In Florida, SLEV transmission is sporadic with periods of low (enzootic) and high (epidemic) activity. SLEV was detected by sentinel chickens every year before introduction of WNV (1988–2007) (Figure). Since 2001, limited SLEV activity has been reported (10); SLEV may be in a natural decline, or transmission of WNV may influence SLEV cycles, as has been suggested in California (4).

In 2006, a total of 2,901 adult sentinel chickens were maintained at 275 sites of potential enzootic arbovirus transmission in 34 Florida counties. Blood was collected weekly from each chicken during peak transmission months (July–December) and tested with hemagglutination inhibition assay, immunoglobulin M antibody-capture ELISA, or plaque reduction neutralization test, as previously described (11). Sites with confirmed SLEV seroconversions were targeted for sample collection. For the first time since 2001, SLEV sentinel seroconversions (n = 40) exceeded WNV seroconversions (n = 30) (10).

In central and south Florida, 5 partner agencies targeted a subset (n = 15) of sentinel chicken sites with recent confirmed arbovirus transmission activity for cloacal swab collection from 95 chickens. During the weekly scheduled bleeding of the flocks, 1,338 cloacal swabs were collected in viral culturettes (Becton Dickinson, Franklin Lakes, NJ, USA); 529 swabs were retrospectively processed for molecular detection assays and virus isolation in Vero cells, as previously described (12). Viral RNA was extracted from cloacal swabs and first-passage cell cultures and amplified with real-time reverse transcription–PCR (RT-PCR) TaqMan assays for WNV and SLEV, as previously described (13). Two SLEV strains, FL06-S569 and FL06-S650, were detected by RT-PCR and cultured in Vero cells. Fourteen additional SLEV strains were obtained from the Florida Department of Health, Bureau of Laboratories–Tampa archive for phylogenetic analysis (Table).

To characterize SLEV strains, we amplified the envelope region using previously described primers (9) and the SuperScript III 1-step RT-PCR system (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions. Sequences were aligned by using ClustalW 1.6 and phylogenetic trees drawn by using the maximum parsimony method, with 1,000 bootstrap replicates, in MEGA 4.0 software (14), including 60 other SLEV envelope sequences available in GenBank (9,15) and 4 related flavivirus outgroups (accession nos.: WNV NY99, AF196835; Japanese encephalitis virus, EF571853; Kunjin virus, AY274505; Murray Valley encephalitis virus, AF161266).

The phylogenetic tree places FL06-S569 and FL06-S650 into genotype VA (Appendix Figure). This analysis further supports classification of SLEV into 7 lineages and 13 clades (IA-IB, IIA-IIE, III, IV, VA-VB, VI, VII), as previously suggested (9). FL06-S569 and FL06-S650 share 98% sequence identity with SLEV strains from South America, including Brazil (BeAn247377, BeAn242587) and Peru (75D90). Two nucleotide mismatches (silent transition mutations at positions 1083, 1404) were noted in the envelope region within the FL06-S569 and FL06-S650 isolates.

Envelope gene sequences were previously published for 6 Florida strains (9), and 9 additional archived Florida isolates were analyzed for this study. Reference strain FL72-M022 was isolated from an opossum from the Florida panhandle in 1972. FL72-M022 shares 97%–98% sequence identity with strains from Brazil (BeAn246262, BeAr23379, and BeH203235) and is placed in genotype VB. In contrast, SLEV reference strains isolated in Florida during 1952 and 1985 share 97%–99% homology with strains collected in Tampa Bay during 1962 (TBH-28, GHA-3) and in Mexico (65V310). The last large outbreak of SLEV in Florida occurred during 1990. Envelope sequence analysis demonstrated that strains isolated during 1989 and 1990 shared 98% homology with USA (V 2380-42), Guatemala (78A28), Tennessee (TNM 4-711), or Texas strains (83V4953, PVI-2419, 98V3181).

Despite detection of SLEV after the introduction of WNV, SLEV had not recently been cultured by existing statewide surveillance methods in Florida (10). Experimental evidence suggests that WNV cross-protective immunity in wild bird species may limit subsequent SLEV infections (3). In 2006, sentinel seroconversions supported this hypothesis; limited WNV activity may have enabled increased transmission of SLEV during the fall (Figure).

Partner agencies in the Florida Arbovirus Surveillance Network used a targeted strategy to preferentially sample sentinels located in “hot zones” of SLEV transmission activity for virus isolation and molecular analysis. Sequence analysis of reference strains and the 2006 SLEV isolates has shown the circulation of genotype V SLEV strains in Florida. The 2006 isolates do not represent a recent extension of the geographic range of strains of SLEV from Brazil because 1 genotype V strain was also collected during field studies in 1972. Instead, they support periodic circulation and maintenance of South American SLEV genotypes in Florida, where the diverse ecosystem may allow for evolution of the virus and periodic seeding of SLEV into the United States where the human population may have no immunity to the virus.

On the basis of placement into multiple lineages (IIA-IID, VA, VB) (Appendix Figure), our data support the hypothesis that persistence of SLEV in Florida may differ from its activity in other regions of the United States. For example, the same or highly similar strains of SLEV can be locally maintained for more than a decade in California and Texas (15), whereas genetically similar strains of SLEV appear to be infrequently isolated, or maintained at levels below detection, over extended periods in Florida. Our findings suggest periodic introduction of different SLEV genotypes to Florida from the eastern United States and other countries (Mexico, Panama, and Brazil), with distinct North American (lineage II) genotypes isolated in epidemic years. The role of South American genotypes in enzootic or epidemic cycles of SLEV is unknown. In Florida, only the detection of North American genotypes has previously been reported (9,15), but the isolation of South American strains in 1972 and 2006 suggests a mechanism for the continued maintenance of SLEV. Further isolation and characterization of SLEV strains is needed to improve understanding of the mechanism(s) that favor the amplification of North vs. South American genotypes in Florida.