The study was carried out from 1999 to 2000 in the Monte San Miguel Forest in the middle Magdalena Valley of Colombia (6° 23′ 30′′N; 74° 21′ 41′′ W; 50 m elevation). This is a lowland tropical rainforest surrounded by cattle ranches created by deforestation. Mean minimum and maximum daily temperatures are 23°C and 33°C (overall mean of 29°C), respectively, and annual rainfall averages 2,700 mm. Mean relative humidity is 80%. Generally, the peaks of the rainy seasons occur in April–May and October–November. Numerous previous isolations of subtype ID VEEV from sentinel hamsters (9) indicate that this forest site is a stable enzootic focus.

Hamster-baited traps were used for detection of natural VEEV vectors. These traps were a version of the Trinidad No. 10 trap (19) with the following modifications: 1) the metal can comprising the trap opening was replaced by a polyvinyl chloride pipe, 9 cm in diameter; 2) the cylindrical animal cage was enlarged to 11 cm in diameter and 12 cm in height; 3) the roof was constructed from plexiglass; and 4) the opening for mosquito aspiration was a simple buttonhole sewn into the polyester collection net (Figure). The traps were baited with adult golden Syrian hamsters obtained from a colony maintained at the Instituto Nacional de Salud in Bogota. Baited traps were suspended approximately. 1.5 m above the ground and placed in transects at 10-m intervals. Carrots and rat chow were provided for food and water. The traps were checked each morning between 0600 and 0800 h, and some were also checked in the evening between 1700 and 1900 h. Mosquitoes were removed from the traps by using an aspirator, and the daily or semi-daily collections from each trap were frozen as a single pool in a plastic bottle immersed in liquid nitrogen vapor. When hamsters within the traps became moribund or died, serum samples were obtained by cardiac puncture or their hearts were dissected aseptically and frozen for virus isolation.

To confirm VEEV infection in dead or moribund hamsters, virus was isolated from a 10% heart tissue suspension in Eagle’s minimal essential medium (MEM), supplemented with 20% fetal bovine serum (FBS) and antibiotics. The suspension was prepared in a Ten Broeck tissue grinder and centrifuged at 15,000 x g for 5 min; 200 μL of the supernatant was added to a 25-cm² flask containing a monolayer of Vero cells and adsorbed for 1 h at 37°C; 6 mL of additional MEM containing 2% FBS was then added. Cultures were incubated at 37°C for 5 days or until cytopathic effects were evident.

Mosquito pools from traps in which hamster infection with VEEV was confirmed were assayed for infectious virus. Pools containing 1–40 individuals of each mosquito species were triturated with a Minibeadbeater (BioSpec Products, Inc., Bartlesville, OK) or a Ten Broeck tissue grinder containing 1.0 mL of MEM supplemented with 20% FBS, penicillin, streptomycin, and amphotericin B. The triturated pool was centrifuged for 5 min at 15,000 x g, and 200 µl of the supernatant was added to a 10-mL plastic tube or a 25-cm² cell culture dish containing a monolayer of Vero cells and 2–5 mL of MEM. Cultures were monitored for cytopathic effects for 5 days.

Viruses isolated from hamster heart tissue suspensions and mosquito pools were characterized antigenically by using immunofluorescence of infected cells and a panel of monoclonal antibodies described previously (20). Subtype ID VEEV isolates were further characterized genetically by reverse transcription–polymerase chain reaction (PCR) amplification of an 856-nucleotide portion of the PE2 (sometimes called p62) envelope glycoprotein precursor gene as described previously (8), followed by single-stranded conformation polymorphism (SSCP) or sequence analysis (9). For SSCP analysis, PCR products were purified on agarose gel by using the QIAquick Gel Extraction Kit (QIAGEN, Valencia, CA). A 2-μl volume of the PCR amplicon DNA suspension was mixed with 8 μl of SSCP loading buffer (95% formamide, 0.05% bromophenol blue, and 0.05% xylene cyanol). The DNA was heated to 95°C for 5 minutes, rapidly cooled on ice, loaded onto an 8% polyacrylamide gel, and underwent electrophoresis in 1X Tris-borate EDTA buffer at room temperature for 20 h at 8 mA. Single-stranded DNA products were visualized by using silver staining (21). SSCP patterns were compared by measuring the migration of single-stranded DNA of the various isolates in comparison to one another and to a standard DNA ladder.

Previous studies of VEE complex enzootic transmission have each identified a single, principal mosquito species in a given geographic region. All of these species, including Cx. portesi (14), Cx. cedecei (15), Cx. aikenii sensu lato (ocossa and panocossa) (16,17), and Cx. taeniopus (18) are members of the Spissipes section of the subgenus Melanoconion within the genus Culex (37). Previous studies of enzootic VEEV transmission in the Catatumbo region of northeastern Colombia also suggested that Cx. pedroi might be the principal vector, based on abundance in active foci (38). Cx. vomerifer from Iquitos, Peru, also has been shown to be susceptible to infection by several strains of VEEV (28), but was only tested after mosquitoes ingested 8 log₁₀ PFU/mL from viremic hamsters, a viremia titer at least 100 times greater than that generated by experimentally infected rodent reservoir hosts (30,31). Our findings of at least three enzootic vectors of subtype ID VEEV in Colombia contrast with the findings of all previous studies of enzootic VEEV vectors, which suggested that enzootic VEEV strains are each adapted to a single, principal vector species (13,18,39–41). In Colombia, subtype ID VEEV appears to utilize efficiently both Cx. vomerifer and Cx. pedroi in the Magdalena Valley. Cx. adamesi, which is usually less abundant in the Monte San Miguel Forest, appears to serve as a secondary vector.

All three of the mosquito species that we identified as VEEV vectors are members of the Spissipes section of the subgenus Culex (Melanoconion), bringing the total to seven confirmed vectors within this section of closely related mosquitoes. The genetic or ecologic basis for the exclusive use of these mosquitoes by enzootic VEE complex viruses deserves further study. Hypotheses to explain this phenomenon include possible shared, derived characteristics of the Spissipes section, such as particularly high susceptibility to infection by enzootic VEE complex viruses, a particularly high degree of association with the Proechimys spp (35). and other small mammalian reservoir hosts (13), or both. Mosquito longevity and population sizes in habitats that support large populations of reservoir hosts may also favor transmission by members of the Spissipes section (35).

Identification of the principal enzootic vectors (Cx. vomerifer and Cx. pedroi) of subtype ID VEEV strains believed to be closely related to epizootic progenitors will allow us to assess the role of these mosquitoes in the generation of mutations that mediate VEE emergence by enhancing equine viremia and infection of epizootic mosquito vectors such as Ae. taeniorhynchus. The hypothesis that epizootic VEEV is not recovered from sylvatic foci because these strains lose their fitness for the enzootic vectors (25) can also be tested in the two principal vectors that we identified.