Introduction

Aedes mosquitoes are known vectors of pathogenic flaviviruses (genus Flavivirus, family Flaviviridae), including dengue virus and yellow fever virus. Furthermore, insect-specific flaviviruses (ISFs) have been detected in these mosquito species in numerous regions of the world, including European countries [2, 7, 14, 16, 24, 25, 31]. ISFs comprise a distinct group of flaviviruses that do not infect mammalian cells, replicate only in mosquitoes or mosquito cell lines, and are not associated with disease in humans. They are probably primordial flaviviruses [8, 27]. Some ISFs have been shown to produce DNA forms of their genomic RNA; integrated sequences related to ISFs have been detected in A. aegypti and A. albopictus mosquitoes [9, 31].

In 2010, West Nile virus (WNV) emerged in Greece and caused large outbreaks of human infections [20]. During entomological surveys conducted in the summer months of 2010-2014, WNV (lineage 2) was detected every year in Culex spp. mosquitoes [11, 18, 19, 21–23]. Furthermore, Culex theileri flavivirus, an ISF, has been detected in Culex spp. mosquitoes in Greece [23]. There is increasing interest in the interactions of ISFs with pathogenic flaviviruses (including WNV) in mosquitoes, and studies have produced contradictory results. It appears that these interactions depend on several factors, including the species of mosquitoes, the flaviviruses involved, the timing of infection, and the load of each virus [4, 5, 13, 15]. Concerning the role of Aedes spp. in the WNV life cycle, vector competence studies have shown that Aedes (Ochlerotatus) caspius is an inefficient vector of WNV in the laboratory [3], while Aedes vexans may play a role in WNV enzootic cycles [12, 30]. The vast majority of Aedes spp. mosquitoes in the rice fields in Thessaloniki, northern Greece, are Aedes caspius, representing 28.5 % of the trapped mosquitoes. The other mosquito species are Anopheles pseudopictus (21 %), Culex modestus (19.5 %), C. pipiens (16.8 %), Anopheles hyrcanus (9.6 %) and Anopheles sacharovi (4.3 %) [8]. Since Aedes mosquitoes in Greece have never been tested for flaviviruses, the aim of the present study was to check whether Aedes spp. mosquitoes collected during WNV outbreaks were infected by the virus and to investigate whether they were carrying any other pathogenic flavivirus or ISF.

Materials and methods

Following the emergence of WNV in 2010 in Greece, mosquito surveillance studies were conducted every year, targeting mainly Culex spp. mosquitoes. During 2010 and 2012, 1173 Aedes spp. mosquitoes were transported in dry ice to the National Reference Centre for Arboviruses in order to be tested for WNV infection (in 2011, only Culex spp. mosquitoes were studied). The mosquitoes were collected using CO2–baited light traps and were identified at the genus level using dichotomous determination keys [10]. Of these, 469 were collected in August and September 2010 (soon after the initiation of the WNV outbreak) at 33 sites in six regional units (Nomenclature of Territorial Units for Statistics 3 level: Imathia, Kilkis, Larisa, Pella, Pieria, Thessaloniki), and 704 were collected in October 2012 at nine sites in five regional units (Chalkidiki, Imathia, Pieria, Serres, Thessaloniki) (Table 1, Fig. 1).

Table 1 Areas in Greece where insect-specific flaviviruses were detected in 2010 and 2012
Full size table
Fig. 1
figure 1

Locations where insect-specific flaviviruses were detected. The sites where OCFV and AeVeV were detected are marked with a white circle and a triangle, respectively (the black circle indicates the site where the non-indentified virus was detected)

Full size image

The Aedes spp. mosquitoes were grouped into 53 pools of up to 100 individuals that were sorted according to the collection site and date. RNA was extracted using the RNeasy Mini Kit (QIAGEN, Hilden, Germany), and a reverse transcriptase (RT)-nested PCR using flavivirus generic primers was applied [26]. Positive samples were tested by PCR with three additional sets of primers (primer sequences are available upon request) to obtain sequences of a 520-bp fragment of the envelope (E) protein gene and 500-bp and 650-bp fragments of the genes for the non-structural (NS) proteins NS3 and NS5, respectively. All flavivirus-positive pools were re-tested without the reverse transcription (RT) step in order to check for possible DNA forms. Sequencing of the PCR products was performed using a BigDye Terminator v3.1 Cycle Sequencing Kit in a 3130 ABI Genetic Analyzer (Life Technologies, Carlsbad, CA, USA). Multiple alignment of the sequences was performed using Clustal W, and phylogenetic trees were constructed by the neighbor-joining method with 1,000 bootstrap replicates in MEGA6 [29]. Virus isolation was attempted in Vero E6 cells.

Results

Eight (8/53, 15.1 %) pools of Aedes spp. mosquitoes were found to contain flavivirus sequences. Using the NCBI BLAST search tool [1], five sequences showed the highest similarity to Ochlerotatus caspius flavivirus (OCFV), which was previously detected in Portugal [14] and Italy [6], two sequences were 100 % identical to Aedes vexans flavivirus (AeVeV), previously detected in Italy (GQ476997) [6] and the Czech Republic (JN802280) [7], and one sequence was too short (<60 bp) to be indentified accurately (Table 1). When the eight positive pools were retested by PCR without the reverse transcription step, the five OCFV-positive pools were negative, suggesting that they were RNA forms, while the remaining three pools were positive, suggesting that they included DNA forms. However, it cannot be excluded that RNA forms were also present in the samples. Virus isolation attempted in Vero E6 cells from two OCFV-positive pools was not successful. The Aedes albopictus C6/36 cell line, which is suitable for the replication of ISFs, was not available.

The percentage of OCFV-positive mosquito pools was 9.43 % (5/53 pools), with the highest percentage observed in 2010 in the regional unit of Thessaloniki (4/31, 12.9 %) (Table 1, Fig. 1). The exact collection sites and dates of the OCFV-positive pools are shown in Table 2. Two pools of Culex spp. mosquitoes (100 mosquitoes per pool) collected in 2013 in which OCFV sequences were detected are also shown in Table 2 (marked with an asterisk).

Table 2 Spatial and temporal data for the ISF-positive mosquito pools collected in Greece
Full size table

Most of the OCFV-positive Aedes mosquitoes (4/5 pools) were collected in August 2010 in the regional unit of Thessaloniki, while all Aedes mosquitoes (23 pools) collected in September 2010 at the same sites were negative. Additional sequences from the partial E, NS3 and NS5 genes were obtained from three of the five OCFV-positive samples. The mean genetic differences among the Greek OCFV flaviviruses in partial fragments of the E, NS3 and NS5 genes at the nucleotide level were 0.4 %, 1.4 % and 0.2 %, respectively, and 0 %, 0.2 % and 0 % at the amino acid level. In all three phylogenetic trees, the Greek OCFV strains constitute a distinct subclade, clustering with related sequences obtained from Portugal [14], Spain (referred to as Mediterranean Ochlerotatus flavivirus) [31] and Finland (Hanko virus, HANKV) [16] (Fig. 2A, B, and C). The genetic differences at the nucleotide and amino acid levels between the Greek OCFV strains and the related viruses from the Iberian Peninsula (Spain and Portugal) and Finland are shown in Table 3. In all fragments, the nucleotide sequence identity is >84 %, which is the threshold value for flaviviruses to be classified within the same species [17]. The nucleotide sequences obtained from the present study have been deposited in the GenBank database under the accession numbers KM245094-KM245102.

Fig. 2
figure 2

Neighbor-joining phylogenetic trees based on partial fragments of the genes encoding the A) E (507 nt), B) NS3 (461 nt), and C) NS5 (460 nt) proteins of OCFV. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The percentage of replicate trees in which the sequences clustered together in the bootstrap test (1000 replicates) is shown next to the branches. Sequences from the present study are indicated. Evolutionary analysis was conducted in MEGA6 [29]

Full size image
Table 3 Mean nucleotide and amino acid differences between the Greek OCFV strains and the Iberian OCFV and Hanko virus in partial fragments of genes encoding the E, NS3 and NS5 proteins
Full size table

Discussion

In order to determine whether Aedes mosquitoes in Greece were infected by WNV and/or other flaviviruses during WNV outbreaks, a generic flavivirus RT nested PCR was applied to test genetic material extracted from Aedes spp. mosquitoes collected in various locations in northern Greece. Of the 53 Aedes spp. pools tested, eight (15.09 %) were found to be flavivirus positive. Five of the eight sequences obtained from the generic PCR showed highest similarity (>90 %) to OCFV detected in Portugal [14]. This was confirmed by phylogenetic analysis of partial genome fragments corresponding to the E, NS3 and NS5 genes: the Greek OCFV sequences clustered together with OCFV sequences from Spain [31] and Portugal [14], with HANKV being included in the same clade [16].

The genetic difference among the Greek OCFV strains was 0.2-1.4 %, and these strains differed by 2.2-4.1 % from the Iberian strains, and by 6.2-11.1 % from HANKV. The differences among strains varied depending on the genome region, with the NS3 gene being the most variable; however, the threshold value of 84 % [17] was not overcome even in the NS3 gene. Nucleotide sequence differences were observed even between sequences detected in pools from mosquitoes collected in the same area. For example, the pools Thessaloniki 23/10 and Thessaloniki 25/10 consisted of mosquitoes collected on the same night in two locations in the Thessaloniki region only 10 km apart. According to Kuno et al. (1998), all these viruses should be considered members of one virus species. Blitvich and Firth [4] suggested HANKV as the name of these related viruses.

There was a drastic difference in the prevalence between 2010 and 2012, since all of the OCFV-positive pools were collected in 2010 (5 of 44 pools, 11.36 %), while all mosquitoes collected in 2012 were OCFV negative. It has to be mentioned that the collection of 2010 was performed from mid-August to late September, while the collection of 2012 was conducted from late September to early October. A seasonality of activity was observed, since all OCFV-positive Aedes mosquitoes were collected in August 2010, while all (23 pools) Aedes mosquitoes collected in September 2010 at the same sites were negative. Furthermore, OCFV-positive Culex spp. mosquitoes were collected in the same season (late July and mid-August of 2013 (Table 3). The fact that similar OCFV sequences were taken from Culex mosquitoes suggests that Culex spp. can also be infected by OCFV. Similar findings were reported also in other countries in southern Europe [7].

Since 1975, when the first ISF (cell fusing agent virus, CFAV) was isolated [28], the number of ISFs has increased dramatically [4], mainly through entomological surveys in which the detection of pathogenic flaviviruses was attempted using PCR protocols with generic primers. So far, ISFs of the OCFV and HANKV group have been detected in Portugal [14], Spain [31], Italy [6], Finland [16] and Greece, suggesting that they are widely distributed in Europe. There are several issues concerning ISFs that need to be clarified, such as their spatial and temporal distribution, factors associated with their prevalence, their complete genome sequences, and their possible effect on human health, either directly or through their interactions with other flaviviruses in the mosquitoes.