Published Ahead of Print 8 May 2013.

2013, 87(14):7882. DOI: 10.1128/JVI.00710-13. J. Virol.

Subhajit Biswas, Daniel Candotti and Jean-Pierre Allain

Infection
May Contribute to Occult Hepatitis B Virus
and In Vitro Protein Prevent Its Excretion
Specific Amino Acid Substitutions in the S
http://jvi.asm.org/content/87/14/7882
Updated information and services can be found at:
These include:
REFERENCES
http://jvi.asm.org/content/87/14/7882#ref-list-1
This article cites 30 articles, 8 of which can be accessed free at:
CONTENT ALERTS
more articles cite this article),
Receive: RSS Feeds, eTOCs, free email alerts (when new
http://journals.asm.org/site/misc/reprints.xhtml Information about commercial reprint orders:
http://journals.asm.org/site/subscriptions/ To subscribe to to another ASM Journal go to:

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m


o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Specic 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 Allain
a
Department of Haematology, University of Cambridge, Cambridge, United Kingdom
a
; National Health Service Blood & Transplant, Cambridge, United Kingdom
b
Occult hepatitis B virus (HBV) infection (OBI) is dened 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 specic amino acid substitutions in the S protein from OBI in HBsAg production and excretion was
examined in vitro. HBsAg was quantied in culture supernatants and cell extracts of HuH-7 cells transiently transfected
with plasmids containing the S gene of eight HBsAg

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 immunouorescence 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 ex-
pressed HBsAg according to different patterns. However, the P178R substitution present in all cloned sequences of two
OBI strains may contribute signicantly to the OBI phenotype.
O
ccult hepatitis B virus (HBV) infection/carriage (OBI) is
characterized by the presence of very lowlevels of HBVDNA
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 Bsurface antigen (anti-HBs), outside the preseroconver-
sion window period (1). OBI represents a particular form of per-
sistent 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 immu-
nodeciency 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 substitu-
tions in the S protein affecting HBsAg detection with commer-
cial immunoassays (5, 6), mutations in regulatory elements
negatively affecting virus replication (7, 8), and mutations af-
fecting 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 (46, 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 vitro-
transfected hepatocyte cell lines and infected mice that might con-
tribute to the OBI phenotype (5, 12, 13).
In the present study, the potential role of OBI-specic amino
acid substitutions in the S protein, within and outside the MHR,
on HBsAg production and excretion was examined in vitro.
MATERIALS AND METHODS
Cloning and sequencing of the HBV S protein coding region, HBsAg
expression plasmid, and site-directed mutagenesis. HBV DNA was iso-
lated from randomly selected plasma samples from 18 previously charac-
terized blood donors with OBI and eight HBsAg

blood donors as con-

trols (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 amplied, 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 amplied 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, respec-
tively). 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 HCMVIE
Received 12 March 2013 Accepted 30 April 2013
Published ahead of print 8 May 2013
Address correspondence to Daniel Candotti, dc241@cam.ac.uk, or Jean-Pierre
Allain, jpa1000@cam.ac.uk.
* 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
7882 jvi.asm.org Journal of Virology p. 78827892 July 2013 Volume 87 Number 14

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

gene promoter replaced the natural pre-S2/S promoter, with the tran-
scriptional initiation site being located 75 nucleotides upstream of the 5=
end of the S mRNAin order to limit variation in mRNAtranscription and
consequently examine S protein expression through HBsAg quantica-
tion. The HBV DNA insert and ligation sites of selected clones (1 to 19
clones/sample) were sequenced, and S amino acid sequences were de-
rived. Amino acid substitutions in S protein sequences of OBI and control
clones were identied by comparison with a consensus sequence of the
respective genotype obtained by aligning sequences from HBsAg

blood
donors (95 sequences per genotype).
Site-directed mutagenesis (SDM) of pcDNA3.1-HBV clones was per-
formed using the QuikChange site-directed mutagenesis kit (Agilent
Technologies, La Jolla, CA) according to the manufacturers instructions.
The correct introduction of a mutation was veried by sequencing.
Cells and transfection. pcDNA3.1-HBV plasmids were transfected
into HuH-7 cells (3 g plasmid/3 10
5
cells) using FuGENE HD trans-
fection reagent (Promega, Madison, WI) as described previously (10). For
each sample, at least two independent transfection experiments were per-
formed in triplicate.
Analysis of extracellular and intracellular HBsAg production. Ex-
tracellular (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 quantication.
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 1PBS, 1%Triton X-100, 10 units/ml
DNase I, and complete proteinase inhibitor/EDTA (Roche, Basel, Swit-
zerland). Cell lysates were centrifuged at 13,000 rpm for 15 min, and
supernatants were collected as cytosolic extracts. HBsAg in extracts was
quantied as described above.
The total amount of HBsAg in each transfected culture of 3 10
5
HuH-7 cells was calculated as the sumof HBsAg present in ICextract and
ECsupernatant. For each sample, the mean ICand ECHBsAg 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 efciency calculated
from the percentage of cells expressing -galactosidase activity.
Immunouorescence detection of intracellular HBsAg. HuH-7 cells
grown on glass coverslips and transfected with pcDNA3.1-HBV plasmids
were xed at 72 h posttransfectionusing 4%paraformaldehyde for 30 min
at room temperature. The xed cells were washed with 1PBS, 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 immunouorescence
microscopy (60oil immersion lens and 10objective) for detection of
intracellular HBsAg in cells at a magnication of 600.
Transmembrane structure predictionof HBsAg variants. The trans-
membrane 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 Re-
source 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 Tukeys multiple
comparison (post hoc) test. Differences between HBsAg productions (to-
tal or EC) for OBI strains before and after SDM repair were compared
using Students 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 signicant.
TABLE 1 Characteristics of HBsAg

control and OBI donors

a
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 Control Malaysia M 30 1.6 10
7
8.3 10
2
Pos Neg C
M86 Control Malaysia F 19 3.1 10
8
2.1 10
4
Pos Neg C
M88 Control Malaysia M 23 1.1 10
8
3.9 10
4
Pos Neg B
M90 Control Malaysia F 22 1.6 10
8
1.2 10
5
Pos Neg B
M92 Control Malaysia M 49 1.0 10
7
1.1 10
3
Pos Neg C
M95 Control Malaysia M 26 1.5 10
3
1.4 10
4
Pos Neg C
E3476 Control Egypt NA NA 6 3.3 10
5
Pos Neg D
E3789 Control Egypt NA NA 31 1.1 10
5
Pos Neg D
HK01556 OBI Hong Kong M 47 5 Neg Pos Neg B
HK3110 OBI Hong Kong F 49 Not detectable Neg Pos Neg B
HK3475 OBI Hong Kong M 51 37 Neg Pos 26 B
HK6794 OBI Hong Kong M 59 91 Neg Pos 16 B
HK6921 OBI Hong Kong M 50 Not detectable Neg Pos 19 B
HK8442 OBI Hong Kong M 53 46 Neg Pos 98 C
HK8663 OBI Hong Kong M 53 56 Neg Pos Neg B
R84 OBI Italy M 44 62 Neg Pos Neg D
TW0498 OBI Taiwan M 46 48 Neg Pos Neg B
TW2256 OBI Taiwan M 30 56 Neg Pos 18 B
TW3437 OBI Taiwan F 49 30 Neg Pos Neg B
TW4576 OBI Taiwan M 55 5 Neg Pos Neg B
TW5004 OBI Taiwan F 63 7 Neg Pos Neg B
TW6083 OBI Taiwan M 62 5 Neg Pos Neg C
TW6639 OBI Taiwan M 55 Not detectable Neg Pos Neg B
TW8964 OBI Taiwan M 30 15 Neg Pos Neg B
TW9015 OBI Taiwan M 47 Not detectable Neg Pos Neg B
TW9331 OBI Taiwan M 54 6 Neg Pos Neg B
a
NA, not available.
Impaired S Protein Excretion and Occult HBV Infection
July 2013 Volume 87 Number 14 jvi.asm.org 7883

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

RESULTS
In vitro HBsAg production patterns. Eighteen HBV strains car-
rying 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 concentra-

tion, 10
4
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 eval-
uating the total amount of detectable ICandECHBsAg, the IC/EC
HBsAg ratio, andthe ICdistributionpatternof HBsAg detectedby
immunouorescence assay (IFA).
In HBsAg

controls, the average normalized total HBsAg pro-

duction was 120 ng/3 10
5
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
immunouorescence staining was diffuse and nely granular
across the hepatocyte cytoplasm (Fig. 3). Two genotype C control
clones (M92-cl2 and M95-cl8) showed IC/ECratios of 10.0 (Fig.
2A), andICHBsAg concentratedinthe perinuclear area as densely
packed uorescence (data not shown).
Three distinct patterns of HBsAg production were identied 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 con-
trols, with total HBsAg production of 120 ng/3 10
5
cells
(range, 146 to 621 ng) (Fig. 2B), IC/ECHBsAg ratio of 3 (range,
0.1 and3) (Fig. 2C), anddiffuse granular uorescent 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/ECHBsAg ratio ranging between
noncalculable and 7, with low or no uorescence staining of IC
HBsAg (Fig. 3). For TW6639-cl1 (HBsAg, 46 ng/3 10
5
cells;
IC/EC, 7) and TW2256-cl3 (HBsAg, 31 ng/3 10
5
cells; IC/EC,
2), the ICHBsAg appeared as sparsely distributed but bright green
uorescent 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 uorescence appeared dense, compact, and essentially
limited to the central part of the cytoplasm, adjacent to the nu-
cleus, 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 10
5
cells; IC/EC, 12) presented as diffuse but very compact
uorescence lling the entire cytosol (not shown). No signicant
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
withHBsAg impaired excretioninHuH-7 cells. In pattern 2 OBI
clones, HBsAg production was perfectly detectable and reached
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.
Biswas et al.
7884 jvi.asm.org Journal of Virology

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

FIG2 Detection of intracellular and extracellular levels of HBsAg/S protein by EIAand IFAafter 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 signicant (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 signicance 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/ECratios of 800, 75, and 23, respectively). (D) Relationship between IC/ECratio 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.
July 2013 Volume 87 Number 14 jvi.asm.org 7885

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

signicantly 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 ex-
cretion 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 con-
sensus genotype B sequence obtained from 124 HBsAg

se-
quences (Fig. 1). Substitutions at positions 3 to 5 were not consid-
ered, as they resulted fromthe use of the degenerated primer SPL3
in the initial genomic amplication procedure. The potential as-
sociation of 17 OBI-specic substitutions with pattern 2 was in-
vestigated 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 pres-
ent 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 modied when 14 se-
lected 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 signicant increase of HBsAg produc-
tion (P 0.02 to 0.05, except HK01556-cl2) that was associated
with increased HBsAg excretion in culture supernatants, as re-
ected in the decrease of HBsAg IC/EC ratio observed (7 to 801
versus 0.125 to 4 after SDM). This nding was conrmed in all
four samples by immunouorescence analysis, showing a change
following SDM from the dense intracellular accumulation of
HBsAg associated with pattern 2 to a diffuse uorescence across
the cytoplasm typical of pattern 1 (Fig. 4A). These data were re-
produced in the hepatocyte cell line HepG2 (data not shown). The
P178Rsubstitution was present in the pattern 3 TW8964-cl1 sam-
ple, but the SDM restoration of proline did not produce any per-
ceptible 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 conrm the effect of M75T, Y100S, and P178R on HBsAg
excretion, these three substitutions were introduced into the ge-
FIG 3 Immunouorescence 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 LacZwere used as
negative controls.
TABLE 2 Impact of OBI-specic amino acid substitutions on HBsAg
production pattern in vitro
OBI clone
Substitution
repaired
a
Average total
HBsAg
(range) (ng)
b
IC/EC
ratio
IFA
pattern
c
Deduced
HBsAg
Production
pattern
HK01556-cl2 None 189 (93254) 800 R 2
T75 M 226 (127310) 0.125 WT 1
R105P 72 (47103) 41 R 2
P154S 214 (213229) 464 R 2
A184V 97 (83109) 73 R 2
HK6794-cl2 None 60 (5192) 8 R 2
S100Y 325 (307362) 3 WT 1
R165WL167S 102 (96107) 7 R 2
C182W 39 (2470) 9 R 2
TW9015-cl1 None 91 (56143) 12 R 2
S111PE112G 91 (8299) 13 R 2
E119G 93 (8999) 11 R 2
TW0498-cl2 None 18 (1323) 13 LF 3
R129Q 21 (1923) 22 LF 3
TW0498-cl3 None 135 (62217) 15 R 2
R129Q 63 14 R 2
HK3110-cl4 None 116 (88131) 20 R 2
V159AN160K 188 (184192) 20 R 2
P176SR178P 284 (267302) 2 WT 1
R178P 757 (692809) 4 WT 1
N226I 23 (1825) 22 R 2
HK3475-cl6 None 75 (49103) 7 R 2
R178P 464 (409541) 2 WT 1
TW8964-cl1 None 0.6 (0.30.8) NA
d
NF 3
R178P 2 (1.82.1) NA NF 3
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 uorescence close to the nucleus; LF, low level uorescence; NF, no
uorescence detected.
d
NA, not applicable.
Biswas et al.
7886 jvi.asm.org Journal of Virology

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

notype BHBsAg

control M88-cl4. OBI pattern2 was reproduced

in M88-cl4 following the M75T and P178R changes but not the
Y100S change, as evidenced by IC/EC ratio evaluation and intra-
cellular immunouorescence 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 P111SG112E,
G119E, Q129R, I150T, S154P, and W165R introduced into the
control M88-cl4 did not substantially affect the HBsAg produc-
tion pattern (Table 3).
The effect of amino acid changes at position 178 on M88-cl4
HBsAg phenotype was investigated further (Table 3). Substitu-
tions P178A, P178E, P178K, P178L, and P178Q resulted in a sub-
stantially decreased overall average productionof 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
FIG4 Immunouorescence staining of native and mutated HBsAg in transfected HuH-7 cells. Amino acid substitutions introduced by site-directed mutagen-
esis in OBI (A) and non-OBI (B) sequences are indicated in bold.
TABLE 3 Impact of OBI-specic substitutions introduced in the HBsAg

control M88-cl4 on HBsAg production in vitro

Substitutions introduced
by SDM
Avg total HBsAg (range) (ng)
HBsAg IC/EC
ratio IFA pattern
c
Deduced HBsAg
production
pattern Total
b
EC
None
a
270 (203338) 197 (1672270) 0.33 WT 1
M75T 48 (3956) ND
d
NA
e
R 2
Y100S 475 (456487) 316 (308321) 0.5 WT 1
P111SG112E 482 (456500) 347 (334354) 0.33 WT 1
G119E 402 (359433) 241 (212256) 0.5 WT 1
Q129R 439 (410471) 278 (256302) 0.5 WT 1
I150T 102 (74112) 71 (5280) 0.5 WT 1
S154P 337 (296392) 214 (191243) 0.5 WT 1
W165R 138 (112161) 70 (4984) 1 WT 1
P178R 78 (7681) 5 (56) 15 R 2
P178Q 63 (4576) 11 (1112) 5 WT Indeterminate
P178L 80 (7082) 19 (1821) 3 WT Indeterminate
P178A 31 (2737) 11 (1013) 2 WT Indeterminate
P178K 41 (3148) 12 (1015) 3 WT Indeterminate
P178E 136 (126143) 99 (92104) 0.33 WT 1
a
Wild type M88-cl4 control.
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 uorescence close to the nucleus; NF, no uorescence detected.
d
ND, not detected.
e
NA, not applicable.
Impaired S Protein Excretion and Occult HBV Infection
July 2013 Volume 87 Number 14 jvi.asm.org 7887

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

(IC/EC 0.33) but still compatible with pattern 1 denition and
lower than the pattern 2-related IC/EC ratio of the P178R mutant
(IC/EC 15). The immunouorescence (IF) results obtained
with P178A/K/L/Q did not clearly match either pattern 1 or pat-
tern2 proles andwere consideredindeterminate (Fig. 5). Mutant
P178E showed a pattern 1-related IC/EC HBsAg ratio and IFA
prole 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 arrange-
ments of the S protein in the bilayer lipid membrane of the endo-
plasmic reticulum (ER). Persson and Argos model as well as
SOSUIs and DASs identied the same four transmembrane
(TM) domains for wild-type S. However, SOSUI predicted a fth
TM domain (amino acids 149 to 171) that we considered doubt-
ful. Nevertheless, the three amino acid substitutions identied as
critical to HBsAg excretion had identical locations predicted by
the three models. M75T was located in the cytosolic loop separat-
ing TM1 and TM2, Y100S was at the end of TM2, and P178R was
in TM3.
HBsAg excretiondecit as potential contributor to OBI phe-
notype. Amino acid substitutions impairing HBsAg excretion re-
ected by HBsAg quantication in vitro have been identied in
individual OBI clones. However, HBVs in HBsAg-positive as well
as OBI strains circulate as quasispecies of related variants not nec-
essarily represented by a single clone tested in vitro. The genetic
diversity and the relevance of specic mutations as potential ex-
planation for the OBI phenotype were therefore examined by se-
quencing 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 signicant relationship with the HBsAg phe-
notype (data not shown). The Y100S mutations responsible for
pattern 2 of clone 2 were present in a subcluster of ve clones
associated with clone 2, but Y100F was present in all 12 other
clones. In contrast, in OBIs HK01556, HK3110, and HK3475, ge-
netic 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
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 gen-
erally 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 specic 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
identied three patterns of HBsAg expression (Fig. 1). These pat-
terns were dened by the total HBsAg production estimated by
enzyme immunoassay (EIA), the IC/EC HBsAg ratio, and the in-
tracellular distribution of HBsAg detected by immunouores-
cence (Fig. 2 and 3). However, two genotype C controls (M92-cl2
and M95-cl8) presented an IC/ECratio and IFApattern similar to
OBI pattern 2, but total HBsAg was similar to pattern 1. In addi-
tion, 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 expres-
sion 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 dene the three patterns, might be
due to variations in the effectiveness of the HBsAg expression
vector transfection into HuH-7 cells despite normalization us-
ing 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 phe-
notypes 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
FIG 5 Immunouorescence staining of HBsAg in cells transfected with wild-
type and mutated M88-cl4 sequences. The position and nature of the OBI-
specic amino acid substitutions introduced into the genotype Bcontrol M88-
cl4 are indicated.
Biswas et al.
7888 jvi.asm.org Journal of Virology

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

when HuH-7 cells with an entire HBVgenotype Cgenome were
transfected (data not shown). Nevertheless, total HBsAg, to-
gether with the intracellular HBsAg immunouorescence pat-
tern, 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 immunouorescence
(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 posttran-
scriptional 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 signicantly
higher in OBI than in non-OBI sequences (0.9% and 1.9% in non-
OBI MHR and S protein, respectively) (Fig. 1). However, mutations
D144EandG145R, previously associatedwithreducedHBsAg detec-
tion (15), were present in three pattern 3 clones (TW6639-cl1 and
TW3437-cl5) or in association (HK8663-cl2) and supported the im-
mune detection escape hypothesis.
FIG6 Phylogenetic analysis of the S amino acid sequences of multiple clones fromOBI and non-OBI strains. Phylogenetic analysis was performed as previously
described(10). HBVreference sequences of genotypes/subgenotypes A1-3 andB-Hare identiedby their GenBank accessionnumbers. For clones usedinHuH-7
transfection experiments, the resulting HBsAg excretion pattern is indicated by open (pattern 1), black (pattern 2) or gray (pattern 3) arrows.
Impaired S Protein Excretion and Occult HBV Infection
July 2013 Volume 87 Number 14 jvi.asm.org 7889

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

In individuals carrying pattern 1 and 2 clones, OBI did not
seem related to antigenic variation in HBsAg, since when pro-
duced 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 modied reverse transcriptase region of the HBVpolymerase,
decreasing HBV replication (15). The reduced level of HBsAg ex-
cretion 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-specic
mutation(s) (Table 2 and Fig. 4A) and introducing this muta-
tion(s) into a control/wild-type sequence (Table 3 and Fig. 4B)
identied three mutationsM75T, Y100S, and P178Rthat
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 specic.
The rare M75T substitution was associated with HBsAg excre-
tion defect. This mutation is located in the C-terminus part of the
protein cytosolic loop that contains a putative core-envelope in-
teraction 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 perinu-
clear IFA staining pattern observed for mutant M75T in the pres-
ent 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 termi-
nus in HBsAg secretion independently of its putative role in virion
morphogenesis.
Three genotype B OBI clonesHK01556-cl2, HK6794-cl2,
and TW6639-cl1with HBsAg excretion defect contained the
Y100S mutation(Fig. 1 andFig. 2). WhenrepairedinHK6794-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 re-
tention phenotype of L77R (16). A Y100C substitution was pres-
ent 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, 1921). However, it
was also found in all pattern 2/3 clones of the M92 control, indi-
cating 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 se-
cretion were affected by mutations within three of the putative
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 do-
mains withinthe S protein, resulting inalteredproteinfolding 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 HK3110-
cl4 and HK3475-cl6 OBI clones and in the mutated M88-cl4 con-
trol. Introduction of a positively charged arginine residue in place
of a proline (a residue commonly found as the rst residue of an
alpha helix) may modify this transmembrane domain and affect
the excretion of the modied protein. The presence of a poten-
tially 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 characteriz-
ing OBI pattern 2. These effects of charged residues appeared to
correlate with the level of free energy required to partitioncharged
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, lim-
ited changes induced by nonpolar residues alanine and leucine
and the lack of signicant 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 showthat, 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 contin-
uously rened.
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 re-
mained 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 secre-
tion defect. Similarly, Q129R was found in four OBI clones irre-
spective 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 efcient HBsAg excretion in TW0498-cl3 (Table
2), and the M88-cl4 excretion pattern was not modied 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 alterationof HBsAg
secretioncanbe due not only to a single substitutionbut also more
frequently to a combination of amino acid substitutions in differ-
ent regions of the protein. This is supported by reports showing
both positive and negative transcomplementation effect of S mu-
tations on HBsAg secretion inhibition (13, 16, 17).
It might be difcult to evaluate the exact importance of muta-
tions altering HBsAg secretion in the genesis of OBI, since such
Biswas et al.
7890 jvi.asm.org Journal of Virology

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

mutations appear to be rare and strain specic 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 in-
fected 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 quasi-
species strongly supports anassociationbetweenHBsAg excretion
defect and OBI phenotype in these donors. A more complex situ-
ation was observed in HK6794 diverse quasispecies reected 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 re-
tained 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
transductionpathways (28). Inhumanhepatocytes, 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 in-
ammation, regenerative hyperplasia, transcription deregulation,
aneuploidy, and the eventual development of hepatocellular car-
cinoma (HCC) (30). Altogether, these studies suggest that cytoso-
lic 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
HBVvariants might be at high risk of developing HCCif sufcient
time is allowed to elapse.
Inconclusion, thehighgeneticvariabilityobservedintheSgeneof
OBI strains is associated with heterogeneous patterns of intracellular
and extracellular HBsAg production in vitro. The identication of
three new OBI-specic mutationsM75T, Y100S, and P178R
associated with HBsAg intracellular retention supports further the
hypothesis that the lack of HBsAg secretion may signicantly con-
tribute to the multifactorial occurrence of OBI. M75T and P178R
mutants provide additional indirect support for a role of the cy-
tosolic loop and the third transmembrane domain in HBsAg mor-
phogenesis. Further studies are needed to investigate a potential
pathogenic role of intracellularly retained HBsAg in the develop-
ment of liver disease in OBI carriers.
ACKNOWLEDGMENTS
S. Linfromthe TaiwanBloodServices Foundation, Taipei, Taiwan, andC.
Kit Lin from Hong Kong Red Cross Blood Transfusion Centre, Hong
Kong, Peoples 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, Co-
lombo 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, Trpo C, Villa E, Will H, Zanetti AR, Zoulim F. 2008.
Statements from the Taormina expert meeting on occult hepatitis B virus
infection. J. Hepatol. 49:652657.
2. Raimondo G, Pollicino T, Romano L, Zanetti AR. 2010. A 2010 update
on occult hepatitis B infection. Pathol. Biol. 58:254257.
3. 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 carci-
noma. Hepatology 45:277285.
4. 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.
5. 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. Inuence
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.
6. 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 52:16001610.
7. 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 127:13561371.
8. Fang Y, Teng X, Xu WZ, Li D, Zhao HW, Fu LJ, Zhang FM, Gu HX.
2009. Molecular characterization and functional analysis of occult hepa-
titis Bvirus infection in Chinese patients infected with genotype C. J. Med.
Virol. 81:826835.
9. Hass M, Hannoun C, Kalinina T, Sommer G, Manegold C, Gnther S.
2005. Functional analysis of hepatitis B virus reactivating in hepatitis B
surface antigen-negative individuals. Hepatology 42:93103.
10. 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 po-
tential mechanisms of occurrence. Gut 61:17441753.
11. Salisse J, Sureau C. 2009. A function essential to viral entry underlies the
hepatitis B virus a determinant. J. Virol. 83:93219328.
12. Mello FC, Martel N, Gomes SA, Araujo NM. 2011. Expression of hep-
atitis Bvirus surface antigencontaining Y100Cvariant frequently detected
in occult HBV infection. Hepat. Res. Treat. 2011:695859. doi:10.1155
/2011/695859.
13. 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:
716723.
14. Persson B, Argos P. 1994. Prediction of transmembrane segments in
proteins utilising multiple sequence alignments. J. Mol. Biol. 237:182
192.
15. Torresi J. 2002. The virological and clinical signicance of mutations in
the overlapping envelope and polymerase genes of hepatitis B virus. J.
Clin. Virol. 25:97106.
16. 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. 79:1348313496.
17. 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. 78:32623270.
18. Lofer-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 270:358367.
19. Gutirrez C, Devesa M, Loureiro CL, Len 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.
73:200207.
20. Araujo NM, Vianna COA, Soares CC, Gomes SA. 2008. A unique amino
acid substitution, L215Q, in the hepatitis Bvrus small envelope protein of
a genotype F isolate that inhibits secretion of hepatitis B vrus subviral
particles. Intervirology 51:8186.
21. Motta JS, Mello FC, Lago BV, Perez RM, Gomes SA, Figueiredo FF.
2010. Occult hepatitis B virus infection and lamivudine-resistant muta-
Impaired S Protein Excretion and Occult HBV Infection
July 2013 Volume 87 Number 14 jvi.asm.org 7891

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

tions in isolates from renal patients undergoing hemodialysis. J. Gastro-
enterol. Hepatol. 25:101106.
22. Jenna S, Sureau C. 1999. Mutations in the carboxy-terminal domain of
the small hepatitis B virus envelope protein impair the assembly of hepa-
titis delta virus particles. J. Virol. 73:33513358.
23. Siegler VD, Bruss V. 2013. Role of transmembrane domains of the hep-
atitis B virus small surface proteins in subviral particle biogenesis. J. Virol.
87:14911496.
24. Patient R, Hourioux C, Roingeard P. 2009. Morphogenesis of hepatitis B
virus and its subviral envelope particles. Cell Microbiol. 11:15611570.
25. Bonifacino JS, Cosson P, Shah N, Klausner RD. 1991. Role of po-
tentially charged transmembrane residues in targeting proteins for re-
tention and degradation within the endoplasmic reticulum. EMBO J.
10:27832793.
26. Kalinina T, Riu A, Fischer L, Will H, Sterneck M. 2001. A dominant
hepatitis B virus population defective in virus secretion because of several
S-gene mutations from a patient with fulminant hepatitis. Hepatology
34:385394.
27. 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 secre-
tion by single-amino-acid substitutions in the small envelope protein and
rescue by a novel glycosylation site. J. Virol. 84:1285012861.
28. Walter P, Ron D. 2011. The unfolded protein response: from stress
pathway to homeostatic regulation. Science 334:10811086.
29. 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
10:20232032.
30. 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:
11451156.
Biswas et al.
7892 jvi.asm.org Journal of Virology

o
n

D
e
c
e
m
b
e
r

1
1
,

2
0
1
3

b
y

N
A
T

I
N
S
T

O
F

H
E
A
L
T
H

L
I
B
h
t
t
p
:
/
/
j
v
i
.
a
s
m
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m