Figure 1

Figure 1. MosquiTRAP version 2.0 (Ecovec Ltd., Belo Horizonte, Brazil) (A), BG-Sentinel trap (Biogents, Regensburg, Germany) (B), and an ovitrap (C) used for obtaining mosquitoes in the northwestern borough of Belo Horizonte, Minas...

Traps designed to catch mosquitoes and eggs were installed in an urban residential area in the northwestern borough of Belo Horizonte, Minas Gerais, Brazil. Since 1998, Belo Horizonte has had a high concentration of dengue cases and high rates of vector infestation. Mosquitoes were obtained during 8 weeks (during October 2005–May 2006). Adult mosquitoes were collected by using 2 capture traps (Figure 1, panels A and B): MosquiTRAP version 2.0 (M trap; Ecovec Ltd., Belo Horizonte, Brazil) (7) and the BG-Sentinel trap (BG trap; Biogents, Regensburg, Germany) (8). Eggs were collected with an ovitrap (Figure 1, panel C) and hatched into larvae (9).

An area of 20 blocks was selected for the survey. In each block, 3 representative houses were selected, and each house received 1 type of trap. All traps were installed outdoors in a shaded area that was also protected from rain. Within each block, traps were rotated so that each house had only 1 trap type for a maximum of 1 week. Mosquitoes collected were identified by sex, trap, and epidemiologic week. Eggs collected were counted, identified by epidemiologic week, and hatched into larvae. All samples were stored at –70°C until analyzed (Table 1).

Pools of <30 mosquitoes and <50 larvae were triturated on ice, macerated in 300 µL of Leibowitz L15 medium (GIBCO-BRL, Gaithersburg, MD, USA), and centrifuged at 2,000 × g for 5 min. RNA from each pool was extracted according to a modified protocol (10) and used as template in a reverse transcription–PCR as described (11). To amplify the virus genome, a seminested PCR was conducted with a forward primer (5′-CGA GAA ACC GCG TGT CAA C-3′) designed to amplify a 434-nt region that contains the capsid–premembrane (C-prM) gene and a reverse primer as described (11). The PCR products were then used for sequencing (MegaBACE sequencer; GE Healthcare, Little Chalfont, UK). Sequences and inferred amino acid sequences were aligned with other available DENV-3 sequences.

Alignments were used to construct midpoint-rooted phylogenetic trees by using the neighbor-joining method. The Tamura-Nei statistical model was implemented by using MEGA4.1 software (Arizona State University, Tempe, AZ, USA) with 1,000 bootstrap replicates. These sequences have been deposited in GenBank (accession nos. GQ330909–11 and GU588695–8).

A total of 237 adult A. aegypti mosquitoes (137 females and 100 males) were tested in 25 pools, and 5,573 larvae were tested in 101 pools (Table 1). Fifteen mosquito pools contained only females, and 10 contained only males. Fifteen pools (8 containing females and 7 containing males) were obtained in BG traps, and 10 pools (7 containing females and 3 containing males) were obtained in M traps. Viral RNA was detected in 4 of the 25 A. aegypti mosquito pools analyzed. Two positive pools were obtained in M traps, and 2 positive pools were obtained in BG traps. Viral RNA was detected in 1 of 101 larvae pools tested (Table 1). Minimum infection rates were 16.9% for adult mosquitoes and 0.18% for larvae (Table 2). The higher minimum infection rates for mosquitoes could be explained by the lower number of analyzed specimens.

Figure 2

Figure 2. Phylogenetic tree of dengue type 3 serotypes and sequences from Aedes aegypti mosquitoes and larvae obtained in Belo Horizonte, Minas Gerais, Brazil. The tree is based on a 434-nt sequence of...

These data also confirm vertical transmission by detection of DENV-3 in male mosquitoes hatched from eggs. C-prM sequences obtained showed a high degree of similarity to isolates from the Philippines and the People’s Republic of China and recent isolates from Minas Gerais and Rondônia, Brazil (6,12). Sequence homology values ranged from 98.8% to 99.8%. Phylogenetic analysis confirmed that all 4 samples obtained from adult mosquitoes (female mosquito BH-14/2005, male mosquito BH-1/2006, female mosquito BH-15/2006, and female mosquito BH-17/2006) and larvae (BH-11/2006) were grouped in a well-supported distinct cluster of genotype I (Figure 2).

Our results suggest that the DENV-3 strains circulating in Minas Gerais, Brazil, in 2005 and 2006 may share a common origin. We identified (6) co-circulation of 2 DENV-3 genotypes (I and III) in Brazil (6,12). DENV-3 isolates detected in Rio de Janeiro, Brazil, during the 2001 and 2002 outbreaks and DENV-3 detected in Latin America were assigned to genotype III, which has been associated with DHF outbreaks (13). DENV-3 genotype I identified in outbreaks during 2002–2004 in Minas Gerais (6) also showed associations with DHF. This genotype was also identified in Colombia (14) and French Guiana (14). Sequences showed high similarity to isolates from Southeast Asia and the South Pacific islands and have not been previously reported in South America.

Detection of DENV-3 genotype I in sylvatic animals (15) may support the hypothesis of a sylvatic origin for this genotype in South America. Our analysis of C-prM gene sequences from mosquitoes naturally infected with DENV-3 confirmed circulation of genotype I in Minas Gerais. Additionally, our results indicated that dengue virus sequences in mosquitoes and larvae are highly similar to sequences in DENV-3 isolates from patients who are spatially and temporally related. Although vertical transmission has not been routinely determined (2), our findings also confirmed that vertical transmission of DENV that may be a major factor in virus prevalence and survival in nature.