259

Identification of Hepatitis B Surface Antigen Variants with Alterations outside

the “a” Determinant in Immunized Singapore Infants
Oon Chong-Jin, Chen Wei Ning, Koh Shiuan, Ransome Research Laboratory, Department of Clinical Research,
and Lim Gek Keow Ministry of Health, Singapore General Hospital,
Republic of Singapore

Vaccine-associated hepatitis B surface antigen (HBsAg) mutations have been mainly iden-
tified within the “a” determinant (aa 124–147). Now changes have been detected that are

Downloaded from https://academic.oup.com/jid/article-abstract/179/1/259/881393 by guest on 28 December 2018

located outside the “a” determinant from immunized Singapore infants born to hepatitis B
virus (HBV) carrier mothers. These include Asn116 to Thr116, Val118 to Ala118, Pro120 to
Ser120, Ala159 to Val159, Phe183 to Cys183, and Val184 to Ala184. Decreased binding to
“a” determinant–specific monoclonal antibody was observed for variants Pro120 to Ser120,
Ala159 to Val159, and Phe183 to Cys183. Unlike other HBsAg variants, the Asn116-to-Thr116
HBsAg variant had the wild type threonine in the infant, whereas the mutated asparagine
was found in the mother. Vertical transmission is indicated for Phe183-to-Cys183 and Val184-
to-Ala184 HBsAg variants, as they were found in both the infants and mothers. Detailed
analysis of these HBsAg variants would provide further understanding of the antigenic struc-
ture of HBV.

Active vaccination has remained the most effective way to
control the leading cause of hepatitis, human hepatitis B virus
(HBV). The highly antigenic hepatitis B surface antigen
(HBsAg) is directly involved in inducing the humoral immune
response, which in turn provides protective immunity against
HBV infection. Neutralizing B cell epitopes are believed to be
located from position 124 to 147 of HBsAg, defined as the “a”
determinant [1].
In recent years, however, HBV variants with mutations on
the “a” determinant have been identified following vaccination.
Such mutants are capable of independent replication and lead
to active infection [2–4]. Mutations have been detected at var-
ious positions on the “a” determinant, the most frequently iden-
tified being the glycine-to-arginine change at position 145 of
HBsAg [4–10].
Importantly, some of the mutants existed before the vacci-
nation program was introduced. These include changes at po-
sitions 126, 129, 133, and 145 [8, 10]. Some of the naturally
occurring HBsAg variants are transmissible and able to infect
the vaccinated population [8, 10]. Furthermore, HBsAg vari-
ants have recently been identified on the major hydrophilic loop
of HBsAg (aa 100–160) but outside the conventional “a” de-
Received 15 April 1998; revised 10 August 1998.
Financial support: Singapore National Medical Research Council and
Ministry of Health Singapore.
Consent was obtained from the parents of infants, and human experi-
mentation guidelines were followed in the conduct of the present clinical
research.
Reprints or correspondence: Dr. Oon Chong-Jin, Dept. of Clinical Re-
search, Ministry of Health, Singapore General Hospital, Outram Road,
Republic of Singapore (gcrcwn@sgh.gov.sg).
The Journal of Infectious Diseases 1999; 179:259–63
q 1999 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/99/7901-0037$02.00
terminant [11]. These new mutants occur in both the vaccinated
and unvaccinated population and are not neutralized by the
presently available antibodies.
We have previously reported 41 HBV DNA–positive im-
munized Singapore infants born to HBsAg- and hepatitis B e
antigen (HBeAg)–positive mothers [7]. These infants have been
considered as breakthrough HBV infection carriers, as their
serum HBV DNA has been tested positive by liquid hybridi-
zation assays (Genostics assays; Abbott Laboratories, Abbott
Park, IL) and polymerase chain reaction (PCR). In addition,
antibody to hepatitis B core (anti-HBc) has persistently been
detected. Sequence analysis of their HBsAg indicated amino
acid changes in the “a” determinant in 16 cases, with a pre-
dominant glycine-to-arginine change (12 infants) at position
145.
The aim of this study was to identify mutations outside the
“a” determinant in the remaining 25 breakthrough cases.
Materials and Methods
Subjects and samples. The HBV vaccination program had been
conducted from 1984 to 1996 on 345 infants born to HBV carrier
mothers (seropositive for both HBsAg and HBeAg) [7]. Forty-one
infants became carriers, and 16 of them have previously been re-
ported [7]. Serial blood samples were obtained from the remaining
25 infants at 24 h after birth; at 1, 2, 3, and 6 months; and at 1,
2, 3, 4, 6, 8, 10, 12, and 14 years. Blood samples were also collected
from the respective mothers shortly before delivery.
Serologic testing for HBV markers. HBsAg, antibody to
HBsAg (anti-HBs), HBeAg, antibody to HBeAg, and HBV DNA
were determined by use of commercial products (Abbott Labo-
ratories). Such serologic testing was done over a period of up to
14 years, from 1984 to 1998.
Amplification by PCR and sequence analysis. DNA of 100 mL
260 Oon et al. JID 1999;179 (January)

Table 1. Serologic and molecular analysis of 5 carrier infants with Table 1. (Continued).
HBV mutations outside “a” epitope and 1 infant with wild type whose
mother had mutation.
Infant no., race, Anti- Anti- HBV Infant no., race, Anti- Anti- HBV
subtype, age HBsAg HBs HBc HBeAg DNA Mutation subtype, age HBsAg HBs HBc HBeAg DNA Mutation
1, Chinese, adw 5, Chinese, adw
1 year 1 2 ND ND ND Wild 1 120 14 years 1 2 1 1 ND 159
2 years 1 2 ND ND ND ND Mother 1 ND 1 2 1.7 ND
3 years 1 2 ND ND ND ND 6, Chinese, adr
4 years 1 2 1 1 ND ND 24 h 1 2 ND ND ND ND

Downloaded from https://academic.oup.com/jid/article-abstract/179/1/259/881393 by guest on 28 December 2018

6 years 1 2 1 1 ND Wild 1 120 4 weeks 2 76.9 ND ND ND ND
8 years 1 2 1 1 ND ND 8 weeks 2 27.0 ND ND ND ND
10 years 1 2 1 ND ND ND 1 year 1 2 1 ND ND ND
12 years 1 2 1 ND ND ND 2 years 1 8.0 1 ND ND Wild type
14 years 1 2 1 ND ND Wild 1 120 4 years 1 2 1 1 ND Wild type
Mother 1 ND ND 1 ND Wild 6 years 1 2 1 1 ND Wild type
2, Chinese, adw 8 years 1 2 1 1 ND ND
24 h 1 2 ND ND ND ND 10 years 1 2 1 ND ND ND
4 weeks 2 36.5 ND ND ND Wild type 12 years 1 2 1 ND ND ND
8 weeks 1 2 ND ND ND Wild 1 120 Mother 1 ND ND 1 ND 116
12 weeks 1 2 1 ND ND Wild 1 120
NOTE. All infants were male. ND, not determined. HBsAg, hepatitis B sur-
6 months 1 2 1 ND ND Wild 1 120
face antigen; anti-HBs, antibody to HBsAg; anti-HBc, antibody to hepatitis B
1 year 1 4.0 1 ND ND Wild 1 120
core antigen; HBeAg, hepatitis B e antigen.
2 years 1 2 1 ND ND Wild 1 120
4 years 1 2 1 ND ND Wild 1 120
6 years 1 2 1 1 ND ND of serum was extracted as described [7]. Oligonucleotides corre-
8 years 1 2 1 1 ND ND
sponding to HBV genomic regions of 418–433 (50 primer) and
10 years 1 1.0 1 ND ND ND
12 years 1 1.0 1 ND ND ND 842–822 (30 primer) were used in PCR amplification, and direct
14 years 1 1.0 1 ND ND Wild 1 120 sequencing of the PCR products was done as described [7]. The
Mother 1 ND ND 1 200.2 Wild type antigenicity profile of amino acid sequences was analyzed by the
3, Malay, adw Hopp/Woods method included in the Omiga system (Molecular
24 h 1 2 ND ND ND ND
4 weeks 2 129.0 ND ND ND ND
Group, Cambridge, MA).
8 weeks 2 57.7 ND ND ND ND Binding of HBsAg variants to anti–“a” determinant monoclonal
12 weeks 2 17.2 ND ND ND ND antibody. The binding of HBsAg variants to the “a” determi-
6 months 1 2 1 ND ND ND nant–specific monoclonal antibody 3E7 (Dako, Copenhagen) was
1 year 1 2 1 ND ND 183
assessed in liquid phase. An equivalent nanogram quantity of each
2 years 1 2 1 ND ND 183
4 years 1 2 1 1 ND ND HBsAg (including wild type, Pro120-to-Ser120, Ala159-to-Val159,
6 years 1 2 1 1 268.0 183 and Phe183-to-Cys183) was incubated with different concentra-
8 years 1 2 1 1 ND ND tions of 3E7 antibody for 1 h at 377C. The unbound HBsAg was
10 years 1 2 1 ND ND ND quantified by AUSRIA (Abbott Laboratories) with polyclonal
12 years 1 2 1 ND ND ND
14 years 1 2 1 ND ND 183
antibody–coated beads.
Mother 1 ND ND 1 202.5 183
4, Chinese, adr
24 h 1 2 1 ND ND ND Results
4 weeks 2 67.0 1 ND ND ND
8 weeks 2 34.0 1 ND ND ND Sixteen of the vaccinated infants with breakthrough HBV
12 weeks 2 10.0 1 ND ND ND infections carry HBsAg mutants with altered “a” determinant
6 months 1 2 1 ND ND ND
1 year 1 2 1 ND ND ND
[7]. More extensive sequence analysis was done on the remain-
2 years 1 2 1 ND ND Wild 1 118 1 184 ing breakthrough cases (25 infants) and revealed the presence
8 years 1 2 1 1 ND Wild 1 118 1 184 of new mutations outside the “a” determinant in 5 (table 1).
10 years 1 2 1 1 ND ND
12 years 1 2 1 ND ND ND
Two had the Pro120-to-Ser120 change in HBsAg (infants 1 and
Mother 1 ND ND 1 95.9 184 2). Sera of both infants were quasispecies, as wild type HBV
5, Chinese, adw was also detected. Another infant had the Phe183-to-Cys183
24 h 1 2 ND ND ND ND
4 weeks 1 2 ND ND ND 159
change (infant 3). No wild type virus was detected in this infant,
8 weeks 1 2 ND ND ND ND and wild type virus also was not found in the serum of infant
12 weeks 1 2 ND ND ND ND 5, which carried an HBsAg variant Ala159 to Val159. In the
6 months 1 2 1 ND ND ND
1 year 1 2.2 1 ND ND ND
case of infant 4, a mixture of wild type and HBsAg variants
2 years 1 2 1 ND ND ND with Thr118-to-Lys118 and Val184-to-Ala184 mutations was
4 years 1 2.3 1 1 ND ND identified in the serum. Whether the two amino acid substi-
6 years 1 2 1 1 ND ND
8 years 1 2 1 1 ND 159
tutions had occurred in the same virus has yet to be determined.
9 years 1 2 1 1 72.8 ND Mutations Pro120 to Ser120, Ala159 to Val159, and Phe183 to
JID 1999;179 (January) HBsAg Variants in Singapore Infants
Cys183 appeared to be stable in the respective infants, as they
were still detected 14 years after birth (table 1).
Serum samples from mothers (except for that of infant 5 with
undetectable PCR product) were then analyzed (table 1). Wild
type HBV was detected in mothers 1 and 2, suggesting that
the observed Pro120-to-Ser120 mutation had emerged follow-
ing vaccination. Conversely, Phe183-to-Cys183 and Val184-to-
Ala184 HBsAg variants that were detected in infants 3 and 4
were preexisting in their respective mothers (table 1), pointing
to their vertical transmission. The variant Val118 to Ala118
was detected only in infant 4, suggesting that it had emerged
following vaccination. Serum analysis also led to the detection
of the Asn116-to-Thr116 HBsAg variant in infant 6. In this
case, the mother carried the mutated residue (asparagine),
whereas the wild type HBsAg was found in the infant (table
1).
Computer analysis of the changes in antigenicity due to these
newly identified mutations was limited to those located on the
major hydrophilic loop (aa 118, 120, and 159). Results shown
in figure 1 predicted alterations in the antigenicity profile to
various extents compared with wild type HBsAg.
To assess the significance of the alterations outside the “a”
determinant, binding of these HBsAg variants to the “a” de-
terminant–specific monoclonal antibody 3E7 (Dako) was done.
At the antibody concentration that achieved 100% neutrali-
zation of the wild type HBsAg, 89%, 85.4%, and 85.3% neu-
tralization by this antibody was observed for the same nano-
gram quantity of variants Pro120 to Ser120, Ala159 to Val159,
and Phe183 to Cys183, respectively. Our results, therefore, sug-
gested a decreased binding of these three HBsAg variants to
the “a” determinant–specific monoclonal antibody, leading
possibly to their escape from HBV vaccines. The serum of infant
4, carrying the mixture of wild type HBsAg and variants Thr118
to Lys118 and Val184 to Ala184, was insufficient for this
analysis.
Discussion
We have previously reported that 41 of 345 infants in Sin-
gapore born to HBV carrier mothers had breakthrough HBV
infection despite receiving hepatitis B immune globulin at birth
and the full course of a plasma-derived HBV vaccine [7]. Sixteen
of these infected infants had mutations on the “a” determinant
of HBsAg. In the present study, a more extensive analysis of
the other 25 infants, mutations outside the antigenic “a” de-
terminant were detected in 6 infants, while wild type HBsAg
was found in the remaining 19.
Thus far, vaccine-escape HBsAg mutants have been char-
acterized by the coexistence of serum HBsAg and anti-HBs [7,
9, 10]. In the present study, the immune pressure by anti-HBs
has resulted in selection of variants such as Val118 to Ala118
(infant 4) and Pro120 to Ser120 (infants 1 and 2). Similar co-
existence was also observed in infant 6; however, the mutation
261
Asn116 to Thr116 occurred in an unusual way in that the mu-
tated residue in the mother was reversed to wild type in the
child. Such reversion from mutated to wild type HBsAg fol-
lowing vaccination has not been reported, and its detection is
intriguing, considering the fact that hepatitis B immune glob-
ulin is believed to neutralize wild type HBsAg.
Some of the HBsAg variants in our study were negative for
anti-HBs (infants 1 and 5). In infant 1, the absence of anti-
HBs could be because serum analysis began at a late date (1
Downloaded from https://academic.oup.com/jid/article-abstract/179/1/259/881393 by guest on 28 December 2018
year of age), as anti-HBs was usually detected within the first
12 weeks after birth in other infants (table 1). The detection of
the HBsAg variant in infant 5 at 4 weeks, despite the low level
of anti-HBs at 1 year of age, may be due to the excessive amount
of wild type HBsAg from HBV infection that totally absorbed
the anti-HBs.
Mutations outside the conventional “a” determinant are rel-
atively uncommon. Their identification in our report could be
explained by the fact that a significant proportion of the anti-
HBs that developed following HBV vaccination recognized the
“a” determinant. Their importance may be strengthened by
recent detection of such mutants in vaccinated as well as un-
vaccinated persons [3, 12]. One such variant has a Trp156-to-
Leu156 substitution and failed to bind to anti-HBs [12]. The
mutation Pro120 to Ser120 detected in our study has been
shown to affect the binding of antibodies to the second loop
of the “a” determinant [13]. Consistently, a decreased binding
to the “a” determinant–specific monoclonal antibody was also
demonstrated in our study for HBsAg variant Pro120 to Ser120,
as well as for Ala159 to Val159 and Phe183 to Cys183.
aa 159 is important for the serologic differences among var-
ious HBV subtypes. An Ala159-to-Gly159 change is frequently
detected in ayw1 and ayw2, whereas an Ala159-to-Val159
change is only found in adrq2 [12]. However, the same change
observed in infant 5 was of subtype adw. Together with the
decreased binding to the “a” determinant–specific antibody and
the predicted alteration in its antigenicity profile (figure 1), this
HBsAg variant may well be able to escape HBV vaccines.
In addition to the mutations on the major hydrophilic loop,
the Phe183-to-Cys183 mutation and the double mutations
(Thr118 to Lys118 and Val184 to Ala184) were identified in
our study. The decreased binding of the HBsAg Phe183-to-
Cys183 variant to “a” determinant–specific antibody was in-
triguing and should warrant further investigation, as it was
located outside the major hydrophilic loop. Both mutations
were preexisting in the serum of the respective mothers, sug-
gesting that they were vertically transmitted. Moreover, they
appeared to be stable over time, as they were still detected 8
and 14 years after the infants were born (table 1). The Phe183-
to-Cys183 mutation has not been detected in Southeast Asia
but was present in all serum samples analyzed from Central
America [14]. These two mutations coincided with a previously
mapped class I HLA-A2–restricted T cell epitope [15]. However,
detailed analysis, including HLA typing and measurement of
Figure 1. Antigenicity profile of HBsAg variants with mutation on major hydrophilic loop (aa 118, 120, and 159). Prediction analysis was done by Hoop/Woods method (Omiga system), with 7 aa
as window that corresponds to average size of epitope. Predicted alterations of antigenicity were seen for mutant 118 (T118KHBsAgadw), 120 (P120SHBsAgadw), and 159 (A159VHBsAgadw) compared
with profile of wild type (WTHBsAgadw).

Downloaded from https://academic.oup.com/jid/article-abstract/179/1/259/881393 by guest on 28 December 2018

JID 1999;179 (January) HBsAg Variants in Singapore Infants
cytotoxic T lymphocyte activity, is needed to assess their func-
tional significance.
In conclusion, HBsAg variants with changes outside the “a”
determinant have been identified in 6 immunized Singapore
infants, from the same pool as previously reported [7]. These
include Asn116 to Thr116, Val118 to Ala118, Pro120 to Ser120,
Ala159 to Val159, Phe183 to Cys183, and Val184 to Ala184.
The significance of these mutants is suggested by the decreased
binding of some of them to “a” determinant–specific mono-
clonal antibody. Identification of these new HBsAg variants
and their functional analysis may provide further understand-
ing of the antigenic structure of HBV.
Acknowledgments
We thank the Singapore National Medical Research Council and
the Shaw Foundation for their support.
References
1. Carman WF, Zanetti AR, Karayiannis P, et al. Vaccine induced escape mutant
of hepatitis B virus. Lancet 1990; 336:325–9.
2. Ogata N, Zanetti AR, Yu M, Miller RH, Purcell RH. Infectivity and path-
ogenicity in chimpanzees of a surface gene mutant of hepatitis B virus
that emerged in a vaccinated infant. J Infect Dis 1997; 175:511–23.
3. Carman WF, Korula J, Wallace L, MacPhee R, Mimms L, Decker R. Ful-
minant reactivation of hepatitis B due to envelope protein that escaped
detection by monoclonal HBsAg ELISA. Lancet 1995; 345:1406–7.
4. Hsu HY, Chang MH, Ni YH, Lin HH, Wang SM, Chen DS. Surface gene
mutants of hepatitis B virus in infants who develop acute or chronic
infections despite immunoprophylaxis. Hepatology 1997; 26:786–91.
263
5. Protzer-Knolle U, Naumann U, Bartenschlager R, et al. Hepatitis B virus
with antigenically altered hepatitis B surface antigen is selected by high-
dose hepatitis B immune globulin after liver transplantation. Hepatology
1998; 27:254–63.
6. Carman WF, Trautwein C, Van Deursen FJ, et al. Hepatitis B virus envelope
variation after transplantation with and without hepatitis B immune glob-
ulin prophylaxis. Hepatology 1996; 24:489–93.
7. Oon CJ, Lim GK, Ye Z, et al. Molecular epidemiology of hepatitis B virus
vaccine variants in Singapore. Vaccine 1995; 13:699–702.
8. Okamoto H, Yano K, Nozaki Y, et al. Mutations within the S gene of hepatitis
Downloaded from https://academic.oup.com/jid/article-abstract/179/1/259/881393 by guest on 28 December 2018
B virus transmitted from mothers to babies immunized with hepatitis B
immune globulin and vaccine. Pediatr Res 1992; 32:264–8.
9. Lee PI, Chang LY, Lee CY, Huang LM, Chang MW. Detection of hepatitis
B surface gene mutation in carrier children with or without immunopro-
phylaxis at birth. J Infect Dis 1997; 176:427–30.
10. Oon CJ. Evolution and transmission of hepatitis B virus mutants. In: Zuck-
erman AJ, ed. Hepatitis B in the Asia-Pacific region. R Coll Physicians
Lond 1997; 1:177–90.
11. Carman WF, Van Deursen FJ, Mimms LT, et al. The prevalence of surface
antigen variants of hepatitis B virus in Papua New Guinea, South Africa,
and Sardinia. Hepatology 1997; 26:1658–66.
12. Norder H, Hammas B, Lee SD, et al. Genetic relatedness of hepatitis B viral
strains of diverse geographical origin and natural variations in the primary
structure of the surface antigen. J Gen Virol 1993; 74:1341–8.
13. Wallace LA, Echevarria JE, Echevarria JM, Carman WF. Molecular char-
acterization of envelope antigenic variants of hepatitis B virus from Spain.
J Infect Dis 1994; 170:1300–3.
14. Arauz-Ruiz P, Norder H, Visona KA, Magnius LO. Molecular epidemiology
of hepatitis B virus in Central America reflected in the genetic variability
of the small S gene. J Infect Dis 1997; 176:851–8.
15. Nayersina R, Fowler P, Guilhot S, et al. HLA A2 restricted cytotoxic T
lymphocyte responses to multiple hepatitis B surface antigen epitopes
during hepatitis B virus infection. J Immunol 1993; 150:4659–71.