Reverse Transcriptase Activity of Hepatitis B Virus Polymerase in Eukaryotic Cell Extracts in Vitro

Reverse Transcriptase Activity of Hepatitis B Virus Polymerase in Eukaryotic Cell Extracts In Vitro

Daniel Favre Helpatitis, Lausanne, Switzerland

Summary Zusammenfassung: In vitro reverse Transkriptase Aktivität der In hepadnaviruses, reverse transcription is primed by the viral Hepatitis B Virus Polymerase in eukariotischen Zellextrakten reverse transcriptase (RT) and requires the specific interaction In Hepadnaviren wird die reverse Transkription von der viralen between the RT and the viral RNA encapsidation signal termed reversen Transkriptase (RT) ausgeführt. Hier findet eine spezifi- ε. To study the activity of the RT in vitro, the current procedure sche Interaktion zwischen der RT und dem viralen Verpackungs- uses in vitro translated duck hepatitis B virus polymerase, but signal ε statt. Um die Aktivität der RT in vitro zu untersuchen, not the hepatitis B virus polymerase itself, in the rabbit reticulo- wird im aktuellen Protokoll in vitro translatierte Entenhepatitis cyte lysate expression system. B Virus Polymerase, aber nicht die Hepatitis B Virus Polyme- Here, the hepatitis B virus (HBV) polymerase has been suc- rase selbst, in einem Kaninchen Retikulozytenlysat Expressions- cessfully expressed in a translational extract that was obtained system eingesetzt. from monolayer human hepatocyte cells HuH-7. The translated In dieser Arbeit wurde die Hepatitis B Virus (HBV) Polymerase polypeptide retained the RNA-directed polymerase (reverse erfolgreich in einem translationalen Extrakt aus einer einlagigen transcriptase) activity on the viral RNA template containing the Kultur der humanen Hepatozytenzellen HuH-7 exprimiert. Das ε signal. We suggest that the reverse transcription event of the translatierte Polypeptid behielt die RNA-gerichtete Polymerase viral RNA coding for the polymerase and containing an ε struc- (reverse Transkriptase) Aktivität auf dem viralen RNA Template ture is concomitant to the translation of the viral polymerase to mit einem ε Signal. Wir schlagen vor, dass die reverse Transkrip- the messenger RNA. In contrast to the duck polymerase, only a tion der viralen RNA, welche die Polymerase kodiert und eine fraction of the reverse transcribed complementary DNA (cDNA) ε Struktur besitzt, gleichzeitig mit der Translation der viralen was covalently bound to the HBV polymerase in this system. Polymerase in die Boten-RNA (mRNA) abläuft. Im Gegensatz When the ε signal was missing on the mRNA, the translated zur Entenpolymerase war nur ein geringer Anteil der revers full-length HBV polymerase could not reverse transcribe the vi- transkribierten komplementär DNA (cDNA) in diesem System ral RNA template. A truncated HBV polymerase that was lack- kovalent an die HBV Polymerase gebunden. Wenn die mRNA ing the YMDD catalytic active site for the initiation of reverse kein ε Signal enthielt, konnte die translatierte HBV Polymera- transcription was unable to reverse transcribe the viral mRNA se das virale RNA Template nicht revers transkribieren. Eine template containing the ε signal. The reverse transcription ac- verkürzte HBV Polymerase, welcher die katalytische Aktiv- tivity could also be partially inhibited by employing nucleoside domäne YMDD zur Initiation der reversen Transkription analogues, such as 2’-3’-dideoxy-3’-thiacytidine (3TC; lamivu- fehlte, konnte das virale mRNA Template trotz ε Signal nicht dine) in the expression system. revers transkribieren. Die reverse Transkriptionsaktivität The procedure described here provides a method for the in vitro konnte auch partiell durch den Einsatz von Nukleosidanaloga, screening of new anti-HBV compounds directed against wild- wie 2’-3’-Dideoxi-3’-thiacytidin (3TC; Lamivudin) im Expressi- type and mutants of this crucial viral protein, the HBV polymer- onssystem inhibiert werden. ase, without the use of animals (ducks) or animal extracts (rab- Diese Arbeit beschreibt eine Methodik mit der in vitro bit reticulocyte lysate). neue anti-HBV Wirkstoffen gegen Wildtyp und Mutanten dieses zentralen viralen Proteins, der HBV Polymerase, ohne den Ein- satz von Tieren (Enten) oder tierischen Extrakten (Kaninchen Retikulozytenlysat) gesucht werden können.

Keywords: hepatitis B virus, HBV, viral polymerase, reverse transcriptase, hepatocyte, cell extract, HuH-7, translation, RNA, reverse transcription, inhibitors

Received 3rd March 2008; received in final form and accepted for publication 28th June 2008

Altex 25, 3/08 197 Fa v r e

1 Introduction polymerase is translated in vitro in a first of the RT reaction, was unable to reverse step and thereafter the reverse transcrip- transcribe the template RNA despite the Hepatitis B virus (HBV) is a major pub- tion activity is analysed in a subsequent ε signal. The reverse transcriptase activ- lic health problem with over 350 million reaction involving radiolabelled ribonucle- ity was tributary to the new translation of chronically infected people worldwide. otides. To date, all attempts at producing the HBV polymerase, since no RT activ- HBV, as a member of the hepadnavirus the biologically active HBV polymerase ity was detected in the presence of both family, is an enveloped virus with par- itself in the reticulocyte lysate system have the mRNA and an inhibitor of the initia- tially double-stranded DNA. The viral in- not been fully convincing (Jeong et al., tion of the protein synthesis. As expected, fection is associated with the development 1996; Kim and Jung, 1999). It was hypoth- the use of various dideoxynucleotides for of hepatocellular carcinoma and liver cir- esised that this might be due to the lack the inhibition of the reverse transcriptase rhosis. Although HBV is a DNA virus, the of additional protein factors or divalent activity revealed that the incorporation of mode of replication involves reverse tran- cations (Jeong et al., 1996; Li and Tyrrell, the first nucleotides was a T followed by scription of an RNA pregenome, a process 1999). Finally, several attempts have also G and A, respectively. The additional use that occurs in intracellular viral cores. been made to produce the HBV polymer- of the nucleoside analogue 2’-3’-dideoxy- The unique virus-encoded reverse tran- ase in insect cells by employing the bacu- 3’-thiacytidine triphosphate (3TC; lami- scriptase (RT) is able to initiate DNA lovirus expression system (in which only a vudine) inhibited the RT reaction by about synthesis de novo, using the RT itself as a very minor percentage of the polymerase is 40% in this system. The RT activity of protein primer (Hu and Seeger, 1996). The biologically active) (Lanford et al., 1997), the HBV polymerase was also obtained protein priming requires the specific inter- in E. coli (Jeong et al., 1996) or in yeast in the rabbit reticulocyte lysate in which action between the RT and a short RNA (Tavis and Ganem, 1993) in order to pro- the translation reaction was concomitantly signal, termed ε, located at the 5’ end of the duce large amounts for further biochemi- performed with the RT reaction; however pregenomic RNA (pgRNA) that serves as a cal and structural studies. the efficiency was lower than that obtained template for the reverse transcription (Nas- A novel and simple in vitro system for in the translational extracts prepared from sal and Rieger, 1996). For the polymerase the rapid analysis of the HBV polymerase eukaryotic cells grown as monolayers. protein of DHBV, but not that of HBV, the is thus needed. Recently, we developed The above results support the conclu- priming reaction for reverse transcription an efficient cell-free translational system sion that, in addition to the presence of could be efficiently reconstituted by in using eukaryotic cells grown as monolay- the very conserved YMDD catalytic mo- vitro translation of the protein in the rab- ers (Favre and Trépo, 2001). This expres- tif in the viral polypeptide and the pres- bit reticulocyte lysate (Wang and Seeger, sion system is particularly suited for the ence of an ε signal on the template RNA, 1992). This event required the presence translation of exogenous viral mRNAs the reverse transcriptase activity of the of a functional ε signal on the mRNA originating from various sources. By em- HBV polymerase is only detected dur- (Beck and Nassal, 1998). In the case of ploying an in vitro transcribed viral RNA ing a reaction in which the translation HBV, some reports show that the puri- coding for the full length HBV polymer- and the reverse transcriptase reactions are fied polymerase protein alone efficiently ase, which also comprised an ε structure concomitantly performed. The generation reverse transcribes the mRNA containing at the 5’ end, the HBV polypeptides with of an in vitro assay for the study of the the ε signal (Jeong et al., 1996; Lanford the expected sizes of 94 and 81 kDa were reverse transcriptase activity of the HBV et al., 1997). The catalytic site is located successfully translated in a cytoplasmic polymerase will undoubtedly allow broad in the YMDD nucleus of the polymerase extract that was obtained from the hepa- in vitro screening for new antiviral mol- (Jeong et al., 1996) and is susceptible to tocyte cell line HuH-7. When the trans- ecules directed against this important pro- mutation during antiviral therapy. In in- lation of the HBV polymerase was con- tein. Moreover, it might be employed for fected patients, the HBV polymerase is comitantly performed with the reverse the ad hoc in vitro screening of antiviral expressed by internal initiation and acts transcription reaction, the incorporation molecules directed against mutants of the preferentially in cis on the pgRNA (Chang of the radiolabelled nucleotide into the HBV polymerase that do not or poorly re- et al., 1989). The viral polymerase exhibits nascent DNA molecule was shown to be spond to the treatment with 3TC or other RNA dependent-DNA polymerase, DNA a fast and efficient process. Surprisingly, inhibitory molecules. This represents dependent-DNA polymerase, and RNase H only a small percentage of this DNA was another novel approach to the putative activity (Roychoudhury et al., 1991). Be- covalently bound to the newly synthesised prevention of the development of liver cause of problems in obtaining sufficient HBV polymerase. This newly synthesised cancer that can thus reduce or replace the amounts of purified RT proteins in vitro, it reverse transcript was indeed HBV DNA, use of ducks, ducklings and rabbit reticu- has been difficult to study the hepadnavi- as revealed by Southern blotting analysis. locyte lysate. rus functions by biochemical and structur- The activity was dependent on the pres- al analyses. The most useful system to date ence of an ε signal on the viral RNA, since 2 Materials and Methods has been a cell-free translation system, the the absence of ε did not allow full length rabbit reticulocyte lysate, developed for HBV polymerase to reverse transcribe Plasmids the study of the cognate duck hepatitis B the viral RNA. Moreover, a truncated Plasmid HH3, used for in vitro expres- virus (DHBV) reverse transcription activ- HBV polymerase lacking the catalytic site sion of full-length HBV reverse tran- ity (Wang, 1992). In this system, the viral YMDD, which is required for the onset scriptase, was constructed by cloning the

198 Altex 25, 3/08 Fa v r e

HBV polymerase coding sequence from the cells were preincubated for 30 to 45 exogenous mRNA (final concentration: the ayw strain into plasmid pSP64 that min at 37°C under 5% CO2 with DMEM 5 to 10 µg per ml) were added. Transla- was digested with restriction endonucle- lacking methionine (Sigma #21013 sup- tion reactions were carried out at 30°C for ase SacI and filled with Klenow enzyme. plemented with cysteine). The cells were 60 min. As controls, translation reactions Plasmid HH3 contains the SP6 promoter then washed with washing buffer (20 using rabbit reticulocyte lysate were per- cassette upstream of the HBV polymer- mM Hepes [pH 7.4], 33 mM NH4Cl, 7 formed according to the manufacturer’s ase gene (Fig. 1). mM KCl, 150 mM sucrose) and there- instructions (Promega). after the cytoplasmic membranes of the In vitro transcription cells were lysed for 90 s by using 100 µg In vitro reverse transcription assay To synthesise the RNA message used for lysolecithin, palmitoyl (Avanti Polar Li- Reverse transcription assays were per- in vitro translation of the HBV polymer- pids; stock at 10 mg/ml in chloroform/ formed as described above for the in vit- ase gene, the plasmid HH3 was linearised methanol (1:1) at -20ºC) per ml in washing ro translation in the presence or absence with the restriction endonuclease PvuII, buffer, on ice. Following the complete re- of exogenous mRNA with the addition which cuts outside of the viral sequenc- moval of lysolecithin from the Petri dish, of [α-32P]radiolabelled deoxyribonu- es and leaves the ε signal downstream the cells were scraped into extraction cleotide triphosphate (3,000 Ci/mM, of the intact polymerase open reading buffer containing 100 mM Hepes-KOH 0.2 µM), and with 40 µM non-radiola- frame (ORF), or with FspI, which cuts [pH 7.4], 120 mM potassium acetate belled methionine instead of [35S]me- right at the stop codon of the polymer- [pH 7.4], 2.5 mM magnesium acetate, thionine, referred to here as a coupled ase ORF and thus no ε is present on the 1 mM dithiotreitol, 2.5 mM ATP, 1 mM translation/reverse transcription reaction. transcript. The in vitro transcription reac- GTP, 100 µM S-adenosyl-methionine, Coupled translation/reverse transcription tion of capped mRNAs was performed 1 mM spermidine, 20 mM creatine phos- reactions were carried out at 30°C for as described elsewhere (Svitkin et al., phate, 100 mM sucrose, 40 µM hemin 60 min. Where indicated, the reactions 1994). Following the transcription reac- and 40 µM of each essential amino acid were performed in the presence of TMN tion, plasmid DNA was removed with except methionine (Promega). The cells buffer (100 mM Tris [pH7.4], 20 mM DNase I and RNA was purified by phe- were then passed ten times through a MgCl2, 30 mM NaCl), as described else- nol and chloroform extraction followed 25-gauge needle, and the lysate was cen- where (Seigneres et al., 2001). by passage into a Sephadex G-50 column trifuged 5 min later at 4°C and 800g for and precipitation, as described elsewhere 2 min. A Petri dish of 10 cm-in-diameter DNA extraction, electrophoresis and (Svitkin et al., 1994). For controls, the provides 0.2 ml of translational extract Southern blot analysis mRNA coding for the DHBV polymerase obtained from about 107 cells. The ex- Where indicated, the DNA products was transcribed from linearised plasmid tracts were then frozen at -70°C until fur- from in vitro polymerase reactions were pHP containing the DHBV polymerase ther use. Where indicated, the translation digested with proteinase K (1 mg/ml) in gene under the control of the SP6 pro- extracts were treated with micrococcal TES buffer (10 mM Tris [pH 7.8], 5mM moter, as described elsewhere (Wang nuclease to hydrolyse the endogenous EDTA, 0.5% SDS) for 2 h at 56°C. DNA and Seeger, 1992). This plasmid was lin- mRNAs prior to translation, as described was extracted first with phenol and then earised with restriction enzyme SalI. The elsewhere (Favre and Trépo, 2001). with chloroform. Then it was ethanol resulting transcribed mRNA codes for precipitated in the presence of glycogen full-length DHBV polymerase protein In vitro translation (2 µg) and resuspended in water. DNA and contains an ε signal at the 3’-end. [35S]-labelled HBV polymerase was products were separated on 1% agarose translated in vitro by employing eu- gels and transferred to a nylon membrane Cell culture and generation of transla- karyotic cytoplasmic extracts that were (Hybond N+, Amersham). The mem- tional extracts obtained from cells grown as monolay- brane was hybridised with a genomic The HuH-7 hepatocyte cell line (Nakaba- ers, as described elsewhere (Favre et al., HBV DNA radiolabelled probe (Favre et yashi et al., 1982) was grown in Eagle’s 2001) with a minor modification: be- al., 2003). Acrylamide gel electrophore- Modified Essential Medium supple- fore the translational reaction, one µl of sis was then performed. mented with 10% foetal calf serum, 1% 40 mg/ml reactivated, biologically active sodium pyruvate, 2 mM L-glutamine and creatine kinase in 50% (vol/vol) glycerol SDS-PAGE antibiotics (100 U/ml penicillin, 100 mi- stored at –20°C (Favre and Muellhaupt, Reactions were disrupted in electro- crograms/ml streptomycin). The chicken 2005) was added to 200 µl of freshly phoresis sample buffer containing 2% hepatoma (LMH) and the baby hamster thawed translational extract. This also sodium dodecyl sulphate (SDS) and kidney (BHK) cell lines were grown as allowed the optimal in vitro regeneration 2% 2-mercaptoethanol and were heated described elsewhere; the translational of the energy regenerating system based to 100°C for 5 min. Proteins were sepa- extract was prepared essentially as de- on ATP and creatine phosphate (Favre rated by 10% SDS-polyacrylamide gel scribed (Favre and Trépo, 2001). In order and Trépo, 2001). In vitro translation in electrophoresis (SDS-PAGE). Gels con- to allow efficient incorporation of radi- 20 µl was carried out by mixing 15 µl taining both the stacking and the resolv- olabelled amino acids such as methionine of the extract in which [35S]methionine ing portions of the gel were fixed with in the in vitro translated polypeptides, (translation grade; >1,000 Ci/mmol) and 30% methanol plus 10% acetic acid,

Altex 25, 3/08 199 Fa v r e

Fig. 1: HBV polymerase expression plasmid HH3 The phage SP6 promoter employed for the in vitro transcription of the HBV polymerase gene is located on a plasmid linearised with restriction enzyme Pvu II. Restriction enzyme Fsp I generates a linearised plasmid allowing transcription of mRNA lacking the ε signal required for the reverse transcriptase activity of the polymerase polypeptide. Due to the cloning strategy, the HBV polymerase coding sequence begins with ATG followed by 6 codons for histidine and the CCC codon coding for Pro 2 in the HBV polymerase. The position of Tyr 63 and the YMDD catalytic site are indicated. washed in water, dried, and autoradio- as suggested elsewhere (Li and Tyrrell, rapidly migrating codon is also attributed graphed (15 min to 2 h exposure on Ko- 1999). No polypeptide was translated to initiation at a second AUG codon lo- dak X-ray film). without the addition of the exogenous cated 50 amino acids downstream of the HBV polymerase mRNA in the extracts first AUG codon, as suggested elsewhere that were previously treated with micro- (Wang and Seeger, 1992). 3 Results coccal nuclease (Fig. 2, lanes 3 and 8). As a control, the mRNA coding for the duck 3.2 Expression and characteri- 3.1 Hepatitis B virus polymerase HBV polymerase was also efficiently sation of HBV polymerase is translated in vitro translated, with a major apparent molecu- expressed in an in vitro in eukaryotic cell extracts lar weight of 86 kDa (Fig. 2, lanes 7 and coupled translation-reverse In an attempt to establish an in vitro assay 10), which is in good agreement with the transcription system for HBV polymerase, the enzyme was ex- expected size of 90.5 kDa calculated from The initial aim was to develop an in vitro pressed in translational extracts obtained the predicted amino acid product of 836 system originating from eukaryotic cells, from eukaryotic cells grown as monolay- amino acids. The appearance of a more but not from rabbit reticulocytes, for the ers. For this, cytoplasmic extracts were prepared from the hepatocyte cell line HuH-7 and from the baby hamster kidney cell line BHK, treated or not with micrococcal nuclease in order to hydrolyse the endogenous mRNAs, and used for the translation of mRNA coding for the HBV polymerase. As illustrated in Figure 2, the HBV polymerase is translated in both cell extracts at the expected molecular masses of 94 and 81 kDa (Fig. 2, lanes 2, 4, 6 and 9). The 94-kDa component corresponds to the full-length HBV polymerase protein. This is consistent with a predicted MW for 845 amino acids. The appearance of the second, more rapidly migrating protein product can be attributed to the initia- Fig. 2: In vitro translation in cytoplasmic extracts obtained from HuH-7 and BHK tion at a second AUG codon, which is lo- cells grown as monolayers cated 113 amino acids downstream of the Extracts were previously treated with micrococcal nuclease to hydrolyze endogenous first AUG codon of the polymerase open mRNAs (lanes 3, 4, 8-10), or were not treated (lanes 1, 2, 5-7). Translation of exogenous reading frame. A minor polypeptide of HBV polymerase mRNA (lanes 2, 4, 6 and 9). Translation of exogenous DHBV polymerase 40 kDa was also consistently expressed. mRNA (lanes 7, 10). Control reactions without exogenous mRNAs (lanes 1, 3, 5, 8). The latter polypeptide might arise from [35S]methionine-labelled polypeptides were resolved in 10% SDS-PAGE followed by the amino-terminal portion of the protein, autoradiography. Mr, relative molecular mass (kDa).

200 Altex 25, 3/08 Fa v r e

analysis of the reverse transcriptase ac- HBV polymerase mRNA and the reverse of translation, 2’-deoxythimidine, 3’-5- tivity of the HBV polymerase. The HBV transcription on the HBV mRNA have diphosphate (pTp) (Skup and Millward, polymerase was expressed in transla- to be performed concomitantly in order 1977) during the coupled translation/re- tional extracts in the presence of all 20 to generate a strong radiolabelled signal verse transcription reaction. The addition non-radioactive amino acids, and also during reverse transcription. of this translation inhibitor did not allow with the concomitant presence of a [32P]- It was then determined, whether the the generation of a [32P]-radiolabelled radiolabelled deoxynucleotide triphos- [32P]-radiolabelled signal, which was ob- signal during the concomitant transla- phate for the incorporation in the newly tained after the coupled in vitro translation/reverse transcription reaction (Fig. synthesised DNA that is generated by tion/reverse transcription reaction, was 3b, lane 3). However, in the absence of the reverse transcription on the HBV indeed due to the de novo translation of the translational inhibitor, the genera- polymerase mRNA. Since the transla- the HBV polymerase protein and was not tion of the [32P]-radiolabelled signal was tional extracts still contain endogenous due to the presence of a putative contami- obtained (Fig. 3b, lane 2). Dimethylsul- deoxynucleotides, no exogenous dNTPs nating endogenous reverse transcriptase phoxide, in which pTp was resuspended, were added in the coupled translation- activity in the HuH-7 cell extracts. To had no effect on the generation of the reverse transcription reaction. Moreover, test this hypothesis, the HBV polymer- signal when the HBV polymerase mRNA it is known that divalent cations are re- ase mRNA was translated in the presence was present (Fig. 3b, lane 4). This result quired for the HBV polymerase activity. of the specific inhibitor of the initiation thus suggested that the de novo transla- Thus, the extracts were employed without prior treatment with micrococcal nu- Fig. 3a Fig. 3b clease, in order to avoid the presence of the calcium chelating agent EGTA in the reactions. As shown in Figure 3a, a strong signal is obtained with the concomitant translation/reverse transcription of the HBV polymerase, with a [32P]-radiolabelled signal present in both the stacking and the resolving portions of the gel (Fig. 3a, lane 3). This signal was totally absent when no exogenous HBV polymerase mRNA was added to the translational reaction (Fig. 3a, lane 1). The two 81 and 94 kDa bands corresponding to the HBV polypeptides might be masked by the radioactive signal. The presence of additional ions such as Mn++ and NaCl has been shown to increase the biologi- Fig. 3: [α-32P]dNTP labelling of expressed HBV polymerase in a coupled translation/ cal activity of the DHBV polymerase reverse transcription assay (Seigneres et al., 2001). In contrast, the a) Coupled translation/reverse transcription of the HBV polymerase mRNA. HBV presence of these two ions has a minor polymerase mRNA was transcribed in vitro from linearised plasmid HH3 and thereafter inhibitory effect on the biological activ- translated in a coupled translation/reverse transcription reaction in the absence of [35S] ity of the HBV polymerase in this assay methionine, but in the presence of all 20 unlabelled amino acids. During translation, (Fig. 3a, lane 4). additional [α-32P]dGTP or [α-32P]dTTP was added for the reverse transcription of the Separate control reactions were also newly synthesised minus-stranded HBV DNA, thus generating a coupled translation/ performed with the in vitro translation reverse transcription reaction. The products of the coupled reaction are shown in lanes 3 of the HBV polymerase for 1 h at 30°C and 4, in the absence (lane 3) or presence (lane 4) of additional TMN buffer. The products performed in the presence of all 20 non- of the control reactions without polymerase expression mRNA (lanes 1 and 2), and with radiolabelled amino acids, followed in a the absence (lane 1) or presence (lane 2) of TMN buffer, are also presented. Similar subsequent reaction by the reverse tran- results were obtained with the use of the three other radiolabelled dNTPs. scription reaction itself for 1 h at 30°C, b) De novo translation of HBV polymerase protein is necessary for biological activity. with the separate addition of each of the Coupled translation/reverse transcription reactions were performed with HBV polymerase four [32P]-labelled deoxynucleotides. mRNA (lanes 2 to 4) and in the absence (lanes 2 and 4) or the presence (lane 3) of When the translation and the reverse additional 2’-deoxythimidine 3’,5’-diphosphate (pTp; 1 mM) in order to inhibit the initiation transcription were performed in two of translation. During translation, additional [α-32P]dTTP was employed for the reverse separate steps, no radiolabelled signal transcription of the newly synthesized minus-stranded HBV DNA Control reaction with was obtained after SDS-PAGE analy- DMSO that has served to suspend pTp (lane 4). Polypeptides were resolved in 10% sis (not shown). Thus, this experiment SDS-PAGE followed by autoradiography for 15 to 30 min. Both the concentrating (C) and revealed that both the translation of the resolving (R) portions of the gel were analysed. Mr, relative molecular mass (kDa).

Altex 25, 3/08 201 Fa v r e

tion of the HBV polymerase mRNA is a that the radiolabelled nucleic acid was the gel, thus probably reflecting the fact prerequisite for obtaining a specific 32[ P]- composed of a major nucleotide at least that the [32P]-radiolabelled HBV DNA radiolabelled signal, and that no contami- 1,000 bases long, as revealed by migra- was trapped in a large molecular weight nating endogenous reverse transcriptase tion in urea-acrylamide gel (Fig. 4a, complex (Fig. 4a, lane 1). activity was responsible for the radioac- lane 2; purified probe without further en- The reverse transcribed, [32P]-radi- tive signal. Finally, micrococcal nuclease zymatic treatments). The hydrolysis of olabelled HBV polymerase DNA was treatment of the cell extracts had no del- the radiolabelled DNA with DNase I for further analysed by agarose gel electro- eterious effect on the HBV polymerase 2 h completely abolished the radioactive phoresis. For this, the [32P]-radiolabelled activity (not shown). signal (Fig. 4a, lane 5). Treatment with HBV DNA was extracted or not with The purified, reverse transcribed DNA proteinase K had no effect on the appear- organic solvents after the coupled trans- was then shown to be solely composed ance of the radioactive signal (Fig. 4a, lation/reverse transcription reaction, pre- of a [32P]-radiolabelled DNA. To show lane 3). Treatment with pronase abol- cipitated and analysed by autoradiogra- this, the radiolabelled HBV nucleic acid ished the signal, thus probably reflecting phy after migration in an agarose gel and was extracted with organic solvents af- the contamination of this enzyme with transfer on a nitrocellulose membrane. ter generation in a coupled translation/ some contaminating DNase activities A major, high molecular weight, [32P]- reverse transcription reaction, precipi- (Fig. 4a, lane 4). Finally, the analysis of radiolabelled band was obtained with tated in the presence of glycogen, resus- a crude preparation of the translation/ the analysis of the crude reaction that pended in water, and subjected to vari- reverse transcription reaction revealed a was not extracted with organic solvents ous enzymatic treatments. This revealed radioactive signal that did not enter into (Fig. 4b, lane 3). After extraction of the

Fig. 4a Fig. 4b Fig. 4c

Fig. 4: Analysis of HBV reverse transcribed DNA The reverse transcribed HBV polymerase DNA was synthesized in the in vitro coupled translation/reverse transcription assay as described in the Fig. 3a, by using [α-32P]dTTP as a radiolabel. a) The nucleic acids were extracted with organic solvents (phenol/chloroform; ph/chl) and precipitated. Purified, radiolabelled HBV polymerase DNA was then incubated with various enzymes and finally analysed by urea-acrylamide (4%) gel electrophoresis and autoradiography. Extraction with organic solvents alone (lane 2). Extraction with organic solvents followed by treatment with proteinase K (lane 3), pronase (lane 4), DNase I (lane 5). Crude, coupled translation/reverse transcription reaction without further extraction of the radiolabelled nucleic acids (lane 1), revealing that the majority of the radioactive signal did not enter into the gel. b) After the coupled translation/reverse transcription assay using HBV polymerase mRNA, the reactions were incubated or not with proteinase K (lane 5) or DNase I (lane 6), as indicated. The nucleic acids were thereafter extracted (ph/chl +) or not (ph/chl -) with organic solvents, precipitated, migrated in 1% agarose gel, transferred to a nitrocellulose membrane, and finally autoradiographed. Control reactions were performed in the absence of exogenous HBV polymerase mRNA (lanes 1 and 2). In the absence of extraction with organic solvents, a substantial portion of the radioactive signal could barely enter into the agarose gel. The position of the front of migration is indicated. c) The coupled translation/reverse transcription reactions were thereafter incubated with DNase I (lane 2), RNase A (lane 3) or proteinase K (lane 4), and the proteins were analysed by 10% SDS-PAGE followed by autoradiography. Both the concentrating (C) and resolving (R) portions of the gel were analysed. Mr, relative molecular mass (kDa).

202 Altex 25, 3/08 Fa v r e

samples with organic solvents followed Fig. 5a Fig. 5b by ethanol precipitation, a major radiolabelled DNA band was obtained (Fig. 4b, lane 4). Interestingly, an additional signal of higher molecular weight appeared with treatment of the coupled translation/ reverse transcription reaction with proteinase K prior to the extraction of the nucleic acids with organic solvents (Fig. 4b, lane 5). The majority of the radioactive bands disappeared after a 2-hour DNase I treatment (Fig. 4b, lane 6). As controls, translation/reverse transcription reactions were performed in extracts that were not supplemented with exogenous HBV polymerase mRNA. Only unspe- cific bands were obtained (Fig. 4b, lanes 1 and 2). In parallel, the coupled translation/reverse transcription reactions were also Fig. 5: Analysis of the HBV polymerase DNA by Southern blot incubated with DNase I, RNase A or a) The reverse transcribed HBV polymerase DNA that is synthesized in the in vitro proteinase K after the completion of the coupled translation/reverse transcription reaction can be employed as a radiolabelled coupled translation/reverse transcription molecular probe for the detection of the HBV polymerase DNA. Reactions were performed reaction. The proteins were thereafter an- as described in Fig. 3a, followed by treatment with proteinase K. The [32P]-radiolabelled, alysed by 10% SDS-PAGE followed by reverse transcribed DNA was thereafter extracted with organic solvents, purified, and autoradiography. This revealed that the employed as a molecular probe for detection in a Southern blot analysis of the Xho I/ radioactive signal was nearly completely Nco I and the Hin dIII/Eco RI DNA fragments of the HBV polymerase digested in plasmid eliminated after hydrolysis of the nucleic HH3, as indicated. These 1,250 and 870 bp corresponding fragments, excised from the acids with DNase I, with the major ex- N-terminal and C-terminal coding sequence of the HBV polymerase, respectively, are both ception of the presence of a radiolabelled hybridised by the radiolabelled reverse transcribed, HBV polymerase DNA probe. polypeptide of around 90 kDa and vari- b) The newly reverse transcribed HBV polymerase DNA can be hybridized by a ous minor polypeptide bands ranging be- radiolabelled HBV DNA probe. The coupled translation/reverse transcription reactions tween 48 and 65 kDa (Fig. 4c, lane 2). were performed with HBV polymerase mRNA using all 20 unlabelled amino acids, and The hydrolysis of the RNAs using RNase were then treated with proteinase K (lane 5), DNase I (lane 6) or RNase A (lane 7), or not A did not reduce the intensity of the ra- treated (lanes 3 and 4). Control reactions were performed in the absence of exogenous dioactive signal (Fig. 4c, lane 3) when HBV polymerase mRNA (lanes 1 and 2). Where indicated, the nucleic acids were then compared to the control reaction per- extracted with organic solvents, precipitated, migrated in 1% agarose gel, transferred to a formed with HBV polymerase and with- nitrocellulose membrane, and finally analyzed by Southern blotting using a radiolabelled out further enzymatic treatments (Fig. 4c, HBV DNA probe followed by autoradiography. Sizes are indicated in kilobases (kb). lane 1). However, treatment with RNase A also generated minor radioactive bands for the detection of various HBV DNA blot analysis using a radiolabelled probe ranging between 48 and 65 kDa. Finally, fragments in a Southern blot analysis. As composed of the genomic HBV DNA treatment with proteinase K did not in- shown in Figure 5a, both the XhoI/NcoI was performed. In a first step, the reverse terfere with the presence of a strong ra- and the HindIII/EcoRI fragments digest- transcript was generated by coupling the dioactive signal (Fig. 4c, lane 4) when ed from plasmid HH3 were recognised translation/reverse transcription reaction compared to the reaction without further by the reverse transcribed, [32P]-radiola- in the presence of the HBV polymerase treatment (Fig. 4c, lane 1). belled probe. This revealed that the probe mRNA, but in the absence of any radi- We further tested whether the [32P]-ra- was specifically HBV polymerase DNA. olabelled deoxynucleotide. The reactions diolabelled signal obtained in the coupled The reverse transcription performed by were further treated or not with various translation/reverse transcription reaction the HBV polymerase generated a large enzymes followed by extraction of the was indeed specifically the reverse tran- DNA fragment, since the 870 bp-long nucleic acids with organic solvents and scribed HBV polymerase DNA. To test HindIII/EcoRI fragment from plasmid precipitation. The nucleic acids were then this, the [32P]-radiolabelled, reverse tran- HH3 was hybridized. separated by agarose gel electrophoresis, scribed DNA was treated with proteinase In order to confirm more precisely the transferred to a nitrocellular membrane, K, extracted with organic solvents, pre- fate of the reverse transcribed DNA af- and finally hybridized with a radiola- cipitated in the presence of glycogen and ter synthesis in the coupled translation/ belled HBV DNA probe. This revealed finally employed as a molecular probe reverse transcription reaction, a Southern that the HBV DNA probe specifically hy-

Altex 25, 3/08 203 Fa v r e

bridized to the HBV DNA sequences that reverse transcription of this mRNA re- the detection of HBV polymerase DNA were reverse transcribed in the coupled sulted in a polypeptide incompetent of fragments in a Southern blot analysis in translation/reverse transcription reaction. reverse transcription activity on the HBV a coupled translation/reverse transcription Indeed, a specific major band was recog- polymerase mRNA (Fig. 6, lane 3). As a reaction (not shown), thus suggesting that nized (Fig. 5b, lanes 3 to 7). Larger addi- positive control, the mRNA coding for the the HBV polymerase per se was unable to tional bands were also obtained (Fig. 5b, HBV polymerase and containing an intact reverse transcribe either an RNA lacking lanes 4, 5 and 7). Importantly, proteinase ε signal was translated into a polypeptide an ε signal or a putatively contaminating K treatment prior to the extraction of the providing reverse transcription activity on plasmid DNA in the coupled translation/ nucleic acids revealed the presence of an the HBV polymerase mRNA (Fig. 6, lane reverse transcription reaction. additional higher band, thus revealing 2). Thus, the absolute dependence on an ε On an other hand, the presence of the the reverse transcript covalently bound signal on the mRNA was proven for the catalytic YMDD motif in the viral RNA- to the HBV polymerase (Fig. 5b, lane reverse transcription activity. Moreover, dependent DNA polymerase is a prerequi- 5). Moreover, the treatment (15 min) of this result confirmed that the biologically site for its activity (Poch et al., 1989). We the nucleic acids with DNase I led to the active, full-length HBV polymerase was therefore tested, whether an YMDD motif extinction of the larger nucleotide bands not able to reverse transcribe putative was necessary for the reverse transcription and to the diminution of the intensity of contaminating HH3 plasmid DNA in the activity in the HBV polymerase. To test the lower signals (Fig. 5b, lane 6), thus RT reaction. Indeed, HBV polymerase this hypothesis, an HBV coding sequence probably reflecting the fact that a long- mRNA lacking an ε signal was unable containing a point mutation at amino er incubation period in the presence of to synthesise a radiolabelled molecular acid position 152 (new stop codon TAG DNase I might be required for the total DNA probe that could be employed for instead of TGG) was generated. After extinction of this signal. The hydrolysis linearisation of the plasmid with restric- of the nucleic acids with RNase A did tion endonuclease Pvu II (thus maintain- not affect the intensity of the radioactive ing the ε signal on the mRNA), in vitro signal (Fig. 5b, lane 7). Analysis of the transcription followed by in vitro transla- crude reverse transcribed HBV polymer- tion, a truncated HBV polymerase protein ase DNA revealed a signal that did not of about 18 kDa was synthesised (not enter into the agarose gel, but remained in shown). When the mRNA containing the the well (Fig. 5b, lane 3). This can be ex- point mutation was employed in the cou- plained by the fact that the newly reverse pled translation/reverse transcription re- transcribed DNA might be surrounded action, no [32P]-radiolabelled polypeptide by proteins. As an internal control, the was obtained (not shown), even thought endogenous nucleic acids present in ex- the mRNA contained an ε structure at the tracts obtained from HuH-7 cells were 3’ end. This suggested that the ε structure analysed after extraction or not with or- alone did not confer reverse transcription ganic solvents followed by precipitation. activity to the HBV polymerase in the ab- This revealed that the HBV DNA probe sence of the catalytic YMDD motif. did not hybridise to contaminating HBV DNA sequences in this cellular extract 3.4 HBV polymerase activity can (Fig. 5b, lanes 1 and 2). be obtained in cell extracts other than hepatocytes 3.3 The HBV polymerase Fig. 6: Absence of reverse transcription Translational extracts were also prepared requires an ε structure on the activity of the HBV polymerase on HBV from other eukaryotic cell lines growing mRNA and a catalytic YMDD polymerase mRNA lacking an ε signal as monolayers, such as LMH and baby amino acid sequence for reverse Coupled translation/reverse transcription hamster kidney (BHK). The extracts were transcription activity reactions were performed as described then employed in the coupled translation/ It was crucial to determine whether the ε in Fig. 3a, in the presence of an mRNA reverse transcription assay by using the signal was necessary for the reverse tran- encoding full length HBV polymerase but mRNA coding for the HBV polymer- scription activity of the HBV polymerase lacking the ε signal at the 3’-end (lane 3). ase. This revealed that in addition to the in the coupled translation/reverse tran- Control reaction with mRNA containing HuH-7 cell extracts (Fig. 2a), the use of scription assay. To test this hypothesis, the ε signal at the 3’-end (lane 2). both LMH and BHK cell extracts for the plasmid HH3 was linerarised with restric- Control reaction without exogenous HBV coupled translation/reverse transcription tion enzyme Fsp I. The in vitro transcrip- polymerase mRNA (lane 1). Proteins were reaction of the HBV polymerase could tion of the linearised plasmid generated an analyzed by 10% SDS-PAGE followed by generate a specific radioactive signal mRNA without the ε signal at the 3’ end. autoradiography. Both the concentrating (data not shown). The in vitro translation of this transcript (C) and resolving (R) portions of the gel We then tested whether the reticulo- generated a full length HBV polymerase were analyzed. Mr, relative molecular cyte lysate expression system could be (not shown). However, the translation/ mass (kDa). employed for the study of the biological

204 Altex 25, 3/08 Fa v r e

activity of the human HBV polymerase. polymerase mRNA generated a single ra- analogues on the HBV polymerase ac- For this, the coupled translation/reverse diolabelled polypeptide band of about 90 tivity in vitro. For this, the extracts were transcription assay was performed in kDa (Fig. 7, lane 2). This suggested that employed in coupled translation/reverse lysate supplemented with all 20 amino only the priming reaction was obtained transcription reactions using the HBV acids and with mRNAs coding for either with DHBV mRNA in a coupled trans- polymerase mRNA in the presence or ab- the DHBV or HBV polymerases. Both lation/reverse transcription reaction. The sence of the various inhibitors. The level polymerases were efficiently translated addition of TMN buffer in the coupled of inhibition of [32P]-labelled DNA syn- with their expected molecular masses in translation/reverse transcription reaction thesis was assessed by quantifying the this system, as seen with the [35S]methio- abolished the generation of radioactive incorporation of [α-32P]dCTP into the nine incorporation into the newly synthe- signals with both mRNAs (Fig. 7, lanes nascent viral DNA. Although no exog- sised polypeptides (data not shown). The 3 and 6). enous dNTPs were added in the reactions synthesised HBV polymerase generated a and since there is a pool of endogenous smeary radioactive signal during the cou- 3.5 HBV polymerase activity dNTP-TPs in the HuH-7 extracts (as pled translation/reverse transcription on can be inhibited in vitro with was shown elsewhere with the use of the the HBV polymerase mRNA (Fig. 7, lane dideoxynucleotides triphosphate DHBV RNA in the reticulocyte lysate; Le 5), however with a 90% lower efficiency and nucleoside triphosphate Guerhier et al., 2001), the results showed when compared to the activity obtained analogues that the incorporation into the nascent vi- in the HuH-7 cell extract (compare with Using the HuH-7 cell extract, we ana- ral DNA was significantly inhibited with Fig. 3a). Interestingly, the coupled trans- lysed the inhibitory effect of dideoxy- the presence of ddNTPs triphosphate lation/reverse transcription of the DHBV nucleotides and nucleoside triphosphate (Fig. 8). The most potent inhibition was obtained with the use of ddTTP-TP (Fig. 8, lane 6) with nearly 100% inhibition, when compared to the control reaction performed in the presence of the exogenous HBV polymerase mRNA but without additional ddNTP-TP (Fig. 8, lane 2) (mean of three experiments). This was followed by ddGTP-TP (Fig. 8, lane 5; 90% inhibition), ddATP-TP (Fig. 8, lane 3; 80% inhibition) and ddCTP-TP (Fig. 8, lane 4; 70% inhibition), respectively. This result was consistent with an order of nucleotide addition 5’-dT-dG-dA during the reverse transcription of the HBV polymerase. The use of 2’-3’-dideoxy-3’-thiacytidine triphosphate (3TC-TP; lamivudine Fig. 7: Analysis of the HBV polymerase Fig. 8: Inhibition of HBV polymerase triphosphate) could act as an inhibitor activity in the rabbit reticulocyte lysate activity by dideoxynucleotides of the elongation activity of the HBV system triphosphate and nucleoside polymerase in the coupled translation/ The mRNAs coding for either DHBV triphosphate analogue in vitro revere transcription reaction. To test this polymerase (lanes 2, 3) or HBV Coupled translation/reverse transcription hypothesis, increasing concentrations polymerase (lanes 5, 6) were employed reactions were performed as described of 3TC-TP were employed. The results 32 32 with the use of [α- P]dGTP or [α- P] in Fig. 3a. Coupled translation/reverse showed that 50% inhibition of dCTP- dTTP, respectively, as a radiolabel. transcription reactions were performed in TP incorporation into the nascent DNA Coupled translation/reverse transcription the presence of [α-32P]dCTP plus various was obtained with the use of 2.5 µM of reactions were also performed with combinations of unlabelled ddNTPs, as 3TC-TP (mean of three experiments; not additional TMN buffer (lanes 3 and 6). indicated in the Figure. Weak labelling was shown). Thus, the results showed that a Proteins were analysed by 10% SDS- obtained with the use of ddTTP followed chain terminator of the reverse transcrip- PAGE followed by autoradiography. Both by ddGTP, ddATP and finally ddTCP, in tion activity of the HBV polymerase the concentrating (C) and resolving (R) agreement with the synthesis of a primer could be analysed in vitro in eukaryotic portions of the gel were analysed. Mr, with an authentic sequence. Proteins were cell extracts. We thereafter tested wheth- relative molecular mass (kDa). analysed by 10% SDS-PAGE followed by er other nucleotide analogues or inhibi- autoradiography. Both the concentrating tory molecules could inhibit the activity (C) and resolving (R) portions of the gel of the HBV polymerase in vitro. This were analysed. Mr, relative molecular revealed that candidate inhibitors such mass (kDa). as PFA (foscarnet), DXG-TP or DAPD-

Altex 25, 3/08 205 Fa v r e

TP possessed 50% inhibitory concentra- Table 1: RRL (rabbit reticulocyte lysate) tions (IC ) of 0.5 mM, 0.8-2 µM and 50 Scientific article Number of animals 0.15-0.30 µM, respectively. When the conventional inhibitor 3TC-TP (lamivu- Offensperger et al., 1993 18 ducklings dine triphosphate) was employed, 50% Severini et al., 1995 One duck inhibition of reverse transcriptase activ- Rajagopalan et al., 1996 3 woodchucks ity was obtained at a concentration of Howe et al., 1996 24 ducklings 2.5 µM (data not shown). These inhibito- Luscombe et al., 1996 18 ducklings ry concentrations are relevant for reduc- Zoulim et al., 1996 13 ducklings tion of hepadnaviral reverse transcription Hafkemeyer et al., 1996 11 ducklings in vitro, as compared to other similar in- Lofgren et al., 1996 20 ducks hibitory studies such as the rabbit reticu- Colledge et al., 1997 Primary duck hepatocytes locyte lysate system. Rahn et al., 1997 Primary duck hepatocytes Witcher et al., 1997 3 woodchucks Cullen et al., 1997 25 woodchucks 4 Discussion Cavanaugh et al., 1997 56 transgenic mice Aguesse-Germon et al., 1998 23 ducklings, RRL, primary duck hepatocytes 4.1 The HBV polymerase Nicoll et al., 1998 18 ducklings and the role of an Genovesi et al., 1998 23 woodchucks in vitro expression system Lin et al., 1998 18 ducks Several animal studies have been per- Seifer et al., 1998 1 woodchuck, RRL formed in the past for the screening of new antiviral molecules that inhibit the Seignères et al., 2000 29 ducklings hepadnavirus replication, either in vivo Tomita et al., 2000 30 ducks with studies involving the sacrifice of Nicoll et al., 2000 16 ducklings the experimental animals (Addison et al., Seignères et al., 2001 Primary fetal duck hepatocytes, RRL 2002; Cao and Tavis, 2006; Cavanaugh Le Guerhier et al., 2001 67 ducklings et al., 1997; Colledge et al., 1997; Cullen Urban et al., 2001 4 ducks et al., 1997; Fourel et al., 1994; Genovesi Addison et al., 2002 14 ducks et al., 2000; Hafkemeyer et al., 1996; Seignères et al., 2003 141 ducks, primary fetal duck hepatocytes, RRL Howe et al., 1996; Lin et al., 1998; Lof- Zhang et al., 2004 48 ducklings gren et al., 1996; Luscombe et al., 1996; Hu et al., 2004 Nicoll et al., 2000; Nicoll et al., 1998; Robaczewska et al., 2005 RRL, primary duck hepatocytes Offensperger et al., 1993a; Rahn et al., Wang et al., 2005 16 ducklings 1997; Rajagopalan et al., 1996; Seigneres Jacquard et al., 2006 RRL et al., 2000; Severini et al., 1995; Tomita Guo et al., 2007 < 36 ducks ; recombinant mice (n= ?) et al., 2000; Urban et al., 2001; Witcher et al., 1997) or in vivo with additional in vitro studies using the rabbit reticulo- a biologically active HBV polymerase 4.2 General considerations cyte lysate system (Aguesse-Germon et in vitro, it might be envisioned that the on the synthesis of hepatitis al., 1998; Doong et al., 1991; Guo et al., molecular responses of the polymerase virus polymerases 2007; Jacquard et al., 2006; Le Guerhier to antiviral compounds can be studied Analysis of the HBV polymerase has et al., 2001; Offensperger et al., 1993b; on a broad scale, by means of reactions been hampered for many years due the Robaczewska et al., 2005; Seifer et al., on biochips, for example. Hundreds or inability to efficiently express this func- 1998; Seigneres et al., 2001; Seigneres thousands of putative inhibitory mol- tional enzyme in a recombinant system in et al., 2003; Shaw et al., 1996; Zoulim et ecules can be selected for their specific vitro. For this reason, the investigations al., 1996) (Tab. 1). activity in inhibiting the initiation and the have mostly relied on studies of the duck Reliable ways and means can be em- elongation of the HBV polymerase. This hepatitis B virus (DHBV) system. For ployed in order to investigate human will undoubtedly lead to a better com- the polymerase protein of DHBV, but not biology and health issues in vitro, for prehension of the processes involved in that of HBV, one of the systems described example in the field of viral infections. the inhibition of this medically important has shown that the priming reaction and The assessment method builds up from enzyme by reducing, refining and replac- primer elongation can be artificially re- molecules and cells to the individual via ing (3Rs) the use of animals and animal constituted by in vitro translation of the tissues and organs. It is in strong contrast extracts. protein in the rabbit reticulocyte lysate to the top-down approach of the animal (Wang and Seeger, 1992). The other sys- model, which faces at the outset the full tem packages an active fusion protein of complexity of the animal. By producing DHBV polymerase in a virus-like particle

206 Altex 25, 3/08 Fa v r e

from the yeast retrotransposon Ty1 (Tavis tain free deoxynucleotides (Gaillard et cells, it was not feasible to hydrolyse the and Ganem, 1993). These two systems al., 2002). Since the reverse transcription RNA template, because this would have yield a polymerase that possesses accu- reaction of the HBV polymerase occurs led to the absence of HBV polymerase rate protein-primed reverse transcriptase in the cytoplasm of infected cells, we expression. Anyway, the presence of the activity that synthesises minus-strand considered that the generation of trans- HBV polymerase mRNA in a coupled DNA originating at ε and DR1 (Tavis lational extracts without further equili- translation/reverse transcription reaction et al., 1994). The polymerase mRNA bration and complementation, and also performed in the presence of the specific employed contained a 3’ copy of ε, but without the hydrolysis of the endogenous inhibitor of the initiation of protein syn- no 5’ copy of this sequence. Recently, a mRNAs, was providing an expression thesis pTp, resulted in the total lack of a third polymerase expression system that system that would approach the natural radiolabelled signal, thus confirming that utilises purified human HBV polymerase cellular conditions present in eukaryotic no putative unexpected endogenous re- that was expressed via the baculovirus- cells very closely. verse transcriptase activity was present in insect cell expression system has been We have shown that both the HBV the reactions. Thus, the translation of the described. The purified HBV polymerase and DHBV polymerases were efficiently HBV polymerase mRNA was a prereq- was active for in vitro priming and re- translated in the cytoplasmic extracts that uisite for the generation of a radioactive verse transcriptase reactions. Some other were obtained from eukaryotic cells, as signal (Fig. 3b). reports claimed that the human HBV revealed with the [35S]methionine label- Second, the reverse transcriptase activ- polymerase was active in the rabbit re- ling of the polypeptides (Fig. 2). For the ity generated a large DNA transcript that ticulocyte lysate expression system (Kim HBV polymerase, three major polypep- was sensitive to DNase I treatment (Fig. and Jung, 1999) or in E. coli (Jeong et tides of 94, 81 and 40 kDa were syn- 4a). The reverse transcript was covalently al., 1996). In this respect, a major break- thesised. These results were identical to bound to protein, as revealed by South- through was performed with the coupling those obtained with the translation of the ern blot analysis of the nucleic acids af- of the transcription and translation of HBV polymerase in the rabbit reticulo- ter the proteinase K treatment (Fig. 4b). the HBV polymerase mRNA in the rab- cyte lysate (Li and Tyrrell, 1999). Thus, Surprisingly, it appeared that a substan- bit reticulocyte lysate expression system the expression system described here was tial portion of the reverse transcript was (Li and Tyrrell, 1999). Importantly, both a strong candidate for the in vitro genera- not covalently bound to protein. Indeed, the priming and the elongation reactions tion of an HBV polymerase that would there are multiple potential reasons why were performed in this system by the possess inherent and efficient reverse full-length HBV polymerase is not fully concomitant transcription and translation transcriptase activity. In initial trials, covalently bound to the minus strand of of the mRNA in the presence of either a the HBV polymerase was translated in the reverse transcript. This might be due protein priming buffer or a polymerisa- a first reaction involving the mRNA and to the inherent differences in the assay tion buffer, respectively, containing an all unlabelled amino acids, and thereafter systems employed for the generation of a additional radiolabelled deoxynucleotide. the reverse transcriptase assay was per- biologically active HBV polymerase, as In addition to the generation of radiola- formed in a second reaction in which ad- well as for the DHBV polymerase. belled polypeptide bands at 94, 81 and ditional radiolabelled dNTPs were add- Third, a single radioactive polypeptide 40 kDa, radioactivity appeared at the top ed. It turned out that this procedure did of 94 kDa was obtained after the diges- of the gel in both the stacking and re- not allow the generation of radiolabelled tion of the nucleic acids with DNase I, solving portions. The polymerase gener- reverse transcripts. It finally turned out thus suggesting that this polypeptide ated DNA fragments of around 200 and that the stumbling block was the step- was HBV polymerase covalently bound 400 bp, the latter seemingly representing wise procedure initially employed. Cir- to primed dTTP (Fig. 4c). A substantial a DNA product with a length twice that cumstantial evidence confirms that the amount of radioactive signal was ob- of the original RNA template (Li and reverse transcribed DNA is indeed the tained after the digestion of the proteins Tyrrell, 1999). expected HBV polymerase DNA. with proteinase K. These striking fea- First, a strong radioactive signal was tures might be explained by the fact that 4.3 Coupling translation with solely obtained under conditions in which the DNA could somehow enter into the reverse transcription in vitro the translation reaction was tightly cou- protein gel due to the net negative charge Recently, we generated an in vitro trans- pled to the reverse transcription reaction of the unbound, radiolabelled reverse lation system from eukaryotic cells that per se, as revealed by the incorporation of transcribed DNA. In this respect, we can were grown as monolayers (Favre and radiolabelled dNTPs into the newly syn- ask ourselves, whether the claims of the Trépo, 2001). Since this expression sys- thesised DNA (Fig. 3a). It has been shown presence of extension products of the tem is particularly suited for the trans- that the removal of the endogenous tem- HBV polymerase in previously published lation of viral mRNAs originating from plate from purified HBV polymerase by articles (Tavis and Ganem, 1993) were in various sources, it was thus employed nuclease treatments resulted in the loss of fact not due to the presence of DNA mol- for the generation of a biologically ac- the nucleotide priming activity. Because ecules covalently bound to the polymer- tive HBV polymerase in vitro. It has been of the inherent procedure employed for ase, but instead of free radiolabelled, re- shown that translational extracts from the expression of the HBV polymerase verse transcribed HBV DNA molecules. HepG2 and HuH-7 cell lines still con- in extracts obtained from eukaryotic Fourth, we have shown that the newly

Altex 25, 3/08 207 Fa v r e

reverse transcribed DNA was specifically 4.4 Inhibition of the HBV the rabbit reticulocyte lysate showed the HBV polymerase DNA, since it could polymerase in vitro striking differences (Fig. 7). A single be specifically employed as a molecular We have shown that the use of about polypeptide band was observed with the probe for the detection of HBV DNA in a 2.5 µM 3TC-TP in the assay inhibited DHBV polymerase, thus suggesting that Southern blot analysis (Fig. 5). 50% of the reverse transcription activity only the priming reaction took place. For the polymerase of DHBV, but not of the HBV polymerase. This inhibitory In contrast, the expression of the HBV that of HBV, the reverse transcription activity closely resembles that obtained polymerase resulted in a smear, without reaction could be initially artificially using the DHBV polymerase in the rab- the appearance of a defined polypeptide. reconstituted by in vitro translation, bit reticulocyte lysate (Seigneres, 2003). Moreover, it was also intriguing to ob- and this activity relies on the specific It is clear that the exact determination of serve that the DHBV polymerase was ε structure either in cis or in trans on the the IC50 measured in our work requires solely active in the initiation step but mRNA. By contrast, the HBV polymer- further investigation, because the assay not in the elongation step in this coupled ase activity of the HBV polymerase ex- was based on the presence of the endog- translation/reverse transcription assay. pressed from recombinant baculoviruses enous deoxynucleotides for the reverse These observations might be explained could be obtained in the absence of the transcription activity. For a more precise by the fact that several differences might ε signal, however to a lesser extent than measure of the inhibitory activity of the exist between the HBV and the DHBV with an ε signal on the mRNA (Lanford antiviral compounds, the translational polymerases. These are for example the et al., 1997). The same phenomenon has extracts should first be passed through a inherent differences in the assay systems been demonstrated in Xenopus oocytes sepharose column in order to remove low employed, and also the intrinsic behav- in the absence of ε as well, however molecular weight molecules (such ions iour of these two polymerases. without characterisation of the priming and nucleotides), and thereafter supple- reaction in this system (Seifer and Stand- mented for standardisation with various 5 Concluding remarks ring, 1993). The requirements for in vitro molecular compounds, as described else- priming are less stringent than in vivo, where (Ochoa and de Haro, 1979; Pelham In conclusion, this study provides the since the 5’ copy of ε in HBV genomic and Jackson, 1976). Our initial attempts first evidence that HBV polymerase pro- replication is absolutely required, as to standardise the translational extracts duced in translational extracts obtained shown by mutation studies (Nassal and by passage through a column and supple- from eukaryotic cells grown as monol- Rieger, 1996). In contrast to the results mentation failed to generate extracts in ayers could initiate reverse transcrip- obtained in these two systems, we have which efficient in vitro translation on ex- tion in vitro in the absence of other viral shown that the deletion of the ε signal ogenous mRNAs occurred. Further work proteins. We favour a model in which a on the mRNA coding for the full-length is thus needed to circumvent this tech- mechanism is active for the regulation HBV polymerase did result in a complete nical problem. This peculiar minor bias of the ribosome and the HBV polymer- lack of reverse transcription activity (Fig. kept in mind, we have analysed the fate ase on the same mRNA template. Once 6). Thus, it appeared that the interaction of various putative candidate molecules translation is completed, the polymerase of the HBV polymerase with the viral ε to inhibit the activity of HBV polymer- immediately initiates (-) strand synthesis. stem-loop structure on the mRNA during ases containing mutations in the YMDD A similar mechanism has been proposed the coupled translation/reverse transcrip- catalytic site. The hepadnavirus polymer- for poliovirus (Herold and Andino, 2001) tion reaction was necessary to induce or ases are sensitive to mutations in respect and barley yellow dwarf virus (Barry and activate the reverse transcription activity to the viral replication cycle. It has been Miller, 2002). of the HBV polymerase, at least in trans- shown that the YVDD and YIDD mu- The expression system described in this lational extracts obtained from human tants fail to respond to lamivudine thera- work might represent an important initial hepatocytes. py. Plasmids HH3 containing these point step towards the detailed understanding Finally, significant levels of polymer- mutations in the YMDD catalytic sites of hepadnaviral genomic replication. The ase mRNA have been shown to co-purify were constructed. In preliminary experi- availability of a biologically active HBV with the polymerase, albeit in a highly ments, we were able to shown that the polymerase in a coupled translation/ degraded state. Interestingly, only 1% promising antiviral molecules DAPD-TP reverse transcription assay provides a of the purified HBV polymerase protein and DXG-TP were effective as inhibitory means for the in vitro screening of inhibi- expressed with the baculovirus system compounds for the inhibition of the HBV tory molecules against the reverse tran- displayed reverse transcription activity. polymerase containing the mutations in scription activity of the HBV polymer- Thus, we can ask, whether this residual the YMDD catalytic site (D. Favre, un- ase. Moreover, it may be employed for reverse transcription activity was indeed published observations). Further work is the screening of antiviral molecules di- not due to the potential priming and the needed to study these inhibitory activities rected against mutants of the polymerase, extension reactions of the HBV polymer- in more detail. such as the YIDD or the YVDD mutants ase on these remaining RNA molecules We have observed that the expression (instead of YMDD in the catalytic site of that potentially contain an ε signal. of the HBV and DHBV polymerases in the wild-type polymerase). Such inhibi-

208 Altex 25, 3/08 Fa v r e

tors could provide novel approaches to Cavanaugh, V. J., Guidotti, L. G. and and membrane insertion of Escherichia antiviral therapy without the use of large Chisari, F. V. (1997). Interleukin-12 coli OmpF porin. Eur. J. Biochem. 222, amounts of laboratory animals in the inhibits hepatitis B virus replication 625-630. screening process. in transgenic mice. J. Virol. 71, 3236- Gaillard, R. K., Barnard, J., Lopez, V. et This scientific article is a contribution 3243. al. (2002). Kinetic analysis of wild- to the 3Rs, for the following reasons: Chang, L. J., Pryciak, P., Ganem, D. and type and YMDD mutant hepatitis B • With the use of this in vitro procedure, Varmus, H. E. (1989). Biosynthesis of virus polymerases and effects of de- it can reduce the number of animals the reverse transcriptase of hepatitis B oxyribonucleotide concentrations and animal extracts that are employed viruses involves de novo translational on polymerase activity. Antimicrob. for the study of the HBV polymerase. initiation not ribosomal frameshifting. Agents Chemother. 46, 1005-1013. • It is refining the procedure employed Nature 337, 364-368. Genovesi, E. V., Lamb, L., Medina, I. et in the rabbit reticulocyte lysate with Colledge, D., Locarnini, S. and Shaw, al. (2000). Antiviral efficacy of lobu- duck HBV polymerase: an ameliorated T. (1997). Synergistic inhibition of cavir (BMS-180194), a cyclobutyl- procedure can be performed in eukary- hepadnaviral replication by lamivu- guanosine nucleoside analogue, in the otic extracts by doing the concomitant dine in combination with penciclovir woodchuck (Marmota monax) model transcription/translation of the HBV in vitro. Hepatology 26, 216-225. of chronic hepatitis B virus (HBV) in- polymerase. This was not possible in Cullen, J. M., Smith, S. L., Davis, M. G. fection. Antiviral. Res. 48, 197-203. the rabbit reticulocyte lysate with duck et al. (1997). In vivo antiviral activ- Guo, Q., Zhao, L., You, Q. et al. (2007). or human HBV polymerase. ity and pharmacokinetics of (-)-cis-5- Anti-hepatitis B virus activity of wo- • It can replace the use of the latter ani- fluoro-1- [2-(hydroxymethyl)-1,3-ox- gonin in vitro and in vivo. Antiviral mals and animal extracts; there is no athiolan-5-yl]cytosine in woodchuck Res. 74, 16-24. further need for ducks, ducklings and hepatitis virus-infected woodchucks. Hafkemeyer, P., Keppler-Hafkemeyer, A., rabbit reticulocyte lysate. Antimicrob. Agents Chemother. 41, al Haya, M. A. et al. (1996). Inhibition 2076-2082. of duck hepatitis B virus replication by Doong, S. L., Tsai, C. H., Schinazi, R. 2’,3’-dideoxy-3’-fluoroguanosine in References F. et al. (1991). Inhibition of the rep- vitro and in vivo. Antimicrob. Agents Addison, W. R., Walters, K. A., Wong, W. lication of hepatitis B virus in vitro by Chemother. 40, 792-794. W. et al. (2002). Half-life of the duck 2’,3’- dideoxy-3’-thiacytidine and re- Herold, J. and Andino, R. (2001). Polio- hepatitis B virus covalently closed cir- lated analogues. Proc. Natl. Acad. Sci. virus RNA replication requires genome cular DNA pool in vivo following inhi- U S A 88, 8495-8499. circularization through a protein-pro- bition of viral replication. J. Virol. 76, Favre, D., Berthillon, P. and Trépo, C. tein bridge. Mol. Cell 7, 581-591. 6356-6363. (2001). Removal of cell-bound lipo- Howe, A. Y., Robins, M. J., Wilson, J. Aguesse-Germon, S., Liu, S. H., Cheval- proteins: a crucial step for the efficient S. and Tyrrell, D. L. (1996). Selective lier, M. et al. (1998). Inhibitory effect infection of liver cells with hepatitis inhibition of the reverse transcription of 2’-fluoro-5-methyl-beta-L-arabino- C virus in vitro. C. R. Acad. Sci. 324, of duck hepatitis B virus by binding furanosyl-uracil on duck hepatitis B 1141-1148. of 2’,3’-dideoxyguanosine 5’-triphos- virus replication. Antimicrob. Agents Favre, D. and Muellhaupt, B. (2005). Re- phate to the viral polymerase. Hepatol- Chemother. 42, 369-376. activation of creatine kinase by dithio- ogy 23, 87-96. Barry, J. K. and Miller, W. A. (2002). A threitol prior to use in an in vitro trans- Hu, J. and Seeger, C. (1996). Expression -1 ribosomal frameshift element that lation extract. ALTEX 22, 259-264. and characterization of hepadnavirus requires base pairing across four kilo- Favre, D., Petit, M. A. and Trépo, C. reverse transcriptases. Methods Enzy- bases suggests a mechanism of regulat- (2003). Latent hepatitis B virus (HBV) mol. 275, 195-208. ing ribosome and replicase traffic on a infection and HBV DNA integration is Jacquard, A. C., Brunelle, M. N., Pi- viral RNA. Proc. Natl. Acad. Sci. U S associated with further transformation choud, C. et al. (2006). In vitro charac- A 99, 11133-11138. of hepatoma cells in vitro. ALTEX 20, terization of the anti-hepatitis B virus Beck, J. and Nassal, M. (1998). Forma- 131-142. activity and cross-resistance profile of tion of a functional hepatitis B virus Favre, D. and Trépo, C. (2001). Trans- 2’,3’-dideoxy-3’-fluoroguanosine. An- replication initiation complex involves lational extracts active biologically in timicrob. Agents Chemother. 50, 955- a major structural alteration in the vitro obtained from eukaryotic monol- 961. RNA template. Mol. Cell. Biol. 18, ayer cells: a versatile method for viral Jeong, J. H., Kwak, D. S., Rho, H. M. 6265-6272. RNA studies. J. Virol. Methods 92, and Jung, G. (1996). The catalytic Cao, F. and Tavis, J. E. (2006). Suppres- 177-181. properties of human hepatitis B virus sion of mRNA accumulation by the Fourel, D., Bernadac, A. and Pages, J. polymerase. Biochem. Biophys. Res. duck hepatitis B virus reverse tran- M. (1994). Involvement of exposed Commun. 223, 264-271. scriptase. Virology 350, 475-483. polypeptide loops in trimeric stability Kim, Y. and Jung, G. (1999). Active hu-

Altex 25, 3/08 209 Fa v r e

man hepatitis B viral polymerase ex- cation in hepatocytes and bile duct epi- and nonsense polymerase mutants of pressed in rabbit reticulocyte lysate thelial cells in vivo. J. Gastroenterol. human hepatitis B virus. J. Virol. 65, system. Virus Genes 19, 123-130. Hepatol. 15, 304-310. 3617-3624. Lanford, R. E., Notvall, L., Lee, H. and Nicoll, A. J., Colledge, D. L., Toole, J. J. Seifer, M., Hamatake, R. K., Colonno, R. Beames, B. (1997). Transcomplemen- et al. (1998). Inhibition of duck hepa- J. and Standring, D. N. (1998). In vitro tation of nucleotide priming and re- titis B virus replication by 9-(2- phos- inhibition of hepadnavirus polymer- verse transcription between independ- phonylmethoxyethyl)adenine, an acy- ases by the triphosphates of BMS- ently expressed TP and RT domains clic phosphonate nucleoside analogue. 200475 and lobucavir. Antimicrob. of the hepatitis B virus reverse tran- Antimicrob. Agents Chemother. 42, Agents Chemother. 42, 3200-3208. scriptase. J. Virol. 71, 2996-3004. 3130-3135. Seifer, M. and Standring, D. N. (1993). Le Guerhier, F., Pichoud, C., Jamard, Ochoa, S. and de Haro, C. (1979). Regu- Recombinant human hepatitis B virus C. et al. (2001). Antiviral activity of lation of protein synthesis in eukaryo- reverse transcriptase is active in the ab- beta-L-2’,3’-dideoxy-2’,3’-didehy- tes. Annu. Rev. Biochem. 48, 549-580. sence of the nucleocapsid or the viral dro-5- fluorocytidine in woodchucks Offensperger, W. B., Offensperger, S., replication origin, DR1. J. Virol. 67, chronically infected with woodchuck Walter, E. et al. (1993a). Suramin pre- 4513-4520. hepatitis virus. Antimicrob. Agents vents duck hepatitis B virus infection Seigneres, B., Aguesse-Germon, S., Pi- Chemother. 45, 1065-1077. in vivo. Antimicrob. Agents Chemoth- choud, C. et al. (2001). Duck hepatitis Li, Z. and Tyrrell, D. L. (1999). Ex- er. 37, 1539-1542. B virus polymerase gene mutants as- pression of an enzymatically active Offensperger, W. B., Offensperger, S., sociated with resistance to lamivudine polymerase of human hepatitis B virus Walter, E. et al. (1993b). In vivo in- have a decreased replication capacity in an coupled transcription-translation hibition of duck hepatitis B virus in vitro and in vivo. J. Hepatol. 34, system. Biochem. Cell Biol. 77, 119- replication and gene expression by 114-122. 126. phosphorothioate modified antisense Seigneres, B., Martin, P., Werle, B. et al. Lin, E., Luscombe, C., Colledge, D. et oligodeoxynucleotides. Embo J. 12, (2003). Effects of pyrimidine and pu- al. (1998). Long-term therapy with the 1257-1262. rine analog combinations in the duck guanine nucleoside analog penciclovir Pelham, H. R. and Jackson, R. J. (1976). hepatitis B virus infection model. An- controls chronic duck hepatitis B virus An efficient mRNA-dependent transla- timicrob. Agents Chemother. 47, 1842- infection in vivo. Antimicrob. Agents tion system from reticulocyte lysates. 1852. Chemother. 42, 2132-2137. Eur. J. Biochem. 67, 247-256. Seigneres, B., Pichoud, C., Ahmed, S. S. Lofgren, B., Vickery, K., Zhang, Y. Y. and Poch, O., Sauvaget, I., Delarue, M. and et al. (2000). Evolution of hepatitis B Nordenfelt, E. (1996). 2’,3’-dideoxy- Tordo, N. (1989). Identification of four virus polymerase gene sequence during 3’-fluoroguanosine inhibits duck hepa- conserved motifs among the RNA- famciclovir therapy for chronic hepati- titis B virus in vivo. J. Viral Hepat. 3, dependent polymerase encoding ele- tis B. J. Infect Dis. 181, 1221-1233. 61-65. ments. Embo J. 8, 3867-3874. Severini, A., Liu, X. Y., Wilson, J. S. Luscombe, C., Pedersen, J., Uren, E. Rahn, J. J., Kieller, D. M., Tyrrell, D. L. and Tyrrell, D. L. (1995). Mechanism and Locarnini, S. (1996). Long-term and Gati, W. P. (1997). Modulation of inhibition of duck hepatitis B vi- ganciclovir chemotherapy for congeni- of the metabolism of beta-L-(-)-2’,3’- rus polymerase by (-)- beta-L-2’,3’- tal duck hepatitis B virus infection in dideoxy-3’-thiacytidine by thymidine, dideoxy-3’-thiacytidine. Antimicrob. vivo: effect on intrahepatic-viral DNA, fludarabine, and nitrobenzylthioino- Agents Chemother. 39, 1430-1435. RNA, and protein expression. Hepa- sine. Antimicrob. Agents Chemother. Shaw, T., Mok, S. S. and Locarnini, S. A. tology 24, 766-773. 41, 918-923. (1996). Inhibition of hepatitis B virus Nakabayashi, H., Taketa, K., Miyano, K. Rajagopalan, P., Boudinot, F. D., Chu, DNA polymerase by enantiomers of et al. (1982). Growth of human hepato- C. K. et al. (1996). Pharmacokinetics penciclovir triphosphate and metabolic ma cell lines with differentiated func- of (-)-2’-3’-dideoxy-3’-thiacytidine basis for selective inhibition of HBV tions in chemically defined medium. in woodchucks. Antimicrob. Agents replication by penciclovir. Hepatology Cancer Research 42, 3858-3863. Chemother. 40, 642-645. 24, 996-1002. Nassal, M. and Rieger, A. (1996). A Robaczewska, M., Narayan, R., Sei- Skup, D. and Millward, S. (1977). Highly bulged region of the hepatitis B virus gneres, B. et al. (2005). Sequence- efficient translation of messenger RNA RNA encapsidation signal contains the specific inhibition of duck hepatitis B in cell-free extracts prepared from replication origin for discontinuous virus reverse transcription by peptide L-cells. Nucleic Acids Res. 4, 3581- first-strand DNA synthesis. J. Virol. 70, nucleic acids (PNA). J. Hepatol. 42, 3587. 2764-2773. 180-187. Svitkin, Y. V., Pause, A. and Sonenberg, Nicoll, A., Locarnini, S., Chou, S. T. et al. Roychoudhury, S., Faruqi, A. F. and Shih, N. (1994). La autoantigen alleviates (2000). Effect of nucleoside analogue C. (1991). Pregenomic RNA encap- translational repression by the 5‘ lead- therapy on duck hepatitis B viral repli- sidation analysis of eleven missense er sequence of the human immunode-

210 Altex 25, 3/08 Fa v r e

ficiency virus type 1 mRNA. J. Virol. gift of plasmid HH3. This work was per- 68, 7001-7007. formed over several years and was sup- Tavis, J. E. and Ganem, D. (1993). Ex- ported by grants of the Institut National pression of functional hepatitis B virus de la Santé et de la Recherche Médicale polymerase in yeast reveals it to be the (INSERM, France ; “Poste Vert pour sole viral protein required for correct chercheurs étrangers”), the Swiss League initiation of reverse transcription. Proc. against Vivisection and for the Rights of Natl. Acad. Sci. U S A 90, 4107-4111. the Animals (Geneva), the Naef Founda- Tavis, J. E., Perri, S. and Ganem, D. tion for in vitro research (Geneva), the (1994). Hepadnavirus reverse tran- Hartmann-Müller Stiftung (Zürich), and scription initiates within the stem-loop the Rentenanstalt-Swiss Life Stiftung of the RNA packaging signal and em- (Zürich). Point mutants Y63D, L180M, ploys a novel strand transfer. J. Virol. M204I, M204V, L180M/M204I, L180M/ 68, 3536-3543. M204V, N236T, N238D and N236T/ Tomita, T., Yokosuka, O., Tagawa, M. et N238D in plasmid HH3 have been con- al. (2000). Decrease of wild-type and structed and sequenced and are available precore mutant duck hepatitis B virus for all researchers in the field, upon re- replication during lamivudine treat- quest. ment in white Pekin ducks infected with the viruses. J. Hepatol. 32, 850- 858. Correspondence to Urban, S., Fischer, K. P. and Tyrrell, D. Dr. Daniel Favre L. (2001). Efficient pyrophosphoroly- Helpatitis sis by a hepatitis B virus polymerase Rte de Berne 52 may be a primer-unblocking mecha- 1003 Lausanne nism. Proc. Natl. Acad. Sci. U S A 98, Switzerland 4984-4989. e-mail: [email protected] Wang, G. H. and Seeger, C. (1992). The reverse transcriptase of hepatitis B virus acts as a protein primer for viral DNA synthesis. Cell 71, 663-670. Wang, W. L. (1992). [Expression and significance of Pre-S1 and Pre-S2 in human primary hepatic carcinoma (PHC)]. Zhonghua Zhong Liu Za Zhi 14, 83-86. Witcher, J. W., Boudinot, F. D., Baldwin, B. H. et al. (1997). Pharmacokinetics of 1-(2-fluoro-5-methyl-beta-L-ara- binofuranosyl)uracil in woodchucks. Antimicrob. Agents Chemother. 41, 2184-2187. Zoulim, F., Dannaoui, E., Borel, C. et al. (1996). 2’,3’-dideoxy-beta-L-5-fluorocytidine inhibits duck hepatitis B virus reverse transcription and suppresses viral DNA synthesis in hepatocytes, both in vitro and in vivo. Antimicrob. Agents Chemother. 40, 448-453.

Acknowledgements We gratefully acknowledge Fabien Zoulim and Christian Trépo for infra- structure support and for the generous

Altex 25, 3/08 211