Occult Hepatitis B Virus Infection in Chronic Liver Disease: Full-Length Genome and Analysis of Mutant Surface Promoter

GASTROENTEROLOGY 2004;127:1356 1371 Occult Hepatitis B Virus Infection in Chronic Liver Disease: Full-Length Genome and Analysis of Mutant Surface Promoter VAISHALI CHAUDHURI,* RUCHI TAYAL,* BAIBASWATA NAYAK,* SUBRAT KUMAR ACHARYA, and SUBRAT KUMAR PANDA* *Department of Pathology and Department of Gastroenterology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi, India Background & Aims: Genome sequence of hepatitis B virus (HBV) from occult chronic infection is scarce. Fiftysix (9.4%) of 591 patients seronegative for hepatitis B surface antigen (HBsAg) with chronic liver disease were positive for HBV DNA. The complete HBV genome from 9 of these patients (S1 S9) and 5 controls positive for HBsAg (SWT.1 SWT.5) were analyzed. Methods: Overlapping genome fragment amplification, cloning, and sequencing was performed on these cases. Functional analysis of surface promoter was conducted using fusion construct. Results: All patients with occult infection except one (S8) had a low viral titer. Eight patients had infection with genotype A (S1 S5, SWT.1 2, SWT.5) and 6 had infection with genotype D (S6 S9, SWT.3 4). S4 and S5.1 of genotype A had the characteristic nucleotide deletions in core and pre-s1 region seen in genotype D. The major observations in patients with occult HBV infection were as follows: frequent quasispecies variation, deletions in pre-s2/s region affecting the surface promoters (nt 3025 54) and pre-s protein (S3, S5, S6, S8), truncated precore (S6, S8, S7.1) and core (S9) owing to stop signal, alternate start codon for the Polymerase gene (S3, S9), and YMDD mutation (S1, S4, S9) in patients not on antiviral therapy. HBsAg and core proteins could be shown immunohistochemically in 3 of 5 liver biopsy specimens available. The mutant surface promoters (pre-s2 and S) on functional analysis showed alterations in HBsAg expression. Conclusions: These changes in the regulatory region with possible alterations in the ratio of large and small surface proteins along with other mutations in the genome may decrease the circulating HBsAg level synergistically, making the immunodetection in serum negative. Hepatitis B virus (HBV) infection has significant morbidity and mortality worldwide and its prevalence varies from region to region. India lies in the intermediate prevalence zone for HBV endemicity with an estimated infection load of 40 million people. 1 The diagnosis of chronic infection by HBV is characterized by the persistent presence of viral envelope protein (hepatitis B surface antigen [HBsAg]) in the blood. HBsAgnegative HBV infection or occult HBV infection is a recently recognized entity, whose exact magnitude, pathogenesis, and clinical relevance in various populations is unclear. 2 Occult HBV infection is defined as HBV DNA detectable by PCR [polymerase chain reaction] among patients negative for HBsAg 3 and has been classified into seropositive and seronegative infection depending on positivity for anti core antibody (HBc) and anti- HBs. 4 Usually individuals with occult HBV infection have low viremia. 5 However, the diagnostic approach and mechanism of anti HBc-/anti HBs-negative serology in individuals with occult HBV infection have not been elucidated clearly. A progressive decrease in HBV load as well as replication and various relevant mutations have been implicated in the explanation of HBsAg negativity in such infections. 6 The cause in 47% (216 of 458) of acute liver failure patients in India 7 is unidentifiable. Forty-one percent (13 of 32) of these patients were positive for HBV DNA by polymerase chain reaction (PCR) (unpublished results, 2002). Considering that chronically infected individuals are the major reservoirs of HBV infection, de novo infection with naturally occurring mutants of HBV from this population might be causing such occult HBV infection. The seronegativity in these patients may be caused by mutations, which alters either the immunoreactivity of various HBV proteins 8 or the quantity of HBsAg in serum. 9 Earlier studies have focused on the analysis of limited genome region 10,11 in occult HBV infection. It is elusive as to which mutations and variations and this can affect HBsAg detection. Therefore, the current study was aimed at assessing the extent of occult HBV infection among patients with chronic liver disease, at analyzing sequence changes of these patients with mutant HBV with respect to those patients with wild-type HBV in the Abbreviations used in this paper: aa, amino acid; ATG, antithymocyte globulin; fluc, firefly luciferase; nt, nucleotide; PCR, polymerase chain reaction; S-rLUC, S gene-renilla luciferase. 2004 by the American Gastroenterological Association 0016-5085/04/$30.00 doi:10.1053/j.gast.2004.08.003

November 2004 GENOME ANALYSIS OF HBV IN OCCULT INFECTION 1357 same geographic region, and at studying the functional effect of commonly observed pre-s2/s promoter deletions on envelope protein expression. Materials and Methods Between January 1992 and December 1999, consecutive patients diagnosed with chronic liver disease without any ascribable cause (such as alcohol abuse, hepatotoxic drug, Wilson s disease, biliary tract disease, venous outflow obstruction, iron overload, autoimmune disease, 1 antitrypsin deficiency, and HBV and hepatitis C virus [HCV] infection) attending the liver clinic at the All India Institute of Medical Sciences, New Delhi, India, were included in the study. The diagnosis of chronic liver disease was made by conventional clinical, biochemical, imaging, and endoscopic criteria. 12 Histologic confirmation of chronic liver disease was obtained wherever possible (provided patients coagulogram was normal and informed consent could be obtained). Cryptogenic chronic liver disease was diagnosed when all known identifiable causes were excluded by relevant investigations. Occult HBV infection in patients with cryptogenic liver disease was diagnosed when PCR detected HBV DNA in sera for at least 2 different regions (surface and core) in the absence of detectable HBsAg. 3 The patients with occult HBV infection were included in the study. Sera samples were selected randomly from the patients with occult HBV infection for complete genome analysis. The genome sequence in patients with occult HBV infection was compared with the genome sequence of HBsAg-positive chronic liver disease patients (controls in the same age group) and with full-length wild-type HBV sequences present in the data bank. Each patient with occult HBV infection was subjected to a thorough investigation of the clinical, hematologic, biochemical, and liver function test profiles (serum bilirubin level, alanine transaminase level, aspartate transaminase level, alkaline phosphatase level, total protein level, serum albumin level, and prothrombin time). Each patient also had conventional imaging (ultrasonography, contrast-enhanced computed tomography) and upper gastrointestinal endoscopy. Liver biopsy examination was performed using an 18-gauge Menghini s aspiration needle (Boston Scientific, Natlick, MA). Serologic Assay The serum samples were investigated for HBV infection using commercial micro enzyme-linked immunosorbent assay kits for HBV envelope protein, HBsAg, hepatitis B e antigen (HBeAg), antibody against HBeAg (anti-hbe), anti-hbc (Organon Teknika, Boxtel, RM, The Netherlands), and anti-hcv (Xcyton, Bangalore, India). The tests were performed according to the manufacturer s instructions. Each sample was tested for autoantibodies by Varelisa double-stranded DNA antibody detection kit (Pharmacia and Upjohn Diagnostics, Freiburg, Germany). Anti nuclear factor, anti smooth muscle antibodies, and liver kidney microsomal antibody were tested by in house immunofluorescence procedure (developed in our laboratory). 13 Sera from these patients negative for all the earlier-mentioned tests were investigated for the presence of HBV DNA (PCR) and HCV RNA (reverse-transcriptase nested PCR) by previously described methods. 14 Patients who tested negative for HBsAg, HCV RNA, and other causative factors implicated in the cause of chronic liver disease but positive for HBV DNA by PCR were labeled as occult HBV infection as per the diagnostic criteria described earlier. 3,4 Sera samples from these patients were retested for the presence of HBsAg by using another commercial kit (Monolisa Biorad, Hercules, CA) to reconfirm absence of HBsAg in their sera. Isolation and Detection of HBV DNA From Serum HBV DNA was extracted by hot phenol method. Briefly, 100 L of serum sample was diluted to 500 L with milli-q grade distilled water. To this, an equal volume of Tris-saturated phenol (ph 8) was added, mixed with gentle inversion of the tube, and incubated at 65 C for 2 hours. The tubes were centrifuged at 12,000 rpm at 4 C for 5 minutes. The supernatant was collected carefully and extracted with Tris-saturated phenol (ph 8) followed by chloroform isoamyl alcohol (24:1). DNA was precipitated from the aqueous phase with absolute ethanol and one-third volume 7.5 mol/l ammonium acetate at 70 C for 1 hour, washed with 70% alcohol, air dried, and dissolved in 30 L Tris.Cl/EDTA (10 mmol/l Tris.Cl, ph 8.0; 1 mmol/l EDTA, ph 8.0). The dissolved DNA was diluted for 3 log dilutions, 10 1 to 10 3, and analyzed by PCR. HBV genome detection was performed using primers for the surface gene. 15 Template DNA (5 L) was suspended in 25 Lof PCR mix containing Gibco BRL buffer (20 mmol/l Tris-HCl [ph 8.4], 50 mmol/l KCl) and 0.3 U Taq DNA Polymerase (Gibco BRL, Bethesda, MD), 200 mol/l each of deoxyadenosine triphosphate, deoxycytidine triphosphate, deoxyguanosine triphosphate, and deoxythymidine triphosphate (Gibco BRL), 2.5 mmol/l magnesium chloride, and 15 pmole of HBV-specific oligonucleotide primers. Amplification conditions were as described in Table 1. Samples positive for surface gene were reconfirmed by PCR amplification of the core gene (Table 1). 16 Control samples included normal sera, HBV-positive sera, and negative controls. Positive and negative controls were included at the extraction steps and at both rounds of amplification. The sensitivity of these PCR assays was evaluated using serial dilutions of a World Health Organization standard (NIBSC, code 97/746). Samples were processed in duplicate for core and surface region and only repeatedly positive samples were considered for amplification of other regions. Every possible care was taken to avoid contamination. 17 PCR products were electrophoresed on a 2% agarose gel and stained with ethidium bromide. Complete HBV Genome Amplification by Overlapping PCR Full-length amplification was attempted using a single set of primers: P1/P2 or P3/P4. 18 The method used by Gunther et al. 18 for amplification of full-length genome

1358 CHAUDHURI ET AL. GASTROENTEROLOGY Vol. 127, No. 5 Table 1. Detection Primers for Hepatitis B Virus in Serum Name Nucleotide position Polarity Sequence 5=-3= Region Amplification condition MD 06 636 CTT GGA TCC TAT GGG 94 C, 4 min - 1 cycle AGT GG 94 C, 40 s MD03 735 CTC AAG CTT CAT CCA Surface 50 C, 1 min TAT A 73 C, 1 min - 35 cycles 73 C, 8 min - 1 cycle PS4 1820 AAC TTT TTC ACC TCT 94 C, 4 min - 1 cycle GCC TAA TCA Core 94 C, 40 s PA2 2454 CTA ACA TTG AGA TTC (outer) 55 C, 1 min CCG AGA TTG 73 C, 2 min - 35 cycles 73 C, 8 min - 1 cycle PIS4 1836 CCG GAT CCT CTC ATG 94 C, 4 min - 1 cycle TTC ATG T Core 94 C, 40 s PIA2 2395 CGA AGC TTG AGA TCT (nested) 56 C, 1 min TCG TCT 73 C, 2 min - 35 cycles 73 C, 8 min - 1 cycle was successful only with the wild-type HBV (Figure 1A and B) but failed repeatedly with the mutants. Hence, the method of overlapping genome fragment amplification had to be the method of choice in amplifying the entire genome from occult HBV cases. 16 The entire HBV genome from randomly selected patients with occult HBV infection and HBsAg-positive controls were amplified in overlapping fragments. The amplification was performed using KlenTaq (PE Cetus, Foster City, CA) and Pfu polymerase (in ratio 8:1) (Stratagene, LaJolla, CA) to prevent misincorporation. HBV-specific oligonucleotide primers and conditions are described in Table 2. Primers were selected from the conserved region of all genotypes. Nested PCR was performed by using 2 L of the first step product as a template in a 25 L PCR mix (Table 2). Cloning and Sequencing of PCR Products The PCR-amplified genomic segments were purified from agarose gel by QIA quick gel extraction kit (Qiagen, Hilden, Germany) and cloned into pgem-t easy vector (Promega, Madison, WI) as per the manufacturer s instructions. Figure 1. (A) Lane M, Hind III marker; lane 1, positive (HBsAg-/ HBeAg-positive) control; lane 2, negative control; lanes 3 6, HBsAgpositive samples. (B) Lane M, Hind III marker; lane 1, negative control; lane 2, HBsAg-positive control; lane 3, positive (HBsAg-/ HBeAg-positive) control; lanes 4 7, S2, S3, S4, and S9. Positive clones were identified by plasmid miniprep isolation and restriction digestion with EcoRI (Amersham Pharmacia Biotech, Uppsala, Sweden). Double-stranded DNA templates were prepared using plasmid miniprep kit (Qiagen), was denatured, and used for seque. Multiple independent clones (5 12) from each case were sequenced in both directions by Sanger s dideoxy chain termination method with T7, SP6, and internal in-house designed primers (Table 3). The manual sequencing reactions were performed using Sequenase version 2.0 polymerase (USB/Amersham Pharmacia Biotech). For patients with wild-type serology, direct sequencing of PCR product was performed. In wild-type serology, direct sequencing was used with manual screening of the peaks to check for variations. The sequencing was performed several times and with different primers to achieve a consensus sequence for each subfragment. Major confirmed variations between clones from the same isolate were called variants. If 2 different clones of the same region of subfragment had no variations they were regarded as a single variant. Each region was named alphabetically for convenience: V (PIS102 PIA7), W (PIS103 PIA104), X (PIS104 PIA100), Y (PIS100 PIA102) (for isolate S3 a different sense primer SQ7 was used as the start site because PS100 was deleted). Between 2 overlapping subfragments, there was considerable overlap excluding the primer sequence. The subfragments were joined by autoassembler version 2.1 (Perkin Elmer). They were put as the same variant if the overlapping sequence matched exactly. Initially manual sequencing was performed for isolates S2, S3, S4, S5, and S9. However, later automated DNA sequencer (ABI Prism 310; PE Applied Biosystems, Foster City, CA) with big dye terminator reaction and POP6 were used for sequencing of clones from isolates S1, S6, S7, S8, and SWT.1 SWT.5. Nucleotide and Amino Acid Analysis The consensus nucleotide and amino acid sequence for HBV DNA isolated from occult HBV infection and HBsAgpositive controls were aligned using the CLUSTAL program of DNASTAR (Laser Gene, Madison, WI) with 109 HBV sequences from GenBank (NCBI) using the BLAST alignment

November 2004 GENOME ANALYSIS OF HBV IN OCCULT INFECTION 1359 Figure 2. Phylogenetic analysis of variants of S1 S9 and SWT.1 SWT.5 with genotype A and D isolates from Genbank. search. Significant alterations were looked for in amino acid sequences and/or regulatory DNA elements. Genotyping of the obtained HBV-DNA sequences was performed by phylogenetic comparison with the representative genotypes (HBVADW2 X02763, genotype A), 19 HPBADW1 (D00329, genotype B), HPBADRA (M12906, genotype C), XXHEPAV (X02496, genotype D), HHVBE4 (X75664, genotype E), and HHVBF (X75663, genotype F) over the entire genome with SEQBOOT (boot strapping was performed in 1000 replicates) followed by DNADIST and FITCH algorithms in the PHYLIP version 3.5c package (J. Felsenstein, University of Washington, Seattle, WA). The old conventional EcoRI position was taken as nucleotide position 1, 19,20 as compared with the new numbering system 21 that is not in use universally. Effect of the Promoter Deletions on the Expression of the S Gene Frequent deletions were observed in the pre-s and S promoter as well as the interpromoter regions (nt 3025 54) in patients (S3, S5, S6, and S8) with occult HBV infection in comparison with wild-type isolates. To investigate the possibility of altered gene regulation caused by these altered promoter regions, functional assays of these mutants were performed on 4 (3 isolates carrying surface-promoter deletions and 1 with deletion of the interpromoter region). One wild-type isolate (ayw3) (SWT.3) was used as control. The activity of mutant and wild-type promoters driving surface gene expression was analyzed by S gene-renilla luciferase (S-rLUC) inframe fusion construct in prl-null vector (Promega). Briefly, the 2 surface promoters along with the S gene antithymocyte globulin (ATG) was amplified from nucleotide position 2698 175 of HBV (from the subfragment PIS104/ PIA100 described earlier) using primers SPF (Kpn I)/SPR (NheI) (Table 2). PIS104/PIA100= region fragments amplified from S6 (XS6.4, XS6.5, XS6.6, and XS6.7), which had nucleotide deletions in the regions the same as S8 (XS8.1 and XS8.2) were not investigated separately. Renilla luciferase (r- LUC) gene amplified (Table 2) and fused downstream of the S gene ATG in-frame in a modified prl-null vector (Promega) (prl null- pre-s/s promoter-s-rluc). These prl null- pre-s/s promoter-s-rluc constructs were confirmed, purified by a plasmid purification kit (Qiagen), and used for transfection of HepG2 cells. Cell Culture and Transfection For in vivo analysis of the promoter activity, the reporter plasmid constructs (prl null- pre-s/s promoter-s-rluc) were transfected to human hepatoma cell line (HepG2). Exponentially growing HepG2 cells were seeded onto 30-mm tissue culture dishes at a density of 1 10 6 cells per plate and incubated to yield 50% 70% confluency. The cells were transfected with 3 g ofprl null- pre-s/s promoter-s-rluc construct and cotransfection of 1 g ofpsg-fluc (firefly luciferase) in 10 L of Lipofectamine Plus Reagent (Life Technologies, Bethesda, MD) as per the manufacturer s instructions.

1360 CHAUDHURI ET AL. GASTROENTEROLOGY Vol. 127, No. 5 Table 2. Primers for Amplification of the Full-Length HBV Genome by PCR With Overlapping Fragments Name (nt) Position Sequence 5= to 3= Region Conditions PS102 972 CCT ATT GAT TGG AAA GTA TGT CAA ENHCII/X/Prec 94 C, 4 min-1 cycle PA7 1770 TAT GCC TAC AGC CTC CTA ATA CAA Outer primer 94 C, 40 s; PIS102 996 CGT ATT GTG GAT CCT TTG GGT TT ENHCII/X/Prec 56 C, 1 min; PIA7 1723 TCA AGC TTC TCC CAG TCT TTA AAC Nested primer 73 C, 2.5 min - 35 cycles 73 C, 8 min - 1 cycle PS103 1642 GCC CAA TGT CTT ACA TAA GAG GAC CORE 94 C, 4 min - 1 cycle PA104 2462 AAA GTT TCC CAC CTT ATG AGT CCA Outer primer 94 C, 40 s PIS103 1667 CTT GGA TCC TCT GTA ATG TCA CORE 56 C, 1 min; PIA104 2444 CCA AGC TTT ACT AAC ATT GAG Nested primer 73 C, 2.5 min - 35 cycles 73 C, 8 min - 1 cycle PS104 2313 CCC TAT CTT ATC AAC ACT TCC POL/Pre-S/S 94 C, 4 min - 1 cycle PA100 249 GAA GTC CAC CAC GAG TCT AGA Outer primer 94 C, 40 s; PIS104 2330 TTC CGG ATC CTA CTG TTG TTA POL/PRE-S/S 55 C, 1 min; PIA100 222 GGT ATT GTG AGG AAG CTT GTC Nested primer 73 C, 2.5 min - 35 cycles 73 C, 8 min - 1 cycle PS100 3163 ATC CTC AGG CCA TGC AGT Pre-S/S 94 C, 4 min - 1 cycle PA102 1178 CGT CAG CGA ACA CTT GG Outer primer 94 C, 40 s; PIS100 15 CCA CCA AAC TCT TCA GGA TCC PRES/S 56 C, 1 min; PIA102 1134 AAC GGG GTA AAG CTT CAG ATA Nested primer 73 C, 2.5 min - 35 cycles 73 C, 8 min - 1 cycle HBVCTF 1803 CAC CAG CAC CAT GCA ACT TT Quantitation primer (core region) 50 C, 2 min - 1 cycle HBVCTR 1911 TCA ATG TCC ATG CCC CAA A 94 C, 10 min - 1 cycle 94 C, 15 s, 60 C, 1 min - 40 cycles SPF 2698 TGG TAC CTT ATT ATC CAG AT Surface promoter-rluc fusion primer 94 C, 4 min - 1 cycle SPR 175 TCC TGA TGT GAT GCT AGC CAT 94 C, 40 s; 56 C, 1 min; 73 C, 2.5 min - 35 cycles 73 C, 8 min - 1 cycle psg-fluc (firefly luciferase) was used as an internal optimized equalization control. Assays for both Renilla (rluc) and firefly luciferase (fluc) activity were performed at 72 hours posttransfection, using commercial dual luciferase kits (Promega) and TD 20/20 Luminometer (Promega), as per the manufacturer s guidelines. Immunohistochemical Studies in Liver Tissue Standard immunohistochemical staining was performed in the liver tissue obtained by liver biopsy examination for detection of HBsAg and HBcAg using monoclonal antibodies (Dakopats, Copenhagen, Denmark). The staining was performed on formalin-fixed paraffin sections of needle biopsy specimens of the liver. The antigens were detected by the streptavidin-biotin complex technique with 3,3= diaminobenzidine tetrahydrochloride (D-0426; Sigma, St. Louis, MO) as chromogen. Brown coloration was taken as a positive staining. HBsAg was localized in the cytoplasm and HBcAg was localized in the nucleus. Quantitative HBV-DNA Estimation HBV-DNA concentration in serum samples of patients with occult HBV infection in whom detailed genomic analysis was performed was estimated by a real-time PCR method using SYBR Green/taq man protocol in an ABI Prism 7700 Table 3. Internal Primers for Sequencing of HBV Genome Name nt. Position Polarity Sequence 5= to 3= Region SQ7 203 AGG CGG GTT TTT CTT GTT GAC AA Surface P550 552 TAT GTT TCC CTC ATG TT Surface P1120 1138 TGA ACC TTT ACC CCG TT Enh I SQS106 1267 TCC ATA CTG CGG AAC TCC TAG Enh I SQS1 1930 GGA GCA TCT GTG GAG TTA CTC Core SQA106 2098 CCC AGG TAG CTA GAG TCA T Core P2720 2717 TAG TTA ATC ATT ACT TC Pre-S1 promoter SQA107 2880 TCG GGA AAG AAT CCC AGA GG Pre-S1 P2900 2887 CTT TCC CGA TCA TCA GTT Pre-S1 Taqman probe (SKPB) 1827 ACC TCT GCC TAA TCA T Core

November 2004 GENOME ANALYSIS OF HBV IN OCCULT INFECTION 1361 Table 4. Demographics and Biochemical Profile of the Patients With Occult and Wild-Type (Control Group) Infection Groups Group 1, total no. of patients (n 56) Group 2 patients included for sequence analysis (n 9) Age (y) Sex (M:F) Serum bilirubin (mg%) Aspartate transaminase (IU/dL) Alanine transaminase (IU/dL) Serum albumin level (g%) Diagnosis 33.39 14.13 2.1:1 2.21 1.66 140.5 92.44 121.125 90.96 3.63 1.07 23 cirrhosis 33 chronic hepatitis 29.5 10.6 3:1 1.78 1.08 188.625 111.428 136.37 112.79 3.76 0.58 5 cirrhosis 4 chronic hepatitis Control (wild-type group) 26 6.04 5:nil 1.46 0.808 103.8 62.6 108 66.26 3.2 0.78 5 chronic hepatitis NOTE. All parameters in the 2 groups were compared statistically using Mann Whitney test and found to be insignificant. Sequence Detector (PE Applied Biosystems). A primer-probe combination was designed on Primer Express software (PE Applied Biosystems) (Tables 2 and 3).The efficiency of amplification of the primers was first standardized by SYBR green assay. The real-time PCR was performed with the 2 Universal Taqman Master Mix in a 25- L reaction volume containing 15 pmole of HBV CTF and HBV CTR primers (Table 2) and 250-nm probe (SKPB probe) (Table 3), and 5 templates. The rest of the volume to 25 L was made with sterile water. The reaction conditions were the same as in the SYBR green assay, the only difference being that now the reporter was changed to Fam (reporter dye) for nontemplate control, unknown, and standard samples. To examine the analytic sensitivity and validity of this PCR, a negative water control, a negative serum control, normal HBsAg-positive serum, a control HBV-DNA standard (NBISC standard code 97/746) of 10 5 copies/ml, and an in-house standard with 10-fold serial dilutions (10 2 10 10 ) were analyzed. The amplification curve was generated and adjusted to the threshold cycle value and a standard curve was plotted. Each reaction was put in triplicate. Results During the period from 1992 to 2001, 591 patients were diagnosed to have cryptogenic chronic liver disease, of whom 255 patients had chronic hepatitis and the remaining 336 patients had cirrhosis. HBV DNA alone was detected in 33 (13%) and 23 (7%) of these patients with chronic hepatitis and cirrhosis, respectively. Similarly, HCV RNA was positive in 45 (17.6%) and 38 (11.3%) patients in these 2 categories. Both HBV DNA as well as HCV RNA was detected in 10 (3%) and 2 (0.5%) patients, respectively. Sixty-five of the 591 patients, including the 56 HBV-DNA positive cryptogenic chronic liver disease patients, were positive for immunoglobulin G anti-hbc. Anti-HBe was detected in none of these patients. These 56 patients positive for anti-hbc immunoglobulin G and HBV DNA were categorized as occult HBV infection. Sera from 9 of the earlierdescribed patients (5 with cirrhosis and 4 with chronic hepatitis) with occult HBV infection were selected randomly for analysis of complete HBV genome (S1 S9). The sera of these 9 patients were tested for the presence of anti-hbsag by using a commercial enzyme-linked immunosorbent assay kit (Organon Teknika) and were found negative for anti-hbsag. Sera from 5 HBsAg-positive patients with histologically confirmed chronic hepatitis were used as controls (SWT.1 SWT.5). Complete genomic analysis of the HBV DNA also was performed in the control group for comparison with any mutation. The demographic and biochemical profiles of these 9 randomly selected patients with occult HBV infection were no different from the entire group of occult HBVinfected patients and the control group of 5 HBsAgpositive patients included for HBV DNA sequencing (SWT.1 SWT.5) (Table 4).The 9 patients (S1 S9) consisted of 6 men and 3 women and had a mean age of 29.5 10.6 years (Tables 4 and 5). In 5 of these 9 patients, liver biopsy specimens were available and immunohistochemistry detected the presence of HBsAg and HBcAg in 3 of them (Table 5). Sequencing of several clones rather than direct sequencing of the PCR product from each isolate led to the detection of a diverse population of variants. 22,23 Eight had infection with genotype A (S1 S5, SWT.1 2, SWT.5) and 6 had infection with genotype D (S6 S9, SWT.3 4) although S4 and S5.1 of genotype A had the characteristic genotype D, 6 nucleotide deletions in core, and 33 nucleotide deletions in the pre-s1 region, respectively 24 (Genbank Accession numbers AY161138 63; SWT.1 5, Genbank Accession numbers AY373428 AY373432) (Figure 2). Mutations and Single Nucleotide Variations in the HBV Genome All 9 patients with occult infection showed quasispecies variation. Sequencing data revealed shortening of the HBV genome in all the variants of isolates S3, S4,

1362 CHAUDHURI ET AL. GASTROENTEROLOGY Vol. 127, No. 5 Table 5. Serum Markers of the Patients Whose HBV Isolates (S1 S9) Were Sequenced Patient isolate Age (y) Sex Diagnosis Anti-HBc Anti-HBe Anti-HBs Aspartate transaminase (IU/L) (at time of sera collection) Viral DNA load (genome copies/ml) Immunostain for HBsAg and HBcAg S1 36 F Cirrhosis 136 10 4 No Bx S2 18 M CH 212 10 3 Positive S3 32 M Cirrhosis 150 10 3 Positive S4 28 F Cirrhosis 30 10 4 No Bx S5 28 M CH 123 10 4 Negative S6 53 M CH 190 10 3 No Bx S7 19 M Cirrhosis 190 10 4 No Bx S8 32 M Cirrhosis 130 10 7 Positive S9 18 F CH 424 10 4 Negative CH, chronic hepatitis; No Bx, biopsy specimen not available. S8, S9, and in a few variants of S5 (S5.2, S5.1) and S6 (S6.4, S6.5, S6.6, S6.7) (Table 6). One isolate from the control group SWT.4 had deletion-covering nt. 3043 3096. Long-stretch deletions were observed in the pre- S1/S2 gene in occult infection at a much higher frequency than in the control group (Figure 3, Table 6). Three unique single nucleotide changes were observed in isolates from occult infection at position 754 (A3 C) in S1, S2, S3, S4, S5, and S9.1; 2577(A3G) in S2, S3, and S5; 2578 (T3C) in S1, S2, S3, and S5, in comparison with full-length sequences in Genbank, and the control group from local geographic isolates (SWT.1 SWT.5). In a single variant S5.2, 3 simultaneous single nucleotide deletions at position 3221, 9, and 80 changed the reading frame from aa 4 to 30 in the pre-s2 protein, but the rest of the sequence had homology to wild type. In pre-s1 protein, aa 19 26 has been reported as a B-cell antigenic epitope. 25 Substitutions in this domain were identified in isolates S9 (S16F) and S3 (P26S). In variant S3.3, G3084A point mutation terminates pre-s1 gene at pre-s1 aa 76 and pre-s2 ATG is abrogated because of internal deletion-covering nt. 3086 3221 and nt. 1 15, but the S ATG was intact. Only 1 insertion of a single amino acid (valine) was observed in isolate S1.1 in the small surface protein in between position 88 and 89. This change lies in the corresponding N terminal region of the reverse-transcriptase domain gene in which an arginine residue was inserted between aa. 444 445, without alteration in the downstream sequences. In variants S6.2 and S6.5, pre-s1/s2/s is terminated prematurely at aa 68 of S protein owing to T363A alteration and in variant S6.3 at aa 6 of S gene owing to C171A point mutation. In variant S6.7, the pre-s1 ATG (ATG-ATT) was ablated owing to G2855T mutation in pre-s1 gene and the pre-s2 gets terminated owing to stop codon at aa 68 of S protein. In case S6, however, there were variants (S6.1, S6.4, S6.6) with the intact surface region to compensate for deletion mutants. A short stretch in pre-s2 protein, from aa (28 51) has been mapped as a T-cell epitope. 26 Amino acid changes were concentrated in genotype A isolates over this region (Table 6). The immunodominant a determinant (aa 124 147) 27 of HBsAg had no variation in any of the isolates, except S6 and variant S9.1 (Table 7). Both these (S6 and S9.1) isolates are of genotype D origin and had 2 consecutive substitutions at aa M133I and Y134H. The same isolates also showed a genotype A specific substitution at T130N. In isolate S2, the glycosylation motif of the a determinant showed a substitution at aa C147R. 28 Given the importance of cys-cys disulphide bond in maintaining the conformation required for HBsAg antigenicity, loss of cysteine can result in alteration of HBsAg structure. 29,30 No HBV variant was found to contain mutations within both loops of the a determinant region simultaneously. The much-documented G145R, vaccine escape mutation, 31,32 was not found in any of these isolates. E164K mutation also was observed in the vicinity of the a determinant region. In isolate S8, the viral load was higher (10 7 copies/ml). This isolate had deletion between the pre-s1 and pre-s2/s promoter (nt. 3034 3090), which brings the promoters closer. The surface protein region (aa 185 215) had a higher incidence of variations (Table 7). This region lies outside the mapped T-cell (aa 28 51) or B-cell epitope (aa 124 148) 33 and could be tolerant to random mutations. Because of the limited sample size in the current study, it could not be concluded whether the mutations in the aa 185 215 region represent a mutational hotspot or results from host-selective pressure in the Indian population, reflecting a different immunogenic make-up. Within the precore/core region, the e suppressive phenotype 34 was found in 3 isolates (S6, S8, S7.2) (Table 6) (Figure 4). In isolates S6 and S8, a stop signal at aa 28 was introduced owing to nucleotide substitution G1896A. A stop signal in variant S7.2 of isolate S7 was observed at aa 18 of the e antigen owing to C1864T (CAA TAA) substitution. The rest of the variants of this

Table 6. Summary of the Amino Acid Changes in Structural Genes of Isolate S1 S9 Isolate Variant Genbank accession numbers Genome length Gen Pre-S1 Pre-S2 S Precore Core S1 S1.1 AY161138 3224 A S5T; A54Q; I74V; T86A; S89P a ; T90A1; I91V P94S S1.2 AY161139 3221 A S5T; A54Q; I74V; T86A; S89P a ; T90A1; I91V P94S S2 S2.1 AY161140 3221 A G35K; F45L; A54Q; I74V; T86A; S89P a ; T90A a ; I91V S2.2 AY161141 322 a A G35K; F45L; A54Q; I74V; T86A; S89P a ; T90A a ; I91V S3 S3.1 AY161142 3080 A P26S; A54Q; I74V; T86A; S89P a ; deletion pre-s1 95 pre-s2 23 S3.2 AY161143 3071 A P26S; A54Q; I74V; deletion (pre-s1 78- pre-s2 9) S3.3 AY161144 3071 A Truncated at amino acid position 76 S4 S4.1 AY161145 3215 A S5T; A54Q; I74V; T86A; S89P a ; T90A a ; I91V; P94S S4.2 AY161146 3215 A S5T; A54Q; I74V; T86A; S89P a ; T90A a ; I91V; P94S S5 S5.1 AY161147 3004 D Deletion in pre-s1 58- pre-s1 119 A7TV a ; V32L; A35V a ; A47S a ; A53V a ; T54P A7TV a ; V32L; A35V a ; A47S a ; A53V a ; T54P F22S; V32L; N37I; S40P; H41L; A47S a ; T49I a ; A53V a ; T54P F22L; P23L; V32L; A47S a ; T49I a ; A53V a ; T54P; Deletion of start codon; V32L; A47S a ; T49I a ; A53V a ; T54P Deletion of start codon; V32L; A47S a ; T49I a ; A53V a ; T54P Abrogated owing to deletion of start codon V32L; A47S a ; A53V a ; T54P V32L; A47S a ; A53V a ; T54P S45P; 88 Ins V 89 b, P105L; I195M; M198I S45P; I195M; M198I L32I; P68A; P70T; C147R; S204N L9H; L32I; P66A; P70T; C147R; S204N N3D N3D R24S; I25F; F83Y; L88H; L89A; I195M; M198I F83Y; L88H; L89A; I195M; M198I R16G; Y21H; P23L; P34A; T54P C65W; P66A; L84R; W199R S5.2 AY161148 3217 A P92S; Alteration in reading frame C65W; P66A; I82N; from position 4 30; T54P L84R; I86N; L89S; W199R S5.3 AY161149 3220 A P92S; T54P G43E; S55P; P70A; L84R; W199R T13S a ; P15S; V17F T13S; P15S; V17F T13S; P15S; V17 F Y6H; V27A; R151C; S156F; R173L Y6H; F9I; V27A; A41G; I97F; R151C; S156F; R173L F9Y; I97F a ; W62C, E77Q; L119V, T142A; T161L; R165C F9Y; I97F a ; W62C; E77Q; L119V; T142A F9Y; W62C; E77Q; L119V; T142A I97F a ; Deletion of amino acid 151 152 a ; D153R a I97F a ; Deletion of amino acid 151 152 a ; D153R a V27D V27D V27D November 2004 GENOME ANALYSIS OF HBV IN OCCULT INFECTION 1363

Table 6. (continued) Genbank Isolate Variant accession numbers Genome length Gen Pre-S1 Pre-S2 S Precore Core S6 S6.1 AY161150 3182 D T7I; M133I; F134H; A194L Stop codon at position 28 E64D; M66I; A69S; I80T; Q178K S6.2 AY161151 3182 D Y21N Stop codon at amino acid 68 Stop codon at position 28 E64D; M66I; A69S, I80T; Q178K S6.3 AY161152 3182 D T7I; Stop codon at amino acid 6 Stop codon at position 28 E64D; M66I; A69S, I80T; Q178K S6.4 AY161153 3143 D Deletion in pre-s1 65 78; L20P; Y21N F20S; M133I; F134H; A194L Stop codon at position 28 E64D; M66I; A69S, I80T; Q178K S6.5 AY161154 3134 D Deletion in pre-s1 64 80 S6.6 AY161155 3143 D Deletion in pre-s1 65 78; S6.7 AY161156 3134 D Pre-S1 ATG mutated L20P; Y21N Q15R; stop codon at amino acid 68 F20S; M133I; F134H; A194L Stop codon at position 28 Stop codon at position 28 Q15R; stop codon at amino acid 68 Stop codon at position 28 S7 S7.1 AY161157 3182 D T57A; S204R; M205L F4L I3T; S7.2 AY161158 3182 D T57A; S204R; M205L F4L; stop Q178K codon at 18 S8 S8.1 AY161159 3125 D Deletion in pre-s1 61 80 S8.2 AY161160 3125 D Deletion in pre-s1 61 80 H41P E164K Stop codon at 28 H41P E164K Stop codon at 28 S9 S9.1 AY161161 3179 D S6F; A79T c ; P105S; S114T c ; M133I; F134H; C144W; M197I S9.2 AY161162 3179 D S6F; A79T c ; P105S; S114T c ; W156R; S187C; S193A S9.3 AY161163 3179 D S6F; A79T c ; P105S; S114T c ; S187C; S193A a Mutation is similar to wild-type genotype D in genotype A isolate. b Insertion of a residue. c Mutation is similar to wild-type genotype D in genotype A isolate. E64D; M66I; A69S, I80T; Q178K E64D; M66I; A69S, I80T; Q178K E64D; M66I; A69S, I80T; Q178K E64D, M66I, A69S, I80T E64D, M66I, A69S, I80T Stop codon at 88 Stop codon at 88 Stop codon at 88 1364 CHAUDHURI ET AL. GASTROENTEROLOGY Vol. 127, No. 5

November 2004 GENOME ANALYSIS OF HBV IN OCCULT INFECTION 1365 Figure 3. Schematic representation showing types of variations in the pre-s1 S2 region affecting the S promoter. Top, the essential S-promoter region ( ), transcription factor binding sites ( ), B-cell epitopes and start sites of the pre S1/S2/S ATG ( ). Regions corresponding to wild type in all variants are shown by a continuous line, deletions are shown by a textured pattern box, and changes in reading frame are shown by slashes. isolate had wild-type precore. In isolate S9, all variants had a deletion of T at nt. 2164 and 2169, which shifts the reading frame and the core protein is truncated at aa 88. Within the core protein, amino acid changes were distributed randomly (Figure 4). Isolate S4 belonged to genotype A, yet had a stretch of 6 nucleotides (2356 2362) missing, as was characteristic of genotype D. 24 A similar recombinant genotype has been reported recently in the precore region in the Japanese population. 35 In the X protein, the C terminal end, which is known to be critical for transactivation, 36 was of wild-type origin in all isolates except amino acid changes in S3 (E109K, F112L, W120R, A146S, and P147S). In 2 isolates (S3 and S9), the polymerase gene lacked the regular P ATG codon owing to nucleotide substitutions. Therefore, it may be using upstream ATG codons (J ATG and/or C2 ATG). 37 The internal deletions of the pre-s1/s2 region coincide with the spacer region of the polymerase gene, which is tolerant to deletions and substitutions. M552V mutation in isolates S1, S4, and D554Y in isolate S9 was observed in the conserved YMDD domain, even though the patients were not on antiviral therapy. Single nucleotide deletions at position 1350 in isolate S5 is manifested as frame shift (801 813) in the RNase H domain of the polymerase gene and the protein is terminated at aa 814. Distinct Mutations in the Transcriptional Regulatory Regions In the S promoter, the binding site for cellular transcription factors SpA (nt. 3007 3020) was disrupted owing to point mutations in all isolates of S1, S2, S3, and S4 (G3013C), and SpB (nt. 3078 3090) was disrupted in S3.2 ( nt. 3086 15), S3.3 ( nt. 3086 15), S5.1 ( nt. 3025 3208), S6.4, S6.6 ( nt. 3046 3084), S6.5 ( nt. 3041 3089), and S8.1 and S8.2 ( nt. 3034 3090), owing to internal nucleotide deletions described earlier in these isolates. The CCAAT element, which is known to bind nuclear factor- transcription factor and necessary in S- promoter activity, 38 was deleted in isolate S3, and the S5.1.A T2782C change in the conserved TATA box

1366 CHAUDHURI ET AL. GASTROENTEROLOGY Vol. 127, No. 5 Table 7. Changes Seen in Hepatitis B Virus S Protein in Isolates S1 S9 Patient isolate HBV genotype Upstream changes 1 109 110 123 a determinant Downstream changes 1st loop 2nd loop 124 138 139 147 148 160 161 226 Changes in polymerase protein corresponding to downstream S protein I195M; M198I M552V; V555L S1 A T45P; 88V89 (Insertion); P105L a S2 A L9H a ; P66A; P70T C147R S204N No change S3 A N3D M552V; V555L S4 A R24S a ; I25F a ; L32I; L32Y a ; D33S; F83Y; L88H; L89A S5 A G43E a ; S55P a ; C65W; P66A; I68T P70A; I82N a ; L84R; I86N a ; C90S a S6 D Q18R a ; F20S a T130N M133I; F134H I195M; M198I; P214S T670I T114S W199R; I213L V555A; F569Y S193L No change S7 D T57A S204R; L205M S561N; no change S8 D E164K G520E S9 D P105S S114T R122K (subtype-specific change) T130N; M133I a ; F134H a C149W a ; W156R a ; A159G F161Y A168V; S187C; S193A; V194A; M197I a Y506D a ; S507P a ; L512P a ; F543L a ; F549C a ; D554Y a a Mutant at this position exists as a minority population in all the clones sequenced of the particular isolate. (TATATAA CATATAA) in all variants of isolate S9 could disrupt TATA binding protein binding. The enhancer I element, which is known to bind a number of liver-specific transcription factors, 39 was conserved in all isolates. The enhancer II is highly specific and is embedded in the core promoter. It binds to several liver-specific and ubiquitous factors. 40 Of the 3 elements that are essential to enhancer II function, box contained several mutations (S2 [A1645G], S3 [C1653T, A1659T], S5 [T1651C], S7 [A1652G], S9 [C1643A]). NRE and box contained mutations in isolate S3 (G1626, C1709A). The C1653T mutation has been documented previously to be associated with fulminant hepatitis. 41 The precore promoter has been mapped from nucleotide position 1751 1778. The Figure 4. Schematic representation of variations observed in the precore/core region. Start sites of precore and pregenome RNA is indicated by arrows. Sites for DR1 and RNA encapsidation are marked. Codons corresponding to wild type are marked above the continuous line, representing the precore/core open reading frame. Corresponding mutation in the variants (S1 S9) are marked on the lower part.

November 2004 GENOME ANALYSIS OF HBV IN OCCULT INFECTION 1367 earlier documented A1762T/G1764T, which is implicated in the e suppressive phenotype, was not found in any isolate, but G1764T/C1766G mutation coexisted in isolates S5 and S8.2. Mutations in the TATTA box-like motif, which have been shown to bind transcription factors HNF4 and COUP-TF1, were found in isolates S3 (A1772G), S6 (A1775G), and S7 (A1778T). S4 and S9 were the only isolates with wild-type precore promoter. The rest had mutations in this region. The DR1 and DR2 regions were conserved in all isolates. In isolate S3, S5 G1888A, and in isolate S6, S8 G1896A mutation may lead to alteration in the encapsidation stem loop structure. Functional Assay of the Altered Pre-S and Pre-S2/S Promoters The mutation (deletion) observed in the surface promoter region in S3.1, S3.2, S5, and in the interpromoter (deletion) region in S8 isolate could be responsible for down-regulating HBsAg synthesis in these patients. Therefore, functional assay of these mutants was performed. Functional assay of 1 wild type (ayw3) also was performed as control. S6 and S8 had similar deletion and therefore a functional assay of S6 isolate was not performed. The rluc gene was cloned downstream of nucleotide position 2698 175, encompassing the pre-s1 promoter, pre-s2/s promoter, and fused in frame to the S gene start codon. Expression of the rluc gene in the same frame as the envelope protein would provide quantitative evidence of promoter strength in regulating S gene expression. Comparison of rluc gene activity of variants with respect to wild-type promoter showed down-regulation as 3-fold in S3.1, 7-fold in S3.2, and 6-fold in S5.1 in comparison with wild type (Figure 5). Surprisingly, there was a 2.5-fold increase in rluc gene activity in variant S8.1. This particular patient also had a high viral load (10 7 viral copies/ml). A similar deletion in the pre-s region previously was described in a high viral load chronic liver disease patient 42 and also the wild-type isolate SWT.4 nt. ( 3043 3096). All the values were adjusted by use of internal control. Discussion Occult HBV infection is a problem in many parts of the world affecting blood banking, causative diagnosis of chronic hepatitis, and immunization. The present study analyzed the sequence changes in the HBV genome isolated from patients with occult HBV infection and compared it with the wild-type HBV genome prevalent in India. The collection of serum samples at the initial presentation ensured that naturally occurring variants Figure 5. Comparative analysis of the expression levels of the rluc gene from HBsAg-rLUC fusion constructs in promoterless vector prlnull. The wild-type (ayw3 subtype) and variants S1 (nt. 3134 54), S3.2 ( nt. 3086 15), S3.3 ( nt. 3086 15), S5.1 ( nt. 3025 3208), and S8.1 ( nt. 3034 3090) were analyzed. The experiments were equalized by cotransfection with psg-fluc. and not drug-related mutants were selected. The present study was cross-sectional rather than longitudinal. 43 Therefore, it was unclear whether the mutations that were documented arose de novo or were acquired. The possible explanations regarding HBsAg-negative HBV infection has been mutations in the structural protein altering the antigenic properties, thus escaping detection by routine tests. 44,45 Alternatively, the quantity of HBsAg in serum is decreased, and might just be enough for viral assembly but below the sensitivity of standard enzyme-linked immunosorbent assay. In 3 of the cases (S3.1, S3.2, S3.3, S5.1, S6.4, S6.4, S6.5, S6.6, S6.7) with occult infection, long-stretch deletions were observed in the pre-s1/s2 gene, which overlaps the surface promoters. These deletions down-regulate the expression of surface proteins as was observed in our functional studies (Figure 5). In 2 additional patients (S8 and SWT.4), the deletions involve interpromoter regions. In 1 of these cases (S8), there was overexpression of the major (small) HBsAg (Figure 5). For assembly of the envelope particles, specific ratio of major (small) HBsAg and large HBsAg is required. In either of the earlier-described circumstances this ratio is altered, which may alter the HBsAg assembly and secretion 46 in comparison with the wild phenotype, leading to quantitative decrease in HBsAg and HBV in the serum. Cellular accumulation could be documented in one of these (S3) cases with the deletion in pre-s region. Biopsy specimens were not available in the other 2 (S5 and S6). The HBV population defective in one gene may coexist with a wild-type population, helping in rescue of the virus. 47 These

1368 CHAUDHURI ET AL. GASTROENTEROLOGY Vol. 127, No. 5 alterations may be part of a dynamic process in which the existing mutants gradually may die out to be replaced by fresh mutants originating from the wild type. The emergence of quasispecies in these circumstances ensures the persistence of the virus. 48 Further mutation in the YMDD domain, which characteristically was described in the course of lamivudine therapy, 49 also was found in 3 of our patients (S1, S4, and S9) with occult HBV infection. These patients were not on antiviral therapy. In the past, it was believed that YMDD mutation helped in the development of drug resistance. However, YMDD domain mutations are being described increasingly in HBV infection in the absence of antiviral therapy. 50 53 It also is known that polymerase with YMDD mutation results in impaired replication efficiency of the virus. 54,55 This also may decrease the production of HBV and escape detection. This argues well with YMDD mutation in a significant percentage of occult HBV cases as observed in 3 of 9 (33%) in the present study and 6 of 12 (50%) in an earlier study 50. Together, the 2 major changes (deletion in the surface promoter region and YMDD mutation in the polymerase gene) were observed in 6 of 9 cases in the present study, making them the most common observations in occult HBV infection Other sequence variations with known and unknown consequences also were observed. In most isolates (S1, S2, S3, S4), the majority of the sequences matched genotype A and showed variations (point mutations) characteristic of wild genotype D (Table 6). 22 In patient S5, mixed infection with genotype A (S5.2, S5.3) and D (S5.1) (the 33-nucleotide deletion at the N terminus characteristic of genotype D) were detected. Jeantet et al. 56 showed that a mixed infection with replication-incompetent genotype D predominates over replication-competent genotype A, leading to intracellular retention of surface protein and seronegativity for HBsAg. The possible cause for seronegativity in patient S5 might be similar to the earlier-described case. 56 In isolate S4 (genotype A), the core gene resembled wild genotype D. Similarly, in S6 and S9.1, which are of genotype D, genotype A specific substitution at T130N in the a determinant region (Table 7) was observed. These could be owing to recombination between genotype A and D infecting the patient simultaneously. Sequencing of a limited number of clones might not have revealed the entire repertoire of HBV quasispecies present. Currently, it is not known whether genotype plays a specific role in the outcome of infection. 57,58 The protective immunity of the HBV vaccine is associated with a neutralizing antibody specific for the a determinant (aa 124 148). 59 In 2 isolates (S6 and S9), the a determinant had the same consecutive substitution M133I and Y134H. Because the same mutation was observed in 2 different isolates, they could be coexisting and compensatory in nature in maintaining the B-cell epitope. It is unknown as to whether the double-mutation M133I and Y134H also has a role in vaccine escape. It has been observed that mutations in residues apart from aa 141, 144, and 145 in the a determinant could be recognized, albeit at a lower sensitivity, by a panel of monoclonal antibodies directed against the wild type. 60 Isolate S8 merits particular attention because the patient had a high viral (HBV) load and the surface promoter activity was increased by 2.5-times the wild type (Figure 5). The increase in expression level probably could be owing to the transcription factor binding sites, which remain intact despite deletion in the interpromoter region, as well as shortening of the gap between the pre-s1 and pre-s2/s promoter, leading to a commonly observed synergistic effect. 61 This also was observed in isolate SWT.4 with wild-type serology, which carried a deletion from nt ( nt. 3043 3096). It is quite possible that the increased expression of the small HBsAg protein and altered large protein do not cooperate optimally in production and secretion of free HBsAg aggregates. The other mutations in this particular isolate also are significant to understanding plausible reasons for the occult phenotype. An E164K mutation was observed in the vicinity of the a determinant region of surface protein, corresponding to G520E in the polymerase protein. The 2 changes could have acted in combination to produce antigenically altered HBsAg protein and increased replication competency of HBV polymerase. According to the proposed topologic model of HBsAg, 62 amino acids 125 131 and 158 169 are essential for antigenicity of HBsAg. The mutation from glutamic acid to an amino acid carrying positive charge (lysine) could have a significant effect. There have been no functional studies yet reported on HBV polymerase G520E mutation showing increased replication competence. Isolate S8 also had a G1896A mutation, which leads to a more stable stem loop structure and precore defect. Loss of HBsAg from serum has been correlated with a G1896 mutation. 63 Three of 4 genotype D isolates had a precore defect, which could abolish HBeAg synthesis (S6, S8, and S7.2). HBeAg or anti-hbe was not detected in any of the patients. S9 had a premature truncation of the core protein at aa 88, which could lead to assembly of defective virus particles. In the core protein, V27A/D substitution, as observed in isolates S2 and S5, have been reported to affect T-cell receptor contact site and an