Genetic characterization of a novel picorna-like virus in Culex spp. mosquitoes from Mozambique

The mosquito collection and viral metagenomic analysis were performed in our previous study on Culex mosquitoes [17]. Briefly, mosquitoes were collected from Cuacua village in the Zambezi Valley of Mozambique from October–November 2014 using CDC light traps, and the genus was determined morphologically. The mosquito pools (up to 20 mosquitoes per pool) were homogenized mechanically. The total RNA was extracted with TRIzol LS reagent (Invitrogen, Life Technologies, USA) according to the manufacturer’s instructions. The samples were pre-amplified by Sequence Independent Single Primer Amplification (SISPA) and submitted to the SciLifeLab for library preparation and sequencing. The sample was sequenced on an Ion Torrent PGM sequencer with an Ion 318™ chip v2 and a max read length of 400 bp. The high-quality reads were mapped to the host genomes (Anopheles, Aedes and Culex genomes), and unmapped reads were then classified using BLASTn and BLASTx with an E-value cutoff of 1e-03. The virus-related sequences were extracted and assembled with the de novo assembler SPAdes [18].

Extracted RNA was used to synthesize cDNA using Superscript III (Invitrogen) as recommended by the supplier. First-strand synthesis was initiated with random primers. To sequence the genome gaps between the HTS assembled contigs, primers were designed, and the PCR reactions were carried out in a 25-μl reaction using the thermal profile as follows: 95 °C for 10 min; followed by 35 cycles of 95 °C for 30 s, 60–62 °C for 30 s and 72 °C for 1–1.5 min; 72 °C for 7 min. The list of the primers used in this study is enclosed as a Additional file 1. The positive PCR products were sequenced at Macrogen Europe (Macrogen Inc.) using Sanger sequencing. After the amplification and genetic characterization, Bowtie2 [19] was used to assess the genome coverage in the original HTS dataset by mapping the viral metagenomic reads back to the assembled, near full-length viral genome.

The extreme genomic 5′ and 3′ ends were analyzed using rapid amplification of the cDNA ends (RACE). To determine the 3′ end of the viral genome, a total of 1.2 μg of total RNA were annealed with a 0.5 μM of poly (A) specific primer (AP) in a 20 μl volume by heating at 65 °C for 5 min and chilling on ice for 2 min. First-strand cDNA was synthesized with Superscript III at 42 °C for 50 min followed by an enzyme inactivation at 70 °C for 15 min. A gene-specific forward primer (GSP-F) and universal amplification primer (UAP) were used to amplify the viral 3′ end sequences using AmpliTaq Gold DNA polymerase. For the 5′ genomic ends, 1.6 μg of total RNA and 0.5 μM of GSP-RT primer were mixed in a total volume of 20 μl, incubated at 65 °C for 5 min and chilled on ice for 2 min. First-strand synthesis was carried out by Superscript III as described previously. The RNA templates from the cDNA:RNA hybrids were degraded with 1.5 U of RNase H at 37 °C for 20 min. The excess primers from the cDNA reaction were removed using the GeneJet PCR purification kit (Thermo Fisher Scientific). The cDNA was recovered in 15 μl of elution buffer and mixed with 5 μl of 5× tailing reaction buffer, 2.5 μl of 2.5 mM dCTP and 2 μl of terminal deoxynucleotidyl transferase (Tdt) (Thermo Fisher Scientific) for 5′ end tailing at 37 °C for 15 min and then 10 min at 65 °C to inactivate the reaction. For the first round of amplification, Platinum SuperFi DNA Taq polymerase (Thermo Fisher Scientific) was used. In short, 2 μl of cDNA was used in a 25-μl reaction with 10 pmols of each GSP-reverse primer and an AUAP forward primer. The following thermo profile was used to carry out the amplification reactions: 95 °C for 2 min; 35 cycles of 95 °C for 10 s, 60–62 °C for 10 s and 68 °C for 2 min; 68 °C for 5 min and then cooling to 4 °C. The amplification products from the first round were diluted 20-fold in 1 mM EDTA. For the second round of amplification, 2 μl of diluted product and 10 pmols of GSP2- reverse primer and AUAP forward primer were used in a final volume of 25 μl, and the reaction was carried out as before. The amplified PCR products were visualized, and the bands were purified using GeneJet Gel Extraction Kit (Thermo Fisher Scientific). The purified PCR products were cloned into the pJET1.2 vector using the CloneJET PCR cloning kit (Thermo Fisher Scientific) and sequenced at Macrogen Europe.

The open reading frame (ORF) of the viral genome was predicted using ORF finder at NCBI (https://​www.​ncbi.​nlm.​nih.​gov/​orffinder). The conserved domains (helicase, protease and RNA-dependent RNA polymerase (RdRP)) in the predicted ORF amino acid sequence were analyzed by multiple sequence alignment with other members of the order Picornavirales using ClustalW. The pairwise identity percentage matrix was generated with MegaAlign 9.0.4 (DNASTAR). To determine the phylogenetic relationships, the predicted RdRP region of different viruses belonging to the order Picornavirales were obtained from GenBank. ClustalW alignment of 421 amino acids (aa) corresponding to 2650–3071 aa positions were used and the phylogeny was generated using the Maximum Likelihood method based on the JTT matrix-based model with MEGA 7.0.26 software [20]. All positions containing gaps and missing data were eliminated, resulting a total of 256 aa positions of RdRP region in the final dataset. The statistical significance of the tree topologies was evaluated by 500 bootstraps. Viral sequences used in the multiple sequence alignment and phylogeny are summarized in Table 1.

Extraction of the total RNA from Culex and Mansonia mosquito pools was performed as described above. Up to one microgram of total RNA was used for first-strand cDNA synthesis with Superscript III and random hexamers, and RT-PCR was performed with AmpliTaq Gold DNA polymerase. The PCR primers used in the assay were 8F (5’-CGACCTAGGACTTATCCAGC-3′) and 7R (5’-ACAATCTAGTGCCTCCTTCTG-3′), with an expected amplicon size of 577 bp. The PCR program was used as follows: 95 °C for 10 min, 35 cycles at 95 °C for 30 s, 60 °C for 30 s and 72 °C for 1 min, and a final extension at 72 °C for 5 min.

The viral genomic sequence was found to be A/U rich (A- 32.23%, U- 31.02%, G- 21.79% and C- 14.94%). An in silico analysis of the identified nucleotide sequence of CuPV-1 showed that the genomic RNA contains a single large open reading frame (ORF) oriented from the 5′ to 3′ end. This large ORF consists of 9339 nt, encoding a 3112-amino acid protein and accounting for 95.58% of the CuPV-1 genome (Fig. 1). It has a predicted molecular mass of 352.21 kDa and theoretical isoelectric point (pI) of 5.78. The ORF was appended by a partial 5′ and 3′ UTRs which are 361 and 41 nt respectively. No large ORFs were found in the inverse orientation of the CuPV-1 genome, suggesting that the CuPV-1 genome is a positive-strand RNA virus.

Conserved domains of structural proteins were found on 5′ end of the CuPV-1 ORF. An rhv_like domain (Picornavirus capsid protein domain_like, cd00205) was identified by NCBI BLAST conserved domain suite with an E-value of 3.73e-19, at amino acids 682–876. The deduced amino acid multiple sequence analysis of the insect picorna-like viruses, including CuPV-1, revealed that CuPV-1 contains key motifs that are known to be present in the capsid proteins of picornaviruses. The conserved motifs identified on the amino acid sequences were: YXGX₈VX₄HX₉F for 1C (VP3), FXRG and DDFX₇GXP for 1D (VP1) (Additional file 2, A and B). The cleavage site for 1B/1C (NX/DXP) was also detected in CuPV-1 (Fig. 1). No conserved motifs for 1A (VP2) (NXNXFQXG) and the leader protein, the most variable region of the insect picorna-like virus genomes, were identified.

The comparison of the CuPV-1 non-structural proteins with other picorna-like viral proteins was performed to identify the similarities and conserved regions of the putative helicase, protease and RdRP (Fig. 2).

Helicase: Three conserved helicase regions were recognized in the deduced amino acid sequences of the predicted CuPV-1 ORF ranging from 1624 to 1797. A conserved RNA-helicase domain (pfam00910, E-value 2.83E-11) was identified in this region. The highly conserved consensus sequence within the first domain, GXXGXGKS, was found between amino acid positions 1650–1657, although ‘K’ was substituted for ‘G’ at the 1653 amino acid position. The last two conserved domains deviated somewhat from the consensus. The highly conserved amino acids were QX₅DD and KGX₄SX₅STN, while the equivalents in CuPV-1 were HX₅DD and KDX₄PX₅TSN, respectively (Fig. 2a).

Protease: The deduced amino acid sequence of the CuPV-1’s ORF from 2331 to 2508 is similar to the protease sequence of the other picorna-like viruses [10, 21]. The conserved motif, GXCG, was found at 2469–2472, and the equivalent motif of GXHXXG, SXHXXG was found at 2485–2490. The amino acids are thought to form a catalytic triad of the protease, with the presence of H²³⁴⁹, D²³⁸⁸ and C²⁴⁷¹ in this region (Fig. 2b).

RNA-dependent RNA polymerase (RdRP): Eight conserved domains, found between 2650 and 3071 in the deduced amino acid sequence of CuPV-1, correspond to those recognized previously, and an RNA_dep_RNAP domain (cd01699; E-value 2.52e-44) was also found in this region (Fig. 2c). This showed that the CuPV-1 RdRP belongs to a superfamily of positive-strand RNA eukaryotic viruses. The conserved or equivalent domains I-VIII in the RdRP of CuPV-1 are located in amino acids between 2743 and 3041 (Table 3). The RdRP amino acid sequence identity between CuPV-1 and closely related virus (HplV-35) was found to be 57.2%. The amino acid sequence identity matrix with RdRP regions and complete polyproteins was described in the Additional file 3.

To determine the phylogenetic relationship of CuPV-1 in other members of the Picornavirales order, phylogenetic analysis was performed using the highly conserved RdRP region including I-VIII domains. The virus formed a clade with known members of iflaviruses, such as Hubei picorna-like virus 35 and Hubei picorna-like virus 34 (unclassified picorna-like viruses), Sacbrood virus (iflavirus) and Hubei arthropod virus 1 (HuAV-1) (Fig. 3). Other clades consisted of viruses that belonged to the families, such as Dicistroviridae, Picornaviridae, Secoviridae, Marnaviridae and newly proposed Polycipiviridae. This suggests that CuPV-1 belongs to the Iflaviridae viral family.

A total of 340 mosquitoes were collected in the Zambezi Valley, a central region of Mozambique. The specimens included mosquitoes of two species, Culex (159) and Mansonia (181), in 23 pools (13 Culex and 10 Mansonia). These mosquito pools were screened with RT-PCR, using primers designed for the selected CuPV-1 RdRP region. Seven pools were positive for CuPV-1, 5 from Culex and 2 from Mansonia. Each pool included 1–20 individuals and included both male and female mosquitoes (Table 4). The overall minimal infection rate (MIR), which was expressed as the number of positive pools per 1000 mosquitoes, was 0.20 (7/340), while the specific MIR for Culex was 0.31 (5/159) and for Mansonia was 0.11 (2/181). The Sanger sequencing of the positive PCR products showed that pools 9, 12, and 13 were identical to CuPV-1, however, Culex pools 3 and 5 showed minor sequence variation compared to both CuPV-1 (97% identity) and each other (97% identity).

The Culex metagenomic dataset, after the quality check, was mapped to the CuPV-1 genome using Bowtie2 to estimate the read coverage in the original data set. Among the 1,684,319 reads in the dataset, 80,636 reads (4.83% of total reads) were aligned throughout the genome with a coverage range of 9–9784 (Fig. 1), with a lower coverage towards the ends of the genome.