Substitution of proline 306 in the reverse transcriptase domain of hepatitis B virus regulates replication

Hepatitis B virus (HBV) belongs to the family Hepadnaviridae, which has a double-stranded DNA genome with a single-stranded gap. The genome of HBV encodes the envelope protein (S/pre-S), the core protein (C/pre-C), the polymerase (P) and a transactivating protein (X). In addition, a number of regulatory elements, such as the promoters and enhancers of the envelope proteins and the core proteins have been identified (Moolla et al., 2002). The replication cycle of HBV is a well-organized process, which involves alternating use of DNA and RNA in the viral genome (Will et al., 1987). HBV has a very small genome of approximately 3·2 kb in length and two-thirds of the genome encodes its own DNA polymerase. The polymerase is composed of four domains, namely, the terminal protein, the spacer protein, the reverse transcriptase (RT) and the RNase H. Apart from the spacer protein, which has not been shown to be associated with specific functions, all the other domains have been assigned specific functions, and, by alignment of the amino acid sequences with similar enzymes from other retroviruses and hepadnaviruses, conserved and consensus regions have been identified (Radziwill et al., 1990). The active site of RT has been defined as containing five functional motifs, AE (Poch et al., 1989).

In a previous study with two naturally occurring HBV isolates, we showed that a mutation at residue 306 of RT that changed proline to serine (rtP306S) could change the replicative competency of an HBV isolate with a low level of replication (#2-18) to one with a high level (Lin et al., 2001b). Because rt306 is located close to the C terminus of RT and does not overlap with other open reading frames of HBV, the change in replicative competency of the mutant was not related to changes in other genes and thus was relatively simple to analyse. Based on homologous modelling studies of HBV RT, we suggested that rt306 was located at the connecting loop or hinge region between two α-helices in the thumb subdomain that interact with the DNA template-primer. In this case, substitution of rtP306S was likely to affect the precise conformation of the two α-helices and their interactions with the DNA template-primer and, consequently, impair the polymerase activity. To analyse the impact of the amino acid residue at rt306 further, here we constructed 11 HBV mutants in which the proline residue at rt306 (rtP306) was substituted with different amino acids using site-directed mutagenesis. The replicative competencies of these mutants were studied and analysed.

Construction of nucleic acid mutants encoding residue rt306.
A recombinant plasmid (p56) containing a full-length HBV isolate (#56) of high replication competency was used as the wild-type virus clone (Lin et al., 2001b) and primers were designed to induce mutations at the nucleic acid sequences encoding residue rt306, leading to the substitution of proline with the following amino acids: alanine, arginine, aspartic acid, glycine, glutamic acid, leucine, phenylalanine, serine, threonine, tyrosine and valine. These amino acids were selected based on varying properties of the side chains, such as hydrophobicity, charge, polarity and size. Mutagenesis was carried out using the Genetailor Site-Directed Mutagenesis system (Invitrogen) following the manufacturer's instructions. In brief, 200 ng p56 DNA was methylated with 8 U methylase in a total volume of 32 µl, and 2 µl methylated DNA was used as the template for PCR using paired primers (Table 1). The reverse primer MutR was used together with one of the forward primers to generate the amino acid substitutions, e.g. SerF was used to substitute proline with serine. After amplification, 2 µl PCR product was used to transform Escherichia coli DH5α-T1, and transformants were screened by growth on bacterial culture plates using ampicillin for selection. All mutants were sequenced to confirm the induced mutations.

Table 1. Primer sequences used for site-directed mutagenesis Mutated or deleted codons are underlined.

Construction of deletion mutants at various residues in RT.
To analyse the effects of deletion of various residues on the activity of RT, four deletion mutants were constructed using site-directed mutagenesis as described above (the primers used are shown in Table 1), with the following residues deleted: methionine at rt204 (rtΔM204), leucine at rt271 (rtΔL271), proline at rt306 (rtΔP306) and leucine at rt336 (rtΔL336). The methionine deleted in mutant rtΔM204 was located at the second position of the highly conserved YMDD motif that forms part of the active site of RT, while rtΔL271, rtΔP306 and rtΔL336 were all located close to the C terminus of RT, not overlapping the other coding regions of HBV. All deletion mutants were used to transfect HepG2 cells and their replicative competencies were quantified by real-time PCR (see below).

Cell culture, transfection and extraction of HBV DNA from intracellular particles.
HepG2 cells were cultured, and recombinant plasmids used for transfection were extracted and purified using the Qiagen Maxiprep kits as reported previously (Lin et al., 2001b). Wild-type and mutated HBV DNA was released from the recombinant plasmids by digestion with SapI [1 U (µg DNA)¹] for 18 h, followed by extraction and purification. HBV DNA (25 µg) from each clone was used to transfect HepG2 cells in 60 mm plates using the calcium phosphate precipitation method as reported previously (Lin et al., 2001a). Duplicate plates were used for all samples and 10 µg reporter plasmid DNA expressing secreted alkaline phosphate was cotransfected into each cell culture as an internal control to normalize the transfection efficiency between plates. Cells were collected at 72 h after transfection and vector pUC18 DNA was used as the mock transfection control. Transfection experiments were done three times on separate days. Wild-type recombinant DNA (p56) was used as the reference when transfection experiments were carried out on different days and the final results of the mutated recombinant plasmids used for transfection were compared with the reference result and normalized.

Cells were washed twice with chilled PBS and lysed with 600 µl lysis buffer (10 mM Tris/HCl, pH 8·0, 1 mM EDTA, 1 % NP-40, 8 % sucrose) as reported by Günther et al. (1995) with some modifications. After centrifugation at 12 000 g for 2 min, clear supernatants were collected and 6 µl 1 M magnesium acetate, 12 µl DNase I (5 mg ml¹) and 3 µl RNase A (20 mg ml¹) were added, followed by incubation at 37 °C for 30 min to digest the remaining DNA and RNA. After centrifugation at 12 000 g for 1 min, the supernatants were collected, 16 µl 0·5 M EDTA and 130 µl 35 % PEG 8000 in 1·75 M NaCl were added and the mixture was kept on ice for 1 h to precipitate the core particles. After spinning at 9000 g for 5 min, the pellets were resuspended and again digested with DNase I [0·1 µg (µg DNA)¹] and incubated at 37 °C for 10 min to ensure full digestion of the remaining DNA used for transfection. The resuspended core particles were then digested with proteinase K and nucleic acids were extracted and precipitated. At this stage, DNA was further digested with HpaII. Full-length HBV DNA (#56 or the various mutants) released from plasmid originating from E. coli is digested by HpaII into seven fragments at nt 509, 1572, 2035, 2332, 3006 and 3042, respectively, while due to DNA methylation at sequences of CCGG, viral DNA derived from cells is resistant to this enzyme (Reyna-Lopez et al., 1997). A pair of primers flanking one of the six restriction sites of HpaII was used for PCR (Fig. 1) to ensure that the amplified DNA product was derived from the viral core particles within the transfected cells and not from remaining contaminating plasmid HBV DNA used for transfection.

(19K): Fig. 1. HpaII sites in the HBV #56 genome and primers and probes used in real-time PCR for virus copy number assays. In #56 and the constructed mutants, there were six HpaII sites located at nt 509, 1572, 2035, 2332, 3006 and 3042 (indicated by arrowheads). The location of primers flanking the HpaII site at nt 509 (HBVF and HBVR), together with the acceptor probe (HBVA) and donor probe (HBVD) used for real-time PCR, are shown.

Quantification of HBV DNA from core particles.
Copy numbers of HBV DNA extracted from the core particles as above were assayed by real-time PCR using a Diagnostics LightCycler (Roche) following the manufacturer's guidelines. As shown in Fig. 1, a pair of primers flanking nt 509 was used: HBVF, 5'-GCCCGTTTGTCCTCTAATTCC-3' (nt 468488) and HBVR, 5'-GGGAAAGCCCTACGAACCACT-3' (nt 698718). The PCR product was 251 bp. The donor fluorescein probe (HBVD, 5'-ACCTGCACAACTCCTGCTCAfluorescein-3', nt 523542) and the acceptor LightCycler-Red 640 probe (HBVA, 5'-LCRedTGAGGCCCACTCCCATAGGT-phosphate-3', nt 637656) specific for the 251 bp product were designed according to the manufacturer's guidelines (Roche). Mock-transfected cellular extracts were expected to be negative and the copy number of HBV DNA extracted from the wild-type virus-transfected cells in each experiment was set as the reference for comparison. The HBV copy number from cells transfected with each mutant was divided by the corresponding wild-type virus copy number and the percentage increase or decrease in replicative competency of each mutant was calculated.

HBV surface antigen (HBsAg) and HBV e antigen (HBeAg) assay.
Supernatants were collected 72 h after cell transfection and were assayed for the HBsAg and HBeAg using diagnostic ELISA kits (Kehua Co.).

Cloning of the enhancer I fragments and CAT assay.
Five different enhancer I fragments spanning nt 9571318 (Bock et al., 2000), according to the nucleotide sequence of wild-type HBV, were amplified separately by PCR for the wild-type HBV (rtP306) and four constructed polymerase mutants with mutations at rt306 (rtP306G, rtP306S, rtP306D and rtP306E). PCR products were cloned at the restriction sites KpnI and BglII of reporter vector pCAT3-promoter (Promega). The primers used were: E1, 5'-GGGGTACCCTTCCTGTTAACAGGCCTAT-3' (starting at nt 957), and E2, 5'-GAAGATCTTTTCCGCGAGAGGACGACAGA-3' (terminating at nt 1318).

Huh-7 cells were cultured in Dulbecco's modified Eagle's minimal essential medium supplemented with 10 % fetal calf serum. Two micrograms of the enhancer I constructs was used to transfect cells using Lipofectamine 2000 (Invitrogen) with 1·0 µg vector pCMVβ (Promega) expressing β-galactosidase, used as the internal standard. Cells were harvested and lysed 48 h after transfection and, after normalization by comparison with intracellular β-galactosidase, the CAT activity of the lysates was assayed using a CAT ELISA kit following the manufacturer's instructions (Roche). Experiments were carried out three times and data are shown as means±SD.

Replication competency of mutants with substitutions at rt306
The mean replication competencies of all 11 mutants in three separate experiments are shown in Table 2. Although all mutants showed varying replication competencies compared with wild-type virus, a marked reduction was found when rtP306 was substituted with glycine or threonine, with replication competencies of 1·96 and 4·51 % compared with the virus copy number of the wild-type virus. Mutation of rtP306 to alanine, arginine or tyrosine also impaired the replication competency of the mutant viruses substantially. On the other hand, mutation of rtP306 to leucine, aspartic acid or valine increased the replication competency slightly, while substitution of rtP306 with glutamic acid resulted in a significant enhancement of replication competency by 9·4-fold. Enhanced replication of the glutamic acid-substituted mutant and decreased replication of the serine, glycine and other substituted mutants were confirmed by the higher or lower levels of HBsAg and HBeAg detected in the cell-culture supernatants (Fig. 2).

Table 2. Replication competencies of mutants substituted at rt306

(38K): Fig. 2. HBsAg and HBeAg levels in the supernatants of cells transfected with various mutants substituted at amino acid rt306. The wild-type virus (#56) contains proline at rt306 and the various amino acids substituted for rtP306 are indicated. Supernatants of transfected cells were collected 72 h after transfection and assayed for HBsAg and HbeAg. Compared with the wild-type virus, significantly decreased levels of HBsAg in mutants rtP306G and rtP306S were found (P<0·01). Significantly increased (rtP306T and rtP396R) or decreased (rtP306Y, rtP306G and rtP306S) levels of HBeAg were also found (P<0·01).

Replication competency of mutants with deletions in RT
All deletion mutants showed greatly reduced replication, irrespective of whether the deletion was at the active site of RT (rtΔM204) or located at the C terminus (rtΔL271, rtΔP306 and rtΔL336). The HBV DNA copy number from wild-type HBV (#56)-transfected cells was 2·46x10⁷, while in deletion mutants rtΔM204, rtΔL271, rtΔP306 and rtΔL336, the virus copy numbers were 0·0084x10⁷, 0·011x10⁷, 0·016x10⁷ and 0·011x10⁷, respectively. Since these values corresponded to 0·3, 0·4, 0·6 and 0·4 % of the value for wild-type virus, these deletion mutants were considered to be replication defective.

Transcriptional activities of the enhancer I mutant
As shown in Fig. 3, compared with the empty vector, all cloned enhancer I fragments showed transcriptional function. More importantly, there was no difference in the level of transcriptional activities of enhancer I fragments with mutations corresponding to the substituted residues 306 of RT, irrespective of whether these mutants showed enhanced (rtP306D and rtP306E) or decreased (rtP306S and rtP306G) replication efficiency.

(12K): Fig. 3. Transcriptional activities of five enhancer I mutants. Five different enhancer I fragments (nt 9571318) corresponding to wild-type rtP306 and to mutants rtP306S, rtP306G, rtP306D and rtP306E were each cloned into the vector pCAT3-promoter. These constructs were used to transfect Huh-7 cells and the vector pCMVβ expressing β-galactosidase was used as an internal standard. Cells were harvested and lysed 48 h after transfection. After normalization against intracellular β-galactosidase, the CAT activities of the lysates were determined. The position of the codon corresponding to rt306 (nt 10471049) is indicated.

In a previous study with two naturally occurring HBV adw subtype isolates, we showed that a single amino acid change from proline to serine at rt306 decreased the replication competence of the virus (Lin et al., 2001a). Based on homologous modelling of RT of HBV, we suggested that rtP306 is located at the connecting loop between two α-helices of the thumb subdomain, equivalent to α-helices H and I in HIV-1 RT that interact with the DNA template-primer and modulate the translocation of DNA during polymerization (Ding et al., 1998; Das et al., 2001). Because proline has a rigid conformation with low flexibility, this amino acid might be necessary to constrain the relative conformation of the two helices to maintain a proper interaction of the template-primer with the polymerase and its precise position at the polymerase active site. To study further the impact of the amino acid residue at rt306, we constructed 11 HBV RT mutants in which rt306 was substituted with amino acids with different side-chain properties. In this study, we used real-time PCR to monitor the virus copy numbers of intracellular HBV core particles. To ensure that the virus copy numbers detected were not from contaminating plasmid HBV DNA used for the transfection, we used HpaII enzyme digestion followed by amplification using real-time PCR primers covering the restriction site of this enzyme. By this means, plasmid-derived HBV DNA should be cut by HpaII and therefore not amplified by the primers, ensuring that the virus copy numbers detected truly represented the replicative competency of the mutants. Marked changes in the replication competency of mutants with substitutions at rt306 further demonstrated the essential role of this residue.

Apart from substitutions of the amino acid at rt306, the remaining nucleotide sequences of the full-length HBV genome were identical in all mutants. The higher level of HBsAg detected by the two mutants with enhanced replication competency (glutamic acid and leucine) and lower levels of HBsAg detected with the mutants with reduced replication competency further supported the conclusion that changes of the amino acid at rt306 modulated HBV replication. It should be noted that the codon corresponding to rt306 overlaps the enhancer I region (nt 10471049) of the HBV genome (Bock et al., 2000). Thus, one could speculate that the varying replication efficiencies of the HBV mutants were due to alterations in the activities of the enhancer I mutants. To study the possible effects of mutated enhancer I at codon rt306, we cloned five different enhancer I fragments, representing wild-type and mutants with both enhanced replicative efficiency (rtP306D and rtP306E) and decreased replication efficiency (rtP306S and rtP306G). The results indicated that the mutation at codon 306 did not affect the transcriptional activities of enhancer I (Fig. 2); thus, the effect of mutated sequences of HBV enhancer I on HBV replication could be excluded.

Although the low replication competency of rtP306G could be explained by the simple structure of glycine, rendering higher flexibility at this hinge region and resulting in impaired binding of the α-helices with the DNA template-primer, this mechanism is difficult to correlate with the reduced replication competency of the rtP306T mutant. Therefore, the impact on the structure of substitutions of rtP306 appears to be relatively complex and cannot be explained solely by the current structural model of HBV RT, which contains only the catalytic domain of HBV RT. Because rtP306 is located at the interface between the thumb and the palm subdomain and the RNase H domain is located at the C terminus of the thumb subdomain, a change of rtP306 to other amino acids could alter the interactions of the thumb subdomain with the palm subdomain and/or the RNase H domain via either hydrophilic or hydrophobic interactions of the side chains of the amino acids. It could be assumed that enhanced replication of the rtP306E mutant is due to interaction of its hydrophilic side chain with the side chains of other amino acids located in RNase H. A change in these interactions could potentially interfere with interactions between the enzyme and the DNA template-primer and consequently affect the enzymic activity of HBV polymerase. Alternatively, the residue at the thumbpalm interface may play an additional role(s) in virus replication, such as involvement in modulating the conformational changes of the polymerase during the various steps of the replication process. Recently, Fisher et al. (2003) reported the biochemical basis of how HIV RT copies both RNA and DNA templates. They found that mutations proximal to the minor groove binding track in the thumb region of HIV RT (N265D or N255D) led to loss of processing polymerization on viral DNA or both RNA and DNA templates, and the mechanism was described as being due to either a loss of template-strand-specific hydrogen bonding or a local change in conformation (Das et al., 2001). However, detailed three-dimensional structural information about HBV RT is not yet available. In addition, the HepG2 cell transfection system (Sells et al., 1984) used in this study reflects only part of the virus replication cycle (i.e. DNA synthesis); therefore, an appropriate cell-culture system and genome-based structural and functional analysis of RT from more HBV isolates are necessary to elucidate the regulatory elements of this enzyme.

This work was supported by the Chinese State Basic Science Foundation Grant no.1999054105, Shanghai Municipal Basic Research Fund Grant no. 01JC14019 and China National Natural Science Foundation Grant no. 39630020. We thank Professor Chris Burrell (Adelaide University, Australia) for reading and editing this manuscript.