New insights into the evolutionary rate of hepatitis B virus at different biological scales

Abstract

The evolutionary rates of hepatitis B virus (HBV) estimated using contemporary sequences are 10(2) to 10(4) times higher than those derived from archaeological and genetic evidence. This discrepancy makes the origin of HBV and the time scale of its spread, both of which are critical for studying the burden of HBV pathogenicity, largely unresolved. To evaluate whether the dual demands (i.e., adaptation within hosts and colonization between hosts) of the viral life cycle affect this conundrum, the HBV quasispecies dynamics within and among hosts from a family consisting of a grandmother, her 5 children, and her 2 granddaughters, all of whom presumably acquired chronic HBV through mother-to-infant transmission, were examined by PCR cloning and next-generation sequencing methods. We found that the evolutionary rate of HBV between hosts was considerably lower than that within hosts. Moreover, the between-host substitution rates of HBV decreased as transmission numbers between individuals increased. Both observations were due primarily to changes at nonsynonymous rather than synonymous sites. There were significantly more multiple substitutions than expected for random mutation processes, and 97% of substitutions were changed from common to rare amino acid residues in the database. Continual switching between colonization and adaptation resulted in a rapid accumulation of mutations at a limited number of positions, which quickly became saturated, whereas substitutions at the remaining regions occurred at a much lower rate. Our study may help to explain the time-dependent HBV substitution rates reported in the literature and provide new insights into the origin of the virus.

Importance: It is known that the estimated hepatitis B virus (HBV) substitution rate is time dependent, but the reason behind this observation is still elusive. We hypothesize that owing to the small genome size of HBV, transmission between hosts and adaptation within hosts must exhibit high levels of fitness trade-offs for the virus. By studying the HBV quasispecies dynamics for a chain of sequentially infected transmissions within a family, we found the HBV substitution rate between patients to be negatively correlated with the number of transmissions. Continual switching between hosts resulted in a rapid accumulation of mutations at a limited number of genomic sites, which quickly became saturated in the short term. Nevertheless, substitutions at the remaining regions occurred at a much lower rate. Therefore, the HBV substitution rate decreased as the divergence time increased.

FIG 1

Phylogenetic relationships of HBV sequences derived from the study family. Red and blue lineages are sequences from the second and third serum samples, respectively, of the individuals. Deduced NGS consensus sequences are labeled. The neighbor-joining tree was constructed using the Kimura two-parameter substitution model implemented in MEGA 5.0 (21) and was rooted by HBV-C (accession no. NC_003977; not shown). Numbers close to the nodes are for 1,000 bootstrap replications. For comparison purposes, HBV genomes with the highest identity score (99%) were included, and they were all distantly related to sequences derived from the study family.

FIG 2

Time of HBV divergence (in years) versus genetic distance derived from cloning sequences. The genetic distances were derived from the whole genome (A) and nonsynonymous (B) and synonymous (C) sites of the nonoverlapping regions. Diamonds show HBV divergence within hosts. Solid lines represent linear regressions performed with all data, with correlation coefficients (R²) of 0.22 (P < 10⁻⁵; Pearson correlation), 0.24 (P < 10⁻⁶), and 0.19 (P < 10⁻⁴) for panels A, B, and C, respectively. Dotted lines represent linear regressions performed without the diamond points (between-host comparisons), with R² values of 0.00 (P = 0.56), 0.01 (P = 0.29), and 0.09 (P < 10⁻²) for panels A, B, and C, respectively. The dashed line in panel A is based on power regression (R² = 0.50; P < 10⁻¹⁴). The regression equations for panel A are D = 0.0046 × 4.95 × 10⁻⁵ × T for linear regression and D = 0.0013 × T^0.425 for power regression, where D is the genetic distance and T is the time of divergence, in years.

FIG 3

Numbers of transmissions versus HBV substitution rates derived from cloning sequences. HBV substitution rates were derived from the whole genome (A) and nonsynonymous (B) and synonymous (C) sites of the nonoverlapping regions. The substitution rates within hosts (number of transmissions = 0) were all significantly higher (P < 10⁻⁴; Wilcoxon rank sum test) than those between hosts (number of transmissions = 1, 2, 3, or 4). The substitution rate between hosts decreased as the transmission number increased in both panels A (τ = −0.45; P < 10⁻⁷; Kendall's rank correlation) and B (τ = −0.26; P < 10⁻²), whereas the substitution rate was almost constant in panel C (τ = −0.08; P = 0.33). +, P < 0.1; *, P < 0.05; **, P < 10⁻²; ***, P < 10⁻⁵ (Wilcoxon rank sum test).

FIG 4

Genetic variations in the HBV genome at different evolutionary scales. A sliding window of genetic variations among HBV-A, -B, -C, and -D genotypes (HBV-A/B/C/D; black line), within HBV genotype B (HBV-B; blue line), and among individuals within the study family (red line) was used. The genomic structure of HBV is shown at the bottom. P, polymerase; C, core protein; preC + C, precore protein; X, X protein; S1 + S2 + S, large surface protein; S2 + S, medium surface protein; S, surface protein. The thin gray lines indicate variations within individuals. The window size was 100 nucleotides, and the step size was 10 nucleotides. The 3′ half of the X ORF and the 5′ region of the precore ORF, which span nucleotides 1680 to 1850, were excluded from this analysis. The genetic variations within the family were derived from the NGS data set.

FIG 5

Colonization-adaptation trade-off model of HBV evolution. The intensity of the gray background represents the strength of host immunity against HBV, which is weak during the tolerance/noninflammatory phase and increases in the clearance/inflammatory phase. The blank areas in circles represent viral strains with high replicative ability, whose proportions decrease as host immunity increases. The colored areas in circles represent mutants that escaped from or adapted to host immune selection, whose proportions increase with elevated host immunity. Solid arrows indicate evolvement of HBV within individuals, and dashed arrows show transmission between individuals. In most cases, the substitution rates (μ) of HBV were measured from T1 to C1 or from C1 to C2, which included many changes not contributing to long-term evolution. Therefore, the μ value was inflated. As we compared HBV sequences with many transmissions between hosts, i.e., T1 to T4 or C1 to C4, such an effect could be minimized. See the text for details.