Phenotypic assay of a hepatitis B virus strain carrying an rtS246T variant using a new strategy

Case Report

The patient selected was a 35-year-old Chinese man with a chronic HBeAg-positive HBV infection. He had never been treated with nucleot(s)ide analogs, and was enrolled in a registered clinical study [Guo et al., 2009]. The patient had received 0.5 mg ETV monotherapy daily for more than 3 years. The dynamics of his serum viral load and aminotransferase activities are shown in Figure 1. No HBeAg to anti-HBe conversion was observed during the whole period of therapy.

Sequence Analysis of Serum HBV DNA

Viral DNA was extracted from 200 µl of serum using the High Pure Viral Nucleic Acid Kit (Roche Diagnostics, Mannheim, Germany), according to the manufacturer's instructions. The final volume of eluate was 50 µl and 1 µl was used for subsequent PCR amplifications. Primers F₂₅₀ and R_1,058 were used to amplify an 808-bp fragment encompassing domains A–E of the HBV reverse transcriptase. The amplicons were sequenced with an ABI PRISM® 3100 Genetic Analyzer (Applied Biosystems Inc., Foster City, CA). All PCR analyses in the present study were performed using PrimeStar HS DNA Polymerase (Takara Biotechnology, Dalian, China). Nine primers (all shown in the 5′ to 3′ direction) used are listed in Table I.

Table I. Sequences of Primers.

• Numbers indicate the positions on the HBV genome to which the primers would anneal (defining EcoRI as position 1).
• Underlined are sequences that anneal to the corresponding HBV sequence.
• Boxed F_c-p-e-1,818 and F_c-p-e-1,816 are sequences that can anneal to both PCH-9/3091 and pEGFP-N1.
• Boxed F_pch9-egfp and _Rpch9-egfp are sequences that can anneal to PCH-9/3091.

HBV 1.1× Plasmid Construction

Fragment Substitution Reaction (FSR)

Plasmid pHBV1.3/1, which contains 1.3 copies of the HBV genome DNA (GenBank accession number: GU451682) from the serum of the patient described above, was constructed previously [Hu et al., 2009]. Plasmid PCH-9/3091, constructed by Nassal [1992] and containing a CMV-early-promoter-driven HBV 1.1× genome, was kindly provided by Lin Nan (Southwest Hospital, Chongqing, China). Figure 2 shows the procedure used to construct p1818-LL, a construct containing the HBV DNA from the patient. Fragment 1,818–803 (designated conventionally, with the ‘C’ corresponding to that in the EcoR I site of genotype A as position 1, although there is no EcoR I site in the genome of this strain), about 2.2 kb in length, was amplified from pHBV1.3/1 using primers F_c-p-e-1,818 and R₈₀₃. After gel purification with the QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany), 800 ng of the fragment was used as the mega-primers in a 50 µl FSR system, together with 50 ng of PCH-9/3091, 4 µl of 2.5 mM dNTPs, 1.25 U of PrimeStar HS DNA Polymerase, and buffer. The cycling parameters used were: Initial denaturation at 94°C for 2 min, followed by 18 cycles of denaturation at 94°C for 15 sec, annealing at 55°C for 15 sec, and extension at 72°C for 4 min. The resultant FSR product was purified with a PCR Purification Kit (Roche Diagnostics), digested with Dpn I (Promega, Madison, WI), and then transformed into competent JM109 cells. That the clone contained the replaced fragment was confirmed by sequencing and it was designated p1818. Plasmid p1818-LL was generated by replacing nucleotides (nt) 599–1,825 of p1818 with a fragment (nt 599–1,825) amplified from pHBV1.3/1 with the same method. The fragment amplified from pHBV1.3/1 with F_c-p-e-1,816 and R₈₀₃ was also used to replace the corresponding region in PCH-9/3091 by FSR to generate the plasmid p1816 (not shown in Fig. 2). The distance from the TATA box of the CMV promoter to nt 1,818 in p1818 is 2 nt shorter than in p1816 (Fig. 4A). This 2-nt deletion is expected to cause p1818 to transcribe pgRNA from nt 1,818. p1818 and p1816 were used to evaluate the influence of different pgRNA transcription start sites on the replication capacity of the construct.

Insertion of a CMV–EGFP Fragment Downstream From HBV DNA

A CMV–EGFP DNA fragment containing the CMV immediate early promoter, the gene encoding enhanced green fluorescent protein (EGFP), and the SV40 polyA signal was amplified from the plasmid pEGFP-N1 (Clonetech, Mountain View, CA) with primers F_pch9-egfp and R_pch9-egfp. The plasmid PCH9–EGFP, carrying CMV–EGFP downstream from HBV 1.1×, was generated with the CMV–EGFP DNA fragment using the same FSR method described in Fragment Substitution Reaction section, except that the FSR extension was performed at 72°C for 4 min 30 sec for 30 cycles. The CMV–EGFP DNA fragment was also inserted into p1818-LL to generate the plasmid pLL–EGFP. pLL–EGFP and PCH9–EGFP were the constructs used in susceptibility assays.

Cell Culture and Transfection

HepG2 cells were grown at 37°C under 5% CO₂ in modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum. For transfection, the cells were seeded into six-well plates or 3.5-cm-diameter plates and allowed to adhere overnight. On the following day, when the cells were 60–70% confluent, the culture medium was replaced with fresh medium and 4 µg of HBV construct was transfected into the cells in each well using 8 µl Lipofectamine 2000 (Invitrogen, Carlsbad, CA), according to the instructions provided by the supplier. The culture medium was changed every 2 days, and the cells were harvested on day 5 after transfection.

Viral DNA Analysis

Intracellular viral core DNA was extracted as described previously, with modifications [Hu et al., 2009]. Briefly, cells from one well or plate were lysed with 400 µl of lysis buffer (10 mM Tris–HCl [pH 8.0], 1 mM EDTA, 1% NP-40, and 2% sucrose) at 37°C for 10 min. After centrifugation at 15,000g for 3 min, the supernatant was transferred to a fresh microcentrifuge tube and 1 M MgCl₂ was added to achieve a final Mg²⁺ concentration of 10 mM. DNase I (Promega; added to 40 U/ml) and RNase A (Calbiochem, Darmstadt, Germany; added to 1 mg/ml) were added to digest the contaminating plasmid and RNA, respectively. After 3 hr at 37°C, the reaction was terminated with 10 mM EDTA, and then 140 µl of 35% polyethylene glycol 8000 containing 1.5 M NaCl was added. After incubation for 1 hr on ice, the viral nucleocapsids were pelleted by centrifugation at 11,000g for 5 min at 4°C, and then digested (3 hr at 40°C) in 400 µl of buffer containing 1 mg/ml proteinase K (Promega). The digestion mixture was extracted twice with phenol. The DNA was precipitated with ethanol and dissolved in TE buffer (10 mM Tris–HCl [pH 8.0], 1 mM EDTA). The DNA was detected with an HBV DNA probe using the DIG DNA Labeling and Detection Kit (Roche Diagnostics), according to the manufacturer's instructions.

ETV Susceptibility Assay

For the drug susceptibility assays, six-well culture plates were seeded with 7.0 × 10⁵ HepG2 cells/well. Twenty hours after seeding, the cells were transfected with 4 µg of plasmid DNA using Lipofectamine 2000 transfection reagent. One day after transfection, the cells were fed with fresh medium containing 0, 0.2, 0.5, 2, 5, or 10 nM ETV (Bristol-Myers Squibb, New York, NY). The medium containing ETV was changed on day 3, and the cells were harvested 5 days after their transfection. To assess the proportion of transduced cells (containing plasmid), cells expressing EGFP within 15 selected randomly fields of 200× magnification on an X71 microscope (Olympus, Tokyo, Japan) were counted before harvest. The counts in all the wells were divided by the counts in the well to which no ETV was added, to calculate the ratio of transduced cells in each well relative to the number of cells in the ETV-free well. The intracellular HBV replicative intermediates were analysed by Southern blotting and quantified with the Quantity One software (Bio-Rad). All experiments were performed in triplicate. GraphPad Prism software (GraphPad Software, Inc., San Diego, CA) was used to determine the best-fit values for individual dose–response equations.

Strategy for HBV 1.1× Plasmid Construction

A new strategy was developed to construct HBV 1.1. The strategy is characterized by a method called the fragment substitution reaction (FSR). The principle of this method resembles that of site-directed mutagenesis developed by Stratagene. In practice, a DNA fragment (1 bp to several kb in size) located between two primers (e.g., 25 bp in length) that can anneal to a plasmid template can be considered ‘one’ mutation. When the two primers of a DNA fragment both anneal to a plasmid, DNA polymerase starts to extend the two strands using the plasmid as the template. The extension products will form a plasmid-like structure with one nick in each strand, as is the case in a site-directed mutagenesis reaction. After the plasmid template is removed by Dpn I digestion, the transformation of the remaining DNA will generate clones containing the substituted fragment.

To construct HBV 1.1×, fragment 1,818–803 was first amplified from pHBV1.3/1 using primers F_c-p-e-1,818 and R₈₀₃. Primer F_c-p-e-1,818 (Table I and Fig. 2B.2) contains the HBV sequence of nt 1,818–1,840 and the end part of the CMV promoter from the plasmid pEGFP-N1. In PCH-9/3091, part of the CMV promoter is identical to the 22 bp at the 5′ end of primer F_c-p-e-1,818, whereas the subsequent 16 bp, which comprises a Sal I site and the HBV sequence at nt 1,809–1,817, is different from the middle part of primer F_c-p-e-1,818 (Fig. 2B.2). The sequence of primer R803 is identical to the corresponding region in PCH-9/3091. With FSR, plasmid p1818 was obtained, which contains the substituted fragment nt 1,818–803 from pHBV1.3/1. Using p1818 as the template, another FSR was performed with fragment 599–1,825 amplified from pHBV1.3/1, producing plasmid p1818-LL, which contains one copy of the HBV genome from the patient serum plus less than 200 bp of the HBV sequence carried over from PCH-9/3091. A fragment containing the CMV promoter, the EGFP gene, and the SV40 polyA signal was then amplified from pEGFP-N1, and the fragment inserted downstream from HBV 1.1 in both p1818-LL and PCH-9/3091 by FSR. The resultant plasmids were designated pLL–EGFP and PCH9–EGFP, respectively.

Fragments can be Substituted Effectively Into a Plasmid With the Fragment Substitution Reaction

More than 20 experiments using FSR have been performed in our laboratory, including FSR with fragments amplified directly from serum containing HBV DNA, and most of them were successful (data not shown). When a fragment with high homology to the template is used as the mega-primers, such as the substitution of one genotype of HBV DNA for another, FSR is very efficient. The newly synthesized recombinants can be observed readily on a gel (Fig. 3A,B). When a fragment with relatively lower homology to the template was used, as in the FSR of CMV–EGFP, the FSR products could not be detected on a gel (Fig. 3C), despite an additional 12 cycles of reaction (30 cycles total; see Insertion of a CMV-EGFP Fragment Downstream From HBV DNA section) being added.

After the transformation of the FSR products, three clones were usually picked for sequencing. In most cases, two of the three clones or three of the three clones contained a correctly substituted fragment. Figure 4 shows the sequence alignment of four FSR products with their templates. The sequence downstream from the TATA box in p1818 and p1816 (A) and the sequence downstream from nt 803 in p1818-LL differ from that in PCH-9/3091, confirming that these regions have been substituted.

p1818 and p1816 Replicate HBV at Similar Levels in HepG2 Cells

PCH-9/3091 has been reported to contain an extra pgRNA transcription start site at nt 1,816 [Liu et al., 2004]. To analyse whether this extra transcription start site affects the level of replication, plasmids p1818 and p1816 were generated. P1818 and p1816 were each designed so that their transcription start sites coincide with nt 1,818 and nt 1,816, respectively (Fig. 4A); so, p1818 was expected to generate pgRNA transcripts starting from nt 1,818, whereas p1816 generated them from nt 1,816. p1816 or p1818 (4 µg) was transfected into HepG2 cells and the intracellular core replicative intermediates were analysed by Southern blotting. As shown in Figure 5, after the cell numbers were normalized to the cell viability, no significant difference was observed between the replication levels of p1816 and p1818. When the replication capacity of p1816 was set as 100%, the replication level of p1818 was 104% (mean value).

Sequence Analysis of Prevalent HBV Reverse Transcriptase at Different Time Points Following ETV Treatment

HBV reverse transcriptase fragments were amplified from the sera of the selected patient at 0, 28, 76, 124, and 144 weeks after ETV treatment and directly sequenced. All were shown to encode identical amino acid sequences (data not shown). No previously reported ETV-resistance-associated substitution, including M204V/I, L180M, I169T, S202G/I, M250V, or T184G, was found in the sequences. The nucleotide sequences of these fragments were classified as genotype B according to the classification system of Norder et al. [2004]. When compared with the alignment data of the 116 genotype B HBV strains described previously [Hu et al., 2009], these sequences from our patient encoded a threonine at reverse transcriptase amino acid 246 (numbering according to the nomenclature of Stuyver et al., 2001) instead of the serine present in the 116 genotype B strains. Amino acid 246 is located one codon 5′ to the boundary of the conserved subdomain E in HBV polymerase, as defined by Stuyver et al. [2001].

ETV Susceptibility of the Patient Isolate

As described above, an HBV 1.1× genomic recombinant plasmid (pLL–EGFP) was generated from the HBV genome DNA from the serum of a patient. To test the susceptibility of the isolate from the patient, HepG2 cells were transfected with pLL–EGFP and treated with various concentrations of ETV. The intracellular replicative intermediates were assayed by Southern blotting (Fig. 6A). The ETV IC₅₀ of the isolate calculated from the dose–inhibition curve (Fig. 6C) was 0.99 ± 0.09 nM. As a wild-type control, PCH9–EGFP was transfected into HepG2 cells and their susceptibility to ETV was tested (Fig. 6B,C). This laboratory strain yielded an IC₅₀ of 0.58 ± 0.10 nM. These results suggest that the HBV isolated from this patient was not resistant to ETV.

FSR is an Effective Method for DNA Cloning

In this study, a new strategy for constructing HBV 1.1× recombinants with a CMV promoter to drive HBV pgRNA transcription was developed. The core of this strategy is the FSR method. FSR allows a fragment (the “substituted fragment”, SF) in a plasmid to be substituted with another fragment (the “target fragment”, TF). The homology between SF and TF can be 0–100%, which allows great flexibility. When the homology is 0, this method essentially becomes a cloning technique, with the same effect as the traditional digestion-and-ligation cloning technique. A non-homologous fragment, CMV–EGFP–PolyA, was successfully inserted into plasmids PCH-9/3091 and pLL1818, as in the construction of plasmids PCH–EGFP and pLL–EGFP, respectively. In this experiment, EGFP was expressed from the HBV 1.1× plasmid, although there was no appropriate site downstream from HBV 1.1×. Two 24-bp sequences (Table I) were added to the 5′ end of the primers for CMV–EGFP amplification respectively. These two sequences could anneal to the two sequences separated by 6-bp, downstream from HBV 1.1×. The CMV–EGFP PCR fragment was then inserted into the plasmids by FSR. Geiser et al. [2001] have used this method to insert fragments of 34–1,117 bp into plasmids. The present results demonstrate that longer fragments can also be inserted into a plasmid without restriction enzyme digestion or ligation. Recently, Stech et al. [2008] have used this method to clone complete sets of functional gene segments from influenza A strains, showing that fragments of up to 2,341 bp can be cloned effectively with this method.

When SF and TF have a relatively high degree of homology, as in the case of HBV fragment substitution, FSR can substitute the fragment with high efficiency. In the FSR for fragments 1,818–803 and 599–1,825, the differences between these two TFs and their SFs were 12.4% (including a 33-bp deletion of nt 2,897–2,929 from the genotype D HBV DNA in PCH-9/3091) and 10.0%, respectively. As shown in Figure 3A,B, 18 cycles of amplification resulted in electrophoresis-detectable FSR products. The transformation of FSR products digested with Dpn I led to clones with a low background of the template plasmid. There was always one or more correctly substituted clone in the three clones tested.

Advantages and a Disadvantage of FSR in HBV Phenotypic Assay

The major advantage of FSR over conventional cloning is that in FSR, the sequence specificity around the cloning site need not be considered, i.e., the presence of an appropriate restriction site is no longer a limitation. Most of the methods available currently to generate replication-competent HBV DNA constructs involve digestion and ligation process. Some restriction sites used in these methods include but are not limited to Sap I [Gunther et al., 1995; Yang et al., 2004], Nco I [Durantel et al., 2004], and Spe I [Hu et al., 2009]. Although these restriction sites are rather conserved in many HBV genomes, the high heterogeneity of HBV [Medley et al., 2001; De Maddalena et al., 2007] still makes it possible that these sites change in other HBV genomes. For example, the Sap I site is present in some HBV genomes [Gunther et al., 1995] and two Spe I sites have been found in some genotype G HBV genomes [Hu et al., 2009], thus making corresponding methods not feasible in some situations. The Nco I site seems to be conserved highly among all HBV genomes; however, sequences around the Nco I site are not conserved (data not shown). That could be a problem for designing desirable primers. In contrast with these methods, the problems arisen from restriction digestion are circumvented totally in FSR. This feature would provide FSR with good flexibility in the construction of recombinant HBV DNA.

Furthermore, with its high efficiency of homologous fragment substitution, FSR provides a simple way of cloning clinical HBV isolates. When a plasmid template already contains HBV 1.1×, two rounds of FSR can generate a plasmid containing an HBV 1.1× derived completely from serum HBV DNA with two appropriate PCR fragments; e.g., one fragment nt 1,818–3,215–1,654 (from the start of the minus strand to a certain position downstream from the start of the plus strand) for the first FSR and the other fragment nt 1,654–1,988 (amplified using the plus strand as the template) for the second FSR. If a region of HBV DNA, such as a reverse transcriptase region, is to be investigated, one round of FSR would be sufficient to produce a plasmid containing that region derived from a serum HBV DNA, and all that must be done is to select a pair of primers whose sequences are identical in SF and TF. Although there is no appropriate restriction site, the HBV DNA fragments from the serum HBV can be cloned in a simple way and do not interrupt any of the elements or open reading frames necessary for HBV replication in vitro. This method is especially useful in the analysis of HBV drug resistance, where an assay for the effect of reverse transcriptase mutations on the replication capacity and drug susceptibility of HBV are most often conducted. To the best of our knowledge, this is the first study to report methods resembling FSR to generate constructs that can be used in an HBV susceptibility assay.

A potential disadvantage of this method is that artificial mutations can occur during FSR, because Taq DNA polymerase is used in the reaction. However, this disadvantage is limited for two reasons. One is that the FSR amplifies DNA in a linear manner, not in an exponential manner like traditional PCR. Therefore, the entire effective FSR product is synthesized using the initial plasmid as the template, unlike a PCR product, which is mainly composed of DNA amplified from newly synthesized template molecules after several cycles. Therefore, even if a mutation occurs in a certain cycle of FSR, the mutation will not accumulate in subsequent cycles, and the end product of FSR may still be made up of the correct molecule. The other reason is that the TF itself, as the mega-primer, need not be synthesized in FSR. When the TF is long enough, the major part of the newly synthesized molecule is the backbone of the plasmid template. Lethal mutations in the backbone of the plasmid will be excluded through transformation, and the mutations retained in the backbone have only a minor probability of significantly affecting the subsequent experiment. For the reason of sensitivity, the initial DNA fragments for FSR have to be obtained by PCR. At present, potential artificial mutations during PCR cannot be avoided completely. An effective way to decrease artificial mutations is to use high-fidelity DNA polymerase in PCR.

rtS246T Does not Confer Significant ETV Resistance in vitro

Two characteristics of the dynamics of the serum HBV DNA load of our patient prompted the investigation of the viruses in the patient's serum. First, his serum HBV load was never reduced to a level <10⁴ copies/ml during 156 weeks of therapy (Fig. 1), representing a partial virological response to ETV, which is uncommon in nucleot(s)ide-naïve patients. Second, his viral load increased by more than one log₁₀ during therapy (Fig. 1), whereas no typical ETV-resistant mutant was detected. One hypothesis to explain these characteristics is that the HBV strains infecting this patient may be primarily resistant to ETV to some extent. Sequence analysis of the HBV reverse transcriptase at different time points identified a variant at amino acid 246, at which position most genotype B HBV strains analysed have a serine, rather than the threonine observed in the strain studied here. However, a susceptibility test of the patient's HBV DNA indicated that the strain containing the reverse transcriptase sequence was not resistant to ETV. The two strains yielded IC₅₀ values of 0.99 ± 0.09 and 0.58 ± 0.10 nM. These values are close to the ETV IC₅₀ (1 nM) of the wild-type HBV reported by Levine et al. [2002] with a pCMV–HBV plasmid, and nearly fall within the IC₅₀ range (0.6–10.3 nM) calculated when several decades of wild-type strains of different genotypes were tested [Baldick et al., 2008]. The susceptibility data suggest that the S246T variant does not confer resistance to ETV, and the primary resistance of HBV strains to ETV does not explain the characteristics of the viral load dynamics observed in our patient very well. The rtS246C substitution has been observed in another nucleot(s)ide-naïve ETV-treated patient [Colonno et al., 2006]. However, a phenotype assay of HBV containing amino acid 246C found no significant resistance to ETV. These data, together with the fact that rtS246N does appear in some wild-type genotype F and H strains, although not in the genotype B strains assayed (data not shown), suggest that position rt246 could be a polymorphic site unrelated to anti-viral resistance.

Another possible explanation for the partial response observed in our patient is that the patient did not administer ETV as prescribed. The concentrations of ETV in the patient's serum at different time points, were not determined, so the possibility of a poor compliance with physician-recommended treatment cannot be ruled out.