Download Correct binding of viral X protein to UVDDB-p127 cellular

Oncogene (2000) 19, 4427 ± 4431
ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00
www.nature.com/onc
SHORT REPORT
Correct binding of viral X protein to UVDDB-p127 cellular protein is
critical for ecient infection by hepatitis B viruses
Delphine Sitterlin1,3, FrancËoise Bergametti1, Pierre Tiollais1, Bud C Tennant2 and
Catherine Transy*,1
1
Institut Pasteur, Unite de recombinaison et expression geÂneÂtique (INSERM U163), 28 rue du Dr Roux 75724 Paris Cedex 15,
France; 2Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, C2-005 Veterinary Medical Center,
Ithaca, New York, NY 14853-6401, USA
A fully e€ective treatment of chronic human hepatitis B
virus (HBV) infection is still missing and HBV remains
the ®rst etiological agent of liver cancer. Although the
viral regulatory X protein is essential for infection, its
mode of action remains obscure, due the lack of an in
vitro infection system. In the accompanying study, we
showed the functional importance of interaction between
X and the host protein UVDDB-p127, in the transactivation and apoptotic properties of the viral protein. Here,
we addressed the biological role of X-UVDDB interaction in the infectious process using a genetic approach in
the woodchuck virus closely related to HBV. We show
that (i) mutations in X, which markedly a€ect UVDDBbinding, also abolished productive infection in woodchucks, (ii) in the few cases where mutant viruses led to
infection, compensatory mutations had occurred in the X
gene of the viral progeny, which restored correct
UVDDB-binding. We conclude that ecient viral
replication in vivo requires proper X-UVDDB interaction. The interaction may thus provide a novel
therapeutic target for the treatment of hepatitis B.
Oncogene (2000) 19, 4427 ± 4431.
Keywords: hepatitis B virus; X protein; UV-damaged
DNA binding protein; virus-host interaction; therapeutic target
Viruses have evolved numerous and complex mechanisms of host cell metabolism diversion for the bene®t of
their multiplication. When unraveled, these mechanisms, which mostly rely on interactions between viral
and cellular macromolecules, may not only shed light
on biological cell pathways and on pathogenic
processes, but also help to identify therapeutic targets.
Chronic infection by the human hepatitis B virus
(HBV), which dramatically increases the risk of liver
cancer (Beasley et al., 1981), a€ects an overwhelming
number of people worldwide, recently estimated at 350
million by the World Health Organization (Kane,
1998). The currently available therapy, namely interferon a treatment, is only partially e€ective (Krogsgaard, 1998). Furthermore, control of viremia by
*Correspondence: C Transy
3
Current address: EMBL, Meyerhofstr., D-69117 Heidelberg,
Germany
Received 2 May 2000; revised 22 June 2000; accepted 27 June 2000
administration of viral polymerase inhibitors may
prove dicult, given the recurrent emergence of
resistant variants (Zoulim and Trepo, 1998). Similar
diculties have been encountered in anti-human
immunode®ciency virus (HIV) therapy. As in the case
of HIV infection, control of chronic HBV infection and
of its evolution to liver cancer would likely bene®t
from combination therapies targeting di€erent steps of
viral replication.
Despite the early cloning of the 3.2 kb HBV genome,
which revealed only four functional open reading
frames (Galibert et al., 1979), full understanding of
the virus life cycle has not been achieved so far.
Indeed, due to the lack of permanent cell lines
permissive for infection, we still poorly understand
the ®rst steps of the infection process: virion binding to
an as yet non-identi®ed cellular receptor, virus entry,
transport of the nucleocapsid to the nucleus and
conversion of the circular partially double-stranded
DNA genome into a covalently closed circular form
(ccc DNA), which can be transcribed by the cellular
machinery. In contrast, it has been possible to unravel
the downstream steps of viral replication because
several hepatoma cell lines support the replication of
transfected recombinant genomes. As a result, encapsidation of the viral transcript, termed pregenome, which
occurs in host cell cytoplasm, as well as its reverse
transcription mediated by the viral polymerase, are
now understood in great detail (reviewed by Nassal
and Schaller, 1996).
Besides HBV, the hepadnavirus family includes
viruses infecting woodchucks (WHV), ground squirrels
(GSHV) and more distantly related members, such as
the duck hepatitis B virus (DHBV) (reviewed by
SchoÈdel et al., 1989). In contrast to the non-oncogenic
avian hepadanaviruses, the mammalian oncogenic
hepadnaviruses possess a regulatory gene termed X,
which encodes a small protein of *17 kDa in size.
Since the X gene is dispensable for the replication of in
vitro transfected hepadnaviral genomes (Blum et al.,
1992; Zoulim et al., 1994) and has no homolog in the
DHBV genome, X was ®rst considered an accessory
protein. Quite the opposite could be concluded
however, when X-null engineered WHV mutants
turned out to be unable to trigger infection in
woodchucks (Chen et al., 1993; Zoulim et al., 1994).
Since these pioneering observations the mechanisms,
which underlie the essential function carried out by X
protein in the viral life cycle and its possible
implication in the virally induced carcinogenesis, have
Pivotal role of X-UVDDB interaction in hepatitis B viruses infection
D Sitterlin et al
4428
remained unexplored. In sharp contrast, numerous
studies have addressed the functional properties of the
viral protein out of the context of viral infection.
Among the multiple activities reported for the viral
protein, two have received wide experimental support:
Transactivation of transcription, in an indirect and
promiscuous way (reviewed by Rossner, 1992) and
stimulation of the apoptotic pathway (Bergametti et
al., 1999; Chirillo et al., 1997; Kim et al., 1998; Su and
Schneider, 1997). For either activity, distinct requirements have been reported depending on the studies and
so far no unifying molecular mechanism has emerged.
In the accompanying study of WHV X protein
mutants (WHx), selected for markedly altered binding
to the host protein UVDDB, we obtained strong
evidence for the functional importance of proper XUVDDB interaction in both transactivation and
apoptosis (Sitterlin et al., submitted). Our choice of a
genetic approach in the WHV virus, was guided by the
unique possibility o€ered by this model, to perform
experimental infection with recombinant viruses, after
assaying their replication capacity in in vitro transfection experiments (Chen et al., 1993; Zoulim et al.,
1994). Thus in the present study, we further analysed
the biological relevance of the interaction by addressing its role in the context of viral infection in vivo.
WHV genomes carrying mutant WHx alleles are
functional in vitro
In the accompanying study, we created WHx single
and double substitution mutants, which fell within
three classes with respect to their binding to the cellular
protein UVDDBp127 (Table 1): Class 1, represented by
WX1 mutant, exhibited a markedly increased binding
as compared to the wild-type protein; class 2,
represented by WX7 mutant showed a close-to-wt
binding; class 3, represented by WX2, WX3, WX6 and
WX15 mutants displayed severely decreased or abolished interaction with UVDDB. As summarized in
Table 1, class 2 mutant is competent for both
transactivation activity and stimulation of apoptosis
whereas, class 1 and class 3 mutants are cis-inactive
molecules with respect to both functions. However, in
contrast to class 3 mutants, which show both cis and
trans loss-of-function, class 1 mutant, which binds
UVDDB more eciently than wt X, has gained a
trans-dominant negative activity over the wt molecule.
At the nucleotide level, mutations in the X gene were
designed in such a way that the viral polymerase
remained unchanged (except in the case of WX1 where
Serine 872 was substituted by a Threonine residue).
Similarly the transcription factor binding sites, which
map to the enhancer I region of the virus (Fourel et al.,
1996), were not a€ected. This allowed us to replace the
wt X gene in the WHV8 infectious genome by its
mutant counterparts, without a€ecting a functional
feature other than X itself. We next transfected these
mutant viral genomes into HepG2 hepatoma cells,
which are permissive for WHV replication. As shown
in Figure 1A, mutant genomes produced the same
transcript species as the wt genome construct and in
similar amounts, except WHV15, which showed
marked reduction in pregenomic transcript accumulation. Note however, that only a threefold di€erence
was obtained between WHV15 and wt WHV (data not
shown) when the raw signals were normalized using a
b-galactosidase transcript, produced by a co-transfected pCMVb-gal vector, which is only marginally
sensitive to X-mediated transactivation. Results of
three independent experiments indicated that mutant
genomes could sustain pregenomic RNA encapsidation
(Figure 1A). In our experimental conditions, only the
single stranded intermediate form of viral DNA was
detectable (Figure 1B), consistent with previous
observations made with similar WHV constructs and
the same recipient HepG2 cell line (Wei et al., 1998).
After normalization, it appeared that this viral DNA
species, which likely corresponds to minus strand DNA
synthesis, was decreased ®vefold in the case of WHV15
genome as compared to wt and other mutant genomes
(Figure 1B and Table 1). Altogether, these results
indicate the UVDDB-binding and transactivation
activities of WHx are dispensable for in vitro WHV
replication, in agreement with previous results obtained
with recombinant viruses unable to encode the X
protein (Blum et al., 1992; Zoulim et al., 1994).
WHx mutations conferring markedly altered
UVDDB-binding abolish or delay productive infection in
woodchuck
Since all mutant genomes were competent for replication in vitro, we next investigated their infectious
potentials in vivo. For each virus species, a group of
three woodchucks received intrahepatic injection of
recombinant DNA. Serum samples were collected at 2week intervals, starting 4 weeks post-injection. Sera
were assayed for WHsAg antigenemia and for the
presence of anti-WHc (capsid) and anti-WHs (envelope) antibodies (Table 2). Moreover, presence of viral
DNA was assessed in each serum sample by PCR
ampli®cation (Figure 2A). All three woodchucks
injected with wt DNA developed anti-WHc response
Table 1 Comprehensive summary of phenotypes associated with WHx mutations
Genome
Mutation
WHV wt
WHV7
WHV1
WHV2
WHV3
WHV6
WHV15
±
Q95R
V71D
N86D
H91R
Q95H
N86D; Q95H
a
In vitro activities of WHxa
UVDDB-bindingb Class Transactivation
Ù
Ù
Ú
Ø
Ø
Ø
Ø
2
2
1
3
3
3
3
+
+
7
7
7
7
7
Apoptosis
+
+
7
7
7
7
7
trans-dominant
+
7
7
ND
ND
Functional features of WHV genomes
Replicationc Infectiond Revertants
+
+
+
+
+
+
+/7
+
+
D
7
D
7
7
7
+
7
+
7
7
Sitterlin et al. (2000). bHorizontal, upward and downward arrows indicate wt, increased and decreased binding respectively. cTransfected
genomes. dD refers to delayed infection as compared to wt genome
Oncogene
Pivotal role of X-UVDDB interaction in hepatitis B viruses infection
D Sitterlin et al
received WHV1 and WHV3 mutants, signs of infection
were observed in two and one animal respectively, the
other animals remaining negative during the follow-up
period. In addition, in those animals that showed
infection markers, infection was delayed, as compared
with the overall kinetics of infection observed with wt
and WHV7 genomes (Table 2 and Figure 2A).
Animals, which received WHV2, WHV6 and WHV15
mutants remained negative for infection markers
during the follow-up period. It should be mentioned
that in our experimental conditions, the viral DNA
content of positive serum samples, was largely above
the detection limit since saturating PCR signals were
still obtained with 20-fold less input (data not shown).
Thus, only the WHV7 class 2 mutant virus, which
encodes an X mutant protein with wt UVDDBbinding, displayed an infectious potential similar to
that of wt genome (Table 1). In contrast, mutations of
WHx protein, which signi®cantly alter the binding to
UVDDB (class 1 and class 3 mutants) appeared to
compromise the infectious potential of the virus,
infection being either abolished or strongly delayed
(Table 1).
4429
Delayed infection occurring with mutant viruses is
associated with the emergence of X mutations, which
restore correct interaction with UVDDB
Figure 1 WHV genomes carrying mutant X alleles are functional in vitro. All WHV sequences were derived from pWHV8,
which contains a full-length genome from the WHV8 from the
WHV8 infectious viral strain (Girones et al., 1989). Numbering
starts at the ®rst A of the unique EcoRI site. The SacII-NsiI
fragments (positions 1593 ± 1910) carrying the introduced mutations were cloned into pWHV8. To create functional recombinant
viruses, intermediate constructs were generated by subcloning the
ApaI ± EcoRI fragments from either the monomeric site-directed
mutant or the wild-type WHV8 genomes (nucleotides 887 ± 3323)
into pBluescript II KS+ vector. The full-length genomes (EcoRI
fragment) were then cloned into the unique EcoRI site of the
isogenic intermediate constructs, generating 1.5 mer WHV
genomes, which contain two copies of the 887 ± 3223 region. (A)
Northern blot analysis of viral transcript accumulation and
pregenomic RNA encapsidation. HepG2 hepatoma cells were
transfected with the indicated recombinant viral genomes (10 mg)
and 1 mg of pCMV-b-Gal (Clontech), used as a control of
transfection eciency. Total RNA was prepared 36 h posttransfection and core particles extracted 96 h post-transfection,
as described by Wei et al. (1998). RNA was separated in 0.8%
formaldehyde-containing agarose gels. Following gel transfer, the
membrane was hybridized with a 32P-labeled WHV DNA probe.
Arrows indicate the positions of pregenomic, S and X transcripts.
Note that the unusual abundance of X transcript results from the
structure of the viral expression construct as previously reported
(Wei et al., 1998). The band visible below the S transcript
corresponds to background hybridization with 18S rRNA. (B)
Southern analysis of viral DNA synthesis. Encapsidated viral
DNA was separated in a 1.2% agarose gel. Lin. ss: singlestranded DNA. Signals seen at the bottom of the gel result from
transfected plasmid DNA cleaved by DNase treatment
and one animal detectable WHsAg antigenemia. In
addition, viral DNA was detectable in the sera of these
animals before appearance of serological markers
(Figure 2a). Virion concentration in the serum of these
animals was in the range of 105/ml, when titrated using
known amounts of recombinant viral DNA (data not
shown). All animals, which received WHV7 mutant,
showed viral DNA in their sera and two of them
developed an anti-WHc response. In the groups, which
We next characterized the X gene carried by the in vivo
produced viruses by sequencing PCR-ampli®ed viral
DNA from positive serum samples. In contrast to the
viruses produced after injection of wt and WHV7
genomes, WHV3 and WHV1 viral progenies di€ered
from the initially injected DNAs (Figure 2B and Table
1). Viral genomes produced in WHV3-injected animal
were wild-type in the X region (Figure 2B). Analysis of
WHV1 progeny revealed two species: one corresponding to the reversion toward wild-type sequence (GAC
codon ?GTC codon) and the other corresponding to a
new mutation at this codon (GAC?GCC), which
substituted the D residue by a A residue (i.e. a residue
with properties similar to that of the wt V residue).
Emergence of the latter species excluded that delayed
infection was due to cross-contamination by wt DNA
at the time of injection. Furthermore, it o€ered the
possibility to examine whether the newly acquired
mutation restored a correct binding of WHx protein to
UVDDB. When assayed for UVDDB interaction, this
new WHx version indeed exhibited a close-to-wt
binding phenotype (only 50% above the level of wt
protein-binding as compared to the fourfold increased
binding shown by WX1 mutant, data not shown).
Altogether these data clearly show that transcription
and reverse transcription of the mutant viral DNAs
occurs in vivo, a strong selective pressure then
operating for the restoration of proper WHx-UVDDB
interaction.
Concluding remarks
In the accompanying study (Sitterlin et al., submitted),
we showed that proper X-UVDDB interaction is a
prerequisite for in vitro X-mediated activities. Our
present ®ndings allow us to extend this functional
connection to viral replication capacity in vivo (Table
1). Speci®cally, only class 2 mutant, which showed
Oncogene
Pivotal role of X-UVDDB interaction in hepatitis B viruses infection
D Sitterlin et al
4430
Table 2 Follow-up of serological infection markers in injected
woodchucks
Transfected Incubation
No
genome
period a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
WHV wt
WHV1
WHV2
12
8
12
20
20
WHV3
WHV6
WHV7
WHV15
18
22
10
WHV serology
AntiAntiWHsAg
WHc Abc WHs Abc
b
7
7
+ (12 and 14)
+/7 (18)
7
+ (18 and 20)
7
7
7
7
7
+/7 (16)
7
7
7
7
7
7
7
7
7
+
+
+ (18)
+
7
+
7
7
7
7
7
+
7
7
7
+
+
7
7
7
7
+/7 (18)
+/7
+ (22)
+/7 (22)
7
7
7
7
7
7
7
+ (20)
7
7
7
+
+ (12)
7
7
7
7
Each DNA species was injected to a group of three animals. Starting
from 4 weeks post-injection, sera were collected at 2 week intervals,
during a 24-week period. Serological assays were performed as
described (Wei et al., 1998). Time is expressed in weeks. aNumbers
refer to the time at which, at least one serological marker was
detected positive. bNumber in brackets indicate time at which
samples were found positive. cAb: antibodies. Numbers in brackets
indicate time at which samples became positive
close-to-wt interaction with UVDDB, remained functional in contrast to the class 3 and class 1 mutants
with severely decreased and markedly increased
UVVDB-binding eciencies, respectively. Thus, three
properties are functionally linked to the ability of X to
engage a normal interaction with UVDDB: (i)
transactivation potential; (ii) proapoptotic e€ect; (iii)
infectious potential in woodchucks of recombinant
viruses (Table 1). In this respect, the in vivo selection of
a class 1-virus pseudorevertant is particularly signi®cant. Here, the new mutation led the decrease of
UVDDB-binding anity, from the high level associated with impaired function in the initial mutant,
toward a close-to-wt level. However, it should be
emphasized that if correct interaction with UVDDB
appears to be necessary for X function, it is clearly not
sucient per se, since C-terminal truncations of the
protein, which preserve the UVDDB-binding domain
inactivate transactivation (Kumar et al., 1996) and
infectious potential (Chen et al., 1993).
The emergence of revertant or pseudorevertant
viruses (Table 1) indicates that transcription and
reverse transcription could be initiated with several of
the mutant genomes. Since no revertant was observed
in similar studies using genomes unable to encode X
protein (Chen et al., 1993; Zoulim et al., 1994), it might
suggest that in our study, at least some mutants had
retained a residual activity sucient to sustain
replication initiation. Alternatively, X protein might
be qualitatively dispensable for the transcription and
reverse transcription steps of the virus life cycle in
agreement with the results of in vitro transfection
experiments. In the latter hypothesis, ecient virus
replication in the context of natural infection might
Oncogene
Figure 2 WHx-UVDDB interaction is critical for productive
infection. Viral DNA was extracted from 200 ml serum using the
QIAamp blood kit (Qiagen) in a ®nal volume of 50 ml. PCR
ampli®cations were performed using 5 ml of DNA sample and the
primers, oligo 1+ (5'-TGGTTAGGAATTTCCCTCA-3') (positions 827 ± 848) and oligo 27 (5'-GCGAGCAGCCATGGAAAGGACGT-3') (positions 1514 ± 1492). (A) The sera of the indicated
woodchucks were analysed by PCR for the presence of viral
DNA. PCR products were run in 1% agarose gels. The injected
recombinant genome is shown on the top and the time is
indicated on the right in weeks post-injection. Note that in the
absence of positive serological markers, infection was considered
signi®cant when viral DNA was detected at least at two successive
time points. (B) Analysis of viral progeny produced in
woodchucks. The X regions containing the initial mutations were
PCR-ampli®ed using oligo 1+ and oligo 37 (5'-GGTTACAGAAGTCGCATGCA-3') as primers and then and directly
sequenced. The corresponding positions in the WHV genome
are indicated on the right. Filled arrowheads mark the positions
where the in vivo-produced virus showed a nucleotide change as
compared to the injected mutant genome. In the case of WX3, the
G nucleotide at position 1773 reverted to an A, converting back
the CGT triplet coding for Arginine into the wild-type CAT
triplet coding for Histidine. In the case of woodchuck no 4, which
was injected with WX1 mutant, the A nucleotide at position 1717
was changed into a C nucleotide converting the GAC codon
(Aspartic acid) into a GCC codon (Alanine). In woodchuck no 6,
the A nucleotide at position 1717 was converted either into a C
nucleotide as in woodchuck no 4 or into the wild-type T
nucleotide. In viral progeny from WX7 mutant the mutated
codon that substituted the wild-type glutamine residue for an
Arginine residue was preserved (empty arrowhead)
require X to overcome an otherwise limiting step, such
as viral transcription level or ccc DNA ampli®cation.
Although future studies are required for a detailed
understanding of viral X protein mechanism of action,
our work pinpoints X-UVDDB interaction as a
potential therapeutic target. Drugs targeting proteinprotein interactions are still scarce. However, the
Pivotal role of X-UVDDB interaction in hepatitis B viruses infection
D Sitterlin et al
potent and widely used immunosuppressive drugs,
cyclosporin A and FK506 provide interesting examples.
Both drugs lead to inhibition of calcineurin, a critical
intracellular component of the immune response, via
the modulation of a protein-protein interaction (Ho et
al., 1996). Our description of a dominant-negative X
mutant indicates that inhibitors of X activity might be
derived from the X molecule itself. We will thus
explore the possibility of developing anti-hepatitis B
virus drugs, based on their ability to inhibit XUVDDB interaction.
Acknowledgments
D Sitterlin was supported by fellowships from the
`MinisteÁre de l'Education Nationale et de la Recherche'
and the `Fondation pour la Recherche MeÂdicale'. We
thank S Whiteside, M Rosbash, F Lehembre, M Flajolet
and Y Wei for helpful suggestions and critical reading of
the manuscript. We thank C-A Renard for support.
4431
References
Beasley RP, Lin CC, Hwang LY and Chien CS. (1981).
Lancet, 2, 1129 ± 1133.
Bergametti F, Prigent S, Luber B, Benoit A, Tiollais P,
Sarasin A and Transy C. (1999). Oncogene, 18, 2860 ±
2871.
Blum HE, Zhang ZS, Galun E, von WeizsaÈcker F, Garner B,
Liang TJ and Wands JR. (1992). J. Virol., 66, 1223 ± 1227.
Chen HS, Kaneko S, Girones R, Anderson RW, Hornbuckle
WE, Tennant BC, Cote PJ, Gerin JL, Purcell RH and
Miller RH. (1993). J. Virol., 67, 1218 ± 1226.
Chirillo P, Pagano S, Natoli G, Puri PL, Burgio VL, Balsano
C and Levrero M. (1997). Proc. Natl. Acad. Sci. USA., 94,
8162 ± 8167.
Fourel G, Ringeisen F, Flajolet M, Tronche F, Pontoglio M,
Tiollais P and Buendia MA. (1996). J. Virol., 70, 8571 ±
8583.
Galibert F, Mandart E, Fitoussi F, Tiollais P and Charnay P.
(1979). Nature, 281, 646 ± 650.
Girones R, Cote PJ, Hornbuckle WE, Tennant BC, Gerin
JL, Purcell RH and Miller RH. (1989). Proc. Natl. Acad.
Sci. USA, 86, 1846 ± 1849.
Ho S, Clipstone N, Timmermann L, Northrop J, Graef I,
Fiorentino D, Nourse J and Crabtree GR. (1996). Clin.
Immunol. Immunopathol., 80, S40 ± S45.
Kane MA. (1998). Vaccine, 16, S104 ± S108.
Kim H, Lee H and Yun Y. (1998). J. Biol. Chem, 273, 381 ±
385.
Krogsgaard K. (1998). J. Viral. Hepat., 5, 389 ± 397.
Kumar V, Jayasuryan N and Kumar R. (1996). Proc. Natl.
Acad. Sci. USA, 93, 5647 ± 5652.
Nassal M and Schaller H. (1996). J. Viral. Hepatitis, 3, 217 ±
226.
Rossner MT. (1992). J. Med. Virol., 36, 101 ± 117.
Sitterlin D, Bergametti F and Transy C. (2000). Oncogene,
19, 4417 ± 4426.
SchoÈdel F, Sprengel R, Weimer T, Fernholz D, Schneider R
and Will H. (1989). Advances in Viral Oncology, Vol. 8.
Klein G (ed.). Raven Press: New York, pp 73 ± 102.
Su F and Schneider RJ. (1997). Proc. Natl. Acad. Sci. USA,
94, 8744 ± 8749.
Wei Y, Tennant B and Ganem D. (1998). J. Virol., 72, 6608 ±
6613.
Zoulim F, Saputelli J and Seeger C. (1994). J. Virol., 68,
2026 ± 2030.
Zoulim F and Trepo C. (1998). J. Hepatol., 29, 151 ± 168.
Oncogene