from June to September, whereas the CLBOV infection rates were highest in June (3.28) and then decreased slightly from July to September (Table 1 and Figure 2). WNV and CLBOV infection rates in Cx. pipiens followed similar seasonal patterns, with the highest infection rates occurring in June when mosquito counts were low (Table 1 and Figure 2).

Both insect-specific flaviviruses occurred widely in the five-county study area. CxFV was detected from all 21 sites in the plains sampled in 2006 and 2007, and CLBOV was recorded from 20 of these sites. The viruses also occurred in foothills sites but were not recorded from montane sites where mosquito counts are very low (Table 2). Site-specific infection rates for WNV were compared with site-specific CxFV and CLBOV infection rates by graphing, but there were no apparent associations (data not shown).

Phylogenetic analyses.

Bayesian analysis was conducted using a 1-kb region of the NS5 gene sequence to assess phylogenetic relationships between insect-specific flavivirus isolates from Culex spp. mosquitoes in Colorado and other selected flaviviruses (Figure 3). Neighbor-joining, maximum parsimony, and maximum likelihood analyses resulted in similar tree topologies (data not shown). The CxFV isolate from this study was most similar to CxFV isolates from Texas and was grouped with other CxFV isolates from Iowa, Japan, Mexico, and Guatemala (Figure 3). The Colorado CLBOV isolate shared the closest phylogenetic relationships with the CLBOV sequences detected in California and Alberta, Canada,32,33 encoding 97% NS5 amino acid sequence similarity.

Phylogenetic tree inferred from Bayesian analysis and midpoint rooted, showing relationships between insect-specific flaviviruses from Culex species mosquitoes collected in Colorado (shaded) and other flaviviruses based on a 1-kb segment of the NS5 gene. The numbers at the nodes represent clade credibility values and the scale bar indicates the number of estimated substitutions per site. GenBank accession numbers for the sequences used in the analysis are listed in parentheses.

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Phylogenetic tree inferred from Bayesian analysis and midpoint rooted, showing relationships between insect-specific flaviviruses from Culex species mosquitoes collected in Colorado (shaded) and other flaviviruses based on a 1-kb segment of the NS5 gene. The numbers at the nodes represent clade credibility values and the scale bar indicates the number of estimated substitutions per site. GenBank accession numbers for the sequences used in the analysis are listed in parentheses.

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

Phylogenetic tree inferred from Bayesian analysis and midpoint rooted, showing relationships between insect-specific flaviviruses from Culex species mosquitoes collected in Colorado (shaded) and other flaviviruses based on a 1-kb segment of the NS5 gene. The numbers at the nodes represent clade credibility values and the scale bar indicates the number of estimated substitutions per site. GenBank accession numbers for the sequences used in the analysis are listed in parentheses.

Citation: The American Society of Tropical Medicine and Hygiene 85, 1; 10.4269/ajtmh.2011.10-0474

• Download Figure
• Download figure as PowerPoint slide

Based on Bayesian analysis, the insect-specific flavivirus clade contains two subclades, which correspond with insect host genus. The first subclade contains CLBOV, Quang Binh virus, and the CxFV isolates, which have all been detected in Culex mosquitoes. The second subclade contains Kamiti River virus, cell fusing agent virus, and Aedes flavivirus, which have all been described in Aedes mosquitoes. Interestingly, Nounané virus, which was recently isolated from Uranotaenia mosquitoes in Côte d'Ivoire,8 seems to only replicate in insect cells and based on our phylogenetic analysis, groups with the arthropod-borne flaviviruses.

Virus isolation.

Isolation of insect-specific flaviviruses from Culex in Colorado was attempted by blind passages in Vero, DF-1, and C6/36 cell cultures. Vero (mammalian) and DF-1 (avian) cells inoculated with mosquito homogenate supernatants did not exhibit any CPE after one to four passages. RNA extractions were performed on Vero and DF-1 cell culture medium after each passage, and these tested negative by RT-PCR using virus-specific primers. In contrast, CxFV-infected Cx. pipiens homogenate supernatants caused minor growth inhibition for C6/36 (mosquito) cells after several passages. RNA extracted from cell culture medium was CxFV-positive by RT-PCR after each passage. CLBOV-infected Cx. tarsalis homogenate supernatants caused apparent CPE in C6/36 cells on 5 dpi only after 10 passages. RNA extractions of C6/36 cell culture medium for passages one to three were negative by RT-PCR using virus-specific primers, but after the fourth passage, they were positive. Spot slides of CxFV- and CLBOV-infected C6/36 cells were tested by IFA with JE, SLE, and WNV antibodies based on a previous report,14 but antigens were undetectable. IFA using virus-specific antibodies for CxFV and CLBOV produced positive results, confirming C6/36 cell infections.

CxFV infection in a Cx. pipiens laboratory colony.

After detection of CxFV from adult mosquitoes collected in 2006 and 2007, our Cx. pipiens laboratory colony, established during the summer of 2005 from egg rafts collected in Fort Collins, was tested in March of 2007 and found to be infected with CxFV by RT-PCR and virus isolation. A subset of mosquitoes had been taken from the colony in September of 2005 and stored at −80°C. These stored specimens also tested positive for CxFV, indicating that the mosquitoes were infected at the time that the colony was established. Our initial RNA extractions were performed using a column-based method that is used for screening field-collected pools; however, we compared this with a Trizol extraction method for individual colony mosquitoes and found that Trizol extraction provided more sensitive detection. Briefly, males and females of various ages were aspirated from the colony, and RNA was extracted from individual mosquitoes using the column-based method and the Trizol method. Samples were tested by standard RT-PCR, with 1 of 28 mosquitoes testing positive for CxFV using the column-based RNA extraction method and 6 of 16 mosquitoes testing positive for CxFV using the Trizol RNA extraction method. To examine viral maintenance within this naturally infected colony, total RNA was extracted from individual egg rafts, individual fourth instar larvae, and individual adult (male and female) mosquitoes using Trizol and tested by a qRT-PCR assay with virus-specific primers. All life stages were found to be positive for CxFV RNA.

Trends for insect-specific flaviviruses in Culex spp. in time and space.

We found that infection rates for CxFV and CLBOV varied by Culex species, month, and site. In 2006, there was a strong species-specific pattern with CxFV being detected almost exclusively in Cx. pipiens and all records of CLBOV coming from Cx. tarsalis. In 2007, we collected and examined far greater numbers of Culex mosquitoes, and both insect-specific flaviviruses were detected in Cx. tarsalis as well as Cx. pipiens. These findings suggest that CxFV and CLBOV circulate in both species of mosquitoes in northern Colorado, although we cannot entirely rule out the possibility that a body part from one species sometimes was accidentally combined with a pool from the other species. Culex flavivirus has previously been detected in a variety of Culex spp., including Cx. pipiens,5,15 Cx. tarsalis,15 Cx. quinquefasciatus Say,5,12,14 Cx. tritaeniorhynchus,5 and Cx. restuans Theobald.14 These findings indicate that CxFV occurs in numerous Culex species. We also tested a subset of Ae. vexans and Culiseta spp. pools, but these were negative with the pan-flavivirus primers used. Other studies also found Aedes species pools to be negative for CxFV RNA.5,15

We detected CLBOV RNA in both Cx. tarsalis and Cx. pipiens from Colorado and isolated the virus from several Cx. tarsalis pools. A portion of the NS5 gene of this virus shares a high similarity to viral sequences found in Cx. tarsalis mosquitoes in Alberta and other western provinces of Canada.32,33 Although a virus isolate has not yet been obtained from Alberta, detection of this sequence in a number of Canadian collections of Culex species suggests that CLBOV may have a wide geographic range. Field studies in California also indicate that CLBOV is present in Cx. tarsalis and Cx. pipiens populations there (Brault A, unpublished data).

Our data suggest that CxFV and CLBOV are prevalent in Cx. tarsalis and Cx. pipiens throughout the northeastern Colorado plains. In 2006, CxFV RNA was detected in April and every month from June to October (Table 1). In 2007, CxFV was detected in both species from June to September. In Texas, Kim and others14 recorded CxFV-positive mosquito pools during February and March, but continued surveillance from April to August resulted in only negative pools. In Iowa, CxFV RNA was detected from July to October but not in May or June, possibly because of low numbers of Culex collected early in the season.15

Our study yielded very high infection rates for CxFV in Cx. pipiens. Other studies have reported variable infection rates for CxFV, with minimum infection rates (MIR) per 1,000 Cx. quinquefasciatus ranging from 4.7 in Guatemala12 to 20.8 in Mexico,13 and overall MIRs in Iowa ranging from 1.2 in Cx. tarsalis to 10.3 for Cx. pipiens.15 Both our study (Table 2) and the Iowa study recorded substantial site-specific variability for CxFV infection rates in a given species. Additional research is needed to determine the mechanisms resulting in spatial and temporal variability in CxFV infection rates.

We present the first data regarding seasonality of CLBOV, which were recorded in April 2006 and then every month from June to September in 2006 and 2007 (Table 1). Infection rates with CLBOV for months when at least 100 Cx. tarsalis or Cx. pipiens were examined ranged from 2.15 to 36.79. As a crude comparison, in Alberta, Canada, CLBOV was detected in 67 of 140 (48%) Cx. tarsalis mosquito pools tested from 2003 to 2005.32,33

Comparison of seasonal infection rates for CxFV, CLBOV, and WNV.

Sample sizes for Culex mosquitoes in 2007 were adequate to compare the seasonal (monthly) patterns for infection rates with CxFV, CLBOV, and WNV from June to September (Figure 2). This produced some unexpected and intriguing results. For instance, the infection rates in Cx. tarsalis for WNV and CxFV followed similar patterns, gradually increasing throughout the study period (Figure 2A and B), whereas the CLBOV infection rate showed an opposite, slightly decreasing trend from June to September (Figure 2C). The infection rates in Cx. pipiens with WNV, CxFV, and CLBOV followed the same general decreasing trend from June to September (Figure 2D–F), but the infection rates observed for CxFV were much higher compared with WNV and CLBOV. It is important to consider that high infection rates can be an artifact of low sample size, which is indicated by larger error bars when mosquito counts are low (Figure 2). We also recognize that the data for Cx. pipiens in our study should be interpreted with care, because use of CDC light traps can underestimate the abundance of this species compared with efforts that also include gravid traps.34

Natural maintenance of CxFV and CLBOV.

Culex flavivirus and CLBOV were detected in both male and female mosquito pools in our studies. Other studies investigating insect-specific flaviviruses have reported similar findings.5,7,11 Presence of insect-specific flaviviruses in mosquitoes of both sexes suggests vertical transmission as a mechanism of viral maintenance in nature.11 Evidence supporting vertical transmission for KRV includes that the first isolates came from adult Ae. macintoshi mosquitoes that were collected as larvae and pupae from flooded dambos,4 and that female Ae. aegypti, infected with KRV by oral exposure, transmitted the virus to their offspring.35 To investigate how insect-specific flaviviruses are maintained in mosquito populations, quantitative RT-PCR was performed on specimens from our laboratory colony of Cx. pipiens that was established from egg rafts collected in Fort Collins in 2005 and later found to be naturally and persistently infected with CxFV. Total RNA was extracted from individual egg rafts, single larvae, and single adults, with high proportions of all stages testing positive for CxFV. These results provide more evidence supporting vertical transmission as a means of viral maintenance in natural populations.

Vertical transmission seems to play a role in viral maintenance for other viruses in the genus Flavivirus. Vertical transmission of a flavivirus was first described in Senegal with isolation of Koutango virus from a male Ae. aegypti.36 Since then, there have been numerous descriptions of flaviviruses isolated from larvae or male mosquitoes, including JEV from Cx. tritaeniorhynchus,37 yellow fever from Ae. aegypti,38 and WNV from Cx. univittatus Theobald,39 and laboratory studies also have shown vertical transmission of flaviviruses.40–43 It is thought that flaviviruses are vertically transmitted at the time of oviposition during fertilization through the micropyle, as the fully developed egg passes through the oviduct.44 This is much less efficient compared with true transovarial transmission, where the virus infects the developing egg.45 Filial infection rates seen with vertically transmitted flaviviruses in mosquitoes are usually low (less than 1%) compared with much higher rates seen with bunyaviruses.46 There are few data regarding the transmission dynamics of insect-specific flaviviruses, which have been found to replicate only in invertebrate cells. The inability to infect and replicate in vertebrate cells indicates that this group of flaviviruses has a distinct transmission cycle compared with the arthropod-borne flaviviruses, which are maintained between arthropod vectors and vertebrate hosts. Studies are planned to further investigate transmission dynamics of CxFV in our naturally infected Cx. pipiens colony.

If CxFV and CLBOV are indeed maintained in nature exclusively by vertical transmission, one might expect very high infection rates throughout the study area. This, however, was not observed by us or in other studies on CxFV. An important consideration with viral surveillance is the sensitivity of the assay used. Initial studies with the CxFV-infected laboratory colony after a column-based RNA extraction and standard RT-PCR resulted in detection of few CxFV-positive specimens. Although this method of RNA extraction was efficient and effective for mosquito pools, RNA extractions were performed using Trizol (Invitrogen) to increase sensitivity of RNA detection in individual mosquitoes from the laboratory colony, and this was followed by quantitative RT-PCR. These methods revealed a higher infection rate in the naturally infected colony. These observations suggest that viral titers in individual mosquitoes may sometimes be below the threshold of sensitivity for certain testing methods; however, we are confident that the column-based RNA extraction method combined with standard RT-PCR is adequate for detection of viral RNA in mosquito pools. It is unclear at this time why insect-specific flavivirus infection rates observed in this study and other studies are lower than would be expected if vertical transmission was an efficient means for viral maintenance in nature.

In conclusion, we detected two insect-specific flaviviruses in Culex spp. mosquitoes collected in northern Colorado: Culex flavivirus, first isolated in Japan,5 and Calbertado virus, a novel flavivirus. Based on 2 years of field studies, we have provided the first description of the seasonal dynamics of these viruses in Colorado. The common occurrence of these viruses in natural Culex populations raises questions regarding possible interactions with other flaviviruses that do cause disease in humans, such as WNV.3 Future studies are planned to determine how insect-specific flaviviruses may interact with arboviruses in a coinfected mosquito and how this may potentially impact vector competence.