Figure 1 shows the structure of the recombinant DENV-2 infectious clone and the amino acid sequence of DENV-2 NS1. A series of mutants of NS1 (three Cys sites [C4A, C55A, C291A], two N-linked glycosylation sites [N130A, N207A], a glutamic acid site [E173A], and a C-terminal truncated fragment [ΔC]), were constructed and incorporated into the infectious DNEV-2 clone. Mutant RNAs were synthesized in vitro and transfected into BHK-21 cells. Mutants N130A and E173A showed no significant diff erences compared to DENV-2 TSV01 (WT) viruses, but C4A, C55A, C291A, and ΔC each failed to cause CPE in transfected cells and did not generate any visible plaques (Figure 2A). In particular, the N207A mutation, which was altered at the second NS1 glycosylation site, produced fewer infectious viral particles and resulted in smaller plaques. Two possibilities might be a dvanced to account for the phenotype of no plaque generation: the first is that C4A, C55A, C291A, and ΔC mutant RNA could not generate any viral particles; the secondis that although mutant RNAs were able to generate infectious particles that could not form plaques. To further investigate the pr oduction of infectious viral particles following transfection, supernatant solutions harvested from transfected cells were used for the infection of BHK-21 cells, which was then followed by immunofluorescent assay (IFA) analysis using an anti-E mAb at 24-h post infection. Consistently, no positive IFA response could be detected for C4A, C55A, C291A, or the ΔC mutant (Figure 2B), demonstrating that all of these mutations impair the production of infectious viral particles. In addition, N130A, N207A, and E173A mutants each showed an approximately 5-20-fold reduction in terms of the specific infectivity compared to WT RNA (Table 1).
Figure 1. Schematic diagram of DENV2-TSV01 and the amino acid sequence (aa807-1159) of the NS1 protein. Systematic site-directed mutagenesis of the N-glycosylation sites and Cys sites is shown by arrows. The C-terminal region of the DENV-2 NS1 protein (aa333-352) that was deleted to generate the C-terminal truncated mutation (ΔC) is underlined. Dotted lines denote the formation of disulfide bonds between Cys residues.
Figure 2. A: Plaque morphology of mutant and WT viruses in BHK-21 cells; B: IFA of Vero cells 24 h after infection with mutant and WT viruses. Red indicates E protein in infected Vero cells (N130A, E173A, N207A and WT). The IFA scale bar represents 20 μm. Other mutants (C4A, C55A, C291A and ΔC) generated no visible fluorescence.
RNA titera CPE N130A 2.80 × 107 + N207A 1.20 × 107 + E173A 5.45 × 107 + C4A < 102 (undetectable) - C55A < 102 (undetectable) - C291A < 102 (undetectable) - ΔC < 102 (undetectable) - WT 2.60 × 108 +
Table 1. Determination of wild type and all mutants titer
For mutants N130A, E173A, and N207A, we also analyzed growth kinetics on Vero cells. As shown in Figure 3, no significant differences between the mutant and WT viruses were observed in terms of viral titer at 24-, 48-, or 72-h p.t. Only at 96-and 120-h p.t. did the mutant viruses show slightly slower growth than WT viruses. Similar results were observed in qRT-PCR analyses of the supernatant solutions harvested from cells infected with mutants N130A, E173A, N207A, and with the WT (data not shown).
Figure 3. Comparison of viral growth kinetics on Vero cells. Cells were infected in 12-well plates with mutant and WT viral stocks (MOI of 0.1). Following sample collection (100 μL; 24, 48, 72, 96, and 120 h), virus titers were determined by plaque assays in BHK-21 cells.
For mutants C4A, C55A, C291A, and ΔC, we attempted to select for any revertants by blind passaging of the supernatant solution from transfected cells. However, even in the supernatant solution from the first passage, the intracellular viral RNAs were below the limit of detection. The RNA copies that were detect able in the P0 supernatant solution may have been residual input RNAs (Table 2). Collectively, these results demonstrated that the C4, C55, C291, and C-terminal residues of NS1 are essential for viral replication.
Mutants P0 (supernatant) P1 (supernatant) RNA Copies Ct RNA Copies Ct C4A 5.84×103 27.10 < 103
31.54 C55A 9.09×103 26.42 < 103
31.20 C291A 2.61×104 24.78 < 103
31.83 ΔC 2.04×105 21.64 < 103
30.55 Negative control < 103
34.24 < 103
Table 2. RNA copy analysis for the mutants
To further determine whether the defect in viral replication was due to the abolition of RNA synthesis, a luciferase-reporting replicon of DENV-2 (RlucRep) TSV01 strain was used to quantify viral translation and RNA synthesis with respect to the NS1 mutants. All the replicons generated a similar level of luciferase activity at 2-6 h p.t. (Figure 4), indicating that none of the mutations affected the translation of the input replicon RNA and that they all possessed similar transfection efficiencies. N130A and E173A revealed typical two-peak luciferase patterns while N207A replicated a bit slowly when compared with WT, replication of other mutants were declined significantly in cells. Taken together, this was consistent with the results obtained using the genome-length RNAs. These results indicated that the Cys sites at C4, C55, and C291 and the C-terminal residues of NS1 are required for viral RNA synthesis.
Figure 4. Analysis of the mutant and WT replicons based on DENV2-TSV01-Rep. Equal amounts of viral replicon RNA were electroporated into BHK-21 cells. At the indicated time points, the luciferase signals were measured. Each data point represents the average of three replicates. Error bars show standard deviations.
The results showed that, as in the case of WT RNA, mutations N130A, E173A, and N207A each gave rise to an NS1 band corresponding to about 90 kDa, representing the dimeric form of NS1. In contrast, no NS1 expression could be observed in cells transfected with C4A, C55A, C291A, or ΔC mutants, on account of viral replication disruption. In the case of the N130A and N207A mutant viruses, it was notable that the NS1 dimer migrated slightly faster during electrophoresis than the NS1 dimer of either the WT or E173A viruses, most likely due to deglycosylation at the corresponding positions (Figure 5). Although the level of NS1 expression in N207A was less than in N130A, E173A, or WT viruses, it is clear that glycosylation at positions 130 and 207 is not essential for viral replication.
Figure 5. Western blot analysis of NS1 in transfected cells. Cells transfected with each viral genome-length RNA were lysed, and the intracellular NS1 proteins were probed with an anti-NS1 antibody. The samples were treated with 1×sample buffer without reducing reagents at 37℃ (A), or incubated with 1×sample buffer containing reducing reagent (2-mercaptoethanol) at 100℃ for 4 min (B) before electrophoresis. The dimeric and monomeric forms of NS1 are shown.