Arboviruses can have benign, deleterious, or beneficial effects on the vector. We tested the hypothesis that oral infection with La Crosse virus (LACV) will have little to no effect on mosquito longevity and fecundity, a prediction of low virulence selected in a system with frequent vertical transmission. We tested the effects of infection in native Ochlerotatus triseriatus Say and invasive Stegomyia albopicta Skuse (Diptera: Culicidae). We artificially fed adult female mosquitoes of each species with either LACV-infected or uninfected bovine blood and determined adult longevity and fecundity. For females fed LACV-infected blood, bodies and legs, respectively, were separately homogenized and assayed by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) to determine the LACV infection and dissemination rates. Ochlerotatus triseriatus had a higher infection and dissemination rate than St. albopicta. For both species, female size had no effect on infection status. Infection status also had no effect on longevity or fecundity for both species. We suggest that the high frequency of vertical transmission may have selected for strains of the virus with low virulence in two vectors, in spite of their different evolutionary histories with the virus.
Vector-borne diseases present a major challenge to wildlife, domestic animal and public health globally, and extensive efforts have been devoted towards a better understanding of the dynamics of their transmission (Gratz, 1999). Arthropod-borne (arbo) viruses are among the most diverse and numerous vector-borne disease-causing agents (Gubler, 1998), and a large proportion of the world's human population is at risk of infection. Understanding how arboviruses interact with their vectors can provide insights into the transmission dynamics and evolution of virulence both of which are essential for the development of effective and sustainable vector control strategies. Historically, it was thought that natural selection should favour arboviruses that have little to no impact on the vector (Burnet & White, 1972). Although this is supported in some systems, many arboviral infections have deleterious effects on the vectors that transmit them (Rennie et al., 2000; Bennett et al., 2008; Lambrechts & Scott, 2009).
In mosquito vector systems, parasitic infections have adverse effects on larval development as well as adult fecundity, survival and blood-feeding behaviour (Turell et al., 1982, 1985; Faran et al., 1987; Scott & Lorenz, 1998; Moncayo et al., 2000; Mahmood et al., 2004; Barnard et al., 2007; Styer et al., 2007; Marciel-de-Freitas et al., 2011). Overall, parasite infection generally reduces adult mosquito survival, although the magnitude of these effects varies depending on experimental conditions, taxonomic relationships of viruses and mosquitoes, as well as the mode of infection, with only horizontal infections reducing adult longevity (Ferguson & Read, 2002; Lambrechts & Scott, 2009).
La Crosse virus (LACV; family Bunyaviridae: genus Bunyavirus) is a member of the California serogroup viruses with substantial socioeconomic impact in the United States. LACV is one of the leading causes of childhood encephalitis in the U.S. and La Crosse encephalitis is an emerging disease in the Appalachian region (Beaty & Marquardt, 1996; Haddow & Odoi, 2009; Leisnham & Juliano, 2012). LACV is transmitted primarily by the eastern treehole mosquito, Ochlerotatus (formerly Aedes) triseriatus Say, Oc. hendersoni Cockerell and secondarily by the invasive Asian tiger mosquito Stegomyia (formerly Aedes) albopicta Skuse (Watt et al., 1973; Paulson et al., 1989; Gerhardt et al., 2001; Hughes et al., 2006). LACV is one of the few arboviruses that can be transmitted both horizontally (from vertebrate host to mosquito and mosquito to mosquito through mating) and vertically (from mother to offspring). The transovarial transmission of LACV was one of the first documented cases of this phenomenon, and is frequent and important in the transmission and maintenance of LACV in nature (Miller et al., 1977; Hughes et al., 2006). This mode of transmission is an overwintering mechanism for LACV and mosquito offspring infected with LACV vertically from their mothers can act as reservoir hosts and transmit the virus to a vertebrate host during their first bloodmeal (Mullen & Durden, 2002). Thus, any fitness consequences LACV may have on the mosquito vector such as reduced longevity and fecundity would directly affect both horizontal and vertical transmission rates of the virus. For these reasons, both theory and empirical data suggest that parasites transmitted both vertically and horizontally are selected for decreased virulence on the vector host than those that are solely or predominately transmitted horizontally (Ewald, 1987; Ebert & Herre, 1996; Lipsitch et al., 1996), and LACV can act as a model system to test these predictions.
Previous studies have shown neutral, beneficial or deleterious effects of LACV on fitness traits in Oc. triseriatus. Transovarial infection in Oc. triseriatus did not affect immature development time, fecundity, survival, hatching success and sex ratio (Patrician & Defoliart, 1985). However, eggs infected with LACV had higher mortality during overwintering and terminated diapause earlier than uninfected eggs (McGaw et al., 1998). Benefits to Oc. triseriatus infected with LACV have been found including higher insemination rates of transovarially infected females compared with uninfected females (Gabitzsch et al., 2006; Reese et al., 2009). Possible benefits that may increase LACV transmission include female Oc. triseriatus transovarially infected with LACV probed more during blood feeding and engorged less blood than uninfected Oc. triseriatus (Grimstad et al., 1980).
Although the effects of transovarially infected LACV on Oc. triseriatus have been investigated in many previous studies, the experimental mode of infection (horizontal versus vertical) can alter the effects the arbovirus has on the mosquito vector (Lambrechts & Scott, 2009). It has been suggested that oral infection should result in more costs to the vector and therefore may result in more deleterious effects on the vector hosts (Tesh & Shroyer, 1980; Lambrechts & Scott, 2009). As most studies investigating the costs associated with LACV infection focused on the effects of transovarial infection, it is difficult to determine if the benign or beneficial results are as a result of the experimental system or low virulence of the virus in the vector, which should be selected for in a system with frequent vertical transmission.
In this study, we determined the effects of oral infections of LACV on adult longevity and fecundity of Oc. triseriatus and St. albopicta. LACV has been isolated in the upper Midwest and southeastern regions of the United States, occupied by high densities of Oc. triseriatus and St. albopicta, respectively (Jones et al., 1999]. Ochlerotatus triseriatus is the primary vector of LACV, but both species are able to transmit the virus horizontally and vertically (Watt et al., 1973; Tesh & Gubler, 1975; Hughes et al., 2006). The historic and current distribution of Oc. triseriatus overlaps with the confirmed California encephalitis cases in humans in the U.S., most of which are predominately in the upper Midwest (Paulson et al., 1989). Stegomyia albopicta was first recorded in the Midwestern U.S. in the mid 1980s but only became more prevalent in those regions through the mid-1990s (Moore & Mitchell, 1997), giving it a relatively shorter exposure to LACV compared with Oc. triseriatus. Here we tested the hypothesis that oral infection of LACV will have little or no effect on mosquito longevity and fecundity, a prediction of low virulence resulting from selection in a system with frequent vertical transmission. We also addressed this hypothesis in the context of evolutionary histories between LACV and the two mosquito species and predicted that the severity of LACV-induced effects on mosquito fitness traits will be greater in St. albopicta than in Oc. triseriatus owing to its shorter evolutionary history with LACV.
Materials and methods
Larval and adult rearing
Experiments were conducted using an F6 progeny of field-collected St. albopicta larvae from discarded tyres in St. Louis, MO, and F5 progeny of field-collected Oc. triseriatus from Illinois. All samples were collected from public lands. Immature stages of each species were reared separately in enamel-rearing pans containing 1.5 L of deionized (DI) water. After hatching, 200 (∼24 h old) first instar larvae of either St. albopicta or Oc. triseriatus were placed into rearing pans with 10 pans total for each species, 5 replicate pans that would be fed LACV infected or uninfected blood (Control) as adults, respectively. On the first day of the experiment, each pan received larvae and 100 mg of a 1 : 1 ratio of lactalbumin-brewer's yeast for resources. On days 4 and 7, all larvae were filtered from the water and added to fresh pans with DI water to prevent waste accumulation, and each pan received 60 mg of crushed Tetramin for additional resources. The pans were housed in an environmental chamber at 25 °C with a L;D 16 ; 8 h photoperiod.
Each day, pupae from each replicate were transferred into vials with DI water until ecolosion. Newly eclosed adults were transferred daily into 1-L paperboard cages with mesh screening. Females from the same species, blood-feeding treatment and replicate that emerged on the same day were housed in the same paperboard cage with up to 15 females per cage. To achieve 1 : 1 sex ratios in the adult cages, males that emerged within 5 days of females from the same species, blood treatment and replicate were housed in the same paperboard cage with up to 30 adults total. Adults were housed at 25 °C with a L;D 16 ; 8 h with continuous access to 10% sucrose solution. Seven-day-old adult females were allowed access to a bloodmeal.
Preparation of infectious bloodmeals
In order to prepare the infected blood for feeding trials, tissue culture flasks (25 cm2) with confluent monolayers of baby hamster kidney cells were inoculated with 250 μL media (Leibovitz L-15 media, 10% fetal bovine serum, 1% penicillin/streptomycin) containing LACV at a multiplicity of infection of 0.01 viruses per cell. LACV virus inoculum was incubated for 1 h at 37 °C to allow virus attachment and entry after which media was added to a total volume of 10 mL/flask and the flasks held at 37 °C for 48 h. LACV-infected bloodmeals consisted of a 1 : 1 mixture of citrated bovine blood and freshly harvested media from LACV infected cell cultures. For LACV-uninfected treatments (Control) a 1 : 1 mixture of citrated bovine blood and uninfected media was prepared. Seven day old adults were allowed to blood feed for 45 min on citrated bovine blood that was either infected (5.7 ± 0.5 log10 plaque forming unit equivalents LACV/mL) or uninfected (Control) with LACV (LACV/human/1960, GenBank accession numbers EF485030-EF485032), delivered through an artificial membrane feeder (Hemotek®, Lancashire, U.K.). The virus titre of the infected blood in our experiment was in the range of those used in other laboratory studies of infection of LACV with Oc. triseriatus (Paulson et al., 1989; Gabitzsch et al., 2006) but somewhat higher than LACV titres found in two primary vertebrate hosts of La Crosse in nature, the sentinel gray squirrel (Sciuris carolinensis) and chipmunk (Tamias striatus), (1.0–3.39 log10 PFU's/0.2 mL), (Ksiazek & Yuill, 1977). If females failed to blood feed on the first attempt, we provided an additional opportunity to blood-feed on day 10 after emergence using the same procedures as above. On each blood-feeding day, we preserved LACV-infected blood samples at −80 °C to determine the titre of the bloodmeals.
After blood feeding, all mosquitoes were cold anaesthetized at 4 °C, and fully engorged blood-fed females were placed individually into individual 0.5-L paperboard cages with mesh screening and a 30-mL plastic cup lined with moist paper towel for an oviposition substrate. The blood-fed females had access to 10% sucrose solution and were maintained for 14 days. Females fed infected blood that survived this period were sacrificed on day 14 post-blood feeding, immediately placed in individual vials and stored at −80 °C for later analysis of infection status. Females fed uninfected blood were also sacrificed on day 14 but stored in freezer at −25 °C along with oviposition cups from all females. Eggs were counted for each female to determine fecundity.
We used a subset of 50 females from each species treatment combination to investigate the treatment effects on adult longevity. We randomly selected five females from each replicate (fed both infected or uninfected blood) to live beyond the 14-day extrinsic incubation period. Each day, females were checked daily for mortality and any dead adults were immediately preserved in either −80 or 0 °C, depending on the blood-feeding treatment. From any females that laid eggs, (Oc. triseriatus-infected treatments n = 52, uninfected treatments n = 54; St. albopicta-infected treatments n = 139, uninfected treatments n = 97), all eggs were stored and counted from oviposition cups.
Samples were retrieved from −80 °C, and all fully engorged blood-fed adult females were dissected to remove wings, and legs were dissected from females that were fed LACV-infected blood. The wings were mounted onto slides and measured to determine wing length, a proxy for adult female size. The legs and bodies of females fed infected blood were separately homogenized initially in 0.2 mL of media, with a subsequent addition of 0.8 mL of media. RNA extractions were performed on the homogenate using QIAamp® virus BioRobot® 9604 kits according to the manufacturer's instructions (Qiagen, Valencia, CA, U.S.A.).
Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) was used to detect LACV RNA in the body and legs in order to determine infection and disseminated infections, respectively. qRT-PCR was first conducted on individual body homogenates and later on the legs of samples that had LACV-positive bodies. These small (S) segment of LACV was amplified using the following primers and probe: Forward (5′-TCTGTGACATCCACTTAT-3′, Reverse 3′-AAACTTTCAAATTCTACCC-5′), probe (ATTCACAGAGTCAAGCAAGGCA). Each polymerase chain reaction was 50 μL in total and consisted of 5 μL of template RNA, 25 μL of one-step RT-PCR Master Mix, 1.25 μL of RNAase inhibiter, 0.5 μL of each primer and probe, and 17.25 μL of DI water. The PCR conditions on the thermocycler were 48 °C for 30 min, 95 °C for 10 min, and 40 cycles of 95 °C for 15 s, 59 °C for 1 min, and 72 °C for 1 min. Mosquitoes with infected bodies and legs represent disseminated infections, whereas mosquitoes with infected bodies and uninfected legs represent non-disseminated infections.
A Pearson's chi-square test was performed on infection status by species to determine the relative susceptibility of each species to LACV infection and disseminated infections. For each species, two separate logistic regressions were performed to determine the relationship between size and infection status while controlling for random replicate effects (PROC GLIMMIX, SAS 9.1; SAS Inc., Cary, NC, U.S.A.). A continuous variable was run against the probability of LACV infection (infected = 1, uninfected = 0) and infection state (disseminated infection = 1, non-disseminated infection = 0) for each species.
anova was also used to investigate the effect of LACV exposure and infection status (control, undetected but fed infected blood, non-disseminated infections, disseminated infections) on fecundity and longevity of the two mosquito species. Follow-up post hoc tests with a Tukey's correction were run to detect any pair-wise differences by species and treatment. A non-parametric survival analysis (PROC LIFETEST, SAS 9.1; SAS Inc.) was further used to investigate the effects of species and LACV exposure and infection status on adult female survival probability. A log-log transformation was applied to the data set for survival analysis. Follow-up pair-wise contrasts of means were adjusted for multiple comparisons using the Sidak method (Sidak, 1967).
Standardized linear regressions were run for each species to determine whether LACV body titre was a good predictor of female fecundity and longevity (PROC GLM, SAS 9.1; SAS Inc.).
Overall, blood-feeding rates of Oc. triseriatus with the artificial membrane feeder were lower compared with St. albopicta. Ochlerotatus triseriatus had higher infection and dissemination rates than St. albopicta (χ2 = 137.7, P < 0.001). LACV infection and dissemination rates for Oc. triseriatus (n = 80) were 66.2% and 45% respectively, and the corresponding values for St. albopicta (n = 223) were 42.2% and 18.4%. For both species, female size did not affect infection or dissemination of either St. albopicta or Oc. triseriatus (Table 1).
Table 1. Logistic regression testing the relationship between adult size and infection and dissemination in Stegomyia albopicta and Ochlerotatus triseriatus
Species and infection status
Infected females are compared against uninfected females, while disseminated infected females are compared against non-disseminated infected females.
St. albopicta infection
St. albopicta dissemination
Oc. triseriatus infection
Oc. triseriatus dissemination
The anova on fecundity indicates that there is a significant effect of species on fecundity, whereas infection status and species × infection status was not significant (Table 2, Fig. 1). Stegomyia albopicta had higher fecundity (41.1 ± 1.9 eggs/female) than Oc. triseriatus (35.6 ± 1.6).
Table 2. anova results of species, infection status and the interaction on fecundity
Species × treatment
Non-parametric survival analyses indicated significant effects of species and species × treatment interaction on survivor function estimates (Table 3). Comparisons of survival distributions showed significantly steeper declines in the survivor function estimates of Oc. triseriatus than St. albopicta (Fig. 2). Within a species (Oc. triseriatus or St. albopicta), there were no significant differences in survival function estimates among mosquitoes exposed to LACV-infected or uninfected blood and different infection statuses (uninfected, non-disseminated and disseminated infections) (Fig. 2). Rather, the species × treatment interaction was attributable to comparisons of mosquito survival function estimates between species. Specifically, Oc. triseriatus with non-disseminated and disseminated LACV infections had significantly steeper declines in the survival function estimates than St. albopicta individuals exposed to uninfected blood. All other comparisons were not significant.
Table 3. PROC LIFETEST results of species, infection status and the interaction on longevity
Species × treatment
Linear regression analyses indicated no significant relationship between LACV body titre and fecundity for both mosquito species (St. albopicta; r2 = 0.016 F1,3 = 0.05, P = 0.8370; Oc. triseriatus; r2 = 0.06, F1,3 = 0.4, P = 0.55; all t < 7.26, d.f. = 1, P > 0.05). There also was no significant relationship between LACV body titre and adult female survival for both species (St. albopicta; r2 = 0.78, F1,3 = 8.30, P = 0.2127; Oc. triseriatus; r2 = 0.074, F1,3 = 0.48, P = 0.5124; all t < 18.14, d.f. = 1, P > 0.05).
A vector-borne parasite should be less virulent to their vectors than to their main host in order to succeed in its spread (Ewald, 1994; Elliot et al., 2003). However, the magnitude of the effects on the vector host can vary, and the trade-off between transmission and virulence can select for benign, slightly deleterious, or advantageous effects on the vector (Anderson & May, 1982; Frank, 1996; Levin, 1996). In our study we found that oral infection of LACV had no adverse effects on fecundity and survival of both Oc. triseriatus and St. albopicta.
Across many mosquito-vector systems, the virulence of the parasite on the vector depends on many factors including spatial heterogeneity, mobility of vector and veretebrate hosts, parasite competition, environmental quality, parasite load and strain, vector and parasite taxonomy, and mode of infection (Patrician & Defoliart, 1985; Scott & Lorenz, 1998; Elliot et al., 2003; Mahmood et al., 2004; Alto et al., 2005, 2008a; Lambrechts & Scott, 2009). A meta-analysis suggests that infection of arboviruses of the genus Bunyavirus (which includes La Crosse) have little to no effects on survival of Aedes mosquitoes. However, the experimental mode of infection altered the effects of an arbovirus on mosquito host with horizontal (oral or intrathoracic inoculation) but not vertical (transovarial) infection significantly reducing mosquito survival (Lambrechts & Scott, 2009). The experimental mode of infection has been suggested to affect the mosquito host differently, with horizontally transmitted infections posing more tissue damage, change in blood composition and higher viral load (Tesh & Shroyer, 1980). However, most studies investigating the fitness costs of Bunyavirus infection on Oc. triseriatus relied on transovarial infection making it difficult to disentangle the effects of vector, parasite taxonomy and mode of infection on virulence in the vector (Patrician & Defoliart, 1985; Lambrechts & Scott, 2009). In previous studies, no effects of transovarial infection were found on Oc. triseriatus across all life stages (Patrician & Defoliart, 1985). Our study also illustrates no adverse effects of LACV infection on Oc. triseriatus or St. albopicta when orally infected, indicating similar effects of the virus on the vector regardless of infection mode. Thus, it appears that in this particular genus of arbovirus, a less virulent strain may have been selected for in the vectors. However, LACV effects on fecundity and adult survival were measured in a laboratory environment, so it is unclear whether LACV has detrimental effects on these traits in nature or whether other traits, such as behaviour, may be adversely affected by LACV.
Transmission cycles that are primarily maintained vertically in nature should select for lower virulence of the pathogen because deleterious effects on the vector fitness (such as longevity and fecundity) directly affect transmission rates (Ewald, 1987; Ebert & Herre, 1996; Lipsitch et al., 1996). Overall, members of the genus Bunyavirus that have benign effects on mosquito survival (Lambrechts & Scott, 2009) also tend to have relatively higher vertical transmission rates (Turell, 1988). For LACV, vertical transmission is very frequent for Oc. triseriatus (71%) and St. albopicta (52%), (Miller et al., 1977; Hughes et al., 2006), is an important mechanism for its maintenance in nature, and thus is likely a strong selective agent. In the present study, LACV infection had no significant effect on survival and fecundity of St. albopicta and Oc. triseriatus and similar findings have been reported for Oc. triseriatus (Patrician & Defoliart, 1985). By contrast, LACV-infected Oc. triseriatus eggs suffered greater mortality during diapause than uninfected eggs, but the hatching rate of viable eggs was increased in infected eggs (McGaw et al., 1998). In the present study, the presence of delayed effects of virus infection on fertility was not assessed, which could be inferred if hatchability differed among LACV-infected and uninfected mosquitoes. Other studies found that Oc. triseriatus infected with LACV were compromised in their ability to blood feed (Grimstad et al., 1980). It is possible these traits experience weaker selection relative to survival and fecundity, which may have greater relative effects on transmission of LACV. Although there are some reports of negative effects of arboviruses on the mosquito vectors in cycles where vertical transmission is common (Tesh & Shroyer, 1980; Turell et al., 1982), viruses that are mostly or exclusively horizontally transmitted more frequently experience negative effects on fitness parameters of the mosquito vector including survival and fecundity (Faran et al., 1987; Scott & Lorenz, 1998; Mahmood et al., 2004; Suchman et al., 2006; Styer et al., 2007; Lambrechts & Scott, 2009; Reiskind et al., 2010). Our results and previous studies suggest that it is the high frequency of vertical transmission of LACV that has selected for lower virulence in vector hosts, rather than effects of the experimental mode of transmission or the genus of mosquito Aedes, which are affected by horizontally transmitted viruses (Suchman et al., 2006; Reiskind et al., 2010; Marciel-de-Freitas et al., 2011).
Overall, Oc. triseriatus had higher infection and dissemination rates than St. albopicta, consistent with other studies (Hughes et al., 2006). These differences in dissemination rates suggest St. albopicta has a significant midgut escape barrier and further confirms Oc. triseriatus as the primary vector of LACV where this species is abundant.
Interestingly, there was no effect of female size on infection status. Size has been considered an important factor in vectorial capacity and can be used in models to predict transmission patterns. It can directly affect both the population density through its positive relationship on fecundity, it affects blood-feeding behaviour (Klowden, 1995), female mosquito life span (Nasci, 1991) and it may possibly affect intrinsic factors within the mosquito that affect vector competence (Grimstad & Walker, 1991).
Many factors affect mosquito adult size such as environmental conditions, viral strain and genetics; and how these factors affect the relationship between size and arbovirus transmission may vary depending on the source of variation (Dye, 1992). Size is often associated with infection when adult size differences are environmentally induced, although the patterns are not consistent. Smaller Oc. triseriatus adults from resource-deprived larval conditions were more likely to become infected and transmit LACV (Grimstad & Haramis, 1984; Patrician & DeFoliart, 1985; Grimstad & Walker, 1991; Paulson & Hawley, 1991). However, when reared under larval competitive conditions with St. albopicta, larger Oc. triseriatus were more likely to become infected and disseminated with LACV virus (Bevins, 2008). For St. albopicta, some studies found no relationship between size and Sindbis virus (SINV) infection status (Muturi et al., 2011) while others found a negative relationship between adult mosquito size and infection and dissemination rates for SINV and dengue-2 viruses (Alto et al., 2005, 2008a, 2008b). In this study, the size differences were not environmentally induced as all larvae were reared under identical conditions. Therefore, the size differences were likely genetic, which is possibly why differences were not detected with size and infection status.
Different effects of LACV infection on survival and fecundity in Oc. triseriatus and St. albopicta were expected owing to their different evolutionary histories with the arbovirus. Other arbovirus infections including Eastern Equine Encephalomyelitis (EEE) and chikungunya virus illustrate similar or different effects of the virus on native vs. invasive species (Moncayo et al., 2000; Reiskind et al., 2010). However, as vertical transmission of LACV is very frequent in both species (Miller et al., 1977; Hughes et al., 2006), it probably is a very strong selective agent on the virulence of the arbovirus towards the vector. The benign nature of this arbovirus in the traits we measured to two species with different evolutionary histories may be an adaptation the virus has with its vectors owing to overall frequent vertical transmission. Our findings in this model system may be applicable to other arboviruses where vertical transmission plays an important role in their maintenance in nature such as Rift Valley Fever (Mullen & Durden, 2002).
In conclusion, the lack of adverse effects of LACV infection on longevity and fecundity in both species suggests a low virulence of LACV on the vector host. Although other factors have been suggested (Lambrechts & Scott, 2009), our results and others suggest that LACV probably has been selected for low virulence with its vectors because of the high frequency of vertical transmission. Vertical transmission is probably a strong selective agent as a general low virulence has evolved and is exhibited even in St. albopicta, which shares a relatively short evolutionary relationship with the virus.
The authors thank S.A. Juliano for providing us with Aedes triseriatus and Stegomyia albopicta eggs to initiate laboratory colonies of mosquitoes. La Crosse virus was kindly provided by the Centers for Disease Control and Prevention, Diagnostic and Reference Laboratory, Arbovirus Diseases Branch. We thank Nina Krasavin and Margarita Dmitrieva for their assistance with the RNA extractions and David Bettinardi for assistance with the daily maintenance of the experiment. We thank C.T. Smartt for her comments on the manuscript. This study was supported by Illinois Waste Tire and Emergency Public Health Funds.