JOURNAL OF VIROLOGY, Apr. 2010, p. 3454–3463 0022-538X/10/$12.00 doi:10.1128/JVI.02164-09 Copyright © 2010, American Society for Microbiology. All Rights Reserved.
Vol. 84, No. 7
Distinct Hepatitis B Virus Dynamics in the Immunotolerant and Early Immunoclearance Phases䌤† Hurng-Yi Wang,1 Ming-Hung Chien,1 Hsiang-Po Huang,2 Hsiao-Chi Chang,1 Chung-Che Wu,3 Pei-Jer Chen,1 Mei-Hwei Chang,1,4,5* and Ding-Shinn Chen1* Institute of Clinical Medicine, National Taiwan University, Taipei, Taiwan1; Department of Medical Research, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan2; School of Medicine, National Taiwan University, Taipei, Taiwan3; Department of Pediatrics, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan4; and Hepatitis Research Center, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan5 Received 13 October 2009/Accepted 7 January 2010
Little is known about hepatitis B virus (HBV) diversity changes within a host during the immunotolerant phase of chronic HBV infection. Such knowledge, nevertheless, may help in understanding how host immunity and HBV interact at the early stage of infection. In this study, serial serum samples were collected from a long-term (>17 years) follow-up cohort of seven patients, and multiple copies of the full-length viral genome from serially sampled sera were recovered and analyzed. Viral genetic diversity was positively correlated with host immunity, represented by levels of alanine aminotransferase (ALT), but was negatively correlated with the viral copy number. During the immunotolerant phase, when the host immunity was feeble (ALT < 20 U/liter), viral nucleotide diversity decreased while copy numbers increased. Rates of evolutionary change derived for different patients were in a very narrow range (1.6 ⴛ 10ⴚ5 to 5.4 ⴛ 10ⴚ5/site/year). As the disease progressed toward the immunoclearance phase (ALT > 20 U/liter), viral diversity increased but copy numbers decreased. Evolutionary rates varied among patients in accordance with their levels of ALT, ranging from 9.6 ⴛ 10ⴚ6 to 3.2 ⴛ 10ⴚ4/site/year. More than half (19/32 sites) of positively selected sites resided in immune epitopes, suggesting their possible role in host immunity. Our results demonstrate that host immunity is a dominant factor in HBV evolution. Different selective forces, including immune-mediated positive selection and virus-mediated negative selection, operate in tandem in shaping viral population dynamics within a host. release of another virion, was proposed as another factor determining the long-term substitution rate (18). The HBV generation time was estimated to range from 10 to 100 days, with an average of 24.8 days (28), which is 6 and 10 times longer than those of HCV (27) and HIV (31), respectively. Therefore, many features of viral dynamics in CHI may differ from those of other viral diseases. In addition, CHI is distinctive from other chronic viral diseases by having a prolonged immune tolerance phase without an acute reaction if an individual is infected perinatally. During the early stage of CHI, HBV quasispecies are not or are only minimally subjected to host immune selection (3, 6). In teenaged patients, CHI enters an immunoclearance phase with an active host immune response to control viral replication (5). Studies of CHI found associations of viral variants, including nucleotide and structural variations such as insertions and deletions, with disease progression (4). However, previous studies focused on the relatively later period of infection, e.g., during active hepatitis close to or after hepatitis Be antigen (HBeAg) seroconversion (usually after the age of 20 years if infected perinatally) or during the development of cirrhosis or hepatocellular carcinoma (HCC) (usually after the age of 40 years) (37). The evolution of HBV in the immunotolerant phase during the first 2 decades within a host is largely unknown. To study the evolution of viral populations in this stage of CHI, we selected a cohort of carriers who have been monitored regularly for more than 17 years, beginning before the age of 10 years. For each patient, multiple full-length HBV sequences
In the study of chronic viral disease, the viral diversity change is established as an important factor in the pathogenesis of hepatitis C virus (HCV) and human immunodeficiency virus (HIV) (14, 23, 30, 34). Because hepatitis B virus (HBV) uses reverse transcriptase for replication, the error-prone nature of polymerase makes the rate of evolutionary change much closer to those of RNA viruses than to those of DNA viruses (21). From this aspect, it was proposed that the evolutionary dynamics of HBV should be similar to those of HCV and HIV (48). Nevertheless, little is known about HBV dynamics in chronic HBV infection (CHI), especially in the early stage of infection. While mutations introduced during replication are an inherent and common driver of all viral variations, the viral generation time, defined as the average length of time from the release of a virion until it infects another cell and causes the
* Corresponding author. Mailing address for Mei-Hwei Chang: Department of Pediatrics, National Taiwan University Hospital, National Taiwan University, 7 Chung-Shan South Road, Taipei 100, Taiwan. Phone: 886-2-23125131. Fax: 886-2-23114592. E-mail:
[email protected]. Mailing address for Ding-Shinn Chen: Institute of Clinical Medicine, National Taiwan University Hospital, National Taiwan University, 7 Chung-Shan South Road, Taipei 100, Taiwan. Phone: 886-2-23123456, ext. 67086. Fax: 886-223709820. E-mail:
[email protected]. † Supplemental material for this article may be found at http://jvi .asm.org/. 䌤 Published ahead of print on 20 January 2010. 3454
VOL. 84, 2010
HOST IMMUNITY AND HBV DIVERSITY
3455
TABLE 1. Clinical features and HBV DNA levels of the seven patients Value or description for patient: Characteristic A
Sex Age at enrollment (first sampling) (yr) Second sampling age (yr) Average ALT level (U/liter) during first and second samplings Third sampling age (yr) Average ALT level (U/liter) during second and third samplings SD of ALT Mean (SD) HBV DNA (log10 copies/ml)
B
C
D
E
F
G
Male 7.4 16.8 17.0
Male 6.2 14.6 12.5
Male 8.9 15.9 7.9
Male 8.9 13.8 8.5
Female 8.8 14.8 10.8
Female 5.4 12.4 13.7
Female 9.5 16.1 16.8
26.3 74.0
26.3 28.6
27.5 21.9
26.4 20.4
25.8 27.0
23.4 21.1
27.5 34.5
45 7.80 (0.72)
10.5 8.11 (0.61)
10.9 7.04 (0.42)
were recovered from three time points, spanning the immunotolerant phase without alanine aminotransferase (ALT) elevation to the early immunoclearance phase with mild ALT elevation. We estimated the viral nucleotide diversity, mutation rate, and effective population size within hosts between the two stages. Based on our results, a scenario of HBV evolution during the early stage of infection is proposed. MATERIALS AND METHODS Study subjects. Seven patients were selected from the Department of Pediatrics, National Taiwan University Hospital (NTUH). Among them, six patients
7.2 8.50 (0.19)
13.4 8.36 (0.08)
6.3 8.09 (0.30)
17 7.02 (0.85)
were presumed to have acquired HBV perinatally, as their mothers were hepatitis B surface antigen (HBsAg) positive at the time of enrollment. Only patients with genotype B were used to ensure that differences found in viral evolution were not due to genotypic differences. These patients have been monitored for an average of 18.3 years, starting before the age of 10 years. Their ALT, HBsAg, HBeAg, anti-HBs, and anti-HBe levels were monitored regularly at 6-month intervals. A guideline was followed for ALT measurement to ensure that the measurements were consistent during the last 2 decades (see the supplemental material for details). These patients were all HBsAg and HBeAg positive, with no HBeAg seroconversion observed during the study. No antiviral therapy was administered throughout the course of this study. All patients were enrolled by community surveys before their becoming aware of CHI. Therefore, this cohort is ideal for studying HBV evolution under natural conditions. We selected the
FIG. 1. Profiles of ALT levels (shaded region; left y axis) and nucleotide diversity values () (red line; right y axis) derived for different ages (x axis) of the seven patients in this study.
3456
WANG ET AL.
J. VIROL.
following three time points to survey the evolutionary changes in HBV quasispecies: I, the time of enrollment; III, the most recent clinical visit; and II, one time point between points I and III, before the elevation of ALT (see Results). HBV cloning and sequencing. For each patient, HBV DNA was extracted, and the virus titer was quantified from 100 l of serum sample at the core facility of NTUH. We first amplified the approximately 3.2-kb HBV genome. The PCR conditions and primers used are listed in Table S1 in the supplemental material. Amplification was performed with Stratagene Easy-A High Fidelity PCR cloning enzyme (Stratagene) and 5 ng of DNA template in a 50-l reaction mix. PCR products were purified with a PCR-M purification kit (Viogene, Taipei, Taiwan), cloned into a Topo XL PCR cloning vector (Invitrogen), and transformed into competent cells according to the manufacturer’s protocols. Clones were picked for a PCR check, and 12 positive clones per cloning reaction were subjected to sequencing. We designed six internal primers for full-length HBV sequencing (see Table S1 in the supplemental material). The sequencing reactions were performed using a Big Dye Terminator kit and then loaded into a model 3730 xl sequencer (Applied Biosystems) at the core facility of NTUH. The base positions were numbered according to a reference sequence (GenBank accession no. NC_003977). Sequence analyses. Sequences were assembled using the Seqman program of the Lasergene package (Lasergene, Madison, WI), with visual inspection, and then aligned using Clustal_W (38). Viral nucleotide diversity () was calculated by custom C⫹⫹ software based on the Libsequence package (39). represents the number of variable nucleotide sites in a sample of sequences divided by sequence length. For nonoverlapping regions of HBV, was calculated separately for silent (synonymous; S) and amino acid-altering (nonsynonymous; A) sites. The evolutionary rate (), given as the number of changes per site per year, and the viral effective population size (Ne) were estimated using an established Bayesian Markov chain Monte Carlo (MCMC) approach (11) that incorporates the time of viral sampling. This approach estimates the rate of nucleotide change and dynamics of the population genetic diversity from serially sampled sequence data (10). In addition, an estimate of the posterior distribution of genealogies related to the sequence data was concurrently inferred for each alignment. From these distributions, the times to the origin for viruses isolated from each patient at each time point (the time to the most recent common ancestor [tMRCA]) were obtained. The analyses were performed using the GTR⫹⌫ model of nucleotide substitutions with a relaxed molecular clock model supplemented in BEAST (11). We performed two independent runs with 2 ⫻ 106 MCMC steps, of which the first 10% were discarded as burn-in. The results were compared to confirm that both sampled the same distribution and then were combined. The rate of HBV evolution () was also estimated by comparing mean pairwise nucleotide differences within and between DNA samples recovered from different time points (16). Test for positive selection. To detect positive selection, two approaches were applied. Selection can be detected by comparing numbers of synonymous changes per synonymous site (Ks) and numbers of nonsynonymous changes per nonsynonymous site (Ka). The ratio Ka/Ks, or , is used to assess selection. A Ka/Ks ratio of ⬎1 indicates positive selection, whereas a Ka/Ks ratio of ⬍1 indicates negative selection. To identify the putative amino acid residues under positive selection, two pairs of likelihood-based models implemented in the PAML package (46) were used. The first pair includes M1a (nearly neutral) and M2a (positive selection) (45), while the second pair includes M7 (beta) and M8 (beta and ) (46). The result was given after running different models in codeml. Because some parts of the HBV genome encode more than one protein via different open reading frames (ORFs), these regions are not suitable for the Ka/Ks-based test. Therefore, we applied the codon-based test only to nonoverlapping coding regions of the HBV genome. Amino acid sites experiencing positive selection in which a mutation to one particular amino acid is advantageous can also be inferred when a mutant accumulates in the population faster than expected by random genetic drift. We carried out a computer simulation to identify these sites and to estimate the selective coefficient. The details of the simulation are given in the supplemental material. In short, we assumed that HBV’s behavior could be described by a discrete-generation model. Supposing that from time t1 to t2 the frequency of a mutant changes from f1 to f2, then under conditions of drift, the probability can be formulated as follows: P(time ⫽ t2 ⫺ t1 兩 ft1⫽ f1, ft2⫽ f2, n ⫽ Ne) (26).
RESULTS Patterns of changes in viral diversity at different stages of CHI. Clinical information for all patients is listed in Table 1
TABLE 2. List of nucleotide diversity for the seven patients at different time points Parameter and patient
Indicated value (%) at time point I
II
III
0.257 0.237 0.330 0.186 0.113 0.361 0.262 0.008
0.196 0.206 0.140 0.103 0.096 0.309 0.145
0.508 0.227 0.453 0.227 0.309 0.185 0.422 0.021
Ab A B C D E F G P value vs time point II
0.323 0.242 0.268 0.054 0.027 0.215 0.288 0.032
0.134 0.161 0.099 0.030 0.063 0.214 0.161
0.497 0.188 0.296 0.269 0.296 0.054 0.377 0.021
Sb A B C D E F G P value vs time point II
0.191 0.384 0.479 0.289 0.288 0.575 0.654 0.344
0.383 0.672 0.472 0.213 0.335 0.192 0.096
0.710 0.384 0.958 0.288 0.670 0.671 0.574 0.023
a
A B C D E F G P value vs time point II
a
Nucleotide diversity derived from the entire HBV genome. Nucleotide diversities of synonymous (S) and nonsynonymous (A) sites derived from nonoverlapping regions of the HBV genome. b
and includes the age at time of enrollment and HBV sampling, the average ALT levels between different sampling points, and virus titers. The mean HBV DNA loads were all higher than 107 copies/ml, indicating active viral replication. In general, the average ALT level between the second and third points was twofold greater than that between the first and second points. The ALT profiles during the follow-up and the nucleotide diversities () at different sampling points are shown in Fig. 1. Although varied in different individuals, a general trend emerged when the values were aligned. Specifically, except for patient F, values at the second point were lower than those derived from the first (P ⫽ 0.008; Wilcoxon signed rank paired test) and third (P ⫽ 0.021) points (Table 2). Namely, HBV diversity was high at a “relative early” stage of infection, decreased during the immune tolerance phase, and rose again toward the immunoclearance phase with elevated ALT levels (Fig. 1). Nucleotide diversities at synonymous (S) and nonsynonymous (A) sites in nonoverlapping regions of the HBV genome were estimated separately. While A was significantly reduced (P ⫽ 0.032) from the first to the second points, S was not (P ⫽ 0.344). From the second to the third points, both A and S significantly increased, with P values of 0.021 and 0.023, respectively. Several studies proposed that selective pressures exerted by
VOL. 84, 2010
HOST IMMUNITY AND HBV DIVERSITY
3457
FIG. 3. Negative correlation between HBV DNA level (log10 HBV DNA/ml) and nucleotide diversity (P ⫽ 0.016; tau ⫽ ⫺0.381; Kendall’s rank sum correlation).
FIG. 2. Correlations of average ALT levels and nucleotide diversity values () derived from the entire HBV genome (P ⫽ 0.002; tau ⫽ 0.604; Kendall’s rank sum correlation) (A), nucleotide diversity values for nonsynonymous sites (A) of nonoverlapping regions (P ⫽ 0.002; tau ⫽ 0.604) (B), and nucleotide diversity values for synonymous sites (S) of nonoverlapping regions (P ⫽ 0.101; tau ⫽ 0.341) (C).
both host immunity and viral load are important factors influencing viral diversity in chronic viral diseases (14, 25, 36). We next examined the relationships among them. Since cellular immunity assays were not available for this analysis, we used the ALT level as an indicator to reflect host immune activity, as it was shown that ALT level provides some estimate of the strength of an immunological response against viral infection (14, 28). There was only a weak correlation between nucleotide diversities and spot ALT levels at the time the sequence was sampled (P ⫽ 0.112; Kendall’s rank sum correlation) (see Fig. S1 in the supplemental material). Since ALT can be influenced by many factors and since changes in diversity between different time points may result from the cumulative effect of immune activity, a single point measurement might not represent the long-term immune status of a host. We therefore used average ALT values between different time points to assess the correlation. By using the average, we assumed that the stochastic effect could be minimized, and the value itself should represent the long-term trend of host immunity. Because there was no ALT information before enrollment, we plotted only the average ALTs and HBV nucleotide diversities derived from the second and third points. The correlation was highly significant (P ⫽ 0.002 and tau ⫽ 0.604) (Fig. 2A). To further decide whether this correlation was contributed by all sites in the genome or by specific sites or regions, we assessed the correlation for nucleotide diversities at synonymous (S) and nonsynonymous (A) sites, separately. The average ALT level was significantly correlated with A (P ⫽ 0.002 and tau ⫽ 0.604) (Fig. 2B), but the correlation between the average ALT level and S was very weak (P ⫽ 0.101) (Fig. 2C). There was a negative correlation between the viral load (log10 copies/ml) and , i.e., the higher the viral load the lower the nucleotide diversity, with P ⫽ 0.016 and tau ⫽ ⫺0.381 (Fig. 3). This negative relationship was also observed when synon-
3458
WANG ET AL.
J. VIROL. TABLE 3. Population parameters for the seven patients Value (HPD) for patient:
Parameter A
B
C
Parameters for the first to the third time points a Pairwise difference method Bayesian MCMC method Ne tMRCA (yr)
1.0 ⫻ 10⫺4 8.6 ⫻ 10⫺5 (4.0 ⫻ 10⫺5, 1.0 ⫻ 10⫺4) 1.7 ⫻ 103 34 (22 to 51)
1.3 ⫻ 10⫺6 8.0 ⫻ 10⫺6 (2.4 ⫻ 10⫺7, 1.9 ⫻ 10⫺5) 4.7 ⫻ 105 174 (29 to 560)
2.7 ⫻ 10⫺5 3.0 ⫻ 10⫺5 (1.3 ⫻ 10⫺7, 7.1 ⫻ 10⫺5) 5.8 ⫻ 104 160 (23 to 510)
Parameters for the first to the second time points a Ne tMRCA (yr)
2.2 ⫻ 10⫺5 (1.0 ⫻ 10⫺7, 5.5 ⫻ 10⫺5) 1.8 ⫻ 104 63 (13 to 679)
1.6 ⫻ 10⫺5 (1.2 ⫻ 10⫺7, 4.1 ⫻ 10⫺5) 1.5 ⫻ 105 125 (12 to 466)
3.0 ⫻ 10⫺5 (1.1 ⫻ 10⫺7, 8.9 ⫻ 10⫺5) 2.8 ⫻ 104 161 (10 to 550)
Parameters for the second to the third time points a Ne tMRCA (yr)
1.3 ⫻ 10⫺4 (2.0 ⫻ 10⫺6, 2.7 ⫻ 10⫺4) 2.9 ⫻ 103 21 (10 to 248)
1.8 ⫻ 10⫺5 (1.9 ⫻ 10⫺7, 4.2 ⫻ 10⫺5) 1.1 ⫻ 105 94 (14 to 327)
5.9 ⫻ 10⫺5 (8.2 ⫻ 10⫺6, 1.0 ⫻ 10⫺4) 2.1 ⫻ 104 33 (13 to 74)
a
is the rate of change, represented as the number of changes per site per year.
ymous and nonsynonymous sites were analyzed separately (see Fig. S2 in the supplemental material). Evolutionary rate and effective population size vary in the two phases. The HBV evolutionary rates () showed great variations among patients (Table 3). Evolutionary rates derived by the pairwise difference method (16) were lowest for patient D (1.1 ⫻ 10⫺6/site/year) and highest for patient G (1.2 ⫻ 10⫺4/site/year). Applying a Bayesian MCMC approach, ranged from 8.0 ⫻ 10⫺6/site/year for patient B to 2.1 ⫻ 10⫺4/site/year for patient G. Most of the 95% highest posterior densities (HPDs) from different patients did not overlap, indicating that variations in among patients were significant. The evolutionary rates yielded by different methods were in good agreement with each other, with a correlation coefficient (R2) of 0.67. We thus consider our estimations to be authentic. The viral effective population size (Ne) was largest in patient B, at 4.7 ⫻ 105, and smallest in patient G, at 1.6 ⫻ 103 (Table 3). The times to the origin for viruses isolated from each patient (tMRCAs) ranged from 20 years (95% HPD, 17 to 23 years) in patient E to 235 years (95% HPD, 23 to 2,051 years) in patient F. Most patients, except patients E and G, had tMRCAs longer than their age at the most recent clinical visit, suggesting that part of the viral diversity might have been inherited from the donor (most likely from the mother through vertical transmission). We next estimated Ne and for different intervals, from the first to the second time points and from the second to the third time points. Because in some cases values calculated from different time intervals by the pairwise difference method were negative due to small differences between different samplings, we applied only the evolutionary rate derived from the Bayesian approach. values from the first and second time points were in a very narrow range, with the lowest in patient B, at 1.6 ⫻ 10⫺5/site/year, and the highest in patient G, at 5.4 ⫻ 10⫺5/site/year, and their 95% HPDs largely overlapped. The effective population sizes ranged from 1.8 ⫻ 104 (patient A) to 1.5 ⫻ 105 (patient B). Furthermore, although the 95% HPDs covered a wide time interval, the tMRCAs were all greater than the ages of the patients at the time of the second sam-
pling. In contrast, rates of change derived from the second to third time points showed great differences among patients, with the highest in patient G, at 3.2 ⫻ 10⫺4/site/year, and the lowest in patient F, at 9.6 ⫻ 10⫺6/site/year. Effective population sizes exhibited ⬎100-fold differences among patients, from the largest of 2.2 ⫻ 105 (patient F) to the smallest of 9.0 ⫻ 102 (patient G). In addition, patients A, E, and G had tMRCAs shorter than their ages at the third sampling. Putative sites under immune selection. Our computer simulation identified 36 nucleotide changes altering 38 amino acids, as two nucleotides changed the amino acids in two reading frames, but covering only 32 amino acid sites, as some of them arose in different individuals independently or two mutations occurred in the same codon, with significantly higher frequencies of change (P ⬍ 0.05) than expected by random genetic drift (see Materials and Methods) (Table 4). Of these amino acid sites, 19 are in known T-cell epitopes (9), 1 causes the precore stop mutation, and the other 12 are in regions with no known immune epitopes. Since the HBV genome contains four ORFs encoding 1,613 amino acids, with 621 of these in known T-cell epitopes, the proportion of putative selected sites residing in epitopes was significantly higher than that in nonepitope regions (P ⫽ 0.014; Fisher’s exact test) (Table 5). In addition, we noted that the nonstructural protein, i.e., polymerase, exhibited smaller numbers of selected sites than structural proteins, i.e., surface and core, with P values of 0.019 and 0.032, respectively. (The X protein was difficult to categorize by this structural and nonstructural dichotomy, as its function is not fully understood [1].) Among 36 mutations, the frequencies of 6 increased between the first and second time points, those of 29 increased between the second and third time points, and that of 1 increased in both intervals (Table 4). Since patient-years from the first to second and from the second to third time points were 49.3 and 78.8 years, respectively, significantly (P ⫽ 0.014) greatly accelerated amino acid changes between the second and third time points correlated well with elevated ALT levels during this period. The selection coefficients of different mutants ranged from 0.007 to 0.036, with an average of 0.020.
VOL. 84, 2010
HOST IMMUNITY AND HBV DIVERSITY
3459
TABLE 3—Continued Value (HPD) for patient: D
E
F
G
1.1 ⫻ 10⫺6 1.2 ⫻ 10⫺5 (2.5 ⫻ 10⫺7, 2.6 ⫻ 10⫺5) 1.4 ⫻ 105 111 (23 to 357)
2.9 ⫻ 10⫺5 8.1 ⫻ 10⫺5 (4.3 ⫻ 10⫺5, 1.2 ⫻ 10⫺4) 3.0 ⫻ 103 20 (17 to 23)
4.4 ⫻ 10⫺5 1.2 ⫻ 10⫺5 (1.1 ⫻ 10⫺7, 4.9 ⫻ 10⫺5) 3.9 ⫻ 105 235 (23 to 2,051)
1.2 ⫻ 10⫺4 2.1 ⫻ 10⫺4 (1.1 ⫻ 10⫺4, 5.0 ⫻ 10⫺4) 1.6 ⫻ 103 22 (19 to 25)
3.2 ⫻ 10⫺5 (2.0 ⫻ 10⫺7, 8.9 ⫻ 10⫺5) 2.6 ⫻ 104 61 (5 to 204)
2.8 ⫻ 10⫺5 (1.9 ⫻ 10⫺7, 7.2 ⫻ 10⫺5) 4.4 ⫻ 104 43 (6 to 159)
5.2 ⫻ 10⫺5 (4.4 ⫻ 10⫺7, 1.3 ⫻ 10⫺4) 4.1 ⫻ 104 93 (9 to 306)
5.4 ⫻ 10⫺5 (4.2 ⫻ 10⫺7, 1.4 ⫻ 10⫺4) 2.0 ⫻ 104 46 (7 to 167)
2.7 ⫻ 10⫺5 (1.6 ⫻ 10⫺7, 4.5 ⫻ 10⫺5) 7.4 ⫻ 104 57 (14 to 146)
1.4 ⫻ 10⫺4 (4.5 ⫻ 10⫺6, 3.1 ⫻ 10⫺4) 2.7 ⫻ 103 21 (11 to 42)
9.6 ⫻ 10⫺6 (1.8 ⫻ 10⫺7, 2.7 ⫻ 10⫺5) 2.2 ⫻ 105 189 (21 to 1,542)
3.2 ⫻ 10⫺4 (1.8 ⫻ 10⫺4, 6.7 ⫻ 10⫺4) 9.0 ⫻ 102 13 (12 to 16)
There were five nucleotide changes (altering six amino acids) for which selection coefficients could not be given, because they reached fixation within the host during follow-up. We added a “greater-than” symbol to their coefficients to indicate that their selection coefficients may have been greater than those numbers. Selective amino acid sites were also assessed by the likelihood approach implemented in PAML (Table 6). In total, two likelihood ratio tests recognized nine amino acid sites under selection, with values of ⬎1, and eight of these sites overlapped in the results of our computer simulation. Thus, there were a total of 33 amino sites putatively under positive selection. DISCUSSION Patterns of change in viral diversity from immunotolerant to immunoclearance phase. During a chronic viral infection, such as HCV or HIV infection, the quasispecies are usually largely reduced upon transmission, probably due to host immune selection or severe genetic drift, and gradually restored after the emergence of escape mutants (12, 14). Therefore, the reduction in viral diversity after a decade of CHI is very peculiar. Significant diversity reduction at nonsynonymous rather than synonymous sites resembles the outcome of natural selection, as it is the only evolutionary force which targets a specific region or site. Selection pressures for the virus come from either competition between viral variants, i.e., differences in replicative efficiency, host immune activity, or both. These possibilities, unfortunately, are confounded under most circumstances. By taking advantage of the immune tolerance phase of CHI, we may now dissect their relative contributions in more detail. It is generally assumed that there is no immune selection during the tolerance phase of CHI (6). This notion is partially supported by the observation that very few infiltrative lymphocytes were found in the livers of asymptomatic HBV carrier children infected in the perinatal period (3). In addition, on average, ALT levels in hepatitis-free populations of different age groups were never ⬎20 U/liter in previous surveys (13, 41).
Therefore, low levels of ALT (⬍20 U/liter) before the second sampling suggest inactive or minimal immune activity and, in turn, imply small or no selective pressure from the host. During this phase, viral strains with higher replicative ability increased their copy numbers and outcompeted those carrying amino acid-altering mutations with lower replicative efficiencies. Consequently, nucleotide diversity at nonsynonymous sites (A) was reduced with an increase in viral load. From the second to third time points, the elevated ALT level (⬎20 U/liter) indicated increased host immunity. Positive selection would favor different mutations under host immune pressures (44) and would result in an increased A and accelerated amino acid changes. In addition to selection, nucleotide diversity () can also be modified in accordance with variations in viral population size and the mutation rate (). Since and the viral copy number are negatively correlated (Fig. 3) and a reduced effective population size (Ne) during this period was shown (Table 3), the only factor left is the change in , which is measured in generation time. The HBV generation time is largely decided by the half-lives of infected cells, which range from about 10 to 100 days and are strongly influenced by the strength of a host’s immunity (28). If we assume that the replication error introduced by viral polymerase is constant over time, an almost 10-fold reduction in generation time is equivalent to a 10-fold increase in measured on a yearly basis. Namely, from the second to the third time point, the increased host immunity speeded up the HBV life cycle by shortening its generation time, which in turn accelerated the rate of change and increased viral nucleotide diversity at both nonsynonymous and synonymous (S) sites. Although the above interpretation explains the observed viral diversity changes in CHI during the follow-up, one issue which remains to be determined is when and how the viral diversity at the first time point was generated. It is possible that viral diversity was inherited from the donor (most likely the mother) and gradually decreased during the immune tolerance phase because of viral competition. The estimated times of viral origin (tMRCAs) for most patients were longer than the ages at which HBV was sampled (Table 3), which seems to
3460
WANG ET AL.
J. VIROL. TABLE 4. Summary of selected amino acid sites in different ORFs
Amino acid changea Epitope sites Polymerase 616-630H M617L
Patient
Intervalb
Frequency of change From
To
Selection coefficientc
E
2
0.00
0.42
0.010
A E
2 2
0.00 0.00
0.43 0.42
0.019 0.010
A A
2 2
0.25 0.00
0.86 0.43
0.021 0.019
C
1
0.25
0.71
0.016
C F G
1 2 2
0.33 0.33 0.00
0.71 0.58 1.00
0.025 0.007 ⬎0.036
E
2
0.00
0.42
0.010
F
1
0.08
0.75
0.035
A C G
2 1 2
0.00 0.42 0.00
0.71 0.71 1.00
0.023 0.012 ⬎0.036
A F
2 2
0.00 0.75
0.57 1.00
0.023 ⬎0.013
E
2
0.00
0.42
0.010
A G
2 2
0.00 0.00
0.57 0.58
0.023 0.022
A
2
0.00
0.43
0.019
A G C F G G F F
2 2 1 1, 2 2 2 2 1
0.08 0.00 0.17 0.33 0.00 0.00 0.75 0.17
0.71 0.67 0.71 0.92 1.00 1.00 1.00 0.83
0.024 0.022 0.025 0.018 ⬎0.036 ⬎0.036 ⬎0.013 0.027
Surface I84T L108Vh L182H Y399S
G G G E
2 2 2 2
0.00 0.00 0.00 0.00
0.67 0.67 0.75 0.42
0.022 0.022 0.024 0.010
X protein A85T H86P
A A
2 2
0.00 0.00
0.57 0.43
0.023 0.019
Core W28stop
G
2
0.00
0.92
0.032
Surface 188-196H,C L195S 215-223H,C G218E L233R 271-280C Q275H 298-311H Q303R Q303H T300Ae 389-397H C395Y X protein 92-100C K95Q 115-123C N118T T118Nf N118T 126-140H K130 M V131I M127Ig Core 79-98H I88V G92V 149-168H P164Q P159L 176-185H,C T176A G182C
Nonepitope sites Polymerase K73N L288Rh N480D N480D N480Se C602G K743N R841K
Sequenced
KMCFRKLPVNRPIDW –L––––––––––––– VLQAGFFLL –––––––S– –––––––S– FLGGTPVCL –––E––––– ––––––––R LLDYQGMLPV ––––H––––– CTTPAQGTSMFPSC –––––R–––––––– –––––H–––––––– ––A––––––––––– LLPIFFCLW ––––––Y–– VLHKRTLGL –––Q––––– CVFNEWEEL –––T––––– –––N––––– –––T––––– EVRLKVFVLGGCRHK ––––MI––––––––– –I––––––––––––– PHHTALRQAILCWGELMNLA –––––––––V–––V–––––– VSFGVWIRTPPAYRPPNAPI –––––––––––––––Q–––– ––––––––––L––––––––– TVVVRRRGRSP A––––––C–––
a
C, cytotoxic T-lymphocyte epitope; H, helper T-lymphocyte epitope. Interval 1 is from the first to the second time point, and interval 2 is from the second to the third time point. Selection coefficients with the symbol “⬎” indicate that the mutant reached 100% frequency. d The epitope positions are based on the work of Desmond et al. (9), and the sequences were derived from the consensus amino acid sequence of this study. e An A-to-G change at base 530 caused amino acid changes in the polymerase (N480S) and surface (T300A) proteins. f The consensus amino acid at this position is Asn, but patient C had a high frequency of Thr at the first time point and Asn became dominant at the second time point. g The consensus amino acid at this position is Val, but patient F was polymorphic, with Met and Ile. h A T-to-G change at base 3169 caused amino acid changes in the surface (L108V) and polymerase (L288R) proteins. b c
favor this scenario. However, since the lower bounds of 95% confidence intervals (95% HPDs) of tMRCA for most patients were generally smaller than their ages, the possibility that diversity was generated during HBV colonization cannot be
ruled out. Within 3 to 6 months of the initial infection, HBV can infect 40% to almost 100% of ⬃2 ⫻ 1011 hepatocytes and produce virus (3, 20, 28). This enormous population expansion should also be accompanied by an inflation in viral nucleotide
VOL. 84, 2010
HOST IMMUNITY AND HBV DIVERSITY
TABLE 5. Comparisons of immune-selected sites in different open reading frames and in epitopes versus nonepitope regions No. of selected sites
No. of amino acids
Protein
Within epitopes
Outside epitopes
Within epitopes
Outside epitopes
Total Polymerase Surface X Corec
19 1 7 5 6
12 6 4 2 0
621 239 213 43 126
992 605 188 112 87
Fisher’s exact P value
0.014a 0.019b 0.003b 0.032b
a Test for an excess of selected sites within immune epitopes versus nonepitope regions. b Test for an excess of selected sites against polymerase. c The precore stop mutant was removed from this analysis because it did not fit in the epitope and nonepitope dichotomy.
diversity. The above two scenarios are, of course, not mutually exclusive, and both suggest that the reduction in nucleotide diversity may have occurred long before the follow-up. Nevertheless, their relative significances have important consequences for the long-term adaptation of HBV. The second hypothesis implies that only a small portion of a diverse population was transmitted and that the majority was lost, i.e., there was a genetic bottleneck upon HBV transmission. This may cause implications for preadapted viral strains during colonization. By lowering the chance of their transmission, genetic drift has the potential to prevent the accumulation of adaptive changes in a population (47), thereby impeding the long-term adaptation of HBV. If the diversity in the population is inherited from the donor, such a reduction in fitness is not necessary, and the adaptive mutants have a greater chance of being transmitted (12). Further studies focusing on direct mother-to-infant transmission may help to resolve this issue. Immune selection plays a critical role in HBV evolution during CHI. Contrasting with the previous finding that HBeAg-positive hosts with normal ALT levels have little viral evolution (33, 43), we identified 33 amino acid sites which are potential candidates for immune selection. Although immune-selected mutations typically require the demonstration that a naturally occurring mutated T-cell epitope fails to trigger an immune response, several lines of evidence suggest that they are, one way or another, responsible for virus and host immune interactions. First, significantly more selected sites resided in immune epitopes than in nonepitope regions. Second, compared with the nonstructural (polymerase) protein, more selected sites resided in structural (surface and core) proteins (Table 5).
3461
Finally, the occurrences of accelerated amino acid changes corresponded well with changes in ALT profiles. Since these patients were all HBeAg positive during the time of the study, our analysis provides evidence that immune selection may have occurred long before the observed HBeAg seroconversion. In addition, since the average ALT level for six of seven patients was ⬍40 U/liter but ⬎20 U/liter, these results illustrate the occurrence of host immune selection even with a slight increase in average ALT level. Although the cutoff of 20 U/liter may seem arbitrary, the ALT levels of hepatitis-free populations of different age groups have never exceeded this level (13, 41). To this extent, an ALT level of 20 U/liter may represent the action of host immunity against HBV. Compared to the case for HIV, the evidence for HBV escape mutants due to host immunity is limited (25, 33, 44). This discrepancy is probably rooted in the weaker immune response and longer generation time in the latter than the former. The estimated selection coefficients were all ⬍0.05, and a sequence usually contained more than one mutation in different epitopes, supporting previous observations that T-cell responses to HBV are weak (7, 43) but probably multiply specific (44). Since the generation time of HBV is, on average, nine times longer than that of HIV, then if the population size and selective coefficient are equal, the rate of frequency change for an advantageous mutation would be nine times lower in the former than in the latter. Therefore, multiple sequences derived from long-term follow-up of individuals are necessary to identify such mutants. In addition, because of the compact nature of its genome, it should be reasonable to challenge HBV’s ability to tolerate many mutations simultaneously under multiple selection regimens. In contrast, a small number of changes among several epitopes may help HBV to gain a slight advantage under feeble immune pressure. In order to identify possible escape mutants, whole-genome sequencing is thus necessary. Short-term follow-up which focuses only on the dominant variant or part of the HBV genome without exploring the full spectrum of diversity of viral quasispecies may fail to recognize such mutants. Mutations at the same amino acid sites emerged independently from different hosts in the late stage of this study (the third time point), implying their advantage under immune pressure. However, these mutations were absent from the early stages of this study (the first and second time points), suggesting that their superiority may have either been compromised (44) in the absence of host immune pressure or been reversed under evolving host immunity (24). For example, an N118T (or T118N) mutation in the X protein was due to a transversion of
TABLE 6. Log likelihood values and parameter estimates for nonoverlapping regions of HBVa Model code
pb
l
ts/tv
Parameter estimates
M1a M2a
2 4
⫺4,544.6 ⫺4,536.5
3.19 3.3
M7 M8
2 4
⫺4,545.2 ⫺4,536.1
3.2 3.3
p0 ⫽ 0.681, p1 ⫽ 0.319, 0 ⫽ 0.078 (1 ⫽ 1) p0 ⫽ 0.696, p1 ⫽ 0.293, P ⫽ 0.010, 0 ⫽ 0.108 (1 ⫽ 1), ⫽ 6.065 p0 ⫽ 0.073 p0 ⫽ 0.988, p1 ⫽ 0.012, P ⫽ 0.203, ⫽ 5.793
a
Positively selected site(s)c
P, K743; X, N118, V127; C, W28, P159, P164 P, C602, M617, Y673, K743; X, N118, V127; C, W28, P159, P164
Only nonoverlapping regions of the HBV genome were used. Since surface protein is 100% overlapped with polymerase, it was not included in this analysis. Number of parameters in the distribution. Positive selection sites were identified using Bayes empirical Bayes (BEB) analysis, with those at a probability of ⬎95% shown in bold. c P, polymerase; X, X protein; C, core open reading frame. b
3462
WANG ET AL.
adenine (A) to cytosine (C) at position 1726, which is within the enhancer II region of precore mRNA. It has been shown that mutations at this position not only alter the immune epitope but also change the viral load (8, 17). Therefore, complex selection forces, including immune-mediated positive selection and virus-mediated negative selection, operate in tandem in shaping HBV evolution. Variations in effective population size and evolutionary rate among hosts. In CHI, according to different estimations, 5% to almost 100% of ⬃2 ⫻ 1011 hepatocytes can be infected with about 25 to 50 covalently closed circular DNA (cccDNA) per cell, which translates to at least 1011 to 1013 replicating particles per day (3, 20, 28). Thus, the estimated HBV effective population sizes (Ne), ranging from 103 to 105, are 106- to 1010-fold smaller than the actual size. Several factors may cause a reduction in the effective population size. One of the most important reasons is that continual selection reduces viral Ne. Based on our analysis, both positive and negative selection pressures were detected at different phases of the disease. HBV is apparently under tremendous selective pressure during an infection. Another possible explanation is that HBV populations do not evolve randomly (panmixia) but, rather, with some population structure. There is evidence that HBV populations are spatially structured, because in healthy chronic carriers, HBsAg-positive cells were usually separated from each other, with intervening areas of HBsAg-negative hepatocytes, and rarely occurred in large clumps (2, 22, 32). Therefore, HBV populations might exhibit a metapopulation structure. In this model, HBVs within the host are divided into local populations (hepatocytes here), and extinct populations (via cell death) are recolonized (regenerated cells) by viral particles from other local populations. In CHI, depending on the strength of the host’s immune activity, about 0.3% to 3% of hepatocytes are killed and replenished every day (28, 40). It is known that such patterns of colonization and replacement can dramatically reduce Ne (35, 42), because only small fractions of local populations (hepatocytes) contribute to the long-term intrahost population of HBV. Finally, changes in population size over time are another possible cause of a small Ne. When population size fluctuates, Ne is dominated by the smallest population size in the process, because it is given by the harmonic mean of population sizes at different stages (19). In our study, patient B had an average ALT level that was higher than that of patient E but exhibited a larger Ne. Despite the ALT measurement, however, patient B had a smaller standard deviation, which meant fewer fluctuations in host immune activity, and perhaps in viral population size, and was able to maintain a larger Ne than patients with similar average ALT levels but larger standard deviations, e.g., patient E. It seems likely that a combination of factors acts together to reduce the intrahost effective population size of HBV. Although evolutionary rates estimated for different patients varied dramatically, they actually captured the entire spectrum of HBV mutation rates estimated from previous studies, from the lowest of 1.5 ⫻ 10⫺5/site/year to the highest of 7.7 ⫻ 10⫺4/site/year (15, 29, 48). The results shown in Tables 1 and 3 demonstrate how HBV evolutionary rates varied in accordance with host immune status. Immune pressure may influence
J. VIROL.
HBV evolutionary rates in two ways. First, it can drive advantageous (supposedly escape) mutants to high frequencies or fixation in a short time, which increases the evolutionary rate. Second, an elevated immune reaction accelerates hepatocyte death and replenishment and shortens the HBV generation time, which, in turn, increases the evolutionary rate. Therefore, it is important to have knowledge of the immune status in advance before explaining the estimated rate of change. Moreover, HBV evolutionary rates derived from patients at different disease stages or with different immune statuses should be compared with caution. In summary, the interplay between viral replication and host immunity explains the pattern of HBV dynamics within the host during a relatively early stage of infection. That is, without immune selection, competition between peers increases the viral load and decreases nucleotide diversity; in contrast, host immunity accelerates viral evolution and decreases copy numbers but increases diversity. Our observation, if true, may have important implications for the study of long-term HBV adaptation and therapeutic design for CHI. ACKNOWLEDGMENTS We thank Lu Yun, Eleanore Chen, and Shi-Hui Chen for technical support. We also thank Shiou-Hwei Yeh, Chun-Jen Liu, and Taichung Tseng for constructive comments. This study was supported by NHRI grants to D.-S.C. (NHRI-EX97CD9201), M.-H. Chang (NHRI-EX97-9418BI), and H.-Y.W. (NHRIEX99-9934BI). REFERENCES 1. Bouchard, M. J., and R. J. Schneider. 2004. The enigmatic X gene of hepatitis B virus. J. Virol. 78:12725–12734. 2. Camilleri, J. P., C. Amat, M. Chousterman, J. P. Petite, A. Duboust, A. Boddaert, and A. Paraf. 1977. Immunohistochemical patterns of hepatitis B surface antigen (HBsAg) in patients with hepatitis, renal homografts recipients and normal carriers. Virchows Arch. A 376:329–341. 3. Chang, M. H., L. Y. Hwang, H. C. Hsu, C. Y. Lee, and R. P. Beasley. 1988. Prospective study of asymptomatic HBsAg carrier children infected in the perinatal period: clinical and liver histologic studies. Hepatology 8:374–377. 4. Chen, B. F., C. J. Liu, G. M. Jow, P. J. Chen, J. H. Kao, and D. S. Chen. 2006. High prevalence and mapping of pre-S deletion in hepatitis B virus carriers with progressive liver diseases. Gastroenterology 130:1153–1168. 5. Chen, D. S. 1993. From hepatitis to hepatoma: lessons from type B viral hepatitis. Science 262:369–370. 6. Chen, D. S. 1993. Natural history of chronic hepatitis B virus infection: new light on an old story. J. Gastroenterol. Hepatol. 8:470–475. 7. Chisari, F. V. 1997. Cytotoxic T cells and viral hepatitis. J. Clin. Invest. 99:1472–1477. 8. Chou, Y. C., M. W. Yu, C. F. Wu, S. Y. Yang, C. L. Lin, C. J. Liu, W. L. Shih, P. J. Chen, Y. F. Liaw, and C. J. Chen. 2008. Temporal relationship between hepatitis B virus enhancer II/basal core promoter sequence variation and risk of hepatocellular carcinoma. Gut 57:91–97. 9. Desmond, C. P., A. Bartholomeusz, S. Gaudieri, P. A. Revill, and S. R. Lewin. 2008. A systematic review of T-cell epitopes in hepatitis B virus: identification, genotypic variation and relevance to antiviral therapeutics. Antivir. Ther. 13:161–175. 10. Drummond, A. J., G. K. Nicholls, A. G. Rodrigo, and W. Solomon. 2002. Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data. Genetics 161:1307–1320. 11. Drummond, A. J., and A. Rambaut. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7:214. 12. Edwards, C. T., E. C. Holmes, D. J. Wilson, R. P. Viscidi, E. J. Abrams, R. E. Phillips, and A. J. Drummond. 2006. Population genetic estimation of the loss of genetic diversity during horizontal transmission of HIV-1. BMC Evol. Biol. 6:28. 13. Elinav, E., I. Z. Ben-Dov, E. Ackerman, A. Kiderman, F. Glikberg, Y. Shapira, and Z. Ackerman. 2005. Correlation between serum alanine aminotransferase activity and age: an inverted U curve pattern. Am. J. Gastroenterol. 100:2201–2204. 14. Farci, P., I. Quinti, S. Farci, H. J. Alter, R. Strazzera, E. Palomba, A. Coiana, D. Cao, A. M. Casadei, R. Ledda, R. Iorio, A. Vegnente, G. Diaz, and P. A. Tovo. 2006. Evolution of hepatitis C viral quasispecies and hepatic injury in
VOL. 84, 2010
15. 16. 17.
18.
19. 20.
21. 22.
23.
24.
25.
26.
27.
28.
29.
30. 31.
perinatally infected children followed prospectively. Proc. Natl. Acad. Sci. USA 103:8475–8480. Fares, M. A., and E. C. Holmes. 2002. A revised evolutionary history of hepatitis B virus (HBV). J. Mol. Evol. 54:807–814. Fu, Y. X. 2001. Estimating mutation rate and generation time from longitudinal samples of DNA sequences. Mol. Biol. Evol. 18:620–626. Gu ¨nther, S., N. Piwon, A. Iwanska, R. Schilling, H. Meisel, and H. Will. 1996. Type, prevalence, and significance of core promoter/enhancer II mutations in hepatitis B viruses from immunosuppressed patients with severe liver disease. J. Virol. 70:8318–8331. Hanada, K., Y. Suzuki, and T. Gojobori. 2004. A large variation in the rates of synonymous substitution for RNA viruses and its relationship to a diversity of viral infection and transmission modes. Mol. Biol. Evol. 21:1074–1080. Hartl, D. L., and A. G. Clark. 2007. Principles of population genetics. Sinauer Associates, Sunderland, MA. Hews, S., S. Eikenberry, J. D. Nagy, and Y. Kuang. 2010. Rich dynamics of a hepatitis B viral infection model with logistic hepatocyte growth. J. Math Biol. 60:573–590. Holmes, E. C. 2008. Evolutionary history and phylogeography of human viruses. Annu. Rev. Microbiol. 62:307–328. Hsu, H. C., Y. H. Lin, M. H. Chang, I. J. Su, and D. S. Chen. 1988. Pathology of chronic hepatitis B virus infection in children: with special reference to the intrahepatic expression of hepatitis B virus antigens. Hepatology 8:378–382. Kearney, M., F. Maldarelli, W. Shao, J. B. Margolick, E. S. Daar, J. W. Mellors, V. Rao, J. M. Coffin, and S. Palmer. 2009. Human immunodeficiency virus type 1 population genetics and adaptation in newly infected individuals. J. Virol. 83:2715–2727. Leslie, A. J., K. J. Pfafferott, P. Chetty, R. Draenert, M. M. Addo, M. Feeney, Y. Tang, E. C. Holmes, T. Allen, J. G. Prado, M. Altfeld, C. Brander, C. Dixon, D. Ramduth, P. Jeena, S. A. Thomas, A. St. John, T. A. Roach, B. Kupfer, G. Luzzi, A. Edwards, G. Taylor, H. Lyall, G. Tudor-Williams, V. Novelli, J. Martinez-Picado, P. Kiepiela, B. D. Walker, and P. J. Goulder. 2004. HIV evolution: CTL escape mutation and reversion after transmission. Nat. Med. 10:282–289. Lim, S. G., Y. Cheng, S. Guindon, B. L. Seet, L. Y. Lee, P. Hu, S. Wasser, F. J. Peter, T. Tan, M. Goode, and A. G. Rodrigo. 2007. Viral quasi-species evolution during hepatitis Be antigen seroconversion. Gastroenterology 133: 951–958. Liu, Y., J. McNevin, J. Cao, H. Zhao, I. Genowati, K. Wong, S. McLaughlin, M. D. McSweyn, K. Diem, C. E. Stevens, J. Maenza, H. He, D. C. Nickle, D. Shriner, S. E. Holte, A. C. Collier, L. Corey, M. J. McElrath, and J. I. Mullins. 2006. Selection on the human immunodeficiency virus type 1 proteome following primary infection. J. Virol. 80:9519–9529. Neumann, A. U., N. P. Lam, H. Dahari, D. R. Gretch, T. E. Wiley, T. J. Layden, and A. S. Perelson. 1998. Hepatitis C viral dynamics in vivo and the antiviral efficacy of interferon-alpha therapy. Science 282:103–107. Nowak, M. A., S. Bonhoeffer, A. M. Hill, R. Boehme, H. C. Thomas, and H. McDade. 1996. Viral dynamics in hepatitis B virus infection. Proc. Natl. Acad. Sci. USA 93:4398–4402. Osiowy, C., E. Giles, Y. Tanaka, M. Mizokami, and G. Y. Minuk. 2006. Molecular evolution of hepatitis B virus over 25 years. J. Virol. 80:10307– 10314. Pawlotsky, J. M. 2003. Hepatitis C virus genetic variability: pathogenic and clinical implications. Clin. Liver Dis. 7:45–66. Perelson, A. S., A. U. Neumann, M. Markowitz, J. M. Leonard, and D. D. Ho.
HOST IMMUNITY AND HBV DIVERSITY
32.
33.
34.
35. 36. 37.
38.
39. 40.
41.
42. 43.
44.
45.
46. 47.
48.
3463
1996. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science 271:1582–1586. Ray, M. B., V. J. Desmet, J. Fevery, J. De Groote, A. F. Bradburne, and J. Desmyter. 1976. Distribution patterns of hepatitis B surface antigen (HBsAg) in the liver of hepatitis patients. J. Clin. Pathol. 29:94–100. Rehermann, B., C. Pasquinelli, S. M. Mosier, and F. V. Chisari. 1995. Hepatitis B virus (HBV) sequence variation of cytotoxic T lymphocyte epitopes is not common in patients with chronic HBV infection. J. Clin. Invest. 96:1527–1534. Ross, H. A., and A. G. Rodrigo. 2002. Immune-mediated positive selection drives human immunodeficiency virus type 1 molecular variation and predicts disease duration. J. Virol. 76:11715–11720. Slatkin, M. 1977. Gene flow and genetic drift in a species subject to frequent local extinctions. Theor. Popul. Biol. 12:253–262. Stumpf, M. P., and O. G. Pybus. 2002. Genetic diversity and models of viral evolution for the hepatitis C virus. FEMS Microbiol. Lett. 214:143–152. Thompson, A., S. Locarnini, and K. Visvanathan. 2007. The natural history and the staging of chronic hepatitis B: time for reevaluation of the virus-host relationship based on molecular virology and immunopathogenesis considerations? Gastroenterology 133:1031–1035. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25:4876–4882. Thornton, K. 2003. Libsequence: a C⫹⫹ class library for evolutionary genetic analysis. Bioinformatics 19:2325–2327. Tsiang, M., J. F. Rooney, J. J. Toole, and C. S. Gibbs. 1999. Biphasic clearance kinetics of hepatitis B virus from patients during adefovir dipivoxil therapy. Hepatology 29:1863–1869. Vasudha, K. C., A. N. Kumar, and T. Venkatesh. 2006. Studies on the age dependent changes in serum adenosine deaminase activity and its changes in hepatitis. Indian J. Clin. Biochem. 21:116–120. Wakeley, J. 2004. Metapopulation models for historical inference. Mol. Ecol. 13:865–875. Webster, G. J., S. Reignat, D. Brown, G. S. Ogg, L. Jones, S. L. Seneviratne, R. Williams, G. Dusheiko, and A. Bertoletti. 2004. Longitudinal analysis of CD8⫹ T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy. J. Virol. 78:5707–5719. Whalley, S. A., D. Brown, G. J. Webster, R. Jacobs, S. Reignat, A. Bertoletti, C. G. Teo, V. Emery, and G. M. Dusheiko. 2004. Evolution of hepatitis B virus during primary infection in humans: transient generation of cytotoxic T-cell mutants. Gastroenterology 127:1131–1138. Wong, W. S., Z. Yang, N. Goldman, and R. Nielsen. 2004. Accuracy and power of statistical methods for detecting adaptive evolution in protein coding sequences and for identifying positively selected sites. Genetics 168: 1041–1051. Yang, Z. 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol. Biol. Evol. 24:1586–1591. Yuste, E., S. Sanchez-Palomino, C. Casado, E. Domingo, and C. LopezGalindez. 1999. Drastic fitness loss in human immunodeficiency virus type 1 upon serial bottleneck events. J. Virol. 73:2745–2751. Zhou, Y., and E. C. Holmes. 2007. Bayesian estimates of the evolutionary rate and age of hepatitis B virus. J. Mol. Evol. 65:197–205.
