A total of 6,791 mosquitoes collected as eggs at 151 study sites in Wisconsin, Minnesota, and Iowa (Figure 1) were reared and tested for LACV antigen by immunofluorescence assay (IFA). Of these, 309 (4.6%) were positive. Viral RNA was amplified by RT-PCR from one to three mosquitoes from the selected sites listed in Table 1. Four LACV isolates from 1960, 1978, 2006 and 2007 were also examined in this study. The viruses from 2006 and 2007 were isolated from mosquitoes collected in the field. L, M, and S viral RNA (see Amplicon Cloning and Sequencing) was also amplified from the two virus isolates as well as directly from the two infected mosquitoes. The L, M, and S sequences from the viruses and the RNA amplified directly from the mosquitoes were identical (data not shown).

The numbers of sequences analyzed and the number of segregating sites in each segment are shown in Table 2. The greatest nucleotide diversity (π) was seen in the M segment, twice that in the S segment and thrice that in the L segment. The distributions of these polymorphisms are shown in Figure 2. What is most noteworthy is that all three segments had more replacement than synonymous substitutions. In the L segment the diversity among replacement substitutions (π_a) was actually 3.24 times larger than the diversity among synonymous substitutions (π_s). The location and amino acid replacements are listed in Table 3. These trends suggest that some form of positive selection is operating on amino acid substitutions in all three segments.

The program Tipdate [28] estimated the molecular evolutionary rate (substitutions/site), the absolute molecular evolution rate (substitutions/site/year) of each segment and the age of the dataset (the time in years since the sequences evolved from a common ancestral sequence)(Table 4). The absolute evolution rate was most rapid in the S segment, 480 times greater than the rate in the L segment and 4.8 times greater than the rate in the M segment. Both the M and S segments appear to be of similar ages, while the L segment appears to predate both by ~400,000 years.

The haplotype grouping system was determined through a conservative phylogenetic analysis. The system identified three S haplotypes based on seven polymorphic sites, five of which were nonsynonymous mutations. The three haplotypes identified in the M segment were based on twelve polymorphic sites, seven of which were nonsynonymous. For the L segment, two haplotypes were identified based on thirteen polymorphic sites, twelve of which were nonsynonymous substitutions (Figure 3).

Maximum parsimony phylogenetic trees were established using amplified sequences from each of the three segments. Comparison of the clades on the three maximum parsimony trees provides evidence for the potential for transmission of reassortant viruses by the infected Ae. triseriatus (Figures 4, 5, 6). If there were no reassortants, the three genome segments from each infected mosquito would have appeared in the same clade. A number of mosquitoes contained viral genome segments that clustered into different clades in each of the trees. For example, the S segment from the sample MCBB/La Crosse/2004 was in haplotype #2 (red), the M segment in haplotype #2 (predominantly red) and the L segment in haplotype #1 (mixture of red and blue). Another example is the LACV RNA from the mosquito collected in NFCS/Winona/2004. The S segment was in haplotype #3 (purple), the M segment in haplotype #2 (predominantly red), and the L segment in haplotype #1 (mixture of red and blue). These suggest that segment reassortment had occurred. The distribution of the sequences in the phylogenetic trees for all three segments would be identical if reassortment had not occurred; however, the phylogenetic trees are highly variable when the S, M and L segment tree topologies are compared.

A linkage disequilibrium analysis was performed within and among the S, M, and L segments. Figure 7 is a heat diagram in which low disequilibrium coefficients are represented by light yellow squares and high disequilibrium coefficients are represented by red squares. The matrix is read according to the nucleotide position of segregating sites displayed along the diagonal. For example in Figure 7, the lowest square connects sites S22 (segregating site 22 from the S segment) and S86 and it is red. This corresponds to an r² of 1.00 and these sites are in complete linkage disequilibrium. In contrast, squares linking site S359 with all other sites are light yellow indicating that all sites in S are in equilibrium with S359. The triangles along the diagonal in Figure 7 contain many red squares indicating that many sites within a segment are in disequilibrium. Thus our coverage of each of the segments appears adequate.

The squares in Figure 7 indicate patterns of disequilibrium among segments. In contrast to the large amounts of disequilibrium found within each of the segments, there is very little disequilibrium among segments. Between S and M there are 192 (12 S sites × 16 M sites) possible interactions but only two of these are in disequilibrium: S359 with M12 and M126. Otherwise 99% of possible interaction between S and M are in equilibrium indicating extensive reassortment between these segments. Between S and L there are again 192 possible interactions but only two in disequilibrium; these are S232 and L427. All other possible interaction between S and L are in equilibrium indicating reassortment between these segments. Between M and L there are again 256 possible interactions but only two of these are in disequilibrium: M179 with L312 and L314. All other possible interaction between M and L are in equilibrium indicating reassortment between these segments.

An independent heterogeneity χ² analysis (Table 5) was performed to test this pattern. There were 3 S clades, 3 M clades and 2 L clades; thus there were 18 possible segment combinations corresponding to each row in Table 5. The observed column is the number of times that a segment combination occurred in the 45 samples. Eight of the combinations were in disequilibrium but 10 were in equilibrium (in bold) supporting an inference of frequent reassortment. In total, eleven of the 45 (24.4%) samples were in linkage equilibrium

Both phylogenetic and linkage disequilibrium analyses revealed that LACV RNA genome segments had undergone reassortment in 24% of mosquitoes and isolates analyzed. This is remarkable and illustrates the exceptional evolutionary potential and genetic diversity of Bunyaviridae viruses in nature. One possible reason for this could be the ability of Ae. triseriatus to become dually infected. When mosquitoes ingest two different LACV isolates simultaneously or sequentially within four hours, 100% become dually infected [17]. Even at 48 hours post-initial bloodmeal, 27% of mosquitoes that ingest a second virus become dually infected before a barrier to superinfection develops. In addition, when transovarially-infected mosquitoes ingested a bloodmeal containing a heterologous LACV, 19% became dually infected [18]. These experiments suggest that dual infection can occur frequently through both oral and transovarial infection and therefore increase the possibility of segment reassortment in vectors. The newly evolved viruses are also efficiently transmitted [17]. These experiments were performed in a controlled laboratory setting, but they demonstrate the potential for segment reassortment to occur frequently in nature.

Although the analyses demonstrate the potential for reassortment, most of the sequences used were from RNA amplified directly from the infected mosquitoes and not from virus isolates. The reassortment frequency detected in this study could have resulted from analysis of RNA quasispecies sequences in the mosquito. However the L, M, and S sequences obtained from the virus isolates as well as those directly amplified from the infected mosquitoes in 2006 and 2007 were identical. This suggests that 1) the genome sequence obtained by direct amplification of the viral RNA from the mosquito is the dominant viral sequence in the mosquito as well as in infectious virus and 2) that the estimation of reassortment frequency was not confounded by potential RNA quasispecies in the mosquitoes. Estimating the frequency of reassortment of LACV in nature would be improved by analysis of plaque-purified viruses isolated from the mosquitoes, preferably from their saliva or ovaries, which are the epidemiologically significant organs of transmission.

In this regard, we were unsuccessful in isolating LACV from most field mosquitoes. The reasons for this are unknown; however, there are several potential explanations for this. Eggs were collected in the field and stored in a hot warehouse for variable periods of time awaiting shipment to Colorado. As soon as the eggs reached AIDL, they were placed in the insectary, hatched and reared. Environmental factors in the collection and shipping process could contribute to loss of virus titer. An additional complication could have been the isolation method. In previous studies, virus was isolated by inoculation of samples into suckling mouse brains. Cell culture assays are likely not as sensitive. Low virus titer, titer loss during processing, and insensitive isolation methods, likely contributed to the inability to isolate virus from mosquitoes.

Aedes triseriatus eggs were collected from five oviposition traps in each of 151 sites in Minnesota (n = 37), Wisconsin (n = 108) and Iowa (n = 6). Sites were established in areas where LACV encephalitis cases occurred or areas that contained clusters of people judged by the La Crosse County Public Health Department to be at risk for infection (e.g. wooded areas adjacent to houses with children, schools, or playgrounds). Mosquito eggs that had entered diapause in fall 2000 were collected in the spring of 2001. Mosquito eggs were also collected between mid-June and August of 2004, 2006 and 2007. Eggs were collected in Crawford, La Crosse, Monroe, Vernon, Lafayette and Iowa counties in Wisconsin; Winona, Houston, and Grant counties in Minnesota; and Clayton and Allamakee counties in Iowa (Figure 1). Eggs were transported to the insectaries at the Arthropod-borne and Infectious Diseases Laboratory (AIDL) at Colorado State University (CSU); Fort Collins, CO. Eggs were hatched immediately and reared to adults.

To determine if mosquitoes were infected, mosquito heads were severed, squashed onto acid-washed microscope slides, and fixed in acetone. Heads were assayed for LACV antigen by direct IFA using LACV-specific polyclonal antiserum [30].

Viral RNA from 40 mosquitoes was analyzed, including 34 field collected mosquitoes from 2004 and six field collected mosquitoes from 2000.

Previously isolated LACV strains were also used in the analysis. The 1960 LACV isolate was isolated originally from the brain of a child who died from LACV encephalitis in La Crosse, WI and it was passed five times in suckling mouse brains (SMB). A 1978 LACV (78V-8853) was isolated from an Ae. triseriatus mosquito from Rochester, MN and passed once in Vero cells and twice in SMB. LACV was isolated from mosquitoes collected in the field in WI and MN in 2006 and 2007.

The LACV-positive mosquitoes were triturated with a pellet pestle (Fisher Scientific) in a 1.5 ml microcentrifuge tube containing 1 ml of minimum essential medium (MEM) (Gibco), 2% fetal bovine serum, 200 μg/ml penicillin/streptomycin, 200 μg/ml fungicide, 7.1 mM sodium bicarbonate, and 1× nonessential amino acids. The homogenate was centrifuged for 10 minutes at 500 × g to form a pellet.

Cell monolayers of Vero cells were grown in six-well plates at 37°C in an atmosphere of 5% CO₂. Supernatant from the centrifuged mosquito homogenate (0.2 ml) was added to one well in a six-well plate, incubated at 37°C for one hour. Following the incubation, 5 ml of medium were added to each well.

The virus isolates from 2006 and 2007 were plaque purified using monolayers of Vero cells in six-well plates [31]. Virus isolates were serially diluted 10^-1 to 10^-6 and 200 μl of each virus dilution was added to individual wells and incubated at 37°C for 1 hour. The virus inoculum was removed and 5 ml of overlay was added to the well. After six days of incubation at 37°C in 5% CO₂, 200 μl of the detection solution, methylthiazolyldiphenyl-tetrazolium bromide (MTT) (5 mg/ml in PBS), was added to each well. The plates were incubated overnight and visible plaques were picked and placed in 1 ml of MEM with 0.2% FBS for 1 hr at 37°C. An aliquot of the medium from the wells was added to Vero cells, and the presence of virus confirmed by detection of cytopathic effect.

The posterior half of each mosquito abdomen was individually homogenized in 500 μl of Trizol (Invitrogen, Carlsbad, CA), using a pellet pestle (Fisher Scientific, Pittsburg, PA), and then total RNA was extracted according to manufacturer's instructions.

The medium and cells from wells with plaque purified virus were removed and placed in a 15 ml conical tube and centrifuged at 3000 rpm for 10 minutes. The supernatant was removed and the cell pellet was resuspended in 500 μl of Trizol (Invitrogen, Carlsbad, CA). Total RNA was extracted according to manufacturer's instructions.

RNA from the 1960 and 1978 LACV isolates was prepared by infection of C6/36 cell cultures at a multiplicity of infection of 0.01. Three days post-infection, cells were scraped into the medium, centrifuged and cell pellets were resuspended in 500 μl of Trizol for RNA extraction.

Portions of the LACV S, M, and L RNA segments were transcribed to cDNA using Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA) and amplified by PCR using Ex Taq DNA polymerase (Takara, Shiga, Japan) according to manufacturer's instructions. The primers specific for the S segment (forward: 5'-GCAAATGGATTTGA TCCTGATGCAG-3', reverse: 5'-CTTAAGGCCTTCTTCAGG TATTGAG-3') amplified a 462 nucleotide region (nucleotides 144 to 604) of the nucleocapsid and NSs genes. This region was selected because it was the most variable region of the published S sequences. The S segment is 984 nucleotides in length, so the amplified region encompasses almost half the entire segment. The primers specific for the M segment (forward: 5'-CCAAAAGCAACAAAAGAAAGA-3', reverse: 5'- CTGAAGGCATGAT GCAAAG-3') amplified a highly variable 411 nucleotide region in the 5' half of the G1 gene (nucleotides 1585 to 1995) [32]. The primers specific for the L segment (forward: 5'-GCATGTGTAGCCAAGGATATCGATG-3', reverse: 5'-CAGTCTTGCACCAGG GTGCTGTAAG-3') amplified a 487 nucleotide region (nucleotides 140 to 626). These primers also were selected to amplify the most variable region of the L segment. Primers specific for the Ae. triseriatus ribosomal protein RpL34 mRNA were used as a positive control. PCR was performed as follows: 94°C for 5 minutes, 35 cycles of 94°C for 1 minute, 55°C for 1 minute and 72°C for 1 minute followed by a final extension at 72°C for 8 minutes.

PCR products were separated by electrophoresis in 1% agarose gels with TAE buffer, visualized with ethidium bromide, excised and extracted using the Powerprep Express Gel Extraction kit (Marligen Biosciences, Ijamsville, MD) according to manufacturer's instructions. PCR products were inserted into the pCR4-TOPO cloning vector (Invitrogen, Carlsbad, CA) and resulting plasmids were used to transform competent TOP10 E. coli cells (Invitrogen, Carlsbad, CA). Cells were grown on LB agar containing ampicillin (50 μg/ml) and kanamycin (50 μg/ml). Colonies were screened for inserts by PCR amplification using the original primers and positive products were purified using QIAquick spin columns (Qiagen, Valencia, CA). Three to five cDNA clones per segment from each mosquito were sequenced in both directions using the ABI PRISM dye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA) and the ABI 310 DNA automated sequencer at Macromolecular Resources, CSU. A 415 nucleotide region of the S segment (nucleotides 190–604), a 358 nucleotide region of the M segment (nucleotides 1637–1994), and a 447 nucleotide region of the L segment (nucleotides 179–625) were sequenced.

Genetic haplotypes were established for each of the three segments through maximum parsimony analysis, sequence identity matrix, neighbor joining distance matrix, and ratio of synonymous to nonsynonymous substitutions.