|Home | About | Journals | Submit | Contact Us | Français|
Adult mosquitoes (Diptera: Culicidae) were collected in 2007 and tested for specific viruses, including West Nile virus, as part of the ongoing arbovirus surveillance efforts in the state of Iowa. A subset of these mosquitoes (6,061 individuals in 340 pools) was further tested by reverse transcription-polymerase chain reaction (RT-PCR) using flavivirus universal primers. Of the 211 pools of Culex pipiens (L.) tested, 50 were positive. One of 51 pools of Culex tarsalis Coquillet was also positive. The flavivirus minimum infection rates (expressed as the number of positive mosquito pools per 1,000 mosquitoes tested) for Cx. pipiens and Cx. tarsalis were 10.3 and 1.2, respectively. Flavivirus RNA was not detected in Aedes triseriatus (Say) (52 pools), Culex erraticus (Dyar & Knab) (25 pools), or Culex territans Walker (one pool). Sequence analysis of all RT-PCR products revealed that the mosquitoes had been infected with Culex flavivirus (CxFV), an insect-specific virus previously isolated in Japan, Indonesia, Texas, Mexico, Guatemala and Trinidad. The complete genome of one isolate was sequenced, as were the envelope protein genes of eight other isolates. Phylogenetic analysis revealed that CxFV isolates from the United States (Iowa and Texas) are more closely related to CxFV isolates from Asia than those from Mexico, Guatemala, and Trinidad.
The majority of viruses in the genus Flavivirus are arthropod-borne viruses that are transmitted between vertebrate hosts by mosquitoes or ticks (ICTV 2005). Other viruses in this genus have a vertebrate host, but no known arthropod vector. Viruses in the final group replicate in mosquitoes but have no apparent vertebrate host and are therefore assumed to be insect-specific. Three insect-specific flaviviruses have been identified, and these are Culex flavivirus (CxFV), cell fusing agent virus (CFAV) and Kamiti River virus (KRV) (Stollar and Thomas 1975, Crabtree et al. 2003, Sang et al. 2003, Hoshino et al. 2007). In addition, flavivirus-related DNA known as cell silent agent is integrated into the genomes of some Aedes spp. mosquitoes (Crochu et al. 2004). Phylogenetically, insect-specific flaviviruses occupy the most divergent lineage, suggesting that they are primordial flaviviruses that emerged before the other members of this genus (Cook and Holmes 2006, Hoshino et al. 2007).
The first insect-specific flavivirus to be discovered was CFAV, originally isolated from Aedes aegypti (L.) cell cultures in 1975 (Stollar and Thomas 1975, Cammisa-Parks et al. 1992), and later discovered in natural populations of Ae. aegypti, Aedes albopictus (Skuse), and Culex spp. mosquitoes in Puerto Rico, and Ae. aegypti in Thailand (Cook et al. 2006, Kihara et al. 2007). In Puerto Rico, CFAV was isolated from both female and male mosquitoes suggesting that the virus can be transovarially transmitted in nature (Cook et al. 2006). CFAV causes cytopathic effect (CPE), including massive syncytium formation, in Ae. albopictus (C6/36) cells, but it does not replicate in mice or in any of the vertebrate cell lines that have been tested (Stollar and Thomas 1975, Crabtree et al. 2003). KRV, the second insect-specific flavivirus to be discovered, was isolated from immature Aedes macintoshi Marks collected in Kenya in 1999 (Crabtree et al. 2003, Sang et al. 2003). KRV replicates in Ae. albopictus, Ae. aegypti and Aedes pseudoscutellaris (Theobald) cell cultures, but not in mice or vertebrate cell cultures. KRV causes CPE in C6/36 cells but not Ae. aegypti and Ae. pseudoscutellaris cells. The 3′ untranslated region (UTR) of KRV is unusual; it is nearly twice as long as the 3′UTRs of all other flaviviruses and seems to have formed as the result of an almost complete duplication of a primordial KRV 3′UTR (Gritsun and Gould 2006).
CxFV was first isolated from Culex spp. mosquitoes collected in Japan and Indonesia in 2003 and 2004 (Hoshino et al. 2007). Nine isolations were made from Cx. pipiens (n = 6), Culex quinquefasciatus Say (n = 2), and Culex tritaeniorhynchus Giles (n = 1). Isolations were made from both female and male mosquitoes. The authors observed that the isolates could replicate in C6/36 cells, but not in African green monkey kidney (Vero) or baby hamster kidney cells. Moderate CPE was periodically observed in C6/36 cells inoculated with CxFV that had been passed at least twice in cell culture. However, CPE was usually absent in cells infected with the original inoculum or virus passed once.
More recently, CxFV has been identified in the Western Hemisphere, with isolates obtained from Guatemala (Morales-Betoulle et al. 2008), the Yucatan Peninsula of Mexico (Farfan-Ale et al. 2009), Trinidad, Texas (Kim et al. 2009), and Colorado (B. G. Bolling, personal communication). In Guatemala, CxFV was detected by reverse transcription-polymerase chain reaction (RT-PCR) in eight of 210 pools of female Cx. quinquefasciatus collected in 2006 (Morales-Betoulle et al. 2008). The CxFV minimum infection rate (MIR), expressed as the number of positive mosquito pools per 1,000 mosquitoes tested, was 4.7. This strain of CxFV (designated Izabal) did not cause CPE in C6/36 cells. In Mexico, CxFV was detected by RT-PCR in 145 of 210 pools (MIR 20.8) of Cx. quinquefasciatus collected in 2007 (Farfan-Ale et al. 2009). Like the Izabal strain, CxFV strains from Mexico do not induce CPE in C6/36 cells. The virus was detected in both female and male mosquitoes. In Texas, CxFV was detected in 37 of 388 pools of female Cx. quinquefasciatus or Culex restuans Theobald collected in 2008 (Kim et al. 2009). Some CxFV strains from Texas produced marked CPE and syncytia formation in C6/36 cells, whereas others did not. Phylogenetic analyses of the envelope protein (E) genes have shown that CxFV can be separated into two distinct clades. One clade contains CxFV isolates from Texas and Asia, and the second clade contains CxFV isolates from Latin America and the Caribbean (Kim et al. 2009).
In the current study, mosquitoes collected in Iowa in 2007 were examined by RT-PCR using flavivirus-specific primers. Of the 340 pools tested, 51 were positive for CxFV RNA. A subset of mosquito homogenates was processed by virus isolation in C6/36 cells, and 12 isolates were obtained. The genomic RNA of one isolate and the E genes of another eight isolates were sequenced. This report presents the sequence and phylogenetic data.
Collections were made in multiple sites in 13 counties in the state of Iowa from May through October 2007 (Fig. 1). Mosquitoes were sampled using CDC light, gravid, and mosquito magnet traps set from dusk until the next morning. Mosquitoes were transported on ice to the Medical Entomology Laboratory at Iowa State University (ISU), where they were sorted on chill tables into groups of up to 50 according to species, sex, date, and study site, and then stored at −80°C. Due to the difficulty in morphologically distinguishing adult Cx. pipiens and Cx. restuans individuals, these mosquitoes were grouped and are referred to as the Cx. pipiens group as described previously (DeGroote et al. 2007, Sucaet et al. 2008).
Mosquitoes were homogenized in 1.8 ml of diluent (minimum essential media containing 10% fetal bovine serum, l-glutamine, penicillin, streptomycin, and Amphotericin B) as described previously (Farfan-Ale et al. 2009). Total RNA was extracted from mosquito homogenates using the QIAamp viral RNA extraction kit (QIAGEN, Valencia, CA), and analyzed by RT-PCR using the flavivirus-specific primers: FU2: 5′-GCT GAT GAC ACC GCC GGC TGG GAC AC-3′ and cFD3: 5′-AGC ATG TCT TCC GTG GTC ATC CA-3′, which target a 845-nt region of the NS5 gene (Kuno et al. 1998). Complementary DNAs were generated using SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA), and PCRs were performed using Taq polymerase (Invitrogen) as described previously (Farfan-Ale et al. 2009). An aliquot of each PCR product was examined by electrophoresis, visualized with ethidium bromide, and extracted using the PureLink gel extraction kit (Invitrogen). Purified DNAs were sequenced using a 3730×1 DNA sequencer (Applied Biosystems, Foster City, CA) at ISU.
An aliquot (100 μl) of selected mosquito homogenates was added to 2 ml of Liebovitz's L15 medium (Invitrogen) supplemented with 2% fetal bovine serum, l-glutamine, penicillin, streptomycin, and Amphotericin B. Samples were filtered and inoculated onto subconfluent monolayers of Ae. albopictus C6/36 cells in 75-cm2 flasks. Cells were incubated for at least 1 h at room temperature on an orbital shaker to allow attachment of the virus. Another 12 ml of L15 maintenance medium was added to each flask, and cells were incubated at 28°C for 7 d. After two additional blind passages, supernatants were collected and tested for the presence of CxFV.
Primers for the RT-PCR amplification and sequencing of CxFV were designed using the genomic sequences of a Japanese CxFV isolate (strain NIID-21) and a Mexican CxFV isolate (strain CxFV-Mex07) (Hoshino et al. 2007, Farfan-Ale et al. 2009). The resulting PCR products were sequenced and used to design additional primers. In total, 24 pairs of overlapping primers were used, and primer sequences are available upon request. The extreme 5′ and 3′ ends of the CxFV genome were determined by 5′ rapid amplification of cDNA ends (RACE) and 3′ RACE, respectively, as described previously (Farfan-Ale et al. 2009).
As part of the ongoing arbovirus surveillance efforts in Iowa, mosquitoes are collected throughout the state each year and select species are assayed by RT-PCR using primers specific for West Nile, St. Louis encephalitis, La Crosse and/or Western equine encephalitis viruses. A subset of mosquitoes that had tested negative for these viruses was randomly selected for the current study. Overall, 6,061 mosquitoes (340 pools) collected in 13 Iowa counties from May through October 2007 were chosen. There were 4,847 Cx. pipiens group (211 pools), 859 Cx. tarsalis Coquillet (51 pools), 239 Ae. triseriatus (Say) (52 pools), 115 Culex erraticus (Dyar & Knab) (25 pools) and one Cx. territans Walker (one pool) (Table 1). All pools were comprised of female mosquitoes.
Mosquito pools were screened by RT-PCR using flavivirus-specific primers, and 51 pools tested positive (Table 1). Of these, 50 pools consisted of Cx. pipiens group and one pool consisted of Cx. tarsalis. The flavivirus MIRs for Cx. pipiens group and Cx. tarsalis were 10.3 and 1.2, respectively. Because of the considerable variation in pool sizes, bias corrected maximum likelihood estimation (MLE) values also were calculated using the PooledInfRate statistical software package (Biggerstaff 2006). The resulting MIR and MLE values were consistently very similar; therefore, MLE data are not shown. The flavivirus MIR for Cx. pipiens group showed considerable geographic variation (Table 2). For the five counties from which >200 Cx. pipiens group females were collected, the flavivirus MIR was 3.4–28.3. Flavivirus RNA was not detected in any Ae. triseriatus, Cx. erraticus, or Cx. territans. All PCR products were partially sequenced and shown to correspond to CxFV. Pairwise alignments of the resulting sequences revealed that they share >98% identity. The mosquitoes positive for CxFV RNA were collected from July through October (Table 3). The monthly MIRs for CxFV in Cx. pipiens group collected in this time period ranged from 7.2 to 61.0. CxFV was not detected in any mosquitoes collected in May or June.
Homogenates from 12 pools containing CxFV RNA were filtered and an aliquot of each was inoculated on C6/36 cells. One week later, an aliquot of each supernatant was inoculated onto fresh monolayers of C6/36 cells. In total, three blind passages were performed. Mild CPE and/or clumping of cells were observed in all cell cultures with the exception of mock-infected control cultures. CxFV RNA was detected by RT-PCR in all supernatants collected after the third blind passage (except for the control cultures). Taken together, these data indicate the CxFV had been isolated from all 12 homogenates tested.
The complete genomic sequence of one isolate (designated CxFV-Iowa07) was determined using a combination of RT-PCR, 5′ and 3′ RACE (GenBank accession FJ663034). The genomic RNA contains a single 10,089-nt open reading frame that is flanked by 5′ and 3′UTRs of 91 and 654 nt, respectively. In the 5′ RACE experiments, 12 cDNA clones were analyzed, and the length and sequence of all clones were identical. Genomic sequence data are available for four other CxFV isolates; two isolates are from Japan (strains Tokyo and NIID-21), one isolate is from Mexico (strain CxFV-Mex07), and one isolate is from Texas (strain TX 24518) (Hoshino et al. 2007, Farfan-Ale et al. 2009, Kim et al. 2009). Pairwise alignment of the CxFV-Iowa07 genomic sequence to the other CxFV genomic sequences revealed that it is most closely related to isolates from Japan and Texas (97.6–98.9% identical). There is 90.9% identity between the genomes of CxFV-Iowa07 and CxFV-Mex07. The predicted amino acid sequence of the CxFV-Iowa07 polyprotein precursor also is more closely related to the polyprotein precursors of CxFV isolates from Japan and Texas (≥99.0% identical, ≥99.4% similar) than to the isolate from Mexico (97.0% identical, 98.4% similar).
A phylogenetic tree was constructed with Bayesian methods using the complete genomic sequence of CxFV-Iowa07, and the genomic sequences of 24 other flavivirus isolates (Fig. 2). Phylogenetic trees also were generated using neighbor-joining (NJ), maximum parsimony (MP), and maximum likelihood (ML) methods. In the Bayesian tree, CxFV-Iowa07 shares a close phylogenetic relationship with CxFV isolates from Japan, Texas, and Mexico, consistent with the identity/similarity estimates. These viruses, together with the other insect-specific flaviviruses (CFAV and KRV) comprise a distinct clade (denoted as I). The posterior support for this topological arrangement is 100%. The other viruses separate into two additional clades. Clade II contains tick/vertebrate viruses and clade III contains mosquito/vertebrate viruses. Viruses with no known vector are present in both clades II and III. This topological arrangement is consistent to that observed in several other flavivirus phylogenetic studies (Cook and Holmes 2006). The ML and NJ trees were identical to the Bayesian tree, but the MP tree placed Modoc virus between the insect-specific flaviviruses and the other flaviviruses (data not shown).
After the completion of the CxFV-Iowa07 genomic sequencing experiments, the E genes of eight other CxFV isolates from Iowa were also sequenced (GenBank accessions FJ663026–FJ663033). This region comprises 1,281 nucleotides and corresponds to nucleotides 938–2,218 of the genomic RNA of the CxFV-Iowa07 strain. Pairwise alignments of these sequences revealed that they share 97.7 to 99.9% identity. The E gene sequences of 21 other isolates are also available in the GenBank database (Hoshino et al. 2007, Morales-Betoulle et al. 2008, Farfan-Ale et al. 2009, Kim et al. 2009). These isolates are from Japan (n = 9), Indonesia (n = 1), Texas (n = 7), Mexico (n = 1), Guatemala (n = 1), and Trinidad (n = 2). The nucleotide sequence of the E gene of CxFV-Iowa07 was aligned to the homologous regions of a representative isolate from each of these geographic locations. As observed with the genomic sequence alignments, this analysis demonstrated that CxFV-Iowa07 is most closely related to CxFV isolates from Japan, Indonesia, and Texas (≥97.7% identity) (Table 4). In contrast, the E gene of CxFV-Iowa07 shares no more than 90.1% identity to the homologous regions of CxFV isolates from Mexico, Guatemala, and Trinidad. Similarly, the deduced amino acid sequence of the E protein of CxFV-Iowa07 shares greater identity and similarity to the homologous regions of CxFV isolates from Japan, Indonesia, and Texas than to those from Mexico, Guatemala, and Trinidad.
A phylogenetic tree was constructed with Bayesian methods by using the E gene sequences of 30 CxFV isolates, including nine isolates from Iowa (Fig. 3). Phylogenetic trees also were generated using NJ, MP, and ML methods. In the Bayesian tree, the CxFV isolates formed two distinct clades (denoted as clade 1 and 2). Clade 1 is composed of CxFV isolates from Mexico, Guatemala, and Trinidad (denoted as the Latin American/Caribbean clade). The posterior support for this clade is 0.87. Clade 2 is composed of CxFV isolates from Iowa, Texas, Japan, and Indonesia (denoted as the U.S./Asian clade). The posterior support for this clade is 0.65. Clade 2 is further separated into three nested clades (denoted as 2a to 2c). Eight CxFV isolates from Iowa grouped in clade 2c, along with all the CxFV isolates from Texas. One CxFV isolate from Iowa (CxFV-Iowa07) clustered in 2b, along with several isolates from Japan. Clade 2a was composed of CxFV isolates from Japan and Indonesia. Whereas these three subclades have reasonably high posterior support (0.98, 0.94, and 0.77, respectively), their relationship is less certain. The ML tree showed the same topological features as the Bayesian tree. In the MP tree, clades 2a and 2b are nearest neighbors, rather than 2b and 2c. The Iowa/Texas sequences do not form a monophyletic clade in the NJ tree but instead branch off tightly just ancestral to the 2a/2b cluster (data not shown).
The placement of CxFV-Iowa07 in the phylogenetic tree constructed using E gene sequences is interesting. The isolate clusters in clade 2b; all other isolates from the United States belong to clade 2c. Nucleotide sequence alignments revealed that CxFV-Iowa07 and all other members of clade 2b share 19 mutations that are not found in any isolates in clade 2c. These mutations are located in the following nucleotide positions in the CxFV-Iowa genome: 1,108, 1,114, 1,276, 1,297, 1,318, 1,438, 1,534, 1,615, 1,645, 1,692, 1,786, 1,825, 1,840, 2,050, 2,065, 2,096, 2,143, 2,146 and 2,185. Ten of these mutations are also present in all isolates in clade 2a. All isolates in the Latin American/Caribbean clade share 88 mutations that are not found in any isolates in the United States/ Asian clade (data not shown).
We report the isolation and sequence determination of CxFV isolates from Culex spp. mosquitoes collected in Iowa. CxFV also was recently isolated from mosquitoes collected in Japan, Indonesia, Guatemala, Mexico, Trinidad, and Texas (Hoshino et al. 2007, Morales-Betoulle et al. 2008, Farfan-Ale et al. 2009, Kim et al. 2009). In addition, CFAV was recently discovered in natural mosquito populations in Puerto Rico and Thailand (Cook et al. 2006, Kihara et al. 2007), and KRV was isolated from mosquitoes collected in Africa (Crabtree et al. 2003, Sang et al. 2003). Together, these data indicate that insect-specific viruses are common and widely distributed in nature. One explanation as to why CxFV was not identified until recently is because it causes minimal (if any) CPE in mosquito cells and does not replicate in mammalian cells. As a result, the virus could easily remain undetected in arbovirus surveillance studies in which mosquitoes are screened for viruses by suckling mouse brain inoculation or by virus isolation in Vero or C6/36 cells.
The phylogenetic analyses demonstrate that CxFV isolates from Iowa are most closely related to CxFV isolates from Asia and Texas. These findings are consistent with the E gene sequence alignments which reveal that CxFV isolates from Iowa share ≥97.7% identity to isolates from Asia and Texas, but no more than 90.1% identity to those from Guatemala, Mexico, and Trinidad. It is surprising that CxFV isolates from Iowa have a closer genetic and phylogenetic relationship to isolates from the Eastern Hemisphere than to isolates from most of the Western Hemisphere locations mentioned above. One explanation is that CxFV has spread “horizontally” across the globe, possibly due to the displacement of mosquitoes by human activity (i.e., air and sea transportation) (Lounibos 2002). Despite these genetic differences, these isolates still represent a single viral species. According to Kuno and colleagues, a viral species is defined as a group of viruses with >84% nucleotide sequence identity among them (Kuno et al. 1998).
CxFV RNA was detected in mosquitoes collected from July through October but was not detected in any mosquitoes collected in May or June (Table 3). The absence of CxFV RNA in the May and June collections could reflect the lower numbers of Culex spp. mosquitoes collected in these months. Only 8.3% of the total number of Cx. pipiens group was collected in May and June. These findings also could reflect the lower numbers of Culex spp. mosquitoes collected at specific study sites during these months. For example, no Cx. pipiens were collected in Polk County in May. This county yielded the highest overall CxFV MIR when compared with all other counties from which >200 Cx. pipiens group females were collected. In Texas, CxFV was detected in mosquitoes from November to March but not April to August, even though mosquitoes were abundant at these times (Kim et al. 2009). Although this would seem to contradict the apparent transmission cycle of CxFV, these data indicate that CxFV infection could be seasonal in some areas.
Mosquitoes infected with one virus are often refractory or less susceptible to subsequent infection with closely related viruses due to the phenomenon of superinfection exclusion (Davey et al. 1979, Sundin and Beaty 1988). Thus, future research should be performed to determine whether mosquitoes infected with CxFV are resistant to subsequent infection with other flaviviruses. These experiments could lead to the development of novel strategies that block the transmission of medically important flaviviruses.
We thank Brendan Dunphy, Patrick Jennings, and Dan Au for mosquito identification and Erica Hellmich for sample preparation. This study was supported by grant 5R21AI067281-02 from the National Institutes of Health. Mosquito samples used in this research were collected as part of the Iowa Agriculture and Home Economics Experiment Station, Ames, IA, Project 5111, which is supported by the Iowa Department of Public Heath and by Hatch Act and State of Iowa funds.