In vitro and in vivo characterization of a West Nile virus MAD78 infectious clone

The viral determinants governing the varied neuropathogenicity of different West Nile virus (WNV) strains are poorly understood. Here, we generated an infectious clone (WNV-MADIC) of the non-pathogenic strain WNV-MAD78 and compared its replication to that of parental WNV-MAD78 and a WNV-MAD78 infectious clone (WNV-MADTX-UTRs) containing the 5′ and 3′ untranslated regions (UTRs) of the pathogenic strain WNV-TX. All three viruses replicated at similar rates and caused similar lethality in mice. Thus, the infectious clone is indistinguishable from parental virus in replication and neurovirulence, and the UTRs alone do not account for the increased virulence of WNV-TX compared to WNV-MAD78.

and then used as a template for reverse transcription (RT) with an oligo-dT primer. The 3 0 end of the resulting cDNA was PCR-amplified using the oligo-dT primer and a sense primer corresponding to position 10743 within the WNV-MAD78 genome. Sequence analysis identified 20 nucleotides immediately adjacent to the added polyA nucleotides that aligned with the 3 0 end of another lineage 2 strain of WNV, WNV-956, suggesting that this sequence represented the exact 3 0 end of WNV-MAD78. To confirm the sequence of the WNV-MAD78 3 0 UTR, we repeated the amplification of this region using the 10743 primer and an antisense primer complimentary to the exact 3 0 end of the WNV-MAD78 genome. Sequence analysis of the resulting RT-PCR product identified a total of 97 nucleotides that were not reported previously ( Fig. 1a and accession number KJ909513). Alignment of the complete 3 0 UTR sequence of WNV-MAD78 with that of other WNV strains indicated that the WNV-MAD78 3 0 UTR diverges from most lineage 1 and 2 strains (Fig. 1b). In contrast, the WNV-MAD78 5 0 UTR is highly conserved (Fig. 1b).
To generate a full-length infectious clone of WNV-MAD78, we utilized the strategy outlined in Fig. 1c. First, total RNA extracted from WNV-MAD78-infected Vero cells was used as template for RT-PCR to generate eight WNV-MAD78 cDNAs spanning the entire virus genome (Fig. 1c, Stage 1). During PCR, an MluI restriction site followed by the T7 promoter was inserted immediately upstream of the 5 0 end of the genome (Fig. 1d). Several other restriction sites were engineered at various locations within the cDNAs to facilitate cloning but were not incorporated into the final construct. Additionally, a NotI restriction site was inserted at the 3 0 end of the genome b Fig. 1 Generation of a WNV-MAD78 infectious clone. (b) Eight-to twelve-week-old C57BL/6 or Ifnar -/-mice were infected with WNV-MAD78 or WNV-MAD IC (n = 7) by subcutaneous injection of 100 PFU into the left rear footpad. Mice were monitored daily and euthanized when body weight loss was[20 % or they reached clinical scores of 4 or above (1, no paresis; 2, mild paresis; 3, frank paresis; 4, severe paresis; 5, true paresis; 6, moribund) Generation of a WNV-MAD78 infectious clone 3115 (Fig. 1d). Each cDNA was blunt-end ligated into the cloning vector pVL-blunt (a kind gift from Vincent Lee) [8]. In Stage 2, the eight WNV fragments were subcloned as indicated in Fig. 1c to generate four plasmids encoding the entire genome. The full-length infectious clone (pWNV-MAD IC ; accession number KJ909514) was generated by assembling the four WNV segments into the very-low-copy plasmid pWSK29 [9] as outlined in Fig. 1c, Stage 3. Sequencing of the final full-length construct identified four silent mutations compared to the previously published WNV-MAD78 sequence (Fig. 1e). To determine if these mutations were introduced during the cloning process, we sequenced the corresponding regions of our WNV-MAD78 stock. Of the four mutations noted, three were present in our WNV-MAD78 stock (Fig. 1e), indicating that only one (T5339C) arose during the cloning process.
To generate virus, linear pWNV-MAD IC was used as template for in vitro transcription, and 12 lg of the resulting RNA was used to transfect 1 9 10 6 Vero cells using the NeonÒ Transfection System (Invitrogen) set to the following conditions: 1150 V, 20 ms, and 2 pulses. After seven days, culture supernatant containing WNV-MAD IC was collected and subsequently passaged once in Vero cells to generate a working stock. The presence of a cytosine nucleotide at position 5339 within the recovered virus (Fig. 1e) confirmed that it was WNV-MAD IC and not parental WNV-MAD78.
To assess the biological properties of WNV-MAD IC , we compared the growth kinetics of the recovered virus to that of the parental strain. Vero cells were infected with WNV-MAD78 or WNV-MAD IC and infectious particle production assessed by plaque assay (Fig. 2a). Similar levels of infectious particles were observed at all times, with peak  (Fig. 2a). Moreover, plaques for both viruses developed at similar rates on Vero cells and were visible at 6 days postinoculation. To assess virulence, wild-type C57BL/6 and interferon receptor knockout (Ifnar-/-) mice, which are resistant and highly susceptible, respectively, to WNV-MAD78 [7], were inoculated subcutaneously with 100 PFU of WNV-MAD78 or WNV-MAD IC (Fig. 2b). No differences in survival were observed. In wild-type mice, neither strain caused mortality or weight loss, though one mouse inoculated with WNV-MAD IC showed mild signs of disease. In contrast, in the absence of IFN signaling, infection with either strain was 100 % lethal. Thus, the virus recovered from the infectious clone was indistinguishable from the parental strain.
Recently, Suthar et al. generated a similar infectious clone, herein referred to as WNV-MAD TX-UTRs , which contains the 5 0 and 3 0 UTRs of a virulent WNV strain, WNV-TX [10]. The WNV-TX UTRs have notable differences from those of WNV-MAD78. Within the 5 0 UTR, WNV-TX contains an extra adenosine at position 51, and the WNV-TX 3 0 UTR is 78 nucleotides longer and shares only 73.2 % sequence identity with that of WNV-MAD78 (Fig. 1b). Specific sequences and structural folds within the UTRs are critical for WNV replication [11][12][13][14][15], and previous work suggests that the UTRs are virulence determinants [16]. Thus, sequence differences among the UTRs of various WNV strains may alter the structure and function of these regions and thereby influence the pathogenicity of the resulting virus. To assess the effect of the TX-UTRs on WNV-MAD78 fitness, we compared WNV-MAD78, WNV-MAD IC , and WNV-MAD TX-UTRs replication in A549 cells, human brain cortical astrocytes (HBCAs), and C6/36 cells, models for WNV infection within the periphery, the neurovascular unit, and the mosquito vector, respectively. In all cell types, similar titers were observed for all three viruses throughout the infection (Fig. 3a-c), consistent with the initial characterization of WNV-MAD TX-UTRs within A549 cells [10]. Thus, the TX-UTRs are neither beneficial nor detrimental to WNV-MAD78 replication in vitro. However, we consistently observed that WNV-MAD TX-UTRs plaques on Vero cells were larger than those of WNV-MAD IC or parental WNV-MAD78 (Fig. 3d), suggesting that the presence of TX-UTRs may marginally enhance viral spread in some mammalian cells.
When inoculated in the periphery, parental WNV-MAD78 and WNV-MAD TX-UTRs are completely avirulent in mice [10]. In contrast, approximately 60 % of mice inoculated intracranially with parental WNV-MAD78 succumb to infection [17], making intracranial (IC) inoculation a good model for assessing virulence. To determine if the WNV-TX UTRs affected WNV-MAD78 virulence, we compared WNV-MAD78, WNV-MAD IC , and WNV-MAD TX-UTRs in 4-week-old outbred Swiss Webster mice (Harlan Laboratories) inoculated intracranially with 10 pfu of virus or a sham PBS control. All viruses exhibited similar lethality (Fig. 3e, p = 0.766), with 80-90 % of infected animals succumbing to infection by 9 days after inoculation. Morbidity correlated directly with mortality, and no animals that lost weight or showed signs of disease recovered (data not shown). Thus, the virulence phenotypes of the infectious clones are indistinguishable from each other and from that of parental WNV-MAD78.
In summary, we have generated an infectious WNV-MAD78 clone (WNV-MAD IC ) that contains the authentic UTRs of the viral RNA and is indistinguishable from parental WNV-MAD78 in replication and neurovirulence. Replacing the WNV-MAD78 UTRs with those of a pathogenic strain of WNV had no effect on the biological phenotype of WNV-MAD78. This finding is consistent with a study of another low-pathogenicity WNV strain, Kunjin virus, in which replacing both the Kunjin UTRs with those of WNV-NY did not enhance virulence in mice [16], though replacing the 5 0 UTR alone did increase pathogenicity. Thus, divergence among the UTRs of various WNV strains may not be a determining factor in virulence, as long as they are maintained as a cognate pair.