Hepatitis A virus adaptation to cellular shutoff is driven by dynamic adjustments of codon usage and results in the selection of populations with altered capsids

Abstract

Hepatitis A virus (HAV) has a highly biased and deoptimized codon usage compared to the host cell and fails to inhibit host protein synthesis. It has been proposed that an optimal combination of abundant and rare codons controls the translation speed required for the correct capsid folding. The artificial shutoff host protein synthesis results in the selection of variants containing mutations in the HAV capsid coding region critical for folding, stability, and function. Here, we show that these capsid mutations resulted in changes in their antigenicity; in a reduced stability to high temperature, low pH, and biliary salts; and in an increased efficacy of cell entry. In conclusion, the adaptation to cellular shutoff resulted in the selection of large-plaque-producing virus populations.

Importance: HAV has a naturally deoptimized codon usage with respect to that of its cell host and is unable to shut down the cellular translation. This fact contributes to the low replication rate of the virus, in addition to other factors such as the highly inefficient internal ribosome entry site (IRES), and explains the outstanding physical stability of this pathogen in the environment mediated by a folding-dependent highly cohesive capsid. Adaptation to artificially induced cellular transcription shutoff resulted in a redeoptimization of its capsid codon usage, instead of an optimization. These genomic changes are related to an overall change of capsid folding, which in turn induces changes in the cell entry process. Remarkably, the adaptation to cellular shutoff allowed the virus to significantly increase its RNA uncoating efficiency, resulting in the selection of large-plaque-producing populations. However, these populations produced much-debilitated virions.

FIG 1

Workflow of the processes of adaptation to changing cellular shutoff conditions, including forward adaptation to 0.05 μg/ml of AMD (F0.05NA still-nonadapted, F0.05A adapted, and F0.05LA long-adapted populations) and to 0.2 μg/ml of AMD (F0.2NA still-nonadapted, F0.2A adapted, and F0.2LA long-adapted populations), as well as the reversion to absence (R0NA still-nonadapted and R0A adapted populations) and 0.05 μg/ml (R0.05NA still-nonadapted and R0.05A adapted populations) of AMD, respectively.

FIG 2

Changes in capsid folding measured through the analysis of the antigenic structure of the different populations selected during the processes of adaptation to changing AMD concentrations. (A through D) Ratio between virus recognition by MAb H7C27 and that by MAb K24F2. (E through H) Ratio between virus recognition by MAb K34C8 and thab by MAb K24F2. Populations tested are described in Fig. 1. Values represent the mean ± standard error of three different virus stocks. Statistically significant differences (P < 0.001) between pairs of populations within each panel are indicated by different letters; populations with the same letter are not significantly different.

FIG 3

Changes in capsid folding measured through the analysis of the stability under extreme conditions of the different populations selected during the processes of adaptation to changing AMD concentrations. (A through D) Virus stability after 5 min at 61°C. (E through H) Virus stability after 1 h at pH 2. (I through L) Virus stability after 4 h in the presence of 1% bile salts. Stability under a particular treatment was measured as the loss of virus infectivity due to this treatment, and it is expressed as the log₁₀ reduction. Populations tested are described in Fig. 1. Values represent the mean ± standard error of three different virus stocks. Statistically significant differences (particular P levels are described in the text) between pairs of populations within each panel are indicated by different letters; populations with the same letter are not significantly different.

FIG 4

Efficiency of the cell entry process of the different populations selected during the processes of adaptation to changing AMD concentrations. (A through D) Percent binding to FRhK-4 cells. (E through H) Uncoating time 50 (UT₅₀), or the required time for an uncoating of 50% of the virus population, in hours. Populations tested are described in Fig. 1. Values represent the mean ± standard error of three different virus stocks. Statistically significant differences (particular P levels are described in the text) between pairs of populations within each panel are indicated by different letters; populations with the same letter are not significantly different.

FIG 5

Plaques of the populations adapted to different AMD concentrations. (A) L0; (B) F0.05A; (C) F0.05LA; (D) F0.2A; (E) F0.2LA; (F) mock-infected cells. Populations are described in Fig. 1.

FIG 6

Anticodon usage variation in the capsid region of populations adapted to different conditions of cellular shutoff compared with the parental type L0 population. Codons pairing with anticodons abundantly used by the cell are rarely used by parental type HAV (rare codons), while codons pairing with anticodons not abundantly used by the cell are commonly used by the HAV parental type. Codon usage deoptimization (left column) is defined by a decreased use of codons pairing with anticodons abundantly used by the cell, which are replaced by codons pairing with anticodons not abundantly used by the cell. In contrast, codon usage optimization (right column) is defined by an increased use of codons pairing with anticodons abundantly used by the cell, which correlates with a decrease of use of codons pairing with anticodons not abundantly used by the cell. A total of 63 anticodons exist; thus, a total of 50 anticodons in F0.05A, 38 in F0.05LA, 43 in R0A, 46 in F0.2A, 30 in F0.2LA, and 32 in R0.05A populations do not vary their use in respect to the L0 population. Populations are described in Fig. 1.