Specific Amino Acid Substitutions in the S Protein ... - Journal of Virology

Specific Amino Acid Substitutions in the S Protein Prevent Its Excretion In Vitro and May Contribute to Occult Hepatitis B Virus Infection Subhajit Biswas,a* Daniel Candotti,a,b Jean-Pierre Allaina Department of Haematology, University of Cambridge, Cambridge, United Kingdoma; National Health Service Blood & Transplant, Cambridge, United Kingdomb

Occult hepatitis B virus (HBV) infection (OBI) is defined as low plasma level of HBV DNA with undetectable HBV surface antigen (HBsAg) outside the preseroconversion window period. The mechanisms leading to OBI remain largely unknown. The potential role of specific amino acid substitutions in the S protein from OBI in HBsAg production and excretion was examined in vitro. HBsAg was quantified in culture supernatants and cell extracts of HuH-7 cells transiently transfected with plasmids containing the S gene of eight HBsAgⴙ controls and 18 OBI clones. The intracellular (IC)/extracellular (EC) HBsAg production ratio was ⬃1.0 for the majority of controls. Three IC/EC HBsAg patterns were observed in OBI strains clones: pattern 1, an IC/EC ratio of 1.0, was found in 5/18 OBI clones, pattern 2, detectable IC but low or undetectable EC HBsAg (IC/EC, 7.0 to 800), was found in 6/18 OBIs, and pattern 3, low or undetectable IC and EC HBsAg, was found in 7/18 clones. Intracellular immunofluorescence staining showed that in pattern 2, HBsAg was concentrated around the nucleus, suggesting retention in the endoplasmic reticulum. The substitution M75T, Y100S, or P178R was present in 4/6 pattern 2 OBI clones. Site-directed-mutagenesis-corrected mutations reversed HBsAg excretion to pattern 1 and, when introduced into a control clone, induced pattern 2 except for Y100S. In a control and several OBIs, variants of a given quasispecies expressed HBsAg according to different patterns. However, the P178R substitution present in all cloned sequences of two OBI strains may contribute significantly to the OBI phenotype.

O

ccult hepatitis B virus (HBV) infection/carriage (OBI) is characterized by the presence of very low levels of HBV DNA in plasma and/or in liver with hepatitis B surface antigen (HBsAg) being undetectable using the most sensitive commercial assays, with or without antibodies to hepatitis core antigen (anti-HBc) or hepatitis B surface antigen (anti-HBs), outside the preseroconversion window period (1). OBI represents a particular form of persistent or chronic infection that is encountered globally, albeit at different frequencies depending on endemicity and genotype (2). OBI is a potential source of HBV transmission by transfusion and organ transplantation, can reactivate in association with immunodeficiency or immunosuppressive treatments, and might be a risk factor for liver cancer (2). Several mechanisms have been proposed as responsible for OBI, such as imperfect control by the host immune system (3, 4), multiple amino acid substitutions in the S protein affecting HBsAg detection with commercial immunoassays (5, 6), mutations in regulatory elements negatively affecting virus replication (7, 8), and mutations affecting posttranscriptional mechanisms regulating S protein expression (9, 10). A high frequency of mutations has been observed especially within the major hydrophilic region (MHR) of S protein from OBI strains, and these may alter antigenicity, HBV infectivity, cell tropism, and virion morphogenesis (4–6, 11). In addition, recent studies reported amino acid substitutions in the MHR of OBI strains that impaired virion and/or S protein secretion in in vitrotransfected hepatocyte cell lines and infected mice that might contribute to the OBI phenotype (5, 12, 13). In the present study, the potential role of OBI-specific amino acid substitutions in the S protein, within and outside the MHR, on HBsAg production and excretion was examined in vitro.

7882

jvi.asm.org

Journal of Virology

MATERIALS AND METHODS Cloning and sequencing of the HBV S protein coding region, HBsAg expression plasmid, and site-directed mutagenesis. HBV DNA was isolated from randomly selected plasma samples from 18 previously characterized blood donors with OBI and eight HBsAg⫹ blood donors as controls (10). Fifteen, two, and one OBI donors were infected with HBV genotype B, C, and D, respectively. Two HBsAg⫹ control samples contained genotype B, four genotype C, and two genotype D strains. A detailed description of OBI and HBsAg⫹ control donors is provided in Table 1. The whole HBV genome was initially PCR amplified, and a consensus viral sequence was obtained by direct sequencing of the PCR products as previously described (4, 10). A subgenomic fragment (positions 156 to 1803) containing the S coding region and the HBV posttranscriptional regulatory element (PRE) was further amplified by using primers SPL3 (5=-gcgcgcgctagcacCATGGGGARCAYCRYATCRGGA-3=; nucleotides [nt] 156 to 177) and SPL2 (5=-gcctttgcaagcttCASACCAATTTATGCCTA C-3=; nt 1803 to 1785) (lowercase indicates NheI and HindIII restriction sites introduced for cloning purpose into SPL3 and SPL2 primers, respectively). Amplicons were cloned into the pcDNA3.1⫹ expression vector (Invitrogen, Paisley, United Kingdom) under the control of the human cytomegalovirus (HCMV) immediate early (IE) promoter. The HCMV IE

Received 12 March 2013 Accepted 30 April 2013 Published ahead of print 8 May 2013 Address correspondence to Daniel Candotti, [email protected], or Jean-Pierre Allain, [email protected]. * Present address: School of Life Sciences, University of Lincoln, Lincoln, United Kingdom. S.B. and D.C. contributed equally to this work. Copyright © 2013, American Society for Microbiology. All Rights Reserved. doi:10.1128/JVI.00710-13

p. 7882–7892

July 2013 Volume 87 Number 14

Impaired S Protein Excretion and Occult HBV Infection

TABLE 1 Characteristics of HBsAg⫹ control and OBI donorsa Donor

Status

Origin

Gender

Age (years)

HBV DNA load (IU/ml)

HBsAg (IU/ml) level

Anti-HBc

Anti-HBs level (IU/L)

HBV genotype

M38 M86 M88 M90 M92 M95 E3476 E3789 HK01556 HK3110 HK3475 HK6794 HK6921 HK8442 HK8663 R84 TW0498 TW2256 TW3437 TW4576 TW5004 TW6083 TW6639 TW8964 TW9015 TW9331

Control Control Control Control Control Control Control Control OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI OBI

Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Egypt Egypt Hong Kong Hong Kong Hong Kong Hong Kong Hong Kong Hong Kong Hong Kong Italy Taiwan Taiwan Taiwan Taiwan Taiwan Taiwan Taiwan Taiwan Taiwan Taiwan

M F M F M M NA NA M F M M M M M M M M F M F M M M M M

30 19 23 22 49 26 NA NA 47 49 51 59 50 53 53 44 46 30 49 55 63 62 55 30 47 54

1.6 ⫻ 107 3.1 ⫻ 108 1.1 ⫻ 108 1.6 ⫻ 108 1.0 ⫻ 107 1.5 ⫻ 103 6 31 ⬍5 Not detectable 37 91 Not detectable 46 56 62 48 56 30 ⬍5 7 ⬍5 Not detectable 15 Not detectable 6

8.3 ⫻ 102 2.1 ⫻ 104 3.9 ⫻ 104 1.2 ⫻ 105 1.1 ⫻ 103 1.4 ⫻ 104 3.3 ⫻ 105 1.1 ⫻ 105 Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg

Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos Pos

Neg Neg Neg Neg Neg Neg Neg Neg Neg Neg 26 16 19 98 Neg Neg Neg 18 Neg Neg Neg Neg Neg Neg Neg Neg

C C B B C C D D B B B B B C B D B B B B B C B B B B

a

NA, not available.

gene promoter replaced the natural pre-S2/S promoter, with the transcriptional initiation site being located 75 nucleotides upstream of the 5= end of the S mRNA in order to limit variation in mRNA transcription and consequently examine S protein expression through HBsAg quantification. The HBV DNA insert and ligation sites of selected clones (1 to 19 clones/sample) were sequenced, and S amino acid sequences were derived. Amino acid substitutions in S protein sequences of OBI and control clones were identified by comparison with a consensus sequence of the respective genotype obtained by aligning sequences from HBsAg⫹ blood donors (ⱖ95 sequences per genotype). Site-directed mutagenesis (SDM) of pcDNA3.1-HBV clones was performed using the QuikChange site-directed mutagenesis kit (Agilent Technologies, La Jolla, CA) according to the manufacturer’s instructions. The correct introduction of a mutation was verified by sequencing. Cells and transfection. pcDNA3.1-HBV plasmids were transfected into HuH-7 cells (3 ␮g plasmid/3 ⫻ 105 cells) using FuGENE HD transfection reagent (Promega, Madison, WI) as described previously (10). For each sample, at least two independent transfection experiments were performed in triplicate. Analysis of extracellular and intracellular HBsAg production. Extracellular (EC) HBsAg production was tested in culture supernatants 72 h posttransfection using the Murex HBsAg v3 enzyme immunoassay (DiaSorin, Dartford, United Kingdom). A standard curve was generated by testing serial dilutions of a commercial calibrated recombinant HBsAg (Source BioScience, Nottingham, United Kingdom) for quantification. To evaluate intracellular (IC) HBsAg production, supernatant-free cell monolayers were washed three times with 1⫻ phosphate-buffered saline (PBS) and lysed in 500 ␮l of 1⫻ PBS, 1% Triton X-100, 10 units/ml DNase I, and complete proteinase inhibitor/EDTA (Roche, Basel, Switzerland). Cell lysates were centrifuged at 13,000 rpm for 15 min, and supernatants were collected as cytosolic extracts. HBsAg in extracts was quantified as described above. The total amount of HBsAg in each transfected culture of 3 ⫻ 105


HuH-7 cells was calculated as the sum of HBsAg present in IC extract and EC supernatant. For each sample, the mean IC and EC HBsAg production was calculated from 2 to 4 independent transfection experiments each, including three replicates. For comparison, HBsAg production in each culture was normalized according to the transfection efficiency calculated from the percentage of cells expressing ␤-galactosidase activity. Immunofluorescence detection of intracellular HBsAg. HuH-7 cells grown on glass coverslips and transfected with pcDNA3.1-HBV plasmids were fixed at 72 h posttransfection using 4% paraformaldehyde for 30 min at room temperature. The fixed cells were washed with 1⫻ PBS, blocked with 1% bovine serum albumin (BSA)– 0.5% Triton X-100 in PBS for 30 min and incubated for 90 min in the presence of Alexa Fluor 488-labeled mouse anti-HBsAg monoclonal IgG (Invitrogen). Cells were washed and coverslips were mounted on slides using Prolong Gold antifade reagent with DAPI (Invitrogen). Cells were examined by immunofluorescence microscopy (60⫻ oil immersion lens and 10⫻ objective) for detection of intracellular HBsAg in cells at a magnification of ⫻600. Transmembrane structure prediction of HBsAg variants. The transmembrane distribution of S protein was predicted using the SOSUI server (http://bp.nuap.nagoya-u.ac.jp/sosui/). Other programs were used to predict the S protein transmembrane domains: DAS, TopPred, and HMMTOP (topology prediction programs, Expasy Bioinformatics Resource Portal [http://www.expasy.org/tools/). Results were compared to the generally used model by Persson and Argos (14). Statistical analysis. IC and EC HBsAg production for M88 cl4 wild type (wt) and its different mutants at position 178 were compared by means of one-way analysis of variance (ANOVA) with Tukey’s multiple comparison (post hoc) test. Differences between HBsAg productions (total or EC) for OBI strains before and after SDM repair were compared using Student’s t test (two-tailed for unpaired data), and the variances of the data were measured by the F test. P values of ⬍0.05 were considered statistically significant.

jvi.asm.org 7883

Biswas et al.

FIG 1 Alignment of amino acid sequences deduced from cloned S genes of genotype B to D non-OBI (HBsAg⫹) and OBI HBV strains. Sequences of HBsAg⫹

and OBI clones were aligned with consensus sequences derived from 124 wild-type/HBsAg⫹ genotype B sequences, 95 genotype C sequences, and 150 genotype D sequences. Residues identical to the reference consensus are indicated by dots. Site-directed mutagenesis of residues in gray boxes resulted in a change in the HBsAg excretion pattern; no change was observed when residues in open boxes were corrected.

RESULTS

In vitro HBsAg production patterns. Eighteen HBV strains carrying multiple amino acid substitutions in the S coding region were selected from a repository of viral strains from blood donors with previously characterized OBI (Table 1 and Fig. 1). Eight HBsAg⫹ strains from blood donors (HBsAg plasma concentration, ⬎104 IU/ml) were used as non-OBI controls. One clone per OBI and control sample was randomly selected for transfection in HuH-7 cells and HBsAg production in vitro characterized by evaluating the total amount of detectable IC and EC HBsAg, the IC/EC HBsAg ratio, and the IC distribution pattern of HBsAg detected by immunofluorescence assay (IFA). In HBsAg⫹ controls, the average normalized total HBsAg production was ⬎120 ng/3 ⫻ 105 cells (range, 123 to 613 ng) at 72 h posttransfection (Fig. 2). Six control clones showed an IC/EC HBsAg ratio ranging between 0.2 and 3.0. HBsAg intracellular immunofluorescence staining was diffuse and finely granular across the hepatocyte cytoplasm (Fig. 3). Two genotype C control clones (M92-cl2 and M95-cl8) showed IC/EC ratios of ⬎10.0 (Fig. 2A), and IC HBsAg concentrated in the perinuclear area as densely packed fluorescence (data not shown). Three distinct patterns of HBsAg production were identified in cells transfected with OBI clones (Fig. 2A to D). Pattern 1 (n ⫽ 5/18) was similar to the features observed in the majority of controls, with total HBsAg production of ⬎120 ng/3 ⫻ 105 cells

7884

jvi.asm.org

(range, 146 to 621 ng) (Fig. 2B), IC/EC HBsAg ratio of ⱕ3 (range, 0.1 and 3) (Fig. 2C), and diffuse granular fluorescent staining of IC HBsAg. Pattern 3 (n ⫽ 7/18) was characterized by total HBsAg of ⬍50 ng (0.6 to 46 ng) and an IC/EC HBsAg ratio ranging between noncalculable and 7, with low or no fluorescence staining of IC HBsAg (Fig. 3). For TW6639-cl1 (HBsAg, 46 ng/3 ⫻ 105 cells; IC/EC, 7) and TW2256-cl3 (HBsAg, 31 ng/3 ⫻ 105 cells; IC/EC, 2), the IC HBsAg appeared as sparsely distributed but bright green fluorescent dots in the cytoplasm (not shown). Pattern 2 (n ⫽ 6/18) was characterized by total HBsAg ranging between 60 and 189 ng nearly exclusively found intracellularly. The IC/EC HBsAg ratio ranged between 7 and 800 (median, 10). The intracellular HBsAg fluorescence appeared dense, compact, and essentially limited to the central part of the cytoplasm, adjacent to the nucleus, in 5 of 6 pattern 2 OBIs, as observed with M92-cl2 and M95-cl8 (Table 2). The IC HBsAg for TW9015-cl1 (HBsAg, 91 ng/3 ⫻ 105 cells; IC/EC, 12) presented as diffuse but very compact fluorescence filling the entire cytosol (not shown). No significant association of HBsAg production patterns with HBV genotype, plasma HBV DNA load, or anti-HBs status was observed (Table 1 and Fig. 2A). Figure 2D summarizes the relation between HBsAg production and IC/EC ratio in controls and patterns 1 to 3. Characterization of amino acid substitutions associated with HBsAg impaired excretion in HuH-7 cells. In pattern 2 OBI clones, HBsAg production was perfectly detectable and reached

Journal of Virology

FIG 2 Detection of intracellular and extracellular levels of HBsAg/S protein by EIA and IFA after transfection of cloned S sequences in HuH-7 cells. (A) Average individual HBsAg production in HuH-7 cells transfected with 8 non-OBI controls and 18 OBI clones. Standard deviations are for three or more culture wells; the three different production patterns observed in OBI samples are indicated. (B) Comparison of total HBsAg production observed for the non-OBI and the three OBI patterns. Statistically significant (P ⬍ 0.05) differences determined by 1-way ANOVA/Tukey test are indicated. (C) Comparison of HBsAg IC/EC ratios observed for each control and OBI patterns. No statistical significance was observed due to three outliers with remarkably high ratios (OBI pattern 2 isolate HK01556-cl2 and non-OBI isolates M92-cl2 and M95-cl8, showing IC/EC ratios of 800, 75, and 23, respectively). (D) Relationship between IC/EC ratio and total HBsAg production. Geometric means are indicated with dashed vertical and horizontal lines; gray squares and circles identify non-OBI controls and pattern 1 OBIs with diffuse granular IF, black squares and triangles identify non-OBI controls and pattern 2 OBIs with dense aggregated IF, and open diamonds identify pattern 3 OBIs with low/undetectable HBsAg in IF.


jvi.asm.org 7885

Biswas et al.

flected in the decrease of HBsAg IC/EC ratio observed (7 to 801 versus 0.125 to 4 after SDM). This finding was confirmed in all four samples by immunofluorescence analysis, showing a change following SDM from the dense intracellular accumulation of HBsAg associated with pattern 2 to a diffuse fluorescence across the cytoplasm typical of pattern 1 (Fig. 4A). These data were reproduced in the hepatocyte cell line HepG2 (data not shown). The P178R substitution was present in the pattern 3 TW8964-cl1 sample, but the SDM restoration of proline did not produce any perceptible change, possibly due to the apparently extremely low HBsAg production that characterizes pattern 3 OBIs (Table 2). The P178R substitution was not found in the sequences of 369 strains (124 genotype B, 95 genotype C, and 150 genotype D) from HBsAg⫹ donors, whereas M75T and Y100S were present in one HBsAg⫹ sequence each. To confirm the effect of M75T, Y100S, and P178R on HBsAg excretion, these three substitutions were introduced into the ge-

TABLE 2 Impact of OBI-specific amino acid substitutions on HBsAg production pattern in vitro Average total HBsAg (range) (ng)b

Deduced HBsAg IC/EC IFA Production ratio patternc pattern

HK01556-cl2 None T75 M R105P P154S A184V

189 (93–254) 226 (127–310) 72 (47–103) 214 (213–229) 97 (83–109)

800 0.125 41 464 73

R WT R R R

2 1 2 2 2

HK6794-cl2

60 (51–92) 325 (307–362) 102 (96–107) 39 (24–70)

8 3 7 9

R WT R R

2 1 2 2

TW9015-cl1 None 91 (56–143) S111P⫹E112G 91 (82–99) E119G 93 (89–99)

12 13 11

R R R

2 2 2

TW0498-cl2 None R129Q

18 (13–23) 21 (19–23)

13 22

LF LF

3 3

TW0498-cl3 None R129Q

135 (62–217) 63

15 14

R R

2 2

HK3110-cl4

None V159A⫹N160K P176S⫹R178P R178P N226I

116 (88–131) 188 (184–192) 284 (267–302) 757 (692–809) 23 (18–25)

20 20 2 4 22

R R WT WT R

2 2 1 1 2

HK3475-cl6

None R178P

75 (49–103) 7 464 (409–541) 2

R WT

2 1

NF NF

3 3

FIG 3 Immunofluorescence microscopy of HuH-7 cells expressing non-OBI and OBI HBsAg. Cell nucleus were stained with DAPI (blue) and HBsAg was detected with Alexa Fluor 488-labeled mouse anti-HBsAg monoclonal IgG (green). Cells transfected with a reporter plasmid expressing LacZ were used as negative controls.

significantly higher levels than that in pattern 3 OBI clones (P ⫽ 0.003) but was lower than that in both pattern 1 OBI clones (P ⫽ 0.011) and controls (P ⫽ 0.007). Nevertheless, the high IC/EC ratio associated with pattern 2 suggested an impaired HBsAg excretion in transfected HuH-7 cells that was investigated further here. The deduced S amino acid sequences of the six OBI genotype B pattern 2 clones (HK01556-cl2, TW0498-cl3, HK3110-cl4, HK3475-cl6, TW9015-cl1, and HK6794-cl2) showed 12 to 25 amino acid substitutions (mean, 17) compared to the HBV consensus genotype B sequence obtained from 124 HBsAg⫹ sequences (Fig. 1). Substitutions at positions 3 to 5 were not considered, as they resulted from the use of the degenerated primer SPL3 in the initial genomic amplification procedure. The potential association of 17 OBI-specific substitutions with pattern 2 was investigated by restoring the wild-type consensual residues using SDM (Table 2). Eight of the selected substitutions were unique to the pattern 2 sequences studied (M75T, P105R, K160N, W165R, L176P, W182C, V184A, and I226N), whereas one (S167L) and six (Y100S, P111S, G112E, G119E, S154P, and P178R) were also present in OBI pattern 1 and pattern 3 sequences, respectively. In addition, Q129R and A159V were found in sequences associated with all three patterns. The HBsAg production pattern was not modified when 14 selected amino acids were restored to the wild-type residues by SDM individually or in combinations (Table 2). In contrast, restoration of methionine 75 in sample HK01556-cl2, tyrosine 100 in sample HK6794-cl2, and proline 178 in samples HK3110-cl4 and HK3475-cl6 resulted in a significant increase of HBsAg production (P ⫽ 0.02 to 0.05, except HK01556-cl2) that was associated with increased HBsAg excretion in culture supernatants, as re-

7886 jvi.asm.org

OBI clone

Substitution repaireda

None S100Y R165W⫹L167S C182W

TW8964-cl1 None R178P

0.6 (0.3–0.8) 2 (1.8–2.1)

NAd NA

a Mutations in OBI sequences (in bold) were repaired by SDM. None, the native OBI sequence was tested. b Cumulative amount of HBsAg (ng) in cytosol and culture supernatant. c IFA detection of intracellular S protein. WT, diffused granular pattern; R, retained dense packed fluorescence close to the nucleus; LF, low level fluorescence; NF, no fluorescence detected. d NA, not applicable.

Journal of Virology


FIG 4 Immunofluorescence staining of native and mutated HBsAg in transfected HuH-7 cells. Amino acid substitutions introduced by site-directed mutagenesis in OBI (A) and non-OBI (B) sequences are indicated in bold.

notype B HBsAg⫹ control M88-cl4. OBI pattern 2 was reproduced in M88-cl4 following the M75T and P178R changes but not the Y100S change, as evidenced by IC/EC ratio evaluation and intracellular immunofluorescence analysis (Table 3 and Fig. 4B). In addition, the introduction of the mutation P178R by SDM in an entire HBV genotype C genome changed the HBsAg production pattern 1 initially observed in HuH-7 transfected cells to pattern 2 (data not shown). In contrast, the substitutions P111S⫹G112E, G119E, Q129R, I150T, S154P, and W165R introduced into the

control M88-cl4 did not substantially affect the HBsAg production pattern (Table 3). The effect of amino acid changes at position 178 on M88-cl4 HBsAg phenotype was investigated further (Table 3). Substitutions P178A, P178E, P178K, P178L, and P178Q resulted in a substantially decreased overall average production of HBsAg (31 to 80 ng) compared to M88-cl4 wt (270 ng), as observed with P178R (78 ng). However, the corresponding IC/EC HBsAg ratios were slightly higher (IC/EC ⫽ 2 to 5) than observed for M88-cl4 wt

TABLE 3 Impact of OBI-specific substitutions introduced in the HBsAg⫹ control M88-cl4 on HBsAg production in vitro Substitutions introduced by SDM a

None M75T Y100S P111S⫹G112E G119E Q129R I150T S154P W165R P178R P178Q P178L P178A P178K P178E

Avg total HBsAg (range) (ng) Totalb

EC

HBsAg IC/EC ratio

270 (203–338) 48 (39–56) 475 (456–487) 482 (456–500) 402 (359–433) 439 (410–471) 102 (74–112) 337 (296–392) 138 (112–161) 78 (76–81) 63 (45–76) 80 (70–82) 31 (27–37) 41 (31–48) 136 (126–143)

197 (167–2270) NDd 316 (308–321) 347 (334–354) 241 (212–256) 278 (256–302) 71 (52–80) 214 (191–243) 70 (49–84) 5 (5–6) 11 (11–12) 19 (18–21) 11 (10–13) 12 (10–15) 99 (92–104)

0.33 NAe 0.5 0.33 0.5 0.5 0.5 0.5 1 15 5 3 2 3 0.33

IFA patternc

Deduced HBsAg production pattern

WT R WT WT WT WT WT WT WT R WT WT WT WT WT

1 2 1 1 1 1 1 1 1 2 Indeterminate Indeterminate Indeterminate Indeterminate 1

a

Wild type M88-cl4 control. Cumulative amount of HBsAg (ng) in cytosol and culture supernatant. c IFA detection of intracellular S protein. WT, diffused granular pattern; R, retained dense packed fluorescence close to the nucleus; NF, no fluorescence detected. d ND, not detected. e NA, not applicable. b


jvi.asm.org 7887

Biswas et al.

as OBI strains circulate as quasispecies of related variants not necessarily represented by a single clone tested in vitro. The genetic diversity and the relevance of specific mutations as potential explanation for the OBI phenotype were therefore examined by sequencing multiple clones from several donor samples containing pattern 2 clones (Fig. 6). The diversity of patterns was shown in one control (M92; two clones with pattern 2 and one clone with pattern 3) and three OBIs: TW0498 (one pattern 2 and one pattern 3), TW8964 (one pattern 1 and one pattern 3) and HK6794 (3 pattern 1, one pattern 2, and one pattern 3) (Fig. 6). In OBI HK6794, the average amino acid diversity between clones was 20 residues (range, 11 to 32 residues) without a significant relationship with the HBsAg phenotype (data not shown). The Y100S mutations responsible for pattern 2 of clone 2 were present in a subcluster of five clones associated with clone 2, but Y100F was present in all 12 other clones. In contrast, in OBIs HK01556, HK3110, and HK3475, genetic diversity was very limited (Fig. 6), and each of the 8 to 12 clones contained Y100S (HK01556) or P178R (HK3110 and HK3475). For these three OBI strains, it is likely that these two mutations shown to prevent HBsAg excretion contribute to the OBI phenotype. DISCUSSION

FIG 5 Immunofluorescence staining of HBsAg in cells transfected with wildtype and mutated M88-cl4 sequences. The position and nature of the OBIspecific amino acid substitutions introduced into the genotype B control M88cl4 are indicated.

(IC/EC ⫽ 0.33) but still compatible with pattern 1 definition and lower than the pattern 2-related IC/EC ratio of the P178R mutant (IC/EC ⫽ 15). The immunofluorescence (IF) results obtained with P178A/K/L/Q did not clearly match either pattern 1 or pattern 2 profiles and were considered indeterminate (Fig. 5). Mutant P178E showed a pattern 1-related IC/EC HBsAg ratio and IFA profile similar to that of M88-cl4 wt. As shown in Fig. 1, P178Q was present in a pattern 1 and a pattern 3 OBI clone (TW5004-cl3 and TW3437-cl5, respectively). Several modeling programs predicted very similar arrangements of the S protein in the bilayer lipid membrane of the endoplasmic reticulum (ER). Persson and Argos’ model as well as SOSUI’s and DAS’s identified the same four transmembrane (TM) domains for wild-type S. However, SOSUI predicted a fifth TM domain (amino acids 149 to 171) that we considered doubtful. Nevertheless, the three amino acid substitutions identified as critical to HBsAg excretion had identical locations predicted by the three models. M75T was located in the cytosolic loop separating TM1 and TM2, Y100S was at the end of TM2, and P178R was in TM3. HBsAg excretion deficit as potential contributor to OBI phenotype. Amino acid substitutions impairing HBsAg excretion reflected by HBsAg quantification in vitro have been identified in individual OBI clones. However, HBVs in HBsAg-positive as well

7888

jvi.asm.org

High mutation rates in the HBsAg in OBI strains from various geographical origins have been reported (4, 5, 8, 10). Some of these mutations have been functionally associated with structural changes in the HBsAg, leading to impaired detection by current immunoassays (5, 6). However, the extremely low viral load generally observed in OBI carriers suggests that S mutations may also negatively affect either viral replication or HBsAg production in circulation. To test the second hypothesis, the effect of amino acid substitutions specific to genotype B to D OBI strains on HBsAg expression was investigated in vitro. The functional analysis of clones derived from 18 previously characterized OBI strains with S mutations not present in 369 HBV genotype B to D strains infecting HBsAg⫹ blood donors identified three patterns of HBsAg expression (Fig. 1). These patterns were defined by the total HBsAg production estimated by enzyme immunoassay (EIA), the IC/EC HBsAg ratio, and the intracellular distribution of HBsAg detected by immunofluorescence (Fig. 2 and 3). However, two genotype C controls (M92-cl2 and M95-cl8) presented an IC/EC ratio and IFA pattern similar to OBI pattern 2, but total HBsAg was similar to pattern 1. In addition, pattern 2 TW9015-cl1 and pattern 3 TW6639-cl1 and TW2256-cl3 IFA appeared to be intermediate between pattern 2 and 3. Nevertheless, there were clear differences in HBsAg expression properties between HBsAg⫹/wild-type controls and OBI strains and between OBIs (Fig. 2A to D). The differences in total HBsAg production between HBV strains, used as a criterion to define the three patterns, might be due to variations in the effectiveness of the HBsAg expression vector transfection into HuH-7 cells despite normalization using cotransfection with a reporter plasmid and the mean (⫾ standard deviation [SD]) of total IC and EC HBsAg in triplicate (Fig. 2). In addition, using a CMV promoter-driven expression vector, protein overexpression may be responsible for the phenotypes observed in vitro. Similarly, discrete variations in HBsAg production may be masked by high protein expression levels. However, similar IC and EC HBsAg levels were obtained

Journal of Virology


FIG 6 Phylogenetic analysis of the S amino acid sequences of multiple clones from OBI and non-OBI strains. Phylogenetic analysis was performed as previously described (10). HBV reference sequences of genotypes/subgenotypes A1-3 and B-H are identified by their GenBank accession numbers. For clones used in HuH-7 transfection experiments, the resulting HBsAg excretion pattern is indicated by open (pattern 1), black (pattern 2) or gray (pattern 3) arrows.

when HuH-7 cells with an entire HBV genotype C genome were transfected (data not shown). Nevertheless, total HBsAg, together with the intracellular HBsAg immunofluorescence pattern, clearly differentiated patterns 1 and 2 from pattern 3 (Fig. 2B and Fig. 3). For the latter, a low total HBsAg level (⬍50 ng) was supported by no or low IC HBsAg immunofluorescence (Fig. 3). Further studies are needed to establish whether the lack of HBsAg detection in pattern 3 OBIs is related to HBsAg production being terminated by an uncharacterized posttranscriptional mechanism, as previously suggested (9, 10), or to mutant HBsAg unrecognized by the EIA and IFA (5, 6). The


average amino acid substitution rates in the MHR and over the S protein were similar in OBI sequences irrespective of HBsAg pattern (MHR, 6.4% in pattern 1 versus 8.6% in pattern 2 versus 9.1% in pattern 3; S protein, 5.7% in pattern 1 versus 7.2% in pattern 2 versus 6.9% in pattern 3) but significantly higher in OBI than in non-OBI sequences (0.9% and 1.9% in nonOBI MHR and S protein, respectively) (Fig. 1). However, mutations D144E and G145R, previously associated with reduced HBsAg detection (15), were present in three pattern 3 clones (TW6639-cl1 and TW3437-cl5) or in association (HK8663-cl2) and supported the immune detection escape hypothesis.

jvi.asm.org 7889

Biswas et al.

In individuals carrying pattern 1 and 2 clones, OBI did not seem related to antigenic variation in HBsAg, since when produced in vitro, it was detected with commercial assays. Mutations in other HBV regions might be responsible for OBI in individuals carrying pattern 1 clones. Mutations in the S gene may also result in a modified reverse transcriptase region of the HBV polymerase, decreasing HBV replication (15). The reduced level of HBsAg excretion observed in pattern 2 OBIs may also contribute to the genesis of OBI by maintaining HBsAg level below the detection limit of currently used serological assays and by concomitantly affecting virion release as suggested by recent studies (5, 12, 13, 16, 17). SDM experiments that involved repairing an OBI-specific mutation(s) (Table 2 and Fig. 4A) and introducing this mutation(s) into a control/wild-type sequence (Table 3 and Fig. 4B) identified three mutations—M75T, Y100S, and P178R—that contributed to decreased HBsAg excretion in vitro. The M75T mutation was unique to HK01556-cl2 but was also present in the consensus sequence of one genotype C HBsAg⫹ control (n ⫽ 1/369; 0.27%). The Y100S mutation was observed in one genotype A2, six genotype B, and four genotype D OBI strains (n ⫽ 11/176; 6.25%), and in one genotype D HBsAg⫹ control (0.27%). The P178R mutation was detected in the consensus sequence of three genotype B and two genotype D OBI strains representing 2.8% of 176 OBI strains previously studied (data not shown). The P178R mutation was absent in 369 HBsAg⫹ sequences of genotypes A to D and was considered OBI specific. The rare M75T substitution was associated with HBsAg excretion defect. This mutation is located in the C-terminus part of the protein cytosolic loop that contains a putative core-envelope interaction domain important for both virions and HBsAg secretion (16). In previous studies, replacement of R73, R78, and R79 by uncharged residues, and the naturally occurring mutation L77R reduced HBsAg secretion (16, 18). The highly restricted perinuclear IFA staining pattern observed for mutant M75T in the present study (Fig. 4A) was consistent with the HBsAg retention in the ER-Golgi reported for mutant L77R (16). Mutant M75T provides additional indirect support to a role of the cytosolic loop C terminus in HBsAg secretion independently of its putative role in virion morphogenesis. Three genotype B OBI clones—HK01556-cl2, HK6794-cl2, and TW6639-cl1—with HBsAg excretion defect contained the Y100S mutation (Fig. 1 and Fig. 2). When repaired in HK6794-cl2, excretion of HBsAg was recovered and changed from pattern 2 to pattern 1. However, Y100S alone did not cause HBsAg retention when introduced in M88-cl4 control (Table 3) or when naturally present in the T75M-repaired HK01556-cl2 clone (Table 2 and Fig. 4), suggesting that the negative effect of mutant Y100S on HBsAg excretion may be corrected by other, yet-uncharacterized, S envelope mutation(s) in M88-cl4 and HK01556-cl2. Similarly, the single mutation W74L has been reported to suppress the retention phenotype of L77R (16). A Y100C substitution was present in two pattern 3 OBI clones (HK8663-cl2 and TW3437-cl5) in agreement with previous reports associating this substitution with HBsAg-negative phenotype in OBI cases (12, 19–21). However, it was also found in all pattern 2/3 clones of the M92 control, indicating that Y100C does not play a direct role in reducing total HBsAg amounts or HBsAg reactivity with commercial assays, as suggested by Mello and coauthors (12). Previous studies have shown that both virions and HBsAg secretion were affected by mutations within three of the putative

7890

jvi.asm.org

transmembrane (TM) alpha-helix domains TM1, TM2, and TM4 of the S protein (18, 22, 23). Mutations in TM2 and TM4 may affect (i) possible intramolecular interactions between TM domains within the S protein, resulting in altered protein folding and defective insertion into the ER membrane, or (ii) intermolecular interactions with the peptide chains of other S proteins essential for HBsAg morphogenesis (11, 23, 24). In the present study, the P178R substitution in the N-terminal part of the putative TM3 prevented HBsAg protein excretion in the two distinct HK3110cl4 and HK3475-cl6 OBI clones and in the mutated M88-cl4 control. Introduction of a positively charged arginine residue in place of a proline (a residue commonly found as the first residue of an alpha helix) may modify this transmembrane domain and affect the excretion of the modified protein. The presence of a potentially charged residue within the TM domain of a typical integral membrane protein was shown to result in retention and rapid degradation in the ER (25). This is in agreement with the reduced total amount of HBsAg and its intracellular location characterizing OBI pattern 2. These effects of charged residues appeared to correlate with the level of free energy required to partition charged chains into a lipid bilayer (25), and it may explain the similar, albeit lessened, effect of substitutions with other strongly polar residues, including lysine and glutamine (Table 3). However, limited changes induced by nonpolar residues alanine and leucine and the lack of significant change with glutamate (Table 3) suggest that these phenotypic changes may be also dependent on their position within the transmembrane sequence and on the nature of the amino acid side chains. Nevertheless, these data show that, like the three other transmembrane domains, TM3 plays a role in the morphogenesis of HBsAg. However, the topology of the HBsAg carboxy-terminal transmembrane domains is not precisely known, and structural predictions rely on models that are continuously refined. Defective HBsAg secretion was also reported to be associated with substitutions in the MHR of OBIs (5, 13, 17, 26, 27). Some of these substitutions were observed in the OBI sequences studied here, but their negative effect if any on HBsAg phenotype remained unclear. For example, serine at position 126, previously associated with a moderate decrease of HBsAg secretion (5), was present in OBI pattern 1 TW5004-cl3, with no evidence of a secretion defect. Similarly, Q129R was found in four OBI clones irrespective of their HBsAg excretion pattern (pattern 1 clone TW4576-cl3, pattern 2 clone TW0498-cl3, and pattern 3 clones TW6639-cl1 and TW8964-cl1). Moreover, the correction R129Q did not restore efficient HBsAg excretion in TW0498-cl3 (Table 2), and the M88-cl4 excretion pattern was not modified by Q129R (Table 3). The substitution G145A was reported to impair HBsAg secretion in a genotype A OBI strain, but it showed no obvious effect in an Asian strain in another study (5, 13). This substitution was present in genotype B OBI TW0498-cl3 clone with pattern 2. Similarly, D144E, previously reported to impair HBsAg detection/ secretion as mentioned above, was also observed in pattern 1 HK6921-cl3. Together, these data suggest that alteration of HBsAg secretion can be due not only to a single substitution but also more frequently to a combination of amino acid substitutions in different regions of the protein. This is supported by reports showing both positive and negative transcomplementation effect of S mutations on HBsAg secretion inhibition (13, 16, 17). It might be difficult to evaluate the exact importance of mutations altering HBsAg secretion in the genesis of OBI, since such

Journal of Virology


mutations appear to be rare and strain specific in most cases. In addition, the HBsAg phenotype associated with these mutations was generally examined from individual clones that may not be representative of the variants constituting the quasispecies in infected individuals, as observed for M75T carriage in HK01556 sample (data not shown). In contrast, the presence of P178R in every closely related variant within HK3110 and HK3475 quasispecies strongly supports an association between HBsAg excretion defect and OBI phenotype in these donors. A more complex situation was observed in HK6794 diverse quasispecies reflected by diversity in the patterns of HBsAg phenotype, suggesting that the HBsAg phenotype of only few clones may not truly represent the overall strain phenotype. At present, the possible pathogenic role of intracellularly retained HBsAg in the development of liver disease in individuals with OBI can be only speculated on. However, impaired secretion of mutated misfolded/unfolded proteins is known to induce ER stress that can affect host cell physiology by activating intracellular transduction pathways (28). In human hepatocytes, accumulation of HBV surface proteins in the ER has been reported to cause the formation of ground glass hepatocytes and to induce oxidative stress, cellular DNA damage, and mutations (29). In a transgenic mouse model, the accumulation of nonsecretable HBsAg particles within the hepatocyte ER appeared to cause severe and prolonged liver dysplasia and damage accompanied by continual liver inflammation, regenerative hyperplasia, transcription deregulation, aneuploidy, and the eventual development of hepatocellular carcinoma (HCC) (30). Altogether, these studies suggest that cytosolic retention of the HBV surface proteins in hepatocytes may have an oncogenic potential by inducing hepatocyte stress response pathways that may stimulate transformation processes. In this context, long-term OBI carriers infected with secretion-defective HBV variants might be at high risk of developing HCC if sufficient time is allowed to elapse. In conclusion, the high genetic variability observed in the S gene of OBI strains is associated with heterogeneous patterns of intracellular and extracellular HBsAg production in vitro. The identification of three new OBI-specific mutations—M75T, Y100S, and P178R— associated with HBsAg intracellular retention supports further the hypothesis that the lack of HBsAg secretion may significantly contribute to the multifactorial occurrence of OBI. M75T and P178R mutants provide additional indirect support for a role of the cytosolic loop and the third transmembrane domain in HBsAg morphogenesis. Further studies are needed to investigate a potential pathogenic role of intracellularly retained HBsAg in the development of liver disease in OBI carriers. ACKNOWLEDGMENTS S. Lin from the Taiwan Blood Services Foundation, Taipei, Taiwan, and C. Kit Lin from Hong Kong Red Cross Blood Transfusion Centre, Hong Kong, People’s Republic of China, are thanked for providing OBI samples studied. S. Biswas was supported in part by a grant from NHSBT England and by a grant from Novartis Diagnostics & Vaccines.

REFERENCES 1. Raimondo G, Allain JP, Brunetto MR, Buendia MA, Chen DS, Colombo M, Craxi A, Donato F, Ferrari C, Gaeta GB, Gerlich WH, Levrero M, Locarnini S, Michalak T, Mondelli MU, Pawlotsky JM, Pollicino T, Prati D, Puoti M, Samuel D, Shouval D, Smedile A, Squadrito G, Trépo C, Villa E, Will H, Zanetti AR, Zoulim F. 2008.


2. 3.

4.

5.

6. 7. 8.

9. 10.

11. 12.

13.

14. 15. 16.

17. 18. 19.

20.

21.

Statements from the Taormina expert meeting on occult hepatitis B virus infection. J. Hepatol. 49:652– 657. Raimondo G, Pollicino T, Romano L, Zanetti AR. 2010. A 2010 update on occult hepatitis B infection. Pathol. Biol. 58:254 –257. Pollicino T, Raffa G, Costantino L, Lisa A, Campello C, Squadrito G, Levrero M, Raimondo G. 2007. Molecular and functional analysis of occult hepatitis B virus isolates from patients with hepatocellular carcinoma. Hepatology 45:277–285. Candotti D, Grabarczyk P, Ghiazza P, Roig R, Casamitjana N, Iudicone P, Schmidt M, Bird A, Crookes R, Brojer E, Miceli M, Amiri A, Li C, Allain JP. 2008. Characterization of occult hepatitis B virus from blood donors carrying genotype A2 or genotype D strains. J. Hepatol. 49:537– 547. Huang CH, Yuan Q, Chen PJ, Zhang YL, Chen CR, Zheng QB, Yeh SH, Yu H, Xue Y, Chen YX, Liu PG, Ge SX, Zhang J, Xia NS. 2012. Influence of mutations in hepatitis B virus surface protein on viral antigenicity and phenotype in occult HBV strains from blood donors. J. Hepatol. 57:720 – 729. El Chaar M, Candotti D, Crowther RA, Allain JP. 2010. Impact of hepatitis B virus surface protein mutations on the diagnosis of occult hepatitis B virus infection. Hepatology 52:1600 –1610. Chaudhuri V, Tayal R, Nayak B, Acharya SK, Panda SK. 2004. Occult hepatitis B virus infection in chronic liver disease: full-length genome and analysis of mutant surface promoter. Gastroenterology 127:1356 –1371. Fang Y, Teng X, Xu WZ, Li D, Zhao HW, Fu LJ, Zhang FM, Gu HX. 2009. Molecular characterization and functional analysis of occult hepatitis B virus infection in Chinese patients infected with genotype C. J. Med. Virol. 81:826 – 835. Hass M, Hannoun C, Kalinina T, Sommer G, Manegold C, Günther S. 2005. Functional analysis of hepatitis B virus reactivating in hepatitis B surface antigen-negative individuals. Hepatology 42:93–103. Candotti D, Lin CK, Belkhiri D, Sakuldamrongpanich T, Biswas S, Lin S, Teo D, Ayob Y, Allain JP. 2012. Occult hepatitis B virus infection in blood donors from South East Asia: molecular characterization and potential mechanisms of occurrence. Gut 61:1744 –1753. Salisse J, Sureau C. 2009. A function essential to viral entry underlies the hepatitis B virus “a” determinant. J. Virol. 83:9321–9328. Mello FC, Martel N, Gomes SA, Araujo NM. 2011. Expression of hepatitis B virus surface antigen containing Y100C variant frequently detected in occult HBV infection. Hepat. Res. Treat. 2011:695859. doi:10.1155 /2011/695859. Martin CM, Welge JA, Rouster SD, Shata MT, Sherman KE, Blackard JT. 2012. Mutations associated with occult hepatitis B virus infection result in decreased surface antigen expression in vitro. J. Vir. Hepat. 19: 716 –723. Persson B, Argos P. 1994. Prediction of transmembrane segments in proteins utilising multiple sequence alignments. J. Mol. Biol. 237:182– 192. Torresi J. 2002. The virological and clinical significance of mutations in the overlapping envelope and polymerase genes of hepatitis B virus. J. Clin. Virol. 25:97–106. Chua PK, Wang RY, Lin MH, Masuda T, Suk FM, Shih C. 2005. Reduced secretion of virions and hepatitis B virus (HBV) surface antigen of a naturally occurring HBV variant correlates with the accumulation of the small S envelope protein in the endoplasmic reticulum and Golgi apparatus. J. Virol. 79:13483–13496. Khan N, Guarnieri M, Ahn SH, Jisu L, Zhou Y, Bang G, Kim KH, Wands JR, Tong S. 2004. Modulation of hepatitis B virus secretion by naturally occurring mutations in the S gene. J. Virol. 78:3262–3270. Loffler-Mary H, Dumortier J, Klentsch-Zimmer C, Prange R. 2000. Hepatitis B virus assembly is sensitive to changes in the cytosolic S loop of the envelope proteins. Virology 270:358 –367. Gutiérrez C, Devesa M, Loureiro CL, León G, Liprandi F, Pujol FH. 2004. Molecular and serological evaluation of surface antigen negative hepatitis B virus infection in blood donors from Venezuela. J. Med. Virol. 73:200 –207. Araujo NM, Vianna COA, Soares CC, Gomes SA. 2008. A unique amino acid substitution, L215Q, in the hepatitis B vírus small envelope protein of a genotype F isolate that inhibits secretion of hepatitis B vírus subviral particles. Intervirology 51:81– 86. Motta JS, Mello FC, Lago BV, Perez RM, Gomes SA, Figueiredo FF. 2010. Occult hepatitis B virus infection and lamivudine-resistant muta-

jvi.asm.org 7891

Biswas et al.

22. 23. 24. 25.

26.

tions in isolates from renal patients undergoing hemodialysis. J. Gastroenterol. Hepatol. 25:101–106. Jenna S, Sureau C. 1999. Mutations in the carboxy-terminal domain of the small hepatitis B virus envelope protein impair the assembly of hepatitis delta virus particles. J. Virol. 73:3351–3358. Siegler VD, Bruss V. 2013. Role of transmembrane domains of the hepatitis B virus small surface proteins in subviral particle biogenesis. J. Virol. 87:1491–1496. Patient R, Hourioux C, Roingeard P. 2009. Morphogenesis of hepatitis B virus and its subviral envelope particles. Cell Microbiol. 11:1561–1570. Bonifacino JS, Cosson P, Shah N, Klausner RD. 1991. Role of potentially charged transmembrane residues in targeting proteins for retention and degradation within the endoplasmic reticulum. EMBO J. 10:2783–2793. Kalinina T, Riu A, Fischer L, Will H, Sterneck M. 2001. A dominant hepatitis B virus population defective in virus secretion because of several

7892

jvi.asm.org

27.

28. 29.

30.

S-gene mutations from a patient with fulminant hepatitis. Hepatology 34:385–394. Ito K, Qin Y, Guarnieri M, Garcia T, Kwei K, Mizokami M, Zhang J, Li J, Wands JR, Tong S. 2010. Impairment of hepatitis B virus virion secretion by single-amino-acid substitutions in the small envelope protein and rescue by a novel glycosylation site. J. Virol. 84:12850 –12861. Walter P, Ron D. 2011. The unfolded protein response: from stress pathway to homeostatic regulation. Science 334:1081–1086. Hsieh YH, Su IJ, Wang HC, Chang WW, Lei HY, Lai MD, Chang WT, Huang W. 2004. Pre-S mutant surface antigens in chronic hepatitis B virus infection induce oxidative stress and DNA damage. Carcinogenesis 10:2023–2032. Chisari FV, Klopchin K, Moriyama T, Pasquinelli C, Dunsford HA, Sell S, Pinkert CA, Brinster RL, Palmiter RD. 1989. Molecular pathogenesis of hepatocellular carcinoma in hepatitis B virus transgenic mice. Cell 59: 1145–1156.

Journal of Virology