In vitro activity of hepatitis B virus polymerase - CiteSeerX

Download PDF
2 downloads 0 Views 540KB Size Report
Comment
Present address: MorphoSys GmbH, Am Klopferspitz 19, 82152. Martinsried, Germany. ‡ Present address: Rhone-Poulenc Rorer, Dagenham, Essex RM10.
Journal of General Virology (1998), 79, 1121–1131. Printed in Great Britain ...................................................................................................................................................................................................................................................................................

In vitro activity of hepatitis B virus polymerase : requirement for distinct metal ions and the viral epsilon stem–loop Margit Urban,† David J. McMillan, Gail Canning, Anne Newell,‡ Emma Brown, John S. Mills§ and Ray Jupp Roche Discovery Welwyn, 40 Broadwater Road, Welwyn Garden City, Herts AL7 3AY, UK

Hepadnaviruses have a complex replication cycle which includes reverse transcription of the pregenomic RNA. The initial step in this process in hepatitis B virus (HBV) requires the viral polymerase to engage a highly stable region of secondary structure within the pregenomic RNA termed the epsilon stem–loop. While reverse transcriptases belonging to the retrovirus family use a specific cellular tRNA as primer, HBV polymerase utilizes a tyrosine residue located within its own N terminus. Therefore, the first deoxyribonucleotide is covalently coupled to HBV polymerase prior to extension of the DNA strand by conventional reverse transcription. We have expressed HBV polymerase in a baculovirus and following purification have found it to be active with respect to protein-priming

Introduction Infection with hepatitis B virus (HBV) affects an estimated 5 % of the world population and is thus a medically significant problem. Primary infection frequently results in acute disease characterized by hepatocellular injury and inflammation. In about 5–10 % of adult patients and up to 90 % of young children, the immune system is not able to clear the virus and the infection progresses into chronic hepatitis, with high risk Author for correspondence : Ray Jupp. Present address : RhonePoulenc Rorer, Dagenham, Essex RM10 7XS, UK. Fax ­44 181 919 2005. e-mail ray.jupp!rp-rorer.co.uk † Present address : MorphoSys GmbH, Am Klopferspitz 19, 82152 Martinsried, Germany. ‡ Present address : Rhone-Poulenc Rorer, Dagenham, Essex RM10 7XS, UK. § Present address : Celltech Therapeutics Ltd, 216 Bath Road, Slough SL1 4EN, UK.

0001-5304 # 1998 SGM

and reverse transcription of copurified RNA. Importantly, we found both of these processes to be critically dependent on the presence of the epsilon stem–loop. The metal ion preferences of HBV polymerase were also investigated for both the proteinpriming and reverse transcription activities of this enzyme. Reverse transcription was dependent on magnesium, with an optimal concentration of 5 mM. However, protein-priming was strongly favoured by manganese ions and was optimal at a concentration of 1 mM. Thus, using manganese as sole source of metal ions our activity assay is restricted to the protein-priming event and will allow the search for novel antivirals specifically blocking this unique mechanism.

for liver cirrhosis and eventually liver carcinoma. The estimated number of chronic HBV carriers exceeds 300 million worldwide (for review see Hollinger, 1996). A large number of antiviral agents under evaluation as potential therapeutics for chronic hepatitis B were originally developed as therapies for other virus infections such as herpes simplex virus, cytomegalovirus and human immunodeficiency virus (HIV). Their major drawbacks are the rapid rebound of virus replication after termination of drug therapy as well as the appearance of resistant strains of virus (for review see Hoofnagle & Lau, 1997 ; Schalm et al., 1995). Interferon-α is currently the only licensed treatment for chronic hepatitis B. It has been shown to be beneficial to some patients ; however, the overall response rate is less than 40 % (Lok et al., 1994 ; Woo & Burnakis, 1997). Therefore, the development of novel antivirals specific for HBV remains an important goal. Hepadnaviruses have a complex replication cycle that includes reverse transcription of the pregenomic RNA, which is believed to start in the cytoplasm of the host cell and conclude within the viral capsid. In common with the reverse

BBCB

M. Urban and others

transcriptases of retroviruses, HBV polymerase has intrinsic RNA-dependent reverse transcriptase, DNA-dependent DNA polymerase as well as RNaseH activity. However, the mechanism of initiation of reverse transcription is unique. While HIV polymerase uses a specific cellular tRNA as primer, HBV polymerase utilizes a tyrosine residue located within its own N terminus. The covalent linking of the first deoxyribonucleotide to the tyrosine residue within polymerase polypeptide as part of the HBV replication strategy was recently reviewed (Nassal & Schaller, 1996 ; Ganem, 1996). A crucial step in this process is the specific binding of HBV polymerase to a highly stable region of secondary structure within the pregenomic RNA termed the epsilon stem–loop. Although epsilon is present within the 5«- and 3«-terminal repeats of the pregenomic RNA, the specific interaction of HBV polymerase with the 5« copy is a prerequisite for initiation of reverse transcription. A sequence in the bulge region of epsilon is used as template for the addition of the first four nucleotides of minus-strand DNA. Next, the primed polymerase complex is translocated to the 3«-terminal repeat where the short oligomer binds to a complementary sequence close to but not within the 3« copy of epsilon. Elongation of minus-strand DNA resumes in the direction of the 5« end of the pregenomic RNA. Besides its role in initiation of reverse transcription, the 5« epsilon loop serves as a signal for encapsidation. Interaction of polymerase with epsilon is essential and sufficient for the addition of core protein dimers and formation of viral capsids. This ensures that neither the polymerase nor the pregenomic RNA can be packaged in the absence of the other. Whether protein-priming takes place prior to or after completion of the viral capsid is an open question. All further steps of virus replication, however, occur separated from the cytoplasm of the host cell within the environment of the viral capsid. Analysis of HBV polymerase has been hampered for many years due to the inability to express functional enzyme in a recombinant system. Investigation of hepadnavirus replication has mostly relied on studies of the duck hepatitis B virus (DHBV) system. In contrast to the human counterpart, DHBV polymerase can be expressed in an active form either by in vitro translation or using the yeast retrotransposon Ty1 system (Tavis & Ganem, 1993 ; Wang & Seeger, 1992). In these systems, the presence of the epsilon stem–loop is essential for protein-primed polymerase activity, although the position dependency of the loop is largely absent. Recombinant DHBV polymerase can readily engage a stem–loop located at the 3« end in cis or even on a separate RNA molecule in trans (Wang & Seeger, 1992 ; Wang et al., 1994). Although numerous attempts have been made, HBV polymerase could not be synthesized in an active form in similar recombinant expression systems. Informative studies on HBV polymerase have come from transfection of cloned wild-type and mutant HBV genomes into cell lines supporting virion formation and required elaborate downstream analysis of the replication

BBCC

products (for example Hirsch et al., 1990 ; Junker-Niepmann et al., 1990 ; Rieger & Nassal, 1996). Some studies have demonstrated that active polymerase can be released from virions, albeit in small quantities (Bavand et al., 1989 ; Shin & Rho, 1995). Expression of recombinant HBV polymerase in Xenopus oocyte lysates resulted in reverse transcriptase activity, but the template and products have not been fully characterized (Seifer & Standring, 1993). Recently, Lanford et al. (1995, 1997) reported the expression of functional HBV polymerase using the baculovirus system We have also used the baculovirus system to express HBV polymerase. Following purification, HBV polymerase was found to be active with respect to protein-priming and reverse transcription of copurified RNA. The availability of recombinant HBV polymerase facilitated further characterization of the biochemical properties of this enzyme and the role in vitro of the epsilon stem–loop. In contrast to the studies of Lanford et al. the HBV polymerase activity measured in our system was critically dependent on the presence of an epsilon stem–loop, supplied either in cis at the 3« end of the polymerase mRNA or when expressed from a separate construct in trans. In addition, using manganese as sole source of metal ions our activity assay is restricted to the protein-priming event and will allow the search for novel antivirals specifically blocking this unique mechanism.

Methods + Cells. The Sf9 cell line of Spodoptera frugiperda was used for the propagation of wild-type and recombinant Autographa californica nucleopolyhedrovirus (AcNPV). Cells were grown in TC100 medium (GibcoBRL) supplemented with 10 % foetal calf serum (FCS) in spinner cultures (Summers & Smith, 1987). For the expression of recombinant proteins Tn5B 1-4 cells of Trichoplusia ni were grown in Excell 405 medium (JRH Biosciences) supplemented with 1 % FCS. The cell line was adapted for spinner culture and kindly provided by W. DePinto (Hoffmann-LaRoche., Nutley, USA). + Construction of recombinant baculoviruses. All HBV sequences were cloned from plasmid pCH3}3097 containing the full-length ayw strain of HBV (Bartenschlager et al., 1992 ; Junker et al., 1987). Nucleotide positions (nt) given refer to the nomenclature of Galibert et al. (1979), beginning at the unique EcoRI restriction site. HBV polymerase was amplified by PCR as two fragments, one encoding amino acids (aa) 2–545 (nt 2312–761) and enclosed by primers creating 5« NdeI and 3« NarI sites and the other encoding aa 546–832 (nt 762–1622) enclosed by 5« NarI and 3« SnaBI–SalI sites. Both fragments were subcloned sequentially into the bacterial expression vector pDS56}RBSII-6His (Stueber et al., 1990). The resulting plasmid, pDS56-PolWT*-6His, contained full-length polymerase sequence with a silent mutation creating the NarI site at nt 760. For plasmid pDS56-PolEWT*-6His a 213 bp fragment covering the DR1 motif and the epsilon stem–loop (nt 1819–2031) was amplified by PCR and cloned via SnaBI}SalI at the 3« end of the polymerase open reading frame. The baculovirus transfer vector pBlueBac-PolEWT*-FLAG was created as follows. The sequence encoding the N-terminal 519 aa of HBV polymerase (nt 2312–681) was amplified by PCR using pDS56-PolWT* as template, a 3« primer spanning the SpeI site at nt 681 and a 5« primer

In vitro activity of HBV polymerase designed to introduce an NcoI site and the sequence encoding the FLAG purification tag (YKDDDDK). The fragment encoding the C-terminal part of polymerase plus the epsilon stem–loop was subcloned from of pDS56-PolEWT* using SpeI}HindIII. Both fragments were cloned by three-way ligation into NcoI}HindIII-digested transfer vector pBlueBacIII (Invitrogen). Plasmid pBlueBac-PolEMut-FLAG encoding HBV polymerase with a mutation at the active site changing YMDD to YMND (aa 539–841) was constructed following the same protocol, but using a mismatched primer in the first PCR step. Plasmid pBlueBac-PolEWTFLAG represents a modification of pBlueBac-PolEWT*-FLAG where the number of nucleotides separating the polymerase ORF and epsilon stem–loop resembles the wild-type situation. The recombinant baculoviruses were produced by co-transfecting the plasmid DNA and AcNPV genomic DNA (Invitrogen) into Sf9 cells using Lipofectin reagent (GibcoBRL) according to the manufacturer’s protocol. Recombinant virus clones were identified by plaque assay in Sf9 cells and staining with XGal, and were purified by several rounds of plaque purification. For construction of the recombinant baculoviruses expressing polymerase without stem–loop (Pol-WT) and the stem–loop alone (E) the Bac-to-Bac baculovirus expression system (GibcoBRL Life Technologies) was used following the manufacturer’s protocol. For pFASTBac-PolWT-FLAG the complete NcoI–HindIII HBV fragment from pBlueBac-PolEWT-FLAG was subcloned into the modified multiple cloning site of pFASTBac1 (GibcoBRL Life Technologies) and subsequently the 740 bp SphI–HindIII stem–loop fragment was replaced by the respective 400 bp fragment from pDS56-PolWT-6His lacking additional non-coding HBV sequences. pFastBac-E was created by amplifying the stem–loop sequence from nt 1819 to 1992 by PCR and cloning it into the XbaI site of pFastBac1. All plasmids were verified by DNA sequencing. + Expression and purification of recombinant proteins. Tn5B 1-4 cells were grown to a cell density of approximately 2¬10' cells}ml and infected with the respective recombinant baculovirus at a multiplicity of 5 p.f.u. per cell at 28 °C. After 48 h incubation cells were harvested by centrifugation, washed twice with PBS and frozen at ®80 °C. The cell pellet of a 1 litre spinner culture was resuspended in a total of 40 ml lysis buffer [PBS supplemented with 10 % glycerol, 0±5 % NP40, 5 mM DTT, 50 U}ml RNasin (Promega), 1 tablet of protein inhibitor cocktail (Boehringer Mannheim)], passed through a French press (SLM}Aminco) at a pressure of 16 000 p.s.i. and clarified by immediate centrifugation at 60 000 g for 30 min at 4 °C. The clarified extract was incubated with 1 ml anti-FLAG M2 affinity resin (Kodak Scientific Imaging Systems) for 30 min at 4 °C under constant stirring. After transfer to a column the beads were washed subsequently with 5 column vols of low salt buffer (100 mM Tris–HCl pH 7±5, 30 mM NaCl, 10 % glycerol), high salt buffer (100 mM Tris–HCl pH 7±5, 1 M NaCl, 10 % glycerol) and again with low salt buffer at a flow rate of 1±0 ml}min. Bound proteins were eluted with 0±1 M glycine pH 3±0, 10 % glycerol, collected in 1 ml fractions and immediately adjusted to pH 7±5 by addition of 50 µl neutralization buffer (1 M Tris–HCl pH 8±7, 4 % Triton X-100, 0±1 M DTT, 50 U}ml RNasin). Purification was monitored by SDS–PAGE and immunoblotting using an antibody against the C-terminal 133 aa of HBV polymerase strain ayw. The fractions containing HBV polymerase were pooled and stored at –80 °C until further use. + HBV polymerase activity assay. The standard polymerase activity assay was performed with 20 µl of purified HBV polymerase corresponding to approximately 1}100 of the purified material from 1 litre of spinner culture. Unless stated otherwise, conditions were adjusted to 5 mM MgCl , 50 mM Tris–HCl pH 7±5, 0±15 µg}µl actinomycin D, # 0±1 mM unlabelled dNTPs and either 0±5 µl [α-$#P]dTTP (3000 µCi}

mmol ; Amersham) or 0±25 µl [α-$#P]dTTP plus 0±25 µl [α-$#P]dGTP at a final volume of 30 µl. After incubation for 2 h at 30 °C the reaction was stopped by addition of 4¬ SDS–PAGE disruption buffer (50 mM Tris–HCl pH 6±8, 2 % SDS, 5 % glycerol, 5 % β-mercaptoethanol, 0±1 % bromophenol blue). The samples were separated on a 10 % polyacrylamide gel, and the gels fixed in 40 % methanol, 10 % acetic acid, dried, and autoradiographed. Quantification was achieved using a PhosphoImager plus ImageQuant software from Molecular Dynamics Ltd.

Results Activity of HBV polymerase expressed in insect cells

In an attempt to establish an in vitro activity assay for HBV polymerase, the enzyme was expressed in insect cells using the baculovirus expression system. As illustrated in Fig. 1 (a), fulllength polymerase of HBV strain ayw was expressed under control of the polyhedrin promoter and with an N-terminal FLAG epitope allowing rapid purification by affinity chromatography. At the C terminus, the HBV sequence was extended to include the DR1 motif and the epsilon stem–loop structure, either at a distance resembling the wild-type situation (PolE-WT) or directly cloned downstream to the polymerase open reading frame (PolE-WT*). An additional construct carrying a mutation (YMDD to YMND) served as a negative control (PolE-Mut). This motif is part of the active site of polymerase and substitution of the aspartic acid at position 540 was shown to result in inhibition of protein-priming as well as elongation activity (Lanford et al., 1997 ; Radziwill et al., 1990). Purified polymerase was analysed in immunoblot assays using a polyclonal rabbit serum developed against the Cterminal 133 aa of HBV polymerase (Fig. 1 b). A prominent polypeptide could be detected at the expected molecular mass of 96 kDa ; smaller immunoreactive polypeptides most likely represent breakdown products. No difference in terms of expression and purification could be detected between PolEWT, PolE-WT* and PolE-Mut. On average, only 5–10 % of the total polymerase expressed in the cells was soluble under the conditions described in Methods. Approximately 1}100 of the purified material from a 1 litre spinner culture was used for each immunoblot and activity assay. As shown in Fig. 1 (b), the analysis of wild-type enzyme (PolE-WT) in our activity assay resulted in a signal corresponding in size to HBV polymerase as well as a smear to higher molecular mass products. This pattern is in agreement with a covalent linkage of the first nucleotide to polymerase in the protein-priming reaction, and extension of minus-strand DNA during reverse transcription. As expected, mutation of the active site (PolE-Mut) resulted in complete loss of the signal (Fig. 1 b). No difference between PolE-WT and PolE-WT* could be detected. Thus the distance between the 3« end of the polymerase open reading frame and the epsilon stem–loop had no apparent effect on enzyme activity. A number of control reactions were carried out to further characterize the reaction and prove specificity (Fig. 1 b).

BBCD

M. Urban and others (a)

(b)

Fig. 1. Expression of HBV polymerase in insect cells. (a) Schematic representation of HBV polymerase-expressing baculovirus constructs. The open reading frame of HBV polymerase is represented by an open box, the FLAG epitope used for affinity purification by the shaded section at the 5« end, and the DR1 motif plus epsilon by the black box plus stem–loop structure at the 3« end. Expression was under control of the polyhedrin promotor (Ph-P), shown as an open triangle. PolE-WT and PolE-WT* contain wild-type HBV polymerase sequence ; PolE-Mut carries a mutation in the active site, changing YMDD to YMND. (b) Activity of recombinant HBV polymerase. Wild-type (PolE-WT) and mutant (PolE-Mut) HBV polymerase were expressed in insect cells and purified via their FLAG epitope by immunoaffinity chromatography. The first blot (IB) shows the immunoblot analysis using a HBV polymerase-specific rabbit antiserum. The following autoradiographs represent in vitro activity assays, performed with wild-type (PolE-WT) or mutant HBV polymerase (PolE-Mut), wild-type polymerase under 20 mM EDTA (­EDTA), 100 µg/ml actinomycin D (­ActinoD), 20 µM aphidicolin (­Aphidicolin) or 1 mM PFA (­PFA), respectively, or wild-type polymerase tested after pretreatment with 6 µg/ml RNaseA for 10 min at 30 °C (­RNaseA). The assay was done with [32P]dTTP, dGTP, dATP, dCTP, 10 mM MgCl2, pH 7±5, at 30 °C for 2 h. The positions of HBV polymerase (pol) and higher molecular mass products (extension products) are indicated.

Addition of EDTA resulted in a complete loss of activity, demonstrating that both protein-priming and reverse transcription require divalent metal ions. The signal was unchanged by actinomycin D but ablated following prior digestion with RNaseA, indicating that RNA rather than DNA is utilized as template. Since no external template was added to the assay, HBV polymerase used RNA copurified from the insect cell. Inclusion of aphidicolin, an inhibitor of polymerase α, in the reaction buffer had no effect, demonstrating that the signal was not due to contamination with baculoviral DNA polymerases. The antiviral drug foscarnet (PFA) resulted in loss of the smear to higher molecular mass products but had no effect on the initial priming event. This is in agreement with previous reports that PFA inhibits elongation but not the protein-priming reaction of HBV polymerase (Mason et al., 1987 ; Wang & Seeger, 1992). In general, we noticed a slight batch-to-batch variation in the labelling of polymerase (see Fig.

BBCE

1 b). This difference between batches is most likely due to a variation of the ratio of free enzyme to enzyme with attached RNA or minus-strand DNA. In summary, this in vitro system demonstrated that HBV polymerase expressed in insect cells showed specific protein-priming as well as reverse transcription activity using copurified RNA as template. Optimal conditions for polymerase activity

Various parameters of this in vitro assay were analysed for their effect on HBV polymerase. First, polymerase activity was measured between pH 6±0 and 9±5 at 30 °C for 2 h. Polymerase activity was optimal at pH 7±5, with over half maximal activity observed from pH 6±5 to 8±7 (Fig. 2 a). The effect of temperature on polymerase activity was also examined and maximal activity was obtained at 30 °C (data not shown). To further characterize the enzyme, polymerase activity was measured over a time course from 1 min to 3 h (Fig. 2 a,b). At longer

In vitro activity of HBV polymerase

(a)

(b)

Fig. 2. Characteristics of in vitro HBV polymerase activity. (a) Time course and pH profile of recombinant HBV polymerase. The in vitro assay was done with wild-type HBV polymerase (PolE-WT) and [32P]dTTP, dGTP, dATP, dCTP, 10 mM MgCl2, at 30 °C. For the time course reactions were carried out at pH 7±5 for 0–180 min ; for the pH profile reaction mixtures were incubated for 2 h at an increasing pH of 6±0–9±5. Filled bars represent total polymerase activity (labelled pol plus extension products) ; open bars represent protein-primed polymerase (labelled pol only). Quantification was done using a PhosphoImager plus ImageQuant software. The maximal value obtained in each assay was considered 100 %. (b) Elongation of copurified minusstrand DNA. The autoradiograph of the time course described under (a) is shown (time course). At time-points 5, 15, 60 and 180 min the assay was repeated with inhibition of strand elongation by substitution of dATP and dCTP with ddATP and ddCTP (­ddAC), or by addition of 1 mM PFA (­PFA). The positions of HBV polymerase (pol) and higher molecular mass products (extension products) are indicated.

reaction times, a progressive increase in the signal and a shift to higher molecular mass products was observed. After 2 h maximal activity was achieved, possibly because the polymerase had become inactive or because all of the template RNA was exhausted. Interestingly, even after 1 min higher molecular mass products were detected, most likely representing minusstrand DNAs whose synthesis had already been initiated within the insect cell and had been copurified due to their covalent linkage to polymerase (Fig. 2 b). To test this hypothesis, dATP and dCTP were substituted with their respective didesoxyribonucleotides to limit the addition of labelled nucleotides to the extension products and the time course was repeated. Under these conditions, polymerase as well as higher molecular mass products were labelled over time with increasing signal strength but constant molecular mass profile, supporting our hypothesis (Fig. 2 b). In contrast, if the assay was performed using PFA, the signal was restricted to the lowest molecular mass fragment detected, in accordance with PFA inhibiting the elongation activity of HBV polymerase

but not the protein-priming activity (Fig. 2 b). Therefore, separation of the protein-priming activity from reverse transcription could be achieved by addition of the pyrophosphate analogue PFA. Coupling of deoxynucleotide to polymerase was stronger with PFA than without the drug, possibly due to its enhancing effect on the protein-priming activity of HBV polymerase. In summary, maximal polymerase activity was normally seen after 2 h incubation at a temperature of 30 °C and a pH of 7±5. Influence of metal ions on HBV polymerase activity

Reverse transcriptase enzymes normally require divalent metal ions as co-factors for polymerase activity. Since the protein-priming and elongation functions of HBV polymerase may have distinct requirements for divalent metal ions, we wanted to measure the effect of various divalent metal ions on these activities. The effect of Mg#+ concentration on polymerase activity was tested under standard assay conditions with all four dNTPs present (extension assay) or with dTTP as

BBCF

M. Urban and others (a) (b)

Fig. 3. Influence of metal ions on in vitro HBV polymerase activity. (a) Effect of Mg2+ (upper autoradiographs) and Mn2+ (lower autoradiographs) concentration on activity of recombinant HBV polymerase. The assay was done with wild-type polymerase (PolE-WT) either with [32P]dTTP only (initiation assay) or with [32P]dTTP, dGTP, dATP, dCTP (extension assay) at pH 7±5, 30 °C for 2 h. Mg2+ was tested in the range 0±1–50±0 mM ; Mn2+ was tested in the range 0±01–5±0 mM. The first lane in each autoradiograph shows the reaction without metal ions. The position of polymerase is marked by an arrow. (b) Line graph of the in vitro HBV polymerase activity assay shown in (a). Total polymerase activity is plotted against the concentration of either Mg2+ (E/D) or Mn2+ (+/*). Filled symbols (E/+) represent results form the initiation assay ; open symbols (D/*) represent results from the extension assay. Quantification was done using a PhosphoImager plus ImageQuant software. The maximal value obtained was considered 100 %.

(b) (a)

Fig. 4. Competitive effect of Mn2+ and Mg2+ on in vitro HBV polymerase activity. (a) The in vitro activity assay was done with wild-type polymerase (PolE-WT) with [32P]dTTP, dGTP, dATP, dCTP, pH 7±5, at 30 °C for 2 h. The Mg2+ concentration was constant at 5 mM ; Mn2+ was increased from 0 to 0±2 mM. The positions of HBV polymerase (pol) and higher molecular mass products (extension products) are indicated. (b) Histogram of the in vitro activity assay shown in (a). Filled bars represent total polymerase activity (labelled pol plus extension products), open bars protein-primed polymerase (labelled pol only). The results in the presence of 0±2 mM Mn2+ and absence of Mg2+ are also shown in the graph, although they were omitted in (a). Quantification was done using a PhosphoImager plus ImageQuant software. The maximal value obtained was considered 100 %.

the only nucleotide (initiation assay). The assay was repeated several times and a representative result is shown in Fig. 3. A wide concentration range was accepted by polymerase with the optimum around 5 mM and at least half maximal activity

BBCG

at 1 and 20 mM. No significant difference in requirements of protein-priming and reverse transcription could be detected. We believe that the slower migrating products observed in the initiation assay are a consequence of polymerase incorporating

In vitro activity of HBV polymerase (a)

(b)

Fig. 5. Influence of the epsilon stem–loop on in vitro HBV polymerase activity. (a) Schematic representation of the respective baculovirus constructs. PoIE-WT and Pol-WT contain the HBV polymerase open reading frame with and without the stem–loop sequence at the 3« end, respectively. E carries the DR1 motif plus the epsilon stem–loop. Symbols are as indicated in Fig. 1 (a). (b) Comparison of wild-type polymerase with epsilon in cis (PoIE-WT), mutant polymerase with epsilon in cis (PolE-Mut), wildtype polymerase in the absence of epsilon (Pol-WT) and wild-type polymerase in the presence of epsilon in trans (Pol-WT­E). The first panel shows an immunoblot analysis using an HBV polymerase-specific rabbit serum. The second panel depicts an autoradiograph representing an in vitro activity assay performed with [32P]dTTP, dGTP, dATP, dCTP and 5 mM MgCl2 (extension assay). The third panel shows an activity assay with [32P]dTTP and 1 mM MnCl2 (initiation assay). Both assays were done at pH 7±5, 30 °C for 2 h. The positions of HBV polymerase (pol) and higher molecular mass products (extension products) are indicated.

labelled dTTP into an extending chain that had been initiated within the baculovirus. For some polymerases, including HIV reverse transcriptase, other divalent metal ions can substitute for magnesium or even result in enhanced activity. To test the preference of HBV polymerase the assay was repeated with Mn#+ instead of Mg#+ (Fig. 3 b). Interestingly, in the initiation assay a very strong labelling of polymerase between the broad concentration range 0±1–5±0 mM MnCl was achieved, with # the optimal activity measured at 1 mM. The signal strength in the extension assay was lower than that observed in the initiation assay. This is most likely due to polymerase priming from adjacent positions on the template, thereby incorporating unlabelled nucleotides, as observed previously (Lanford et al., 1995). In clear contrast to magnesium, no significant labelling of higher molecular mass products could be detected, apart from a slight shift of the signal at a low concentration of around 0±1 mM in the extension assay. Therefore, manganese is a very potent cofactor for the initiation of reverse

transcription (Fig. 3 b) but it is not able to support the elongation activity of HBV polymerase to a significant extent in our in vitro assay. To further investigate the molecular basis for the inability of manganese to act as a cofactor for reverse transcription, the assay was repeated at a constant concentration of 5 mM MgCl and addition of increasing amounts of MnCl . A dose# # dependent reduction of labelling of higher molecular mass products was observed, with significantly reduced incorporation at 0±2 mM Mn#+ (Fig. 4 a,b). Therefore, manganese was acting as a dominant inhibitor of reverse transcription of HBV polymerase in our in vitro assay. At the same time, the presence of magnesium ions reduced the efficiency of proteinpriming since the signals obtained in this assay were lower that those obtained using manganese ions as sole source of divalent metal ions (Fig. 4 b). A number of other divalent metal ions were tested for their capacity to act as cofactors of either protein-priming or reverse

BBCH

M. Urban and others

transcription activity of HBV polymerase. Co#+, Ca#+, Cd#+, Ni#+, Zn#+ and Fe#+ were used as sole supply of divalent metal ions at concentrations of 0±05, 0±5 and 5±0 mM (data not shown). Cobalt, calcium, nickel and zinc were not able to support polymerase activity. Cadmium acted as a weak cofactor at its optimal concentration of 5±0 mM. It resembled manganese, as only protein-priming but no strand extension could be detected. The overall signal, however, was approximately 50-fold lower. In contrast, iron was able to support the initiation as well as elongation activities of polymerase. At the optimal concentration of 0±5 mM the overall signal was about 7-fold lower compared to magnesium. In summary, the choice of metal cofactor allowed the separation of initiation and elongation activities of HBV polymerase. Manganese as sole source of divalent metal ions was the most potent cofactor for protein-priming, while reverse transcription was most efficiently supported by magnesium ions. Impact of the epsilon stem–loop on HBV polymerase activity

To investigate the importance of the epsilon RNA stem–loop for HBV polymerase activity, the incorporation of labelled nucleotides by polymerase was measured in the presence and absence of epsilon. Two further recombinant baculoviruses were constructed allowing separate expression of full-length HBV polymerase (Pol–WT) and the epsilon stem–loop (E) (Fig. 5 a). Insect cells were infected either with PolE-WT (loop in cis), Pol-WT (no loop) or coinfected with PolWT and E (loop in trans). HBV polymerase was purified and the concentration of each enzyme adjusted by immunoblot analysis (Fig. 5 b). Initially, the enzymes were tested in our extension assay supplying all four dNTPs and magnesium as the divalent metal ion. Strong activity was detected when the stem–loop was provided in cis at the 3« end of the mRNA encoding polymerase. No activity was observed in the complete absence of the stem–loop, although the presence of the stem–loop in trans resulted in a very weak signal (Fig. 5 b). Addition of in vitro-transcribed epsilon RNA to the activity assay had no effect on the result (data not shown). When the enzymes were tested in the more sensitive initiation assay with manganese ions and [$#P]dTTP as the only nucleotide, a faint signal was detected in the absence of stem–loop, while coexpression of epsilon loop in trans could restore activity to a level approximately 2- to 4-fold lower than that obtained with stem–loop in cis. In conclusion, significant polymerase activity was dependent on the presence of the epsilon stem–loop. Moreover, a preference of polymerase for the stem–loop in a cis compared to a trans location was indicated.

Discussion Investigation of hepadnavirus polymerase has been hampered for a long time by the lack of a suitable recombinant expression system for production of functional enzyme. HBV

BBCI

polymerase expressed in our baculovirus system exhibits a clear preference for reverse transcribing RNAs containing an epsilon stem–loop. Moreover, these RNA molecules are more readily reverse transcribed when epsilon is in a cis configuration compared to a trans configuration. Although the shorter RNA containing epsilon in a trans configuration was able to promote protein-priming, it may have been too short to have supported extensive strand elongation. A suboptimal ratio between polymerase and the RNA molecules containing epsilon in the trans configuration may have also contributed to this effect. Interestingly, in a culture system capable of producing HBV virions, HBV polymerase reverse transcribed RNA that contained the epsilon stem–loop in a cis configuration 10-fold more efficiently than RNAs containing epsilon in a trans configuration (Bartenschlager et al., 1990). In this respect, our expression system closely mimicked that which occurs during the normal HBV replicative cycle. Interestingly, no signal was detected if RNA containing epsilon was supplied to polymerase in the trans configuration after polymerase had been translated (data not shown). These data suggest that polymerase may require the epsilon stem–loop for correct folding or that polymerase is required to adopt a specific conformation to engage the stem–loop which is not supported in the in vitro assay. In the later case, a chaperon(s) capable of facilitating the engagement of polymerase with epsilon and present in the insect cells may have been lost during purification. In general, the stringency of hepadnaviral polymerase with respect to template seems to be dependent on the system under investigation and the requirements for in vitro systems seem to be more relaxed. However, the DHBV epsilon stem–loop was also shown to be essential and sufficient for production of DHBV polymerase with significant activity. In the absence of stem–loop, only a limited amount of polymerase activity was observed (Tavis & Ganem, 1996 ; Wang et al., 1994). In contrast, Lanford et al. (1997) found no difference in the ability of HBV polymerase expressed and purified from their baculovirus system to reverse transcribe RNAs that either contained or did not contain the epsilon stem–loop. One of the reasons for this difference in stringency could be the presence of different viral or cellular proteins interacting with the RNA–polymerase complex. Recombinant DHBV polymerase requires interaction with cellular chaperons in order to be able to bind epsilon. The current hypothesis is that a chaperon complex, including Hsp90 and p23, binds to DHBV polymerase and induces a conformational change which allows the subsequent incorporation of stem–loop RNA (Hu & Seeger, 1996). Although the expression systems used by Lanford et al. and by us are similar, there are a number of differences in experimental protocol in addition to the different strain of insect cells used to express polymerase. In this report, HBV polymerase was expressed in a cell line derived from Trichoplusia ni while Sf9 cells from Spodoptera frugiperda were used by Lanford et al. These cell lines may differ in their spectrum and balance of chaperons, thereby influencing the

In vitro activity of HBV polymerase

conformation of HBV polymerase and its specificity for viral epsilon RNA compared to cellular RNAs. Whilst the same factor(s) that facilitate polymerase activity in HepG2 cells may not be present in Trichoplusia ni cells, there may be co-factors that are able to substitute that are slightly different or reduced compared to those present in Sf9 cells. In E. coli, strain differences are frequently observed to affect the activity of recombinant eukaryotic proteins. In our hands, considerably lower yields of polymerase were obtained from Sf9 cells and this protein was of lower quality compared to that obtained from Trichoplusia ni cells ; for these reasons we have not pursued the experimental validation of this assumption (data not shown ; Ayola et al., 1993). HBV replicates in the cytoplasm of hepatocytes and if HBV reverse transcriptase was able to indiscriminately copy cellular RNAs, highly recombinogenic nucleic acids would be produced with the potential to alter cellular functions or induce cell death. Retroviruses have a similar problem and they synthesize their reverse transcriptase as an inactive precursor which is not activated until after the enzyme is packaged into the virion, effectively isolating active enzyme from cellular RNAs. In contrast, HBV circumvents this problem by using a polymerase which is only able to engage}reverse transcribe RNAs containing the viral epsilon stem–loop. Moreover, this mechanism should also increase the packaging and replication efficiency of the pregenomic RNA. At least for DHBV polymerase, this RNA–protein interaction alters the conformation of polymerase and this conformational change seems to be necessary for conversion of the enzyme into an active form (Tavis & Ganem, 1996). Interestingly, conversion to the active conformation of the enzyme was reversible following degradation of epsilon. Almost all enzymes involved in the replication of nucleic acid require divalent metal ions as cofactors for optimal activity. Able to bind to the DNA or RNA template, the nucleotide substrate or the enzyme itself, they can serve a variety of functions including allosteric control of the enzyme structure, regulation of template and substrate complexes and catalysis of the phosphoryl transfer reaction (Black & Cowan, 1995). The metal ion preferences of HBV polymerase were investigated for both the protein-priming and reverse transcription activities of this enzyme. These functions displayed a clear difference in their preference for metal ions. Reverse transcription was dependent on magnesium, with an optimal concentration of 5 mM. The only other metal ion able to support reverse transcription, albeit with reduced efficiency, was iron. In contrast, manganese ions not only failed to act as effective cofactors of reverse transcription but inhibited the magnesium-driven reaction. However, protein-priming was strongly favoured by manganese ions and was optimal at a concentration of 1 mM. Although magnesium could act as a cofactor for initiation of transcription, the efficiency was significantly lower compared to manganese. Cadmium and iron were the only other metal ions found to act as cofactors of

protein-priming. The reasons for these differences are unclear. It is conceivable that Mn#+ interacts with polymerase in such a way to make the tyrosine residue of polymerase more available for covalent linkage to the first deoxynucleotide, and thus Mn#+ is a better co-factor compared to Mg#+ in this reaction. By contrast, the conformation induced by Mn#+ is not appropriate for elongation whereas that induced by Mg#+ facilitates chain extension. However Mg#+, in gaining this ability, has suffered a reduction in its ability to prime polymerase, which might not be surprising given our hypothesis requiring two distinct conformations. HIV polymerase, like most other reverse transcriptases, displays a characteristic preference for magnesium as divalent cation for maximal reverse transcriptase activity. Interestingly, a single amino acid substitution at position 89 alters the preference from magnesium to manganese (Prasad et al., 1991). However, the metal ion preferences of HBV polymerase more closely resemble those of the polymerases from adenovirus and certain bacteriophages such as Φ29 and PRD1 (for review see Salas et al., 1991). In these systems, manganese ions are superior in stimulating initiation and inferior in supporting elongation of DNA synthesis compared to magnesium (Caldentey et al., 1992 ; Esteban et al., 1992 ; Leith et al., 1989 ; Pronk et al., 1994). The increase in protein-priming efficiency using manganese as cofactor is reported to be about 50- to 100-fold for Φ29, 6- to 10-fold for PRD1 and about 3-fold for adenovirus polymerase compared to magnesium. In the case of PRD1 and adenovirus polymerase, manganese ions completely failed to support elongation of DNA after protein-priming. As both enzymes are able to use Mn#+ as cofactor for the elongation of oligonucleotide-primed DNA, the effect is probably not due to an intrinsic block of strand elongation but rather the failure to perform the switch from protein-priming to elongation (Pronk et al., 1994). These metal ion preferences seem to be typical characteristics of polymerases able to use the hydroxyl group of a protein as a primer for DNA synthesis. In all these systems, the protein-priming step is followed by translocation of polymerase to another site. Adenovirus and the bacteriophage polymerases use a slide back mechanism for the synthesis of the first repetitive nucleotides, while HBV polymerase pauses after addition of four bases prior to strand transfer form the 5« epsilon loop to the 3« DR1 motif. Therefore, the affinity of polymerase for template has to be high for the initiation step but subsequently low to release the enzyme and allow translocation. This process most likely involves a conformational change in the structure of the enzyme with potentially manganese stabilizing the initiating complex and magnesium preferentially stabilizing the elongation complex. Since the protein-priming reaction of HBV polymerase can occur prior to encapsidation and reverse transcription subsequently, these different activities of HBV polymerase could occur in distinct environments containing a different balance of metal ions. Our system provides the opportunity to separate different

BBCJ

M. Urban and others

functions of HBV polymerase and allows their detailed investigation in an in vitro assay. Manganese acts as potent cofactor for initiation and at the same time inhibits elongation of DNA synthesis. This could provide the basis for establishing a protein-priming assay to screen for compounds which specifically block this unique mechanism. We thank Heinz Schaller (Zentrum fu$ r Molekulare Biologie, University of Heidelberg, Germany) for the kind gift of pCH3}3097 ; Wanda DePinto (Hoffmann-La-Roche, Nutley, USA) for the Tn5B 1-4 insect cell line ; Georg Schmid and Nathalie Schaub (Hoffmann-La-Roche, Basel, Switzerland) for large-scale insect cell culture, Markus Ha$ nggi (HoffmannLa-Roche, Basel, Switzerland) for the HBV polymerase-specific rabbit serum ; Anne Broadhurst (Roche Products Ltd, Welwyn Garden City, UK) for valuable advice in protein purification ; and Phil Wong Kai In (Roche Products Ltd, Welwyn Garden City, UK) for critical comments on the manuscript.

References Ayola, B., Kanda, P. & Lanford, R. E. (1993). High level expression and phosphorylation of hepatitis B virus polymerase in insect cells with recombinant baculoviruses. Virology 194, 370–373. Bartenschlager, R., Junker-Niepmann, M. & Schaller, H. (1990). The P gene product of hepatitis B virus is required as a structural component for genomic RNA encapsidation. Journal of Virology 64, 5324–5332. Bartenschlager, R., Kuhn, C. & Schaller, H. (1992). Expression of the Pprotein of the human hepatitis B virus in a vaccinia virus system and detection of the nucleocapsid-associated P-gene product by radiolabelling at newly introduced phosphorylation sites. Nucleic Acids Research 20, 195–202. Bavand, M., Feitelson, M. & Laub, O. (1989). The hepatitis B virusassociated reverse transcriptase is encoded by the viral pol gene. Journal of Virology 63, 1019–1021. Black, C. B. & Cowan, J. A. (1995). Magnesium-dependent enzymes in nucleic acid biochemistry. In The Biological Chemistry of Magnesium, pp. 139–158. Edited by J. A. Cowan. New York : VCH. Caldentey, J., Blanco, L., Savilahti, H., Bamford, D. H. & Salas, M. (1992). In vitro replication of bacteriophage PRD1 DNA. Metal

activation of protein-primed initiation and DNA elongation. Nucleic Acids Research 20, 3971–3976. Esteban, J. A., Bernad, A., Salas, M. & Blanco, L. (1992). Metal activation of synthetic and degradative activities of phi 29 DNA polymerase, a model enzyme for protein-primed DNA replication. Biochemistry 31, 350–359. Galibert, F., Mandart, E., Fitoussi, F., Tiollais, P. & Charnay, P. (1979).

Nucleotide sequence of the hepatitis B virus genome (subtype ayw) cloned in E. coli. Nature 281, 646–650. Ganem, D. (1996). Hepadnaviridae : the viruses and their replication. In Fields Virology, 3rd edn, pp. 2703–2737. Edited by B. N. Fields, D. M. Knipe & P. M. Howley. Philadelphia : Lippincott–Raven. Hirsch, R. C., Lavine, J. E., Chang, L. J., Varmus, H. E. & Ganem, D. (1990). Polymerase gene products of hepatitis B viruses are required for

genomic RNA packaging as well as for reverse transcription. Nature 344, 552–555. Hollinger, F. B. (1996). Hepatitis B virus. In Fields Virology, 3rd edn, pp. 2739–2807. Edited by B. N. Fields, D. M. Knipe & P. M. Howley. Philadelphia : Lippincott–Raven.

BBDA

Hoofnagle, J. H. & Lau, D. (1997). New therapies for chronic hepatitis B. Journal of Viral Hepatitis 4(Suppl. 1), 41–50. Hu, J. & Seeger, C. (1996). Hsp90 is required for the activity of a hepatitis B virus reverse transcriptase. Proceedings of the National Academy of Sciences, USA 93, 1060–1064. Junker, M., Galle, P. & Schaller, H. (1987). Expression and replication of the hepatitis B virus genome under foreign promoter control. Nucleic Acids Research 15, 10117–10132. Junker-Niepmann, M., Bartenschlager, R. & Schaller, H. (1990). A short cis-acting sequence is required for hepatitis B virus pregenome encapsidation and sufficient for packaging of foreign RNA. EMBO Journal 9, 3389–3396. Lanford, R. E., Notvall, L. & Beames, B. (1995). Nucleotide priming and reverse transcriptase activity of hepatitis B virus polymerase expressed in insect cells. Journal of Virology 69, 4431–4439. Lanford, R. E., Notvall, L., Lee, H. & Beames, B. (1997). Transcomplementation of nucleotide priming and reverse transcription between independently expressed TP and RT domains of the hepatitis B virus reverse transcriptase. Journal of Virology 71, 2996–3004. Leith, I. R., Hay, R. T. & Russell, W. C. (1989). Adenovirus subviral particles and cores can support limited DNA replication. Journal of General Virology 70, 3235–3248. Lok, A. S. (1994). Treatment of chronic hepatitis B. Journal of Viral Hepatitis 1, 105–124. Mason, W. S., Lien, J., Petcu, D. L., Coates, L., London, W. T., O’Connel, A., Aldrich, C. & Custer, R. P. (1987). In vivo and in vitro studies on

duck hepatitis B virus replication. In Hepadnaviruses, pp. 3–16. Edited by W. S. Robinson, K. Koike & H. Will. New York : A. R. Liss. Nassal, M. & Schaller, H. (1996). Hepatitis B virus replication – an update. Journal of Viral Hepatitis 3, 217–226. Prasad, V. R., Lowy, I., de los Santos, T., Chiang, L. & Goff, S. P. (1991). Isolation and characterization of a dideoxyguanosine tri-

phosphate-resistant mutant of human immunodeficiency virus reverse transcriptase. Proceedings of the National Academy of Sciences, USA 88, 11363–11367. Pronk, R., Van Driel, W. & Van der Vliet, P. C. (1994). Replication of adenovirus DNA in vitro is ATP-independent. FEBS Letters 337, 33–38. Radziwill, G., Tucker, W. & Schaller, H. (1990). Mutational analysis of the hepatitis B virus P gene product : domain structure and RNase H activity. Journal of Virology 64, 613–620. Rieger, A. & Nassal, M. (1996). Specific hepatitis B virus minus-strand DNA synthesis requires only the 5« encapsidation signal and the 3«proximal direct repeat DR1. Journal of Virology 70, 585–589. Salas, M. (1991). Protein-priming of DNA replication. Annual Review of Biochemistry 60, 39–71. Schalm, S. W., de Man, R. A., Heijtink, R. A. & Niesters, H. G. (1995).

New nucleoside analogues for chronic hepatitis B. Journal of Hepatology 22, 52–56. Seifer, M. & Standring, D. N. (1993). Recombinant human hepatitis B virus reverse transcriptase is active in the absence of the nucleocapsid or the viral replication origin, DR1. Journal of Virology 67, 4513–4520. Shin, H. J. & Rho, H. M. (1995). Release of the hepatitis B virusassociated DNA polymerase from the viral particle by the proteolytic cleavage. Journal of Biological Chemistry 270, 11047–11050. Stueber, D., Matile, H. & Garotta, G. (1990). System for high-level production in Escherichia coli and rapid purification of recombinant proteins : application to epitope mapping, preparation of antibodies, and structure–function analysis. In Immunological Methods, pp. 121–152. Edited by I. Lefkovits & B. Pernis. New York : Academic Press.

In vitro activity of HBV polymerase Summers, M. D. & Smith, G. E. (1987). A Manual of Methods for

Wang, G. H. & Seeger, C. (1992). The reverse transcriptase of hepatitis

Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agricultural Experiment Station Bulletin no. 1555, pp. 1–48. Tavis, J. E. & Ganem, D. (1993). Expression of functional hepatitis B virus polymerase in yeast reveals it to be the sole viral protein required for correct initiation of reverse transcription. Proceedings of the National Academy of Sciences, USA 90, 4107–4111. Tavis, J. E. & Ganem, D. (1996). Evidence for activation of the hepatitis B virus polymerase by binding of its RNA template. Journal of Virology 70, 5741–5750.

B virus acts as a protein primer for viral DNA synthesis. Cell 71, 663–670. Wang, G. H., Zoulim, F., Leber, E. H., Kitson, J. & Seeger, C. (1994).

Role of RNA in enzymatic activity of the reverse transcriptase of hepatitis B viruses. Journal of Virology 68, 8437–8442. Woo, M. H. & Burnakis, T. G. (1997). Interferon alfa in the treatment of chronic viral hepatitis B and C. Annual Pharmacotherapy 31, 330–337. Received 5 November 1997 ; Accepted 13 January 1998

BBDB