Zika virus (ZIKV) is a medically important member of the Flaviviridae family that cycles between insect (Aedes mosquitoes) and mammalian (primate) hosts. Understanding what governs viral fitness in each host is crucial for developing approaches to prevent disease. Now, a study in Nature Microbiology describes how deep mutational scanning (DMS), by speeding up virus evolution, can rapidly identify key determinants of host tropism and potentially aid vaccine development.

Credit: Konstantin Nechaev/Alamy

In DMS, a cDNA library is generated that encodes every possible amino acid (or codon) variant for a protein (or protein region) of interest. Variants are then selected for in a given environment and can be identified by deep sequencing and computational analysis. Setoh et al. applied DMS to the 204 C-terminal residues of the ZIKV E protein, which is involved in receptor-mediated infection, and the resulting library was transfected into mosquito cells (C3/36) or primate cells (Vero). Two highly enriched variants were detected in C3/36 cells (K316Q and S461G) and three in Vero cells (Q350L, T397S and a synonymous substitution at R416).

To investigate how the mosquito-selected substitutions affected viral replication in mammalian cells, the authors created a double mutant virus strain, 316Q/461G. Whereas the replication efficiency of 316Q/461G was similar to that of wild-type (WT) virus in C3/36 cells (which grow at 28 oC), it was substantially lower than WT virus in Vero cells and all other mammalian cell lines tested (which grow at 37 oC). Subsequent analysis revealed that 316Q/461G could bind to and infect mammalian cells but a temperature-sensitive defect in virion formation prevented efficient viral replication, secretion and spread. Furthermore, 316Q/461G replicated less efficiently in induced pluripotent stem cell (iPSC)-derived human brain organoids, which developed similarly to mock-infected organoids. By contrast, organoids infected with WT virus displayed abnormalities consistent with previous studies.

Next, Setoh et al. showed that 316Q/461G replication was attenuated in IFNAR–/– mice, which lack the innate antiviral response and are therefore susceptible to viral infection. However, 316Q/461G infection elicited a protective immune response against a lethal dose of the African ZIKV MR766 strain. Taken together, these observations suggest that 316Q/461G has the potential to be developed as a vaccine, even for pregnant or immunocompromised individuals. Importantly, the substitutions were stably maintained in both mosquito and mammalian cells after 5 passages and the virus can be grown to high titres in C3/36 cells, further supporting its vaccine potential.

This study demonstrates that DMS can — within a few days — identify not only amino acids with host-specific roles in replication but also virus strains with the potential to be developed as vaccines.