|Home | About | Journals | Submit | Contact Us | Français|
Hepatitis B virus (HBV) of a novel genotype (J) was recovered from an 88-year-old Japanese patient with hepatocellular carcinoma who had a history of residing in Borneo during the World War II. It was divergent from eight human (A to H) and four ape (chimpanzee, gorilla, gibbon, and orangutan) HBV genotypes, as well as from a recently proposed ninth human genotype I, by 9.9 to 16.5% of the entire genomic sequence and did not have evidence of recombination with any of the nine human genotypes and four nonhuman genotypes. Based on a comparison of the entire nucleotide sequence against 1,440 HBV isolates reported, HBV/J was nearest to the gibbon and orangutan genotypes (mean divergences of 10.9 and 10.7%, respectively). Based on a comparison of four open reading frames, HBV/J was closer to gibbon/orangutan genotypes than to human genotypes in the P and large S genes and closest to Australian aboriginal strains (HBV/C4) and orangutan-derived strains in the S gene, whereas it was closer to human than ape genotypes in the C gene. HBV/J shared a deletion of 33 nucleotides at the start of preS1 region with C4 and gibbon genotypes, had an S-gene sequence similar to that of C4, and expressed the ayw subtype. Efficient infection, replication, and antigen expression by HBV/J were experimentally established in two chimeric mice with the liver repopulated for human hepatocytes. The HBV DNA sequence recovered from infected mice was identical to that in the inoculum. Since HBV/J is positioned phylogenetically in between human and ape genotypes, it may help to trace the origin of HBV and merits further epidemiological surveys.
Worldwide, an estimated 400 million people are infected with hepatitis B virus (HBV) persistently, of whom three quarters live in the Southeast and Far East Asia, and one million die of decompensated cirrhosis and/or hepatocellular carcinoma (HCC) annually (8, 15). HBV is the smallest animal DNA virus and has a genome made of approximately 3,200 nucleotides (nt) that contains four open reading frames for P, C, S, and X genes; they code for DNA polymerase/reverse-transcriptase, core protein, surface protein, and X protein, respectively (49). The S gene is divided into preS1 and preS2 regions and the small S gene, and the C gene splits into PreC and C.
Eight genotypes of HBV have been recognized by a sequence divergence of >8% in the entire genome and named by capital alphabet letters (A to H) in the order of discovery (3, 26, 29, 42). HBV genotypes are further classified into subgenotypes, such as B1/Bj and B2-5/Ba (44), as well as C1/Cs, C2/Ce, and C3-5 (36). A systematic nomenclature is proposed for designating HBV subgenotypes using Arabic numbers, such as A1, A2, and A3 (25). HBV genotypes have distinct geographical distribution (16, 23). Genotype A is prevalent in Africa, Europe and India, genotypes B and C are common in Asia, and genotype E is common in sub-Saharan Africa. Genotypes F and H are restricted to Central and South American continents, whereas genotype D is distributed all over the world. HBV genotypes have clinical application, and they influence severity and progression of liver disease and the response to antiviral therapies. Previous reports indicate that HCC is more frequent in the patients infected with genotype C than B (7, 47), and interferon is more effective in those infected with genotype B than C in Asia and more effective in those infected with genotype A than D in Europe (18, 34, 51).
Recently, a ninth genotype (I) was tentatively proposed for HBV strains detected in Laos (31). These strains are phylogenetically similar to aberrant Vietnamese strains that display complex recombination over the genome (10). In the present study, an HBV isolate was recovered from a Japanese patient with HCC, who was involved in military actions in Borneo during the World War II. The isolated strain was compared against eight human (A to H) and four ape (chimpanzee, gorilla, gibbon, and orangutan) genotypes and was provisionally designated genotype J. The new genotype was assigned based on a sequence divergence of 10.7 to 15.7% from other genotypes, a unique phylogenetic position between human and ape genotypes, and the absence of strong evidence of recombination.
A Japanese man, 88 years old, developed HCC in 2006. He had a history of residing in Borneo during the World War II. No HBV infections were recorded in his family members. In October 1996, he was diagnosed with chronic hepatitis B. Hepatitis B surface antigen (HBsAg) was detected in serum, and the aspartate aminotransaminase and alanine aminotransferase levels were elevated to 83 and 73 U/liter, respectively (normal levels, <30 U/liter for both). Thereafter, the transaminase levels were normalized, and he had been monitored as an asymptomatic HBV carrier. In August 2000, the level of a tumor marker (des-γ-carboxy prothrombin) was elevated to 52 mAU/ml (normal, <40 mAU/ml), while another tumor marker (alpha-fetoprotein) remained within normal range (<10 ng/ml) as alanine aminotransferases. In October 2006, a tumor (4.3 by 4.1 cm) was detected in the liver by ultrasonography, and he received treatment with transarterial embolization. Des-γ-carboxy prothrombin was elevated to 419 mAU/ml, while the aminotransferase levels remained within normal limits. Hepatitis B e antigen (HBeAg) was negative, and the corresponding antibody (anti-HBe) was detected in his serum. The subtype of HBsAg in this serum was ayw.
HBV DNA was extracted from his serum specimen obtained in 2006, and the full-length genome sequence was determined for phylogenetic and biological analyses. An informed consent had been obtained from the patient, and the study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the institution's human research committee.
HBeAg and anti-HBe were determined by enzyme-linked immunosorbent assay (ELISA) with commercial kits (HBeAg EIA; Institute of Immunology, Tokyo, Japan), and subtypes of HBsAg by ELISA with commercial kits (HBsAg Subtype EIA; Institute of Immunology). Hepatitis B core-related antigen (HBcrAg) was determined by chemiluminescence enzyme immunoassay (13). The method allows more sensitive detection of core protein and, as was shown in previous studies, HBcrAg levels reflect HBV DNA loads and well correlate with intrahepatic covalently closed circular DNA (cccDNA) levels. The measurement of serum HBcrAg is a useful noninvasive tool for monitoring intrahepatic HBV viral status (52). HBV DNA was quantified by the S gene-targeted real-time detection PCR with a sensitivity of 100 copies/ml (equivalent to 20 IU/ml) (1). However, due to small volumes of sera available from the challenged mice, HBV DNA was extracted from 10-fold-diluted specimens, resulting in reduced assay sensitivity in the present study (1,000 copies/ml [200 IU/ml]).
HBV DNA was extracted by using the QIAamp DNA blood kit (Qiagen, GmbH, Hilden, Germany) from 100 μl of serum that had been stored at −80°C. The complete genome sequence of an HBV/J isolate recovered from the patient was determined by the strategy previously reported (43). In brief, two sets of primers were designed to amplify overlapping fragments (A and B) covering the entire HBV genome (stat not shown). Nested PCR was carried out for 35 cycles (95°C, 30 s; 57°C, 30 s; and 72°C, 2 min) using TaKaRa LA Taq polymerase (Takara Biochemicals, Kyoto, Japan). Amplified fragments were inserted into the pGEM-T Easy vector (Promega, Madison, WI), and cloned in DH5a cells (Toyobo, Osaka, Japan). Obtained HBV DNA clones were confirmed to have the sequence identical to the major-clone consensus sequence determined directly on PCR products by Prism BigDye (Applied Biosystems, Foster City, CA) in the ABI 3100 automated sequencer.
Full-length sequences of HBV isolates were aligned with use of the CLUSTAL W software program (48) (available at www.ebi.ac.uk), and the alignment was confirmed by visual inspection. Genetic distances were estimated by the six-parameter method, and phylogenetic trees were constructed with the neighbor-joining method (35). To confirm the reliability of phylogenetic trees, bootstrap resampling and reconstruction were carried out 1,000 times using the program of the Hepatitis Virus Database (39). All 1,440 complete genomes available in the DDBJ/GenBank served as references for the initial alignment in the present study. Divergence in the nucleotide sequence between a strain of provisional genotype J and previously reported strains was estimated by using MEGALIGN v.6.00 (Lazergene package; DNASTAR, Inc., Madison, WI).
Evidence of possible recombination was investigated by using the software packages Simmonic 2005 v1.6 and SimPlot v3.5.1, both implementing PHYLIP (Phylogeny Inference Package v3.68; J. Felsenstein, Department of Genome Sciences, University of Washington, Seattle [distributed by the authors]) (19, 40).
Severe combined immunodeficiency mice transgenic for the urokinase-type plasminogen activator gene (uPA+/+/SCID+/+ mice) with the liver repopulated with human hepatocytes (chimeric mice) were purchased from Phoenix Bio Co., Ltd. (Hiroshima, Japan). Human serum albumin was measured by ELISA with commercial assay kits (Eiken Chemical Co., Ltd., Tokyo, Japan) for estimating the extent of repopulation. The research complied with all relevant federal guidelines and institutional policies.
Freshly prepared liver tissues were snap-frozen in isopentane precooled in liquid nitrogen. Frozen specimens were cut at 5 to 6 μm by cryostat, mounted on glass slides, air dried, and fixed in 100% acetone at room temperature for 10 min. Sections were blocked with antibody diluent (Dako, Tokyo, Japan) and stained for hepatitis B core antigen (HBcAg). They were incubated with rabbit anti-HBc (Dako) at room temperature for 1 h, washed in phosphate-buffered saline, and then incubated with goat anti-rabbit immunoglobulin G conjugated with Cy3 (Chemicon International, Inc., Temecula, CA) or goat anti-human albumin antibody labeled with fluorescein isothiocyanate (Bethyl Laboratories, Inc., Montgomery, TX). Sections were washed with phosphate-buffered saline and observed in a fluorescence microscope (Eclipse E800M; Nikon, Tokyo, Japan).
The nucleotide sequence data reported in the present study will appear in the DDBJ/EMBL/GenBank databases under accession no. AB486012.
HBV DNA was extracted from serum of a patient with HCC. It was named JRB34 (“J” for Japanese; “R” after the southernmost island [Ryukyu] where the patient has spent most of his life now exceeding 90 years; and “B” for Borneo where he is suspected to have contracted the HBV infection). The entire nucleotide sequence was determined for the JRB34 isolate of genotype J (HBV/J). It had a genomic length of 3,182 nt, which consisted of envelope gene containing preS1 region (nt 2848 to 3171, coding for 108 amino acids [aa]), preS2 region (nt 3172 to 154 [55 aa]), and the small S gene (nt 155 to 835 [226 aa]), X gene (nt 1374 to 1838 [154 aa]), preC region (nt 1814 to 1897 [27 aa]), C gene (nt 1901 to 2452 [183 aa]), and P gene (nt 2307 to 1623 [832 aa]).
The complete genome sequence of the JRB34 strain obtained in the present study was compared against those of 1,440 HBV genomes registered in the Viral Hepatitis Database (39). Estimated nucleotide sequence divergence of the JRB34 strain from four ape and nine human genotypes is summarized in the Table Table1.1. The mean divergence by genotypes ranged from 10.7 and 10.9% (from orangutan and gibbon, respectively) to 15.6 and 15.7% (from genotypes F and H, respectively). Surprisingly, the minimum divergence of 9.9% was observed in comparison with a nonhuman HBV isolate from Hilobates agilis gibbon confiscated in Taiwan in 1993 (AY330917) (41). Since the sequence divergence from any documented genotypes, including recently proposed genotype I, exceeded 8%, the JRB34 strain was tentatively classified into a novel genotype J of HBV.
In the phylogenetic tree constructed on 1,440 complete genome EMBL/DDBJ/GenBank database entries, the HBV/J strain was positioned distinctively from all known human genotypes (data not shown). It was closest to the cluster formed by gibbon- and orangutan-derived strains. However, including recombinant strains in such analyses may significantly affect the overall phylogenetic topology. This possibility was ruled out by reconstruction of the phylogeny using nonrecombinant HBV strains that further confirmed the phylogenetic peculiarity of the studied JRB34 strain (see Fig. S1 in the supplemental material). A total of 44 representative reference strains were further selected for establishing the consistency. Thus, phylogenetic topology indicating genotype-specific clustering is shown in the Fig. Fig.1.1. Hence, using various sets of references, we confirmed that genotype J undoubtedly differed phylogenetically from all other known genotypes.
To investigate possible recombination in the JRB34 genome, a window scanning analysis of aligned HBV genomes was performed by means of Simplot and Simmonics software packages. Both Bootscanning by SimPlot and GroupScanning by Simmonics showed similar output results. However, the methodological approach is different between these two software packages; GroupScanning provides more robust analysis of the phylogenetic relation between the examined strain and clusters of reference strains, whereas SimPlot does this comparison between the examined strain and parametrically generated consensus of the reference strains. The results obtained by SimPlot therefore can be significantly affected by selected parameters for the generation of consensus. This is especially undesirable when a new genotype strain (for which no references are available among known genotypes) is being analyzed (40). Figure Figure22 shows genome-wide distance scanning and GroupScanning plots for the JRB34 strain in comparison with a reference set consisting of 228 nonrecombinant HBV isolates retrieved from the public database (the phylogenetic tree is shown in Fig. S1 in the supplemental material). It is evident that the JRB34 strain was divergent from all known genotypes, and the closest genetic neighbors were estimated by distance and phylogenetic association scanning were the gibbon genotype (in preS, S, and P genes) and genotype C (in the core gene). However, no significant evidence of recombination between these two ape and human genotypes was revealed by the used methods. Homology scan carried out by SimPlot using the same set of reference sequences gave concordant results.
Phylogenetic relationship between the JRB34 strain and other genotypes was further analyzed in four open reading frames. In the small S gene, subgenotype C4 recovered from Australian aborigines (43) changed its phylogenetic topology from the branch of human genotypes to a branch intermediate between orangutan and gibbon strains (Fig. (Fig.3A).3A). Remarkably, genotype J and C4 strains joined together to create a clade between orangutan and gibbon strains. In contrast, genotype J clustered with human genotypes in the phylogenetic analysis of the C gene and was closely related to genotype C; it took a position outside genotype I strains, however (Fig. (Fig.3B).3B). Genotype J was closer to gibbon and orangutan genotypes in the phylogenetic trees constructed on P and large S genes (data not shown), demonstrating its topology similar to that in the analysis of the entire genome (Fig. (Fig.11).
The amino acid sequence of HBV/J was compared against those of other genotypes over three different areas of the genome. The amino acid sequence in the preC gene and arginine-rich domain in the carboxy-terminal sequence in the C gene were well conserved by genotype J (Fig. (Fig.4).4). In the preS1 region, genotype J had a deletion of 11 aa as gibbon and chimpanzee genotypes (Fig. (Fig.5).5). This deletion was shared by one of the two HBV/C4 isolates from Australian aborigines, as well as all HBV/D isolates. Amino acid sequence in the S gene of genotype J was the same as those of aborigine isolates of subgenotype C4; they would share antigenic epitopes of HBsAg. Amino acids at codons 122 and 160 were arginine (with G as nt 365) and lysine (with G as nt 479), respectively, which was consistent with subtype ayw of HBsAg from this patient (27).
Five domains (A to E) of DNA polymerase/reverse transcriptase in the P gene were preserved well in HBV/J, and it did not have mutations in the Tyr-Met-Asp-Asp motif in the domain C that determines the sensitivity to lamivudine (data not shown). HBV/J possessed A1762T/G1764A double mutations in the core promoter and G1896A stop codon mutation in the preC region, which was compatible with an HBeAg-minus phenotype of HBV recovered from the patient positive for anti-HBe.
Two chimeric mice that had been transplanted with human hepatocytes were inoculated with 104 HBV DNA copies of genotype J. In both mice, HBV DNA in a high titer (105 copies/ml) appeared in the circulation at week 7, plateaued at high levels (106 to 108 copies/ml), and stayed detectable until 22 weeks of observation after the inoculation (Fig. (Fig.6A).6A). HBsAg and HBcrAg became detectable at week 7 and kept increasing in concentrations until week 15 when they reached a plateau at high levels (Fig. (Fig.6B).6B). HBV strains recovered from mice at the last day of follow-up were identical in the complete genome sequence to the JRB34 strain used for inoculation.
The liver from chimeric mice infected with HBV/J was stained for HBcAg by immunofluorescence (Fig. (Fig.7A).7A). The staining for HBcAg was confined to areas where mouse liver had been replaced for human hepatocytes, and the same areas were stained for human albumin (Fig. (Fig.7B).7B). Colocalization of HBcAg and human hepatocytes was demonstrated by double staining for HBcAg and human albumin (Fig. (Fig.7C).7C). Finally, expression and replication of the JRB34 strain were confirmed by successful detection of cccDNA and HBV RNA in the liver tissue from both sacrificed mice (see Fig. S2A and B in the supplemental material).
An HBV isolate (JRB34) was recovered from a male, 88-year-old Japanese patient with HCC and sequenced over the entire genome. In the full-genome sequence, the JRB34 strain had 10.9 to 15.7% divergence from 1,440 HBV strains retrieved from the DDBJ/EMBL/GenBank. The divergence exceeds 8% that has been defined originally for distinguishing between four genotypes (A to D) (29) and later for an additional four genotypes (E to H) (3, 26, 42). Phylogenetically, the sequence of JRB34 was closer to ape than human HBV genotypes. No significant evidence of recombination with eight known human and four ape genotypes was revealed by the GroupScanning analysis (40) and phylogenetic analyses. These lines of evidence have qualified the JRB34 strain to represent a possible new HBV genotype. To further confirm the epidemiological significance of this strain, capable of establishing new infections, two chimeric mice were each inoculated with 104 copies of JRB34 HBV DNA. They both were successfully infected with sharp increases in HBV DNA and HBsAg in serum several weeks after the inoculation. Replication in the chimeric mice was also confirmed by detection of cccDNA and HBV RNA in their liver tissues.
Recently, an HBV isolate from Vietnam (VH24 [accession no. AB231908]) was reported as a ninth human genotype (I) (12). However, VH24 differed by only 7.0% ± 0.4% from HBV isolates of genotype C and possessed complex recombination with genotypes A and G in three genomic areas. A number of sporadic HBV isolates have been reported to date that contain recombination between human genotypes (4, 24, 40), as well as between human and ape genotypes (21). Only a few recombinant variants, however, became widely spread in human populations, developing their own specific distributions and epidemiologies. This is particularly demonstrated for the B/C recombinant designated as a distinct subgenotype; Ba/B2-5 now accounts for the majority of genotype B strains in mainland Asia (44). Likewise, the C/D recombinant prevails in Tibet and northern China (50). To avoid assigning a new genotype for every newly discovered sporadic recombinant HBV variant, evidence of intergenotypic recombination should be carefully eliminated (14). However, in some cases, designation of a new genotype is proposed by a potential epidemiological significance of a novel genetic variant. Recently, a study carried out in Laos described a number of strains closely related phylogenetically with the Vietnamese genotype I strains, thereby suggesting their epidemiological significance (31). The JRB34 strain documented in the present study was genetically and phylogenetically distinct from any previously published strains, including those of genotype I from Vietnam and Laos. To avoid possible misconceptions in the future, the strain is provisionally designated genotype J.
HBV of distinct genotypes can infect great apes in the wild, including chimpanzee, gorilla, orangutan and gibbons (9, 20, 37, 51). HBV genotypes of chimpanzee and gorilla, as well as those of orangutan and gibbon, cocluster in agreement with their geographical distribution in Africa and Southeast Asia, respectively (41). Genotype J represented by the JRB34 strain clustered with gibbon/orangutan genotypes. In a phylogenetic analysis of the S region/gene sequence, JRB34 belonged to a nonhuman HBV group but was closely related to an HBV isolate of subgenotype C4 (AB048704) recovered from an Australian aborigine; C4 is most divergent from other subgenotypes of genotype C (43). In the phylogenetic analysis of the C gene, however, JRB34 clustered with human genotypes and closely related to genotype C, including C4, and was positioned outside genotype I strains (Fig. (Fig.4).4). Taken together, genotype J is phylogenetically close to gibbon/orangutan genotypes in the entire genome and to genotype C (C4 in particular) in the S and C genes. However, despite observed interchangeable relatedness with gibbon and genotype C/I strains, no strong evidence of recombination was confirmed in the JRB34.
In the sequence of C gene, carboxyl-terminal arginine-rich region, required for binding with HBV DNA, was preserved in JRB34. It had the G1896A stop codon in the precore region that aborts the translation of HBeAg (5, 30) and A1762T/G1764A double mutations in the core promoter that interfere with the transcription of HBeAg by downregulating preC mRNA (28, 45); they are compatible with the HBeAg− anti-HBe+ phenotype of the patient from whom JRB34 was isolated. Since the double mutations are detected frequently in HBV DNA sequences from patients with HCC (17, 33), it could be implicated in hepatocarcinogenesis of the patient from whom JRB34 was isolated. It is not certain, however, if precore and core-promoter mutations had existed in HBV transmitted to the patient who is presumed to have been infected 60 years ago. Since amino acid sequences constituting antigenic loops of HBsAg (6) were the same as those of Australian aborigine isolates of C4, they would share antigenic epitopes of HBsAg. The amino acids at codons 122 and 160 were arginine (with G at nt 365) and lysine (with G at nt 479), respectively (27), in agreement with subtype ayw of HBsAg from this patient. Five domains (A to E) of DNA polymerase/reverse transcriptase in the P gene were preserved well in HBV/J, and it did not have mutations in the Tyr-Met-Asp-Asp motif in the domain C that determines the sensitivity to lamivudine (2).
How and when the patient contracted infection with HBV/J is not certain. It is very unlikely, however, that he acquired infection in Japan via perinatal or horizontal transmission. There are no wild primates in Okinawa, where the patient was originally from, and the prevalent human HBV genotypes are limited to B (60%), C (39%), and sporadic cases of A (1%) (32). Furthermore, HBV/J was not found among patient's family members who are currently alive (data not shown). The phylogenetic position within open reading frames of JRB34 in between gibbon/orangutan genotypes and human genotype C gives a clue where and when the patient had contracted HBV infection. He was drafted to Borneo during World War II (1939 to 1945); the island in the Southeast Asia is inhabited by gibbons and orangutans and has a local population mainly infected with genotypes B or C. Zoonotic infection of HBV has been previously reported (11, 46), and HBV of genotype E was recovered from a chimpanzee captured in West Africa where this genotype is common. There is a possibility that JRB34 of genotype J had been transmitted to the study patient in Borneo during the war (38).
The origin of genotype J in gibbon/orangutan or human inhabitants in Borneo is not certain but very likely. HBV DNA and/or HBsAg was detected in 26% (55/213) and 20% (58/297) of gibbons and orangutans, respectively, captured in Southeast Asia (38). HBV is also endemic in people living there, with a prevalence of HBsAg at 2 to 8%. There would be high chances for cross-species transmission of HBV where it prevails both in human beings and nonhuman primates. Phylogenetic analysis for close relationship between human and nonhuman HBV genotypes has indicated geographical influence rather than association with particular species (41).
It is remains to be determined whether genotype J and ape-derived strains originate from species-specific convergent evolution of distant strains or whether they have diverged from a single common ancestor sometime in the past and evolved independently thereafter. The validity of cross-species infection or species-specific evolution for genotype J would be verified by sequence analysis of HBV DNA from gibbons and humans living in Borneo. If they turn out to be the same, cross-species infection will be justified. Should genotype J be restricted to human beings, in converse, species-specific infection will be confirmed.
In conclusion, a novel HBV genotype was identified in the Ryukyu isolate and provisionally named genotype J. Phylogenetic analyses over the full-length sequence and open reading frames indicate a close relationship of genotype J with gibbon/orangutan genotypes and human genotype C. The index patient would have been infected with HBV/J while he resided in Borneo inhabited by gibbons and orangutans. Although only one HBV isolate of genotype J (JRB34) has been identified, this may be only the tip of an iceberg. It would be worthwhile to examine the genotype of HBV infecting people and gibbons, as well as orangutans, living in Borneo and neighboring countries for mapping the epidemiology of genotype J and finding any clinical relevance.
This study was supported in part by a grant-in-aid from the Ministry of Health, Labor and Welfare of Japan and a grant-in-aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
Published ahead of print on 29 July 2009.
†Supplemental material for this article may be found at http://jvi.asm.org/.