Arboviral screening of invasive Aedes species in northeastern Turkey: West Nile virus circulation and detection of insect-only viruses

Abstract

Background: The recent reports of Aedes aegypti and Ae. albopictus populations in Turkey, in parallel with the territorial expansion identified in several surrounding countries, have raised concerns about the establishment and re-establishment of these invasive Aedes mosquitoes in Turkey. This cross-sectional study was performed to detect Aedes aegypti and Ae. albopictus in regions of recent incursions, and screen for viral pathogens known to be transmitted elsewhere by these species.

Methodology: Mosquitoes were collected at several locations in Artvin, Rize and Trabzon provinces of the Black Sea region during 2016-2017, identified morphologically, pooled and analyzed via generic or specific nucleic acid amplification assays. Viruses in positive pools were identified by product sequencing, cell culture inoculation and next generation sequencing (NGS) in selected specimens.

Principal findings: The study group comprised 791 specimens. Aedes albopictus was the most abundant species in all locations (89.6%), followed by Ae. aegypti (7.8%) and Culex pipiens (2.5%). Mosquitoes were screened for viruses in 65 pools where fifteen (23.1%) were reactive. The infecting strains was identified as West Nile virus (WNV) in 5 pools (7.7%) with Ae. albopictus or Cx. pipiens mosquitoes. The obtained WNV sequences phylogenetically grouped with local and global lineage 1 clade 1a viruses. In 4 (6.2%) and 6 (9.2%) pools, respectively, cell fusing agent virus (CFAV) and Aedes flavivirus (AEFV) sequences were characterized. NGS provided a near-complete AEFV genome in a pool of Ae. albopictus. The strain is provisionally called "AEFV-Turkey", and functional analysis of the genome revealed several conserved motifs and regions associated with virus replication. Merida-like virus Turkey (MERDLVT), a recently-described novel rhabdovirus, was also co-detected in a Cx. pipiens pool also positive for WNV.

Conclusions/significance: Invasive Aedes mosquitoes are established in certain locations of northeastern Turkey. Herein we conclusively show the role of these species in WNV circulation in the region. Biosurveillance is imperative to monitor the spread of these species further into Asia Minor and to detect possible introduction of pathogens.

Fig 1. Map of the locations used for mosquito collection in the study.

(Red: countries/territories with Ae. aegypti and Ae. albopictus; Orange: countries/territories with Ae. albopictus). Blue dots represent sampling locations. The baseline map has been prepared using Natural Earth raster + vector map data in the public domain (URL: www.naturalearthdata.com. Accessed: April 2019), which is freely available for personal, educational, and commercial use. Current information on Aedes species were obtained from the European Centre for Disease Prevention and Control websites (https://ecdc.europa.eu/en/publications-data/aedes-aegypti-current-known-distribution-june-2018; https://ecdc.europa.eu/en/publications-data/aedes-albopictus-current-known-distribution-june-2018; Accessed: December 2018).

Fig 2. The maximum likelihood analysis of the partial West Nile virus NS5 sequences (777 nt).

The tree is constructed using Maximum Likelihood method with the General Time Reversible (GTR) model, Gamma distributed with Invariant sites (G+I) for 1000 replications. The sequences characterized in this study are given in bold and indicated with a symbol, GenBank accession number, hosting mosquito species and pool code. Global virus strains are indicated by GenBank accession number, strain/isolate name and country of detection. Viruses previously characterized in Turkey are indicated with blue letters. Bootstrap values higher than 60 are provided. Japanese encephalitis virus strain GP78 is included as an outgroup.

Fig 3. The maximum likelihood analysis of the partial flavivirus NS5 sequences (355 nt).

The tree is constructed using Maximum Likelihood method with the General Time Reversible (GTR) model, Gamma distributed with Invariant sites (G+I) for 1000 replications. The sequences characterized in this study are given in bold and indicated with a symbol, GenBank accession number, hosting mosquito species and pool code. Global virus strains are indicated by GenBank accession number, virus and strain/isolate name. Viruses previously characterized in Turkey are indicated with blue letters. Bootstrap values higher than 60 are provided.

Fig 4. The maximum likelihood analysis of the near-complete polyprotein coding region (10766 nt) of selected insect-specific flaviviruses.

The tree is constructed using Maximum Likelihood method with the General Time Reversible (GTR) model, Gamma distributed with Invariant sites (G+I) for 1000 replications. The sequence characterized in this study are given in bold and indicated with a symbol, GenBank accession number, virus and isolate/strain name. Global virus strains are indicated by GenBank accession number, virus and strain/isolate name. Viruses previously characterized in Turkey are indicated with blue letters. Bootstrap values higher than 60 are provided.

Fig 5. Plots of similarity of the near-complete polyprotein coding alignment (9931 nt) of AEFV-Turkey (MK251047), with individual functional units indicated (GapStrip: On, Reps: 1000, Kimura (2-parameter), T/t: 2.0).

The curves indicate comparisons between the target and reference genomes (AEFV-Bangkok: KJ741266 AEFV-Narita21: NC012932, AEFV-SPFLD: KC181923). Each point plotted is the percent identity within a sliding window 200 bp wide centered on the position plotted, with a step size between points of 20 bp.