Host alternation of chikungunya virus increases fitness while restricting population diversity and adaptability to novel selective pressures

Abstract

The mechanisms by which RNA arboviruses, including chikungunya virus (CHIKV), evolve and maintain the ability to infect vertebrate and invertebrate hosts are poorly understood. To understand how host specificity shapes arbovirus populations, we studied CHIKV populations passaged alternately between invertebrate and vertebrate cells (invertebrate ↔ vertebrate) to simulate natural alternation and contrasted the results with those for populations that were artificially released from cycling by passage in single cell types. These CHIKV populations were characterized by measuring genetic diversity, changes in fitness, and adaptability to novel selective pressures. The greatest fitness increases were observed in alternately passaged CHIKV, without drastic changes in population diversity. The greatest increases in genetic diversity were observed after serial passage and correlated with greater adaptability. These results suggest an evolutionary trade-off between maintaining fitness for invertebrate ↔ vertebrate cell cycling, where maximum adaptability is possible only via enhanced population diversity and extensive exploration of sequence space.

FIG. 1.

Alternately passaged populations experience the greatest global increases in fitness. Relative fitness was determined by competition assays comparing p7 CHIKV populations obtained from serial passage (BHK, HeLa, or C6/36) or alternating passage (C6/36 ↔ BHK, C6/36 ↔ HeLa) with their p1 progenitors. The progenitor of both alternating series was the C6/36 p1 population. Competitions were performed on BHK (black bars), C6/36 (white bars), or HeLa (gray bars) cells. Triplicate wells were infected with a 1:1 mix of p1 or p7 populations and the marked clone reference. Forty-eight hours after infection, the output/input ratio of p1 or p7 populations/marked clone was determined. Fitness is represented as the ratio of passages/marked clone. A value of 1 indicates equal fitness; values greater than 1 show that the passage is more fit than the marked clone. Error bars show standard deviations. Asterisks denote significant differences in mean fitness of passaged competitors versus that of their p1 progenitor (Student's t test, *, P < 0.05; **, P < 0.001).

FIG. 2.

Increases in genetic diversity are greatest when no alternating passages are imposed. CHIKV RNAs from supernatants of passages were reverse transcribed, and amplicons flanking the E1 gene were cloned into TOPO vectors and sequenced. The mutation frequency (average number of mutations per E1 gene) was calculated by dividing the number of nucleotide polymorphisms in all clones by the number of nucleotides sequenced and then multiplying by the CHIKV E1 length (1,317 nt). When the same polymorphism was present in multiple RNAs in the population, it was counted each time that it occurred. The E1 gene was sequenced a mean of 71 times (∼90,000 nucleotides/population). Each frequency was adjusted by subtracting the background error rate, estimated from mutations in TOPO-cloned CHIKV plasmid DNA. Numbers above the lines denote P values (χ² test with Yates' correction) comparing frequencies.

FIG. 3.

Alternately passaged CHIKV populations contain more viable genomes than serially passaged populations. (A) The mutation frequency for molecular clones, representing all RNAs in the supernatant, was compared to that for RNAs isolated from plaques, representing only replication-competent RNAs. The molecular clone/plaque clone ratio (indicated above the bars) is calculated by dividing the mutation frequency in molecular clones by that for plaque clones from the same population. A mean of 91,189 (range, 41,773 to 141,805) nucleotides were sequenced for molecular or plaque clone RNAs. (B) The genome/PFU ratio was determined by dividing the number of CHIKV genomes measured by quantitative RT-PCR (4 replicates) by the number of PFU in passages of known titer, determined by plaque assay (duplicate titrations). Differences in ratios were compared using the χ² test with Yates' correction; line above the bars shows P value between populations.

FIG. 4.

Genetic distance increases most after serial passage in vertebrate cells. The percentage of clones representing individual RNA molecules that contain 0, 1, or ≥2 mutations compared to the consensus was measured across the same 717-bp region of the E1 gene for all populations. The mean number of RNAs sequenced per population was 94 (range, 63 to 124), corresponding to 67,577 nt/population. Numbers above the bars denote P values comparing the distribution of clones with 0, 1, and ≥2 mutations in the p1 versus p7 populations (Mann-Whitney test).

FIG. 5.

Increased genetic diversity and distance of serially passaged populations correlates with increased resistance to antiviral compounds. The percentage of particles resisting antiviral compound treatment on HeLa cells was calculated by dividing the mean CHIKV titer from wells containing the antiviral cocktail by the mean titers in wells lacking antivirals multiplied by 100. HeLa cells were pretreated with an antiviral compound cocktail (100 μM ribavirin, 50 μg/ml 5-fluorouracil, and 50 μM azacytidine) or were mock treated with cocktail-free medium, and then CHIKV from each population was adsorbed to duplicate wells (0.1 PFU/cell) for 2 h in cocktail-free medium. After absorption, treated wells were overlaid with the antiviral compound cocktail. Viruses recovered 72 h after inoculation were assayed by titration. Error bars denote standard deviations. Numbers above the lines denote statistical differences in the percentage of resistant particles (χ² test with Yates' correction).

FIG. 6.

Serially passaged populations better resist neutralization than alternately passaged populations. The percentage of particles resisting neutralization on cell types in which passages were conducted was calculated by dividing the number of CHIKV plaques in wells with a standard concentration of neutralizing antibody by the number of plaques in wells without neutralizing antibody and multiplying by 100. Numbers above the lines denote P values comparing frequencies (χ² test with Yates correction).