Every 2 weeks during July 13–September 13, 2009, we collected adult female mosquitoes by using 35 miniature light traps from the US Centers for Disease Control and Prevention (Atlanta, GA, USA) baited with carbon dioxide; the traps were set within the regular mosquito surveillance area of the River Dalälven floodplains in central Sweden (7). Mosquitoes were kept cold on a chilled table during morphologic identification and stored at −80°C. Legs were removed from mosquitoes morphologically identified as Cx. pipiens/Cx. torrentium and used for DNA extraction, enabling identification of the individual specimens to species by using a previously described molecular method (6).

RNA was extracted from 668 mosquito bodies without legs (301 Cx. torrentium, 367 Cx. pipiens) and from 74 pools of mosquitoes pooled by collection trap and week (290 Cs. morsitans); pool sizes ranged from 1 to 19 mosquitoes. The mosquitoes were processed for RNA extraction, real-time reverse transcription PCR (rRT-PCR), and virus isolation on Vero cells as previously described (3).

First, all samples were screened by SINV rRT-PCR by pooling 2–10 RNA extractions by species and collection week. Of 81 total pools, 14 were positive for SINV RNA. Nine positive pools were from Cx. torrentium mosquitoes, and 3 and 2 pools, respectively, were from Cx. pipiens and Cs. morsitans mosquitoes. Second, all individual samples from the SINV-positive pools from the first screening were subjected to another rRT-PCR, so that individual (Culex) or smaller pools of (Culiseta) mosquitoes were ultimately tested for SINV RNA. The second rRT-PCR showed 16 samples positive for SINV RNA. One of the positive Cx. torrentium pools, which contained samples from 10 mosquitoes, included samples from 3 SINV RNA–positive mosquitoes. Thus, 11 of 301 individual Cx. torrentium and 3 of 367 individual Cx. pipiens mosquitoes were positive for SINV RNA. For Cs. morsitans mosquitoes, 2 of 74 pools were positive for SINV RNA (Table 1).

Because individual Cx. torrentium and Cx. pipiens mosquitoes were tested, the IRs were calculated as (no. positive individual mosquitoes/total no. tested) × 1,000; differences between the species were tested for significance by using a χ² test. The actual species-specific IR differed significantly between species (p = 0.01): 36.5, 8.2, and 21 infections/1,000 mosquitoes for Cx. torrentium, Cx. pipiens, and mixed species (Cx. pipiens and Cx. torrentium), respectively. Cs. morsitans mosquitoes were tested in pools, so we calculated the minimum IR (MIR) as (no. positive pools/total no. mosquitoes tested) × 1,000; the calculated MIR was 6.9 infections/1,000 Cs. morsitans mosquitoes. A comparison of MIR and maximum-likelihood estimates of infection gave similar estimates.

Figure

Figure. Consensus tree of the partial E2 envelope glycoprotein gene for Sindbis virus constructed by using MrBayes (http://mrbayes.sourceforge.net/). The phylogram includes 16 Sindbis virus strains isolated from mosquitoes collected in central...

SINV was successfully isolated from all 16 rRT-PCR–positive mosquito samples (Table 2). Part of the SINV E2 envelope glycoprotein gene was sequenced as previously described (8). The phylogenetic analysis included all genomic sequences obtained in this study together with previously published sequences for 63 other virus strains (8). The results showed that all new strains belonged to the SINV-I genotype and have close relationships with strains from Europe, the Middle East, and South Africa (Figure).

We describe information on the actual occurrence and IR of SINV-I in the enzootic mosquito vectors Cx. torrentium and Cx. pipiens, reliably identified to species level. The significantly higher SINV-I IR observed for field-caught Cx. torrentium than Cx. pipiens mosquitoes is a key addition to the previous findings of the extreme susceptibility of Cx. torrentium mosquitoes to SINV-I (4). Experimental transmission studies showed that all infected Cx. torrentium mosquitoes could transmit the virus upon refeeding on a susceptible animal (4). Thus, the observed natural IR of 36.5 infections/1,000 Cx. torrentium mosquitoes translates to 36.0 mosquitoes/1,000 being able to transmit SINV-I. In contrast, because only one third of infected Cx. pipiens mosquitoes can transmit SINV-I upon refeeding, the observed natural IR of 8.2 infections/1,000 Cx. pipiens mosquitoes translates to only 2.0 mosquitoes/1,000 being able to transmit the virus (4).

The observed SINV-I IRs are very high for both species, and the IR for Cx. torrentium is among the highest ever reported for mosquitoes. This could partly be attributed to the fact that single mosquitoes were analyzed, as compared with the more common technique of pooling. This higher IR would have remained undetected if only pooled mosquitoes were analyzed, even though our original pools consisted of only 10 individual mosquitoes. Cx. pipiens mosquitoes are generally considered a secondary enzootic vector of SINV because, as in Sweden, they are less frequently found infected in nature in South Africa, Israel, and Saudi Arabia, where Cx. univittatus mosquitoes are the main SINV vector (9–11).

Because of the exceptionally high SINV IR for field-caught Cx. torrentium mosquitoes in Sweden, the outstanding vector competence results for the species in the laboratory (4), and its status as the dominating Culex species in SINV-endemic areas of Europe (12), Cx. torrentium can now be identified as the main enzootic vector of SINV-I in Scandinavia. In areas of Sweden and Finland where clinical SINV-I infections are most prevalent, Cx. torrentium mosquitoes account for >90% of the Cx. pipiens/Cx. torrentium population; in central Europe, where the virus is more uncommon in mosquitoes and no human cases have been observed, both species are equally common (12–14). Thus, a large population of Cx. torrentium mosquitoes may be a prerequisite for the intense enzootic transmission of SINV-I that is needed to increase the risk for spillover infections in humans. Whether the Cx. torrentium mosquitoes are also to be considered a vector of other mosquitoborne bird viruses remains to be investigated. In continental Europe, West Nile virus and Usutu virus are emerging, and it is unknown if Cx. torrentium has a vector role for these viruses is unknown because of the lack of transmission experiments and isolation attempts from reliably identified female Cx. pipiens/Cx. torrentium mosquitoes. Knowledge from such experiments and isolation attempts would be especially valuable for northern and central Europe where Cx. torrentium is the dominating candidate enzootic vector species for bird-associated viruses (12).