Arbovirus high fidelity variant loses fitness in mosquitoes and mice

Abstract

The error rate of RNA-dependent RNA polymerases (RdRp) affects the mutation frequency in a population of viral RNAs. Using chikungunya virus (CHIKV), we describe a unique arbovirus fidelity variant with a single C483Y amino acid change in the nsP4 RdRp that increases replication fidelity and generates populations with reduced genetic diversity. In mosquitoes, high fidelity CHIKV presents lower infection and dissemination titers than wild type. In newborn mice, high fidelity CHIKV produces truncated viremias and lower organ titers. These results indicate that increased replication fidelity and reduced genetic diversity negatively impact arbovirus fitness in invertebrate and vertebrate hosts.

Fig. 1.

(A) Generation of RNA mutagen-resistant CHIKV. WT CHIKV was serially passaged in 100 μM ribavirin (white), 10 μg/mL 5-FU (gray), or no compound (black) at 0.1 pfu/cell. p0 shows the initial WT stock titer. Mean titers of harvests ± SD are shown (Student's t test, n = 3, *P < 0.05, **P < 0.001). (B) Schematic of mutagen passage series showing appearance of nsP4 C483Y mutation. Selected replicates from the 20 triplicate passages were sequenced by flanking CHIKV nsP4 483. WT CHIKV is shown in white, the emergence of Y is represented by black/white flasks, and replacement of WT by C483Y is shown in black. (C and D) Resistance of CHIKV WT and C483Y to mutagens. Matched titers of WT (black line) and C483Y (gray line) were inoculated in triplicate in HeLa cells with increasing concentrations of ribavirin (C) or 5-FU (D). Mean titers ± SD are shown (Student's t test, n = 3, *P < 0.05, **P < 0.001).

Fig. 2.

(A and B) Mutation frequencies of CHIKV WT and C483Y. Five independent stocks per virus were generated in DMEM (A) and two independent replicates were generated in 200 μM ribavirin (B). A mean of 65 partial E1 sequences (≈55,550 nucleotides⋅replicate⁻¹) were obtained. The mean mutation frequencies (number of polymorphisms per 10,000 nt sequenced) ± SEM represent the sum of all replicates; the same pattern of lower mutation frequency for C483Y (gray bar) than for WT (black bar) was observed in each replicate. (X², n = 5 (A), n = 2 (B); **P < 0.001, *P < 0.1). (C–E) Genome-wide deep sequencing of CHIKV WT and C483Y. Nine cDNA libraries spanning the CHIKV genome were sequenced. For each cDNA library (open circle), the percentage of the total quality-filtered reads presenting 0 (C), 1 (D), or 2 (E) mutations compared with consensus were plotted as a function of C483Y (x axis) and WT (y axis). The dashed line delineates where equal numbers of mutations for WT and C483Y would fall. The percentage values ± SD for all libraries combined are noted below graphs. (F–I) One-step growth curves and RNA synthesis. BHK and C6/36 cells were inoculated in triplicate with CHIKV WT (black) or C483Y (gray) at 1 pfu/cell and supernatant titers (F and G) were determined by Vero plaque assay and genome copy measures (H and I) were determined from the same samples by quantiative RT-PCR. Mean values ± SD and P values (repeated measures ANOVA, n = 3) are shown.

Fig. 3.

Competition assays comparing relative fitness of CHIKV WT and C483Y. (A) Direct assays: Viruses were mixed at a 1:1 ratio and inoculated in triplicate into BHK or C6/36 at 0.1 pfu/cell for three passages, at which point nsP4 483 was sequenced. The abundance of each competitor was measured visually as the height of the nucleotide encoding either WT (G nt) or C483Y (A nt) in sequencing chromatograms. (B) Indirect assays: Each virus was mixed in a 1:1 ratio with the marked competitor presenting a silent SacII restriction site and inoculated in triplicate in BHK or C6/36 at 0.1 pfu/cell. The progeny RNA was RT-PCR amplified, and restriction fragment length polymorphism assays were performed to determine the abundance of each competitor. Fitness is represented as the output:input ratio of the CHIKV nsP4 483 variant:marked competitor. A fitness value >1 indicates that WT or C483Y is more fit than marked competitor. Mean values ± SD are shown (Student's t test, n = 3, **P < 0.001). For comparison, viruses adapted to BHK and C6/36 cells via serial passage are included (16).

Fig. 4.

Infection of WT CHIKV and C483Y in invertebrate and vertebrate hosts. (A–D) CHIKV titers in mosquitoes: A. aegypti from the Rockefeller colony (A and B) or a F₂₃ colony from Kedougou, Senegal (C and D), were assayed 7 d after ingestion of infectious bloodmeals containing ≈5 log₁₀ pfu/mL CHIKV WT or C483Y. Back titrations of bloodmeals and single mosquitoes harvested after feeding showed no significant differences in ingested titers. CHIKV pfu equivalents in homogenized bodies and legs/wings (A and C) and saliva samples (B and D) were measured by using quantitative RT-PCR. The limit of detection (0.3 log₁₀ pfu equivalents) is shown as 0. Dots represent individual mosquito titers, lines show mean cohort titers (Student's t test, n = 16 for Rockefeller, n = 9 for Senegal; *P < 0.05, **P < 0.01, ***P < 0.001). (E) CHIKV tissue titers in mice: 8-d-old C56BL/6 mice were inoculated s.c. with 160 pfu of CHIKV WT (black) or C483Y (gray) and organs were sampled on day 5. Dot show individual organs; lines indicate means (Student's t test, n = 4; ***P < 0.0001) The y intercept denotes the limit of detection (1.9 log₁₀ pfu/mL). (F) Mean CHIKV viremias in mice inoculated with 560 pfu/mouse of WT or C483Y. Day 0 values show inocula measured by back titration. (Student's t test, n = 5; ***P < 0.0001).