Use of Envelope Domain III Protein for Detection and Differentiation of Flaviviruses in the Free State Province,South Africa

Abstract

The presence of the mosquito-borne flavivirus species West Nile virus (WNV) and Wesselsbron virus (WESSV) in southern Africa is well established; however, their true prevalence remains unknown. To date, the presence of tick-borne flaviviruses has not been confirmed in this region. Serological assays using reagents that can be handled in a biosafety level 2 or lower facility were developed and evaluated for the detection and differentiation of tick- and mosquito-borne flaviviruses in the Free State province of South Africa. A total of 2393 serum samples from a variety of species including humans, cattle, and sheep were tested using Kunjin virus (KUNV) cell lysate antigen for the detection of anti-flavivirus antibodies in an indirect immunoglobulin G (IgG) enzyme-linked immonosorbent assay (ELISA). To further differentiate positive reactors on the KUNV assay for antibodies against tick- or mosquito-borne flaviviruses, recombinant envelope domain III (r-EDIII) proteins of Langat virus (LGTV), WNV, and WESSV were expressed in a bacterial expression system and used in ELISA. A total of 722 samples were positive using the KUNV assay, of which 71, 457, and 431 were positive using the r-LGTVEDIII, r-WNVEDIII, and r-WESSVEDIII assays, respectively. A total of 70 samples were reactive using the KUNV assay but not using any of the other assays, suggesting that there are possibly other flaviviruses circulating in the Free State province for which specific r-EDIII assays were not available. Collectively, the results suggest a strong presence of flaviviruses co-circulating in the Free State province with an abundance of mosquito-borne flaviviruses. There is evidence suggesting the presence of tick-borne flaviviruses, but it has yet to be confirmed. The EDIII protein is a useful tool that can be used in the detection and differentiation of flaviviruses in resource-limited laboratories, but virus neutralization assays are suggested for accurate confirmation of results.

Introduction

The genus Flavivirus is a diverse group that comprises of mosquito-borne and tick-borne viruses, as well as viruses for which there are no known vectors (Porterfield 1980). Some of the members belonging to this genus are significant medical pathogens causing disease, including encephalitis, hemorrhagic fever, and even death. The principal vectors of mosquito-borne flaviviruses are primarily members of the Culex genus; however, species belonging to the Aedes genus serve as vectors for other mosquito-borne flaviviruses such as yellow fever virus (YFV) (Gaunt et al. 2001). The mosquito-borne group of flaviviruses is subdivided into seven groups, some of which have been implicated as major human pathogens (Weissenböck et al. 2010). One such group is the Japanese encephalitis group, which includes Japanese encephalitis virus (JEV), Kunjin virus (KUNV), Usutu virus (USUV), St. Louis encephalitis virus (SLEV), and West Nile virus (WNV) (Halstead and Jacobson 2003). Cases involving WNV outbreaks in humans, horses, and birds have been reported in Africa, Asia, and Europe (Karabatsos 1985) and more recently in the United States (Briese et al 1999, Centers for Disease Control and Prevention 1999). Until recently, all South African isolates of WNV belonged to lineage 2, which was presumed to comprise less virulent strains, and this was supported by the lower number of cases reported in this region compared to those reported in the Northern Hemisphere. However, phylogenetic comparisons (Burt et al. 2002) and transmission of neuroinvasive West Nile fever (Venter et al. 2010) showed that pathogenic and neuroinvasive strains do exist in lineage 2, and WNV may be underestimated in South Africa. Moreover, a WNV isolate belonging to lineage 1 was recently detected in South Africa from the brain of a pregnant mare that had aborted her young and died. The animal had been reported to show signs of neurological disease (Venter et al. 2011). Strains belonging to lineage 2 have also emerged in Hungary in birds of prey (Bakonyi et al. 2006) and in Greece from Culex mosquitoes (Papa et al. 2011).

The yellow fever group is comprised of 10 flaviviruses, including YFV and Wesselsbron virus (WESSV). WESSV is not responsible for large outbreaks of human disease as is YFV, and rather it is implicated as a causative agent of abortion and death of newborn offspring of cattle, sheep, and goats in southern Africa (Blackburn and Swanepoel 1980). The virus has also been isolated in other parts of Africa, such as Nigeria and the sub-Saharan African country of Madagascar (Tomori et al. 1981, Morvan et al. 1990). Swanepoel and Coetzer (1994) reported on an unusual case of laboratory infection of the virus, where the subject developed symptoms suggestive of encephalitis after spraying the virus into their eyes. The virus was first isolated from infected sheep in 1955 in the northern district of the Free State in South Africa, in a small rural town of Wesselsbron (Weiss et al. 1956). Infection has been documented in other parts of Africa from a variety of hosts including humans (Swanepoel 1989).

Tick-borne flaviviruses are mainly transmitted by the Ixodes spp. and include various medically significant pathogens (Lindquist and Vapalahti 2008). Although ticks of the Ixodes spp. are considered the principal vectors of tick-borne encephalitis virus (TBEV), the virus has been isolated from 18 different tick species (Gould et al. 2003) and has also been reported as the most significant cause of viral neurological disease in Europe and Asia (Thiel 2005). Other tick-borne flaviviruses implicated as the cause of encephalitis in humans include louping ill virus (LIV), Powassan virus (POWV), and deer tick virus (DTV), which is a subtype of POWV (Gritsun et al. 2003). Tick-borne flaviviruses are known to be prevalent in Europe but their presence in southern Africa has not been established. However, in a study that investigated the presence of other tick-borne viruses in suspected cases of Crimean–Congo hemorrhagic fever virus (CCHFV) infection, serological evidence suggested the presence of tick-borne flaviviruses in southern Africa (Burt et al. 1996). The finding is also supported by the natural occurrence of the Ixodes tick species in this region. Additionally, there are diagnostic services for tick-borne infections such as CCHFV and Rickettsia spp., but these tests confirm infections in 10% of the cases and the remaining infections are undiagnosed. This raises the question of whether the remaining undiagnosed cases could be due to previously unidentified flaviviruses.

The true prevalence of WESSV and WNV in South Africa is not known and this can largely be due to the high degree of serological cross-reactivity within the genus Flavivirus resulting from common group epitopes on the envelope (E) protein (Gubler 1998). The envelope domain III (EDIII) of flaviviruses was, however, demonstrated to have the ability to differentiate the tick-borne from the mosquito-borne group when used in immunoassays (Holbrook et al. 2004). Furthermore, these viruses are classified as biosafety level 3 pathogens, and this restricts performing culture and virus neutralization techniques in a resource-limited laboratory such as ours. Hence, recombinant proteins are safe reagents that have important applications in diagnosis and surveillance. In this study, we aim to use the recombinant EDIII (r-EDIII) of flaviviruses to differentiate between and determine the seroprevalence of tick-borne flaviviruses, WNV and WESSV in the Free State province of South Africa.

Materials and Methods

Selection of viruses and amplification of genes.

KUNV strain MRM16 was selected as a representative native Flavivirus antigen for detection of antibodies directed against both tick- and mosquito-borne flaviviruses in immunoassays and can be handled in a biosafety level 2 laboratory; thus, it meets the requirements of our laboratory setting. A tick-borne Flavivirus Langat virus (LGTV) strain TP21 and two mosquito-borne flaviviruses WNV strain SA93/01 and WESSV strain SAH177 were selected for the preparation of r-EDIII antigens representative of mosquito- and tick-borne Flavivirus for serological assays. LGTV is a tick-borne Flavivirus species that can be handled in a biosafety level 2 laboratory. Assays based on the WNVEDIII and WESSVEDIII proteins were developed for differentiation of homologous antibodies that will indicate the presumptive prevalence of these viruses in the Free State region. Both viruses are biosafety level 3 pathogens; therefore, only extracted RNA was handled in the laboratory.

For the preparation of the r-EDIII protein, the genomic RNA of LGTV was extracted and primers targeting the EDIII genes of LGTV (LGTVD3F, 5′-AAA GGA TTG ACA TAC ACA GTC TGT GA-3′; LGTVD3R, 5′-CTT CTG AAA CCA CTG ATG GTT CAG G-3′, nucleotides (nt) 1928–1954 and nt 2155–2131, respectively, relative to the Langat virus strain TP21), WNV (WNVD3F, 5′-AAG GGA ACA ACA TAT GGA GTA TG-3′; WNVD3R, 5′-AGA TTT GTG CCA GTG GTG GTT T-3′, nt 1858–1880 and nt 2166–2144, respectively, relative to West Nile virus strain SA93/01) and WESSV (WESSVD3F, 5′-AAA GGT TCA ACC TAC TCA ATG-3′; WESSD3R, 5′-TGA TCT TTG CCA GTG GTA T-3′, nt 1826–1846 and nt 2125–2106, respectively, relative to Wesselsbron virus SAH177) were identified and flanked with sites for BamHI and HindIII restriction endonucleases (Promega, Madison, WI) on the 5′ and 3′ positions, respectively. Using RNA as template, an RT-PCR assay was performed using Titan^® one-step reverse transcriptase PCR kit (Roche, Indianapolis, IN) according to the manufacturer's instructions.

Sequence alignment and phylogenetic analyses

To determine similarities in the EDIII proteins of flaviviruses, the amino acid sequences of the following viruses were aligned using MAFFT version 7 program (Katoh and Standley 2013) with applicable GenBank accession numbers WNV (EF429198.1), JEV (AEN14607.1), LGTV (EU790644.1), LIV (NC001809.1), POWV (AAL32144.1), WESSV (DQ859058.1), YFV (AAD45531.1), KUNV (AF196505.1), and TBEV (KC414090.1) as shown in Table 1. The amino acid sequences were used to construct a neighbor-joining phylogenetic tree using MEGA version 4 software (Tamura et al. 2007). Details of the isolates are provided in the phylogenetic tree (Fig. 1).

FIG. 1.

Phylogenetic comparison of the EDIII protein of various flaviviruses. The neighbor-joining tree was constructed by using MEGA version 4 software (Tamura et al. 2007) and 1000 bootstrap replicates by using the maximum composite likelihood algorithm. Scale bar indicates nucleotide substitutions.

Table 1.

Amino Acid Sequences of the EDIII Proteins of Various Flaviviruses

Symbols denote the homology between amino acids: ( ^*) identical, (:) conserved substitutions, (.) semi-conserved substitutions. A space denotes variable substitutions.

GenBank accession numbers are: WNV, EF429198.1; JEV, AEN14607.1; LGTV, EU790644.1; LIV, NC001809.1; POWV, AAL32144.1; WESSV, DQ859058.1; YFV, AAD45531.1; KUNV, AF196505.1; TBEV, KC414090.1.

WNV, West Nile virus; JUNV, Kunjin virus; JEV, Japanese encephalitis virus; LIV, louping ill virus; TBEV, tick-borne encephalitis virus; LGTV, Langat virus; POWV, Powassan virus; WESSV, Wesselsbron virus; YFV, yellow fever virus.

KUNV cell lysate antigen

Vero cells (ATCC no. CCL-81™) were cultured in 1×Eagle minimum essential media (EMEM) (BioWhittaker^®, Portsmouth, NH) containing nonessential amino acids (BioWhittaker), 2 mM L-glutamine, and 100 U/mL penicillin-streptomycin (Sigma-Aldrich, St. Louis, MO). The medium was enriched with 10% fetal bovine sera (FBS; Delta Bioproducts, Johannesburg, South Africa) and discarded upon confluency of over 90%. A 1-mL preparation of KUNV diluted 10-fold from a 10% mouse brain suspension was used to inoculate the cells. Cells were incubated for 3–4 days and harvested from the flasks before cytopathic effects (CPE) occurred. Following centrifugation, supernatants were discarded and pellets resuspended in borate-buffered saline (BBS) (pH 9). Cells were clarified, and pellets were further resuspended in a reduced volume of BBS, 1% Triton X-100, and 0.1% sodium dodecyl sulfate (SDS) and sonicated in a sonicator bath (Branson Ultrasonics, Danbury, CT). Cell debris was removed by centrifugation at 4000×g for 10 min, and supernatant was aliquoted and stored at −70°C until needed.

Expression and characterization of proteins

Each of the amplicons was initially cloned into pGem-T-Easy^® vectors (Promega). Genes derived from LGTV and WESSV were subcloned into a pQE-80L expression vector (Qiagen, Hilden, Germany), and the gene derived from WNV was subcloned into a pCold-TF expression vector (Takara Bio, Paris, France), which allows the co-expression of the target protein with a 48-kD chaperone. The resulting pQE-80L–LGTVEDIII, pQE-80L–WESSVEDIII, and pCOLD-TF–WNVEDIII constructs were used to transform competent Escherichia coli OverExpress C43 (DE3) cells (Lucigen, Middleton, WI). Mock antigens were prepared using cells transformed with pCOLD-TF and pQE-80L plasmids lacking any of the Flavivirus EDIII genes. The E. coli transformants were propagated in Luria Bertani (LB) broth containing 100 μg/mL ampicillin, incubated overnight at 37°C, and used as starter cultures. Cultures of bacterial cells were harvested and lysed by resuspending the pellet in 5 mL of BugBuster Protein Extraction Reagent (Novagen, Madison, WI) per gram of pelleted cells. All the proteins were expressed with a 6XHis-tag for facilitation of protein purification. Cell lysates were loaded onto an 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel, and proteins were separated by running the gel at 150 V for 80 min. The gel was stained with Coomassie stain for visualization. Expressed proteins were further evaluated for solubility properties and purified from the crude lysates using Protino^® Ni-TED packed columns (Macherey-Nagel, Duren, Germany) under native conditions with buffers supplied in the kit according to manufacturers' instructions. Briefly, proteins expressed with a 6XHis-tag were allowed to bind to the Protino^® columns by eluting cell lysates through the columns. Columns were washed with 1× lysis-equilibration-wash (LEW) buffer and subsequently eluted with 1× elution buffer containing 250 mM imidazole (buffers were supplied in the kit). The recombinant proteins were characterized by western blot analyses using the Pierce^® Fast Western Blot Kit (Thermo Scientific, Waltham, MA) according to the manufacturer's instructions and as previously described (Samudzi et al. 2012).

Serum samples

Human sera with antibodies positive for WNV, YFV, and LIV confirmed using hemagglutination inhibition test by an external laboratory were obtained with informed consent from volunteers with a history of infection. Mouse hyperimmune ascitic fluid (MHAF) against TBEV was obtained from Dr. Robert Tesh from the Arbovirus Reference Center (University of Texas Medical Branch at Galveston) and MHAF against WNV, WESSV, Zika virus (ZIKV), dengue virus serotype (DENV-1), Spondweni virus (SPOV), YFV, LIV, WNV, Kadam virus (KADV), and Banzi virus (BANV) were collectively obtained from Professor Janusz Paweska from the Center for Emerging and Zoonotic Diseases at the National Institute for Communicable Diseases in South Africa. Serum samples collected between 2008 and 2011 from 196 human patients in the Free State province presenting with a tick bite associated with acute illness, and no diagnoses were obtained anonymously from the routine serology laboratory of the Medical Microbiology Department at the University of the Free State. A total of 2052 serum samples collected between 2007 and 2012 from cattle submitted for routine screening for brucellosis in four districts within the Free State province and serum samples from 88 sheep, four horses, 11 rats (Aethomys namaquensis), five mice (Mastomys coucha and Rhabdomys pumilio), two shrews (Elephantulus myurus), 2 bats (species unidentified), 9 rooihartbees (hartebeest, Alcelaphus buselaphus), nine eland (Tragelaphus oryx), five warthogs (Phacochoerus africanus), three buffalos (Syncerus caffer), two gemsbuck (Oryx gazella), two giraffes (Giraffa camelopardalis), and two zebras (Equus quagga) were obtained from the Free State Veterinary Services Laboratory in Bloemfontein, South Africa or collected at various sites in the Free State.

Enzyme-linked immunosorbent assay

Following the preparation of the KUNV cell lysate antigen and r-EDIII proteins of WNV, LGTV, and WESSV, indirect immunoglobulin G (IgG) enzyme-linked immunosorbent assays (ELISA) were developed for the detection of anti-Flavivirus antibodies in serum samples from the species mentioned above. For each assay, the antigen was diluted in 1× phosphate-buffered saline (PBS) in an optimal dilution determined by performing checkerboard titrations, used to coat 96-well PolySorp immuno-plates (Nunc™, Roskilde County, Denmark), and incubated overnight at 4°C. Wells were washed with 1% Tween-20/PBS between incubations of 1 h at 37°C. After blocking wells with 200 μL of 10% skimmed milk in PBS, serum samples diluted 100-fold in 2% skimmed milk/PBS were added to each well. Anti-species secondary antibodies conjugated to horseradish peroxidase (HRP; Zymed Inc., California, and KPL, Washington DC) were diluted 4000-fold and added to the wells, except for serum samples collected from wildlife and rodents, where Pierce^® recombinant protein A/G conjugate HRP (Thermo Scientific, Waltham, MA) was used at a 10,000-fold dilution. The reactions were visualized by addition of 2,2′-azino-di(3-ethylbenzthiazoline-6-sulfonate) (ABTS^®) peroxidase substrate (KPL), and optical density (OD) readings were read at 405 nm using a Sunrise absorbance reader (Tecan, Durham, NC). Cutoff values for each assay were determined by calculating the mean net OD (sample OD on test antigen − sample OD on mock antigen)+2 (KUNV, r-WNVEDIII, r-WESSVEDIII) or 3 (r-LGTVEDIII) standard deviations, depending on the intensity of the background.

Specificity testing

The specificity of immunoassays was evaluated by cross-reacting KUNV antigen and recombinant proteins with samples containing antibodies against flaviviruses that are known to be endemic in southern Africa (WNV, WESSV), those that have previously been reported in southern Africa (SPDV, BANV) (Smithburn et al. 1959, McIntosh 1986), as well as nonendemic flaviviruses that are known to be significant human pathogens (DENV, YFV, TBEV), including other tick- and mosquito-borne flaviviruses representing additional serogroups (KADV, ZIKV).

Results

Amino acid sequences of EDIII

To investigate diversity based on predicted amino acid sequences of the Flavivirus EDIII protein, the amino acid sequences of the EDIII proteins of various flaviviruses were aligned (Table 1). The results were further employed to construct a phylogenetic tree (Fig. 1). On the basis of predicted amino acid data, the viruses formed distinct lineages in agreement with previous studies based on complete or partial sequence data. Although WNV and WESSV are serologically cross-reactive, they belong to distinct serogroups, whereas the tick-borne viruses are more closely related.

Enzyme-linked immunosorbent assay

The 13-kD r-EDIII proteins of LGTV, WESSV, and WNV were expressed in a bacterial expression system and purified from the soluble fraction (not shown). The ability of KUNV cell lysate antigen to detect antibodies against tick- and mosquito-borne flaviviruses is illustrated in Figure 2. All species that were confirmed to be positive for flaviviruses using a hemagglutination inhibition assay performed at an external laboratory were detected using this assay, and it was discriminatory against sera that were negative for flaviviral antibodies. To ascertain if the LGTVEDIII recombinant protein would differentiate between tick-borne and mosquito-borne flaviviruses, the protein was reacted with sera positive for antibodies against tick-borne TBEV, LIV, KADV, and seven mosquito-borne flaviviruses (Fig. 3).

FIG. 2.

Validation of Kunjin virus (KUNV) cell lysate antigen for detection of anti-Flavivirus antibodies from various species in enzyme-linked immunosorbent assay (ELISA). The cutoff value of the assay was determined to be net OD₄₀₅=0.23, as indicated by the horizontal line.

FIG. 3.

Evaluation of the LGTVEDIII recombinant protein for specificity against mosquito-borne flaviviruses in enzyme-linked immunosorbent assay (ELISA). The cutoff value of the assay was determined to be net OD₄₀₅=0.28, as indicated by the horizontal line.

The assay was able to differentiate antibodies directed against the three tick-borne flaviviruses from those against the mosquito-borne and Flavivirus-negative sera. Similarly, the recombinant WNVEDIII recombinant protein was instrumental in differentiating antibodies raised against tick-borne flaviviruses from those raised against WNV as well as sera with no anti-Flavivirus antibodies (Fig. 4). Though the WNVEDIII immunoassay was able to differentiate between anti-WESSV and anti-WNV antibodies, demonstrating a differentiation potential to species level, the assay was not able to differentiate antibodies raised against DENV. DENV is, however, not endemic in southern Africa.

FIG. 4.

Evaluation of the WNVEDIII recombinant protein for differentiation between antibodies raised against mosquito-borne and tick-borne flaviviruses in enzyme-linked immunosorbent assay (ELISA). The cutoff value of the assay was determined to be net OD₄₀₅=0.226, as indicated by the horizontal line.

The recombinant WESSVEDIII antigen also showed similar results. The assay was able to differentiate between anti-WESSV and anti-WNV antibodies as well as sera raised against a tick-borne Flavivirus. Antibodies that are negative for Flavivirus were also not detected. Antibodies raised against YFV could not, however, be differentiated (Fig. 5). This was not a major limitation for this study because YFV is not endemic in southern Africa.

FIG. 5.

Evaluation of the WESSVEDIII recombinant protein for differentiation between antibodies raised against mosquito-borne and tick-borne flaviviruses in enzyme-linked immunosorbent assay (ELISA). The cutoff value of the assay was determined to be net OD₄₀₅=0.23, as indicated by the horizontal line.

Subsequent to their validation, the four assays were further used to detect anti-Flavivirus antibodies in a variety of species, as summarized in Table 2. The initial screening for anti-Flavivirus antibodies was performed using the KUNV immunoassay, and samples that were confirmed as positive based on the cutoff value were further tested using the more specific recombinant LGTV, WNV, and WESSV assays.

Table 2.

Summary of Samples Tested on Serological Assays for Presence of Flavivirus Antibodies

Species tested	Total number of samples	KUNV assay	LGTVEDIII assay	WNVEDIII assay	WESSVEDIII assay
Human	196	67	18	47	54
Cattle	2052	595	39	372	376
Sheep	88	31	14	26	1
Horses	5	3	0	3	0
Rats	11	0	ND	ND	ND
Mice	5	0	ND	ND	ND
Shrews	2	0	ND	ND	ND
Bats	2	0	ND	ND	ND
Rooihaartbees (hartebeest)	9	8	0	1	0
Eland	9	8	0	1	0
Warthog	5	5	0	4	0
Buffalo	3	2	0	1	0
Gemsbuck	2	2	0	2	0
Giraffe	2	0	ND	ND	ND
Zebra	2	1	0	0	0
Total	2393	722	71	457	431

ND, no further tests done as samples were negative for antibody against Kunjin virus (KUNV).

Screening of field samples

As shown in Table 2, a total of 722 samples were positive in the KUNV assay and were further tested using the r-EDIII assays. While a large proportion (40.8%) of the samples was only detected using one of the EDIII recombinant protein assays, the rest of the samples were either detected by multiple r-EDIII assays (43.5%) or the KUNV assay only (9.8%), as illustrated in Figure 6. A total of 34.2% of the samples reacted against both WNVEDIII and WESSVEDIII antigens with mono-detections of the WNVEDIII (22.5%) and WESSVEDIII (17.9%) assays. Mono-detection of recombinant WNVEDIII and WESSVEDIII further proves the potential of the r-EDIII protein to differentiate between antibodies raised against the serologically cross-reactive WNV and WESSV. Table 3 outlines the number of samples that were differentiated by the r-EDIII proteins of WNV and WESSV. Samples that were detected on both the LGTVEDIII and WNVEDIII assays accounted for 2.5%, whereas those detected on LGTVEDIII and WESSVEDIII accounted for 2.8%. A small proportion of samples (4%) were detected on all three r-EDIII assays, and the mono-detection of LGTVEDIII assay was the lowest at 0.6%. The remaining 5.6% of samples not illustrated in Figure 6 could not be tested further due to insufficient sample.

FIG. 6.

Profiles of the detection of samples positive on the Kunjin virus (KUNV) assay by the r-EDIII proteins of Langat virus (LGTV), West Nile virus (WNV), and Wesselsbron virus (WESSV).

Table 3.

Differentiation of Field Samples by r-EDIII Proteins of WNV and WESSV

Assay result	WESSV positive	WESSV negative
WNVEDIII positive	246	207
WNVEDIII negative	178	91

WNV, West Nile virus; WESSV, Wesselbron virus.

To further understand the putative prevalence of flaviviruses in the Free State province, samples detected using the immunoassays developed in this study were correlated to their site of collection and the results were analyzed per district (Fig. 7). All districts that were surveyed showed similar distribution for samples that were detected using the KUNV assay only, and they ranged between 8% and 12% in each district. Samples that were detected using the recombinant WNVEDIII were abundant in all districts (35–50%), with the highest number in the Xhariep and Lejweleputswa regions. Similarly, there was an ample representation of the samples detected on the recombinant WESSVEDIII assay in all the districts (28–55%), with dominance in the Motheo (47%) and Thabo Mofutsanyane (54%) districts. Detection observed using the LGTVEDIII was the lowest in all regions (average 5%), and no detection was observed in the Thabo Mofutsanyane district. The Xhariep district however, showed a high presence of recombinant LGTVEDIII-detected samples (12%).

FIG. 7.

Correlation of the detection of anti-Flavivirus immunoglobulin G (IgG) antibody results to the geographical distribution in four districts of the Free State province. Various symbols have been used to represent the number of samples detected on each of the assays developed in this study: () KUNV; () WNVEDIII; () WESSVEDIII; () LGTVEDIII; () no samples were collected in this district.

Discussion

The presence of mosquito-borne flaviviruses in southern Africa is known, but their prevalence has not been fully established. In addition, cases of Flavivirus disease may have been underreported in this region due to limited routine diagnostic services. An assay that can detect antibodies against flaviviruses could have a role in diagnosis and in serological surveillance of various species. Due to the high cross-reactivity between members of this genus, it is essential to have an assay that can differentiate between them. The r-EDIII protein of flaviviruses is a useful tool that has been demonstrated to have potential in differentiating between flaviviruses (Holbrook et al. 2004). This domain has been reported to comprise predominantly subcomplex and type-specific epitopes that mainly induce neutralizing antibodies (Oliphant et al. 2007) and has also been observed to be highly antigenic with primary linear epitopes (Beasley and Barrett 2002) on the basis of X-ray crystallography (Rey et al. 1995).

EDIII is probably involved in viral receptor binding (Crill and Roehrig, 2001) and can be expressed as a recombinant protein that has the ability to fold as a functional fragment (Martina et al. 2007). Although this protein has the ability to differentiate between antibodies directed against tick- and mosquito-borne flaviviruses, our study has suggested that it has potential to differentiate between WESSV and WNV, which are known to occur in southern Africa. The potential of WESSVEDIII and WNVEDIII assays to differentiate between antibodies, as well as the high proportions of discrimination between field samples detected by these assays, further shows the potential of the EDIII protein as a tool for surveillance. The gold standard for differentiation between flaviviruses is the neutralization assay, but this could not be performed because our laboratory is limited to a biosafety level 2 facility and both WNV and WESSV are biosafety level 3 pathogens. Therefore, the r-EDIII protein is a valuable tool that may be beneficial for surveillance in low-resource laboratories that cannot perform neutralization assays due to viruses classified above level 2.

The amino acid sequences alignments of EDIII proteins of WNV and WESSV showed a homology of 36% (Table 1). Additionally, the two viruses belong to different serogroups within the genus, and this is also consistent with the phylogenetic tree that was constructed based on the amino acid sequences of various flaviviruses (Fig. 1). The phylogenetic comparison between the EDIII protein of WNV and that of WESSV positions these viruses in distinct lineages of their respective serogroups, namely JEV and YFV. The r-WNVEDIII has previously been shown to have potential in differentiating homologous antibodies from those raised against viruses in the same serogroups, namely JEV, St. Louis encephalitis virus, and Murray Valley encephalitis virus (Beasely et al. 2004). This further substantiates the discriminatory potential of the EDIII protein in immunoassays.

The EDIII proteins of KUNV and WNV grouped together on the phylogenetic tree similar to a study that used phylogenetic analyses to indicate that certain strains of WNV are more similar to KUNV and form a separate lineage to other WNV strains, suggesting the addition of more subtypes in the WNV group (Hall et al. 2006). Both viruses are associated with febrile illness with rash or mild encephalitis (Scherret et al. 2001) as well as neurological disease in horses (Badman et al. 1984, Venter et al. 2010). The limitation that was observed with the r-WNVEDIII was the detection of antibodies raised against DENV. Though this may be a major problem in regions where DENV is endemic, it was not believed to play a major role in the surveillance of field samples in this study because DENV is not endemic in southern Africa.

In contrast to WNV, WESSV has not been associated with a fatal outcome of disease, and cases involving human infection with WESSV may be underreported. In a serological survey of human sera performed in South Africa, WESSV was recorded in 32% of samples collected in the subtropical region of KwaZulu Natal province and 0.7% in the southern Cape (McIntosh 1980). Recently, WESSV was isolated from two human cases in which an initial diagnosis of Rift Valley fever virus (RVFV) infection was suspected due to an outbreak of the RVFV that occurred in South Africa, with human and livestock infections, between 2010 and 2011. These isolates were included in a phylogenetic analysis of collective WESSV isolates based on the sequence data of the NS5 gene (Weyer et al. 2013). The study identified two clades of WESSV circulating in southern Africa. Clade I included isolates collected from South Africa and Zimbabwe, whereas clade II was comprised exclusively of isolates from the KwaZulu Natal province of South Africa. Our study also shows a high serodetection of human samples on the r-WESSV assay (54/196).

On the basis of the specificity testing performed on the r-WESSVEDIII in ELISA, none of the antibodies raised against other flaviviruses tested were detectable by this antigen except for the YFV antibodies, which is not endemic in southern Africa. Moreover, WNV is known to be the only other endemic flavivirus in this region and could be distinguished from WESSV. Additionally, humans from whom samples were collected did not have a history of YFV vaccination, which rules out YFV as a cause of seropositivity. The limited diagnostic service for WESSV disease and lack of awareness of the infection may account for mild, self-limiting cases of infection being undiagnosed.

The isolation of WESSV from samples collected during the recent RVFV outbreak indicates that there was circulation of other mosquito-borne viruses during this period (Weyer et al. 2013). This is not uncommon because the outbreak occurred during a period of high rainfall. In our study, we observed that a considerable amount of the bovine serum samples that were positive on the KUNV anti-Flavivirus antibody assay was collected after the 2010 outbreak of RVFV (National Institute for Communicable Diseases Bulletin 2011). A total of 537 samples were collected in this period, of which 281 were positive on the assay. This increases the serodetection for flaviviruses to 52% compared to the 21% (314/1515) that was observed in samples that were collected prior to the outbreak.

Similarly, the ovine samples that were collected after the outbreak showed a more positive trend than those collected prior. Of the 88 ovine samples tested, 40 were collected after the outbreak with a seropositivity of 68% (27/40), which is much higher than the 8% (4/48) collected before the outbreak. Ovine samples that were positive for the r-LGTV assay were all collected after the outbreak with a seropositivity of 35% (14/40) compared to 0% (0/48) before the outbreak. This suggests that tick-borne flaviviruses were also circulating during the outbreak, particularly in sheep than in cattle. Although seropositivity of ovine samples seem to have increased after the outbreak on the r-WESSV and r-WNV assays, the seropositivity of bovine sera remained constant regardless of the outbreak. All human serum samples were collected prior to the 2010 RVFV outbreak.

On the basis of specificity testing, the r-LGTVEDIII protein was able to react with antibodies raised against heterologous tick-borne flaviviruses and only excluded those raised against mosquito-borne flaviviruses. Our in-house specificity testing for r-LGTVEDIII against seven samples positive for mosquito-borne flaviviruses belonging to various serotypes within the group and three samples positive for tick-borne flaviviruses provided repeatable data that excluded tick-borne flaviviruses from mosquito-borne flaviviruses. This observation positions the LGTVEDIII assay as a reliable tool for the detection of anti-tick-borne Flavivirus antibodies and is consistent with the grouping of tick-borne flaviviruses in the phylogenetic tree. Holbrook and co-workers (2004) also reported that members of the tick-borne serocomplex of flaviviruses are highly cross-reactive and cannot be differentiated on the EDIII protein.

The r-EDIII proteins prepared in this study suggest the circulation of WNV, WESSV, and tick-borne flaviviruses in the Free State province that has not been previously established. However, accurate identification of these samples requires confirmation by the virus neutralization assay. Flaviviruses seem to be widely distributed in this province, and their co-circulation has been suggested in various farms where samples would be collectively positive for WNV, WESSV, and tick-borne flaviviruses. Antibodies raised against mosquito-borne flaviviruses were most abundant. This was not alarming because the presence of WNV and WESSV has been established in southern Africa, and their transmission by the same mosquito vector would explain the high degree of co-circulation of WNV and WESSV in all districts. Although neutralization assays are needed for confirmation, the detection of antibodies by r-LGTVEDIII suggests that tick-borne flaviviruses may be circulating in southern Africa. The detection of these antibodies was not only made in samples from livestock, but also in human patients that had had a tick bite and no diagnosis, suggesting, but not confirming, that tick-borne flaviviruses may have played a role in their illness. Future studies will focus on developing molecular techniques to detect isolates of novel mosquito-borne flaviviruses and confirm the presence of tick-borne flaviviruses.

Conclusion

The r-EDIII protein of flaviviruses is a safe, cost-effective, and significant tool for the detection of flaviviruses, especially in low-resource laboratories. The protein can be used for differentiation between the serologically cross-reactive WNV and WESSV as well as tick-borne flaviviruses. Although neglected, the putative prevalence of the WESSV in the Free State is suggested. Together with WNV and WESSV, tick-borne flaviviruses may also be circulating on the farms located in the Free State. The presence of other flaviviruses is also suggested, but this data will have to be confirmed using the virus neutralization test.

Footnotes

Acknowledgments

The authors wish to thank the Free State Veterinary Laboratory in Bloemfontein for submission of samples, and Prof. Janusz Paweska and Dr. Robert Tesh for donating mouse hyperimmune ascitic fluid used as controls. The authors also wish to thank the Poliomyelitis Research Foundation and the National Health Laboratory Services (NHLS) Research Trust for project funding as well as the NHLS Universitas where the study was performed.

Author Disclosure Statement

No competing financial interests exist.

References

Badman

, Campbell

, Aldred

. Arbovirus infection in horses—Victoria. Comm Dis Intell (Aust.), 1984; 17:5–6.

Bakonyi

, Ivanics

, Erdelyi

, Ursu

, et al. Lineage 1 and 2 strains of encephalitic West Nile virus, central Europe. Emerg Infect Dis, 2006; 12:618–623.

Beasely

, Barrett

ADT

. Identification of neutralizing epitopes within structural domain III of the West Nile virus envelope protein. J Virol, 2002; 76:13097–13100.

Beasely

, Holbrook

, Travassos Da Rosa

, Coffey

, et al. Use of recombinant envelope protein subunit antigen for specific serological diagnosis of West Nile Infection. J Clin Microbiol, 2004; 42:2759–2765.

Blackburn

, Swanepoel

. An investigation of flavivirus infections of cattle in Zimbabwe Rhodesia with particular reference to Wesselsbron virus. J Hyg Camb, 1980; 85:1–33.

Briese

, Xi-Yu

, Huang

, Grady

, et al. Identification of Kunjin/West Nile-like flavivirus in brains of patients with New York encephalitis. Lancet, 1999; 354:1261–1262; erratum in Lancet 1999; 354:1650.

Burt

, Spencer

, Leman

, Patterson

, et al. Investigation of tick-borne viruses as pathogens of human in South Africa and evidence of Dugbe infection in a patient with prolonged thrombocytopenia. Epidemiol Infect, 1996; 116:353–361.

Burt

, Grobbelaar

, Leman

, Anthony

, et al. Phylogenetic relationships of southern African West Nile virus isolates. Emerg Infect Dis, 2002; 8:820–826.

Centers for Disease Control and Prevention (CDC). Outbreak of West Nile-like viral encephalitis-New York 1999. MMWR, 1999; 48:845–849.

10.

Crill

, Roehrig

. Monoclonal antibodies that bind to domain III of dengue virus E glycoprotein are the most efficient blockers of virus adsorption to Vero cells. J Virol, 2001; 75:7769–7773.

11.

Ceausu

, Erscoiu

, Calistru

, Ispas

, et al. Clinical manifestations in the West Nile virus outbreak. Rom J Virol, 1997; 48:3–11.

12.

Gaunt

, Sall

, de Lamballerie

, Falconar

, et al. Phylogenetic relationships of flaviviruses correlate with their epidemiology, disease association and biogeography. J Gen Virol, 2001; 82:1867–1876.

13.

Gould

, de Lamballerie

, Zanotto

, Holmes

. Origins, evolution and vector/host coadaptations within the genus Flavivirus. Adv Virus Res, 2003; 59:227–314.

14.

Gritsun

, Nuttall

, Gould

. Tick-borne flaviviruses. Adv Virus Res, 2003; 61:317–371.

15.

Gubler

. The global pandemic of dengue/dengue hemorrhagic fever: Current status and prospects for the future. Ann Acad Med, 1998; 27:227–234.

16.

Hall

, Scherret

, Mackenzie

. Kunjin virus: An Australian variant of West Nile?. Ann NY Acad Sci, 2006; 1:153–160.

17.

Halstead

, Jacobson

. Japanese encephalitis. Adv Virus Res, 2003; 61:103–138.

18.

Holbrook

, Shope

, Barret

ADT

. Use of recombinant E protein domain III-based enzyme-linked immunosorbent assays for differentiation of tick-borne encephalitis serocomplex flaviviruses form mosquito-borne flaviviruses. J Clin Microbiol, 2004; 42:4101–4110.

19.

Karabatsos

, ed. International Catalogue of Arboviruses, Including Certain Other Viruses of Vertebrates, 3 ^rd ed. San Antonio, TX: American Society of Tropical Medicine and Hygeine, 1985.

20.

Katoh

, Standley

. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Mol Biol Evol, 2013; 30:772–780.

21.

Lindquist

, Vapalahti

. Tick-borne encephalitis. Lancet, 2008; 371:1861–1871.

22.

Martina

, Koraka

, van den Doel

, van Amerongen

, et al. Immunization with West Nile virus envelope domain III protects mice against lethal infection with homologous and heterologous virus. Vaccine, 2007; 26:153–157.

23.

McIntosh

. The epidemiology of arthropod-borne viruses in southern Africa, DSc dissertation. Pretoria: University of Pretoria, 1980:144–159.

24.

McIntosh

. Mosquito-borne virus diseases of man in southern Africa. S Afr Med J, 1986; Oct 11(Suppl):69–72.

25.

Morvan

, Fontenille

, Digoutte

, Coulanges

. The Wesselsbron virus, a new arbovirus for Madagascar. Arch Inst Pasteur Madagascar, 1990; 57:183–192.

26.

National Institute for Communicable Diseases (NICD). The outbreak response unit, Special Pathogens Unit (SPU), South African field epidemiology and laboratory training programme (SAFELTP), National Institute for Communicable Diseases (NICD) of the National Health Laboratory Services (NHLS). Rift Valley fever outbreak: Interim report on the 2011 Rift Valley fever (RVF) outbreak in South Africa. 2011. Last updated on 20 May 2011. Accessed on 15 April 2013 . Available at http://www.nicd.ac.za/?page=rift_valley_fever_outbreak&id=94

27.

Porterfield

. Antigenic characteristics and classification of Togaviridae In: Schlesinger

, ed. The Togaviruses. New York: Academic Press, 1980:13–46.

28.

Oliphant

, Nybakken

, Austin

, Xu

, et al. Diamond MS. Induction of epitope-specific neutralizing antibodies against West Nile virus. J Virol, 2007; 81:11828–11839.

29.

Papa

, Bakonyi

, Xanthopoulou

, Vazquez

, et al. Genetic characterization of West Nile virus lineage 2, Greece, 2010. Emerg Infect Dis, 2011; 17:920–922.

30.

Rey

, Heinz

, Mandl

, Kunz

, et al. The envelope glycoprotein from tick-borne encephalitis virus at 2 A resolution. Nature, 1995; 375:291–298.

31.

Samudzi

, Leman

, Paweska

, Swanepoel

, et al. Bacterial expression of Crimean-Congo hemorrhagic fever virus nucleoprotein and its evaluation as a diagnostic reagent in an indirect ELISA. J Virol Methods, 2012; 179:70–76.

32.

Scherret

, Poidinger

, Mackenzie

, Broom

, et al. The relationships between West Nile and Kunjin viruses. Emerg Infect Dis, 2001; 7:697–705.

33.

Smithburn

, Paterson

, Heymann

, Winter

. An agent related to Uganda S virus from man and mosquitoes in South Africa. S Afr Med J, 1959; 33:959–962.

34.

Swanepoel

. Wesselsbron virus disease In: Monath

, ed. The Arboviruses: Epidemiology and Ecology. Boca Rota: CRC Press Inc., 1989:32–57.

35.

Swanepoel

, Coetzer

JAW

. Rift valley fever In: Coetzer

JAW

, Thomson

, Tustin

, ed. Infectious Diseases of Livestock with Special Reference to Southern Africa. Cape Town: Oxford University Press, 1994:688–717.

36.

Tamura

, Dudley

, Nei

, Kumar

. MEGA4: Molecular evolutionary genetics analysis (MEGA) software, version 4.0. Mol Biol Evol, 2007; 24:1596–1599.

37.

Thiel

. Family Flaviviridae. In: Fauquet

, ed. Virus Taxonomy: Eighth Report of the International Committee on Taxonomy of Viruses. Amsterdam-Boston-Heidelberg-London-New York-Oxford: Elsevier Academic Press. 2005:981–998.

38.

Tomori

, Monath

, O'Connor

, Lee

, et al. Arbovirus infections among laboratory personnel in Ibadan, Nigeria. Am J Trop Med Hyg, 1981; 30:855–861.

39.

Venter

, Steyl

, Human

, Weyer

, et al. Transmission of West Nile virus during horse autopsy. Emerg Infect Dis, 2010; 16:573–575.

40.

Venter

, Human

, van Niekerk

, William

, et al. Fatal neurological disease and abortion in mare infected with lineage 1 West Nile virus, South Africa. Emerg Infect Dis, 2011; 1534–1536.

41.

Weiss

, Haig

, Alexander

. Wesselsbron virus—a virus not previously described associated with abortion in domestic animals. Onderstepoort J Vet, 1956; 27:183–195.

42.

Weissenböck

, Hubálek

, Bakonyi

, Nowotny

. Zoonotic mosquito-borne flaviviruses: worldwide presence of agents with proven pathogenicity and potential candidates for future emerging diseases. Vet Microbiol, 2010; 140:271–280.

43.

Weyer

, Thomas

, Leman

, Grobbelaar

, et al. Human cases of Wesselsbron disease, South Africa 2010–2011. Vector Borne Zoonotic Dis, 2013; 13:330–336.