A Reverse Genetics System for Zika Virus Based on a Simple Molecular Cloning Strategy

Abstract

The Zika virus (ZIKV) has recently attracted major research interest as infection was unexpectedly associated with neurological manifestations in developing foetuses and with Guillain-Barré syndrome in infected adults. Understanding the underlying molecular mechanisms requires reverse genetic systems, which allow manipulation of infectious cDNA clones at will. In the case of flaviviruses, to which ZIKV belongs, several reports have indicated that the construction of full-length cDNA clones is difficult due to toxicity during plasmid amplification in Escherichia coli. Toxicity of flaviviral cDNAs has been linked to the activity of cryptic prokaryotic promoters within the region encoding the structural proteins leading to spurious transcription and expression of toxic viral proteins. Here, we employ an approach based on in silico prediction and mutational silencing of putative promoters to generate full-length cDNA clones of the historical MR766 strain and the contemporary French Polynesian strain H/PF/2013 of ZIKV. While for both strains construction of full-length cDNA clones has failed in the past, we show that our approach generates cDNA clones that are stable on single bacterial plasmids and give rise to infectious viruses with properties similar to those generated by other more complex assembly strategies. Further, we generate luciferase and fluorescent reporter viruses as well as sub-genomic replicons that are fully functional and suitable for various research and drug screening applications. Taken together, this study confirms that in silico prediction and silencing of cryptic prokaryotic promoters is an efficient strategy to generate full-length cDNA clones of flaviviruses and reports novel tools that will facilitate research on ZIKV biology and development of antiviral strategies.

Keywords: ZIKV; cryptic promoter silencing; full-length molecular clone; plasmid toxicity; reporter virus; subgenomic replicon.

Figure 1

Construction and stability of synthetic full length Zika virus (synZIKV) cDNA clones. (A) Schematic representation of the synZIKV MR766 construct and the four fragments used to assemble the genome. The 5′ and 3′UTRs are indicated with bold black lines, the promoter for the T7 RNA polymerase with a black arrow. Restriction sites used for the assembly of the fragments are indicated. An enlargement of fragment #1 is shown below with putative CEPs (score > 0.85) indicated by red arrow heads. CEP 1 was not mutated (indicated with the pink arrow head). (B) Same as in panel (A) but for synZIKV-H/PF/2013. (C) Restriction patterns of pFK-synZIKV constructs obtained after digest with EcoRI (MR766) or XmnI (H/PF/2013) and agarose gel electrophoresis. Plasmids were analysed directly after assembly (original prep) and after five passages (P5) in E. coli (five DNA clones of P5 are shown).

Figure 2

Replication kinetics of viruses obtained with the full-length synZIKV clones. (A) Replication kinetics of the two synZIKV clones as determined by plaque assay. VeroE6 cells were transfected with in vitro transcribed synZIKV RNAs and virus contained in culture supernatant at different time points after transfection was measured. Mean ± SEM of two independent experiments is shown. (B,C) Comparison of replication kinetics of synZIKV and parental viruses. Huh7 cells were infected with either ZIKV using a multiplicity of infection (MOI) of 1. Supernatants from infected cells were harvested at indicated times post-infection and titres were determined by plaque assay. Mean ± SEM of three independent experiments is shown. (D) Comparison of plaque morphology of synZIKV and the parental viruses. (E,F) Replication kinetics of passaged synZIKVs. Virus stocks were prepared as described in Materials and methods (P0). Huh7 cells were infected with MOI = 0.1 of P0 virus, cell culture supernatants were collected 72 h post-infection (P1) and passaged two more times by infection of Huh7 cells (P2–P3) in 72 h intervals. Huh7 cells were then infected using a MOI of 0.01 of P0 and P3 virus, respectively and virus titres were measured at indicated time points by plaque assay.

Figure 3

Construction and characterization of synZIKV-R2A reporter virus genomes. (A) Schematic representation of the synZIKV-R2A reporter virus genomes. For both strains the R2A reporter cassette (light red) was inserted into the wild-type pFK-synZIKV plasmids via MLuI/KpnI restriction sites. The NotI/NruI sites flanking the RLuc gene allow for the exchange of the reporter gene. (B) Immunofluorescence analysis of VeroE6 cells transfected with synZIKV-R2A in vitro transcripts. Cells were grown on coverslips, fixed 72 h and 96 h after transfection and stained with E-specific antibody (green). Nuclear DNA was counterstained with DAPI (grey). Scale bar = 15 μm. (C) Replication kinetics of the synZIKV-R2A reporter viruses in VeroE6 cells. After electroporation (EPO) cells were harvested at given time points and RLuc activity was determined. Values were normalized to the 4 h-value reflecting transfection efficiency. Mean ± SEM of three independent experiments is shown. Replication deficient mutants containing two mutations affecting the active site of the RNA-dependent-RNA polymerase in NS5 (GAA) served as negative controls. (D) VeroE6 cells were transfected with synZIKV-R2A RNAs, cell culture supernatants were collected 72 h post- transfection (P0) and passaged three times by infection of VeroE6 cells (P1-P3) in 72 h intervals. Culture supernatants obtained from each passage were used to inoculate Huh7 cells. In the case of supernatant obtained directly from transfected VeroE6 cells (P0), Huh7 cells were inoculated with undiluted (undil) or 1:10 diluted supernatant. After 72 h cells were harvested and RLuc activity in cell lysates was determined. Mean ± SEM from two independent experiments is shown. (E) Virus titres as determined by plaque assay for each synZIKV-R2A passage; values are mean ± SEM of two independent experiments. (F) Stability of the reporter gene. SynZIKV-R2A viruses released into culture supernatants were harvested after each passage as described in panel D, RNA was isolated and the region encompassing the RLuc coding sequence was amplified by using random hexamer primers for reverse transcription and specific primers for subsequent PCR. The ~1350 bp long DNA fragment in the P0 virus sample corresponds to the reporter gene, while the ~250 bp long fragment corresponds to the WT sequence. (G) Antiviral assay using synZIKV-R2A viruses. VeroE6 cells were inoculated with a 1:10 dilution of a P0 stock and one hour later the medium was replaced with DMEM containing the indicated amount of 2′CMC. RLuc activity was measured in cell lysates 72 h post-infection. Mean ± SEM from two independent experiments is shown.

Figure 4

Construction and characterization of synZIKV-FP635 reporter viruses suitable for live cell imaging. (A) Schematic representation of the synZIKV-FP635 reporter genomes. The FP635 gene fused at the 3′ end to the coding sequence of the SV40 NLS (not indicated) was inserted into the synZIKV constructs via NotI/NruI restriction sites. (B) Detection of E-antigen by immunofluorescence analysis of VeroE6 cells 96 h post-transfection with synZIKV-FP635 RNAs. The FP635 signal (red) was detected by its fluorescence. Note the accumulation of FP635 in distinct nuclear sites, most likely corresponding to nucleoli. Nuclear DNA was counterstained with DAPI (grey). Scale bar = 15 μm. (C) Quantification of E- and FP635-positive VeroE6 cells 96 h post-transfection of synZIKV-FP635 RNAs. Results show the mean from two independent experiments ± SEM. At least 150 cells per condition were counted.

Figure 5

Properties of synZIKV sub-genomic reporter replicons. (A) Schematic representation of the synZIKV-sgR2A subgenomic reporter replicons. The reporter cassette (grey) was inserted into the synZIKV genomes via the MLuI and AgeI restriction sites and replaces the region encoding the structural proteins. (B) RLuc activity in Huh7 cells transfected with wild-type or replication-deficient (mutant GAA) synZIKV-sgR2A replicon RNAs measured at given times post-transfection. Shown RLuc values were normalized to the 4 h value to correct for transfection efficiency. Mean ± SEM of three independent experiments is presented. (C) Western blot showing the abundance of ZIKV NS3 and NS4B proteins in Huh7 cells transfected with synZIKV-sgR2A replicon RNAs. Cells were lysed at indicated times post-transfection and ZIKV-specific antibodies were used to detect viral proteins. β actin served as loading control. Numbers on the left refer to the positions of marker proteins that are given in kilodalton (kDa). (D) Immunofluorescence analysis of Huh7 cells 48 h post-transfection of synZIKV-sgR2A RNAs. Cells were stained with a dsRNA- (green) and a NS3-specific antibody (red). Nuclear DNA was stained with DAPI (grey). Scale bars = 15 μm. Boxed areas indicate regions that are shown in the left panels as enlargements.