IFIT1 is rapidly evolving and exhibits disparate antiviral activities across 11 mammalian orders

Abstract

Mammalian mRNAs possess an N7-methylguanosine (m7G) cap and 2'O methylation of the initiating nucleotide at their 5' end, whereas certain viral RNAs lack these characteristic features. The human antiviral restriction factor IFIT1 recognizes and binds to specific viral RNAs that lack the 5' features of host mRNAs, resulting in targeted suppression of viral RNA translation. This interaction imposes significant host-driven evolutionary pressures on viruses, and many viruses have evolved mechanisms to evade the antiviral action of human IFIT1. However, little is known about the virus-driven pressures that may have shaped the antiviral activity of IFIT1 genes across mammals. Here, we take an evolution-guided approach to show that the IFIT1 gene is rapidly evolving in multiple mammalian clades, with positive selection acting upon several residues in distinct regions of the protein. In functional assays with 39 IFIT1s spanning diverse mammals, we demonstrate that IFIT1 exhibits a range of antiviral phenotypes, with many orthologs lacking antiviral activity against viruses that are strongly suppressed by other IFIT1s. We further show that IFIT1s from human and a bat, the black flying fox, inhibit Venezuelan equine encephalitis virus (VEEV) and strongly bind to Cap0 RNAs. Unexpectedly, chimpanzee IFIT1, which differs from human IFIT1 by only 8 amino acids, does not inhibit VEEV infection and exhibits minimal Cap0 RNA-binding. In mutagenesis studies, we determine that amino acids 364 and 366, the latter of which is rapidly evolving, are sufficient to confer the differential anti-VEEV activity between human and chimpanzee IFIT1. These data suggest that virus-host genetic conflicts have influenced the antiviral specificity of IFIT1 across diverse mammalian orders.

Figure 1.. Evolutionary analysis of IFIT1 from clades of mammals reveals rapid evolution.

A. PAML analysis on NCBI-derived coding sequences that were aligned and then analyzed using CodeML. Likelihood ratio tests were performed to compare model 7 vs model 8 and model 8 vs model 8a to determine the presence of positive selection. N=number of sequences input, and numbers under “positively selected sites” represent residue number of the reference sequence. B. FUBAR analysis on aligned coding sequences using HyPhy software and DataMonkey. N=number of sequences input, and numbers under “positively selected sites” represent residue number of the reference sequence. C. MEME analysis on aligned coding sequences using HyPhy software and the DataMonkey application. N=number of sequences input, and numbers under “sites of episodic positive selection” represent residue number of the reference sequence. D. Diagram of Primate IFIT1 domains and location of positively selected sites determined by PAML, FUBAR, and MEME.

Figure 2.. Mutagenesis of rapidly evolving residue 193 in human IFIT1 reveals mutational resiliency.

A. Depiction of the human IFIT1 protein with the TPR4 loop location (orange) shown. B. Solved crystal structure of IFIT1 bound to RNA with TPR4 loop forming a “lid” over the exit of the RNA-binding tunnel (PDB:5udj). C. Zoom in on the TPR4 loop and RNA-binding tunnel exit illustrating the location of the TPR4 loop and residue 193 in relation to bound RNA in the human IFIT1 crystal structure. D. (Top) Saturating mutagenesis screen in cells expressing human IFIT1 with residue 193 mutated to every possible residue and challenged with VEEV infection. (Top) (Bottom) Western blot from lysates of Huh7.5 cells expressing IFIT1 point mutants.

Figure 3.. IFIT1 ortholog screens reveal extensive heterogeneity in mammalian IFIT1 antiviral function.

A. TimeTree illustrating the evolutionary relationship of IFIT1 orthologs selected for ortholog screen. Scale bar represents divergence time of 10 million years. B. Graph of protein sequence identity of IFIT1 orthologs used in screen relative to human IFIT1. C. Dot plot representing the relative infection (compared to control cells) from an ectopic overexpression screen in which Huh7.5 cells expressing 39 different IFIT1 mammalian orthologs were challenged with 1.0 MOI VEEV-GFP for 4 h. Infectivity was quantified by flow cytometry. Red line denotes 30% relative infection. n=2 independent experiments. D. Same as C, for VSV-GFP (1 MOI, 4 h infection). n=3 independent experiments. In C, D, silhouettes represent the species with a relative inhibition of at least 70% in order from most (top) to least (bottom) inhibitory..

Figure 4.. Validation of selected IFIT1s from ortholog screen.

A. TimeTree illustrating the evolutionary relationship of IFIT1 orthologs selected for screen validation and follow-up. Scale bar represents divergence time of 10 million years. B. Western blot from cells expressing HA-tagged IFIT1 orthologs used in C-F. Image is representative of three independent replicates. Quantification of band intensity was performed using LiCOR Image Studio C. Infection of IFIT1 ortholog-expressing Huh7.5 cells with VEEV-GFP (MOI 2, 4h) D. Infection of IFIT1 ortholog-expressing Huh7.5 cells with VSV-GFP (MOI 2, 4h) E. Infection of IFIT1 ortholog-expressing Huh7.5 cells with PIV3-GFP (MOI 2, 10h) F. Infection of IFIT1 ortholog-expressing Huh7.5 cells with SINV-GFP (MOI 2, 10h). G. Heat map summarizing the infection data shown in C-F and complementary ortholog expression data shown in B. In C-F, data represent mean ± SD, n=3 independent experiments. Statistical significance was determined by one-way ANOVA with Dunnett’s test. ns p > 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Figure 5.. IFIT1 proteins exhibit species-specific Cap0 RNA binding.

A. RNA EMSA with Cap0 VEEV RNA probes (41nt of VEEV TC-83 strain 5’ UTR) at 50nM incubated with increasing concentrations of the indicated purified IFIT1 proteins. Images are one representative image from three independent replicates. B. Plot of the band shift intensity from EMSAs in A. Data represent mean ± SD, n=3 independent experiments. Band intensity was quantified by ImageLab software (BioRad). C. Area Under the Curve (AUC) analysis calculated from raw data in B. Statistical significance was tested by one-way ANOVA with correction for multiple comparisons. Data represent mean ± SD, n=3 independent experiments. ns, p > 0.05, *, p < 0.05, **, p < 0.01, ***, p < 0.001, and **** p < 0.0001.

Figure 6.. Mutagenesis uncovers genetic determinants of primate IFIT1 antiviral function.

A. Diagram and chart describing amino acids that differ between human and chimpanzee IFIT1 as well as their domain location. B. Effects of primate IFIT1 mutant expression on VEEV-GFP infection. Huh7.5 cells were infected with an MOI of 2 for 4 h and infectivity was quantified by flow cytometry. Data represent mean ± SD, n=3 independent experiments. Statistical significance was determined by one-way ANOVA with Dunnett’s test. ns, p > 0.05; ***, p < 0.001; and **** p < 0.0001. Western blotting for HA-tag (IFIT1) and GAPDH (Loading control) was also performed to determine expression levels of IFIT1 mutants. C. ClustalOmega protein sequence alignment of residues 361–367 of IFIT1 from 20 primate species. D. Effects of primate IFIT1 double or triple mutants on VEEV-GFP infection. Huh7.5 cells were infected with an MOI of 2 for 4 h and infectivity was quantified by flow cytometry. Data represent mean ± SD, n=3 independent experiments. Statistical significance was determined by one-way ANOVA with Dunnett’s test. ns, p > 0.05; and **** p < 0.0001. Western blotting for HA-tag (IFIT1) and GAPDH (Loading control) was also performed to determine relative IFIT1 mutant protein abundance. E. Structure of human IFIT1 bound to RNA (left) or AlphaFold predicted structure of chimpanzee IFIT1 (right) visualizing location of residues 364 and 366 (red) that confer antiviral activity. Graphics were generated using ChimeraX. F. Zoom in of (E).