Kuo, M. Y.-P., J. Goldberg, L. Coates, W. Mason, J. Gerin, and. J. Taylor. ... Alan. R. Liss, Inc., New York. 28. Nakabeppu, Y., K. Ryder, and D. Nathans. 1988.
JOURNAL OF VIROLOGY, Sept. 1990, p. 4051-4058 0022-538X/90/094051-08$02.00/0 Copyright © 1990, American Society for Microbiology
Vol. 64, No. 9
Characterization of Hepatitis Delta Antigen: Specific Binding Hepatitis Delta Virus RNA
to
BAKER,"12 SUGANTHA GOVINDARAJAN 14 MICHAEL M. C. LAI' 2* Howard Hughes Medical Institute' and Departments of Microbiology2 and Pathology,4 University of Southern California, School of Medicine, Los Angeles, California 90033-1054, and Institute of Molecular Biology, Academia Sinica, JU-HUNG LIN,"2'3 MING-FU CHANG,"2t SUSAN C.
AND
Nankang, Taiwan 115293 Received 9 August 1989/Accepted 26 May 1990
It has previously been shown that human hepatitis virus delta antigen has an RNA-binding activity (Chang et al., J. Virol. 62:2403-2410, 1988). In the present study, the specificity of such an RNA-protein interaction was demonstrated by expressing various domains of the delta antigen in Escherichia coli as TrpE fusion proteins and testing their RNA-binding activities in a Northwestern protein-RNA immunoblot assay and RNA gel mobility shift assay. Hepatitis delta virus (HDV) RNA bound specifically to the delta antigen in the presence of an excess amount of unrelated RNAs and a relatively high salt concentration. Both genome- and antigenomesense HDV RNAs and at least two different regions of HDV genomic RNA bound to the delta antigen. Surprisingly, these two different regions of HDV genomic RNA could compete with each other for delta antigen binding, although they do not have common nucleotide sequences. In contrast, this binding could not be competed with by other viral or cellular RNA. Since both the genomic and antigenomic HDV RNAs had strong intramolecular complementary sequences, these results suggest that the binding of delta antigen is probably specific for a secondary structure unique to the HDV RNA. By expressing different subdomains of the delta antigen, we found that the middle one-third of delta antigen was responsible for binding HDV RNA. Neither the N-terminal nor the C-terminal domain bound HDV RNA. Binding between the delta antigen and HDV RNA was also demonstrated within the HDV particles isolated from the plasma of a human delta hepatitis patient. This in vivo binding resisted treatment with 0.1% sodium dodecyl sulfate and 0.5% Nonidet P-40. In addition,
showed that the antiserum from a human patient with delta hepatitis reacted with all three subdomains of the delta antigen, indicating that all of the domains are immunogenic in vivo. These studies demonstrated the specific interaction between delta antigen and HDV RNA. we
Hepatitis delta virus (HDV) is a defective virus associated with severe hepatitis in humans (1, 13, 19). It requires hepatitis B virus as a helper virus (32). The HDV particle is a 36-nm enveloped virus containing the hepatitis B virus surface antigen on the virion surface. Inside the envelope is the HDV-specific delta antigen (HDAg) (32). The HDAg is also detected in the nuclei of infected hepatocytes (31). Delta antigen is the only HDV-specific structural protein in the virion. The viral genome is a circular single-stranded RNA of approximately 1.7 kilobases (kb) (20, 26, 41) which has extensive intramolecular complementary sequences. Thus, its RNA structure resembles that of viroid and some plant virusoid RNAs. HDV RNA has a unique autocatalytic cleavage and ligation activity (24, 34, 35, 43, 44). It also has five open reading frames which can potentially code for proteins of more than 100 amino acids each (26). One of these open reading frames, in the antigenomic sense, encodes delta antigen (8, 23, 26, 41). The remaining open reading frames have not been demonstrated to encode any proteins. The delta antigens in the liver or serum of delta hepatitis patients are composed of two protein species of approximately 27 and 24 kilodaltons (kDa) (2, 4, 29, 42, 45). The same two proteins are detected in the liver and serum of chimpanzees experimentally infected with HDV (2, 45). Recently, it has been demonstrated that these two proteins Corresponding author. t Present address: Department of Biochemistry, University, College of Medicine, Taipei, Taiwan.
derived from two different HDV RNA species with different coding capacities (Y.-P. Xia, M.-F. Chang, D. Wei, S. Govindarajan, and M. M. C. Lai, Virology, in press). The HDAg is phosphorylated and is localized in the nuclei both in infected animals (14) and in transfected cells (8). It also has an RNA-binding activity (8). However, the specificity of this RNA-protein interaction has not yet been demonstrated. Since HDAg appears to be required for the replication of HDV RNA (22), the interaction between HDAg and HDV RNA probably plays an important role in HDV RNA synthesis. Prevailing evidence suggests that HDV RNA replicates through a rolling-circle mechanism. This evidence includes the detection, in HDV-infected hepatocytes (9, 27), of HDV-specific RNA of longer than genomic length, which represents intermediates of RNA replication, and the ability of HDV RNA to cleave and ligate itself autocatalytically (24, 34, 35, 43, 44). Conceivably, the interaction of HDAg with HDV RNA might regulate these activities or RNA replicawere
tion. To further characterize the interaction between the HDAg and HDV RNA, we have constructed TrpE fusion proteins containing various domains of HDAg and tested their RNAbinding activities. In this report, we demonstrate that this interaction is very specific and occurs between HDV RNA and the middle one-third of HDAg. Furthermore, this interaction can be demonstrated within the HDV particle. MATERIALS AND METHODS Construction of plasmids expressing TrpE-HDAg fusion proteins. The 0.9-kb PstI-SstII insert of plasmid pECE-d-E,
*
National Taiwan
4051
4052
J. VIROL.
LIN ET AL.
Sma I digestion
Pst I,Sst II digestion
T4
IdBAP
DNA polymerase
Sma I/Sst II
Junction
-ATC
sequences
-TAG GGG
CCC
|
Pst I/Sma I
GGG------[TGA--------ACT----GGG]
GGG GATCCC
CTA-
FIG. 1. Schematic diagram of the construction of a TrpE-HDAg plasmid vector for expression of HDAg. pATH-D was designed so that it expresses, in Escherichia coli, a fusion protein of TrpEHDAg containing amino acids 11 to 214 of HDAg. Boxes with shaded areas represent TrpE sequences. Solid boxes indicate the HDAg coding region. The HDAg stop codon is represented by an asterisk. The solid circle represents the simian virus 40 origin of replication. The sequences at the junctions of the pATH-2 vector and HDV insert are shown. BAP, Bacterial alkaline phosphatase.
which contains a 1.3-kb EcoRI fragment of HDV cDNA (26), was blunt-ended by using T4 DNA polymerase and subcloned, in frame, into the SmaI site of pATH-2 (11) DNA to generate pATH-D (Fig. 1). pATH-D expresses a TrpEHDAg fusion protein containing amino acids 11 to 214 of HDAg (26). The 0.7-kb StuI-XbaI fragment was removed from the pATH-D, and the remaining DNA was filled in, by using the Klenow fragment of DNA polymerase I, and ligated with DNA ligase to generate pATH-N, which contains sequences representing the N-terminus of the HDAg (amino acids 11 to 78). The purified 0.7-kb StuI-XbaI fragment of pATH-D was subcloned, in frame, into the SmaIXbaI site of a pATH-2 vector to generate pATH-MC, which contains sequences representing the middle portion and carboxyl terminus of the HDAg (amino acids 79 to 212). A 0.4-kb SmaI-XbaI fragment was removed from the plasmid MC, and the remaining DNA was filled in by using the Klenow fragment of DNA polymerase I before self-ligation to generate pATH-M, which contains
sequences
represent-
ing only the middle portion of the HDAg (amino acids 79 to 163). A 0.3-kb SalI insert was removed from pATH-D, and the remaining DNA was filled in by using the Klenow fragment of DNA polymerase I and religated to generate pATH-X. A 0.46-kb SmaI fragment was further removed from the pATH-X plasmid, and the remaining DNA was religated to generate pATH-C, which represents the Cterminus of the HDAg (amino acids 164 to 212). Construction of other plasmids. The plasmid pHD-P, which contains the protein-conding domain of HDV RNA (7), was constructed by excising an EcoRI fragment (positions 965 to
487 through 1683/0) from plasmid pS29, which contains the full-length cDNA of HDV RNA (Sall-Sall, 1,683 base pairs [bp]). The fragment was cloned into the EcoRI site of pTZ18U (U.S. Biochemicals). The plasmid pHD-V, which contains the viroidlike domain (7) of HDV RNA, was constructed by inserting a SalI-EcoRI fragment (positions 487 to 965), derived from pS29, into the polylinker of pTZ-18U between the EcoRI and Sall sites. Transcription by T7 RNA polymerase of pHD-P, linearized by Hindlll, yielded an RNA of 1.2 kb, corresponding to the protein-coding domain (7) of the genomic-sense HDV RNA. Transcription of pHDV, linearized by Sall, yielded a genomic-sense HDV RNA subfragment corresponding to the viroidlike domain (7). Preparation of TrpE-HDAg fusion proteins and HDAgspecific antiserum. The preparation of TrpE-HDAg fusion proteins and immunization schedule followed the procedures described by Hardy and Strauss (17) with minor modifications. Briefly, a culture of MC1061 cells containing plasmids encoding the hybrid TrpE-HDAg fusion proteins was induced with 3-indoleacrylic acid, bacterial lysates were prepared, and the TrpE-HDAg fusion proteins were concentrated in the insoluble fraction. For the purpose of immunization, the insoluble fraction of TrpE-HDAg fusion protein D (approximately 500 p.g) was purified by electrophoresis on a 7.5% polyacrylamide gel containing 0.1% sodium dodecyl sulfate (SDS). The fusion protein was identified by staining with Coomassie blue and excised from the gel. The gel slice was homogenized in 1 ml of phosphatebuffered saline and then emulsified with an equal volume of Freund complete adjuvant. This mixture was injected subcutaneously along the back of New Zealand White rabbits weighing 3 to 4 kg. The rabbits were boosted every 2 weeks by injection in the hindleg with 200 to 400 ,ug of gel-purified fusion protein D in incomplete Freund adjuvant. Rabbits were bled from the ear vein 1 week following boost injections. Serum was separated, and antibody titer was determined by enzyme-linked immunosorbent assay. RNA-protein binding assays by the Northwestern blot procedure. The Northwestern protein-RNA blot analysis was carried out as described previously (8). Briefly, TrpE-HDAg fusion proteins which contained different portions of the HDAg were prepared as described above. The fusion proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) (7.5% polyacrylamide) and then electrotransferred to a nitrocellulose membrane at 200 mA overnight at 4°C. The membrane was then incubated in standard binding buffer (SBB; 10 mM Tris hydrochloride [Tris-HCl, pH 7.0], 1 mM disodium EDTA, 50 mM NaCl, 0.02% bovine serum albumin, 0.02% Ficoll, and 0.02% polyvinylpyrrolidone). The binding buffer also contained a 32P-labeled RNA transcribed in vitro (105 cpm/ml per 1.7 kb of RNA; specific activity, 2 x 107 cpm/,Lg) and 100 pLg of nonspecific total cellular RNA isolated from DBT cells, a murine astrocytoma cell line, per ml as a nonspecific competitor (36). The probes used in this study included the full-length HDV genomic RNA, viroidlike domain (EcoRI-SalI fragment in the region between nucleotides 487 and 965) (26), a protein-coding domain (SalI-PstI fragment in the region between nucleotides 965 and 654 through 1683/0), the full-length antigenomic RNA of HDV, and mouse hepatitis virus (MHV) nucleocapsid RNA (37). The T7 polymerase transcription procedures were done as described previously (44). 32P-labeled RNA bound to the fusion proteins on the nitrocellulose membrane was visualized by autoradiography, and the amount of radioactivity bound to the membrane was determined by
VOL. 64, 1990
scintillation counting of specific bands excised from the nitrocellulose filter. RNA-protein binding studies by the RNA mobility shift assay. The RNA mobility shift assay was carried out by a modified procedure as described previously (18, 30, 38). Briefly, 20 ,ug of partially purified and heat-denatured TrpEHDAg fusion protein was incubated in 10 ,ul of buffer containing 2.5 mM Tris-HCl (pH 7.4), 100 mM NaCl, 2 mM MgCl2, 1 mM dithiothreitol, and 20% glycerol with 40 U of RNasin, 10 ,ug of tRNA, and 10 p.g of total cellular RNA at 4°C for 5 min. If required, unlabeled RNA competitors (usually 500 ng in 2 ,ul) were added, and the reaction mixes were incubated for an additional 10 min at 4°C. Subsequently, 1 ,ul (10 ng) of 32P-labeled RNA (approx. 20,000 cpm) was added, and the reaction mixture was incubated for an additional 30 min at 4°C. Then, electrophoresis was carried out in 1, 0.8, or 0.6% low-melting-point agarose gels with 10 mM sodium phosphate (pH 7.0) as the running buffer, maintaining the voltage at about 20 to 25 V at 4°C, for 15 h. The gels were then dried and visualized by autoradiography. Slot-blot RNA analysis. HDV genomic RNA from the plasma of an HDV-infected patient was prepared by disrupting hepatitis delta virions purified from the patient's serum with the lysis buffer containing 80 mM Tris-HCl (pH 6.8), 0.1 M 2-,-mercaptoethanol, and 0.1% (wt/vol) SDS, followed by immunoprecipitation with HDAg-specific antiserum prepared against TrpE-HDAg fusion protein. The final pellet containing HDV RNA was suspended in buffer (10 mM TrisHCl [pH 7.4], 60 mM NaCl, 1 mM EDTA, and 1% SDS), followed by extraction with phenol-chloroform (1:1) and ethanol precipitation. The ethanol-precipitated RNA was denatured in lOx SSC (1 x SSC contains 0.15 M NaCl and 15 mM sodium citrate)-13% formaldehyde at 65°C for 15 min and bound to a nitrocellulose membrane by filtration with a slot-blot apparatus (Schleicher & Schuell). The nitrocellulose membrane was then baked under vacuum for 1.5 h at 80°C and prehybridized in sealed plastic bags in 50% formamide-5 x SSC-50 mM sodium phosphate (pH 7.0)-1% SDS-5x Denhardt solution (lx Denhardt solution contains 0.02% each bovine serum albumin, polyvinylpyrrolidone, and Ficoll), and 250 p.g of denatured salmon sperm DNA per ml for 4 h at 42°C. Filters were then hybridized at 42°C overnight in the same solution with 32P-labeled DNA probes of HDV monomer RNA, which were prepared by using random hexamers as primers (12). After hybridization, filters were washed in 0.1 x SSC-0.1% SDS at 50°C and exposed to X-ray film. RESULTS Construction of plasmid DNA encoding TrpE-HDAg fusion protein and generation of antiserum against delta antigen. To facilitate studies of the protein structure and RNA-binding capacity of HDAg, plasmid DNA capable of expressing a TrpE-HDAg fusion protein was constructed. We used a bacterial expression vector containing the TrpE promoter (11) and inserted an HDV cDNA fragment encoding the delta antigen (Fig. 1). This plasmid encodes a protein consisting of the N-terminus of the TrpE protein and all of HDAg except for the N-terminal 10 amino acids. Bacteria transformed by this vector were induced by 3-indoleacrylic acid, and TrpEHDAg fusion protein was partially purified from the insoluble protein fraction. PAGE analysis of the insoluble fraction showed a predominant protein of 60 kDa (Fig. 2A). This protein was isolated from the gel and used for immunization
HEPATITIS DELTA VIRUS ANTIGEN AND RNA
4053
B.
A.
tkd)
M :*
I
1 2
3
_
_
kkd)
:7
97-It c68-s 29-
-
43-i 18x814-
29-%b.
FIG. 2. Expression of the TrpE-HDAg fusion protein D and specificity of the antiserum prepared against this fusion protein. (A) TrpE-HDAg fusion protein D containing a nearly full-length HDAg sequence was expressed as described in Materials and Methods and analyzed by SDS-PAGE (lane 1). The polyacrylamide gel was stained with Coomassie blue and dried. Molecular mass markers (lane M) were run in parallel. The arrow indicates the TrpE-HDAg fusion protein of 60 kDa. (B) Antiserum prepared against the TrpEHDAg fusion protein was used to immunoprecipitate in vitrotranslated HDAg. The [35S]methionine-labeled, in vitro-translated HDAg was prepared as described previously (8). The immunoprecipitated proteins were separated by SDS-PAGE and then exposed to an X-ray film. Lane 1, In vitro-translated HDAg; lane 2, in vitro-translated HDAg immunoprecipitated with preimmune serum; lane 3, in vitro-translated HDAg immunoprecipitated with antiserum prepared against TrpE-HDAg fusion protein. The arrowhead indicates the immunoprecipitated product of HDAg. Molecular mass markers are given on the left (in kilodaltons).
of rabbits. The antiserum obtained specifically precipitated the HDAg which was translated in vitro from the HDV RNA (Fig. 2B), confirming that this TrpE fusion protein contained the delta antigen. Construction and characterization of plasmids encoding different domains of HDAg as TrpE fusion proteins. We have previously shown that the delta antigen has an RNA-binding activity in vitro (8). However, the specificity of this RNAprotein interaction has not yet been studied. Analysis of the predicted amino acid sequence of HDAg showed that this protein, particularly its N-terminal two-thirds, had a highly basic nature, with 31.5% of the amino acid residues being arginine or lysine (Fig. 3). Furthermore, the protein also contained two stretches of sequences resembling the leucine zipper motif (16, 21, 28), which has been found in many 1
MSBSEDB
51
LIIUGIIGB
GCGBDILEW VSG (EEL ELDLBEEE IImENP DGEPP
AZ14Et4E
IDAP PL BGGFTDEE
101
D1REBLBaJ MQISSG-S
151
GVNPLEGGSB GAPGCaVS M2GVPESPFA BTG
201
PADPPFSPQS CBE
SLS.
ZEBBAGSVG EL&I TF DIBG SQEPtWDILF
FIG. 3. Predicted amino acid sequence of HDAg. The amino acid sequence of HDAg is predicted from the published sequence (26). The basic amino acids arginine (R) and lysine (K) are underlined. Leucine residues of leucine zipperlike motifs (3, 10, 25) are indicated by asterisks.
J. VIROL.
LIN ET AL.
4054
B.
A,
A. 2 aa 214
HDAg
TrpE
N M C MC
N M C MC
Fusiorn Protein (KDna) 60
"A. v
46 M
C
_
49 aa
|Y~4 4 aa
D
43 52
a
D.
C.
C.
B
Ir
46
5n
N
M
C
MC
N
M
C
N M C
*-68
FIG. 4. Structure and expression of the TrpE-HDAg fusion proteins D, N, M, C, and MC. (A) Schematic diagram of the structures of the TrpE-HDAg fusion proteins. The number of amino acids (aa) and predicted molecular mass for each mutant protein are indicated. (B) TrpE-HDAg fusion proteins were induced in E. coli, partially purified as described in Materials and Methods, and analyzed by SDS-PAGE. Coomassie blue staining of the gel is shown. Molecular mass markers (in kilodaltons) were run in parallel. (C) Western blot analysis of TrpE-HDAg fusion proteins was carried out with antiserum from a human delta hepatitis patient and detected by '25I-labeled protein A as described before (8). The autoradiogram is shown.
nucleic acid-binding proteins. This basic domain and leucine zipper sequence may account for the ability of HDAg to bind RNA. We therefore attempted to identify domains of HDAg which are responsible for the RNA-binding activity. For this purpose, HDAg was divided into three separate domains. These domains were fused with the bacterial TrpE protein (see Materials and Methods and Fig. 4A). These fusion proteins were induced by 3-indoleacrylic acid and shown to be the predominant proteins in the insoluble fraction of the bacteria (Fig. 4B). The electrophoretic mobility of the N-terminal fusion protein varied slightly depending on the electrophoretic conditions. These proteins were transferred to nitrocellulose membrane and detected by Western procedures with antiserum from a human delta hepatitis patient. The results in Fig. 4C show that the antibody bound to all three proteins, suggesting that all of them contain epitopes of the delta antigen expressed during infection in humans. Furthermore, since all three fusion proteins were detected by the serum from a delta hepatitis patient, this result indicates that all three protein domains are immunogenic during in vivo infection. Determination by Northwestern procedure of RNA-binding
FIG. 5. Northwestern blot analysis of HDAg with various RNA species. The different TrpE-HDAg fusion proteins were separated by SDS-PAGE, electrotransferred to nitrocellulose membranes, and then incubated with various 32P-labeled RNAs as described in Materials and Methods. The 32P-labeled RNAs used were (A) HDV genomic RNA, (B) MHV RNA; (C) HDV antigenomic RNA, and (D) the viroidlike domain of the HDV RNA. N, M, C, and MC represent the TrpE-HDAg fusion proteins containing the N-terminus, middle portion, C-terminus, and middle portion plus Cterminus of HDAg, respectively, as shown in Fig. 3.
domain on HDAg. The three plasmids expressing different TrpE-HDAg fusion proteins were used to determine the specificity of the interactions between HDAg and HDV RNA. The various fusion proteins were separated by SDSPAGE and transferred to a nitrocellulose membrane, which was then incubated with 32P-labeled RNA from various sources. When 32P-labeled full-length HDV genomic RNA was used, the RNA bound specifically to the middle domain (M) and did not bind to the N- or C-terminus of HDAg (Fig. 5A). Consistently, the combined middle and C-terminal domains also bound RNA. In contrast, when a murine coronavirus (MHV) RNA of comparable size and of the same specific activity was tested, practically no [32P]RNA bound to any of these fusion proteins (Fig. 5B). Since all of these binding studies were done in the presence of an excess amount (100 ,ig/ml) of total cytoplasmic RNA isolated from a mouse cell line as a nonspecific competitor RNA, this result suggests that HDV RNA binds specifically to the middle domain of the HDAg. The binding was equally effective when the binding study was performed in the presence of 150 mM NaCl, indicating the specificity of this interaction (data not shown). To determine the possible binding sites of the protein on the HDV RNA, we tested the antigenomic-sense HDV RNA and both the viroidlike domain (7) (HDV nucleotides 487 to 965), which includes all of the HDV sequences required for the cleavage of genomic-sense and anti-genomic-sense HDV RNA, and the protein-coding domain of HDV RNA (nucleotides 965 to 487) in this binding assay. The results showed that all three RNAs bound to the middle domain of the TrpEHDAg fusion protein (Fig. 5C and D, and data not shown). The relative binding efficiencies of different RNAs tested are
VOL. 64, 1990
HEPATITIS DELTA VIRUS ANTIGEN AND RNA
4055
TABLE 1. Relative binding of different domains of HDAg to various RNA species" Relative binding (% of control)
RNA or sequence
Expt
1
Genomic RNA MHVRNA
2
Antigenomic RNA Viroidlike domain Protein-coding domain
N
M
C
MC
3.5
76.0
0.3
100.0
0.8
4.8
0.1
0.1
3.2
100.0
0.1
b
3.4 3.5
62.9 45.2
0.5 0.1
a RNA-protein binding assays were performed by the Northwestern blot procedure as shown in Fig. 5. The TrpE fusion proteins N, M, C, and MC represent fusions of bacterial TrpE with the N-terminus, middle portion, C-terminus, and middle portion plus C-terminus, respectively, of the HDAg, as shown in Fig. 4. The 32P-labeled RNA bound to each protein was identified by autoradiography, cut from the gel, and quantitated by scintillation counting. In experiment 1, the relative binding was calculated as the percentage of total counts of genomic RNA bound to the fusion protein containing the middle and C-terminal domains. In experiment 2, total counts of antigenomic RNA bound to the fusion protein containing the middle portion was used as the normalization standard. MHV RNA is the MHV RNA encoding the nucleocapsid protein (37). b-, Not determined.
summarized in Table 1. Since the HDV RNA had a very high percentage of intramolecular complementarity, the antigenomic RNA had a primary sequence and secondary structure very similar to those of the HDV genomic RNA. This result further indicates that the middle domain of HDAg binds specifically to the HDV RNA. The finding that both the viroidlike and the protein-coding domains bound to the same extent to HDV RNA suggests that there are at least two different binding sites or that the two domains have a common secondary structure which is unique to HDV RNA. Determination of specificity of HDAg-HDV RNA interactions by RNA mobility shift assay. To further determine the specificity of the HDAg-RNA interactions, we used an RNA mobility shift assay to study the RNA-protein interaction. The 32P-labeled HDV RNA was incubated with the various TrpE-HDAg fusion proteins in the presence of excess cellular RNA and then analyzed by agarose gel electrophoresis under nondenaturing conditions. Figure 6A shows that the HDV RNA and the middle domain of HDAg formed a slow-moving RNA-protein complex. So did the middle- plus A 1 2 3 4 5 6
BI 2 3 4 5
carboxyl-terminal domains and the total HDAg. In contrast, the N- and C-terminal portions did not form such an RNAprotein complex. This result confirmed the result obtained by the Northwestern procedure. The specificity of such an RNA-protein interaction was further studied by competition assays. Figure 6B shows that the RNA-protein complex formed between the genomic RNA and the HDAg could be competed with by an excess of the cold genomic RNA. Furthermore, the antigenomic RNA could form a complex with the HDAg (lanes 4 and 5), while the coronavirus (mouse hepatitis virus) RNA did not. The nature of binding sites on the genomic RNA was studied further. Figure 7 shows that the viroidlike domain of the HDV RNA could form an RNA-protein complex with HDAg, confirming the results obtained from the Northwestern procedure. This complex formation could be competed with by an excess of the unlabeled viroidlike RNA itself or by total HDV RNA. In contrast, it could not be competed with by either MHV RNA or the RNA representing the plasmid vector (pT7). Most surprisingly, this complex could be competed with by the protein-coding domain (nucleotides 965 to 487) of HDV RNA, which did not overlap the viroidlike domain. Since these two domains do not share 1 2 3 4 5 6 7
k
,.
kA
-A
I,
L*4.
FIG. 6. RNA mobility shift assay. 32P-labeled HDV genomic RNA was complexed with different TrpE-HDAg fusion proteins and then separated by electrophoresis on 0.6% low-melting-point agarose gels. After electrophoresis, the gel was dried and autoradiographed. (A) Lanes: 1, 32P-labeled RNA only; 2, N-terminal domain of HDAg; 3, middle domain; 4, C-terminal domain; 5, middle- plus C-terminal domains; 6, full-length HDAg. (B) Lanes: 1, 32P-labeled genomic RNA only; 2, 32P-labeled genomic RNA and full-length HDAg; 3, 32P-labeled genomic RNA plus 50-fold excess of cold genomic RNA and HDAg; 4, 32P-labeled antigenomic RNA only; 5, 32P-labeled antigenomic RNA and HDAg.
E1I
FIG. 7. Competitive binding of HDAg with the viroidlike domain of HDV RNA. 32P-labeled viroidlike domain of HDV genomic RNA transcribed from pHD-V was complexed with the full-length HDAg and various cold RNA species (50-fold excess). The RNA-protein complexes were separated by electrophoresis on 1% low-meltingpoint agarose gels. Lanes: 1, [32P]RNA only; 2, [32P]RNA and HDAg; 3, [32P]RNA, HDAg, and excess cold viroidlike domain of HDV RNA; 4, [32P]RNA, HDAg, and excess cold full-length HDV RNA; 5, [32P]RNA, HDAg, and excess cold protein-coding domain of HDV RNA; 6, [32P]RNA, HDAg, and excess cold MHV (coronavirus) RNA; 7, [32P]RNA, HDAg, and excess cold plasmid RNA transcribed from pT7 vector itself.
4056
J. VIROL.
LIN ET AL.
1 2 3
4
-.r
A
1
2
B
(kb)
_I
-C -1 -2
_10
-3
w:
2p0-1,7--
FIG. 8. Lack of binding of single-stranded and double-stranded MHV RNA. 32P-labeled plus-strand mRNA 7 of MHV was transcribed in vitro from the plasmid construct containing the 3' end 1.7 kb of the MHV genomic sequence (C. K. Shieh, unpublished) and hybridized with the antisense construct of the same MHV sequence. After hybridization, the double-stranded RNA was used for binding with the near-full-length TrpE-HDAg fusion protein. Lanes: 1, positive-strand RNA only; 2, positive-strand RNA and HDAg; 3, double-stranded RNA only; 4, double-stranded RNA and HDAg.
nucleotide sequence (26), this result suggests that the binding of the HDAg to HDV RNA is probably specific not for HDV primary sequence but for the secondary structure of the HDV RNA. Since the entire HDV RNA could form a double-stranded rod structure because of strong intramolecular hydrogen bonding (20), we examined whether the binding of HDAg to HDV RNA was the result of the double-strandedness of the RNA. We synthesized both plus- and minus-strand RNA of MHV mRNA 7 (1.7 kb) and then hybridized them in vitro to form a double-stranded RNA. This RNA was used to interact with HDAg. As shown in Fig. 8, double-stranded RNA migrated more slowly than the single-stranded RNA, but neither of them bound to HDAg. This result suggests that the HDAg did not bind to the double-stranded RNA but rather to a secondary structure unique to HDV RNA. Binding of HDAg to HDV RNA in vivo. To determine whether the specific interaction between HDAg and HDV RNA observed in vitro also occurred in the hepatitis delta virion, we examined the status of HDV RNA in HDV virions isolated from the plasma of a patient with acute delta hepatitis. Electrophoretic analysis of the RNA showed that only one HDV-specific RNA species was detectable (Fig. 9A). This RNA migrated slightly faster than the monomeric RNA of 1.7 kb in a glyoxal-agarose gel, consistent with the interpretation that this RNA is a circular RNA (20, 23, 26, 41). To determine whether this RNA was bound to the delta antigen within the virus particle, the purified virus particle was disrupted with 0.1% SDS and then immunoprecipitated with the rabbit antiserum prepared against the TrpE-HDAg fusion protein. Immunoprecipitation was performed in the presence of 0.5% Nonidet P-40. The immunoprecipitate was extracted with phenol-chloroform (1:1), transferred to a nitrocellulose membrane, and probed with an HDV-specific DNA probe. The HDV-specific RNA was precipitated with this antiserum but not by the preimmune serum (Fig. 9B). A control in vitro-transcribed HDV RNA without HDAg could not be precipitated (Fig. 9B). This result indicates that the HDV RNA was bound to the HDAg inside the hepatitis delta
FIG. 9. Analysis of HDV RNA from the plasma of a delta hepatitis patient. (A) Northern (RNA) blot analysis of the HDV RNA. HDV RNA (lane 2) was prepared from the patient's plasma as described in Materials and Methods. The probe used was a 3p_ labeled monomer cDNA of HDV RNA (26). RNA size markers of 2.0 and 1.7 kb (lane 1) were made by in vitro transcription with T7 RNA polymerase of the pT7 plasmids containing HDV cDNA inserts of respective sizes and run in parallel. (B) Slot-blot analysis of HDV RNA that was immunoprecipitated with HDAg-specific antiserum. The patient's plasma was treated with 0.1% SDS and then precipitated with various antisera. The immunoprecipitate was extracted with phenol-chloroform, transferred onto a nitrocellulose membrane by filtration, and probed with 32P-labeled HDV cDNA. Slot 1 was HDV RNA transcribed in vitro and immunoprecipitated by HDAg-specific antiserum. Slots 2 and 3 represent HDV RNA from the plasma of a delta hepatitis patient which was immunoprecipitated with either preimmune serum (slot 2) or HDAg-specific antiserum prepared against fusion protein D (slot 3). C, In vitrotranscribed HDV RNA without immunoprecipitation (positive control).
virion by a bond which was stable after treatment with 0.1% SDS and 0.5% Nonidet P-40.
DISCUSSION Previous studies have suggested that HDAg, which is the only protein encoded by HDV, has an RNA-binding ability (8). The current study demonstrated the specificity of such an RNA-protein interaction. HDV RNA binds specifically to HDAg. Unrelated viral (MHV) RNA does not bind to HDAg. Although we have not exhaustively tested every kind of RNA, the fact that all of the binding studies performed in this report were done in the presence of a large excess of nonspecific competitor RNA isolated from uninfected mammalian cells suggests that this binding is very specific to the HDV RNA. This specific binding was demonstrated by two independent methods: a Northwestern blot procedure and an RNA mobility shift assay. The binding occurred at a relatively high salt concentration. Furthermore, binding could be competed with by homologous RNAs but not by other RNAs. These results suggest that there is a specific interaction between HDV RNA and HDAg. An additional interesting finding was that this specific binding was not the result of a specific nucleotide sequence but rather of a specific RNA conformation unique to HDV RNA. This conclusion was deduced from the finding that the binding of the viroidlike domain of HDV RNA to HDAg could be competed with by
VOL. 64, 1990
the protein-coding domain, which does not share sequence with the viroidlike domain. It is not clear what this RNA conformation is. The most logical explanation is that both domains form a similar rod structure because of intramolecular complementarity (20, 41); however, this study has ruled out the simple double-stranded structure as the basis of RNA-protein binding. Previous analysis has suggested that HDV RNA is capable of forming a more complex structure (7, 43). Thus, a tertiary structure or an alternative RNA conformation may be responsible for the specificity of this RNA-protein interaction. The function of the HDAg-HDV RNA interaction is not clear. This specific binding is consistent with the previous
finding that HDAg is localized in the nucleus, where HDV RNA replicates in infected hepatocytes (8, 15, 39). Furthermore, it has been demonstrated, by HDV cDNA transfection, that HDAg is required for the replication of HDV RNA (22). Thus, the specific binding of HDAg to HDV RNA may play an important role in HDV RNA replication. The binding domain on the HDAg protein is also specific; only the middle domain of the protein binds HDV RNA. It is interesting that this domain is highly hydrophilic and contains many basic amino acid residues (Fig. 3). However, the basic nature of the middle domain of the HDAg does not seem to be sufficient for its specific binding to HDV RNA, since the N-terminal domain of the protein is equally basic and yet it does not bind to the RNA. Therefore, the specific interaction between HDAg and HDV RNA probably has specific sequence requirements. Significantly, the HDV RNA sequence encoding the middle domain of the HDAg
has recently been shown to be conserved among clinical isolates of HDV (Y.-C. Chao, M.-F. Chang, I. Gust, and M. M. C. Lai, Virology, in press). Thus, the binding of HDAg to HDV RNA is an important function for HDV RNA replication. Both the N-terminal and middle domains of HDAg contain a stretch of leucine zipperlike sequences (Fig. 3). This type of sequence has been shown to induce the dimerization of the proteins (10, 33) and, as a result, is responsible for the binding of proteins to DNA (16, 21, 28). Indeed, both the N-terminal and middle domains can form dimers (unpublished observation), and yet only the middle domain binds HDV RNA. Thus, dimer formation is not sufficient for the specific interaction between HDAg and HDV RNA. In contrast, the C-terminus is relatively free of charged amino acids. No nucleic acid-binding properties were detected with this domain. HDAg has been shown to be an internal viral structural protein of HDV, i.e., it can only be detected when the viral
envelope is disrupted with detergents (5, 6, 32). However, HDAg does not form a classical nucleocapsid structure. The data presented in this report show that at least some of the HDAg indeed binds to the circular HDV RNA via a bond which is resistant to treatment with 0.1% SDS and 0.5% Nonidet P-40. Whether this interaction between HDAg and HDV RNA is the same as that observed in vitro is not yet clear. One additional piece of information on the properties of the HDAg was revealed from this study. Serum from a delta hepatitis patient detected all three domains of HDAg in an immunoblotting procedure. Thus, all three domains appear to be immunogenic in humans. This finding has recently been demonstrated independently (40). These bacterial TrpEHDAg fusion proteins may be suitable as a diagnostic probe in the future.
HEPATITIS DELTA VIRUS ANTIGEN AND RNA
4057
ACKNOWLEDGMENTS This work was supported by Public Health Service grant AI-26741 (to M.M.C.L.) from the National Institutes of Health and a grant from the National Science Council of the Republic of China. M.M.C.L. is an Investigator of the Howard Hughes Medical Institute. M.-F.C. is partially supported by a Smoot Fellowship awarded by Norris Cancer Center of the University of Southern California. S.C.B. is a postdoctoral fellow of the Arthritis Foundation. LITERATURE CITED 1. Arico, S., M. Aragona, M. Rizzetto, F. Caredda, A. Zanetti, G. Marinucci, S. Diana, P. Farci, M. Arnone, N. Caporaso, A. Ascione, P. Dentico, G. Pastore, G. Raimondo, and A. Craxi. 1985. Clinical significance of antibody to the hepatitis delta virus in symptomless HBsAg carriers. Lancet ii:356-358. 2. Bergmann, K. F., and J. L. Gerin. 1986. Antigens of hepatitis delta virus in the liver and serum of humans and animals. J. Infect. Dis. 154:702-705. 3. Biedenkapp, H., U. Borgmeyer, A. E. Sippel, and K.-H. Klempnauer. 1988. Viral myb oncogene encodes a sequence-specific DNA-binding activity. Nature (London) 335:835-837. 4. Bonino, F., K. H. Heermann, M. Rizzetto, and W. H. Gerlich. 1986. Hepatitis delta virus: protein composition of delta antigen and its hepatitis B virus-derived envelope. J. Virol. 58:945-950. 5. Bonino, F., B. Hoyer, E. Ford, J. W.-K. Shih, R. H. Purcell, and J. L. Gerin. 1981. The delta agent: HBsAg particles with delta antigen and RNA in the serum of an HBV carrier. Hepatology 1:127-131. 6. Bonino, F., B. Hoyer, J. W.-K. Shih, M. Rizzetto, R. H. Purcell, and J. L. Gerin. 1984. Delta hepatitis agent: structural and antigenic properties of the delta-associated particle. Infect. Immun. 43:1000-1005. 7. Branch, A. D., B. J. Benenfield, B. M. Baroudy, F. V. Wells, J. L. Gerin, and H. Robertson. 1989. A UV-sensitive structural element in a viroid-like domain of the hepatitis delta virus. Science 243:649-652. 8. Chang, M.-F., S. C. Baker, L. H. Soe, T. Kamahora, J. G. Keck, S. Makino, S. Govindarajan, and M. M. C. Lai. 1988. Human hepatitis delta antigen is a nuclear phosphoprotein with RNAbinding activity. J. Virol. 62:2403-2410. 9. Chen, P.-J., G. Kalpana, J. Goldberg, W. Mason, B. Werner, J. L. Gerin, and J. Taylor. 1986. Structure and replication of the genome of the hepatitis delta virus. Proc. Natl. Acad. Sci. USA 83:8774-9778. 10. Dang, C. V., M. Mcguire, M. Buckmire, and W. M. F. Lee. 1989. Involvement of the 'leucine zipper' region in the oligomerisation and transforming activity of human c-myc protein. Nature (London) 337:664-666. 11. Dieckmann, C. D., and A. Tzagoloff. 1984. Assembly of the mitochondrial membrane system. J. Biol. Chem. 260:1513-1520. 12. Feinberg, A. P., and B. Vogelstein. 1983. A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-13. 13. Govindarajan, S., K. P. Chin, A. G. Redeker, and R. L. Peters. 1984. Fulminant B viral hepatitis: role of delta agent. Gastroenterology 86:1416-1420. 14. Govindarajan, S., B. Lim, and R. L. Peters. 1984. Immunohistochemical localization of the delta antigen associated with hepatitis B virus in liver biopsy sections embedded in Araldite. Histopathology 8:63-67. 15. Gowans, E. J., B. M. Baroudy, F. Negro, A. Ponzetto, R. H. Purcell, and J. L. Gerin. 1988. Evidence for replication of hepatitis delta virus RNA in hepatocyte nuclei after in vivo infection. Virology 167:274-278. 16. Halazonetis, T. D., K. Georgopoulos, M. E. Greenberg, and P. Leder. 1988. c-jun dimerizes with itself and with c-fos, forming complexes of different DNA binding affinities. Cell 55:917-924. 17. Hardy, W. R., and J. H. Strauss. 1988. Processing the nonstructural polyproteins of Sindbis virus: study of the kinetics in vivo by using monospecific antibodies. J. Virol. 62:998-1007. 18. Heaphy, S., C. Dingwall, I. Ernberg, M. J. Gait, S. M. Green, J. K. A. D. Lowe, M. Singh, and M. A. Skinner. 1990. HIV-1 regulator of virion expression (Rev) protein binds to an RNA
4058
19.
20. 21. 22. 23.
24.
25.
26.
27.
28. 29.
30.
31.
32.
J. VIROL.
LIN ET AL.
stem-loop structure located within the Rev response element region. Cell 60:685-693. Jacobson, I. M., J. L. Dienstag, B. C. Werner, D. B. Brettler, P. H. Levine, and I. K. Mushahwar. 1985. Epidemiology and clinical impact of hepatitis D virus infection. Hepatology 5: 188-191. Kos, A., R. Dikema, A. C. Arnberg, P. H. van der Merde, and H. Schelekens. 1986. The HDV possesses a circular RNA. Nature (London) 323:558-560. Kouzarides, T., and E. Ziff. 1988. The role of the leucine zipper in the fos-jun interaction. Nature (London) 336:646-651. Kuo, M. Y.-P., M. Chao, and J. Taylor. 1989. Initiation of replication of the human hepatitis delta virus genome from cloned DNA: role of delta antigen. J. Virol. 63:1945-1950. Kuo, M. Y.-P., J. Goldberg, L. Coates, W. Mason, J. Gerin, and J. Taylor. 1988. Molecular cloning of hepatitis delta virus from an infected woodchuck liver: sequence, structure, and applications. J. Virol. 62:1855-1861. Kuo, M. Y.-P., L. Sharmeen, G. Dinter-Gottlieb, and J. Taylor. 1988. Characterization of self-cleaving RNA sequences on the genome and antigenome of human hepatitis delta virus. J. Virol. 62:4439-4444. Landschulz, W. H., P. F. Johnson, and S. L. McKnight. 1988. The leucine zipper: a hypothetical structure common to a new class of DNA-binding proteins. Science 240:1759-1764. Makino, S., M.-F. Chang, C.-K. Shieh, T. Kamahora, D. M. Vannier, S. Govindarajan, and M. M. C. Lai. 1987. Molecular cloning and sequencing of a human hepatitis delta virus RNA. Nature (London) 329:343-346. Makino, S., M.-F. Chang, C.-K. Shieh, T. Kamahora, D. M. Vannier, S. Govindarajan, and M. M. C. Lai. 1987. Molecular biology of a human hepatitis delta virus RNA, p. 549-564. In W. Robinson, K. Koike, and H. Will (ed.), Hepadna viruses. Alan R. Liss, Inc., New York. Nakabeppu, Y., K. Ryder, and D. Nathans. 1988. DNA-binding activities of three murine jun proteins: stimulation by fos. Cell 55:907-915. Pohl, C., B. M. Baroudy, K. F. Bergmann, P. J. Cote, R. H. Purcell, J. Hoofnagle, and J. L. Gerin. 1987. A human monoclonal antibody that recognizes viral polypeptides and in vitro translation products of the genome of the hepatitis D virus. J. Infect. Dis. 156:622-629. Query, C. C., R. C. Bentley, and J. D. Keene. 1989. A specific 31-nucleotide domain of Ul RNA directly interacts with the 70K small nuclear ribonucleoprotein component. Mol. Cell. Biol. 9:4872-4881. Rizzetto, M., M. G. Canese, S. Arico, 0. Crivelli, F. Bonino, C. G. Trepo, and G. Verme. 1977. Immunofluorescence detection of a new antigen-antibody system (delta/anti-delta) associated to the hepatitis B virus in the liver and in the serum of HBsAg carriers. Gut 18:997-1003. Rizzetto, M., B. Hoyer, M. G. Canese, J. W.-K. Shih, R. H.
33.
34. 35. 36.
37. 38.
39.
40.
41.
42.
43.
44. 45.
Purcell, and J. L. Gerin. 1980. Delta agent: association of delta antigen with hepatitis B surface antigen and RNA in serum of delta-infected chimpanzees. Proc. Natl. Acad. Sci. USA 77: 6124-6128. Schuermann, M., M. Neuberg, J. B. Hunter, T. Jenuwein, R.-P. Ryseck, R. Bravo, and R. Muller. 1989. The leucine repeat motif in Fox protein mediates complex formation with Jun/AP-1 and is required for transformation. Cell 56:507-516. Sharmeen, L., M. Y.-P. IKuo, G. Dinter-Gottlieb, and J. Taylor. 1988. Antigenomic RNA of human hepatitis delta virus can undergo self-cleavage. J. Virol. 62:2674-2679. Sharmeen, L., M. Y.-P. Kuo, and J. Taylor 1989. Self-ligating RNA sequences on the antigenome of human hepatitis delta virus. J. Virol. 63:1428-1430. Stohlman, S. A., R. S. Baric, G. N. Nelson, L. H. Soe, L. M. Welter, and R. J. Deans. 1988. Specific interaction between coronavirus leader RNA and nucleocapsid protein. J. Virol. 62:4288-4295. Stohlman, S. A., and M. M. C. Lai. 1979. Phosphoproteins of murine hepatitis viruses. J. Virol. 32:672-675. Surowy, C. S., V. L. Van Santen, S. M. Scheib-Wixted, and R. A. Spritz. 1989. Direct, sequence-specific binding of the human U1-70k ribonucleoprotein antigen protein to loop 1 of Ul small nuclear RNA. Mol. Cell. Biol. 9:4179-4186. Taylor, J., W. Mason, J. Summers, J. Goldberg, C. Aldrich, L. Coates, J. L. Gerin, and E. Gowans. 1987. Replication of human hepatitis delta virus in primary culture of woodchuck hepatocytes. J. Virol. 61:2891-2895. Wang, J.-G., R. W. Jansen, E. A. Brown, and S. M. Lemon. 1990. Immunogenic domains of hepatitis delta virus antigen: peptide mapping of epitopes recognized by human and woodchuck antibodies. J. Virol. 64:1108-1116. Wang, K.-S., O.-L. Choo, A. J. Weiner, J.-H. Ou, R. C. Najarian, R. M. Thayer, G. T. Mullenbach, K. J. Denniston, J. L. Gerin, and M. Houghton. 1986. Structure, sequence and expression of the hepatitis delta viral genome. Nature (London) 323:508-513. Weiner, A. J., Q.-L. Choo, K.-S. Wang, S. Govindarajan, A. G. Redeker, J. L. Gerin, and M. Houghton. 1988. A single antigenomic open reading frame of the hepatitis delta virus encodes the epitope(s) of both hepatitis delta antigen polypeptides p248 and p278. J. Virol. 62:594-599. Wu, H.-N., Y.-J. Lin, F.-P. Lin, S. Makino, M.-F. Chang, and M. M. C. Lai. 1989. Human hepatitis delta virus RNA subfragments contain an autocleavage activity. Proc. Natl. Acad. Sci. USA 86:1831-1835. Wu, H.-N., and M. M. C. Lai. 1989. Reversible cleavage and ligation of hepatitis delta virus RNA. Science 243:652-654. Zyzik, E., A. Ponzetto, B. Forzani, C. Hele, K. Heermann, and W. H. Gerlich. 1987. Proteins of hepatitis virus in serum and liver, p. 565-577. In W. Robinson, K. Kolike, and H. Will (ed.), Hepadna viruses. Alan R. Liss, Inc., New York.