Hep3B Human Hepatoma Cells Support Replication of the Wild-Type and a 5′-End Deletion Mutant GB Virus B Replicon

Abstract

Hepatitis C virus (HCV) and GB virus B (GBV-B) replicons have been reported to replicate only in Huh7 cells. Here we demonstrate that subpopulations of another human hepatoma cell line, Hep3B, are permissive for the GBV-B replicon, showing different levels of enhancement of replication from those of the unselected parental cell population. Adaptive mutations are not required for replication of the GBV-B replicon in these cells, as already demonstrated for Huh7 cells. Nonetheless, we identified a mutant replicon in one of the selected cell lines, which, although lacking the 5′ end proximal stem-loop, is able to replicate in Hep3B cells as well as in Huh7 cells. This mutant indeed shows a higher replication efficiency than does wild-type replicon, especially in the Hep3B cell clone from which it was originally recovered. This indicates that the stem-loop Ia is not necessary for replication of the GBV-B replicon in human cells, unlike what occurs with HCV, and that its absence can even provide a selective advantage.

The simian flavivirus GB virus B (GBV-B) is in principle a useful surrogate model for human hepatitis C virus (HCV) (3): the genome organization and enzymatic activities are similar in the two viruses (6, 7, 16, 19-23), and GBV-B offers the advantage of infecting primates more suitable for research than the chimpanzee, which is the only host for HCV (3, 5). A valuable achievement for drug discovery purposes would be the construction of viable chimeric HCV/GBV-B viruses with the host range of GBV-B. An intermediate step, important to further validate the use of GBV-B as a model for HCV and as a scaffold for chimeric constructs, is the development of comparable cell-based systems for GBV-B and for HCV.

We have recently described a subgenomic dicistronic GBV-B replicon system (9, 18) working in Huh7-derived human hepatoma cell lines and similar to that available for HCV (2, 4, 14), which allows analysis of the replication and the synthesis/processing of the viral nonstructural proteins. So far no successful attempt at identifying cells different from Huh7 cells as the host for HCV or GBV-B has been described (1, 14). Nonetheless, the availability of more than one host cell type would be important to study the molecular features underlying cell permissiveness for these viruses and as an additional tool in the drug discovery process.

In this paper we describe the identification of cells permissive to GBV-B replicons derived from the human hepatoma Hep3B cell line and the spontaneous occurrence of an interesting adapted mutant replicon lacking the stem-loop structure at the immediate 5′ terminus.

MATERIALS AND METHODS

Cell lines and culture conditions.

Human hepatoma cell lines HepG2, Hep3B, Huh7 (our laboratory stocks), and derived cell lines (9) were grown in high-glucose Dulbecco's modified Eagle's medium (Life Technologies) supplemented with 2 mM l-glutamine, 100 U of penicillin per ml, 100 μg of streptomycin per ml, and 10% fetal bovine serum. Cells were subcultivated twice a week with a 1:5 split ratio. Cells transfected with neomycin-resistant constructs were selected and used, whenever appropriate, in the presence of 0.250 mg of G418 per ml. Nonetheless they were tested for resistance up to 1 mg of G418 per ml.

Plasmids.

The neo-RepB plasmid (9) encodes a subgenomic dicistronic GBV-B replicon bearing the neomycin phosphotransferase (neo) gene, which confers resistance to G418. The 4-29Δ-neo-RepB mutant construct was obtained by replacing the BamHI-AscI fragment of neo-RepB with an equivalent fragment produced by primer-based mutagenesis with a sense primer spanning the deletion of nucleotides 4 to 29 identified in the resident replicon molecules of the cell line B6/Hep3B, one of the Hep3B-derived cell lines identified in this study. The GBV-B wild-type replicon bearing the β-lactamase reporter gene was described previously (9, 18), and the corresponding nucleotide-4-to-29 deletion mutant was constructed with a similar strategy. HCV replicon plasmids were kindly provided by Giacomo Paonessa and Giovanni Migliaccio.

Sequence analysis.

Sequencing was performed as described previously (9). Sequencing of the 5′ end of the GBV-B replicon lacking nucleotides 4 to 29, extracted from B6/Hep3B cells, was performed on the reverse transcription-PCR product obtained with the 5′-RACE kit (Invitrogen) following the manufacturer's instructions.

Transfection of GBV-B replicon RNA and monitoring of replication.

Human hepatoma HepG2, Hep3B, and Huh7 cells as well as lines derived from Huh7 cells and Hep3B cells were used to test the replication of GBV-B replicon constructs. Linearized plasmids encoding replicons were in vitro transcribed by T7 RNA polymerase as described previously (9). Confluent cells from 15-cm-diameter plates were split 1:2. Cells were recovered after 24 h in 5 ml of medium, washed twice with 40 ml of cold diethyl pyrocarbonate-treated phosphate-buffered saline, filtered with Cell Strainer filters (Falcon), and diluted in cold diethyl pyrocarbonate-treated phosphate-buffered saline at a concentration of 10⁷ cells/ml. Then 2 × 10⁶ cell aliquots were subjected to electroporation with 10 μg of in vitro-transcribed RNA by two pulses at 0.35 kV and 10 μF with a Bio-Rad Genepulser II. Immediately after the electric pulses, cells were diluted in 8 ml of complete Dulbecco's modified Eagle's medium.

In the experiments to select G418-resistant clones, the cells were divided into three plates 15 cm in diameter, and on the following day, the selecting antibiotic G418 (Sigma G-9516) was added at 0.250 mg/ml. After about 4 weeks, surviving cell clones could be picked and expanded by growing them in individual plates. When quantitative PCR was used to measure transient replication after transfection, 1 × 10⁵ to 2 × 10⁵ cells/well were plated in six-well plates. After 3 days, total RNA was purified with the RNeasy kit (Qiagen) following the manufacturer's instructions, and 10 μl out of 100 μl of total RNA recovered was used in each TaqMan reaction.

Test of putative inhibitors of GBV-B replication.

Cell lines carrying GBV-B replicons were used to test the effect of putative inhibitors on replication. To test the effect of 2′-C-methyladenosine, 10⁵ cells/well were plated into a series of wells of six-multiwell plates (Falcon) in the absence of G418. After 16 h the medium was discarded, and increasing concentrations of the test compound in fresh medium were added to each series of wells. Controls were run with the specific compound solvent at the same concentration used to test the compound (0.25% dimethyl sulfoxide). Cells were grown up to 3 days in the presence of compound or solvent, avoiding cell confluence, and finally lysed to extract total RNA with RNeasy as described in the kit protocol (Qiagen); 10 out of 100 μl of total RNA was used in each TaqMan reaction.

TaqMan analysis was performed as described below with intracellular 18S rRNA as an endogenous reference. Monitoring the amount of mRNA coding for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used to exclude cytotoxicity effects. In the experiments with alpha interferon (IFN-α), the protocol was the same as described above, with the following differences: 1× culture medium was used as the specific solvent control; the ABI-Prism 6700 nucleic acid workstation was used for automated RNA purification and TaqMan reaction assembly, following the manufacturer's instructions, and starting from 5 × 10³ cells plated in individual wells of a 96-well plates. Human alpha-2b interferon (Intron A3) was purchased from Schering-Plough, and 2′-C-methyladenosine was kindly provided by Anne Eldrup and Marija Prhavc (ISIS Pharmaceuticals) and David B. Olsen (Merck Research Laboratories).

TaqMan quantification of GBV-B RNA.

GBV-B RNA was quantified by a real-time 5′ exonuclease PCR (TaqMan) assay with a primer/probe set that recognized a portion of the GBV-B 5′ untranslated region (5′UTR). The primers GBV-B-F3 (GTAGGCGGCGGGACTCAT) and GBV-B-R3 (TCAGGGCCATCCAAGTCAA) and probe GBV-B-P3 (6-carboxyfluorescein-TCGCGTGATGACAAGCGCCAAG-N,N,N′,N′-tetramethyl-6-carboxyrhodamine) were selected with the Primer Express software (PE Applied Biosystems). The primers were used in a 10 pmol/50-μl reaction, and the probe was used in a 5 pmol/50-μl reaction. The reactions were performed with One-step PCR Master mix (Applied Biosystems) and included a 30-min reverse transcription step at 48°C, followed by 10 min at 95°C and by 40 cycles of amplification with the universal TaqMan standardized conditions (15 s at 95°C denaturation step followed by 1 min at 60°C annealing/extension step). RNA was quantified by absorbance at 260 nm and stored at −80°C. All reactions were run in duplicate by using the ABI Prism 7900 Sequence Detection System (PE Applied Biosystems). Whenever necessary, RNA extracted from cells transfected with RNA of replication-defective Rep-GAA mutant constructs was used as the calibrator. Results from two independent experiments were analyzed with the comparative Ct method.

In vitro translation.

In vitro-transcribed RNA (1 μg), prepared as described above, was translated with the rabbit reticulocyte lysate system (Promega kit) by incubation for 1 h at 30°C under the conditions suggested by the manufacturer in a 30-μl final volume. [³⁵S]methionine (Promix; Amersham Pharmacia Biotech) was incorporated as the radioactive tracer. Aliquots of the in vitro translation reaction were analyzed by sodium dodecyl sulfate-10% polyacrylamide gel electrophoresis (SDS-10% PAGE) followed by treatment of the gel with Amplify (Amersham Pharmacia Biotech) and X-ray film exposure. The gels were also scanned with a Storm 820 PhosphorImager (Molecular Dynamics), and a densitometric analysis of the radioactive bands was obtained with the Image Quant software.

Nucleotide sequence accession number.

The nucleotide sequence of the GBV-B neo-RepB construct is available in the DDBJ/EMBL/GenBank nucleotide sequence database under accession number AJ428955.

RESULTS AND DISCUSSION

Identification of Hep3B-derived cell lines sustaining replication of GBV-B neo-RepB replicon RNA.

In the initial characterization of the GBV-B replicon system (9), we screened various cell lines for the ability to replicate selectable GBV-B replicon RNAs but could not identify any permissive line besides Huh7 cells. The human hepatoma cell line HepG2 was particularly unsuitable for this system because of its spontaneous tendency to form loose clusters of cells that might be confused with arising resistant clones. Hep3B cells, instead, formed very tiny and slowly growing colonies surviving selection in a background of slowly dying sensitive cells, which at first we did not identify as positive cell clones. After acquiring a deeper experience with the experimental system, we could confirm that the HepG2 cell line is unable to support GBV-B replication, whereas we could demonstrate that Hep3B cells are an acceptable host, as detailed below.

We transfected 10 μg of in vitro-transcribed GBV-B neo-RepB replicon RNA in 2 × 10⁶ Hep3B cells by electroporation and applied G418 selection; 30 antibiotic-resistant colonies could be identified, and 11 colonies were expanded as individual cell lines to be further characterized. TaqMan analysis, performed on RNA from three of these cell lines (B9/Hep3B cells, B10/Hep3B cells, and B11/Hep3B cells) at a growth phase far from confluence to avoid replication inhibition, demonstrated the presence of neo-RepB RNA (not shown). The specific RNA amount detected accounted for a number of GBV-B genomes per μg of total RNA, 2.6 × 10⁴, 1.3 × 10⁵, and 3.7 × 10⁵, respectively.

Effect of inhibitors of neo-RepB replication in neo-RepB/Hep3B cell lines.

GBV-B replicon-positive B11/Hep3B cells were tested for susceptibility to the human α-IFN currently used as therapy for hepatitis C. The compound, already proven effective against the GBV-B replicon in Huh7 cells (B76/Huh7 cells) (9), was tested with the same protocol in Hep3B cells (B11/Hep3B cells). In this case the GBV-B neo-RepB RNA replication was inhibited by IFN-α with an apparent higher potency (50% effective concentration [EC₅₀] of about 0.2 U/ml), as shown in Fig. 1, with respect to the activity in Huh7 host cells, where the EC₅₀ is around 1 U/ml (9).

FIG. 1.

Effect of IFN-α on the replication of GBV-B neo-RepB replicon RNA in Hep3B-derived replicon cell line B11/Hep3B. Replicon RNA amount is reported as a percentage of the values obtained in the absence of test compounds. Data fitting and EC₅₀ calculation were carried out with the use of Kaleidagraph.

B11/Hep3B cells was also tested for susceptibility to the nucleoside 2′-C-methyladenosine, known to be active against the HCV replicon in Huh7 cells and, as a triphosphate, against the HCV NS5B polymerase in vitro (8). The compound, tested in parallel against the GBV-B replicon in Hep3B cells (B11/Hep3B cells) and in Huh7 cells (B76/Huh7 cells), inhibited GBV-B neo-RepB RNA replication in both host cells (Fig. 2) with a potency comparable to that shown against HCV replicon in Huh7 cells (EC₅₀ of about 100 nM).

FIG. 2.

Effect of 2′-C-methyladenosine on the replication of GBV-B neo-RepB replicon RNA in the Hep3B-derived replicon cell line B11/Hep3B (A) and in Huh7-derived cell line B76.1/Huh7 (B). Replicon RNA amount is reported as a percentage of the values obtained in the absence of test compounds. Data fitting and EC₅₀ calculation were carried out with the use of Kaleidagraph.

IFN-α is a protein producing an antiviral effect through a complex mechanism of cascade events, including interactions among cellular macromolecular components. Hence, the different environments of Hep3B cells and Huh7 cells might account for the observed difference in potency between the two cell lines. On the contrary, unless differences in transport or metabolism do occur in the different cell types, a difference in potency is not necessarily expected, and actually was not observed, for a low-molecular-weight inhibitor as the 2′-C-methyladenosine directly interacting with a viral component. Nonetheless, in a screening to identify novel inhibitors, it might be useful to examine more than one clone derived from a certain cell line to minimize the possibility of misinterpretation of results.

Neo-RepB/Hep3B cells show different levels of permissiveness.

The Hep3B-derived cell lines were then cured with α-IFN, as described previously (9), to eliminate resident GBV-B sequences and ascertain whether the derived cured lines were enhanced, compared to the Hep3B parental cell line, for the ability to sustain replication of GBV-B. TaqMan analysis confirmed the efficacy of the curing procedure (not shown). Cured cell lines were then retransfected, in parallel with parental Hep3B cell population, with the in vitro-synthesized GBV-B neo-RepB replicon RNA. G418-based selection was performed to evaluate the number of resistant clones in a long-term experiment.

In this selection experiment we could appreciate an increase in the overall colony number in the cured cells compared to the unselected Hep3B population by visual inspection upon staining with crystal violet (not shown), but a precise count of the colony number was not feasible due to the peculiar colony phenotype. The efficiency of transient replication was nonetheless analyzed by TaqMan at 3 days posttransfection. In Table 1 TaqMan results are shown, in which RNA amounts are reported as the ratio to the background levels obtained after transfecting the neo-RepB-GAA replication-defective mutant. This experiment demonstrated quantitatively that the cured cell lines were all enhanced for the GBV-B replicon with respect to the unselected Hep3B cell population, although at different levels. No direct correlation was observed between the copy number of the GBV-B replicon in a certain cell clone before the cure and the permissiveness level of that clone after the cure. This was not surprising, according to the results obtained with both HCV (17) and GBV-B (C. Traboni, unpublished data) in Huh7 cells. All parental and cured cell lines were also challenged with HCV replicon RNA, both the wild-type Con1 sequence (14) and a version containing the A227T cell-adaptive mutation in NS5A (G. Paonessa, unpublished data). Unfortunately, none of those cells proved to be a host for HCV RNA.

TABLE 1.

neo-RepB GBV-B replicon RNA in Hep3B and derived cell lines

Cell line	RNAa	Cured cell line	RNA increaseb (fold)
B9/Hep3B	2.6 × 104	Cured-B9/Hep3B	13.6
B11/Hep3B	1.3 × 105	Cured-B11/Hep3B	5.4
B10/Hep3B	3.7 × 105	Cured-B10/Hep3B	NT
B6/Hep3B	NT	Cured-B6/Hep3B	2.7

^aResident replicon RNA in stable cell lines is expressed as genome equivalents/μg of total RNA; NT, not tested.

^bReplicon RNA detected at 4 days posttransfection is expressed as increase relative to transfection of the neo-RepB-GAA replication-defective control. In unselected Hep3B cells, the increase was 1.6-fold.

Replicon resident in B6/Hep3B cells shows an unusual mutation.

GBV-B replicon RNA was extracted and fully sequenced from three cell lines representative of different categories, B6/Hep3B cells, B9/Hep3B cells, and B11/Hep3B cells, whose corresponding cured lines showed a low, medium, and high enhanced phenotype, respectively, upon retransfection with the wild-type GBV-B replicon. The results are summarized in Table 2. Only silent mutations were found in the replicons from the two more enhanced cell lines, suggesting that the observed increased permissiveness was actually due to features of the specific host cell clones. We considered the third cell line, B6/Hep3B cells, supporting replication of wild-type replicon almost at the same level than unselected population, as the best candidate to harbor adaptive replicon mutants.

TABLE 2.

Mutations in neo-RepB GBV-B replicon rescued from Hep3B-derived cell lines

Cell line	Nucleotides affected	Genea	Amino acid
6B/Hep3B	4-29Δ	5′ UTR	NAb
9B/Hep3B	A insertion at 1876-1880 A run	EMCV IRES	NA
	A2780G	NS3 (hel)c	Ala296, silent
11B/Hep3B	T1827G	EMCV IRES	NA
	T2375A	NS3 (pro)d	Ser161, silent

^aEMCV, encephalomyocarditis virus.

^bNA, not applicable.

^chel, helicase domain.

^dpro, protease domain.

Surprisingly, sequencing the replicon from that line revealed a 26-nucleotide deletion located at 5′ end, spanning nucleotides 4 to 29. According to the predicted secondary structure of the GBV-B 5′UTR (19), the region of nucleotides 4 to 29 corresponds to the Ia stem-loop plus the segment between the Ia and Ib stem-loops (see Fig. 3). We performed a computer-based secondary-structure prediction to check if a structure similar to that eliminated by the deletion of nucleotides 4 to 29 might have been created in the mutant RNA replicon molecule. According to this prediction, the deletion of the 4-to-29 sequence only produces the neat removal of the Ia stem-loop structure plus the downstream linear region without modifying the adjacent Ib stem-loop at all (see Fig. 3). This result is independent of the length of the genome fragment analyzed (data not shown) and suggests that the Ia stem-loop is not functionally replaced by a new structure. The residual stem-loop Ib seems quite different from both the Ia structure of wild-type GBV-B and stem-loop I of HCV (see Fig. 3).

FIG. 3.

Secondary structure prediction of 5′-end RNA of HCV, GBV-B, and the GBV-B 4-29Δ mutant. The Mfold program of the GCG sequence analysis package (Accelrys, Inc.) was used to derive the structures shown. HCV and GBV-B portions starting at nucleotide 1 and ending between nucleotides 60 and 500 were analyzed. Part of each of the three derived structures is shown, which includes stem-loops Ia and Ib and excludes the sequence structure from stem-loop II onwards.

4-29Δ mutation does not significantly affect IRES-dependent translation.

A mutant replicon molecule including the deletion of nucleotides 4 to 29 (4-29Δneo-RepB) identified in the GBV-B replicon RNA extracted from B6/Hep3B cells was constructed by means of primer-based mutagenesis.

We compared the efficiency of internal ribosome entry site (IRES)-mediated translation from wild-type neo-RepB and 4-29Δneo-RepB RNAs by performing an in vitro time course experiment with rabbit reticulocyte extract and fractionating the translation products on SDS-PAGE (Fig. 4). A slight increase in the GBV-B IRES-specific translation product (neomycin phosphotransferase protein) was detectable by visual inspection in samples translated from the mutant RNA compared to the wild-type RNA. A ratio of 1.4 to 1.5 times the protein amount from the mutant versus the wild-type construct was confirmed by densitometric analysis of the signals corresponding to the neomycin phosphotransferase protein normalized to the internal control, GBV-B NS3 protein, whose translation from the replicon RNA is dependent on the encephalomyocarditis virus IRES. However, the error intrinsic to the indirect read-out of the experiment suggests that weak enhancement may also not be really significant.

FIG. 4.

Effect of deletion 4-29Δ on GBV-B IRES-mediated translation in vitro. Autoradiogram of SDS-10% PAGE showing the analysis of in vitro-translated 4-29Δ-neoRepB (lanes 1 to 4) and wild-type neo-RepB RNA (lanes 5 to 8). Lanes 1 and 5 represent not-incubated in vitro reactions (time zero); lanes 2 to 8 are reactions incubated at 30°C for 20 min (2 and 6), 40 min (lanes 3 and 7), and 60 min (lanes 4 and 8). The neomycin phosphotransferase protein (NPT), translated under GBV-B IRES control, and the GBV-B NS3 protein internal control, depending on the encephalomyocarditis virus IRES located in the second cistron of the replicon molecule, are indicated.

The intracellular IRES activity of the wild-type and mutant constructs was checked by the β-lactamase reporter system, already used in the experiment with the wild-type replicon in Huh7 cells (9, 18). Unfortunately, Hep3B cells are not stained with this system, and we only got results with Huh7 cells and cB76/Huh7 cells (9), which did not show significant differences in the IRES-dependent translation between the wild-type and mutant replicons at 4 and 24 h posttransfection (not shown).

The effect of an experimental deletion spanning nucleotides 1 to 21 of the GBV-B 5′UTR, including the Ia step-loop, in in vitro translation of the chloramphenicol acetyltransferase (CAT) reporter gene was reported to produce an increase in translation efficiency similar to our results (19). In conclusion, we can say that if the deletion of nucleotides 4 to 29 in the GBV-B 5′UTR may not increase the translation efficiency; nonetheless, it does not decrease it, confirming that the region involved in the deletion is not part of the IRES (19).

Also in the case of HCV, data available in literature on the effect of IRES-mediated translation, derived from in vitro experiments performed with different reporters and extract sources, consistently indicated that the deletion of nucleotides 5 to 20 in HCV does not significantly affect the efficiency of the in vitro reaction (10, 13, 15) and that intracellular IRES-dependent translation was only moderately reduced compared to that in the wild type (10).

4-29Δ mutation is compatible with replication of neo-RepB in both Hep3B cells and Huh7 cells and confers an advantage versus the wild type.

The RNA in vitro transcribed from the 4-29Δneo-RepB mutant replicon plasmid was also tested for replication efficiency in the cured cell line cB6/Hep3B and in parental Hep3B cells (Fig. 5) in parallel with the wild-type neo-RepB replicon. Unselected Huh7 cells and derived lines selected as permissive to wild-type neo-RepB (9) and to genomic GBV-B replicon (C. Traboni, unpublished data), that is, the cB76.1/Huh7 and cfl2/Huh7 lines, respectively, were also compared to the Hep3B-derived cells as hosts for the mutant and wild-type replicons. The results of TaqMan analysis at 4 days posttransfection (Fig. 5) indicate that replication occurs in all cell types tested, confirming that this spontaneous mutation is compatible with replication in these lines. Control samples run in the presence of IFN-α showed almost complete suppression of replication. The effect of IFN-α on all the samples confirms that the higher level of intracellular replicon RNA detected in the cells transfected with the mutant construct depends on the efficiency of replication and not on higher stability of the corresponding RNA.

FIG. 5.

Effect of deletion 4-29Δ on the replication of neo-RepB in different cell lines. Intracellular wild-type neo-RepB RNA (light grey bars) and mutant 4-29Δ-neoRepB RNA (dark grey bars) were measured 4 days after transfection by TaqMan. Data are expressed as genome equivalents (G.E.) per microgram of total RNA.

Interestingly, in both unselected Huh7 and Hep3B cell populations, the mutant RNA replicated better than the wild-type molecule, suggesting that the mutation is adaptive. In the Huh7-derived cured cell lines cB76.1/Huh7 and cfl2/Huh7, generated from cells originally hosting nondeleted replicons, the 4-29Δneo-RepB mutant did not perform better than the wild-type replicon. Moreover, the highest gain in replication with respect to the wild-type replicon (25-fold) occurred in the cured cell line cB6/Hep3B, derived from that originally bearing the deleted replicon (B6/Hep3B). This seems to indicate that, besides the advantage conferred by the mutation in the cell types challenged, in the specific B6/Hep3B cell clone, unknown host factors may also play some role in the improvement of replication efficiency.

These observations raise interest in studying the interactions between viral or host factors and the 5′-terminal stem-loop in the genome of GBV-B and similar viruses. Unfortunately, limited data are available for HCV in the literature (12), whereas no information has been published for GBV-B, making it difficult to elaborate a hypothesis. Moreover, no spontaneous mutation similar to that identified by us had ever been found before in either the HCV or GBV-B replicon system. Nonetheless, the finding that the 4-29Δneo-RepB deletion mutant is able to replicate and is even improved with respect to the wild type was unexpected on the basis of the results obtained with similar deletions experimentally introduced in HCV replicons, where the removal of the Ia stem-loop, independently of the presence or absence of the first five nucleotides, actually abolishes replication (10, 15).

On the other hand, it is also interesting that in the case of bovine viral diarrhea virus (BVDV), it was reported that the 5′-proximal stem-loop is absolutely required for replication (as well as for translation) (24), but data obtained with chimeras between BVDV and HCV point in a different direction (11). In that study, replacement of the Ia stem-loop of BVDV, starting with genome nucleotide 1, with that of HCV was not compatible with productive infection in cultured cells susceptible to wild-type BVDV. Viable pseudorevertants could however be identified in which the initial BVDV genome 3- to 4-nucleotide sequence was restored, suggesting that the very first nucleotides of the positive-stranded genome RNA (corresponding to the very last nucleotides of the negative-stranded form) play a crucial role in some replication step of that virus.

Interestingly, in our spontaneous mutant, the deletion of nucleotides 4 to 29 actually leaves in place the first three nucleotides of the GBV-B genome, which might be indispensable for its replication, similarly to what was described for BVDV. It has to be considered that in the case of the GBV-B replicon, the host cells used are of human origin; that is, they derive from a species which has not been univocally proved to be a natural host for the virus, and hence we cannot exclude that the behavior of the mutant replicon in tamarin cells might be different. Experiments with intrahepatic injection of a full-length construct bearing the deletion of nucleotides 4 to 29 in tamarins or other GBV-B host species would also bear interest for evaluating the effect of the deletion in the context of the complete genome in an environment supporting all the steps of the viral life cycle.

Nonetheless, an issue might have been revealed by these data regarding the role of the proximal 5′ stem-loop in the replication process of HCV and GBV-B. Our occasional observation would deserve a broader experimental analysis in the perspective of validating the equivalence of key features of the two viruses, which are crucial to construct viable chimeras.

The replication of subgenomic replicons of GBV-B, which is a hepatotropic simian virus, has already been reported to occur in the human hepatoma cell line Huh7 with no need of mutations with respect to a genomic sequence infecting tamarins (9, 18). In summary, here we show that a second human hepatoma line, Hep3B, is permissive for GBV-B replication. Moreover, we identified a spontaneous deletion mutation in the 5′ untranslated region that unexpectedly allowed better replication than the wild-type sequence in these cells and in Huh7 cells, providing an interesting starting point to study the cis- and trans-acting elements important for GBV-B replication. A practical advantage of the availability of Hep3B cells as alternative host cells consists of the possibility of comparing the performance of the same replicon in two different cell environments, which might be useful in drug discovery programs. Unfortunately, the cell lines described in this paper, selected for permissiveness to the GBV-B replicon, do not also sustain HCV replication. However, the positive results obtained with GBV-B encourage more extensive direct efforts to identify cell lines other than Huh7 that are able to replicate the HCV genome.

ADDENDUM IN PROOF

After acceptance of the manuscript, Zhu et al. reported the replication of HCV subgenomes in nonhepatic human cells and mouse hepatoma cells (Q. Zhu, J.-T. Guo, and C. Seeger, J. Virol. 77:9204-9210, 2003).